Sleisenger and Fordtran's Gastrointestinal and Liver Disease, 9th Edition

Sleisenger and Fordtran’s

Ninth Edition

Gastrointestinal and Liver Disease

pathophysiology/diagnosis/management Mark Feldman, MD

William O. Tschumy Jr., MD, Chair of Internal Medicine Director, Internal Medicine Residency Program Medical Director, Research Services Texas Health Presbyterian Hospital Dallas Clinical Professor of Internal Medicine University of Texas Southwestern Medical School Dallas, Texas

Lawrence S. Friedman, MD Professor of Medicine Harvard Medical School Professor of Medicine Tufts University School of Medicine Chair, Department of Medicine Newton-Wellesley Hospital Assistant Chief of Medicine Massachusetts General Hospital Boston, Massachusetts

Lawrence J. Brandt, MD

Professor of Medicine and Surgery Albert Einstein College of Medicine Emeritus Chief, Division of Gastroenterology Montefiore Medical Center Bronx, New York

1600 John F. Kennedy Blvd. Ste 1800 Philadelphia, PA 19103-2899

SLEISENGER AND FORDTRAN’S GASTROINTESTINAL AND LIVER DISEASE: PATHOPHYSIOLOGY/ DIAGNOSIS/MANAGEMENT

ISBN: 978-1-4160-6189-2

Copyright © 2010, 2006, 2002, 1998, 1993, 1989, 1983, 1978, 1973 by Saunders, an imprint of Elsevier Inc. Chapters 32 and 106 are in the public domain. Mayo Clinic reserves all rights to original Mayo drawings in Chapter 120. All rights reserved.  No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permissions may be sought directly from Elsevier’s Rights Department: phone: (+1) 215 239 3804 (US) or (+44) 1865 843830 (UK); fax: (+44) 1865 853333; e-mail: [email protected]. You may also complete your request on-line via the Elsevier website at http://www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. With respect to any drug or pharmaceutical products identified, readers are advised to check the most current information provided (i) on procedures featured or (ii) by the manufacturer of each product to be administered, to verify the recommended dose or formula, the method and duration of administration, and contraindications. It is the responsibility of practitioners, relying on their own experience and knowledge of their patients, to make diagnoses, to determine dosages and the best treatment for each individual patient, and to take all appropriate safety precautions. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data Sleisenger and Fordtran’s gastrointestinal and liver disease : pathophysiology, diagnosis, management / [edited by] Mark Feldman, Lawrence S. Friedman, Lawrence J. Brandt.—9th ed.    p. ; cm.   Rev. ed. of: Sleisenger & Fordtran’s gastrointestinal and liver disease / edited by Mark Feldman, Lawrence S. Friedman, Lawrence J. Brandt. 8th ed. c2006.   Includes bibliographical references and index.   ISBN 978-1-4160-6189-2   1.  Gastrointestinal system—Diseases.  2.  Liver—Diseases.  I.  Sleisenger, Marvin H.  II.  Feldman, Mark, 1947-  III.  Friedman, Lawrence S. (Lawrence Samuel), 1953-  IV.  Brandt, Lawrence J.  V.  Sleisenger & Fordtran’s gastrointestinal and liver disease.  VI.  Title: Gastrointestinal and liver disease.   [DNLM: 1. Gastrointestinal Diseases. 2. Liver Diseases. WI 140 S632 2010]   RC801.G384 2010   616.3′3—dc22 2009014222 Acquisitions Editor: Druanne Martin Developmental Editor: Anne Snyder Publishing Services Manager: Frank Polizzano Project Manager: Jeff Gunning Project Management Assistance: Lee Ann Draud, Robin Hayward Design Direction: Steve Stave Printed in Canada Last digit is the print number:  9  8  7  6  5  4  3  2  1

Working together to grow libraries in developing countries www.elsevier.com | www.bookaid.org | www.sabre.org

This edition is dedicated to our grandchildren Noah and Jordan Feldman, Shayna and Olivia Feldgus, Christopher Friedman, Zachary and Noah Fishkind, and Chloe Jane King

Contributors Julian A. Abrams, MD

Assistant Professor of Clinical Medicine, Department of Medicine, Columbia University College of Physicians and Surgeons; Division of Digestive and Liver Diseases, Columbia University Medical Center; Cancer Epide­ miology Program, Herbert Irving Comprehensive Cancer Center, New York, New York Adenocarcinoma and Other Tumors of the Stomach

Nezam H. Afdhal, MD

Associate Professor of Medicine, Harvard Medical School; Attending Physician, Beth Israel Deaconess Medical Center, Boston, Massachusetts Gallstone Disease

Rakesh Aggarwal, MD, DM

Additional Professor, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India Hepatitis E

Karin L. Andersson, MD, MPH

Instructor in Medicine, Harvard Medical School; Staff Hepatologist, Gastrointestinal Unit, Massachusetts General Hospital, Boston, Massachusetts Acalculous Biliary Pain, Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis, and Polyps of the Gallbladder

Jane M. Andrews, MBBS, PhD, FRACP

Clinical Associate Professor, Department of Medicine, University of Adelaide Faculty of Health Sciences; Senior Consultant in Gastroenterology, Department of Gastroenterology and Hepatology, Royal Adelaide Hospital, Adelaide, South Australia, Australia Small Intestinal Motor and Sensory Function and Dysfunction

Paul Angulo, MD

Professor of Medicine and Section Chief, Hepatology, Division of Digestive Diseases and Nutrition, University of Kentucky Medical Center, Lexington, Kentucky Primary Biliary Cirrhosis

Fernando Azpiroz, MD, PhD

Professor of Medicine, Autonomous University of Barcelona; Chief, Department of Gastroenterology, University Hospital Vall d’Hebron, Barcelona, Spain Intestinal Gas

Bruce R. Bacon, MD

James F. King, MD, Endowed Chair in Gastroenterology and Professor of Internal Medicine, Saint Louis Univer­ sity School of Medicine; Director, Division of Gastro­ enterology and Hepatology, Saint Louis University Hospital, St. Louis, Missouri Hemochromatosis

Christina Wood Baker, PhD

Instructor in Psychiatry, Harvard Medical School; Depart­ ment of Psychiatry, Massachusetts General Hospital, Boston, Massachusetts Eating Disorders

William F. Balistreri, MD

Dorothy M. M. Kersten Professor of Pediatrics and Asso­ ciate Chair, Subspecialty Education, Department of Pediatrics, University of Cincinnati College of Medicine; Director Emeritus, Pediatric Liver Care Center, and Medical Director Emeritus, Liver Transplantation; Program Director, Fellowship in Transplant Hepatology, Division of Pediatric Gastroenterology, Hepatology, and Nutrition, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio Other Inherited Metabolic Disorders of the Liver

Todd H. Baron, MD

Professor of Medicine, Mayo Clinic College of Medicine; Consultant, Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota Endoscopic Treatment of Pancreatic Disease; Endoscopic and Radiologic Treatment of Biliary Disease

Bradley A. Barth, MD, MPH

Assistant Professor of Pediatrics, University of Texas Southwestern Medical School; Attending Physician, Pediatric Gastroenterology, Children’s Medical Center, Dallas, Texas Anatomy, Histology, Embryology, and Developmental Anomalies of the Pancreas

Anne E. Becker, MD, PhD, ScM

Associate Professor of Psychiatry and Associate Professor of Medical Anthropology, Harvard Medical School; Director, Eating Disorders Clinical and Research Program, Massachusetts General Hospital, Boston, Massachusetts Eating Disorders

Alex S. Befeler, MD

Associate Professor of Internal Medicine and Medical Director for Liver Transplantation, Saint Louis Univer­ sity School of Medicine, St. Louis, Missouri Tumors and Cysts of the Liver

Kfir Ben-David, MD

Assistant Professor, University of Florida College of Medicine; Chief of Minimally Invasive Gastroeso­ phageal and Bariatric Service, Shands at University of Florida; Director of Bariatric Surgery, Malcolm Randall VA Medical Center, Gainesville, Florida Appendicitis

L. Ashley Blackshaw, PhD

Affiliate Professor of Medicine, University of Adelaide Faculty of Health Sciences; Principal Research Fellow, Royal Adelaide Hospital, Adelaide, South Australia, Australia Small Intestinal Motor and Sensory Function and Dysfunction

Boris Blechacz, MD, PhD

Instructor in Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota Tumors of the Bile Ducts, Gallbladder, and Ampulla

vii

viii

Contributors Lawrence J. Brandt, MD

Professor of Medicine and Surgery, Albert Einstein College of Medicine; Emeritus Chief, Division of Gastroen­ terology, Montefiore Medical Center, Bronx, New York Vascular Lesions of the Gastrointestinal Tract; Intestinal Ischemia

George A. Bray, MD

Boyd Professor, Pennington Biomedical Research Center, Louisiana State University, Baton Rouge, Louisiana Obesity

Robert S. Bresalier, MD, FACP

Professor of Medicine, Birdie J. and Lydia J. Resoft Distin­ guished Professor in Gastrointestinal Oncology, and Director, Gastrointestinal Cancer Research Laboratory, Department of Gastroenterology, Hepatology, and Nutri­ tion, University of Texas M. D. Anderson Cancer Center, Houston, Texas Colorectal Cancer

Robert S. Britton, PhD

Associate Research Professor, Division of Gastroenterol­ ogy and Hepatology, Saint Louis University School of Medicine; Saint Louis University Hospital, St. Louis, Missouri Hemochromatosis

Simon J. Brookes, MD

Professor of Human Physiology, Department of Human Physiology and Center for Neuroscience, Flinders Uni­ versity School of Medicine, Adelaide, South Australia, Australia Colonic Motor and Sensory Function and Dysfunction

Alan L. Buchman, MD, MSPH

Professor of Medicine and Surgery and Medical Director of Intestinal Rehabilitation/Transplant Center, Division of Gastroenterology, Northwestern University Feinberg School of Medicine, Chicago, Illinois Short Bowel Syndrome

J. Steven Burdick, MD

Associate Professor of Medicine, University of Texas Southwestern Medical School; Staff Physician and Director of Endoscopy, Parkland Health and Hospital System; Staff Physician, Zale Lipshy University Hospi­ tal and St. Paul University Hospital, Dallas, Texas Anatomy, Histology, Embryology, and Developmental Anomalies of the Pancreas

Robert L. Carithers, Jr., MD

Professor of Medicine, University of Washington School of Medicine; Director, Liver Care Line, and Medical Director, Liver Transplant Program, University of Washington Medical Center, Seattle, Washington Alcoholic Liver Disease

Julie G. Champine, MD

Professor of Radiology, University of Texas Southwestern Medical School; Chief of Radiology, Parkland Health and Hospital System, Dallas, Texas Abdominal Abscesses and Gastrointestinal Fistulas

Francis K. L. Chan, MD

Professor of Medicine, Chinese University of Hong Kong; Chief of Gastroenterology and Hepatology, Department of Medicine and Therapeutics, Prince of Wales Hospital, Hong Kong, China Treatment of Peptic Ulcer Disease

Joseph G. Cheatham, MD

Instructor in Clinical Medicine, Uniformed Services Uni­ versity for Health Sciences, Bethesda, Maryland; Fellow in Gastroenterology, Walter Reed Army Medical Center, Washington, DC Hepatitis A

Shivakumar Chitturi, MD, FRACP

Senior Staff Specialist in Gastroenterology and Hepatol­ ogy, Canberra Hospital, Canberra, Australian Capital Territory, Australia Liver Disease Caused by Drugs

Daniel C. Chung, MD

Associate Professor of Medicine, Harvard Medical School; Director, Gastrointestinal Cancer Genetics Service, Massachusetts General Hospital, Boston, Massachusetts Cellular Growth and Neoplasia

Raymond T. Chung, MD

Associate Professor of Medicine, Harvard Medical School; Director of Hepatology and Medical Director, Liver Transplant Program, Massachusetts General Hospital, Boston, Massachusetts Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess

Robert R. Cima, MD

Associate Professor of Surgery, Mayo Clinic College of Medicine; Consultant in Colon and Rectal Surgery, Mayo Clinic, Rochester, Minnesota Ileostomy, Colostomy, and Pouches

Robert H. Collins, Jr., MD

Professor of Internal Medicine and Medical Oncology, University of Texas Southwestern Medical Center, Dallas, Texas Gastrointestinal Lymphomas

Ian J. Cook, MD, FRACP

Conjoint Professor of Medicine, Department of Medicine, University of New South Wales Faculty of Medicine; Senior Staff Specialist in Gastroenterology and Director, Gastroenterology Department, St. George Hospital, Sydney, New South Wales, Australia Colonic Motor and Sensory Function and Dysfunction

Diane W. Cox, PhD, FCCMG, FRSC

Professor of Medical Genetics, University of Alberta Faculty of Medicine, Edmonton, Alberta, Canada Wilson Disease

Sheila E. Crowe, MD

Professor of Medicine, Division of Gastroenterology and Hepatology, University of Virginia School of Medicine, Charlottesville, Virginia Helicobacter pylori

Albert J. Czaja, MD

Professor of Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota Autoimmune Hepatitis

Brian G. Czito, MD

Associate Professor of Radiation Oncology, Duke Univer­ sity School of Medicine, Durham, North Carolina Radiation Injury

Contributors Ananya Das, MD, DM, FASGE

Professor of Medicine, Mayo Clinic College of Medicine; Associate Chair and Director of Endoscopy, Department of Medicine, Mayo Clinic Arizona, Scottsdale, Arizona Tumors of the Esophagus

Fredric Daum, MD

Professor of Pediatrics and Clinical Scholar of Medicine, Stony Brook University Medical Center School of Medicine, Stony Brook; Chief, Division of Pediatric Gastroenterology, Hepatology, and Nutrition, Winthrop University Hospital, Mineola, New York Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine

Gary L. Davis, MD

Director, General and Transplant Hepatology, Baylor University Medical Center, Dallas, Texas Hepatitis C

Paul A. Dawson, PhD

Professor, Department of Internal Medicine, Section of Gastroenterology, Wake Forest University School of Medicine, Winston-Salem, North Carolina Bile Secretion and the Enterohepatic Circulation

Mark H. DeLegge, MD

Professor of Medicine, Medical University of South Carolina, Charleston, South Carolina Nutrition in Gastrointestinal Diseases

George D. Demetri, MD

Associate Professor of Medicine, Harvard Medical School; Director, Ludwig Center at Dana-Farber/Harvard Cancer Center, Boston, Massachusetts Gastrointestinal Stromal Tumors (GISTs)

Kenneth R. DeVault, MD

Professor of Medicine, Mayo Clinic College of Medicine; Chair, Division of Gastroenterology and Hepatology, Mayo Clinic, Jacksonville, Florida Symptoms of Esophageal Disease

Adrian M. Di Bisceglie, MD, FACP

Professor of Internal Medicine, Saint Louis University School of Medicine, St. Louis, Missouri Tumors and Cysts of the Liver

Philip G. Dinning, PhD

Research Fellow, Department of Medicine, University of New South Wales Faculty of Medicine, Sydney, New South Wales, Australia Colonic Motor and Sensory Function and Dysfunction

Iris Dotan, MD

Lecturer, Sackler School of Medicine; Head, IBD Center, Department of Gastroenterology and Liver Diseases, Sovrasky Medical Center, Tel Aviv, Israel Mucosal Immunity

Douglas A. Drossman, MD

Professor of Medicine and Psychiatry and Co-Director, UNC Center for Functional GI and Motility Disorders, Division of Gastroenterology and Hepatology, Univer­ sity of North Carolina School of Medicine; Attending Physician, UNC Hospitals, Chapel Hill, North Carolina Biopsychosocial Issues in Gastroenterology

David E. Elliott, MD, PhD

Professor and Director, Division of Gastroenterology and Hepatology, University of Iowa Carver College of Medicine, Iowa City, Iowa Intestinal Infections by Parasitic Worms

B. Joseph Elmunzer, MD

Clinical Lecturer in Internal Medicine, University of Michigan Medical School, Ann Arbor, Michigan Biliary Tract Motor Function and Dysfunction

Grace H. Elta, MD

Professor of Medicine, University of Michigan Medical School, Ann Arbor, Michigan Biliary Tract Motor Function and Dysfunction

Silvia Degli Esposti, MD

Associate Clinical Professor of Medicine and Director, Center for Gastrointestinal Services, Alpert Medical School of Brown University, Providence, Rhode Island Gastrointestinal and Hepatic Disorders in the Pregnant Patient

Michael B. Fallon, MD

Professor of Medicine, Dan and Lillie Sterling Professor of Gastroenterology, and Director, Division of Gastroenter­ ology, Hepatology, and Nutrition, University of Texas Health Science Center at Houston Medical School; Chief of Service, Gastroenterology and Hepatology, Memorial Hermann Medical Center, Houston, Texas Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome, and Systemic Complications of Liver Disease

Geoffrey C. Farrell, MD, FRACP

Professor of Hepatic Medicine, Australian National University Medical School; Gastroenterology and Hepatology Unit, Canberra Hospital, Canberra, Austra­ lian Capital Territory, Australia Liver Disease Caused by Drugs

James J. Farrell, MD

Associate Professor of Medicine, David Geffen School of Medicine at UCLA; Director, Pancreaticobiliary Endos­ copy, UCLA Medical Center, Los Angeles, California Digestion and Absorption of Nutrients and Vitamins

Richard J. Farrell, MD

Assistant Professor of Medicine, Harvard Medical School; Associate Physician, Gastroenterology Division, Beth Israel Deaconess Medical Center, Boston, Massachusetts Celiac Disease and Refractory Celiac Disease

Jordan J. Feld, MD, MPH

Assistant Professor of Medicine, University of Toronto Faculty of Medicine; Hepatologist, University Health Network, Toronto Western Hospital, Toronto, Ontario, Canada Hepatitis Caused by Other Viruses

Mark Feldman, MD

William O. Tschumy Jr., MD, Chair of Internal Medicine, Director of Internal Medicine Residency Program, and Medical Director of Research Services, Texas Health Presbyterian Hospital Dallas; Clinical Professor of Inter­ nal Medicine, University of Texas Southwestern Medical School, Dallas, Texas Gastritis and Gastropathies

ix

x

Contributors Carlos Fernández-del Castillo, MD

Associate Professor of Surgery, Harvard Medical School; Director, Pancreas and Biliary Surgery Program, Massachusetts General Hospital, Boston, Massachusetts Tumors of the Pancreas

Lincoln E. Ferreira, MD, PhD

Director, Digestive Endoscopy Unit, University Hospital, Universidade Federal de Juiz de Fora, Juiz de Fora, Brazil Endoscopic Treatment of Pancreatic Disease

Paul Feuerstadt, MD

Gastroenterology Fellow, Albert Einstein College of Medi­ cine; Montefiore Medical Center, Bronx, New York Intestinal Ischemia

Robert J. Fontana, MD

Associate Professor of Medicine and Medical Director of Liver Transplantation, University of Michigan Medical School, Ann Arbor, Michigan Acute Liver Failure

Chris E. Forsmark, MD

Professor of Medicine and Chief, Division of Gastroenter­ ology, Hepatology, and Nutrition, University of Florida College of Medicine, Gainesville, Florida Chronic Pancreatitis

Jeffrey M. Fox, MD, MPH

Assistant Clinical Professor of Medicine, Division of Gas­ troenterology, University of California, San Francisco, School of Medicine, San Francisco; Chief, Department of Gastroenterology, Kaiser Permanente, San Rafael, California Diverticular Disease of the Colon

Amy E. Foxx-Orenstein, DO

Associate Professor of Medicine, Mayo Clinic College of Medicine; Consultant, Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota Ileus and Pseudo-obstruction

Frank K. Friedenberg, MD

Professor of Medicine, Temple University School of Medi­ cine; Attending Physician, Temple University Hospital, Philadelphia, Pennsylvania Gastroesophageal Reflux Disease

Lawrence S. Friedman, MD

Professor of Medicine, Harvard Medical School; Professor of Medicine, Tufts University School of Medicine, Boston; Chair, Department of Medicine, NewtonWellesley Hospital, Newton; Assistant Chief of Medicine, Massachusetts General Hospital, Boston, Massachusetts Chronic Abdominal Pain; Acalculous Biliary Pain, Acalculous Cho­ lecystitis, Cholesterolosis, Adenomyomatosis, and Polyps of the Gallbladder

Ralph A. Gianella, MD

Mark Brown Professor of Medicine, University of Cincin­ nati College of Medicine; Attending Physician, Univer­ sity Hospital, Cincinnati, Ohio Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning

Gregory G. Ginsberg, MD

Professor of Medicine, University of Pennsylvania School of Medicine; Director of Endoscopic Services, Gastroen­ terology Division, University of Pennsylvania Health System, Philadelphia, Pennsylvania Foreign Bodies, Bezoars, and Caustic Ingestions

Robert E. Glasgow, MD

Associate Professor, Department of Surgery, University of Utah School of Medicine, Salt Lake City, Utah Treatment of Gallstone Disease

Gregory J. Gores, MD

Professor of Medicine, Mayo Clinic College of Medicine; Chair, Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota Tumors of the Bile Ducts, Gallbladder, and Ampulla

David A. Greenwald, MD

Associate Professor of Clinical Medicine, Albert Einstein College of Medicine; Associate Division Director, Gastroenterology, and Fellowship Program Director, Montefiore Medical Center, Bronx, New York Protein-Losing Gastroenteropathy

Heinz F. Hammer, MD

Associate Professor of Internal Medicine and Gastroenter­ ology, Medical University Graz; Attending Gastroenter­ ologist, Division of Gastroenterology and Hepatology, Medical University Graz, Graz, Austria Maldigestion and Malabsorption

William V. Harford, Jr., MD

Professor of Internal Medicine, University of Texas South­ western Medical School; Director, Gastrointestinal Endoscopy, and Staff Physician, Veterans Affairs Medical Center, Dallas, Texas Diverticula of the Pharynx, Esophagus, Stomach, and Small Intestine; Abdominal Hernias and Gastric Volvulus

David J. Hass, MD

Clinical Instructor, Yale University School of Medicine; Clinical Faculty, Yale-New Haven Hospital, New Haven, Connecticut Complementary and Alternative Medicine

E. Jenny Heathcote, MB BS, MD, FRCP, FRCP(C)

Professor of Medicine, University of Toronto Faculty of Medicine; Hepatologist, University Health Network, Toronto Western Hospital, Toronto, Ontario, Canada Hepatitis Caused by Other Viruses

Maureen Heldmann, MD

Associate Professor of Radiology, Louisiana State University Health Sciences Center School of Medicine Shreveport; Director, Body CT and MRI, Louisiana State University Health Sciences Center, Shreveport, Louisiana Intestinal Obstruction

Christoph Högenauer, MD

Associate Professor of Internal Medicine, Medical University Graz; Attending Gastroenterologist, Division of Gastroenterology and Hepatology, Medical University Graz, Graz, Austria Maldigestion and Malabsorption

Christopher D. Huston, MD

Associate Professor, Department of Medicine and Depart­ ment of Microbiology and Molecular Genetics, Univer­ sity of Vermont College of Medicine; Attending Physician, Fletcher Allen Health Care, Burlington, Vermont Intestinal Protozoa

Steven H. Itzkowitz, MD

Professor of Medicine and Director, GI Fellowship Program, Mount Sinai School of Medicine, New York, New York Colonic Polyps and Polyposis Syndromes

Contributors Rajeev Jain, MD

Clinical Assistant Professor of Medicine, University of Texas Southwestern Medical School; Chief of Gastroen­ terology, Texas Health Presbyterian Hospital Dallas, Dallas, Texas Gastrointestinal and Hepatic Manifestations of Systemic Diseases

Dennis M. Jensen, MD

Professor of Medicine, David Geffen School of Medicine at UCLA and Veterans Affairs Greater Los Angeles Healthcare System; Staff Physician, UCLA Medical Center and West Los Angeles Veterans Affairs Medical Center, Los Angeles, California Gastrointestinal Bleeding

Robert T. Jensen, MD

Chief, Cell Biology Section, Digestive Diseases Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland Endocrine Tumors of the Pancreas and Gastrointestinal Tract

D. Rohan Jeyarajah, MD, FACS

Director of Surgical Oncology and Hepatobiliary Fellow­ ship Program, Methodist Dallas Medical Center, Dallas, Texas Diverticula of the Pharynx, Esophagus, Stomach, and Small Intestine; Abdominal Hernias and Gastric Volvulus

Ramon E. Jimenez, MD

Assistant Professor of Surgery, University of Connecticut Medical School, Farmington; Associate Attending Phy­ sician, Hartford Hospital, Hartford, Connecticut Tumors of the Pancreas

Ellen Kahn, MD

Professor of Pathology and Pediatrics, New York Univer­ sity School of Medicine, New York; Director of Labora­ tories, North Shore Gastroenterology Associates PC, Great Neck, New York Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine

Peter J. Kahrilas, MD

Gilbert H. Marqvardt Professor in Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois Esophageal Neuromuscular Function and Motility Disorders

Patrick S. Kamath, MD

Professor of Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota Portal Hypertension and Gastrointestinal Bleeding

David A. Katzka, MD

Professor of Medicine, University of Pennsylvania School of Medicine; Division of Gastroenterology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania Esophageal Disorders Caused by Medications, Trauma, and Infection

Jonathan D. Kaunitz, MD

Professor of Medicine, David Geffen School of Medicine at UCLA; Staff Physician, West Los Angeles Veterans Affairs Medical Center, Los Angeles, California Gastric Secretion

Ciarán P. Kelly, MD

Professor of Medicine, Harvard Medical School; Chief, Blumgart Internal Medicine Firm, and Director, Gastro­ enterology Fellowship Training, Beth Israel Deaconess Medical Center, Boston, Massachusetts Celiac Disease and Refractory Celiac Disease; Antibiotic-Associated Diarrhea, Pseudomembranous Enterocolitis, and Clostridium difficile-Associated Diarrhea and Colitis

Seema Khan, MB BS

Associate Professor of Pediatrics, Jefferson Medical College of Thomas Jefferson University, Philadelphia, Pennsyl­ vania; Pediatric Gastroenterologist, Nemours/Alfred I. duPont Hospital for Children, Wilmington, Delaware Eosinophilic Disorders of the Gastrointestinal Tract

Arthur Y. Kim, MD

Assistant Professor of Medicine, Harvard Medical School; Assistant in Medicine, Infectious Diseases Unit, Massachusetts General Hospital, Boston, Massachusetts Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess

Michael B. Kimmey, MD

Clinical Professor of Medicine, University of Washington School of Medicine, Seattle; President, Tacoma Diges­ tive Disease Center, Tacoma, Washington Complications of Gastrointestinal Endoscopy

Kenneth L. Koch, MD

Professor of Medicine and Chief, Section on Gastro­ enterology, and Director, Digestive Health Center, Wake Forest University School of Medicine, Winston-Salem, North Carolina Gastric Neuromuscular Function and Neuromuscular Disorders

Kris V. Kowdley, MD

Clinical Professor of Medicine and Director, Center for Liver Disease, Virginia Mason Medical Center, Seattle, Washington Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis

Krzysztof Krawczynski, MD, PhD

Distinguished Consultant and Team Leader, Centers for Disease Control and Prevention, Atlanta, Georgia Hepatitis E

Robert C. Kurtz, MD

Professor of Clinical Medicine, Weill Medical College of Cornell University; Chief, Gastroenterology-Nutrition Service, and Attending Physician, Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, New York Tumors of the Small Intestine

J. Thomas Lamont, MD

Professor of Medicine, Harvard Medical School; Chief of Gastroenterology, Beth Israel Deaconess Medical Center, Boston, Massachusetts Antibiotic-Associated Diarrhea, Pseudomembranous Enterocolitis, and Clostridium difficile-Associated Diarrhea and Colitis

Charles S. Landis, MD, PhD

Fellow in Transplant Hepatology, Washington, Seattle, Washington Vascular Lesions of the Gastrointestinal Tract

University

of

xi

xii

Contributors Anne M. Larson, MD, FACP

Associate Professor of Internal Medicine and Medical Director of Liver Transplant Program, University of Texas Southwestern Medical Center, Dallas, Texas Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation

James Y. W. Lau, MD

Professor of Surgery, Chinese University of Hong Kong; Director of Endoscopy, Endoscopy Centre, Prince of Wales Hospital, Hong Kong, China Treatment of Peptic Ulcer Disease

Edward L. Lee, MD

Professor and Chair, Department of Pathology, Howard University College of Medicine; Howard University Hospital, Washington, DC Gastritis and Gastropathies

Anthony J. Lembo, MD

Associate Professor of Medicine, Harvard Medical School; Physician, Beth Israel Deaconess Medical Center, Boston, Massachusetts Constipation

Mike A. Leonis, MD, PhD

Assistant Professor of Pediatrics, University of Cincinnati College of Medicine; Staff Physician, Division of Gastroenterology, Hepatology, and Nutrition, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio Other Inherited Metabolic Disorders of the Liver

Michael D. Levitt, MD

Professor of Medicine, University of Minnesota Medical School; Staff Physician, Minneapolis Veterans Affairs Medical Center, Minneapolis, Minnesota Intestinal Gas

James H. Lewis, MD, FACP, FACG

Professor of Medicine, Georgetown University School of Medicine; Director of Hepatology, Georgetown Univer­ sity Hospital, Washington, DC Liver Disease Caused by Anesthetics, Toxins, and Herbal Preparations

Hsiao C. Li, MD

Assistant Professor of Internal Medicine–Hepatology/ Oncology, University of Texas Southwestern Medical School, Dallas, Texas Gastrointestinal Lymphomas

Gary R. Lichtenstein, MD

Professor of Medicine, University of Pennsylvania School of Medicine; Director, Center for Inflammatory Bowel Diseases, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania Ulcerative Colitis

Rodger A. Liddle, MD

Professor of Medicine, Duke University School of Medicine; Chief, Division of Gastroenterology, Duke University Hospital, Durham, North Carolina Gastrointestinal Hormones and Neurotransmitters

Steven D. Lidofsky, MD, PhD

Professor of Medicine and Pharmacology, University of Vermont College of Medicine; Director of Hepatology, Fletcher Allen Health Care, Burlington, Vermont Jaundice

Keith D. Lindor, MD

Professor of Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota Primary Biliary Cirrhosis

Caroline Loeser, MD

Assistant Professor of Medicine, Section of Digestive Dis­ eases, Yale University School of Medicine; Attending Physician, Yale–New Haven Hospital, New Haven, Connecticut Ulcers of the Small and Large Intestine

John D. Long, MD

Associate Professor of Medicine, Wake Forest University School of Medicine; Director, GI Neuromuscular Disor­ ders Program, Wake Forest University Baptist Medical Center, Winston-Salem, North Carolina Anatomy, Histology, Embryology, and Developmental Anomalies of the Esophagus

Mark E. Lowe, MD, PhD

Professor of Pediatrics, University of Pittsburgh School of Medicine; Chief, Pediatric Gastroenterology, Hepatol­ ogy, and Nutrition, Children’s Hospital of Pittsburgh of UPMC, Pittsburgh, Pennsylvania Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood

Emmy Ludwig, MD

Assistant Professor of Clinical Medicine, Weill Medical College of Cornell University; Assistant Attending Phy­ sician, Gastroenterology-Nutrition Service, Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, New York Tumors of the Small Intestine

Matthias Maiwald, MD

Associate Professor of Medical Microbiology, Department of Microbiology and Infectious Diseases, Flinders Uni­ versity School of Medicine, Northern Territory Clinical School; Director of Microbiology, Department of Pathol­ ogy, Royal Darwin Hospital, Tiwi, Northern Territory, Australia Whipple’s Disease

Carolina Malagelada, MD

Research Associate, Autonomous University of Barcelona; Staff Physician, Digestive Diseases Department, Vall d’Hebron University Hospital, Barcelona, Spain Nausea and Vomiting

Juan-R. Malagelada, MD

Professor, Autonomous University of Barcelona; Chair, Digestive Diseases Department, Vall d’Hebron Univer­ sity Hospital, Barcelona, Spain Nausea and Vomiting

Peter W. Marcello, MD, FACS, FASCRS

Vice Chair, Department of Colon and Rectal Surgery, Lahey Clinic, Burlington, Massachusetts Diseases of the Anorectum

Lawrence A. Mark, MD, PhD

Assistant Professor of Dermatology and Charles W. Lewis Investigator, Indiana University School of Medicine; Wishard Dermatology Chief, Wishard Health Services, Indianapolis, Indiana Oral Diseases and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease

Contributors Paul Martin, MD

Professor of Medicine and Chief, Division of Hepatology, University of Miami Leonard M. Miller School of Medi­ cine, Miami, Florida Liver Transplantation

Joel B. Mason, MD

Professor of Medicine and Nutrition, Tufts University School of Medicine; Staff Physician, Divisions of Gas­ troenterology and Clinical Nutrition, Tufts University Medical Center, Boston, Massachusetts Nutritional Assessment and Management of the Malnourished Patient

Jeffrey B. Matthews, MD

Christian R. Holmes Professor and Chair, Department of Surgery, University of Cincinnati College of Medicine, Cincinnati, Ohio Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm

Lloyd Mayer, MD

Professor and Chair, Immunology Institute; Chief, Henry Janowitz Division of Gastroenterology; and Chief, Divi­ sion of Clinical Immunology, Mount Sinai School of Medicine, New York, New York Mucosal Immunity

Craig J. McClain, MD

Professor of Medicine, Professor of Pharmacology and Toxicology, and Associate Vice President for Transla­ tional Research, University of Louisville School of Med­ icine; Chief of Gastroenterology, Louisville Veterans Affairs Medical Center, Louisville, Kentucky Alcoholic Liver Disease

George B. McDonald, MD

Professor of Medicine, University of Washington School of Medicine; Member and Head, Gastroenterology/ Hepatology Section, Fred Hutchinson Cancer Research Center, Seattle, Washington Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation

Frederick H. Millham, MD

Associate Clinical Professor of Surgery, Harvard Medical School; Chair, Department of Surgery, NewtonWellesley Hospital, Newton, Massachusetts Acute Abdominal Pain

Joseph P. Minei, MD, MBA

Professor and Chair, Division of Burn, Trauma, and Critical Care, Department of Surgery, University of Texas Southwestern Medical Center; Surgeon-in-Chief, Parkland Health and Hospital System, Dallas, Texas Abdominal Abscesses and Gastrointestinal Fistulas

Ginat W. Mirowski, DMD, MD

Adjunct Associate Professor, Indiana University School of Dentistry, Indianapolis, Indiana Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease

Joseph Misdraji, MD

Assistant Professor of Pathology, Harvard Medical School; Assistant Pathologist, Massachusetts General Hospital, Boston, Massachusetts Embryology, Anatomy, Histology, and Developmental Anomalies of the Liver

John Morton, MD, MPH

Associate Professor of Surgery, Director of Bariatric Surgery and Surgical Quality, and Section Chief of Mini­ mally Invasive Surgery, Stanford University School of Medicine, Stanford, California Bariatric Surgery

Sean J. Mulvihill, MD

Professor and Chair, Department of Surgery, and Ross Anderson Presidential Endowed Chair in Surgery, Uni­ versity of Utah School of Medicine; Senior Director for Clinical Affairs, Huntsman Cancer Institute, University of Utah, Salt Lake City, Utah Treatment of Gallstone Disease

Moises Ilan Nevah, MD

Assistant Professor of Medicine, University of Texas Health Science Center at Houston Medical School, Houston, Texas Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome, and Systemic Complications of Liver Disease

Jeffrey A. Norton, MD

Professor of Surgery, Stanford University School of Medi­ cine; Chief, Surgical Oncology, Stanford University Medical Center, Stanford, California Endocrine Tumors of the Pancreas and Gastrointestinal Tract

Kjell Öberg, MD, PhD

Professor of Endocrine Oncology, Department of Medical Sciences, Uppsala University; Chair, Center of Excel­ lence Endocrine Tumors, Department of Endocrine Oncology, University Hospital, Uppsala, Sweden Gastrointestinal Carcinoid Tumors (Gastrointestinal Neuroendocrine Tumors) and the Carcinoid Syndrome

Jacqueline G. O’Leary, MD, MPH

Medical Director, Inpatient Liver and Transplant Unit, Baylor University Medical Center, Dallas, Texas Hepatitis C

Seamus O’Mahony, MD, FRCP

Senior Lecturer in Gastroenterology, University College Cork, National University of Ireland; Consultant Gastro­ enterologist, Cork University Hospital, Cork, Ireland Enteric Microbiota and Small Intestinal Bacterial Overgrowth

Susan R. Orenstein, MD

Professor Emerita, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania Eosinophilic Disorders of the Gastrointestinal Tract

Roy C. Orlando, MD

Mary Kay and Eugene Bozymski and Linda and William Heizer Distinguished Professor of Medicine and Adjunct Professor of Cell and Molecular Physiology, University of North Carolina School of Medicine, Chapel Hill, North Carolina Anatomy, Histology, Embryology, and Developmental Anomalies of the Esophagus

Mark T. Osterman, MD, MSCE

Assistant Professor of Medicine, University of Pennsylvania School of Medicine; Division of Gastro­ enterology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania Ulcerative Colitis

xiii

xiv

Contributors Stephen J. Pandol, MD

Professor of Medicine, David Geffen School of Medicine at UCLA; Staff Physician, Veterans Affairs Greater Los Angeles Healthcare System, Los Angeles, California Pancreatic Secretion

John E. Pandolfino, MD

Associate Professor of Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois Esophageal Neuromuscular Function and Motility Disorders

Abhitabh Patil, MD

Assistant Professor of Medicine, Rush University Medical Center, Chicago, Illinois Vascular Diseases of the Liver

John H. Pemberton, MD

Professor of Surgery, Mayo Clinic College of Medicine; Consultant in Colon and Rectal Surgery, Rochester, Minnesota Ileostomy, Colostomy, and Pouches

V. S. Periyakoil, MD

Director, Stanford University Hospice and Palliative Med­ icine Fellowship Program, Stanford University School of Medicine; Associate Director of Palliative Care Ser­ vices, Veterans Affairs Palo Alto Health Care System, Palo Alto, California Palliative Care for Patients with Gastrointestinal and Hepatic Disease

Robert Perrillo, MD

Associate Director, Hepatology Division, and Director, Transplant Hepatology Fellowship, Baylor University Medical Center, Dallas, Texas

B. S. Ramakrishna, MD, DM, PhD

Professor of Gastroenterology, Christian Medical College, Vellore, Tamil Nadu, India Tropical Diarrhea and Malabsorption

Mrinalini C. Rao, PhD

Professor and Vice President of Faculty Affairs, Depart­ ment of Physiology and Biophysics, University of Illi­ nois College of Medicine at Chicago, Chicago, Illinois Intestinal Electrolyte Absorption and Secretion

Satish S. C. Rao, MD, PhD, FRCP(Lon)

Professor of Medicine, University of Iowa Carver College of Medicine; Director, Neurogastroenterology and Gastrointestinal Motility, University of Iowa Hospitals and Clinics, Iowa City, Iowa Fecal Incontinence

Andrea E. Reid, MD, MPH

GI/Hepatology/Nutrition Section, Veteran Affairs Medical Center, Washington, DC Nonalcoholic Fatty Liver Disease

John F. Reinus, MD

Professor of Clinical Medicine, Albert Einstein College of Medicine; Chief of Clinical Hepatology, Montefiore Medical Center, Bronx, New York Gastrointestinal and Hepatic Disorders in the Pregnant Patient

David A. Relman, MD

Professor of Medicine and Professor of Microbiology and Immunology, Stanford University School of Medicine; Chief, Infectious Diseases Section, Veterans Affairs Palo Alto Health Care System, Palo Alto, California

Hepatitis B and D

Whipple’s Disease

David A. Peura, MD

Joel E. Richter, MD

Emeritus Professor of Medicine, Division of Gastroenterol­ ogy and Hepatology, University of Virginia School of Medicine, Charlottesville, Virginia Helicobacter pylori

Patrick R. Pfau, MD

Associate Professor of Medicine, Section of Gastroenterol­ ogy and Hepatology, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin Foreign Bodies, Bezoars, and Caustic Ingestions

Daniel K. Podolsky, MD

Professor of Internal Medicine, Doris and Bryan Wilden­ thal Distinguished Chair in Medical Science, and Presi­ dent, University of Texas Southwestern Medical Center, Dallas, Texas Cellular Growth and Neoplasia

Jonathan Potak, MD

Assistant Professor of Medicine, Mount Sinai School of Medicine, New York, New York Colonic Polyps and Polyposis Syndromes

Daniel S. Pratt, MD

Assistant Professor of Medicine, Harvard Medical School; Liver-Biliary-Pancreas Center, Massachusetts General Hospital, Boston, Massachusetts Liver Chemistry and Function Tests

Deborah Denise Proctor, MD

Professor of Medicine and Director, Inflammatory Bowel Disease Program, Section of Digestive Diseases, Yale University School of Medicine; Attending Physician, Yale-New Haven Hospital, New Haven, Connecticut Ulcers of the Small and Large Intestine

Richard L. Evans Chair and Professor, Department of Medicine, Temple University School of Medicine, Philadelphia, Pennsylvania Gastroesophageal Reflux Disease

Eve A. Roberts, MD, FRCPC

Adjunct Professor of Pediatrics, Medicine, and Pharmacol­ ogy, University of Toronto Faculty of Medicine; Associ­ ate, Division of Gastroenterology, Hepatology, and Nutrition, Hospital for Sick Children, Toronto, Ontario, Canada Wilson Disease

Hugo R. Rosen, MD, FACP

Waterman Endowed Chair in Liver Research, Professor of Medicine and Immunology, Division Head of Gastroen­ terology and Hepatology, and Program Director, Hepati­ tis C Research Center, University of Colorado Denver Health Sciences Center, Denver, Colorado Liver Transplantation

Andrew S. Ross, MD

Digestive Disease Institute, Virginia Mason Medical Center, Seattle, Washington Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis

Jayanta Roy-Chowdhury, MD

Professor of Medicine and Genetics, Albert Einstein College of Medicine, Bronx, New York Liver Physiology and Energy Metabolism

Namita Roy-Chowdhury, MD

Professor of Medicine and Genetics, Albert Einstein College of Medicine, Bronx, New York Liver Physiology and Energy Metabolism

Contributors Bruce A. Runyon, MD

Professor of Medicine, Division of Gastroenterology and Hepatology, Loma Linda University School of Medicine; Chief of Liver Service, Loma Linda University Medical Center, Loma Linda, California Ascites and Spontaneous Bacterial Peritonitis

Michael A. Russo, MD

Assistant Professor of Pediatrics, Division of Pediatric Gastroenterology, Hepatology, and Nutrition, University of Texas Southwestern Medical School, Dallas; Attend­ ing Physician, Children’s Medical Center of Dallas at Legacy, Plano, Texas Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum

Hugh A. Sampson, MD

Professor of Pediatrics and Immunology, Mount Sinai School of Medicine; Staff Physician, Mount Sinai Hos­ pital, New York, New York Food Allergies

Bruce E. Sands, MD

Associate Professor of Medicine, Harvard Medical School; Interim Chief, Gastrointestinal Unit, and Medical Co-Director, MGH Crohn’s and Colitis Center, Massachusetts General Hospital, Boston, Massachusetts Crohn’s Disease

George A. Sarosi, Jr., MD

Associate Professor of Surgery, Robert H. Hux, MD, Pro­ fessor of Surgery, and Surgical Residency Program Director, University of Florida College of Medicine; Assistant Chief, Surgical Service, North Florida/South Georgia Veterans Affairs Medical Center, Gainesville, Florida Appendicitis

Thomas J. Savides, MD

Professor of Clinical Medicine, University of California, San Diego, School of Medicine; Clinical Service Chief, Gastroenterology, University of California, San Diego, Medical Center, La Jolla, California Gastrointestinal Bleeding

Lawrence R. Schiller, MD

Clinical Professor of Internal Medicine, University of Texas Southwestern Medical School; Director, Gastro­ enterology Fellowship Program, and Attending Physi­ cian, Digestive Health Associates of Texas, Baylor University Medical Center, Dallas, Texas Diarrhea

Mitchell L. Schubert, MD

Professor of Medicine and Physiology, Virginia Common­ wealth University Health System; Chief of Gastro­ enterology, McGuire Veterans Affairs Medical Center, Richmond, Virginia Gastric Secretion

Joseph H. Sellin, MD

Professor of Medicine and Director, GI Fellowship Program, Baylor College of Medicine; Chief, Division of Gastroenterology, Ben Taub General Hospital, Houston, Texas Diarrhea; Intestinal Electrolyte Absorption and Secretion

M. Gaith Semrin, MD, MBBS

Assistant Professor of Pediatric Gastroenterology, Hepato­ logy, and Nutrition, University of Texas Southwestern Medical School; Attending Physician, University of Texas Southwestern Medical Center, Dallas, Texas Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum

Vijay H. Shah, MD

Professor of Medicine and Physiology, Mayo Clinic College of Medicine, Rochester, Minnesota Portal Hypertension and Gastrointestinal Bleeding

Fergus Shanahan, MD, FRCP(UK), FRCPI, FACP, FRCP(C)

Professor and Chair, Department of Medicine, University College Cork, National University of Ireland; Consultant Gastroenterologist, Cork University Hospital, Cork, Ireland Enteric Microbiota and Small Intestinal Bacterial Overgrowth

Corey A. Siegel, MD

Assistant Professor, Dartmouth Medical School, Hanover; Director, IBD Center, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire Crohn’s Disease

Maria H. Sjogren, MD, MPH, FACP

Associate Professor of Medicine, Georgetown University School of Medicine, Washington, DC; Associate Profes­ sor of Preventive Medicine, Uniformed Services Univer­ sity of the Health Sciences F. Edward Hébert School of Medicine, Bethesda, Maryland; Staff Physician, Department of Gastroenterology/Hepatology, Walter Reed Army Medical Center, Washington, DC Hepatitis A

Rhonda F. Souza, MD

Associate Professor of Medicine, University of Texas Southwestern Medical School; Staff Physician, Veterans Affairs North Texas Health Care System, Dallas, Texas Barrett’s Esophagus

Stuart Jon Spechler, MD

Professor of Medicine and Berta M. and Cecil O. Patterson Chair in Gastroenterology, University of Texas South­ western Medical Center; Chief, Division of Gastroenter­ ology, Veterans Affairs North Texas Health Care System, Dallas, Texas Barrett’s Esophagus

William M. Steinberg, MD

Clinical Professor of Medicine, George Washington University School of Medicine and Health Sciences, Washington, DC Acute Pancreatitis

William E. Stevens, MD

Clinical Faculty, Department of Internal Medicine, Division of Gastroenterology, Presbyterian Hospital of Dallas, Dallas, Texas Vascular Diseases of the Liver

Andrew H. Stockland, MD

Assistant Professor of Radiology, Mayo Clinic College of Medicine; Consultant, Diagnostic Radiology, Mayo Clinic, Rochester, Minnesota Endoscopic and Radiologic Treatment of Biliary Disease

xv

xvi

Contributors Neil H. Stollman, MD

Associate Clinical Professor of Medicine, Division of Gas­ troenterology, University of California, San Francisco, School of Medicine, San Francisco; Chief, Division of Gastroenterology, Alta Bates Summit Medical Center, Oakland, California Diverticular Disease of the Colon

Frederick J. Suchy, MD

Professor and Chair, Department of Pediatrics, Mount Sinai School of Medicine; Pediatrician-in-Chief, Kravis Children’s Hospital, New York, New York Anatomy, Histology, Embryology, Developmental Anomalies, and Pediatric Disorders of the Biliary Tract

Jan Tack, MD, PhD

Professor of Medicine and Chair, Department of Patho­ physiology, University of Leuven; Clinic Head, Division of Gastroenterology, University Hospital Leuven, Leuven Belgium Dyspepsia

Nicholas J. Talley, MD, PhD

Professor of Medicine, Mayo Clinic College of Medicine; Division of Gastroenterology and Hepatology, Mayo Clinic, Jacksonville, Florida Irritable Bowel Syndrome

Scott Tenner, MD, MPH

Associate Professor of Medicine, State University of New York Health Science Center at Brooklyn; Director, Medical Education and Research, Maimonides Medical Center, Brooklyn, New York Acute Pancreatitis

Associate Professor of Medicine, Australian National University Medical School; Senior Staff Specialist in Gastroenterology and Hepatology, Canberra Hospital, Canberra, Australian Capital Territory, Australia Liver Disease Caused by Drugs

Professor and Vice Chair, Department of Internal Medi­ cine, University of Texas Southwestern Medical Center, Dallas, Texas Gastrointestinal and Hepatic Manifestations of Systemic Diseases

Richard H. Turnage, MD

Professor and Chair, Department of Surgery, University of Arkansas for Medical Sciences; UAMS Medical Center, Little Rock, Arkansas Intestinal Obstruction

Hospital–

Nimish Vakil, MD

Clinical Professor of Medicine, University of Wisconsin School of Medicine and Public Health, Madison; Gastroenterologist, Aurora Health Care, Waukesha, Wisconsin Peptic Ulcer Disease

Visiting Research Assistant Professor, Department of Physiology and Biophysics, University of Illinois College of Medicine at Chicago, Chicago, Illinois Intestinal Electrolyte Absorption and Secretion

Arnold Wald, MD

Professor of Medicine, Section of Gastroenterology and Hepatology, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin Other Diseases of the Colon and Rectum

David Q.-H. Wang, MD, PhD

Assistant Professor of Medicine, Harvard Medical School; Gastroenterologist, Beth Israel Deaconess Medical Center, Boston, Massachusetts Gallstone Disease

Timothy C. Wang, MD

Dorothy L. and Daniel Silberberg Professor of Medicine, Columbia University College of Physicians and Sur­ geons; Chief, Digestive and Liver Diseases, Columbia University Medical Center, New York, New York Adenocarcinoma and Other Tumors of the Stomach

David C. Whitcomb, MD, PhD

Professor of Medicine, Cell Biology and Physiology, and Human Genetics, University of Pittsburgh School of Medicine, Pittsburgh; Chief, Division of Gastroenter­ ology, Hepatology, and Nutrition, University of Pitts­ burgh Medical Center, Allison Park, Pennsylvania Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood

Professor of Medicine, University of Alabama School of Medicine at Birmingham; Division of Gastroenterology and Hepatology, UAB Hospital, Birmingham, Alabama Gastrointestinal Consequences of Infection with Human Immunodeficiency Virus

Leonard Prosnitz Professor of Radiation Oncology, Duke University School of Medicine; Chief, Department of Radiation Oncology, Duke University Hospital, Durham, North Carolina Radiation Injury

Gavitt Woodard, BS

Research Coordinator, Stanford University School of Medicine, Stanford, California Bariatric Surgery

Stephan G. Wyers, MD

Constipation

Jayashree Venkatasubramanian, PhD

Whipple’s Disease

Christopher G. Willett, MD

Dwain L. Thiele, MD

Gastroenterologist, Beth Israel Deaconess Needham, Needham, Massachusetts

Professor and Consultant in Gastrointestinal Pathology, John Radcliffe Hospital, Oxford, United Kingdom

C. Mel Wilcox, MD

Narci C. Teoh, MBBS, PhD, FRACP

Sonal P. Ullman, MD

Axel von Herbay, MD

Assistant Professor of Surgery, University of Chicago Pritzker School of Medicine; General Surgery, Univer­ sity of Chicago Medical Center, Chicago, Illinois Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm

Joseph C. Yarze, MD, FACP, FACG, FASGE, AGAF

Consultant, Gastroenterology Associates of Northern New York; Medical Director, GI Center, Glens Falls Hospital, Glens Falls, New York Chronic Abdominal Pain

Foreword It is a pleasure to write the foreword for the 9th edition of Sleisenger and Fordtran’s Gastrointestinal and Liver Disease: Pathophysiology/Diagnosis/Management, the classic textbook that has maintained its leadership position and progressively enhanced its reputation as the “go-to” book in its field. The extraordinary effort required to create a textbook that comprehensively and authoritatively presents the state of knowledge in an arena as broad and everexpanding as gastroenterology is a daunting task but one that has been achieved with distinction in this 9th edition. The excellence and reliability exhibited by this considerably revised and updated edition amply validates the lasting value of the series. Almost 40 years ago, Drs. Marvin Sleisenger and John Fordtran gathered a group of highly respected, thoughtful, and clinically excellent gastroenterologists to create what became the first edition of Gastrointestinal and Liver Disease. The book was a success from the onset. Subsequent editions have continued to successfully portray the accelerating evolution of gastroenterology. Gastroenterologists, internists, surgeons, pathologists, residents, and students from diverse backgrounds have relied on Gastrointestinal and Liver Disease as a thoughtful, extensively referenced, and user-accessible textbook. The images are well chosen and superbly produced. Important landmarks are noted, evaluated, and placed into proper perspective. The 9th edition is ably edited by Drs. Mark Feldman, Lawrence Friedman, and Lawrence Brandt—all three highly respected clinicians and educators. This monumental undertaking continues a grand tradition and succeeds in providing authoritative overviews of the status of gastro­ enterology in the early 21st century. The editors are to be applauded for their efforts. The creation of a textbook requires considerable judgment in defining the current state of the art for a diverse range of disorders and in selecting authors best qualified to evaluate and lead the field. The editors have chosen the authors well for the 9th edition, with an admixture of established leaders and younger colleagues who represent the next generation that will advance gastroenterology. A textbook differs in important ways from a journal. In a textbook, the state of knowledge is presented at a given, and set, time. A line is drawn in the sand. A successful chapter in a textbook defines the state of knowledge regarding a subject, provides a clear definition of what is known,

outlines how the knowledge evolved, suggests what challenges may be ahead, and offers a framework of advice as to the most efficient and effective means of establishing the diagnosis and developing a plan for treatment. The goal is to provide a foundation and a guide that will serve us well in the varied situations we encounter in clinical practice. There surely have been outstanding advances in gastroenterology to incorporate into the 9th edition. Notable examples are the further emergence of advanced diagnostic and therapeutic endoscopy, introduction of liver and intestinal transplantation, development of more precise diagnostic genetic and serologic tests, and game-changing results derived from the advent of minimally invasive surgery. New innovative and effective therapeutic agents have been developed. These days, drugs are increasingly designed to solve a specific identified problem. Increasingly, therapeutic agents provide the ability to modulate inflammation, inhibit fibrosis, and regulate the processes that lead to cell death. In the 9th edition of Gastrointestinal and Liver Disease, we are provided a sturdy platform to aid us now and direct us toward the future. What lies ahead is likely to be exciting beyond our most ambitious speculations. Knowledge will continue to evolve. Innovative, often life-changing therapies will be discovered. New challenges will emerge that must be faced. As time goes by, we all will undoubtedly rely even more on electronically transmitted information (including the online version of this textbook, in which the references will be updated regularly). The role for a comprehensive, respected medical textbook, however, is far from over. We salute those who wrote and those who edited Gastrointestinal and Liver Disease. Through their efforts we are provided with renewed confidence that tangible progress is being achieved on many fronts. An effective textbook reminds us of the past, firmly grounds us in the present, and guides our expectations and hopes for the future. Gastrointestinal and Liver Disease is such a book. We, the readers and our patients, will benefit from the guide provided. Willis C. Maddrey, MD University of Texas Southwestern Medical Center Dallas, Texas

xvii

Preface The 9th edition of Sleisenger and Fordtran’s Gastrointestinal and Liver Disease is the second edition for which the three of us have served together as editors and the first for which neither Dr. Fordtran nor Dr. Sleisenger has participated. We remain grateful to their inspiring examples as visionaries and superb editors and are pleased to present an edition that we believe is worthy of their high standards and commitment to excellence. This edition is the first for which a single purchase will provide all readers with both a copy of the book and access to the online version of the text; the latter will be updated regularly to keep readers abreast of new developments in the rapidly evolving field of gastroenterology and liver disease. Moreover, providing readers with online access has allowed us to reduce the bulk of the hard copy of the book by listing only 15 or so key references per chapter and referring the reader to the online version for the complete list of references cited in each chapter and linked to PubMed. We are confident that readers will appreciate the improved “portability” of the book. Like its predecessors, the 9th edition presents a critical overview of the state of gastrointestinal practice and its scientific basis by eminent authorities in their respective fields. With authors from at least 12 countries on 4 continents, the book is truly international in scope. We have been aided in our preparation of the 9th edition by the valuable feedback provided by our colleagues, trainees, reviewers, and readers; this has led to incremental refinements and enhancements since the highly acclaimed 8th edition, which received the first prize for a textbook in Gastroenterology from The British Medical Association in 2007. Attributes of the book, we believe, are its logical organization, minimized redundancies, polished writing, incorporation of color figures into the body of the text, inclusion of clear, visually appealing algorithms to summarize clinical decision making, and practical approaches to patient management. As always, we have tried to incorporate new, highquality scientific evidence as the basis of rational treatment, while alerting readers to emerging developments that hold promise for new approaches to patient management. To keep the book fresh and to ensure critical review of fields for which controversy exists, we have again rotated the authorship of many chapters, with approximately one third of the authors being new to this edition. We feel particularly fortunate to have enlisted contributors who are renowned authorities in their fields. Each has been chosen for his or her expertise and skills in communication, and it was a remarkable privilege for us to work with such accomplished colleagues. Reflecting recent trends in the practice of gastroenterology, three new chapters have been added to the 9th edition. First, a chapter by Drs. Woodard and Morton is devoted to Bariatric Surgery and complements the chapter on Obesity by Dr. Bray. This new chapter is an indication of the emerging role of the gastroenterologist not only in the selection of patients for bariatric surgery but also in the management of postsurgical complications and eventually in the endoscopic management of obesity. Second, a chapter on Barrett’s Esophagus by Drs. Spechler and Souza has been separated from the chapter on Gastroesophageal Reflux Disease by Drs. Richter and Freidenberg, again reflecting the

importance and controversies concerning the premalignant nature of Barrett’s esophagus and emerging approaches to its treatment. Third, a new chapter on Endoscopic Treatment of Pancreatic Disease complements the chapter on Endoscopic and Radiologic Treatment of Biliary Disease, both by Dr. Baron, with two different co-authors (Drs. Ferreira and Stockland, respectively), and reflects the expanding role of endoscopic interventions in the management of pancreatic disorders. Additionally, two chapters in the 8th edition have been merged into a single chapter in the 9th edition—Foreign Bodies, Bezoars, and Caustic Injury, by Drs. Pfau and Ginsberg. The organization of the 9th edition is otherwise similar to that of the 8th edition with the exception that the section on Nutrition in Gastroenterology (Section II) has now been placed, more logically we believe, before the section on Symptoms, Signs, and Biopsychosocial Issues (Section III). In Section I on Biology of the Gastrointestinal Tract, we are delighted to welcome Drs. Dotan and Mayer, who provide a comprehensive overview of Gastrointestinal Immunology and Inflammation. In Section III, we welcome Dr. Millham (Acute Abdominal Pain), Dr. Yarze (Chronic Abdominal Pain), Dr. Tack (Dyspepsia), Dr. Rao (Fecal Incontinence), and Drs. Savides and D. Jensen (Gastrointestinal Bleeding), all recognized experts in their respective fields. We are also happy to welcome back Dr. Drossman, who has provided a masterly discussion of Biopsychosocial Issues in Gastroenterology. Section IV deals with Topics Involving Multiple Organs and includes an expanded discussion of Eosinophilic Disorders of the Gastrointestinal Tract, with a section on eosinophilic esophagitis, by Drs. Khan and Orenstein, and an expanded chapter on Preparation for and Complications of Gastrointestinal Endoscopy by Dr. Kimmey. Among the many other noteworthy contributors in this section are Drs. Li and Collins (Gastrointestinal Lymphomas), Dr. Demetri (Gastrointestinal Stromal Tumors), Dr. Oberg (Carcinoid Tumors), Drs. R. Jensen and Norton (Pancreatic and Gastrointestinal Endocrine Tumors), Drs. Larson and McDonald (Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation), and Drs. Willett and Czito (Radiation Injury). The next six sections cover the Esophagus, Stomach and Duodenum, Pancreas, Biliary Tract, Liver, and Small and Large Intestine. Among the new authors are Drs. Kahrilas and Pandolfino (Esophageal Motor and Sensory Function and Motor Disorders), Drs. Kaunitz and Schubert (Gastric Secretion), Drs. Crowe and Peura (Helicobacter pylori), Dr. Vakil (Peptic Ulcer Disease), Drs. Wang and Afdhal (Gallstone Disease), Dr. Andersson (Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis, and Gallbladder Polyps), Drs. Blechacz and Gores (Tumors of the Bile Ducts, Gallbladder, and Ampulla), Dr. Misdraji (Anatomy, Histology, Embryology, and Developmental Anomalies of the Liver), Dr. Pratt (Liver Chemistry and Function Tests), Drs. O’Leary and Davis (Hepatitis C), Drs. Nevah and Fallon (Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome, and Systemic Complications of Liver Disease), Drs. Belefer and Di Bisceglie (Hepatic Tumors and Cysts), Dr. Ramakrishna (Tropical Malabsorption and

xix

xx

Preface Tropical Diarrhea), Dr. Foxx-Orenstein (Ileus and Pseudoobstruction), Drs. Ludwig-Miller and Kurtz (Small Intestinal Neoplasms), and Dr. Marcello (Diseases of the Anorectum). The book concludes with an important section on Palliative, Complementary, and Alternative Medicine. The 9th edition of this book remains a living tribute to its founding editors, Drs. Marvin Sleisenger and John Fordtran, and the current editors are proud to strive to uphold the tradition of excellence established by our predecessors. We are particularly grateful to Druanne Martin, Anne Snyder, and Jeffrey Gunning and their colleagues at Elsevier for their

commitment to this classic book, which has been a labor of love for the editors. We believe that the 9th edition of Sleisenger and Fordtran’s Gastrointestinal and Liver Disease remains the ideal resource for all who seek excellence in the care of patients with gastrointestinal and liver disorders. Mark Feldman, MD Lawrence S. Friedman, MD Lawrence J. Brandt, MD

CHAPTE R

1

Gastrointestinal Hormones and Neurotransmitters Rodger A. Liddle

CHAPTER OUTLINE Cellular Communication  3 Neural Regulation of the GI Tract  5 Peptide Hormones of the GI Tract  6 Synthesis, Post-Translational Modification, and Secretion  6 Gastrin  6 Cholecystokinin  7 Secretin  7 Vasoactive Intestinal Polypeptide  8 Glucagon  8 Glucose-Dependent Insulinotropic Polypeptide  8 Pancreatic Polypeptide Family  9 Substance P and the Tachykinins  9 Somatostatin  9 Motilin  9 Leptin  10 Ghrelin  10 Other Chemical Messengers of the Gastrointestinal Tract  10 Acetylcholine  10 Catecholamines  11 Dopamine  11 Serotonin  11 Histamine  11

Cells throughout the gastrointestinal (GI) tract receive information in many forms, including chemical messengers that emanate from other cells. The initial stimulus for hormone secretion is the ingestion of food. Food provides central neural stimulation in the form of thought (anticipation) and sight, chemical stimulation in the form of odor and taste, nutrient stimulation of the epithelial cells lining the GI tract, and mechanical stimulation. These processes all stimulate the release of peptides and other transmitters from cells of the mucosa into the nearby space, where they act locally, or into the bloodstream, where they circulate to distant target tissues. Therefore, chemical messengers from the GI tract can have far-reaching effects throughout the body.

CELLULAR COMMUNICATION Chemical transmitters of the gut are produced by discrete cells of the GI mucosa and can be classified as endocrine, paracrine, synaptic (“neurocrine”), or autocrine (Fig. 1-1). Specialized signaling cells that secrete transmitters into the

Nitric Oxide  12 Adenosine  12 Cytokines  13 Signal Transduction  13 G Protein–Coupled Receptors  13 G Proteins  13 Receptors Not Coupled to G Proteins  14 Hormone and Transmitter Regulation of Gastrointestinal Growth  15 Growth Factor Receptors  15 Epidermal Growth Factor  16 Transforming Growth Factor-α  16 Transforming Growth Factor-β  16 Insulin-Like Growth Factors  16 Fibroblast Growth Factor and Platelet-Derived Growth Factor  16 Trefoil Factors  16 Other G Protein–Coupled Receptors  16 Taste Receptors  17 Intraluminal Releasing Factor Regulation of Gastrointestinal Hormones  17 Gastrointestinal Peptides That Regulate Satiety and Hunger  18 Enteroinsular Axis  18

blood are known as endocrine cells, and the transmitters they produce are known as hormones. Hormones bind to specific receptors on the surface of target cells at remote sites and regulate metabolic processes.1 In contrast with endocrine cells that act on distant target tissues, other signaling cells of the GI tract may produce transmitters that act on neighboring cells. This process is known as paracrine signaling and is typical of cells that produce somatostatin.2 Paracrine transmitters are secreted locally and cannot diffuse far. They bind to receptors on nearby cells to exert their biological actions. These actions are limited because they are taken up rapidly by their target cells, destroyed by extracellular enzymes, and adhere to extracellular matrix, all of which limit their ability to act at distant sites. Because paracrine signals act locally, their onset of action is generally rapid and can be terminated abruptly. By comparison, endocrine signaling takes much longer, and termination of signaling requires clearance of hormone from the circulation. A third form of signaling in the GI tract is neurotransmission. The enteric nervous system is a complex and sophisticated array of nerve cells and ganglia that is intimately involved in all aspects of GI function. When neurons of the

3

4

Section I  Biology of the Gastrointestinal Tract Table 1-1  Hormones and Transmitters of the Gastrointestinal Tract

Endocrine

Autocrine

Paracrine

Neurocrine

Figure 1-1.  Examples of cell-to-cell communication by chemical transmitters in the gastrointestinal tract. Hormones are secreted from endocrine cells into the blood, where they are carried to distant targets. Paracrine cells secrete transmitters into the paracellular space and act locally. Neurons secrete chemical transmitters or peptides into synapses or onto other cell types. Autocrine transmitters bind to receptors on the cell from which they originate.

GI tract are activated, signals in the form of neurotransmitters are released from the nerve terminals. These synapses deliver neurotransmitters to nerves, muscle cells, epithelial and secretory cells, and other specialized cells of the GI tract. Neurotransmitters are critical for the processes of digestion including the coordination of gut motility and secretion. Many of the same transmitters are produced by endocrine, paracrine, and neural cells. For example, cholecystokinin (CCK) is produced by typical endocrine cells of the upper small intestine and is secreted into the bloodstream on ingestion of a meal. However, CCK is also abundant in nerves of the GI tract and brain. In neural tissue, CCK functions as a neurotransmitter, although when secreted into the blood, CCK is a classic GI hormone. This conservation of transmitters allows the same messenger to have different physiologic actions at different locations and is made possible by the manner in which the transmitter is delivered to its target tissues. Endocrine cells secrete many different hormones into the blood, and their actions depend on the specificity of the receptor on the target tissues. In contrast, in synaptic transmission, the variety of neurotransmitters is more limited, and the specificity of action is dependent on the precise location at which the nerves synapse with the target cells. The concentration of signaling molecules can be adjusted quickly because the transmitter can be rapidly metabolized. In the synaptic cleft, transmitters are either rapidly destroyed or taken back up by the secretory neuron. Concentrations of these transmitters can be regulated rapidly by changes in their rate of synthesis, secretion, or catabolism. Many peptide transmitters have extremely short halflives (generally on the order of minutes), which allows the rapid initiation and termination of signaling. Endocrine transmitters of the GI tract consist predominantly of peptides (e.g., gastrin, secretin). Paracrine transmitters can be peptides, such as somatostatin, or nonpeptides, such as histamine, that act locally on neighboring cells. Neurotransmitters can be peptides, such as vasoactive intestinal polypeptide (VIP) and tachykinins, or small molecules, such as acetylcholine and norepinephrine, that are secreted, or nitric oxide (NO), which simply diffuses across the syn-

Peptides That Function Mainly as Hormones Gastrin Glucose-dependent insulinotropic peptide (GIP) Glucagon and related gene products (GLP-1, GLP-2, glicentin, oxyntomodulin) Insulin Motilin Pancreatic polypeptide Peptide tyrosine tyrosine (PYY) Secretin Peptides That May Function as Hormones, Neuropeptides, or Paracrine Agents Cholecystokinin (CCK) Corticotropin-releasing factor (CRF) Endothelin Neurotensin Somatostatin Peptides That Act Principally as Neuropeptides Calcitonin gene-related peptide (CGRP) Dynorphin and related gene products Enkephalin and related gene products Galanin Gastrin-releasing peptide (GRP) Neuromedin U Neuropeptide Y Peptide histidine isoleucine (PHI) or peptide histidine methionine (PHM) Pituitary adenylate cyclase–activating peptide (PACAP) Substance P and other tachykinins (neurokinin A, neurokinin B) Thyrotropin-releasing hormone (TRH) Vasoactive intestinal peptide (VIP) Peptides That Act as Growth Factors Epidermal growth factor Fibroblast growth factor Insulin-like factors Nerve growth factor Platelet-derived growth factor Transforming growth factor-β Vascular endothelial growth factor Peptides That Act as Inflammatory Mediators Interferons Interleukins Lymphokines Monokines Tumor necrosis factor-α Peptides That Act on Neurons Cholecystokinin Gastrin Motilin Nonpeptide Transmitters Produced in the Gut Acetylcholine Adenosine triphosphate (ATP) Dopamine γ-Aminobutyric acid (GABA) Histamine 5-Hydroxytryptamine (5-HT, serotonin) Nitric oxide Norepinephrine Prostaglandins and other eicosanoids Newly Recognized Hormones or Neuropeptides Amylin Ghrelin Guanylin and uroguanylin Leptin

aptic cleft. The major transmitters and hormones of the GI tract are listed in Table 1-1. Criteria for establishing whether a candidate transmitter functions as a true hormone requires the following: (1) that the peptide be released into the circulation in response to

Chapter 1  Gastrointestinal Hormones and Neurotransmitters a physiologic stimulus; and (2) that the target tissue response can be reproduced by infusing the transmitter into the blood, thereby producing the same blood levels that occur physiologically. If an identical target tissue response is elicited, the hormonal effect of the transmitter has been proved. These criteria have been satisfied for a limited number of GI hormones, including gastrin, CCK, secretin, motilin, and glucose-dependent insulinotropic peptide (GIP). Somatostatin is the prototype of a paracrine transmitter. However, depending on its location, somatostatin may also exert endocrine and neural actions. For example, intestinal somatostatin is released into the local circulation following ingestion of fat and acts on the stomach as an enterogastrone to inhibit gastric acid secretion. Some cells release messengers locally and possess cell surface receptors for the same messengers, thus enabling those cells to respond to their own secreted products. This mode of transmission, known as autocrine signaling, has been demonstrated for several growth factors and has been implicated in the growth of certain cancers, including colorectal cancer (see Chapter 3).3

NEURAL REGULATION OF THE GI TRACT The enteric nervous system plays an integral role in the regulation of gut mucosal and motor function.4 It is organized into two major plexuses (Fig. 1-2). The myenteric plexus lies between the external longitudinal and internal circular muscle layers. The submucosal plexus lies between the circular muscle layer and the mucosa. Although the enteric nervous system receives input from the central and autonomic nervous systems, it can function independently. Nerves of the myenteric plexus project fibers primarily to the smooth muscle of the gut, with only a few axons extending to the submucosal plexus. Most of the fibers of the submucosal plexus project into the mucosa and the submucosal and myenteric plexuses. Various peptide and nonpeptide neurotransmitters are found in the enteric nervous system. Studies using immunohistochemical staining have

Serosa Circular muscle

Longitudinal muscle

Submucosa

Myenteric plexus

Submucosal plexus

Muscularis mucosa Mucosal nerves Mucosa Figure 1-2.  Organization of the enteric nervous system. The enteric nervous system is composed of two major plexuses, one submucosal and one located between the circular and longitudinal smooth muscle layers. These neurons receive and coordinate neural transmission from the GI tract and central nervous system.

localized neurotransmitters to specific neurons in the GI tract. γ-Aminobutyric acid is found primarily in the myenteric plexus and is involved in regulating smooth muscle contraction. Serotonin is found within the plexus and functions as an interneuron transmitter. Adrenergic neurons originate in ganglia of the autonomic nervous system and synapse with enteric neurons. Peptides such as neuropeptide Y (NPY) are often secreted from the same adrenergic neurons and generally exert inhibitory effects, such as vasoconstriction.5 Other adrenergic neurons containing somatostatin project to the submucosal plexus, where they inhibit intestinal secretion. Coexistence of peptides and neurotransmitters in the same neurons is not unusual; in fact, the interplay among transmitters is critical for coordinated neural regulation.6 For example, the peptides VIP and peptide histidine isoleucine (PHI) are commonly found together, as are the tachykinins substance P and substance K, where they have complementary effects. Somatostatin is found in interneurons that project caudally. The inhibitory action of somatostatin is consistent with a role in causing muscle relaxation in advance of a peristaltic wave. The abundance of VIP in the myenteric plexus also suggests that its inhibitory actions are important for smooth muscle relaxation in gut motility. VIP neurons that project from the submucosal plexus to the mucosa most likely stimulate intestinal fluid secretion. Other neurons that innervate the mucosa contain acetylcholine. Mucosal cells of the intestine contain receptors for both VIP and acetylcholine, allowing these transmitters to exert synergistic effects, because VIP increases intracellular cyclic adenosine monophosphate (cAMP) levels and acetylcholine increases intracellular calcium in the target cell. Bipolar neurons that project to the mucosa and myenteric plexus act as sensory neurons and often contain substance P, calcitonin gene-related peptide (CGRP), and acetylcholine as neurotransmitters. These neurons participate in pain pathways and modulate inflammation. The ability of hormones to act on nerves locally within the submucosa of the intestine and affect more distant sites on nerves such as the vagus expands the potential organs that may be regulated by gut hormones.7 Chemical and mechanical stimuli cause the release of hormones from endocrine cells of the intestinal mucosa. These interactions initiate a wide variety of secretomotor responses, many of which are mediated by enteric neurons. Secretomotor circuits consist of intrinsic primary afferent neurons with nerve endings in the mucosa and extension through the myenteric and submucosal plexi. This circuitry allows nerves to stimulate mucosal cells to secrete fluid and electrolytes and at the same time stimulate muscle contraction. The same motor neurons also have axons that supply arterioles and can initiate vasodilator reflexes. Extrinsic primary afferent neurons can be of the vagus, with somal bodies in the nodose ganglia and axons that reach the gut through the vagus nerve, or of the spinal nerves of the thoracic and lumbar regions, whose cell bodies lie in the dorsal root ganglia. Information conducted by extrinsic primary afferent neurons includes pain, heat, and sensations of fullness or emptiness. These neurons are also targets for hormones. For example, the satiety effect of CCK in the bloodstream is mediated through the vagus nerve.8 Specific CCK receptors have been identified on the vagus, and blockade of these receptors abolishes the satiation induced by peripheral CCK. Endocrine, paracrine, and neural transmitters existing within the lamina propria modulate effects on the gut immune system.7 Lymphocytes, macrophages, mast cells, neutrophils, and eosinophils are potential targets for endo-

5

6

Section I  Biology of the Gastrointestinal Tract crine and neural transmitters and participate in the inflammatory cascade. Moreover, inflammatory mediators can act directly on enteric nerves. Serotonin released from endocrine cells is involved in intestinal anaphylaxis and stimulates vagal afferent fibers that possess the 5-hydroxytryptamine 3 (5-HT3) receptor.

Gene 5′

Exon 1

intron

Exon 2

intron

Exon 3

3′

transcription mRNA

poly A

Cap site translation

PEPTIDE HORMONES OF THE GI TRACT SYNTHESIS, POST-TRANSLATIONAL MODIFICATION, AND SECRETION

The expression of peptides is regulated at the level of the gene that resides on defined regions of specific chromosomes. The genes for most of the known GI peptides have now been identified. Specific gene regulatory elements determine if and when a protein is produced and the particular cell in which it will be expressed. Gut hormone gene expression is generally linked to peptide production and regulated according to the physiologic needs of the organism. For example, the production of a hormone may increase when gut endocrine cells are stimulated by food, changes in intraluminal pH, exposure to releasing factors, or other transmitters or hormones. These factors may simultaneously stimulate hormone secretion and increase gene expression. Ultimately, hormones are secreted into the circulation, where they can bind to receptors on target tissues. Once a biological response is elicited, signals may then be sent back to the endocrine cell to “turn off” hormone secretion. This negative feedback mechanism is common to many physiologic systems and avoids excess production and secretion of hormone. All GI peptides are synthesized via gene transcription of DNA into messenger RNA (mRNA) and subsequent translation of mRNA into precursor proteins known as preprohormones. Peptides that are to be secreted contain a signal sequence that directs the newly translated protein to the endoplasmic reticulum, where the signal sequence is cleaved and the prepropeptide product is prepared for structural modifications.9 These precursors undergo intracellular processing and are transported to the Golgi apparatus and packaged in secretory granules. Further modifications in peptide structure may occur within the Golgi apparatus (e.g., sulfation) that is important for the bioactivity of many peptide hormones, such as CCK. Secretory granules may be targeted for immediate release or stored in close proximity to the plasma membrane for release following appropriate cell stimulation. When GI endocrine cells are stimulated, mature hormone is secreted into the paracellular space and is taken up into the bloodstream. For many hormones, such as gastrin and CCK, multiple molecular forms exist in blood and tissues. Although there is only a single gene for these peptides, the different molecular forms result from differences in pretranslational or post-translational processing (Fig. 1-3). A common mechanism of pretranslational proc­ essing includes alternative splicing of mRNA, which generates unique peptides from the same gene. Post-translational changes include cleavage of precursor molecules. Enzymatic cleavage of the signal peptide produces a prohormone. Other post-translational features that result in mature GI peptides include peptide cleavage to smaller forms (e.g., somatostatin), amidation of the carboxyl terminus (e.g., gastrin), and sulfation of tyrosine residues (e.g., CCK). These processing steps are usually critical for biolo­ gical activity of the hormone. For example, sulfated CCK is 100-fold more potent than its unsulfated form. The vast biochemical complexity of gastroenteropancreatic hor-

Prepropeptide

Signal spacer peptide Post-translational processing

Propeptide

Peptide AB Peptide A Figure 1-3.  Schematic representation of the production of gastrointestinal peptides. The genetic information is transcribed into mRNA, which is translated to a prepropeptide. Subsequent enzymatic cleavage produces peptides of various lengths. mRNA, messenger RNA.

mones is evident in the different tissues that secrete these peptides. As GI peptides are secreted from endocrine as well as nervous tissue, the distinct tissue involved often determines the processing steps for production of the peptide. Many hormone genes are capable of manufacturing alternatively spliced mRNAs or proteins that undergo different post-translational processing and ultimately produce hormones of different sizes. These modifications are important for receptor binding, signal transduction, and consequent cellular responses.10 It has become possible to express human genes in other species. By introducing specific hormone-producing genes into pigs or sheep, human hormones have been produced for medicinal use.11 With the rapid sequencing of the human genome, it is likely that novel methods of gene expression will expand the therapeutic use of human proteins. Moreover, drugs are being developed that inhibit the transcription of DNA into mRNA or that block the gene elements responsible for turning on specific hormone production (e.g., antisense oligonucleotides).12 This technology is based on the principle that nucleotide sequences bind to critical DNA regions and prevent transcription into mRNA. Similarly, oligonucleotides can be made to interact with mRNA and alter (or inhibit) translation of a protein product. These principles may be applicable to the treatment of the growing list of diseases that result from aberrant protein processing.13,14

GASTRIN

As discussed in more detail in Chapter 49, gastrin is the major hormone that stimulates gastric acid secretion. Subsequently, gastrin was found to have growth-promoting effects on the gastric mucosa and possibly some cancers.15 Human gastrin is the product of a single gene located on chromosome 17. The active hormone is generated from a precursor peptide called preprogastrin. Human preprogastrin contains 101 amino acids (AAs), including a signal peptide (21 AAs), spacer sequence (37 AAs), gastrin component (34 AAs), and a 9-AA extension at the carboxyl terminus. The enzymatic processing of preprogastrin

Chapter 1  Gastrointestinal Hormones and Neurotransmitters produces all the known physiologically active forms of gastrin. Preprogastrin is processed into progastrin and gastrin peptide fragments of various sizes by sequential enzymatic cleavage. The two major forms of gastrin are G34 and G17, although smaller forms exist. The common feature of all gastrins is an amidated tetrapeptide (Try-Met-Asp-Phe-NH2) carboxyl terminus, which imparts full biological activity. Modification by sulfation at tyrosine residues produces alternative gastrin forms of equal biological potency. A nonamidated form of gastrin known as glycine-extended gastrin is produced by colonic mucosa. Glycine-extended gastrin has been shown in animal models to stimulate proliferation of normal colonic mucosa and enhance the development of colorectal cancer. It is not known whether local production of this form of gastrin contributes to human colon carcinogenesis, and the receptor for glycine-extended gastrin has not been identified.16 Most gastrin is produced in endocrine cells of the gastric antrum.17 Much smaller amounts of gastrin are produced in other regions of the GI tract, including the proximal stomach, duodenum, jejunum, ileum, and pancreas. Gastrin has also been found outside the GI tract, including in the brain, adrenal gland, respiratory tract, and reproductive organs, although its biological role in these sites is unknown. The receptors for gastrin and CCK are related and constitute the so-called gastrin-CCK receptor family. The CCK-1 and CCK-2 (previously known as CCK-A and -B) receptor complementary DNAs were cloned from the pancreas and brain, respectively, after which it was recognized that the CCK-2 receptor is identical to the gastrin receptor of the stomach.18 The CCK-1 receptor is present in the gallbladder and, in most species, in the pancreas. The CCK-1 receptor has a 1000-fold higher affinity for CCK than for gastrin. The CCK-1– and CCK-2–gastrin receptors have more than 50% sequence homology and respond differentially to various receptor antagonists and to gastrin. Gastrin is released from specialized endocrine cells (G cells) into the circulation in response to a meal. The specific components of a meal that stimulate gastrin release include protein, peptides, and amino acids. Gastrin release is profoundly influenced by the pH of the stomach. Fasting and increased gastric acidity inhibit gastrin release, whereas a high gastric pH is a strong stimulus for its secretion. Hypergastrinemia occurs in pathologic states associated with decreased acid production, such as atrophic gastritis. Serum gastrin levels can also become elevated in patients on prolonged acid-suppressive medications, such as his­ tamine receptor antagonists and proton pump inhibitors. Hypergastrinemia in these conditions is caused by stimulation of gastrin production by the alkaline pH environment. Another important but far less common cause of hypergastrinemia is a gastrin-producing tumor, also known as Zollinger-Ellison syndrome (see Chapter 32). The gastrin analog, pentagastrin, has been used clinically to stimulate histamine and gastric acid secretion in diagnostic tests of acid secretory capacity (see Chapter 49).

CHOLECYSTOKININ

CCK is a peptide transmitter produced by I cells of the small intestine and is secreted into the blood following ingestion of a meal. Circulating CCK binds to specific CCK-1 receptors on the gallbladder, pancreas, smooth muscle of the stomach, and peripheral nerves to stimulate gallbladder contraction and pancreatic secretion, regulate gastric emptying and bowel motility, and induce satiety.19 These effects serve to

coordinate the ingestion, digestion, and absorption of dietary nutrients. Ingested fat and protein are the major food components that stimulate CCK release. CCK was originally identified as a 33–amino acid peptide. However, since its discovery larger and smaller forms of CCK have been isolated from blood, intestine, and brain. All forms of CCK are produced from a single gene by posttranslational processing of a preprohormone. Forms of CCK ranging in size from CCK-58 to CCK-8 have similar biological activities.20 CCK is the major hormonal regulator of gallbladder contraction. It also plays an important role in regulating mealstimulated pancreatic secretion (see Chapter 56) In many species, this latter effect is mediated directly through receptors on pancreatic acinar cells but in humans, in whom pancreatic CCK-1 receptors are less abundant, CCK appears to stimulate pancreatic secretion indirectly through enteropancreatic neurons that possess CCK-1 receptors. In some species, CCK has trophic effects on the pancreas, although its potential role in human pancreatic neoplasia is speculative. CCK also has been shown to delay gastric emptying.21 This action may be important in coordinating the delivery of food from the stomach to the intestine. CCK has been proposed as a major mediator of satiety and food intake, an effect that is particularly noticeable when food is in the stomach or intestine. CCK inhibits gastric acid secretion by binding to CCK-1 receptors on somatostatin (D) cells in the antrum and oxyntic mucosa. Somatostatin acts locally to inhibit gastrin release from adjacent G cells and directly inhibits acid secretion from parietal cells.22 Clinically, CCK has been used together with secretin to stimulate pancreatic secretion for pancreatic function testing. It is also used radiographically or scintigraphically to evaluate gallbladder contractility. There are no known diseases of CCK excess. Low CCK levels have been reported in individuals with celiac disease who have reduced intestinal mucosal surface area and in those with bulimia nervosa.23,24 Elevated levels of CCK have been reported in some patients with chronic pancreatitis (see Chapter 59), presumably because of reduced pancreatic enzyme secretion and interruption of negative feedback regulation of CCK release.25

SECRETIN

The first hormone, secretin, was discovered when it was observed that intestinal extracts, when injected intravenously into dogs, caused pancreatic secretion.26 Secretin is released by acid in the duodenum and stimulates pancreatic fluid and bicarbonate secretion, leading to neutralization of acidic chyme in the intestine (see Chapter 56). Secretin also inhibits gastric acid secretion (see Chapter 49) and intestinal motility. Human secretin is a 27–amino acid peptide and, similar to many other GI peptides, is amidated at the carboxyl terminus. It is the founding member of the secretin-glucagonVIP family of structurally related GI hormones. Secretin is selectively expressed in specialized enteroendocrine cells of the small intestine called S cells.27 The secretin receptor is a member of a large family of G protein–coupled receptors (GPCRs) that is structurally similar to receptors for glucagon, calcitonin, parathyroid hormone, pituitary adenylate cyclase–activating peptide (PACAP), and vasoactive intestinal polypeptide (VIP). One of the major physiological actions of secretin is stimulation of pancreatic fluid and bicarbonate secretion (see Chapter 56). Pancreatic bicarbonate, on reaching the duodenum, neutralizes gastric acid and raises the duodenal pH, thereby “turning off” secretin release (negative feedback). It

7

8

Section I  Biology of the Gastrointestinal Tract has been suggested that acid-stimulated secretin release is regulated by an endogenous intestinal secretin-releasing factor.28 This peptide stimulates secretin release from S cells until the flow of pancreatic proteases is sufficient to degrade the releasing factor and terminate secretin release. Although the primary action of secretin is to produce pancreatic fluid and bicarbonate secretion, it is also an enterogastrone, a substance that is released when fat is present in the GI lumen and that inhibits gastric acid secretion. In physiologic concentrations, secretin inhibits gastrin release, gastric acid secretion, and gastric motility.29 The most common clinical application of secretin is in the diagnosis of gastrin-secreting tumors,30 as discussed in Chapter 32.

Proglucagon

Glucagon

Pancreas

Glucagon

GLP-1

GLP-2

Small intestine

GLP-1

GLP-2

Figure 1-4.  Different post-translational processing of glucagon in the pancreas and small intestine. The glucagon gene transcript is transcribed and translated into a prohormone (proglucagon) capable of producing glucagon and glucagon-like peptides (GLP-1 and GLP-2). However, only glucagon is produced in the pancreas because of specific processing. In the small intestine, GLP-1 and GLP-2 are the primary products.

VASOACTIVE INTESTINAL POLYPEPTIDE

VIP is a neuromodulator that has broad significance in intestinal physiology. VIP is a potent vasodilator that increases blood flow in the GI tract and causes smooth muscle relaxation and epithelial cell secretion.31,32 As a chemical messenger, VIP is released from nerve terminals and acts locally on cells bearing VIP receptors. VIP belongs to a family of GI peptides, including secretin and glucagon, that are structurally related. The VIP receptor is a G protein–coupled receptor that stimulates intracellular cAMP generation. Like other GI peptides, VIP is synthesized as a precursor molecule that is cleaved to an active peptide of 28 amino acids. VIP is expressed primarily in neurons of the peripheral-enteric and central nervous systems and is released along with other peptides, including primarily PHI and/or PHM (see Table 1-1).33 VIP is an important neurotransmitter throughout the central and peripheral nervous systems.34 Because of its wide distribution, VIP has effects on many organ systems; most notably, in the GI tract, VIP stimulates fluid and electrolyte secretion from intestinal epithelium and bile duct cholangiocytes.35,36 VIP, along with NO, is a primary component of nonadrenergic, noncholinergic nerve transmission in the gut.37 GI smooth muscle exhibits a basal tone, or sustained tension, caused by rhythmic depolarizations of the smooth muscle membrane potential. VIP serves as an inhibitory transmitter of this rhythmic activity, causing membrane hyperpolarization and subsequent relaxation of GI smooth muscle. Accordingly, VIP is an important neuromodulator of sphincters of the GI tract, including the lower esophageal sphincter and sphincter of Oddi. In certain pathologic conditions, such as achalasia and Hirschsprung’s disease, the lack of VIP innervation is believed to play a major role in defective esophageal relaxation and bowel dysmotility, respectively.38,39 Unlike GI endocrine cells that line the mucosa of the gut, VIP is produced and released from neurons and it is likely that most measurable VIP in serum is of neuronal origin. Normally, serum VIP levels are low and do not appreciably change with a meal. However, in pancreatic cholera, also known as Verner-Morrison syndrome and manifested by watery diarrhea, hypokalemia, and achlorhydria,40 VIP levels can be extraordinarily high.35 VIP-secreting tumors usually produce a voluminous diarrhea41 (see Chapter 32).

GLUCAGON

Glucagon is synthesized and released from pancreatic alpha cells and from intestinal L cells of the ileum and colon. Pancreatic glucagon is a 29–amino acid peptide that regulates glucose homeostasis via gluconeogenesis, glycogenolysis, and lipolysis and is counterregulatory to insulin. The gene for glucagon encodes not only preproglucagon but also

glucagon-like peptides (GLPs). This precursor peptide consists of a signal peptide, a glucagon-related polypeptide, glucagon, and GLP-1 and GLP-2. Tissue-specific peptide processing occurs through prohormone convertases that produce glucagon in the pancreas and GLP-1 and GLP-2 in the intestine (Fig. 1-4).42 Glucagon and GLP-1 regulate glucose homeostasis.43 Glucagon is released from the endocrine pancreas in response to a meal and binds to G protein–coupled receptors on skeletal muscle and the liver to exert its glucoregulatory effects. GLP-1 stimulates insulin secretion and augments the insulin-releasing effects of glucose on the pancreatic beta cell (see later, “Enteroinsular Axis”). GLP-1 analogs have been developed for the treatment of type II diabetes mellitus. A long-acting human GLP-1 analog improves beta cell function and can lower body weight in patients with type II diabetes.44,45 GLP-2 is an intestinal growth factor and may have therapeutic implications in the maintenance of the GI mucosal mass and the reversal of villus atrophy.

GLUCOSE-DEPENDENT INSULINOTROPIC POLYPEPTIDE

GIP was discovered based on its ability to inhibit gastric acid secretion (enterogastrone effect) and was originally termed gastric inhibitory polypeptide. It was subsequently shown that the effects on gastric acid secretion occur only at very high concentrations that are above the physiologic range. However, GIP has potent effects on insulin release that (like GLP-1) potentiates glucose-stimulated insulin secretion.46 Based on this action, GIP was redefined as glucose-dependent insulinotropic polypeptide. GIP is a 42–amino acid peptide produced by K cells in the mucosa of the small intestine. GIP is released into the blood in response to ingestion of glucose or fat. In the presence of elevated blood glucose levels, GIP binds to its receptor on pancreatic beta cells, activating adenylate cyclase and other pathways that increase intracellular calcium concentrations, leading to insulin secretion. Importantly, however, the effects on insulin secretion occur only if hyperglycemia exists; GIP does not stimulate insulin release under normoglycemic conditions. GIP receptors are also expressed on adipocytes through which GIP augments triglyceride storage, which may contribute to fat accumulation. Based on the insulinotropic properties of GIP, coupled with its effects on adipocytes, it has been proposed that GIP may play a role in obesity and development of insulin resistance associated with type II diabetes mellitus.47 Consistent with this proposal was the experimental finding that mice lacking the GIP receptor do not gain weight when placed on a high-fat diet.48 It remains

Chapter 1  Gastrointestinal Hormones and Neurotransmitters to be seen whether GIP antagonists can be used to treat obesity. In rare circumstances, receptors for GIP may be aberrantly expressed in the adrenal cortex, resulting in fooddependent Cushing’s syndrome.49,50

PANCREATIC POLYPEPTIDE FAMILY

Originally isolated during the preparation of insulin, pancreatic polypeptide (PP) is the founding member of the PP family.51 The PP family of peptides includes NPY and peptide tyrosine tyrosine (PYY), which were discovered because of the presence of a C-terminal tyrosine amide.52,53 PP is stored and secreted from specialized pancreatic endocrine cells (PP cells),54 whereas NPY is a principal neurotransmitter found in the central and peripheral nervous systems.55 PYY has been localized to enteroendocrine cells throughout the GI tract but is found in greatest concentrations in the ileum and colon.56 The PP-PYY-NPY family of peptides functions as endocrine, paracrine, and neurocrine transmitters in the regulation of a number of actions that result from binding to one of five receptor subtypes.57 PP inhibits pancreatic exocrine secretion, gallbladder contraction, and gut motility.58 PYY inhibits vagally stimulated gastric acid secretion and other motor and secretory functions.59 An abbreviated form of PYY lacking the first two amino acids of the normally produced 36 amino acid peptide, PYY3-36, has been shown to reduce food intake when administered to humans, indicating that intestinally released peptide may play a role in regulating meal size.60 NPY is one of the most abundant peptides in the central nervous system and, in contrast to PYY3-36, is a potent stimulant of food intake.61 Peripherally, NPY affects vascular and GI smooth muscle function.62

SUBSTANCE P AND THE TACHYKININS

Substance P belongs to the tachykinin family of peptides, which includes neurokinin A and neurokinin B. The tachykinins are found throughout the peripheral and central nervous systems, and are important mediators of neuropathic inflammation.63 Tachykinins, as a group, are encoded by two genes that produce preprotachykinin A and pre­ protachykinin B. Common to both is a well-conserved C-terminal pentapeptide. Transcriptional and translational processing produce substance P, neurokinin A, and/or neurokinin B, which are regulated in large part by alternative splicing. These peptides function primarily as neuropeptides. Substance P is a neurotransmitter of primary sensory afferent neurons and binds to specific receptors in lamina I of the spinal cord.64 Three receptors for this family of peptides have been identified—NK-1, NK-2, and NK-3.65 Substance P is the primary ligand for the NK-1 receptor, neurokinin A for the NK-2 receptor, and neurokinin B for the NK-3 receptor. However, all these peptides can bind and signal through all three receptor subtypes. Substance P has been implicated as a primary mediator of neurogenic inflammation. In the intestine, Clostridium difficile–initiated experimental colitis results from toxininduced release of substance P and consequent activation of the NK-1 receptor.66 These inflammatory sequelae can be blocked by substance P receptor antagonists. Substance P receptors are more abundant in the intestine of patients with ulcerative colitis and Crohn’s disease.67

SOMATOSTATIN

Somatostatin is a 14–amino acid cyclic peptide that was initially identified as an inhibitor of growth hormone secretion. Since its discovery, it has been found in almost every organ in the body and throughout the GI tract. In the gut, somatostatin is produced by D cells in the gastric and intes-

tinal mucosa and islets of the pancreas, as well as enteric neurons.68 Somatostatin has a number of pharmacologic effects that are mostly inhibitory. In the stomach, somatostatin plays an important role in regulating gastric acid secretion.69 In the antrum, D cells are open to the lumen, where they are directly exposed to acid. A low gastric pH stimulates D cells that lie in close proximity to gastrin-producing cells to secrete somatostatin and inhibit gastrin release (see Chapter 49). Reduced gastrin secretion decreases the stimulus for acid production and the pH of the stomach contents rises. Thus, some of the inhib­ itory effects of gastric acid on gastrin release (see earlier, “Gastrin”) are mediated by somatostatin. Somatostatin release is also influenced by mechanical stimulation, dietary components of a meal, including protein, fat, and glucose, and other hormones and neurotransmitters.70 Muscarinic stimulation appears to be the most important neural stimulus to somatostatin secretion. At least five somatostatin receptors have been identified that account for divergent pharmacologic properties.71 For example, receptor subtypes 2 and 3 couple to inhibitory G proteins but receptor subtype 1 does not. In addition, only somatostatin receptor subtype 3 inhibits adenylate cyclase. The inhibitory effects of somatostatin are mediated by a decrease in cAMP, Ca2+ channel inhibition, or K+ channel opening. In the gut, somatostatin has broad inhibitory actions. In addition to effects on gastric acid, somatostatin reduces pepsinogen secretion. Somatostatin profoundly inhibits pancreatic enzyme, fluid, and bicarbonate secretion and reduces bile flow.72 The effects of somatostatin on gut motility are largely inhibitory, with the exception that it stimulates the migrating motor complex, possibly through effects on motilin. Somatostatin also reduces intestinal transport of nutrients and fluid, reduces splanchnic blood flow, and has inhibitory effects on tissue growth and proliferation.73,74 Because of its varied physiologic effects, somatostatin has several clinically important pharmacologic uses. Many endocrine cells possess somatostatin receptors and are sensitive to inhibitory regulation. Therefore, somatostatin and more recently developed somatostatin analogs are used to treat conditions of hormone excess produced by endocrine tumors, such as acromegaly, carcinoid tumors, and islet cell tumors (including gastrinomas).75 Its ability to reduce splanchnic blood flow and portal venous pressure has led to somatostatin analogs being useful in treating esophageal variceal bleeding (see Chapter 90).76 The inhibitory effects on secretion have been exploited by using somatostatin analogs to treat some forms of diarrhea and reduce fluid output from pancreatic fistulas. Many endocrine tumors express abundant somatostatin receptors, making it possible to use radiolabeled somatostatin analogs, such as octreotide, to localize even small tumors throughout the body.

MOTILIN

Motilin is a 22–amino acid peptide produced by endocrine cells of the duodenal epithelium.77 Motilin is secreted into the blood in a periodic and recurrent pattern that is synchronized with the migrating motor complex (MMC) under fasting conditions. Elevations in blood motilin levels regulate the phase III contractions that initiate in the antroduodenal region and progress toward the distal gut. Motilin secretion is not stimulated by eating. Motilin binds to specific receptors on smooth muscle cells of the esophagus, stomach, and small and large intestines through which it exerts propulsive activity.78 Agonists to the motilin receptor such as erythromycin have pronounced

9

10

Section I  Biology of the Gastrointestinal Tract effects on GI motility, which occasionally produces undesired side effects of abdominal cramping and diarrhea.79 However, motilin agonists may be useful to treat conditions of impaired gastric and intestinal motility and are being investigated for the treatment of constipation-predominant irritable bowel syndrome.80

LEPTIN

Leptin is a 167–amino acid protein that is secreted primarily from adipocytes. Blood leptin levels reflect total body fat stores.81 Its primary action appears to be to reduce food intake. Leptin is a member of the cytokine family of signaling molecules. Five different forms of leptin receptors have been reported.82 A short form of the receptor appears to transport leptin from the blood across the blood-brain barrier, where it has access to the hypothalamus. A long form of the leptin receptor is located in hypothalamic nuclei, where leptin binds and activates the Janis kinase signal transduction and translation system (JAK STAT).83 Small amounts of leptin are produced by the chief cells of the stomach and by the placenta, and are present in breast milk. Peripheral administration of leptin reduces food intake. However, this effect is reduced as animals become obese. Interestingly, when injected into the central nervous system, obese animals respond normally to leptin and reduce food intake, suggesting that leptin “resistance” in obesity occurs at the level of the leptin receptor that transports leptin across the blood-brain barrier.84 Leptin’s ability to reduce food intake occurs within the brain by decreasing NPY (a potent stimulant of food intake) and by increasing α– melanocyte-stimulating hormone (α−MSH), an inhibitor of food intake.85 Peripherally, leptin acts synergistically with cholecystokinin to reduce meal size.86 In obese rats lacking the leptin receptor, the synergistic effects of leptin plus CCK to reduce meal size are lost, but could be restored with genetic reconstitution of the leptin receptor in the brain.87 One might expect loss of leptin-CCK synergy on meal size in those rare cases of human obesity caused by leptin receptor defects or even with leptin resistance. Blood levels of leptin increase as obesity develops and leptin appears to reflect total fat content.88 At the cellular level, large adipocytes produce more leptin than small adipocytes. Because of its effects on food intake, it was initially thought that exogenous leptin could be used therapeutically to treat obesity. However, only a very modest effect on weight loss has been demonstrated in clinical trials. Leptin deficiency has been reported as a cause of obesity in a few families, but this condition is extremely rare.89,90 Mutation of the leptin receptor has been described as a cause of obesity in at least one family.91

GHRELIN

Ghrelin is a 28–amino acid peptide produced by the stomach and is the natural ligand for the growth hormone secretagogue (GHS) receptor.92 When administered centrally or peripherally ghrelin stimulates growth hormone secretion, increases food intake, and produces weight gain.93,94 Circulating ghrelin levels increase during periods of fasting or under conditions associated with negative energy balance, such as starvation or anorexia. In contrast, ghrelin levels are low after eating and in obesity. Ghrelin appears to play a central role in the neurohormonal regulation of food intake and energy homeostasis. The gastric fundus is the most abundant source of ghrelin, although lower amounts of ghrelin are found in the intestine, pancreas, pituitary, kidney, and placenta. Ghrelin is produced by distinctive endocrine cells known as P/D1

cells95,96 that are of two types, open and closed. The open type is exposed to the lumen of the stomach, where it comes into contact with gastric contents, whereas the closed type lies in close proximity to the capillary network of the lamina propria.97 Both cell types secrete hormone into the bloodstream. Based on its structure, ghrelin is a member of the motilin family of peptides and, like motilin, ghrelin stimulates gastric contraction and enhances stomach emptying. The observations that circulating ghrelin levels increase sharply before a meal and fall abruptly after a meal suggest that it serves as a signal for initiation of feeding. The effects of food on plasma ghrelin levels can be reproduced by ingestion of glucose and appear to be unrelated to the physical effects of a meal on gastric distention. Circulating ghrelin levels are low in states of positive energy balance such as obesity and are inversely correlated with body mass index.98,99 Conversely, ghrelin levels are high in fasting, cachexia, and anorexia. Importantly, weight loss increases circulating ghrelin levels.100 Ghrelin released from the stomach acts on the vagus nerve to exert its effects on feeding. However, it is also active when delivered to the central nervous system and, in this location, ghrelin activates NPY and agouti-related protein-producing neurons in the arcuate nucleus of the hypothalamus, which is involved in the regulation of feeding.94,101 Gastric bypass patients do not demonstrate the premeal increase in plasma ghrelin that is seen in normal individuals.102 This lack of ghrelin release may be one of the mechanisms contributing to the overall effectiveness of gastric bypass surgery for inducing weight loss. Prader-Willi syndrome is a congenital obesity syndrome characterized by severe hyperphagia, growth hormone deficiency, and hypogonadism. Although obesity is ordinarily associated with low ghrelin levels, patients with PraderWilli syndrome have high circulating ghrelin levels that do not decline after a meal.103,104 The levels of ghrelin in this syndrome are similar to those that can stimulate appetite and increase food intake in individuals receiving infusions of exogenous ghrelin, suggesting that abnormal ghrelin secretion may be responsible for the hyperphagia in PraderWilli syndrome.105

OTHER CHEMICAL MESSENGERS OF THE GASTROINTESTINAL TRACT The enteric nervous system, through intrinsic and extrinsic neural circuits, controls GI function. This control is mediated by various chemical messengers, including motor and sensory pathways of the sympathetic and parasympathetic nervous systems. The parasympathetic preganglionic input is provided by cholinergic neurons and elicits excitatory effects on GI motility via nicotinic and muscarinic receptors. Sympathetic input occurs through postganglionic adrenergic neurons.

ACETYLCHOLINE

Acetylcholine is synthesized in cholinergic neurons and is the principal regulator of GI motility and pancreatic secretion. Acetylcholine is stored in nerve terminals and released by nerve depolarization. Released acetylcholine binds to postsynaptic muscarinic and/or nicotinic receptors. Nicotinic acetylcholine receptors belong to a family of ligand-gated ion channels and are homopentamers or heteropentamers composed of α, β, γ, δ, and ε subunits.106 The

Chapter 1  Gastrointestinal Hormones and Neurotransmitters α subunit is believed to be the mediator of postsynaptic membrane depolarization following acetylcholine receptor binding. Muscarinic receptors belong to the heptahelical GPCR family. There are five known muscarinic cholinergic receptors (M1 to M5). Muscarinic receptors can be further classified based on receptor signal transduction, with M1, M3, and M5 stimulating adenylate cyclase and M2 and M4 inhibiting this enzyme. Acetylcholine is degraded by the enzyme acetylcholinesterase, and the products may be recycled through high-affinity transporters on the nerve terminal.

CATECHOLAMINES

The primary catecholamine neurotransmitters of the enteric nervous system include norepinephrine and dopamine. Norepinephrine is synthesized from tyrosine and released from postganglionic sympathetic nerve terminals that innervate enteric ganglia and blood vessels. Tyrosine is converted to dopa by tyrosine hydroxylase. Dopa is initially converted into dopamine by dopa decarboxylase and packaged into secretory granules. Norepinephrine is formed from dopamine by the action of dopamine β-hydroxylase in the secretory granule. After an appropriate stimulus, norepinephrine-containing secretory granules are released from nerve terminals and bind to adrenergic receptors. Adrenergic receptors are G protein–coupled, have seven typical membrane-spanning domains, and are of two basic types, α and β. α-Adrenergic receptors are further classified into α1A, α1B, α2A, α2B, α2C, and α2D. Similarly, β receptors include β1, β2, and β3. Adrenergic receptors are known to signal through various G proteins, resulting in stimulation or inhibition of adenylate cyclase and other effector systems. Norepinephrine signaling is terminated by intracellular monoamine oxidase or by rapid reuptake by an amine transporter. The actions of adrenergic receptor stimulation regulate smooth muscle contraction, intestinal blood flow, and GI secretion.

DOPAMINE

Dopamine is an important mediator of GI secretion, absorption, and motility and is the predominant catecholamine neurotransmitter of the central and peripheral nervous systems. In the central nervous system, dopamine regulates food intake, emotions, and endocrine responses and, peri­ pherally, it controls hormone secretion, vascular tone, and GI motility. Characterization of dopamine in the GI tract has been challenging for several reasons. First, dopamine can produce inhibitory and excitatory effects on GI motility.107 Generally, the excitatory response, which is mediated by presynaptic receptors, occurs at a lower agonist concentration than the inhibitory effect, which is mediated by postsynaptic receptors. Second, localization of dopamine receptors has been hampered by identification of dopamine receptors in locations that appear to be species specific.108 Third, studies of dopamine in GI tract motility have often used pharmacologic amounts of this agonist. Therefore, the interpretation of results has been confounded by the ability of dopamine to activate adrenergic receptors at high doses. Classically, dopamine was thought to act via two distinct receptor subtypes, type 1 and type 2. Molecular cloning has now demonstrated five dopamine receptor subtypes, each with a unique molecular structure and gene locus.108 Dopamine receptors are integral membrane GPCRs, and each receptor subtype has a specific pharmacologic profile when exposed to agonists and antagonists. After release from the nerve terminal, dopamine is cleared from the synaptic cleft by a specific dopamine transporter.

SEROTONIN

Serotonin has long been known to play a role in GI neurotransmission.109 The GI tract contains more than 95% of the total body serotonin, and serotonin is important in various processes, including epithelial secretion, bowel motility, nausea and emesis.110 Serotonin is synthesized from tryptophan, an essential amino acid, and is converted to its active form in nerve terminals. Secreted serotonin is inactivated in the synaptic cleft by reuptake via a serotonin-specific transporter. Most plasma serotonin is derived from the gut, where it is found in mucosal enterochromaffin cells and the enteric nervous system. Serotonin mediates its effects by binding to a specific receptor. There are seven different serotonin receptor subtypes found on enteric neurons, enterochromaffin cells, and GI smooth muscle (5-HT1 to 5-HT7). The actions of serotonin are complex (Fig. 1-5).111 It can cause smooth muscle contraction through stimulation of cholinergic nerves or relaxation by stimulating inhibitory NO-containing neurons.110 Serotonin released from mucosal cells stimulates sensory neurons, initiating a peristaltic reflex and secretion (via 5-HT4 receptors) and modulates sensation through activation of 5-HT3 receptors.109 The myenteric plexus contains serotoninergic interneurons that project to the submucosal plexus and ganglia extrinsic to the bowel wall. Extrinsic neurons activated by serotonin participate in bowel sensation and may be responsible for abdominal pain, nausea, and symptoms associated with irritable bowel syndrome. Intrinsic neurons activated by serotonin are primary components of the peristaltic and secretory reflexes responsible for normal GI function. Serotonin may also activate vagal afferent pathways and, in the central nervous system, modulates appetite, mood, and sexual function. Because of these diverse effects, it is not surprising that selective serotonin reuptake inhibitor drugs (SSRIs), commonly used to treat depression and anxiety, have pro­minent GI side effects when compared with placebo treatment. Serotonin and its receptor have been implicated in the pathogenesis of motility disorders of the GI tract.112 Characterization of specific serotonin receptor subtypes has led to the development of selective agonists and antagonists for the treatment of irritable bowel syndrome and chronic constipation and diarrhea. For example, 5-HT3 receptor antagonists, which reduce intestinal secretion, are used to treat diarrhea-predominant irritable bowel syndrome. 5-HT4 receptor agonists elicit prokinetic effects and are used to treat constipation-predominant irritable bowel syndrome and other motility disorders.113,114 Serotonin can also be enzymatically converted to melatonin by serotonin N-acetyltransferase.115 Other than the pineal gland, the GI tract is the major source of the body’s melatonin. Melatonin is produced in enterochromaffin cells and released into the blood after ingestion of a meal. A number of actions on the GI tract have been described for melatonin, including reducing gastric acid and pepsin secretion, inducing smooth muscle relaxation, and preventing epithelial injury through an antioxidant effect.116 It has been proposed that melatonin released after a meal may contribute to postprandial somnolence.117

HISTAMINE

In the GI tract, histamine is best known for its central role in regulating gastric acid secretion (see Chapter 49) and intestinal motility. Histamine is produced by enterochromaffin-like cells of the stomach and intestine as well as enteric nerves. Histamine is synthesized from l-histidine by histidine decarboxylase and activates three

11

12

Section I  Biology of the Gastrointestinal Tract

CNS Longitudinal muscle 5-HT3 Excitatory motor neuron

5-HT4

5-HT4

Myenteric plexus

5-HT3

Inhibitory motor neuron

Circular muscle

5-HT3

Intrinsic primary

Figure 1-5.  Role of serotonin in the enteric nervous system. This model illustrates the location of 5hydroxytryptamine3 (5-HT3) and 5-HT4 receptor subtypes in the GI tract. CNS, central nervous system. (Modified from Talley NJ. Serotoninergic neuroenteric modu­ lators. Lancet 2001; 358:2061-8).

afferent neuron 5-HT3

5-HT4

Sensory neuron

Extrinsic Submucosal afferent plexus neuron 5-HT3 Mucosa

Enterochromaffin cells

NO bound to guanylyl cyclase NO synthase Arginine

Rapid diffusion of NO

GTP

cGMP

Relaxation

Nitric oxide Smooth muscle cell Activated neuron Figure 1-6.  Nitric oxide (NO) signals smooth muscle relaxation. NO, synthesized from arginine by nitric oxide synthase, diffuses across the plasma membrane into smooth muscle cells. NO binds to and activates guanylyl cyclase, which converts GTP to cGMP. cGMP causes smooth muscle relaxation. (Modified from Alberts B, Bray D, Lewis J, et al, editors. Molecular biology of the cell. 4th ed. New York: Garland Science; 2002. p 831.)

GPCR subtypes. H1 receptors are found on smooth muscle and vascular endothelial cells and are linked to phospholipase C (PLC) activation. As such, the H1 receptor mediates many of the allergic responses induced by histamine. H2 receptors are present on gastric parietal cells, smooth muscle, and cardiac myocytes. H2 receptor binding stimulates Gs (G proteins that stimulate adenylate cyclase) and activates adenylate cyclase. H3 receptors are present in the central nervous system and GI tract enterochromaffin cells. These receptors signal through Gi and inhibit adenylate cyclase.118 Histamine can also interact with the N-methyl-daspartate (NMDA) receptor and enhance activity of NMDAbearing neurons independently of the three known histamine receptor subtypes. Unlike other neurotransmitters, there is no known transporter responsible for termination of histamine’s action. However, histamine is metabolized to telemethylhistamine by histamine N-methyltransferase and is then degraded to telemethylimidazoleacetic acid by monoamine oxidase B and an aldehyde dehydrogenase.

known as endothelial NOS and neuronal NOS, respectively, and are constitutively active. Small changes in NOS activity can occur through elevations in intracellular calcium. The inducible form of NOS (type II) is apparent only when cells become activated by specific inflammatory cytokines. This form of NOS is capable of producing large amounts of NO and is calcium-independent. NOS is often colocalized with VIP and PACAP in neurons of the enteric nervous system.120 NO, being an unstable gas, has a relatively short half-life. Unlike most neurotransmitters and hormones, NO does not act via a membrane-bound receptor. Instead, NO readily diffuses into adjacent cells to activate guanylate cyclase directly (Fig. 1-6). NO activity is terminated by its oxidation to nitrate and nitrite. Many enteric nerves use NO to signal neighboring cells and induce epithelial secretion, vasodilation, or muscle relaxation. NO is also produced by macrophages and neutrophils to help kill invading organisms.121

NITRIC OXIDE

Adenosine is an endogenous nucleoside that acts through any of four GPCR subtypes.122 Adenosine causes relaxation of intestinal smooth muscle and stimulates intestinal secretion. Adenosine can also cause peripheral vasodilation and

NO is a unique chemical messenger produced from larginine by the enzyme nitric oxide synthase (NOS).119 Three types of NOS are known. Types I and III are also

ADENOSINE

Chapter 1  Gastrointestinal Hormones and Neurotransmitters activation of nociceptors that participate in neural pain pathways.

Extracellular

CYTOKINES

Cytokines are a group of polypeptides produced by various immunomodulatory cells and are involved in cell proliferation, immunity, and inflammation. Cytokines are induced by specific stimuli, such as toxins produced by pathogens, and often elicit a complex response involving other cellular mediators to eradicate the foreign substance. Cytokines may be categorized as interleukins (ILs), tumor necrosis factors (TNFs), lymphotoxins, interferons, colony-stimulating factors (CSFs), and others.123 Interleukins can be further subtyped into at least 35 separate substances, IL-1 to IL-35. There are two TNFs, TNF-α and TNF-β, which are also known as lymphotoxin-α. Interferons are produced during viral or bacterial infection and come in two varieties, interferon-α (also known as leukocyte-derived interferon or interferon-β) and interferon-γ. Interferon-α is produced by T lymphocytes and is used clinically for the treatment of viral hepatitis (see Chapters 78 and 79). The major CSFs are granulocyte mononuclear phagocyte CSF, mononuclear phagocyte CSF, and granulocyte CSF. These agents are used for chemotherapy-induced neutropenia and marrow support after bone marrow transplantation. Chemokines initiate and propagate inflammation and are of two groups, CXC (α chemokines) and CC (β chemokines). Other cytokines, such as transforming growth factor-β (TGF)-β and platelet-derived growth factor (PDGF), have proliferative effects.

Intracellular

Figure 1-7.  Molecular structure of a typical heptahelical G protein–coupled receptor. The amino terminus is extracellular and of variable length. It often contains N-linked glycosylation sites (Y) important in ligand binding. There are seven membrane-spanning domains and intracellular loops that contain sites for G protein binding and possible phosphorylation residues (orange circles).

Ligand

Extracellular

SIGNAL TRANSDUCTION Cells live in a constantly changing milieu. The structure and biochemical nature of this environment are dynamic and, for cells to function normally, they must be able to access this changing information. The biochemical mediators of this information are cell surface receptors and transmitters. Receptors transduce signals from the extracellular space to the intracellular compartment. Each step in the process from receptor activation to receptor desensitization, internalization, and resensitization represents a potential regu­ latory checkpoint and possible target for therapeutic intervention. Cell surface receptors include GPCRs, ion channels, and enzyme-coupled receptors.

α

β γ

Effector

Intracellular events Figure 1-8.  Hormones (ligands) bind to specific G protein–coupled receptors at a unique location within the receptor-binding pocket. On binding, the receptor conformation is altered so that a specific G protein α subunit is activated. G protein activation leads to dissociation of the α subunit from the βγ subunit and activation of effector pathways. These effectors include adenylate cyclase, ion channels, and an array of other systems. Intracellular

G PROTEIN–COUPLED RECEPTORS

GPCRs are seven membrane-spanning domain proteins associated with a heterotrimeric G protein (Fig. 1-7). The membrane regions consist of α-helical domains with a conserved structural motif.124 GPCRs contain an extracellular amino terminus and an intracellular carboxyl terminus (see Fig. 1-7). When stimulated by the appropriate chemical messenger, the GPCR undergoes a conformational change and couples to a specific G protein. The first crystal structure of a GPCR, for rhodopsin, was elucidated in 2000.125 The three-dimensional structure of the rhodopsin receptor reveals a highly organized heptahelical transmembrane component with a portion of the C-terminus perpendicular to the seventh and final membrane-spanning domains of the protein.

G PROTEINS

G proteins are molecular intermediaries that initiate the intracellular communication process on ligand binding to its GPCR (Fig. 1-8).126 G proteins are composed of three

subunits—α, β, and γ—and are classified according to their α subunit. They activate various effector systems, including adenylate cyclase, guanylate cyclase, phospholipases, and specific ion channels.127 G proteins that stimulate adenylate cyclase are classified as Gs; those that inhibit adenylate cyclase are called Gi.128 When an agonist binds to a Gs-coupled receptor, a conformational change occurs, allowing the receptor to associate with the Gαs subunit. Under basal (unstimulated) conditions, Gαs is bound to guanosine diphosphate (GDP); however, with agonist binding, GDP is released and replaced with guanosine triphosphate (GTP). The Gs-GTP complex then activates adenylate cyclase, resulting in the generation of cAMP from adenosine triphosphate (ATP) within the cell cytoplasm. cAMP phosphorylates effector proteins that ultimately lead to responses such as secretion, cell movement, and growth. Receptor activation also initiates

13

14

Section I  Biology of the Gastrointestinal Tract Table 1-2  Classification of G Protein α Subunits and Their Signaling Pathways Class

Signaling

Gαs Gαi and Gαo

Adenylate cyclase, calcium channels Adenylate cyclase, cyclic guanosine monophosphate, phosphodiesterase, c-Src, STAT 3 Phospholipase C-β Sodium-hydrogen exchange

Gαq Gα12 and Gα13

the dissociation of the α subunit from the βγ subunits. However, the βγ subunits remain tightly associated and themselves participate in a vast array of cellular signals. For example, not only can βγ subunits activate GPCR kinases, adenylate cyclase, and ion channels, they induce receptor desensitization and stimulate Ras-mediated mitogenactivated protein (MAP) kinase.129,130 The Gαs-GTP complex is gradually inactivated by guanosine triphosphatase (GTPase), which converts GTP to GDP. This enzymatic conversion occurs spontaneously by the G protein, which is itself a GTPase. The conversion of GTP to GDP terminates G protein stimulation of adenylate cyclase and is one way whereby the basal condition is restored. Certain GPCRs activate an inhibitory G protein (Gαi) that inhibits cAMP accumulation and antagonizes the effects of Gs-coupled events. In this manner, GPCRs can maintain fine control of the cellular cAMP concentration and subsequent intracellular signaling. Members of this GPCR family also activate phospholipases and phosphodiesterases, and are often involved with ion channel regulation. Other GPCRs couple with Gq and G12 (see Table 1-2). The Gq family of G protein subunits regulates the production of inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG).131 Following α subunit dissociation from βγ, when the α subunit reverts to the GDP-bound form, it reassociates with βγ. With reestablishment of the αβγ heterotrimer, along with other mechanisms of desensitization, receptor signaling via the separate subunits ceases.

Effector Systems

Following receptor occupation, G protein subunits cause activation of enzymes or other proteins, ultimately resulting in intracellular signaling events (Table 1-2). Enzymes such as adenylate cyclase or phospholipase C generate specific second messengers such as cAMP or IP3 and DAG. Some G proteins couple directly to specific ion channels, such as potassium or calcium channels, and initiate changes in ion permeability. The effector systems are not well understood for some receptors, such as those involved with cell growth and differentiation. Other G proteins such as Go may activate the phosphoinositide system. When bound to hormone, receptors that couple to Go activate PLC, which acts on inositol phospholipids found in the cell membrane. PLC can cause the hydrolysis of phosphatidylinositol 4,5-bisphosphate, generating 1,2-DAG and IP3. DAG and IP3 can regulate cell metabolism by increasing intracellular calcium levels.

Receptor Desensitization

To ensure the rapidity of hormone signaling, shortly after receptor stimulation, a series of events is initiated that ultimately acts to turn off signaling. The principal events in this process involve receptor desensitization and internalization, which reestablish cell responsiveness.

Phosphorylation of the receptor is one of the initial events involved in turning off the signal after agonist binding and occurs through binding of arrestin-like molecules, which uncouple the receptor from the G protein.132 This uncoupling and subsequent receptor internalization (sequestration) continue the process of signal termination and eventually lead to the reestablishment of cell responsiveness.

Receptor Resensitization

Internalization or sequestration of the receptor occurs within minutes of receptor occupancy. Agonist-activated receptors are phosphorylated by G protein–coupled receptor kinases at specific intracellular sites, which causes G protein uncoupling and initiates receptor endocytosis. GPCR endocytosis is followed by receptor dephosphorylation, recycling, and down-regulation. Chronic exposure of cells to high concentrations of hormones frequently leads to a decrease in cell surface–binding sites. This reduction in surface receptor expression is termed down-regulation and is the result of receptor internalization. The mechanisms used by the cell that distinguish receptor internalization and recycling from down-regulation are not clear. However, long-term agonist exposure to some receptors has been shown to activate signaling molecules that may be important in receptor down-regulation.

RECEPTORS NOT COUPLED TO G PROTEINS Enzyme-Coupled Receptors

Receptor Tyrosine Kinases Unlike GPCRs, where ligand-receptor interaction causes activation of a G protein intermediary, some ligand receptors possess intrinsic protein tyrosine kinase activity. These membrane-spanning cell surface receptors catalyze the transfer of phosphate from ATP to target proteins. Such receptors are structurally unique in that they contain glycosylated extracellular binding domains, a single transmembrane domain, and a cytoplasmic domain. The cytoplasmic domain contains a protein tyrosine kinase region and substrate region for agonist-activated receptor phosphorylation. With activation, these receptors may phosphorylate themselves or be phosphorylated by other protein kinases.133 In general, receptor tyrosine kinases exist in the cell membrane as monomers. However, with ligand binding, these receptors dimerize, autophosphorylate, and initiate other intracellular signal transduction pathways. Most receptor tyrosine kinases couple, via ligand binding, to Ras and subsequently activate MAP kinase. MAP kinase is then able to modulate the regulation of other cellular proteins, including transcription factors. Members of the receptor tyrosine kinase family include the insulin receptor, growth factor receptors (vascular endothelial growth factor, PDGF, epi­ dermal growth factor [EGF], fibroblast growth factor [FGF], insulin-like growth factor I [IGF] I, macrophage-CSF, nerve growth factor), and receptors involved in development.134 Receptor tyrosine kinases are discussed further in Chapter 3 in relation to cellular growth and neoplasia. Activated tyrosine kinase receptors participate in a number of intracellular signaling events that involve the phosphorylated cytoplasmic domain. Specific phosphorylated tyrosine residues serve as binding sites for Src homology regions 2 and 3 (SH2 and SH3 domains). The result of SH2 domain binding is activation or modulation of the signaling protein that contains this binding domain. In this manner, receptor tyrosine kinases activate diverse signaling pathways.135

Chapter 1  Gastrointestinal Hormones and Neurotransmitters Receptor Guanylate Cyclases Receptor guanylate cyclases use cyclic GMP (cGMP) as a direct intracellular mediator. These cell surface receptors contain an extracellular ligand-binding region, a single transmembrane domain, and a cytoplasmic guanylate cyclase catalytic domain.136 Ligand stimulation of a receptor guanylate cyclase results in activation of cGMP-dependent protein kinase, which is a serine-threonine protein kinase. The atrial natriuretic peptide (ANP) receptor is a representative receptor guanylate kinase, which mediates the potent smooth muscle relaxing activity of ANP. Nonreceptor Tyrosine Kinases Some cell surface receptors involved in inflammation and hematopoietic cell regulation work through tyrosine kinases but do not contain a cytoplasmic catalytic domain. The Src family of kinases is the primary component of this receptor signaling system.137 Receptor Tyrosine Phosphatases Leukocyte regulation is modulated by surface receptors whose function is to remove phosphate groups from specific phosphotyrosines. CD45 is a surface protein found in white blood cells that participates in T and B cell activation.138 CD45 contains a single membrane-spanning domain and a cytoplasmic region with tyrosine phosphatase activity. Depending on the substrate, dephosphorylation of signaling proteins may result in reduced or enhanced activity. Receptors in this family are important in inflammation and immune regulation and have been shown to participate in GI development, growth, and cancer. Receptor Serine-Threonine Kinases TGF-β (see “Growth Factor Receptors”) receptors are a unique group of surface proteins that are involved in various cell functions, including chemotaxis, inflammation, and proliferation. These receptors contain a single membrane domain and a cytoplasmic serine-threonine kinase region. Receptor stimulation initiates activation of the serinethreonine kinase and subsequent modulation of cellular protein function.139

are normally secreted in response to food ingestion and mediate many of the nutrient effects on the GI tract. They play a key role in cellular proliferation. Alterations in intestinal proliferation are manifested by atrophy, hyperplasia, dysplasia, or malignancy (see Chapter 3). Growth factors that have important effects on the GI tract include peptides of the EGF, TGF-β, IGF, FGF, and PDGF families, hepatocyte growth factors, trefoil factors, and many cytokines (including interleukins).140

GROWTH FACTOR RECEPTORS

Growth factors regulate cellular proliferation by interacting with specific cell surface receptors. These receptors are membrane proteins that possess specific binding sites for the growth factor ligand. An unusual form of signaling occurs when the ligand interacts with its receptor within the same cell. For example, PDGF receptors present on the intracellular surface of fibroblast cell lines are activated by intracellular ligand. This process is known as intracrine signaling. Most peptide growth factors, however, interact with receptors on different cells to regulate proliferation. Growth factor receptors can be single polypeptide chains containing one membrane-spanning region, such as the receptor for EGF, or they may be composed two subunit heterodimers, with one subunit containing a transmembrane domain and the other residing intracellularly but covalently bound to the transmembrane subunit (Fig. 1-9). Heterodimers may also dimerize to form a receptor composed of four subunits (e.g., IGF receptor). Binding of the ligand to its receptor usually causes aggregation of two or more receptors and activation of intrinsic tyrosine kinase activity. Growth factor receptors also have the ability to autophosphorylate when bound to ligand. In addition, receptor tyrosine kinase activity may phosphorylate other intracellular proteins important in signal transduction. Autophosphorylation attenuates the receptor’s kinase activity and often leads to down-regulation and internalization of the receptor. Mutation of the receptor at its autophos-

Proteoglycan

Ion Channel–Coupled Receptors

Ion channel–coupled receptors are involved in rapid signaling between cells. Ion channel signaling is particularly important in nerve cells and other electrically excitable tissues such as muscle. In nerve cells, a relatively small number of neurotransmitters are released that act directly on ion channel proteins, causing them to open or close. Ion channels are selective to specific anions or cations and, when open, allow the flow of those particular ions across the plasma membrane according to the concentration inside and outside the cell. This flow of ions regulates the excitability of the target cell, which can trigger cellular responses such as neurotransmission, muscle contraction, electrolyte and fluid secretion, and hormone release.

HORMONE AND TRANSMITTER REGULATION OF GASTROINTESTINAL GROWTH Growth of GI tissues is a balance between cellular proliferation and senescence. Many factors participate in maintenance of the GI mucosa. Nutrients and other luminal factors stimulate growth of the intestinal mucosa and are necessary to maintain normal digestive and absorptive functions. Hormones and transmitters serve as secondary messengers that

Plasma membrane

Cytoplasm

EGF receptor

Insulin receptor IGF-I receptor

PDGF TGFβ-I TGFβ-II TGFβ-III receptor receptor receptor receptor Tyrosine kinase Serine/Threonine kinase

Figure 1-9.  Growth factor receptors in the gastrointestinal tract. Schematic examples of growth factor receptor families are depicted in relation to the cell surface. Receptor regions that contain kinase activity are shown in boxes. On activation, these receptors have the ability to autophosphorylate or phosphorylate other proteins to propagate intracellular cell signaling. EGF, epidermal growth factor; IGF-I, insulin-like growth factor I; PDGF, platelet-derived growth factor; TGF, transforming growth factor (β-I, -II, -III). (Modified from Podolsky DK: Peptide growth factors in the gastro­ intestinal tract. In Johnson LR, editor. Physiology of the gastrointestinal tract. New York: Raven Press; 1994. p 129.)

15

16

Section I  Biology of the Gastrointestinal Tract phorylation site may lead to constitutive receptor activity and cellular transformation. Growth factor receptors may couple to various intracellular signaling pathways, including adenylate cyclase, phospholipase C, calciumcalmodulin protein kinases, MAP kinase, and nuclear transcription factors. Thus, growth factors play important and varied roles in most cells of the GI tract. It is not surprising, therefore, that mutations in growth factor receptors or downstream signaling proteins can lead to unregulated cell growth and neoplasia (see Chapter 3). An important action of growth factors is their ability to modulate the expression of transacting transcription factors that can regulate expression of many other genes.141 Early response genes such as jun and fos are activated rapidly after ligand binding and control the expression of many other genes involved in cellular proliferation. Other important transcriptional factors include c-myc and nuclear factor κB (NF-κB). The latter is found in the cytoplasm in an inactive form and, following ligand binding, translocates to the nucleus, where it activates other transcription factors. NF-κB is a key target for strategies to regulate cellular proliferation and inflammation. In its phosphorylated form Rb-1, originally identified in retinoblastoma, is an inhibitor of cellular proliferation that complexes with the tran­ scription factor p53. Dephosphorylation of Rb-1 releases p53, which activates other genes leading to cellular proliferation. Almost all growth factors of the GI tract exert paracrine effects. However, many growth factors also possess autocrine and even intracrine actions. It has become apparent that growth factors and other signaling molecules secreted into the lumen of the gut can have important local biological actions. Distant effects of growth factors found in the circulation may be important for growth of certain types of cancers, particularly lung and colon cancer.

EPIDERMAL GROWTH FACTOR

EGF was the first growth factor to be discovered. It is the prototype for a family of growth factors that are structurally related and have similarly related receptors. Other members of the family include TGF-α, amphiregulin, and heparinbinding EGF. EGF is identical to urogastrone (originally isolated from urine), which was shown to inhibit gastric acid secretion and promote healing of gastric ulcers. EGF is secreted from submaxillary glands and Brunner’s glands of the duodenum. It is likely that EGF interacts with luminal cells of the GI tract to regulate proliferation. EGF has important trophic effects on gastric mucosa, and the wide distribution of EGF receptors suggests that EGF has mitogenic actions on various cells throughout the gut. The EGF receptor has been reported to be responsible for gastric hyperplasia in patients with Ménétrier’s disease.142 Moreover, two patients were effectively treated with a mono­ clonal antibody that blocks ligand binding to the EGF receptor.143 EGF receptors are considered important targets for the experimental treatment of human cancer based on the evidence that they play a critical role in the growth and survival of certain tumors. Monoclonal antibodies as well as small tyrosine kinase inhibitors have been undergoing clinical evaluation for the treatment of human tumors.144

TRANSFORMING GROWTH FACTOR-α

TGF-α is produced by most epithelial cells of the GI tract and acts through the EGF receptor. Therefore, it shares trophic properties with EGF. It is believed to play a key role in gastric reconstitution after mucosal injury. Moreover, it

appears to be important in intestinal neoplasia because most gastric and colon cancers produce TGF-α (see Chapters 54 and 123).

TRANSFORMING GROWTH FACTOR-β

A family of TGF-β peptides exerts various biological actions, including stimulation of proliferation, differentiation, embryonic development, and formation of extracellular matrix.139 In contrast with the TGF-α receptor, there are three distinct TGF-β receptors (see Fig. 1-9).145 TGF-β modulates cell growth and proliferation in nearly all cell types and can enhance its own production from cells. It is likely that TGF-β plays a critical role in inflammation and tissue repair. TGF-β augments collagen production by recruitment of fibroblasts through its chemoattractant properties. This action can have beneficial or deleterious effects, depending on its site of deposition and abundance. For example, TGF-β may play a key role in the development of adhesions following surgery.146

INSULIN-LIKE GROWTH FACTORS

Alternative splicing of the insulin gene produces two structurally related peptides, IGF I and IGF II.147 IGFs signal through at least three different IGF receptors. The IGF I receptor is a tyrosine kinase, and the IGF II receptor is identical to the mannose 6-phosphate receptor. Although the exact function of IGFs in the GI tract is not clearly understood, they have potent mitogenic activity in intes­ tinal epithelium. IGF II appears to be critical for embryonic development.

FIBROBLAST GROWTH FACTOR AND PLATELET-DERIVED GROWTH FACTOR

At least seven related FGFs have been identified.148 These peptides have mitogenic effects on various cell types, including mesenchymal cells, and likely play an important role in organogenesis and neovascularization.149 Although not unique to the GI tract, PDGF is one of the most thoroughly studied growth factors. It is important for fibroblast growth, and its receptor is expressed in the liver and throughout the GI tract, where it appears to promote wound healing.

TREFOIL FACTORS

Trefoil factors (pS2, spasmolysin, and intestinal trefoil factor, also known as TTF1, 2, and 3, respectively) are a family of proteins expressed throughout the GI tract.150 They share a common structure, having six cysteine residues and three disulfide bonds, creating a cloverleaf appearance that stabilizes the peptide within the gut lumen. The pS2 peptide is produced in the gastric mucosa, spasmolysin is found in the antrum and pancreas, and intestinal trefoil factor is produced throughout the small and large intestines. These peptides are produced by mucous neck cells in the stomach or goblet cells in the intestine and are secreted onto the mucosal surface of the gut. It is likely that trefoil factors act on the apical surface of the epithelial cells, where they have growth-promoting properties on the GI mucosa.

OTHER G PROTEIN–COUPLED RECEPTORS

Other peptides signaling through GPCRs may also have growth-promoting effects. Three important examples include gastrin, CCK, and gastrin-releasing peptide (GRP). Gastrin stimulates the growth of enterochromaffin-like cells of the stomach and induces proliferation of the oxyntic mucosa containing parietal cells.151 Gastrin binds to CCK-2

Chapter 1  Gastrointestinal Hormones and Neurotransmitters receptors of the stomach and activates PLC and Ras pathways, which ultimately results in activation of protein kinase C and MAP kinase, respectively. MAP kinase, which can also be activated by tyrosine kinase receptors typical of growth factors, causes the phosphorylation of transcription factors that are involved in cellular proliferation. In some cells, cAMP and protein kinase A exert synergistic effects on cellular growth through activation of nuclear transcription factors such as cAMP-responsive element binding (protein) (CREB). However, in other cells, cAMP antagonizes proliferation. Therefore, depending on the cell type, the effects of growth factors such as EGF, IGF, and PDGF may be enhanced by hormones that stimulate cAMP production. Certain colon cancer cells possess CCK-2 receptors and respond to the proliferative effects of gastrin. Moreover, gastrin may be produced by some colon cancers, enabling it to exert an autocrine effect to promote cancer growth.152 Whether circulating gastrin initiates colon cancer development is unknown. CCK binds preferentially to the CCK-1 type receptor, which is abundant in gallbladder, the pancreas of many species, brain, and peripheral nerves of the gut. In the rodent, but not human, pancreas, CCK causes hypertrophy and hyperplasia of pancreatic acinar cells. Similar to the effects of gastrin, CCK activates phospholipase C and small GTP-binding proteins to activate MAP kinase. In animal models, CCK can promote pancreatic cancer growth.16 GRP (the mammalian analog of bombesin) was first recognized for its ability to stimulate gastrin secretion from the stomach. Neurons containing GRP are abundant in the oxyntic mucosal of the proximal stomach and, based on studies with a specific GRP antagonist, appear to play a major role in the cephalic phase of gastric acid secretion.153 It was later appreciated that GRP stimulates proliferation of G cells. GRP has received considerable attention for its growth-promoting effects on small cell lung cancer, pancreatic cancer, and certain colon cancers.154

TASTE RECEPTORS

The GI tract contains specialized taste receptor cells that detect chemicals and toxins. These cells are best characterized in the tongue, where they are concentrated in taste buds. Taste receptor cells can detect nutrients such as proteins, fats, sugars, and salt at submolar concentrations and other chemicals such as drugs and toxins at submicromolar concentrations.155 Detection of these chemical signals is important for several reasons. Distinguishing among chemicals can warn of dangerous chemical ingestants and induce a vomiting reflex, thus protecting the organism against poisoning. Alternatively, sensing various foods through taste can be pleasant and encourage food intake as well as facilitate digestion by stimulating salivary, gastric, and pancreatic secretions. Taste receptor cells contain a wide number of receptors and ion channels that serve as targets for different tastants.156 The membrane proteins can respond to molecules as simple as ions or as complex as fatty acids and proteins. Ions and amino acids activate taste receptor cells by directly interacting with ion channels. Such interactions often cause membrane depolarization and induce Ca2+ entry into the cell. Other types of molecules such as sweet and bitter tastants activate G protein–coupled receptors and stimulate production of second messengers, such as cAMP or inositol trisphosphate, which stimulate intracellular signaling processes. A large number of taste receptors have been identified. A family of receptors that detects sweet compounds, including certain l amino acids, are known as T1R1, 2, and 3. A larger

family, known as T2Rs, consisting of over 30 members, mediates bitter gustatory signals. The T1R and T2R families of receptors are G protein–coupled receptors that couple to α subunits, including α-gustducin and α-transducin. Interestingly, although taste receptors are abundant in the tongue, T2Rs, α-gustducin, and α-transducin have been identified in the gastric and intestinal mucosa and in the pancreas. In the intestine, T2Rs have been localized in some enteroendocrine and brush border cells. It is possible that chemical messengers stimulate T2Rs in these cells to secrete hormones or induce cellular responses, such as production of nitric oxide.157

INTRALUMINAL RELEASING FACTOR REGULATION OF GASTROINTESTINAL HORMONES Most GI hormones are secreted into the blood following the ingestion of a meal. However, the exact mechanism whereby luminal nutrients stimulate hormone secretion is unknown. Although the apical surface of most enteric endocrine cells is exposed to the intestinal lumen (“open cells”), it is unclear whether nutrients interact with specific receptors on the surface of endocrine cells or whether they are absorbed and then stimulate hormone secretion. It has been recognized that specific releasing factors for GI hormones are present in the lumen of the gut (Fig. 1-10). CCK was the first hormone shown to be regulated by an intraluminal releasing factor.158,159 Luminal CCK-releasing factor was purified from intestinal washings and shown to stimulate CCK release when instilled into the lumen of animals. Diazepam-binding inhibitor has also been shown to stimulate CCK release, as has a pancreatic peptide known as monitor peptide.160,161 Secretin may also be regulated by an intraluminal releasing factor.28 The existence of these releasing factors underscores the significance of bioactive peptides within the lumen of the gut.

CCK-RF

Trypsin

Food

Pancreas

Monitor peptide

CCK Figure 1-10.  Regulation of cholecystokinin (CCK) secretion by intraluminal releasing factors. Endocrine cells containing CCK are stimulated by trypsinsensitive releasing factors (CCK-RF) present in the lumen of the GI tract. Releasing factors secreted from the intestine are responsible for negative feedback regulation of pancreatic secretion. Under basal conditions, local trypsin inactivates CCK-RF; however, with ingestion of nutrients that compete as substrates for trypsin, CCK-RF is available to stimulate CCK secretion. Monitor peptide, a pancreatic releasing factor, may contribute to sustained CCK release and pancreatic secretion after a meal.

17

18

Section I  Biology of the Gastrointestinal Tract GASTROINTESTINAL PEPTIDES THAT REGULATE SATIETY AND HUNGER During a meal, ingested nutrients interact with cells of the mouth and GI tract. Endocrine cells of the stomach and small intestine possess receptors that are linked to the secretion of GI hormones. GI peptides (see Chapters 6 and 8) are then released into the surrounding space, where they exert paracrine actions or are taken up into the circulation, where they function as hormones.162 Each of these transmitters facilitates the ingestion, digestion, absorption, or distribution of nutrients that are essential for the organism. Some GI hormones control the size of an ingested meal and are known as satiety signals. Satiety hormones share several qualities.163 First, they decrease meal size. Second, blocking their endogenous activity leads to increased meal size. Third, reduction of food intake is not the result of an aversion to food. Fourth, secretion of the hormone is caused by ingestion of food that normally causes cessation of eating (Table 1-3). Most satiety signals interact with specific receptors on nerves leading from the GI tract to the hindbrain. CCK is one of the most extensively studied satiety hormones. In a time- and dose-dependent manner, CCK reduces food intake in animals and humans,164 an effect that is mediated by CCK-1 receptors residing on vagus nerve endings.165 The effect of CCK on food intake is a proven physiologic action because administration of a CCK receptor antagonist induces hunger and results in larger meal sizes. CCK also delays the rate at which food empties from the stomach, which may explain why the satiety actions of CCK are most apparent when the stomach is distended. Together, these findings indicate that CCK provides a signal for terminating a meal. GLP-1 is produced by L cells of the ileum and colon and is released in response to food in the intestine. Although the primary action of GLP-1 is to stimulate insulin secretion, it also delays gastric emptying. Moreover, infusion of GLP-1 increases satiety and produces feelings of fullness, thereby reducing food intake without causing aversion.166 GLP-1 receptors are found in the periventricular nucleus, dorsal medial hypothalamus, and arcuate nucleus of the hypothalamus, which are important areas in the regulation of hunger. Like CCK, central administration of GLP-1 suppresses food intake. PYY is also produced by L cells of the ileum and colon. Two forms of PYY are released into the circulation, PYY1-36 and PYY3-36. PYY1-36 binds to all subtypes of the neuropeptide Y family of receptors, whereas PYY3-36 has strong affinity for the Y2 receptor. When administered to animals, PYY3-36 causes a reduction in food intake, and mice lacking the Y2 receptor are resistant to the anorexigenic effects of PYY3-36, indicating that PYY3-36 signals satiety through this

Table 1-3  Gastrointestinal Peptides That Regulate Satiety and Food Intake Reduce food intake

Increase food intake

Cholecystokinin (CCK) Glucagon-like peptide-1 Peptide tyrosine tyrosine (PYY3-36) Gastrin-releasing peptide Amylin Apolipoprotein A-IV Somatostatin

Ghrelin

receptor.167 PYY3-36 has been shown in humans to decrease hunger scores and caloric intake.168 Interestingly, most of the GI peptide receptors involved in satiety are also found in the brain, where they mediate similar satiety effects. This may represent conservation of peptide signals that serve similar purposes. Leptin is referred to as an adiposity signal because it is released into the blood in proportion to the amount of body fat and is considered a long-term regulator of energy balance. Together with CCK, leptin reduces food intake and produces a greater reduction in body weight than either agent alone.86 Therefore, it appears that long-term regulators of energy balance can affect short-term regulators through a decrease in meal size, which may promote weight reduction. Hunger and initiation of a meal are intimately related. Ghrelin is intriguing because it is the only known circulating GI hormone that has orexigenic effects.102 Produced by the stomach, ghrelin levels increase abruptly before the onset of a meal and decrease rapidly after eating, suggesting that it signals initiation of a meal. Consistent with this role are studies demonstrating that administration of antighrelin antibodies or a ghrelin receptor antagonist suppresses food intake.169 It is not known if ghrelin is responsible for the hunger pains and audible bowel sounds that occur in people who are hungry. Bariatric surgery, in particular Roux-en-Y gastric bypass, is the most effective procedure for long-term weight loss in morbid obesity. Although it had been assumed that weight loss accompanying this procedure was the result of reduced gastric capacity and calorie malabsorption, recent evidence of reduced ghrelin release and exaggerated PYY release after a meal has suggested that hormonal factors may contribute to reduced calorie intake.170

ENTEROINSULAR AXIS GI hormones play an important role in the regulation of insulin secretion and glucose homeostasis. These hormones control processes that facilitate the digestion and absorption of nutrients, as well as disposal of nutrients that have reached the bloodstream. In particular, gut peptides control postprandial glucose levels through three different mechanisms: (1) stimulation of insulin secretion from pancreatic beta cells; (2) inhibition of hepatic gluconeogenesis by suppression of glucagon secretion; and (3) delaying the delivery of carbohydrates to the small intestine by inhibiting gastric emptying.171 Each of these actions reduces the blood glucose excursions that normally occur after eating. Approximately 50% of the insulin released after a meal is the result of GI hormones that potentiate insulin secretion.172 This interaction is known as the enteroinsular axis and the gut peptides that stimulate insulin release are known as incretins. The major incretins are GLP-1 and GIP. GLP-1 not only stimulates insulin secretion but also increases beta cell mass, inhibits glucagon secretion, and delays gastric emptying. GIP stimulates insulin secretion when glucose levels are elevated and decreases glucagonstimulated hepatic glucose production.173 Thus, on ingestion of a meal, glucose, as it is absorbed, stimulates GLP-1 and GIP secretion. Circulating glucose then stimulates beta cell production of insulin, and this effect is substantially augmented by incretins acting in conjunction with glucose to increase insulin levels. Postprandial hyperglycemia may also be controlled by delaying the delivery of food from the stomach to the small

Chapter 1  Gastrointestinal Hormones and Neurotransmitters Table 1-4  Gastrointestinal Peptides That Regulate Postprandial Blood Glucose Levels Stimulate Insulin Release Glucagon-like peptide-1 Glucose-dependent insulinotropic peptide Gastrin releasing peptide Cholecystokinin (potentiates amino acid–stimulated insulin release) Gastrin (in presence of amino acids) Vasoactive intestinal peptide (potentiates glucose-stimulated insulin release) Pituitary adenylate cyclase–activating peptide (potentiates glucosestimulated insulin release) Motilin Delay Gastric Emptying Cholecystokinin Amylin Secretin Inhibit Glucagon Release Amylin

intestine, allowing the rise in insulin to keep pace with the rate of glucose absorption. Several gut hormones that delay gastric emptying have been shown to reduce postprandial glucose excursions (Table 1-4).171 Amylin (islet amyloid polypeptide) is a 37–amino acid peptide synthesized primarily in the beta cells of the pancreatic islets together with insulin. Although it was originally recognized for its ability to form amyloid deposits in association with beta cell loss, it has more recently been found to suppress glucagon secretion, delay gastric emptying, and induce satiety.174 Insulin resistance in obese patients is associated with increased levels of both insulin and amylin. Type II diabetes mellitus is characterized by high circulating insulin levels and insulin resistance. In addition, insulin levels do not increase appropriately after a meal and significant hyperglycemia occurs, which is consistent with an impaired incretin effect. GIP secretion is preserved in type II diabetes; however, the insulinotropic effect of GIP is reduced.175 Although the precise cause is unknown, the defect in GIP-stimulated insulin release is most pronounced in the late phase of insulin secretion. In contrast to GIP, GLP-1 secretion has been shown to be reduced in insulinresistant type II diabetics. The lower GLP-1 levels are caused by impaired secretion rather than increased degradation of the hormone.176 Unlike GIP, the insulin response to infusion

of GLP-1 is preserved, indicating that the beta cell can respond normally to this incretin hormone. These observations suggest that GLP-1 administration could be a viable treatment for the hyperglycemia associated with diabetes.177 The growing evidence that beta cell failure may develop in type II diabetes supports the use of incretin hormones, such as GLP-1, or agents that delay GLP-1 degradation by the enzyme dipeptidyl peptidase-4 (DPP-4) to enhance beta cell function.178,179

KEY REFERENCES

Batterham RL, Cowley MA, Small CJ, et al. Gut hormone PYY(3-36) physiologically inhibits food intake. Nature. 2002; 418:650-4. (Ref 60.) Boehning D, Snyder SH. Novel neural modulators. Annu Rev Neurosci. 2003; 26:105-31. (Ref 119.) Burdick JS, Chung E, Tanner G, et al. Treatment of Ménétrier’s disease with a monoclonal antibody against the epidermal growth factor receptor. N Engl J Med 2000; 343:1697-701. (Ref 142.) Cummings DE, Weigle DS, Frayo RS, et al. Plasma ghrelin levels after diet-induced weight loss or gastric bypass surgery. N Engl J Med 2002; 346:1623-30. (Ref 100.) de Herder WW, Lamberts SW. Somatostatin and somatostatin analogues: Diagnostic and therapeutic uses. Curr Opin Oncol. 2002; 14:53-7. (Ref 75.) Drucker DJ. Enhancing incretin action for the treatment of type 2 diabetes. Diabetes Care 2003; 26:2929-40. (Ref 175.) Joseph IM, Zavros Y, Merchant JL, Kirschner D. A model for integrative study of human gastric acid secretion. J Appl Physiol. 2003; 94:160218. (Ref 70.) Miyawaki K, Yamada Y, Ban N, et al. Inhibition of gastric inhibitory polypeptide signaling prevents obesity. Nat Med. 2002; 8:738-42. (Ref 48.) Nakazato M, Murakami N, Date Y, et al. A role for ghrelin in the central regulation of feeding. Nature. 2001; 409:194-8. (Ref 94.) Pennefather JN, Lecci A, Candenas ML, et al. Tachykinins and tachykinin receptors: a growing family. Life Sci. 2004; 74:1445-63. (Ref 65.) Schlessinger J. Cell signaling by receptor tyrosine kinases. Cell. 2000; 103:211-25. (Ref 133.) Thomas RP, Hellmich MR, Townsend CM Jr, Evers BM. Role of GI hormones in the proliferation of normal and neoplastic tissues. Endocr Rev 2003; 24:571-99. (Ref 73.) Tracey KJ. The inflammatory reflex. Nature. 2002; 420:853-9. (Ref 123.) Vilsboll T, Zdravkovic M, Le-Thi T, et al. Liraglutide, a long-acting human glucagon-like peptide-1 analog, given as monotherapy significantly improves glycemic control and lowers body weight without risk of hypoglycemia in patients with type 2 diabetes. Diabetes Care 2007; 30:1608-10. (Ref 44.) Woods SC. GI satiety signals I. An overview of GI signals that influence food intake. Am J Physiol Gastrointest Liver Physiol 2004; 286:G7-13. (Ref 163.) Full references for this chapter can be found on www.expertconsult.com.

19

CHAPTE R

2

Mucosal Immunity Iris Dotan and Lloyd Mayer

CHAPTER OUTLINE Distinct Immune Responses in Gut-Associated Lymphoid Tissue  21 Controlled/Physiologic Inflammation  21 Oral Tolerance  22 Unusual Immunoglobulins of Gut-Associated Lymphoid Tissue  23 Immunoglobulin A  23 Immunoglobulin M  24 Other Immunoglobulins  24 Physiology of Gut-Associated Lymphoid Tissue: The Intestinal Barrier  24

Mucosal immunity refers to immune responses that occur at mucosal sites. The demands on the mucosal immune system are distinct from their systemic counterparts. At mucosal sites, the outside world is typically separated from the inner world by a single layer of epithelium. The mucosal immune system exists at a number of sites, including the respiratory tract (especially the upper respiratory tract), the urogenital tract, the mammary glands, and even the eye and ear. Regardless of the site, unique lymphoid and other cell populations are required to handle a wide array of environmental stimuli. Together, these sites are called mucosaassociated lymphoid tissue, or MALT.1-5 However, the site that is most often associated with mucosal immunity is the intestinal tract. The intestinal tract is unique in several aspects. In contrast to other mucosal sites, it is the least sterile, containing billions to trillions of microorganisms, mainly bacteria. These organisms, along with ingested food, represent an enormous antigenic load that must be tolerated to maintain the status quo in the intestine. This unusual environment and the demands associated with it have resulted in the development of a distinct immune system designated as gut-associated lymphoid tissue (GALT). The specific characteristics and peculiarities of GALT reflect the unique milieu in which it needs to function. To maintain homeostasis in the intestine, one of the most important tasks of GALT is to distinguish potentially harmful antigens, such as pathogenic bacteria or toxins, from ones that may benefit the body, such as those derived from food or commensal bacteria. To achieve homeostasis, unusual cell types, immunoglobulins (Igs), and secreted mediators need to function in a coordinated fashion. In contrast to the systemic immune system, whose focus is to act quickly within seconds of encountering a foreign antigen, the GALT is poised to respond but is predominantly tolerant, rejecting harmful antigens but allowing beneficial or harmless ones to persist without evoking immune responses such as allergic reactions or inflammation. The unique ways in which GALT performs in its demanding environment are the focus of this chapter, along with the consequences of

Functional Anatomy of Gut-Associated Lymphoid Tissue  25 Peyer’s Patches and Microfold Cells  25 Intestinal Epithelial Cells  25 Antigen Presentation in the Gut  26 Intestinal Mononuclear Cells  28 Intraepithelial Lymphocytes  28 Lamina Propria Mononuclear Cells  28 T Cell Differentiation  28 Dendritic Cells  29 Gut-Associated Lymphoid Tissue: Relevant Chemokines  29

abnormal GALT function that result in intestinal allergic or inflammatory diseases (discussed in other chapters).

DISTINCT IMMUNE RESPONSES IN GUT-ASSOCIATED LYMPHOID TISSUE As noted, the hallmark of mucosal, in contrast to systemic, immunity is suppression, exemplified by two linked phenomena, controlled or physiologic inflammation and oral tolerance. These processes are mediated by a unique anatomy, distinct resident cell populations, and selective antibody isotypes.

CONTROLLED/PHYSIOLOGIC INFLAMMATION

Within the lamina propria exist billions of activated plasma cells, memory T cells, memory B cells, macrophages, and dendritic cells.6,7 In fact, given the large surface area of the gastrointestinal (GI) tract and the resident cell populations that inhabit this space, the gut is the largest lymphoid organ in the body. Still, in contrast to activated lymphocytes in the peripheral immune system, significant inflammation is not present in the intestine. This phenomenon has been termed controlled/physiologic inflammation (Fig. 2-1). The entry and activation of the cells into the lamina propria is antigen-driven. Germ-free mice have few cells in their lamina propria. However, within hours to days following colonization with normal intestinal flora (without pathogens), there is a massive influx and activation of cells.8-11 Despite the persistence of an antigen drive (luminal bacteria), the cells fail to develop into aggressive inflammation-producing lymphocytes and macrophages. Bacteria or their products play a role in this persistent state of activation12 and likely contribute to the controlled inflammatory process as well. The failure to produce gastrointestinal pathology, despite the activation state of intestinal lymphocytes, is probably the consequence of regulatory mechanisms. The failure of lamina propria lymphocytes (LPLs) to generate “normal”

21

22

Section I  Biology of the Gastrointestinal Tract IELs

↓Cellular responses

Bacteria

DC Tight junctions Goblet cell

Treg

Defensins HBD-2, 3, 4 SlgA J Macrophage Plasma cell

Lymphocytes LPMC MAdCAM-1

α4β7 integrin Blood vessels

Figure 2-1.  Mechanisms for damping the mucosal immune response. The intestine uses a number of distinct mechanisms to dampen mucosal immune responses. The major source of antigen in the intestine is the commensal bacterial flora, but both innate and adaptive responses control local responses. Physical barriers such as mucins secreted by goblet cells and tight junctions between epithelial cells prevent invasion by luminal flora. Defensins such as HBD-2, -3, and -4 are thought to maintain the sterility of the crypt, whereas SIgA produced by local plasma cells prevents attachment and invasion by luminal bacteria, thereby reducing the antigenic load. Even in the face of antigenic challenge, the lymphocytes, macrophages, and dendritic cells in the intestine are programmed not to respond as a consequence of decreased expression of pattern recognition receptors (e.g., Toll-like receptors) and a decrease in the ability of lymphocytes to be activated through their antigen receptor. DC, dendritic cell; HBD, human β-defensin; IELs, intraepithelial lymphocytes; LPMC, lamina propria mononuclear cells; MAdCAM, mucosal addressin cell adhesion molecule; SIgA, secretory immunoglobulin A, a dimer with a connecting J chain; Treg, T regulatory cells (formerly known as suppressor T cells).

antigen-receptor mediated responses is an important factor in controlled inflammation (e.g., lack of expansion, despite a state of activation). LPLs respond poorly when activated via their T cell receptor (TCR), failing to proliferate although they still can produce cytokines.13,14 This is key to the maintenance of controlled inflammation.

ORAL TOLERANCE

The most recognized phenomenon equated with mucosal immunity and associated with suppression is oral tolerance.15-20 Oral tolerance can be defined as the active, antigen-specific nonresponse to antigens administered orally.18,21,22 Disruption of oral tolerance may result in food allergies (see Chapter 9) and food intolerances such as celiac disease (see Chapter 104). Part of the explanation for oral tolerance relates to the properties of digestion per se, where large potentially antigenic macromolecules are degraded so that potentially immunogenic substances are rendered nonimmunogenic. However, approximately 2% of dietary proteins enter the draining enteric vasculature intact.23 How does the body regulate the response to these potential antigens that have bypassed complete digestion? This is achieved by oral tolerance. Factors affecting the induction of oral tolerance are the age of the host, genetic factors, the nature of antigen, the tolerogen’s form, and dose. In addition, the state of the intestinal barrier affects oral tolerance and, when barrier function is reduced, oral tolerance decreases as well. Oral tolerance is difficult to achieve in the neonate, probably because of the rather permeable intestinal barrier that

exists in the newborn, as well as the immaturity of the mucosal immune system. Within 3 months of age (in the mouse), oral tolerance can be induced and many previous antibody responses to food antigens are suppressed. The limited diet in the newborn may further serve to protect the infant from generating a vigorous response to food antigens. Furthermore, the intestinal flora has been demonstrated to affect the development of oral tolerance. Probiotics such as Lactobacillus GG given to mothers before delivery and during lactation have provided protection against the development of atopic eczema in their offspring.24 The effects of probiotics on oral tolerance are probably mediated through modulation of cytokine responses,25 a positive effect on intestinal barrier function and restitution of tight junctions,26-27 suppression of intestinal inflammation via downregulation of Toll-like receptor (TLR) expression,28,29 and secretion of metabolites that may inhibit inflammatory cytokine production by mononuclear cells. The role of genetic factors in oral tolerance has been suggested in murine models in which certain strains develop tolerance more easily than others.30 The nature and form of the antigen also play a significant role in tolerance induction. Protein antigens are the most tolerogenic whereas carbohydrates and lipids are much less effective at inducing tolerance.31 The way the antigen is delivered is also critical. For example, a protein delivered in soluble form (e.g., ovalbumin) is tolerogenic, whereas aggregation of this protein reduces its potential to induce tolerance. This phenomenon may be associated with an alteration in the sites of antigen sampling.6 Exposure or

Chapter 2  Mucosal Immunity prior sensitization to an antigen through an extraintestinal route also affects the development of tolerance responses. Finally, the dose of antigen administered is critical to the form of oral tolerance generated. In mouse models, high doses of antigen are associated with clonal deletion or anergy.32 In this setting, tolerance is not transferable; transfer of T cells from tolerized animals does not lead to the transfer of tolerance. The mechanism underlying T cell deletion is possibly Fas-mediated apoptosis.33 On the other hand, low doses of antigen activate regulatory-suppressor T cells.34,35 Increasing numbers of such T cells occur, both in CD4 and CD8 lineages. Th3 cells were the initial regulatorysuppressor cells described as mediators of oral tolerance.35-37 These cells appear to be activated in Peyer’s patches and secrete transforming growth factor-β (TGF-β). This cytokine plays a dual role in mucosal immunity; it is the most potent suppressor of T and B cell responses while also promoting the production of IgA (TGF-β is the IgA switch factor).38-41 The production of TGF-β by Th3 cells elicited by low-dose antigen administration helps explain an associated phenomenon of oral tolerance, namely, bystander suppression. Whereas oral tolerance is antigen specific, the effector arm is antigen non-specific. If an irrelevant antigen is coadministered systemically with the tolerogen, suppression of T and B cell responses to that irrelevant antigen will also occur (hence, bystander suppression). Secreted TGF-β suppresses the response to the coadministered antigen. T regulatory 1 (Treg1, or Tr1) cells may also participate in bystander suppression and oral tolerance by producing interleukin-10 (IL-10), another potent immunosuppressive cytokine.42-44 Evidence for the activation of CD4+, CD25+ regulatory T cells during oral tolerance also exists, although their exact role in this process is still being investigated.45-49 Tolerance studies carried out in mice depleted of CD4+, CD25+ T cells coupled with neutralization of TGF-β have demonstrated that CD4+, CD25+ T cells and TGF-β together are involved in the induction of oral tolerance, partly through regulation of the expansion of antigen-specific CD4+ T cells.50 The ability to identify regulatory CD4+, CD25+ T cell subpopulations was enhanced by the recognition that these cells express the transcription factor FoxP3. Because not every cell within the CD4+, CD25+ population is a naturally occurring Treg cell, the ability to use FoxP3 as a marker of these Treg cells has been a major breakthrough in our ability to study these cells.51,52 Importantly, the absence of CD4+ T regulatory cell activity in mice results in inflammatory bowel disease (IBD), although this has not been demonstrated in humans.53-56 Preliminary data also support a role for antigen-specific CD8+ T cells in oral tolerance,57-61 as well as in the regulation of mucosal immune responses. Specifically, in vitro activation of human CD8+ peripheral blood T cells by normal intestinal epithelial cells (IECs) results in the expansion of CD8+, CD28− T cells with regulatory activity. Moreover, in the lamina propria of IBD patients, such CD8+, CD28− cells were significantly reduced, supporting a role for these epithelial-induced T regulatory (TrE) cells in the control of intestinal inflammation.62 Another factor affecting tolerance induction is the state of the intestinal barrier. In addition to the failure to generate tolerance in the neonatal period (when intestinal permeability is higher), several other states of barrier dysfunction are associated with aggressive inflammation and a lack of tolerance. During anaphylaxis, increased intestinal permeability caused by the disruption of tight junctions allows luminal antigens to pass through paracellular spaces.63-65 Treatment with interferon-γ (IFN-γ) can disrupt the mucosal barrier in mice and they fail to develop tolerance in response to

E E

E

L

L L

L

Figure 2-2.  M cell. This transmission electron micrograph from the noncolumnar region of the Peyer’s patch epithelium shows a cross-sectional view of a microfold (M) cell, as well as associated microvillus-covered epithelial cells and at least three lymphoid cells (L). Note the attenuated cytoplasm of the M cell (between arrows) that bridges the surface between microvillus-covered epithelial cells, forming tight junctions with them and producing a barrier between the lymphoid cells and the intestinal lumen (×9600). B, B cell; E, epithelial cell. (From Owen RL, Jones AL: Epithelial cell specialization within human Peyer’s patches: An ultrastructural study of intestinal lymphoid follicles. Gastroenterology 1974; 66:189-203.)

ovalbumin feeding. Even more interesting is the failure of N-cadherin–dominant negative mice to suppress mucosal inflammation (loss of controlled inflammation),66 possibly because of the enormous antigenic exposure resulting from the leaky barrier in these mice. Lastly, oral tolerance may also be associated with the cells serving as the antigen-presenting cell (APC; see later), as well as the site of antigen uptake. Orally administered reovirus type III is taken up in mice by microfold (M) cells expressing reovirus type III–specific receptors (Fig. 2-2).67 This epithelial uptake by M cells induces an active IgA response to the virus. Reovirus type I, on the other hand, infects intestinal epithelial cells (IECs) adjacent to M cells and this uptake induces tolerance to the virus. Thus, the route of entry (M cell versus IEC) of a specific antigen may dictate the type of immune response generated (IgA versus tolerance). Interestingly, poliovirus vaccine, one of the few oral vaccines effective in humans, binds to M cells, which may account for its ability to stimulate active immunity in the gut.68

UNUSUAL IMMUNOGLOBULINS OF GUT-ASSOCIATED LYMPHOID TISSUE IMMUNOGLOBULIN A

The unique antibody, secretory IgA (SIgA), is the hallmark of MALT-GALT immune responses (Fig. 2-3). Although IgG

23

24

Section I  Biology of the Gastrointestinal Tract Secretory component J chain Light chain

Heavy chain Figure 2-3.  Secretory immunoglobulin A (IgA) complex. Two IgA molecules are linked by a J chain and stabilized by secretory component (polymeric Ig receptor) to form dimeric secretory IgA.

Secretory IgA SC

Epithelial cell

mers are bound together by a J chain (also produced by plasma cells). Subsequently, the dimer binds to secretory component (SC), also known as the polymeric immunoglobulin receptor (pIgR), a highly specialized 55-kd glycoprotein produced by IECs. SC (pIgR) is expressed on the basolateral aspect of the IEC and binds only to dimeric IgA or to IgM (also polymerized with J chain; see later). Once bound to SC, SIgA is actively transported within vesicles to the apical membrane of the IEC. The vesicle fuses with the apical membrane and the SC-IgA complex is released into the lumen. Once in the lumen, SC serves its second function, protection of the SIgA dimer from degradation by luminal proteases and gastric acid. SIgA and SIgM are the only antibodies that can bind SC and therefore withstand the harsh environment of the GI tract. In addition to its unique form, SIgA is also unique as an immunoglobulin in that it is anti-inflammatory in nature. SIgA does not bind classic complement components but rather binds to luminal antigens, preventing their attachment to the epithelium or promoting agglutination and subsequent removal of the antigen in the mucus layer overlying the epithelium.69,72 This binding of luminal antigens by SIgA reflects immune exclusion, as opposed to the non­ specific mechanisms of exclusion exerted by the epithelium, the mucous barrier, proteolytic digestion, and other mechanisms.

IMMUNOGLOBULIN M

As noted, IgM is the other antibody capable of binding SC (pIgR). Like IgA, IgM uses J chains produced by plasma cells to form polymers—in the case of IgM, a pentamer. SC binds to the Fc portion of the antibody formed during polymerization. The ability of IgM to bind SC may be important in individuals with IgA deficiency in which secretory IgM (SIgM) may compensate for the absence of IgA in the lumen.

OTHER IMMUNOGLOBULINS

Polymeric Ig receptor

IgA plasma cell Polymeric IgA IgA

J chain

Figure 2-4.  Assembly and secretion of dimeric immunoglobulin A (IgA). IgA and J chain produced by IgA-committed plasma cells (bottom) dimerize to form polymeric IgA, which covalently binds to membrane-bound polymeric Ig receptor produced by epithelial cells (top). This complex is internalized, transported to the apical surface of the epithelial cell, and secreted into the lumen. SC, secretory component.

is the most abundant isotype in the systemic immune system, IgA is the most abundant antibody in mucosal secretions.69-71 In fact, given the numbers of IgA-positive plasma cells and the size of the MALT system, IgA turns out to be the most abundant antibody in the body. SIgA is a dimeric form of IgA produced by plasma cells in the lamina propria and transported into the lumen by a specialized pathway through the intestinal epithelium (Fig. 2-4). Two IgA mono-

Whereas SIgA is the major antibody isotype produced in GALT, IgG has been detected as well.73-74 The neonatal Fc receptor (FcRN), expressed by IECs, might serve as a bidirectional transporter of IgG75,76 and may be important in control of neonatal infections and IgG metabolism. In IBD, marked increases in IgG in the lamina propria and lumen have been observed.77 Even IgE production may play an important role in intestinal diseases in GALT. CD23 (low-affinity IgE Fc receptor) has been reported to be expressed by IEC. One model has suggested that CD23, or FcεRII, may play a role in facilitated antigen uptake and consequent mast cell degranulation in food allergy. In this setting, IgE transcytosis and mast cell degranulation may be associated with fluid and electrolyte loss into the lumen, an event that is intimately associated with allergic reactions in the gut and airways.78,79

PHYSIOLOGY OF GUT-ASSOCIATED LYMPHOID TISSUE: THE INTESTINAL BARRIER The cells, structures, and mediators separating the intestinal lumen from the lamina propria function as a physical barrier. However, this physical barrier is a biologically active structure that constantly interacts with its everchanging environment. The intestinal barrier changes not only on a daily basis but also over the years. Many barrier mechanisms are not fully developed at birth, and evidence exists to support less restricted antigen transport in neonates compared with adults, specifically in animals.

Chapter 2  Mucosal Immunity Physiologic factors operative in the upper GI tract influence the antigenic load that reaches the major sites of GALT in the small and large bowels. Detailed exploration of these factors are beyond the scope of this chapter and are discussed elsewhere, but include proteolysis, gastric acidity, and peristalsis. The mucous coat lining the intestinal tract is composed of a mixture of glycoproteins (mucins). The protein core of mucins is enriched in serine, threonine, and proline residues, and carbohydrate moieties are attached via Nacetylgalactosamine residues. At least six different mucin species have been identified,80 each with a distinct carbohydrate and amino acid composition. Mucus protects the intestinal wall by several mechanisms. Its stickiness and competitive binding to glycoprotein receptors decrease the ability of microorganisms to penetrate the intestine.81 It also generates a stream that moves luminal contents away from epithelial cells. Underneath the mucous layer, the physical barrier that prevents penetration of antigen across the intestinal epithelium consists of the actual epithelial cell (the transcellular route) and the tight intercellular spaces (the paracellular route) regulated by tight junction (TJ) complexes (e.g., zona occludens) and the subjunctional space.82 Of the two structures, TJs have the greater role in preventing macromolecular diffusion across the epithelium, because these junctions exclude almost all molecules present in the lumen (see Chapter 96). The barrier formed by the TJ is a dynamic structure, preserved even when epithelial cells themselves are damaged; this feature might be crucial for the prevention of intestinal inflammation (e.g., as seen in idiopathic IBD). The epithelial cells themselves serve as a physical barrier in several ways—their microvilli are at a distance of about 25 nm from each other and are negatively charged. Thus, a negatively charged luminal molecule would be repelled from passage even if its diameter were well below 25 nm. However, intact antigens may traverse the epithelium by fluid phase endocytosis and enter the circulation.83

FUNCTIONAL ANATOMY OF GUT-ASSOCIATED LYMPHOID TISSUE To accomplish the two major goals of the mucosal immune system in the intestine (maintenance of homeostasis and clearance of pathogens), several key features have been identified. Compartmentalization of cells into distinct regions and sites, despite being millimeters away from each other, is a hallmark of the GALT. Cell populations and the immune response in the epithelium, subepithelial region, lamina propria (LP), Peyer’s patches, and mesenteric lymph node (MLN) may differ substantially. The cells residing in these compartments differ not only topographically but also phenotypically and functionally, depending on the anatomic site in GALT. Cells with distinct phenotypes and functions are attracted to specific sites in GALT.

PEYER’S PATCHES AND MICROFOLD CELLS

The follicle-associated epithelium (FAE), which contains microfold (M) cells, is a specialized epithelium overlying the only organized lymphoid tissue of GALT, Peyer’s patches (PPs). The M cell, in contrast to the adjacent absorptive epithelial cell, has few microvilli, a limited mucin overlayer, a thin elongated cytoplasm, and a shape that forms a pocket surrounding subepithelial, T, B, macrophages, and dendritic cells (see Fig. 2-2). M cells are capable of taking up large particulate antigens from the lumen and transport-

ing them intact into the subepithelial space.84-86 M cells contain few lysosomes, so little or no processing of antigen occurs.87 M cells are exposed to the lumen, thus having a larger area for contact with luminal contents than adjacent epithelial cells. The M cell expresses several unique lectinlike molecules that help promote binding to specific pathogens—the prototype being poliovirus.88 Antigens that bind to the M cell and are transported to the underlying PP generally elicit a positive (SIgA) response. Successful oral vaccines bind to the M cell and not to the adjacent epithelium. Thus, M cells appear to be critical for the positive aspects of mucosal immunity. The M cell is a conduit to PPs. Antigens transcytosed across the M cell and into the subepithelial pocket are taken up by macrophages and DCs and carried into PPs. Once in the patch, TGF-β–secreting T cells promote B cell isotype switching to IgA.89 Importantly, there is a clear relationship between M cells and PPs. The induction of M cell differentiation has been shown to be dependent on direct contact between the epithelium and B lymphocytes in PPs.90 M cells do not develop in the absence of PPs. For example, M cells have not been identified in B cell–deficient animals in whom there are no PPs.91 Even though M cells and PPs may be involved in oral tolerance,92-94 PP-deficient mice are capable of developing tolerance after oral administration of soluble antigen.95 After activation in PPs, lymphocytes are induced to express specific integrins (α4β7), which provide a homing signal for mucosal sites (where the ligand is MadCAM-1).96-98 Cells then travel to MLNs and subsequently into the main intestinal lymphatic drainage system, the thoracic duct, and finally into the systemic circulation (Fig. 2-5). There, mucosally activated lymphocytes with their mucosal addressins circulate in the bloodstream to exit in high endothelial venules in various mucosal sites. Those bearing α4β7 molecules exit in the MALT-GALT lamina propria where they undergo terminal differentiation. Chemokines and their receptors (see later) as well as adhesion molecules and ligands may help direct this trafficking pattern.

INTESTINAL EPITHELIAL CELLS

The epithelium is composed of a single layer of columnar cells. These IECs are derived from the basal crypts and differentiate into absorptive villous or surface epithelium, secretory goblet cells, neuroendocrine cells, or Paneth cells (see Chapter 96). In addition to their function as a physical barrier in GALT discussed earlier, IECs contribute to innate and adaptive immunity in the gut and may play a key role in maintaining intestinal homeostasis.

Antigen Trafficking Across Intestinal Epithelial Cells

The ability of intact antigen to cross the lipid bilayer at the surface of the IEC (underneath the microvilli) is limited. However, invagination of apical membranes occurs regularly, allowing macromolecules to be carried into the cell within membrane-bound compartments. Binding to the surface of the cell depends on the structure of the antigen and the chemical composition of the microvillous membrane. For instance, bovine serum albumin binds less efficiently to the surface of the IEC than bovine milk protein and, as a consequence, is transported less efficiently.99 In addition, structural alterations in an antigen caused by proteolysis might also affect its binding, because this will change the physicochemical characteristics of the molecule.100 Several factors influence the transport of antigens from the apical to the basolateral surface of IECs. The rate of

25

26

Section I  Biology of the Gastrointestinal Tract Intestine Tonsil

Villi

Lumen

Follicle-associated epithelium

Mammary gland

Peyer’s patch Figure 2-5.  Mucosal lymphocyte migration. Following antigenic stimulation, T and B lymphocytes migrate from the intestine to the draining mesenteric lymph nodes, where they further differentiate and then reach the systemic circulation via the thoracic duct. Cells bearing the appropriate mucosal addressins then selectively home to mucosal surfaces which constitute the common mucosal immune system, including the intestine.

Circulation Mesenteric lymph node

vesicular passage to the basolateral membrane depends on the rate of endocytosis, the proportion of vesicles trafficking to the lysosome, and the speed of travel of membrane-bound compartments. Lysosomally derived enzymes determine the rate of breakdown of products contained in membrane compartments. These include proteases such as cathepsins B and D (found throughout the length of the intestine, particularly in the mid and distal thirds of the small intestine), as well as those enzymes that catalyze carbohydrate breakdown, such as acid phosphatase and mannosidase. It is the degree to which the organellar contents encounter such enzymes (in the lysosome or in endocytic vesicles) that determines the rate of intracellular destruction of macromolecules.101 Although cathepsins are capable of catalyzing antigens, they may not completely digest the protein, which may require further proteolysis by peptidases in the cytoplasm.

Recognition of Pathogen-Associated Molecular Patterns by Pattern Recognition Receptors

Classic APCs in the systemic immune system possess the innate capacity to recognize components of bacteria and viruses, called pathogen-associated molecular patterns (PAMPs). Receptors for these PAMPs are expressed on the cell surface (e.g., TLRs) and inside the cell (e.g., NOD2 [see later]). Despite the fact that IECs live adjacent to large numbers of luminal flora, IECs retain the ability to recognize components of these bacteria. In general, proinflammatory responses are normally down-regulated (i.e., expression of the lipopolysaccharide [LPS] receptor, TLR4, is absent) and expression of some of these pattern recognition receptors is maintained, such as TLR5, which recognizes bacterial flagellin. TLR5 is expressed basolaterally, so it is poised to identify organisms such as Salmonella species that have invaded the epithelial layer.102 Following invasion and engagement of TLR5, the IEC is induced to secrete a broad array of cytokines and chemokines that attract inflammatory cells to the local environment to control the spread of infection. In contrast to invading pathogens, some bacteria are probiotic and induce the IEC to produce anti-inflammatory cytokines (e.g., IL-10) and to increase the expression of peroxisome proliferator-activated receptor-γ (PPAR-γ).103 Furthermore, other bacterial products help promote the barrier and IEC differentiation (e.g., products of Bacteroides thetaiotaomicron).

Genitourinary tract Bronchus-associated lymphoid tissue

Thoracic duct

The significance of the ability of IECs to recognize PAMPs via surface TLRs, such as TLR5, or via intracellular nuclear oligomerization domain 1, 2 (NOD1, 2), has been increasingly recognized over the past decade. The latter ability has been shown to contribute to intestinal inflammation, because about 25% of patients with Crohn’s disease have mutations in the NOD2-CARD15 gene, interfering with their ability to mount an appropriate immune response to bacterial stimuli (discussed further in Chapter 111).104-108 In addition, TLRs that are weakly expressed by normal IECs are expressed at higher levels on IECs obtained from patients with IBD.109 Expression of different TLRs by IECs, as well as their contribution to innate and adaptive T and B cell responses in intestinal inflammation and homeostasis, has been demonstrated in several murine models.110,111 TLR expression by professional APCs is also down-regulated in the lamina propria. This finding, along with others described earlier, contribute to the immunologic nonresponsiveness of GALT. The importance of TLR and NOD2CARD15 expression and signaling in the intestine has been reviewed.112

ANTIGEN PRESENTATION IN THE GUT Effective immune responses to antigenic proteins depend on the antigen being presented to T lymphocytes by APCs. APCs internalize, digest, and then couple a small fragment of the antigen to a surface glycoprotein (major histocompatibility complex [MHC] class II, or HLA-D in humans) that eventually interacts with a TCR. Several cells in GALT can act as APCs, including B cells, macrophages, and dendritic cells. The ability of these cells to present antigen depends on the expression of class II MHC molecules on their surface. Class II MHC molecules are also present on the epithelium of the normal small intestine and, to a lesser extent, colonocytes in humans113 and rodents.114 In vitro studies have demonstrated that enterocytes isolated from rat and human small intestine can present antigens to appropriately primed T cells.115-117 This raises the possibility that IECs might present peptides to GALT T cells beneath the epithelium. Thus, IECs are capable of antigen trafficking and proc­ essing as well as antigen presentation to cells in the lamina propria in the appropriate context. Importantly, increased expression of MHC class II molecules by IECs has been

Chapter 2  Mucosal Immunity Luminal bacteria Food or bacterial antigen Inflammation ↑paracellular transport

Tight junction Stress response/ homeostasis? Autoregulatory or suppressor population?

γδ TCR

MICA/MICB

αβ TCR CD2

CD4+ or CD8+ IEL αβ TCR CD8 CD8+ CD28− IEL

β2m CD1d gp180

αβ TCR CD28 CD4+ CD25+ LPL

CD4+ IEL

MHC class I CD1d β2m gp180

αβ TCR CD8

MHC class II CD86 CD58

αβ TCR CD8

αβ TCR CD2

CD8+ CD28− LPL

CD4+ LPL

Homeostasis?

CD8+ IEL

Cytolytic or suppressor activity?

αβ TCR CD28+ CD4+ LPL

Tolerance? Autoregulatory or suppressor population?

Inflammation?

Figure 2-6.  Normal intestinal epithelial cell (IEC). The IEC is shown to express classic MHC molecules (classes I and II) that have the potential to present conventional antigen to local T cell populations and a broad array of nonclassic class I molecules (e.g., CD1d, MICA/MICB, and β2m [shown in the figure] and MR-1, ULBP, HLA-E, and FcRn [not shown]), which have the potential to present unconventional antigens to unique T cell populations. In addition, alternate pathways of activation appear to be functional in the intestine (e.g., activation via a CD58-CD2 interaction) and classic costimulatory molecules are not expressed on IECs, although CD86 may be induced in patients with ulcerative colitis. Other members of the B7 family are expressed (B7h and B7H-1) and may play a role in local T cell activation. β2m associates with MHC class I, CD1d, HLA-E, HLA-G, and FcRn. β2m, β2 microglobulin; gp180, membrane glycoprotein 180 (a CD8 ligand); IEL, intraepithelial lymphocyte; LPL, lamina propria lymphocyte, MHC, major histocompatibility complex; MICA/MICB, MHC class I–related chains A and B; TCR, T cell receptor.

reported in patients with IBD.118,119 Such overexpression would be expected to increase their potential to activate T lymphocytes and this has been reported.120,121 Drugs used to treat patients with IBD, such as 5-aminosalicylic acid (5-ASA) preparations, may reduce IEC MHC class II expression on IEC.122 In addition, IECs from normal individuals or IBD patients express a variety of costimulatory molecules required for T cell activation (Fig. 2-6). These include intercellular adhesion molecule 1 (ICAM-1), which binds to leukocyte function-associated antigen 1 (LFA-1) on the T cell and to B7 (CD80) on the APC. B7, which binds CD28 and cytotoxic T-lymphocyte antigen 4 (CTLA-4),123,124 has been shown to be expressed by IECs of patients with ulcerative colitis (UC). Interestingly, unique expression of these costimulatory molecules by IECs may be involved in the distinct regulation of mucosal responses. Failure of naive T cells to engage CD28 by B7 family members may result in T cell tolerance. This may be less of an issue in GALT, in which cells express the memory phenotype.125,126 Because small intestine IECs do not express B7-1 (CD80),127 activation of naive T cells by IECs is not probable, aiding in the down-regulation of T-cell

responses. On the other hand, increased expression of B7 during intestinal inflammation may serve to augment T cell stimulation.128 MHC class I and nonclassic class I molecules are also expressed by IECs. Thus, antigen presentation to unique T cell populations is possible and has been reported by several groups.116,129-135 Specifically, CD1d expressed on human IECs is able to present antigen, in a complex with CEACAM5, to CD8+ T cells.136-140 The role of other nonclassical class I molecules expressed by IECs (e.g., MR-1, TL, Hmt-1, MICA/B, HLA-E, HLA-G) is still unclear.141-144 In humans, IECs specifically activate CD8+ regulatory T cells.116 These regulatory cells may be involved in local tolerance and interaction with intraepithelial lymphocytes, which are CD8+ T cells (see later). The role of IECs in the regulation of mucosal immunity is best demonstrated in studies with tissues obtained from patients with IBD. IECs derived from IBD patients, in contrast to normal IECs, stimulate CD4+ T cells rather than regulatory CD8+ cells.120-121,145 Furthermore, oral antigen administration does not result in tolerance in IBD patients, but instead produces active immunity.146

27

28

Section I  Biology of the Gastrointestinal Tract INTESTINAL MONONUCLEAR CELLS Juxtaposed to the IEC are two unusual intestinal lymphocyte populations, each very different from the other—IELs and LPLs. The clear compartmentalization of these two distinct cell populations correlates with their ability to respond to distinct microenvironmental cues.

INTRAEPITHELIAL LYMPHOCYTES

IELs in the small intestine are largely (>98%) T cells and are characterized by being mostly CD8+ cells,147-154 including CD8+ αα T cells and CD4+, CD8+ double-positive and CD4−, CD8− double-negative cells. Greater numbers of these cells also express the γδ TCR, in contrast to the αβ TCR expressed by T cells in the systemic immune system.155 Approximately half of murine small bowel IELs express the γδ TCR,156 whereas the murine and human large intestines contain primarily αβ CD4+ or CD8+ T cells similar to those found in the systemic immune system. Based on their phenotype, IELs have been classified into two subsets, a and b. Subpopulation a includes TCR αβ T cells selected in the thymus with conventional MHC class I and II expression. Subpopulation b includes TCR αβ CD8+ αα, TCR γδ DP, and TCR γδ DN cells. Both subpopulations have been shown to be cytolytic, killing via granzyme or by engagement of Fas. They both also secrete Th1 cytokines such as IL-2, TNF-α, and IFN-γ. However, antigen-specific type a IELs can transfer protection against a variety of pathogenic organisms, whereas type b IELs are not able to transfer immunologic protection and do not possess immunologic memory. This is possibly caused by the activation of type b IELs by IECs in situ by nonclassical MHC molecules, rather than by the polymorphic MHC-expressed molecules on professional APCs that activate type a IELs.156 IELs express a variety of activation markers and are CD45 RO+ (memory cells). IELs also express GALT-specific integrin αEβ7.157-158 Expression of this integrin is induced by TGF-β, and its ligand on IECs is E-cadherin, which is involved in cell signaling and cytoskeletal rearrangement.158 When isolated, IELs are difficult to activate through their TCR and they barely proliferate, even in response to potent stimuli.153 On the other hand, they may be activated by alternative pathways, such as via CD2. Type a IELs secrete cytokines such as IL-7 that are dif­ ferent from the ones secreted by their peripheral blood counterparts.152,159-161 Functionally it has been suggested that IELs potentially kill epithelial cells that have undergone some form of stress, such as infection, transformation, or invasion by other cells.154-156,162 Alternatively, it has been proposed that IELs are active in suppressing local immunity, although the evidence that they actually function in luminal antigen recognition is weak. IELs do not travel in and out of the epithelium. Rather, the epithelial cells grow over the IELs as they move from the crypt to the surface. Thus, IELs likely serve as sentinels for epithelial integrity.

LAMINA PROPRIA MONONUCLEAR CELLS

The LP is the major effector site in GALT. Recently, it has been suggested that the LP may also be an inductive site, because antigen presentation by professional and nonprofessional APCs may occur in the LP itself. The LP is also considered a graveyard for activated lymphocytes. Lamina propria lymphocytes (LPLs) are more prone to undergo apoptosis compared with their peripheral counterparts.

This may be a regulatory mechanism limiting the potentially inflammatory effects of activated lymphocytes. Indeed, a major defect reported in Crohn’s disease (CD) is the resistance of LPL to undergo apoptosis when activated inappropriately (see later). Clearly, GALT operates under a distinct set of rules compared with the systemic immune system. This is reflected not only in its functional anatomy (no organized structure) but also in its responses and regulation. As noted, highly specialized cells mediate these effects, some detected only in GALT. Lamina propria mononuclear cells (LPMNCs) are a heterogeneous group of cells.163,164 The most populous cell type is the IgA-positive plasma cell, but there are more than 50% T and B cells (together comprising the LPL population), macrophages, and dendritic cells. In contrast to IELs, which express αEβ7, LPLs express the mucosal addressin α4β7. Similar to IEL, they express an activated memory phenotype and do not proliferate in response to engagement of the TCR. Alternate pathways of LPL activation are mainly via CD2 and CD28.159,165,166 Down-regulating the ability of LPLs to respond to stimulation via the TCR (i.e., antigen) may be another mechanism involved in dampening immune responses to normal luminal contents, along with the increased tendency for LPLs to undergo apoptosis if activated inappropriately. The mechanism underlying this latter apoptotic phenomenon possibly relates to engagement of the death receptor Fas and its ligand on activated LPLs, and by the imbalance between the intracellular anti- and pro-apoptotic factors, Bcl2 and Bax. Defects in this proapoptotic balance have been reported in patients with Crohn’s disease.167,168 These observations all contribute to the normal scenario in the lamina propria, called controlled physiologic inflammation (see earlier). This state of inflammation is the norm in the gut, whereas it would be considered to be indicative of disease in any other organ. When regulatory mechanisms go awry—an increase in cell recruitment coupled with a decrease in apoptosis—the result is uncontrolled inflammation, such as that observed in IBDs.

T CELL DIFFERENTIATION

As noted, there is an organized lymphoid structure in the LP, the Peyer’s patch (see Fig. 2-5). There, B and T lymphocytes interact with antigen sampled via M cells in the FAE. Activation and maturation of T lymphocytes from naive Th0 cells to distinct biased subpopulations is strongly influenced by the microenvironment. Specifically, contact with dendritic cells (DCs), professional APCs within GALT, and their secreted mediators will skew T lymphocytes to one of several effector cells. IL-2, IFN-γ and TNF-α–secreting T helper 1(Th1) cells develop from Th0 cells when DCs secrete the IL-12/p35-40 heterodimer.169 This heterodimer induces activation and phosphorylation of the transcription factor STAT-4 (signal transducer and activator of transcription factor 4).170 STAT-4 in turn induces IFN-γ expression and production. IFN-γ then induces activation of STAT-1 and consequently of T-bet; this is the master transcription factor that induces Th1 cytokine production and IL-12 receptor β2 expression while simultaneously suppressing Th2 cytokine production. Thus, a cycle promoting Th1 and suppressing Th2 responses is created. Overactivation of T-bet is possibly an essential step for Th1-mediated mucosal diseases, such as those seen in some patients with Crohn’s disease.170 Another Th1-promoting cytokine is IL-18, mediating its effects by augmenting IL-12 receptor β2 expression on T cells and by AP-1-(c-fos/c-jun)–dependent transactivation

Chapter 2  Mucosal Immunity of the IFN-γ promoter. It also activates nuclear factor κB (NF-κB) in T cells.169 In contrast, when IL-4 is secreted, STAT-6 is activated, followed by activation of the transcription factor GATA-3, which is capable of promoting the expression of several Th2 cytokines, including IL-4, IL-5, and IL-13.171 In addition to IL-4, IL-13 also plays an important role in Th2 development and IgE synthesis in an IL-4–independent fashion. These Th2 cytokines, which also include IL-6, -9, and -10, appear to play a role in the development of food allergies (see Chapter 9). IL-5 induces B cells expressing surface IgA to differentiate into IgA-producing plasma cells. IL-6 causes a marked increase in IgA secretion, with little effect on IgM or IgG synthesis.172 Thus, in the normal state in GALT, a Th2 bias might exist. However, despite the presence of classic Th1 and Th2 cells, the dominant tone in the intestine is one of regulation. Regulatory T cells, such as Th3 and Tr1 cells (see earlier, “Oral Tolerance”), develop in an environment in which IL-10 is predominant. Th3 and Tr1 cells secrete the regulatory cytokines TGF-β and IL-10, respectively. In addition to these regulatory cells, other cells such as CD4+, CD25+, Foxp3+, and even CD8+ T regulatory cells have been identified. All these cell populations may be involved in the induction of oral tolerance and controlled inflammation in GALT (see earlier).

DENDRITIC CELLS

DCs play an important role in tolerance and immunity in the gut. DCs continuously migrate within lymphoid tissues and present self antigens (likely from dying apoptotic cells to maintain self-tolerance) as well as non– self-antigens.173 Within the lamina propria of the distal small intestine, DCs express the chemokine receptor CX3CR1 and form transepithelial dendrites that allow direct sampling of luminal antigen.174 It has been suggested that IECs expressing chemokine (C-C motif) ligand 25 (CCL25), the ligand for chemokine (C-C motif) receptor 9 (CCR9) and for CCR10, attract DCs to the small bowel, whereas CCL28, the ligand for CCR3 and CCR10, attracts DCs to the colon.175-177 DCs process internalized antigens more slowly than macrophages,178 which probably contributes to local tolerance.179,180 Tolerance induction by DCs is associated with their degree of maturation at the time of antigen presentation to T cells (e.g., immature DCs activate Tregs), downregulation of costimulatory molecules CD80 and CD86, production of the suppressive cytokines IL-10, TGF-β and IFN-α, and interaction with the costimulatory molecule CD200.181-183

GUT-ASSOCIATED LYMPHOID TISSUE: RELEVANT CHEMOKINES Many of the chemokines secreted in GALT are produced by IECs, one more piece of evidence for their active participation in the regulation of intestinal immune responses. This is especially true in IBDs, in which the secretion of IECderived chemokines and cytokines is increased, mainly because of enhanced bacterial translocation and IFN-γ production, contributing to the augmentation of mucosal inflammation. Of the chemokines secreted, those secreted by IECs have the capacity to attract inflammatory cells, such as lymphocytes, macrophages, and DCs. Chemokine (C-C motif) ligand 5 (CCL5, formerly called RANTES), secreted predominantly by macrophages, can be

produced by human IECs as well.184 CCL5 may have a role in innate as well as adaptive mucosal immunity185 and, interestingly, increased CCL5 expression has been demonstrated in the mucosa of patients with IBD.186-189 Several CXC cytokines are constitutively expressed by lymphocytes, endothelial cells, and human colonic IECs.190,191 These include CXCL9, also known as monokine induced by interferon-γ (MIG), CXCL10, also known as interferon-γ inducible protein 10 (IP-10), which is a chemokine that appears to promote Th1 responses and therefore may be relevant in Crohn’s disease, and CXCL11, also known as IFN-γ–inducible T cell α chemoattractant (ITAC). Expression of CXCL9, 10, and 11, and their polarized basolateral secretion, increases after IFN-γ stimulation. CXC chemokines attract Th1 cells expressing high levels of CXCR3 (CXC receptor 3).192 They also contribute to natural killer (NK) T cell chemotaxis and increased cytolytic responses193 and activate subsets of DCs.194 By attracting CD4+ Th1 cells that produce IFN-γ, up-regulation of expression and secretion of CXC chemokines occurs, because IECs express IFN-γ receptors. This appears to contribute to a positive feedback loop that may be relevant in inflammatory states, specifically in IBD and celiac disease. Importantly, blockade of the CXCR3-CXCL10 axis has been shown to be beneficial in ameliorating murine colitis195 and is currently being investigated in human IBD. In contrast to the inflammation-related CXCR3 receptor, a tissue-specific chemokine receptor, CCR9, is constitutively expressed on small bowel IELs and LPLs.196-198 CCL25, also known as thymus-expressed cytokine (TECK), is the ligand for CCR9 and is differentially expressed in the jejunal and ileal epithelium, where levels of expression decrease from the crypt up to the villous epithelium.199 In murine models, it was shown that CCL25-CCR9 is associated with selective localization of MLN-activated CD8 αβ+ lymphocytes coexpressing αEβ7 to the small intestine.200 CCL25 expression by IECs has been shown to be increased in inflamed small bowel in Crohn’s disease patients, with increased CCR9 expression by peripheral blood lymphocytes (PBLs) and decreased expression by LPLs,197 thus supporting its role in the specific attraction of peripheral lymphocytes to the small bowel in CD. This key chemokine/chemokine receptor pair (CCL25-CCR9) has also been used as a target for therapeutic intervention in CD, with preliminary positive results.201 Fractalkine is a unique chemokine expressed by IECs. It combines the properties of chemokines and adhesion molecules. Fractalkine attracts NK cells, monocytes, CD8+ T lymphocytes and, to a lesser extent, CD4+ T lymphocytes that express CX3CR1.202 Expression of fractalkine is increased in CD, specifically in the basolateral aspect of IECs.203 Mucosa-associated epithelial chemokine (MEC) may also have a role in intestinal immunity. This chemokine and its receptors, CCR3 and CCR10, are expressed by colonic IECs. CD4+ memory lymphocytes and eosinophils are attracted by MEC in vitro, although its function in vivo has not yet been demonstrated.204 MDC-CCL2 is a chemokine that is constitutively expressed and secreted by colonic IECs. MDC-CCL2 is unique in that it attracts CCR4+ Th2 cytokine-producing lymphocytes. Polarized basolateral secretion of MDC-CCL2 from stimulated colonic IEC lines has been reported.205 The specific recruitment of lymphocytes that preferentially secrete anti-inflammatory (Th2) cytokines supports a role for the IEC in orchestrating normal mucosal homeo­stasis, and adds to the accumulating evidence that these cells possess the ability to regulate mucosal immune responses.

29

30

Section I  Biology of the Gastrointestinal Tract The chemokine CCL20, also known as macrophage inflammatory protein-3α (MIP-3α), is unique in its ability to attract immature DCs specifically, as well as memory CD4+ T lymphocytes.206-208 CCL20 is also expressed and produced in the human small intestine, mainly in the FAE, and by colonic IECs, and has been suggested to be the mediator of lymphocyte adhesion to the α4β7 ligand MAdCAM1.206 CCL20 expression and secretion are increased in colonic IECs derived from IBD patients.209 Its stimulated secretion is polarized to the basolateral compartment, supporting its ability to attract immune cells into the LP. Mucosal memory T cells and IECs express CCR6, the cognate receptor for CCL20. CCR6 and CCR9 are coexpressed in T cells expressing the α4β7 integrin, characteristic of mucosal lymphocytes. This may suggest that in inflammatory states, and to some extent in the normal state, CCL20 and CCL25 expression by IECs attracts CCR6- or CCR9-positive lymphocytes that are activated in mesenteric lymph nodes, enter the peripheral blood, and then are recruited to the intestinal mucosa, where they can undergo activation-induced apoptosis, if they are aberrantly activated, or terminal differentiation.

KEY REFERENCES

Groux H, O’Garra A, Bigler M, et al. A CD4+ T-cell subset inhibits antigen-specific T-cell responses and prevents colitis. Nature 1997; 389:737-42. (Ref 42.) Hugot JP, Chamaillard M, Zouali H, Lesage S, et al. Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn’s disease. Nature 2001; 411:599-603. (Ref 104.) Kerneis S, Bogdanova A, Kraehenbuhl JP, Pringault E. Conversion by Peyer’s patch lymphocytes of human enterocytes into M cells that transport bacteria. Science 1997; 277:949-52. (Ref 90.)

Kraus TA, Toy L, Chan L, et al. Failure to induce oral tolerance to a soluble protein in patients with inflammatory bowel disease. Gastroenterology 2004; 26:1771-8. (Ref 146.) MacDonald TT. T cell immunity to oral allergens. Curr Opin Immunol 1998; 10:620-7. (Ref 20.) Mowat AM, Viney JL. The anatomical basis of intestinal immunity. Immunol Rev 1997; 156:145-66. (Ref 2.) Neurath MF, Finotto S, Glimcher LH. The role of Th1/Th2 polarization in mucosal immunity. Nat Med 2002; 8:567-73. (Ref 169.) Neurath MF, Weigmann B, Finotto S, et al. The transcription factor T-bet regulates mucosal T cell activation in experimental colitis and Crohn’s disease. J Exp Med 2002; 195:1129-43. (Ref 170.) Neutra MR. Current concepts in mucosal immunity. V. Role of M cells in transepithelial transport of antigens and pathogens to the mucosal immune system. Am J Physiol 1998; 274:G785-91. (Ref 85.) Niess JH, Reinecker HC. Lamina propria dendritic cells in the physiology and pathology of the gastrointestinal tract. Curr Opin Gastroenterol 2005; 21:687-91. (Ref 182.) Ogura Y, Bonen DK, Inohara N, et al. A frameshift mutation in NOD2 associated with susceptibility to Crohn’s disease. Nature 2001; 411:603-6. (Ref 105.) Perera L, Shao L, Patel A, et al. Expression of nonclassical class I molecules by intestinal epithelial cells. Inflam Bowel Dis 2007; 13: 298-307. (Ref 141.) Sakaguchi S, Toda M, Asano M, et al. T cell-mediated maintenance of natural self-tolerance: Its breakdown as a possible cause of various autoimmune diseases. J Autoimmun 1996; 9:211-20. (Ref 45.) Singh UP, Venkataraman C, Singh R, et al. CXCR3 axis: Role in inflammatory bowel disease and its therapeutic implication. Endocr Metab Immune Dis Drug Targets 2007; 7:111-23. (Ref 195.) Steinman RM, Hawiger D, Nussenzweig MC. Tolerogenic dendritic cells. Annu Rev Immunol 2003; 21:685-711. (Ref 173.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

3

Cellular Growth and Neoplasia Daniel C. Chung and Daniel K. Podolsky

CHAPTER OUTLINE Mechanisms of Normal Cell Homeostasis  31 Cellular Proliferation  31 Programmed Cell Death and Senescence  32 Signaling Pathways That Regulate Cellular Growth  32 Intestinal Tumor Development: Multistep Formation and Clonal Expansion  34 Neoplasia-Associated Genes  35 Oncogenes  35 Tumor Suppressor Genes  37 DNA Repair Genes  40 Oncogenic Signaling Pathways  40

Neoplasia in the gastrointestinal (GI) tract remains one of the most common types of diseases that gastroenterologists confront. Advances in our understanding of the cellular and molecular basis of GI neoplasia have provided a foundation for the development of novel diagnostic and therapeutic approaches. Although some features are tissue site–specific, many mechanisms of tumorigenesis are common to all sites throughout the GI tract. This chapter reviews mechanisms of normal cell growth and the fundamental cellular and molecular alterations that result in malignant transformation. The common principles discussed in this chapter provide the framework for consideration of specific GI neoplasms in later chapters.

MECHANISMS OF NORMAL CELL HOMEOSTASIS CELLULAR PROLIFERATION

Neoplasia is the ultimate result of the disruption of exquisite mechanisms regulating normal cell growth. Growth is determined by the balance of cellular proliferation, differentiation, senescence, and programmed cell death. Proliferation occurs as cells traverse the cell cycle (Fig. 3-1). In preparation for cell division, there is a period of deoxyribonucleic acid (DNA) synthesis, designated the S phase. After an intervening gap period, designated the G2 phase, actual mitosis occurs in the M phase. After another intervening gap period, the G1 phase, DNA replication can begin again. The commitment to proceed through DNA replication and cell division occurs during the G1 phase at the so-called start or restriction (R) point. Cells may exit this cycle of active proliferation before reaching the R point and enter a quiescent phase, G0. Cells can subsequently re-enter the cell cycle from the G0 state (see Fig. 3-1). The duration of each

Environmental Mutagenesis  41 Chemical Carcinogenesis  41 Dietary Factors  41 Biological Features of Tumor Metastasis  42 Epithelial-Mesenchymal Transition  42 Angiogenesis and Lymphangiogenesis  42 Metastasis Genes  43 Molecular Medicine: Current and Future Approaches in Gastrointestinal Oncology  43 DNA-Based Approaches  43 Oncofetal Proteins  44

cell cycle phase as well as the overall length of the cycle vary among cell types. Regulation of cell cycle progression appears to be achieved principally by cyclins and cyclin-dependent kinase activity at the G1/S and G2/M phase transitions. Cyclin proteins are classified on the basis of their structural features and temporal expression patterns during the cell cycle (see Fig. 3-1). Cyclins A and B are expressed predominantly during the S and G2 phases. In contrast, cyclins D and E proteins are most active during the G1 phase.1 Overexpression of cyclin D1 in fibroblasts results in more rapid entry of cells into the S phase. Cyclin D1 is frequently overexpressed in a number of GI and non-GI malignancies, including those originating from the oral cavity, esophagus, breast, and bladder.2 Each cyclin forms a complex with a cyclin-dependent kinase (cdk) in a cell cycle–dependent fashion. Cyclins function as catalysts for cdk activity (see Fig. 3-1). Cdks physically associate with cyclins through their catalytic domains. The cyclin-cdk complexes regulate cell cycle progression through phosphorylation of key target proteins, including the retinoblastoma gene product (pRb) as well as the Rb family members p130 and p107.3 The final result is progression out of G1 into the S phase of the cell cycle. The cell cycle is also regulated by multiple cdk inhibitors; p21CIP1/WAF1 and p27KIP1 are inhibitors of cyclin E/cdk2. Originally discovered to be part of the complex containing cyclin D1 and cdk4/6, p21CIP1/WAF1 is transcriptionally activated by the TP53 tumor suppressor gene product (see Fig. 3-1).4 p16INK4A is another cdk inhibitor that specifically inhibits cdk4 and cdk65 and is part of a larger family of related inhibitors that includes p14, p15, and p18. p16INK4A is frequently inactivated in esophageal squamous cell cancers and pancreatic ductal adenocarcinomas, a finding that is consistent with its function as a tumor suppressor gene.6,7 p16INK4A disrupts the complex of cyclin D1

31

32

Section I  Biology of the Gastrointestinal Tract CIP1/WAF1

p21

INK4A

p16

–

KIP1

or p27

–

+ cdk4/6

cdk2

cyc D1

cyc E

pRb pRb E2F

P P P

G1

GO

E2F

S

Cyclin A

M

G2 Cyclin B Figure 3-1.  Regulation of the cell cycle by cyclins, cdks, and cdk inhibitors. In the normal cell cycle, DNA synthesis (in which chromosomal DNA is duplicated) occurs in the S phase, whereas mitosis (in which nuclei first divide to form a pair of new nuclei, followed by actual cellular division to form a pair of daughter cells) takes place in the M phase. The S and M phases are separated by two gap phases, the G1 phase after mitosis and before DNA synthesis, and the G2 phase following the S phase. During these gap phases, the cell is synthesizing proteins and metabolites, increasing its mass, and preparing for the S phase and M phase. Cell cycle progression is regulated primarily at two points, the G2/M and G1/S boundaries, through the coordinated activities of cyclins and cyclindependent kinases (cdks), which in turn are negatively regulated by cdk inhibitors (Ink4 and Cip/kip families). The mid-G1 phase is characterized by the interaction between cyclin D1 and cdk4/6. This complex hyperphosphorylates the retinoblastoma protein (pRb) and its family members (e.g., p130). Another important complex at the G1/S boundary is that of cdk2 and cyclin E (cyc E). The result is to release transcription factors such as E2F that are complexed with pRb. In turn, E2F binds to and activates the promoters of genes important in DNA synthesis.

and cdk 4/6, thereby freeing p21CIP1/WAF1 and p27KIP1 to inhibit the activity of cyclin E/cdk2.8

PROGRAMMED CELL DEATH AND SENESCENCE

Apoptosis (or programmed cell death) is an important mechanism that counterbalances cell proliferation, and escape from normal apoptotic mechanisms plays a critical role in oncogenesis. Apoptosis is characterized by distinctive features that include chromatin compaction, condensation of cytoplasm, and mild convolution of the nucleus and cytoplasm. These changes are followed by nuclear fragmentation and marked convolution of the cell surface. Eventually, membrane-bound apoptotic bodies that represent the cellular residue are produced and phagocytosed. Apoptosis is distinguished biochemically by cleavage of doublestranded DNA, which results in fragmented DNA.

Studies of the roundworm Caenorhabditis elegans have led to the initial identification of the gene ced-3, a protease that is the major effector of apoptosis. Two key regulators of ced-3, designated ced-9 and ced-4, were found to prevent or induce apoptosis, respectively.9 The mammalian oncogene bcl-2 shares homology with ced-9 and protects lymphocytes and neurons from apoptosis10; bcl-2 complexes with bax, a protein that by itself contributes to apoptosis.11 Of note, both bcl-2 and bax are part of larger gene families, and the stoichiometric relationships among different combinations of the encoded proteins can determine the balance between cell survival and cell death.12 Two well-defined pathways that trigger apoptosis have been described in detail. One pathway is mediated through membrane-bound death receptors, which include tumor necrosis factor (TNF) receptors, Fas, and DR5, whereas the other pathway involves activation of TP53 expression by environmental insults such as ionizing radiation, hypoxia, or growth factor withdrawal, with a subsequent increase in the bax-to-bcl-2 ratio. Both pathways converge to disrupt mitochondrial integrity and release of cytochrome c (Fig. 3-2). The so-called apoptosome complex (cytochrome c, caspase 9, and Apaf1) then activates downstream caspases, such as caspase 3, eventuating in cell death. Activation of caspases, intracellular cysteine proteases that cleave their substrates at aspartate residues, is a key step in programmed cell death in mammalian cells. Replicative senescence also plays a role in determining overall growth in cell populations. Most primary cells when grown in vitro have a limited replicative potential and eventually undergo senescence.13 In contrast, malignant cells can replicate indefinitely. Up-regulation of the telomerase enzyme is essential to escape from replicative senescence. Telomeres are repetitive DNA sequences at the ends of all chromosomes that regulate chromosomal stability. Telomeres shorten with each cell division and, when they have been reduced to a certain critical length, senescence occurs. Cancer cells are able to maintain their telomere length despite multiple cell divisions through the reactivation of telomerase enzyme activity.14

SIGNALING PATHWAYS THAT REGULATE CELLULAR GROWTH

Cellular proliferation is achieved through transition of cells from G0 arrest into the active cell cycle (see Fig. 3-1). Although progression through the cell cycle is controlled by the regulatory proteins just described, overall proliferation is modulated by external stimuli. Growth factors that bind to specific transmembrane receptors on the cell surface may be especially important. The cytoplasmic tails of these transmembrane receptor proteins produce an intracellular signal after ligand binding. In addition to peptide growth factors, extracellular matrix and cell-cell adhesion molecules have a significant impact on cell proliferation. Although the full spectrum of molecules that play a role in cell-matrix and cell-cell adhesion is still not defined, it is known to include integrins, cadherins, selectins, and proteoglycans. Interactions with these adhesion molecules lead to changes in the cell cytoskeleton, indirectly modulating external growth stimuli. Alterations in cell-matrix or cell-cell interactions are particularly important in contributing to the invasive phenotype characteristic of malignant cells. Interaction of ligands with their receptors at the cell surface induces intracellular signals that ultimately result in alterations in gene transcription. Three important receptor subtypes appear to initiate cellular signaling through

Chapter 3  Cellular Growth and Neoplasia Death receptors (TNF-R1, Fas, DR5)

Apoptotic signal (e.g., ionizing radiation) DD FLIP TP53

Caspase 8 bax, bak

bcl-2 bcl-xl

BID

Mitochondria

Apoptosome

Apaf1, caspase 9, cytochrome c

Caspase 3

Cell death Figure 3-2.  Apoptosis (programmed cell death) counterbalances cellular proliferation to regulate overall tissue growth. A complex interplay of proapoptotic and antiapoptotic molecules results in the downstream activation of caspases that mediate cell death. Some of these signals are initiated through environmental insults that activate the TP53 tumor suppressor gene, and some are initiated through death receptors, including TNF-R1, Fas, and DR5. Death receptors activate caspase 8, which in turn activates BID. In addition, there is an interplay between proapoptotic (bax, bak) and antiapoptotic (bcl-2, bcl-xl) molecules. Both pathways converge on the mitochondria, resulting in the release of cytochrome c and formation of the apoptosome (Apaf1, caspase 9, and cytochrome c). This leads to activation of multiple caspases, in particular caspase 3, and ultimately to cell death. BID, bcl-2 interacting domain; DD, death domain; FLICE, FADD-like IL-1β-converting enzyme; FLIP, FLICE (also known as caspase 8) inhibitory protein; TNF-R1, tumor necrosis factor receptor 1.

ligand-receptor interaction at the cell surface: (1) tyrosine kinases; (2) serine and threonine kinases; and (3) G protein– coupled receptors. The receptors for many peptide growth factors contain intrinsic tyrosine kinase activity within their intracellular tail. After ligand binding, tyrosine kinase activity is stimulated, leading to phosphorylation of tyrosine residues in target proteins within the cell. The full spectrum of proteins phosphorylated by each tyrosine kinase remains to be determined. Most receptors also autophosphorylate tyrosine residues present in the receptors themselves to initiate signaling and, in some cases, this also causes attenuation of their own activity to effect an intramolecular feedback regulatory mechanism. The receptors for many peptide growth factors, including epidermal growth factor (EGF), belong to this receptor class. Other receptors on the cell surface possess kinase activity directed toward serine or threonine residues rather than tyrosine. These receptors also phosphorylate a variety of cellular proteins, leading to a cascade of biological responses.

Multiple sites of serine and threonine phosphorylation are present on many growth factor receptors, including the tyrosine kinase receptors, suggesting the existence of significant interactions among various receptors present on a single cell.15 The transforming growth factor-β (TGF-β) receptor complex is one important example of a serine-threonine kinase–containing transmembrane receptor. Many receptors are members of the so-called sevenmembrane–spanning receptor family. These receptors are coupled to guanine nucleotide binding proteins, and are designated G proteins. G proteins undergo a conformational change that is dependent on the presence of guanosine phosphates.16 Activation of G proteins can trigger a variety of intracellular signals, including stimulation of phospho­ lipase C and the generation of phosphoinositides (most importantly, inositol 1,4,5-triphosphate) and diacylglycerol through hydrolysis of membrane phospholipids, as well as modulation of the second messengers cyclic AMP and GMP.17 Somatostatin receptors exemplify a G protein– coupled receptor prevalent in the GI tract. Binding of growth factors and cytokines to cell surface receptors typically produces alterations in a variety of cellular functions that influence growth. These functions include ion transport, nutrient uptake, and protein synthesis. However, the ligand-receptor interaction must ultimately modify gene expression within the nucleus to affect cell proliferation. The regulation of the content and activity of transcriptional factors within the nucleus is the final step in pathways that translate an external stimulus to a change in cell proliferation. These transcriptional factors modulate the expression of genes that control cell proliferation and phenotype. The Wnt pathway is one important example of a signaling pathway that regulates the cell cycle machinery to control the proliferation of intestinal epithelial cells (Fig. 3-3). Although the details of the specific interactions between the Wnt ligand and its receptor Frz, a member of the seven-membrane receptor family, in the GI tract are not fully clarified, an active Wnt signal ultimately results in the accumulation of β-catenin in the nucleus, where it binds with the transcription factor TCF-4 to activate a set of target genes.18 Inhibition of the Wnt signal in mice can be achieved by deletion of TCF-4 or overexpression of a Wnt inhibitor designated Dickkopf1, which results in dramatic hypoproliferation of the intestinal epithelium.19,20 This hypopro­ liferation appears to be mediated by decreased expression of the TCF-4 target gene c-MYC, which directly represses p21CIP1/WAF1.21 Thus, a Wnt signal stimulates proliferation of intestinal epithelial cells by repressing the cell cycle inhibitor p21CIP1/WAF1. Cyclin D1 has an extremely short half-life (<20 minutes) and is a rate-limiting factor for progression through the G1 phase of the cell cycle (see Fig. 3-1). Consequently, it is one of the most tightly regulated of all cell cycle proteins. Extracellular signals from growth factors, including EGF, colonystimulating factor 1 (CSF-1), platelet-derived growth factor (PDGF), and insulin-like growth factor (IGF), can regulate cellular proliferation by rapidly inducing the expression of the cyclin D1 gene.22 Tissue homeostasis is also maintained by growthinhibiting signals that counterbalance proliferative signals. TGF-β is a potent growth-inhibiting factor that mediates arrest of the cell cycle at the G1 phase. TGF-β not only induces the transcription of the cell cycle inhibitors p15INK4B and p21CIP1/WAF1, but also enhances the inhibitory activity of p27KIP1 on the cyclin E/cdk2 complex (see Fig. 3-1).23 These effects are mediated intracellularly through the Smad family of proteins.

33

34

Section I  Biology of the Gastrointestinal Tract Wnt

Frz β-catenin Figure 3-3.  The Wnt signaling pathway is an important regulator of intestinal epithelial cell proliferation and tumorigenesis. In the absence of a Wnt signal (left panel), β-catenin forms a cytoplasmic complex with APC, Axin, and glycogen synthase kinase-3β (GSK-3β). This β-catenin destruction complex phosphorylates (-P) β-catenin and targets it for degradation via the ubiquitin-mediated proteasomal pathway. In the presence of an active Wnt signal (right panel), β-catenin is stabilized, and excess cytoplasmic β-catenin is translocated to the nucleus, where it interacts with the TCF-4 transcription factor to regulate the expression of many key target genes. VEGF, vascular endothelial growth factor.

Frz DSH

-P

GSK-3β

GSK-3β β-catenin

Axin

Axin APC

β-catenin

β-catenin

APC

β-catenin

P β-catenin

+ β-catenin TCF-4

c-Myc cyclin D1 VEGF

Nucleus Degradation

INTESTINAL TUMOR DEVELOPMENT: MULTISTEP FORMATION AND CLONAL EXPANSION Multiple sequential genetic alterations are required for the transformation of normal intestinal epithelium to frank malignancy. This multistep nature of tumorigenesis is most directly illustrated by the changes that accrue in the development of colonic neoplasia (see Chapters 122 and 123). The accumulation of genetic alterations roughly parallels the progression from normal epithelium through adenomatous polyps (or, in the case of ulcerative colitis, flat dysplastic mucosa) to malignant neoplasia. Studies on the molecular pathogenesis of colon cancer have served as a paradigm for the elucidation of genetic alterations in other GI cancers. For example, a similar progression is also seen in the transition from normal squamous epithelium to metaplastic mucosa (Barrett’s esophagus) through dysplasia to adenocarcinoma of the esophagus. Gastric and pancreatic oncogenesis are each thought to proceed through similar multistep pathways. Models of the multistep or multiple hit process of tumor formation have largely superseded earlier concepts of oncogenesis that discriminated between tumor initiation and subsequent promotion. Initiation was attributed to a single change in a cell that converted it from a normal to a malignant cell. Promotion reflected all the factors that acted after the initiating event to enhance tumor growth. However, oncogenesis occurs through a series of events that result in incremental changes in cell behavior until the cell eventually passes some threshold associated with the malignant phenotype. Nevertheless, there is still some merit in a more limited concept of tumor promotion. A number of factors promote the likelihood of malignant transformation through the stimulation of increased cellular turnover, which increases opportunities for somatic mutations to occur.24 In the GI tract, these promoting factors include dietary constituents (see later) as well as chronic inflammation, which are associated with increased cell proliferation. Thus, a

number of chronic inflammatory conditions increase the site-specific risk of cancer, such as ulcerative colitis (see Chapter 112), chronic gastritis (see Chapters 51 and 54), chronic pancreatitis (see Chapters 59 and 60), Barrett’s esophagus (see Chapters 44 and 46), and chronic hepatitis (see Chapter 94). Although the mechanisms whereby inflammatory processes elicit eventual tumor development are incompletely understood, cytokines produced by inflammatory cells can stimulate tumor cells, leading to activation of nuclear factor-κB (NF-κB) in the tumor cells that can serve to inhibit apoptosis and stimulate proliferation.25 Clonal expansion is also essential to tumor development.26 Whereas germline mutations may lead to altered expression of a gene in all cells in a tissue, subsequent additional somatic mutations generally occur only in a small, largely random subpopulation of cells. Clonal expansion occurs if a specific gene mutation results in a survival advantage for the cell. A second round of clonal expansion occurs when a cell within this population sustains still another genetic alteration, which further enhances its growth properties. After several iterations, a final genetic alteration eventually confers a property that, together with the preceding genetic alterations, makes a cell malignant. Recent evidence has led to the suggestion that cancer stem cells in tumors may play a central role in tumorigenesis. These cells are defined by the capacity for self-renewal and the ability to generate progeny that lack this capacity but manifest the characteristics of the heterogeneous lineages that comprise the tumor.27 Failure to eradicate the cancer stem cell population is posited to underlie tumor recurrences after chemotherapy. Precise identification of the cancer stem cell compartment has been possible for hematologic malignancies. However, comparable definitive proof of their presence in solid tumors, such as intestinal cancers, remains a challenge.28 A genetically unstable environment was thought to be necessary for the development of the multiple alterations that ultimately result in cancer. Genomic instability is

Chapter 3  Cellular Growth and Neoplasia Chromosomal instability pathway

T=7–10 years DNA hypomethylation K-ras

APC Normal

Early adenoma

?

TP53 Advanced adenoma

Invasive cancer

MSH2 MLH1 MSI pathway

Metastases

TGF-βRII IGF-RII BAX E2F-4 T=1–3 years

Figure 3-4.  Multistep model of colorectal cancer based on underlying genetic instability. There are two major pathways, chromosomal instability (top half of figure) and microsatellite instability (bottom half). The progression from normal colonic epithelium to carcinoma is associated with the acquisition of several genetic alterations. In the chromosomal instability pathway, these alterations include the concomitant activation of oncogenes (e.g., K-ras) through a point mutation and inactivation of tumor suppressor genes (e.g., APC, TP53) through a point mutation or deletion. In addition, DNA hypomethylation may be important in this process. An increasing aggregate number of mutations can be correlated with progression from early benign polyp to cancer, as reflected by analysis of polyps by size. In the microsatellite instability model, mutations in DNA mismatch repair genes such as MSH2 and MLH1 create a mutator phenotype in which mutations accumulate in specific target genes (see section on DNA mismatch repair). Tumors develop much more rapidly through this pathway than through the chromosomal instability pathway (horizontal arrows).

observed in almost all cancers, regardless of organ site. Instability of the genome may result from several mechanisms. In colon cancer, there are at least two wellrecognized forms of genetic instability, and they have been termed chromosomal instability and microsatellite instability.29 Chromosomal instability results in tumor cells that display frequent aneuploidy, large chromosomal deletions, and chromosomal duplications. In contrast, tumors that display microsatellite instability are often diploid or neardiploid on a chromosomal level but harbor frequent alterations in smaller tracts of microsatellite DNA (see later discussion on DNA repair). Thus, there are at least two distinct routes to the formation of a colorectal cancer, depending on the nature of the underlying genetic instability (Fig. 3-4).

NEOPLASIA-ASSOCIATED GENES The genes that collectively play an important role in oncogenesis generally lead to disruption of the orderly mechanisms of normal cell proliferation. Insofar as normal cell proliferation appears to depend on a wide variety of genes, it is not surprising that alterations in diverse genes confer part or all the phenotypic features of transformation. Despite this diversity, all these genes that become altered appear to belong to one of three distinct groups: (1) oncogenes, which actively confer a growth-promoting property; (2) tumor suppressor genes, the products of which normally restrain growth or proliferation; and (3) DNA repair genes, which contribute to transformation by fostering genomic instabil­ bity and facilitating mutations in other genes. Activation of oncogenes or inactivation of tumor suppressor genes and

DNA repair genes contributes to malignant transformation (Table 3-1).

ONCOGENES

Typically, oncogenes are genes that encode a normal cellular protein expressed at inappropriately high levels or mutated genes that produce a structurally altered protein that exhibits inappropriately high activity. For example, several genes that encode tyrosine kinase–containing growth factor receptors become oncogenes after a mutation results in unregulated tyrosine kinase activity that is no longer dependent on the presence of the appropriate ligand. The normal cellular genes from which the oncogenes derive are designated proto-oncogenes or cellular oncogenes. More than 80 oncogenes have been isolated, and additional oncogenes continue to be identified. Most of these genes are widely expressed in many different types of tumor cells. Multiple oncogenes (usually in combination with altered tumor suppressor genes) are commonly found within a single tumor. Several mechanisms can lead to oncogene activation. These include gene transduction or insertion, point mutation, gene rearrangement, and gene amplification. Gene transduction and insertion generally result from retroviral infection. Point mutations result in constitutively active oncogene products. Gene rearrangements can result in oncogenic fusion proteins, and gene amplifications lead to uncontrolled overexpression of a normal gene product. The proteins encoded by oncogenes comprise at least four distinct groups—peptide growth factors that may be secreted into the extracellular milieu, protein kinases, signal-transducing proteins associated with the inner cell membrane surface (membrane-associated G proteins), and transcriptional regulatory proteins located in the nucleus.

35

36

Section I  Biology of the Gastrointestinal Tract Table 3-1  Oncogenes, Tumor Suppressor Genes, and DNA Repair Genes Altered in Gastrointestinal Tumors* Gene Oncogenes K-ras c-Myc EGFR ErbB2 B-Raf c-Src c-Kit β-catenin Tumor Suppressor Genes TP53 p16INK4A p14ARF APC DPC4, SMAD4 E-cadherin Rb BRCA2 Axin LKB1 DNA Repair Genes MSH2, MLH1, MSH6, PMS2 MYH

Esophagus

+ +

+ +

Stomach

Biliary tract

Pancreas

Colon

Liver

+ +

+

+ +

+ + +

+

+

+

+ +

+

+

+

+

+ +

+ +

+

+ +

+

+ +

+ + + +

+ +

+

+ +

+

+

+

+

GIST

+ +

*This list is not comprehensive but classifies the most important known oncogene, tumor suppressor gene, and DNA repair gene alterations in GI tumors. Oncogenes are activated in tumors; tumor suppressor genes and DNA repair genes are inactivated. GIST, gastrointestinal stromal tumor.

Oncogenes and Peptide Growth Factors

The transforming effects of enhanced expression of a variety of growth factors have been demonstrated both in vitro and in vivo. Several growth factor–related proteins encoded by oncogenes have now been recognized, including the family of Wnt proteins and Sis, which encodes the β chain of platelet-derived growth factor. It is axiomatic that cells that produce high levels of a growth factor must also express specific receptors activated by that growth factor to yield an autocrine growth-stimulating loop that is often present in neoplasms. For example, several colon-derived cancer cell lines express TGF-β and IGF I and II, as well as their receptors. This autocrine mechanism may be contrasted with overproduction of a growth factor that exerts its influence at a remote cellular target rather than within the tumor itself, exemplified by gastrin produced by gastrinoma cells, which exerts trophic effects on the gastric mucosa but not on the tumor itself (see Chapter 32).

Protein Kinase–Related Oncogenes

The largest family of oncogenes encodes proteins with kinase activity. These oncogenes encompass the full variety of protein kinases, including receptor and nonreceptor tyrosine kinases and cytoplasmic serine and threonine kinases. Many members of this large oncogene group are expressed by neoplasms of the GI tract. A brief consideration of receptor protein tyrosine kinase HER2/Neu/ERBB2, which is related to the EGF receptor, is particularly illustrative. There are four EGF receptor family members (ERBB1-4). The viral v-erb-b2 encodes a truncated form of the EGF receptor that lacks most of the external EGF-binding domain.30 As a result, the receptor no longer requires the presence of the ligand for activation and remains continuously activated, stimulating proliferation. The neu oncogene is derived from a rat cellular protooncogene closely related to the EGF receptor. The oncogene

differs from its normal counterpart by a point mutation that changes a single residue (valine to glutamic acid) within the transmembrane domain, thereby causing activation of the 185-kd tyrosine kinase protein (p185neu).31 The human counterpart (ERBB2) of the neu oncogene is not mutated but is overexpressed or amplified in a variety of adenocarcinomas, including those arising in the stomach, breast, and prostate.32 In addition, ERBB2 expression increases progressively in the transition from normal esophageal mucosa through the dysplastic state characteristic of Barrett’s esophagus to esophageal adenocarcinoma.33 In contrast with the receptor type of tyrosine kinase that possesses intrinsic catalytic activity, many other receptors and membrane proteins lack self-contained signaling activity. Instead, they are coupled to nonreceptor tyrosine kinases on the cytoplasmic side of the plasma membrane that act as signal transducers. A number of oncogenes associated with neoplasms of the GI tract, most notably the colon, are members of the src family of nonreceptor tyrosine kinases. Members of the src family are approximately 60-kd phosphoproteins (v-src) that possess inherent tyrosine kinase activity and associate with the inner surface of the plasma membrane. Autophosphorylation of the normal c-src leads to attenuation of its kinase activity, thereby providing inherent regulation to limit unrestrained activity.34 Increased levels of c-src activity have been found in colonic cancer tissue and colon cancer–derived cell lines.35 Activating mutations of c-src have been identified in a subset of advanced, metastatic colon cancers.36

Signal Transduction–Related Oncogenes (Membrane-Associated G Proteins)

Intermediate steps that effectively translate ligand-receptor binding to an intracellular signal are essential in mediating functional responses of the cell. Mutations in genes that

Chapter 3  Cellular Growth and Neoplasia encode key proteins that participate in signal transduction can also lead to cellular transformation. G proteins regulate signaling of the large family of G protein–coupled receptors (GPCRs) through the exchange of guanosine triphosphate (GTP) with guanosine diphosphate (GDP). Altered ras genes, a family of proteins related to the G proteins, are among the most commonly detected oncogenes in GI tract cancers. The ras family contains three genes: H-ras, K-ras, and N-ras. All three encode 21-kd proteins. Post-translational modification of the carboxy-terminal end of the protein results in plasma membrane localization. Point mutations that result in amino acid substitutions at critical hot spot positions (residues 12, 13, 59, and 61) convert the normal gene into an oncogene. Almost all ras mutations in GI malignancies that have been identified occur in the K-ras oncogene, and the frequency of mutations varies greatly among different GI tumor types. The highest frequency is found in tumors of the exocrine pancreas; more than 90% of these tumors possess mutations in the K-ras gene.37 Ras genes activated through point mutation have been identified in approximately 50% of colonic cancers as well as large benign colonic polyps.38 In contrast, fewer than 10% of colonic adenomas smaller than 1 cm have K-ras mutations (see Fig. 3-4). Most oncogenic mutations in ras cause biochemical changes that maintain it in the active, GTP-bound state by reducing guanosine triphosphatase (GTPase) activity or by destabilizing the inactive GDP-bound form. However, several ras mutants retain significant GTPase activity; therefore, other mechanisms that convert ras to a transforming protein may be involved.39 The GTPase-activating protein (GAP) induces a 500-fold increase in the GTPase activity of the normal ras protein, and some mutant ras proteins are resistant to this modifying protein.40 In the presence of GAP, ras oncogenic activity correlates strongly with its reduced GTPase activity. A functional consequence of ras activation is the phosphorylation of key serine and threonine kinases. One important downstream signaling target of ras is B-raf. In colon cancers without an identifiable K-ras mutation, 20% possess an activating B-raf mutation,41 consistent with the concept that activation of an oncogenic pathway can be achieved through an alteration in any of several sequential components of a particular pathway.

Nuclear Oncogenes

Many cellular oncogenes encode proteins that localize to the nucleus. In essence, these nuclear oncogene products are the final mediators of signal transduction pathways that are also affected by cytoplasmic and plasma membranebound oncoproteins, because they regulate the expression of certain genes that enhance cellular proliferation and suppress normal differentiation. In general, most nuclear oncoproteins immortalize primary cells and cooperate with other oncoproteins, especially ras, to cause transformation. Many nuclear oncoproteins belong to the class of transcription factors that alter the expression of specific target genes. Although a detailed understanding of the mechanisms whereby the nuclear oncoproteins regulate transcription is still lacking, homo- and heterodimerization of these proteins through well-defined motifs is important in the process. Other domains confer DNA-binding specificity which is critical in the regulation of genes involved in cell cycle control. The role of nuclear oncogenes that encode transcriptional regulatory proteins and that are involved in protein-protein interactions is illustrated by the myc family. The c-Myc

protein product is involved in critical cellular functions, such as proliferation, differentiation, apoptosis, transformation, and transcriptional activation of key genes.42 Frequently, c-Myc is overexpressed in many GI cancers. The protein contains several important domains. The carboxy terminal contains a helix-loop-helix motif that mediates binding to other proteins, such as Max.43 These heterodimers bind DNA through the basic domain of c-Myc. The amino terminal of c-Myc contains regions critical for transcriptional activation of genes, transformation, and apoptosis.44 c-Myc has been found to be a transcriptional target of the β-catenin/TCF-4 complex in colorectal cancers (see Fig. 3-3), which may explain the overexpression of c-Myc observed in this cancer type.45

TUMOR SUPPRESSOR GENES

The products of tumor suppressor genes prevent the acquisition of the transformed phenotype in vitro and have similar functional properties in vivo. Mutations that disrupt the biological function of these genes are associated with all GI cancers. Germline mutations of this class of gene underlie most of the known inherited cancer syndromes in which a specific gene has been implicated. A number of these genes and their products have been identified and characterized (Table 3-2). Initial recognition of the existence of tumor suppressor genes was derived from analysis of families with a markedly increased incidence of specific tumors. Almost all types of tumors have been found to occur in an inherited form. In the GI tract, hereditary colon cancer, gastric cancer, and pancreatic cancer syndromes are the best described and are discussed elsewhere in this text. Within these relatively rare kindreds, as many as half of first-degree relatives of a proband (and each subsequent generation) develop specific tumors, consistent with a single-gene disorder with an autosomal dominant mode of transmission. Despite the variation in the type of tumor found in different inherited cancer syndromes, a number of features are common to all inherited GI cancer syndromes. Most importantly, the marked increase in risk for a particular tumor is found in the absence of other predisposing environmental factors. In addition, multiple primary tumors often develop within the target tissue, and tumors in these affected members typically arise at a younger age than they do in the general population. Finally, affected individuals are sometimes at risk for some types of tumors outside the GI tract.

Table 3-2  Chromosomal Localization and Function of Several Key Tumor Suppressor Genes in Gastrointestinal (GI) Cancers Chromosome

Gene*

Function

5q 9p 11q

APC p16INK4A MEN1

16q

E-cadherin

17p

TP53

18q

DPC-4, SMAD4

Inhibition of Wnt signaling Cell cycle inhibition Regulation of histone methyltransferase Maintenance of cell-cell interactions Regulation of DNA repair and apoptosis Transduction of transforming growth factor-β signal

*Clinical GI disorders associated with defects in some of these genes are listed in Table 3-4.

37

38

Section I  Biology of the Gastrointestinal Tract Acquired somatic mutation

Germline mutation X

Inherited cancer syndrome

Tumor Acquired somatic mutation

Second somatic mutation X

X

Sporadic cancer

Tumor

Time , Tumor suppressor gene. Figure 3-5.  Knudson’s two-hit hypothesis of tumor suppressor genes. In an inherited cancer syndrome, one chromosome has an inactive tumor suppressor gene locus because of a germline mutation. The counterpart tumor suppressor gene on the remaining paired chromosome is subsequently inactivated by a somatic mutation, leading to formation of carcinoma. In contrast, in a sporadic cancer, the two alleles of the tumor suppressor gene become inactivated through two independent somatic mutations, an unlikely event within a single cell. (Ref. 46).

These observations led Knudson to hypothesize that tumors in familial cancer syndromes might derive from independent mutations in the two alleles of a specific tumor suppressor gene (Fig. 3-5)—that the first mutation was present in one copy of the gene inherited in the germline and therefore present in all cells in affected family members.46 A somatic mutation of the remaining normal allele of the tumor suppressor gene that might occur in any cell would then lead to tumor development, explaining the high incidence of cancer and multiple tumors. The same gene might play a role in the development of the same tumor type in the general population (sporadic cancer), but two independent somatic mutations of each of the two alleles would be required. However, this combination of events should be uncommon and would explain the lower frequency and later age of diagnosis of similar tumors in the general population. Comings was the first to suggest that the relevant gene in a familial cancer syndrome might encode a tumor-suppressing gene product.47 Although this two-hit hypothesis has been generally accepted, there are exceptions. For example, there are data suggesting that a single alteration in just one allele of the Lkb1 tumor suppressor gene that underlies the Peutz-Jeghers syndrome may be sufficient for intestinal polyp formation.48

Loss of Heterozygosity, Allelic Deletion, and Tumor Suppressor Gene Inactivation

Some tumor suppressor genes were first cloned through detection of regions of gene deletion in tumor samples from cancer-prone kindreds by screening of DNA for markers scattered throughout the genome. These deletions targeted the second wild-type allele and served to pinpoint the chromosomal location of the disease-causing gene present on the other allele. Polymorphisms that distinguish between the two different paternal and maternal alleles on a mole­c­ ular level made these discoveries possible. Three types of polymorphisms are recognized—single nucleotide polymorphisms (SNPs), restriction fragment length polymorphisms (RFLPs), and microsatellite polymor-

phisms. SNPs represent single base pair alterations that are typically silent; these are the most abundant type of polymorphism, occurring in approximately every 1000 base pairs throughout the genome.49 RFLPs are a unique type of SNP in which the single nucleotide change alters a recognition site for a restriction endonuclease. Thus, digestion of DNA with a restriction endonuclease allows the two different alleles of the same genetic segment inherited from the subject’s two parents to be distinguished (Fig. 3-6). A more widely applicable approach uses polymorphisms present within DNA microsatellite markers. Microsatellite DNA sequences are short repetitive mononucleotide or dinucleotide repeats, such as a poly-A or poly-CA sequence. Microsatellite polymorphisms are much more common than RFLPs in the genome. There is a wide variation in the number of repeats in different alleles, and these differences can be detected through a polymerase chain reaction (PCR; see Fig. 3-6). Although SNPs or microsatellites can be used to detect deletions, SNPs will likely supersede other techniques. Either approach provides a means of assessing whether a specific region of a chromosome is deleted in tumor tissue when compared with normal tissue from the same individual. These losses, termed loss of heterozygosity or an allelic deletion (loss of an allele from one parent), represent an important mechanism of inactivation of one copy of tumor suppressor genes (see Fig. 3-6). When coupled with a preexisting germline mutation, such allelic deletions provide the second hit, which results in a loss of function of both tumor suppressor gene copies. Other mechanisms of tumor suppressor gene inactivation include point mutation or small intragenic deletions that result in premature truncation of the protein product, or promoter hypermethylation. Transcriptional silencing can result from methylation of CpG islands in gene promoters; this has been demonstrated to occur in the gene encoding p16INK4A in esophageal and pancreatic cancers and the gene encoding E-cadherin in gastric cancer.50 Tumor suppressor genes do not function identically in every tissue type. Consequently, inactivation of a particular tumor suppressor gene is tumorigenic only in certain tissues. For example, the tumor suppressor genes RB, BRCA1, and VHL play crucial roles in retinoblastomas, breast cancer, and renal cell cancer, respectively, but are rarely mutated in GI malignancies. Three tumor suppressor genes shown to have a critical role in the pathogenesis of GI malignancies, APC, TP53, and SMAD4, are described below.

Adenomatous Polyposis Coli Gene

Genetic linkage analysis has revealed markers on chromosome 5q21 that are tightly linked to polyp development in affected members of kindreds with the familial adenomatous polyposis (FAP) and Gardner’s syndrome.51 Further work led to the identification of the gene responsible for FAP, the adenomatous polyposis coli (APC) gene.52 As predicted, germline mutations of APC were found in affected patients, and the germline mutations segregate with the disease within a given family.53,54 The full spectrum of adenomatous polyposis syndromes attributable to APC is discussed in detail in Chapter 122. Although these syndromes are relatively rare, studies identifying genetic factors that contribute to these syndromes have provided insight into mechanisms essential to the development of common sporadic colon cancers as well as to tumorigenesis in general.55 Somatic mutations in APC have been found in most sporadic colon polyps and cancers.56,57 Mutations in APC are characteristically identified in the earliest adenomas, indicating that APC plays a critical role as the gatekeeper in the

Chapter 3  Cellular Growth and Neoplasia 5′

C T G C C TA G C C

Allele 1

5′

CTGCCGAGCC

Allele 2

A 1

2 2

1 gene locus

restriction enzyme

gel electrophoresis

digest

autoradiography

2 1

restriction endonuclease sites

B

Normal DNA (N)

Tumor DNA (T)

1

1

2

2 deletion

CA dinucleotide

PCR amplification

Gel electrophoresis

N

T

1

N normal DNA T tumor DNA

2

allelic deletion or LOH

C Figure 3-6.  Genomic polymorphisms facilitate the identification of allelic deletions. A, Single nucleotide polymorphisms (SNPs). SNPs are single base pair changes at two identical positions of a specific chromosome. Most are silent and do not result in alterations in gene products. B, Restriction fragment length polymorphism (RFLP) analysis. Normal homologous chromosomes have sequences that can be recognized by bacterial enzymes, designated restriction endonucleases (REs). These REs cleave DNA at the site of these specific sequences to yield a pattern of restriction fragments. If segments of DNA inherited from the two parents differ by the presence or absence of an RE site, the RE fragments will migrate as bands of different size on gel electrophoresis. After transfer of the restriction fragments to a membrane and hybridization with a specific gene probe, different bands can be detected by autoradiography. In this manner, it can be determined whether both the paternal- and maternal-derived alleles of a gene or a DNA segment are present. C, Loss of heterozygosity (LOH) analysis using microsatellites. Polymorphic microsatellite markers vary in the number of repeats—that is, CAn—between the two alleles. These markers can be amplified using PCR primers that recognize conserved sequences flanking the microsatellite DNA. When tumor DNA is compared with a normal DNA sample, tumors may exhibit a deletion in one of the alleles. This is LOH. CA, cytosine adenine dinucleotide; PCR, polymerase chain reaction.

39

40

Section I  Biology of the Gastrointestinal Tract multistep progression from normal epithelial cell to colon cancer (see Fig. 3-4). The APC gene comprises 15 exons and encodes a predicted protein of 2843 amino acids, or approximately 310  kd. Most germline and somatic APC gene mutations result in a premature stop codon and therefore a truncated protein product. Although mutations are most common in exon 15 of the APC gene, they may occur throughout the gene. Those occurring in the APC amino terminal are associated with a rare variant of FAP, attenuated familial adenomatous polyposis (AFAP).58 Studies have revealed a segregation of certain APC mutations with the phenotype of congenital hypertrophy of the retinal pigment epithelium (CHRPE).59 APC mutations result in functional changes in key protein-protein interactions. As discussed noted, APC is a negative regulator of the Wnt signaling pathway (see Fig. 3-3). Mutant APC proteins are unable to interact with β-catenin, resulting in uncontrolled activation of the Wnt signaling pathway and the subsequent oncogenic phenotype.

TP53 Gene

p53, named for a 53-kD sized gene product, is a nuclear phosphoprotein that plays a key role in cell cycle regulation and apoptosis.60 The p53 protein was first detected in tumors as the product of a mutated gene that was mapped to chromosome 17p, a region found to exhibit loss of heterozygosity in many tumors. Point mutations in TP53 have been identified in as many as 50% to 70% of sporadic colon cancers but only a small subset of colonic adenomas (see Fig. 3-4).61 Point mutations in TP53 have also been found in esophageal squamous carcinoma and adenocarcinoma, gastric carcinoma, pancreatic adenocarcinoma, and hepatocellular carcinoma.60 Interestingly, aflatoxin appears to induce a mutation in a single hot spot codon (codon 249) of TP53 in many hepatocellular carcinomas.62 In addition to the TP53 point mutations in sporadic cancers, germline TP53 mutations have been observed in the Li-Fraumeni syndrome, an autosomal dominant familial disorder in which breast carcinoma, soft tissue sarcoma, osteosarcoma, leukemia, brain tumor, and adrenocortical carcinoma can develop in affected persons.63 p53 is a sequence-specific transcription factor that is induced in conditions of cellular stress, such as ionizing radiation, growth factor withdrawal, or cytotoxic therapy (see Fig. 3-2). As a consequence of genotoxic damage, p53 arrests cells at the G1 phase to facilitate DNA repair or trigger apoptosis. p53 mediates some of these responses through the induction of the p21CIP1/WAF1 inhibitor of the cell cycle or pro-apoptotic genes, including PUMA, and c-Myc appears to play a role in this cell fate decision.64 The functional importance of p53 in colon cancer has been underscored by experiments in which wild-type TP53 was reintroduced into colon cancer cells that had only mutant TP53.65 Repleting cells with p53, the product of TP53, can arrest growth in a cell cycle phase-specific manner.

SMAD4 Gene

SMAD4, also designated deleted in pancreas cancer-4 (DPC-4), is a tumor suppressor gene located on chromosome 18q and is deleted or mutated in most pancreatic adenocarcinomas and a subset of colon cancers. This gene encodes Smad4, an essential intracellular mediator of the growth inhibitory effects of TGF-β. The Smad4 protein has two important domains, the mad homology domains 1 and 2 (MH1 and MH2), which are essential for DNA binding and for oligomerization with other Smad proteins, respectively.66

Mutant Smad4 blocks TGF-β–induced inhibition of pro­ liferation. Germline mutations in SMAD4 result in the juvenile polyposis syndrome. Other genes on chromosome 18q may also be important in colon carcinogenesis.67

DNA REPAIR GENES

Cellular mechanisms have evolved to preserve the fidelity of DNA. Errors can be introduced into the genome through the spontaneous mismatching of nucleotides during normal DNA replication. This occurs most commonly from slippage in microsatellite DNA, which involves regions of mono­ nucleotide (e.g., poly-A) or dinucleotide (e.g., poly-CA) repeats.68 The DNA mismatch repair system corrects these errors. The components of this system have been studied most extensively in prokaryotes and lower eukaryotes, most notably yeast. The enzymes bind mismatched DNA, cut the DNA strand with the mismatched nucleotide, unwind the DNA fragment, fill in the gap with the correct nucleotide, and finally reseal the remaining nick. The human homologs of these DNA mismatch repair genes include hMSH2, hMSH3, hMSH4, hMSH5, hMSH6, hMLH1, hMLH3, hPMS1, and hPMS2, and likely others. The genes hMSH2 and hMLH1 are the two DNA mismatch repair genes that are most frequently mutated at the germline level in the hereditary nonpolyposis colorectal cancer (HNPCC) syndrome, also known as Lynch syndrome.69,70 Mutations can lead to functional alterations that allow strand slippage during replication. Affected cells are called replication error (RER)–positive, in contrast to the RER-negative phenotype.71,72 Because microsatellite DNA sequences are primarily affected by this type of genetic instability, the tumor cells are said to display microsatellite instability (MSI). DNA mismatch repair genes are mutated not only in Lynch syndrome, but also in a subset of sporadic GI cancers, including many arising in the esophagus, stomach, pancreas, and colon. Mechan­ istically, the absence of DNA repair does not cause cancer directly. Rather, the DNA repair defect creates a milieu that permits the accumulation of mutations in a variety of other genes that contain microsatellite DNA sequences, such as the TGF-β type II receptor, BAX, IGF type II receptor, and E2F-4. This MSI pathway represents a novel mechanism for the accumulation of mutations within a tumor (see Fig. 3-4). It is characteristic of all Lynch-related tumors and is observed in approximately 15% of all sporadic colon cancers. Errors can also be introduced when individual nucleotides are damaged by chemical factors; the base excision repair system corrects these types of errors. 8-Oxoguanine residues can result from oxidative DNA damage, and these altered bases will inappropriately pair with adenines, ultimately leading to G:C→T:A mutations if uncorrected. MYH is a DNA glycosylase that participates in the repair of these oxidized guanine nucleotides. An autosomal recessive adenomatous polyposis syndrome caused by germline mutations in the MYH repair gene has been identified.73,74 Interestingly, G:C→T:A mutations in the APC gene were almost universally found in the polyps of patients with germline MYH mutations, indicating that there are important similarities in the molecular pathogenesis of polyps in the MYH and FAP syndromes.

ONCOGENIC SIGNALING PATHWAYS

Individual oncogenes or tumor suppressor genes do not necessarily induce cellular transformation directly but typically function as components of larger oncogenic signaling pathways. Some of the pathways that are particularly relevant for gastrointestinal tumorigenesis include the

Chapter 3  Cellular Growth and Neoplasia Wnt and ras signaling pathways. These are pathways that regulate normal tissue homeostasis but become oncogenic when the signals are transduced in an aberrant or amplified manner. The key features of Wnt signaling are illustrated in Figure 3-4. β-Catenin is translocated from the plasma membrane to the cytoplasm. There, it forms a macromolecular complex with the APC protein, glycogen synthase kinase-3β (GSK-3β), and Axin. Phosphorylation of β-catenin by GSK-3β triggers its degradation. In the presence of an active Wnt signal, β-catenin is stabilized, and it enters the nucleus where it interacts with the transcription factor TCF-4 to upregulate a number of key target genes, including c-Myc, cyclin D1, and VEGF. As discussed earlier, Wnt signaling is essential for regulating proliferation of normal intestinal epithelium, and dysregulated Wnt signaling is an almost universal feature of all colon cancers. The latter can result from a mutation in the APC, Axin, or β-catenin genes, but alterations in the APC tumor suppressor gene are the most common. An alteration in just one of these components is sufficient to activate the entire pathway. Thus, it is essential to consider individual genetic alterations in the context of the overall signaling pathway in which they function. Because pathways are typically not linear, additional levels of complexity arise. There is frequent overlap among pathways, and the distinction between pathways can be somewhat arbitrary. For example, mutations in the K-ras oncogene result in the activation of multiple distinct signaling pathways, including Raf/ERK/MAPK, PI3K/Akt, and NF-κB, all of which play an important role in tumorigenesis (Fig. 3-7). Crosstalk between these effector pathways serves to modulate the cellular responses further. Akt, a target of PI3K, can phosphorylate Raf and thereby regulate signaling through the MAPK pathway.75 Finally, each of these signaling pathways regulates multiple biological processes related to tumorigenesis,76 including cell cycle progression, apoptosis, senescence, angiogenesis, and invasion. Another pathway that plays a particularly important role in gastrointestinal tumors is the cyclooxygenase-2 (COX2) pathway. The enzyme COX-2 is a key regulator of pros­ taglandin synthesis that is induced in inflammation and neoplasia. Although no mutations of COX-2 have been described, overexpression of COX-2 in colonic adenomas

K-ras

RALGDS

Ral

RAF

PI3K

MEK

Akt

JNK

NF-κB

ERK AP-1 Figure 3-7.  Diversity of signaling through K-ras. Oncogenic K-ras can activate multiple signaling pathways. The mechanisms that determine which pathway may be preferentially activated in a given cell type are not fully defined. Crosstalk between these pathways increases the complexity of the signaling networks. These effector pathways can influence cellular biological processes including proliferation, apoptosis, differentiation, and motility.

and cancers is associated with tumor progression and angiogenesis, primarily through the induction of synthesis of prostaglandin E2. Inhibition of COX-2 with a variety of agents (aspirin, nonsteroidal anti-inflammatory drugs, or COX-2 selective inhibitors) is associated with a reduced risk of colorectal adenomas and cancer.77

ENVIRONMENTAL MUTAGENESIS Fundamentally, cancer is a genetic disorder. Environmental factors play an important role in tumorigenesis, but they ultimately lead to the expression of abnormal genes or inappropriate expression of normal genes, the products of which confer the malignant phenotype. Genetic mutation is the common denominator of agents or mechanisms that contribute to the development of neoplasia. Somatic mutations can result from any class of carcinogen, including chemical mutagens and ionizing and ultraviolet radiation. Dietary constituents and their metabolites may act as important environmental mutagens within the GI tract. Viral agents also can lead to disruption of normal genes by entry into the host genome in a position that disrupts normal gene sequences (insertional mutagenesis) or through the introduction of aberrant genes present in the virus’s own genetic material. Viral agents that appear to play a role in oncogenesis in the GI tract through insertional mutagenesis include human papillomavirus in squamous cell cancers of the esophagus and anus, Epstein-Barr virus in gastric lymphoepithelial malignancies, and hepatitis B virus in hepatocellular carcinoma. Ironically, many of these viral oncogenes originated as host cellular genes that were captured at some time in the past when a viral ancestor was present as a lysogen in an ancestral host genome.

CHEMICAL CARCINOGENESIS

Metabolic activation by the host is a key determinant of the carcinogenic potential of many compounds. The initial compound, the procarcinogen, is converted by host enzymes to an electrophilic derivative, which then chemically modifies DNA. Mutations result from errors that occur during DNA replication as a result of distorted base pairs. These mutations, in conjunction with other tumor-promoting factors, facilitate or cause the development of malignancy. Factors that influence the potency of any chemical carcinogen include the equilibrium between the activation of the procarcinogen and deactivation or degradation of the carcinogen.78 Deactivation typically occurs through a conjugation reaction, usually in the liver. These principles are exemplified by experimental colonic carcinomas that arise in rodents fed cycasin, a glucosylated compound present in the cycad nut. The glucose residue of cycasin is cleaved in the rat liver by β-glucosidase to form methylazoxymethanol (MAM), which is subsequently deformylated by enzymes in the liver and colon to give rise to methyldiazonium, a carcinogen. These same metabolites are formed through hepatic enzymatic modification of the compound dimethylhydrazine and result in colon cancer in the rat.

DIETARY FACTORS

Chemical mutagenesis may be especially important in the development of cancers within the GI tract and related organs. The mucosal surfaces from which most primary cancers in the GI tract develop are exposed to a complex mixture of dietary constituents that are potential carcinogens or procarcinogens. The ability of dietary factors to act

41

42

Section I  Biology of the Gastrointestinal Tract as mutagens in humans was demonstrated directly in 1995. The frequency of contamination of foodstuffs with aflatoxins, a fungal metabolite, parallels the incidence of hepatocellular carcinoma in various areas of the world.79 Studies demonstrating that aflatoxins cause mutations in the TP53 gene in hepatocellular carcinoma have provided a com­ pelling link between genes and the environment.79 Nitrates present in many foods appear to be additional dietary constituents that may act as procarcinogens in the GI tract. Diet-derived nitrates can be converted by bacterial action in a hypochlorhydric stomach to nitrites and subsequently to mutagenic nitrosamines.80 These events may underlie the documented correlation between dietary intake of foods high in nitrates and the incidence of gastric cancer in different populations. Other dietary factors may modulate the biological potency of dietary procarcinogens. Variations in the relative and absolute amounts of dietary fats may lead to alterations in the composition of the colonic microflora and their metabolic characteristics, resulting in modulation of the production of enzymes that convert dietary constituents into potentially mutagenic compounds. Changes in dietary fiber content can alter the transit time of luminal contents in the bowel, thereby changing the duration of exposure of the mucosa to potential mutagens. Bile salt content may be an additional luminal factor that can modulate the biolo­gical effect of procarcinogens. Deconjugated bile salts may promote carcinogenesis through mucosal injury and enhanced epithelial proliferation. These mechanisms could explain well-documented correlations between the intake of various dietary constituents and the incidence of colon cancer in certain populations. Populations that have a high fiber intake and resulting fast colonic transit times generally exhibit a lower incidence of colon cancer than populations with low fiber intake and delayed transit. The incidence of colon cancer in Japanese immigrants to the United States who consume a Western diet is much higher than that of native Japanese who consume a traditional Japanese diet.81

BIOLOGICAL FEATURES OF TUMOR METASTASIS The establishment of distant metastasis requires multiple processes, many of which involve alterations in interactions between tumor cells and normal host cells. To metastasize, a cell or group of cells must detach from the primary tumor, gain access to the lymphatic or vascular space, adhere to the endothelial surface at a distant site and penetrate the vessel wall to invade the second tissue site and, finally, proliferate as a second tumor focus. Angiogenesis is necessary for proliferation of the primary tumor and tumor metastases. Tumor cells must also overcome host immune cell killing. As a result, few circulating tumor cells (less than 0.01%) successfully initiate metastatic foci. A “survival of the fittest” view of metastasis has been proposed, in which selective competition favors metastasis of a subpopulation of cells from the primary site.82 Clonal expansion occurs again after formation of a metastatic focus.

EPITHELIAL-MESENCHYMAL TRANSITION

Modulation of tumor cell interactions with adjacent cells and with the extracellular matrix is an essential step as tumor cells invade through the basement membrane and ultimately metastasize to distant sites. A similar process occurs during normal embryogenesis, when polarized epi-

High-grade dysplasia/ carcinoma in situ Basement membrane Loss of E-cadherin Stroma

EMT EMT

Blood vessel Metastases

Figure 3-8.  An epithelial-mesenchymal transition (EMT) provides a model for tumor progression and invasion. Epithelial cells maintain their polarity and boundaries with adjacent cells through many junctional proteins, including E-cadherin. The loss or down-regulation of E-cadherin is a key feature in EMT, wherein epithelial cells can adopt a migratory mesen­ chymal phenotype. In tumor progression, EMT can occur at multiple levels, including the transition from early carcinoma in situ to invasive cancer, as well as the invasion of a tumor cell into blood and lymphatic vessels.

thelial cells no longer recognize the boundaries imposed by adjacent epithelial cells or their basement membrane and adopt features of migratory, mesenchymal cells. This phenomenon, designated epithelial-mesenchymal transition (EMT), has provided a new model for understanding tumor progression (Fig. 3-8). E-cadherin is a critical com­ ponent of adherens junctions that maintain cell-cell interactions, and loss of E-cadherin is one of the key features of the EMT phenotype.83 Mutations in E-cadherin are common in many GI cancers, particularly gastric cancer. E-cadherin gene expression can be down-regulated by the transcriptional repressors Snail, SIP1, and Twist,84-86 but it is not yet clear whether these are relevant in GI cancers. The epithelial basement membrane consists of a dense matrix of collagen, glycoproteins, and proteoglycans and normally does not permit passive penetration of cells. The transmigration of tumor cells through the basement membrane likely involves production of key proteolytic activities. Alternatively, the tumor cell may produce factors capable of activating proenzymes present in the extracellular matrix. For example, the tumor may produce urokinase, itself a protease, or plasminogen activator. Having gained access to the interstitial stromal compartment, tumor cells can then enter lymphatic and blood vessels and metastasize.

ANGIOGENESIS AND LYMPHANGIOGENESIS

Angiogenesis is essential to sustain continued growth of the primary tumor. If new vessels are not developed as the primary tumor expands, cells most distant from available vessels are deprived of an adequate source of nutrition and central necrosis occurs. Neovascularization is also an important permissive factor in facilitating metastatic dissemination of tumors.87 A number of protein growth factors, produced by malignant tumor cells and stromal cells, have been found to be potent stimuli of angiogenesis, including vascular endothelial growth factor A (VEGF-A), basic fibro-

Chapter 3  Cellular Growth and Neoplasia blast growth factor (bFGF) and TGF-β. VEGF-A is perhaps the most critical factor that is up-regulated in most tumor types, including colon cancer. Multiple genetic pathways modulate VEGF-A expression, including Wnt and mutant ras.88 Therapeutic strategies that inhibit VEGF-A have demonstrated some promise for patients with advanced colon cancer.89 Angiogenesis occurs in an ordered series of events. Endothelial cells in the parent vessel are stimulated to degrade the endothelial basement membrane, migrate into the perivascular stroma, and initiate a capillary sprout. The sprout develops into a tubular structure that in turn develops into a capillary network. In vitro models that recapitulate the early events of angiogenesis indicate that this process involves a balance between proteases and protease inhibitors in a manner similar to that during tumor invasion. Indeed, functional parallels between tumor invasion and angiogenesis are evident in their mutual requirement for cellular motility, basement membrane proteolysis, and cell growth. In addition to angiogenesis, lymphangiogenesis plays an important role in tumor metastasis. Some important clues into the molecular basis of tumor lymphangiogenesis have been obtained. VEGF-C or VEGF-D bind to the VEGF receptor-3 on lymphatic endothelial cells to stimulate the formation of new lymphatic vessels.90 This results in the development of new lymphatic channels within the tumor mass and, consequently, the enhanced dissemination of tumor cells to regional lymph nodes.91 Strategies to inhibit tumor lymphangiogenesis are being actively pursued.

METASTASIS GENES

It is likely that properties important to the development of metastasis reflect the effects of genes distinct from oncogenes involved in the initial formation of the tumor. Although no gene specifically associated with metastasis has yet been identified, one gene, designated nm-23, may be a potential metastasis suppressor gene. Levels of nm-23 are reduced in a variety of metastatic tumors and in cell lines with high metastatic potential, compared with primary tumors and cell lines with low metastatic potential, respectively.92 The function of nm-23 is uncertain but may be inferred from its predicted sequence homology to nucleoside diphosphate (NDP) kinases. NDP kinases are involved in microtubule assembly and signal transduction through G

proteins, functions that may be important in the formation of metastases.

MOLECULAR MEDICINE: CURRENT AND FUTURE APPROACHES IN GASTROINTESTINAL ONCOLOGY DNA-BASED APPROACHES Progress in the identification of cancer-associated genes coupled with the inherent power of molecular biological techniques to analyze exquisitely small amounts of DNA and protein are leading to more effective diagnostic markers (Table 3-3). The most immediate application is assessment of cancer risk in members of cancer-prone kindreds. Strategies have been developed to identify germline mutations in patients with a variety of inherited GI cancer syndromes, including FAP, HNPCC, and hereditary gastric cancer (Table 3-4). In most of these conditions, there is no consensus hot spot mutational site, so these tests analyze the full gene through a variety of analytic techniques (see Table 3-3). Genetic testing is a powerful tool to identify high-risk families and to define the cancer risk for individual family members. Application of genetic testing must take into consideration the sensitivity and specificity of the assay as well as issues of patient confidentiality and potential impact on medical insurability. For these reasons, genetic counseling is an essential component of the genetic testing process. Improved detection of sporadic GI cancers and their precursor lesions has also been the focus of research studies. Small numbers of shed cells obtained from stool can be assessed for the presence of mutations in specific tumorassociated genes (K-ras, APC, and TP53) using the PCR assay.93 Detection of ras mutations in DNA extracted from the pancreatic ductal fluid obtained at the time of endoscopic retrograde cholangiopancreatographic evaluation for pancreatic cancer has also been reported.94 The MSI test can be performed on archived colon tumor samples and serves as a useful screening test to identify individuals whose colon cancers may have developed as a manifestation of the Lynch syndrome.95 Loss of hMSH2, hMLH1, or hMSH6 protein by immunohistochemical staining may provide similar information. Studies have suggested that the MSI status of a colon tumor may be predictive of the response to 5-fluorouracil–based chemotherapy.96

Table 3-3  Molecular Diagnostic Techniques for Detection of Cancer-Associated DNA Mutations or Altered Proteins Technique PCR-Based Strategies to Detect DNA Mutations Single-strand conformational polymorphism (SSCP) Denaturing gradient gel electrophoresis (DGGE) Heteroduplex analysis Heteroduplex mismatch cleavage Direct DNA sequencing PCR-Based Strategies to Detect Known Mutations in Genes* Restriction enzyme digestion Allele-specific oligonucleotide hybridization Protein-Based Strategies In vitro translation (IVT) Yeast and bacterial colorimetric assays Immunohistochemistry

Purpose or Strategy Detection of alteration of secondary structure of single-stranded DNA caused by single base mutation Detection of strand dissociation of double-stranded DNA altered by mutations Detection of altered electrophoretic migration caused by mutations Detection of chemical cleavage of mismatches in heteroduplexes Direct detection of altered DNA nucleotide sequence Detection of mismatched primers followed by enzymatic cleavage Hybridization of specific oligonucleotides with wild-type or mutant sequence Detection of truncated protein resulting from nonsense mutation and a premature stop codon Detection of altered colorimetric assay caused by mutation Determination of presence or absence of gene product in tumor sample

43

44

Section I  Biology of the Gastrointestinal Tract Table 3-4  Applications of Molecular Diagnostics for Gastrointestinal Cancers Disorder

Gene(s) Detected

Germline DNA Analysis for Hereditary GI Cancer Syndromes FAP, AFAP APC Lynch, HNPCC hMSH2, hMLH1, hMSH6, hPMS2 MYH polyposis MYH Peutz-Jeghers syndrome LKB1 Cowden’s disease PTEN Juvenile polyposis SMAD4, BMPR1A Hereditary gastric cancer E-cadherin Hereditary pancreatic cancer p16INK4A, BRCA2 MEN1 Menin Molecular Analysis for the Diagnosis of Sporadic GI Cancers Colon cancer Stool DNA testing K-ras, APC, TP53 Tumor DNA MSI testing Tumor immunohistochemistry for hMSH2, hMLH1, hMSH6, hPMS2 protein APC, adenomatous polyposis coli; FAP, familial adenomatous polyposis; AFAP, attenuated FAP; HNPCC, hereditary nonpolyposis colorectal cancer; MEN1, multiple endocrine neoplasia, type 1; MSI, microsatellite instability.

Therapies that target specific signaling pathways are likely to increase as our molecular understanding of GI cancers increases. Antibodies that target EGF receptors and block the EGF receptor signaling pathway have proven thera­ peutic benefit in colon cancer, and their role in treatment strategies has been evolving.97 In addition, small molecule tyrosine kinase inhibitors of the c-KIT oncogene now constitute routine treatment of gastrointestinal stromal tumors (see Chapter 30).98 Molecular techniques may also find a role in the staging of disease. For example, the PCR assay has been used to detect lymph node micrometastases from colon cancer.99 Finally, as more tests for genetic markers become available, monitoring for disease recurrence after surgery may become another important application.

ONCOFETAL PROTEINS

Characterization of malignant and transformed cells has led to the identification of markers that may be useful for the early detection and diagnosis of GI cancers. The most productive approaches have exploited the antigenicity of distinctive cell surface glycoconjugates to prepare antisera or monoclonal antibodies directed against tumor-associated determinants. The first useful marker developed through this approach was the carcinoembryonic antigen (CEA),

which was identified by Gold and coworkers after immunization of rabbits with colorectal cancer tissue.100 The resulting antisera were found to recognize a determinant present in tumor tissue and circulating in blood from patients with colorectal cancer but largely absent from normal colonic mucosa and normal serum. This oncofetal determinant is also expressed in nonmalignant mucosa in association with increased proliferation. On a practical level, however, the CEA concentration is falsely elevated in a variety of inflammatory conditions associated with increased cell turnover, such as ulcerative colitis. In addition, it was noted that CEA could be produced by tumors arising from many sites, particularly those elsewhere in the GI tract (e.g., gastric and pancreatic cancers). This finding underscores the relatively limited tissue specificity of transformation-associated alterations in cell surface determinants. Future strategies to identify new protein markers that may be useful for diagnosis, therapy, or prognosis will rely on emerging proteomic techniques.

KEY REFERENCES

Brown JR, DuBois RN. COX-2: A molecular target for colorectal cancer prevention. J Clin Oncol 2005; 23:2840-55. (Ref 77.) Cech TR. Beginning to understand the end of the chromosome. Cell 2004; 116:273-9. (Ref 14.) Clarke MF, Dick JE, Dirks PB, et al. Cancer stem cells—perspectives on current status and future directions: AACR Workshop on cancer stem cells. Cancer Res 2006 66:9339-44. (Ref 27.) Erichsen HC, Chanock SJ. SNPs in cancer research and treatment. Br J Cancer 2004; 90:747-51. (Ref 49.) Giardiello FM, Brensinger JD, Petersen GM. AGA technical review on hereditary colorectal cancer and genetic testing. Gastroenterology 2001; 121:198-213. (Ref 95.) Herman JG, Baylin SB. Gene silencing in cancer in association with promoter hypermethylation. N Engl J Med 2003; 349:2042-54. (Ref 50.) Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004; 350:2335-42. (Ref 89.) Lin WW, Karin M. A cytokine-mediated link between innate immunity, inflammation, and cancer. J Clin Invest 2007; 117:1175-83. (Ref 25.) Reed JC. Mechanisms of apoptosis. Am J Pathol 2000; 157:1415-30. (Ref 12.) Schlessinger J. Cell signaling by receptor tyrosine kinases. Cell 2000; 103:211-25. (Ref 15.) Sherr CJ. The Pezcoller lecture: Cancer cell cycles revisited. Cancer Res 2000; 60:3689-95. (Ref 1.) Siegel PM, Massague J. Cytostatic and apoptotic actions of TGF-beta in homeostasis and cancer. Nat Rev Cancer 2003; 3:807-21. (Ref 23.) Thiery JP. Epithelial-mesenchymal transitions in tumour progression. Nat Rev Cancer 2002; 2:442-54. (Ref 83.) Ulku AS, Der CJ. Ras signaling, deregulation of gene expression and oncogenesis. Cancer Treat Res 2003; 115:189-208. (Ref 76.) van de Wetering M, Sancho E, Verweij C, et al. The beta-catenin/TCF-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 2002; 111:241-50. (Ref 21.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

4

Nutritional Assessment and Management of the Malnourished Patient Joel B. Mason

CHAPTER OUTLINE Basic Nutritional Concepts  47 Energy Stores  47 Energy Metabolism  47 Protein  49 Carbohydrate  51 Lipids  51 Major Minerals  52 Micronutrients  52 Vitamins  52 Trace Minerals  52 Physiologic and Pathophysiologic Factors Affecting Micronutrient Requirements  58 Starvation  59 Malnutrition  60 Protein-Energy Malnutrition  60 Physiologic Impairments Caused by Protein-Energy Malnutrition  63 Nutritional Assessment Techniques  64

Diligent attention to patients’ nutritional needs can have a major positive impact on medical outcomes. This is particularly true in gastrointestinal and liver disease because many of these conditions, in addition to altering nutrient metabolism and requirements, are prone to interfere with the ingestion and assimilation of nutrients. Nutritional management, however, often continues to be an inadequately or incorrectly addressed component of patient care. In part, inadequate or misdirected attention to nutritional issues occurs because of the failure to distinguish patients who stand to benefit from nutritional care from those whose outcomes will not respond to nutritional intervention. Indeed, the fact that many clinical trials have failed to demonstrate a benefit of nutritional support in hospitalized patients can be attributed to the fact that such a distinction has not been made. The major aim of this chapter, therefore, is to provide the scientific principles and practical tools necessary to recognize patients who will benefit from focused attention to nutritional needs. Some principles regarding the initial management of the severely malnourished patient, as well as nutritional considerations in severe malabsorption, are also provided as examples of how an appreciation for pathophysiology can be translated into clinical care.

Aggressive Nutritional Support in the Hospitalized Patient  69 Malnourished Patients Undergoing Major Surgery  69 Patients Hospitalized with Decompensated Alcoholic Liver Disease  70 Patients Undergoing Radiation Therapy  70 Refeeding Syndrome  70 Mineral Depletion  70 Cardiovascular Complications  70 Glucose Intolerance  70 Gastrointestinal Dysfunction  71 Clinical Recommendations  71 Management of Severe Malabsorption: A Nutritional Perspective  71 Clinical Considerations  71 Treatment  72

BASIC NUTRITIONAL CONCEPTS ENERGY STORES

Endogenous energy stores are oxidized continuously for fuel. Triglyceride present in adipose tissue is the body’s major fuel reserve and is critical for survival during periods of starvation (Table 4-1). The high energy density and hydrophobic nature of triglycerides make it a fivefold better fuel per unit mass than glycogen. Triglycerides liberate 9.3  kcal/g when oxidized and are stored compactly as oil inside the fat cell. In comparison, glycogen produces only 4.1 kcal/g on oxidation and is stored intracellularly as a gel, containing approximately 2  g of water for every gram of glycogen. Adipose tissue is unable to provide fuel for certain tissues, such as bone marrow, erythrocytes, leukocytes, renal medulla, eye tissues, and peripheral nerves, which cannot oxidize lipids and require glucose for their energy supply. During endurance exercise, glycogen and triglycerides in muscle tissue provide an important source of fuel for working muscles.

ENERGY METABOLISM

Energy is required continuously for normal organ function, maintenance of metabolic homeostasis, heat production, and performance of mechanical work. Total daily energy

47

48

Section II  Nutrition in Gastroenterology expenditure (TEE) is composed of three components: resting energy expenditure (~70% of TEE); the energy expenditure of physical activity (~20% of TEE); and the thermic effect of feeding (~10% of TEE), which is the temporary increase in energy expenditure that accompanies enteral ingestion or parenteral administration of nutrients.

Resting Energy Expenditure

Resting energy expenditure (REE) represents energy expenditure while a person lies quietly awake in an interprandial state; under these conditions, approximately 1 kcal/kg body weight is consumed/hour in healthy adults. Energy requirements of specific tissues differ dramatically (Table 4-2). The liver, intestine, brain, kidneys, and heart constitute approximately 10% of total body weight but account for approximately 75% of REE. In contrast, skeletal muscle at rest consumes approximately 20% of REE but represents approximately 40% of body weight, and adipose tissue consumes less than 5% of REE but usually accounts for more than 20% of body weight. Several empirical equations have been generated to estimate resting energy requirements (Table 4-3).1-4 These equations are useful in healthy subjects because they generate values that are usually within 10% of measured values. These equations are much less accurate, however, for persons who are at extremes in weight or who are ill, because alterations in body composition and metabolic stress influence energy expenditure. Protein-energy malnutrition and hypocaloric feeding without superimposed illness each decrease REE to values 10% to 15% below those expected for actual body size, whereas acute illness or trauma predictably increases energy expenditure (see later).

Energy Expenditure of Physical Activity

The effect of physical activity on energy expenditure depends on the intensity and duration of daily activities. Highly trained athletes can increase their TEE 10- to 20-fold during athletic events. The activity factors shown in Table 4-4, each expressed as a multiple of REE, can be used to estimate TEE in active patients. The energy expended during physical activity is equal to REE × activity factor × duration of activity in hours/24 hours. TEE represents the summation of energy expended during all daily activities, including rest periods.

Thermic Effect of Feeding

Eating or infusing nutrients increases metabolic rate. Dietary protein causes the greatest stimulation of metabolic rate, followed by carbohydrate and then fat. A meal containing all these nutrients usually increases metabolic rate by 5% to 10% of ingested or infused calories.

Recommended Energy Intake in Hospitalized Patients

In arriving at a nutritional plan for hospitalized patients, it rarely is necessary to obtain actual measurements with a bedside indirect calorimeter. A number of simple formulas can be used instead and make up in practical value what they lack in accuracy. A few examples follow. Methods Incorporating Metabolic Stress Factors Metabolic stress—that is, any injury or illness that incites some degree of systemic inflammation—will increase the metabolic rate through a variety of mechanisms (see later). The increase in energy expenditure is roughly proportional to the magnitude of the stress.5 Thus, the equations in Table 4-3 may be used to estimate the total energy requirement of an acutely ill patient by multiplying the predicted REE by a stress factor: TEE = REE × stress factor

Table 4-1  Endogenous Fuel Stores in a 70-kg Man Mass Tissue

Fuel Source

Grams

Kilocalories

Adipose Liver

Triglyceride Protein Glycogen Triglyceride Protein Glycogen Triglyceride Glucose Triglyceride Free fatty acids

13,000 300 100 50 6,000 400 250 3 4 0.5

121,000 1,200 400 450 24,000 1,600 2,250 12 37 5

Muscle Blood

In acutely ill hospitalized patients, it is usually not necessary to include an activity factor. Although determination of the degree of stress is subjective, its use generates a caloric goal that closely approximates actual values. Table 4-5 delineates metabolic stress factors that accompany some common conditions and clinical scenarios in inpatients. An alternative and rather simple formula for adult inpatients that also incorporates a gauge of metabolic stress is as follows: • 20-25  kcal/kg of actual body weight (ABW)/day for unstressed or mildly stressed patients • 25-30 kcal/ABW/day for moderately stressed patients • 30-35 kcal/ABW/day for severely stressed patients

Table 4-2  Resting Energy Requirements of a 70-kg Man Tissue Mass Tissue Liver Gastrointestinal tract Brain Kidneys Heart Skeletal muscle Adipose REE, resting energy expenditure.

Energy Consumed

Grams

Percentage Body Weight

kcal/g Tissue/Day

kcal/Day

1,550 2,000 1,400 300 300 28,000 15,000

2.2 3.0 2.0 0.4 0.4 40.0 21.0

0.28 0.15 0.30 1.27 0.80 0.014 0.005

445 300 420 360 235 400 80

Percentage REE 19 13 18 15 10 18 4

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient Table 4-3  Commonly Used Formulas for Calculating Resting Metabolic Rate

Adjusted IBW = IBW + 0.33 ( ABW − IBW )

Harris-Benedict Equation 66 + (13.7 × W) + (5 × H) − (6.8 × A) 665 + (9.6 × W) + (1.8 × H) − (4.7 × A)

Men Women

World Health Organization Formula Age (yr)

Male

Female

0-3 3-10 10-18 18-30 30-60 >60

(60.9 × W) − 54 (22.7 × W) − 495 (17.5 × W) + 651 (15.3 × W) + 679 (11.2 × W) + 879 (13.5 × W) + 987

(60.1 × W) − 51 (22.5 × W) + 499 (12.2 × W) + 746 (14.7 × W) + 996 (8.7 × W) + 829 (10.5 × W) + 596

A, age in years; H, height in centimeters; W, weight in kilograms.

Table 4-4  Relative Thermic Effect of Various Levels of Physical Activity Activity Level Resting Very light Light Moderate Heavy

Examples Standing, driving, typing Walking 2-3 miles/hr shopping, light housekeeping Walking 3-4 miles/hr, biking, gardening, scrubbing floors Running, swimming, climbing, basketball

Activity Factor 1.0 1.1-2.0 2.1-4.0 4.1-6.0 6.1-10.0

Adapted from Alpers DA, Stenson WF, Bier DM. Manual of nutritional therapeutics. Boston: Little, Brown; 1995.

Table 4-5  Metabolic Stress Factors for Estimating Total Energy Expenditure in Hospitalized Patients Injury or Illness Second- or third-degree burns, >40% BSA Multiple trauma Second- or third-degree burns, 20%-40% BSA Severe infections Acute pancreatitis Second- or third-degree burns, 10%-20% BSA Long bone fracture Peritonitis Uncomplicated postoperative state

overestimation of energy requirements and is calculated as follows:

relative Stress Factor* 1.6-1.8 1.5-1.7 1.4-1.5 1.3-1.4 1.2-1.4 1.2-1.4 1.2 1.2 1.1

*A stress factor of 1.0 is assumed for healthy controls. BSA, body surface area.

Also, in patients who have large artifactual increases in weight because of extracellular fluid retention, such as the patient with ascites, the IBW should be used to estimate energy requirements, rather than the ABW. Method Without a Stress Factor Table 4-7 outlines a simple method for estimating total daily energy requirements in hospitalized patients based on body mass index (BMI).6 With this method, energy expressed per kilogram is inversely related to BMI. Common sense needs to be applied when using any means to estimate energy expenditure in hospitalized individuals because illness commonly interjects artifacts into these calculations (e.g., ascites). Over the past two decades, the trend generally has been toward a more conservative approach to caloric delivery in acutely ill patients. One reason for this conservatism is that acute illness and its management often exacerbate preexisting diabetes or produces de novo glucose intolerance. As a result, hyperglycemia is a frequent consequence of enteral, and especially parenteral, nutrition. The issue seems to be particularly germane for intensive care unit (ICU) patients, in whom even modest hyperglycemia results in worse clinical outcomes, usually of an infectious nature. Clinical trials of high quality in surgical ICU (SICU)7 and medical ICU (MICU)8 patients have found that morbidity is substantially and significantly reduced in those randomized to intensive insulin therapy who maintained serum glucose levels below 111 mg/dL compared with those whose glucose values were maintained below 215  mg/dL. Mortality also was significantly lower in those in the SICU randomized to receive tight glucose control, although in the MICU study such reductions in mortality caused by tight glucose control only were realized in those residing in the MICU longer than three days. These clinical observations substantiate years of animal studies showing that even modest hyperglycemia impairs immune function in a variety of tissues.9 The clinical benefits of tight glucose control in the ICU, however, have not always been reproducible,10 and come at the cost of more frequent hypoglycemic episodes,7,8,10 so the issue of how tight glucose control should be remains controversial. Results of a recent meta-analysis of 29 trials in critically ill patients recapitulate the previously observed discrepancies between SICU and MICU patients.11 Overall, the relative risk of septicemia was reduced approximately 25% in those randomized to tight glucose control, but this salutary effect largely was attributable to the SICU patients, in whom the reduction in septicemia was almost 50%. In contrast, no benefit was observed in MICU patients. Also in this meta-analysis, no demonstrable benefit in overall mortality was evident in any of the categories of critically ill patients.

PROTEIN In using this formula, adjustments are necessary when the ABW is a misleading reflection of lean body mass. An adjusted ideal body weight should be substituted for ABW in obese individuals who are more than 30% heavier than their ideal body weight (desirable body weight, more commonly referred to as ideal body weight [IBW], appears in Table 4-6). The use of an adjusted IBW helps prevent an

Twenty different amino acids are found commonly in human proteins. Some amino acids (histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, valine, and possibly arginine) are considered essential because their carbon skeletons cannot be synthesized by the body. Other amino acids (glycine, alanine, serine, cysteine, cystine, tyrosine, glutamine, glutamic acid, asparagine, and aspartic acid) are nonessential in most circumstances because they can be made from endogenous

49

50

Section II  Nutrition in Gastroenterology Table 4-6  Desirable Weight in Relation to Height for Men and Women 25 Years or Older* Men, Medium Frame

Women, Medium Frame

Weight (lb) Height (ft/inches)

5′1″ 5′2″ 5′3″ 5′4″ 5′5″ 5′6″ 5′7″ 5′8″ 5′9″ 5′10″ 5′11″ 6′0″ 6′1″ 6′2″ 6′3″

Range

Midpoint

113-124 116-128 119-131 122-134 125-138 129-142 133-147 137-151 141-155 145-160 149-165 153-170 157-175 162-180 167-185

118.5 122 125 128 131.5 135.5 140 144 148 153 157 161.5 166 171 176

Weight (lb) Height (ft/inches)

Range

Midpoint

4′8″ 4′9″ 4′10″ 4′11″ 5′0″ 5′1″ 5′2″ 5′3″ 5′4″ 5′5″ 5′6″ 5′7″ 5′8″ 5′9″ 5′10″

93-104 95-107 98-110 101-113 104-116 107-119 110-123 113-127 117-132 121-136 125-140 129-144 133-148 137-152 141-156

98.5 101 104 107 110 113 116.5 120 124.5 128.5 132.5 136.5 140.5 144.5 148.5

*Corrected to nude weights and heights by assuming 1-inch heel for men, 2-inch heel for women, and indoor clothing weight of 5 and 3 lb for men and women, respectively. Data from Metropolitan Life Insurance Company. New height standards for men and women. Statistical Bulletin 1959;40:1-4.

Table 4-7  Estimated Energy Requirements for Hospitalized Patients Based on Body Mass Index (BMI)*

Table 4-8  Recommended Daily Protein Intake*

Body Mass Index (BMI; kg/m2)

Clinical Condition

<15 15-19 20-29 ≥30

Energy Requirements (kcal/kg/day)† 35-40 30-35 20-25 15-20

*The lower range within each BMI category should be considered in calculating energy requirements for insulin-resistant or critically ill patients to decrease the risk of hyperglycemia and infection associated with overfeeding. † These values are recommended for critically ill patients and all obese patients; add 20% of the total calories when estimating energy requirements in non–critically ill patients.

precursors or essential amino acids. In disease states, non­ essential amino acids may become essential—so-called conditionally essential amino acids. Thus, infectious morbidity and wound healing appear to be improved in critically ill patients by the inclusion of supplemental glutamine in total parenteral nutrition (TPN) because of cellular depletion of this amino acid. Studies to date suggest that the clinical benefits of providing supplemental quantities of parenteral glutamine only are realized in patients with particularly high severity of illness scores (e.g., high APACHE II or SOFA scores), although the true nature of its efficacy likely will remain controversial until the results of several ongoing clinical trials become known.12 Similarly, it has been shown that cysteine and tyrosine are essential in some patients with cirrhosis13 because of impaired hepatic synthesis. The body of an average 75-kg man contains approximately 12 kg of protein. In contrast to fat and carbohydrate, there is no storage depot for protein, so excess intake is catabo-

Normal Metabolic stress Hemodialysis Peritoneal dialysis

Daily Protein Requirement (g/kg IBW) 0.75 1.0-1.6 1.2-1.4 1.3-1.5

*Additional protein requirements are needed to compensate for excess protein loss in specific patient populations (e.g., patients with burn injuries, open wounds, protein-losing enteropathy, or nephropathy). Lower protein intake may be necessary for patients with renal insufficiency not treated by dialysis and certain patients with liver disease and hepatic encephalopathy. IBW, ideal body weight.

lized and the nitrogen component is excreted. Inadequate protein intake causes net nitrogen losses and, because no depot form of protein exists, there is an obligatory net loss of functioning protein. The U.S. Recommended Daily Allowance (RDA) of protein has been established at 0.8 g/ kg/day, which reflects a mean calculated requirement of 0.6 g/kg/day plus an added factor to take into account the biological variance in requirement observed in a healthy population. Intravenously administered amino acids are as effective in maintaining nitrogen balance as oral protein of the same amino acid composition.14 Individual protein requirements are affected by several factors, such as the amount of nonprotein calories provided, overall energy requirements, protein quality, and the patient’s nutritional status (Table 4-8). Protein requirements increase when calorie intake does not meet energy needs. The magnitude of this increase is directly proportional to the deficit in energy supply. Therefore, nitrogen balance reflects protein intake and energy balance. Thus, correcting

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient a negative nitrogen balance sometimes may be achieved merely by increasing the caloric delivery if the total amount of calories has been inadequate. As metabolic stress (and with it, metabolic rate) increases, nitrogen excretion increases proportionately; quantitatively, the relationship is approximately 2 mg nitrogen (N)/kcal of REE. In part, this increase is explained by the fact that in metabolic stress, a larger proportion of the total substrate that is oxidized for energy is from protein. This has two important implications for managing the nutritional needs of ill patients. The first is that illness, by increasing catabolism and metabolic rate, increases the absolute requirement for protein (see Table 4-8), and does so in a manner that is roughly proportional to the degree of stress. Second, because a greater proportion of energy substrate in acute illness comes from protein, nitrogen balance is more readily achieved if a larger proportion of the total calories are from protein. In healthy adults, as little as 10% of total calories need to come from protein to maintain health, whereas in the ill patient, nitrogen balance is achieved more easily if 15% to 25% of total calories are delivered as protein. Protein requirements also are determined by the adequacy of essential amino acids in the protein source. Inadequate amounts of an essential amino acid result in inefficient uptake, and therefore proteins of low biological quality increase the protein requirement. In normal adults, approximately 15% to 20% of total protein requirements should be in the form of essential amino acids. Additional proteins are needed to compensate for excess loss in specific patient populations (e.g., patients with burn injuries, open wounds, and protein-losing enteropathy or nephropathy). Delivering less protein than is needed often is a necessary compromise in patients with acute renal insufficiency who are not adequately dialyzed, because in this situation the rise in azotemia is directly proportional to protein delivery. Once adequate dialysis is available, protein delivery should be increased to the actual projected need, including additional protein to compensate for losses resulting from dialysis (see Table 4-8). Most patients with hepatic encephalopathy respond to simple pharmacologic measures and therefore do not require a protein restriction; however, those who do not respond may benefit from a modest protein restriction (~0.6 g/kg/day).

Nitrogen Balance

Nitrogen (N) balance commonly is used as a proxy measure of protein balance—that is, whether the quantity of protein (or amino acids) taken in is sufficient to prevent any net loss of protein. N balance is calculated as the difference between N intake and N losses in urine, stool, skin, and body fluids. In the clinical setting, it is calculated as follows for adults: N balance = grams of N administered as nutrition − ( urinary urea N [ g ] + 4 ) Every 6.25  g of administered protein (or amino acids) contains approximately 1  g of N. The additional 4  g of N loss incorporated into the equation is intended to account for the insensible losses from the other sources listed, and because urinary urea N only accounts for about 80% of total urinary nitrogen. N balance is a suitable surrogate for protein balance because approximately 98% of total body N is in protein, regardless of the health of the individual. A positive N balance (i.e., intake greater than loss) represents anabolism and a net increase in total body protein, whereas a negative N balance represents net protein catabolism. For example, a negative N balance of 1  g/day repre-

sents a 6.25-g/day loss of body protein, which is equivalent to a 30-g/day loss of hydrated lean tissue. In practice, N balance studies tend to be artificially positive because of overestimation of dietary N intake and underestimation of losses caused by incomplete urine collections and unmeasured outputs. It is best to wait at least four days after a substantial change in protein delivery before N balance is examined because a labile N pool exists; this tends to dampen and retard changes that otherwise would be observed as a result of altered protein intake.

CARBOHYDRATE

Complete digestion of the principal dietary digestible carbohydrates—starch, sucrose, and lactose—generate monosaccharides (glucose, fructose, and galactose). In addition, approximately 5 to 20 g of indigestible carbohydrate (soluble and insoluble fibers) are consumed daily. All cells are able to generate energy (adenosine triphosphate) by metabolizing glucose to three-carbon compounds via glycolysis or to carbon dioxide and water via glycolysis and the tricarboxylic acid (TCA) cycle. There is no absolute dietary requirement for carbohydrate because glucose can be synthesized from endogenous amino acids as well as glycerol. Nevertheless, carbohydrate is an important fuel because of the interactions between carbohydrate and protein metabolism. Carbohydrate intake stimulates insulin secretion, which inhibits muscle protein breakdown,15 stimulates muscle protein synthesis,16 and decreases endogenous glucose production from amino acids.17 In addition, glucose is the required or preferred fuel for red and white blood cells, the renal medulla, eye tissues, peripheral nerves, and the brain. Once glucose requirements for these tissues are met (∼150 g/day), however, the proteinsparing effects of carbohydrate and fat are similar.18

LIPIDS

Lipids consist of triglycerides (TGs), sterols, and phos­ pholipids. These compounds serve as sources of energy, precursors for steroid hormone, prostaglandin, thromboxane, and leukotriene synthesis, structural components of cell membranes, and carriers of essential nutrients. Dietary lipids are composed mainly of TGs, which contain saturated and unsaturated long-chain fatty acids (FAs) of 16 to 18 carbons. The use of fat as a fuel requires the hydrolysis of endogenous or exogenous TGs and cellular uptake of released FAs. Long-chain FAs are delivered across the outer and inner mitochondrial membranes by a carnitinedependent transport system.19 Once inside the mitochondria, FAs are degraded by beta oxidation to acetyl coenzyme A (CoA), which then enters the TCA cycle. Therefore, the ability to use fat as a fuel depends on normally functioning mitochondria. A decrease in the number of mitochondria or oxidative enzymes associated with aging20 or deconditioning favors the use of carbohydrate as fuel.21

Essential Fatty Acids

Humans lack the desaturase enzyme needed to produce the n-3 (double bond between carbons 3 and 4) and n-6 (double bond between carbons 6 and 7) FA series. Linoleic acid (C18 : 2, n-6) and linolenic acid (C18 : 3, n-3), therefore, should constitute at least 2% and 0.5%, respectively, of the daily caloric intake to prevent essential FA deficiency (EFAD). Before the advent of parenteral nutrition, EFAD was recognized only in infants and manifested as a scaly rash with a specific alteration in the plasma FA profile (see later). Adults previously were thought not to be susceptible to EFAD because of sufficient essential FA stores in adipose tissue. However, an abnormal FA profile in conjunction

51

52

Section II  Nutrition in Gastroenterology Table 4-9  Major Mineral Requirements and Assessment of Deficiency

Mineral

Enteral

Test

Comment

Sodium

0.5-5 g

60-150

Hypovolemia, weakness

Urinary sodium

Potassium

2-5 g

60-100

Serum potassium

Magnesium

300-400 mg

5-15

Weakness, paresthesias, arrhythmias Weakness, twitching, tetany, arrhythmias, hypocalcemia

May not reflect body stores; clinical evaluation is best May not reflect body stores

Serum magnesium

May not reflect body stores

Calcium

800-1200 mg

5-15

Urinary magnesium 24-hr urinary calcium

May not reflect body stores Reflects recent intake

Dual energy radiation absorptiometry Plasma phosphorus

Reflects bone calcium content

Phosphorus

800-1200 mg

20-60

Symptoms or Signs of Deficiency

Laboratory Evaluation

Parenteral (mmol)

Osteomalacia, tetany, arrhythmias Weakness, fatigue, leukocyte and platelet dysfunction, hemolytic anemia, cardiac failure, decreased oxygenation

with a clinical syndrome of EFAD is now known to occur sometimes in adults with severe short bowel syndrome who are on long-term total parenteral nutrition (TPN) that lacks parenteral lipids.22 Adults who have moderate to severe fat malabsorption (fractional fat excretion >20%) from other causes and who are not TPN-dependent also frequently display a biochemical profile of EFAD,23 although whether such a biochemical state carries adverse clinical consequences with it is unclear. Moreover, TPN lacking any source of fat may lead to EFAD in adults if no exogenous source of EFAs is available; the plasma pattern of EFAD may be observed as early as 10 days after glucose-based TPN is started and before the onset of any clinical features.24 In this situation, EFAD is probably attributable to the increase in plasma insulin concentrations caused by TPN, because insulin inhibits lipolysis and therefore the release of endogenous essential FAs. The biochemical diagnosis of EFAD is defined as an absolute and relative deficiency in the two EFAs in the plasma FA profile. The full clinical EFAD syndrome includes alopecia, scaly dermatitis, capillary fragility, poor wound healing, increased susceptibility to infection, fatty liver, and growth retardation in infants and children.

MAJOR MINERALS

Major minerals are inorganic nutrients that are required in large (>100 mg/day) quantities, and are important for ionic equilibrium, water balance, and normal cell function. Malnutrition and nutritional repletion can have dramatic effects on major mineral balance. The evaluation of macromineral deficiency and recommended daily allowance (RDA) of minerals for healthy adults are shown in Table 4-9.

MICRONUTRIENTS Micronutrients (the vitamins and trace minerals) are a diverse array of dietary components that are necessary to sustain health. The physiologic roles of micronutrients are as varied as their composition. Some are used in enzymes as coenzymes or prosthetic groups, others as biochemical substrates or hormones and, in some cases, their functions are not well defined. The average daily dietary intake for each micronutrient required to sustain normal physiologic operations is measured in milligrams or smaller quantities.

May not reflect body stores

In this way, micronutrients are distinguished from macronutrients (carbohydrates, fats, and proteins) and macro­ minerals (calcium, magnesium, and phosphorus). An individual’s dietary requirement for any given micronutrient is determined by many factors, including its bioavailability, the amount needed to sustain its normal physiologic functions, a person’s gender and age, any diseases or drugs that affect the nutrient’s metabolism, and certain lifestyle habits, such as smoking and alcohol use. The U.S. National Academy of Sciences Food and Nutrition Board regularly updates dietary guidelines that define the quantity of each micronutrient that is “adequate to meet the known nutrient needs of practically all healthy persons.” These RDAs underwent revision between 1998 and 2001, and the values for adults appear in Tables 4-10 and 4-11. The formulation of an RDA takes into account the biological variability in the population, and, therefore, RDAs are set 2 SDs above the mean requirement; this allows the requirements of 97% of the population to be met and ingestion of quantities that are somewhat less than the RDA usually are sufficient to meet the needs of a particular individual. A “tolerable upper limit (TUL),” which is “the maximal daily level of oral intake that is likely to pose no adverse health risks,” also has been established for most of the micronu­ trients (see Tables 4-10 and 4-11). Present recommendations for how much of each micronutrient is needed in individ­ uals on TPN are based on far less data than what were available for the development of the RDAs. Nevertheless, it is important to have guidelines, and Table 4-12 provides such recommendations.

VITAMINS

Vitamins are categorized as fat-soluble (A, D, E, K) or watersoluble (all others) (see Table 4-10). This categorization remains physiologically meaningful; none of the fat-soluble vitamins appear to serve as coenzymes, whereas almost all of the water-soluble vitamins appear to function in that role. Also, the absorption of fat-soluble vitamins is primarily through a micellar route, whereas the water-soluble vitamins are not absorbed in a lipophilic phase in the intestine (see Chapter 100).

TRACE MINERALS

Compelling evidence exists for the essential nature of 10 trace elements in humans—iron, zinc, copper, chromium, Text continued on p. 58

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient Table 4-10  Salient Features of Vitamins Vitamin

Deficiency (RDA)*

Toxicity (TUL)†

Assessment of Status

A

Follicular hyperkeratosis and night blindness are early indicators. Conjunctival xerosis, degeneration of the cornea (keratomalacia) and dedifferentiation of rapidly proliferating epithelia are later indications of deficiency. Bitot spots (focal areas of the conjunctiva or cornea with foamy appearance) are an indication of xerosis. Blindness, caused by corneal destruction and retinal dysfunction, may ensue. Increased susceptibility to infection also a consequence (1 µg of retinol equivalent to 3.33 IU of vitamin A; F, 700 µg; M, 900 µg)

Retinol concentration in the plasma, as well as vitamin A concentrations in milk and tears, are reasonably accurate measures of status. Toxicity is best assessed by elevated levels of retinyl esters in plasma. A quantitative measure of dark adaptation for night vision and electroretinography are useful functional tests.

D

Deficiency results in disordered bone modeling called rickets in childhood and osteomalacia in adults. Expansion of the epiphyseal growth plates and replacement of normal bone with unmineralized bone matrix are the cardinal features of rickets; the latter feature also characterizes osteomalacia. Deformity of bone and pathologic fractures occur. Decreased serum concentrations of calcium and phosphate may occur (1 µg equivalent to 40 IU; 5 µg, ages 19-50; 10 µg, age 51-70; 15 µg, age > 70 yr)

In adults, >150,000 µg may cause acute toxicity—fatal intracranial hypertension, skin exfoliation, and hepatocellular injury. Chronic toxicity may occur with habitual daily intake of >10,000 µg—alopecia, ataxia, bone and muscle pain, dermatitis, cheilitis, conjunctivitis, pseudotumor cerebri, hepatic fibrosis, hyperlipidemia, and hyperostosis are common. Single large doses of vitamin A (30,000 µg), or habitual intake of >4500 µg/day during early pregnancy can be teratogenic. Excessive intake of carotenoids causes a benign condition characterized by yellowish discoloration of the skin. Habitually large doses of canthaxanthin, a carotenoid, have the additional capability of inducing a retinopathy (3000 µg) Excess amounts result in abnormally high concentrations of calcium and phosphate in the serum; metastatic calcifications, renal damage, and altered mentation may occur (50 µg)

E

Deficiency caused by dietary inadequacy is rare in developed countries. Usually seen in (1) premature infants, (2) individuals with fat malabsorption, and (3) individuals with abetalipoproteinemia. Red blood cell fragility occurs and can produce hemolytic anemia. Neuronal degeneration produces peripheral neuropathies, ophthalmoplegia, and destruction of the posterior columns of the spinal cord. Neurologic disease is frequently irreversible if deficiency is not corrected early enough. May contribute to hemolytic anemia and retrolental fibroplasia in premature infants. Has been reported to suppress cell-mediated immunity (15 mg) Deficiency syndrome uncommon except in (1) breast-fed newborns, in whom it may cause “hemorrhagic disease of the newborn,” (2) adults who have fat malabsorption or are taking drugs that interfere with vitamin K metabolism (e.g., warfarin, phenytoin, broad-spectrum antibiotics), and (3) individuals taking large doses of vitamin E and anticoagulant drugs. Excessive hemorrhage is the usual manifestation (F, 90 µg; M, 120 µg)

K

Depressed levels of vitamin K–dependent procoagulants and potentiation of oral anticoagulants have been reported, as has impaired leukocyte function. Doses of 800 mg/day have been reported to increase slightly the incidence of hemorrhagic stroke (1000 mg)

Rapid intravenous infusion of vitamin K1 has been associated with dyspnea, flushing, and cardiovascular collapse; this is likely related to the dispersing agents in the dissolution solvent. Supplementation may interfere with warfarin-based anticoagulation. Pregnant women taking large amounts of the provitamin menadione may deliver infants with hemolytic anemia, hyperbilirubinemia, and kernicterus (TUL not established)

The serum concentration of the major circulating metabolite, 25-hydroxyvitamin D, is an excellent indicator of systemic status except in chronic kidney disease, in which the impairment of renal 1-hydroxylation results in dissociation of the monoand dihydroxy vitamin concentrations; measuring the serum concentration of 1,25-dihydroxyvitamin D is then necessary. Plasma or serum concentration of alpha-tocopherol is used most commonly. Additional accuracy is obtained by expressing this value per mg of total plasma lipid. Red blood cell peroxide hemolysis test is not entirely specific but is a useful functional measure of the antioxidant susceptibility of cell membranes.

The prothrombin time is typically used as a measure of functional vitamin K status; it is neither sensitive nor specific for vitamin K deficiency. Determination of undercarboxylated prothrombin in the plasma is more accurate but is less widely available.

53

Table 4-10  Salient Features of Vitamins—cont’d Vitamin

Deficiency (RDA)*

Toxicity (TUL)†

Assessment of Status

Thiamine (vitamin B1)

Classic deficiency syndrome (beriberi) is described in Asian populations consuming polished rice diet. Alcoholism and chronic renal dialysis are also common precipitants. High carbohydrate intake increases need for B1. Mild deficiency commonly produces irritability, fatigue, and headaches. More pronounced deficiency produces various combinations of peripheral neuropathy and cardiovascular and cerebral dysfunction. Cardiovascular involvement (wet beriberi) includes congestive heart failure and low peripheral vascular resistance. Cerebral disease includes nystagmus, ophthalmoplegia, and ataxia (Wernicke’s encephalopathy), as well as hallucinations, impaired short-term memory, and confabulation (Korsakoff’s psychosis). Deficiency syndrome responds within 24 hours to parenteral thiamine but is partially or wholly irreversible after a certain stage (F, 1.1 mg; M, 1.2 mg) Deficiency is usually seen in conjunction with deficiencies of other B vitamins. Isolated deficiency of riboflavin produces hyperemia and edema of nasopharyngeal mucosa, cheilosis, angular stomatitis, glossitis, seborrheic dermatitis, and normochromic, normocytic anemia (F, 1.1 mg; M, 1.3 mg)

Excess intake is largely excreted in the urine, although parenteral doses of >400 mg/day are reported to cause lethargy, ataxia, and reduced tone of the gastrointestinal tract (TUL not established)

The most effective measure of vitamin B1 status is the erythrocyte transketolase activity coefficient, which measures enzyme activity before and after addition of exogenous TPP; red blood cells from a deficient individual express a substantial increase in enzyme activity with addition of TPP. Thiamine concentrations in the blood or urine are also measured.

Toxicity not reported in humans (TUL not established)

The most common method of assessment is determining the activity coefficient of glutathione reductase in red blood cells (the test is invalid for individuals with glucose-6phosphate dehydrogenase deficiency). Measurements of blood and urine concentrations are less desirable methods. Assessment of status is problematic; blood levels of the vitamin are not reliable. Measurement of urinary excretion of the niacin metabolites N-methylnicotinamide and 2-pyridone are thought to be the most effective means of assessment.

Riboflavin (vitamin B2)

Niacin (vitamin B3)

Pyridoxine (vitamin B6)

B12

Pellagra is the classic deficiency syndrome and is often seen in populations in which corn is the major source of energy. Still endemic in parts of China, Africa, and India. Diarrhea, dementia (or associated symptoms of anxiety or insomnia), and a pigmented dermatitis that develops in sun-exposed areas are typical features. Glossitis, stomatitis, vaginitis, vertigo, and burning dysesthesias are early signs. Occasionally occurs in carcinoid syndrome because tryptophan is diverted to other synthetic pathways (F, 14 mg; M, 16 mg) Deficiency usually seen in conjunction with other water-soluble vitamin deficiencies. Stomatitis, angular cheilosis, glossitis, irritability, depression, and confusion occur in moderate to severe depletion; normochromic, normocytic anemia has been reported in severe deficiency. Abnormal EEGs and, in infants, convulsions also have been reported. Sideroblastic anemias are responsive to B6 administration. Isoniazid, cycloserine, penicillamine, ethanol, and theophylline are drugs that can inhibit B6 metabolism (Ages 19-50, 1.3 mg; >50 yr, 1.5 mg for women, 1.7 mg for men) Dietary inadequacy is a rare cause of deficiency, except in strict vegetarians. The vast majority of cases of deficiency arise from loss of intestinal absorption: this may be a result of pernicious anemia, pancreatic insufficiency, atrophic gastritis, small intestinal bacterial overgrowth, or ileal disease. Megaloblastic anemia and megaloblastic changes in other epithelia (see “Folate”) are the result of sustained depletion. Demyelination of peripheral nerves, the posterior and lateral columns of the spinal cord and nerves within the brain may occur. Altered mentation, depression, and psychoses occur. Hematologic and neurologic complications may occur independently. Folate supplementation, in doses exceeding 1000 µg/day, may partly correct the anemia, thereby masking (or perhaps exacerbating) the neuropathic complications (2.4 µg)

Human toxicity is known largely through studies examining hypolipidemic effects; includes vasomotor phenomenon (flushing), hyperglycemia, hepatocellular injury, and hyperuricemia (35 mg)

Chronic use with doses exceeding 200 mg/day (in adults) may cause peripheral neuropathies and photosensitivity (100 mg)

Many useful laboratory methods of assessment exist. The plasma or erythrocyte PLP levels are most common. Urinary excretion of xanthurenic acid after an oral tryptophan load or activity indices of RBC aminotransferases (ALT and AST) all are functional measures of B6-dependent enzyme activity.

A few allergic reactions have been reported from crystalline B12 preparations and are probably caused by impurities, not the vitamin (TUL not established)

Serum or plasma concentrations are generally accurate. Subtle deficiency with neurologic complications, as described in the “Deficiency” column, can best be established by concurrently measuring the concentration of plasma B12 and serum methylmalonic acid because the latter is a sensitive indicator of cellular deficiency.

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient Table 4-10  Salient Features of Vitamins—cont’d Vitamin

Deficiency (RDA)*

Toxicity (TUL)†

Assessment of Status

Folate

Women of childbearing age are the most likely to develop deficiency. The classic deficiency syndrome is megaloblastic anemia. The hemopoietic cells in the bone marrow become enlarged and have immature nuclei, which reflect ineffective DNA synthesis. The peripheral blood smear demonstrates macro-ovalocytes and polymorphonuclear leukocytes with an average of more than 3.5 nuclear lobes. Megaloblastic changes also occur in other epithelia that proliferate rapidly, such as the oral mucosa and gastrointestinal tract, producing glossitis and diarrhea, respectively. Sulfasalazine and diphenytoin inhibit absorption and predispose to deficiency (400 µg of dietary folate equivalent (DFE); 1 µg folic acid = 1 µg DFE; 1 µg food folate = 0.6 µg DFE) Overt deficiency is uncommonly observed in developed countries. The classic deficiency syndrome is scurvy, characterized by fatigue, depression, and widespread abnormalities in connective tissues (e.g., inflamed gingivae, petechiae, perifollicular hemorrhages, impaired wound healing, coiled hairs, hyperkeratosis, and bleeding into body cavities). In infants, defects in ossification and bone growth may occur. Tobacco smoking lowers plasma and leukocyte vitamin C levels (F, 75 mg; M, 90 mg; requirement for cigarette smokers increased by 35 mg/day)

Daily dosage >1000 µg may partially correct the anemia of B12 deficiency and may, therefore, mask (and perhaps exacerbate) the associated neuropathy. Large doses also are reported to lower seizure threshold in individuals prone to seizures. Parenteral administration is rarely reported to cause allergic phenomena from dispersion agents (1000 µg)

Serum folate levels reflect short-term folate balance, whereas RBC folate is a better reflection of tissue status. Serum homocysteine levels rise early in deficiency but are nonspecific because B12 or B6 deficiency, renal insufficiency, and older age may also cause elevations.

Quantities exceeding 500 mg/day (in adults) sometimes cause nausea and diarrhea. Acidification of the urine with vitamin C supplementation, and the potential for enhanced oxalate synthesis, have raised concerns regarding nephrolithiasis, but this has yet to be demonstrated. Supplementation with vitamin C may interfere with laboratory tests based on redox potential (e.g., fecal occult blood testing, serum cholesterol, and serum glucose). Withdrawal from chronic ingestion of high doses of vitamin C supplements should occur gradually over one month because accommodation does seem to occur, raising a concern for rebound scurvy (2000 mg) Toxicity has not been reported in humans, with doses as high as 60 mg/day in children (TUL not established)

Plasma ascorbic acid concentration reflects recent dietary intake, whereas leukocyte levels more closely reflect tissue stores. Plasma levels in women are approximately 20% higher than in men for any given dietary intake.

Diarrhea is reported to occur with doses exceeding 10 g/day (TUL not established)

Whole blood and urine concentrations of pantothenic acid are indicators of status; serum levels are not thought to be accurate.

C (ascorbic and dehydroascorbic acid)

Biotin

Pantothenic acid

Isolated deficiency is rare. Deficiency in humans has been produced experimentally (by dietary inadequacy), by prolonged TPN lacking the vitamin, and by ingestion of large quantities of raw egg white, which contains avidin, a protein that binds biotin with such high affinity that it renders it bio-unavailable. Alterations in mental status, myalgias, hyperesthesias, and anorexia occur. Later, seborrheic dermatitis and alopecia develop. Biotin deficiency is usually accompanied by lactic acidosis and organic aciduria (30 µg) Deficiency rare: reported only as a result of feeding semisynthetic diets or as an antagonist to the vitamin. Experimental isolated deficiency in humans produces fatigue, abdominal pain and vomiting, insomnia, and paresthesias of the extremities (5 mg)

Plasma and urine concentrations of biotin are diminished in the deficient state. Elevated urine concentrations of methyl citrate, 3-methylcrotonylglycine, and 3-hydroxyisovalerate are also observed in deficiency.

*RDA, recommended daily allowance; established for female (F) and male (M) adults by the U.S. Food and Nutrition Board, 1999-2001. In some cases, data are insufficient to establish an RDA, in which case the adequate intake (AI) established by the board is listed. † TUL, tolerable upper level; established for adults by the U.S. Food and Nutrition Board, 1999-2001. ALT, alanine aminotransferase; AST, aspartate aminotransferase; EEG, electroencephalogram; PLP, pyridoxyl 5-phosphate; RBC, red blood cell; TPN, total parenteral nutrition; TPP, thiamine pyrophosphate. Adapted from Goldman L, Ausiello D, Arend W, et al, editors. Cecil Textbook of Medicine. 22nd ed. Philadelphia: WB Saunders; 2004. With permission.

55

56

Section II  Nutrition in Gastroenterology Table 4-11  Salient Features of Trace Minerals Mineral

Deficiency (RDA)*

Toxicity (TUL)†

Assessment of Status

Chromium

Deficiency in humans is only described for patients on long-term TPN whose TPN contained inadequate chromium. Hyperglycemia, or impaired glucose intolerance, is uniformly observed. Elevated plasma free fatty acid concentrations, neuropathy, encephalopathy, and abnormalities in nitrogen metabolism are also reported. Whether supplemental chromium may improve glucose tolerance in mildly glucose intolerant, but otherwise healthy, individuals remains controversial (F, 25 µg; M, 35 µg) Dietary deficiency is rare; it has been observed in premature and low birth weight infants fed exclusively a cow’s milk diet and in individuals on long-term TPN without copper. The clinical manifestations include depigmentation of skin and hair, neurologic disturbances, leukopenia and hypochromic, microcytic anemia, and skeletal abnormalities. The anemia arises from impaired uptake of iron, and is therefore a secondary form of iron deficiency anemia. The deficiency syndrome, except the anemia and leukopenia, is also observed in Menkes disease, a rare inherited condition associated with impaired copper utptake (900 µg) Intake of <0.1 mg/day in infants and 0.5 mg/ day in children is associated with an increased incidence of dental caries. Optimal intake in adults is between 1.5 and 4.0 mg/day (F, 3 mg; M, 4.0 mg)

Toxicity after oral ingestion is uncommon and seems confined to gastric irritation. Airborne exposure may cause contact dermatitis, eczema, skin ulcers, and bronchogenic carcinoma (No TUL established)

Plasma or serum concentration of chromium is a crude indicator of chromium status; it appears to be meaningful when the value is markedly above or below the normal range.

Acute copper toxicity has been described after excessive oral intake and with absorption of copper salts applied to burned skin. Milder manifestations include nausea, vomiting, epigastric pain, and diarrhea; coma and hepatocellular injury may ensue in severe cases. Toxicity may be seen with doses as low as 70 µg/kg/day. Chronic toxicity is also described. Wilson disease is a rare inherited disease associated with abnormally low ceruloplasmin levels and accumulation of copper in the liver and brain, eventually leading to damage to these two organs (10 mg) Acute ingestion of >30 mg/kg body weight of fluoride is likely to cause death. Excessive chronic intake (0.1 mg/kg/day) leads to mottling of the teeth (dental fluorosis), calcification of tendons and ligaments, and exostoses and may increase the brittleness of bones (10 mg) Large doses (>2 mg/day in adults) may induce hypothyroidism by blocking thyroid hormone synthesis. Supplementation with >100 µg/day to an individual who was formerly deficient occasionally induces hyperthyroidism (1.1 mg)

Practical methods for detecting marginal deficiency are not available. Marked deficiency is reliably detected by diminished serum copper and ceruloplasmin concentrations, as well as low erythrocyte superoxide dismutase activity.

Copper

Fluoride

Iodine

In the absence of supplementation, populations relying primarily on food from soils with low iodine content have endemic iodine deficiency. Maternal iodine deficiency leads to fetal deficiency, which produces spontaneous abortions, stillbirths, hypothyroidism, cretinism, and dwarfism. Rapid brain development continues through the second year, and permanent cognitive deficits may be induced by iodine deficiency during that period. In adults, compensatory hypertrophy of the thyroid (goiter) occurs, along with varying degrees of hypothyroidism (150 µg)

Estimates of intake or clinical assessment are used because no reliable laboratory test exists.

Urinary excretion of iodine is an effective laboratory means of assessment. The thyroidstimulating hormone (TSH) level in the blood is an indirect, not entirely specific, means of assessment. Iodine status of a population can be estimated by the prevalence of goiter.

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient Table 4-11  Salient Features of Trace Minerals—cont’d Mineral

Deficiency (RDA)*

Toxicity (TUL)†

Assessment of Status

Iron

The most common micronutrient deficiency in the world. Women of childbearing age constitute the highest risk group because of menstrual blood losses, pregnancy, and lactation. Hookworm infection is the most common cause, worldwide. The classic deficiency syndrome is hypochromic microcytic anemia. Glossitis and koilonychia (spoon nails) are also observed. Easy fatigability often develops as an early symptom, before the appearance of anemia. In children, mild deficiency of insufficient severity to cause anemia is associated with behavioral disturbances and poor school performance (postmenopausal F, 8 mg; M, 8 mg; premenopausal F, 18 mg)

Negative iron balance initially leads to depletion of iron stores in the bone marrow: a bone marrow biopsy and the concentration of serum ferritin are accurate and early indicators of such depletion. As the severity of deficiency proceeds, serum iron (SI) decreases and total iron binding capacity (TIBC) increases; an iron saturation (= SI/TIBC) of <16% suggests iron deficiency. Microcytosis, hypochromia, and anemia ensue in latter stages of the deficient state. Elevated levels of serum ferritin or an iron saturation more than 60% raises the suspicion of iron overload, although systemic inflammation elevates the serum ferritin level regardless of iron status.

Manganese

Manganese deficiency has not been conclusively demonstrated in humans. It is said to cause hypocholesterolemia, weight loss, hair and nail changes, dermatitis, and impaired synthesis of vitamin K–dependent proteins (F, 1.8 mg; M, 2.3 mg) Cases of human deficiency extremely rare; caused by TPN lacking the element or by parenteral administration of sulfite. Reported to result in hyperoxypurinemia, hypouricemia, low urinary sulfate excretion, and CNS disturbances (45 µg) Deficiency is rare in North America but has been observed in individuals on long-term TPN lacking selenium. Such individuals have myalgias and/or cardiomyopathy. Populations in some regions of the world, most notably some parts of China, have marginal intake of selenium. It is in these regions of China that Keshan’s disease is endemic, a condition characterized by cardiomyopathy. Keshan’s disease can be prevented (but not treated) by selenium supplementation (55 µg) Deficiency of zinc has its most profound effect on rapidly proliferating tissues. Mild deficiency causes growth retardation in children. More severe deficiency is associated with growth arrest, teratogenicity, hypogonadism and infertility, dysgeusia, poor wound healing, diarrhea, dermatitis on the extremities and around orifices, glossitis, alopecia, corneal clouding, loss of dark adaptation, and behavioral changes. Impaired cellular immunity also is observed. Excessive loss of gastrointestinal secretions through, for example, chronic diarrhea or fistulas, may precipitate deficiency. Acrodermatitis enteropathica is a rare, recessively inherited disease in which intestinal absorption of zinc is impaired (F, 8 mg; M, 11 mg)

Iron overload typically occurs when habitual dietary intake is extremely high, intestinal absorption is excessive, repeated parenteral administration of iron occurs, or a combination of these factors exists. Excessive iron stores usually accumulate in the reticuloendothelial tissues and cause little damage (hemosiderosis). If overload continues, iron will eventually begin to accumulate in tissues such as the hepatic parenchyma, pancreas, heart and synovium, damaging these tissues (hemochromatosis). Hereditary hemochromatosis arises as a result of homozygosity of a common, recessive, trait. Excessive intestinal absorption of iron is observed in homozygotes (45 mg) Toxicity by oral ingestion is unknown in humans. Toxic inhalation causes hallucinations, other alterations in mentation, and extrapyramidal movement disorders (11 mg)

Molybdenum

Selenium

Zinc

Molybdenum has low toxicity; occupational exposures and high dietary intake are linked to hyperuricemia and gout in epidemiologic studies (2 mg) Toxicity is associated with nausea, diarrhea, alterations in mental status, peripheral neuropathy, loss of hair and nails; such symptoms were observed in adults who inadvertently consumed between 27 and 2400 mg (400 µg)

Acute zinc toxicity can usually be induced by ingestion of >200 mg of zinc in a single day (in adults). It is manifested by epigastric pain, nausea, vomiting, and diarrhea. Hyperpnea, diaphoresis, and weakness may follow inhalation of zinc fumes. Copper and zinc compete for intestinal absorption: chronic ingestion of >25 mg zinc/day may lead to copper deficiency. Chronic ingestion of >150 mg/day has been reported to cause gastric erosions, low high-density lipoprotein cholesterol levels, and impaired cellular immunity (40 mg)

Until the deficiency syndrome is better defined, an appropriate measure of status will be difficult to develop.

No effective clinically available assessment exists. Rare cases of deficiency are associated with hypouricemia, hypermethionemia, and low levels of urinary sulfate with elevated excretion of sulfite, xanthine and hypoxanthine. Erythrocyte glutathione peroxidase activity and plasma, or whole blood, selenium concentrations are the most commonly used methods of assessment. They are moderately accurate indicators of status.

There are no accurate indicators of zinc status available for routine clinical use. Plasma, erythrocyte, and hair zinc concentrations are frequently misleading. Acute illness, in particular, is known to diminish plasma zinc levels, in part by inducing a shift of zinc out of the plasma compartment and into the liver. Functional tests that determine dark adaptation, taste acuity, and rate of wound healing lack specificity.

*Recommended Daily Allowance (RDA) established for female (F) and male (M) adults by the U.S. Food and Nutrition Board, 1999-2001. In some cases, insufficient data exist to establish an RDA, in which case the adequate intake (AI) established by the Board is listed. † Tolerable upper level (TUL) established for adults by the U.S. Food and Nutrition Board, 1999-2001. TPN, total parenteral nutrition. Adapted from Goldman L, Ausiello D, Arend W, et al, editors. Cecil Textbook of Medicine. 22nd ed. Philadelphia: WB Saunders; 2004. With permission.

57

58

Section II  Nutrition in Gastroenterology Table 4-12  Guidelines for Daily Delivery of Parenteral Micronutrients in Adults and Children Micronutrient Vitamin A D E K C Folate Niacin Riboflavin Thiamine B6 B12 Pantothenic acid Biotin Trace Elements Copper Chromium Manganese Zinc Molybdenum Iodine* Selenium Iron

Adults

Children

1000 µg (= 3300 IU) 5 µg (= 200 IU) 10 mg (= 10 IU) 1 mg 100 mg 400 µg 40 mg 3.6 mg 3 mg 4 mg 5 µg 15 mg 60 µg

700 µg 10 µg 7 mg 200 µg 80 mg 140 µg 17 mg 1.4 mg 1.2 mg 1.0 mg 1.0 µg 5 mg 20 µg

0.5-1.5 mg 10-15 µg 0.1 mg 2.5-4.0 mg 15 µg — 100 µg 1-2 mg

20 µg/kg/day 0.2 µg/kg/day 1.0 µg/kg/day 50 µg/kg/day 0.25 µg/kg/day — 2.0 µg/kg/day 1 mg/day

*

Naturally occurring contamination of parenteral nutrition appears to provide sufficient quantities of iodine. Adult vitamin guidelines adapted from American Society of Parenteral and Enteral Nutrition (ASPEN). Board of Directors and the Clinical Guidelines Task Force. Guidelines for the use of parenteral and enteral nutrition in adult and pediatric patients. JPEN J Parenter Enteral Nutr 2002;26:144. Children’s values adapted from Greene HL, Hambidge KM, Schanler R, Tsang RC. Guidelines for the use of vitamins, trace elements, calcium, magnesium, and phosphorus in infants and children receiving total parenteral nutrition: Report of the Subcommittee on Pediatric Parenteral Nutrient Requirements from the Committee on Clinical Practice Issues of the American Society for Clinical Nutrition. Am J Clin Nutr 1988; 48:1324; Am J Clin Nutr 1989; 49:1332; and Am J Clin Nutr 1989; 50:560.

selenium, iodine, fluorine, manganese, molybdenum, and cobalt (see Table 4-11). The biochemical functions of trace elements have not been as well characterized as those of the vitamins, but most of their functions appear to be as components of prosthetic groups or as cofactors for a number of enzymes. Aside from iron, the trace mineral depletion that clinicians are most likely to encounter is zinc deficiency. Zinc depletion is a particularly germane issue to the gastroenterologist because the gastrointestinal tract is a major site for zinc excretion. Chronically excessive losses of gastrointe­ stinal secretions, such as chronic diarrhea in inflammatory bowel disease, is a known precipitant for zinc deficiency, and in this setting zinc requirements often increase several-fold.25 Nevertheless, a biochemical diagnosis of zinc deficiency is problematic, as is true for many of the other essential trace minerals. Accurate laboratory assessment of zinc status is complicated by the very low concentrations of zinc in bodily fluids and tissues, a lack of correlation between serum and red blood cell levels of zinc with levels in the target tissues, and the reality that functional tests have yet to be devised. Furthermore, it is well recognized that in acute illness a shift in zinc occurs from the serum compartment into the liver, further obscuring the diagnostic value of serum zinc levels.26,27 Thus, it is often best simply to proceed with empirical zinc supplementation in patients whose clinical scenario puts them at high risk of zinc deficiency. Some reports have indicated that TPN solutions that deliver several-fold more manganese than what is recommended in Table 4-12 may lead to deposition of the mineral

in the basal ganglia, with extrapyramidal symptoms, seizures, or both.28 Because the content of manganese varies widely in the different trace element mixtures available for TPN compounding, the health professional needs to be mindful of this issue as protocols for TPN admixtures are developed.

PHYSIOLOGIC AND PATHOPHYSIOLOGIC FACTORS AFFECTING MICRONUTRIENT REQUIREMENTS Age

An evolution of physiology continues throughout the life cycle, with an impact on the requirements of certain micronutrients with aging; specific RDAs for older adults now have been developed. The mean vitamin B12 status of most populations, for example, declines significantly with older age, in large part because of the high prevalence of atrophic gastritis and its resultant impairment of protein-bound vitamin B12 absorption.29 Approximately 10% to 15% of the older ambulatory population is thought to have significant vitamin B12 depletion because of this phenomenon, and neuropathic degeneration may occur in older individuals whose plasma vitamin B12 levels are in the low-normal range (150 to 300  pg/mL), even in the absence of hematologic manifestations. For this reason, the use of sensitive indicators of cellular depletion of vitamin B12, such as serum methylmalonic acid levels in conjunction with serum levels of vitamin B12, now are recommended for diagnosis.30 Some experts also suggest that older adults should consume a portion of their vitamin B12 requirement in the crystalline form (i.e., as a supplement) rather than relying only on the naturally occurring protein-bound forms found in food.31 Older adults also require greater quantities of vitamins B6 and D and calcium to maintain health compared with younger adults, and these requirements are reflected in the new RDAs (see Tables 4-10 and 4-11).

Malabsorption and Maldigestion

Both fat- and water-soluble micronutrients are absorbed predominantly in the proximal small intestine, with the only exception being vitamin B12. Diffuse mucosal diseases, which affect the proximal portion of the gastrointestinal tract, are therefore likely to result in multiple deficiencies. Even in the absence of proximal small intestinal disease, however, extensive ileal disease, small bowel bacterial overgrowth, and chronic cholestasis may interfere with the maintenance of adequate intraluminal conjugated bile acid concentrations and thereby may impair absorption of fatsoluble vitamins. Conditions that produce fat malabsorption frequently are associated with selective deficiencies of the fat-soluble vitamins. The early stages of many vitamin deficiencies are not apparent clinically and therefore may go undetected until progression of the deficiency has resulted in significant morbidity. This can be disastrous in conditions such as spinocerebellar degeneration from vitamin E deficiency, which often is irreversible.32 Fat-soluble vitamin deficiencies are well-recognized complications of cystic fibrosis and congenital biliary atresia in which fat malabsorption often is overt, but monitoring also is necessary in conditions associated with more subtle fat malabsorption, such as the latter stages of chronic cholestatic liver disease.33,34 Restitution of vitamin deficiencies sometimes can be difficult when severe fat malabsorption is present and initial correction may require parenteral administration. In severe fat malabsorption, chemically modified forms of vitamins D and E that largely bypass the need for the lipophilic phase

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient of intestinal absorption are commercially available for oral use and can be helpful. The polyethylene glycol succinate form of vitamin E (Nutr-E-Sol) is very effective in patients with severe fat malabsorption who cannot absorb conventional alpha-tocopherol.35 Similarly, hydroxylated forms of vitamin D (1-hydroxyvitamin D [Hectorol] and 1,25dihydroxyvitamin D [Rocaltrol]) can be used in patients resistant to the more conventional forms of vitamin D. Intermittent monitoring of serum calcium levels is indicated in the first few weeks of therapy when hydroxylated forms of vitamin D are administered because they are considerably more potent than vitamin D2 or D3 and risk of vitamin D toxicity exists. In contrast, water-miscible preparations of fat-soluble vitamins, in which a conventional form of vitamin A or E is dissolved in polysorbate 80 (e.g., AquasolE, Aquasol-A), have not been proven to improve overall absorption. At the time of this writing, Aquasol-A is no longer available as an oral supplement. Maldigestion usually results from chronic pancreatic insufficiency, which if untreated frequently causes fat malabsorption and deficiencies of fat-soluble vitamins. Vitamin B12 malabsorption also can be demonstrated in this setting, but clinical vitamin B12 deficiency is rare unless other conditions known to diminish its absorption also are present, such as atrophic gastritis29 or chronic administration of proton pump inhibitors (PPIs).36 Whether the long-term administration of PPIs alone warrants occasional checks of vitamin B12 status is a matter of debate. Regardless, malabsorption of vitamin B12 from atrophic gastritis or with PPIs is confined to dietary sources of vitamin B12. Small supplemental doses of crystalline vitamin B12 are absorbed readily in both cases. Histamine H2 receptor antagonists also inhibit protein-bound vitamin B12 absorption, although the effect generally is believed to be less potent than with the PPIs.37 Many medications may adversely affect micronutrient status. The manner in which drug-nutrient interaction occurs varies; some of the more common mechanisms are described in Table 4-13. A comprehensive discussion of drug-nutrient interactions is beyond the scope of this

Table 4-13  Interactions of Drugs on Micronutrient Status Drug(s)

Nutrient

Mechanism(S)

Dextroamphetamine, fenfluramine, levodopa Cholestyramine

Potentially all micronutrients

Induces anorexia

Vitamin D, folate

PPIs

Vitamin B12

Sulfasalazine

Folate

Isoniazid

Pyridoxine

NSAIDs

Iron

Penicillamine

Zinc

Adsorbs nutrient, decreases absorption Modest bacterial overgrowth, decreases gastric acid, impairs absorption Impairs absorption and inhibits folatedependent enzymes Impairs uptake of vitamin B6 Gastrointestinal blood loss Increases renal excretion

NSAIDs, nonsteroidal anti-inflammatory drugs; PPIs, proton pump inhibitors. From Goldman L, Ausiello D, Arend W, et al, editors. Cecil Textbook of Medicine. 22nd ed. Philadelphia: WB Saunders; 2004. With permission.

chapter and the reader is referred to other references for a detailed discourse on this topic.38

STARVATION During periods of energy and/or protein deficit, an array of compensatory mechanisms serves to lessen the pathophy­s­ iologic impact of these deficits. These responses decrease the metabolic rate, maintain glucose homeostasis, conserve body nitrogen, and increase the uptake of adipose tissue TGs to meet energy needs. To appreciate how acute illness disrupts this compensatory scheme, it is first necessary to understand how the body adapts to starvation in the absence of underlying disease. During the first 24 hours of fasting, the most readily available energy substrates (i.e., circulating glucose, FAs, and TGs, and liver and muscle glycogen) are used as fuel sources. The sum of energy provided by these stores in a 70-kg man, however, is only about 5000 kJ (1200 kcal) and therefore is less than a full day’s requirements. Hepatic glucose production and oxidation decrease, whereas whole-body lipolysis increases, and the latter provides additional FAs and ketone bodies.39 Oxidation of the FAs released from adipose tissue TGs accounts for approximately 65% of energy consumed during the first 24 hours of fasting. During the first several days of starvation, obligate glucose-requiring tissues such as the brain and blood cells, which collectively account for about 20% of total energy consumption, can use only glycolytic pathways to obtain energy. Because FAs cannot be converted to carbohydrate, these glycolytic tissues must use glucose or substrates that can be converted to glucose. Glucogenic amino acids derived from skeletal muscle (chiefly alanine and glutamine) are a major source of substrate for this purpose. Approximately 15% of the REE is provided by oxidation of protein.40 The relative contribution of gluconeogenesis to hepatic glucose production increases as the rate of hepatic glycogenolysis declines because the latter process becomes redundant; after 24 hours of fasting, only 15% of liver glycogen stores remain. During short-term starvation (1 to 14 days), several adaptive responses appear that lessen the loss of lean mass. A decline in levels of plasma insulin, an increase in plasma epinephrine levels, and an increase in lipolytic sensitivity to catecholamines stimulate adipose tissue lipolysis.41,42 The increase in FA delivery to the liver, in conjunction with an increase in the ratio of plasma glucagon-to-insulin concentrations, enhances the production of ketone bodies by the liver. A maximal rate of ketogenesis is reached by three days of starvation, and plasma ketone body concentration is increased 75-fold by seven days. In contrast to FAs, ketone bodies can cross the blood-brain barrier and provide most of the brain’s energy needs by seven days of starvation.43 The use of ketone bodies by the brain greatly diminishes glucose requirements and thus spares the need for muscle protein degradation to provide glucose precursors. If early protein breakdown rates were to continue throughout starvation, a potentially lethal amount of muscle protein would be catabolized in less than three weeks. Similarly, the heart, kidney, and skeletal muscle change their primary fuel substrate to FAs and ketone bodies. Other tissues such as bone marrow, renal medulla, and peripheral nerves switch from full oxidation of glucose to anaerobic glycolysis, resulting in increased production of pyruvate and lactate. The latter two compounds can be converted back to glucose in the liver using energy derived from fat oxidation via the Cori

59

60

Section II  Nutrition in Gastroenterology cycle, and the resulting glucose is available for systemic consumption. This enables energy stored as fat to be used for glucose synthesis. Whole-body glucose production decreases by more than 50% during the first few days of fasting because of a marked reduction in hepatic glucose output. As fasting continues, the conversion of glutamine to glucose in the kidney represents almost 50% of total glucose production. Energy is conserved by a decrease in physical activity because of fatigue and a reduction in REE of approximately 10%, resulting from increased conversion of active thyroid hormone to its inactive form and suppressed sympathetic nervous system activity. During long-term starvation (14 to 60 days), maximal adaptation is reflected by a plateau in lipid, carbohydrate, and protein metabolism. The body relies almost entirely on adipose tissue for its fuel, providing more than 90% of daily energy requirements.44 Muscle protein breakdown decreases to less than 30  g/day, causing a marked decrease in urea nitrogen production and excretion. The decrease in osmotic load diminishes urine volume to 200  mL/day, thereby reducing fluid requirements. Total glucose production decreases to approximately 75 g/day, providing fuel for glycolytic tissues (40  g/day) and the brain (35  g/day) while maintaining a constant plasma glucose concentration. Energy expenditure decreases by 20% to 25% at 30 days of fasting and remains relatively constant thereafter, despite continued starvation. The metabolic response to short- and long-term starvation differs somewhat between lean and obese persons. Obesity is associated with a blunted increase in lipolysis and decrease in glucose production compared with that in lean persons.45,46 In addition, protein breakdown and nitrogen losses are less in obese persons, thereby helping conserve muscle protein.47 The events that mark the terminal phase of starvation have been studied chiefly in laboratory animals. Body fat mass, muscle protein, and the sizes of most organs are markedly decreased. The weight and protein content of the brain, however, remain relatively stable. During this final phase of starvation, body fat stores reach a critical level, energy derived from body fat decreases, and muscle protein catabolism is accelerated. Death commonly occurs when there is a 30% to 50% loss of skeletal muscle protein.48 In humans, it has been proposed that there are certain thresholds beyond which lethality is inevitable—depletion of total body protein between 30% and 50% and of fat stores between 70% and 95%, or reduction of BMI below 13 kg/m2 for men and 11 kg/m2 for women.49,50

MALNUTRITION In the broadest sense, malnutrition implies a sustained imbalance between nutrient availability and nutrient requirements. This imbalance results in a pathophysiologic state in which intermediary metabolism, organ function, and body composition are variously altered. Sustained is an important element of this definition because homeostatic mechanisms and nutrient reserves usually are adequate to compensate for any short-term imbalance. Customarily, the term malnutrition is used to describe a state of inadequacy in protein, calories, or both, and is more precisely called protein-energy malnutrition (or proteincalorie malnutrition). Occasionally, it is used to describe a state of excessive availability, such as a sustained excess of calories (e.g., obesity) or a vitamin (e.g., vitamin toxicity).

PROTEIN-ENERGY MALNUTRITION

There are different pathways whereby protein-energy malnutrition (PEM) may evolve. Primary PEM is caused by inadequate intake of protein and/or calories or, less commonly, when the protein ingested is of such poor quality that one or more essential amino acids becomes a limiting factor in the maintenance of normal metabolism. Secondary PEM is caused by illness or injury. Acute illnesses and injuries increase bodily requirements for protein and energy substrate and they impair the digestion, absorption, and uptake of these nutrients in various ways. Consequently, secondary PEM usually arises from multiple factors. Illness and injury also commonly induce anorexia (see later for mechanisms), and so primary and secondary factors often act in concert to create PEM in the setting of illness. Illness or injury may directly interfere with nutrient assimilation; for example, extensive ileal disease or resection may directly produce fat malabsorption and a caloric deficit. The most common causes of secondary PEM, however, are the remarkable increases in protein catabolism and energy expenditure that occur as a result of a systemic inflammatory response. REE may increase as much as 80% above basal levels in a manner roughly proportional to the magnitude of the inflammatory response, which in turn is roughly proportional to the severity and acuity of the illness. Thus, for example, REE in patients with extensive second- and third-degree burns (the prototype for maximal physiologic stress) may approach twice normal; with sepsis, REE is about 1.5 times normal; and with a localized infection or fracture of a long bone, REE is 25% above normal.5 Such stress factors can be used to construct a formula for predicting the caloric needs of ill individuals (see Table 4-5). Protein catabolism during illness or injury also increases in proportion to the severity and acuity of the insult, and therefore parallels the increase in energy consumption. The magnitude of increase in protein catabolism, however, is proportionately greater than that observed with energy consumption, such that urinary urea N losses, which reflect the degree of protein catabolism in acute illness, are about 2.5 times the basal level with maximal stress.5 This increase in catabolism results in a net loss of protein because the rate of synthesis usually does not rise in concert with the rise in catabolism.51 No known storage form of protein exists in the body and, therefore, any net loss of protein represents a loss of functionally active tissue. A healthy adult typically loses about 12  g  N in the urine/day, and excretion may increase to as much as 30  g/day during critical illness. Because 1  g of urinary N represents the catabolism of approximately 30  g of lean mass, it follows that severe illness may produce a daily loss of up to ~0.5  kg of lean mass as a result of excess protein catabolism. Most of this loss comes from the skeletal muscle, where the efflux of amino acids increases two- to sixfold in critically ill patients.52 Mobilization of amino acids from skeletal muscle appears to be an adaptive response. Once liberated, these amino acids, in part, are deaminated and used for gluconeogenesis; they also are taken up by the liver and other visceral organs. The proteolysis of muscle under stress thus enables the body to shift amino acids from the skeletal muscle (the somatic protein compartment) to the visceral organs (the visceral protein compartment), the functions of which are more critical for immediate survival during illness. Nevertheless, with sustained stress, the limitations of this adaptive response become evident, and even the visceral protein compartment sustains a contraction in mass.44

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient Primary versus Secondary Protein-Energy Malnutrition: A Body Compartment Perspective

The type of tissue lost as malnutrition evolves is critical in determining the pathologic ramifications of weight loss. Over 95% of energy expenditure resides in the lean body mass, which therefore contains the bulk of metabolism that sustains homeostasis. It is the maintenance of this body compartment that is most critical for health. Lean body mass can be subdivided further into somatic and visceral protein compartments, blood and bone cells, and extracellular lean mass, such as plasma and bone matrix (Fig. 4-1). In total or semistarvation in otherwise healthy individuals, adipose tissue predominates as a primary energy source; thus, fat mass contracts to a much greater degree proportional to the loss of lean mass.44 Alterations in metabolism from injury or illness, however, produce a proportionately greater loss of muscle mass so that it matches or exceeds the loss in fat mass.53,54 Although the lean mass that is lost in illness preferentially is from the somatic protein compartment, with sustained stress there also will be a significant contraction of the visceral protein compartment (Table 4-14). The metabolic forces associated

Table 4-14  Body Compartment Losses in Simple Starvation versus Metabolic Stress

Parameter Starvation Metabolic stress

Skeletal Muscle Wasting

Visceral Wasting

Loss of Fat Mass

+ +++

+/−* ++/−*

+++ +++

*

Relatively spared early in the process; can become pronounced with extended starvation or metabolic stress.

Blood cells, bone cells, etc. 7%

Extracellular lean mass (plasma, bone mineral, etc.) 36%

Visceral mass 7%

with acute illness and injury are potent, and restoration of muscle mass is unlikely with nutritional support unless the underlying inflammatory condition is corrected. There is increasing interest in attenuating or reversing net catabolism with the use of exogenous anabolic agents in conjunction with nutrition, although to date it remains unclear whether the clinical benefits of using exogenous growth hormone and other anabolic agents in acute illness outweigh their potential side effects.55,56 Another important ramification of the potency of the catabolic state associated with acute illness is that most of the weight that is gained with the provision of nutritional support is the result of increases in fat mass and body water; only minor increases in lean mass are observed until the inflammatory focus is resolved.57 Cytokines are the most important mediators of the alterations in energy and protein metabolism that accompany illness and injury. In a wide spectrum of systemic illnesses, increased secretion of interleukin-1 (IL-1), tumor necrosis factor-α (TNF-α), IL-6, and interferon-γ (IFN-γ) has been observed to be associated with increased energy expenditure and protein catabolism, as well as the shift of amino acids into the visceral compartments.58-60 Such observations concur with in vitro studies in human cells and animal models that have shown remarkably potent effects of these cytokines in this regard (Table 4-15). In the wasting syndrome associated with cancer, proteolysis-inducing factor and lipid-mobilizing factor are two humoral mediators that appear to be unique to cancer cachexia, contrib­ uting to protein catabolism and loss of adipose tissue, respectively.61

Protein-Energy Malnutrition in Children

Undernutrition in children differs from that in adults because it affects growth and development. Much of our understanding of undernutrition in children comes from observations made in underdeveloped nations in which poverty, inadequate food supply, and unsanitary conditions lead to a high prevalence of PEM. The Waterlow classification of malnutrition takes into account a child’s weight for height (wasting) and height for age (stunting) (Table 4-16).62 The characteristics of the three major clinical PEM syndromes in children—kwashiorkor, marasmus, and nutritional dwarfism—are outlined in Table 4-17.63 Although these three syndromes are classified separately, overlap syndromes often coexist in the same patient. Marasmus Weight loss and marked depletion of subcutaneous fat and muscle mass are characteristic features of children with marasmus. Ribs, joints, and facial bones are prominent and the skin is thin, loose, and lies in folds. In contrast, the visceral protein compartment is relatively spared, a fact that often is reflected by a normal serum albumin level, which in turn sustains normal oncotic pressure in the vascular compartment, thus minimizing edema.

Muscle mass 22% Fat mass 28% Figure 4-1.  Body composition analysis by weight in a healthy adult. The speckled segments and gray segment collectively represent lean body mass. The speckled segments alone represent body cell mass. (From Mason JB: Gastrointestinal cancer; nutritional support. In Kelsen D, Daly J, Kern S, et al, editors. Principles and Practice of Gastrointestinal Oncology. Philadelphia, Pa: Lippincott Williams & Wilkins; 2002. p 127. With permission.)

Kwashiorkor The word kwashiorkor, from the Ga language of West Africa, means “disease of the displaced child” because it commonly was seen after weaning. The presence of peripheral edema distinguishes children with kwashiorkor from those with marasmus and nutritional dwarfism. Children with kwashiorkor also have characteristic skin and hair changes (see sections on hair and skin changes, later). The abdomen is protuberant because of weakened abdominal muscles, intestinal distention, and hepatomegaly, but ascites is rare.

61

62

Section II  Nutrition in Gastroenterology Table 4-15  Major Cytokines That Mediate Hypercatabolism and Hypermetabolism Associated with Metabolic Stress Cytokine

Cell Source

Metabolic Effects

Tumor necrosis factor-α

Monocytes/macrophages, lymphocytes, Kupffer cells, glial cells, endothelial cells, natural killer cells, mast cells

IL-1

Monocytes/macrophages, neutrophils, lymphocytes, keratinocytes, Kupffer cells

IL-6

Monocytes/macrophages, keratinocytes, endothelial cells, fibroblasts, T cells, epithelial cells Lymphocytes, pulmonary macrophages

Decreased FFA synthesis Increased lipolysis Increased peripheral amino acid loss Increased hepatic amino acid uptake Fever Increased ACTH levels Increased hepatic acute-phase protein synthesis Fever Increased acute-phase protein synthesis Fever

IFN-γ

Increased monocyte respiratory burst

ACTH, adrenocorticotropic hormone; FFA, free fatty acid; IL, interleukin; IFN, interferon. Adapted from Smith M, Lowry S. The hypercatabolic state. In: Shils M, Olson J, Shike M, Ross AC, editors. Modern Nutrition in Health and Disease. Baltimore: Williams & Wilkins; 1999. p 1555.

Table 4-16  Waterlow Classification of Protein-Energy Malnutrition in Children Parameter Weight for Height (Wasting) Percentage of median NCHS standard Standard deviation from the NCHS median Height for Age (Stunting) Percentage of median NCHS standard Standard deviation from NCHS median

Normal

Mild

Moderate

Severe

90 to 100

80 to 89

70 to 79

<70

+Z to −Z

−1.1 Z to −2 Z

−2.1 Z to −3 Z

<−3 Z

95 to 105

90 to 94

85 to 89

<85

+Z to −Z

−1.1 Z to −2 Z

−2.1 Z to −3 Z

<−3 Z

NCHS, National Center for Health Statistics; Z, Standard Score.

Table 4-17  Features of Protein-Energy Malnutrition Syndromes in Children Syndrome Parameter

Kwashiorkor

Marasmus

Nutritional Dwarfism

Weight for age (% expected) Weight for height Edema Mood

60-80 Normal or decreased Present Irritable when picked up, apathetic when alone Poor

<60 Markedly decreased Absent Alert

<60 Normal Absent Alert

Good

Good

Appetite

The presence of ascites, therefore, should prompt the clinician to search for liver disease or peritonitis. Children with kwashiorkor typically are lethargic and apathetic, but become very irritable when held. Kwashiorkor is not caused by a relative deficiency in protein intake; rather, it most often occurs when there is physiologic stress, such as infection, in an already malnourished child. Because infection or other acute stress usually is present in kwashiorkor, the metabolic aberrations associated with secondary PEM are in play and contractions of the visceral protein compartment are evident. A decrease in serum proteins such as albumin is common, distinguishing it from pure marasmus. Kwashi-

orkor is characterized by leaky cell membranes, which permit the movement of potassium and other intracellular ions into the extracellular space, causing water movement and edema. Nutritional Dwarfism The child with failure to thrive may be of normal weight for height but have short stature and delayed sexual development. Providing appropriate feeding can stimulate catch-up growth and sexual maturation. The diagnosis of PEM is different in adults than in children, because adults do not grow in height. Therefore,

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient undernutrition in adults causes wasting rather than stunting. In addition, although pure forms of kwashiorkor and marasmus can occur in adults, contemporary studies of adult PEM in industrialized societies typically focus on hospitalized patients with secondary PEM, coexisting illness or injury, and overlapping features of kwashiorkor and marasmus.

PHYSIOLOGIC IMPAIRMENTS CAUSED BY PROTEIN-ENERGY MALNUTRITION

PEM adversely affects almost every organ system, although the brain usually is spared. Almost all the adverse effects are reversible with nutritional restitution. The following discussion is not exhaustive; rather, it emphasizes impairments that commonly translate into overt morbidity or that are important in the diagnosis of PEM. The effects described reflect what occurs in primary PEM; the superimposition of acute illness and secondary PEM often imposes more complexity.

System Effects

Gastrointestinal Tract Although PEM alone produces adverse effects on gastrointestinal structure and function, diminished stimulation of the gastrointestinal tract by a lack of ingested nutrients has an independent effect. Thus, sustained absence of nutrients passing through the intestine of healthy, nutri­ tionally replete, parenterally fed individuals alone results in functional atrophy of the small intestinal mucosa, as evidenced by a loss of brush border enzymes and diminished integrity of the epithelial barrier. Villus atrophy also may be observed with lack of intestinal stimulation but, in the absence of PEM, the degree of structural atrophy is minor.64 The structural and functional deterioration of the intestinal tract, pancreas, and liver from PEM is described best in children. Marked blunting of the intestinal villi is seen and is usually associated with loss of some or all of the brush border hydrolases. Gastric and pancreatic secretions are reduced in volume and contain decreased concentrations of acid and digestive enzymes, respectively. The volume of bile and the concentrations of conjugated bile acids in bile are reduced. Increased numbers of facultative and anaerobic bacteria are found in the upper small intestine, probably explaining the increased proportion of free bile acids in the intestinal lumen. Malabsorption of carbohydrates, fats, and fat- and water-soluble vitamins may occur, and the degree of steatorrhea is proportional to the severity of the PEM, creating a vicious cycle of further malnutrition. The pro­ tuberance of the abdomen that is sometimes evident in advanced malnutrition is thought to arise in part from intestinal hypomotility and gas distention. Cardiovascular System Moderate to severe PEM produces quantitative and qualitative declines in the cardiac muscle. Myocardial mass is diminished, although proportionately less than the loss in body weight. Myofibrillar atrophy, edema and, less commonly, patchy necrosis and infiltration with chronic inflammatory cells are seen in the myocardium. These structural changes are associated with impaired myocardial performance. A decrease in stroke volume, cardiac output, and maximal work capacity may be observed, and are most evident under conditions of increased demand. Such functional impairments are sometimes accompanied by bradycardia and, in conjunction with the factors noted, can lead to low blood pressure.

Immune System The immune system is the most vulnerable to PEM, which explains why several functions of immunity are used diagnostically as indicators of malnutrition, such as total lymphocyte count and delayed skin hypersensitivity. The functional integrity of T lymphocytes, polymorphonuclear leukocytes, and complement is uniformly blunted, whereas impaired B lymphocyte production of antibodies is variably affected. A moderately to severely malnourished patient is an immunocompromised individual. Malnutrition leads to increased susceptibility to infection, which in turn promotes the development of PEM; thus, a vicious cycle is created. Respiratory System The diaphragm and other respiratory muscles undergo structural and functional atrophy, diminishing inspiratory and expiratory pressures and vital capacity. These changes in muscular performance, in conjunction with blunted ventilatory drive, impair the ability to sustain ventilation in the severely malnourished individual. In patients with a tracheostomy, adherence of bacteria to the tracheal epithelium correlates with the severity of PEM, exacerbating other compromises in the immune system described earlier. Endocrine System Although alterations in hormones are common in PEM, many of the changes can be perceived as serving adaptive functions. The inadequate intake of food leads to a decrease in the availability of circulating glucose and amino acids, low circulating levels of insulin, and increased levels of growth hormone. These alterations, in conjunction with the decreased levels of somatomedins and increased levels of cortisol in PEM, promote skeletal muscle catabolism and, at the same time, enhance incorporation of the liberated amino acids into visceral organs. Urea synthesis is inhibited, decreasing nitrogen loss and enhancing the reutilization of amino acids. The enhancement of lipolysis and gluconeogenesis provides a substrate for energy needs. Serum levels of triiodothyronine (T3) and thyroxine (T4) commonly are decreased in conjunction with increased concentrations of reverse T3, resembling the pattern observed in the euthyroid sick syndrome. The decreased concentration of T3 may play a role in decreasing REE and the protein catabolic rate that is observed in primary PEM. Primary gonadal dysfunction is common in adults with moderate to severe PEM, and results in impaired reproductive potential. Decreased circulating levels of testosterone in men and estrogen in women is evident, and amenorrhea is common. Delayed puberty, or loss of menstrual periods, most often occurs when lean body mass drops below a critical threshold. These changes also can be considered physiologic adaptations, because ensuring immediate survival is more critical than the need for sexual maturation in the child or reproduction in the adult.

Other Effects

Wound Healing Well-nourished individuals lay down more collagen at the site of a surgical wound than those with even mild malnutrition. Furthermore, nutritional repletion of the malnourished patient before surgery leads to better wound healing than if nutritional needs are addressed only postoperatively. Skin Undernutrition often causes dry, thin, and wrinkled skin, with atrophy of the basal layers of the epidermis and

63

64

Section II  Nutrition in Gastroenterology hyperkeratosis. Severe malnutrition may cause considerable depletion of skin protein and collagen. Patients with kwashiorkor experience sequential skin changes in different areas. Hyperpigmentation occurs first, followed by cracking and stripping of superficial layers, leaving behind hypopigmented, thin, and atrophic epidermis that is friable and easily macerated. Hair Scalp hair becomes thin, sparse, and is easily pulled out. In contrast, the eyelashes become long and luxuriant and there may be excessive lanugo in children. The hair of children with kwashiorkor develops hypopigmentation, with reddish-brown, gray, or blond discoloration. Adults may lose axillary and pubic hair. Kidneys Renal mass and function often are well preserved during undernutrition. When malnutrition is severe, however, there are decreases in kidney weight, glomerular filtration rate, ability to excrete acid and sodium, and ability to concentrate urine. Mild proteinuria also may occur. Bone Marrow Severe undernutrition suppresses bone marrow red blood cell and white blood cell production, leading to anemia, leukopenia, and lymphocytopenia.

NUTRITIONAL ASSESSMENT TECHNIQUES Overview

The purpose of nutritional assessment is to identify PEM and other nutritional deficits, even when they are not readily discernible. PEM can be subtle, but most cases are detected when a systematic nutritional assessment is performed. An example of subtle but clinically significant PEM is found in Child-Turcotte-Pugh class A alcoholic cirrhotics. These individuals usually appear well nourished. Indeed, one criterion to determine class A status is a normal serum albumin level. Studies of whole-body nitrogen by in vivo neutron activation analysis, however, have demonstrated that more than half of these class A individuals have less than 80% of expected total body protein,65 the threshold level below which patients have increased morbidity associated with malnutrition.54 In otherwise healthy people and in those who are chronically ill, PEM usually is defined by comparing an anthropometric measurement, such as weight for height, to established normative standards (see Table 4-6). In contrast, there is no gold standard to define and measure PEM in the acutely ill patient because most parameters used to assess PEM in otherwise healthy persons are altered by illness; weight and the concentration of serum proteins are prime examples. Despite the inaccuracies inherent in assessing PEM in acutely ill individuals, however, the usefulness of nutritional assessment in this setting has been demonstrated repeatedly. Acutely ill patients who are malnourished sustain higher rates of malnutrition-related morbidity. The presence of PEM, therefore, has a predictive value. Even more importantly, identification of malnourished patients and appropriate nutritional intervention is likely to improve clinical outcome.66-73 Meta-analyses have underscored the importance of performing objective nutritional assessments to categorize inpatients, because individuals who are wellnourished or mildly malnourished seem to realize little benefit from intensive nutritional support.71-73

Specific Tools for Assessing Protein-Energy Malnutrition

A comprehensive nutritional assessment requires a history, physical examination, evaluation of anthropometrics or functional measures of nutritional status, and a panel of laboratory blood tests. Some of the more commonly used assessment tools are weight, height, and other anthropometric measures, including skinfold thickness and midarm measurements, functional measures (e.g., hand grip strength or skin testing for evidence of anergy), serum protein concentrations, (e.g., albumin or prealbumin), complete blood count, including absolute lymphocyte count, and 24-hour urinary creatinine and urea nitrogen levels. Less readily available measures of body composition such as bioelectrical impedance and total body potassium can be helpful in the appropriate setting. Some of these measures lack a high degree of specificity but continue to be useful in clinical care because of their prognostic significance. Because no single parameter is sufficiently sensitive or specific to assess PEM, these tools are most effective when used in combination. If the clinician is undertaking a nutritional assessment solely to determine whether a patient falls into a category of moderate to severe PEM, far less than comprehensive assessment usually will suffice; such simple means of categorization are provided in the following sections. History Weight Loss.  Unintentional weight loss associated with illness is the single most practical predictor of a clinically significant degree of PEM. It is useful to quantify such loss by determining whether the patient has sustained a mild (<5%), moderate (5% to 10%), or severe (>10%) degree of loss over the preceding six months. Because acute illness incites a disproportionately large loss of lean mass, it is not surprising that a more than 10% unintentional loss in body weight usually translates into a 15% to 20% decrease in total body protein.54 This degree of unintentional loss is an important threshold because it is associated with impaired physiology, a poor clinical outcome, and extended hospitalization74-76; it also defines those individuals who will likely benefit from intensive nutritional support. The clinician should nevertheless be mindful that determination of the magnitude of weight loss by history has limited accuracy; one study has found that one third of patients with true weight loss go undetected by history and one quarter of those who had been weight-stable are miscategorized as having undergone weight loss.77 Furthermore, the nutritional significance of changes in body weight can be confounded by changes in hydration status and extracellular fluid accumulation. Because weight loss is an imperfect indicator of PEM, it is useful to obtain other historical clues that can contribute to the identification of these patients (see later). Food Intake.  Has there been a change in habitual diet pattern (number, size, and contents of meals)? What is the reason for altered food intake (e.g., change in appetite, mental status or mood, ability to prepare meals, ability to chew or swallow, gastrointestinal symptoms)? Evidence of Malabsorption.  Are there symptoms that are consistent with malabsorption (e.g., greasy, fatty stools, osteopenia, easy bruisability)? Evidence of Specific Nutrient Deficiencies.  Are there symptoms of specific nutrient deficiencies, including macrominerals, micronutrients, and water? (See Tables 4-9 to 4-11.)

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient Influence of Disease on Nutrient Requirements.  Is the nature of the patient’s underlying illness one that is likely to increase nutrient needs or nutrient losses? Functional Status.  Has the patient’s ability to perform daily activities that determine consumption of wholesome meals changed? Can the patient still shop and prepare meals? Have finances interfered with the ability of the patient to purchase food? Physical Examination Hydration Status.  The patient should be evaluated for signs of dehydration (e.g., hypotension, tachycardia, postural changes, mucosal xerosis, decreased axillary sweat, dry skin), and excess body fluid (e.g., edema or ascites). Tissue Depletion.  A general loss of adipose tissue can be assumed if there is well-defined bony, muscular, and venous outlines and loose skin folds. A fold of skin, pinched between the forefinger and thumb, can reveal the adequacy of subcutaneous fat. The presence of hollowness in the cheeks, buttocks, and perianal area suggests body fat loss. An examination of the temporalis, interosseous, and quadriceps muscles should be made to judge muscle wasting. Muscle Function.  Strength testing of individual muscle groups can be performed to determine if there is generalized or localized muscle weakness. In addition, an evaluation of myocardial function and of respiratory muscle function using spirometry can be done. Specific Nutrient Deficiencies.  Rapidly proliferating tissues (e.g., oral mucosa, hair, skin, gastrointestinal epithelium, bone marrow) often are more sensitive to nutrient deficiencies than tissues that turn over more slowly (e.g., heart, skeletal muscle, brain; see Tables 4-9 to 4-11). Anthropometry Anthropometric techniques are those in which a quantitative measure of the size, weight, or volume of a body part is used to assess protein and calorie status. Historically, one of the most commonly used anthropometric parameters has been weight for height. This is a useful parameter when neither the patient nor family can provide reliable historical information, but is less desirable than a history of unintentional weight loss because it requires the patient’s weight to be judged against a normative standard that has been established in a large control population, and interindividual variability in the population limits the accuracy with which this method correctly predicts PEM in a single person. Table 4-6 displays the 1959 Metropolitan Health desirable body weights that were established with prospective mortality data. The 1959 tables remain preferable to the 1983 tables because of certain assumptions made in the construction of the latter. In the context of the Metropolitan tables, desirable weight for height is defined as that figure associated with maximal longevity. Generally speaking, individuals whose weight is less than 85% of the standard can be considered to have a clinically significant degree of PEM. Of note is that desirable weights in this table are substantially less than average weights in North America. Body mass index (Table 4-18), defined as weight (in kilograms) divided by height (in meters squared), has been supplanting the use of weight for height, in part because it precludes the need to use normative data tables. BMIs that are outside the desirable range (18.5 to 24.9  kg/m2) help identify patients at increased risk of adverse clinical out-

Table 4-18  Classification of Nutritional Status by Body Mass Index in Adults Body Mass Index (kg/m2)

Nutritional Status

<16.0 16.0-16.9 17.0-18.4 18.5-24.9 25.0-29.9 30.0-34.9 35.0-39.9 ≥40

Severely malnourished Moderately malnourished Mildly malnourished Normal Overweight Obese (class I) Obese (class II) Obese (class III)

comes. A BMI modestly above the desirable range has been shown to be predictive of adverse outcomes in the surgical management of many diseases78-80 and in the medical management of conditions such as alcoholic liver disease.81 Similarly, a low BMI has been shown to be a robust independent risk factor in surgical and medical patients.82 Adult patients who are extremely underweight (BMI < 14 kg/m2) are at high risk of death and should be strongly considered for admission to the hospital for the initiation of intensive nutritional support. The BMI, however, like weight for height, is a surrogate and imperfect measure of body composition. A low BMI (<18.5 kg/m2) is interpreted as an indication of PEM and a high BMI (>24.9 kg/m2) is interpreted as excessive fat mass (overweight or obesity). Although BMI is accurate in this regard for the vast majority of adults, it can be just as misleading as any other measure that relies on body weight without a direct evaluation of body composition.83 For example, the individual with excessive fluid accumulation, where the actual fat and body cell mass are less than that implied by the BMI, and in the muscle-bound athlete, where a high BMI is indicative of an extraordinarily large lean mass, are two examples in which the underlying assumptions inherent in the BMI are false. Gender and race also are confounding variables, although the differences are clinically irrelevant; more important are the remarkable changes in body composition that accompany development, making the interpretation of BMI in childhood and adolescence very complex.84 It should be apparent from this discussion that the measurement of relevant body compartments, such as the fat mass or fat-free mass, can reveal important information about nutritional status that is often obscured by measurement of weight alone. Underwater (hydrostatic) weighing, dual energy x-ray absorptiometry (DEXA), air impedance plethysmography, total body potassium, isotopically labeled water dilution, in vivo neutron activation analysis, computed tomography (CT), and magnetic resonance imaging (MRI) are accurate noninvasive or minimally invasive techniques of measuring body compartments.85-91 All are highly effective but because of their expense, lack of accessibility, and impracticality, their use is relegated largely to the sphere of clinical research. A detailed understanding of these tools is beyond the scope of this chapter; however, the primary use of each, with reference to detailed reviews, is outlined in Table 4-19. In the clinical setting, simple but less accurate techniques are used to assess body compartments. An approximate measure of whole-body fat mass can be derived from assessing the thickness of the subcutaneous fat, which, in a normally proportioned adult, contains approximately half of

65

66

Section II  Nutrition in Gastroenterology Table 4-19  Advanced Techniques for Measurement of Body Compartments Primary Use in Body Compartment Analysis

Technique Underwater (hydrostatic) weighing84 Air displacement plethysmography85 Dual energy x-ray absorptiometry86 Total body potassium87 Isotopically labeled water and NaBr dilution88 In vivo neutron activation analysis89 Computed tomography84 Magnetic resonance imaging90

Proportion of body composed of FM Proportion of body composed of LM Proportion of body composed of FM Proportion of body composed of LM Absolute FM and LM; bone density Body cell mass TBW, ICW, ECW Total body protein, absolute FM, absolute LM Regional FM/LM Regional FM/LM

ECW, extracellular water; FM, fat mass; ICW, intracellular water; LM, lean mass; NaBr, sodium bromide; TBW, total body water.

Table 4-20  Normative Standards for Upper Arm Muscle Area and Sum of Triceps and Subscapular Skinfolds Percentile Parameter

Age

5th

50th

85th

2

Upper Arm Muscle Area (cm )* Men 25-29 38 53 45-49 37 55 65-69 33 48 Women 25-29 20 30 45-49 21 32 65-69 22 35 Sum of Triceps and Subscapular Skinfolds (mm) Men 25-29 12 24 45-49 13 29 65-69 12 27 Women 25-29 18 37 45-49 21 46 65-69 22 45

65 66 63 38 45 46 41 43 42 58 68 65

*Mid–upper arm muscle area (cm2) is calculated as follows. For men: Area =

[arm circumference − {π × triceps skinfold}]2

4π

− 10

For women: Area =

[arm circumference − {π × triceps skinfold}]2

4π

− 6.5

Adapted with permission from Frisancho AR. Anthropometric standards for the assessment of growth and nutritional status. Ann Arbor, Mich: University of Michigan Press; 1990.

the body’s adipose stores. The triceps and subscapular sites are used most commonly for this purpose, and it is best to use the sum of the triceps and subscapular folds because sizable interindividual differences exist in fat distribution. Furthermore, as total body fat changes, the subcutaneous fat at each site responds in a different manner. Similarly, midarm muscle circumference (MAMC) provides a measure of skeletal muscle mass. Table 4-20 contains guidelines for

interpretation of skinfold and midarm muscle area based on data from the first two National Health and Nutrition Examination Surveys (NHANES I and II).92 The clinical use of skinfolds and appendicular muscle area has distinct weaknesses. As was true for the weight for height tables, there is considerable interindividual variation in values, so these measurements are more useful in population studies than in an individual. Moreover, these measures are highly operator-dependent.93 Also, although the updated databases defining normative values no longer contain the biases of race and age of older versions, correction factors for hydration and physical activity still are not available. In practice, I have found the most useful clinical role for the measurement of skinfolds and muscle area is in tracking patients with serial measurements over time as a means of monitoring their recovery from disease or response to a clinical intervention. In this manner, the patient is being compared with himself or herself rather than with some normative value. In gastroenterology, the use of skinfolds and muscle area has been of particular usefulness in the assessment and management of cirrhotic patients because cirrhosis corrupts almost all the other common measures of nutritional status. Abnormally low values for triceps skinfold (TSF) and MAMC are independent predictors of mortality in cirrhotics and their incorporation into a Cox regression model improves the prognostic value of the Child-Turcotte score.94 Also, when patients with severe alcoholic hepatitis are treated with anabolic steroids, improvements in MAMC and other measures of the fat-free mass (FFM) correlate with a positive response to treatment.95 Interest continues in bioimpedance analysis (BIA) as an inexpensive, relatively easy, noninvasive, and safe means of assessing FFM, body cell mass (BCM), and total body water (TBW). BCM is sometimes perceived as a more important measure of lean mass than FFM because it does not include nonliving lean mass, such as the blood plasma and bone mineral (see Fig. 4-1). Resistance to electrical flow through the body is measured, which is proportional to fat and bone mineral content because these body components have poor conductivity. Because all other components of the body are suffused with electrolyte-laden water that readily conducts an electrical current, calculations of TBW, FFM, and BCM can be made if one abides by some general assumptions that define the water content of each compartment. Those with expertise in its use have found it useful for monitoring the FFM in outpatient studies of renal dialysis and human immunodeficiency virus (HIV) patients.96,97 Acute illness, however, produces shifts in the total amount of body water and its distribution in the different compartments, rendering the technique largely worthless in the inpatient setting.98 Furthermore, the algorithms used to calculate body com­ position contain assumptions about body water that can change with age, obesity, and disease, so BIA must be revalidated within any population in which it is used. Functional Measures of Protein-Calorie Status Three different techniques have been developed that exploit the fact that skeletal muscle function is impaired in PEM: fist-grip dynamometry (FGD), adductor pollicis electrical stimulation, and respiratory muscle strength evaluation. FGD uses a hand-held dynamometer to measure the maximal fist-grip force that can be elicited. When examined as a surrogate measure of total body protein in patients awaiting gastrointestinal surgery, FGD correlated strongly with in vivo neutron activation analysis and with MAMC.99 Similarly, FGD is excellent for detecting depleted body cell mass in cirrhotic patients,100 a group in which it is notoriously

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient difficult to perform nutritional assessment. As noted, valid indicators of moderate to severe PEM are strong predictors of clinical outcome in acutely ill patients and FGD is effective in this regard. Preoperative patients whose fist grip strength is less than 85% of age-and gender-corrected standards have a twofold increased risk of perioperative complications compared with those whose FGD is normal.101 In patients undergoing surgery for gastrointestinal cancers, FGD had superior sensitivity and specificity in predicting perioperative morbidity and mortality than a widely used discriminant analysis called the prognostic nutritional index.102 FGD holds considerable promise for the rapid and convenient assessment of protein-calorie status in inpatients and outpatients. The technique is nevertheless limited by its need for an alert and cooperative patient. Respiratory muscle strength, typically measured with a bedside spirometer as maximal sustained inspiratory force and/or maximal sustained expiratory force, also has been used as a measure of protein-calorie status,103 but is generally considered not reliable because too many nonnutritional factors may alter its measurement. Electrical stimulation of the ulnar nerve with measurement of various functions of the adductor pollicis muscle also has been explored as an indicator of protein-calorie status. The technique has been shown to distinguish between malnourished subjects and well-nourished controls,104 and to predict postoperative complications in patients awaiting surgery for gastrointestinal malignancies105; abnormal muscle function was observed to resolve with nutritional restitution. Use of this method in an ICU

setting, however, was found to be very problematic106 because patients found it uncomfortable and neuromuscular blocking agents interfered with obtaining adequate data. Much more work is needed to establish this method’s validity and appropriate applications. Although PEM adversely affects the physiology of almost all organ systems, the immune system is in the most sensitive; delayed hypersensitivity skin testing, which assesses the integrity of cell-mediated immunity, has been used most often in this regard. In critically ill patients, skin testing has value in predicting mortality,107 but its interpretation is fraught with confounding variables, such as older age, systemic infection, and major surgery, each of which will independently depress reactivity. Furthermore, reactivity improves in an unpredictable manner with nutritional restitution, so it is not useful for monitoring the progress of patients.108 The value of skin testing in assessing nutritional status is used best as part of an array of parameters that collectively assess nutritional status (see later, “Discrim­ inant Analyses of Protein-Calorie Status”). Biochemical Measures of Protein-Calorie Status Serum Proteins.  The serum concentrations of several proteins that are synthesized in the liver are used as indicators of protein-calorie status—albumin, prealbumin (transthyretin), transferrin, and retinol-binding protein (RBP; Table 4-21). A low concentration of any of these proteins strongly suggests the presence of PEM in an individual who does not have a concurrent illness or injury. Because the half-lives of prealbumin, transferrin, and RBP are considerably shorter

Table 4-21  Hepatically Synthesized Proteins Used for Assessment of Nutritional Status

Serum Protein

Normal Value mean ± SD (range)*

Half-life (days)

Function

Comment†

Maintains plasma oncotic pressure; carrier for small molecules Binds Fe2+ in plasma and transports it to bone marrow

Serum levels are determined by many different processes

Albumin

4.5 (3.5-5.0)

14-20

Transferrin

2.3 (2.0-3.2)

8-9

Transthyretin (prealbumin)

0.30 (0.2-0.5)

2-3

Binds T3 and to a lesser extent T4; carrier for retinol-binding protein

Retinol-binding protein (RBP)

0.0372 ± 0.0073‡

0.5

Transports vitamin A in plasma; binds noncovalently to prealbumin

Iron nutriture influences plasma level; increased during pregnancy, estrogen therapy, and acute hepatitis; reduced in protein-losing enteropathy and nephropathy, chronic infections, uremia, and acute catabolic states; often measured indirectly as total iron-binding capacity Increased in patients with chronic kidney disease on dialysis; reduced in acute catabolic states, after surgery, in hyperthyroidism; serum level is determined by overall energy and nitrogen balance Catabolized in renal proximal tubular cell; with renal disease, RBP increases and half-life is prolonged; low plasma levels in vitamin A deficiency, acute catabolic states, after surgery, in hyperthyroidism

*Units are g/L. Normal range varies among centers; check local values. † All the listed proteins are influenced by hydration and presence of hepatocellular dysfunction. ‡ Normal values are age- and gender-dependent. Table value is for pooled subjects. T3, triiodothyronine; T4, thyroxine. Adapted from Heymsfield S, Tighe A, Wang Z-M. Nutritional assessment by anthropometric and biochemical means. In: Shils M, Olson J, Shike M, editors. Modern Nutrition in Health and Disease. 8th ed. Philadelphia: Lea & Febiger; 1994. p 812.

67

68

Section II  Nutrition in Gastroenterology Table 4-22  Normative Values for Creatinine Excretion Based on Height Women†

Men* Height (cm)

Ideal Creatinine (mg)

Height (cm)

Ideal Creatinine (mg)

157.5 160.0 162.6 165.1 167.6 170.2 172.7 175.3 177.8 180.3 182.9 185.4 188.0 190.5 193.0

1288 1325 1359 1386 1426 1467 1513 1555 1596 1642 1691 1739 1785 1831 1891

147.3 149.9 152.4 154.9 157.5 160.0 162.6 165.1 167.6 170.2 172.7 175.3 177.8 180.3 182.9

830 851 875 900 925 949 977 1006 1044 1076 1109 1141 1174 1206 1240

*Creatinine coefficient (men) = 23 mg/kg of ideal body weight. † Creatinine coefficient (women) = 18 mg/kg of ideal body weight. From Blackburn GL, Bistrian BR, Maini BS, et al: Nutritional and metabolic assessment of the hospitalized patient. JPEN J Parenter Enteral Nutr 1977; 1:11-22.

than that of albumin, it follows that changes in nutritional status will be reflected more promptly in levels of these three than in albumin. A variety of factors alter the serum concentration of each of these proteins. For example, prealbumin levels often are elevated in chronic kidney disease or by glucocorticoid or oral contraceptive administration. The degree to which serum levels of all of these proteins are decreased in cirrhosis increases incrementally with worsening grades of the Child classification, although even patients who are Child class A have a small decrease in albumin compared with healthy individuals.109 All these proteins behave as negative acute-phase reactants—that is, their serum concentrations drop in response to systemic inflammation, roughly proportional to the magnitude of the inflammatory response. This effect severely curtails their reliability as indicators of PEM in the acutely ill patient.110 Nonetheless, with proper respect for their limited accuracy, they still can be useful; for example, prealbumin has been shown to be an efficient rapid means of screening inpatients for PEM on hospital admission.111 Creatinine-Height Index.  The amount of creatinine excreted in the urine over a 24-hour period, corrected for the patient’s height, is an excellent means of assessing total skeletal muscle mass. The relationship holds because a relatively constant percentage (~2%) of muscle creatine is converted to creatinine each day. Values that are more than 20% below gender-and-height-adjusted normative values are indicative of moderate to severe PEM (Table 4-22). Updated normative creatine-height index (CHI) values for children aged 3 to 18 years are available.112 In sick persons, the CHI tends to correlate with simple measures such as unintentional weight loss, as well as with highly accurate measures of skeletal muscle, such as DEXA.113 However, incomplete urine collection, abnormal or unstable renal function, excessive meat or milk ingestion immediately preceding or during the collection, and glucocorticoid administration all can alter creatinine excretion independently of changes in muscle mass. Discriminant Analyses of Protein-Calorie Status.  As noted, many of the parameters used to measure PEM also can predict important clinical outcomes; each parameter,

Table 4-23  Prognostic Indices in Hospitalized Patients Index Likelihood of malnutrition

Prognostic nutritional index Instant nutritional index Hospital prognostic index

Incorporated Parameters

Correlates with

Serum folate, serum vitamin C, serum albumin, lymphocyte count, hematocrit, triceps skinfold, arm muscle circumference, weight Serum albumin, serum transferrin, delayed hypersensitivity, triceps skinfold Serum albumin, lymphocyte count

Duration of hospitalization

Serum albumin, delayed hypersensitivity, presence of sepsis or cancer

Frequency of postoperative complications and mortality Frequency of postoperative infection Hospital mortality

Adapted from Mason J, Rosenberg I. Protein-energy malnutrition. In: Isselbacher K, Braunwald E, Wilson J, et al, editors. Harrison’s principles of internal medicine. 13th ed. New York: McGraw-Hill; 1994. p 440.

however, has its own limitations. Multifactorial indices that incorporate various combinations of these parameters have been developed through the use of discriminant analyses. By combining the strengths of several parameters, the goal is to arrive at a prognostic index that is more accurate for determining whether patients have a substantial degree of PEM and for optimizing the ability to predict which patients will have adverse clinical outcomes because of PEM, thereby identifying those who might benefit from intensive attention to their nutritional needs. Table 4-23 lists the prognostic indices that have been most widely studied, along with the outcomes that each has been shown to predict. Because many of the parameters incorporated into these indices are influenced by the severity of disease and by nutritional status, such indices are more properly thought of as assess-

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient ing an integration of illness severity and the likelihood of malnutrition. Rapid Screening Tools for the Assessment of Targeted Populations.  Inexpensive, rapid, and convenient means of assessing nutritional status that are accurate in identifying patients with PEM are of great value, particularly when large numbers of people need to be evaluated. Two such tools have been developed and extensively validated, the subjective global assessment (SGA) and the Mini-Nutritional Assessment (MNA). Subjective Global Assessment.  This initially was intended for use in surgical inpatients as a means of assessing nutritional status and predicting postoperative infections; for the latter, it was found to be a better predictor than serum albumin concentration, delayed skin hypersen­ sitivity, MAMC, CHI, and the prognostic nutritional index.114 A focused history and physical examination are used to categorize patients as well nourished (category A), having mild or moderate malnutrition (category B), or having severe malnutrition (category C; Table 4-24). In spite of

Table 4-24  Subjective Global Assessment (SGA) of Nutritional Status History Weight change Loss in past 6 months: amount = __________ kg; % loss = __________ Change in past 2 weeks: __________ Increase __________ No change __________ Decrease Dietary intake change: No change ________ Change ________ Duration = ________ weeks Dietary status: __________ Suboptimal solid diet __________ Hypocaloric liquids __________ Starvation Gastrointestinal symptoms (that have persisted for >2 weeks): __________ None __________ Nausea __________ Vomiting __________ Diarrhea __________ Anorexia Functional capacity: __________ No dysfunction __________ Dysfunction Duration = __________ Weeks Type: __________ Working suboptimally __________ Ambulatory but not working __________ Bedridden Effect of disease on nutritional requirements: Primary diagnosis: __________ Metabolic demand: ________ Low stress ________ Moderate stress __________ High stress Physical Examination (Normal, Moderate, or Severe) __________ Loss of subcutaneous fat (triceps, chest) __________ Muscle wasting (quadriceps, deltoids) __________ Ankle or sacral edema __________ Ascites SGA Rating* A = Well nourished B = Mild or moderate malnutrition C = Severe malnutrition *The ranks of A, B, and C in the SGA are assigned on the basis of subjective weighting. A patient with weight loss and muscle wasting who is currently eating well and gaining weight is classified as well nourished. A patient with moderate weight loss (between 5% and 10%), continued compromise in food intake, continued weight loss, progressive functional impairment, and moderate stress because of illness is classified as moderately malnourished. A patient with severe weight loss (>10%), poor nutrient intake, progressive functional impairment, and muscle wasting usually is classified as having severe malnutrition.

the subjective nature of some of its components, there is excellent agreement in independent observers.115 The SGA has been shown to be reliable, even in the hands of first-year medical and surgical residents116 and validated as a pre­ dictor of clinical outcomes in chronically institutionalized older adults and in patients with a variety of medical conditions.117-119 Mini-Nutritional Assessment.  This was developed as a rapidly administered screen to detect PEM in geriatric populations. A combination of history, brief physical examination, and simple anthropometrics (BMI, arm and calf circumference) can be obtained in a few minutes. Subjects receive a score that classifies them as being nourished, malnourished, or at risk of malnutrition. The MNA is a valid means of detecting PEM in older adults who are generally healthy and ambulatory as well as those who are frail and institutionalized120,121; in the chronically institutionalized, it possesses considerable predictive value in projecting future morbidity.122 One disadvantage of the MNA is that it does not screen for overweight or obesity. Other screening tools designed for geriatric populations, such as the Nutrition Screening Initiative, have the ability to screen for under- and overnutrition but have not been as extensively validated as the MNA.123

AGGRESSIVE NUTRITIONAL SUPPORT IN THE HOSPITALIZED PATIENT Aggressive nutritional support, here defined as using whatever means is necessary and practical to meet the nutritional needs of the patient, will not benefit every acutely ill patient. In practice, any acutely ill patient who has moderate to severe malnutrition and is unlikely to be able to meet his or her own nutritional needs within 48 hours is a strong candidate for aggressive nutritional support. Another common indication for aggressive nutritional support is when a well-nourished or mildly malnourished inpatient is judged to be unlikely to meet at least 80% of his or her projected calorie or protein goals for the coming 10 days. However, to date, there is insufficient evidence-based science to prove efficacy of this latter indication. Catabolic forces that accompany acute illness make it difficult to correct nutritional deficits. In those with a high degree of sustained metabolic stress, nutritional support generally will not lead to an increase in the protein compartment of the body. Moreover, a gain in weight may not occur and, when it does, much of the initial gain is from water and an expanded fat mass.124 Despite these limitations, however, even in the absence of weight gain or increases in serum protein levels, a course of nutritional support for an appropriate patient can improve physiologic functions and clinical outcome.125 The following sections cite some clinical scenarios particularly relevant to gastroenterology for which there exists compelling clinical research that aggressive nutritional support provides benefit to the inpatient.

MALNOURISHED PATIENTS UNDERGOING MAJOR SURGERY

Nutritional support can be beneficial for moderately to severely malnourished patients who are scheduled to undergo major surgery. Aggressive nutritional support for seven or more days before surgery reduces perioperative complications, and sometimes mortality, in malnourished

69

70

Section II  Nutrition in Gastroenterology patients.66-73,126-128 In the VA Cooperative Trial,66 which encompassed almost 500 subjects about to undergo major abdominal or thoracic surgery, patients who were categorized as severely malnourished and randomized to receive preoperative TPN realized an almost 90% decrease in noninfectious perioperative complications. No benefits were observed in mildly malnourished or well-nourished individuals. In trials of moderately to severely malnourished patients, preoperative nutrition support generally conveys sizeable benefits: one trial that enrolled 90 patients with gastric or colorectal cancers undergoing surgery demonstrated a 35% decline in overall complications and a significant reduction in mortality.128 The observation that the benefits of preoperative nutritional support are confined to those with a substantial degree of malnutrition is the same conclusion reached by meta-analyses.71,72 Deferring aggressive nutritional support until after surgery does not appear to have the same ability to diminish perioperative complications. 129 Provision of nutrients via an enteral approach also is beneficial. There have been fewer trials done of preoperative enteral support than of preoperative TPN, but it appears that preoperative enteral support conveys the same nutritional130 and clinical131 benefits as TPN. As with TPN, postoperative enteral nutrition in the absence of aggressive preoperative support is less likely to convey benefit to the patient.132

PATIENTS HOSPITALIZED WITH DECOMPENSATED ALCOHOLIC LIVER DISEASE

The prevalence of moderate to severe PEM is so high in patients admitted for acute alcoholic hepatitis and other forms of decompensated alcoholic liver disease65 that it is best to assume that all such patients are malnourished. Furthermore, patients with acute alcoholic hepatitis usually fall far short of their nutritional needs when allowed to eat freely. Clinical trials have demonstrated that the rates of morbidity, mortality, and the speed of recovery are improved with prompt institution of enteral or parenteral nutrition in these patients.67-69,133

PATIENTS UNDERGOING RADIATION THERAPY

The usefulness of aggressive nutrition support in patients undergoing radiation therapy has been studied most extensively in those who have head and neck and esophageal cancers. There is now reasonable evidence in these patients that placement of a percutaneous endoscopic gastrostomy (PEG) tube and administration of supplemental tube feedings during and after the course of radiation therapy prevents further deterioration of nutritional status.134,135 In patients with head and neck cancers, supplemental PEG feedings also have been shown to improve quality of life (QOL). Although improvements in survival or decreased morbidity have not yet been demonstrated, the improved QOL alone may warrant its use in this setting.

REFEEDING SYNDROME Injudicious and overly aggressive nutritional therapy can have adverse clinical consequences, known as the refeeding syndrome, in those who have suffered sustained inadequacy in food intake.136,137 Early evidence of the refeeding syndrome was reported at the end of World War II, when it was found that refeeding chronically semistarved war victims and research volunteers caused cardiac insufficiency and

neurologic complications. More recently, refeeding abnormalities and serious complications have been reported after aggressive refeeding of hospitalized cachectic patients.138 The likelihood that a refeeding syndrome will occur depends on the severity of underlying PEM and the rapidity with which nutritional restitution is undertaken. Thus, the syndrome usually can be avoided merely by approaching nutritional repletion in an incremental fashion over the first several days of nutritional therapy and careful moni­ toring for potential complications, with interventions as indicated.

MINERAL DEPLETION

Of the mineral abnormalities associated with refeeding, phosphate depletion has received the most attention. Refeeding with enteral carbohydrates or glucose-based parenteral formulas stimulates insulin release and intracellular uptake of phosphate.139 This phenomenon can cause the extracellular phosphorus concentration to fall below 1 mg/ dL within hours of initiating nutritional therapy if adequate phosphate is not given. Severe hypophosphatemia, accompanied by muscle weakness, paresthesias, seizures, coma, cardiopulmonary decompensation, and death, has occurred in patients receiving enteral or parenteral nutritional repletion.138,140 Potassium and magnesium are the most abundant intracellular cations. Loss of body cell mass in the malnourished patient causes whole-body potassium and magnesium depletion. However, serum concentrations of these two electrolytes remain normal or near-normal during starvation because of their release from tissue and bone stores. The increases in protein synthesis rates, body cell mass, and glycogen stores during refeeding require increased intracellular potassium and magnesium. In addition, hyperinsulinemia during refeeding increases the cellular uptake of potassium and can cause a rapid decline in its intravascular concentration.141

CARDIOVASCULAR COMPLICATIONS

Chronic undernutrition is associated with decreased cardiac mass, stroke volume, and end-diastolic volume, bradycardia, and fragmentation of cardiac myofibrils.142-145 In addition, carbohydrate refeeding increases the concentration of circulating insulin, which enhances sodium and water reabsorption by the renal tubule.146 Moreover, if hypophosphatemia develops, it can impair oxygen unloading from red blood cells and increase susceptibility to ventricular tachyarrhythmias. A serum phosphate concentration less than 2.0  mg/dL further impedes myocardial performance. These factors put the severely malnourished patient at increased risk of fluid retention and congestive heart failure after nutritional therapy, and may precipitate cardiovascular collapse. The cardiovascular complications of thiamine deficiency (wet beriberi) also have been reported to con­ tribute to cardiovascular compromise in the refeeding syndrome.147 Because thiamine is an essential cofactor for several enzymes in carbohydrate metabolism, it is believed that overzealous administration of carbohydrate to a severely malnourished patient with limited thiamine availability may precipitate an acute thiamine deficiency state. The high concentrations of glucose in TPN seem particularly likely to produce this effect. Ventricular tachyarrhythmias, which can be fatal, occur during the first week of refeeding,148 and may be preceded by prolongation of the QT interval.

GLUCOSE INTOLERANCE

The adaptive changes during starvation enhance the use of FAs and ketone bodies for fuel while glucose is conserved. In addition, the ability of insulin to stimulate glucose uptake

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient and oxidation by peripheral tissues is impaired.141 Thus, refeeding with high-carbohydrate meals or large amounts of parenteral glucose may not be well tolerated initially and may produce marked elevations in the blood glucose level, glucosuria, dehydration, and hyperosmolar coma.149

GASTROINTESTINAL DYSFUNCTION

As noted, starvation and malnutrition cause structural and functional deterioration of the gastrointestinal tract, thereby limiting digestion and absorption.150,151 When malnutrition is severe, oral refeeding has been associated with an increased incidence of diarrhea and death; most of the adverse gastrointestinal consequences of starvation resolve after one to two weeks of refeeding.

CLINICAL RECOMMENDATIONS

The severity of complications during the refeeding of cachectic, chronically semistarved patients dictates caution, particularly during the first week of therapy, when the risk of complications is highest. A careful search for cardiovascular and electrolyte abnormalities should be performed before refeeding. Judicious resuscitation with fluids and electrolytes may be necessary before beginning feedings to prevent congestive heart failure. Vitamin supplementation should be given routinely. Severely malnourished patients are poi­ kilothermic, so a warm ambient temperature and warming blankets may be necessary to increase the core temperature slowly. Oral or enteral tube feedings are preferred over parenteral feeding for nutritional restitution because of their fewer serious complications and enhanced gastrointestinal tract recovery. Isotonic feedings should be given in small amounts at frequent intervals to prevent overwhelming the body’s limited capacity for nutrient processing and to prevent hypoglycemia, which can occur during brief nonfeeding intervals. Parenteral supplementation or TPN may be necessary if the intestine cannot tolerate oral or enteral feeding. Although it is impossible to know the precise nutrient requirements of individual patients, some general guidelines are recommended for the first week of refeeding. Fluid intake for adults initially should be limited to approximately 800 mL/day plus replacement for insensible losses; adjustments are needed in fluid-overloaded or dehydrated patients. Changes in body weight are a useful guide to evaluate the efficacy of fluid administration. Weight gain of more than 0.25 kg/day, or 1.5 kg/week, probably represents fluid accumulation. Daily calorie intake should be approximately 15 to 20 kcal/kg and contain approximately 100 g of carbohydrate and 1.5 g of protein/kg body weight. Sodium should be restricted to approximately 60  mEq or 1.5  g/day, but liberal amounts of phosphorus, potassium, and magnesium should be given to patients who have normal renal function. Daily monitoring of body weight, fluid intake, urine output, and plasma glucose and electrolyte values (including magnesium and phosphate) are critical during early refeeding (first 3 to 5 days) so that appropriate adjustments can be made.

MANAGEMENT OF SEVERE MALABSORPTION: A NUTRITIONAL PERSPECTIVE The medical management of patients with severe malabsorption, such as that which often occurs in short bowel syndrome or pancreatic insufficiency, can be challenging

and frustrating. In many cases, the patient and clinician must be prepared to accept incremental improvements as the management goal rather than complete resolution of symptoms. Great strides and an improved QOL, however, often can be accomplished with only modest improvements in absorptive physiology. Malabsorption syndromes and short bowel syndrome are discussed in greater detail in Chapters 101 and 103, respectively.

CLINICAL CONSIDERATIONS

Initial assessment of the patient with chronic malabsorption is meant to provide a set of rational strategies to improve the patient’s clinical condition and prevent complications. The therapeutic approach depends on the following: the level at which the patient’s intestinal tract functions; the presence of macronutrient, micronutrient, electrolyte, and fluid deficits; the identification of risk factors for future medical complications; the presence of coexisting diseases that hamper the ability to provide nutritional therapy; and an evaluation of factors that affect the patient’s daily activities. A careful review of medical records, operative reports, and radiologic studies is needed to evaluate the absorptive capacity of the intestine by determining the length and status of the remaining intestine, the site of intestinal disease or resection, and the presence of other conditions that reduce intestinal absorption. An assessment of fluid losses through diarrhea, ostomy output, and fistula volume should be made to help determine fluid requirements. Knowledge of fluid losses also is used to calculate intestinal mineral losses by multiplying fluid loss by the estimated electrolyte concentration in intestinal fluid (Table 4-25). In patients who do not respond to treatment as predicted, dynamic studies of intestinal absorptive function may be helpful for adjusting the treatment program. Such studies may include measuring fat, carbohydrate, or nitrogen balance and evaluating ostomy, fecal, or fistula mineral and fluid losses. The urgency for medical intervention is determined by the severity of hemodynamic, electrolyte, and nutritional abnormalities. This requires an evaluation for volume depletion, weight loss, and specific nutrient deficiencies. In addition to standard laboratory tests to evaluate whether micronutrient deficiencies are causes of anemia (iron, folate, vitamin B12), prolonged prothrombin time (vitamin K deficiency), and neurologic deficits (vitamin B12, thiamine, vitamin E), more sophisticated measurements to determine vitamin and trace mineral status can be obtained when such deficiencies are suspected clinically. Water-soluble vitamin

Table 4-25  Electrolyte Concentrations in Gastrointestinal Fluids* Electrolyte (mEq/L) Location

Na

K

Cl

HCO3

Stomach Bile Pancreas Duodenum Mid-small bowel Terminal ileum Rectum

65 150 150 90 140 140 40

10 4 7 15 6 8 90

100 100 80 90 100 60 15

— 35 75 15 20 70 30

*Average values are listed; these can vary considerably from patient to patient.

71

72

Section II  Nutrition in Gastroenterology deficiencies tend to occur as a cluster, as do fat-soluble vitamin deficiencies, so measurements of red blood cell folate and serum 25-hydroxyvitamin D levels are useful as an initial screen because a deficiency of one vitamin raises the suspicion of additional deficiencies. Bone mineral densitometry may be useful to establish a baseline and to screen for unrecognized bone mineral depletion. Metabolic bone disease is common in chronic malabsorptive disorders, particularly so in individuals on long-term TPN.152 An accurate dietary history is important in evaluating nutrient requirements in nutritionally stable patients and for identifying dietary inadequacies in those with nutrient deficiencies. Working in conjunction with a skilled dietitian for the evaluation and treatment phases of management often is invaluable.

TREATMENT

The goals of therapy are to control diarrhea, maintain fluid, electrolyte, and nutritional homeostasis, treat and prevent medical complications, and maximize the QOL. Striving for an optimal QOL rather than insisting on optimal physiology cannot be overemphasized, and often requires tailoring therapy to the specific needs of each patient. For example, maintaining adequate nutritional status with oral feedings at the cost of massive diarrhea and frequent bowel movements may be unacceptable to the patient with an active social or professional life outside the home; in such a patient, parenteral supplementation may be necessary to improve the QOL. Initial therapy often requires subsequent modification using a trial and error approach because of individual variability in absorptive function, continued intestinal adaptation, and the development of new medical complications or disease progression. Continued clinical monitoring is critical so that medical and nutritional therapy can be adjusted when necessary.

Control of Diarrhea

Diarrhea often is caused by a combination of factors, including increased gastrointestinal secretions, decreased intestinal transit time, and osmotic stimulation of water secretion by unabsorbed contents of the food stream. In addition, unabsorbed bile acids and fatty acids are potent stimuli of fluid and electrolyte secretion by the colonic mucosa. Therefore, therapy for diarrhea involves limiting endogenous secretions, slowing motility, and improving solute absorption. Normally, the stomach produces approximately 2.5  L of fluid/day, which is absorbed by the small intestine and colon. Gastric secretion therefore may contribute to diarrhea if there is inadequate absorptive capacity or if there is an excess amount of gastric secretion, such as with ZollingerEllison syndrome. Immediately following extensive resection of distal bowel, a gastric hypersecretory state occurs, probably from loss of inhibitory peptides such as glucagonlike peptide-2 (GLP-2) and peptide YY153,154; recovery from this hypersecretory state occurs over a period of months. Histamine (H2) receptor antagonists or proton pump inhibitors are used to reduce gastric secretions, and the presence of acidic jejunostomy or ileostomy contents after meals is a clear indication for acid reduction therapy.155 Large doses, twice the normal amount used for the treatment of peptic ulcer disease or gastroesophageal reflux, may be required for adequate control in some patients because of reduced drug absorption. The somatostatin analog, octreotide acetate (Sandostatin), can decrease small intestine secretions. Therapy with

octreotide has been shown to decrease ostomy or stool volume, decrease sodium and chloride output, and prolong small intestine transit time in patients with short bowel syndrome.156,157 Octreotide therapy, however, usually does not improve absorption of macronutrients and other minerals; in fact, it may exacerbate the degree of fat malabsorption,158-160 presumably by inhibiting pancreatic secretions. In addition, octreotide is expensive, must be given by subcutaneous injection, can decrease appetite and increase the risk of gallstones,161 and diminishes protein synthesis in the intestinal epithelium and exocrine pancreas.162 Nevertheless, in patients who have persistent large-volume intestinal output despite standard antidiarrheal therapy, a trial of 100  µg of octreotide injected subcutaneously three times daily with meals may be useful. A long-acting form of octreotide, which only requires monthly administration, has not been studied as thoroughly but has been effective in controlling diarrhea in some settings.163 Opiates are the most effective means for slowing intes­ tinal motility and act by delaying gastric emptying, de­ creasing peristalsis of the small intestine and colon, and increasing anal sphincter tone. Loperamide (Imodium) should be tried first because it is metabolized on first pass by the liver and does not easily cross the blood-brain barrier, thereby limiting its side effects and potential for drug dependence. If loperamide is not effective, other opiates such as codeine or deodorized tincture of opium (DTO, 5 to 15 drops every 6 hours) should be considered. In addition, the combination of an anticholinergic drug and an opiate may be beneficial. Capsules containing 25  mg powdered opium and 15  mg of powdered belladonna are a potent combination, although such capsules are not commercially available and require compounding. Diphenoxylate with atropine (Lomotil) is an effective agent, but is rather expensive if large doses are needed. Patients with severe malabsorption often require much larger than conventional doses of antidiarrheal medications. Pushing the dosage to the limits of the therapeutic range, while maintaining diligent attention to the potential for side effects, often is necessary. Foods and medications that cause diarrhea should be avoided. Traditionally, the recommendation has been to decrease or eliminate lactose-containing foods because of the reduction in intestinal lactase with intestinal resection. Patients with jejunostomies and 15 to 150  cm of jejunum remaining, however, can tolerate 20-g lactose loads as milk or yogurt.164 Foods that have laxative effects, such as caffeine-containing drinks and diet products containing osmotically active sweeteners (e.g., sorbitol, xylitol, mannitol), should be avoided. Medications that contain magnesium or sorbitol also can contribute to diarrhea.165 In individuals with short bowel syndrome and uncontrolled diarrhea, soft drinks and fruit juices should be diluted 1 : 1 with water to reduce their high osmotic load.

Enteral Feeding

Patients with nausea, vomiting, abdominal pain, or severe diarrhea may be unable to tolerate enteral feeding, regardless of their intestinal absorptive capacity. Specific foods that cause gastrointestinal complaints should be avoided, but it is important to evaluate the validity of these complaints to prevent the unnecessary withdrawal of nutritious foods. The goal of feeding is to provide the patient with all recommended nutritional requirements. The amount of ingested nutrients needed to reach this goal depends on the normal RDAs, modified by an estimate of absorptive function and intestinal losses; this usually requires ingestion of

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient large amounts of fluid, calories, protein, vitamins, and minerals. Even in patients with severe short bowel syndrome, TPN may not be needed when vitamin and mineral supplements and large amounts of calories and protein are provided enterally.166 Increasing the time that food is in contact with the intestine may enhance absorption in patients with limited absorptive function. For this reason, total dietary intake should be divided into at least six small meals daily. If this is unsuccessful, defined liquid formulas ingested between meals or administered by continuous tube feedings at night may preclude the need for TPN. Patients with severe malabsorption often must ingest 40 to 60  kcal/kg/ day and 1.2 to 1.5  g of protein/kg/day to avoid malnutrition, but this is highly variable. A more practical goal might be to consume however much is necessary to sustain a stable and reasonable weight while avoiding dehydration, electrolyte disorders, and any micronutrient deficiencies. The macro- and micronutrient needs of each patient often are arrived at only by experimentation with different regimens. Fat Intake Fat intake need not be restricted in patients whose colon is not in continuity with the food stream, despite the presence of steatorrhea. Although food preferences and tolerances differ from patient to patient, a high-fat, low-carbohydrate diet is comparable to a low-fat, high-carbohydrate diet with regard to total fluid, energy, nitrogen, sodium, potassium, and divalent ion absorption in patients with short bowel syndrome whose colon is not in continuity.167-169 By retaining fat in the diet, the patient’s ability to achieve caloric goals will be enhanced because of the high caloric density of fat. In contrast, in those who have steatorrhea and a colon in continuity with the food stream, unabsorbed fatty acids are potent stimuli to colonic ion and fluid secretion, and limiting fat intake usually diminishes gastrointestinal symptoms, colonic water secretion, hyperoxaluria, and divalent cation losses.170,171 Patients who have only a portion of the colon in continuity with the food stream behave in a variable manner in this regard. Medium-chain triglycerides (MCTs) can be very useful as a feeding supplement in patients who have impaired fat absorption, because MCTs are hydrolyzed rapidly and do not require bile salts and micelle formation for absorption. The efficacy of MCTs, however, usually is confined to patients whose colon is in continuity with the food stream, apparently because the colon is a site for MCT absorption.172 In contrast, there is usually little improvement in overall

energy absorption in patients with jejunostomies or ileostomies. Many patients do not find pure MCT oil to be palatable (used as a salad oil or as a spread on toast), and it also may cause nausea, vomiting, or abdominal discomfort. Unfortunately, liquid nutritional supplements designed for oral intake that contain most of their fat calories as MCT have been discontinued because of low demand. A homemade substitute nevertheless can be readily produced by mixing a package of Carnation Instant Breakfast, 8 ounces of cold skim milk, and 2 teaspoons of MCT in an electric blender for a few seconds. Predigested Liquid Nutritional Supplements Predigested formulas—monomeric (also called elemental; e.g., Vivonex T.E.N.) and oligomeric (also called semielemental; e.g., Peptamen, Reabilan)—claim benefit for patients with short bowel syndrome. Theoretically, these formulas, which contain nitrogen in the form of free amino acids or small peptides, are absorbed more efficiently than polymeric formulas or whole food. However, when subjected to clinical trials, they have shown no benefit169 or a marginal increase in nitrogen absorption.173 At present, there is insufficient clinical evidence to justify the routine use of expensive predigested formulas in patients with short bowel syndrome. Oral Rehydration Therapy In many patients with severe malabsorption, maintenance of fluid and electrolyte homeostasis constitutes a major problem. These patients may benefit from oral rehydration therapy (ORT), taking advantage of the sodium-glucose cotransporter in the intestinal brush border.174 Frequent ingestion of small-volume feedings of an isotonic glucose or starch-based electrolyte solution stimulates active sodium transport across the intestine, and water follows passively by solvent drag. Intestinal perfusion studies and clinical trials have indicated that sodium and water absorption are maximal from solutions containing 60 to 75  mmol/L of sodium with equimolar concentrations of glucose; this concentration is somewhat lower than what had been suggested in the past.175,176 Nevertheless, many commercially available oral rehydration formulas and sport drinks contain far lower sodium concentrations than what is necessary to optimize water and electrolyte absorption, and should be avoided by patients with short bowel syndrome. Instead, inexpensive and more effective solutions can be made by patients at home (Table 4-26). Daily oral administration of 1 to 2 L of rehydration solutions has been successful in correcting fluid

Table 4-26  Characteristics of Selected Oral Rehydration Solutions* Parameter Product CeraLyte 70 CeraLyte 90 Pedialyte Gatorade Thirst Quencher Gatorade Endurance WHO (2002 version)† Washington University‡ *

Na (mEq/L)

K (mEq/L)

Cl (mEq/L)

Citrate (mEq/L)

Calories (kcal/L)

CHO (g/L)

Osmolarity (mOsm/L)

70 90 45 20 36 75 105

20 20 20 3 10 20 0

60 80 35 11 11 65 100

30 30 30 NA NA 30 10

165 165 100 208 208 ∼51 85

40 40 20 58 58 14 20

235 260 300 NA NA 245 250

Mix solutions with sugar-free flavorings as needed for palatability. WHO (World Health Organization) formula: Each sachet contains 13.5 g glucose, 2.6 g NaCl, 2.9 g trisodium citrate dehydrate, 1.5 g KCl, to be dissolved in 1 L of water. Washington University formula: Mix 34 tsp sodium chloride, 12 tsp sodium citrate, and 3 tbsp + 1 tsp Polycose powder (glucose polymers) in 4 1 4 cups of distilled water. CHO, carbohydrate; N/A, information not available. † ‡

73

74

Section II  Nutrition in Gastroenterology and electrolyte abnormalities and allows intravenous supplementation to be discontinued in patients who have had extensive intestinal resection.177 Major Minerals Calcium and magnesium bind to unabsorbed FAs and may be lost in large quantities in patients with chronic steatorrhea, thus contributing to the frequent occurrence of metabolic bone disease in these individuals. The amount of calcium lost is proportional to the degree of steatorrhea. Magnesium is an integral component of bone mineral and is necessary for the release of parathyroid hormone (PTH) and for its actions on target organs.178,179 Hypocalcemia refractory to aggressive calcium replacement can be caused by chronic hypomagnesemia. Like calcium, total body magnesium status is not well reflected in serum levels of the cation. Consequently, the percentage of magnesium excreted in urine after intravenous infusion may be the best index of body magnesium stores in patients with short bowel syndrome; excretion of less than 80% of infused magnesium suggests whole-body magnesium depletion.180 Maintaining magnesium homeostasis often is difficult because magnesium is poorly absorbed and enteral supplementation with magnesium salts increases diarrhea. Entericcoated magnesium supplements are not used in the setting of intestinal malabsorption because their delayed release reduces contact with the intestine for absorption.181 Soluble magnesium salts, such as magnesium gluconate, are tolerated and absorbed better than other magnesium complexes. In some patients, magnesium is best given in liquid form as magnesium gluconate (Fleming, St. Louis) and can be added to an oral rehydration solution in doses of 18 to 27  mmol (432 to 648 mg of elemental magnesium) daily. This solution should be sipped, not ingested as a bolus, to maximize absorption and avoid diarrhea. In patients with steatorrhea and an intact colon, supplemental calcium should be given routinely to improve calcium availability and reduce excessive oxalate absorption. Plasma levels of calcium usually are maintained by mobilizing bone stores unless there is concurrent magnesium or vitamin D deficiency; urinary calcium excretion, which should be greater than 50 mg in 24 hours, is a more reliable index of calcium status. Most patients with fat malabsorption require approximately 1.5 to 2  g of elemental calcium daily. Trace Minerals With the exceptions of zinc and iron, absorption of trace minerals from ingested foods or liquid formulas often is adequate to prevent overt deficiency syndromes. In patients with malabsorption, zinc deficiency is common and often subclinical. Large dosages of oral zinc supplements may become necessary, because zinc losses are often high and zinc absorption is low. Zinc gluconate is tolerated well and is less likely than zinc sulfate to cause gastric distress. Zinc should not be given with meals because absorption is reduced by several foods, in particular those high in phytates.182 Daily zinc supplementation of 25  mg plus an additional 100 mg/L (or 100 mg/kg) of ostomy or diarrheal output is necessary to maintain zinc homeostasis.25 Thus, many patients require approximately 150  mg of elemental zinc daily. Although zinc ingestion reduces copper absorption and can cause clinically significant copper deficiency,183 additional copper intake usually is not needed. Treating iron deficiency with oral preparations can be difficult. A liquid form of ferrous sulfate (300  mg/5  mL containing 60 mg of elemental iron) mixed in orange juice four times daily has been found to be efficacious. Diluting

ferrous sulfate liquid prevents staining of teeth, and the ascorbic acid present in orange juice enhances iron absorption. Some patients, however, require intermittent administration of parenteral iron, which should be administered in the form of iron sucrose because it is far less likely to produce an anaphylactoid reaction compared with iron dextran.184 Vitamins Patients with fat malabsorption usually can absorb adequate amounts of most water-soluble vitamins from their diet, but have difficulty absorbing fat-soluble vitamins. Vitamin K deficiency rarely is a clinical problem unless patients are receiving antibiotics because colonic bacteria synthesize usable forms of vitamin K (vitamin K2; multiprenyl menaquinones). Large doses of vitamins A, D, and E may be required, however, to maintain normal concentrations of these vitamins. If provision of more than 5 to 10 times the basal requirement of a fat-soluble vitamin still is not effective, it is worthwhile using the more expensive polar forms of fat-soluble vitamins (see discussion earlier, in “Micronutrients” section). Testing blood levels of these vitamins should be used to determine whether repletion has been achieved.

Parenteral Feeding

Just because a patient requires parenteral support for fluid and electrolytes does not necessarily warrant inclusion of all the other nutrients in the intravenous solution. Some patients may be able to meet their other nutritional needs through an oral or enteral route. Long-term parenteral support with fluids and electrolytes (and sometimes with minerals and vitamins) alone is fraught with fewer infectious and metabolic complications than if macronutrients are included in the parenteral solution. Some general guidelines are useful in deciding which patients require parenteral therapy. Although a postduodenal intestinal remnant of less than 100 cm, or the lack of an ileocecal valve or a colonic segment in continuity, are each associated with a much greater likelihood of permanent dependence on parenteral support,185 exceptions abound. Adults in whom urine output is less than 1  L/day are at increased risk for developing renal dysfunction and should receive intravenous fluids. Adequate levels of fat-soluble vitamins and certain minerals (e.g., magnesium, potassium, and zinc) are difficult to maintain with oral feedings in patients with severe steatorrhea or large intestinal fluid output, and they may require parenteral supplementation. Magnesium sulfate can be injected intramuscularly at a dose of 12 mmol (290 mg of elemental magnesium) one to three times/week if attempts at oral therapy are unsuccessful. Intravenous infusion of magnesium is preferred, however, because intramuscular injections are painful and can cause sterile abscesses. In some patients, TPN may be life-saving or may be needed to limit diarrhea and achieve an acceptable QOL.

Nutritional Rehabilitation of the Short Bowel

Whether administration of growth hormone (or, more recently, GLP-2) to patients with short bowel syndrome will enhance the absorption of macronutrients and electrolytes continues to be a matter of controversy,186,187 but pharmacologic management is not the focus of this chapter (see Chapter 103). More relevant to this discussion is that some of these studies have included oral glutamine (at a dosage of approximately 30 g/day) as part of the growth hormone regimen, the rationale being that glutamine is a preferred fuel substrate of the enterocyte. There is, however, no con-

Chapter 4  Nutritional Assessment and Management of the Malnourished Patient trolled evidence that oral glutamine contributes to recovery of absorptive function.186-189

KEY REFERENCES

Duggan C, Fontaine O, Pierce N, et al. Scientific rationale for a change in the composition of oral rehydration solution. JAMA 2004; 291: 2628-31. (Ref 176.) Hamilton C, Seidner DL. Metabolic bone disease and parenteral nutrition: Curr Gastroenterol Rep 2004; 6:335-41. (Ref 152.) Heyland D, MacDonald S, Keefe L, et al. TPN in the critically ill patient: A meta-analysis. JAMA 1998; 16:2013-19. (Ref 72.) Hill G. Body composition research: implications for the practice of clinical nutrition. JPEN J Parenter Enteral Nutr 1992; 16:197-218. (Ref 54.) Hoffer LJ. Metabolic consequences of starvation. In: Shils M, Shike M, Ross AC, et al, editors. Modern nutrition in health and disease. 10th ed. Baltimore: Lippincott Williams & Wilkins; 2006. p 730. (Ref 44.) Klein S. The myth of serum albumin as a measure of nutritional status. Gastroenterology 1990; 99:1845-6. (Ref 110.) Mason JB, Roubenoff R. The unbearable lightness of being … a cirrhotic. Gastroenterology 1993; 105:1911-14. (Ref 133.)

Prijatmoko D, Strauss B, Lambert J, et al. Early detection of protein depletion in alcoholic cirrhosis: role of body composition analysis. Gastroenterology 1993; 105:1839-45. (Ref 65.) Stanga Z, Brunner A, Leuenberger M, et al. Nutrition in clinical practice—the refeeding syndrome: Illustrative cases and guidelines for prevention and treatment. Eur J Clin Nutr 2008; 62:687-94. (Ref 136.) van den Berghe G, Wilmer A, Hermans G, et al. Intensive insulin therapy in the medical ICU. N Engl J Med 2006; 354:449-61. (Ref 8.) van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in the critically ill patients. N Engl J Med 2001; 345:1359-67. (Ref 7.) Wiener R, Wiener D, Larson R. Benefits and risks of tight glucose control in critically ill adults: A meta-analysis. JAMA 2008; 300: 933-44. (Ref 11.) Wischmeyer PE. Glutamine: Role in critical illness and ongoing clinical trials. Curr Opinion Gastroenterol 2008; 24:190-7. (Ref 12.) Wolfe RR. Regulation of skeletal muscle protein metabolism in catabolic states. Curr Opin Clin Nutr Metab Care 2005; 8:61-5. (Ref 51.) Wolman SL, Anderson GH, Marliss EB, Jeejeebhoy KN. Zinc in total parenteral nutrition: requirements and metabolic effects. Gastroen­ terology 1979; 76:458-67. (Ref 25.) Full references for this chapter can be found on www.expertconsult.com.

75

CHAPTE R

5

Nutrition in Gastrointestinal Diseases Mark H. DeLegge

CHAPTER OUTLINE Nutritional Assessment  77 Medical History and Physical Examination  77 Anthropomorphic Measurements  77 Biochemical Measurements  77 Immunologic Tests  78 Muscle Function  78 Global Assessments  78 Research and Investigational Tools  78 Caloric Assessment  78 Protein Assessment  79 Nutrient Substrates  79 Macronutrients  79 Macrominerals  80 Micronutrients  80 Vitamins  81

It has long been known that nutritional status affects clinical outcome. In lean healthy people, death has been associated with weight loss of more than 35%, protein storage loss of more than 30% and fat storage loss of more than 70%. Clinicians spend a considerable amount of time treating patients but give little attention to the signs and symptoms of nutritional inadequacies or imbalances. As practicing clinical gastroenterologists, we have a responsibility to understand the science of nutrition and to become familiar with the nutrition literature. The literature detailing nutrition as medical therapy is controversial. Overall, there is a lack of large, randomized, prospective studies comparing one nutritional therapy with another; thus, meta-analyses often are used to group small studies and allow reasonable conclusions.

NUTRITIONAL ASSESSMENT Determining who is at risk for malnutrition is a complicated science. A proper nutritional assessment provides a mechanism whereby patients who may require nutritional support can be identified and also provides a gauge to monitor the effectiveness of such support (see Chapter 4).

MEDICAL HISTORY AND PHYSICAL EXAMINATION

Reading a patient’s chart, taking a history, and performing a physical examination allows a good understanding of

Nutrition in Specific Disease States  82 Intestinal Failure (Short Bowel Syndrome)  82 Pancreatitis  83 Crohn’s Disease  83 Liver Disease  83 Diverticular Disease  83 Dumping Syndrome  84 Celiac Disease  84 Cancer  84 Obesity  84 Nutritional Therapy  85 Parenteral Nutrition  85 Enteral Nutrition  87 Oral Diet Therapy  95 Probiotics  96

a patient’s disease status and allows the diagnosis of some nutritional deficiencies. Inquiries regarding a patient’s usual body weight (UBW) versus ideal (IBW) or present body weight (PBW) should be noted at the initial patient encounter because these values have been shown in a number of studies to be predictors of morbidity and mor­tality.1,2 Although a simple tool, the most sensitive marker of nutritional risk is the percentage deviation from UBW over the past three to six months.

ANTHROPOMORPHIC MEASUREMENTS

Anthropomorphic measurements allow the estimation of body composition or body stores of energy using relatively simple and inexpensive equipment such as hand-held calipers (Fig. 5-1) and scales. Hand-held calipers allow measurement of a patient’s triceps skinfold (TSF), a marker of body fat stores, and midarm muscle circumference (MAMC), a marker of body protein stores. These measurements are compared with standardized tables to obtain percentages of normal values. Loss of body stores of protein (e.g., reduced MAMC) has been associated with poor patient outcomes. Because of the obesity epidemic, emphasis has been placed on evaluating a patient’s weight and body habitus using the body mass index (BMI), which is defined as weight (in kilograms)/height (in meters squared). BMI values help categorize patients as underweight, normal weight, overweight, and obese (see Table 4-18).

BIOCHEMICAL MEASUREMENTS

The plasma proteins albumin, prealbumin, and transferrin are widely used to assess nutritional status. Albumin has

77

78

Section II  Nutrition in Gastroenterology

Figure 5-1.  Hand-held calipers. This device helps determine body stores of protein or fat.

been overrated as a predictor of protein malnutrition and, because it has a half-life of 21 days, it is a poor indicator of protein malnutrition. Infections, chronic medical conditions, liver disease, acute physiologic changes, and medications affect serum albumin levels through mechanisms not related to protein malnutrition.3 Low serum albumin levels (<3.5 g/dL) on hospital admission have been correlated with poor prognosis and poor surgical outcome.4 Prealbumin has been thought to be a better laboratory marker of nutritional status because of its two-day half-life. Prealbumin often is used by nutrition support teams to assess the adequacy of ongoing nutrition therapy; however, the reliability of prealbumin as a marker of malnutrition has not been validated. Transferrin also has been used to evaluate a patient’s nutritional status because it resides almost completely within the intravascular space and has a half-life of seven days. It is, however, affected by iron deficiency, iron overload, pernicious anemia, and infection. There are no data validating its use as a marker of nutritional status.

IMMUNOLOGIC TESTS

The serum total lymphocyte count has been correlated with changes in nutritional status, especially protein depletion.5 However, no prospective study has shown this to be a reliable marker of a patient’s nutritional status.

MUSCLE FUNCTION

An effective tool to evaluate a patient’s nutritional status over time is to measure his or her endurance or muscle strength. A practical method of assessing muscle strength involves hand grip strength,6 which is proportional to forearm lean muscle mass (Fig. 5-2). The reliability of this test is reduced for patients who are acutely ill or have hand or arm motor abnormalities.

GLOBAL ASSESSMENTS

To date, there is no single tool that is an accurate predictor of nutritional status. A variety of validated nutrition scoring systems, incorporating a number of features of the patient’s medical history and physical examination, has been developed to assess nutritional status. One of these, the subjective global assessment (SGA), is a comprehensive nutritional assessment tool that incorporates weight changes, dietary intake, functional capacity and preliminary medical diagnosis to categorize patients as well-nourished, moderately malnourished, or severely malnourished (see Table 4-24).7 The SGA has been validated in the oncology population.8 Other global nutritional assessment tools include the prognostic nutritional index (PNI), the instant nutritional

Figure 5-2.  Hand-grip dynamometer. This device measures muscle strength and therefore body protein stores.

assessment (INA), the Mini-Nutritional Assessment, the malnutrition universal screening tool, the nutritional risk score, and the nutritional risk index.9

RESEARCH AND INVESTIGATIONAL TOOLS

Densitometry, isotope measurements, electric impedance testing, infrared refractance, and absorptiometry all provide very accurate assessment of body composition, but many of these tests are limited by cost or safety considerations. These tools may prove practical in the future to help assess a patient’s nutritional status.

CALORIC ASSESSMENT Mathematical Equations

Calculation of energy requirements can be obtained through mathematical equations. Many different methods of estimating energy needs have been used over the years, including estimations based on body surface area, body weight, body height, and age of the patient. The most commonly used equation for calculating energy needs is the Harris-Benedict equation.10 The calculation is as follows: Men: Energy needs ( kcal 24 hr ) = 66 + (13.7 × W ) (5 × L) − (6.8 × A ) Women: Energy needs ( kcal 24 hr ) = 655 + (9.6 × W ) + (1.7 × L) − ( 4 .7 × A )

where W = weight in kilograms, A = age, and L = height in centimeters. This calculation of energy needs is multiplied by stress factors to arrive at a patient’s overall caloric needs (Table 5-1) Other caloric assessment formulas commonly used include the Ireton-Jones equation and the Mifflin-St. Jeor formula.11 Calculated energy needs, however, may over- or underestimate a patient’s true energy needs, especially in patients with complicated disease processes that can alter their metabolic rate.12 In these cases, direct measurements of overall energy needs may be more appropriate.

Chapter 5  Nutrition in Gastrointestinal Diseases excretion. To determine a patient’s daily protein needs, the following formula may be used:

Table 5-1  Influence of Disease Severity (Physiologic Stress) on Resting Energy Expenditure (REE) Disease Severity Mild Moderate Severe

REE (%) 10 25 50

Example: REE of 1500 kcal in a patient with moderate disease = (1500 kcal × 0.25) + 1500 kcal = 1875 kcal/day.

Total daily protein needs ( g ) = (24-hour UUN [ g ] × 6.25) + 3 g protein (insensible losses)

NUTRIENT SUBSTRATES MACRONUTRIENTS

Figure 5-3.  MedGem calorimetry device for energy measurement.

Indirect Calorimetry

Direct measurements of a patient’s energy needs can be performed using indirect calorimetry, which measures heat produced by oxidation. A ventilated hood is placed over the resting patient’s head and its oxygen and carbon dioxide content for two hours is analyzed. From this information, the resting energy expenditure (REE) is derived. True caloric needs are calculated by multiplying the REE by an activity or stress factor. In addition, a patient’s respiratory quotient (RQ) is derived. The RQ is equal to Vco2 (volume of CO2)/Vo2 (volume of O2). An RQ of approximately 0.7 or less is an indication of underfeeding whereas an RQ of approximately 1.0 or more is an indication of overfeeding.13 Indirect calorimetry plays a large role in obese or very malnourished patients, for whom determining caloric needs based on total body weight can be inaccurate.14 More recently, simple hand-held devices have been developed with an oxygen sensor for energy measurement. By breathing through the device for 10 minutes, the patient’s REE can be determined (Fig. 5-3).15 This tool has been shown to be as effective as the more expensive and laborious indirect calorimetry device in patients who can spontaneously breathe on their own volition, that is, those who are not on a ventilator.16

PROTEIN ASSESSMENT

Determining a patient’s daily protein needs is an important part of writing their nutrition prescription, especially for patients on enteral nutrition (EN) or parenteral nutrition (PN). Protein catabolism occurs at varying rates and is affected by a patient’s disease status, current nutritional state, and diet. Stressed patients may use 15 to 30 g of nitrogen/day. Total body protein needs can be calculated by assessing a patient’s 24-hour urinary urea nitrogen (UUN)

The macronutrients that fuel human metabolism include carbohydrates, fats, and proteins. The major source of energy in the human diet is carbohydrate, which constitutes almost half of the typical American diet. Carbohydrate is consumed as starch, sucrose, or lactose. Starch is made up of the polysaccharides amylopectin and amylose. Sucrose and lactose are disaccharides. Most digestion occurs in the duodenum and small intestine (see Chapter 100).17 Starch digestion begins in the mouth with the enzyme amylase and continues with the pancreatic enzyme alpha-amylase in the small intestine. Starch is broken down into oligosaccharides. The oligosaccharides and disaccharides are hydrolyzed by the small intestinal brush border to monosaccharides by glucoamylase, sucrase, and amylase. Glucose and galactose are absorbed across the small intestinal mucosa by active transport, whereas fructose moves across by facilitated diffusion.18 Dietary fat is primarily composed of triglycerides, which consist mainly of four long-chain fatty acids—palmitic, stearic, oleic, and linoleic—with smaller amounts of lino­ lenic acid and medium-chain triglycerides (MCTs). Linoleic and linolenic acids are essential fatty acids; that is, they cannot be synthesized from nondietary sources. Essential fatty acid deficiency can result in a clinical syndrome consisting primarily of a scaly erythematous rash.19 Clinically, this may be seen in patients completely dependent on PN for their nutrient intake who receive no fat in their daily PN solution over a period of one to two weeks. Triglycerides are hydrolyzed to free fatty acids and beta-monoglycerides by pancreatic lipase and colipase.20 Because fats are insoluble in water, their digestion requires a unique environment involving an emulsification process whereby bile salts enhance their absorption (see Chapter 100). In water, bile salts form a micelle with a hydrophobic core and a hydrophilic periphery. Micellar contents diffuse across the water layer, intestinal mucosa, and cell membrane and enter the cell. Here they are reesterified to triglycerides and linked to an apoprotein to form a chylomicron. MCTs have a different mechanism of absorption; 30% are absorbed directly as intact triglycerides and the remaining 70% are digested to medium-chain fatty acids, which are soluble in water and directly absorbed by the small intestinal mucosa. MCT-based diets may be helpful for patients with fat malabsorption from pancreatic or liver disease. A diet very high in MCT and low in long-chain triglycerides (LCTs), however, can lead to essential fatty acid deficiency. These patients present with a characteristic diffuse, scaly, erythematous skin rash. Fats provide a long-term energy source for patients who are experiencing starvation. Unfortunately, critically ill catabolic patients lose their ability to use fat for energy efficiently, making them more dependent on their carbohydrate and protein stores. Protein and amino acid metabolism are essential to provide the building blocks necessary to create a variety of body proteins and nitrogen-containing compounds. Proteins are composed of amino acids joined by peptide bonds.

79

80

Section II  Nutrition in Gastroenterology There are 22 amino acids, eight of which are essential: lysine, threonine, leucine, isoleucine, valine, methionine, phenylalanine, and tryptophan; two others, histidine and arginine, are essential for infants and growing children. Dietary proteins are partially digested by pepsin in the stomach to form polypeptides, but most protein digestion is performed in the duodenum and upper jejunum by pancreatic proteases, including trypsin, chymotrypsin, carboxypeptidase, and elastase (see Chapter 100).21 Peptides are further hydrolyzed in the small intestine by aminopeptidases, enzymes found in the brush border of the small intestine, to free amino acids and very small peptides. Large peptides must be hydrolyzed by brush border enzymes for absorption to occur, whereas dipeptides and tripeptides can move intact into mucosal cells.22

MACROMINERALS

Minerals account for only 4% of total body weight, yet they serve as essential cofactors, help maintain fluid osmotic pressures, and provide the proper environment for many chemical reactions (see Chapter 4). Calcium.  This is the most abundant cation in the human body. Bone and teeth contain about 99% of total body calcium. It is ingested in the form of an insoluble salt and must be released from its salt form to its ionized form for absorption (see Chapter 100).23 Calcium absorption occurs along the length of the small intestine and is vitamin D–dependent. In periods of restricted calcium intake, the colon may become more involved in calcium homeostasis by increasing its absorption. In periods of low serum calcium levels, renal excretion of calcium is reduced and bone calcium stores are released. Magnesium and calcium compete with one another for absorption. Unabsorbed dietary fat can interfere with calcium absorption by the formation of soaps in the intestine, which are excreted in the feces. The recommended daily intake of calcium is 800  mg/day for adults and 1200  mg/day for growing children. Calcium requirements are higher for pregnant women and older adults. In addition to bone mineralization, calcium also is important in blood clotting, muscle contraction, and the secretory activity of most endocrine and exocrine cells. Hypocalcemia may result in tetany, paresthesias, hyperreflexia, seizures, and mental status changes. Chronic calcium deficiency will result in rickets in children and osteomalacia in adults. Phosphorus.  This is the second most abundant inorganic element in humans and is stored in bone (85%) and soft tissue. Animal-derived phosphorus is in the form of sodium phosphate and is rapidly absorbed. Cereals and grains contain phosphorus in the form of phytic acid. Phosphorus absorption occurs through the small intestine. The body adapts to low phosphate ingestion by simply reducing urinary excretion losses. Excess dietary magnesium can interfere with phosphorus absorption by forming magnesium-phosphorus complexes within the gastrointestinal tract. Phosphorus deficiency may occur in patients who have been chronically starved and suddenly refed. In this circumstance, the delivery of a glucose load causes a rise in the serum insulin level, which drives serum glucose into metabolically active cells, dragging phosphorus, magnesium, and potassium into the cells and leading to low serum levels of these substances. This disorder is known as refeeding syndrome. The recommended daily intake of phosphorus is 800 to 1200  mg/day. Phosphorus is important in the maintenance and development of skeletal tissue and regulation of body pH. It is an important component of

Table 5-2  Daily Trace Element Requirements Requirement Trace Element

Enteral

Parenteral

Chromium Copper Fluoride Iodine Iron Manganese Molybdenum Selenium Zinc

30 µg 0.9 mg 4 mg 150 µg 18 mg 2.3 mg 45 µg 55 µg 11 mg

10-15 µg 0.3-0.5 mg Not well defined Not well defined Not routinely added 60-100 µg Not routinely added 20-60 µg 2.5-5 mg

From American Society for Parenteral and Enteral Nutrition. Guidelines for the use of parenteral and enteral nutrition in adults and pediatric patients. J Parenter Enteral Nutr 2002; 26:29SA-33SA.

nucleic acids and energy compounds such as adenosine triphosphate (ATP). Phosphate deficiency can result in muscle weakness, paresthesias, seizures, hemolytic anemia, impaired white blood cell function, and tissue hypoxia secondary to shifting of the oxygen-hemoglobin dissociation curve. Magnesium.  This is the second most abundant intracellular cation. Approximately 60% of magnesium is located in bone.24 Skeletal muscle also serves as another large source of magnesium storage. Magnesium absorption increases when magnesium intake is low. Vitamin D may affect absorption, although this relationship is not clear. Patients on a low-protein diet also can have difficulty with magnesium absorption. The recommended daily intake of magnesium is 300 to 500  mg/day. Magnesium is important in providing stability to the structure of ATP and is involved in numerous other enzyme systems. Magnesium deficiency can result in tetany, ataxia, myoclonus, coma, psychosis, cardiac dysrhythmias, and hypotension. Severe hypomagnesemia may be seen in patients with refeeding syndrome.

MICRONUTRIENTS

Essential micronutrients are present in minute or trace amounts in the body, sometimes in quantities less than 100 µg (see Chapter 4). Although trace elements are present in very small amounts, they often have dramatic effects; deficiencies are more common than toxicity. Deficiencies can result from reduced intake, decreased bioavailability, decreased transport proteins, excess excretion, or as the result of certain disease states. Many of these deficiencies develop in patients who are on long-term PN or who are severely malnourished. Assessment of trace element deficiency is extremely difficult (Table 5-2) because serum levels may not accurately reflect body stores. Therefore, clinicians may have to depend largely on physical signs and symptoms to detect micronutrient deficiency. Chromium.  This is important in protein, carbohydrate, and lipid metabolism by serving as an important cofactor in enzymatic breakdown. It is crucial for the synthesis of glucose tolerance factor, a cofactor in insulin action.25 Chromium deficiency is manifested by glucose intolerance. The daily requirement is 50 to 100 µg/day. Copper.  This is important for normal body iron uptake. A microcytic hypochromic anemia can be seen with copper deficiency,26 a result of shortened red blood cell lifespan.

Chapter 5  Nutrition in Gastrointestinal Diseases Copper plays a major role in taste sensation and is important in reducing the potential injurious effects of free radicals. Copper is excreted in bile and should be replaced in those with external biliary drains or excessive diarrhea. The daily requirement is approximately 1.5 to 3 µg/day. Iodine.  This is important in the cellular oxidative processes associated with thyroid function. Iodine deficiency is the most widespread micronutrient deficiency. Deficiency results in weakness, cold intolerance, facial swelling, pallor, thin hair, hoarseness, constipation, reduced sweating, and growth failure. The daily requirement is approximately 150 µg/day. Iron.  This is an oxygen delivery substrate. It is principally absorbed in the duodenum as ferrous iron, and acidification with vitamin C will increase the bioavailability of iron for absorption. Iron deficiency can result in a microcytic anemia as a consequence of reduced intake, blood loss, achlorhydria, or malabsorption. Toxicity results in hemosiderosis. The daily requirement is 1 mg/day. Only 10% of iron given orally is absorbed, so at least 10 mg of oral iron is required daily. Manganese.  This is important in the synthesis of protein, mucopolysaccharides, and prothrombin. Deficiency can lead to a thinning of hair and a lightening of its color. Manganese toxicity can result in neuropsychiatric symptoms. The daily requirement is 3 to 5 mg/day. Selenium.  This is involved in the oxidation of glutathione and degradation of peroxidases. It is excreted mainly in the urine, but substantial stool losses can occur in association with severe malabsorptive disease of the small intestine. The daily requirement is 70 µg for men and 55 µg/day for women. Deficiencies have been reported in patients on long-term PN. Symptoms include myositis, cardiomyopathy (Keshan’s syndrome), and collagen vascular disease (Keshan’s disease). Zinc.  This is an important component of many enzymes. It is involved in protein and lipid synthesis, and insulin activity. Approximately 25% of ingested zinc is absorbed daily in the duodenum and proximal jejunum. Excretion is through the biliary tract, skin, and feces. Zinc deficiency can result in a characteristic skin rash (acrodermatitis), poor wound healing, impaired taste, glucose intolerance, alopecia, depression, and diarrhea. Because body copper levels can be suppressed by zinc loading, zinc has been evaluated for the treatment of early Wilson disease.27 The daily requirement is approximately 10 to 15 mg/day.

VITAMINS

Vitamins (see Chapters 4, 100) are essential micronutrients involved in basic body functions such as growth, tissue maintenance, and metabolism. They are broadly classified as water-soluble or fat-soluble. Absorption of fat-soluble vitamins requires absorption and transport of lipids. Watersoluble vitamins, except vitamin C, are part of a B-complex group (Table 5-3).

Water-Soluble Vitamins

Vitamin C.  This is important in maintaining connective tissue by its effect on the hydroxylation of proline and lysine. It also plays a role in the conversion of tryptophan to serotonin, thus accounting for the fatigue, weakness, and vasomotor instability seen in scurvy. Scorbutic patients present with bleeding gums, joint pains, loosening of teeth,

Table 5-3  Daily Reference Intakes* and Sites of Absorption of Vitamins Vitamin

Recommended dietary allowance*

Absorption Sites

C B1 (thiamine) Riboflavin

75-90 mg 1.1-1.2 mg 1.1-1.3 mg

Niacin

14-16 mg

Pantothenic acid Biotin Folic acid B12 B6 A†

5 mg

Distal small intestine Jejunum Proximal small intestine Stomach, small intestine Jejunum

D

200-600 IU

E K

15 mg/day 90-120 µg/day

30 µg 400 µg 2-4 µg 1.3-1.7 µg 700-900 retinol equivalents

Unknown Jejunum Distal ileum Jejunum Proximal small intestine Duodenum, terminal ileum Mid–small intestine Jejunum, ileum, colon

*Daily Reference Intakes (DRI) were established by the Institute of Medicine between 1997-2001. They are quantitative estimates of nutrient intakes to be used for planning and assessing diets for healthy people. The DRIs include both recommended intakes and tolerable upper intake levels. The RDAs (Recommended Dietary Allowances) are a component of the DRIs and are defined as the daily intake of a nutrient considered sufficient to meet the requirements of 97% to 98% of adults. † A retinol equivalent is 3.3 IU of vitamin A; 1 retinol equivalent = 6 µg β-carotene or 1 µg retinol.

gingivitis, weakness, and hyperkeratosis of hair follicles. Body hair is described as having a corkscrew appearance. Vitamin C is found in many fruits and vegetables, especially oranges, grapefruit, and strawberries. Thiamine (Vitamin B1).  This is widely abundant in grain foods such as whole wheat bread and spaghetti and also in peas, peanuts, potatoes, sausage, and beef. It is important in the synthesis of nicotinamide adenine dinucleotide phosphate (NADPH), energy transformation, and membrane and nerve conduction. Thiamine deficiency can result in anorexia, anemia, ataxia, polyneuritis, beriberi, and Wernicke’s encephalopathy. Riboflavin.  This is found in milk and enriched breads. Absorption is facilitated by bile salts. It is active in its coenzyme derivative forms flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD), which serve as cofactors for a number of oxidative reactions. Deficiency of riboflavin may lead to soreness and burning of the lips, mouth, and tongue, along with photophobia. Riboflavindeficient patients often have a beefy red tongue. Niacin.  This is found in animal foods such as beef, pork, and chicken and in cereal grains, especially wheat, rice, and bran. Niacin is hydrolyzed to niacinamide in the mucosal cells of the small intestine. It is important for the formation of the nucleotides nicotinamide adenine dinucleotide, reduced form (NADH) and NADPH, compounds that serve in a number of electron transport systems. Niacin is used in the treatment of hypercholesterolemia, although the mechanism whereby it reduces cholesterol levels is unclear. Side effects of therapeutic doses of niacin include flushing, liver injury, elevated uric acid levels, and dermatologic

81

82

Section II  Nutrition in Gastroenterology problems.28 Niacin deficiency may result in a constellation of symptoms referred to as pellagra, which is characterized by glossitis, coarse, scaly erythematous skin, diarrhea, and mental confusion. Pantothenic Acid.  This is present in almost all plant and animal foods. Most of the pantothenic acid in food exists as a component of coenzyme A. It is essential for the production of energy from fat, carbohydrate, and protein and is important in wound healing. Biotin.  This is a B-complex vitamin widely distributed in foods and produced by intestinal bacteria; it is important in carboxylase reactions. Deficiency in humans is extraordinarily rare. Folic Acid.  This is present in many foods including leafy vegetables, fruit, and liver. Polyglutamate forms of folate must be hydrolyzed by jejunal mucosal enzymes prior to absorption.29 Folate serves as a carbon donor in a number of synthetic reactions. The maturation of red blood cells and other short-lived cells is folate-dependent. Folate deficiency results in a macrocytic anemia because of decreased deoxyribonucleic acid (DNA) synthesis. Excessive folate ingestion may produce malaise, insomnia, and gastrointestinal distress. Vitamin B12.  This is an important cofactor for the adequate functioning of folate; it is also a cofactor in synthetic reactions. Vitamin B12 belongs to a group of compounds called corrinoids, which contain a cobalt atom in the center of the molecule, with various side chain attachments. Cyanocobalamin is the most common commercially available form. Animal products are the only source of vitamin B12 for humans. Ingested cobalamin is liberated from polypeptides in the stomach by the action of pepsin and becomes attached to R binders. In the small intestine, pancreatic enzymes cleave vitamin B12 from R binders and facilitate the attachment of B12 to intrinsic factor (IF). Intrinsic factor attaches to specialized ileal receptors allowing vitamin B12 to be absorbed. Vitamin B12 is excreted in bile. Malabsorption accounts for most cases of B12 deficiencies in the United States and may be the result of bacterial overgrowth or impaired absorption of the vitamin B12–IF complex because of distal ileal disease or loss. Deficiencies can present as a megaloblastic anemia or peripheral neuropathy. Vitamin B6 (Pyridoxine).  This is found in plant foods, especially bananas and navy beans. The coenzyme form of vitamin B6 is involved in a number of enzyme reactions, especially amino acid metabolism. Patients at risk for vitamin B6 deficiency include alcoholics, dialysis patients, and those on a number of medications, including isoniazid and glucocorticoids.

Fat-Soluble Vitamins

Vitamin A.  This is present in a number of pigmented plants. A complex absorption process involves small intestinal absorption, esterification, and uptake in the liver for attachment to retinal binding protein to permit delivery to target tissues. Vitamin A is essential for vision, cellular differentiation, and integrity of the immune system. Vitamin A is excreted primarily in the bile. Vitamin A toxicity can lead to anorexia, alopecia, birth defects, and liver damage. Vitamin D.  This is present primarily in animal foods such as eggs, liver, butter, and fortified milk. In the presence of

sunlight, vitamin D3 is produced by the skin from provitamin D3. Both the kidney and liver are important in the synthesis of the active form of vitamin D, 1,25-(OH)2D3. The primary function of vitamin D is to increase intestinal absorption of calcium and phosphorus. Renal absorption of calcium also is enhanced. Lack of vitamin D leads to bone resorption to maintain calcium homeostasis. Toxicity can lead to anorexia, nausea, vomiting, renal insufficiency, and failure to thrive. Vitamin E.  This comprises a group of eight compounds found in plants and known as tocols (saturated side chains) or tocotrienols (unsaturated side chains). Alpha-tocopherol provides 75% of the total vitamin E found in humans. The largest store of vitamin E is found in adipose tissue. The principal function of vitamin E is to maintain cellular membrane integrity by protecting it from the effects of oxidation. The major route of vitamin E excretion is the feces or through the skin. Deficiencies of vitamin E are rare and generally occur in the context of severe fat-soluble vitamin malabsorption, such as in short bowel syndrome, cystic fibrosis, pancreatic insufficiency, or advanced liver disease. Deficiencies can lead to neuromuscular dysfunction. Toxicity generally produces nausea, vomiting, and diarrhea. Vitamin K.  This is necessary for the synthesis of four of the 13 factors required for blood coagulation (II, VII, IX, and X). It also is important in the synthesis of four proteins involved in coagulation (C, Z, S, and M). It is found in plant and animal foods and is absorbed from the jejunum, ileum, and colon by an energy-dependent process or diffusion. Of the total ingested vitamin K, 8% is absorbed and its half-life is short, two to three hours. Bacteria in the digestive tract also can synthesize vitamin K. After liver conjugation, the excretion of vitamin K occurs in the bile, feces, and urine. The normal diet contains 300 to 500 mg of vitamin K/day; thus, deficiencies are rare, but can occur in patients with severe malnutrition, fat malabsorption, pancreatic insufficiency, cholestasis, or severe liver disease, or in patients receiving antibiotics. Toxicity is very rare and usually reported only in infants.30

NUTRITION IN SPECIFIC DISEASE STATES Nutritional assessment and directed nutritional therapy are important in the treatment of many gastrointestinal diseases, and familiarity with appropriate nutritional intervention is imperative to obtaining good clinical outcomes.

INTESTINAL FAILURE (SHORT BOWEL SYNDROME)

Intestinal failure or short bowel syndrome results from loss or disease of the intestine, or both, to an extent that precludes adequate digestion and absorption (see Chapter 103); Crohn’s disease (see Chapter 111), intestinal trauma, and intestinal infarction (see Chapter 114) are the most common causes. The patient often presents with weight loss, diarrhea, dehydration, and weakness. Following extensive resection of the small intestine, intestinal rehabilitation (goal of resuming oral nutrition) of the remaining small intestine is more likely to be successful if the colon has been preserved and the ileocecal valve is maintained.31 The nutritional management of short bowel syndrome depends on the amount and location of small bowel removed. Initially, proton pump inhibitors are used to

Chapter 5  Nutrition in Gastrointestinal Diseases reduce gastric hypersecretion and anticholinergic agents are used to slow intestinal transit. Parenteral nutrition is prescribed to meet nutritional needs. Oral feedings are gradually started and the volume of PN is reduced as oral feedings are tolerated. If the patient has had a partial ileal resection and has an intact colon, cholestyramine can be used to reduce bile salt–induced diarrhea. In patients with a small amount of ileum remaining and an intact colon, however, the use of cholestyramine can increase diarrhea by creating a relative bile salt deficiency. Vitamin B12 should be given monthly. In patients with significant small bowel resections (80 to 100  cm remaining), a trial of a smallpeptide, low-fat, enteral formula should be attempted to reduce the amount of PN the patient requires. Later, a polymeric enteral formula can be substituted. Patients with less than 80 cm of small intestine remaining and no colon often are PN-dependent. The use of somatostatin to reduce intestinal secretions and slow transit time remains controversial.32 Anticholinergic therapy should be initiated. Patients may require larger doses of anticholinergics than usually are recommended, because absorption of the oral medication is limited. The use of growth hormone, glutamine, and a rice-based diet in an attempt to cause small bowel mucosal hypertrophy and better absorption is controversial. Early data have suggested very significant improvements in small intestinal absorptive function,33 not confirmed by subsequent studies.34 The use of a glycoprotein (GL-2) also has been postulated as a small intestine mucosal stimulator for improved absorption. A recent prospective evaluation of its efficacy noted a statistically significant reduction in PN use35; the effectiveness of this therapy is still being evaluated in additional studies.

PANCREATITIS

Nutritional support is imperative for patients with severe acute pancreatitis or relapsing chronic pancreatitis (see Chapters 58 and 59). Early PN appears to be associated with a reduction in the complications and mortality associated with acute pancreatitis compared with maintaining the patient on an NPO regimen.36 However, central line catheter sepsis rates are high and hyperglycemia is common. Enteral nutrition also has been used in patients with pancreatitis in contrast to previous beliefs that complete bowel rest was required. It appears that intrajejunal feedings are safe and well tolerated.37 A standard, fat-containing, polymeric enteral formula can be used.38 Randomized, prospective trials have shown a reduction in overall patient complications, hospital length of stay, and total hospital charges compared with the use of PN.39,40 The use of jejunal feeding in patients with chronic pancreatitis has been described to improve weight and reduce abdominal pain associated with eating.41 Gastric feedings also have been used successfully in patients with severe acute pancreatitis,42 but are still a topic of investigation.

CROHN’S DISEASE

Crohn’s disease (see Chapter 111) is sometimes associated with malnutrition.43 These patients often are hypermetabolic and may have anorexia because of nausea and abdo­ minal pain. Deficiencies of magnesium, selenium, potassium, and zinc are common in inflammatory bowel disease (IBD) as a result of losses in diarrheal fluids and through fistula tracts.44 Dietary therapy in IBD always has been considered important, but no specific diet can be recommended. Fat restriction may be important in patients with ileal disease or those who have undergone an ileal resection. The use of EN can be an important component of IBD therapy for

patients who cannot eat. Enteral nutrition has not proven superior to PN in inducing remissions in IBD,45 although it is less costly and associated with fewer complications. Enteral nutrition alone has not proven superior to drug therapy for the treatment of Crohn’s disease.46 The use of PN in IBD should be restricted to patients who have not responded to conservative medical therapy (EN and medications) or in whom EN cannot be delivered.

LIVER DISEASE

Nutritional deficiencies are common among patients with liver disease, mainly from decreased dietary intake, but also as a result of altered metabolism, decreased nutrient storage, and increased nutrient requirements (see Chapters 72, 92, and 93). Decreased nutrient intake is secondary to anorexia and nausea and is more common in patients with cirrhosis.47 Decreased bile salt production results in an intolerance to high-fat foods and the development of fat-soluble vitamin malabsorption. In addition, hypoalbuminemia results in edema of the mucosa of the small intestine, further compromising nutrient absorption. Depletion of muscle mass occurs secondary to a lack of adequate glucose stores and a dependency on protein stores for energy. Normal serum amino acid concentrations are altered, with a rise in aromatic amino acids (tyrosine, phenylalanine, and methionine) and a fall in branched-chain amino acids (valine, leucine, and isoleucine). The aromatic amino acids normally are removed by the liver; it is postulated that the rise in aromatic amino acids precipitates hepatic encephalopathy because they act as false neurotransmitters. Moreover, branched-chain amino acids are used preferentially as a protein source by patients in liver failure because they require minimal liver catabolism. Unfortunately, studies have failed to demonstrate improved outcomes in liverfailure patients fed a branched-chain amino acid–fortified diet or enteral solution.48 There is a general tendency to limit protein intake in patients with cirrhosis to prevent encephalopathy; however, these patients have an increased protein demand and further limiting their protein intake will only accelerate protein calorie malnutrition. It is preferable to feed patients according to their protein needs and treat encephalopathy with medications as it develops. Parenteral nutrition should be used with caution in patients with liver failure, because immune dysfunction places these patients at increased risk for catheter sepsis. In addition, the lack of liver glycogen stores can lead to episodes of hypoglycemia when patients are rapidly tapered off PN. Nutritional support prior to liver transplantation has been shown to improve patient outcome, especially in patients who are significantly malnourished prior to the transplantation.49

DIVERTICULAR DISEASE

Patients with diverticular disease (see Chapter 117) often are provided with incorrect nutritional information. They are told to avoid nuts or foods that contain seeds because of fear that the hard small particles may lodge in diverticula and precipitate diverticulitis. There are no clinical data to support this concept,50 and most data suggest that a highfiber diet will reduce the occurrence of symptomatic diverticular disease.51 Patients hospitalized for complicated diverticular disease can remain symptom-free on a highfiber diet.52 Fiber intake should be at least 25  g/day and should be provided as insoluble fiber, such as that contained in wheat bran, bran muffins, and fiber-based cereals. The use of probiotics has had some success in the treatment of and prevention of diverticulitis, although more work needs to be done in this area.53

83

84

Section II  Nutrition in Gastroenterology DUMPING SYNDROME

needs for a prolonged period. It is in these patients that an improved quality of life may occur.

Dumping syndrome is common after partial gastrectomy and vagotomy. Hypertonic gastric contents empty rapidly into the small intestine, and consequently up to 25% of plasma volume suddenly may be transferred to the small intestine.54 Nausea, cramping, diaphoresis, and palpitations develop. Nutritional therapy for dumping syndrome aims to deliver a lower osmolarity solution to the small intestine by the frequent ingestion of small meals containing fat, protein, and complex carbohydrates, but limited in simple sugars. Fluid intake should be restricted and separated from solid food intake to avoid rapid gastric transit. High pectin-containing foods (bananas, oranges) will slow gastric output.

OBESITY

Obesity is a disease that has had explosive growth, simultaneously affecting many individuals in a community without regard to age, gender, or ethnic origin (see Chapter 6).59 Traditionally, the gastroenterologist has been involved in the endoscopic management of postbariatric surgical complications, including stomal stenosis, gastrointestinal bleeding, and fistulization (see Chapter 7). Previous endoscopic technologies used to treat obesity endoscopically, such as the gastric balloon, had limited exposure in the United States and were removed from the market because of associated complications, such as balloon deflation with migration and resultant small intestinal obstruction. Internationally, however, balloon endoscopic therapy continues, with a number of favorable outcomes reported.60,61 Patients with obesity-related gastrointestinal disease, such as gastroeso­phageal reflux disease, always have been a staple of the gastroenterologist’s practice. Today, gastroenterologists find themselves on the front line with other physicians attempting to combat obesity and its associated complications, including assessment of patients for risk, identification of those who may benefit from weight loss therapy, and determination of the appropriate weight loss intervention, including management of therapy-associated complications. Assessing patients with obesity generally takes into consideration the BMI, waist circumference, and comorbid disease states. A person with a BMI more than 30 kg/m2 is considered obese. Patients who are overweight (BMI, 25 to 30 kg/m2) with comorbid diseases should be offered dietary management, exercise, and behavioral therapy. Patients with a BMI more than 30  kg/m2 also should be offered medical management (Table 5-4). If these patients have comorbid diseases, they also should be offered pharmacotherapy (weight loss medications). Patients with a BMI more than 35 kg/m2 should be offered medical management and pharmacotherapy. Surgical therapy is recommended for patients with a BMI more than 35  kg/m2 with significant comorbid diseases and who have not responded to medical management and pharmacotherapy. Patients with a BMI more than 40 kg/m2 should be offered medical management, pharmacotherapy, and surgery. An enlarged waist circumference also places a patient at higher medical disease risk, separate from comorbid diseases, and correlates well with total body fat.62 Men with a waist circumference more than 40 inches or women with a waist circumference more than 35 inches have a higher risk, similar to the presence of significant comorbid diseases in a man with a BMI more than 30  kg/m2. These patients would be offered pharmacotherapy in addition to medical

CELIAC DISEASE

Small intestinal mucosal injury and consequent malabsorption in celiac disease (see Chapter 104) occur when a susceptible patient ingests gluten-containing foods such as wheat, barley, rye, or possibly oats. Patients, especially younger ones, present with classic signs of malabsorption, including diarrhea, cramping, and marked weight loss and often develop folate, iron, and fat-soluble vitamin deficiencies. The primary treatment is a gluten-free diet. Wheat starch free of gliadin forms the basis for most bread in a gluten-free diet. Corn, rice, and buckwheat are allowed. Most patients will improve with dietary management. The vast majority of commercially available EN products, if required, are gluten-free.

CANCER

Protein-calorie malnutrition is a common problem in cancer patients. Cancer cachexia is the consequence of multiple metabolic abnormalities induced by the tumor. Appetite stimulation has been used successfully in cancer patients with mild malnutrition.55 The routine use of aggressive nutritional support in all patients receiving chemotherapy and radiation is controversial. Prospective randomized studies have failed to show improved tolerance to chemotherapy with the use of nutritional support,56 and PN also has failed to show an improvement in morbidity from radiation therapy. Parenteral nutrition has been shown to be beneficial for patients with gastrointestinal obstruction from primary or metastatic tumors.57 Parenteral nutrition also has been found to be beneficial in patients following bone marrow transplantation who have developed severe gastrointestinal mucositis.58 Enteral nutrition has been used successfully in patients with head and neck cancer to prevent weight loss, reduce hospitalizations, and reduce interruptions in chemotherapy and radiotherapy. In summary, the use of nutritional support in the cancer patient should be restricted to those patients with a reasonable life expectancy who are likely to be unable to maintain their nutritional

Table 5-4  Targeted Obesity Treatment Based on Body Mass Index Body Mass Index (kg/m2) Treatment

25-26.9

27-29.9

Diet, physical activity, behavior therapy Pharmacotherapy Surgery

With comorbidity* -

With comorbidity* With comorbidity* -

30-34.9 + + -

35-39.9 + + With comorbidity*

>40 + + +

*Comorbidities of three or more of the following: hypertension, smoking, elevated low-density lipoprotein (LDL), reduced high-density lipoprotein (HDL), impaired fasting serum glucose level, family history of coronary artery disease, age (men > 45 yr; women > 55 yr); or, comorbidity of any of the following: coronary artery disease, atherosclerotic disease, type 2 diabetes mellitus, sleep apnea. +, treatment indicated; -, treatment not indicated.

Chapter 5  Nutrition in Gastrointestinal Diseases management, similar to the management of a patient with a BMI more than 30 kg/m2 with significant comorbid disease. Surgical management of obesity in the United States typically involves a Roux-en-Y gastric bypass, vertical banded gastroplasty, or gastric banding (see Fig. 7-1). In general, most weight is lost in the first year.63 Gastric bypass surgery has a reasonable five-year outcome for maintenance of successful weight loss, but there is an associated morta­ lity (0.5% to 2%) and complications associated with surgical therapy, some of them very significant.64,65

NUTRITIONAL THERAPY PARENTERAL NUTRITION

For patients with a nonfunctioning gastrointestinal tract, nutrients can be delivered directly into the venous system via a process referred to as parenteral nutrition (PN); nutrients can be delivered into a central vein (central parenteral nutrition, CPN) or a peripheral vein (peripheral parenteral nutrition, PPN). PN delivers a solution consisting of water, electrolytes, amino acids, carbohydrates, fats, proteins, vitamins, and trace elements. These compounds are mixed and delivered over a period of time, usually 12 to 24 hours. Table 5-5 details a typical CPN formula. A CPN solution is six times more concentrated than blood (1800 to 2400 mOsm/L) and generally consists of approximately 30 to 50 g of protein and 1000 to 1200  cal/L. Determination of caloric and protein needs is based on a prior nutritional assessment.66 Table 5-6 provides an approximation of a patient’s daily protein and calorie needs based on the severity of the disease processes. These needs could be used instead of more complicated

Table 5-5  Sample Central Parenteral Nutrition Order* Component

Amount

Amino acids Dextrose Lipids

220 kcal/L (55 g protein) 555 kcal/L (163 g carbohydrates) 400 kcal/L (40 g of lipids), total: 1175 kcal/L 70 mEq/L 35 mEq/L 5 mEq/L 5 mEq/L 15 mmol/L To balance Industry standard Industry standard Individualized

Sodium Potassium Calcium Magnesium Phosphorus Chloride, acetate Multivitamins (MVI-13) Trace elements (MTE-5) Drug additives/L (heparin, insulin, H2 blockers)

*Volume: 2000 mL (83 mL/hr over 24-hr infusion).

Table 5-6  Calorie and Protein Needs Based on Degree of Physiologic Stress

Physiologic Stress Mild Moderate Severe

Calorie Needs (kcal/kg/day)

Protein Needs (g/kg/day)

25-28 28-32 32-35

0.8-1.0 1.0-1.5 1.5-2.0

formulas presented earlier in this chapter for determining calorie and protein requirements. Overall, daily water requirements are estimated at 20 to 30 mL/kg.

Central Parenteral Nutrition Formulation

To create a PN formula, one must first determine the protein, carbohydrate, and fat content required by the patient. A stepwise approach is effective (Table 5-7). Each component of PN has a defined caloric content, with protein = 4 kcal/g, carbohydrate = 3.4  kcal/g, and fat = 10  kcal/g. Once the macronutrient components of a PN formula have been defined, electrolytes, trace elements, multivitamins, and some medications such as insulin, heparin, and H2 blockers can be added. Any of the components can be increased or decreased based on a patient’s laboratory values and comorbid disease processes. In addition, serum glucose levels should be maintained below 120 mg/dL by the addition of insulin to the solution.67 Water is added to meet the daily volume needs of the patient.

Peripheral Parenteral Nutrition Formulation

PPN prescriptions account for an increasing percentage of all PN prescriptions written each day in this country, perhaps because hospital patients often are on short-term PN therapy. Fifty percent of patients in the hospital are on PN for less than 10 days and 75% for less than 14 days.68 Use of PPN would allow potential reduction of the risk and complications associated with central venous access because it is delivered via a peripheral vein. Formulation of PPN requires more attention to the solution osmolarity than to its actual caloric or protein content. A hyperosmolar solution can cause a chemical thrombophlebitis, resulting in patient discomfort and loss of peripheral venous access. In general, PPN solutions should be maintained at 1000 mOsmol/L or lower.69 Of all of nutrient components, carbohydrates have the most significant impact on PPN osmolarity. Because of the osmolarity issues, the caloric content of PPN/mL is limited (Table 5-8). To

Table 5-7  Stepwise Approach to Writing a Parenteral Nutrition Order* Caloric Contents of Nutrient Substances Protein, 4 kcal/g Fat, 10 kcal/g Carbohydrates, 3.4 kcal/g Estimated Daily Needs for This Patient Calories: 30 kcal/kg = 2100 kcal Protein: 1.2 g/kg = 84 g Fluids: 30 mL/kg = 2100 mL Steps 1. Add protein (1.2 g/kg/day) to the PN mixture.   84 g of protein needed   1 g of protein = 4 cal (total, 326 kcal)   2100 kcal − 326 kcal = 1774 kcal still required 2: Add lipids (1.0-1.5 g/kg/day).   70 g fat = 700 kcal   1774 residual calories − 700 kcal = 1074 kcal still required 3. Add carbohydrates (3-5 gm/kg/day).   1074 kcal/3.4 cal/g carbohydrate = 315.8 g 4. Make total volume.   30 mL/kg = 2100 mL Additional Additives Electrolytes, minerals, vitamins added (see PN example formula for details) Drug additives: Histamine 2 blockers, insulin, heparin *Order is written for a 70-kg man with moderate physiologic stress (see Table 5-5). PN, parenteral nutrition.

85

86

Section II  Nutrition in Gastroenterology Table 5-8  Glucose Concentrations and Osmolarity of Intravenous Solutions Glucose Concentration (%) 5 10 20 40 50 60 70

kcal/L

Osmolarity (mOsm/L)

170 340 680 1360 1700 2040 2380

252 505 1010 2020 2575 3030 3535

approximate a patient’s daily calorie and protein needs, a larger volume of PPN than CPN must be delivered to maintain the solution’s lower osmolality. The addition of heparin, hydrocortisone, or lipids to the PPN solution may reduce the incidence of chemical thrombophlebitis.70

Administration

The average CPN solution is composed of approximately 25% to 30% solute. At initiation, the CPN formula should be infused over 24 hours. Patients with glucose intolerance or those at risk for refeeding syndrome (see complications below) should have their PN infused at half their daily caloric needs for the first 24 hours. This ratio may be increased to full caloric needs over the next 24 to 72 hours with monitoring of serum glucose, electrolyte, magnesium, and phosphate levels and fluid tolerance.71 Central PN is infused via a large central vein, whereas less hypertonic PPN may be infused via a peripheral vein. In either case, the port or lumen of the catheter used for PN must be used solely for PN infusion. Use of the PN port or lumen for blood drawing or infusion of other solutions dramatically increases the risk of catheter infection.

Laboratory Testing

Following PN initiation, some laboratory values need to be monitored. In the first few days following initiation of PN infusion, serum electrolytes, magnesium, phosphorus, calcium, and blood urea nitrogen (BUN) should be monitored closely. After stabilization, these values can be checked weekly. In the home-parenteral-nutrition (HPN) patient on long-term PN, checking of these routine laboratory tests sometimes is decreased in frequency to every other week or once a month. Complete blood counts should be followed monthly. In other clinically relevant cases, serum zinc, selenium, copper, chromium, vitamin B12, and vitamin B6 levels may need to be monitored. Iron is not a standard additive for PN solutions and patients who are PN-dependent require monitoring for the development of iron deficiency anemia.

Metabolic Complications

Metabolic complications may develop as a consequence of the glucose, amino acids, lipid, vitamin, electrolyte, or mineral content of the PN solution.72 Hyperglycemia is the most common complication and is directly related to the dextrose content of the PN, the patient’s glucose tolerance, and the rate of PN infusion. Critically ill patients and patients with preexisting glucose intolerance require the most aggressive monitoring of serum glucose. The serum glucose level should be maintained below 120  mg/dL. In the critical care setting, tight control of serum glucose levels has been shown to improve patient morbidity and mortal-

ity.73 Patients who develop hyperglycemia should first be maintained on a sliding scale of regular insulin. Of the total amount of sliding scale insulin required over 24 hours, two thirds should be added to the next day’s PN formula. Further adjustments in insulin dosing may be required on a daily basis and some patients in the hospital may require a separate insulin drip. Failure to control blood glucose levels results in an increase of infectious complications, such as catheter sepsis. In some hospitalized patients, blood glucose control may be difficult, even with an insulin drip. In such cases, the caloric load of the PN solution, specifically carbohydrate and fat calories, is reduced; this is known as permissive underfeeding.74 In these patients, the risk of hyperglycemia and its consequences has been determined to be more than the risk of temporarily underfeeding the patient. During permissive underfeeding, a patient’s full daily protein needs continue to be met. Refeeding syndrome is a common metabolic consequence of PN resulting from sudden provision of a large amount of glucose calories to a patient who previously was malnourished. With PN infusion, the metabolism of these patients attempts to become rapidly anabolic. Insulin production is increased, pushing potassium, phosphorus, and magnesium into intracellular compartments, with resultant hypokalemia, hypophosphatemia, and hypomagnesemia.75 Sodium retention and large fluid shifts also can occur and may place the patient at risk of developing congestive heart failure.76 Abnormal liver biochemical test results are common after initiation of PN and typically feature elevations of serum aminotransferase levels up to twice normal. Greater elevations in aminotransferase levels and associated hyperbilirubinemia warrant investigation. Liver disease such as viral hepatitis, nonalcoholic steatohepatitis, sclerosing cholangitis, primary biliary cirrhosis, autoimmune hepatitis, and hemochromatosis need to be excluded. Right upper quadrant ultrasonography usually can eliminate the diagnosis of a liver mass, cholelithiasis, or biliary sludge. Acalculous cholecystitis needs to be excluded because the normal intestinal stimulation of gallbladder contraction is lost when a patient has no enteral intake. True PN-induced liver disease presents as a fatty infiltration of the liver that is especially pro­minent in the periportal areas, and may respond to a reduction in a patient’s total daily carbohydrate or total calorie infusion.77 It has been shown that repeated episodes of catheter sepsis will increase the probability of a deterioration in liver biochemical test results in patients on PN. Patients with short bowel syndrome are more likely to develop PN-associated liver failure, although the reason for this is unclear. Current research suggests that choline deficiency may play a role in the development of liver disease associated with PN use.79 Unfortunately, choline is not available for infusion in the United States. In addition, the lipid source of the PN solution is believed to contribute to this complication. In the United States, the lipid content of the PN solution is soy-based and composed of long-chain triglycerides (omega-6 fatty acids). In early reports, the use of an omega-3–containing lipid source (fish oil) had been shown to be effective in reversing PN-induced liver disease.80 Patients who develop significant complications with PN use may be candidates for small bowel transplantation. These complications include progressive liver failure, repeated catheter sepsis, or thrombosis of major venous systems that precludes obtaining central venous access. The arrival of tacrolimus as an immunosuppressive agent has improved small bowel transplant outcomes. Current fiveyear survival rates for patients receiving small bowel trans-

Chapter 5  Nutrition in Gastrointestinal Diseases plants are close to 50%.81 However, the five-year survival rate of patients on HPN are still better than the five-year survival rate for patients receiving small bowel transplants. Quality-of-life differences between HPN and small bowel transplantatation still are being explored.

Vascular Access Devices

Vascular access devices have developed substantially over the past 40 years. Anatomically, the subclavian and internal jugular veins provide the safest and easiest central venous access. The subclavian vein often is chosen for long-term access, such as for HPN, because of reduced catheterassociated complications compared with other central venous access sites. Multilumen catheters allow for the infusion of a number of fluids and medications at the same time. Tunneled Silastic catheters (Hickman, Broviac; CR Bard, Murray Hill, NJ) commonly are used for longer term vascular access. They have a Dacron cuff to induce fibrotic tissue adherence, which is believed important in preventing bacterial migration up the catheter. The catheter tip should be positioned in the distal third of the subclavian vein, not in the right atrium. Implantable ports are placed subcutaneously, usually on the chest wall. They require a specialized access needle to allow blood drawing or fluid infusion. It is believed by some investigators that implantable ports are associated with a reduction in infectious and thrombotic complications compared with tunneled catheters.83 Because they are implanted in the chest wall, these ports require a more extensive procedure for bedside removal than tunneled catheters. Peripherally inserted central catheters (PICCs) have been used for PN infusion both in the hospital and at home. The PICC line is generally placed in an upper extremity. These catheters are long, allowing the tip of the catheter to be positioned into a central vein. PICC lines are associated with a reduction in major insertion complications, such as pneumothorax, compared with standard centrally inserted catheters; however, they have more complications, such as infection, thrombosis, and displacement.84 The use of PICCs for home PN infusion has not yet been evaluated in a prospective study. Some patients with continued difficulty maintaining central venous access have limited options. The use of an arteriovenous (AV) shunt may be considered an alternative, because it provides the patient with a skin level approach to the vascular system in an area of high blood flow. There are no prospective studies evaluating the use of an AV shunt for the delivery of PN. Other less commonly used and greater risk venous access sites include transhepatic, transvertebral, and direct puncture of the inferior vena cava or right atrium.

Central Venous Catheter Complications

Central venous catheter complications occur, with incidence rates of 1% to 20%.85 Complications of subclavian vein catheterization include hemothorax, pneumothorax, brachial plexus injury, hematoma, and subcutaneous emphysema. Common long-term catheter complications include sepsis, thrombosis, and catheter occlusion. Catheter infection generally occurs from touch contamination and often involves coagulase-negative staphylococcus. Currrent hospital environments also have led to a surge in vancomycin-resistant enterococcus. The major mechanism of catheter contamination is tracking of organisms from the skin to the subcutaneous tissues and catheter tip. Other organisms frequently causing catheter infection include gram-negative bacteria and fungi. In the home setting, the more time spent teaching the patient about the

care and operation of the central venous access device, the less likely the patient is to develop complications.86 Diagnosing catheter infections can be difficult. The white blood cell count may not be elevated in a patient with a venous catheter infection. Peripheral blood cultures can be negative and catheter tip culture is a much more sensitive method of documenting catheter infections. Generally, bacterial infections of catheters can be treated with the catheter in place, whereas fungal catheter infections and tunnel infections of the catheter tract require catheter removal for effective treatment.87 Often, broad-spectrum antibiotic or antifungal treatment of catheter infections is initiated once the diagnosis is suspected and definitive therapy introduced when the organism is identified. The addition of heparin (1000 units/L) to each bag of PN solution can prevent subclinical thrombus formation to which bacteria or fungi can attach, thereby potentially reducing the risk of catheter sepsis. When treating central venous catheter infections, importance is given to the actual contact time between the bacteria or fungus in the catheter and the antibiotic or antifungal agent. Some believe in locking a small amount of antibiotic into an infected line to try and reduce central line decontamination.88 Others advocate flushing the catheter daily with absolute alcohol (100%) to sterilize the catheter and reduce infectious events.89 Catheter-induced thrombosis occurs secondary to irritation of the blood vessel wall. The thrombus usually is composed of fibrin. Precipitation of medication in the catheter occurs less commonly. Symptoms of central vein thrombus formation include neck pain, neck swelling, anterior chest wall venous distention, and reduced catheter function. Flushing of the central venous catheter with saline has proven as beneficial as flushing with heparin for the prevention of catheter occlusion with a thrombus.90 Treatment of fibrin thrombus formation requires a thrombolytic agent, such as streptokinase, given as a bolus or continuous infusion.91 Medication or precipitate occlusions can be treated by the instillation of small amounts of sodium hydroxide or hydrochloric acid, depending on whether the drug is acidic or basic, respectively.92

ENTERAL NUTRITION Enteral Access

In a patient who can eat and drink, EN support focuses on the use of nutritional supplements, dietary counseling, and appetite stimulation. In those patients who will not or cannot eat secondary to some dysfunction of the gastrointestinal tract, an enteral route or feeding tube is necessary to provide feedings. In this situation, obtaining enteral access becomes the foundation of any attempt to provide EN. The radiologist, gastroenterologist, or surgeon usually places enteral access devices (Table 5-9). This can be done at the bedside, fluoroscopically, endoscopically, or in the operating room, depending on the specific device and the expertise available. Medicare trends from 1997 to 2000 have shown a significant increase in enteral access procedures in the United States. The greatest increase in enteral access procedures was among radiologists, closely followed by gastroenterologists. The percentage of surgeons performing enteral access has decreased over time.93 Nasoenteric Tube Access Nasoenteric tube placement techniques have been developed for placement at the bedside, with endoscopy, with fluoroscopy, or during surgery. These techniques all have their indications, benefits, and risks. The final position of

87

88

Section II  Nutrition in Gastroenterology Table 5-9  Enteral Access Methods Type of Access Surgical or Percutaneous Gastrostomy Gastrojejunostomy Jejunostomy Nasal or Oral Nasal or oral gastric tube Nasal or oral gastrojejunal tube Nasal or oral small bowel tube

Purpose

duration of Need (mo)

Gastric feeding Gastric decompression Gastric decompression Gastric feeding Jejunal feeding Jejunal feeding

>1

Gastric feeding Gastric decompression Gastric decompression Gastric feeding Jejunal feeding Jejunal feeding

<1

Table 5-10  Complications of Nasogastric or Nasojejunal Tubes Aspiration pneumonia Nasal mucosal ulceration Otitis media Pharyngitis Pneumothorax Sinusitis Tracheoesophageal fistula Tube migration Tube obstruction

an enteral access tube is the stomach for gastric feedings, or the jejunum for small bowel feedings. A patient who is intolerant of gastric feedings, such as a patient with gastroparesis, gastric outlet obstruction, or who has had the stomach surgically removed, will receive small bowel feedings. Nasogastric and nasojejunal tubes have similar complications (Table 5-10). The use of small bowel feedings to prevent tube feeding aspiration events is a complicated and contentious issue. Some studies have shown fewer aspiration episodes in patients whose feedings were placed directly into the small bowel than when placed into the stomach.94,95 A prospective trial by Neumann and DeLegge96 directly compared the incidence of aspiration episodes with gastric feedings and small bowel feedings in the intensive care unit and found no difference. It took longer to initiate small bowel feedings, however, because of the difficulty in obtaining adequate tube position. A 2002 consensus conference determined that small bowel feeding is recommended for the prevention of aspiration pneumonia in critically ill patients,97 but this debate certainly will continue. Small bowel feedings should be initiated in patients known to have gastroparesis, patients who are intolerant of gastric feedings, or patients who have had a witnessed tube feeding aspiration event with gastric feedings. Bedside nasoenteric tube placement is the most common enteral access technique used in the hospital and long-term care environments. A nasogastric (NG) or nasojejunal (NJ) tube may be placed, a decision based on concerns with tube feeding tolerance and aspiration risk. There are many techniques available for passing bedside NG tubes. Typically, an 8- to 12-Fr NG tube is lubricated and passed into the stomach

>1 >1

<1 <1

with the patient’s head flexed; the patient ingests sips of water to assist in passing the tube into the stomach.98 Many centers advise bedside auscultation for confirmation of an adequate position of the NG tube before its use; however, this can be misleading because a tube in an inappropriate location (e.g., lung, pleural cavity, esophagus) may be misinterpreted as being in proper position by improper bedside auscultatory techniques. Roubenoff and Ravich99 have reported on a technique to avoid intrapulmonary placement of nasoenteric tubes in high-risk patients. They suggested measuring a length of the tube from the earlobe to the xiphoid process before insertion. Once the tube is passed to this length, an anterior-posterior plain film is obtained to determine that the tube is in the esophagus before passing it further into the stomach or small intestine. Every patient should have a plain film to confirm proper positioning of a NG or NJ tube before initiating feedings.100 It is not unusual to be faced with a patient who is comatose and therefore unable to assist with the passage of a nasoenteric tube. In this case, the tube can again be passed at the bedside after tube lubrication and head flexion. The patient is monitored for coughing and wheezing consistent with a bronchial placement. Auscultation of the abdominal cavity and an abdominal plain film can confirm proper tube location. The difficulty of passing a nasoenteric tube at the bedside blindly into the small intestine has prompted development of a variety of techniques. Thurlow has promoted the use of a stylet-filled NJ tube that is advanced with a corkscrew motion.101 Zaloga has confirmed the reliability of this technique, with a more than 90% success rate.102 Ugo and associates have reported on placing the patient in the right lateral decubitus position and tracking the position of the tube into the small bowel by using a stethoscope and auscultation, a technique that resulted with an 83% success rate.103 Lord and coworkers have promoted the use of unweighted feeding tubes for these bedside passages; their success rate for spontaneous small bowel placement was far more than that documented for weighted tubes (92% vs. 56%).104 The hypothesis here was that the weight at the tip of the feeding tube actually was an impediment to the tube’s spontaneous passage through the pylorus. In Lord’s study, metoclopramide was used to promote passage of the tube from the stomach into the small intestine. Others have reported on positioning a tube beyond the pylorus with the use of pharmacologic agents; results with metoclopramide ranged from no benefit105,106 to a 90% success rate.107,108 Silva

Chapter 5  Nutrition in Gastrointestinal Diseases Table 5-11  Endoscopic Methods of Nasoenteric Tube Placement Methods

Technique

Drag-and-pull

Suture on end of a tube pulled with forceps into position Tube pushed into position over a guidewire Tube pushed through biopsy channel of endoscope into small bowel Tube passed over guidewire placed through a nasal endoscope

Over-the-guidewire Through-the-scope Nasal endoscopy

and colleagues, in a comprehensive literature review, noted that metoclopramide given intravenously or intramuscularly was effective in promoting successful nasoenteric tube placement into the small intestine.109 Griffith and associates have confirmed the usefulness of another promotility agent, erythromycin, for promoting successful NJ tube placement at the bedside in critically ill patients.110 Newer technologies have been developed for NJ tube placement. Self-propelled NJ tubes have a spiral tip at the distal end. It is hypothesized that the stomach can propel this type of tip through the pylorus easier than with a standard straight tip. Berger and coworkers have reported a 50% success rate in passing the pylorus with the self-propelled feeding tube in 105 critically ill patients.111 Magnetically steered NJ tubes also have been reported to show reasonably successful rates of proper placement.112 Concurrent use of narcotics decreased the likelihood of successful tube placement. If attempts fail to pass an NJ tube blindly at the bedside, fluoroscopic or endoscopic methods of passage are then required. The preference of technique is center-dependent. In centers with available C-arm fluoroscopy and modified fluoroscopy beds, fluoroscopic passage of NJ tubes can be done at the patient’s bedside. Success of fluoroscopic guidance of NJ tube passage can approach 100%.113 In institutions without bedside fluoroscopic capabilities, however, transport of patients to the radiology suite, especially critically ill patients, can be time-consuming, expensive, and hazardous.114 In these cases, bedside endoscopic passage of NJ tubes is preferable. NJ feeding tubes can be placed endoscopically at the bedside with moderate sedation. Table 5-11 lists the techniques for endoscopic passage of nasoenteric tubes at the bedside. The “drag-and-pull” method has the longest history. In this technique, a suture or another grabbable item is attached to the end of a NJ tube and used to drag the NJ tube into position in the small intestine with a grasping forceps. Difficulty usually occurs in releasing the suture from the grasping forceps. A second common technique, the “over-the-guidewire” technique, requires the initial endoscopic placement of a guidewire into the small intestine. The endoscope is removed and the guidewire is left in place. A feeding tube subsequently is passed blindly or with fluoroscopic assistance over the guidewire and into position in the small intestine. Patrick and colleagues have reported a 94% success rate using this technique.115 Fang and associates have described the use of an ultrathin endoscope to perform nasal endoscopy and place a guidewire into the small intestine. After the endoscope is removed, a NJ tube is passed over the guidewire into position.116 With this technique, no sedation is required. Other methods of endoscopic NJ placement are used less frequently.

Nasoenteric tube placement is the most common method of enteral access. Unfortunately, nasoenteric tubes may fail early because of tube occlusion or dislodgment, with subsequent interruption of tube feeding and medication regimens. Therefore, nasoenteric tubes should be used in patients requiring NG or NJ access for less than one month. The tip of the NJ tube may be anchored onto the small bowel mucosa using an endoscopic clipping device, a practice that seems to add a few days to the projected longevity of NJ tubes, presumably by reducing the risk for NJ tube migration.117 Patients who have experienced repeated, early failure of nasoenteric tubes should receive more permanent enteral access, such as a percutaneous endoscopic gastrostomy, percutaneous endoscopic jejunostomy, surgical gastrostomy, or surgical jejunostomy. The decision to use a jejunostomy (J) tube warrants specific instructions regarding the tube’s care. The lumen of a J tube is usually much smaller than that of a gastric tube and therefore is prone to clogging. Jejunal feeding tubes should never be checked for residual content because this measure is a poor indicator of the residual content of the small bowel. In addition, checking residuals through these small-bore tubes increases their probability of clogging. J tubes should be flushed after every tube feeding and medication instillation. Only liquid medications, or completely dissolved medications, should be placed through a J tube to reduce the chances of tube occlusion. Care should be taken to stop tube feedings during infusion of certain medications, such as theophylline or potassium chloride, products that are known to coagulate tube feedings and obstruct the J tube. Percutaneous Endoscopic Enteral Access If a patient will require enteral access for longer than one month, percutaneous procedures are preferred. The endoscopic procedures include percutaneous endoscopic gastrostomy (PEG), percutaneous endoscopic gastrojejunostomy (PEG/J), and direct percutaneous endoscopic jejunostomy (DPEJ). All these procedures require the use of moderate or deep sedation and can be performed in the endoscopy suite, in the operating room, or at the bedside. In comparison to NG access, PEG has been shown to be a more reliable enteral access tube, allowing patients to receive more calories daily because of a reduction in tube dysfunction.118 Percutaneous Endoscopic Gastrostomy.  PEG was developed by Gauderer and coworkers in the early 1980s.119 The procedure involves the placement of a percutaneous gastrostomy tube after endoscopic transillumination of the abdominal wall for an appropriate gastrostomy site. The use of prophylactic antibiotics before the procedure is important to prevent postprocedure infections.120 Placement of a PEG tube can be accomplished by the Sachs-Vine (push) or Ponsky (pull) techniques, depending on physician preference. Prospective evaluations of PEG placement have found this procedure to be associated with few procedure-related complications.121 An older technique, the push-dilator technique, has started to make a resurgence. In this technique, the endoscope is used as a viewing port during PEG tube placement. T fasteners are placed percutaneously to attach the wall of the stomach to the abdominal wall; an incision is made in the abdominal wall and a fistula into the stomach is created and sequentially dilated. Ultimately, a PEG tube with a balloon internal bolster is passed through the newly created insertion site, as is done with balloon replacement tubes (Fig. 5-4).122 PEG tubes are indicated for patients who will be unable to consume sufficient nutrition for longer than one month,

89

90

Section II  Nutrition in Gastroenterology PEG tube placement for this patient population. One retrospective analysis has suggested that mortality among dementia patients does not differ, regardless of whether a PEG tube is placed,130 although this study did not address whether the use of PEG tubes in this population for hydration and medication delivery would be appropriate. The use of PEG in the dementia population remains a subject of great debate131 and ultimately a prospective randomized trial will be required to answer the question definitively.

Figure 5-4.  Introductory percutaneous endoscopic gastrostomy (PEG) kit. This kit uses T fasteners to tack the stomach to the abdominal wall, followed by PEG tube tract serial dilations.

despite a functional gastrointestinal tract. Patients requiring PEG placement often are older and have numerous comorbid diseases. In one study of older patients referred for PEG placement, there was a 48% mortality at seven days if the patient had a prior aspiration episode or urinary tract infection, or was older than 75 years, compared with a 4% mortality if none of these risk factors was present.123 In addition to providing access for nutrition, PEG tubes are indicated for hydration and administration of medications as well as for gastric decompression. Some of the more common medical indications for PEG placement are described here. Cancer.  One area of oncology in which PEG tubes are beneficial is head and neck cancer. The benefit of PEG tubes in this setting was illustrated in a retrospective study that had 40% (32 of 88) of head and neck cancer patients receiving a PEG tube prior to chemotherapy and radiotherapy. Those who received a PEG lost an average of 3.1 kg compared with 7 kg of weight loss for those without a PEG. The same PEG group had significantly fewer hospitalizations for dehydration and malnutrition and had no interruption in treatment of their cancer compared with the group which did not receive PEG.124 Stroke.  Data support the use of PEG tubes in patients with stroke-related dysphagia. In one study, the authors reported a 1-, 8-, and 48-month survival of 78%, 35%, and 27%, respectively, when the most common indication was a hemispheric stroke.125 Compared with NG feeding, early PEG placement was found to be associated with a lower incidence of ventilator-associated pneumonia.126 In contrast, Dennis and colleagues have reported an increase in the death rate and poor outcome in stroke patients randomized to PEG tube feedings as opposed to nasoenteric tube feedings.127 In this same study, patients receiving early tube feeding by any means had a reduction in overall mortality compared with patients who did not receive early enteral nutrition. Dementia.  Dementia is a frequent disorder of older adults and a common indication for referral for PEG. More than 36,000 older patients with dementia receive a PEG tube each year.128 The benefit of providing EN in these patients, however, is less clear.129 No large, prospective, randomized trials have been performed to evaluate the usefulness of

Disabling Neurologic Conditions.  The use of PEG tubes for the patient with disabling motor neuron disease, such as amyotrophic lateral sclerosis (ALS) and multiple sclerosis (MS), has been examined. Both these diseases are associated with progressive dysphagia, leading to reduced oral intake, weight loss, and an increasing risk of oral aspiration. Langmore and coworkers did an extensive literature search of patients with ALS receiving PEG and only identified retrospective studies and prospective cohort studies.132 Using these data, they concluded that survival and nutritional status could be improved with PEG placement in ALS. There was a trend for improved quality of life although this required further study. Small case reports have demonstrated an improvement in comorbid disease states, such as pressure ulcer healing, in patients with MS and dysphagia who receive tube feedings.133 PEG Procedure.  PEG kits are commercially available from a number of manufacturers. The most common sizes for adult patients range from 16 to 24 Fr. Most tubes are made of silicone, although some are constructed of polyurethane. In general, PEG tubes start to degrade one to two years after placement, usually from yeast implantation and degradation of the PEG tube wall.134 PEG tubes are less likely to clog than nasoenteric tubes because of their larger size. Obstructed PEG tubes may be cleared by flushing them with warm water. In some cases, pancreatic enzymes mixed in a bicarbonate solution also can be effective.135 There are no data to support the use of juices, soft drinks, or meat tenderizers to unclog a PEG tube. Commercially available PEG tube cleaning brushes also are available. Relative contraindications for PEG placement include the presence of gastric varices, severe obesity, major gastric resection, significant disease of the gastric or abdominal wall, ascites, and coagulopathy. Absolute contraindications include the inability to transilluminate the anterior abdominal wall and an ineffective digital intrusion of the abdominal wall to locate a safe gastric access site.136 Once a PEG tube malfunctions, degrades, or is no longer needed, it can be removed at the bedside with a traction pull force of seven to 10 pounds; such removable PEG tubes are labeled as traction removal tubes.137 Some PEG tubes, labeled as endoscopic removal tubes, have a stiff internal bolster and can be removed only with an endoscope. Although there is an increase in cost with the use of endoscopic removal PEG tubes because of the need for a repeat upper endoscopy at removal, they may be safer to use in patients who are confused or combative and at risk for pulling their PEG tube out after initial placement. Some authors have suggested cutting the PEG tube at the abdominal wall level and allowing the remaining tube and internal bolster to pass through the gastrointestinal tract; this should not be done because there have been cases reported of these cut internal bolsters leading to a small bowel obstruction.138 Most post-PEG complications arise from a patient’s comorbidities, such as poor wound healing, aspiration, or coagulopathy.139-141 Caregivers must raise the head of the

Chapter 5  Nutrition in Gastrointestinal Diseases patient’s bed to 30 to 45 degrees during and for one hour after feeding to reduce the risk of aspiration.142 The most common complication is peristomal wound infection.143 Excessive tightening of the PEG tube external bolster against the abdominal wall can cause tissue ischemia, wound leakage, and necrotizing fasciitis.144 To minimize this complication, the external bolster of the PEG tube should be maintained 1 to 2 cm from the anterior abdominal wall to avoid tissue compression and wound breakdown.145 Peristomal wound infections often are treated for seven days with an oral antibiotic such as cephalexin to cover skinrelated microorganisms. The infected area should also have daily topical cleansing with or without antibiotic ointment. The PEG tube should be removed in cases of worsening infection. Other common complications include pneumoperitoneum, fever, ileus, cutaneous or gastric ulceration, and tube extrusion or migration.146-148 Pneumoperitoneum is common after PEG placement and is not diagnostic of a perforation. Careful clinical evaluation of a patient after PEG placement with questionable peritonitis from perforation or leak should include a contrast study through the PEG tube to avoid unnecessary surgical exploration.149 Major complications are rare and include intra-abdominal bleeding or hematoma formation, peritonitis, necrotizing fasciitis, gastric or colonic perforation, and hepatogastric, gastrocolic, and colocutaneous fistula. A colocutaneous fistula results from the inadvertent placement of a percutaneous feeding tube through the colon before it enters the stomach. When a colocutaneous fistula is documented, the tube should be removed and the patient’s condition should be monitored for appropriate closure of the fistulous tract. If the tract does not heal, surgery may be warranted to repair the fistula. If a PEG tube dislodges within four weeks of placement, fluoroscopy can be used to replace it at the bedside or endoscopy may be repeated. The concern in this situation is that the abdominal and gastric wall have separated from one another when the PEG tube was dislodged. If the PEG tube is dislodged more than four weeks after placement, the tract may be mature enough to replace the PEG tube blindly at the bedside without fluoroscopy or endoscopy. Proper placement can be confirmed by aspirating gastric contents from the replacement PEG tube. If there is any question that the replacement tube tip is not in the gastric lumen, a contrast radiologic study through the PEG tube should be obtained. Replacement PEG tubes are divided broadly into two categories, replacement gastrostomy tubes and low-profile devices. Replacement gastrostomy tubes usually have a balloon-type internal bolster (Fig. 5-5). These balloon tubes can be inserted blindly through the gastrostomy site into the gastric lumen. The balloon is inflated to serve as the internal bolster and an external bolster is slid down the external tube against the abdominal wall to keep the PEG tube from migrating. There also are replacement PEG tubes with a distensible internal bolster (Fig. 5-6). The internal bolster is stretched with a stylet and pushed blindly through the gastrostomy site and the stylet is then removed, allowing the internal bolster to assume its previous shape. The direction of the gastrostomy tract should be known so that the stylet does not damage or rupture the tract. PEG tubes also may be replaced with low-profile gastrostomy devices (Fig. 5-7). These devices provide a skin level access to the gastric lumen, and may be particularly useful for disoriented patients who may habitually tug at their bedclothes and pull out their tube connections. The internal bolster of these low-profile devices may be an inflatable balloon or a distensible internal bolster that requires a stylet

Figure 5-5.  Replacement gastrostomy tube.

Figure 5-6.  Replacement percutaneous endoscopic gastrostomy device with distensible internal bolster.

Figure 5-7.  Low-profile percutaneous endoscopic gastrostomy device.

for placement. Low-profile PEG tubes come in predetermined lengths and the gastrostomy tract length must be measured in order to choose the correct length of the device. To access the low-profile device for feeding or gastric decompression, a separate access tube must be used to engage a valve in the top of the device. Although these tubes are cosmetically appealing, the small internal diameter of the access tubing and the valve make them more prone to valve and access tube occlusion.

91

92

Section II  Nutrition in Gastroenterology Percutaneous Endoscopic Gastrojejunostomy.  In patients in whom small bowel feedings are desired, endoscopic percutaneous small bowel access may be obtained by two methods. With the first method, percutaneous endoscopic gastrojejunostomy (PEG/J), a PEG is placed in the standard fashion, after which various techniques may be used to place a jejunal feeding tube through the PEG into the small bowel.150-152 Usually, a 9- or 12-Fr J tube is passed through the existing PEG and into position in the small bowel over a guidewire (Fig. 5-8). DeLegge and colleagues have reported a 100% success rate and no major complications using the over-the-guidewire technique for PEG/J placement, with a procedure time of 26 minutes.150 This PEG/J system allowed for concurrent gastric decompression and small bowel feeding. The average longevity of this tube system was approximately 120 days when patients who died from comorbid diseases were excluded from the analysis. A onepiece gastrojejunostomy (G/J) system does exist and is generally used as a replacement device that is passed through a preexisting PEG tube tract (Fig. 5-9). This tube can be dragged into position by a suture at the end of the

G/J system during endoscopy or passed over a guidewire during endoscopy or fluoroscopy. The internal bolster on this system is a balloon. The management of PEG/J tubes is similar to that of PEG tubes. Jejunal tubes need to be flushed aggressively to avoid clogging. Reported clogging rates of J tubes have ranged from 3.5% to 35%.153,154 Administration of semidissolved medications and bulking medications such as fiber through the J tube, and checking J tube residuals, leads to an increased incidence of tube occlusion. In contrast, medications may be given through the gastrostomy tube because of its larger diameter; the gastrostomy tube also may be used for decompression of gastroparesis or gastric outlet obstruction. Complications of PEG/J tubes include those already discussed for a PEG tube. In addition, the jejunal tube may migrate in a retrograde direction or become kinked so that it no longer functions. Tube migration occurs most commonly in patients who have persistent vomiting or in cases in which the J tube has been placed improperly.150,155 Direct Percutaneous Jejunostomy.  The second method of jejunal access, direct percutaneous endoscopic jejunostomy (DPEJ), directly places a J tube into the small bowel using an endoscope. This procedure requires the use of an enteroscope or a pediatric colonoscope to reach a puncture position beyond the ligament of Treitz. Success with this procedure has been reported by Shike and associates;156,157 minor complications included local site infection, but no instance of peritonitis or bowel infarction. One of the difficulties with DPEJ placement is the frequent migration of the small bowel away from the introducer trocar needle once an adequate entry site on the abdominal wall has been located. Varadarajulu and DeLegge158 have resolved this problem with the use of a two-needle stick technique (Fig. 5-10). In this procedure, a smaller, sharper, 19-gauge needle (finder needle) first is passed through an appropriate abdominal site into the small intestine. This finder needle is grasped by a snare, thus anchoring the small bowel against the abdominal wall. The larger introducer catheter is passed alongside the 19-gauge needle into the small bowel without pushing the small bowel into the abdominal cavity. The

Figure 5-8.  Percutaneous endoscopic gastrojejunostomy showing overthe-guidewire placement technique.

Figure 5-9.  Combination gastrojejunostomy balloon feeding tube (one piece).

Figure 5-10.  Finder needle insertion and snare fixation in two-needle stick technique.

Chapter 5  Nutrition in Gastrointestinal Diseases snare is removed from the 19-gauge needle and placed around the introducer catheter. A guidewire is passed through the introducer catheter into the small bowel, where it is then grasped by the snare and pulled out via the oral cavity. A J tube is attached to the guidewire and pulled into place in the small bowel similar to how a PEG tube is placed. Adequate positioning of the internal bolster of the J tube is confirmed with endoscopic visualization. Knowing what the patient’s needs are for jejunal access can help decide which enteral access is appropriate. Shortterm access is probably best achieved by a nasojejunal (NJ) tube. Patients requiring jejunal access for less than six months or those requiring concomitant gastric decompression would do well with a PEG/J system. Long-term jejunal access (i.e., more than six months) is best achieved with a DPEJ tube. Comparisons of PEG/J and DPEJ for jejunal access have found fewer tube-related complications, such as jejunal tube migration and jejunal tube occlusion, with the DPEJ tube.159,160 Immediately after DPEJ placement, it is helpful to leave the J tube unclamped so that the substantial amount of air that was insufflated during the procedure may escape and thereby decompress the small bowel. Management of DPEJ tubes is otherwise similar to that of PEG tubes. Complications and technical failures have been presented in three retrospective series on DPEJ outcomes. Technical failure rates ranged from 12% to 28%. Complications included bleeding, abdominal wall abscess, colonic perforations, peristomal infections, enteric ulcers, small bowel torsion, and intraperitoneal leakage; tube-related malfunctions similar to those with PEG tubes also have occurred.157-159 Surgical Enteral Access A number of studies have compared surgical gastrostomy with PEG and have shown cost savings, operative time savings, or a reduction in morbidity with PEG.160,161 In the standard surgical gastrostomy tube placement, a gastrotomy is formed and a gastric tube is placed into the gastric lumen. The gastric wall is then fixed to the abdo­ minal wall. Surgical gastrostomy first was described by Stamm in 1894 and has not changed significantly since that description.162 Surgical jejunostomy was performed first by Bush in 1858 in a patient with nonoperable cancer163 and, in 1891, Witzel described the most well-known technique for jejunostomy, which subsequently has undergone a number of modifications.164 Laparoscopic placement of J tubes and G tubes began in the early 1990s because it was believed that these procedures would be associated with less morbidity and operative stress than the standard surgical jejunostomy and gastrostomy. It was soon learned that these laparoscopic techniques did not add any significant advantage over standard surgical gastrostomy or jejunostomy with regard to operative time or associated procedure morbidity. Needle catheter jejunostomy (NCJ) involves the placement of a 5- or 7-Fr catheter into the jejunum via a submucosal tunnel. It was hypothesized that this technique would have fewer procedure-related complications than standard jejunostomy because of the smaller entrance created to the jejunum. Multiple studies have reported reduced infectious complications of NCJ compared with standard surgical jejunostomy; however, there is a significant increase in tube occlusions and dislodgment with the smaller NCJ. Fluoroscopic Percutaneous Enteral Access Placement of percutaneous gastrostomy and gastrojejunostomy tubes with fluoroscopic guidance has continued to gain acceptance since the introduction of this technique in

the early 1980s.165,166 These procedures usually are performed by radiologists in the fluoroscopy suite. After topical anesthesia to the abdominal wall, occasionally with additional moderate sedation, the inferior margin of the liver is identified by ultrasonography and marked on the patient’s abdominal skin surface. An NG tube is passed into the stomach, after which the stomach is insufflated and punctured with an introducer catheter; some but not all radiologists then attach the stomach to the anterior abdominal wall with T fasteners. A guidewire is placed into the stomach through the introducer and the puncture site is serially dilated over a guidewire to a size of 10 to 14 Fr. A gastrostomy tube is passed over the guidewire, through the dilated puncture site, and into the stomach or the small intestine if a gastrojejunostomy tube is desired. This fluoroscopic approach to enteral access has a reported technical success rate of over 95%.167 Most of the complications involve inadvertent puncture of contiguous abdominal organs or separation of the abdominal and gastric wall during gastrostomy tract dilation; the latter may lead to peritonitis, intraperitoneal leakage, and even death. Frequent occlusion of these feeding tubes, because of their typically smaller internal lumen size, has been shown to be avoidable if larger gastrostomy tubes (18 to 22 Fr) are used.168

Enteral Feeding

Patients may receive their tube feedings by bolus, intermittent, or continuous methods. Bolus feeding delivery allows a relatively large volume of tube feeding (200 to 400 mL) to be delivered over a short period. Intermittent feedings are delivered over a few hours by pump or by gravity drip using a bedside pole. Intermittent feedings may be practical for patients who cannot tolerate bolus feedings, and who do not require the precise delivery method of continuous enteral pump feedings. Continuous feedings usually are delivered over 12 to 24 hours by a mechanical pump. Patients receiving small bowel feedings are almost always fed using continuous feedings. An intermittent or continuous feeding regimen, rather than the rapid bolus method, may be used to limit the risk of tube-feed aspiration. For patients receiving gastric feedings, gastric residuals should be checked regularly and feeding intolerance monitored in all cases. Any gastric residual more than 400  mL should be followed closely.169 Repeated gastric residuals of more than 400 mL require the tube feedings to be stopped; in these cases, small bowel feedings may be necessary. Other indicators of tube feeding intolerance include persistent nausea and vomiting, abdominal distention, abdominal pain, and a dilated small intestine or colon seen on radiologic imaging.170 In general, it is a constellation of these signs, symptoms, and findings that should be present before intolerance to tube feeding is declared. There is no good consensus on when to begin feeding a patient after PEG placement because it is thought that a patient develops transient gastroparesis or there is potential for intra-abdominal PEG tube leakage immediately after tube placement. Brown and coworkers randomized patients to begin feedings 3 or 24 hours after PEG placement and found no differences in tolerance or complications that required discontinuation of tube feedings. Wound infections were more common in the delayed feeding group, however.171 Water Requirements After choosing an enteral formula, one must pay attention to the amount of free water a patient receives each day. All commercial enteral formulas contain a certain amount of free water. The more calorie-concentrated a formula, the less free water is contained in the formula.

93

94

Section II  Nutrition in Gastroenterology Table 5-12  Examples of Enteral Formulations Type

Examples

Fiber-containing (1.0 cal/mL) Free amino acids (elemental, low fat, 1.0 cal/mL) Polymeric (intact protein, 2.0 cal/mL) Polymeric (standard tube feedings, intact protein, 1 kcal/kg/day)

Jevity Vivonex TEN

Specialty Formulas Glucose intolerance (1.0 cal/mL, increased fructose component of carbohydrates) Hepatic formulation (increased branched-chain amino acids) Immune-enhancing (1.0 cal/mL; arginine-, glutamine-, omega-3 fatty acid–fortified) Pulmonary (1.5 cal/mL, low carbohydrate) Renal (2.0 cal/mL, increased essential amino acids)

Magnacal TwoCal HN (high nitrogen) Nutren 1.0 Isocal Osmolite HN (high nitrogen) Glucerna Nutra-Hep Impact Crucial Oxepa Pulmocare Amin-Aide

Advancement of Tube Feedings Once initiated, advancement of tube feedings is an imperfect science. In our center, patients are begun on continuous tube feeding at a rate of 30  mL/hr. Feedings are advanced at a rate of 10 to 20 mL every six hours until the patient’s goal rate is reached. Signs of tube feeding intolerance prompt temporary cessation of tube feeding or a reduction in the tube feeding rate. When patients reach their goal rates, they may be maintained on continuous 24-hour tube feedings or changed to 18- or 12-hour continuous tube feedings, intermittent tube feedings, or bolus tube feedings. Enteral Formulations Many formulations for enteral feeding are available including blenderized, polymeric, predigested, specialty, modular, and supplemental regimes (Table 5-12). Blenderized Formulations.  Blenderized formulations are combinations of vitamins and minerals added to table foods. Because of this, they have more fiber and high viscosity and osmolarity. They require a functional gastrointestinal tract. They are not recommended for use in small-caliber feeding tubes because of their propensity for causing tube clogging. Standard Polymeric Formulations.  Lactose-free, glutenfree, polymeric formulations are the basic feeding formulas designed for long-term use. These formulations are denoted as polymeric, because the macronutrient components are intact and not predigested. Standard formulations contain 15% to 20% calories from proteins, 45% to 60% calories from carbohydrates, and 30% to 40% calories from fats. Generally, these formulations provide 1 kcal/mL, although they may be concentrated to 1.5 to 2 kcal/mL. As the calorie content per milliliter volume of tube feeding increases, the free water content of the formula decreases. Most 1 cal/mL enteral formulas are 80% to 85% free water. Formulas that are higher in protein are designated as HN (high nitrogen). Predigested Formulations.  Predigested enteral formulations, also called elemental, semi-elemental, or small

peptide formulations, are designed for patients with limited digestive capacity. The protein and carbohydrate components of these formulations have been broken down into smaller substrates for easier absorption, and because of the presence of multiple smaller particles they are highly osmotic. These formulas also generally are low in fat content or contain a significant amount of MCTs to improve gastro­ intestinal tolerance for patients with fat malabsorption disorders. Specialty Formulations.  Specialty formulations are designed for patients with special nutritional requirements based on specific disease processes, such as diabetes, renal failure, hepatic failure, pulmonary disease, or severe stress or trauma. There are no data to show that these specialty formulations improve survival when used for their intended disease state. Immune-Enhancing Formulas.  Immune-enhancing formulas contain higher amounts of arginine, glutamine, omega-3 fatty acids, and nucleotides, substances shown to be important in immune modulation.. More recently, other immuneenhancing nutrients, such as borage (starflower) oil, have been examined. Immune-enhancing formulas may be useful for patients who: • Are about to undergo elective gastrointestinal surgery • Have blunt and penetrating torso trauma • Have an anticipated prolonged need for a mechanical ventilator • Are about to undergo head and neck surgery • Have major burns • Are critically ill If possible, administration of an immune-enhancing formula should be initiated for five to seven days prior to elective surgery. In such cases, feedings should be advanced as tolerated until 1500  mL are administered daily or more than 50% to 60% of calculated nutrient goals are met. Evidence suggests that, with their use, there is a reduction in subsequent infectious complications, anti­ biotic needs, ventilator days, episodes of multiple organ dysfunction, and decreased hospital stay.172 Some newer immune-enhancing formulas are showing promise in changing clinical outcomes for patients with acute respiratory distress syndrome.173 Modular Feedings.  Modular feedings consist of individual nutritional components (carbohydrate, protein, and fat) that can be mixed in various proportions to create a custom enteral formulation. Modular formulas rarely are used today. Supplemental Regimens.  Supplements are taken with or between meals for those who cannot otherwise meet all their caloric needs.174 Most supplements are available in liquid, ready-to-drink formulations. They are palatable and usually available in a 1- or 1.5-cal/mL concentration. Enteral Feeding Complications Gastrointestinal side effects of tube feedings are reported in 15% to 30% of patients and include nausea, vomiting, abdominal distention, abdominal cramping, and diarrhea. Nausea, vomiting, and abdominal distention often can be resolved by slowing the delivery rate of the feeding. Diarrhea is the most common complication.175 Its pathophysiology is complex. The most common cause of diarrhea in patients on enteral feeding is Clostridium difficile enterocolitis resulting from concurrent antibiotic use. The second most common cause of diarrhea in this group of patients is

Chapter 5  Nutrition in Gastrointestinal Diseases medications. Frequently, medications are changed from tablet to liquid form for easy instillation through the feeding tube. These liquid medications often have a sorbitol base, which is a known cathartic. Magnesium-containing medications, hypertonic medications, and promotility agents also may promote diarrhea. High-osmolarity tube feedings often are cited as a cause of diarrhea, although studies have demonstrated tolerance of tube feedings, the osmolality of which is as high as 600 to 700 mOsm/kg. There are no data to support the recommendation that commercial enteral formulations be diluted in an attempt to improve their gastrointestinal tolerance. For patients with small bowel malabsorption, especially fat malabsorption, predigested, lower fat formulas may improve absorption and reduce diarrhea. Hypoalbuminemia may lead to small bowel wall edema and resultant diarrhea. There are no data to support the use of intravenous albumin to improve diarrhea in these patients. In these circumstances, the use of anticholinergic agents to slow bowel motility may help improve absorption, thus decreasing diarrhea. Fiber supplementation may improve diarrhea, although fiber’s effect on diarrhea remains controversial. Studies have provided evidence for and against fiber’s efficacy in treating diarrhea associated with tube feeding.176,177 There are a number of fiber-supplemented commercial enteral formulas available. Some fiber contained in enteral formulations may be termed a prebiotic. Metabolic complications are less common with enteral than with parenteral feeding. Dehydration and fluid shifts may occur with formulas of high concentration, especially if insufficient water is supplied. Hyperglycemia may occur with high rates of carbohydrate delivery in patients with glucose intolerance. Medication delivery also may be affected by concurrent tube feeding. Phenytoin administration is affected because phenytoin binds to the enteral formula178 and forms a phenytoin–tube feeding complex that adheres to the wall of the feeding tube. Ciprofloxacin also has been shown to bind with tube feedings, reducing its absorption. Vitamin K, present in many enteral formulas, may make a patient more resistant to the effects of warfarin.

ORAL DIET THERAPY

A general diet is designed to provide optimal nutrition to patients who do not require a therapeutic diet. A healthful diet contains a variety of foods that are low in fat and cholesterol, contain moderate salt content, and has an abundance of fruits, grains, and vegetables. Low-carbohydrate diets have been promoted for weight loss,179 but the longterm outcomes of this dietary regimen for weight loss vary significantly among individuals (see Chapter 6).

carbohydrates and should be used with caution in diabetic patients. Soft Diets.  These are designed for patients who cannot tolerate a regular diet, usually because of an oral, pharyngeal, or esophageal anatomic lesion, such as pharyngeal or esophageal cancer. Soft diets are often used in a progression from a liquid to solid diet. Soft diets are believed to reduce gas and nausea in postoperative patients, although there are no data supporting this concept. For patients with poor dentition, a soft diet can provide adequate calories, protein, and nutrients without having to rely on any significant mastication. Fiber- and Residue-Restricted Diets.  These are used for patients with gastrointestinal strictures and are presumed to reduce the risk of obstruction while prolonging transit time.181 Carbohydrate intake is reduced and well-cooked vegetables, refined cereals, and breads are used. Although these diets are commonly prescribed, there are no data supporting their use in the gastrointestinal stricture patient population. High-fiber diets include soluble and insoluble fibers, which have a wide range of metabolic and physiologic effects. They are used to reduce intraluminal colon pressures in patients with diverticulosis, although no data support this concept. They also may be useful in diabetes by delaying glucose absorption, in cardiovascular disease by lowering serum cholesterol and serum triglyceride levels, and in the prevention of colon cancer. This diet emphasizes foods such as vegetables, fruits, legumes, and whole-grain breads and cereals. Postgastrectomy Diets.  These supply enough calories and nutrients to promote tissue healing while avoiding the dumping syndrome. Liquids are avoided during meals and simple carbohydrates are restricted. A diet high in fat and proteins should be provided and small, frequent feedings promoted. Post–Gastric Bypass Diets.  These focus initially on allowing sips of oral nutritional supplements, with additional vitamin and mineral supplements. Over time, the patient progresses to small, frequent, regular meals. The limitation in oral intake is related to the size of the gastric pouch. Lactose-Restricted Diets.  These are designed to avoid the bloating, flatulence, diarrhea, and cramping associated with the ingestion of dairy products. They are used in patients who are lactose-intolerant and avoid foods containing milk or other lactose-containing dairy products, such as breads, candies, cold cuts, salad dressings, sugar substitutes, instant drink mixes, and commercial sauces or gravy.

Clear Liquid Diets.  These supply fluid and energy in a form that creates a minimal amount of residue. They are meant to avoid a high osmolar delivery to the gastrointestinal tract, which would result in fluid shifts and associated nausea and diarrhea. Clear liquid diets generally contain an abundance of carbohydrates but little protein or fat. There is but sparse evidence to suggest that a clear liquid diet is better tolerated than any other diet following surgery. It is known that early feeding after abdominal or thoracic surgery reduces postoperative complications and hospital length of stay.180

Low-Fat Diets.  These are used to minimize diarrhea and steatorrhea associated with fat malabsorption, especially in patients with pancreatic or biliary dysfunction. In patients who are to be on low-fat diets for prolonged periods, fatsoluble vitamins (A, D, E, K) need to be supplemented. MCTs may be used to substitute for some long-chain triglycerides. MCTs have 6- to 12-carbon fatty acid chains and high aqueous solubility, and do not require bile salts for absorption in the small intestine.182 Unfortunately, most MCTbased products are not very palatable.

Full Liquid Diets.  These are indicated for patients who are unable to chew, swallow, or digest solids. They are largely milk-based and should not be used for lactoseintolerant patients. They contain a large amount of simple

Diabetic Diets.  These are meant to maintain the blood glucose level as close to normal as possible and also to control serum lipid levels. Calories are provided to maintain a reasonable body weight. Carbohydrates represent 55% to

95

96

Section II  Nutrition in Gastroenterology 60% of total calories. Complex unrefined carbohydrates serve as the bulk of the fiber in the diet. Protein represents 12% to 20% of total calories. Fat is restricted because of related cardiovascular disease and should be less than 25% to 30% of total calories.183 Specialty Diets.  Specialty diets for celiac, renal, and liver disease have been discussed earlier. Consultation with a dietitian or nutrition support team can be crucial to the effective dietary management of these diseases.

PROBIOTICS

Probiotics are microorganisms that on ingestion may help improve microbial balance and the health of the host. Probiotics must survive passage through the gastrointestinal tract and have the ability to adhere to human epithelial cells to populate the intestine. Three basic mechanisms are believed responsible for the action of probiotics: 1. Alteration of the microbial flora of the bowel 2. Improvement of intestinal barrier function 3. Modulation of the immune function of the host Probiotics adhere to the colon mucosa, thereby reducing sites at which pathologic bacteria can attach. Probiotics also may release bacteriocins that are detrimental to pathogenic bacteria, but not to the organisms of the ingested probiotic.184 Probiotics have been shown to decrease the pro­ duction of inflammatory cytokines, such as tumor necrosis factor-α, interferon-γ, and interleukin-1, and to increase the production of the anti-inflammatory cytokines interferon-α and interleukin-10.185 Probiotics enhance the intestinal barrier function by binding with Toll-like receptors on the epithelial cells and activating protein kinase C, which results in tightening of the tight junctions of the epithelial cells.186 Probiotics mainly have been used as treatment in four areas of gastrointestinal disease: symptomatic diverticular disease and recurrent diverticulitis (see Chapter 117), irritable bowel syndrome (IBS; see Chapter 118), inflammatory bowel disease (see Chapters 111 and 112), and infectious diarrhea (see Chapter 104).

Diverticulitis

The theory behind the prophylactic or active treatment of diverticulitis with probiotics hinges on the concept that an altered microflora in an area of diverticulosis can set up an area of chronic inflammation and that probiotics can be used to alter this flora and reduce the recurrence rate of diverticulitis. One study using the Nissle 1917 strain of E. coli has demonstrated a reduction in diverticulitis episodes in patients with a history of relapsing diverticular disease.187 A study by Tursi and colleagues has compared the relapse rate of symptomatic diverticular disease in 90 patients randomized to mesalamine, Lactobacillus casei, or mesalamine and L. casei. The group receiving the combination of mesalamine and L. casei had significantly fewer recurrences of symptomatic diverticular disease than the group randomized to mesalamine or L. casei alone.188

Irritable Bowel Syndrome

The anaerobic breakdown of carbohydrates and protein by bacteria is known as fermentation. Fermentation is very active in the ascending colon and transverse colon and minimally active in the left colon.189 It is believed that fermentation substrates, especially hydrogen and carbon dioxide gas, are responsible for some of the cramping and pain associated with IBS. In an open label trial, Lactobacillus plantarum DSM9843 was administered to patients with IBS, with a reduction of flatulence and abdominal pain.190

Biopsies of the rectum were able to recover the probiotic in 34% of samples whereas culture of the feces recovered the probiotic in 83% of samples. In another open-label trial, VSL-3 (Lactobacillus, Bifidobacterium, and Streptococcus) was able to reduce pain in 81% of patients with IBS. In a prospective randomized trial, Saggioro evaluated 40 patients with IBS given probiotic A (Lactobacillus plantarum and Bifidobacterium breve), probiotic B (Lactobacillus plantarum and Lactobacillus acidophilus), or placebo.191 Both probiotic combinations A (49%) and B (45%) improved pain compared with placebo (29.5%) at four weeks. These findings were statistically significant.

Inflammatory Bowel Disease

In three controlled trials, capsules containing a nonpathogenic strain of E. coli (Nissle 1917) was shown to be as effective as mesalamine for maintenance of remission in patients with ulcerative colitis (UC).192,193 Total proctocolectomy with ileoanal anastomosis is a common surgical procedure for UC that may be complicated by pouchitis, an inflammatory process of the ileal reservoir. In one study, 40 consecutive patients were randomized to VSL-3 or a placebo immediately after this surgical procedure and treatment was continued for one year.194 Of patients on VSL-3, 19% had an occurrence of pouchitis in the first year following surgery compared with 40% treated with placebo (P < 0.05). More studies of this and other probiotics are indicated in IBD and for the postoperative management of these patients.

Infectious Diarrhea

The rationale for using probiotics in infectious diarrhea is that they act against enteric pathogens by competing for available nutrients and binding sites. They make the intestinal contents acidic, produce a number of chemicals that are bacteriostatic, and increase the specific and nonspecific immune systems. Allen and associates have reviewed the effect of probiotics in 23 studies of infectious diarrhea evaluating 1917 participants,195 and showed that probiotics reduce the risk of developing diarrhea after an initial exposure and reduce the duration of diarrhea in those with infectious diarrhea.

KEY REFERENCES

Abou-Abassi S, Craig K, O’Keefe SJD. Hypocaloric jejunal feeding is better than total parenteral nutrition in acute pancreatitis: Results of a randomized comparative study. Am J Gastroenterol 2002; 97:225662. (Ref 40.) American Society for Parenteral and Enteral Nutrition. Guidelines for the use of parenteral and enteral nutrition in adult and pediatric patients. JPEN J Parenter Enteral Nutr 2002; 26(Suppl):1SA-138SA. (Ref 71.) Bistrian BR. Jonathan Rhoads lecture. Clinical aspects of essential fatty acid metabolism. JPEN J Parenter Enteral Nutr 2003; 27:168-75. (Ref 19.) Buchman A. Total parenteral nutrition–associated liver failure. JPEN J Parenter Enteral Nutr 2002; 26(Suppl):S43-8. (Ref 79.) Crook MA, Hally Panteli JV. The importance of refeeding syndrome. Nutrition 2001; 17:632-7. (Ref 75.) DeLegge MH. Enteral access: The foundation of feeding. JPEN J Parenter Enteral Nutr 2001; 25(Suppl):S8-13. (Ref 116.) DeLegge MH, Alsolaiman MM, Barbour E, et al. Short bowel syndrome: Parenteral nutrition versus intestinal transplantation. Where are we today? Dig Dis Sci 2007; 52:876-92. (Ref 82.) DeLegge MH, Drake LM. Nutritional assessment. Gastroenterol Clin North Am 2007; 36:1-22. (Ref 9.) DeMaria EJ, Murr M, Byrne TK, et al. Validation of the obesity surgery mortality risk score in a multicenter study proves it stratifies mortality risk in patients undergoing gastric bypass for morbid obesity. Ann Surg 2007; 246:578-82. (Ref 64.) Floch MH, White J. Diverticulitis: New concepts and new therapies. J Clin Gastroenterol 2005; 39:355-6. (Ref 53.)

Chapter 5  Nutrition in Gastrointestinal Diseases Garrow D, Pride P, Moran W, et al. Feeding alternatives in patients with dementia: Examining the evidence. Clin Gastro Hep 2007; 5:1372-8. (Ref 131.) Glynn CC, Greene GW, Winkler MF, Albina JE. Predictive versus measured energy expenditure using limits-of-agreement analysis in hospitalized, obese patients. JPEN J Parenter Enteral Nutr 1999; 23:147-54. (Ref 14.) McClave SA, DeMeo MT, DeLegge MH, et al. North American summit on aspiration in the critically ill patient: Consensus statement. JPEN J Parenter Enteral Nutr 2002; 26:S80-5. (Ref 97.) McClave SA. Lukan JK. Stefater JA. Poor validity of residual volumes as a marker for risk of aspiration in critically ill patients. Crit Care Med 2005; 33:324-30. (Ref 169.)

Morrill AC, Chinn CD. The obesity epidemic in the United States. J Public Health Policy 2004; 25:353-66. (Ref 59.) Preiser JC, Devos P, Van den Berghe G. Tight control of glycaemia in critically ill patients. Curr Opin Clin Nutr Metab Care 2002; 5:533-7. (Ref 67.) Scolapio JS. Effect of growth hormone, glutamine and diet on body composition in short bowel syndrome. JPEN J Parenter Enteral Nutr 1999; 23:309-12. (Ref 34.) Full references for this chapter can be found on www.expertconsult.com.

97

CHAPTE R

6

Obesity George A. Bray

CHAPTER OUTLINE Definition  100 Body Mass Index  100 Central Adiposity  100 Prevalence and Costs  100 Causes  101 Energy Imbalance: Epidemiologic Model  101 Environmental Agents  101 Host Agents  102 Pathology and Pathophysiology  103 Fat Cell as an Endocrine Cell  103 Visceral Fat  103 Complications and Consequences  103 Death  103 Diseases and Disorders  104 Benefits of Weight Loss  106

Obesity is a chronic disease that is increasing in prevalence. Both the World Health Organization1 and the National Heart, Lung, and Blood Institute2 of the U.S. National Institutes of Health have labeled obesity an epidemic. More than 30% of adult Americans are now obese3 and the prevalence of obesity in children and adults has increased more than 50% in the past decade. The progress of this epidemic is shown in maps indicating obesity rates in all states and regions of the United States. In 1991, 18% of the states (9/50) had obesity rates exceeding 15%, and by 1998 this rate had increased to 78% (39/50; Fig. 6-1). This epidemic is a time bomb for the future development of diabetes and its many complications.4 As a disease, obesity has its pathology rooted in the enlargement of fat cells, the secretory products of which produce most of the pathogenic changes that result in the complications associated with obesity. The remaining changes are a consequence of the fat mass per se.4 Phy­ sicians and the health care system have two strategies to deal with this problem. First, we can prevent the development of obesity, or treat it before complications develop. Alternatively, we can wait until complications develop and then treat these. Many physicians would prefer to wait until the comorbidities arise, because there are more treatment options for diabetes, hypertension, and heart disease than for obesity. In one long-term trial, the incidence of new cases of diabetes was reduced to zero over two years in patients who lost weight and maintained a weight loss of 12% or more, compared with an incidence of 8.5% for new cases of diabetes in those who did not lose weight.5 One reason that most physicians are reluctant to treat patients for obesity is that they have a limited number of successful options. There are only two agents currently approved by the U.S. Food and Drug Administration (FDA) for long-term use, sibutramine and orlistat. As monotherapy, either agent can produce weight loss of 8% to 10%. To

Evaluation  106 Clinical History  106 Physical Examination  106 Laboratory Studies  107 Prevention  108 Treatment  108 Diets  108 Lifestyle Modification  109 Exercise  109 Pharmacotherapy  110 Surgery  113

achieve the maximal reduction in the incidence of diabetes, however, weight loss needs to exceed 12%, a goal that cannot be achieved easily with current therapy other than surgery. Obesity is a stigmatized disease.4 A common view shared by the public and by health professionals alike is that obese people are lazy and weak-willed. It is not unusual to hear statements such as, “If fat people just had will power, they would push themselves away from the table and not be so fat.” This view is supported by the perception in advertising that thin women are more attractive than full-figured women; the declining relative weights of center-fold models in Playboy and other publications and of women who are winners of the Miss America contest substantiate this view. Another issue that aggravates the problem of treating obesity is the negative perception that surrounds the use of appetite suppressants.4 Amphetamine, the first widely used weight loss drug, is addictive. This concern about addiction has been transferred to other drugs, whether warranted or not. With all treatments for obesity, weight loss slows and then stops. This so-called plateau effect arises when homeostatic mechanisms in the body come into play and stabilize weight, although at a lower level than the starting level. A similar phenomenon also is observed in the treatment of hypertension; when an antihypertensive drug produces a drop in blood pressure, there is a plateau at a new, lower level.3 The antihypertensive drug has not lost its effect when the plateau occurs, but its effect is being counteracted by physiologic mechanisms. In the treatment of obesity, this new plateau in body weight often is viewed as a therapeutic failure for the weight loss drug or other treatment. Cessation of weight loss often prompts patients to think that they are “cured” and they stop treatment only to regain weight. Finally, many treatments for obesity have provided unwanted side effects.4 A recent example is what befell

99

100

Section II  Nutrition in Gastroenterology

Figure 6-1.  Prevalence of obesity in adults ages 20 to 74 years and of overweight children from 1960 to 2006 using data from the National Center for Health Statistics. Body weight and height were measured. Obesity in adults is defined as a body mass index (BMI) >30 kg/m2. In children, overweight is >95th percentile of weight for height, which is equivalent to a BMI >30  kg/m2. The percentage of obese adults older than 20 years has been estimated since 1991, and these maps show the changing pattern. There has been a rapid and impressive increase in the percentage of obese people, both men and women, in the United States. Data of the Behavioral Risk Factor Surveillance System. (From the Centers for Disease Control and Prevention. U.S. obesity trends 1985-2007. Available at http://www.cdc.gov/nccdphp/dnpa/ obesity/trend/maps/index.htm.)

No data

1991

1996

2003

2006

<10%

many patients who took the combination of fenfluramine and phentermine. The aortic regurgitant lesions that occurred in up to 25% of the patients treated with this combination of drugs led many physicians to say “I told you so” and “I’m certainly glad I didn’t use those drugs.” Much of this suspicion will subside with time, but there will remain residual concerns about such potential side effects of antiobesity agents among physicians, government regulators, and the public.

DEFINITION BODY MASS INDEX

Throughout the past 50 years, there has been a steady rightward, upward shift in the distribution curve for body weight. This trend can be traced most effectively using the body mass index (BMI), defined as the weight in kilograms divided by the height (in meters) squared [W/(H)2], which provides a useful operating definition of overweight. A normal BMI is between 18.5 and 25 kg/m2. A BMI between 25 and 29.9  kg/m2 is operationally defined as overweight, and individuals with a BMI higher than 30 kg/m2 are obese, with special consideration for muscle builders and other resistance-trained athletes. Above a BMI of 30  kg/m2, BMI is specific, but not very sensitive when compared with body fat for determining obesity.6 Values below a BMI of 30 kg/m2 are less specific, but still represents a good starting point in evaluating overweight people. BMI also provides a good risk measure for obesity.1,2,4

CENTRAL ADIPOSITY

The waist circumference is a practical measure of central adiposity that is a surrogate for more precise measures of visceral fat, such as computed tomography (CT) or magnetic

10%–14%

15%–19%

20%–24%

≥25%

≥30%

resonance imaging (MRI) scans of the abdomen. When BMI and waist circumference were used to predict the risk of hypertension, dyslipidemia, and the metabolic syndrome, the waist circumference was shown to be a better predictor than the BMI.7

PREVALENCE AND COSTS Using the BMI, it is clear that obesity is increasing in prevalence. This increase began in the 1980s and continues at present, although recent data suggest that it may be abating somewhat.3 Obesity affects children as well as adults. We are now seeing a rise in the prevalence of type 2 diabetes in adolescents that is directly related to the rising prevalence of obesity. Obesity has a higher prevalence in Hispanic and African American populations.8,9 Both mean height and weight increased in adults ages 20 to 74 years between 1960 and 2002. Men increased in height from a mean of 68 inches (172.7 cm) to 69.5 inches (176.5 cm) and women from a mean of 63 to 64 inches (160 cm to 162.6 cm) during this period. For men, weight rose from a mean of 166.3 to 191 pounds (75.4 to 86.8 kg) and for women from a mean 140.2 to 164.3 pounds (63.7 to 74.6  kg), for an average increase of BMI from 25.2 to 28 kg/m2 for men and from 24.8 to 28.2  kg/m2 for women during this 42-year period. The increase in weight was greater in older men than in younger men, but the reverse was true for women, with older women gaining less weight than younger women. Similar effects are seen in children, with the weight of 10-year-old boys rising from a mean of 74.2 pounds (33.7 kg) in 1963 to 85 pounds (38.6 kg) in 2002, and of 10-year-old girls rising from a mean of 77.4 pounds (35.2  kg) to 88 pounds (40  kg) in this same interval. These increases in weight were associated with increases in BMI for boys and

Chapter 6  Obesity girls. For 7-year-old boys, BMI increased from a mean of 15.8 to 17.0 kg/m2 between 1963 and 2002 and for 7-year-old girls it rose from a mean of 15.8 to 16.6 kg/m2. For 16-yearold boys, it rose from 21.3 to 24.1 kg/m2 and for girls, from 21.9 to 24.0 kg/m2 in this same interval. Obesity is expensive, representing between 3% and 8% of health care budgets in various countries.4,10 Hospital costs and use of medication also escalate with increasing BMI. In a large health maintenance organization, mean annual costs were 25% higher in participants with a BMI between 30 and 35 kg/m2 and 44% higher in those with a BMI greater than 35  kg/m2 than in those with a BMI between 20 and 25 kg/m2.11 Costs for lifetime treatment of hypertension, hypercholesterolemia, type 2 diabetes, heart disease, and stroke in men and women with a BMI of 37.5  kg/m2 was $10,000 higher than for men and women with a BMI of 22.5 kg/m2, according to data from the National Center for Health Statistics and the Framingham Heart Study.12

CAUSES

glucocorticoids (Table 6-1). Although neurotropic viruses can produce obesity in experimental animals, the most intriguing human example is the presence of antibodies to adenovirus 36 in some obese people, and the fact that this virus can produce obesity in marmosets, a nonhuman primate. Exposure to estrogen-like compounds in utero may enhance the risk of obesity later in life. In affluent Western societies, foods, particularly foods high in fat, are abundant. In addition, portion sizes have increased, providing more energy to people with each portion. Toxins are an interesting potential group of agents for which more research is needed. Viruses are known to produce obesity and their potential role in obesity needs to be studied further. Physical activity may have gradually decreased, thereby reducing energy expenditure. Some have described the current environment as a virulent or toxic environment that has heightened the risk for obesity for people who are genetically susceptible to becoming obese.15 For the genetically susceptible host, this excess of food energy, environmental toxins, and viruses, along with a reduced level of physical activity, may lead to an accumulation of fat in fat cells.

ENVIRONMENTAL AGENTS

ENERGY IMBALANCE: EPIDEMIOLOGIC MODEL

We become obese because over an extended period of time, we ingest more carbon- and nitrogen-containing compounds as food than we need for daily energy expen­ diture. We and other animals thus obey the First Law of Thermodynamics, which describes this energy imbalance. Unfortunately, however, this important law of nature fails to inform us about such important issues as how food intake is regulated, where fat is stored, why men store fat differently than women, and how genes control these processes.4,13 An epidemiologic model is a good way to look at energy balance and to conceptualize obesity as a disease.14 In an epidemiologic model, environmental agents act on a host to produce a disease (Fig. 6-2). Disease is a function of the virulence of the agent and susceptibility of the host. For obesity, the environmental agents include drugs, food, toxins, physical inactivity, viruses, and other people.4,13,15 Among the important drugs that cause weight gains are some antipsychotic drugs (e.g., clozapine, olanzapine), some antidepressants (e.g., amitriptyline), some antidiabetic drugs (insulin, sulfonylureas, and thiazolidinediones), and

Intrauterine Factors.  Several environmental intrauterine events influence postnatal weight and lifetime weight gain and fatness.4 These include maternal diabetes, maternal smoking,16 and intrauterine nutrition, all of which increase

Food

Drugs

Toxins

Host

Ease of inactivity

Obesity

Viruses

Figure 6-2.  Epidemiologic model of obesity. Environmental factors are shown interacting with the susceptible host to produce the disease called obesity.

Table 6-1  Drugs That Produce Weight Gain and Potential Alternatives Category

Drugs that Cause Weight Gain

potential Alternatives

Neuroleptics

Thioridazine, olanzepine, quetiapine, resperidone, clozapine

Molindone, haloperidol, ziprasodone

Amitriptyline, nortriptyline Imipramine, mitrazapine Paroxetine Valproate, carbamazepine, gabapentin Insulin, sulfonylureas, thiazolidinediones

Protriptyline Bupropion, nefazadone Fluoxetine, sertraline Topiramate, lamotrigine, zonisamide Acarbose, miglitol, metformin, pramlintide exenatide — Inhalers, decongestants Calcium channel blockers Angiotensin-converting enzyme inhibitors Barrier methods, nonsteroidal anti-inflammatory drugs

Antidepressants   Tricyclics   Monoamine oxidase inhibitors   Selective serotonin-reuptake inhibitors Anticonvulsants Antidiabetic drugs Antiserotonin agents Antihistamines β-Adrenergic blockers α-Adrenergic blockers Steroid hormones

Pizotifen Cyproheptadine Propranolol Terazosin Contraceptives, glucocorticoids, progestational steroids

101

102

Section II  Nutrition in Gastroenterology an individual’s risk for increased body weight and diabetes later in life. Offspring of mothers who smoke during pregnancy are at increased risk of weight gain in their first decades of life,16 as are infants of diabetic mothers and small-for-gestational-age infants. Neonatal Environment.  Infants who are breast-fed for more than three months may have a reduced risk of future obesity.17 Children who get more sleep tend to weigh less when they enter school than those who sleep less.4 Adiposity Rebound.  Adiposity rebound occurs in childhood at the age that the BMI stops falling and begins to rise. Early adiposity rebound predicts future obesity.18 Drug-Induced Weight Gain.  In our current practice of medicating much of society, it would not be surprising to find drugs that produce weight gain. Table 6-1 is a list of medications that produce weight gain when used to treat other diseases, such as psychosis, depression, allergies, and diabetes. Also listed in the table are possible alternatives that can be used to avoid such weight gain. In most cases, there are alternative strategies that can be used to treat a patient when weight gain is closely associated with the initiation of a new medication for one of these conditions. When the binding of drugs listed in Table 6-1 to receptors in the brain was examined, several—including the histamine (H1), α1Aadrenergic, and serotonin (5-HT2C and 5-HT6) receptors— could explain much of the differences in weight gain associated with the atypical antipsychotic drugs shown in Table 6-1.19 Diet.  Portion size, fat intake, and consumption of beverages sweetened with sucrose (table sugar) or highfructose corn syrup have all been implicated in the current obesity epidemic.4,20 Consumption of soft drinks predicted future weight gain in children and adults, and also may be related to the risk for the metabolic syndrome and gout.21 Physical Inactivity.  Low levels of physical activity, such as watching television, correlate with weight gain. In a 10-year study of individuals 20 to 74 years of age in the National Health and Examination Survey (NHANES I), those with low levels of recreational activity gained more weight than those with higher levels of activity.22 Similarly, low levels of baseline energy expenditure predicted weight gain in the Pima Indians,23 and exercise capacity and body composition predict mortality among men with diabetes.24 Time spent watching television correlates with percentage of overweight children4 and the more television watched, the greater the risk of overweight and obesity. Smoking.  Smokers have lower body weights than nonsmokers, and cessation of smoking generally is associated with weight gain.4,25 Two explanations have been offered for the effect of smoking on body weight. First, smoking is thermogenic; that is, the metabolic rate during the act of smoking is higher than when the subject is not smoking. Second, smoking reduces hunger and changes taste perceptions; smokers tend to eat less. Viruses.  One laboratory has reported that obese humans have higher antibody levels to one strain of adenovirus (AM-36).26 As noted, this virus can produce obesity when given to nonhuman primates and this viral antibody, as a marker of viral infection, is found in the circulation of many obese people.

HOST AGENTS Genetic Causes

Several genes have a strong relation to the development of obesity.27 The melanocortin-4 receptor gene, leptin gene, pro-opiomelanocortin (POMC) gene, and agouti gene all have significant effects on body fat and fat stores. There are five melanocortin receptors, MC4 and MC being primarily in areas of the brain that affect feeding. Genetic abnormalities in this receptor may account for up to 6% of cases in early-onset, severely obese children.28 Absence of leptin or an ineffective leptin receptor is associated with massive obesity in human beings and animals. Leptin has the dual effect of reducing food intake and increasing energy expenditure, both of which favor loss of body fat. Treatment of leptin-deficient children with leptin decreased their body weight and hunger, indicating the importance of leptin in normal subjects (see later, “Neurophysiologic Factors”). Heterozygotes for leptin deficiency have low but detectable serum leptin levels and have increased adiposity, indicating that low levels of leptin are associated with increased hunger and gain in body fat. Leptin also can increase energy expenditure. Thus, when leptin is given and caloric intake simultaneously is reduced, the leptin attenuates the decreases in thyroid hormones and 24-hour energy expenditure that result, because leptin normally falls with reduced caloric intake.29 There are several rare clinical syndromes of obesity with a genetic basis.30 The Prader-Willi syndrome is the most common.31 This disease is transmitted as an abnormality on chromosome 15 and is characterized by a floppy baby who has difficulty feeding. These children are mentally slow, short in stature, have hypotonia and hypogonadism, and are obese. The Bardet-Biedl syndrome can result from changes at many different genetic loci. These children have visual impairment caused by abnormalities in the retina, are mentally slow, and have increased numbers of digits (polydactylism). In at least one pedigree, the genetic defect was caused by a fault in the chaperonin-like gene, which produces a product involved in folding proteins in the reticular endoplasm.32 The epidemic of obesity is occurring on a genetic background that does not change as fast as the epidemic has been exploding. Nonetheless, it is clear that genetic factors play an important role in the development of obesity, for which more than 100 genes have so far been implicated.30

Neurophysiologic Factors

A number of peptides play an important role in the development of obesity.33 The discovery of leptin in 1994 opened a new window for understanding control of food intake and body weight. The response of leptin-deficient children to treatment with leptin revealed the critical role that this peptide plays in the control of energy balance.27 Leptin enters brain tissue, probably by transport across the bloodbrain barrier, where it acts on receptors in the arcuate nucleus to regulate, in a conjugate fashion, the production and release of at least four peptides.34 Leptin inhibits the production of neuropeptide Y (NPY) and agoutirelated peptide (AGRP), both of which increase food intake; it also enhances production of POMC, the source of α-melanocyte stimulating hormone (α-MSH), which reduces food intake. Three other brain peptide systems also have been linked to the control of feeding. Melanin-concentrating hormone (MCH) is found in the lateral hypothalamus and decreases food intake when injected into the ventricular system of the brain.35 Orexin (also called hypocretin) was identified in a search for G protein-linked peptides that affect food intake34;

Chapter 6  Obesity it increases food intake and plays a role in sleep. Endocannabinoids (anandamide and 2-arachidonoyl glycerol) also increase food intake by acting on cannabinoid-1 (CB1) receptors. There are two cannabinoid receptors that originally were identified when tetrahydrocannabinol, the active ingredient of marijuana, was used to identify its endogenous receptor. These receptors are located in the brain and in many peripheral tissues. An antagonist to the CB1 receptor has served as the basis for a new antiobesity drug.36 Intestinal peptides, including cholecystokinin, pancreatic polypeptide, and polypeptide YY, reduce food intake,4 whereas ghrelin, a small peptide produced in the stomach, stimulates food intake.33 Metabolism of fatty acids in the brain may be another important control point. An experimental chemical that blocks fatty acid synthase in the brain was noted to result in significant weight loss and accumulation of malonyl coenzyme A. In pursuing this system, 5′-adenosine monophosphate kinase (AMPK), an enzyme that is activated or inhibited in relation to the ratio of adenosine monophosphate (AMP) to adenosine triphosphate (ATP), was thought perhaps to be the underlying central point in this control system.37

Genes

Environment

↓ Activity

↑ Food intake

Excess fat stores

Diseases due to increased fat cell size Diabetes mellitus

NAFLD

GB disease

CVD

Cancer

Diseases due to increased fat mass Social stigmatization

Osteoarthritis

Sleep apnea

Figure 6-3.  Pathogenesis of health problems associated with obesity. The mass of fat and the responses to products produced by fat cells can explain most of the diseases that result from prolonged obesity. CVD, cardiovascular disease. GB, gallbladder; NAFLD, nonalcoholic fatty liver disease.

PATHOLOGY AND PATHOPHYSIOLOGY The pathology and pathophysiology of obesity lie in the changes in the fat cells that store fat. Enlarged fat cells are the hallmark of this process. These enlarged cells produce effects through their increased mass, which increases the wear and tear on joints and makes overweight individuals obvious candidates for stigmatization. In addition, enlarged fat cells produce many adipokine products that affect distant cells. In addition to enlarged fat cells, some individuals also have an increased number of fat cells.14,38

FAT CELL AS AN ENDOCRINE CELL

Two mechanisms are integral to understand the pathophysiology of health problems associated with obesity. The first is the increased fat mass, which explains the stigmatization of physically obvious obesity and accompanying osteoarthritis and sleep apnea (Fig. 6-3).14 The second is the consequences of the increased number of peptides produced by the enlarged fat cells that act on distant organs. The discovery of leptin catapulted the fat cell into the arena of endocrine cells. In addition to leptin, fat cells secrete, for example, cytokines, angiotensinogen, adipsin (complement D), and metabolites, such as free fatty acids and lactate. Angiotensinogen is an enzyme involved in the control of blood pressure. Adipsin was one of the first substances secreted by fat cells to be identified, and turned out to be a component of the coagulation system. In contrast to these fat cell products, the release of adiponectin, the most abundant peptide produced by fat cells, is decreased in obesity.39 High levels of adiponectin are associated with insulin sensitivity and low levels with insulin resistance. These products of the fat cell in turn modify the metabolic processes of other organs in the host. For the susceptible host, these metabolic changes lead in turn to hyperinsulinemia, atherosclerosis, hypertension, and physical stress on bones and joints.

VISCERAL FAT

A considerable body of data suggests that visceral fat has a stronger relationship with the complications associated with obesity than does total body fat.7 Moreover, central

Table 6-2  National Cholesterol Education Program Adult Treatment Panel (ATP) III Modified Criteria for Metabolic Syndrome* parameter Waist circumference   Female   Male HDL cholesterol   Female   Male Fasting glucose Triglycerides Blood pressure

ATP III Modified Criterion >35 inches (>88 cm) >40 inches (>102 cm) <50 mg/dL (<1.29 mmol/L) <40 mg/dL (<1.03 mmol/L) ≥110 mg/dL (≥6.2 mmol/L) ≥150 mg/dL ≥130/85 mm Hg

*The metabolic syndrome is present when any three of the five criteria listed in the table are abnormal. HDL, high-density lipoprotein. Adapted from Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA 2001; 285:2486-97; and Park YW, Zhu S, Palaniappan L, et al. The metabolic syndrome: Prevalence and associated risk factor findings in the U.S. population from the Third National Health and Nutrition Examination Survey, 1988-1994. Arch Intern Med 2003; 163:427-36.

adiposity is one of the key components of the metabolic syndrome, the diagnostic criteria of which are based on the recommendations of the National Cholesterol Education Program Adult Treatment Panel III (Table 6-2).40,41

COMPLICATIONS AND CONSEQUENCES DEATH

Mortality and BMI have a J-shaped relationship in essentially all studies.4 Among the more than 90,000 women in the Women’s Health Initiative, there was a graded increase

103

104

Section II  Nutrition in Gastroenterology in the risk of death as BMI increased from normal to a BMI > 40 kg/m2.42 In the more than 527,000 men and women in the NIH-AARP American cohort of individuals ages 50 to 71 years, the risk of death was increased both in those who were overweight and in those who were obese.43 In an even larger Korean study, both overweight and obesity in men and women were related to higher death rates compared with normal-weight subjects.44 Another American cohort of more than 80,000 men and women was monitored for over 14.7 years and over 1.23 million person-years of follow-up. Excluding the first five years of death in men and women, those younger than 55 years showed a risk of death that was directly related to BMI, beginning at a BMI of 21.0  kg/m2 in women and 23.0  kg/m2 in men; in those older than 55 years, the increase in mortality occurred at a higher BMI (25 kg/m2 in women and 30 kg/m2 in men).45

DISEASES AND DISORDERS Disorders Related to Enlarged and Visceral Fat Cells

Excess body fat, particularly visceral fat, increases the risk for a number of diseases as a metabolic consequence of the enlarged fat cells or as a result of the increased mass of fat.4,14,46 Diabetes The risk of diabetes rises as BMI increases and is particularly steep when the BMI is more than 30  kg/m2. Weight gain in middle age, independent of attained weight, increases the risk of impaired glucose tolerance45 and increases the risk of heart disease. Blood pressure increases linearly with BMI and hypertension is present in approximately half of very obese subjects at initial evaluation.47 In the Nurses Health Study, the BMI values at age 18 years and at midlife were positively associated with the occurrence of hypertension.4 Lipid Derangements Dyslipidemia, characterized by a low high-density lipoprotein (HDL) cholesterol and high triglyceride level, is more common in obesity, particularly with central adiposity48 and, when accompanied by hypertension and an elevated serum glucose level, meet the National Cholesterol Education Program criteria for the so-called metabolic syndrome. A meta-analysis of 21 cohort studies has suggested that the adverse effects of obesity on blood pressure and lipids account for approximately half of the excess risk of coronary heart disease.49 Cardiovascular Diseases Because coronary heart disease accounts for nearly half of all deaths in our society, its relationship to obesity is particularly important.50 In one study, an increase in BMI of 1.1 kg/m2 increased the risk for major cardiovascular disease by 6%.51 Obesity also increases the risk of congestive heart failure52 and atrial fibrillation.53 Much of this increased risk of heart disease is associated with central adiposity.7 The INTERHEART study of patients from 52 countries showed that abdominal adiposity accounts for 20% of the population’s attributable risk for a first myocardial infarction.54 Hypertension Blood pressure often is increased in overweight individuals.55 In the Swedish Obese Subjects Study,48 hypertension was present at baseline in 44% to 51% of subjects. For each decline of 1 mm Hg in diastolic blood pressure, the risk of myocardial infarction decreased an estimated 2% to 3%.

Obesity and hypertension interact with cardiac function. In overweight individuals, ventricular eccentric dilation occurs, whereas hypertension in normal-weight people produces concentric hypertrophy of the heart, with uniform thickening of ventricular walls. Increased preload and stroke work are associated with hypertension. The combination of overweight and hypertension leads to thickening of the ventricular wall and larger heart volume, and thus to a greater likelihood of cardiac failure. Kidney Disease Obesity may affect the kidney in several ways. First, an obesity-related glomerulopathy characterized as focal segmental glomerulosclerosis has increased significantly, from 0.2% of biopsies collected between 1986 and 1990 to 2.0% of specimens taken between 1996 and 2000.56 Second, overweight patients also are at increased risk for kidney stones.57 Finally, BMI is related to the risk of end-stage renal disease. In a study from the Kaiser Permanente Group of Northern California, Hsu and colleagues58 found that a higher BMI is a progressively greater risk factor for end-stage renal disease that persists even after correcting for multiple potential confounding factors, including baseline blood pressure or diabetes mellitus. Gallbladder Disease Cholelithiasis is the primary hepatobiliary pathology associated with overweight,59 part of the explanation for which is increased cholesterol turnover related to total body fat.60 Cholesterol production is linearly related to body fat; approximately 20  mg of additional cholesterol is synthesized for each kilogram of extra body fat. Thus, a 10-kg increase in body fat leads to the daily synthesis of as much cholesterol as contained in the yolk of one egg. The increased cholesterol is in turn excreted in the bile, where high cholesterol concentrations relative to bile acids and phospholipids increase the likelihood of precipitation of cholesterol gallstones in the gallbladder (see Chapter 65). Additional factors, however, such as nidation conditions, also are involved in determining whether gallstones form.60 During weight loss, the likelihood of gallstone formation increases because the flux of cholesterol mobilized from fat is increased through the biliary system. Diets with moderate levels of fat that trigger gallbladder contraction and thus empty its cholesterol content may reduce this risk. Similarly, the use of bile acids, such as ursodeoxycholic acid, may be advisable if the risk of gallstone formation is thought to be increased. Liver Disease Nonalcoholic fatty liver disease (NAFLD) is the term given to describe a constellation of liver abnormalities associated with overweight, including hepatomegaly, elevated liver biochemical test results, and abnormal liver histology, including steatosis, steatohepatitis, fibrosis, and cirrhosis (see Chapter 85).61 NAFLD may reflect increased very lowdensity lipoprotein (VLDL) production associated with hyperinsulinemia. A study using a cross-sectional analysis of liver biopsies has suggested that in overweight patients, the prevalences of steatosis, steatohepatitis, and cirrhosis are approximately 75%, 20%, and 2% respectively.62 Data from the Homeostasis Assessment Model (HOMA), a means used to develop mathematical models to describe glucose regulation, have shown that the more marked the insulin resistance, the higher the prevalence of severe steatosis.63 Using ultrasound for diagnosing increased liver fat, Hamaguchi and colleagues64 found that in a Japanese population there was a 10% incidence of new cases of NAFLD after a

Chapter 6  Obesity mean follow-up of 414 days, and that this was predicted by the presence of the metabolic syndrome. If increased fat in the liver is suspected, an ultrasound of the liver can provide a quantitative estimate that is much better than serum liver biochemical test results. Gastroesophageal Reflux Disease Overweight also may be a contributing factor in gastroesophageal reflux disease (GERD) (see Chapter 43). A total of nine studies examined the association of GERD with BMI,65 six of which showed a statistically significant association. Erosive esophagitis and esophageal adenocarcinoma also were more common in obesity. The odds ratio for GERD was 1.43 in the overweight group (BMI, 25 to 29.9  kg/m2) compared with the normal-weight group and rose to 1.94 when the BMI was higher than 30 kg/m2.66 Cancer Certain forms of cancer are significantly increased in obesity.67-69 Obese men face increased risk for neoplasms of the colon, rectum, and prostate, whereas in women, cancers of the reproductive system and gallbladder are more common than in nonobese women. One explanation for the increased risk of endometrial cancer in overweight women is the increased production of estrogens by stromal cells in adipose tissue. This increased production is related to the degree of excess body fat and accounts for a major source of estrogen production in postmenopausal women. Breast cancer is not only related to total body fat, but also may have a more important relationship to central body fat,70 which may help explain why breast cancer risk is increased at age 75 in women in the highest versus the lowest quartile of BMI.71 Increased visceral fat as measured by CT shows an important relationship to the risk of breast cancer. The Nurses Health Study has added significant insight to the relationship of body weight and breast cancer. Women who gained 25  kg or more after age 18 were at increased risk of breast cancer (relative risk [RR], 1.45; P < 0.001), and women who gained 10 kg or more after menopause were at increased risk for breast cancer compared with women whose weight remained stable. Women who achieved and maintained a 10-kg or more weight loss and who did not use postmenopausal hormones were at lower risk than those who maintained a stable weight.72 Finally, a pooling project with data from 13 cohort studies found that the relative risk of renal cell carcinoma was increased to 2.10 in those with a BMI > 30  kg/m2 compared with those with a BMI < 23 kg/m2.73 Endocrine Effects A variety of endocrine changes are associated with obesity, including the polycystic ovary syndrome (PCOS), which is characterized by hirsutism, oligomenorrhea, and marked insulin resistance, hyperactivity of the adrenal glands (Cushing’s syndrome), and reduced fertility in men and women. In the Nurses Health Study, as BMI increased, the relative risk of infertility rose. Compared with the reference group, which had a BMI of 20 to 21.9  kg/m2, the relative risk of infertility was 1.7 for a BMI of 26 to 27.9 kg/m2 and 2.7 for a BMI above 30 kg/m2.74 Obesity influences the outcome of pregnancy. Increasing prepregnancy body weight was associated with a significant and weight-related increase in the likelihood of cesarean delivery. Infant pre-term birth weight was higher in smaller women. Heavier women had heavier babies, but no increased risk of low birth weight infants. Low birth weight infants were less likely in heavier women and also in those who gain more weight during their pregnancy.75 Weight gain of

more than18.6  kg (41 pounds) also increases the risk of cesarean delivery. The risk of postpartum urinary tract infection also appears to be increased in overweight women, based on an observational study of 60,167 women.76 In a large retrospective study from Scotland, nulliparous women, compared with multiparous women, had increased elective preterm delivery, neonatal death, and infant weights less than 1000 g, and these effects were greatest in women with a BMI > 35 kg/m2.77 Pneumonia Community-acquired pneumonia may be an additional risk related to being obese. In the Health Professionals Followup Study and the Nurses Health Study II, the risk of pneumonia increased as BMI increased.78 Significant weight gain in women after age 18 years also increased the risk of pneumonia.

Disorders Associated with Increased Fat Mass

Obstructive Sleep Apnea The chief disturbance of pulmonary function in obese individuals results from a decrease in residual lung volume because of increased abdominal pressure on the diaphragm,79 although fat distribution, independent of total fat, also affects ventilatory capacity. In contrast to the relatively benign effects of excess weight on respiratory function, overweight often is associated with sleep apnea, which can be severe and associated with significant reduction in nocturnal oxygen saturation.79,80 Sleep apnea is considerably more common in men than in women. People with sleep apnea have an increased snoring index and increased maximal nocturnal sound intensity. An interesting hypothesis is that the increased neck circumference and fat deposits in the pharyngeal area may lead to the obstructive sleep apnea of overweight. Diseases of the Bones, Joints, Muscles, Connective Tissue, and Skin Osteoarthritis is significantly increased in overweight individuals. The osteoarthritis that develops in the knees and ankles may be directly related to the trauma associated with the degree of excess body weight,81 but the increased osteoarthritis in other non–weight-bearing joints suggests that some components of the overweight syndrome alter cartilage and bone metabolism, independent of weight bearing. Increased osteoarthritis accounts for a significant component of the health cost of being overweight. Increased body weight also produces disability from joint disease. Using the Behavioral Risk Factor Surveillance System telephone survey data on individuals older than 45 years, Okoro and colleagues82 found that class 3 overweight (BMI > 40  kg/m2) was associated with disability in individuals who reported arthritis and those who did not report arthritis. Even lighter weight obese individuals had an increased likelihood of disability compared with normal-weight respondents.83,84 Rheumatoid arthritis and BMI have a paradoxical relationship. In a study of rheumatoid arthritis that accrued 123 deaths in 3460 patient-years of observation, the BMI was found to be inversely related to mortality, although the study period was relatively short and the number of subjects relatively small.85 Several skin changes are associated with excess weight.86 Stretch marks, or striae, are common and reflect the pressure on the skin from expanding lobular deposits of fat. Acanthosis nigricans refers to a deepening of the pigmentation in the folds of the neck, knuckles, and extensor surfaces that occurs in many overweight individuals. In normal-weight

105

106

Section II  Nutrition in Gastroenterology individuals, this can be a sign of increased risk of malignancy or insulin resistance but, when associated with obesity, such risks are unusual. Hirsutism in obese women may reflect altered increased androgen production, which can impair menstrual cycles and ovulation.4 Psychosocial Dysfunction Overweight is stigmatized in both children and adults87-89; that is, overweight individuals are exposed to the consequences of public disapproval of their fatness. The disapproval of obesity has, if anything, worsened over the past 40 years.90 Overweight children have a negative self-image and also a significant decrease in physical and social functioning compared with normal-weight children. A study using the Medical Outcomes Study Short-form Health Survey (SF-36) found that obese people had profound abnormalities in health-related quality of life.91 Overweight women appear to be at greater risk of psychological dysfunction, compared with overweight men, possibly because of increased societal pressures on women to be thin.92 A systematic review has shown that in four of eight studies that met criteria for inclusion in the investigation, obese subjects had an increased risk of dementia.93

BENEFITS OF WEIGHT LOSS

Weight loss improves a person’s health outlook in many ways.4 Weight loss reduces the risk of death in obese patients treated with bariatric surgery. In one study from Sweden94 with more than 4000 patients, half of whom received one of three bariatric operations (see later), there was a reduction in mortality of 24% after 10.9 years. Another study from Utah95 compared over 7000 patients who received a gastric bypass (see later) with those matched for weight and age and showed a reduction in mortality of 56% after 7.1 years.95 Weight reduction also reduces the risk from diseases that result from obesity. The Diabetes Prevention Program (DPP) is a clear example of a reduction in the risk of developing diabetes with weight loss. After an average of 2.8 years of follow-up in 3234 individuals with impaired glucose tolerance, those who were randomized to the intensive lifestyle treatment lost 7% of their body weight and showed a 58% reduction in the risk of developing diabetes.96 Blood pressure also benefits from weight loss. In the Framingham study, a modest weight loss of at least 6.8 kg or more led to a 28% reduction in the risk of hypertension in middle-aged adults and a 37% reduction in older adults.97 In a clinical trial using lifestyle interventions to lower blood pressure (TOHP II), the risk of being hypertensive was reduced 42% at 6 months and 22% at 18 months. In those who maintained a weight loss of 4.5 kg for 30 months, the risk of hypertension was reduced 65%.98 Apneic episodes also are influenced by changes in weight. Relative to stable weight, a 10% weight loss predicted a 26% reduction in the apnea-hyperpnea index.99 In a systematic review of long-term weight loss studies in obese adults, dietary and lifestyle approaches and pharmacologic interventions improved markers of cardiovascular disease, particularly in patients with cardiovascular risk factors at the beginning of the study.100 Quality of life also improves following weight loss.101

EVALUATION The hazards of excess weight and the benefits of weight loss point to the need for careful evaluation of the overweight patient. The National Heart, Lung, and Blood Institute

has provided an algorithm for evaluating the overweight patient. It is a useful framework for viewing the information that is collected during the evaluation of individual patients (Fig. 6-4). The basic components in the evaluation of any overweight or obese patient are a record of the historical events associated with the patient’s weight problem, a physical examination, and an appropriate laboratory assessment. Here, I have used the criteria recommended by the U.S. Preventive Services Task Force102 and also taken into account reports from the National Heart, Lung, and Blood Institute2 and the World Health Organization.1 The importance of evaluating overweight individuals has increased as the epidemic of overweight has worsened, and the number of patients potentially needing treatment has increased.

CLINICAL HISTORY

It is important to identify specific events associated with the increase in the patient’s body weight. Has there been a sudden increase in weight, or has body weight been rising steadily over a long period of time? Three categories of weight gain are identified: <5  kg (<11 pounds); 5 to 10  kg (11 to 22 pounds); and >10 kg (>22 pounds). In addition to total weight gain, the rate of weight gain after age 20 years needs to be considered when deciding the degree of risk for a given patient. Weight gain is associated with an increased risk to health and the more rapidly the patient is gaining weight, the more concerned the health care provider should be. Successful and unsuccessful weight loss programs undertaken by the patient also should be identified. A sedentary lifestyle increases the risk of early death. It is important to determine whether the patient comes from a family in which overweight is common, the usual setting, or whether she or he has become overweight in a family in which few people are overweight. The latter setting suggests a need to search for environmental factors that may be contributing to weight gain. Studies have shown that alteration in the melanocortin-4 receptor occurs in 2.5% to 5.5% of children and adolescents with a BMI > 30 kg/m2.28 This genetic defect is among the most common of those associated with any chronic disease, and evaluation for this defect may become important in the treatment of overweight people.

PHYSICAL EXAMINATION

The first step in the clinical examination of the overweight patient is to determine vital signs, which include BMI and waist circumference, as well as pulse and blood pressure.103 Accurate measurement of height and weight is the initial step in the clinical assessment,104 because these are needed to determine the BMI (see earlier). The BMI has a reasonable correlation with body fat and a curvilinear relationship to risk. Risk arbitrarily has been subdivided by cut points derived from data collected on whites. It is now clear, however, that different ethnic groups have different percentages of body fat for the same BMI105 and BMI thus needs to be interpreted in an ethnically specific context. An Asian Conference selected lower levels of BMI to define overweight (BMI < 23 kg/m2) and obesity (BMI > 25 kg/m2); the same BMI presumably carries a different level of risk in various populations. These differences need to be taken into consideration when making clinical judgments about the degree of risk for the individual patient. During weight loss, body weight is more useful than the BMI, because height doesn’t change during this period, and the need to use the height squared makes it more difficult for the physician and patient to calculate. After BMI, waist circumference is the second vital sign in the evaluation of the overweight individual. Waist

Chapter 6  Obesity Patient encounter Hx of BMI ≥25? No BMI measured in past 2 years?

Measure Wt, Ht, and waist circumference Calculate BMI

Yes

BMI ≥25 or waist circumference >88 cm (F) >102 cm (M)

Yes

BMI ≥30 or waist circumference Yes >88 cm (F) >102 cm (M) and >1 risk factor

Assess risk factors

No Hx of BMI ≥25?

No

Yes

Does patient want to lose wt?

No Reinforcement/ educate on weight management

Devise goals and strategy for weight loss and risk factor control

Advise to maintain weight/address other risk factors

Yes

No

Yes

Progress made/goal achieved? No

Examination Treatment

Periodic weight check

Maintenance counseling: Diet Behavior therapy Exercise

Assess reasons for failure to lose weight

Figure 6-4.  Algorithm for the diagnosis and treatment of obesity developed by the National Heart, Lung and Blood Institute (NHLBI). BMI, body mass index; Ht, height; Hx, history; Wt, weight. (From Carpenter KM, Hasin DS, Allison DB, Faith MS. Relationships between obesity and DSM-IV major depressive disorder, suicide ideation, and suicide attempts: Results from a general population study. Am J Public Health 2000; 90:251-7.)

circumference is the easiest measurement to evaluate central adiposity. It is determined using a metal or nonstretchable plastic tape. Measurements at the level of the umbilicus or at the midpoint between the lower rib and suprailiac crest are the two most common locations. Waist circumference is a good strategy for following the clinical progress of weight loss and is particularly valuable if patients become more physically active. Physical activity may slow loss of muscle mass and thus slow weight loss, whereas fat continues to be mobilized. Waist circumference can help in making these distinctions. The relationship of central fat to risk factors for health varies among populations as well as within them. Japanese Americans and Indians from South Asia have relatively more visceral fat and are thus at higher risk for a given BMI or total body fat than whites. Even though the BMI is below 25 kg/ m2, central fat may be increased, particularly in Asian populations, and may increase the risk of disease.106 Central adiposity is important, particularly with a BMI between 22 and 29  kg/m2. Blood pressure should be measured carefully. Hyper­ tension is amenable to improvement with diet107 and is an important criterion for diagnosis of the metabolic syndrome. The patient should sit quietly for 5 minutes before measuring the blood pressure with a calibrated instrument to increase the accuracy of measurement. The blood pressure criteria from the Seventh Joint National Commission recommendations should be followed. A normal blood pres-

sure is less than 120/80  mm Hg. Prehypertension was defined by this group as a systolic blood pressure (SBP) of 120 to 139 mm Hg and diastolic BP (DBP) of 80 to 89 mm Hg. Hypertension is then a SBP/DBP of 140/90 mm Hg and clearly needs treatment if such blood pressure values are confirmed. Individuals with prehypertension need to be carefully observed. Acanthosis nigricans (see earlier) in normal-weight individuals may signify increased insulin resistance or malignancy, but this is not usually the case in obesity. If this is suspected, however, further evaluation is necessary.

LABORATORY STUDIES

Serum lipids, glucose, C-reactive protein (now measured as high-sensitivity CRP [hs-CRP]), and other values indicated from the history and physical examination should be measured. An increased fasting glucose, low HDL cholesterol, and high triacylglycerol levels are atherogenic components of the metabolic syndrome. Along with elevated blood pressure, it is possible to categorize the patient as having the metabolic syndrome by using criteria proposed by the National Cholesterol Education Program (see Table 6-2). An individual has the metabolic syndrome if three of the five criteria are abnormal. Measurement of LDL cholesterol also is important because it may need treatment independently of obesity or central adiposity. A positive hs-CRP assay along with an elevated serum LDL cholesterol level is a clear risk factor for heart disease.

107

108

Section II  Nutrition in Gastroenterology PREVENTION Studies designed to prevent obesity have been conducted in children and in adults.4 For children and adolescents, many school-based programs have been tried and, although there have been some promising results, the long-term impact of such programs has been small. In one successful study in children, a reduction in television watching slowed their gain in BMI.108 Another study has shown that decreasing children’s consumption of carbonated beverages, primarily soft drinks, was associated with slower weight gain than that of children who were not given this advice.109 In studies involving adults, however, there are few successful preventive programs.

TREATMENT Realism is one important aspect of treatment for obesity. For most forms of treatment, including behavior therapy, diet, and exercise, weight loss levels off at less than 10% below baseline. For many patients this is a frustrating experience, because their dream weight would require a loss of almost 30% of their body weight.110 A weight loss of less than 17% would be a disappointment with participation in a weight loss program. Yet, other than surgery, a weight loss of 10% is the expected outcome. It is important for the patient and physician to realize that an initial weight loss of 10% of body weight should be considered a success and that this amount of weight loss lessens the health risks of obesity.100 Because obesity left to itself will lead to a number of associated diseases, there are two therapeutic strategies: (1) wait until associated diseases develop (e.g., diabetes, hypertension, or dyslipidemia) and treat them individually; or (2) treat the obesity itself, thus reducing the risk of developing diabetes, hypertension, and other associated diseases. The second approach is preferable.

DIETS

To lose weight, a person must consume fewer calories (eat less food) than the body needs for daily activities. Many diet plans are available for overweight individuals (Table 6-3).111 These can be categorized as those that are low in fat (Ornish, Jenny Craig), those that are low in carbohydrate (Atkins, South Beach), those that restrict most nutrients, the so-called balanced deficit diets (Weight Watchers, Volumetrics, Slim Fast, eDiets), those that highlight one type of food or another (e.g., the low glycemic

index diet), or diets that highlight special foods (e.g., the grapefruit diet). The efficacy of dietary counseling versus control therapy has been examined in a meta-analysis.112 A random effects model of 46 studies of dietary counseling showed a maximum net treatment effect of −1.9 BMI units (95% confidence interval [CI]: −2.3, −1.5) or approximately a 6% weight loss over 12 months. There was a loss of about 0.1 BMI unit/month for the 12 months of active treatment and a regain of about 0.02 to 0.03 BMI unit/month during subsequent phases of the program. There were many different strategies used in the studies reviewed in this meta-analysis, but there was no clear basis for selecting one dietary approach over another.

Popular Diets

Low-Fat Diets Low-fat diets are a standard strategy to help patients lose weight. One benefit from a very low level of fat intake is the slowing or reversal of coronary artery disease.113 A metaanalysis of five randomized controlled trials of low-fat diets, however, has shown that these diets produce significant weight loss, but no more so than the control diets.114 Subsequently, 48,835 women were randomly assigned in a large clinical trial to low-fat or control diets.115 Weight loss was 2.2 kg below baseline at year one and 0.6 kg at an average of 7.5 years of follow-up. At both time points, weight loss was significantly less in the women on the low-fat diet compared with those on the normal-fat diet, and there was a clear relationship between the decrease in percentage of fat ingested and weight loss (P < 0.001 for trend). A metaanalysis of weight loss studies has found that over the first six months, low-fat diets produce weight loss and that heavier individuals lose more weight than lighter weight individuals.116 Low Energy Density Diets The theory behind the use of low energy density diets117 is that filling the stomach with low-fat, high-fiber foods (low energy density) reduces hunger and produces satiety. Conversely, in experimental settings, people eat more food when it is more energy dense—that is, has more calories per unit weight. Low-Carbohydrate Diets The most popular diets are the low-carbohydrate, highprotein, high-fat diets. Daily carbohydrate intake in some of these diets is as low as 13 g; when carbohydrate intake is less than 50 g/day, ketosis develops. In short-term metabolic ward studies, patients did not increase the intake

Table 6-3  Nutritional Content of Several Popular Diets

Diet Atkins eDiets Jenny Craig Ornish Slim-Fast South Beach Volumetrics Weight Watchers

Protein (%)

Fat (%)

Saturated Fat (%)

29 24 20 16 21 22 22 20

60 23 18 6 22 39 23 24

20 5 7 1 6 9 7 7

Adapted from Rating the Diets from Atkins to Dr. Sears Zone. Consum Rep 2005; 70:18-22.

Carbohydrate (%)

Fiber (g/1000 kcal)

Daily Servings of Fruits and Vegetables

11 53 62 77 57 38 55 56

12 19 16 31 21 19 20 20

6 12 6 17 12 3 14 11

Chapter 6  Obesity of other foods to compensate for the lower calories in a very low carbohydrate diet.118 Several randomized clinical trials have reported greater weight loss in the low-carbohydrate diet group up to six months, but not at one year.119 Four clinical trials have compared the effect of macronutrient composition on weight loss in one-year studies120,121 and two-year studies.122,123 The two one-year studies compared the Atkins, Zone, and Ornish diets, with the Weight Watchers diet in one study120 or the LEARN manual in another study.121 One two-year study compared the Atkins diet, a Mediterranean-style diet, and the American Heart Association low-fat diet122 and the other compared four diets: 20% or 40% fat with 15% or 25% protein.123 In the first one-year study,120 160 participants were randomly assigned to each diet. After one year, there was no significant difference in the weight loss of patients using any of the four diets. The weight loss was as follows: −3.9 ± 6 kg (−8.58 ± 13.2 pounds) with the Atkins diet; −4.9 ± 6.9  kg with the Zone diet (CI, 10.78 ± 15.18 pounds); −4.6 ± 5.4 kg (−10.12 + 11.88 pounds) with the Weight Watchers diet; and −6.6 ± 9.3 kg (−14.52 ± 20.46 pounds) with the Ornish diet— thus, no differences among the diets. The principal determinant of weight loss was the degree of adherence to the diet, not the diet itself. In the second one-year trial, 311 premenopausal women were randomly assigned to one of four diets.121 In this study, the Atkins diet produced more weight loss at 12 months (−0.7 kg) compared with the other three diets (Zone, −1.6 kg; LEARN, −2.6 kg; Ornish, −2.2 kg), a difference that was not statistically different. Also in this study, adherence to the diet was the principal determinant of success. The first two-year study compared a Mediterranean diet, Atkins diet, and low-fat diet in a group composed of 90% men living in a town in Israel. At the end of two years, the weight loss was −4.7 kg for the low-carbohydrate group, −2.9  kg for the low-fat group, and −4.4  kg for the Mediterranean diet group. After reaching a plateau at six to seven months, the group on the Mediterranean diet had another drop in weight equal to that of the low-carbohydrate group.122 The second two-year study was the most complex, using a 2 × 2 design that had four carbohydrate levels (35%, 45%, 55%, and 65%) resulting from the two fat levels (20% or 40%) and the two protein levels (15% or 25%).123 In this study, 811 individuals were randomized to one of the four diets. The weight loss at one year averaged 7% across diets, with no significant differences. Thereafter, there was a small weight regain. Attendance at support, education, and diet groups strongly predicted success. The authors concluded that “the content of dietary fat, carbohydrate, and protein had little influence on body weight loss over two years in obese people.”

Very Low-Calorie Diets

Very low-calorie diets (i.e., diets with an energy level below 800  kcal/day), can be used for rapid weight loss prior to major surgery. In other settings, the weight rebound that usually occurs at the end of a program with very lowcalorie diets may not make them worth the effort for some people, and may deter them from using similar diets in the future. A systematic review of 29 studies of weight loss programs using a very low-calorie diet that lasted more than two years124 found that participants in the very lowcalorie diet program lost significantly more weight than those eating hypoenergetic balanced diets. The very lowcalorie diets, however, have been replaced largely by portion-controlled diets, in which the calories from bever-

ages, bars, or frozen meals provided at breakfast or lunch are fixed by the manufacturer. In a four-year study, this approach resulted in early initial weight loss, which then was maintained.125

Commercial Programs

A number of commercial and self-help programs, including Overeaters Anonymous, Take Off Pounds Sensibly (TOPS), Weight Watchers, Jenny Craig, Herbalife, OPTIFAST, LA Health, and eDiets, are available to the consumer. Tsai and Wadden126 have examined the effectiveness of a number of these programs. The Weight Watchers program is done in groups, in contrast to Jenny Craig and LA Weight Loss, in which clients are seen individually. Jenny Craig uses prepackaged food and Weight Watchers and LA Weight loss use diet plans. All three programs encourage physical activity. In one trial lasting two years and including 423 subjects, the participants in the intervention group attended the Weight Watchers meetings and experienced a mean weight loss of 5.3% at one year and 3.2% at two years, compared with 1.5% at one year and 0% weight loss for the control group that received the self-help intervention with two visits to a dietitian.127

LIFESTYLE MODIFICATION

A basic strategy in helping obese patients lose weight is through lifestyle changes. The first step in this process is to determine whether the individual is really ready to make lifestyle changes. Patients often have a dream weight that involves a weight loss of nearly 30% of their initial body weight.110 An initial loss of 5% to 10% of body weight is a more realistic goal, because it will significantly reduce many of the health hazards described above, if they are present.100 Behavioral strategies include helping patients learn to monitor their eating behavior by recording what is eaten, the setting in which it is eaten, and the situations that trigger eating. With this information, the health care provider can help a patient change his or her eating habits. Patients should be encouraged to use a defined eating plan. People who are successful in losing weight and maintaining weight loss tend to monitor their behavior, eat low-fat diets, increase their physical activity, and practice positive self-thinking and techniques for stress reduction, as documented by the National Weight Loss Registry.128 Use of the Internet is a promising new tool.129

EXERCISE

Exercise is one strategy for balancing energy intake and expenditure, whether as a primary treatment for weight loss or for prevention of weight regain. Walking expends approximately 100 kcal/mile. A deficit of 3500 kcal (500 kcal/day) maintained for one week should result in a loss of 0.45 kg (1 pound). To obtain this effect from exercise alone, an individual would need to walk five miles/day, seven days/ week. For this reason, exercise alone has not been very effective as a primary weight loss technique.4 Moderate to vigorous exercise for 60 min/day, six days/week, produced more weight loss (−1.4 kg in women; −1.8 kg in men) than that of a nonexercise group in over 12 months.130 A metaanalysis of weight loss trials lasting at least one year has found that exercise-alone groups had minimal weight loss. Use of resistance training, as opposed to aerobic exercise, may help retain lean body mass and reduce the associated fall in resting energy expenditure.131 For individuals wanting to monitor their exercise, inexpensive pedometers can be worn on the belt. A mile is about 2000 steps, and increasing the number of monitored

109

110

Section II  Nutrition in Gastroenterology Table 6-4  Drugs Approved by the U.S. Food and Drug Administration for the Treatment of Obesity Drug

Trade Name

Dosage

Pancreatic Lipase Inhibitor* Orlistat Xenical Norepinephrine-Serotonin Reuptake Inhibitor* Sibutramine Meridia (United States); Reductil (Europe and other countries) Noradrenergic Drugs† Diethylpropion Tenuate, Tepanil Tenuate Dospan Phentermine Adipex-P, Fastin, Phentercot Ionamin Benzphetamine Didrex Phendimetrazine Bontril XR, Plegine Plegine, Prelu-2, X-Trozine

DEA Schedule

120 mg three times daily before meals

—

5-15 mg daily

IV

25 mg twice daily 75 mg every morning 15 to 37.5 mg daily 15-30 mg daily 25-50 mg daily to three times daily 35 mg twice to three times daily 105 mg daily

IV IV III III

DEA, Drug Enforcement Administration. *Approved for long-term use; †Approved for short-term use.

steps walked each day is a good way to encourage walking.132

PHARMACOTHERAPY

Because all medications inherently have more risks than diet and exercise, medications should only be used when the benefits justify the risk. Current medications for the treatment of obesity can be divided into two broad categories: (1) those that act primarily on the central nervous system to reduce food intake; and (2) those that act primarily outside the brain. Wherever the primary site of action may be, however, the net effect must be a reduction in food intake, an increase in energy expenditure, or both. There currently are several drugs available in the United States to treat obesity133-135 (Table 6-4).

Mechanisms of Drug Action

The brain plays a central role in regulating food intake by receiving and processing information from the environment and internal milieu.34 A number of neurotransmitter systems, including monoamines, amino acids, and neuropeptides, are involved in modulating food intake. The monoamines include norepinephrine, serotonin, dopamine, and histamine, as well as certain amino acids. The serotonin system has been one of the most extensively studied of the monoamine pathways. Its receptors modulate both the quantity of food eaten and macronutrient selection. Stimulation of the serotonin receptors in the paraventricular nucleus reduces fat intake, with little or no effect on the intake of protein or carbohydrate. This reduction in fat intake is pro­bably mediated through 5-HT2C receptors, because its effect is attenuated in mice that cannot express the 5-HT2C receptor. Sibutramine blocks reuptake of serotonin and norepinephrine. Lorcaserin is a drug in clinical trials that acts directly on serotonin receptors in the brain. Stimulation of α1-adrenergic receptors also reduces food intake,134 as shown by the α1 agonist phenylpropanolamine. Some of the α1 receptor antagonists used to treat hypertension produce weight gain, further indicating a role for this receptor in weight control. In contrast, stimulation of α2adrenergic receptors increases food intake in experimental animals, and a polymorphism in the α2a adrenoceptor has been associated with reduced metabolic rate in humans. Activation of β2 receptors in the brain reduces food intake. These receptors can be activated by β agonists, which release norepinephrine in the vicinity of these receptors or block the reuptake of norepinephrine. Sibutramine is both a serotonin and norepinephrine reuptake inhibitor and also uses this mechanism.

Histamine receptors also can modulate feeding. Stimulation of the H1 receptor in the central nervous system reduces feeding. Experimentally, this has been addressed by modulating the H3 autoreceptor, which controls histamine release. When the autoreceptor is stimulated, histamine secretion is reduced and food intake increases. Blockade of this H3 autoreceptor decreases food intake. The histamine system is important in control of feeding because some psycho­ active drugs bind to histamine receptors and produce weight gain.19 The opioid receptors were the first group of peptide receptors shown to modulate feeding. They also modulate fat intake. Both mu and kappa opioid receptors can stimulate feeding. Stimulation of the mu opioid receptors increases the intake of dietary fat in experimental animals. Two other peptides, corticotropin-releasing hormone (CRH) and the closely related urocortin, reduce food intake and body weight in experimental animals. The endocannabinoid system is the most recent addition to the central controllers of feeding.36 Tetrahydrocanna­ binol, isolated from the marijuana plant, stimulates food intake. Isolation of the cannabinoid receptor was followed by the identification of two fatty acids, anandamide and 2-arachidonoylglycerol, which are endogenous ligands for this receptor; infusion of either ligand into the brain stimulates food intake. The CB1 receptor is a preganglionic receptor, meaning that its activation inhibits synaptic transmission. Antagonists to this receptor have been shown to reduce food intake and lead to weight loss.36 In addition to the drugs that act on the central nervous system, there are also drugs that act peripherally.33 Thus, for example, blockade of intestinal lipase by orlistat will produce weight loss. A second drug in this class, cetilistat, is in clinical trials. Pancreatic and intestinal peptides modulate food intake and also are candidates for treatment targets. Intestinal glucagon-like peptide-1 (GLP-1) acts on the pancreas, intestine, and brain to reduce food intake and slow gastric emptying. GLP-1 and exenatide, a drug that works by this GLP-1 mechanism, increase insulin secretion from the pancreas, reduce glucagon release from the pancreas, and reduce food intake by acting on GLP-1 receptors in the brain. Amylin is secreted from the pancreatic beta cell and can reduce food intake. Pramlintide is an example of a drug that works by mimicking the effects of amylin, which reduce food intake by acting on receptors in the brain.

FDA-Approved Medications

The FDA has approved several drugs for the treatment of obesity, shown in Table 6-4. Two of them, sibutramine and

Chapter 6  Obesity orlistat, are approved for long-term use (12 months); the others are approved for up to a few weeks, which is usually interpreted as 12 weeks. Sibutramine.  Sibutramine (Meridia in the United States and Reductil in Europe and other countries) has been marketed in the United States since March 1998. It is a selective reuptake inhibitor of serotonin and norepinephrine into neurons, but does not act on any known receptors. Sibutramine promotes satiety, but may also increase energy expenditure by blocking the reduction in metabolic rate that accompanies weight loss. In a randomized, placebocontrolled, six-month dose-ranging study of 1047 patients, there was a clear dose-response effect in dosages of 1 to 30  mg/day. In this study, 67% of subjects treated with sibutramine achieved a 5% weight loss from baseline, and 35% lost 10% or more.136 In a one-year trial of 456 patients who received sibutramine (10 or 15  mg/day) or placebo, 56% of those who stayed in the trial for 12 months lost at least 5% of their initial body weight, and 30% of the patients lost 10% of their initial body weight while taking the 10-mg dose.137 In a trial of patients who initially lost weight eating a very low-calorie diet before being randomized to sibutramine (10  mg/day) or placebo, sibutramine produced additional weight loss, whereas the placebo-treated patients regained weight.134,138 Sibutramine also is effective for weight maintenance. The Sibutramine Trial of Obesity Reduction and Maintenance (STORM) Trial began with a six-month open-label phase to induce weight loss using 10 mg/day of sibutramine. Patients who lost more than 8 kg then were randomized to sibutramine or placebo.139 During the 18-month double-blind phase of this trial, the placebo-treated patients steadily regained weight, maintaining only 20% of their weight loss at the end of the trial. In contrast, the subjects treated with sibutramine maintained 80% of their weight loss after two years.139 The blood pressure levels of the sibutramine-treated patients were still higher than in patients treated with placebo (see later). Clinical trials with sibutramine have shown that about 75% of patients treated with 15  mg/day of sibutramine achieved more than 5% weight loss, and 80% of those maintained that loss for 2 years if they stayed on the drug. In a meta-analysis of clinical trials with sibutramine,135 the drug produced a weighted mean weight loss of 6.35 ± 6.47 kg (−13.9 pounds) compared with 2.18 ± 5.23 kg (−4.8 pounds) for the placebo group, giving a net effect, or what is often called the placebo-subtracted weight loss, of −4.16 kg (95% CI: −4.73 to −3.59). About 5% of patients do not tolerate sibutramine because of adverse effects on blood pressure and pulse. Some patients (∼25%) are nonresponders. This drug, like other sympathomimetic drugs, produces a small increase in mean heart rate and mean blood pressure, as observed in clinical trials; however, the blood pressure response is variable. Sibutramine should be used with caution in patients with cardiovascular disease and in individuals taking selective serotonin reuptake inhibitors. It should not be used within two weeks of taking monoamine oxidase inhibitors or with other noradrenergic agents. A subset (∼5%) of patients appears to be sensitive to the blood pressure effects and cannot tolerate the drug. Other side effects, including dry mouth, insomnia, and asthenia, are similar to those of other noradrenergic drugs. Sibutramine is not associated with valvular heart disease, primary pulmonary hypertension, or substance abuse. Orlistat.  Orlistat (Xenical) is available by prescription in a dosage of 120 mg three times daily and Alli is offered over-

the-counter (OTC) at a lower dosage, 60  mg three times daily; both drugs should be taken before meals. Orlistat inhibits the enzymatic action of pancreatic lipase. In a twoyear trial with orlistat,140 patients received a hypocaloric diet that was 500 kcal/day less than their calculated requirements for the first year, and a diet that was calculated to maintain body weight in the second year. By the end of the first year, the placebo-treated patients lost 6.1% of their initial body weight and the drug-treated patients lost 10.2%. At the end of the second year, the patients who were switched from orlistat to placebo after one year gained weight, from 10% to 6% below baseline, a gain of 4%. Patients switched from placebo to orlistat lost weight, from 6% to 8.1% below baseline (a loss of 2.1%), an amount essentially identical to the 7.9% weight loss in the patients treated with orlistat for the full 2 years. In a three-year study,141 a very low-energy diet was used for eight weeks, and subjects who lost a minimum of 5% of their body weight were randomized to lifestyle or lifestyle plus orlistat. Weight loss continued to decline for three months and remained below randomization levels at 12 months in the orlistat group, but had risen above randomization level by six months in the lifestyle controls. At the end of three years, those on orlistat were still 2.4 kg lighter than the controls. Clinical trials show that ∼70% of patients will achieve more than 5% weight loss and at 2 years, 70% of them will have maintained that loss. There are clinical trials documenting orlistat use for up to four years.142 One advantage to orlistat’s use is its beneficial effect on LDL cholesterol. Because orlistat blocks fat absorption, the LDL reduction is about twice that seen with weight loss alone. A meta-analysis has shown that the orlistat-treated patients had a weighted mean weight loss of −5.70 ± 7.28 kg (−12.6 pounds) compared with −2.40 ± 6.99  kg (−5.3 pounds), giving a net, or placebo-subtracted weighted mean, weight loss of −2.87 (95% CI −3.21 to −2.53) (−6.4 pounds).135 In regard to safety, orlistat is poorly absorbed; all its side effects are those expected from inhibition of lipase in the intestine. It can produce fecal incontinence, anal leakage, bloating, and borborygmi, but these tend to occur early in treatment and deter very few patients. It also can lower levels of fat-soluble vitamins. A multivitamin taken when orlistat is not taken can prevent the reduction in fat-soluble vitamin levels. Sympathomimetic Amines.  Four sympathomimetic drugs have been approved by the FDA.134 Two (phentermine, diethylpropion) are schedule IV drugs and the other two (benzphetamine and phendimetrazine) are schedule III drugs (see Table 6-4). These drugs are only approved for a few weeks of use, which usually is interpreted as up to 12 weeks. Phentermine is not available in Europe. Obtaining written informed consent if phentermine is prescribed for longer than 12 weeks is good medical practice, because of the paucity of published reports on the long-term use of phentermine. Cannabinoid CB1 Antagonists.  Rimonabant is a cannabinoid receptor antagonist to the CB-1 receptor and initially was approved by European regulatory authorities. Marketing approval in Europe was withdrawn on October 23, 2008 by the European Medicines Agency (EMEA), and Sanofi-Aventis stopped development of this drug on October 5, 2008.

Non–FDA-Approved Medications

Fluoxetine.  Fluoxetine is a selective serotonin reuptake inhibitor (SSRI) that blocks serotonin transporters, thus pro-

111

112

Section II  Nutrition in Gastroenterology longing the action of serotonin. Fluoxetine is approved by the FDA for the treatment of depression. Fluoxetine at a dosage of 60  mg/day (three times the usual dose for treatment of depression) was effective in reducing food intake4 and body weight in overweight patients. A meta-analysis of six studies using fluoxetine has shown a wide range of results with a mean weight loss in one study of 14.5 kg and a weight gain of 0.40 kg in another.133 In the meta-analysis by Avenell and colleagues,143 the weight loss at 12 months was 0.33 kg (95% CI, −1.49 to 0.82 kg). Goldstein and colleagues reviewed the trials with fluoxetine that included one 36-week trial in type II diabetic subjects, one 52-week trial in subjects with uncomplicated overweight, and two 60-week trials in subjects with dyslipidemia, diabetes, or both.144 A total of 719 subjects were randomized to fluoxetine and 722 to placebo. Six months of treatment were completed by 522 subjects on fluoxetine and 504 subjects on placebo. Weight losses in the placebo and fluoxetine groups at six months were 2.2 and 4.8  kg and at one year were 1.8 and 2.4 kg, respectively. The regain of 50% of the lost weight during the second six months of treatment with fluoxetine makes this drug inappropriate for the long-term treatment of obesity. Fluoxetine, however, although not a good drug for long-term treatment of obesity, may be preferred for the treatment of depressed obese patients over some of the tricyclic antidepressants, which are associated with significant weight gain. Bupropion.  Bupropion is a norepinephrine and dopamine reuptake inhibitor approved for the treatment of depression and for help in smoking cessation. Two multicenter clinical trials, one in obese subjects with depressive symptoms and one in uncomplicated overweight patients, have tested this drug. In the study of overweight patients with depressive symptom ratings of 10 to 30 on a Beck Depression Inventory, 213 patients were randomized to 400 mg/day of bupropion and 209 subjects were assigned to placebo over a 24-week period. In the bupropion group, 121 subjects completed the trial and lost 6.0% ± 0.5% of initial body weight; the 108 subjects in the placebo group who completed the trial lost 2.8% ± 0.5% (P < 0.0001).145 A study in uncomplicated overweight subjects randomized 327 subjects to bupropion, 300  mg/day, bupropion 400  mg/day, or placebo in equal proportions.146 At 24 weeks, 69% of those randomized remained in the study and the percent losses of initial body weight were 5% ± 1%, 7.2% ± 1%, and 10.1% ± 1% for the placebo, bupropion 300-mg, and bupropion 400-mg groups, respectively (P < 0.0001). The placebo group was randomized to the 300- or 400-mg group at 24 weeks and the trial was extended to week 48. By the end of the trial, the dropout rate was 41%, and the weight losses in the bupropion 300and 400-mg groups were 6.2% ± 1.25% and 7.2% ± 1.5% of initial body weight, respectively.146 Thus, it appears that nondepressed subjects may respond to bupropion with more weight loss than those with depressive symptoms. Topiramate.  Topiramate is an antiepileptic drug that was discovered to be associated with weight loss in its clinical trials for epilepsy.147 Weight losses of 3.9% of initial weight were seen at three months and losses of 7.3% of initial weight were seen at one year. Bray and colleagues reported a six-month, placebo-controlled, dose-ranging study of topiramate in which 385 obese subjects were randomized to placebo or topiramate at 64, 96, 192, or 384 mg/day.148 These doses were gradually reached by a tapering increase and were reduced in a similar manner at the end of the trial. Weight loss from baseline to 24 weeks was 2.6%, 5%, 4.8%, 6.3%, and 6.3% in the placebo, 64-, 96-, 192-, and 384-mg

groups, respectively. The most frequent adverse events were paresthesias, somnolence, and difficulty with concentration, memory, and attention. This trial was followed by two other multicenter trials.149,150 The first trial randomized 1289 obese subjects to placebo or topiramate 89, 192, or 256 mg/day. This trial was terminated early because of the sponsor’s decision to pursue a time-release form of the drug. The 854 subjects who completed one year of the trial before it was terminated lost 1.7%, 7%, 9.1%, and 9.7% of their initial body weight in the placebo, 89-, 192-, and 256-mg groups, respectively. Subjects in the topiramate groups also had significant improvement in blood pressure and glucose tolerance.149 The second trial enrolled 701 subjects who were treated with a very low-calorie diet to induce an 8% loss of initial body weight.150 The 560 subjects who achieved an 8% weight loss were randomized to topiramate 96 or 192  mg/day, or placebo. This study also was terminated early. At the time of termination, 293 subjects had completed 44 weeks of the trial. The topiramate groups lost 15.4% and 16.5% of their baseline weight and the placebo group lost 8.9%.150 Topiramate also is effective in producing weight loss in diabetic patients.151 Although topiramate still is available as an antiepileptic drug, the development program to obtain an indication for overweight was terminated by the sponsor because of the associated adverse events. Zonisamide.  Zonisamide is an antiepileptic drug that has serotonergic and dopaminergic activity in addition to inhibiting sodium and calcium channels. Weight loss was noted in the clinical trials for the treatment of epilepsy, again suggesting the drug as a potential agent for weight loss. Gadde and colleagues tested this possibility by performing a 16-week randomized controlled trial in 60 obese subjects.152 Subjects were placed on a calorie-restricted diet and randomized to zonisamide or placebo. Zonisamide was started at 100 mg/day and increased to 400 mg/day. At 12 weeks, the dosage in those subjects who had not lost 5% of initial body weight was increased to 600  mg/day. The zonisamide group lost 6.6% of initial body weight at 16 weeks compared with 1% in the placebo group. Thirtyseven subjects completing the 16-week trial elected to continue the trial for 32 weeks, 20 in the zonisamide group and 17 in the placebo group. At the end of 32 weeks, the 19 subjects in the zonisamide group lost 9.6% of their initial body weight compared with 1.6% for the 17 subjects in the placebo group.152 Metformin.  Metformin is a biguanide approved for the treatment of diabetes mellitus that reduces hepatic glucose production, decreases glucose absorption from the gastrointestinal tract, and enhances insulin sensitivity. In clinical trials in which metformin was compared with sulfonylureas, it produced weight loss.134 In one French trial, called BIGPRO, metformin was compared with placebo in a oneyear multicenter study of 324 middle-aged subjects with upper body adiposity and the metabolic syndrome (insulin resistance syndrome). The subjects on metformin lost significantly more weight (1 to 2 kg) than those in the placebo group, and the study concluded that metformin may have a role in the primary prevention of type 2 diabetes.153 In a meta-analysis of three of these studies, Avenell and colleagues143 reported a weighted mean weight loss at 12 months of 1.09 kg (95% CI, −2.29 to 0.11 kg). The best trial of metformin for obesity, however, is the Diabetes Prevention Program (DPP) study of individuals with impaired glucose tolerance. This study included a double-blind comparison of metformin 850 mg twice daily

Chapter 6  Obesity with placebo. During the 2.8 years of this trial, the 1073 patients treated with metformin lost 2.5% of their body weight (P < 0.001) compared with the 1082 patients treated with placebo, and the conversion from impaired glucose tolerance to diabetes was reduced by 31% compared with placebo. In the DPP trial, metformin was more effective in reducing the development of diabetes in persons in the subgroup who were most overweight and in the younger members of the cohort.96 Although metformin does not produce enough weight loss to qualify as a weight loss drug (FDA criteria require 5% weight loss), it would appear to be a very useful choice for overweight individuals who have diabetes or are at high risk for diabetes. One area in which metformin has found use is in treating overweight women with the polycystic ovary syndrome, in whom the modest weight loss may contribute to increased fertility and reduced insulin resistance.154 Pramlintide.  Amylin is a peptide found in the beta cells of the pancreas that is cosecreted along with insulin to circulate in the blood. Amylin and insulin are deficient in type 1 diabetics, in whom beta cells are destroyed immunologically. Pramlintide, a synthetic amylin analog, has a prolonged biological half-life.155 Pramlintide is approved by the FDA for the treatment of diabetes. Unlike insulin and many other diabetic medications, pramlintide use is associated with weight loss. In a study in which 651 subjects with type 1 diabetes were randomized to placebo or subcutaneous pramlintide 60 µg three or four times daily, along with an insulin injection, the hemoglobin A1c level decreased 0.29% to 0.34% and weight decreased −1.2 kg relative to placebo.156 Maggs and colleagues analyzed the data from two one-year studies in insulin-treated type 2 diabetic subjects randomized to pramlintide 120 µg twice daily or 150 µg three times daily.157 Weight decreased by 2.6  kg and hemoglobin A1c decreased 0.5%. When weight loss was then analyzed by ethnic group, African Americans lost 4  kg, whites lost 2.4 kg, and Hispanics lost 2.3 kg; the improvement in diabetes correlated with the weight loss, suggesting that pramlintide is effective in ethnic groups with the greatest burden from overweight. The most common adverse event was nausea, which usually was mild and confined to the first four weeks of therapy. Exenatide.  GLP-1 is derived from the processing of the preproglucagon peptide, which is secreted by L cells in the terminal ileum in response to a meal. Increased GLP-1 inhibits glucagon secretion, stimulates insulin secretion, stimulates gluconeogenesis, and delays gastric emptying.155 It has been postulated to be responsible for the superior weight loss and improvement in diabetes seen after gastric bypass surgery for overweight.158 GLP-1 is rapidly degraded by dipeptidyl peptidase-4 (DPP-4), an enzyme that is elevated in obese individuals. Bypass operations for overweight increase GLP-1, but do not change the levels of DPP-4.155 Exenatide (exendin-4) is a 39–amino acid peptide produced in the salivary gland of the Gila monster lizard. It has 53% homology with GLP-1, but has a much longer half-life. Exenatide is approved by the FDA for treatment of type 2 diabetics who are inadequately controlled while being treated with metformin or sulfonylureas. In humans, exenatide reduces fasting and postprandial glucose levels, slows gastric emptying, and decreases food intake by 19%.155 The side effects of exenatide in humans are headache, nausea, and vomiting that are lessened by gradual dose escalation. Several clinical trials of 30 weeks’ duration have been reported using exenatide at a daily

dosage of 5 µg twice daily subcutaneously.159-161 In one trial with 377 type 2 diabetic subjects who were failing maximal sulfonylurea therapy, exenatide produced a fall of 0.74% more in hemoglobin A1c than placebo. Fasting glucose levels also decreased and there was a progressive weight loss of 1.6 kg.161 The interesting feature of this weight loss is that it occurred without change in lifestyle, diet, or exercise. In a 26-week randomized controlled trial, exenatide produced a 2.3-kg weight loss compared with a gain of 1.8 kg in the group receiving insulin glargine.162 The FDA has issued a notice about potential risks of pancreatitis.

Drug Combinations

The first important clinical trial combining drugs that acted by separate mechanisms used phentermine and fenfluramine.163 This trial showed that by using combination therapy, a highly significant weight loss was achieved of almost 15% below baseline and with fewer side effects than those seen with the individual agents. This combination became very popular, but because of reports of aortic valvular regurgitation associated with its use, fenfluramine was withdrawn from the market worldwide on September 15, 1997.164 Several other combinations of existing drugs are now under development, including phentermine with topiramate (Qnexa), phentermine with zonisamide, and naltrexone with bupropion. Initial data have been published on all three combinations, but longer term studies are needed to evaluate the potential drug-drug interactions and side effects produced.

SURGERY

Surgical intervention for obesity has become ever more popular (see Chapter 7).165 The Swedish Obese Subjects Study has offered surgical intervention for obese Swedish patients aimed at reducing their obesity through a gastrointestinal operation. The control group included obese Swedish patients who did not get surgical treatment but were treated with the best alternatives in the Swedish health care system. The effect of weight change on dyslipidemia, blood pressure, and serum insulin levels in the surgically treated group two and 10 years after surgery was compared with these parameters in the control group.166 There was a graded effect of weight change on HDL cholesterol, triglycerides, systolic and diastolic blood pressure, insulin, and glucose. This surgery has now been shown to reduce mortality.94,95 A comparison of surgically and nonsurgically treated patients has shown that weight loss improves long-term health outcomes, but at a cost of significant short-term health problems.167

KEY REFERENCES

Avenell A, Brown TJ, McGee MA, et al. What interventions should we add to weight reducing diets in adults with obesity? A systematic review of randomized controlled trials of adding drug therapy, exercise, behaviour therapy or combinations of these interventions. J Hum Nutr Diet 2004; 7:293-316. (Ref 143.) Bray GA. The metabolic syndrome and obesity. Totowa, NJ: Humana Press; 2007. (Ref 4.) Bray G, Greenway F. Pharmacological treatment of the overweight patient. Pharmacol Rev 2007: 59:151-84. (Ref 134.) Christakis NA, Fowler JH. The spread of obesity in a large social network over 32 years. N Engl J Med 2007; 357:370-9. (Ref 15.) Dansinger ML, Tatsioni A, Wong JB, et al. Meta-analysis: The effect of dietary counseling for weight loss. Ann Intern Med 2007; 147:41-50. (Ref 112.) Farooqi IS, O’Rahilly S. Genetic factors in human obesity. Obes Rev 2007; 8(Suppl 1):37-40. (Ref 27.) Freedman DM, Ron E, Ballard-Barbash R, et al. Body mass index and all-cause mortality in a nationwide US cohort. Int J Obes (Lond) 2006; 30:822-9. (Ref 45.)

113

114

Section II  Nutrition in Gastroenterology Klein S, Burke LE, Bray GA, et al; American Heart Association Council on Nutrition, Physical Activity, and Metabolism. Clinical implications of obesity with specific focus on cardiovascular disease: A statement for professionals from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism: Endorsed by the American College of Cardiology Foundation. Circulation 2004; 110:2952-67. (Ref 46.) Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002; 346:393-403. (Ref 96.) National Institutes of Health. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults— The Evidence Report. National Institutes of Health. Obes Res 1998; 6(Suppl 2):51S-209S. (Ref 2.)

Poirier P, Giles TD, Bray GA, et al; American Heart Association; Obesity Committee of the Council on Nutrition, Physical Activity, and Metabolism. Obesity and cardiovascular disease: pathophysiology, evaluation, and effect of weight loss: An update of the 1997 American Heart Association Scientific Statement on Obesity and Heart Disease from the Obesity Committee of the Council on Nutrition, Physical Activity, and Metabolism. Circulation 2006; 113:898-918. (Ref 50.) Rucker D, Padwal R, Li SK, et al. Long term pharmacotherapy for obesity and overweight: updated meta-analysis. BMJ 2007; 335:1194-9. (Ref 135.) Sjöström L, Narbro K, Sjöström CD, et al. Effects of bariatric surgery on mortality in Swedish obese subjects. N Engl J Med 2007; 357:741-52. (Ref 94.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

7

Bariatric Surgery Gavitt Woodard and John Morton

CHAPTER OUTLINE Epidemiology of Morbid Obesity  115 Bariatric Surgery as Treatment for Morbid Obesity  115 Preoperative Evaluation  115 Efficacy  116 Effects on Mortality and Morbidity  116 Comorbidity Resolution  117

EPIDEMIOLOGY OF MORBID OBESITY Morbid obesity is the leading public health crisis of the industrialized world (see Chapter 6).1,2 The prevalence of obesity in the United States continues to rise at an alarming rate, with two thirds of adults currently considered overweight, half of whom are obese.3 Being overweight is defined by the body mass index (BMI): normal BMI = 25 kg/m2; BMI for obesity > 30 kg/m2; BMI for morbid obesity > 40 kg/m2; and BMI for supermorbid obesity > 50 kg/m2. Rising rates of obesity are seen across the United States in men and women and in all major racial, ethnic, and socioeconomic groups.4 Morbid obesity reduces life expectancy by five to 20 years and, for the first time in history, it is predicted that the current generation may have a shorter life expectancy than the last.5,6

BARIATRIC SURGERY AS TREATMENT FOR MORBID OBESITY

Bariatric surgery remains the only effective and enduring treatment for morbid obesity. Since 1997, the number of bariatric surgical procedures in the United States has grown sevenfold as evidence has proven their safety and efficacy.7 Weight loss operations can be classified as malabsorptive or restrictive (Fig. 7-1). Roux-en-Y gastric bypass (RNYGB), which accounts for 88% of bariatric procedures in the United States, is restrictive and malabsorptive. Biliopancreatic diversion–duodenal switch (BPD-DS), the other malabsorptive procedure, is not as commonly performed in the United States because of its higher risk profile. Purely restrictive procedures include the laparoscopic adjustable gastric banding (LAGB), vertical banded gastroplasty (VBG), gastrectomy, and sleeve gastrectomy procedures, all of which reduce the size of the stomach so it is unable to accommodate more than a few ounces of food. VBG is no longer performed commonly because of its potential for staple line dehiscence and subsequent weight gain. Advantages, disadvantages and complications of the major weight loss operations are shown in Tables 7-1 and 7-2.

Surgical Complications  118 Volume Effect and Center of Excellence Movement  119

PREOPERATIVE EVALUATION To qualify for bariatric surgery, patients must meet the 1991 NIH consensus criteria, which include having a BMI > 40 kg/m2 or a BMI > 35 kg/m2 with obesity-related comorbidities, and at least six months of documented, medically supervised weight loss attempts.8 Some bariatric surgeons require patients to lose additional weight through diet and exercise between the time of initial bariatric surgery con­ sultation and the date of operation. This additional required preoperative weight loss is not correlated with comorbidity resolution or complication rates, but is associated with shorter operative times and greater weight loss at one year after the surgery; therefore, it should be encouraged in all patients.9,10 Prior to surgery, patients must complete an extensive screening process, including consultation with a surgeon, psychologic evaluation, nutrition consultation, chest roentgenography, electrocardiography, pulmonary function testing, a sleep study, and an esophagogastroduodenoscopy (EGD). The EGD was recommended by the European Association for Endoscopic Surgery to detect and treat any upper gastrointestinal lesions that may cause postoperative complications or influence the decision of which type of bariatric surgery should be performed.11 In a study of 272 gastric bypass patients who underwent preoperative EGD, 12% had clinically significant preoperative findings that included erosive esophagitis (3.7%), Barrett’s esophagus (3.7%), gastric ulcer (2.9%), erosive gastritis (1.8%), duodenal ulcer (0.7%), and gastric carcinoid (0.3%); 1.1% had more than one lesion. Given that 12% of patients who eventually underwent RNYGB had clinically significant preoperative findings, and only 67% of them had upper gastrointestinal symptoms, it is important to perform EGD preoperatively, because the excluded distal stomach cannot be evaluated easily after a RNYGB procedure.12 Patients undergoing LAGB surgery also should undergo a preoperative EGD, especially to evaluate gastroesophageal reflux disease (GERD). Gastric banding leads to satiety and weight loss by creating a small restrictive stomach with a slow

115

116

Section II  Nutrition in Gastroenterology gastric emptying time. Patients who overfill their small stomach pouch post-LAGB can force food and stomach acid back up into their lower esophagus, thereby worsening any preexisting GERD.13,14 In addition, overzealous banding adjustment may lead to pseudoachalasia with an increased pressure zone below the lower esophageal sphincter, furthering any incompetence.

EFFICACY

Roux-en-Y Gastric Bypass (RNYGB)

Vertical Banded Gastroplasty (VBG)

The steep rise in the use of bariatric surgery can be attributed to its proven efficacy as a treatment for morbid obesity. Two meta-analyses have provided strong validation that bariatric surgery leads to successful weight loss and mortality reduction.15,16 A meta-analysis by Buchwald and colleagues that included 22,094 patients found the mean percentage of excess weight loss (EWL) for all patients to be 61.2%.15 EWL is highest for VBG (68.2%), lower for RNYGB (61.6%), and lowest for LAGB (47.5%). A meta-analysis by Maggard and associates found similar weight loss trends at three or more years postoperatively, with the greatest weight loss achieved after the malabsorptive procedures of BPD-DS (53 kg) and RNYGB (42 kg), and less weight loss after the restrictive LAGB (35 kg) and gastroplasty (32 kg).16

EFFECTS ON MORTALITY AND MORBIDITY

Laparoscopic Adjustable Gastric Band (LAGB)

Biliopancreatic Diversion (BPD) with Duodenal Switch

Figure 7-1.  Types of weight loss operations. (From the American Society for Bariatric Surgery. The story of surgery for obesity, 2005. Available at www.asbs.org.)

Such substantial weight loss is associated with a clear reduction in long-term mortality. A retrospective cohort study of 9,949 RNYGB patients matched to 9,628 severely obese controls found that having RNYGB surgery reduces the adjusted long-term mortality from any cause of death by 40%.17 Among RNYGB patients, mortality was decreased 56% from coronary artery disease, 92% from diabetes, and 60% from cancer. In another study, a 14% decrease in cancer incidence was shown in patients who underwent RNYGB, the biggest reductions in which were seen among types of cancers that are considered obesity-related: eso­ phageal adenocarcinoma (2% reduction), colorectal cancer (30% reduction), breast cancer (9%), uterine cancer (78%), non-Hodgkin’s lymphoma (46%), and multiple myeloma (54%).18 The lower cancer risk of patients after RNYGB

Table 7-1  Advantages and Disadvantages of Bariatric Surgery Procedures for Weight Loss

Procedure RNYGB

Weight Loss Mechanism

Mean Weight Loss (kg) Advantages

Disadvantages

Reversible?

Good weight loss

Vitamin deficiencies; internal hernias Less weight loss; may require adjustments; foreign body

LAGB

Restriction and malabsorption Restriction

VGB

Restriction

Relatively low risk of vitamin deficiencies; low rate of complications Ease of construction

BPD

Malabsorption

Greatest weight loss

No longer performed Staple dehiscence High risk of malabsorption and vitamin deficiencies

After 12 mo

After 36 mo

No

43.5

41.5

Band may be removed

30.2

34.8

Yes

32.1

32.0

No

51.9

53.1

BPD, biliopancreatic diversion; LAGB, laparoscopic adjustable gastric banding; NR, not reported; RNYGB, Roux-en-Y gastric bypass; VGB, vertical banded gastroplasty. Data from Maggard MA, Shugarman LR, Suttorp M, et al. Meta-analysis: Surgical treatment of obesity. Ann Intern Med 2005; 142:547-59.

Chapter 7  Bariatric Surgery Table 7-2  Complication Rates (%) for Bariatric Surgery Procedures Procedure RNYGB LAGB VGB BPD

Mortality Rate 1.0 0.4 0.2 NR, but at least ≥RNYGB

GI Symptoms

Reflux

Vomiting

16.9 7.0 17.5 37.7

10.9 4.7 2.2 NR

15.7 2.5 18.4 5.9

Nutritional Abnormalities 16.9 NR 2.5 NR, but high potential

Reoperation

Leak Rate

1.6 7.7 11.3 4.2

2.2 NR 1.0 1.8

Bleeding 2.0 0.3 0.7 0.2

BPD, biliopancreatic diversion; GI, gastrointestinal; LAGB, laparoscopic adjustable gastric banding; NR, not reported; RNYGB, Roux-en-Y gastric bypass; VGB, vertical banded gastroplasty. Data from Maggard MA, Shugarman LR, Suttorp M, et al. Meta-analysis: Surgical treatment of obesity. Ann Intern Med 2005; 142:547-59.

presumably was caused by weight loss, which has been shown in many studies to reduce cancer incidence. Furthermore, once obese patients lose weight, they may have better access to needed health surveillance, such as Pap smears and colonoscopy. Finally, given that increased BMI leads to worse surgical oncologic outcomes, it may be surmised that with weight loss, a better surgical outcome may be anticipated. Overall, bariatric surgery dramatically improves survival and decreases mortality from all disease-related causes of death. Only rate of deaths not caused by disease, including deaths resulting from accidents and suicide, increased after bariatric surgery and were 58% higher in the RNYGB patients17; it has been speculated that alcohol abuse may explain why accidents and suicides were higher in the surgical group. One study demonstrated altered alcohol metabolism after gastric bypass surgery, with the gastric bypass patients having a greater peak alcohol level and a longer time for the alcohol level to reach a zero blood level than in controls.30 Another study of bariatric surgery candidates found that 9% reported having attempted suicide and 19% reported having abused alcohol preoperatively.19 There is concern that this vulnerable patient population has additional difficulty with the psychological adjustments to weight loss, further supporting the need for psychological counseling before and after surgery.20,21 In addition to benefiting from a decreased mortality, bariatric patients benefit from decreased morbidity. Morbidly obese patients suffer from more intense gastrointestinal symptoms (e.g., abdominal pain, heartburn) and sleep disturbances than normal-weight patients. By six months after RNYGB, however, the frequency and severity of gastrointestinal symptoms of many morbidly obese patients have decreased to levels seen in normal weight patients. Dysphagia is common in morbidly obese patients, all of whom experience increased intra-abdominal pressure. Dysphagia is the only gastrointestinal symptom that worsens after RNYGB, probably from further increase in esophageal pressure because of overeating and overfilling of the restrictive small gastric pouch; this observation again underscores the need for preoperative and postoperative education regarding diet.27 Quality of life, as measured by the validated SF36 survey, improves greatly after RNYGB surgery. Preoperatively, morbidly obese patients score significantly lower than U.S. population norms in the categories of general health, vitality, physical functioning, bodily pain, emotional, and social functioning. As soon as three months following RYNGB, these same patients scored no differently than U.S. norms in these categories.28

COMORBIDITY RESOLUTION

Weight loss surgery is a singular medical intervention that can reverse or improve the numerous medical conditions associated with obesity. RNYGB results in a substantial reduction in cardiac risk factors with the following resolution rates: diabetes (82%), hypertension (70%), and hyperlipidemia (63%).29 Gastric bypass has assembled the most evidence of comorbidity resolution; however, all weight loss operations result in some degree of improvement. The meta-analysis by Buchwald and coworkers15 found that bariatric surgery reverses, ameliorates, or eliminates major cardiovascular risk factors: Hypertension was resolved by banding (38%), gastroplasty (73%), gastric bypass (75%), and BPD-DS (81%). Diabetes was resolved by banding (48%), gastroplasty (68%), gastric bypass (84%), and BPD-DS (98%). Hyperlipidemia was improved by banding (71%), gastroplasty (81%), gastric bypass (94%), and BPD-DS (99%). One study found resolution of all conventional abnormal risk factors, including serum levels of total cholesterol, lowdensity lipoprotein, high-density lipoprotein (HDL), triglyceride, high-sensitivity C-reactive protein, homocysteine, and lipoprotein A at one year after RNYGB.30 The Swedish Obese Subjects (SOS) Study has provided further demonstration of the ameliorative effect of bariatric surgery. At 10 years of follow-up, surgically treated obese patients had 25% reduction in hypertension, 43% improvement in HDL, and 75% reduction in diabetes compared with the medically treated group.31 Beyond the significant improvement in cardiac risk factors, weight loss surgery also provides substantial relief of the many medical problems that obesity causes. One study found that a leading digestive health complaint, GERD, is cured or improved at a 96% rate.29 Other studies also have documented a highly significant reduction in GERD symptoms after bariatric surgery.27,32 One study of patients with severe GERD prior to RNYGB found signi­ ficant declines in use of proton pump inhibitors (44% to 9%) and H2 blockers (60% to 10%) at postoperative times ranging from six to 36 months.33 GERD resolution rates following RNYGB are so robust that RNYGB is a suggested treatment for recalcitrant GERD in morbidly obese patients, especially given the lack of efficacy of antireflux surgery in the obese.34,35 Prior to surgery, morbidly obese patients report significantly more symptoms of abdominal distress, including pain, gnawing sensations, nausea, vomiting, and abdominal distention than normal controls. After RNYGB, these symptoms are reduced to levels comparable with those of normal controls.27,32 Irritable bowel syndrome (IBS) is a constella-

117

118

Section II  Nutrition in Gastroenterology tion of symptoms (see Chapter 118), including abdominal pain or discomfort, altered bowel habits (diarrhea, consti­ pation), increased flatus, and bloating or distention. Preoperative morbidly obese patients rate their severity of IBS symptoms significantly higher than controls. Following RNYGB, these same patients rated the severity of their symptoms significantly lower than they did preoperatively and at levels equivalent to those of control IBS patients. Nonalcoholic fatty liver disease (NAFLD) comprises a histologic spectrum of fatty liver ranging from simple steatosis to portal fibrosis, nonalcoholic steatohepatitis (NASH), and cirrhosis.36 The most advanced forms of NAFLD are strongly associated with the metabolic syndrome—that is, obesity and hypertension, hypertriglyceridemia, and diabetes mellitus—and NAFLD is the most prevalent liver disease in the United States.37,38 In a postmortem series of obese nondrinkers, hepatic steatosis was present in 76% and NASH was present in 18.5%.39 Mounting evidence suggests that current bariatric surgical procedures actually may be beneficial for NAFLD.40 One study documented that histologic improvements as measured by steatosis (89.7% improvement), hepatocellular ballooning (58.9% improvement), and centrilobular-perisinusoidal fibrosis (50% improvement) occur within a mean period of 18 months after RNYGB. The diagnosis of NASH was made in 58.9% of patients preoperatively and none of them were found to have NASH in postoperative liver biopsies.41 Weight loss surgery has been shown to improve sleep apnea with the following rates: banding (95%), gastroplasty (77%), gastric bypass (87%), and BPD-DS (95%).15 Another study has demonstrated that after RNYGB, patients have significantly fewer sleep disturbances with regard to falling asleep, insomnia, and feeling rested on awakening.27 Morbidly obese patients undergoing RNYGB have demonstrated an 88% resolution or improvement of joint disease.29 Furthermore, weight loss surgery has been demonstrated to eliminate or improve obesity-associated venous stasis disease, gout, asthma, pseudotumor cerebri, urinary incontinence, and infertility.29

SURGICAL COMPLICATIONS The risk of operative mortality and complications may temper some enthusiasm for bariatric surgery. Based on a meta-analysis by Maggard and colleagues,16 mortality rates were shown to depend on the procedure performed. The average mortality rates for the different procedures are banding (0.4%, 0.01% to 2.1%), gastroplasty (0.2%, 0% to 16.8%), gastric bypass (1%, 0.2% to 2.5%), and BPD-DS (0.9%, 0.01% to 1.3%). Complications of bariatric procedures include anastomotic leak or stenosis, pulmonary embolus, gastrointestinal bleeding, nutritional deficiencies, wound complications, bowel obstructions, ulcers, hernias, respiratory, cardiac, and implant device-related complications. Among the different surgical procedures, the rate of complications is proportional to the amount of weight loss produced by each operation: banding (7%), gastroplasty (18%), gastric bypass (17%), and BPD-DS (38%).16,42 Marginal ulcers are estimated to occur in 1% to 16% of gastric bypass patients.43,44 Perforated marginal ulcers occur in 1% of RNYGB patients. Ulcer perforation is linked to smoking and use of nonsteroidal anti-inflammatory drugs (NSAIDs) or glucocorticoids.45 The use of nonabsorbable sutures, as opposed to absorbable sutures, for the inner layer of the gastrojejunal anastomosis is associated with increased

ulcer incidence.46 The presence of H. pylori also increases risk for marginal ulcers.47 Postoperatively, it is common practice for bariatric surgeons to begin a six-month ulcer prophylaxis program with proton pump inhibitors. If a marginal ulcer is recalcitrant to medical therapy, the possibility of a gastric-gastric fistula must be entertained. If a gastricgastric fistula is present with a marginal ulcer, then surgical correction is mandated. Gastrointestinal bleeding occurs postoperatively in 2.0% of RNYGB, 0.7% of VBG, 0.3% of LAGB, and 0.2% of BPD-DS procedures. Another potential complication is an anastomotic gastric pouch or duodenal leak, which occurs after 2.2% of RNYGB, 1.0% of VBG, and 1.8% of BPD-DS procedures.16 Nutritional and vitamin deficiencies and electrolyte abnormalities occur in 16.9% of RNYGB patients and 2.5% of patients having VBG.16 Patients who do not take daily vitamins postoperatively or patients who experience frequent vomiting are at increased risk of developing such deficiencies, most common of which are protein, iron, vitamin B12, folate, calcium, and the fat-soluble vitamins A, D, E, and K.48 The parietal cells of the stomach produce intrinsic factor, which is necessary for vitamin B12 absorption in the terminal ileum. Patients who undergo RNYGB may develop vitamin B12 deficiency because RNYGB separates the parietal cells in the fundus and body of the stomach from the smaller gastric pouch, which receives ingested food. There is, therefore, no contact between ingested food and intrinsic factor until the intersection of the Roux limb in the jejunum.49,50 In addition, following RNYGB, the parietal cells of the stomach often cease to produce intrinsic factor, presumably because the fundus no longer has any contact with food.51 It has been shown that restrictive bariatric surgery does not cause vitamin B12 deficiency because the parietal cells remain in contact with the ingested food.52 Fat-soluble vitamin deficiencies are most often seen following BPD-DS operations because food has very little exposure to the biliary and pancreatic secretions necessary for fat digestion, and there is little exposure of food to the ileum, where fat is normally absorbed. Calcium and folate deficiency can occur because they are absorbed in the duodenum and proximal jejunum. These segments of the digestive tract are commonly bypassed in gastric bypass and duodenal switch surgery. The fat-soluble vitamin D is necessary for calcium absorption, and so vitamin D deficiency will further contribute to any calcium deficiency.48 Thiamine deficiency may lead to Wernicke’s encephalopathy, a syndrome of confusion, ataxia, ophthalmoplegia, and impaired short-term memory. If thiamine deficiency is suspected, the patient should be given intravenous or intramuscular thiamine immediately in order to increase the chances of symptom resolution.53 Meta-analysis has revealed that gastroesophageal reflux occurs postoperatively in 10.9% of RNYGB, 2.2% of VBG, and 4.7% of LAGB patients.16 Approximately half of LAGB patients will experience some degree of heartburn and acid regurgitation.54 Pseudoachalasia and esophageal dysmotility are late complications of LAGB and usually reverse on removal of the gastric band.55 Among RNYGB patients, postoperative dysphagia is significantly worse than normal weight controls, but not significantly worse than the patient’s matched preoperative symptoms.27 A high incidence of gallstone formation has been well documented when morbidly obese patients undergo rapid surgically induced weight loss.56 It was shown in a doubleblind, randomized, placebo-controlled trial that a daily dosage of 600 mg ursodiol for the first six months after surgery reduces the incidence of gallstones to 2%.57 It is

Chapter 7  Bariatric Surgery therefore recommended that all bariatric patients take ursodiol for six months postoperatively to reduce the risk of this largely preventable complication. Beyond the type of procedure, there are identified risk factors for complications after bariatric surgery, including older age, male gender, greater BMI, comorbidities, and Medicare insurance status.58-62 The increased risk for Medicare patients is beyond age, given that eligibility for Medicare is disability, which may affect outcomes. Although patients with the most risk factors carry the highest risk for surgery, they also may derive the most benefit from bariatric surgery, given the disease burden they carry.63 Of note, complications may not affect long-term weight loss, which is the outcome that best predicts long-term mortality risk.9

VOLUME EFFECT AND CENTER OF EXCELLENCE MOVEMENT Many of the reported studies regarding morbidity and mortality had been completed prior to the many improvements in current surgical technique.28 In addition, surgeon and hospital experience can mitigate the risks associated with weight loss surgery. The best demonstrated and most protective effect against complications is an experienced surgeon and hospital.63-66 Clearly, there is benefit in having this complex and demanding surgery performed by experienced and committed surgeons operating in a dedicated health care facility. None of the previously mentioned perioperative risk factors can be modified, with the exception of the volume status of the surgeon and hospital. For gastric bypass surgery, it has been demonstrated that a high-volume surgeon and high-volume hospital lead to decreased morbidity and mortality.63-66 In the United States, this volume outcome effect has been recognized by the Centers for Medicare and Medicaid Services, which now require Medicare patients to undergo surgery only at a Bariatric Surgery Center of Excellence.67 Numerous criteria enable a center to become a Bariatric Surgery Center of Excellence, but the primary criteria are a surgeon volume of more than 50 cases and hospital volume exceeding 125 cases annually. Although a referral

to a Bariatric Surgery Center of Excellence may lead to decreased morbidity and mortality, this referral pattern must be balanced with appropriate and sufficient access to care for a vulnerable population without other therapeutic options.

KEY REFERENCES

Adams TD, Gress RE, Smith SC, et al. Long-term mortality after gastric bypass surgery. N Engl J Med 2007; 357:753-61. (Ref 17.) Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery: a syste­ matic review and meta-analysis. JAMA 2004; 292:1724-37. (Ref 15.) Fontaine KR, Redden DT, Wang C, et al. Years of life lost due to obesity. JAMA 2003; 289:187-93. (Ref 5.) Hagedorn JC, Encarnacion B, Brat GA, Morton JM. Does gastric bypass alter alcohol metabolism? Surg Obes Relat Dis 2007; 3:543-8. (Ref 30.) Klein S, Mittendorfer B, Eagon JC, et al. Gastric bypass surgery improves metabolic and hepatic abnormalities associated with nonalcoholic fatty liver disease. Gastroenterology 2006; 130:1564-72. (Ref 40.) Maggard MA, Shugarman LR, Suttorp M, et al. Meta-analysis: surgical treatment of obesity. Ann Intern Med 2005; 142:547-59. (Ref 16.) Nguyen NT, Goldman C, Rosenquist CJ, et al. Laparoscopic versus open gastric bypass: A randomized study of outcomes, quality of life, and costs. Ann Surg 2001; 234:279-89; discussion 289-91. (Ref 28.) Nguyen NT, Paya M, Stevens CM, et al. The relationship between hospital volume and outcome in bariatric surgery at academic medical centers. Ann Surg 2004; 240:586-93. (Ref 64.) Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of overweight and obesity in the United States, 1999-2004. JAMA 2006; 295:1549-55. (Ref 1.) Olshansky SJ, Passaro DJ, Hershow RC, et al. A potential decline in life expectancy in the United States in the 21st century. N Engl J Med 2005; 352:1138-45. (Ref 6.) Santry HP, Gillen DL, Lauderdale DS. Trends in bariatric surgical procedures. JAMA 2005; 294:1909-17. (Ref 7.) Schauer PR, Ikramuddin S, Gourash W, et al. Outcomes after laparoscopic Roux-en-Y gastric bypass for morbid obesity. Ann Surg 2000; 232:515-29. (Ref 29.) Sjöström L, Lindroos AK, Peltonen M, et al. Lifestyle, diabetes, and cardiovascular risk factors 10 years after bariatric surgery. N Engl J Med 2004; 351:2683-93. (Ref 31.) Sugerman HJ, Brewer WH, Shiffman ML, et al. A multicenter, placebocontrolled, randomized, double-blind, prospective trial of prophy­ lactic ursodiol for the prevention of gallstone formation following gastric-bypass-induced rapid weight loss. Am J Surg 1995; 169:91-6. (Ref 57.) Williams DB, Hagedorn JC, Lawson EH, et al. Gastric bypass reduces biochemical cardiac risk factors. Surg Obes Relat Dis 2007; 3:8-13. (Ref 58.) Full references for this chapter can be found on www.expertconsult.com.

119

CHAPTE R

8

Eating Disorders Anne E. Becker and Christina Wood Baker

CHAPTER OUTLINE Epidemiology  121 Causative Factors  121 Satiety  122 Appetite  123 Energy Storage  123 Onset and Course  123 Diagnosis and Evaluation  123 Anorexia Nervosa  124 Bulimia Nervosa  124 Eating Disorder Not Otherwise Specified  125 Binge Eating Disorder  125 Night Eating Syndrome and Nocturnal Sleep-Related Eating Disorder  125

Eating disorders are mental illnesses characterized by disturbances in weight control, body image, and/or dietary patterns. Diagnostic categories include (1) anorexia nervosa (AN); (2) bulimia nervosa (BN); and (3) eating disorder not otherwise specified (EDNOS; Fig. 8-1; Table 8-1). Several variants of EDNOS, such as binge-eating disorder and night eating syndrome, are well-described in the literature. The focus of this chapter is eating disorders in adults; other disturbances in eating that typically have onset in infancy and early childhood, such as pica, rumination syndrome, and feeding disorder of early childhood, are not discussed here. Although eating disorders are classified as mental illnesses, their associated behaviors commonly result in and present with medical sequelae, many of which are gastrointestinal. Because associated chronic undernutrition, overweight, and/or purging behaviors often result in serious medical complications that can be chronic, individuals with eating disorders benefit from the ongoing care of a multidisciplinary treatment team. Indeed, eating disorders (AN and BN) are among the mental disorders with the highest mortality risk.1

EPIDEMIOLOGY Eating disorders have been described across diverse global settings, although epidemiologic data are best established for populations in North America and Europe. The incidence rate for AN is approximately eight cases/100,000 population/year, with a point prevalence of AN estimated at 0.3% in young women of the United States and Western European general populations. BN is more common than AN, with an incidence of 12 cases/100,000 population/year in the United States and Western Europe2 and a 12-month prevalence of 0.5% among adult women in the United States. Lifetime prevalence estimates for U.S. women based on the National Comorbidity Study Replication (NCS-R) are

Purging Disorder  125 Differential Diagnosis  125 Nutritional and Medical Evaluation  126 Nutritional Evaluation  126 Gastrointestinal Abnormalities Associated with Eating Disorders  128 Management  131 Psychiatric Treatment  131 Weight Management  134 Medical Management of Gastrointestinal Symptoms of Patients with Eating Disorders  135

0.9% for AN and 1.5% for BN; U.S. men have a lifetime prevalence of 0.3% for AN and 0.5% for BN.3 The most common presentation of an eating disorder in outpatient settings is EDNOS, although fewer prevalence data are available for this broad and heterogeneous category. One large study from Portugal reported a prevalence of 2.37% for EDNOS in female students in grades 9 to 12.4 Relatively high prevalence rates also are reported for specific symptoms associated with disordered eating. For example, in 2007, 6.4% of school-going female adolescents in the United States reported vomiting or laxative use, and 16.3% reported fasting within the previous month to lose weight.5 Within the diagnostic category of EDNOS, there is great interest in a clinical variant termed binge-eating disorder (BED). The lifetime prevalence of BED in the United States is 3.5% for adult women and 2.0% for adult men. Additional variants of disordered eating (that are not classified as distinct disorders in the Diagnostic and Statistical Manual of Mental Disorders, DSM) include the night eating syndrome (NES) and nocturnal sleep-related eating disorder (NSRED). The prevalence of NES in young adult women has been reported as 1.6%6 and of NSRED as 0.5%.7 Eating disorders occur across ethnically and socioeconomically diverse populations, but each of the eating disorders is more common in women than in men. Males account for less than 10% of individuals with AN, 10% of those with BN, 34% of those with NES, and 40% of those with BED.8,9

CAUSATIVE FACTORS Although incompletely understood, the cause of eating disorders is almost certainly multifactorial, with psychodevelopmental,10 sociocultural,11 and genetic12 contributions to risk. For example, exposure to risk factors for dieting appears to elevate risk for BN,13 just as childhood exposure

121

122

Section II  Nutrition in Gastroenterology Eating disorders

Eating disorder, not otherwise specified

Anorexia nervosa

Bulimia nervosa

Unwilling to maintain minimal healthy weight Excessive fear of weight gain Distorted perception of weight or body and/or related medical danger Amenorrhea

Recurrent binge-eating Recurrent behavior to purge or neutralize excessive intake or control weight Excessive concern with weight or body shape

Binge/purge subtype

Restricting subtype

Recurrent binge-eating and/or purging

Restrictive pattern of eating without binge-eating or purging

Purging subtype

Non-purging subtype

Recurrent self-induced vomiting, and/or abuse of laxatives, enemas, diuretics, or stimulants

Behaviors to neutralize excessive intake limited to food restriction and exercise

Atypical or subthreshold symptoms Examples: Normal weight, recurrent purging, but no bingeeating Restrictive eating, with normal weight Binge-eating and purging once weekly Chewing and spitting out food

Binge-eating disorder Recurrent binge-eating without recurrent purging or other behaviors to neutralize the excess energy intake Binge-eating that is associated with distress and often with secrecy and shame Binge-eating that is often rapid and irrespective of hunger or satiety

Night-eating syndrome Evening hyperphagia Morning anorexia Nighttime wakings, snacking during nighttime wakings

Figure 8-1.  Key diagnostic features of the eating disorders.

Table 8-1  Behaviors Used to Neutralize Excessive Food Intake or to Prevent Weight Gain Purging behaviors Self-induced vomiting (including syrup of ipecac use) Laxative and/or enema abuse Diuretic abuse Stimulant abuse (e.g., caffeine, ephedra, methylphenidate, cocaine) Non-purging behaviors Excessive physical activity Fasting, skipping meals, restrictive eating pattern Inappropriate withholding or underdosing of insulin (among individuals with diabetes mellitus)

to negative comments about weight and shape elevate risk for BED.14 Body dissatisfaction in a social context in which thinness,15,16 self-efficacy, and control are valued may be an important means whereby dieting is initiated and disordered eating attitudes and behaviors ensue. Dietary restraint may precipitate a cycle of hunger, binge eating, and purging.17 Among numerous risk correlates, childhood gastrointestinal (GI) complaints have been found associated with earlier age of onset and greater severity of BN in a retrospective study,18 and picky eating and digestive prob­ lems were found prospectively associated with AN in adolescence.10 It has been suggested that physiologic vulnerabilities potentially may increase risk for an eating disorder. Neurobiologic targets have been identified as possibly playing a role in the pathogenesis of AN, BN, and BED. Researchers have studied the psychobiology of eating disorders and the

neurophysiologic correlates and determinants of energy intake, hunger, and satiety for decades. Findings highlight the multifactorial and phenotypically diverse nature of eating behavior. For example, energy intake is influenced by complex interactions among signaling molecules from peripheral systems (e.g., GI peptides, vagal stimulation) and central nervous system (CNS) neuropeptides and neuroamines. As is true in the search for treatments for obesity, it is unlikely that single mechanisms will become the basis for therapeutic interventions for eating disorders; however, the greater our understanding of the physiology of ingestive behavior, the more likely we are to establish integrated therapy models in the future. There is a vast literature on this topic, and what follows is simply a brief overview of the more commonly investigated mechanisms.

SATIETY

Serotonin has long been a focus of attention for its possible role in disrupted satiety. There is substantial evidence that altered 5-hydroxytryptamine (5-HT; serotonin) functioning contributes to dysregulated appetite, mood, and impulse control in eating disorders and persists after recovery from AN and BN, possibly reflecting premorbid vulnerability.19,20 There also is evidence that cholecystokinin (CCK) levels are altered in eating disorder populations. Findings for AN are inconsistent: There is some evidence that young women with AN have high levels of pre- and postprandial CCK, which may impede treatment progress by contributing to postmeal nausea and vomiting,21,22 whereas other reports have shown decreased CCK compared with that of controls.23 In those with BN, there is consistent evidence for an impaired satiety response, characterized by a blunted postprandial CCK response and satiety, as well as delayed

Chapter 8  Eating Disorders gastric emptying.24,25 In contrast, individuals with BED and obesity do not differ in postmeal CCK responses from those with obesity but no BED.26 The relationships between CCK, binge eating, and BMI need further clarification. Lastly, protein tyrosine (PYY) functioning also appears to be dysregulated in BN and AN, but not in BED. Young women with AN have higher levels of PYY, the intestinally derived anorexigen that elicits satiety, compared with controls, perhaps con­tributing to reduced food intake.27 In individuals with BN, expected elevations in PYY after meals are blunted,28,29 possibly playing a role in impaired satiety. A recent report found no differences between BED and non-BED groups in fasting levels and postmeal changes in PYY.30

APPETITE

The orexigenic peptide ghrelin is of interest for its role in eating disorders because it influences secretion of growth hormone (GH), stimulates appetite and intake, induces adiposity, and is implicated in signaling to the hypothalamic nuclei involved in energy homeostasis. There are two consistent findings in the literature examining ghrelin and AN: (1) circulating ghrelin levels are elevated, likely a consequence of prolonged starvation31,32; and (2) GH and appetite responses to ghrelin are blunted, suggesting altered ghrelin sensitivity.33,34 In BN, plasma levels of ghrelin are normal or elevated28,29; of most interest is the postprandial blunted response (i.e., reduced suppression of ghrelin35). Investigations of ghrelin functioning in individuals with BED have reported lower circulating levels of pre- and postmeal ghrelin, possibly reflecting down-regulation in response to chronic overeating and smaller decreases in ghrelin after eating.30

ENERGY STORAGE

Leptin and adiponectin are hormonal signals associated with longer term regulation of body fat stores. Leptin is also directly implicated in satiety through its binding to the ventral medial nucleus of the hypothalamus, an area termed the satiety center. Leptin and adiponectin are both altered in patients with eating disorders. A number of studies have found evidence for hyperadiponectinemia and hypoleptinemia in populations of underweight AN with reversal following restoration of weight36,37; increased adiponectin levels may act protectively to support energy homeostasis during food deprivation. Individuals with BN also exhibit decreased plasma levels of leptin, which are inversely correlated with length of illness and severity of symptoms.38 The mechanism of altered leptin functioning in BN is unclear because blunted postmeal leptin levels are not observed in individuals with BED.26 There are other mechanisms of interest, including but not limited to neuropeptide Y, GLP-1 and GLP-2, orexins A and B, the endocannabinoids, resistin (adipose tissue–specific secretory factor), pancreatic polypeptide, and brain-derived neurotrophic factor, but more research is necessary to clarify their roles in the pathophysiology of eating disorders. One high priority for research is clarifying whether observed psychobiological abnormalities are antecedents or consequences of disturbed eating behavior that return to normal after recovery, because this information would shed light on cause and possible treatment targets. Patients often report intense discomfort after eating as a reason that they continue to restrict intake. The discomfort may be dismissed as perceptual or psychological in the absence of any positive medical findings to support the symptoms, but there may in fact be disruptions in CNS or peripheral signals contributing to reported symptoms.

ONSET AND COURSE AN and BN most commonly have their onset in adolescence,39 and BED usually manifests in the early 20s,40 but eating disorders can occur throughout most of the lifespan and appear to be increasing in frequency in middle-aged and older women.41,42 Diagnostic migration from one eating disorder category to another is common.43 Lifetime comorbidity of AN, BN, and BED with other psychiatric disorders is high at 56.2%, 94.5%, and 63.6%, respectively.3 Mortality associated with AN and BN combined is five times higher than expected and is one of the highest mortality rates among mental disorders.1 Some data support the chronicity of AN, reporting that slightly under half of survivors with AN make a full recovery, with 60% attaining a normal weight and 47% regaining normal eating behavior; 34% improve but only achieve partial recovery, whereas 21% follow a chronic course.44 Other data suggest that recovery rates for AN may be more favorable than previously believed,3 with one large twin cohort study reporting a five-year clinical recovery rate of 66.8%.45 In contrast, after a five-year follow-up of 216 patients with BN and EDNOS, 74% and 78% of patients, respectively, were still in recovery.46 In a six-year longitudinal study of patients with BED, 43% of individuals continued to be symptomatic.47 In summary, despite well-established treatments available for the eating disorders, up to 50% of treated individuals continue to be symptomatic.48 Some prevention strategies developed for eating disorders show promise, including programs that induce cognitive dissonance about the “thin ideal.”49,50

DIAGNOSIS AND EVALUATION A substantial percentage of individuals with eating disorders in the United States do not receive treatment for this problem.3 Despite clear diagnostic criteria for the eating disorders, clinical detection often is problematic and up to 50% of cases may go unrecognized in clinical settings. Moreover, individuals with eating disorders are often reluctant to disclose their symptoms,51 and those with BN and BED can have a normal physical examination. Although individuals with AN are underweight by definition, this is easily missed in clinical settings. Even when noted on evaluation, the medical seriousness of low weight is frequently unappreciated.52 Finally, when an eating disorder is suspected or confirmed, patients may decline or avoid mental health care. Indeed, a feature of AN can be denial of the medical seriousness of symptoms.8 Given that many individuals with eating disorders initially present in primary care or medical subspecialty settings, recognition of clinical signs and symptoms across diverse health care settings will facilitate appropriate referrals and make diagnostic evaluation and treatment plans more efficient. One study has reported that individuals with BN are more likely to seek help for their GI complaints prior to seeking treatment for their eating disorder.53 Thus, familiarity with the diagnostic features and gastrointestinal complications of eating dis­orders will help to identify the most appropriate interventions, including the full spectrum of treatment resources available, for a comprehensive treatment plan. Formal screening for eating disorders can be timeconsuming, and although shorter measures are being developed,54 these have many limitations in clinical settings.55

123

124

Section II  Nutrition in Gastroenterology When an eating disorder is suspected, however, a directed clinical interview about restrictive or binge eating and inappropriate compensatory measures to control weight (see Table 8-1) is essential in determining the scope and severity of symptoms that underlie specific GI complaints and pose medical risk. Accurate and timely diagnosis of eating disorders is challenging for several reasons. First, patients may be unreliable reporters of their history and BN and BED may be present without any abnormal physical findings. Next, some dieting and exercise behaviors are appropriate, and discerning pathologic behavior that is consistent with a clinically significant eating disorder can be difficult. There is con­ siderable overlap in symptoms among the eating disorders; diagnostic specificity, however, is critical to effective management. Given the frequent reluctance of patients to disclose symptoms of an eating disorder, targeted history taking may be essential to making a prompt diagnosis. In some cases, an eating disorder may not be suspected nor the diagnosis confirmed until physical findings suggestive of purging are detected, a suggestive pattern is noticed in weight changes, and/or there is difficulty gaining weight, notwithstanding appropriate nutritional treatment and the exclusion of other potential causes for low weight.

ANOREXIA NERVOSA

Anorexia nervosa is characterized by an unwillingness or incapacity to maintain a minimally normal weight (commonly described as at least 85% of expected weight or a body mass index [BMI] of 17.5 kg/m2), fear of gaining weight (despite being thin), a disturbance in the way weight is experienced (e.g., a denial of the medical seriousness of being underweight or feeling fat despite emaciation), and amenorrhea (in postmenarcheal females). Individuals with AN typically restrict their food selections and caloric intake, but approximately half of those with AN also routinely binge-eat and/or engage in inappropriate compensatory behaviors, such as induced vomiting or laxative use to prevent weight gain (see Fig. 8-1). AN is subdivided further into restricting type (i.e., those who primarily control their weight through dieting, fasting, or exercising) and bingeeating–purging type (i.e., those who routinely purge calories to control weight and may or may not routinely binge-eat).8 In middle-aged and older women, new-onset AN may present in conjunction with difficulty making life transitions and fear of aging.42 The diagnosis of AN may be delayed when patients present to a GI specialty practice without disclosing their concerns and behaviors relating to weight. Presentation with GI complaints, even if related to real symptoms or disease, can sometimes prove to be a red herring, drawing attention away from and delaying diagnosis of an eating disorder. One study evaluated 20 consecutive patients presenting to a GI practice who were ultimately diagnosed with an eating disorder, and found that patients did not receive a diagnosis of an eating dis­ order for an average of 13 months after presentation. Notably, all patients stated a desire to gain weight and denied attempts to lose weight via exercise, purging, or dietary restriction.56 Individuals with AN are not always able or willing to frame their difficulty maintaining a healthy weight as intentional; thus, the diagnosis initially may be unsuspected and delayed.

BULIMIA NERVOSA

The clinical hallmark of BN is recurrent binge eating accompanied by inappropriate compensatory behaviors to control weight or to purge calories consumed during a binge. On

average, these behaviors must occur twice weekly for at least three months to meet diagnostic criteria.8 Also intrinsic to the diagnosis of BN is the excessive influence of weight and/or shape on self-image. By definition, binge eating is consumption of an unusually large amount of food during a “discrete period of time” (i.e., not overeating or grazing all day), accompanied by the feeling that the eating cannot be controlled.8 Many patients describe an emotional numbing during the period of eating. For some, this state appears to motivate the bingeing. Most clinicians are familiar with self-induced vomiting as the primary purging behavior, but individuals with BN often use alternative or additional means of preventing weight gain, including abuse of laxatives and/or enemas, diuretics (especially among health care workers), stimulants (including methylphenidate, cocaine, ephedra, and caffeine), underdosing of insulin (for those with diabetes mellitus), fasting or restrictive eating, and excessive exercise (see Table 8-1). Intentional consumption of gluten to promote weight loss in adolescents with celiac disease also has been reported.57 Whereas most compensatory behaviors to prevent weight gain fall within the diagnostic subtype of purging BN, excessive exercise and fasting are behaviors categorized within the subtype of nonpurging BN.8 Because of the absence of the more classic purging behaviors and because of their frequent indistinct nature, this variant of BN often is challenging to identify. As with overeating and dieting, it frequently is difficult to determine the line between culturally normative and pathologic behavior with excessive exercise. Generally, clinical suspicion should be raised when an individual continues to exercise despite an injury or illness, or if he or she is exercising routinely in excess of what a coach is recommending for the team. It is recommended that clinicians explore the presence of purging behaviors if an eating disorder is suspected. Although it is not certain that a patient will respond candidly, individuals probably are more likely than not eventually to disclose information about symptoms when asked.51 Some patients report feeling relieved when clinicians pose such questions if they previously had not been able to discuss their symptoms. On occasion, however, patients report learning about techniques from clinicians’ questions, so it is advisable to provide a psychoeducational context for the questions (e.g., by conveying serious physical consequences associated with the behavior, such as ipecac use) and to avoid introducing information about a dangerous behavior (e.g., underdosing insulin), depending on the clinical context. Patients also benefit from learning that treatment is available and that their clinicians understand the illness. Whereas all these purging and other behaviors aimed at neutralizing calorie intake and controlling weight can pose medical risks when chronic, some of them pose more immediate, and potentially lethal, consequences. Patients should be informed of these acute life-threatening risks and steps should be taken to eradicate such behaviors immediately. For example, because of the serious neurotoxicity, cardiotoxicity, and risk of death associated with repeated syrup of ipecac ingestion,58 its ongoing use is a clinical emergency and may require immediate hospitalization. Many patients are unaware of the serious risk associated with syrup of ipecac use. Similarly, ephedra, now banned in the United States, poses risk of stroke or adverse cardiac events, even in young adults.59 Some ephedra-free supplements marketed as weight loss agents also may be proarrhythmic and pose medical risks.60 Although patients find it difficult to abstain from purging behaviors, they may be willing to

Chapter 8  Eating Disorders substitute less immediately harmful behaviors while treatment is initiated. Individuals with BN are excessively concerned with weight and body shape. For example, they may be preoccupied with a running mental tally of calories and plans to neutralize them to prevent weight gain. Their self-image frequently is poor and anchored to their weight. It is not unusual for individuals with AN or BN to weigh themselves daily, even several times each day, and experience fluctuations in self-esteem and mood based on the result.

EATING DISORDER NOT OTHERWISE SPECIFIED

EDNOS covers a broad range of clinical manifestations, including atypical symptoms, symptoms of BED, symptoms consistent with NES, and subthreshold, yet clinically significant, eating disorders. Although EDNOS is a residual category, it nonetheless is the most commonly diagnosed eating disorder in outpatient settings, and therefore there is much interest in refining diagnostic categories of eating disorders.61 EDNOS currently includes eating disorders that do not meet threshold criteria for duration or frequency for AN or BN as well as BED, NES, purging disorder and other atypical variants. Diagnostic classification of eating disorders will be updated in the DSM-V, with an expected publication date of 2012.

BINGE EATING DISORDER

BED is a variant of EDNOS, although there is a substantial literature to support its prevalence and consistent response to specific therapeutic strategies. Like BN, BED is characterized by recurrent binge-eating. To meet provisional criteria for BED, the binge eating episodes must occur two days per week, on average, for at least six months. Unlike BN, however, BED is not associated with recurrent inappropriate compensatory behaviors to prevent weight gain. BED is distinguished from nonpathologic overeating by several possible associated symptoms, including rapid eating, eating irrespective of hunger or satiety, eating alone because of shame, and negative feelings after a binge.8 Apart from overweight or obesity, BED patients frequently present without any specifically associated physical findings. Although in some cases binge eating associated with BED may cause or perpetuate weight gain, many with BED develop symptoms only after they have become overweight. Individuals with BED frequently are distressed enough about their symptoms to seek medical help, although they may present seeking a solution to their weight gain rather than their binge eating. A substantial percentage of patients seeking weight loss treatment will have comorbid BED or NES. Therefore, medical subspecialists are likely to encounter these patients before they have been diagnosed with BED.

NIGHT EATING SYNDROME AND NOCTURNAL SLEEP-RELATED EATING DISORDER

NES is a pathologic eating pattern that may be considered a variant of EDNOS. First described in 1955,62 it, too, is characterized by recurrent bouts of overeating—but not necessarily bingeing—without associated inappropriate compensatory behaviors to prevent weight gain. As such, some individuals may appear to meet criteria for NES and BED, but these are distinct syndromes with relatively little overlap,7,63 and proposed criteria for the syndrome exclude a concomitant diagnosis of BN or BED.64 There is no clear consensus on core criteria for NES, although most investigators propose morning anorexia, evening hyperphagia, and sleep disturbance (operationalized in various ways); some

propose the additional criterion of eating in relation to sleep disturbance, such as during a nighttime awakening.65 In one study, NES in obese subjects was associated with an average of 3.6 awakenings/night compared with just 0.3 awakenings/night for matched controls. Subjects with NES ate during 52% of their awakenings, taking in a mean of 1134 kJ/ episode, considerably less than the usual intake of a binge associated with BN or BED.64 NES also can occur in nonobese individuals66 but is more common in the obese and may contribute to poor outcome in weight loss treatment programs.64,67 NSRED is also characterized by nighttime snacking, but individuals typically are totally or partially unconscious (e.g., they are in stage 3 or 4 sleep) during the snacking and frequently do not remember it.7

PURGING DISORDER

Emerging evidence raises the possibility of an additional distinctive eating disorder variant characterized by recurrent purging symptoms in the absence of clinically significant binge pattern eating. The proposed name for this diagnostic category is purging disorder. Crossover between this variant and BN appears to be rare, lending support to the hypothesis that this represents a distinctive clinical phenomenon68; data also support comparable severity to BN. Lifetime prevalence of purging disorder has been estimated as 1.1% to 5.3% of young adult women. Course, outcome, and treatment strategies for purging disorder require further research.69

DIFFERENTIAL DIAGNOSIS

The differential diagnosis of the eating disorders includes evaluation and exclusion of medical causes of weight loss, weight gain, anorexia, hyperphagia, vomiting, and other associated symptoms. These considerations are especially germane in cases of atypical or early- or late-onset eating disorders.42 Medical causes of appetite and/or weight loss include hyperthyroidism, diabetes, malignancy, and infectious diseases, among the systemic disorders, and substance abuse, depression, dementia, delirium, and psychosis. Illnesses associated with weight gain include hypothyroidism, Cushing’s disease, and organic brain disease. The differential diagnosis of hyperphagia is broad and includes PraderWilli syndrome, dementia (including Alzheimer’s disease), and intracranial lesions. Hyperphagia also has been associated with the use of certain medications, particularly many of the psychotropic agents (e.g., lithium, valproate, tricyclic antidepressants, mirtazapine, and conventional and atypical antipsychotic agents), pregnancy,70 and poststarvation refeeding.71 Psychiatric illnesses associated with loss of appetite and weight loss include major depression, anxiety, and substance-use disorders. Moreover, comorbid psychiatric illness is common among those with eating disorders,3 and frequently complicates their diagnosis and treatment. Thus, identi­fication of excessive concern with weight and food intake, unrealistic or inappropriate weight goals, or resistance to attempts to restore normal weight and/or limit excessive exercise can be helpful in distinguishing an eating disorder from another psychiatric illness or in revealing the presence of an underlying comorbid eating disorder. Because individuals with BN and EDNOS can have an unremarkable physical examination on presentation, the diagnosis may remain obscure until the patient discloses his or her symptoms, or until the clinician suspects an eating disorder based on other elements of the clinical history (e.g., weight fluctuations or menstrual irregularities). Although there potentially is much phenomenologic overlap among the eating disorders and individuals do cross

125

126

Section II  Nutrition in Gastroenterology Table 8-2  Distinguishing Features of Eating Disorders

Eating Disorder

Physical Signs Included in Diagnostic Criteria

Restrictive Pattern Eating

Binge Pattern Eating

AN

Amenorrhea, underweight

Typically

May occur in AN

BN

None (patients generally are of normal weight or overweight) None (patients frequently are overweight or obese) Nighttime awakenings (patients range from normal weight to obese)

May occur as behavior to control weight No

None (patients range from underweight to obese)

Frequently

Must occur twice weekly for at least 3 mo Must occur at least 2 days/wk for at least 6 mo Not intrinsic to the disorder; snacks of high-calorie foods typically contain fewer calories than a binge Frequently

BED NES

EDNOS

No

Purging and Other Behaviors to Control Weight or Neutralize Effects of Calorie Intake

Excess Concern with Body Image or Weight

Purging may occur in up to half of those with AN Must occur to meet diagnostic criteria

Yes

No

Not intrinsic to current research diagnostic criteria8 Not intrinsic to the syndrome

No

Frequently

Yes

Frequently occurs but not necessary to meet diagnostic criteria

AN, anorexia nervosa; BED, binge-eating disorder; BN, bulimia nervosa; EDNOS, eating disorder, not otherwise specified; NES, night-eating syndrome.

over from one diagnostic category to another (Table 8-2),43 categories are mutually exclusive, according to DSM-IV* diagnostic criteria.8 Even though a transdiagnostic approach to eating disorders classification and treatment has been proposed,72 differing responses to treatment make it desirable to establish a clear diagnosis to optimize care. A weight criterion distinguishes anorexia nervosa from bulimia nervosa in some cases. Individuals who are substantially underweight (e.g., 85% or less of expected body weight) and who otherwise meet criteria for AN most likely should be classified as having AN, even if bingeing, purging, or both are present. Individuals with BED or BN also can have symptom overlap. BN is distinguished by recurrent purging and other behaviors directed at neutralizing excessive calorie intake so as to prevent weight gain, as well as an excessive concern with weight.

NUTRITIONAL AND MEDICAL EVALUATION

In addition to excluding medical causes of weight and appetite changes, medical evaluation for confirmed or suspected eating disorders includes obtaining a full history of the patient’s eating behaviors, with attention to daily number of calories ingested, purging behavior (e.g., vomiting, use of ipecac, or laxative use), and exercise patterns. Often, medical evaluation will be guided by an assessment of nutritional status, which includes determination of the appropriateness of weight for height, age, and gender.

NUTRITIONAL EVALUATION

There are several established means for evaluating nutritional status in the office, central to which is measuring weight and height (see Chapter 5). Assessment of the appropriateness of weight for height is one of the key factors intrinsic to determining the urgency of medical and psychiatric care. For patients with AN, it is important not to rely on self-reported weight, given the strong possibility of an inaccurate report. Individuals with AN often go to great effort to conceal their low weights. For example, some patients “water load” prior to a clinical encounter, some attach weights to themselves, and others layer loose and bulky clothing to create the illusion of being of normal weight. Assessment of weight, therefore, should factor in the possibility that a patient may wish to conceal a low weight or weight loss. Some clinicians will find it helpful to have a scale in a private area (i.e., not in a hallway) and a clear and consistent protocol for weighing patients with AN. This might include asking them to void prior to being weighed, to change into a hospital gown, and to remove heavy jewelry. When patients have a history of consuming water prior to an appointment to increase their measured weight, it may be helpful to check a urine specific gravity in order to adjust interpretation of the office weight. Standard means of evaluating the appropriateness of weight for height include use of the BMI. This is calculated as follows: 2 BMI = weight ( in kg ) height ( in m )

*The multiaxial diagnostic system inherent in the Diagnostic and Statistical Manual (of Mental Disorders (DSM) reflects several dimensions relevant to health, mental health, and functioning.8 Axis I refers to psychiatric illnesses exclusive of personality disorders and mental retardation. Axis II refers to personality disorders and mental retardation. Axis III refers to medical conditions that are potentially relevant to the mental disorder(s). Axis IV refers to acute and chronic psychosocial stressors. Axis V refers to the level of psychological, social, and occupational functioning.

Although BMI may not be an appropriate standard for evaluating a healthy weight status in professional athletes, with relatively high lean muscle mass, and in some ethnic groups (e.g., Polynesians may have a different cut point for obesity73), BMI generally is appropriate for men and women aged 18 years or older. A BMI within the range of 18.5 to

Chapter 8  Eating Disorders 24.9 kg/m2 for men and women is considered normal. A BMI of 17.5 kg/m2 or less is the threshold for meeting the underweight criterion for AN in the ICD-10 (International Classification of Diseases 10).8,74 A BMI in the range of 25 to 29.9 kg/m2 is consistent with overweight, and a BMI higher than 30 kg/m2 reflects obesity.75 An alternative means for evaluating weight for height that is especially useful to assess nutritional compromise and weight recovery in those with eating disorders expresses the patient’s weight as a percentage of expected body weight. The formula is as follows: Expected body weight (%) = (patient weight expected weight for height and gender ) × 100% Expected weight for height = 100 pounds + 5 pounds per inch above 5 feet ± 10% for women, 106 pounds + 6 pounds per inch above 5 feet ± 10% for men and, if the patient is shorter than 5 feet, then the same number of pounds per inch is subtracted for each inch below 5 feet.76 Although this is a linear equation (compared with the quadratic equation for BMI) and may be less useful at extreme heights, it is straightforward to calculate. Moreover, conceptually, this formula may be easier for patients and families to understand, especially in setting weight goals or limits. A 90% to 110% range of expected body weight is considered within the normal range and is a good place to begin for setting weight gain goals for patients with AN. Within this range, the goal will be refined by clinical history (including the patient’s history of baseline, minimal and maximal weights), whether and when menses return, and medical parameters, such as reversal of bone loss. Patients below 85% of expected body weight likely meet the weight criterion for AN, and those below 75% of expected body weight are seriously nutritionally compromised and generally require inpatient care.77 For patients who are overweight or obese (>110% or >120% expected body weight, respectively), it may not be realistic or desirable to set weight goals within the normal range. Considerations in weight management are discussed subsequently. Seriously nutritionally compromised patients require inpatient care for both efficacy and safety of weight management. For underweight patients without this degree of compromise, the primary goals of nutritional management are increasing caloric requirements to regain weight, ensuring adequate intake and balance of macro- and micronu­ trients, and reestablishing a dietary pattern of three meals daily. Patients’ diets are supplemented routinely with calcium (if dietary intake is inadequate) and multivitamins containing vitamin D. Some may require additional dietary guidance and adjustments because many patients restrict not only calories, but specific foods or food groups as well. For patients with BN, BED, and EDNOS, dietary intervention includes moderating excessive calorie intake and establishing a pattern of eating that is less vulnerable to emotional cues and excessive hunger. Many patients with eating disorders are quite knowledgeable about nutrition and commonly wish to avoid meeting with a nutritionist. Conversely, information from a nutritional assessment is invaluable to the treatment team and even well-informed patients are likely to benefit from reinforcement of more healthful food choices, meal patterns, and appropriate intake.

Medical Evaluation

Medical evaluation includes a clinical history with special attention to weight fluctuations and any purging or other

Figure 8-2.  Dental erosions resulting from chronic vomiting. (Adapted with permission from the Department of Psychiatry, Massachusetts General Hospital, Boston.)

inappropriate behaviors to neutralize calorie intake to control weight (see Table 8-1). Ascertainment of syrup of ipecac use (as an emetic) and nonadherence to insulin protocols in patients with diabetes mellitus is essential, given the potentially lethal sequelae of these behaviors.78 Symptoms of medical complications of undernutrition, overnutrition, excessive exercise, or purging should be assessed and a menstrual history should be clarified. Physical examination includes a comprehensive assessment of potential complications of nutritional deficiencies, underweight, overweight, excessive exercise, and purging behaviors. If an eating disorder is suspected, physical examination may reveal signs to confirm nutritional compromise (e.g., bradycardia, hypotension, hypothermia, lanugo, breast tissue atrophy, muscle wasting, peripheral neuropathy) or to suggest chronic purging (e.g., Russell’s sign, an excoriation on the dorsum of the hand from chronic scraping against the incisors); hypoactive or hyperactive bowel sounds; an attenuated gag reflex79; dental erosion (perimolysis; Fig. 8-2)80; or parotid hypertrophy (Fig. 8-3).81 Medical complications of behaviors associated with AN, BN, BED, and EDNOS are potentially serious and are too numerous to review in detail here; selected complications are listed in Table 8-3. Complications that are common and/ or associated with serious morbidity should be actively sought on physical examination and laboratory studies, so that appropriate interventions can be initiated. Examples of such important and common findings include abnormal vital signs (e.g., hypotension, orthostatic hypotension, bradycardia, hypothermia), low weight or overweight, osteopenia or osteoporosis,82 and dental pathology (e.g., perimolysis [erosion of the tooth enamel], caries, or both).81,83,84 Cardiac complications can be lethal and include prolonged QT interval, QT dispersion, ventricular arrhythmias, and cardiac syncope.85,86 Neurologic findings in AN include cortical atrophy and increased cerebral ventricular size.87 Endocrinologic abnormalities include menstrual abnormalities, low serum estradiol levels, low serum testosterone levels, hypercortisolism, and euthyroid sick syndrome, with resultant hypotension and cold intolerance.88 Reported complications of eating disorders during pregnancy include miscarriage, inadequate weight gain of the mother, intrauterine growth retardation, premature delivery, infants of low birth weight and low Apgar scores, and perinatal death.89-92

127

128

Section II  Nutrition in Gastroenterology b-human chorionic gonadotropin and possibly a serum prolactin level therefore are recommended. Additional studies such as follicle-stimulating hormone (FSH) to evaluate ovarian function or neuroimaging studies to exclude a pituitary lesion may be indicated in some clinical scenarios. Bone densitometry using dual-energy x-ray absorptio­ metry (DEXA) scans of the hip and spine are useful in identifying bone loss and can be repeated after a year to assess further bone loss if disease continues. Osteopenia and osteoporosis may be present in as many as 90% and 40%, respectively, of women with AN, and are associated with risk of fractures and kyphosis.103,104 An electrocardiogram is recommended to evaluate patients with eating disorders because idiopathic QT prolongation can occur with AN,105 and QT intervals may be prolonged in those with BN and EDNOS, even in the absence of hypokalemia.106 Use of pharmacologic agents that can prolong the QT interval (e.g., olanzapine or desipramine), as well as purging that leads to hypokalemia, may further increase the risk of cardiac arrhythmia in this patient population. Abuse of ipecac may result in potentially fatal cardiotoxicity and arrhythmias.78 Figure 8-3.  Patient with parotid hypertrophy resulting from chronic vomiting. (Adapted with permission from the Department of Psychiatry, Massachusetts General Hospital, Boston.)

Laboratory Evaluation

Whereas choice of laboratory studies to evaluate medical complications of eating disorders will depend on the clinical history and presentation, it is useful to obtain serum electrolyte levels among individuals in whom AN or BN is suspected or confirmed. For example, hypokalemia occurred in 4.6% of a large sample of outpatients with eating disorders in one study93 and in 6.8% of individuals with BN in another moderately sized sample.94 In the latter study, hypokalemia was significantly more common in patients with BN than in those without BN. Although assessment for hypokalemia is not efficient for identifying occult cases of BN, it will assist in the identification and monitoring of individuals at risk for cardiac arrhythmias secondary to their eating disorder. Hypochloremia, hypomagnesemia, hyponatremia, hypernatremia, and hyperphosphatemia also are seen in patients with eating disorders.88,94-96 In addition, for patients with AN, a serum glucose determination is recommended to identify hypoglycemia, which can be severe in this population.97 Although hyperamylasemia reportedly is common in BN (i.e., in 25% to 60% of cases), laboratory analysis of serum amylase generally is not clinically useful for detecting BN or gauging the severity of bingeing and purging symptoms.98 An elevated serum amylase level in a patient with AN or BN often reflects increased salivary isoamylase activity98,99; however, pancreatitis should be considered, when clinically appropriate, given its occurrence in this patient population. A complete blood count is recommended to assess for anemia, neutropenia, leukopenia, and thrombocytopenia among patients with AN. A retrospective study of 67 patients with AN found that 27% had anemia, 17% had neutropenia, 36% had leukopenia, and 10% had thrombocytopenia.100 Evaluation of the cause of amenorrhea is suggested, even if it is presumed to be related to decreased pulsatility of gonadotropin-releasing hormone secondary to weight loss.88 Menstrual irregularities are common among women with eating disorders, but women with symptomatic eating disorders still may be menstruating at presentation101 and women with AN can become pregnant102; a quantitative

GASTROINTESTINAL ABNORMALITIES ASSOCIATED WITH EATING DISORDERS GI signs and symptoms are common in those with eating disorders (Tables 8-3, 8-4). It has been asserted that the most dramatic changes in bodily function caused by AN are in the GI tract.107 There is also evidence that many individuals with eating disorders may present with a GI complaint prior to seeking treatment for an eating disorder. In one small retrospective study, 8 of 13 inpatients with eating disorders had sought care for a GI complaint, and 6 of them had sought such GI care before tending to their eating disorder.108 Several cross-sectional studies of hospital inpatients with eating disorders have suggested that 78% to 98% have GI symptoms.109-113 For example, constipation is a frequently reported symptom in AN and BN; in a study of 28 inpatients with an eating disorder, 100% of patients with AN and 67% of patients with BN had constipation.112 Nausea, vomiting, gastric fullness, bloating, diarrhea, and decreased appetite also are seen commonly in AN and bloating, flatulence, decreased appetite, abdominal pain, borborygmi, and nausea commonly are reported in BN. In one study of 43 inpatients with severe bulimia nervosa, 74% reported bloating, 63% reported constipation, and 47% reported nausea; borborygmi and abdominal pain also were more frequent than in the comparison group of healthy controls.110 Moreover, certain GI symptoms have been shown to be more common in dieters (specifically, abdominal pain, bloating, and diarrhea)114 and in those with binge eating (nausea, vomiting, and bloating) than in normal controls.115 A large study of obese individuals with GI symptoms found a strong association between BED and abdominal pain and bloating, after adjusting for BMI.116 Finally, a study of 101 consecutive women admitted to an inpatient eating disorders program found the vast majority of study participants (98%) to have functional gastrointestinal disorders (FGIDs). Among these, 52% had irritable bowel syndrome, 51% had functional heartburn, 31% had functional abdominal bloating, 24% had functional constipation, 23% had functional dysphagia, and 22% had functional anorectal pain; 52% of respondents met criteria for three or more FGIDs. Whereas the authors found psychological predictors for several of the FGIDs, they were not associated with functional

Chapter 8  Eating Disorders Table 8-3  Selected Clinical Features and Complications of Behaviors in Patients with Eating Disorders Clinical Feature or Complication Associated with Weight Loss and Food Restriction or binge-eating in anorexia nervosa

Associated with Purging or refeeding Behaviors in anorexia nervosa, bulimia nervosa, or EDNOS

Cardiovascular

Arrhythmia Bradycardia Congestive heart failure (in refeeding syndrome) Decreased cardiac size Diminished exercise capacity Dyspnea Hypotension Mitral valve prolapse Orthostasis Prolonged QT interval QT dispersion Syncope

Ventricular arrhythmia Cardiomyopathy (with ipecac use) Prolonged QT interval Orthostasis Syncope

Dermatologic

Brittle hair Dry skin Hair loss Hypercarotenemia Lanugo

Russell’s sign (knuckle lesions from repeated scraping against the incisors)

Oral, pharyngeal

Cheilosis

Dental erosion and caries Sialadenosis Pharyngeal and soft palatal trauma Angular cheilitis Perimolysis Vocal fold pathology

Gastrointestinal*

Anorectal dysfunction Delayed gastric emptying Elevated liver enzyme levels Elevated serum amylase levels Gastroesophageal reflux Hepatic injury Pancreatitis Prolonged whole-gut transit time Rectal prolapse Slow colonic transit Superior mesenteric artery syndrome

Abdominal pain Acute gastric dilatation Barrett’s esophagus Bloating Constipation Delayed gastric emptying Diarrhea Dysphagia Elevated liver enzyme levels Elevated serum amylase levels Esophageal bleeding Esophageal ulcers, erosions, stricture Gastroesophageal reflux Mallory-Weiss tear

System Affected

During refeeding: Acute gastric dilatation, necrosis, and perforation Elevated liver enzyme levels Hepatomegaly Pancreatitis

Endocrine and metabolic

Amenorrhea Euthyroid sick syndrome Hypercholesterolemia Hypocalcemia Hypoglycemia Hyponatremia Hypothermia Low serum estradiol, low serum testosterone levels Osteopenia, osteoporosis Pubertal delay, arrested growth As part of the refeeding syndrome: Hypomagnesemia Hypophosphatemia

Gastroesophageal reflux Gastric necrosis and perforation Hematemesis Pancreatitis Prolonged intestinal transit time Rectal bleeding Rectal prolapse Hypercholesterolemia Hyperphosphatemia Hypochloremia Hypoglycemia Hypokalemia Hypomagnesemia Hyponatremia Hypophosphatemia Metabolic acidosis Metabolic alkalosis Secondary hyperaldosteronism

129

130

Section II  Nutrition in Gastroenterology Table 8-3  Selected Clinical Features and Complications of Behaviors in Patients with Eating Disorders—cont’d Clinical Feature or Complication Associated with Weight Loss and Food Restriction or binge-eating in anorexia nervosa

Associated with Purging or refeeding Behaviors in anorexia nervosa, bulimia nervosa, or EDNOS

Acute kidney injury Amenorrhea Atrophic vaginitis Breast atrophy Infertility Pregnancy complications (including low birth weight, premature birth, and perinatal death)

Abnormal menses Azotemia Pregnancy complications (including low birth weight infant)

Neurologic

Cognitive changes Cortical atrophy Delirium (in refeeding syndrome) Peripheral neuropathy Ventricular enlargement

Stroke (associated with ephedra use) Neuropathy (with ipecac use) Reduced or absent gag reflex

Hematologic

Anemia Leukopenia Neutropenia Thrombocytopenia

System Affected Genitourinary and reproductive

EDNOS, eating disorder, not otherwise specified. *Gastrointestinal complications associated with binge pattern eating in any of the eating disorders, are not all listed, and include weight gain, acute gastric dilatation, gastric rupture, gastroesophageal reflux, increased gastric capacity, and increased stool volume.

Table 8-4  Common Gastrointestinal Symptoms in Patients with Eating Disorders Abdominal pain Belching Bloating Borborygmi Changes in appetite Constipation Diarrhea Dyschezia Flatulence Nausea Vomiting

abdominal bloating or functional dysphagia in this study population.109 Specific GI findings are commonly associated with eating disorders. Delayed whole-gut transit time98 and delayed gastric emptying appear to be common among inpatients with AN or BN.24,117-120 Delayed colonic transit has been reported in AN.121 Mild esophagitis is common (e.g., 22% of a case series of 37 consecutive patients) in patients with chronic BN, but more serious esophageal disease is rare.122,123 Abnormal esophageal motor activity has been reported in AN and BN.124,125 Barrett’s esophagus, Mallory-Weiss tears, and gastroesophageal reflux have been reported in association with chronic vomiting associated with BN.126 Unusual GI manifestations and catastrophic complications have been described in case reports of patients with eating disorders, including acute gastric dilation, gastric necrosis and perforation, and occult gastrointestinal bleeding (attributed to transient intestinal ischemia in the setting of endurance running).127-129 Rectal bleeding and rectal prolapse have been reported in patients with AN and BN.130,131

Other studies have found abnormal gastric function in patients with eating disorders, including diminished gastric relaxation (in patients with BN),132 bradygastria (in patients with AN, BN, and EDNOS),133 and higher gastric capacity (in patients with BN).134 Some evidence suggests that the GI abnormalities associated with eating disorders may be related to the duration or presence of active eating disorder symptoms. Physiologic sequelae of disordered eating, such as contracted or expanded gastric capacity, altered gastric motility, delayed large bowel transit (through reflex pathways),135 and possibly blunted postprandial cholecystokinin release, may perpetuate symptoms that exacerbate the excessive body image concern driving abnormal eating patterns.24 There is evidence that subjective reports of GI symptoms do not correlate well with physiologic data in patients with eating disorders.136 GI findings associated with eating disorders are listed in Table 8-3. In one study,137 elevated liver biochemical test results were documented in 4.1% of 879 patients presenting for treatment of an eating disorder. A probable cause distinct from the eating disorder was identified for 47% of the study participants but, for the remaining 53% of subjects, the abnormal results could not be attributed to a condition other than their eating disorder. Elevated enzymes were seen in underweight and normal-weight study participants. The results of this study suggest that abnormal liver biochemical tests are neither a specific nor a common marker for an eating disorder, and other possible causes should be excluded before attributing such abnormality to the eating disorder.137 In contrast, a study of 163 adolescent and young adult women outpatients with AN or EDNOS and low weight (excluding those with acute illness, alcohol abuse, hepatitis from viral or other known causes, and on medications associated with elevated liver enzyme levels) found elevated aminotransferase levels (alanine aminotransferase and/or aspartate aminotrans­ferase) in 19.6% of AN patients with a BMI lower than 16 kg/m2; 8.7% of AN patients with BMI higher than 16 kg/m2, and 15.2% of low-weight EDNOS

Chapter 8  Eating Disorders patients.138 Elevated liver biochemical test results and hepatomegaly also are observed on the initiation of refeeding in AN.137,139 There also are several case reports of severe liver dysfunction or damage in patients with AN attributed to malnutrition and associated hypoperfusion.140-142 Although many of the common GI complications of eating disorders are relatively benign, others, such as acute gastric dilation, gastric necrosis, and gastric rupture,143-147 although uncommon, are serious or even catastrophic. Esophageal rupture is another potentially catastrophic risk with chronic vomiting.139 Acute pancreatitis has been reported in patients with AN and BN,145,148,149 and also can be associated with refeeding in AN.138 In addition, there is a case report of severe steatosis resulting in fatal hepatic failure in a patient with severe AN150 and one of death resulting from duodenal obstruction secondary to a binge in a patient with BN.151 Both help-seeking and diagnosis may be delayed or com­ plicated by an undisclosed or unrecognized eating dis­ order.127,145,152 Conversely, esophageal dysfunction can be obscured by bulimic symptoms79,124 and can be misdiagnosed as AN.125 Superior mesenteric artery (SMA) syndrome can complicate AN and occurs when the support of the SMA is lost with weight loss and the duodenum is compressed between the aorta and the SMA. Because it manifests with vomiting, a concurrent diagnosis can be missed if this symptom is attributed to the eating disorder.153

MANAGEMENT Optimally, management of patients with eating disorders includes integration of mental health, nutrition, and primary care (Fig. 8-4). Occasionally, medical subspecialty consultation and care are helpful. Multidisciplinary management is desirable for several reasons. First, patients are at risk of medical, psychological, and nutritional complications of their disease. Second, patients commonly selectively avoid care essential to their ultimate recovery. For example, a patient may wish to avoid the detection of an injury so that she or he can continue to participate in a team sport; another may find it difficult to undergo the psychological work necessary to address antecedents of her illness; or another may wish to bypass active weight management. Conversely, a patient may attempt to pursue relief for specific medical complications to the exclusion of appropriate psychological or nutritional therapies. It often is helpful, if not essential, to establish a treatment agreement at the outset of care for a patient with an eating disorder. This is particularly relevant for patients for whom the severity of their symptoms may compromise medical and psychological health to a degree that hospital level care is likely during the course of treatment. A treatment agreement allows the caregivers to establish initial treatment goals and criteria for which they may wish to adjust the level of intensity of care. This will allow transparency of expectations for the patient, facilitate a rapid response to emerging crises, and help avoid split opinions among the team during the course of care. A treatment agreement also clarifies for patients the contingencies for nonadherence or poor health. As part of the initiation of care, patients should be asked to give permission for open communication among the members of the clinical team. If a patient cannot agree to this, it signals potential difficulties in providing coordinated care, and the lack of agreement should be reconciled. Depending on the patient’s age and circumstances, a plan for how and what information will be shared with parents also should be established.

Evaluation

Education and involvement of patient Identification of most appropriate level of care

Outpatient

Partial hospitalization

Identification of interdisciplinary team and referrals

Primary care clinician(s)

Medical specialty care

Nutritionist

Inpatient

Residential care

Referrals

Mental health clinician(s)

Psychotherapist

Psychopharmacologist

Adjunctive psychotherapist(s) (e.g., group or family) Division of tasks (e.g., weighing, medical parameter surveillance, dietary plan, monitoring symptom frequency and severity) Consensus on whether there will be any weight, symptom severity, or symptom frequency parameters that necessitate a more intensive level of care, adjunctive treatment, or restriction of activities Plan for how often and how to communicate

Obtain permission from the patient for communication Draft treatment agreement if warranted

Initiate treatment Reevaluate and adjust care as necessary Figure 8-4.  Key components of team management of patients with an eating disorder.

PSYCHIATRIC TREATMENT

Psychiatric treatment generally begins with psychotherapy. In many cases, pharmacotherapy is useful as an adjunctive treatment for BN and BED. Active weight management is indicated for AN, and there is a role for weight loss treatment in some patients with BED. Usually, psychotherapy can be used to support weight management goals, although optimally it should be coordinated with the efforts of the nutritionist and primary care clinician on the team. Regardless of the mode of psychotherapy chosen, specific behavioral strategies directed at establishing normal eating patterns and drawing the patient’s attention to triggers for abnormal patterns can augment treatment. Among these, patients are encouraged to identify and avoid emotion-, schedule-, and food-related triggers to episodes of bingeing

131

132

Section II  Nutrition in Gastroenterology

Time

Location

Food consumed

Feelings about Eating Before

During

After

Triggers

7 a.m.

At home

Coffee, 1/2 bagel

Hungry

OK

Good—only ate 1/2 of the bagel

None; it was breakfast time

9 a.m.

At desk

Diet cola

Avoiding a snack

OK

Glad to avoid food

None

11 a.m.

At desk

Diet cola

Hungry—trying to avoid eating

OK

Glad to avoid food

None

1:30 p.m.

Cafeteria

Salad with fat-free dressing on the side and a diet cola

Very hungry— trying not to eat too much

OK since dressing was low-cal

Glad to avoid a bigger lunch than I had

None

2:00 p.m.

At desk

3 mini candy bars, and a cookie

Starving

Gross

Gluttonous

Candy looked tempting and I was hungry

4:00 p.m.

In meeting

2 cups of coffee

Still hungry

OK

Good; glad to avoid food

Hungry again

7:00 p.m.

Walking home

Pizza (3 slices)

Starving: upset about interaction at work; want comfort food

Temporarily distracted

Terrible—way too many calories

Boss reprimanding me

8:45 p.m.

At home in front of TV

Most of a pint of ice cream, a box of cookies, 3 donuts, and a spoonful of peanut butter

Still upset; feeling that I’ve already blown it for the day and may as well eat more and then purge

Numb

Disgusted with myself for bingeing but relieved after purging

Still upset about work; knowing there was unfinished cookie dough ice cream in freezer; roommate left donuts; alone

Figure 8-5.  Food journal page from a hypothetical patient with bulimia nervosa.

and to plan three regular meals and two between-meal snacks to prevent excessive hunger. Finally, a food journal (Fig. 8-5) kept for a few days and reviewed in a treatment session will help many patients identify relationships among psychosocial stressors, hunger, and symptoms and may provide a concrete framework from which to relate symptoms to other psychological concerns.

Psychotherapy

A variety of psychotherapies have established efficacy for the eating disorders. Recent guidelines and reviews have summarized findings from empirical studies and highlighted the paucity of recommendations for treatment of AN and EDNOS.77,154-157 Cognitive behavioral therapy (CBT) and interpersonal therapy (IPT) have received a great deal of research attention for the treatment of eating disorders. CBT is a structured, manual-based approach that addresses the relationships among thoughts, feelings, and behaviors; IPT is another short-term therapy focused on present-day interpersonal events and roles in relationships. The choice of psychotherapeutic modality will be guided by the diagnosis, medical and psychiatric comorbidities, desirability of targeting the eating disorder symptoms versus bro­adening the therapeutic goals, treatment history, patient strengths and preferences, and the availability of care. Initial recommendations should be evidence-based when possible; however, clinical judgment is important for identifying individual needs and situations in which alternative treatment choices are appropriate.154 In practice, patients

with AN or EDNOS can benefit from a flexible approach to treatment that uses appropriate components of the various therapeutic modalities because there is a dearth of empirical data to enable evidence-based recommendations and because some patients do not respond to evidencebased treatments.158 There is limited empirical evidence for AN treatments. The one consistent finding is that family therapy focused on parental control of nutrition emerges as the treatment of choice for adolescents, particularly those who are younger and who have a shorter duration of illness.155,159 Although evidence-based recommendations are limited, guidelines do suggest therapies to be considered for the psychological treatment of AN: cognitive analytic therapy (CAT), CBT, IPT, focal psychodynamic therapy and family interventions focused explicitly on eating disorders.154 A study comparing CBT, IPT and nonspecific supportive clinical management in the treatment of underweight AN outpatients found that the supportive treatment produced better global outcomes than IPT and was superior to CBT, over 20 weeks, in its impact on global functioning.160 The efficacy of CBT for underweight individuals remains unclear, but it appears useful as a posthospitalization treatment for AN, contributing to improved outcomes and relapse prevention in adults after weight restoration.161 Factors consistently predicting treatment outcome have not been identified.155 A number of treatments for BN have strong empirical support. CBT and IPT have been found effective, with CBT superior at reducing behavioral symptoms.156 CBT leads to

Chapter 8  Eating Disorders faster improvement in symptoms, with better outcomes at the end of treatment, but at follow-up assessment there are no differences between CBT and IPT.162 All guidelines recommend CBT (16 to 20 sessions over four to five months) as the first-line treatment of choice for BN,77,154,156 but not all patients respond to CBT, and IPT is an effective alternative. CBT and IPT can be delivered in a group format as well as individually.163,164 Other promising treatment options with preliminary empirical support include dialectical behavior therapy (DBT, an approach developed for borderline personality disorder that focuses on assisting patients in developing skills to regulate affect165) and a manual-based guided self-change approach.166 For a subset of patients, self-help or guided self-help with an evidence-based CBT manual167 is an appropriate starting point for treatment in a steppedcare approach154 or if other treatments are not available.168 A variety of factors have been shown to be associated with treatment outcome in BN, but two emerge consistently— severity (higher frequency of binge eating) and duration of illness are associated with poorer outcomes.156 There are limited data to guide treatment decisions for the large proportion of individuals with eating disorders who are diagnosed with EDNOS. The main exception is the subgroup of those with BED. As with BN, some individuals will benefit from an evidence-based self-help program as a first step in treatment or if other treatments are not available.77,154,168 Studies have found that self-help intervention, delivered in a variety of ways (with varying levels of professional or peer support), leads to better outcomes when compared with control groups, with reductions in binge eating, binge days, and psychological features associated with BED (for a review, see Brownley and colleagues157 and Sysko and Walsh168). After consideration of self-help, American Psychiatric Association (APA)77 and National Institute for Clinical Excellence (NICE)154 guidelines recommend CBT adapted for BED as an initial treatment choice. Group CBT has been found effective for treating binge eating in overweight individuals.169,170 There is some support for individually based CBT, although methodologic limitations preclude firm conclusions. Group IPT and adapted DBT are options to consider if CBT is not a good match for the individual or is unavailable. In one study, IPT was found to lead to similar abstinence rates as CBT at one-year follow-up.170 DBT has shown promising results, with a recovery rate of 56% at six months after treatment in one randomized controlled trial (RCT).171 It is important to note that treatments for BED usually do not result in weight loss, but they may still be of benefit with regard to weight by preventing further weight gain.157 This issue of dual treatment goals—weight loss and reducing binge eating—is explored in depth later in this chapter (see “Weight Management”). Further research is needed to establish and replicate factors associated with treatment outcome. Across all diagnoses and treatments, there has been little attention to differential outcomes by socioeconomic factors. Future studies are needed to explore whether treatment efficacy differs by gender, age, race, ethnicity, socioeconomic status, or cultural group.155,156 Given the frequent psychiatric comorbidity associated with eating disorders, as well as psychosocial risk correlates, some patients with an eating disorder will benefit from psychodynamic psychotherapy and a flexible and eclectic approach depending on patient capabilities, goals, treatment history, and other psychosocial considerations.

Pharmacotherapy

Pharmacologic management has an adjunctive role for the treatment of BN and BED. Of numerous agents that have

been studied, only one, fluoxetine, has U.S. Food and Drug Administration (FDA) approval for an eating disorder (bulimia nervosa). There is insufficient empirical support for efficacy of any agent in treating the primary symptoms of AN. Similarly, there are no clinical trial data to support recommendations for the pharmacologic management of EDNOS (with the exception of trials addressing BED and NES). Finally, there are not adequate available clinical trial data to support recommendations for pharmacologic management of eating disorders in children and adolescents.172 Among a variety of agents evaluated for treatment of the primary symptoms of AN, several have been studied because of their association with weight gain; of these, none is in routine clinical use. Although some data have suggested that olanzapine may be beneficial in promoting clinical improvement in AN,173,174 a recent RCT combining olanzapine with CBT versus placebo with CBT did not demonstrate significant between-group differences in improvement in BMI.175 Given the lack of data supporting efficacy and safety in patients with AN, no pharmacologic agents currently can be generally recommended to promote weight gain in this patient population. Pharmacologic agents associated with weight gain for other indications should be used judiciously and with a candid discussion with the patient about the anticipated risks and benefits of appetite and weight changes. Finally, if such an agent is selected, symptoms should be monitored carefully to look for onset, recurrence, or increase in bingeing or purging behaviors. Other agents may have a limited role in the management of AN but do not have FDA approval for this indication. Fluoxetine has not been found to be effective for treating the primary symptoms of AN in underweight patients176 and has unclear benefit in stabilizing weight-recovered patients with AN.177,178 Sertraline (50 to 100 mg/day) was associated with significant clinical improvements in a small, open, controlled trial of patients with AN.179 Comorbid psychiatric illness is common among patients with AN and may improve with pharmacologic management, but depressive symptoms in severely underweight patients may not respond as well to antidepressant medication as in normal-weight patients. Notwithstanding the very limited role for psychotropic medication in the management of AN, patients likely will need calcium and vitamin D supplementation if dietary sources are inadequate.88 Although oral contraceptive agents may mitigate some of the symptoms of hypoestrogenemia associated with AN, they do not protect against bone loss in this population.150,180 It is useful for clinicians to bear in mind that weight restoration is the treatment of choice for underweight individuals with AN for medical stabilization, and probably also as a prerequisite to developing the psychological insight necessary for recovery. In contrast to the limitations of medication management for AN, a number of medications have established shortterm modest efficacy for the treatment of BN, although remission rates are low.181 CBT has better efficacy than medication to reduce the symptoms associated with BN, but there is some support for augmenting psychotherapy with medication, and this is fairly routine clinical practice. It is optimal to use pharmacotherapy as an adjunct to, rather than a substitute for, psychotherapy; psychotherapy, however, may not be available or beneficial to all patients. Some evidence supports treatment with fluoxetine (60 mg/ day) alone in a primary care setting.182 Fluoxetine (60 mg/ day) also has been found superior to placebo for treating bulimic symptoms in patients who have not responded adequately to CBT or IPT.183

133

134

Section II  Nutrition in Gastroenterology Of medications with established efficacy in treating BN, only fluoxetine has FDA approval for this indication. Fluoxetine (60 mg/day) generally is well tolerated in this patient population and has been shown to be effective for symptom reduction and for maintenance therapy for up to 12 months.184,185 Desipramine and imipramine (both at conventional antidepressant dosages, as tolerated) also have efficacy in symptom reduction, but are not as well tolerated in this patient population.186 Topiramate has shown efficacy in reduction of binge and purge symptoms in two short-term RCTs in individuals with BN.187-189 Other agents that have demonstrated at least some efficacy (but with less data available) are trazodone,190 ondansetron (in patients with severe BN),191 and sertraline.192 Flutamide has shown some efficacy in reducing binge (but not purge) frequency in one small RCT, but was associated with hepatotoxicity and teratogenicity and cannot be recommended for the treatment of BN.193 A number of studies has investigated the efficacy of naltrexone in treating bulimic symptoms,186 but only at higher doses was it superior to placebo and in reducing symptoms in patients who had previously not responded to alternative pharmacotherapy.194 Monitoring of liver biochemical test results is essential when this drug is used. Other medications with efficacy are relatively contraindicated for those with BN given their potential adverse effects. For example, bupropion was associated with a seizure risk of 5.8% during a clinical trial195 and there have been case reports of spontaneous hypertensive crises in patients with BN who were taking monoamine oxidase (MAO) inhibitors.196 Finally, although fluvoxamine has shown some efficacy for BN relapse prevention in one RCT,197 another RCT combining fluvoxamine with stepped-care psych­ otherapy not only did not show efficacy of this agent, but also reported grand mal seizures in participants on the active drug.198 Several trials have investigated the efficacy of phar­ macologic treatment of BED. Of the selective serotonin reuptake inhibitors (SSRIs), sertraline,199 fluvoxamine,200 citalopram,201 and fluoxetine202 have shown some efficacy in reducing symptoms associated with BED in RCTs. In addition, atomoxetine,203 orlistat in combination with CBT,204 and zonisamide205 have shown efficacy in an RCT, although zonisamide was not well tolerated in this BED study population and significantly greater binge remission rates in the orlistat group were not maintained at the three-month post-treatment follow-up.206 Two agents, sibutramine207 and topiramate,208 have shown efficacy in reducing symptoms of BED (the latter in BED co-morbid with obesity) in mul­ tisite placebo-controlled trials. None of these medications has FDA approval for the treatment of BED. Notwithstanding some efficacy of medication, studies have suggested that CBT is a superior treatment and that augmentation of CBT with medication may not enhance treatment response.209

WEIGHT MANAGEMENT

Active weight management is a cornerstone of treatment for AN. As essential as weight gain is to reduce or reverse the medical and cognitive sequelae of severe undernutrition, it is one of the great challenges in the successful treatment of this illness. By definition, individuals with AN are unreasonably fearful of gaining weight and many of them remain unconvinced of the serious medical impact of their selfstarvation. Ideally, patients can be engaged in the process of weight recovery by identifying some clear benefit (e.g., permission to remain on an athletic team or participate in a performance, or to avoid a compulsory medical leave from school or work). Such behavioral reinforcement can be an essential adjunct to a nutritional plan that provides bal-

anced nutrition and calories adequate for weight gain and reestablishes routine meals. Outpatient weight recovery is best addressed with the collaboration of a nutritionist experienced in the treatment of AN. As calories are added and foods are reintroduced into the diet, patients may initiate or increase compensatory behaviors (e.g., exercise, purging) to control weight gain. If possible, behavioral restrictions on exercise can be implemented if patients are not meeting weight gain goals. Caloric supplements often are added as snacks to help patients meet nutritional and weight gain goals. Patients with early satiety and delayed gastric emptying have a particularly difficult time adding calories because gastrointestinal discomfort and bloating enhance their concerns about feeling and being “fat.”210,211 If supportive psychotherapy, nutritional guidance, behavioral reinforcements and limits, nutritional supplements, and restricted exercise do not result in adequate weight gain, a higher intensity of care may be indicated. Supervised meals, partial hospitalization (e.g., a structured day treatment program, often including a 12-hour day of various treatment modalities, during which the patient returns to his or her own home in the evenings), or even increasing the frequency of outpatient therapy appointments, may be sufficient to promote weight gain. However, if bingeing and purging symptoms are emerging or increasing or if the patient is losing weight, inpatient care may be required for weight restoration. Even in this setting, behavioral methods to promote weight gain are preferred to nasogastric feeding or total parenteral nutrition. The latter options are avoided if possible, but in the setting of severe malnutrition they may be necessary. Severely malnourished patients—especially those below 70% to 75% of expected body weight—require inpatient care for refeeding. Patients with AN are at particularly high risk of refeeding syndrome, which can occur with any means of refeeding (see Chapters 4 and 5).212 Refeeding syndrome, typically associated with hypophosphatemia in the setting of depletion and cellular shifts in the early weeks following refeeding, can result in delirium, congestive heart failure, and death.213 Risk for refeeding syndrome can be reduced for at-risk patients by using an initially low-calorie prescription that is advanced slowly. During at least the first two weeks of refeeding, serum electrolyte, phosphorus, and magnesium levels should be monitored closely (e.g., six to eight hours after feeding begins, then daily for a week, then at least every other day until the patient is stabilized214). Heart rate, respiratory rate, lower extremity edema, and signs of congestive heart failure also should be evaluated daily for at least a week and then gradually at longer intervals as the patient stabilizes, and cardiac telemetry should be used to monitor heart rhythm during the first two weeks so that supplementation and other appropriate measures can be instituted if hypophosphatemia or other signs of refeeding syndrome develop. Delirium may occur in the second week of refeeding, or later, and may last for several weeks.215-218 Some experimental data have suggested that a healthful dieting intervention may be beneficial in reduction of bulimic symptoms219 but, conventionally, weight loss treatment has been discouraged in patients with BN because dieting can stimulate bingeing and purging. Weight loss is often a primary or secondary treatment goal for individuals with BED because of comorbid obesity. Models of binge eating have proposed that dietary restriction is an antece­ dent to binge eating; thus, there has been debate about the optimal means and order of addressing co-occurring binge eating and obesity. Most data, however, have shown that a variety of weight loss approaches do not exacerbate binge

Chapter 8  Eating Disorders eating and may help reduce symptoms; one prospective study found no evidence that a reduced calorie diet precipitated binge eating in women with obesity.220 Behavioral weight loss treatment (BWLT)221 and very low-calorie diets (VLCDs)222,223 have been found effective for reducing symptoms of BED. Available evidence also supports that CBT and BWLT are equally effective in terms of binge eating outcomes, although the rates of change differ for binge eating and weight loss. Binge eating decreases faster with CBT, whereas weight decreases more rapidly with BWLT, so treatment priorities may inform recommendations. The addition of exercise to treatment for BED is associated with greater decreases in binge eating and BMI.224 Although a number of studies have found that treating binge eating does not translate to weight loss, some studies have found that reductions in binge eating can assist in modest weight loss among those with BED, especially when complete remission is achieved.225 BED is common in individuals presenting for obesity surgery, so there is much interest in clarifying how BED affects the outcome of bariatric surgery and, conversely, how surgery might influence binge eating behavior. The prevalence of BED in preoperative gastric bypass patients has been found to range from 2% to 49%, and up to 64% of bariatric surgery candidates have binge eating behaviors.226 The most current studies examining presurgery binge eating, postsurgery binge eating, and long-term weight outcomes suggest that a presurgery history of binge eating does impart risk for poorer long-term weight outcome.226,227 Postsurgical binge eating usually is seen in patients who reported binge eating prior to surgery and is rare among those without presurgery binge eating.226 Patients who continue to experience binge eating after surgery have poorer weight outcomes. One study has shown that a history of binge eating before Roux-en-Y gastric bypass is associated with significantly less weight loss at one- and two-year follow-ups.227 Many patients with binge eating will have positive outcomes after bariatric surgery, but adjunctive treatment is likely to be important for optimizing outcomes and preventing relapse. There are few data on bulimia or self-induced vomiting and bariatric surgery. Cases of AN developing after bariatric surgery have been reported.228,229

MEDICAL MANAGEMENT OF GASTROINTESTINAL SYMPTOMS OF PATIENTS WITH EATING DISORDERS

Individuals with eating disorders are likely to have cooccurring GI symptoms for which consultation may be sought. GI complaints are the most common somatic complaint among adolescents with partial eating disorders.230 Similarly, childhood GI complaints may influence later risk or timing or onset and severity for an eating disorder. In some cases, behaviors associated with eating disorders result in serious GI complications. In other cases, GI symptoms may be mild and not correlate with underlying pathology, but may compromise efforts to nutritionally rehabilitate the patient. Given the evidence that restrictive eating, binge pattern eating, and purging behaviors may underlie or exacerbate some of the GI symptoms, concurrent management of the eating disorder is integral to prevent worsening of the GI manifestations of illness. Careful differential diagnosis also is necessary to avoid misattribution of symptoms to an eating disorder and to detect primary GI pathology that may be obscured by an eating disorder. Available data suggest that individuals with an eating disorder are significantly more likely to seek GI specialty care than healthy controls.53

Moreover, presentation to a GI practice rather than to an eating disorder specialty practice results in delayed diagnosis and a greater number of clinical tests than controls with slow transit constipation.56 Because the GI consul­tation may precede help-seeking related to the primary symptoms of the eating disorder, the patient’s care will benefit from identification of an associated eating disorder, evaluation of its severity, and appropriate counsel about the necessity of team management and referrals to mental health, nutritional, and primary care clinicians. A case series of individuals with coexisting comorbid eating dis­orders and celiac disease has illustrated how synchronous GI and eating disorders reciprocally influence management. For example, celiac disease can mimic, exacerbate, or promote recovery from an eating disorder, whereas an eating disorder can reduce adherence to treatment for celiac disease.57 Subjective reports of Gl symptoms may not reliably indicate pathology122,231; moreover, they may be mediated by affect103 or body image concerns.93 Thus, when patients complain of bloating and constipation, it is useful to determine to what extent these complaints stem from fear of gaining weight or reflect decreased GI motility. A number of studies have evaluated improvement in GI function after nutritional rehabilitation (Table 8-5). These studies have yielded mixed results, and conclusions have been limited by small sample sizes and nonrandomized design. For example, in one study,122 gastric emptying improved in patients with restricting type AN, but did not improve in patients with binge eating–purging type AN after a 22-week treatment period of increasing dietary intake up to 4000 cal/day and CBT. Self-reported GI symptom scores improved after treatment in this same study, but remained abnormal and did not correlate with gastric emptying as evaluated with ultrasound.122 Another study of a mixed sample of adolescents and adults with AN did not demonstrate significant improvement of gastric emptying after weight gain (N = 6), despite normalization of heart rate and blood pressure.232 Other studies have suggested that nutritional rehabilitation is associated with improved gastric emptying in inpatients with AN, but it is unclear whether such improvement is related to refeeding per se, or to weight gain.97,122,106,233 Constipation is a frequent complaint of patients with AN and BN and may have multiple causes. Colonic transit appears to be delayed in patients with constipation and AN, but colonic transit has been shown to return to normal within three to four weeks of refeeding in hospitalized patients with AN.107,121 In one study, however, anorectal dysfunction in anorexic patients with severe constipation did not significantly improve with refeeding. The investigators suggested that abnormal defecatory perception thresholds and expulsion dynamics in AN may have contributed to the patients’ unremitting constipation.121 Laxative abuse occurs in AN234 and BN.235 Some patients use laxatives as their chief method of purging and may gradually escalate their daily dose to very large amounts. Although the relationship of laxative abuse to colonic dysfunction remains controversial,236-238 it has been observed that patients with chronic laxative abuse complain of consti­ pation while tapering off their laxatives. Rectal prolapse has been described with AN and BN and is thought to be linked to constipation, laxative use, excessive exercise, and increased intra-abdominal pressure secondary to selfinduced vomiting.116,117 Delayed intestinal transit and its associated clinical symptoms present a particularly interesting clinical challenge in patients with eating disorders. Studies to date, however, have been small, short term, and not randomized, so only limited conclusions regarding management can be

135

136

Section II  Nutrition in Gastroenterology Table 8-5  Selected Studies of Effects of Nutritional Rehabilitation on Gastrointestinal Symptoms Associated with Eating Disorders in Adults condition and Method of Assessment

Study Population

Results

109

Gastrointestinal symptoms using a survey (GISS)

16 with AN and 12 healthy volunteers

110

Response of GI symptoms in BN to nutritional and psychotherapeutic inpatient treatment using GISS Eating Disorders Inventory; Zung Depression Inventory

43 inpatients with severe bulimia; 32 healthy volunteers as untreated comparison group

111

GI symptoms and gastric emptying using a double isotope technique to measure gastric emptying and self-reporting of GI symptoms

14 adult inpatients with AN (13 women, 1 man) and 14 normal male controls

120

Gastric emptying in BN and AN using gamma cameraassessed gastric emptying

121

Colonic transit and motility using radiopaque markers and anorectal manometry

22 patients with AN (12 patients on self-selected diet; 10 patients on refeeding diet); 10 with BN (untreated comparison group); 10 controls (second untreated comparison group) 13 adult inpatients with AN and chronic constipation and 20 age-matched healthy female controls

135

Constipation using radiopaque markers and anorectal manometry

12 adult inpatient women with AN and constipation and 12 healthy female controls

136

Gastric dysmotility using ultrasonographic gastric emptying test, bowel symptom questionnaire, and psychological assessments

23 inpatients with AN (12 with binge/purge AN and 11 with restrictive AN) and 24 age-matched healthy controls

80% of patients reported one or more serious GI complaints; significant improvement noted in most symptoms after nutritional rehabilitation, but patients remained more symptomatic on GISS than comparison group 95% of patients had two or more GI complaints; patients had more GI symptoms (nausea, dysphagia, heartburn, borborygmi, belching, bloating, flatulence, abdominal pain, constipation, diarrhea) on admission compared with controls; patients remained more symptomatic than comparison group, even after nutritional rehabilitation, despite some improvement Upper GI symptoms were present in 78% of AN patients; gastric emptying was significantly slower in AN patients than in comparison group pretreatment; gastric emptying of liquids and solids was significantly faster after treatment completion (n = 11) (returned to normal in six, remained slow in two, and faster than normal in three); no change in GI symptoms and gastric emptying for three patients who did not gain weight Gastric emptying time of solid meal in AN subjects on self-selecting diet was significantly longer than that of controls; gastric emptying times of AN subjects on refeeding diet and BN subjects not significantly different from controls; weight gain among refed AN subjects had no significant effect on gastric emptying of solid meal Colonic transit significantly slower for patients within three wk of admission compared with those with >three wk of treatment; no significant differences between patients in hospital >three wk and controls; two of four patients with slow colonic transit who were restudied after six weeks in hospital achieved normal transit time; no significant differences between AN patients and controls with respect to anorectal manometry 8 of 12 AN patients had slow colonic transit times that normalized after four wk of refeeding; 5 of 12 patients had anorectal dysfunction that did not normalize with refeeding Significantly delayed gastric emptying and higher self-reported symptom scores in patients compared with controls pretreatment; no correlation found between gastric symptoms and emptying or psychological tests; self-reported GI symptoms improved significantly following treatment but remained mostly in pathologic range; significantly improved gastric emptying found in treatment completers with restricting but not binge/purge AN; gastric emptying time in treatment completers still longer than controls at baseline

reference

AN, anorexia nervosa; BN, bulimia nervosa; GI, gastrointestinal.

Chapter 8  Eating Disorders drawn. Existing data suggest that reestablishing regular food intake or weight gain will improve delayed gastric emptying and slowed colonic transit, although this may not be sufficient to restore normal GI function. Patients may resist active weight management or cessation of their disordered pattern of eating, despite their having a serious eating disorder and associated GI complications. This resistance may be exacerbated by early satiety, abdominal pain, bloating, or constipation, all of which may reinforce the patient’s excessive concern with weight or conviction that his or her diet needs to be further restricted. Management of symptoms is complicated further because subjective symptom reports correlate imperfectly with pathology and some of the complaints may be mediated by psychiatric symptoms or illness, including depression, anxiety, or distorted body image. Because refeeding and establishing normal and healthful dietary patterns are both treatment goals and likely to improve symptoms, careful nutritional rehabilitation is a reasonable and conservative initial step in management of suspected delayed gastric emptying and slow colonic transit for inpatients with AN or BN. Patients are likely to benefit from the support and reassurance that many of the GI symptoms commonly associated with eating disorders (e.g., bloating, constipation, nausea, vomiting, and diarrhea) will improve as eating and weight return to normal. Additional management strategies include dietary changes to reduce bloating (e.g., promoting smaller, more frequent meals, encouraging consumption of liquids earlier in the meal, and possibly providing a percentage of calories, no more than 25% to 50%, in liquid form initially).58,95 Various prokinetic agents have been used to manage delayed gastric emptying in AN, although metoclopramide is difficult to tolerate in frequent or high dosage, domperidone is not available in the United States except in compounding pharmacies, and cisapride is no longer being manufactured. Moreover, existing data do not support a recommendation for their use for gastric motility complaints in AN.211 Some clinicians have reservations about treating the constipation that follows laxative abuse with laxatives. Although it does not make sense to reproduce purging behavior using cathartics to treat constipation in patients with chronic laxative abuse, some patients will benefit from a thoughtful bowel regimen to reduce discomfort and bloating that otherwise might induce relapse of laxative abuse. Increasing fluid intake, dietary fiber, and possibly the addition of stool softeners and bulk-forming laxatives would be reasonable and conservative initial treatment. Osmotic laxatives initially may be necessary for symptom relief in some cases.239 Management of constipation may require anorectal retraining if a result of anorectal dysfunction.121 Some patients may benefit from symptomatic relief of gastroesophageal reflux or esophagitis with antacids or H2 antagonists; proton pump inhibitors may be required for relief of more severe symptoms.58 Although this may be appropriate clinically, the underlying cause and exacerbation of the GI complaint should be made clear to the patient and also actively addressed in psychotherapeutic treatment when related to the eating disorder. Mild elevation of serum aminotransferase levels secondary to malnutrition in AN likely will likely remit with weight restoration. Elevated serum levels of liver enzymes in severely ill patients may be an indication of refeeding syndrome or reflect AN-related hypoperfusion, and require emergent evaluation and intervention.58,128 Although many GI symptoms may be related to restrictive eating, binge pattern eating, or purging, some GI complaints

will require diagnostic evaluation. Anecdotal reports of catastrophic GI complications of the eating disorders, as well as primary GI illness that arises coincidentally with an eating disorder or mimics an eating disorder, suggest that complaints should be evaluated in their specific clinical context. Acute gastric dilation may be unsuspected in the absence of clinical history of binge eating.130 If acute gastric dilation is confirmed in the setting of refeeding or in the presence of a history of an eating disorder with binge eating, nasogastric decompression and fluid resuscitation are necessary. If these are not effective, laparotomy may be necessary.113,114,130 In addition, symptoms that persist after nutritional rehabilitation may require additional diagnostic evaluation. Eating disorders commonly are associated with GI symptoms and severe eating disorders can be associated with serious GI complications. Patients with eating disorders commonly present to primary and specialty care settings with GI symptoms or illness. In such cases, patients should be engaged in the concept of team management of their eating disorder and associated medical and nutritional complications at the same time as their GI complaint is addressed. More clinical trial data are needed to clarify treatment strategies for GI complaints associated with eating disorders. However, management of GI symptoms in patients with an eating disorder can be guided by several key considerations. Primary GI illness should be excluded, and the possibility considered that an eating disorder is obscuring or mimicking a primary illness. If GI symptoms appear to be associated with the eating disorder, nutritional rehabilitation in combination with psychotherapeutic care should be considered as an initial step. Nutritional rehabilitation often will require inpatient-level care and monitoring for serious potential complications, such as refeeding syndrome and acute gastric dilation. During treatment of an eating disorder, resistance to weight gain, to eating normally, and to cessation of bingeing and purging is common, so the possibility that body image or emotional symptoms mediate GI complaints should be considered in the treatment plan.

KEY REFERENCES

American Psychiatric Association. Treatment of patients with eating disorders. 3rd edition. American Psychiatric Association. Am J Psychiatry 2006; 163(Suppl):4-54. (Ref 77.) American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed, text revision. Arlington, Va: American Psychiatric Association; 2000. (Ref 8.) Brownley KA, Berkman ND, Sedway JA, et al. Binge eating disorder treatment: A systematic review of randomized controlled trials. Int J Eat Disord 2007; 40:337-48. (Ref 157.) Bulik CM, Berkman ND, Brownley KA, et al. Anorexia nervosa: A systematic review of randomized controlled trials. Int J Eat Disord 2007; 40:310-20. (Ref 155.) Emmanuel AV, Stern J, Treasure J, et al. Anorexia nervosa in a gastrointestinal practice. Eur J Gastroenterol Hepatol 2004; 16:1135-42. (Ref 56.) Hadley SJ, Walsh BT. Gastrointestinal disturbances in anorexia nervosa and bulimia nervosa. Curr Drug Targets CNS Neurol Disord 2003; 2:1-9. (Ref 211.) Harris EC, Barraclough B. Excess mortality of mental disorder. Br J Psychiatry 1998; 173:11-53. (Ref 1.) Hudson JI, Hiripi E, Pope HG Jr, Kessler RC. The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biol Psychiatry 2007; 61:348-58. (Ref 3.) Keel PK. Purging disorder: Subthreshold variant or full-threshold eating disorder? Int J Eat Disord 2007; 40(Suppl):S89-94. (Ref 69.) National Institute for Clinical Excellence (NICE). Eating disorders—core interventions in the treatment and management of anorexia nervosa, bulimia nervosa and related eating disorders. NICE Clinical Guideline No 9. London: NICE; 2004. (Ref 154.)

137

138

Section II  Nutrition in Gastroenterology Niego SH, Kofman MD, Weiss JJ, Geliebter A. Binge eating in the bariatric surgery population: A review of the literature. Int J Eat Disord 2007; 40:349-59. (Ref 226.) Shapiro JR, Berkman ND, Brownley KA, et al. Bulimia nervosa treatment: A systematic review of randomized controlled trials. Int J Eat Disord 2007; 40:321-36. (Ref 156.) Striegel-Moore RH, Franko DL, May A, et al. Should night eating syndrome be included in the DSM? Int J Eat Disord 2006; 39:544-9. (Ref 65.)

Winstead NS, Willard SG. Gastrointestinal complaints in patients with eating disorders. J Clin Gastroenterol 2006; 40:678-82. (Ref 53.) World Health Organization. The ICD-10 classification of mental and behavioural disorders: Clinical descriptions and diagnostic guidelines. Geneva: WHO; 1992. (Ref 74.) Zerbe K. Integrated treatment of eating disorders. New York: WW Norton; 2008. (Ref 158.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

9

Food Allergies Hugh A. Sampson

CHAPTER OUTLINE Background, Definitions, and Prevalence  139 Definitions  139 Prevalence  139 Pathogenesis  139 Clinical Features  142 IgE-Mediated Disorders  143

BACKGROUND, DEFINITIONS, AND PREVALENCE The first recorded account of food allergy was provided by Hippocrates, but it was not until 1921 that the classic experiment of Prausnitz initiated investigation on a scientific level and established the immunologic basis of allergic reactions.1 In this experiment, Prausnitz injected serum from his patient, Kustner, who was allergic to fish, into his own skin; the next day he injected fish extract into the same areas and into control sites. Local reactions proved sensitivity could be transferred by a factor in serum from an allergic to a nonallergic person. In 1950, Loveless demonstrated the inaccuracy of diagnosing food allergy by history in her report of the first blinded, placebo-controlled food trials in patients with milk allergy.2 In the following three decades, standardized protocols for the evaluation of food allergy were developed, and the double-blind, placebo-controlled oral food challenge (DBPCFC) emerged as the accepted standard for the diagnosis of food allergy.3

DEFINITIONS

Terminology used by investigators in the field of food allergy differs slightly in different parts of the world. The following represents current terminology in the United States. An adverse food reaction is a generic term indicating any untoward reaction occurring after the ingestion of a food or food additive and may be the result of toxic or nontoxic reactions. Toxic reactions will occur in any exposed individual upon ingestion of a sufficient dose. Nontoxic reactions depend on individual susceptibilities and may be immune-mediated (food allergy or food hypersensitivity) or non–immune-mediated (food intolerance). Food intolerances comprise most adverse food reactions and are cate­ gorized as enzymatic, pharmacologic, or idiopathic food intolerances. Secondary lactase deficiency, an enzymatic intolerance, affects the vast majority of adults, whereas most other enzyme deficiencies are rare inborn errors of metabolism and thus primarily affect infants and children. Pharmacologic food intolerances are present in individuals who are abnormally reactive to substances such as vasoactive amines, which are normally present in some foods (e.g., tyramine in aged cheeses). Confirmed adverse food reactions for which the mechanism is not known are generally

Mixed IgE- and Non–IgE-Mediated Disorders  143 Non–IgE-Mediated Disorders  145 Diagnosis  146 Treatment and Natural History  148

classified as idiopathic intolerances. Food allergies usually are characterized as IgE-mediated or non–IgE-mediated; the latter are presumed to be cell-mediated.

PREVALENCE

About 6% of young children and 3.5% of adults in the United States have food allergies.4 The prevalence of food allergies is greatest in the first few years of life and decreases over the first decade. The most common food allergens in young children include milk (2.5%), egg (1.5%), peanut (0.8%), wheat (∼0.4%), and soy (∼0.4%). Other than peanut, most childhood food allergies are outgrown by the end of the first decade. Almost all infants who develop cow’s milk allergy do so in the first year of life, with about two thirds experiencing IgE-mediated reactions and 35% going on to develop other food allergies.5 Peanut, tree nut, sesame, and seafood allergies tend to be lifelong, but about 20% of young children with peanut allergy develop clinical tolerance.6 Food allergies may persist after childhood into adulthood or develop in adulthood, with the most common food allergies in adults consisting of shellfish (2%), peanut (0.6%), tree nuts (0.4%), and fish (0.4%).7 About 5% of the U.S. population experiences allergic reactions to raw fruits and vegetables. Most of these reactions are caused by crossreactivity between homologous proteins in pollens, such as ragweed, and certain fruits and vegetables, such as melons and bananas (oral allergy syndrome), in adolescents and adults who have seasonal allergic rhinitis. The prevalence of food allergies appears to be increasing.8 Studies from the United States and United Kingdom have indicated that the prevalence of peanut allergy has doubled in young children during the past decade.9,10 In addition, children with atopic disorders have a higher prevalence of food allergies; for example, 35% to 40% of children with moderate to severe atopic dermatitis have IgE-mediated food allergy.11

PATHOGENESIS Gut-associated lymphoid tissue (GALT), a component of the mucosal immune system, lies juxtaposed to the external environment and acts to differentiate organisms and foreign proteins that are potentially harmful from those that are not. Unlike the systemic immune system, which recognizes

139

140

Section II  Nutrition in Gastroenterology Table 9-1  Physiologic and Immunologic Barriers of the Gastrointestinal Tract Physiologic Barriers Block penetration of ingested antigens Epithelial cells—one cell layer of columnar epithelium Glycocalyx—coating of complex glycoprotein and mucins that traps particles Intestinal microvillus membrane structure—prevents penetration Tight junctions joining adjacent enterocytes—prevent penetration even of small peptides Intestinal peristalsis—flushes trapped particles out in the stool Break down ingested antigens Salivary amylases and mastication Gastric acid and pepsins Pancreatic enzymes Intestinal enzymes Intestinal epithelial cell lysozyme activity Immunologic Barriers Block penetration of ingested antigens Antigen-specific SigA in intestinal lumen Clear antigens penetrating the gastrointestinal barrier Serum antigen-specific IgA and IgG Reticuloendothelial system SigA, secretory immunoglobulin A; IgG, immunoglobulin G.

relatively small quantities of antigen and mounts a brisk inflammatory response to neutralize them, the mucosal immune system regularly encounters enormous quantities of antigen and generally functions to suppress immune reactivity to harmless foreign antigens (e.g., food proteins, commensal organisms), only mounting a brisk protective response to dangerous pathogens when appropriate. A single-cell layer of columnar intestinal epithelial cells (IECs) separates the external environment from the loosely organized lymphoid tissue of the lamina propria. A highly efficient gastrointestinal mucosal barrier has evolved, however, that provides an enormous surface area for proc­ essing and absorbing ingested food and discharging waste products.12 This barrier uses physiologic and immunologic barriers to prevent the penetration of foreign antigens (Table 9-1). The physiologic barrier is composed of the following: epithelial cells, joined by tight junctions and covered with a thick mucus layer that traps particles, bacteria, and viruses; trefoil factors (TFFs, 7 to 12 kd), protease-resistant proteins secreted by mucus-secreting cells of the stomach (TFF1, TFF2) and intestine (TFF3) that help strengthen and promote restoration of the barrier; and luminal and brush border enzymes, bile salts, and extremes of pH, all serving to destroy pathogens and render antigens nonimmunogenic. Innate (natural killer [NK] cells, polymorphonuclear leukocytes, macrophages, epithelial cells, and Toll-like receptors) and adaptive immune (intraepithelial and lamina propria lymphocytes, Peyer’s patches, secretory immunoglobulin A [S-IgA] and cytokines) responses provide an active barrier to foreign antigens. Developmental imma­ turity of various components of the intestinal barrier and immune system reduces the efficiency of the infant mucosal barrier. For example, the activity of various enzymes is suboptimal in the newborn period and the S-IgA system is not fully mature until four years of age. This immature state of the mucosal barrier may play a role in the increased prevalence of gastrointestinal infections and food allergies seen in the first few years of life. In addition, studies have shown that alteration of the physiologic barrier function, such as gastric acidity, can lead to increased IgE sensitization in children and adults.13

Despite the evolution of this complex mucosal barrier, about 2% of ingested food antigens are absorbed and transported throughout the body in an immunologically intact form, even through the normal mature intestine.12 In an elegant series of experiments performed more than 75 years ago, Walzer and colleagues used sera from food-allergic patients to sensitize volunteers passively and demonstrated that immunologically intact antigens cross the mucosal barrier and disseminate rapidly throughout the body.14-16 Increased gastric acidity and the presence of food in the intestine decrease antigen absorption, whereas hypochlorhydria (e.g., H2 blocker– and proton pump inhibitor– induced) and ingestion of alcohol increase antigen absorption.15 The immunologically intact proteins that elude the intestinal barrier usually do not provoke adverse reaction, because most individuals have developed tolerance, but in a sensitized individual, allergic reactions will occur. Although more common in the developing GALT of young children, it is clear that cellular and IgE-mediated allergic responses to foods can develop at any age. The dominant response in GALT is suppression, or tolerance. The means whereby the immune system is educated to avoid sensitization to ingested food antigens is not well understood, but studies have suggested that antigenpresenting cells, especially IECs and various dendritic cells, and regulatory T cells play a central role.12 High-dose tolerance is caused by lymphocyte anergy, resulting from antigen T cell receptor ligation in the absence of costimulatory signals, whereas low-dose tolerance is mediated by regulatory T cells. Five different regulatory T cells have been identified in conjunction with intestinal immunity: (1) Th3 cells, a population of CD4+ cells that secrete transforming growth factor-β (TGF-β); (2) Th1 cells, CD4+ cells that secrete interleukin-10 (IL-10); (3) CD4+,CD25+ regulatory T cells; (4) CD8+ suppressor T cells; and (5) γδ T cells.17 IECs have been shown to be nonprofessional antigen-presenting cells (APCs) that can process luminal antigen and present it to CD4+ T cells on class II major histocompatibility complex (MHC) molecules. IECs, however, lack a second signal necessary to activate T cells, thus suggesting there is another mechanism whereby these cells can induce tolerance to food antigens. Extracellular proteins that are internalized by professional APCs (e.g., monocytes, macrophages, dendritic cells) into vesicles are processed and displayed by class II MHC molecules to CD4+ T cells, whereas proteins in the cytosol of nucleated cells are processed and displayed by class I MHC molecules to CD8+ T cells. IECs also can present lipid and glycolipid antigens to CD8+ suppressor T cells by a nonclassic—that is, non-MHC, class I molecule (CD1d)—and other novel membrane molecules that interact with CD8+ T cells (see Fig. 9-1). In addition, dendritic cells residing within the lamina propria and noninflammatory environment of Peyer’s patches express IL-10 and IL-4, which favor the generation of tolerance. It has been suggested that T cells primed in the local mucosal environment induce tolerance, whereas T cells primed in the mesenteric lymph nodes, either from antigen reaching the nodes in lymph or carried there by circulating dendritic cells, differentiate and travel to the mucosa, where they induce local immune responses.18 Recently, the unique role of the oral mucosa and its relation to tolerance induction via Langerhans cells has been increasingly appreciated.19 It is likely that the commensal bowel flora also play a role in shaping the mucosal immune response. It is estimated that there are 1012 to 1014 bacteria/g of colonic tissue, which means that there are more bacteria in the colon than cells in the body.12 Bowel flora is largely established in the first 24 hours after birth, is dependent on maternal flora, genet-

Chapter 9  Food Allergies IgE-associated

Non–IgE-associated Soluble proteins

Particulate proteins

Glycocalyx

M cell

IEC IEC IEC IEC IEL

IgE receptor

Mast cell

T Mφ Peyer's patch

LPL

B

Lamina propria

Th3 cell

Mφ

Histamine B cell

IgE

B

Th cell

IgG

? TNF-α TNF-α IL-5 IL-4

ics, and local environment, and is relatively stable throughout life. The importance of bowel flora in the development of oral tolerance induction is suggested by the fact that mice raised in a germ-free environment from birth fail to develop normal tolerance.20 Studies in which lactating mothers and their offspring were fed Lactobacillus suggest that probiotics may be beneficial in preventing some atopic disorders, such as eczema,21 but results from other studies are not consistent. IECs also may play a central regulatory role in determining the rate and pattern of uptake of ingested antigens. Studies in sensitized rats have indicated that intestinal antigen transport proceeds in two phases.22 In the first phase, transepithelial transport occurs via endosomes, is antigen-specific and mast cell–independent, and occurs 10 times faster in sensitized rats compared with nonsen­ sitized control animals. Antigen-specific IgE antibodies bound to the mucosal surface of IECs via FcεRII are responsible for this accelerated allergen entry.23 In the second phase, paracellular transport predominates. Loosening of the tight junctions occurs as a result of factors released by mast cells activated in the first phase. Whereas the first antigen-specific pathway involves antibody, the second nonspecific pathway most likely involves cytokines. Consistent with this concept, IECs express receptors for a number of cytokines (IL-1, IL-2, IL-6, IL-10, IL-12, IL-15, granulocyte-monocyte colony-stimulating factor [GM-CSF], and interferon-γ [IFN-γ]), and have been shown to be functionally altered by exposure to these cytokines. Although the development and mechanistic features of non–IgE-mediated food-allergic responses are poorly understood, the development of IgE-mediated responses has been well characterized. Sensitivity to allergens (generally glycoproteins) is the result of a series of molecular and cellular interactions involving APCs, T cells, and B cells.24 APCs present small peptide fragments (T cell epitopes) in conjunction with MHC class II molecules to T cells. T cells bearing the appropriate complementary T cell receptor (TCR) will bind to the peptide-MHC complex. This interactive first signal leads to T cell proliferation and cytokine generation and the generation of a second signal (e.g., IL-4) that promotes an IgE response (Th2-like cell activation).

Figure 9-1.  Immunopathogenesis of food allergies. Massive quan­ tities of food proteins are processed in the intestinal tract to non­ immunogenic peptides and amino acids. As described in the text, however, a small amount of immunogenic protein passes through the intestinal barrier. Intestinal epithelial cells (IECs) normally process soluble proteins for presentation to appropriate T-helper (Th1 or Th2) and regulatory (Th3) T-cells. Protective IgA and IgG antibody responses are generated, and systemic T-cell responses are down-regulated. In IgE-associated disorders, food-specific IgEproducing B cells are activated. IgE antibodies adhere to the surface of mast cells, and release histamine and other mediators if surfacebound IgE encounters the food antigen. IgE also binds to FcεR on intestinal IECs, thereby expediting antigen transfer through IECs. In non–IgE-mediated disorders, antigen-presenting cells and/or T cells are activated to secrete TNF-α (dietary protein–induced enterocolitis syndrome) or IL-4 and/or IL-5 (allergic eosinophilic gastroenteritis). M cells overlying Peyer’s patches are believed to play a major role in processing particulate protein and pathogens. FcεR, Fc epsilon receptor; Ig, immunoglobulin; IL-4, -5, interleukin-4, -5; LPL, lamina propria lymphocyte; Mf, macrophage; TNF-α, tumor necrosis factor-α.

These cells and their products, in turn, interact with B cells bearing appropriate antigen-specific receptors, leading to isotype switching and the generation of antigen-specific IgE. At all stages, a number of specific cytokines are secreted that modulate the cell interactions. The antigen-specific IgE then binds to surface receptors of mast cells, basophils, macrophages, and other APCs, arming the immune system for an allergic reaction upon the next encounter with the specific antigen. A breakdown in mucosal integrity, caused by infection or other inflammatory processes, leads to increased intestinal permeability, which results in antigens bypassing the normal tolerogenic presentation by IECs and, under some circumstances, leads to allergic sensitization. Oral tolerance of humoral and cellular immunity has been demonstrated in rodents and humans. Feeding of keyhole limpet hemocyanin to human volunteers resulted in T cell tolerance but priming of B cells at both mucosal and systemic sites.25 The failure of human infants to develop oral tolerance or the breakdown of oral tolerance in older individuals results in the development of food allergy. Young infants are more prone to develop food-allergic reactions because of the immaturity of their immunologic system and, to some extent, the gastrointestinal (GI) tract (see Table 9-1). Exclusive breast-feeding promotes the development of oral tolerance and may prevent some food allergy and atopic dermatitis.26,27 The protective effect of breast milk appears to be the result of several factors, including decreased content of foreign proteins, the presence of S-IgA (which provides passive protection against foreign protein and pathogens), and the presence of soluble factors (e.g., prolactin), which may induce earlier maturation of the intestinal barrier and the infant’s immune response. The antibacterial activity of breast milk is well established, but the ability of breast milk S-IgA to prevent food antigen penetration is less clear. Low concentrations of food-specific IgG, IgM, and IgA antibodies commonly are found in the serum of normal persons. Food protein–specific IgG antibodies tend to rise in the first months following the introduction of a food and then generally decline, even though the food protein continues to be ingested. Persons with various inflammatory bowel disorders (e.g., celiac disease, food allergy) frequently have high levels of food-specific IgG and IgM antibodies,

141

142

Section II  Nutrition in Gastroenterology although there is no evidence that these antibodies are pathogenic. Increased lymphocyte proliferation or IL-2 production following food antigen stimulation in vitro is frequently seen in patients with inflammatory bowel disorders, but it also occurs in normal persons.28 Antigen-specific T-cell proliferation in vitro alone does not represent a marker of immunopathogenicity, but simply reflects response to antigen exposure. In genetically predisposed individuals, as noted, antigen presentation leads to excessive Th2 responsiveness (lymphocytes secreting IL-4, IL-5, IL-10, and IL-13), resulting in increased IgE production and expression of Fc epsilon (FcεI) receptors on a variety of cells.24 These IgE antibodies bind high-affinity FcεI receptors on mast cells, basophils, and dendritic cells, as well as low-affinity FcεII (CD23) receptors on macrophages, monocytes, lymphocytes, eosinophils, and platelets. When food allergens penetrate mucosal barriers and reach IgE antibodies bound to mast cells or basophils, the cells are activated and mediators (e.g., histamine, prostaglandins, and leukotrienes) are released that induce vasodilation, smooth muscle contraction, and mucus secretion, and lead to symptoms of immediate hypersensitivity. These activated mast cells also may release a variety of cytokines (e.g., IL-4, IL-5, IL-6, tumor necrosis factor-α [TNF-α], platelet-activating factor), which may induce the IgEmediated late-phase inflammation. Various symptoms have been associated with IgE-mediated allergic reactions— generalized: shock; cutaneous: urticaria, angioedema, and a pruritic morbilliform rash; oral and GI: lip, tongue, and palatal pruritus and swelling, vomiting, and diarrhea; and upper and lower respiratory: ocular pruritus and tearing associated with nasal congestion, laryngeal edema, and wheezing. A rise in the plasma histamine level has been associated with the development of these symptoms after blinded food challenges.29 In IgE-mediated GI reactions, endoscopic observation has revealed local vasodilation, edema, mucus secretion, and petechial hemorrhage.30 Cellmediated hypersensitivity reactions are believed responsible for allergic eosinophilic esophagitis and gastroenteritis (AEE, AEG). It is believed that activated T cells secrete IL-5 and other cytokines, attracting eosinophils and inducing the inflammatory response that causes the delayed onset of symptoms.31 Expansion studies of T cells from biopsy specimens of milk-induced AEG patients have revealed large numbers of CD4+ Th2 cells.32 In summary, the GI tract processes ingested food into a form that can be absorbed and used for energy and cell growth. During this process, nonimmunologic and immunologic mechanisms help destroy or block foreign antigens (e.g., bacteria, viruses, parasites, food proteins) from entering the body proper. Despite this elegant barrier, antigenically intact food proteins enter the circulation, but in the normal host are largely ignored by the immune system, which has become “tolerized” to these nonpathogenic substances.

CLINICAL FEATURES As depicted in Table 9-2, a number of GI food hypersensi­ tivity disorders have been described. Clinically, these disorders are generally divided into two main categories: IgE-Th2–mediated and non–IgE (cell)-mediated hypersen­ sitivities. There are a number of other disorders, however, that may result in symptoms similar to food-allergic reactions; these must be excluded during the evaluation (Table 9-3).

Table 9-2  Gastrointestinal Food Hypersensitivities IgE-Mediated Food Hypersensitivities Gastrointestinal allergy Infantile colic (minor subset) Oral allergy syndrome Mixed IgE- and Non–IgE-Mediated Hypersensitivities Allergic eosinophilic esophagitis Allergic eosinophilic gastritis Allergic eosinophilic gastroenteritis Allergic eosinophilic proctocolitis Non–IgE-Mediated Food Hypersensitivities Dietary protein-induced enterocolitis syndrome Dietary protein-induced enteropathy Celiac disease Dermatitis herpetiformis Mechanism Unknown Cow’s milk–induced occult gastrointestinal blood loss and iron deficiency anemia of infancy Gastroesophageal reflux disease Infantile colic (subset) Inflammatory bowel disease (?)

Table 9-3  Disorders That Must Be Differentiated from Food Hypersensitivities Food Intolerances Enterotoxigenic bacteria Vibrio cholerae, toxigenic Escherichia coli, Clostridium difficile Metabolic disorders Acrodermatitis enteropathica Hypo- or abetalipoproteinemia Primary carbohydrate malabsorption: lactase deficiency, sucrase deficiency Transient fructose and/or sorbitol malabsorption Postinfection malabsorption (secondary disaccharidase deficiency, villus atrophy, bile salt deconjugation) Bacterial: Shigella, Clostridium difficile Parasitic: Giardia, Cryptosporidium Viral: Rotavirus Anatomic Abnormalities Hirschsprung’s disease (especially with enterocolitis) Ileal stenosis Intestinal lymphangiectasia Short bowel syndrome Other Disorders Chronic nonspecific diarrhea of infancy Cystic fibrosis Inflammatory bowel disease Tumors Neuroblastoma Zollinger-Ellison syndrome (gastrinoma)

Long before IgE antibodies were identified, studies of food hypersensitivity focused on radiologic changes associated with immediate hypersensitivity reactions. In one of the first of these reports, hypertonicity of the transverse and pelvic colon and hypotonicity of the cecum and ascending colon were noted following wheat feeding to an allergic patient.33 In a later report, gastric retention, hypermotility of the small intestine, and colonic spasm were observed in 4 patients studied after administering barium containing specific food allergens.34 In a third study, fluoroscopy was used to compare the effect of barium contrast with and without food allergens in 12 food-allergic children35; gastric hypotonia and retention of the allergen test meal, prominent pylorospasm, and increased or decreased peristaltic activity of the intestines were noted.

Chapter 9  Food Allergies In the late 1930s, the rigid gastroscope was used to observe reactions in the stomachs of allergic patients. One study evaluated patients with GI food allergy or wheezing exacerbated by food ingestion and control subjects.36 Thirty minutes after a food allergen was placed on the gastric mucosa, patients with GI food allergy had markedly hyperemic and edematous patches of thick gray mucus and scattered petechiae at these sites, similar to those reported earlier by Walzer, in passively sensitized intestinal mucosal sites.15 Only mild hyperemia of the gastric mucosa was noted in patients with wheezing provoked by food ingestion. Subsequent studies confirmed these earlier observations and established an IgE-mediated mechanism for the reactions30; they demonstrated food-specific IgE antibodies and increased numbers of intestinal mast cells prior to challenge in food-allergic patients compared with normal controls, and significant decreases in stainable mast cells and tissue histamine content following a positive food challenge.

IgE-MEDIATED DISORDERS

The IgE-mediated food-induced GI allergic responses comprise two major symptom complexes: pollen-food allergy (oral allergy) syndrome and gastrointestinal allergy. These disorders are distinguished by their rapid onset, usually within minutes to an hour of ingesting the offending food. In addition, simple laboratory tests that detect food-specific IgE antibodies, such as prick skin tests and in vitro tests of serum food-specific IgE antibodies (e.g., ImmunoCAP; Phadia, Portage, Mich) often are useful in determining which foods are responsible for the patient’s symptoms.

Pollen-Food Allergy Syndrome

The pollen-food allergy syndrome (oral allergy syndrome) is a form of immediate contact hypersensitivity confined predominantly to the oropharynx and rarely involving other target organs.37 Symptoms include the rapid onset of pruritus and angioedema of the lips, tongue, palate and throat, generally followed by a rapid resolution of symptoms, and most commonly associated with the ingestion of various fresh (uncooked) fruits and vegetables. Symptoms result from local IgE-mediated reactions to conserved homologous proteins (sequences of amino acids in peptide backbones shared by plant pollens and fruit and vegetable proteins that remain unchanged through evolution) that are heat-labile (i.e., readily destroyed by cooking) and shared by certain fruits, vegetables, and some plant pollens.38 Patients with seasonal allergic rhinitis (hay fever) secondary to birch or ragweed pollen sensitivity often are afflicted with this syndrome. For example, in up to 50% of patients with ragweed-induced allergic rhinitis, ingestion of melons (e.g., watermelon, cantaloupe, honeydew) and bananas will provoke oral symptoms,39-41 whereas in birch pollen– allergic patients, symptoms may develop following the ingestion of raw potatoes, carrots, celery, apples, hazelnuts, and kiwi. Diagnosis is based on classic history and positive prick skin tests (e.g., prick and prick—pricking the fresh fruit or vegetable with a needle and then pricking the skin of the patient) with the implicated fresh fruits or vegetables.42

Gastrointestinal Allergy

Gastrointestinal allergy is a relatively common form of IgEmediated hypersensitivity, which generally accompanies allergic manifestations in other target organs (e.g., skin, airway) and results in a variety of symptoms.4 Symptoms typically develop within minutes to two hours of consum-

ing a food and consist of nausea, abdominal pain, cramps, vomiting, and/or diarrhea. In some infants, frequent ingestion of a food allergen appears to induce partial desensitization of gastrointestinal mast cells resulting in a subclinical reaction, with the only symptom reported being poor appetite and periodic abdominal pain. Diagnosis is established by clinical history, evidence of food-specific IgE antibodies (positive skin prick tests or serum food-specific IgE anti­ bodies), resolution of symptoms following complete elimination of the suspected food, and recurrence of symptoms following oral food challenges. GI allergy is common in IgEmediated food allergies, with more than 50% of children experiencing abdominal symptoms during double-blind, placebo-controlled food challenges.28

Infantile Colic

Infantile colic is an ill-defined syndrome of paroxysmal fussiness characterized by inconsolable agonized crying, drawing up of the legs, abdominal distention, and excessive gas. It generally develops in the first two to four weeks of life and persists through the third to fourth months of life.43 Various psychosocial and dietary factors have been implicated in the cause of infantile colic, but trials in bottle-fed and breast-fed infants have suggested that IgE-mediated hypersensitivity occasionally may be a pathogenic factor, possibly in 10% to 15% of colicky infants. Diagnosis of food-induced colic is established by the implementation of several brief trials of hypoallergenic formula. In infants with food allergen–induced colic, symptoms are generally shortlived, so prolonged restricted diets are generally unnecessary. Periodic rechallenges should be done every three to four months to determine when eliminated foods can be returned to the infant’s diet.

MIXED IgE- AND NON–IgE-MEDIATED DISORDERS Allergic eosinophilic esophagitis, gastroenteritis, and proctocolitis (AEE, AEG, AEP) may be caused by IgE- and/or non–IgE-mediated food allergies and are characterized by eosinophilic infiltration of the esophagus, stomach, and/or intestinal walls with peripheral eosinophilia in up to 50% of patients (see Chapter 27 for a more complete discussion).31,39-41 In the esophagus, basal hyperplasia and papillary lengthening are seen. The eosinophilic infiltrate may involve the mucosal, muscular, and/or serosal layers of the stomach or small intestine. Eosinophilic invasion of the muscular layer leads to thickening and rigidity of the stomach and small intestine, which may manifest as obstruction, whereas infiltration of the serosa commonly results in eosinophilic ascites. In most children with AEE-AEG, foodinduced IgE- and non–IgE-mediated reactions have been implicated in pathogenesis.44 Patients with IgE-mediated food-induced symptoms generally have atopic disease (atopic dermatitis, allergic rhinitis, and/or asthma), elevated serum IgE concentrations, positive skin prick tests to various foods and inhalants, peripheral blood eosinophilia, iron deficiency anemia, and hypoalbuminemia.

Allergic Eosinophilic Esophagitis

AEE manifests predominantly in young children, especially boys, with reflux or vomiting, irritability, food refusal, early satiety, and failure to thrive,45,46 whereas adults are more likely to present with reflux, epigastric or chest pain, dysphagia, and food impaction.43,47 Foodinduced AEE was first demonstrated in a group of 10 children with postprandial abdominal pain, early satiety or

143

144

Section II  Nutrition in Gastroenterology food refusal, vomiting or retching, failure to thrive, and refractoriness to standard medical therapy (4 of 10 had undergone Nissen fundoplication).44 Following six to eight weeks of an amino acid-based formula (Neocate) plus corn and apples, symptoms completely resolved in eight patients and were markedly improved in two others. Esophageal bio­psies revealed a marked reduction or clearing of the eosinophilic infiltrate and significant improvements in the basal zone hyperplasia and length of the vascular papillae. Symptoms could be reproduced with the introduction of certain foods. Some patients appear to have an association of pulmonary and esophageal inflammation, and some report seasonal esophageal symptoms.31,48 AEE appears to have increased in prevalence over the past decade, an observation some authors believe may be explained by the increased early use of antacids and pro­kinetic agents in young infants with symptoms of reflux. Because murine models of food-induced anaphylaxis require the use of antacids for sensitization,49,50 it is thought that antireflux medications may further compromise the young infant’s intestinal barrier function. In a cohort study of 152 adults using H2 blockers or proton pump inhibitors for three months, 10% of patients experienced an increase in foodspecific IgE and 15% developed de novo IgE to specific foods.51 Untersmayr and colleagues have noted that the use of antacid medications can lead to food sensitivity in children and adults.13 Diagnosis of AEE is based on a suggestive history, the demonstration of an eosinophilic infiltrate in the esophageal mucosa (>20 eosinophils/high-power field [×40]), and the absence of GERD, as evidenced by a normal pH monitoring study of the distal esophagus or lack of response to highdose proton pump inhibitors.52 Multiple biopsies are necessary because of the potential patchiness of the lesions; a single esophageal biopsy specimen has a sensitivity of 55%, whereas taking five biopsy specimens increases sensitivity to 100%.53 Esophagoscopy may reveal mucosal rings, furrowing, ulcerations, whitish papules (which represent eosinophilic abscesses), or strictures, but endoscopic findings are normal in at least one third of patients with AEE. There is some evidence suggesting that atopy patch testing may be useful in identifying foods responsible for the allergic inflammation, but further studies are necessary to confirm these early reports.54 Elimination of suspect foods for six to ten weeks should lead to resolution and normalization of esophageal histology, although clinical symptoms should improve substantially in three to six weeks.41,55,56 Challenges consist of reintroducing the suspected food allergen and evaluating for recurrence of symptoms and/or eosinophilic infiltrate on biopsy. If food allergens are not identified as provoking agents, oral glucocorticoids generally are required to alleviate symptoms. Although symptoms usually respond to glucocorticoid therapy, recurrence of symptoms is frequent when steroids are discontinued.57 Topical glucocorticoid therapy with swallowed fluticasone spray or viscous budesonide has been shown to induce remission in 50% to 80% of patients, but esophageal candidiasis may occur in up to 20% of patients using this form of treatment.58,59 If exacerbations recur, a daily regimen of low-dose prednisone or prednisolone or prednisone every other day may be successful in suppressing symptoms.60 Recent evidence suggests that anti–IL-5 may be useful in this disorder.61

Allergic Eosinophilic Gastroenteritis

AEG manifests with abdominal pain, nausea, vomiting, diarrhea, and weight loss.62 Generalized edema secondary to hypoalbuminemia may occur in some infants and young

children with marked protein-losing enteropathy, often in the presence of minimal gastrointestinal symptoms (e.g., occasional vomiting and diarrhea).63 Rarely AEG may manifest as pyloric stenosis in infants with outlet obstruction and postprandial projectile emesis.64 The immunopathogenesis of AEG is not known, but is believed to involve primarily cell-mediated mechanisms. A subset of patients have exacerbations of symptoms following the ingestion of food to which they have specific IgE antibodies, but most reactions do not appear to involve this mechanism. Peripheral blood T cells from all AEG patients evaluated have been shown to secrete excessive amounts of Th2 cytokines, IL-4, and IL-5 in vitro, compared with normal controls,65 and T cells expanded from duodenal biopsies of AEG patients express Th2 cytokines in vitro following antigen stimulation.32 The diagnosis of AEG is dependent on a suggestive history, gastrointestinal biopsy specimens demonstrating a prominent eosinophilic infiltration, and peripheral eosinophilia, which occurs in about 50% of patients. Lesions are not uniform; therefore, multiple biopsies are often necessary.62 Allergy skin testing may be helpful in some cases to identify causative foods, but often a therapeutic trial of an elemental diet for six to 10 weeks is necessary to determine whether food allergy is provoking the disorder. In a study of children with AEG and protein-losing enteropathy, institution of an amino acid-based formula therapy brought about resolution of symptoms and normalization of intestinal histology.63 As with AEE, if no sensitization is found, a trial of glucocorticoids is recommended, although relapses frequently occur when they are discontinued. The long-term prognosis of this disorder is not well characterized. In one series of children with AEG and protein-losing enteropathy, follow-up for 2.5 to 5.5 years revealed persistence of foodresponsive disease.

Allergic Eosinophilic Proctocolitis

AEP generally presents in the first few months of life and is most often secondary to cow’s milk or soy protein hypersensitivity. Over half of reported cases now occur in breastfed infants because of food antigens that are passed in maternal breast milk.66,67 Affected infants usually appear healthy, often have normally formed stools, and generally are evaluated because of the presence of gross or occult blood in their stools. Blood loss typically is minor but occasionally can produce anemia. Lesions generally are confined to the distal large bowel and consist of mucosal edema, with infiltration of eosinophils in the epithelium and lamina propria. In severe cases with crypt destruction, neutrophils also are prominent. The immunologic mechanism underlying this disorder is not known, but is believed to involve a cell-mediated reaction. There is no evidence that IgE antibodies are involved in this disorder and therefore skin prick testing or evaluation of food-specific IgE antibodies is not helpful. Diagnosis can be established when elimination of the responsible allergen leads to resolution of hematochezia, generally with dramatic improvement within 72 hours of appropriate food allergen elimination. Complete clearing and resolution of mucosal lesions may take up to one month. Reintroduction of the allergen leads to recurrence of symptoms within several hours to days. Sigmoidoscopic findings vary and range from areas of patchy mucosal injection to severe friability, with small aphthoid ulcerations and bleeding. Colonic biopsy reveals a prominent eosinophilic infiltrate in the crypt epithelia and lamina propria. Children with cow’s milk and soy protein–induced proctocolitis usually outgrow their protein sensitivity (i.e., become clinically tolerant

Chapter 9  Food Allergies within six months to two years of allergen avoidance), but occasionally refractory cases are seen.

NON–IgE-MEDIATED DISORDERS

Some gastrointestinal food-allergic disorders are clearly not IgE-mediated, and are believed to be the result of various cell-mediated mechanisms. Consequently, tests for evidence of food-specific IgE antibodies are of no value to identify the responsible food in these disorders. These non– IgE-mediated hypersensitivities are believed to result from different abnormal antigen processing and/or cell-mediated mechanisms, and may be divided into the following syndromes: dietary protein-induced enterocolitis and dietary protein-induced enteropathy.68

Dietary Protein-Induced Enterocolitis Syndrome

Dietary protein-induced enterocolitis syndrome is a dis­ order most commonly seen in young infants, presenting between one week and three months of age, with protracted vomiting and diarrhea that not infrequently results in dehydration.69,70 About one third of infants with severe diarrhea develop acidosis and transient methemoglobinemia. Cow’s milk and/or soy protein most often are responsible, but enterocolitis secondary to egg, wheat, rice, oat, peanut, nuts, chicken, turkey, and fish sensitivities has also been reported in older individuals.71 Breast-fed babies almost never develop symptoms while breast-feeding, but may be sen­ sitized through food proteins passed in the breast milk and experience a reaction on the first few feedings of the whole food.72,73 Similar reactions to seafood (e.g., shrimp, crab, lobster), with symptoms developing about two to four hours following ingestion, often are reported in adults. Stools frequently contain occult blood, polymorpho­ nuclear neutrophils, and eosinophils. Jejunal biopsies reveal flattened villi, edema, and increased numbers of lymphocytes, eosinophils, and mast cells. Food challenges generally result in vomiting and diarrhea within one to three hours, and result in hypotension in about 15% of cases. The immunopathogenesis of this syndrome remains unknown. Some studies suggest that food antigen-induced secretion of TNF-α from local mononuclear cells (e.g., macrophages, dendritic cells) may account for the reaction.74 Other studies indicate that the disorder may be caused by lower expression of type 1 TGF-β receptors than type 2 receptors, suggesting differential contributions of each receptor to the diverse biological activities of TGF-β in the intestinal epithelium.75 Some studies have suggested that atopy patch testing with the suspected food may be useful in distinguishing which children will develop symptoms following ingestion, but most such evidence is not convincing.76 Diagnosis can be established when elimination of the responsible allergen leads to resolution of symptoms within 72 hours and oral challenge provokes symptoms.72 Secondary disaccharidase deficiency may persist longer, however, and may result in ongoing diarrhea for up to two weeks. Oral food challenges consist of administering 0.3 to 0.6 g/kg body weight of the suspected protein allergen while monitoring the peripheral blood white cell count. Vomiting generally develops within one to four hours of administering the challenge food, whereas diarrhea or loose stools often develop after four to eight hours. In conjunction with a positive food challenge, the absolute neutrophil count in the peripheral blood will increase at least 3500 cells/mm3 within four to six hours of developing symptoms, and neutrophils and eosinophils may be found in the stools. About 15% of food antigen challenges lead to profuse vomiting, dehydration, and hypotension, so they must be performed under medical supervision.

Dietary Protein-Induced Enteropathy

Dietary protein-induced enteropathy (excluding celiac disease) frequently manifests in the first several months of life with diarrhea (mild to moderate steatorrhea in about 80%) and poor weight gain.67,77 Symptoms include protracted diarrhea, vomiting in up to two thirds of patients, failure to thrive, and malabsorption, demonstrated by the presence of reducing substances in the stools, increased fecal fat, and abnormal d-xylose absorption. Cow’s milk sensitivity is the most frequent cause of this syndrome, but it also has been associated with sensitivities to soy, egg, wheat, rice, chicken, and fish. The diagnosis is established by identifying and excluding the responsible allergen from the diet, which should result in resolution of symptoms within several days to weeks. On endoscopy, patchy villus atrophy is evident and biopsy reveals a prominent mononuclear round cell infiltrate and a small number of eosinophils, similar to celiac disease, but generally much less extensive. Colitic features such as mucus and gross or microscopic hematochezia usually are absent, but anemia occurs in about 40% of affected infants and protein loss occurs in most. Complete resolution of the intestinal lesions may require 6 to 18 months of allergen avoidance. Unlike celiac disease, loss of protein sensitivity and clinical reactivity frequently occurs, but the natural history of this disorder has not been well studied.

Celiac Disease

Celiac disease (CD) is a more extensive enteropathy leading to malabsorption (see details in Chapter 104). Total villus atrophy and extensive cellular infiltrate are associated with sensitivity to gliadin, the alcohol-soluble portion of gluten found in wheat, rye, and barley. CD is strongly associated with HLA-DQ2 (α1*0501, β1*0201), which is present in more than 90% of CD patients.78 The incidence of CD has been reported as 1 in 250 in the United States. The striking increase in CD in Sweden compared with genetically similar Denmark,79 and the variation in prevalence associated with changes in patterns of gluten feeding in Sweden,80 strongly implicate environmental factors (e.g., feeding practices) in the cause of this disorder.81 The intestinal inflammation in CD is precipitated by exposure to gliadin and is associated with increased mucosal activity of tissue transglutaminase (tTG), which deamidates gliadin in an ordered and specific fashion, creating epitopes that bind efficiently to DQ2 and are recognized by T cells.82 Initial symptoms often include diarrhea or frank steatorrhea, abdominal distention and flatulence, weight loss, and occasionally nausea and vomiting. Oral ulcers and other extraintestinal symptoms secondary to malabsorption are not uncommon. Villus atrophy of the small bowel is a characteristic feature of CD patients who are ingesting gluten. IgA antibodies to gluten are present in more than 80% of adults and children with untreated CD.83 In addition, patients generally have increased IgG antibodies to a variety of foods, presumably the result of increased food antigen absorption. Diagnosis has been dependent on demonstrating biopsy evidence of villus atrophy and an inflammatory infiltrate, resolution of biopsy findings after six to 12 weeks of gluten elimination, and recurrence of biopsy changes following gluten challenge. Revised diagnostic criteria have been proposed, however, that require greater dependency on serologic studies. Quantitation of IgA tTG antibodies may be used for screening in children older than two years. Diagnosis of CD, however, requires an intestinal biopsy showing clear-cut evidence of villus atrophy plus resolution of symptoms on a gluten-free diet, with serologic follow-up showing disappearance of the

145

146

Section II  Nutrition in Gastroenterology antibodies to confirm the diagnosis further.84,85 Once the diagnosis of CD is established, lifelong elimination of gluten-containing foods is necessary to control symptoms and possibly to avoid the increased risk of gastrointestinal malignancy.86

Dermatitis Herpetiformis

Dermatitis herpetiformis (DH) is a chronic blistering skin disorder associated with a gluten-sensitive enteropathy. It is characterized by a chronic, intensely pruritic, papulo­ vesicular rash symmetrically distributed over the extensor surfaces and buttocks.87,88 The histology of the intestinal lesion is almost identical to that seen in CD, although villus atrophy and the inflammatory infiltrate are generally milder and T-cell lines isolated from intestinal biopsy specimens of DH patients produce significantly more IL-4 than T cell lines isolated from CD patients.89 Although many patients have minimal or no gastrointestinal complaints, biopsy of the small bowel generally confirms intestinal involvement. Elimination of gluten from the diet generally leads to resolution of skin symptoms and normalization of intestinal findings over several months. Administration of sulfones, the mainstay of therapy, leads to rapid resolution of skin symptoms, but has almost no effect on intestinal symptoms.

Other Gastrointestinal Disorders

Several other disorders have been suggested to be caused by food protein hypersensitivity. Ingestion of pasteurized whole cow’s milk by infants younger than six months may lead to occult GI blood loss and occasionally to iron deficiency anemia.90,91 Substitution of heat-processed infant formula (including cow’s milk–derived formulas) for whole cow’s milk generally leads to resolution of symptoms within three days. Gastroesophageal reflux (GER) in young infants may be the result of food-induced AEE. In a study of 204 infants younger than one year with GER (diagnosed with a 24-hour esophageal pH test and esophageal biopsy),85 42% were diagnosed with cow’s milk–induced reflux by blinded milk challenges. These infants experienced resolution of GER and normalization of pH studies once cow’s milk was eliminated from the diet.92 Constipation also has been reported to be caused by milk allergy,93 although the underlying mechanism is not clear. Circumstantial evidence suggests a possible role of food allergy in inflammatory bowel disease (Crohn’s disease and ulcerative colitis), but convincing evidence of an immunopathogenic role remains to be established.

DIAGNOSIS The diagnosis of food allergy is a clinical exercise involving a careful history, physical examination, and selective laboratory studies. Various tests are used for the evaluation of food hypersensitivity (see Sicherer and Sampson4 and Sampson94). In some cases, the medical history may be useful in diagnosing food allergy (e.g., acute anaphylaxis after the isolated ingestion of peanuts). Fewer than 50% of reported food-allergic reactions, however, can be verified by a double-blind, placebo-controlled food challenge. Information useful in establishing that a food-allergic reaction has occurred and in constructing an appropriate oral food challenge includes the following: (1) food presumed to have provoked the reaction; (2) quantity of the suspected food ingested; (3) length of time between ingestion and development of symptoms; (4) type of symptoms provoked; and (5) whether similar symptoms developed on other occasions when the food was eaten. Although any food may induce an allergic reaction, a few foods are responsible for the vast majority of reactions (Table 9-4). Figure 9-2 depicts a standard approach for evaluating and managing adverse food reactions. If an IgE-mediated disorder is suspected, selected skin prick tests or quantification of food-specific IgE antibodies (e.g., ImmunoCAP) followed by an appropriate exclusion diet and blinded food challenge are warranted. If a non–IgE-mediated GI hypersensitivity disorder is suspected, laboratory and endoscopic studies (with or without oral food challenges) are required to arrive at the correct diagnosis (see earlier). Table 9-5 compares the Table 9-4  Foods Responsible for Most Food Hypersensitivity Disorders IgE-MEDIATED FOOD HYPERSENSITIVITIES*

NON–IgE-MEDIATED FOOD HYPERSENSITIVITIES†

Milk Egg Peanuts Shellfish Tree nuts Sesame Fish Soy Wheat

Barley Beef, lamb Egg Fish Milk Shellfish Soy Wheat White potato

*Listed in order of overall prevalence. † Listed alphabetically.

Table 9-5  Differentiating Non–IgE-Mediated Gastrointestinal Food Hypersensitivities

PARAMETER

AEE, AEG

DIETARY PROTEININDUCED ENTEROPATHY

Age of onset Duration Food proteins implicated

1 mo and older ≥1 yr Cow’s milk, egg, soy, wheat, barley

1-18 mo 18-36 mo Cow’s milk, soy, wheat, barley

2 wk-9 mo 9-36 mo Cow’s milk, soy

1 wk-3 mo 6-18 mo Cow’s milk, soy, breast milk*

Moderate to severe Prominent† Minimal Minimal to moderate

Moderate Variable Moderate Moderate

Moderate Prominent Severe Moderate

None None Rare Moderate to severe

Clinical Features Failure to thrive or weight loss Vomiting Diarrhea Hematochezia

*Food proteins in breast milk (most often cow’s milk or egg protein). † Retching or gastroesophageal reflux. AEE, allergic eosinophilic esophagitis; AEG, allergic eosinophilic gastroenteritis.

DIETARY PROTEININDUCED ENTEROCOLITIS

DIETARY PROTEININDUCED PROCTOCOLITIS

Chapter 9  Food Allergies Clinical History

Adverse food reaction likely (possible foods identified)

Toxic

Adverse food reaction unlikely (unless a non–IgE-mediated food hypersensitivity is possible)

Nontoxic

Educate

Food intolerance

Laboratory studies

Finished

Food hypersensitivity

Suspect Non–IgE-mediated

Suspect IgE-mediated Skin-prick tests

Laboratory studies and/or endoscopy

Suggestive Symptoms persist; look for other cause

Positive

Negative

History of anaphylaxis

Finished (unless a non–IgE-mediated food hypersensitivity is possible)

No

Yes

Restrict food

Elimination diet Symptoms improve

No symptoms occur

Resume regular diet

Finished

Symptoms recur Eliminate food from diet

Open challenge with foods least likely to provoke symptoms

Symptoms

Short list of food likely to provoke symptoms; blinded challenges

No symptoms; add food back

Symptoms; restrict food

Figure 9-2.  Algorithm for the evaluation and management of adverse food reactions.

main features of four non–IgE-mediated food-allergic disorders. An exclusion diet eliminating all foods suspected by history and/or skin testing (for IgE-mediated disorders) should be conducted for one to two weeks in suspected IgE-mediated disorders, food-induced enterocolitis, and benign eosinophilic proctocolitis. Exclusion diets may need to be extended for as long as 12 weeks in other suspected GI hypersensitivity disorders (e.g., food protein-induced enteropathy, AEE, or AEG) and may require the use of ele-

mental diets (e.g., Vivonex, Neocate One+, or EleCare) to exclude all antigens. If no improvement is noted and dietary compliance is ensured, it is unlikely that food allergy is involved. Before undertaking blinded food challenges (single- or double-blind), suspect foods should be eliminated from the diet for 7 to 14 days before challenge and even longer in some disorders when secondary disaccharidase deficiency may have developed, as noted earlier. Prescribing elimination diets, like prescribing medications,

147

148

Section II  Nutrition in Gastroenterology may have adverse effects (e.g., malnutrition or eating dis­ orders) and should not be done in the absence of evidence that they are likely to be beneficial.

TREATMENT AND NATURAL HISTORY Once the diagnosis of food hypersensitivity is established, strict elimination of the offending allergen is the only proven therapy. Patients must be taught to scrutinize food labels to detect potential sources of hidden food allergens.95 Drugs such as H1 and H2 antihistamines and glucocorticoids modify symptoms to food allergens but, overall, have minimal efficacy or unacceptable side effects. Anti–IL-5 antibodies have shown promise in the treatment of eosinophilic disorders.61 The prevalence of food hypersensitivity is greatest in the first few years of life, but most young children outgrow their food hypersensitivity within three to five years, except possibly for IgE-mediated hypersensitivities to peanuts, nuts, and seafood.28 Although younger children are more likely to outgrow food hypersensitivity, older children and adults also may lose their food hypersensitivity (i.e., develop clinical tolerance and be able to ingest the food without symptoms) if the responsible food allergen can be identified and eliminated from the diet for a period of time.96,97 Gastrointestinal food allergies affect about 4% of children younger than three years and about 1% of the general population. Current research in this field is providing new information regarding the pathogenesis of these disorders and should lead to the development of new diagnostic and therapeutic algorithms. In the interim, specific food hypersensitivities must be diagnosed carefully, and patients must be educated to avoid ingesting the responsible food allergens.

KEY REFERENCES

Breiteneder H, Clare Mills EN. Plant food allergens—structural and functional aspects of allergenicity. Biotechnol Adv 2005; 23:395-9. (Ref 38.)

Chehade M, Magid MS, Mofidi S, et al. Allergic eosinophilic gastroenteritis with protein-losing enteropathy: intestinal pathology, clinical course, and long-term follow-up. J Pediatr Gastroenterol Nutr 2006; 42:516-21. (Ref 63.) Chehade M, Mayer L. Oral tolerance and its relation to food hypersensitivities. J Allergy Clin Immunol 2005; 115:3-12. (Ref 12.) Furuta GT, Liacouras CA, Collins MH, et al. Eosinophilic esophagitis in children and adults: A systematic review and consensus recommendations for diagnosis and treatment. Gastroenterol 2007; 133:1342-63. (Ref 52.) Khan S, Orenstein SR. Eosinophilic gastroenteritis. Gastroenterol Clin North Am 2008; 37:333-48. (Ref 62.) Leffler DA, Kelly CP. Update on the evaluation and diagnosis of celiac disease. Curr Opin Allergy Clin Immunol 2006; 6:191-6. (Ref 78.) Novak N, Haberstok J, Bieber T, Allam JP. The immune privilege of the oral mucosa. Trends Mol Med 2008; 14:191-8. (Ref 19.) Nowak-Wegrzyn A, Sampson HA, Wood RA, Sicherer SH. Food proteininduced enterocolitis syndrome caused by solid food proteins. Pediatr 2003; 111:829-35. (Ref 71.) Rothenberg ME. Eosinophilic gastrointestinal disorders (EGID). J Allergy Clin Immunol 2004; 113:11-28. (Ref 31.) Sampson HA. Food allergy. Part 2: Diagnosis and management. J Allergy Clin Immunol 1999; 103:981-99. (Ref 94.) Sampson HA, Sicherer SH, Birnbaum AH. AGA technical review on the evaluation of food allergy in gastrointestinal disorders. American Gastroenterological Association. Gastroenterol 2001; 120:1026-40. (Ref 68.) Sicherer SH. Food protein-induced enterocolitis syndrome: Clinical perspectives. J Pediatr Gastroenterol Nutr 2000; 30:S45-9. 2000. (Ref 70.) Sicherer SH, Sampson HA. Food allergy. J Allergy Clin Immunol 2006; 117:S470-5. (Ref 4.) Vercelli D. Immunoglobulin E and its regulators. Curr Opin Allergy Clin Immunol 2001; 1:61-5. (Ref 24.) Wickens K, Black PN, Stanley TV, et al. A differential effect of 2 pro­ biotics in the prevention of eczema and atopy: A double-blind, randomized, placebo-controlled trial. J Allergy Clin Immunol 2008; 122:788-94 (Ref 21.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

10 Acute Abdominal Pain Frederick H. Millham

CHAPTER OUTLINE Anatomy  151 Visceral Pain  151 Somatic-Parietal Pain  152 Referred Pain  152 Clinical Evaluation  153 Approach to Acute Care  154 History  154 Physical Examination  155 Laboratory Data  156 Imaging Studies  156 Other Diagnostic Tests  156 Causes  156 Acute Appendicitis  157

Acute abdominal pain is a common complaint that brings patients to emergency departments. As many as 1 in 20 emergency department visits is for abdominal pain.1 Approximately half of these patients have nonspecific findings or “gastroenteritis.”2 The other half have a more serious disorder that warrants further evaluation and treatment. A small proportion of patients has a life-threatening disease. Therefore, the evaluation of acute abdominal pain must be efficient and lead to an accurate diagnosis early in the presentation so that the treatment of patients who are seriously ill is not delayed and patients with self-limited disorders are not overtreated. This chapter discusses the anatomic factors that determine how abdominal pain is perceived, a systematic approach to the evaluation of abdominal pain, and common and special circumstances encountered in evaluating patients with acute abdominal pain.

ANATOMY Physiologic determinants of pain include the nature of the stimulus, the type of receptor involved, the organization of the neural pathways from the site of injury to the central nervous system, and a complex interaction of modifying influences on the transmission, interpretation, and reaction to pain messages.3,4 Sensory neuroreceptors in abdominal organs are located in the mucosa and muscularis of hollow viscera, on serosal structures such as the peritoneum, and within the mesentery.5 In addition to nociception (the perception of noxious stimuli), sensory neuroreceptors are involved in the regulation of secretion, motility, and blood flow via local and central reflex arcs.6 Although sensory information conveyed in this manner usually is not perceived, disordered regulation of these gastrointestinal functions (secretion, motility, blood flow) can cause pain. For example, patients with irritable bowel syndrome perceive

Acute Biliary Disease  157 Small Bowel Obstruction  158 Acute Diverticulitis  158 Acute Pancreatitis  158 Perforated Peptic Ulcer  159 Acute Mesenteric Ischemia  159 Abdominal Aortic Aneurysm  160 Abdominal Compartment Syndrome  160 Other Intra-Abdominal Causes  160 Extra-Abdominal Causes  160 Special Circumstances  161 Pharmacologic Management  161

pain as a result of heightened sensitivity of intestinal afferent neurons to normal endogenous stimuli that results in altered gut motility and secretion (see Chapter 118).7 Abdominal pain is transmitted by two distinct types of afferent nerve fibers, unmyelinated C fibers and myelinated A-δ fibers. These two types of nerve fibers result in the perception of two different types of abdominal pain (visceral and somatic-parietal pain, respectively); interplay between the two systems results in a third type of pain, referred pain.

VISCERAL PAIN

Visceral pain is transmitted by C fibers that are found in muscle, periosteum, mesentery, peritoneum, and viscera. Most painful stimuli from abdominal viscera are conveyed by this type of fiber and tend to be dull, cramping, burning, poorly localized, and more gradual in onset and longer in duration than somatic pain. Because abdominal organs transmit sensory afferents to both sides of the spinal cord, visceral pain is usually perceived to be in the midline, in the epigastrium, periumbilical region, or hypogastrium (Fig. 10-1). Visceral pain is not well localized because the number of nerve endings in viscera is lower than that in highly sensitive organs such as the skin and because the innervation of most viscera is multisegmental. The pain is generally described as cramping, burning, or gnawing. Secondary autonomic effects such as sweating, restlessness, nausea, vomiting, perspiration, and pallor often accompany visceral pain. The patient may move about in an effort to relieve the discomfort. The afferent fibers that mediate painful stimuli from the abdominal viscera follow the distribution of the autonomic nervous system (Fig. 10-2). The cell bodies for these fibers are located in the dorsal root ganglia of spinal afferent nerves. On entering the spinal cord, these fibers branch into the dorsal horn and tract of Lissauer, where afferent nerves from adjacent spinal segments travel in a cephalad direction and caudally over one or two spinal segments

151

152

Section III  Symptoms, Signs, and Biopsychosocial Issues

A

1 2 3 1

2

Abdominal visceral nociceptors also respond to various chemical stimuli. Chemical nociceptors are contained mainly in the mucosa and submucosa of the hollow viscera. These receptors are activated directly by substances released in response to local mechanical injury, inflammation, tissue ischemia and necrosis, and noxious thermal or radiation injury. Such substances include H+ and K+ ions, histamine, serotonin, bradykinin and other vasoactive amines, substance P, calcitonin gene-related peptide, prostaglandins, and leukotrienes.11,12 Accumulation of nociceptor-reactive substances may change the microenvironment of the injured tissue and thereby reduce the pain threshold. The sensation of pain to a given stimulus is thus increased, and otherwise innocuous stimuli become painful. For example, the application of chemical irritants or pressure to normal gastric mucosa is not painful, whereas the application of the same stimuli to inflamed or injured gastric mucosa causes pain.

SOMATIC-PARIETAL PAIN 3

Figure 10-1.  Localization of visceral pain. Pain arising from organ areas depicted in 1, 2, and 3 is felt in the epigastrium, midabdomen, and hypogastrium, respectively, as shown in A. The arrow in A indicates biliary pain that is referred to the right scapular area.

before terminating on dorsal horn cells in laminae I and V. The dorsal horn cells within laminae I and V are the primary projection neurons for ascending pain pathways. From the dorsal horn, second-order neurons transmit nociceptive impulses via fibers that pass across the anterior commissure and ascend the spinal cord in the contralateral spinothalamic tract. These fibers project to the thalamic nuclei and the reticular formation nuclei of the pons and medulla. The thalamic nucleus sends third-order neurons to the somatosensory cortex, where the discriminative aspects of pain are perceived. The reticular formation nucleus sends neurons to the limbic system and frontal cortex, where the emotional aspects of pain are interpreted.8,9 Abdominal visceral nociceptors respond to mechanical and chemical stimuli. The principal mechanical signal to which visceral nociceptors are sensitive is stretch; cutting, tearing, or crushing of viscera does not result in pain. Visceral stretch receptors are located in the muscular layers of the hollow viscera, between the muscularis mucosa and submucosa, in the serosa of solid organs, and in the mesentery (especially adjacent to large vessels).5,10 Mechanoreceptor stimulation can result from rapid distention of a hollow viscus (e.g., intestinal obstruction), forceful muscular contractions (e.g., biliary pain or renal colic), and rapid stretching of solid organ serosa or capsule (e.g., hepatic congestion). Similarly, torsion of the mesentery (e.g., cecal volvulus) or tension from traction on the mesentery or mesenteric vessels (e.g., retroperitoneal or pancreatic tumor) results in stimulation of mesenteric stretch receptors.

Somatic-parietal pain is mediated by A-δ fibers that are distributed principally to skin and muscle. Signals from this neural pathway are perceived as sharp, sudden, welllocalized pain, such as that which follows an acute injury. These fibers convey pain sensations through spinal nerves. Stimulation of these fibers activates local regulatory reflexes mediated by the enteric nervous system and long spinal reflexes mediated by the autonomic nervous system, in addition to transmitting pain sensation to the central nervous system.13 Somatic-parietal pain arising from noxious stimulation of the parietal peritoneum is more intense and more precisely localized than visceral pain. An example of this difference occurs in acute appendicitis, in which the early vague periumbilical visceral pain is followed by the localized somatic-parietal pain at McBurney’s point that is produced by inflammatory involvement of the parietal peritoneum. Somatic-parietal pain is usually aggravated by movement or vibration. The nerve impulses that mediate such pain travel in somatic sensory spinal nerves. The fibers reach the spinal cord in the peripheral nerves that correspond to the cutaneous dermatomes of the skin, the sixth thoracic (T6) to first lumbar (L1) vertebra. Lateralization of the discomfort of parietal pain is possible because only one side of the nervous system innervates a given part of the parietal peritoneum. Reflexive responses, such as involuntary guarding and abdominal rigidity, are mediated by spinal reflex arcs involving somatic-parietal pain pathways. Afferent pain impulses are modified by inhibitory mechanisms at the level of the spinal cord. Somatic A-d fibers mediate touch, vibration, and proprioception in a dermatomal distribution that matches the visceral innervation of the injured viscera and synapse with inhibitory interneurons of the substantia gelatinosa in the spinal cord. In addition, inhibitory neurons that originate in the mesencephalon, periventricular gray matter, and caudate nucleus descend within the spinal cord to modulate afferent pain pathways. These inhibitory mechanisms allow cerebral influences to modify afferent pain impulses.9,14

REFERRED PAIN

Referred pain is felt in areas remote from the diseased organ and results when visceral afferent neurons and somatic afferent neurons from a different anatomic region converge on second-order neurons in the spinal cord at the same spinal segment. This convergence may result from the innervation, early in embryologic development, of adjacent

Chapter 10  Acute Abdominal Pain Heart Midbrain

Medulla

Larynx Trachea Bronchi Lungs

Vagus nerve

C1

Superior cervical ganglion

Esophagus Stomach

T1 Celiac ganglion

Abdominal blood vessels Liver Bile ducts Pancreas

Superior mesenteric ganglion L1 Inferior mesenteric ganglion

Adrenal Small intestine Large intestine

S1 Kidney Pelvic nerves

Bladder

Reproductive organs

structures that subsequently migrate away from each other. As such, referred pain can be understood to refer to an earlier developmental state. For example, the central tendon of the diaphragm begins its development in the neck and moves craniocaudad, bringing its innervation, the phrenic nerve, with it.15 Figure 10-3 shows how diaphragmatic irritation from a subphrenic hematoma or splenic rupture may be perceived as shoulder pain (Kehr’s sign).9

CLINICAL EVALUATION Effective evaluation of a patient with acute abdominal pain (an acute abdomen) requires careful but expeditious history taking and physical examination (often repeated serially) and, in many cases, informed use of imaging studies. When a carefully performed history and physical examination are paired with appropriate and timely imaging, an accurate diagnosis can often be determined relatively quickly. Inadequate clinical evaluation or poor selection of imaging methods leads to unnecessary delay, often resulting in a

Figure 10-2.  Pathways of visceral sensory innervation. The visceral afferent fibers that mediate pain travel with autonomic nerves to communicate with the central nervous system. In the abdomen, these fibers include vagal and pelvic parasympathetic nerves and thoracolumbar sympathetic nerves. Sympathetic fibers (red lines); parasympathetics (blue lines).

To brain C B Spinal cord A Visceral afferent first-order neuron A B Spinal cord second-order neuron Somatic afferent first-order neuron C

Figure 10-3.  Demonstration of the neuroanatomic basis of referred pain. Visceral afferent fibers that innervate the diaphragm can be stimulated by local irritation (e.g., subdiaphragmatic abscess [circle]). These visceral afferent fibers (A) synapse with second-order neurons in the spinal cord (B) as well as somatic afferent fibers (C) arising from the left shoulder area (cervical roots 3 to 5). The brain interprets the pain to be somatic in origin and localizes it to the shoulder.

153

Section III  Symptoms, Signs, and Biopsychosocial Issues poor outcome. Currently, common entities such as appendicitis, cholecystitis, and diverticulitis can be diagnosed with almost complete accuracy; patients with other diseases require an orderly and efficient evaluation and judicious selection of imaging studies.

APPROACH TO ACUTE CARE

When approaching a patient with acute abdominal pain, the physician should begin with a rapid assessment of the patient’s overall physiologic state. Quickly assessing the three domains according to the mnemonic “ABC” will help identify patients who are unstable and therefore require expedited treatment: A: Airway: Is the patient able to maintain an airway? Does an impaired sensorium endanger the patient’s airway or pose a risk for aspiration of vomit or oral secretions? B: Breathing: How effectively is the patient breathing? Are breaths rapid and shallow? Is the use of accessory muscles evident? Does the patient appear tachypneic? C: Circulation: Circulation encompasses three areas of assessment: (1) Is the patient in shock, as suggested by pallor, cyanosis, mottling, prostration, hypotension, tachycardia, or other signs of hypoperfusion? (2) Has intravenous access been established? (3) Is there evidence of active bleeding? If hemodynamic instability is apparent, including clinical evidence of shock, surgical consultation should be sought immediately, and consideration should be given to endotracheal intubation and resuscitation early in the encounter. The adage in acute care surgery that “death begins in radiology” should be a reminder that hemodynamic resuscitation should precede diagnostic imaging. Patients who are in shock demand urgent care and should not be sent for imaging studies without aggressive resuscitation and monitoring.

Chronology

The time courses of several common causes of acute abdominal pain are diagrammed in Figure 10-4. The rapidity of onset of pain is often a measure of the severity of the underlying disorder. Pain that is sudden in onset, severe, and well localized is likely to be the result of an intra-abdominal catastrophe such as a perforated viscus, mesenteric infarction, or ruptured aneurysm. Affected patients usually recall the exact moment of onset of their pain. Progression is an important temporal factor in abdominal pain. In some disorders, such as gastroenteritis, pain is self-limited, whereas in others, such as appendicitis, pain is progressive. Colicky pain has a crescendo-decrescendo pattern that may be diagnostic, as in renal colic. The duration of abdominal pain is also important. Patients who seek evaluation of abdominal pain that has been present for an extended period (e.g.,

D

B

C Severity

154

A

HISTORY

Despite the advances made in clinical imaging, history taking remains the most important component of the initial evaluation of the patient with acute abdominal pain.16 Characteristic features of the pain associated with various common causes of acute abdominal pain are shown in Table 10-1. Attention to these features can lead to a rapid clinical diagnosis or exclusion of important diseases in the differential diagnosis, thus enhancing the reliability and effectiveness of subsequent diagnostic testing.2

Time Figure 10-4.  Patterns of acute abdominal pain. A, Many causes of abdominal pain subside spontaneously with time (e.g., gastroenteritis). B, Some pain is colicky (i.e., the pain progresses and remits over time); examples include intestinal, renal, and biliary pain (colic). The time course may vary widely from minutes in intestinal and renal pain to days, weeks, or even months in biliary pain. C, Commonly, acute abdominal pain is progressive, as in acute appendicitis or diverticulitis. D, Certain conditions have a catastrophic onset, such as ruptured abdominal aortic aneurysm.

Table 10-1  Comparison of Common Causes of Acute Abdominal Pain Cause

ONSET

LOCATION

CHARACTER

DESCRIPTOR

RADIATION

Appendicitis

Gradual

None

++

Acute Acute Gradual Sudden

Constricting Boring Ache Burning

Scapula Midback None None

++ ++ to +++

Small bowel obstruction Mesenteric ischemia, infarction Ruptured abdominal aortic aneurysm Gastroenteritis Pelvic inflammatory disease Ruptured ectopic pregnancy

Gradual Sudden

Periumbilical area Periumbilical area

Diffuse early; localized late Localized Localized Localized Localized early, diffuse late Diffuse Diffuse

Ache

Cholecystitis Pancreatitis Diverticulitis Perforated peptic ulcer

Periumbilical area early; RLQ late RUQ Epigastrium, back LLQ Epigastrium

Cramping Agonizing

None None

++ +++

Sudden

Abdomen, back, flank

Diffuse

Tearing

None

+++

Gradual Gradual Sudden

Periumbilical area Either LQ, pelvis Either LQ, pelvis

Diffuse Localized Localized

Spasmodic Ache Sharp

None Upper thigh None

+ to ++ ++ ++

+ = mild; ++ = moderate; +++ = severe; LLQ = left lower quadrant; LQ = lower quadrant; RLQ = right lower quadrant; RUQ = right upper quadrant.

INTENSITY

+++

Chapter 10  Acute Abdominal Pain weeks) are less likely to have an acute life-threatening illness than patients who present within hours to days of the onset of their symptoms.

lent vomitus suggests more distal small bowel or colonic obstruction. A constellation of findings may indicate a particular disease entity.

Location

Past Medical History

The location of abdominal pain provides a clue to interpreting the cause. As noted, a given noxious stimulus may result in a combination of visceral, somatic-parietal, and referred pain, thereby creating confusion in interpretation unless the neuroanatomic pathways are considered. For example, the pain of diaphragmatic irritation from a left-sided subphrenic abscess may be referred to the shoulder and misinterpreted as pain from ischemic heart disease (see Fig. 10-3). Changes in location may represent progression from visceral to parietal irritation, as with appendicitis, or represent the development of diffuse peritoneal irritation, as with a perforated ulcer.

Intensity and Character

Acute abdominal pain usually follows one of three patterns. Pain that is prostrating, physically incapacitating the sufferer, is usually caused by a severe, life-threatening disease such as a perforated viscus, ruptured aneurysm, or severe pancreatitis. By contrast, patients with obstruction of a hollow viscus, as in intestinal obstruction, renal colic, or biliary pain, present with the gradual onset of cramping pain that follows a sinusoidal pattern of intense pain alternating with a period of relief. Nausea and vomiting are characteristic symptoms associated with this group of disorders. The obstructed viscus need not be the intestine for nausea or vomiting to occur, as in the case of a kidney stone. The third pattern is of gradually increasing discomfort, usually vague and poorly localized at the start, but becoming more localized as the pain intensifies. This picture is usually caused by inflammation, as with acute appendicitis or diverticulitis. Some disorders, such as acute cholecystitis, may start out as colicky pain but evolve into a constant pain as cystic duct obstruction leads to gallbladder inflammation. The clinician should be cautious, however, in assigning too much importance to a patient’s description of the pain; exceptions are common, and a given descriptor may be attributable to a number of conditions.

Aggravating and Alleviating Factors

The relationship of the pain to positional changes, meals, bowel movements, and stress may yield important diagnostic clues. Patients with peritonitis, for example, lie motionless, whereas those with renal colic may writhe in an attempt to find a comfortable position. Sometimes, certain foods exacerbate pain. A classic example is the relationship between the intake of fatty foods and the development of biliary pain. Pain associated with duodenal ulcer often is alleviated by meals. By contrast, patients with gastric ulcer or chronic mesenteric ischemia may report exacerbation of pain with eating. Patients often self-medicate to alleviate symptoms. A history of chronic antacid use or of nonsteroidal anti-inflammatory drug use, for example, may suggest the presence of peptic ulcer disease.

Associated Symptoms

Information regarding changes in constitutional symptoms (e.g., fever, chills, night sweats, weight loss, myalgias, arthralgias), digestive function (e.g., anorexia, nausea, vomiting, flatulence, diarrhea, constipation), jaundice, dysuria, changes in menstruation, and pregnancy should be solicited from the patient. A careful review of these symptoms may reveal important diagnostic information. For example, clear vomitus suggests gastric outlet obstruction, whereas fecu-

A careful review of the patient’s other medical problems often sheds light on the presentation of acute abdominal pain. Previous experience with similar symptoms suggests a recurrent problem. Patients with a history of partial small bowel obstruction, renal calculi, or pelvic inflammatory disease are likely to have recurrences. A patient whose presentation suggests intestinal obstruction, and who has no prior surgical history, deserves special attention because of the likelihood of surgical pathology, such as a hernia or malignancy. Patients with a systemic illness such as scleroderma, systemic lupus erythematosus, nephrotic syndrome, porphyria, or sickle cell disease often have abdominal pain as a manifestation of the underlying disorder. Abdominal pain also may arise as a side effect of a medication taken for another disease.

PHYSICAL EXAMINATION

The physical examination of the patient with acute abdom­ inal pain begins with an assessment of the patient’s appearance and airway, breathing, and circulation (ABC), as described earlier. The patient’s ability to converse, breathing pattern, position in bed, posture, degree of discomfort, and facial expression should be noted. A patient lying still in bed, in the fetal position and reluctant to move or speak, with a distressed facial expression, is likely to have peritonitis. On the other hand, a patient who writhes and changes position frequently likely has purely visceral pain, as in intestinal obstruction or gastroenteritis. Tachypnea may be a sign of metabolic acidosis caused by shock. Atrial fibrillation noted on physical examination or electro­ cardiogram may suggest mesenteric arterial embolus. All patients should undergo a careful, systematic examination, regardless of the differential diagnosis suggested by the history.

Abdominal Examination

Examination of the abdomen is central to the evaluation of a patient with acute abdominal pain and should begin with careful inspection. The entire abdomen, from the nipple line to the thighs, should be exposed. Obese patients should be asked whether the degree of protrusion of the abdominal wall is more than usual. Asthenic patients may feel distended but have relatively little apparent abdominal protrusion. Assessment for the presence of bowel sounds and their character should precede any maneuvers that will disturb the abdominal contents. Before concluding that an abdomen is silent, the examiner should listen for at least two minutes and in more than one quadrant of the abdomen. Experienced listeners will distinguish the high-pitched churning of a mechanical small intestinal obstruction from the more hollow sounds of toxic megacolon (like dripping in a cavern). The examiner should begin to palpate the abdomen with the head of the stethoscope while carefully watching the patient’s facial expression. If tenderness is detected, an assessment for rebound tenderness should be carried out next to look for evidence of peritonitis. Rebound tenderness may be elicited by jarring the patient’s bed or stretcher or by finger percussion. Palpation is performed next. If pain is emanating from one particular region, that area should be palpated last to detect involuntary guarding and muscular rigidity. Patients with a rigid abdomen rarely reveal any additional findings (such as a mass) on physical exam­ ination. Because these patients usually have a surgical

155

156

Section III  Symptoms, Signs, and Biopsychosocial Issues emergency, abdominal examination can be done more completely once the patient is under anesthesia, just before laparotomy.

Genital, Rectal, and Pelvic Examinations

The pelvic organs and external genitalia should be examined in every patient with acute abdominal pain. The rectum and vagina provide additional avenues for gentle palpation of pelvic viscera. Gynecologic pathology should be excluded in all women with acute abdominal pain.

LABORATORY DATA

The history and physical examination findings generally are not sufficient to establish a firm diagnosis in a patient with acute abdominal pain. All patients with acute abdominal pain should have a complete blood count, with a differential count, and urinalysis. The determination of serum electrolyte, blood urea nitrogen, creatinine, and glucose levels is useful for assessing the patient’s fluid and acid-base status, renal function, and metabolic state and should be done for every patient with acute abdominal pain who presents to an emergency department. Urine or serum pregnancy testing must be performed in all women of reproductive age with abdominal pain. Liver biochemical tests and serum amylase levels should be ordered for patients with upper abdominal pain or with jaundice. Leukocytosis, particularly when associated with immature band forms, is an important finding. Metabolic acidosis, an elevated serum lactate level, or a depressed bicarbonate level are all associated with tissue hypoperfusion and shock. Patients who manifest these findings are likely to require urgent surgical intervention or intensive care.

IMAGING STUDIES Computed Tomography

The development of high-speed helical computed tomog­ raphy (CT) scanning has revolutionized the evaluation of acute abdominal pain. In many conditions, such as appendicitis, CT scanning can almost eliminate diagnostic uncertainty. In the pre-CT era, history taking and physical examination alone had a specificity of approximately 80%; by contrast, the sensitivity and specificity of CT scanning for acute appendicitis are 94% and 95%, respectively.17 A negative CT scan in the setting of acute abdo­minal pain has considerable value in excluding common disorders. The question arises as to whether CT scanning should be a standard part of the evaluation in all patients with acute abdominal pain. Several arguments as to why CT should not be routine have been raised. First, CT scanning can be performed in a number of ways, and the most efficacious method must be chosen in any given clinical setting. For example, a patient with suspected renal colic should have a limited, non–contrast-enhanced, renal calculus protocol CT; obtaining a standard oral and intravenous contrast CT in this case may obfuscate rather than illuminate the pathology. Alternatively, a patient in whom arterial occlusive disease is suspected should undergo CT arteriography using a bolus intravenous contrast technique. A radiologist should be consulted regarding the selection of the most appropriate CT study in a given patient. Second, some diseases, such as acute cholecystitis and cholangitis, remain relatively invisible on CT. A patient with right upper quadrant pain who is suspected of having either of these diagnoses should undergo an ultrasound examination of the right upper quadrant as the primary diagnostic test. Third, as noted earlier, a patient who is unstable or exhibits signs of shock should be evaluated by a surgeon before any

imaging study is considered. In a patient with suspected trauma or hemoperitoneum, the focused abdominal sonogram for trauma (FAST; see later), which can be done at the bedside in the emergency department, is a preferable approach. The presence of shock and fluid in the abdomen is an indication for immediate laparotomy, and further diagnostic maneuvers, including CT, add little value to the patient’s care. A final consideration regarding the role of CT in the evaluation of acute abdominal pain is radiation exposure. Particularly for patients younger than 35 years and those who have required multiple examinations, abdominal CT may increase the lifetime risk of cancer.18 Additionally, unless a life-threatening condition is suspected, CT is best avoided in a pregnant patient, in whom ultrasound examination or magnetic resonance imaging (MRI) may provide a suitable alternative.

Ultrasonography

A FAST is a rapid, reliable, bedside test to detect fluid in the abdominal cavity. Although its main usefulness is for the evaluation of injured persons, this examination also aids in the diagnosis of any condition that results in free intraperitoneal fluid. Although not part of the formal FAST series, imaging of the aorta can be added, allowing a rapid assessment for aortic aneurysm. General beside ultrasound is likely to be used increasingly by nonradiologists in the future. A Swedish study has demonstrated that the diagnostic accuracy of emergency abdominal examinations by surgeons is increased significantly when an ultrasound examination is added to the evaluation.19

OTHER DIAGNOSTIC TESTS

Other diagnostic imaging modalities such as MRI and radionuclide scanning (e.g., 99mTc-labeled hydroxyl iminodiacetic acid [HIDA] scan) and endoscopy usually take a secondary role in the evaluation of the patient with acute abdominal pain. Use of these tests is generally guided by the results of CT or ultrasound. Angiography may be useful not only for establishing a diagnosis of visceral ischemia, but also for delivering therapy aimed at improving or reestablishing blood flow. Diagnostic peritoneal lavage, although seldom used now, is useful when a patient is too unstable from a cardiopulmonary standpoint to tolerate radiographic imaging. The finding of leukocytes in the lavage effluent in an unstable patient may, in extreme circumstances, constitute sufficient grounds for laparotomy. In a patient who is unstable and deteriorating and has signs of an acute abdomen, laparotomy as a diagnostic maneuver should be considered if imaging is considered prohibitively risky. An overall approach to the patient with acute abdominal pain is illustrated in Figure 10-5.

CAUSES Acute abdominal pain is usually defined as pain of less than one week in duration. Patients usually seek attention within the first 24 to 48 hours, although some may endure longer periods of abdominal discomfort. The most common reason for a patient to seek emergency department evaluation of abdominal pain is so-called nonspecific abdominal pain. Between 25% and 50% of all patients who visit an emergency department for abdominal pain will have no specific disease identified. The distribution of the causes of abdominal pain in patients who present to an emergency department is shown in Table 10-2.

Chapter 10  Acute Abdominal Pain Acute abdominal pain

Most likely diagnoses

Evaluation

ABC Prostration; hemodynamically unstable

Yes

No RLQ pain (gradual onset); RLQ tenderness, localized rebound tenderness

Yes

No Gradual onset of RUQ cramping pain; history of postprandial discomfort

Yes

Resuscitation Urgent surgical consultation Consider FAST examination Consider laparotomy

Perforated viscus Severe pancreatitis Ruptured spleen/ hemoperitoneum Ruptured AAA

Appendix protocol CT

Appendicitis*

In female patients, consider pelvic US or CT

Tubo-ovarian abscess Ovarian torsion Ectopic pregnancy Cholelithiasis Cholecystitis Bile duct obstruction Cholangitis

RUQ ultrasound

No Nausea, vomiting, obstipation, constipation, abdominal distention; prior surgery

Yes

Upright abdominal film or oral contrast CT scan

Small bowel obstruction

No Sudden onset, diffuse pain; involuntary guarding, rebound tenderness, peritonitis

Yes

Upright abdominal film or oral contrast CT scan

Perforated viscus Diverticulitis Mesenteric infarction Acute pancreatitis

Table 10-2  Causes of Acute Abdominal Pain in Patients Presenting to an Emergency Department CAUSE Nonspecific abdominal pain Appendicitis Bowel obstruction Urologic disease Biliary disease Diverticular disease Pancreatitis Medical illness Other

PATIENTS (%) 35 17 15 6 5 4 2 1 15

From Irvin TT. Causes of abdominal pain in 1190 patients admitted to a British Surgical Service. Br J Surg 1989; 76:1121-5.

ACUTE APPENDICITIS Acute appendicitis is a ubiquitous problem. In adult patients younger than 60 years, acute appendicitis accounts for 25% of admissions to the hospital from the emergency department for abdominal pain.20 The overall incidence of appendicitis is approximately 11/10,000 population, with a lifetime risk of 8.6% for men and 6.7% for women.21 Typically, acute appendicitis begins with prodromal symptoms of anorexia, nausea, and vague periumbilical pain. Within 6 to 8 hours, the pain migrates to the right lower quadrant and peritoneal signs develop. In uncomplicated appendi­ citis, a low-grade fever to 38°C and mild leukocytosis are usually present. A higher temperature and white blood cell

Figure 10-5.  An approach to the urgent evaluation of abdominal pain. Specific complaints and physical examination findings are coupled with appropriate radiologic imaging. AAA, abdominal aortic aneurysm; ABC, airway, breathing, circulation; CT, computed tomography; FAST, focused abdominal sonogram for trauma; RLQ, right lower quadrant; RUQ, right upper quadrant; US, ultrasound. *For left lower quadrant pain, the most likely diagnosis is diverticulitis.

count are associated with perforation and abscess formation. The mnemonic PANT can help the novice remember the classic progression of symptoms in appendicitis—pain followed by anorexia followed by nausea followed by temperature elevation. Uncommon presentations of acute appendicitis, however, are common, and the wary physician will not reject a diagnosis of acute appendicitis simply on the basis of the patient’s history and physical examination alone. Whereas plain abdominal radiographs are not diagnostic and have little role in the diagnosis of acute appendicitis, CT has dramatically improved the accuracy of diagnosis in patients with acute appendicitis. The finding of an appendiceal diameter larger than 6 mm has positive and negative predictive values of 98%.22 Other CT signs of acute appendicitis include periappendiceal fat inflammation, presence of fluid in the right lower quadrant, and failure of contrast dye to fill the appendix23; these findings have lower degrees of specificity. Traditionally, an erroneous diagnosis of appendicitis, reflected by the finding of normal pathology at surgical exploration, was as high as 33%.24 The addition of CT has reduced the false-negative rate to approximately 6% for men and 10% for women.25 As noted earlier, CT does entail radiation exposure,18 and some authorities advocate avoiding CT in children and adolescents,26 in whom a higher degree of diagnostic uncertainty is tolerated in favor of lower radiation exposure (see Chapter 116).

ACUTE BILIARY DISEASE

Biliary disease accounts for approximately 5% to 7% of emergency department visits for abdominal pain.2,20 Most

157

158

Section III  Symptoms, Signs, and Biopsychosocial Issues patients in this group present at some point on the spectrum between biliary pain and acute cholecystitis. Biliary pain is a syndrome of right upper quadrant or epigastric pain, usually postprandial, caused by transient obstruction of the cystic duct by a gallstone. Biliary pain is self-limited, generally lasting less than 6 hours. Acute cholecystitis is, in most cases, caused by persistent obstruction of the cystic duct by a gallstone. The pain of acute cholecystitis is almost indistinguishable from that of biliary pain, except that it is persistent. The pain usually is a dull ache and is localized to the right upper quadrant or epigastrium and may radiate around the back to the right scapula. Nausea, vomiting, and low-grade fever are common. On examination, right upper quadrant tenderness, guarding, and Murphy’s sign (inspiratory arrest on palpation of the right upper quadrant) are diagnostic of acute cholecystitis. The white blood cell count is usually mildly elevated but may be normal. Mild elevations in serum total bilirubin and alkaline phosphatase levels are typical. The role of gallstones in the etiology of biliary pain and acute cholecystitis makes ultrasound evaluation of the right upper quadrant the key diagnostic test. Demonstration of gallstones may suggest biliary pain, whereas the finding of stones with gallbladder wall thickening, pericholecystic fluid, and pain on compression of the gallbladder with the ultrasound probe (sonographic Murphy’s sign) is essentially diagnostic of acute cholecy­s­ titis, as is positive hepatobiliary scintigraphy (e.g., HIDA scan).27 Patients with acute cholecystitis are best managed with cholecystectomy within 48 hours.28-30 Patients who are diabetic, particularly those with a leukocyte count over 15,000/mm3, are at particular risk for gangrenous cholecystitis and should have immediate surgical consultation.31 These patients are likely to require an emergent open cholecystectomy. Patients who present with right upper quadrant pain with jaundice and signs of sepsis should be suspected of having obstruction of the bile duct by a gallstone. Right upper quadrant pain, fever and chills, and jaundice (Charcot’s triad) are suggestive of ascending cholangitis.32 These patients require intravenous fluids, antibiotics, and bile duct drainage, usually by endoscopy (see also Chapters 65, 66, and 67).

SMALL BOWEL OBSTRUCTION

Intestinal obstruction may occur in patients of all ages. In pediatric patients, intussusception, intestinal atresia, and meconium ileus are the most common causes. In adults, about 70% of cases are caused by postoperative adhesions; incarcerated hernias make up most of the remainder. Small bowel obstruction is characterized by sudden, sharp, periumbilical abdominal pain. Nausea and vomiting occur soon after the onset of pain and provide temporary relief of discomfort. Frequent bilious emesis with epigastric pain is suggestive of high (proximal) intestinal obstruction, whereas cramping periumbilical pain with infrequent feculent emesis is more typical of distal intestinal obstruction. Examination reveals an acutely ill, restless patient. Fever, tachycardia, and orthostatic hypotension are common. Abdominal distention is usual. Auscultation characteristically demonstrates hyperactive bowel sounds and audible rushes. The patient’s abdomen is diffusely tender to percussion and palpation, but peritoneal signs are absent, unless a complication such as ischemia or perforation has occurred. Leukocytosis and lactic acidosis suggest intestinal ischemia or infarction. Plain radiographs of the abdomen are diagnostic when they reveal dilated loops of small intestine with airfluid levels and decompressed distal small bowel and colon. Plain abdominal films can be misleading in a patient with

proximal jejunal obstruction, because dilated bowel loops and air-fluid levels may be absent. CT is superior for establishing the diagnosis and location of intestinal obstruction.33 In patients with partial small intestinal obstruction, initial treatment is with bowel rest, intravenous fluids, nasogastric decompression, and close observation. Surgery is required for patients who fail conservative management or have evidence of complete obstruction, especially if ischemia is suspected (see also Chapter 119).

ACUTE DIVERTICULITIS

Acute diverticulitis is a common disease. Approximately 80% of affected patients are older than 50 years,34 but the incidence may be increasing in younger persons.35 Patients with diverticulitis usually present with constant, dull, left lower quadrant pain and fever. They may complain of constipation or obstipation and usually are found to have a leukocytosis. Physical examination demonstrates left lower quadrant tenderness and, in some cases, a left lower quadrant mass. Localized peritoneal signs are frequent. In severe cases, generalized peritonitis may be present, making differentiation from other causes of a perforated viscus difficult. CT is reliable in confirming the diagnosis, with a sensitivity of 97%,36 and should be performed routinely in the emergency evaluation of patients with diverticulitis. Acute diverticulitis presents as a spectrum of disease from mild abdominal discomfort to gross fecal peritonitis, which is an acute surgical emergency. The severity of diverticulitis, as determined by CT, is best described using the Hinchey grading system (see Table 117-2).37 Patients with mild disease and no CT findings of perforation, in the absence of limiting comorbid disease, can generally be treated as an outpatient. Those with Hinchey grade I diverticulitis (localized pericolic abscess or inflammation) frequently require hospitalization for intravenous antibiotics. Patients with Hinchey grade II diverticulitis (pelvic, intraabdominal, or retroperitoneal abscess) should undergo CT-guided drainage of the abscess and receive a course of broad-spectrum intravenous antibiotics. Patients with Hinchey III (generalized purulent peritonitis) and IV (generalized fecal peritonitis) diverticulitis frequently require emergency surgery. Optimal surgical management of patients with Hinchey I or II diverticulitis is a matter of debate (see Chapter 117).

ACUTE PANCREATITIS

Hospital admissions for acute pancreatitis in the United States seem to be increasing. The incidence of acute pancreatitis in California rose from 33 to 43 cases/100,000 between 1994 and 2001.38 Acute pancreatitis typically begins as acute pain in the epigastrium that is constant, unrelenting, and frequently described as boring through to the back or left scapular region. Fever, anorexia, nausea, and vomiting are typical. Patients with pancreatitis usually are more comfortable sitting upright, leaning forward slightly, and are commonly found in this position in the emergency department. Physical examination reveals an acutely ill patient in considerable distress. Patients are usually tachycardic and tachypneic. Abdominal examination reveals hypoactive bowel sounds and marked tenderness to percussion and palpation in the epigastrium. Abdominal rigidity is a variable finding. In rare patients, flank or periumbilical ecchymoses (Grey-Turner’s or Cullen’s sign, respectively) develop in the setting of pancreatic necrosis with hemorrhage. Extremities are often cool and cyanotic, reflecting underperfusion. White blood cell counts of 12,000 to 20,000/mm3 are common. Elevated serum and urine amylase levels are usually present within the first few hours of

Chapter 10  Acute Abdominal Pain pain. Depending on the cause and severity of pancreatitis, serum electrolyte, calcium, and blood glucose levels and liver biochemical test and arterial blood gas results may be abnormal. Abdominal ultrasonography is useful for identifying gallstones as a potential cause of pancreatitis. CT is reserved for patients with severe or complicated pancreatitis. Although most cases of acute pancreatitis are self-limited, as many as 20% of patients have severe disease with local or systemic complications,39 including hypovolemia and shock, renal failure, liver failure, and hypocalcemia. Although a number of prognostic physiologic scales, such as the Sequential Organ Failure Assessment (SOFA) and Acute Physiologic Assessment and Chronic Health Evaluation (APACHE) II scores, have been advocated as measures of the severity of acute pancreatitis, the Ranson score, first published in 1974, remains a useful and widely used checklist for the early assessment of patients with acute pancreatitis.40 The Ranson score consists of five early and six late factors that indicate severe pancreatitis (see Table 58-2). A minority of patients with severe acute pancreatitis present with a profound intra-abdominal catastrophe, usually caused by thrombosis of the middle colic artery or right colic artery, which travels in proximity to the head of the pancreas, with resulting colonic infarction. This process may not be seen clearly on CT scans obtained early in the course of disease and should be suspected in any case marked by rapid hemodynamic collapse. Such patients require immediate laparotomy (see Chapter 58).

PERFORATED PEPTIC ULCER

The epidemiology of peptic ulcer disease continues to change. The overall incidence of peptic ulcer disease has declined significantly since the late 1970s,41,42 and the number of patients requiring hospital admission for severe and complicated peptic ulcer disease has also decreased.42 Better therapeutic modalities, including proton pump inhibitors, eradication of Helicobacter pylori, and endoscopic methods for control of hemorrhage, have reduced the number of patients with peptic ulcer disease who require surgical intervention,43 although the incidence of complicated disease has increased in older adults, in whom morbidity and mortality related to surgery are also increased.42 Patients with a perforated peptic ulcer typically present with the sudden onset of severe diffuse abdominal pain. These patients may be able to specify the precise moment of the onset of symptoms. In the usual case, the afflicted patient presents acutely with excruciating abdominal pain. Abdominal examination reveals peritonitis, with rebound tenderness, guarding, or abdominal muscular rigidity. In such cases, distinguishing perforated ulcer from other causes of a perforated viscus, such as a perforated colonic diverticulum or perforated appendicitis, may not be possible. Older or debilitated patients may present with less dramatic symptoms, with perforation detected by the presence of free intraperitoneal air on an upright abdominal film or CT scan. A perforated peptic ulcer should be suspected in any patient with the sudden onset of severe abdominal pain who presents with abdominal rigidity and free intraperitoneal air. Pneumoperitoneum is identified on an abdominal radiograph in 75% of patients (Fig. 10-6). In equivocal cases, CT of the abdomen usually suggests the diagnosis by demonstrating edema in the region of the gastric antrum and duodenum, associated with extraluminal air. CT may not be diagnostic, however, and patients with diffuse peritonitis or hemodynamic collapse should be explored surgically. Lapa-

Figure 10-6.  This upright chest film of an 80-year-old man with the acute onset of severe epigastric pain demonstrates free intra-abdominal air under the right hemidiaphragm. The patient has pneumoperitoneum as a result of a perforated viscus. At surgery, an anterior duodenal ulcer perforation was found.

rotomy is acceptable as the primary diagnostic maneuver in such patients. Endoscopy is not advisable when the diagnosis of a perforated peptic ulcer is suspected. Insufflation of the stomach can convert a sealed perforation into a free perforation. Survival following emergency surgery for complications of peptic ulcer disease is surprisingly poor. Patients who require surgery for a complication of peptic ulcer disease are generally older and more medically ill than those seen in the past. Sarosi and colleagues have reported a 23% in-hospital mortality rate in a Veterans Administration population,44 and Imhof and associates,45 reporting on a series of German patients with perforated peptic ulcer, found an in-hospital mortality rate of 12.1%, a one-year mortality rate of 28.7%, and a five-year mortality rate of 46.8% (see also Chapters 52 and 53).

ACUTE MESENTERIC ISCHEMIA

Acute mesenteric ischemia can result from occlusion of a mesenteric vessel arising from an embolus, which may emanate from an atheroma of the aorta or cardiac mural thrombus, or from primary thrombosis of a mesenteric vessel, usually at a site of atherosclerotic stenosis. Embolic occlusion is more common in the superior mesenteric artery than the celiac or inferior mesenteric artery, presumably because of the less acute angle of the superior mesenteric artery off the abdominal aorta. Nonocclusive mesenteric ischemia results from inadequate visceral perfusion and can also lead to intestinal ischemia and infarction. Such cases are usually consequent to catastrophic systemic illnesses such as cardiogenic or septic shock. Acute mesenteric embolism, mesenteric thrombosis, and nonocclusive mesenteric ischemia each account for approximately one third of cases of acute mesenteric ischemia and have a combined mortality rate of 60% to 100%.46 The hallmark of the diagnosis of acute mesenteric ischemia is the abrupt onset of intense cramping epigastric and periumbilical pain out of proportion to the findings on abdominal examination. Other symptoms may include diarrhea, vomiting, bloating, and melena. On physical examination, most patients appear acutely ill, but the presentation may be subtle. Shock is present in about 25% of cases.

159

160

Section III  Symptoms, Signs, and Biopsychosocial Issues CT is the best initial diagnostic test. Mesenteric angio­g­ raphy may be useful for determining the cause of intestinal ischemia and defining the extent of vascular disease. Patients with acute embolic or thrombotic intestinal ische­ mia should be referred for immediate revascularization and bowel resection.47 Patients with nonocclusive mesenteric ischemia are best managed by treatment of the underlying shock state. For those with persistent symptoms, laparotomy for resection of infarcted intestine may be necessary. Transcatheter vasodilator therapy may be helpful for patients who are found to have vasospasm on visceral arteriography (see also Chapter 114).47

ABDOMINAL AORTIC ANEURYSM

Rupture of an abdominal aortic aneurysm is heralded by the sudden onset of acute, severe abdominal pain localized to the midabdomen or paravertebral or flank areas. The pain is tearing in nature and associated with prostration, lightheadedness, and diaphoresis. If the patient survives transit to the hospital, shock is the most common presen­ tation. Physical examination reveals a pulsatile, tender abdominal mass in about 90% of cases. The classic triad of hypotension, a pulsatile mass, and abdominal pain is present in 75% of cases and mandates immediate surgical intervention.47

ABDOMINAL COMPARTMENT SYNDROME

Although not usually presenting as acute abdominal pain, abdominal compartment syndrome (ACS) warrants consideration in any patient with an abdominal emergency. First reported in the setting of massive intra-abdominal trauma, ACS, defined as pathologic elevation of intraabdominal pressure, is now recognized as a frequent complication of many severe disease processes. An elevated intra-abdominal pressure may develop in a patient who survives massive volume resuscitation with resulting visceral edema or who has a disease such as severe pancreatitis that can cause visceral or retroperitoneal edema. The elevated intra-abdominal pressure in turn compromises visceral perfusion, with resulting injury and additional edema. The kidney is particularly prone to underperfusion in this setting, and kidney failure may be the first sign of ACS.48 Intra-abdominal pressure can be measured simply by connecting a transducer to a urinary catheter, with the zero reference point at the midaxillary line in a supine patient. The World Society for Abdominal Compartment Syndrome has established a consensus grading scheme for ACS based on the measured bladder pressure. A normal value for bladder pressure is less than 7 mm Hg. Grade I ACS is defined as a pressure of 12 to 15 mm Hg, grade II as 16 to 20 mm Hg, grade III as 21 to 25 mm Hg, and grade IV as greater than 25 mm Hg. Nonsurgical options for treating low-grade ACS include gastric decompression, sedation, neuromuscular blockade, placing the patient in a reverse Trendelenburg position while allowing the hips to remain in a neutral position, and diuretics. In a patient with highgrade ACS, particularly when renal and respiratory function is com­promised, laparotomy and creation of an open abdomen is most effective. Management of the open abdomen requires specific surgical expertise usually found in referral medical centers.49

OTHER INTRA-ABDOMINAL CAUSES

Other intra-abdominal causes of acute abdominal pain include the following: gynecologic conditions such as endometritis, acute salpingitis with or without tubo-ovarian

abscess, ovarian cysts or torsion, and ectopic pregnancy; spontaneous bacterial peritonitis (Chapter 91); functional dyspepsia and peptic ulcer disease (Chapters 13 and 52); infectious gastroenteritis (Chapters 107 and 108); viral hepatitis and other liver infections (Chapters 77 to 82); pyelonephritis; cystitis; mesenteric lymphadenitis; inflammatory bowel disease (Chapters 111 and 112); and functional abnormalities such as irritable bowel syndrome (Chapter 118) and intestinal pseudo-obstruction (Chapter 120).

EXTRA-ABDOMINAL CAUSES

Acute abdominal pain may arise from disorders involving extra-abdominal organs and systemic illnesses. Examples are listed in Table 10-3. Surgical intervention for patients with acute abdominal pain arising from an extra-abdominal or systemic illness is seldom required except in cases of pneumothorax, empyema, and esophageal perforation. Esophageal perforation may be iatrogenic, result from blunt or penetrating trauma, or occur spontaneously (Boerhaave’s syndrome; see also Chapter 45).

Table 10-3  Extra-Abdominal Causes of Acute Abdominal Pain Cardiac Congestive heart failure Endocarditis Myocardial ischemia and infarction Myocarditis Thoracic Esophageal rupture (Boerhaave’s syndrome) Esophageal spasm Empyema Esophagitis Pleurodynia (Bornholm’s disease) Pneumonitis Pneumothorax Pulmonary embolism and infarction Hematologic Acute leukemia Hemolytic anemia Henoch-Schönlein purpura Sickle cell anemia Metabolic Acute adrenal insufficiency (Addison’s disease) Diabetes mellitus (especially with ketoacidosis) Hyperlipidemia Hyperparathyroidism Porphyria Uremia Toxins Hypersensitivity reactions (e.g., to insect bites, reptile venoms) Lead poisoning Infections Herpes zoster Osteomyelitis Typhoid fever Neurologic Abdominal epilepsy Radiculopathy, spinal cord or peripheral nerve tumors, degenerative arthritis of spine, herniated vertebral disk Tabes dorsalis Miscellaneous Familial Mediterranean fever Heat stroke Muscle contusion, hematoma, tumor Narcotic withdrawal Psychiatric disorders

Chapter 10  Acute Abdominal Pain SPECIAL CIRCUMSTANCES Extremes of Age

Evaluation of acute abdominal pain in patients at the extremes of age is a challenge. Historical information and physical examination findings are often difficult to elicit or are unreliable. Similarly, laboratory data may be misleadingly normal in the face of serious intra-abdominal pathology. For these reasons, patients at the extremes of age often are diagnosed late in the course of the disease, thereby resulting in increased morbidity. For example, the perforation rate for appendicitis in the general population averages 10% but exceeds 50% in infants. A carefully obtained history, thorough physical examination, and high index of suspicion are the most useful diagnostic aids. The occurrence of acute abdominal conditions is highly variable in these populations, and a high index of suspicion is required. In the pediatric population, the causes of acute abdominal pain vary with age. In infancy, intussusception, pyelonephritis, gastroesophageal reflux, Meckel’s diverticulitis, and bacterial or viral enteritis are common. In children, Meckel’s diverticulitis, cystitis, pneumonitis, enteritis, mesenteric lymphadenitis, and inflammatory bowel disease are prevalent. In adolescents, pelvic inflammatory disease, inflammatory bowel disease, and the common adult causes of acute abdominal pain predominate. In children of all ages, two of the most common causes of pain are acute appendicitis and abdominal trauma secondary to child abuse. In the older adult population, biliary tract disease accounts for almost 25% of cases of acute abdominal pain and is followed in frequency by nonspecific abdominal pain, malignancy, intestinal obstruction, complicated peptic ulcer disease, and incarcerated hernia. Appendicitis, although rare in older patients, usually manifests late in its course and is associated with high morbidity and mortality rates.

Pregnancy

The gravid woman with acute abdominal pain presents a difficult diagnostic dilemma. Pregnant women develop acute appendicitis and cholecystitis at the same rate as their nonpregnant counterparts. A number of additional diag­ noses, such as placental abruption and pain related to tension on the broad ligament, must be distinguished from nonobstetric diagnoses. Furthermore, the risk of radiation injury to the developing fetus must be considered when imaging studies are planned. Surgery in pregnancy is not rare; approximately 1 in 500 pregnancies will be associated with a nonobstetric general surgical intervention.50 Primary consideration is given to the health of the mother. The middle three months of gestation are optimal for abdominal surgical intervention, because this period presents the lowest risk for teratogenicity and spontaneous labor. Emergency interventions during pregnancy carry a risk of fetal loss that varies with the type of intervention and the age of gestation. Appendicitis occurs in approximately 1 in 2000 pregnancies and is equally distributed among the three trimesters. In later stages of pregnancy, the appendix may be displaced cephalad, with consequent displacement of the signs of peritoneal irritation away from McBurney’s point. Ultrasound or, in challenging cases, MRI may be useful for establishing a diagnosis in this setting. Biliary tract disease is also common during pregnancy. Open or laparoscopic management of these diseases is safe but is associated with a rate of preterm delivery of approximately 12% for appendectomy and 11% for cholecystectomy.51

Immunocompromised Hosts

In addition to diseases that occur in the general population, such as appendicitis and cholecystitis, a number of diseases unique to immunocompromised hosts may manifest with acute abdominal pain, including neutropenic enterocolitis, drug-induced pancreatitis, graft-versus-host disease, pneumatosis intestinalis, and cytomegalovirus (CMV) and fungal infections. Patients infected with human immunodeficiency virus (HIV) can present a particular challenge. When advanced, HIV infection is associated with a number of other diseases that may present as acute abdominal pain. One of the most common abdominal disorders seen in immunocompromised persons in the developing world is primary peritonitis. Affected patients have suppurative peritonitis without a definable source. Spontaneous intes­ tinal perforation, usually secondary to CMV infection, is also common in patients with advanced HIV infection. Tuberculous peritonitis is a consideration in patients from areas in which tuberculosis is common.52 In general, immunocompromised patients may lack the definitive signs of an acute abdominal crisis usually seen in immunocompetent persons; an elevated temperature, peritoneal signs, and leukocytosis may be absent in these cases.

PHARMACOLOGIC MANAGEMENT An unfortunate practice in the care of patients with acute abdominal pain is the delay in administration of narcotics pending definitive surgical assessment. Sir Zachary Cope stated that “Morphine does little or nothing to stop serious intra-abdominal disease, but it puts an efficient screen in front of the symptoms.”53 The practice of delaying relief of pain in a suffering patient, however, does not appear to withstand careful clinical scrutiny. Six studies in which the early administration of analgesia was compared with administration of placebo in patients with acute abdominal pain have shown that the patients who receive analgesics are more comfortable and do not experience a delay in diagnosis.54 Patients with acute abdominal pain frequently are suffering the most intense pain that they have ever experienced and should receive appropriate opioid analgesics early in their care. Patients with acute abdominal processes frequently require antibiotic treatment for peritonitis. When appro­ priate, antibiotic therapy aimed at the likely causative pathogens should be given as soon as a putative diagnosis is reached; little benefit is derived from treating an immunocompetent patient with broad-spectrum antibiotics before a likely source is identified. Patients who are immunocompromised or neutropenic are an exception to this rule. They should receive broad-spectrum antibiotics early in the course of management for acute abdominal pain (see Chapters 37 and 91).

KEY REFERENCES

An G, West M. Abdominal compartment syndrome: A concise clinical review. Crit Care Med 2008; 36:1304-10. (Ref 49.) Addiss DG, Shaffer N, Fowler B, Tauxe RV. The epidemiology of appendicitis and appendectomy in the United States. Am J Epidemiol 1990; 132:910-25. (Ref 21.) Birnbaum BA, Wilson SR. Appendicitis at the millennium. Radiology 2000; 215:337-48. (Ref 23.) Bohner H, Yang Q, Franke C, et al. Simple data from history and phy­ sical examination help to exclude bowel obstruction and to avoid radiographic studies in patients with acute abdominal pain. Eur J Surg 1998; 164:777-84. (Ref 2.)

161

162

Section III  Symptoms, Signs, and Biopsychosocial Issues Cotton M. The acute abdomen and HIV. Trop Doct 2006; 36:198-200. (Ref 52.) Diaz JJ, Bokhari F, Mowery NT, et al. Guidelines for management of small bowel obstruction. J Trauma 2008; 64:1651-64. (Ref 33.) Frossard JL, Steer ML, Pastor CM. Acute pancreatitis. Lancet 2008; 371:143-52. (Ref 39.) Jacobs DO. Diverticulitis. N Engl J Med 2007; 357:2057-66. (Ref 34.) Maerz L, Kaplan LJ. Abdominal compartment syndrome. Crit Care Med 2008; 36:S212-15. (Ref 48.) Manterola C, Asutdillo P, Losada H, et al. Analgesia in patients with acute abdominal pain. Cochrane Database Syst Rev 2007; (3):CD005660. (Ref 54.) McGory ML, Zingmond DS, Nanayakkara D, et al. Negative appendectomy rate: Influence of CT scans. Am Surg 2005; 71:803-8. (Ref 25.)

Parangi S, Levine D, Henry A, et al. Surgical gastrointestinal disorders during pregnancy. Am J Surg 2007; 193:223-32. (Ref 50.) Peng WK, Sheikh Z, Nixon SJ, Paterson-Brown S. Role of laparoscopic cholecystectomy in the early management of acute gallbladder disease. Br J Surg 2005; 92:586-91. (Ref 30.) Silen W. Cope’s early diagnosis of the acute abdomen, 18th ed. New York: Oxford University Press; 1991. p 301. (Ref 16.) Terasawa T, Blackmore CC, Brent S, Kohlwes RJ. Systematic review: Computed tomography and ultrasonography to detect acute appen­ dicitis in adults and adolescents. Ann Intern Med 2004; 141:537-46. (Ref 17.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

11 Chronic Abdominal Pain Joseph C. Yarze and Lawrence S. Friedman

CHAPTER OUTLINE Definition and Clinical Approach  163 Abdominal Wall Pain  164 Anterior Cutaneous Nerve Entrapment and Myofascial Pain Syndromes  164 Slipping Rib Syndrome  165 Thoracic Nerve Radiculopathy  165 Functional Abdominal Pain Syndrome  165 Epidemiology  165

The evaluation of any patient with a complaint of abdom­ inal pain is challenging. Abdominal pain can be benign and self-limited or a harbinger of a serious life-threatening disease (see Chapter 10). Chronic abdominal pain poses a particularly challenging clinical problem. Not only is the management of chronic abdominal pain a frequently daunt­ ing task, but also the possibility of overlooking a structural or organic disorder is always a concern. Many disorders discussed elsewhere in this text can produce chronic abdominal pain (Table 11-1). Many of these diagnoses require careful consideration and clinical interrogation, in addition to appropriate diagnostic testing, to discern whether the entity is indeed the cause of the patient’s pain. Diagnosis of a functional gastrointestinal disorder is gener­ ally considered once potential causes of organic chronic abdominal pain have been confidently excluded. Although the causes of chronic abdominal pain are varied, the pathophysiologic pathways that produce chronic pain are common to many of them. This chapter focuses on the neuromuscular causes of chronic abdominal pain and the functional abdominal pain syndrome (FAPS). FAPS serves as a model to illustrate many of the complex issues involved in caring for patients with chronic abdominal pain.

DEFINITION AND CLINICAL APPROACH Abdominal pain is considered chronic when it has been occurring constantly or intermittently over at least six months. Abdominal pain is considered acute when it has been occurring for several days and subacute when it has been occurring more than several days but less than 6 months. These arbitrary definitions are often helpful when formulating a differential diagnosis. The clinician initially must adopt a broad-based approach, which necessarily becomes more focused as the evaluation ensues. Impor­ tantly, although typical patterns of presentation are useful to remember, some patients, especially immunosuppressed and older persons, may present with atypical features. As for acute abdominal pain (see Chapter 10), the initial step in evaluating a patient with chronic abdominal pain is

Pathophysiology  165 Clinical Features  168 Diagnosis and Differential Diagnosis  169 Treatment  169 Role of Laparoscopy with Lysis of Adhesions  170

to elicit a detailed history from the patient. The chronology of the pain, including its abruptness of onset and duration, and its location and possible radiation should be deter­ mined. Visceral pain emanating from the digestive tract is perceived in the midline, because of the relatively sym­ metrical bilateral innervation of the organs, but is diffuse and poorly localized.1 Referred pain is ordinarily located in the cutaneous dermatomes that share the same spinal cord level as the affected visceral inputs.2 The patient should be questioned about the intensity and character of the pain, with the understanding that these parameters are subjective. The patient’s perception of precipitating, exacerbating, or mitigating factors may be useful when diagnostic possibil­ ities are considered. When initially attempting to determine whether the patient’s pain is caused by an organic or functional process, the clinician should search for clues in the patient’s history and physical examination that support or refute the diagnosis of a progressive, serious, chronic underlying illness. Such features in the history include fever, night sweats, appetite change, weight loss, and nocturnal awakening. A complete physical examination is indicated to look for evidence of a systemic disease. The abdominal examination should use a combination of inspection, auscultation, per­ cussion, and palpation. The most critical step for a patient with an acute exacerbation of chronic abdominal pain is to ascertain promptly whether a surgical abdomen is present (see Chapter 10). Although most causes of chronic abdomi­ nal pain do not require immediate surgical treatment, a complication related to a disease process ordinarily asso­ ciated with chronic abdominal pain may present acutely (e.g., intestinal perforation in a patient with inflammatory bowel disease). Furthermore, a patient who has experienced chronic abdominal pain may present with acute pain related to another disease process (e.g., acute mesenteric ischemia in a patient with underlying irritable bowel syndrome [IBS]). The abdomen should be auscultated to detect an abdominal bruit, because the presence of a bruit may suggest chronic mesenteric ischemia (intestinal angina). Abdominal palpation for the presence of organomegaly, masses, and ascites and examination for hernias are particularly per­ tinent. Other physical findings that suggest an underlying

163

164

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 11-1  Differential Diagnosis of Chronic or Recurrent Abdominal Pain Structural (or Organic) Disorders Inflammatory Appendicitis (Chapter 116) Celiac disease (Chapter 104) Eosinophilic gastroenteritis (Chapter 27) Fibrosing mesenteritis (mesenteric panniculitis) (Chapter 37) Inflammatory bowel diseases (Chapters 111 and 112) Pelvic inflammatory diseases Primary sclerosing cholangitis (Chapter 68) Vascular Celiac artery syndrome (Chapter 36) Mesenteric ischemia (Chapter 114) Superior mesenteric artery syndrome (Chapter 14) Metabolic Diabetic neuropathy Familial Mediterranean fever (Chapter 35) Hereditary angioedema Porphyria (Chapter 76) Neuromuscular Anterior cutaneous nerve entrapment syndrome Myofascial pain syndrome Slipping rib syndrome Thoracic nerve radiculopathy Other Abdominal adhesions (Chapter 119) Abdominal neoplasms (Chapters 29-32, 46, 54, 60, 69, 94, 121-123) Anaphylaxis (Chapter 9) Chronic pancreatitis (Chapter 59) Endometriosis (Chapter 124) Gallstones (Chapter 65) Hernias (Chapter 24) Intestinal malrotation (Chapter 96) Intestinal obstruction (Chapter 119) Lactose intolerance (Chapter 101) Peptic ulcer disease (Chapter 52) Small intestinal and pelvic lipomatosis (Chapter 37) Functional Gastrointestinal Disorders Biliary pain (gallbladder or sphincter of Oddi dysfunction)  (Chapter 63) Functional abdominal pain syndrome Functional (nonulcer) dyspepsia (Chapter 13) Gastroparesis (Chapter 48) Irritable bowel syndrome (Chapter 118) Levator ani syndrome (Chapter 125)

organic illness include signs of malnutrition (e.g., muscle wasting or edema), vitamin deficiencies, or extraintestinal processes (e.g., arthropathy or skin changes). Although not entirely specific, the closed eyes sign is often seen in patients with FAPS (see later). Similarly, Carnett’s sign and the hover sign (described later) may be seen in persons with abdominal wall pain. The laboratory evaluation can be helpful, but the clinician must first distill pertinent facets of the history and physical examination to focus the laboratory assessment. Injudicious use of laboratory testing is costly and can confuse the clinical picture and even lead to complications. It is worth emphasizing that an abnormal laboratory test result does not necessarily prove causality in relation to a patient’s chronic pain syndrome. The clinician must exercise the utmost discretion when ordering and interpreting the results of laboratory tests. Endoscopic and imaging studies have important roles in diagnosing and excluding many causes of chronic abdom­ inal pain. Upper endoscopy and colonoscopy, as well as capsule endoscopy, may be indicated in selected cases. Available imaging investigations include barium and radio­

nuclide studies, ultrasonography, computed tomography, magnetic resonance imaging, positron emission tomography (PET), and conventional angiography. The indications for each of these radiologic investigations differ, as do their potential to clarify an individual clinical situation. Endo­ scopic and radiologic testing in specific disorders is dis­ cussed in detail elsewhere in this text.

ABDOMINAL WALL PAIN ANTERIOR CUTANEOUS NERVE ENTRAPMENT AND MYOFASCIAL PAIN SYNDROMES

Anterior cutaneous nerve entrapment syndrome (ACNES) and myofascial pain syndrome (MFPS) are common causes of chronic abdominal wall pain. These syndromes share clinical, diagnostic, and treatment characteristics. The importance of recognizing these syndromes rests in provid­ ing the patient with an accurate diagnosis and effective treatment, as well as avoiding further expensive investiga­ tion and unnecessary surgical intervention. The abdominal wall should be suspected as the cause of symptoms when there is a complaint of chronic and unremitting abdominal pain that is unrelated to eating or bowel function but clearly related to movement. Although ACNES was initially described in the 1970s, it remains a frequently overlooked cause of chronic abdom­ inal pain.3,4 In ACNES, the pain is believed to occur when there is entrapment of a cutaneous branch of a sensory nerve that is derived from a neurovascular bundle emanating from spinal levels T7 to T12. The nerve entrapment may be related to pressure from an intra- or extra-abdominal lesion or to another localized process, such as fibrosis or edema. Pain emanating from the abdominal wall is discrete and localized, in contrast to pain originating from an intraabdominal source, which is diffuse and poorly localized. Patients usually point to the location of their pain with one finger, and the examiner can often localize the area of maximal tenderness to a region less than 2 cm in diameter. During physical examination, the patient often guards the affected area from the examiner’s hands (hover sign).5 Patients often note that activities associated with tightening of the abdominal musculature are associated with an exac­ erbation of pain and, during physical examination, the clini­ cian will note increased localized tenderness to palpation when the patient tenses the abdominal muscles (Carnett’s sign).6 In contrast, an increase in tenderness during relax­ ation of the abdominal musculature suggests an intraabdominal source of pain. In MFPS, pain emanates from myofascial trigger points in skeletal muscle.7 Causative factors include musculo­ skeletal trauma, vertebral column disease, intervertebral disc disease, osteoarthritis, overuse, psychological distress, and relative immobility. The exact pathophysiology of pain in MFPS remains unclear. Chronic abdominal wall pain may occur in patients with MFPS. Pain may be referred from another site, and the identification of trigger points (including those remote from the site of pain) is a useful physical finding. When attempting to identify a trigger point, the examiner uses a single finger to palpate a tender area. This is most often located in the central portion of a muscle belly, which may feel indurated or taut to palpation, and elicits a jump sign.8 This finding refers to a patient’s response by wincing, jerking away, or crying out as the myofascial trigger point is detected. Less commonly, trigger points may be located at sites such as the xiphoid process, costochondral junctions, or ligamentous and tendinous insertions.

Chapter 11  Chronic Abdominal Pain Treatment of ACNES and MFPS, when successful, not only improves symptoms, but also confirms the diagno­ sis.9,10 The treatment strategy depends on the severity of the symptoms. With mild and intermittent symptoms that are reproducibly precipitated by certain movements, simple reassurance and a recommendation to avoid such move­ ments may suffice. Non-narcotic analgesics, nonsteroidal anti-inflammatory drugs (NSAIDs), and application of heat can be used during exacerbations. Physical therapy may be beneficial, although no randomized studies have supported this treatment modality. For severe and persistent symp­ toms, injection therapy with a local anesthetic, with or without a glucocorticoid, is recommended.9-11 In one study of 136 patients in whom the history and physical examina­ tion suggested abdominal wall pain, and in whom benefit was noted with injection therapy, the diagnosis remained unchanged after a mean follow-up of four years in 97% of cases.12 In carefully selected patients with symptoms refrac­ tory to injection therapy, a prospective nonrandomized investigation has suggested that diagnostic laparoscopy with open exploration of abdominal trigger points may be beneficial.13 In this study, after intra-abdominal adhesions in close proximity to trigger points were lysed, subcutane­ ous nerve resection was performed. After a median postop­ erative follow-up of 37 months, 23 of 24 patients (96%) believed that this approach was beneficial in managing their previously intractable pain.

SLIPPING RIB SYNDROME

The slipping rib syndrome (SRS), which was described ini­ tially in the early 20th century,14,15 is an uncommonly rec­ ognized cause of chronic lower chest and upper abdominal pain. SRS ordinarily causes unilateral, sharp, often lancinat­ ing pain in the subcostal region. The acute pain may be followed by a more protracted aching sensation. The syn­ drome is associated with hypermobility of the costal car­ tilage at the anterior end of a false rib (rib 8, 9, or 10), with slipping of the affected rib behind the superior adjacent rib during contraction of the abdominal musculature. This slipping causes pain by a variety of potential mechanisms, including costal nerve impingement and localized tissue inflammation. The key to diagnosis is clinical awareness of the syndrome, in conjunction with use of the hooking maneuver; the clinician hooks his or her examining fingers underneath the patient’s lowest rib and, as the rib is moved anteriorly, the pain is reproduced and an audible pop or click is often heard.16 Conservative therapeutic measures often suffice but, on occasion, costochondral nerve block­ ade, response to which supports the diagnosis, or even surgical rib resection is required.17

THORACIC NERVE RADICULOPATHY

Disease related to thoracic nerve roots T7 through T12 may be responsible for abdominal pain. The disease processes that may cause this problem include neuropathy related to back and spine disorders, diabetes mellitus, and herpes zoster infection.18,19 Obtaining a complete history and per­ forming a careful physical examination of the patient, with attention to the possibility of a systemic disease and abnor­ mal neurologic and dermatologic findings, should lead to the correct diagnosis in most instances. Treatment depends on the specific underlying disease process.

FUNCTIONAL ABDOMINAL PAIN SYNDROME FAPS is a distinct medical disorder. Evidence suggests that the syndrome relates to central nervous system (CNS)

Table 11-2  Rome III Criteria for Functional Abdominal Pain Syndrome* Must include all the following: 1. Continuous or almost continuous abdominal pain 2. No or only occasional relationship of pain with physiologic events (e.g., eating, defecation, menses) 3. Some loss of daily functioning 4. Pain is not feigned (e.g., no malingering) 5. Insufficient symptoms to meet criteria for another functional gastrointestinal disorder that would explain the pain *Criteria fulfilled for the past three months with symptom onset at least six months prior to diagnosis.

amplification of normal regulatory visceral signals, rather than functional abnormalities in the gastrointestinal tract.20,21 The disorder is characterized by continuous, almost con­ tinuous, or at least frequently recurrent abdominal pain that is poorly related to bowel habits and often not well local­ ized. FAPS is properly understood as abnormal perception of normal (regulatory) bowel function rather than a motility disorder. The syndrome appears to be closely related to alterations in endogenous pain modulation systems, includ­ ing dysfunction of descending and cortical pain modulation circuits.21 The Rome III diagnostic criteria for FAPS are shown in Table 11-2.20,21 Studies that included patients who meet diagnostic criteria for FAPS have revealed that only rarely is an organic cause of chronic abdominal pain found during long-term follow-up.22,23 FAPS is commonly associated with other unpleasant somatic symptoms, and, when it persists or dominates the patient’s life, it usually is associated with chronic pain behaviors and comorbid psychological disturbances.24 Patients with FAPS typically define their illness as medical, and their symptoms tend to be more severe and associated with greater functional impairment than those of patients with IBS.24 Psychological disturbances, if present, must be considered as comorbid features of FAPS rather than as part of a primarily psychiatric problem.25 When compared with patients who have chronic back pain, those with chronic abdominal pain report significantly better physical func­ tioning, yet their overall perception of health is significantly worse.26

EPIDEMIOLOGY

Although the epidemiology of FAPS is incompletely known, in the U.S. Householder Survey of Functional Gastrointestinal Disorders, FAPS was estimated to be present in 2% of the sample and was less frequent than IBS (9%).27 A female predominance was noted (F:M = 1.5). Patients with FAPS missed more work days because of illness and had more physician visits than those without abdominal symptoms. A substantial proportion of patients are referred to gastroenterology practices and medical centers; they have a disproportionate number of health care visits and often undergo numerous diagnostic procedures and treatments.

PATHOPHYSIOLOGY

Chronic pain is a multidimensional (sensory, emotional, cognitive) experience explained by abnormalities in neuro­ physiologic functioning at the afferent, spinal, and CNS levels. Unlike acute pain arising from peripheral or visceral injury or disease, chronic functional pain is not associated with increased afferent visceral stimuli from structural abnormalities and tissue damage. FAPS is considered what is termed a biopsychosocial disorder related to dysfunction

165

166

Section III  Symptoms, Signs, and Biopsychosocial Issues MCC

Early life Genetics Environment

Primary somatosensory cortex

Limbic system

Thalamus

ACC Insula

Psychosocial factors Life stress Psychologic state Coping Social support CNS

Reticulothalamic tract Spinothalamic tract

Outcome

ENS

Medications Health provider visits Daily function Quality of life

Physiology Motility Sensation

FAPS Symptom experience Behavior Figure 11-1.  Biopsychosocial model of functional abdominal pain syndrome (FAPS). Consistent with a biopsychosocial model of illness, a person may be predisposed to FAPS because of factors (e.g., genetic, environmental) in early life. The patient’s symptoms and behavioral responses result from the interaction between psychosocial factors (e.g., life stress, social support) and gastrointestinal physiology (i.e., motility and sensation). FAPS relates to dysfunction in the brain-gut neuraxis with abnormal modulation of afferent signals from the gut that influences the symptoms experienced by the patient and that leads to increased use of health care resources and reduced quality of life. CNS, central nervous system; ENS, enteric nervous system.

of the brain-gut axis (see Chapter 21).25 As shown in Figure 11-1, the clinical expression of FAPS is derived from psy­ chological and intestinal physiologic input that interacts via the CNS-gut neuraxis. This model integrates the clinical, physiologic, and psychosocial features of FAPS into a com­ prehensible form, providing the basis for understanding psychological influences and application of psychopharma­ cologic treatments. Research relating to the pathophysiology of painful func­ tional gastrointestinal disorders has focused on the concepts of visceral hypersensitivity and alterations of brain-gut interactions. Visceral hypersensitivity is facilitated by upregulation of mucosal nociceptors and sensitization of vis­ ceral afferent nerves.28 Dysregulation of the brain-gut axis can be manifested as central enhancement of afferent vis­ ceral signals.29 The brain-gut dysregulation can, in turn, be initiated or modified by a variety of events. In a large-scale, prospective, controlled investigation of the development of chronic abdominal pain in women undergoing gynecologic surgery for nonpainful indications, pain developed signi­ ficantly more frequently in the surgical group (15%) than in a nonsurgical control group (4%). The development of chronic abdominal pain in the postoperative setting was predicted only by psychosocial, and not surgical, variables, implying that the development of pain is associated closely with central registration and amplification of the afferent signal. This study lends strong support to the biopsychosocial model, documenting the importance of cognitive and emotional input during the development of postoperative FAPS.

Spinoreticular tract Colon Figure 11-2.  Neuroanatomic pathways that mediate visceral pain sensation. The afferent transmission of visceral abdominal pain involves firstorder neurons that innervate the viscera and subsequently synapse in the dorsal horn of the spinal cord. Second-order neurons ascend from the dorsal horn of the spinal cord via the spinothalamic tract and the spinoreticular and reticulothalamic tracts to link in the thalamus with third-order neurons that then synapse in the limbic system, which contains the insula and anterior cingulate cortex (ACC), and in the primary somatosensory cortex. MCC, midcingulate cortex.

Ascending Visceral Pain Transmission

The afferent transmission of visceral abdominal pain involves first-order neurons that innervate the viscera, carry information to the thoracolumbar sympathetic nervous system, and subsequently synapse in the dorsal horn of the spinal cord. Second-order neurons cross and ascend from the dorsal horn via the spinothalamic and spinoretic­ ular tracts. These second-order neurons synapse in the thalamus with third-order neurons that synapse with the somatosensory cortex (sensory-discriminative component), which is involved in the somatotypic or point-specific local­ ization and intensity of afferent signals, and with the limbic system (motivational-affective component), which contains the anterior cingulate cortex (ACC; Fig. 11-2; see also Chapter 21). The insular cortex receives input from the sensory thalamus and the nucleus tractus solitarius and integrates visceral sensory and emotional information.31 The limbic system serves as a modulator of the pain experience, based on the individual’s emotional state, prior experiences, and cognitive interpretation of the signal. This multicompo­ nent integration of nociceptive information in the CNS explains the variability in the experience and reporting of pain.32 Motivational-affective regions of the CNS are important contributors to the chronic pain experience by modulating afferent sensory information from the intestine. This conceptual scheme of pain modulation has been demonstrated through PET imaging with the use of radiola­ beled oxygen.33 In a group of healthy subjects who immersed their hands in hot water, half were hypnotized to experience the immersion as painful and the other half as not painful or even pleasant. The changes in cortical activation were compared between the two groups, and no difference was found in activity in the somatosensory cortex; however, those who experienced pain had significantly greater activation of the ACC of the limbic system, which is involved in the affective component of the pain experience. Func­ tional brain imaging studies comparing patients with functional gastrointestinal disease and normal controls

Chapter 11  Chronic Abdominal Pain Thalamus Limbic system

ACC

PAG Locus coeruleus Caudal raphe nucleus Noradrenergic pathway Serotonergic pathway

Amygdala Rostral ventral medulla

Opioidergic pathway

Colon Figure 11-3.  The descending endorphin- or enkephalin-mediated inhibitory system. This network includes connections from the sensory cortex and limbic system (via the amygdala and thalamus), which have major links to the midbrain periaqueductal gray (PAG) matter, locus coeruleus, and medullary caudal raphe nucleus. Connections then project to neurons in the dorsal horn of the spinal cord. When activated, this system inhibits afferent impulses from peripheral nociceptive sites (e.g., the colon) to the brain. Endorphin activity, which has opioidergic properties, is facilitated by release of serotonin (serotonergic pathway) and possibly norepinephrine (noradrenergic pathway). ACC, anterior cingulate cortex.

have shown abnormal brain activation mainly in the motivational-affective pain regions, including the prefrontal cortex, ACC, amygdala, and insula.34 These regions gener­ ally show increased activation in patients with chronic pain, thereby suggesting abnormal afferent input as well as central modulation, which could be caused in part by increased attention to visceral stimuli, abnormal cognitive or affective processing of afferent input, or comorbid psy­ chiatric disorders.

Descending Modulation of Pain

According to the gate control theory, afferent transmission of visceral pain can be modulated by descending impulses from the cortex down to the visceral nerves.32 In this model, the central descending control of the gating system occurs primarily through the descending inhibitory system.35 This system is an endorphin- or enkephalin-based neural network that originates from the cortex and limbic system and descends to the spinal cord, with major links in the mid­ brain (periaqueductal gray) and medulla (caudal raphe nucleus; Fig. 11-3). This system inhibits nociceptive pro­ jection directly on the second-order neurons or indirectly via inhibitory interneurons in the spinal cord. Then, the dorsal horn of the spinal cord acts as a gate to modulate (i.e., increase or decrease) transmission of afferent impulses from peripheral nociceptive sites to the CNS. In effect, this descending pain modulation system determines the amount of peripheral afferent input from the gut that is allowed to ascend to the brain. Descending inhibitory systems can be diffuse and, when activated, inhibit pain sensitivity throughout the body—so-called diffuse noxious inhibitory control (DNIC). Patients with chronic pain syn­ dromes, including FAPS, appear to have an impaired ability to activate DNIC.36

Visceral Sensitization

Recurrent peripheral stimulation is thought to up-regulate afferent signals or inhibit descending pain control mecha­

nisms, thereby sensitizing the bowel and producing a state of visceral hyperalgesia (increased pain response to a noxious signal) and chronic pain. Several clinical studies have supported this concept, and the increase in pain appears to occur to a greater degree in patients with func­ tional gastrointestinal disorders than in healthy subjects.37 Furthermore, preoperative treatment with local or regional anesthesia or NSAIDs reduces the severity of postoperative pain,38 suggesting that the CNS response to peripheral injury can be modified by prior reduction of afferent input to the spinal cord and CNS. Conversely, recurrent peripheral injury, such as repeated abdominal operations, may sensi­ tize intestinal receptors, thereby making perception of even baseline afferent activity more painful (allodynia). Visceral sensitization may develop through different mechanisms at one or more levels of the neuraxis, including the mucosal level (via afferent silent nociceptors) and spinal level (spinal hyperexcitability). Patients with IBS may also experience hyperal­gesia. Studies of rectal balloon distention in patients with IBS have demonstrated that a greater proportion of patients report discomfort to balloon distention than normal volunteers at a given volume of inflation; in addition, the intensity of the discomfort in patients is higher than in the normal volunteers.39 Rectal hypersensitivity induced by repetitive painful rectal distention is seen in patients with IBS, but not FAPS.40 This observation supports the conten­ tion that IBS and FAPS are distinct functional gastrointes­ tinal disorders.

Biochemical Mechanisms of Sensitization

The biochemical basis of visceral sensitization is under active study, and this research may identify future targets for therapy. Serotonin (5-hydroxytryptamine [5-HT]) has received considerable attention because the gastrointestinal tract is its main source within the body.41 5-HT is found primarily in mucosal enterochromaffin cells, where it appears to serve as a neurotransmitter of the enteric nervous system (ENS) and as a paracrine molecule that signals other (e.g., vagal) neural activity. 5-HT mediates numerous gas­ trointestinal functions, and modulation of various receptor subtypes, such as 5-HT1, 5-HT3, and 5-HT4, and of 5-HT reuptake affects gastrointestinal sensorimotor function.

Role of the Central Nervous System

Although peripheral sensitization may influence the onset of pain, the CNS is critically involved in the predisposition to and perpetuation of chronic pain. In FAPS, the preemi­ nent role of the CNS is evident by the lack of peripheral motor or sensory abnormalities and the strong association with psychosocial disturbances. In addition, comorbid psy­ chiatric diagnoses, major life stressors, a history of sexual or physical abuse, poor social support, and maladaptive coping all are associated with more severe chronic abdomi­ nal pain and poorer health outcomes.31,42,43 These factors in patients with FAPS and other functional gastrointestinal pain conditions may impair or diminish descending inhibi­ tory pain pathways that act on dorsal horn neurons or may amplify visceral afferent signals.25,36,44 Prospective studies of patients with postinfection IBS (see Chapter 118) and post­ operative FAPS support the importance of the brain in the experience of gastrointestinal pain.30,45 Functional brain imaging has been useful in clarifying brain-gut interaction and has demonstrated that links between emotional distress and chronic pain may be medi­ ated through impairment in the ability of the limbic system to modulate visceral signals. The motivational-affective component of the central pain system, specifically the ACC (see Figs. 11-2 and 11-3), is dysfunctional in patients with

167

168

Section III  Symptoms, Signs, and Biopsychosocial Issues IBS and other chronic painful conditions. Functional mag­ netic resonance imaging (MRI) and PET brain imaging in response to rectal distention in patients with IBS have shown differential activation of the ACC in patients com­ pared with normal subjects46 and increased activation of the thalamus.46,47 Similar results have been found in patients with a history of abuse, somatization, and post-traumatic stress disorder. Furthermore, the return of ACC activity to baseline in depressed patients is associated with clinical improvement48 and predicts response to antidepressant treatment.49 As the pain and emotional distress of a patient with IBS improve, the activity within the ACC changes cor­ respondingly.50 A study of patients with IBS and an abuse history used functional brain MRI to show that during aver­ sive visceral stimulation (rectal balloon distention), differ­ ential activation of regions of the ACC occurred51; areas involved in pain facilitation (posterior and middle cingulate subregions) were stimulated, whereas activity in a region usually associated with pain inhibition (supragenual ante­ rior cingulate) was reduced. This study confirmed a strong association between visceral pain reporting and brain acti­ vation in predetermined brain regions involved in the affective and motivational aspects of the pain experience. The observed synergistic effect of IBS and abuse history on differential ACC activation suggests a mechanism to explain how afferent processing in the CNS can be associ­ ated with reporting of greater pain severity and poorer outcomes in this patient population. This and other research29,52-55 has suggested that dysregulation of central pain modulation is critical and may occur in various medical and psychological conditions. The challenge remains to alter this dysregulated afferent processing network repro­ ducibly and to reverse the findings on functional brain imaging studies (by pharmacologic, psychological, or other therapeutic means), with a concomitant improvement in patient outcomes.

Clinical Implications

The concept of FAPS as a dysregulation of CNS–enteric nervous system function at varying levels of the neuraxis, rather than a purely psychiatric or structural gastrointesti­ nal disorder, suggests that chronic pain results from enhanced pain perception as a result of combinations of the following: (1) activation of silent nociceptors; (2) dorsal horn transmission of impulses stimulated by release of cyto­ kines or other substances; and (3) chronic or frequently recurring psychosocial stresses that influence central pain modulation. By linking psychosocial factors to the patho­ physiology of chronic abdominal pain, this conceptual scheme alters the therapeutic approach from one that is purely psychiatric in nature to one that encompasses a broader array of potential therapies. Early pharmacologic and psychological treatment ultimately may be proven to prevent the development of a subsequent chronic pain syndrome.

CLINICAL FEATURES History

Typically, patients with FAPS are middle-aged and female. The history is one of chronic abdominal pain, often for more than 10 years, and the patient is often in distress at the time of initial consultation. The pain is frequently described as severe, constant, and diffuse. Pain is often a focal point in the patient’s life, may be described in emotional or bizarre terms (e.g., as nauseating or like a knife stabbing), and is not influenced by eating or defecation. The abdominal pain may be one of several painful symptoms or part of a continuum

of painful experiences often beginning in childhood and recurring over time.22 FAPS sometimes coexists with other disorders, and the clinician must determine the degree to which one of these other conditions contributes to the FAPS. Frequently, FAPS will evolve in a patient who has had another well-defined gastrointestinal disorder, but who has been operated on one or more times and, following these operations, has developed chronic abdominal pain. Repetitive surgery in such patients is often performed for alleged intestinal obstruction caused by adhesions. Patients with FAPS often have a psychiatric diagnosis of anxiety, depression, or somatization.24 They may minimize the role of psychological factors, possibly having learned in childhood that attention is more likely received when reporting illness but not emotional distress. A history of unresolved losses is a common feature.56 Symptoms fre­ quently worsen soon after these events and recur on their anniversaries or during holiday seasons. A history of sexual and physical abuse is frequent and is predictive of poor health, refractoriness to medical care, and a high number of diagnostic and therapeutic procedures and health care visits.43 Because patients do not usually volunteer an abuse history, physicians should inquire about this possibility, particularly in those with refractory symptoms.57 Finally, patients with FAPS may report poor social net­ works and exhibit ineffective coping strategies. They feel unable to decrease their symptoms and may “catastrophize”— that is, view their condition in pessimistic and morbid ways without any sense of control over the consequences. These cognitions are associated with greater pain scores that lead to a cycle of more illness reporting, more psychological distress, and poorer clinical outcomes.58 For many, the illness provides social support via increased attention from friends, family, and physicians.

Patient Behavior

Certain behavioral traits are common in patients with FAPS. Often, these patients demand that the physician not only diagnose the problem promptly, but also relieve their chronic symptoms rapidly. They similarly deny a relation­ ship between their problem and psychologically disturbing issues and often attribute depression to pain rather than recognizing it as a primary factor. Frequently, an accompa­ nying spouse or parent takes responsibility for reporting the patient’s history, an observation that suggests the possibility of family dysfunction. A history of narcotic use is not uncommon, as is a request by the patient for such medica­ tion during the initial visit. This type of behavior reflects the patient’s consideration of his or her situation as an acute condition requiring immediate symptom relief, rather than as a chronic condition in which treatment must be directed toward enhancing coping and adaptive strategies.

Physical Examination

Certain physical findings help support a diagnosis of FAPS, yet none is perfectly sensitive or specific. Abdominal palpa­ tion should begin at an area remote from the perceived site of maximal intensity. The patient’s behavior during abdomi­ nal palpation should be noted, with an emphasis on whether a change is noted during distracting maneuvers. Patients with FAPS usually lack signs of autonomic arousal. The presence of multiple abdominal surgical scars without clearly understood indications may suggest chronic pain behaviors that have led to unnecessary procedures. The closed eyes sign may be noted59; when the abdomen is pal­ pated, the patient with FAPS may wince, with her or his eyes closed, whereas those with acute pain caused by organic pathology tend to keep their eyes open in fearful

Chapter 11  Chronic Abdominal Pain anticipation of the examination. Often, the stethoscope sign (i.e., gentle, distracting compression on a painful site of the abdomen with the diaphragm of the stethoscope), elicits a diminished behavioral response in a patient with FAPS, thereby affording a more accurate appraisal of the complaint of pain.

DIAGNOSIS AND DIFFERENTIAL DIAGNOSIS

After obtaining a complete history, performing a thorough physical examination, and paying appropriate attention to psychosocial factors in the patient’s life, the scenario will often point the physician toward a diagnosis of FAPS. A physical examination that does not suggest evidence of organic intra-abdominal pathology, as well as normal results of a battery of routine laboratory tests, lends support to the contention that the patient’s pain is not the result of an identifiable structural disease. Recognition of the diagnostic criteria for FAPS (see Table 11-2), and failure to find evi­ dence of another cause of chronic abdominal pain (see Table 11-1), should lead the physician to a diagnosis of FAPS. If the features of FAPS are absent or atypical, or if concerning abnormalities are found on physical examina­ tion (e.g., abdominal mass, enlarged liver) or on screening laboratory studies (e.g., anemia, hypoalbuminemia), another diagnosis should be considered and pursued accordingly. Not uncommonly, nonspecific abnormalities are found (e.g., a liver cyst) and require determination of their relevance to the patient’s symptoms.

TREATMENT Establishing a Successful Patient-Physician Relationship

Once other diagnoses have been excluded, formation of a successful relationship between the physician and patient with FAPS is necessary for effective management. Several factors must be taken into account to help establish this relationship and move toward successful treatment. An understanding of the psychosocial background is helpful, because a detailed knowledge of this aspect of the patient’s life aids in selecting the most useful treatment strategies. Having an appreciation of the degree of the patient’s under­ standing of the illness is also important, particularly for enhancing the success of a treatment plan. Early in the development of the patient-physician rela­ tionship, it is important to determine whether there are abnormal illness behaviors and associated psychiatric diag­ noses, which are often present in patients with FAPS. The role of the family in relation to the patient’s illness should also be understood. Normally, family experiences with illness lead to emotional support and a focus on recovery. With dysfunctional family interactions, stresses are not managed in an optimal fashion, and diverting attention toward illness serves to reduce family distress.60 Dysfunc­ tion is seen when family members indulge the patient, assume undue responsibility in the patient’s management, or become the spokesperson for the patient. If such family dysfunction is observed, counseling may help the family develop more useful coping strategies. Cultural belief systems must also be understood, because patients may not comply with treatments that are inconsistent with their cultural values. It is important to gain knowledge of the patient’s psychosocial resources (i.e., the availability of social networks) that may assist in buffering the adverse effects of stress and improve the outcome. It is essential for the physician to convey validation of illness to the patient by acknowledging the patient’s illness and the effect it has had on his or her life in a nonjudgmen­ tal fashion. This step is important in ensuring that the

patient understands that the physician considers FAPS to be a medical illness. Empathy is primary, because it acknowledges the reality and distress associated with the patient’s pain. Providing an empathetic approach can provide benefit by improving adherence to a treatment plan, patient satisfaction, and clinical outcomes.61 It does not, however, equate with overreacting to the patient’s wish for a rapid diagnosis and overmedication or performing unnec­ essary diagnostic studies. Education is provided by eliciting the patient’s knowledge of the syndrome, addressing any concerns, explaining the nature of the symptoms, and ensuring understanding in all matters that have been dis­ cussed. It is helpful to reiterate that FAPS is a medical disorder and that symptoms can be attenuated by pharma­ cologic or psychological treatments that modify the regula­ tion of pain control. Reassurance should be provided, because patients may fear serious disease. After the evalu­ ation is complete, the physician should respond to the patient’s concerns in a clear, objective, and nondismissive manner. Both patient and physician must then negotiate the treatment. This approach will enable the patient to contrib­ ute to and take some responsibility for the treatment plan. Within the context of the patient’s prior experience, inter­ ests, and understanding, the physician should provide choices rather than directives. Adherence to a treatment plan is more likely when the patient has confidence that it will benefit him or her and its rationale is understood. Finally, the physician must set reasonable limits in relation to time and effort expended. The key to success is to maintain a trusting relationship, while setting proper boundaries.

Instituting a Treatment Plan

Successful treatment rests on formulating a plan that encom­ passes ongoing interviews to ensure that the patient does not expect a cure. The physician should explain that a real­ istic treatment goal is to attenuate the symptoms and improve daily function. The patient should increase his or her responsibility for the illness by identifying the circum­ stances surrounding episodes of pain, including emotional and cognitive responses. This technique helps the patient achieve insight into aggravating factors and also character­ izes the patient’s coping style. Such information helps identify a strategy for behavioral treatment. The treatment chosen should be based on the severity of symptoms and degree of associated disability. Symptoms that are intermit­ tent and less severe and those that are clearly linked to psychological distress are frequently amenable to psycho­ logical treatment. If the pain is continuous and severe, phar­ macotherapy targeted to achieve central analgesia may be helpful.

Pharmacotherapy

There is a paucity of evidence from prospective, random­ ized, controlled trials to support the use of drug therapy in FAPS. Drug development in the area of functional gastroin­ testinal disorders, particularly FAPS, has been slow. A major reason for this slow progress is the rather empirical process for experimental testing that necessarily occurs in a symptom-based syndrome.62 Pharmacologic brain imaging approaches hold promise as a means to accelerate drug discovery and subsequent development.63 Despite these limitations, some specific medications have been used in the treatment of FAPS (see later). Peripherally acting anal­ gesics (e.g., acetaminophen, aspirin, other NSAIDs) offer little benefit to patients with FAPS, given the pathophysiol­ ogy of the disorder (i.e., a biopsychosocial disorder related to dysfunction of the brain-gut neuraxis). Moreover, narcot­

169

170

Section III  Symptoms, Signs, and Biopsychosocial Issues ics and benzodiazepines should not be prescribed for treat­ ment of FAPS, because of the potential for increased pain sensitivity and a lowering of the pain threshold, respec­ tively. Furthermore, the omnipresent potential for drug dependency with these types of medications must be borne in mind. Importantly, prescribing such medications subor­ dinates the development of more comprehensive treatment strategies to that of providing medication, can be counter­ productive by leading to narcotic-induced potentiation of visceral pain, and can thus result in the narcotic bowel syndrome.63 Narcotic bowel syndrome may occur in patients with other gastrointestinal disorders or in patients with no other structural intestinal disease who have been exposed to high doses of narcotic medication. The clinical scenario is dominated by chronic abdominal pain that continues to worsen, despite the use of escalating doses of narcotics. The keys to successful management of this disorder include the timely recognition of the syndrome followed by the estab­ lishment of an effective physician-patient relationship and graded tapering of the narcotic, with simultaneous insti­ tution of medical therapy to mitigate the effects of opiate withdrawal.63 As in the treatment of other chronic pain disorders, tricy­ clic antidepressants (TCAs) can be helpful in FAPS.64-66 The benefit of these medications is derived from their ability to improve pain directly and treat associated depression. In general, TCAs have been shown to be effective but can cause anticholinergic effects, hypotension, sedation, and cardiac arrhythmias. They can be given in dosages lower than those used to treat major depression (e.g., desipramine, 25 to 100 mg/day at bedtime) to reduce side effects. However, dosage increases may be needed, particularly if the patient has psychiatric comorbidity. There is less evi­ dence for the use of selective serotonin reuptake inhibitors (SSRIs) in FAPS. These medications may cause agitation, sleep disturbance, vivid dreams, and diarrhea but are much safer than TCAs if taken in an overdose. In most cases, administration of a single daily dose (e.g., 20 mg of fluox­ etine, paroxetine, or citalopram) will suffice. Although the efficacy of SSRIs for pain control is not well established, this class of drugs has additional benefits because they are anxiolytic and helpful for patients with social phobia, posttraumatic stress disorder, panic disorder, and obsessional thoughts related to their condition. Drug combinations (e.g., TCAs with SSRIs) have little support for their use in patients with functional gastrointestinal disorders.67 Anticonvulsants such as carbamazepine and gabapentin have been evaluated in other chronic pain syndromes but have no proven efficacy in FAPS. These drugs may find a role as adjunctive agents in the future. As is the case for other peripherally acting analgesics, topical capsaicin would not be expected to be helpful in the management of FAPS.68 Leuprolide acetate may be of benefit for premen­ strual females with FAPS,69 but the consequent reproduc­ tive hormonal effects of this therapy have dampened enthusiasm for this approach. To enhance compliance, especially in the case of TCA use, the physician should explain that these medications work as central analgesics and are not simply being used to treat a psychiatric condition. Investing the time to explain that these drugs induce neurotransmitter changes in the brain and thereby alter pain perception, and that the dosage is usually lower than that typically chosen for treatment of psychiatric disorders, is often helpful. Further, it may be beneficial to emphasize that the lag time for clinical effect may be several weeks; most side effects diminish after a few days and can be reduced by temporarily lowering the dose of the drug.

Mental Health Referral and Psychological Treatments

Patients may be reluctant to see a psychologist or psychia­ trist because they lack knowledge of the benefits of referral, feel stigmatized for being thought to have a psychiatric problem, or see referral as a rejection by the medical physi­ cian. Psychological interventions are best presented as vehi­ cles that are orchestrated in parallel with medical visits and are used to help manage pain and reduce the psychological distress caused by the symptoms. The mental health consultant may recommend any of several types of psychological treatments for pain manage­ ment.21,70 Cognitive-behavioral treatment, which identifies maladaptive thoughts, perceptions, and behaviors, may be beneficial.65 Evidence from functional brain imaging sug­ gests that this psychological intervention decreases activa­ tion from rectal stimulation in the central emotional regions that are typically hyperactive in chronic pain, such as the amygdala, ACC, and frontal cortex.71 Hypnotherapy has been investigated primarily in IBS, where the focus is on relaxation of the gut. A randomized, controlled trial in chil­ dren that included 31 patients with FAPS has concluded that hypnotherapy is superior to standard medical therapy in reducing pain at one year of follow-up.72 Dynamic or interpersonal psychotherapy and relaxation training have less evidence to support their use in FAPS.

ROLE OF LAPAROSCOPY WITH LYSIS OF ADHESIONS The value of laparoscopy with lysis of adhesions (adhesioly­ sis) in patients with chronic abdominal pain continues to be debated. Relevant studies generally have often been retrospective and nonrandomized, with varying criteria for selecting patients and durations of follow-up. Therefore, the role of adhesiolysis is difficult to assess. Prospective observational investigations have shown improvement in 45% to 90% of patients.73-76 Perhaps most provocative is a prospective, blinded, randomized investigation performed by Swank and colleagues in which patients who were found at laparoscopy to have adhesions were randomized to undergo adhesiolysis or no treatment.77 At 12 months of follow-up, patients in both groups reported substantial pain relief and improved quality of life; however, there were no differences between the groups. The authors concluded that laparoscopic adhesiolysis could not be recommended in this setting. Given these somewhat conflicting data, it seems reasonable to withhold laparoscopy in most patients with chronic abdominal pain, with the understanding that, on occasion, the procedure may be of some benefit. The chal­ lenge for the future will be to define which patients will benefit from such intervention.

ACKNOWLEDGMENT

The authors thank Dr. Douglas A. Drossman for his expert discussion of FAPS in previous editions of this text.

KEY REFERENCES

Bixquert-Jiménez M, Bixquert-Pla L. Antidepressant therapy in func­ tional gastrointestinal disorders. Gastroenterol Hepatol 2005; 28:48592. (Ref 66.) Clouse RE, Mayer EA, Aziz Q, et al. Functional abdominal pain syn­ drome. In: Drossman DA, Corazziari E, Delvaux M, et al, editors. Rome III. The functional gastrointestinal disorders. 3rd ed. McLean, Va: Degnon Associates: 2006. p 557. (Ref 21.) Drossman DA, Ringel Y, Vogt B, et al. Alterations in brain activity associated with resolution of emotional distress and pain in a case of severe IBS. Gastroenterology 2003; 124:754-61. (Ref 50.)

Chapter 11  Chronic Abdominal Pain Drossman DA. Brain imaging and its implications for studying centrally targeted treatments in irritable bowel syndrome: A primer for gastro­ enterologists. Gut 2005; 54:569-73. (Ref 29.) Drossman DA. Functional abdominal pain syndrome. Clin Gastroenterol Hepatol 2004; 2:353-65. (Ref 24.) Kuan LC, Li YT, Chen FM, et al. Efficacy of treating abdominal wall pain by local injection. Taiwan J Obstet Gynecol 2006; 45:239-43. (Ref 10.) Lackner JM, Lou Coad M, Mertz HR, et al. Cognitive therapy for irritable bowel syndrome is associated with reduced limbic activity, GI symp­ toms, and anxiety. Behav Res Ther 2006; 44:621-38. (Ref 71.) Mayer EA, Naliboff BD, Craig AD. Neuroimaging of the brain-gut axis: From basic understanding to treatment of functional GI disorders. Gastroenterology 2006; 131:1925-42. (Ref 54.) Peterson LL, Cavanaugh DL. Two years of debilitating pain in a football spearing victim: Slipping rib syndrome. Med Sci Sports Exerc 2003; 35:1634-7. (Ref 17.)

Ringel Y, Drossman DA, Leserman JL, et al. Effect of abuse history on pain reports and brain responses to aversive visceral stimulation: An FMRI study. Gastroenterology 2008; 134:396-404. (Ref 51.) Ringel Y. New directions in brain imaging research in functional gas­ trointestinal disorders. Dig Dis 2006; 24:278-85. (Ref 52.) Sperber AD, Morris CB, Greemberg L, et al. Development of abdominal pain and IBS following gynecological surgery: A prospective, con­ trolled study. Gastroenterology 2008; 134:75-84 (Ref 30.) Swank DJ, Swank-Bordewijk SCG, Hop WCJ, et al. Laparoscopic adhesiolysis in patients with chronic abdominal pain: A blinded randomized controlled multi-centre trial. Lancet 2003; 361:1247-51. (Ref 77.) Vlieger AM, Menko-Frankenhuis C, Wolfkamp SCS, et al. Hypnotherapy for children with functional abdominal pain or irritable bowel syndrome: A randomized controlled trial. Gastroenterology 2007; 133:1430-6. (Ref 72.) Full references for this chapter can be found on www.expertconsult.com.

171

CHAPTE R

12 Symptoms of Esophageal Disease Kenneth R. DeVault

CHAPTER OUTLINE Dysphagia  173 Pathophysiology  173 Differential Diagnosis and Approach  174 Odynophagia  176 Globus Sensation  176 Pathophysiology and Approach  176 Hiccups  177

Symptoms related to the esophagus are among the most common in general medical as well as gastroenterologic practice. For example, dysphagia becomes more common with aging and affects up to 15% of persons age 65 or older.1 Heartburn, regurgitation, and other symptoms of gastroesophageal reflux disease (GERD) also are common. A survey of healthy subjects in Olmsted County, Minnesota, found that 20% of persons, regardless of gender or age, experienced heartburn at least weekly.2 Mild symptoms of GERD rarely indicate severe underlying disease but must be addressed, especially if they have occurred for many years. Frequent or persistent dysphagia or odynophagia suggests an esophageal problem that necessitates investigation and treatment. Other less specific symptoms of possible esophageal origin include globus sensation, chest pain, belching, hiccups, rumination, and extraesophageal complaints, such as wheezing, coughing, sore throat, and hoarseness, especially if other causes have been excluded. A major challenge in the evaluation of esophageal symptoms is that the degree of esophageal damage often does not correlate well with the patient’s or physician’s impression of symptom severity. This is a particular problem in older patients, in whom the severity of gastroesophageal reflux–induced injury to the esophageal mucosa is increased despite an overall decrease in the severity of symptoms.3

DYSPHAGIA Dysphagia, from the Greek dys (difficulty, disordered) and phagia (to eat), refers to the sensation that food is hindered in its passage from the mouth to the stomach. Most patients complain that food sticks, hangs up, or stops, or they feel that the food “just won’t go down right.” Occasionally they complain of associated pain. If asked, “Do you have trouble swallowing?” some patients with dysphagia in the lower esophagus will actually say “no” in that they may only think of swallowing as the transfer of food from the mouth to the esophagus. Dysphagia always indicates malfunction of

Chest Pain of Esophageal Origin  177 Pathophysiology and Approach  177 Heartburn and Regurgitation  178 Pathophysiology and Approach  179 Extraesophageal Symptoms of Gastroesophageal Reflux Disease  179

some type in the oropharynx or esophagus, although associated psychiatric disorders can amplify this symptom.

PATHOPHYSIOLOGY

The inability to swallow is caused by a problem with the strength or coordination of the muscles required to move material from the mouth to the stomach or by a fixed obstruction somewhere between the mouth and the stomach. Occasional patients may have a combination of the two processes. The oropharyngeal swallowing mechanism and the primary and secondary peristaltic contractions of the esophageal body that follow usually transport solid and liquid boluses from the mouth to the stomach within 10 seconds (see Chapter 42). If these orderly contractions fail to develop or progress, the accumulated bolus of food distends the esophageal lumen and causes the discomfort that is associated with dysphagia. In some patients, particularly older adults, dysphagia is the result of low-amplitude primary or secondary peristaltic activity that is insufficient to clear the esophagus. Other patients have a primary or secondary motility disorder that grossly disturbs the orderly contractions of the esophageal body. Because these motor abnormalities may not be present with every swallow, dysphagia may wax and wane (see Chapter 42). Mechanical narrowing of the esophageal lumen may interrupt the orderly passage of a food bolus despite adequate peristaltic contractions. Symptoms vary with the degree of luminal obstruction, associated esophagitis, and type of food ingested. Although minimally obstructing lesions cause dysphagia only with large, poorly chewed boluses of foods such as meat and dry bread, lesions that obstruct the esophageal lumen completely lead to symptoms with solids and liquids. GERD may produce dysphagia related to an esophageal stricture, but some patients with GERD clearly have dysphagia in the absence of a demonstrable stricture, and perhaps even without esophagitis.4 Abnormal sensory perception in the esophagus may lead to the perception of dysphagia, even when the bolus has cleared the esophagus. Because some normal subjects experience the sensation of dysphagia when the distal esophagus

173

174

Section III  Symptoms, Signs, and Biopsychosocial Issues is distended by a balloon, as well as by other intraluminal stimuli, an aberration in visceral perception could explain dysphagia in patients who have no definable cause.5 This mechanism also may apply to the amplification of symptoms in patients with spastic motility disorders, in whom the frequency of psychiatric disorders is increased.6

DIFFERENTIAL DIAGNOSIS AND APPROACH

When faced with a patient who complains of dysphagia, the physician should approach the problem in a systematic way. Most patients can localize dysphagia to the upper or lower portion of the esophagus, although occasional patients with a distal esophageal cause of dysphagia will present with symptoms referred only to the suprasternal notch or higher. The approach to dysphagia can be divided into oropharyngeal and esophageal dysphagia, although con­ siderable overlap may occur in certain groups of patients. In addition, an attempt should be made to determine whether the patient has difficulty only with solid boluses or with liquids and solids.

Oropharyngeal Dysphagia

Processes that affect the mouth, hypopharynx, and upper esophagus produce a distinctive type of dysphagia. The patient often is unable to initiate a swallow and repeatedly has to attempt to swallow. Patients frequently describe coughing or choking when they attempt to eat. The inability to propel a food bolus successfully from the hypopharyngeal area through the upper esophageal sphincter (UES) into the esophageal body is called oropharyngeal, or transfer, dysphagia. The patient is aware that the bolus has not left the oropharynx and locates the site of symptoms specifically to the region of the cervical esophagus. Dysphagia that occurs immediately or within one second of swallowing suggests an oropharyngeal abnormality. At times, a liquid bolus may enter the trachea or nose rather than the esophagus. Some patients describe recurrent bolus impactions that require manual dislodgment. In severe cases, saliva cannot be swallowed, and the patient drools. Abnormalities of speech such as dysarthria or nasal speech may be associated with oropharyngeal dysphagia. Oral pathology should be considered as well. For example, poor teeth or poorly fitting dentures may disrupt mastication and result in an attempt to swallow an overly large or poorly chewed bolus. Loss of salivation—caused by medications, radiation, or primary salivary dysfunction—may result in a bolus that is difficult to swallow. Recurrent bouts of pulmonary infection may reflect spillover of food into the trachea because of inadequate laryngeal protection. Hoarseness may result from recurrent laryngeal nerve dysfunction or intrinsic muscular disease, both of which cause ineffective vocal cord movement. Weakness of the soft palate or pharyngeal constrictors causes dysarthria and nasal speech as well as pharyngonasal regurgitation. Swallowing associated with a gurgling noise may be described by patients with Zenker’s diverticulum. Finally, unexplained weight loss may be the only clue to a swallowing disorder; patients avoid eating because of the difficulties encountered. Potential causes of oropharyngeal dysphagia are shown in Table 12-1. After an adequate history is obtained, the initial test is a carefully conducted barium radiographic examination, which is optimally performed with the assistance of a swallowing therapist (modified barium swallow). If the study is normal with liquid barium, the examination is repeated after the patient is fed a solid bolus in an attempt to bring out the patient’s symptoms and thereby aid in localizing any pathology. If the oropharyngeal portion of the study is

Table 12-1  Causes of Oropharyngeal Dysphagia Neuromuscular Causes* Amyotrophic lateral sclerosis (ALS, Lou Gehrig’s disease) CNS tumors (benign or malignant) Idiopathic UES dysfunction Manometric dysfunction of the UES or pharynx† Multiple sclerosis Muscular dystrophy Myasthenia gravis Parkinson’s disease Polymyositis or dermatomyositis Postpolio syndrome Stroke Thyroid dysfunction Structural Causes Carcinoma Infections of pharynx or neck Osteophytes and other spinal disorders Prior surgery or radiation therapy Proximal esophageal web Thyromegaly Zenker’s diverticulum *Any disease that affects striated muscle or its innervation may result in dysphagia. † Many manometric disorders (hypertensive and hypotensive UES, abnormal coordination, and incomplete UES relaxation) have been described, although their true relationship to dysphagia is often unclear. CNS, central nervous system; UES, upper esophageal sphincter.

normal, the remainder of the esophagus should be examined. This single test usually identifies the problem and directs initial therapy.

Esophageal Dysphagia

Most patients with esophageal dysphagia localize their symptoms to the lower sternum or, at times, the epigastric region. A smaller number of patients will describe a sensation in the suprasternal notch or higher, even though the bolus stops in the lower esophagus. Esophageal dysphagia frequently can be relieved by various maneuvers, including repeated swallowing, raising the arms over the head, throwing the shoulders back, and using the Valsalva maneuver. Motility disorders or mechanical obstructing lesions can cause esophageal dysphagia. To clarify the origin of symptoms of esophageal dysphagia, the answers to three questions are crucial: 1. What type of food or liquid causes symptoms? 2. Is the dysphagia intermittent or progressive? 3. Does the patient have heartburn? On the basis of these answers, distinguishing the several causes of esophageal dysphagia (Table 12-2) as a mechanical or a neuromuscular defect and postulating the specific cause are often possible (Fig. 12-1). Patients who report dysphagia with solids and liquids are more likely to have an esophageal motility disorder than mechanical obstruction. Achalasia is the prototypical esophageal motility disorder in which, in addition to dysphagia, many patients complain of bland regurgitation of undigested food, especially at night, and of weight loss. By contrast, patients with spastic motility disorders such as diffuse esophageal spasm may complain of chest pain and sensitivity to hot or cold liquids. Patients with scleroderma of the esophagus usually have Raynaud’s phenomenon and severe heartburn. In these patients, mild complaints of dysphagia can be caused by a motility disturbance or esophageal inflammation, but severe dysphagia almost always signals the presence of a peptic stricture (see Chapters 35 and 43).

Chapter 12  Symptoms of Esophageal Disease Oropharyngeal (see Table 12-1)

Esophageal (see Table 12-2)

Type of dysphagia

Video swallow study

Abnormal Address specific cause

Type of bolus

Normal Consider other causes (including esophageal dysphagia)

Solids only

Solids and liquids

Character

Character

Progressive

No weight loss

Age >50 or weight loss

Caustic stricture Diverticula Peptic stricture

Carcinoma

Intermittent

Progressive

Intermittent

Eosinophilic esophagitis Esophageal ring

Achalasia Chagas’ disease Scleroderma

Less specific motility disorder

Table 12-2  Common Causes of Esophageal Dysphagia Motility (Neuromuscular) Disorders Primary Disorders Achalasia Diffuse esophageal spasm Hypertensive LES Ineffective esophageal motility Nutcracker (high-pressure) esophagus Secondary Disorders Chagas’ disease Reflux-related dysmotility Scleroderma and other rheumatologic disorders Structural (Mechanical) Disorders Intrinsic Carcinoma and benign tumors Diverticula Eosinophilic esophagitis Esophageal rings and webs (other than Schatzki ring) Foreign body Lower esophageal (Schatzki) ring Medication-induced stricture Peptic stricture Extrinsic Mediastinal mass Spinal osteophytes Vascular compression LES, lower esophageal sphincter.

In patients who report dysphagia only after swallowing solid foods and never with liquids alone, a mechanical obstruction is suspected. A luminal obstruction of sufficiently high grade, however, may be associated with dysphagia for solids and liquids. If food impaction develops,

Figure 12-1.  Diagnostic algorithm for patients with dysphagia. For details of the approach to each type of dysphagia, see the text and tables. Less specific motility disorders include nutcracker esophagus, diffuse esophageal spasm, and other disorders of ineffective esophageal motility. (Modified from Castell DO, Donner MW. Evaluation of dysphagia: A careful history is crucial. Dysphagia 1987; 2:65-71.)

the patient frequently must regurgitate for relief. If a patient continues to drink liquid after the bolus impaction, large amounts of that liquid may be regurgitated. In addition, hypersalivation is common during an episode of dysphagia, thereby providing even more liquid to regurgitate. Episodic and nonprogressive dysphagia without weight loss is characteristic of an esophageal web or a distal esophageal (Schatzki) ring. The first episode typically occurs during a hurried meal, often with alcohol. The patient notes that the bolus of food sticks in the lower esophagus; it often can be passed by drinking large quantities of liquids. Many patients finish the meal without difficulty after the obstruction is relieved. The offending food frequently is a piece of bread or steak—hence the term steakhouse syndrome.7 Initially, an episode may not recur for weeks or months, but subsequent episodes may occur frequently. Daily dysphagia, however, is likely not caused by a lower esophageal ring (see Chapter 41). If solid food dysphagia is clearly progressive, the differential diagnosis includes peptic esophageal stricture and carcinoma. Benign esophageal strictures develop in some patients with GERD. Most of these patients have a long history of associated heartburn. Weight loss seldom occurs in patients with a benign lesion, because these patients have a good appetite and convert their diet to high-calorie soft and liquid foods to maintain weight (see Chapter 43). Patients with carcinoma differ from those with peptic stricture in several ways. As a group, the patients with carcinoma are older and present with a history of rapidly progressive dysphagia. They may or may not have a history of heartburn, and heartburn may have occurred in the past but not the present. Most have anorexia and weight loss (see Chapter 46). True dysphagia may be seen in patients with

175

176

Section III  Symptoms, Signs, and Biopsychosocial Issues pill, caustic, or viral esophagitis; however, the predominant complaint of patients with these acute esophageal injuries is usually odynophagia. Patients may present with food bolus impaction, and eosinophilic esophagitis should be considered in the differential diagnosis of all patients (particularly those who are young) who present with dysphagia (see Chapter 27).8 After a focused history of the patient’s symptoms is obtained, a barium radiograph, including a solid bolus challenge, is often advocated as the first test. Alternatively, many experts have advocated endoscopy as the first test, especially in patients with intermittent dysphagia for solid food suggestive of a lower esophageal ring or with pronounced reflux symptoms. The choice of the initial test should be based on local expertise and the preference of the individual health care provider. If the barium examination demonstrates an obstructive lesion, endoscopy is usually done for confirmation and biopsy. Endoscopy also permits dilation of strictures, rings, and neoplasms. Empirical dilation of the esophagus is often performed in patients with a suggestive history and normal endoscopic examination,9 but the safety and efficacy of this approach have been questioned.10 If the barium examination is normal, esophageal manometry is often performed to look for a motility disorder. Some patients with reflux symptoms and dysphagia, a normal barium study or endoscopy, or both, will respond to a trial of gastric acid suppressive therapy.

Table 12-3  Causes of Odynophagia Caustic Ingestion Acid Alkali Pill-Induced Injury Alendronate and other bisphosphonates Aspirin and other NSAIDs Emepronium bromide Iron preparations Potassium chloride (especially slow-release form) Quinidine Tetracycline and its derivatives Zidovudine Infectious Esophagitis Viral Cytomegalovirus Epstein-Barr virus Herpes simplex virus Human immunodeficiency virus Bacterial Mycobacteria (tuberculosis or Mycobacterium avium complex) Fungal Candida albicans Histoplasmosis Protozoan Cryptosporidium Pneumocystis Severe Reflux Esophagitis Esophageal Carcinoma NSAIDs, nonsteroidal anti-inflammatory drugs.

ODYNOPHAGIA Like dysphagia, odynophagia, or painful swallowing, is specific for esophageal involvement. Odynophagia may range from a dull retrosternal ache on swallowing to a stabbing pain with radiation to the back so severe that the patient cannot eat or even swallow his or her own saliva. Odynophagia usually reflects an inflammatory process that involves the esophageal mucosa or, in rare instances, the esophageal muscle. The most common causes of odynophagia include caustic ingestion, pill-induced esophagitis, radiation injury, and infectious esophagitis (Candida, herpesvirus, and cytomegalovirus; Table 12-3). In these diseases, dysphagia also may be present, but pain is the dominant complaint. Odynophagia is an infrequent complaint of patients with GERD and, when present, usually is associated with severe ulcerative esophagitis. In rare cases, a nonobstructive esophageal carcinoma can produce odynophagia. Because many of the diseases that cause odynophagia have associated symptoms and signs, a carefully taken history can often lead directly to a diagnosis. For example, a teenager who takes tetracycline for acne and in whom odynophagia develops most likely has pill dysphagia, an immunocompromised patient with odynophagia is likely to have an infectious cause, and a patient with GERD is likely to have severe peptic esophagitis. On the other hand, gastrointestinal endoscopy to visualize and obtain biopsies of the esophageal mucosa is required to confirm a specific diagnosis in most patients with odynophagia.

GLOBUS SENSATION Globus sensation is a feeling of a lump or tightness in the throat, unrelated to swallowing. Up to 46% of the general population experience globus sensation at one time or another.11 The sensation can be described as a lump, tight-

ness, choking, or strangling feeling, as if something is caught in the throat. Globus sensation is present between meals, and swallowing of solids or large liquid boluses may give temporary relief. Frequent dry swallowing and emotional stress may worsen this symptom. Globus sensation should not be diagnosed in the presence of dysphagia or odynophagia.

PATHOPHYSIOLOGY AND APPROACH

The detection of physiologic and psychological abnormalities in patients with globus sensation has been inconsistent and controversial. Although frequently suggested, manometrically identifiable UES dysfunction has not been iden­ tified directly as the cause of globus sensation. The UES also does not appear to be hyperresponsive to esophageal distention, acidification, or mental stress.12 Furthermore, esophageal distention can cause globus sensation unrelated to any rise in UES pressure, and stress can induce an increase in UES pressure that is not associated with globus sensation in normal subjects and in patients who complain of globus sensation. Heartburn has been reported in up to 90% of patients with globus sensation,13 yet documentation of esophagitis or abnormal gastroesophageal reflux by esophageal pH monitoring is found in fewer than 25%. Balloon distention of the esophagus produces globus sensation at lower balloon volumes in globus sufferers than in controls; this finding suggests that the perception of esophageal stretch may be heightened in some patients with globus sensation. Psychological factors may be important in the genesis of globus sensation. The most common associated psychiatric diagnoses include anxiety, panic disorder, depression, hypochondriasis, somatization, and introversion.14 Indeed, globus sensation is the fourth most common symptom in patients with somatization disorders.15 A combination of biological factors, hypochondriacal traits, and learned fear

Chapter 12  Symptoms of Esophageal Disease after a choking episode provides a framework for misinterpretation of the symptoms and intensifies the symptoms of globus or the patient’s anxiety.16 The approach to globus sensation involves excluding a more sinister underlying disorder and then offering symptom-driven therapy. A nasal endoscopy to rule out pharyngeal pathology and a barium swallow to rule out a fixed pharyngeal lesion are often helpful.17 If these studies are negative, trials of acid suppression with a proton pump inhibitor, medications directed at visceral sensitivity, or other psychologically based therapies are reasonable. If a patient has heartburn, then acid suppressive therapy is the first step, but reflux may be the cause of globus sensation, even in the absence of heartburn. A trial of a proton pump inhibitor (usually given twice daily, before meals) is diagnostic and therapeutic in some patients. Ambulatory reflux monitoring may show acid or nonacid reflux in some patients.18 Alternatively, if the patient has obvious anxiety and has already failed a trial of acid suppression, therapy directed toward the psychological component of the problem should be considered.

HICCUPS The symptom of hiccups (hiccoughs, singultus) is caused by a combination of diaphragmatic contraction and glottic closure. Therefore, it is not classically an esophageal symptom but is a common complaint in primary care and gastroenterology. Most cases of hiccups are idiopathic, but the symptom has been associated with many conditions (trauma, masses, infections) that affect the central nervous system, thorax, or abdomen. Gastrointestinal causes include GERD, achalasia, gastropathies, and peptic ulcer. Hiccups associated with uremia may be particularly difficult to control. They often occur after a large meal. Because most cases are self-limited, intervention is not usually required. The evaluation of chronic or difficult cases should include selected tests to exclude esophageal, thoracic, or systemic diseases. Because GERD has been associated with hiccups, a trial of acid suppressive therapy may be reasonable in some patients.19 Many agents have been used to suppress hiccups with varying success, including chlorpromazine, nifedipine, haloperidol, phenytoin, metoclopramide, bac­ lofen, and gabapentin.20 Alternative modalities, including acupuncture, also have been tried in refractory cases.21

CHEST PAIN OF ESOPHAGEAL ORIGIN Chest pain of esophageal origin may be indistinguishable to patients and their health care providers from angina pectoris. The esophagus and heart are anatomically adjacent and share innervation. In fact, once cardiac disease is excluded, esophageal disorders are probably the most common causes of chest pain. Of the approximately 500,000 patients in the United States who undergo coronary angio­ graphy yearly for presumed cardiac pain, almost 30% have normal epicardial coronary arteries; of these patients, esophageal diseases may account for the symptoms in 18% to 56%.22 Esophageal chest pain usually is described as a squeezing or burning substernal sensation that radiates to the back, neck, jaw, or arms. Although it is not always related to swallowing, the pain can be triggered by ingestion of hot or cold liquids. It may awaken the patient from sleep and can

worsen during periods of emotional stress. The duration of pain ranges from minutes to hours, and the pain may occur intermittently over several days. Although the pain can be severe, causing the patient to become ashen and to perspire, it often abates spontaneously and may be eased with antacids. Occasionally, the pain is so severe that narcotics or nitroglycerin are required for relief. Close questioning reveals that most patients with chest pain of esophageal origin have other esophageal symptoms; however, chest pain is the only esophageal complaint in about 10% of cases.23 The clinical history does not enable the physician to distinguish reliably between a cardiac and esophageal cause of chest pain. In fact, gastroesophageal reflux may be triggered by exercise24 and cause exertional chest pain that mimics angina pectoris, even during treadmill testing. Symptoms suggestive of esophageal origin include pain that continues for hours, retrosternal pain without lateral radiation, pain that interrupts sleep or is related to meals, and pain relieved with antacids. The presence of other esophageal symptoms helps establish an esophageal cause of pain. As many as 50% of patients with cardiac pain, however, also have one or more symptoms of esophageal disease.25 Furthermore, relief of pain with sublingual nitroglycerin has been shown not to be specific for a coronary origin of pain.26 Cardiac and esophageal disease increase in frequency as people grow older, and both problems may not only coexist but also interact to produce chest pain.

PATHOPHYSIOLOGY AND APPROACH

The specific mechanisms that produce esophageal chest pain are not well understood. Chest pain that arises from the esophagus has commonly been attributed to the stimulation of chemoreceptors (by acid, pepsin, or bile) or mechanoreceptors (by distention or spasm); thermoreceptors (stimulated by cold) also may be involved. Gastroeso­phageal reflux causes chest pain primarily through acid-sensitive esophageal chemoreceptors (see later). Acid-induced dysmotility may be a cause of esophageal pain. Older studies have shown that perfusion of acid into the esophagus in patients with gastroesophageal reflux increases the amplitude and duration of esophageal contractions and induces simultaneous and spontaneous contractions, with the occurrence of pain.27 Diffuse esophageal spasm also has been demonstrated during spontaneous acid reflux. Subsequent studies with modern equipment have shown that such changes in motility are rare.28 In addition, studies using 24-hour ambulatory esophageal pH and motility monitoring have shown that the association between abnormal motility and pain is uncommon, and that spontaneous acid-induced chest pain is rarely associated with abnormalities in esophageal motility.29,30 Patients with chest pain suspected to be esophageal in origin have an increased frequency of esophageal contractions of high amplitude and a slightly increased frequency of simultaneous contractions when compared with a normal control population.31 In addition, intraluminal ultrasound has been able to identify abnormal sustained contractions of the longitudinal smooth muscle in a subset of patients with chest pain.32 How these contractions cause pain is unknown. One possible explanation is that pain occurs when high intramural esophageal tension resulting from altered motility inhibits blood flow to the esophagus for a critical period of time (i.e., myoischemia). MacKenzie and coworkers have found that rates of esophageal rewarming are decreased after infusions of cold water into the esophagus of patients with symptomatic esophageal motility disorders as compared with age-matched controls.33 Because

177

178

Section III  Symptoms, Signs, and Biopsychosocial Issues the rate of rewarming after cold water infusion in patients with Raynaud’s phenomenon correlates directly with blood flow, the authors theorized that esophageal ischemia is the cause of the reduced rate of rewarming. None of the patients with a symptomatic esophageal motility disorder, however, experienced chest pain during the study. Furthermore, the extensive arterial and venous blood supply to the esophagus makes it unlikely that blood flow is compromised after even the most abnormal esophageal contractions. Complicating the relationship between esophageal chest pain and abnormal esophageal contractions is the consistent observation that most of these patients are asymptomatic when the contraction abnormalities are identified. In addition, amelioration of chest pain does not correlate predictably with reduction in the amplitude of esophageal contractions.34 The possibility exists that chest pain–associated motility changes represent an epiphenomenon of a chronic pain syndrome rather than the direct cause of the pain. Other potential causes of esophageal chest pain include excitation of temperature receptors and luminal distention. The ingestion of hot or cold liquids can produce severe chest pain. This association was previously thought to be related to esophageal spasm, but subsequent studies have shown that cold-induced pain produces esophageal aperistalsis and dilatation, not spasm.35 This observation suggests that the cause of esophageal chest pain may be activation of stretch receptors by acute distention. Esophageal distention and pain are experienced during an acute food impaction, drinking of carbonated beverages (in some patients), and dysfunction of the belch reflex.36 In susceptible persons, esophageal chest pain can be reproduced by distention of an esophageal balloon to volumes lower than those that produce pain in asymptomatic persons.37 Therefore, altered pain perception may contribute to the patient’s reaction to a painful stimulus. Panic disorder is a commonly overlooked coexisting condition in patients with chest pain38 and should be sought specifically during history taking. The observation that anxiolytics and antidepressants can raise pain thresholds, as well as improve mood states, may explain why these medications may improve esophageal chest pain in the absence of manometric changes,39,40 The approach to patients with esophageal chest pain has evolved over the years. Before the esophagus is considered to be the cause of chest pain, a cardiac cause must be excluded. Appropriate testing may include an exercise stress test, noninvasive cardiac imaging studies, and coronary angiography. Insufficiency of coronary blood flow with normalappearing epicoronary arteries (microvascular angina) has been suggested as a cause of chest pain in some patients.41 Diagnosing microvascular angina on the basis of a thera­ peutic trial is difficult because the medications reported to improve this condition also have effects on the esophagus; however, the prognosis of patients with microvascular angina is thought to be good. The recognition that chest pain is often associated with GERD has been a major advance in our understanding of esophageal chest pain. Ambulatory pH testing can document pathologic amounts of acid reflux or a correlation between acid reflux and chest pain in up to 50% of patients in whom a cardiac cause has been excluded.42 In addition, a trial of therapy with a proton pump inhibitor produces symptomatic improvement in many such patients.43 The association between chest pain and GERD is easy to recognize when the patient has coexisting reflux symptoms but not so clear when typical reflux symptoms are absent. A 10- to 14-day trial of an oral proton pump inhibitor taken twice daily has been shown to be sensitive and specific for

the diagnosis of esophageal chest pain when compared with ambulatory intraesophageal pH testing (see later).44 If a patient fails this trial, the next practical approach may be a trial of agents such as imipramine or trazodone that raise the pain threshold. Some authorities recommend esophageal testing with stationary manometry at this point to exclude a motility disorder and ambulatory pH testing to exclude reflux unresponsive to the initial trial of the proton pump inhibitor therapy. The advent of a tube-free system for reflux monitoring allows a longer and more comfortable monitoring period, which increases the likelihood of observing a correlation between pain and an acid event.45 If reflux is confirmed by ambulatory pH testing, an additional trial of acid suppressive therapy is warranted. If a spastic motility disorder is discovered on manometry, an attempt at lowering esophageal pressure with nitrates or a calcium channel blocker is appropriate (see Chapter 42).

HEARTBURN AND REGURGITATION Heartburn (pyrosis) is one of the most common gastrointestinal complaints in Western populations.46 In fact, it is so common that many people assume it to be a normal part of life and fail to report the symptom to their health care providers. They seek relief with over-the-counter antacids, which accounts for most of the $1 billion/year sales of these nonprescription drugs. Despite its high prevalence, the term heartburn is frequently misunderstood. It has many synonyms, including indigestion, acid regurgitation, sour stomach, and bitter belching. The physician should listen for these descriptors if the patient does not volunteer a complaint of heartburn. A study from Europe has suggested that using a word-picture description of “a burning feeling rising from the stomach or lower chest up toward the neck” increases the ability to identify patients with reflux.47 The burning sensation often begins inferiorly and radiates up the entire retrosternal area to the neck, occasionally to the back, and rarely into the arms. Heartburn caused by acid reflux may be relieved, albeit only transiently, by the ingestion of antacids, baking soda, or milk. Interestingly, the severity of esophageal damage (esophagitis or Barrett’s esophagus) does not correlate with the severity of heartburn (e.g., patients with severe heartburn may have a normalappearing esophagus, and those with severe esophagitis or Barrett’s esophagus may, at times, have mild or even no symptoms; see Chapters 43 and 44).48 Heartburn is most frequently noted within one hour after eating, particularly after the largest meal of the day. Sugars, chocolate, onions, carminatives, and foods high in fats may aggravate heartburn by decreasing lower esophageal sphincter (LES) pressure. Other foods commonly associated with heartburn—including citrus products, tomato-based foods, and spicy foods—irritate the inflamed esophageal mucosa because of acidity or high osmolarity.49 Beverages, including citrus juices, soft drinks, coffee, and alcohol, also may cause heartburn. Many patients have exacerbation of heartburn if they retire shortly after a late meal or snack, and others say that their heartburn is more pronounced while they lie on their right side.50 Weight gain frequently results in the development of new GERD symptoms and in the worsening of GERD symptoms in patients with preexisting symptoms.51 Activities that increase intra-abdominal pressure, including bending over, straining at stool, lifting heavy objects, and performing isometric exercises, may aggravate heartburn. Running also may aggravate heartburn, and stationary

Chapter 12  Symptoms of Esophageal Disease bike riding may be a better exercise for those with GERD.52 Because nicotine and air swallowing relax LES pressure, cigarette smoking exacerbates the symptoms of reflux.53 Emotions such as anxiety, fear, and worry may exacerbate heartburn by lowering visceral sensitivity thresholds rather than by increasing the amount of acid reflux.54,55 Some heartburn sufferers complain that certain drugs may initiate or exacerbate their symptoms by reducing LES pressure and peristaltic contractions (e.g., theophylline, calcium channel blockers) or by irritating the inflamed esophagus (e.g., aspirin, other nonsteroidal anti-inflammatory drugs, bisphosphonates). Heartburn may be accompanied by the appearance of fluid in the mouth, either a bitter acidic material or a salty fluid. Regurgitation describes return of bitter acidic fluid into the mouth and, at times, the effortless return of food, acid, or bilious material from the stomach. Regurgitation is more common at night or when the patient bends over. The absence of nausea, retching, and abdominal contractions suggests regurgitation rather than vomiting. Water brash is an uncommon and frequently misunderstood symptom that should be used to describe the sudden filling of the mouth with clear, slightly salty fluid. This fluid is not regurgitated material but is secreted from the salivary glands as part of a protective, vagally mediated reflex from the distal esophagus.56 Regurgitation and symptoms similar to water brash can occur in patients with achalasia, who may be misdiagnosed as having GERD. Regurgitation must be distinguished from the syndrome of rumination (see Chapter 14). Rumination is a clinical diagnosis and is best described by the Rome III diagnostic criteria. Patients must have persistent or recurrent regurgitation (not preceded by retching) of recently ingested food into the mouth, with subsequent remastication and swallowing. Supportive criteria include absence of nausea, cessation of the process when the regurgitated material becomes acidic, and content consisting of recognizable food with a pleasant taste in the regurgitant.57 Rumination is essentially a diagnosis of exclusion when there is clinical suspicion. Nocturnal reflux symptoms have particular significance. In a survey of patients with frequent reflux symptoms, 74% reported nocturnal symptoms.58 These nighttime symptoms interrupt sleep and health-related quality of life to a greater degree than daytime reflux symptoms alone. Patients who have prolonged reflux episodes at night also are at increased risk of complications of GERD, including severe reflux esophagitis and Barrett’s esophagus.

PATHOPHYSIOLOGY AND APPROACH

The physiologic mechanisms that produce heartburn remain poorly understood. Although the reflux of gastric acid is most commonly associated with heartburn, the same symptom may be elicited by esophageal balloon distention,59 reflux of bile salts,60 and acid-induced motility disturbances. The best evidence that the pain mechanism is probably related to the stimulation of mucosal chemoreceptors is the sensitivity of the esophagus to acid that is perfused into the esophagus or acid reflux, demonstrated by monitoring of pH. The location of these chemoreceptors is not known. One suggestion is that the esophagus is sensitized by repeated acid exposure, resulting in the production of symptoms from smaller boluses after repeated exposure to acid. This hypersensitivity has been reported to resolve with acid suppressive therapy.61 The correlation between discrete episodes of acid reflux and symptoms, however, is poor. For example, postprandial gastroesophageal reflux is common in healthy people, but

symptoms are uncommon. Intraesophageal pH monitoring of patients with endoscopic evidence of esophagitis typically shows excessive periods of acid reflux, but fewer than 20% of these reflux episodes are accompanied by symptoms.62 Moreover, one third of patients with Barrett’s esophagus, the most extreme form of GERD, are acid-insensitive.63 As patients age, their sensitivity to esophageal acid seems to decline; this finding may explain the common observation that mucosal damage is fairly severe but symptoms are minimal in older patients.64 Therefore, the development of symptoms must require more than esophageal contact with acid. Mucosal disruption and inflammation may be a contributing factor but, on endoscopy, the esophagus appears normal in most symptomatic patients. Other factors that possibly influence the occurrence of heartburn include the acid clearance mechanism, salivary bicarbonate concentration, volume of acid refluxed, as measured by the duration and proximal extent of reflux episodes, frequency of heartburn, and interaction of pepsin with acid (see Chapter 43). In addition, studies in which acid reflux is monitored for more than 24 hours have demonstrated considerable daily variability in esophageal acid exposure.65,66 As noted, heartburn strongly suggests gastroesophageal acid reflux, but peptic ulcer disease, delayed gastric emptying, and even gallbladder disease can produce symptoms similar to those caused by reflux. Regurgitation is not quite as specific for acid reflux as heartburn, and the differential diagnosis of regurgitation should include an esophageal obstruction (e.g., ring, stricture, or achalasia) or a gastric emptying problem (e.g., gastroparesis or gastric outlet obstruction). Some patients have overlap among symptoms of gastroesophageal reflux, dyspepsia, and irritable bowel syndrome (see Chapters 13, 43, and 118).67 The approach to patients with heartburn and regurgitation is discussed extensively in Chapter 43. In brief, published guidelines support an initial trial of acid suppressive therapy, generally with a proton pump inhibitor, as a diagnostic and therapeutic maneuver.68 This concept is cost-effective but plagued by limitations in sensitivity and specificity.69 If the cause of symptoms remains uncertain after a therapeutic trial, ambulatory intraesophageal pH testing is the best test to document pathologic esophageal acid exposure. Endoscopy of the esophagus is reserved for patients with symptoms suggestive of a complication (e.g., dysphagia, weight loss, signs of bleeding), but the predictive value of using a symptom profile to predict esophageal damage is questionable at best. Although not without controversy, most guidelines also suggest endoscopy to screen for Barrett’s esophagus in patients with chronic reflux symptoms70; the risk is particularly increased in older and obese patients.71,72

EXTRAESOPHAGEAL SYMPTOMS OF GASTROESOPHAGEAL REFLUX DISEASE Extraesophageal symptoms of GERD are listed in Table 12-4. Although these symptoms may be caused by esophageal motility disorders, they are most frequently associated with GERD. In patients with extraesophageal symptoms, the classic reflux symptoms of heartburn and regurgitation often are mild or absent (see Chapter 43). Gastroesophageal reflux is thought to cause chronic cough and other extraesophageal symptoms as a result of recurrent microaspiration of gastric contents, a vagally mediated neural reflex or, in many patients, a combination of both. Although bronchodilators lower LES pressure, most persons

179

180

Section III  Symptoms, Signs, and Biopsychosocial Issues Possible extraesophageal manifestation of GERD

Table 12-4  Extraesophageal Manifestations of Gastroesophageal Reflux Disease Asthma Chronic cough Excess mucus or phlegm Globus sensation Hoarseness Laryngitis Pulmonary fibrosis Sore throat

with asthma have gastroesophageal reflux with or without bronchodilator therapy. In animal studies, the instillation of small amounts of acid in the trachea or on the vocal cords73 can produce marked changes in airway resistance, as well as vocal cord ulcers. Direct evidence for aspiration is more difficult to identify in adults and rests primarily on the presence of fat-filled macrophages in sputum,74 radioactivity in the lungs after a tracer is placed in the stomach overnight,75 and a high degree of esophageal or hypopharyngeal acid reflux recorded by 24-hour pH monitoring with dual probes.76 Data from animal and human studies suggest that a neural reflex is another pathophysiologic basis for these symptoms. Acid perfusion into the distal esophagus increases airway resistance in all subjects, but the changes are most marked in patients with both asthma and heartburn.77 Abnormal amounts of acid reflux recorded by prolonged intraesophageal pH monitoring have been identified in 35% to 80% of asthmatic adults.78 Symptoms that suggest refluxinduced asthma include the onset of wheezing in adulthood in the absence of a history of allergies or asthma, nocturnal cough or wheezing, asthma that is worsened after meals, by exercise, or in the supine position, and asthma that is exacerbated by bronchodilators or that is glucocorticoiddependent. In patients with reflux, symptoms strongly suggestive of aspiration include nocturnal cough and heartburn, recurrent pneumonia, unexplained fevers, and an associated esophageal motility disorder. Ear, nose, and throat complaints associated with gastroesophageal reflux include postnasal drip, voice changes, hoarseness, sore throat, persistent cough, otalgia, halitosis, dental erosion, and excessive salivation. Many patients with GERD complain of only head and neck symptoms. Examination of the vocal cords may help in evaluating patients with suspected acid reflux–related extraesophageal problems. Some patients have redness, hyperemia, and edema of the vocal cords and arytenoids. In more severe cases, vocal cord ulcers, granulomas, and even laryngeal cancer, all secondary to GERD, have been reported. Normal results of a laryngeal examination, however, are not incompatible with acid reflux–related extraesophageal symptoms, nor are the aforementioned laryngeal signs specific for a GERD-related pathogenesis. The options in a patient with suspected extraesophageal GERD are to study them with an ambulatory intraesophageal pH test or to initiate a trial of therapy to confirm the diagnosis and treat the symptom (Fig. 12-2). Either approach is reasonable, but many experts favor an initial trial of acid suppressive therapy with a proton pump inhibitor twice daily.79 Ambulatory pH testing is then reserved for those who fail the initial trial, although it is not clear whether pH testing should be done with the patient continuing or discontinuing acid-suppressive therapy (see Chapter 43).

Test first strategy

Exclude underlying cardiac, thoracic, and head/neck disease

Treat first strategy

24-hour pH study

Trial of twice daily PPI

Positive?

Successful?

Yes

No

Trial of twice daily PPI

Successful?

Yes

Yes

24-hour pH study

Positive?

No No

GERD maintenance therapy

Yes

No

Maximize medical therapy or consider antireflux surgery

Consider other diagnosis Figure 12-2.  Suggested approach to patients with extraesophageal manifestations of reflux disease, including noncardiac chest pain. The approach to the exclusion of underlying disease varies, depending on the symptom under evaluation (see text). A proton pump inhibitor (PPI) is given before breakfast and before the evening meal. The duration of the trial depends on the symptom. For example, a 10- to 14-day trial may be sufficient for noncardiac chest pain, whereas a three-month trial may be needed for chronic cough. GERD, gastroesophageal reflux disease.

The association between reflux and extraesophageal symptoms, particularly laryngeal symptoms, has been challenged. In one study, pH monitoring of the hypopharynx and proximal and distal esophagus was performed in patients with presumed acid reflux–related endoscopic laryngeal findings.80 An abnormal result was noted in only 15% of hypopharyngeal probes, 9% of proximal esophageal probes, and 29% of distal esophageal probes, thereby indicating that most patients (70%) with symptoms and signs of laryngeal reflux do not have documentable abnormal acid exposure. That preliminary study was followed by a randomized, placebo-controlled trial of esomeprazole, 40  mg twice daily, in the same patients, with response rates of 42% in those treated with esomeprazole and 46% in those treated with placebo.81 In addition, a randomized controlled trial of therapy with a proton pump inhibitor in asthmatics produced similar results.82 Despite the contradictory data, an early trial of proton pump inhibitor therapy in patients

Chapter 12  Symptoms of Esophageal Disease with symptoms suggestive of extraesophageal GERD is reasonable; however, the patient and physician should not be surprised if this therapy fails.

KEY REFERENCES

Avidan B, Sonnenberg A, Schnell TG, Sontag SJ. There are no reliable symptoms for erosive oesophagitis and Barrett’s oesophagus: Endoscopic diagnosis is still essential. Aliment Pharmacol Ther 2002; 16:735-42. (Ref 48.) Corley DA, Kubo A, Levin TR, et al. Abdominal obesity and body mass index as risk factors for Barrett’s esophagus. Gastroenterology 2007; 133:34-41. (Ref 72.) DeVault KR, Castell DO. Updated guidelines for the diagnosis and treatment of gastroesophageal reflux disease. Am J Gastroenterol 2005; 100:190-200. (Ref 68.) Farup C, Kleinman L, Sloan S, et al. The impact of nocturnal symptoms associated with gastroesophageal reflux disease on health-related quality of life. Arch Intern Med 2001; 161:45-52. (Ref 58.) Fass R, Naliboff BD, Fass SS, et al. The effect of auditory stress on perception of intraesophageal acid in patients with gastroesophageal reflux disease. Gastroenterology 2008; 13:696-705. (Ref 55.) Furuta G, Liacouras C, Collins M, et al. Eosinophilic esophagitis in children and adults: A systematic review and consensus recommendations for diagnosis and treatment. Gastroenterology 2007; 133:134263. (Ref 8.)

Henrikson CA, Howell EE, Bush DE, et al. Chest pain relief by nitroglycerin does not predict active coronary artery disease. Ann Intern Med 2003; 139:979-86. (Ref 26.) Johnson DA, Fennerty MD. Heartburn severity underestimates erosive esophagitis severity in elderly patients with gastroesophageal reflux disease. Gastroenterology 2004; 126:660-4. (Ref 3.) Kiljander TO, Harding SM, Field SK, et al. Effects of esomeprazole 40 mg twice daily on asthma: A randomized, placebo-controlled trial. Am J Resp and Crit Care Med 2006; 173:1091-7. (Ref 82.) Numans ME, Lau J, de Witt NJ, Bonis PA. Short-term treatment with proton-pump inhibitors as a test for gastroesophageal reflux disease. A meta-analysis of diagnostic test characteristics. Ann Intern Med 2004; 140:518-27. (Ref 69.) Pandolfino JE, Richter JE, Ours T, et al. Ambulatory esophageal pH monitoring using a wireless system. Am J Gastroenterol 2003; 98:7409. (Ref 66.) Prakash C, Clouse R. Wireless pH monitoring in patients with noncardiac chest pain. Am J Gastroenterol 2006; 101:446-52. (Ref 45.) Rey E, Moreno-Elola-Olaso C, Artalejo FR, et al. Association between weight gain and symptoms of gastroesophageal reflux in the general population. Am J Gastroenterol 2006; 101:229-33. (Ref 51.) Vaezi MF, Richter JE, Stasney CR, et al. Treatment of chronic posterior laryngitis with esomeprazole. Laryngoscope 2006; 116:254-60. (Ref 81.) Full references for this chapter can be found on www.expertconsult.com.

181

CHAPTE R

13 Dyspepsia Jan Tack

CHAPTER OUTLINE Definition  183 Organic Causes of Dyspepsia  183 Intolerance to Food or Drugs  184 Peptic Ulcer Disease  184 Gastroesophageal Reflux Disease  184 Gastric and Esophageal Cancer  184 Pancreatic and Biliary Tract Disorders  184 Other Gastrointestinal and Systemic Disorders  184 Functional Dyspepsia  185 The Dyspepsia Symptom Complex  185 Epidemiology  187 Pathophysiology  187 Pathogenic Factors  188

DEFINITION Dyspepsia is derived from the Greek words dys and pepse and literally means “difficult digestion.” In current medical terminology, dyspepsia refers to a heterogeneous group of symptoms located in the upper abdomen. Dyspepsia is often broadly defined as pain or discomfort centered in the upper abdomen1,2 and may include multiple and varying symptoms such as epigastric pain, postprandial fullness, early satiation (also called early satiety), anorexia, belching, nausea and vomiting, upper abdominal bloating, and even heartburn and regurgitation. Patients with dyspepsia commonly report several of these symptoms.3 Several consensus definitions of dyspepsia and functional dyspepsia have been proposed. The overlap between symptoms of gastric origin and symptoms of presumed esophageal origin (especially those associated with gastroesophageal reflux disease [GERD]) has remained an area of controversy. With time, definitions of dyspepsia have evolved to become more restrictive and more focused on symptoms thought to arise from the gastroduodenal region and not the esophagus. Earlier definitions considered dyspepsia to consist of all upper abdominal and retrosternal sensations—in effect, all symptoms considered to be referable to the pro­ximal gastrointestinal tract.4 The Rome I and II consensus committees both defined dyspepsia as pain or discomfort centered in the upper abdomen.1,2 Discomfort includes postprandial fullness, upper abdominal bloating, early satiation, epigastric burning, belching, nausea, and vomiting. Heartburn may occur as part of the symptom constellation, but the Rome II committee decided that when heartburn is the predominant symptom, the patient should be con­sidered to have GERD and not dyspepsia. The most recent consensus committee, Rome III, has defined dyspepsia as the presence of symptoms considered by the physician to originate from the gastroduodenal region (Table 13-1).5 Only four symptoms (bothersome postpran-

Approach to Uninvestigated Dyspepsia  189 History and Physical Examination  189 Laboratory Testing  189 Initial Management Strategies  189 Additional Investigations  191 Treatment of Functional Dyspepsia  191 General Measures  191 Pharmacologic Treatment  191 Psychological Interventions  193 Recommendations  193

dial fullness, early satiation, epigastric pain, and epigastric burning) are now considered to be specific for a gastroduodenal origin, although many other symptoms are acknowledged to coexist with dyspepsia. In patients with dyspepsia, additional clinical investigations may identify an underlying organic disease that is the likely cause of the symptoms. In these persons, dyspeptic symptoms are attributable to an organic cause of dyspepsia (Table 13-2). In most people with dyspeptic symptoms, however, no organic abnormality is identified by a routine clinical evaluation, and patients who have undergone a diagnostic investigation (including endoscopy) and have not been found to have an obvious specific cause of their symptoms are said to have functional dyspepsia. The term uninvestigated dyspepsia refers to dyspeptic symptoms in persons in whom no diagnostic investigations have yet been performed and in whom a specific diagnosis that explains the dyspeptic symptoms has not been determined.

ORGANIC CAUSES OF DYSPEPSIA The most important identifiable causes underlying dyspeptic symptoms are peptic ulcer disease and GERD. Malignancies of the upper gastrointestinal tract and celiac disease are less common but important causes of dyspeptic symptoms (see Table 13-2).6-10 The investigation of choice in persons with dyspeptic symptoms is endoscopy, which may identify erosive esophagitis, peptic ulcer, or gastric or esophageal cancer. Systematic studies indicate that approximately 20% of patients with dyspeptic symptoms have erosive esophagitis, 20% have endoscopy-negative GERD, 10% have peptic ulcer, 2% have Barrett’s esophagus, and 1% or less have malignancy.6 Minor findings such as duodenitis or gastritis do not seem to correlate with the presence or absence of dyspeptic symptoms.

183

184

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 13-1  Dyspeptic Symptoms and Their Definitions* Symptom More Specific Postprandial fullness Early satiation Epigastric pain

Epigastric burning Less Specific Bloating in the upper abdomen Nausea Vomiting Belching

Definition An unpleasant sensation perceived as the prolonged persistence of food in the stomach A feeling that the stomach is overfilled soon after starting to eat, out of proportion to the size of the meal being eaten, so that the meal cannot be finished. Previously, the term early satiety was used, but satiation is the correct term for the disappearance of the sensation of appetite during food ingestion. Epigastric refers to the region between the umbilicus and lower end of the sternum, within the midclavicular lines. Pain refers to a subjective, unpleasant sensation; some patients may feel that tissue damage is occurring. Epigastric pain may or may not have a burning quality. Other symptoms may be extremely bothersome without being interpreted by the patient as pain. Epigastric refers to the region between the umbilicus and lower end of the sternum, within the midclavicular lines. Burning refers to an unpleasant subjective sensation of heat. An unpleasant sensation of tightness located in the epigastrium. Bloating should be distinguished from visible abdominal distention Queasiness or sick sensation; a feeling of the need to vomit Forceful oral expulsion of gastric contents associated with contraction of the abdominal and chest wall muscles. Vomiting is usually preceded by and associated with retching, repetitive contractions of the abdominal wall without expulsion of gastric contents. Venting of air from the stomach or the esophagus

*According to the Rome III committee. Adapted from Tack J, Talley NJ, Camilleri M, et al. Functional gastroduodenal disorders. In: Drossman DA, Corazziari E, Delvaux M, et al, editors. Rome III. The Functional Gastrointestinal Disorders. 3rd ed. McLean, Va: Degnon Associates; 2006. p 422.

INTOLERANCE TO FOOD OR DRUGS

Contrary to popular beliefs, ingestion of specific foods such as spices, coffee, or alcohol, or of excessive amounts of food, has never been established as causing dyspepsia.11,12 Although ingestion of food often aggravates dyspeptic symptoms, this effect probably is related to the sensorimotor response to food rather than to specific food intolerances or allergies. Studies have shown that acute ingestion of capsaicin induces dyspeptic symptoms in healthy persons and in those with functional dyspepsia, with greater intensity in the latter group.13 Dyspepsia is a common side effect of many drugs, including iron, antibiotics, narcotics, digitalis, estrogens and oral contraceptives, theophylline, and levodopa. Medications may cause symptoms through direct gastric mucosal injury, changes in gastrointestinal sensorimotor function, provo­ cation of gastroesophageal reflux, or idiosyncratic mech­ anisms. Nonsteroidal anti-inflammatory drugs (NSAIDs) have received the most attention because of their potential to induce ulceration in the gastrointestinal tract. Chronic use of aspirin and other NSAIDs may provoke dyspeptic symptoms in up to 20% of persons, but the occurrence of dyspepsia correlates poorly with the presence of an ulcer. In controlled trials, dyspepsia develops in 4% to 8% of persons treated with NSAIDs, with odds ratios ranging from 1.1 to 3.1 compared with placebo; the magnitude of this effect depends on the dose and type of NSAID.14 Compared with NSAIDs, selective cyclooxygenase-2 inhibitors are associated with a lower frequency of dyspepsia and peptic ulceration.15

PEPTIC ULCER DISEASE

Peptic ulcer is a well-established cause of dyspeptic symptoms and is an important consideration for clinicians in the management of patients who present with dyspepsia. The frequency of peptic ulcer in patients with dyspepsia, however, is only 5% to 10%.6,10,14 Increasing age, NSAID use, and Helicobacter pylori infection are the main risk factors for peptic ulcer (see Chapters 50 and 52).

GASTROESOPHAGEAL REFLUX DISEASE

Erosive esophagitis is a diagnostic marker for GERD, but many patients with symptoms that are attributable to the

reflux of stomach contents into the esophagus have no endoscopic signs of esophageal erosion; this is referred to as nonerosive GERD. Erosive esophagitis is found in approximately 20% of dyspeptic patients, and a similar number of patients may have nonerosive GERD (see Chapter 43).6,10

GASTRIC AND ESOPHAGEAL CANCER

The risk of gastric or esophageal malignancy in patients with dyspeptic symptoms is estimated to be less than 1%.9 The risk of gastric cancer is increased among persons with H. pylori infection, persons with a family history of gastric malignancy, persons with a previous history of gastric surgery, and immigrants from areas endemic for gastric malignancy. The risk of esophageal cancer is increased in men, smokers, persons with a high consumption of alcohol, and those with a long-standing history of heartburn (see Chapters 46 and 54).

PANCREATIC AND BILIARY TRACT DISORDERS

Despite the high prevalence of dyspepsia and gallstones in adults, epidemiologic studies have confirmed that cholelithiasis is not associated with dyspepsia. Therefore, patients with dyspepsia should not be investigated routinely for cholelithiasis, and cholecystectomy in patients with cholelithiasis is not indicated for dyspepsia alone. The clinical presentation of biliary pain is easily distinguishable from that of dyspepsia (see Chapter 65). Pancreatic disease is less prevalent than cholelithiasis, but symptoms of acute or chronic pancreatitis or of pan­ creatic cancer may initially be mistaken for dyspepsia. Pancreatic disorders, however, are usually associated with more severe pain and are often accompanied by anorexia, rapid weight loss, or jaundice (see Chapters 58 to 60).

OTHER GASTROINTESTINAL AND SYSTEMIC DISORDERS

Several gastrointestinal disorders may cause dyspepsia-like symptoms. These include infectious (e.g., Giardia lamblia and Strongyloides stercoralis parasites, tuberculosis, fungal infections, syphilis), inflammatory (e.g., celiac disease, Crohn’s disease, sarcoidosis, lymphocytic gastritis, eosino-

Chapter 13  Dyspepsia Table 13-2  Causes of Dyspepsia Luminal Gastrointestinal Tract Chronic gastric volvulus Chronic gastric or intestinal ischemia Food intolerance Functional dyspepsia Gastroesophageal reflux disease Gastric or esophageal neoplasms Gastric infections (e.g., cytomegalovirus, fungus, tuberculosis, syphilis) Gastroparesis (e.g., diabetes mellitus, postvagotomy, scleroderma, chronic intestinal pseudo-obstruction, postviral, idiopathic) Infiltrative and inflammatory gastric disorders (e.g., Crohn’s disease, eosinophilic gastroenteritis, sarcoidosis, amyloidosis) Irritable bowel syndrome Ménétrier’s disease Peptic ulcer disease Parasites (e.g., Giardia lamblia, Strongyloides stercoralis) Medications Acarbose Aspirin, other nonsteroidal anti-inflammatory drugs (including cyclooxygenase-2 selective agents) Colchicine Digitalis preparations Estrogens Ethanol Gemfibrozil Glucocorticoids Iron Levodopa Niacin Narcotics Nitrates Orlistat Potassium chloride Quinidine Sildenafil Theophylline Pancreaticobiliary Disorders Biliary pain—cholelithiasis, choledocholithiasis, sphincter of Oddi dysfunction Chronic pancreatitis Pancreatic neoplasms Systemic Conditions Adrenal insufficiency Congestive heart failure Diabetes mellitus Hyperparathyroidism Intra-abdominal non-gastrointestinal malignancy Myocardial ischemia Pregnancy Renal insufficiency Thyroid disease

philic gastroenteritis), and infiltrative (e.g., lymphoma, amyloidosis, Ménétrier’s disease) disorders of the upper gastrointestinal tract. Most of these causes will be identi­ fiable by upper gastrointestinal endoscopy with mucosal biopsies. Recurrent gastric volvulus and chronic mesenteric or gastric ischemia may present with dyspeptic symptoms (see Chapters 27, 29, 35, 47, 104, 109, 111, and 114). The symptom pattern associated with gastroparesis (idiopathic, drug-induced, or secondary to metabolic, systemic, or neurologic disorders) is similar to dyspepsia, and the distinction between idiopathic gastroparesis and functional dyspepsia with delayed gastric emptying (see later) is not well-defined (see Chapter 48). Finally, dyspepsia may be the presenting or accom­ panying symptom of acute myocardial ischemia, pregnancy, acute or chronic kidney disease, thyroid dysfunction, adrenal insufficiency, and hyperparathyroidism (see Chapters 35 and 38).

FUNCTIONAL DYSPEPSIA According to the Rome III criteria, functional dyspepsia is defined as the presence of early satiation, postprandial fullness, epigastric pain, and epigastric burning in the absence of organic, systemic, or metabolic disease that is likely to explain the symptoms.

THE DYSPEPSIA SYMPTOM COMPLEX Pattern and Heterogeneity

The dyspepsia symptom complex is broader than the four cardinal symptoms that constitute the Rome III definition. It includes multiple symptoms such as epigastric pain, bloating, early satiation, fullness, epigastric burning, belching, nausea, and vomiting. Although often chronic, the symptoms in patients with functional dyspepsia are mostly intermittent, even during highly symptomatic episodes.3,16 In persons with functional dyspepsia who present to a tertiary care center, the most frequent symptoms are postprandial fullness and bloating, followed by epigastric pain, early satiation, nausea, and belching.17-20 Heterogeneity of symptoms is considerable, however, as shown, for example, in the number of symptoms that patients report (Fig. 13-1). In the general population, the most frequent dyspeptic symptoms are postprandial fullness, early satiation, upper abdominal pain, and nausea.21-23 Weight loss is traditionally considered an alarm symptom, pointing toward potentially serious organic disease. Patients with functional dyspepsia who present to a tertiary care center also have a high frequency of unexplained weight loss,17,18 and population-based studies in Australia and in Europe have established an association between uninve­s­ tigated dyspepsia and unexplained weight loss.22,23

Subgroups

The heterogeneity of the dyspepsia symptom complex is well accepted. Factor analyses of dyspepsia symptoms in the general population and in patients who present to a tertiary care center have not supported the existence of functional dyspepsia as a homogeneous (i.e., unidimensional) condition.22-24 These studies confirmed the hetero­ geneity of the dyspepsia symptom complex but did not provide clinically meaningful subdivisions of the syndrome. Several attempts have been made to identify clinically meaningful dyspepsia subgroups to simplify the intricate heterogeneity of the dyspepsia symptom complex and to guide management. The Rome II consensus committee proposed a classification based on a predominant symptom of pain or discomfort. Although correlations were found between the two subdivisions and the presence or absence of H. pylori infection, the absence or presence of delayed gastric emptying, and response or lack of response to gastric acid suppressive therapy,25,26 the subdivisions have been criticized because of the difficulty in distinguishing pain from discomfort, the lack of a widely accepted definition of predominant, uncertainty concerning overlap between the symptom subgroups, the lack of an association with putative pathophysiologic mechanisms and, especially, the lack of stability of the predominant symptom over short time periods.5,27-30 The Rome III consensus committee has proposed different subdivisions (Fig. 13-2). Studies of patients referred to a tertiary care center and of patients with uninvestigated dyspepsia in the general population have revealed that between 40% and 75% of dyspeptic persons report aggravation of symptoms after ingestion of a meal.23,31,32 Assuming

185

Section III  Symptoms, Signs, and Biopsychosocial Issues 100 90 80 Prevalence (% of patients)

186

70 60 50 40

Absent Mild Moderate Severe

30 20 10 0 Fullness

Bloating

Pain

Nausea

Early satiety

Belching Epigastric Vomiting burning

Figure 13-1.  Prevalence of symptoms and their severity ratings in 674 patients with functional dyspepsia seen at a tertiary referral center. (Unpublished, University Hospital Gasthuisberg, Leuven, Belgium). Uninvestigated dyspepsia (postprandial fullness, early satiation, epigastric pain, epigastric burning) Endoscopy, other investigations

Functional dyspepsia

Postprandial distress syndrome (PDS): Meal-related FD – Early satiation – Postprandial fullness

Organic dyspepsia (e.g., ulcer, esophagitis)

Epigastric pain syndrome (EPS): Meal-unrelated FD – Epigastric pain – Epigastric burning

Figure 13-2.  Classification of uninvestigated dyspepsia, functional dys­ pepsia (FD), and subtypes of functional dyspepsia, according to the Rome III criteria.

that a distinction between meal-related and meal-unrelated symptoms might be pathophysiologically and clinically relevant, the Rome III consensus committee proposed that functional dyspepsia be used as an umbrella term and that postprandial distress syndrome (PDS; meal-related dyspeptic symptoms, characterized by postprandial fullness and early satiation) be distinguished from the epigastric pain syndrome (EPS; meal-unrelated dyspeptic symptoms, characterized by epigastric pain and epigastric burning; Table 13-3).5 Few studies have evaluated the Rome III–based classification of functional dyspepsia. One study of postprandial symptom patterns in persons with functional dyspepsia has provided some support for the distinction between EPS and PDS,32 and a population-based study confirmed the existence of the two distinct subgroups, with less

Table 13-3  Classification of and Diagnostic Criteria for Functional Dyspepsia, Postprandial Distress Syndrome, and Epigastric Pain Syndrome* Functional Dyspepsia† Includes one or more of the following: 1. Bothersome postprandial fullness 2. Early satiation 3. Epigastric pain 4. Epigastric burning and No evidence of structural disease (including at upper endoscopy) that is likely to explain the symptoms Postprandial Distress Syndrome† Must include one or both of the following: 1. Bothersome postprandial fullness, occurring after ordinary-sized meals, at least several times per week 2. Early satiation that prevents finishing a regular meal, at least several times per week Supportive Criteria 1. Upper abdominal bloating or postprandial nausea or excessive belching can be present 2. Epigastric pain syndrome may coexist Epigastric Pain Syndrome† Must include all of the following: 1. Pain or burning localized to the epigastrium of at least moderate severity, at least once per week 2. Pain is intermittent 3. Not generalized or localized to other abdominal or chest regions 4. Not relieved by defecation or passage of flatus 5. Not fulfilling criteria for gallbladder or sphincter of Oddi disorders Supportive Criteria 1. Pain may be of a burning quality, but without a retrosternal component 2. Pain is commonly induced or relieved by ingestion of a meal, but may occur while fasting 3. Postprandial distress syndrome may coexist *According to the Rome III committee. † Criteria fulfilled for the previous 3 months with symptom onset at least 6 months prior to diagnosis. Adapted from Tack J, Talley NJ, Camilleri M, et al. Functional gastroduodenal disorders. In: Drossman DA, Corazziari E, Delvaux M, et al, editors. Rome III. The Functional Gastrointestinal Disorders. 3rd ed. McLean, Va: Degnon Associates; 2006; pp 427-428.

Chapter 13  Dyspepsia than anticipated overlap between EPS and PDS.33 On the other hand, an open-access endoscopy-based study found considerable overlap in endoscopic findings between patients with EPS or PDS and a large group of dyspeptic patients who were not classified with either.34 The validity of the Rome III classification will have to be assessed in additional ongoing and future studies.

Overlap with Heartburn and Irritable Bowel Syndrome

The issue of overlap of dyspepsia with GERD has been a challenging one. Although earlier investigators considered a group of patients with reflux-like dyspepsia,4 the Rome committees did not consider heartburn to arise primarily from the gastroduodenal region, and this symptom was thus excluded from the definition of dyspepsia.2,5 Heartburn commonly occurs along with dyspeptic symptoms, however, both in the general population and in those with a diagnosis of functional dyspepsia.23,27,35 Nevertheless, separating GERD from dyspepsia is hampered by a number of confounding factors, such as the presence of dyspepsia-type symptoms in many patients with GERD36 and difficulties in recognizing heartburn by patients and physicians.37,38 The Rome II consensus committee stated that patients with typical heartburn as a dominant complaint almost invariably have GERD and should be distinguished from patients with dyspepsia.2 Although this distinction is probably valid, it has become clear that the predominant symptom approach does not reliably identify or exclude patients with GERD. The Rome III consensus committee has proposed identification of patients with frequent heartburn, and the suggestion has been made that a word-picture questionnaire be used to facilitate recognition of heartburn by patients and to identify patients with functional dyspepsia who may respond to acid suppressive therapy or in whom pathologic esophageal acid exposure can be demonstrated.39,40 Whereas the Rome II definition for functional dyspepsia excluded patients with predominant heartburn and was unclear about those with nonpredominant heartburn, the Rome III definition stated that heartburn is not a gastroduodenal symptom, although it often occurs simultaneously with symptoms of functional dyspepsia and its presence does not exclude the diagnosis of functional dyspepsia.5 Similarly, the frequent co-occurrence of functional dyspepsia and irritable bowel syndrome (IBS)41 is explicitly recognized and does not exclude a diagnosis of functional dyspepsia.

EPIDEMIOLOGY

Dyspeptic symptoms are common in the general population, with frequencies ranging from 10% to 45%.11,16,23,27,42-44 The frequency of dyspepsia is slightly higher in women, and the influence of age varies among studies. The results of prevalence studies are strongly influenced by the criteria used to define dyspepsia, and several studies included patients with typical symptoms of GERD or did not take into account the presence of dyspepsia-like symptoms in many patients with GERD. When heartburn is excluded, the frequency of uninvestigated dyspepsia in the general population is in the range of 5% to 15%.43,44 Long-term follow-up studies have suggested improvement or resolution of symptoms in approximately half of patients. The annual incidence rate of dyspepsia has been estimated to range from 1% to 6%. Quality of life is significantly affected by dyspepsia, especially functional dyspepsia. Although most patients do not seek medical care, a significant proportion will eventually proceed with a consultation, constituting a major impact on the cost of care.16,45-47 Factors that influence health care–

seeking are the severity of symptoms, fear of underlying serious disease, psychological distress, and lack of adequate psychosocial support (see later).48

PATHOPHYSIOLOGY

Several pathophysiologic mechanisms have been suggested to underlie functional dyspeptic symptoms. These suggested mechanisms include delayed gastric emptying, impaired gastric accommodation to a meal, hypersensitivity to gastric distention, altered duodenal sensitivity to lipids or acid, abnormal intestinal motility, and central nervous system dysfunction.3 The heterogeneity of functional dyspepsia seems to be confirmed in the contribution of one or more of these disturbances in subgroups of patients. The studies that have investigated the pathophysiologic mechanisms of functional dyspepsia predate the Rome III consensus committee and classification. Therefore, most studies define functional dyspepsia according to the Rome I and II consensus definitions.

Delayed Gastric Emptying

Several studies have investigated gastric emptying and its relationship to the pattern and severity of symptoms in patients with functional dyspepsia. The frequency of delayed gastric emptying has ranged from 20% to 50%.3,5 In a meta-analysis of 17 studies involving 868 dyspeptic patients and 397 control subjects, a significant delay in gastric emptying of solids was present in almost 40% of patients with functional dyspepsia.49 Most of the studies, however, were performed in small groups of patients with small groups of control subjects. In the largest studies, gastric emptying of solids was delayed in about 30% of the patients with functional dyspepsia. Most studies failed to find a convincing relationship between delayed gastric emptying and the pattern of symptoms. Three large-scale single-center studies from Europe have shown that patients with delayed gastric emptying for solids are more likely to report postprandial fullness, nausea, and vomiting,20,50,51 although two other large multicenter studies in the United States found no or a very weak association.52,53 Whether delayed gastric emptying causes symptoms or is an epi­ phenomenon is a matter of ongoing controversy.

Impaired Gastric Accommodation

The motor functions of the proximal and distal stomach differ remarkably. Whereas the distal stomach regulates gastric emptying of solids by grinding and sieving the contents until the particles are small enough to pass the pylorus, the proximal stomach serves mainly as a reservoir during and after ingestion of a meal. Accommodation of the stomach to a meal results from a vagally mediated reflex relaxation of the proximal stomach that provides the meal with a reservoir and enables the stomach to handle large intragastric volumes without a rise in intragastric pressure.54 Studies using a gastric barostat, scintigraphy, ultraso­ nography, single photon emission computed tomography (SPECT), or noninvasive surrogate markers (e.g., satiety drinking test) have all suggested the presence of impaired gastric accommodation in approximately 40% of patients with functional dyspepsia.3,5,17,19,54 Insufficient accommodation of the proximal stomach during and after the ingestion of a meal may be accompanied by increased intragastric pressure and activation of mechanoreceptors in the gastric wall, thus inducing symptoms. Although a number of studies found associations between impaired accommodation and both early satiation and weight loss, other studies failed to find such associations. In addition, the mechanisms whereby impaired accommodation can be a cause of

187

188

Section III  Symptoms, Signs, and Biopsychosocial Issues symptoms is still unclear; meal ingestion in the absence of proper relaxation of the proximal stomach may be accompanied by activation of tension-sensitive mechano­ receptors in the proximal stomach. On the other hand, insufficient accommodation of the proximal stomach may force the meal into the distal stomach, thereby causing activation of tension-sensitive mechanoreceptors in a distended antrum.

Hypersensitivity to Gastric Distention

Visceral hypersensitivity, defined as abnormally enhanced perception of visceral stimuli, is considered one of the major pathophysiologic mechanisms in the functional gastrointestinal disorders (see Chapters 11, 21, and 118).55 Several studies have established that as a group, patients with functional dyspepsia are hypersensitive to isobaric gastric distention.3,5,18 The level at which visceral hypersensitivity is generated is unclear, and evidence exists for involvement of tension-sensitive mechanoreceptors as well as alterations at the level of visceral afferent nerves or of the central nervous system.56-58

Altered Duodenal Sensitivity to Lipids or Acid

In healthy subjects and in persons with functional dyspepsia, duodenal perfusion with nutrient lipids, but not glucose, enhances the perception of gastric distention through a mechanism that requires lipid digestion and subsequent release of cholecystokinin.59-61 Duodenal infusion of hydrochloric acid induces nausea in persons with functional dyspepsia but not in healthy subjects, thereby suggesting duodenal hypersensitivity to acid.62 Duodenal pH monitoring using a clipped pH electrode has revealed increased postprandial duodenal acid exposure in patients with functional dyspepsia compared with controls, and this difference has been attributed to impaired clearance of acid.63 On the basis of these observations, it has been proposed that increased duodenal sensitivity to lipids or acid may contribute to the generation of symptoms in patients with functional dyspepsia, but further research in this area is needed.

Other Mechanisms

One study has reported a high frequency of rapid gastric emptying in patients with functional dyspepsia; rapid gastric emptying was correlated with postprandial symptom intensity.64 Other studies that evaluated rapid emptying, however, failed to find such a high frequency or correlation.20,32 Phasic fundic contractions induce transient increases in gastric wall tension, which can be perceived by patients with functional dyspepsia.57 One study has reported lack of suppression of phasic contractility of the proximal stomach after a meal in a subset of patients with functional dyspepsia.65 Evidence also exists that abnormalities in the control of gastric myoelectrical activity, as measured by cutaneous electrogastrography, are found in up to two thirds of patients with functional dyspepsia.66,67 No correlation was found between the symptom pattern and presence of electrogastrographic findings. Small intestinal motor alterations, usually hypermotility with burst activity or clusters and an increased proportion of duodenal retrograde contractions (see Chapter 97), have been reported in patients with functional dyspepsia, but no clear correlation with symptoms has been found.68

PATHOGENIC FACTORS

The cause of symptoms in patients with functional dyspepsia has not been established, but evidence exists for genetic

susceptibility, infectious factors, and psychological factors. The relationship between potential pathogenic factors and putative pathophysiologic mechanisms has not been addressed in depth.

Genetic Predisposition

Population studies have suggested that genetic factors contribute to functional dyspepsia. The risk of dyspepsia is increased in first-degree relatives of patients compared with their spouses.69 In a case-control study, polymorphisms of the GNB3 gene that encodes guanine nucleotide binding protein, beta polypeptide 3 (especially the homozygous GNB3 825C state), were associated with symptoms of functional dyspepsia in blood donors and patients.70 Subsequently, a community-based study in the United States reported that both homozygous variants (CC and TT) of GNB3 were associated with meal-unrelated dyspepsia.71

Infection

Helicobacter pylori Infection Depending on the region and population studied, a variable proportion of patients with functional dyspepsia are infected with H. pylori.3,5 Although H. pylori is associated with a number of organic causes of dyspepsia, little evidence supports a causal relationship between H. pylori infection and functional dyspepsia.72 No consistent differences in the pattern of symptoms or putative pathophysiologic mechanisms have been found between H. pylori–positive and H. pylori–negative subjects.73 The best evidence in support of a role for H. pylori in functional dyspepsia is the small but statistically significant beneficial effect of eradication therapy on symptoms (see later).74 Postinfection Functional Dyspepsia Postinfection functional dyspepsia has been proposed as a possible clinical entity on the basis of a large retrospective study from a tertiary referral center.19 Compared with patients who had functional dyspepsia of unspecified onset, patients with a history suggestive of postinfection functional dyspepsia were more likely to report symptoms of early satiation, weight loss, nausea, and vomiting and had a significantly higher frequency of impaired accommodation of the proximal stomach, which was attributed to dysfunction at the level of gastric nitrergic (nitroxidergic) neurons.19 In a prospective cohort study, development of functional dyspepsia was increased fivefold in patients one year after acute Salmonella gastroenteritis compared with subjects who did not have gastroenteritis.75 Additional studies are required to identify the underlying pathophysiology and risk factors and to determine the long-term prognosis.

Psychosocial Factors

Review of the literature clearly reveals an association between psychosocial factors and functional dyspepsia.3,5,76-80 The most common psychiatric comorbidities in patients with functional dyspepsia are anxiety disorders, depressive disorders, somatoform disorders, and a recent or remote history of physical or sexual abuse.79,80 Psychological distress has long been assumed to be a feature of health care–seeking behavior in patients with functional bowel disorders, including functional dyspepsia. Studies have confirmed an association between dyspeptic symptoms in the general population and psychosocial factors such as somatization, anxiety, and life event stress; this association argues against a mere health care–seeking effect.31,81 Furthermore, the severity of symptoms in patients with functional dyspepsia seen in a tertiary care center is more strongly

Chapter 13  Dyspepsia related to psychosocial factors (especially depression, abuse history, and somatization) than to abnormalities of gastric sensorimotor function (see Chapter 21).82 Although these observations show a close interaction between different psychosocial variables and the presence and severity of symptoms of functional dyspepsia, they do not establish whether the psychosocial factors and dyspeptic symptoms are manifestations of a common predisposition or whether the psychosocial factors play a causal role in the pathophysiology of dyspeptic symptoms. The relationship is unlikely to be simple. A factor analysis of symptoms of functional dyspepsia and their relationship with pathophysiology and psychopathology has clearly demonstrated the heterogeneity and complexity of these interactions. It identified four separate functional dyspeptic symptom factors, of which the factor consisting of epigastric pain was associated with visceral hypersensitivity, several psychosocial dimensions, including somatization and neuroticism, and low health-related quality of life.24 These observations suggest a relationship between psychosocial factors and visceral hypersensitivity in particular. Acutely induced anxiety in healthy volunteers, however, was not associated with increased visceral sensitivity but with decreased gastric compliance and a significant inhibition of meal-induced accommodation.83 In patients with functional dyspepsia, a correlation between anxiety and gastric sensitivity was found in the subgroup of hypersen­ sitive patients, but not in the group as a whole.84 A history of physical or sexual abuse was associated with visceral hypersensitivity in patients with functional dyspepsia.85 Clearly, the role of psychosocial factors in the generation and severity of symptoms, especially in terms of their impact on clinical management, requires further study.

APPROACH TO UNINVESTIGATED DYSPEPSIA Considering the high prevalence of dyspepsia and the large number of persons who present to a physician for their symptoms, the initial aim of management is to decide which patients can be treated empirically and which patients should undergo additional diagnostic evaluation.

HISTORY AND PHYSICAL EXAMINATION

A complete clinical history should be obtained and a physical examination performed in all patients with dyspepsia. The nature of the symptoms, as well as their frequency and chronicity, should be ascertained, particularly with regard to their relationship to the ingestion of meals and the possible influence of specific dietary factors. The onset of symptoms—acute with a gastroenteritis-like episode or more chronic and gradual—is also of interest. The presence and amount of weight loss, if present, needs to be assessed, as should other alarm symptoms, such as anemia, blood loss, and dysphagia. Symptom subgroupings according to the Rome II or III classification have not proved to be of clinical usefulness. In cases of long-standing symptoms, the reason that the patient is seeking health care at this time should be elicited, so that specific fears and concerns can be addressed. Further assessment of symptoms or signs of systemic disorders (e.g., diabetes mellitus, cardiac disease, thyroid disorders) and of the patient’s family and personal history will indicate whether the patient is at risk for specific organic diseases that may present as dyspepsia. Physical findings, such as an abdominal mass or organomegaly, ascites, or fecal occult blood necessitate further evaluation. Specific attention should be given to a history of heartburn, and a word-picture questionnaire may help the

patient recognize the typical symptom pattern.37 Burning pain confined to the epigastrium is a cardinal symptom of dyspepsia and is not considered to be heartburn unless the pain radiates retrosternally. The presence of frequent and typical reflux symptoms should lead to a provisional diagnosis of GERD rather than dyspepsia, and the patient should initially be managed as a patient with GERD (see Chapter 43). On the other hand, overlap of GERD with dyspepsia is probably frequent and needs to be considered when symptoms do not respond to appropriate management of GERD. The possible presence of overlapping IBS should also be assessed, and symptoms that improve with bowel movements or are associated with changes in stool frequency or consistency should lead to a presumptive diagnosis of IBS. The use of prescription and nonprescription medications should be reviewed, and medications commonly associated with dyspepsia (especially NSAIDs) should be discontinued, if possible. In patients in whom NSAIDs cannot be discontinued, a trial of a proton pump inhibitor can be considered, although some guidelines recommend endoscopic evaluation to exclude peptic ulcer (see later).

LABORATORY TESTING

The cost-effectiveness of routine laboratory testing, especially in younger patients with uncomplicated dyspepsia, has not been established. Nevertheless, most clinicians will consider routine tests (complete blood count, routine electrolytes, serum calcium level, liver biochemical tests, and thyroid function tests) in patients older than 45 to 55 years. Other studies such as the serum amylase level, antibodies for celiac disease, stool testing for ova and parasites or Giardia antigen, and a pregnancy test may be considered in certain cases.

INITIAL MANAGEMENT STRATEGIES

In most cases, the patient’s history and physical examination will allow dyspepsia to be distinguished from symptoms suggestive of esophageal, pancreatic, or biliary disease. The history and physical findings, and even the presence of alarm symptoms, are unreliable for distinguishing functional from organic causes of dyspepsia by primary care physicians and by gastroenterologists.6,9,10,86,87 Therefore, most guidelines and recommendations advocate prompt endoscopy when risk factors such as NSAID use, age above a certain threshold, or alarm symptoms are present.88-90 The optimal management strategy for most patients who do not have a risk factor for an organic cause of dyspepsia remains a matter of debate and controversy, and several approaches have been proposed. The options include the following: (1) prompt diagnostic endoscopy, followed by targeted medical therapy; (2) noninvasive testing for H. pylori infection, followed by treatment based on the result (test and treat strategy); and (3) empirical antisecretory therapy. In the two latter strategies, endoscopy is performed in patients who do not respond to treatment or who experience recurrent symptoms after treatment.

Prompt Endoscopy and Directed Treatment

Diagnostic upper gastrointestinal endoscopy allows direct recognition of organic causes of dyspepsia such as peptic ulcer, erosive esophagitis, or malignancy. Endoscopy before any therapy has been instituted is still considered the diagnostic gold standard for patients with an upper gastrointestinal disorder.91 The procedure may also have a reassuring effect on physicians and patients.92-94 Gastric mucosal biopsies allow the diagnosis of H. pylori infection, with subsequent eradication therapy if the result is positive. Endoscopy

189

190

Section III  Symptoms, Signs, and Biopsychosocial Issues is claimed to permit diagnosis of early gastric cancer at a curable stage, but detecting early gastric cancer in a symptomatic individual is relatively rare, and evidence for the claim is weak.95-97 On the other hand, endoscopy is expensive and invasive and may not have as major an impact on treatment as hoped. Patients found to have a peptic ulcer or erosive esophagitis will receive antisecretory therapy. In those with negative upper endoscopy findings, functional dyspepsia and non­ erosive GERD are the likely diagnoses, and both are treated empirically with antisecretory therapy. Still, it is argued that initial empirical antisecretory therapy will only delay endoscopy, because functional dyspepsia and GERD are likely to recur after discontinuation of antisecretory therapy, at which time the patient will be referred for endoscopy anyway. A number of randomized controlled trials have compared prompt endoscopy with empirical noninvasive management strategies. A meta-analysis of five trials that compared initial endoscopy with a test and treat strategy has concluded that initial endoscopy may be associated with a small reduction in the risk of recurrent dyspeptic symptoms but that this gain is not cost-effective.98 Most relevant studies found that the direct and indirect costs are higher with prompt endoscopy and that these costs are not completely offset by a reduction in medication use or the number of subsequent physician visits.99-101 The available data, therefore, do not support early endoscopy as a cost-effective initial management strategy for all patients with uncomplicated dyspepsia. Nevertheless, most available practice guidelines advocate initial endoscopy in all patients above a certain age threshold, usually 45 to 55 years, to detect a potentially curable upper gastrointestinal malignancy.88-90 The rationale for this approach is that the vast majority of gastric malignancies occur in patients older than 45 years and that the rate of cancer detection rises in persons with dyspepsia older than age 45.95-97 Most patients with newly diagnosed gastric cancer, however, are already incurable at the time of diagnosis, and many will have some alarm features that would have warranted immediate endoscopy.97 Early endoscopy is also recommended in patients younger than 45 years who have a family history of gastric cancer, emigrated from a country with a high rate of gastric cancer, or had a prior partial gastrectomy.

Test and Treat for H. pylori Infection

H. pylori infection is causally associated with most peptic ulcers and is the most important risk factor for gastric cancer.102 Because of the involvement of H. pylori in peptic ulcer disease, several consensus panels have advocated noninvasive testing for H. pylori in young patients (younger than 45 to 55 years) with uncomplicated dyspepsia. Patients with a positive test result receive eradication therapy (a proton pump inhibitor and two antibiotics, such as amoxicillin and clarithromycin, taken for 7 to 14 days (see Chapter 50), whereas patients with a negative test result are treated empirically, usually with a proton pump inhibitor. The benefits of this test and treat strategy are the cure of peptic ulcer disease or prevention of future peptic ulcers and symptom resolution in a small subset (approximately 7% above the rate with placebo) of patients with functional dyspepsia who are infected with H. pylori.74,103 Eradication of H. pylori eliminates chronic gastritis and, in theory, may thereby contribute to a reduction in the risk of H. pylori– associated gastric cancer.104 On the other hand, in Western countries, the prevalence of H. pylori infection in patients with uninvestigated dys-

pepsia is rapidly declining, and infection rates are especially low in persons younger than 30 years (10% to 30%). Widespread use of antibiotics has the disadvantages of inducing resistance and occasionally causing allergic reactions. Whether eradication of H. pylori causes or worsens GERD has not been proved and is an ongoing matter of debate.102,105 Furthermore, the accuracy of noninvasive testing for H. pylori depends on the prevalence of H. pylori in the population as well as the sensitivity and specificity of the test. Serologic tests are the least expensive but also the least accurate. If the prevalence of H. pylori in a population is less than 60%, the fecal antigen and urea breath tests for H. pylori are preferred, because their higher accuracy rates lead to a reduction in inappropriate treatment for patients without H. pylori infection (see Chapter 50).106 Randomized placebo-controlled trials have shown only a modest reduction in symptoms of dyspepsia after a test and treat approach in primary care.107-109 A meta-analysis of studies that compared a test and treat strategy with empirical antisecretory therapy in persons with uninvestigated dyspepsia has found little difference in symptom resolution or costs between the two approaches.110 Although earlier models that assumed a higher prevalence rate of H. pylori infection suggested a greater benefit to a test and treat approach,111-113 subsequent economic models have suggested that the test and treat strategy may be equally or less cost-effective than empirical antisecretory therapy.114,115 The test and treat strategy as an initial approach is most likely to be beneficial in areas where the H. pylori infection rate is high.

Empirical Antisecretory Therapy

Initial empirical antisecretory therapy is widely used in primary care for patients with uninvestigated dyspepsia. This approach is attractive because it controls symptoms and heals lesions in most patients with underlying GERD or peptic ulcer disease, and may provide symptomatic benefit in up to one third of patients with functional dyspepsia.116,117 Proton pump inhibitors provide symptomatic relief superior to that of histamine H2 receptor agonists, and the response usually occurs within two weeks of initiating therapy.99 Disadvantages of empirical proton pump inhibitor therapy are a rapid relapse in symptoms after cessation of therapy and the potential for rebound gastric hypersecretion when therapy is discontinued.115,116 Many patients, therefore, will continue to take proton pump inhibitor therapy chronically. As noted earlier, a meta-analysis of studies that compared a test and treat strategy with empirical antisecretory therapy in persons with dyspepsia found little difference in symptom resolution or costs between the two strategies.110 Empirical antisecretory therapy may be equally or more cost-effective.114,115

Recommendations

The optimal cost-effective approach to the initial management of uncomplicated dyspepsia remains unclear, and clinical decisions should take into account specific aspects of a patient’s case and weigh several risk-benefit factors. In a young dyspeptic patient (younger than 45 to 55 years) without alarm features, initial endoscopy cannot be recommended because the yield is low and the test is unlikely to lead to improved outcomes. This position can be reconsidered if the patient is worried about an underlying disease, has a family history of cancer, or has emigrated from an area with a high incidence of gastric or esophageal cancer.

Chapter 13  Dyspepsia In a population with a high prevalence rate (>20%) of H. pylori infection, the test and treat approach remains attractive because it will cure patients with peptic ulcer disease. The tests of choice for H. pylori infection are the urea breath test or the fecal antigen test. H. pylori–positive patients should be given a 7- to 14-day course of H. pylori eradication therapy. In those who are negative for H. pylori, a proton pump inhibitor can be prescribed for one to two months. In populations in which the prevalence of H. pylori infection is low, empirical antisecretory therapy (a proton pump inhibitor for one to two months) appears to be the preferred option. Patients who fail to respond to these initial approaches, and possibly those with recurrent symptoms after cessation of antisecretory therapy, should undergo endoscopy, although the yield is still likely to be low. In patients older than 45 to 55 years without alarm features, most guidelines recommend initial diagnostic endoscopy, although a benefit in the detection of early-stage malignancies remains unproved. In these cases, management will depend on the endoscopic findings and detection of H. pylori infection, but proton pump inhibitor therapy is likely to be prescribed to most patients.

ADDITIONAL INVESTIGATIONS

Additional investigations may be pursued in patients with progressive or refractory dyspepsia that does not respond to the initial management approaches described earlier. Testing for celiac disease and Giardia infection is useful for patients with refractory symptoms, especially when accompanied by weight loss. In patients with severe pain or weight loss, abdominal ultrasonography or computed tomography scans can be used to rule out pancreaticobiliary disease and to screen for stenosis of large abdominal arteries. In cases of severe postprandial fullness, and especially in cases of refractory nausea and vomiting, gastric emptying testing using scintigraphy or a breath test can be considered (see Chapter 48). In cases of a severe delay in gastric emptying, a small bowel series can rule out mechanical obstruction as a contributing factor. In cases of refractory intermittent pain or epigastric burning, esophageal pH with impedance monitoring is useful for diagnosing atypical manifestations of GERD that is not responsive to empirical antisecretory therapy (see Chapter 43). Psychological or psychiatric assessment is recommended in cases of long-standing refractory or debilitating symptoms. Electrogastrography, barostat studies, or simple nutrient challenge tests have been used in pathophysiologic studies but have no established role in the clinical management of dyspeptic patients.

might be advisable.59,60 Similarly, consumption of spicy foods containing capsaicin and other irritants is often discouraged.15 Coffee may aggravate symptoms in some cases119 and, if implicated, should be avoided. Cessation of smoking and alcohol consumption is suggested to be helpful, with no convincing evidence of efficacy.120 The avoidance of aspirin and other NSAIDs is commonly recommended and seems sensible, although not of established value.14,15 If a patient has an apparent coexisting anxiety disorder or depression, appropriate treatment should be considered (see later).

PHARMACOLOGIC TREATMENT

For many but not all patients, pharmacotherapy will be considered. The efficacy of pharmacologic treatments for functional dyspepsia is limited, however.

Acid Suppressive Drugs

In patients with gastroesophageal reflux, a trial of antisecretory therapy often has therapeutic and diagnostic value. Based on meta-analyses of therapeutic outcomes in patients with functional dyspepsia, the efficacy of antacids, sucralfate, and misoprostol has not been demonstrated.121 A metaanalysis of 12 randomized placebo-controlled trials that evaluated the efficacy of H2 receptor antagonists in patients with functional dyspepsia reported a significant benefit over placebo, with a relative risk reduction of 23% and a number needed to treat of 7.121 H2 receptor blockers thus appear to be efficacious in functional dyspepsia. Many of these trials, however, probably included patients with GERD under a broad interpretation of functional dyspepsia, thereby accounting for much of the benefit. A meta-analysis of eight placebo-controlled, randomized trials of proton pump inhibitors for functional dyspepsia also confirmed that this class of agents was superior to placebo, with a number needed to treat of 10 (Table 13-4).121,122 The relative risk reduction (13%) was lower than that for H2 receptor blockers, probably reflecting more stringent entry criteria and better exclusion of patients with GERD. No difference in efficacy was found between halfdose and full-dose proton pump inhibitors, and a double dose of a proton pump inhibitor was also not superior to a single dose. H. pylori status did not affect the response to proton pump inhibitor therapy. Subgrouping of patients with functional dyspepsia using Rome definitions showed a trend for proton pump inhibitor therapy to be most effective in the group with overlapping dyspepsia and reflux, less effective in those with only epigastric pain, and ineffective in those with dysmotility.

Eradication of H. pylori Infection TREATMENT OF FUNCTIONAL DYSPEPSIA GENERAL MEASURES

Reassurance and education are of primary importance in patients with functional dyspepsia. In spite of normal findings at endoscopy, the patient should be given a confident and positive diagnosis. In patients with IBS, a positive physician-patient interaction can reduce health care– seeking behavior, and this approach is probably also valid for patients with functional dyspepsia.118 Lifestyle and dietary measures are usually prescribed to patients with functional dyspepsia, but the impact of dietary interventions has not been studied systematically.11 Having patients eat more frequent, smaller meals seems logical. Because the presence of lipids in the duodenum enhances gastric sensitivity, avoiding meals with a high fat content

A Cochrane meta-analysis has reported a 10% pooled relative risk reduction in dyspepsia for therapy to eradicate H. pylori infection, compared with placebo, at 12 months of follow-up, with a number needed to treat of 14 (Table 13-5).74 Arguments against eradication therapy are the low number of responders and the delayed occurrence of a demonstrable symptomatic benefit. On the other hand, H. pylori eradication can induce sustained remission in dyspepsia, albeit in a small minority of patients.123 Other arguments in favor of the use of eradication therapy are protection against peptic ulcer, presumed protection against gastric cancer, and short-term nature and relatively low cost of the treatment.

Prokinetic Agents

Gastric prokinetic agents are a heterogeneous class of compounds that act through different types of receptors. The

191

192

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 13-4  Meta-Analysis of 10 Randomized Controlled Trials of Proton Pump Inhibitor (PPI) Therapy in Patients with Functional Dyspepsia STUDY/YEAR Blum 2000 Bolling-Stemevald 2002 Farup 1999 Gerson 2005 Peura 2004 (M96) Peura 2004 (M97) Talley 1998 (BOND) Talley 1998 (OPERA) van Zanten 2006 Wong 2002

PPI n/N

PLACEBO n/N

274/395 71/100 6/14 16/21 165/261 164/249 242/423 277/403 49/109 231/301

171/203 79/96 8/10 9/19 104/131 109/133 81/110 71/102 62/115 107/152

RELATIVE RISK 95% CI

RELATIVE RISK 95% CI 0.82 [0.75, 0.90] 0.86 [0.74, 1.01] 0.54 [0.27, 1.06] 1.61 [0.95, 2.74] 0.80 [0.70, 0.90] 0.80 [0.71, 0.91] 0.78 [0.68, 0.89] 0.99 [0.85, 1.14] 0.83 [0.64, 1.09] 1.09 [0.97, 1.23]

Total (95% CI) 2276 1071 Total events: 1495 (PPI), 801 (Placebo) Test for heterogeneity chi-square = 31.33, df = 9, P = 0.0003, I2 = 71.3% Test for overall effect z = 279, P = 0.005

0.87 [0.80, 0.96]

0.1

0.2 0.5 10 1 2 5 Favors PPI Favors placebo

*For each trial, n/N represents the proportion of nonresponders (n) over the total number of patients in that group (N). CI, confidence interval. From Moayyedi P, Soo S, Deeks J, et al. Pharmacological interventions for non-ulcer dyspepsia. Cochrane Database Syst Rev 2006; (4):CD001960.

Table 13-5  Meta-Analysis of 12 Randomized Controlled Trials of Helicobacter pylori Eradication in Patients with Functional Dyspepsia STUDY/YEAR Blum (OCAY) 1998 Froehlich 2001 Gisbert 2004 Gonzalez Carro 2004 Hsu 2001 Kcetz 2003 Koskenpato 2001 Malfertheiner 2003 Martinek 2005 Mazzoleni 2006 McColl 1998 Miwa 2000 Ruiz 2005 Talley (ORCHID) 1999 Talley (USA) 1999 van Zanten 2003 Varannes 2001

TREATMENT n/N

CONTROL n/N

119/164 31/74 13/34 22/47 34/81 67/89 61/77 338/534 5/20 39/46 121/154 33/48 46/79 101/133 81/150 45/75 74/129

130/164 34/70 8/16 31/46 36/80 73/92 63/74 177/266 12/20 40/43 143/154 28/37 64/79 111/142 72/143 55/82 86/124

RELATIVE RISK 95% CI

RELATIVE RISK 95% CI 0.92 [0.81, 1.03] 0.86 [0.60, 1.24] 0.76 [0.40, 1.46] 0.69 [0.48, 1.00] 0.93 [0.66, 1.33] 0.95 [0.81, 1.11] 0.93 [0.80, 1.08] 0.95 [0.85, 1.06] 0.42 [0.18, 0.96] 0.91 [0.79, 1.06] 0.85 [0.77, 0.93] 0.91 [0.70, 1.18] 0.72 [0.58, 0.89] 0.97 [0.85, 1.11] 1.07 [0.86, 1.34] 0.89 [0.70, 1.14] 0.83 [0.68, 1.00]

Total (95% CI) 1934 1632 Total events: 1230 (Treatment), 1163 (Control) Test for heterogeneity chi-square = 17.69, df = 16, P = 0.34, I2 = 9.5% Test for overall effect z = 4.58, P < 0.00001

0.90 [0.86, 0.94]

0.1 0.2 0.5 1 2 5 10 Favors Treatment Favors Control *For each trial, n/N represents the proportion of nonresponders (n) over the total number of patients in that group (N). CI, confidence interval. From Moayyedi P, Soo S, Deeks J, et al. Eradication of Helicobacter pylori for non-ulcer dyspepsia. Cochrane Database Syst Rev 2006; (2):CD002096.

efficacy of available prokinetic agents in functional dyspepsia has been controversial.121,124,125 A meta-analysis, based mainly on studies of domperidone and cisapride, suggested superiority of prokinetic agents over placebo in patients with functional dyspepsia, with a relative risk reduction of 33% and a number needed to treat of six121; separate analy-

ses have also suggested efficacy for cisapride and domperidone individually.124 Metoclopramide and domperidone are dopamine receptor agonists with a stimulatory effect on upper gastrointestinal motility. Unlike metoclopramide, which may cause serious neurologic adverse effects, domperidone—which is not approved by the U.S. Food

Chapter 13  Dyspepsia and Drug Administration—does not cross the blood-brain barrier. Cisapride facilitates the release of acetylcholine in the myenteric plexus via 5-hydroxytryptamine4 (5-HT4) receptor agonism and accelerates gastric emptying. The available trials with these drugs, however, were often of poor quality, concerns were raised about publication bias, and cisapride has been withdrawn from the market because of cardiac safety concerns.121 Unfortunately, more recent studies with other types of prokinetic agents have generally not demonstrated symptomatic relief in patients with functional dyspepsia.125 The motilin receptor agonist, ABT-229, was actually found to worsen symptoms compared with placebo.126 Mosapride, which like cisapride is a mixed 5-HT4 receptor agonist and 5-HT3 receptor antagonist, demonstrated no benefit when compared with placebo in a large European study.127 The 5-HT4 receptor agonist tegaserod, 6 mg twice daily, was evaluated in two phase 3 randomized controlled trials in women with dysmotility-like functional dyspepsia. The two primary endpoints were the percentage of days with satisfactory symptom relief and the symptom severity on a composite average daily severity score. Statistical significance for both endpoints was obtained in one study but not in the other, and the overall therapeutic gain was small.128 The drug was well tolerated in this program but was withdrawn from the market because of an increased frequency of cardiovascular ischemic events. Itopride is a dopamine D2 antagonist and acetylcholinesterase inhibitor that was intensively studied in functional dyspepsia. A phase 2 placebo-controlled trial found significantly more responders to itopride, based on a global efficacy measure.129 No significant improvement in symptoms compared with placebo was observed, however, in two subsequent phase 3 trials.130

Antidepressants

Antidepressants are commonly used for the treatment of functional gastrointestinal disorders that do not respond to initial conventional approaches. Although systematic reviews suggest that anxiolytics and antidepressants, especially tricyclic antidepressants, may have some benefit in patients with functional gastrointestinal disorders, including functional dyspepsia (pooled relative risk reduction of 45%), the available trials are small and of poor quality, and publication bias cannot be excluded.131,132 A multicenter controlled trial of the tricyclic antidepressant desipramine in patients with functional bowel disorders failed to show benefit in an intention-to-treat analysis, but symptomatic improvement was obtained in a per-protocol analysis.133 Most of the enrolled patients, however, seemed to have IBS, and the number of patients with functional dyspepsia in the trial is unclear. The mechanism of action of antidepressants is also unclear; symptomatic relief from these medications appears to be independent of the presence of depression,133 and no significant effects of antidepressants on visceral sensitivity have been established in functional dyspepsia.134,135 The selective serotonin reuptake inhibitor paroxetine enhanced gastric accommodation in healthy subjects, but clinical studies evaluating this class of agents in functional dyspepsia are lacking. A large controlled trial with the selective serotonin and norepinephrine reuptake inhibitor venlafaxine in functional dyspepsia failed to show any benefit.135

Other Pharmacotherapeutic Approaches

Based on a meta-analysis of four trials, bismuth salts seemed efficacious, but the analysis had marginal statistical signi­ ficance.121 Simethicone was superior to placebo in one

controlled trial.136 Various studies reported an improvement in symptoms during treatment with mixed herbal prep­ arations, Chinese herbal preparations, or artichoke leaf extract.137-139 The data suggest that some of these preparations are effective, but the basis for the improvement remains to be determined. One study reported that the chronic administration of red pepper was more effective than placebo in decreasing the intensity of dyspeptic symptoms in patients with functional dyspepsia.140

New Drug Development

Fundic relaxants and visceral analgesics to reverse impaired gastric accommodation and visceral hypersen­sitivity are other attractive approaches for treating sensorimotor dis­ orders of the upper gastrointestinal tract. Although nitrates, sildenafil, and sumatriptan can relax the proximal stomach, they seem less suitable for therapeutic application in functional dyspepsia.54,125 A number of serotonergic drugs are also able to enhance gastric accommodation, including 5-HT1A, 5-HT3, and 5-HT4 receptor agonists.54,125 A clinical trial with a newly developed 5-HT1A receptor agonist R137696 in functional dyspepsia failed to show any symptomatic benefit.141 Acotiamide (Z-338) is a novel compound that enhances acetylcholine release via antagonism of the M1 and M2 muscarinic receptors (see Chapter 49). In a pilot study, acotiamide showed potential to improve symptoms and quality of life through a mechanism that may involve enhanced accommodation.142 Visceral hypersensitivity is another attractive target for drug development. The principal drug classes under evaluation are neurokinin receptor antagonists and peripherally acting kappa opioid receptor agonists. The kappa opioid agonist fedotozine showed potential efficacy in functional dyspepsia, but development of this drug was discontinued.143 More recently, asimadoline, another kappa opioid receptor agonist, failed to improve symptoms in a small pilot study.144

PSYCHOLOGICAL INTERVENTIONS

Studies have shown that patients with functional dyspepsia have a higher prevalence of psychosocial comorbidities, although the role of psychosocial factors in symptom generation remains unclear. Based in part on these comorbidities, psychological interventions such as group support with relaxation training, cognitive therapy, psychotherapy, and hypnotherapy have been used in patients with functional dyspepsia. A systematic review of clinical trials of psychological interventions for functional dyspepsia found that all trials claimed benefit from psychological interventions, with effects persisting for longer than one year, but all studies were limited by inadequate statistical analysis.145 The authors concluded that the evidence to confirm the efficacy of psychological interventions in functional dyspepsia is insufficient.

RECOMMENDATIONS In patients with functional dyspepsia who have mild or intermittent symptoms, reassurance, education, and some dietary changes may be sufficient (Figs. 13-3 and 13-4). Drug therapy can be considered for patients with more severe symptoms or those who do not respond to reassurance and lifestyle changes. Testing for H. pylori infection is recommended and, if the results are positive, eradication therapy can be prescribed. An immediate impact on symptoms is unlikely, however, and any potential benefit is observed mainly over longer follow-up. Both proton pump inhibitors

193

194

Section III  Symptoms, Signs, and Biopsychosocial Issues Uninvestigated dyspepsia

Clinical evaluation: History and physical examination Determine reason for presentation If patient has any of the following: Age >45-55 years Alarm features: Unexplained weight loss Bleeding Unexplained anemia Dysphagia Protracted vomiting Change in character of chronic symptoms Fear of cancer or organic disease

Consider: Dietary indiscretion Medication-induced dyspepsia Cardiac disease Conditions associated with gastroparesis Hepatobiliary disorders Other systemic disease

If patient has: Age <45-55 years and No alarm features or Chronic, mild symptoms or Prior full evaluation

See Figure 13-4

Endoscopy with biopsy for Helicobacter pylori infection

Normal endoscopy result indicating functional dyspepsia

Organic disease: PUD, GERD, cancer

Treat as indicated

Reassurance Lifestyle changes Treat H. pylori infection, if present Trial of PPI for 2-4 weeks, especially if heartburn or epigastric pain or burning Trial of prokinetic agent if postprandial fullness and early satiation Consider IBS (altered bowel habits, abdominal pain)

Symptoms persist Consider depression, psychosocial issues Trial of antidepressant

Refractory/disabling symptoms: Psychological therapies Hypnotherapy Referral to psychiatrist or psychologist Consider analgesics

Figure 13-3.  Management algorithm for patients with dyspepsia. Patients younger than 45 to 55 years who do not have alarm features should be evaluated as in Figure 13-4. GERD, gastroesophageal reflux disease; IBS, irritable bowel syndrome; PPI, proton pump inhibitor; PUD, peptic ulcer disease.

and prokinetic agents can be used as initial pharmaco­ therapy. The symptom pattern may help in determining the most appropriate initial choice of therapy, but a change in drug class is advisable in case of an insufficient therapeutic response. A two- to (preferably) four-week trial of a proton pump inhibitor should be given to all patients with coexisting heartburn and also to those with epigastric pain or burning. If the drug provides symptomatic relief, treatment should be interrupted, and intermittent or chronic therapy with a proton pump inhibitor (or H2 receptor antagonist) tried for patients with repeated relapses. In patients with postprandial fullness and early satiation, a prokinetic agent with an attractive safety profile (e.g., domperidone, where available) can be considered. Metoclopramide generally should not be used because of the risk of serious adverse events. (Cisapride is generally not available and should also be avoided.) Although combinations of a proton pump inhib­itor and prokinetic agent may have additive symptomatic effects, in theory, therapy with a single drug is preferable.

In patients with bothersome symptoms that persist despite these initial therapies, a trial of a low-dose tricyclic anti­ depressant may be considered, even in the absence of apparent anxiety or depression. Higher doses can be considered for patients with significant anxiety or depression. It seems advisable to avoid selective serotonin and nor­ epinephrine reuptake inhibitors. A trial of simethicone, medically prescribed herbal preparations with apparent benefit in controlled trials, or bismuth salts also may be considered for otherwise refractory patients. In patients with debilitating epigastric pain, symptomatic analgesics, even possibly opioids, can be considered after appropriate exclusion of organic disease. Referral to a psychiatrist or psychotherapist can be considered for patients with obvious coexisting psychiatric disease, a history of physical or sexual abuse, or a debili­ tating impact of severe symptoms on daily life activities. Motivated patients may benefit from psychological approaches such as psychotherapy, hypnotherapy, cognitive behavioral therapy, or relaxation therapy.

Chapter 13  Dyspepsia KEY REFERENCES

Dyspepsia in patients <45-55 years

Alarm features Family history or ethnic Yes EGD risk of GI cancer Excessive worry No Noninvasive Helicobacter pylori test Urea breath test Fecal antigen test Serology, if above unavailable

H. pylori positive: Eradication therapy

H. pylori negative: Reassurance Lifestyle changes

Symptoms persist: PPI trial for 2-4 weeks

Improvement: Stop PPI

If symptoms relapse: Intermittent or continuous PPI Consider EGD

No improvement: EGD with biopsy for H. pylori

Normal endoscopy result indicating functional dyspepsia: If H. pylori test is still positive, treat with second-line regimen

Treat as in Fig. 13-3

Organic disease: GERD PUD Cancer

Bisschops R, Karamanolis G, Arts J, et al. Relationship between symptoms and ingestion of a meal in functional dyspepsia. Gut 2008; 57:1495-503. (Ref 32.) Camilleri M, Dubois D, Coulie B, et al. Prevalence and socioeconomic impact of upper gastrointestinal disorders in the United States: Results of the U.S. Upper Gastrointestinal Study. Clin Gastroenterol Hepatol 2005; 3:543-52. (Ref 44.) Delaney B, Ford AC, Forman D, et al. Initial management strategies for dyspepsia. Cochrane Database Syst Rev 2005; (4):CD001961. (Ref 99.) Enck P, Dubois D, Marquis P. Quality of life in patients with upper gastrointestinal symptoms: Results from the Domestic/International Gastroenterology Surveillance Study (DIGEST). Scand J Gastroenterol Suppl 1999; 231:48-54. (Ref 45.) Haycox A, Einarson T, Eggleston A. The health economic impact of upper gastrointestinal symptoms in the general population: Results from the Domestic/International Gastroenterology Surveillance Study (DIGEST). Scand J Gastroenterol Suppl 1999; 231:38-47. (Ref 46.) Ofman JJ, MacLean CH, Straus WL, et al. Meta-analysis of dyspepsia and nonsteroidal anti-inflammatory drugs. Arthritis Rheum 2003; 49:508-18. (Ref 14.) Moayyedi P, Delaney BC, Vakil N, et al. The efficacy of proton pump inhibitors in nonulcer dyspepsia: A systematic review and economic analysis. Gastroenterology 2004; 127:1329-37. (Ref 122.) Moayyedi P, Soo S, Deeks J, et al. Eradication of Helicobacter pylori for non-ulcer dyspepsia. Cochrane Database Syst Rev 2006; (2):CD002096. (Ref 74.) Moayyedi P, Soo S, Deeks J, et al. Pharmacological interventions for non-ulcer dyspepsia. Cochrane Database Syst Rev 2006; (4):CD001960. (Ref 121.) Soo S, Moayyedi P, Deeks J, et al. Psychological interventions for nonulcer dyspepsia. Cochrane Database Syst Rev 2005; (2):CD002301. (Ref 145.) Tack J, Bisschops R, Sarnelli G. Pathophysiology and treatment of functional dyspepsia. Gastroenterology 2004; 127:1239-55. (Ref 3.) Tack J, Talley NJ, Camilleri M, et al. Functional gastroduodenal disorders. Gastroenterology 2006; 130:1466-79. (Ref 5.) Talley NJ, Vakil NB, Moayyedi P. American Gastroenterological Association technical review on the evaluation of dyspepsia. Gastroenterology 2005; 129:1756-80. (Ref 88.) Talley NJ, Vakil N. Practice Parameters Committee of the American College of Gastroenterology. Guidelines for the management of dyspepsia. Am J Gastroenterol 2005; 100:2324-37. (Ref 89.) Vakil N, Moayyedi P, Fennerty MB, Talley NJ. Limited value of alarm features in the diagnosis of upper gastrointestinal malignancy: Systematic review and meta-analysis. Gastroenterology 2006; 131: 390-401. (Ref 9.) Full references for this chapter can be found on www.expertconsult.com.

Treat as indicated

Figure 13-4.  Proposed initial test and treat management algorithm for patients with uncomplicated dyspepsia (no alarm features) based on results of noninvasive testing for Helicobacter pylori. Although this strategy may be cost-effective in patients of any age, consensus recommendations of panels of experts endorse use of this strategy only in younger patients (younger than 45 to 55 years). In communities in which the prevalence of H. pylori infection is low, an initial trial of therapy with a proton pump inhibitor may be more cost-effective than an initial test and treat strategy. EGD, esophagogastroduodenoscopy; GERD, gastroesophageal reflux disease; GI, gastrointestinal; PPI, proton pump inhibitor; PUD, peptic ulcer disease.

195

CHAPTE R

14 Nausea and Vomiting Juan-R. Malagelada and Carolina Malagelada

CHAPTER OUTLINE Pathophysiology  197 Clinical Characteristics of Vomiting  198 Causes  198 Acute Vomiting  198 Chronic or Relapsing Vomiting  200 Nausea and Vomiting During Pregnancy  201 Functional Vomiting  201 Cyclic Vomiting Syndrome  202 Superior Mesenteric Artery Syndrome  203 Rumination Syndrome  203 Diagnostic Evaluation  204 Acute Vomiting  204 Chronic Vomiting  204

Nausea, retching, and vomiting may occur separately or together. When they occur together they are often in sequence, as manifestations of the various physiologic events that integrate the emetic reflex. Vomiting is a complex act that requires central neurologic coordination, whereas nausea and retching do not imply activation of the vomiting reflex. When nausea, retching, or vomiting manifest as iso­ lated symptoms, their clinical significance may differ from the stereotypical picture of emesis.1,2 Nausea is an unpleasant subjective sensation that most people have experienced at some point in their lives and usually recognize as a feeling of impending vomiting in the epigastrium or throat. Retching consists of spasmodic and abortive respiratory movements with the glottis closed. When part of the emetic sequence, retching is associated with intense nausea and usually, but not invariably, culminates in the act of vomiting. Vomiting is a partially voluntary act of forcefully expelling gastric or intestinal content through the mouth. Vomiting must be differentiated from regurgitation, an effortless reflux of gastric contents into the esophagus that sometimes reaches the mouth but is not usually asso­ ciated with the forceful ejection typical of vomiting (see Chapter 12).

PATHOPHYSIOLOGY The mechanism of vomiting has been well characterized in experimental animals and humans (Fig. 14-1).3 Neurologic coordination of the various components of vomiting is pro­ vided by the emetic center (or vomiting center) located in the medulla, specifically in the dorsal portion of the lateral reticular formation in the vicinity of the fasciculus soli­ tarius. The afferent neural pathways that carry activating signals to the emetic center arise from many locations in the body. Afferent neural pathways arise from various sites

Complications  206 Emetic Injuries to the Esophagus and Stomach  206 Spasm of the Glottis and Aspiration Pneumonia  206 Fluid, Electrolyte, and Metabolic Alterations  206 Nutritional Deficiencies  206 Treatment  207 Correction of Metabolic Complications  207 Pharmacologic Treatment  207 Gastric Electrical Stimulation  209

along the digestive tract—the pharynx, stomach, and small intestine. Afferent impulses from these organs are relayed at the solitary nucleus (nucleus tractus solitarius) to the emetic center. Afferent pathways also arise from nondiges­ tive organs such as the heart and testicles. Pathways from the chemoreceptor trigger zone (CTZ) located in the area postrema on the floor of the fourth ventricle activate the emetic center. Despite its central location, the CTZ is outside, at least in part, the blood-brain barrier and serves primarily as a sensitive detection apparatus for circulating endogenous and exogenous molecules that may activate emesis. Finally, pathways arise from other central nervous system structures, including the cortex, brainstem, and vestibular system, via the cerebellum. The circuitry of the emetic reflex involves multiple receptors.4 The following elements are the most relevant to clinical issues: 1. Stimulation of the 5-hydroxytryptamine3 (5-HT3) sero­ tonin receptor provokes release of dopamine, which in turn stimulates dopamine D2 receptors in the emetic center, thereby activating the emetic sequence. This sequence is the basis for the pharmacodynamic action of antiemetic agents, such as ondansetron, a 5-HT3 receptor inhibitor that is effective in the treatment of chemotherapy-induced vomiting,5 and metoclopramide, a dopamine D2 receptor antagonist.6 2. Histamine H1 and muscarinic M1 receptors, which are abundant in the vestibular center and solitary nucleus, constitute the preferred pharmacologic targets for inhib­ iting motion sickness, vestibular nausea, and pregnancyrelated emesis.7 3. Cannabinoid CB1 receptors in the dorsal vagal complex inhibit the emetic reflex.8,9 Cannabinoid agonists also modulate 5-HT3 ion channels. Thus, the CB and 5-HT3 receptor systems colocalize and interact in the brainstem.10 4. Neurokinin-1 (NK-1) receptors located in the area pos­ trema and the solitary nucleus bind to substance P and are part of the terminal emetic pathways. NK-1 antago­

197

198

Section III  Symptoms, Signs, and Biopsychosocial Issues Drugs/chemicals: dopamine agonists, cancer chemotherapy, apomorphine, digoxin Chemoreceptor trigger zone (CTZ) Emetic center

Medulla

CLINICAL CHARACTERISTICS OF VOMITING Receptors for dopamine (D2) and serotonin (5-HT3)

Vagal afferents

Sympathetic afferents

Spinal cord

GI tract Stomach

Other trigger areas: Pharynx Coronary arteries Peritoneum Bile ducts Cortex, thalamus, hypothalamus Vestibular apparatus (motion sickness) Triggers (e.g., staphylococcal enterotoxin)

Figure 14-1.  Schematic of the proposed neural pathways that mediate vomiting. GI, gastrointestinal; 5-HT, 5-hydroxytryptamine.

nists reduce emesis induced by peripherally and cen­ trally acting emetogens. 5-HT3 receptors appear to be involved to a greater extent in centrally induced emesis than in peripherally induced emesis. Therefore, NK-1 receptor antagonists appear to be more efficacious than 5HT3 receptor inhibitors and other known antiemetic drugs in reducing vomiting induced by a variety of causes. Conversely, they may have less potent antinausea effects.87 When activated, the emetic center sets into motion, through neural efferents, the various components of the emetic sequence.11 First, nausea develops as a result of acti­ vation of the cerebral cortex; the stomach relaxes concomi­ tantly, and antral and intestinal peristalsis are inhibited. Second, retching occurs as a result of activation of spasmodic contractions of the diaphragm and intercostal muscles combined with closure of the glottis. Third, the act of vomiting occurs when somatic and visceral components are activated simultaneously. The components include brisk contraction of the diaphragm and abdominal muscles, relaxation of the lower esophageal sphincter, and a forceful retrograde peristaltic contraction in the jejunum that pushes enteric content into the stomach and from there toward the mouth.12 Simultaneously, protective reflexes are activated. The soft palate is raised to prevent gastric content from entering the nasopharynx, respiration is inhibited momen­ tarily, and the glottis is closed to prevent pulmonary aspira­ tion, which is a potentially serious complication of vomiting. Other reflex phenomena that may accompany this picture include hypersalivation, cardiac arrhythmias, and passage of gas and stool rectally.

Certain clinical features may be characteristic of specific causes of vomiting. Nausea and vomiting that occur in the morning or with an empty stomach and with emission of mucoid material (swallowed saliva) or gastroenteric secre­ tions are characteristic of vomiting produced by direct acti­ vation of the emetic center or CTZ. This type of emesis is most typical of pregnancy, drugs, toxins (e.g., alcohol abuse), or metabolic disorders (diabetes mellitus, uremia). Psychogenic vomiting also may exhibit these characteris­ tics. Clinical tradition holds that excessive nocturnal post­ nasal drip may be responsible for this type of vomiting, although direct evidence for this association is lacking. Vomiting that occurs outside the immediate postprandial period and that is characterized by evacuation of retained and partially digested food is typical of slowly developing gastric outlet obstruction or gastroparesis.13 Pseudovomitus, in which totally undigested food that has not been exposed to gastric juice is expelled, may occur in long-standing acha­ lasia or with a large Zenker’s diverticulum. Bilious vomiting is commonly seen after multiple vomiting episodes occur in close succession because of retrograde entry of intestinal material into the stomach. It is also characteristic of patients with a surgical enterogastric anastomosis, in whom the gastric contents normally include bile-stained enteric reflux­ ate. Vomitus with a feculent odor suggests intestinal obstruc­ tion, ileus associated with peritonitis, or long-standing gastric outlet obstruction. Vomiting that develops abruptly without preceding nausea or retching (projectile vomiting) is characteristic of, but not specific for, direct stimulation of the emetic center, as may occur with intracerebral lesions (tumor, abscess) or increased intracranial pressure.14

CAUSES In clinical practice, establishing the cause of vomiting promptly is critical, because specific treatment may be fea­ sible. Acute (less than 1 week) and chronic vomiting should be considered separately, because the respective causes gen­ erally differ. Patients with chronic vomiting tend to consult a specialist after being symptomatic for some time, whereas patients with severe acute vomiting require immediate medical attention. Causes of nausea and vomiting are listed in Table 14-1.

ACUTE VOMITING

In the patient with acute vomiting, the following two ques­ tions must be answered immediately: 1. Is emergency action required? The patient must be assessed for shock, hypokalemia, other serious electro­ lyte disturbances, hollow viscus perforation, organ infarction, cerebral edema, and poisoning. 2. Is the female patient pregnant? In fertile females, preg­ nancy must be considered first. Once these two issues are addressed, a number of potentially emergent diagnostic possibilities should be considered.

Acute Intestinal Obstruction

Vomiting may be a presenting feature of intestinal obstruc­ tion caused by an incarcerated hernia or stool impaction; the latter entity is seen in older, debilitated, or mentally retarded persons. Distal duodenal and proximal jejunal neo­ plasms (adenocarcinoma, lymphoma, leiomyosarcoma, car­

Chapter 14  Nausea and Vomiting Table 14-1  Principal Causes of Nausea and Vomiting Abdominal Causes Mechanical obstruction Gastric outlet obstruction Small bowel obstruction Motility disorders Chronic intestinal pseudo-obstruction Functional dyspepsia Gastroparesis Other intra-abdominal causes Acute appendicitis Acute cholecystitis Acute hepatitis Acute mesenteric ischemia Crohn’s disease Gastric and duodenal ulcer disease Pancreatitis and pancreatic neoplasms Peritonitis and peritoneal carcinomatosis Retroperitoneal and mesenteric pathology Drugs* Aspirin and other nonsteroidal anti-inflammatory drugs Antidiabetic agents Antigout drugs Antimicrobials Acyclovir Antituberculosis drugs Erythromycin Sulfonamides Tetracycline Cancer chemotherapy Cis-platinum Cytarabine Dacarbazine Etoposide 5-Fluorouracil Methotrexate Nitrogen mustard Tamoxifen Vinblastine Cardiovascular drugs Antiarrhythmics Antihypertensives Beta blockers Calcium channel blockers Digoxin Diuretics Central nervous system drugs Antiparkinsonian drugs (levodopa and other dopamine agonists) Anticonvulsants Gastrointestinal medications Azathioprine Sulfasalazine Narcotics Oral contraceptives Theophylline Infectious Causes Acute gastroenteritis Viral Bacterial Nongastrointestinal (systemic) infections

Metabolic and Endocrine Causes Acute intermittent porphyria Addison’s disease Diabetic ketoacidosis Diabetes mellitus Hyperparathyroidism and other causes of hypercalcemia Hyperthyroidism Hyponatremia Hypoparathyroidism Pregnancy Nervous System Causes Demyelinating disorders Disorders of the autonomic system Hydrocephalus Congenital malformations Increased intracranial pressure Low-pressure hydrocephalus Intracerebral lesions with edema Abscess Hemorrhage Infarction Neoplasm Labyrinthine disorders Labyrinthitis Ménière’s disease Motion sickness Meningitis Migraine headaches Otitis media Seizure disorders Visceral neuropathy Other Causes Anxiety and depression Cardiac disease Congestive heart failure Myocardial infarction, ischemia Radiofrequency ablation Collagen vascular disorders Scleroderma Systemic lupus erythematosus Cyclic vomiting syndrome Eating disorders Ethanol abuse Functional disorders Hypervitaminosis A Intense pain Paraneoplastic syndrome Postoperative state Postvagotomy Radiation therapy Starvation

*Partial list.

cinoid) may cause gastric outlet or intestinal obstruction that manifests as acute or chronic vomiting. Proximal intes­ tinal obstruction may be particularly difficult to diagnose because the obstructing lesion may be overlooked or unreachable by conventional upper gastrointestinal endos­ copy and yet may present without the typical picture of dilated fluid-filled loops of small bowel (air-fluid levels) on plain abdominal films (see Chapter 119).

Gastric Outlet Obstruction

In the past, peptic ulcer disease was a major cause of gastric outlet obstruction (see Chapter 52). Before the 1980s, 12% of patients with a peptic ulcer presented with gastric outlet obstruction, either as a direct consequence of a pyloric channel ulcer with associated edema and pylorospasm or, more commonly, as a result of marked deformity of the entire antroduodenal region in the setting of long-standing

199

200

Section III  Symptoms, Signs, and Biopsychosocial Issues ulcer disease. Obstruction caused by a peptic ulcer can occur abruptly with acute vomiting or insidiously, mimick­ ing the clinical picture of gastroparesis (see Chapter 48). As the incidence of peptic ulcer disease has declined sharply and patients are treated early and more effectively in the course of the disease, peptic ulcer disease has become a much less frequent cause of gastric outlet obstruction. Gastric volvulus is a relatively uncommon but important cause of acute vomiting; symptoms may be relapsing as a result of intermittent volvulus formation and spontaneous resolution. Paraesophageal and post-traumatic diaphrag­ matic hernias also may predispose to acute vomiting as a result of obstruction (see Chapter 24).15 Both acute and chronic pancreatitis, with associated inflammatory masses, necrosis, pseudocysts, or secondary infection, may lead to gastric outlet obstruction at the duodenum or, less commonly, the antrum and pylorus (see Chapters 58 and 59). Similarly, gastric, duodenal, or pancreatic malignancies (adenocarcinoma, lymphoma, cystic pancreatic neoplasms) may cause gastric outlet obstruction, sometimes manifesting as acute vomiting (see Chapters 29-32, 54, 60, and 121).

antiarrhythmics), antibiotics, levodopa and its derivatives, theophylline, opiates, and azathioprine. Patients on multi­ drug regimens pose a special challenge in determining which drug (or drugs) is the culprit.

Intestinal Infarction

Nausea and vomiting may be the sole or predominant mani­ festation of neurologic disorders. Meningeal inflammation is another potential cause. Nausea and vomiting may be associated with vertigo in patients with vestibular or cere­ bellar disorders and motion sickness. Migraine headaches may be accompanied by nausea and vomiting with little or no headache, making the diagnosis difficult. Ictal vomiting is a rare manifestation, most often associated with right temporal lobe epilepsy.20 Intracerebral lesions associated with increased intracranial pressure, interference with intracerebral fluid flow, or direct compression of the emetic center may manifest with nausea and vomiting. Projectile vomiting is a common but not invariable feature of intrace­ rebral lesions.

A diagnosis of intestinal infarction should be considered in any patient with acute vomiting.16 Intestinal infarction may occur with a paucity of physical signs but requires expedi­ tious management. The diagnosis is more common in patients with vascular disorders and thrombotic diatheses and in older adults (see Chapter 114).

Extraintestinal Causes

Extraintestinal causes of vomiting usually do not present a challenging diagnostic problem because the primary condi­ tion is generally clinically apparent. Myocardial infarction may manifest initially as acute vomiting because of afferent connections between the heart and the emetic center. Renal colic and biliary pain similarly may manifest with intense vomiting, although the localization of the pain and other characteristic features usually make these diagnoses evident (see Chapter 65). Ovarian or testicular torsion may manifest initially with intense vomiting. Intraperitoneal or retroperitoneal inflammatory condi­ tions, including acute appendicitis, bowel perforation, acute pancreatitis and, in general, any cause of acute abdominal pain, may be associated with vomiting. On occa­ sion, vomiting may be so intense (and, rarely, the only symptom) as to cause diagnostic confusion (see Chapters 10, 26, 37, and 116).

Toxins and Drugs

Vomiting caused by toxins and drugs is common but usually not difficult to diagnose. Alcohol abuse and various types of poisoning should be considered. Cancer chemotherapy is associated with a high likelihood of nausea and vomiting, although routine administration of antiemetic agents before chemotherapy often prevents nausea and vomiting. Vomiting also can be induced by radiotherapy. Chemotherapeutic agents and combinations of agents vary in their propensity to cause nausea and vom­ iting (see Table 14-1).17,18 The list of drugs that can induce nausea and vomiting is lengthy (see Table 14-1). Some classes of drugs and indi­ vidual agents are particularly common culprits in clinical practice, especially aspirin and other nonsteroidal antiinflammatory drugs (NSAIDs; their emetic effect is attenu­ ated partially by coadministration of a proton pump inhibitor; see Chapter 52), cardiovascular drugs (digitalis,

Metabolic Causes

Metabolic causes of vomiting include diabetic ketoacidosis, hyponatremia, and hypercalcemia. Diabetic gastroparesis associated with visceral neuropathy is usually associated with chronic relapsing nausea and vomiting (see later). The clinical onset of diabetic gastroparesis, however, may be abrupt (see Chapter 48). Addison’s disease also may mani­ fest clinically with acute vomiting.

Infectious Causes

Vomiting may be caused by acute gastritis or gastroenteritis caused by a virus or bacterium, including bacterial toxins, such as that produced by Staphylococcus.19 During the early stages of the illness, nausea and vomiting may be the pre­ dominant or even exclusive clinical manifestation (see Chapters 51 and 107).

Neurologic Causes

Postoperative Nausea and Vomiting

Postoperative nausea and vomiting is generally a therapeu­ tic rather than a diagnostic problem. About one third of patients who do not receive antiemetic prophylaxis will experience nausea and vomiting after surgery.21 The risk is highest with abdominal, gynecologic, strabismus, and middle ear surgery, and is three times as common in women as in men. The differential diagnosis includes complica­ tions of surgery, such as intestinal perforation, peritonitis, and electrolyte disturbances. Cardiac disease (“silent” myo­ cardial infarction, congestive heart failure) also may mani­ fest as nausea and vomiting in the postoperative period.

CHRONIC OR RELAPSING VOMITING

In patients with chronic or relapsing vomiting, the same causes of acute vomiting discussed earlier must be consid­ ered, but with important differentiating features. Additional considerations include pregnancy, functional vomiting, cyclic vomiting syndrome, and pseudovomiting.

Partial Intestinal Obstruction

In contrast to acute complete intestinal obstruction, partial intestinal obstruction may be associated with relapsing vomiting over long periods of time. Abdominal pain and distention may accompany the clinical picture but may wax and wane as intestinal flow is intermittently interrupted and spontaneously restored. The clinical presentation of longstanding partial intestinal obstruction and chronic intesti­ nal pseudo-obstruction (an intestinal motor disorder) may be similar. In fact, exclusion of occult partial intestinal

Chapter 14  Nausea and Vomiting obstruction is a prerequisite for the diagnosis of pseudoobstruction (see Chapters 119 and 120). Stenotic Crohn’s disease, neoplasms of the intestine, and ischemic strictures are the main causes of partial mechanical intestinal obstruc­ tion (see Chapters 111, 114, and 121). Chronic adhesions from surgery or pelvic inflammatory disease are also poten­ tial causes of intestinal obstruction, although establishing their pathogenic role is sometimes difficult. Advanced intra-abdominal cancer is another important cause of intestinal obstruction.22 In older, debilitated, and mentally retarded individuals, constipation may lead to a picture of intestinal obstruction when the colon becomes impacted with stool and ileal outflow is partially impeded (see Chapter 18).23

Gastric Outlet Obstruction

When partial and sustained over time, gastric outlet obstruc­ tion is an important cause of chronic vomiting. In the past, this presentation was common in patients with peptic ulcer disease (see earlier). With the advent of early endoscopic diagnosis of peptic ulcer and effective endoscopic manage­ ment of gastric outlet obstruction, this presentation of peptic ulcer is now uncommon.

Gastrointestinal Motility Disorders

Gastroparesis and chronic intestinal pseudo-obstruction may produce chronic vomiting.13,24 Recurrent vomiting, sometimes with symptom-free periods, is a major compo­ nent of the clinical picture of gastroparesis. As in partial gastric outlet obstruction, abdominal pain is absent, the stomach may become markedly dilated, and the vomitus may contain partially digested food, but these findings are not constant. Some patients with neuropathic gastroparesis, as is associated with diabetes mellitus, may vomit repeat­ edly, even with an empty stomach; epigastric pain may occur. Nausea and vomiting may be presenting features of intestinal pseudo-obstruction, but other symptoms and signs associated with small bowel dysmotility, such as abdominal pain and distention, are usually present. The distinction between primary and secondary forms of gastro­ paresis and chronic intestinal pseudo-obstruction often requires specific diagnostic tests (see later; see also Chapters 48 and 120).

Neurologic Disorders

Neurologic disorders are an important and sometimes clini­ cally elusive source of chronic nausea and vomiting, even when the various neuropathies responsible for gastroparesis and chronic intestinal pseudo-obstruction are excluded. Foremost among neurologic causes of chronic or relapsing vomiting is migraine, particularly atypical forms without an aura or family history and with delayed or no headache. Hydrocephalus and lesions that compress or irritate the emetic center in the base of the brain also may account for chronic vomiting.

NAUSEA AND VOMITING DURING PREGNANCY

Nausea occurs in more than half of all normal pregnancies and frequently is associated with vomiting.25 These symp­ toms tend to develop early in pregnancy, peak around 9 weeks of gestation, and rarely continue beyond 22 weeks of gestation. Nausea with vomiting is more common in women with multiple gestations than with a single gestation. The origin of nausea and vomiting during pregnancy remains unclear, although hormonal and psychological influences appear to contribute.26,27 Gastric dysrhythmias have been documented by electrogastrography (see Chapter 48). The

symptoms may occur even before a woman realizes that she is pregnant; therefore, a pregnancy test must be obtained in any fertile woman with a complaint of nausea and vomiting.28 Nausea and vomiting tend to occur primarily, although not exclusively, in the morning, before food is ingested. The symptoms may warrant pharmacotherapy to alleviate the discomfort they produce but must be regarded as a normal manifestation of pregnancy.29 The prognosis for mother and child is excellent. Drugs that may be used safely to treat nausea and vomiting during pregnancy, as based on published data, include vitamin B6, ondansetron and related 5-HT3 antagonists,30 metoclopramide, and doxylamine, an antihistamine with antiemetic properties available in some European countries.31 Other antiemetics also may be safe, but specific evidence in support of their use is not available. Ancillary nonpharmacologic measures may be helpful.32,33 Hyperemesis gravidarum refers to unusually severe nausea and vomiting that leads to complications (e.g., dehy­ dration, electrolyte imbalance, malnutrition). Multiparous overweight women are at increased risk.34 The syndrome appears to represent an exaggeration of the common nausea and vomiting of pregnancy, and hormonal and psychologi­ cal factors also are thought to contribute to the pathogenesis. Hyperthyroidism has been reported in some affected persons. The manifestations generally develop in, and may continue beyond, the first trimester. Fluid and electrolyte replacement therapy may be required, together with anti­ emetic drugs. Glucocorticoids, erythromycin, and powdered ginger root have been reported to be helpful in patients with hyperemesis gravidarum. Behavioral modification and other psychotherapeutic techniques have been reported to be helpful as well. Occasionally, enteral or parenteral nutrition may need to be prescribed to prevent severe malnutrition.35 Patients with hyperemesis gravidarum, however, do not have an increased risk of toxemia of pregnancy or spontane­ ous abortion, and the condition does not lead to an increased rate of adverse fetal consequences.36 Severe vomiting may accompany acute fatty liver of preg­ nancy, a serious but uncommon condition that occurs in the third trimester of pregnancy (in contrast to hyperemesis gravidarum).37 Headache, general malaise, and manifesta­ tions of preeclampsia (hypertension, edema, proteinuria) are common accompanying features. Progression to hepatic failure and disseminated intravascular coagulation may occur rapidly. Therefore, measurement of serum liver bio­ chemical test levels is advisable in women in whom severe nausea and vomiting develop late in pregnancy. The detec­ tion of elevated serum aminotransferase levels may warrant a liver biopsy, which characteristically discloses microve­ sicular steatosis. The differential diagnosis of acute fatty liver of pregnancy includes fulminant viral hepatitis and drug-induced hepatitis. If the diagnosis of acute fatty liver is confirmed, the pregnancy should be terminated immediately to prevent maternal and fetal death (see Chapter 38).

FUNCTIONAL VOMITING

Consensus criteria for functional vomiting by the Rome III Committee on Functional Gastrointestinal Disorders include one or more episodes of vomiting per week for 3 months, with the onset of symptoms at least six months prior to diagnosis. Eating disorders, rumination, self-induced vomit­ ing, major psychiatric disorders, chronic cannabinoid use, and organic causes of vomiting (i.e., with a definable struc­ tural or physiologic basis) should be excluded (see Chapters 8 and 21).38

201

202

Section III  Symptoms, Signs, and Biopsychosocial Issues Epidemiologic studies have suggested that occasional vomiting is not uncommon in otherwise healthy persons. Specifically, population-based data indicate that vomiting once a month or more occurs in approximately 2% to 3% of the general population. Only a small minority of these persons probably fulfill the criteria for functional vomiting. The diagnostic evaluation of a patient with suspected functional vomiting should be directed toward excluding the usual organic causes of vomiting. Special motility tests are usually necessary to differentiate functional vomiting from gastroparesis or intestinal pseudo-obstruction. In our experience, a radionuclide gastric emptying test and gastrointestinal manometry are useful, because a normal or minimally affected gastric emptying test result excludes gastroparesis. Unfortunately, the reverse is not true because gastric emptying may be abnormally prolonged when the test is performed in a patient with severe nausea of any cause. If gastrointestinal manometry is performed in a patient with protracted vomiting, the detection of strong antral phasic waves and a normal intestinal pressure pattern, during fasting and postprandially, makes the diagnosis of unrecognized intestinal pseudo-obstruction almost unten­ able. Electrogastrography, another technique that has been applied to the evaluation of gastrointestinal dysmotility, may be of limited usefulness in patients with otherwise unexplained vomiting because the test may not establish whether gastric dysrhythmias are a cause or consequence of nausea and vomiting; furthermore, gastric dysrhythmias may occur in patients with a normal gastric emptying rate. Vomiting is an uncommon manifestation of gastroeso­ phageal reflux, which may or may not be detected at endos­ copy, depending in part on whether or not esophagitis is present. The definition of functional vomiting excludes major psy­ chiatric disorders. Underlying anxiety or depression, however, often plays an important role in the patient’s illness and may need to be addressed. Nutritional deficien­ cies and metabolic imbalances, if present, must be cor­ rected, but antiemetic medications tend to be ineffective in these patients. Specific dietary therapy adds little to man­ agement, because patients are already likely to avoid offend­ ing foods that may worsen their symptoms. Psychosocial support is essential, and reports suggest that cognitive and social skills training may be helpful. Psychotherapy, behav­ ioral therapy, and psychotropic agents all are used in prac­ tice, even in the absence of formal studies demonstrating their efficacy.

CYCLIC VOMITING SYNDROME

First recognized in the 19th century, cyclic vomiting syn­ drome is characterized by clustered episodes of vomiting that last from one day to three weeks (average, six days). The vomiting episodes tend to be stereotypical, with a pre­ dictable onset and duration separated by asymptomatic or almost asymptomatic intervals that range from two weeks to six months; sometimes, mild to moderate dyspeptic symptoms persist between episodes of vomiting. Some patients describe a prodromal phase resembling that associ­ ated with a migraine.39 The Rome III committee’s definition of cyclic vomiting syndrome requires three or more discrete episodes of vomiting (with no apparent explanation) during the preceding year.38 A personal or family history of migraine is supportive of the diagnosis of cyclic vomiting syndrome, particularly in children. Also, in the pediatric age group, various mito­ chondrial, ion channel, and autonomic disorders have been

associated with intermittent episodes of vomiting and may need to be excluded. Similarly, food allergy (sensitivity to cow’s milk, soy, or egg white protein) or food intolerances (to chocolate, cheese, nuts, or monosodium glutamate) may manifest with vomiting spells and should be excluded (see Chapter 9). Cyclic vomiting syndrome may occur in adults of any age, although the disorder is uncommon in older adults. There is no gender predilection. A history of migraine headaches is elicited in only one fourth of patients, and abdominal pain may be an accompanying feature in two thirds of affected persons.40 Transient fever and diarrhea also may occur. In some women, the vomiting episodes are linked to the menstrual cycle. Although cyclic vomiting syndrome has features that suggest an episodic central nervous system disorder, such as migraine or cluster headaches, studies have suggested that a high percentage of these patients have underlying intestinal motor disturbances.41 An association between chronic cannabis abuse and cyclic vomiting has been described.42 A useful diagnostic feature is the asso­ ciated urge to take hot baths or showers during the active phase of the illness. Patients who discontinue cannabis recover completely. Diagnostic evaluation of cyclic vomiting should proceed along the lines described for chronic vomiting, with an emphasis on excluding neurologic diseases, chronic partial small bowel obstruction, and disordered gastric emptying. If gastrointestinal manometric studies are abnormal, a lapa­ roscopic full-thickness biopsy of the small bowel should be considered to diagnose genetic and acquired myogenic or neurologic causes of chronic intestinal pseudo-obstruction. Testing for mitochondrial disorders (chiefly mitochondrial neurogastrointestinal encephalopathy [MNGIE]; see Chapter 35) and food allergies or intolerances (see Chapter 9) should be considered as well. The psychological aspects of cyclic vomiting syndrome require special consideration. Clinicians should refrain from the temptation to attribute cyclic vomiting to purely psychological factors. Careful studies have shown that only one in five adult patients with cyclic vomiting syndrome has an anxiety disorder or other psychiatric disease. Patients may note, however, that tension and stress precipitate epi­ sodes of vomiting. The treatment of cyclic vomiting syndrome is mostly empirical; formal therapeutic trials have not been con­ ducted. Dehydration and metabolic complications may require admission to the hospital and intravenous corrective measures. Conventional antiemetics are used but rarely abort an episode of vomiting. That some patients have a personal or family history of migraine headaches has stimulated the use of antimigraine drugs, especially serotonin 5-HT1 agonists (e.g., sumatrip­ tan), given by a subcutaneous, transnasal, or oral route. Such drugs are relatively contraindicated in patients with a history of ischemic heart disease, ischemic stroke, and uncontrolled hypertension. Similarly, beta receptor block­ ers such as propranolol have been used as preventive therapy and reportedly have helped some patients by reduc­ ing the frequency of or abolishing vomiting spells.43 Antidepressants, serotonin reuptake inhibitors, or tricyclics also have been used, although evidence from clinical trials is lacking.44 Other agents that have been reported anecdotally to help include cyproheptadine, naloxone, carnitine, valproic acid, and erythromycin. Even though habitual cannabis abuse may apparently induce cyclic vomiting syndrome, other reports have emphasized the therapeutic value of marijuana smoking in patients with the syndrome.40

Chapter 14  Nausea and Vomiting SUPERIOR MESENTERIC ARTERY SYNDROME

Although some objective basis exists for the superior mes­ enteric artery (SMA) syndrome, the diagnosis tends to be applied inappropriately to patients with functional vomit­ ing or cyclic vomiting syndrome, who then unfortunately are subjected to unnecessary surgery.45,46 The SMA branches off the aorta at an acute angle, travels in the root of the mesentery, and crosses over the duodenum. usually just to the right of the midline. In some persons, possibly because the angle between the aorta and the SMA is or becomes more acute than normal, the duodenum is partially obstructed and the patient becomes symptomatic, usually when precipitating factors accentuate the vascular compres­ sion of the duodenum. Such precipitating factors include increased lordosis (as may occur with use of a body cast), loss of abdominal muscle tone, rapid weight loss, and abdominal surgery followed by prolonged bed rest. A some­ what analogous situation has been described in conjunction with ulcer disease, pancreatitis, or other intra-abdominal inflammatory conditions that may compress the mesenteric vessels. Symptoms associated with the SMA syndrome include epigastric fullness and pressure after meals, nausea and vomiting (often bilious because the obstruction occurs distal to the ampulla of Vater), and midabdominal pain. Some patients obtain relief from adopting a prone or knee-chest position. The diagnosis is supported by imaging tests (upper gas­ trointestinal barium contrast study or computed tomogra­ phy [CT] scan), which show dilatation and stasis proximal to the duodenum where the SMA crosses (Fig. 14-2). The appearance may be misleading, however, because duodenal dilatation may be caused by atony rather than mechanical

Figure 14-2.  Upper gastrointestinal series showing sharp cutoff of contrast material at the third portion of the duodenum (secondary to compression by the superior mesenteric artery) in a patient with the superior mesenteric artery syndrome. (Courtesy of Dr. Mark Feldman, Dallas, Tex.)

obstruction.46 As noted, the SMA syndrome is often overdi­ agnosed. Before surgical correction is considered, stasis proximal to the site of duodenal obstruction should be dem­ onstrated on contrast studies and, in some cases, scinti­ graphic tests. In specialized centers, intestinal manometry may be performed and demonstrates characteristic patterns that distinguish mechanical obstruction from a motility dis­ order. Finally, a feeding catheter should be passed across the obstruction into the proximal jejunum (with endoscopic assistance, if required) to demonstrate that vomiting does not occur when the obstruction is bypassed and, if neces­ sary, to replete the patient’s nutritional status. If the SMA syndrome has been precipitated by any of the factors indicated earlier, the precipitating factor should be corrected, if possible. If the syndrome has developed acutely, patience is required because the condition may self-correct with gastric decompression combined with intravenous fluid replacement. Only infrequently, in wellinvestigated patients with chronic relapsing episodes of SMA syndrome, should corrective surgery be undertaken. The surgical technique recommended most commonly is a proximal duodenojejunostomy; a gastrojejunostomy may not be effective because the proximal duodenum is not decompressed by this approach.

RUMINATION SYNDROME

Rumination resembles vomiting but does not involve an integrated somatovisceral response coordinated by the emetic center. Rather, it consists of the repetitive effortless regurgitation of small amounts of recently ingested food into the mouth followed by rechewing and reswallowing or expulsion.47,48 Characteristically, nausea and autonomic manifestations (e.g., hypersalivation, cutaneous vasocon­ striction, sweating) that usually accompany vomiting are absent. In many ruminators, the process begins while the person is eating or immediately following completion of a meal. In some ruminators, rumination ceases when the regurgitated material becomes noticeably acidic. Others continue to ruminate for hours, however. In infants, in whom rumination was first described, rumination is rela­ tively common, and typically develops between three and six months of age. The rumination process occurs without apparent distress to the ruminator and ceases when the baby is distracted by other events or sleeps, but undernu­ trition and dehydration, which can lead to serious com­ plications, may occur. In adults, rumination occurs in men and women with equal frequency and at any age. According to the Rome III committee,38 rumination constitutes a distinct and unique category of functional gastroduodenal disorders. The clinical significance of rumination varies. Some otherwise healthy persons ruminate frequently without con­ sidering the practice abnormal. Others, under pressure from family or friends, consult a health care provider, who may mistakenly interpret rumination as habitual vomiting. Some ruminators seek medical attention because of the concern that they are unable to control the process. Physicians unaware of rumination may mistakenly attribute the symp­ toms to gastroesophageal reflux, achalasia, or gastroparesis, and thereby delay making the correct diagnosis and institut­ ing appropriate management. Alternatively, rumination fre­ quently is associated with heartburn, epigastric discomfort, and changes in bowel habits in patients who have concomi­ tant gastroesophageal reflux disease, functional dyspepsia, or irritable bowel syndrome, respectively. Weight loss may occur and suggest a possible eating disorder. The diagnosis of rumination involves several steps. First, typical features of rumination, as described earlier,

203

204

Section III  Symptoms, Signs, and Biopsychosocial Issues should be confirmed by careful history taking. Second, organic diseases, chiefly achalasia or other esophageal motility disorders, gastric outlet obstruction, and gastropa­ resis, should be excluded (see earlier). Detection of esopha­ gitis at endoscopy does not exclude rumination. Third, diagnostic tests for rumination can be performed (see later). Fourth, the coexistence of rumination with another func­ tional disorder should be considered. Rumination may be diagnosed in most patients by its typical clinical features. In equivocal cases, however, the diagnosis may be confirmed by combined upper gastroin­ testinal manometry and 24-hour ambulatory esophageal pH testing. The study may show rapid oscillations in esopha­ geal pH induced by the repeated regurgitation and reswal­ lowing of gastric contents. These oscillations typically cluster in the first one or two hours after ingestion of a meal. More definitive evidence of rumination is provided by the concurrence of declines in esophageal pH and sharp phasic pressure spikes recorded in the antrum and duodenum on manometry.48 The spikes correspond to abrupt increments in intra-abdominal pressure as the patient forces subdia­ phragmatic intragastric content toward the esophagus through a relaxed lower esophageal sphincter. The fre­ quency of positive manometric findings in ruminators, however, may not be high. In one study, only one third of patients showed an abnormal manometric study, with char­ acteristic features.49 An association between rumination and anorexia nervosa or bulimia has been reported. In one study, 20% of patients with bulimia were found to ruminate, although they tended to expel rather than reswallow the regurgitated portion of the meal. In patients with bulimia, rumination may be a learned behavior used for controlling weight without resorting to (or in addition to) frank vomiting (see Chapter 8). The pathophysiology of rumination syndrome has been elucidated only partially. Most likely, rumination repre­ sents an adaptation of the belch reflex. The spurting retro­ grade movement of gastric content probably occurs through learned transient relaxation of the lower esophageal sphinc­ ter in combination with a voluntary increase in intraabdominal pressure. Alternatively, the diaphragmatic crura may relax, thereby allowing the normal postprandial increase in intragastric pressure to overcome the resistance of the lower esophageal sphincter. Ruminators require sig­ nificantly lower fundic pressures to induce lower esopha­ geal sphincter relaxation and have increased sensitivity to balloon distention of the proximal stomach as compared with healthy control subjects. The treatment of rumination involves several steps. Patients with heartburn and endoscopic evidence of esopha­ gitis should be treated with a proton pump inhibitor. Reas­ surance and careful explanation of the phenomenon may permit some patients to control rumination on their own. Behavior modification is the most effective therapy and may be accomplished by teaching the patient special diaphrag­ matic breathing techniques. The rumination behavior is eliminated by these habit reversal techniques, because rumination and the competing response (diaphragmatic breathing) cannot be performed at the same time.

DIAGNOSTIC EVALUATION ACUTE VOMITING

There are a number of diagnostic tests that can be used.

Basic Tests

As noted earlier, the evaluation of a patient with acute vomiting should begin with a carefully obtained history and a physical examination that focuses on the patient’s volume status. An algorithm for the management of the patient with acute vomiting is shown in Figure 14-3. A urine pregnancy (human chorionic gonadotropin) test should be performed in all women of childbearing potential with acute vomiting. Routine blood studies should include a complete blood count, tests of kidney function, thyroid function tests, liver biochemical tests, electrolyte, glucose, and serum amylase and lipase levels and, in some cases, arterial blood gases to assess the patient’s acid-base status.

Imaging Tests

Plain abdominal radiographs, lying and standing, should be obtained. If the films suggest small bowel obstruction, further testing to ascertain the cause of obstruction (includ­ ing abdominal exploration) should be undertaken (see Chapter 119). If plain abdominal films are negative, additional tests can be considered. Upper endoscopy can be performed to look for mucosal lesions, ulcer, neoplasia, and gastric outlet or duodenal obstruction. CT of the abdomen can be performed to look for painless appendicitis, acute intestinal ischemia, obstruction, or pseudo-obstruction. Ultrasound may substi­ tute for CT but tends to be less revealing in this setting. Magnetic resonance imaging of the brain can be performed to look for a mass lesion or other neurologic causes of vomiting.

Additional Tests

If these test results are negative, further testing can be considered, including blood levels for drugs and toxins— specifically, digoxin, opiates, theophylline, ethanol, and carbamazepine—cultures of blood or body fluids if an infection is suspected, analysis of the cerebrospinal fluid following lumbar puncture, and serologic tests for viral hepatitis, if indicated. If appropriate, blood levels of corti­ sol, corticotropin-releasing factor, and catecholamines can be determined.

CHRONIC VOMITING

As noted, a detailed clinical history and careful physical examination (primarily to exclude other diagnoses) are central to the diagnosis of functional dyspepsia, functional vomiting, cyclic vomiting syndrome, and rumination syn­ drome. Upper gastrointestinal endoscopy or an upper gas­ trointestinal barium study, and often both, are the tests of choice for partial gastric outlet obstruction and partial duo­ denal obstruction. CT of the abdomen is also useful for establishing the presence of partial intestinal obstruction secondary to an intrinsic intestinal lesion or an intraabdominal disease that can cause intestinal obstruction. CT provides information on the degree of bowel dilatation, thickness of the bowel wall, and point of transition of the caliber of the intestinal lumen. Intra-abdominal masses, as well as retroperitoneal pathology (e.g., pancreatitis, appen­ dicitis, peritonitis, infarction), can be detected by CT. In contrast, plain radiographs of the abdomen often are unreli­ able, particularly in the presence of fluid-filled loops of bowel. A barium contrast study of the upper gastrointestinal tract and small intestine may be performed after CT to iden­ tify the site of partial obstruction more precisely or to provide an estimate of the gastrointestinal transit time. Barium contrast studies may suggest a diagnosis of achala­ sia, gastroparesis (missed by endoscopy), or neoplasm. In

Chapter 14  Nausea and Vomiting Obtain relevant medical history, basic blood tests, and pregnancy test, if applicable

Chemotherapy, radiotherapy, or surgery?

Suspicion of gastrointestinal or systemic infection?

Suspicion of drugor toxin-induced emesis?

Yes

Neurologic or vestibular manifestations?

Electrolyte or glucose imbalance?

Yes

Standard management with 5-HT3 antagonist, glucocorticoids

Yes Correct metabolic derangements. Consider testing for adrenal insufficiency Antiemetic agent, e.g., metoclopramide 0.1-1 mg/kg/6 hr intravenously

Remove offending agent. If uncertain, perform toxicology screen or measure drug level. Central antiemetics

Yes

No

MRI/CT of the brain, other neurologic and ENT studies, if necessary

Confirm by cultures, serologic testing, imaging studies, as appropriate. Antiemetic agent, e.g., metoclopramide 0.1-1 mg/kg/6 hr intravenously Gastrointestinal obstruction suspected?

No

Yes

No

No

Motion sickness: antihistamine or muscarinic M1 blockers Other neurologic disorder : central antiemetics

Yes

Yes Abdominal CT, upper endoscopy, or UGI series

Investigate possible motility disorder, other less common causes

No

Mechanical obstruction confirmed?

Yes

Specific treatment

Figure 14-3.  Algorithm for the management of a patient with acute vomiting. Possible treatments are italicized. CT, computed tomography; ENT, ear, nose, and throat; 5-HT, 5-hydroxytryptamine; MRI, magnetic resonance imaging; UGI, upper gastrointestinal.

occasional cases, an enteroclysis study (where still per­ formed), in which barium is infused directly into the small bowel via a nasoduodenal tube, may detect abnormalities missed on conventional barium studies. A higher diagnostic yield may be obtained by CT enterography. This radiologic procedure makes use of thin CT sections and large amounts of an oral neutral enteric contrast to allow better resolution of intestinal wall morphology and evaluation of individual loops of intestine without superimposition of the loops.50 Magnetic resonance enterography is an alternative to CT enterography and has the advantage of not exposing patients to radiation.51 Magnetic resonance imaging of the head is used to diagnose central nervous system lesions that may cause vomiting, including slow-growing tumors, hydrocephalus, and inflammatory, vascular, and ischemic lesions. Motility tests are useful for evaluating motor disorders, such as gastroparesis and chronic intestinal pseudo-

obstruction, that are relatively uncommon but important causes of nausea and vomiting. Various tests are available (see Chapters 48, 97, and 120).

Esophageal Manometry

Esophageal manometry is used to assess the motor activity of the esophagus. Patients with esophageal motility disor­ ders occasionally may present with vomiting. Achalasia may produce pseudovomiting and progress unrecognized for years. Similarly, manometry may detect diffuse esopha­ geal spasm and other motor disturbances of the smooth muscle portion of the esophagus that may present with or without characteristic symptoms (see Chapter 42).

Measurement of Gastric Emptying

Radioscintigraphy is the preferred and most accurate method of assessing gastric emptying. Ideally, dual markers (one for solids and one for liquids) should be used and the

205

206

Section III  Symptoms, Signs, and Biopsychosocial Issues test performed with a dual-headed gamma camera. Alterna­ tive methods of assessing gastric emptying include gastric ultrasound to assess emptying of a liquid meal and the 13C breath test with octanoid acid, a fatty acid that is labeled with a stable isotope and incorporated into a test meal. The rate at which 13CO2 is exhaled reflects the rate of gastric emptying and subsequent duodenal absorption of the lipid marker (see Chapter 48).

Cutaneous Electrogastrography

Electrogastrography (EGG) with cutaneously placed elec­ trodes identifies dysrhythmia (e.g., bradygastria or tachygas­ tria) of the gastric pacemaker and changes in the frequency of pacemaker activity in response to feeding. Potential advantages of this test are its noninvasiveness and relative simplicity. Disadvantages include its unreliability because of poor signal and artifact and a lack of correlation with clinical symptoms; identified abnormalities may or may not be related to the patient’s symptoms. Certain EGG anomalies also may be secondary to nausea rather than the cause of nausea, although this issue is still subject to debate (see Chapter 48).

Gastrointestinal Manometry

Gastrointestinal manometry is probably the most reliable physiologic test for assessing motor disturbances of the upper gastrointestinal tract. Intraluminal pressure changes are recorded via a pressure-sensitive catheter at sites located in the antrum and small bowel. The test, however, is cum­ bersome, expensive, and technically challenging to perform and is available at only a few centers that specialize in gastrointestinal motility disorders. Manometry may distinguish myogenic from neurogenic forms of pseudoobstruction and may help detect partial small bowel obstruc­ tion on the basis of wave pattern analysis (see Chapters 48, 119, and 120).

Autonomic Function Tests

Autonomic function tests can be used to assess sympathetic function, as with the tilt table test (an orthostatic challenge to blood pressure and cardiac rate regulation) and the cold hand test (a pain reflex test in which the hand is immersed in cold water to produce vasoconstriction and, normally, a significant increase in systolic arterial pressure). Parasym­ pathetic function also may be assessed by measuring varia­ tions in the RR interval on the electrocardiogram in response to bradycardia induced by deep respiration (via the vasova­ gal reflex) and by a voluntary Valsalva maneuver. The results of such tests can help distinguish visceral autonomic neuropathies (e.g., caused by amyloidosis or diabetes mellitus) from a central autonomic disorder (e.g., ShyDrager syndrome, pandysautonomia).

COMPLICATIONS Vomiting, particularly when protracted or recurring, can lead to a number of potentially life-threatening complications.

EMETIC INJURIES TO THE ESOPHAGUS AND STOMACH

Chronic protracted vomiting often produces esophagitis. The endoscopic severity may range from mild erythema to erosions and ulcerations. Characteristically, the esophagitis extends uniformly throughout the body of the esophagus,

as opposed to esophagitis associated with gastroesophageal reflux disease, which tends to be more pronounced distally. Patients often experience heartburn or retrosternal pain after an acute bout of vomiting. By contrast, patients with chronic vomiting rarely complain of chest symptoms, and the esophagitis associated with long-standing vomiting is often asymptomatic. Abrupt retching or vomiting episodes also may induce longitudinal mucosal and even transmural lacerations at the level of the gastroesophageal junction. When the lacerations are associated with acute bleeding and hematemesis, the clinical picture is described as the Mallory-Weiss syndrome (see Chapter 19). Boerhaave’s syndrome refers to spontane­ ous rupture of the esophageal wall, with free perforation and secondary mediastinitis, and carries a high mortality rate.52 It is more common in alcoholics, although esophageal rupture may develop in any person during vomiting (see Chapter 45). Multiple purpuric lesions also may appear on the face and upper neck after prolonged episodes of vomiting, probably because of repetitive increases in intrathoracic pressure and rupture of blood vessels. Dental caries and erosions may result from chronic vomiting.

SPASM OF THE GLOTTIS AND ASPIRATION PNEUMONIA

Spasm of the glottis and transient asphyxia may develop during vomiting as a result of irritation of the pharynx by acidic or bilious material. Similarly, vomiting during inser­ tion of a nasogastric tube or during endoscopy, when the patient’s consciousness is diminished, or in an older person or patient with a depressed cough reflex, may be associated with aspiration of gastric contents into the bronchi, with resulting acute asphyxia and a subsequent risk of aspiration pneumonia.53 Aspiration is more likely to occur when the stomach contains food or enteric secretions than when it is empty.

FLUID, ELECTROLYTE, AND METABOLIC ALTERATIONS

Fluid, electrolyte, and metabolic abnormalities may develop rapidly after protracted vomiting. The clinical picture is that of dehydration, hypotension, hemoconcentration, oliguria, muscle weakness, and cardiac arrhythmias. Hypochloremic alkalosis is usually the first metabolic abnormality to develop and is attributable to loss of fluid and hydrogen and chloride ions. Hypokalemia is usually present as a result of loss of potassium ions in the vomitus and renal potassium wasting because of alkalosis. Hyponatremia may occur in severe cases because of loss of sodium and release of anti­ diuretic hormone in an attempt to conserve intravascular volume. This pattern of metabolic derangements associated with chronic vomiting may alert the clinician to the possi­ bility of chronic functional or self-induced vomiting, despite the patient’s denial. A diagnostic clue is that metabolic alkalosis secondary to vomiting is typically associated with low urinary chloride excretion, which suggests extrarenal loss of chloride.

NUTRITIONAL DEFICIENCIES

Nutritional deficiencies may result from reduced caloric intake or loss of nutrients in the vomitus. Regardless of cause, nausea and vomiting may result in malnutrition, weight loss, and deficiency states that require correction (see Chapter 4).

Chapter 14  Nausea and Vomiting TREATMENT Effective management of the patient with nausea and vomit­ ing requires correction of clinically relevant metabolic com­ plications, pharmacologic therapy, and treatment of the underlying cause.

CORRECTION OF METABOLIC COMPLICATIONS

Patients with acute, severe, or repeated vomiting may become dehydrated rapidly and experience metabolic imbalances, secondary circulatory collapse, and kidney failure. If oral intake is not possible, intravenous fluids and electrolytes should be administered promptly. Adequate replacement generally consists of a normal saline solution in volumes sufficient to correct deficits (and in addition to maintenance fluids) with potassium supplementation (60 to 80 mEq/24 hr). The saline can be administered with glucose (e.g., 5% dextrose in normal saline), and in some cases a 10% glucose solution may be required. When oral intake can be resumed, glucose-containing fluids are preferred because they are easily absorbed from the intestine. A low-fat solid diet can be introduced gradually. Patients with long-standing chronic vomiting are at risk of developing malnutrition. Therefore, enteral or parenteral feeding should be considered when the patient is not able to resume adequate oral nourishment after five to eight days. Although enteral nutrition is a good option, even orogastro­ jejunal catheters placed with guidewires may be dislodged during episodes of vomiting. For long-term treatment, home parenteral nutrition may be required (see Chapter 5).

PHARMACOLOGIC TREATMENT

Drugs used to treat nausea and vomiting belong to one of two main categories, central antiemetic agents and peripheral prokinetic agents. Some drugs share both mechanisms of action, with variable predominance of one or the other.54-57

Central Antiemetic Agents

Central antiemetic agents are classified according to the predominant receptor on which the drug acts. Dopamine D2 Receptor Antagonists Benzamides.  The main antiemetic effect of benzamides (e.g., metoclopramide, clebopride) is exerted centrally in the emetic center through antagonism of the dopamine D2 receptor. These agents also stimulate peripheral 5-HT4 receptors, thereby facilitating the release of acetylcholine and acting as antroduodenal prokinetic agents.6,57 Side effects limit the use of these drugs. Metoclopramide, if administered rapidly by the intravenous route, may cause acute restlessness and anxiety. Repeated oral administration may induce somnolence in some patients. In about 1% of treated patients, distressing extrapyramidal effects, includ­ ing dystonic reactions and tremor, may appear and limit their use, particularly at high doses. Older patients are at particular risk of tardive dyskinesia.58 Metoclopramide may prolong the QT interval and thus has an arrhythmogenic potential. The most common indications for these drugs are nausea and vomiting of pregnancy, postoperative nausea and vom­ iting, and chemotherapy- and radiotherapy-induced nausea and vomiting. Because of their associated gastric prokinetic action, the drugs can be used for gastroparesis related to diabetes mellitus, prior vagotomy, and prior partial gastrec­ tomy.59-61 The standard dose of metoclopramide is 10 to 20 mg three or four times daily orally or intravenously.

Benzimidazole Derivatives.  Domperidone is the chief rep­ resentative of this class of antiemetics.62 The drug crosses the blood-brain barrier poorly and acts primarily as a peripheral dopamine D2 receptor antagonist. It blocks the receptors centrally in the area postrema (which is partly outside the blood-brain barrier) and in the stomach, where D2 receptor inhibition decreases proximal gastric relaxation and facilitates gastric emptying.63,64 Although domperidone is a weaker antiemetic than metoclopramide, it may be particularly useful for the management of nausea and vomit­ ing secondary to treatment with levodopa in Parkinson’s disease, because it antagonizes the proemetic side effects of levodopa without interfering with its antiparkinsonian action in brain centers protected by the blood-brain barrier. The standard dose is 10 to 20 mg three or four times daily orally. A review of the use of domperidone in the treatment of diabetic gastroparesis has concluded that the drug is probably useful but has not been properly evaluated by well-designed controlled trials.63 Domperidone (as well as benzamides) may increase the release of prolactin and occasionally is associated with breast tenderness and galactorrhea. Phenothiazines and Butyrophenones The phenothiazines (chlorpromazine, perphenazine, pro­ chlorperazine, promethazine, thiethylperazine) and butyro­ phenones (droperidol and haloperidol) also block D2 dopaminergic receptors and, in addition, block muscarinic M1 receptors. Phenothiazines also block histamine H1 receptors.65,66 These drugs tend to induce relaxation and somnolence and are generally used parenterally or as sup­ positories in patients with acute intense vomiting of central origin, as occurs with vertigo, migraine headaches, and motion sickness. They are also useful for patients with vom­ iting secondary to toxic agents and chemotherapy and after surgery.66,67 Droperidol also has been used as an adjunct to standard sedation during endoscopic procedures and, in combination with morphine, is used to reduce postopera­ tive pain, nausea, and vomiting.67,68 Safety concerns, includ­ ing common extrapyramidal effects, however, have limited the use of all these agents.69 Antihistamines and Antimuscarinic Agents Antihistamines and antimuscarinic agents act primarily by blocking histamine H1 receptors (cyclizine, diphenhy­ dramine, cinnarizine, meclizine, hydroxyzine) and musca­ rinic M1 receptors (scopolamine) at a central level.70 Promethazine belongs to the phenothiazine class but acts as an antihistaminic H1 and antimuscarinic agent with strong sedative properties. Cyclizine and diphenhydrinate are used commonly to treat motion sickness and have been shown to decrease gastric dysrhythmia. Therefore, their antiemetic effect may be mediated in part by their periph­ eral action. A standard antiemetic dose of cyclizine is 50 mg, given three times daily orally or as a 100-mg sup­ pository. The main indication is nausea and vomiting asso­ ciated with motion sickness and vestibular disease. Cyclizine is useful for postoperative and other forms of acute vomiting.71 Some of these drugs are also used as anti­ pruritic agents. Drowsiness is the major limiting side effect, particularly for the older agents, but this effect may be advantageous in the treatment of acute vomiting. The anti­ cholinergic effects are potentially troublesome in patients with glaucoma, prostatic hyperplasia, and asthma. Serotonin Antagonists Serotonin 5-HT3 receptor antagonists (ondansetron, granis­ etron, dolasetron, tropisetron) are potent antiemetics that

207

208

Section III  Symptoms, Signs, and Biopsychosocial Issues selectively block 5-HT3 receptors in the emetic center and in gastric wall receptors that relay afferent emetic impulses through the vagus nerve.72 In addition to their antiemetic effect, they have a modest gastric prokinetic action.73 The main indication for this class of drugs is nausea and vomit­ ing associated with chemotherapy or radiation therapy or following surgery.5,74-76 Ondansetron appears to be safe in pregnancy.77 Ondansetron may be given as a single dose of 8 to 32 mg, intravenously in a dose of 0.15 mg/kg every eight hours, or orally in a dose of 12 to 24 mg every 24 hours, in three divided doses. Headache is a common side effect. Glucocorticoids The antiemetic mechanism of glucocorticoids is not well understood. It may relate to inhibition of central prostaglan­ din synthesis, release of endorphins, or altered synthesis or release of serotonin. The principal indication is treatment of nausea and vomiting in the postoperative period or as a result of chemotherapy or radiation.78-80 Glucocorticoids also may be used to reduce cerebral edema and hence allevi­ ate vomiting secondary to increased intracranial pressure. Dexamethasone is the formulation used acutely, in doses ranging from 8 to 20 mg intravenously and 4 mg every six hours orally. Side effects are uncommon because treatment is usually administered for short periods. In diabetic patients, however, careful monitoring of blood glucose levels is required. In patients with a history of peptic ulcer or with a gastroenteric anastomosis, concurrent administra­ tion of a gastric antisecretory agent is advisable. In practice, dexamethasone often is used in combination with another antiemetic agent, such as metoclopramide or a 5-HT3 antagonist.81 Cannabinoids Synthetic cannabinoids are becoming part of the standard therapeutic ornamentation.82 Two oral formulations are available, nabilone and dronabinol. Both are approved by the U.S. Food and Drug Administration (FDA) for use in chemotherapy-induced nausea and vomiting refractory to conventional antiemetic therapy. The combination of a dopamine antagonist and a cannabinoid may be particularly effective in preventing nausea that has a major negative impact on a patient’s quality of life.83,84 Mood-enhancing properties make cannabinoids attractive to patients, but these drugs are potentially more toxic than conventional antiemetic agents. Hypotension and psychotropic reactions are relatively common side effects. These drugs should be used with caution in older adults and in patients with a history of mental illness.85,86 Neurokinin-1 Receptor Antagonists NK-1 receptor antagonists, which inhibit substance P/NK-1, are potent antiemetic agents. Two formulations are avail­ able, aprepitant (oral) and fosaprepitant (parenteral), and several others (e.g., casopitant, vestipitant, netupitant) are undergoing evaluation. These drugs appear to provide better protection against postoperative vomiting but not nausea when compared with 5-HT3 antagonists.87 NK-1 receptor antagonists may be particularly useful when combined with other drugs such as 5-HT3 antagonists and dexamethasone and are approved by the FDA for use in preventing vomiting in patients undergoing cancer chemotherapy.88 Adjuvant Agents and Therapies Patients with acute nausea and vomiting associated with chemotherapy, radiotherapy, and surgery often have anxiety, which may exacerbate their symptoms. Therefore, the anx­

iolytic effects of benzodiazepines such as lorazepam and alprazolam may potentiate the antiemetic action of agents such as 5-HT3 receptor antagonists and glucocorticoids that are devoid of psychotropic effects. Acupuncture, acustimu­ lation, and acupressure also have been shown to decrease the nausea associated with motion sickness induced by illusory self-motion and nausea associated with cancer chemotherapy.89-91

Gastric Prokinetic Agents

Serotonin 5-HT4 Receptor Agonists Cisapride and cinitapride are drugs in the benzamide class that share the peripheral 5-HT4 agonist effect of metoclo­ pramide (also a benzamide) without the dopamine D2 antagonist action that is primarily responsible for the poten­ tially troublesome central side effects of metoclopramide. Although cisapride and cinitapride lack central depressant effects, they retain antiemetic properties because of some 5-HT3 properties.92 Cisapride is a potent gastric prokinetic agent at doses of 5 to 20 mg three to four times daily in adults. Dosing adjustments are not needed in older adult patients. Unfortunately, cisapride carries a significant risk of precipitating serious cardiac ventricular arrhythmias, especially in patients concomitantly taking drugs that prolong the QT interval.92 Thus, cisapride has been with­ drawn from the market in many countries, although in others, including the United States, it may still be pre­ scribed with certain restrictions. Cinitapride is analogous to cisapride, but at a dose of 1 mg orally three times daily has not been associated with cardiac arrhythmias. It is not yet available in the United States.93 Tegaserod, a partial 5-HT4 agonist with prokinetic action, was considered potentially useful for the treatment of gas­ troparesis and functional dyspepsia but had to be with­ drawn from the market because of reported cardiovascular events; it may be obtained only under exceptional circum­ stances.94 The main indication for 5-HT4 agonist drugs is the management of nausea and vomiting associated with gastroparesis, intestinal pseudo-obstruction, and functional dyspepsia.95-97 Motilin Receptor Agonists Motilin receptor agonists include the antibiotic erythromy­ cin and other agents—none of which is commonly avail­ able—that act as motilin receptor ligands on smooth muscle cells and enteric nerves. The pharmacodynamic effects in humans are dose-dependent. At low doses (0.5 to 1 mg/kg as an intravenous bolus), erythromycin induces sweeping gastric and intestinal peristaltic motor activity that resem­ bles phase III of the interdigestive migrating motor complex but may empty the stomach inefficiently (see Chapters 48 and 97).98 At higher doses of 200 mg intravenously used in clinical practice, antral activity becomes intense and empties the stomach rapidly, although the burst of motility does not always migrate down the small intestine.99,100 A simultaneous increase in small bowel contractions may induce abdominal cramps and diarrhea. Curiously, when used clinically as an antibiotic, erythromycin may cause nausea and vomiting. In clinical practice, erythromycin may be used to treat acute nausea and vomiting associated with gastroparesis (diabetic, postsurgical, or idiopathic)92,100 and to clear the stomach of retained food, secretions, and blood prior to endoscopy. Erythromycin may be administered intrave­ nously in boluses of 200 to 400 mg every four to five hours. The lower doses are more appropriate for patients with pseudo-obstruction, which is associated with reduced inter­ digestive sweeping motor activity in the small bowel.

Chapter 14  Nausea and Vomiting Erythromycin is not suitable for prolonged treatment, because its efficacy by the oral route is uncertain and its inherent antibiotic properties carry the potential risk of complications, including pseudomembranous colitis. New synthetic motilin agonists devoid of antibiotic activity are in development. Ghrelin is a peptide structurally and functionally related to motilin that acts to accelerate postprandial gastric emptying. Ghrelin receptor agonists may have a future therapeutic role as prokinetic agents for the treatment of gastroparesis.101,102

GASTRIC ELECTRICAL STIMULATION

Gastric pacing involves application of high-energy currents that entrain the gastric slow waves and generate phasic contractions. This technique may achieve correction of gastroparesis and amelioration of nausea and vomiting but requires external energy sources that limit the autonomy of the patient and is not practical. An alternative approach, gastric neurostimulation, delivers brief low-energy impulses to the stomach with the use of an implantable neurostimu­ lator similar to devices used to control chronic pain. The gastric neurostimulator does not entrain slow wave activity and produces little or no acceleration of gastric emptying, but it appears to relax proximal gastric tone and to produce significant improvement in nausea and vomiting, as well as the patient’s nutritional status.103 The device is approved for humanitarian use only by the FDA. Gastric neurostimula­ tion, however, is not without risk, and commonly reported complications include electrode dislodgement, infection, and bowel obstruction. The device may be recommended with caution for long-standing (at least one year’s duration) refractory gastroparesis.104,105

KEY REFERENCES

Abell TL, Adams KA, Boles RG, et al. Review article: Cyclic vomiting syndrome in adults. Neurogastroenterol Motil 2008; 20:269-84. (Ref 40.)

Allen JH, de Moore GM, Heddle R, Twartz JC. Cannabinoid hypereme­ sis: Cyclical hyperemesis in association with chronic cannabis abuse. Gut 2003; 53:1566. (Ref 42.) Apfel CC, Korttila K, Abdalla M, et al. IMPACT Investigators. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004; 350:2441-51. (Ref 21.) Einarson A, Maltepe C, Navioz Y, et al. The safety of ondansetron for nausea and vomiting of pregnancy: A prospective comparative study. BJOG 2004; 111:940-3. (Ref 77.) Flake ZA, Scalley RD, Bailey AG. Practical selection of antiemetics. Am Fam Physician 2004; 69:1169-74. (Ref 54.) Gourcerol G, Leblanc I, Leroi AM, et al. Gastric electrical stimulation in medically refractory nausea and vomiting. Eur J Gastroenterol Hepatol 2007; 19:29-35. (Ref 105.) Gralla RJ, Osoba D, Kris MG, et al: Recommendations for the use of antiemetics: Evidence-based, clinical practice guidelines. American Society of Clinical Oncology. J Clin Oncol 1999; 17:2971-94. (Ref 55.) Grunberg SM, Deuson RR, Mavros P, et al. Incidence of chemotherapyinduced nausea and emesis after modern antiemetics. Cancer 2004; 100:2261-8. (Ref 18.) Imperato F, Canova I, Basili R, et al. Hyperemesis gravidarum—etiology and treatment. Clin Ter 2003; 154:337-40. (Ref 34.) Jewell D. Nausea and vomiting in early pregnancy. Clin Evid 2003; Jun:1561-70. (Ref 28.) Malagelada JR. Chronic idiopathic intestinal pseudo-obstruction. Curr Treat Options Gastroenterol 2000; 3:335-40. (Ref 24.) O’Brien MD, Bruce BK, Camilleri M. The rumination syndrome: Clinical features rather than manometric diagnosis. Gastroenterology 1995; 108:1024-9. (Ref 49.) Pandolfino JE, Howden CW, Kahrilas PJ. Motility modifying agents and management of disorders of gastrointestinal motility. Gastroenterol­ ogy 2000; 2(Suppl 1):S32-47. (Ref 60.) Sewell DD, Jeste DV. Metoclopramide-associated tardive dyskinesia. An analysis of 67 cases. Arch Fam Med 1992; 1:271-8. (Ref 58.) Sugumar A, Singh A, Pasricha PJ. A systematic review of the efficacy of domperidone for the treatment of diabetic gastroparesis. Clin Gastroenterol Hepatol 2008; 6:726-33. (Ref 63.) Tack J, Talley NJ, Camilleri M, et al. Functional gastroduodenal disor­ ders. Gastroenterology 2006; 130:1466-79. (Ref 38.) Full references for this chapter can be found on www.expertconsult.com.

209

CHAPTE R

15 Diarrhea Lawrence R. Schiller and Joseph H. Sellin

CHAPTER OUTLINE Definition  211 Pathophysiology  212 Osmotic Diarrhea  212 Secretory Diarrhea  213 Complex Diarrhea  214 Clinical Classification  215 Acute versus Chronic Diarrhea  216 Large-Volume versus Small-Volume Diarrhea  216 Osmotic versus Secretory Diarrhea  216 Watery versus Fatty versus Inflammatory Diarrhea  216 Epidemiologic Features  216 Differential Diagnosis  216 Evaluation of the Patient  218 History  218 Physical Examination  219

Diarrhea is a universal human experience. The average American has an estimated 0.65 episodes of acute gastrointestinal illness per year.1 For most persons, episodes of diarrhea last a day or two and rapidly subside without medical intervention. For others, diarrhea lasts for more than a few days or is complicated by fever, prostration, or rectal bleeding. Such persons are likely to visit their physicians. Over 3.5 million outpatient visits for diarrhea occur each year.2 Most patients can be managed successfully as outpatients; however, more than 150,000 hospital admissions each year are for gastroenteritis. Over the course of a year, chronic diarrhea (liquid stools for more than four weeks) may occur in 5% of the population and is thus a major cause of disability for Americans.3 In developing countries, acute infectious diarrhea remains an important cause of morbidity and mortality, particularly among children. Diarrhea is a symptom, not a disease, and therefore may occur in dozens of conditions. The evaluation and management of diarrhea can be complex and time-consuming. Research efforts since 1970 have uncovered some fundamental mechanisms of diarrhea and thereby permit a rational approach to its diagnosis and management.

DEFINITION Most patients consider increased fluidity of stool as the essential characteristic of diarrhea.4 Stool consistency is difficult to quantitate and visual scales may be helpful for patients to use in describing their diarrhea.5 Researchers also have used stool frequency or stool weight as a surrogate marker of diarrhea. Three or more bowel movements daily are considered to be abnormal, and the upper limit of stool weight is generally agreed to be 200 g daily in Western

Further Evaluation of Acute Diarrhea  219 Further Evaluation of Chronic Diarrhea  220 Treatment  226 Empirical Therapy of Acute Diarrhea  226 Empirical Therapy of Chronic Diarrhea  226 Selected Diarrheal Syndromes  227 Irritable Bowel Syndrome and Functional Diarrhea  227 Microscopic Colitis  228 Postsurgical Diarrhea  228 Bile Acid–Induced Diarrhea  229 Diarrhea in Hospitalized Patients  229 Factitious Diarrhea  230 Idiopathic Secretory Diarrhea  231 Diarrhea of Obscure Origin  231

countries. Although stool weight often is cited as a “scientific” definition of diarrhea, diarrhea should not be defined solely in terms of fecal weight. Some persons have increased fecal weight as a result of fiber ingestion but do not complain of diarrhea because their stool consistency is normal. For example, stool output can be as great as 300 g when a high-fiber diet is consumed, as is customary in some developing countries. Conversely, about 20% of patients referred for evaluation of diarrhea may have a normal stool weight. Whether this is a result of hyperdefecation (i.e., more frequent passage of formed stool) or a change in consistency (i.e., passage of small-volume loose stools) is unclear. In a study of the objective determinants of decreased fecal consistency,4 the ability of water-insoluble fecal solids, such as those derived from dietary fiber or bacterial cell walls, to hold or bind fecal water correlated well with fecal consistency. Too little water-holding capacity to bind all the water present resulted in loose stools, but when fecal solids had sufficient water-holding capacity to bind all the water present, stools remained thick or formed. Fecal consistency correlated best with the ratio of the water-holding capacity of insoluble solids to the total amount of water present and not simply to the amount of fecal water, further supporting the concept that stool weight should not be the sole criterion for diarrhea. Fecal incontinence may be reported as “bad diarrhea” by some patients, especially older adults.6 Although many incontinent patients have loose stools, their major problem is with the mechanisms of continence and not with intestinal fluid or electrolyte absorption. Accordingly, all patients who complain of diarrhea should be asked about the presence of fecal incontinence. If incontinence is frequent, especially in the absence of rectal urgency or loose stools, the patient should be evaluated for incontinence and not for diarrhea (see Chapter 17).

211

212

Section III  Symptoms, Signs, and Biopsychosocial Issues PATHOPHYSIOLOGY Diarrhea frequently represents a protective response to a variety of intestinal insults and assaults. Normally, the intestine absorbs most of the fluid that it secretes, and intestinal motility provides a favorable milieu for water, electrolyte, and nutrient absorption. When infectious agents, toxins, or other noxious substances are present in the intestine, fluid secretion and motility are stimulated to expel the unwanted material, thereby producing diarrhea. This protective response is valuable acutely but, when chronic, it is inappropriate and no longer serves an adaptive purpose. Historically, diarrhea was thought to be primarily a motility disorder. An improved understanding of intestinal elec­ trolyte transport since 1970 has shifted the emphasis to epithelial function, rather than motility. Clearly, however, epithelial and motor functions are altered in a coordinated fashion to produce diarrhea.7 Diarrhea usually is the result of an excess of stool water rather than a decrease in the water-holding capacity of fecal solids, with the implication that water transport within the intestine is abnormal.4 Normally, the small intestine and colon absorb 99% of both oral intake and endogenous secretions from the salivary glands, stomach, liver, and pancreas— a total fluid load of approximately 9 to 10 L daily (Fig. 15-1). Diarrhea results from a disruption of this normally finetuned mechanism; reduction of water absorption by as little as 1% can result in diarrhea. Therefore, to understand the pathogenesis of diarrhea, one needs to understand normal water absorption by the intestine and the abnormalities that can impair water absorption. Water itself is not actively transported but moves across the intestinal mucosa secondary to osmotic forces generated by the transport of solutes—that is, electrolytes and nutrients (see Chapters 99 and 100). The molecular pathways of ion and nutrient transport across the mucosa have been well characterized and are regulated by a complex communication system of extracellular and intracellular messengers that maintain fluid equilibrium throughout a wide range of physiologic conditions. Normally, absorption and secretion take place simultaneously, but absorption is quantitatively greater. Either a decrease in absorption or an increase in secretion leads to additional fluid within the lumen and diarrhea. Disruption of epithelial electrolyte transport or its regulatory system by toxins, drugs, hormones, and cytokines is a major cause of diarrhea. Diarrhea resulting from disordered electrolyte transport is known as secretory diarrhea, even though it is more commonly caused by reduced absorption than by net secretion.8 Another major cause of diarrhea is ingestion of some poorly absorbed, osmotically active substance (e.g., magnesium ion, lactulose) that retains fluid within the lumen to maintain osmotic equilibration with body fluids, thereby reducing water absorption. Diarrhea resulting from this mechanism is known as osmotic diarrhea (Table 15-1). Few clinical situations produce pure secretory or osmotic diarrhea, but considering some conditions in which one or the other mechanism predominates is useful before considering more complex processes.

Upper tract

Volume absorbed Small intestine

6L

1.5 L

Jejunum 2.5 L Ileum

Colon

1.4 L

1.5 L

0.1 L Figure 15-1.  Fluid loads along the gastrointestinal tract. Each day, close to 10 L of fluid composed of ingested food and drink and secretions from the salivary glands, esophagus, stomach, pancreas, bile duct, and duodenum pass the ligament of Treitz. The jejunum absorbs approximately 6 L and the ileum 2.5 L, leaving about 1.5 L to pass into the colon each day. The colon absorbs more than 90% of this load, leaving approximately 0.1 L in the feces. Therefore, the overall efficiency of water absorption is 99%. Reduction of this efficiency by as little as 1% may lead to diarrhea. (From Schiller LR. Chronic diarrhea. In: McNally P, editor. GI/Liver Secrets. 2nd ed. Philadelphia, Pa: Hanley & Belfus; 2001. p 411.)

Table 15-1  Secretory versus Osmotic Diarrhea type of diarrhea Secretory diarrhea

CAUSES

EXAMPLES

Exogenous secretagogues Endogenous secretagogues

Enterotoxins (e.g., cholera) Neuroendocrine tumors (e.g., carcinoid syndrome) Congenital chloridorrhea Intestinal resection, diffuse intestinal mucosal disease Diffuse mesenteric atherosclerosis Intestinal hurry following vagotomy Magnesium ingestion

Absence of ion transporter Loss of intestinal surface area Intestinal ischemia

OSMOTIC DIARRHEA

Ingestion of poorly absorbed cations and anions or poorly absorbed sugars or sugar alcohols (e.g., mannitol, sorbitol) accounts for most osmotic diarrheas.9 Ions that are poorly absorbed include magnesium, sulfate, and phosphate. These ions are transported actively by mechanisms that are saturated at low intraluminal ion concentrations and passively

10 L

10 L

Rapid intestinal transit Osmotic diarrhea

Ingestion of poorly absorbed agent Loss of nutrient transporter

Lactase deficiency

Chapter 15  Diarrhea by mechanisms that are slow. Together, these processes limit total absorption to a fraction of the amount that can be ingested. Because neither the small intestine nor colon can maintain an osmotic gradient, unabsorbed ions (and their counter ions) that remain in the intestinal lumen obligate retention of water to maintain an intraluminal osmolality equal to that of body fluids (about 290 mOsm/kg). Therefore, approximately 3.5 mL of water (1000 mL/kg divided by 290 mOsm/kg) are retained for every 1 mOsm of retained ions or molecules.9-12 Sugars and sugar alcohols are the other category of substances that cause osmotic diarrhea.12 Monosaccharides, but not disaccharides, can be absorbed intact across the apical membrane of the intestine. When disaccharides such as sucrose and lactose are ingested, absence of the appropriate disaccharidase will preclude absorption of the disaccharide or its component monosaccharides (see Chapter 101). The most common clinical syndrome of disaccharidase deficiency is acquired lactase deficiency, which accounts for lactose intolerance in many adults.13 Lactase is present in the brush border of the small intestine of most immature mammals but disappears in adult mammals, including 70% of adult humans.14 The main exceptions are persons from the northern European gene pool, who typically maintain lactase activity into adult life. Lactase activity often falls with age even in this group, however. Congenital deficiency of lactase is rare and seems to be the result of a mutation in a gene distinct from that for lactase-phlorizin hydrolase (the gene affected in adult lactase deficiency).15 Acquired deficiencies also may be associated with mucosal diseases of the upper small intestine. Congenital sucrase and trehalase deficiencies are rare and prevent the adequate digestion of sucrose (table sugar) and trehalose (a sugar found in mushrooms), respectively. Lactulose is a synthetic disaccharide that cannot be hydrolyzed by the human intestine and is not absorbed intact in more than trace amounts. It thereby causes an osmotic diarrhea when given in sufficient quantity to overwhelm the metabolic capacity of colonic bacteria (about 80 g/day). The essential characteristic of osmotic diarrhea is that it disappears with fasting or cessation of ingestion of the offending substance. This characteristic has been used clinically to differentiate osmotic diarrhea from secretory diarrhea, which typically continues with fasting. Electrolyte absorption is not impaired in osmotic diarrhea, and electrolyte concentrations in stool water are usually low.10-12

SECRETORY DIARRHEA

Secretory diarrhea has many causes, and the mechanism of this type of diarrhea is always net secretion of anions (chloride or bicarbonate) or net inhibition of sodium absorption.16 The stimuli for secretion arise from the intestinal lumen, subepithelial space, or systemic circulation and substantially alter the messenger systems that regulate ion transport pathways. In some cases, congenital absence of a specific transport molecule limits sodium or chloride absorption and results in diarrhea; in others, lack of sufficient absorptive surface area limits electrolyte, particularly sodium, absorption critically. The most common cause of secretory diarrhea is infection.16 Enterotoxins from a host of infectious agents (primarily bacteria but also parasites and viruses) interact with receptors that modulate intestinal transport and lead to increased anion secretion. Enterotoxins also may block specific absorptive pathways in addition to stimulating secretion. Most enterotoxins inhibit Na+-H+ exchange in the small intestine and colon, thereby blocking one of the important driving forces for electrolyte and fluid absorption.17,18

Peptides, such as vasoactive intestinal peptide, produced by endocrine tumors, cause secretory diarrhea by stimulating secretion by epithelial cells, as do peptides released from subepithelial neurons and inflammatory cells (see Chapter 32).19 Neurotransmitters such as acetylcholine or serotonin (5-hydroxytryptamine, 5-HT) and other modulators such as histamine and inflammatory cytokines also are potent secretory stimuli.20,21 Most of these endogenous regulators of inte­stinal transport elicit diarrhea by altering intracellular messengers, such as cyclic adenosine monophosphate (cAMP), cyclic guanosine monophosphate (cGMP), and calcium, that control specific transport pathways.22 In addition, peptides and other regulators may affect the synthesis, localization, and degradation of individual transport proteins. Exogenous agents, such as drugs and some poisons, lead to secretory diarrhea, presumably by interacting with intracellular regulators or intracellular messengers of the enterocytes. The absence or disruption of a specific absorptive pathway may cause diarrhea. For example, rare congenital syndromes, such as congenital chloridorrhea and congenital sodium diarrhea, are caused by the absence of a specific transport molecule.23 In chloridorrhea, Cl−-HCO3− exchange in the ileum and colon is defective, thereby transforming chloride into a poorly absorbed ion. Diarrhea resulting from chloridorrhea can be reduced by limiting oral chloride intake or chloride secretion (i.e., by reducing gastric acid secretion with a proton pump inhibitor) or by stimulating chloride absorption in the colon by enhancing short-chain fatty acid absorption.24 Several transporter defects have been proposed for congenital sodium diarrhea.25 More commonly, a relative shift in the balance of absorptive and secretory pathways may contribute to diarrhea in clinical settings associated with epithelial injury or changes in cell proliferation. For example, viral gastroenteritis and celiac disease may disproportionately compromise transporters (e.g., disaccharidases, Na+-coupled absorption) that mediate absorption on the villous surface, whereas secretory pathways in the crypt are unchanged or increased (see Chapters 104 and 107). For intestinal fluid and electrolyte absorption to be complete, the intestine must have an adequate surface area and adequate contact time with luminal contents. Substantial loss of surface area, as in celiac disease or inflammatory bowel disease (IBD) or after resective surgery, may compromise water absorption. Even though the reserve absorptive capacity in the small intestine and colon is large, sufficiently long surgical resections inevitably cause diarrhea. In some cases, the problem is temporary, because over time the intestine may improve its capacity for absorption by the process of adaptation.26 Such compensation is impossible following resection of certain segments of the intestine with highly specific absorptive functions that simply cannot be assumed by other segments of the bowel. For example, ileocecal resection is followed by permanent inability to absorb sodium chloride against a concentration gradient27 and, if sufficient ileum is resected, by failure to absorb vitamin B12, intrinsic factor, and normal amounts of conjugated bile acids (see Chapter 103). Abnormal motility may lead to diarrhea that has secretory and osmotic components.8 For fluid and electrolyte absorption to be complete, the contact time between luminal contents and the epithelium must be sufficient to permit absorption. In some patients, abnormal motility produces intestinal “hurry.”28,29 Because rapid transit prevents adequate time for absorption, diarrhea results despite intact mucosal absorptive capacity, as measured by intestinal perfusion studies during which contact time is maximized by

213

Section III  Symptoms, Signs, and Biopsychosocial Issues

Although classification of diarrhea as osmotic or secretory may be instructive in thinking about the pathophysiology of diarrhea, cases of pure secretory or pure osmotic diarrhea are uncommon. Most clinically important diarrhea is complex in pathogenesis; rather than being produced by a single pathophysiologic mechanism, several mechanisms are involved. Causes may include the effects of substances released by enteric endocrine cells, cytokines released by local and remote immunologically reactive cells, activity of the enteric nervous system, and peripherally released peptides and hormones (paracrine, immune, neural, and endocrine systems; see Chapter 1). Further complicating the understanding of diarrhea is the recognition that certain mediators affect not only epithelial or muscle function, but also each other. For example, enteric nerves may stimulate mast cells, and products released from mast cells (particularly histamine) may alter enteric neuron functions.38 A single agonist, such as prostaglandin, may have multiple simultaneous effects on epithelial function, muscle contraction, and the paracellular pathway, thereby leading to effects on ion transport, motility, and mucosal permeability.39 Therefore, a number of modulators and

e

Neu ral

Systems

ine

COMPLEX DIARRHEA

ne mu Im

r doc En

rapid infusion of fluid into the intestine.8 In some patients with intestinal hurry, the oral-cecal transit time may be as short as 10 minutes. Under such circumstances, the diarrhea is exacerbated by malabsorption of nutrients that produces an osmotic component to diarrhea. In disorders such as diabetes mellitus and postvagotomy diarrhea, intestinal hurry has been linked to abnormal enteric nervous system function.30 In other clinical settings, such as amyloidosis, postprandial diarrhea, and irritable bowel syndrome (IBS), enteric nervous system dysfunction is suspected, but unproved.31,32 Many endocrine diarrheas, such as those caused by peptide-secreting tumors or hyperthyroidism, may lead to diarrhea not only by affecting intestinal electrolyte transport, but also by accelerating intestinal motility.33 Conversely, slow intestinal transit may lead to a secretory diarrhea by promoting small intestinal bacterial overgrowth.34,35 Excess bacteria in the small intestine disrupt digestion and may alter electrolyte transport. The best documented example of diarrhea related to this mechanism is scleroderma. Although diabetes mellitus is often suspected of causing diarrhea by slow transit and stasis, as seen in scleroderma, such a pathophysiology is not always established (see Chapter 102).36,37 Evaluation of the role of intestinal motility in the pathogenesis of diarrhea has been limited by the lack of the necessary tools to measure the interactions among motility, propulsive forces, and transit time. Except for intestinal perfusion studies, during which the effect of motility on electrolyte transport is eliminated, no methodology exists to dissociate the effects of intestinal transport and motility on net absorption.8 Thus, consensus has not been achieved on whether too much or too little motility causes diarrhea, nor on how luminal factors may alter intestinal smooth muscle function. Reduced intestinal blood flow has an important but as yet poorly defined role in diarrhea. Whether mesenteric ische­ mia has a direct effect on absorption or whether low blood flow prompts secondary responses (e.g., via cytokines or neurotransmitters) that affect fluid transport and produce a secretory diarrhea is not clear. Radiation enteritis also produces an abnormal intestinal microcirculation associated with persistent diarrhea that may be difficult to treat (see Chapters 39 and 114).

Para cri n

214

PINES

Paracellular pathway

Blood flow Epithelium

Permeability

Transport

Muscle

Motility

Metabolism

Figure 15-2.  PINES regulatory system in the intestine. The regulatory system of the intestine integrates paracrine, immune, neural, and endocrine systems and produces coordinated changes in mucosal and muscular function that permit adaptive responses to changing conditions. The regulatory system can widen or narrow the paracellular pathway that governs passive transmucosal permeability of electrolytes, accelerate or retard the transepithelial transport of nutrients and electrolytes by affecting membrane channels and pumps, alter motility by relaxing or contracting the various muscle layers in the intestine, and increasing or decreasing mucosal blood flow, thereby influencing intestinal metabolism. Diarrhea may be an appropriate response to acute infection. Maladaptive responses may be responsible for chronic diarrhea. (From Sellin JH. Functional anatomy, fluid and electrolyte absorption. In: Feldman M, Schiller LR, editors. Gastroenterology and hepatology. The Comprehensive Visual Reference, vol 7: Small Intestine. Philadelphia, Pa: Current Medicine; 1997. p 1.11.)

effectors contribute to the final clinical picture. A full appreciation of the pathophysiology of diarrhea requires consideration of a regulatory system known as PINES (paracrine, immune, neural, and endocrine system modulators; Fig. 15-2). An example of the complexity of the pathophysiology of a diarrheal syndrome is cholera. Cholera is often cited as the paradigm of a pure secretory diarrhea: cholera toxin targets the epithelial cell, increases the second messenger, cAMP, which opens apical chloride channels to stimulate chloride secretion, and results in diarrhea; however, the actual mechanism whereby cholera induces diarrhea is far more complex.40 Cholera toxin stimulates endocrine cells and neural elements that reinforce its direct secretory effect on enterocytes.41 In addition, cholera toxin causes distinct changes in intestinal motility. Other toxins produced by Vibrio cholerae target tight junctions and thereby alter mucosal permeability (Fig. 15-3; see Chapter 107).42 Another example of dysregulation of PINES is IBD.43 Diarrhea in patients with IBD involves more than just exudation into the lumen as a result of destruction of the mucosa. Intact enterocytes are barraged by multiple secretagogues released by immune cells in the intestine and by bacterial toxins that may influence enterocyte function. Although initial models of diarrhea in IBD suggested altered fluid transport driven by chloride secretion, more recent studies have demonstrated that the diarrhea in IBD is mediated by an antiabsorptive effect associated with down-regulation of

Chapter 15  Diarrhea Enterocyte

Lumen ACE V. cholerae

?

Subepithelial space and lamina propria

ZOT Brefeldin A CT

Adenylate cyclase Macrophage

Cl – +

Na+

cAMP

PG 5HT VIP

–

Cl – Myenteric plexus

Enteric neuron

EC cell

Smooth muscle cell

Figure 15-3.  Pathophysiology of cholera. Vibrio cholerae produces several toxins that interact with adenylate cyclase in the enterocyte and several elements of the regulatory system of the intestine, including enteric neurons and enterochromaffin (EC) cells, to produce a secretory state and voluminous diarrhea. In addition to the classic enterotoxin, cholera toxin (CT), the bacterium also produces zona occludens toxin (ZOT), which increases the permeability of the tight junction between enterocytes, and accessory cholera enterotoxin (ACE), which has unclear effects on enterocytes. In addition to cyclic adenosine monophosphate (cAMP) generated by adenylate cyclase in response to CT, secretory stimuli include prostaglandin (PG), serotonin (5HT), and vasoactive intestinal polypeptide (VIP) released by macrophages, EC cells, and enteric neurons. CT has been shown to activate adenylate cyclase on the basolateral membrane of the cell in experiments in which the transcellular transport of CT to the basolateral membrane was inhibited by Brefeldin A (jagged arrow). (From Sellin JH. Functional anatomy, fluid and electrolyte absorption. In: Feldman M, Schiller LR, editors. Gastroenterology and Hepatology. The Comprehensive Visual Reference, vol 7: Small Intestine. Philadelphia, Pa: Current Medicine; 1997. p 1.14.)

sodium channels and pumps.44-46 The pathophysiology of diarrhea in IBD is even more complex if we consider the role of luminal bacteria. Bacterial proteins, such as flagellin, may stimulate the production of cytokines, such as interleukin (IL)-8, which further attract inflammatory cells.47 Cytokines and immune cells also may influence tight junction barrier function and enterocyte secretory and absorptive pathways directly.48,49 Conversely, epithelial cells may secrete cytokines, such as IL-6, that enhance polymorphonuclear leukocyte (neutrophil) function.50 IBS is another example of a disorder with a complex pathophysiology. A constellation of factors, such as altered motility,32 bile acid malabsorption,51 and compromised rectal reservoir capacity,52 may aggravate symptoms in patients with IBS. At a more fundamental level, alterations in mast cell or enterochromaffin cell number, serotonin content, and serotonin reuptake and transport may contrib­ ute to the development of diarrhea (see Chapter 118).53-55 Complex pathophysiology also may be observed in malabsorption syndromes and functional disorders, particularly those characterized by rapid transit. Failure to absorb carbohydrates may lead to osmotic diarrhea, but failure to

absorb long-chain fatty acids may complicate matters by impairing electrolyte absorption by the colon.12,56 Common postprandial functional diarrhea probably involves an interplay between motility and transport functions. Diarrhea caused by food allergy also involves activation of immunologic, paracrine, and neural mechanisms that regulate vascular permeability, electrolyte transport, and motility (see Chapters 9 and 101).57

CLINICAL CLASSIFICATION Diarrhea can be classified in one of several ways: time course (acute vs. chronic), volume (large vs. small), pathophysiology (secretory vs. osmotic), stool characteristics (watery vs. fatty vs. inflammatory), and epidemiology. For the clinician, the classification used is useful only if it delineates a diagnostic and management approach in a given patient. In this regard, no single scheme is perfect; the experienced physician uses all of these classifications to facilitate patient care.

215

216

Section III  Symptoms, Signs, and Biopsychosocial Issues ACUTE VERSUS CHRONIC DIARRHEA

The time course of diarrhea can help direct management. Acute diarrhea (
LARGE-VOLUME VERSUS SMALL-VOLUME DIARRHEA

Differentiation of the cause of diarrhea on the basis of the volume of individual stools (rather than the total daily stool output) rests on the premise that the normal rectosigmoid colon functions as a storage reservoir. When that reservoir capacity is compromised by inflammatory or motility disorders involving the left colon, frequent small-volume bowel movements ensue. If the source of diarrhea is in the right colon or small bowel and if the rectosigmoid reservoir is intact, individual bowel movements are less frequent and larger. Therefore, frequent, small, painful stools may point to a distal colonic site of pathology, whereas painless largevolume stools suggest a right colonic or small bowel source. Although patients have difficulty quantifying stool volume accurately, the distinction between small- and large-volume stools may guide further diagnostic studies. The daily total stool output may also provide hints about the cause. IBS often results in normal or only slightly elevated 24-hour stool weights, whereas diarrhea of other causes may produce greater stool weights. The stool weight can be estimated by the patient’s history: patients with diarrhea that produces dehydration (in the absence of vomiting or limited oral intake) typically have stool weights more than 1000 g and therefore are unlikely to have IBS (see Chapter 118).

OSMOTIC VERSUS SECRETORY DIARRHEA

Distinguishing diarrhea that results from intestinal malabsorption of ingested nonelectrolytes (osmotic diarrhea) from diarrhea that results from malabsorption or secretion of electrolytes (secretory diarrhea) helps separate the small number of cases of osmotic diarrhea from the much larger number of cases of secretory diarrhea. This distinction is based on the measurement of stool electrolyte concentrations.10 In secretory diarrhea, sodium, potassium, and accompanying anions account almost entirely for stool osmolality, whereas in osmotic diarrhea poorly absorbable solutes within the lumen of the intestine account for much of the osmotic activity of stool water (see later discussion). Because osmotic diarrhea is caused by the ingestion of some poorly absorbed substance, it abates with fasting. Secretory diarrhea typically continues during fasting, although stool output may decrease modestly because of reduced endogenous secretions.

WATERY VERSUS FATTY VERSUS INFLAMMATORY DIARRHEA

When diarrhea is chronic (>four weeks), the differential diagnosis can overwhelm even the most experienced clinician. By characterizing stools as watery, fatty, or inflamma-

Table 15-2  Likely Causes of Diarrhea in Well-Defined Patient Groups or Settings Travelers Bacterial infection (mostly acute) Protozoal infections (e.g., amebiasis, giardiasis) Tropical sprue Epidemics and Outbreaks Bacterial infection Epidemic idiopathic secretory diarrhea (e.g., Brainerd diarrhea) Protozoal infection (e.g., cryptosporidiosis) Viral infection (e.g., rotavirus) Diabetic Patients Altered motility (increased or decreased) Associated diseases Celiac disease Pancreatic exocrine insufficiency Small intestinal bacterial overgrowth Drug side effects (especially acarbose, metformin) Patients with Acquired Immunodeficiency Syndrome Drug side effects Lymphoma Opportunistic infections (e.g., cryptosporidiosis, cytomegalovirus, herpes virus, Mycobacterium avium complex) Institutionalized and Hospitalized Patients Clostridium difficile toxin–mediated colitis Drug side effects Fecal impaction with overflow diarrhea Ischemic colitis Tube feeding

tory on the basis of simple stool tests, evaluation of the patient can be expedited by limiting the number of conditions that must be considered in the differential diagnosis.3 Watery diarrhea implies a defect primarily in water absorption as a result of increased electrolyte secretion or reduced electrolyte absorption (secretory diarrhea) or ingestion of a poorly absorbed substance (osmotic diarrhea). Fatty diarrhea implies defective absorption of fat and perhaps other nutrients in the small intestine. Inflammatory diarrhea implies the presence of one of a limited number of inflammatory or neoplastic diseases involving the gastrointestinal tract.

EPIDEMIOLOGIC FEATURES

One of the most useful clinical approaches to narrowing the differential diagnosis is to relate diarrhea to its setting. For example, a soccer mom and a backpacker from Nepal conceivably could have the same cause of the diarrhea but are more likely to have different causes. Some common clinical scenarios and the diagnoses that should be considered are shown in Table 15-2.

DIFFERENTIAL DIAGNOSIS Many gastrointestinal and systemic diseases may present with diarrhea. To facilitate the differential diagnosis, the clinician should divide diarrheal diseases into acute and chronic and further subdivide chronic diarrhea by stool characteristics—watery, inflammatory, and fatty (Table 15-3). Acute diarrhea is defined as lasting less than four weeks, although many cases last shorter than four days.58 The usual cause is infection by bacteria, viruses, protozoa, or multicellular parasites (Table 15-4). Acute diarrhea also can be caused by food poisoning, food allergies, and medications.

Chapter 15  Diarrhea Table 15-3  Differential Diagnosis of Diarrhea

Table 15-4  Infections That Cause Diarrhea

Acute Diarrhea Infection (see Table 15-4) Bacteria Parasites Protozoa Viruses Food allergies Food poisoning Medications Initial presentation of chronic diarrhea Chronic Diarrhea Fatty Diarrhea Malabsorption syndromes Mesenteric ischemia Mucosal diseases (e.g., celiac disease, Whipple’s disease) Short bowel syndrome Small intestinal bacterial overgrowth Maldigestion Inadequate luminal bile acid concentration Pancreatic exocrine insufficiency Inflammatory Diarrhea Diverticulitis Infectious diseases Invasive bacterial infections (e.g., tuberculosis, yersiniosis) Invasive parasitic infections (e.g., amebiasis, strongyloidiasis) Pseudomembranous colitis (Clostridium difficile infection) Ulcerating viral infections (e.g., cytomegalovirus, herpes simplex virus) Inflammatory bowel diseases Crohn’s disease Ulcerative colitis Ulcerative jejunoileitis Ischemic colitis Neoplasia Colon cancer Lymphoma Radiation colitis Watery Diarrhea Osmotic diarrhea Carbohydrate malabsorption Osmotic laxatives (e.g., Mg+2, PO4−3, SO4−2) Secretory diarrhea Bacterial toxins Congenital syndromes (e.g., congenital chloridorrhea) Disordered motility, regulation Diabetic autonomic neuropathy Irritable bowel syndrome Postsympathectomy diarrhea Postvagotomy diarrhea Diverticulitis Endocrinopathies Addison’s disease Carcinoid syndrome Gastrinoma Hyperthyroidism Mastocytosis Medullary carcinoma of the thyroid Pheochromocytoma Somatostatinoma VIPoma Idiopathic secretory diarrhea Epidemic secretory (Brainerd) diarrhea Sporadic idiopathic secretory diarrhea Ileal bile acid malabsorption Inflammatory bowel disease Crohn’s disease Microscopic colitis Collagenous colitis Lymphocytic colitis Ulcerative colitis Laxative abuse (stimulant laxatives) Medications and toxins (see Table 15-5) Neoplasia Colon carcinoma Lymphoma Villous adenoma in rectum Vasculitis

Bacteria Aeromonas spp. Campylobacter spp. Clostridium difficile Escherichia coli (enterotoxigenic, enteroinvasive, enterohemorrhagic) Pleisiomonas spp. Salmonella spp. Shigella spp. Viruses Adenovirus Norovirus Rotavirus Parasites or Protozoa Cryptosporidia Cyclospora Entamoeba histolytica Giardia lamblia Microsporidia

Table 15-5  Medications and Toxins Associated with Diarrhea Acid-reducing agents (e.g., histamine H2 receptor antagonists, proton pump inhibitors) Antacids (e.g., those that contain magnesium) Antiarrhythmics (e.g., quinidine) Antibiotics (most) Anti-inflammatory agents (e.g., 5-aminosalicylates, gold salts, NSAIDs) Antihypertensives (e.g., β-adrenergic receptor blocking drugs) Antineoplastic agents (many) Antiretroviral agents Colchicine Heavy metals Herbal products Prostaglandin analogs (e.g., misoprostol) Theophylline Vitamin and mineral supplements NSAIDs, nonsteroidal anti-inflammatory drugs.

Diseases that lead to chronic diarrhea may present with an acute onset and therefore must be considered when acute diarrhea becomes persistent (see Chapter 107). Chronic watery diarrhea may be caused by ingestion of poorly absorbed, osmotically active substances (osmotic diarrhea) or, more commonly, conditions that cause secretory diarrhea. Ingestion of any of a limited number of osmotic agents, such as magnesium, phosphate, and sulfate laxatives, or poorly absorbed carbohydrates, causes osmotic diarrhea. By contrast, chronic secretory diarrhea, in which electrolyte malabsorption leads to retention of fluid within the lumen, is associated with many clinical conditions (see Table 15-3). Although IBD typically produces diarrhea characterized by the presence of blood and pus, other diseases of inflammation without ulceration, such as microscopic colitis, cause diarrhea with the characteristics of chronic secretory diarrhea. Diarrhea in such cases is thought to be mediated by secretion of cytokines and other inflammatory mediators (see Chapter 124). Chronic watery diarrhea can also be caused by the ingestion of drugs or poisons (Table 15-5).59-61 Identification of drugs as the cause of diarrhea depends on recognizing that the initiation of drug ingestion and the onset of diarrhea

217

218

Section III  Symptoms, Signs, and Biopsychosocial Issues occurred coincidentally. Such a temporal correlation, however, is not always easy to identify and requires a detailed and carefully taken history. The pathophysiology of drug-induced diarrhea is complex and has not been carefully studied. Some drugs may activate specific receptors and transporters; for example, caffeine, like theophylline, may increase intracellular cAMP activity and fluid secretion. Clinically, this phenomenon can be seen in cases of what has been called “Starbucks diarrhea.” Erythromycin interacts with the motilin receptor, thereby stimulating propulsive motor activity in the gastrointestinal tract. Other antibiotics may alter the bacterial flora in the colon and lead to impaired colonic salvage of malabsorbed carbohydrate or overgrowth of toxin-producing Clostridium difficile. Some drugs such as cocaine may interfere with blood flow to the intestine. Chemotherapeutic agents are associated with a high frequency of diarrhea, which may result from disruption of the delicate balance between enterocyte proliferation and apoptosis, leading to what has been termed an apop­ totic enteropathy. A diverse group of drugs (e.g., aspirin, mycophenolate mofetil, gold) can incite an inflammatory process in the intestine that may cause diarrhea. The problem of detecting drug-induced diarrhea is more difficult in patients with surreptitious laxative abuse, because these patients deliberately conceal vital information about the cause of their problem (see later discussion of factitious diarrhea).62 Another category of chronic watery diarrhea involves disordered motility or dysregulation of intestinal function. Problems such as postvagotomy diarrhea, postsympathectomy diarrhea, diabetic autonomic neuropathy, amyloidosis, and probably diarrhea-predominant IBS belong in this category. In these situations, the diarrhea has the characteristics of a secretory diarrhea, because of primary dysregulation of electrolyte transport or of altered motility that speeds luminal fluid past absorptive sites in the intestine (see Chapters 35, 53, and 118). Another large category of watery diarrhea is diarrhea caused by endocrine dysfunction. Relatively common endocrine disturbances, such as hyperthyroidism and Addison’s disease, can be complicated by chronic secretory diarrhea. Much rarer endocrine tumors also produce diarrhea, typically by altering electrolyte absorption or speeding intestinal transit. The rarity of these tumors makes the pretest probability of finding these conditions low, especially in the absence of liver metastases, and therefore screening tests often are falsely positive (see later and Chapters 31, 32, and 35). Other tumors cause watery diarrhea by obstructing bowel, blocking lymphatic drainage, interfering with absorption, or causing electrolyte secretion. Examples of such conditions include colon carcinoma (bowel obstruction), lymphoma (lymphatic obstruction in the small bowel and mesentery), and villous adenomas of the rectum (secretion of a large amount of potassium-rich gelatinous fluid into the lumen). Villous adenomas found more proximally in the colon rarely cause this type of diarrhea (see Chapters 29, 122, and 123). The last category of chronic watery diarrhea is idiopathic secretory diarrhea. This rubric includes two entities, epidemic secretory diarrhea (also known as Brainerd diarrhea) and sporadic idiopathic secretory diarrhea. Both disorders are protracted but self-limited conditions (see later discussion of idiopathic secretory diarrhea).63,64 Chronic inflammatory diarrhea is the designation for a diverse group of infectious or idiopathic inflammatory and neoplastic processes. Stools are characterized by the presence of mucus and pus and are usually associated

with ulceration of the mucosa. Idiopathic IBD, including ulcerative colitis and Crohn’s disease, typically produces such stools. Less commonly, other inflammatory conditions such as diverticulitis or ulcerative jejunoileitis may be associated with blood or pus in the stool, as may infectious diseases that are invasive or ulcerating. Infections that cause chronic inflammatory diarrhea include bacterial infections, such as tuberculosis, yersiniosis, and Clostridium difficile– associated colitis, viral infections that ulcerate, such as cytomegalovirus and herpes simplex virus, and invasive parasitic infections, such as strongyloidiasis. In the immunocompromised person, a broader range of infectious agents should be considered. Noninfectious diseases that cause chronic inflammatory diarrhea include ischemic colitis, and neoplasms, such as colon cancer or lymphoma, that are complicated by ulceration of the mucosa (see Chapters 29, 107 to 112, 114, 115, 117, and 123). Chronic fatty diarrhea results from malabsorption or maldigestion. Malabsorption syndromes caused by mucosal diseases, such as celiac disease or Whipple’s disease, typically produce fatty diarrhea. Short bowel syndrome or postresection diarrhea can also present with this pattern, although if the resection is relatively limited, the diarrhea may be watery secondary to nutrient or bile-acid malabsorption. Small intestinal bacterial overgrowth causes steatorrhea by deconjugation of bile acids. Mesenteric ischemia affecting the small intestine may impair intestinal absorption of fat, but weight loss is more often attributed to sitophobia (“fear of eating”) because of postprandial pain. Maldigestion as a result of pancreatic exocrine insufficiency or inadequate duodenal bile acid concentration produces steatorrhea. Although fatty, the stools may not be very loose in maldigestive conditions because, in the absence of fat digestion, triglyceride remains intact and has little effect on colonic electrolyte absorption. By contrast, malabsorption in the presence of normal digestion may produce fairly voluminous diarrhea because of the cathartic action of free fatty acids in the colon (see Chapters 59, 101, 103 to 106, and 114).61

EVALUATION OF THE PATIENT HISTORY

A carefully taken medical history is the key to the evaluation of a patient presenting with diarrhea. An essential feature is the duration of symptoms. Patients with acute diarrhea (
Chapter 15  Diarrhea infectious diarrhea, visible blood in the stool is highly specific for infection with an invasive organism.65 Watery stools suggest an osmotic or secretory process, and the presence of oil or food particles is suggestive of malabsorption, maldigestion, or intestinal hurry. The phenomenon of floating stools generally represents an increase in the gas content rather than the fat content of the stools. The physician should also ask about the relationship of defecation to meals or fasting, passage of stool during the day versus the night, and presence of fecal urgency or incontinence. Urgency and incontinence are not indicative of voluminous diarrhea but suggest a problem with rectal compliance or with the muscles regulating continence. Nocturnal diarrhea that awakens the patient from sleep strongly suggests an organic rather than a functional disorder such as IBS. Other coexisting symptoms such as abdominal pain, flatulence, bloating or gaseous distention, cramps, fever, and weight loss should be noted. Excessive flatus suggests increased fermentation of carbohydrate by colonic bacteria as a result of ingestion of poorly absorbable carbohydrate or malabsorption of carbohydrate by the small intestine. Because iatrogenic causes of diarrhea, such as drugs, previous surgery, or radiation therapy, are common, the physician should explore the history thoroughly for prior abdominal surgeries and ingestion of prescription drugs and over-the-counter remedies, including nutritional and herbal therapies. The patient’s diet should be reviewed thoroughly because diarrhea may result from ingestion of large quantities of poorly absorbable carbohydrates, such as fructose, or sugar alcohols, such as sorbitol or mannitol, which may be consumed in fruit juices and soda (which contain fructose and high-fructose corn syrup) or as dietetic, sugar-free candies and chewing gums (which contain sorbitol and mannitol).66 Excessive coffee consumption also may be associated with diarrhea. Epidemiologic clues also should be pursued (see Table 15-2). For example, recent foreign travel, particularly to undeveloped countries, makes the diagnosis of travelers’ diarrhea likely. The globalization of commerce has increased the frequency of once exotic infections in those without grossly obvious exposures.67 The physician also should consider the patient’s residence in a rural or urban environment, the source of the patient’s drinking water, and the patient’s occupation, sexual orientation, and use of alcohol or illicit drugs. Potential secondary gains from illness or a history of attempted weight loss and fixation on body image should raise the possibility of laxative abuse (see later). The patient’s history is essential in differentiating patients with IBS from those with other functional disorders or organic conditions that cause diarrhea. Current definitions of IBS emphasize the presence of abdominal pain associated with defecation.68 Additional factors that suggest a diagnosis of IBS include a long history that usually extends back to adolescence or young adulthood, passage of mucus, and exacerbation of symptoms by stress. Factors that argue against a diagnosis of IBS include the recent onset of diarrhea, especially in older patients, diarrhea that awakens the patient from sleep, weight loss, the presence of blood in the stool, and stool weights more than 400 to 500 g daily. Abnormal blood test results, such as a low hemoglobin level, low serum albumin concentration, or high erythrocyte sedimentation rate, also argue against a diagnosis of IBS (see Chapter 118). Painless diarrhea should no longer be considered a form of IBS. The Rome III committee has defined functional diarrhea as “at least 3 months, which need not be consecutive, in the preceding 6 months of liquid (mushy) or watery stools more than three-quarters of the time; and no abdominal

pain.”68 Obviously, many patients with chronic diarrhea will not have a readily defined cause of diarrhea identified when first seen and could be characterized as having functional diarrhea. Physicians should not rush to make this diagnosis without exploring alternative possibilities, particularly those that can produce episodic and variable diarrhea, such as small intestinal bacterial overgrowth and carbohydrate malabsorption. Functional diarrhea must also be distinguished from idiopathic secretory diarrhea (see later).

PHYSICAL EXAMINATION

Physical findings are usually more useful in determining the severity of diarrhea than in determining its cause. The patient’s volume status can be assessed by looking for orthostatic changes in blood pressure and pulse. Fever and other signs of toxicity should be noted. A careful abdominal examination is important, with particular attention to the presence or absence of bowel sounds, abdominal distention, localized or generalized tenderness, masses, and an enlarged liver. On occasion, the physical examination may provide more direct evidence of the cause of diarrhea. Characteristic physical findings may be seen in mastocytosis (urticaria pigmentosa), amyloidosis (macroglossia, waxy papules, pinch purpura), Addison’s disease (increased pigmentation), glucagonoma (migratory necrotizing erythema), carcinoid syndrome (flushing), Köhlmeier-Degos disease (malignant atrophic papulosis), and celiac disease (dermatitis herpetiformis). Peripheral neuropathy and orthostatic hypotension may be the only clues to a diagnosis of amyloidosis. A thyroid nodule with cervical lymphadenopathy may be the only lead to the presence of medullary carcinoma of the thyroid. Tremor and other systemic signs should lead to consideration of hyperthyroidism. Right-sided heart murmurs, as well as an enlarged hard liver, may be present with carcinoid syndrome. Evidence of arthritis may be noted in IBD, Whipple’s disease, and some enteric infections. Lymphadenopathy might suggest acquired immunodeficiency syndrome (AIDS) or lymphoma. Signs of peripheral vascular disease with or without an abdominal bruit may suggest chronic mesenteric ischemia. Evidence of chronic liver disease may suggest advanced primary sclerosing cholangitis in a patient with colitis. A careful rectal examination may disclose defective sphincter or pelvic floor muscle function that could produce fecal incontinence.

FURTHER EVALUATION OF ACUTE DIARRHEA

Most cases of acute diarrhea are caused by infectious diseases that have a limited course, from a few days to a few weeks, and do not require a physician’s intervention unless the patient’s immune system is compromised or the patient develops complications of volume depletion or other evidence of severe toxicity, including inability to ingest fluid, frequent vomiting, and debilitating muscle or joint pain.58 When these complications are present or when the diarrhea has persisted for more than a few days, a more comprehensive evaluation is warranted. In such patients, a complete blood count should be done to look for anemia, hemoconcentration, or an abnormal white blood cell count. Patients with a viral cause of diarrhea usually have normal white blood cell and differential counts or lymphocytosis, but those with bacterial infections, particularly caused by organisms that invade the intestinal mucosa, have a leukocytosis with an excess of immature white blood cells. Neutropenia, however, can occur in patients with salmonellosis. Measurements of serum electrolyte concentrations

219

220

Section III  Symptoms, Signs, and Biopsychosocial Issues and blood urea nitrogen and serum creatinine levels can be used to assess the extent of fluid and electrolyte depletion and its effect on kidney function. Stool samples should be examined for white blood cells to identify inflammatory diarrhea.69 The standard method for detecting white blood cells in stool is with a Wright stain and microscopy. The accuracy of the test depends on the experience and skill of the observer, because false-positive and false-negative results are common. Tests for the neutrophil products, calprotectin and lactoferrin, are sensitive and specific for the detection of neutrophils in stool and may be a useful alternative to microscopy.70 Studies have suggested that stool cultures are unlikely to grow pathogenic bacteria in the absence of fecal leukocytes; therefore, a Wright stain of stool or a fecal lactoferrin assay can be used to decide which stool samples should be sent for bacterial culture, thereby minimizing expense. This approach may be of more value in outpatients than in patients hospitalized with diarrhea, because the latter have toxicity or have failed to resolve spontaneously within a few days and must have stool cultures sent. Routine stool cultures are of little use for hospitalized patients in whom acute diarrhea develops while the patient is in the hospital; testing for Clostridium difficile toxin is likely to be more helpful. The diagnostic value of examination of stool for ova and parasites depends on the pretest probability of a parasitic infection and the experience of the observer. Enzyme-linked immunosorbent assays (ELISAs) for giardiasis and cryptosporidiosis and serologic testing for amebiasis are more accurate tests than stool microscopy and should be ordered, even in the absence of fecal leukocytes.71 Patients who have been treated with antibiotics in the preceding three months or those in whom diarrhea develops in an institutional setting should be tested for C. difficile toxin.72 With the increase in community acquired C. difficile infections, physicians should consider this treatable cause of acute diarrhea, even in the absence of a prior history of antibiotic use (see Chapters 107 to 109). Abdominal radiographs should be obtained in toxic patients to assess for colitis and to look for evidence of ileus or megacolon. Proctoscopy or flexible sigmoidoscopy also should be considered in patients who are clearly toxic with infection, have blood in the stool, or have persistent acute diarrhea. Sigmoidoscopy is probably adequate as an early investigation in such cases of severe acute diarrhea. In patients with AIDS-related diarrhea, colonoscopy is preferable because a substantial proportion of infections and lymphomas may be present only in the right colon,73 although this approach has been called into question.74 If sigmoidoscopy or colonoscopy is done, mucosal biopsy specimens should be obtained, even if the mucosa does not appear to be grossly inflamed, because pathologic examination can identify important clues to facilitate a specific diagnosis.75 An algorithm for the evaluation of patients with acute diarrhea is shown in Figure 15-4.

FURTHER EVALUATION OF CHRONIC DIARRHEA

Because the differential diagnosis of chronic diarrhea is more extensive than that of acute diarrhea, evaluation of patients with chronic diarrhea is more complex.76,77 Initially, the physician should categorize the diarrhea as watery, inflammatory, or fatty (Fig. 15-5). In addition to the history, physical examination, and routine laboratory tests already mentioned, analysis of a stool sample can be used to categorize the diarrhea and thereby limit the number of conditions to be considered in the differential diagnosis.

History Duration Epidemiology Travel Food Water

Stool characteristics Watery Bloody

Abdominal pain Acute colitis Inflammatory bowel disease

Concurrent diseases Drugs

Physical examination General Fluid balance Fever Nutrition Rash Initial assessment Toxic Nontoxic Prolonged course Short duration Blood in stools No bleeding Tender Not tender Dehydrated

Abdomen Tenderness Distention Organomegaly

Rectal examination Fecal occult blood test

Symptomatic therapy Oral rehydration Antidiarrheal drugs No response

Response

Fluid/electrolyte repletion

Laboratory evaluation Complete blood count Hemoconcentration WBC, differential counts Sigmoidoscopy or colonoscopy with mucosal biopsy

Empirical antibiotic therapy Specific therapy

Serum chemistries Electrolytes Urea nitrogen Creatinine Ameba serology*

Stool tests Ova and parasite exam* Giardia antigen* Clostridium difficile toxin* Fecal WBCs Positive

Negative

Stool culture *In appropriate epidemiologic circumstances

Figure 15-4.  Algorithm for the evaluation of patients with acute diarrhea. WBC, white blood cells. (From Schiller LR. Diarrhea. Med Clin North Am 2000; 84:1259-74.)

Stool analysis can be obtained on a random sample or a timed collection (i.e., 24-, 48-, or 72-hour stool sample). The value of analyzing a timed collection is that the stool weight and hence the output of stool components, such as fat, can be measured accurately. The daily stool weight is perhaps the best clue to the potential metabolic impact of the diarrhea. In the absence of a timed collection, however, assessments of other stool characteristics on a random, or spot, collection still provide many clues to the correct diagnosis.3 These assessments include stool sodium and potassium concentrations, stool pH, a fecal occult blood test, and an examination of stool for white blood cells or a test for the presence of a surrogate marker, such as fecal lactoferrin or calprotectin.70 In appropriate circumstances, stool samples can also be analyzed for fat content and for laxatives, including magnesium, phosphate, sulfate, bisacodyl, and anthraquinones (see later discussion of factitious diarrhea). Although stool collections are often viewed by patients and physicians as being messy and distasteful, they usually can be done easily and successfully at home or in the hospital. Perhaps the biggest hurdle is in dealing with laboratories that are inexperienced or uninterested in stool analysis. Commercially available collection units that fit into a commode and allow separation of stool and urine facilitate the collection, as does the use of preweighed plastic or metal containers and a small refrigerator or picnic cooler to keep the specimens cold. Patients should continue

Chapter 15  Diarrhea History Onset Congenital Abrupt Gradual Pattern Continuous Intermittent

Stool characteristics Watery Bloody Fatty Fecal incontinence

Duration

Abdominal pain Inflammatory bowel disease Irritable bowel syndrome Ischemia Weight loss Malabsorption Neoplasm

Epidemiology Travel Food Water

Aggravating factors Diet Stress Mitigating factors Diet OTC drugs Rx drugs

Iatrogenic diarrhea Drugs Radiation Surgery Factitious diarrhea Laxatives

Systemic diseases Hyperthyroidism Diabetes mellitus Collagen-vascular diseases Tumor syndromes AIDS lg deficiencies

Previous evaluation

Physical examination General Fluid balance Nutrition Fever

Skin Flushing Rashes Dermatographism

Thyroid Mass

Chest Wheezing

Heart Murmur

Abdomen Hepatomegaly Mass Ascites Tenderness

Routine laboratory tests Complete blood count Anemia Leukocytosis

Anorectal Sphincter competence Fecal occult blood test Fissure Fistula

Extremities Edema

Chemistry screen Fluid/electrolyte status Nutritional status Serum protein/globulins

Stool analysis Weight

Electrolytes Osmotic gap

pH (Low in carbohydrate malabsorption)

Fecal occult blood test

Stool WBCs

Fat output Sudan stain Quantitative fat determination

Laxative screen

Categorize Watery diarrhea

Secretory

Inflammatory diarrhea

Fatty diarrhea

Osmotic

Figure 15-5.  Algorithm for the initial evaluation of patients with chronic diarrhea. AIDS, acquired immunodeficiency syndrome; Ig, immunoglobulin; OTC, over-the-counter; Rx, prescription; WBCs, white blood cells. (From Fine KD, Schiller LR. AGA technical review on the evaluation and management of chronic diarrhea. Gastroenterology 1999; 116:1464-86.)

their regular activities and should consume a regular diet, including a fat intake of 80 to 100 g of fat daily, during the collection. Keeping a diary of food and liquid ingested facilitates estimation of the patient’s fat and calorie intake by a dietitian. When evaluating the results of timed stool collections, it is important to recognize that patients often do not eat diets this high in fat. During the collection, diagnostic tests that might alter stool output or composition, such as barium radiograph studies, should be avoided and only essential medication should be given. Any antidiarrheal medications should be withdrawn. For most patients with diarrhea, a 48-hour collection is sufficient. If stool

output is not representative during that time, the collection can be extended. Occasionally, stool output is measured during fasting; if the diarrhea is caused by some ingested substance, fasting should abolish the diarrhea. Persistence of diarrhea during fasting is one criterion for secretory diarrhea. Measurement of stool sodium and potassium concentrations allows the physician to calculate an osmotic gap in stool water. The osmotic gap is calculated by subtracting twice the sum of the sodium and potassium concentrations from 290 mOsm/kg, the osmolality of stool in the body.10 The concentration is doubled to account for anions that

221

Section III  Symptoms, Signs, and Biopsychosocial Issues 300

mmol/L

222

250

Unmeasured osmoles

200

Bicarbonate/ organic anions

150

Chloride

100

Potassium Sodium

50

0 Secretory Osmotic diarrhea diarrhea Figure 15-6.  Fecal electrolytes and the fecal osmotic gap. The osmolality of colonic fluid is in equilibrium with body fluids and is approximately 290 mOsm/kg. The total concentration of electrolytes, therefore, cannot exceed 290 mmol/L. In secretory diarrhea, almost all the osmotic activity of colonic contents is caused by electrolytes and, therefore, the estimate of electrolyte content by the formula 2 × ([Na+] + [K+]), is approximately 290 mmol/L. In osmotic diarrhea, electrolytes account for only a small part of the osmotic activity; unmeasured osmoles resulting from the ingestion of a poorly absorbed substance account for most of the osmotic activity, and the calculated osmotic gap will be high. (From Schiller LR. Chronic diarrhea. In: McNally P, editor. GI/Liver Secrets. 2nd ed. Philadelphia, Pa: Hanley & Belfus; 2001. p 411.)

accompany these cations. A small osmotic gap (<50 mOsm/ kg), which signifies that the osmolality of stool water is attributable mostly to incompletely absorbed electrolytes, is characteristic of secretory diarrhea (Fig. 15-6). On the other hand, a large osmotic gap (>100 mOsm/kg) indicates that much of the stool osmolality is composed of nonelectrolytes. A large gap is characteristic of an osmotic diarrhea, usually resulting from ingestion of some poorly absorbed substance, such as magnesium salts. When the sum of sodium and potassium concentrations doubled is higher than 290 mOsm/kg, ingestion of a poorly absorbed multivalent anion, such as phosphate or sulfate, is likely.10 Such a negative osmotic gap is the result of an excess of cations obligated by multivalent anions. The actual measurement of stool osmolality is of value only in detecting samples that have been contaminated by the addition of water or hypotonic urine. Such samples have an osmolality lower than 290 mOsm/kg. Stool osmolality tends to rise once the stool has been collected because of continuing bacterial fermentation in vitro.12 Therefore, measured osmolality should not be used to calculate the fecal osmotic gap. The pH of stool water provides useful information about the possibility of carbohydrate malabsorption.10 Carbohydrate that reaches the colon is promptly fermented by the bacterial flora, with release of CO2 and H2 gases and shortchain fatty acids. As a result of fermentation, the pH is acidic, usually dropping to less than 6, a finding that indirectly indicates excess carbohydrate fermentation in the colon. Fecal occult blood testing and examination of stool for leukocytes allow one to identify inflammatory diarrhea resulting from colitis or malignancy. Other diarrheal condi-

tions that cause occult bleeding include lymphoma of the small intestine, celiac disease (fecal occult blood in 50% of cases), and refractory sprue (fecal occult blood in 70% of cases).78 Stool fat output can be measured quantitatively by chemical means on a timed (48- to 72-hour) collection or estimated qualitatively by use of a Sudan stain on a random specimen. Steatorrhea is defined as excessive loss of fat in the stool (more than 7 g, or 9% of intake, for 24 hours). This definition, however, may not be valid for the diagnosis of fat malabsorption or maldigestion in all patients with chronic diarrhea. In one study,79 diarrhea induced with laxatives produced mild steatorrhea in 35% of normal subjects. In patients with diarrhea, fat excretion in the range of 7 to 14 g/24 hr has a low specificity for the diagnosis of defective fat absorption. Fat excretion more than 14 g/24 hr, however, strongly indicates a problem with fat absorption.79 Fat intake during a quantitative collection should be estimated from diet diaries, because patients with diarrhea frequently have anorexia or early satiation that may reduce their fat intakes substantially, thereby reducing fat excretion. For a valid study, patients should consume 70 to 100 g of fat daily for a few days before and during the timed collection. Measurement of fat excretion as a measure of malabsorption also can be compromised by ingestion of the lipase inhibitor orlistat or the fat substitute olestra.80 When only a random sample of stool is available, qualitative estimation of fat excretion by means of a Sudan stain of a fecal smear may be helpful.81 Semiquantitative methods can be applied to measure the number and size of fat globules, and these methods produce results that correlate well with quantitative collections (see Chapter 101). In patients in whom surreptitious laxative ingestion is suspected, stool water can be analyzed for laxatives by chemical or chromatographic methods. As currently done commercially, this analysis is subject to error.82 If positive, the test for the laxative should be repeated on another stool sample to confirm the finding before confronting the patient with this discovery. Stool samples can also be assayed with a chemical test for carbohydrate (anthrone reagent)12 and for α1-antitrypsin clearance to detect protein-losing enteropathy.83 These tests have limited clinical usefulness and should not be used routinely for the initial evaluation of a patient with chronic diarrhea. Once the stool analysis is completed, chronic diarrhea can be categorized as watery (and secretory or osmotic), inflammatory, or fatty. This classification allows more direct evaluation of the cause of diarrhea.

Chronic Watery Diarrhea

Secretory diarrhea has a broad differential diagnosis, as indicated previously (see Table 15-3), and a wide investigative net must be cast to identify a specific cause (Fig. 15-7). Infection should be excluded by stool culture for bacteria and special tests for other organisms. The patient’s human immunodeficiency virus (HIV) status should be clarified at this point, because patients with AIDS are more likely than others to have an infectious cause of chronic diarrhea (see Chapter 33).84 Although most bacteria that cause diarrhea are cleared spontaneously within four weeks, some organisms, such as Aeromonas and Plesiomonas, may produce chronic diarrhea.85,86 Special culture techniques may be needed to detect these organisms. Special techniques are also required to find other pathogens. For example, coccidia and microsporidia require special microbiologic techniques such as polymerase chain reaction methodology for optimal detection.87 Giardiasis and cryptosporidiosis are sometimes

Chapter 15  Diarrhea Secretory diarrhea

Exclude infection Bacterial pathogens “Standard” Aeromonas Plesiomonas

Other pathogens “Standard” ova + parasites Microsporidia Giardia

Exclude structural disease Small bowel series

Sigmoidoscopy or colonoscopy with biopsy

CT scan of abdomen and pelvis CT enterography

Small bowel biopsy and aspirate for quantitative culture

Selective testing Plasma peptides Gastrin Calcitonin VIP Somatostatin

Urine 5-HIAA Metanephrines Histamine

Other tests TSH ACTH stimulation Serum protein electrophoresis Immunoglobulins

Trial of bile acid-binding resin for bile acid diarrhea

difficult to diagnose by a standard examination for ova and parasites, and ELISA testing for Giardia antigen and cryptosporidium antigen should be ordered.71 Examination of mucosal biopsy specimens with special stains or electron microscopy may be needed to find pathogens. Small intestinal bacterial overgrowth may result in secretory diarrhea, presumably caused by toxins, as well as fatty diarrhea caused by bile salt deconjugation (see later). The glucose-hydrogen breath test (see later) can be used to screen for this condition, but the gold standard for diagnosis of small intestinal bacterial overgrowth remains quantitative culture of a small bowel aspirate, if available (see Chapter 102).88 Structural diseases, such as short bowel syndrome, gastrocolic or enteroenteric fistula, mucosal diseases, IBD, and tumors, including lymphomas, should be sought by means of radiographic and endoscopic techniques. Small bowel radiographs remain an important method for detecting structural small bowel diseases. Computed tomography (CT) or magnetic resonance imaging (MRI), particularly CT or MRI enterography, is of value for detecting not only small bowel and colonic diseases, but also diseases extrinsic to the bowel that can cause diarrhea, such as pancreatic tumors. Visualization and biopsy of the mucosa of the small bowel by endoscopy or enteroscopy can be valuable, although whether push enteroscopy adds much to standard esophagogastroduodenoscopy for the evaluation of diffuse small bowel diseases is uncertain.89 Diseases that may be detected by small intestinal biopsy include celiac disease, Crohn’s disease, giardiasis, intestinal lymphoma, eosinophilic gastroenteritis, tropical sprue, Whipple’s disease, lymphan­

Figure 15-7.  Algorithm for the evaluation of chronic secretory diarrhea. ACTH, adrenocorticotropic hormone; CT, computed tomography; 5-HIAA, 5-hydroxyindoleacetic acid; TSH, thyroid-stimulating hormone; VIP, vasoactive intestinal peptide. (From Fine KD, Schiller LR. AGA technical review on the evaluation and management of chronic diarrhea. Gastroenterology 1999; 116:1464-86.)

giectasia, abetalipoproteinemia, amyloidosis, mastocytosis, and various infectious processes. Patients with many of these disorders usually, but not always, present with steatorrhea (see Chapters 27, 29, 35, 101, 104 to 109, and 111). The role of wireless capsule endoscopy in the diagnosis of diarrhea caused by small bowel disease is evolving rapidly.90 Subtle lesions not appreciated by other diagnostic modalities can be seen, and new methods of deep enteroscopy (double-balloon and spiral overtube enteroscopy) allow access to and biopsy of most of these abnormalities (see Chapter 19). Sigmoidoscopy or colonoscopy can be used to visualize the colon and permit directed biopsies. Because colonic causes of chronic secretory diarrhea tend to produce diffuse changes throughout the colon, sigmoidoscopy usually is adequate for this purpose.91 Colonoscopy is preferable if the patient is older (and thus requires screening for colon cancer), has blood in the stool, is suspected of having right colonic or ileal disease, or has AIDS.73,92 Chronic disorders that can be diagnosed by inspection of the colonic mucosa include pseudomelanosis coli, polyps, tumors, Crohn’s disease, ulcerative colitis, amebiasis, and nonspecific ulceration. All patients with undiagnosed chronic secretory diarrhea should have mucosal biopsy specimens obtained from the colon, even when the mucosa appears grossly normal.3 Random biopsies should include multiple samples from several locations to give the pathologist the best chance of making a diagnosis. Diseases in which the colonic mucosa appears normal endoscopically, but which can be diagnosed histologically, include microscopic colitis (lymphocytic and collagenous colitis, see later),

223

224

Section III  Symptoms, Signs, and Biopsychosocial Issues amyloidosis, granulomatous infections, and schistosomiasis (see Chapters 35, 110 to 112, and 124). The diagnostic yield of colonoscopy or sigmoidoscopy with biopsy in patients referred for chronic diarrhea is approximately 15% to 30%. The next level of investigation is selective testing for diarrhea caused by peptide-secreting tumors, an intellectually interesting form of chronic watery diarrhea that is quite rare. The pretest probability of having a peptide-secreting tumor in a patient with chronic diarrhea is so low that screening these patients with a panel of serum peptide levels is far more likely to produce a false-positive than a truepositive result.93 Testing for elevated serum peptide levels or urinary metabolites of endocrine mediators, such as 5-hydroxyindoleacetic acid (5-HIAA), metanephrine, and histamine, should be limited to those patients who have chronic diarrhea with symptoms and signs consistent with a tumor syndrome, such as flushing or a large hard liver in carcinoid syndrome, ulcer disease suggestive of ZollingerEllison syndrome, headache, flushing, and urticaria pigmentosa in mastocytosis, or patients who have a CT scan that shows a tumor.19 Scintigraphy using radiolabeled octreotide, especially combined with positron-emission tomography (PET) or CT, also can be used to identify peptide-secreting tumors (see Chapters 31 and 32). More common endocrinologic diseases that cause diarrhea are diabetes mellitus, hyperthyroidism, and Addison’s disease. In many cases, other symptoms and signs, such as an enlarged thyroid or skin pigmentation characteristic of Addison’s disease, suggest the presence of these conditions. Blood glucose, thyroid-stimulating hormone, and serum cortisol levels before and after injection of an adrenocorticotropic hormone analog should be measured selectively in patients who might have one of these conditions. Other blood tests that may be relevant in evaluating secretory diarrhea include serum protein electrophoresis and immunoglobulin electrophoresis. Selective immunoglobulin A (IgA) deficiency may present with recurrent intestinal infections such as giardiasis, whereas combined variable immune deficiency can be associated with a variety of puzzling intestinal findings that sometimes mimic celiac disease.94 Testing for HIV and HIV-2 may be appropriate (see Chapters 33 and 35). Osmotic diarrhea has a much more limited differential diagnosis, and its evaluation is much simpler (Fig. 15-8).9 For practical purposes, osmotic diarrhea is caused by one of three conditions—ingestion of exogenous magnesium, consumption of poorly absorbable carbohydrates, or carbohydrate malabsorption. Ingestion of other osmotically active substances is unusual. Fortunately, these conditions can be differentiated by taking a careful history and performing simple stool tests. Magnesium can be measured directly in stool water by atomic absorption spectrophotometry.11 Excretion of more than 15 mmol (30 mEq) of magnesium daily or concentrations in stool water of more than 44 mmol/L (90 mEq/L) strongly suggests magnesium-induced diarrhea. The ingestion may be intentional, as in a patient with surreptitious laxative ingestion, or accidental, as in a patient who uses magnesium-containing antacids or mineral supplements. Ingestion of poorly absorbed carbohydrates or carbohydrate malabsorption typically leads to a low fecal pH because of bacterial fermentation in the colon. A fecal pH lower than 6 is highly suggestive of carbohydrate malabsorption.10,12 More generalized malabsorption that involves fecal loss of amino acids and fatty acids in addition to carbohydrate may produce a somewhat higher pH (e.g., pH = 6 to 7.5). Isolated carbohydrate malabsorption is usually

Osmotic diarrhea

Stool analysis Low pH (Carbohydrate malabsorption)

High magnesium output (Inadvertent ingestion, laxative abuse)

Dietary review Breath H2 test (lactose) Lactase assay Figure 15-8.  Algorithm for the evaluation of chronic osmotic diarrhea. (From Fine KD, Schiller LR. AGA technical review on the evaluation and management of chronic diarrhea. Gastroenterology 1999; 116:1464-86.)

caused by ingestion of a poorly absorbable carbohydrate, such as lactose in a person with lactase deficiency. Other common causes include ingestion of poorly absorbed sugar alcohols that are used as artificial sweeteners, such as sorbitol or mannitol, or excessive ingestion of sugars with a limited absorption capacity, such as fructose.66 Therapeutic use of inhibitors of carbohydrate absorption, such as acarbose, may also lead to carbohydrate malabsorption.95 Because fermentation produces not only short-chain fatty acids that acidify the stool, but also carbon dioxide and hydrogen, complaints of gas and bloating by the patient are clinical clues to the presence of carbohydrate malabsorption, although these symptoms are fairly nonspecific (see Chapters 16, 100, and 101).96 Once the clinical picture or stool analysis suggests carbohydrate malabsorption, a careful review of the patient’s diet may indicate the likely source. In some persons, a hydrogen breath test with lactose as the sugar substrate can confirm lactose intolerance as the diagnosis.97 In this test, a previously fasting patient ingests 25 g of lactose dissolved in water, and exhaled breath is assayed for hydrogen content at baseline and at intervals for several hours. Because hydrogen is not a normal product of human metabolism, any increase in breath hydrogen concentration is the result of bacterial fermentation and indicates that unabsorbed lactose has reached the colon. This principle also has been applied to the assessment of sucrase deficiency after administration of a sucrose load and to the assessment of fructose malabsorption after a fructose load.98 Breath hydrogen testing has been adapted to detect small intestinal bacterial overgrowth with the use of glucose, a substrate that ordinarily should be absorbed completely before reaching the colon.88 Lactulose, a nonabsorbable but easily fermented disaccharide, also has been used to detect small intestinal bacterial overgrowth, but because of the wide variability of intestinal transit time, use of lactulose for this purpose is problematic. Lactulose can be used as a substrate for determining the oral-cecal transit time. Breath hydrogen testing after administration of d-xylose has been advocated as a screening test for generalized intestinal malabsorption.99 For most purposes breath hydrogen testing provides only supportive evidence when the pretest likelihood of a particular diagnosis is high (see Chapter 101). Once a specific cause of osmotic diarrhea has been postulated, a therapeutic trial of an elimination diet can confirm the diagnosis.

Chapter 15  Diarrhea Inflammatory diarrhea

Fatty diarrhea

Exclude structural disease

Exclude structural disease Small bowel series

Sigmoidoscopy or colonoscopy with biopsy

CT scan of abdomen and pelvis CT enterography

Small bowel biopsy

Exclude infection Bacterial pathogens “Standard” Aeromonas Plesiomonas Tuberculosis

Small bowel series Exclude pancreatic exocrine insufficiency

Other pathogens Parasites Viruses

Figure 15-9.  Algorithm for the evaluation of chronic inflammatory diarrhea. CT, computed tomography. (From Fine KD, Schiller LR. AGA technical review on the evaluation and management of chronic diarrhea. Gastroenterology 1999; 116:1464-86.)

Chronic Inflammatory Diarrhea

Patients with chronic diarrhea and white blood cells or blood in the stool are classified as having inflammatory diarrhea. These characteristics indicate that the mucosa is disrupted and inflamed. Diagnostic considerations include IBD, infections, pseudomembranous enterocolitis, mesenteric ischemia, radiation enteritis, and neoplasia. Because these conditions may produce a secretory diarrhea without markers of inflammation in the stool, they must be considered in the differential diagnosis of secretory diarrhea as well (see Chapters 39, 104 to 112, 114, 115, and 121 to 124). Sigmoidoscopy or colonoscopy should be undertaken to look initially for structural changes, because colitis is a common cause of inflammatory diarrhea and a neoplasm can be life threatening (Fig. 15-9). Sigmoidoscopy can detect most causes of inflammatory diarrhea but can miss disorders localized to the right colon and ileum. Because preparation for this test is simpler than that for colonoscopy and the frequency of complications is lower, sigmoidoscopy is preferred by some physicians. Others prefer to examine the entire colon and terminal ileum in patients with inflammatory diarrhea, especially if occult blood is detected in the stool.92 The choice of test depends on the circumstances of the individual patient. For example, a patient older than 50 years who has not undergone previous screening for colon cancer would benefit by undergoing colonoscopy in this setting. Whichever test is selected, biopsy specimens must be obtained from the colon to aid in making the correct diagnosis. Infection can cause chronic inflammatory diarrhea or aggravate existing inflammatory diarrhea caused by ulcerative colitis or Crohn’s disease. The pathogens most likely to cause chronic inflammatory diarrhea are C. difficile, cytomegalovirus, Entamoeba histolytica, Yersinia spp., and Mycobacterium tuberculosis. C. difficile and cytomegalovirus are notorious for causing exacerbations of IBD (see Chapters 111 and 112).100 In addition to biopsies, appro­ priate cultures and serologic tests should be obtained to exclude these infections.

CT scan of abdomen and pelvis CT enterography

Secretin test

Small bowel biopsy and aspirate for quantitative culture

Stool chymotrypsin or fecal elastase activity

Figure 15-10.  Algorithm for the evaluation of chronic fatty diarrhea. (From Fine KD, Schiller LR. AGA technical review on the evaluation and management of chronic diarrhea. Gastroenterology 1999; 116:1464-86.)

Chronic Fatty Diarrhea

Steatorrhea implies the disruption of fat solubilization, digestion, or absorption in the small intestine. The evaluation of chronic fatty diarrhea is designed to distinguish maldigestion, inadequate luminal breakdown of triglyceride, from malabsorption, inadequate mucosal transport of the products of digestion (see Chapter 101). The major causes of maldigestion are pancreatic exocrine insufficiency (e.g., chronic pancreatitis) and lack of bile (e.g., advanced primary biliary cirrhosis). Mucosal diseases (e.g., celiac disease) are the usual causes of malabsorption. The absolute amount of steatorrhea and the fecal fat concentration (grams of fat/100 g of stool) provide clues to the cause of steatorrhea.101 The degree of steatorrhea tends to be higher with maldigestion (as in pancreatic insufficiency)— often more than 30 g fat/day—than with mucosal disease, because of the greater disruption of fat assimilation with maldigestion. Additionally, the fecal fat concentration tends to be higher with maldigestion than with mucosal disease because mucosal disease often is associated with poor fluid and electrolyte absorption, so that the stool fat content is diluted by unabsorbed water. Also, because fat digestion usually is intact in mucosal disease, triglycerides are broken down to fatty acids in the small intestine and pass into the colon, where they inhibit electrolyte and water absorption and thus further dilute the fat content of stool.56 By contrast, maldigestion caused by pancreatic and biliary disorders typically reduces triglyceride hydrolysis and does not result in delivery of excess fatty acids to the colon or inhibition of fluid and electrolyte absorption. Therefore, the unabsorbed fat is dispersed in a smaller stool volume and is thus more concentrated. A fecal fat concentration more than 9.5 g/100 g in a patient with suspected maldigestion strongly suggests a pancreatic or biliary cause of steatorrhea. The further evaluation of patients with chronic fatty diarrhea is relatively straightforward (Fig. 15-10). The first step is to look for a structural problem involving the small bowel or pancreas. The evaluation may include small bowel radiography or endoscopy with small bowel biopsy, CT, and MR cholangiopancreatography. When a small bowel biopsy is performed, luminal contents should be aspirated and a sample sent for quantitative culture to exclude small intestinal bacterial overgrowth. Because celiac disease is the

225

226

Section III  Symptoms, Signs, and Biopsychosocial Issues most common cause of mucosal disease that leads to malabsorption, tissue transglutaminase antibodies and endomysial antibodies should be determined (see Chapters 102 and 104).102 If no intestinal abnormalities are discovered or radiographic evidence of chronic pancreatitis is detected, abnormal pancreatic exocrine function should be considered. Available tests of pancreatic function all have limitations. The secretin stimulation test, in which exogenous secretin is used to stimulate the pancreas and bicarbonate output is measured by aspiration of duodenal contents, is the most time-honored of these tests but is rarely performed because of its complexity.103 For the test to be done properly, the duodenum and stomach have to be intubated and the samples obtained from the duodenum need to be analyzed promptly. Attempts to update the secretin stimulation test by combining it with endoscopic retrograde cholangiopancreatography have been attempted, but these modifications have not been adopted widely.104 Determination of pancreatic enzyme concentrations in stool has been advocated as a simpler screening test for pancreatic exocrine insufficiency. Direct measurement of stool chymotrypsin activity has poor sensitivity and specificity in patients with chronic diarrhea.105 Measurement of fecal elastase has only somewhat better reliability.106 In reality, the best way to determine pancreatic exocrine insufficiency may be a therapeutic trial of pancreatic enzyme supplementation. If such a trial is conducted, high doses of enzymes should be prescribed and some objective measurement, such as fecal fat excretion or weight gain, should be monitored to assess response (see Chapter 59).107 Inadequate bile salt solubilization of dietary fat can usually be inferred from the patient’s history or physical examination (e.g., cholestatic jaundice, ileal resection, or a known enterocolic fistula). If proof of the mechanism is required, analysis of a postprandial duodenal aspirate can demonstrate reduced conjugated bile acid concentrations. This test may not be available outside specialized centers, and a therapeutic trial of exogenous conjugated bile acids may be the best way of establishing the diagnosis. Supplementation with bile acids reduces steatorrhea and can often improve the patient’s nutritional status without aggravating diarrhea.108

TREATMENT The most important treatment of diarrhea is to ensure that fluid and electrolyte deficits are replenished with intravenous fluids or oral rehydration therapy. Although oral rehydration therapy is a convenient, inexpensive therapeutic option in industrialized countries, its major impact has been in decreasing the morbidity and mortality from cholera and other infectious diarrheas in less developed countries.109 Because nutrient absorption enhances sodium and fluid absorption in the jejunum, even when other forms of sodium absorption are impaired, orally ingested saline solutions that contain glucose, amino acids, or more complex nutrients that can be hydrolyzed at the brush border or intraluminally will be absorbed readily. Although the earliest oral rehydration solutions used glucose to accelerate sodium absorption, cereal-based oral rehydration solutions are now thought to be superior.110 Modifications to the formula have included hypo-osmolarity and use of amylase-resistant starch to enhance production of short-chain fatty acids in the colon, thereby stimulating colonic water and elec­ trolyte absorption.111,112 Although oral rehydration solutions

increase fluid and electrolyte absorption, they are not designed to reduce stool output, and stool weight actually may increase with use of these solutions. Use of oral rehydration solutions is precluded in patients who are vomiting frequently. Most sport drinks (e.g., Gatorade) are designed to replenish modest electrolyte losses from sweat and do not contain enough sodium to replace losses in diarrhea adequately. These solutions can be used if additional sources of sodium and absorbable nutrients (e.g., pretzels or crackers) are ingested concomitantly. Solutions that approximate the World Health Organization oral rehydration solution or cereal-based rehydration solutions more closely are available commercially (e.g., Rehydralyte, Resol, Ricalyte).

EMPIRICAL THERAPY OF ACUTE DIARRHEA

Because infection is a frequent cause of acute diarrhea, empirical trials of antibiotic therapy are often considered by physicians.113 If the prevalence of bacterial or protozoal infection is high in a community or a specific situation, empirical use of an antibiotic is logical, as in the treatment of travelers’ diarrhea with a fluoroquinolone or rifaximin, even without bacteriologic proof of infection.114 Antibiotic therapy also is often used empirically for more severely ill patients while bacterial culture results are pending. This approach has been called into question with the observation that patients in whom hemolytic-uremic syndrome develops in response to infection with Escherichia coli are more likely than those without hemolytic-uremic syndrome to have received empirical antibiotic therapy, although the occurrence of the syndrome may relate to the specific antibiotic and dose prescribed.115 Experts also advise against empirical antibiotic treatment of salmonellosis unless enteric fever is present.116 For patients with persistent diarrhea (lasting more than one week), an empirical trial of metronidazole or nitazoxanide for a protozoal infection is sometimes considered (see Chapters 107 and 109).117 Nonspecific antidiarrheal agents can reduce stool frequency, stool weight, and coexisting symptoms, such as abdominal cramps (Table 15-6). Opiates such as loperamide or diphenoxylate with atropine frequently are prescribed.118 The concern that these antiperistaltic agents slow the clearance of pathogens from the intestine has largely not been substantiated. Intraluminal agents, such as bismuth subsalicylate (Pepto-Bismol) and adsorbents (e.g., kaolin), also may help reduce the fluidity of bowel movements. Race­ cadotril, a drug that inhibits enkephalinase and thereby increases the effects of endogenous opiates on the mu opiate receptor, is available for the treatment of acute diarrhea in some countries.119

EMPIRICAL THERAPY OF CHRONIC DIARRHEA

Empirical therapy is used in patients with chronic diarrhea in three situations: (1) as temporizing or initial treatment before diagnostic testing; (2) after diagnostic testing has failed to confirm a diagnosis; and (3) when a diagnosis has been made but no specific treatment is available or specific treatment has failed to produce a cure. Generally, empirical antibiotic therapy is less useful for chronic diarrhea than for acute diarrhea, because infection is a much less likely cause. Although some clinicians try an empirical course of metronidazole or a fluoroquinolone before committing a patient to extensive diagnostic testing, this approach is not supported by data and is not recommended. In the appropriate clinical setting, therapeutic trials of pancreatic enzyme replacement and conjugated bile acid supplementation in patients with unexplained steatorrhea may be diagnostic and therapeutic (see earlier). By contrast,

Chapter 15  Diarrhea Table 15-6  Nonspecific Drug Therapy for Chronic Diarrhea DRUG CLASS

AGENT

DOSE

Opiates (mu opiate receptor selective)

Codeine Diphenoxylate Loperamide Morphine Tincture of opium Racecadotril* (acetorphan)

15-60 mg four times daily 2.5-5 mg four times daily 2-4 mg four times daily 2-20 mg four times daily 2-20 drops four times daily 1.5 mg/kg three times daily

Clonidine Octreotide Cholestyramine Colesevelam Colestipol Calcium polycarbophil Psyllium

0.1-0.3 mg three times daily 50-250 µg three times daily (subcutaneously) 4 g one to four times daily 3 tabs twice daily 4 g one to four times daily 5-10 g daily 10-20 g daily

Enkephalinase inhibitor (delta opiate receptor effects) Alpha-2 adrenergic agonist Somatostatin analog Bile acid-binding resin Fiber supplements *Not yet approved in the United States.

when pancreatic enzyme supplements or bile acid– binding resins are tried empirically for so-called idiopathic chronic diarrhea, they rarely yield satisfactory results (see later). Symptomatic treatment with an opiate often is necessary in patients with chronic diarrhea because specific treatment may not be available.118 Potent opiates such as codeine, opium, or morphine are underused in the management of these patients, largely because of fear of abuse. In fact, these agents are rarely abused by patients with chronic diarrhea, especially if a few simple measures are taken. First, the patient needs to be informed about the abuse potential of the medication and should be warned not to increase the dose without consulting the physician. Second, the dose should be low initially and titrated up until efficacy is achieved. Third, use of the opiate should be monitored closely, and the prescription should not be refilled until an interval appropriate with the anticipated usage has passed. Other agents that are sometimes used as nonspecific antidiarrheal agents include octreotide and clonidine. Octreotide, a somatostatin analog, has been shown to improve diarrhea in patients with the carcinoid syndrome and other endocrinopathies, dumping syndrome, chemotherapy-induced diarrhea, and AIDS.120 The benefit in other diarrheal diseases is less clear. Clonidine, an α-adrenergic agent that has effects on intestinal motility and transport,121 may have a special role in diabetic diarrhea, but its hypotensive effect limits its usefulness in many patients with diarrhea.122 Interest in and evidence to support the use of probiotics, ostensibly good bacteria (e.g., certain strains of lactobacilli), as therapy for diarrhea has been increasing. By modifying the colonic flora, these agents may stimulate local immunity and speed the resolution of travelers’ diarrhea, antibioticassociated diarrhea, and infantile diarrhea.123-125 Herbal remedies for diarrhea include those containing berberine (goldenseal, barberry), which appears to stimulate fluid and electrolyte absorption, and arrowroot, the mechanism of which is unknown.126,127 Stool-modifying agents such as psyllium alter stool consistency but do not reduce stool weight.128,129 They can be helpful in patients with coexisting fecal incontinence and in some patients with low stool weights (see earlier). The change from watery to semiformed stools may be sufficient to alleviate symptoms. In addition, pectin may delay transit through the proximal intestine and increase luminal vis­ cosity, thus serving as an adjunctive empirical treatment.

SELECTED DIARRHEAL SYNDROMES IRRITABLE BOWEL SYNDROME AND FUNCTIONAL DIARRHEA

Undoubtedly, the most common diagnosis made in patients with chronic diarrhea is IBS, yet only a fraction of patients with chronic diarrhea actually meet the current criteria for the diagnosis of IBS, for which abdominal pain is a central feature68,130 and diagnostic criteria have changed. In the past, painless diarrhea was considered to be a subtype of IBS, but such patients are now excluded from a diagnosis of IBS and are labeled as having functional diarrhea instead. Diarrhea in patients with IBS and functional diarrhea tends to be variable and sometimes alternates with periods of constipation.68 When measured, daily stool output is relatively low, typically less than 400 g/24 hr. Consistency varies from loose to soft and rarely is watery. The diarrhea does not awaken patients from sleep. Rectal urgency and fecal incontinence may be pronounced, especially during periods of psychological stress. Weight loss and evidence of chronic illness are uncommon (see Chapter 118). In some patients, IBS seems to be a late consequence of acute gastroenteritis.131 When typical features of IBS are absent, other diagnoses should be considered. An alternative diagnosis is carbohydrate malabsorption which can produce diarrhea of variable severity, depending on the amount of the malabsorbed carbohydrate consumed.132 In addition, cramps, excessive flatus, and bloating may be present with carbohydrate malabsorption. A carefully taken dietary history and a stool pH value lower than 6 distinguish this disorder from IBS. Caution should be exercised when using breath testing to confirm lactose intolerance in a patient with chronic diarrhea.133 Another condition that can simulate IBS is small intestinal bacterial overgrowth. The diagnosis of this condition is complex and controversial (see earlier).88 Bile acid– induced diarrhea also can vary in severity, depending on the rate of delivery of bile acids to the colon. Response to a therapeutic trial of a bile acid–binding resin may be a reasonable diagnostic test for this condition (see later). Celiac disease is another diagnosis that often is considered but is found in only 3% to 5% of patients who meet criteria for a diagnosis of IBS.134 Most patients with other causes of chronic diarrhea have been misdiagnosed as IBS at some point before the correct diagnosis is discovered.

227

228

Section III  Symptoms, Signs, and Biopsychosocial Issues MICROSCOPIC COLITIS

The terminology associated with microscopic colitis has been confusing. The term collagenous colitis was used initially to describe the histologic findings of subepithelial fibrosis and inflammation in the rectal mucosa of a woman with chronic watery diarrhea who had a normal gross appearance of the mucosa by proctoscopy. Subsequently, the term microscopic colitis was introduced to describe the histologic findings of mucosal inflammation without fibrosis in patients with chronic diarrhea and a normal endoscopic appearance to the colonic mucosa. The term lymphocytic colitis was coined later to emphasize the presence of intraepithelial lymphocytosis in these patients. The clinical and pathologic spectrum of microscopic colitis has been reviewed in the 2000s (see Chapter 124).135,136 The interrelationship between the different histologic diagnoses is not entirely clear. Because their clinical presentations are so similar and their pathologic appearances differ only in the presence or absence of a thickened subepithelial collagen table, current usage now includes lymphocytic colitis and collagenous colitis as histologic subtypes of microscopic colitis. Therefore, microscopic colitis is defined as a syndrome of watery diarrhea characterized by a normal colonoscopic appearance and histologic changes of lymphocytic-plasmacytic inflammation in the lamina propria and intraepithelial lymphocytosis, with or without thickening of the subepithelial collagen table. Microscopic colitis is a fairly common cause of chronic diarrhea of obscure origin in the general population, as well as at referral centers. At a tertiary referral center, microscopic colitis was discovered in 10% of patients with chronic diarrhea, with an even division between the lymphocytic and collagenous subtypes.93 A population-based epidemiologic study in Sweden has shown that the annual incidence of microscopic colitis is similar to that of Crohn’s disease, and the disorder was diagnosed in 10% of patients who presented with chronic nonbloody diarrhea.137 The keys to making the diagnosis are remembering to obtain biopsies of normal-appearing colonic mucosa in patients who present with chronic watery diarrhea, taking a sufficient number of biopsy specimens (6 to 10) to avoid sampling error, and having a skilled pathologist review the biopsy slides. Like idiopathic IBS, the cause or causes of microscopic colitis remain unknown. The disease occurs most frequently in middle-aged women and often occurs in association with autoimmune diseases, such as arthritis and hypothyroidism. A history of nonsteroidal anti-inflammatory drug use is frequent. Most fascinating is the tight linkage of lymphocytic and collagenous colitis with human leukocyte antigen (HLA)-DQ2 and HLA-DQ1,3 (including the HLADQ1,3 subtypes HLA-DQ1,7, HLA-DQ1,8, and HLA-DQ1,9), as in celiac disease, suggesting the possibility that similar immune mechanisms are involved in the pathogenesis of microscopic colitis and celiac disease.138 In fact, microscopic colitis has been associated with celiac disease and may be a cause of persistent diarrhea in patients with celiac disease who are treated with a gluten-free diet.139 Gluten is almost certainly not the causative antigen in microscopic colitis, however, because many patients with celiac disease continue to have histologic evidence of lymphocytic colitis despite treatment with a gluten-free diet, and elimination of gluten from the diet is not effective in other patients with microscopic colitis.140 Bacterial antigens in the colonic lumen, rather than dietary antigens, might play an important pathogenic role in microscopic colitis. Whatever the cause, mucosal inflammation is largely responsible for the diarrhea of microscopic colitis. Colonic

perfusion studies have shown that absorption of water and salt is impaired in lymphocytic colitis and collagenous colitis.141 In vitro studies using human colon specimens have demonstrated decreases in sodium chloride absorption accompanied by changes in diffusion and the function of tight junctions.142 Colonic water absorption correlates inversely with the cellularity of the lamina propria, but not with the thickness of the collagen table. Net secretion of water and salt by the colon is not noted frequently, however. Typical stool weights of 500 to 1000 g/24 hr are consistent with little or no fluid absorption by the colon. Bile acid malabsorption also may play a role in the pathogenesis of diarrhea in this condition.143 A Cochrane analysis of interventions for treating col­ lagenous colitis has concluded that budesonide has the best evidence for efficacy.144 Less well-established agents for the treatment of collagenous colitis include bismuth subsalicylate, mesalamine, and prednisone. Studies of budesonide in the treatment of lymphocytic colitis are fewer in number, and the overall evidence in support of a benefit is weaker.145 Microscopic colitis remits in most patients with time, and symptomatic therapy with an antidiarrheal drug may be an appropriate option.146 Bile acid sequestrants also may improve diarrhea in patients with this condition.143

POSTSURGICAL DIARRHEA

Gastrointestinal and biliary tract surgical procedures produce a number of changes in intestinal function that may lead to diarrhea. Although surgery for peptic ulcer is performed much less commonly now than in the past, other operations performed on the gastrointestinal and biliary tracts continue to be complicated by diarrhea.

Diarrhea after Gastric Surgery

For many years, peptic ulcer was treated surgically by vagotomy with pyloroplasty or antrectomy. The introduction of highly selective vagotomy in the 1980s led to a decrease in the frequency of postoperative diarrhea. The more traditional surgeries are still done for obstructing or malignant ulcer disease. In addition, the use of gastric bypass surgery for the treatment of obesity has increased dramatically. This operation and other bariatric surgeries are commonly associated with digestive symptoms, including diarrhea.147 Diarrhea also can occur as a complication of laparoscopic antireflux surgery, presumably because of accidental vagotomy.148 The most common syndrome seen after gastric surgery is dumping syndrome, a condition characterized by postprandial flushing, hypotension, diarrhea, and hypoglycemia (see Chapter 53). This syndrome results from unregulated gastric emptying, osmotic shifts, and the rapid release of peptide hormones from the intestine149 and can be treated successfully with a modified diet, antidiarrheal drugs, and the somatostatin analog octreotide.150 Gastric surgery may also predispose patients to small intestinal bacterial overgrowth, abnormally rapid intestinal transit, bile acid malabsorption, and pancreatic exocrine insufficiency as a result of poor stimulation of the pancreas or inadequate mixing of pancreatic enzymes with intestinal contents.

Diarrhea after Bowel Resection

Loss of intestinal surface area promotes malabsorption of fluid, electrolytes, or nutrients, depending on the portion of the bowel resected. The bowel is endowed with an excess of surface area for absorption under ordinary circumstances. Diarrhea develops after resection of the small intestine

Chapter 15  Diarrhea that leaves insufficient surface area for normal absorption, so-called short bowel syndrome (see Chapter 103).151 The process of intestinal adaptation may improve intestinal electrolyte absorption with time but cannot overcome defects in specialized functions.152 For example, removal of the ileocecal area limits the ability of the intestine to absorb sodium against its electrochemical gradient, a defect that cannot be compensated for elsewhere in the gastrointestinal tract (see Chapter 99).27 Similarly, resection of the terminal ileum results in a permanent reduction in vitamin B12 absorption, thereby necessitating parenteral vitamin B12 replacement. Ileal resection also may compromise conjugated bile acid absorption and result in bile acid–mediated fluid and electrolyte secretion by the colon.153 Various approaches can be used to correct nutritional deficiencies and intestinal dysfunction in patients with short bowel syndrome (see Chapter 103).

Ileostomy Diarrhea

Normally, 1 to 1.5 L of fluid enters the colon from the small intestine each day; an ileostomy diverts this fluid from the body. Adaptation eventually results in a decrease in flow from the small intestine to an average of 750 mL daily, provided that the patient has a sufficient length of functioning small bowel. This excess daily fluid loss usually is readily overcome by an increase in oral intake by the patient. Patients with ileostomies tolerate abnormally increased fluid losses poorly, however, and are at risk of dehydration under such circumstances. Ileostomy diarrhea is said to be present when fluid losses exceed 1000 mL daily (see Chapter 113).154 Causes of ileostomy diarrhea include stomal stenosis, partial small bowel obstruction, small intestinal bacterial overgrowth, recurrent IBD, recurrent tumor proximal to the stoma, medication-associated diarrhea, and intraperitoneal infection. In most cases, however, no specific cause is identified. A special circumstance occurs in patients with an ileal pouch formed to create a continent ileostomy or an ileoanal anastomosis after colectomy for ulcerative colitis, in whom inflammation of the pouch, so-called pouchitis, caused by bacterial overgrowth or recurrent IBD may develop.155 This condition is treated with antibiotics, such as metronidazole, with probiotics, such as Lacto­ bacillus spp., or with an anti-inflammatory drug, such as mesalamine. Idiopathic ileostomy diarrhea is treated with antidiarrheal drugs; high doses of potent opiates may be necessary. Octreotide is used if control with potent opiates is suboptimal. If ileostomy output exceeds 2000 mL daily, supplemental oral rehydration solution or intravenous fluid may have to be provided to prevent dehydration and to maintain normal urine output.

Postcholecystectomy Diarrhea

Postcholecystectomy diarrhea is relatively common, occurring in up to 20% of patients after cholecystectomy.156,157 It usually occurs shortly after cholecystectomy, but the onset may be delayed, perhaps in response to some additional unknown disturbance that develops over time. The conventional explanation for postcholecystectomy diarrhea relates to changes in the enterohepatic cycling of bile acids, but the evidence in support of this mechanism is limited. The gallbladder provides a reservoir for bile acids at night when they are not needed for digestion of fat. When the gallbladder is removed, the enterohepatic cycling of bile acids continues at night, and a substantial portion of the bile acid pool remains within the small bowel at all times. Every 90 minutes during fasting, the migrating myoelectric complex

passes through the small intestine and rapidly sweeps intestinal contents, including much of the bile acid pool in these cases, past the specialized absorptive sites in the ileum and into the colon. Increased concentrations of bile acids in the colonic lumen may inhibit fluid and electrolyte absorption and accelerate transit. Alternatively, bacterial deconjugation of bile acids in the small bowel may increase and thereby decrease ileal absorption. Some, but not all, studies have confirmed increased fecal bile acid excretion in patients with postcholecystectomy diarrhea. Postcholecystectomy diarrhea is best treated with bile acid binders taken at bedtime and perhaps at other times during the day as well. Opiate antidiarrheal agents can also be helpful for refractory cases.

BILE ACID–INDUCED DIARRHEA

The importance of bile acid malabsorption as a mechanism for producing chronic secretory diarrhea in the absence of ileal resection or disease is controversial.98,153,158 Bile acid malabsorption has been well described as the mechanism of diarrhea when ileal disease or resection allows excessive amounts of conjugated bile acid to enter the colon. Concentrations of bile acid in the colon higher than 3 to 5 mmol/L can inhibit electrolyte absorption and stimulate secretion by the colonic mucosa.159 How often this mechanism produces chronic watery diarrhea when there is no overt ileal disease or resection is unclear. In rare congenital cases, primary bile acid malabsorption is caused by mutations in the ileal sodiumdependent bile acid transporter gene (see Chapter 64), but most cases of adult-onset bile acid malabsorption are not associated with a discrete molecular defect.160-162 Accelerated transit through the small bowel and colon could account for impaired ileal bile acid absorption in the absence of ileal disease or molecular defects.163 Studies from Europe and the United States have indicated that bile acid malabsorption is common in patients with idiopathic chronic diarrhea. Reports of the therapeutic effect of bile acid–sequestering resins in this setting, however, are discrepant. In many European studies, a high proportion of patients with idiopathic secretory diarrhea responded to therapeutic doses of bile acid–sequestering resins, whereas in American studies no consistent beneficial effect has been observed. The clinical implications of these reports are clear. First, performing a sophisticated test for bile acid malabsorption, such as the selenium-75 labeled homotaurocholic acid (SeHCAT) retention test (not available in the United States) or fecal radiolabeled bile acid study is not worthwhile, because diarrhea itself may make the test abnormal. Second, such a test may not be predictive of successful therapy with a bile acid–sequestering resin. In patients with undiagnosed idiopathic diarrhea, therefore, it makes more sense to prescribe an empirical trial of a bile acid–sequestering resin. If the medication controls the diarrhea, bile acid malabsorption may be playing a role. Because such agents may also bind toxins or other luminal agonists, the possibility of a nonspecific effect must be considered.

DIARRHEA IN HOSPITALIZED PATIENTS

Diarrhea frequently develops during hospitalization, particularly in severely ill patients hospitalized for protracted periods. Common causes of diarrhea in this setting include medications, especially antibiotics, tube feedings, intestinal ischemia, and fecal impaction. Diarrhea is a side effect of many medications, including those frequently used in hospitalized patients (see Table 15-5).59 Antibiotic therapy is particularly likely to cause

229

230

Section III  Symptoms, Signs, and Biopsychosocial Issues diarrhea by at least two main mechanisms, impairing carbohydrate metabolism by the colonic bacterial flora and facilitating overgrowth of C. difficile and production of toxins by the bacteria.164 In some cases, erythromycin produces diarrhea by its motilin-like effect on gastrointestinal transit. Impaired bacterial metabolism can cause diarrhea by allowing carbohydrates and associated water to remain in the intestinal lumen.165 Ordinarily, all dietary fiber and about 20% of wheat starch evade digestion and absorption and reach the colon. Colonic bacteria ferment these carbohydrates to short-chain fatty acids, hydrogen, and carbon dioxide. These fermentation products and associated water are rapidly absorbed by the colon, and therefore diarrhea does not result. By contrast, when an antibiotic kills some of the normal colonic flora, fermentation decreases; undigested fiber and carbohydrates as well as water are retained in the lumen, thereby leading to an osmotic diarrhea. In some persons, intestinal transit may be modified by illness or by other drugs given concomitantly, thus leading to greater delivery of carbohydrate to the colon and further aggravating the diarrhea. Such diarrhea should subside when the patient is fasting. C. difficile-related diarrhea is a more serious concern.166 Hospitalized patients and residents of nursing homes are likely to be colonized by this organism; approximately 20% of hospitalized patients become colonized with C. difficile. Physical proximity and poor hand hygiene are major factors in its spread within institutions. Factors precipitating diarrhea in this condition include antibiotic therapy, chemotherapy, and altered immunity, including reduced gastric acidity resulting from administration of a proton pump inhibitor.167 C. difficile-related disease ranges in severity from modest diarrhea to life-threatening colitis, particularly when patients are colonized with a virulent strain of C. difficile (see Chapter 108). Patients may have severe pain, abdominal tenderness, and marked polymorphonuclear leukocytosis, with an increased percentage of immature forms. In less severely ill patients suspected of having this infection, stool should be tested for C. difficile toxin and, if the result is positive, appropriate antibiotic therapy should be initiated (see Chapter 108). In more severely ill patients, analysis of stool samples for toxin may not be sufficient,168 and sigmoidoscopy is performed to identify the characteristic findings of pseudomembranous colitis. Occasionally, the colitis is evident only more proximally in the colon, and colonoscopy is required to confirm the diagnosis. Therapy with metronidazole or vancomycin often suppresses the diarrhea, but recurrence is common because the organism forms spores. Probiotics, toxin-absorbing resins, and vaccines are being studied as ways to reduce recurrence.72 Diarrhea also may be a complication of enteral nutrition, although it is often the result of coexisting problems.169 Tube feeding, although more physiologic than parenteral nutrition, is still different from the normal presentation of nutrients to the intestine by oral ingestion of food, and the regulatory system of the gastrointestinal tract may not adapt to tube feeding. Some tube feeding formulas are hypertonic and may induce diarrhea by a mechanism similar to that of dumping syndrome. In such cases, a change in formula to one that is isotonic may be of benefit. In other cases, slowing the rate of infusion and thereby decreasing the delivery of nutrients to the intestine may be helpful, but this approach may be of limited value if the patient’s nutritional needs are not being met at the slower rate of infusion. Addition of an antidiarrheal agent such as loperamide or tincture of opium to the tube feeding may be necessary, although this approach

Table 15-7  Groups of Patients Predisposed to Laxative Abuse GROUP

CHARACTERISTICS

Bulimia

Usually adolescent to young women; concerned about weight or manifesting an eating disorder; may binge eat, vomit, or exercise excessively to neutralize excessive food intake Patients who relish being a diagnostic challenge; may undergo extensive testing repeatedly Dependent child or adult poisoned with laxatives by parent or caregiver to show effectiveness as caregiver; may have history of sibling who died with chronic diarrhea Patients may have disability claim pending; illness may induce concern or caring behavior in others

Munchausen syndrome Polle’s syndrome (Munchausen’s syndrome by proxy) Secondary gain

has limitations, especially for patients in whom ileus is a risk (see Chapter 120). Intestinal ischemia may develop in some hospitalized patients, especially those with hypotension or shock.170 These patients are at risk of developing bloody diarrhea caused by ischemic colitis or more profound diarrhea if small intestinal ischemia develops (see Chapter 114). The risk of fecal impaction is increased in older adults, patients on prolonged bowel rest, and those taking constipating drugs. Paradoxical or overflow diarrhea with incontinence may be the first clue to an impaction. Hospitalized patients in whom diarrhea develops should undergo a digital rectal examination to exclude fecal impaction.171

FACTITIOUS DIARRHEA

Surreptitious laxative abuse should be considered for persons in whom diarrhea remains undiagnosed, especially in those who fit into one of the following four categories (Table 15-7): (1) patients with anorexia nervosa or bulimia who use laxatives as a way of adjusting body weight (see Chapter 8); (2) those who use laxative-induced illness for secondary gain, such as disability income, or to generate concern in other persons; (3) patients with Munchausen’s syndrome, who feign illness to confound physicians; and (4) persons who are being poisoned with laxatives by their caregivers.172,173 A high index of suspicion is required for detecting laxative abuse. Physicians usually assume that patients are being truthful, but up to 15% of patients who undergo an evaluation for chronic diarrhea may be abusing laxatives surreptitiously.174 Clues to laxative abuse may be found during the evaluation for chronic diarrhea. For example, hypokalemia may suggest ingestion of stimulant laxatives, such as senna. Detection of pseudomelanosis coli, a brownish pigmentation of the colonic mucosa, suggests chronic ingestion of anthracene laxatives, such as senna or cascara (see Chapter 124). The presence of a large fecal osmotic gap suggests magnesium ingestion. In patients who belong to one of the groups listed in Table 15-7 and in those with diarrhea that remains undiagnosed after evaluation, stool samples should be analyzed for laxatives with standardized methods. Most laxatives can be detected by spectrophotometry or chromatography, but the accuracy of commercial analysis has been called into question.82 Because some patients exaggerate stool volume by

Chapter 15  Diarrhea adding urine or water, stool osmolality should be measured as well; a value lower than 290 mOsm/kg suggests dilution of the stool with water or hypotonic urine. Admixture of stool with hypertonic urine often leads to an impossibly high fecal osmolality (typically, >600 mOsm/kg) and to a negative fecal osmotic gap because of high concentrations of sodium and potassium in the urine. A negative fecal osmotic gap also can be noted when phosphate or sulfate osmotic laxatives are ingested. In today’s legal environment, unauthorized room searches for laxatives should not be conducted. When a diagnosis of laxative abuse is made, an effort should be made to confirm the diagnosis with repeated stool analyses before discussion with the patient or family. The patient should be confronted with the findings, but not before plans for the aftermath are made. Psychiatric consultation should follow the discussion with the patient because some persons who abuse laxatives become suicidal after being discovered and all patients who abuse laxatives need counseling. In cases of laxative administration by a parent or caregiver, legal proceedings should be instituted to separate the patient from the abuser. Few outcome studies of the effect of discovery of laxative abuse are available. In one study of 11 patients seen at the Cleveland Clinic, 6 said that they were improved and 5 claimed no benefit; 4 of these 5 unimproved patients sought further medical attention elsewhere for chronic diarrhea.175

IDIOPATHIC SECRETORY DIARRHEA

When an exhaustive evaluation fails to reveal a cause of chronic diarrhea and stool analysis suggests a secretory diarrhea, the diagnosis of idiopathic secretory diarrhea is made. This condition often starts suddenly in a previously healthy person and is differentiated from the many similar acute diarrheal illnesses by persisting beyond four weeks. It occurs in two forms, epidemic and sporadic.63,64 The epidemic form of secretory diarrhea occurs in outbreaks seemingly linked to contaminated food or drink.63,176-180 The initial description of this condition resulted from an outbreak in Brainerd, Minnesota, thus giving this condition its common appellation, Brainerd diarrhea. Several outbreaks have been described in the literature in different communities and even on a cruise ship. Although the epidemiology suggests an infectious cause, no causative agent has been identified in these outbreaks. Sporadic idiopathic secretory diarrhea affects persons in a fashion identical to that of the epidemic form but does not seem to be acquired easily by family members or others.64 Many affected persons give a history of travel, but to destinations not usually associated with travelers’ diarrhea. Diarrhea begins abruptly and reaches its maximum intensity soon after onset. Weight loss of up to 20 pounds is characteristic and almost always occurs within the first few months of illness and not thereafter. Empirical trials of antibiotics and bile acid–binding resins are ineffective. Nonspecific opioid antidiarrheal agents may provide symptomatic improvement. Both forms of idiopathic secretory diarrhea have a selflimited course and usually resolve within two years of onset. The resolution of idiopathic secretory diarrhea occurs gradually over two to three months. Understanding this natural history can be a solace to patients, who may otherwise feel mired in an unending illness. Idiopathic secretory diarrhea may share several clinical characteristics with functional diarrhea but in general has a more discrete onset and is associated with higher stool volumes.

Table 15-8  Frequent Diagnoses in Patients with Diarrhea of Obscure Origin Bile acid-induced diarrhea Carbohydrate malabsorption Chronic idiopathic secretory diarrhea Fecal incontinence Functional diarrhea Iatrogenic diarrhea (drugs, surgery, radiation) Irritable bowel syndrome Microscopic colitis Pancreatic exocrine insufficiency Peptide-secreting tumors Small intestinal bacterial overgrowth Surreptitious laxative ingestion

DIARRHEA OF OBSCURE ORIGIN

Physicians sometimes fail to make a specific diagnosis in patients with chronic diarrhea, despite an elaborate evaluation, and may refer these patients to centers interested in this condition. Common diagnoses resulting from reevaluation of these patients are shown in Table 15-8. Although unusual or obscure conditions that require special tests might be expected to predominate in this group of patients, most of the eventual diagnoses are straightforward and could have been made sooner.93 Fecal incontinence and iatrogenic diarrhea could be recognized with a careful history. Surreptitious laxative ingestion and microscopic colitis could be diagnosed with an appropriate index of suspicion and testing (e.g., laxative screen and colonic biopsy, respectively). Bile acid–induced diarrhea, small intestinal bacterial overgrowth, pancreatic exocrine insufficiency, and carbohydrate malabsorption could be discovered with a detailed history and a properly conducted therapeutic trial. Peptide-secreting tumors are rare, but serum peptide assays and scanning techniques (e.g., CT scanning and octreotide scanning) are widely available. Failure to make a diagnosis typically results from failure to appreciate the evidence at hand and to think through the differential diagnosis of chronic diarrhea.

KEY REFERENCES

DuPont AW, Sellin JH. Ileostomy diarrhea. Curr Treat Options Gastroenterol 2006; 9:39-48. (Ref 154.) Fernandez-Banares F, Esteve M, Salas A, et al. Systematic evaluation of the causes of chronic watery diarrhea with functional characteristics. Am J Gastroenterol 2007; 102:2520-8. (Ref 98.) Fine KD, Schiller LR. AGA technical review on the evaluation and management of chronic diarrhea. Gastroenterology 1999; 116:146486. (Ref 3.) Fine KD, Seidel RH, Do K. The prevalence, anatomic distribution, and diagnosis of colonic causes of chronic diarrhea. Gastrointest Endosc 2000; 51:318-26. (Ref 91.) Floch MH, Walker WA, Guandalini S, et al. Recommendations for probiotic use-2008. J Clin Gastroenterol 2008; 42 Suppl 2:S104-8. (Ref 125.) Kelly CP, LaMont JT. Clostridium difficile—more difficult than ever. N Engl J Med 2008; 359:1932-40. (Ref 72.) Longstreth GF, Thompson WG, Chey WD, et al. Functional bowel disorders. Gastroenterology 2006; 130:1480-91. (Ref 68.) McMahan ZH, DuPont HL. Review article: The history of acute infectious diarrhoea management—from poorly focused empiricism to fluid therapy and modern pharmacotherapy. Aliment Pharmacol Ther 2007; 25:759-69. (Ref 113.) Murphy C, Hahn S, Volmink J. Reduced osmolarity oral rehydration solution for treating cholera. Cochrane Database Syst Rev 2004; (4):CD003754. (Ref 111.) Schiller LR. Review article: Anti-diarrhoeal pharmacology and therapeutics. Aliment Pharmacol Ther 1995; 9:87-106. (Ref 118.)

231

232

Section III  Symptoms, Signs, and Biopsychosocial Issues Schiller LR. Secretory diarrhea. Curr Gastroenterol Rep 1999; 1:389-97. (Ref 16.) Schiller LR. Management of diarrhea in clinical practice: Strategies for primary care physicians. Rev Gastroenterol Disord 2008; 7(Suppl 3):S27-38. (Ref 77.) Sellin JH. A practical approach to treating patients with chronic diarrhea. Rev Gastroenterol Disord 2007; 7(Suppl 3):S19-26. (Ref 76.)

Theilman NM, Guerrant RL. Clinical practice. Acute infectious diarrhea. N Engl J Med 2004; 350:38-47. (Ref 58.) Westergaard H. Bile acid malabsorption. Curr Treat Options Gastroenterol 2007; 10:28-33. (Ref 153.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

16 Intestinal Gas Fernando Azpiroz and Michael D. Levitt

CHAPTER OUTLINE Composition and Volume of Gastrointestinal Gas  233 Gas Metabolism and Excretion  233 Mouth to Stomach  233 Small Intestine  233 Colon  234 Elimination  236 Intestinal Propulsion, Accommodation, and Tolerance of Gas  236

Intraluminal gas homeostasis is a complex phenomenon involving production, consumption, and movement of gas within the intestine and diffusion of gas between the intes­ tine and the blood.

COMPOSITION AND VOLUME OF GASTROINTESTINAL GAS Five gases—N2, O2, CO2, H2, and methane (CH4)—account for more than 99% of intestinal gas; many additional gases are present in trace concentrations. In a study of 10 healthy fasting subjects, the composition of gas within the gastroin­ testinal tract was assessed using a washout technique in which a rapid infusion of argon into the jejunum was used to flush out the intestinal gases.1 N2 was usually predomi­ nant, O2 was present in low concentrations, and the con­ centrations of CO2, H2, and CH4 were highly variable. The volume of each gas within the intestinal lumen reflects the balance between the input and output of that gas. Input may result from swallowing, chemical reactions, bacterial fer­ mentation, and diffusion from the blood, whereas output involves belching, bacterial consumption, absorption into the blood, and anal evacuation. Measurements of intestinal gas volume, originally obtained using a body plethysmo­ graph and later using a washout technique, have indicated that the volume of intestinal gas is approximately 200 mL in healthy subjects. Similar data have been reported using a specifically designed and validated computed tomography (CT) technique.2 In the fasting state, the healthy gastrointes­ tinal tract contains only about 100 mL of gas distributed almost equally among six compartments—stomach, small intestine, ascending colon, transverse colon, descending colon, and distal (pelvic) colon. Postprandially, the volume of gas increases by 65%, primarily in the pelvic colon (Figs. 16-1 and 16-2).

GAS METABOLISM AND EXCRETION Gas enters the stomach primarily via air swallowing and leaves the stomach via belching, absorption, or emptying

Clinical Gas Problems  237 Repetitive Eructation  237 Voluminous Flatus  237 Excessively Odoriferous Flatus  238 Impaired Gas Evacuation  238 Abdominal Bloating and Distention  238 Pneumatosis Cystoides Intestinalis  240

into the duodenum. In the proximal small bowel, the inter­ action of bicarbonate and acid produces large quantities of CO2. Gas-producing bacteria in the colon ferment unab­ sorbed substrates, releasing hydrogen, CO2, and sulfurcontaining gases. Other bacteria use these gases. A fraction of each bacterial gas diffuses into the blood and is elimi­ nated in the breath, with the remainder eliminated via the anus. The diffusivity of a gas across the mucosa of the gastro­ intestinal tract depends on its solubility in water; for a given partial pressure difference, CO2 diffuses much more rapidly than H2, CH4, N2, and O2. The rate and direction of diffusion of each gas is a function of the diffusivity, partial pressure difference between lumen and blood, and exposure of the gas to the mucosal surface.3 H2 and CH4 absorbed from the bowel are excreted in expired air, and breath analysis provides a simple means of assessing the volume of these gases in the gastrointestinal tract. Figure 16-3 schematically depicts the processes regulating the volume and com­ position of gas in various segments of the gastrointestinal tract.

MOUTH TO STOMACH

The absence of the gastric bubble in subjects with advanced achalasia indicates that air swallowing (rather than intra­ luminal production) is the major source of stomach gas. Ultrafast CT studies have shown that an average of 17.7 mL of air (14 mL of N2) is swallowed with a 10-mL bolus of liquid.4 Most of this swallowed air presumably is regur­ gitated, because neither absorption of intestinal gas nor excretion in flatus can account for the relatively enormous quantities of N2 that would be swallowed with the daily ingestion of 1500 mL of liquid.5 Swallowed air contains minimal CO2, and CO2 diffuses from blood into the stomach bubble. The Po2 of swallowed air is higher than that of blood, and O2 is absorbed from the stomach.3

SMALL INTESTINE

In the upper small intestine, CO2 is liberated from the inter­ action of bicarbonate and acid. Bicarbonate is delivered to the intestinal lumen via biliary, pancreatic, and small intes­ tinal secretions, and the acid is delivered via gastric secre­ tion (about 30 mEq/hr after meals6) or fatty acids released during digestion of triglycerides (about 100 mEq of acid/30 g

233

Section III  Symptoms, Signs, and Biopsychosocial Issues

Figure 16-1.  Computed tomography image analysis of abdominal gas content (green) in a healthy subject in the supine position. An anterior view is shown on the left. Note in the lateral view on the right that in the supine position, most luminal gas is located close to the anterior abdominal wall. (From Accarino A, Perez F, Azpiroz F, et al. Intestinal gas and bloating: Effect of prokinetic stimulation. Am J Gastroenterol 2008; 103:2036-42.)

70

Table 16-1  Carbohydrate-Containing Foods That May Be Malabsorbed in the Healthy Human Small Intestine

Fast Fed

Gas volume (mL)

234

Food

0 Stomach Small bowel

Right colon

TransLeft Pelvic verse colon colon colon Figure 16-2.  Abdominal gas volume in various segments of the gastrointestinal tract in the fasting and fed states as determined by computed tomography volumetric analysis. The postprandial increment in intestinal gas is located predominantly in the pelvic portion of the colon. (From Perez F, Accarino A, Azpiroz F, et al. Gas distribution within the human gut: Effect of meals. Am J Gastroenterol 2007; 102:842-9.)

of fat). The Pco2 of duodenal contents after a meal rises to 300 mm Hg in control subjects and 500 mm Hg in patients with duodenal ulcers,7 which indicates that CO2 comprises about 40% and 66% of duodenal gas in these two states, respectively. This high Pco2 causes the luminal Pn2 to fall below that of blood, and N2 diffuses from blood into the lumen. Although CO2 is absorbed rapidly during its passage through the small intestine, a study using CT-based volu­ metric analysis has shown that in the postprandial period, small bowel gas volume remains constant, whereas gas increases in the pelvic colon (see Fig. 16-2). This colonic increment occurs 99 ± 22 minutes after the meal, earlier than would be expected for gas derived from colonic fer­ mentation of food substrates. Thus, gas (N2 or CO2) of pro­ximal origin presumably is propelled caudally to increase the volume of gas in the pelvic colon.8

COLON

In the colon, luminal bacteria produce and consume intes­ tinal gases, and this activity frequently is the primary deter­ minant of anal gas output. The composition of the colonic microflora, which varies considerably among individuals, is determined by early environmental conditions as well as factors encountered later in life, such as antibiotic and dietary exposures.9 For example, chronic ingestion of high doses of an intestinally malabsorbed disaccharide (lactulose by patients with constipation or lactose by persons with intestinal lactase deficiency) results in diminished breath H2 excretion following a challenge dose of the same disac­

Complex carbohydrates (wheat, corn, potatoes) Dairy products (milk, ice cream, cottage cheese, yogurt) Dietetic candies and chewing gum Grains, fruits, vegetables Legumes (baked beans, soy beans) Soft drinks, honey

Malabsorbed Carbohydrate Resistant and retrograded starch (see text) Lactose Mannitol, sorbitol, xylitol Fiber (hemicellulose, pectin, gums, mucilage) Stachyose, raffinose Fructose

charide.10 This phenomenon appears to result from the colonic proliferation of organisms such as Bifidobacterium spp. that ferment lactose or lactulose via non-H2 releasing pathways.11 The production of the bacterial gases CO2, H2, and CH4 may indirectly increase luminal gas volumes by reducing luminal Pn2 to a value less than that of blood, thereby causing N2 to diffuse from the blood to the lumen. Similarly, the low Po2 of colonic gas results in diffusion of O2 from the blood to the lumen; however, O2 is rapidly consumed by intestinal organisms (see Fig. 16-3).

Fermentation of Unabsorbed Substrates

The colon harbors large number of hydrogen-producing bacteria that ferment undigested substrates (carbohydrates and protein), with the release of hydrogen and carbon dioxide.12-14 Newborn infants, as well as germ-free rats, excrete no H2, whereas H2 is detected within hours of bacte­ rial contamination of the gastrointestinal tract. Therefore, bacterial metabolism is the sole source of intestinal H2.15 Various carbohydrates are incompletely absorbed by the intestines of healthy subjects (Table 16-1). Most of the world’s adult population malabsorbs lactose as a result of a genetically programmed reduction in lactase synthesis. Fruits and vegetables (particularly legumes) contain indi­ gestible oligosaccharides such as stachyose and raffinose that are readily fermented by colonic bacteria.16 A fraction of the complex carbohydrate in wheat (pasta, white bread), oats, potatoes, and corn is not absorbed in the small bowel.17 In part, this malabsorption reflects the presence of starch in a physical form that resists amylase digestion. Resistance to amylase is further enhanced when starches are refrigerated and then reheated, a process that results in crystallization (retrogradation) of the starch.18 White rice flour is the only complex carbohydrate that is almost totally absorbed. A

Chapter 16  Intestinal Gas Swallowed air

Small intestine Stomach 1 2 Eructation O2 + N2 3 O2 N2 8

11

9

Sulfide Acetate

Colon Bacteria 10

Bacteria

H2 CO2 CH4 Trace gases

5 H+ + HCO3– 4

CO2

N2 6

CH4

Bacteria 7 Fermentable substrates

N2, O2, CO2, H2, CH4, trace gases

sizable fraction of the healthy population cannot completely absorb the large quantity of fructose present in soft drinks. The poor absorption of sorbitol has led to its use as a lowcalorie sugar substitute; however, this compound is readily fermented by colonic bacteria. Although fermentable fiber is commonly assumed to provide substrate for gas produc­ tion, the standard dose of a psyllium results in a minimal increase in H2 excretion.19 Some components of normal meals interfere with absorption of nutrients and thereby increase colonic gas production; for example, fiber increases starch malabsorption,20 and a pancreatic amylase inhibitor in beans slows starch digestion and absorption.21 The high fasting H2 excretion observed in small intestinal bacterial overgrowth or untreated celiac disease has been attributed to fermentation of large quantities of mucus secreted by the intestine in these conditions.22 CO2 may be the predominant gas in flatus, and the con­ centration of this gas tends to be highest during periods of rapid anal gas evacuation. The positive correlation between the concentrations of H2 and CO2 in flatus suggests that these two gases arise from shared mechanisms. Because H2 clearly is derived from bacterial fermentation, presumably flatus CO2 has a similar origin.

8

Figure 16-3.  Mechanisms of entry and elimination of intestinal gases. Air is swallowed (1) and a sizable fraction is eructated (2). Some oxygen in swallowed air diffuses into the gastric mucosa (3). The reaction of acid and bicarbonate in the duodenum yields copious CO2 (4), which diffuses into the blood (5), while N2 diffuses into the lumen (6) down the gradient established by CO2 production. In the colon, bacterial metabolism of fermentable substrates releases CO2, H2, and CH4, as well as a variety of trace gases (7). Fractions of these bacteria-derived gases are absorbed and metabolized or excreted in expired air (8). In addition, a large proportion of H2 is consumed by other bacteria to reduce sulfate to sulfide, CO2 to acetate (9), and CO2 to CH4 (10), thereby reducing the net volume of gas derived from bacterial metabolism. N2 and O2 diffuse from the blood into the colonic lumen down a gradient created by the production of gas by bacteria (11). Gas ordinarily is propelled through the gastrointestinal tract and excreted per rectum. The net result of all of these processes determines the volume and composition of intestinal gas.

Gas-Consuming Flora

Some colonic bacteria consume intraluminal gases (H2, CO2, and O2), and this catabolism may account for a considerable proportion of intraluminal gas disposal.14,23,24 The vast majority of H2 released in the intestine normally is con­ sumed by other bacteria, and the marked interindividual differences in breath H2 excretion observed after ingestion of a given dose of lactulose probably reflects differences in bacterial H2 consumption rather than differences in pro­ duction. Sulfate-reducing bacteria use H2 to reduce sulfate to sulfide. Methanogenic bacteria use H2 to reduce CO2 to CH4.25-27 This reaction uses 5 moles of gas to produce 1 mole of CH4, and thus results in a reduction in bowel gas. Although methanogens are present in the feces of almost all adults, only about 40% of adults have sufficient concentra­ tions of methanogens (106/g) to yield detectable breath CH4 concentrations.28 Sulfate-reducing bacteria and methano­ gens compete for H2, and feces of subjects usually contain one or the other type of organism. Analysis of flatus, however, has shown that subjects who excrete methane also can excrete appreciable H2S, suggesting that sulfatereducing bacteria remain active in the proximal colons of these subjects.

235

236

Section III  Symptoms, Signs, and Biopsychosocial Issues The fraction of H2 that escapes consumption is deter­ mined by the efficiency of fecal stirring and the quantity and type of H2-consuming bacteria. Because a higher hydro­ gen tension is maintained in poorly stirred feces, a greater fraction of total H2 production is consumed in the semisolid feces of the left colon than in the liquid feces of the right colon. Consumption is markedly enhanced by the presence of methanogens, which oxidize H2 more rapidly than other H2-consuming bacteria. Studies of strongly methanogenic feces have suggested that if similar high concentrations of methanogens were present throughout the colon, almost no H2 would escape consumption. CH4-producing bacteria, however, are normally present in a high concentration only in the left colon.29 As a result, H2 liberated in the right colon is not acted on by methanogens until it reaches the left colon, thus explaining the reduced but still appreciable breath H2 excretion observed in most CH4-producing sub­ jects.30 The inability of some subjects to increase breath H2 excretion after intestinal carbohydrate malabsorption prob­ ably reflects extremely efficient consumption of H2 by meth­ anogens rather than a failure to produce H2. Methanogenesis has several clinical implications. Persons who produce large quantities of CH4 have stools that con­ sistently float on water; thus, floating stools are not a reliable indicator of steatorrhea. Methanogenesis tends to be high in subjects with constipation. The most likely explanation for this association is that slow colonic transit in constipated subjects enhances the proliferation of methanogens. Based on studies showing that exposure to CH4 slows small bowel transit in dogs, however, it has been proposed that CH4 might actually be a cause, rather than an effect of, constipa­ tion. Lastly, early reports described a higher than expected prevalence of methane production in patients with colon cancer, but this finding was not confirmed by subsequent studies.

Odoriferous Gases

None of the quantitatively important gases has an odor, and the unpleasant odor of feces results from gases present in trace quantities. The intensity of the noxious odor of flatus samples positively correlates with the concentrations of hydrogen sulfide and methanethiol.31 Hydrogen sulfide is released during bacterial metabolism of sulfate, cysteine, and mucin; therefore, both exogenous and endogenous com­ pounds supply the substrate for this reaction. Methionine appears to be the favored substrate for methanethiol produc­ tion. In addition to their noxious odor, hydrogen sulfide and methanethiol have toxicities similar to that of cyanide. The colonic mucosa protects itself from the damaging effect of these compounds via a highly developed system that metab­ olizes these gases to thiosulfate.32 This detoxification mech­ anism is so efficient that negligible quantities of these gases enter the blood perfusing the colon, and hydrogen sulfide and methanethiol of intestinal origin are not excreted in the breath.33 By contrast, an odoriferous sulfur-containing gas (allyl methyl sulfide) derived from garlic is not metabolized by the intestinal mucosa and is absorbed from the gut and excreted in expired air.33

ELIMINATION

The intestinal gas that is not absorbed or metabolized is eliminated via anal evacuation. Despite the many beliefs concerning which foods cause gas, scientific data concern­ ing diet-related increases in rectal gas evacuation are sparse. Collection of gas using a rectal tube has shown a postpran­ dial excretion rate of 15 mL/hr with a low-fiber diet, 93 mL/ hr with a normal diet, 140 mL/hr after ingestion of Brussels

sprouts, and 176 mL/hr with a pork and beans diet.16 Another study in which 24-hour gas evacuation was mea­ sured has shown that with a normal diet containing 200 g of beans, 705 mL (range, 476 to 1491 mL) was excreted per anus, and 50% of this gas was hydrogen, whereas with a fiber-free diet, the evacuation rate fell to 214 mL, with a low hydrogen content.34 The average frequency of passages of gas per rectum by healthy subjects is roughly 10 times daily, with an upper limit of normal of about 20 times daily.35 Neither age nor gender significantly correlates with flatus frequency.16 The composition of flatus is highly variable, depending on fer­ mentative activity in the colon. In the presence of ferment­ able residues in the colon, the volume of flatus increases as a result of more rapid production of H2, CO2, and methane (in subjects with methanogenic flora), whereas N2 and O2 concentrations decline. Because absorbed H2 and CH4 are not metabolized, breath excretion of these gases equals their rates of absorption. Breath H2 excretion is the product of the alveolar ventilation rate and alveolar H2 concentration. Because alveolar ventilation is relatively constant in seden­ tary persons, the end-alveolar breath H2 concentration can be used as a simple indicator of total breath H2 excretion. Long-term simultaneous measurements of rectal and breath H2 excretion have been made in adult subjects maintained in an air-tight environment.36 Normally, breath H2 excretion averages about 50% of total (breath plus rectal) excretion; however, this percentage increases to 65% when production is low and decreases to 20% when production is high. Because the intestinal absorption process for H2 is not satu­ ratable, the decreasing proportion of H2 excreted in the breath with rapid production presumably is a result of more rapid propulsion of this gas to the anus.

INTESTINAL PROPULSION, ACCOMMODATION, AND TOLERANCE OF GAS The rate of propulsion of gas in the intestine toward the anus is a crucial determinant of the volume of gas present in each segment of the gastrointestinal tract and the total volume in all segments at any moment. Gas transit deter­ mines the residence time of gas in the intestinal lumen; hence, the absorption and bacterial consumption of gas are influenced by transit time, as is the composition of gas evacuated from the rectum.37 Intestinal gas transit and toler­ ance have been measured using a gas challenge test, in which a mixture of gases is constantly infused into the jejunum and anal gas output is quantified (Fig. 16-4). An initial dose-response study using infusion rates of up to 30 mL/min (1.8 L/hr) has shown that most healthy subjects evacuate gas as rapidly as it is infused, with little or no discomfort.38 The intestine actively propels gas, with transit of gas more effectively in the erect than in the supine posture.39 Transit of gas, like that of solids and liquids, is modulated by a series of reflex mechanisms; intraluminal nutrients, particularly lipids, delay gas transit,40 whereas mechanical stimulation of the intestine (e.g., mild rectal distention) has a strong prokinetic effect.41 Gas is moved along the gastrointestinal tract far more rapidly than solids and liquids, but the type of motor activity that determines gas transit is not known. Infusion of gas does not induce detectable changes in small intestinal motility as recorded by manometry.42 By contrast, studies in which tonic (sus­ tained motor) activity in the intestine was measured by means of a barostat have suggested that gas infusion induces

Chapter 16  Intestinal Gas 6000

Gas infused or evacuated (mL)

CLINICAL GAS PROBLEMS Patients frequently complain of a generic problem with gas, and the initial step for the clinician is to determine whether the patient is referring to chronic eructation, passage of excessive gas, odoriferous flatus, impaired anal gas evacua­ tion, or abdominal bloating and distention, because each of these complaints has a different pathophysiology and treatment.

30 mL/min

REPETITIVE ERUCTATION Pathophysiology

12 mL/min

4 mL/min 1 mL/min 0 0

180 Infusion period (min)

Figure 16-4.  Dose-response study in different groups of subjects (n = 6 to 12) showing gas evacuation during continuous intestinal infusion of gas. Dotted lines indicate infusion rates. Solid lines indicate passage of rectal gas. Note that gas evacuation closely parallels the infusion rates. Values are mean ± standard error; blue, 1 mL/min; green, 4 mL/min; purple, 12 mL/min; red, 30 mL/min. (From Serra J, Azpiroz F, Malagelada J-R. Intestinal gas dynamics and tolerance in humans. Gastroenterology 1998; 115:542-50.)

tonic changes—contraction orad to the infusion site and relaxation distal to the collection site. Conceivably, move­ ment and displacement of large masses of low-resistance gas is produced by subtle changes in tonic activity and capaci­ tance of the intestine that do not affect the movement of solids and liquids.43 Gas boluses infused into the left colon have been shown to elicit forceful peristaltic contractions that precede small gas expulsions,44 but this type of phasic event has not been recorded during continuous gas infusion with a barostat located inside the rectum. Therefore, these phasic events could be a response to focal distention pro­ duced by abrupt delivery of intraluminal gas. Gas transit is normally effective, but when an appreciable amount of gas is retained within the gastrointestinal tract (meteorism), subjects may develop abdominal distention and symptoms. Different experimental models of gas reten­ tion have been used to show that although abdominal distention is related to the volume of gas within the gut, perception of abdominal symptoms depends both on intes­ tinal motor activity and on the intraluminal distribution of gas.45,46 When the propulsion of jejunal gas is inhibited with glucagon, gas accumulation is not associated with symp­ toms, suggesting that the reduction in bowel tone reduces awareness of the increase in bowel gas. Gas retention is better tolerated in the colon than in the small intestine. Retention of gas in the intestine stimulates an abdominal accommodation reflex that adapts the muscular activity of the anterior abdominal wall and the diaphragm to the volume load. Considerable colonic gas retention produces relatively small increments in girth in healthy persons, because the anterior abdominal wall contracts and the diaphragm relaxes.47 Therefore, this abdominophrenic coor­ dination determines the increase in abdominal girth that results from an increase in intestinal contents.

The occasional belch expels gas swallowed with ingested solids or liquids. Repetitive eructation results from the inad­ vertent aspiration of air into the hypopharynx, most of which is immediately expelled (with inexplicable satisfac­ tion to the subject), and only a small fraction enters the stomach. Frequently, the process is triggered by emotional stress or dyspeptic symptoms that patients misinterpret as excessive gas in the gastrointestinal tract. If appreciable swallowed air enters the stomach, discomfort may increase, resulting in more air swallowing—that is, a vicious cycle develops. Thus, chronic eructation is almost always a behavioral disorder, and radiographic and endoscopic eval­ uation should be reserved for patients who have associated symptoms or signs suggestive of thoracic or abdominal pathology.48,49 Difficulty with eructation after a fundoplica­ tion for gastroesophageal reflux disease results in the gasbloat syndrome (see Chapter 43).

Treatment

The only potentially effective treatment is to provide patients with a clear-cut pathophysiologic explanation for their repetitive belching—that air swallowing rather than gas production in the gastrointestinal tract is the basis of the problem. Patients should be instructed to refrain from belching. Although many patients continue to belch, dis­ tress is diminished by an understanding of the benign nature of chronic eructation.

VOLUMINOUS FLATUS Pathophysiology

Excessive flatulence usually results from intestinal flora proficient at producing gas (most likely because of reduced gas consumption), particularly when associated with a diet rich in fermentable residues. Most subjects who complain of excessive passage of rectal gas have no apparent bowel disease and produce excess gas on a seemingly normal diet. In a minority of persons, however, excess flatulence may be attributable to a condition that results in carbohydrate malabsorption (e.g., lactose malabsorption or celiac disease). A well-documented case of severe flatulence secondary to air swallowing has been reported.50 Frequent eructation and large volumes of gas in the stomach and small bowel suggest that the problem is caused by air swallowing. Gas chromatographic analysis of flatus collected via a rectal tube can differentiate air swallowing (N2 predominant) from intraluminal production (H2, CO2, and CH4 predominant) as the source of the gas, but this test is too complex to use in clinical practice. Excessive passage of gas per anus may be a source of social embarrassment, but most affected persons do not complain of abdominal discomfort, because the normal intestine is able to propel and evacuate large gas loads

237

238

Section III  Symptoms, Signs, and Biopsychosocial Issues without symptoms. Abdominal discomfort generally indi­ cates associated irritable bowel syndrome (IBS).51

Treatment

Because there is confusion as to what constitutes “nor­ mality” with regard to the passage of gas per rectum, a useful approach is to have the patient keep a daily record of gas passages. Greater than 20 passages daily is abnormal. Some subjects have an identifiable condition that results in intestinal carbohydrate malabsorption (e.g., lactose malab­ sorption or celiac disease), and treatment of this condition reduces gas production. Most flatulent patients, however, have no demonstrable intestinal absorptive defect, and their gas is generated from carbohydrates that normally escape absorption in the small intestine. Gas production can be reduced by a diet that contains foods that provide minimal substrate to the colonic bacteria, including the following: proteins such as meat, fowl, fish and eggs; carbohydrates such as gluten-free bread, rice bread, and rice; and selected vegetables and fruits, including lettuce, tomatoes, cherries, and grapes. Foods thought to increase gas include legumes, Brussels sprouts, onions, celery, carrots, raisins, bananas, fermentable fiber, and complex starches, such as wheat and potatoes. Most patients note a dramatic decrease in the passage of gas if they consume a diet devoid of foods known to produce gas; however, long-term adherence to such a diet is difficult. Commercial preparations of β-galactosidase (Beano) are touted to enhance the digestion of the indigestible oligosaccharides present in legumes and other vegetables52; however, efficacy has only been demonstrated for the liquid preparation; tablets containing this enzyme may not be effective. Activated charcoal has been reported to reduce breath H2 excretion,53 but a second study has shown that charcoal does not bind H2 (nor any other quantitatively important intestinal gas) and does not reduce breath H2 excretion.54 Antibiotics, particularly rifaximin, have been claimed to reduce intestinal gas production,55 although decreased rectal gas excretion has not been demonstrated. Given the problems associated with chronic antibiotic therapy as well as the lack of clear-cut benefit, it seems inadvisable to use antibiotics to manipulate the flora of flatulent patients.

EXCESSIVELY ODORIFEROUS FLATUS Pathophysiology

The odor of rectal gas results from trace gases, rather than the gases present in large volumes in flatus. As noted, sul­ fur-containing gases are associated with an offensive odor. Excessively malodorous flatus may be related to an overly proficient sulfate-reducing flora and the availability of sul­ fate-containing substrates (e.g., cruciferous vegetables, some amino acids) in the colon. These gases are rapidly absorbed across the mucosa (half-time in the lumen of less than one minute56; therefore, the passage of these gases in flatus is sensitive to the time between their release from the fecal mass and arrival at the rectum.

Treatment

Theoretically, a diet low in sulfur-containing compounds (e.g., cruciferous vegetables, beer, protein) should limit the production of sulfur-containing gases; however, the ability of dietary manipulations to reduce gas odor is totally anec­ dotal. Dietary measures that reduce total gas production (see earlier) also may be helpful. Several commercial devices use activated charcoal to adsorb odoriferous gases. These

devices consist of charcoal-impregnated pads, underwear, and cushions. The efficacy of these devices has been tested by infusing hydrogen sulfide at the anus and measuring the fraction of this gas absorbed by the device. Cushions have been shown to be less effective than the pads or under­ wear.31,57 Orally administered products that have been tested for their ability to reduce the release of sulfur gases include activated charcoal (eight 260-mg tablets daily), which is ineffective,58 and the maximal dosage of bismuth subsalicy­ late, which is effective. The potential toxicity of long-term administration of bismuth subsalicylate, however, probably precludes this mode of therapy.

IMPAIRED GAS EVACUATION

In contrast to the patients with excessive flatus, some patients complain of gas retention associated with impaired fecal evacuation. Normally, evacuation of gas results from a mild increase in intra-abdominal pressure coupled with anal relaxation.59 Incoordination of this process produces functional outlet obstruction that may be associated with the sensation of difficult gas evacuation and constipation (see Chapter 18). Fecal retention prolongs colonic fermenta­ tion and could increase gas production. Patients with gas retention resulting from impaired anal evacuation may benefit from biofeedback treatment, which can help improve gas elimination and fecal evacuation.

ABDOMINAL BLOATING AND DISTENTION Pathophysiology

Symptoms commonly attributed to too much gas, such as abdominal bloating and distention, are among the most frequently encountered gastrointestinal complaints.60 Bloat­ ing is an ambiguous term that refers to subjective sensations of a swollen abdomen, full belly, abdominal pressure, or excess gas. Abdominal distention refers to an objective increase in girth. Distention usually develops following meals or at the end of the day and resolves after an overnight rest. Tape measure and x-ray measurements have demon­ strated objectively a clear-cut increase in abdominal girth with episodes of bloating, a finding confirmed by mea­ surements using inductance plethysmography and CT imaging.61-65 Some patients with IBS, particularly those with rectal hypersensitivity, however, complain of bloating in the absence of objective distention.52,63 Understanding the complex relationship between gas and bloating complaints is important, because rational treatment is based on the altered physiology. A major question is to what extent subjective bloating and objective distention are associated with or caused by an increased rate of production or volume of intestinal gas. Patients with IBS (who frequently complain of bloating) have been reported to have increased gas production caused by small intestinal bacterial overgrowth55 or malabsorption (see Chapter 118).66 This finding, however, has not been supported by other well-designed studies.67 A one-week study68 of hourly breath H2 concentrations during waking hours has shown no differences between patients with IBS and healthy subjects, despite a greater perception of bloating by the patients with IBS. Another study,69 in which total (breath plus anal) gas excretion was measured in a small number of patients with IBS and healthy subjects housed in a gas-tight environment, has demonstrated that the patients with IBS excrete more H2 but the total of H2 plus CH4 does not differ between the two groups. Multiple studies have attempted to determine whether the gastrointestinal tracts of patients with bloating contain excessive gas. Intestinal gas washout studies have shown

Chapter 16  Intestinal Gas terior diameter of the abdomen and a significant diaphrag­ matic descent, but only a modest increment in intestinal gas content.65 Electromyographic studies have shown that the abdomen normally adapts to an increase in contents via a coordinated abdominophrenic response.47 Patients with IBS have an incoordination (dyssynergia) between diaphrag­ matic contraction and anterior wall relaxation in response to increased intra-abdominal volume loads.80,81 The result is that minor increases in bowel contents that are well tolerated by healthy subjects cause discomfort, a sense of abdominal distention, and an increase in abdominal girth in patients with bloating.

6

Perception score

Healthy IBS

3

Treatment

0 –400

0

400

1600

Gas retained (mL) Figure 16-5.  Individual perception scores and gas retention after a twohour intestinal gas infusion in patients with irritable bowel syndrome (IBS; circles) and healthy subjects (squares). Perception of abdominal symptoms was scored on a seven-point scale (0 to 6). Gas retained is the volume infused minus the volume evacuated. Broken lines represent the upper limits for perception (score 3) and gas retention (400 mL) in healthy subjects. Only 4 of 20 healthy subjects progressively retained gas in a volume exceeding 400 mL. By contrast, most (18 of 20) patients with IBS retained gas (>400 mL) and/or developed moderate to severe symptoms (>3 perception score). (From Serra J, Azpiroz F, Malagelada J-R. Impaired transit and tolerance of intestinal gas in the irritable bowel syndrome. Gut 2001; 48:14-9.)

similar volumes of bowel gas in patients with IBS and healthy controls.2,70,71 Although some studies using plain abdominal radiography72,73 found that intra-abdominal gas content was greater in patients with IBS than in healthy subjects (by 54% to 118%), no significant correlations were observed between intestinal gas content and bloating. The most compelling evidence for the lack of an association between bloating and the volume of bowel gas was provided by a study in which a validated CT technique was used to show that most patients with bloating have normal volumes of bowel gas. Excessive gas was observed, however, in patients with severe motility disorders.65,74 Although the volume of bowel gas seemingly is normal in persons with bloating, multiple studies using intestinal gas infusion have shown consistently that these patients have impaired handling of the infused gas. In response to an exogenous gas load, these patients exhibit gas retention, abdominal symptoms, or both (Fig. 16-5).2,70,71 These abnor­ malities apparently reflect impaired reflex control of gas transit40,51,75,76 as well as the frequently demonstrated hyper­ sensitivity to bowel distention that characterizes patients with IBS.77-79 Therefore, gas transit studies seem to provide a sensitive method of identifying subtle intestinal motor disturbances not detectable by conventional diagnostic tests. Although total gas volume is not increased in patients with IBS, disturbances in gas propulsion may lead to local­ ized gas accumulations (e.g., the splenic flexure syndrome) that cause symptoms in the hyperreactive intestines of patients with bloating. A study of patients with IBS in which CT images were compared under basal conditions and during an episode of severe distention has demonstrated that the sensation of distention is associated with an increase in the anteropos­

Because patients with bloating and abdominal distention seem to have a common variant of IBS, the basic approach to treatment should be similar to that prescribed for IBS (see Chapter 118). These patients, however, may have a disorder resulting from a number of altered pathophysiologic mecha­ nisms. A hypersensitive gut, for example, may be associated with impaired anal evacuation, particularly in patients with constipation-predominant IBS, and symptoms will worsen if gas production is increased. In these patients, the treat­ ment strategy may need to be modified.82 Many therapies have been claimed to relieve bloating, but the few wellcontrolled studies were directed toward symptoms of IBS in general, not bloating specifically. Nonpharmacologic Therapies Intestinal clearance of perfused gas is increased by mild exercise and the erect posture, which may explain anecdotal observations that activity (as opposed to resting in the supine posture) improves bloating symptoms.39,83,84 Although intestinal gas volumes appear to be normal in patients with bloating, the sensitivity of their intestines to normal volumes of bowel contents suggests that limiting gas production to a minimum may be beneficial. Therefore, dietary manipula­ tions to reduce gas production (described earlier) may be beneficial. Low-fiber85 and low-residue diets also have been reported to improve symptoms.86 More than 20 reports on the use of probiotics to reduce symptoms of IBS have been published. The results are promising but inconsistent, possibly depending on the bacterial species used, doses, duration of treatment, and endpoints used for evaluation. Hypnosis has been reported to reduce symptoms of IBS, including bloating.87 Pharmacologic Therapies Although studies have suggested that antibiotics, particu­ larly rifaximin, can reduce symptoms of IBS,55 IBS may first appear after antibiotic therapy.88,89 Until more data are avail­ able, using antibiotics to treat bloating seems inadvisable. Simethicone has defoaming properties that eliminates bubbles that might trap gas,90 but it does not reduce the volume of gas. The effectiveness of this compound in the treatment of gas symptoms remains controversial.91 Neostigmine, a potent prokinetic agent, has been reported to reduce abdominal symptoms resulting from an intestinal infusion of gas. Chronic administration of pyridostigmine improves symptoms in patients complaining of bloating but has only marginal effects on intestinal gas content.71,74 In placebo-controlled trials, the prokinetic agents metoclo­ pramide and the restricted drug cisapride have produced statistically significant reductions in complaints of abdomi­ nal distention.92,93 Tegaserod (no longer available in the United States) also has been shown to reduce bloating and distention in some, but not all, controlled trials carried out with this agent.94

239

240

Section III  Symptoms, Signs, and Biopsychosocial Issues As noted, patients with bloating have reduced tolerance to gas infused in the intestine, and the perception of intes­ tinal distention by healthy subjects is reduced when intes­ tinal motor activity is inhibited by glucagon. Therefore, drugs that reduce intestinal motility theoretically could enhance the tolerance of the hypersensitive gut to normal volumes of gas. Anticholinergic drugs such as hyoscyamine, dicyclomine, or scopolamine inhibit intestinal motility and may reduce the response to bowel gas. A meta-analysis of the efficacy of smooth muscle relaxants in the treatment of IBS46,95 has concluded that these drugs are superior to placebo in the management of symptoms, specifically abdominal pain and distention. These drugs, however, may enhance gas retention, and anecdotal evidence suggests that their administration may actually worsen abdominal disten­ tion in some patients. Peppermint oil has an antispasmodic effect on the gastrointestinal tract because of the calcium channel blocker activity of its active constituent, menthol.96 Although some studies have reported improvement in abdominal distention and reduction in flatus emission with peppermint oil, a meta-analysis97 has indicated that its benefit in IBS is questionable. Drugs that act on the efferent nerves of the intestine also might be effective. For example, low-dose tricyclic antidepressants and possibly selective serotonin reuptake inhibitors have proven useful in the treatment of abdominal pain and symptoms of IBS, an effect probably related to their antinociceptive action (see Chapter 118).98

PNEUMATOSIS CYSTOIDES INTESTINALIS

Pneumatosis cystoides intestinalis and coli is a condition characterized by the presence of gas-filled cysts in the wall of the small bowel, colon, or both (see Chapter 124). The cysts may be asymptomatic or associated with diarrhea, bloating, or abdominal pain. Many patients with pneuma­ tosis have extremely high breath H2 concentrations, a finding indicative of high luminal concentrations of H2.99,100 The feces of three patients with pneumatosis of the colon were found to have unusually low concentrations of H2consuming organisms, and a patient with pneumatosis limited to the small intestine had small bowel contents that produced but could not consume H2. Therefore, the high luminal H2 of these subjects appears to reflect H2 production that is relatively unopposed by H2 consumption. An association between pneumatosis and chronic adminis­ tration of chloral hydrate seemingly is explained by the ability of chloral hydrate to inhibit H2 consumption by intestinal flora.101 How a high luminal H2 tension results in pneumatosis is controversial. One proposal is that the high luminal H2 results in supersaturation of tissue with H2. As a result, H2 bubbles form via a process similar to that which results in tissue collections of gas in deep sea divers.102 A second theory proposes that small intramural gas collections nor­ mally occur with some frequency, but are quickly absorbed

into the circulation. In the presence of high H2 production, rapid diffusion of luminal H2 into the cyst dilutes other cyst gases (e.g., N2).100 Thus, the cyst N2 tension remains lower than or equal to that in the blood. As a result, N2 in the cyst cannot be absorbed and the cyst persists. The most effective treatment to eliminate the cysts is the administration of high concentrations of O2 via inhalation.103 This maneuver reduces the blood N2 tension to a value below that of the cyst, allowing N2 to diffuse from the cyst into the blood, with resolution of the cyst. Other forms of therapy that may be effective are heliox (a low-density gas mixture), antibiotics that inhibit H2 production (ciprofloxacin has been used successfully in a patient with small intestinal bacterial overgrowth and pneumatosis cystoides intesti­ nalis), and dietary manipulations, such as lactose restric­ tion, that reduce the delivery of fermentable substrate to the colonic bacteria.

KEY REFERENCES

Accarino A, Perez F, Azpiroz F, et al. Intestinal gas and bloating: Effect of prokinetic stimulation. Am J Gastroenterol 2008; 103:2036-42. (Ref 74.) Agrawal A, Houghton LA, Lea R, et al. Bloating and distention in irri­ table bowel syndrome: The role of visceral sensation. Gastroenterol­ ogy 2008; 134:1882-9. (Ref 63.) Azpiroz F, Malagelada J-R. Abdominal bloating. Gastroenterology 2005; 129:1060-78. (Ref 60.) Bredenoord AJ, Smout AJ. Physiologic and pathologic belching. Clin Gastroenterol Hepatol 2007; 5:772-5. (Ref 5.) Houghton LA, Lea R, Agrawal A, et al. Relationship of abdominal bloat­ ing to distention in irritable bowel syndrome and effect of bowel habit. Gastroenterology 2006; 131:1003-10. (Ref 62.) Levitt MD, Furne J, Aeolus MR, et al. Evaluation of an extremely flatu­ lent patient: Case report and proposed diagnostic and therapeutic approach. Am J Gastroenterol 1998; 11:2276-81. (Ref 50.) Levitt MD, Furne J, Olsson S. The relation of passage of gas an ab­ dominal bloating to colonic gas production. Ann Intern Med 1996; 124:422-4. (Ref 19.) Levitt M, Olsson S. Pneumatosis cystoides intestinalis and high breath H2 excretion: Insights into the role of H2 in this condition. Gastroen­ terology 1995; 108:1560-5. (Ref 100.) Passos MC, Tremolaterra F, Serra J, et al. Impaired reflex control of intestinal gas transit in patients with abdominal bloating. Gut 2005; 54:344-8. (Ref 75.) Perez F, Accarino A, Azpiroz F, et al. Gas distribution within the human gut: Effect of meals. Am J Gastroenterol 2007; 102:842-9. (Ref 8.) Posserud I, Stotzer PO, Bjornsson ES, et al. Small intestinal bacterial overgrowth in patients with irritable bowel syndrome. Gut 2007; 56:802-8. (Ref 67.) Salvioli B, Serra J, Azpiroz F, et al. Origin of gas retention and symptoms in patients with bloating. Gastroenterology 2005; 128:574-9. (Ref 51.) Suarez F, Furne J, Springfield J, et al. Insights into human colonic physiology obtained from the study of flatus composition. Am J Physiol 1997; 272:G1028-33. (Ref 14.) Suarez FL, Springfield J, Levitt MD. Identification of gases responsible for the odour of human flatus and evaluation of a device purported to reduce this odour. Gut 1998; 43:100-4. (Ref 31.) Tremolaterra F, Villoria A, Azpiroz F, et al. Impaired viscerosomatic reflexes and abdominal wall dystony associated with bloating. Gas­ troenterology 2006; 130:1062-8. (Ref 81.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

17 Fecal Incontinence Satish S. C. Rao

CHAPTER OUTLINE Epidemiology  241 Health Care Burden  241 Pathophysiology  242 Functional Anatomy and Physiology of the Anorectum  242 Pathogenic Mechanisms  243 Evaluation  245 Clinical Features  245

Fecal incontinence is usually defined as the involuntary passage of fecal matter through the anus or the inability to control the discharge of bowel contents. The severity of incontinence can range from occasional unintentional elimination of flatus to the seepage of liquid fecal matter or the complete evacuation of bowel contents. Consequently, the problem has been difficult to characterize from an epidemiologic and pathophysiologic standpoint, but undoubtedly causes considerable embarrassment, a loss of self-esteem, social isolation, and a diminished quality of life.1

EPIDEMIOLOGY Fecal incontinence affects people of all ages, but its prevalence is disproportionately higher in middle-aged women, older adults, and nursing home residents. Estimates of the prevalence of fecal incontinence vary greatly and depend on the clinical setting, definition of incontinence, frequency of occurrence, and influence of social stigma and other factors.2 Both the embarrassment and social stigma attached to fecal incontinence make it difficult for subjects to seek health care; consequently, treatment is often delayed for several years. Fecal incontinence not only causes significant morbidity in the community, but also consumes substantial health care resources. In a U.S. householder survey, frequent leakage of stool or fecal staining for more than one month were reported by 7.1% and 0.7% of the population, respectively.3 In the United Kingdom, two or more episodes of fecal incontinence per month were reported by 0.8% of patients who presented to a primary care clinic.4 In an older self-caring population (older than 65 years), fecal incontinence occurred at least once a week in 3.7% of subjects and in more men than women (ratio of 1.5 : 1).5 The frequency of fecal incontinence increases with age, from 7% in women younger than 30 years to 22% in women in their seventh decade.6,7 By contrast, 25% to 35% of institutionalized patients and 10% to 25% of hospitalized geriatric patients have fecal incontinence.1 In the United States, fecal incontinence

Physical Examination  246 Diagnostic Testing  246 Treatment  250 Supportive Measures  250 Specific Therapies  252 Treatment of Subgroups of Patients  256

is the second leading reason for placement in a nursing home. In a survey of 2570 households, comprising 6959 individuals, the frequency of at least one episode of fecal incontinence during the previous year was 2.2%; among affected persons, 63% were women, 30% were older than 65 years, 36% were incontinent of solid stool, 54% were incontinent of liquid stool, and 60% were incontinent of flatus.1 Furthermore, in another prospective survey of patients who attended either a gastroenterology or a primary care clinic, over 18% reported fecal incontinence at least once a week.8 Only one third had ever discussed the problem with a physician. When stratified for the frequency of episodes, 2.7% of patients reported incontinence daily, 4.5% weekly, and 7.1% monthly.8 In another survey, fecal incontinence was associated with urinary incontinence in 26% of women who attended a urology-gynecology clinic.9 A high frequency of mixed fecal and urinary incontinence was also reported in nursing home residents. Persons with incontinence were 6.8 times as likely to miss work or school, and missed an average of 50 work or school days per year, compared with those without incontinence or other functional gastrointestinal symptoms.3

HEALTH CARE BURDEN The cost of health care related to fecal incontinence includes measurable components such as the evaluation, diagnostic testing, and treatment of incontinence, the use of disposable pads and other ancillary devices, skin care, and nursing care. Approximately $400 million/year is spent on adult diapers,8 and between $1.5 and $7 billion/year is spent on care for incontinence among institutionalized older patients.1,2,10 In a long-term facility, the annual cost for a patient with mixed fecal and urinary incontinence was $9,711.11 In the outpatient setting, the average cost per patient (including evaluation) has been estimated to be $17,166.12 In addition, these persons incur costs that cannot be easily measured and that result from their impaired quality of life and social dysfunction.7 Fecal incontinence

241

242

Section III  Symptoms, Signs, and Biopsychosocial Issues is most likely to affect a person’s quality of life significantly and lead to increased use of health care, predominantly in women with moderate to severe symptoms.

PATHOPHYSIOLOGY FUNCTIONAL ANATOMY AND PHYSIOLOGY OF THE ANORECTUM

A structurally and functionally intact anorectal unit is essential for maintaining normal continence of bowel contents (see Chapters 98 and 125).13 The rectum is a hollow muscular tube composed of a continuous layer of longitudinal muscle that interlaces with the underlying circular muscle. This unique muscle arrangement enables the rectum to serve both as a reservoir for stool and as a pump for emptying stool. The anus is a muscular tube 2 to 4 cm in length that at rest forms an angle with the axis of the rectum (Fig. 17-1). At rest, the anorectal angle is approximately 90 degrees; with voluntary squeeze, the angle becomes more acute, approximately 70 degrees; and during defecation the angle becomes obtuse, about 110 to 130 degrees. The anal sphincter consists of two muscular components—the internal anal sphincter (IAS), a 0.3- to 0.5-cm thick expansion of the circular smooth muscle layer of the rectum, and the external anal sphincter (EAS), a 0.6- to 1.0-cm thick expansion of the levator ani muscles. Morphologically, both sphincters are separate and heterogenous.14 The IAS is composed predominantly of slow-twitch, fatigueresistant smooth muscle and generates mechanical activity with a frequency of 15 to 35 cycles/min as well as ultraslow waves at 1.5 to 3 cycles/min.13 The IAS contributes approximately 70% to 85% of the resting anal sphincter pressure but only 40% of the pressure after sudden distention of the rectum and 65% during constant rectal distention; the remainder of the pressure is provided by the EAS.15 Therefore, the IAS is responsible chiefly for maintaining anal continence at rest. The anus is normally closed by the tonic activity of the IAS. This barrier is reinforced during voluntary squeeze by the EAS. The anal mucosal folds, together with the expansive anal vascular cushions (see later), provide a tight seal.16 These barriers are augmented by the puborectalis muscle, which forms a flap-like valve that creates a forward pull and reinforces the anorectal angle.13

Anorectal angle Levator ani muscle Puborectalis shelf

Symphysis pubis

Coccyx Internal anal sphincter muscle External anal sphincter muscle Posterior Anterior Figure 17-1.  Sagittal diagrammatic view of the anorectum. (From Rao SSC. Pathophysiology of adult fecal incontinence. Gastroenterology 2004; 126:S14-22.)

The anorectum is richly innervated by sensory, motor, and autonomic nerves and by the enteric nervous system. The principal nerve to the anorectum is the pudendal nerve, which arises from the second, third, and fourth sacral nerves (S2, S3, S4), innervates the EAS, and subserves sensory and motor function.17 A pudendal nerve block creates a loss of sensation in the perianal and genital skin and weakness of the anal sphincter muscle but does not affect rectal sensation.15 A pudendal nerve block also abolishes the rectoanal contractile reflexes (see later), an observation that suggests that pudendal neuropathy may affect the rectoanal contractile reflex response. The sensation of rectal distention is most likely transmitted along the S2, S3, and S4 parasympathetic nerves. These nerve fibers travel along the pelvic splanchnic nerves and are independent of the pudendal nerve.13 How humans perceive stool contents in the anorectum is not completely understood. Earlier studies failed to demonstrate rectal sensory awareness.13 Subsequent studies have confirmed that balloon distention is perceived in the rectum and that such perception plays a role in maintaining continence.16,18 Furthermore, sensory conditioning can improve hyposensitivity19,20 and hypersensitivity21 of the rectum. Mechanical stimulation of the rectum can produce cerebral evoked responses,22 thereby confirming that the rectum is a sensory organ. Although organized nerve endings are not present in the rectal mucosa or myenteric plexus, myelinated and unmyelinated nerve fibers are present.13 These nerves most likely mediate the distention or stretch-induced sensory responses as well as the viscerovisceral,22 rectoanal inhibitory, and rectoanal contractile reflexes. The sensation of rectal distention is most likely transmitted via the parasympathetic nervi erigentes along the S2, S3, and S4 splanchnic nerves. Rectal sensation and the ability to defecate can be abolished completely by resection of the nervi erigentes.23 If parasympathetic innervation is absent, rectal filling is perceived only as a vague sensation of discomfort. Even persons with paraplegia or sacral neuronal lesions may retain some degree of sensory function, but almost no sensation is felt if lesions occur in the higher spine.15,18,24 Therefore, the sacral nerves are intimately involved in the maintenance of continence. The suggestion has been made that bowel contents are sensed periodically by anorectal sampling,25 the process whereby transient relaxation of the IAS allows the stool contents from the rectum to come into contact with specialized sensory organs in the upper anal canal. Specialized afferent nerves may exist that subserve sensations of touch, temperature, tension, and friction, but the mechanisms are incompletely understood.13 Incontinent patients appear to sample rectal contents less frequently than continent subjects. The likely role of anal sensation is to facilitate discrimination between flatus and feces and the fine-tuning of the continence barrier, but its precise role has not been well characterized. Rectal distention is associated with a fall in anal resting pressure known as the rectoanal inhibitory reflex. The amplitude and duration of this relaxation increases with the volume of rectal distention. This reflex is mediated by the myenteric plexus and is present in patients in whom the hypogastric nerves have been transected and in those with a spinal cord lesion. The reflex is absent after transection of the rectum, but it may recover.18 Although the rectoanal inhibitory reflex may facilitate discharge of flatus, rectal distention is also associated with a rectoanal contractile response, a subconscious reflex effort to prevent release of rectal contents, such as flatus.26,27 This contractile response involves contraction of the EAS and is

Chapter 17  Fecal Incontinence mediated by the pelvic splanchnic and pudendal nerves. The amplitude and duration of the rectoanal contractile reflex also increases with rectal distention, up to a maximum volume of 30 mL. Abrupt increases in intra-abdominal pressure, as caused by coughing or laughing, are associated with an increase in anal sphincter pressure. A number of mechanisms, including reflex contraction of the puborectalis, may be involved. The blood-filled vascular tissue of the anal mucosa also plays an important role in producing optimal closure of the anus. An in vitro study has shown that even during maximal involuntary contraction, the internal sphincter ring is unable to close the anal orifice completely, and a gap of approximately 7 mm remains. This gap is filled by the anal cushions, which may exert pressures of up to 9 mm Hg and thereby contribute 10% to 20% to the resting anal pressure.26

PATHOGENIC MECHANISMS

Fecal incontinence occurs when one or more mechanisms that maintain continence is disrupted to the extent that other mechanisms are unable to compensate. Therefore, fecal incontinence is often multifactorial.2,27 In a prospective study, 80% of patients with fecal incontinence had more than one pathogenic abnormality (Fig. 17-2).13 Although the pathophysiologic mechanisms often overlap, they can be categorized under four broad groups, as summarized in Table 17-1.

cushions may lead to a poor seal and an impaired sampling reflex. These changes may cause passive incontinence or fecal seepage (see later), often under resting conditions. Both sphincters may be defective in many patients. The extent of muscle loss can influence the severity of incontinence.13 The most common cause of anal sphincter disruption is obstetric trauma, which may involve the EAS, IAS, or pudendal nerves. Why most women who have sustained an obstetric injury in their 20s or 30s typically do not present with fecal incontinence until their 50s, however, is unclear. In a prospective study, 35% of primiparous (normal antepartum) women showed evidence of anal sphincter disrup-

Anal sphincter dysfunction Pudendal neuropathy Impaired rectal sensation Poor rectal compliance Other

Abnormal Anorectal and Pelvic Floor Structures

Anal Sphincter Muscles Disruption or weakness of the EAS muscle causes urgerelated or diarrhea-associated fecal incontinence. In contrast, damage to the IAS muscle or anal endovascular

0

20

40

60

80

100

Frequency (%) Figure 17-2.  Relative frequencies of the common mechanisms that lead to fecal incontinence.

Table 17-1  Mechanisms, Causes, and Pathophysiology of Fecal Incontinence mechanism

CAUSES

Abnormal Anorectal or Pelvic Floor Structures Anal sphincter muscle Hemorrhoidectomy, neuropathy, obstetric injury Puborectalis muscle Aging, excessive perineal descent, trauma Pudendal nerve Excessive straining, obstetric or surgical injury, perineal descent Nervous system, spinal cord, autonomic Avulsion injury, spine surgery, diabetes mellitus, nervous system head injury, multiple sclerosis, spinal cord injury, stroke Rectum Aging, inflammatory bowel disease, irritable bowel syndrome, prolapse, radiation Abnormal Anorectal or Pelvic Floor Function Impaired anorectal sensation Autonomic nervous system disorders, central nervous system disorders, obstetric injury Fecal impaction Dyssynergic defecation Altered Stool Characteristics Increased volume and loose consistency Drugs, bile salt malabsorption, infection, inflammatory bowel disease, irritable bowel syndrome, laxatives, metabolic disorders Hard stools, retention Drugs, dyssynergia Miscellaneous Physical mobility, cognitive function Aging, dementia, disability Psychosis Willful soiling Drugs* Anticholinergics Antidepressants Caffeine Laxatives Muscle relaxants Food intolerance Fructose, lactose, or sorbitol malabsorption *Pathophysiology is noted for each class of drugs.

PATHOPHYSIOLOGY Sphincter weakness, loss of sampling reflex Obtuse anorectal angle, sphincter weakness Sphincter weakness, sensory loss, impaired reflexes Loss of sensation, impaired reflexes, secondary myopathy, loss of accommodation Loss of accommodation, loss of sensation, hypersensitivity Loss of stool awareness, rectoanal agnosia Fecal retention with overflow, impaired sensation Diarrhea and urgency, rapid stool transport, impaired accommodation Fecal retention with overflow Multifactorial changes Multifactorial changes Constipation Altered sensation, cconstipation Relaxation of sphincter tone Diarrhea Relaxation of sphincter tone Diarrhea, flatus

243

244

Section III  Symptoms, Signs, and Biopsychosocial Issues tion after vaginal delivery.28,29 Other important risk factors include a forceps-assisted delivery, prolonged second stage of labor, large birth weight, and occipitoposterior presentation.13 A prospective study of 921 primiparous women has shown that the frequencies of fecal incontinence at 6 weeks and 6 months postpartum are 27% and 17%, respectively, in subjects with vaginal delivery and a sphincter tear; 11% and 8%, respectively, in subjects with vaginal delivery but without a tear; and 10% and 7.6%, respectively, in subjects who underwent cesarean section.30 This study showed clearly that the occurrence and severity of fecal incontinence were attributable to an anal sphincter tear that occurred at the time of vaginal delivery. Episiotomy is believed to be a risk factor for anal sphincter disruption. In one study, medial episiotomy was asso­ ciated with a ninefold higher risk of anal sphincter dysfunction.31 Regardless of the type of delivery, however, incontinence of feces or flatus occurred in a surprisingly large percentage of middle-aged women, thereby suggesting that age-related changes in the pelvic floor may predispose to fecal incontinence. Aging affects anal sphincter function.32 In men and women older than 70 years, sphincter pressures decrease by 30% to 40% compared with younger persons.33 Also, in all age groups, anal squeeze pressure is lower in women than men,33 with a rapid fall after menopause.34 Estrogen receptors have been identified in the human striated anal sphincter, and ovariectomy in rats leads to atrophy of the striated anal sphincter muscle.13,35 These observations suggest that the strength and vigor of the pelvic floor muscles are influenced by hormones. Pudendal nerve terminal motor latency (PNTML) is prolonged in older women, and pelvic floor descent is excessive on straining.36 These mechanisms may contribute to progressive damage to the striated anal sphincter muscle. Aging is also associated with increased thickness and echogenicity of the IAS.37 Other causes of anatomic disruption include anorectal surgery for hemorrhoids, fistulas, and fissures. Anal dilation or lateral sphincterotomy may result in incontinence because of fragmentation of the anal sphincters.38 Hemorrhoidectomy can cause incontinence by inadvertent damage to the IAS39 or loss of endovascular cushions. Accidental perineal trauma or a pelvic fracture may also cause direct sphincter trauma that leads to fecal incontinence,40 but anoreceptive intercourse is not associated with anal sphincter dysfunction.41 Finally, IAS dysfunction may also occur because of myopathy, degeneration, or radiotherapy.13 Puborectalis Muscle The puborectalis muscle is also important for maintaining continence by forming a flap valve mechanism.42 Studies using three-dimensional ultrasound have shown that 40% of women with fecal incontinence have major abnormalities, and another 32% have minor abnormalities of the puborectalis muscle, compared with 21% and 32%, respectively, of asymptomatic parous controls.43 Also, assessment of puborectalis function by a perineal dynamometer revealed impaired puborectalis (levator ani) contraction in patients with fecal incontinence, and this finding was an independent risk factor for and correlated with the severity of fecal incontinence.44 Furthermore, improvement in puborectalis strength following biofeedback therapy was associated with clinical improvement, in part because the upper portion of the puborectalis muscle receives its innervations from branches of the S3 and S4 sacral nerves rather than the pudendal nerve. Therefore, the puborectalis muscle and EAS have separate neurologic innervations. Consequently, pudendal blockage does not abolish voluntary con-

traction of the pelvic floor45 but completely abolishes EAS function.15 Nervous System Intact innervation of the pelvic floor is essential for maintaining continence. Sphincter degeneration secondary to pudendal neuropathy and obstetric trauma may cause fecal incontinence in women.28 The neuropathic injury is often sustained during childbirth, probably as a result of stretching of the nerves during elongation of the birth canal or direct trauma during the passage of the fetal head. The nerve damage is more likely to occur when the fetal head is large, the second stage of labor is prolonged, or forceps are applied, especially with a high-forceps delivery or prolonged labor. The role of extrinsic autonomic innervation is somewhat controversial. Animal studies have shown that the pelvic nerves convey fibers that relax the rectum.46 Consequently, these nerves may play a role in accommodating and storing feces and gas. Damage to the pelvic nerves may lead to impaired accommodation and rapid transit through the rectosigmoid region, thereby overwhelming the continence barrier mechanisms. Sympathetic efferent activity, as studied by stimulating the presacral sympathetic nerves, tends to relax the IAS, whereas parasympathetic stimulation may cause contraction of the anal sphincter. The upper motor neurons for voluntary sphincter muscle lie close to those that innervate the lower limb muscles in the parasagittal motor cortex, adjacent to the sensory representation of the genitalia and perineum in the sensory cortex.13 Consequently, damage to the motor cortex from a central nervous system (CNS) lesion may lead to incontinence. In some patients with neurogenic incontinence, the sensory and motor nerve fibers may be damaged, resulting in sensory impairment.47 This damage can impair conscious awareness of rectal filling as well as the associated reflex responses in the striated pelvic floor sphincter muscles. Approximately 10% of patients with fecal incontinence may have a lesion more proximal than the intrapelvic or perianal nerves. The primary abnormality in these patients is cauda equina nerve injury,48 which may be occult and not evident through clinical evaluation. These patients have a prolongation of nerve conduction along the cauda equina nerve roots without an abnormality in PNTML.49 In a minority of patients, however, a combination of peripheral and central lesions is present. Other disorders such as multiple sclerosis, diabetes mellitus, and demyelination injury (or toxic neuropathy from alcohol or traumatic neuropathy) may also lead to incontinence.13 Rectum The rectum is a compliant reservoir that stores stool until social conditions are conducive to its evacuation.2 If rectal wall compliance is impaired, a small volume of stool material can generate a high intrarectal pressure that can overwhelm anal resistance and cause incontinence.50 Causes include radiation proctitis, ulcerative colitis, or Crohn’s disease, infiltration of the rectum by tumor, and radical hysterectomy.51 Similarly, rectal surgery, particularly pouch surgery,52 and spinal cord injury53 may be associated with loss of rectal compliance.

Abnormal Anorectal and Pelvic Floor Function

Impaired Anorectal Sensation An intact sensation not only provides a warning of imminent defecation, but also helps distinguish among formed stool, liquid feces, and flatus. Older persons,54 those who are physically and mentally challenged, and children with

Chapter 17  Fecal Incontinence fecal incontinence55 often show blunted rectal sensation. Impaired rectal sensation may lead to excessive accumulation of stool, thereby causing fecal impaction, megarectum (extreme dilatation of the rectum), and fecal overflow. Causes of impaired sensation include neurologic damage such as multiple sclerosis, diabetes mellitus, and spinal cord injury.53 Less well known is that analgesics (particularly opiates) and antidepressants also may impair rectal sensation and produce fecal incontinence. The importance of the rectum in preserving continence has been demonstrated conclusively through surgical studies in which preservation of the distal 6 to 8 cm of the rectum, along with its parasympathetic nerve supply, helped subjects avoid incontinence.56 By contrast, rectal sensation and the ability to defecate can be abolished completely by resection of the nervi erigentes (see earlier).23 An intact sampling reflex allows an individual to choose whether to discharge or retain rectal contents. Conversely, an impaired sampling reflex may predispose a subject to incontinence.25 The role of the sampling reflex in maintaining continence, however, remains unclear. In children who have undergone colonic pull-through surgery (see Chapter 113), some degree of sensory discrimination is preserved.57 Because the anal mucosal sensory zone is absent in these children, the suggestion has been made that sensory receptors, possibly located in the puborectalis muscle, may play a role in facilitating sensory discrimination. Also, traction on the muscle is a more potent stimulus for triggering defecation and a sensation of rectal distention. Because abolition of anal sensation by the topical application of 5% lidocaine does not reduce resting sphincter pressure (although it affects voluntary squeeze pressure but does not affect the ability to retain saline infused into the rectum), the role of anal sensation in maintaining fecal continence has been questioned.13

of large-volume liquid stools, which often transit the hindgut rapidly, continence can only be maintained through intact sensation and a strong sphincteric barrier. Similarly, in patients with bile salt malabsorption, lactose or fructose intolerance, or rapid dumping of osmotic material into the colon, colonic transit of gaseous and stool contents is too rapid and can overwhelm the continence mechanisms (see Chapters 15 and 101).2

Dyssynergic Defecation and Incomplete Stool Evacuation In some patients, particularly older adults, prolonged retention of stool in the rectum or incomplete evacuation may lead to seepage of stool or staining of undergarments.54 Most of these patients show obstructive or dyssynergic defecation,58 and many of them also exhibit impaired rectal sensation, whereby anal sphincter and pudendal nerve function is intact but the ability to evacuate a simulated stool is impaired. Similarly, in older adults and in children with functional incontinence, the prolonged retention of stool in the rectum can lead to fecal impaction. Fecal impaction may also cause prolonged relaxation of the IAS, thereby allowing liquid stool to flow around impacted stool and to escape through the anal canal (see Chapter 18).55

CLINICAL FEATURES

Descending Perineum Syndrome In women with long-standing constipation and a history of excessive straining for many years (perhaps even without prior childbirth), excessive straining may lead to progressive denervation of the pelvic floor muscles.59 Most of these patients demonstrate excessive perineal descent and sphincter weakness, which may lead to rectal prolapse; however, fecal incontinence is not an inevitable consequence. Whether or not incontinence develops will depend on the state of the pelvic floor and the strength of the sphincter muscles.

Altered Stool Characteristics

The consistency, volume, and frequency of stool and the presence or absence of irritants in stool also may play a role in the pathogenesis of fecal incontinence.2 In the presence

Miscellaneous Mechanisms

Various medical conditions and disabilities may predispose to fecal incontinence, particularly in older adults. Immo­ bility and lack of access to toileting facilities are primary causes of fecal incontinence in this population.60 Several drugs may inhibit sphincter tone. Some are used to treat urinary incontinence and detrusor instability, including anticholinergics such as tolterodine tartarate (Detrol) and oxybutynin (Ditropan) and muscle relaxants such as baclofen (Lioresal), and cyclobenzaprine (Flexeril). Stimulants such as caffeinated products, fiber supplements, or laxatives may produce fecal incontinence by causing diarrhea.13

EVALUATION The evaluation of the patient with fecal incontinence includes a detailed clinical assessment and appropriate physiologic and imaging tests of the anorectum. These three sources of information are complementary and should provide useful data regarding the severity of the problem, underlying causative factors, and impact on the patient’s quality of life. On the basis of this information, appropriate treatment strategies can be designed. The first step in the evaluation of a patient with fecal incontinence is to establish a trusting relationship with the patient and assess the duration and nature of the symptoms, with specific attention to whether the leakage consists of flatus, liquid stool, or solid stool and to the impact of the symptoms on the quality of the patient’s life (Table 17-2). Because many people misinterpret fecal incontinence as diarrhea or urgency,61 a detailed characterization of the symptom(s) is important. The clinician should ask about the use of pads or other devices and the patient’s ability to discriminate between formed or unformed stool and gas (the lack of such discrimination is termed rectal agnosia).2 An obstetric history; history of coexisting conditions such as diabetes mellitus, pelvic radiation, neurologic problems, or

Table 17-2  Features of the History That Should Be Elicited from a Patient with Fecal Incontinence Onset and precipitating event(s) Duration and timing Severity Stool consistency and rectal urgency History of fecal impaction Coexisting problems (e.g., diarrhea, inflammatory bowel disease) Drugs, caffeine, diet Past history—spine surgery, urinary incontinence, back injury, diabetes mellitus, neurologic disorders Clinical subtypes—passive or urge incontinence or fecal seepage Obstetric history—use of forceps, tears, presentation of the infant, repairs

245

246

Section III  Symptoms, Signs, and Biopsychosocial Issues spinal cord injury; dietary history; and history of coexisting urinary incontinence are important. A prospective stool diary can be useful. The circumstances under which incontinence occurs should also be determined. Such a detailed inquiry may facilitate the recognition of the following types of fecal incontinence: 1. Passive incontinence, the involuntary discharge of fecal matter or flatus without any awareness. This pattern suggests a loss of perception or impaired rectoanal reflexes, with or without sphincter dysfunction. 2. Urge incontinence, the discharge of fecal matter or flatus despite active attempts to retain these contents. The predominant causes of this pattern are disruption of the sphincter function and a decrease in rectal capacity to retain stool. 3. Fecal seepage, the undesired leakage of stool, often after a bowel movement, with otherwise normal continence and evacuation. This condition results primarily from incomplete evacuation of stool or impaired rectal sensation.54,58 Sphincter function and pudendal nerve function are mostly intact. Although overlap exists among these three types, by determining the predominant pattern, useful insights can be gained regarding the underlying mechanism(s) and preferred management. Symptom assessment, however, may not correlate well with manometric findings (see later). In one study, leakage had a sensitivity of 98.9%, specificity of 11%, and positive predictive value of 51% for detecting a low resting anal sphincter pressure on manometry.62 The positive predictive value for detecting a low anal squeeze pressure was 80%. Therefore, for an individual patient with fecal incontinence, the history and clinical features alone are insufficient to define the pathophysiology, and objective testing is essential63,64 (see later). On the basis of the clinical features, several grading systems have been proposed. A modification of the Cleveland Clinic grading system65 has been validated by the St. Mark’s investigators66 and provides an objective method of quantifying the degree of incontinence. It can also be useful for assessing the efficacy of therapy. This grading system is based on seven parameters that include the following: (1-3) the character of the anal discharge as solid, liquid, or flatus; (4) the degree of alterations in lifestyle; (5, 6) the need to wear a pad or take antidiarrheal medication; and (7) the ability to defer defecation. The total score ranges from 0 (continent) to 24 (severe incontinence). As noted earlier, however, clinical features alone are insufficient to define the pathophysiology. The use of validated questionnaires such as the SCL-90R (Symptom Checklist-90-R) and SF-36 (Short-Form 36) surveys may provide additional information regarding psychosocial issues and the impact of fecal incontinence on the patient’s quality of life.

PHYSICAL EXAMINATION

A detailed physical examination, including a neurologic examination, should be performed in any patient with fecal incontinence, because incontinence may be secondary to a systemic or neurologic disorder. The focus of the examination is on the perineum and anorectum. Perineal inspection and digital rectal examination are best performed with the patient lying in the left lateral position and with good illumination. On inspection, the presence of fecal matter, prolapsed hemorrhoids, dermatitis, scars, skin excoriations, or a gaping anus and the absence of perianal creases may be noted. These features suggest sphincter weakness or chronic skin irritation and provide clues regarding the underlying cause.2 Excessive perineal descent or rectal prolapse can be

demonstrated by asking the patient to attempt defecation. An outward bulge that exceeds 3 cm is usually defined as excessive perineal descent (see Chapter 18).67 Perianal sensation should be checked. The anocutaneous reflex examines the integrity of the connections between the sensory nerves and the skin; the intermediate neurons in spinal cord segments S2, S3, and S4; and the motor innervation of the external anal sphincter. This reflex can be assessed by gently stroking the perianal skin with a cotton bud in each perianal quadrant. The normal response consists of a brisk contraction of the external anal sphincter (“anal wink”). An impaired or absent anocutaneous reflex suggests either afferent or efferent neuronal injury.2 After inserting a lubricated, gloved index finger into the anus and rectum, one should assess the resting sphincter tone, length of the anal canal, strength of the puborectalis sling, acuteness of the anorectal angle, strength of anal sphincter squeeze, and elevation of the perineum during voluntary squeeze. Also, the presence of a rectocele or impacted stool may be noted. The accuracy of the digital rectal examination has been assessed in several studies. In one study of 66 patients, digital rectal examination by an experienced surgeon correlated somewhat with resting sphincter pressure (r = 0.56; P < 0.001) or maximum squeeze pressure (r = 0.72; P < 0.001).68 In a study of 280 patients with various anorectal disorders, a reasonable correlation was reported between digital examination and manometric findings, but the sensitivity, specificity, and positive predictive values of digital examination were low.69 In another study of 64 patients, the agreement between digital rectal examination and resting or squeeze pressure was 0.41 and 0.52, respectively.70 These data suggest that digital rectal examination provides only an approximation of sphincter strength. The findings are influenced by many factors, including the size of the examiner’s finger, technique used, and cooperation of the patient. One study has shown that trainees lack adequate skills for recognizing the features of fecal incontinence on digital rectal examination.71 Therefore, digital rectal examination is not reliable and is prone to interobserver differences. Digital rectal examination can identify patients with fecal impaction and overflow but is not accurate for diagnosing sphincter dysfunction and should not be used as the basis for decisions regarding treatment.2

DIAGNOSTIC TESTING

The first step in assessing a patient with fecal incontinence is to determine whether the incontinence is secondary to diarrhea or independent of stool consistency. If diarrhea coexists with incontinence, appropriate tests should be performed to identify the cause of the diarrhea (see Chapter 15). Such testing may include flexible sigmoidoscopy or colonoscopy to exclude colonic mucosal inflammation, a rectal mass, or stricture and stool studies for infection, volume, osmolality, electrolytes, fat content, and pancreatic function. Biochemical tests should be performed to look for thyroid dysfunction, diabetes mellitus, and other metabolic disorders. Breath tests may be considered for lactose or fructose intolerance or small intestinal bacterial overgrowth.2 A history of cholecystectomy may suggest bile salt malabsorption and prompt a therapeutic trial of a bile salt– binding agent. Specific tests are available for defining the underlying mechanisms of fecal incontinence and are often used in complementary fashion. The most useful tests are anorectal manometry, anal endosonography, the balloon expulsion test, and PNTML.2,72-74

Chapter 17  Fecal Incontinence Anorectal Manometry and Sensory Testing

Anorectal manometry is a simple and useful method for assessing IAS and EAS pressures (Fig. 17-3) as well as rectal sensation, rectoanal reflexes, and rectal compliance. Several types of probes and pressure recording devices are available. Each system has distinct advantages and drawbacks. A water-perfused probe with multiple closely spaced sensors is commonly used.2 Increasingly, a solid-state probe with microtransducers or air-filled miniaturized balloons is used. A novel solid-state probe with 36 circumferential sensors spaced at 1-cm intervals, with a 4.2-mm outer diam100

Squeeze

Rest

Anal canal 1.0 cm (mm Hg)

Anal canal 2.5 cm (mm Hg)

Rectum (mm Hg)

80

A

60 40 20 0 100 80 60 40 20 0 100 80

eter (Sierra Scientific Instruments, Los Angeles) has been reported to provide higher resolution than older style probes.75 This device uses a novel pressure transduction technology (TactArray) that allows each of the 36 pressure sensing elements to detect pressure over a length of 2.5 mm and in each of 12 radially dispersed sectors. The data can be displayed in isobaric contour plots that can provide a continuous dynamic representation of pressure changes, although anal sphincter pressures are higher than those recorded with water-perfused manometry. A high-definition manometry system with 256 circumferentially arrayed sensors in a 5-cm probe76 has become available and may provide anal sphincter pressure profiles and topographic changes of even higher fidelity (Fig. 17-4). Anal sphincter pressures can be measured by stationary or station pull-through techniques.73,74 Resting anal sphincter pressure predominantly represents IAS function, and voluntary anal squeeze pressure predominantly represents EAS function. Patients with fecal incontinence have low resting and low squeeze pressures (see Figs. 17-3 and 17-4), indicating IAS and EAS weakness.2,69 The duration of sustained squeeze pressure provides an index of sphincter muscle fatigue. The ability of the EAS to contract reflexively can be assessed during abrupt increases in intra-abdominal pressure, as when the patient coughs. This reflex response causes the anal sphincter pressure to rise above that of the rectal pressure to preserve continence. The response may be triggered by receptors in the pelvic floor and mediated through a spinal reflex arc. In patients with a spinal cord

60 Posterior Maximal & Sustained Squeeze

40 20

Anterior

Posterior

0 13:29:30 100

13:30:00 Squeeze

13:30:30

13:31:00

Rest

Anal canal 1.0 cm (mm Hg)

Anal canal 2.5 cm (mm Hg)

Rectum (mm Hg)

80

B

60 40

A

20 0 100

Maximal Squeeze

80 60 40 20 0 100 80 60

B

40 20 0 13:29:30

13:30:00

13:30:30

13:31:00

Figure 17-3.  Anorectal manometry profiles in (A) a healthy normal subject in whom resting (internal anal sphincter) and squeeze (external anal sphincter) pressures are normal and (B) a patient with fecal incontinence in whom resting and squeeze pressures are weak. Upper tracings, Rectal pressure activity; middle tracings, anal pressure activity at 2.5 cm; lower tracings, anal pressure activity at 1.0 cm from the anal margin.

low pressure area

3D: SAGITTAL VIEW 2D: UNFOLDED VIEW Figure 17-4.  High dynamic anal sphincter vector topography showing pressure changes during maximal squeeze in three-dimensional (3D) sagittal view (left) and two-dimensional (2D) unfolded view (right). A, Changes in a healthy control subject. B, Changes in a subject with fecal incontinence. The subject with incontinence has significant weakness of the anal sphincter, with an asymmetrical squeeze and a change in some vectors (pre­ dominantly yellow and green), whereas the healthy subject shows a robust squeeze (orange and red) and symmetrical decrease in sphincter diameter.

247

248

Section III  Symptoms, Signs, and Biopsychosocial Issues lesion above the conus medullaris, this reflex response is preserved, even though voluntary squeeze may be absent, whereas in patients with a lesion of the cauda equina or sacral plexus, both the reflex and voluntary squeeze responses are absent.2,77,78

Rectal Sensory Testing

Rectal balloon distention with air or water can be used to assess sensory responses and compliance of the rectal wall. By distending a balloon in the rectum with incremental volumes, the thresholds for first perception, first desire to defecate, and urgent desire to defecate can be assessed. A higher threshold for sensory perception indicates reduced rectal sensitivity.2,77,79 The balloon volume required for partial or complete inhibition of anal sphincter tone also can be assessed. The volume required to induce reflex anal relaxation is lower in incontinent patients than in controls.80 Because sampling of rectal contents by the anal mucosa may play an important role in maintaining continence,25 quantitative assessment of anal perception using electrical or thermal stimulation has been advocated but is not used clinically.2 Rectal compliance can be calculated by assessing the changes in rectal pressure during balloon distention with air or fluid.73,81 Rectal compliance is reduced in patients with colitis,50 patients with a low spinal cord lesion, and diabetic patients with incontinence but is increased in those with a high spinal cord lesion. Anorectal manometry can provide useful information regarding anorectal function.72,73,82 The American Motility Society has provided consensus guidelines and minimal standards for manometry testing.74 Although there are insufficient data regarding normal values, overlap betwen healthy subjects and patients with fecal incontinence,69 and large confidence intervals for test reproducibility,83 manometry testing can be useful for the individual patient with fecal incontinence.74 Manometric tests of anorectal function may also be useful for assessing objective improvement following drug therapy, biofeedback therapy, or surgery.84-86

Imaging the Anal Canal

Anal Endosonography Anal endosonography is performed by using a 7- to 15-mHz rotating transducer with a focal length of 1 to 4 cm.87 The test provides an assessment of the thickness and structural integrity of the EAS and IAS and can detect scarring, loss of muscle tissue, and other local pathology (Fig. 17-5).88 Higher frequency (10- to 15-mHz) probes that provide

better delineation of the sphincter complex have become available.88 After vaginal delivery, anal endosonography has revealed occult sphincter injury in 35% of primipara women; most of these lesions were not detected clinically. In another study, sphincter defects were detected in 85% of women with a third-degree perineal tear compared with 33% of subjects without a tear.89 In studies that compared electromyography (EMG; see later) mapping with anal endosonography, the concordance rate for identifying a sphincter defect was high.90,91 The technique is, however, operatordependent and requires training and experience.73 Although endosonography can distinguish internal from external sphincter injury, it has a low specificity for demonstrating the cause of fecal incontinence.2 Because anal endoso­ nography is more widely available, less expensive, and certainly less painful than EMG, which requires needle insertion, it is the preferred technique for examining the morphology of the anal sphincter muscles. Magnetic Resonance Imaging Endoanal magnetic resonance imaging (MRI) has been shown to provide superior imaging with excellent spatial resolution, particularly for defining the anatomy of the EAS.92,93 One study,94 but not another,92 has shown that MRI is less accurate than anal endosonography. A major contribution of anal MRI has been the recognition of external sphincter atrophy, which may adversely affect sphincter repair95 (see later). Atrophy also may be present without pudendal neuropathy.96 The addition of dynamic pelvic MRI using fast imaging sequences or MRI colpocystography, which involves filling the rectum with ultrasound gel as a contact agent and having the patient evacuate while lying inside the magnet, may define the anorectal anatomy more precisely.97 The use of an endoanal coil significantly enhances the resolution and allows more precise definition of the sphincter muscles. Comparative studies of costs, availability, technical factors, clinical utility, and role in treatment decision making are warranted.

Defecography

Defecography uses fluoroscopic techniques to provide morphologic information about the rectum and anal canal.98 It is used to assess the anorectal angle, measure pelvic floor descent and length of anal canal, and detect the presence of a rectocele, rectal prolapse, or mucosal intussusception. Approximately 150 mL of contrast material is placed into the rectum, and the subject is asked to squeeze or cough and

EAS

Figure 17-5.  Anal endosonograms. A, Normal healthy subject with intact, hypoechoic internal anal sphincter (IAS) and intact, thicker, and hyperechoic external anal sphincter (EAS). B, Subject with fecal incontinence secondary to an obstetric injury causing a large anterior sphincter defect that involves the IAS and EAS and spans the circumference between the 10 and 2 o’clock positions (arrows).

IAS

A

B

Chapter 17  Fecal Incontinence

Normal

100 µV O

Time

A

P

Amplitude (µV)

P

O

Time

3 msec

Patient

100 µV

3 msec

B

Figure 17-6.  Pudendal nerve terminal motor latency time in a normal subject (A) and a patient with fecal incontinence and pudendal neuropathy (B). Compared with the pudendal terminal motor nerve latency time in the normal subject, the patient’s tracing shows a delayed onset (O) and peak (P). µV, microvolts; msec, milliseconds.

expel the contrast. Although defecography can detect a number of abnormalities, these abnormalities can also be seen in otherwise asymptomatic persons,73,99 and their presence correlates poorly with impaired rectal evacuation. Agreement between observers in the measurement of the anorectal angle is also poor. Whether one should use the central axis of the rectum or the posterior wall of the rectum when measuring the angle is unclear. The functional significance of identifying morphologic defects has been questioned. Although defecography can confirm the occurrence of incontinence at rest or during coughing, it is most useful for demonstrating rectal prolapse2,100 or poor rectal evacuation (see Chapter 18). In selected patients, magnetic resonance defecography may evaluate evacuation and identify coexisting problems such as a rectocele, enterocele, cystocele, or mucosal intussusception.88

Balloon Expulsion Test

Normal subjects can expel a 50-mL water-filled balloon101 or a silicone-filled artificial stool from the rectum in less than one minute.2 Most patients with fecal incontinence have little or no difficulty with evacuation, but patients with fecal seepage58 and many older persons with fecal incontinence secondary to fecal impaction54 demonstrate impaired evacuation. In these patients, a balloon expulsion test may help identify coexisting dyssynergia or a lack of coordination between the abdominal, pelvic floor, and anal sphincter muscles during defecation. One study has shown a high frequency of dyssynergia in residents of nursing homes (see Chapter 18).102

Neurophysiologic Testing

Electrical recording of the muscle activity from the anal sphincter (EMG) is a useful technique for identifying sphincter injury as well as denervation-reinnervation potentials that can indicate neuropathy.22,73 EMG can be performed using a fine wire needle electrode or a surface electrode, such as an anal plug. Abnormal EMG activity, such as fibrillation potentials and high-frequency spontaneous discharges, provides evidence of chronic denervation, which commonly is seen in patients with fecal incontinence secondary to pudendal nerve injury or cauda equina syndrome.103 The PNTML measures the neuromuscular integrity between the terminal portion of the pudendal nerve and the anal sphincter. Injury to the pudendal nerve leads to denervation of the anal sphincter muscle and muscle

weakness. Therefore, measurement of the nerve latency time can help distinguish muscle injury from nerve injury as the cause of a weak sphincter muscle. A disposable electrode (St. Mark’s electrode; Dantec, Denmark) is used to measure the latency time.104 A prolonged nerve latency time suggests pudendal neuropathy (Fig. 17-6). Women who have delivered vaginally with a prolonged second stage of labor or have had forceps-assisted delivery have been found to have a prolonged PNTML compared with women who delivered by cesarean section or spontaneously.105,106 An American Gastroenterological Association technical review did not recommend PNTML,73 although an expert review has noted that patients with pudendal neuropathy generally have a poor surgical outcome.107 A normal PNTML does not exclude pudendal neuropathy, because the presence of a few intact nerve fibers can lead to a normal result, whereas an abnormal latency time is significant. PNTML may be useful in the assessment of patients prior to anal sphincter repair and is particularly helpful in predicting the outcome of surgery. The integrity of the peripheral component of efferent motor pathways that control anorectal function can also be assessed by recording the motor evoked potentials (MEPs) of the rectum and anal sphincter in response to magnetic stimulation of the lumbosacral nerve roots (translumbar magnetic stimulation [TLMS] and transsacral magnetic stimulation [TSMS]).22,108,109 The technique is based on Faraday’s principle, which states that in the presence of a changing electrical field, a magnetic field is generated. Consequently, when a current is discharged rapidly through a conducting coil, a magnetic flux is produced around the coil. The magnetic flux causes stimulation of neural tissue. Magnetic stimulation of the lumbosacral roots (TLMS and TSMS) may allow more precise localization of the motor pathways between the brain and the anal sphincter as well as subcomponent analysis of the efferent nervous system between the brain and sphincter. Electrical or magnetic stimulation of the lumbosacral nerve roots facilitates measurement of the conduction time within the cauda equina and can diagnose sacral motor radiculopathy as a possible cause of fecal incontinence.110,111 One study has shown that translumbar MEP and transsacral MEP of the rectum and anus provides delineation of peripheral neuromuscular injury in subjects with fecal incontinence108 (Fig. 17-7) and can reveal hitherto undetected changes in patients with back injury.

249

Section III  Symptoms, Signs, and Biopsychosocial Issues LEFT

RIGHT Spinal cord injury

Spinal cord injury Amplitude

250

Healthy

Healthy

Time Figure 17-7.  Anal motor evoked potential (MEP) responses following translumbar magnetic stimulation in a subject with fecal incontinence and a history of spinal cord injury (upper tracings) and in a healthy normal subject (lower tracings). The MEP responses on the left and right sides are shown separately. When compared with the healthy subject, the incontinent subject with spinal cord injury shows an MEP response with a prolonged onset time on the left and right sides and a smaller amplitude of the MEP response on the right side. These features indicate that bilateral lumbospinal neuropathy is the cause of fecal incontinence.

Saline Infusion Test

The saline infusion test assesses the overall capacity of the defecation unit to maintain continence during conditions that simulate diarrhea.72,80,82 With the patient lying on the bed, a 2-mm plastic tube is introduced approximately 10 cm into the rectum and taped in position. Next, the patient is transferred to a commode. The tube is connected to an infusion pump and 800 mL of warm saline (37°C) is infused into the rectum at a rate of 60 mL/min. The patient is instructed to hold the liquid for as long as possible. The volume of saline infused at the onset of first leak (defined as a leak of at least 15 mL) and the total volume retained at the end of infusion are recorded. Most normal subjects should retain most of the infused volume without leakage, whereas patients with fecal incontinence or patients with impaired rectal compliance, such as those with ulcerative colitis,112 leak at much lower volumes. The test is also useful for assessing objective improvement of fecal incontinence after biofeedback therapy.85

Clinical Utility of Tests for Fecal Incontinence

In one prospective study, history taking alone could detect an underlying cause in only 9 of 80 patients (11%) with fecal incontinence, whereas physiologic tests revealed an abnormality in 44 patients (55%).113 In a large retrospective study of 302 patients with fecal incontinence, an underlying pathophysiologic abnormality was identified, but only after manometry, EMG, and rectal sensory testing were performed.114 Most patients had more than one pathophysiologic abnormality. In another large study of 350 patients, incontinent patients had lower resting and squeeze sphincter pressures, a smaller rectal capacity, and earlier leakage following saline infusion in the rectum.82 Nevertheless, results of a single test or a combination of three different tests (anal manometry, rectal capacity, saline continence test) provided a low discriminatory value between continent and incontinent patients. This finding emphasizes the wide range of normal values and the ability of the body to compensate for the loss of any one mechanism involved in fecal incontinence. In a prospective study, anorectal manometry with sensory testing not only confirmed a clinical impression, but also provided new information that was not detected clinically.72 Furthermore, the diagnostic information obtained from these studies can influence both the management and outcome of patients with incontinence. A single abnorma­ lity was found in 20% of patients, whereas more than one abnormality was found in 80% of patients. In another study,

abnormal sphincter pressure was found in 40 patients (71%), whereas altered rectal sensation or poor rectal compliance was present in 42 patients (75%).113 These findings were confirmed by another study, which showed that phy­ siologic tests provided a definitive diagnosis in 66% of patients with fecal incontinence.114 Still, on the basis of the test results alone, it is not possible to predict whether an individual patient is continent or incontinent. Consequently, an abnormal test result must be interpreted in the context of the patient’s symptoms and the results of other complementary tests. Tests of anorectal function provide objective data and define the underlying pathophysiology. Table 17-3 summarizes the key tests, information gained from them, and evidence to support their clinical use.

TREATMENT The goal of treatment for patients with fecal incontinence is to restore continence and improve their quality of life. Strategies that include supportive and specific measures may be used. An algorithmic approach to the evaluation and management of patients with fecal incontinence is presented in Figure 17-8.

SUPPORTIVE MEASURES

Supportive measures such as avoiding offending foods, ritualizing bowel habit, improving skin hygiene, and instituting lifestyle changes may serve as useful adjuncts to the management of fecal incontinence. Obtaining a comprehensive history (see Table 17-1), performing a detailed physical examination, and requesting that the patient keep a prospective stool diary2,73 can provide important clues regarding the severity and type of incontinence as well as predisposing conditions, such as fecal impaction, dementia, neurologic disease, inflammatory bowel disease, or dietary factors (e.g., carbohydrate intolerance). If present, these conditions should be treated or corrected. In the management of older or institutionalized patients with fecal incontinence, the availability of personnel experienced in the treatment of fecal incontinence, timely recognition of soiling, and immediate cleansing of the perianal skin are of paramount importance.60 Hygienic measures such as changing undergarments, cleaning the perianal skin immediately following a soiling episode, use of moist tissue paper (baby wipes) rather than dry toilet paper, and use of barrier creams such as zinc oxide and calamine lotion

Evaluates presence of fecal retention; inexpensive and widely available Simple, inexpensive, bedside assessment of ability to expel a simulated stool; identifies dyssynergic defecation

Colonic transit study with radiopaque markers Balloon expulsion test (BET)

Directly visualizes the colon to exclude mucosal lesions (e.g., solitary rectal ulcer syndrome, inflammation, malignancy)

Identifies megacolon, megarectum, stenosis, diverticulosis, extrinsic compression, and intraluminal masses

Simultaneously evaluates global pelvic floor anatomy and dynamic motion; reveals sphincter morphology and pathology outside the anorectum Identifies excessive amount of stool in the colon; simple, inexpensive, widely available

Invasive, risks related to procedure (perforation, bleeding) and sedation

Lack of standardization of interpretation, lack of controlled studies Lack of standardization, embarrassment, radiation exposure, lack of controlled studies

Interobserver bias; scars difficult to identify Radiation exposure, embarrassment, availability, interobserver bias, inconsistent methodology Expensive, lack of standardization, availability

Minimally invasive, low sensitivity, interobserver differences Lack of standardization, training, controlled studies, and availability Inconsistent methodology, validity has been questioned Lack of standardization

Invasive, painful; not widely available Inaccurate, frequent artifacts

Lack of standardization

WEAKNESSES

*Evidence-based summary. EAS, external anal sphincter; EMG, electromyography; IAS, internal anal sphincter; MRI, magnetic resonance imaging.

Endoscopy Flexible sigmoidoscopy and colonoscopy

Barium enema

Plain abdominal film

MRI

Defecography

Imaging Anorectal ultrasonography

Visualizes IAS and EAS defects, thickness, and atrophy and puborectalis muscles Detects prolapse, intussusception, obtuse anorectal angle, and pelvic floor weakness, as well as rectoceles and megarectum

Quantifies nerve conduction time of entire spinoanal and spinorectal pathways; minimally invasive

Translumbar and transsacral motor evoked potentials

Surface EMG

Needle EMG

Pudendal nerve terminal motor latency (PNTML)

clinical use and STRENGTHS

Clinical Use

Quantifies EAS and IAS pressures; identifies rectal hyposensitivity, rectal hypersensitivity, impaired rectal compliance, dyssynergic defecation Quantifies spike potentials and re-innervation pattern indicating neuropathy or myopathy Displays EMG activity; can provide information on normal or weak muscle tone Measures latency of terminal portion of pudendal nerve, simple to perform

Physiologic Anorectal manometry

TEST

Table 17-3  Diagnostic Tests for Fecal Incontinence*

Poor

Poor

Poor

Fair

Fair

Good

Good

Good

Fair

Fair

Fair

Fair

Good

quality of EVIDENCE

Indicated in patients with unexplained diarrhea and seepage and in subjects older than age 50 yr

Not recommended for routine evaluation but useful in older adults and children with incontinence and fecal impaction Not recommended as part of routine evaluation

Used as an adjunct to other tests

Useful and complementary with other tests

Most widely available

Useful for identifying patients with fecal seepage and older persons with impaction Normal BET does not exclude dyssynergia; should be interpreted in the context of other anorectal test results

Promising noninvasive test; more objective and higher yield than PNTML

Conflicting data; correlation with other tests and surgical outcome unclear

Used largely for neuromuscular training

Useful but used largely in research laboratories

Useful for detecting anal sphincter weakness, altered rectal sensation and accommodation, dyssynergia

COMMENTS

Chapter 17  Fecal Incontinence 251

252

Section III  Symptoms, Signs, and Biopsychosocial Issues History, physical examination (including digital rectal examination)

Diarrhea + incontinence

Obstetric, surgical, neurologic injury

Local anorectal problem Appropriate treatment (see Chapter 125)

Flexible sigmoidoscopy, colonoscopy, and/or barium enema + routine blood tests

Suspected rectal prolapse Clinically confirmed

Not confirmed

Positive

Defecography + MRI

All test results negative Trial of loperamide, diphenoxylate and atropine, or other antidiarrheal agent Improved

Not improved

Anorectal manometry + Anal endosonography + − Balloon expulsion test + − Neurophysiology tests (EMG/PNTML/MEP testing)

Weak sphincter or sphincter defect + No or mild neuropathy Neuromuscular training Figure 17-8.  Algorithm for the evaluation and management of patients with fecal incontinence. EMG, electromyography; MEP, motor evoked potential; MRI, magnetic resonance imaging; PNTML, pudendal nerve terminal motor latency.

Weak sphincter or sphincter defect + neuropathy Neuromuscular training or colostomy

Surgery

Normal ? Factitious incontinence

Impaired sensation

Dyssynergic defecation + impaired evacuation

Neuromuscular training

Neuromuscular training

If ineffective Sphincteroplasty or sphincter repair Sacral nerve stimulation Artificial bowel sphincter Colostomy

(Calmoseptine; Calmoseptine, Huntington Beach, Calif) may help prevent skin excoriation. Perianal fungal infections should be treated with topical antifungal agents. More importantly, scheduled toileting with a commode at the bedside or bedpan and supportive measures to improve the general well-being and nutritional status of the patient may prove effective. Stool deodorants (e.g., Bedside Care Perineal Wash, Minneapolis; Derifil, Integra, Plainsboro, NJ; Devrom, Parthenon, Salt Lake City) can help disguise the smell of feces. In an institutionalized patient, ritualizing the bowel habit and instituting cognitive training may prove beneficial. Using these measures, short-term (3- to 6-month) success rates of up to 60% have been reported in case series.115 Patients in whom these measures fail have been shown to have a higher mortality rate than those without incontinence and than those with incontinence who respond to these measures.116 Other supportive measures include dietary modifications, such as reducing caffeine or fiber intake. Caffeinecontaining coffee enhances the gastrocolic (or gastroileal) reflex, increases colonic motility,117 and induces fluid secretion in the small intestine.118 Therefore, reducing caffeine consumption, particularly after meals, may help lessen postprandial urgency and diarrhea. Brisk physical activity, particularly after meals or immediately after waking, may precipitate fecal incontinence because these physiologic events are associated with increased colonic motility.119

Decreased rectal reservoir Rectal augmentation surgery

Acute exercise can enhance colonic motor activity and transit.120 A food and symptom diary may identify dietary factors that cause diarrheal stools and incontinence; frequent culprits are lactose and fructose, which may be malabsorbed.121 Eliminating food items containing these constituents may prove beneficial.2 Fiber supplements such as psyllium are often advocated in an attempt to increase stool bulk and reduce watery stools. In a single casecontrolled study, psyllium led to a modest improvement,122 but fiber supplements can potentially worsen diarrhea by increasing colonic fermentation of unabsorbable fiber.

SPECIFIC THERAPIES Pharmacologic Therapy

The antidiarrheal agents loperamide hydrochloride (Imodium) and diphenoxylate and atropine sulfate (Lomotil ) remain the mainstays of drug treatment for fecal incontinence, although other drug treatments have been proposed.2,123 In placebo-controlled studies, loperamide, 4 mg three times daily, has been shown to reduce the frequency of incontinence, improve stool urgency, and increase colonic transit time,84 as well as increase anal resting sphincter pressure124 and reduce stool weight. Clinical improvement was also reported with diphenoxylate and atropine,125 but objective testing showed no improvement in the ability of the patient to retain saline or spheres in the

Chapter 17  Fecal Incontinence rectum. Although most patients benefit from antidiarrheal agents temporarily, many report cramping, lower abdominal pain, or difficulty with evacuation after a few days. Therefore, careful titration of the dose is required to produce the desired result. Idiopathic bile salt malabsorption may be an important underlying cause of diarrhea and fecal incontinence (see Chapter 15).126 Patients with this problem may benefit from titrated doses of ion exchange resins such as cholestyramine (Questran) or colestipol (Colestid). Alosetron (Lotronex), a 5-hydroxytryptamine3 receptor antagonist used for the treatment of irritable bowel syndrome and diarrhea, may serve as an adjunct to the therapy of fecal incontinence, but use of the drug is restricted because of side effects (see Chapter 118).127 Postmenopausal women with fecal incontinence may benefit from estrogen replacement therapy.128 An openlabeled study has shown that oral amitriptyline, 20 mg, is useful in the treatment of patients with urinary or fecal incontinence without evidence of a structural defect or neuropathy.129 Suppositories or enemas may also have a role in the treatment of incontinent patients with incomplete rectal evacuation or in those with postdefecation seepage. In some patients, constipating medications alternating with periodic enemas may provide more controlled evacuation of bowel contents, but these interventions have not been tested prospectively.

Neuromuscular Training

Pressure (mm Hg)

Neuromuscular training, usually referred to as biofeedback therapy, improves symptoms of fecal incontinence, restores quality of life, and improves objective parameters of ano­ rectal function. Biofeedback training is useful in patients with a weak sphincter or impaired rectal sensation. The method is based on operant conditioning techniques whereby an individual acquires a new behavior through a learning process of repeated reinforcement and instant feedback.2,130 The goals of neuromuscular training in a patient with fecal incontinence are as follows: (1) to improve the strength of the anal sphincter muscles; (2) to improve the coordination between the abdominal, gluteal, and anal 100 90 80 70 60 50 40 30 20 10 0 100 90 80 70 60 50 40 30 20 10 0

Squeeze

Rest

100 90 80 70 60 50 40 30 20 10 0 100 90 80 70 60 50 40 30 20 10 0

14:09:00

A

Time

sphincter muscles during voluntary squeeze and following rectal perception; and (3) to enhance anorectal sensory perception. Because each goal requires a specific method of training, the treatment protocol should be customized for each patient on the basis of the underlying pathophysiologic mechanism(s). Neuromuscular training often is performed using visual, auditory, or verbal feedback techniques, and the feedback is provided by a manometry or EMG probe placed in the anorectum.2,130 When a patient is asked to squeeze, the anal sphincter contraction is displayed as an increase in anal pressure or EMG activity. This visual cue provides instant feedback to the patient. The aim of rectoanal coordination training is to achieve a maximum voluntary squeeze in less than two seconds after a balloon is inflated in the rectum. In reality, this maneuver mimics the arrival of stool in the rectum and prepares the patient to react appropriately by contracting the right group of muscles.2,130 Patients are taught how to squeeze their anal muscles selectively without increasing intra-abdominal pressure or inappropriately contracting their gluteal or thigh muscles. Also, this maneuver identifies sensory delay and trains the individual to use visual clues to improve sensorimotor coordination.131,132 Sensory training of the rectum educates the patient to perceive a lower volume of balloon distention but with the same intensity as they had felt earlier with a higher volume. This goal is achieved by repeatedly inflating and deflating a balloon in the rectum. These neuromuscular training techniques must be used together with pelvic muscle strengthening (modified Kegel exercises) and other supportive measures to achieve sustained improvement of bowel function. A component analysis—muscle training, sensory training, or both—is most effective; whether Kegel exercises alone are more effective than the use of multiple approaches has not been determined. Predicting how many neuromuscular treatment sessions will be required is often difficult. Most patients seem to require between four and six training sessions (Fig. 17-9).2,85,130 Studies that used a fixed number of treatment

Squeeze Rest

9:20:00

B

Time

Figure 17-9.  Anal manometric pressure tracings in a patient with fecal incontinence before (A) and after (B) neuromuscular training (biofeedback) while squeezing and at rest. Before neuromuscular training, the patient has a weak and poorly sustained squeeze and makes multiple ineffective attempts to squeeze. After six sessions of training, the ability to generate and sustain the squeeze has improved significantly.

253

254

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 17-4  Outcome of Neuromuscular Training (Biofeedback Therapy) and/or Exercises for Fecal Incontinence in Adults* REFERENCE

SUBJECTS (F/M)

TREATMENT

CONTROL

OUTCOME

Manometric BFB + rectal sensory training + coordination training (weekly, 4 wk) BFB + electrical stimulation (augmented) (weekly, 12 wk)

Sham training (crossover design)

Treatment improved symptoms

Vaginal manometric biofeedback

Greater symptom improvement in treated group than control group (P < 0.001) Treatment improved symptoms more than PFMT alone (77% vs. 41%; P = 0.001) NSD between groups

131

17/8

137

40/0

138

83/25

BFB + PFMT + sensory training (biweekly, 12 wk)

PFMT

142

60/0

BFB

143

49/0

BFB (weekly, 12 wk) + electrical stimulation BFB + home exercises

144

159/12

Four groups:

NA

1. Education + advice 2. As per group 1 + PFMT 3. As per group 2 + manometric

Electrical stimulation

Both groups improved; NSD between groups in symptoms and QOL ~54% improved in all groups NSD between groups in symptoms and QOL

BFB

4. As per group 3 + home BFB 145

107/13

(biweekly, 6 sessions, 3 mo) Three groups: 1. PFMT 2. PFMT + anal ultrasound BFB 3. PFMT + manometric BFB (monthly, 5 sessions)

NA

NSD between groups in symptoms, QOL, and manometry changes

*Selected randomized controlled trials. BFB, biofeedback (using electromyography probe unless otherwise specified); F, females; M, males; NA, not applicable; NSD, no significant difference; PFMT, pelvic floor muscle training; QOL, quality of life. Adapted from Norton C. Fecal incontinence and biofeedback therapy. Gastroenterol Clin North Am 2008; 37:587-604.

sessions, often less than three, showed a less favorable improvement response than those that titrated the number of sessions on the basis of the patient’s performance.133,134 In one study, periodic reinforcement with neuromuscular training at six weeks, three months, and six months was thought to confer additional benefit85 and long-term improvement.135 In the literature on fecal incontinence,136-146 the terms improvement, success, or cure have been used interchangeably, and the definition of each term has been inconsistent. In uncontrolled studies, subjective improvement has been reported in 40% to 85% of patients.2,133 Table 17-4 summarizes selected randomized controlled trials of neuromuscular training in patients with fecal incontinence.130,131,137,138,142-145 A Cochrane review of 11 randomized, controlled trials has concluded that no method of training is better than any other method.147 Whether biofeedback is superior to conservative management is also unclear. In the most recent randomized controlled trial,138 108 patients were randomized to receive either six sessions of EMG biofeedback (n = 44) or Kegel exercises (n = 64) plus supportive therapy. After treatment, 77% of patients who received biofeedback reported adequate relief of symptoms compared with 41% of those who did Kegel exercises (P < 0.001). The number of episodes of incontinence was not different between groups in an intention-to-treat analysis, but a trend toward improvement (P = 0.042) was observed in a per-protocol analysis.138 This study suggests that biofeedback is superior to Kegel exercises. The technique of neuromuscular training has not been standardized, and the use of this treatment is largely restricted to specialized centers. The manometric parameters obtained at baseline do not appear to predict the clinical

response to biofeedback treatment.148 Similarly, the patient’s age, presence of sphincter defects, or presence of neuropathy do not predict outcome.149 Therefore, criteria used for selection, motivation of the individual patient, enthusiasm of the therapist, and severity of incontinence each may affect the outcome.2,130,133,134,144 Despite the lack of a uniform approach and the inconsistencies in the reported outcomes of randomized controlled trials, neuromuscular training seems to confer benefit (see Table 17-4). Therefore, neuromuscular training should be offered to all patients with fecal incontinence who have failed supportive measures and especially to older patients, patients with comorbid illnesses, and those for whom reconstructive surgery is being considered. Severe fecal incontinence, pudendal neuropathy, and an underlying neurologic disorder are associated with a poor response to biofeedback therapy.150-152 One study has suggested that neuromuscular training may be most beneficial in patients with urge incontinence.153 Biofeedback also seems to be useful for patients who have undergone anal sphincteroplasty,154 postanal repair (see later),155 or low-anterior resection156 and children who have undergone correction of a congenital anorectal anomaly.157

Plugs, Sphincter Bulkers, and Electrical Stimulation

Disposable anal plugs have been used to help occlude the anal canal temporarily.158 Unfortunately, many patients are unable to tolerate prolonged insertion of the device.159,160 A plug may be useful for patients with impaired anal canal sensation, those with neurologic disease,161 and those who are institutionalized or immobilized. In some patients with fecal seepage, insertion of an anal plug made of cotton wool may prove beneficial162; the recommended wear

Chapter 17  Fecal Incontinence Table 17-5  Success Rates of Surgical Interventions for Fecal Incontinence PROCEDURE Current Anal sphincter repair Sacral nerve stimulation Dynamic gracilis neosphincter Artificial bowel sphincter Fecal diversion Evolving Injection of biomaterials Radiofrequency therapy (Secca procedure) Rectal augmentation

OUTCOME MEASURES

SUCCESS RATE (%)

quality OF EVIDENCE

Clinical, physiologic Complete continence Improvement in continence by ≥50% Restoration of continence Full continence NA

50-66* 40-75 75-100 42-85 50-100§ No data

Good† Very good

Cessation of leakage or improvement in continence Improvement in continence by ≥50% Avoidance of stoma

66 (short term)¶

Poor

84 64

Poor Poor

Poor‡ Good NA

*5-year success rates fall to 50%. † Derived from a Cochrane review, but in some cases data were extrapolated from only one study. ‡ Based on a systematic review of case series; no comparative studies available. § Explantation rates in case series of approximately 50%. ¶ No difference in continence scores compared with preoperative scores on long-term follow-up. NA, not available. Adapted from Gladman MA. Surgical treatment of patients with constipation and fecal incontinence. Gastroenterol Clin North Am 2008; 37:605-25, with permission.

time (although not formally tested) is up to 12 hours.130 Diapers generally are thought to be unsatisfactory for providing security or comfort, protecting the skin, or disguising odor. Many people with fecal incontinence choose not to wear a pad. Small anal dressings may be useful for people with minor soiling contained between the buttocks but can become costly if several dressings are needed each day. Bulking the anal sphincter to augment its surface area and thereby provide a better seal for the anal canal has been attempted with a variety of agents, including autologous fat,163 glutaraldehyde-treated collagen,164 and synthetic macromolecules.165 These materials usually are injected submucosally at the site where the sphincter is deficient or circumferentially if the whole muscle is degenerated or fragmented. Studies have shown definite improvement in the short term in patients with passive fecal incontinence. The experience with these techniques, however, is limited, and controlled and long-term outcome studies have not been done. Newer and better designed anal plugs are currently being tested. Electrical stimulation of striated muscle at a frequency sufficient to produce a tonic involuntary contraction (usually 30 to 50 Hz) can increase muscle strength, conduction rate of the pudendal nerve, and size of motor units, encourage neuronal sprouting, and promote local blood flow.166,167 Stimulation at lower frequencies (typically 5 to 10 Hz) can modulate autonomic function, including sensation and overactivity. Studies of electrical stimulation for fecal incontinence, however, generally have been small and uncontrolled and have been confounded by the effects of exercise, biofeedback, or other interventions. A Cochrane review of four randomized controlled trials with 260 participants concluded that electrical stimulation may have some effect.168 One study has found that anal electrical stimulation with anal biofeedback produces short-term benefits greater than those with biofeedback alone,137 whereas another study found no additional benefit to electrical stimulation over exercises and biofeedback alone.142 Also, patients have been shown to improve equally with stimulation at 1 and 35 Hz.130 Two randomized controlled trials have reported that biofeedback and electrical stimulation are equally effective.169,170 Therefore, whether electrical stimulation by itself is helpful remains unclear.

Surgical Therapy Surgery should be considered for selected patients who have failed conservative measures or biofeedback therapy. The choice of surgical procedure must be tailored to the need of the individual patient and can be described under four broad clinical categories: (1) simple structural defects of the anal sphincters; (2) weak but intact anal sphincters; (3) complex disruption of the anal sphincter complex; and (4) extrasphincteric abnormalities. Table 17-5 summarizes success rates for these surgical procedures.171 In most subjects, particularly those with obstetric trauma, overlapping sphincter repair is often sufficient. The torn ends of the sphincter muscle are plicated together and to the puborectalis muscle. Overlapping sphincter repair, as described by Parks and McPartlin,172 involves a curved incision anterior to the anal canal with mobilization of the external sphincter, which is divided at the site of the scar; the scar tissue is preserved to anchor the sutures, and overlap repair is carried out using two rows of sutures. If an internal anal sphincter defect is identified, a separate imbrication (overlapping repair) of the internal anal sphincter may be undertaken. Symptom improvement with a frequency in the range of 70% to 80% has been reported, although one study reported an improvement rate of only approximately 50%.172-176 Furthermore, some patients may experience problems with evacuation after surgery. In patients with incontinence caused by a weak but intact anal sphincter, postanal repair has been tried.177 The anorectal angle is made more acute via an intersphincteric approach, thereby improving continence. The long-term success of this approach ranges from 20% to 58%.178 In patients with severe structural damage of the anal sphincter and significant incontinence, construction of a neosphincter has been attempted using two approaches: (1) use of autologous skeletal muscle, often the gracilis and rarely the gluteus107,179; and (2) use of an artificial bowel sphincter (ABS).180 The technique of stimulated gracilis muscle transposition (dynamic graciloplasty) has been tested in many centers.181,182 This technique uses the principle that a fast-twitch, fatigable skeletal muscle, when stimulated over a long period of time, can be transformed into a slow-twitch fatigable muscle that can provide a sustained, sphincter-like muscle response. Such continuous

255

256

Section III  Symptoms, Signs, and Biopsychosocial Issues stimulation is maintained by an implanted pacemaker. When the subject has to defecate or expel gas, an external magnetic device is used to switch off the pacemaker tem­ porarily. Rates of clinical improvement with this approach have ranged from 38% to 90% (mean 67%).171 The other approach to neosphincter construction has been to implant an ABS. The ABS consists of an implanted inflatable cuffed device that is filled with fluid from an implanted balloon reservoir, which is controlled by a subcutaneous pump. The cuff is deflated to allow defecation. In one series of 24 carefully selected patients, almost 75% reported satisfactory results, although some had the device explanted.183 Both approaches (dynamic graciloplasty and ABS) require major surgery and are associated with revision rates that approach 50%. At medium-term follow up, 50% to 70% of patients have a functioning new sphincter. Several groups have reported their experiences with the ABS in small numbers of patients with overall improvement in continence in approximately 50% to 75% of patients.184,185 A randomized controlled trial has demonstrated that ABS is better than conservative treatment in improving continence.186 Longterm outcome studies, with median follow-up periods of approximately seven years, however, have documented success rates of less than 50%, explantation rates as high as 49%, and infection rates of up to 33%.131,187 Additionally, evacuation problems occur in 50% of patients. Rectal augmentation is a novel approach to correct the physiologic abnormalities in a subgroup of patients with intractable fecal incontinence secondary to reservoir or rectal sensorimotor dysfunction.188 Candidates have low rectal compliance and heightened rectal sensation (rectal hypersensitivity). The procedure involves the creation of a side-to-side ileorectal pouch, or ileorectoplasty, that involves incorporating a 10-cm patch of ileum on its vascular pedicle into the anterior rectal wall to increase rectal capacity and compliance.189 In 11 subjects, at medium-term follow up (4.5 years), rectal capacity was increased, with an associated improvement in bowel symptoms (increased ability to defer defecation and reduced frequency of episodes of incontinence) and in patients’ quality of life.190

Other Procedures

Radiofrequency energy can be delivered deep to the mucosa of the anal canal via multiple needle electrodes with use of a specially designed probe (Secca System; Rayfield Technology, Houston) inserted into the anal canals of patients with fecal incontinence.191 The proposed mechanism of action is heat-induced tissue contraction and remodeling of the anal canal and distal rectum. In one study, symptomatic improvement was sustained at two and five years after treatment.192 A multicenter trial has confirmed the improvements in continence and quality of life, at least in the short term (at six months). Complications include ulceration of the mucosa and delayed bleeding.193 Interestingly, no changes were seen in the results of anorectal manometry, PNTML measurement, or anal endosonography. Results of a randomized controlled trial of this method completed in the United States are pending. The Malone, or antegrade continent, enema procedure194 consists of fashioning a cecostomy button or appendicostomy195 to allow periodic antegrade washout of the colon. This approach may be suitable for children and for patients with neurologic disorders.195-197 If none of these techniques is suitable or all have failed, a colostomy remains a safe, although aesthetically less preferable, option for many patients.107,198-200 It is particularly suitable for patients with spinal cord injury, immobilized patients, and those with severe skin problems or other com-

plications. A colostomy should not be regarded as a failure of medical or surgical treatment.171 For many patients with fecal incontinence, the restoration of a normal quality of life and amelioration of symptoms can be rewarding. The use of a laparoscopic-assisted approach, a trephine colostomy, may help to fashion a stoma with minimal morbidity for the patient.201 In one study, the total direct costs were estimated to be $31,733 for a dynamic graciloplasty, $71,576 for a colostomy including stoma care, and $12,180 for conventional treatment of fecal incontinence.202 No controlled studies have compared surgical management with pharmacologic therapy or biofeedback therapy. Similarly, no controlled studies of the different surgical approaches have been published. Because the outcome of most procedures ranges from significant improvement initially to a less satisfactory result in the long term, no single procedure is universally accepted. In the future, a better understanding of the underlying pathophysiology and development of safer and better techniques, followed by prospective controlled trials, may allow selection of younger patients with well-defined sphincter defects for appropriate surgery.

Sacral Nerve Stimulation

Sacral nerve stimulation (SNS) has emerged as a useful treatment option in selected patients, although how SNS improves fecal incontinence remains unclear.203 The benefit may relate to direct effects peripherally on colorectal sensory or motor function or to central effects at the level of the spinal cord or brain.204 Earlier studies were performed in subjects with a morphologically intact anal sphincter, but subsequent reports have described the treatment in patients with EAS defects,205 IAS defects,206 and cauda equina syndrome207 or spinal injuries.208 The technique of SNS consists of two phases. The first phase is a temporary trial phase of two weeks during which electrodes are implanted in the second or third sacral nerve roots and the nerves are stimulated with a neurostimulator device. If the patient reports satisfactory improvement of symptoms, a permanent neurostimulator device is placed in the second phase (Fig. 17-10). Initial reports of SNS have described marked improvements in clinical symptoms and quality of life and marginal effects on physiologic parameters.176,209 The results of multicenter studies of SNS have reported marked and sustained improvement in fecal incontinence and quality of life.210-212 A randomized controlled trial has found SNS to be superior to supportive therapy (pelvic floor exercises, bulking agents, and dietary manipulation),213 but long-term outcomes are not yet available. A morphologically intact anal sphincter may not be a pre­ requisite for success with SNS, and patients with EAS defects of less than 33% can be treated effectively with this method.214 A systematic review of the published outcomes of trials of SNS has revealed that 40% to 75% of patients achieve complete continence, and 75% to 100% experience improvement, with a low (10%) frequency of adverse events.215 An evidence-based summary of current therapies for fecal incontinence is shown in Table 17-6.

TREATMENT OF SUBGROUPS OF PATIENTS Patients with Spinal Cord Injury

Patients with a spinal cord injury demonstrate delayed colonic motility or anorectal dysfunction that may manifest as incontinence, seepage, difficulty with defecation, or rectal hyposensitivity.216 Anal sphincter pressures and rectal compliance are low in these patients, but the correla-

Chapter 17  Fecal Incontinence soiling with stool, followed by the periodic administration of enemas or the use of laxatives or lavage solutions at convenient intervals.2 A cecostomy procedure also may be appropriate.217 In some patients, colostomy may be the best option.198

Patients with Fecal Seepage

Because patients with fecal seepage show dyssynergic defecation with impaired rectal sensation, neuromuscular conditioning with biofeedback techniques to improve dyssynergia can be useful (see Chapter 18).58,218 Therapy that consists of sensory conditioning and rectoanal coordination of the pelvic floor muscles to evacuate stools more completely has been shown to reduce the number of fecal seepage events substantially and to improve bowel function and anorectal function by objective measures.58

Older Patients

Figure 17-10.  Plain abdominal film showing a nerve stimulator device located in the right lower quadrant along with electrodes (radiopaque) permanently implanted into the sacral nerves. This patient presented with fecal incontinence and underwent a colonic transit study that revealed significant retention of radiopaque markers, which were located mostly in the distal colon, suggesting anorectal outlet dysfunction.

Table 17-6  Treatment Options for Fecal Incontinence* TREATMENT Pharmacologic treatment Loperamide Diphenoxylate and atropine Amitriptyline Cholestyramine Neuromuscular training (biofeedback) Surgical treatment Sphincteroplasty Dynamic graciloplasty Artificial bowel sphincter Colectomy Novel treatments Anal plugs Sphincter bulking agents Sacral nerve stimulation Radiofrequency therapy (Secca procedure)

quality OF EVIDENCE Fair Fair Poor Poor Good Fair Fair Fair Poor Poor Poor Good Poor

*Evidence-based summary.

tion between manometric findings and bowel dysfunction is poor. Studies of translumbar and transsacral MEPs have shown profound neuromuscular dysfunction affecting the entire spinoanal and spinorectal pathways.109 Patients with a spinal cord injury may have fecal incontinence because of a supraspinal lesion or lesion of the cauda equina.77,78 In the former group, the sacral neuronal reflex arc is intact, and the cough reflex is preserved. Therefore, reflex defecation is possible through digital stimulation or with suppositories. In patients with a low spinal cord or cauda equina lesion, digital stimulation may not be effective because the defecation reflex is often impaired. In these cases, management consists of antidiarrheal agents to prevent continuous

Fecal incontinence is a common problem in older adults and may be a marker of declining health and increased mortality in patients in nursing homes.60 In one study, fecal incontinence developed in 20% of nursing home residents during a 10-month period after admission, and long-lasting incontinence was associated with reduced survival.116 In one report, immobility, dementia, and the use of restraints that precluded a patient from reaching the toilet in time were the most important risk factors for the development of fecal incontinence.219 Usual mechanisms of incontinence include impaired anorectal sensation, weak anal sphincter, and weak pelvic floor muscles. Decreased mobility and lowered sensory perception are common causes of incon­ tinence.220 Many of these patients have fecal impaction and overflow.54,221 Fecal impaction, a leading cause of fecal incontinence in institutionalized older adults, results largely from a person’s inability to sense and respond to the presence of stool in the rectum. A retrospective screening of 245 permanently hospitalized geriatric patients222 has revealed that fecal impaction (55%) and laxatives (20%) are the most common causes of diarrhea and that immobility and fecal incontinence are strongly associated with fecal impaction and diarrhea. One study has shown that impaired anal sphincter function (a risk factor for fecal incontinence), decreased rectal sensation, and dyssynergia are seen in up to 75% of nursing home residents with fecal incontinence.36,223 Stool softeners, saline laxatives, and stimulant laxatives are frequently administered as prophylactic treatment to prevent constipation and impaction. In a study of institutionalized older patients, the use of a single osmotic agent with a rectal stimulant and weekly enemas to achieve complete rectal emptying reduced the frequency of fecal incontinence by 35% and the frequency of soiling by 42%.224 If fecal impaction is not relieved by laxatives and better toileting, a regimen of manual disimpaction, tap water enemas two or three times weekly, and rectal suppositories should be considered.225 In the presence of impaired sphincter function and decreased rectal sensation, however, liquid stools may be counterproductive. Similarly, neuromuscular training to improve dyssynergia in older adults, ritualizing the patient’s bowel habit, improving mobility, and cognitive training may be useful.60

Children

Incontinence is seen in 1% to 2% of otherwise healthy 7-year-old children.226 It is caused by functional fecal retention (previously described as encopresis), functional nonretentive fecal incontinence,227 congenital anomalies, developmental disability, or mental retardation.

257

258

Section III  Symptoms, Signs, and Biopsychosocial Issues In children with functional fecal retention, the bowel movements are irregular, often large, bulky, and painful. Consequently, when the child experiences an urge to defecate, he or she assumes an erect posture, holds the legs close together, and forcefully contracts the pelvic and gluteal muscles. Over time, this conscious suppression of defecation leads to excessive rectal accommodation, loss of rectal sensitivity, and loss of the normal urge to defecate. The retained stools become progressively more difficult to evacuate, thereby leading to a vicious cycle. The ultimate result is overflow incontinence, with seepage of mucus or liquid stool around an impacted fecal mass. This aberrant behavior may lead to the unconscious contraction of the external sphincter during defecation and cause dyssynergic defecation.221,228 By contrast, functional nonretentive fecal incontinence represents the repeated and inappropriate passage of stool at a place other than the toilet by a child older than four years with no evidence of fecal retention. According to criteria established by the Rome III consensus committee,227 children with functional nonretentive fecal incontinence often pass stools daily in the toilet but in addition have almost complete stool evacuations in their underwear more than once a week. They have no palpable abdominal or rectal fecal mass nor evidence of fecal retention on an abdominal x-ray, and colonic radiopaque marker studies are normal.229 The frequency of daytime and nighttime enuresis is higher (40% to 45%) in children with functional nonretentive fecal incontinence than in those with fecal retention. Children with functional nonretentive fecal incontinence have significantly more behavioral problems and more externalizing or internalizing of psychosocial problems than controls. The goals of treatment are to remove any fecal impaction, restore a normal bowel habit, including passage of soft stools without discomfort, and ensure self-toileting and passage of stools at appropriate places.229 Disimpaction is best accomplished with oral medication or enemas. High doses of polyethylene glycol 3350 (1 to 1.5 g/kg/day for three days) have been shown to be effective.230 Once disimpaction has been achieved, the treatment should focus on preventing a recurrence through dietary interventions, behavioral modification, and laxatives. Treatment of functional nonretentive fecal incontinence is based on education, a nonaccusatory approach, regular toilet use with rewards, and referral to a psychologist. Successful resolution of symptoms may require prolonged treatment and follow-up.231,232 Resolving parental conflicts and psycho­ social stressors and alleviating the fear of painful bowel movements may be critical to a successful outcome.218,233 The most common congenital anomalies are neural tube defects, such as meningomyelocele or spina bifida, and anal atresia (imperforate anus; see Chapter 96). Children with a

neural defect or malformation may benefit from behavioral therapy, including a stimulated defecation program (see earlier).234 Anal atresia is best treated by surgery, but about 20% may have unsatisfactory results.229 Surprisingly, children with anorectal malformations seem to cope well with their illness.235 Children with mental retardation or those with a developmental delay may be slow or never achieve full bowel control and require life-long supportive therapy.

ACKNOWLEDGMENT

I am most grateful for the excellent secretarial assistance of Ms. Kimberly Klein. This work was supported in part by Grant 1RO1 DK57100 from the National Institutes of Health.

KEY REFERENCES

Bharucha AE, Fletcher JG, Harper CM, et al. Relationship between symptoms and disordered continence mechanisms in women with idiopathic fecal incontinence. Gut 2005; 54:546-55. (Ref 63.) Bharucha AE, Zinsmeister AR, Locke GR, et al. Prevalence and burden of fecal incontinence: A population-based study in women. Gastroenterology 2005; 129:42-9. (Ref 7.) Borello-France D, Burgio KL, Richter HE, et al. Fecal and urinary incontinence in primiparous women. Obstet Gynecol 2006; 108:863-72. (Ref 30.) Diamant NE, Kamm MA, Wald A, et al. AGA technical review on anorectal testing techniques. Gastroenterology 1999; 116:735-60. (Ref 73.) Gladman MA. Surgical treatment of patients with constipation and fecal incontinence. Gastroenterol Clin North Am 2008; 37:605-25. (Ref 171.) Kamm MA. Obstetric damage and faecal incontinence. Lancet 1994; 344:730-3. (Ref 28.) Leung FW, Schnelle JF. Urinary and fecal incontinence in nursing home residents. Gastroenterol Clin North Am 2008; 37:697-707. (Ref 60.) Norton C, Cody JD, Hosker G. Biofeedback and/or sphincter exercises for the treatment of faecal incontinence in adults. Cochrane Database System Rev 2006; (3):CD002111. (Ref 147.) Rao SSC. Practice guidelines: Diagnosis and management of fecal incontinence. Am J Gastroenterol 2004; 99:1585-604. (Ref 2.) Rao SSC. Pathophysiology of adult fecal incontinence. Gastroenterology 2004; 126:S14-22. (Ref 13.) Read NW, Abouzekry L, Read MG, et al. Anorectal function in elderly patients with fecal impaction. Gastroenterology 1985; 89:959-66. (Ref 54.) Remes-Troche J, Rao SSC. Neurophysiological testing in anorectal disorders. Gastroenterol Hepatol 2008; 2:323-35. (Ref 22.) Savoye-Collet C, Koning E, Dacher J. Radiologic evaluation of pelvic floor disorders. Gastroenterol Clin North Am 2008; 37:553-67. (Ref 88.) Scott SM, Gladman MA. Manometric, sensorimotor, and neurophysiologic evaluation of anorectal function. Gastroenterol Clin North Am 2008; 37:511-38. (Ref 79.) Tjandra JJ, Chan MK, Yeh CH, et al. Sacral nerve stimulation is more effective than optimal medical therapy for severe fecal incontinence: A randomized, controlled study. Dis Colon Rectum 2008; 51:494-502. (Ref 213.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

18 Constipation Anthony J. Lembo and Sonal P. Ullman

CHAPTER OUTLINE Definition and Presenting Symptoms  259 Epidemiology  260 Prevalence  260 Incidence  260 Public Health Perspective  260 Risk Factors  260 Gender  260 Age  260 Ethnicity  262 Socioeconomic Class and Education Level  262 Diet and Physical Activity  262 Medication Use  262 Colonic Function  262 Luminal Contents  262 Absorption of Water and Sodium  263 Diameter and Length  263 Motor Function  263 Innervation and the Interstitial Cells of Cajal  263 Defecatory Function  264 Size and Consistency of Stool  264 Classification  264 Pathophysiology  264 Normal-Transit Constipation  264 Slow-Transit Constipation  265 Defecatory Disorders  265 Disorders of the Anorectum and Pelvic Floor  266 Rectocele  266 Descending Perineum Syndrome  266 Diminished Rectal Sensation  267 Rectal Prolapse and Solitary Rectal Ulcer Syndrome  267

Constipation affects a substantial portion of the Western population and is particularly prevalent in women, chil­ dren, and older adults. Many persons with constipation do not seek medical attention, but because constipation affects between 2% and 28% of the population, it results in over $6.9 billion in medical costs annually and is one of the most common reasons for an office visit to a physician. For most affected persons, constipation is intermittent and requires no or minimal intervention, such as fiber sup­ plements or other dietary modifications. For others, con­ stipation can be challenging to treat and have a negative impact on quality of life. In these cases, specific causes of constipation, such as systemic or structural diseases, must be excluded, although constipation most commonly results from disorders of function of the colon or rectum. An under­ standing of the pathophysiology of constipation is funda­ mental to effective management. Treatment of chronic constipation begins with lifestyle modifications, if appropriate, and therapy with fiber. Osmotic and stimulant laxatives, stool softeners, emollients,

Systemic Disorders  268 Hypothyroidism  268 Diabetes Mellitus  268 Hypercalcemia  268 Nervous System Disease  268 Loss of Conscious Control  268 Parkinson’s Disease  268 Multiple Sclerosis  268 Spinal Cord Lesions  268 Structural Disorders of the Colon, Rectum, Anus, and Pelvic Floor  269 Obstruction  269 Disorders of Smooth Muscle  269 Disorders of Enteric Nerves  269 Medications  270 Psychological Disorders  270 Depression  270 Eating Disorders  271 Denied Bowel Movements  271 Clinical Assessment  271 History  271 Physical Examination  271 Diagnostic Tests  272 Tests to Exclude Systemic Disease  272 Tests to Exclude Structural Disease of the Intestine  272 Physiologic Measurements  272 Treatment  274 General Measures  274 Specific Therapeutic Agents  276 Other Forms of Therapy  282

and enemas sometimes are required to treat refractory con­ stipation. Newer agents and nonpharmacologic approaches offer further options for the treatment of constipation.

DEFINITION AND PRESENTING SYMPTOMS The definition of constipation varies among people, and it is important to ask patients what they mean when they say “I am constipated.” Most persons are describing a percep­ tion of difficulty with bowel movements or a discomfort related to bowel movements. The most common terms used by young healthy adults to define constipation are straining (52%), hard stools (44%), and the inability to have a bowel movement (34%).1 The definition of constipation also varies among physi­ cians and other health care providers. The traditional medical definition of constipation, based on the 95% lower confidence limit for healthy adults in North America and

259

260

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 18-1  Rome III Criteria for Functional Constipation Two or more of the following six must be present*: Straining during at least 25% of defecations Lumpy or hard stools in at least 25% of defecations Sensation of incomplete evacuation for at least 25% of defecations Sensation of anorectal obstruction/blockage for at least 25% of defecations Manual maneuvers to facilitate at least 25% of defecations (e.g., digital evacuation, support of the pelvic floor) Fewer than three defecations/wk *Criteria fulfilled for the previous three months with symptom onset at least six months prior to diagnosis. In addition, loose stools should rarely be present without the use of laxatives, abdominal pain is not required, and there should be insufficient criteria for irritable bowel syndrome. These criteria may not apply when the patient is taking laxatives.

the United Kingdom,2 has been three or fewer bowel move­ ments/week. Reports of stool frequency, however, are often inaccurate and do not correlate with complaints of constipa­ tion.3 In an attempt to standardize the definition of constipa­ tion, a consensus definition was initially developed by international experts in 1992 (Rome I criteria)4 and was revised in 1999 and in 2006 (Rome II and III criteria, respec­ tively; Table 18-1).5,6 The Rome criteria incorporate the multiple symptoms of constipation, of which stool frequency is only one, and require that a minimum of two symptoms be present at least 25% of the time. Unlike the Rome I criteria, the Rome II criteria include symptoms suggestive of outlet obstruction (e.g., a sensation of anorectal blockage or obstruction and use of maneuvers to facilitate defecation). The Rome III criteria allow patients to have occasional loose stools and require that symptoms be present during the previous three months, with an onset at least six months earlier. When abdominal pain or discomfort is the predominant symptom, irritable bowel syndrome (IBS), rather than constipation, should be considered to be the diagnosis (see Chapter 118). Intermittently loose stools unrelated to laxative use also suggest a diagnosis of IBS. Although distinguishing IBS from constipation alone is important, the symptoms and pathophysiology of these entities overlap substantially.

EPIDEMIOLOGY PREVALENCE

The prevalence of constipation ranges from 2% to 28% of the population in Western countries (Table 18-2)7-17 and varies depending on the demographics of the population, definition of constipation (e.g., self-reported symptoms, fewer than three bowel movements/week, or the Rome cri­ teria), and method of questioning (e.g., postal questionnaire, interview). Some studies have attempted to identify subcat­ egories of constipation based on the symptom pattern. In general, the prevalence is highest when constipation is self-reported9 and lowest when the Rome II criteria for con­ stipation are applied. When the Rome II criteria are used to diagnose constipation, the effects of gender, race, socioeco­ nomic status, and level of education on the prevalence of constipation are reduced.10

INCIDENCE

Little is known about the incidence of constipation in the general population. Talley and colleagues studied 690 non­ elderly residents of Olmsted County, Minnesota, at baseline

and after 12 to 20 months.18 Constipation, defined as fre­ quent straining at stool and passing hard stool, a stool fre­ quency of fewer than three stools/week, or both, was present in 17% of respondents on the first survey and 15% on the second survey. The rate of new constipation in this study was 50/1000 person-years, whereas the disappearance rate was 31/1000 person-years. Robson and colleagues found that 12.5% of older persons (mean age, 83 years) entering a nursing home had constipation and that constipation devel­ oped in 7% over three months of follow-up.19

PUBLIC HEALTH PERSPECTIVE

Constipation results in more than 2.5 million physician visits, 92,000 hospitalizations, and several hundred million dollars of laxative sales/year in the United States.20 Eightyfive percent of physician visits for constipation lead to a prescription for laxatives or cathartics.21 The cost of testing alone in patients with constipation has been estimated to be $6.9 billion annually.22 Among patients with constipa­ tion seen in a tertiary referral center, the average cost of a medical evaluation was $2,252, with the greatest cost attrib­ uted to colonoscopy.23 In an analysis of physician visits for constipation in the United States between 1958 and 1986, 31% of patients who required medical attention were seen by general and family practitioners, followed by internists (20%), pediatricians (15%), surgeons (9%), and obstetricians-gynecologists (9%). Only 4% of patients were seen by gastroenterologists, sug­ gesting that few such patients were deemed to need advice from a specialist.20,21 In a National Canadian Survey, 34% of persons who reported constipation had seen a physician for their symptoms.9

RISK FACTORS Risk factors for constipation in the United States include female gender, advanced age, nonwhite ethnicity, low levels of income and education, and low level of physical activ­ ity.3,8,11,24 Other risk factors include use of certain medica­ tions and particular underlying medical disorders (see later). Diet and lifestyle also may play a role in the develop­ ment of constipation (Table 18-3).

GENDER

The prevalence of self-reported constipation is two to three times higher in women than in men,10-12,16 and infrequent bowel movements (e.g., once a week) are reported almost exclusively by women.25 In one study of 220 normal sub­ jects eating their normal diets, 17% of women, but only 1% of men, passed less than 50 g of stool daily.26 The reason for the female predominance is unknown. A reduction in levels of steroid hormones has been observed in women with severe idiopathic constipation, although the clinical signifi­ cance of this finding is dubious.27 An overexpression of progesterone receptors on colonic smooth muscle cells has been reported to down-regulate contractile G proteins and up-regulate inhibitory G proteins.28 In addition, overexpression of progesterone receptor B on colonic muscle cells, thereby making them more sensitive to physiologic concentrations of progesterone, has been pro­ posed as an explanation for severe slow-transit constipation in some women.29

AGE

The prevalence of self-reported constipation among older adults ranges from 15% to 30%, with most,7,21,24,30,31 but not

Telephone interview Mailed questionnaire Mailed questionnaire

United States8 Canada9 Spain* 10,018 1,149 349

5,430

690

328

1,897

14,407 42,375 835

15,014 563

SAMPLE SIZE

RII, FC; RII, OD SR SR, RI, RII

RI, FC; RI, D

Straining and hard stools or frequency < three/wk SR; RI, FC; RI, OD

SR SR Straining and hard stools or frequency < three/wk Stool type and frequency

SR SR

DEFINITION OF CONSTIPATION

12.5, SR; 18.3, FC; 11.0, OD 3.6, FC; 13.8, D 4.6, FC; 4.6, OD 27.2, SR; 16.7, RI; 14.9, RII 29.5, SR; 19.2, RI; 14.0, RII

15 to >45 (mean, 49) 18 to >70 18 to >65 18-65

24.1

—

— 3.5 17.4

12.8 7.3

prevalence (%)

30-64

65-93

25-69

25-74 <40 to >80 30-64

12-74 Mean, 24 (65% students)

AGE range (yr)

FC: M, 17; F, 16 OD M, 6; F, 17 FC: M, 2.4 F, 4.8; D: M, 11.5; F, 16 — — SR: M, 18.4; F, 35.4; RI: M, 12.0; F, 21.0; RII: M, 8.3: F, 21.1

—

M, 0.6; F, 3.5

M, 8.06; F, 20.8 M, 1.3; F, 4.9 —

M, 7.0; F, 18.2 —

prevalence BY GENDER (%)

*From Garrigues V, Galvez C, Ortiz V, et al. Prevalence of constipation: Agreement among several criteria and evaluation of the diagnostic accuracy of qualifying symptoms and self-reported definition in a population-based survey in Spain. Am J Epidemiol 2004; 159:520-6. D, dyschezia; F, female; FC, functional constipation; M, male; OD, outlet delay; RI, Rome I criteria; RII, Rome II criteria; SR, self-report.

Mailed questionnaire

Mailed questionnaire

Face-to-face interview with questionnaire Mailed questionnaire

Face-to-face interview Questionnaire administered in person Face-to-face interview Face-to-face interview Mailed questionnaire

SURVEY METHOD

United States17

Olmsted County, Minn

16

Olmsted County, Minn15

East Bristol, UK14

United States11 United States12 Olmsted County, Minn13

United States3 Chapel Hill, NC1

LOCATION OF STUDY (REFERENCE)

Table 18-2  Population-Based Studies of the Prevalence of Constipation

Chapter 18  Constipation 261

262

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 18-3  Risk Factors for Constipation Advanced age Female gender Low level of education Low level of physical activity Low socioeconomic status Nonwhite ethnicity Use of certain medications (see Table 18-4)

all,8,9,12,17 studies showing an increase in prevalence with age. Constipation is particularly problematic in nursing home residents, among whom constipation is reported in almost half and 50% to 74% use laxatives on a daily basis.32,33 Similarly, hospitalized older patients appear to be at high risk of developing constipation. A study of patients on a geriatrics ward in the United Kingdom showed that up to 42% had a fecal impaction.34 Older adults also tend to seek medical assistance for con­ stipation more commonly than their younger counterparts. In an analysis of physician visits for constipation in the United States between 1958 and 1986, the frequency was about 1% in persons younger than 60, between 1% and 2% in those 60 to 65, and between 3% and 5% in those older than 65 years.21 Constipation in older adults is most commonly the result of excessive straining and hard stools30 rather than a decrease in stool frequency. In a community sample of 209 people ages 65 to 93 years, the main symptom used to describe constipation was the need to strain at defecation; 3% of men and 2% of women reported that their average bowel fre­ quencies were less than three/week.29 Possible causes for the increased frequency of straining in older adults include decreased food intake, reduced mobility, weakening of abdominal and pelvic wall muscles, chronic illness, psychological factors, and medications, particularly painrelieving drugs.19,32 Constipation is also common in children younger than 4 years.33 For example, in Great Britain, the frequency of a consultation for constipation in general practice was 2% to 3% for children ages 0 to 4, approximately 1% for women ages 15 to 64, 2% to 3% for both genders ages 65 to 74, and 5% to 6% for patients ages 75 years or older. Fecal retention with fecal soiling is a common cause of impaired quality of life and the need for medical attention in childhood.

ETHNICITY

In North America, constipation is reported more commonly by nonwhites than whites. In a survey of 15,014 persons, the frequency of constipation in whites was 12.2%, com­ pared with 17.3% in nonwhites.3 Both groups demonstrate similar age-specific increases in prevalence.7 In developing countries, constipation is less common among the native populations, in whom stool weights are three to four times more than the median of 106 g daily in Britain.26 In rural Africa, constipation appears to be rare.

SOCIOECONOMIC CLASS AND EDUCATION LEVEL

The prevalence of constipation is influenced by socioeco­ nomic status. In population-based surveys, subjects with a lower income status have higher rates of constipation as compared with those who have a higher income.3,6-8 In a survey of approximately 9000 Australians, men and women

of lower socioeconomic status were more likely to report constipation than those of higher socioeconomic status.35 Similarly, persons who have less education tend to have an increased prevalence of constipation as compared with those who have more education.3,8,9,11,16

DIET AND PHYSICAL ACTIVITY

Cross-sectional studies have not linked low intake of fiber with constipation,29,36 yet data suggest that increased con­ sumption of fiber decreases colonic transit time and increases stool weight and frequency.22 An analysis from the Nurses Health Study, which assessed the self-reported bowel habits of 62,036 women between the ages of 36 and 61 years, demonstrated that women who were in the highest quintile of fiber intake and who exercised daily were 68% less likely to report constipation than women who were in the lowest quintile of fiber intake and exercised less than once a week.24 Although other observational studies have supported a protective effect of physical activity on consti­ pation, results from trials designed to test this hypothesis are conflicting. In a trial designed to assess the effect of regular exercise on chronic constipation, symptoms did not improve after a four-week exercise program.37 In healthy sedentary subjects, a nine-week program of progressively increasing exercise had no consistent effect on whole-gut transit time or stool weight.38 Dehydration has been identified as a potential risk factor for constipation. Some but not all observational studies have found an association between slowed intestinal transit time and dehydration.36,39 Although patients with constipa­ tion are advised routinely to increase their intake of fluid, the benefit of increased fluid intake has not been investi­ gated thoroughly.

MEDICATION USE

Persons who use certain medications are at a substantially higher risk of constipation. In a review of 7251 patients with chronic constipation (and nonconstipated controls) from a general practice database, medications that were signifi­ cantly associated with constipation were opioids, diuretics, antidepressants, antihistamines, antispasmodics, anticon­ vulsants, and aluminum antacids (Table 18-4).40 The use of aspirin or other nonsteroidal anti-inflammatory drugs in the older population is associated with a small but significantly increased risk of constipation.14

COLONIC FUNCTION LUMINAL CONTENTS

The main contents of the colonic lumen are food residue, water and electrolytes, bacteria, and gas. Unabsorbed food entering the cecum contains carbohydrates that are resistant to digestion and absorption by the small intestine, such as starches and nonstarch polysaccharides (NSPs). Some of the unabsorbed carbohydrate serves as substrate for bacterial proliferation and fermentation, yielding short-chain fatty acids and gas (see Chapter 16). On average, bacteria repre­ sent approximately 50% of stool weight.41 In an analysis of feces from nine healthy subjects on a metabolically controlled British diet, bacteria constituted 55% of the total solids, and fiber represented approximately 17% of the stool weight.42 A meta-analysis of the effect of wheat bran on colonic function has suggested that bran increases stool weight and decreases mean colonic transit time in healthy volunteers.42 The effect of bran may be the result primarily of increased

Chapter 18  Constipation Table 18-4  Secondary Causes of Constipation Mechanical Obstruction Anal stenosis Colorectal cancer Extrinsic compression Rectocele or sigmoidocele Stricture Medications Antacids Anticholinergic agents (e.g., antiparkinsonian drugs, antipsychotics, antispasmodics, tricyclic antidepressants) Anticonvulsants (e.g., carbamazepine, phenobarbital, phenytoin) Antineoplastic agents (e.g., vinca derivatives) Calcium channel blockers (e.g., verapamil) Diuretics (e.g., furosemide) 5-Hydroxytryptamine3 antagonists (e.g., alosetron) Iron supplements Nonsteroidal anti-inflammatory drugs (e.g., ibuprofen) Mu-opioid agonists (e.g., fentanyl, loperamide, morphine) Metabolic and Endocrinologic Disorders Diabetes mellitus Heavy metal poisoning (e.g., arsenic, lead, mercury) Hypercalcemia Hyperthyroidism Hypokalemia Hypothyroidism Panhypopituitarism Pheochromocytoma Porphyria Pregnancy Neurologic and Myopathic Disorders Amyloidosis Autonomic neuropathy Chagas’ disease Dermatomyositis Intestinal pseudo-obstruction Multiple sclerosis Parkinsonism Progressive systemic sclerosis Shy-Drager syndrome Spinal cord injury Stroke

bulk within the colonic lumen; the increased bulk stimu­ lates propulsive motor activity. The particulate nature of some fibers also may stimulate the colon. For example, ingestion of coarse bran, 10 g twice daily, was shown to reduce colonic transit time by about one third, whereas ingestion of the same quantity of fine bran led to no signifi­ cant decrease.41 Similarly, ingestion of inert plastic particles similar in size to coarse bran increased fecal output by almost three times their own weight and decreased colonic transit time.43

ABSORPTION OF WATER AND SODIUM

The colon avidly absorbs sodium and water (see Chapter 99). Increased water absorption can lead to smaller, harder stools. The colon extracts most of the 1000 to 1500 mL of fluid that crosses the ileocecal valve, and leaves only 100 to 200 mL of fecal water daily. Less reabsorption of electro­ lytes and nutrients takes place in the colon than in the small intestine, and sodium-chloride exchange and short-chain fatty acid transport are the principal mechanisms for stimu­ lating water absorption. Colonic absorptive mechanisms remain intact in patients with constipation. One proposed pathophysiologic mechanism in slow-transit constipation is that the lack of peristaltic movement of contents through the colon allows more time for bacterial degradation of stool

solids and increased NaCl and water absorption, thereby decreasing both stool weight and frequency.44 The volume of stool water and quantity of stool solids seem to be reduced proportionally in constipated persons.45

DIAMETER AND LENGTH

A wide or long colon may lead to a slow colonic transit rate (see Chapter 96). Although only a small fraction of patients with constipation have megacolon or megarectum, most patients with dilatation of the colon or rectum report constipation. Colonic width can be measured on barium enema films. A width of more than 6.5 cm at the pelvic brim is abnormal and has been associated with chronic constipation.46

MOTOR FUNCTION

Colonic muscle has four main functions (see also Chapter 98): (1) delays passage of the luminal contents so as to allow time for the absorption of water; (2) mixes the contents and allows contact with the mucosa; (3) allows the colon to store feces between defecations; and (4) propels the contents toward the anus. Muscle activity is affected by sleep and wakefulness, eating, emotion, the contents of the colon, and drugs. Nervous control is partly intrinsic and partly extrin­ sic by the sympathetic nerves and the parasympathetic sacral outflow. Transit of contents along the colon takes hours or days (longer than transit in other portions of the gastrointestinal tract). In a study of 73 healthy subjects, the mean colonic transit time was 35 hours.47 In another similar study, the mean colonic transit time in healthy volunteers was 34 hours, with an upper limit of normal of 72 hours.48 Scintigraphic studies in constipated subjects have shown that overall transit of colonic contents is slow. In some patients, the rate of movement of contents is approximately normal in the ascending colon and hepatic flexure but delayed in the transverse and left colon. Other patients show slow transit in the right and left sides of the colon.49 Colonic propulsions are of two basic types, low-amplitude propagated contractions (LAPCs) and high-amplitude prop­ agated contractions (HAPCs).50 The frequency and duration of HAPCs are reduced in some patients with constipation. In one study, 14 chronically constipated patients with proved slow transit of intestinal contents and one or fewer bowel movements weekly were compared with 18 healthy subjects. Four of the patients had no peristaltic movement, whereas peristaltic movement was normal in all the healthy subjects during a 24-hour period. Peristaltic movements in other subjects with constipation were fewer in number and shorter in duration, and thus passed for a shorter distance along the colon, as compared with the findings in the healthy controls. All the healthy subjects reported abdominal dis­ comfort or an urge to defecate during peristaltic movements, and two defecated, whereas only four of the 14 subjects with constipation experienced any sensation during such move­ ments, and none defecated.51

INNERVATION AND THE INTERSTITIAL CELLS OF CAJAL

Proximal colonic motility is under the involuntary control of the enteric nervous system, whereas defecation is volun­ tary. Slow-transit constipation may be related to autonomic dysfunction.52,53 Histologic studies have shown abnormal numbers of myenteric plexus neurons involved in excit­ atory or inhibitory control of colonic motility, thereby resulting in decreased amounts of the excitatory transmitter substance P54 and increased amounts of the inhibitory trans­

263

264

Section III  Symptoms, Signs, and Biopsychosocial Issues mitters vasoactive intestinal polypeptide (VIP) or nitric oxide (NO).55 Interstitial cells of Cajal (ICCs) are the intestinal pace­ maker cells and play an important role in regulating gastro­ intestinal motility. They facilitate the conduction of electrical current and mediate neural signaling between enteric nerves and muscles. ICCs initiate slow waves throughout the gastrointestinal tract. Confocal images of ICCs in patients with slow-transit constipation show not only reduced numbers but also abnormal morphology of ICCs, with irregular surface markings and a decreased number of dendrites. In patients with slow-transit constipa­ tion, the number of ICCs has been shown to be decreased in the sigmoid colon56 or the entire colon.57,58 Pathologic examination of colectomy specimens of 14 patients with severe intractable constipation has revealed decreased numbers of ICCs and myenteric ganglion cells throughout the colon.59

DEFECATORY FUNCTION

The process of defecation in healthy persons begins with a predefecatory period, during which the frequency and amplitude of propagating sequences (three or more succes­ sive pressure waves) are increased. Stimuli such as waking and meals (gastroileal reflex, also referred to as gastrocolic reflex) can stimulate this process. This predefecatory period is blunted, and may be absent, in patients with slow-transit constipation.50 The gastroileal reflex also is diminished in persons with slow-transit constipation. Stool is often present in the rectum before the urge to defecate arises. The urge to defecate is usually experienced when stool comes into contact with receptors in the upper anal canal. When the urge to defecate is resisted, retrograde movement of stool may occur and transit time increases throughout the colon (see Chapter 98).60 Although the sitting or squatting position seems to facili­ tate defecation, the benefit of squatting has not been studied in patients with constipation. Full flexion of the hips stretches the anal canal in an anteroposterior direction and straightens the anorectal angle, thereby promoting emptying of the rectum.61 Contraction of the diaphragm and abdomi­ nal muscles raises intrapelvic pressure, and the pelvic floor relaxes simultaneously. Striated muscular activity expels rectal contents, with little contribution from colonic or rectal propulsive waves. Coordinated relaxation of the puborectalis muscle, which maintains the anorectal angle, and external anal sphincter at a time when pressure is increasing in the rectum results in expulsion of stool (Fig. 18-1). The length of the colon emptied during spontaneous def­ ecation varies but most commonly extends from the descending colon to the rectum.62 When the propulsive action of smooth muscle is normal, defecation usually requires minimal voluntary effort. If colonic and rectal waves are infrequent or absent, however, the normal urge to defecate may not occur.51

During straining

Puborectalis muscle

Anorectal angle Descent of the pelvic floor

Figure 18-1.  Physiology of defecation. Defecation requires relaxation of the puborectalis muscle with descent of the pelvic floor and straightening of the anorectal angle during straining, as well as relaxation of the internal anal sphincter. (From Lembo A, Camilleri M. Chronic constipation. N Engl J Med 2003; 349:1360-8.)

Human stools may vary in consistency from small hard lumps to liquid. The water content of stool determines con­ sistency. Rapid colonic transit of fecal residue leads to diminished water absorption and (perhaps counterintui­ tively) an increase in the bacterial content of the stool. The Bristol Stool Scale25 is used in the assessment of constipa­ tion and is regarded as the best descriptor of stool form and consistency (Fig. 18-2). Stool consistency appears to be a better predictor of whole-gut transit time than of defecation frequency or stool volume.64

CLASSIFICATION Mechanical small and large bowel obstruction, medications, and systemic illnesses can cause constipation, and these causes of secondary constipation must be excluded, espe­ cially in patients presenting with a new onset of consti­ pation (see Table 18-4). Most often, however, constipation is caused by disordered function of the colon or rectum (functional constipation). Functional constipation can be divided into three broad categories—normal-transit con­ stipation, slow-transit constipation, and defecatory or rectal evacuation disorders (Table 18-5). In a study of more than 1000 patients with functional constipation who were evaluated at the Mayo Clinic, 59% were found to have normal-transit constipation, 25% had defecatory disorders, 13% had slow-transit constipation, and 3% had a com­ bination of a defecatory disorder and slow-transit constipation.65

SIZE AND CONSISTENCY OF STOOL

In a study of normal subjects who were asked to expel single hard spheres of different sizes from the rectal ampulla, the intrarectal pressure and time needed to pass the objects varied inversely with their diameters. Small hard stools are more difficult to pass than large soft stools. When larger stimulated stools were tested, a hard stool took longer to expel than a soft silicone rubber object of approximately the same shape and volume. Similarly, more subjects were able to expel a 50-mL water-filled compressible balloon than a hard 1.8-cm sphere.63

PATHOPHYSIOLOGY NORMAL-TRANSIT CONSTIPATION

In normal-transit constipation, stool travels along the colon at a normal rate.66 Patients with normal-transit constipation may have misperceptions about their bowel frequencies and often exhibit psychosocial distress.67 Some patients have abnormalities of anorectal sensory and motor function indistinguishable from those in patients with slow-transit

Chapter 18  Constipation Table 18-5  Clinical Classification of Functional Constipation CATEGORY

FEATURES

CHARACTERISTIC FINDINGS

Normal-transit constipation

Incomplete evacuation; abdominal pain may be present but not a predominant feature Infrequent stools (e.g., ≤1/wk); lack of urge to defecate; poor response to fiber and laxatives; generalized symptoms, including malaise and fatigue; more prevalent in young women Frequent straining; incomplete evacuation; need for manual maneuvers to facilitate defecation

Normal physiologic test results

Slow-transit constipation

Defecatory disorders (pelvic floor dysfunction, anismus, descending perineum syndrome, rectal prolapse)

Whole gut transit time

Type of stool

Description

Pictorial representation

Long transit (e.g., 100 hours) Type 1

Separate hard lumps, like nuts, hard to pass

Type 2

Sausage shaped but lumpy

Type 3

Like sausage but with cracks on its surface

Type 4

Like sausage or snake, smooth and soft

Type 5

Soft blobs with clear-cut edges (passed easily)

Type 6

Fluffy pieces with ragged edges, a mushy stool

Type 7

Watery, no solid pieces

Abnormal balloon expulsion test and/or rectal manometry

similar to those seen in persons with IBS.71 In patients with more severe symptoms, the pathophysiology includes delayed emptying of the proximal colon and fewer HAPCs after meals. Colonic inertia is a term used to describe the disorder in patients with symptoms at the severe end of the spectrum. In this condition, colonic motor activity fails to increase after a meal,72 ingestion of bisacodyl,73 or adminis­ tration of a cholinesterase inhibitor such as neostigmine.74

DEFECATORY DISORDERS

Entirely liquid

Short transit (e.g., 10 hours) Figure 18-2.  Bristol Stool Form Scale. Common stool forms and their consistency in relation to whole-gut transit time are shown. (From Heaton KW, Radvan J, Cripps H, et al. Defecation frequency and timing, and stool form in the general population: A prospective study. Gut 1992; 33:818-24.)

constipation.68 Whether increased rectal compliance and reduced rectal sensation are effects of chronic constipation or contribute to the failure of the patients to experience an urge to defecate is unclear. Most patients, however, have normal physiologic testing. IBS with constipation differs from normal-transit constipation in that abdominal pain is the predominant symptom in IBS (see Chapter 118).

SLOW-TRANSIT CONSTIPATION

Retention in colon of >20% of radiopaque markers five days after ingestion

Slow-transit constipation is most common in young women and is characterized by infrequent bowel movements (less than one bowel movement/week). Associated symptoms include abdominal pain, bloating, and malaise. Symptoms are often intractable, and conservative measures such as fiber supplements and osmotic laxatives are usually ineffec­ tive.69,70 The onset of symptoms is gradual and usually occurs around the time of puberty. Slow-transit constipa­ tion arises from disordered colonic motor function. Patients who have mild delays in colonic transit have symptoms

Defecatory disorders arise from failure to empty the rectum effectively because of an inability to coordinate the ab­ dominal, rectoanal, and pelvic floor muscles. Many patients with defecatory disorders also have slow-transit constipa­ tion75 Defecatory disorders are also known as anismus, dyssynergia, pelvic floor dyssynergia, spastic pelvic floor syndrome, obstructive defecation, or outlet obstruction. These disorders appear to be acquired and may start in childhood. They may be a learned behavior to avoid some discomfort associated with the passage of large hard stools or pain associated with attempted defecation in the setting of an active anal fissure or inflamed hemorrhoids. Patients with defecatory disorders commonly have inappropriate contraction of the anal sphincter when they bear down (Fig. 18-3). This phenomenon can occur in asymptomatic subjects but is more common among patients who complain of difficult defecation.76 Some patients with a defecatory disorder are unable to raise intrarectal pressure to a level sufficient to expel stool, a disturbance that manifests clinically as failure of the pelvic floor to descend on straining.77 Defecatory disorders are particularly common in older patients with chronic constipation and excessive straining, many of whom do not respond to standard medical treat­ ment.78 Defecatory disorders rarely are associated with structural abnormalities such as rectal intussusception, an obstructing rectocele, megarectum, or excessive perineal descent.79 Patients with defecatory disorders may report infrequent bowel movements, ineffective and excessive straining, and the need for manual disimpaction; however, symptoms, particularly in the case of pelvic floor dysfunction, do not correlate with physiologic findings.80 For a diagnosis of a defecatory disorder, a Rome working group81 has specified the criteria listed in Table 18-6. In patients with this dis­ order, constipation is functional and caused by dysfunction of the pelvic floor muscles as determined by physiologic tests. Pelvic floor dyssynergia is a subset of these patients in which the anal sphincter fails to relax more than 20% of its basal resting pressure during attempted defecation, despite the presence of adequate propulsive forces in the rectum.

265

Section III  Symptoms, Signs, and Biopsychosocial Issues

EMG Pressure cm H2O

Control subject

Cough

Pressure cm H2O EMG

Strain

200 µV

Ext. sphincter

100 Anal canal

50 0 Cough (x 3)

Constipated patient

266

Strain

5 sec

200 µV

Ext. sphincter

150 100 50 0

Anal canal 5 sec

Figure 18-3.  Electromyography (EMG) and pressure tracings during defecation in a normal (control) subject and a constipated patient with a defecatory disorder. In both the control subject and constipated patient, a cough produces a rise in pressure. When a normal subject strains (upper tracing), EMG activity of the external anal sphincter is inhibited and pressure in the anal canal falls. In a constipated patient with a defecatory disorder, EMG activity of the anal sphincter is not inhibited on straining, and pressure within the anal canal increases (lower tracing). This paradoxical contraction has been termed anismus, anal dyssynergia, and spastic perineum. (From Preston DM, Lennard-Jones JE. Anismus in chronic constipation. Dig Dis Sci 1985; 30:413-8.)

Table 18-6  Rome III Criteria for Functional Defecation Disorders81* The patient must satisfy diagnostic criteria for functional constipation (see Table 18-1). During repeated attempts to defecate, the patient must have at least two of the following: Evidence of impaired evacuation, based on balloon expulsion test or imaging Inappropriate contraction of pelvic floor muscles (i.e., anal sphincter or puborectalis) or less than 20% relaxation of basal resting sphincter pressure by manometry, imaging, or EMG Inadequate propulsive forces assessed by manometry or imaging *Criteria fulfilled for the previous three months with symptom onset at least six months prior to diagnosis. EMG, electromyography.

Functional fecal retention (FFR) is the most common def­ ecatory disorder in children. It is a learned behavior that results from withholding defecation, often because of fear of a painful bowel movement.82 The symptoms are common and may result in secondary encopresis (fecal incontinence) because of leakage of liquid stool around a fecal impaction. FFR is the most common cause of encopresis in childhood (see Chapter 17).83

DISORDERS OF THE ANORECTUM AND PELVIC FLOOR RECTOCELE

A rectocele is the bulging or displacement of the rectum through a defect in the anterior rectal wall. In women, the perineal body supports the anterior rectal (posterior vaginal)

wall above the anorectal junction, and a layer of fascia runs from the rectovaginal pouch of Douglas to the perineal body and adheres to the posterior vaginal wall. The anterior rectal wall is unsupported above the level of the perineal body, and the rectovaginal septum can bulge anteriorly to form a rectocele (Fig. 18-4). Rectoceles can arise from damage to the rectovaginal septum or its supporting struc­ tures during vaginal childbirth. These injuries are exacer­ bated by repetitive increases in intra-abdominal pressure and the long-term effects of gravity. Prolapse of other pelvic organs may be present. For example, urinary incontinence, as well as a previous hysterectomy, has been reported to be more common in patients with a rectocele than in patients with difficult defecation but no demonstrable rectocele.84 Studies using defecating proctography (see later) have shown that rectoceles are common in symptomless healthy women and may protrude as much as 4 cm from the line of the anterior rectal wall without causing bowel symptoms, although 2 cm is the generally accepted lower limit of a rectocele that may be regarded as clinically significant.85 Symptomatic patients report the inability to complete fecal evacuation, perineal pain, sensation of local pressure, and appearance of a bulge at the vaginal opening on straining. Women may report the need to use their thumb or fingers to support the posterior vaginal wall to complete defeca­ tion.84 Women also may report the need to use a finger to evacuate the rectum digitally. Defecating proctography can be used to demonstrate a rectocele, measure its size, and determine whether barium becomes trapped within the rectocele. In one study, trap­ ping of barium in rectoceles changed with the degree of rectal emptying and was related to the size of the rectocele86; however, the size of the rectocele or degree of emptying on defecation has not been shown to correlate with the outcome of surgical repair.87,88 Asymptomatic women with rectoceles do not require sur­ gical treatment. Kegel exercises (designed to strengthen the pelvic floor muscles that support the urethra, bladder, uterus, and rectum) and instructions to avoid repetitive increases in intra-abdominal pressure may help prevent progression of the rectocele. Surgery should be considered only for patients in whom contrast is retained during defe­ cography and patients in whom constipation is relieved with digital vaginal pressure to facilitate defecation.89 Surgi­ cal repair can be performed by endorectal, transvaginal, or transperineal approaches. Other types of genital prolapse may also be present, and collaboration between the surgeon and gynecologist may be appropriate. In carefully selected patients surgical repair benefits approximately 75% of patients. In a review of 89 women who underwent a com­ bined transvaginal and transanal rectocele repair for symp­ toms of obstructive defecation, the repair was successful in 71% of patients, as assessed by the absence of symptoms after one year.90 Reduction in the size of the rectocele, as judged by defecating proctography, does not appear to cor­ relate clearly with improvement in symptoms.88

DESCENDING PERINEUM SYNDROME

In the descending perineum syndrome, the pelvic floor descends to a greater extent than normal (1 to 4 cm) when the patient strains during defecation, and rectal expulsion is difficult. The anorectal angle is widened as a result of pelvic floor weakness, and the rectum is more vertical than normal. The perineal body is weak (thereby facilitating for­ mation of a rectocele), and the lax muscular support favors intrarectal mucosal intussusception or rectal prolapse. The pelvic floor may not provide the resistance necessary for extrusion of solid stool through the anal canal. A common

Chapter 18  Constipation

Levator plate Rectum

A

Perineal body

Vagina

Rectovaginal septum

Rectocele

B

Figure 18-4.  Development of a rectocele. A, Normal anatomy of the female pelvis. The levator plate is almost horizontal, supporting the rectum and vagina. The perineal body provides support for the lower posterior vaginal wall; above it lies the rectovaginal septum. B, Weakness of the pelvic floor leads to a more vertical levator plate. The perineal body is attenuated, which favors the formation of a rectocele. The laxity of the pelvic floor also favors rectal mucosal prolapse. (From Loder PB, Phillips RKS. Rectocele and pelvic floor weakness. In: Kamm MA, Lennard-Jones JE, editors. Constipation. Peterfield, England: Wrightson Biomedical; 1994. p 281.)

reason for pelvic floor weakness is trauma or stretching during parturition. In some cases, repeated and prolonged defecation appears to be a damaging factor. Symptoms include constipation, incomplete rectal evacuation, exces­ sive straining and, less commonly, digital rectal evacua­ tion.91 Electrophysiologic studies show partial denervation of the striated muscle and evidence of pudendal nerve damage. Histologic examination of operative specimens of the pelvic floor muscles confirms loss of muscle fibers.

DIMINISHED RECTAL SENSATION

The urge to defecate depends in part on tension within the rectal wall (determined by the tone of the circular muscle of the rectal wall), rate and volume of rectal distention, and size of the rectum. Some patients with constipation appear to feel pain normally as the rectum is distended to the maximal tolerable volume, but they fail to experience an urge to defecate with intermediate volumes.92 In a study of women with severe idiopathic constipation, a higher than normal electrical stimulation current applied to the rectal mucosa was required to elicit pain, thereby suggesting a possible rectal sensory neuropathy.93 Rectal hyposensitivity (RH) is defined as insensitivity of the rectum to balloon distention on anorectal physiologic investigation, although the pathophysiology of RH is not entirely clear. Constipation is the most common presenting symptom of RH. In an investigation of 261 patients with RH, 38% had a history of pelvic surgery, 22% had a history of anal surgery, and 13% had a history of spinal trauma.94

RECTAL PROLAPSE AND SOLITARY RECTAL ULCER SYNDROME

Full-thickness rectal prolapse and solitary rectal ulcer syn­ drome are part of a spectrum of defects that arise from weakening of the pelvic floor. Some patients may complain of many fruitless visits to the bathroom, with prolonged straining in response to a constant desire to defecate. The patient has a sense of incomplete evacuation and may spend an hour or more daily on the toilet. The infrequent passage of small hard stools is common, as are other features of a

functional bowel disorder, such as abdominal pain and distention. Rectal prolapse refers to complete protrusion of the rectum through the anus (see Chapter 125). Occult (asymp­ tomatic) rectal prolapse has been found in 33% of patients with clinically recognized rectoceles and defecatory dys­ function.95 Rectal prolapse can be detected easily on phys­ ical examination by asking the patient to strain as if to defecate. A laparoscopic rectopexy—in which the prolapsed rectum is raised and secured with sutures to the adjacent fascia—is the recommended treatment.96 Solitary rectal ulcer syndrome is a rare disorder character­ ized by erythema or ulceration generally of the anterior rectal wall as a result of chronic straining (see Chapter 115). Mucus and blood may be passed when the patient strains during defecation.97,98 Endoscopic findings may include erythema, hyperemia, mucosal ulceration, and polypoid lesions. Misdiagnosis may occur because of the heteroge­ neous findings and misleading name of the syndrome (an ulcer need not be present). In a study of 98 patients with solitary rectal ulcer syndrome, 26% were initially diag­ nosed incorrectly. In patients with a rectal ulcer or mucosal hyperemia, the most common misdiagnoses were Crohn’s disease and ulcerative colitis. In those with a polypoid lesion, the most common misdiagnosis was a neoplastic polyp.99 Histology of full-thickness specimens of the lesion reveals extension of the muscularis mucosa between crypts and disorganization of the muscularis propria. Defecogra­ phy, transrectal ultrasonography, and anorectal manometry are helpful in the diagnosis. Varying degrees of rectal prolapse exist in association with solitary rectal ulcer syndrome. Rectal prolapse and paradoxical contraction of the puborectalis muscle can lead to rectal trauma because of the high pressures generated within the rectum. In addition, rectal mucosal blood flow is reduced.100 Medical treatment may be difficult, and a single optimal therapy does not exist. The patient should be advised to resist the urge to strain. Bulk laxatives and dietary fiber may be of some benefit.101 Surgery may be required; rectopexy is performed most commonly. Of patients who undergo

267

268

Section III  Symptoms, Signs, and Biopsychosocial Issues surgery for solitary rectal ulcer syndrome with rectal pro­ lapse, 55% to 60% report long-term satisfaction, although a colostomy is eventually required in approximately one third of patients.102 Repair of a rectal prolapse may aggravate constipation. Biofeedback appears to be a promising mode of therapy for patients with solitary rectal ulcer syndrome.103

SYSTEMIC DISORDERS HYPOTHYROIDISM

Constipation is the most common gastrointestinal com­ plaint in patients with hypothyroidism. The pathologic effects are caused by an alteration of intestinal motor func­ tion and possible infiltration of the intestine by myxedema­ tous tissue. The basic electrical rhythm that generates peristaltic waves in the duodenum decreases in hypothy­ roidism, and small bowel transit time is increased.104 Myxedema megacolon is rare but can result from myx­ edematous infiltration of the muscle layers of the colon. Symptoms include abdominal pain, flatulence, and constipation.105

DIABETES MELLITUS

The mean colonic transit time is longer in diabetics than in healthy controls. In one study, the mean total colonic transit time in 28 diabetic patients (34.9 ± 29.6 hours; mean ± SD) was significantly longer than that in 28 healthy subjects (20.4 ± 15.6 hours; P < 0.05).106 Among the 28 diabetic patients, 9 of 28 (32%) met the Rome II criteria for consti­pation and 14 of 28 (50%) had cardiovascular autonomic neuropathy. The mean colonic transit times in diabetic patients with and without cardiovascular auto­ nomic neuropathy were similar. By contrast, a previous study reported that asymptomatic diabetic patients with cardiovascular autonomic neuropathy had significantly longer whole-gut transit times (although still within the range of normal) than a control group without evidence of neuropathy.107 In another study, diabetic patients with mild constipation demonstrated delayed colonic myoelec­ trical and motor responses after ingestion of a standard meal, whereas diabetics with severe constipation had no increases in these responses after food. Neostigmine increased colonic motor activity in all diabetic patients, suggesting that the defect was neural rather than muscular (see Chapter 35).108

HYPERCALCEMIA

Constipation is a common symptom of hypercalcemia resulting from hyperparathyroidism.109 It also may be a manifestation of hypercalcemia caused by other conditions, such as sarcoidosis or malignancy involving bone (see Chapter 35).

NERVOUS SYSTEM DISEASE LOSS OF CONSCIOUS CONTROL

A decrease in or complete loss of bodily perception as a result of cerebral disability or dementia may lead to defeca­ tory failure, possibly because of inattention.

PARKINSON’S DISEASE

Constipation occurs frequently in patients with Parkinson’s disease (PD). In a study of 12 patients with PD compared with normal controls, slow colonic transit, decreased phasic

rectal contractions, weak abdominal wall muscle contrac­ tion, and paradoxical anal sphincter contraction on defeca­ tion were all features in patients with PD and frequent constipation.110 Loss of dopamine-containing neurons in the central nervous system is the underlying defect in PD; a defect in dopaminergic neurons in the enteric nervous system also may be present. Histopathologic studies of the myenteric plexuses of the ascending colon in 11 patients with PD and constipation revealed that in 9 patients, the number of dopamine-positive neurons was one tenth or less the number in control subjects. Dopamine concentrations in the muscularis externa were significantly lower in patients with PD than in controls (P < 0.01).111 Another possible contributor to constipation is the inabil­ ity of some patients with PD to relax the striated muscles of the pelvic floor on defecation. This finding is a local mani­ festation of the extrapyramidal motor disorder that affects skeletal muscle. Preliminary observations suggest that injec­ tion of botulinum toxin into the puborectalis muscle is a potential therapy for this type of outlet dysfunction consti­ pation in patients with PD.112,113

MULTIPLE SCLEROSIS

Constipation is common among patients with multiple sclerosis (MS). In an unselected group of 280 patients with MS, the frequency of constipation (defined as diminished bowel frequency, digitation to facilitate defecation, or the use of laxatives) was approximately 43%. Almost 25% of the subjects passed fewer than three stools/week, and 18% used a laxative more than once a week. Constipation cor­ related with the duration of MS but preceded the diagnosis of MS in 45% of subjects. Constipation did not correlate with immobility or the use of medications.114 In another questionnaire study of 221 patients with MS, the frequency of constipation was as high as 54%.115 Constipation in patients with MS can be multifactorial and related to a reduction in postprandial colonic motor activity, limited physical activity, and medications with constipating side effects. Patients with advanced MS and constipation have evi­ dence of a visceral neuropathy. In a group of patients with advanced MS and severe constipation, all had evidence of disease in the lumbosacral spinal cord and decreased com­ pliance of the colon. Motor and electrophysiologic measure­ ments have shown that the usual increase in colonic motor activity after meals is absent. Among less severely affected patients, slow colonic transit and manometric evidence of pelvic floor muscular and anal sphincter dysfunction have been demonstrated. Patients may have fecal inconti­ nence.116,117 Therapy with biofeedback has been reported to relieve constipation and fecal incontinence, although in a study of 13 patients with MS who underwent biofeedback for either constipation or incontinence, only 38% improved (see Chapter 17).118

SPINAL CORD LESIONS Lesions Above the Sacral Segments

Spinal cord lesions or injury above the sacral segments lead to an upper motor neuron disorder, with severe constipa­ tion. The resulting delay in colonic transit affects the recto­ sigmoid colon primarily.119,120 In a study of patients with severe thoracic spinal cord injury, colonic compliance was abnormal, with a rapid rise in colonic pressure on instilla­ tion of relatively small volumes of fluid. Motor activity after meals did not increase, but the colonic response to neostigmine was normal, thereby suggesting absence of myopathy.

Chapter 18  Constipation Studies of anorectal function in patients with severe trau­ matic spinal cord injury have shown that rectal sensation to distention is abolished, although a dull pelvic sensation is experienced by some patients at maximum levels of rectal balloon distention. Anal relaxation on rectal disten­ tion is exaggerated and occurs at a lower balloon volume than in normal subjects. Distention of the rectum leads to a linear increase in rectal pressure, without the plateau at inter­mediate values seen in normal subjects, and ends in high-pressure rectal contractions after a relatively small volume (100 mL) has been instilled into the balloon. As expected, the rectal pressure generated by straining is lower in patients than in control subjects and is less with higher than lower spinal cord lesions. Patients demonstrate a loss of conscious external anal sphincter control, and the sphincter does not relax on straining, suggesting that in normal subjects, descending inhibitory pathways are present.121 These findings explain why some patients with spinal cord lesions experience not only constipation, but also sudden uncontrollable rectal expulsion with incon­ tinence. Other patients cannot empty the rectum in response to laxatives or enemas, possibly because of failure of the external anal sphincter to relax, and they may require manual evacuation. Electrical stimulation of anterior sacral nerve roots S2, S3, and S4 via electrodes implanted for urinary control in para­ plegic patients leads to a rise in pressure within the sigmoid colon and rectum and contraction of the external anal sphincter. Contraction of the rectum and relaxation of the internal anal sphincter persist for a short time after the stimulus ceases. By appropriate adjustment of the stimulus in one study, it was possible for 5 of 12 paraplegic patients to evacuate feces completely and for most of the others to increase the frequency of defecation and reduce the time spent emptying the rectum.122 In another series, left-sided colonic transit time decreased with regular sacral nerve stimulation.123

Lesions of the Sacral Cord, Conus Medullaris, Cauda Equina, and Nervi Erigentes (S2 to S4)

Neural integration of anal sphincter control and rectosig­ moid propulsion occurs in the sacral segments of the spinal cord. The motor neurons that supply the striated sphincter muscles are grouped in Onuf’s nucleus at the level of S2. There is evidence that efferent parasympathetic nerves that arise in the sacral segments enter the colon at the region of the rectosigmoid junction and extend distally in the inter­ muscular plane to reach the level of the internal anal sphincter and proximally to the midcolon via the ascending colonic nerves, which retain the structure of peripheral nerves (see Chapter 98).124 Damage to sacral segments of the spinal cord or to efferent nerves leads to severe constipation. Fluoroscopic studies show a loss of progression of contractions in the left colon. When the colon is filled with fluid, the intraluminal pres­ sure generated is lower than normal, in contrast with the situation after higher lesions of the spinal cord. The distal colon and rectum may dilate, and feces may accumulate in the distal colon. Spasticity of the anal canal can occur. Loss of sensation of the perineal skin may extend to the anal canal, and rectal sensation may be diminished. Rectal wall tone depends on the level of the spinal lesion. In a study of 25 patients with spinal cord injury, rectal tone was signifi­ cantly higher than normal in patients with acute and chronic supraconal lesions but significantly lower than normal in patients with acute and chronic conal or cauda equina lesions.125

STRUCTURAL DISORDERS OF THE COLON, RECTUM, ANUS, AND PELVIC FLOOR OBSTRUCTION

Anal atresia in infancy, anal stenosis later in life, or obstruction of the colon may manifest as constipation. Obstruction of the small intestine generally manifests as abdominal pain and distention, but constipation and inability to pass flatus also may be features (see Chapters 96 and 119).

DISORDERS OF SMOOTH MUSCLE Myopathy Affecting Colonic Muscle

Congenital or acquired myopathy of the colon usually mani­ fests as pseudo-obstruction. The colon is hypotonic and inert (see Chapter 120).

Hereditary Internal Anal Sphincter Myopathy

Hereditary internal anal sphincter myopathy is a rare condi­ tion characterized by constipation with difficulty in rectal expulsion and episodes of severe proctalgia fugax, defined as the sudden onset of brief episodes of pain in the anorectal region.126-128 Three affected families have been reported. The mode of inheritance appears to be autosomal dominant with incomplete penetrance. In symptomatic persons, the inter­ nal anal sphincter muscle is thickened, and resting anal pressure is increased greatly. In two of the described patients, treatment with a calcium channel blocker improved pain but had no effect on constipation. In another family, two patients were treated by internal anal sphincter strip myectomy; one showed marked improvement and one had improvement in the constipation but only slight improve­ ment in the pain. Examination of the muscle strips showed myopathic changes with polyglucosan bodies (glucose poly­ mers) in the smooth muscle fibers and increased endomysial fibrosis.

Progressive Systemic Sclerosis

Progressive systemic sclerosis (scleroderma) may lead to constipation. In patients with progressive systemic sclerosis and constipation, 9 of 10 had no increase in colonic motor activity after ingestion of a 1000-kcal meal. Histologic examination of colonic specimens from these subjects revealed smooth muscle atrophy of the colonic wall (see Chapter 35).129

Muscular Dystrophies

Muscular dystrophies usually are regarded as disorders of striated muscle, but visceral smooth muscle also may be abnormal. In myotonic muscular dystrophy, a condition in which skeletal muscle fails to relax normally, megacolon may be found, and abnormal function of the anal sphincter is demonstrable.130 Cases associated with intestinal pseudoobstruction have been reported (see Chapter 120).131

DISORDERS OF ENTERIC NERVES Congenital Aganglionosis or Hypoganglionosis

Congenital absence or reduction in the number of ganglia in the colon leads to functional colonic obstruction with proximal dilatation, as seen in Hirschsprung’s disease and related conditions (see Chapter 96). In Hirschsprung’s disease, ganglion cells in the distal colon are absent because of an arrest in the caudal migration of neural crest cells in the intestine during embryonic development. Although most patients present during early childhood, often with

269

270

Section III  Symptoms, Signs, and Biopsychosocial Issues delayed passage of meconium, some patients with a rela­ tively short segment of involved colon present later in life.132 Typically, the colon narrows at the area that lacks ganglion cells, and the bowel proximal to the narrowing is usually dilated. Two genetic defects have been identified in patients with Hirschsprung’s disease—a mutation in the RET (rearranged during transfection) proto-oncogene, which is involved in the development of neural crest cells, and a mutation in the gene that encodes the en­ dothelin B receptor, which affects intracellular calcium levels.133,134 Hypoganglionosis is reported when small, sparse myen­ teric ganglia are seen. Neuronal counts can be made on full-thickness tissue specimens and compared with pub­ lished reference values obtained from autopsy material. Establishing the diagnosis of hypoganglionosis is not easy, because of variations in the normal density of neurons.135 Quantitative declines in the number of neurons in the enteric nervous system also are seen in patients with severe slow-transit constipation and characterized morphologi­ cally as oligoneuronal hypoganglionosis.136

Congenital Hyperganglionosis (Intestinal Neuronal Dysplasia)

Congenital hyperganglionosis, or intestinal neuronal dys­ plasia, is a developmental defect characterized by hyperpla­ sia of the submucosal nerve plexus. Clinical manifestations of the disease are similar to those seen in Hirschsprung’s disease and include young age of onset and symptoms of intestinal obstruction (see Chapter 96). In contrast to func­ tional constipation, affected children do not have symptoms of soiling or evidence of a fecaloma.137 A multicenter study of interobserver variation in the histologic interpretation of findings in children with constipation caused by abnormali­ ties of the enteric nervous system has shown complete agreement in the diagnosis of Hirschsprung’s disease but accord in only 14% of children with colonic motility disor­ ders other than aganglionosis. Some of the clinical features and histologic changes previously associated with congeni­ tal hyperganglionosis may be age-related and evolve to normal as children age.135 A diagnosis of congenital hyper­ ganglionosis can be made on the basis of hyperganglionosis of the submucous plexus with giant ganglia and at least one of the following features in rectal biopsy specimens: (1) ectopic ganglia; (2) increased acetylcholinesterase (AChE) activity in the lamina propria; and (3) increased AChE nerve fibers around the submucosal blood vessels. Most patients with congenital hyperganglionosis respond to conservative treatment, including laxatives. Internal anal sphincter myectomy may be performed if conservative management fails.138

Acquired Neuropathies

Chagas’ disease, which results from infection with Trypanosoma cruzi, is the only known infectious neuropathy. The reason for neuronal degeneration in this disorder is unclear but may have an immune basis.139 Patients present with progressively worsening symptoms of constipation and abdominal distention resulting from a segmental mega­ colon that may be complicated by sigmoid volvulus (see Chapter 109). Paraneoplastic visceral neuropathy may be associated with malignant tumors outside the gastrointestinal tract, particularly small cell carcinoma of the lung and carcinoid tumors. Pathologic examination of the affected intestine reveals neuronal degeneration or myenteric plexus inflam­

mation.140 An antibody against a component of myenteric neurons has been identified in some patients with this disorder (see Chapter 120).141 Disruption of the ICCs has been associated with a case of small cell lung carcinoma– related paraneoplastic colonic motility disorder.142

Neuropathies of Unknown Cause

Severe acute neuropathies that present mainly with obstruc­ tive symptoms and not principally with constipation have been described. As noted earlier, neuropathic features affecting the colon may occur in some patients with severe idiopathic constipation.

MEDICATIONS Constipation may be a side effect of a drug or preparation taken long term. Drugs commonly implicated are listed in Table 18-4. Common offenders include opioids used for chronic pain, anticholinergic agents including antispas­ modics, calcium supplements, some tricyclic antidepres­ sants, phenothiazines used as long-term neuroleptics, and antimuscarinic drugs used for parkinsonism.

PSYCHOLOGICAL DISORDERS Constipation may be a symptom of a psychiatric disorder or a side effect of its treatment (see Chapter 21). Healthy men who are socially outgoing, energetic, and optimis­ tic—and not anxious—and who described themselves in more favorable terms than others have heavier stools than men without these personality characteristics.143 Psy­ chological factors associated with a prolonged colonic transit time in constipated patients include a highly depressed mood state and frequent control of anger.144 In one study, women with constipation had higher soma­ tization and anxiety scores than healthy controls, and the psycholo­gical scores correlated inversely with rectal mucosal blood flow (used as an index of innervation of the distal colon).145 In a study that assessed psycholo­ gical characteristics of older persons with constipation, a delayed colonic transit time was related significantly to symptoms of somatization, obsessive-compulsiveness, depression, and anxiety.36 In a study of 28 consecutive female patients undergoing psychological assessment for intractable constipation, 60% had evidence of a current affective disorder. One third reported distorted attitudes toward food. Patients with slow-transit constipation reported more psychosocial distress on rating scales than those with normal-transit constipation.146

DEPRESSION

For some patients, constipation can be a somatic manifesta­ tion of an affective disorder. In a study of patients with depression, 27% said that constipation developed or became worse at the onset of the depression.147 Constipation can occur in the absence of other typical features of severe depression, such as anorexia or psychomotor retardation with physical inactivity. Psychological factors are likely to influence intestinal function via autonomic efferent neural pathways.145 In an analysis of 4 million discharge records of U.S. military veterans, major depression was associated with constipation, and schizophrenia was associated with both constipation and megacolon.148

Chapter 18  Constipation EATING DISORDERS

Patients with anorexia nervosa or bulimia often complain of constipation, and a prolonged whole-gut transit time has been demonstrated in patients with these disorders.149 Colonic transit time returns to normal in most patients with anorexia nervosa once they are consuming a balanced diet and gaining weight for at least three weeks.150 Pelvic floor dysfunction is found in some patients with an eating disorder and does not improve with weight gain and a balanced diet.151 Anorexia nervosa should be considered as a possible diag­ nosis in a young underweight woman who presents with constipation. Patients with an eating disorder often resort to the regular use of laxatives as treatment for constipation or to facilitate weight loss or relieve the presumed conse­ quences of binge eating. Treatment of such patients is directed at the underlying eating disorder (see Chapter 8).

DENIED BOWEL MOVEMENTS

Patients may deny or fail to report defecation when solid inert markers have been demonstrated to disappear from the abdomen by radiologic examination, proving that elimina­ tion has occurred. Such patients who deny that defecation has occurred despite evidence to the contrary need skilled psychiatric help.

CLINICAL ASSESSMENT HISTORY

It is important to determine exactly what the patient means when he or she reports constipation. A detailed history that includes the duration of symptoms, frequency of bowel movements, and associated symptoms such as abdominal discomfort and distention should be obtained. The history should include an assessment of stool consistency, stool size, and degree of straining during defecation. The pres­ ence of warning symptoms or signs, such as unintentional weight loss, rectal bleeding, change in the caliber of the stool, severe abdominal pain, and family history of colon cancer, should be elicited. A long duration of symptoms that have been refractory to conservative measures is suggestive of a functional colorectal disorder. By contrast, the new onset of constipation may indicate a structural disease. Physicians should always evaluate the patient for a structural disease in this situation. A dietary history should be obtained. The amount of daily fiber and fluid consumed should be assessed. Many patients tend to skip breakfast,152 and this practice may exacerbate constipation, because the postprandial increase in colonic motility is greatest after breakfast.72,153,154 Although caffein­ ated coffee (150 mg of caffeine) stimulates colonic motility, the ingestion of a meal has a greater effect.155 A patient’s past medical history must be reviewed. Obstet­ ric and surgical histories are particularly important. Neuro­ logic disorders also may explain some cases of constipation. A carefully taken drug history, including the use of overthe-counter laxatives and herbal medications, and their fre­ quency of intake, is important. A detailed social history may provide useful information as to why the patient has sought help for constipation at this point in time; potentially relevant behavioral back­ ground information also may be obtained. In patients with IBS, the frequency of a history of sexual abuse is increased as compared with healthy controls.156 In a survey of 120 patients with dyssynergia, 22% reported a history of sexual abuse, and 32% reported a history of physical abuse. Bowel

dysfunction adversely affected sexual life in 56% and social life in 76% of patients.157 The physician should be alert to manifestations of depression, such as insomnia, lack of energy, loss of interest in life, loss of confidence, and a sense of hopelessness. A history of physical or sexual abuse may not emerge during the initial visit, but if the physician evinces no surprise at whatever is revealed, indicates that distressing events are common in patients with intestinal symptoms, and maintains a sensitive, encouraging attitude, the full story often gradually emerges during subsequent visits, provided that there is privacy, confidentiality, and adequate time (see Chapters 21 and 118).

PHYSICAL EXAMINATION

The patient’s general appearance or voice may point to a clinical diagnosis of hypothyroidism, parkinsonism, or depression. The general physical examination should exclude major central nervous system disorders, especially spinal lesions. If spinal disease is suspected, the sacral der­ matomes should be examined for loss of sensation. The abdomen should be examined for distention, hard feces in a palpable colon, or an inflammatory or neoplastic mass. If the abdomen appears distended, a hand should be passed under the lumbar spine while the patient is lying supine to exclude anterior arching of the lumbar spine as a cause of postural bloating. The rectal examination is paramount in evaluating a patient with constipation. Placing the patient in the left lateral position is most convenient for performing a thor­ ough rectal examination. Painful perianal conditions and rectal mucosal disease should be excluded, and defecatory function should be evaluated. The perineum should be observed both at rest and after the patient strains as if to have a bowel movement. Normally, the perineum descends between 1 and 4 cm during straining. Descent of the perineum with the patient in the left lateral position below the plane of the ischial tuberosities (i.e., >4 cm) usually suggests excessive perineal descent. A lack of descent may indicate the inability to relax the pelvic floor muscles during defecation, whereas excessive perineal descent may indi­ cate descending perineum syndrome. Patients with descend­ ing perineum syndrome strain excessively and achieve only incomplete evacuation because of lack of straightening of the anorectal angle. Excessive laxity or descent of the perineum usually results from previous childbirth or exces­ sive straining. Eventually, excessive descent of the perineum may result in injury to the sacral nerves from stretching, a reduction in rectal sensation, and ultimately incontinence resulting from denervation.91 Rectal prolapse may be detected when the patient is asked to strain. The perianal area should be examined for scars, fistulas, fissures, and external hemorrhoids. A digital rectal exami­ nation should be performed to evaluate the patient for the presence of a fecal impaction, anal stricture, and rectal mass. A patulous anal sphincter may suggest prior trauma to the anal sphincter or a neurologic disorder that impairs sphincter function. Other important functions that should be assessed during the digital examination are summarized in Table 18-7. Specifically, the inability to insert the exam­ ining finger into the anal canal may suggest an elevated anal sphincter pressure, and tenderness on palpation of the pelvic floor as it traverses the posterior aspect of the rectum may suggest pelvic floor spasm. The degree of descent of the perineum during attempts to strain and expel the exam­ ining finger provides another way of assessing the degree of perineal descent. A thorough history and physical examina­ tion can exclude most secondary causes of constipation (see Table 18-4).

271

272

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 18-7  Clinical Clues to an Evacuation Disorder History Prolonged straining to expel stool Assumption of unusual postures on toilet to facilitate stool expulsion Support of perineum, digitation of rectum, or application of pressure to the posterior vaginal wall to facilitate rectal emptying Inability to expel enema fluid Constipation after subtotal colectomy for constipation Rectal Examination (with patient in left lateral position) Inspection Anus “pulled” forward during attempts to simulate strain during defecation Anal verge descends <1 cm or >4 cm (or beyond ischial tuberosities) during attempts to simulate straining at defecation Perineum balloons down during straining; rectal mucosa partially prolapses through anal canal Palpation High anal sphincter tone at rest precludes easy entry of examining finger (in absence of a painful perianal condition such as an anal fissure) Anal sphincter pressure during voluntary squeeze only minimally higher than anal tone at rest Perineum and examining finger descend <1 cm or >4 cm during simulated straining at defecation Puborectalis muscle tender to palpation through rectal wall posteriorly, or palpation reproduces pain Palpable mucosal prolapse during straining “Defect” in anterior wall of the rectum, suggestive of rectocele Anorectal Manometry and Balloon Expulsion (with patient in left lateral position) Average resting anal sphincter tone >80 cm water (>59 mm Hg) Average anal sphincter squeeze pressure >240 cm water (>177 mm Hg) Failure of balloon expulsion from rectum despite addition of 200-g weight to the balloon

DIAGNOSTIC TESTS The high prevalence of bowel symptoms in the population implies that the symptoms are only a nuisance for most people and do not signify serious disease. Therefore, an investigation is not necessary for most patients who com­ plain of one or more of these symptoms, especially adoles­ cents and young adults. Investigations may be indicated for one of two reasons: (1) to exclude a systemic illness or structural disorder of the gastrointestinal tract as a cause of constipation; or (2) to elucidate the underlying pathophysi­ ologic process when the symptoms are unresponsive to simple treatment.

TESTS TO EXCLUDE SYSTEMIC DISEASE

Determination of the hemoglobin level, erythrocyte sedi­ mentation rate, and biochemical screening test levels, including thyroid function, serum calcium, glucose, and other appropriate investigations, are indicated if the clinical picture suggests that the symptoms may result from an inflammatory, neoplastic, metabolic, or other systemic disorder.

TESTS TO EXCLUDE STRUCTURAL DISEASE OF THE INTESTINE

A barium enema study reveals the width and length of the colon and excludes an obstructing lesion severe enough to cause constipation. When fecal impaction is present, a limited enema study with a water-soluble contrast agent

outlines the colon and fecal mass without aggravating the condition. A barium examination of the small bowel is indicated only if obstruction or pseudo-obstruction involv­ ing the small bowel is suspected (see Chapters 119 and 120). Endoscopy allows direct visualization of the colonic mucosa. The yield of colonoscopy in the absence of “alarm” symptoms in patients with chronic constipation is low and is comparable with that for asymptomatic patients who undergo colonoscopy for colon cancer screening. Therefore, a colonoscopy is recommended only when there has been a recent change in bowel habits, blood in stools, or other alarming symptoms (e.g., weight loss, fever).158,159 All adults older than 50 years who present with constipation should undergo a colonoscopy, flexible sigmoidoscopy and barium enema, or computed tomographic colonography to screen for colorectal cancer, as widely recommended. A flexible sigmoidoscopy is probably sufficient for the evaluation of constipation in patients younger than 50 years without alarm symptoms (e.g., weight loss, recent onset of severe constipation, rectal bleeding) or a family history of colon cancer.

PHYSIOLOGIC MEASUREMENTS

Physiologic testing is unnecessary for most patients with constipation and is reserved for patients with refractory symptoms who do not have an identifiable secondary cause of constipation or in whom a trial of a high-fiber diet and laxatives has not been effective. An American Gastroentero­ logical Association Technical Review on Anorectal Testing Techniques160 has recommended the following investiga­ tions in patients with refractory constipation: symptom diaries to establish a diagnosis of constipation and monitor the efficacy of treatment; colonic transit study to confirm the patient’s complaint of constipation and assess colonic motility for slow transit and regional delay; anorectal manometry to exclude Hirschsprung’s disease and to com­ plement other tests of pelvic floor dysfunction; and surface electromyography (EMG) to evaluate anal sphincter func­ tion and facilitate biofeedback training. Tests of possible value include the following: defecation proctography to document the patient’s inability to defecate; balloon expul­ sion test to document the inability to defecate; and rectal sensory testing to help distinguish functional from neuro­ logic disorders as a cause of constipation.

Measurement of Colonic Transit Time

Radiopaque Markers The normal colonic transit time is less than 72 hours. Measurement of colonic transit time is performed only when objective evidence of slow transit is needed to confirm a patient’s history or as a prelude to surgical treatment. Colonic transit time is measured by performing abdominal radiography 120 hours after the patient has ingested radi­ opaque markers in a gelatin capsule (Fig. 18-5). Before the study, patients should be maintained on a high-fiber diet and should avoid laxatives, enemas, or medications that may affect bowel function. Retention of more than 20% of the markers at 120 hours is indicative of prolonged colonic transit. Because the markers are eliminated only with def­ ecation, the process of measuring colonic transit is discon­ tinuous, and the result of a transit measurement should be regarded with caution, taking recent defecation into account. If the markers are retained exclusively in the sigmoid colon and rectum, the patient may have a defecatory disorder. The presence of markers throughout the colon, however, does not exclude the possibility of a defecatory disorder because delayed colonic transit can result from a defecatory disor­ der. Measurements of transit through different segments of

Chapter 18  Constipation during childbirth or an episiotomy.161 Paradoxical anal sphincter contraction is common in patients with a recto­ cele, suggesting that straining and attempts at emptying against a contracted pelvic floor may facilitate development of a rectocele. The limitations of defecography include vari­ ability among radiologists in interpreting studies, inhibition of normal rectal emptying because of patient embarrass­ ment, and differences in texture between barium paste and stool. Confirmatory studies are needed before a decision about management can be made on the basis of the radio­ graphic findings alone. Importantly, identified anatomic abnormalities are not always functionally relevant. For example, a rectocele is only relevant if it fills preferentially (i.e., instead of the rectal ampulla) and fails to empty after simulated defecation. Magnetic resonance defecography may offer advantages over standard barium defecography162 but is not yet widely available.

Figure 18-5.  Colonic transit study; abdominal film. This constipated patient had ingested 20 inert ring markers 120 hours previously and 20 cube-shaped markers 72 hours previously. Most of the markers are still present, indicating slow whole-gut transit.

the colon are of doubtful value in planning treatment, except for megarectum, in which all the markers move rapidly to the rectum and are retained there. Wireless Motility Capsule The wireless pH and pressure recording capsule (SmartPill, Buffalo, NY) is a novel ambulatory technique for assessing colonic transit without radiation. Colonic transit measure­ ments with the wireless capsule technique have been shown to correlate well with the radiopaque marker test and also allow an assessment of gastric and small bowel transit.161 If surgical treatment for severe constipation is being con­ sidered (see later), studies of gastric emptying, small bowel transit, and segmental colonic transit times are valuable for confirming slow transit and correlating abnormalities with therapeutic outcome. In particular, scintigraphic studies of gastrointestinal transit are indicated.50 Generally, abnormal gastric or small bowel motility precludes surgical treatment of constipation.

Tests to Assess the Physiology of Defecation

Defecography Defecography is performed by instilling thickened barium into the rectum. With the patient sitting on a radiolucent commode, films or videos are taken during fluoroscopy with the patient resting, deferring defecation, and straining to defecate. This procedure evaluates the rate and comple­ teness of rectal emptying, anorectal angle, and amount of perineal descent. In addition, defecography can identify structural abnormalities, such as a large rectocele, internal mucosal prolapse, or intussusception. A rectocele repre­ sents a herniation, usually of the anterior rectal wall into the lumen of the vagina, and usually results from trauma

Balloon Expulsion Test When the rectum is distended with a balloon, the internal anal sphincter relaxes. The inability to evacuate a 50- to 60-mL inflated balloon in the rectum163 while sitting on the toilet for two minutes, with the addition of 200 g of weight to the end of the balloon,93 suggests a defecatory disorder. The balloon expulsion test is an effective and useful screen­ ing tool for identifying patients with a defecatory disorder who do not have pelvic floor dyssynergia. In one study of 359 patients with constipation, the balloon expulsion test was abnormal in 21 of 24 patients with pelvic floor dys­ synergia and an additional 12 of 106 patients without pelvic floor dyssynergia. (The diagnosis of pelvic floor dyssynergia was confirmed by manometric and defecographic findings according to the Rome II criteria.164) Anorectal Manometry Anorectal manometry can provide useful information about patients with severe constipation by assessing the resting and maximum squeeze pressure of the anal sphincters, pres­ ence or absence of relaxation of the anal sphincter during balloon distention of the rectum (rectoanal inhibitory reflex), rectal sensation, and ability of the anal sphincter to relax during straining.93,158,165 Patients with a defecatory dis­ order commonly have inappropriate contraction of the anal sphincter when they bear down. The absence of the recto­ anal inhibitory reflex raises the possibility of Hirschsprung’s disease. A high resting anal pressure suggests the presence of an anal fissure or anismus, the paradoxical contraction of the external anal sphincter in response to straining or pressure within the anal canal. Rectal hyposensitivity suggests a neurologic disorder; however, the volume of rectal content needed to induce rectal urgency also may be increased in patients with fecal retention, and the results of rectal sensitivity testing need to be interpreted with caution. Electromyographic Testing of Striated Muscle Activity EMG studies of the external anal sphincter and puborectalis muscles using concentric needle or surface electrode record­ ings generally are not essential and are rarely indicated. An exception is the use of EMG in patients with suspected spinal cord or cauda equina lesions, in whom bilateral or unilateral dysfunction of the external anal sphincter can be demonstrated. Rectal Sensitivity and Sensation Testing Rectal sensitivity to distention can be measured by intro­ ducing successive volumes of air into a rectal balloon and recording the volume at which the stimulus is first per­

273

274

Section III  Symptoms, Signs, and Biopsychosocial Issues ceived, the volume that produces an urge to defecate, and the volume above which further addition of air can no longer be tolerated owing to discomfort. These measure­ ments are not of value in the routine investigation of con­ stipation but are of research interest. The threshold current needed to elicit sensation when the rectal mucosa is stimu­ lated electrically by a current passed between bipolar elec­ trodes can be used as a test of sensory nerve function, but the test has not been established for general use.93

TREATMENT Initial treatment of constipation is based on nonpharmaco­ logic interventions. If these measures fail, then pharmaco­ logic agents may be used. Figure 18-6 provides an algorithm for the evaluation and treatment of patients with severe constipation. If a defecatory disorder is present, initial treat­ ment should include biofeedback; up to 75% of patients with disordered evacuation respond to biofeedback, and many do not respond well to fiber supplementation or oral laxatives. Otherwise, the initial treatment should include increased fluid, exercise, and intake of fiber, either through changes in diet or use of commercial fiber supplements.

Treat secondary causes of constipation (see Table 18-4)

Patients who do not improve with fiber should be given an osmotic laxative, such as milk of magnesia or polyethy­ lene glycol. The dose of the osmotic laxative should be adjusted until soft stools are attained. Stimulant agents, such as bisacodyl or senna derivatives, should be reserved for patients who do not respond to fiber or osmotic laxatives.

GENERAL MEASURES Reassurance

Some people are raised in childhood to believe that a daily bowel movement is essential for health or derive this opinion from advertisements, and they worry if their bowel habit is irregular or less frequent. They can be helped by being told that an irregular bowel habit and other defecatory symptoms are common in the healthy general population and that their symptoms are not harmful. Such reassurance may be all that they need. Patients who are concerned that their symptoms may indicate disease may be helped by appropriate investigation to relieve their fears.

Lifestyle Changes

The need to set aside an unhurried and, if possible, regular time for defecation and always to respond to a defecatory

History and physical examination

Medication history

Stop/change medication(s)

Supplement diet with 20 g fiber/d No response Colonic transit study

Normal transit

Slow transit

Fiber (>20 g/d), osmotic laxative, stimulant laxative

Assess for defecatory disorder (e.g., anorectal manometry, balloon expulsion test)

No response Assess for defecatory disorder (e.g., anorectal manometry, balloon expulsion test) Normal Treat as for irritable bowel syndrome

Abnormal Abnormal

Normal Defecography

Evacuation disorder

Rectal anatomic defect

Slow-transit constipation

Repair of prolapse Osmotic laxative, stimulant laxative, or rectocele Biofeedback, physical prokinetic agent; therapy, consultation rarely colectomy with psychologist and/or dietitian Figure 18-6.  Algorithm for the evaluation and treatment of severe constipation.

Chapter 18  Constipation urge should be stressed. If patients experience difficulty in expulsion of stool, they should be advised to place a support approximately 6 inches in height under their feet when sitting on a toilet seat so that the hips are flexed toward a squatting posture. For persons with an inactive lifestyle, activity should be encouraged. The use of constipating drugs should be avoided.

Psychological Support

Constipation may be aggravated by stress or may be a mani­ festation of emotional disturbance (e.g., previous sexual abuse; see Chapter 21). For such patients, an assessment of the person’s circumstances, personality, and background and supportive advice may help more than any physical measures of treatment. Behavioral treatment (see later) offers a physical approach with a psychological component and is often acceptable and beneficial. Psychological treat­ ment is needed only when it would be indicated in any circumstance, not specifically for constipation.

Fluid Intake

Dehydration or salt depletion is likely to lead to increased salt and water absorption by the colon, leading in turn to the passage of small hard stools. Although dehydration is generally accepted as a risk factor for constipation, unless a person is clinically dehydrated, no data support the notion that increasing fluid intake improves constipation.166

Dietary Changes and Fiber Supplementation

After studying the dietary and stool patterns of rural Afri­ cans in the early 1970s, Burkitt and colleagues speculated that a deficiency in dietary fiber was contributing to consti­ pation and other colonic diseases in Western societies.167 Since then, studies have shown that when nonconstipated persons increase their intake of dietary fiber, stool weight increases in proportion to their baseline stool weight and frequency of defecation and correlates with a decrease in colonic transit time.168 Every gram of wheat fiber ingested yields approximately 2.7 g of stool expelled. It follows that when an increased intake of dietary fiber leads to an increase in stool weight in constipated subjects who pass small stools, the resulting stool weight may still be lower than normal. For this reason, the therapeutic results of a high-

fiber diet are often disappointing as a treatment for consti­ pation. In a study of 10 constipated women who took a supplement of wheat bran, 20 g/day, average daily stool weight increased from approximately 30 to 60 g/day, with only half of patients achieving a normal average stool weight. Bowel frequency increased from a mean of two to three bowel movements weekly.169 In a controlled, crossover trial, 24 patients took 20 g of bran or placebo daily for four weeks. Although bran was more effective than placebo in improving bowel frequency and oroanal transit rate, the occurrence and severity of constipation experienced by the patients did not differ between the two treatment periods.170 This result probably reflects the observation that patients complain mainly of difficulty in defecation, rather than a decreased frequency of bowel movements. In a series of constipated patients, about half were reported to have gained some benefit from a bran supplement of 20 g daily.171 Dietary fiber appears to be effective in relieving mild to moderate43 but not severe constipation,69 especially if severe constipation is associated with slow colonic transit, evacu­ ation disorders, or medications. Although dietary modifica­ tion may not succeed, all constipated subjects should be advised initially to increase their dietary fiber intake as the simplest, most physiologic, and cheapest form of treatment. Patients should be encouraged to take about 25 g of NSPs daily by eating whole-wheat bread, unrefined cereals, plenty of fruit and vegetables and, if necessary, a supplement of raw bran, either in breakfast cereals or with cooked foods. Specific dietary counseling often is needed to achieve a satisfactory increase in dietary fiber. Because of side effects, adherence with fiber supplemen­ tation is poor, especially during the first several weeks of therapy. Side effects include abdominal distention, bloat­ ing, flatulence, and poor taste. Most controlled studies of the effect of fiber have shown that the minimum supplemen­ tation needed to consistently alter bowel function or colonic transit time significantly is 12 g/day. To improve adherence, patients should be instructed to increase their dietary fiber intake gradually over several weeks to approximately 20 to 25 g/day. If results of therapy are disappointing, commer­ cially packaged fiber supplements should be tried (Table 18-8). Fiber and bulking agents are concentrated forms of

Table 18-8  Commercial Fiber Products AGENT

starting DAILY DOSe (g)

Methylcellulose

4-6

Psyllium

4-6

Polycarbophil

4-6

Guar gum

3-6

COMMENTS Semisynthetic cellulose fiber that is relatively resistant to colonic bacterial degradation and tends to cause less bloating and flatus than psyllium Made from ground seed husk of the ispaghula plant; forms a gel when mixed with water, so an ample amount of water should be taken with psyllium to avoid intestinal obstruction; undergoes bacterial degradation, which may contribute to side effects of bloating and flatus; allergic reactions such as anaphylaxis and asthma have been reported but are rare Synthetic fiber made of polymer of acrylic acid, which is resistant to bacterial degradation Soluble fiber extracted from seeds of the leguminous shrub Cyamopsis tetragonoloba

275

276

Section III  Symptoms, Signs, and Biopsychosocial Issues NSPs based on wheat, plant seed mucilage (ispaghula), plant gums (sterculia), or synthetic methylcellulose deriva­ tives (methylcellulose, carboxymethylcellulose; see later). Some patients, particularly women with markedly delayed colonic transit, find that fiber aggravates abdominal distention. Bran also may be unhelpful in young people with megacolon and in older subjects, in whom it may lead to fecal incontinence. For these patients, a reduction in fiber intake may relieve symptoms.

solution. Guar gum is approved for use in a number of foods and cosmetics and as a supplement. When used in high doses, guar gum has been reported to cause intestinal obstruction.

SPECIFIC THERAPEUTIC AGENTS Commercial Fiber Products

Poorly Absorbed Ions Magnesium, sulfate, and phosphate ions are poorly absorbed by the gut and thereby create a hyperosmolar intraluminal environment. Their primary mode of action appears to be osmotic, but they may have other possible effects with unclear consequences, such as increasing prostaglandin concentrations in the stool.176 In mildly constipated patients, regular use of magnesium hydroxide is a useful and safe laxative. Stool weight increases by 7.3 g for each additional millimole of soluble magnesium excreted.177 Standard doses of magnesium hydroxide (see Table 18-9) contain 40 to 80 mmol of magnesium ion and typically produce a bowel movement within 6 hours. Magnesium sulfate is a more potent laxative that tends to produce a large volume of liquid stool. Patients may complain about this compound because it often leads to abdominal distention and the sudden passage of a liquid foul-smelling stool. Use of mag­ nesium in older adults is limited by adverse effects such as flatulence, abdominal cramps, and magnesium toxicity. A small percentage of magnesium is actively absorbed in the small intestine; the remainder draws water into the intestine along an osmotic gradient.178 Hypermagnesemia can occur in patients with renal failure and in children. Hypermagnesemia-induced paralytic ileus is a rare compli­ cation,179 and hypermagnesemia with coma has occurred in a normal 6-week-old infant given 16 2-mL doses of milk of magnesia.180 Severe toxicity with coma also has occurred in a chronically constipated child given an enema containing 32.5 mg of magnesium sulfate.181 Patients with renal insuf­ ficiency or cardiac dysfunction can experience electrolyte and volume overload from the absorption of magnesium or phosphorus. Even patients who are otherwise healthy can experience these complications, in addition to dehydration, as a result of excessive use. Phosphate can be absorbed by the small intestine, and a substantial dose must be ingested to produce an osmotic laxative effect. One commercial preparation, Fleet Phospho Soda, contains 48 g (400 mmol) of monobasic sodium phosphate and 18 g (130 mmol) of dibasic sodium phosphate/100 mL, resulting in a hypertonic solution. Hyperphosphatemia can occur, especially in patients with renal insufficiency. In addition, a rare but serious form of acute kidney injury has been associated with sodium phos­ phate solution used before colonoscopy, even in patients with normal baseline renal function. Risk factors include hypertension, advanced age, volume depletion, and use of angiotensin-converting enzyme inhibitors or nonsteroidal anti-inflammatory drugs.182,183 The preparation is no longer available over the counter in the United States.

Methylcellulose Methylcellulose is a semisynthetic NSP of varying chain length and degree of methylation. Methylation reduces bac­ terial degradation in the colon. One study of constipated patients with an average daily fecal weight of only 35 g showed an increase in fecal solids with 1, 2, and 4 g of methylcellulose/day, but fecal water increased only with the 4-g dose. Bowel frequency in this group of patients increased from an average of two to four stools weekly, but the patients did not report marked improvement in consis­ tency or ease of passage of stools (see Table 18-8).172 Ispaghula (Psyllium) Ispaghula is derived from the husks of an Asian plant, has high water-binding capacity, is fermented in the colon to a moderate extent, and increases bacterial cell mass. It is available as effervescent suspensions, granules, and a powder. The suspensions, which are popular, need to be drunk quickly before the husk absorbs water. The granules may be stirred briskly in a half-glass of water and swallowed at once; carbonated water may be preferred. Some people like to swallow the solid granules and then drink a glass of water. Ispaghula (3.4 g as Metamucil) has been shown to increase fecal bulk to the same extent as methylcellulose 1 to 4 g daily in constipated subjects. Although both stool dry and wet weights increased, the total weekly weights remained less than those of a healthy control group without treatment. In an observational study, 149 patients were treated with psyllium in the form of Plantago ovata seeds, 15 to 30 g daily, for a period of at least 6 weeks. The response to treat­ ment was poor among patients with slow colonic transit or a disorder of defecation, whereas 85% of patients without abnormal physiologic testing results improved or became symptom-free. Nevertheless, the authors recommend that a trial of dietary fiber be undertaken before diagnostic testing is performed.69 Ispaghula taken by mouth can cause an acute allergic immunoglobulin E–mediated response, with facial swell­ ing, urticaria, tightness in the throat, cough, and asthma.173 Workers who inhale the compound during manufacture or preparation can have a similar reaction174 (see Table 18-8). Calcium Polycarbophil Calcium polycarbophil is a hydrophilic polyacrylic resin that is resistant to bacterial degradation and thus may be less likely to cause gas and bloating. In patients with IBS, calcium polycarbophil appears to improve global symptoms and ease of stool passage175 but not abdominal pain (see Table 18-8). Guar Gum Guar gum is a natural high molecular weight polysaccharide extracted from the seed of the leguminous shrub Cyamopsis tetragonoloba. It hydrates rapidly to form a highly viscous

Other Laxatives

The main groups of laxatives other than fiber are osmotic agents and stimulatory laxatives; stool softeners and emol­ lients are additional therapeutic agents (see later) (Tables 18-9 and 18-10).

Poorly Absorbed Sugars Lactulose.  Lactulose is a nonabsorbable synthetic disaccha­ ride that consists of galactose and fructose linked by a bond resistant to lactase. Therefore, lactulose is not absorbed by the small intestine but undergoes fermentation in the colon to yield short-chain fatty acids, hydrogen, and carbon dioxide, with consequent lowering of the fecal pH. When

Chapter 18  Constipation Table 18-9  Laxatives Commonly Used for Constipation TYPE OF LAXATIVE

GENERIC NAME(S)

DOSE

COMMENTS

Osmotic Laxatives Poorly Absorbed Ions Magnesium

Magnesium hydroxide

15-30 mL once or twice daily

Sulfate

Magnesium citrate Magnesium sulfate Sodium sulfate

150-300 mL every day 15 g every day 5-10 g every day

Hypermagnesemia can occur in patients with renal failure and in children.

Phosphate

Sodium phosphate

0.5-10 mL with 12 oz of water

Lactulose Sorbitol Mannitol

15-30 mL once or twice daily 15-30 mL once or twice daily 15-30 mL once or twice daily

Polyethylene glycol electrolyte

17-34 g once or twice daily

Cascara sagrada Senna

325 mg (or 5 mL) at bedtime 1-2 7.5-mg tablets daily

Castor oil Bisacodyl Phenolphthalein

15-30 mL at bedtime 5-10 mg at bedtime 30-200 mg at bedtime

Sodium picosulfate Docusate sodium Mineral oil

5-15 mg at bedtime 100 mg twice daily 5-15 mL at bedtime

Enemas, Suppositories

Phosphate enema Mineral oil retention enema Tap water enema Soapsuds enema Glycerin suppository Bisacodyl suppository

120 mL 100 mL 500 mL 1500 mL 60 g 10 mg

Chloride Channel Activator

Lubiprostone

8-24 µg twice daily

Poorly Absorbed Sugars Disaccharides Sugar alcohols

Polyethylene glycol

Stimulant Laxatives Anthraquinones

Ricinoleic acid Diphenylmethane Derivatives

Stool Softeners Emollients

normal subjects take lactulose 20 g (30 mL) daily, none of the sugar is detectable in the stool. In larger doses, some of the sugar passes though the colon unchanged and acts as an osmotic laxative. The recommended dose of lactulose for adults is 15 to 30 mL once or twice daily. The time to onset of action is longer than that for other osmotic laxatives, and two or three days are required for lactulose to achieve an effect. Some patients report that lactulose is effective initially but then loses its effect, perhaps because the intestinal flora are altered in response to the medication.184 Adverse effects related to lactulose include abdominal distention or dis­ comfort, presumably as a result of colonic gas production. Cases of lactulose-induced megacolon have been reported. In a group of young, chronically constipated volunteers who reported fewer than three stools a week, lactulose

Sulfate is generally not used by itself as a laxative agent. Hyperphosphatemia can occur, especially in patients with renal failure. Gas and bloating are common side effects. Sorbitol is commonly used as a sweetener in sugar-free products. In older adults, sorbitol has an effect similar to that of lactulose but has a lower cost. Tends to cause less bloating and cramps than other agents; tasteless and odorless, can be mixed with noncarbonated beverages. Typically used to prepare colon for diagnostic examinations and surgery; also available as powder without electrolytes for regular use (MiraLax) Cause apoptosis of colonic epithelial cells phagocytosed by macrophages; result in lipofuscin-like pigmented condition known as pseudomelanosis coli; no definitive association established between anthraquinones and colon cancer or myenteric nerve damage (cathartic colon) Cramping is common. Has effects in small intestine and colon Removed from U.S. market because of teratogenicity in animals Likely has effects only on colon Efficacy in constipation not well established. Long-term use can cause malabsorption of fat-soluble vitamins, anal seepage, and lipoid pneumonia in patients predisposed to aspiration of liquids. Serious damage to rectal mucosa can result from extravasation of enema solution into the submucosa; hypertonic phosphate enemas and large-volume water or soapsuds enemas can lead to hyperphosphatemia and other electrolyte abnormalities if enema is retained; soapsuds enemas can cause colitis. Increases secretion in the intestines

increased bowel frequency and percentage of stool moisture and softened the stools when compared with a control syrup that contained only sucrose. The effectiveness of lactulose was dose-dependent.185 The effect of lactulose in older patients has been studied in two double-blind, placebo-controlled trials. In one trial, only about half of patients were found to be truly consti­ pated, and, among these patients, lactulose was effective in 80%, as compared with 33% of those who received placebo (glucose) (P < 0.01).186 The second trial was conducted in a nursing home over 8 to 12 weeks in 42 older patients with constipation.187 The initial dose of lactulose was 30 mL/day, and the dose was reduced temporarily or permanently to 15 mL, depending on bowel frequency. Lactulose showed an advantage over placebo (a 50% glucose syrup) by increas­ ing the mean number of bowel movements each day and

277

278

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 18-10  Grade of Evidence for the Use of Laxatives According to the American College of Gastroenterology Task Force on Chronic Constipation LAXATIVE Bulking agents   Psyllium   Calcium polycarbophil   Bran Stool softeners Lubricants Osmotic laxatives   PEG   Lactulose   Milk of magnesia Stimulant laxatives Prokinetic agent   Tegaserod‡ Chloride channel activator   Lubiprostone

GRADE of evidence* B B †

B C A A †

B A §

*Grade A: Based on two or more randomized controlled trials (RCTs) with adequate sample sizes and appropriate methodology. Grade B: Based on evidence from a single RCT of high quality or conflicting results from high-quality RCTs or two or more RCTs of lesser quality. Grade C: Based on noncontrolled trials or case reports. † Insufficient data. ‡ Removed from the U.S. market. § Not yet graded. PEG, polyethylene glycol. Data from Brandt LJ, Prather CM, Quigley EM, et al. Systematic review on the management of chronic constipation in North America. Am J Gastroenterol 2005; 100(Suppl 1):S5-21.

markedly reducing episodes of fecal impaction (P < 0.015) and the need for enemas. Sorbitol.  Sorbitol is used widely in the food industry as an artificial sweetener. Ingestion of as little as 5 g causes a rise in breath hydrogen, and 20 g produces diarrhea in about half of normal subjects.188 Sorbitol is as effective as lactulose and less expensive. A randomized, doubleblind, crossover trial of lactulose, 20 g/day, and sorbitol, 21 g/day, in ambulant older men with chronic constipation showed no difference between the two compounds with regard to frequency or normality of bowel movements or patient preference.189 The frequency of side effects was similar except for nausea, which was more common with lactulose. Mannitol is another sugar alcohol that can be used as a laxative. Polyethylene Glycol Polyethylene glycol (PEG) has emerged as a safe and effec­ tive treatment for chronic constipation. It is an isosmotic laxative that is metabolically inert and able to bind water molecules, thereby increasing intraluminal water retention.190 PEG is not metabolized by colonic bacteria. Solutions containing PEG and electrolytes typically are administered orally to prepare the colon for diagnostic examinations or surgery. Ingestion of PEG leads to an increase in stool volume and softer stools, which may become liquid depending on the volume of PEG consumed. PEG is excreted unchanged in the feces. Electrolytes are added to PEG solutions used for colonic lavage before co­ lonoscopy to avoid the potential adverse effects associated with drinking large volumes of a fluid, such as dehydration and electrolyte imbalance. PEG (with electrolytes) is also available as a powder that is mixed in smaller doses with water for regular use to treat constipation.

Several studies have demonstrated the efficacy of PEG in the treatment of chronic constipation. In a trial in which 70 ambulatory outpatients were treated for four weeks with a PEG-electrolyte solution, 250 mL once or twice daily, at the end of the four weeks, bowel frequency had increased to normal, hard stools were uncommon, and straining on def­ ecation was experienced by fewer than 20% of patients, compared with 80% before treatment. The patients were then randomized to continue PEG or a placebo for 20 weeks in a dose of one or two packets daily, as determined to give the best result. In every parameter examined, the active treatment gave significantly improved results over placebo without adverse clinical or laboratory events. At the end of follow-up, complete remission of constipation was reported by 77% and 20% of patients randomized to PEG and placebo, respectively. The dropout rate of 46% in the placebo group, mostly secondary to treatment failure, was notable.191 In another randomized multicenter trial that compared standard and maximum doses of two PEG formulations of different molecular weights, PEG 3350 and PEG 4000, in 266 outpatients, most patients had their first stool within one day of initiating PEG treatment, and stool consistency improved in both treatment groups. The lowest dose of PEG produced the most normal stool consistency, whereas higher doses produced more liquid stools.192 Low-dose PEG appears to be more effective than lactulose in the treatment of chronic constipation.193,194 A study of 307 patients with chronic constipation who were randomized to 17 g of PEG or placebo for six months showed continued benefit of PEG compared with placebo and no electrolyte abnormalities or intestinal malabsorption.194 PEG is approved by the FDA for children, and a doseranging trial in children with constipation, ages six months to 15 years, has suggested that PEG is a potentially useful treatment, provided that the dose is adjusted to the child’s age.195 PEG solutions may be useful for the short-term treat­ ment of fecal impaction. In one study,196 16 severely ill patients, ages 26 to 87 years, who, despite treatment with various laxatives, had not had a bowel movement in the hospital for 5 to 23 days, were treated with PEG. All had a fecal impaction on clinical examination. They were advised to drink 1 L of a PEG-electrolyte solution, taken as two portions of 500 mL, each over 4 to 6 hours. The regimen was repeated on a second and third day, if neces­ sary. The full dose was taken by 12 patients on the first day, and the remainder took at least one half of the recommended dose; only 8 patients needed treatment on the second day and 2 patients on the third day. The treatment was highly effective and, after the last dose, most patients were passing moderate or large volumes of soft stool, with resolution of impaction. No adverse side effects, apart from abdominal rumbling, occurred, and only 1 patient, who was paraplegic, experienced fecal incontinence. Successful treatment with PEG has been described in outpatients with refractory constipation, older adults (with administration of PEG by mouth or by a nasogastric tube), and children with fecal impaction,197 although children have had difficulty drinking the large volume of fluid. The most common adverse effects of PEG include abdomi­ nal bloating and cramps.190 The medication is generally well tolerated, but cases of fulminant pulmonary edema have been reported after administration of PEG solution by naso­ gastric tube, with one fatality.198,199 In each case, the patient had emesis, suggesting aspiration of PEG. PEG also may delay gastric emptying.200

Chapter 18  Constipation Stimulant Laxatives Stimulant laxatives increase intestinal motility and intesti­ nal secretion. They begin working within hours and often are associated with abdominal cramps. Stimulant laxatives include anthraquinones (e.g., cascara, aloe, senna) and diphenylmethanes (e.g., bisacodyl, sodium picosulfate, phe­ nolphthalein). Castor oil is used less commonly because of its side effect profile. The effect of stimulant laxatives is dose-dependent. Low doses prevent absorption of water and sodium, whereas high doses stimulate secretion of sodium, followed by water, into the colonic lumen. Stimulant laxatives sometimes are abused, especially in patients with an eating disorder,201 even though at high doses they have only a modest effect on calorie absorption. Although a cathartic colon (i.e., a colon with reduced motil­ ity) has been attributed to prolonged use of stimulant laxa­ tives, no animal or human data support this effect. Rather, cathartic colon, as seen on a barium enema examination, is probably a primary motility disorder. Overall, stimulant laxatives are well tolerated if used in doses that produce normal, soft, formed stools. They act rapidly and are particularly suitable for use in a single dose for temporary constipation. Most clinicians are cautious about recommending indefinite daily dosing of stimulant laxatives for chronic constipation. Large doses produce abdominal cramping and liquid stools. Stimulant laxatives vary widely in clinical effectiveness, and some patients with severe constipation are not helped by stimulant laxatives. Anthraquinones.  Anthraquinones, such as cascara, senna, aloe, and frangula, are produced by a variety of plants. The compounds are inactive glycosides that when ingested, pass unabsorbed and unchanged down the small intestine and are hydrolyzed by colonic bacterial glycosi­ dases to yield active molecules. These active metabolites increase the transport of electrolytes into the colonic lumen and stimulate myenteric plexuses to increase intestinal motility. The anthraquinones typically induce defecation six to eight hours after oral dosing. Anthraquinones cause apoptosis of colonic epithelial cells, which then are phagocytosed by macrophages and appear as a lipofuscin-like pigment that darkens the colonic mucosa, a condition termed pseudomelanosis coli202 (see Chapter 124 and Figures 124-7 and 124-8). Whether anthra­ quinone laxatives given over the long term cause adverse functional or structural changes in the intestine is contro­ versial. Animal studies have shown neither damage to the myenteric plexus after long-term administration of senno­ sides203 nor a functional defect in motility.204 A case-control study in which multiple colonic mucosal biopsy specimens were examined by electron microscopy showed no dif­ ferences in the submucosal plexuses between patients taking an anthraquinone laxative regularly for one year and those not taking one.205 An association between use of anthraquinones and colon cancer or myenteric nerve damage and the development of cathartic colon has not been established.206 Senna has been shown in controlled trials to soften stools207 and to increase the frequency and wet and dry weights of stool. The formulations available for clinical use vary from crude vegetable preparations to purified and stan­ dardized extracts to a synthetic compound. Castor Oil.  Castor oil comes from the castor bean. After oral ingestion, it is hydrolyzed by lipase in the small intes­ tine to ricinoleic acid, which inhibits intestinal water absorption and stimulates intestinal motor function by dam­ aging mucosal cells and releasing neurotransmitters.206 Cramping is a common side effect.

Diphenylmethane Derivatives.  Diphenylmethane com­ pounds include bisacodyl, sodium picosulfate, and phenol­ phthalein. After oral ingestion, bisacodyl and sodium picosulfate are hydrolyzed to the same active metabolite, but the mode of hydrolysis differs. Bisacodyl is hydrolyzed by intestinal enzymes and thus can act in the small and large intestines. Sodium picosulfate is hydrolyzed by colonic bacteria. Like anthraquinones, the action of sodium picosulfate is confined to the colon, and its activity is unpre­ dictable because its activation depends on the bacterial flora. The effects of bisacodyl, and presumably sodium picosul­ fate, on the colon are similar to those of the anthraquinone laxatives. When applied to the colonic mucosa, bisacodyl induces an almost immediate, powerful, propulsive motor activity in healthy and constipated subjects, although the effect is sometimes reduced in the latter.208 The drugs also stimulate colonic secretion. Like the anthraquinone laxatives, bisacodyl leads to apop­ tosis of colonic epithelial cells, the remnants of which accu­ mulate in phagocytic macrophages, but these cellular remnants are not pigmented.209 Aside from these changes, bisacodyl does not appear to cause adverse effects with long-term use.210 Bisacodyl is a useful and predictable laxative, especially suitable for single-dose use in patients with temporary constipation. Its possible effect on the small bowel is a disadvantage, in contrast to anthraquinones and sodium picosulfate. Long-term use of bisacodyl or related agents is sometimes necessary for patients with chronic severe constipation. In the doses used, liquid stools and cramps tend to result, and it is difficult to adjust the dose to produce soft, formed stools. Phenolphthalein inhibits water absorption in the small intestine and colon by effects on eicosanoids and the Na+/K+ATPase pump present on the surface of enterocytes. The drug undergoes enterohepatic circulation, which may prolong its effects. It has been removed from the U.S. market because it is teratogenic in animals.

Stool Softeners and Emollients

Docusate Sodium Although the detergent dioctyl sodium sulfosuccinate (docusate sodium) is available as a stool softener, further study of its efficacy is needed. The compound stimulates fluid secretion by the small and large intestines but does not increase the volume of ileostomy output or the weight of stools in normal subjects.211,212 A double-blind, crossover trial has shown benefit in 5 of 15 older constipated subjects, as judged by patients and their caregivers, and a significant increase in bowel frequency.213 In a multicenter, doubleblind, randomized trial in adults, docusate sodium was less effective than psyllium for the treatment of chronic idio­ pathic constipation.214 Mineral Oils Mineral oils alter the stool by undergoing emulsification into the stool mass and providing lubrication for the passage of stool. Long-term use can cause intestinal malabsorption of fat-soluble vitamins, anal seepage, and lipoid pneumonia in patients predisposed to aspiration of liquids.

Enemas and Suppositories

Compounds may be introduced into the rectum to stimulate contraction by distention or chemical action, soften hard stools, or both. Serious damage to the rectal mucosa can result from extravasation of the enema solution into the submucosal plane. The anterior rectal mucosa is the site

279

280

Section III  Symptoms, Signs, and Biopsychosocial Issues most vulnerable to trauma from the tip of a catheter intro­ duced through the backward-angulated anal canal (see Chapter 125). The enema nozzle should be directed poste­ riorly after the anal canal has been passed. Phosphate Enemas Hypertonic sodium phosphate enemas are often effective. They cause distention and stimulation of the rectum. A histologic study in normal subjects showed that a single hypertonic phosphate enema caused disruption of the surface epithelium in 17 of 21 biopsy specimens. Scanning electron microscopy showed patchy denudation of the surface epithelium, with exposure of the lamina propria and absence of goblet cells. The proctoscopic appearance of the mucosa was abnormal in every case but returned to normal within one week.215 Therefore, superficially damaged mucosa appears to heal rapidly. Phosphate enemas are used widely, but no conclusive evidence supports their use. Phosphate enemas, if given to a patient who cannot evac­ uate it promptly, can lead to dangerous hyperphosphatemia and hypocalcemic tetany; one patient, age 91 years, died after a single phosphate enema,216 and coma developed in an adult who was given six phosphate enemas at hourly intervals without evacuation.217 Severe hyperphosphatemia, hypocalcemia, and seizure have been reported in a 4-yearold child with normal renal function after retention of two phosphate enemas.218 The use of phosphate enemas in children 3 years and younger is not recommended.219,220 Saline, Tap Water, and Soapsuds Enemas Saline, tap water, or soapsuds enemas can be effective mainly by distending the rectum and softening feces. Stool evacuation typically occurs two to five minutes following administration. A saline enema does no damage to the rectal mucosa and may be effective.215 Water enemas and soapsuds enemas also may be used, but with large volumes, danger­ ous water intoxication can occur if the enema is retained. Large-volume water or soapsuds enemas also can lead to hyperphosphatemia and other electrolyte disturbances if the enema is retained. Soapsuds enemas can cause rectal mucosal damage and necrosis. Stimulant Suppositories and Enemas Glycerin can be administered as a suppository and is often clinically effective. The rectum is stimulated by an osmotic effect. The effect of glycerin, if any, on the rectal mucosa is unknown. Bisacodyl, 10 mg, is available as a suppository that appears to act topically by stimulating enteric neurons.182 In normal subjects, a single bisacodyl suppository or an enema containing 19 mg of bisacodyl in 100 or 200 mL of water produced marked changes in 23 of 25 rectal mucosal biopsy specimens. The epithelium of the surface and within the crypt was altered; with use of the enema, the surface epithelium was absent.215 The regular use of bisacodyl sup­ positories thus appears unwise. Oxyphenisatin (Veripaque), which is no longer available in the United States, is a stimu­ lant enema that was used in the past mainly before diagnos­ tic procedures. When given by mouth, this compound led to some cases of chronic hepatitis.

Chloride Channel Activator

Lubiprostone is a novel bicyclic fatty acid that activates the chloride 2 channel, thereby increasing intestinal fluid secre­ tion and transit221 without altering serum electrolyte levels. In two phase III randomized, placebo-controlled trials, lubiprostone, 24 µg twice daily, increased the number of spontaneous bowel movements in patients with chronic constipation as defined by the Rome II criteria. Lubipros­

tone also significantly decreased straining, improved stool consistency, and reduced overall severity of symptoms. The frequency of spontaneous bowel movements increased in men and women, as well as older patients, who took the drug. A rebound effect after withdrawal of the drug was not evident.222 The most common side effects were nausea, headache, and diarrhea. Lubiprostone, 24 µg twice daily, was approved by the FDA in 2006 for the treatment of men and women with chronic constipation and in 2008, in a dose of 8 µg twice daily, for women who have IBS with constipation.

Prokinetic Agents

Prokinetic agents induce increased contractility in a segment of the gastrointestinal tract. Stimulation of the 5hydroxytryptamine4 (5-HT4) receptor on afferent nerves in the wall of the gastrointestinal tract induces peristaltic con­ traction of the intestine. Several 5-HT4 agonists have been tested for the treatment of constipation. Cisapride, a benzo­ diazepine, has had variable results in treating constipa­ tion.223 Potentially lethal cardiac dysrhythmias led to its withdrawal from the commercial market in the United States in July 2000, although it still remains available through a limited access program. Newer 5-HT4 agonists such as prucalopride and TD-5108 appear promising as future treatments for chronic constipation. Tegaserod Tegaserod, a partial 5-HT4 agonist, is an aminoguanidine indole derivative of serotonin that is structurally different from cisapride. Because of cardiovascular safety concerns, tegaserod was withdrawn from the market in April 2007. The frequency of cardiovascular events in previous clinical trials was 13 in 13,614 (0.11%) compared with 1 in 7031 (0.01%) in control subjects. The cardiovascular events reported were myocardial infarction (n = 3), sudden cardiac death (n = 1), unstable angina (n = 6), and stroke (n = 3). The decision of the U.S. Food and Drug Administration (FDA) to withdraw the drug has been the subject of debate.224 In a randomized, double-blind, placebo-controlled trial, 1348 subjects with chronic constipation were randomized to receive tegaserod 2 mg twice daily, tegaserod 6 mg twice daily, or placebo for 12 weeks. Response was defined as an increase in the number of complete spontaneous bowel movements (CSBMs) of one or more/week as compared with baseline. During the first 4 weeks of treatment, response rates were 41.4%, 43.2%, and 25.1% for tegaserod 2 mg twice daily, tegaserod 6 mg twice daily, and placebo recipi­ ents, respectively. The effect was maintained over 12 weeks. The median time to first CSBM was significantly shorter in patients treated with tegaserod 6 mg twice daily (73 ± 45 hours) and 2 mg twice daily (117 ± 66 hours) than in those treated with placebo (229 ± 123 hours). Tegaserod also improved the frequency and consistency of stools and reduced straining. No rebound effect was seen after with­ drawal of tegaserod. Diarrhea was more common with tega­ serod 2 and 6 mg twice daily (4.5% and 7.3%, respectively) than with placebo (3.8%). Most cases of diarrhea occurred within the first week of therapy and lasted a median of 2.0 days. Treatment with tegaserod did not result in any cases of electrolyte imbalance.225 Tegaserod had also been used in women with constipation-predominant IBS (see Chapter 118).226 Prucalopride Prucalopride, a full 5-HT4 agonist, is a benzofuran deriva­ tive that induces strong contractions in the proximal colon in dogs and accelerates colonic transit in healthy humans

Chapter 18  Constipation and in patients with functional constipation.227 Three large, 12-week, randomized, placebo-controlled phase III trials of similar design that evaluated the efficacy and safety of prucalopride 2 mg or 4 mg once daily versus placebo in patients with chronic constipation have been published.228-230 Patients enrolled in these studies were required to have at least two CSBMs/week, in combination with straining, a sensation of incomplete evacuation, or hard stools, at least 25% of the time. In one of these studies, the percentage of patients achieving more than three CSBMs/week was 30.9% for those receiving prucalopride 2 mg and 28.4% for those receiving prucalopride 4 mg, compared with 12.0% in the placebo group (P < 0.001 for both comparisons). All other secondary efficacy endpoints, including patients’ satisfac­ tion with their bowel function and treatment and their per­ ception of the severity of their constipation symptoms, were improved significantly at week 12 with the use of 2 or 4 mg of prucalopride as compared with placebo. When the results of the phase III studies (N = 1924) were combined, the per­ centage of patients with an average of at least three CSBMs/ week over the 12-week treatment period was 23.6%, 24.7%, and 11.3% for prucalopride 2 mg, prucalopride 4 mg, and placebo, respectively (P < 0.005). The most frequent adverse effects were headaches, nausea, and diarrhea. No cardio­ vascular side effects were observed, nor were any electro­ cardiographic abnormalities reported. TD-5108 TD-5108 is also a full 5-HT4 agonist. In a four-week phase II trial, 401 patients with chronic constipation (less than three CSBMs/week during a two-week baseline period) were randomized to receive TD-5108, 15, 30, or 50 mg or placebo, once daily. The percentages of patients achieving three or more CSBMs for all four weeks were 27%, 19%, 21%, and 5%, respectively (P = 0.0006 for all TD-5108 groups com­ pared with placebo).231

Peripheral Mu-Opioid Antagonists

Peripherally acting opioid antagonists have been shown to reverse opioid-induced bowel dysfunction without revers­ ing analgesia or precipitating central nervous system with­ drawal signs. Preclinical data suggest that these agents may be effective in the treatment of constipation. Methylnaltrex­ one is a peripherally acting mu-opioid receptor antagonist that was approved by the FDA in 2008 for the treatment of opioid-induced constipation in patients with a late-stage, advanced illness who are receiving an opioid on a continu­ ous basis to relieve pain. Approval was based on two phase III trials. One of the studies involved 133 patients with a life expectancy of less than six months and fewer than three bowel movements in the week prior to treatment or no bowel movements for more than two days. Patients were randomized to receive methylnaltrexone, 0.15 mg/kg sub­ cutaneously, or placebo every other day for two weeks, followed by a three-month open-label treatment period. In this study, 47% of patients reported having a bowel move­ ment within four hours of starting methylnaltrexone com­ pared with 15% of those who received placebo (P < 0.001). Methylnaltrexone did not appear to precipitate opioid with­ drawal symptoms or affect central analgesia.232 Results of long-term studies are awaited. Alvimopan is another mu-opioid receptor antagonist approved by the FDA to accelerate bowel recovery follow­ ing surgery. Alvimopam also appears to improve symptoms of opioid-induced constipation. Patients who received alvimopam, 0.5 mg or 1 mg twice daily or 1 mg once daily, had significantly more spontaneous bowel movements than those receiving placebo.233 Other symptoms of constipation

also improved. The most common side effects were abdomi­ nal pain, nausea, and diarrhea.

Other Agents

Colchicine, a drug used for gout, and misoprostol, a prosta­ glandin analog, have been used to treat patients with severe chronic constipation. Treatment of chronic constipation with colchicine has been studied in a randomized, placebocontrolled, double-blind crossover trial in which colchicine increased the frequency of bowel movements as compared with placebo; however, abdominal pain was greater during administration of colchicine than placebo.234 Data for misoprostol are limited, and side effects of the drug are common.235 Cholinergic Agents Cholinergic agents also have been used to treat constipation. Bethanechol, a cholinergic agonist, appears to benefit patients in whom constipation results from therapy with tricyclic antidepressants; however, data to support its use in patients with other causes of constipation are limited. A single intravenous dose of neostigmine, a cholinesterase inhibitor, has been shown to be remarkably effective in decompressing the colon in patients with acute colonic pseudo-obstruction236 (see Chapter 120), but controlled studies of this class of drugs have not been completed in patients with normal-transit or slow-transit constipation. Side effects, such as bradycardia, increased salivation, vomiting, and abdominal cramping, are common. Botulinum Toxin Clostridium botulinum toxin type A (Botox), a potent neu­ rotoxin that inhibits presynaptic release of acetylcholine, has been injected intramuscularly into the puborectalis muscle to treat defecatory disorders. Preliminary data suggest that botulinum toxin may be effective for treating patients with defecatory disorders in which spastic pelvic floor dysfunction causes outlet delay,237 including those who also have Parkinson’s disease.112,113 One study showed that 19 of 24 patients reported improvement in symptoms and physiologic measurements of pelvic floor function at two months.238 Controlled trials have not yet been per­ formed, however, and this approach is not recommended in lieu of biofeedback, for which clinical experience is greater (see later). Newer Agents A newer approach to treating constipation involves using neurotrophins, a multigene family of proteins that includes nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and neurotrophin-3 (NT-3).239,240 These factors promote growth of subpopulations of sensory neurons and modulate synaptic transmission at developing neuromuscular junctions in Xenopus nerve muscle cul­ tures.241 A specific factor, R-metHuNT-3, has been shown to increase stool frequency and facilitate passage of stool when administered to constipated patients. In healthy persons, R-metHuNT-3 administered subcutaneously has been shown to accelerate gastric, small bowel, and colonic transit. The effects on stool frequency are observed within three days of the start of treatment and last up to five days after cessation of treatment. R-metHuNT-3 has been well toler­ ated, although half of patients treated in the two studies experienced injection site reactions or paresthesias, pre­ sumably by stimulating noncholinergic excitation and sup­ pressing nitrergic inhibition. Linaclotide is a novel, investigational, minimally absorbed guanylate cyclase C agonist that has been shown to reduce

281

282

Section III  Symptoms, Signs, and Biopsychosocial Issues visceral pain and promote intestinal secretion and colonic transit in animal models. In health volunteers, linaclotide was safe and well tolerated. In women with IBS with constipation, oral linaclotide, 1000 µg daily, significantly accelerated ascending colon transit and improved bowel function.241 In a four-week phase IIb study, 307 patients with chronic constipation with less than three spontaneous bowel movements/week were randomized to placebo or linaclotide, 75, 150, 300, or 600 µg. Within 24 hours, 55% of patients who received 300 µg and 59% of patients who received all the linaclotide doses had a spontaneous bowel movement, compared with 37% of those who received placebo. By week four, the mean number of spontaneous bowel movements/week was 5.5 in those who received 300 µg of linaclotide, compared with 3.9 for those who received placebo. Linaclotide was well tolerated; diarrhea was the only dose-dependent adverse effect. No evidence of rebound constipation was noted in the post-treatment period.242

OTHER FORMS OF THERAPY Defecation Training

Defecation training typically involves three to five treatment sessions, each lasting at least 30 minutes. During these ses­ sions, the normal defecation process is taught, and miscon­ ceptions are dispelled. Patients are encouraged to give a detailed description of their bowel symptoms, prompted by a sympathetic listener who is familiar with the full range of problems experienced by those with defecatory dysfunc­ tion. This process is in itself therapeutic because it enables patients to discuss symptoms that otherwise might be regarded as a private burden. Recommendations regarding the proper amount of fiber intake are often given. For patients with infrequent defecation, the importance of developing a regular bowel habit and not ignoring a call to defecate is emphasized. For those who spend excessive time in the bathroom because of ineffective straining, a regimen of less frequent visits to the bathroom and more effective defecation is recommended. The optimum posture for def­ ecation, including the benefit of raising the feet above floor level when using a Western-type toilet, is described. Patients are encouraged to practice what they are taught; that they may be able to help themselves often gives patients new self-confidence. At each visit, patients are encouraged to reduce any dependence on laxatives, enemas, and sup­ positories. Progress is praised at successive treatment sessions.

Anorectal Biofeedback

Anorectal biofeedback typically follows defecation training. During anorectal biofeedback, patients receive visual or auditory feedback, or both, on the functioning of their anal sphincter and pelvic floor muscles. Biofeedback can be used to train patients to relax their pelvic floor muscles during straining and to coordinate this relaxation with abdominal maneuvers to enhance entry of stool into the rectum. Bio­ feedback can be performed with an electromyographic or anorectal manometry catheter. Simulated evacuation with a balloon or silicone-filled artificial stool is commonly taught to patients to emphasize normal coordination of successful defecation.243 Patient education and rapport between the therapist and the patient are integral components of successful biofeedback.244 Patients typically complete from six sessions in six weeks to three sessions/day for 10 successive days. A systematic review of biofeedback studies performed up to 1993 revealed an overall success rate of 67%, although

controlled studies were lacking.245 Biofeedback may be less effective for patients with descending perineum syndrome than for those with spastic pelvic floor disorders.91 In a review of 38 biofeedback studies, psychological factors were found to influence the response to biofeedback.246 More recently, several controlled trials have supported the efficacy of biofeedback.247-249 Patients with pelvic floor dys­ synergia who failed fiber, 20 g/day, plus enemas or sup­ positories were randomized to five weekly biofeedback sessions (n = 54) or PEG, 14.6 to 29.2 g/day, plus five weekly counseling sessions on preventing constipation (n = 55). At six months, major improvement was reported by 80% of patients who underwent biofeedback compared with 22% of the laxative-treated patients (P < 0.001). The benefits of biofeedback were sustained at 12 and 24 months and pro­ duced greater reductions in straining, sensations of incom­ plete evacuation and anorectal blockage, use of enemas and suppositories, and abdominal pain (all P < 0.01). Stool fre­ quency increased in both groups. All biofeedback-treated patients reporting major improvement were able to relax the pelvic floor and defecate a 50-mL balloon at 6 and 12 months.248 In another controlled trial, 77 patients with dys­ synergic defecation were randomized to biofeedback, sham therapy, or standard therapy for three months. Patients who received biofeedback were significantly more likely to correct dyssynergia, improve the defecation index, decrease the balloon expulsion time, increase the number of CSBMs/ week, and decrease the use of digital maneuvers. Global bowel satisfaction was also higher for patients treated with biofeedback than for those in the other groups.249 Originally, biofeedback training was intensive and initi­ ated during admission to the hospital,250 but subsequent experience has shown that training as an outpatient is sat­ isfactory. A small comparative trial has shown no difference in outcome with or without use of an intrarectal balloon or with home training.251 Results are similar when training is conducted with or without access to a visual display of muscular activity. In the absence of a visual display, the instructor gives continuous information and encouragement to the patient and assesses the effect of instruction by observing how the patient strains and by sensing the effec­ tiveness of straining through gentle tension on a rectal balloon. Many patients do not complete defecation training. Of those who do complete the training, most continue to report improvement in symptoms up to two years after completion of training.250,251 Symptoms reported to improve with defe­ cation training include bowel frequency, straining, abdomi­ nal pain, bloating, and need for laxatives.252 Physiologic measurements before and after treatment have shown that training results in appropriate relaxation of the puborectalis and external anal sphincter muscles,253-255 increase in intra­ rectal pressure,75 a widened rectoanal angle on straining during defecation, an increased rate of rectal emptying, an increased rate of colonic transit, and increased rectal mucosal blood flow. Role of Physiologic and Anatomic Investigation Most published series have restricted defecation training and anorectal biofeedback to patients with a defecatory disorder (i.e., paradoxical contraction of the pelvic floor muscles). At one center, however, such training appeared to benefit a high proportion of unselected patients with idiopathic constipation, regardless of the results of investi­ gation of colonic transit or pelvic floor dysfunction, includ­ ing patients with slow colonic transit.254,256 In another series, the results of treatment did not depend on the presence or absence of a rectocele, intussusception, or perineal

Chapter 18  Constipation descent.252 Other investigators, however, have shown that patients who fail to respond to defecation training and biofeedback have a greater degree of perineal descent than those who respond.91 Defecation training has benefited some patients in whom constipation developed after hysterectomy257 and some patients with solitary rectal ulcer syndrome.258

Complementary and Alternative Medical Therapies

Many complementary and alternative therapies are used by patients with constipation,259 but clinical studies are limited and generally of poor quality (see Chapter 127). A system­ atic review of approximately 90 trials of acupuncture for the treatment of chronic constipation identified in the Chinese Biomedical Database is in progress.260

Sacral Nerve Stimulation

Uncontrolled data suggest that sacral nerve stimulation may be helpful for patients with severe constipation.261 Further investigation is in progress.

Surgery

The goal of surgical treatment for patients with severe constipation is to increase bowel frequency and ease of defecation; a possible additional benefit is relief of abdomi­ nal pain and distention. Procedures may be divided into three groups—partial or total colectomy, construction of a stoma, and anorectal operations undertaken to improve defecatory function. Colectomy Colectomy for constipation produces variable results. A review of 32 published studies of surgery for chronic con­ stipation has found considerable variability in rates of patient satisfaction (39% to 100%).262 The most common complications following surgery are small bowel obstruc­ tion, diarrhea, and incontinence; however, diarrhea and incontinence tend to improve after the first year following surgery. Selection of Patients for Colectomy.  Preoperative psy­ chological assessment is essential, because poor results are common among patients who are psychologically dis­ turbed.263 Because the aim of surgery is to increase bowel frequency, slow colonic transit must be demonstrated by an objective method. Also, defecatory function must be assessed, inasmuch as the inability to expel stool from the rectum may be a major factor in causing symptoms. Finally, a generalized intestinal dysmotility or pseudo-obstruction syndrome should be excluded, as much as possible, by appropriate radiographic study of the small intestine and, when available, studies of gastric emptying and small bowel transit. Series in which these steps have been taken to select a homogeneous group of patients have shown the best results, although longer follow-up is awaited. For example, at one center, only 74 of 1009 patients referred for possible surgi­ cal treatment of chronic constipation underwent surgery. Measurement of intestinal transit and tests of pelvic floor function revealed that 597 patients had no quantifiable abnormality and that 249 patients had pelvic floor dysfunc­ tion without slow colonic transit. Colectomy with an ileorectal anastomosis was performed in 52 patients with demonstrable slow colonic transit and normal defecatory function. The operation also was performed in 22 patients with slow colonic transit and pelvic floor dysfunction after the latter had been treated by a training program. Of the 74 patients treated surgically, 97% were satisfied with the result, and 90% had a good or improved quality of life after

a mean follow-up of 56 months. There was no operative mortality, but 7 patients had a subsequent episode of small bowel obstruction.41 Type of Operation.  Several series have shown that the results of colectomy with cecorectal or ileosigmoid anasto­ mosis are inferior to those for a subtotal colectomy with an ileorectal anastomosis.264 Occasional reports have described proctocolectomy with ileoanal anastomosis and construc­ tion of an ileal pouch, usually following failure of colec­ tomy and ileorectal anastomosis.265 In one patient, ileorectal anastomosis failed because the rectum had a larger than normal capacity.266 Laparoscopic subtotal colectomy appears to be as effective as an open approach.267,268 Construction of a Stoma A colostomy is occasionally performed for slow-transit constipation because it is reversible and the results of colectomy are uncertain. Most patients report subjective improvement after a colostomy performed as a primary procedure for slow-transit constipation or for neurologic disease.89 Many patients, however, continue to require laxa­ tives or regular colonic irrigation. An ileostomy occasionally is performed after failure of colectomy and ileorectal anastomosis for slow-transit con­ stipation, either because constipation persists or because severe diarrhea and incontinence occur. Patients who do not benefit from colectomy with ileorectal anastomosis are likely to be those with a generalized disorder of gut motility or those with a psychological disturbance. Creation of a continent catheterizable appendicostomy, through which antegrade enemas can be administered, can sometimes benefit patients with paraplegia and severe con­ stipation and incontinence. Such a procedure can decrease the time and medication needed for bowel care; most of the experience is in children.269 Operations for Defecatory Disorders The stapled transanal rectal resection (STARR) procedure has been used with some success, particularly for patients who also have a rectocele and intussusception.270-272 Pubo­ rectalis or internal anal sphincter muscle division is un­successful in patients with slow-transit constipation.273 Procedures to correct a rectocele should be considered only for patients who have evidence of retained contrast during defecating proctography or in women in whom constipation is relieved with digital vaginal pressure.89

KEY REFERENCES

Bharucha AE, Wald A, Enck P, Rao S. Functional anorectal disorders. Gastroenterology 2006; 130:1510-18. (Ref 81.) Camilleri M, Bharucha AE, Ueno R, et al. Effect of a selective chloride channel activator, lubiprostone, on gastrointestinal transit, gastric sensory, and motor functions in healthy volunteers. Am J Physiol Gastrointest Liver Physiol 2006; 290:G942-7. (Ref 221.) Camilleri M, Kerstens R, Rykx A, Vandeplassche L. A placebocontrolled trial of prucalopride for severe chronic constipation. N Engl J Med 2008; 358:2344-54. (Ref 228.) Chiarioni G, Whitehead WE, Pezza V, et al. Biofeedback is superior to laxatives for normal transit constipation due to pelvic floor dyssyn­ ergia. Gastroenterology 2006; 130:657-64. (Ref 248.) Dipalma JA, Cleveland MV, McGowan J, Herrera JL. A randomized, multicenter, placebo-controlled trial of polyethylene glycol laxative for chronic treatment of chronic constipation. Am J Gastroenterol 2007; 102:1436-41. (Ref 194.) Fletcher JG, Busse RF, Riederer SJ, et al. Magnetic resonance imaging of anatomic and dynamic defects of the pelvic floor in defecatory disorders. Am J Gastroenterol 2003; 98:399-411. (Ref 162.) Higgins PD, Johanson JF. Epidemiology of constipation in North America: A systematic review. Am J Gastroenterol 2004; 99:750-9. (Ref 10.)

283

284

Section III  Symptoms, Signs, and Biopsychosocial Issues Holzer B, Rosen HR, Novi G, et al. Sacral nerve stimulation in patients with severe constipation. Dis Colon Rectum 2008; 51:524-9. (Ref 261.) Lembo A, Camilleri M. Chronic constipation. N Engl J Med 2003; 349:1360-8. (Ref 2.) Locke GR 3rd, Pemberton JH, Phillips SF. AGA technical review on constipation. American Gastroenterological Association. Gastroenter­ ology 2000; 119:1766-78. (Ref 22.) Longstreth GF, Thompson WG, Chey WD, et al. Functional bowel dis­ orders. Gastroenterology 2006; 130:1480-91. (Ref 6.) Markowitz GS, Stokes MB, Radhakrishnan J, D’Agati VD. Acute phos­ phate nephropathy following oral sodium phosphate bowel purga­ tive: An underrecognized cause of chronic renal failure. J Am Soc Nephrol 2005; 16:3389-96. (Ref 183.)

Rao SS, Seaton K, Miller M, et al. Randomized controlled trial of bio­ feedback, sham feedback, and standard therapy for dyssynergic def­ ecation. Clin Gastroenterol Hepatol 2007; 5:331-8. (Ref 249.) Steinman TI, Samir AE, Cornell LD. Case records of the Massachusetts General Hospital. Case 27-2008. A 64-year-old man with abdominal pain, nausea, and an elevated level of serum creatinine. N Engl J Med 2008; 359:951-60. (Ref 182.) Thomas J, Karver S, Cooney GA, et al. Methylnaltrexone for opioidinduced constipation in advanced illness. N Engl J Med 2008; 358:2332-43. (Ref 232.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

19 Gastrointestinal Bleeding Thomas J. Savides and Dennis M. Jensen

CHAPTER OUTLINE Initial Assessment and Management of Acute Gastrointestinal Bleeding  285 History  285 Physical Examination  286 Laboratory Studies  288 Clinical Determination of the Bleeding Site  288 Hospitalization  288 Resuscitation  288 Initial Medical Therapy  288 Endoscopy  289 Endoscopic Hemostasis  291 Radiologic Imaging  292 Surgery  293 Upper Gastrointestinal Bleeding  293 Epidemiology  293 Risk Factors and Risk Stratification  293

The annual rate of hospitalization for any type of gastro­ intestinal (GI) hemorrhage in the United States is estimated to be 350 hospital admissions/100,000 population, with more than 1,000,000 hospitalizations annually.1 Approxi­ mately 50% of admissions for GI bleeding are for upper GI (UGI) bleeding (from the esophagus, stomach, duodenum), 40% are for lower GI (LGI) bleeding (from the colon and anorectum), and 10% are for obscure bleeding (from the small intestine). This chapter focuses on overt GI bleeding that may be severe and prompts the patient to seek medical attention. The source of most GI bleeds can be suspected by the clini­ cal symptoms and physical examination and confirmed by upper or lower endoscopy. Initial management focuses on medical resuscitation, followed by endoscopic diagnosis and interventions to stop acute bleeding and prevent recurrent bleeding. Pharmacologic therapy is playing an increasingly important role in the management of UGI bleeding from peptic ulcers and varices. Optimal patient outcomes depend on successful medical resuscitation, precise endoscopic diagnosis, and appropriate use of thera­ peutic endoscopy. Severe gastrointestinal bleeding is defined as documented gastrointestinal bleeding (i.e., hematemesis, melena, hema­ tochezia, or positive nasogastric lavage) accompanied by shock or orthostatic hypotension, a decrease in the hema­ tocrit value by at least 6% (or a decrease in the hemoglobin level of at least 2 g/dL), or transfusion of at least two units of packed red blood cells. Most patients with severe gastro­ intestinal bleeding are admitted to the hospital for resusci­ tation and treatment. Overt bleeding implies visible signs of blood loss from the GI tract. Hematemesis is defined as vomiting of blood, which is indicative of bleeding from the esophagus, stomach, or duodenum. Hematemesis includes vomiting of bright red blood, which suggests

Upper Endoscopic Technique  293 Peptic Ulcer  294 Other Causes  303 Varices  306 Lower Gastrointestinal Bleeding  308 Risk Factors and Risk Stratification  309 Mortality  309 Diagnostic and Therapeutic Approach  309 Causes and Management  311 Obscure Overt Gastrointestinal Bleeding  315 Causes and Diagnostic and Therapeutic Approaches  315 Diagnostic Tests  318 Obscure Occult Gastrointestinal Bleeding and Iron Deficiency Bleeding  320 Fecal Occult Blood  320 Iron Deficiency Anemia  321

recent or ongoing bleeding, and dark material (coffeeground emesis), which suggests bleeding that stopped some time ago. Melena is defined as black tarry stool and results from degradation of blood to hematin or other hemochromes by intestinal bacteria. Melena can signify bleeding that originates from UGI, small bowel, or proximal colonic source. Melena generally occurs when 50 to 100 mL or more of blood is delivered into the GI tract (usually the upper tract), with passage of characteristic stool occurring several hours after the bleeding event.2,3 Hematochezia refers to bright red blood per rectum, and suggests active UGI or small bowel bleeding, or distal colonic or anorectal bleeding. Occult gastrointestinal bleeding refers to subacute bleeding that is not clinically visible. Obscure gastrointes­ tinal bleeding is bleeding from a site that is not apparent after routine endoscopic evaluation with esophagogastro­ duodenoscopy (upper endoscopy) and colonoscopy, and possibly small bowel radiography. An algorithm for the initial management of acute, severe UGI bleeding is shown in Figure 19-1.

INITIAL ASSESSMENT AND MANAGEMENT OF ACUTE GASTROINTESTINAL BLEEDING HISTORY

Initial assessment of the patient with acute GI bleeding includes medical history taking, obtaining vital signs, per­ forming a physical examination, including a rectal examina­ tion, and nasogastric lavage. During history taking, patients should be questioned about risk factors and historical fea­ tures that help identify diagnostic possibilities for the bleed­ ing source (Table 19-1). Bleeding from a peptic ulcer should be suspected in patients with a history of an ulcer or those

285

286

Section III  Symptoms, Signs, and Biopsychosocial Issues Severe upper gastrointestinal bleeding

History and physical examination Onset in hospital, syncope, shock, comorbidities, hematochezia Admission to intensive care unit

Hemodynamic resuscitation (ongoing)

Type and crossmatch, complete blood count, chemistry panel, liver biochemical tests, coagulation tests (transfusions as indicated)

Electrocardiogram and chest x-ray

Nasogastric (or orogastric) tube and gastric lavage

Gastroenterology consultation

Proton pump inhibitor may be started before endoscopy if peptic ulcer is suspected

If patient is known or suspected to have chronic liver disease, consider beginning octreotide (bolus and infusion) Figure 19-1.  Algorithm for the initial management of severe upper gastrointestinal (UGI) bleeding. The steps in the algorithm may take place simultaneously or in varying orders and in the emergency department depending on the clinical situation.

taking daily aspirin or other nonsteroidal anti-inflammatory drugs (NSAIDs). Patients who have known or suspected liver disease or who are alcoholics should be suspected of bleeding related to portal hypertension. Patients with heavy alcohol intake, a feeding or chronic nasogastric tube, or a history of gastroesophageal reflux disease are at risk of erosive esophagitis. Patients who have had prior surgical repair of an abdominal aortic aneurysm should be consid­ ered to have a fistula from the graft to the duodenum until proven otherwise. Patients on an anticoagulant such as warfarin should be evaluated for the possibility of excessive anticoagulation. Prior radiation to the abdomen should raise the possibility of radiation enteritis or colitis. Weight loss suggests possible malignancy. Abdominal pain may suggest malignancy, inflammatory bowel disease, or ische­ mic colitis. A change in stool caliber suggests colon cancer or a colonic stricture. Chest pain or syncope suggests pos­ sible cardiovascular complications related to blood loss.

PHYSICAL EXAMINATION

On initial evaluation, physical examination should focus on the patient’s vital signs, with attention to signs of hypovole­ mia such as hypotension, tachycardia, and orthostasis. The abdomen should be examined for surgical scars, tenderness, and masses. Signs of chronic liver disease include spider

Upper endoscopy (generally within 6–12 hours of arrival)

Specific endoscopic treatment (see Figure 19-2); for ulcer, begin proton pump inhibitor if not already started

angiomata, palmar erythema, gynecomastia, ascites, spleno­ megaly, caput medusae, and Dupuytren’s contracture. The skin, lips, and buccal mucosa should be examined for telangiectasias, which are suggestive of hereditary hemor­ rhagic telangiectasia (HHT), or Osler-Weber-Rendu disease. Pigmented lip lesions may suggest Peutz-Jeghers syndrome. Purpuric skin lesions may suggest Henoch-Schönlein purpura. Acanthosis nigricans may suggest underlying malignancy, especially gastric cancer. The patient’s feces should be observed to identify melena or maroon and red stool; however, the subjective description of stool color varies greatly among patients and physicians.4 Nasogastric or orogastric tube placement to aspirate and visually characterize gastric contents can be useful to deter­ mine the presence or absence of large amounts of red blood, coffee-ground material, or nonbloody fluid. Occult blood testing of a nasogastric tube aspirate is not useful, however, because trauma from the nasogastric tube may cause suffi­ cient, although scant, bleeding to cause a false-positive result. Patients who have coffee-ground emesis or fresh bloody emesis that is witnessed do not require placement of a nasogastric tube for diagnostic purposes but may need a nasogastric tube to help clear the gastric blood for better endoscopic visualization and to minimize the risk of aspiration.

Chapter 19  Gastrointestinal Bleeding Table 19-1  Suspected Source of Gastrointestinal Bleeding as Suggested by a Patient’s History SUSPECTED SOURCE OF BLEEDING

PATIENT HISTORY

Nasopharynx

History of nasopharyngeal radiation Recurrent epistaxis Prior nasopharyngeal malignancy Hemoptysis Gastroesophageal reflux disease Heartburn Heavy alcohol use Odynophagia Pill ingestion Traumatic nasogastric tube placement Dysphagia Weight loss Alcohol binge Vomiting Large hiatal hernia Cirrhosis Chronic liver disease Heavy alcohol use Chronic kidney disease Epigastric discomfort Frequent aspirin or nonsteroidal anti-inflammatory drug use History of peptic ulcer disease Early satiation Weight loss Prior severe acute unexplained bleeding Prior abdominal aortic aneurysm surgical repair with synthetic graft Recent endoscopic sphincterotomy Recent liver biopsy or cholangiography Pancreatitis, pseudocyst Recent pancreatography Hereditary nonpolyposis colorectal cancer History of intra-abdominal metastatic cancer Intermittent small intestinal obstruction Recurrent unexplained gastrointestinal bleeding Weight loss Unexplained gastrointestinal bleeding since childhood Use of aspirin or other nonsteroidal anti-inflammatory drug Frequent nosebleeds Hereditary hemorrhagic telangiectasia (Osler-Weber-Rendu disease) Age > 60 years Hematochezia without abdominal pain History of diverticulosis Change in bowel habits Personal or family history colon neoplasia Subacute bleeding Weight loss Cardiovascular disease Hematochezia with abdominal pain or discomfort Family history of IBD Bloody diarrhea History of ulcerative colitis Family history of IBD History of Crohn’s disease Chronic abdominal discomfort Hematochezia with anal pain Dripping blood with bowel movements Hematochezia with normal bowel movements Recent colonoscopy with polypectomy Use of anticoagulants or antiplatelet drugs Age > 70 yr Cardiovascular disease Recurrent bleeding of variable severity Prior intestinal surgical anastomosis History of abdominal radiation therapy

Lungs Esophageal ulceration

Esophageal cancer Mallory-Weiss tear Cameron’s erosions Esophageal or gastric varices or portal hypertensive gastropathy Gastric angiodysplasia Peptic ulcer Gastric cancer Primary aortoenteric fistula Secondary aortoenteric fistula Ampulla of Vater Bile ducts Pancreatic ducts Small intestine malignancy

Meckel’s diverticulum Small intestine or colon ulcerations Small intestine telangiectasias Small intestine angiodysplasia Colonic diverticulosis Colonic neoplasia

Ischemic colitis Ulcerative colitis Crohn’s disease Anal fissure Hemorrhoids Postpolypectomy ulcer Colonic or small intestinal angioectasias Anastomotic ulceration Radiation enteritis or proctitis IBD, inflammatory bowel disease.

287

288

Section III  Symptoms, Signs, and Biopsychosocial Issues LABORATORY STUDIES

Blood from the patient with acute GI bleeding should be sent for standard hematology, chemistry, liver biochemical, and coagulation studies and for typing and crossmatching for packed red blood cells. The hematocrit value immedi­ ately after the onset of bleeding may not reflect blood loss accurately because over 24 to 72 hours there is equilibration of red blood cells in the vascular space with extravascular fluid and hemodilution resulting from intravenous admin­ istration of saline.5 The mean corpuscular volume (MCV) is an important indicator of the chronicity of blood loss; an MCV lower than 80 fL suggests chronic GI blood loss and iron deficiency, which can be confirmed by the finding of low blood iron, total iron-binding capacity (TIBC), and ferritin levels. A low MCV and negative fecal occult blood test result raise the possibility of celiac disease. A high MCV (>100 fL) suggests chronic liver disease or folate or vitamin B12 deficiency. An elevated white blood cell count may occur in more than half of patients with UGI bleeding and has been associated with greater severity of bleeding.6 A low platelet count can contribute to the severity of bleed­ ing and suggests chronic liver disease or a hematologic disorder. The blood urea nitrogen (BUN) and serum creatinine levels can help assess the patient for hemoconcentration (elevated levels) or chronic kidney disease, which may lead to chronic anemia because of decreased erythropoietin production. In patients with UGI bleeding, the BUN level typically increases to a greater extent than the serum crea­ tinine level because of increased intestinal absorption of urea after the breakdown of blood proteins by intestinal bacteria.7 The prothrombin time (PT) and international normalized ratio (INR) assess whether a patient has impairment of the extrinsic coagulation pathway. Values can be elevated in chronic liver disease or with the use of warfarin. Liver biochemical test levels may indicate the presence of acute or chronic liver disease; a low serum albumin level suggests possible chronic liver disease, malnutrition, or protein loss via the intestine or kidney.

CLINICAL DETERMINATION OF THE BLEEDING SITE

Presentation with hematemesis, coffee-ground emesis, or nasogastric lavage with return of a large amount of blood or coffee-ground emesis indicates an UGI source of bleeding. A small amount of coffee-ground material or pink-tinged fluid that clears easily may represent mucosal trauma from the nasogastric tube rather than active bleeding from an UGI source. A clear (nonbloody) nasogastric aspirate does not necessarily indicate a more distal GI source bleeding, because 16% of patients with actively bleeding UGI lesions have a clear nasogastric aspirate.8 The presence of bile in the nasogastric aspirate makes UGI bleeding unlikely but can be seen with an intermittently bleeding UGI source. Melena generally indicates an UGI source but can be seen with small intestinal or proximal colonic bleeding. Hema­ tochezia generally implies a colonic or anorectal source of bleeding unless the patient is hypotensive, which could indicate a severe, brisk UGI bleed with rapid transit of blood through the GI tract.4 Maroon-colored stool can be seen with an actively bleeding UGI source or a small intestinal or proximal colonic source.

HOSPITALIZATION

On the basis of the patient’s initial history, physical exami­ nation, and laboratory test results, the location of bleeding

(upper or lower), suspected bleeding lesion, and severity of bleeding can be predicted. Patients with severe GI bleeding require hospitalization, whereas those who present with only mild acute bleeding (self-limited hematochezia or infrequent melena) and who are hemodynamically stable (not suspected to be volume depleted), have normal blood test results, and can be relied on to return to the hospital if symptoms recur may be candidates for semiurgent outpa­ tient endoscopy rather than direct admission to the hospi­ tal.9,10 On the other hand, patients should be hospitalized in an intensive care unit if they have large amounts of red blood in the nasogastric tube or per rectum, have unstable vital signs, or have had severe acute blood loss that may exacerbate other underlying medical conditions. Patients who have had an acute GI bleed but are hemodynamically stable can be admitted to a monitored bed (step-down unit) or standard hospital bed, depending on their clinical condition. Several small studies have suggested that urgent endos­ copy performed in the emergency department in patients with suspected UGI bleeds can help determine optimal hos­ pital placement; however, widespread implementation of this practice is unlikely.11,12

RESUSCITATION

Resuscitation efforts should be initiated at the same time as initial assessment in the emergency department and continue during the patient’s hospitalization. At least one large-bore (14- or 16-gauge) catheter should be placed intra­ venously, and two should be placed when the patient has ongoing bleeding. Normal saline is infused as fast as needed to keep the patient’s systolic blood pressure higher than 100 mm Hg and pulse lower than 100/min. Patients are transfused with packed red blood cells, platelets, and freshfrozen plasma as necessary to keep the hematocrit value higher than 24%, platelet count higher than 50,000/mm3, and prothrombin time less than 15 seconds, respectively. A GI endoscopist should be consulted as soon as possible to expedite the patient’s assessment and determine the optimal timing of endoscopy. In hospitals with a liver transplanta­ tion program, the transplantation hepatology service should also be notified if the patient is known to have advanced liver disease and is a potential transplant candidate. The patient’s vital signs should be monitored frequently, as appropriate to the level of hospitalization. Laboratorydetermined hematocrit values (not fingerstick hematocrit values, which are less reliable) should be obtained every four to eight hours until the hematocrit value is stable. In patients with active bleeding, an indwelling urinary catheter should be placed to monitor the patient’s urine output. Endotracheal intubation should be considered in patients with active ongoing hematemesis or with altered mental status to prevent aspiration pneumonia. Patients who are older than 60 years, have chest pain, or have a history of cardiac disease should be evaluated for myocardial infarc­ tion with electrocardiography and serial troponin measure­ ments. A chest x-ray should also be considered.

INITIAL MEDICAL THERAPY

Administration of a proton pump inhibitor (PPI) is useful for reducing rebleeding rates in patients with peptic ulcer disease (see later). Starting a PPI in the emergency depart­ ment or intensive care unit (ICU) before endoscopy is per­ formed in patients with severe UGI bleeding has become a common practice but is still controversial.13 Several clinical studies and meta-analyses have shown that infusion of a high-dose PPI before endoscopy accelerates the resolution

Chapter 19  Gastrointestinal Bleeding of endoscopic stigmata of bleeding in ulcers (see later) and reduces the need for endoscopic therapy but does not result in improved clinical outcomes in the transfusion require­ ment, rebleeding rate, need for surgery, or death rate.14-17 Patients with a strong suspicion of portal hypertension and variceal bleeding should be started empirically on intrave­ nous octreotide (bolus followed by infusion [see later and Chapter 90]), which can reduce the risk of rebleeding to a rate similar to that associated with endoscopic therapy (Fig. 19-2; also see Fig. 19-1).18,19

ENDOSCOPY

GI endoscopy will identify the bleeding site and permit therapeutic hemostasis in most patients with GI bleeding. Endoscopy should be done only when it is safe to do so and

when the information obtained from the procedure will influence patient care. Ideally, the patient should be hemo­ dynamically stable, with a heart rate of less than 100 beats/ min and a systolic blood pressure higher than 100 mm Hg. Respiratory insufficiency, altered mental status, or ongoing hematemesis indicates the need for endotracheal intubation before emergency upper endoscopy to stabilize the patient and protect the airway. Coagulopathy and thrombocytope­ nia should be corrected with transfusions prior to endos­ copy. Proper medical resuscitation will not only allow safer endoscopy, but also ensure a better diagnostic examination for lesions, such as varices, that are volume dependent and will allow more effective hemostasis because of the correc­ tion of coagulopathy (Figs. 19-3 and 19-4; also see Figs. 19-1 and 19-2).

Upper endoscopy

Major stigmata (active bleeding, NBVV, or clot)

Oozing

Flat pigmented spot or clean-based ulcer

Combination endoscopic hemostasis (e.g., epinephrine injection and multipolar electrocoagulation)

Hemoclip or thermal hemostasis

Oral PPI and early discharge

Oral PPI twice daily High-dose PPI (IV bolus plus infusion for 72 hr), followed by oral PPI

Figure 19-2.  Algorithm for the endoscopic and medical management of severe ulcer hemorrhage following hemodynamic stabilization. IV, intravenous; NBVV, nonbleeding visible vessel; PPI, proton pump inhibitor; UGIB, upper gastrointestinal bleed.

Severe hematochezia Ongoing hemodynamic resuscitation History, physical examination, nasogastric tube Consult gastroenterologist + surgeon Oral or nasogastric tube colonic purge

Anoscopy Colonoscopy (or flexible sigmoidoscopy)

Source identified (see Figure 19-4)

No source identified

Upper endoscopy or push enteroscopy

Source identified: Treat appropriately (see Figures 19-1 and 19-2)

Source identified: Arteriographic embolization or surgery

No source identified: RBC scintigraphy Angiography

No source identified: Consider repeat endoscopic studies, capsule endoscopy, balloon enteroscopy, or surgery

Figure 19-3.  Algorithm for the management of severe hematochezia. RBC, red blood cell.

289

290

Section III  Symptoms, Signs, and Biopsychosocial Issues Severe hematochezia Ongoing hemodynamic resuscitation History, physical examination, nasogastric tube History of cirrhosis, ulcers, melena, or hematemesis Upper endoscopy and/or push enteroscopy Source identified: Treat

No source identifed

History of hemorrhoids, pelvic or abdominal radiation, colitis, diarrhea

No identifiable risk factors, painless hematochezia

Anoscopy and flexible sigmoidoscopy Source identified: Treat

No source identified

Colonic purge and urgent colonoscopy

No source identified: Push enteroscopy

Source identified: Treat

Source identified: Treat

No source identified: Capsule endoscopy or RBC scintigraphy or angiography

Source identified: Treat (may require balloon enteroscopy)

No source identified: Balloon enteroscopy or surgery

Figure 19-4.  Algorithm for the management of severe hematochezia modified according to the patient’s history. RBC, red blood cell.

Patients with active hemorrhage (i.e., a high-volume bloody gastric lavage or ongoing hematochezia) should undergo emergency endoscopy soon after medical resuscita­ tion. Where emergency endoscopy should be performed in the hospital depends somewhat on local circumstances, but, in general, emergency endoscopy is best performed once the patient has reached an ICU bed, rather than in the emer­ gency department, because resources (personnel, medica­ tions, space) are more readily available in the ICU. Patients suspected of having cirrhosis or an aortoenteric fistula or who rebleed in the hospital should undergo emergent endoscopy, usually within six hours of admission or rebleeding. Patients who are hemodynamically stable without evidence of ongoing bleeding can undergo urgent endoscopy (within 12 hours), often in the GI endoscopy unit rather than the ICU. Middle-of-the-night endoscopy should be avoided, except for the most severely bleeding or highrisk patients, because well-trained endoscopy nurses, endo­ scopic equipment, and surgical backup may not be available at night. In the rare patient with massive bleeding and refractory hypotension, endoscopy can be performed in the operating room, with the immediate availability of surgical management, if necessary. In patients with severe UGI bleeding, gastric lavage with a large (34-Fr) orogastric tube should be performed to evacu­ ate blood and clots from the stomach to prevent aspiration and allow good endoscopic visualization. Special lavage systems can help remove blood rapidly. The intravenous

administration of erythromycin (a gastric prokinetic agent) 30 to 90 minutes before upper endoscopy to induce gastric contraction and push blood from the stomach into the small intestine helps endoscopic visualization.20,21 Therapeutic single- or double-channel endoscopes with large-diameter suction channels should be used to allow quick removal of fresh blood from the GI tract during endoscopy. Addition­ ally, a water pump can be used to irrigate target lesions through an accessory channel and dilute blood for suction­ ing, both of which facilitate visualization. Using iced saline lavage to prevent or decrease UGI bleeding is of no particu­ lar value and may impair coagulation and cause hypother­ mia. Gastric lavage with lukewarm tap water is as safe as lavage with sterile saline and much less expensive. In patients with severe hematochezia and suspected active colonic bleeding, urgent colonoscopy can be under­ taken after a rapid purge (see Figs. 19-3 and 19-4).22,23 Patients should receive 4 to 8 L of polyethylene glycol purge orally or via a nasogastric tube over four to six hours until the rectal effluent is clear of stool, blood, and clots. Addi­ tional polyethylene glycol bowel purge may be required in some patients, particularly those with active bleeding, severe constipation, or the onset of hematochezia in the hospital. Metoclopramide, 10 mg, may be given intrave­ nously before the purge and repeated every four to six hours to facilitate gastric emptying and reduce nausea. In patients with severe or ongoing active hematochezia, urgent co­ lonoscopy should be performed within 12 hours, but only

Chapter 19  Gastrointestinal Bleeding Severe overt obscure gastrointestinal bleeding Hematochezia

Melena

Urgent colonoscopy after colonic purge

Upper endoscopy and/or push enteroscopy

Source identified: Treat

No source identified

Source identified: Treat

Source identified: Treat

No source identified: Colonoscopy with examination of terminal ileum

No source identified: Capsule endoscopy

Source identified No source identified: Balloon endoscopy In proximal small intestine

In distal small intestine

Deep enteroscopy*

Retrograde ileoscopy (via balloon enteroscopy or colonoscopy)

Source identified: Treat or laparotomy and intraoperative enteroscopy

No source identified: Supportive care

Figure 19-5.  Algorithm for the management of severe overt obscure gastrointestinal bleeding. *Deep enteroscopy includes double-balloon enteroscopy, single-balloon enteroscopy, and spiral enteroscopy.

after thorough cleansing of the colon. Patients with mild or moderate self-limited hematochezia should undergo co­ lonoscopy within 24 hours of admission, and a colonic purge is also recommended in this situation to cleanse the colon thoroughly. Patients with maroon stool in whom there is pretest uncertainty about the bleeding source should be considered for an urgent polyethylene bowel preparation as well. Co­ lonoscopy immediately after push enteroscopy (see later), while the patient is still sedated, will expedite a patient’s care if push enteroscopy does not provide a diagnosis (and is also indicated for colon cancer screening in patients older than 50 years; Fig. 19-5). Wireless small bowel capsule endoscopy (or capsule endoscopy; see later) can be useful in patients with overt GI bleeding who have normal push enteroscopy and colonos­ copy results and in whom a small bowel source of bleeding is suspected.24 Capsule endoscopy has the advantages of directly visualizing the small intestine to identify potential sources or active bleeding. Disadvantages are that the pro­ cedure takes eight hours to complete and additional time to download and review the images, does not permit therapeu­ tic hemostasis, and is difficult to perform in inpatients because of limited availability of staff trained to place the capsule. A follow-up endoscopic procedure, such as singleor double-balloon enteroscopy or retrograde ileoscopy, may be indicated for definitive diagnosis and treatment if a focal bleeding site is found on capsule endoscopy. Complications related to emergency endoscopy and endo­ scopic hemostasis may occur in up to 1% of patients,

depending on the type of endoscopy and treatment per­ formed.25,26 The most common complications include GI tract perforation, aspiration pneumonia, induced hemor­ rhage, an adverse medication reaction, hypotension, and hypoxia (see Chapter 40).

ENDOSCOPIC HEMOSTASIS

Thermal contact probes have been the mainstay of endo­ scopic hemostasis since the 1970s. These probes come in diameters of 7 and 10 Fr and in lengths that can fit through panendoscopes, enteroscopes, or colonoscopes. Contact probes can physically tamponade a blood vessel to stop bleeding and interrupt underlying blood flow, and thermal energy is then applied to seal the underlying vessel (coap­ tive coagulation). The most commonly used probe is a mul­ tipolar electrocoagulation (MPEC) probe, also referred to as a bipolar electrocoagulation probe, with which heat is created by current flowing between intertwined electrodes on the tip of the probe. Animal studies in which MPEC probes were used to stop bleeding in mesenteric vessels have shown that optimal coagulation occurs with lowpower settings (12 to 16 W) applied for a moderate amount of time (8 to 10 seconds), with moderate pressure on the bleeding site.27 Heater probes can provide a predetermined amount of joules of energy, which does not vary with tissue resistance. Animal studies have shown that heater probes can effectively coagulate arteries up to 2 mm in diameter, a diameter considerably larger than most secondary or ter­ tiary branches of arteries (usually 1 mm) found in resected bleeding human peptic ulcers.28,29 The main risk of using a

291

292

Section III  Symptoms, Signs, and Biopsychosocial Issues thermal probe is perforation with excessive application of coagulation or pressure, especially in acute or nonfibrotic lesions. Thermal probes can also cause a coagulation injury that can make lesions larger and deeper and may induce delayed bleeding in patients with a coagulopathy. Injection therapy is performed most commonly with the use of a sclerotherapy needle to inject epinephrine, diluted to a concentration of 1 : 10,000 or 1 : 20,000, submucosally into or around the bleeding site or stigma of hemorrhage (see later). The advantages of this technique are its wide availability, relatively low cost, and safety in patients with a coagulopathy. Additionally, it is associated with a lower risk of perforation and thermal burn damage than the thermal techniques. The disadvantage of epinephrine injec­ tion is that it is not as effective for definitive hemostasis as thermal coagulation, hemoclipping (see later), or combina­ tion therapy.30,31 Injection therapy can also be performed with a sclerosant, such as ethanolamine or alcohol, but these agents are associated with increased tissue damage and other risks. Endoscopic hemoclips (or clips) have been available since 1974, and have become popular as technical improvements have been introduced.32 Hemoclips serve to apply mechani­ cal pressure to a bleeding site, as is done with surgical clips or sutures. Endoscopic hemoclips differ from surgical clips, however, in that they do not have as much compressive strength, and the currently available clips do not close com­ pletely but leave a small space between the prongs. Animal studies have shown that the first-generation hemoclips could not stop bleeding in vessels larger than a diameter of 1 mm.33 Subsequent hemoclips have been larger and stronger and have had a grasp and release mechanism that improves endoscopic deployment and hemostasis. By not causing significant thermal damage, hemoclips are espe­ cially useful for patients with malnutrition or coagulopa­ thy.34 Nevertheless, hemoclips can also be difficult to deploy depending on the location of the bleeding site, the degree of fibrosis of the underlying lesion, and limitations to endo­ scopic access. Band ligation is a technique in which mucosal (with or without submucosal) tissue is suctioned into a cap placed on the end of the endoscope, and a rubber band is rolled off the cap and over the lesion to compress its base. This tech­ nique is widely used for the treatment of esophageal varices (see Chapter 90) and occasionally can be used for other bleeding lesions. An advantage of band ligation is that it is relatively easy to perform; however, a disadvantage is that sufficient mucosa must be suctioned into the cap for ligation to be successful. Depending on the manufacturer, some band ligation devices can only fit on diagnostic endoscopes, and switching from a larger therapeutic endoscope to a smaller diagnostic endoscope during a case is timeconsuming and inefficient.

RADIOLOGIC IMAGING

Angiography may be used to diagnose and treat severe bleeding, especially when the cause cannot be determined by upper and lower endoscopy. Angiography generally is diagnostic of extravasation into the intestinal lumen only when the arterial bleeding rate is at least 0.5 mL/min.35 The sensitivity of mesenteric angiography is 30% to 50% (with higher sensitivity rates for active GI bleeding than for recur­ rent acute or chronic occult bleeding), and the specificity is 100%.36 An advantage of angiography is that it permits therapeutic intra-arterial infusion of vasopressin or trans­ catheter embolization for hemostasis if active bleeding is detected, without the need for bowel cleansing. Neverthe­ less, the rate of major complications, including hematoma

formation, femoral artery thrombosis, contrast dye reac­ tions, acute kidney injury, intestinal ischemia, and transient ischemic attacks, is 3%.37 Another disadvantage of angiog­ raphy is that it usually does not identify the specific cause of bleeding, only its location. Radionuclide imaging is occasionally helpful for patients with unexplained GI bleeding, although it is used less fre­ quently now than in the past because of the widespread use of endoscopy and lack of availability of nuclear medicine services for emergencies, particularly at night and on week­ ends. Radionuclide imaging can be performed relatively quickly and may help localize the general area of bleeding and thereby guide subsequent endoscopy, angiography, or surgery. The technique involves injecting a radiolabeled substance intravenously into the patient’s bloodstream and then performing serial scintigraphy to detect focal collec­ tions of radiolabeled material. Radionuclide imaging has been reported to detect bleeding at a rate of 0.04 mL/min.38 The two tracers used for radionuclide imaging for bleeding (bleeding scans) are technetium sulfur colloid and techne­ tium pertechnetate–labeled autologous red blood cells. Technetium sulfur colloid is cleared rapidly from the blood­ stream and is therefore useful for identifying acute, active bleeding. Technetium pertechnetate–labeled red blood cells remain in the circulation for up to 24 hours and there­ fore can be used for repeated scanning in patients with intermittent bleeding. A comparative study of patients with suspected GI bleeding has found technetium per­ technetate–labeled red blood cell scans to be more sensitive, specific, and accurate than technetium sulfur colloid scans.39 The overall rate of a diagnostic radionuclide scan is approximately 45%, with a 78% accuracy rate in the local­ ization of the true bleeding site.40,41 Up to 25% of bleeding scans suggest a site of bleeding that proves to be incor­ rect.41-43 The rate of true-positive scans is higher for active bleeding with hemodynamic instability than for less severe bleeding.44 The most common reason for a false-positive result is rapid transit of luminal blood, so that labeled blood is detected in the colon even though it originated from a more proximal site in the GI tract. Caution is recommended in using the results of delayed scans to localize and target lesions for resective surgery.45 Technetium pertechnetate scintigraphy can identify ectopic gastric mucosa in a Meckel’s diverticulum. This diagnosis should be considered in a pediatric or young adult patient with unexplained GI bleeding. The positive predictive value, negative predictive value, and overall accuracy of a so-called Meckel’s scan has been reported to be higher than 90% in young patients.46,47 In patients older than 25 years, however, Meckel’s scans are much less sensitive (<50%). Caution is recommended in interpreting a negative Meckel’s scan result in an adult patient with GI bleeding.48 Conventional radiographic imaging is usually not needed in a patient with GI bleeding but occasionally may provide some important information. In patients with a prior abdom­ inal aortic aneurysm repair and graft, computed tomogra­ phy (CT) with intravenous contrast can identify inflammation between the graft and duodenum and thus suggest graft fistulization into the duodenum.49 In selected patients, an abdominal CT scan can also identify a mass lesion, such as an intra-abdominal tumor, or small bowel abnormalities that may suggest a cause of bleeding. Advances in CT scan technology have permitted CT enterography and CT angiog­ raphy to be performed, with promising results.50,51 Barium radiography is not indicated (and contraindicated if endoscopy or angiography is planned) in the emergency setting. Barium studies were used in the past to identify

Chapter 19  Gastrointestinal Bleeding small bowel diverticula and ulcerations, but this approach has been supplanted by CT enterography and capsule endoscopy.

SURGERY

Most patients admitted for acute GI bleeding have bleeding of mild to moderate severity and do not need surgical con­ sultation. In selected patients with severe, ongoing GI bleed­ ing in whom a diagnosis is not made by urgent endoscopy or colonoscopy, surgical consultation during the hospital­ ization is recommended. Also, patients who have massive hemorrhage and cannot be stabilized hemodynamically should undergo emergency angiography or urgent surgical exploration (either without prior endoscopy or with emer­ gency endoscopy in the operating room; see later). Patients with bleeding that cannot be controlled with endoscopy or angiography may require surgery.

UPPER GASTROINTESTINAL BLEEDING EPIDEMIOLOGY

UGI bleeding is a common medical emergency that accounts for more than 500,000 hospital admissions each year, or approximately 170 patients/100,000 population/year.1 Of the potential causes of severe UGI bleeding, peptic ulcer is the most common, accounting for approximately 40% of cases (Table 19-2).52,53 Despite advances in medical therapy, ICU care, endoscopy, and surgery, the mortality rate of 5% to 10% for severe UGI bleeding has not changed since the 1970s.1,52-56 The lack of decline in the mortality rate may be explained by an increase in the proportion of older patients with GI bleeding, who may die as a result of worsening of other medical conditions rather than from exsanguina­ tion, and an increase in the number of patients with cir­ rhosis and variceal bleeding. Bleeding is self-limited in 80% of patients with UGI hem­ orrhage, even without specific therapy.54,57 Of the remaining 20% who continue to bleed or rebleed, the mortality rate is 30% to 40%.8 Patients at high risk for continuous bleeding or for rebleeding potentially can benefit the most from acute medical, endoscopic, and surgical therapy.

RISK FACTORS AND RISK STRATIFICATION

Scoring tools have been developed to try to identify patients with nonvariceal UGI bleeding at greatest risk for mortality and rebleeding. These tools could be used to triage patients

Table 19-2  Causes of Severe Upper Gastrointestinal Bleeding in the UCLA CURE Database cause Peptic ulcer Gastric or esophageal varix Esophagitis No cause found Upper gastrointestinal tract tumor Angioma* Mallory-Weiss tear Erosions Dieulafoy’s lesion Other

FREQUENCY (%) 38 16 13 8 7 6 4 4 2 2

*Angioectasia and telangiectasia. CURE, Center for Ulcer Research and Education; UCLA, University of California, Los Angeles.

to a higher level of hospital care or more urgent endoscopy. They can be divided into those that use purely clinical parameters available on a patient’s presentation to the hos­ pital and those that incorporate clinical parameters and endoscopic findings. Pre-endoscopy scoring systems for nonvariceal bleeding include the Blatchford Score, the Clinical Rockall Score, and an artificial neural network score. The Blatchford Score uses pre-endoscopy variables such as blood pressure, BUN level, hemoglobin level, heart rate, syncope, melena, liver disease, and heart failure to assess a patient’s risk for needing clinical interventions to control bleeding (e.g., blood transfusions, endoscopic therapy, or surgery).58 The clinical Rockall Score uses the patient’s age, shock, and coexisting illnesses.59 The artificial neural network instru­ ment uses 21 clinical variables to help predict the presence of stigmata of recent hemorrhage at endoscopy (see later) and the need for endoscopic therapy.60 The most commonly used postendoscopy scoring system is the Complete Rockall Score (Table 19-3).59 The Rockall Score was developed after an audit of hospitalizations for UGI bleeding at 74 hospitals in England and validated in another audit of 45 hospitals in 1994, to identify risk factors for death or rebleeding. The Complete Rockall Score includes the Clinical Rockall Score (pre-endoscopy vari­ ables—the patient’s age, shock, and coexisting illnesses) and endoscopic findings, including endoscopic stigmata of recent bleeding. The Rockall score after endoscopic therapy correlates well with mortality but does not seem to correlate as well with the risk of rebleeding.61-63 The Rockall risk stratification schemes can be used not only to identify patients at highest risk for poor outcomes, but also to identify patients at low risk for poor outcomes (i.e., Rockall scores of 0 to 2) who should be considered for early discharge from the hospital.64 Other scoring systems to predict outcomes from UGI bleeding after endoscopy include the Baylor Scoring System and the Cedars-Sinai Bleeding Index.65-68 In general, all these scoring systems are better at determining mortality than rebleeding.69

UPPER ENDOSCOPIC TECHNIQUE

A therapeutic endoscope should be used to allow adequate aspiration of blood and the use of large accessories. Water irrigation should be available. Patients should be resus­ citated medically prior to endoscopy (see earlier), and, if active bleeding is severe, consideration should be given to prophylactic endotracheal intubation or use of an esopha­ geal overtube to minimize the risk of airway aspiration. Once the endoscope is inserted, the first thing to look for is blood in the GI tract lumen. Examining all the nonbloody mucosa quickly is often best to document that these areas are free of any lesions. Then, any liquid blood that can be aspirated should be removed. Aspiration of blood can be aided by water irrigation to dilute the blood. If large clots that cannot be removed with suction are present, the patient can be turned onto his or her back or right side, provided that the airway is adequately protected against aspiration. Raising the head of the bed can also help move a clot dis­ tally from the gastric fundus. Any visualized adherent fresh blood or clot should be followed to find its origin. If too much blood is present in the stomach to allow detection of a bleeding lesion, administration of a prokinetic agent (e.g., erythromycin) should be considered, lavage should be repeated with a large orogastric tube, or the examination should be repeated in the next 24 hours if the patient has stabilized. If bleeding from the duodenum is suspected but not identified with a forward-viewing endoscope, a side-

293

294

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 19-3  Rockall Scoring System for Upper Gastrointestinal Tract Bleeding Points VARIABLE

0

1

2

3

Age (yr) Pulse rate (beats/min) Systolic blood pressure (mm Hg) Comorbidity

<60 <100 Normal

60-79 >100 >100

>80 — <100

— — —

None

—

Diagnosis

Mallory-Weiss tear or no lesion observed No stigmata or dark spot in ulcer base

All other diagnoses

Ischemic heart disease, cardiac failure, other major illness Malignant lesions

Renal failure, hepatic failure, metastatic cancer —

Blood in upper gastrointestinal tract, adherent clot, visible vessel, active bleeding

—

Endoscopic stigmata of recent hemorrhage

TOTAL SCORE 0 1 2 3 4 5 6 7 ≥8

—

FREQUENCY (% OF TOTAL)

REBLEEDING RATE (%)

MORTALITY RATE (%)

4.9 9.5 11.4 15.0 17.9 15.3 10.6 9.0 6.4

4.9 3.4 5.3 11.2 14.1 24.1 32.9 43.8 41.8

0 0 0.2 2.9 5.3 10.8 17.3 27.0 41.1

Modified from Rockall TA, Logan RF, Devlin HB, Northfield TC. Selection of patients for early discharge or outpatient care after acute upper gastrointestinal haemorrhage. National Audit of Acute Upper Gastrointestinal Haemorrhage. Lancet 1996; 347:1138-40.

viewing duodenoscope should be used to examine the duo­ denal wall and ampulla.

PEPTIC ULCER

In the past, peptic ulcer, most commonly gastric or duode­ nal ulcer, was identified as the leading cause of severe UGI bleeding in the United States, accounting for 50% of UGI bleeds and approximately 100,000 hospitalizations/year.70,71 Some data have suggested that the incidence of bleeding peptic ulcer decreased between 1993 and 2002, whereas the proportion of ulcers caused by NSAIDs increased.72 Other data related to peptic ulcer bleeding between 1990 and 2000, however, found no change in overall bleeding rates but an increase in the rate of bleeding in the subgroup of older patients taking NSAIDs (Fig. 19-6).73 The mortality rate associated with peptic ulcer bleeding is 5% to 10%.52,53 The estimated costs of hospitalization for peptic ulcer bleeding is estimated to be more than $2 billion/year in the United States (see Chapter 52).74 Clinical and endoscopic factors in patients with peptic ulcer bleeding associated with increased morbidity and mortality are shown in Table 19-4. Knowledge of these risk factors can be used to identify patients at high risk for rebleeding and assist in planning the timing of endoscopy.

Table 19-4  Factors Predictive of a Poor Prognosis after Hemorrhage from Peptic Ulcer Age > 60 yr Bleeding onset in hospital Comorbid medical illness Shock or orthostatic hypotension Fresh blood in nasogastric tube Coagulopathy Multiple transfusions required Higher lesser curve gastric ulcer (adjacent to left gastric artery) Posterior duodenal bulb ulcer (adjacent to gastroduodenal artery) Endoscopic finding of arterial bleeding or visible vessel

No diagnosis or >1 type of lesion No diagnosis or >1 type of lesion 59%

Other

33% 39%

7% 22%

Pathogenesis

Peptic ulcers are most commonly caused by a decrease in mucosal defense mechanisms attributable to aspirin or other NSAIDs, Helicobacter pylori infection, or both.75,76 In one large multicenter study of patients with severe peptic ulcer bleeding, 57% of those with bleeding from a gastric ulcer (n = 2057) took aspirin or other NSAID, and 45% were infected with H. pylori, whereas 53% of those with a bleed­ ing duodenal ulcer (n = 2033) took aspirin or other NSAID, or both, and 50% were infected with H. pylori.77 Of the patients with a bleeding peptic ulcer in this study, 10% had no obvious cause for the ulcer (H. pylori–negative, no aspirin or other NSAID use, no cancer, no gastrinoma).

11%

Ulcers

17%

12% Gastroesophageal varices

Other 1983–1992 (n = 945)

2000–2008 (n = 300)

Figure 19-6.  The frequencies of major causes of severe upper gastrointestinal bleeding during two time periods in patients seen at the University of California, Los Angeles, Center for Ulcer Research Education. (All differences between the two time periods are statistically significant; P < 0.05.) Note that in the more recent period, the overall number of cases of severe upper gastrointestinal bleeding and the percentage of cases caused by peptic ulcer have declined.

Chapter 19  Gastrointestinal Bleeding H. pylori infection is common, with a prevalence of over 80% of the population in many developing countries and 20% to 50% in industrialized countries.78 H. pylori gastritis most commonly involves the antrum and pre­ disposes patients to duodenal ulcers, whereas gastric body– predominant gastritis is associated with gastric ulcers. The lifetime risk of peptic ulcer disease from H. pylori infec­ tion ranges from 3% in the United States to 25% in Japan (see Chapter 50). NSAIDs are the most widely used medication in the United States, with 11% of the adult population using NSAIDs on a daily basis.79 NSAIDs, including aspirin, pre­ dominantly cause ulceration by inhibiting cyclooxygenasemediated prostaglandin synthesis and thereby impairing mucosal protection, rather than causing direct topical injury.76 Gastroduodenal ulcers are found at endoscopy in 15% to 45% of patients who take NSAIDs regularly.80,81 Gastric ulcers are approximately four times as common as duodenal ulcers in patients who take NSAIDs.82 In a large study of patients with UGI hemorrhage and NSAIDassociated ulcers, however, gastric and duodenal ulcers occurred with equal frequencies.77

Histopathology

In a landmark study by Swain and colleagues, the patho­ logic examination of 27 surgically resected bleeding gastric ulcers with endoscopically visible vessels revealed an un­ derlying artery in 96% of specimens.28 Approximately 50% of the vessels protruded above the surface of the ulcer, whereas the other 50% had clot in continuity with a breach in the vessel wall. The bleeding arteries had a mean di­ ameter of 0.7 mm, with a range of 0.1 to 1.18 mm.

Endoscopic Risk Stratification

Endoscopy not only detects a peptic ulcer, but also can be used to evaluate the ulcer for stigmata associated with an increased risk of rebleeding. The Forrest classification is used to categorize findings during endoscopic evaluation of bleeding peptic ulcers, as follows: active spurting bleeding (Forrest IA); oozing bleeding (Forrest IB); pigmented protu­ berance or nonbleeding visible vessel (NBVV; Forrest IIA); adherent clot (Forrest IIB); flat pigmented spot (Forrest IIC); and clean-based ulcer (Forrest III).83 Overall interobserver agreement among experts for classifying these stigmata of recent bleeding is only fair to moderate, with poor agree­ ment for NBVVs.84,85 Endoscopic stigmata of recent hemorrhage from an ulcer are shown in Figure 19-7, and the risk of rebleeding asso­ ciated with each stigma is shown in Figure 19-8. Patients at high risk of rebleeding without treatment are those with active arterial bleeding (90%), an NBVV (50%), or an adherent clot (33%).86,87 These patients benefit from endoscopic hemostasis (see later). An endoscopically identified NBVV that has a translucent (pearl or whitish) color has a higher risk of rebleeding than a darkly colored pigmented pro­tuberance (clot), because the translucent stigma likely represents the arterial wall.88,89 A multi­ variate analysis of predictors of persistent or recurrent bleeding in patients with nonvariceal UGI bleeding is shown in Table 19-5. Patients with major stigmata of ulcer hemorrhage (spurting, NBVV, or adherent clot) benefit most from endoscopic hemostasis, whereas those with a flat spot or clean ulcer base do not. Patients with oozing bleeding and no other stigma (e.g., a clot or NBVV) have an intermediate risk of rebleeding and may benefit from

*

A

C

B

D

Figure 19-7.  Endoscopic stigmata of recent peptic ulcer bleeding. A, Active bleeding with spurting. B, Visible vessel (arrow) with an adjacent clot. C, An adherent clot. D, Slight oozing of blood after washing in the center of an ulcer without a clot or visible vessel.

295

Section III  Symptoms, Signs, and Biopsychosocial Issues

Risk (%)

296

100 90 80 70 60 50 40 30 20 10 0

were treated with an intravenous histamine H2 antagonist and cessation of aspirin and other NSAIDs.88,90,91 Natural history studies of untreated NBVVs have found that these lesions resolve over four days and adherent clots tend to resolve over two days.92

90%

50% 33% 10%

7%

3%

Clean Active Non- Adherent Oozing Flat ulcer bleeding bleeding clot spot base visible vessel Figure 19-8.  Rebleeding rates without endoscopic therapy or administration of a proton pump inhibitor in patients with ulcers demonstrating various stigmata of recent hemorrhage. (From the University of California, Los Angeles, Center for Ulcer Research Education database, unpublished.)

Doppler Probe Ultrasound Portable Doppler ultrasound probes can be passed through the working channel of an endoscope and applied to an ulcer to determine if blood flow is present beneath a stigma in the ulcer base.93,94 The presence of a blood flow signal correlates with the risk of rebleeding before and after endo­ scopic therapy. Conflicting results have been reported, however, as to whether use of Doppler ultrasound improves the outcome of endoscopic hemostasis in patients with acute peptic ulcer bleeding.95,96 A decision-analysis study has found that Doppler ultrasound is the preferred costminimizing strategy over conventional endoscopic therapy alone in patients with acute peptic ulcer bleeding,97 but the area remains one of active investigation.

Endoscopic Hemostasis

Table 19-5  Independent Risk Factors for Persistent or Recurrent Gastrointestinal Tract Bleeding

RISK FACTOR Clinical Factors Health status (ASA class 1 vs. 2-5) Comorbid illness Shock (systolic blood pressure < 100 mm Hg) Erratic mental status Ongoing bleeding Age ≥ 70 yr Age > 65 yr Transfusion requirement Presentation of Bleeding Hematemesis Red blood on rectal examination Melena Laboratory Factors Coagulopathy Initial hemoglobin ≤ 10 g/dL Endoscopic Factors Ulcer location on superior wall of duodenum Ulcer location on posterior wall of duodenum Active bleeding High-risk stigmata Ulcer size ≥ 2 cm Ulcer location high on lesser curve Diagnosis of gastric or duodenal ulcer Clot over ulcer

range of ODDS RATIOs FOR INCREASED RISK 1.94-7.63 1.6-7.63 1.2-3.65 3.21 3.14 2.23 1.3 NA 1.2-5.7 3.76 1.6 1.96 0.8-2.99 13.9 9.2 2.5-6.48 1.91-4.81 2.29-3.54 2.79 2.7 1.72-1.9

ASA, American Society of Anesthesiologists; NA, not applicable. Data from Barkun A, Bardou M, Marshall JK. Consensus recommendations for managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2003; 139:843-57.

endoscopic hemostasis but not from high-dose PPI infu­ sion (see later). The risk of rebleeding from a peptic ulcer decreases sig­ nificantly 72 hours after the initial episode of bleeding. This conclusion is based on studies in which serial endoscopies were performed and only active bleeding was treated endo­ scopically, with all other stigmata observed. All patients

Active Bleeding and Nonbleeding Visible Vessels Many well-conducted randomized controlled trials, metaanalyses, and consensus conferences have concluded that endoscopic hemostasis with epinephrine injection or coap­ tive thermal probe therapy significantly decreases the rates of ulcer rebleeding, urgent surgery, and mortality in patients with high-risk stigmata, such as active bleeding and NBVVs.98-101 The rebleeding rates for peptic ulcers with various endoscopic stigmata are shown in Figure 19-8. These rebleeding rates are based on studies performed before the widespread use of high-dose infusions of PPIs and predominantly used injection therapy, MPEC therapy, or a combination of injection and thermal probe therapy. In general, for the lesions at highest risk, including those with active bleeding (90% risk of ongoing bleeding) or NBVVs (50% risk of ongoing bleeding), endoscopic hemo­ stasis alone decreases the rebleeding rate to approximately 15% to 30% (Table 19-6). The adjunctive administration of a high-dose intravenous PPI (e.g., pantoprazole, 80-mg bolus and 8 mg/hr for 72 hours) decreases this rate even further, as discussed in the next section. Intravenous formulations of pantoprazole, lansoprazole, and esomeprazole are avail­ able in the United States. The most commonly used treatment for ulcer bleeding worldwide is epinephrine injection therapy, because it is widely available, easy to perform, safe, and inexpensive. Therapy with epinephrine alone seems to be more effective when used in high doses (13 to 20 mL) than in low doses (5 to 10 mL).102 Injection of epinephrine results in a fivefold increase in circulating plasma epinephrine levels but rarely is thought to cause clinically significant cardiovascular events.103 Although epinephrine injection alone is effective com­ pared with placebo, numerous studies and meta-analyses have shown that the addition of a thermal or mechanical hemostatic modality further decreases the rates of rebleed­ ing, surgery, and mortality significantly.30,104,105 Several studies have suggested that the only benefit to adding epi­ nephrine injection to thermal probe therapy is in patients with active bleeding and that no benefit is seen in patients with NBVVs.106,107 Mechanical endoscopic clips have not been studied as well as injection and thermal probe techniques but seem to be more effective than epinephrine injection alone and have shown mixed results when compared with thermal probe therapy.108-111 In a meta-analysis of outcomes for ulcer

Chapter 19  Gastrointestinal Bleeding Table 19-6  Endoscopic Stigmata of Recent Ulcer Hemorrhage ENDOSCOPIC APPEARANCE

FREQUENCY (%)

RISK OF REBLEEDING (%)

12 22 10 14 10 32

90 50 33 10 7 3

Active arterial bleeding Visible vessel Adherent clot Oozing without stigmata Flat spot Clean ulcer base

RISK OF REBLEEDING AFTER ENDOSCOPIC HEMOSTASIS (%)* 15-30 15-30 0-5 0-5 NA NA

*Reduction in bleeding risk is without the administration of a proton pump inhibitor. NA, not applicable.

Table 19-7  Endoscopic Technical Parameters for Using Multipolar Electrocoagulation in the Treatment of Bleeding Lesions* Peptic Ulcer

Epinephrine Injection Probe size|| Pressure¶ Power setting (W)** Pulse duration (sec) Endpoint

ACTIVE BLEEDING

NONBLEEDING VISIBLE VESSEL

ADHERENT CLOT

MALLORYWEISS TEAR

DIEULAFOY’S LESION

GASTRIC ANGIOECTASIA

COLON DIVERTICULum with VISIBLE VESSEL

Yes† Large Firm 12-15 8-10 Bleeding stops

No Large Firm 12-15 8-10 Flat vessel

Yes‡ Large Firm 12-15 8-10 Flat stigma

Maybe Large or small Moderate 10-15 4 Bleeding stops

Yes Large Firm 10-15 8-10 Flat vessel

No Large Light 10-15 2 White

Maybe§ Large or small Light 10-15 2 Flat vessel

COLON ANGIOECTASIA No Large or small Light 10-15 2 White

*These guidelines from UCLA CURE have been derived from experimental and randomized endoscopic studies. Power, pressure, and duration settings must be reduced for small, acute, or deep bleeding lesions. Epinephrine (1 : 20,000) injected in 1-mL aliquots into each of 4 quadrants should be used to control bleeding initially, followed by coagulation. ‡ Epinephrine (1 : 20,000) injected in 1-mL aliquots into each of 4 quadrants should be injected around clot initially, followed by piecemeal snare resection and treatment of underlying stigmata. § Colonic diverticulum with active bleeding can be treated with epinephrine (1 : 20,000) injected into the neck or base. If a visible vessel is seen at the neck, it can be treated with multipolar electrocoagulation. || Large probe is 10 Fr (3.2-mm diameter) and fits through a 3.8-mm endoscope channel. Small probe is 7 Fr (2.4 mm) and fits through a 2.8-mm endoscope channel. ¶ Pressure is the tamponade pressure exerted en face or tangentially via the contact probe directly on the lesion. **Power setting using BICAP II generator. Power settings are general guidelines and may vary based on the generator used. CURE, Center for Ulcer Research Education; UCLA, University of California, Los Angeles; W, watts. †

hemorrhage, application of hemoclips was shown to be superior to epinephrine injection alone but comparable to thermocoagulation.31 Hemoclips have the advantage of being able to be used in patients with severe coagulopathy without the risk of inducing bleeding. The disadvantage of hemoclips is that they can be difficult to deploy, depend­ ing on the position of the endoscope in approaching an ulcer. Adherent Clots An adherent clot is generally defined as a blood clot over an ulcer that is resistant to several minutes of vigorous target jet water irrigation. The rebleeding rate for ulcers with an adherent clot with medical therapy alone is 8% to 35%, with most large studies reporting rebleeding rates of 30% to 35%.112-115 Randomized controlled studies have shown that endoscopic treatment of adherent clots can decrease the rebleeding rate to less than 5% (see Table 19-6). A metaanalysis has found that endoscopic therapy is superior to medical therapy for preventing recurrent bleeding from peptic ulcers with an adherent clot, but no differences in the need for surgery, duration of hospitalization, number of transfusions, or mortality rate are observed.116 These studies were performed prior to the widespread use of PPIs, which also decrease rates of rebleeding.

Clean-Based Ulcers Patients with clean-based ulcers at endoscopy after target irrigation have a rebleeding rate of less than 5%. Laine and colleagues have found no difference in outcomes between patients who immediately resumed eating and those who waited several days before they resumed eating after an UGI bleed.117 Longstreth and Feitelberg showed that selected low-risk patients with clinically mild UGI bleeds and clean-based ulcers can be discharged safely to home with a significant savings in cost.9,10

Techniques for Endoscopic Hemostasis

Active Bleeding The technique used at the University of California, Los Angeles (UCLA) Center for Ulcer Research and Education (CURE) for actively spurting ulcer bleeding is to inject 0.5to 1.0-mL aliquots of epinephrine (1 : 20,000) via a sclero­ therapy needle into four quadrants of the ulcer within 1 to 2 mm of the bleeding site (Table 19-7). When combination therapy is performed, coagulation is performed with a large 10-Fr multipolar probe. After epinephrine injection, the thermal probe is placed directly on the bleeding site to tamponade the site and stop the bleeding, and coagulation is applied with long (10-second) pulses and firm pressure at a low (12- to 15-W) power setting (Fig. 19-9). The probe

297

298

Section III  Symptoms, Signs, and Biopsychosocial Issues

A

B

C

D

E

F

Figure 19-9.  Actively bleeding gastric ulcer treated with a combination of epinephrine injection, multipolar electrocoagulation, and hemoclip placement. A, Clot with oozing of blood is seen. B, After injection of epinephrine, the oozing has subsided; the edge of the ulcer is seen inferior to the clot. C, Multipolar electrocoagulation is applied with a probe. D, The appearance of the ulcer after electrocoagulation; some oozing is noted at the 7 o’clock position at the edge of the crater. E, A single hemoclip has been applied; bleeding has ceased entirely. F, A second hemoclip has been applied.

is then removed slowly from the ulcer (sometimes with gentle irrigation to prevent pulling coagulated tissue), and thermal coagulation is repeated as needed to stop bleeding and flatten any underlying visible vessel. Epinephrine injec­ tion can be repeated if rebleeding persists. With successful endoscopic hemostasis, the rebleeding rate can be decreased to 30% with monotherapy and 15% with combination therapy (see Table 19-6). Alternatively, injection of epi­ nephrine followed by hemoclip placement directly across the actively bleeding site is also effective, although some investigators recommend that clips be placed prior to injec­ tion of epinephrine to allow placement of the clip directly on the vessel rather than on a submucosal epinephrinefilled cushion. Nonbleeding Visible Vessel In contrast to active arterial bleeding, no significant dif­ ference in results between thermal therapy alone and com­ bination thermal and epinephrine injection therapy is seen with NBVVs. We use the same technique as that used to stop active bleeding; visible vessels are flattened using a large probe, firm pressure, and a low power setting (Fig. 19-10). Hemoclipping can also be effective for preventing rebleeding from an NBVV if the clip is placed across the NBVV and a high-dose PPI is administered intravenously

for 72 hours (Fig. 19-11).77,118 With successful endoscopic hemostasis, the rebleeding rate can be reduced to 30% with injection alone and 10% to 15% with thermal coagulation, hemoclipping, or combination therapy (see Table 19-6). Adherent Clot Our current recommendations for treating an adherent clot on an ulcer are first to inject epinephrine (1 : 20,000) in 1-mL increments in four quadrants around the pedicle of the clot and then use a rotatable cold snare to guillotine the clot piecemeal, without pulling it off the base, until an underlying stigma of hemorrhage is identified in the ulcer base or a 3-mm or smaller clot pedicle is left. Coagulation or hemoclipping is performed if active bleeding, a visible vessel, or residual pedicle is seen. Figure 19-12 shows an example of combination treatment for an adherent clot. The combination technique decreases the rebleeding rate from up to 35% (with medical therapy alone) to 5%. Adherent clots are considered a high-risk stigma, and administration of a high-dose PPI is recommended after endoscopic hemostasis.115,116 Oozing of Blood from an Ulcer without Other Stigmata Minor bleeding from the edge or base of an ulcer (without other stigmata) that continues despite water irrigation and

Chapter 19  Gastrointestinal Bleeding

A

B

C

D

Figure 19-10.  A, Epinephrine injection and multipolar electrocoagulation for hemostasis of a chronic gastric ulcer (thick arrow) with a nonbleeding visible vessel (thin arrow). B, The nonbleeding visible vessel is injected with epinephrine, after which blanching and swelling of the surrounding mucosa occur. C, A multipolar electrocoagulation probe is applied with firm pressure and coagulation. D, After completion of treatment, the visible vessel has been coagulated and flattened.

A

B

C

Figure 19-11.  A, A gastric ulcer with a nonbleeding visible vessel (arrow) treated by endoscopy with epinephrine injection (B) and hemoclip placement (C).

observation suggests the need for endoscopic treatment. The rebleeding rate for ulcers with persistent oozing treated medically varies from 10% (UCLA CURE) to 27% (Hong Kong). Monotherapy with probes or epinephrine injection reduces the rebleeding rate to less than 5%. In patients with oozing, the bleeding arteries may be small and the outcomes better than those in patients with active

arterial bleeding.119 Patients with oozing and no other stigmata of hemorrhage (e.g., a clot or NBVV) can be treated effectively with epinephrine injection alone and have an added benefit from combination therapy. After successful endoscopic hemostasis, patients with oozing and no other stigmata do not benefit from administration of a high-dose PPI.

299

300

Section III  Symptoms, Signs, and Biopsychosocial Issues

A

B

C

D

E

F

G

H

Figure 19-12.  A, Endoscopic treatment of a duodenal ulcer with an adherent clot. B, The clot was injected with epinephrine, followed by piecemeal snare polypectomy to trim away the clot (C-E), after which an underlying vessel was revealed (F; arrow). G, H, Two endoscopic hemoclips are placed across the visible vessel.

Clean-Based Ulcers Patients with clean-based ulcers at endoscopy have a rebleeding rate of less than 5% and therefore do not require endoscopic therapy. If the patient has a clean-based gastric ulcer, biopsies from the ulcer edge should be con­ sidered to exclude underlying malignancy as well as to assess for H. pylori infection (see Chapter 54). These patients can be fed after the endoscopy and treated with oral acid suppression medication; they do not require continued hospitalization unless indicated for other medical problems.

Assessing for Helicobacter pylori Infection

In a patient with a bleeding gastric or duodenal ulcer, endo­ scopic mucosal biopsies of the normal appearing antrum and midbody greater curvature should be obtained to assess for the presence of H. pylori infection. Biopsies can be obtained safely after successful endoscopic hemostasis; however, bleeding reduces the sensitivity of rapid urease testing (see Chapter 50).

Pharmacologic Therapy

Acid Suppression Medication In vitro studies have shown that a luminal gastric pH higher than 6.8 is required for normal clotting function (platelet aggregation and fibrin formation) and that a pH less than 5.4 almost abolishes platelet aggregation and plasma coagula­ tion.120 Platelet aggregates lyse at an acidic pH, an effect that is enhanced by the presence of pepsin. Therefore, reducing the risk of acute bleeding and rebleeding from a peptic ulcer is theoretically possible by maintaining a gastric pH higher than 6. Intravenous H2 receptor antagonists can raise the intragastric pH acutely, but tolerance to these agents devel­ ops rapidly and the pH usually returns to 3 to 5 within 24 hours. Several studies have shown that in normal subjects, administration of an intravenous PPI can consistently keep gastric pH higher than 4 (and often 6) over a 72-hour infu­ sion in contrast to an intravenous H2 receptor antago­ nist.121,122 Trials of intravenous H2 receptor antagonists for the prevention of recurrent ulcer bleeding have shown no definite benefit.123,124

Chapter 19  Gastrointestinal Bleeding Several studies of PPIs have shown that these agents are effective in reducing rebleeding rates from peptic ulcer. In a study from India, patients with endoscopic high-risk stig­ mata of peptic ulcer bleeding (active bleeding, NBVV, clot, oozing) who did not undergo endoscopic hemostasis were randomized to omeprazole, 40 mg orally twice daily, or placebo. The rebleeding rate in the omeprazole-treated group was 11% compared with 36% in the placebo-treated group (P < 0.001).125 Another study from the same investiga­ tors showed that omeprazole, 40 mg orally twice daily for five days, decreased the rebleeding rate after endoscopic hemostasis with injection therapy for ulcers with active bleeding, an NBVV, or a clot from 21% in the placebotreated group to 7% in the oral omeprazole-treated group (P = 0.02).126 In a study from Hong Kong, patients who had undergone successful endoscopic hemostasis for active bleeding or an NBVV were randomized to high-dose intra­ venous omeprazole, 80-mg bolus, followed by 8 mg/hour or placebo. The 30-day rebleeding rate was 6.7% in the omeprazole-treated group compared with 22.5% in the placebo-treated group (P < 0.05).127 The same investigators from Hong Kong found that the 30-day rebleeding rate in patients with an adherent clot or NBVV who received intra­ venous omeprazole alone was 12% compared with 1% in those who received intravenous omeprazole and underwent endoscopic hemostasis (P < 0.05).128 Another study from Hong Kong found that starting intravenous omeprazole before upper endoscopy in patients with UGI bleeding resulted in a decrease in the number of high-risk stigmata found and the need for endoscopic therapy, but no differ­ ence in clinical outcomes such as the number of units trans­ fused, frequency of recurrent bleeding, or rates of surgery and death.129 Systematic and Cochrane reviews of the clinical effective­ ness and cost-effectiveness of PPIs in acute UGI bleeding by Leontiadis and colleagues have found that PPI treatment initiated after endoscopic diagnosis of peptic ulcer bleeding significantly reduces the rates of rebleeding and surgery compared with placebo or H2 receptor blockers and that the benefit is more pronounced in Asian than in non-Asian studies.130-132 PPI treatment was associated with decreased mortality in the Asian studies as well as in patients with high-risk endoscopic stigmata. The initiation of PPI treat­ ment prior to endoscopy significantly reduced the pro­ portion of patients with stigmata of recent hemorrhage at index endoscopy compared with placebo or H2 receptor blockers but did not reduce the rate of mortality, rebleeding, or surgery. Some caution is advised in generalizing the results of PPI trials in Asian patients with peptic ulcer hemorrhage to heterogeneous non-Asian populations. The Asian patients are generally more responsive than heterogeneous popula­ tions or whites to PPIs.133 Asian patients have a smaller average parietal cell mass, are slower metabolizers of PPIs, and often have H. pylori infection, all of which increase the effectiveness of PPIs. These factors may explain the lower mortality rates in Asians compared with non-Asians in meta-analyses of PPI trials for peptic ulcer hemorrhage. A number of issues related to PPIs and UGI bleeding are still unresolved. Whether a PPI should be given before or after endoscopy is uncertain. Although some small random­ ized studies have not shown pre-endoscopy administration of a PPI to improve clinical outcomes (although the number of high-risk stigmata that require treatment is reduced), most modeling studies have suggested that pre-endoscopy administration of a PPI is cost-effective.14,16,17,129,132 Intrave­ nous administration of a PPI by high-dose continuous drip

or intermittent bolus infusion is also a matter of contro­ versy, but most data favor continuous drip, with small comparative studies suggesting that continuous infusion decreases the rate of rebleeding and need for surgery com­ pared with intermittent dosing.134 Whether administration of an oral PPI is as effective as intravenous administration is unclear, although studies have shown that high-dose oral PPI administration (e.g., omeprazole, 40 mg twice daily) reduces rebleeding to rates that would be expected from endoscopic hemostasis. In addition, studies have shown that the increase in intragastric pH with high-dose oral PPI administration is almost identical (although delayed by one hour) to that with intravenous PPI administration.125,135 Whether intravenous administration of a PPI alone is suf­ ficient therapy (without endoscopic hemostasis) in patients with recent UGI bleeding and some stigmata of hemorrhage, such as an NBVV, oozing, or clot, is controversial. In an Asian study, Sung and colleagues reported that the 30-day rebleeding rate with intravenous PPI administration alone (12%) is similar to that in previous studies of endoscopic hemostasis, although they also found that the rebleeding rate with a combination of endoscopic therapy and an intra­ venous PPI is even lower (1%).136 Finally, because almost all the major studies of PPIs in acute peptic ulcer bleeding have been conducted in Asian populations, who have a greater pharmacodynamic response to PPIs than non-Asian populations, studies in non-Asian populations are needed to confirm the Asian data. One large international study has confirmed the benefit of high-dose intravenous PPI admin­ istration in high-risk patients with active arterial bleeding, a NBVV, or an adherent clot in a study of a predominantly white population.137 Somatostatin and Octreotide A meta-analysis has suggested that intravenous administra­ tion of somatostatin or its long-acting form octreotide decreases the risk of rebleeding from peptic ulcers when compared with placebo or an H2 receptor blocker.138 The proposed mechanisms of action include reductions in splanchnic and gastroduodenal mucosal blood flow, decreases in gastrointestinal motility, inhibition of gastric acid secretion, inhibition of pepsin secretion, and gastric mucosal cytoprotective effects. These drugs have not been studied, however, in the era of endoscopic therapy or PPI use and therefore cannot be considered for routine use.139 Somatostatin or octreotide can be considered in patients with severe ongoing bleeding who are not responsive to endoscopic therapy, an intravenous PPI, or both, and are not surgical candidates, although their effectiveness in these patients is uncertain. Intravenous octreotide may also be useful in patients with portal hypertension and peptic ulcer hemorrhage as an adjunct to endoscopic hemostasis and a PPI (see Chapter 90).

Second-Look Endoscopy

Routine repeat, or second-look, endoscopy 24 hours after initial endoscopic hemostasis, with additional endoscopic hemostasis if persistent high-risk endoscopic stigmata are found, has been proposed as a way to improve patient out­ comes. The results of four prospective randomized trials have yielded conflicting results, with no benefit in the majority of studies.140-143 Therefore, routine second-look endoscopy is not recommended for most patients with peptic ulcer bleeding.100 The exception to this recommenda­ tion is for patients who had an incomplete initial endo­ scopic examination because of excessive blood that obscured the view or technical problems with hemostasis. Patients

301

302

Section III  Symptoms, Signs, and Biopsychosocial Issues with clinically significant rebleeding also should undergo a second endoscopy. For patients who are treated with epinephrine injection alone, second-look endoscopy and repeat treatment are considered a routine part of the man­ agement.119 Endoscopic retreatment with administration of more epinephrine for major stigmata of hemorrhage or oozing is required in 20% to 25% of patients with peptic ulcer hemorrhage.

Rebleeding after Endoscopic Treatment

The risk of rebleeding from peptic ulcers that started bleeding in the outpatient setting and required endoscopic hemostasis is greatest in the first 72 hours after diagnosis and treatment. Such patients, therefore, should be kept on a PPI in high doses for at least 72 hours after endoscopic hemostasis, after which they can be switched to a standard dose. Before the widespread use of intravenous PPIs, the rebleeding rate after endoscopic hemostasis of actively bleeding ulcers or those with a NBVV was as high as 30%; now, with the use of PPIs, the rate is less than 10%. The difference between ulcer hemorrhage that starts in the outpatient setting and hemorrhage that starts in the inpatient setting is substantial (Table 19-8). Because the time to rebleeding can be much longer for inpatient (than outpatient) ulcer hemorrhage and the risk of rebleeding is high, combination endoscopic hemostasis and high-dose intravenous PPI administration for at least 72 hours ought to be considered. Further studies are warranted in this highrisk group to define the optimal management. If rebleeding from a peptic ulcer is severe, an urgent repeat endoscopy (rather than surgery) should be performed. A large, well-designed, randomized trial from Hong Kong has found that when endoscopic hemostasis is repeated in patients with hemodynamically significant rebleeding after initial endoscopic hemostasis, 73% of patients achieve sus­ tained hemostasis and do not require surgery.144 The overall mortality rate was the same in both groups, but the rate of complications was significantly higher in the surgical group. Factors that predicted failure of endoscopic retreatment included an ulcer size of at least 2 cm and hypotension on initial presentation.

Angiography and Surgery

Patients with recurrent bleeding despite two sessions of endoscopic hemostasis should be considered for angio­ graphic embolization or surgical therapy. Interventional

Table 19-8  Comparison of Outpatient and Inpatient Onset of Peptic Ulcer Bleeding* Onset PARAMETER Frequency (%) American Society of Anesthesiologists comorbidity score† Time to rebleeding (%)   ≤72 hr   4-7 days   8-30 days   >30 days

OUTPATIENT

INPATIENT

80-90 ≤3

10-20 >3

70-80 10-15 1-5 0

40-50 15-20 15-20 5-10

*Data from the UCLA CURE database. † 1 point signifies a healthy person; 5 points signifies high likelihood of mortality within 24 hr. CURE, Center for Ulcer Research Education; UCLA, University of California, Los Angeles

angiography with embolization has become widely avail­ able. Several retrospective series have reported no signifi­ cant difference between angiography with embolization and surgery in rates of rebleeding and mortality, despite the older age of and more serious medical problems in patients treated by angiography than in those treated by surgery.145,146 These studies suggest that angiography can be considered after failure of endoscopic therapy. If embolization therapy does not control the bleeding, surgery remains an option. Acute surgical intervention is indicated for patients who have exsanguinating bleeding and those who cannot be medically resuscitated. Surgery should also be considered if the endoscopist does not feel comfortable treating a large or pulsating visible vessel (e.g., one in a deep, posterior duodenal ulcer that may represent the gastroduodenal artery). Another indication for surgery is a locally confined bleeding malignant ulcerated mass. The ultimate choice between interventional angiography and surgery often depends on local availability and exper­ tise, as well as the patient’s medical suitability to undergo surgery safely.

Immediate Postendoscopic Management

High-Risk Endoscopic Stigmata Patients who have undergone endoscopic hemostasis for active arterial bleeding, an NBVV, or an adherent clot should be observed in the hospital for 72 hours while they receive high-dose intravenous infusions of a PPI. After suc­ cessful endoscopic treatment and recovery from moderate sedation, the patient can be started on a liquid diet, with subsequent advancement of the diet. Ideally, NSAIDs or warfarin should be withheld for as long as safely possible to help allow the ulcer to heal. For patients with severe atherosclerotic cardiovascular disease who require aspirin, however, a dose of 81 mg/day should be started within seven days to prevent stroke and myocardial infarction. Intermediate-Risk Patients Patients with oozing from an ulcer and no other stigmata (e.g., spurting, NBVV, or clot), severe comorbidity, or shock on presentation should undergo endoscopic hemostasis. Ini­ tiation of an oral PPI and observation in the hospital for 24 to 48 hours after endoscopic hemostasis are recommended. Such patients do not benefit from a high-dose intravenous PPI after successful endoscopic hemostasis. Low-Risk Endoscopic Stigmata Patients with a clean-based ulcer or flat spot in the ulcer base can generally resume a normal diet immediately, begin an oral PPI once daily, and be discharged from the emer­ gency department or hospital when stable.117 Several studies have shown that patients at low risk for rebleeding on the basis of these endoscopic findings and a stable clinical status can avoid hospitalization entirely or be discharged early.9,10,64,147 Generally these patients are young and hemo­ dynamically stable, have no severe coexisting medical ill­ nesses, a hemoglobin level higher than 10 mg/dL, and normal coagulation parameters, had the onset of ulcer bleed­ ing outside the hospital, have good social support systems at home in case rebleeding occurs, and have a clean-based ulcer or ulcer with a flat spot on initial endoscopy. Aspirin and Clopidogrel For patients with severe atherosclerotic cardiovascular disease, aspirin and clopidogrel may need to be restarted soon after hemostasis is achieved to prevent myocardial infarction or stroke. The patient’s cardiologist or neurologist should be consulted in these high-risk cases.

Chapter 19  Gastrointestinal Bleeding Prevention of Recurrent Ulcer Bleeding

Helicobacter pylori Infection All patients with peptic ulcer bleeding should be tested for H. pylori infection and, if the result is positive, should receive antibiotic therapy in standard fashion (see Chapter 50).78 One caveat is that bleeding can lead to a false-negative rapid urease test result, and the patient may need to undergo an alternative method of testing for H. pylori in this setting. Antibiotic therapy does not need to be started urgently and can be initiated on an outpatient basis when the patient has resumed a normal diet. Patients who are H. pylori–positive and who will need long-term PPI treatment because of the concomitant need for aspirin or other NSAID do not neces­ sarily need to be treated for H. pylori infection, because recurrent ulceration will be prevented by the PPI. In patients who are found to have an H. pylori–induced ulcer, confir­ mation of the eradication of H. pylori after treatment is recommended. Aspirin, Other Nonsteroidal Anti-inflammatory Drugs, and Clopidogrel Ideally, patients with ulcer bleeding caused by aspirin or another NSAID should stop the drug. If the patient is also positive for H. pylori, the organism should be eradicated with antibiotics (see Chapter 50).148 In patients with a history of ulcer bleeding who are H. pylori–positive and need to continue taking low-dose aspirin (81 mg daily), eradication of H. pylori alone results in ulcer rebleeding rates similar to those associated with daily PPI therapy (if H. pylori is not eradicated).149 By con­ trast, in patients with a history of ulcer bleeding who are H. pylori–positive and need to continue full-dose NSAID therapy, eradication of H. pylori alone leads to a signifi­ cantly higher rebleeding rate than use of a daily PPI in conjunction with the NSAID. In patients with ulcer bleeding who do not have H. pylori infection but who need to con­ tinue daily aspirin, co-therapy with a daily PPI significantly reduces the rebleeding rate compared with placebo in com­ bination with aspirin.150 Patients who require an antiplatelet medication and have a history of ulcer bleeding will have less chance of recurrent bleeding if they take aspirin 81 mg and a PPI daily com­ pared with clopidogrel alone.151 Patients who require an NSAID after an ulcer bleed may be considered for a selective cyclooxygenase-2 (COX-2) inhibitor. Selective COX-2 inhibitors cause fewer ulcers than nonselective NSAIDs but are associated with a greater rate of cardiovascular complications. Because selective COX-2 inhibitors result in rebleeding rates similar to those associated with NSAID and PPI cotherapy, their use may not be worth the cardiovascular risk.152

Repeat Endoscopy to Confirm Gastric Ulcer Healing

Repeat upper endoscopy should be considered in patients with a gastric ulcer after 6 to 10 weeks of acid suppressive therapy to confirm healing of the ulcer and absence of malig­ nancy (see Chapter 54). In areas of the world in which the population is at intermediate risk for gastric cancer, 2% to 4% of repeat upper endoscopies to confirm ulcer healing have been reported to disclose gastric cancer.153-155 Some experts have suggested that when the index endoscopy with biopsies is negative for malignancy and the ulcer appears benign endoscopically, a follow-up endoscopy is unneces­ sary.156 A small retrospective study has found that when gastric cancer is detected on repeat endoscopy to evaluate gastric ulcer healing, survival is no better than that for patients who did not undergo the recommended follow-up endoscopy.153

OTHER CAUSES Esophagitis

Patients with severe erosive esophagitis can present with hematemesis or melena. A multivariate analysis from a center in France, in which 8% of all UGI bleeding was caused by erosive esophagitis, found that independent risk factors for bleeding esophagitis were grade 3 or 4 (moderate to severe) esophagitis by the Savary-Miller grading system (see Chapter 43), cirrhosis, a poor performance status, and anticoagulant therapy.157 A history of heartburn was obtained in only 38% of patients. Severe bleeding from gastroesopha­ geal reflux–induced esophagitis is treated medically with a PPI (see Chapter 43). Upper endoscopy is critical to diag­ nosing severe erosive esophagitis, but endoscopic therapy generally has no role unless a focal ulcer with a stigma of recent hemorrhage is found. These patients should be treated with a daily PPI for 8 to 12 weeks and undergo repeat endoscopy to exclude underlying Barrett’s esophagus (see Chapter 44). Patients can sometimes present with mild UGI bleeding from esophagitis not related to gastroesophageal reflux disease, such as infections (e.g., Candida, herpes simplex virus, cytomegalovirus) or pill-induced esophagitis. Endos­ copy with biopsies and brushings is critical for making these diagnoses and determining the appropriate pharma­ cologic therapy (see Chapter 45).

Ulcer Hemorrhage in Hospitalized Patients

Hemorrhage from an ulcer or erosions in hospitalized patients typically falls into two categories. The classic cause is stress-related mucosal injury (SRMI, or stress ulcers), which is characterized by diffuse bleeding from erosions and superficial ulcers. The second category is inpatient ulcers, which are large, focal, chronic-appearing ulcers that are painless and present with severe inpatient UGI hemorrhage manifested by hematochezia, melena, or bloody emesis. On emergency endoscopy, focal inpatient ulcers often are actively bleeding or demonstrate a visible vessel or adherent clot and are marked by high rebleeding rates, despite combination endoscopic therapy, and delayed healing on a high-dose PPI. SRMI occurs in the UGI tract of severely ill inpatients in an ICU and is likely caused by a combination of decreased mucosal protection and mucosal ischemia. SRMI usually occurs in the stomach but can also be seen in the duode­ num, esophagus, and even rectum. Diffuse oozing is common, and patients have a poor prognosis and high rebleeding rate, often related to impaired wound healing and multiple organ failure. Bleeding from SRMI is now uncommon, with a frequency of approximately 1.5% patients in an ICU. The two main risk factors are severe coagulopathy and mechanical ventila­ tion for longer than 48 hours.158 The frequency of clinically significant GI bleeding with either or both of these risk factors is 3.7% compared with 0.1% when neither risk factor is present. Other proposed risk factors include a history of UGI bleeding, sepsis, an ICU admission longer than seven days, occult GI bleeding for more than five days, and treatment with high-dose glucocorticoids. ICU patients with risk factors for bleeding are the main target groups for pharmacologic prevention of bleeding SRMI. Therapy with an H2 receptor antagonist has been shown to decrease the rate of clinically significant bleeding in ICU patients at high risk of SRMI.159 Well-designed and adequately powered studies that compare H2 receptor block­ ers and PPIs are few in number, but one large multicenter study found that prophylactic treatment with oral omepra­ zole or intravenous cimetidine results in similar bleeding

303

304

Section III  Symptoms, Signs, and Biopsychosocial Issues rates but that omeprazole is more effective than cimetidine in maintaining the luminal gastric pH above 4.160 A potential harmful effect of gastric acid suppression to prevent stress ulcers is that a decrease in gastric pH may allow proli­ feration of bacteria in the stomach and the potential for aspiration and ventilator-associated pneumonia; however, randomized trials in which acid suppression (with an H2 receptor blocker or antacids) and sucralfate (which does not lower gastric pH) were compared have not shown convincingly that lowering gastric pH increases the risk of pneumonia.161,162 Generally, if a patient with SRMI or an inpatient ulcer is supported hemodynamically and medically, the lesion will heal as the patient’s overall medical status improves. Because SRMI is diffuse, endoscopic therapy is generally not feasible. By contrast, focal inpatient ulcer hemorrhage often requires endoscopic hemostasis for severe hemor­ rhage; however, rebleeding rates are higher and healing rates are slower than those in patients in whom bleeding starts before hospitalization (see Table 19-8).163,164 A study in which epinephrine injection plus hemoclip placement was compared with epinephrine injection plus MPEC in a cohort of patients who had a high frequency of in-hospital ulcers found a significantly lower rebleeding rate in the group that underwent injection and hemoclip placement.118 Figure 19-9 illustrates combination treatment using injec­ tion therapy, bipolar probe coagulation, and endoscopic hemoclipping for an inpatient ulcer hemorrhage.

Dieulafoy’s Lesion

A Dieulafoy’s lesion is a large (1- to 3-mm) submucosal artery that protrudes through the mucosa, is not associated with a peptic ulcer, and can cause massive bleeding. It usually is located in the gastric fundus, within 6 cm of the gastroesophageal junction, although lesions in the duode­ num, small intestine, and colon have been reported. The cause is unknown and congenital and acquired (related to mucosal atrophy or an arteriolar aneurysm) causes are thought to occur. Dieulafoy’s lesion can be difficult to identify at endoscopy because of the intermittent nature of the bleeding; the over­ lying mucosa may appear normal if the lesion is not bleed­ ing. An NBVV or adherent clot without an ulcer may be seen on endoscopy. If a massive UGI bleed seems to be emanating from the stomach, a careful inspection of the proximal stomach should be carried out to look for a protuberance that might be a Dieulafoy’s lesion. Endoscopic Doppler ultrasound has been used to help identify a Dieulafoy’s lesion that is not visualized on endoscopy.165 Because of the difficulty of identifying the bleeding site, we recommend that if a Dieulafoy’s lesion is found and treated, the site be marked with submucosal injection of ink to tattoo the area in case of rebleeding and the need for retreatment. Endoscopic hemostasis of a Dieulafoy’s lesion can be performed with injection therapy, a thermal probe, or clip device or by band ligation.165-169,170 Large case series have reported an initial hemostasis rate of approximately 90%, with the need for surgery in 4% to 16% of cases.171 Rebleed­ ing after successful hemostasis appears to be rare. Although all the endoscopic hemostasis techniques seem to be effec­ tive, perforation and delayed rebleeding have been reported after band ligation (see Chapter 36).

Mallory-Weiss Tears

Mallory-Weiss tears are mucosal or submucosal lacerations that occur at the gastroesophageal junction and usually extend distally into a hiatal hernia (Fig. 19-13). Patients generally present with hematemesis or coffee-ground emesis

Figure 19-13.  Endoscopic appearance of a Mallory-Weiss tear with mild oozing. Note that the tear starts at the gastroesophageal junction (large arrow) and extends distally into the hiatal hernia (small arrow).

and typically have a history of recent nonbloody vomiting followed by hematemesis, although some patients do not recall any vomiting. The tear is thought to result from increased intra-abdominal pressure, possibly in combina­ tion with a shearing effect caused by negative intrathoracic pressure above the diaphragm, which is often related to vomiting in patients with a history of alcohol abuse. Mallory-Weiss tears have been reported in patients who vomit while taking a bowel purge before colonoscopy.171 Endoscopy usually reveals a single tear that begins at the gastroesophageal junction and extends several millimeters distally into a hiatal hernia sac. Occasionally, more than one tear is seen. A retroflexed view in the stomach may provide better visualization than the forward viewing position. The bleeding stigmata of Mallory-Weiss tears can include a clean base, oozing, or active spurting. Usually, the bleed­ ing is self-limited and mild, but occasionally it can be severe. Superficial (mucosal) Mallory-Weiss tears can start healing within hours and can heal completely within 48 hours. Although approximately 50% of patients hospitalized with UGI bleeding from a Mallory-Weiss tear receive blood transfusions, the tear manifests as mild, self-limited hematemesis in most patients, who do not seek medical care.172 The rebleeding rate among patients hospitalized for a Mallory-Weiss tear is approximately 10%; risk factors for rebleeding include shock at presentation and active bleed­ ing at endoscopy.173 Because of the risk of continued and recurrent bleeding, patients with active bleeding from a Mallory-Weiss tear should undergo endoscopic therapy, which can be performed successfully with epinephrine injection, MPEC, hemoclip placement, or band ligation. Randomized trials that compared MPEC and medical therapy with an H2 receptor antagonist have found that endoscopic therapy reduces the rates of rebleeding, blood transfusions, and emergency surgery.174 Our current endoscopic technique for treating actively bleeding Mallory-Weiss tears in patients without portal hypertension or esophageal varices is to apply endoscopic hemoclips to stop the bleeding and close the tear. If hemo­ clips are not available, MPEC at a low power setting and with light pressure for one to two seconds is recommended. The management of patients with esophageal varices caused by portal hypertension who also have a Mallory-Weiss tear should be targeted toward the esophageal varices, with

Chapter 19  Gastrointestinal Bleeding

Figure 19-14.  Endoscopic appearance of Cameron’s lesions. Note that these linear ulcerations (arrows) are located at the distal end of a hiatal hernia.

esophageal band ligation or variceal sclerotherapy (see later and Chapter 90). Patients with a Mallory-Weiss tear are often treated with antiemetics if they have nausea or vomiting and a PPI to accelerate mucosal healing. The PPI may allow better natural hemostasis by raising the gastric pH to improve coagulation and possibly speed the healing of the tear, but this approach has not been well studied and most MalloryWeiss tears heal within days anyway. Therefore, long-term treatment with a PPI is generally not needed.

Cameron’s Lesions

Cameron’s lesions are linear erosions or ulcerations in the proximal stomach at the end of a large hiatal hernia, near the diaphragmatic pinch (Fig. 19-14).175 Cameron’s lesions are thought to be caused by mechanical trauma and local ischemia as the hernia moves against the diaphragm and only secondarily by acid and pepsin. They can be a source of acute UGI bleeding but more commonly may present as slow GI bleeding and iron deficiency anemia. Cameron’s lesions are a common cause of obscure GI bleeding (see later) and not uncommonly are missed by an unsuspecting endoscopist. Endoscopic management has been reported.176 The long-term medical management is usually with iron supplements and an oral PPI (see Chapter 36).177,178 Occasionally surgical repair of the hiatal hernia may be needed.

Upper Gastrointestinal Malignancy

Malignancy accounts for 1% of severe UGI bleeds. The tumors are usually large, ulcerated masses in the esophagus, stomach, or duodenum. Endoscopic hemostasis with MPEC, laser, injection therapy, or hemoclips can temporarily control acute bleeding in most patients and allow time to determine the appropriate long-term management.179,180 Patients with an ulcerated subepithelial mass (which usually is a gastrointestinal stromal tumor or leiomyoma) should undergo surgical resection of the mass to prevent rebleeding and to prevent the risk of metastasis. Angiogra­ phy with embolization should be considered for patients with severe UGI bleeding caused by malignancy who do not respond to endoscopic therapy. External beam radiation can

Figure 19-15.  Endoscopic appearance of gastric antral vascular ectasia (GAVE), or watermelon stomach. The pattern seen in this view is considered classic, with rows of ectatic mucosal blood vessels emanating from the pylorus.

provide palliative hemostasis for patients with bleeding from advanced gastric or duodenal cancer (see Chapter 54).

Gastric Antral Vascular Ectasia

Gastric antral vascular ectasia (GAVE), also described as watermelon stomach, is characterized by rows or stripes of ectatic mucosal blood vessels that emanate from the pylorus and extend proximally into the antrum (Fig. 19-15). The cause is uncertain, and the lesion may represent a response to mucosal trauma from contraction waves in the antrum. GAVE has been associated with cirrhosis and scleroderma (see Chapters 35 and 90). Patients with GAVE who do not have portal hypertension demonstrate linear arrays of angio­ mas (classic GAVE), whereas those with portal hypertension have more diffuse antral angiomas.181 The diffuse type of antral angiomas and occasionally classic GAVE are some­ times mistaken for gastritis by an unsuspecting endoscopist. Such cases are a common cause of obscure GI bleeding in referral centers (see later).48 Patients usually present with iron deficiency anemia or melena, with a mildly decreased hematocrit value sugges­ tive of a slow UGI bleed. GAVE is most commonly reported in older women181 and also seems to be more common in patients with end-stage renal disease. Endoscopic hemostasis with thermal heat modalities such as laser, MPEC, or argon plasma coagulation has been used successfully. Endoscopic hemostasis and ablation with thermal modalities can result in good palliation with an increase in the hematocrit value and a decrease in the need for blood transfusions and hospitalization.181,182 Usually, several sessions, approximately four to eight weeks apart, are needed to achieve eradication of the lesions and a reduction in bleeding from the antral ectasias. Endoscopic therapy with argon plasma coagulation has been shown to be equally (80%) effective in cirrhotic and noncirrhotic patients with GAVE.183 Pilot studies have demonstrated that mucosal band ligation, radiofrequency ablation, and cryo­ therapy can also lead to eradication of GAVE in selected patients.184-186

305

306

Section III  Symptoms, Signs, and Biopsychosocial Issues Placement of a transjugular intrahepatic portosystemic shunt (TIPS) in patients with portal hypertension and cir­ rhosis does not decrease bleeding from GAVE or diffuse antral angiomas. Patients who have ongoing severe chronic bleeding from GAVE rarely may require surgical antrectomy to control symptoms (see Chapters 36 and 90).187

Portal Hypertensive Gastropathy

Portal hypertensive gastropathy (PHG) is caused by increased portal venous pressure and severe mucosal hyperemia that results in ectatic blood vessels in the proximal gastric body and cardia and oozing of blood. Less severe grades of PHG appear as a mosaic or snakeskin pattern and are not asso­ ciated with bleeding.188 Usually, patients with severe PHG present with chronic blood loss, but they occasionally can present with acute bleeding. Severe PHG with diffuse bleeding is treated by measures that decrease portal pressure, usually with β-adrenergic receptor blockers or possibly with placement of a TIPS or surgical portacaval shunt. Endoscopic management has no role unless an obvious focal bleeding site is identified. The best treatment is liver transplantation (see Chapters 36 and 90).

Hemobilia

Hemobilia may occur in patients who have experienced liver trauma, undergone a liver biopsy, manipulation of the hepatobiliary system, as occurs with endoscopic retrograde cholangiopancreatography (ERCP), percutaneous transhe­ patic cholangiography, or TIPS, or have hepatocellular carcinoma or a biliary parasitic infection.189 Patients may present with a combination of GI bleeding and elevated liver biochemical test levels. The diagnosis can be confirmed by using a side-viewing duodenoscope to identify bleeding from the ampulla (Fig. 19-16). Ongoing or recurrent bleed­ ing is treated with arterial embolization via arteriography.

Hemosuccus Pancreaticus

Hemosuccus pancreaticus is a rare form of UGI bleeding that occurs most commonly in patients with acute pancreatitis,

chronic pancreatitis, pancreatic pseudocyst, or pancreatic cancer or after ERCP with pancreatic duct manipulation. It can also result from rupture of a splenic artery aneurysm into the pancreatic duct.190 CT can demonstrate pancreatic pathology if previously unsuspected. Endoscopy with a side-viewing duodenoscope reveals blood coming out of the ampulla. Management of severe hemorrhage is usually with angiographic embolization or surgery.

Postsphincterotomy Bleeding

Bleeding following endoscopic sphincterotomy occurs in approximately 2% of patients (see Chapter 40).191 Potential risk factors include coagulopathy, use of anticoagulants, portal hypertension, renal failure, and the type and length of sphincterotomy. Successful hemostasis of postsphinc­ terotomy bleeding is usually achieved with endoscopic methods such as injection of epinephrine, hemoclips, or MPEC (see Chapter 40).

Aortoenteric Fistula

Bleeding from an aortoenteric fistula is usually acute and massive, with a high mortality rate.192 A primary aortoen­ teric fistula is a communication between the native ab­ dominal aorta (usually an atherosclerotic abdominal aortic aneurysm) and, most commonly, the third portion of the duodenum.193 Often, a self-limited herald bleed occurs hours to months before a more severe, exsanguinating bleed. Occasionally, the diagnosis of an aortoenteric fistula is sus­ pected by a history of an abdominal aortic aneurysm or by palpation of a pulsatile abdominal mass. The diagnosis can be difficult to make on endoscopy in the absence of active bleeding. Demonstration of an aortic aneurysm on ab­ dominal CT scan (with intravenous contrast) suggests the diagnosis of a fistula.49 Secondary aortoenteric fistulas usually occur between the small intestine and an infected abdominal aortic surgical graft. The fistula typically occurs between the third portion of the duodenum and the proximal end of the graft but may occur elsewhere in the GI tract. The fistula usually forms between three and five years after graft placement. Patients often experience a herald bleed that is mild and self-limited, and occasionally intermittent, before massive bleeding occurs.194 A secondary fistula can also occur between the third part of the duodenum and an endovascular stent, in which case the fistula occurs as a result of pressure from the stent against the duodenum, infection of the stent, or pos­ sibly expansion of the native aneurysm.195 Patients with an acute UGI bleed and a history of an aortic aneurysm repair should undergo urgent CT with intrave­ nous contrast, push enteroscopy to evaluate the third portion of the duodenum for compression or blood, as well as to exclude other bleeding sources, and a vascular surgery consultation. CT may show inflammation around the graft and may demonstrate the fistula. Surgical treatment is required to remove the infected graft. Therapeutic endos­ copy plays no role in the management of bleeding from an aortoenteric fistula (see Chapter 36).

VARICES

Figure 19-16.  Endoscopic appearance of the ampulla of Vater and hemobilia. Note the fresh red blood exuding from the ampulla of a patient who earlier that day had undergone a percutaneous liver biopsy.

Variceal hemorrhage is an important cause of UGI bleeding and is discussed in more detail in Chapter 90. Esophageal variceal bleeding related to portal hypertension is the second most common cause of severe UGI bleeding (after peptic ulcer disease). The acute mortality rate with each bleed is approximately 30%, and the long-term survival rate is less than 40% after one year with medical management alone.196 Despite advances in medical therapy, endoscopic hemostasis, and portosystemic shunt procedures, overall

Chapter 19  Gastrointestinal Bleeding long-term survival rates have not improved for patients with variceal bleeding. Liver transplantation, however, can improve survival in selected patients. Survival in nontrans­ planted patients with variceal bleeding is heavily influ­ enced by the severity of underlying liver disease, with poorer survival rates for patients with Child (or Child-Pugh) class C cirrhosis than for those with Child class A or B cirrhosis (see Chapter 90). Bleeding gastric varices are a difficult therapeutic problem because, in contrast to bleeding esophageal varices, most available nonsurgical treatments are ineffective, except when isolated gastric varices are found without accompany­ ing esophageal varices, as occurs with splenic vein throm­ bosis and often in association with pancreatitis or pancreatic cancer. The diagnosis of splenic vein thrombosis can be made with Doppler ultrasound, magnetic resonance imaging, or angiography. Bleeding from gastric varices caused by splenic vein thrombosis is treated by splenectomy.

Medical Management of Acute Variceal Bleeding

Somatostatin and its long-acting analog octreotide cause selective splanchnic vasoconstriction and lower portal pressure, without causing the cardiac complications seen with vasopressin (even in combination with nitroglycerin). Studies have shown mixed results as to whether soma­ tostatin is more effective than placebo in managing vari­ ceal bleeding, but it seems to be at least as effective as vasopressin and much safer. A meta-analysis has shown that vasoactive drugs (e.g., octreotide, somatostatin, terli­ pressin [a long-acting vasopressin analog]) are as effective as sclerotherapy for controlling variceal bleeding and cause fewer adverse events.19 No studies have shown a survival benefit to vasopressin or somatostatin in patients with variceal bleeding. Given the potential ability of octreotide to control acute variceal hemorrhage, its low toxicity, and its availability in the United States (unlike somatostatin and terli­pressin), octreotide appears to be the pharmacologic drug of choice as an adjunct to endoscopic therapy for the treatment of variceal hemorrhage. The dose of octreotide for acute variceal hemorrhage is a 50-µg bolus followed by a continuous infusion of 50  µg/hour for up to five days. Patients with a prolonged prothrombin time that does not correct with fresh frozen plasma may benefit from infusion of human recombinant factor VIIa. In one uncontrolled trial, a single 80-µg/kg dose of recombinant factor VIIa norma­ lized the prothrombin time in all 10 patients within 30 minutes, with immediate control of bleeding in all patients.197 In a large, randomized, placebo-controlled study, administration of recombinant factor VII in addition to endoscopic hemostasis decreased rebleeding rates in patients with Child class B and C cirrhosis who had bled from varices.198 Because recombinant factor VIIa is expen­ sive, its use should be reserved for patients with severe ongoing bleeding and irreversible coagulopathy, pending the results of additional clinical and cost-effectiveness studies. Up to 20% of cirrhotic patients who are hospitalized with GI bleeding have a bacterial infection at the time of admis­ sion to the hospital, and infection develops during the hos­ pitalization in up to 50%. Meta-analyses have suggested that administration of an antibiotic to cirrhotic patients with variceal bleeding is associated with a decrease in the rates of mortality and bacterial infections.199,200 The optimal type and duration of antibiotic is unknown. The most commonly prescribed antibiotics are fluoroquinolones, including oral norfloxacin, 400 mg twice daily, intravenous ciprofloxacin, 400 mg every 12 hours, intravenous levofloxacin, 500 mg

every 24 hours, and intravenous ceftriaxone, 1 g every 24 hours, administered for seven days.

Balloon Tamponade

Balloon tamponade of varices is seldom used now to control variceal bleeding; it may be used to stabilize a patient with massive bleeding prior to definitive therapy. Varices lie in the esophageal and gastric submucosa and are amenable to physical tamponade. Three types of tamponade balloons are available. The Sengstaken-Blakemore tube has gastric and esophageal balloons, with a single aspirating port in the stomach. The Minnesota tube also has gastric and esopha­ geal balloons and has aspiration ports in the esophagus and stomach. The Linton-Nicholas tube has a single large gastric balloon and aspiration ports in the stomach and esophagus. Most reports suggest that balloon tamponade provides initial control of bleeding in 85% to 98% of cases, but vari­ ceal rebleeding recurs soon after the balloon is deflated in 21% to 60% of patients.201 The major problem with tampon­ ade balloons is a 30% rate of serious complications, such as aspiration pneumonia, esophageal rupture, and airway obstruction. Patients should be intubated before placement of a tamponade balloon to minimize the risk of pulmonary complications. Clinical studies have not shown a significant difference in efficacy between vasopressin administration and balloon tamponade.

Endoscopic Sclerotherapy

Endoscopic variceal sclerotherapy involves injecting a sclerosant into or adjacent to esophageal varices. The most commonly used sclerosants are ethanolamine oleate, sodium tetradecyl sulfate, sodium morrhuate, and ethanol. Cyanoacrylate is a glue that when injected into esophageal or gastric varices, effectively stops bleeding, but it is difficult to use and is not approved by the U.S. Food and Drug Administration. Various techniques are used; their common goal is to achieve initial hemostasis and perform sclerotherapy on a weekly basis until all varices are obliterated. Esophageal varices are much more amenable than gastric varices to eradication with endoscopic therapy. Prospective randomized trials have shown mixed results but suggest improved immediate hemostasis and a reduc­ tion in acute rebleeding with sclerotherapy compared with medical therapy alone for bleeding esophageal varices.202-205 Hemostasis can be achieved in 85% to 95% of cases, with a rebleeding rate of 25% to 30%.206 Complications of endo­ scopic variceal sclerotherapy include esophageal ulcers, which can bleed or perforate, esophageal strictures, medi­ astinitis, pleural effusions, aspiration pneumonia, acute respiratory distress syndrome, chest pain, fever, and bacte­ remia and account in part for the use of esophageal variceal band ligation as the preferred endoscopic therapy for vari­ ceal bleeding.

Endoscopic Band Ligation

The technique of endoscopic band ligation is similar to that used for band ligation of internal hemorrhoids. A rubber band is placed over a varix, which subsequently undergoes thrombosis, sloughing, and fibrosis. Prospective, random­ ized, controlled trials have shown that endoscopic band ligation is as effective as sclerotherapy in achieving initial hemostasis and reducing the rate of rebleeding from esopha­ geal varices. Acute hemostasis generally can be achieved in 80% to 85% of cases, with a rebleeding rate of 25% to 30%. Subsequent studies have shown that band ligation is associ­ ated with fewer local complications, especially esophageal strictures, and requires fewer endoscopic treatment sessions

307

308

Section III  Symptoms, Signs, and Biopsychosocial Issues than sclerotherapy.206 A meta-analysis has reported that variceal band ligation reduces the rates of rebleeding, overall mortality, and death from bleeding compared with sclero­ therapy.207 Band ligation, however, may be more technically difficult to perform than sclerotherapy during active vari­ ceal bleeding. Devices used for band ligation allow up to 10 bands to be placed, without the need to remove the endo­ scope to reload the banding device. The strategy is to control active bleeding and place two bands on each esophageal variceal column, one distally near the gastroesophageal junction and another 4 to 6 cm proximally.

Transjugular Intrahepatic Portosystemic Shunt

Placement of a transjugular intrahepatic portosystemic shunt (TIPS) is an interventional radiologic procedure in which an expandable metal stent is placed via percutaneous insertion between the hepatic and portal veins, thereby cre­ ating an intrahepatic portosystemic shunt. TIPS is effective for the short-term control of bleeding gastroesophageal varices.208,209 Initially envisioned as a bridge to liver trans­ plantation, it has been used with increased frequency in nontransplantation situations. Randomized trials that have compared TIPS with endoscopic sclero­therapy suggest that TIPS is more effective for the long-term prevention of rebleeding.210 The main problems with TIPS are a rate of shunt occlusion of up to 80% (less with polytetrafluoro-eth­ ylene-coated stents) within one year and the development of new or worsening hepatic ence­phalopathy in approximately 20% of patients.211 TIPS does not prolong the survival of patients with variceal bleeding compared with endoscopic treatment. In the management of acute variceal bleeding, TIPS is reserved for patients who fail endoscopic treatment.

Portosystemic Shunt Surgery

A variety of portosystemic shunt operations can be per­ formed to reduce portal venous pressure. When compared with sclerotherapy, surgical shunts decrease the rebleeding rate significantly but do not improve survival.206,212-215 Some groups have suggested that survival can be improved by using a combination of endoscopic sclerotherapy and surgi­ cal portosystemic shunt rescue for those who rebleed despite sclerotherapy. Surgical shunts may be associated with hepatic encephalopathy and can make future liver trans­ plantation technically more difficult, but they have an advantage over endoscopic variceal therapy in reducing portal hypertension and treating gastric variceal bleeding. Surgical shunts are considered for selected patients who have failed endoscopic therapy and who are not expected to become candidates for liver transplantation (see Chapters 90 and 95).

LOWER GASTROINTESTINAL BLEEDING LGI bleeding generally signifies bleeding from the colon or anorectum. The annual incidence of LGI bleeding is approx­ imately 20 cases/100,000 population, with an increased risk in older adults.216 The rate of hospitalization for LGI bleed­ ing is lower than that for UGI bleeding. Most patients are older than 70 years. Patients usually present with painless hematochezia and a decrease in their hematocrit value but without orthostasis. If orthostasis is associated with hema­ tochezia, a briskly bleeding UGI source should be excluded (see earlier). Severe painless hematochezia results from a foregut source in approximately 15% of patients.217 The sites of origin within the GI tract of severe hematochezia at UCLA CURE are shown in Figure 19-17.

Small intestine 5% (n = 30) No source identified 3% (n = 18)

Colon 75% (n = 486)

UGI tract 17% (n = 113) Figure 19-17.  The frequencies of the sources of severe hematochezia in patients seen at the University of California, Los Angeles, Center for Ulcer Research Education. Note that in most cases (75%), severe hematochezia is from the colon, 17% is from an upper gastrointestinal (UGI) (esoph­agus, stomach, or duodenum) source, and 5% is from a small intestinal source.

Table 19-9  Causes of Severe Hematochezia (%) Study LESION Diverticulosis Colon cancer or polyps Colitis Ischemic colitis IBD Noninfectious colitis Infectious colitis Angioectasia Postpolypectomy Rectal ulcer Hemorrhoids Anorectal source (unspecified) Radiation colitis Other Unknown

reference 218

reference 219

UCLA CURE

30 18

33 21

30 6

17 NP NP NP

17 7 4 5

21 12 9 0

NP 7 6 NP NP 4

1 6 NP 1 20 3

0 3 8 6 14 0

0 8 16

0.5 3 0

3 6 0

CURE, Center for Ulcer Research Education; IBD, inflammatory bowel disease; NP, not provided; UCLA, University of California, Los Angeles.

Patients with LGI bleeding initially should be resuscitated medically. After they have been stabilized, they generally should undergo colonoscopy after a polyethylene glycol purge.22 In early reports, urgent colonoscopy resulted in a diagnosis in approximately 70% of cases.40,218 In more recent reports, however, the combination of urgent colonoscopy and, if necessary, push enteroscopy, anoscopy, and capsule endoscopy has resulted in a diagnosis in 95% of cases (see Fig. 19-4).217 The most common causes of LGI bleeding are shown in Table 19-9. Diverticulosis is generally the most common cause of acute LGI bleeding, occurring in approximately 30% of cases. Colonic polyps or cancer, colitis, and anorec­ tal disorders each account for approximately 20% of cases.219 In most cases, acute LGI bleeding will stop spontane­ ously, thereby allowing nonurgent diagnosis and treatment. For patients with ongoing or recurrent hematochezia, urgent diagnosis and treatment are required to control the bleeding. In a large series of patients at the UCLA Medical Center and Wadsworth Veterans Administration Hospital, 64% of patients with severe hematochezia required a therapeutic intervention to control continued bleeding or rebleeding22; 39% underwent endoscopic hemostasis, 1% underwent angiographic embolization, and 24% underwent surgery.

Chapter 19  Gastrointestinal Bleeding Table 19-10  Clinical Predictors of Severe Acute Lower Gastrointestinal Bleeding*

TOTAL risk POINTS† 0 1-3 ≥4

FREQUENCY (%)

RISK OF SEVERE BLEEDING (%)

NEED FOR SURGERY

MORTALITY rate (%)

HOSPITAL DAYS

mean NUMBER OF UNITS TRANSFUSED (PACKED RED BLOOD CELLS)

6 75 19

6 43 79

0 1.5 7.7

0 2.9 9.6

2.8 3.1 4.6

0 1 3

*Severe lower gastrointestinal bleeding is defined as continued bleeding within the first 24 hr of hospitalization (transfusion of two or more units of packed red blood cells and/or hematocrit value drop of 20% or more) and/or recurrent bleeding after 24 hr of stability (need for additional transfusions, further hematocrit value decrease of 20% or more, or readmission to the hospital for lower gastrointestinal bleed within 1 wk of discharge). † Risk factors (1 point each): aspirin use; more than two comorbid illnesses; heart rate ≥ 100 beats/min; nontender abdominal examination; rectal bleeding within the first 4 hr of evaluation; syncope; systolic blood pressure ≤ 115 mm Hg. Data from Strate LL, Saltzman JR, Ookubo R, et al. Validation of a clinical prediction rule for severe acute lower intestinal bleeding. Am J Gastroenterol 2005; 100:1821-7.

RISK FACTORS AND RISK STRATIFICATION

Nonselective NSAIDs increase the risk of LGI bleeding com­ pared with placebo.220,221 The main risk factors for NSAIDassociated LGI bleeding appears to be an age of 65 years or older and prior history of LGI bleeding.222 It is uncertain whether the use of long-term selective COX-2 inhibitors is associated with a lower risk of LGI bleeding than nonselec­ tive COX-2 inhibitors. Table 19-10 shows clinical factors that are predictive of severe LGI bleeding (defined as continued bleeding within the first 24 hours of hospitalization, with a transfusion requirement of at least two units of packed red blood cells or a decrease in the hematocrit value of 20% or more) or recurrent bleeding after 24 hours of stability (defined as the need for additional transfusions, a further decrease in the hematocrit value of at least 20%, or readmission for LGI bleeding within one week of discharge). Predictive factors include tachycardia, hypotension, syncope, a nontender abdomen, rectal bleeding on presentation, aspirin use, and more than two comorbid illnesses.223,224 Such a prognostic scoring system could identify patients at the highest risk for severe LGI bleeding, who account for 19% of patients with LGI bleeding and who might benefit most from urgent co­ lonoscopy, although this benefit remains unproved. A single-institution case series of 94 patients admitted for LGI bleeding has identified similar risk factors.225 This study found that 39% of all cases of LGI bleeding requiring hos­ pitalization were severe, as defined by the passage of red blood after the patient left the emergency department, with hypotension or tachycardia, or the need for a transfusion of more than two units of packed red blood cells during the hospitalization. Independent risk factors for severe LGI bleeding were an initial hematocrit value of 35% or lower, abnormal vital signs (a systolic blood pressure less than 100 mm Hg or a heart rate higher than 100 beats/min) on admission, and gross blood on initial rectal examination. Artificial neural networks also have been used to develop prediction models for severe LGI bleeding.226,227 The main problem with the use of such networks from a clinical point of view is that a large number of variables need to be entered into a computer program for analysis, thus limiting wide­ spread routine clinical use.

MORTALITY

A large United States database study of 227,000 patients with a discharge diagnosis of LGI bleeding in 2002 reported an overall mortality rate of 3.9%.219 Multivariate analysis has found that independent predictors of in-hospital mortal­ ity are age older than 70 years, intestinal ischemia, at least two comorbid illnesses, bleeding during a hospitalization

for an unrelated condition, coagulopathy, hypovolemia, transfusion of packed red blood cells, and male gender. Colorectal polyps and hemorrhoids were associated with a lower mortality risk. The low risk of death from LGI bleed­ ing identified in this study is consistent with data from smaller series such as those from Kaiser San Diego (2.4%) and the University of California, San Francisco (3.2%).216,225 The Kaiser study also found an increased risk of death with in-hospital LGI bleeding.

DIAGNOSTIC AND THERAPEUTIC APPROACH

Patients with hematochezia should undergo the same careful history taking, physical examination, and laboratory testing described earlier for the general approach to the patient with acute GI bleeding (see Table 19-1). The history should focus specifically on identifying sources of LGI bleeding. Diverticular bleeding should be suspected in patients with painless severe acute hematochezia and a history of diverticulosis. Ischemic colitis may be suspected in those with painful acute hematochezia with mild abdom­ inal discomfort. A recent polypectomy suggests a postpol­ ypectomy bleed. A history of recent antibiotic use or inflammatory bowel disease and bloody diarrhea suggests colitis. The acute onset of bright red blood that drips from the anus suggests a hemorrhoidal bleed. Prior radiation sug­ gests radiation proctitis. Patients should be medically resuscitated. Because LGI bleeding is generally less severe than UGI bleeding, blood transfusions may not be needed. Most patients should undergo initial evaluation with colonoscopy after bowel preparation, although in selected cases anoscopy or flexible sigmoidoscopy without any bowel cleansing or after an enema may be performed. Other diagnostic tests, including radionuclide bleeding scans or angiography, may be used in selected cases or when colonoscopy fails to detect a source of bleeding.

Anoscopy

Anoscopy can be useful for patients in whom actively bleed­ ing internal hemorrhoids or other anorectal disorders (e.g., fissures, fistulas, proctitis) are suspected and allows imme­ diate treatment with rubber band ligation (see Chapter 125). Most patients, however, especially if older than 50 years, will also require colonoscopy, at least electively, to evaluate the remainder of the colon.

Flexible Sigmoidoscopy

Flexible sigmoidoscopy can evaluate the rectum and left side of the colon for a bleeding site and can be performed without a standard colonoscopy bowel preparation.

309

310

Section III  Symptoms, Signs, and Biopsychosocial Issues Although not adequate for evaluation of the anal canal, flex­ ible sigmoidoscopy alone will result in a diagnosis in approximately 9% of cases.228 If the distal colon can be adequately cleansed with enemas, an urgent flexible sig­ moidoscopy can be useful for patients suspected of having a solitary rectal ulcer, ulcerative colitis, radiation proctitis, ischemic colitis, postpolypectomy bleeding (in the rectosig­ moid), or internal hemorrhoids (see Chapters 39, 112, 115, 122, 124, and 125). Therapeutic hemostasis can be provided with injection therapy, hemoclip placement, band ligation, and MPEC. Monopolar electrocautery (e.g., argon plasma coagulation, snare polypectomy, or hot biopsy forceps) should not be used if a bowel preparation has not been administered to avoid the risk of ignited flammable colonic gas (see Chapter 16).

patient is having hematochezia with abdominal pain. One study from France reported that CT accurately identified 17 of 19 LGI bleeding sites, including diverticula, tumors, angi­ omas, and varices.231 Multidetector CT has been shown to be more accurate than technetium-tagged red blood cell scanning in patients with LGI bleeding.232 CT colonography is being used increasingly to screen persons for colonic polyps and cancer and may be of some benefit in patients with LGI bleeding. CT colonography detects large polyps (>1 cm) or cancers with a sensitivity rate of 90%.233 Faster multidetector scanners also allow CT angiography to be performed, as well as evaluation of the small bowel. This capability could allow detection of masses and vascular lesions and is a potential advantage of CT angiography over other radiologic imaging techniques.

Nuclear Scintigraphy

Colonoscopy

Nuclear scintigraphy involves injecting a radiolabeled sub­ stance into the patient’s bloodstream and performing serial scintigraphy to detect focal collections of radiolabeled mate­ rial (see earlier). This technique has been reported to detect bleeding at a rate as low as 0.04 mL/min,38 with an overall positive diagnostic rate of approximately 45% and a 78% accuracy rate for localizing the true bleeding site.40 The disadvantages of radionuclide bleeding scans are that delayed scans may be misleading, and determining the specific cause of bleeding often depends on endoscopy or surgery. False-positive results are most likely to occur when transit of luminal blood is rapid, so that radiolabeled blood is detected in the colon, even though it originated in the UGI tract. Radionuclide scanning may be helpful in cases of obscure GI bleeding (see later) or prior to angiography to help localize a lesion, particularly if an early scan (e.g., 30 minutes to four hours after injection of the radiolabeled material) is positive for red blood cell extravasation.

Angiography

Angiography is most likely to detect a site of bleeding when the rate of arterial bleeding is at least 0.5 mL/min.35 The diagnostic yield depends on patient selection, the timing of the procedure, and the skill of the angiographer, with posi­ tive results in 12% to 69% of cases. An advantage of angi­ ography is that embolization can be performed to control some bleeding lesions. Major complications, however, occur in 3% of cases and include bowel ischemia, hema­ toma formation, femoral artery thrombosis, contrast dye reactions, acute kidney injury, and transient ischemic attacks.37 Other disadvantages of angiography are the absence of active bleeding in most patients at the time of angiography, expense of the test, and inability to determine the specific lesion responsible for bleeding in many cases. A small retrospective case series of 11 patients with colonic bleeding who underwent angiographic emboliza­ tion reported that the bleeding ceased in 10, mesenteric ischemia developed in 7, and 6 died.229 Another study of 65 patients with acute LGI bleeding who did not undergo colonoscopy as a first diagnostic step found that diagnostic angiography provided little additional clinical information because the bleeding stopped spontaneously in most patients. Moreover, angiography did not help guide sub­ sequent surgery and was associated with a complication rate of 11%.230

Computed Tomography and Computed Tomography Colonography

Multidetector CT can identify abnormalities in the colon that could be a source of bleeding, such as diverticulosis, colitis, masses, and varices. CT is often performed if the

Urgent colonoscopy following a rapid bowel purge has been shown to be safe, provide important diagnostic information, and allow therapeutic intervention.22,217 Patients usually ingest 4 to 8 L of polyethylene glycol solution orally or via a nasogastric tube over four to six hours until the rectal effluent is clear of stool, blood, and clots. Metoclopramide, 10 mg, may be given intravenously before the purge and repeated every three to four hours to facilitate gastric empty­ ing and reduce nausea. Sodium phosphate bowel prepara­ tions probably should be avoided in patients with suspected LGI bleeding because of potential risks of the high phos­ phate and sodium loads. Urgent colonoscopy for LGI bleeding generally is per­ formed 6 to 36 hours after the patient is admitted to the hospital. Because most bleeding stops spontaneously, colo­ noscopy often is performed semielectively on the day after initial hospitalization to allow the patient to receive blood transfusions and the bowel preparation on the first day of hospitalization. The overall rate of detecting a presumed or definite cause of LGI bleeding by colonoscopy ranges from 48% to 90%, with an average of 68%, based on a review of 13 studies.40 The problem with interpreting these data, however, is that making a definite diagnosis of the cause of the bleeding is often not possible unless a bleeding stigma such as active bleeding, a visible vessel, an adherent clot, mucosal friabil­ ity or ulceration, or the presence of fresh blood limited to a specific segment of the colon is seen. The optimal time for performing urgent bowel prepara­ tion and colonoscopy is unknown. Theoretically, the sooner endoscopy is performed, the higher the likelihood of finding a lesion, such as a bleeding diverticulum or polyp stalk, that might be amenable to endoscopic hemostasis. A retrospec­ tive study from the Mayo Clinic, however, has suggested that in patients with diverticular bleeding, the timing of endoscopy (0 to 12 hours, 12 to 24 hours, or more than 24 hours after admission) is not associated significantly with the finding of active bleeding or other stigmata that would prompt colonoscopic hemostasis.234 Early colonos­ copy (soon after admission) has been associated with a shorter length of hospitalization, principally because of improved diagnostic yield rather than therapeutic interven­ tions.235 A consensus on a single approach to patients with severe hematochezia has not been reached, and the approach used depends on local resources and expertise. In large centers, the approach detailed in Figure 19-4 is recom­ mended. With use of an urgent endoscopic approach for diagnosis and treatment, the diagnostic yield of definitive and presumptive bleeding sites is more than 90%, and the estimated direct costs are significantly less than the costs associated with an elective evaluation.23

Chapter 19  Gastrointestinal Bleeding Barium Enema

Emergency barium enema has no role in patients with LGI bleeding. This test is rarely diagnostic because it cannot demonstrate vascular lesions and may be misleading if only diverticula are seen. It fails to detect 50% of polyps larger than 10 mm. In addition, the barium contrast liquid can make urgent colonoscopy more difficult by impairing visu­ alization.236 Subsequent colonoscopy is needed for any sus­ picious lesions seen on barium enema or for lesions that require therapy.

Role of Surgery

Surgical management is rarely needed in patients with LGI bleeding because most bleeding is self-limited or easily managed with medical or endoscopic therapy. The main indications for surgery are malignancy, diffuse bleeding that fails to cease with medical therapy (as in ischemic or ulcer­ ative colitis), and recurrent bleeding from a diverticulum. Therefore, most stable patients can be managed on a medical service rather than a surgical service.

CAUSES AND MANAGEMENT

Visualizing active bleeding during colonoscopy is not always possible. Unlike the case with angiography or nuclear red blood cell scanning, however, colonoscopy permits identification of stigmata of recent hemorrhage (visible vessels, clots, or spots) and provides information on the location of the lesion and on risk stratification. The earlier the colonoscopy is carried out, the higher the chance of detecting an actively bleeding lesion or stigmata of recent hemorrhage. A definite diagnosis of a bleeding lesion can usually be made if active bleeding, a visible vessel, or a clot is seen. A presumptive diagnosis of the cause of bleeding can be made if a lesion that is a potential cause of bleeding is seen and no other possible sources are identified by anos­ copy, full colonoscopy with intubation of the terminal ileum, and, in some cases, push enteroscopy.23,217

Diverticulosis

Colonic diverticula are herniations of colonic mucosa and submucosa through the muscular layers of the colon (see Chapter 117). Histopathologically, diverticula in the colon are actually pseudodiverticula, because they do not contain all layers of the colonic wall. Diverticula form when colonic tissue is pushed out by intraluminal pressure at points of entry of the small arteries (vasa recta), where they penetrate the circular muscle layer of the colonic wall. The entry points of the vasa recta are areas of relative weakness through which the mucosa and submucosa can herniate when intraluminal pressure is increased. Diverticula vary in diameter from a few milliliters to several centimeters and are located most commonly in the left colon. Most colonic diverticula are asymptomatic and remain uncomplicated. Bleeding may occur from vessels at the neck or base of a diverticulum.237 In our experience with definitive diverticu­ lar hemorrhage (see later), bleeding was from the base in 52% and from the neck in 48% of diverticula.217 Diverticula are common in Western countries, with a fre­ quency of 50% in older adults.238 By contrast, diverticula are found in fewer than 1% of continental African and Asian populations.239 It has been hypothesized that the regional differences in prevalence rates can be explained by the low amount of dietary fiber in Western diets. Presum­ ably, a low-fiber diet results in less stool content, longer fecal transit times, increased colonic muscle contraction and, ultimately, increased intraluminal pressure, which results in the formation of propulsion diverticula. Further­ more, diverticula increase in frequency with advanced age,

possibly because of weakening of the colonic wall and muscle tone. Diverticular bleeding develops in an estimated 3% to 5% of patients with diverticulosis.240 Although most diverticula are in the left colon, several series have sug­ gested that diverticula in the right colon are more likely to bleed.237,240-242 Two thirds of definitive diverticular bleeds (with stigmata of hemorrhage) emanate from the region of the splenic flexure of the colon or proximally.217 Diverticular hemorrhage should be classified carefully based on findings at colonoscopy, angiography, or surgery,23 particularly in the case of older patients with severe hema­ tochezia who are likely to have colonic diverticulosis. Definitive diverticular hemorrhage is diagnosed when stig­ mata of recent hemorrhage (e.g., active bleeding, visible vessel, adherent clot) are seen on colonoscopy or active bleeding is demonstrated on angiography or nuclear red blood cell scanning, with later confirmation of a diverticu­ lum in that location as the source of bleeding by colonos­ copy or surgery. Presumptive diverticular hemorrhage is diagnosed when colonoscopy reveals diverticulosis without stigmata and no other significant lesions are seen in the colon and by anoscopy, terminal ileum examination, and push enteroscopy. The term incidental diverticulosis is used when another lesion is identified as the cause of hema­ tochezia and colonic diverticulosis is evident. In a large, prospective cohort study in which the management algo­ rithm shown in Figure 19-4 was used in our institutions to classify patients with hematochezia, colonic diverticulosis was incidental in 52%, presumptive diverticular hemor­ rhage occurred in 31%, and definitive diverticular hemor­ rhage was established in 17% of cases.217 Patients with diverticular bleeding typically are older, have been taking aspirin or other NSAIDs, and present with painless hematochezia.243 In at least 75% of patients with diverticular bleeding, the bleeding stops spontaneously, and these patients require transfusion of fewer than four units of packed red blood cells. In one surgical series, surgi­ cal segmental colonic resection was needed in 60% of patients, most of whom had had continued bleeding despite trans­fusion of four units of blood.241 Patients who under­ went resection for a bleeding diverticulum had a rebleeding rate of 4%. Among patients who stopped bleeding spontane­ ously, the rebleeding rate from colonic diverticulosis has been reported to range from 25% to 38% over the next four years, with most patients having mild rebleeding.216,241 These data, however, are not based on colonoscopic docu­ mentation of diverticular bleeding, and the actual rate of rebleeding appears to be lower. In a large, prospective, cohort study of patients with documented colonic diverticu­ lar hemorrhage (definitive or presumptive) by our group, the overall rate of rebleeding was 18% in four years—9% from recurrent diverticular hemorrhage and 9% from other GI sources.217 Occasionally, urgent colonoscopy reveals a stigma of recent bleeding, such as active bleeding, a visible vessel, a clot, or blood limited to one segment of the colon. As noted, earlier colonoscopy for LGI bleeding is likely to result in a greater frequency of finding stigmata of recent diverticular bleeding, although a small case series study from the Mayo Clinic did not find any difference in the rate of detection of these stigmata whether colonoscopy was performed between 0 and 12 hours, 12 and 24 hours, or more than 24 hours from the time of hospital admission.234 Stratifying the risk of diverticular rebleeding by applying the same endoscopic stigmata used in high-risk peptic ulcer bleeding (active bleeding, visible vessel, and clot) has been attempted, but the natural history associated with each of these stigmata is unknown (Fig. 19-18). The pigmented pro­

311

312

Section III  Symptoms, Signs, and Biopsychosocial Issues

A

B

C

Figure 19-18.  Endoscopic stigmata of recent colonic diverticular bleeding. A, Active bleeding (arrow). B, Adherent clot (arrow). C, Nonbleeding visible vessel (arrow).

tuberance found on the edge of some diverticula is usually clot at the edge of a ruptured blood vessel on histopathol­ ogy.244 The UCLA CURE group found that among 17 patients with stigmata of recent diverticular hemorrhage (active bleeding in 6, visible vessel in 4, and adherent clot in 7), the rates of rebleeding (53%) and emergency surgery (35%) were high.23 Endoscopic Hemostasis Colonoscopic hemostasis of actively bleeding diverticula has been reported using MPEC, epinephrine injection, hemoclips, fibrin glue, or combinations of epinephrine and MPEC or hemoclips.23,244-249 If fresh red blood is seen in a focal segment of colon, that segment should be irrigated vigorously with water to remove the blood and detect the underlying bleeding site. If the bleeding is coming from the edge of a diverticulum or a pigmented protuberance is seen on the edge, a sclerotherapy needle can be used to inject epinephrine (diluted 1 : 20,000 in saline) in 1-mL aliquots submucosally into four quadrants around the bleeding site. Subsequently, MPEC at a low power setting (10 to 15 W) and light pressure can be carried out for a one-second pulse duration to cauterize the diverticular edge and stop bleeding or flatten the visible vessel, or hemoclips can be applied. A nonbleeding adherent clot can be injected with 1 : 20,000 epinephrine into four quadrants, 1 mL/quadrant, after which the clot can be removed piecemeal by guillotining it with a cold polyp snare until it extends 3 mm above the diverticulum. The underlying stigma is treated with MPEC or hemoclips (see earlier). After endoscopic hemostasis of a bleeding diverticulum is completed, a permanent submucosal tattoo should be placed around the lesion to allow identification of the site in case colonoscopy or surgery is repeated for recurrent bleeding. After colonoscopic hemostasis, patients are told to avoid aspirin and other NSAIDs and to take a daily fiber supplement on a long-term basis. In 2000, Jensen and the UCLA CURE group published their results on urgent colonoscopy for the diagnosis and treatment of severe diverticular hemorrhage23 and reported that 20% of patients with severe hematochezia had endo­ scopic stigmata, suggesting a definitive diverticular bleed. This group of patients, who underwent colonoscopic hemostasis, had a rebleeding rate of 0% and an emergency hemicolectomy rate of 0% compared with 53% and 35%, respectively, in a historical control group of patients who had high-risk stigmata but did not undergo colonoscopic hemostasis. No rebleeding had occurred after three years of follow-up in the patients who underwent colonoscopic

hemostasis. Other studies also have suggested that endo­ scopic hemostasis is effective for immediate diverticular bleeding, although subsequent rebleeding and the need for surgery have been reported in some cases.250 Angiography and Surgery Angiographic embolization can be performed in selected cases of diverticular bleeding, but with a risk of bowel infarction, contrast reactions, and acute kidney injury. One study found that routine angiography prior to surgical resection is not helpful in reducing the overall risk of complications.230 Surgical resection for diverticular bleeding is rarely needed and is reserved for recurrent bleeding. The decision to operate is best guided by colonoscopic, angiographic, or nuclear medicine studies that demonstrate the likely segment of colon from which the bleeding is emanating and by the presence of medical comorbidities. Diverticular bleeding is usually mild, and the risks of surgical com­ plications are increased in older patients. Blind subtotal colectomy, often performed in the past when a definite bleeding site could not be identified, should be avoided, if possible.

Colitis

The term colitis refers to any form of inflammation of the colon. Severe LGI bleeding may be caused by ischemic colitis, inflammatory bowel disease, or possibly infectious colitis. Ischemic colitis can present as painless or painful hema­ tochezia with mild left-sided abdominal discomfort. The painless subtype usually results from mucosal hypoxia, and is thought to be caused by hypoperfusion of the intramural vessels of the intestinal wall, rather than by large-vessel occlusion, which is often painful and clinically more severe and has worse outcomes. The incidence of ischemic colitis is estimated to be 4.5 to 44 cases/100,000 person-years.251 Most cases do not have a recognizable cause. Risk factors associated with ischemic colitis have been reported to include older age, shock, cardiovascular surgery, congestive heart failure, chronic obstructive pulmonary disease, ileos­ tomy, colon cancer, abdominal surgery, irritable bowel syn­ drome, constipation, laxative use, oral contraceptive use, and use of an H2 receptor antagonist.251-254 The superior mesenteric artery supplies blood to the right colon (cecum, ascending colon, hepatic flexure, proximal transverse colon, and midtransverse colon) whereas the inferior mesenteric artery supplies blood to the left colon (distal transverse colon, splenic flexure, descending colon, sigmoid colon,

Chapter 19  Gastrointestinal Bleeding and rectum). The colon has an abundant blood supply, but the watershed area between the superior and inferior mes­ enteric arteries has the fewest collateral vessels and is at most risk for ischemia. The colon normally receives 10% to 35% of cardiac output, and ischemia can occur if blood flow decreases by more than 50%. Although ischemia is most likely to occur in the watershed area of the splenic flexure, it can occur anywhere in the colon.255 The diagnosis of ischemia is usually made by colonos­ copy and is suspected by the demonstration of thumbprint­ ing on plain film radiographs or colonic wall thickening on CT scans. The colonoscopic appearance of the mucosa includes erythema, friability, and exudate. Mucosal biopsy specimens may suggest ischemic changes but generally are used to exclude infectious colitis or Crohn’s disease. Ische­ mic colitis generally resolves in a few days and generally does not require colonoscopic hemostasis or antibiotic therapy. In the UCLA CURE experience, approximately 10% of patients with ischemic colitis and severe hemato­ chezia had a focal ulcer with a major stigma of hemorrhage on urgent colonoscopy. The recommended treatment in these cases is epinephrine injection and hemoclipping (see Chapter 114). In a large retrospective series from Kaiser, no episodes of rebleeding from ischemic colitis occurred over a four-year follow-up period.216 On the other hand, inpa­ tients in whom ischemic colitis develops or those with large-vessel mesenteric ischemia usually have worse out­ comes, including higher rates of rebleeding, perforation, need for surgery, and death. Inflammatory bowel disease that involves the colon can rarely cause severe acute LGI bleeding. In a case series from the Mayo Clinic, most of these patients had Crohn’s disease, and most were successfully treated medically.256 Three of the 31 patients in this series underwent endoscopic therapy with epinephrine injection alone or with MPEC for an adherent clot or an oozing ulcer. These 3 patients had no rebleeding, but 23% of the other patients had rebleeding at a median of three days (range, 1 to 75 days) after the initial bleed. Thirty-nine percent of the patients with severe bleeding eventually required surgery (see Chapters 111 and 112). Infectious colitis should be excluded in any patient with severe LGI bleeding and colitis. LGI bleeding can occur with infection caused by Campylobacter jejuni, Salmonella, Shigella, enterohemorrhagic Escherichia coli (O157:H7), cytomegalovirus, or Clostridium difficile. Significant blood loss is rare except in patients with severe coagulopathy. Diagnosis is made by stool cultures and flexible sigmoidos­ copy or colonoscopy. Treatment is with medical manage­ ment; the use of antibiotics depends on the causative organism. Endoscopic management generally has no role in infectious colitis (see Chapter 107).

Postpolypectomy Bleeding

Bleeding occurs after approximately 1% of colonoscopic polypectomies. The bleeding occurs most commonly five to seven days after polypectomy but can occur from 1 to 14 days after the procedure; it generally is self-limited and mild to moderate, with 50% to 75% of patients requiring blood transfusions.257-260 Reported risk factors for postpoly­ pectomy bleeding include a large polyp size (more than 2 cm), thick stalk, sessile type, location in the right colon, use of warfarin or heparin, and use of aspirin or other NSAID. In most cases of delayed postpolypectomy bleeding, an ulceration at the site of the polypectomy is found on colonoscopy at the time of bleeding. In patients with severe bleeding, stigmata of recent hemorrhage may be found in the ulceration.261 Figure 19-19 illustrates a postpolypectomy

Figure 19-19.  Endoscopic appearance of post-polypectomy bleeding in the colon. Bleeding occurred seven days after a snare polypectomy for a large pedunculated polyp. Note the nonbleeding visible vessel (arrow) in the ulcerated polypectomy site.

ulcer with evidence of recent hemorrhage. Endoscopic man­ agement techniques for delayed postpolypectomy ulcer bleeding depend on the stigma found and are similar to those used for peptic ulcer hemorrhage, including epineph­ rine injection, thermal coagulation, hemoclip placement, and combination therapy.

Colon Polyps and Cancer

Patients with colon polyps and cancer can present with acute hematochezia. Often, these patients have a microcytic iron deficiency anemia consistent with slow GI blood loss (see later) before more overt bleeding occurs. At colonos­ copy, epinephrine can be injected into the lesion to slow active bleeding, and hemoclips can be applied to treat stig­ mata of hemorrhage on ulcerated lesions that cannot be resected endoscopically. When possible, colon polyps can be removed to stop bleeding. Surgical resection is usually required to prevent rebleeding from a large ulcerated sessile lesion (see Chapters 122 and 123).

Radiation Proctitis

Radiation proctitis usually causes mild chronic hematoche­ zia but occasionally can cause acute severe LGI bleeding. Ionizing radiation can cause acute and chronic damage to the normal colon and rectum when used to treat gyneco­ logic, prostatic, bladder, or rectal tumors. Acute self-limited diarrhea, tenesmus, abdominal cramping, and rarely bleed­ ing will develop for a few weeks in approximately 75% of patients who have received a radiation dose of 4000 cGy. Chronic radiation effects occur 6 to 18 months after comple­ tion of treatment and manifest as bright red blood with bowel movements. Bowel injury resulting from chronic radiation is related to vascular damage, with subsequent mucosal ischemia, thickening, and ulceration. Much of this damage is thought to result from chronic hypoxic ischemia and oxidative stress. Flexible sigmoidoscopy or colonoscopy reveals telangiec­ tasias, friability, and sometimes ulceration in the rectum (Fig. 19-20). At times, active bleeding is noted, and usually numerous telangiectasias are seen. Internal hemorrhoids are often seen as well and frequently are considered in the dif­ ferential diagnosis of the rectal bleeding. Treatment initially focuses on avoidance of aspirin and other NSAIDs, consumption of a high-fiber diet, and iron

313

314

Section III  Symptoms, Signs, and Biopsychosocial Issues medical therapy, those with severe bleeding are likely to require some form of endoscopic or surgical treatment (see Chapter 125).

Anal Fissures

Patients with an anal fissure usually present with extremely painful bowel movements but can present with hematoche­ zia. Generally, the hematochezia is mild and is noticed with wiping; rarely, hematochezia is moderate to severe. Treat­ ment focuses on resolving the anal fissure, rather than using specific hemostasis techniques. A topical calcium channel blocker (such as 2% topical diltiazem cream) along with fiber supplementation, stool softeners, and sitz baths will heal most anal fissures (see Chapter 125).

Rectal Varices

Figure 19-20.  Endoscopic appearance of radiation proctitis. Note the diffuse oozing and telangiectasias.

supplementation if the patient is anemic. Medical therapy with topical or oral 5-aminosalicylic acid (mesalamine), sucralfate, or glucocorticoids can be tried but is not gener­ ally effective.262 Thermal therapy can be successful, includ­ ing repeated treatments with MPEC or argon plasma coagulation.263 Topical formalin applied directly to the rectal mucosa can reduce bleeding,264 as can the use of hyperbaric oxygen.265 Antioxidant vitamins, such as vita­ mins E and C, also have been reported to decrease bleeding from chronic radiation proctitis (see Chapter 39).266

Colonic Angioectasias

Colonic bleeding from angioectasias, an important cause of LGI bleeding, is discussed in the section on small bowel and obscure bleeding. When angioectasias are the cause of bleeding in the colon, they are often multiple, making endo­ scopic hemostasis a challenge (see Chapter 36).

Internal Hemorrhoids

Hemorrhoidal bleeding is characterized by bright red blood per rectum that can coat the outside of the stool, may drip into the toilet bowel, can be seen on tissue after wiping, and often appears as a large amount of fresh blood in the toilet. Usually, bleeding is mild, intermittent, and self-limited but occasionally, severe, transfusion-requiring bleeding may occur from hemorrhoids.267 In a large study of patients with hematochezia discharged from the hospital, 20% were thought to have had bleeding from hemorrhoids.219 In the UCLA CURE series of patients hospitalized for severe hema­ tochezia (see earlier), internal hemorrhoids were the second most common cause (see Table 19-9).217 Hemorrhoids were documented by urgent anoscopy and colonoscopy after a colonic preparation. The diagnosis can be made with anoscopy, sigmoidoscopy, or colonoscopy, especially if per­ formed while bleeding is ongoing. The treatment of hemorrhoids usually starts with medical therapy consisting of fiber supplementation, stool softeners, lubricant rectal suppositories (with or without glucocor­ ticoids), and warm sitz baths. Anoscopic therapy can also be used and includes injection sclerotherapy, rubber band ligation, cryosurgery, infrared photocoagulation, MPEC, and direct current electrocoagulation. Although most patients with mild hemorrhoidal bleeding respond to

Ectopic varices may develop in the rectal mucosa between the superior hemorrhoidal veins (portal circulation) and middle and inferior hemorrhoidal veins (systemic circula­ tion) in patients with portal hypertension. On sigmoidos­ copy, rectal varices are seen during retroflexion as vascular structures located several centimeters above the dentate line and extending into the rectum; they are distinct from inter­ nal hemorrhoids. The frequency of rectal varices increases with the degree of portal hypertension. Approximately 60% of patients with a history of bleeding esophageal varices have rectal varices. The treatment of rectal varices is similar to that for esophageal varices, with sclerotherapy, band liga­ tion, or a portosystemic shunt (see Chapter 90).268-270

Rectal Dieulafoy’s Lesions

Dieulafoy’s lesions are large submucosal arteries without overlying mucosal ulceration that can cause massive bleed­ ing. They can occur anywhere in the GI tract, although usually in the stomach (see earlier). Bleeding Dieulafoy’s lesions in the rectum, which have been treated successfully with endoscopic hemostasis, have been described in several reports.271,272

Rectal Ulcers

Several case series have described seriously ill hospitalized patients with the sudden onset of painless severe hemato­ chezia from a solitary or multiple rectal ulcers located 3 to 10 cm above the dentate line. In one series of 19 cases from Taiwan, 2.7% of patients evaluated for severe hematochezia were diagnosed with acute hemorrhagic rectal ulcer syn­ drome.273 The patients had a mean age of 71 years and had been hospitalized for other medical problems from 3 to 14 days (average 7.5 days) prior to the onset of bleeding. All developed hypotension and required transfer to an ICU and blood transfusions. Colonoscopy revealed an equal number of cases of multiple and solitary ulcers located 1 to 7 cm from the dentate line; most of the ulcers were large (more than 1 cm) and circumferential or geographic in appear­ ance. The patients were treated with combinations of thermal coagulation, injection therapy, and suture ligation and had a mortality rate of 26% because of multiorgan failure. The pathology of the lesions revealed necrosis sug­ gestive of mucosal ischemia, as seen with gastric stress ulcers. This entity appears to be a different disease from solitary rectal ulcer syndrome, colitis cystica profunda, infectious ulcers, radiation ulcer, NSAID ulcers, or constipation-induced stercoral ulcer and can be considered a type of stress ulcer of the rectum, similar to that seen in the duodenum, in extremely ill hospitalized patients. Solitary or multiple painless rectal ulcers were the third most common cause of severe hematochezia in the UCLA CURE study (see Table 19-9). In contrast to solitary rectal

Chapter 19  Gastrointestinal Bleeding Table 19-11  Causes of Obscure Gastrointestinal Bleeding

Figure 19-21.  Endoscopic appearance of bleeding from a solitary rectal ulcer with a visible vessel (arrow) seen on a retroflexed view.

ulcer syndrome (see Chapter 124), they occur in older patients with severe constipation, ICU patients, and persons who are bedridden and, on colonoscopy, are chronicappearing, large, and single or multiple. They often have stigmata of recent hemorrhage and can be treated endo­ scopically (Fig. 19-21).274 Patients with inpatient hemato­ chezia from a rectal ulcer have a higher rate of rebleeding than those who present from home (see Chapters 115 and 124).

OBSCURE OVERT GASTROINTESTINAL BLEEDING Obscure GI bleeding is commonly defined as GI bleeding of uncertain cause after a nondiagnostic upper endoscopy, colonoscopy, and barium small bowel follow-through.275 Obscure GI bleeding may have an overt or occult presenta­ tion. Overt obscure GI bleeding refers to patients who have visible acute GI bleeding (e.g., melena, maroon stool, hema­ tochezia) and a nondiagnostic upper endoscopy, colonos­ copy, and small bowel series. Occult obscure GI bleeding refers to patients with a positive fecal occult blood test result, usually in association with unexplained iron defi­ ciency anemia. In most large series, a diagnostic lesion is not found on upper endoscopy and colonoscopy in 5% of hospitalized patients with overt GI bleeding, and in 75% of these patients a bleeding site is found in the small intestine. In patients with obscure GI bleeding, the following pos­ sibilities exist: (1) the lesion was within reach of a standard endoscope and colonoscope but not recognized as the bleed­ ing site (e.g., Cameron’s lesions, angioectasias, or internal hemorrhoids); (2) the lesion was within reach of the endo­ scope and colonoscope but was difficult to visualize (e.g., a blood clot obscured visualization of the lesion, varices became inapparent in a hypovolemic patient, or a lesion was hidden behind a mucosal fold) or present intermittently (e.g., Dieulafoy’s lesion, angioectasias); or (3) the lesion was in the small intestine beyond the reach of standard endo­ scopes (e.g., neoplasm, angioectasias, diverticulum). In several series, approximately 50% of patients referred to a tertiary medical center for evaluation of obscure bleeding were found to have a lesion within reach of standard endo­ scopes (i.e., a missed lesion or difficult to see lesion that

Upper Gastrointestinal Tract* Cameron’s lesions Dieulafoy’s lesions Gastric antral vascular ectasia Small Intestine Angioectasias Aortoenteric fistula Dieulafoy’s lesion Diverticulosis Meckel’s diverticulum Neoplasm Pancreatic or biliary disease Ulceration Colon Angioectasias Diverticulosis Hemorrhoids Varices *After exclusion of common causes of upper gastrointestinal bleeding.

accounted for the bleeding). Causes of obscure GI bleeding are shown in Table 19-11.276 In a patient with recurrent severe unexplained hemato­ chezia, without hypotension, a colonic source should be suspected and a repeat colonoscopy is warranted (by the same or a different endoscopist). Colonic lesions that can bleed profusely and then stop, such as diverticulosis or hemorrhoids, should be considered. In patients with recur­ rent severe melena, upper endoscopy to re-examine the esophagus, stomach, and duodenum, as well as the proxi­ mal jejunum, for a missed or unrecognized lesion should be considered. If an upper GI endoscopic examination is repeated, push enteroscopy is recommended (see later). Once it is certain that a bleeding lesion in the UGI or LGI tract was not missed, the evaluation should focus on the small intestine. In the past, the principal imaging modality of the small intestine was barium radiography, but this technique has been limited by the length, mobility, and motility of the small bowel and by overlying loops of bowel. Because small bowel bleeding is often intermittent, nuclear medicine bleeding scanning or angiography has limited value in the diagnostic evaluation. Since the late 1990s, diagnostic options for evaluating the small intestine have expanded greatly and have been revolutionized by the development of new small bowel imaging techniques, including wireless video capsule endoscopy, deep enteros­ copy, and CT enterography, which now allow greater visu­ alization and more therapeutic options than in the past.

CAUSES AND DIAGNOSTIC AND THERAPEUTIC APPROACHES

A number of lesions can cause obscure bleeding (see Table 19-11). The cause of bleeding is more likely to be a tumor, Meckel’s diverticulum, or Crohn’s disease in patients younger than 40 years and angioectasias or an NSAIDinduced ulcer in those ages 40 years and older.

Angioectasia

A variety of vascular lesions may cause bleeding from the GI tract (see Chapter 36). Angioectasias, sometimes referred to as angiodysplasias, are aberrant blood vessels found throughout the GI tract that develop with advancing age. They are distinct from arteriovenous malformations, which are congenital, and angiomas, which are neoplastic. Telan­ giectasia is the lesion that results from dilatation of the

315

316

Section III  Symptoms, Signs, and Biopsychosocial Issues terminal aspect of a vessel. Any of the vascular lesions may cause overt or obscure GI bleeding in adults, particularly in older adults and those who take antiplatelet and anticoagu­ lant drugs. Acquired vascular lesions (angioectasias and telangiectasias) occur in association with various disorders, such as chronic kidney disease, cirrhosis, rheumatoid dis­ orders, and severe heart disease.48 Although angioectasias may present as overt bleeding, they often manifest as occult bleeding or iron deficiency anemia. The most common loca­ tions are the colon and small intestine. The histopathology of angioectasias in the colon is char­ acterized by ectatic, dilated submucosal veins.277,278 A pro­ posed mechanism for their formation in the colon is that partial, intermittent, low-grade obstruction of submucosal veins during muscular contraction and distention of the cecum results in dilatation and tortuosity of the submucosal veins. Over time, the increased pressure also results in dila­ tation of the venules, capillaries, and arteries of the mucosal vasculature. Finally, precapillary sphincters can become incompetent, thereby causing arteriovenous communica­ tions to develop and possibly resulting in local mucosal ischemia. Because angioectasias can occur elsewhere in the GI tract, other mechanisms are postulated, including a response to mucosal irritation or local ischemia, as occurs after radiation. Most angioectasias occur in patients older than 60 years and can involve any segment of the GI tract. Usually, the lesions are multiple in a given segment of intestine. Approx­ imately 20% (and probably more) of patients have angioec­ tasias in at least two sections of the GI tract.279,280 In studies of asymptomatic persons who underwent co­ lonoscopy, angioectasias were found in 1% to 3%.281,282 In these individuals, the angioectasias were mostly in the right colon, with the following distribution: cecum, 37%; ascend­ ing colon, 17%; transverse colon, 7%; descending colon, 7%; sigmoid colon, 18%; and rectum, 14%. Among asymp­ tomatic persons found incidentally to have colonic angioec­ tasias, no bleeding occurred during a three-year follow up. Several conditions appear to be associated with an increased frequency of angioectasias. Patients with chronic kidney disease and uremia have an increased rate of intes­ tinal angioectasias. A study of patients with and without chronic kidney disease who had obscure GI bleeding found angioectasias as the presumptive source in 47% compared with 18% of those without kidney disease.283 The increased risk of bleeding from angioectasias in patients with chronic kidney disease may be associated with uremia-induced platelet dysfunction. Von Willebrand’s disease (congenital or acquired) also has been associated with bleeding angioectasias. Von Wil­ lebrand’s factor is needed for effective platelet aggregation. A small case series of 10 patients with bleeding GI angioectasias and von Willebrand’s disease suggested an association between the two disorders.284 Subsequently, a well-controlled prospective study found that almost all patients with bleeding GI (upper GI and colonic) angioecta­ sias, as opposed to nonbleeding angioectasias or bleeding diver­ticulosis, had acquired von Willebrand’s disease asso­ ciated with selective loss of the largest multimeric forms of von Willebrand’s factor, as well as with aortic stenosis.285 Because the large von Willebrand multimers promote primary hemostasis in a microcirculation characterized by high shear forces, as occurs in angioectasias, the loss of the large multimers may explain why bleeding occurs in some patients with angioectasias. Aortic stenosis has been associated with GI bleeding from angioectasias (Heyde’s syndrome).286 This association is controversial because both conditions are common and an

association may not imply cause and effect.287 Nevertheless, aortic stenosis has been shown to be associated with an acquired form of von Willebrand’s disease in 67% to 92% of patients because of mechanical disruption of von Willebrand proteins during passage through the stenotic aortic valve; the acquired von Willebrand’s disease, in turn, increases the risk of bleeding from angioectasias.288,289 Several series have reported the cessation of bleeding from angioectasias after aortic valve replacement, even though the angioectasias persisted, an observation consistent with the hypothesis that bleeding was the result of the damaged von Willebrand factors that normalized after aortic valve replacement.290 Because many older persons with bleeding from intestinal angioectasias have cardiovascular disease, but not severe aortic stenosis, other cardiovascular disorders such as mild to moderate aortic stenosis, aortic sclerosis, hypertrophic cardiomyopathy, and peripheral vascular disease may result in sufficiently high shear rates to disrupt von Willebrand factors and contribute to bleeding angioectasias.290 On endoscopy, angioectasias appear as 2- to 10-mm red lesions, with arborizing ectatic blood vessels that emanate from a central vessel (Fig. 19-22). Application of pressure on an angioectasia with an endoscopic probe may cause the lesion to blanch. One study has suggested that sedation of a patient with a narcotic during endoscopy can make visualization of angioectasias difficult because of transient hypotension that leads to decreased filling or causes vasoconstriction and that reversal with naloxone, a narcotic antagonist, can make the angioectasia more prominent.291 In practice, however, this maneuver is unlikely to be useful clinically and might make the patient more uncomfortable. Endoscopic treatment of angioectasias can be performed with various modalities, including injection therapy with epinephrine, thermal probe coagulation, argon plasma coagulation, and band ligation. Assessing efficacy can be difficult, given the heterogeneity of patients affected and intermittent nature of the blood loss. One series of 16 patients with transfusion-requiring angioectasias found no difference in the frequency of continued bleeding (50%), whether treatment was with surgery, endoscopic therapy, or blood transfusions alone, presumably because of the diffuse locations of the angioectasias.292 In another study of 33 patients with iron deficiency anemia and small bowel angioectasias seen on push enteroscopy, no changes in clinical or endoscopic findings were found in most patients one year after endoscopic therapy.293 By contrast, in another study of patients with GI bleeding suspected from small bowel angioectasias, treatment with electroco­ agulation led to a significant decrease (but not elimination of the need for) blood transfusions compared with observa­ tion alone.294 In a pilot study of double-balloon enteroscopy, endoscopic treatment was performed in approximately one half of patients with angioectasias, and rebleeding rates during follow-up were similar in the treated and nontreated patients.295 In a small case series, hormonal therapy with estrogen was suggested to have a benefit in controlling bleeding from telangiectasias in patients with chronic kidney disease.296 Case reports have also suggested that estrogen decreases bleeding in patients with hereditary hemorrhagic telan­ giectasia (Osler-Weber-Rendu disease; see later) and von Willebrand’s disease. A multicenter, randomized, controlled trial involving 72 patients, however, found no difference between an estrogen-progesterone combination and placebo in the rates of rebleeding, which were 39% and 46%, respectively.297 Therefore, routine use of hor­

Chapter 19  Gastrointestinal Bleeding

A

B

Figure 19-22.  Endoscopic appearance of jejunal angioectasia before (left) and after (right) multipolar probe electrocoagulation.

mones for managing bleeding from angioectasias cannot be recommended. Most patients with intermittently bleeding GI angioec­ tasias require medical treatment in addition to endoscopic hemostasis. Use of medications that can exacerbate chronic low-level bleeding, in particular aspirin, other NSAIDs, warfarin, and clopidogrel, should be avoided or at least minimized. Many patients can be managed with chronic administration of iron (orally or intravenously) and occasionally may need erythropoietin injections as well to maintain adequate blood counts, despite ongoing bleeding.

Hereditary Hemorrhagic Telangiectasia

HHT, also known as Osler-Weber-Rendu disease, is a hered­ itary condition characterized by diffuse telangiectasias and large arteriovenous malformations. The most striking clinical feature is telangiectasias on the lips, oral mucosa, and fingertips. Additionally, up to one third of patients have pulmonary, hepatic, or cerebral arteriovenous malfor­ mations (AVMs). Patients generally present with recurrent severe nosebleeds, GI bleeding, and iron deficiency anemia. Usually. the epistaxis, rather than GI bleeding, causes the more profound blood loss and anemia. HHT can be lifethreatening because of embolic strokes or brain abscesses related to the pulmonary and cerebral AVMs. Symptoms of HHT generally develop in childhood or early adulthood. HHT is inherited as an autosomal dominant trait, with varying phenotypic expression. Mutations occur in at least four genes (ENG [encodes endoglin], ALK-1 [encodes activin receptor-like kinase 1], MADH4, and HHT3) that encode proteins needed to maintain the integrity of the vascular endothelium; defects in these proteins allow the formation of AVMs. The diagnosis of HHT is based on four criteria: (1) spon­ taneous and recurrent epistaxis; (2) multiple mucocutane­ ous telangiectasias; (3) visceral AVMs (GI, pulmonary, brain, liver); and (4) a first-degree relative with HHT.298 Genetic testing to detect mutations in the ENG, ALK-1, or MAHD4 genes may be helpful in selected cases. Patients suspected of having HHT should be screened for cerebral and pulmonary AVMs, and family members of the patient should consider genetic testing.

Telangiectasias can occur anywhere in the small intestine in patients with HHT. In a case series in which capsule endoscopy was performed in 32 patients with and 48 patients without HHT who were being evaluated for small bowel bleeding, small bowel telangiectasias were found in 81% of patients with HHT compared with 29% of those without HHT.299 The telangiectasias were evenly distributed throughout the small bowel, but all actively bleeding lesions were found in the duodenum or proximal jejunum and within reach of a standard push enteroscope. The detec­ tion of five or more telangiectasias had a sensitivity of 75% and a positive predictive value of 86% for a diagnosis of HHT. The treatment of HHT is generally focused on the control of acute bleeding (epistaxis and GI bleeding), prevention of rebleeding, and treatment of anemia (with iron supple­ ments). Patients with GI bleeding should undergo endos­ copy (or push enteroscopy) and colonoscopy to look for any GI tract lesions that may be bleeding. Focal GI tract bleeding can be treated with endoscopic coagulation. Hormonal therapy also has been reported as a treatment for small bowel bleeding in HHT.300 Patients who have symp­ tomatic or large cerebral or pulmonary AVMs should be considered for radiologic embolization of these lesions (see Chapter 36).

Blue Rubber Bleb Nevus Syndrome

Blue rubber bleb nevus syndrome is a rare syndrome characterized by venous malformations in the skin, soft tissues, and GI tract. Bleeding usually occurs in childhood and continues into adulthood and results in chronic iron deficiency requiring iron replacement and transfusions. On endoscopy, lesions appear as large protuberant polypoid venous blebs; they can occur anywhere in the GI tract, but especially in the small bowel and colon, and can be treated by endoscopic band ligation or surgical resection (see Chapter 36).301,302

Meckel’s Diverticulum

A Meckel’s diverticulum is a congenital blind intestinal pouch that results from incomplete obliteration of the vitel­ line duct during gestation.303 Characteristic features of Meckel’s diverticula have been described by the “rule of

317

318

Section III  Symptoms, Signs, and Biopsychosocial Issues twos”: They occur in 2% of the population, are found within 2 feet of the ileocecal valve, are 2 inches long, result in a complication in 2% of cases, have two types of ectopic tissue (gastric and pancreatic) within the diverticulum, present clinically most commonly at age 2 (with intestinal obstruction), and have a male-to-female ratio of more than 2 : 1. The most common complications of Meckel’s diver­ ticula are bleeding, obstruction, and diverticulitis, which can occur in children or adults. Histopathologic evaluation of bleeding diverticula reveal ectopic gastric mucosa, which can lead to acid secretion and ulceration, in 75% of patients. The diagnostic test for a Meckel’s diverticulum is a 99mTcpertechnetate scan (Meckel’s scan), because technetium pertechnetate has an affinity for gastric mucosa. Meckel’s scans have a high specificity (almost 100%) and positive predictive value but can be negative in the 25% to 50% of patients in whom the diverticulum does not contain ectopic gastric mucosa.304 The accuracy of the Meckel’s scan can be improved with administration of an H2 receptor antagonist for 24 to 48 hours before the test. Meckel’s diverticula also have been diagnosed by capsule endoscopy and doubleballoon enteroscopy (via an oral or rectal approach; see Chapter 96).

Nonsteroidal Anti-Inflammatory Drug-Induced Small Intestinal Erosions and Ulcers

Mucosal erosions or ulcers that can be seen on capsule endoscopy develop in 25% to 55% of patients who take full-dose nonselective NSAIDs.305-309 Patients who take selective COX-2 inhibitors have lower rates of mucosal ulcers on capsule endoscopy (see Chapter 115).

Small Intestinal Neoplasms

Tumors of the small intestine comprise only 5% to 7% of all GI tract neoplasms but are the most common cause of obscure GI bleeding in patients younger than 50 years.310 The most common small intestine neoplasms are adenomas (usually duodenal), adenocarcinomas (Fig. 19-23), carcinoid tumors (usually ileal), gastrointestinal stromal tumors, lym­ phomas, hamartomatosis polyps (Peutz-Jeghers syndrome), and juvenile polyps (see Chapters 29 to 31, 121, and 122).

Small Intestinal Diverticula

The duodenum is the most common site of small intestinal diverticula. In one large series,311 79% of small intestinal

Figure 19-23.  Ileal adenocarcinoma detected on deep balloon enteroscopy in a patient with a history of hereditary nonpolyposis colorectal cancer who had obscure overt gastrointestinal bleeding. The lesion was initially visualized on a video capsule endoscopy study.

diverticula occurred in the duodenum, 18% were in the jejunum or ileum, and only 3% were in all three segments— duodenum, jejunum, and ileum. Duodenal diverticula are noted in up to 20% of the population, with an increasing frequency with age.311-314 They usually are located along the medial wall of the second part of the duodenum within 1 to 2 cm of the ampulla of Vater. Bleeding from a duodenal diverticulum appears to be rare. Several reports have described bleeding from a duodenal diverticulum that was managed endoscopically.314,315 Jejunal and ileal diverticula occur in 1% to 2% of the population, are most commonly associated with a motility disorder and small intestinal bac­ terial overgrowth, and only rarely have been associated with bleeding (see Chapter 23).

Dieulafoy’s Lesion of the Small Intestine

Several reports have described Dieulafoy’s lesions of the duodenum, jejunum, and ileum.316 Most affected persons are younger than 40 years, in contrast to those with gastric Dieulafoy’s lesions, who tend to be older. The lesions are often challenging to find and in the past were detected by angiography and intraoperative endoscopy. Currently, capsule endoscopy can localize these lesions, which can be treated via a single- or double-balloon enteroscope.

DIAGNOSTIC TESTS Radiologic Studies

Barium small bowel follow-through has long been used to evaluate the small intestine because it is relatively easy to perform and readily available, but it has a low yield for determining the cause of obscure GI bleeding because of its limited ability to distend the bowel and visualize mucosal surfaces. Barium enteroclysis involves placement of a naso­ enteric catheter to infuse contrast at a variable rate and to insufflate air to produce an air contrast barium radiograph. Barium enteroclysis is more accurate than a barium small bowel follow-through, with reported rates of positive find­ ings of 10% to 20% in patients with obscure bleeding.317 Conversely, enteroclysis is more difficult for the patient to tolerate, requires more involvement by the radiologist to perform, and is not widely available. A major limitation of barium studies is the inability to visualize mucosal angioec­ tasias, which are a common cause of small bowel bleeding (see earlier). Barium studies are not recommended for patients with acute bleeding because the barium can make urgent endoscopy, colonoscopy, or angiography more diffi­ cult to perform because of residual contrast in the GI tract. CT of the abdomen has the advantage of imaging extralu­ minal structures as well as mucosal and intramural lesions in the small bowel. High-quality abdominal CT (with and without oral contrast) can show thickening of the small bowel, suggestive of Crohn’s disease or malignancy. Standard CT is less accurate than barium enteroclysis for the diagnosis of low-grade bowel obstruction, mucosal ulcerations, and fistulas. CT enteroclysis using a multidetec­ tor scanner provides better views of the small intestine than standard CT. Because placement of a nasoduodenal tube is usually required, patients sometimes receive moderate sedation for CT enteroclysis.318 CT enterography with a high volume of an oral contrast agent to distend the small bowel may have a diagnostic yield similar to that for CT entero­ clysis, without the need for a nasoduodenal tube. Magnetic resonance imaging (MRI) enteroclysis and enterography have also been described, but preliminary studies suggest that results to date are inferior to those with a multidetector CT. MRI techniques have the advantage of not exposing the patient to radiation.

Chapter 19  Gastrointestinal Bleeding Nuclear medicine studies and angiography can be used to evaluate obscure GI bleeding. A Meckel’s (99mTcpertechnetate) scan is useful for the diagnostic evaluation of a Meckel’s diverticulum, as discussed earlier. Radionu­ clide scanning with technetium-labeled red blood cells has limited benefit because of its poor ability to localize the bleeding site in the small bowel. Angiography can be useful for patients with active, acute, small bowel bleeding because of the possibility of therapeutic embolization. Small case series have described provocative angiography, in which heparin or another anticoagulant is administered to provoke GI bleeding that has been intermittent. The technique increases the yield of detecting a bleeding lesion but at the risk of causing a life-threatening complication.319

Endoscopy

Push Enteroscopy Push enteroscopy can be performed with a colonoscope (160 to 180 cm in length) or dedicated push enteroscope (220 to 250 cm in length). These endoscopes can be used to evaluate the esophagus, stomach, duodenum, and proximal jejunum approximately 50 to 150 cm beyond the ligament of Treitz. Insertion is often limited by looping of the endo­ scope in the stomach. Push enteroscopy identifies a poten­ tial bleeding site in approximately 50% of patients, and approximately 50% of the lesions found are within reach of a standard upper endoscope, suggesting that the lesion was missed or unrecognized on the initial examination.276,320 The overall diagnostic yield of push enteroscopy is approxi­ mately 40%, with a range of 3% to 80% in various studies; the most commonly detected lesions are angioectasias.275 In our UCLA CURE hemostasis experience in patients with recurrent, overt, severe obscure GI bleeding manifesting as melena, the diagnostic yield has been 80%.48 The lesions were categorized as those missed by upper endoscopy, those in the duodenum (first to fourth portion), and those in the jejunum; most lesions were in reach of a push enteroscope. Focal lesions were treated endoscopically, biopsied, or tat­ tooed. Patients in whom a diagnosis was not made by push enteroscopy underwent further studies (see Fig. 19-5). Intraoperative Endoscopy and Surgical Exploration Surgical exploration of the small intestine can be performed when other studies are nondiagnostic. At surgery, the small bowel should be palpated (“running the bowel”) to detect mass lesions. In general, a standard exploratory laparotomy or laparoscopy is performed first to lyse any adhesions and to look for obvious tumors, a Meckel’s diverticulum, or large vascular lesions. The small bowel is usually extracted through the abdominal incision to allow the surgeon to assist with advancement of an endoscope within the lumen of the GI tract, which allows mucosal visualization as well as transillumination. Any endoscope can be used (panen­ doscope, pediatric colonoscope, or push enteroscope), depending on the route of access. The endoscope can be passed transorally for a natural orifice luminal examination or via an enterotomy with use of a sterile endoscope. Because air insufflation will distend the entire small intes­ tine and thereby make laparoscopic or open visualization difficult, the surgeon should pinch the intestine, manually or with an atraumatic clamp, distal to the tip of the endo­ scope, to trap enough air to permit visualization. Addition­ ally, insufflation of the bowel with carbon dioxide rather than room air allows faster diffusion of air out of the bowel. The surgeon helps advance the endoscope by pleating the small bowel over the endoscope. Any lesion identified can be addressed surgically or endoscopically, depending on the nature of the lesion. Most series report

complete enteroscopy of the entire small bowel in 50% to 75% of cases.321,322 The diagnostic yield of intraoperative enteroscopy ranges from 58% to 88%, but rebleeding after intraoperative enteroscopy has also been reported in 13% to 60% of patients.275 The moderate performance character­ istics, as well as risks of surgical exploration, limit this procedure as a diagnostic tool, but in selected patients, combined endoscopic and surgical evaluation can be useful and definitive. Capsule Endoscopy With capsule endoscopy, the patient ingests a pill camera that transmits images of the small intestine over the course of approximately eight hours. In patients with severe, recurrent GI bleeding, this technique can identify a transi­ tion point at which fresh blood appears in the small bowel and thereby possibly detect a potential bleeding site. Capsule endoscopy does not permit the application of therapy and can only localize a lesion in the small bowel on the basis of the time of passage down the small intestine, as determined by sensors on the abdomen and telemetry. The information can be useful, however, in directing subsequent therapeutic procedures such as deep (balloon) enteroscopy, angiography, or surgery. Although capsule endoscopy may occasionally detect gastric, duodenal, or colonic lesions, it is not a substitute for upper endoscopy and colonoscopy. Compared with small bowel barium studies, capsule endoscopy has significantly improved detection rates for small bowel lesions (67% vs. 8%) and findings that influ­ ence clinical management (42% vs. 6%).323,324 A small series has found capsule endoscopy to be superior to CT entero­ clysis for the diagnosis of obscure GI bleeding because of its ability to identify angioectasias.325 An evaluation of published studies that have compared push enteroscopy with capsule endoscopy in patients with obscure bleeding (79% overt, 21% occult) has found the average rate of positive findings to be 23% for push enter­ oscopy and 63% for capsule endoscopy.275 A similar result was found in a meta-analysis of published trials and abstracts; the diagnostic yield for push enteroscopy was 28% and that for capsule endoscopy was 63%.324 A random­ ized trial that compared push enteroscopy with capsule endoscopy as a first-line approach to obscure GI bleeding reported identification of a bleeding source in 24% of the push enteroscopy examinations and 50% of the capsule studies (P = 0.02).326 In this study, capsule endoscopy missed lesions in 8% of patients; all the missed lesions were within reach of a standard upper endoscope. Capsule endoscopy was compared with intraoperative endoscopy in one study of 47 patients who underwent both procedures, primarily for obscure overt GI bleeding.327 Using intraoperative endoscopy as the gold standard, capsule endoscopy had a sensitivity of 95%, specificity of 75%, positive predictive value of 95%, and negative predictive value of 85%. Most of the bleeding lesions were angioectasias. Several studies have found that the diagnostic yield of capsule endoscopy increases in the setting of ongoing or recent (less than two weeks) overt GI bleeding or severe chronic GI bleeding (hemoglobin < 10 g/dL, iron deficiency anemia, or more than one overt bleeding episode).327-329 In a study from Greece of 34 patients who had active, mild to moderate overt GI bleeding, and negative upper endoscopy and colonoscopy results, and who underwent an urgent capsule endoscopy study while still in the hospital, the diagnostic rate was 92%, as defined by the identification of a bleeding lesion (18 angioectasias, 3 ulcers, 2 tumors) or

319

320

Section III  Symptoms, Signs, and Biopsychosocial Issues the segment of intestine with bleeding (11 patients).24 By contrast, the same group from Greece found that the diag­ nostic yield of capsule endoscopy in patients with obscure occult bleeding and iron deficiency anemia was 57% (angio­ ectasias in 24%, multiple jejunal or ileal ulcers in 12%, multiple erosions in 8%, solitary ulcers in 6%, polyps in 4%, and other tumors in 4%).330 Deep Enteroscopy of the Jejunum and Ileum Specially designed, ultraflexible, 200-cm-long enteroscopes are used in conjunction with an overtube to advance the endoscope by pleating the small intestine over it. The avail­ able systems include a double-balloon endoscope (with a balloon on the tip of the endoscope and another balloon on the overtube), a single-balloon system (a balloon on the overtube only), and a spiral overtube (no balloon used). All enteroscopes work by pleating the small intestine over the endoscope. These enteroscopes can be inserted orally (ante­ grade) and advanced into the proximal to midileum or inserted rectally (retrograde) and advanced to the distal to midileum. Rarely, a complete enteroscopy of the small intestine to the cecum can be performed via the antegrade approach. Deep enteroscopy allows not only visualization but also interventions such as biopsy, hemostasis, and tat­ tooing of lesions. The endoscopes used for deep enteros­ copy have standard working channels that allow the passage of accessories such as biopsy forceps, MPEC thermal probes, endoscopic hemoclips, and injection needles that fit through a standard colonoscope. The risks of deep enteroscopy are similar to those for push enteroscopy or colonoscopy, with the additional small risk of pancreatitis. The preponderance of available data is derived from studies using the doubleballoon enteroscopy system. A compilation of 12 case series of double-balloon enter­ oscopy for obscure bleeding in 723 patients found an overall diagnostic yield of 65% (Table 19-12).275 Comparative studies of capsule endoscopy and double-balloon enteros­ copy have revealed a slightly higher diagnostic yield for capsule endoscopy. The agreement between the two approaches in one large multicenter study of 115 patients was 74% for angioectasias, 96% for ulcers, 94% for polyps, and 96% for other large tumors.331 Another comparative study found that for patients with obscure bleeding, the agreement between the two was 92%, but the yield in a given segment of intestine in patients with polyposis was only 33% for capsule endoscopy compared with 67% for double-balloon enteroscopy; however, capsule endoscopy may detect polyps beyond the reach of the double-balloon enteroscope.332

Table 19-12  Small Intestinal Lesions Found in 488 Patients during Double-Balloon Enteroscopy for Obscure Gastrointestinal Bleeding LESION None Angioectasias Ulcerations Malignancy Other

FREQUENCY, % (RANGE) 40 31 13 8 6

(0-57) (6-55) (2-35) (3-26) (2-22)

Data from Raju GS, Gerson L, Das A, Lewis B. American Gastroenterological Association (AGA) Institute technical review on obscure gastrointestinal bleeding. Gastroenterology 2007; 133:1697-717.

Overall Approach to the Patient with Overt Obscure Gastrointestinal Bleeding

For patients with unexplained overt GI bleeding and nega­ tive upper endoscopy and colonoscopy results, capsule endoscopy is generally recommended as the next step. If capsule endoscopy reveals a lesion in the proximal jejunum, push enteroscopy can be performed. If a lesion is found in the mid–small intestine, deep enteroscopy or surgery may be considered, depending on the nature of the lesion. A lesion in the terminal ileum may prompt deep enteroscopy via the colonic route. If no lesion is detected on capsule endoscopy, but a high suspicion for a lesion remains, capsule endoscopy should be repeated or deep enteroscopy performed. With the increased avail­ ability of deep enteroscopy equipment, deep enteroscopy could become the preferred initial diagnostic step before capsule endoscopy. Modeling studies have suggested that this approach might be a cost-effective strategy,333,334 but the question ideally should be addressed in a randomized study. The UCLA CURE group’s algorithm for the management of patients who have had unexplained, severe, overt GI bleeding, with a history of melena and the need for blood transfusions, is shown in Figure 19-5. For such patients, the diagnostic yield is over 80%.48

OBSCURE OCCULT GASTROINTESTINAL BLEEDING AND IRON DEFICIENCY BLEEDING FECAL OCCULT BLOOD

Occult GI bleeding is usually detected with a routine fecal occult blood test (FOBT), with no visible blood in the stool and with or without iron deficiency. Normal fecal blood loss is 0.5 to 1.5 mL/day.335 Many FOBTs are available for detect­ ing increased amounts blood in the stool and are described in detail in Chapter 123. The approach to the patient with a positive FOBT result depends on why the test was obtained. If the FOBT was obtained for colon cancer screening in a patient older than 50 years, the patient should undergo colonoscopy and pos­ sibly upper endoscopy. Upper endoscopy should also be considered in a patient with a positive FOBT result who does not have iron deficiency anemia; this recommendation is based on the results of a study of 248 patients with fecal occult blood in whom more lesions were found in the UGI tract by upper endoscopy (mostly esophagitis, gastritis, and ulcers) than in the colon by colonoscopy (mostly large adenomas and cancer).336 If the FOBT was performed for iron deficiency anemia, the patient should be evaluated with upper endoscopy and colonoscopy. If both examina­ tions are negative, the small bowel should be imaged, as described earlier, with capsule endoscopy, possibly fol­ lowed by deep enteroscopy if a lesion is detected on capsule endoscopy. Although colon cancer screening with FOBT generally is based on six samples of spontaneously passed stool, a posi­ tive FOBT result, not uncommonly, may be found when stool is obtained during digital examination of the rectum. Although a digital rectal examination could potentially cause trauma to the anal canal, several studies have found no increase in the false-positive rate of FOBTs when stool is obtained from a digital examination.337,338 Therefore, a positive FOBT result should be approached in the same manner regardless of the method by which the stool sample was obtained. Additionally, a single negative FOBT result

Chapter 19  Gastrointestinal Bleeding on a digital rectal examination is not considered adequate colon cancer screening and does not reduce a patient’s chances of having advanced neoplasia.339

IRON DEFICIENCY ANEMIA

Iron deficiency anemia is common, with a frequency of 2% to 5% in adult men and postmenopausal women.340 Iron deficiency anemia represents 4% to 13% of all referrals for outpatient gastroenterology consultation.341 The approach to iron deficiency anemia depends on the patient’s gender and the result of an FOBT. Young women with iron deficiency anemia should be considered to have menstrual blood loss as the cause of anemia and, depending on clinical circumstances, may not need a GI evaluation. By contrast, men and postmenopausal women with iron defi­ ciency anemia should always be evaluated for a GI cause of iron deficiency. Iron deficiency anemia should be considered in patients with a low mean corpuscular volume (MCV) and anemia. In iron deficiency anemia, the serum iron concentration is decreased, and the level of transferrin (TIBC) is increased. A transferrin saturation index (serum iron divided by TIBC) lower than 15% is a sensitive indicator of iron deficiency anemia. A serum ferritin level lower than 15 ng/mL has a sensitivity of 59% and specificity of 99% for iron defi­ ciency, whereas a cutoff ferritin level of 41 ng/mL has a sensitivity and specificity of 98%.342 A bone marrow aspi­ rate can provide information about body stores of iron but is rarely necessary. Iron deficiency can result from overt or occult blood loss (from GI tract luminal lesions, menses, epistaxis, pulmonary lesions, or urinary tract lesions), intestinal iron malabsorp­ tion (as in celiac disease or gastric atrophy or after gastric bypass surgery), treatment with erythropoietin (because of excess iron requirements), and red blood cell destruction (hemolysis). The GI evaluation of a patient with iron deficiency should focus on endoscopy (upper and lower) to detect treatable lesions, especially malignancies. Recognizing iron malab­ sorption from the GI tract as a cause of iron deficiency is especially important. The duodenum is the site of iron absorption in the small intestine. Most dietary iron is in the ferric form, but only the ferrous form of iron can be absorbed by the duodenum. Ascorbic acid at a low pH is required to release nonheme iron and convert it to the ferrous form for absorption in the small intestine.343 Several studies have shown that 20% to 30% of patients with iron deficiency anemia have gastric atrophy and therefore do not produce an acid milieu that facilitates iron absorption.340,344,345 Iron deficiency anemia has also been associated with H. pylori infection.346 These studies suggest that gastric biopsies should be obtained during upper endoscopy in patients with unexplained iron deficiency anemia (see Chapters 50, 51, and 100). Celiac disease commonly manifests as iron deficiency anemia, primarily because of iron malabsorption resulting from blunted duodenal villi. Patients with celiac disease have been reported to have higher rates of positive FOBT results than healthy controls, but subsequent studies in which radiolabeled red cells were used did not find a true increase in blood loss.347,348 Nevertheless, it is possible that the cause of iron deficiency anemia in patients with celiac disease is multifactorial. Any patient who is evaluated for iron deficiency anemia and undergoes upper endoscopy should have duodenal biopsy samples obtained to look for celiac disease (see Chapter 104).

Patients who have undergone Roux-en-Y gastric bypass surgery are at high risk of iron malabsorption because of bypass of the duodenum, where most iron is absorbed. These patients can present with severe unexplained iron deficiency without occult blood in the stool. They often have extremely low body stores of iron and require intrave­ nous iron supplementation. The differential diagnosis of iron deficiency anemia includes anemia of chronic disease and thalassemia. In anemia of chronic disease, both the serum iron level and TIBC are low, with a normal serum ferritin level. Patients with thalassemia have a family history of anemia, spleno­ megaly, target cells on peripheral blood smear, and normal serum ferritin levels. Patients with unexplained iron deficiency anemia should undergo upper endoscopy and colonoscopy to rule out a GI tract lesion that may cause chronic blood loss. In a prospec­ tive study of 100 patients with iron deficiency anemia, GI tract lesions were found in 62 patients, with 36 having lesions in the UGI tract (mostly ulcers), 25 in the colon (mostly cancer), and 1 in both the UGI tract and colon.349 In patients with unexplained iron deficiency anemia who undergo upper endoscopy, duodenal biopsy specimens should be obtained to rule out celiac disease as a cause of iron malabsorption. Gastric biopsy samples also should be obtained to rule out gastropathy and H. pylori infection. Depending on the severity of iron deficiency anemia, even without a positive FOBT result, evaluation of the small intestine for a bleeding lesion, as discussed earlier, should be considered. If a specific cause of anemia is not identified, patients should be advised to avoid antiplatelet and antico­ agulant drugs and take supplemental iron.

KEY REFERENCES

Barkun A, Bardou M, Marshall JK. Consensus recommendations for managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2003; 139:843-57. (Ref 100.) Gralnek IM, Barkun AN, Bardou M. Management of acute bleeding from a peptic ulcer. N Engl J Med 2008; 359:928-37. (Ref 139.) Jensen DM, Kovacs TO, Jutabha R, et al. Randomized trial of medical or endoscopic therapy to prevent recurrent ulcer hemorrhage in patients with adherent clots. Gastroenterology 2002; 123:407-13. (Ref 115.) Jensen DM, Machicado GA. Diagnosis and treatment of severe hemato­ chezia. The role of urgent colonoscopy after purge. Gastroenterology 1988; 95:1569-74. (Ref 22.) Jensen DM, Machicado GA, Jutabha R, Kovacs TO. Urgent colonoscopy for the diagnosis and treatment of severe diverticular hemorrhage. N Engl J Med 2000; 342:78-82. (Ref 23.) Khuroo MS, Yattoo GN, Javid G, et al. A comparison of omeprazole and placebo for bleeding peptic ulcer. N Engl J Med 1997; 336:1054-8. (Ref 125.) Laine L. Multipolar electrocoagulation in the treatment of active upper gastrointestinal tract hemorrhage. A prospective controlled trial. N Engl J Med 1987; 316:1613-17. (Ref 174.) Lau JY, Leung WK, Wu JC, et al. Omeprazole before endoscopy in patients with gastrointestinal bleeding. N Engl J Med 2007; 356:163140. (Ref 129.) Lau JY, Sung JJ, Lam YH, et al. Endoscopic retreatment compared with surgery in patients with recurrent bleeding after initial endo­ scopic control of bleeding ulcers. N Engl J Med 1999; 340:751-6. (Ref 144.) Lau JY, Sung JJ, Lee KK, et al. Effect of intravenous omeprazole on recurrent bleeding after endoscopic treatment of bleeding peptic ulcers. N Engl J Med 2000; 343:310-16. (Ref 127.) Raju GS, Gerson L, Das A, Lewis B. American Gastroenterological Asso­ ciation (AGA) Institute technical review on obscure gastrointestinal bleeding. Gastroenterology 2007; 133:1697-717. (Ref 275.) Rockey DC, Koch J, Cello JP, et al. Relative frequency of upper gastro­ intestinal and colonic lesions in patients with positive fecal occultblood tests. N Engl J Med 1998; 339:153-9. (Ref 336.)

321

322

Section III  Symptoms, Signs, and Biopsychosocial Issues Rockey DC, Cello JP. Evaluation of the gastrointestinal tract in patients with iron-deficiency anemia. N Engl J Med 1993; 329:1691-5. (Ref 349.) Strate LL, Ayanian JZ, Kotler G, Syngal S. Risk factors for mortality in lower intestinal bleeding. Clin Gastroenterol Hepatol 2008; 6:100410. (Ref 219.)

Sung JJ, Chan FK, Lau JY, et al. The effect of endoscopic therapy in patients receiving omeprazole for bleeding ulcers with nonbleeding visible vessels or adherent clots: A randomized comparison. Ann Intern Med 2003; 139:237-43. (Ref 136.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

20 Jaundice Steven D. Lidofsky

CHAPTER OUTLINE Bilirubin Metabolism and Measurement  323 Metabolism  323 Measurement  324 Differential Diagnosis  325 Isolated Disorders of Bilirubin Metabolism  325 Liver Disease  326 Obstruction of the Bile Ducts  329 Diagnostic Approach to Jaundice  330 History and Physical Examination  330

Jaundice (icterus), is a condition characterized by yellow discoloration of the skin, conjunctivae, and mucous membranes as a result of widespread tissue deposition of the pigmented metabolite bilirubin. Although jaundice is generally associated with liver and biliary tract disease, it has many causes. It is thus not surprising that the diagnosis and management of jaundice have challenged clinicians for centuries. Attempts to classify icteric syndromes appeared as early as the treatises of Hippocrates. By the time of Osler, distinctions were already made between biliary tract obstruction and nonobstructive causes of jaundice. In the latter part of the twentieth century, elucidation of the molecular mechanisms that underlie bilirubin metabolism, as well as the development of more sophisticated biochemical and imaging techniques, made it possible to pinpoint the cause of jaundice in most cases. Despite the impressive array of tools available today, it should be emphasized that to minimize risk to the patient, an effective approach to jaundice requires selection of diagnostic and therapeutic modalities on the basis of a careful assessment of the likelihood of possible underlying diseases. This chapter covers four major areas: (1) bilirubin metabolism; (2) differential diagnosis of jaundice; (3) role of the history, physical examination, and routine biochemical tests in narrowing the differential diagnosis and the usefulness of selected laboratory and hepatobiliary imaging studies; and (4) therapeutic approaches to the management of jaundice.

BILIRUBIN METABOLISM AND MEASUREMENT METABOLISM

Bilirubin, a hydrophobic and potentially toxic compound, is a tetrapyrrole that is an end product of heme degradation. Bilirubin metabolism has been reviewed in depth elsewhere1,2 and is summarized briefly in Figure 20-1. Each day, a healthy adult produces approximately 4 mg/kg of bilirubin (i.e., almost 0.5 mmol in a 70-kg person). Most

Initial Laboratory Studies  331 Overall Approach  332 Imaging Studies  332 Other Studies  334 Therapeutic Approaches  334 Biliary Obstruction  334 Other Conditions  335

bilirubin (70% to 80%) is derived from degradation of hemoglobin from senescent erythrocytes, and a minor component arises from premature destruction of newly formed erythrocytes in the bone marrow or circulation (i.e., ineffective erythropoiesis). Most of the remaining 20% to 30% is formed from breakdown of hemoproteins, such as catalase and cytochrome (CYP family) oxidases, in hepatocytes. Although nonhemoglobin heme-containing proteins are also present in extrahepatic tissues, their mass is so small or their turnover rate so slow (as for myoglobin) that their overall contribution to bilirubin production is minimal. The breakdown of heme to bilirubin occurs by a two-step process. First, heme is converted to biliverdin by heme oxygenase, which functions predominantly as an integral membrane protein of the smooth endoplasmic reticulum. Second, biliverdin is converted rapidly to bilirubin by the cytosolic protein biliverdin reductase. Catabolism of erythrocyte-derived hemoglobin to bilirubin takes place prim­arily in reticuloendothelial cells in the spleen, liver, and bone marrow. By contrast, free hemoglobin, haptoglobinbound hemoglobin, and methemalbumin are catabolized to bilirubin predominantly in hepatocytes. Bilirubin circulates in plasma tightly, but noncovalently, bound to albumin. Excretion of bilirubin requires con­ version to water-soluble conjugates by hepatocytes and subsequent secretion into bile. Bilirubin metabolism and elimination is a multistep process for which several inherited disorders have been identified (see later). Bilirubin is taken up across the sinusoidal (basolateral) membrane of hepatocytes by a carrier-mediated mechanism. The uptake of bilirubin is inhibited competitively by certain organic anions such as sulfobromophthalein (BSP) and indocyanine. Bilirubin uptake has been suggested to be mediated by a liver-specific sinusoidal organic anion transport protein, (OATP1B1, SLC21A6), but this is not entirely certain.3,4 After uptake, bilirubin is directed by cytosolic binding proteins (e.g., glutathione S-transferase B, fatty acid binding protein) to the endoplasmic reticulum, where it is conjugated with uridine diphosphate (UDP)–glucuronic acid by the enzyme bilirubin UDP–glucuronyl transferase (B-UGT).

323

324

Section III  Symptoms, Signs, and Biopsychosocial Issues jugated by bacterial enzymes in the terminal ileum and colon and converted to colorless tetrapyrroles called urobilinogens. Up to 20% of urobilinogens are resorbed and ultimately excreted in bile and urine.

Hemoglobin Other hemoproteins Br Alb-Br Plasma Br

BrG

Urine MRP3?

OATP?

B-UGT BrG

Hepatocytes

MRP2

BrG Bile Figure 20-1.  Overview of bilirubin formation, metabolism, and transport. Heme from hemoglobin and other hemoproteins is converted to biliverdin and then to bilirubin (Br), predominantly in reticuloendothelial cells in the bone marrow and spleen. Br is released into plasma (in its unconjugated form), where it is tightly but reversibly bound to albumin (Alb). Br is then taken up at the sinusoidal membrane of hepatocytes, possibly via a member of the organic anion transporter (OATP) family. Br is conjugated, via the activity of bilirubin UDP-glucuronyl transferase (B-UGT), to form bilirubin mono- and di-glucuronides (BrG). Biliary secretion of BrG occurs at the canalicular membrane by the multispecific organic anion transporter MRP2. Under physiological conditions, the vast majority of BrG is eliminated in bile. Small amounts of BrG are transported at the sinusoidal membrane back into plasma, possibly via the multispecific organic anion transporter MRP3. Plasma BrG enters the renal circulation, where it undergoes glomerular filtration and elimination into urine. Thus, under normal conditions, at least 95% of bilirubin in plasma is present in the unconjugated form. If abnormally high concentrations of BrG are retained over a prolonged period, BrG-Alb complexes, which do not dissociate and cannot undergo glomerular filtration, are formed. MRP, multidrug resistanceassociated protein.

Conjugation converts hydrophobic bilirubin into a watersoluble form suitable for excretion. Conjugated bilirubin is then directed primarily toward the canalicular (apical) membrane, where it is transported into the bile canaliculus by an adenosine triphosphate (ATP)-dependent export pump. The responsible protein, multidrug resistance– associated protein-2 (MRP2, ABCC2), appears to function as a multispecific transporter of various organic anions (including BSP, glutathione, and conjugated bile salts).5 Small amounts of bilirubin glucuronides are secreted across the sinusoidal membrane via a pathway postulated to be mediated by a distinct multispecific organic ion export pump, MRP3 (ABCC3)6; conjugated bilirubin in plasma undergoes renal elimination (see Fig. 20-1). This pathway may be upregulated in disorders characterized by cholestasis (impaired bile flow). With prolonged cholestasis (or a metabolic disorder of conjugated hyperbilirubinemia; see later), an increasing proportion of conjugated bilirubin in plasma becomes covalently bound to albumin, and this covalently bound bilirubin cannot be excreted into urine. Approximately 80% of bilirubin in human bile is in the form of diglucuronides. Almost all the rest is in the form of monoglucuronides, and only trace amounts are unconjugated. Resorption of conjugated bilirubin by the gallbladder and intestine is negligible; however, bilirubin can be decon-

MEASUREMENT

The normal bilirubin concentration in the serum of adults is lower than 1 to 1.5 mg/dL. In general, jaundice is not evident until the serum bilirubin concentration exceeds 3 mg/dL. In healthy persons, most bilirubin circulates in its unconjugated form; less than 5% of circulating bilirubin is present in conjugated form. In cholestatic conditions, the proportion of unconjugated bilirubin may increase as a consequence of upregulated MRP3 expression. The importance of accurate measurement of bilirubin is underscored by its incorporation as a critical variable in scoring systems such as the Model for End-stage Liver Disease (MELD), which provide estimates of survival in various acute and chronic liver disorders.7 Serum bilirubin is detected conventionally by the diazo van den Bergh reaction. With this colorimetric method, bilirubin is cleaved by compounds such as diazotized sulfanilic acid to form an azodipyrrole that can be assayed by spectrophotometry. Conjugated bilirubin is cleaved rapidly (directly) by diazo reagents. By contrast, unconjugated bilirubin reacts slowly with diazo reagents because the site of chemical cleavage is rendered inaccessible by internal hydrogen bonding. Therefore, reliable measurement of total bilirubin concentration requires the addition of an accelerator compound, such as ethanol or urea, which disrupts this hydrogen bonding and facilitates the cleavage of unconjugated bilirubin by the diazo reagent. The concentration of the indirect bilirubin fraction is calculated by subtracting the direct bilirubin concentration (measured in the absence of the accelerator compound) from that of the total bilirubin concentration (measured in the presence of the accelerator compound). Although the direct bilirubin concentration is influenced by changes in conjugated bilirubin levels, the two are not equivalent. Similarly, indirect bilirubin is not equivalent to unconjugated bilirubin. In particular, reliance on direct and indirect bilirubin measurements can lead to errors in the diagnosis of isolated disorders of bilirubin metabolism (e.g., suspected Gilbert’s syndrome; see later). Many clinical laboratories have abandoned measurements of direct and indirect bilirubin and instead use automated reflectance spectroscopic assays that provide more accurate estimates of conjugated and unconjugated bilirubin. These assays are useful clinically in the management of physiologic jaundice of the newborn (see later), in which neurotoxicity may result from the passage of unconjugated bilirubin across the blood-brain barrier (kernicterus). In disorders characterized by prolonged cholestasis, however, such assays may underestimate the conjugated bilirubin concentration, because they do not accurately detect albumin-bound conjugated bilirubin (so-called delta bilirubin). Indeed, if an isolated disorder of bilirubin metabolism is suspected, the diagnosis may require more sophisticated chromatographic techniques that precisely measure the concentrations of unconjugated, monoglucuronidated, and diglucuronidated bilirubin, as well as conjugated bilirubin-albumin complexes.2 In practice, these techniques are not widely used. Even with such accurate methods, measurements of con­jugated and unconjugated bilirubin will not distinguish hepatic disorders from biliary obstruction. Therefore, in most cases, these tests are of limited use.

Chapter 20  Jaundice DIFFERENTIAL DIAGNOSIS Jaundice can result from an increase in the formation of bilirubin or a decrease in the hepatobiliary clearance of bilirubin. From a practical standpoint, conditions that produce jaundice can be classified under the broad categories of isolated disorders of bilirubin metabolism, liver disease, and obstruction of the bile ducts (Table 20-1).

ISOLATED DISORDERS OF BILIRUBIN METABOLISM Unconjugated Hyperbilirubinemia

Three basic mechanisms can lead to isolated unconjugated hyperbilirubinemia: (1) increased bilirubin production; (2) decreased hepatocellular uptake of unconjugated bilirubin; Table 20-1  Differential Diagnosis of Jaundice and Hyperbilirubinemia Isolated Disorders of Bilirubin Metabolism Unconjugated Hyperbilirubinemia Increased bilirubin production (e.g., hemolysis, ineffective erythropoiesis, blood transfusion, resorption of hematomas) Decreased hepatocellular uptake (e.g., drugs such as rifampin, Gilbert’s syndrome [secondary mechanism]) Decreased conjugation (e.g., Gilbert’s syndrome, Crigler-Najjar syndrome, physiologic jaundice of the newborn, drugs such as indinavir, atazanavir) Conjugated or Mixed Hyperbilirubinemia Dubin-Johnson syndrome Rotor’s syndrome Liver Disease Hepatocellular Dysfunction Acute or subacute hepatocellular injury (e.g., viral hepatitis, hepatotoxins [such as ethanol, acetaminophen, Amanita phalloides]; drugs such as isoniazid, phenytoin; ischemia such as in hypotension, vascular outflow obstruction; metabolic disorders such as Wilson disease, Reye’s syndrome; pregnancy-related as in acute fatty liver of pregnancy, pre-eclampsia) Chronic hepatocellular disease (e.g., viral hepatitis; hepatotoxins such as ethanol, vinyl chloride, vitamin A; autoimmune hepatitis; metabolic disorder such as hemochromatosis, Wilson disease, nonalcoholic fatty liver disease, α1-antitrypsin deficiency; celiac disease) Hepatic Disorders with Prominent Cholestasis Diffuse infiltrative disorders (e.g., granulomatous diseases such as mycobacterial infections, sarcoidosis, lymphoma, Wegener’s granulomatosis; amyloidosis; malignancy) Cholangiocyte injury (e.g., primary biliary cirrhosis; graft-versus-host disease; drugs such as erythromycin, trimethoprimsulfamethoxazole; cystic fibrosis) Miscellaneous conditions (e.g., benign recurrent intrahepatic cholestasis; drugs such as estrogens, anabolic steroids; total parenteral nutrition; bacterial infections; paraneoplastic syndromes; intrahepatic cholestasis of pregnancy; benign postoperative cholestasis) Obstruction of the Bile Ducts Choledocholithiasis Diseases of the Bile Ducts Inflammation, infection (e.g., primary sclerosing cholangitis, AIDS cholangiopathy, hepatic arterial chemotherapy, postsurgical strictures) Neoplasms (e.g., cholangiocarcinoma) Extrinsic Compression Neoplasms (e.g., pancreatic carcinoma, metastatic lymphadenopathy, hepatocellular carcinoma, ampullary adenoma, lymphoma) Pancreatitis Vascular enlargement (e.g., aneurysm, cavernous transformation of the portal vein [portal cavernoma]) AIDS, acquired immunodeficiency syndrome.

and (3) decreased bilirubin conjugation. In each of the resulting conditions, liver function is otherwise normal, and the results of standard biochemical liver tests other than the serum bilirubin concentration are normal. Increased Bilirubin Production Processes that can generate excessive bilirubin production include hemolysis, ineffective erythropoiesis, and resorption of a hematoma.2 Jaundice may thus complicate the clinical course of patients with hemolytic anemias, megaloblastic anemia from folate or vitamin B12 deficiency, iron deficiency anemia, sideroblastic anemia, and polycythemia vera. With these disorders, bilirubin concentration does not generally exceed 4 to 5 mg/dL. Jaundice can follow massive blood transfusions, because the foreshortened lifespan of transfused erythrocytes leads to excessive hemoglobin release. Hyperbilirubinemia resulting from resorption of hematomas and blood transfusions also may develop in patients who have experienced major trauma.8 Decreased Bilirubin Uptake A decrease in hepatocellular uptake of bilirubin can be seen with certain drugs. For example, the antituberculosis agent rifampin has been shown to inhibit bilirubin uptake by hepatocytes competitively and may produce jaundice by inhibiting the transport protein OATP1B1 (SLC21A6); similar effects may be produced by the immunosuppressive drug cyclosporine A.9,10 Decreased bilirubin uptake also may contribute to phenotypic expression of the hereditary disorder Gilbert’s syndrome, in which the predominant abnormality is impaired bilirubin conjugation resulting from reduced B-UGT activity.11 Decreased Bilirubin Conjugation Three autosomally inherited disorders of unconjugated hyperbilirubinemia are attributable to impaired bilirubin conjugation (Table 20-2). The most common of these dis­ orders is Gilbert’s syndrome, which has a prevalence of approximately 10% in white populations. The disorder is entirely benign and rarely produces clinical jaundice. Serum bilirubin levels may rise two- to threefold with fasting or dehydration but are generally below 4 mg/dL. Patients with Gilbert’s syndrome typically present during or after adolescence, when isolated hyperbilirubinemia is detected as an incidental finding on routine multiphasic biochemical screening. The molecular basis of Gilbert’s syndrome has been linked to a reduction in transcription of the B-UGT gene UGT1A1 as a result of mutations in the promoter region and, less commonly, in the coding region.1 Mutations in the coding region of UGT1A1 appear to be responsible for Crigler-Najjar syndrome.12 In type I CriglerNajjar syndrome, B-UGT activity is absent, and many patients die of kernicterus in the neonatal period (see Table 20-2). Phototherapy (see later) is required to prevent kernicterus, and liver transplantation can be lifesaving. Persons with type II Crigler-Najjar syndrome have markedly reduced B-UGT activity, with serum bilirubin levels between those of patients with Gilbert’s syndrome and those with type I Crigler-Najjar syndrome (see Table 20-2). In contrast to patients with type I Crigler-Najjar syndrome, those with type II Crigler-Najjar syndrome are not ill during the neonatal period and may not be diagnosed until early childhood. Although the degree of jaundice can wax and wane, most patients with type II Crigler-Najjar syndrome experience a fall in serum bilirubin levels to 2 to 5 mg/dL with phenobarbital, an agonist for the constitutive androstane receptor CAR, which increases expression of UGT1A1

325

326

Section III  Symptoms, Signs, and Biopsychosocial Issues Table 20-2  Hereditary Disorders of Bilirubin Metabolism and Transport Syndrome CRIGLER-NAJJAR TYPE I

CRIGLER-NAJJAR TYPE II

6%-12% UGT1A1 ↓Bilirubin conjugation

Very rare UGT1A1 No bilirubin conjugation

Uncommon UGT1A1 ↓↓Bilirubin conjugation

≤3 in absence of fasting or hemolysis, almost all unconjugated Usually normal, occasional ↑lipofuscin

Usually >20 (range, 17-50), all unconjugated Normal

Usually <20 (range, 6-45), almost all unconjugated Normal

Other distinguishing features

↓Bilirubin concentration with phenobarbital

No response to phenobarbital

↓Bilirubin concentration with phenobarbital

Prognosis

Normal

Usually normal

Treatment

None

Death in infancy if untreated Phototherapy as a bridge to liver transplantation

PARAMETER

GILBERT’S

Incidence Gene affected Metabolic defect

Plasma bilirubin (mg/dL) Liver histology

Phenobarbital for ↑↑bilirubin concentration

DUBIN-JOHNSON

ROTOR’S

Uncommon MRP2 Impaired canalicular export of conjugated bilirubin Usually <7, about half conjugated

Rare Unknown Impaired canalicular export of conjugated bilirubin Usually <7, about half conjugated

Coarse pigment in centrilobular hepatocytes ↑Bilirubin concentration with estrogens; ↑↑urinary coproporphyrin I/III ratio; slow BSP elimination kinetics with secondary rise Normal

Normal

Avoid estrogens

None available

Mild ↑urinary coproporphyrin I/III ratio; very slow BSP* elimination kinetics without secondary rise Normal

*No longer available. BSP, sulfobromophthalein; MRP2, multidrug resistance–associated protein-2 gene; UGTIA1, bilirubin UDP-glucuronyl transferase gene.

and thus increases B-UGT activity.13 Such patients have normal life expectancies and do not manifest neurologic impairment. A related disorder of bilirubin metabolism is physiologic jaundice of the newborn. This syndrome, which is believed to result from delayed developmental expression of B-UGT, is characterized by transient jaundice that generally resolves rapidly in the neonatal period. A brief course of phototherapy may be required to prevent kernicterus. B-UGT is inhibited competitively by the viral protease inhibitors indinavir and atazanavir, which produce hyperbilirubinemia in more than 25% of patients who receive these agents.14,15

Conjugated or Mixed Hyperbilirubinemia

A selective decrease in bilirubin secretion into the bile canaliculus may produce conjugated or mixed hyperbiliru­ binemia (i.e., an increase in conjugated and unconjugated bilirubin concentrations). Such a defect underlies two autosomally inherited disorders, Dubin-Johnson syndrome and Rotor’s syndrome. Each of these disorders is associated with a benign clinical course. In Dubin-Johnson syndrome, the molecular defect has been linked to an absence of expression of or impaired canalicular membrane targeting of the multispecific organic anion transporter MRP2.5 Interestingly, in Dubin-Johnson syndrome and in selected cholestatic disorders (e.g., primary biliary cirrhosis), compensatory up-regulation of the sinusoidal export protein MRP3 has been reported.16 Up-regulation of MRP3 may prevent hepatocellular overload by potentially toxic organic anions that are normally secreted by MRP2. The molecular basis of Rotor’s syndrome is unknown and does not appear to involve mutations in MRP2.17 In both Dubin-Johnson and Rotor’s syndromes, global hepatic function is preserved. Serum bilirubin levels are elevated, but serum levels of

other commonly measured liver biochemical tests are normal. Dubin-Johnson and Rotor’s syndromes can be distinguished biochemically and histologically (see Table 20-2). In Dubin-Johnson syndrome, hepatocytes contain a characteristic black pigment that is not seen in Rotor’s syndrome. This pigment is believed to result from lysosomal deposition of aromatic amino acid metabolites that are putative substrates for MRP2.5 Liver biopsy is generally unnecessary in the diagnostic evaluation of patients suspected to have Dubin-Johnson or Rotor’s syndrome, however, because neither disorder is associated with an adverse clinical outcome.

LIVER DISEASE

Jaundice is a common feature of generalized hepatic dysfunction. In contrast to isolated disorders of bilirubin metabolism, icteric liver disease is characterized by an increase in serum bilirubin concentration that generally occurs in association with abnormalities in other standard biochemical liver test results. The extensive differential diagnosis of icteric liver disease is outlined briefly here. In the discussion that follows, disorders in which hyperbilirubinemia and jaundice are simply manifestations of global hepatocellular dysfunction will be distinguished from those for which cholestasis is a major or predominant manifestation. The latter are often difficult to distinguish clinically from obstruction of the bile ducts.

Acute Hepatocellular Dysfunction

Generalized impairment of hepatocellular function can result from acute or chronic liver injury. A clue to such disorders is the presence of elevated serum activities of alanine aminotransferase (ALT) and aspartate aminotrans-

Chapter 20  Jaundice ferase (AST) (see later and Chapter 73). Among conditions that produce acute or subacute hepatocellular injury are viral hepatitis, exposure to hepatotoxins, hepatic ischemia, and certain metabolic derangements. Acute viral hepatitis often is heralded by anorexia, malaise, myalgias, or discomfort in the epigastrium or right upper abdominal quadrant before jaundice develops (see Chapters 77 to 81). Five major hepatitis viruses have been isolated. Hepatitis A and E viruses are transmitted enterally. Each typically produces a self-limited illness that does not progress to chronic liver disease. By contrast, hepatitis B, C, and D viruses are transmitted parenterally, and illness produced by these agents can be prolonged and may lead to chronic disease. Major risk factors for hepatitis B, C, and D include injection drug use, exposure to blood products, and unprotected sexual exposures. The diagnosis of each of these disorders is aided by serologic testing (see later). Many drugs and toxins produce hepatocellular injury (see Chapters 86 and 87). In particular, ingestion of acetaminophen (in large quantities) or of the mushroom Amanita phalloides may lead to hepatocellular necrosis and jaundice within several days after exposure. Toxic liver injury can have a fulminant course associated with a high mortality rate (see Chapter 93). In patients who survive, jaundice generally resolves and hepatic function recovers completely in those without preexisting liver disease. Certain drugs can produce idiosyncratic hepatocellular injury and jaundice, and these are discussed extensively elsewhere in this text (see Chapter 86). Alcoholic hepatitis should be a diagnostic consideration in the jaundiced patient with ethanol dependency, particularly when hepatomegaly and fever are present (see Chapter 84). Laboratory studies may help distinguish this entity from most other acute liver diseases (see later). Jaundice related to hepatic ischemia may result from hypotension, hypoxia, hyperthermia, or hepatic venous outflow obstruction (see Chapter 83). Thrombosis of the hepatic vein (Budd-Chiari syndrome) or sinusoidal obstruction syndrome (hepatic veno-occlusive disease) should be suspected in a patient who presents with the rapid onset of ascites and hepatomegaly; the latter syndrome is more commonly associated with jaundice and is a complication of certain cytotoxic agents, particularly in the setting of hematopoietic cell transplantation (see also Chapter 34). Wilson disease, an inherited disorder of hepatobiliary copper secretion, may manifest de novo with clinical features indistinguishable from those of acute viral hepatitis (see Chapter 75). The disease should be a diagnostic consideration in patients younger than 40 years, particularly when neurologic abnormalities are present or KayserFleischer rings are seen on slit-lamp examination of the eye. Hemolytic anemia is a part of the spectrum of Wilson disease and contributes to the disproportionate hyperbilirubinemia often present in these patients. The diagnosis of Wilson disease is confirmed by biochemical testing and liver copper analysis (see later). Reye’s syndrome, a disorder of fatty infiltration of the liver associated with impaired mitochondrial metabolism of fatty acids, may produce jaundice as a manifestation of acute liver failure (see Chapter 86 and 93). It usually follows a viral illness in children, has been associated with the ingestion of aspirin, and is heralded by nausea and vomiting; its incidence has declined markedly as a result of public health campaigns advocating the avoidance of aspirin in children.

Chronic Hepatocellular Dysfunction

In contrast with acute liver disease, jaundice does not typically develop in chronic conditions associated with hepa-

tocellular injury unless cirrhosis is present. A major cause of cirrhosis is chronic viral hepatitis, which should be a diagnostic consideration in patients with risk factors for parenteral exposure to causative agents. Diagnosis is aided by serologic testing (see later). Cirrhosis is part of the spectrum of nonalcoholic fatty liver disease, which is emerging as the most common cause of chronic hepatocellular injury in industrialized nations; major risk factors are obesity and diabetes mellitus (see Chapter 85). A similar histologic picture of steatohepatitis and sinusoidal fibrosis is found in the setting of alcoholic liver disease (see Chapter 84). Toxic injury by other compounds is less likely to produce cirrhosis, although cirrhosis has been described as a manifestation of industrial exposure to vinyl chloride and as a consequence of chronic ingestion of large quantities of vitamin A (see Chapter 87). Certain hereditary metabolic liver diseases may progress to cirrhosis. Hemochromatosis, a disorder characterized by excessive intestinal iron absorption with resulting hepatocellular iron accumulation and injury, is the most common of these (see Chapter 74). Although most affected persons are asymptomatic, the presence of diabetes mellitus, arthritis, or deep pigmentation in a jaundiced person should heighten suspicion for the disorder. The diagnosis is confirmed by detection of mutations in the HFE gene or by hepatic iron analysis. Hepatocellular copper overload and injury in Wilson disease also may progress to cirrhosis (see Chapter 75). As noted, the diagnosis should be suspected in younger persons, and the disease confirmed by biochemical testing and liver copper analysis. In a jaundiced patient with chronic obstructive pulmonary disease, α1-antitrypsin deficiency should be suspected (see Chapter 76). In this disorder, mutant α1-antitrypsin is misfolded and accumulates in the endoplasmic reticulum of hepatocytes, proteasomal degradation is impaired, and liver injury results. The diagnosis can be confirmed by laboratory testing and liver biopsy (see later). Autoimmune hepatitis, a disease that may be associated with systemic complaints such as malaise, fever, and arthralgias, is more common in women than in men (see Chapter 88). The diagnosis is aided by serologic testing and liver biopsy (see later). Celiac disease (see Chapter 104) may manifest as otherwise unexplained chronic liver disease—although rarely, if ever, with jaundice.

Hepatic Disorders with Prominent Cholestasis

Intrahepatic cholestatic disorders are characterized by impaired bile formation in the absence of widespread hepatocellular injury or biliary tact obstruction. Biochemical abnormalities demonstrate predominant elevation of the serum alkaline phosphatase activity relative to ALT and AST activities. The clinical presentation of these disorders may mimic obstruction of the bile ducts and can generate great diagnostic confusion. Such disorders can be categorized histologically as those associated with infiltration of the liver, those associated with injury to cholangiocytes within intrahepatic bile ductules, and those in which major histologic changes are not evident. Infiltrative Diseases Infiltrative diseases of the liver disrupt the network of intrahepatic bile ductules and are often associated with striking cholestasis. Granulomatous diseases of the liver can be caused by the following: infections, such as tuberculosis, Mycobacterium avium complex infection (particularly in an immunocompromised host), leprosy, brucellosis, Q fever, syphilis, fungal diseases, parasitic diseases, and mononucleosis; toxins, such as beryllium, quinidine, allopurinol, and sulfonamides; and systemic disorders,

327

328

Section III  Symptoms, Signs, and Biopsychosocial Issues including sarcoidosis, lymphoma (in particular, Hodgkin’s disease), and Wegener’s granulomatosis (see Chapters 35 and 82). The most common of these disorders that produce jaundice are tuberculosis and sarcoidosis.18 Granulomatous diseases should be suspected when jaundice accompanies fever of undetermined origin; other nonspecific symptoms include night sweats and weight loss. Physical examination usually reveals hepatosplenomegaly; right upper quadrant abdominal tenderness is uncommon. Lymphadenopathy often is seen in sarcoidosis and is also a clue to an infectious cause or lymphoma. The presence of erythema nodosum suggests mycobacterial disease, sarcoidosis, or syphilis. Elevations in the blood eosinophil count should heighten suspicion of sarcoidosis, parasitic disease, or drug toxicity. Radiographic chest abnormalities often provide a clue to the diagnosis of sarcoidosis or mycobacterial infection. Ultimately, diagnosis may require liver biopsy, if other tissue is not available. Jaundice is an unusual manifestation of amyloidosis but, when present, invariably is accompanied by marked hepatomegaly.19 The diagnosis of amyloidosis also may be suspected if clinical evidence of involvement of other organs is detected, such as macroglossia, malabsorption, congestive heart failure, or peripheral neuropathy. With renal involvement, proteinuria will be present. Otherwise, no specific biochemical clues may be found. Amyloidosis often can be detected on rectal valve or abdominal wall fat pad biopsy, but if the results are negative, liver biopsy specimens are diagnostic for hepatic amyloidosis. Jaundice resulting from extensive neoplastic replacement of hepatic parenchyma usually is heralded by anorexia and weight loss. Noninvasive imaging studies generally lead to the diagnosis (see later). Disorders Involving Cholangiocyte Injury A variety of disorders can lead to cholangiocyte damage and progressive loss of bile ductules. These encompass the differential diagnosis of vanishing bile duct syndrome, a term that has been employed to describe cholestatic conditions that are associated with a paucity of small intrahepatic bile ducts. In many of these disorders, the cholangiocyte is a target of an immune-mediated inflammatory response, as is characteristic of primary biliary cirrhosis (see Chapter 89) and is part of the spectrum of graft-versus-host disease encountered in organ transplant recipients (see Chapter 34). Primary biliary cirrhosis is a disease that occurs predominantly in women. In patients with jaundice, pruritus is also usually present and fatigue is common. The skin is often hyperpigmented in patients with advanced primary biliary cirrhosis, and detection of xanthelasma or xanthomata related to hypercholesterolemia is highly suggestive of the diagnosis. Serologic testing (antimito­chondrial antibodies) is generally sufficient to establish a diagnosis of primary biliary cirrhosis, but liver biopsy may be necessary to confirm the diagnosis in selected cases (see later). Graftversus-host disease is a complication of hematopoietic and solid organ transplantation; by contrast, hepatic involvement is rare after liver transplantation. Jaundice related to graft-versus-host disease develops in approximately 10% of hematopoietic cell transplant recipients.20 Certain drugs also can produce cholestasis with inflammation of the portal tracts (see Chapter 86). These include erythromycin (particularly the estolate salt), trimethoprim-sulfamethoxazole, amoxicillin–clavulanic acid, and terbinafine.21 Clinical features that may heighten suspicion of drug-induced cholestasis include arthralgias, rash, and peripheral eosinophilia; cholestasis generally resolves within several months following discontinuation of the causative drug. The most common

inherited disorder of cholangiocyte injury is cystic fibrosis, a systemic disease that affects secretory epithelia and is linked to mutations in the CFTR gene, which encodes the cystic fibrosis transmembrane conductance regulator ion channel protein. Cholestatic hepatobiliary disease occurs in at least 30% of adults with this disorder (see Chapter 76). Cholestasis with Minimal Histologic Abnormalities Jaundice may accompany conditions characterized by minimal hepatocellular injury and histologic abnormalities. Several mechanisms contribute to cholestasis in these conditions, including mutations in the genes that encode transport proteins involved in bile formation, conditions that interfere with the function of such proteins, and reduction in transport protein expression. Benign recurrent cholestasis is an autosomal recessive disorder associated with mutations in the genes that encode transport proteins involved in bile formation—the familial intrahepatic cholestasis 1 protein (FIC1, ATP8B1) and the bile salt export pump (BSEP, ABCB11; see Chapters 64 and 76).22,23 FIC1 (ATP8B1) is a P-type ATPase found in cholangiocytes as well as on the canalicular membrane of hepatocytes and is believed to function as a “flippase” for aminophospholipids such as phosphatidylserine. FIC1 dysfunction appears to alter the characteristic canalicular membrane lipid asymmetry and to increase cholesterol extraction into the canaliculus, thereby impairing BSEP activity.24 BSEP is an ATP-dependent bile salt export pump on the canalicular membrane of hepatocytes, and its activity provides a major driving force for bile formation.25 Factors that interfere with BSEP activity or expression lead to cholestasis.16 Mutations in FIC1 and BSEP also are responsible for progressive familial intrahepatic cholestasis types 1 and 2, two morbid pediatric cholestatic disorders (see Chapter 76). Patients with benign recurrent cholestasis typically experience recurrent episodes of malaise and pruritus in association with jaundice; fever and abdominal pain are uncommon.26 The first episode of jaundice commonly occurs before the second decade of life. During periods of jaundice, laboratory abnormalities include elevations in serum alkaline phosphatase and aminotransferase levels but, as in other cholestatic disorders, the elevation in serum alkaline phosphatase characteristically predominates. When performed during an icteric episode, liver biopsy findings are generally confined to centrilobular cholestasis. Portalbased inflammatory cell infiltrates are uncommon (and, if present, are mild), and hepatocellular necrosis is not observed. Cholestatic episodes may last up to several months and are separated by periods of clinical remission. Although the patient’s quality of life may be affected adversely, the disease does not progress histologically, and liver failure does not occur. A number of drugs produce histologically bland intrahepatic cholestasis (see Chapter 86). Estrogens reduce bile formation principally by inhibiting bile salt secretion.16 Estrogens down-regulate the sinusoidal bile salt uptake protein sodium taurocholate cotransporting peptide (NTCP, SLC10A1) and competitively inhibit BSEP. Other potential mechanisms of estrogen-induced cholestasis include inhi­ bition of the hepatocellular plasma membrane sodiumpotassium pump, an important modulator of solute transport from blood to bile, and impaired acidification of intracellular organelles, with disruption of the targeting of organic anion transporters to their proper membrane domains.27 Cholestasis related to the use of oral contraceptives usually develops within two months of the initiation of

Chapter 20  Jaundice therapy. Jaundice is generally accompanied by pruritus, but fever, rash, and arthralgias are absent. Cholestasis resolves promptly with discontinuation of the drug. Anabolic steroids can produce a syndrome clinically indistinguishable from estrogen-induced cholestasis. The clinical features of cholestasis associated with total parenteral nutrition may also resemble those of estrogen and anabolic steroid-induced cholestasis, but progressive hepatic fibrosis has also been described.28 The syndrome is believed to be related, in part, to an alteration in the enterohepatic circulation and to diminished neuroendocrine stimulation of bile flow. Cholestasis and jaundice also may develop during bacterial infections, likely because of the down-regulation of the transporters NTCP and MRP2 by tumor necrosis factor-α and interleukin (IL)-1β, as well as IL-1β–dependent downregulation of BSEP.29 Sepsis-related cholestasis in the critically ill patient may be difficult to distinguish from obstruction of the bile ducts; abdominal pain and pruritus are generally absent in patients with sepsis-related chole­ stasis. However, depending on the severity of illness and response to antibiotic therapy, imaging studies may be required to exclude intrahepatic abscesses or biliary tract obstruction (see Chapter 82). Jaundice resulting from intrahepatic cholestasis has been reported as a paraneoplastic phenomenon (i.e., in the absence of malignant infiltration of the liver) in patients with lymphoma and renal cell carcinoma. The latter, referred to as Stauffer’s syndrome, classically is associated with hepatosplenomegaly.30 Cholestasis and hepatosplenomegaly resolve after nephrectomy. The pathogenesis of this disorder is uncertain but may relate to tumor-derived secretion of cytokines such as IL-6,31 which down-regulates MRP2 and BSEP and reduces NTCP activity.29 Atypical Presentations of Cholestasis Viral hepatitis rarely may cause profound cholestasis with marked pruritus.32 Unless the patient has risk factors for viral hepatitis, no features reliably distinguish this disorder from other cholestatic syndromes or biliary tract obstruction. A high level of suspicion and appropriate serologic tests will help establish the diagnosis. Alcoholic hepatitis manifesting as fever, jaundice, abdominal pain, and leukocytosis also may be difficult to distinguish from obstruction of the bile ducts. Occasionally, the increase in serum alkaline phosphatase levels is greater than the increase in aminotransferase levels.33 Liver biopsy may be required to confirm the diagnosis if there is a high clinical index of suspicion.

Jaundice in Pregnancy

Several cholestatic disorders are associated with pregnancy (see Chapter 38). Jaundice uncommonly may accompany hyperemesis gravidarum, a generally self-limited disorder of the first trimester, but liver failure is not a feature of this illness.34 Intrahepatic cholestasis of pregnancy typically occurs in the third trimester and manifests with pruritus; it occasionally is associated with jaundice. Cholestasis generally resolves within two weeks of delivery and tends to recur with subsequent pregnancies. Polymorphisms in the genes encoding the canalicular phospholipid transporters MDR3 (ABCB4), BSEP, FIC1, and MRP2s, and nuclear receptors that modulate their expression have been associated with this disorder,35-38 and these mutations may result in increased sensitivity to the inhibitory effects of estrogens on bile formation (see earlier). A far more serious syndrome is acute fatty liver of pregnancy, which typically occurs in the third trimester. The characteristic histologic features of

microvesicular steatosis in hepatocytes resembles Reye’s syndrome. Jaundice, when present, usually is accompanied by nausea, abdominal pain, and encephalopathy. The disorder may be fatal unless obstetrical delivery is performed promptly. Preeclampsia, a microvascular disorder of the third trimester, is heralded by hypertension and proteinuria and affects the liver in approximately 10% of cases. A particularly severe form, the (HELLP) syndrome (hemolysis, elevated liver enzyme levels, and a low platelet count), is treated by prompt obstetric delivery.

Jaundice in the Critically Ill Patient

The diagnosis of jaundice in the critically ill patient often presents a major challenge to intensivists and their consultants. In this setting, concerned advocates for the patient, including family and friends, may regard jaundice as the cause rather than a manifestation of the underlying problems, and the persistence of jaundice can be a source of dismay and frustration. Indeed, under ideal conditions in which the patient recovers, resolution of jaundice may lag behind disease remission by days or weeks. Therefore, the management of such individuals requires a careful search for reversible causes of jaundice and a great deal of patience. Possible predisposing factors to jaundice in critically ill patients include hepatic ischemia, blood transfusions, massive trauma, hepatotoxic pharmacologic agents, parenteral nutrition, and occult sepsis.39 Moreover, jaundice can be exacerbated by renal insufficiency, which leads to decreased excretion of conjugated bilirubin into urine (see Fig. 20-1) (see Chapter 35).

OBSTRUCTION OF THE BILE DUCTS

Obstructive disorders of the biliary tree include occlusion of the bile duct lumen, intrinsic disorders of the bile ducts, and extrinsic compression.

Choledocholithiasis

The most common cause of biliary obstruction is choledocholithiasis. Cholesterol gallstones that obstruct the bile ducts typically originate in the gallbladder, migrate into the common bile duct, and occlude the ampulla of Vater or produce partial obstruction in a ball valve fashion (see Chapter 65). In patients with unconjugated hyperbilirubinemia, calcium bilirubinate stones, so-called black pigment gallstones, form in the gallbladder and also may form in situ at any level of the biliary tree. Brown pigment gallstones, a distinct type of bilirubinate stone, also form in situ within the biliary tree. Obstruction of the bile ducts by these stones leads to repeated bouts of cholangitis (recurrent pyogenic cholangitis) in patients from certain regions of Asia and in patients with prior biliary tract surgery (see Chapter 68).

Diseases of the Bile Ducts

Intrinsic narrowing of the bile ducts occurs in inflam­ matory, infectious, or neoplastic biliary disease. Congenital disorders of the bile ducts, including cysts and biliary atresia, are discussed in Chapter 62. Primary sclerosing cholangitis, an inflammatory disorder of the bile ducts, is characterized by focal and segmental biliary strictures and is discussed extensively in Chapter 68. Jaundice is an unusual complication of a similar disorder characterized by focal narrowing and localized obstruction of the bile ducts (AIDS [acquired immunodeficiency syndrome] cholan­ giopathy) in patients with AIDS (see Chapter 33). Biliary strictures also may follow hepatic arterial infusion of certain chemotherapeutic agents40 or result from surgical

329

330

Section III  Symptoms, Signs, and Biopsychosocial Issues injury to the bile duct or hepatic artery. Neoplasms of the biliary tree, including cholangiocarcinoma, are discussed in detail in Chapter 69.

and (4) development of a strategy for treatment or further testing if unexpected diagnostic possibilities arise.

Extrinsic Compression

The patient’s history and physical examination provide important clues regarding the cause of jaundice (Table 20-3). A history of biliary surgery, fever, especially when accompanied by rigors, and abdominal pain, particularly in the right upper quadrant, is suggestive of biliary obstruction with cholangitis. Obstructive jaundice from gallstone disease or malignant neoplasms is more common in older adults than in younger persons. Symptoms compatible with a viral prodrome, such as anorexia, malaise, and myalgias, make viral hepatitis a strong diagnostic possibility, as does a history of a known infectious exposure, injection drug use, or prior transfusions of blood products. A carefully taken history may suggest that environmental hepatotoxins, ethanol, or medications underlie the patient’s cholestatic liver disease. Furthermore, a family history of jaundice or liver disease suggests the possibility of hereditary hyperbilirubinemia or genetic liver disease. All clues must be interpreted with caution, because fever and abdominal pain accompany diseases other than biliary obstruction, and viral hepatitis may occur coincidentally in patients with a history of prior biliary surgery. Conversely, anorexia and malaise are not specific for viral hepatitis, and gallstones frequently develop in patients with chronic liver disease. Nonetheless, when clues are evaluated in the context of the physical findings and routine laboratory tests, jaundice can be characterized correctly as obstructive or nonobstructive in at least 75% of cases.42

Extrinsic compression of the biliary tree may result from neoplastic involvement or inflammation of surrounding viscera. Rarely, marked enlargement of the surrounding vasculature (e.g., arterial aneurysms, cavernous transformation of the portal vein [portal cavernoma]) can compress the bile ducts (see Chapter 83). Painless jaundice is a classic feature of carcinoma of the head of the pancreas (see Chapter 60). Occasionally, hepatocellular carcinoma or periportal lymph nodes enlarged by metastatic tumor or lymphoma obstructs the extrahepatic bile ducts. Pancreatitis may also produce extrinsic biliary compression as a result of edema, pseudocyst formation, or fibrosis (see Chapters 58 and 59). Rarely, gallstones in the cystic duct or infundibulum of the gallbladder compress the common hepatic duct (Mirizzi’s syndrome) and produce jaundice.41

DIAGNOSTIC APPROACH TO JAUNDICE A general algorithm for evaluating the patient with jaundice is depicted in Figure 20-2. The sequential approach involves the following: (1) a carefully taken patient history, thorough physical examination, and screening laboratory studies; (2) formulation of a working differential diagnosis; (3) selection of specialized tests to narrow the diagnostic possibilities;

HISTORY AND PHYSICAL EXAMINATION

History, physical examination, routine laboratory tests

Alkaline phosphatase or aminotransferases elevated?

ERCP or THC

Evaluate for hemolysis, hereditary hyperbilirubinemia

Yes

Therapeutic intervention Biliary obstruction

No

Biliary tract obstruction a consideration? Dilated bile ducts

No

Biochemical studies for specific causes of liver disease

Positive

Specific therapy

Negative

Yes

Observe, consider liver biopsy

Abdominal ultrasonography or CT Nondilated bile ducts

No biliary obstruction

High

Clinical likelihood of biliary obstruction

Low

Intermediate Dilated bile ducts

Consider MRCP or EUS

Nondilated bile ducts

Figure 20-2.  Algorithm for the eva­luation and management of jaundice and hyper­bilirubinemia. CT, computed tomography; ERCP, endoscopic retrograde cholan­giopancreatography; EUS, endoscopic ultrasound; MRCP, magnetic resonance cholangiopancreatography; THC, transhepatic cholangiography.

Chapter 20  Jaundice Table 20-3  Clues to the Differential Diagnosis of Jaundice: Biliary Obstruction versus Liver Disease PARAMETER

BILIARY OBSTRUCTION

LIVER DISEASE

History

Abdominal pain Fever, rigors Prior biliary surgery Older age

Physical examination

Fever Abdominal tenderness Palpable abdominal mass Abdominal surgical scar Predominant elevation of serum alkaline phosphatase relative to aminotransferases* Prothrombin time (INR) normal or normalizes with vitamin K administration Leukocytosis Elevated serum amylase or lipase level

Anorexia, malaise, myalgias (viral prodrome) Known viral exposure History of blood product receipt or of injection drug use Exposure to known hepatotoxin Family history of liver disease Spider angiomata Stigmata of portal hypertension (e.g., prominent abdominal veins, splenomegaly, ascites) Asterixis Predominant elevation of serum aminotransferase levels relative to alkaline phosphatase Prolonged prothrombin time that does not normalize with vitamin K administration Thrombocytopenia Serologies indicative of specific liver disease

Laboratory studies

*Except early after acute obstruction when the opposite pattern may be seen transiently. INR, international normalized ratio.

The clues offered by the physical examination also are important in the patient with jaundice. High fever or abdominal tenderness (particularly in the right upper quadrant) suggests cholangitis, and a palpable abdominal mass suggests a neoplastic cause of obstructive jaundice. The rare finding of silver stools, resulting from the combination of blood and lack of bile, suggests an ampullary neoplasm. An abdominal scar in the midline or right upper quadrant may be the only clinical clue to prior biliary surgery. The presence of cirrhosis may be suggested by signs of portal hypertension, such as ascites, splenomegaly, and prominent abdominal veins, or other physical findings of liver disease, such as spider angiomata, gynecomastia, and asterixis. Certain physical findings may suggest specific liver diseases, as for hyperpigmentation in hemochromatosis, xanthomas in primary biliary cirrhosis, and Kayser-Fleischer rings in Wilson disease.

INITIAL LABORATORY STUDIES

Essential laboratory studies in the patient with jaundice include serum total bilirubin, alkaline phosphatase, ALT, and AST levels, a complete blood count, and the prothrombin time (see Chapter 73). Serum alkaline phosphatase activity reflects a number of related enzymes of overlapping substrate specificity. Alkaline phosphatase is associated predominantly with the apical domain of the plasma membrane of hepatocytes and cholangiocytes. Under physiologic conditions, this protein is cleaved enzymatically from a glycolipid anchor and released into bile, and small amounts are released from the sinusoidal (basolateral membrane) into plasma as well. Biliary obstruction and intrahepatic cholestasis increase the synthesis and basolateral release of alkaline phosphatase, and serum alkaline phosphatase activity increases. An increase in serum alkaline phosphatase activity, however, also may reflect release of alkaline phosphatase isoenzymes from extrahepatic tissues. Therefore, other more specific markers, such as the serum activities of the canalicular enzymes gamma glutamyl transpeptidase, or 5′-nucleotidase (or alternatively, alkaline phosphatase isoenzymes), are measured to confirm the hepatobiliary origin of an elevated serum alkaline phosphatase level when other liver biochemical test results (e.g., total bilirubin, ALT, AST) are normal. In a jaundiced patient, a predominant increase in (hepatic) alkaline phosphatase activity relative to levels of the serum aminotransferases suggests the possibility of biliary tract obstruction. Intra­

hepatic cholestatic disorders can produce an identical biochemical picture. The aminotransferases—ALT, a cytosolic enzyme found predominantly in hepatocytes, and AST, isozymes of which are found in both the cytosol and mitochondria of parenchymal cells of liver and several other tissues—are ordinarily detected in serum in low concentrations. Hepatocellular injury caused by ischemia, toxins, or immunemediated responses to foreign antigens such as viral proteins greatly increases serum aminotransferase activity. Predominant elevation of serum aminotransferase activity in comparison with alkaline phosphatase activity suggests that jaundice is the result of intrinsic hepatocellular disease. A serum activity of AST that is less than 10 times the upper limit of normal and that exceeds ALT activity by at least a factor of 2 is usually suggestive of alcoholic liver disease (see Chapter 84), but there are exceptions to these generalizations. For example, transient biliary obstruction from choledocho­lithiasis associated with cholangitis may cause a brief but dramatic elevation (exceeding 10 to 20 times normal) of serum aminotransferase activity.43 A complete blood count provides complementary information concerning the cause of jaundice. Leukocytosis may be a clue to the presence of biliary tract obstruction or other inflammatory disorder that may be associated with cholestasis. The presence of anemia leaves open the possibility that a hemolytic disorder is responsible for bilirubin overload. Thrombocytopenia is a characteristic finding in cirrhosis and appears to result from reduced platelet production from decreased hepatocyte synthesis of thrombopoietin or from increased platelet consumption from splenic sequestration associated with portal hypertension. The prothrombin time is a measure of the plasma activities of coagulation factors I, II, V, VII, and X, each of which is synthesized by hepatocytes. Prolongation of the prothrombin time (and an associated increase in the international normalized ratio, INR) can result from impaired hepatic synthesis of these proteins and from deficiency of vitamin K, which is required as a cofactor for essential posttranslational modification of factors II, VII, IX, and X. Efficient absorption of vitamin K by the small intestine requires an intact enterohepatic circulation of bile salts (hence, an unobstructed biliary tree). Exogenous administration of vitamin K will generally normalize a prolonged prothrombin time in patients with obstructive jaundice and intrahepatic cholestasis but not in patients with liver disease caused by hepatocellular injury.

331

332

Section III  Symptoms, Signs, and Biopsychosocial Issues OVERALL APPROACH

Integration of the patient’s history, physical examination, and laboratory study results will provide an estimate of the likelihood that obstructive jaundice is present. For example, an asymptomatic patient with hyperbilirubinemia who has an unremarkable physical examination, normal serum alkaline phosphatase and aminotransferase levels, normal platelet count, and normal prothrombin time is unlikely to have liver disease or biliary obstruction. In this patient, further testing for specific disorders, such as isolated defects in bilirubin metabolism or hemolysis, is warranted (see Fig. 20-2). Alternatively, if the history, physical examination, and laboratory study results suggest the possibility of biliary obstruction, an imaging study of the biliary tree is appropriate. Selection of the appropriate imaging study depends on the likelihood of bile duct obstruction and the diagnostic accuracy, cost, complication rate, and availa­ bility of each test (see later), especially if therapeutic intervention at the time of the study is anticipated.

IMAGING STUDIES Abdominal Ultrasonography

Abdominal ultrasonography is usually the initial imaging test in the evaluation of hepatobiliary disease because it determines the caliber of the extrahepatic biliary tree and reveals intra- or extrahepatic mass lesions. The sensitivity of abdominal ultrasonography for the detection of biliary obstruction in jaundiced patients ranges from 55% to 91%, and the specificity ranges from 82% to 95%.44-48 Ultrasonography also can demonstrate cholelithiasis, although bile duct stones may not be well seen, and intrahepatic spaceoccupying lesions more than 1 cm in diameter. Ultrasonography has the advantages of being noninvasive, portable (invaluable in the evaluation of the critically ill patient), and relatively inexpensive (Table 20-4). The major disadvantages are that the procedure is operator-dependent and interpretation may be difficult in obese patients or patients with overlying bowel gas. An additional caveat is that in patients with cirrhosis and other conditions associated with poorly compliant hepatic parenchyma, such as primary sclerosing cholangitis, intrahepatic ducts may not dilate with biliary obstruction.

Computed Tomography

Computed tomography (CT) of the abdomen with intravenous contrast is an alternative noninvasive means of evaluating the possibility of biliary tract obstruction. Abdominal CT permits accurate measurement of the caliber of the biliary tree, with sensitivity and specificity rates of 63% to 96% and 93% to 100%, respectively, for detecting biliary obstruction; these rates are comparable with those for ultrasonography.44-47 Abdominal CT detects intrahepatic space-occupying lesions as small as 5 mm, is not operatordependent, and provides technically superior images in obese persons and in those in whom the biliary tree is obscured by bowel gas. The caveats that apply to the accuracy of ultrasonography for the diagnosis of biliary obstruction also apply to abdominal CT. Abdominal CT also lacks portability, it is more expensive than ultrasonography, and the requirement for the use of intravenous contrast is a potential contraindication in the setting of kidney failure (see Table 20-4).

Magnetic Resonance Cholangiopancreatography

Magnetic resonance cholangiopancreatography (MRCP) is a technical refinement of standard magnetic resonance imaging that permits rapid clear-cut delineation of the

biliary tree without the need for intravenous contrast. MRCP appears to be superior to conventional ultrasound or CT for the detection of biliary tract obstruction and now plays a major role as a diagnostic test in this setting (see Table 20-4). Moreover, standard magnetic resonance imaging can be performed during the same examination if there is a question of a hepatobiliary or pancreatic mass or if a contrast allergy precludes CT. For detection of obstruction of the bile ducts, the sensitivity of MRCP is 82% to 100% and the specificity is 94% to 98%.49-52 Its expense is higher than that of ultrasound or CT and comparable with that of ERCP.53

Endoscopic Retrograde Cholangiopancreatography

Endoscopic retrograde cholangiopancreatography (ERCP) permits direct visualization of the biliary tree. ERCP is more invasive than ultrasonography and CT (see Table 20-4). The procedure involves passage of an endoscope into the duodenum, introduction of a catheter into the ampulla of Vater, and injection of contrast medium into the bile duct; sedation and analgesia are necessary. ERCP is highly accurate in the diagnosis of biliary obstruction, with sensitivities of 89% to 98% and specificities of 89% to 100%.47,54,55 In addition to providing radiographic images, ERCP permits biopsy and brushings for cytology of distal biliary and periampullary lesions. Moreover, if a focal cause of biliary obstruction is identified (e.g., choledocholithiasis, biliary stricture), maneuvers to relieve obstruction (e.g., sphincterotomy, stone extraction, stricture dilation, stent placement) can be performed during the same session (see Chapter 70). Acquisition of biopsy specimens and therapeutic inter­ ventions via ERCP are limited largely to lesions distal to the bifurcation of the right and left hepatic bile ducts. The technical success rate of ERCP is higher than 90%; the technique fails when the ampulla of Vater cannot be cannulated, as may be the case in patients with prior abdominal surgery and altered anatomy (e.g., gastric bypass, choledochojejunostomy). Rates of morbidity and mortality from untoward events, such as respiratory depression, aspiration, bleeding, perforation, cholangitis, and pancreatitis, are 3% and 0.2%, respectively, in patients undergoing ERCP.55 These rates are higher when interventional procedures are carried out.56

Percutaneous Transhepatic Cholangiography

Percutaneous transhepatic cholangiography (THC) is a procedure that complements ERCP. Percutaneous THC requires the passage of a needle through the skin and subcutaneous tissues into the hepatic parenchyma and advancement into a peripheral bile duct. When bile is aspirated, a catheter is introduced through the needle, and radiopaque contrast medium is injected. Sensitivity and specificity rates of percutaneous THC for the diagnosis of biliary tract obstruction are 98% to 100% and 89% to 100%, respectively, and are comparable with those for ERCP.57,58 Like ERCP, interventional procedures, such as balloon dilation and stent placement, can be performed at the time of percutaneous THC to relieve focal obstructions of the biliary tree (see Chapter 70). Percutaneous THC is potentially technically advantageous when the level of biliary obstruction is proximal to the common hepatic duct or altered anatomy precludes ERCP (see earlier). Percutaneous THC may be technically challenging in the absence of dilatation of the intrahepatic bile ducts; in this situation, multiple passes may be required, and visualization of the biliary tree may be unsuccessful in up to 10% of attempts.59 Rates of morbidity and mortality as a result of bleeding, perforation, and cholangitis are 3% and 0.2%, respectively, in patients undergoing percutaneous THC. Percutaneous THC is more

Chapter 20  Jaundice Table 20-4  Imaging Studies for the Evaluation of Jaundice TEST

SENSITIVITY (%)

SPECIFICITY (%)

MORBIDITY (%)

MORTALITY (%)

Abdominal US

55-91

82-95

0

0

Abdominal CT

63-96

93-100

See disadvantages

0

MRCP

82-100

94-98

See disadvantages

ERCP

89-98

89-100

3

0.2

Percutaneous THC

98-100

89-100

3.5

0.2

EUS

89-97

67-98

See disadvantages

0

ADVANTAGES AND DISADVANTAGES Advantages—noninvasive, portable Disadvantages—bowel gas may obscure bile duct; difficult in obese persons, operator-dependent Advantages—noninvasive, higher resolution than ultrasound, not operator-dependent Disadvantages—potential for contrast-induced nephrotoxicity, anaphylaxis Advantages—noninvasive, imaging of bile ducts superior to ultrasound and CT Disadvantages—requires breath holding, may miss small-caliber bile duct disease Advantages—provides direct imaging of bile ducts; permits direct visualization of periampullary region and acquisition of tissue distal to bifurcation of hepatic ducts; permits simultaneous therapeutic intervention, especially useful for lesions distal to bifurcation of hepatic ducts Disadvantages—requires sedation, cannot be performed if altered anatomy precludes endoscopic access to ampulla (e.g., Roux-en-Y loop); has complications (e.g., pancreatitis) Advantages—provides direct imaging of bile ducts, permits simultaneous therapeutic intervention, especially useful for lesions proximal to common hepatic duct Disadvantages—more difficult with nondilated intrahepatic bile ducts; has complications Advantages—imaging of bile ducts superior to ultrasound and CT, permits needle aspiration of suspected neoplasms Disadvantages—requires sedation

CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasonography; MRCP, magnetic resonance cholangiopancreatography; THC, transhepatic cholangiography; US, ultrasonography.

expensive than abdominal ultrasonography and CT (see Table 20-4).

Endoscopic Ultrasonography

Endoscopic ultrasonography (EUS) also can detect obstruction of the bile duct and major intrahepatic bile ducts, with a sensitivity and specificity comparable with those of MRCP.49,60,61 EUS has the potential advantage of permitting biopsy of suspected malignant lesions, and under appropriate circumstances, the operator can proceed directly to ERCP for definitive biliary decompression (see Table 20-4). The risk of diagnostic EUS is comparable with that of diagnostic upper endoscopy; when needle biopsy is used, the mortality rate is approximately 0.1%.62 EUS may be most useful in circumstances in which the patient is thought to be at high risk for complications of ERCP or percutaneous THC.

Nuclear Imaging Studies Nuclear scintigraphy of the biliary tree, although helpful in the diagnosis of cholecystitis, is not sufficiently sensitive to justify its routine use in the diagnostic evaluation of jaundice.45,46 Furthermore, hepatic uptake of radiolabeled derivatives of iminodiacetic acid (e.g., HIDA) is limited when the serum bilirubin level exceeds 7 to 10 mg/ dL.63 One exception to this generalization is in the evaluation of a potential bile leak, an uncommon cause of jaundice following biliary surgery, in which scintigraphy has an accuracy rate as high as 87%.64

Suggested Strategies for Imaging

The order of imaging studies depends largely on the clinical likelihood of obstructive jaundice (see Fig. 20-2). Several diagnostic strategies have been compared by clinical deci-

333

334

Section III  Symptoms, Signs, and Biopsychosocial Issues

A

B

Figure 20-3.  Liver histology in biliary tract obstruction. A, Prominent bile duct proliferation (arrows) and a mixed portal-based inflammatory infiltrate are evident. Periportal hepatocytes show feathery degeneration (arrowheads), indicative of cholate stasis, cytological changes caused by prolonged cholestasis (Hematoxylin and eosin, ×200). B, The periportal bilirubin-stained region (arrow) surrounded by necrotic cells represents a bile infarct (Hematoxylin and eosin, ×40).

sion analysis.65 If the probability of biliary obstruction is approximately 20%, the positive and negative predictive values of a strategy that uses ultrasonography as the initial test is estimated to be 96% and 98%, respectively. This strategy compares favorably with one that uses ERCP as the initial test. Alternatively, if the probability of biliary obstruction is 60%, a strategy that uses ultrasonography as the first test would yield a positive predictive value of 99%, whereas the negative predictive value would fall to 89%. The implication is that if the level of suspicion for biliary tract obstruction is high and an ultrasound does not show dilated bile ducts, further studies to visualize the biliary tree should be pursued. Therefore, in jaundiced patients in whom biliary obstruction is a possibility, abdominal ultrasonography (or CT) is an appropriate initial approach. If the bile ducts are dilated, the biliary tree should be imaged directly with ERCP (or percutaneous THC) and appropriate therapy undertaken if biliary obstruction is found. If the bile ducts are not dilated on abdominal ultrasonography (or CT), the next step depends on the clinical likelihood of biliary obstruction. If the likelihood of biliary obstruction is thought to be low, the patient should be evaluated for intrinsic liver disease (see later). If the likelihood of biliary obstruction is believed to be intermediate, MRCP or EUS is a reasonable option for imaging the biliary tree before an evaluation for a hepatic disorder is undertaken. Among patients in whom biliary obstruction is believed to be likely, ERCP (or percutaneous THC) should be considered as the next step. If ERCP or percutaneous THC does not show biliary obstruction, the patient should be evaluated for cholestatic liver disease. The decision to use ERCP versus percutaneous THC will be influenced by various factors (see Table 20-4), including the availability of each procedure at a particular facility, presence or absence of dilated bile ducts on initial imaging, and suspected level of biliary obstruction. Under most circumstances, ERCP should be the procedure of choice, because it is comparable with percutaneous THC in accuracy, technical success rate, and frequency of major complications; tends to be more widely available; and may offer better postprocedure tolerability (e.g., no need for an external biliary drainage tube).

OTHER STUDIES Serologic Testing

When imaging studies do not suggest biliary obstruction, jaundiced patients with biochemical evidence of hepatocel-

lular dysfunction or cholestasis should be evaluated for underlying liver disease. Depending on the disorder suspected, screening laboratory studies may include viral serologies (including those for hepatitis B and C and, if the disease is acute, hepatitis A), serum levels of iron, transferrin, and ferritin (for hemochromatosis), ceruloplasmin (for Wilson disease), antimitochondrial antibodies (for primary biliary cirrhosis), antinuclear antibodies, smooth muscle antibodies, and serum protein electrophoresis or serum immunoglobulins (for autoimmune hepatitis), and tissue transglutaminase antibodies (for celiac disease). Confirmation of these diagnoses, as well as elucidation of diagnoses not revealed by serologic analysis, may be made by liver biopsy—and small bowel biopsy in the case of celiac disease.

Liver Biopsy

Liver biopsy provides precise information regarding hepatic lobular architecture and extent and pattern of fibrosis, and is most helpful for patients with persistent and undiagnosed jaundice. With special histologic stains and, if appropriate, quantification of copper or iron content, liver biopsy permits the diagnosis of viral hepatitis, fatty liver disease, hemochromatosis, Wilson disease, primary biliary cirrhosis, granulomatous hepatitis, and neoplasms. Occasionally, liver biopsy specimens provide clues to otherwise unsuspected biliary tract obstruction, the histologic features of which are shown in Figure 20-3; however, liver histology may be entirely normal in acute biliary obstruction. Liver biopsy is associated with a low but definite complication rate, predominantly from bleeding and perforation, and the need for hospitalization in 1% of cases; the mortality rate is approximately 0.01% (see Chapter 73).66

THERAPEUTIC APPROACHES BILIARY OBSTRUCTION

In the patient with obstruction of the bile ducts, therapy is typically directed at relieving the obstruction. Interventional endoscopic or radiologic approaches include sphincterotomy, balloon dilation of focal strictures, and placement of drains or stents; the alternatives are surgical (see Chapter 70). The therapeutic strategy chosen will depend, in part, on the location and likely cause of the obstructing lesion. Focal intrahepatic strictures may be amenable to an interventional radiologic approach, whereas lesions distal to the bifurcation of the hepatic ducts may be more suitably

Chapter 20  Jaundice managed endoscopically (e.g., sphincterotomy for choledocholithiasis); mass lesions may require surgery.

OTHER CONDITIONS

When jaundice is caused by liver disease, the optimal treatment is directed toward the underlying cause (e.g., cessation of ethanol, discontinuation of the offending drug, administration of antiviral agents, phlebotomy for hemochromatosis, immunosuppressive agents for autoimmune hepatitis). Therapy for hyperbilirubinemia per se is generally not necessary in adults, because the neurotoxicity of bilirubin is confined to disorders characterized by extreme elevations of unconjugated bilirubin in neonates and infants, such as physiologic jaundice of the newborn or type I CriglerNajjar syndrome. In these special cases, the risk of neurotoxicity can be reduced with phototherapy, in which exposure to blue or green light produces photoisomerization of bilirubin to more water-soluble enantiomers that do not require conjugation for excretion in bile.67,68 Preliminary observations have suggested that orlistat, which increases intestinal fat excretion, may trap unconjugated bilirubin intraluminally and may augment phototherapyor phenobarbital-induced reduction of unconjugated hyperbiliru­binemia in children with type I or type II CriglerNajjar syndrome, respectively.69 The choleretic bile acid ursodeoxycholic acid (ursodiol) has been studied as a treatment for several cholestatic disorders.70 Ursodeoxycholic acid improves biochemical indices and has been suggested to slow disease progression in primary biliary cirrhosis (see Chapter 89). Ursodeoxycholic acid has been shown to improve biochemical markers and clinical outcomes in patients with intrahepatic cholestasis of pregnancy,71,72 and pilot studies have suggested that it may be helpful in improving biochemical indices of cholestasis related to parenteral nutrition73 and cystic fibrosis74 and in preventing cholestasis following hematopoietic cell transplantation.75 By contrast, although initial pilot studies suggested that ursodeoxycholic acid reversed cholestasis in primary sclerosing cholangitis, a long-term benefit of this agent in this disorder has not been demonstrated to date in randomized controlled trials (see Chapter 68). In addition to the specific treatments outlined, cholestatic dis-

orders may lead to impaired absorption of fat-soluble vitamins (A, D, E, and K), and supplementation is recommended. The management of pruritus caused by cholestasis is discussed in Chapter 89.

KEY REFERENCES

Borst P, de Wolf C, van de Wetering K. Multidrug resistance-associated proteins 3, 4, and 5. Pflugers Arch 2007; 453:661-73 (Ref 6). Bosma PJ. Inherited disorders of bilirubin metabolism. J Hepatol 2003; 38:107-17 (Ref 1). Brienza N, Dalfino L, Cinnella G, et al. Jaundice in critical illness: Promoting factors of a concealed reality. Intensive Care Med 2006; 32:267-74 (Ref 39). Dennery PA, Seidman DS, Stevenson DK. Neonatal hyperbilirubinemia. N Engl J Med 2001; 344:581-90 (Ref 67). Fevery J. Bilirubin in clinical practice: A review. Liver Int 2008; 28:592605 (Ref 2). Geier A, Wagner M, Dietrich CG, Trauner M. Principles of hepatic organic anion transporter regulation during cholestasis, inflammation and liver regeneration. Biochim Biophys Acta 2007; 1773:283-308 (Ref 29). Kullak-Ublick GA, Stieger B, Meier PJ. Enterohepatic bile salt transporters in normal physiology and liver disease. Gastroenterology 2004; 126:322-42 (Ref 16). Labori KJ, Bjornbeth BA, Raeder MG. Aetiology and prognostic implication of severe jaundice in surgical trauma patients. Scand J Gastroenterol 2003; 38:102-8 (Ref 8). Luketic VA, Shiffman ML. Benign recurrent intrahepatic cholestasis. Clin Liver Dis 2004; 8:133-49 (Ref 26). Malchow-Moller A, Gronvall S, Hilden J, et al. Ultrasound examination in jaundiced patients. Is computer-assisted preclassification helpful? J Hepatol 1991; 12:321-6 (Ref 42). Nies AT, Keppler D. The apical conjugate efflux pump ABCC2 (MRP2). Pflugers Arch 2007; 453:643-59 (Ref 5). Paulusma CC, Groen A, Kunne C, et al. Atp8b1 deficiency in mice reduces resistance of the canalicular membrane to hydrophobic bile salts and impairs bile salt transport. Hepatology 2006; 44:195-204 (Ref 24). Paumgartner G, Beuers U. Ursodeoxycholic acid in cholestatic liver disease: Mechanisms of action and therapeutic use revisited. Hepatology 2002; 36:525-31 (Ref 70). Richter JM, Silverstein MD, Schapiro R. Suspected obstructive jaundice: A decision analysis of diagnostic strategies. Ann Intern Med 1983; 99:46-51 (Ref 65). Stieger B, Meier Y, Meier PJ. The bile salt export pump. Pflugers Arch 2007; 453:611-20 (Ref 25). Full references for this chapter can be found on www.expertconsult.com.

335

CHAPTE R

21 Biopsychosocial Issues in Gastroenterology Douglas A. Drossman

CHAPTER OUTLINE Conceptualization of Gastrointestinal Illness  337 Biomedical Model  337 Biopsychosocial Model  338 Early Life  339 Early Learning  339 Culture and Family  340 Psychosocial Environment  340 Life Stress and Abuse  341 Psychological Factors  341 Coping and Social Support  342 Brain-Gut Axis  342 Stress and Gastrointestinal Function  342 Role of Neurotransmitters  343 Regulation of Visceral Pain  343

This chapter presents an integrated overview of how psy­ chosocial factors relate to gastrointestinal function, disease susceptibility, clinical illness, and outcomes, and it pro­ vides an integrated approach to the care of the patient with gastrointestinal illness.

CONCEPTUALIZATION OF GASTROINTESTINAL ILLNESS BIOMEDICAL MODEL

In the practice of medicine, feeling confused, even stuck, when discrepancies exist between what we observe and what we expect is not uncommon. This experience may occur when diagnosing and caring for a patient who has symptoms that do not match our understanding of the degree of disease. In Western civilization, the traditional understanding of illness (the personal experience of ill health or bodily dysfunction, as determined by current or previous disease as well as psychosocial, family, and cul­ tural influences) and disease (abnormalities in structure and function of organs and tissues)1 has been termed the biomedical model.2 This model adheres to two premises. The first is that any illness can be linearly reduced to a single cause (reductionism). Therefore, identifying and modifying the underlying cause is necessary and sufficient to explain the illness and ultimately lead to cure. The second is that an illness can be dichotomized to a disease, or organic dis­ order, which has objectively defined pathophysiology, or a functional disorder, which has no specifically identifiable pathophysiology (dualism). This dichotomy presumes to

Effects of Stress on Immune Function and Disease Susceptibility  345 Cytokines and the Brain  346 Symptom Experience and Behavior  346 Outcome  346 Clinical Applications  346 History  347 Evaluating the Data  347 Diagnostic Decision Making  347 Treatment Approach  348 Establishing a Therapeutic Relationship  348 Psychopharmacologic Treatment  348 Psychotherapy and Behavioral Treatments  349 Physician-Related Issues  349

distinguish medical (organic) from psychological (func­ tional) illness or relegates functional illness to a condition with no cause or treatment. The limitations of this model are now becoming evident; modern research has shown a blurring of this dichotomy,3,4 as illustrated by the following case history.

Case 1 Ms. L, a 42-year-old woman, presents to her new physician with a 20-year history of mid to lower abdominal pain with nausea and occasional vomiting. She states, “I can’t live anymore with this pain.” Her bowel function is normal, and her weight is stable. She is unable to work, believes that the symptoms have taken over her life, and perceives no sense of control over her symptoms or any ability to decrease them. She has a long-standing history of major losses, and depression, and a history of sexual and physical abuse. She requests narcotics for pain relief because “that is the only treatment that works.” She also requests that the physician expeditiously “find the cause of the pain and remove it.” The record shows frequent emergency room visits and several hospital admissions, during which extensive diagnostic studies (including upper gastrointestinal [GI] series with small bowel follow-through, upper endoscopy, colonoscopy, capsule study, computed tomography [CT] of the abdomen and pelvis, pelvic ultrasound, and laparoscopy) were negative, and increasing doses of narcotics were given for pain relief. A prior cholecystectomy did not document gallstones, and a hysterectomy was done years earlier for endometriosis. On this occasion, an upper endoscopy shows only a positive rapid urease test for Helicobacter pylori, but a two-week

337

338

Section III  Symptoms, Signs, and Biopsychosocial Issues course of antibiotics and a proton pump inhibitor for H. pylori infection is without benefit. The patient is then referred to a psychiatrist, who makes a diagnosis of major depression but cautions that other possible medical disease should be excluded. On the next clinical visit, the patient requests narcotic pain medication for relief, but the physician suggests instead ongoing psychological care. In response, the patient requests referral to a medical center.

This case of a patient with a severe functional GI dis­ order5,6 can be challenging when approached from the biomedical model. In addition to difficulties in diagnosis and management, strong feelings may occur that are mal­ adaptive to the physician-patient relationship, for several reasons.7 First, the physician and patient approach the problem using a functional-organic dichotomy. With no evidence of a structural (organic) diagnosis to explain the symptoms for over 20 years, the patient still urges that further diagnostic studies be done to “find and fix” the problem, and the physician orders an upper endoscopy. However, failure to find a specific structural cause for medical symptoms is the rule rather than the exception in ambulatory care. In a study involving 1000 ambulatory internal medicine patients,8 only 16% of 567 new com­ plaints (and only 11% for abdominal pain) over a three-year period were eventually found to have an organic cause, and only an additional 10% were given a psychiatric diagnosis. This patient has functional abdominal pain syndrome,9,10 one of 27 adult functional GI disorders11 that comprise over 40% of a gastroenterologist’s practice (see Chapter 11).12 Mutual acceptance of this entity as a real diagnosis is the key to beginning a proper plan of care. Because functional GI disorders do not fit into a biomedical construct (i.e., they are seen as an illness without evident disease),4 the risk that unneeded and costly diagnostic tests will be ordered to find the cause continues, as illustrated by the physician’s ordering another upper endoscopy. This approach may deflect attention away from the direction of proper management. Second, psychosocial features are evident, including major loss, depression, abuse history, and maladaptive thinking (i.e., catastrophizing and perceived inability to manage the symptoms), in addition to the possible develop­ ment of narcotic bowel syndrome (see Chapter 11), which

adversely influence the clinical outcome and are amenable to proper treatment.13-15 These features are ignored or mini­ mized, however, reflecting that patients and physicians tend to view psychosocial factors as separate from, and often less important than, medical illness.16 In reality, the psychoso­ cial features are so relevant to the illness presentation that by addressing them the patient may improve. Ultimately, the physician, possibly recognizing that these issues are important, may feel unable to manage the problems and refers Ms. L to a psychiatrist. In turn, the psychiatrist also approaches the problem dualistically, noting the depression but also indicating uncertainty and even concern as to whether a medical diagnosis has been overlooked. These competing viewpoints may only confuse the patient. Third, difficulties exist in the physician-patient interac­ tion. The patient’s and physician’s goals and expectations for care are at odds. Whereas the patient wants a quick fix, the physician sees the condition as chronic and ultimately requiring psychiatric intervention, not narcotics, which could do harm.15 In response, the patient requests referral to another facility. This maladaptive interaction and poor communication relating to differing understandings of the illness and its treatment could have been avoided by addressing these differing views and mutually negotiating a plan of diagnosis and care. This “vicious cycle” of ineffective care (Fig. 21-1) results from the limitations imposed by the biomedical model. The vicious cycle occurs not only among patients with func­ tional GI disorders, but also among patients with organic disorders such as inflammatory bowel disease (IBD). In such cases, pain and diarrhea are not explained by the degree of disease activity, and the patient likely has irritable bowel syndrome (IBS) as well (so-called IBD-IBS; see Chapters 111, 112, and 118).3 The reality is that (1) medical disorders and patient symptoms are inadequately explained by struc­ tural abnormalities; (2) psychosocial factors predispose to the onset and perpetuation of illness and disease, are part of the illness experience, and strongly influence the clinical outcome regardless of diagnosis; and (3) successful applica­ tion of this understanding and proper management require an effective physician-patient relationship.

BIOPSYCHOSOCIAL MODEL

The biopsychosocial model2,16 proposes that illness and disease result not from a single cause, but from simultane­ ously interacting systems at the cellular, tissue, organism,

Anxiety Hypervigilance Figure 21-1.  Patient-physician vicious cycle. The vicious cycle relates to three components: (1) functional-organic dichotomy in which the diagnosis of a functional gastrointestinal disorder is not recognized, and the effort is directed toward further tests to identify an organic disease; (2) limited ability to identify and address underlying psychosocial factors that contribute to the illness; and (3) an impaired patient-physician relationship, with a lack of shared decision making about diagnosis and treatment. As indicated in this diagram, the risk of the vicious cycle is for increased testing, high health care costs, many referrals, and mutual dissatisfaction in care until the cycle is broken. (Adapted from Longstreth GF, Drossman DA. Severe irritable bowel and functional abdominal pain syndromes: Managing the patient and health care costs. Clin Gastroenterol Hepatol 2005; 3:397-400.)

Patient experience Pain Disability Expects cure

Urgent visits, requests tests

Physician “Organic” cause focus “Drained” Biased attitudes

Patient cognitions Dissatisfaction Pessimism Maladaptive coping Lack of control

Referrals, tests, drugs, surgery, iatrogenic harm, high costs

Chapter 21  Biopsychosocial Issues in Gastroenterology interpersonal, and environmental levels. Furthermore, psy­ chosocial factors have direct physiologic and pathologic consequences, and vice versa. For example, change at the subcellular level (e.g., human immunodeficiency virus infection or susceptibility to IBD) has the potential to affect organ function, the person, the family, and society. Simi­ larly, a change at the interpersonal level, such as the death of a spouse, can affect psychological status, cellular immu­ nity, and, ultimately, disease susceptibility.17 The model also explains why the clinical expression of biological substrates (e.g., alterations in oncogenes) and associated responses to treatment vary among patients. The biopsycho­ social model is consistent with emerging scientific data about the mechanisms of disease and clinical care and is assumed to be valid in this discussion. Figure 21-2 provides the framework for understanding the mutually interacting relationship of psychosocial and biological factors in the clinical expression of illness and disease. Early life factors (e.g., genetic predisposition, early learning, cultural milieu) can influence an individual’s later psychosocial environment, physiologic functioning, and disease (pathologic) expression, as well as reciprocal inter­ actions via the brain-gut (central nervous system [CNS]– enteric nervous system [ENS]) axis. The product of this brain-gut interaction will affect symptom experience and behavior, and ultimately the clinical outcome. Figure 21-2 will serve as a template for the outline and discussion that follows.

Early life Genetics Environment Culture

Psychosocial environment Life stress Psychological factors Coping Social support Brain

Gut

Physiology Pathology FGID Illness Disease Symptoms Behavior

Outcome Medications Health care visits Daily function Quality of life Figure 21-2.  Systems (biopsychosocial) model of gastrointestinal illness. The relationship between early life, psychosocial factors, physiology and pathology, symptom experience and behavior, and outcome are presented as interacting systems. This figure is the template for the discussion in this chapter. FGID, functional gastrointestinal disease.

EARLY LIFE EARLY LEARNING

Early learning in regard to gastrointestinal issues involves developmental aspects and physiologic conditioning.

Developmental Aspects

At or perhaps even before birth, a person’s genetic composi­ tion and interactions with the environment begin to affect later behaviors and susceptibility to illness. For example, the earliest interactions involve feeding and elimination. According to psychoanalytic theory, situations of conflict arise early in which the child’s innate impulses (e.g., to eat or defecate) confront external environmental (i.e., parental) constraints; normal personality development involves suc­ cessful resolution of these conflicts. The complex behaviors of feeding and elimination—sources of intense gratification to the infant—must gradually be controlled by the growing child according to the prevailing mores of family and society. To varying degrees during development, these adopted constraints remain in conflict with desires for immediate gratification. With increased motor control of these functions, the child can defy or comply with environ­ mental constraints by choosing to eat, resist eating, bite, defecate, or withhold stool. When and how these behaviors are displayed will depend on the child’s needs and the quality and intensity of the environmental influences that affect them. Behaviors learned during this period are con­ sidered pivotal in the child’s personality, development, and later interaction with the environment, including the devel­ opment of autonomy, of distinguishing right from wrong, and of disciplining impulses in a socially acceptable manner. Conversely, failure to resolve these early conflicts may make the adult vulnerable in situations that tax these character traits. Thus, the obstinate (as in “obstipation”) person who withholds or resists when feeling controlled may have been influenced by unresolved interactions around the control of elimination. Certain GI disorders may be influenced by learning diffi­ culties or emotionally conflicting or challenging interac­ tions that occur early in life. These disorders include rumination syndrome (see Chapter 14),18 anorexia nervosa (see Chapter 8),19 functional (psychogenic) vomiting (see Chapter 14),20 and constipation (see Chapter 18), all of which can develop based on early conditioning experiences. Dis­ orders of anorectal function (e.g., pelvic floor dyssynergia and encopresis) also may have resulted from learning dif­ ficulties relating to bowel habit21 or abuse (see Chapters 17 and 18).22 Encopretic children may withhold stool out of fear of the toilet, to struggle for control, or to receive attention from parents.23 One study has proposed that encop­ retic children are more likely than controls to have been toilet-trained by coercive techniques, have psychological difficulties, and have poor rapport with their mothers.24 Well-designed studies have supported the role of early modeling of symptom experience and behavior in the clini­ cal expression of GI symptoms and disorders.14 In particu­ lar, childhood sexual and physical abuse can have physical consequences, thereby affecting the development or severity of functional GI disorders,25 and early family attention toward GI symptoms and other illnesses can influence later symptom reporting, health behaviors, and health care costs.26

Physiologic Conditioning

Early conditioning experiences may also influence physio­ logic functioning and possibly the development of psycho­

339

340

Section III  Symptoms, Signs, and Biopsychosocial Issues physiologic disorders. Psychophysiologic reactions involve psychologically induced alterations in the function of target organs without structural change. They often are viewed as physiologic concomitants of an affect such as anger or fear, although the person is not always aware of the affect. The persistence of an altered physiologic state or the enhanced physiologic response to psychologic stimuli is considered by some as a psychophysiologic disorder. Visceral functions such as the secretion of digestive juices and motility of the gallbladder, stomach, and intestine can be classically con­ ditioned,27 even by family interaction. Classic conditioning, as described by Pavlov, involves linking a neutral food or unconditioned stimulus (sound of a bell) with a conditioned stimulus (food) that elicits a conditioned response (saliva­ tion). After several trials, the unconditioned stimulus is able to produce the conditioned response. By contrast, operant conditioning involves the development of a desired response through motivation and reinforcement. Playing basketball is an example; accuracy improves through practice, and the correct behavior is reinforced by the reward of scoring a basket. Consider the following case.

Case 2 A young child, Johnny, wakes up on the day of a school examination with anxiety and “flight-fight” symptoms of tachycardia, diaphoresis, abdominal cramps, and diarrhea. The parent keeps the child home because of a “tummy-ache” and allows him to stay in bed and watch television. Several days later, when the child is encouraged to go back to school, the symptoms recur.

In this case, the parent focused on the abdominal discom­ fort as an illness that required absence from school rather than as a physiologic response to a distressing situation. The child then avoided the feared situation. Repetition of the feared situation not only may lead to a conditionally enhanced psychophysiologic symptom response, but also may alter the child’s perception of the symptoms as illness, thereby leading to health care–seeking behaviors later in life (illness modeling).28 For example, in two studies,27,29 patients with IBS recalled more parental attention toward their ill­ nesses than those with IBS who did not seek health care; they stayed home from school and saw physicians more often and received more gifts and privileges. Somatic responses to stressful situations may be reduced when the parent openly solicits and responds to the thoughts and feelings of the child, thus making these thoughts and feel­ ings acceptable.

CULTURE AND FAMILY

Social and cultural belief systems modify how a patient experiences illness and interacts with the health care system.30 The meaning that an individual attributes to symptoms may be interpreted differently, even within the same ethnic group. Qualitative ethnographic studies con­ ducted in New York City in the mid–twentieth century among white immigrants highlighted important cultural dif­ ferences in pain behaviors.31,32 In these studies, first- and second-generation Jews and Italians were observed to embellish the description of pain by reporting more symp­ toms in more bodily locations and with more dysfunction and greater emotional expression. As noted by one Italian patient, “When I have a headache, I feel it’s very bad and it makes me irritable, tense, and short-tempered.” By contrast, the Irish might minimize the description of the pain: “It was a throbbing more than a pain,” and the “Old Americans”

(Protestants) were stoic. These behaviors related to family attitudes and mores surrounding illness either reinforce or extinguish attention-drawing symptom reporting. As a further elaboration, whereas Italians were satisfied to hear that the pain was not a serious problem, the Jewish patients needed to understand the meaning of the pain and its future consequences, with the latter possibly relating to cultural influences on the importance of the acquisition of knowl­ edge within the culture. Geographic differences are also noted with regard to worries and concerns about having IBD—for example, Southern European patients, such as Italian and Portuguese patients, report more and greater degrees of concern than their northern European counterparts.33 These cultural influences can shape health seeking and the respective roles of physicians and patients. From a global standpoint, 70% to 90% of all self-recognized ill­ nesses are managed outside traditional medical facilities, often with self-help groups or religious cult practitioners providing a substantial portion of the care.34 Rural cultural groups, including Mexicans living on the American border, more often will go to a community healer (e.g., a curandero) first despite access to a standard medical facility.35 When given the option, the Romani (gypsies) will select only the top physicians (ganzos) to take care of a family member.36 Among whites, one third see practitioners of unconven­ tional treatments (e.g., homeopathy, high colonic enemas, crystal healing) at a frequency that exceeds the number of primary care visits, and most do not inform their physicians of these treatments.37 Conversely, physicians may judge the appropriateness of patient behaviors on the basis of their own cultural biases and make efforts to show them the right way without first understanding the patient’s illness schema. Considering the hot-cold theory of illness practiced by some Puerto Ricans, if a clinician prescribes a hot medicine (not related to temperature) for a hot illness, the patient might not take that medicine. From a diagnostic standpoint, health care providers in the rural south need to be familiar with root working, a form of voodoo magic practiced by some rural African Americans.38 Acknowledging and addressing the patient’s beliefs can be therapeutic for the patient. Psychosocial factors that relate to illness can be culturally determined and affect clinical management. In China, com­ municating psychological distress is stigmatizing,34 so when a person is in distress, reporting physical symptoms (soma­ tization) is more acceptable,39 whereas in Southern Europe, emotional expression is not only assumed but also is a reinforcer of family support.32 In some nonliterate societies, individuals freely describe hallucinations that are fully accepted by others in the community.31 In fact, the meaning of the hallucinations, not their presence, is the focus of interest, particularly when reported by those in a position of power. Conversely, in Western societies, in which the emphasis is on rationality and control, hallucinations are viewed as stigmatizing, a manifestation of psychosis until proved otherwise. These factors and influences indicate the importance of inquiring about the patient’s understanding of the onset, beliefs about cause, clinical course, and desired or expected treatment of an illness. They affect adherence to treatment, at least, and possibly the response to treatment.

PSYCHOSOCIAL ENVIRONMENT As a child moves into adulthood, genetics, culture, early learning, and other environmental influences are integrated into the individual’s unique personality and behavioral

Chapter 21  Biopsychosocial Issues in Gastroenterology style. Life stress, his or her current psychological state, including the presence of a psychiatric diagnosis, and coping style and degree of social support will, in combina­ tion, determine the physiologic functioning of the intestine, susceptibility to and activity of disease, illness perceptions and behaviors, and clinical outcome.

LIFE STRESS AND ABUSE

Unresolved life stress, such as the loss of a parent, an abor­ tion, a major personal catastrophic event or its anniversary, or daily life stresses (including having a chronic illness), may influence an individual’s illness in several ways, including the following: (1) producing psychophysiologic effects (e.g., changes in motility, blood flow, body fluid secretion, or bodily sensations, thereby exacerbating symp­ toms); (2) increasing one’s vigilance toward symptoms; and (3) leading to maladaptive coping and greater illness behaviors and health care seeking.13,40 Despite numerous methodologic limitations in studying the relationship of such psychosocial factors to illness, disease, and their out­ comes, such factors clearly can exacerbate functional GI disorders14 and symptoms of certain structural disorders, such as IBD.41 Although the scientific evidence that such factors are causative in the development of pathologic diseases is compelling, based on retrospective studies and psychoimmunologic mechanisms, this conclusion is not fully established. Nevertheless, the negative impact of stressful life events on a person’s psychological state and illness behaviors requires the physician to address them in the daily care of all patients. A history of physical or sexual abuse strongly influences the severity of the symptoms and clinical outcome.42 When compared with patients without a history of abuse, patients with a history of abuse who are seen in a referral gastroen­ terology practice reported 70% more severe pain (P < 0.0001) and 40% greater psychological distress (P < 0.0001), spent over 2.5 times more days in bed in the previous three months (11.9 vs. 4.5 days, P < 0.0007), had almost twice as poor daily function (P < 0.0001), saw physicians more often (8.7 vs. 6.7 visits over six months, P < 0.03), and even underwent more surgical procedures (4.9 vs. 3.8 proce­ dures, P < 0.04) unrelated to the GI diagnosis.43 Life stress and abuse history have physiologic and behavioral effects that amplify the severity of the condition. Several possible mechanisms help explain the relation­ ship between a history of abuse and poor outcome.25 These mechanisms include the following: (1) susceptibility to developing psychological conditions that increase the perception of visceral signals or its noxiousness (central hypervigilance and somatization); (2) development of psy­ chophysiologic (e.g., autonomic, humoral, immunologic) responses that alter intestinal motor or sensory function or promote inflammation; (3) development of peripheral or central sensitization from increased motility or physical trauma (visceral hyperalgesia or allodynia); (4) an abnormal appraisal of and behavioral response to physical sensations of perceived threat (response bias); and (5) development of maladaptive coping styles that lead to increased illness behavior and health care seeking (e.g., catastrophizing). Physiologically, in patients with IBS and a history of abuse, rectal distention produces more pain reporting with greater activation of the dorsal anterior cingulate cortex44 compared with patients with IBS and no history of abuse; the pain and activation of the brain subside after treatment (see later).45

PSYCHOLOGICAL FACTORS

As shown in Figure 21-2, along with life stress and abuse, a mix of concurrent psychosocial factors can influence GI physiology and susceptibility to developing a pathologic

condition and its symptomatic and behavioral expression, all of which affect the outcome. The psychological factors relate to long-standing (also called trait) features (e.g., per­ sonality and psychiatric diagnosis) and more modifiable state features (e.g., psychological distress and mood). The latter features are amenable to psychological and psycho­ pharmacologic interventions. In addition, coping style and social support provide modulating (buffering) effects.

Personality

Are there specific personalities associated with GI disor­ ders? During the psychoanalytically dominated era of psy­ chosomatic medicine (1920 to 1955), certain psychological conflicts were believed to underlie development of person­ alities that expressed specific psychosomatic diseases (e.g., asthma, ulcerative colitis, essential hypertension, duodenal ulcer).46 In the biologically predisposed host, disease would develop when environmental stress was sufficient to acti­ vate the psychological conflict. The idea that personality features specifically relate to causation of medical disease (albeit in a biologically predisposed host), however, is too simplistic. Currently, investigators view personality and other psychological traits as enablers or modulators of illness, along with other contributing factors such as life stress, social environment, and coping.

Psychiatric Diagnosis

The co-occurrence of a psychiatric diagnosis in patients with a medical disorder (comorbidity) is common, and the psychiatric diagnosis aggravates the clinical presentation and outcome of the medical disorder. The most common psychiatric diagnoses seen among patients with chronic GI disorders are depression (including dysthymia) and anxiety (including panic attacks), and these psychiatric disorders are often amenable to psychopharmacotherapeutic or psy­ chological treatment.14 When psychiatric disorders and personality traits adversely affect an individual’s experience and behavior to the point of interfering with interactions involving family, social peers, and physicians, these disorders and traits must be attended to. These conditions can include the following: somatization disorder, characterized by a fixed pattern of experiencing and reporting numerous physical complaints beginning early in life; factitious disorder, or possibly Munchausen’s syndrome, in which a patient surreptitiously simulates illness (e.g., ingesting laxatives, causing GI bleed­ ing, feigning symptoms of medical illness) to obtain certain effects (e.g., to receive narcotics or operations and proce­ dures); and borderline personality disorder, in which the individual demonstrates unstable and intense (e.g., overly dependent) interpersonal relationships, experiences marked shifts in mood, and exhibits impulsive (e.g., suicidal, self-mutilating, sexual) behaviors.47 It is important for the physician to recognize these patterns to avoid maladaptive interactions, to maintain clear boundaries of medical care (e.g., not to overdo studies based on the patient’s requests) and, when necessary, to refer the patient to a mental health professional skilled in the care of patients with these conditions.

Psychological Distress

Even for a previously healthy person, having an illness can cause psychological distress, which is understood as tran­ sient and modifiable symptoms of anxiety, depression, and other mood disturbances. Psychological distress also affects the medical disorder and its outcome; it lowers the pain threshold48 and influences health care seeking for patients with a functional bowel disturbance and those with a struc­ tural disease.14 Psychosocial difficulties may not be recog­

341

342

Section III  Symptoms, Signs, and Biopsychosocial Issues nized by the patient, even if evident to the health care provider. When patients with IBS who seek health care are compared with those who do not see a physician, the former group report greater psychological difficulties but also may deny the role of these difficulties in their illnesses.49 This pattern may develop early in life. The young child, Johnny, described earlier, who was becoming conditioned to report somatic symptoms when distressed, may not recognize or communicate the association of symptoms with the stressful antecedents, because these antecedents were not acknowl­ edged or attended to within the family. The ability to become consciously aware of one’s own feelings is believed to be a cognitive skill that goes through a developmental process similar to that which Piaget described for other cognitive functions.50 This development, however, may be suppressed in oppressive family environments. Alexithymia (from the Greek, “absence of words for emo­ tions”) describes patients who have chronic difficulties rec­ ognizing and verbalizing emotions. Alexithymia is believed to develop in response to early traumatic experiences such as abuse, severe childhood illness, or deprivation. Patients with alexithymia may express strong emotions, such as anger or sadness, in relation to their illnesses, but they know little about the psychological basis for these feelings and cannot link them with past experience or current illness.51 This lack of understanding limits their ability to regulate emotions, use coping strategies effectively, and adjust to their chronic condition. Their tendency to communicate emotional distress through somatic symptoms and illness behavior, rather than verbally, appears to be associated with more frequent physician visits and a poorer prognosis.52 Difficulties in awareness of one’s own emotional levels has been reported in the practice of gastroenterology in patients with functional as well as structural GI disorders.53,54

COPING AND SOCIAL SUPPORT

Coping and social support modulate (by buffering [turning down] or enabling [turning up and amplifying]) the effects of life stress, abuse, and morbid psychological factors on the illness and its outcome. Coping has been defined as “efforts, both action-oriented and intrapsychic, to manage (i.e., master, tolerate, minimize) environmental and internal demands and conflicts that tax or exceed a person’s resources.”55 In general, emotion-based coping (e.g., denial or distraction), although possibly adaptive for acute over­ whelming stresses, is not effective for chronic stressors, whereas problem-based coping strategies (e.g., seeking social support or reappraising the stressor) involves efforts to change one’s response to the stressor and is more effective for chronic illness. Patients with Crohn’s disease who do not engage in emotion-based coping, such as social diver­ sion or distraction, are less likely to relapse.56 For GI diag­ noses of all types, we have found57 that a maladaptive emotional coping style, specifically catastrophizing, along with the perceived inability to decrease symptoms, led to higher pain scores, more physician visits, and poorer func­ tioning over the subsequent one-year period. Catastrophiz­ ing is also associated with more difficult interpersonal relationships,58 predicts postoperative pain,59 and contrib­ utes to greater worry and suffering in patients with IBS.60 Therefore, efforts made through psychological treatments to improve a person’s appraisal of the stress of illness and ability to manage symptoms is likely to improve health status and outcome.14 Social support through family, religious, and community organizations and other social networks can have similar benefits in reducing the impact of stressors on physical and mental illness, thereby improving ability to cope with the

illness.61 Using IBD as an example, patients who have sat­ isfactory social support are able to reduce the psychological distress related to their conditions,62 and good social support improves health-related quality of life after surgery.63

BRAIN-GUT AXIS The combined functioning of GI motor, GI sensory, and CNS activity is considered the brain-gut axis, and dysregulation of this system’s homeostasis explains altered GI function­ ing, GI symptoms, and functional GI disorders to a great extent. The brain-gut axis is a bidirectional and integrated system in which thoughts, feelings, memories, and environ­ mental influences can lead to neurotransmitter release that affects sensory, motor, endocrine, autonomic, immune, and inflammatory function.64,65 Conversely, altered functioning or disease of the GI tract can reciprocally affect mental functioning. In effect, the brain-gut axis is the neuroana­ tomic and neurophysiologic substrate of the clinical appli­ cation of the biopsychosocial model. This section will address known and potential mecha­ nisms for communication between brain (i.e., the psychoso­ cial environment) and gut (physiology and pathology) through the central and enteric nervous systems (CNS-ENS, or brain-gut axis) and its consequences on peripheral systems (e.g., motility, sensation, inflammation, symptoms, and clinical outcome).

STRESS AND GASTROINTESTINAL FUNCTION

An observational relationship between stress and GI func­ tion has been part of the writings of poets and philosophers for centuries.66 Stress is difficult to understand and study; no definition is entirely satisfactory. The human organism functions in a constantly changing environment. Any influ­ ence on one’s steady state that requires adjustment or adap­ tation can be considered stress. The term is nonspecific and encompasses the stimulus and its effects. The stimulus can be a biologic event such as infection, a social event such as a change of residence, or even a disturbing thought. Stress can be desirable or undesirable. Some stimuli, such as pain, sex, or threat of injury, often elicit a predictable response in animals and humans. By contrast, life events and many other psychological processes have more varied effects. For example, a change of jobs may be of little concern to one person but a crisis to another, who perceives it as a personal failure. A stimulus may produce a variety of responses in different persons or in the same person at different times. The effect may be nonobservable, a psychological response (anxiety, depression), a physiologic change (diarrhea, dia­ phoresis), the onset of disease (asthma, colitis), or any com­ bination of these. A person’s interpretation of events as stressful or not and his or her response to stress depend on prior experience, attitudes, coping mechanisms, personal­ ity, culture, and biological factors, including susceptibility to disease. The nonspecific nature of the term stress pre­ cludes its use as a distinct variable for research. Healthy subjects commonly have abdominal discomfort or change in bowel function when they are upset or dis­ tressed,67 a fact usually taken into account by clinicians who manage patients. Clinical reports and psychophysiologic studies in animals and humans support these observations. Cannon noted a cessation in bowel activity in cats when reacting to a growling dog.68 Pavlov first reported that psychic factors affect gastric acid secretion via the vagus nerve in dogs.69 In humans, Beaumont,70 Wolf and Wolff,71 and Engel and colleagues72 observed changes in the color of

Chapter 21  Biopsychosocial Issues in Gastroenterology the mucosa and secretory activity of a gastric pouch or fistula in response to psychological and physical stimuli. Gastric hyperemia and increased motility and secretion were linked to feelings of anger, intense pleasure, or aggres­ sive behavior toward others. Conversely, mucosal pallor and decreased secretion and motor activity accompanied fear or depression, states of withdrawal (i.e., giving up behavior), or disengagement from others. Subsequently, studies were reported that showed the effects of experimental stress on physiologic functioning in almost all segments of the GI tract. Complicated cognitive tasks produce the following: (1) high-amplitude, highvelocity esophageal contractions73; (2) increased chymo­ trypsin output from the pancreas74; (3) reduction of phase II intestinal motor activity75; and (4) prolongation of phase III activity of the migrating myoelectric complex (MMC)74 in the small intestine (see Chapters 42, 56, and 97). Experimen­ tally induced anger increases motor and spike potential activity in the colon, and this change is greater in patients with functional bowel disorders (see Chapter 98).76 In one study,48 psychological stress was more likely than physical stress to produce propagated contractions in the colon, which presumably would be associated with the develop­ ment of diarrhea.77 Physical or psychological stress also can lower the pain threshold, particularly in patients with IBS.

ROLE OF NEUROTRANSMITTERS

The richly innervated nerve plexuses and neuroendocrine associations of the CNS and ENS provide the hard wiring for reciprocal activity between brain and gut. The mediation of these activities involves neurotransmitters and neuropeptides found in the CNS and intestine, including corticotropin-releasing factor (CRF), vasoactive intestinal polypeptide (VIP), 5-hydroxytryptamine (5-HT) and its congeners, substance P, nitric oxide (NO), cholecystokinin (CCK), and enkephalin. Depending on their location, these substances have integrated activities on GI function and human behavior. For example, the stress hormone CRF has central stress modulatory and peripheral gut physiologic effects. It produces gastric stasis and an increase in the colonic transit rate in response to psychologically aversive stimuli78 and can increase visceral hypersensitivity79 and alter immune functioning.64 Thus, CRF appears to be active in stress-induced exacerbations of IBS80 and in cyclic vomit­ ing syndrome (see Chapters 14 and 118).81

REGULATION OF VISCERAL PAIN

An important example of brain-gut axis function relates to the regulation of visceral pain.

Transmission to the Central Nervous System

Figure 21-3 shows ascending afferent pathways from the colon. After visceral stimulation by colonic dilatation, firstorder visceral neurons are stimulated and then project to the spinal cord, where they synapse with second-order neurons and ascend to the thalamus and midbrain (see Chapter 11). Of the several supraspinal pathways (spinotha­ lamic, spinoreticular, and spinomesencephalic), the spino­ thalamic tract shown on the right in Figure 21-3 terminates in the medial thalamus and projects as a third-order neuron to the primary somatosensory cortex. This pathway is important for sensory discrimination and localization of visceral and somatic stimuli (i.e., determining the location and intensity of pain). The spinoreticular tract (middle pathway) conducts sensory information from the spinal cord to the brainstem (reticular formation). Notably, this region is involved mainly in the affective and motivational properties of visceral stimulation—that is, the emotional

Primary somatosensory cortex

MCC

pACC

Thalamus Insula

Reticulothalamic

Spinomesencephalic Dorsal reticular nucleus

Spinothalamic Spinoreticular Spinal cord

Colon

Figure 21-3.  Visceral pain transmission to the central nervous system. This figure shows the ascending visceral pathways from the intestine to soma­ tosensory and limbic structures in the brain via spinal and midbrain pathways. See text for details. MCC, midcingulate cortex; pACC, perigenual anterior cingulate cortex. (From Drossman DA. Functional abdominal pain syndrome. Clin Gastroenterol Hepatol 2004; 2:353-65.)

component of pain. The reticulothalamic tract projects from the reticular formation to the medial thalamus on the left and then to the cingulate cortex. This is shown in Figure 21-3 as divided into subcomponents, including the perigenual anterior cingulate cortex (pACC), MCC, and insula, which are involved with the processing of noxious visceral and somatic information. This multicomponent integration of nociceptive information, dispersed to the somatotypic-intensity area (lateral sensory cortex) and to the emotional or motivational-affective area of the medial cortex, explains variability in the experience and reporting of pain. This conceptual scheme of pain modulation through sensory and motivational-affective components has been supported through positron emission tomography (PET) imaging using radiolabeled oxygen.82 In healthy subjects who immersed their hands in hot (47°C) water, hypnotic suggestion could make the experience painful or pleasant. Notably, no differences were observed between the two groups in somatosensory cortical activation, but in hypno­ tized subjects who experienced the hand immersion as painful, the activation of the anterior cingulate cortex (ACC) was higher. Thus, the hypnotic suggestion differen­tiated the functioning of these two pain systems. The suggestion of unpleasantness is specifically encoded in the anterior midcingulate portion of the ACC, an area involved with negative perceptions of fear and unpleasantness; this is the area associated with functional pain syndromes (see later).

Amplification of Visceral Afferent Signals

The evidence is growing that visceral inflammation and injury can amplify ascending visceral pathways. An increase in the sensitivity of peripheral receptors or the excitability of spinal or higher CNS pain regulatory systems may produce hyperalgesia (increased pain response to a noxious signal), allodynia (increased pain response to non-noxious

343

344

Section III  Symptoms, Signs, and Biopsychosocial Issues or regulatory signals), or chronic pain.65 Chronic pain results from prior episodes of recurrent pain events that become generalized to a persistent symptom presentation. The CNS response (hyperalgesia) to peripheral injury may be reduced by a preemptive reduction in the afferent input to the spinal cord and CNS (e.g., by analgesia or a local anesthetic).83 Visceral hypersensitivity is well demonstrated in imma­ ture (i.e., neonate) animals, in which inflammation or injury to nerve fibers can alter the function and structure of periph­ eral neurons84 and thereby result in a greater pain response to visceral distention when the animals mature and become adults.85 In humans, repetitive balloon inflations can later lead to an enhanced pain response to rectal distention (vis­ ceral hypersensitivity),86 as shown by the observation that patients with IBS have a greater degree of and more pro­ longed pain after a colonoscopy. Perhaps the best clinical model for the effects of inflam­ mation on visceral hypersensitivity is the postinfection IBS model (PI-IBS).87 PI-IBS results from an inflammationinduced altered mucosal immune response that sensitizes visceral afferent nerves88 in a setting of emotional distress.89 The CNS amplification of the visceral signals that occur in psychologically distressed persons probably raises the afferent signals to conscious awareness and an enhanced perception of symptoms.90 In multivariate analyses, both enterochromaffin cell hyperplasia (with increased produc­ tion of 5-HT) and depression are equally important predic­ tors of the development of PI-IBS (risk ratio, 3.8 and 3.2, respectively).91 These data support the contention that for PI-IBS to become clinically expressed, evidence of brain-gut dysfunction must exist, with visceral sensitization and high levels of psychological distress. Some of the effect may result from stress-mediated activation of mast cells, which in turn leads to neural sensitization.92

Descending Modulation

Figure 21-4 shows the central descending inhibitory system that is believed to originate in the pACC, an area rich in opioids.93 Activation of this region from visceral afferent activity may down-regulate afferent signals via descending corticofugal inhibitory pathways. Descending connections from the ACC and the amygdala to pontomedullary net­ works, including the periaqueductal gray (PAG), rostral ventral medulla (RVM), and raphe nuclei, activate inhibi­ tory pathways via opioidergic, serotonergic, and noradren­ ergic systems94 to the dorsal horn of the spinal cord. The dorsal horn acts like a gate to increase or decrease the pro­ jection of afferent impulses arising from peripheral nocicep­ tive sites to the CNS (see Chapter 11). Psychological treatments and antidepressants are thought to activate these descending pathways.

Psychological Distress and Its Influence on Central Amplification

Psychological disturbances may amplify the pain experi­ ence (i.e., CNS sensitization).95 Whereas peripheral sensiti­ zation may influence the onset and short-term continuation of pain, the CNS appears involved in the predisposition and perpetuation of pain, thereby leading to a more severe chronic pain condition. Empirical data have supported the hypothesis that a comorbid psychiatric diagnosis, major life stress, history of sexual or physical abuse, poor social support, and maladaptive coping are associated with more severe and more chronic abdominal pain and a poorer health outcome.43,96 Figure 21-5 demonstrates this associa­ tion for functional and structural GI disorders. For most

ACC

Thalamus PAG Locus coeruleus

Caudal raphe nucleus Noradrenergic Serotonergic Spinal cord

Amygdala Rostral ventral medulla Opioidergic Colon

Figure 21-4.  Descending transmission of pain regulatory signals. This figure demonstrates the corticofugal descending inhibitory pathways from the central nervous system to the spinal cord. The descending pathway is consistent with the gate control theory of pain modulation. See text for details. ACC, anterior cingulate cortex; PAG, periaqueductal gray. (From Drossman DA. Functional abdominal pain syndrome. Clin Gastroenterol Hepatol 2004; 2:353-65.)

patients with mild-to-moderate symptoms, environmental and bowel-related factors (e.g., intestinal infection, inflam­ mation, or injury, diet, hormonal factors) can lead to afferent excitation and up-regulation of afferent neuronal activity. Patients with moderate to severe symptoms also have impaired central modulation of pain as a result of various psychosocial factors, with decreased central inhibitory effects on afferent signals at the level of the spinal cord (disinhibition). Knowing the severity of the disorder and the purported site of action (i.e., intestine, brain, or both) can help when choosing an approach to treatment. When peripheral influences on severity predominate, medica­ tions, surgery, or other modalities that act on the intestine are primary treatment considerations. As pain symptoms become more severe, however, behavioral and psychophar­ macologic treatments need to be added.

Cingulate Mediation of Psychosocial Distress and Pain

The relationship between psychosocial distress and painful GI symptoms appears to be mediated through impairment in the ability of the cingulate cortex to process incoming visceral signals. The ACC, which is involved in the motiva­ tional and affective components of the limbic, or medial, pain system, is dysfunctional in patients with IBS and other chronic painful conditions like fibromyalgia97,98 and may be similarly involved with structural diagnoses, such as IBDIBS.3 The pACC, an area rich in opioids associated with emotional encoding and down-regulation of pain, and the dorsal ACC (also called the rostral or anterior MCC), may be activated to varying degrees in response to painful stimuli. The dorsal ACC, along with the amygdala, is associ­ ated with unpleasantness, fear, and an increase in responses to motor pain.99 When PET and functional magnetic reso­ nance imaging (fMRI) are used to evaluate the response of the ACC to rectal distention or to the anticipation of disten­ sion, patients with IBS display preferential activation of the

Chapter 21  Biopsychosocial Issues in Gastroenterology Bowel injury Infection

Hormones Peptides

Diet

Life stress Afferent excitation

Psychiatric diagnosis

Poor coping Abuse

Disinhibition Mild Moderate Severe Figure 21-5.  Brain-gut influences on symptom severity. This figure conceptualizes brain-gut influences on symptom severity (horizontal axis). With mild to moderate symptoms, gut-related factors (e.g., infection, inflammation, bowel injury, hormones, and peptides) lead to afferent excitation and upregulation of afferent neuronal activity. For the smaller group of patients with moderate to severe symptoms, central modulation of pain is impaired, leading to decreased central inhibitory effects on afferent signals at the level of the spinal cord (disinhibition). Factors that contribute to this effect may include life stress and abuse, comorbid psychiatric diagnoses and poor coping. Knowing the purported site of action (intestine, brain, or both) can help in determining the treatment approach, such as whether to use medications targeting the intestine or brain. (From Drossman DA. The biopsychosocial continuum in visceral pain. In: Pasricha PJ, Willis D, Gehhart GF, editors. Chronic abdominal and visceral pain: Theory and practice. New York: Informa Healthcare; 2006.)

MCC and less activation of the pACC than controls.97,100,101 In IBS, activation of the descending inhibitory pain pathway that originates in the opioid-rich pACC might be supplanted by activation of the MCC, the area associated with fear and unpleasantness. Similar findings occur in patients with somatization102 and post-traumatic stress disorder (PTSD).103 A history of abuse amplifies GI pain via central dysregula­ tion. A history of abuse in a patient with IBS leads to greater dorsal ACC activation and reporting of pain with rectal distention than either condition alone.44 This finding sup­ ports observational studies that patients with GI disorders of any type who have a history of abuse report more pain and have poorer health behaviors than those with only the GI diagnosis.43 With clinical recovery, CNS activity appears to return to normal (i.e., reduced MCC and increased insular activation).45,104 These observations have been supported by a study of the Gulf War and related health issues.105 The report portrays a strong relationship between the deployment of soldiers to a war zone with traumatic exposure to injury, mutilation, or dead bodies and the ensuing development of medical and psychological symptoms and syndromes. In fact, clusters of several medical symptoms were noted, such as those termed Gulf War syndrome, which includes IBS, chronic fatigue, and chemical sensitivity syndrome, in addition to PTSD and cognitive impairments. Therefore, the psychological effects of abuse or wartime exposure may produce disruption in central pain modula­ tion systems and in brain circuits at the interface of emotion and pain.44 This change leads to a lowering of sensation thresholds, with a loss of the brain’s ability to filter bodily sensations. The result is an increase in physical and psy­ chological symptoms and more intense pain and syndromes (e.g., IBS, fibromyalgia, headache, widespread body pain), a condition that has been variably described as somatiza­ tion, comorbidity, or just extraintestinal functional GI symptoms.105 The data suggest that psychological and anti­ depressant treatments are potentially beneficial for more severe forms of chronic pain, in which CNS contributions are thought to be preeminent.

EFFECTS OF STRESS ON IMMUNE FUNCTION AND DISEASE SUSCEPTIBILITY

Stressful experiences can also affect peripheral immune function and ultimately susceptibility to disease; con­ versely, peripheral immune and inflammatory mediators can affect psychological functioning. The discipline of psy­ choneuroimmunology began in the 1970s, when alterations of immune function were first documented in astronauts after splashdown.106 Later, in vitro experiments showed effects of acute stress on lymphoproliferative activity, inter­ feron production, and DNA repair.107,108 Clinical studies have shown that chronically distressed persons (e.g., longterm caregivers of spouses with dementia) have impaired cellular immunity and higher frequencies of depression and respiratory infections compared with matched controls,109 and persons under psychological distress have a higher frequency of respiratory infections after intranasal virus inoculation.110 In patients with multiple myeloma, stress management treatment has been associated with an increased number of natural killer (NK) cells, a significantly lower mortality rate after six years, and a trend toward fewer tumor recurrences as compared with a control group of patients who did not undergo stress management treatment.111 The principal mediators of the stress-immune response include CRF and the locus-coeruleus-norepinephrine (LC-NE) systems in the CNS. These systems are influenced by numerous positive and negative feedback systems that allow behavioral and peripheral adaptations to stress.112 The peripheral limb of the CRF system is the hypothalamicpituitary-adrenal (HPA) axis, a negative feedback system involved in psychoneuroimmunologic regulation. In the HPA system, inflammatory cytokines, primarily tumor necrosis factor-α (TNF-α), interleukin (IL)-1, and IL-6, liberated during inflammation, stimulate the paraven­ tricular nucleus of the hypothalamus to secrete CRF. CRF stimulates the pituitary gland to release adrenocorticotropic hormone (ACTH; corticotropin), which, in turn, stimulates the adrenal glands to release glucocorticoids. Finally, the glucocorticoids suppress inflammation and cytokine

345

346

Section III  Symptoms, Signs, and Biopsychosocial Issues production, thereby completing the negative feedback loop.113 CRF is recognized to be important via central and periph­ eral pathways in stress-related modulation of GI motor and sensory function114 and may be involved in the generation or maintenance of pain-related symptoms sensitive to mod­ ulation by psychological stress.115 Furthermore, disruptions in the HPA system can lead to behavioral and systemic disorders as a result of increased (e.g., Cushing’s syndrome, depression, susceptibility to infection) or decreased (e.g., adrenal insufficiency, rheumatoid arthritis, chronic fatigue syndrome, PTSD) HPA axis reactivity.112 Inflammatory GI disorders (e.g., IBD) also may be affected through this stressmediated system.41 The clinical evidence to support a role for stress in activa­ tion of human IBD, or other structural disorders, is limited because of methodologic limitations of available studies. The evidence from clinical observational data, however, is compelling. For example, in one study of 62 patients with ulcerative colitis who were followed prospectively for more than five years,116 27 experienced an exacerbation, and the psychosocial predictors of disease exacerbation were assessed. Patients who scored high on a measure of longterm perceived stress had an increased actuarial risk of an exacerbation of ulcerative colitis (hazard ratio = 2.8; 95% confidence interval [CI], 1.1 to 7.2), and exacerbation was not associated with short-term perceived stress or with the extent, duration, or severity of the colitis.

CYTOKINES AND THE BRAIN

Stress might have proinflammatory effects, but intestinal inflammation also may affect behavior reciprocally via acti­ vation of cytokines. Many behavioral features (e.g., fever, fatigue, anorexia, depression) of chronic inflammatory dis­ eases may result from the central effects of peripherally activated inflammatory cytokines.117,118 Sickness behavior refers to the coordinated set of behavioral changes that develop in sick persons during the course of IBD, cancer, infection, or other catabolic disorders associated with cyto­ kine activation. At the molecular level, these changes result from the effects of proinflammatory cytokines, such as IL-1 and TNF-α, in the brain. Peripherally released cytokines act on the brain via a fast transmission pathway, which involves primary afferent nerves that innervate the site of inflamma­ tion, and a slow transmission pathway, which involves cytokines that originate from the choroid plexus and cir­ cumventricular organs and diffuse into the brain paren­ chyma. At the behavioral level, sickness behavior appears to be the expression of a central motivational state that reorganizes the organism’s priorities to cope with infectious pathogens.

SYMPTOM EXPERIENCE AND BEHAVIOR The product of the interacting effects of brain and gut relate to the clinical expression of illness and disease—namely, the symptom experience and subsequent illness-related behaviors. The meaning of illness, the perceived effect of alterations in body image (e.g., a colostomy), social accept­ ability, the degree of functional impairment and its implica­ tions at work and at home, and the likelihood of surgery or untimely death must all be dealt with by the patient. How well the patient adapts, in addition to the quality of the physician’s involvement, is crucial to the patient’s psycho­ logical well-being and clinical course. Some chronically ill patients regress and become dependent. Their continued

symptoms, restricted activity, and health care tax family, friends, and physician, all of whom may feel helpless to provide enough emotional or medical assistance. Con­ versely, other patients resist help to avoid acknowledging their imposed dependence. The family must then deal with feelings of guilt and anger, the expressions of which, although unavoidable, are not usually socially permitted. Often, the physician carries the burden of the feelings of the patient and family and must reconcile the two. In most cases, the problems are worked out, and the patient estab­ lishes a pattern of coping. However, if the patient has limited capacity to cope psychologically with the illness, the disorder is particularly incapacitating, or the interper­ sonal family relationships are unstable, additional efforts by the physician and ancillary personnel (e.g., psychological counselors, social workers, peer support groups) will be required.

OUTCOME The severity or activity of GI disease is not sufficient to explain a patient’s health status and its consequences fully. Early life conditioning and current psychological difficul­ ties also influence a variety of outcome measures, including symptoms, health care seeking, quality of life—a global measure of the patient’s perceptions, illness experience, and functional status that incorporates social, cultural, psycho­ logical, and disease-related factors—and health care costs, in some cases more than disease-related factors.14 When the effects of early life conditioning on health care visits and costs were assessed, the children of parents diagnosed with IBS had significantly more ambulatory care visits for all medical (12.3 vs. 9.8; P < 0.0001) and GI symptoms (0.35 vs. 0.18; P < 0.0001), and outpatient health care costs over three years were also higher ($1979 vs. $1546; P < 0.0001) than those for a comparison group of children whose parents did not have IBS.119 Presumably, the parents with IBS were vigilant, and responded more often, to the symptoms (GI or non-GI) of their children and more readily sought health care for them. In adults, psychosocial factors affect symptom reporting and health care seeking. In one survey of 997 subjects with IBD who belonged to the Crohn’s and Colitis Foundation of America,120 the number of physician visits was related to psychosocial factors (e.g., psychological distress, perceived well-being, and physical functioning), whereas the severity of symptoms was not found to predict this outcome. Other important psychosocial predictors of poorer health outcome (e.g., symptom severity, phone calls, doctor visits, daily function, and health-related quality of life) for functional or structural disorders include a history of sexual or physical abuse,43 maladaptive illness beliefs,121,122 ineffective coping strategies (e.g., catastrophizing), and perceived inability to decrease symptoms.57

CLINICAL APPLICATIONS The data relating to the scientific basis for an association between psychosocial factors and GI illness and disease require that the physician obtain, organize, and integrate psychosocial information to achieve optimal care. The recommendations offered here are particularly useful for patients who have chronic illness or major psychosocial difficulties. More comprehensive discussions of techniques

Chapter 21  Biopsychosocial Issues in Gastroenterology for obtaining and analyzing the data and of the interview process are found elsewhere.47,123,124

HISTORY

The physician’s dialogue with the patient is the most impor­ tant asset for enhancing the physician-patient relationship, developing a diagnosis, and formulating treatment and is often underused. Consider the information obtained in this office interview. Physician: “How can I help you?” (looking at chart) Patient: “I developed a flare-up of my Crohn’s … the pain, nausea, and vomiting, when I came back from vacation.” (pause, looks pensive) Physician: “Was the pain like what you had before?” (interrupting) Patient: “Yes, well almost, I think.” Physician: “Was it made worse by food?” (looks up) Patient: “Yes.” Physician: “Did you have fever? or diarrhea?” (leaning forward) Patient: “Well yes, I think. … I didn’t take my temperature.” (looks down) Physician: “So you had fever and diarrhea?” Patient: “Uh no, well, they were a little loose … I guess.” In this exchange, some relevant information was not elic­ ited and, because of the interruptions and leading questions, the accuracy of the information after the first question is uncertain. Furthermore, the nonverbal communication did not facilitate an effective physician-patient interaction. The medical history should be obtained through a patientcentered nondirective interview during which the patient is encouraged to tell the story in his or her own way, so that the events contributing to the illness unfold naturally.125 Open-ended questions are used initially to generate hypoth­ eses, and additional information is obtained with facilitat­ ing expressions—“Yes?,” “Can you tell me more?” Repeating the patient’s previous statements, head nodding, or even silent pauses with an expectant look can facilitate history taking. Avoid closed-ended (yes-no) questions at first, although they can be used later to characterize the symp­ toms further. Never use multiple-choice or leading ques­ tions, because the patient’s desire to comply may bias the responses. The traditional medical and social histories should not be separated, but elicited together, so that the medical problem is described in the context of the psychosocial events surrounding the illness. The setting of symptom onset or exacerbation should always be obtained. At all times, the questions should communicate the physician’s willingness to address the biological and psychological aspects of the illness: Physician: “How can I help you?” (concerned, looking at patient) Patient: “I developed a flare-up of my Crohn’s … the pain, nausea, and vomiting, when I came back from vacation.” (pause) Physician: “Yes?” Patient: “I was about to start my new position as floor supervisor and thought I’d take a vacation to get prepared, and then all this happened.” Physician: “Oh, I see.” (pause) Patient: (continues) “I started getting that cramping feeling right here (points to lower abdomen) and then it got worse after eating. So I knew I’d be obstructed again if I didn’t get in to see you.” Physician: “Hmm. Any other symptoms?” Patient: “Well, I felt warm, but didn’t take my temperature.”

Physician: “What was your bowel pattern like?” Patient: “They started getting loose when I was on vacation. Now they’re slowing down. I haven’t gone today.” The number of verbal exchanges is the same, yet the patient offers more information. The clinical features are clearer and additional knowledge of an association of symp­ toms with beginning a new job situation is obtained. This interview method also encourages patient self-awareness and allows consideration of possible behavioral treatments (e.g., stress reduction techniques, job change, and counsel­ ing) that may ameliorate future flare-ups of the patient’s symptoms. The historical information should be obtained from the perspective of the patient’s understanding of the illness. Important questions to ask include the following126: “What do you think is causing this problem?” “Why did it happen now?” “What type of treatment do you think you should receive?” “What do you fear most about your illness?”126

EVALUATING THE DATA

The physician must assess the relative influences of the biological, psychological, and social dimensions on the illness. Determining whether psychosocial or biological processes are operative in an illness is unnecessary and possibly countertherapeutic. Usually, both are important, and treatment is based on determining which is identifiable and remediable. A negative medical evaluation is not suf­ ficient for making a psychosocial diagnosis. Table 21-1 lists several questions to consider in the assessment and evalu­ ation of the patient.47

DIAGNOSTIC DECISION MAKING

Deciding which tests to order will depend on their clinical usefulness. A number of questions should be considered: Is a test safe and cost-effective? Will the results make a difference in treatment? Patients who are persistent in their requests for further studies, or who challenge their physician’s competence, may tempt the physician to schedule unneeded studies or surgery out of uncertainty or out of feeling that he or she needs to do something. This temptation can be avoided by basing decisions on the objective evaluation of data (e.g., blood in the stool, fever, abnormal serum chemistry values) rather than solely on the patient’s illness behavior. The case of Ms. L, the patient with persistent and unex­ plained abdominal pain, is an example familiar to the gas­ troenterologist. The urge to work up a patient with chronic abdominal pain must be tempered by the evidence that

Table 21-1  Questions to Consider in the Clinical Evaluation of the Patient Does the patient have acute or chronic illness? What is the patient’s life history of illness? Why is the patient coming for medical care now? What are the patient’s perceptions and expectations? Does the patient exhibit abnormal illness behavior? What is the impact of the illness on the patient? Is there a concurrent psychiatric diagnosis? Are there cultural or ethnic influences? How does the family interact around the illness? What are the patient’s other psychosocial resources? How extensive should the evaluation be? Should the patient be referred to a psychiatric consultant?

347

348

Section III  Symptoms, Signs, and Biopsychosocial Issues an adequate initial evaluation considerably reduces the likelihood of finding an overlooked cause later. Here, the clinical approach is not medical diagnosis, but psychosocial assessment and treatment of the chronic pain.6,126 Factors associated with or exacerbating chronic pain symptoms include the following: (1) a recent disruption in the family or social environment (e.g., child leaving home, argument); (2) major loss or anniversaries of losses (e.g., death of a family member or friend, hysterectomy, interference with the outcome of pregnancy); (3) history of sexual or physical abuse; (4) onset or worsening of depression or other psychi­ atric diagnosis; and (5) a hidden agenda (e.g., narcoticseeking behavior, laxative abuse, pending litigation, disability). Although psychiatric consultation and treatment may be needed, it is important that the physician should continue to be involved in the patient’s care and be vigilant about the development of new findings. At times, decisions must be made with incomplete or nonspecific information. Particularly for chronic symptoms, when studies are unrevealing and the patient is clinically stable, it is wise to tolerate the uncertainty in diagnosis and observe the patient for new developments over a period of time. Experienced physicians usually make diagnostic and treatment decisions based on the degree of change in the condition over weeks or months, rather than on one or two occasions. Psychiatric consultation should be considered when additional psychological data could clarify the illness or improve patient care. Examples include the following: (1) identification of a psychiatric diagnosis for which specific treatment (e.g., psychopharmacologic agents) may be benefi­ cial; (2) serious impairment of the patient’s level of psycho­ social functioning (e.g., inability to work); or (3) consideration of invasive diagnostic or therapeutic strategies on the basis of the patient’s complaints, without clear indications from the medical data.

TREATMENT APPROACH ESTABLISHING A THERAPEUTIC RELATIONSHIP

The physician establishes a therapeutic relationship when he or she does the following: (1) elicits and validates the patient’s beliefs, concerns, and expectations; (2) offers empathy when needed; (3) clarifies the patient’s misunder­ standings; (4) provides education; and (5) negotiates the plan of treatment with the patient.125 This strategy must be individualized, because patients vary in the degree of negotiation and participation they require. The physician must be nonjudgmental, show interest in the patient’s well-being, and be prepared to exercise effective communi­ cation skills.123

Eliciting, Evaluating, and Communicating the Role of Psychosocial Factors

A sensitive and nonjudgmental interview facilitates a patient’s disclosure of psychosocial information. At times, the patient may be unwilling or unable to discuss this infor­ mation, particularly on the first visit or in the context of seeing the gastroenterologist for a medical problem. The fear of disapproval and lack of trust often prevent the patient from sharing intimate thoughts and feelings, an obstacle that can be overcome by a good physician-patient relationship. When the patient is unwilling or unable to accept the role of psychosocial factors in illness, the physician can still obtain such information indirectly, by inference, and should

not attempt to provide the patient with insight. If the patient is asked whether the problem is just “in my head,” the physician explains that illness is rarely either mental or physical; that understanding all factors, including the patient’s feelings, is important; and that many chronic con­ ditions are associated with depression or unrealistic fears. Consistent with the biopsychosocial model of illness, dis­ cussing psychosocial and biological factors in terms of cau­ sation (e.g., by stating “It is common for stress to cause your problems”) or exclusion (e.g., by stating “The workup is negative; it must be stress”) is not helpful.

Providing Reassurance

A patient’s fears and concerns require reassurance. If the reassurance is premature, inadequate, or inappropriate, it will be perceived as insincere or as a lack of thoroughness by the physician. The physician should respond to the patient’s needs and requests empathically but not go along when doing so would not be in the patient’s best interest. For example, disability may be a disincentive to helping the patient reestablish wellness and return to gainful employ­ ment. If the patient does not qualify for disability, the physi­ cian should be clear about it.

Recognizing the Patient’s Adaptations to Chronic Illness

Illness is associated with certain benefits for the patient, such as increased attention and support, release from usual responsibilities, and possibly social and financial compen­ sation. For some patients, more may be lost by giving up the state of illness than gained by wellness, and improve­ ment may be slow. The patient can be helped by improving his or her psychosocial adjustment to the illness (e.g., improving coping strategies).

Reinforcing Healthy Behaviors

Sometimes, complaints of physical distress are a maladap­ tive effort to communicate emotional distress or to receive attention.127 The physician may unwittingly reinforce this behavior in several ways: (1) by paying a great deal of atten­ tion to the patient’s complaints, to the exclusion of other aspects; (2) by acting on each complaint by ordering diag­ nostic studies or giving a prescriptive medication; or (3) by assuming total responsibility for the patient’s well-being. The patient learns to keep the physician’s interest by reporting symptoms rather than by trying to improve, per­ petuating the cycle of symptom recitation and passive interaction. To encourage a patient to take more responsibility for his or her care and have a heightened sense of control, the physician offers a choice among several treatments or helps design an exercise program. The physician should limit discussion about symptoms (often termed the organ recital) to what is needed to satisfy medical concerns and focus instead on adaptations to the illness rather than the cure. I often find it best not to ask about the patient’s symptom (e.g., “How is your pain?”), because the question puts my attention on the fact that the patient is having symptoms. Rather, I ask about the symptoms in the context of the patient’s health-promoting behaviors (e.g., “What are you doing to manage your pain?”).

PSYCHOPHARMACOLOGIC TREATMENT

Psychopharmacologic or psychotropic agents act on neu­ rotransmitter receptors in the brain-gut regulatory pathways that target serotonergic, dopaminergic, opioidergic, and noradrenergic receptor sites and produce various effects in the following ways: (1) by reducing visceral afferent signal­ ing from painful GI conditions; (2) by treating GI pain by

Chapter 21  Biopsychosocial Issues in Gastroenterology facilitating central down-regulating pathways; (3) depend­ ing on the agent, by modifying diarrhea or constipation; (4) by reducing anxiety, depression, nausea, and loss of appe­ tite; and (5) in higher doses, by treating major depression or other psychiatric disorders.128 The tricyclic antidepressants (TCAs) in relatively low doses can reduce chronic pain129,130 through peripheral and central mechanisms, including activation of corticofugal pain inhibitory pathways by endorphins. They can also treat major and secondary depressive symptoms when used in full antidepressant doses. They may have anti­ histaminic and anticholinergic effects that could lead to nonadherence. Selective serotonin reuptake inhibitors (SSRIs) are not as established as TCAs as antinociceptive agents. They are often used, however, to reduce anxiety and are particularly helpful for psychiatric comorbid conditions, including major depression, panic disorder, and other high-anxiety conditions (e.g., obsessive-compulsive disorder, PTSD, social phobia). The serotonin-norepinephrine receptor inhibitors (SNRIs) are a more recently introduced class of antidepressants that appear to be particularly helpful for the treatment of painful conditions. The presence of any syndromes associated with deterioration in daily function (e.g., inability to work) or of vegetative symptoms (e.g., poor appetite, weight loss, sleep disturbance, decreased energy and libido) subserved by central brain monoamine function, are reasonable indica­ tions for a therapeutic trial. They should be considered even if the patient denies feelings of sadness (masked depres­ sion). Treatment should be increased to full therapeutic levels over two to three weeks and maintained for six to nine months. Poor clinical responses may be the result of relatively low doses.131 Anxiolytic agents, particularly the benzodiazepines, are frequently used to ameliorate anxiety disorders or used for a short term for acute anxiety, particularly if the anxiety is associated with stress-induced flare-ups of bowel distur­ bance. Their potential benefit should be balanced with the long-term risks of sedation, drug interactions, habituation, and rebound after withdrawal. Antipsychotic drugs, or neuroleptics, include the pheno­ thiazines (e.g., chlorpromazine), butyrophenones (e.g., halo­ peridol), and newer classes of safer antipsychotic agents (e.g., quetiapine, olanzapine) and are used primarily for treating disturbances of thought, perception, and behavior in psychotic patients. Increasingly, they are being used in lower doses to promote sleep, achieve anxiolytic and anti­ depressant effects, and augment the analgesic effects of anti­ depressants; they may have a role in pain management.132 Opiates have little role in treating patients with chronic pain or psychosocial disturbance because of their potential for abuse, dependency, and narcotic bowel syndrome (see Chapters 11 and 120).15

PSYCHOTHERAPY AND BEHAVIORAL TREATMENTS

The need for adjunctive care by a psychiatrist, psychologist, or other mental health professional should be determined through an assessment of the personal, social, and economic hardship of the illness, rather than identification of a spe­ cific psychiatric diagnosis per se. Referral is based on the likelihood of improved function, mood, or coping style fol­ lowing the intervention. The patient must see psychological care as relevant to personal needs, rather than “to prove I’m not crazy.” Psychological treatments for GI disorders have been reviewed elsewhere.14

Cognitive-behavioral treatment can be effective for patients with functional GI symptoms, chronic pain, depres­ sion, and bulimia. It involves the patient in identifying stressors and thoughts that increase mental distress and learning new ways of coping by restructuring these thoughts. Interpersonal or dynamic psychotherapy is recommended for patients motivated to address and adjust interpersonal difficulties associated with exacerbations of symptoms. For the patient who would benefit from peer support or who has interpersonal difficulties or limited finances, group therapy can be considered. Crisis intervention is designed to help the individual over a particularly difficult period in a few sessions. It can help the patient with chronic illness and an identifiable cause of recent deterioration in function. Family or marital counseling is indicated when difficulties in family interactions interfere with health-promoting activ­ ities. Finally, other behavioral treatments (e.g., relaxation training, generalized biofeedback, meditation, and hypno­ sis) are safe, noninvasive, and cost-effective methods designed to reduce anxiety levels, teach patients how to engage in health-promoting behaviors, give the patient greater responsibility and control over his or her health care, and improve pain control.

PHYSICIAN-RELATED ISSUES

Patients’ psychosocial difficulties may affect the physician’s attitudes and behaviors7 and, if unrecognized, may adversely affect the patient’s care. Physicians are uncomfortable making decisions in the face of diagnostic uncertainty,133 because the assumption is that more knowledge will make the illness more treatable. Nonetheless, many clinical treat­ ments are undertaken for symptoms or psychosocial con­ cerns that are not based on a specific diagnosis, particularly for patients with unexplained complaints who demand a diagnosis or for those who are thought to be litigious. Here, the physician risks overdoing the diagnostic evaluation or instituting unneeded or harmful treatments in such patients (sometimes called furor medicus).134 Alternatively, the physician may not believe that the complaints are legitimate and may then exhibit behaviors (e.g., referring the patient to a psychiatrist) that the patient will recognize as a rejection. Awareness of the psychosocial dimensions of illness and their effects on the genesis of symptoms legitimizes the complaints and puts them into better perspective. Some patients and physicians can experience interpersonal con­ flicts. The patient, feeling helpless and out of control, may behave in ways interpreted by the physician as dependent and demanding; these feelings may lead to blame or stigma­ tization by the physician. The physician must understand that these behaviors are part of the patient’s maladaptive communication style and should not be perceived as criti­ cal. In our role as physicians, we may have expectations for rapid relief with displays of gratitude from patients. Some patients, however, develop psychosocial adaptations to chronic illness (e.g., family attention, control through the illness behaviors, disability) that delay ultimate improve­ ment, and these patients may not acknowledge the physi­ cian’s efforts. In these situations, treatment is best refocused from efforts to cure to efforts to improve daily function, despite continued symptoms. Gratification can be obtained from the personal effort rather than from the patient’s com­ ments of gratitude. Finally, each physician must set per­ sonal limits in time and energy toward the care of patients who are particularly challenging. Limiting the length of office visits, allocating part of the patient’s care to other health care workers and, when necessary, saying “no” to the patient are all important methods for achieving a balance

349

350

Section III  Symptoms, Signs, and Biopsychosocial Issues between the physician’s personal needs and benefit to the patient.

KEY REFERENCES

Creed F, Levy R, Bradley L, et al. Psychosocial aspects of functional gastrointestinal disorders. In: Drossman DA, Corazziari E, Delvaux M, et al, editors. Rome III: The functional gastrointestinal disorders. 3rd ed. McLean, Va: Degnon Associates; 2006. pp 295-368. (Ref 14.) Drossman DA. Presidential Address: Gastrointestinal illness and bio­ psychosocial model. Psychosom Med 1998; 60:258-67. (Ref 16.) Drossman DA. Brain imaging and its implications for studying centrally targeted treatments in IBS: A primer for gastroenterologists. Gut 2005; 54:569-73. (Ref 95.) Drossman DA. Severe and refractory chronic abdominal pain: Treatment strategies. Clin Gastroenterol Hepatol 2008; 6:978-82. (Ref 6.) Drossman DA, Leserman J, Mitchell CM, et al. Health status and health care use in persons with inflammatory bowel disease: A national sample. Dig Dis Sci 1991; 36:1746-55. (Ref 120.) Drossman DA, Ringel Y. Psychosocial factors in ulcerative colitis and Crohn’s disease. In: Sartor BR, Sandborn WJ, editors. Kirsner’s inflammatory bowel disease. 6th ed. London: WB Saunders; 2004. pp 340-356. (Ref 41.) Grover M, Drossman DA. Psychotropic agents in functional gastrointes­ tinal disorders. Curr Opin Pharmacol 2008; 8:715-23. (Ref 128.) Jones MP, Dilley JB, Drossman D, Crowell MD. Brain-gut connections in functional GI disorders: Anatomic and physiologic relationships. Neurogastroenterol Motil 2006; 18:91-103. (Ref 65.)

Kellow JE, Azpiroz F, Delvaux M, et al. Principles of applied neurogas­ troenterology: Physiology/motility-sensation. In: Drossman DA, Corazziari E, Delvaux M, et al, editors. Rome III: The functional gastrointestinal disorders. 3rd ed. McLean, Va: Degnon Associates; 2006. pp 89-160. (Ref 86.) Leserman J, Drossman DA. Relationship of abuse history to functional gastrointestinal disorders and symptoms: Some possible mediating mechanisms. Trauma Violence Abuse 2007; 8:331-43. (Ref 25.) Levenstein S, Prantera C, Varvo V, et al. Stress and exacerbation in ulcerative colitis: A prospective study of patients enrolled in remis­ sion. Am J Gastroenterol 2000; 95:1213-20. (Ref 116.) Mayeux R, Drossman DA, Basham KK, et al. Gulf war and health: Physi­ ologic, psychologic, and psychosocial effects of deployment-related stress. Washington, DC: National Academies Press; 2008. (Ref 105.) Naliboff BD, Derbyshire SWG, Munakata J, et al. Cerebral activation in irritable bowel syndrome patients and control subjects during recto­ sigmoid stimulation. Psychosom Med 2001; 63:365-75. (Ref 97.) Ringel Y, Drossman DA, Leserman JL, et al. Effect of abuse history on pain reports and brain responses to aversive visceral stimulation: An FMRI study. Gastroenterology 2008; 134:396-404. (Ref 44.) Spiller RC. Postinfectious irritable bowel syndrome. Gastroenterology 2003; 124:1662-71. (Ref 88.) Wood JD, Grundy D, Al-Chaer ED, et al. Fundamentals of neurogastro­ enterology: Basic science. In: Drossman DA, Corazziari E, Delvaux M, et al, editors. Rome III: The functional gastrointestinal disorders. 3rd ed. McLean, Va: Degnon Associates; 2006. pp 31-87. (Ref 80.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

22 Oral Disease and OralCutaneous Manifestations of Gastrointestinal and Liver Disease Ginat W. Mirowski and Lawrence A. Mark

CHAPTER OUTLINE Disorders of the Mouth and Tongue  353 Mucocutaneous Disorders  354 Candidiasis  354 Human Immunodeficiency Virus Infection  355 Ulcerative Diseases  356 Vesiculobullous Diseases  357 Lichen Planus  359 Cutaneous Manifestations of Intestinal Disease  359 Vascular and Connective Tissue Disorders  360 Cutaneous Manifestations of Gastrointestinal Malignancies  363

DISORDERS OF THE MOUTH AND TONGUE Xerostomia (dry mouth) is a common complaint with destruction or atrophy of the salivary glands as a result of autoimmune disease (Sjögren’s syndrome), after radiation therapy, or as a consequence of a variety of medications, such as anticholinergics, H1 antihistamines, tricyclic anti­ depressants, hypnotics, sedatives, antihypertensives, anti­ psychotics, antiparkinson agents, and diuretics.1 Sjögren’s syndrome is an autoimmune disease that is clas­ sified by the triad of xerostomia, keratoconjunctivitis sicca (dry eyes), and arthritis.2,3 It may be characterized as primary when no other disorders are diagnosed or secondary when connective tissue disease, such as rheumatoid arthritis or systemic lupus erythematosus, is present. The oral mani­ festations of Sjögren’s syndrome are caused by the irrever­ sible destruction of the salivary glands by a lymphocytic infiltrate that results in diminished or absent saliva. The lack of saliva is associated with difficulty chewing, odyno­ phagia, and diminished taste and smell, as well as mucosal erythema, increased incidence of dental caries, oral can­ didiasis, and salivary gland calculi. Sucking mints and chewing gum may help by increasing salivary flow, which assists in the the removal of debris. Patients with xerostomia should avoid sweets and acidic foods and beverages and be encouraged to sip water and suck ice chips frequently. Preparations containing 1% sodium carboxymethyl cellu­

Polyposis Syndromes  363 Internal Malignancy and Related Disorders  364 Cutaneous Metastases  365 Cutaneous Manifestations of Liver Disease  365 Drug-Induced Liver Disease in Patients with Skin Disease  366 Parasitic Diseases of the Intestine and Skin  367 Dermatitis Herpetiformis and Celiac Disease  367 Vitamin Deficiencies  368

lose may be used to moisten the oral cavity. Salivary stimu­ lants such as cevimeline (Evoxac), 30 mg three times daily, or pilocarpine (Salagen), 5 mg four times daily, are effective sialogogues. Glossitis, inflammation of the tongue, occurs in a hetero­ geneous group of disorders that includes nutritional defi­ ciencies, chemical irritants, drug reactions, iron deficiency, pernicious anemia, amyloidosis, sarcoidosis, infections, and vesiculoerosive diseases. Sometimes, no underlying cause can be detected.3 Patients may complain of lingual pain (glossodynia) or burning sensation (glossopyrosis). Loss of filiform papillae results in a spectrum of changes, from patchy erythema with or without erosive changes to a completely smooth, atrophic, erythematous surface (Fig. 22-1). Atrophic glossitis is a sign of protein-calorie malnu­ trition and muscle atrophy and is commonly found in older adults. Median rhomboid glossitis manifests as an asymp­ tomatic, well-defined erythematous patch in the midposte­ rior dorsum of the tongue.4 Glossodynia (burning sensation or pain in tongue) in the absence of clinical or histologic evidence of glossitis may be associated with anxiety or depression. Although it is found most commonly in postmenopausal women, hor­ monal therapy is of no value.5 Hypnosis has been found to improve glossodynia when a psychogenic component or when organic disease is present.6 Serologic evaluation for hypomagnesemia and for vitamin B12 or folate deficiency,

353

354

Section IV  Topics Involving Multiple Organs

Figure 22-1.  Glossitis in a patient with diabetes mellitus and malabsorption. The tongue is smooth (depapillated) and red; angular cheilitis is present.

as well as a complete medication history, may occasionally yield a correctable cause.7 Hypogeusia (diminished sense of taste) and dysgeusia (distortion of normal taste) are other complaints that are sometimes associated with glossitis. Hypogeusia and dys­ geusia have been attributed to various neurologic, nutri­ tional, and metabolic disorders and to a large number of medications. The evidence supporting these associations is tenuous.8,9 How taste buds are affected by aging is not understood. Tobacco smokers, denture wearers, and patients with anxiety or other psychiatric disorders commonly complain of hypogeusia and dysgeusia. Radiation therapy to the head and neck may result in altered taste. The therapy is empirical and includes identifying and correcting any associated condition. Patients may be treated with zinc supplementation, a low-dose anxiolytic, or an antidepres­ sant medication such as a selective serotonin reuptake inhibitor (SSRI).10 Paradoxically, tricyclic antidepressant medications block responses to a wide range of taste stimuli and may contribute to clinical reports of hypogeusia and dysgeusia.11,12 Geographic tongue (benign migratory glossitis) is charac­ terized by patchy loss of filiform papillae forming irregular, moving configurations that resemble geographic landmarks on a map. Geographic tongue is reported to occur in up to 4% of the population. Patients may complain of pain or difficulty in eating acidic, spicy, or salty foods. Recurrent episodes are common and may represent pustular psoriasis. Some patients may present with an exfoliative cheilitis and/or migratory annular plaques and papules on any of the oral mucosal surfaces, representing geographic mucositis in ectopic locations. Histologically, spongiosus and neutro­ philic microabscesses are found in the epithelium, with no evidence of candidiasis. Treatment consists of topical anes­ thetics, benzocaine (Orabase) or aluminium hydroxide and magnesium hydroxide (Maalox) protective coatings, and topical glucocorticoids, along with control of the underlying cutaneous psoriasis if present.13 Geographic tongue has no known associations with malignancy.14 Black hairy tongue is another common entity. The dorsal surface of the tongue may appear yellow, green, brown, or black because of exogenous pigment trapped within elon­ gated keratin strands of filiform papillae.15 Acquired black hairy tongue is seen most commonly in chronic smokers and often follows a course of systemic antibiotics, the use of hydrogen peroxide, or drinking coffee or tea.14 Off-label treatment consists of 25% podophyllum or topical tretinoin (Retin-A) gel.15 Chronic débridement with a tongue scraper may also be helpful.

Figure 22-2.  Oropharyngeal candidiasis. Multiple white and yellow plaques are seen on the soft and hard palate, uvula, and tongue. Lesions may be seen in the hypopharynx during passage of an endoscope.

MUCOCUTANEOUS DISORDERS CANDIDIASIS

Candida spp. (chiefly Candida albicans) are part of the normal flora in almost half of the population. Oral candidia­ sis or candidosis (moniliasis, thrush) typically appears as white curd-like patches or as red (atrophic) or white and red friable lesions on any mucosal surface (Fig. 22-2). Many newborns experience initial overgrowth of Candida before colonization of the gastrointestinal (GI) tract. Candidiasis often occurs during or after antibiotic or glucocorticoid therapy, in denture wearers, pregnant women, and older adults, and in patients with anemia, diabetes mellitus, Hashimoto’s thyroiditis, Cushing’s disease, or familial hypoparathyroidism. Immunosuppression caused by human immunodeficiency virus (HIV; see later), other debilitating illnesses, or cancer chemotherapy may lead to candidiasis (see Chapter 33). Candida albicans remains the predomi­ nant species cultured. However, C. glabrata, C. krusei, and other azole-resistant species must be considered in resistant cases. Oral candidiasis is also associated with xerostomia, whatever the cause. Topical therapy is most effective in patients with no underlying chronic conditions (see Chapter 45) and may entail the use of the following: (1) nystatin (Mycostatin), 100,000-U vaginal tablet dissolved orally three to five times daily; (2) clotrimazole (Mycelex), 10-mg troche to be dissolved orally five times daily; or (3) clotrimazole, 500-mg vaginal tablet, to be dissolved orally at bedtime. Topical agents are effective in the absence of immuno­ suppression, whereas oral antifungal agents are needed in immunocompromised patients (see Chapters 33 and 34). In denture wearers, adjunctive measures, including regular denture cleaning, soaking in a dilute bleach solution, and taking the dentures out overnight, are important for clear­ ing. When dysphagia and/or upper GI bleeding accompany oral thrush, concurrent candidal esophagitis should be con­ sidered (see Chapter 45). Systemic candidiasis may result when normal barriers to infection are lost. Microthrombi, resulting from obstruction of cutaneous and systemic

Chapter 22  Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease vessels, lead to local necrosis and manifest as small necrotic papules and ulcerations that are easily visible on the skin and mucosa.

HUMAN IMMUNODEFICIENCY VIRUS INFECTION

Oral and cutaneous complications are common in patients with HIV infection (see Chapter 33).16 These manifestations cause significant morbidity and can provide valuable diag­ nostic and prognostic information. Frequently, the first and most common HIV-associated infection of the mouth is can­ didiasis. The history and physical findings usually establish the diagnosis. The presence of spores, pseudohyphae, or hyphal forms on a smear (potassium hydroxide, periodic acid–Schiff, Papanicolaou), culture, or biopsy confirms the diagnosis. Oral candidiasis in HIV should be treated sys­ temically. Systemic therapy involves the use of oral azole preparations (fluconazole or itraconazole). Amphotericin B given intravenously is also effective, but usually not neces­ sary. Treatment for one to two weeks is usually effective, even in the late stages of HIV infection. Frequent recur­ rences may require chronic or repeated treatment. The like­ lihood of clinical relapse is dependent on the degree of immunosuppression and the duration of therapy. As adjunc­ tive measures, mouth rinses with chlorhexidine gluconate (Peridex), Listerine, or hydrogen peroxide–saline may be of some benefit.17,18 Periodontal disease and alveolar bone loss occur com­ monly in the general population, as well as in individuals with or at risk for AIDS. The association of cardiovascular disease, diabetes mellitus, low birth weight, other systemic diseases, and periodontal disease is a growing field of investigation. Hairy leukoplakia (oral hairy leukoplakia, HL) appears as corrugated white lesions on the lateral borders of the tongue (Fig. 22-3). HL is usually asymptomatic and may be an early sign of HIV infection. The epithelium in patients with HL is infected with Epstein-Barr virus.19 The severity of HL does not correlate with the stage of HIV disease. However, the presence of HL in an HIV-infected person has prognostic implication. Analysis of 198 cases of HL has demonstrated that the median time to onset of AIDS is 24 months, and the median time to death in the era prior to highly active antiretroviral therapy (HAART) was 41 months.20 Other mucosal white lesions, such as oral leukoplakia (Fig. 22-4), can resemble HL lesions; biopsy confirmation should be considered if the diagnosis of HL is in doubt. HL may be confused with candidiasis (which coexists in about half of cases). A prudent first step in management is the

Figure 22-3.  Hairy leukoplakia involving the tongue in a patient with AIDS. (Courtesy of Dr. Sol Silverman, Jr, DDS, and Dr. Victor Newcomer.)

administration of anticandidal therapy. The suspicion of HL justifies a discussion of its implications and suggestion of HIV testing. Although HL occurs predominantly in HIV-infected homosexual and bisexual men, it also has been found in renal and other organ transplant recipients. Because HL is usually asymptomatic, treatment is elective. HL responds to oral acyclovir, topical retinoic acid, and podophyllum. When treatment is discontinued, HL usually returns. Kaposi’s sarcoma (KS) is a common consequence of HIV infection and is associated with human herpesvirus 8. A significant decline in the incidence of KS occurred during 1996 and 1997, which corresponded to the introduction of HAART. Although KS is usually found on the skin, more than half of patients also have intraoral lesions.21 The first sign of KS occurred in the mouth in 22% of patients and, in another 45%, KS occurred in the mouth and skin simul­ taneously.20,21 The cutaneous lesions of KS appear as asymp­ tomatic red to purple, oval macules that develop into papules, plaques, or nodules. They rarely ulcerate, except on the lower extremities and genitalia. Edema often accom­ panies cutaneous lesions, especially on the lower extremi­ ties or on the face. Oral lesions may vary in appearance from minimal, asymptomatic, flat, purple or red macules to large nodules. The hard palate is the most frequent location, followed by the gingiva and tongue (Fig. 22-5).22 The dif­ ferential diagnosis of KS includes purpura, hemangioma, coagulation defects, and bacillary angiomatosis. Diagnosis is established by biopsy. Treatment approaches are mainly palliative but include topical alitretinoin (9-cis-retinoic acid) gel, imiquimod, radiation therapy, chemotherapy

Figure 22-4.  Oral leukoplakia and associated squamous carcinoma.

Figure 22-5.  Kaposi’s sarcoma involving the palate. (Courtesy of Dr. Sol Silverman, Jr, DDS, and Dr. Victor Newcomer.)

355

356

Section IV  Topics Involving Multiple Organs (including intralesional injections), and surgery.23 Patients with cutaneous KS may have asymptomatic visceral lesions (see Chapter 33). Lymphoma may involve the oral cavity and skin in HIV patients and, rarely, may be the first sign of HIV or lymphoma. Other conditions associated with HIV infection include the following: extensive oral, genital, or cutaneous warts; recurrent aphthae; chronic mucocutaneous herpes simplex virus (HSV) infections; lymphocytic infiltrates of major salivary glands, leading to secondary Sjögren’s syndrome; drug reactions, including drug-induced Stevens-Johnson syndrome; Bartonella infections (bacillary angiomatosis and its associated peliosis hepatis); premature and progressive periodontal disease; and acute necrotizing ulcerative gingivitis.24

ULCERATIVE DISEASES

Aphthous ulcers (canker sores, recurrent aphthous ulcers [RAUs]) are painful shallow ulcers, often covered with a grayish-white or yellow exudate and surrounded by an erythematous margin. They appear almost exclusively on unkeratinized oral mucosal surfaces (Table 22-1). Rarely, RAUs may occur in the esophagus, upper and lower GI tracts, and anorectal epithelium. RAUs develop at some time in 25% of individuals in the general population and recur at irregular intervals. Three clinical forms of aphthous ulcers are recognized—minor aphthae (most common), major aphthae (less common), and herpetiform aphthae (least common). Minor aphthae typically are smaller than 5 mm and heal in one to three weeks (Fig. 22-6A). Major aphthae may exceed 6 mm (see Fig. 22-6B) and require months to heal, often leaving scars. Herpetiform aphthae are 1 to 3 mm in diameter, occur in clusters of 10 to hundreds of ulcers, and resolve quickly.42 The cause of RAU is thought to be multifactorial, with precipitating factors including the following: (1) immuno­ logic abnormalities such as celiac disease and increased allergen presentation caused by decreased constitutive oral barriers (putatively from sodium lauryl sulfate use in dental products); (2) chronic trauma, such as from ill-fitting dentures; (3) vitamin or mineral deficiencies, such as iron, folate, and vitamin B12; (4) genetic predisposition; (5) stress and anxiety; (6) allergies to food or medication, such as to cyclooxygenase 2 (COX-2) inhibitors or sertraline; and (7) xerostomia.25 Helicobacter pylori infection may be associ­ ated with RAUs as eradication of H. pylori appears to be associated with a reduction of recurrences, as well as a decrease in the number of ulcers and days of symptoms.26 Morphologically identical aphthous lesions may be seen in inflammatory bowel disease (IBD; see later) and Behçet’s syndrome. The workup for recurrent aphthous ulcers includes a complete blood count (CBC), erythrocyte sedi­

mentation rate, serum iron, ferritin, folate, and B12 levels, potassium hydroxide (KOH) stain, Tzanck smear, viral culture, biopsy of coexisting skin lesions to exclude HSV, and colonoscopy to address the possibility of inflammatory bowel disease (IBD). Histologically, lesional tissue shows ulcerated mucosa with chronic mixed inflammatory cells. Management of aphthous ulcers includes palliative and curative measures. First, vitamin deficiencies, if found, should be treated. Otherwise, patients should be advised to use multivitamins with iron and avoid crusty, salty, or spicy foods to minimize irritation of oral lesions. Soft tooth­ brushes, repair of dentition, and other measures to avoid unnecessary oral trauma should be instituted. Analgesics and topical anesthetics such as 2% viscous lidocaine may be helpful, along with bismuth subsalicylate (Kaopectate) and sucralfate to protect lesions and accelerate healing. Aphthous ulcers can be treated effectively with a potent topical glucocorticoid, such as fluocinonide (Lidex) or clo­ betasol (Temovate) gel or ointment. Second-line therapy includes colchicine, 0.6 mg three times daily; cimetidine, 400 to 800 mg/day; azathioprine, 50 mg/day; or thalido­ mide, 200 mg/day (U.S. Food And Drug Administration [FDA]–approved for HIV patients). Short courses of sys­ temic prednisone (20 to 60 mg/day) are reproducibly effec­ tive when more conservative approaches are not satisfactory. An elimination diet may be helpful for patients with allergic reactions to certain foods or medications, including a trial of sodium lauryl sulfate–free dental products.27 A glutenfree diet is recommended for patients with gluten-sensitive enteropathy (see Chapter 104). Infection with HSV commonly produces painful vesicles and ulcers on the genitalia, eyes, lips, mouth, and skin. Primary herpetic gingivostomatitis is caused by HSV type 1 (or, occasionally, type 2). Primary infection occurs in up to 90% of the population before puberty. The illness is often mild and mistaken for a routine upper respiratory tract infection; it may include varying degrees of fever, malaise, and adenopathy, together with oral and gingival ulcers.

A

Table 22-1  Distinctions between Aphthous and Herpetic Oral Ulcers CONDITION

MUCOSA

LOCATION

Aphthous ulcers

Unkeratinized

Herpes simplex virus ulcers

Keratinized

Lateral tongue, floor of the mouth, labial and buccal mucosa, soft palate, pharynx Gingiva, hard palate, dorsal tongue

B Figure 22-6.  A, Multiple minor aphthous ulcers. B, A major aphthous ulcer.

Chapter 22  Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease Lesions may appear on the lips. They generally heal in one to two weeks. Management is palliative, but acyclovir, 400 mg three times a day, may shorten the course and reduce severity. Secondary bacterial infection is common and can be treated topically. Recurrent orolabial herpes simplex is caused by reactiva­ tion of HSV that has been dormant in regional ganglia, with no associated increase in HSV antibody titers. Episodes may be precipitated by fever, sunlight, and physical or emo­ tional stress. Recurrences vary in frequency and severity. Typically, the lesions involve the lips (cold sores) and are preceded by several hours of prodromal symptoms such as burning sensation, tingling, or pruritus. Vesicles then appear but soon rupture, leaving small, irregular, painful ulcers. Coalescence of ulcers, crusting, and weeping of lesions are common. Intraoral recurrent herpetic ulcers occur on kera­ tinized mucosa (i.e., hard palate or gingiva; see Table 22-1). They appear as shallow, irregular, small ulcerations and may coalesce. Labial and oral herpetic ulcers normally heal in less than two weeks. Recurrent HSV is the most common cause of recurrent erythema multiforme. In immunocompromised patients, HSV can affect any mucocutaneous surface and can appear as large, irregular, pseudomembrane-covered ulcers. This is especially true in HIV-infected persons, in whom all perineal and orolabial ulcerations should be considered manifestations of HSV until proven otherwise (see Chapters 33 and 125). Care should be taken to avoid ocular autoinoculation. Herpes simplex is usually diagnosed from the history and clinical findings. A history of a prodrome or of vesicles, the site of lesions, and the reappearance of lesions in the same location help differentiate herpes from other ulcerative dis­ orders. A cytologic smear (Tzanck) showing multinucleate giant cells is suggestive, although viral cultures and mono­ clonal antibody staining of smears are more sensitive and specific tests for diagnosing HSV infection. Topical acyclo­ vir is of little benefit in recurrent labial herpes and is of limited benefit in recurrent genital HSV. Systemic acyclovir is regularly used for treatment of primary or recurrent attacks in immunosuppressed patients (2 g orally in divided doses, or 5 mg/kg intravenously three times daily until lesions heal). Famciclovir, 125 mg twice daily, or valacyclo­ vir, 500 mg twice daily, are also available in the United States. Oral treatment should optimally begin within the first few hours of the prodrome. Suppression of recurrences may be accomplished with acyclovir, 200 mg orally three times daily or 400 mg twice daily. Acyclovir is used for the prevention of recurrent oral and genital herpes associ­ ated with bone marrow transplantation (see Chapter 34). Antivirals are also used to prevent recurrent herpes infections in other immunocompromised patients such as those with leukemia or HIV infection, or after solid organ transplantation. Herpes zoster is caused by a reactivation of the varicella virus. The oral lesions can resemble the ulcers of aphthous stomatitis, except for the following features: the ulcers are unilateral; lip and/or skin lesions may coexist; and the onset is sudden, acutely painful, and often associated with fever. High dosages of acyclovir (4 g/day orally), famciclovir, 500 mg every 8 hours, or valacyclovir, 1 g every 8 hours, at the onset may be helpful in accelerating healing. Cytomegalovirus (CMV) affects 40% to 80% of adults, as evidenced by serologic studies. However, symptomatic disease occurs mainly after solid organ or bone marrow transplantation or in HIV-infected persons. Other than reti­ nitis, mucosal ulcers are the main consequence of CMV infection. Skin is rarely affected. Characteristic features of biopsy specimens include intranuclear and intracytoplas­

mic inclusions. Treatment includes use of ganciclovir, val­ ganciclovir, cidofovir, or foscarnet and treatment of any underlying immunodeficiency. Acute necrotizing ulcerative gingivitis is an acute inflam­ matory and necrotic infection affecting the interdental papillae, and commonly affects healthy young adults. Treat­ ment consists of surgical débridement, oral rinses, and sys­ temic antibiotics. Cutaneous amebiasis is a rare complication of amebic dysentery. Undermined ulcers of the perineum and genita­ lia may result from direct inoculation from the colon and rectum to the anus after contact with contaminated stool or from external inoculation after intercourse or hand (scratch­ ing) contact. Dissemination may result from invasion of the colonic mucosa, with hematogenous spread to the liver and then to the lungs or chest wall. In long-standing disease, the ulcer borders may become vegetative and proliferative and resemble a squamous cell carcinoma. Because primary car­ cinomas of the rectum, colon, and cervix may be second­ arily colonized with Entamoeba histolytica, it is important that either possibility be ruled out.

VESICULOBULLOUS DISEASES Pemphigoid is a general term for heterogeneous blistering disorders characterized by bullae and ulcers affecting the mucosa of the oral cavity, pharynx, esophagus, anus, con­ junctiva, and skin. Oral findings appear as highly inflamed (erythematous) mucosa on the buccal mucosa and gingival mucosa. Two types of pemphigoid have been identified, bullous pemphigoid and cicatricial (mucous membrane) pemphigoid. Patients with bullous pemphigoid typically have skin lesions, and about one third also have mucous membrane lesions. All patients with cicatricial pemphigoid have mucosal lesions, and about one third also have skin lesions (tense bullae). Ocular symblepharon (i.e., adhesion between the tarsal and bulbar conjunctiva) commonly occurs with cicatricial pemphigoid. Potentially fatal upper GI bleeds because of esophageal involvement by pem­ phigoid has been reported.28 Immunofluorescent staining of involved mucosa and skin shows linear deposition of antibody and complement in the basement membrane zone. Serum antibodies against 230- and 180-kd antigens located at the squamous epithelial basement membrane have been documented. Patients with serum IgG and IgA antibodies are more likely to respond to systemic medications. Treatment ranges from low-dose to high-dose prednisone for patients without contraindications to glucocorticoid use. Alternative therapies for patients with contraindications or systemic toxicities to glucocorticoids include dapsone, tetracycline and nicotinamide in combination, azathio­ prine, chlorambucil, plasma exchange, intravenous immune globulin (IVIG), cyclosporine, cyclophosphamide given orally or in a pulse-dosing format, methotrexate, and infliximab.29 Pemphigus vulgaris differs from pemphigoid in that bullous skin lesions, which are often flaccid, are more severe (and can be life-threatening if untreated) and oral involvement can be extensive. Mucosal involvement can cause poor nutrition and severe pain. Half of patients with pemphigus vulgaris present with oral lesions, and oral lesions occur in almost 100% of patients during the illness. Autoantibodies mediate a loss of cell to cell adhesion. IgG antibodies and complement on the surface of squamous epithelial cells are diagnostic. Indirect immunofluorescence detects circulating IgG antibodies in most patients with

357

358

Section IV  Topics Involving Multiple Organs

Figure 22-7.  Characteristic lesions resulting from skin fragility caused by epidermolysis bullosa dystrophica.

pemphigus vulgaris. Treatment consists of various regimens of topical or systemic prednisone, sometimes supplemented with cytotoxic or immunosuppressive drugs. Paraneoplastic pemphigus shares features of pemphigus vulgaris and erythema multiforme.30 It is associated with various malignancies, including GI malignancies, lympho­ mas and leukemias, thymomas, and soft tissue sarcomas. Five features characterize paraneoplastic pemphigus: (1) painful mucosal erosions and a polymorphous skin erup­ tion; (2) intraepidermal acantholysis, keratinocyte necrosis, and vacuolar interface reaction; (3) deposition of IgG and C3 intercellularly and along the epidermal basement mem­ brane zone; (4) serum autoantibodies that bind to skin and mucosa epithelium in a pattern characteristic of pemphigus, as well as binding to simple, columnar, and transitional epithelia; and (5) immunoprecipitation of a complex of four proteins (250, 230, 210, and 190 kd) from keratinocytes by the autoantibodies. The prognosis of paraneoplastic pem­ phigus depends on the associated underlying malignancy, and successful treatment is predicated on successful elimi­ nation of the underlying malignancy. Epidermolysis bullosa (EB) is a heterogeneous group of rare inherited disorders of skin fragility (Fig. 22-7). They are characterized by the formation of blisters with minimal trauma and are divided into dystrophic (scarring), junc­ tional, and simplex forms. Oral erosions, premature caries, and gingival involvement, as well as GI disease, are common in the dystrophic form but also occur in some patients with the junctional form. In addition to oral erosions, esophageal strictures are the most common GI complication in dystro­ phic EB.31 They may be narrow or broad and most com­ monly occur in the upper third of the esophagus, but also may be found in the lower third. The esophageal strictures are probably induced by repeated trauma from food and/or refluxed gastric contents; therefore, strict adherence to a soft food diet remains a mainstay of management. Although dilations with bougienage historically have been shunned because of an unacceptable risk of increasing esophageal stenosis over the long term, evidence supports the use of balloon dilation as a safe and efficacious method of palliating esophageal strictures without this risk. Surgical excision, feeding gastrostomy, and colonic interpositioning have been effectively used in dystrophic EB patients with severe esophageal strictures. Esophageal webs in the post­

cricoid area have also been described. Anal stenosis and constipation (with or without stenosis) are frequent in patients with dystrophic EB. Junctional EB has been uniquely associated with pyloric atresia. Anemia and growth retardation frequently develop in patients with severe dystrophic and junctional EB, partly because of GI and oral complications. Patients with clinical lesions identical to the dystrophic forms of EB but with no family history and an adult onset have been identified; their condition is called acquired EB or EB acquisita (EBA). EBA, like pemphigus and pem­ phigoid, is an autoimmune disease. The autoantibodies in EBA are directed against type VII collagen.32 The diagnosis of EBA is established by routine histology and direct immu­ nofluorescence examination of skin biopsy specimens. Patients may have significant mucosal involvement, like patients with cicatricial pemphigoid, especially oral and esophageal disease. Coexistent Crohn’s disease has been reported in a number of patients with EBA. Treatment is with immunosuppressive agents. Erythema multiforme (EM) is an acute, benign mucocuta­ neous eruption associated with underlying infections (especially HSV). It is often preceded or accompanied by low-grade fever, malaise, and symptoms suggesting an upper respiratory tract infection. The eruption consists of alternating pink and red target lesions on the elbows, knees, palms, and soles and of shallow, broad oral erosions. Patients with EM may only have oral involvement. Variable degrees of nonspecific erythema are found, with or without ulcers. Crusting, hemorrhagic, and moist lip ulcers may be present. Severe oral and pharyngeal pain, secondary bacte­ rial and fungal infections, and bleeding are common com­ plications. The diagnosis is made by clinical characteristics, ruling out other specifically diagnosable diseases, and by response to treatment. The biopsy reveals a nonspecific interface reaction. Oral EM can be self-limited or chronic, and often the inciting process goes unidentified. Manage­ ment includes palliative measures and elimination of any offending agent. Often, glucocorticoids and/or other immu­ nosuppressive drugs are needed. Recurrences and flares have variable patterns. Herpes-associated erythema multi­ forme lesions are treated with episodic or suppressive antiviral therapy with acyclovir, valacyclovir, famciclovir, or foscarnet.33 Stevens-Johnson syndrome (between 10% and 30% skin sloughing) and toxic epidermal necrolysis (more than 30% skin sloughing) are diagnosed when severe, acute targetoid lesions and skin sloughing occur in association with eye, skin, and mucous membrane involvement. Diffuse oral and pharyngeal ulceration may prevent oral intake. At endos­ copy, the esophagus may show diffuse erythema, friability, and whitish plaques that can be mistaken for candidiasis. Diffuse gastric and duodenal erythema and friability may be present without esophageal involvement. The colonoscopic appearance may resemble severe ulcerative or pseudo­membranous colitis. However, colonic biopsies show extensive necrosis and lymphocytic infiltration, without crypt abscesses or neutrophils. This pattern is reminiscent of graft-versus-host disease (see Chapter 34). The mucosa of large portions of bowel may slough in Stevens-Johnson syndrome, accounting for reports of hematemesis, melena, and intestinal perforation. Treatment largely consists of discontinuation of offending pharmaceu­ tical agents (often anticonvulsants), hospital admission to a burn unit, and supportive care by a multiteam approach. Some evidence suggests that IVIG is beneficial when administered early in the course and that glucocorticoids actually have a negative effect on outcomes.34

Chapter 22  Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease

Figure 22-8.  The erosive form of oral lichen planus involving the buccal mucosa. Note the lace-like keratoses, erythema, and ulceration.

LICHEN PLANUS Lichen planus is a common, chronic inflammatory disorder involving the mucosa and skin. The disease usually begins in adulthood, and two thirds of patients are women. Oral lesions appear as white, lace-like and/or punctate patterns on any mucosal surface (Fig. 22-8). Mucosal erythema or ulceration is common. The lesions are small, flat-topped, pruritic, violaceous papules. Oral lesions can be asymptom­ atic or severe oral pain may develop. Topical and/or sys­ temic glucocorticoids are effective in decreasing the signs and symptoms in almost all cases of oral and cutaneous lichen planus. In rare refractory cases, systemic retinoids are necessary. Topical tacrolimus is an effective steroidsparing treatment alternative. Esophageal lichen planus may present with progressive dysphagia and odynophagia, upper GI bleeding, strictures and squamous cell carcinoma. The endoscopic findings include erythema, ulcers, proximal esophageal webs, and erosions throughout the esophagus. An increased prevalence of chronic liver disease, including chronic active hepatitis C and primary biliary cirrhosis, has been reported in patients with lichen planus. Oral lichen planus may be associated with an increased risk of squa­ mous cell carcinoma arising in areas of atrophy or erosion, regardless of treatment.35,36

CUTANEOUS MANIFESTATIONS OF INTESTINAL DISEASE Both Crohn’s disease and ulcerative colitis may be accom­ panied by cutaneous manifestations (see Chapters 111 and 112). Skin lesions are more common (up to 44%) and often more specific in Crohn’s disease than in ulcerative colitis. It is rare for cutaneous involvement by Crohn’s disease to appear before symptomatic bowel disease. The most common cutaneous complication of Crohn’s disease is granulomatous inflammation of the perianal or perifistular skin, which occurs by direct extension from underlying diseased bowel. Metastatic Crohn’s disease refers to rare ulcerative lesions, plaques, or nodules that occur at sites distant from the bowel. Such lesions favor intertriginous areas, such as the retroauricular and inframammary regions. On histologic study, local cutaneous extension and meta­ static Crohn’s disease show sarcoid-like granulomatous inflammation, and both occur with greater frequency in patients with colonic involvement by Crohn’s disease.37 Oral manifestations of Crohn’s disease occur in 4% to 14% of patients and include aphthae (see Fig. 22-6), lip

Figure 22-9.  Pyostomatitis vegetans in a patient with ulcerative colitis. A biopsy specimen revealed microabscesses.

fissures, cobblestone plaques, cheilitis, mucosal tags, and perioral erythema. Patients may also complain of metallic dysgeusia. Aphthosis occurs in approximately 5% of patients with Crohn’s disease, and the lesions are indis­ tinguishable, clinically and histologically, from typical aphthae. Aphthosis and perianal-perifistular ulcerations are not seen in ulcerative colitis. Granulomatous cheilitis is a rare condition with recurrent lip swelling that leads to enlargement and firmness of the lips. A biopsy shows noncaseating granulomas. In rare cases associated with Crohn’s disease,38 this condition may be a component of Melkersson-Rosenthal syndrome (scrotal tongue, lip swelling, with or without facial palsy and migraine), or may be idiopathic. Pyostomatitis vegetans (Fig. 22-9), and its cutaneous counterpart, pyoderma vegetans, is characterized by pus­ tules, erosions, and vegetations involving the labial mucosa of the upper and lower lips, buccal mucosa, and gingival mucosa, as well as the skin of the axillae, genitalia, trunk, and scalp. Both pyostomatitis vegetans and pyoderma veg­ etans are specific markers of IBD (Crohn’s and ulcerative colitis) and may precede the GI symptoms by months to years. Histologically, intraepithelial and subepithelial eosinophilic miliary abscesses are characteristic. Superficial pustules coat the friable, erythematous, and eroded mucosa of the oral cavity, least commonly the floor of the mouth and tongue. Symptoms may be severe or minimal. Eosino­ philia and anemia are common. Diagnosis is made from biopsy findings, and treatment is with topical or systemic glucocorticoids, dapsone, or sulfasalazine. Erythema nodosum is a common inflammatory disorder of the subcutaneous fat, with a marked predilection for women. Lesions characteristically appear as 1 mm or larger, shiny, tender, deep red nodules on the anterior shins. The pathogenesis is unknown. The causes of erythema nodosum are infections, especially streptococcal, systemic fungal, and tuberculous, medications (especially oral contracep­ tives), and leukemias. Erythema nodosum develops in 7% of patients with Crohn’s disease and 4% of patients with ulcerative colitis. In addition, GI infections with Yersinia enterocolitica, Shigella flexneri, and Campylobacter jejuni have been associated with erythema nodosum. Treatment of the underlying disease, strict bed rest, and elevation of the legs, as well as the use of anti-inflammatory drugs or potassium iodide, are effective. Pyoderma gangrenosum is a noninfectious ulcerative cutaneous disorder of unknown pathogenesis (Fig. 22-10). The classic lesion is a tender or painful ulcer with an elevated, dusky purple border that is widely undermined. One or multiple lesions may occur. Lesions begin as small

359

360

Section IV  Topics Involving Multiple Organs

Figure 22-10.  Pyoderma gangrenosum in a patient with ulcerative colitis.

papulopustules that break down very rapidly. Pathergy, the appearance of new ulcers at sites of minor trauma or surgery, is often present. The diagnosis is one of exclusion, in that infectious and other causes of ulceration, including factitia, must be ruled out. Most cases of pyoderma gan­ grenosum occur in patients with no underlying disease. Pyoderma gangrenosum develops in approximately 5% of patients with ulcerative colitis and 1% of patients with Crohn’s disease. The bowel disease may be subclinical when the skin lesions appear and therefore bowel evalua­ tion, especially of the rectum and distal colon, is essential in cases of pyoderma gangrenosum. If the disorder is associated with underlying bowel disease, therapy of the bowel disease may lead to improvement of the skin lesions. The usual management of pyoderma gangrenosum includes local wound care, high-dose systemic glucocor­ ticoids, or steroid-sparing immunosuppressive agents, such as azathioprine, mycophenolate mofetil, methotrexate, and cyclosporine.37 Bowel bypass syndrome (bowel-associated dermatosisarthritis syndrome) occurs in up to 50% of patients who have undergone an intestinal bypass procedure for obesity (an operation not performed currently) and, rarely, in patients with other forms of bowel disease. In addition to polyarticular arthritis, characteristic skin lesions develop in 80% of patients, usually erythematous macules evolving over several days to form vesiculopustular lesions on a purpuric base. A skin biopsy shows a perivascular and diffuse infiltrate of neutrophils. The pathogenesis of this disorder seems to be overgrowth of bacteria in a blind loop or pouch of bowel, which leads to the development of anti­ bodies against bacterial peptidoglycans. Immune complexes are formed, sometimes in the form of cryoproteins that appear to be responsible for the symptom complex. Chronic antibiotic therapy, anti-inflammatory drugs (including glu­ cocorticoids), and correction of the bowel bypass all are variably beneficial.

VASCULAR AND CONNECTIVE TISSUE DISORDERS Immune complex vasculitis of small vessels (leukocytoclas­ tic vasculitis) appears on the skin of dependent sites as crops of palpable purpura and is mediated by deposition of immune complexes in postcapillary venules (Fig. 22-11; see Chapter 36). Although GI involvement can occur in any case of small vessel vasculitis, it occurs in 50% to 75% of patients

Figure 22-11.  Cryoglobulinemic vasculitis caused by a drug eruption. This type of vasculitis may also be seen in patients with chronic hepatitis C, although generally not as severe as shown here.

Figure 22-12.  Skin lesions of Henoch-Schönlein purpura.

with Henoch-Schönlein purpura (Fig. 22-12).39 Vasculitic hemorrhage, bowel wall edema, and intussusception affect the jejunum and ileum most commonly. Direct immuno­ fluorescence of early skin lesions reveals deposits of IgG in most cases of small vessel vasculitis and deposits of IgA in Henoch-Schönlein purpura. Polyarteritis nodosa, sometimes associated with hepatitis B, is a vasculitis of the medium-sized and small arteries. Arterial lesions of the abdominal viscera can lead to infarc­ tion of the gut, liver, and gallbladder and to ischemic pan­ creatic necrosis, as well as to GI infarcts or perforation. Involvement of the appendix, gallbladder, or pancreas can simulate acute appendicitis, cholecystitis, or pancreatitis. Cutaneous involvement occurs in 25% of cases, most typi­ cally manifesting as nodules, 5 to 10 mm in size, distributed along the course of the superficial arteries. A mottled livedo vascular pattern is also frequently seen. Malignant atrophic papulosis (Degos’ disease, KöhlmeierDegos syndrome, progressive arterial mesenterial vascular occlusive disease, or disseminated intestinal and cutaneous thromboangiitis) is a rare multisystem vasculopathy dis­ order that is occasionally familial. Cutaneous lesions are the initial manifestations, appearing most commonly in early adulthood. They appear as crops of asymptomatic, pink, 2- to 5-mm papules that rapidly become umbilicated

Chapter 22  Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease

Figure 22-13.  Malignant atrophic papulosis (Degos’ disease) with cutaneous lesions of different stages.

Figure 22-14.  Flat telangiectasias of the lips and vermillion border in a patient with hereditary hemorrhagic telangiectasia (Osler-Weber-Rendu disease).

and develop a characteristic atrophic, depressed, porcelain white center (Fig. 22-13). These lesions represent cutaneous infarcts. Similar infarcts are seen in the small bowel in almost all cases. Although GI involvement may initially be asymptomatic or nonspecific, an acute abdominal catastro­ phe eventually occurs, often necessitating laparoscopy or laparotomy. Perforation of the intestine is usually found, along with multiple white, yellowish, or rose-colored flat or slightly depressed patches below an intact serosa, usually in the small intestine. The intestinal disease is recurrent and eventually is often fatal. Cerebral and peripheral nerve infarcts develop in approximately 20% of patients, leading to neurologic complications that can include hemiparesis, aphasia, cranial neuropathies, monoplegia, sensory distur­ bances, and seizures. Histologic study reveals that the infarctive lesions of the skin, gut, and nervous system are consequences of noninflammatory thromboses. The patho­ genesis of Degos’ disease is unknown, but identical lesions have been reported in systemic lupus erythematosus and in a patient without systemic lupus erythematosus with anticardiolipin antibodies and a lupus anticoagulant. Treatment has been attempted with antithrombotic agents such as aspirin, ticlopidine, and dipyridamole, with limited success. Hereditary hemorrhagic telangiectasia (HHT), or OslerWeber-Rendu disease, is a group of autosomal dominant disorders characterized by vascular lesions including telan­ giectases, arteriovenous malformations, and aneurysms of the skin and internal organs (lung, brain, and GI tract). Epi­ staxis and GI hemorrhage are the most common complica­ tions (see Chapter 36); the incidence of frequent epistaxis ranges from 81% to 96%. The skin lesions are 1- to 3-mm macular telangiectases of the face, lips, tongue, conjunctiva, fingers, chest, and feet (Fig. 22-14). Skin lesions appear later

Figure 22-15.  Fingertip lesion in a patient with the blue rubber bleb nevus syndrome.

than the epistaxis, usually in the second or third decade of life. In the fifth to sixth decades, recurrent upper and lower GI hemorrhage may occur. Vascular malformations have been reported in the GI tract (46%), liver (26%), lungs (14%), central nervous system (12%), genitourinary tract (1.9%), and almost every other organ system in the body. Management of the GI bleeding may be difficult, but the use of bipolar electrocoagulation or laser techniques has been beneficial (see Chapter 19). Associated von Willebrand factor deficiency may be present, and therapy with desmo­ pressin has been successful in treating massive GI bleeding. There is presently no treatment to prevent the development of telangiectatic lesions in patients with HHT. Chronic therapy with estrogen and progesterone may reduce bleed­ ing from GI telangiectases. Skin and oral lesions similar to those seen in HHT are found in some patients with the CREST (calcinosis, Raynaud’s phenomenon, esophageal dysfunction, sclerodactyly, and telangiectasia) variant of scleroderma. The extent of cutaneous sclerosis may be limited in patients with CREST, so differentiation by phys­ ical examination may be difficult. Epistaxis is uncommon in patients with CREST and almost universal in patients with HHT. In addition, patients with CREST have anticen­ tromere antibodies in their serum that are not found in patients with HHT. Blue rubber bleb nevus syndrome is a rare disorder of the skin and GI tract composed of a constellation of multiple cutaneous and GI venous malformations. Most cases are sporadic. In affected patients, blue, subcutaneous, com­ pressible nodules develop on the skin (Fig. 22-15). GI vas­ cular malformations are common, especially in the small intestine or colon, and bleeding is an almost universal feature. Acute GI hemorrhage, intussusception, volvulus, bowel infarction, and rectal prolapse have been described. Treatment is primarily surgical or with photocoagulation. Amyloidosis commonly has prominent cutaneous and oral manifestations. Waxy papules around the eyes, nose, and central face—as well as purpura involving the face, neck, and upper eyelids—are frequently noted. If a waxy papule is pinched, hemorrhage will ensue (pinch purpura). Orbital purpura after endoscopy, vomiting, or coughing is almost diagnostic. Macroglossia, increased tongue firmness, enlarged submandibular structures, and lingual indenta­ tions from the teeth occur in 20% to 50% of patients. The macroglossia may interfere with eating and closing the mouth and may cause airway obstruction, especially in the reclining position. The tongue may be enlarged and highly vascular, resulting in bleeding. Recurrent hemor­ rhagic bullae in the mouth are common. Patients may have

361

362

Section IV  Topics Involving Multiple Organs

Figure 22-16.  Characteristic “plucked chicken skin” appearance in a patient with pseudoxanthoma elasticum. (Courtesy of Dr. Victor Newcomer.)

carpal tunnel syndrome, edema, the shoulder pad sign (amyloid deposits in soft tissues around shoulders), GI bleeding (see Chapter 35), peripheral neuropathies, rheuma­ toid arthritis–like deposits in small joints, and cardiac involvement. Congestive heart failure or arrhythmias account for death in 40% of patients with systemic amyloi­ dosis. Diagnosis of amyloidosis can be made by subcutane­ ous fat aspiration or by bone marrow, rectal, skin, or tongue biopsy. Pseudoxanthoma elasticum is a rare disorder character­ ized by aberrant calcification of mature elastic tissue. Skin lesions are usually the initial manifestation, appearing in the second decade as yellow to orange papules (“plucked chicken skin”) on the lateral neck (Fig. 22-16). Skin lesions may progress caudally, involving other flexural areas (e.g., axilla, groin, antecubital and popliteal fossae). Calcification of the elastic tissue of arteries leads to the major complica­ tions—retinal bleeding, intermittent claudication, prema­ ture coronary artery disease, and GI bleeding. From 8% to 13% of patients experience GI bleeding, which is usually from the stomach, and often no specific bleeding point is found. As opposed to the other complications of pseudo­ xanthoma elasticum just noted, GI bleeding tends to occur in younger patients (average age, 26 years), often occurs during pregnancy, and may be recurrent. Skin lesions may not be visible at the time of bleeding. Because apparently normal flexural or scar skin may yield diagnostic findings, a blind skin biopsy may be indicated in a young person with GI bleeding with no other explanation. Lesions identical to those seen on the skin may also be present on the lower lip and the rectal mucosa. Neurofibromatosis type 1 (NF1, von Recklinghausen’s disease) is defined by its cutaneous manifestations of six or more café au lait spots (each with a diameter more than 5 mm in prepubertal persons and more than 15 mm in post­ pubertal persons), multiple soft papules (neurofibromas; Fig. 22-17), or a single plexiform neurofibroma, and freck­ ling of the axillae or inguinal areas. GI involvement occurs in 10% to 15% of patients with NF1. Intestinal neurofibro­ mas may arise at any level of the GI tract, although small intestinal involvement is most common. These tumors are generally submucosal but may extend to the serosa. Dense growths known as plexiform neurofibromatosis of the mesentery or retroperitoneal space may lead to arterial compression or nerve injury. Other tumors may occur in neurofibromatosis. There is an increased incidence of pheo­ chromocytoma, with or without the multiple endocrine neoplasia type IIB syndrome.40 Duodenal and ampullary

Figure 22-17.  Neurofibromatosis.

Figure 22-18.  An adult with urticaria pigmentosa. Reddish-brown frecklelike lesions are characteristic of the adult form of this disease.

carcinoid tumors (sometimes producing obstructive jaun­ dice; see Chapter 31), malignant schwannomas, sarcomas, and pancreatic adenocarcinomas are seen with increased frequency. Clinical manifestations include abdominal pain, constipation, anemia, melena, and an abdominal mass. Serious complications that have been reported include intestinal or biliary obstruction, ischemic bowel, perfora­ tion, and intussusception. Involvement of the myenteric plexus has resulted in megacolon. Mastocytosis is characterized by mast cell infiltration of the bone marrow, skin, liver, spleen, lymph nodes, and GI tract. It occurs in adult and pediatric patients (see Chapter 35). In children, the most common lesions consist of a large red to brown plaque (solitary mastocytoma), multiple red to brown papules or plaques (urticaria pigmentosa), or diffuse cutaneous involvement, with or without flushing or blister­ ing. In adult patients, most have urticaria pigmentosa–type lesions (Fig. 22-18), sometimes with prominent telangiecta­ sia. Lesions often are on the trunk. The most common GI complaint is dyspepsia and often peptic ulcer disease caused by histamine-induced gastric hypersecretion (see Chapter 49). Diarrhea and abdominal pain are also common problems and can be accompanied by malabsorption.41 In children, the lesions usually involute spontaneously, and systemic disease is uncommon. In adults, cutaneous lesions

Chapter 22  Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease may resolve as well, but without improvement in systemic symptoms. In the rare pediatric case with a solitary masto­ cytoma and significant systemic symptoms, excision of the skin lesion may resolve the systemic complications. Extra­ cutaneous involvement should be considered for adult patients with cutaneous mastocytosis because management of symptoms can easily be achieved. Porphyria cutanea tarda has a well-documented associa­ tion with hepatitis C and is described in more detail later. However, in almost all the porphyrin disorders, manifesta­ tions include some cutaneous and gastrointestinal symp­ toms. Skin vesicles, erosions, and bullae with hemolytic anemia and splenomegaly characterize erythropoietic por­ phyria. Erythropoietic protoporphyria patients demonstrate edematous plaques, erythema, and purpura, often with cholelithiasis and occasionally with hepatic failure. Acute intermittent porphyria has no cutaneous manifestations. In variegate porphyria, patients exhibit both the cutaneous features of porphyria cutanea tarda and the visceral mani­ festations of acute intermittent porphyria. Hereditary coproporphyria is characterized by cutaneous blisters and similar, but milder, symptoms of acute intermittent por­ phyria. Hepatoerythropoietic porphyria is marked by occa­ sional splenomegaly and anemia with subepidermal vesicles and blisters that often progress to scleroderma-like scarring and hyperpigmentation. Connective tissue diseases such as systemic lupus erythe­ matosus (SLE), dermatomyositis (DM), and progressive sys­ temic scleroderma (PSS) all have characteristic skin and GI manifestations (see Chapter 35). SLE patients prototypically have malar erythema with photosensitivity and often ery­ thematous raised patches with follicular plugging (discoid lupus). SLE patients can have oral ulcers, anorexia, nausea, vomiting, peritonitis with ascites, autoimmune hepatitis, and pancreatitis. DM is described in greater detail later, but patients with DM can have oropharyngeal dysphagia and large bowel infarction from vasculopathy (especially in juvenile DM). Patients with PSS often demonstrate gener­ alized sclerotic skin or, less commonly, morphea (sclerotic plaques with ivory-colored centers), matted telangiectasia, and Raynaud’s phenomenon. Esophageal dysfunction is the most common internal symptom, although the small intes­ tine may also be affected, producing constipation, diarrhea, and bloating.

CUTANEOUS MANIFESTATIONS OF GASTROINTESTINAL MALIGNANCIES Cutaneous manifestations may be of importance in recog­ nizing individuals with cancer or from kindred with a high risk for the development of cancer. These cutaneous markers are discussed in three sections: syndromes with GI poly­ posis and skin findings, cutaneous markers of internal malignancy, and cutaneous manifestations of metastatic GI carcinoma.

epidermoid cysts, which usually appear on the back and occur after puberty. True sebaceous cysts (steatocystomas) are not associated with Gardner’s syndrome. The oral mani­ festations of Gardner’s syndrome include the presence of 1- to 10-mm osteomas and multiple unerupted, supernu­ merary teeth. Muir-Torre syndrome is an autosomal dominant syn­ drome with cutaneous sebaceous neoplasms and multiple primary cancers, especially of the proximal colon.42,43 It is part of the cancer family syndrome (CFS), or Lynch II syn­ drome. The most prominent cutaneous manifestation is one or more sebaceous neoplasms of various degrees of differ­ entiation, from benign adenoma to aggressive sebaceous carcinoma. Because cutaneous sebaceous neoplasms are rare, even the presence of one lesion should prompt evalu­ ation for this syndrome. In addition, keratoacanthomas and basal and squamous cell carcinomas of the skin develop in these patients. Multiple primary (low-grade) malignancies are characteristic; in one series, 40 patients had a total of 106 tumors and 1 patient had nine different primary carci­ nomas. The most common location for carcinomas is the GI tract (93%), especially the proximal colon. One or more polyp(s) of the intestine have been described in 38% of patients, but multiple adenomatous polyposis is absent in this syndrome.43 Urogenital carcinomas, especially of the endometrium, bladder, and kidney, occur in 50% of patients. The defects leading to this phenotype are found in the DNA repair genes MLH1 and MSH2 on chromosome 3p.43 Peutz-Jeghers syndrome is an autosomal dominant syn­ drome of GI hamartomas and mucocutaneous hyperpigmen­ tation (Fig. 22-19). The macules appear during infancy and early childhood. Mucosal lesions persist, whereas the cuta­ neous lesions fade over time. The hyperpigmented lesions consist of dark brown 1-mm to 1-cm macules on the lips (95%), buccal mucosa (83%), and acral areas (palms, soles, digits), and around the eyes, anus, and mouth. The most common associated malignancy is duodenal carcinoma; granulosa theca cell tumors of the ovary may be present in up to 20% of female patients. The patient with PeutzJeghers syndrome may carry an overall relative risk of cancer of up to 18. The cause is a mutation in the STK11 gene (see Chapters 3 and 122).44 Confusion may occur between Peutz-Jeghers syndrome and Laugier-Hunziker syndrome, which is a benign oral and acral lentiginous pigmentation disorder. In Laugier-Hunziker syndrome, the pigmented macules involve the lips, tongue, buccal mucosa and nails. Laugier-Hunziker syndrome is best differentiated from Peutz-Jeghers syndrome by its lack of personal or family history for GI cancers and lack of physical finding for lentigines that cross the vermillion border.45,46

POLYPOSIS SYNDROMES

Gardner’s syndrome, or familial adenomatous polyposis, is inherited as an autosomal dominant trait (see Chapter 122). The adenomatous polyposis coli (APC) gene on chromo­ some 5q21 is mutated in the germline of these patients. Cutaneous features of this syndrome occur in more than 50% of affected individuals and often appear before the polyps become symptomatic. Multiple epidermoid cysts (also called inclusion cysts) of the face, scalp, and extremi­ ties appear before puberty. This is in contrast with common

Figure 22-19.  Mucocutaneous pigmentation in a patient with PeutzJeghers syndrome.

363

364

Section IV  Topics Involving Multiple Organs Cowden’s syndrome, or multiple hamartoma syndrome, is an uncommon autosomal dominant syndrome with mul­ tiple mucocutaneous manifestations and an increased risk of malignancy. The diagnostic skin lesions are multiple facial verrucous lesions that histologically are trichilem­ momas. Oral papillomatosis is common. Thyroid disease (goiter, adenoma, cancer) are important components of the syndrome, as is fibrocystic disease of the breast (60%) and breast carcinoma (29%) in affected women. GI lesions occur in at least 40% of patients and consist primarily of multiple polyps, which occur anywhere along the GI tract, most com­ monly the colon. These polyps are usually small and pre­ dominantly hamartomatous. The gene mutation responsible is PTEN (phosphatase and tensin homolog), found on chro­ mosome 10q.47 Cronkhite-Canada syndrome is a rare sporadic syndrome of generalized GI polyposis, mucocutaneous hyperpigmen­ tation, alopecia, malabsorption with malnutrition, and nail dystrophy.48 The mean age at onset is 59 years. Diarrhea, weight loss (usually more than 10 kg), and abdominal pain are the most common symptoms. Nail changes (90% of patients) affect all 20 nails and consist of thinning and split­ ting of nails, onycholysis (separation of the nail from the nail bed), or total shedding of the nails. Alopecia (more than 95% of patients) is usually sudden and involves not only the scalp but also the body hair. Hyperpigmentation occurs in about 85% of patients and has been described as lentigi­ nes that may coalesce, most commonly on the upper extrem­ ities, lower extremities, palms, and soles. The cutaneous changes all resolve with treatment but may resolve sponta­ neously even with continued GI disease. Death occurs in about half of patients as a result of persistent diarrhea or malnutrition. Aggressive nutritional support in the form of total parenteral nutrition has led to complete resolution of the syndrome, suggesting that at least some of the manifesta­ tions are a complication of the metabolic abnormalities caused by the severe diarrhea.

INTERNAL MALIGNANCY AND RELATED DISORDERS

Dermatomyositis is manifested by a violaceous color of the eyelids, often with edema (heliotrope); keratotic papules over the knuckles (Gottron’s papules; Fig. 22-20); a wide­ spread erythema, often with accentuation over the elbows and knees (Gottron’s sign), resembling psoriasis; photosen­ sitivity; and nail cuticle abnormalities, including telangiec­

Figure 22-20.  Dermatomyositis with erythematous plaques, especially over the knuckles (Gottron’s papules). (Courtesy of Dr. Timothy Berger, San Francisco, Calif.)

tases, thickening, roughness, overgrowth, and irregularity. About 25% of patients with dermatomyositis have internal malignancy, particularly in patients older than 50 years.49 Cancers most commonly associated with dermatomyositis are gastric, colorectal, pancreatic, ovarian, and lung cancer and non-Hodgkin’s lymphoma. Nasopharyngeal carcinoma is the most commonly associated malignancy in the Chinese population. There does not appear to be a predilection for either gender. To detect an associated cancer, a complete medical history, physical examination, including rectal, pelvic, and breast examinations, CBC, routine serum chemistry analysis, serum protein electrophoresis, multiple fecal occult blood tests, a urinalysis, chest roentgenography, and mammography (in women) are recommended. Any abnormalities should be investigated further. Extensive blind evaluation of patients with dermatomyositis is not warranted.50,51 Keratosis palmaris et plantaris (Howel-Evans syndrome; tylosis and esophageal cancer) is an adult-onset diffuse hyperkeratosis of the palms and soles that has been described in association with a very high incidence of esophageal carcinoma in several kindred in Liverpool, England. It is an autosomal dominant phenotype caused by loss of heterozy­ gosity of the TOC gene on chromosome 17q.52 The skin lesions appear during adolescence or early adulthood, and the carcinomas appear on the average at 45 years. Esopha­ geal carcinoma develops in almost all patients in these kindred with tylosis.52 Tripe palms, also called acanthosis nigricans of the palms, acanthosis palmaris, pachydermatoglyphy, palmar hyper­ keratosis, and palmar keratoderma, is a paraneoplastic phe­ nomenon characterized by a moss-like or velvety texture with pronounced dermatoglyphics or by a cobbled or hon­ eycombed surface of the palms and fingers. Of reported cases of tripe palms, 91% have occurred in association with neoplasm. Gastric and lung cancers were the most common neoplasms, each accounting for more than 25% of all the malignancies.53 Acanthosis nigricans is a cutaneous finding that manifests with a velvety hyperplasia and hyperpigmentation of the skin of the neck and axillae (Fig. 22-21), often associated with multiple skin tags. Some patients with acanthosis nigricans have internal malignancy, so-called malignant acanthosis nigricans. In these patients, the extent of involve­ ment may be severe, including the hands, genitalia, and oral mucosa. The associated carcinoma is usually present simul­ taneously with the acanthosis nigricans but may not yet be clinically evident. Intra-abdominal adenocarcinomas con­ stitute more than 85% of associated malignancies, with gastric carcinomas representing more than 60%. Survival is short, and more than 50% of patients die in less than one year.

Figure 22-21.  Acanthosis nigricans in the axilla.

Chapter 22  Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease Paraneoplastic acrokeratosis of Bazex is a rare but distinc­ tive syndrome associated with a primary malignant neo­ plasm of the upper aerodigestive tract or metastatic carcinoma to the lymph nodes of the neck. All of the more than 50 patients reported to date have had malignancy, including esophageal carcinoma and one gastric carcinoma with cervical metastases. The skin eruption begins acrally as thickening of the periungual skin and marked nail dys­ trophy. The rash progresses proximally and also involves the tip of the nose and ears. Thickening of the palms and soles ensues initially, with central sparing, which can make walking very painful. Eventually, the face and scalp become involved. Treatment of the underlying carcinoma is usually associated with improvement or resolution of the skin lesions. Hypertrichosis lanuginosa, another rare paraneoplastic syndrome consisting of fine, thin, down-like, unpigmented lanugo-type hair, is typically noted on the face, forehead, ears, nose, axillae, limbs, and trunk. Other manifestations include glossodynia, papillary hypertrophy of the tongue, disturbances of taste and smell, diarrhea, scleroderma, acanthosis nigricans, seborrheic keratoses, adenopathy, and weight loss. Colorectal carcinomas are second only to lung carcinoma in frequency of associated malignancies. Carcinoid tumors produce a number of vasoactive sub­ stances that can induce cutaneous flushing (see Chapter 31). The most common carcinoid tumors (appendix and small bowel) do not produce flushing until the vasoactive sub­ stances reach the systemic circulation. Flushing, therefore, generally denotes metastasis to the liver or a different primary tumor site (e.g., lung or ovary). Glucagonoma of the pancreas often precipitates necrolytic migratory erythema of the skin. The rash is common around orifices, flexural regions, and the fingers. Lesions are typi­ cally papulovesicular, with secondary erosions, crusting, and fissures appearing in a geographic circinate pattern (Fig. 22-22). Patients can also often have weight loss,

diarrhea, anemia, psychiatric disturbances, hypoaminoaci­ demia, and diabetes. The rash typically clears with success­ ful removal of the tumor (discussed in more detail in Chapter 32). Subcutaneous fat necrosis and polyarthralgia is associ­ ated with pancreatic acinar cell carcinoma, pancreatitis, and pancreatic pseudocysts. Most affected persons are men. Deep, subcutaneous, erythematous nodules ranging from 1 to several centimeters in diameter usually appear on the legs. In uncommon cases, the nodules may break down, exuding a creamy material. Arthritis of one or several joints, especially the ankles and knees, may accompany the nodules or occur without skin lesions. Abdominal pain may be absent when the skin lesions or arthritis occur. In addi­ tion to the expected elevations of serum amylase and lipase levels, eosinophilia is common. Histopathologic evaluation of skin lesions usually reveals diagnostic findings—pale staining necrotic fat cells (ghost cells) and deposits of calcium in the necrotic fat. The mortality rate in cases not associated with carcinoma approaches 50%. Subcutaneous swellings, which commonly break down and drain, may also be seen in patients with α1-antitrypsin deficiency. These nodules usually occur on the buttocks or proximal extremities and are often precipitated by trauma. In pancre­ atitis, subcutaneous nodules usually manifest on the ante­ rior shins. A bluish discoloration of the skin (ecchymosis) around the umbilicus, sometimes associated with hemor­ rhagic pancreatitis, is called Cullen’s sign; when a similar process occurs in the flank, it is called the Grey-Turner sign (see Chapter 58 for example). Some cutaneous markers historically thought to be asso­ ciated with internal malignancies have more recently been dismissed as having no direct relationship. These include Bowen’s disease (cutaneous squamous cell carcinoma in situ), skin tags, and the Leser-Trélat sign (sudden appear­ ance of multiple seborrheic keratoses). Sweet’s syndrome (acute febrile neutrophilic dermatoses) might be associated with lymphoproliferative neoplasm instead of GI malig­ nancy, if such an association truly exists.

CUTANEOUS METASTASES

Cutaneous metastases occur rarely with GI adenocarcino­ mas. They may appear anywhere on the skin and are often nonspecific, very firm, dermal or subcutaneous nodules. When metastasis to the umbilicus occurs, however, intraabdominal GI carcinoma is found in more than half of cases and gastric carcinoma in 20%. This lesion is called Sister Mary Joseph’s nodule. Immunoperoxidase markers have assisted pathologists in predicting the primary site of origin from biopsy specimens of metastatic nodules.

CUTANEOUS MANIFESTATIONS OF LIVER DISEASE

Figure 22-22.  Necrolytic migratory erythema, in a patient with glucagonoma, characterized by rapidly eroding, superficial blisters. The lesions are usually localized to the buttocks, groin, perineum, elbows, hands, feet, and perioral area. (Courtesy of Dr. Carl Grunfeld, San Francisco, Calif.)

In addition to jaundice, patients with liver disease may show vascular spider angiomata, corkscrew scleral vessels, palmar erythema, telangiectasia, striae, and caput medusa.54 Patients with hemochromatosis often develop a generalized bronze-brown color with accentuation over sun-exposed sites. Primary biliary cirrhosis may be associated with xan­ thomas that involve the trunk, face, or extremities. Striking plane xanthomas may develop on the palmar creases (see Chapter 89). Patients with sarcoidosis involving the liver or, less commonly, the GI tract may have sarcoid skin lesions (Fig. 22-23).

365

366

Section IV  Topics Involving Multiple Organs

Figure 22-24.  Porphyria cutanea tarda characterized by noninflammatory blisters and erosions of the dorsa of the hands. Affected patients are frequently infected with hepatitis C virus. (Courtesy of Dr. Timothy Berger, San Francisco, Calif.)

Figure 22-23.  Patient with sarcoidosis manifesting as an annular plaque on the face. (Courtesy of Dr. Timothy Berger, San Francisco, Calif.)

Pruritus is a distressing complication of cholestatic, inflammatory, and malignant liver diseases. The itching of liver disease is not relieved by scratching or topical gluco­ corticoids and may be difficult to manage. Amelioration of pruritus with ultraviolet B light treatment, cholestyramine, or rifampin does not help in elucidating the pathogenesis of this distressing condition. Opiate antagonists may relieve pruritus, which suggests that endogenous opioids are involved in its pathogenesis. The pruritus associated with metastatic disease to the liver has been successfully treated with intravenous and oral ondansetron, a 5-HT3 receptor antagonist. Vitamin K is frequently administered to patients with liver disease and hypoprothrombinemia. Cutaneous reactions, although rare, may occur after subcutaneous, intramuscular, or intravenous administration. Large, ery­ thematous, indurated, pruritic plaques occur within a few days to a few weeks. These reactions may be a delayed hypersensitivity reaction, in that dermal testing can repro­ duce the reactions. When tested, patients have been found to be allergic to the vitamin K and not the benzoyl alcohol vehicle. However, vitamin K3 (Synkayvite), which is watersoluble, has not been reported to cause similar reactions. If reactions occur after buttock injections of vitamin K, there is an almost diagnostic tendency of these plaques to spread around the waist and down the thigh, reproducing what has been called a cowboy gun belt and holster pattern. These reaction sites resolve over days to weeks but may persist for months to years. After an erythematous reaction, or without prior reaction, expanding sclerotic plaques with violaceous borders similar to those of morphea have occurred months to years after injections. The latter pattern usually occurs after large parenteral doses of vitamin K. In addition to these local reactions, anaphylaxis after intravenous administra­ tion that may be fatal has been reported. The association between polyarteritis nodosa and hepa­ titis B is well documented. Urticaria and serum sickness occur more commonly in patients with hepatitis B, although both have been reported in association with hepatitis C (see Chapters 78 and 79). Chronic hepatitis C virus is

associated with leukocytoclastic vasculitis with cryoglobu­ linemia. Petechiae and palpable purpura are noted on the skin. Porphyria cutanea tarda (PCT) is a metabolic disorder characterized by skin fragility, blisters, hypertrichosis, and hyperpigmentation in sun-exposed skin (Fig. 22-24). PCT is the most common form of porphyria and is characterized by a deficiency of uroporphyrinogen decarboxylase. Diag­ nosis is typically made with a 24-hour urine collection demonstrating elevated porphyrin levels, specifically uroporphyrin. Alcohol consumption, estrogens, iron, and sunlight all are known to exacerbate PCT. There is a clear and substantial link between PCT and hepatitis C.55 The prevalence of hepatitis C in patients with PCT demonstrates regional variation, ranging from 65% in southern Europe and North America to 20% in northern Europe and Aus­tralia.54 Treatment involves phlebotomy and antima­ larial agents. Lichen planus is a common, idiopathic, inflammatory disorder that can affect skin, hair, mucous membranes, and nails (see earlier). The prototypical presentation of lichen planus is violaceous, polygonal, flat-topped papules of flexural areas of the wrists, arms, and legs. The papules often have an overlying reticulated white scale known as Wickham’s striae. An association between lichen planus and hepatitis C exists but is not as prominent as the link between PCT and hepatitis C.56

DRUG-INDUCED LIVER DISEASE IN PATIENTS WITH SKIN DISEASE Dermatologists frequently consult gastroenterologists for evaluation of patients who are being treated with methotrex­ ate or retinoids, because these medications can cause acute and chronic liver disease (see Chapter 86). Methotrexate is the more commonly used of these medications. It is extremely effective for severe psoriasis and psoriatic arthri­ tis and is also used for cutaneous T cell lymphoma, connec­ tive tissue diseases such as rheumatoid arthritis, and other dermatologic and GI disorders. Methotrexate is usually given as a single weekly dose of 10 to 25 mg, but may be used in higher dosages in selected patients. A grading system for liver biopsies has been established and is gener­

Chapter 22  Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease Table 22-2  Grading System for Liver Biopsy Findings in Patients Taking Methotrexate GRADE

CRITERIA

I

Normal; mild fatty infiltration; nuclear variability, portal inflammation Moderate to severe fatty infiltration; nuclear variability; portal tract expansion, portal tract inflammation, and necrosis Mild fibrosis (fibrosis denotes formation of fibrotic septa extending into the lobules) Moderate to severe fibrosis Cirrhosis (regenerating nodules as well as bridging of portal tracts must be demonstrated)

II IIIA IIIB IV

From Roenigk HH Jr, Auerbach R, Maibach H, et al: Methotrexate in psoriasis: Consensus conference. J Am Acad Dermatol 1998; 38:478-85.

Table 22-3  Guidelines for Continuation of Methotrexate Therapy* Patients with grade I or II changes may continue to receive methotrexate therapy. Patients with grade IIIA change(s) may continue to receive methotrexate therapy but should have a repeat liver biopsy after approximately six more months of methotrexate therapy. Alternative therapy should be considered. Patients with grade IIIB and IV changes should not be given further methotrexate therapy. Exceptional circumstances, however, may require continued methotrexate therapy, with follow-up liver biopsies. *Based on the results of liver biopsy (see Table 22-2). From Roenigk HH Jr, Auerbach R, Maibach H, et al: Methotrexate in psoriasis: Consensus conference. J Am Acad Dermatol 1998; 38:478-85.

ally followed by dermatologists (Table 22-2). Current Amer­ ican Academy of Dermatology guidelines recommend pretreatment liver biopsies and repeated biopsies during therapy, depending on the results of regular liver chemistry tests and determination of other risk factors for hepatic disease (obesity with diabetes, results of prior liver biopsies, and alcohol consumption). Decisions on continuation of treatment are frequently based on the results of these biop­ sies.57 An American College of Gastroenterology committee has made similar recommendations (Table 22-3). Retinoids (e.g., Accutane, Acitretin), derivatives of vitamin A, are currently used for the treatment of certain forms of severe psoriasis, cystic acne, and other disorders of keratinization. Regular evaluation of liver function is required during this treatment. Mild elevations of serum triglyceride, cholesterol, and alanine and aspartate amino­ transferase levels are common (20% to 30% of patients treated), usually transient, or easily managed by reducing the dose. Severe or even fatal hepatitis has been reported, however. Retinoids may be used for patients with psoriasis who were previously treated with methotrexate or who have liver disease contraindicating the use of methotrexate. Limited experience suggests that these patients do not suffer progression of the liver disease with such retinoid therapy. As with methotrexate, during retinoid therapy there is a poor correlation between liver chemistry test and liver his­ tology results. Therefore, pretreatment and intermittent liver biopsies may be required for certain high-risk patients being treated with oral retinoids chronically.

Figure 22-25.  Cutaneous larva migrans characterized by a serpiginous erythematous migratory lesion caused by an infection with dog hookworm. (Courtesy of Dr. Timothy Berger, San Francisco, Calif.)

PARASITIC DISEASES OF THE INTESTINE AND SKIN The larval forms of human and animal nematodes may cause migratory erythematous skin lesions, called creeping eruptions (see Chapter 110). The most common pattern is cutaneous larva migrans, caused by dog and cat hookworms (Fig. 22-25). Pruritic linear papules migrate at a rate of 1 to 2 cm daily on skin sites, usually the feet, buttocks, or back, that have come in contact with fecally contaminated soil. Lesions resolve spontaneously over weeks to months. Larva currens is caused by Strongyloides stercoralis larva migrat­ ing in the skin. It occurs in two forms, a form localized to the perirectal skin in immunocompetent hosts and a dis­ seminated form occurring in immunosuppressed hosts. S. stercoralis has the unique capacity among nematodes to develop into infective larvae within the intestine. These infective larvae may invade the perirectal skin in infected immunocompetent individuals, causing urticarial, erythem­ atous, linear lesions that migrate up to 10 cm a day, usually within 30 cm of the anus. Skin lesions may occur intermit­ tently, making diagnosis difficult. In immunosuppressed hosts, repeated autoinfection through the intestine leads to a tremendous parasite burden (hyperinfection), manifested most commonly by pulmonary disease. In association, dis­ seminated larva currens–type lesions may appear over the whole body, especially the trunk. Petechial or purpuric serpiginous lesions may also occur periumbilically. Parasitic infections are classically considered in the dif­ ferential diagnosis of urticaria. Except for fascioliasis and hydatid disease, however, a direct relationship has rarely been proved. If blood eosinophilia and GI symptoms are absent, stool examination for parasites is rarely beneficial.

DERMATITIS HERPETIFORMIS AND CELIAC DISEASE Dermatitis herpetiformis (DH) is an extremely pruritic skin disorder most commonly appearing during early adulthood (see Chapter 104). The cutaneous eruption consists of urti­ carial, vesicular, or bullous lesions characteristically local­ ized to the scalp, shoulders, elbows, knees, and buttocks.58 The disorder is so pruritic that often all the skin lesions

367

368

Section IV  Topics Involving Multiple Organs

Figure 22-26.  Dermatitis herpetiformis characterized by pruritic, urticarial papules and small blisters concentrated over the shoulders, scalp, and lumbosacral area. (Courtesy of Dr. Timothy Berger, San Francisco, Calif.)

have been excoriated, and the diagnosis must be suspected on the basis of this and the distribution (Fig. 22-26). The diagnosis of DH is established by skin biopsy and direct immunofluorescence examination of the skin. Deposits of IgA are found in the dermal papillae at sites of itching and where vesicles are forming. Patients with DH commonly have an enteropathy indistinguishable from celiac disease. Their human leukocyte antigen (HLA) patterns, including haplotypes B8, DR3, and DQw2; abnormalities of intestinal absorption; antiendomysial and antigliadin antibodies; and small bowel biopsy findings are similar to those of patients with celiac disease, and yet fewer than 5% of patients with DH have symptomatic GI disease. Gluten has been shown to be the dietary trigger of DH. Patients, even those with such minimal bowel disease that the bowel biopsy finding is normal, improve on a gluten-free diet. Reintroduction of gluten in a symptom-free patient on a gluten-free diet leads to the reappearance of pruritus and skin lesions. The IgA antibodies deposited in the skin and causing the eruption have not been proved to originate in the gut, and do not seem to be directed against gluten. The pathogenesis remains unknown. Because it is occasionally difficult to distinguish DH from other blistering skin diseases, a patient with an extremely pruritic eruption may be referred for endoscopy. The finding of an abnormal small intestine consistent with celiac disease in a patient with a pruritic eruption would be highly sug­ gestive of DH (see Chapter 104). The skin lesions of DH respond dramatically to sulfa drugs (dapsone or sulfapyr­ idine), but the gut pathology and skin immunofluorescence are unchanged by sulfa drugs. Treatment with a gluten-free diet leads to gradual clearing of skin lesions, improvement of the intestinal abnormality, disappearance of the IgA from the skin, and decreased dependence on dapsone for control of the cutaneous eruption.59

VITAMIN DEFICIENCIES See also Chapters 4 and 100. Pellagra, a deficiency of niacin, may be related to inade­ quate diet, medication (isoniazid), or the carcinoid syn­ drome.60 The lesions appear symmetrically in sun-exposed areas as brown-red, blistering, or scaling plaques, which may become indurated. Glossodynia, atrophic glossitis and, sometimes, ulcerative gingivostomatitis may be present. In

Figure 22-27.  Infant girl with acrodermatitis enteropathica secondary to nutritional zinc deficiency. She was subsisting on a diet of rice cereal and water. (Courtesy of Dr. Genevieve Wallace, Dallas, Tex.)

Figure 22-28.  Lower extremities of an older man with Whipple’s disease. Perifollicular hemorrhage is apparent. Plasma vitamin C levels were decreased. The skin lesions rapidly disappeared after vitamin C supplementation. (Courtesy of Dr. Mark Feldman, Dallas, Tex.)

addition to dermatitis (with or without oral lesions), diar­ rhea and dementia may occur (the three Ds), as may death if untreated. Deficiencies of zinc (acrodermatitis enteropathica), essen­ tial fatty acids, and biotin all produce a superficial scaling and an occasionally blistering eruption accentuated in the groin and around the mouth (Fig. 22-27). Alopecia is often present. These conditions are most common in children with congenital metabolic abnormalities, in alcoholics with cirrhosis, and in persons on hyperalimentation who have

Chapter 22  Oral Disease and Oral-Cutaneous Manifestations of Gastrointestinal and Liver Disease not been adequately supplemented. Replacement of the deficiency leads to rapid resolution of the dermatitis and alopecia. Acrodermatitis enteropathica also occurs in zincdeficient patients with Crohn’s disease.61 Scurvy, or vitamin C deficiency, manifests as follicular hyperkeratosis and perifollicular hemorrhage, ecchymoses, xerosis, leg edema, poor wound healing, and/or bent or coiled body hairs.62 Large purpuric plaques, especially on the extremities, may occur (Fig. 22-28). Gingivitis with gum hemorrhage occurs only in dentulous patients and com­ monly occurs in the presence of poor oral hygiene and periodontal disease.63 Scurvy is most common in alcoholics, but may occur with Crohn’s or Whipple’s disease. The focus of treatment is to correct the vitamin C deficit and to replete body stores. Symptoms recede promptly and disappear within a few weeks.64 As discussed earlier, a characteristic dermatosis called necrolytic migratory erythema63,64 frequently develops in patients with glucagon-secreting tumors of the pancreas (see Fig. 22-22). A skin biopsy specimen may be highly suggestive, showing psoriasiform hyperplasia, a subcorneal blister containing neutrophils, and hydropic degeneration and necrosis of the subcorneal keratinocytes. Manifestations of this syndrome are discussed in Chapter 32. There are also reports of this syndrome occurring without glucagono­ mas, especially in the setting of cirrhosis and subtotal villous atrophy of the jejunal mucosa.65 Glucagon is there­ fore pro­bably not the cause of the eruption. Infusion of

amino acids has been reported to clear the eruption, despite persistently elevated glucagon levels. Zinc deficiency can cause a similar eruption (acrodermatitis enteropathica), as can biotin-responsive multiple carboxylase deficiency and essential fatty acid deficiency. The eruption seems to be a cutaneous manifestation of several metabolic dis­ orders, but the crucial pathogenic defect has not been determined.66

KEY REFERENCES

Airio A, Pukkala E, Isomaki H. Elevated cancer incidence in patients with dermatomyositis: A population-based study. J Rheumatol 1995; 22:1300-3. (Ref 49.) Delahoussaye AR, Jorizzo JL. Cutaneous manifestations of nutritional disorders. Dermatol Clin North Am 1989; 7:559-70. (Ref 60.) Faure M. Dermatitis herpetiformis. Semin Dermatol 1988; 7:123-9. (Ref 58.) Karaca S, Seyhan M, Senol M, et al. The effect of gastric Helicobacter pylori eradication on recurrent aphthous stomatitis. Int J Dermatol 2008; 47:615-7. (Ref 26.) National Center for Biotechnology Information. Muir-Torre syndrome, 2008. Available at http://www.ncbi.nlm.nih.gov/entrez/dispomim. cgi?id=158320. (Ref 43.) Reunala TL. Dermatitis herpetiformis. Clin Dermatol 2001; 19:728-36. (Ref 59.) Sardana K, Mishra D, Garg V. Laugier Hunziker syndrome. Indian Pediatr 2006; 43:998-1000. (Ref 45.) Tierney EP, Badger J. Etiology and pathogenesis of necrolytic migratory erythema: Review of the literature. MedGenMed. 2004; 6:4. (Ref 64.) Full references for this chapter can be found on www.expertconsult.com.

369

CHAPTE R

23 Diverticula of the Pharynx, Esophagus, Stomach, and Small Intestine William V. Harford, Jr. and D. Rohan Jeyarajah

CHAPTER OUTLINE Zenker’s Diverticulum  371 Diverticula of the Esophageal Body  373 Esophageal Intramural Pseudodiverticula  374 Gastric Diverticula  375

Diverticula are outpouchings from tubular structures. True diverticula involve all layers of the intestinal wall, whereas false diverticula are caused by herniation of mucosa and submucosa through the muscular wall. Many diverticula contain attenuated portions of the muscular wall of the intestine, and hence may be difficult to define as true or false. True diverticula are often assumed to be congenital lesions and false diverticula are assumed to be acquired, but this is not always the case. Some authors reserve the terms false diverticula or pseudodiverticula for diverticula caused by an inflammatory process.

ZENKER’S DIVERTICULUM Ludlow first described a patient with a hypopharyngeal diverticulum in 1767, and in 1877 Zenker and Von Ziemssen reported 23 such patients.

Cause and Pathogenesis

Zenker’s diverticula are acquired. They develop when abnormally high pressures occurring during swallowing lead to protrusion of mucosa through an area of anatomic weakness in the pharynx known as Killian’s triangle. High pressures are generated when the opening of the upper esophageal sphincter (UES) is impaired. In patients with Zenker’s diverticula, several pathophysiologic changes have been documented in the cricopharyngeus. These changes lead to a reduction in compliance and to decreased opening of the UES.1 Killian’s triangle is located where the transverse fibers of the cricopharyngeal sphincter intersect with the oblique fibers of the inferior pharyngeal constrictor muscle. The size of this area of weakness varies among individuals. Relatively large defects may predispose to the development of Zenker’s diverticula.2 Diverticula similar in appearance to Zenker’s diverticula have been reported as a complication of anterior cervical spine surgery.3,4

Duodenal Diverticula  376 Extraluminal Diverticula  376 Intraluminal Diverticula  376 Jejunal Diverticula  377

Clinical Features and Diagnosis

The prevalence of Zenker’s diverticula has been estimated to be between 0.1% and 0.01%. Patients generally present in the seventh or eighth decade of life. Twice as many men as women develop Zenker’s diverticula.1,5 Common presenting symptoms are listed in Table 23-1. Patients with small diverticula may be asymptomatic. Zenker’s diverticula may be discovered incidentally during barium swallow or upper endoscopy carried out for investigation of unrelated problems. Squamous cell cancer may develop in Zenker’s diverticula. The incidence has been estimated to be from 0.4% to 1.5%.1,6,7 If myotomy without diverticulectomy is planned, it is prudent to inspect the lining of the diverticulum carefully for any evidence of cancer. Bleeding may occur from ulcerated Zenker’s diverticula. Aspiration of retained food contents may complicate in­ duction of anesthesia.8 Medications may become lodged in Zenker’s diverticula. Corrosive medications may cause ulceration. Unpredictable absorption of tablets or capsules may also lead to clinical problems.9 Accumulation of radioactive iodine tracer in a Zenker’s diverticulum has been reported to lead to an erroneous diagnosis of metastatic thyroid cancer.10 Videocapsules may also become lodged in Zenker’s diverticula and should be delivered into the stomach with a fiberoptic endoscope when such studies are required.11,12 Intubation of the trachea or the esophagus may be complicated by the presence of a Zenker’s diverticulum. A large diverticulum displaces the lumen of the esophagus. The tip of the intubation instrument is directed preferentially into the diverticulum. At endoscopy, it may be difficult to distinguish the lumen of the diverticulum from the true lumen of the esophagus (Fig. 23-1A). Endotracheal intubation, placement of a nasogastric tube, and intubation of the esophagus for upper endoscopy, endoscopic retrograde cholangiopancreatography, or transesophageal echocardiography may be difficult. Perforation can occur. Intubation in patients with Zenker’s diverticula should be done under

371

372

Section IV  Topics Involving Multiple Organs

A

over the guidewire. An alternative technique consists of passing a forward-viewing endoscope loaded with an overtube. Once the endoscope has been passed into the esophagus, the overtube is advanced, the forwardviewing endoscope is withdrawn, and the side-viewing or ultrasound endoscope is passed through the overtube.13 Zenker’s diverticulum can be suspected from a careful history (see Table 23-1). Barium swallow is the most useful diagnostic study. The radiologist should be alerted in advance, so that proper views are taken (see Fig. 23-1B). Small diverticula may be seen only transiently. Barium swallow in the lateral view using video fluoroscopy is helpful for detecting small diverticula. The opening of a large Zenker’s diverticulum often becomes aligned with the axis of the esophagus. Oral contrast will preferentially fill the diverticulum and will empty slowly. Large diverticula are therefore often obvious, even on delayed images. During endoscopy, Zenker’s diverticulum should be suspected if, on entering the pharynx, the upper esophageal sphincter cannot be located. In such cases, the endoscopy should be stopped and the patient should be sent for a barium study.

Treatment and Prognosis

B Figure 23-1.  Zenker’s diverticulum. A, Endoscopic view. It is often difficult to distinguish the lumen of the esophagus from the lumen of the diver­ ticulum. B, Barium esophagogram showing a diverticulum large enough to cause esophageal obstruction when it fills. (A, Courtesy of Dr. David Langdon, Arlington, Texas; B, Courtesy of Dr. Charles E. Pope, Seattle, Wash.)

Table 23-1  Presenting Symptoms in Patients with a Zenker’s Diverticulum Dysphagia Regurgitation Choking Aspiration Voice changes Halitosis Weight loss

direct vision. When a large Zenker’s diverticulum causes marked anatomic distortion or when intubation with a sideviewing endoscope is required, direct intubation is not prudent. In such cases, a forward-viewing endoscope can be used to pass a soft-tipped guidewire into the esophageal lumen. The guidewire is then back-loaded into the endoscope and the endoscope is advanced into the esophagus

Patients with small asymptomatic or minimally symptomatic diverticula can be followed, because progressive enlargement is uncommon.2 Patients with large and symptomatic Zenker’s diverticula should be offered treatment. Zenker’s diverticula may be treated by open surgical procedures or by transoral endoscopic techniques with rigid or flexible fiberoptic instruments. Open surgery for Zenker’s diverticula can be performed as an outpatient procedure.14 An open surgical approach is the safest alternative for patients with large (>5 cm) diverticula that extend into the thorax.15 Damage to mediastinal structures is best avoided by optimal exposure. Large diverticula can be resected, inverted, or suspended (diverticulopexy). Resection of small diverticula is not required. UES myotomy should always be part of the procedure. If diverticula are resected without myotomy, there is an increased risk of postoperative leaks and an increased frequency of recurrence.16,17 Complications of open surgery include leaks with mediastinitis, esophagocutaneous fistula, and vocal cord paralysis from recurrent laryngeal nerve injury. Endoscopic treatment has become the predominant technique for the management of Zenker’s diverticula since the introduction of stapling for myotomy in 1993. Compared with open surgical approaches, endoscopic approaches are associated with lower complication rates, shorter anesthesia times, and shorter hospital stays.7 Endoscopic techniques are suitable for patients with medium-sized diverticula (2 to 5 cm). Specially designed rigid diverticuloscopes (e.g., Weerda, van Overbeek) and conventional flexible fiberoptic endoscopes have been used. The diverticuloscope is used to provide optimal visualization of the lumen of the esophagus and diverticulum and the septum between them (Fig. 23-2). This septum is comprised of the posterior wall of the esophagus and the anterior wall of the diverticulum, and includes the UES. The muscular layers of this septum are incised, resulting in ablation of the UES and restoration of a single lumen. The incision can be performed by a number of techniques. With a rigid diverticuloscope, CO2 laser and surgical staplers have been used. Stapling has become the predominant technique since it was reported to be safer and more effective than CO2 laser.7,18 With CO2 laser, fusion of the cut edges of the incision is relied on to prevent leaks. With endoscopic staplerassisted myotomy, a double row of staples is placed along

Chapter 23  Diverticula of the Pharynx, Esophagus, Stomach, and Small Intestine Diverticuloscope

Esophageal lumen

Figure 23-3.  Endoscopic view of a midesophageal diverticulum. These diverticula are most apparent when the esophagus is well insufflated. Zenker’s diverticulum Figure 23-2.  Weerda diverticuloscope. The instrument is positioned to expose the common wall between the lumen of the esophagus and the Zenker’s diverticulum.

the cut edges, reducing the risk of perforation and bleeding. Stapling may not be technically feasible if the diverticulum is short (<3 cm), because not enough of the stapler will fit into the diverticulum. Modifications of the stapler and other techniques may improve results in short diverticula.19 Conversion from an endoscopic to an open approach may be necessary in about 10% of patients. Overall, 90% of patients have long-term satisfactory results after endoscopic stapling.17 Complications of endoscopic procedures include bleeding, perforation, and leaks, but these are uncommon if a stapler-assisted technique is used. Flexible fiberoptic endoscopic techniques also have a role in the treatment of Zenker’s diverticula. Rigid diverticuloscopes cannot be used in patients who have limited neck extension or limited ability to open their mouth.15 Flexible fiberoptic techniques do not require general anesthesia. To improve exposure and stabilize the diverticulum, a transparent cap may be attached to the tip of the endoscope.20-22 Another device developed for this purpose is a soft diverticuloscope used as an overtube.23,24 A variety of techniques are used to perform the endoscopic myotomy, including needle knife, argon plasma coagulation, and monopolar forceps. Several sessions may be required to achieve an adequate myotomy. Complications of fiberoptic flexible techniques include cervical and mediastinal air dissection, which are common, as well as perforation and mediastinitis. Persistent or recurrent symptoms have been reported in 10% to 15% of cases.

DIVERTICULA OF THE ESOPHAGEAL BODY Cause and Pathogenesis

Diverticula of the esophageal body are most commonly located in the middle or lower third of the esophagus (Fig. 23-3). Diverticula located near the diaphragmatic hiatus are called epiphrenic diverticula (Fig. 23-4). Con­ genital bronchopulmonary-foregut malformations can communicate with the esophagus and present as esophageal diverticula.25 Traction diverticula are often related to mediastinal inflammation associated with tuberculosis

Figure 23-4.  Barium esophagogram showing an epiphrenic diverticulum that is immediately above the stomach. In this projection, the diverticulum may be confused with a hiatal hernia. (Courtesy of Dr. Charles A. Rohrman and Dr. Charles E. Pope.)

and histoplasmosis. Enlarged mediastinal lymph nodes from lung malignancies can also lead to traction diverticula. Epiphrenic diverticula are acquired. About 80% are associated with motility disorders, such as achalasia, diffuse esophageal spasm, hypertensive lower esophageal sphincter, and nonspecific motility disorders.26-28 Epiphrenic diverticula have been reported as a complication of obesity surgery.29

Clinical Features and Diagnosis

Congenital and traction diverticula are usually asymptomatic. Bronchopulmonary fistulae can develop, leading to cough, pneumonia, and recurrent bronchopulmonary infections.30 Midbody and distal esophageal diverticula are also

373

374

Section IV  Topics Involving Multiple Organs usually asymptomatic. If symptoms are not present at diagnosis, they rarely occur during follow-up. When symptoms occur, the most common are dysphagia, food regurgitation, reflux, weight loss, and chest discomfort.31 Dysphagia may be caused by an underlying motility disorder or by extrinsic compression of the esophagus by a large diverticulum, with preferential filling.32,33 Regurgitation and aspiration of the contents of the diverticulum may complicate induction of anesthesia. Perforation may occur during nasogastric intubation or upper gastrointestinal endoscopy. An epiphrenic diverticulum may be mistaken for a diaphragmatic hernia or duplication cyst on chest radiography. Diagnosis is best made by barium swallow, which serves to visualize the diverticulum and localizes it more precisely than endoscopy (see Fig. 23-4). Squamous cell carcinoma has been reported in epiphrenic diverticula.34 As with Zenker’s diverticula, accumulation of radioactive iodine tracer in esophageal diverticula has been mistaken for metastatic thyroid cancer.35

Treatment and Prognosis

Asymptomatic diverticula of the esophagus need no treatment. Only those patients with symptoms clearly related to their diverticula should be treated. Preoperative endoscopy and manometry are advisable. It can be difficult to pass a manometry catheter beyond the diverticulum and into the stomach, but documentation of achalasia or diffuse esophageal spasm is helpful for guiding treatment.36 Large diverticula may be inverted or resected. Given the high prevalence of associated motility disorders, esophageal myotomy is performed in most if not all cases.27,28,37 Small diverticula can be treated by myotomy without resection. To prevent gastroesophageal reflux, a nonobstructing fundoplication is usually done. Surgery for esophageal diverticula may be done by open, laparoscopic, or laparoscopic combined with thoracoscopic techniques.27,28,37,38 Epiphrenic diverticula are often amenable to a laparoscopic approach, which has the advantages of a short hospital stay and a quick return to normal activities.

Figure 23-5.  Esophageal intramural pseudodiverticula. A, Barium esopha­ gogram showing small outpouchings. B, Endoscopic view. The tiny openings of the pseudodiverticula are seen in this patient, who also has a distal esophageal peptic stricture.

A

ESOPHAGEAL INTRAMURAL PSEUDODIVERTICULA Esophageal intramural pseudodiverticula (EIP) are flaskshaped outpouchings from the lumen of the esophagus, ranging in size from 1 to 4 mm. They may occur in any segment of the esophagus or in the entire esophagus. Only about 200 cases have been reported, but EIP are more common than the small number of published case reports would imply. EIP have been demonstrated in about 1% of barium swallow studies.39 In autopsy studies, the incidence has been reported to be as high as 55%.40,41

Cause and Pathogenesis

EIP are abnormally dilated ducts of submucosal glands. They are thought to be acquired, and are often associated with conditions that cause chronic esophageal inflam­ mation. The ducts may become dilated because of peri­ ductal inflammation or fibrosis.40,42 Gastroesophageal reflux, chronic candidiasis, previous caustic ingestion, esophageal cancer, and a single case of eosinophilic esophagitis have all been associated with EIP.39,43-45 Esophageal strictures are also commonly associated with EIP.41,46 Marked thickening of the esophageal wall has been noted in some cases by CT or endoscopic ultrasound.47

Clinical Features and Diagnosis

Patients are found to have EIP most commonly in their sixth or seventh decades. The condition is slightly more common in men than in women.41 EIP are discovered most commonly on a barium swallow done for dysphagia or heartburn (Fig. 23-5A). EIP may also be an incidental finding in patients without related symptoms. EIP are usually segmental in distribution, but may be diffuse. Stricture is noted in most cases.41 Tracking or communication between adjacent pseudodiverticula is not uncommon if it is looked for carefully.48 The differential diagnosis on barium swallow examination includes esophageal ulceration. Cancer must be excluded by upper endoscopy if a stricture is present.

B

Chapter 23  Diverticula of the Pharynx, Esophagus, Stomach, and Small Intestine Although the endoscopic appearance of EIP is characteristic (see Fig. 23-5B), the openings of EIP are small and are often missed. EIP located within an area of stricture are particularly difficult to see at endoscopy. Symptoms, when present, are generally related to the associated condition, such as stricture, cancer, acid reflux, or candidiasis, rather than to the EIP. There have been case reports of perforation of EIP leading to mediastinitis.49

During endoscopy, juxtacardiac diverticula are best seen on a retroflexed view. Juxtacardiac diverticula may be missed on barium study unless lateral views are taken. On CT scans, they may appear as air- or contrast-filled supra­renal masses and can be mistaken for a necrotic adrenal mass.61,62 Intramural diverticula do not usually cause symptoms. They are often mistaken for ulcers on barium studies.

Treatment and Prognosis

Treatment and Prognosis

EIP per se do not require treatment. Treatment should be directed at the underlying condition, such as stricture, acid reflux, or candidiasis. EIP may persist, even if treatment relieves symptoms.50

Intramural diverticula require no intervention. Juxtacardiac diverticula almost never need treatment. A clear association with a specific symptom complex should be firmly established before considering resection. Complications such as ulceration, bleeding, and cancer are very rare. If bleeding occurs, treatment may be challenging. The use of hemoclips

GASTRIC DIVERTICULA Gastric diverticula are found in less than 1% of upper gastrointestinal x-rays or autopsies.51 Juxtacardiac diverticula make up 75% of all gastric diverticula. These are most often located near the gastroesophageal junction, on the posterior aspect of the lesser curvature.52 They are most commonly found in middle-aged patients, although cases have been reported in children and adolescents.53-55 They range in size from 1 to 3 cm in diameter (Fig. 23-6). Intramural or partial gastric diverticula are formed by the projection of the mucosa of the stomach through the muscularis. These diverticula are found most commonly on the greater curvature (Fig. 23-7).56,57 Deformities caused by peptic ulcers or other inflammatory processes can resemble prepyloric diverticula on barium studies or at endoscopy. Gastric diverticula have been reported as complication of obesity surgery.58,59

Clinical Features and Diagnosis

Juxtacardiac diverticula are almost always asymptomatic. Rarely, patients may complain of pain or dyspepsia attributable to a diverticulum. Reproduction of pain by probing the diverticulum with a biopsy forceps during upper gastro­ intestinal endoscopy has been reported.60

A

Figure 23-6.  Juxtacardiac gastric diverticulum. This wide-mouthed diver­ ticulum was seen on a retroflexed view of the cardia. The mucosa within the diverticulum was normal.

B

Figure 23-7.  Inverted gastric diverticulum. A, On endoscopy, the diverticulum protrudes into the lumen of the antrum, thus simulating a pancreatic rest or mass. B, A biopsy forceps used to evert the mucosa confirmed that the protrusion is an inverted diverticulum.

375

376

Section IV  Topics Involving Multiple Organs has been reported.63 Bleeding that cannot be controlled with endoscopic techniques may require referral for surgery. Perforation or cancer may be treated by diverticulectomy or partial gastrectomy, respectively. If a patient with a juxtacardiac diverticulum requires referral for surgery, it may be prudent to mark the diverticulum with India ink, because localization by the surgeon can be difficult. Laparoscopic diverticulectomy has been reported.54

DUODENAL DIVERTICULA Duodenal diverticula can be extraluminal or intraluminal.

EXTRALUMINAL DIVERTICULA Cause and Pathogenesis

Extraluminal duodenal diverticula are noted in about 5% of upper gastrointestinal x-rays, and in about 25% of endoscopic retrograde cholangiopancreatography (ERCP) studies or autopsies.64,65 They are thought to be acquired. They arise in an area of the duodenal wall where a vessel penetrates the muscularis or where the dorsal and ventral pancreas fuse in embryologic development. Approximately 75% are located within 2 cm of the ampulla and are termed juxtapapillary diverticula (JPD).

Clinical Features and Diagnosis

Duodenal diverticula are typically diagnosed on upper gastrointestinal x-rays (Fig. 23-8). They are easily missed on endoscopy unless a side-viewing endoscope is used. The sensitivity of computed tomography (CT) and magnetic resonance imaging (MRI) for duodenal diverticula is low.66-69 If a diverticulum is suspected on CT or MRI, the diagnosis can be clarified by having the patient drink water and repeating the scan. The presence of an air-fluid level in

the structure will clarify the diagnosis. A duodenal diverticulum may be mistaken for a pancreatic pseudocyst, peripancreatic fluid collection, cystic pancreatic tumor, hypermetabolic mass, or distal common bile duct stone on ultrasound, CT, MRI, or positron emission tomography (PET)-CT.70-74 Although extraluminal duodenal diverticula are relatively common, only a few are associated with clinical problems. Problems associated with extraluminal duodenal diverticula include perforation or diverticulitis, bleeding, acute pancreatitis, and common bile duct stones. Duodenal diverticulitis may present as a free or contained perforation. Patients present with pain in the upper abdomen, often radiating to the back, and may have signs and symptoms of sepsis. An abdominal CT scan may reveal thickening of the duodenum, retroperitoneal air, phlegmon, or abscess. The findings are usually nonspecific and the diagnosis is often not made until exploratory laparotomy.69,75 Bleeding has been reported from Dieulafoy-like lesions or ulcers within diverticula.76-78 Bleeding from duodenal diverticula may be very difficult to diagnose, requiring examination with a side-viewing endoscope or angiography. In some patients, the site of bleeding is discovered only at laparo­ tomy and duodenotomy. Patients with multiple duodenal diverticula may develop bacterial overgrowth and malabsorption (see Chapter 102).79 Juxtapapillary diverticula have been associated with common bile duct stones, cholangitis, and recurrent pancreatitis.80-85 The presence of a juxtapapillary diverticulum has been shown to lead to sphincter of Oddi dysfunction.86 Delayed emptying of the common bile duct may occur, even after sphincterotomy. Stasis within diverticula can result in local bacterial overgrowth, favoring deconjugation of bilirubin and thus increasing the risk of primary common bile duct stones.81 JPD do not appreciably increase the difficulty of cannulation or the risk of complications at ERCP unless the papilla is not visible within the diverticulum.83,87 Several techniques have been described to overcome difficulties associated with an ampulla situated deep within a diverticulum.88,89

Treatment and Prognosis

Extraluminal duodenal diverticula rarely require therapeutic intervention. Resection of duodenal diverticula should never be done for vague abdominal complaints. Bleeding, diverticulitis, and perforation are the most common pro­ blems associated with duodenal diverticula. Endoscopic control of bleeding from diverticula has been accomplished using various techniques, including bipolar cautery, epinephrine injection, and hemoclips.76-78,90 If the diagnosis is not made preoperatively, surgical control of bleeding can be accomplished through a duodenotomy. Damage to the pancreatic and biliary ducts may occur during surgery in patients with periampullary diverticula. Most patients with perforation or diverticulitis undergo laparotomy for diagnosis. The usual surgical treatment is drainage and resection of the involved diverticulum, if feasible. If the diagnosis is made preoperatively, successful conservative therapy by percutaneous drainage and anti­ biotics is possible.75,91

INTRALUMINAL DIVERTICULA

Figure 23-8.  Upper gastrointestinal radiograph showing multiple large duodenal diverticula.

Intraluminal duodenal diverticula (windsock diverticula) are single saccular structures that originate in the second portion of the duodenum. They are connected to the entire circumference or only to part of the wall of the duodenum and may project as far distally as the fourth part of the

Chapter 23  Diverticula of the Pharynx, Esophagus, Stomach, and Small Intestine

Ampulla of Vater

A

Ampulla of Vater

Figure 23-9.  Intramural duodenal diverticulum (windsock diverticulum). A, The diverticulum is attached to the entire duodenal circumference. B, The diverticulum is attached to only part of the duodenal circumference.

B

duodenum. There is often a second opening located eccentrically in the sac (Fig. 23-9). Both sides of the diverticulum are lined by duodenal mucosa. Fewer than 100 cases have been reported.

Cause and Pathogenesis

During early fetal development, the duodenal lumen is occluded by proliferating epithelial cells and later recanalized (see Chapter 47). Abnormal recanalization may lead to a duodenal diaphragm or web. An incomplete or fenestrated diaphragm may not produce obstructive symptoms in childhood. Over time, peristaltic stretching may transform the diaphragm into an intraluminal diverticulum.

Clinical Features and Diagnosis

Intraluminal diverticula may become symptomatic at any age. The most common symptoms are those of incomplete duodenal obstruction.92,93 Obstruction may be precipitated by retention of vegetable material or foreign bodies within the diverticulum. In one report, a 41-year-old man was found to have two marbles, swallowed during childhood, retained in an intraluminal diverticulum.94 Pancreatitis and bleeding have also been reported.95,96 The typical radiographic appearance is that of a barium-filled globular structure of variable length, originating in the second portion of the duodenum, with its fundus extending into the third portion, and outlined by a thin, radiolucent line. The CT appearance has been reported as a ring-like soft tissue density in the lumen of the second portion of the duodenum, outlined with oral contrast and containing oral contrast and a small amount of air (halo sign).97 At endoscopy, an intraluminal diverticulum is a sac-like structure with an eccentric aperture or a large, soft, polypoid mass if the diverticulum is inverted orad.96,98 Endoscopic diagnosis may be difficult. A long sac may be mistaken for the duodenal lumen, whereas an inverted diverticulum may be mistaken for a large polyp. Gastric retention or dilation of the duodenal bulb may result from chronic partial obstruction caused by the diverticulum.

JEJUNAL DIVERTICULA Diverticula of the small bowel (apart from duodenal and Meckel’s diverticula) are most commonly found in the proximal jejunum. About 80% of jejunoileal diverticula arise in

the jejunum, 15% in the ileum, and 5% in both; small bowel diverticula have been found in about 0.5% to 5% of small bowel x-rays and autopsies.99,100 They are commonly multiple and can vary from a few millimeters to 10 cm in length. They are usually located on the mesenteric border of the small bowel. Small bowel diverticula generally lack a true muscular wall and are considered to be acquired.

Cause and Pathogenesis

The cause of jejunoileal diverticula is largely unknown. Many patients have an underlying intestinal motility disorder. Periodic elevated intraluminal pressures can lead to herniation through areas of weakness at the mesenteric border where blood vessels penetrate the muscularis. Visceral neuropathies and myopathies, including progressive systemic sclerosis, can lead to chronic atrophy and fibrosis of the intestinal wall, with resultant herniation and diverticula formation.101

Clinical Features and Diagnosis

Jejunal diverticula are best diagnosed by upper gastrointestinal radiography with small bowel follow-through. They may also be found by CT.102 Jejunal diverticula most commonly occur on the mesenteric border of the bowel, in contrast to Meckel’s diverticula, which occur on the antimesenteric border. Jejunal diverticula have been discovered by doubleballoon enteroscopy, and diverticular bleeding has been treated with this technique.103,104 On the other hand, if a patient is known or discovered to have small bowel diverticula, it may be prudent to avoid enteroscopy because of the risk of perforation.105 Many cases of jejunoileal diverticulosis are asymptomatic or associated with nonspecific symptoms for which patients may not seek medical attention. About 40% of cases are discovered incidentally.100 In 1881, Sir William Osler wrote about a patient with jejunal diverticula who, for years, “had suffered much from loud rumbling noises in his belly, particularly after each meal. So loud were they that it was his habit, shortly after eating, to go out and take a walk to keep away from people, as the noises could be heard at some distance.” Various symptoms and clinical problems may occur with jejunal diverticula.106 The most common clinical features are recurrent abdominal pain, early satiety, and bloating. Loud borborygmi and intermittent diarrhea may occur. These symptoms are likely caused by an underlying motility disorder. Malabsorption may result from associated bacte-

377

378

Section IV  Topics Involving Multiple Organs rial overgrowth (see Chapters 101 and 102).100,107,108 Patients with jejunal diverticulosis and severe dysmotility develop a syndrome of intestinal pseudo-obstruction. Patients with this clinical picture may periodically have free intraperitoneal air (pneumoperitoneum) without overt perforation. If such patients are otherwise well, they should be carefully observed. Laparotomy is often not necessary. Bleeding from small bowel diverticula may be difficult to localize.109,110 If a source of bleeding is discovered in the small bowel at angiography, it may be useful to leave a small catheter within the feeding vessel as the patient is taken to the operating room. When the patient is explored, a small amount of dye can be injected through the catheter, staining the involved bowel. This may help the surgeon localize an otherwise obscure lesion. Capsule endoscopy has been reported to discover small bowel diverticula.111,112 Diverticulitis may result in free perforation or an abscess contained within the mesentery.113 Preoperative diagnosis is difficult. The finding of an inflammatory mass in the mesentery should raise the possibility of small bowel diverticula.114 Because jejunal diverticula usually project into the mesentery, they can be difficult to detect, even at surgery. Large enteroliths can form in jejunal diverticula and lead to erosion, with bleeding, diverticulitis, perforation, or intestinal obstruction.115-117 If a small bowel volvulus is found in an adult, small bowel diverticulosis should be considered, because it appears that there is an association between the two conditions.118

Treatment and Prognosis

There is no specific treatment for symptoms related to intestinal dysmotility. The use of oral antibiotics to treat associated bacterial overgrowth may lead to improvement in bloating and diarrhea, as well as malabsorption (see Chapters 101 and 102). In patients with bleeding, perforation, or diverticulitis, limited surgical resection of the section of bowel with the offending diverticulum should be the goal, but this may be difficult to localize with precision.119,120 In patients with symptoms of chronic intestinal pseudo-obstruction, surgery should generally be avoided, although carefully selected patients may benefit.121 If a long segment of bowel is resected in an attempt to remove all the diverticula, the patient may

not only be left with a short bowel syndrome (see Chapter 103), but underlying dysmotility may also involve the remaining intestine, compromising its function and leading to severe disability.

KEY REFERENCES

Aly A, Devitt PG, Jamieson GG. Evolution of surgical treatment for pharyngeal pouch. Br J Surg 2004; 91:657-64. (Ref 17.) Ames JT, Federle MP, Pealer KM. Perforated duodenal diverticulum: Clinical and imaging findings in eight patients. Abdom Imaging 2009; 34:135-9. (Ref 69.) Christiaens P, De RW, Van OA, et al. Treatment of Zenker’s diverticulum through a flexible endoscope with a transparent oblique-end hood attached to the tip and a monopolar forceps. Endoscopy 2007; 39:137-40. (Ref 21.) Costamagna G, Iacopini F, Tringali A, et al. Flexible endoscopic Zenker’s diverticulotomy: Cap-assisted technique vs. diverticuloscopeassisted technique. Endoscopy 2007; 39:146-52. (Ref 24.) Ferreira LE, Simmons DT, Baron TH. Zenker’s diverticula: Pathophy­ siology, clinical features, and flexible endoscopic management. Dis Esophagus 2008; 21:1-8. (Ref 5.) Fintelmann F, Levine MS, Rubesin SE. Jejunal diverticulosis: Findings on CT in 28 patients. AJR Am J Roentgenol 2008; 190:1286-90. (Ref 102.) Liu CY, Chang WH, Lin SC, et al. Analysis of clinical manifestations of symptomatic acquired jejunoileal diverticular disease. World J Gastroenterol 2005; 11:5557-60. (Ref 106.) Macari M, Faust M, Liang H, Pachter HL. CT of jejunal diverticulitis: Imaging findings, differential diagnosis, and clinical management. Clin Radiol 2007; 62:73-7. (Ref 114.) Panteris V, Vezakis A, Filippou G, et al. Influence of juxtapapillary diverticula on the success or difficulty of cannulation and complication rate. Gastrointest Endosc 2008; 68:903-10. (Ref 87.) Rabenstein T, May A, Michel J, et al. Argon plasma coagulation for flexible endoscopic Zenker’s diverticulotomy. Endoscopy 2007; 39: 141-5. (Ref 22.) Seaman DL, de la Mora LJ, Gostout CJ, et al. A new device to simplify flexible endoscopic treatment of Zenker’s diverticulum. Gastrointest Endosc 2008; 67:112-15. (Ref 23.) Sen P, Kumar G, Bhattacharyya AK. Pharyngeal pouch: Associations and complications. Eur Arch Otorhinolaryngol 2006; 263:463-8. (Ref 1.) Tedesco P, Fisichella PM, Way LW, Patti MG. Cause and treatment of epiphrenic diverticula. Am J Surg 2005; 190:891-4. (Ref 27.) Varghese TK Jr, Marshall B, Chang AC, et al. Surgical treatment of epiphrenic diverticula: A 30-year experience. Ann Thorac Surg 2007; 84:1801-9. (Ref 37.) Vogelsang A, Preiss C, Neuhaus H, Schumacher B. Endotherapy of Zenker’s diverticulum using the needle-knife technique: Long-term follow-up. Endoscopy 2007; 39:131-6. (Ref 20.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

24 Abdominal Hernias and Gastric Volvulus D. Rohan Jeyarajah and William V. Harford, Jr.

CHAPTER OUTLINE Diaphragmatic Hernias  379 Gastric Volvulus  383 Inguinal and Femoral Hernias  385 Other Ventral Hernias  388 Incisional Hernias  389

A hernia is a protrusion of an organ or structure into an opening or pouch. Abdominal wall hernias protrude through the retaining walls of the abdomen and have two parts, the orifice or defect in the aponeurotic wall of the abdomen, and the hernia sac, which consists of peritoneum and abdominal contents. Internal hernias are contained within the abdominal cavity and do not always have a hernia sac. Abdominal wall hernias are external if the sac protrudes through the abdominal wall or interparietal if the sac is contained within the abdominal wall. Hernias are reducible when the protruding organ can be returned to the abdomen and irreducible or incarcerated when it cannot. A hernia is strangulated when the vascular supply of the protruding organ is compromised and, as a consequence, the organ becomes ischemic or necrotic. An incarcerated hernia is generally repaired because there is danger of strangulation. There can be difficulty determining whether a hernia is incarcerated or strangulated; therefore, all incarcerated hernias are treated with surgical intervention. In a Richter’s hernia, only one side of the bowel (usually the antimesenteric) protrudes through the hernia orifice. Thus, strangulation may occur without intestinal obstruction. Richter’s hernias may occur in various locations.

DIAPHRAGMATIC HERNIAS Diaphragmatic hernias may occur through the esophageal hiatus, through other congenital openings (such as the foramina of Bochdalek or Morgagni), or through posttraumatic defects. Most diaphragmatic hernias are sliding hernias of the stomach through the esophageal hiatus.

Cause and Pathogenesis

Sliding hiatal hernias (type 1) occur when the gastroesophageal junction and some portion of the stomach are displaced above the diaphragm. The orientation of the stomach axis is unchanged. The cause of sliding hiatal hernias is not known. The frequency of sliding hiatal hernias increases with age. The phrenoesophageal membrane anchors the gastroesophageal junction to the diaphragm. Hiatal hernias

Epigastric and Umbilical Hernias  390 Spigelian Hernias  390 Pelvic and Perineal Hernias  391 Lumbar Hernias  392 Internal Hernias  392

may be caused by age-related deterioration of this membrane, combined with normal positive intra-abdominal pressure and with traction of the esophagus on the stomach as the esophagus shortens during swallowing. Paraesophageal hernias (type 2) occur when the stomach protrudes through the esophageal hiatus alongside the esophagus. The gastroesophageal junction typically remains in a normal position at the level of the diaphragm because there is preservation of the posterior phrenoesophageal ligament with normal anchoring of the gastroesophageal junction.1 The entire stomach can pass into the chest (Fig. 24-1A). Gastric volvulus (see later) may result. The omentum, colon, or spleen may also herniate. Patients with paraesophageal hernias may have a congenital defect in the diaphragmatic hiatus anterior to the esophagus. Most paraesophageal hernias contain a sliding hiatal component in addition to the paraesophageal component, and are thus mixed diaphragmatic hernias (type 3; see Fig. 24-1B).2 A barium study is often obtained to diagnose these defects. Very specific questioning of the radiologist with respect to two critical points will allow the clinician to make an accurate diagnosis: 1. Does the gastroesophageal junction lie at or above the hiatus? 2. Does the stomach or any other viscera lie above the gastroesophageal junction? For example, if the gastroesophageal junction is above the hiatus and there is stomach above it, the patient has a type 3 defect. Congenital diaphragmatic hernias result from failure of fusion of the multiple developmental components of the diaphragm (Fig. 24-2). The diaphragm is derived from the septum transversum (separating the peritoneal and pericardial spaces), the mesentery of the esophagus, the pleuroperitoneal membranes, and muscle of the chest wall. Morgagni hernias form anteriorly at the sternocostal junctions of the diaphragm and Bochdalek hernias posterolaterally at the lumbocostal junctions of the diaphragm (Fig. 24-3).3 Bochdalek hernias manifest immediately after birth and are commonly associated with pulmonary hypoplasia. Post-traumatic diaphragmatic hernias are caused by blunt trauma (e.g., motor vehicle accidents) in about 80% of cases,

379

380

Section IV  Topics Involving Multiple Organs

A Figure 24-3.  Bochdalek hernia. This plain chest radiograph (lateral view) shows a Bochdalek hernia as a small opacity in the posterior chest at the level of the diaphragm (arrows). (Courtesy of Dr. Nahid Eshaghi, Dallas, Tex.)

B Figure 24-1.  A, Paraesophageal hernia. This barium radiograph shows a paraesophageal hernia complicated by an organoaxial volvulus of the stomach (see Fig. 24-6). The gastroesophageal junction remains in a relatively normal position below the diaphragm (arrow). The entire stomach has herniated into the chest and the greater curvature has rotated anteriorly and superiorly. B, Combined hernia in a different patient. A retroflexed endoscopic view of the proximal stomach shows the endoscope traversing a sliding hiatal hernia adjacent to a large paraesophageal hernia. (A courtesy of Dr. Herbert J. Smith, Dallas, Tex.)

and to penetrating trauma (e.g., stab wounds or gunshots) in the remainder. During blunt trauma, abrupt changes in intra-abdominal pressure may lead to large rents in the diaphragm. Penetrating injuries often cause only small lacerations. Blunt trauma is more likely than penetrating trauma to lead eventually to herniation of abdominal contents into the chest because the defect is usually larger. The right hemidiaphragm is somewhat protected by the liver during blunt trauma. Thus, 70% of diaphragmatic injuries from blunt trauma occur on the left side.4-6 Diaphragmatic injury may not result in immediate herniation but, with time, normal negative intrathoracic pressure may lead to gradual enlargement of a small diaphragmatic defect and protrusion of abdominal contents through the defect.7 Stomach, omentum, colon, small bowel, spleen, and even kidney may be found in a post-traumatic diaphragmatic hernia.

Incidence and Prevalence

1

1

2

3

3

Figure 24-2.  Congenital diaphragmatic hernias. Diagram of diaphragm viewed from below with areas of potential herniation shown. 1, Sternocostal foramina of Morgagni anteriorly. 2, Esophageal hiatus. 3, Lumbocostal foramina of Bochdalek posteriorly. Arrows indicate direction of herniation.

In the United States and Canada, a large proportion of adults undergoing upper gastrointestinal barium radiographs are found to have a small hiatal hernia. About 90% to 95% of hiatal hernias found by radiograph are sliding (type 1) hernias; the remainder are paraesophageal (type 2) or mixed (type 3).2 Most sliding hiatal hernias are small and of little clinical significance. Patients with symptomatic paraesophageal hernias are most often middle-aged to older adults. Congenital hernias occur in about 1/2,000 to 10,000 births.8,9 Those hernias manifesting in neonates are most often Bochdalek hernias. With the routine use of prenatal ultrasound, congenital diaphragmatic hernias (CDHs) can be discovered in the prenatal period. The presence of intraabdominal contents in the chest during fetal development results in significant hypoplasia of the lung. It is the degree of pulmonary dysfunction, not the presence of the hernia per se, that determines the child’s prognosis. Prenatal measures are then taken to prepare for the pulmonary hypoplasia that invariably accompanies a large CDH. Only a few Bochdalek hernias are first discovered in adulthood.10 Bochdalek hernias occur on the left side in about 80% of cases (see Fig. 24-3).11 Right-sided Bochdalek hernias usually

Chapter 24  Abdominal Hernias and Gastric Volvulus contain liver in the right chest. Morgagni hernias make up about 2% to 3% of surgically treated diaphragmatic hernias.12,13 Although thought to be congenital, they usually manifest in adults and occur on the right side in 80% to 90% of cases. The incidence of post-traumatic diaphragmatic hernia is uncertain. Diaphragmatic injury occurs in about 5% of patients with multiple traumatic injuries.5,6

Clinical Features and Diagnosis

Many patients with small simple sliding hiatal hernias are asymptomatic. The main clinical significance of the sliding hiatal hernia is its contribution to gastroesophageal reflux (see Chapter 43). In addition to heartburn and regurgitation, patients with large sliding hiatal hernias may complain of dysphagia or discomfort in the chest or upper abdomen. In a prospective, population-based study, the risk of iron deficiency anemia in adults was found to be increased by almost threefold.14 With chest radiography, a hiatal hernia may be noted as a soft tissue density or an air-fluid level in the retrocardiac area. Hiatal hernias are most often diagnosed on upper gastrointestinal barium studies. Computed tomography (CT) scanning can demonstrate the proximal stomach above the diaphragmatic hiatus. At endoscopy, the gastroesophageal junction is noted to be proximal to the impression of the diaphragm. Cameron lesions or linear erosions may develop in patients with sliding hiatal hernias, particularly large hernias (see Chapters 19 and 52). These mucosal lesions are usually found on the lesser curve of the stomach at the level of the diaphragmatic hiatus (Fig. 24-4). This is the location of the rigid anterior margin of the hiatus formed by the central tendon of the diaphragm. Mechanical trauma, ischemia, and peptic injury have been proposed as the cause of these lesions. The prevalence of Cameron lesions in patients with hiatal hernias who undergo endoscopy has been reported to be about 5%, with the highest prevalence in the largest hernias. Cameron lesions may cause acute or chronic upper gastrointestinal bleeding.15 The presence of Cameron lesion(s) and occult gastrointestinal bleeding may prompt repair of the hiatal defect to aid healing of this defect. Patients with paraesophageal or mixed hiatal hernias are rarely completely asymptomatic if closely questioned. About half of patients with paraesophageal hernias have gastroesophageal reflux.2,16,17 Other symptoms include dysphagia, chest pain, vague postprandial discomfort, and shortness of breath, and a substantial number of patients will have chronic gastrointestinal blood loss.18-20 If the hernia is complicated by gastric volvulus, acute abdominal

Figure 24-4.  Cameron lesion. A large hiatal hernia is seen on endoscopic retroflexed view, with a Cameron lesion at the level of the diaphragmatic hiatus.

pain and retching will occur, often progressing rapidly to a surgical emergency (see later, “Gastric Volvulus”). A paraesophageal or mixed hiatal hernia may be seen on a chest radiograph as an abnormal soft tissue density (often with a gas bubble) in the mediastinum or left chest. Upper gas­ trointestinal radiography is the best diagnostic study (see Fig. 24-1A). CT scanning can demonstrate that part of the stomach is in the chest. Lack of filling the gastric lumen with contrast or gastric wall thickening with pneumatosis can increase suspicion for a volvulus and associated gastric necrosis. Paraesophageal hernias are usually obvious on upper gastrointestinal endoscopy (see Fig. 24-1B), but the paraesophageal com­ponent of a large mixed hernia may be missed. Endoscopy may be difficult if the hernia is asso­ ciated with gastric volvulus. The clinical presentation of congenital diaphragmatic hernias varies greatly, from death in the neonatal period to an asymptomatic serendipitous finding in adults. Newborns with Bochdalek hernia have respiratory distress, absent breath sounds on one side of the chest, and a scaphoid abdomen.11 Most of these neonates are diagnosed in utero with routine use of prenatal ultrasound. Serious chromosomal anomalies are found in 30% to 40% of cases; the most common of these are trisomy 13, 18, and 21. Pulmonary hypoplasia occurs on the side of the hernia, but some degree of hypoplasia may also occur in the contralateral lung. Pulmonary hypertension is common. The major causes of mortality in infants with Bochdalek hernias are respiratory failure and associated anomalies. Prenatal diagnosis may be made sonographically by visualizing stomach or loops of bowel in the chest. The diagnosis of congenital diaphragmatic hernia in the prenatal period will make the pregnancy high risk. Pediatric surgeons are available at delivery to initiate extracorporeal membrane oxygenation (ECMO), because the neonate will commonly need complete cardiopulmonary bypass because of the lack of pulmonary function. The hernia will then be repaired using a large mesh prosthesis once the child has stabilized from a pulmonary standpoint. In older children and adults, a Bochdalek hernia may manifest as an asymptomatic chest mass. The differential diagnosis includes mediastinal or pulmonary cyst or tumor, pleural effusion, or empyema. Symptoms, when present, are caused by herniation of the stomach, omentum, colon, or spleen. About half of adult patients present with acute emergencies caused by incarceration. Gastric volvulus is common (see later). Other patients may have chronic intermittent symptoms, including chest discomfort, shortness of breath, dysphagia, nausea, vomiting, and/or constipation. The diagnosis may be suspected on a chest radiograph, particularly a lateral view. The key finding is a posterior chest mass because the defect of Bochdalek is posterior (see Fig. 24-3) as opposed to the Morgagni defect, which is anterior (Fig. 24-5). The diagnosis may be confirmed by barium upper gastrointestinal radiography or CT scanning.8,10,11 Morgagni hernias are most likely to manifest in adult life. They may contain omentum, stomach, colon, or liver. Bowel sounds may be heard in the chest if bowel has herniated through the defect. As with Bochdalek hernias, the diagnosis is often made by chest radiography, particularly the lateral view, because Morgagni hernias are anterior (see Fig. 24-5A and B). The contents of the hernia can be confirmed with barium radiography or CT scanning (see Fig. 24-5C and D). The differential diagnosis is similar to that of Bochdalek hernias. Many patients have no symptoms or nonspecific symptoms, such as chest discomfort, cough, dyspnea, or upper abdominal distress. Gastric, omental, or intestinal

381

382

Section IV  Topics Involving Multiple Organs

B

A

D

C Figure 24-5.  Morgagni hernia. A, A mass is noted in the right chest on this chest radiograph (posteroanterior view). B, Lateral chest radiograph shows that the mass is in the anterior chest. C, Barium enema shows that a portion of the transverse colon is the hernia (top left). D, Computed tomographic scan shows a contrast-filled colon in the right anterior chest (11 o’clock position).

incarceration with obstruction and/or ischemia may cause acute symptoms.12,13 Post-traumatic diaphragmatic hernias cause respiratory or abdominal symptoms. After serious trauma, rupture of the diaphragm is often masked by other injuries.4 Penetrating injuries between the fourth intercostal space and the umbilicus should raise the level of suspicion of a diaphragmatic injury. Respiratory or abdominal symptoms manifesting several days to weeks after injury should suggest the possibility of a missed diaphragmatic injury. The diaphragm must be closely inspected to detect injury at the time of exploratory laparotomy because these injuries can be easily missed. Careful examination of the chest radiograph or CT scan is important, but is diagnostic in only half of cases. The use of rapid helical CT, especially with sagittal recon-

struction, has facilitated the diagnosis.5,6 In patients on ventilatory support after trauma, positive intrathoracic pressure may prevent herniation through a diaphragmatic tear. However, on attempted ventilator weaning, herniation may occur, causing respiratory compromise. Symptoms may also manifest long after injury. Delays of more than 10 years are not uncommon.7 In such cases, the patient may not connect the acute illness with remote trauma.

Treatment and Prognosis

Simple sliding hiatal hernias do not require treatment. Patients with symptomatic giant sliding hiatal hernias, paraesophageal, or mixed hernias should be offered surgery. When closely questioned, most patients with type 2 or 3 hernias will have symptoms.1 In the past, paraesophageal

Chapter 24  Abdominal Hernias and Gastric Volvulus hernias were thought to be a surgical emergency. However, it is now clear that the risk of progression to gastric necrosis is lower than initially believed.17 However, many experts suggest that surgery should be offered to all patients with paraesophageal hernias because some complication will develop in about 30% of patients if left untreated.2,18-20 In general, a selective approach to patients with large para­ esophageal hernias is warranted; those with any symptoms that may be attributable to the hernia should be offered surgical intervention. The extent of the preoperative evaluation needed for paraesophageal hernia repair is controversial. Many surgeons recommend routine preoperative evaluation with esophageal manometry and ambulatory esophageal pH monitoring because of the high prevalence of associated gastroesophageal reflux and esophageal motility disorders. The object of the evaluation is to determine which patients should have a fundoplication and whether to perform a complete or partial wrap. However, complete manometry is frequently not possible in these patients, and anatomic distortions make it difficult to place the pH probe in the correct location, making esophageal pH monitoring unreliable.21-24 The main use of manometry is to ensure that the patient has an excellent primary peristaltic wave rather than to identify the lower esophageal sphincter (LES) pressure. Patients with dysphagia should be studied to ensure that significantly abnormal motility is not present. Many surgeons routinely add a fundoplication to all repairs to prevent postoperative reflux esophagitis and to fix the stomach in the abdomen. However, in patients with motility disorders, the surgeon may elect to perform a loose posterior wrap or simply a gastrostomy or gastropexy to fix the stomach intraabdominally. Addition of gastropexy may, in fact, reduce the recurrence rate after hernia repair.25 The principles of surgery for hiatal or paraesophageal hernias include three main elements: (1) reduction of the hernia from the mediastinum or chest with excision of the hernia sac; (2) reconstruction of the diaphragmatic hiatus with simple closure or use of prosthetic mesh; and (3) fixation of the stomach in the abdomen with a wrap, gastropexy, or gastrostomy tube. These elements can be accomplished laparoscopically or via open operation and may be approached through the abdomen or the chest. Most patients are approached laparoscopically with a shorter hospital stay and less postoperative pain26 and an equivalent risk of recurrence. Reduction of chronic paraesophageal hernias from the chest can be difficult and may be approached through a combined thoracoscopic and abdominal procedure. Injury to the lung can occur with vigorous traction; however, as the diaphragmatic defect is central, rather than peripheral, as in the traumatic defect, intense lung adhesions are usually not present. Resection of the hernia sac can result in violation of the left chest, requiring chest tube placement. Reconstruction of the diaphragm can be performed by placing nonabsorbable sutures anterior or posterior to the esophagus.22,23 The use of prosthetic mesh has resulted in fewer recurrences.27-31 Fixation of the stomach in the abdomen is usually achieved by using a wrap, which provides some bolstering effect at the hiatus to keep the stomach in the abdomen and can reduce postoperative gastroesophageal reflux. Additional use of gastropexy, with suturing of the stomach to the abdominal wall or tube placement, may result in fewer recurrences.25 Patients with sliding hiatal or paraesophageal hernias may have shortening of the esophagus. This makes it difficult to restore the gastroesophageal junction below the diaphragm without tension, a key factor in decreasing recurrence. In such cases, an extra length of neoesophagus can be constructed from the proximal stomach (Colles-

Nissen procedure).32 In this situation, a stapler is fired parallel to the axis of the esophagus along a bougie that is passed into the stomach, creating a lengthened esophagus. Alternatively, transmediastinal dissection of the esophagus for more than 5 cm into the chest will usually result in adequate intra-abdominal length of esophagus, without the need for additional stapling.33 Paraesophageal and mixed hernias can be repaired through the chest or abdomen, with open or laparoscopic techniques.2,18-20,26,34,35 Compared with open repair, laparoscopic repair is associated with less blood loss, fewer overall complications, and shorter hospital stay, and return to normal activities is faster. Long-term results are probably equal with either approach. Potential surgical complications include esophageal and gastric perforation, pneumothorax, and liver laceration; potential long-term complications may include dysphagia if the wrap is too tight or gastroesophageal reflux if the fundoplication breaks down or migrates into the chest. When examined closely, recurrence after paraesophageal hernia repair is 25% to 30%.36,37 However, the clinical impact of a recurrence may be minimal, because most of these patients remain symptom-free. Like other gastric ulcers, Cameron ulcers or erosions are initially treated with antisecretory medication (see Chapter 53). However, Cameron lesions may persist or recur despite antisecretory medication in about one third of patients, in which case surgical repair of the associated hernia may be required.15 The first priority of treatment for infants with Bochdalek hernias is adequate ventilatory support. Newer techniques of ventilation such as high-frequency oscillation and ECMO are very helpful in some cases. Ventilatory support allows infants to be stabilized before diaphragmatic repair. From 39% to 77% of infants survive the neonatal period after repair, but a significant number have long-term neurologic and musculoskeletal problems, and as many as 50% experience gastroesophageal reflux.11 Laparoscopic repair of Bochdalek hernias has been reported.10 Morgagni hernias have been repaired through the chest or abdomen, using open, thoracoscopic, and/or laparoscopic techniques.12,13,38-40 Acute diaphragmatic ruptures may be approached from the abdomen during exploratory laparotomy or through the chest. Diagnostic laparoscopy has been used in patients who are thought to have a high risk of diaphragmatic injury (e.g., after a stab wound to the lower chest). Chronic posttraumatic diaphragmatic hernias may be associated with extensive adhesions and lack of a peritoneal hernia sac. In such cases, repair is best done through the chest or by a combined thoracoscopic-abdominal approach, although laparoscopic repair has been reported.5,41

GASTRIC VOLVULUS Gastric volvulus occurs when the stomach twists on itself. Gastric volvulus rarely occurs unless there is an associated diaphragmatic hernia. Paré described the first case of gastric volvulus in 1579 in a patient who had a diaphragmatic injury from a sword wound. Gastric volvulus may be transient and produce few symptoms, or it may lead to obstruction and ischemia.

Cause and Pathogenesis

The stomach is normally fixed in position by ligamentous attachments to the spleen, liver, and diaphragm. When there is normal intestinal rotation, the duodenum is fixed to the retroperitoneum, which results in pexis of the distal

383

384

Section IV  Topics Involving Multiple Organs 1

A

B

In mesenteroaxial volvulus, the stomach folds on its short axis, running across from the lesser curvature to the greater curvature (see Fig. 24-6, 1A and 1B), with the antrum twisting anteriorly and superiorly. In rare cases, the antrum and pylorus rotate posteriorly. Mesenteroaxial volvulus is more likely than organoaxial volvulus to be incomplete and intermittent, and to manifest with chronic symptoms. Mixed mesenteroaxial and organoaxial volvulus has also been reported.42

Incidence and Prevalence 2

A

B

The incidence and prevalence of gastric volvulus are unknown. It is difficult to estimate how many cases are intermittent and undiagnosed. About 15% to 20% of cases occur in children younger than 1 year of age, most often in association with a congenital diaphragmatic defect. The peak incidence in adults is in the fifth decade. Men and women are equally affected.43-45

Clinical Features and Diagnosis 3

A

B

Figure 24-6.  Pathogenesis of gastric volvulus. 1A, Axis for potential mesenteroaxial volvulus bisecting the lesser and greater curvatures. 1B, Mesenteroaxial volvulus resulting from anterior rotation of the antrum along this axis. 2A, Axis for potential organoaxial volvulus passing through the body of the stomach. 2B, Organoaxial volvulus resulting from anteriorsuperior rotation of the antrum along this axis. 3A, Axis for potential organoaxial volvulus passing through the gastroesophageal junction and the pylorus. 3B, Organoaxial volvulus resulting from anterior-superior rotation of the antrum and posterior-inferior rotation of the fundus along this axis. (Adapted from Carter R, Brewer LA 3rd, Hinshaw DB. Acute gastric volvulus. A study of 25 cases. Am J Surg 1980; 140:101-6.)

stomach. Laxity of these ligamentous attachments, elevation of the left hemidiaphragm, or fixation of an otherwise mobile stomach to specific point can result in volvulus. For example, focal adhesions, gastric tumor, or masses in adjacent organs may predispose to gastric volvulus. In two thirds of cases, volvulus occurs above the diaphragm in association with a paraesophageal or mixed diaphragmatic hernia. In the other third of cases, the volvulus occurs below the diaphragm. Gastric volvulus may be mesenteroaxial (40%) or organoaxial (60%; Fig. 24-6). In organoaxial volvulus, the stomach twists along its long axis. This axis usually passes through the gastroesophageal and gastropyloric junctions. The antrum rotates anteriorly and superiorly and the fundus posteriorly and inferiorly, twisting the greater curvature at some point along its length (see Fig. 24-6, 3A and 3B). Less commonly, the long axis passes through the body of the stomach itself, in which case the greater curvature of the antrum and fundus rotate anteriorly and superiorly (Fig. 24-7; see Fig. 24-6, 2A and 2B). This type of volvulus is commonly associated with a diaphragmatic hernia. Organoaxial volvulus is usually an acute event. Vascular compromise and gastric infarction may occur.

Acute gastric volvulus causes sudden severe pain in the upper abdomen or lower chest. Persistent unproductive retching is common. In cases of complete volvulus, it is impossible to pass a nasogastric tube into the stomach. Hematemesis is rare, but may occur because of an esophageal tear or gastric mucosal ischemia.45 The combination of pain, unproductive retching, and inability to pass a nasogastric tube is called Borchardt’s triad. Symptoms of acute gastric volvulus may be mistaken for a myocardial infarction or an abdominal catastrophe such as biliary obstruction or acute pancreatitis.43,44 If the volvulus is associated with a diaphragmatic hernia, physical examination may reveal evidence of the stomach in the left chest. Plain chest or abdominal films will show a large gas-filled viscus in the chest. A barium upper gastrointestinal radiograph will confirm the diagnosis. Upper endoscopy may show twisting of the gastric folds (Fig. 24-8). Endoscopy is not prudent if gastric ischemia is suspected. Chronic gastric volvulus is associated with mild and nonspecific symptoms such as dysphagia, epigastric discomfort or fullness, bloating, and heartburn, particularly after meals. Symptoms may be present for months to years.43,45 It is likely that a substantial number of cases are unrecognized. The diagnosis should be suspected in the proper clinical setting if an upper gastrointestinal radiograph or CT scan shows a large diaphragmatic hernia, even if the stomach is not twisted at the time of the radiograph.46

Treatment and Prognosis

Acute gastric volvulus is an emergency. Nasogastric decompression should be performed if possible. If signs of gastric infarction are not present, acute endoscopic detorsion may be considered. Using fluoroscopy, the endoscope is advanced to form an alpha loop in the proximal stomach. The tip is passed through the area of torsion into the antrum, or duodenum if possible, avoiding excess pressure. Torque may then reduce the gastric volvulus.47 The risk of gastric rupture should be weighed against the possible benefit of temporary detorsion. Surgery for gastric volvulus may be done by open or laparoscopic techniques. In recent years, there has been a trend toward laparoscopic repair.44,48 After the torsion is reduced, the stomach is fixed by gastropexy or tube gastrostomy. Associated diaphragmatic hernia must be repaired.45,49 Combined endoscopic and laparoscopic repair or simple endoscopic gastropexy by placement of a percutaneous gastrostomy tube has been reported.47-52 Chronic gastric volvulus is treated in the same manner as acute volvulus. The surgeon may elect to treat an associated paraesophageal

Chapter 24  Abdominal Hernias and Gastric Volvulus

GE

A

C

B

Figure 24-7.  Gastric volvulus with paraesophageal hernia. A, Chest radiograph showing gas-filled mass in the mediastinum. B, Barium examination showing that the greater curvature and lesser curvature of the stomach are reversed in position (upside-down stomach). C, Plain film, reconstructed from a computed tomographic scan, showing an upside-down stomach in another patient. (A and C courtesy of Dr. Mark Feldman, Dallas, Tex.)

component in the usual manner, with repair of the diaphragm and wrap, if the patient is clinically stable. Acute gastric volvulus has carried a high mortality in the past. However, in one reported series, there were no major complications or deaths in 36 patients with gastric volvulus, including 29 who presented acutely.

INGUINAL AND FEMORAL HERNIAS Cause and Pathogenesis

Figure 24-8.  Gastric volvulus. Twisting of the gastric folds at the point of torsion is noted in this endoscopic view of a gastric volvulus.

The abdominal wall is protected from hernia formation by several mechanisms. In the lateral abdominal wall, there are layers of muscles that together with intervening fascia, provide support. These muscles travel at oblique angles to

385

386

Section IV  Topics Involving Multiple Organs each other and therefore handle forces in various planes, affording greater support than if they were parallel to each other. In the central abdomen, the bulky rectus abdominis muscles provide a barrier to herniation. Abdominal wall hernias occur in areas in which these muscles and fascial layers are attenuated, and they can be congenital or acquired. In the groin, there is an area that is prone to herniation bound by the rectus abdominis muscle medially, the inguinal ligament laterally, and the pubic ramus inferiorly. The aponeurosis of the transverses abdominis muscle provides the deep layer for this area. In this area, the external and internal oblique muscles thin to a fascial aponeurosis only, so that there is no muscular support of the transverse abdominal fascia and the peritoneum. Upright posture causes intra-abdominal pressure to be constantly directed to this area. During transient increases in abdominal pressure, such as occur with coughing, straining, or heavy lifting, reflex abdominal muscle wall contraction narrows the myopectineal orifice and tenses the overlying fascia (shutter mechanism).53 For this reason, hernias are not more common in laborers than in sedentary persons. However, conditions that chronically increase intra-abdominal pressure (e.g., obesity, pregnancy, and ascites) are associated with an increased risk of hernia. Chronic muscle weakness and deterioration of connective tissue (caused by aging, systemic disease, malnutrition, or smoking) promote hernia formation. During embryologic development, the spermatic cord and testis in men (the round ligament in women) migrate from the retroperitoneum through the anterior abdominal wall to the inguinal canal, along with a projection of peritoneum (processus vaginalis). The defect in the abdominal wall (internal inguinal ring) associated with this process represents an area of potential weakness through which an indirect inguinal hernia may form (Fig. 24-9). The processus vaginalis may persist in up to 20% of adults, further predisposing to hernia formation. Direct inguinal hernias do not pass through the internal ring but rather protrude through defects in an area called Hesselbach’s triangle, bounded by the rectus abdominis muscle, the inferior epigastric artery, and the inguinal ligament (see Fig. 24-9). Therefore, indirect inguinal hernias travel with the spermatic cord (or round ligament) and are found lateral to the inferior epigastric vessels; direct hernias are found in the floor of the inguinal canal—an area supported only by the weak transversalis fascia—and are medial to the epigastric vessels. Femoral hernias pass through the opening associated with the femoral artery and vein. They manifest inferior to the inguinal ligament and medial to the femoral artery (see Fig. 24-9).53 Clinical examination cannot easily differentiate indirect from direct inguinal hernias.54 The importance of distinguishing these two entities preoperatively is not critical because the operative approach and repair is identical. However, it is important to diagnose femoral hernias accurately because they can be mistaken for lymph nodes in the groin. Misdiagnosis of an incarcerated loop of bowel in a femoral defect as a lymph node can lead to fine-needle aspiration of the mass and bowel injury. The omentum, colon, small bowel, and bladder are the most common contents of groin hernias, although the appendix, Meckel’s diverticulum, fallopian tube, and ovary have been reported to herniate. In a Richter’s hernia, only the antimesenteric side of the bowel protrudes. In this situation, the patient can have compromise of the bowel without evidence of a bowel obstruction, as is usually present when bowel is incarcerated in a hernia.

Spigelian hernia External oblique

Indirect inguinal hernia

Spermatic cord

Inguinal ligament

Direct inguinal hernia

Internal inguinal ring Femoral artery

Femoral hernia

Femoral vein

Figure 24-9.  Anatomic diagram of inguinal, femoral, and Spigelian hernias. The external oblique muscle has been omitted and the spermatic cord (the round ligament in women) is retracted. Spigelian hernia occurs through defects in the fused aponeurosis of the internal oblique and transverse abdominal muscles. Indirect inguinal hernia occurs through the internal inguinal ring. Direct inguinal hernia occurs through defects in the transversalis fascia in Hesselbach’s triangle. Femoral hernia occurs inferior to the inguinal ligament and medial to the femoral vein and femoral artery.

Incidence and Prevalence

The overall incidence of groin hernias in American men is 3% to 4% if determined through interview, and about 5% if determined by physical examination. The incidence increases with age, from 1% in men younger than 45 years to 3% to 5% in those older than 45 years. About 750,000 groin hernia repairs are done annually in the United States. Of these, 80% to 90% are done in men. Indirect inguinal hernias account for about 65% to 70% of groin hernias in men and women. In men, direct inguinal hernias account for about 30% and femoral hernias for about 1% to 2%, whereas in women the opposite is true. Groin hernias are somewhat more common on the right than on the left side.55 Congenital hernias are more common in males because they represent a patent processus vaginalis. These pediatric hernias are commonly bilateral.

Clinical Features and Diagnosis

Many groin hernias are asymptomatic. The most common symptom is a mass in the inguinal or femoral area that enlarges when the patient stands or strains. An incarcerated hernia may produce constant discomfort. Strangulation causes increasing pain. Symptoms of bowel obstruction or ischemia may occur. In a Richter-type hernia, pain from bowel strangulation may occur without symptoms of obstruction, as only one wall of the intestine is involved in the hernia. The patient should be questioned about risk factors for hernia formation (e.g., chronic cough, constipation, and symptoms of prostate disease). These factors, if not corrected prior to herniorrhaphy, can lead to recurrence.56-58

Chapter 24  Abdominal Hernias and Gastric Volvulus prolonged Valsalva maneuver to demonstrate even small defects.

Treatment and Prognosis

Figure 24-10.  Plain radiograph of a 28-year-old man with a giant incarcerated inguinal hernia. (Courtesy of Dr. Michael J. Smerud, Dallas, Tex.)

On physical examination, inguinal hernias present as a soft mass in the groin. The mass may be larger on standing or straining. It may be slightly tender. It may be possible to palpate the fascial defect associated with the hernia. The patient should be examined upright, the examiner’s finger should be inserted into the femoral canal, and a prolonged Valsalva maneuver should be initiated. It is normal to feel a small impulse against the examining finger with coughing; however, when a hernia is present, a prolonged Valsalva maneuver will result in the protrusion of the sac against the examiner’s finger. Direct and indirect hernias may be difficult to distinguish. Groin hernias may also be noted on a plain abdominal radiograph (Fig. 24-10), barium radiograph, sonogram, or CT scan. Femoral hernias are more difficult to diagnose than other groin hernias. Two thirds of femoral hernias manifest as surgical emergencies. The correct diagnosis is often not made before surgery. The neck of femoral hernias is usually small. Even a small femoral hernia that is difficult to palpate may cause obstruction or strangulation. Richter’s hernias are most common in the femoral area, further complicating the diagnosis. Femoral hernias are most common in women, in whom clinicians may have a lower level of suspicion for hernia than in men. Femoral hernias also occur in children.59 Delay in diagnosis, strangulation, and need for emergency surgery are common.60-62 Any mass below the inguinal ligament and medial to the femoral artery should raise the suspicion of femoral hernia. Femoral hernias are commonly mistaken for femoral adenopathy or groin abscess. Obviously, bedside drainage of an incarcerated femoral hernia must be avoided and therefore liberal use of sonography or CT scanning is useful for distinguishing a hernia from adenopathy, abscess, or other mass.63 The radiologist should perform these examinations with and without a

Many surgeons recommend repair of direct and indirect inguinal hernias, even if asymptomatic, but this is controversial. A study by the American College of Surgeons has shown that males with minimally symptomatic groin hernias can be safely watched.64,65 This study randomized 720 male patients to elective hernia repair or watchful waiting. Only 2 of the 364 patients in the watchful waiting arm study developed complications related to their hernia in 4.5 years. This suggests that minimally symptomatic patients can be watched safely and have their hernia repaired when symptoms increase. Femoral hernias must be repaired promptly because the risk of strangulation is very high.60-62 Groin hernias can be repaired using various techniques. Historically, tissue repairs have been performed. However, several studies have shown a decreased recurrence rate with the use of mesh resulting in tension-free repairs.66-70 These can be performed by open surgery or laparoscopically. The traditional tissue-based repairs were performed exclusively until the 1990s. There are two key components to successful hernia repair: (1) high ligation of the hernia sac, which treats the direct defect; and (2) repair of the floor of the canal, which treats the indirect defect. Even if there is no direct component, a repair of the floor is routinely undertaken. These repairs involve approach to the inguinal canal through a small incision parallel to the inguinal ligament and centered over the internal inguinal ring. Dissection is continued through the external oblique muscle, exposing the internal inguinal ring. The cord structures are then isolated and explored thoroughly to identify an indirect hernia sac. This is ligated and transected. The floor of Hesselbach’s triangle is then reinforced and strengthened by apposing the lateral border of the rectus abdominis aponeurosis to the inguinal ligament (Bassini or Shouldice repair) or to Cooper’s ligament (McVay repair).70-72 Tissue repairs inherently are not tension-free and pose a greater risk of recurrence than tension-free mesh repairs (see later). However, in cases in which there is probable contamination (e.g., in a strangulated hernia), it is important to perform a primary tissue repair and not a mesh repair because there is a high risk of mesh infection. Open mesh repairs are most commonly performed as described by Lichtenstein.66-68 These can be performed under local, regional, or general anesthetic.73,74 The two major components of successful repair remain, with high ligation of the sac; however, the floor is repaired using synthetic mesh to bridge the gap between the conjoint tendon (the edge of the rectus aponeurosis) and inguinal ligament. The mesh can be sutured or stapled in place. Mesh plug repairs have also been developed.75-77 In these cases, minimal dissection is undertaken and the mesh plug, which looks like a badminton shuttlecock, is laid into the defect and tacked in place with a few sutures. The mesh causes fibroblast ingrowth and scarring that leads to strengthening of the floor of the inguinal canal. Mesh repairs have the advantage of being somewhat simpler to perform than tissue repairs and have less tension, less acute pain, and a decreased rate of recurrence.58,66,76,78 Most inguinal hernia repairs in the United States are currently done with mesh.55 Bilateral, very large, or complex abdominal hernias can be repaired with a large mesh that reinforces the entire ventral abdominal wall. This is called giant prosthetic reinforcement of the visceral sac (GPRVS), or the Stoppa procedure.79-81

387

388

Section IV  Topics Involving Multiple Organs Repair of groin hernias may be done with open or laparoscopic techniques.82-86 Several series have compared open hernia repair with laparoscopic repair. The largest and most recent study was performed by the Veterans Cooperative group.86 Almost 1700 patients were followed for two years after being randomized to open versus laparoscopic repair of inguinal hernias. Patients who had their hernias repaired laparoscopically had less pain initially and returned to work one day sooner than those who had open repair. However, the recurrence rate was higher in the laparoscopic group (10% vs. 4% in the open group) and complication rates were higher and more serious in the laparoscopic group compared with the open repair group. In another multicenter, prospective, randomized study performed in the United Kingdom, open repair, primarily using mesh, was compared with laparoscopic repair.82 The recurrence rate after laparoscopic repair was 7% compared with 0% after open repair. As in the U.S. study, patients returned to normal activities more quickly after laparoscopic repair than after open repair. The overall complication rate was lower after laparoscopic repair, but three serious complications occurred after laparoscopy, and none after open repair.84 The results of the Veterans Cooperative group trial has changed the face of hernia repair in the United States. Patients with primary groin hernias are treated with open mesh repair unless they have a strong preference for a laparoscopic approach. Those with recurrent hernias or bilateral hernias can be considered for laparoscopic repair, which can be performed effectively in experienced hands.

Complications and Recurrence

Elective groin hernia repair has a mortality rate of less than 0.001% and serious complications are unusual.86-89 Lacerations of the bowel, bladder, or blood vessels may occur, particularly during laparoscopic repair, and may cause serious consequences if not detected early. Damage to the bowel may also occur during reduction of an incarcerated hernia. Minor acute complications include acute urinary retention, seroma, hematoma, and infection. Serious infection occurs in less than 1% of cases. Damage to the spermatic cord may lead to ischemic orchitis. Tissue dissection predisposes to thrombosis of the venous drainage of the testis. Symptoms are swelling and pain of the cord and testis. The condition persists for 6 to 12 weeks and may result in testicular atrophy. Fortunately, this is a rare complication, occurring after about 0.04% of tissue repairs.90 Hydrocele or vas deferens injury occurs in less than 1% of cases. Damage to sensory nerves is not uncommon during inguinal hernia surgery, and can be related to the division or preservation of the ilioinguinal nerve as it traverses the inguinal canal.87,88,91,92 Chronic paresthesias and pain are reported by about 10% of patients, either caused by deafferentation or to a neuroma. This can often be treated by local nerve block or desensitization therapy.93 Some recurrent hernias are actually indirect hernias missed during the first hernia repair. The risk of recurrence is related to conditions that lead to tissue deterioration, such as malnutrition, liver or renal failure, steroid therapy, and malignancies. Patients with scrotal hernias and recurrent hernias are at higher risk for recurrence or rerecurrence, respectively.57 Recurrent hernias are also more common among smokers than nonsmokers. In patients with cirrhosis and no ascites or moderate ascites, inguinal hernia repair is reported to be safe, although the recurrence rate is increased.94 It is our preference to manage ascites aggressively prior to elective herniorrhaphy. Recurrence rates are higher with laparoscopic hernia repair compared with open

herniorrhaphy.84,86 The routine use of mesh has reduced recurrences, because the learning curve for mesh repair is quicker than for laparoscopic or tissue repair. Recurrence rates are higher after repair of recurrent and femoral hernias than after primary repair of inguinal hernias. Overall, recurrence rates are higher after tissue repairs than after tensionfree mesh repairs.58,69 For inguinal hernias, the most favorable reported recurrence rates for Canadian and Cooper’s ligament repairs have been about 1.5% to 2% for primary repairs and about 3% for repair of recurrent hernia.70,72 Reported recurrence rates for mesh repairs vary from 0% to 4% for primary repairs and can approach 14% for repairs of recurrent hernias.82,86

Inguinal Hernias and Colorectal Cancer Screening

Some practitioners recommend that patients age 50 years or older with inguinal hernias be screened for colorectal neoplasms before hernia repair. Several older prospective studies using sigmoidoscopy or barium enema to screen middle-aged or older men with inguinal hernias have reported the prevalence of polyps to be from 4% to 26% and of colorectal cancer to be from 2.5% to 5%.95,96 However, more recent data have clearly shown that there is no increased risk of colorectal cancer in patients who have groin hernias. In a prospective study of colonoscopy for screening of asymptomatic U.S. veterans, the prevalence of polyps was 37.5% and of colorectal cancer 1%.97 Thus, the prevalence of colorectal neoplasms is substantial in middleaged or older men with or without inguinal hernias. In several more recent studies, the risk of colorectal cancer was found to be similar in patients with hernias (4% to 5%) compared with a control group that did not have hernias (3% to 4%).98-100 Large inguinal hernias, particularly incarcerated hernias, may cause difficulty during sigmoidoscopy or colonoscopy. In such patients, it may be advisable to defer the examination until after the hernia repair. Incarceration of fiberoptic endoscopes within hernias has been reported.101,102

Inguinal Hernias and Benign Prostatic Hyperplasia

Inguinal hernia and symptomatic benign prostatic hyperplasia coincide in 9% to 25% of men. Straining to void may cause worsening of inguinal hernia. Conversely, the risk of postoperative urinary retention after hernia repair is increased by prostatic hypertrophy, and older male patients with any symptoms of prostate disease should be counseled on the risk of urinary retention after hernia repair.103 With the advent of improved medical therapy for benign prostatic hyperplasia, most patients can be managed with medical therapy prior to herniorrhaphy. If elective inguinal hernia repair and transurethral prostatic resection are required, there are some surgeons that would consider performing these procedures concurrently.104,105 However, infection of mesh can be a significant problem, and therefore we recommend sequential surgery.

OTHER VENTRAL HERNIAS Patients often mistake diastasis recti for abdominal hernia. Diastasis recti is a separation of the rectus abdominis muscles without a defect in the abdominal fascia. This condition does not require repair. Diastasis recti can be demonstrated as a midline defect exaggerated by a Valsalva maneuver. No fascial ring can be palpated and the defect is often very wide and long. True ventral hernias include incisional, epigastric, umbilical, and Spigelian hernias.

Chapter 24  Abdominal Hernias and Gastric Volvulus INCISIONAL HERNIAS

Incisional hernias, as the name implies, are hernias that occur after a prior operation. Incisional hernias include postlaparotomy hernias, parastomal hernias, and trocar site hernias.

Cause and Pathogenesis

Incisional hernias are caused by two main factors, patientrelated and surgery-related. Patient-related factors include conditions that may increase intra-abdominal pressure, such as obesity, collagen vascular diseases, history of aneurysms, nutritional factors, and ascites.94,106,107 Conditions that impair healing, such collagen vascular disease in patients receiving glucocorticoid therapy, can also increase postoperative hernia formation. Surgery-related factors include the type and location of the incision. For example, it is more common for hernias to develop after a vertical midline incision than after a transverse incision.106 This has led some surgeons to use transverse incisions in patients who are predisposed to hernias (e.g., patients with Crohn’s disease who are on glucocorticoids or other immunosuppressants). Development of a wound infection postoperatively can lead to a higher incidence of hernia formation. Placement of a stoma results in an intentional creation of a hernia through which the intestine runs. By placing these intentional hernias within the rectus muscle, rather than lateral to the rectus, the defect can be somewhat controlled. This results in a lower rate of parastomal hernias. Trocar hernias have become a more common occurrence with the increased use of laparoscopic surgery. The rate of hernia formation is related to the size of the trocar used, with trocars larger than 10 mm in diameter more commonly associated with hernia formation, and to the location of the trocar placement on the abdominal wall. Lateral trocar placement has a lower chance of hernia formation than midline placement. The fascial defect should be closed carefully if large trocars are used.108

Incidence and Prevalence

Incisional hernias are common after laparotomy. When followed carefully over a long period, as many as 20% of patients can be found to develop a hernia. This incidence increases to 35% to 50% of cases when there is wound infection or dehiscence.106,107 Up to 50% of such hernias manifest more than one year after surgery. Vertical incisions are more likely to be complicated by hernias than transverse incisions. Obesity, advanced age, debility, sepsis, postopera­ tive pulmonary complications, and glucocorticoid use also increase the risk.109 Trocar site hernias are estimated to occur after 0.5% of laparoscopic cholecystectomies.110 They usually occur at the site of the largest trocar, which is typically larger than 10 mm in diameter. Parastomal hernias are reported to occur in as many as 50% of cases after stoma placement.111,112 Specific measures are taken at the time of surgery to decrease the incidence of hernia formation. For example, the smallest fascial defect is created within the rectus sheath, rather than lateral to it. The use of mesh in primary stoma placement may reduce the incidence of subsequent hernia formation; however, this routine use of mesh is controversial. Conditions that led to bowel dilation prior to stoma placement (e.g., obstruction) can result in subsequent bowel shrinkage after stoma placement. This can increase the space between the bowel wall and the fascia, facilitating hernia formation.

Clinical Features and Diagnosis

Incisional hernias can cause chronic abdominal discomfort. Because the fascial defect of incisional hernias is usually

large, strangulation is unusual even with incarceration. Reduced ability to voluntarily increase intra-abdominal pressure interferes with defecation and urination. Lordosis and back pain may occur. Large incisional hernias may lead to eventration disease. With the loss of integrity of the abdominal wall, the diaphragm cannot contract against the abdominal viscera, but rather forces the viscera into the hernia. The diaphragm thus becomes inefficient. The hernia tends to enlarge. The viscera may lose the so-called right of domain in the abdominal cavity. Surgeons need to be careful about reducing and repairing these large hernias because the acute increase in abdominal pressure can lead to pulmonary failure and reduced venous return, resulting in an effective abdominal compartment syndrome.113,114 Techniques have been developed whereby the intra-abdominal cavity is insufflated gradually with air through a surgically placed indwelling catheter. This technique gradually stretches and expands the abdominal wall, preparing for successful reduction of the hernia contents. Trocar site hernias usually cause pain and a bulge at the trocar site. Because of the small opening, it is more likely that intra-abdominal contents could become strangulated in the defect. Richter’s hernia and small intestinal volvulus have been reported.108,110 Parastomal hernias often interfere with ostomy function and the fit of appliances. Incarceration and strangulation of bowel may occur.111 Diagnosis of a hernia can be difficult if the defect is small or, tender, or in an obese patient. A useful adjunct to diagnosis can be ultrasound or CT scanning in these situations. The physician requesting the ultrasound or CT should communicate his or her suspicions carefully to the radiologist because specific maneuvers can be performed by the radiologist to demonstrate the defect. For example, ultrasound can be performed with the patient in an upright position.

Treatment and Prognosis

Incisional hernias are best repaired with prosthetic mesh because the recurrence rate is substantially lower than after traditional tissue repair.106,107 The key element in hernia repair is to achieve a tension-free repair. In general, a nonabsorbable mesh is used to bridge the gap between the fascial edges. Every attempt is made to place a layer of peritoneum or hernia sac between the abdominal contents and the mesh. However, if this cannot be done, special double-sided mesh is available with expanded polytetrafluoroethylene on one side. This material does not stick to bowel and therefore is unlikely to erode into the intestine.113,114 If eventration disease is suspected, the abdominal wall may need to be stretched by repeated progressive pneumoperitoneum before repair. Recurrences of incisional hernia are reported in 4% to 29% of cases.106,115,116 Laparoscopic repair of ventral defects can be performed. There is some suggestion that laparoscopic repair results in fewer recurrences and lower morbidity.117 Laparoscopic repair is performed by insufflating the abdomen and gradually creating a working space by lysing adhesions carefully. Doublesided mesh is then placed in the retroperitoneal position and fixed by tacks and sutures.118 This can result in the sensation of a residual hernia, caused by retention of fluid in the hernia sac between the mesh and the skin, which can be frustrating for the clinician and patient. Chronic pain at suture or tack sites appears to be a greater issue with laparoscopic hernia repair when compared with open repair.119 Small and minimally symptomatic parastomal hernias may be treated with a modified ostomy belt. If surgery is necessary, there are several modes of treatment. The stoma can be relocated to the other side of the abdomen or to another

389

390

Section IV  Topics Involving Multiple Organs quadrant of the abdomen. Primary repair of the parastomal defect is no longer considered adequate treatment and therefore mesh placement is advocated. A piece of mesh shaped with a keyhole defect, through which the stoma can be exteriorized, can be used.111,112 There are reports of this being done laparoscopically.120,121 To decrease the incidence of trocar site hernias, it is recommended that trocar ports be removed under direct vision and the defects sutured closed, particularly those defects related to trocars that are larger than 10 mm in diameter. Other risk factors for trocar site hernias include age older than 60 years, obesity, and increased operative time.110 Newer prosthetic materials that are biodegradable have become available. Pig mucosa infused with a collagen matrix can be used in the place of mesh in patients in whom there has been contamination (e.g., when bowel resection is necessary). These substrates are thought to be degradable and cause an influx of fibroblasts, resulting in a vigorous scar that can provide strength similar to mesh. With time, these substrates are degraded, leaving only autologous tissue. However, recurrence is still a significant issue and can occur in more than 30% of patients.122

EPIGASTRIC AND UMBILICAL HERNIAS Cause and Pathogenesis

Epigastric hernias occur through midline defects in the aponeurosis of the rectus sheath (linea alba) between the xiphoid and the umbilicus. The defects are usually small and frequently multiple. Because of the location in the upper part of the abdominal wall, it is unusual for bowel to become incarcerated in epigastric hernias. More commonly, preperitoneal fat or omentum protrude through these hernias.123 Umbilical hernias in infants are congenital (see Chapter 96). They often close spontaneously. There is an increased incidence of congenital umbilical hernias in children of African descent.123 In general, these defects will close spontaneously by 4 years of age. If they are still evident after this age, surgical repair is indicated. In adults, umbilical hernias may develop consequent to increased intra-abdominal pressure because of ascites, pregnancy, or obesity.

Incidence and Prevalence

Epigastric hernias are found in 0.5% to 10% of autopsies. Many are asymptomatic or undiagnosed during life. They generally occur in the third through fifth decades. They are more common in men than in women.123 Umbilical hernias occur in about 30% of African American infants and 4% of white infants at birth, and are present in 13% and 2%, respectively, by 1 year of age.124 Umbilical hernias are more common in low birth weight infants than in those of normal weight. Umbilical hernias occur in 20% of patients with cirrhosis and ascites.125

Clinical Features and Diagnosis

The main symptom of epigastric hernia is upper abdominal pain. The pain is usually localized to the abdominal wall, rather than the deep visceral pain that accompanies intestinal pathology. A specific tender nodule or point of tenderness can be palpated in the nonobese patient. Diagnosis may be difficult, particularly in obese patients. However, symptoms are sometimes mistaken for those of a peptic ulcer or biliary disease. Sonography and CT may be helpful in the diagnosis.123,126 Umbilical hernias in children are usually asymptomatic. However, incarceration and strangulation may occur in chil-

dren and adults. Spontaneous rupture of umbilical hernias may occur in patients with ascites and, rarely, in pregnant women.125,127 Skin changes with maceration and ulceration generally occur prior to frank rupture. Therefore, the findings of skin changes in a patient with an umbilical hernia should warrant urgent repair. Care must be taken when performing a therapeutic paracentesis in patients with umbilical hernias; the hernia must be reduced and kept reduced during the paracentesis because strangulation of umbilical hernias may occasionally be precipitated by rapid removal of ascites.128

Treatment and Prognosis

If surgery is performed for epigastric hernia, the linea alba should be widely exposed because multiple defects, called Swiss cheese defects, may be found. Umbilical hernias are most often left untreated in children because complications are unusual, and they usually close spontaneously if smaller than 1.5 cm in diameter. Repair should be considered if they are larger than 2 cm or if they are still present after 4 years of age.124 Repair of umbilical hernias should be recommended for adults if they are even minimally symptomatic or difficult to reduce. Techniques for repair of all abdominal wall defects rely on a tension-free repair to decrease the risk of recurrence. Open or laparoscopic techniques can be used to achieve this end. Data support the routine use of mesh in repair of these defects, because this results in a decrease in recurrences.129 Mesh is always used in laparoscopic repair. Once complications develop in patients with umbilical hernias, the prognosis worsens significantly. Those patients requiring bowel resection at the time of umbilical herniorrhaphy had a 29% mortality compared with no mortality in those that did not require bowel resection.130 Repair of umbilical hernias in patients with cirrhosis and ascites is a difficult clinical problem. In general, ascites should be aggressively controlled. If this is not possible, consideration should be given to transjugular intrahepatic portosystemic shunt (TIPS) or liver transplantation (see Chapters 91 and 95). Spontaneous rupture of umbilical defects in patients with ascites portends a poor prognosis, with a mortality of 60%.125,128,131,132 Laparoscopic techniques and earlier repair of hernias in patients with cirrhosis should be considered, because the morbidity of elective repair appears not to be as high as once thought. Outcome after surgical repair is directly dependent on nutritional status and control of ascites. Control of ascites may require frequent paracentesis to keep the abdomen flat to allow healing. Topical sealants can be used to decrease the risk of leakage.127

SPIGELIAN HERNIAS Cause and Pathogenesis

Spigelian hernias occur through defects in the fused aponeurosis of the transverse abdominal muscle and internal oblique muscle, lateral to the rectus sheath, most commonly just below the level of the umbilicus (see Fig. 24-9). This area is called the Spigelian fascia, named after the Belgian anatomist Adrian van den Spieghel. This fascia is where the linea semilunaris, the level at which the transversus abdominis muscle becomes aponeurosis rather than muscle, meets the semicircular line (of Douglas). The epigastric vessels penetrate the rectus sheath in this area. The com­bination of all these anatomic features can lead to a potential defect and a Spigelian hernia. The Spigelian fascia is covered by the external oblique muscle; therefore, Spigelian hernias do not penetrate through all layers of the abdominal wall.123

Chapter 24  Abdominal Hernias and Gastric Volvulus Incidence and Prevalence

Spigelian hernias (SHs) are rare. Approximately 1000 cases have been reported.133-137 The largest series of patients included 81 patients.138 SHs are twice as common in females as in males and are more common on the left side of the abdomen (60% left and 40% right).139 SHs generally occur in patients older than 40 years.123,140

Clinical Features and Diagnosis

Spigelian hernias can be difficult to diagnose because the external oblique muscle overlies the defect in the deeper fascia. Only 75% of patients with SH are correctly diagnosed before surgery.139 Therefore, the examiner must have a high degree of suspicion when a patient complains of pain at the lateral edge of the rectus, inferior to the umbilicus. Careful examination will suggest that the pain originates in the abdominal wall and not in the peritoneal cavity. This determination is critical because SH can be mistaken for conditions such as acute appendicitis and diverticulitis.141,142 Frequently, only omentum is present in the hernia, but large or small bowel, ovary, appendix, or fallopian tube may herniate. A Richter’s hernia and bowel obstruction caused by incarcerated small intestine may occur.140,143 The differential diagnosis includes rectus sheath hematoma, lipoma, or sarcoma. Sonography and CT are the most useful adjuncts for diagnosing SH.123,144,145 An astute radiologist will perform these studies using various techniques, such as the Valsalva maneuver, to increase detection of even a small SH.

Treatment and Prognosis

Spigelian hernias may be approached by open or laparoscopic techniques. Laparoscopy can be helpful as a diagnostic tool in patients who are suspected of having SH, even if open repair is anticipated.146 The hernia can be best identified from within the peritoneal cavity. Preperitoneal laparoscopic techniques can be used, with the advantage of staying outside the peritoneal cavity.135 Intraperitoneal laparoscopic repair can be performed using mesh that is coated on one side so as not to stick to the underlying bowel.147 Laparoscopy results in decreased pain and decreased length of hospital stay compared with open techniques.148 However, these hernias are so rare that the surgeon should elect to repair them well in the manner with which they are the most experienced.137,149 As with other hernias, most SHs are closed using mesh repairs, a technique that appears to have a lower recurrence rate than primary repair.138,149

PELVIC AND PERINEAL HERNIAS Cause and Pathogenesis

Most pelvic and perineal hernias occur in older female patients. Obturator hernias occur through the greater and lesser obturator foramina. The obturator foramen is larger in women than in men, and is ordinarily filled with fat. Marked weight loss thus predisposes to herniation. Sciatic hernias occur through the foramina formed by the sciatic notch and the sacrospinous or sacrotuberous ligaments. Abnormal development or atrophy of the piriform muscle may predispose to sciatic hernia. Sciatic hernias may contain ovary, ureter, bladder, or large or small bowel.150-152 Perineal hernias occur in the soft tissues of the perineum and are very rare. They may be primary or postoperative. Primary perineal hernias occur anteriorly through the uro-

genital diaphragm or posteriorly through the levator ani muscle or between the levator ani and coccygeus muscles. Secondary perineal hernias occur most often after surgery, such as abdominal-perineal resection, pelvic exenteration, perineal prostatectomy, resection of the coccyx, or hysterectomy. Radiation therapy, wound infection, and obesity predispose to the development of secondary perineal hernias.153-155

Incidence and Prevalence

Pelvic hernias are rare. Obturator hernias typically occur in older, cachectic, multiparous women. About 600 cases have been reported.150 In Japan, obturator hernias account for about 1% of all hernia repairs, but in the West, they account for 0.07% of all hernias.156,157 Sciatic hernias are even less common than obturator hernias, with fewer than 100 cases reported.123 They are most common in women. Perineal hernias are also rare. Primary perineal hernias are most common in middle-aged women. Anterior perineal hernias do not occur in men.155 Secondary perineal hernias occur after less than 3% of pelvic exenterations and less than 1% of abdominal-perineal resections for rectosigmoid cancer.153,154

Clinical Features and Diagnosis

Obturator hernias (OHs) occur almost exclusively in older women and are more common on the right side.156,158,159 OHs commonly cause lower abdominal pain. Up to 88% of patients present with symptoms of small bowel obstruction.123,156 Because the hernia orifice is small, Richter’s hernia and strangulation are common, and 50% of patients present with bowel necrosis.160 There are three signs specific for incarcerated OH: 1. Obturator neuralgia. This manifests as paresthesia that extends along the medial aspect of the thigh. 2. Howship-Romberg sign, caused by pressure on the obturator nerve, resulting in paresthesias and pain in the hip and inner thigh. The pain is diminished by hip flexion and increased by hip extension, adduction, or medial rotation. This sign is seen in 25% to 50% of patients with OH and is considered pathognomonic. 3. Hannington-Kiff sign, elicited by percussing the adductor muscle above the knee. Absence of the normal adductor reflex contraction is a strong indicator of obturator nerve impingement caused by an OH.161 Occasionally, a mass may be palpable in the upper medial thigh or in the pelvis on pelvic or rectal examination. The diagnosis is difficult, often delayed, and usually not made preoperatively. Preoperative diagnosis is sometimes evident on a ultrasonogram or CT scan.150,157,161-164 Sciatic foramen hernias may manifest as a mass or swelling in the gluteal or infragluteal area, but are generally difficult to palpate because they occur deep to the gluteal muscles. Chronic pelvic pain may occur caused by incarceration of a fallopian tube and/or ovary. Impingement on the sciatic nerve may also produce pain radiating to the thigh. Intestinal or ureteral obstruction may occur. The differential diagnosis includes lipoma or other soft tissue tumor, cyst, abscess, and aneurysm. The diagnosis is often difficult and is made only at laparotomy or laparoscopy.123 In women, primary perineal hernias manifest anteriorly in the labia majora (pudendal hernia) or posteriorly in the vagina.155 In men, they manifest in the ischiorectal fossa. Primary and postoperative perineal hernias are usually soft and reducible. Most patients complain of a mass that produces discomfort on sitting. Because the orifice of the hernia is usually wide, incarceration is rare. If the bladder is

391

392

Section IV  Topics Involving Multiple Organs involved, urinary symptoms may occur. Postoperative perineal hernias may be complicated by cutaneous ulceration. The differential diagnosis includes sciatic hernia, tumor, hematoma, cyst, abscess, and rectal or bladder prolapse.153,155

Latissimus dorsi 12th rib

Treatment and Prognosis

The treatment of pelvic hernias is surgical. Laparoscopic repair of obturator and sciatic hernias has been reported.151,165 However, most patients with pelvic hernias present with an acute surgical condition, often bowel obstruction. Therefore, it is often necessary to perform an open procedure to manage the problem.166 The prognosis is poor when patients present with an acute illness. Nutritional depletion, advanced age, and poor medical health are all confounding variables. Repair of perineal hernias can be complex. When bowel resection is required, mesh placement is usually not used because of the high risk of infection. The advent of newer bioabsorbable products has allowed these materials to be used in these contaminated fields. Peritoneal flaps or muscle advancement flaps can be used to perform tissue repairs of these defects.156,160

LUMBAR HERNIAS Cause and Pathogenesis

Lumbar hernias can occur in two separate triangular areas of the flank. The superior triangle (Grynfeltt’s lumbar triangle) is bound by the 12th rib superiorly, the internal oblique muscle inferiorly, and the sacrospinous muscles medially.167 The inferior triangle (Petit’s lumbar triangle) is bound by the latissimus dorsi muscle posteriorly, the external oblique muscle anteriorly, and the iliac crest inferiorly. Of all lumbar hernias, 20% are congenital; congenital lumbar hernias are commonly bilateral.168 Lumbar hernias are more common on the left than on the right side. This may be because the liver pushes the right kidney inferiorly in development, leading to protection of the lumbar triangles. Grynfeltt’s hernias are more common than Petit’s hernias. There is a 2 : 1 male predominance.169 Pseudohernia may occur in the lumbar area as the result of paresis of the thoracodorsal nerves. This is caused by loss of muscle control and tone, but there is no associated fascial defect. Causes include diabetic neuropathy, herpes zoster infection, and syringomyelia.170 Incisional and post-traumatic hernias also occur in the lumbar area. Flank incisions are used to access the retroperitoneum for procedures such as nephrectomies. These may be true hernias or pseudohernias caused by postoperative muscle paralysis. Motor vehicle accidents are the most common cause of post-traumatic lumbar hernias. If a lumbar hernia is found after a motor vehicle accident, it is critical to assume that the patient has other intra-abdominal injuries. These patients should undergo urgent laparotomy because more than 60% of them will have major intra-abdominal injuries (Fig. 24-11).171-173

Incidence and Prevalence

Lumbar hernias are rare. About 300 cases have been reported.123

Clinical Features and Diagnosis

Lumbar incisional hernias generally present as a large bulge that may produce discomfort. These are especially evident when the patient strains or is in the upright position. Because of the large size of the defect, incarceration is not common. Inferior and superior lumbar triangle hernias may

External oblique Internal oblique Iliac crest Sacrospinalis

Figure 24-11.  Anatomic diagram of Petit’s and Grynfeltt’s hernias, which are lumbar hernias. The inferior triangle hernia, Petit’s hernia (thick arrow), is bounded by the latissimus dorsi muscle, the external oblique muscle, and the iliac crest. The superior triangle hernia, Grynfeltt’s hernia (thin arrow), is bounded by the 12th rib, the internal oblique muscle, and the sacrospinalis muscle.

occur through small defects and can manifest with incarceration (24%) and strangulation (18%).169 The differential diagnosis includes lipoma, renal tumor, abscess, and hematoma. Bowel, mesentery, spleen, ovary, and kidney have been reported to herniate. Occasionally a small lumbar hernia may impinge on a cutaneous branch of a lumbosacral nerve, causing pain referred to the groin or thigh. CT scanning may aid in the diagnosis of lumbar hernia.174

Treatment and Prognosis

Closure of large lumbar hernias, as well as superior and inferior lumbar triangle hernias, often requires the use of prosthetic mesh or an aponeurotic flap. Identifying fascia with good tensile strength and repairing the defect with mesh in a tension-free manner is critical to preventing recurrence.175 Fixation of mesh to bony structures, such as the rib or the iliac crest, may be required. Preperitoneal as well as transperitoneal laparoscopic repair has been reported, and can result in less pain and quicker return to activity.176-179 Large and symptomatic lumbar pseudohernias should be treated by managing the underlying condition. Resolution has been reported following treatment of herpes zoster.170

INTERNAL HERNIAS Internal hernias are protrusions into pouches or openings within the abdominal cavity, rather than through the abdominal wall. Internal hernias may be the result of developmental anomalies or may be acquired.180 Commonly, internal hernias develop after earlier abdominal surgery (e.g., after a Roux-en-Y gastric bypass procedure).

Cause and Pathogenesis

Internal hernias caused by developmental anomalies include paraduodenal, foramen of Winslow, mesenteric, and supra-

Chapter 24  Abdominal Hernias and Gastric Volvulus vesical hernias. During gestation, the intestines are extraabdominal. During fetal development, the mesentery of the duodenum, ascending colon, and descending colon becomes fixed to the posterior peritoneum. These segments of the bowel become reperitonealized and they attach to the retroperitoneum. Anomalies of mesenteric fixation may lead to abnormal openings through which internal hernias may occur. The extreme example of this is a complete intestinal malrotation, in which the ligament of Treitz does not assume its appropriate location to the left of the spine. This condition predisposes to midgut volvulus and can lead to extensive mesenteric ischemia (see Chapter 96).181,182 Lesser anomalies of fixation lead to defects such as paraduodenal and supravesical hernias. Abnormal mesenteric fixation may lead to abnormal mobility of the small bowel and right colon, which facilitates herniation. During fetal development, abnormal openings may occur in the pericecal, small bowel, transverse colon, or sigmoid mesentery, as well as the omentum, leading to mesenteric hernias. Unusual hernias can occur on structures such as the broad ligament.183 Paraduodenal hernias (PDHs) are thought to occur because of anomalies in fixation of the mesentery of the ascending or descending colon. PDHs occur on the left side in 75% of cases and have a 3 : 1 male predominance.184,185 Patients most commonly present in the fourth decade. In cases of left PDH, an abnormal foramen (fossa of Landzert) occurs through the mesentery close to the ligament of Treitz, leading under the distal transverse and descending colon, posterior to the superior mesenteric artery. Small bowel may protrude through this fossa and become fixed in the left upper quadrant of the abdomen. The mesentery of the colon thus forms the anterior wall of a sac that encloses a portion of the small intestine. Right PDH occurs in the same fashion through another abnormal foramen (fossa of Waldeyer), leading under the ascending colon.186 Foramen of Winslow hernias may occur when this foramen is abnormally large, particularly if there is abnormal mesenteric fixation of the small bowel and right colon. Most commonly, the right colon is abnormally fixed to the retroperitoneum, resulting in a patulous foramen of Winslow. Abnormally mobile small bowel and colon may herniate through the foramen of Winslow into the lesser sac. Symptoms of small bowel or colonic obstruction may occur; these may be intermittent as the hernia reduces spontaneously. Impingement on the portal structures can occur, but rarely results in obstruction of the bile duct or the portal vein.187 Gastric symptoms may also occur if the herniated bowel becomes distended because the herniated bowel loops are located in the lesser sac, posterior to the stomach. Mesenteric hernias occur when a loop of intestine protrudes through an abnormal opening in the mesentery of the small bowel or colon. These mesenteric defects are thought to be developmental in origin, although they may also be acquired as a result of surgery, trauma, or infection. The most common area for such an opening is in the mesentery of the small intestine, most often near the ileocolic junction. Defects have been reported in the mesentery of the appendix, sigmoid colon, and a Meckel’s diverticulum.188-190 The intestine finds its way through the defects through normal peristaltic activity. Various lengths of intestine may herniate posteriorly to the right colon, into the right paracolic gutter (Fig. 24-12). Compression of the loops may lead to obstruction of the herniated intestine. Strangulation may occur by compression or by torsion of the herniated segment. Obstruction may be acute, chronic, or intermittent. The herniated bowel may also compress arteries in the margins of the mesenteric defect, causing ischemia of nonherniated

Figure 24-12.  Internal (pericecal) hernia with strangulation. This computed tomographic scan shows a mass of infarcted small intestine in the right side of the abdomen (white arrow). The area of herniation (open arrow to the right of the spine) shows twisting of the small bowel as it passes through the mesentery. (Courtesy of Dr. Michael J. Smerud, Dallas, Tex.)

intestine. Similar defects may occur in the mesentery of the small bowel, transverse mesocolon, omentum, and sigmoid mesocolon. There are three types of mesenteric hernias involving the sigmoid colon. Transmesosigmoid hernias have no true sac. They occur through both layers of the mesocolon. Generally, the bowel becomes trapped in the left gutter, lateral to the sigmoid colon. Intermesosigmoid hernias are hernias that occur within the leaves of the sigmoid colon. This results in the hernia contents being contained within the mesentery of the sigmoid colon, generally posterior to the sigmoid colon. Intersigmoid hernias occur between the retroperitoneal fusion plane, between the sigmoid colon mesentery and the retroperitoneum. These hernias are contained in the retroperitoneum and generally lift and dissect the sigmoid colon on its mesentery out of the left gutter.180,191 Supravesical hernias protrude into abnormal fossae around the bladder. They are classified as internal or external supravesical hernias. Internal supravesical hernias occur within the abdomen and thus are internal hernias. They may extend anterior, lateral, or posterior to the bladder. External supravesical hernias occur outside the abdominal wall and appear much like indirect inguinal hernias. They usually contain small bowel but may contain omentum, colon, ovary, or fallopian tube.192-194 Acquired internal hernias may occur as a complication of surgery or trauma if abnormal spaces or mesenteric defects are created. Adhesions can create spaces into which bowel may herniate. Division of mesentery to create conduits, such as Roux-en-Y limbs, can lead to defects within the mesentery or around the reconstruction, which can result in herniation. With the increased popularity of the Roux-en-Y gastric bypass (RYGB) procedure for morbid obesity, there has been an increased incidence of unusual hernias related specifically to this surgery (see later).108 Retroanastomotic hernias may occur after gastrojejunostomy, colostomy or ileostomy, ileal bypass, or vascular bypass when an abnormal space may be created into which small bowel, colon, or omentum may herniate. The most common retroanastomotic hernia occurs after gastrojejunostomy, usually after gastric resection with Billroth II reconstruction. The afferent loop, efferent loop, or both, protrude into the space posterior to the anastomosis. Efferent loop

393

394

Section IV  Topics Involving Multiple Organs hernias are about three times as common as afferent loop hernias, likely caused by the limited length of the afferent loop and the tethering effect of fixed structures involved in the afferent loop. For example, after a Bilroth II anastomosis, the afferent loop is connected to the duodenum, which is fixed, and the efferent loop is connected to the remainder of the small intestine. The efferent loop is therefore more mobile and can herniate into potential spaces. Colostomy, ileostomy, ileal bypass, and vascular bypass procedures (e.g., aortofemoral bypass) may also lead to the creation of a space into which organs can protrude. Renal transplant procedures are extraperitoneal, but an unrecognized inadvertent rent in the peritoneum can lead to pararenal intestinal herniation.180,191,195 Hernias after RYGB procedures have become more common with the increasing demand for this operation. These can be internal or external hernias through the incision or port sites (see earlier). Small bowel obstruction related to internal hernias after RYGB occurs in approximately 1.8% to 4% of patients.108,195 The incidence of small bowel obstruction is slightly higher after laparoscopic RYGB (3%) than after open RYGB (2%).108 There are three potential spaces created during the RYGB that can result in internal herniation. The Peterson defect occurs to the right of the jejunum as it traverses the mesentery of the transverse colon to reach the pouch of the stapled stomach. The endoscopist encounters this as a narrowing that occurs in the Roux limb at around 40 to 60 cm distal to the pouch-jejunum anastomosis. The jejunojejunostomy mesenteric defect occurs between the divided leaves of the small intestinal mesentery. The mesentery is divided to create the Roux limb, which is brought up to the gastric pouch. The two edges of the transected mesentery are then sewn together to prevent this defect. However, despite these measures, a defect can develop, resulting in herniation of intra-abdominal contents. The transverse mesocolic defect occurs through the defect in the transverse mesocolon through which the jejunal limb is brought to reach the stomach pouch. The Peterson and transverse mesocolic defects can be avoided by placing the jejunal limb in an antecolic position. In this case, the jejunum is not placed through a rent in the transverse mesocolon, but rather is brought anterior to the transverse colon. Although this makes intuitive sense, it is not always possible to achieve enough length of small intestinal mesentery to ensure an antecolic anastomosis without tension.196 Hernias can occur in the mesentery of the colon after colonoscopy.197 This likely occurs as a rent develops in the sigmoid mesocolon with insufflation of the colon. Hernias may occur through the broad ligament of the uterus, most commonly through tears occurring during pregnancy because 85% of these hernias occur in parous women. Other cases may be developmental or caused by surgery (e.g., uteropexy or salpingo-oophorectomy).198,199

Incidence and Prevalence

Internal hernias are rare. They are found in 0.2% to 0.9% of autopsies, but a substantial proportion of these remain asymptomatic.180 About 4% of bowel obstructions are caused by internal hernias. Internal hernias occur most often in adults. Although half of developmental internal hernias are paraduodenal hernias, 1% or fewer of all cases of intestinal obstruction are caused by paraduodenal hernias. About 500 cases have been reported. They are more common in males than in females. They may occur in children or adults, but typically manifest between the third and sixth decades of life; most (75%) paraduodenal hernias occur on the left

side.184,185,200-202 Fewer than 200 cases of foramen of Winslow hernia have been reported.203,204 Mesenteric hernias are rare and can occur at any age.180,191 Fewer than 100 cases of internal supravesical hernia have been reported. They are more common in men than in women. Almost all reported cases have occurred in adults, most commonly in the sixth or seventh decade.193,194 Similarly, fewer than 100 cases of broad ligament hernias have been reported.198,199 Postgastroenterostomy internal hernias have become less common because the frequency of surgery for peptic ulcer disease has declined. Other postanastomotic internal hernias are also rare.180 Internal hernias related to RYGB procedures have become more common because surgeries for morbid obesity have become more widely performed. Small bowel obstruction–related to internal hernias in most patients occurs with an incidence of 1.5% to 4% after RYGB.108,195

Clinical Features and Diagnosis

Any of the various forms of internal hernias may manifest with symptoms of acute or chronic intermittent intestinal obstruction. The diagnosis is difficult in patients with chronic symptoms and is rarely made preoperatively in patients who present with acute obstruction and strangulation.180,191 Intestinal obstruction, which may be low grade, chronic, and recurrent, or may be high grade and acute, develops in approximately half of patients with paraduodenal hernias.184,201 Upper gastrointestinal tract contrast radiography has been shown to have excellent accuracy.200 Barium radiographs may show the small bowel to be bunched up or agglomerated, as if it were contained in a bag, and displaced to the left or right side of the colon. Small bowel is often absent from the pelvis. The colon may be deviated by the internal hernia sac. Bowel proximal to the hernia may be dilated.180,205 However, barium radiographs may be normal if the hernia has reduced at the time of the study. Endoscopy is not reliable for the diagnosis of paraduodenal hernias. Displacement of the mesenteric vessels can be noted if CT scanning with intravenous contrast or arteriography is performed.184,200,202 However, CT scanning may miss a paraduodenal hernia unless specific attention is paid to the relationship of the small intestine to the colon and mesenteric vessels. In hernias of the foramen of Winslow, small bowel herniates behind the portal structures in about two thirds of cases; in the remaining cases, the right colon herniates into the lesser sac. Herniation of the gallbladder has been reported. Patients may have symptoms of gastric or proximal intestinal obstruction, even in the case of colon herniation, because of pressure of the herniated bowel on the stomach. Occasionally, an epigastric mass is palpable. Plain abdominal radiographs may show the stomach displaced anteriorly and to the left. Contrast enema may show displacement of the cecum into the epigastrium. Sonography may show a mass in the lesser sac.180,203,204 CT scanning is accurate for the diagnosis of foramen of Winslow hernias. The herniated bowel is posterior to the stomach within the lesser sac. There may be associated dilation of the biliary tree or portal vein narrowing caused by compression of the portal structures. Rarely is there any physiologic consequence to this finding.187 Mesenteric hernias are difficult to diagnosis preoperatively. Symptoms and signs are those of acute or chronic intermittent bowel obstruction or of acute strangulation. Plain abdominal radiographs may show evidence of bowel obstruction or displacement of the normal gas pattern. For example, with hernias through the sigmoid mesentery, the small intestine gas pattern lies laterally to the sigmoid gas

Chapter 24  Abdominal Hernias and Gastric Volvulus pattern. This finding, in association with bowel obstruction, may increase the suspicion for an internal hernia. Internal supravesical hernias produce symptoms of bowel obstruction. Associated symptoms of bladder compression occur in about 30% of cases. Anterior supravesical hernias may result in a suprapubic mass or tenderness. About 50% of patients also have an inguinal hernia. Barium radiography or abdominal CT scanning with oral contrast may be helpful in the diagnosis.193,194 Hernias of the broad ligament of the uterus cause symptoms of bowel obstruction in about 50% of cases. Other cases are discovered incidentally at surgery. Small bowel, sigmoid colon, appendix, omentum, and ureter have been reported to herniate. CT scanning may show dilation of small bowel and deviation of the uterus.198,199 Retroanastomotic hernias cause symptoms and signs similar to those of other internal hernias. Postgastrojejunostomy hernias cause symptoms of gastric outlet obstruction. The efferent loop herniates most often. Afferent loop hernias are one cause of the afferent loop syndrome (see Chapter 53). About 50% of postgastrojejunostomy hernias occur within the first month after surgery, 25% occur during the first year, and the rest occur later. The physical examination is not specific. The serum amylase level is often elevated with afferent limb obstruction. Plain abdominal radiographs may show gastric distention and a fluid-filled loop. Barium upper gastrointestinal radiographs are most useful for documenting efferent limb obstruction versus afferent limb obstruction. Sonography or CT scanning may show dilation of the afferent limb. Biliary scintigraphy will show excretion of radionuclide into the biliary tree but retention of the tracer in an obstructed afferent limb. The clinical presentation of post-RYGB hernias is similar to that of other internal hernias. Most commonly, bowel obstruction is present. Herniation of the afferent limb of the jejunojejunostomy (the limb that carries pancreaticobiliary secretions) can present an interesting diagnostic dilemma because this loop does not carry food material. Therefore, vomiting may not occur. As a consequence, herniation of the afferent limb may present with biliary obstruction and pancreatitis, rather than classic bowel obstruction. CT scans, biliary scintigrams, and plain films will show evidence of duodenal distention, with lack of progression of radionuclide into the distal small intestine on biliary scintigraphy. Herniation of the distal small intestine manifests with signs and symptoms of a bowel obstruction. Strictures at the base of the Roux limb can present with a similar obstructive syndrome. However, findings of a more distal bowel obstruction should increase suspicion for an internal hernia. Adhesive small intestinal obstructions are uncommon after laparoscopic RYGB. Therefore, internal hernias are the most common cause of small intestinal obstructions after RYGB.108,195

Treatment and Prognosis

Symptomatic internal hernias require surgery.180,191 Laparoscopic repair is preferred if the hernia is detected prior to complications.186,193 Once the patient has developed signs and symptoms of bowel obstruction, it is reasonable to

explore the patient, reduce the hernia, ensure that the bowel is viable, and repair the defect. Acute obstruction leads to strangulation, bowel ischemia, and death if not promptly treated. Paraduodenal hernias are usually corrected by incising the enclosing mesentery. Care must be taken to avoid injuring the superior or inferior mesenteric arteries because they follow an abnormal course within the border of the hernia. Sometimes, the small bowel can be reduced through the opening of the hernia without incising the mesentery.184-186 Thereafter, the paraduodenal defect must be closed. This may involve performing a formal Ladd’s procedure if the hernia is associated with a true malrotation (see Chapter 96). If there is a patulous paraduodenal space, a simple resection of the hernia sac and plication of the defect can afford adequate repair. Once incarceration has occurred, mortality can be higher than 20%.184,201 Therefore, it is recommended that all paraduodenal hernias be repaired electively, if possible.

KEY REFERENCES

Ballem N, Parikh R, Berber E, Siperstein A. Laparoscopic versus open ventral hernia repairs: 5-year recurrence rates. Surg Endosc 2008; 22:1935-40. (Ref 115.) Champion JK, Williams M. Small bowel obstruction and internal hernias after laparoscopic Roux-en-Y gastric bypass. Obes Surg 2003; 13:596-600. (Ref 195.) Fitzgibbons RJ Jr, Giobbie-Hurder A, Gibbs JO, et al. Watchful waiting vs repair of inguinal hernia in minimally symptomatic men: A randomized clinical trial. JAMA 2006; 295:285-92. (Ref 65.) Gerson LB, Triadafilopoulos G. Is colorectal cancer screening necessary in the preoperative assessment of inguinal herniorrhaphy? A casecontrol study. Am J Gastroenterol 2001; 96:1914-17. (Ref 98.) Gourgiotis S, Vougas V, Germanos S, Baratsis S. Acute gastric volvulus: Diagnosis and management over 10 years. Dig Surg 2006; 23:169-72. (Ref 44.) Israelsson LA. Parastomal hernias. Surg Clin North Am 2008; 88:113-25. (Ref 111.) Landreneau RJ, Del PM, Santos R. Management of paraesophageal hernias. Surg Clin North Am 2005; 85:411-32. (Ref 18.) Larson DW, Farley DR. Spigelian hernias: Repair and outcome for 81 patients. World J Surg 2002; 26:1277-81. (Ref 138.) Marsman HA, Heisterkamp J, Kazemier G. Re: Management in patients with liver cirrhosis and an umbilical hernia. Surgery 2008; 143:695. (Ref 131.) Moldrem AW, Papaconstantinou H, Broker H, Megison S, Jeyarajah DR. Late presentation of intestinal malrotation: An argument for elective repair. World J Surg 2008; 32:1426-31. (Ref 182.) Neumayer L, Giobbie-Hurder A, Jonasson O, et al. Open mesh versus laparoscopic mesh repair of inguinal hernia. N Engl J Med 2004; 350:1819-27. (Ref 86.) Oelschlager BK, Pellegrini CA, Hunter J, et al. Biologic prosthesis reduces recurrence after laparoscopic paraesophageal hernia repair: A multicenter, prospective, randomized trial. Ann Surg 2006; 244:481-90. (Ref 27.) Rathore MA, Andrabi SI, Bhatti MI, et al. Metaanalysis of recurrence after laparoscopic repair of paraesophageal hernia. JSLS 2007; 11:45660. (Ref 36.) Salameh JR. Primary and unusual abdominal wall hernias. Surg Clin North Am 2008; 88:45-60. (Ref 123.) Skandalakis LJ, Androulakis J, Colborn GL, Skandalakis JE. Obturator hernia. Embryology, anatomy, and surgical applications. Surg Clin North Am 2000; 80:71-84. (Ref 150.) Turner PL, Park AE. Laparoscopic repair of ventral incisional hernias: Pros and cons. Surg Clin North Am 2008; 88:85-100. (Ref 107.) Full references for this chapter can be found on www.expertconsult.com.

395

CHAPTE R

25 Foreign Bodies, Bezoars, and Caustic Ingestions Gregory G. Ginsberg and Patrick R. Pfau

CHAPTER OUTLINE Gastrointestinal Foreign Bodies  397 Epidemiology  397 Pathophysiology  398 History and Physical Examination  399 Diagnosis  399 Treatment  400 Specific Foreign Bodies  401 Procedure-Related Complications  404 Bezoars  404 Epidemiology  404 Clinical Features  405

Gastrointestinal foreign bodies (GIFBs) are comprised of food bolus impactions and intentionally and unintentionally ingested or inserted foreign objects. Bezoars are ingested materials (food or other materials) that accumulate in a normal or abnormal stomach. Caustic ingestions present following intentional or unintentional ingestion of acid or alkaline materials, which may result in acute and/or chronic injury to the esophagus and stomach. These topics are discussed in detail in this chapter.

GASTROINTESTINAL FOREIGN BODIES GIFBs are a fairly common problem encountered by gastroenterologists. Most resolve without serious clinical sequelae.1 Older studies have suggested that between 1500 and 2750 deaths occurred in the United States secondary to GIFBs.2-4 More recent studies have suggested the mortality from GIFBS to be significantly lower, with no deaths reported in over 850 adults and only one death in approximately 2200 children with reported GIFB.5-11 However, regardless of imprecise morbidity and mortality rates, serious complications and deaths occur as a consequence of foreign body ingestions.12-14 Thus, because of their frequent occurrence and potential for negative consequences, it is important to understand which patients are at risk for GIFBs, know how to diagnose and treat GIFBs, and deal with their complications.

EPIDEMIOLOGY

GIFBs may result from unintentional or intentional ingestion. The most common patient group that unintentionally ingests foreign bodies is children, particularly those between ages 6 months and 3 years. Children account for 80% of true foreign body ingestions.15 Children’s natural oral curiosity

Diagnosis  405 Treatment  405 Caustic Ingestions  406 Epidemiology  406 Pathophysiology  406 Clinical Features  406 Diagnosis  407 Treatment  407 Late Complications  408

leads to placing objects in their mouth and occasionally swallowing them. Coins are the most common objects swallowed by children but other frequently swallowed objects include marbles, small toys, crayons, nails, and pins.6,10,16 Accidental ingestion may also occur in adults with dental covers or dentures because of loss of tactile sensation during swallowing.17 A not uncommon occurrence is a patient mistakenly ingesting one’s own dentures.18 Patients with altered mental status or sensorium, including the very old, demented, or intoxicated, are at risk for accidental foreign body ingestions (Fig. 25-1). Accidental coin ingestion has been noted in college-aged adults during a tavern beer drinking game called “quarters,” in which the coin becomes lodged in the esophagus.19 Finally, those in certain occupations, such as roofers, carpenters, seamstresses, and tailors are at risk of accidental ingestion when nails or pins are held in the mouth during work. The most common groups that intentionally ingest foreign bodies are psychiatric patients and prisoners.20 Ingestion in these groups of patients is often done for secondary gain; they often ingest multiple objects, multiple times, and often the most complex foreign bodies. Iatrogenic foreign bodies are increasing in prevalence because of complications from capsule endoscopy, migrated stents (esophageal, enteral, and biliary), and migrated enteral access tubes and bolsters.21,22 Esophageal food impaction is the most common GIFB requiring medical attention in the United States, with an incidence of 16/100,000.23 The vast majority (75% to 100%) of patients with an esophageal food impaction have an underlying predisposing esophageal pathology,24-25 most often peptic strictures, Schatzki’s rings, and, increasingly, eosinophilic esophagitis.26 Other causes that contribute to esophageal food impactions include altered surgical anatomy following esophagectomy, fundoplication, or

397

398

Section IV  Topics Involving Multiple Organs Hypopharynx Upper esophageal sphincter Level of aortic arch Gastroesophageal junction Pylorus Duodenum

Figure 25-1.  Endoscopic image of a bottle opener (in the stomach) ingested by an intoxicated patient. Ileocecal valve

Rectum and anus Figure 25-3.  Gastrointestinal areas of luminal narrowing and angulation that predispose to foreign body impaction and obstruction.

Figure 25-2.  Endoscopic image of bratwurst with sauerkraut impacted in the esophagus while the patient was tailgating at a football game.

bariatric surgery and motility disorders such as achalasia and diffuse esophageal spasm.27 Food impactions most commonly occur in adults in their fourth or fifth decade of life but are becoming more prevalent in young adults because of the rising incidence of eosinophilic esophagitis. Cultural and regional dietary habits influence GIFBs. Fish bone injury is common in Asian countries and the Pacific rim, whereas impactions caused by meats, including hot dogs, pork, beef, and chicken, are common in the United States (Fig. 25-2).28,29 Symptomatic rectal foreign bodies are more often the result of insertion through the anus rather than oral ingestion and transit. This is reported most commonly in young adult males.30 Rectal foreign bodies that come to medical attention are most commonly inserted with the intention of autoeroticism but may present following consensual sexual acts or sexual assault.31 Less common, but still prevalent, causes of rectal foreign bodies include concealment of illegal drugs during smuggling efforts, loss of objects during attempts by the patient to relieve constipation, and even reports of falling on objects.32

PATHOPHYSIOLOGY

The vast majority (approximately 80% to 90%) of gastrointestinal foreign bodies pass through the gastrointestinal (GI) tract without any clinical sequelae and cause no harm to the

patient.1,33 The remaining 10% to 20% of GIFBs will require endoscopic intervention and 1% of GIFBs may require operative therapy.5,34 True foreign bodies and food impactions can cause significant morbidity, with the most serious complications being bowel perforation or obstruction and ensuing death.3 Thus, it is important to understand the conditions in which complications associated with GIFBs are apt to occur to help stratify therapeutic interventions. Perforation and obstruction from GIFB can occur in any part of the digestive tract, but are more apt to occur in areas of narrowing, angulation, anatomic sphincters, or previous surgery (Fig. 25-3).35 The posterior pharynx is the first area in which foreign bodies may become entrapped and cause complications. In the hypopharynx, short sharp objects such as fish bones and toothpicks may lacerate the mucosa or become lodged.36,37 Once in the esophagus, there are four areas of narrowing at which food boluses and foreign bodies become lodged, including the upper esophageal sphincter, level of the aortic arch, level of the main stem bronchus, and gastroesophageal junction. These areas are all luminal narrowings of 23 mm or less.38 However, food impactions and foreign bodies more commonly lodge in the esophagus at areas of pathology, including rings, webs, or strictures. Multiple esophageal rings associated with eosinophilic esophagitis (see Chapter 27) contribute to esophageal food impaction at an increasing prevalence in young adults.26,39,40 Similarly, esophageal motor abnormalities (see Chapter 42) such as diffuse esophageal spasm, achalasia, or segmental variation in peristalsis may lead to food or foreign body impaction in the esophagus.41-44 Foreign body and food impaction in the esophagus generally have the highest incidence of overall adverse events, with the complication rate being directly proportional to how long the object is lodged in the esophagus.

Chapter 25  Foreign Bodies, Bezoars, and Caustic Ingestions Serious complications of esophageal foreign bodies include perforation, abscess, mediastinitis, pneumothorax, fistula formation, and cardiac tamponade.45,46 Once a GIFB passes through the esophagus, the vast majority will pass through the entire GI tract without further difficulty or complication. Exceptions are sharp, long, and large objects. Sharp or pointed objects may have a perforation rate as high as 35%. Large objects (larger than 2.5 cm [1 inch] in diameter) may not be able to pass through the pylorus. Long objects (longer than 5 cm [2 inches]) such as pens, pencils, and eating utensils, may not negotiate around the duodenal sweep. Objects may become obstructed in the small intestine at the ligament of Treitz or at the ileocecal valve. Adhesions, postinflammatory strictures, and surgical anastomoses within the small intestine may also be sites where foreign bodies lodge and become obstructed. However, most objects, even sharp ones, rarely cause damage once in the small intestine and colon because the bowel naturally protects itself through peristalsis and axial flow; these tend to keep the foreign body concentrated in the center of fecal residue, with the blunt end leading and the sharp end trailing.47,48 Inserted rectal objects are often tenaciously retained because of anal sphincter spasm and edema, making spontaneous passage of the object difficult. The angulation and valves of Houston may also impede passage of objects through the rectum.

HISTORY AND PHYSICAL EXAMINATION

The history from children or noncommunicative adults is often unreliable. Most gastric and up to 20% to 30% of esophageal foreign bodies in children are asymptomatic.49 Most of these present after having been witnessed or suspected by a parent, caregiver, or older sibling. However, in up to 40% of cases, there is no history of a witnessed ingestion.50 Thus, symptoms are often subtle in children, presenting as drooling, not wanting to eat, and failure to thrive. For communicative adults, history of the timing and type of ingestion is usually reliable. Patients are able to relate exactly what they ingested, when they ingested it, and symptoms of pain or/and obstruction. Patients with esophageal food bolus impactions are symptomatic, with complete or intermittent obstruction. They are unable to drink liquids or retain their own oral secretions. Sialorrhea is common. Ingestion of an unappreciated small, sharp, object, including obscured fish or animal bones, may cause odynophagia or a persistent foreign body sensation because of mucosal laceration. The type of symptoms can aid in determining whether an esophageal foreign object is still present. If the patient presents with dysphagia, odynophagia, or dysphonia, there is an 80% likelihood that a foreign body is present, causing at least partial obstruction. Symptoms of drooling and inability to handle secretions are indicative of a neartotal obstruction of the esophagus. If the symptoms are restricted to retrosternal chest pain or pharyngeal discomfort, less than 50% of patients will still have a foreign body present.51 Patient localization of where an ingested foreign object is lodged is not accurate, with only a 30% to 40% correct localization in the esophagus and essentially a 0% accuracy for foreign bodies in the stomach.49,52 Once the object reaches the stomach, small intestine, or colon, the patient will not report symptoms unless a complication occurs, such as obstruction, perforation, or bleeding. Patients with rectal foreign bodies are frequently asymptomatic,31 but obtaining an accurate history from the patient may be difficult because of the embarrassment associated with its insertion. Presentation is often after the patient or

another person has made multiple attempts to remove the object.36 Symptoms may include anorectal pain, bleeding, and pruritus, with a small number of patients presenting with more serious complications, including obstruction, perforation, and peritonitis. Past medical history is useful to identify previous foreign body ingestion, because these repeat offenders are likely to ingest multiple and more complex foreign objects. A history of dysphagia in a person with a food impaction or esophageal foreign body suggests a high likelihood of underlying esophageal pathology. Physical examination does little to secure the diagnosis or location of a retained foreign body. However, physical examination is crucial to identify already developed complications related to foreign body ingestion. Assessment of the patient’s airway, ventilatory status, and risk for aspiration are crucial prior to initiating therapy to remove a GIFB. A neck and chest examination looking for crepitus, erythema, and swelling can suggest a proximal perforation. An abdominal examination should be performed to evaluate for signs of perforation or obstruction.

DIAGNOSIS Diagnostic Imaging

Plain films of the chest and abdomen are recommended for patients presenting with suspected foreign body ingestion to determine the presence, type, number, and location of foreign objects present. Both anteroposterior and lateral films are needed, because lateral films will aid in determining if a foreign body is in the esophagus or the trachea53 and may detail foreign bodies obscured by the overlying spine in an anteroposterior film. Biplanar neck films are recommended if there is a suspected object or complication in the hypopharynx or cervical esophagus. Plain films are also useful in identifying complications such as free air, aspirations, or subcutaneous emphysema (Fig. 25-4).54 Unfortunately, radiography is unable to diagnose non­ radiopaque objects such as plastic, glass, or wood and may miss small bones or metal objects. The false-negative rate for plain film investigation of foreign bodies is as high as 47%, with false-positive rates up to 20%. Thus, if there is

Figure 25-4.  Radiograph of the chest demonstrating pneumomediastinum and bilateral pneumothoraces in a patient who developed esophageal perforation secondary to food impaction left untreated for longer than 24 hours.

399

400

Section IV  Topics Involving Multiple Organs continued clinical suspicion or symptoms, the individual should undergo further clinical investigation.55 Use of plain films in children is more controversial because of the inability of the child to give a history and the radiation associated with the x-rays. Some have suggested mouth to anus screening films to detect the presence of foreign bodies in children. To limit radiation, hand-held metal detectors have been used, with a sensitivity higher than 95% for the detection and localization of metallic foreign bodies.56 Barium studies are generally not recommended for the evaluation of GIFBs. Aspiration of hypertonic contrast agents in patients with complete or near-complete esophageal obstruction may lead to aspiration pneumonitis.57 Furthermore, barium may delay or impair the performance of a therapeutic endoscopic intervention by interfering with endoscopic visualization.58 Also, even if a barium study is considered normal, an endoscopy is still recommended if symptoms persist or the suspicion of a foreign body is high.36 Imaging such as computed tomography (CT) or magnetic resonance imaging (MRI) rarely is needed for the diagnosis of GIFBs. However, CT has been found to detect foreign bodies missed by other modalities59 and may aid in detecting complications of foreign body ingestion, such as perforation or abscess, prior to the use of endoscopy.60

Endoscopy

Diagnostic endoscopy provides the most precise means to diagnose suspected foreign bodies or food impactions. This ensures an almost 100% diagnostic accuracy for objects within the reach of the endoscope, including nonradiopaque objects and objects obscured by overlying bony structures that are not visualized by radiography. Endoscopy also allows the most accurate diagnosis of the underlying pathology, such as esophageal strictures, which may have led to a food impaction or impacted esophageal foreign body. Endoscopy also allows visualization of mucosal defects, abrasions, or ulcerations that may have resulted from the foreign body. Diagnostic endoscopy is also linked to the most efficacious therapy for GIFBs, the use of therapeutic endoscopy to remove or treat the object. Diagnostic upper endoscopy for foreign bodies is contraindicated when there are clinical or radiographic signs of perforation. Once an ingested foreign object has passed the ligament of Treitz, endoscopy is generally not indicated, because these objects will typically pass unimpeded with notable exceptions (see later). Similarly, most small (less than 2.5-cm) blunt objects in an adult patient’s stomach do not require endoscopic retrieval because most will pass without complication.

TREATMENT Nonendoscopic Methods

Treatment of gastrointestinal foreign bodies should always be planned with the knowledge that 80% to 90 % of GIFBs will spontaneously pass through the GI tract without complication.5,8 This has led some investigators to suggest that all foreign bodies can be managed with conservative observation.61,62 Although conservative management is effective in most cases of GIFB, it is more appropriate to perform selective endoscopy for treatment based on the location, size, and type of foreign body ingested.20,63 A number of medical therapies have been considered as primary treatment of esophageal foreign bodies and food impactions. The smooth muscle relaxant glucagon is the most widely used and studied drug for the treatment of

esophageal food and foreign object impactions. Glucagon, given in doses of 0.5 to 2.0 mg, can produce relaxation of the lower esophageal sphincter by as much as 60%, with the potential to permit passage of the impacted food or foreign body.64,65 Success with glucagon ranges from 12% to 58% in treating food impactions.66-68 Glucagon may cause nausea, vomiting, and abdominal distention. In addition, glucagon has little effect when a fixed obstruction is present, preventing passage of the foreign body. Nifedipine and nitroglycerin are not recommended because of hypotensionrelated side effects. The use of gas-forming agents such as carbonated beverages or preparations consisting of sodium bicarbonate and citric acid have been described for treating esophageal impactions. These agents are purported to release carbon dioxide gas to distend the lumen and act as a piston to push the object from the esophagus into the stomach.69 However, the effectiveness of this method is doubtful, and perforations have been reported associated with the use of gasforming objects.70 Similarly, the meat tenderizer papain is not recommended for the treatment of esophageal meat impactions because of lack of efficacy and risk of complications, including perforation and mediastinitis.71,72 Radiologic methods have been described for the treatment of esophageal foreign bodies. Under fluoroscopic guidance, Foley catheters, suction catheters, wire baskets, and magnets have been used to retract objects.73 The most commonly described device is the Foley catheter; the balloon tip of the catheter is passed distal to the object, inflated, and then the object is withdrawn into the oropharynx. Success of Foley catheter extraction of esophageal foreign bodies under fluoroscopy has been described as more than 90%. However, all radiographic methods suffer from lack of control of the object, particularly at the level of the upper esophageal sphincter and hypopharynx. Complications may include nosebleeds, laryngospasm, aspiration, perforation, and even death.74 Radiographic methods are generally recommended only if flexible endoscopy is not available.

Endoscopic Methods

Flexible endoscopy has become the treatment of choice for gastrointestinal food impactions and foreign bodies because it is safe and highly efficacious. Multiple large series have reported the success rate for endoscopic treatment of GIFBs to be more than 95%, with complication rates of less than 5%.5,58,63,75-78 The risk for complications is increased when sharp or multiple objects are ingested and when the ingestion is intentional as opposed to accidental. Because most GIFBs pass spontaneously without causing symptoms, it is important to understand the indications and timing for endoscopic intervention. Generally, all foreign bodies lodged in the esophagus require urgent intervention. The risk for an adverse outcome from an esophageal foreign body or food impaction is directly related to how long the object or food dwells in the esophagus.79 Ideally, no object should be left in the esophagus longer than 24 hours. Once in the stomach, most ingested objects will pass spontaneously and the risk of complications is much lower, thus making observation acceptable. There are notable exceptions, as follows. Sharp and pointed objects are associated with perforation rates as high as 15% to 35%.79 Objects longer than 5 cm and round objects wider than 2 cm also may not be passed and should be removed from the stomach with an endoscope at presentation or if they have not progressed in three to five days. If a more complex or sharp object has progressed beyond the stomach and cannot be retrieved, periodic radiographs should be obtained to docu-

Chapter 25  Foreign Bodies, Bezoars, and Caustic Ingestions Table 25-1  Equipment for Treatment and Removal of Gastrointestinal Foreign Bodies and Food Impactions ENDOSCOPES

OVERTUBES

Flexible endoscope Rigid endoscope Laryngoscope

Standard esophageal overtube 45- to 60-cm foreign body overtube

ACCESSORY EQUIPMENT Retrieval net Grasping forceps Dormia basket Polypectomy snare Transparent vacuum cap Latex protector hood Kelly or McGill forceps

ment progression through the GI tract.80 The patient should then be followed for any symptoms suggestive of obstruction or perforation, such as fever, tachycardia, abdominal pain, or distention. The type of sedation selected to facilitate endoscopy for the management of food impactions and ingested foreign objects should be individualized. Although conscious sedation is adequate for the treatment of most food impactions and simple foreign bodies in the adult population, anesthesia assistance may be required for uncooperative patients or patients who have swallowed multiple complex objects (see Chapter 40). For management of impactions and ingestions below the level of the laryngopharynx, flexible endoscopy is preferred.81 Rigid esophagoscopy and flexible nasoendoscopes can be used, but provide no additional benefit and are often available to only a few endoscopists.82,83 Laryngoscopes with the aid of a Kelly or McGill forceps can be useful for proximal foreign bodies and small sharp objects in the hypopharynx. Availability of and familiarity with multiple endoscopic retrieval devices for the removal of foreign bodies and food impactions is critical (Table 25-1). An endoscopy suite and/ or travel cart should be equipped with at least a grasping forceps, polypectomy snare, Dormia basket, and retrieval net.84 Overtubes of 45 and 60 cm in length should be available to the endoscopist. An overtube allows protection of the airway, multiple exchanges of the endoscope, and mucosal protection from sharp objects. The longer 60-cm overtube enables retrieval of sharp and complex objects from the stomach, bypassing the lower esophageal sphincter. An alternative adjunct for extraction of sharp objects is a latex protection hood, which fits onto the tip of the endoscope85,86 (see later). When planning for extraction of complex, sharp, or pointed ingested objects, and when opportunity permits, it may be valuable to go through an ex vivo dry run on a similar object when considering retrieval devices and extraction technique.5 Success and speed of retrieval of the foreign body have been shown to be directly related to endoscopist experience.87 When personnel or facilities are not available to accomplish relief endoscopically, consideration should be given to transferring the patient to another center.

SPECIFIC FOREIGN BODIES Food Impaction

Food impaction is the most common ingested foreign body in the United States.27 The most common foods to cause

impactions in the United States are meat products, including beef, hot dogs, and chicken. Fish bone impactions are more common in coastal areas and Asian countries. Imbibing alcohol while eating large cuts of meat may increase the risk for food impactions and has led to the terms backyard barbecue syndrome and steakhouse syndrome. Given that food boluses may pass spontaneously, the need for endoscopic intervention is based on the persistence of symptoms. Patients with signs of complete or near-complete obstruction with drooling or excessive salivation should undergo urgent upper GI endoscopy. Endoscopic intervention should be achieved at the latest within 24 hours of onset of symptoms and more ideally within the first 6 to 12 hours. An increased risk for complications is thought to be proportional to the duration of esophageal food impaction.1,17,88 The primary method to treat food impaction is the push method, with success rates well over 90% and with minimal complications.24 Before the food impaction is pushed into the stomach, an attempt to steer the endoscope around the food into the stomach should be made. Generally, if the endoscope can be passed around the food impaction into the stomach, the food impaction can be safely pushed into the stomach without difficulty. This also allows assessment of any obstructive esophageal pathology beyond the food impaction. Even if the endoscope cannot steer around the food impaction, gentle pushing pressure can be safely attempted. Larger boluses of impacted meat can be broken apart with the endoscope or an accessory prior to pushing the smaller pieces into the stomach safely. Eosinophilic esophagitis has increasingly been associated with esophageal food impactions (see Chapter 27). Reports indicate that food impaction in patients with eosinophilic esophagitis can be treated effectively and safely with the push method.39 However, care should be taken to minimize inducing mucosal tears.89 Food impactions that cannot be gently pushed into the stomach must be dislodged and withdrawn. Retrograde removal can be achieved with various retrieval devices, including snares, baskets, and forceps. Initial manual disruption of the food bolus into smaller pieces typically makes removal easier. An oroesophageal overtube is useful in such cases because it protects the airway and allows multiple exchanges of the endoscope during retrieval. A dedicated food bolus retrieval net can be useful for removing large pieces of food without the use of an overtube because the food can be satisfactorily secured within the net, thus reducing the risk of aspiration of the ingestate.90 Transparent plastic hoods or caps, such as those used to perform variceal band ligation and endoscopic mucosal resection, have been used successfully for the removal of large, tightly impacted meat boluses. With the cap secured to the tip of the endoscope, the device can be used to suction the food into the vacuum chamber and to withdraw the bolus per os.91,92 More than 75% of patients with food impactions have associated esophageal pathology.5,25 In addition, approximately half of patients with food bolus impactions have abnormal 24-hour pH studies and/or esophageal manometry. If an esophageal stricture or Schatzki’s ring is present after the food bolus is cleared, it can be safely and effectively dilated concurrently if circumstances allow. More often, mucosal abrasions or erythema exists from the food dwelling in the esophagus for an extended period and dilation is delayed for 2 to 4 weeks, during which patients should be prescribed proton pump inhibitor therapy. When multiple esophageal rings are present, biopsies should be obtained to evaluate for eosinophilic esophagitis.

401

402

Section IV  Topics Involving Multiple Organs

A

B

Figure 25-5.  A, A razor blade (in the stomach) ingested by a prisoner. B, Removal of the razor blade with a grasping forceps and overtube.

Nail and spoon

Figure 25-6.  Endoscopic image of a long spoon and nail that became impacted in the duodenal sweep after being swallowed by a psychiatric patient.

Sharp and Pointed Objects

Sharp and pointed objects may cause a perforation in up to 15% to 35% of patients and account for one third of all perforations from GI foreign bodies.93 Sharp and pointed objects, particularly toothpicks and animal bones, are the most likely ingested foreign objects to cause a perforation that necessitates surgical management.94 Patients with psychiatric illness and incarcerated patients are more likely to ingest more complex and multiple sharp and pointed objects, such as razor blades (Fig. 25-5), pins, needles, and writing and eating utensils (Fig. 25-6). Sharp and pointed objects retained in the esophagus are considered a medical emergency and should be removed within 6 to 12 hours. Moreover, any sharp and pointed object within the reach of the endoscope should be removed as well if this can be safely done (Fig. 25-7). When removing sharp and pointed objects, the foreign body should be grasped and oriented so that the pointed end trails on withdrawal to reduce the risk of perforation and mucosal laceration.95 For sharp and pointed objects, retrieval is best achieved with a grasping forceps, polypectomy snare, or biliary stone retrieval basket.87 All these devices can secure the object; orient the device used as described earlier. Retrieval nets

tend to shear in the removal of sharp objects and may compromise visualization. Use of an overtube should be considered to protect the esophagus and oropharynx. Long pointed objects can be grasped and directed into the overtube; the entire assembly, including the sharp and pointed object, endoscope, and overtube are then removed in unison. An alternative to an overtube for the extraction of sharp and pointed objects is a retractable latex hood that can be affixed to the tip of the endoscope (Fig. 25-8). When the endoscope is pulled back through the lower esophageal sphincter, the hood flips over the grasped object and protects the mucosa during withdrawal.85,96 Although associated with an increased risk of perforation, most sharp or pointed objects beyond reach of the endoscope will pass unimpeded and be eliminated through the GI tract without complication. Because of the increased risk of perforation, sharp and pointed objects should be followed by serial daily radiographs to ensure progression. If a sharp or pointed object fails to progress over three days, operative intervention should be considered.

Long Objects

Ingested objects longer than 5 cm (2 inches), and especially those longer than 10 cm (4 inches), have difficulty passing through the pylorus and duodenal sweep and can get hung up, causing obstruction or perforation at these locations (see Fig. 25-7). The most commonly ingested long objects are pens, pencils, toothbrushes, and eating utensils. Grasping forceps and polypectomy snares are the most commonly used devices to secure and remove long objects. Long objects should be grasped at one end and oriented longitudinally to permit removal. For extraction of long objects, use of the 60 cm overtube endoscope assembly, as described earlier, should be considered.

Blunt Objects: Coins, Disc Batteries, Magnets, and Bread Tabs

Small blunt objects, such as pieces of toys and coins, are the most commonly ingested objects by children. Disc (button) battery and magnet ingestions are uncommon but pose unique potential dangers. Blunt objects in the esophagus should be removed promptly. Impacted coins can result in pressure necrosis of the esophageal wall, resulting in perforation and fistula. A coin of any size can become lodged in the esophagus of children but ingested coins, in

Chapter 25  Foreign Bodies, Bezoars, and Caustic Ingestions

A

B

Figure 25-7.  Endoscopic images of shards of broken glass in the prepyloric antrum. (Courtesy of Dr. Jamie Anderson and Dr. Tushar Dharia, Dallas, Tex.)

A

B

Figure 25-8.  A, A latex protector hood with the hood pulled back. This position enables full visualization and allows the endoscopist to grasp a sharp object easily. B, As the protector hood is pulled back through the lower esophageal sphincter, the hood flips forward, protecting the gastrointestinal mucosa from the sharp object.

particular dimes and pennies measuring 17 and 18 mm, will usually pass through the adult esophagus. Polyp or dedicated retrieval nets allow capture and secure removal.87 Grasping forceps and biliary stone retrieval baskets are also effective. Standard biopsy forceps and snares are not recommended because they fail to secure coins reliably during extraction. If it is difficult to capture a blunt object in the esophagus, it is safe to push the object into the stomach, where there is more room to negotiate. Once a small blunt object enters the stomach, conservative outpatient management is appropriate for most patients.97 Exceptions to this include patients with surgically altered digestive tract anatomy and those who have ingested large blunt objects. In adults, the pylorus will allow passage of most blunt objects up to 25 mm in diameter, which includes all coins except half-dollars (30 mm) and silver dollars (38 mm). Otherwise, once in the stomach, a regular diet is appropriate, with radiographic monitoring every one to two weeks to confirm progression or elimination. If after three to four weeks a blunt object has not passed, endoscopic removal should be performed.98 Disc batteries are now contained in many small toys and electronic devices that are accessible to young children. Disc battery ingestion is of particular concern because batteries contain an alkaline solution, which can cause rapid liquefaction necrosis in the esophagus. Disc battery ingestion occurs most commonly in younger children but as few

as 10% will become symptomatic.99 Therefore, any clinical suspicion of a disc battery in the esophagus should prompt emergent endoscopy. Grasping forceps and snares are generally ineffective for disc battery removal, but use of a retrieval net permits successful removal in almost 100% of cases.100 Once in the stomach or small intestine, disc batteries rarely cause clinical problems and can be observed radiographically, with 85% passing through the GI tract within 72 hours.101 Small, brightly colored, coupling magnets have become popular as children’s toys. Ingested magnets within the reach of the endoscope should also be removed on an urgent basis. Although a single magnet will rarely be a cause of symptoms, concern exists if multiple magnets are ingested or if magnets were ingested with other metal objects. This can result in magnetic attraction and coupling between interposed loops of bowel with subsequent pressure necrosis, fistula formation, and bowel perforation.102,103 Removal should be performed urgently when the magnets are more likely to be within reach of a standard endoscope; this can be achieved with grasping forceps, retrieval net, or basket. Magnetic attraction to metallic retrieval devices may ease the task of removal. Bread tabs or bread bag clips are another otherwise seemingly innocuous blunt object that when ingested (usually unknowingly) have been associated with a high risk for gastrointestinal tract complications.104 Bread bag clips,

403

404

Section IV  Topics Involving Multiple Organs introduced in the 1950s, are ubiquitous in developed societies. Bleeding, bowel obstruction, and perforation have all been described. The small bowel is the most common site of impaction, where the arms of the clip tenaciously grasp the mucosa. Esophageal and large bowel injury have also been described. Management is problematic in that ingestion is typically not detected until complications arise. Bread bag clips are radiolucent and, as such, are not detected by conventional radiography. When recognized at endoscopy, an attempt at removal is justified, using a gasping forceps; however, operative intervention is commonly required. 52.13mm

Narcotic Packets

Ingested packets of illicit narcotics in the GI tract present in two general groups, termed body stuffers and body packers. Body stuffers are drug users or traffickers who quickly ingest small amounts of drugs, but in poorly wrapped or contained packages that are prone to leakage. Body packers are the mules used by drug smugglers who ingest large quantities of carefully prepared packages intended to withstand GI transit.105,106 These patients may present with intestinal obstruction because of the packages or symptoms related to the drug ingested. The latter may result in serious toxicity and death in 5% of these mules.107 Suspected patients are typically uncooperative and accompanied by law enforcement agents. Diagnosis is initiated with plain film radiology or CT scan, with multiple round or tube-shaped packets seen. Endoscopic removal is contraindicated because of the high risk of package perforation with resultant drug overdose.1 Observation on a clear liquid diet is recommended. Operative intervention is indicated when bowel obstruction or drug leakage is suspected.

Colorectal Foreign Bodies

Ingested objects uncommonly become lodged in the colo­ rectum. More commonly, colorectal foreign bodies were inserted into the rectum intentionally or unintentionally. Radiographs should be obtained prior to attempting removal of colorectal foreign bodies for better visualization of the location, orientation, and configuration of the object (Fig. 25-9). To avoid health care provider injury, attempts at manual removal or digital rectal examination should be deferred until the presence of a sharp or pointed object has been excluded. Manual digital extraction may be successful for the removal of small, blunt, palpable objects in the distal rectum. Conscious sedation may be adequate for manual removal in some patients, but examination and extraction under general anesthesia may be required in others to allow greater anal sphincter relaxation and successful object extraction. Nonpalpable and sharp or pointed objects should be removed under direct visualization with the use of a rigid proctoscope or flexible sigmoidoscope.108 Standard retrieval devices can be used as described earlier for the upper digestive tract. A latex hood or overtube can be particularly useful in removing long, sharp, pointed objects to protect the rectal mucosa from laceration and to overcome the tendency of the anal sphincter to contract on attempted removal of objects. Although conscious sedation will often facilitate removal, general anesthesia can allow maximum dilation of the anal sphincter to help remove larger and more complex objects.109 Operative intervention is indicated for any suspected complications secondary to a rectal or colon foreign body, including perforation, abscess, and obstruction. Complica-

229.20mm SUPINE

Figure 25-9.  A self-introduced rectal foreign body is apparent in a 73-yearold man presenting with lower abdominal pain. (Courtesy of Dr. William Beaujohn, Plano, Tex.)

tions are more common when the object is proximal to the rectum.110

PROCEDURE-RELATED COMPLICATIONS

Although the reported complication rate associated with endoscopic removal of gastrointestinal foreign bodies and food impactions is low (0% to 1.8%), it is thought to be much higher in practice.5,9,23,24,58,78 Perforation is the most feared complication, although aspiration and sedationrelated cardiopulmonary complications may also occur (see Chapter 40). Factors that increase the risk for complications include removal of sharp and pointed objects, an uncooperative patient, multiple and/or deliberate ingestion, and extended duration of time from food impaction or foreign body ingestion.12

BEZOARS Bezoars are collections of indigestible material that accumulate in the gastrointestinal tract, most frequently in the stomach. The three most common types of bezoars encountered are phytobezoars, composed of vegetable matter; trichobezoars, made up of hair or hair-like fibers; and medication bezoars (pharmacobezoars) (Fig. 25-10; Table 25-2).

EPIDEMIOLOGY

Phytobezoars are the most common type of bezoar. Offending fruits and vegetables include celery, pumpkin, prunes, raisins, leeks, beets, and persimmon.5 All these foods contain large amounts of insoluble and indigestible fibers, such as cellulose, hemicellulose, lignin, and fruit tannin.111 A phytobezoar develops when large quantities are ingested and accumulate.

Chapter 25  Foreign Bodies, Bezoars, and Caustic Ingestions seen in patients with gastric bezoars. Patients with diabetes or end-stage renal disease, and patients on mechanical ventilation, are all at greater risk for bezoar formulation.115

CLINICAL FEATURES

Figure 25-10.  Endoscopic image of a pharmacobezoar in a patient with a history of a pancreaticoduodenectomy who had obstructive symptoms. The pills were removed with an endoscopic net, with subsequent relief of the patient’s symptoms.

Table 25-2  Oral Pharmacologic Agents Associated with Medication Bezoar Formation Nonabsorbable antacids Bulk laxatives Cardiovascular medications Nifedipine Verapamil Procainamide Vitamins and minerals Vitamin C Vitamin B12 Ferrous sulfate Miscellaneous agents Sucralfate Guar gum Cholestyramine Enteral feeding formulations Theophylline Sodium polystyrene sulfonate (Kayexalate) resin

Trichobezoars occur most commonly in young women and children from ingestion of large amounts of hair, carpet fiber, or clothing fiber. Trichobezoars are more often associated with psychiatric disorders, mental retardation, or pica.112 Medication bezoars occur with fiber-containing medications, resin-water products, or extended-release medications designed to resist digestion.113 Medication bezoars can result in decreased pharmacologic efficacy when the active agent is trapped in the bezoar and cannot be absorbed or, alternatively, increased toxicity when the contents of a large gastric medication bezoar are released all at once into the small intestine. The vast majority of patients with bezoars (other than trichobezoars) have a predisposing factor that decreases emptying of gastric contents. Prior gastric surgery is evident in as many as 70% to 94% of patients with bezoars. Retained gastric contents may be observed in up to 65% to 80% of patients who have undergone a vagotomy with pyloroplasty.52 Bezoar formation after surgery results from delayed gastric emptying, decreased gastric accommodation, and reduced acid-peptic activity.114 Gastroparesis is commonly

Patients with gastric bezoars may be asymptomatic, but most (80%) have vague symptoms of epigastric discomfort.115 Associated anorexia, nausea, vomiting, weight loss, and early satiety may also be present. Bezoars can cause gastric ulceration secondary to pressure necrosis. Bezoarinduced gastric ulcers can cause bleeding and gastric outlet obstruction.116 Bezoars may also accumulate in the small bowel and usually present with mechanical obstruction. Rapunzel syndrome is a term used to describe trichobezoars located primarily in the stomach that extend past the pylorus and into the duodenum, causing bowel obstruction or even jaundice or pancreatitis because of obstruction at the level of the ampulla of Vater.117,118

DIAGNOSIS

The history is helpful in the diagnosis of bezoar, with a focus on the amount and types of food or medications consumed. A history of previous bezoar, gastric surgery, or gastric dysmotility should be considered. Physical examination usually assists little in the diagnosis, although occasionally a palpable abdominal mass may be appreciated. Halitosis may be present because of the putrefying materials of the bezoar residing in the stomach. Baldness and a patchy hair pattern may be present in patients who suffer from trichotillomania; they are consistently ingesting their own hair. A plain abdominal radiograph may demonstrate the outline of the bezoar. On contrast radiography, a gastric bezoar classically presents as filling defects within the stomach.111 Plain films and contrast studies will detect only 25% of bezoars detected at upper endoscopy, however. Therefore, gastric bezoars are definitively diagnosed with upper endoscopy. On upper endoscopy, phytobezoars are a dark brown, green, or black mass of amorphous vegetable material in the stomach. Trichobezoars tend to have a hard, blackened, and almost concrete appearance. Medication bezoars will be seen as whole pills or pill fragments in the midst of the material (see Fig. 25-10).

TREATMENT

Smaller bezoars may be treated with conservative medical management; usually, this consists of a liquid diet for a short period of time and a prokinetic agent to promote gastric emptying.111 Chemical dissolution, most commonly with cellulase, has been reported successful in up to 85% of patients with small bezoars.116 Cellulase can be taken as a tablet or instilled into the stomach as a liquid via an endoscope or nasogastric tube. Nasogastric lavage may aid in the physical dissolution of small bezoars. For larger bezoars and bezoars resistant to medical therapy, endoscopic therapy may be effective. The endoscope is used to fragment the bezoar into smaller pieces. Fragmentation can be performed with the endoscope itself, with accessory devices such as forceps or snares, or with the instillation of saline or water flushes through the endoscope. The fragments of the bezoar can be pushed into the small bowel or removed by mouth. If most of the bezoar is to be removed, an overtube is recommended to facilitate frequent passes of the endoscope and to protect the airway. Mechanical disruption and endoscopic removal will be successful in 85% to 90% of gastric bezoars. Resistant gastric bezoars may be treated with mechanical lithotripsy, electro-

405

406

Section IV  Topics Involving Multiple Organs hydraulic lithotripsy, Nd:YAG laser, or a needle-knife sphincterotome.119-121 Operative intervention may be needed if endoscopic therapy fails or if there is a complication related to the bezoar, such as a perforation, obstruction, or bleeding. Trichobezoars typically require surgery more often than phytobezoars. Gastric bezoars are usually removed via a small gastrostomy.116,118 Small bowel bezoars are removed via an enterotomy or can be transmurally milked to the cecum, where they rarely cause a problem in the larger diameter colon. When operative intervention is contemplated, care must be made to exclude multiple bezoars in more than one location. Preventing bezoar recurrence is as important as active treatment. If the underlying causes of bezoar formation are not corrected, recurrence is likely. Avoidance of high-fiber and other nondigestible foods should be followed. A starting dose of the enzymatic dissolution medication, cellulase, can be taken prophylactically by patients who have frequently recurring bezoars. Probiotics may be useful for patients with underlying motility disorders. In particularly refractory patients with recurring gastric bezoars, repeated periodic endoscopy with physical disruption of food material may prevent larger and clinically significant bezoar formation.

CAUSTIC INGESTIONS

Figure 25-11.  Barium esophagogram showing a stricture in the upper esophagus, with narrowing of the midesophagus, several weeks after a caustic ingestion. (Courtesy of Dr. Robert N. Berk, University of California, San Diego, Calif.)

EPIDEMIOLOGY

Approximately 5000 caustic ingestions are reported annually in the United States.122 Most caustic ingestions occur as accidental ingestions by children younger than 6 years.123 Caustic ingestions in adults occur as suicide attempts, in patients with mental health problems, and in the intoxicated person as a result of alcohol or recreational drug ingestion. Adults may tend to have more serious injuries because of ingesting larger amounts of caustic substances as compared with children, who will spit out or throw up the caustic agent that was swallowed. Broadly, two types of caustic agents are most commonly ingested, alkali agents or acidic agents. Alkali agents are most commonly household cleaners, including drain, toilet bowl, and oven cleaners. Lye ingestion is an alkali ingestion that contains sodium or potassium hydroxide. Alkali solutions are often odorless and tasteless, which can result in large amounts being swallowed accidentally. Finally, as noted, disc batteries may also cause alkaliinduced damage. Acid ingestion usually comes from swallowing toilet bowl cleaner, swimming pool cleaner, or battery acid. Acid ingestion often causes immediate pain, which results in the agent being rapidly expelled. Household bleach may contain both acid and alkali products, but rarely causes severe injury because of their diluted concentration.

PATHOPHYSIOLOGY Alkali

Alkaline ingestion causes a liquefactive necrosis that very rapidly extends through the mucosa, submucosa, muscularis of the esophagus, and stomach.124 Vascular thrombosis occurs following the necrosis. The initial alkali injury can be transmural and can result in perforation, mediastinitis, and peritonitis.125 External sloughing and ulceration occur a few days after ingestion. Finally, extensive granulation

tissue, fibroblastic activity, and collagen deposition occur over weeks, leading to chronic stricture formation (Fig. 25-11). With alkali ingestion, the esophagus is affected the most, with some limitation of damage in the stomach because of neutralization by acid; a minority of patients have damage in the small intestine as well.126 The degree of injury is also dependent on the agent ingested, its quantity, and how long the GI tract was exposed to the agent.127

Acid

Acidic agents cause a coagulative necrosis, with thromboses of mucosal blood vessels and a more limited superficial necrosis. Acidic agents are more apt to damage the stomach, particularly the antrum, more than the esophagus (Fig. 25-12). Furthermore, acidic agents tend to be ingested in a smaller quantity because of their offensive taste and immediate pain. Thus, acidic ingestions are associated with less overall damage compared with alkali agents.

CLINICAL FEATURES

Patients may present with oropharyngeal pain, epigastric pain, chest pain, dysphagia, or odynophagia. Oropharyngeal involvement can cause sialorrhea and drooling. Hoarseness, stridor, and dyspnea suggest injury to the epiglottis, larynx, and upper airway. Persistent chest or back pain may suggest esophageal perforation and mediastinitis, whereas severe abdominal pain can be related to gastric perforation and peritonitis. Of importance is that early signs and symptoms do not always correlate with the amount of caustic injury and likelihood for late complications.128 On physical examination, patients may have evidence of burns to the oral cavity with edema, ulceration, and exudate, but as many as 20% to 45% of patients will have normal physical examinations.129

Chapter 25  Foreign Bodies, Bezoars, and Caustic Ingestions

A

B

Figure 25-12.  Caustic injury to the esophagus and stomach by acid. A, After the ingestion of acid, the squamous mucosa of the esophagus has sloughed in a linear pattern. The esophageal mucosa is edematous and has a bluish discoloration. B, The gastric mucosa in this patient is hemorrhagic and edematous. (From Wilcox MC: Atlas of Clinical Gastrointestinal Endoscopy. Philadelphia: WB Saunders; 1995. p 85.)

Acute caustic ingestion

Table 25-3  Endoscopic Grade of Caustic Injury

Airway management and resuscitation

Chest and abdominal radiographs

GRADE

ENDOSCOPIC FINDINGS

I IIA IIB III

Edema and erythema Hemorrhage, erosions, blisters, ulcers with exudate Circumferential ulceration Multiple deep ulcers with brown, black, or gray discoloration Perforation

IV

Perforation EGD Grade of I or IIA injury No therapy

IIB or Ill Observe for perforation and stricture formation, with appropriate therapy

IV Surgery

Figure 25-13.  Algorithm for the approach to acute caustic injury. For the definitions of endoscopic grades of injury, see Table 25-3. EGD, esophagogastroduodenoscopy.

DIAGNOSIS

Radiologic images such as chest x-ray and abdominal films will not aid in the direct diagnosis or grading of severity of injury, but will indicate the presence of perforation by showing a pneumomediastinum, pneumothorax, or pneumoperitoneum. CT of the neck, chest, and/or abdomen should be considered when a high degree of suspicion remains for perforation, despite negative plain films. If perforation is present, surgery rather than endoscopy should be performed emergently (Fig. 25-13). Given the fact that symptoms and the physical examination may not match the degree of injury after a caustic ingestion, an upper endoscopy examination should be performed in the first 24 to 48 hours after ingestion in patients

without perforation.130 An upper endoscopy allows diagnosis of injury to the GI tract, permits grading of the degree of injury, establishes a prognosis, and can guide therapy (see Fig. 25-13). It is important to note that up to 40% to 80% of patients with a reported caustic ingestion will have no evidence of injury on endoscopic examination.131 The degree of injury seen on endoscopic examination can be graded and provides prognostic information (Table 25-3). Grades I and IIA burns, which correspond to first- and second-degree burns, will usually heal without sequelae.126 However, strictures will develop in 70% to 100% of patients with grade IIB injury, which causes circumferential ulceration, and grade III injury, with associated necrosis.132 Grade IV injury, with perforation, carries a mortality rate of up to 65% and requires urgent surgery.

TREATMENT

Initial management should address the ABCs of resusci­ tative management: airway, breathing, and circulation. Regional poison control should be contacted at 1-800-2221222. Once initially stabilized, management is based on the clinical status of the patient and grade of injury seen on endoscopy. Asymptomatic patients who have a normal endoscopic examination or only grade I or IIA injury can be started on oral intake in the first 24 to 48 hours and usually discharged within that same time frame. Clinically ill patients with hypotension, respiratory distress, and grade IIB (circumferential ulceration) or III necro-

407

408

Section IV  Topics Involving Multiple Organs sis on endoscopy should be admitted to an intensive care unit and managed with intravenous fluid resuscitation and close monitoring for evidence of perforation. Laryngoscopy should be performed in patients with respiratory distress. A patient with an edematous necrotic laryngopharynx should not undergo endotracheal intubation and will need a tracheotomy to maintain an airway. Emergency surgery is required with esophagectomy or gastrectomy for perforation. Colonic interposition is sometimes required. Operator and institutional experience affects mortality and morbidity of emergent esophagogastrectomy and is best undertaken at referral centers when circumstances allow.133 The need for and timing of operative intervention in patients with severe ulceration or necrosis without clear evidence of perforation remains controversial. Comparative analyses are difficult in this patient population. Some authorities have suggested that early operative exploration decreases mortality,134 but others cite lower mortality rates and complete healing in patients with nonoperative supportive care.127 As such, management must be considered on an individualized basis. Inducing emesis or placing a nasogastric tube to clear or dilute the GI tract of the caustic agent is contraindicated because it may re-expose the esophagus, oropharynx, and airway to the caustic agent. Induced retching and vomiting may increase the risk of perforation. The use of neutralizing agents is not recommended because they have not been shown to be efficacious, may lead to increased thermal injury, and may also promote retching and emesis.135 In addition, the routine use of glucocorticoids124 and systemic antibiotics122 is not recommended.

Figure 25-14.  Barium radiograph of a chronic antral stricture caused by a caustic ingestion. (Courtesy of Dr. Robert N. Berk, San Diego, Calif.)

Caustic exposure Seconds Necrosis Perforation

LATE COMPLICATIONS

Up to one third of caustic ingestion patients will develop esophageal stricture after initial recovery (see Fig. 25-11). Stricture formation presents most commonly at two months after injury but can occur at any time from two weeks to many years after the initial injury.126 Stricture formation occurs more commonly following more severe (grade IIB or III) injuries (see Table 25-3). The primary treatment of esophageal strictures secondary to caustic ingestion is frequent dilation. Endoscopic management of caustic strictures must be deliberate, with gradual and incremental progressive dilation to 15 mm or until symptom relief is obtained.136 The perforation rate is 0.5% for endoscopic dilation of chronic caustic strictures and as many as 10% to 50% of patients will eventually require operative intervention. Esophageal resection with an esophagogastric anastomosis, esophagojejunostomy, or colonic interposition may be considered.137 Case reports have described successful treatment with the use of temporary esophageal stents soon after caustic ingestion, but there is insufficient evidence to support recommendations for routine prophylactic stenting or for prophylactic early endoscopic dilation to prevent caustic-related strictures.135 Antral and pyloric strictures may also occur after caustic injury (Fig. 25-14). Antral and pyloric stenoses will usually develop one to six weeks after caustic ingestion, but can also occur years later.122 The risk of antral stenosis is also related to the degree of injury. Endoscopic dilation with the addition of acid suppression is successful in many patients, but many others will require antrectomy. Alkaline caustic ingestion, in particular, is associated with an increased risk for squamous cell cancer of the esophagus. Patients with a history of lye ingestion have a 1000-fold increased risk of developing esophageal cancer,

24-72 hours Ulceration 14-21 days Fibrosis Weeks-years Stricture Decades Carcinoma

Figure 25-15.  Sequence of the consequences of caustic injury to the gastrointestinal tract as a function of time after ingestion.

with a lag time from injury of approximately 40 years.138 Periodic endoscopic surveillance is advocated every one to three years, beginning 20 years after the caustic ingestion. Figure 25-15 summarizes the time course of complications from caustic ingestions reviewed in this section.

KEY REFERENCES

Alzakem AM, Soundappan SSV, Jefferies H, et al. Ingested magnets and gastrointestinal complications. J Pediatr Child Health 2007; 40:497-8. (Ref 102.) Clarke DL, Buccimazza I, Anderson FA, et al. Colorectal foreign bodies. Colorectal Dis 2005; 7:98-103. (Ref 32.) Gmeiner D, von Rahden BHA, Meco C, et al. Flexible versus rigid endoscopy for treatment of foreign body impaction in the esophagus. Surg Endosc 2007; 21:2026-29. (Ref 81.) Harikumar R, Humar S, Kumar B, et al. Rapunzel syndrome: A case report and review of literature. Tropical Gastroenterol 2007; 28:37-8. (Ref 118.)

Chapter 25  Foreign Bodies, Bezoars, and Caustic Ingestions Kerlin P, Jones D, Remedios M, et al. Prevalence of eosinophillic esophagitis in adults with food bolus obstruction of the esophagus. J Clin Gastroenterol 2007; 41:356-61. (Ref 26.) Lake JP, Essani R, Petrone P, et al. Management of retained colorectal foreign bodies: Predictors of operative intervention. Dis Colon Rectum 2004; 47:1694-8. (Ref 110.) Lin HH, Lee SC, Chu HC, et al. Emergency endoscopic management of dietary foreign bodies in the esophagus. Am J Emerg Med 2007; 25:662-5. (Ref 28.) Morrissey SK, Thakar SJ, Weaver ML, Farah K. Bread bag clip ingestion: A rare cause of upper gastrointestinal bleeding. Gastroenterol Hepatol 2008; 4:499-500. (Ref 104.)

Rodriguez-Hermosa JI, Codina-Cazador A, Sirvent JM, et al. Surgically treated perforations of the gastrointestinal tract caused by ingested foreign bodies. Colorectal Dis 2008; 10:701-7. (Ref 94.) Weissberg D, Refaely Y. Foreign bodies in the esophagus. Ann Thorac Surg 2007; 84:1854-7. (Ref 83.) Wong KKY, Fang CX, Tam PHK. Selective upper endoscopy for foreign body ingestion in children: An evaluation of management protocol after 282 cases. J Pediatr Surg 2006; 41:2016-18. (Ref 63.) Zhou JH, Jiang YG, Wang RW, et al. Management of corrosive esophageal burns in 149 cases. J Thorac Cardiovasc Surg 2005; 130:449-55. (Ref 133.) Full references for this chapter can be found on www.expertconsult.com.

409

CHAPTE R

26 Abdominal Abscesses and Gastrointestinal Fistulas Joseph P. Minei and Julie G. Champine

CHAPTER OUTLINE Abdominal Abscesses  411 Pathophysiology  411 Bacteriology  412 Diagnosis and Treatment  413 Outcomes  419 Gastrointestinal Fistulas  419

ABDOMINAL ABSCESSES PATHOPHYSIOLOGY

The development of an intra-abdominal abscess (IAA) occurs as a result of a host response to intra-abdominal bacterial contamination secondary to, or in conjunction with, various pathologic clinical entities. In 60% to 80% of cases, intra-abdominal abscess formation is associated with perforated hollow viscera, whether secondary to inflammatory disease such as appendicitis or diverticulitis or as a consequence of penetrating or blunt trauma to the abdomen.1-12 Other conditions associated with IAA formation include inflammatory bowel disease and complications of elective surgery (Table 26-1). Abscesses associated with solid organs such as the pancreas or liver are discussed in Chapters 58, 61, and 82. Clinical risk factors for the development of an IAA fall into two general categories: (1) factors related to the intraabdominal source of infection found at the time of surgery for peritonitis and that can be considered local factors (see Chapter 37); and (2) factors that may have been present prior to surgery (e.g., preexisting comorbidities) or are related to generalized care of the patient during surgery, which can be considered systemic factors. Table 26-2 lists the local and systemic factors associated with increased risk of abscess formation postoperatively. There is a delicate balance of opposing forces in the peritoneal cavity between bacterial factors and the host’s defense mechanisms, which attempt to clear bacterial contamination and localize infection (Table 26-3). These two opposing forces are often influenced by the presence of adjuvant factors (e.g., foreign material, fibrin) in the peritoneal cavity that often tip the balance toward bacterial infection with abscess formation (see Chapter 37). Once bacteria gain access to the peritoneal cavity through perforation of the intestinal wall, several factors come into play that determine whether an active infection is initiated. The typical bacteria that make up intra-abdominal infections have the ability to adhere to peritoneal surfaces and selectively grow and use host nutrients. These bacteria can undergo metabolic processes that are adapted to the host

Definitions and Classification  419 Pathophysiology  419 Diagnosis  420 Treatment  420 Outcomes  424

environment (e.g., obligate anaerobic metabolism). Furthermore, these bacteria have the capacity to resist antibiotic attack. Bacterial synergy plays an important role in the development of intra-abdominal infection (see later, “Bacteriology”).6 The peritoneum uses a number of host defenses to combat bacterial contamination.6,13 The balance of host defense factors in the setting of adjuvant factors determines whether contamination continues on to infection. Lymphatic clearance of bacteria is a major defense process that is so efficient that abscess formation occurs only when adjuvant substances such as hemoglobin, barium, or necrotic tissue are present.14 These adjuvant substances may block lymphatics (barium, fecal particulate matter), provide bacterial nutrients (iron from hemoglobin), or impair bacterial killing. Shortly after bacterial contamination, peritoneal macrophages are the predominant phagocytic cells. These cells are also cleared by the lymphatic system. As bacteria proliferate, polymorphonuclear leukocytes invade and become more numerous. The resultant peritoneal inflammation leads to an increase in splanchnic blood flow, with protein and fluid exudation into the peritoneal cavity. Procoagulatory effects of the inflammatory process and reduced levels of plasminogen activator activity enhance fibrin deposition and lead to entrapment of bacteria and localization of infection.13 These peritoneal defense mechanisms can have adverse effects. Lymphatic clearance of bacteria may be so brisk and effective that it results in a systemic response to bacteremia and sepsis. The exudation of fluid into the peritoneal cavity can lead to hypovolemia and shock; it can also dilute the opsonins needed in phagocytosis. Fibrin entrapment of bacteria can impair antimicrobial penetration and phagocytic migration with the potential to localize infection and lead to abscess formation.13 However, attempts to alter this balance of defense mechanisms are still not fully understood. In a study using a rodent intraperitoneal abscess model, recombinant tissue plasminogen activator (rt-PA) was used to increase intra-abdominal fibrinolytic activity. Rats treated with rt-PA had significantly fewer abscesses than controls, but they had significantly more bacteremic

411

412

Section IV  Topics Involving Multiple Organs Table 26-1  Causes of Intra-abdominal Abscesses Abdominal trauma Appendicitis Cholecystectomy and other operations Crohn’s disease Diverticulitis Neoplastic disease Pancreatitis Perforated hollow viscus (e.g., duodenal or gastric ulcer)

Table 26-2  Clinical Risk Factors for Intra-abdominal Abscess Formation Systemic Factors Chronic glucocorticoid use Increasing age Malnutrition Preexisting organ dysfunction Transfusion Underlying malignancy Local Factors Delay in performing surgery for underlying disease Formation of an ostomy Nonappendiceal source of infection Severity of illness, infection Severity of trauma

Table 26-3  Factors Influencing Transition from Bacterial Contamination to Infection BACTERIAL FACTORS

ADJUVANT FACTORS

HOST DEFENSE FACTORS

Adherence capacity Invasiveness Metabolic systems Resistance to antibiotics

Barium Blood Fecal matter Fibrin Foreign material Necrotic tissue

Fibrin sequestration Lymphatic clearance Lymphocyte response Neutrophil influx Peritoneal macrophages

Adapted from Fartham EH, Schoffel U. Epidemiology and pathophysiology of intra-abdominal infections (IAI). Infection 1998; 26:329-34.

episodes and higher mortality rates.15 Further work in this area by the same group used similar rodent models of intraperitoneal infection to study the role of a hyaluronic acid solution in abdominal adhesion and abscess formation. In bacterial peritonitis, intraperitoneal hyaluronic acid solution in the presence of antibiotics reduced the development of adhesions and abscess formation without increasing mortality.16 Possible mechanisms of action include mechanical separation of wound surfaces, improvement of peri­ toneal healing, modulation of the inflammatory response, and enhanced fibrinolysis.17 The potential of hyaluronatebased agents to reduce intra-abdominal adhesions and abscesses in abdominal surgery and sepsis is a promising new concept. Studies have suggested that the formation of adhesions is a complicated process that is not only dependent on surface apposition but is also under the tight control of positive and negative T cell costimulation.18 This was exemplified by a clinical trial19 of more than 1700 patients undergoing abdominopelvic surgery of the intestine, most for complica-

tions of inflammatory bowel disease. Patients were randomized to an adhesion barrier (Seprafilm [modified sodium hyaluronic acid and carboxymethyl cellulose], Genzyme, Cambridge, Mass) or control (no intervention) placed at the time of abdominal closing. Although abdominal and pelvic abscess rates were not different between the groups, there was a statistically higher rate of postoperative fistula formation and peritonitis in the Seprafilm group. This was particularly noted in patients who had the adhesion barrier wrapped around fresh intestinal anastomoses. Extensive experience has now accumulated with the use of Seprafilm. A recent meta-analysis20 of eight studies and over 4000 patients has shown that Seprafilm clearly decreases the severity and extent of postoperative adhesions. However, this came at the cost of increased rates of abdominal abscesses and anastomotic leaks. The effect was enhanced in the setting of surgery for inflammatory bowel disease. Newer preparations of hyaluronate-based membranes have also been associated with increased formation of postoperative intra-abdominal abscesses.21 Thus, the evidence for Seprafilm increasing postoperative abscess formation after surgery for acute peritonitis appears to be growing. The use of Seprafilm in this clinical situation to reduce postoperative abscess formation cannot be recommended.

BACTERIOLOGY

The bacteriologic factors associated with IAA formation depend on the circumstances of the initial peritoneal contamination. Patients who have abscesses that form in association with community-associated secondary peritonitis, such as perforated appendicitis or as a complication of penetrating abdominal trauma, often have very different microbial flora from those of the patient who has been in the intensive care unit (ICU) for a prolonged period and has been exposed to broad-spectrum antibiotics. The typical abscess that forms as a complication of secondary bacterial peritonitis, defined as loss of integrity of the gastrointestinal (GI) tract, is a mixed aerobic and anaerobic infection. In studies of isolates from subphrenic,22 retroperitoneal,23 and diverticular abscesses,24 a range of 2.9 to 3.7 bacterial isolates per abscess was recovered. The most common aerobes were Escherichia coli and Enterococcus species (range, 1.3 to 1.6 isolates per specimen). The most common anaerobes were Bacteroides fragilis and Peptostreptococcus species, which accounted for 50% to 75% of all anaerobes isolated. Other Bacteroides species and Clostridium species made up the remainder of anaerobes isolated (range, 1.7 to 2.1 isolates per specimen). In all three studies, most abscesses contained mixed aerobic and anaerobic flora (60% to 75%); the minority contained aerobic isolates only (10% to 20%) or anaerobic isolates only (15% to 20%). The number of anaerobic isolates always was higher than the number of aerobic isolates. Bacteroides species are important microbes in the formation of IAA. The existence of specific repeating negatively and positively charged cell wall polysaccharides on B. fragilis leads to a host response that results in the formation of an IAA. This host response is T cell–mediated and abscess formation can be experimentally prevented by vaccination with these repeating polysaccharide units. This vaccination does not appear to be antigen-specific in the traditional sense. Rather, the protective ability of these polysaccharides is conferred by, and perhaps specific for, a motif of oppositely charged groups. Vaccination with B. fragilis capsular polysaccharide complex significantly reduced the mortality rate and intra-abdominal abscess formation in a rat cecal ligation and puncture model.25 The cellular mechanism of IAA formation by B. fragilis has been elucidated.25 B. fragilis

Chapter 26  Abdominal Abscesses and Gastrointestinal Fistulas capsular polysaccharide complex adheres to peritoneal mesothelial cells and interacts with T cells and peritoneal macrophages to produce proinflammatory cytokines and chemokines, with subsequent expression of intercellular adhesion molecule-1 (ICAM-1) on host cells and recruitment of polymorphonuclear leukocytes to the abdominal cavity. Thus, the role of the capsular polysaccharide complex is to promote adhesion of B. fragilis to the peritoneal wall and coordinate the cellular events leading to the development of abscesses. The bacteria associated with intra-abdominal infections and abscesses in patients in the ICU who have been subjected to broad-spectrum antimicrobial selection pressure are different from those in patients with abscesses that result from secondary bacterial peritonitis. Thus, the microbiologic agents that cause tertiary peritonitis, defined as persistent intra-abdominal sepsis with or without a discrete focus of infection, generally after an operation for secondary peritonitis, are no longer E. coli and B. fragilis (see Chapter 37). Rather, nosocomial infections with resistant gramnegative organisms, Enterococcus species, and/or yeast are more common.26,27 The microbiologic analysis of abscesses in severely ill patients (Acute Physiology and Chronic Health Evaluation [APACHE] II score > 15) revealed that 38% had monomicrobial infections. The most common organisms were Candida (41%), Enterococcus (31%), and Enterobacter (21%) spp. and Staphylococcus epidermidis (21%); E. coli and Bacteroides spp. accounted for only 17% and 7%, respectively.28

Figure 26-1.  Computed tomography scan shows a large left upper quadrant abscess (closed arrow) with an enhancing wall and mild adjacent inflammatory changes. The thin collapsed bowel is noted laterally (open arrow).

DIAGNOSIS AND TREATMENT

The optimal management of the patient with an IAA includes the following: (1) accurate diagnosis and localization of the collection; (2) removal or control of the source of peritoneal contamination; (3) drainage of any established collections; (4) elimination of residual contamination of the peritoneum through antimicrobial therapy; and (5) physiologic support of the patient.13 The symptoms and signs of IAA are nonspecific, and a high level of vigilance is needed to make the diagnosis. Fever and elevated leukocyte count are frequent but nonspecific findings. Abdominal pain, tenderness to palpation, distention, and a palpable mass are also common findings. Suspicion of the presence of an IAA warrants further diagnostic imaging.

Diagnostic Imaging

Computed Tomography Computed tomography (CT) with intravenous and oral contrast medium is the imaging modality of choice for the diagnosis of most abdominal abscesses. Administration of intravenous contrast medium is useful to demonstrate the enhancing wall of an abscess and to define adjacent vascular anatomic characteristics clearly. Intravenous contrast medium is also necessary to evaluate completely for hepatic, splenic, pancreatic, or renal abscesses. Use of oral contrast medium is imperative to help differentiate an abscess from fluid-filled nonopacified bowel that may mimic an abscess. Ideally, oral contrast medium should be administered up to two hours before the study to opacify small and large bowel completely. Because inability to administer bowel contrast medium orally can limit the diagnostic ability of CT, alternative administration through a nasogastric or other enteric tube should be considered in patients unable to drink oral contrast medium. Rectal contrast medium may also be administered if initial images are equivocal. The CT diagnosis of abdominal abscess is suggested by identification of a loculated fluid density in an extraluminal

Figure 26-2.  Computed tomography scan demonstrates a right lower quadrant mass with an air-fluid level (closed arrow) containing an appendicolith (open arrow) compatible with a periappendiceal abscess. Prominent inflammatory changes are noted in the adjacent right abdominal wall.

location. Extraluminal gas within an abdominal mass is highly suggestive of an abscess, although necrotic tumors and resolving hematomas may occasionally exhibit this finding. Wall enhancement and adjacent inflammation favor the likelihood of infection in fluid collections (Fig. 26-1). Any fluid collection on CT should be clearly differentiated from nonopacified bowel. Delayed images are often necessary to allow bowel to opacify fully and to allow the investigator to distinguish an abscess from bowel confidently. The fluid in an abscess may occasionally be higher in density when proteinaceous material is present or when the collection represents an infected hematoma. Phlegmonous inflammatory tissue does not exhibit fluid density; rather, it is solid in appearance, often with inhomogeneous enhancement. In some cases, the CT appearance can suggest the cause of the abscess. Periappendiceal abscesses commonly have a characteristic location in the right lower quadrant adjacent to the cecum and may demonstrate an appendicolith (Fig. 26-2). Peridiverticular abscesses are often associated

413

414

Section IV  Topics Involving Multiple Organs

Figure 26-3.  Peridiverticular abscess demonstrated by a computed tomographic scan showing a small interloop fluid collection (long arrow) with an adjacent thickened sigmoid colon harboring small diverticula (short arrow).

with an inflamed adjacent colon demonstrating diverticula (Fig. 26-3). Abscesses associated with Crohn’s disease may demonstrate adjacent thickened small bowel. Although the CT appearance may strongly suggest an abscess, the diagnosis can be made with certainty only by obtaining a sample of the fluid through diagnostic aspiration. CT can be used to guide diagnostic aspiration or percutaneous abscess drainage and has advantages over other modalities in accurately identifying intervening structures, particularly the location of bowel loops. Ultrasonography Ultrasonography can be used to diagnose abdominal abscesses, especially abscesses in the liver, spleen, or pelvis, because of the good visualization of these areas it provides. However, the usefulness of ultrasonography can be limited in the midabdomen, in which visualization is not optimal as a result of blocking of sound waves by bowel gas. Surgical dressings may also inhibit visualization by ultrasound. Unlike CT, portable ultrasonography can be performed and may be warranted in the initial imaging evaluation when the patient’s condition precludes transportation. The classic ultrasonographic appearance of an abscess is a localized rounded or oval area of decreased echogenicity with internal debris and a thick irregular wall (Fig. 26-4). Most abscesses exhibit fluid characteristics on ultrasonography, but some may appear solid as a result of thick debris. Internal septations may be seen and are better identified by ultrasound than by CT. Gas within an abscess is suggested when areas of increased echogenicity are present, with posterior shadowing. The shadowing behind a gas collection tends to be less distinct than the more defined shadowing identified behind calculi on ultrasound. There is considerable overlap of the ultrasonographic appearance of infected and sterile fluid collections, and diagnostic aspiration is necessary to differentiate them. Ultrasound can be used for guidance during some percutaneous drainage procedures; however, poor visualization of intervening structures such as bowel in the midabdomen may limit its usefulness in some anatomic areas. Superficial and large abscesses tend to be more amenable to ultrasound guidance than smaller and deeper abscesses. Abdominal Plain Films Abdominal films demonstrating nonspecific mass effect in a patient with a suspected abscess can suggest the diag-

Figure 26-4.  Abdominal ultrasonogram of a typical abscess (arrow) demonstrating central decreased echogenicity, thickened wall, and debris.

nosis of an abdominal abscess. The diagnosis is further substantiated if gas is visualized in an extraluminal location. A localized ileus may also be seen. The abdominal plain film is, overall, less sensitive and specific than CT, and significant abscesses can be obscured by overlying normal structures. Chest Radiography Sympathetic thoracic changes caused by an IAA may be demonstrated on chest radiography but are nonspecific. Abdominal abscesses can be associated with an elevated hemidiaphragm, pleural effusions, and atelectasis. Liver, splenic, and subphrenic abscesses are more likely to be associated with changes on the chest radiograph than abscesses in the mid- and lower abdomen. Nuclear Imaging Gallium 67 (67Ga) nuclear imaging has been used to localize abscesses throughout the body. However, uptake of 67Ga is nonspecific for infection and can be demonstrated in tumors. Furthermore, 67Ga can accumulate in normal structures such as the colon, limiting its use in the abdomen. Although indium 111 (111In)–labeled leukocyte scanning has a higher specificity for infection, the test is of limited usefulness in acute infections because of the time constraints of the imaging procedure. Initial images for 111In-labeled leukocyte scans typically are not obtainable until 18 hours after administration of the tracer, and delayed images for up to 72 hours may be necessary to make the diagnosis. Nuclear imaging techniques are not a first-line diagnostic study for an IAA, but can be helpful in further clarifying equivocal findings seen on the CT scan or when the initial CT scan is negative. Magnetic Resonance Imaging Magnetic resonance imaging (MRI) has limited usefulness in the diagnosis of abdominal abscesses because of scan length, limited availability, cost, artifacts, and nonspecific findings. Fluid-filled bowel can be difficult to differentiate from an abscess on MRI. Oral contrast agents can be used

Chapter 26  Abdominal Abscesses and Gastrointestinal Fistulas in MRI to differentiate bowel from intra-abdominal abscess, but these agents are not widely utilized.

Mode of Drainage

Once an IAA is diagnosed and localized, a decision must be made regarding the optimal drainage technique and route. Percutaneous abscess drainage (PAD) has been shown to have equivalent success rates and less risk compared with surgery,29 although no randomized prospective studies are available. Assuming the availability of a safe route to the abscess, as occurs in 85% to 90% of cases, PAD should be the drainage procedure of choice.30-36 In the past, multiloculated, poorly organized, and multiple abscesses were not considered good candidates for PAD because of higher failure rates in these cases. Indications for PAD have now broadened to include these more challenging circumstances in many institutions, although longer drainage duration and multiple interventions may be necessary to obtain success.30-37 Although abscesses associated with enteric fistulae have lower success rates with PAD, successful PAD can be achieved in many cases.31,38 Rates of spontaneous closure up to 57% have been reported when aggressive catheter management has been combined with nutritional support.39 Interloop or intramesenteric collections are often not accessible percutaneously and surgery is often necessary. PAD is not appropriate for uncontained perforations or diffuse peritonitis.31,32,35 If surgery is chosen as the drainage mode of IAA (see later), an extraperitoneal approach is desirable to prevent contamination of the entire abdo­ minal cavity.40 When feasible, some pancreatic abscesses or walled-off pancreatic necrosis (WOPN) can be drained endoscopically (see Chapter 61). Percutaneous Abscess Drainage Continuing advances in diagnostic imaging and percutaneous catheter development have allowed PAD of abdominal abscesses combined with systemic antibiotic therapy to become the standard initial treatment of abdominal absces­ ses.30-36,41,42 Success rates for PAD range from 70% to 93%.29,31,33,36-38,42 Most abdominal and pelvic abscesses can be safely accessed percutaneously. A safe route into the abscess should be chosen that avoids major vascular structures, bowel, and adjacent organs. In extreme circumstances, the liver and the stomach can be traversed.35 Small and large bowel should not be traversed with a catheter, making interloop abscesses often inaccessible percutaneously. With increasing experience of interventional radiologists, indications for PAD have expanded to include multiple abscesses, multiloculated abscesses, poorly defined collections, and more challenging access routes.30-36,38,43 Adjunctive thrombolytics can be used safely to increase success in septated abscesses or when thick debris is encountered.44 An inflammatory phlegmon without demonstrable fluid collection is not appropriate for percutaneous drainage. Some small fluid collections, typically less than 3 cm, also may not require a catheter and can be managed through percutaneous aspiration for diagnosis, followed by antibiotic therapy. Catheter management is generally preferred for larger collections in most institutions; however, one-step percutaneous needle aspiration of abdominal and pelvic abscesses combined with systemic antibiotics has also been advocated as an alternative to catheter placement in larger abscesses.45 Contraindications to PAD include lack of a safe access route and uncorrectable coagulopathy.32 Coagulation studies and correction of any coagulopathy are recommended before the procedure to reduce the risk of uncontrollable hemorrhage. Guidance for PAD can be accomplished with a variety of imaging modalities including CT, ultrasonography, and

fluoroscopy, or a combination of modalities. The imaging modality selected is dependent on the location and size of the abscess as well as operator preference. The most common imaging modality is CT because of its widespread use for the initial diagnosis of abdominal abscess and its superb visualization of bowel and vascular anatomy. Ultrasonography can provide more real-time visualization during catheter insertion and can be useful when extreme angling of the route is needed.34,35 After a safe percutaneous route is identified, the cavity is accessed using a trocar method or a needle and guidewire method. The tract is then dilated to a diameter approxi­ mating that of the planned catheter, and the catheter is advanced into the cavity. A sump-type double-lumen catheter is the most common catheter used. A 12- or 14-Fr catheter size is generally adequate to drain most abscesses, although a larger catheter size may be necessary for an abscess associated with a large amount of debris or hemorrhage.38 The catheter position should be confirmed by repeat imaging to ensure that all catheter side holes are within the abscess. The cavity is usually aspirated dry, followed by flushing with sterile saline solution to clear any residual debris. The catheter is then placed to suction drainage and secured to the skin. A sample of the fluid is generally saved for Gram stain and culture. Postdrainage Management The catheter is flushed daily with sterile saline solution to maintain patency. Catheter output and character should be documented daily. Clinical status should be monitored for adequate response by assessing temperature and leukocyte counts. PAD endpoints and decisions to obtain follow-up imaging studies depend on the clinical response, catheter drainage, and presence of suspected enteric communications. If the clinical response has been satisfactory and the catheter drainage has diminished to less than 20 mL/day, the catheter can be safely removed. If clinical response is inadequate, repeat imaging is warranted. Persistently high catheter output raises suspicion of a fistula. A catheter study performed by instilling water-soluble contrast medium through the catheter under fluoroscopy is the best method to assess for an internal fistula as the cause of high drainage output. If a fistula is located, the catheter can be repositioned adjacent to the opening into the bowel for better control of bowel effluent. Poor clinical response can also be caused by catheter dislodgment from the major abscess cavity, un­drained loculations, multiple abscesses, or new abscesses. Repeat CT can evaluate for these possible causes of poor clinical response and guide additional percutaneous interventions when appropriate. Thick debris may occlude the catheter and inhibit daily flushing. In this case, the catheter can be exchanged for a larger catheter.41,46 Complications of Percutaneous Abscess Drainage The complication rate of PAD ranges from 4% to 15%.29,38,41,47 Complications include transient sepsis, organ injury, hemorrhage, pneumothorax, peritonitis, empyema, and pain. Recurrent Abscesses Intra-abdominal abscess recurrence rates range from 1% to 9%.29,31,41,48 Even when an abscess recurs, repeat secondary PAD should be considered and can be curative. Success rates for secondary PAD up to 91% have been achieved in recurrent abscesses, although the mean duration of drainage to achieve success was significantly longer with the secondary procedure.31

415

416

Section IV  Topics Involving Multiple Organs

A

B

C Figure 26-5.  A, Computed tomography (CT) scan demonstrates a left subphrenic abscess postsplenectomy (arrow). B, A catheter (arrow) is placed within the abscess. C, CT scan several days later demonstrates the catheter in the subphrenic space (arrow), with no residual abscess. The catheter was subsequently removed.

Drainage of Specific Abscesses Subphrenic Abscesses.  Subphrenic abscesses can be drained percutaneously, with careful attention to technique (Fig. 26-5). Avoidance of the pleural space is optimal to prevent pneumothorax and seeding of infection to the chest. The pleural space typically extends to the level of the eighth thoracic vertebra (T8) anteriorly, T10 laterally, and T12 pos­ teriorly. These guidelines can be used to prevent traversing of the pleural space. Some subphrenic fluid collections may not allow an extrapleural approach, in which case surgical risks should be weighed against the increased risk of empyema and pneumothorax posed by transpleural PAD. The safety of a transpleural approach has been debated.48,49 Pelvic Abscesses.  Anterior access to pelvic abscesses can be limited by intervening bowel, bladder, uterus, or vascular structures. A posterior transgluteal approach through the sciatic notch50 with the patient in the prone position has been used to drain deep pelvic fluid collections that are not accessible to an anterior approach (Fig. 26-6). Care must be taken to avoid the gluteal vasculature and the sciatic nerve. Ultrasound-guided transvaginal and transrectal drainage techniques have also been increasingly used for drainage of deep pelvic abscesses that are not accessible through other routes.51,52 A comparison of transrectal and transvaginal techniques has demonstrated better patient tolerance of the transrectal drainage route.53 Appendiceal Abscesses.  Periappendiceal abscesses can often be suggested by the CT appearance (Fig. 26-7A; see Fig. 26-2). Percutaneous abscess drainage has been increas-

Figure 26-6.  Prone computed tomography scan demonstrates transgluteal catheter placement through the sciatic notch into a deep pelvic abscess.

ingly accepted as the initial management of sepsis associated with a periappendiceal abscess, allowing the surgeon to perform a subsequent appendectomy, often laparoscopically, on an elective basis (see Fig. 26-7B).54,55 Peridiverticular Abscesses.  Percutaneous drainage of peridiverticular abscesses has also been increasingly accepted. Drainage can allow initial control of symptoms and obviate a diverting colostomy by allowing a one-stage rather than a two-stage procedure.56

Chapter 26  Abdominal Abscesses and Gastrointestinal Fistulas

A

B

Figure 26-7.  A, CT scan demonstrates a right lower quadrant abscess (closed arrow) with an appendicolith (open arrow). B, The same patient after placement of a percutaneous abscess drainage catheter (closed arrow). The abscess has resolved. The appendicolith (open arrow) is seen medial to the catheter.

Surgical Management

Surgical management of an infected patient is indicated when the patient is not a candidate for PAD secondary to multiple or interloop abscesses, or inability to access the cavity. Alternatively, surgery may be indicated for failure of PAD to control the source of infection. In these cases, care must be taken to adequately plan the route for surgical abscess drainage adequately. If possible, an extraperitoneal approach, such as the posterior approach to a left subphrenic abscess with 12th rib resection,57 will allow dependent drainage of the collection without contaminating the remainder of the peritoneal cavity. There are two groups of patients who pose difficult surgical challenges. These patients have an overwhelming intraabdominal infection noted at their first operation with significant bowel inflammation and edema,58 or they have failed initial therapy aimed at controlling secondary peritonitis and are now being managed for tertiary peritonitis.59,60 In both cases, these patients will often be treated with an open abdomen. Simply defined, the term open abdomen refers to a surgical technique in which the midline fascia is purposely not closed as part of a planned approach to severe abdominal infection that requires multiple repeat surgical débridements and washouts. These patients tend to be older, more critically ill, as measured by APACHE scoring, and tend to have more significant organ dysfunction scores. The use of the open abdomen with repeated washout strategy has been facilitated by the development of the vacuum-assisted closure device (VAC; KCI, San Antonio, Tex). This device consists of a porous foam pad connected to subatmospheric suction under an occlusive dressing. The VAC can be applied directly to the open abdomen and allows for continuous suction therapy to clear abdominal exudative fluids, maintain tension on the abdominal wall, allow protection of the skin from repeated dressing changes, and decrease patient discomfort associated with multiple daily dressing changes. Although large well-designed trials are still lacking, use of the VAC strategy has allowed improved wound management that has led to earlier definitive closure of the abdomen after using an open abdomen technique with repeat abdominal washouts.12,61 Whether this technique will ultimately decrease complications and improve survival is yet to be defined.62

Antibiotic Treatment

Once an intra-abdominal infection has been diagnosed and source control has been obtained with percutaneous or surgical techniques, elimination of residual infection within the peritoneum is carried out by the use of antibiotics. In general, antibiotics are effective only after an abscess has been drained because of a number of factors, including poor penetration of antibiotics into abscess cavities,63 very high bacterial counts in the abscess cavity (>108 colony-forming units [CFUs]/mL) that may alter bactericidal activity,64 and the fact that pus has an acidic pH and low Po2 from necrotic tissue and a poor blood supply.65 After abscess drainage, the initial choice of antibiotic should be based on the clinical picture and Gram stain findings of the abscess fluid. In an otherwise healthy individual who has a secondary bacterial peritonitis or abscess, antibiotic selection should be directed to the common organisms isolated from that type of abscess, primarily coliforms such as E. coli and anaerobes such as B. fragilis. This could be a second-generation cephalosporin, β-lactamase inhibitor– extended-spectrum penicillin derivative combination, or classic combination therapy with an aminoglycoside and antianaerobe. Recent drug shortages and product discontinuation by major pharmaceutical companies have limited the role of second-generation cephalosporins as first-line therapy. Furthermore, two meta-analyses have suggested that aminoglycosides plus anti-anaerobe combinations are less efficacious than a variety of newer antibiotic comparators.66,67 Although there was no difference in all-cause mortality or infection-specific mortality, there were statistically significant decreases in the clinical and microbiologic success rates in the aminoglycoside group. Because of significant resistance of B. fragilis to clindamycin, the antianaerobe of choice is metronidazole.68 Studies have documented the equivalence of broad-spectrum singleagent regimens such as carbapenems, extended-spectrum penicillin–β-lactamase inhibitor combinations, and fluoroquinolone or third-generation cephalosporin-metronidazole combinations (Table 26-4).4,11,69-73 The Therapeutic Agents Committee of the Surgical Infection Society (SIS) has developed a position paper for the antibiotic treatment of intraabdominal infection and noted that there was level 1 evidence to state that no regimen is superior to another.74 Newer agents that have been approved by the U.S. Food

417

418

Section IV  Topics Involving Multiple Organs Table 26-4  Antibiotic Choices in the Treatment of Intra-abdominal Infection Single-Agent Therapy Second-Generation Cephalosporins, Bacteroides Species–Active Cefoxitin Carbapenems Ertapenem Imipenem-cilastatin Meropenem Extended-Spectrum Penicillin–β-Lactamase Inhibitor Combinations Ampicillin sodium/sulbactam Piperacillin sodium–tazobactam Ticarcillin disodium–clavulanate Glycylcyclines Tigecycline Combination Therapy Antiaerobes

Aminoglycosides Tobramycin Gentamicin Third-Generation Cephalosporins Ceftriaxone Cefotaxime Ceftazidime (Pseudomonas coverage) Cefoperazone Fluoroquinolones Ciprofloxacin Levofloxacin Moxifloxacin Fourth-Generation Cephalosporins Cefepime antianaerobes

Metronidazole Clindamycin

and Drug Administration (FDA) after showing noninferi­ ority to standard antibiotic regimens in patients with intra-abdominal infections include a new carbapenem, ertapenem75,76; a fluoroquinolone, moxifloxacin77; and tigecycline, which is the first of a new class of antibiotics (glycylcyclines) based on tetracycline.78,79 In a severely ill patient with postoperative tertiary peritonitis and an elevated APACHE II score, the choice of empirical therapy can be more difficult. As noted earlier, infections in this population are often monomicrobial. Therefore, the results of the Gram stain can be of great importance in choosing initial antibiotics. Because many of these patients have already been exposed to broad-spectrum antibiotics, antibiotic selection must be made with knowledge of previous prescriptions and information on the resistance patterns in the ICU in which the patient is housed. Attention must also be paid to the underlying organ dysfunction of any individual patient, which can also affect antibiotic selection (e.g., the use of aminoglycosides in the setting of renal dysfunction should be avoided). A broadspectrum gram-negative coverage with a desirable sensitivity pattern should be considered. This would include a choice of carbapenems, extended-spectrum penicillins, or fluoroquinolones as appropriate therapy. Combination with a β-lactamase inhibitor–extended-spectrum penicillin is desirable under these circumstances. Data also suggest that the addition of an aminoglycoside does not add further activity in the treatment of intra-abdominal infections in this patient population.74 A recent Cochrane database analysis80 has shown that the addition of an aminoglycoside to broad-spectrum β-lactam therapy versus β-lactam monotherapy alone offered no survival benefit or increase in

clinical cure. Furthermore, the addition of aminoglycoside therapy was at the expense of an increased risk of nephrotoxicity. If gram-positive organisms are found on Gram stain, vancomycin therapy should be strongly considered for the treatment of potential methicillin-resistant Staphylococcus or Enterococcus spp. (see later). There is continued debate over the proper treatment of Enterococcus species isolated from IAA fluid. In an otherwise healthy population with minimal comorbid conditions, although no definitive conclusions are indicated by the literature on this topic, data support an antibiotic selection that does not specifically cover Enterococcus spp.81-83 This conclusion is based on the high success rate of regimens that do not have anti-Enterococcus spp. coverage, as high as that of regimens that do provide anti-Enterococcus coverage. For a critically ill patient in the ICU who has an elevated APACHE II score, the isolation of Enterococcus spp. takes on a different meaning, however. Several studies have shown that in a population of patients with an elevated APACHE II score, comorbid conditions, and early organ dysfunction, Enterococcus spp. isolation is an independent risk factor for treatment failure.74 Here, anti-Enterococcus spp. therapy is an important part of antibiotic selection. Combination therapy with a cell wall–specific antibiotic such as ampicillin in conjunction with an aminoglycoside has been shown to be synergistic as anti-Enterococcus spp. therapy.84 Recent development of resistance to β-lactam antibiotics as well as aminoglycosides has led to increasing use of vancomycin as anti-Enterococcus spp. therapy. Unfortunately, this has led to the development of new strains of Enterococcus spp. that carry plasmids encoding for vancomycin resistance (vancomycin-resistant enterococci, VRE). Once the organism has been identified and its sensitivity pattern reported, antibiotic selection can be focused. It is important to follow the patient’s response to abscess drainage and antibiotics. Continued deterioration, with repeated fever and white blood cell count elevation, should prompt a search for an explanation. Repeat CT scanning is warranted in this situation to look for an area of undrained or new infection.37,43,85 A second cause of a poor response is microbial resistance to the antibiotic selection. Thought must be given to broadening the antibiotic selection further in this case. Another reason for poor response to therapy is the possibility of fungal superinfection. As noted, Candida spp. infections constitute approximately 20% to 40% of infections in the setting of postoperative tertiary peritonitis.27,28 Candida, like Enterococcus species, do not need to be treated in secondary peritonitis (see Chapter 37), but in high-risk surgical patients with intra-abdominal infections, data support the use of fluconazole prophylaxis, which prevented invasive intra-abdominal Candida spp. infections and resulting sepsis in this group of patients with complicated conditions.86 Candida spp. are notoriously difficult to culture from blood and deep tissues. When they are isolated in this high-risk population, they should be aggressively treated with a systemic antifungal agent—amphotericin B, fluconazole,87 or the newer class of antifungals, the echinocandins, which show increased activity to nonalbicans Candida spp. Duration of antibiotic therapy depends on the underlying patient condition and the adequacy of, and response to, invasive drainage techniques. A classic study has evaluated the risk of recurrent sepsis after the termination of antibiotics.88 In the group of patients who were afebrile with a persistent leukocytosis at the end of therapy, there was a 33% recurrence rate of IAA. When both fever and leukocytosis were present, recurrent IAA occurred in 57%. However,

Chapter 26  Abdominal Abscesses and Gastrointestinal Fistulas when the patient was afebrile and had a normal leukocyte count, there was no IAA recurrence. These data, which have been confirmed, suggest that antimicrobial treatment of IAA be continued until the patient has a normal leukocyte count and is afebrile.89 This is the present recommendation of the Surgical Infection Society.74 Although duration of antibiotic therapy may not be altered, the use of oral conversion strategies may shorten the length of hospital stay and reduce cost. Studies have suggested that appropriate oral conversion therapy is as effective as intravenous therapy in those able to tolerate oral intake and that length of hospital stay, although not significantly different, tends to be shorter in the oral conversion group.4,71

OUTCOMES

Outcomes after treatment of IAA is dependent on a number of factors. The mortality rate has been reported to range from less than 5% for simple secondary bacterial peritonitis to approximately 65% or higher for complicated tertiary peritonitis.26,27,62,90-92 Simple abscesses associated with perforated appendicitis that responds to surgical drainage and antibiotics have a low mortality rate. Higher mortality rates occur in older patients, in those who have complex abscesses, high APACHE II and multiple organ dysfunction (MOD) scores, or a therapeutic delay, male gender, and those who use glucocorticoids or are otherwise immunosuppressed.60 Other risk factors include multiple reoperations to control intra-abdominal sepsis, malnutrition, poor physiologic reserve, high New York Heart Association class, and MOD syndrome.93 Retrospective studies have suggested that planned reexploration of the abdomen may be more successful in controlling intra-abdominal sepsis for patients with complicated infections. Furthermore, patients treated with an open abdominal technique may also have improved outcomes. Prospective randomized trials on techniques to manage complicated intra-abdominal sepsis do not presently exist. As noted, patients who show evidence of MOD have a particularly poor outcome, which has been thought to be secondary to the inability to control intra-abdominal infection. It has been suggested that continued intra-abdominal infection is another manifestation of organ failure and not a cause74—that is, patients die with infection, not of infection. Aggressive surgical, antibiotic, and supportive care is required in this group of patients, and patients may benefit from defined clinical pathways that minimize variability in practice.8,89 Studies have suggested a beneficial role for drotrecogin alfa (activated; recombinant human activated protein C [rhAPC]) in the treatment of severe intra-abdominal infections. The data from the initial PROWESS trial94 was reviewed by a surgical evaluation committee to evaluate the outcomes of surgical patients, with special emphasis on those with severe sepsis of abdominal origin.95 There was an absolute risk reduction of 9.1% and a relative risk reduction in 28-day mortality of 30% in those undergoing abdominal procedures. Although this did not reach statistical significance because of relatively small numbers, the group with an APACHE II score higher than 25 had an absolute mortality risk reduction of 18.2%. The rate of serious bleeding complications was not different between groups. Sub­ sequently, an analysis of the INDEPTH database,96 which is made up of five clinical trials of the efficacy of rhAPC, demonstrated a significant risk reduction in mortality (10.7% absolute, 34% relative) in surgical patients treated with rhAPC, over half of whom had intra-abdominal infections. This effect was only seen in those with an APACHE II score higher than 25. This finding was supported in a separate retrospective matched cohort study of patients

with septic shock of abdominal origin.97 These data, taken together, are highly suggestive of a benefit of rhAPC in the treatment of critically ill patients with severe sepsis of abdominal origin.

GASTROINTESTINAL FISTULAS DEFINITIONS AND CLASSIFICATIONS

A fistula is any abnormal anatomic connection between two epithelialized surfaces, a definition that includes many clinical entities. Because of this, fistulas are generally classified by anatomic and physiologic methods. Anatomic classifications rely on sites of fistula origin and drainage point. Inherent in this anatomic classification system is whether the fistula is internal or external. Physiologic classifications rely on fistula output in a 24-hour period (Table 26-5). Both fistula classifications are used clinically when describing a fistulous tract (e.g., a high-output enterocutaneous fistula). Compared with fistulas connected to the skin that are obvious, internal fistulas may be difficult to diagnose, depending on the organs involved. This would be the case, for example, in a cholecystoduodenal fistula, which might first be manifested by gallstone ileus. In a colovesical fistula, the presenting signs are urinary tract infection, fecaluria, and pneumaturia. Fistulas arising in the abdomen can originate from any epithelialized surface of a hollow viscus or drainage duct in the GI or genitourinary (GU) tract, liver, or pancreas. This chapter focuses on GI fistulas; for specific discussions of fistulas arising from the pancreatic or biliary duct, see Chapters 58, 59, 61, and 70.

PATHOPHYSIOLOGY

GI fistulas can occur spontaneously or postoperatively. Spontaneous fistulas account for 15% to 25% of fistulas and arise in association with inflammatory processes, cancer, and radiation treatment.98-106 Inflammatory processes include diverticulitis, inflammatory bowel disease, peptic ulcer disease, and appendicitis. These fistulas can be internal or external and, depending on cause and anatomic variations, have different rates of spontaneous closure. The remaining 75% to 85% of fistulas are almost always postoperative, external, and iatrogenic in origin.107-112 These occur after cancer surgery, emergency surgery in which bowel cannot be adequately prepared and cleansed, trauma surgery in which injuries may be missed, and reoperative surgery in which extensive lysis of adhesions and partial-thickness bowel injury occur. Risk factors for the formation of spontaneous or postoperative fistulas include malnutrition, sepsis, shock or hypotension, vasopressor therapy, glucocorticoid therapy, associated disease states, and technical difficulties with a surgical anastomosis.113 It is important to determine the cause of fistula formation because it often determines therapy. Fistulas that arise in

Table 26-5  Fistula Classification Anatomic Internal (e.g., ileocolic, colovesical) External (e.g., enterocutaneous) Physiologic High output (>500 mL/day) Moderate output (200-500 mL/day) Low output (<200 mL/day)

419

420

Section IV  Topics Involving Multiple Organs inflammatory bowel disease or from direct involvement of intestinal cancer are unlikely to close spontaneously and often require surgical correction. On the other hand, a postoperative low-output fistula arising from a partial anastomotic dehiscence frequently closes with appropriate conservative management. Conditions associated with the nonhealing of GI fistulas are listed in Table 26-6.

DIAGNOSIS

Once a fistula is suspected, early management should be directed to confirming the diagnosis. The anatomic site of origin and underlying cause can be determined when the patient’s condition is stabilized. One simple bedside maneuver to confirm the presence of an external fistula as the cause of suspicious postoperative wound drainage is to give the patient oral charcoal. The fistula can then be confirmed by the presence of charcoal in the suspicious drainage. Once confirmed, exact anatomic origins can be determined by radiographic contrast studies. These can include administration of contrast medium orally or rectally (depending on the site of suspicion) to define the site of origin via the bowel lumen (Fig. 26-8). Alternatively, contrast can be injected retrograde into the drainage site (fistulography) and followed to its site of origin in the bowel (Fig. 26-9). Internal fistulas can be diagnosed by injecting contrast medium into a hollow viscus (e.g., urinary bladder) with opacification of another viscus (e.g., rectosigmoid; Fig. 26-10).

ment strategy when treating GI fistulas. Once a diagnosis of an enterocutaneous fistula is confirmed (e.g., with enteral charcoal), early management is directed to fluid and electrolyte replacement. This can be a daunting task if the fistula has a high output (more than 500 mL/day; see Table 26-5). Output in excess of 1000 mL/day is not uncommon if the fistula originates in the proximal small bowel. To prevent intravascular volume depletion and electrolyte imbalance, fluid and electrolyte replacement must be a priority and should be addressed before more detailed diagnostic fistula studies are undertaken. Administration of replacement fluids should take into account the volume and electrolyte

TREATMENT

Treatment may be nonsurgical or surgical. Generally, nonsurgical treatment is the cornerstone of the early manage-

Table 26-6  Conditions Associated with Nonhealing Fistulas* Foreign body within the fistula tract (see Chapter 25) Radiation enteritis involving the affected bowel (see Chapter 39) Infection or inflammation at the fistula origin Epithelialization of the fistula tract Neoplasm at the fistula origin Distal obstruction of intestine *The acronym FRIEND can be used to remember these conditions.

A

Figure 26-8.  Lateral view of the rectosigmoid region on a barium enema radiography. A colovesical fistula (open arrow) secondary to diverticulitis is present. Diverticular disease (solid arrow) can be seen in the sigmoid colon. The bladder is shown with contrast pooling from the barium enema (long solid arrow).

B

Figure 26-9.  A, Fistulogram obtained through a percutaneous catheter (open arrow) demonstrates a fistulous tract (straight solid arrow) from the small bowel (curved solid arrow) to the skin. B, The same patient with the catheter (open arrow) advanced near the opening in the small bowel (solid arrow).

Chapter 26  Abdominal Abscesses and Gastrointestinal Fistulas

A

B

Figure 26-10.  Rectovesical fistula in a patient with Crohn’s disease, pneumaturia, and urinary tract infection. A, Catheter in the bladder with contrast beginning to fill the bowel. B, Contrast has filled the sigmoid colon and rectum through the fistulous tract. (Courtesy of Dr. Mark Feldman, Dallas, Tex.)

content lost through the fistula. Generally, fistula output is iso-osmotic and high in potassium. Therefore, output should be replaced milliliter for milliliter with a balanced salt solution that contains added potassium. If difficulties are encountered when managing electrolyte imbalances, a sample of fistula fluid can be sent to the laboratory for electrolyte determination. Subsequent electrolyte replacement can then be formulated on the basis of laboratory results. A second cornerstone of the early management strategy in the treatment of enterocutaneous fistulas is establishment of adequate drainage of external fistulas. This may require minor surgical maneuvers, such as opening a recent surgical incision to allow adequate drainage. As noted, percutaneous catheters are often essential in controlling a fistula. This point requires early attention because if a fistula cannot be controlled, pooling of fistula contents within the abdominal cavity can lead to infection with abscess formation and sepsis. Because most enterocutaneous fistulas occur postoperatively, some ingenuity may be required when trying to protect the skin from the caustic effects of the fistula output. Most acute postoperative enterocutaneous fistulas decompress through the surgical incision. As the incision shows signs of infection and drainage, it must be opened. A reopened incision that is draining intestinal contents is not amenable to simple placement of an ostomy bag to collect the drainage. An experienced enterostomal therapist should be consulted when dealing with this difficult problem. A recent adjunct in the management of enterocutaneous fistulas has been local wound care with the VAC system, described earlier. The use of the VAC device has made management of these difficult wounds more standardized. By applying the VAC device, control of the effluent and the open wound can be managed simultaneously.107,114-116 The VAC system has been able to protect the skin and decrease the fistulous output by the use of nonporous plugs to prevent fistula drainage, despite subatmospheric pressure applied by suction on the wound to aid in wound healing.

The use of the VAC system can effectively convert a highoutput fistula into a low-output state. This wound care strategy has allowed easier nonoperative management of fistulas, thus restoring bowel continuity resulting in the initiation of enteral feeding (see Chapter 5). Once the patient’s condition is stabilized in regard to fluids and electrolytes and the fistula is adequately drained, attention is turned to anatomic and diagnostic considerations to plan further therapy. Table 26-7 lists some prog­ nostic factors important in determining whether the fistula has a high or low rate of spontaneous closure. Spontaneous fistula closure is more likely for low-output fistulas, fistulas secondary to surgical complications, and fistulas arising anatomically in the proximal small intestine. Fistulas that ultimately require surgical closure are more often associated with high output, jejunal origin, and ongoing sepsis. Wellnourished patients without infectious complications are also more likely to experience spontaneous closing.102,117-120 When spontaneous closure is likely, nutritional evaluation and support must be aggressively pursued (see Chapters 4 and 5). The causes of malnutrition in the patient with a GI fistula are multifactorial, including underlying disease states, lack of protein intake, protein losses through the fistula, and underlying sepsis with hypercatabolism.113

Nutrition

Total parenteral nutrition (TPN) seems to be the natural first choice for a patient with an enterocutaneous fistula. Soon after diagnosis, aggressive caloric support must be given. Once the anatomic origin of the fistula is determined, route of feeding is considered. Not all patients must be placed on TPN, however. In a study of 335 patients with external fistulas, 85% were managed solely with enteral feedings. In a subgroup of uncomplicated fistulas, 50% healed spontaneously with this mode of nutritional therapy alone.121 Enteral feeding has been shown to enhance mucosal proliferation and villous growth through direct and indirect mechanisms. Nutrients in contact with the bowel mucosa also provide

421

422

Section IV  Topics Involving Multiple Organs Table 26-7  Prognostic Indicators of Successful Spontaneous Fistula Closure PARAMETER

SPONTANEOUS CLOSURE LIKELY

SURGICAL CLOSURE MORE LIKELY TO BE NEEDED

Output (mL/day) Age (yr) Site Nutritional status Cause

<500 <40 Proximal small bowel Well nourished Anastomotic breakdown

Anatomic characteristics Duration

Long fistulous tract Acute

≥500 >65 Distal small bowel or colon Malnourished Malignancy, inflammatory or infectious disease, complete anastomotic dehiscence Eversion of mucosa Chronic

Adapted from Berry SM, Fischer JE. Enterocutaneous fistulas. Curr Probl Surg 1994; 31:469-566; and Rombeau JL, Rolandelli RH. Enteral and parenteral nutrition in patients with enteric fistulas and short bowel syndrome. Surg Clin North Am 1987; 67:551-71.

direct stimulation to the enterocyte, and feedings high in glutamine may be particularly beneficial because glutamine is the main source of energy of the enterocyte.122 Furthermore, nutrients within the gut lumen release gut-derived hormones that have an indirect trophic effect on the intestinal mucosa. TPN, in contrast, has been shown to lead to gut mucosal atrophy. This may be the result, in part, of the fact that standard TPN solutions do not contain glutamine because it crystallizes out of solution. In a recent small study, 28 patients were fed an oral glutamine solution and the remainder of caloric intake was supplemented with TPN. Spontaneous resolution of fistula drainage was 13 times more likely in the oral glutamine-supplemented group.123 Despite the recent advances in enteral feeding of patients with GI fistulas, TPN remains the mainstay of nutritional support for most patients because they are unable to absorb sufficient calories enterally.124 In one study, Rose and associates reviewed 114 consecutive patients with GI fistulas, all treated with TPN and conservative therapy.125 They found that 61% of the fistulas closed spontaneously in an average of 26 days. The remainder continued on to surgical extirpation of the fistulous tract. The decision to support the patient with a GI fistula with enteral or parenteral nutrition has to be based on anatomic and physiologic considerations. If the fistula has a low output and is anatomically distal in the intestine, a trial of enteral feedings should be pursued. If the fistula is in the proximal intestine and distal access to the intestine has been established, as in many postoperative fistulas in which a feeding jejunostomy has been placed at the time of surgery, enteral feeding into the distal bowel should be considered. Along with this, infusion of the proximal fistula drainage into the distal bowel should be considered. Reinfusion of succus entericus into the distal bowel has been shown to make fluid and electrolyte management easier, as well as decrease the output of the proximal fistula.105,126 It is not mandatory to provide full nutritional support via the enteral route to obtain the benefits of enteral feeding. Protein and caloric requirements can be supplemented by TPN.

Somatostatin Analogs

Another potential adjunct to TPN in the management of the patient with a GI fistula is the use of a long-acting somatostatin analog such as octreotide or lantreotide. Octreotide has been shown to decrease fistula output by three mechanisms. First, it inhibits the release of gastrin, cholecystokinin, secretin, motilin, and other GI hormones. This inhibition decreases secretion of bicarbonate, water, and pancreatic enzymes into the intestine, subsequently decreasing intestinal volume. Second, octreotide relaxes intestinal smooth muscle, thereby allowing for a greater intestinal capacity. Third, octreotide increases intestinal water and electrolyte absorption.127

Initial studies evaluating the effect of octreotide on spontaneous intestinal fistula closure were uncontrolled, used historical controls, or were unblinded. These studies suggested that octreotide decreases fistula output, leads to improved spontaneous fistula closure rates, decreases time to spontaneous closure, and reduces mortality rate. However, randomized, placebo-controlled, double-blind studies using strict entry criteria had less favorable findings for octreotide.128-130 These studies, which had relatively small group sizes, showed no significant effects of octreotide on fistula closure, complication, or mortality rate. One consistent finding in some studies has been an improvement in healing time with octreotide, perhaps by decreasing a high-output fistula to a low-output fistula.131-133 Currently, the role of octreotide is limited to occasional use for high-output fistulas.

Management of Crohn’s Disease

Historically, conservative management of fistulas associated with Crohn’s disease has been uniformly unrewarding, as most abdominal and perianal fistulas required surgical correction. The observation that tumor necrosis factor-α (TNF-α) production in the intestinal mucosa is increased in patients with Crohn’s disease134 has led to the development and clinical investigation of chimeric monoclonal anti­bodies against TNF-α (e.g., infliximab [Remicade]) for the treatment of Crohn’s disease. In a randomized, multicenter, double-blind, placebo-controlled trial of 94 Crohn’s disease patients with draining abdominal or perianal fistulas of at least three months’ duration, 68% of patients receiving infliximab (5 mg/kg) had at least a 50% reduction in draining fistulas compared with 26% of patients receiving placebo. Furthermore, 55% of patients receiving infliximab had closure of all fistulas as compared with 13% of patients assigned to the placebo group.135 Initial trials with short-term infusions revealed the salutary effects of infliximab to be transient in most patients.136 A subsequent prospective, randomized, double-blind, placebo-controlled study revealed the benefit of a maintenance infusion of infliximab given every eight weeks. The group that received maintenance therapy had significantly longer periods without fistula drainage compared with controls (more than 40 weeks vs. 14 weeks in controls) and at more than one year on maintenance infliximab infusions, 36% were fistula-free versus 19% of controls.137 This was associated with a decreased rate of subsequent hospitalization and surgical and nonsurgical procedures.138 Subsequent data analysis from these and other trials has shown that the use of infliximab is associated with better outcomes for perianal than for abdominal wall fistulas,139 for external than for internal fistulas140 and, when given as early fistula treatment, rather than late in the disease process.141 The use of infliximab has not been associated with increased infec-

Chapter 26  Abdominal Abscesses and Gastrointestinal Fistulas New fistula confirmed

Volume replacement

Electrolyte correction

Acid/base balance correction

Nutrition support

Collect drainage; protect skin; apply VAC

Adequate drainage?

No

Yes Septic?

Yes

Surgical drainage or PAD

No Anatomic/diagnostic studies

Anatomically favorable for spontaneous closure? (see Table 26-7) No

Yes

Complex fistula, prior radiation, cancer

Crohn’s disease

Proximal intestine/high output

Distal intestine/low output

Supportive care until surgical closure safe

Trial of anti-TNF agent

Distal enteral access?

Conservative management (enteral feeding)

No TPN support

Open >6 weeks; surgery

Yes Distal enteral feeding, reinfuse fistula output

Spontaneous closure

Figure 26-11.  Algorithm depicting management of gastrointestinal fistulas. See text for details. PAD, percutaneous abscess drainage; TNF, tumor necrosis factor, alpha; TPN, total parenteral nutrition; VAC, vacuum-assisted closure.

tious complications in fistula patients.142 For the initial management of fistulas in Crohn’s disease, a trial of an anti–TNF-α monoclonal antibody regimen should be considered and has been supported by a recent Cochrane database review.143 There are a number of small series that advocate for the use of aggressive immunosuppression of fistulizing Crohn’s patients with methotrexate144 or tacro­ limus145 when other nonoperative therapy has failed. Fistula formation in Crohn’s disease is discussed in greater detail in Chapter 111.

scopic use of fibrin glue and other occlusive plugs. Although reports are limited to case series at this time, a variety of techniques, including fistuloscopy, fluoroscopy, and endoscopy, have been used to cannulate fistula tracts.146-149 Once cannulated, the tracts are débrided and then occluded with fibrin glue, collagen plugs, or gelatin sponges. Results have been encouraging in the small series evaluated, and the technique may serve as a useful adjunct for fistulas refractory to conservative management.

Adjunctive Techniques

Surgical therapy remains the mainstay of management of the complex fistula that is not a candidate for conservative management or has had a prolonged course of conservative

Another nonoperative approach to the management of refractory fistulas includes the percutaneous and endo-

Surgical Therapy

423

424

Section IV  Topics Involving Multiple Organs management (longer than six weeks) without resolution of fistulous output.102,105 Indications for early surgery include inability to control the fistula without surgical drainage, sepsis or abscess formation, distal intestinal obstruction, bleeding, and persistence of fistulous output not responsive to conservative management. Some more complex fistulas may require surgery to remove mesh or other foreign bodies before closure can be undertaken. The goal of surgical therapy is to resect the involved bowel and restore intestinal continuity.150 This surgery allows the patient to start eating through normal routes. Minimally invasive surgery was shown to be an option in selected patients in the surgical management of intestinal fistulas.146,151 Fistula recurrence after surgery has been noted in 10% to 20% of postoperative patients. Factors associated with fistula recurrence after surgery include Crohn’s disease, poor nutritional status, complex fistula associated with an open abdomen or mesh implantation or infection, advanced underlying disease states, and oversewing the fistula instead of resection and reanastomosis.99-101,107,119,150 Although innovative therapy and supportive care have resulted in improving spontaneous closure rates, management of these difficult problems requires a multidisciplinary approach that includes a nutritional support service, entero­ stomal therapist, surgeon, invasive radiologist, and gastroenterologist. An algorithm to manage GI fistulas is presented in Figure 26-11.

OUTCOMES

Early morbidity and mortality in the management of external fistulas result from initial fluid and electrolyte derangements that go unchecked. However, the major cause of mortality in patients with GI fistulas is sepsis with multiple organ failure. The typical setting for septic complications is provided by complex fistulas for which there is inadequate or uncontrolled drainage. In this setting, pooling of enteric contents occurs within the abdominal cavity and acts as a nidus of infection. Therefore, as noted, aggressive attempts must be made to ensure that fistulous drainage is well controlled. The mortality rate from all causes in patients with fistulas ranges from 10% to 30%.99,100,102,105,107,118,119,152 Higher mortality rates are seen in those who are malnourished, have had previous irradiation therapy, or have complex fistulas associated with a postoperative abdominal wall dehiscence.111 A second major cause of mortality in patients with GI fistulas is severe underlying disease, most often cancer. Often, patients who are termi-

nally ill secondary to malignancy forgo further operative procedures.153

KEY REFERENCES

Alivizatos V, Felekis D, Zorbalas A. Evaluation of the effectiveness of octreotide in the conservative treatment of postoperative enterocutaneous fistulas. Hepatogastroenterology 2002; 49:1010-2. (Ref 133.) Beck DE, Cohen Z, Fleshman JW, et al. A prospective, randomized, multicenter, controlled study of the safety of Seprafilm adhesion barrier in abdominopelvic surgery of the intestine. Dis Colon Rectum 2003; 46:1310-19. (Ref 19.) Behm BW, Bickston SJ. Tumor necrosis factor-alpha antibody for maintenance of remission in Crohn’s disease. Cochrane Database Syst Rev 2008; (1):CD006893. (Ref 143.) Eggimann P, Francioli P, Bille J, et al. Fluconazole prophylaxis prevents intra-abdominal candidiasis in high-risk surgical patients. Crit Care Med 1999; 27:1066-72. (Ref 86.) Farthmann EH, Schoffel U. Epidemiology and pathophysiology of intra-abdominal infections (IAI). Infection 1998; 26:329-34. (Ref 6.) Haffejee AA. Surgical management of high output enterocutaneous fistulae: A 24-year experience. Curr Opin Clin Nutr Metab Care 2004; 7:309-16. (Ref 100.) Lynch AC, Delaney CP, Senagore AJ, et al. Clinical outcome and factors predictive of recurrence after enterocutaneous fistula surgery. Ann Surg 2004; 240:825-31. (Ref 150.) Martinez JL, Luque-de-Leon E, Mier J, et al. Systematic management of postoperative enterocutaneous fistulas: Factors related to outcomes. World J Surg 2008; 32:436-443. (Ref 118.) Mazuski JE, Sawyer RG, Nathens AB, et al. The Surgical Infection Society Guidelines on Antimicrobial Therapy for Intra-Abdominal Infections: An Executive Summary. Surg Infect (Larchmt) 2002; 3(3):161-73. (Ref 74.) Paul M, Silbiger I, Grozinsky S, Soares-Weiser K, Leibovici L. Beta lactam antibiotic monotherapy versus beta lactam-aminoglycoside antibiotic combination therapy for sepsis. Cochrane Database Syst Rev 2006; (1):CD003344. (Ref 80.) Payen D, Sablotzki A, Barie PS, et al. International integrated database for the evaluation of severe sepsis and drotrecogin alfa (activated) therapy: Analysis of efficacy and safety data in a large surgical cohort. Surgery 2006; 140:726-39. (Ref 96.) Present DH, Rutgeerts P, Targan S, et al. Infliximab for the treatment of fistulas in patients with Crohn’s disease. N Engl J Med 1999; 340:1398405. (Ref 135.) Sands BE, Anderson FH, Bernstein CN, et al. Infliximab maintenance therapy for fistulizing Crohn’s disease. N Engl J Med 2004; 350:87685. (Ref 137.) Wainstein DE, Fernandez E, Gonzalez D, et al. Treatment of high-output enterocutaneous fistulas with a vacuum-compaction device. A tenyear experience. World J Surg 2008; 32:430-5. (Ref 116.) Wong PF, Gilliam AD, Kumar S, et al. Antibiotic regimens for secondary peritonitis of gastrointestinal origin in adults. Cochrane Database Syst Rev 2005; (2):CD004539. (Ref 67.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

27 Eosinophilic Disorders of the Gastrointestinal Tract Seema Khan and Susan R. Orenstein

CHAPTER OUTLINE Eosinophil: Role in Health and Disease  425 Eosinophils and the Gastrointestinal Tract  425 Clinical Entities  427 Eosinophilic Esophagitis  427 Eosinophilic Gastroenteritis  428 Food Protein–Induced Enterocolitis  429 Eosinophilic Proctitis  429 Evaluation  429 History and Physical Examination  430 Laboratory Evaluation  430 Allergy Evaluation  430 Radiologic Evaluation  430 Endoscopy and Pathology  430

Eosinophilic inflammation of the gastrointestinal (GI) tract occurs in primary eosinophilic GI disease (EGID), as well as secondary to other diseases. The classification of primary EGID is traditionally according to the sites of inflammation in the GI tract (Table 27-1).1,2 The best characterized of these EGIDs, eosinophilic esophagitis (EE) and eosinophilic gastroenteritis (EG), affect all ages and exhibit prominent eosinophilic tissue infiltration. Other noteworthy diagnoses within the spectrum of EGIDs, such as food protein–induced enterocolitis (FPIEC) and eosinophilic proctitis (EP), are uniquely pediatric diagnoses, and the pathology may be characterized by a mixed inflammatory infiltrate including dense tissue eosinophilia. The cause of these disorders may not be yet well understood, but they all have a strong association with allergies, and many respond to nutritional management (see Chapter 9).3,4 The collaborative efforts of gastroenterologists, allergists, and immunologists have made significant advances in understanding the immunopathogenesis of EGID in recent years. The publication of guidelines for the diagnosis and management of EE also became possible through such efforts.5 Eosinophilic inflammation also occurs secondarily in the GI tract in inflammatory bowel disease (IBD), autoimmune diseases, reactions to medications,6,7 infections, hypereosinophilia syndrome (HES), tumors, and after solid organ transplantation.8,9 These disorders should be considered in the differential diagnosis of the primary eosinophilic diseases and are briefly reviewed in this chapter. We first review the molecular mechanisms and key mediators in the inflammatory process in EGID, then discuss the clinical characteristics of the most important EGIDs, and finally suggest an approach to diagnostic evaluation and treatment.

Differential Diagnosis  431 Infections  431 Medications  432 Connective Tissue Disease and Vasculitis  432 Intestinal Polyps  432 Hypereosinophilia Syndrome  432 Inflammatory Bowel Disease  432 Celiac Disease  433 Transplantation  433 Treatment  433 Diet  433 Therapeutic Agents  434 Esophageal Dilation and Surgery  435

EOSINOPHIL: ROLE IN HEALTH AND DISEASE The eosinophil, a bilobed nucleated granulocyte, differen­ tiates from myeloid progenitor cells into its mature form containing brilliant birefringent cationic granules with a high affinity for the acidic dye eosin. It matures mainly under the influence of the hematopoiesis-specific tran­scription factors GATA-1, GATA-2, and PU.I, and c/EBP (enhancer-binding protein family).10 Cytokines, interleukin-3 (IL)-3, IL-5, and granulocyte-macrophage colony-stimulating factor (GM-CSF) strongly influence further development. IL-5, in particular, plays a role in the eosinophil’s differentiation and release from the bone marrow into the peripheral circulation, where it constitutes 2% to 4% of the granulocyte pool and has a circulating half-life of only 8 to 12 hours. The eosinophil then moves into resident tissues, mainly the GI tract, thymus, hematopoietic organs, and mammary glands. In the GI tract, eosinophils survive for about 1 week and finally undergo apoptosis.11

EOSINOPHILS AND THE GASTROINTESTINAL TRACT

As noted, eosinophils spend most of their lifespan in tissues, rather than circulating. The GI tract is the main nonhematopoietic organ in which eosinophils reside in the healthy state. In the GI tract, eosinophils are not homogeneously distributed. Highest concentrations are found in the cecum, ascending colon, and appendix. The esophageal epithelium is unique in being devoid of eosinophils under noninflammatory conditions.11,12 Eosinophils are normally present in the lamina propria of the gut, but the number of eosinophils

425

426

Section IV  Topics Involving Multiple Organs Table 27-1  Proposed Classification and Differentiation of Primary and Secondary Eosinophilic Gastrointestinal Diseases PRIMARY EGID Esophagitis (EE) Allergic (atopic) Idiopathic (nonatopic)

Gastroenteritis (EG) Mucosal Muscular Serosal

Colitis Food protein–induced enterocolitis Eosinophilic proctitis

SECONDARY EGID AND/OR DIFFERENTIAL DIAGNOSIS Esophagitis Isolated esophageal disorder GERD Extraesophageal disorders EG HES Autoimmune disease, vasculitis Connective tissue disease (e.g., scleroderma) Medications Transplantation Leiomyomatosis Gastroenteritis Infections (especially parasitic) HES IBD Celiac disease Autoimmune disease, vasculitis Connective tissue disease (e.g., scleroderma) Medications Transplantation Inflammatory fibroid polyps, polyposis Colitis Parasitic infections IBD Autoimmune disease, vasculitis Connective tissue disease (e.g., scleroderma) Medications Transplantation Juvenile polyps, polyposis, adenomas

EE, eosinophilic esophagitis; EG, eosinophilic gastroenteritis; EGID, eosinophilic gastrointestinal disease; GERD, gastroesophageal reflux disease; HES, hypereosinophilic syndrome; IBD, inflammatory bowel disease.

regarded as pathologic for various sites along the GI tract is debated. An array of stimulatory and proinflammatory factors mediate eosinophilic inflammation (Fig. 27-1). In the case of eosinophilic GI inflammation, an antigen exposure stimulates eosinophil synthesis, rolling, adhesion, diapedesis, and trafficking to the site of insult. The recruitment of eosinophils into GI segments is regulated by differential pathways involving a family of cell adhesion receptors called integrins. Investigations have shown that eosinophil movement into the small intestine and large intestine are controlled by α4β7-integrin and β2-integrin pathways, respectively.10,13,14 Eosinophils function as antigen-presenting cells and also affect the inflammatory process through specific eosinophil-derived granule proteins (EDGPs). These EDGPs include eosinophil cationic protein (ECP), eosinophilderived neurotoxin (EDN), eosinophil peroxidase (EPO), and major basic protein (MBP). These cationic proteins are cytotoxic to the human intestinal epithelium, possess antiviral and ribonuclease activity, and trigger degranulation of mast cells and release of cytokines (IL-1, IL-3, IL-4, IL-5, IL-13, GM-CSF, tumor necrosis factor-α [TNF-α], transforming growth factors), chemokines such as eotaxin-1 and RANTES (regulated on activation, normal T cell expressed and secreted), lipid mediators (leukotrienes, platelet-activating factor), and neuromediators (substance P, vasoactive intestinal polypeptide, nerve growth factor).15 Investigators have used various experimental models to explore the mechanisms whereby eosinophils mediate GI disease. An important puzzle includes the localization of the instigation of eosinophilic responses within the GI tract. The route of allergen exposure may determine the localization of the response. For example, oral or intragastric allergen exposure does not initiate EE but, in anesthetized mice, exposure to repeated challenges of aeroallergens induces marked EE in addition to lung eosinophilia. Interestingly, however, such aeroallergen challenge does not provoke eosinophilic inflammation in the stomach or small intestine

Respiratory tract

Allergens

GI tract

Figure 27-1.  Pathogenesis of eosinophilic gastrointestinal disorders. Allergens (aeroallergens or food allergens) stimulate a cascade of events that results in activation of Th2 cells, mast cells, and eotaxin and promotes eosinophil synthesis in the bone marrow and eosinophil trafficking to the site of insult through the processes of rolling, adhesion, and diapedesis. Interleukin-5 (IL-5), a Th2 cytokine, and eotaxin, in synchrony with eosinophil-selective transcription factors (GATA-1 and -2, c/ERB) and growth factors such as GM-CSF, stimulate eosinophil synthesis and mobilization from the bone marrow. IL-5 and eotaxin also influence eosinophil trafficking or recruitment to the sites of allergen exposure, and it is here at the target sites that eosinophils undergo activation and degranulation that can result in tissue injury. GI, gastrointestinal; GM-CSF, granulocyte-macrophage colony-stimulating factor; Th2, T helper 2; TNF-α, tumor necrosis factor-α. (Modified from Rothenberg ME: Eosinophilic gastrointestinal disorders [EGID]. J Allergy Clin Immunol 2004; 113:11-28.)

Th2 cells → IL-3, IL-4, IL-5, IL-13

Bone marrow

GM-CSF

GATA-1&2 c/ERB

Eotaxin Mast cells → TNF-α

Eosinophil activation and degranulation

Peripheral eosinophilia

Recruitment

Chapter 27  Eosinophilic Disorders of the Gastrointestinal Tract of the mice.14,16 In human EE, therefore, sensitization likely occurs via the respiratory tract, with subsequent exposure to oral allergens leading to a hypersensitivity response and esophageal eosinophil infiltration. Experimental studies have suggested the mechanism of this link between the lung and esophagus via T helper 2 (Th2) allergic responses in the lung and esophagus.14,16,17 Th2 cells (see Chapter 2 for more details) produce an array of cytokines, of which IL-5 is the most specific for eosinophils, inducing eosinophil growth, differentiation, activation, and survival, and enhancing responsiveness to chemoattractants such as eotaxin-1, eotaxin-2, and eotaxin-3 (eosinophil selective chemokines structurally distinguished from others on the basis of conserved cysteines). Further studies using the murine model of EE have demonstrated an important role for IL-5, IL-13, and eotaxin in this disorder.18-20 In IL-5–deficient mice, the allergen-induced EE response is ablated and, in the absence of eotaxin, it is attenuated. Furthermore, the absence of IL-5 reduces the esophageal eosinophilia induced by oral allergens after sensitization to aeroallergens, but the absence of IL-5 does not reduce intestinal eosinophilia, strongly suggesting a differential recruitment of eosinophils in EE and EG. In humans the esophageal infiltrate in EE also includes increased numbers of T cells and mast cells and increased IL-5, TNF-α, and eotaxin.21,22 Recent evidence that IL-13 delivery into the lung induces EE further implicates Th2 cells and cytokines in the immunopathogenesis of EE.17,23 Furthermore, esophageal biopsies in EE are notable for sharing some remodeling features with airway disease in asthma, such as increased profibrotic cytokines, signaling molecules, increased vascularity, and vascular activation.24,25 In a placebo-controlled experiment using another murine model, mice challenged with oral, encapsulated ovalbumin developed peripheral blood eosinophilia and antigenspecific immunoglobulin E (IgE) and IgG1 antibodies. Their eosinophil-predominant cellular infiltrate was largely localized in the lamina propria throughout the small intestine but was also present in the esophagus, stomach, and Peyer’s patches. The mice developed gastromegaly, dysmotility, and cachexia, thought to be correlates of human EGID.26 In a mouse model of the homozygous lyp gene mutation (lyp protects against lymphocyte apoptosis), increased levels of Th2 cytokines and IgE are observed in association with clinical features of bloating, intestinal distention, wasting, splenomegaly, and increased intestinal eosinophilia.27 Some patients with EE have also been characterized as having a unique genomic transcript comprised of an increased expression of the gene encoding eosinophilspecific chemoattractant eotaxin-3 compared with healthy patients.18,28

CLINICAL ENTITIES EOSINOPHILIC ESOPHAGITIS

In the esophagus, attention to eosinophilic infiltration has focused on the epithelium, rather than on the lamina propria. This squamous epithelium normally is devoid of eosinophils, but various disorders cause eosinophils to infiltrate the esophageal epithelium. In general, such esophageal eosinophilic infiltration is considered to be secondary to an extraesophageal cause (e.g., parasitic infections, autoimmune diseases, vasculitis, HES, medications) or an esophageal cause (e.g., gastroesophageal reflux disease [GERD])29

or to be primary EE. Primary EE may be divided into allergic or idiopathic cases, depending on whether identifiable allergens play a role (see Table 27-1). Occasional patients presenting with apparent EE have marked eosinophilic inflammation of other segments of the GI tract, and designation as EE secondary to EG or as a form of primary EE is a matter of semantics. Primary EE, rarely diagnosed until the mid-1990s, currently represents an important esophageal disorder, particularly in children, but increasingly in adults. The emergence of this disease has paralleled the increasing incidence of allergies and asthma. Whether the increasing number of cases of EE is to the result of increased recognition of EE or a truly increased incidence of EE is still debatable.30-32 Although some of the previous lack of recognition in adults may have resulted from failure to biopsy intact-appearing esophageal mucosa, the actual prevalence of EE in adults has increased, as evident in some recent reviews.33-39 Similarly, although the routine biopsying of even normalappearing mucosa by pediatric gastroenterologists may account for some of the predominance of EE in school-age children, it is likely that this age group currently does experience more EE, with another peak in young adulthood.40,41 Like allergic disorders, EE’s prevalence also varies markedly in different locales and perhaps in different seasons.42 The last 10 years or so have seen a dramatic increase in the diagnosis of EE around the world; this may be the result of both improved recognition and an actual increase in new cases akin to other allergic disorders. The incidence of pediatric EE in Australia increased from 0.05 to 0.89/10,000 during 1995 to 2004, according to a retrospective review.43 In the U.S. Northeast, a twofold increase in incidence was observed in a four-year period at Children’s Hospital of Philadelphia.44 In the U.S. Midwest, a fourfold increase in prevalence was reported.45 The prevalence of EE was reported as 15/100,000 inhabitants in one region of Switzerland, perhaps an underestimation because of limited expertise in its diagnosis.46 Markedly more prevalent in males, the disorder occurs in females as well. Duration of symptoms before diagnosis may vary from just a few days in those presenting with sudden episodes of food impaction to many years in those with GERD-like symptoms.4,47 Like allergies in general, EE clusters in families and an autosomal dominant pattern of inheritance have been proposed on the basis of the 10% rate of familial clustering.48 A personal history of atopy in association with EE in children and adults occurs in 50% to 80% cases, with food allergy accounting for an allergic diathesis in up to 90% children; 39% report a family history of allergies.49-52 In one series, symptoms of chronic respiratory disease were seen in 62% of patients.40 In support of the theory that aeroallergens promote the disease are cases of EE with symptomatic and biopsy-proven exacerbations during pollen season and resolution during winter months.53,54 The first report of EE in 1978 and subsequent reports have characterized the phenotype of EE.35,40,44,55,56 Symptoms of EE are similar to those of GERD, but respond poorly to antireflux medical and surgical therapy.57 Whereas younger children with EE frequently present with GERD-like symptoms, feeding problems, and abdominal pain,58,59 adolescents and adults present with obstructive presentations such as dysphagia or food impactions, with or without strictures.36,47 The degree to which these presentations represent actual structural obstruction versus dysmotility is unclear and appears to vary among patients.60 Esophageal biopsy of patients without any GI symptoms, such as some patients presenting for evaluation of respiratory symptoms, has disclosed unsuspected EE.37,61

427

428

Section IV  Topics Involving Multiple Organs Because the symptoms of EE resemble those associated with GERD and because esophageal epithelial eosinophils were previously recognized as accompanying GERD, differentiating EE from GERD is the primary diagnostic challenge, particularly because of the high prevalence of the latter. Laboratory studies, including elevations of peripheral eosinophil counts, IgE levels, and positive radioallergo­ sorbent tests (RASTs) or skin prick and patch tests to food antigens, suggest an allergic predisposition in a considerable number of the patients, but these tests are neither very sensitive nor specific. An upper endoscopy and esophageal biopsies substantiate the diagnosis of EE. The brief history of recognition of EE has prevented clear definition of its natural history, but EE appears to be a chronic disease with a waxing and waning course, as suggested by a noteworthy relapse rate of 80% in an eight-year follow-up of children with EE and similarly high rate of recurrent symptoms and chronic therapy in adults.62,63 It is thought that the esophageal wall is fragile and weakened in patients with chronic EE, thus predisposing to endoscopyrelated and spontaneous perforation.36,64,65 Another study of the long-term (mean, 7.2 years) follow-up of EE in adults has reported persistence of eosinophilic inflammatory infiltrate (albeit significantly reduced compared with baseline) and dysphagia and a high rate of subepithelial fibrosis and sclerosis, perhaps the mechanistic link to esophageal strictures in EE.41 Although not supported by any published data, EE patients are also considered to have a low quality of life. This information raises questions regarding the pros and cons of treatment for asymptomatic patients with EE. The current literature does not clearly identify any malignant potential of the disease.66,67

EOSINOPHILIC GASTROENTERITIS

Eosinophilic gastroenteritis is a heterogeneous disorder affecting children and adults characterized by the presence of an intense eosinophilic infiltrate on histopathology of one or multiple segments, from the esophagus to the rectum.68 These eosinophilic infiltrates not only may involve various sites down the length of the GI tract, but also may occupy various sites through the depth of the wall. These inconsistencies from case to case promote unpredictability in presenting symptoms, which range from pain to dysmotility, bleeding, obstruction, or ascites.69-72 Since the initial report of EG seven decades ago, reports of EG have emerged from different parts of the world, including North America, Europe, Australia, and Asia.72-77 These reports provide important information regarding epidemiology, disease characteristics, and management. The diagnosis of EG is rare, with an approximate incidence of 1/100,000, but it is also possible that physicians make the diagnosis of EG infrequently because of the inaccessibility of much of the length of the small bowel and of the deeper layers of the luminal wall. Therefore, the literature on EG has been somewhat anecdotal. Retrospective review of an 18-year period at a hospital in China identified 15 patients with EG, including 2 children.74 Histologic evaluation established the diagnosis in 13, and radiologic findings, combined with eosinophilic ascites, suggested it in the remaining 2 patients. In one of the largest series, Talley and colleagues compared laboratory and clinical data on 40 adults diagnosed with EG during a 30-year period with data on 10 other patients with similar GI symptoms but no tissue eosinophilia.73 EG is most commonly diagnosed between the second and sixth decades of life, and is rarely diagnosed in infants.1,74,78 Unlike EE, which favors males, EG does not appear to manifest a significant gender disparity. It is associ-

ated with asthma and allergies in 40% to 50% of cases.79,80 Peripheral eosinophilia may be seen in up to 80% of cases, but is not a prerequisite for diagnosis.81 In published reports, the stomach (26% to 81%) and small intestine (28% to 100%) are the predominantly affected areas, but the esophagus, large intestine, and rectum may be affected as well.1,80 The depth of infiltration varies and leads to the broad spectrum of clinical manifestations in patients with EG. The classification of EG on the basis of depth of eosinophilic infiltration proposed by Klein and associates is currently the one most cited in publications.82

Mucosal Eosinophilic Gastroenteritis

Those with mucosal inflammation usually present with common, albeit nonspecific, complaints of abdominal pain, nausea, vomiting, diarrhea, fecal occult blood loss, anemia, or protein-losing enteropathy. Because of the nonspecific nature, these clinical presentations may be confused with irritable bowel syndrome, dyspepsia, peptic ulcer, pancreatitis, acute appendicitis, or IBD.83 Eosinophilic enteritis presenting in an adult as intussusception, and treated effectively with the nonsurgical option of prednisone, has been observed.76 Frequently, atopy and high IgE levels coexist.73,74 An example of eosinophilic gastritis is shown later (see Fig. 51-8B).

Muscular Eosinophilic Gastroenteritis

Signs and symptoms of gastric outlet and intestinal obstruction are common in those with muscular EG.84,85 The presentation of gastric outlet obstruction mimicking hypertrophic pyloric stenosis has been reported in infancy and adulthood.86,87 A hypoallergenic diet has been shown to alleviate the condition in infants (Fig. 27-2). Enteric strictures are rare, but can occur in children and adults with EG.88 As a sort of amalgam between the mucosal and muscular forms, one patient was reported to have eosino-

Figure 27-2.  Radiograph from an upper gastrointestinal series demonstrating a string sign (arrow) suggestive of gastric outlet obstruction in a child with eosinophilic gastroenteritis. (From Khan S, Orenstein SR: Eosinophilic gastroenteritis masquerading as pyloric stenosis. Clin Pediatr [Phila] 2000; 39:55-7.)

Chapter 27  Eosinophilic Disorders of the Gastrointestinal Tract philic inflammation of myenteric plexus and the lamina propria on colonic biopsies, producing functional intestinal obstruction.89

Serosal Eosinophilic Gastroenteritis

Involvement of the serosal layer occurs in 10% of cases of EG and typically presents as ascites. The serosal form of EG, compared with other types, is reported to be associated with significant bloating, a higher level of peripheral eosinophilia, and a better response to glucocorticoid therapy.73,90,91 The natural history of EG remains somewhat vague, although recent pediatric data has supported a protracted course, and hence the need for long-term treatment strategies, including dietary restrictions and repeated use of glucocorticoids.78

over 20 years had blood-tinged stools, and one third of them were observed to have painful defecation and eczema.99 The diagnosis may be confused with colic or GERD in infants who also present with irritability or vomiting. In a prospective study, 18% of 40 breast-fed infants presenting with rectal bleeding were diagnosed to have cow’s milk allergy on the basis of milk elimination and provocation; association with food-specific IgE and positive SPTs was uncommonly noted.100 Endoscopic examination reveals focal rectal mucosal erythema, erosions, and lymphoid nodular hyperplasia (Fig. 27-3A). Histopathologic examination shows prominent eosinophil infiltration in the mucosa and lamina propria, at least 6 eosinophils/high-power field, and/or eosinophils invading crypts or the muscularis mucosae (see Fig.

FOOD PROTEIN–INDUCED ENTEROCOLITIS

Enterocolitis signifies an inflammatory process involving the small and large intestines. FPIEC represents a symptom complex of severe vomiting and diarrhea that usually pre­ sents in infancy as a reaction to ingested proteins. The onset of symptoms is in the first few weeks of life. The trigger is most often ingestion of cow’s milk protein-based formula, but approximately half of infants also react to soy. Other food proteins, including rice, oats, and chicken, have also been implicated in individual cases.92-94 The responsible dietary antigens in the maternal diet are thought to sensitize via breast milk. The profuse, often bloody and mucoid, diarrhea is associated with weight loss and malnutrition in an ill-appearing patient. An association with methemoglobinemia noted in several cases was attributed to increased heme oxidation caused by an elevation of nitrite levels in the intestine in severe intestinal inflammation.95,96 Cow’s milk protein sensitivity is a general term used to refer to the various clinical states caused by ingestion of the offending milk protein, but without distinguishing between the specific clinicopathologic entities, including FPIEC and EP. Typically, FPIEC is caused by non-IgE mediated delayed food protein hypersensitivity, so allergy testing with skin prick tests (SPTs) and RASTs are negative; patch tests have not been adequately studied in this diagnosis.97,98 Patients lack evidence for other causes of eosinophilia, such as infections, inflammatory bowel disease, and ischemia. The diagnosis rests on clinical criteria and resolution after elimination of the causal milk and soy proteins from the diet. Most infants do well when their milk is changed to an extremely hydrolyzed formula that digests the intact proteins into small polypeptides of sizes that do not engender the hypersensitivity response. Up to 90% of such infants can tolerate milk by 3 years of age. Any milk challenge should be performed under medical supervision because of the risk of serious reactions leading to shock. Criteria for a failed challenge include vomiting, diarrhea, gross or occult fecal blood, fecal leukocytes, fecal eosinophils, and elevated white blood cell count.

EOSINOPHILIC PROCTITIS

EP uniquely affects children younger than 2 years. These children present with bloody stools, either alone or in association with diarrhea. The condition has been reported in infants receiving cow’s milk and soy protein–based formulas, as well as in exclusively breast-fed infants. A few infants suffer from eczema, but otherwise these children lack any systemic symptoms. In contrast to children with enterocolitis, these babies generally appear well. In a review, all 95 exclusively breast-fed infants evaluated for proctitis

A

B Figure 27-3.  A, Eosinophilic proctitis. This is an endoscopic image of the rectum in an infant presenting with guaiac positive stools and anemia. Mucosal nodularity with central umbilication characteristic of nodular lymphoid hyperplasia is evident, findings often associated with food allergies. B, Photomicrograph of a rectal mucosal biopsy shows increased numbers of eosinophils in the lamina propria that are forming aggregates and, occasionally, encroaching on the epithelium and crypts (Hematoxylin and eosin, ×40). (Courtesy of Dr. Robert Garola, Department of Pathology, Wilmington, Del.)

429

430

Section IV  Topics Involving Multiple Organs 27-3B).101 There is an excellent response to elimination diets using hydrolyzed or elemental formulas; by 1 year of age most infants can tolerate a rechallenge with the offending food proteins.

EVALUATION Evaluation of patients suspected of EGID is undertaken to exclude differential diagnostic possibilities, establish the definitive diagnosis, and assess potential complications associated with the diagnosis. The comprehensive evaluation of a patient with EGID, initiated by the demonstration of tissue eosinophilia, is often unconventional and may not conform to the sequence discussed here. Depending on the condition, different diagnostic algorithms pertain. In EE, for example, the diagnosis may be first suspected on examination of biopsies performed during evaluation for suspected GERD, and more basic laboratory investigations into EGID follow. Alternatively, in serosal EG, for example, analysis of peritoneal fluid provides the clearest clue, and histopathologic examination may not be required. Finally, some cases of typical FPIEC or EP in infants are treated for one or two weeks empirically, using the response to dietary antigen elimination as the diagnostic test and undertaking little in the way of formal investigation. The following aspects of evaluation usually complement each other.

HISTORY AND PHYSICAL EXAMINATION

The personal and family history should include information pertaining to adverse effects of foods and environmental antigens, manifesting as GI, respiratory (e.g., asthma), or skin (e.g., eczema) reactions. Stigmata of atopic disease may be present on physical examination. Although most patients with EE and EP appear well and have a normal nutritional status, those with active EG and with FPIEC may present with malnutrition or failure to thrive and appear ill.

LABORATORY EVALUATION

Peripheral eosinophilia in the context of GI symptoms is a useful clue to EGID, but the absence of eosinophilia does not exclude these diagnoses.102 It is important to note that circulating eosinophils represent a balance between bone marrow production and tissue infiltration. Moreover, the frequently observed fluctuations in peripheral eosinophil concentrations may be caused by the effects of the circadian rhythm.103 To exclude important secondary causes for GI eosinophilia, evaluation should generally include stool or duodenal aspirate for ova and parasites. In those with ascites, paracentesis may provide the only clue to the diagnosis in the form of ascitic fluid eosinophilia. Hopefully, our increasing knowledge will soon allow simple, reliable, and relatively noninvasive testing involving markers of active eosinophil inflammation (e.g., fecal ECP) for monitoring disease course and response to treatment.104

ALLERGY EVALUATION

Immunologic evidence of underlying allergy is usually lacking in most EGIDs, except those mediated by IgE antibody and typically presenting with immediate reactions or accompanied by eczema or asthma (see Chapter 9). In addition to the lack of sensitivity of allergy testing in diagnosing EGIDs, there is a lack of specificity, with a high rate of false-positive results. The most commonly available tests, SPTs and RASTs, are used to detect IgE antibody specific to inhaled and ingested allergens. SPTs are sensitive, so negative SPTs are useful in confirming the absence

of IgE-mediated reactions, if good-quality food extracts are used. In those older than 1 year, SPTs are associated with a high negative predictive value but a positive predictive value of 50% or lower. Atopy patch testing, used for non-IgE cell–mediated immunologic reactions, is gaining popularity in Europe, but has not found a routine place in most U.S. centers.105 The use of patch tests in combination with SPT in EE has been shown to identify food allergies with greater accuracy, thus leading to greater success in dietary therapy than SPT alone.52,106,107 RAST or enzyme-linked immunosorbent assays detect circulating IgE antibody against specific food antigens. The quantitative CAP fluorescent enzyme immunoassay (FEIA) has been found accurate for predicting symptomatic food hyper­sensitivity when compared with the gold standard of double-blind, placebo-controlled food challenges.108 Non-IgE–mediated allergies, often implicated in the pathogenesis of EGID, pose a particularly difficult challenge; the diagnosis rests on the results of elimination diets, selected oral food challenges, and biopsies. Important to note is that double-blind, placebo-controlled food challenges in the research setting and open challenges in clinics have limited usefulness for EGIDs, because delayed hypersensitivity reactions may not be apparent for a few days. An alternative approach, therefore, uses diagnostic trials of therapy with dietary restrictions such as elimination diets, the six-food elimination diet, or oligoantigenic or elemental (amino acid–based) diets.44,109

RADIOLOGIC EVALUATION

Barium esophagography is an important study in the evaluation of dysphagia and may demonstrate EE strictures, which, in contrast to the distal location of GERD strictures, are usually located in the proximal and midesophagus. High-resolution endoscopic ultrasound demonstrates increased esophageal wall thickness with expansion of mucosa, submucosa, and muscularis propria in children with EE, compared with healthy controls.110 It provides insight into the full-thickness inflammation that can lead, in turn, to esophageal dysmotility and obstructive presentations. Radiographic changes in patients with EG of the stomach may include an irregular and lacy antral surface on an upper GI series. Furthermore, a string sign may be demonstrated in gastric outlet obstruction because of antral EG (see Fig. 27-2).87 Eosinophilic infiltration of the small bowel manifests as thickening of the circular folds and wall.111 Computed tomography (CT) may show nodular, irregular, and thickened intestinal folds in the affected segments in EG.77 Deep infiltration may result in rigid bowel loops, simulating lymphoma.112 Abdominal ultrasonography is useful for detecting ascites.

ENDOSCOPY AND PATHOLOGY

Although the macroscopic appearance of the esophagus in EE may be normal, most cases are associated with one or more characteristic findings. These include furrowing, vertical lines, rings, granularity, crepe paper appearance, and whitish exudates (Fig. 27-4A).40,113-115 Esophageal biopsies sampled from areas of whitish exudates or specks contain a significantly higher eosinophilic density compared with areas without these findings.116 A small-caliber noncompliant esophagus, also termed a defiant esophagus, furrowing, and mucosal shearing on dilation are noted in some young adults with EE.33,117 Multiple esophageal biopsies should be procured along the length of the esophagus to ensure satisfactory sampling. An eosinophil density higher than 15 eosinophils/high-power field (eos/hpf) strongly suggests the diagnosis of EE, but must not be relied on as the sole diagnostic criterion. The diagnosis of EE is supported by the

Chapter 27  Eosinophilic Disorders of the Gastrointestinal Tract

A

B Figure 27-4.  A, Upper endoscopic view of the characteristic furrowing and white specks on gross appearance in a patient with eosinophilic esophagitis. B, Photomicrograph of an esophageal mucosal biopsy specimen in eosinophilic esophagitis illustrates numerous intraepithelial eosinophils (≈100 eos/hpf), degranulation of eosinophils, and juxtaluminal clustering of eosinophils. Increased numbers of lymphocytes are also seen. Basal layer proliferation and increased eosinophils in the lamina propria are other features (Hematoxylin and eosin, ×10).

proper clinical context, in which a patient presents with signs of esophageal dysfunction and a poor response to a trial of proton pump inhibitor therapy.118 In contrast, the esophageal eosinophil density in patients with GERD is generally less than 7 eos/hpf.5,44 A mean eosinophil density more than or equal to 7 eos/ hpf provides a sensitivity of 61%, specificity of 96%, and a predictive value for failure to respond to antireflux therapy of 86%. A mean eosinophil density less than 7 eos/hpf provides an 85% predictive value for successful antireflux therapy.119 Differentiation of EE from GERD based on eosinophil density may be even clearer in the proximal esophagus than in the distal esophagus and might be compelling in light of the proposed triggers (dietary from above in EE and acid from below in GERD), but this is not a universal finding. Other typical histologic features

include a preferential juxtaluminal location of eosinophils, degranulating eosinophils, eosinophil abscesses, elongated papillae, and prominent basal layer hyperplasia (see Fig. 27-4B).5,118 Eosinophilic gastroenteritis may present with a spectrum of gastric macroscopic abnormalities, which may include erythema, focal erosions, ulcerations, and pseudopolyps.120 The gross endoscopic abnormalities in EG are most striking in the mucosal form, and include thickening of folds, erythema, and friability.75,121 The precise histologic criteria required to diagnose EG are ambiguous (see Fig. 51-8B). Normally, mucosal eosinophils may be found in low numbers in the stomach and reach higher densities (up to 30 eos/hpf) in the appendix, terminal ileum, cecum, and proximal colon.122 Furthermore, the diagnosis may be elusive, either because of patchy disease distribution or the mucosa being spared altogether, as in muscular and serosal types of EG. Degranulated eosinophils are noted in the intestinal mucosa accompanying histologic damage in EG. Although invasive, laparoscopy or open surgical exploration is most helpful for establishing the diagnosis of muscular and serosal disease. The gross findings described in serosal EG are ascites, whitish nodules, and thickening of the parietal and visceral peritoneum.91 Histologic descriptions from some case series of FPIEC feature prominent eosinophilia in colonic or small bowel biopsies and also include nonspecific findings of crypt abscesses and a diffuse inflammatory cell infiltrate in colonic biopsies, as well as variable villous injury, acute inflammation, and prominent eosinophilia in small bowel biopsies.95 Endoscopic findings in EP demonstrate disease in the rectum and sigmoid. There may be focal areas of erythema, friability, ulcerations, and lymphoid nodular hyperplasia. Histopathology consistently reveals intense eosinophilic infiltration of the mucosa, with the eosinophil concentration varying between 6 and 20 eos/hpf, and it often features degranulated eosinophils (see Fig. 27-3A and B). Characteristics of chronicity and granulomas are absent, facilitating the exclusion of IBD.100,101

DIFFERENTIAL DIAGNOSIS INFECTIONS Parasitic Infestations

Invasive helminthic infections frequently result in tissue and peripheral eosinophilia, which reflects an immunologic response to tissue migration; however, when migration ceases, the eosinophilia often resolves (see Chapter 110). Tissue eosinophilia is associated with hookworms (Ancy­ lostoma caninum), pinworms (Enterobius vermicularis), Eustoma rotundatum, Giardia lamblia, Anisakis, Trichi­ nella spiralis, Ascaris, Trichuris, and Schistosoma.123-126 Eosinophilic ascites in the absence of gastroenteritis has occurred with Toxocara canis and Strongyloides ster­ coralis.127-129 Fasciola hepatica can cause eosinophilia, right upper quadrant pain, fever, and hepatomegaly.130,131 Although peripheral eosinophilia is usually absent in Giardia infestations, the diagnosis is aided by stool studies (ova and cysts, Giardia antigen) and careful histologic examination of a duodenal aspirate, which has the highest diagnostic yield (90%). The larvae of Anisakis may be identified at endoscopy in the stomach in an area of mucosal edema and, in some cases, ulceration (see Chapter 51).132,133 This parasitic infestation may be underdiagnosed; a Spanish

431

432

Section IV  Topics Involving Multiple Organs study has suggested that up to 80% of patients thought to have idiopathic EG have evidence of exposure to Anisakis, contrasted with 10% of control subjects.134

Helicobacter pylori Infection

The diagnosis of EG and concurrent Helicobacter pylori on gastric biopsies is a rare occurrence. The role of the bacterial infection in producing tissue eosinophilia is not known and may be a coincidental finding in those evaluated for dyspepsia.135-137

MEDICATIONS

A drug allergy may result in eosinophilic involvement of the gut.138,139 For example, a hypersensitivity reaction to carbamazepine, producing EE and resolution with drug withdrawal, has been reported. Other medications reported to induce intestinal eosinophilia include tacrolimus,9 gemfibrozil,140 enalapril,141 and interferon-α.142

CONNECTIVE TISSUE DISEASE and VASCULITIS

Several connective tissue and vasculitic disorders are reported to be associated with GI eosinophilia (see Chapter 35).6 Gastrointestinal symptoms in patients with systemic lupus erythematous have been described as being caused by eosinophilic enteritis. The signs and symptoms of the underlying disease (e.g., lupus) allow for the classification of such secondary EGIDs despite the pathology, which is similar to that seen in the primary EGIDs reviewed earlier.143 In another example of a vasculitic disease, Churg-Strauss syndrome, an eosinophilic infiltrate involves the small arteries and veins; granulomas can be found in the lungs, heart, kidneys, and subcutaneous tissues and may also occur in the stomach, small bowel, and colon.144 The diagnosis of vasculitis can be substantiated by biopsy of involved organs (skin, muscle).

INTESTINAL POLYPS Juvenile Polyps

Juvenile polyps, also known as retention or hyperplastic polyps, are typically benign colonic tumors diagnosed most commonly in children aged 2 to 10 years who present with painless hematochezia (see Chapter 122). The polyps are usually pedunculated, solitary, and located in the rectosigmoid colon. The surface of the polyp appears lobulated because of multiple mucin retention cysts. Adenomatous transformation of a juvenile polyp is rare. Occasionally, a heavy eosinophil infiltrate is present in the stroma of a juvenile polyp, as well as in mucosal biopsies of grossly normal colon in the same patient. Our understanding of the role of eosinophils in juvenile polyps is poor.

Inflammatory Fibroid Polyps

These benign localized lesions should not be confused with EG. They typically are found in the stomach, followed by the small and large intestine, but they rarely may be in the esophagus.145 They originate in the submucosa and typically appear as polyps or nodules; peripheral eosinophilia is absent, and a history of allergy is unusual.146 Many other terms have been used to refer to these lesions, including fibroma, inflammatory pseudotumor, submucosal granu­ loma, and localized EG.6 Histologically, the stroma in these lesions is characterized by a concentric arrangement of proliferating spindle cells, which may be fibroblasts or endothelial cells, although their exact nature remains controversial, surrounding arborizing capillaries, with a variable eosinophil infiltration. These lesions are relatively rare, with a slight male preponderance; although they may

appear at any age, they are most common in the sixth and seventh decades. Most patients present with obstructive symptoms that depend on the site of the lesion; gastric outlet obstruction and small bowel intussusception are common manifestations. Surgical excision is curative in symptomatic patients, and recurrence has not been reported. Therapy with glucocorticoids is not indicated. One retrospective study of the spectrum of colonic neoplasms revealed the most prominent stromal eosinophilia in adenomas; only 5% of hyperplastic polyps had any eosinophil infiltration, and invasive adenocarcinomas had a striking absence of any eosinophilia.147

HYPEREOSINOPHILIA SYNDROME

Occasionally, hypereosinophilia syndrome (HES), a multisystem disorder, can involve the gut and be confused with EG. The diagnostic criteria established in 1975 and still in use today are the following: blood eosinophilia exceeding 1500 cells/µL for more than six consecutive months, absence of an underlying cause of hypereosinophilia despite extensive evaluation, and presence of organ damage or dysfunction related to hypereosinophilia.148 By definition, eosinophilic infiltration of multiple organs outside the abdomen excludes the diagnosis of primary EGID. The heart, skin, and central nervous system are the major targets, with more than 50% of patients presenting with complications in one or more of these sites. Recent studies have indicated that the condition can be classified as myeloproliferative or lymphocytic, providing evidence for the existence of discrete hematologic disorders underlying these variants.148,149 Anemia and thrombocytopenia are often present. Congestive heart failure with endocardial fibrosis (and valvular incompetence), venous and arterial thromboembolism, neuropsychiatric disturbances, mononeuritis multiplex, and fever are common clinical features. The prognosis is poor in patients with prominent organ involvement, with a 25% three-year mortality rate without treatment. The prognosis tends to be better in patients who have angioedema. Imatinib mesylate (Gleevec) is the treatment of choice for those with the myeloproliferative variant and the FIP1L1-PDGFRAα (F/P) fusion gene. Recognition of imatinib-resistant mutations has led to newer treatment options, including the development of tyrosine kinase inhibitors and anti–IL-5 monoclonal antibodies.150-152 Other treatment strategies include glucocorticoids (about one third respond), hydroxyurea, cyclosporine A, and interferon-α. In the presence of malignant transformation, chemotherapy, bone marrow transplantation, or stem cell transplantation may be considered.

INFLAMMATORY BOWEL DISEASE

In contrast to the primary eosinophilic diseases, IBD is a secondary eosinophilic disorder and is treated in greater detail in Chapters 111 and 112. Aspects specifically related to eosinophil infiltration, however, are useful to consider here. Eosinophils may be elevated in the peripheral blood and inflamed tissue in patients with Crohn’s disease and ulcerative colitis. Eosinophils are not pathognomonic of IBD, but represent a major component of the inflammatory infiltrate in active IBD.153 The increased concentrations of eosinophil granular proteins in intestinal fluid and mucosa in IBD suggest a pathogenic role for activated eosinophils. It is suggested that ulcerative colitis is a predominantly Th2-associated disease accompanied by overproduction of IL-5154 and increased serum eotaxin levels.155 Patients with ulcerative colitis seem to experience a higher prevalence of allergies than controls

Chapter 27  Eosinophilic Disorders of the Gastrointestinal Tract (52% vs. 18%). However, whether infants with allergic intestinal diseases are at risk for developing IBD in later life is controversial.156 Crohn’s disease is characterized by transmural inflam­ mation believed to be mediated by Th1-type cytokines. Immunohistochemical analysis shows expression of TNF-α in several cell lines, including eosinophils. There is an increased serum eotaxin level in patients with Crohn’s disease,157 and fecal excretion of ECP is elevated during disease flare-ups.158

Table 27-2  Recommended Therapeutic Options for Eosinophilic Gastrointestinal Disorders DISORDER Eosinophilic esophagitis

CELIAC DISEASE

Celiac disease, regarded as an immune-mediated hypersensitivity to gluten-containing grains, is characterized by variable small intestinal villous atrophy with crypt hyperplasia, intraepithelial lymphocytes, and increased cellularity of the lamina propria (see Chapter 104). Activated eosinophils are one type of infiltrating inflammatory cells in the lamina propria in celiac disease, suggesting a role for eosinophils in this disorder.159 A concurrent diagnosis of EE and celiac disease has been recognized in recent case reports, but do not provide sufficient understanding to distinguish between a pathogenic association and mere coexistence.160,161 Glucocorticoids, in addition to a gluten-free diet, were required for treatment of the EGID, perhaps pointing to the coexistence of two discrete disorders.

Eosinophilic gastroenteritis

TRANSPLANTATION

Eosinophilic GI inflammation after solid organ transplantation is being increasingly reported (see Chapter 34).8,9,162 The precise roles of immunosuppression, therapy for rejection, and viral infections are yet to be determined. Proposed mechanisms include an imbalance of Th1 and Th2 lymphocytes as a result of immunosuppressive therapy promoting tissue eosinophilia and de novo food allergies after transplantation. In a recent study in 54 pediatric recipients of 57 liver transplants, 28% of patients developed peripheral eosinophilia.162 Of 23 patients who had an endoscopic evaluation, 6 also developed EG. Those with eosinophilia were significantly younger, had more rejection episodes, were more commonly managed with tacrolimus-based immunosuppression, and experienced more frequent episodes of detectable Ebstein-Barr viremia. Patients with EG were more frequently retransplanted.

TREATMENT Thus far, the management of EE and EG has been largely guided by several case reports, case series, and expert opinion, which provide support for various treatments, including special diets, glucocorticoids, and investigational agents such as anti-IL5 antibody (Table 27-2). The use of mast cell inhibitors, antihistamines, and leukotriene antagonists has dwindled, simply because of the lack of supporting evidence. When this chapter was written, most of the medical treatments under discussion were not approved by the U.S. Food and Drug Administration, but their status could change as a result of ongoing research trials, and new agents are being proposed. Obstructive manifestations of EGIDs require the adjunctive therapeutic use of periodic esophageal dilations for esophageal strictures and of surgery for intestinal obstruction.

DIET

The strong association of EGID with food allergies (see Chapter 9) prompted the use of restrictive or elemental

Food protein–induced enterocolitis

Eosinophilic proctitis

RECOMMENDED THERAPEUTIC OPTIONS Diet Elemental diet (free amino acid formula) Oligoantigenic diet Six-food elimination diet Elimination diet Glucocorticoids Systemic (e.g., prednisone) Topical (e.g., fluticasone, oral viscous budesonide) Leukotriene receptor antagonist (e.g., montelukast) Anti–interleukin-5 antibody Diet Elemental diet (free amino acid formula) Oligoantigenic diet Six-food elimination diet Elimination diet Glucocorticoids Systemic (e.g., prednisone, methylprednisolone) Topical (e.g., budesonide) Mast cell inhibitors Disodium cromoglycate Ketotifen Leukotriene receptor antagonist (e.g., montelukast) Anti–interleukin-5 antibody Anti-IgE antibody Diet Extensively hydrolyzed casein formula Free amino acid formula Consideration of delayed weaning to grains and high-risk solid foods Diet Maternal dietary elimination of milk and soy (if infant is breast-fed) Extensively hydrolyzed casein formula Free amino acid formula

diets.163 The degree of allergen restriction ranges from the provision of protein nitrogen exclusively as amino acids, through the use of protein hydrolysate formulas, consisting of free amino acids and peptides of varying chain lengths, manufactured via enzymatic hydrolysis of casein or whey proteins, to the simple elimination of one or several wholefood proteins via careful reading of labels. Other modifications include oligoantigenic diets, allowing choice of those foods considered unlikely to be food allergens (e.g., broccoli, apple, corn, sweet potato, olive oil, salt, sugar, lamb) and the six-food elimination diet (eliminating only wheat, milk, soy, egg, peanut, tree nuts, fish, and shellfish) to exclude the most common food allergens. Because of the unreliability of allergy testing or selection based on allergy prevalence, however, the most reliable test of an allergic basis for the EGID is the relief of symptoms within one or two weeks of beginning consumption of an amino acid–based diet. In practice, therefore, dietary therapy may be limited by the challenge of identifying a few responsible food allergens, the restricted nature of the diet, and poor palatability and expense (particularly with the use of amino acid–based diets). The type of EGID, age and willing-

433

434

Section IV  Topics Involving Multiple Organs ness of the patient, and results of any allergen testing will determine the type of dietary management in an individual case of EGID. Both retrospective and prospective open-label, uncontrolled studies have indicated that elemental diets can lead to clinical and histologic improvement in children and adolescents with EE.44,51,56 The six-food elimination diet and an elemental diet positively affect clinical symptoms, but esophageal histology appears to be more responsive to an elemental diet.50,109 In EG, case reports suggest efficacy of exclusively elemental diets,78,87 but they have not yet been examined in any prospective fashion. Most infants with EP respond to extensively hydrolyzed formulas or, in the case of breast-fed infants, maternal dietary restriction of both milk and soy. The rationale for eliminating milk and soy is the coexistence of milk and soy protein intolerance in 30% to 50% of cases. A considerable proportion of infants who are unresponsive to these basic dietary changes will show complete resolution of their symptoms when an elemental diet is instituted. Infants with FPIEC and patients with more severe cases of EG are more apt to benefit from initiation of treatment with amino acid– based diets, or even with an initial period of stabilization using intravenous fluids.93,95

THERAPEUTIC AGENTS Glucocorticoids

Glucocorticoids, by virtue of their potent anti-inflammatory actions, are an effective and powerful treatment for EGIDs. Mechanisms of action include inhibition of eosinophil growth factors, IL-3, IL-5, and GM-CSF. Several uncontrolled studies have provided evidence for beneficial antiinflammatory effects during short-term and long-term treatment.40,44,73 Long-term use of glucocorticoids is undesirable, however, because of serious side effects, which include fluid and electrolyte disturbances, glucose intolerance, cushingoid state, growth suppression, bone demineralization, pituitary and adrenocortical hyporesponsiveness, posterior subcapsular cataracts, and various infections. Glucocorticoids are useful as first-line therapy to induce remission, similar to their use in IBD at a dose equivalent to prednisone, 1 to 2 mg/kg/day. This is particularly the case in patients with nonallergic EGID or those refractory to dietary therapy. In a study examining the effects of oral glucocorticoids in children with EE, all 20 children showed clinical and histologic improvement at four weeks and half of them remained well at one-year follow-up.164 An important development in the treatment of EE with regard to efficacy and tolerance is the successful use of topical fluticasone, a relatively safe alternative to systemic glucocorticoids.165,166 A randomized, double-blind, placebo controlled study has shown swallowed fluticasone to be more effective in inducing histologic remission (eos/hpf ≤ 1) in EE; its effects were more impressive in younger patients, in nonallergic EE, and in the proximal esophagus. Suggested doses for fluticasone are 440 to 880 µg/day in children and 880 to 1760 µg/day in adolescents and adults, administered in two to four divided doses. Side effects associated with fluticasone include esophageal candidiasis, herpes esophagitis, and epistaxis.167 A head-to-head comparison of prednisone and fluticasone in a randomized fashion resulted in similar clinical and histologic response rates, although with far greater side effects with the former. Both treatment groups suffered relapse of symptoms by six months of follow-up.168 Topical viscous budesonide is also an effective treatment option for EE, particularly in children with swallowing

disorders or in those who find the metered-dose inhaler difficult to use.169 Recommended doses are 1 mg daily for children younger than 10 years and 2 mg/day for those 10 years and older. Another form of topical glucocorticoid delivery with potential application in EGID is non– enteric-coated budesonide administered at a dosage of 9 mg daily in patients with EG affecting the ileum and right colon.170,171

Mast Cell Inhibitors

Oral disodium cromoglycate and ketotifen have shown limited success as treatment options for patients with EGID.172 There are only sporadic case reports of the use of these agents, and most reports used them in combination with other treatment modalities.173,174 Sodium cromoglycate may be an effective agent in EG at a dosage of 200 mg orally four times daily. Ketotifen (also an antihistamine), administered in dosages of 2 to 4 mg/day for 1 to 4 months, has been effective in improving symptoms and peripheral and intestinal eosinophilia in patients with EG.

Antihistamines

Support for use of conventional antihistamines (H1 receptor antagonists) in EGID comes from the murine model of EE induced by aeroallergens and reports of GI tissue eosinophilia in association with seasonal allergies.16,54,175 A study of children with dyspepsia described clinical benefit of antihistamines in 50% of all patients with duodenal eosinophilia. Because of insufficient evidence for the efficacy of other antihistamines in EGID and the fact that a significant proportion of cases are idiopathic, antihistamines probably cannot be recommended as a mainstay of therapy.

Leukotriene Receptor Antagonists

Montelukast selectively and competitively antagonizes the leukotriene receptor Cys-LT1 expressed on bronchial smooth muscle cells and eosinophils. Montelukast thereby blocks the actions of LTD4, a potent and specific eosinophil chemoattractant. Experience with the use of montelukast in EGID is limited to a few reports in EG and EE.176-179 Although some reports disclose persistent tissue eosinophilia, they indicate that montelukast in initial dosages up to 100 mg daily (and maintenance dosages of 20 to 40 mg daily for several months) induces an improvement in peripheral eosinophilia and symptoms.178 Montelukast is approved by the U.S. Food and Drug Administration for use in children 1 year and older with asthma and allergic rhinitis. It has the potential for being a relatively safe and effective steroidsparing therapy for EE.

Anti–Interleukin-5 Therapy (Mepolizumab)

Clinical trials have described reduced peripheral eosinophil counts and clinical benefit with mepolizumab in certain eosinophilic disorders, such as HES and EE. An open-label trial in four patients with HES suggested the efficacy and safety of mepolizumab, a humanized monoclonal antibody against IL-5.152 Three intravenous doses of anti-IL-5 at fourweek intervals lowered peripheral eosinophilia during the 12 weeks of therapy and improved clinical and quality of life measurements. In the patient with HES and EE who began the trial tolerating only a liquid diet, a 10-fold reduction in tissue eosinophilia and significant improvement in vomiting and dysphagia occurred. These data clearly need confirmation in larger randomized, controlled studies.151,180-182

Anti–IgE Therapy (Omalizumab)

The humanized anti-IgE monoclonal antibody omalizumab, known to be effective therapy against allergic rhinitis and

Chapter 27  Eosinophilic Disorders of the Gastrointestinal Tract asthma, has also been described to have positive effects in EG by improving peripheral and tissue eosinophilia, serum IgE, and symptom scores.182

Other Novel and Emerging Treatments

As understanding of the role of eosinophils in EGID continues to evolve, investigation of the effects of novel agents targeting eosinophils proceeds. These agents include anti-CCR3 antibodies against eosinophil-selective adhesion molecules, a monoclonal eotaxin antibody (CAT-213), and therapeutic agents to enhance eosinophil apoptosis.183

ESOPHAGEAL DILATION AND SURGERY

Dilation of esophageal strictures may be considered as initial therapy for symptomatic relief in patients with EE presenting with dysphagia and food impactions.184 Esophageal mucosal rents and extensive linear abrasions observed in some patients on withdrawal of the endoscope may be exaggerated by bougienage. Therefore, dilation should be undertaken cautiously after pretreatment with topical or oral steroids to reduce the risk of complications.5 In one study, 7 of 13 patients with EE experienced transient relief lasting less than three months, requiring repeated dilations with limited success.185 Most patients reported chest pain after dilation, but none had esophageal perforation despite extensive mechanical trauma. Patients with EG presenting with GI obstruction or perforation are usually treated by surgery. Resection of the obstructing segment is successful in relieving obstruction, but symptoms may persist or recur, warranting close follow-up and adjunctive medical management.186,187

KEY REFERENCES

Canani RB, Ruotolo S, Auricchio L, et al. Diagnostic accuracy of the atopy patch test in children with food allergy-related gastrointestinal symptoms. Allergy 2007; 62:738-43. (Ref 105.) Chehade M, Sicherer SH, Magid MS, et al. Multiple exudative ulcers and pseudopolyps in allergic eosinophilic gastroenteritis that

responded to dietary therapy. J Pediatr Gastroenterol Nutr 2007; 45:354-357. (Ref 75.) Collins MH. Histopathologic features of eosinophilic esophagitis. Gastrointest Endosc Clin N Am 2008; 18:59-71. (Ref 118.) Foroughi S, Foster B, Kim N, et al. Anti-IgE treatment of eosinophilassociated gastrointestinal disorders. J Allergy Clin Immunol 2007; 120:594-601. (Ref 182.) Fox VL. Eosinophilic esophagitis: Endoscopic findings. Gastrointest Endosc Clin N Am 2008; 18:45-57. (Ref 115.) Fulkerson PC, Rothenberg ME. Origin, regulation and physiological function of intestinal oeosinophils. Best Pract Res Clin Gastroenterol 2008; 22:411-23. (Ref 10.) Furuta GT, Liacouras CA, Collins MH, et al. Eosinophilic esophagitis in children and adults: A systematic review and consensus recommendations for diagnosis and treatment. Gastroenterology 2007; 133:134263. (Ref 5.) Hogan SP, Mishra A, Brandt EB, et al. A pathological function for eotaxin and eosinophils in eosinophilic gastrointestinal inflammation. Nat Immunol 2001; 2:353-60. (Ref 26.) Liacouras CA, Spergel JM, Ruchelli E, et al. Eosinophilic esophagitis: A 10-year experience in 381 children. Clin Gastroenterol Hepatol 2005; 3:1198-1206. (Ref 44.) Loscher T, Saathoff E. Eosinophilia during intestinal infection. Best Pract Res Clin Gastroenterol 2008; 22:511-36. (Ref 126.) Sampson HA, Ho DG. Relationship between food-specific IgE concentrations and the risk of positive food challenges in children and adolescents. J Allergy Clin Immunol 1997; 100:444-51. (Ref 108.) Schaefer ET, Fitzgerald JF, Molleston JP, et al. Comparison of oral prednisone and topical fluticasone in the treatment of eosinophilic esophagitis: A randomized trial in children. Clin Gastroenterol Hepatol 2008; 6:165-73. (Ref 168.) Sicherer SH. Food protein-induced enterocolitis syndrome: Case presentations and management lessons. J Allergy Clin Immunol 2005; 115149-56. (Ref 95.) Stein ML, Collins MH, Villanueva JM, et al. Anti–IL-5 (mepolizumab) therapy for eosinophilic esophagitis. J Allergy Clin Immunol 2006; 118:1312-19. (Ref 181.) Talley NJ. Gut eosinophilia in food allergy and systemic and autoimmune diseases. Gastroenterol Clin North Am Jun 2008; 37:307-32. (Ref 6.) Full references for this chapter can be found on www.expertconsult.com.

435

CHAPTE R

28 Protein-Losing Gastroenteropathy David A. Greenwald

CHAPTER OUTLINE Definition and Normal Physiology  437 Pathophysiology  437 Clinical Manifestations  438 Diseases Associated with Protein-Losing Gastroenteropathy  440 Diseases without Mucosal Erosions or Ulcerations  440 Diseases with Mucosal Erosions or Ulcerations  441

DEFINITION AND NORMAL PHYSIOLOGY Protein-losing gastroenteropathy describes a diverse group of disorders associated with excessive loss of serum proteins into the gastrointestinal (GI) tract. This excess serum protein loss can result in hypoproteinemia. In 1947, Maimon and colleagues postulated that fluid emanating from the large gastric folds in patients with Ménétrier’s disease was rich in protein. In 1949, Albright and colleagues discovered, using intravenous infusions of albumin, that hypoproteinemia resulted from excessive catabolism of albumin rather than decreased albumin synthesis.1 By 1956, Kimbel and colleagues demonstrated an increase in gastric albumin production in patients with chronic gastritis; a year later, Citrin and colleagues2 were able to show that the GI tract was the actual site of excess protein loss in patients with Ménétrier’s disease. They showed that the excess loss of intravenously administered radioiodinated albumin could be explained by the appearance of labeled protein in the gastric secretions of such patients. Subsequent research using 131I-labeled polyvinylpyrrolidone, 51Cr-labeled albumin, and other radiolabeled proteins, as well as immunologic methods measuring enteric loss of α1-antitrypsin (α1-AT), has further characterized the role of the GI tract in the metabolism of serum proteins. In fact, GI tract loss of albumin normally accounts for only 2% to 15% of the total body degradation of albumin, but in patients with severe protein-losing GI disorders, this enteric protein loss may extend to up to 60% of the total albumin pool.3-5 Under physiologic conditions, most endogenous proteins found in the lumen of the GI tract are derived from sloughed enterocytes and from pancreatic and biliary secretions.6 Studies of serum protein loss into the GI tract measured by various methods (e.g., 67Cu-ceruloplasmin, 51Cr-albumin, or a1-AT clearance) have shown that daily enteric loss of Drs. Karen Kim and Thomas Brasitus contributed to previous versions of this chapter.

Diseases with Lymphatic Obstruction or Elevated Lymphatic Pressure  441 Diagnosis  441 Laboratory Tests  441 Approach to the Patient with Suspected Protein-Losing Gastroenteropathy  442 Treatment and Prognosis  443

serum proteins accounts for less than 1% to 2% of the serum protein pool in healthy individuals, with enteric loss of albumin accounting for less than 10% of total albumin catabolism. In normal subjects, the total albumin pool is approximately 3.9 g/kg in women and 4.7 g/kg in men, with a half-life of 15 to 33 days and a rate of hepatic albumin synthesis of 0.15 g/kg/day, equaling the rate of albumin degradation.7 Excess proteins that enter the GI tract are metabolized by existing proteases much like other peptides, broken down to constituent amino acids, and then reabsorbed. In healthy individuals, GI losses play only a minor role in total protein metabolism, and serum protein levels reflect the balance between protein synthesis and metabolism. However, this balance can be altered markedly in patients with protein-losing gastroenteropathy.8

PATHOPHYSIOLOGY Excessive plasma protein loss across the GI epithelium can result from several pathologic alterations of healthy mucosa. Mucosal injury can result in increased permeability to plasma proteins; mucosal erosions and ulcerations can result in the loss of an inflammatory, protein-rich exudate; and lymphatic obstruction or increased lymphatic hydrostatic pressure can result in direct leakage of lymph, which contains plasma proteins. Changes in vascular permeability can affect the concentration of serum proteins in the interstitial fluid, thereby influencing the amount of enteric mucosal protein loss.9 Hypoproteinemia seen in GI dis­ orders can therefore be classified into three groups: (1) increased mucosal permeability to proteins as a result of cell damage or cell loss; (2) mucosal erosions or ulcerations; and (3) lymphatic obstruction. Examining the pathogenesis of protein-losing gastroenteropathy, Bode and colleagues have suggested that the condition might be related to loss of heparan sulfate proteins that are normally present on the surface of intestinal epithelial cells.10,11 Heparan sulfate proteoglycans appear to affect the

437

438

Section IV  Topics Involving Multiple Organs intestinal barrier by having large extracellular domains that bind to the plasma membrane, known as syndecans, or are attached to a membrane glycolipid, called a glypican.12 These syndecans are important in the maintenance of tight intercellular junctions (see Chapters 2 and 96). A myriad of diseases are associated with protein-losing gastroenteropathy. These are listed in Table 28-1 and discussed in more detail elsewhere in this text.13-63 Mice that were genetically altered to lack syndecans or other heparan sulfate proteins have alterations to the normal tight intercellular barrier, and leak protein via paracellular pathways into the intestinal lumen (Fig. 28-1). Moreover, treatment of such mice with proinflammatory cytokines such as tumor necrosis factor-a (TNF-a) or interferon-γ leads to significantly defective intercellular junctions and even greater protein loss into the intestine.10 The combination of a syndecan-deficient state and exposure to proinflammatory cytokines leads to even greater albumin flux and protein loss. Finally, reintroduction of heparin sulfate or other syndecans abolishes the protein loss into the lumen of the bowel. The loss of serum proteins in patients with protein-losing gastroenteropathy is independent of their molecular weight, and therefore the fraction of the intravascular pool degraded daily remains the same for various proteins, including albumin, immunoglobulin G (IgG), IgA, IgM, and ceruloplasmin.8 In contrast, patients with nephrotic syndrome preferentially lose low molecular weight proteins such as albumin. As proteins cross into the GI tract, synthesis of new proteins occurs in a compensatory fashion. Proteins that enter the GI tract are metabolized into constituent amino acids by gastric, pancreatic, and small intestinal enzymes, reabsorbed by specific transporters, and recirculated. When the rate of gastric or enteric protein loss, or both, exceeds the body’s capacity to synthesize new protein, hypoproteinemia develops.6 Hypoalbuminemia, for example, is common in protein-losing gastroenteropathy and results when there is an imbalance between hepatic albumin synthesis, which is limited and can increase only by 25%, and albumin loss, with reductions in the total body albumin pool and albumin half-life.9 Adaptive changes in endogenous protein catabolism may compensate for excessive enteric protein loss, resulting in unequal loss of specific proteins. For example, proteins such as insulin, clotting factors, and IgE have rapid catabolic turnover rates (short half-lives) and, as such, are relatively unaffected by GI losses, because rapid synthesis of these proteins ensues. On the other hand, proteins such as albumin and most gamma globulins, except IgE, are limited in their ability to respond to GI losses, so protein loss from the gut will be manifested by hypoproteinemia (hypoalbuminemia and hypoglobulinemia).8 Other factors also can contribute to the excessive enteric protein loss seen in various diseases. These include impaired hepatic protein synthesis and increased endogenous degradation of plasma proteins. In addition to hypoproteinemia, protein-losing gastroenteropathy can result in reduced concentrations of other serum components, such as lipids, iron, and trace metals.8 Lymphatic obstruction can result in lymphocytopenia, with resultant alterations in cellular immunity.

CLINICAL MANIFESTATIONS Hypoproteinemia and edema are the principal clinical manifestations of protein-losing gastroenteropathy. Most other

Table 28-1  Disorders Associated with Protein-Losing Gastroenteropathy Diseases without Mucosal Erosions or Ulcerations AIDS-associated gastroenteropathy13 Acute viral gastroenteritis14 Allergic gastroenteropathy15 Celiac disease16 Cobalamin deficiency17 Collagenous colitis18 Cytomegalovirus infection19 Eosinophilic gastroenteritis20 Giant hypertrophic gastropathy (Ménétrier’s disease)21,22 Giardiasis, schistosomiasis, nematodiasis, strongyloidiasis Helicobacter pylori gastritis Henoch-Schönlein purpura23 Hypertrophic hypersecretory gastropathy Intestinal parasitosis24-26 Lymphocytic colitis18 Lymphocytic gastritis Postmeasles diarrhea21 Small intestinal bacterial overgrowth27 Systemic lupus erythematosus28-30 Tropical sprue31 Vascular ectasia (gastric, colonic)32 Whipple’s disease33 Diseases with Mucosal Erosions or Ulcerations α Chain disease34 Amyloidosis35 Behçet’s disease 36 Carcinoid syndrome Crohn’s disease37,38 Duodenitis39 Erosive gastritis39 Gastrointestinal carcinomas Graft-versus-host disease40 Helicobacter pylori gastritis41-43 Idiopathic ulcerative jejunoileitis44 Infectious diarrhea (e.g., Clostridium difficile,45 Shigella spp46) Ischemic colitis Kaposi’s sarcoma47 Lymphoma Multiple myeloma Neurofibromatosis48 Nonsteroidal anti-inflammatory drug enteropathy49 Sarcoidosis50 Toxic shock syndrome (Streptococcus pyogenes)51 Ulcerative colitis52 Waldenström’s macroglobulinemia53 Diseases with Lymphatic Obstruction or Elevated Lymphatic Pressure Cardiac disease54-56 Congestive heart failure, constrictive pericarditis, tricuspid regurgitation, Fontan procedure (see text) Crohn’s disease37,38 Intestinal endometriosis57 Intestinal lymphangiectasia (congenital, acquired)58,59 Lymphatic-enteric fistula24 Lymphoma, including mycosis fungoides Mesenteric tuberculosis and sarcoidosis50 Mesenteric venous thrombosis60 Neoplastic disease involving mesenteric lymphatics Portal hypertensive gastroenteropathy61 Post-transplant lymphoproliferative disease62 Retroperitoneal fibrosis Sclerosing mesenteritis63 Superior vena cava thrombosis Systemic lupus erythematosus29,30 Tuberculosis peritonitis Whipple’s disease33 AIDS, acquired immunodeficiency syndrome.

Chapter 28  Protein-Losing Gastroenteropathy Normal mouse intestine

Syndecan-1–deficient

Ions, nutrient solutes, proteins, bacteria, toxins Intestinal lumen

Solutes, serum proteins

Lamina propria

A

Solutes, serum proteins

Syndecan-1–deficient + TNF-α and interferon-γ (or increased venous pressure)

B

Syndecan-1–deficient + TNF-α and interferon-γ + heparin sulfate analogs Ions, nutrient solutes, proteins, bacteria, toxins

Solutes, serum proteins

C

Solutes, serum proteins

Solutes, serum proteins

D

Solutes, serum proteins

Figure 28-1.  Diagrams illustrating the factors that contribute to intestinal integrity in the mouse. A, The normal mouse intestine makes an effective barrier against free diffusion of certain ions, nutrient solutes, proteins, bacteria, and toxins to separate the intestinal lumen (outside) from the lamina propria (inside) effectively. B, As noted by Bode and colleagues,10 syndecan-1–deficient mice have decreased intestinal barrier function as a result of defective intercellular junctions and increased paracellular leaks (dashed line) or increased transcellular protein transport (solid line). C, Syndecan-1–deficient mice that were treated with inflammatory cytokines such as tumor necrosis factor-α (TNF-α) and interferon-γ or surgically to increase their portal venous pressure have massively defective intercellular junctions and large intercellular protein leaks (dashed lines), consistent with protein-losing enteropathy. D, Infusions of heparin sulfate analogs completely reverse the intestinal barrier dysfunction seen in syndecan-1–deficient mice treated with inflammatory cytokines. See text for more details. (From Lencer WI. Patching a leaky intestine. N Engl J Med 2008; 359:526-8, with permission.)

439

440

Section IV  Topics Involving Multiple Organs Table 28-2  Clinical Manifestations of Protein-Losing Gastroenteropathy Symptoms and Signs Edema (dependent, upper extremity, facial, macular; unilateral in lymphangiectasia) Diarrhea Retinal detachment (in lymphangiectasia)65 Laboratory Abnormalities Hypoproteinemia Hypoalbuminemia Decreased serum gamma globulins (IgG, IgA, IgM) Decreased serum proteins—ceruloplasmin, α1-antitrypsin, fibrinogen, transferrin, hormone-binding proteins Decreased serum lipoproteins Evidence of fat malabsorption Evidence of carbohydrate malabsorption Evidence of fat-soluble vitamin malabsorption or deficiency Altered cellular immunity66 Lymphocytopenia Ig, immunoglobulin.

clinical features reflect the underlying disease process and, as such, the clinical presentation of patients with proteinlosing gastroenteropathy is varied (Table 28-2). Hypoproteinemia, the most common clinical sequela, is manifested by a decrease in serum levels of albumin, gamma globulins (IgG, IgA, and IgM, but not IgE), fibrinogen, lipoproteins, a1-AT, transferrin, and ceruloplasmin.8 Levels of rapid turnover proteins, such as retinal binding protein and prealbumin, are typically preserved, despite hypoproteinemia.64 Dependent edema is frequently a clinically significant issue, and results from diminished plasma oncotic pressure. Anasarca is rare in protein-losing gastroenteropathy. Unilateral edema, upper extremity edema, facial edema and macular edema (with reversible blindness), and bilateral retinal detachments have been seen as a consequence of intestinal lymphangiectasia.65 Despite a decrease in serum gamma globulin levels, increased susceptibility to infections is uncommon. Although clotting factors may be lost into the GI tract, coagulation status typically remains unaffected. Circulating levels of proteins that bind hormones, such as cortisol and thyroid-binding proteins, may be substantially decreased, but levels of circulating free hormones are not significantly altered. Most of the clinical findings in patients with proteinlosing diseases are the result of the underlying disease state and are not cased by the protein loss itself. For example, small bowel disorders with protein loss as a feature, such as celiac disease or tropical sprue, may be associated with malabsorption and resultant diarrhea, fat-soluble vitamin deficiencies, and anemia. Lymphatic obstruction, as occurs with lymphangiectasia, may be seen as lymphocytopenia or abnormal cellular immunity.66

DISEASES ASSOCIATED WITH PROTEIN-LOSING GASTROENTEROPATHY Diseases associated with protein-losing gastroenteropathy can be divided into three broad categories: (1) diseases without GI mucosal erosions or ulcerations; (2) diseases with GI mucosal erosions or ulcerations; and (3) diseases leading to elevated lymphatic and interstitial pressure (see Table 28-1). More than one of these mechanisms may be operative in some disease states, as is the situation for some infectious diseases.

DISEASES WITHOUT MUCOSAL EROSIONS OR ULCERATIONS

Diseases that damage the GI epithelium without causing erosions or ulcers may lead to surface epithelial cell shedding, resulting in excess protein loss. Lesions of the small intestine that cause malabsorption are often associated with enteric leakage of plasma proteins. Protein loss also may be caused by alterations in vascular permeability caused by vascular injury, such as in lupus vasculitis, IgE-mediated inflammation from an allergic response, infection (parasitic, viral, bacterial overgrowth), increased intercellular permeability, or increased capillary permeability.24-30

Ménétrier’s Disease

Giant hypertrophic gastropathy (Ménétrier’s disease; see Chapter 51) is the most common gastric lesion causing severe protein loss.21,22 Patients usually have dyspepsia, postprandial nausea, emesis, edema, and weight loss and are found to have hypoproteinemia. Prominent and thick gastric folds with substantial mucus and protein-rich exudates are seen; normal gastric glands are replaced by mucussecreting cells, reducing the number of parietal cells and resulting in hypochlorhydria or achlorhydria. An increase in intercellular permeability results in protein loss. In this disorder, tight junctions between cells are wider than those found in healthy subjects, and it is believed that proteins traverse the gastric mucosa through these widened spaces. Histamine (H2) receptor antagonists, anticholinergic agents, and octreotide may be used to improve symptoms, but patients with persistent abdominal pain or severe unrelenting protein loss require subtotal or total gastrectomy. A possible causal relationship appears to exist between Helicobacter pylori infection and Ménétrier’s disease with protein-losing gastroenteropathy, because resolution of the hypoproteinemia and return of the gastric folds to their normal configuration may occur with eradication of the organism.41-43

Helicobacter pylori Gastritis

H. pylori gastritis in the absence of Ménétrier’s disease (see Chapter 50) has been associated with protein-losing gastropathy and responds to eradication of H. pylori infection.41-43 Some of these patients may have gastric erosions through which protein may be lost.

Allergic Gastroenteropathy

Although allergic gastroenteropathy (see Chapters 9 and 27) is often considered a disease of childhood, it may be seen in adults as well. This syndrome is manifest by symptoms including abdominal pain, vomiting, and sporadic diarrhea; findings include hypoproteinemia, iron deficiency anemia, and peripheral eosinophilia. Serum levels of total protein and albumin, as well as IgA and IgG, will be markedly reduced, whereas levels of IgM and transferrin will be only moderately diminished. Characteristic histology of the small bowel in patients with this disorder includes a marked increase in the number of eosinophils in the lamina propria, and Charcot-Leyden crystals may be found on stool examination.15

Systemic Lupus Erythematosus

Systemic lupus erythematosus (SLE) is a systemic autoimmune disease not infrequently associated with proteinlosing gastroenteropathy (Fig. 28-2).29,30 Mesenteric vasculitis can result in intestinal ischemia, edema, and altered intestinal vascular permeability. In addition, gastritis and mucosal ulcerations, both of which may contribute to excess protein loss, can develop in patients with SLE. Protein-

Chapter 28  Protein-Losing Gastroenteropathy

A

B

Figure 28-2.  A, Computed tomography (CT) scan of the abdomen in a 29-year-old woman with severe watery diarrhea and diffuse, nonradiating abdominal pain. The serum albumin level was 2.9 g/dL and creatinine level was 0.6 mg/dL. Stool studies were negative for pathogens. The CT scan shows diffuse small bowel wall thickening. The titer of antinuclear antibodies was 1:1280, and she was started on methylprednisolone. Her symptoms improved rapidly, with much less diarrhea and resolution of her abdominal pain. B, Repeat CT scan five days later showed marked improvement of bowel wall thickening, at which time the serum albumin level was 3.4 g/dL. Outpatient renal biopsy confirmed changes consistent with those of systemic lupus erythematosus.

losing gastroenteropathy may be the initial clinical presentation of SLE. Therapy with systemic glucocorticoids, as well as other immunomodulatory agents such as azathioprine and cyclophosphamide, can lead to remission with resolution of clinical symptoms, including protein-losing gastroenteropathy.29,30

DISEASES WITH MUCOSAL EROSIONS OR ULCERATIONS

Mucosal erosions or ulcerations resulting in protein-losing gastroenteropathy can be localized or diffuse and can be caused by benign or malignant disease (see Table 28-1). The severity of protein loss depends on the degree of cellular loss and the associated inflammation and lymphatic obstruction. Diffuse ulcerations of the small intestine or colon, as seen with Crohn’s disease, ulcerative colitis, and pseudomembranous colitis, can result in severe protein loss.37,38,52 Hypoalbuminemia is common in patients with GI tract malignancies; although this is often the result of a decrease in albumin synthesis, excessive enteric protein loss has been reported. Protein-losing gastroenteropathy also has been related to cancer therapy including chemotherapy, radiation-related injury, and bone marrow transplantation.

DISEASES WITH LYMPHATIC OBSTRUCTION OR ELEVATED LYMPHATIC PRESSURE

Lymphatic obstruction results in dilation of intestinal lymphatic channels and can result in rupture of lacteals rich in plasma proteins, chylomicrons, and lymphocytes. When central venous pressure is elevated, such as in congestive heart failure or constrictive pericarditis, bowel wall lymphatic vessels become congested, resulting in a loss of protein-rich lymph into the GI tract.54-56 Tortuous, dilated mucosal and submucosal lymphatic vessels are also seen in patients with primary intestinal lymphangiectasias (Fig. 28-3). These patients often present by 30 years of age with edema, hypoproteinemia, diarrhea, and lymphocytopenia from both lymphatic leakage and rupture.58,59 Retroperitoneal processes such as adenopathy, fibrosis, and pancreatitis can also impair lymphatic drainage.

Figure 28-3.  Intestinal lymphangiectasia. This small intestine biopsy specimen was obtained from a patient with protein-losing enteropathy. It shows focal lymphangiectasia (i.e., two villi are involved and two are spared), consistent with an acquired (secondary) lymphangiectasia. A more diffuse lymphangiectasia would favor a congenital type of lymphangiectasia. (Courtesy of Dr. Edward Lee, Washington, DC.)

An association between protein-losing gastroenteropathy and heart disease is seen after the Fontan procedure, a sur­ gical correction for a congenital, univentricular heart. The surgery creates a wide anastomosis between the right atrium and pulmonary artery, and protein-losing gastroenteropathy has been noted in up to 15% of patients in the ensuing 10 years.67 Hemodynamic studies in such patients reveal increased central venous pressures.

DIAGNOSIS LABORATORY TESTS

Because hypoproteinemia and edema are seen in other disorders in addition to protein-losing gastroenteropathy, documentation of excessive protein loss into the GI tract is important. Patients with unexplained hypoproteinemia in

441

442

Section IV  Topics Involving Multiple Organs the absence of proteinuria, liver disease, and malnutrition should be investigated for evidence of protein-losing gastroenteropathy. The gold standard for diagnosing proteinlosing gastroenteropathy, measurement of the fecal loss of radiolabeled, intravenously administered macromolecules such as 51Cr-albumin, has significant limitations, such as exposure to radioactive material and a 6- to 10-day collection period. Therefore, this test is not clinically useful.68 α1-AT is a useful marker of intestinal protein loss. a1-AT is a 50-kd glycoprotein similar in size to albumin (67 kd). a1-AT, like albumin, is synthesized in the liver and is neither actively absorbed nor secreted in the intestine; it is also resistant to luminal proteolysis. a1-AT is normally present in the stool in low concentrations.68-71 Enteric protein loss can be demonstrated by quantifying the concentration of a1-AT in the stool or by measuring its clearance from the plasma; the latter is the more reliable indicator. Therefore, the optimal test is to measure the clearance of a1-AT from the plasma during a 72-hour stool collection, with a1-AT plasma clearance expressed in milliliters/day using this formula: α1-AT plasma clearance = ([daily stool volume] × [stool α1-AT ]) serum α1-AT Plasma clearance of a1-AT can also be used to monitor response to therapy. An a1-AT clearance in excess of 24 mL/day in patients without diarrhea is abnormal. Diarrhea alone can increase a1-AT clearance; thus, an a1-AT clearance exceeding 56 mL/ day in patients with diarrhea is considered abnormal. In addition, there is an inverse correlation between a1-AT plasma clearance and serum albumin concentration; as serum albumin levels fall below 3 g/dL, the clearance of a1-AT exceeds 180 mL/day. In infants, meconium can interfere with a1-AT results (false-positives) because of the higher concentration of a1-AT in meconium and, therefore, this test should not be performed on infants suspected of having protein-losing enteropathy.68-71 Intestinal bleeding also leads to false elevations of a1-AT clearance. In patients who test positive for fecal occult blood, interpretation of a1-AT clearance can be difficult because of increased clearance rates.68-71 Finally, a1-AT is degraded by pepsin at a gastric pH below 3 and thus cannot be relied on to measure gastric protein loss (false-negatives); the use of a proton pump inhibitor to prevent peptic degradation of a1-AT in the stomach may allow detection of protein-losing gastropathy. Nuclear studies are available to aid in the diagnosis of protein-losing gastroenteropathy; these include technetium99m (99mTc)–labeled human serum albumin (99mTc-HSA), 99m Tc-labeled methylene diphosphonate (99mTc-MDP), 99mTclabeled dextran scintigraphy, 99mTc-labeled human immunoglobulin, and indium-111 (111In)–labeled transferrin.72-74 Nuclear imaging may be useful to quantify protein loss or localize a site-specific area of protein loss and can be helpful in establishing a diagnosis when the a1-AT clearance results are equivocal. Of these tests, 99mTc-labeled dextran scintigraphy may be more sensitive than 99mTc-HSA, although neither test is widely available. Studies in children and adults have used 99mTc-HSA for detecting the specific site of gastric or enteric protein loss, and this test can also be used to monitor response to therapy. 99mTc-labeled human immunoglobulin and 111In-labeled transferrin also may help quantify and localize protein loss into the GI tract.75,76 Magnetic resonance imaging (MRI) has been described as a useful tool for the diagnosis of primary protein-losing gastroenteropathy, readily characterizing lesions that may be

associated with protein loss into the gut, such as dilated mesenteric lymphatics in the abdomen and prominent subcutaneous lymphatics in the extremities.77

APPROACH TO THE PATIENT WITH SUSPECTED PROTEIN-LOSING GASTROENTEROPATHY

The diagnosis of protein-losing gastroenteropathy is usually made on the basis of an increase in a1-AT clearance, in the absence of confounding variables just discussed, with nuclear testing such as 99mTc-HSA helping confirm and quantitate the extent and location of the disorder in certain patients, and directing the evaluation to a specific organ (Fig. 28-4). Testing to confirm protein loss from the GI tract is critical to establishing the diagnosis of proteinlosing gastroenteropathy because many other diseases can present with edema and hypoproteinemia without enteric protein loss. Examples include nephrotic syndrome, cirrhosis, malignancy, eating disorders including bulimia and anorexia, malnutrition, and diuretic or laxative abuse. Following confirmation of enteric protein loss, further evaluation is necessary to identify the underlying disease process. Initial evaluation should include a thorough history and physical examination. Blood testing typically would include a complete blood count with differential (specifically looking for eosinophilia) and red cell indices, electrolytes, calcium, magnesium, serum protein electrophoresis and immunophoresis, C-reactive protein, erythrocyte sedimentation rate, antinuclear antibody (ANA) and rheumatoid factor, coagulation studies, human immunodeficiency virus testing, iron and iron-binding capacity, and thyroid studies. In those patients with diarrhea, a 72-hour fecal fat determination may be useful, if not performed earlier, as well as collection of stool specimens for ova and parasites, Giardia

Initial evaluation

Normal α1-AT clearance

Increased α1-AT clearance

Nephrotic syndrome Liver disease (cirrhosis) Malignancy Eating disorders Diuretic/laxative abuse Malnutrition

Steatorrhea

No steatorrhea

SBFT or capsule endoscopy

Endoscopy/biopsy

Abnormal

Normal

Small bowel biopsy

US CT Cardiac evaluation Laparoscopy Lymphangiogram, if available

Figure 28-4.  Approach to the patient with protein-losing gastroenteropathy. Initial evaluation includes complete history and physical examination, laboratory evaluation (see text), and determination of a1-antitrypsin (a1-AT) plasma clearance. CT, computed tomography; SBFT, small bowel follow-through; US, ultrasonography.

Chapter 28  Protein-Losing Gastroenteropathy antigen, Clostridium difficile toxin, and Charcot-Leyden crystals if peripheral eosinophilia is present. A chest radiograph may reveal granulomatous disease or evidence of cardiomegaly. Electrocardiography or echocardiography may be indicated if increased venous pressure is suspected. In the presence of steatorrhea, diagnostic studies should concentrate on the upper GI tract, and radiologic evaluation of the small intestine, including capsule endoscopy, might be performed. Esophagogastroduodenoscopy and colonoscopy may help detect mucosal inflammation, ulceration, neoplastic disease, or other abnormalities. Biopsies of abnormal-appearing areas should be taken; random biopsies also may have a yield, because conditions such as collagenous or lymphocytic colitis can appear endoscopically normal. Barium studies of the small and large bowel may demonstrate ulcers and mucosal abnormalities. Disorders that might lead to lymphatic obstruction such as fibrosis, pancreatic diseases, or malignancies can be evaluated by computed tomography or MRI of the abdomen and pelvis. Videocapsule endoscopy is useful in evaluating for protein-losing gastroenteropathy to identify the presence of intestinal lymphangietases.78 Lymphangiography may be considered for selected patients, but this test is rarely performed in most centers. When the diagnosis remains unclear, exploratory laparotomy to exclude the possibility of occult malignancy is sometimes appropriate.

TREATMENT AND PROGNOSIS Because protein-losing gastroenteropathy is a syndrome and not a specific disease, treatment is directed at correction of the underlying disease. Protein loss may be offset in part by a high-protein diet, and a diet lower in fat appears to have a beneficial effect on albumin metabolism. Moreover, octreotide may be useful for some patients with protein-losing gastroenteropathy to decrease fluid secretion and protein exudation from the bowel.79 There is some suggestion in experimental mouse models that infusion of heparin analogs may restore intestinal mucosal tight junctions and prevent protein loss across the surface of the bowel; further clinical work is needed to define efficacy.12 For diseases affecting the stomach, such as giant hypertrophic gastropathy (Ménétrier’s disease), gastrectomy reverses protein loss. However, evidence of an infection with H. pylori should be sought before surgical consideration and treated if present (see Chapter 50).41,42 Protein loss from the small bowel should be treated according to the individual disease process present. For example, diseases involving bacterial pathogens such as small intestinal bacterial overgrowth and Whipple’s disease should be treated with appropriate antibiotic therapy (see Chapters 102 and 106), whereas inflammatory processes such as Crohn’s disease or lupus may require immunosuppressive therapy, including glucocorticoids, budesonide, cyclosporine, or cyclophosphamide, or a combination.30,80,81 In the colon, protein loss seen in diseases such as ulcerative colitis and collagenous colitis may require long-term immuno­ modulators or surgery, and infectious colitides need anti­ biotic treatment. Malignancy-induced enteric protein loss

requires cancer-specific therapy. Enteric protein loss and lymphocytopenia seen in cardiac diseases (e.g., congestive heart failure and constrictive pericarditis) can be ameliorated with medical and surgical management of the underlying cardiac condition.55,67,82 Acquired intestinal lymphangiectasia should be treated by correction of the primary disease, whereas congenital intestinal lymphangiectasia can be partially controlled with dietary restrictions. Enteric protein loss in patients with the latter condition can be reduced by a low-fat diet enriched with medium-chain triglycerides, which do not require lymphatic transport and therefore do not stimulate lymph flow.83 Supportive care can reduce the incidence of secondary symptoms. Diuretics typically are not indicated because the edema is caused by a decrease in plasma oncotic pressure; however, diuretics may reduce dependent edema from hypoalbuminemia, thereby improving comfort. Support stockings, if used appropriately, can reduce lower extremity edema in patients with lymphedema and hypoalbuminemia. Exercise and adequate ambulation should be encouraged to reduce the risk of venous thrombosis. Meticulous skin care is critical to prevent skin breakdown and cellulitis. Although these measures do not affect enteric protein loss, they can minimize secondary complications. Most causes of the protein-losing disorders of the GI tract are easily detectable and treatable, and many can be cured. As such, the goal of therapy in protein-losing gastroenteropathy is to identify the cause and direct dietary, medical, or surgical intervention, or a combination, at the underlying disease.8 With reversal or control of the primary disease, a significant proportion of patients will have a partial or complete remission of enteric protein loss, edema, and other associated conditions.

KEY REFERENCES

Bode L, Salvestrini C, Park PW, et al. Heparan sulfate and syndecan-1 are essential in maintaining murine and human intestinal epithelial barrier function. J Clin Invest 2008; 118: 229-38. (Ref 10.) Freeman HJ, Sleisenger MH, Kim YS. Human protein digestion and absorption: Normal mechanisms and protein-energy malnutrition. Clin Gastroenterol 1983; 12:357-78. (Ref 6.) Landzberg BR, Pochapin MB. Protein-losing enteropathy and gastro­ pathy. Curr Treat Options Gastroenterol 2001; 4:39-49. (Ref 8.) Lencer WI. Patching a leaky intestine. N Engl J Med 2008; 359:526-8. (Ref 11.) Rychik J. Protein-losing enteropathy after Fontan operation. Congenit Heart Dis 2007; 2:288-300. (Ref 56.) Sato T, Chiguchi G, Inamori M, et al. Protein-losing gastroenteropathy and gastric polyps: Successful treatment by Helicobacter pylori eradication. Digestion 2007; 75:99. (Ref 42.) Strygler B, Nicar MJ, Santangelo WC, et al. Alpha 1-antitrypsin excretion in stool in normal subjects and in patients with gastrointestinal disorders. Gastroenterology 1990; 99:1380-7. (Ref 70.) Takeda H, Ishihama K. Fukui T, et al. Significance of rapid turnover proteins in protein-losing gastroenteropathy. Hepatogastroenterology 2003; 50:1963-5. (Ref 64.) Touibia N, Schubert ML. Menetrier’s disease. Curr Treat Options Gastroenterol 2008; 11:103-8. (Ref 22.) Wang S, Tsai S, Lan J. Tc-99m albumin scintigraphy to monitor the effect of treatment in protein-losing gastroenteropathy. Clin Nucl Med 2000; 25:197-9. (Ref 72.) Yazici Y, Erkan D, Levine DM, et al. Protein-losing enteropathy in systemic lupus erythematosus: Report of a severe, persistent case and review of pathophysiology. Lupus 2002; 11:119-23. (Ref 29.) Full references for this chapter can be found on www.expertconsult.com.

443

CHAPTE R

29 Gastrointestinal Lymphomas Hsiao C. Li and Robert H. Collins, Jr.

CHAPTER OUTLINE Background  445 General Principles of Lymphoma Management  446 Diagnosis  446 Staging and Prognostic Assessment  447 Treatment  447 Gastric Lymphomas  447 Gastric Marginal Zone B Cell Lymphoma of MucosaAssociated Lymphoid Tissue  447 Diffuse Large B Cell Lymphoma of the Stomach  451 Uncommon Gastric Lymphomas  453 Small Intestinal Lymphomas  453 Marginal Zone B Cell Lymphoma of MALT Type  453

Lymphomas are solid malignancies of the lymphoid system and are subdivided into Hodgkin’s and non-Hodgkin’s lymphomas (NHLs). It was estimated that in 2008, there would be 8,220 and 66,120 new diagnoses of Hodgkin’s and NHL, respectively, in the United States.1 The gastrointestinal tract is very rarely involved with Hodgkin’s lymphoma and will not be discussed in this chapter. There are 0.8 to 1.2 new cases of primary gastrointestinal NHL/100,000 persons/ year.2 This accounts for 30% to 50% of all extranodal NHLs, making the gastrointestinal (GI) tract the most common site of extranodal NHL. Lymphomas that involve the GI tract but have the bulk of the disease in nodal areas are managed in a similar fashion to those that do not involve the GI tract. This chapter deals with primary gastrointestinal lymphoma, which is usually defined as “a lymphoma that has presented with the main bulk of disease in the GI tract, with or without involvement of the contiguous lymph nodes, necessitating direction of treatment to that site.”3

BACKGROUND Lymphomas are malignancies of the immune system. In broad terms, the immune system can be thought of as a highly structured and tightly regulated interaction between lymphoid and nonlymphoid tissues aimed at protecting the host from harmful agents.4 Lymphoid cells are produced in the bone marrow and thymus and then arrayed in the lymphoid tissues, which include the lymph nodes, spleen, Waldeyer’s ring, and mucosa-associated lymphoid tissue (MALT; see Chapter 2). The GI tract lymphoid tissue is MALT, typified by the Peyer patches of the terminal ileum. MALT contains B cells at various stages of differentiation,

Diffuse Large B Cell Lymphoma  453 Mantle Cell Lymphoma  453 Follicular Lymphoma  453 Burkitt’s Lymphoma  454 Immunoproliferative Small Intestinal Disease  455 Enteropathy-Type Intestinal T Cell Lymphoma  457 Uncommon Small Intestinal Lymphomas  459 Other Sites  459 Immunodeficiency-Related Lymphomas  459 Post-transplantation Lymphoproliferative Disorders  459 Human Immunodeficiency Virus–Associated Non-Hodgkin’s Lymphoma  459

organized into different zones (Fig. 29-1A). B cells that have encountered antigen diffusing across the mucosa enter the germinal center of MALT and undergo repeated immunoglobulin gene mutations (somatic mutations)5; the resultant B cell subclones whose immunoglobulins are highly specific for antigen have a survival advantage over B cells whose immunoglobulins are less specific. These more specific B cells then leave the germinal center, enter the circulation, differentiate into memory B cells or antibody-producing plasma cells, and return to the intestinal mucosa. Memory B cells reside in the marginal zone of MALT. Some marginal zone B cells occupy the epithelial tissue that covers the Peyer patches; these cells are called intraepithelial marginal zone B cells. B cells that have not encountered antigen make up the mantle zone of MALT. T cells play a role in the coordination and delivery of the immune system and thus are also found in MALT (see Fig. 29-1A). Therefore, MALT is composed of B and T cells at various stages of differentiation; immune cells at a given stage of differentiation have characteristic histologic, immunophenotypic, and genetic features. Malignant transformation may occur in a cell at any one of these particular stages of differentiation, leading to a malignancy with distinct clinical pathologic features (see Fig. 29-1B). This way of understanding lymphomas has led to the World Health Organization (WHO) lymphoma system, which recognizes at least 28 different clinical pathologic entities.6 Most lymphomas of the GI tract are B cell lymphomas, with most of these resulting from transformation of marginal zone B cells, classified by the WHO system as extranodal marginal zone B cell lymphomas. However, B cell lymphomas can also arise from other cells of MALT, such as centrocytes of the germinal center (follicular lymphomas) or cells of the mantle zone (mantle cell lymphoma). The

445

446

Section IV  Topics Involving Multiple Organs

B GC

T

MZ

Mar

A

B

Figure 29-1.  A, Normal mucosa-associated lymphoid tissue (MALT) of small intestine. The T zone is situated toward the serosal aspect (T). Intraepithelial B cells are also present (B). B, Large B cell lymphoma of the small intestine. Note the infiltration and expansion of the mucosa by the neoplastic cells, with atrophy of the native epithelial structures. GC, germinal center; Mar, pale external marginal zone; MZ, dark surrounding mantle zone. (Courtesy of Dr. Pamela Jensen, Dallas, Tex.)

precise histogenesis of large B cell lymphomas likely varies from case to case. T cell lymphomas of the GI tract are less common and usually involve malignant transformation of intraepithelial T cells in patients with celiac disease (see Chapter 104). GI lymphomas most commonly involve the stomach or small intestine, although the oral pharynx, esophagus, colon, or rectum may be involved uncommonly. In developed countries, the stomach is the most common site of involvement (approximately 60% of cases), but in the Middle East, the small intestine is the most common site of GI involvement. Table 29-1 lists the GI lymphomas discussed in this chapter. Clinicians dealing with GI lymphoma are faced with a specific pathologic diagnosis of a lymphoma occurring in a specific site and, in some cases, modified by important patient characteristics, such as human immunodeficiency virus (HIV) infection. This chapter discusses the main clinicopathologic entities that a clinician may encounter. Certain of these GI lymphomas have particular features that warrant more extensive discussion; these include gastric extranodal marginal zone B cell lymphoma of MALT type, gastric diffuse large B cell lymphoma, immunoproliferative small intestinal disease, and enteropathy-type intestinal T cell lymphoma. The other lym­phomas that may occur in the GI tract are covered in less detail. Before proceeding to discussion of each entity, we briefly review certain general principles that apply to all subtypes.7

Table 29-1  Gastrointestinal Lymphomas Gastric Lymphomas B cell Marginal zone B cell lymphoma of MALT type Diffuse large B cell lymphoma Uncommon types Small Intestinal Lymphomas B cell Non-IPSID Marginal zone B cell lymphoma of MALT type Diffuse large B cell lymphoma Mantle cell lymphoma (multiple lymphomatous polyposis) Follicular lymphoma Burkitt’s lymphoma IPSID T cell Enteropathy-type intestinal T cell lymphoma Other types not associated with enteropathy Other Sites Waldeyer’s ring Esophagus Liver Pancreas Biliary tree Colon Rectum Immunodeficiency-Related Lymphoma Post-transplantation HIV-associated HIV, human immunodeficiency virus; IPSID, immunoproliferative small intestinal disease; MALT, mucosa-associated lymphoid tissue.

GENERAL PRINCIPLES OF LYMPHOMA MANAGEMENT DIAGNOSIS

Because of the many subtypes of NHL, lymphoma should be diagnosed and categorized by a hematopathologist. Sufficient tissue is required for an accurate diagnosis. In the GI tract, this often means multiple endoscopic biopsies. Fineneedle aspiration biopsy is not considered sufficient for diagnosis because it only permits analysis of the morphol-

ogy of individual cells and not an in-depth examination of the background milieu in which those cells reside. The minimal pathologic workup should include light microscopy and immunophenotypic analysis, either by flow cytometry or immunohistochemistry. Staining for immunoglobulin light chains assists in the documentation of monoclonality, with a clear-cut light chain restriction (κ/γ ratio or γ/κ ratio of 10 : 1 or more) strongly suggesting B cell lym-

Chapter 29  Gastrointestinal Lymphomas Table 29-2  Staging Systems for Gastrointestinal Lymphoma TNM STAGING SYSTEM (MODIFIED FOR GASTRIC LYMPHOMA)

ANN ARBOR STAGING SYSTEM

TUMOR INVOLVEMENT

T1 N0 M0 T2 N0 M0 T3 N0 M0

IE IE IE

Mucosa, submucosa Muscularis propria Serosa Perigastric or peri-intestinal lymph nodes More distant regional lymph nodes Invasion of adjacent structures

STAGE

LUGANO STAGING SYSTEM11

I

Confined to GI tract (single primary or multiple, noncontiguous)

II

Extending into abdomen   II1 = local nodal involvement

T1-3 N1 M0

IIE

  II2 = distant nodal involvement

T1-3 N2 M0

IIE

Penetration of serosa to involve adjacent organs or tissues Disseminated extranodal involvement or concomitant supradiaphragmatic nodal involvement

T4 N0 M0

IE

T1-4 N3 M0

IIIE

T1-4 N0-3 M1

IVE

IIE IV

Lymph nodes on both sides of the diaphragm Distant metastases (e.g., bone marrow or additional extranodal sites)

GI, gastrointestinal; TNM, tumor node metastasis. Modified from Zucca E, Bertoni F, Roggero E, Cavalli F. The gastric marginal zone B-cell lymphoma of MALT type. Blood 2000; 96:410-9.

phoma. Occasionally, molecular genetic analysis by Southern blot testing or polymerase chain reaction (PCR) assay is indicated to document monoclonal immunoglobulin or T cell receptor gene rearrangements, or to assess characteristic oncogene rearrangements. Microarray analysis of gene expression profiles may yield important prognostic information, but such analyses are not commonly performed in commercial laboratories and thus are not part of the routine pathologic workup.8

STAGING AND PROGNOSTIC ASSESSMENT

The extent of involvement by NHL is assessed by careful history and physical examination; computed tomography (CT) of the neck, chest, abdomen, and pelvis; positron emission tomography (PET) in cases of high-grade NHL; bone marrow examination; and endoscopic ultrasonography for GI lymphomas.9 Waldeyer’s ring is often involved in gastrointestinal lymphomas, and examination of the upper airway is therefore indicated. The Ann Arbor staging system (Table 29-2),10 which was originally developed for Hodgkin’s lymphoma but is also used for NHL, is deemed by many to be inadequate for staging of GI lymphomas, and several alternative systems have been proposed, two of which are also shown in Table 29-2.11 Prognosis is assessed by defining the distinct lymphoma subtype and evaluating clinical features, including tumor stage, age of the patient, performance status, and serum lactate dehydrogenase (LDH) level. The International Prognostic Index, a model used to predict outcome in patients with aggressive NHL,12 can also be used to assess the prognosis of patients with diffuse large B cell lymphoma of the stomach.13

TREATMENT

Treatment varies according to lymphoma subtype and stage, but it should be noted that the best treatment for many gastrointestinal lymphomas remains controversial. Whereas many large controlled trials have defined the best treatment for many nodal lymphomas, this is not the case for GI lymphomas. Thus, many treatment recommendations are based on small case series and extrapolation from results with nodal lymphomas. With this caveat in mind, we present

the current consensus regarding treatment of the various GI lymphomas.

GASTRIC LYMPHOMAS Primary gastric lymphomas account for 5% of gastric neoplasms, with an increasing worldwide trend.14 The stomach is the most common extranodal site of lymphoma in developed countries.15 Most of these lymphomas are classified as marginal zone B cell lymphoma of the MALT type16 or as diffuse large B cell lymphoma.

GASTRIC MARGINAL ZONE B CELL LYMPHOMA OF MUCOSA-ASSOCIATED LYMPHOID TISSUE

Extranodal marginal zone B cell lymphoma of MALT, also known as MALT lymphoma, comprises about 8% of all NHLs.17 These lymphomas arise from malignant transformation of B cells from the marginal zone of MALT.18 They may arise from MALT that exists under normal physiologic circumstances (e.g., in Peyer’s patches of the gut) or from MALT that has been acquired in sites of inflammation associated with infection or an autoimmune process. For example, gastric tissue normally does not contain MALT but may acquire it in response to chronic Helicobacter pylori infection (see Chapters 50 and 51).19 Malignant transformation occurs in a small percentage of patients with acquired gastric MALT and results in a lymphoma with generally indolent behavior. The malignant process appears to be driven to a large degree by chronic H. pylori infection, because eradication of the infection leads to regression of the lymphoma in most cases.20

Epidemiology

Gastric marginal zone B cell lymphoma of MALT represents approximately 40% of gastric lymphomas.15 The incidence varies according to the incidence of H. pylori in the population being assessed; the incidence in northeastern Italy, where the rate of H. pylori infection is very high, is roughly 13 times the incidence in the United Kingdom.21 The inci-

447

448

Section IV  Topics Involving Multiple Organs dence in H. pylori–infected individuals is from 1 in 30,000 to 1 in 80,000.22 The median age at diagnosis is approximately 60 years, with a wide age range. The male-to-female ratio is equal.

Cause and Pathogenesis

Helicobacter pylori Infection Several lines of evidence support the key role of H. pylori in the development of gastric MALT lymphoma (see Chapter 50). Infection by H. pylori is present in approximately 90% of cases of gastric MALT lymphoma examined histologically23 and in 98% of cases studied by serology.24,25 The epidemiologic studies cited earlier have shown a close correlation between the prevalence of H. pylori infection and gastric lymphoma in a given population,22 and casecontrol studies have shown an association between pre­ vious H. pylori infection and subsequent development of gastric lymphoma.26 In vitro studies have shown that gastric MALT lymphoma tissue contains T cells that are specifically reactive to H. pylori.27 These H. pylori–reactive T cells support the proliferation of neoplastic B cells.28,29 Many groups have documented the regression of gastric MALT lymphoma after eradication of H. pylori.20,30-33 Of interest, responses of small intestinal and rectal lymphoma to H. pylori eradication have been reported,34,35 although a consistent role of the organism at these nongastric sites is not clear. Lymphomas have also been reported in patients with Helicobacter heilmannii infections, with resolution after eradication of the infection.36 Evidence for Antigen-Driven B Cell Proliferation As noted, the B cell immunoglobulin variable region (V) genes undergo somatic hypermutation during the T cell– dependent B cell response to antigen5; this leads to the production of new antigen receptors with altered antigenbinding affinity. Resultant B cell clones that express higher affinity antigen receptors have a survival advantage over B cell clones containing receptors with lower affinity. Thus, somatic mutation is a marker for antigen-driven selection of B cell clones. Sequence analysis of malignant B cells from gastric MALT lymphoma shows that the immunoglobulin genes have undergone somatic mutation.37-39 As discussed in Chapter 2, helper T cells support the proliferation of various B cells through cytokines and cellto-cell interactions.40 These B cells may be specific for the same antigen as the T cells or may be reactive with other antigens. As noted, CD4+ T cells within gastric MALT lymphoma tissue are reactive with H. pylori antigens. The malignant B cells in gastric MALT lymphoma may be reactive with non–H. pylori antigens and, in fact, may be self-reactive. One study has shown that idiotypic immunoglobulins isolated from gastric MALT lymphoma react with a variety of autoantigens.27 Another study has shown that the immunoglobulin genes from gastric MALT lymphoma are derived from germline genes commonly used in the formation of autoantibodies.38 One study has suggested the presence of selecting antigens common to different patients. Analysis of the DNA and amino acid sequences of the antigen-binding region of tumor-derived immunoglobulin from two separate patients showed that although the DNA sequences differed, the resultant amino acid sequences were almost identical.39 Genetic Studies There are four main chromosomal translocations in extranodal marginal zone lymphomas: t(11;18)(q21;q21), t(14;18) (q32;q21), t(1;14)(p22;q32), and t(3;14)(p14.1;q32). The most common translocation is t(11;18)(q21;q21). Overall, it

is found in 30% of cases, but its incidence varies with disease site: it is more common in cases involving the lung and stomach, but rare in other sites.41 The t(11;18) translocation results in the reciprocal fusion of the API-2 and MALT-1 genes. API-2 is an apoptosis inhibitor, and MALT-1 is involved in nuclear factor κB (NF-κB) activation. MALT lymphomas with this translocation do not respond as well to antibiotic therapy aimed at eradicating H. pylori infection as lymphomas without this infection.42 However, they are also less likely to have other chromosomal translocations or transform to more aggressive large cell lymphomas.43,44 The t(14;18)(q32;q21) variant results in the translocation of the MALT-1 gene on chromosome 18q21 to the immunoglobulin gene heavy chain enhancer region, leading to its overexpression, thus differing from the t(14;18) translocation of follicular lymphoma, which involves the bcl-2 gene. This occurs in about 20% of MALT lymphomas overall, although the incidence varies according to the disease site; it is more common in lymphomas occurring in the salivary glands and ocular adnexa, but rare in the GI tract.45,46 Approximately 5% of gastric MALT lymphomas have a t(1;14)(p22;q32) translocation.47 In this translocation, the bcl-10 gene is brought under the control of the immunoglobulin heavy-chain gene enhancer, deregulating its expression. This translocation has been detected only in patients with MALT lymphomas, but those with it often have concurrent trisomies of chromosomes 3, 12, and 18. It is more commonly found in advanced-stage cases, which are less likely to respond to H. pylori eradication.41 The t(3;14)(p14.1;q32) is the latest translocation to be described and results in the juxtaposition of the transcription factor FOXP1 on 3p14.1, next to the immunoglobulin gene heavy (IGH) chain enhancer region.46 The oncogenic role of the IGH-FOXP1 fusion protein is not yet known. Common Molecular Pathway for MALT Lymphoma Chromosomal Translocations The first three translocations listed earlier all activate nuclear factor-kB (NF-kB), a transcription factor that increases cell activation, proliferation, and survival.47 In unstimulated B and T lymphocytes, NF-kB is sequestered in the cytoplasm because it is bound to IkB, an inhibitory protein. Phosphorylation of IkB targets it for ubiquitination and degradation, thus releasing NF-kB, which then translocates to the nucleus to function as a transcription factor. The pathways through which IkB is phosphorylated are tightly regulated and involve BCL-10 and MALT-1. Excessive BCL-10 or MALT-1 activity occurring as a consequence of t(11;18), t(14;18), or t(1;14) leads to constitutive NF-kB activation.47,48 Model for Pathogenesis of Gastric MALT Lymphoma A model for the pathogenesis of gastric MALT lymphoma suggests that the evolution of the disease is a multistage process, comprising the sequential development of H. pylori gastritis, low-grade B cell lymphoma, and then high-grade B cell lymphoma.47,49 This model is supported by gastric biopsies obtained from patients with chronic gastritis taken years before the onset of lymphoma showing B-lymphocytic clones that later gave rise to a clinically evident lymphoma. In this model, H. pylori infection elicits an immune response in which T and B cells are recruited to the gastric mucosa, where MALT is then formed. H. pylori–specific T cells provide growth help to abnormal B cell clones. The abnormal B cells may not be H. pylori–specific and may even be autoreactive. However, their continued proliferation, initially, depends on T cell help. The pivotal role of H. pylori– reactive T cells in driving B cell proliferation may explain

Chapter 29  Gastrointestinal Lymphomas why tumor cells tend to remain localized and why the tumor regresses after eradication of H. pylori. However, continued B cell proliferation eventually leads to accumulation of additional genetic abnormalities, resulting in autonomous growth and more aggressive clinical behavior. Because only a small percentage of H. pylori–infected individuals develops lymphoma, additional currently unknown environmental, microbial, or genetic factors must play a contributory role. H. pylori strains expressing certain proteins such as CagA have been suggested to play a role in the development of gastric lymphoma, but studies have yielded conflicting results.50,51

Pathology

Gross Appearance and Location Low-grade gastric MALT lymphomas may present as a single lesion or as multiple lesions. Unifocal disease usually presents as ulcerated, protruding, or infiltrating masses, but may also manifest as erosions or simply erythema. They are most commonly located in the antrum, but may also be multifocal. Histology The key histologic feature of low-grade MALT lymphoma is the presence of lymphoepithelial lesions (Fig. 29-2).52,53 These lesions are defined as the unequivocal invasion and partial destruction of gastric glands or crypts by tumor cell aggregates. It should be noted, however, that these lesions can sometimes be seen in cases of florid chronic gastritis. Tumor cells are small to medium-sized lymphocytes, with irregularly shaped nuclei and moderately abundant cytoplasm. The morphology of these cells can vary from small lymphoplasmacytoid cells to monocytoid cells that have abundant pale cytoplasm and well-defined borders. Scattered larger cells or transformed blasts may also be seen. The lymphoma cells infiltrate the lamina propria diffusely and grow around reactive follicles; the germinal centers may be invaded, a phenomenon termed follicular colonization. Because there is a continuous spectrum from the transition of gastritis to lymphoma, diagnosis of borderline cases can be difficult. Various parameters may assist in the distinction, such as the prominence of lymphoepithelial lesions,

Figure 29-2.  Lymphoepithelial lesion characteristic of gastric mucosa– associated lymphoid tissue lymphoma. Cytokeratin stain demonstrates invasion and destruction of some gastric glands by a monomorphic population of lymphocytes. Note for comparison the uninvolved normal glands in the bottom center of the photograph. Special stains (not shown) demonstrated Helicobacter pylori. (Courtesy of Dr. Edward Lee, Washington, DC.)

degree of cytologic atypia, and presence of plasma cells with Dutcher bodies (periodic acid–Schiff [PAS]–positive intranuclear pseudoinclusions). The presence of large cells can add further complexity to the diagnosis.14 The low-grade MALT lymphoma may have scattered large cells, but the tumor is composed predominantly of small cells. At the other end of the spectrum, gastric lymphomas that contain only large cells or only small areas of small cell MALT-like lymphoma should be classified as diffuse large B cell lymphomas (see later).6 In between the ends of this spectrum are low-grade lymphomas in the process of evolving into more aggressive lymphoma, with increasing numbers of large cells being observed with transformation. Some investigators have proposed histologic grading systems to take this into account; these systems appear to be able to assess prognosis on the basis of large cell percentages and clusters, but it is unclear how reproducible the systems are.20,54 Immunophenotype Gastric MALT lymphoma cells have the typical immunophenotype of marginal zone B cells. They express pan-B antigens (CD19, CD20, and CD79a) and lack expression of CD5, CD10, CD23, and cyclin D1.55 Further immunostaining by experienced pathologists can aid in identifying lym­ phoepithelial lesions (see Fig. 29-2) and in distinguishing follicular colonization from follicular lymphoma (a rare occurrence in the stomach; see later). Molecular Tests of Monoclonality Southern blotting or PCR assay of immunoglobulin heavy chain rearrangement can assist in the documentation of monoclonality. It should be noted that B cell monoclonality may be detected in H. pylori–associated gastritis (see Chapter 50). Although monoclonality may predict for later development of lymphoma, monoclonality alone does not allow a diagnosis of lymphoma; thus, molecular tests should always be considered in the context of histologic findings.56

Clinical Features

Symptoms, Signs, and Laboratory Tests The most common symptoms are dyspepsia and epigastric pain. Other less common symptoms include anorexia, weight loss, nausea and/or vomiting, and early satiety.15 Gastric bleeding and B symptoms (fevers, night sweats, weight loss) are rare. Serum levels of serum LDH and β2microglobulin are usually normal.57 Diagnosis and Staging Patients are evaluated by esophagogastroduodenoscopy (EGD). Endoscopic findings include erythema, erosions, and/or ulcers. Diffuse superficial infiltration is typical for MALT lymphoma, whereas masses are more commonly seen in diffuse large B cell lymphoma (Fig. 29-3), an aggressive NHL.58 The most common sites of involvement in the stomach are the pyloric antrum, corpus, and cardia, but biopsies should be taken from all abnormal areas and randomly from each area of the stomach, as well as the duodenum and gastroesophageal junction, because disease is often multifocal.59 Because some lymphomas infiltrate the submucosa without involving the mucosal membrane, biopsies need to be sufficiently deep and large for histopathologic and immunohistochemical analyses. H. pylori infection should be established by histologic studies, breath test, or fecal antigen testing (see Chapter 50).60 Endoscopic ultrasound (EUS) can determine the depth of infiltration and assess for the presence of enlarged perigastric lymph

449

450

Section IV  Topics Involving Multiple Organs Table 29-3  Treatment of Gastric Marginal Zone B Cell Lymphoma of MALT Type* LUGANO STAGE

TREATMENT†‡

I, with disease limited to mucosa and submucosa I, with involvement of muscularis propria or serosa; IIE, II

Antibiotics

IV

Best treatment unknown at this time. Radiation or chemotherapy is probably a better option than surgery (see text). Chemotherapy for symptomatic disease. Local management with radiation or surgery may be indicated in selected cases.

*According to Lugano staging system. † Patients with Helicobacter pylori infection should be treated with antibiotics to clear the infection, regardless of stage (see Chapter 50). ‡ Patients with a high percentage of large cells and disease limited to the mucosa may respond to antibiotics alone, although further study of this issue is necessary. Patients with a high percentage of large cells and more advanced-stage disease should be treated as in Table 29-4 for diffuse large B cell lymphoma. MALT, mucosa-associated lymphoid tissue.

Figure 29-3.  Endoscopic appearance of gastric lymphoma with multiple umbilicated lesions adjacent to the gastroesophageal junction. One large ulceration is seen just beyond the squamocolumnar junction.

nodes.61-63 Additional staging consists of upper airway examination, CT scans of the chest, abdomen, and pelvis, bone marrow aspiration and biopsy, and measurement of the serum LDH level. PET is not usually helpful in gastric MALT lymphoma because of low uptake of fluorodeoxyglucose (FDG).64,65 Staging System and Prognostic Assessment In 1994, an international workshop on the staging of GI tract lymphomas proposed the Lugano staging system,11 a modification of the Blackledge system (see Table 29-2). Approximately 75% of gastric MALT lymphomas are confined to the stomach (stage I) at diagnosis66,67 and behave in a clinically indolent fashion; thus, prognosis is good for most patients, with overall survival rates of 80% to 95% at five years. Prognosis is poor in the rare patient with more advanced disease. Additional features associated with a worse prognosis are deep infiltration of the stomach wall, which is associated with a higher likelihood of regional lymph node involvement,68 and high percentages of large cells on histologic evaluation.54

Treatment

Large, randomized clinical trials have not been performed in MALT lymphoma because of the rarity of the disorder. Therefore, treatment recommendations are based on case series and expert opinion. Wotherspoon and colleagues20 first reported that gastric MALT lymphoma could completely regress by endoscopic, histologic, and molecular criteria after eradication of H. pylori. Numerous studies have confirmed these observations,30,33,69-71 and antibiotics aimed at eradicating H. pylori (see Chapter 50) have become the mainstay of therapy for low-grade gastric MALT lymphoma. Even patients with advanced stages of disease can regress with eradication of H. pylori.67 However, it is important to recognize that the current literature in this field is less than optimal in several respects: older studies are limited by insufficient staging procedures and outdated classification systems, none of the reports in the literature

is a controlled or randomized trial, and longer follow-up is necessary. Nevertheless, the current literature is sufficient to suggest to most experts in the field that early-stage disease is best managed with a trial of antibiotics, reserving more toxic therapies such as radiation, chemotherapy, or surgery for cases without concomitant H. pylori infection or those that do not respond to antibiotics.57,72 Table 29-3 summarizes treatment according to stage, according to the Lugano staging system. Stage I Disease Most patients fall into this category and can be treated with antibiotic therapy aimed at eradication of H. pylori. Any one of the treatment regimens discussed in Chapter 50 may be used. Follow-up endoscopy with multiple biopsies should be done six to eight weeks after completion of therapy to document clearance of infection and to assess disease regression. Regression of lymphoma, but not necessarily complete regression, is usually evident at this examination. Patients with persistence of infection should be treated with a second-line antibiotic regimen (see Chapter 50).73 Histopathology at this examination can predict ultimate response, with biopsies showing only small foci of lymphoma being predictive of subsequent complete regression and biopsies showing diffuse persistent disease predicting a low likelihood of subsequent complete regression.74 Patients are then followed with endoscopy approximately every six months for two years and then yearly. Overall, approximately 75% of patients with stage I disease confined to the mucosa and submucosa will achieve complete remission. The median time to remission is five months, with remission usually occurring within 12 months; however, time to remission has been reported to be as long as 45 months.30,72,75 Of patients in clinical remission, 50% have tumor clones detected by PCR.76 With continued follow-up of these patients, the malignant clone decreases; current studies have suggested that a positive PCR at histologic remission does not predict for subsequent relapse, but longer follow-up of this issue is necessary.71 Approximately 90% of patients who had a complete clinical remission to H. pylori eradication remain in remission, with a median follow-up of three years; however late relapses can occur. Relapse may occur in association

Chapter 29  Gastrointestinal Lymphomas with H. pylori reinfection and can be cured by eradicating the organism again.77 In the absence of H. pylori reinfection, relapse is frequently transient.78 Approximately 25% of patients do not respond to H. pylori eradication. Lack of response is more common in patients with t(11;18) translocation; in one study, 67% of nonresponders harbored this abnormality, whereas only 4% of responders did.42 Lack of response to H. pylori eradication is also seen in patients with translocations t(1;14) and t(1;2)72; lymphomas bearing these translocations can be detected by immunohistochemical staining for nuclear BCL10.79 Lack of response to H. pylori eradication may also indicate higher grade lesions or more extensive disease (i.e., involving the gastric wall and regional lymph nodes). The optimal management of disease unresponsive to H. pylori eradication is not certain. Options include surgical resection, chemotherapy, and radiation. These options are discussed in the section on treatment of stage IIE disease (see later). The management of patients with localized disease but a significant percentage of large cells is also uncertain. More recent studies have documented remission to H. pylori eradication,80,81 in contrast to earlier studies. For example, in one study of 16 patients with stage IE disease with high-grade histology, 10 of the patients had disease regression with H. pylori eradication and were free of lymphoma with a median follow-up of 43 months.80 Five of the 6 remaining patients responded to multiagent chemotherapy. If this approach is taken, the patient should be followed closely and, if the response is suboptimal, treated with one of the approaches discussed in the following section. Occasional cases of gastric MALT lymphoma are H. pylori–negative. As would be expected, these patients are much less likely to respond to antibiotic treatment30; however, optimal management remains undefined. Locally Advanced Disease—Stage I with Involvement of Muscularis or Serosa or Poor Response to H. pylori Eradication (Stage IIE).  Patients with more advanced-stage disease who are H. pylori–positive should also receive antibiotic therapy against H. pylori, but antibiotic therapy alone is usually not sufficient to eradicate the lymphoma. There is currently no consensus regarding the optimal management of this group of patients. Total gastrectomy can cure more than 80% of patients with stage IIE disease but diminishes patients’ quality of life and has not been shown to achieve superior results when compared with more conservative approaches.82,83 Involved field radiation therapy (30 to 40 Gy delivered in four weeks to the stomach and perigastric nodes) produces excellent results with a complete remission rate of 90% to 100% and a five-year disease-free survival of approximately 80%.84-86 Radiation therapy is usually well tolerated and preserves gastric function. Thus, it has become the preferred therapy for patients with advanced-stage disease, those who are negative for H. pylori, and those with persistent disease despite H. pylori treatment.72 Other treatment options in this group include chemotherapy, immunotherapy, or combined chemoim­ munotherapy. Single-agent oral chemotherapy using cyclophosphamide87 or chlorambucil has activity,88 as does treatment with purine analogs.89 Immunotherapy with rituximab, a monoclonal antibody against CD20, is also highly efficacious,90,91 although rituximab given with chemotherapy, a common combination in the treatment of aggressive lymphomas, has not been widely tested in low-grade gastric NHLs. The International Extranodal Lymphoma Study Group is currently evaluating the efficacy of this combination in marginal zone lymphomas.

Stage II or IV Disease Low-grade gastric MALT lymphoma that has spread to distant lymph nodes or extranodal sites should be treated as advanced low-grade NHL. Various regimens are used, most incorporating rituximab. Such disease is usually not considered curable, but is generally indolent, with transient responses to chemotherapy. Asymptomatic patients may be followed expectantly.

DIFFUSE LARGE B CELL LYMPHOMA OF THE STOMACH Epidemiology

Approximately 50% of gastric lymphomas are diffuse large B cell lymphomas (DLBCLs). The incidence may be higher in developing than in developed nations, but clinical features appear to be similar.92,93 The median age is approximately 60 years, with a slight male predominance.15,94

Cause and Pathogenesis

The pathogenesis of gastric diffuse large B cell lymphoma is poorly understood.95 Many large cell tumors have components of low-grade MALT tissue and are assumed to have evolved through transformation of low-grade lesions. Frequently, these bear identical rearranged immunoglobulin genes. According to the WHO classification, this is now referred to as diffuse large B cell lymphoma with areas of marginal zone–MALT-type lymphoma.96 However, other DLBCLs have no evidence of associated low-grade MALT tissue. It is unclear whether de novo gastric DLBCL has a worse prognosis than DLBCL with areas of marginal zone– MALT-type lymphoma.68,97 Most studies have suggested this to be the case, but it is unclear what percentage of large cells confers a poorer prognosis. If the large cell lesions commonly arise from progression of low-grade lesions, then conceivably H. pylori may have a role in the initial pathogenesis. One study has suggested that H. pylori infection is more common in patients whose large cell lesions had a low-grade component.97 As outlined earlier in the discussion of tentative models for H. pylori– induced lymphoma, large cell transformation resulting from genetic events, including loss of p53 and p16, may lead to loss of tumor cells losing their dependence on H. pylori for growth.98 A high incidence of somatic mutations in rearranged immunoglobulin heavy-chain variable genes in one study of diffuse large B cell lymphoma of the stomach has implicated antigen selection in the genesis of the lymphoma. Finally, observation of a response of early-stage large cell lymphomas to H. pylori eradication has suggested a role for the organism, at least in some cases.80,81

Pathology

Diffuse large B cell lymphoma may appear grossly as large ulcers, protruded tumors, or multiple shallow ulcers.99 The most common sites of involvement are the body and antrum of the stomach. Tumors with a low-grade component are more likely to be multifocal than tumors with no low-grade component. Large cell lymphomas typically invade the muscularis propria layer or even more deeply. Microscopic examination reveals compact clusters, confluent aggregates, or sheets of large cells that resemble immunoblasts or centroblasts, most often with a mixture of the two.99 From 25% to 40% of cases show evidence of derivation from MALT, including dense infiltration of centrocyte-like cells in the lamina propria and typical lymphoepithelial lesions.97 Immunophenotypic analysis shows expression of one or more B cell antigens (CD19, CD20, CD22, CD79a) and

451

452

Section IV  Topics Involving Multiple Organs CD45.100 Lesions with evidence of low-grade MALT tissue do not express CD10, consistent with their having evolved from the CD10-negative marginal zone low-grade lesions. Lesions without evidence of MALT may or may not express CD10. Genetic analysis reveals monoclonal immunoglobulin gene rearrangements. Bcl-6 is frequently mutated or rearranged.101 It is worth discussing the evolution in terminology regarding diffuse large B cell lymphomas of the stomach. Many pathologists have referred to lymphomas arising in MALT with high-grade features (with or without a component of low-grade disease) as high-grade gastric MALT lymphomas. However, those involved in the development of the WHO classification were concerned that many clinicians had come to regard the term gastric MALT lymphoma as synonymous with a lesion that responds to antibiotics. This is usually not the case with high-grade lesions arising in MALT. Therefore, those involved in formulating the WHO classification96 agreed to use the term extranodal marginal zone B cell lymphoma of MALT type for low-grade lesions, and the term diffuse large B cell lymphoma for high-grade lesions, leaving out the term MALT. Low-grade lesions involving MALT often contain varying proportions of large cells, with a worse prognosis in relation to increased percentage of large cells. However, at this point, a precise grading system for this situation has not been devised and remains a goal of ongoing research.

Clinical Features

Patients present with epigastric pain (70%) or dyspepsia (30%), symptoms similar to those patients with gastric adenocarcinoma.102,103 Large tumors may cause obstruction. Ulcerating lesions may be associated with GI bleeding. B symptoms (fevers, night sweats, weight loss) and elevated serum LDH concentrations are uncommon. Staging consists of EGD, upper airway examination, CT scans of the chest, abdomen, and pelvis or PET scan, bone marrow aspiration and biopsy, and measurement of the serum LDH level. In addition, endoscopic ultrasonography plays an important role in assessing depth of stomach wall involvement. Lastly, H. pylori infection should be assessed. It is detected in 35% of patients with DLBCL of the stomach and is more common in those with concomitant gastric MALT.104 Most patients have stage I or II disease by the Ann Arbor Staging System (see Table 29-2).105 However, other staging systems have been developed and the use of various systems has made it difficult to compare results of different series. In 1994, a Workshop at the International Conference on Malignant Lymphoma proposed a modification to Blackledge’s system, known as Lugano staging (see Table 29-2).11

Treatment

The optimal management of DLBCL of the stomach is controversial, but the current consensus recommends chemoimmunotherapy with or without radiotherapy as a replacement for surgery (Table 29-4).103 Traditionally, localized disease was approached with surgery alone or surgery followed by radiation and/or chemotherapy for patients with poor prognostic features.106 This approach had the advantage of providing diagnostic and staging information and avoided the risk of perforation or bleeding that was believed to result from treatment with chemotherapy or radiation. Approximately 70% of patients with stage I disease are disease-free five years after surgery. However, several investigators have questioned the role of surgery in the management of localized gastric DLBCL. They noted that with the availability of endoscopy, surgery was no longer necessary for diagnosis and, with the availability of

Table 29-4  Treatment of Diffuse Large B Cell Lymphoma of the Stomach* LUGANO STAGE

TREATMENT

I II, II1, II2, IIE IV

CHOP† × 3-4 cycles + XRT‡ + rituximab§ CHOP × 3-4 cycles + XRT + rituximab CHOP × 6-8 cycles + XRT + rituximab

*According to the Lugano staging system, optimal management of this entity is controversial. However, the developing consensus seems to favor combined chemotherapy and radiation and avoidance of surgery (see text). † Cyclophosphamide, doxorubicin [hydroxydaunorubicin], vincristine [Oncovin], prednisone). ‡ XRT (radiotherapy); usually, 30-40 Gy in 20-30 fractions. § The suggestion for the addition of rituximab in this setting involves extrapolation of randomized data from nodal diffuse large B cell lymphoma.

CT and EUS, surgery was no longer necessary for staging. In addition, the risk of bleeding or perforation during chemotherapy is lower than 5% and only a few of those who bleed require urgent gastrectomy.107 Surgery, however, carries a 5% to 10% risk of mortality and is associated with significant morbidity. Thus, chemotherapy and radiation were investigated as alternatives to surgery. Retrospective studies have shown similar outcomes in patients treated with surgery alone versus chemotherapy alone.94 A prospective randomized study of patients with DLBCL of the stomach who were randomized to surgery, surgery plus radiotherapy, surgery plus chemotherapy, and chemotherapy alone showed improved complete response rates and overall survival for patients who received surgery plus chemotherapy and chemotherapy alone when compared with those who received surgery alone or surgery plus radiotherapy.108 The German Multicenter Study GIT NHL 01/92 was a prospective nonrandomized study of surgery in conjunction with chemotherapy and radiation versus chemotherapy and radiation alone for primary gastric lymphoma in localized stages. Whether the treatment included surgery was left to the discretion of each participating center. There was no difference in survival rate between those who received surgery followed by chemoradiotherapy and those who received chemoradiotherapy alone.109 These results were confirmed in a second, larger, prospective nonrandomized trial, GIT NHL 02/96.110 For patients with advanced-stage nodal DLBCL, the addition of rituximab, a monoclonal antibody against CD20, to CHOP (cyclophosphamide, doxorubicin [hydroxydaunorubicin], vincristine [Oncovin], prednisone) chemotherapy has been shown to improve overall survival when compared with CHOP alone.111-113 This combination has also been administered to patients with gastric DLBCL and found to be safe and effective.114 The necessity of radiation therapy in the management of gastric DLBCL is also controversial. A small retrospective study of patients with stage I or II primary gastric high-grade DLBCL treated with chemotherapy with or without radiotherapy has shown decreased relapse rates in patients who received consolidative radiotherapy.107 However, this study included only 21 patients, of whom 3 relapsed, and it is thought that a prospective randomized trial is needed. Thus, standard management of gastric large B cell lymphoma follows standard management of nodal large B cell lymphomas. The treatment of localized (stage I or II) nodal large B cell lymphoma consists of three to six cycles of combination chemotherapy (typically the CHOP regimen)

Chapter 29  Gastrointestinal Lymphomas given with rituximab, with or without consolidative radiotherapy.115 DLBCL patients with evidence of H. pylori infection should be treated. Response of large cell lymphoma has been reported after eradication of H. pylori.81,116 However, these studies must be considered preliminary, and most patients treated with antibiotics alone have had disease limited to the mucosa; most patients with DLBCL of the stomach have more advanced disease, and antibiotics alone are considered inadequate treatment.

UNCOMMON GASTRIC LYMPHOMAS

B cell lymphomas other than marginal zone or diffuse large B cell may involve the stomach uncommonly. Gastric lymphomas of T cell origin have rarely been reported.117,118

SMALL INTESTINAL LYMPHOMAS Small intestinal lymphomas may be divided into B cell tumors and T cell tumors. The B cell tumors include im­munoproliferative small intestinal disease (IPSID) and various non-IPSID subtypes, including marginal zone B cell lymphoma of MALT, DLBCL, mantle cell lymphoma, follicular lymphoma, and Burkitt’s lymphoma. Relatively few reports have described the various non-IPSID small intestinal lymphomas, and large series have tended to group together all the lymphoma subtypes when cataloguing manifestations and treatment outcomes.119-121 Given the lack of information about these diseases with regard to their behavior in the intestine, it is probably best to consider them in light of the well-described features of their nodal counterparts. Thus, marginal zone and follicular lymphomas are regarded as indolent processes, incurable but controllable by chemotherapy, and often associated with a relatively long survival. Diffuse large B cell lymphomas, mantle cell lymphomas, and Burkitt’s lymphomas are more aggressive processes, which generally require chemotherapy as part of their management. T cell lymphomas of the small intestine are usually enteropathy-type intestinal T cell lymphomas; other forms of T cell lymphoma have been rarely reported. Recent reports have suggested the existence of a rare natural killer (NK) cell or NK-type T cell intestinal lymphoma.122-125

MARGINAL ZONE B CELL LYMPHOMA OF MALT TYPE

Lymphoma arising in the small intestine may have the characteristics of marginal zone B cell lymphoma, with the same histologic and immunophenotypic features described earlier for gastric marginal zone B cell lymphoma.124 However, an association with H. pylori infection has not been documented, although rare responses to antibiotics have been reported. Most cases occur in older patients who present with melena. The disease is usually present as a single annular or exophytic tumor,126 which may be present anywhere in the small intestine; disease is usually confined to the intestine or to local nodes. Treatment is generally surgical. Some patients have received chemotherapy, but few data are available regarding regimens and outcome. It should be noted that in nodal marginal zone lymphoma, chemotherapy is usually reserved for patients with symptoms, because the disease is understood to be slow-growing and sensitive to chemotherapy, but not curable by it. The fiveyear survival rate is approximately 75%. As in gastric marginal zone B cell lymphoma, the small intestinal variety

may have varying components of large cell transformation. This feature probably confers a worse prognosis, but data are scanty.

DIFFUSE LARGE B CELL LYMPHOMA

DLBCL of the small intestine is similar to its gastric counterpart in histology and clinical behavior. Patients may present with abdominal pain, weight loss, obstruction, abdominal mass, bleeding, and/or perforation. The tumor is usually an exophytic or annular lesion. Histologic findings are similar to those described earlier for gastric diffuse large B cell lymphoma, with some patients having a low-grade component and others having only a large-cell component. Approximately half of patients have localized disease, and half have disease spread to regional or distant nodes. Surgery is usually required, and additional therapy includes anthracycline-containing chemotherapy and the anti-CD20 monoclonal antibody, rituximab.119 In addition, radiotherapy is sometimes indicated. Prognosis depends on disease stage and patient factors, such as age and performance status.

MANTLE CELL LYMPHOMA

Mantle cell lymphoma is a relatively recently described subtype of B cell NHL.127 Patients typically present with widespread adenopathy and frequently have bone marrow and extranodal involvement. The GI tract is involved in more than 80% of patients, although not all patients with GI involvement are symptomatic.128 The most common manifestation of GI disease is multiple lymphomatous polyposis, in which multiple lymphoid polyps are present in the GI tract (Fig. 29-4).129,130 The most common site of involvement is the ileocecal region, but any other area may be involved from the stomach to the rectum; occasional patients have involvement of all these regions (Fig. 29-5). Involvement of the GI tract may also occur without the appearance of multiple polyps, and the GI tract as the only site of involvement has been reported. When patients have symptoms related to GI involvement, they usually include pain, obstruction, diarrhea, or hematochezia. It should be noted that multiple lymphomatous polyposis can also be seen with other lymphomas, especially marginal zone B cell lymphomas of MALT and follicular lymphomas. Microscopically, mantle cell lymphoma involves the mucosa and submucosa and the malignant cells have the appearance of small atypical lymphocytes, which may surround benignappearing germinal centers or may efface the lymphoid tissue. The tumor cells express pan-B markers and the T cell marker CD5. The disease is characterized by t(11;14), a translocation that results in rearrangement and overexpression of the bcl-1 gene encoding cyclin D1.131 Patients with obstructive tumor masses require surgical therapy, but the mainstay of treatment is chemotherapy. Although mantle cell lymphoma is initially responsive to chemotherapy, it eventually becomes refractory; median survival is three to five years.

FOLLICULAR LYMPHOMA

Follicular B cell lymphomas of the GI tract are rare.132 The most common presentation is as an obstructing lesion in the terminal ileum. As noted, patients with this diagnosis may also present with the gross appearance of multiple lym­ phomatous polyposis. Microscopically, most follicular lymphomas are composed of small cleaved lymphocytes, or centrocytes (Fig. 29-6), with a varying admixture of large cells. The disease is characterized by t(14;18), a translocation that results in overexpression of the bcl-2 gene.133 Obstructing lesions require surgical management. Chemo-

453

454

Section IV  Topics Involving Multiple Organs

A

B

Figure 29-4.  Endoscopic appearance of mantle cell lymphoma presenting as multiple lymphomatous polyposis in the stomach (A) and in the colon (B).

A

B

Figure 29-5.  Multiple lymphomatous polyposis (mantle cell lymphoma). A, Gross photograph shows numerous small polypoid lesions in the cecum. Additional synchronous and metachronous lesions were present or later developed in the ileum and the duodenum, as well as the rectum and sigmoid colon. B, Low-power photomicrograph of ileum shows multiple discrete sites of mucosal and submucosal involvement by lymphomatous polyposis. (Courtesy of Dr. Edward Lee, Washington, DC.)

therapy and radiation are sometimes indicated for the management of this indolent but incurable disorder.

BURKITT’S LYMPHOMA

Figure 29-6.  Follicular lymphoma, World Health Organization grade II. Neoplastic lymphoid follicles are evident, involving the wall of the small intestine and effacing the normal architecture (Hematoxylin and eosin, low power.) (Courtesy of Dr. Imran Shahab and Dr. Pamela Jensen, Dallas, Tex.)

Burkitt’s lymphoma is a highly aggressive malignancy that in patients who are HIV-negative, presents as an endemic form, observed in Africa, or a sporadic form.134 In the sporadic form, patients usually present with disease in the abdomen, with involvement of the distal ileum, cecum, mesentery, or cecum and mesentery. Burkitt’s tumor cells are monomorphic, medium-sized cells with round nuclei, multiple nucleoli, and basophilic cytoplasm (Fig. 29-7). The involved lymphoid tissue microscopically has a starry sky appearance caused by numerous benign macrophages that have ingested apoptotic tumor cells.135 The tumor cells express B cell–associated antigens and surface immunoglobulin. Most cases have a translocation of the c-myc gene, either to the immunoglobulin heavy-chain region on chro-

Chapter 29  Gastrointestinal Lymphomas

A

B

Figure 29-7.  Burkitt’s lymphoma. A, Diffuse involvement of the small bowel by Burkitt’s lymphoma. Note infiltration around native glandular structures. B, High-power view showing brisk mitotic activity and background macrophages. CD20 immunostain (not shown) was strongly positive within the tumor population. (A, Hematoxylin and eosin, ×20; B, hematoxylin and eosin, ×600.) (Courtesy of Dr. Pamela Jensen, Dallas, Tex.)

mosome 14 or to one of the immunoglobulin light-chain regions on chromosomes 2 or 22, resulting in a t(8;14), t(2;8), or t(8;22) translocation.136 Burkitt’s lymphoma is rapidly fatal without treatment but responds immediately to institution of aggressive chemotherapy. Treatment carries a high risk of tumor lysis syndrome. Cure rates are 50% to 90%, depending on the extent of the disease.137,138

IMMUNOPROLIFERATIVE SMALL INTESTINAL DISEASE Epidemiology

Immunoproliferative small intestinal disease (also known as α heavy-chain disease and as Mediterranean lymphoma) is confined to certain regions of the world, especially North Africa, Israel, and surrounding Middle Eastern and Mediterranean countries.139 IPSID is seen less often in other areas, including Central and South Africa, India and East Asia, and South and Central America. A diagnosis in North America or Europe should be questioned, unless the patient has previously lived in an endemic area. The disease occurs in individuals with lower socioeconomic status who live in conditions of poor hygiene and sanitation.140 The disease generally occurs in the second or third decade of life, although it has been observed in older individuals. The incidence in males and females is equal.

Cause and Pathogenesis

Several observations have led to the belief that IPSID may be initiated by an infectious agent or agents: (1) an association of the disease with lower socioeconomic status and poor sanitation; (2) a high prevalence of intestinal bacterial overgrowth and parasitosis; (3) a decrease in incidence when living conditions have improved in endemic areas; and (4) a response of early lesions to antibiotic therapy. In addition, it is known that bowel flora stimulate IgAproducing cells, and intestinal biopsies from apparently normal individuals from endemic regions have shown an increase in lamina propria lymphocytes and plasma cells, reminiscent of findings in patients with IPSID. One study detected the presence of Campylobacter jejuni in five of seven patients with IPSID, suggesting a role for this organism in the disease.141

As discussed later, IPSID is associated with the production of an unusual IgA heavy-chain protein, called α heavy chain, which is secreted by plasma cells and is detectable in various body fluids.142,143 The plasma cells, which are the predominant histologic feature in the superficial mucosa, possess surface and cytoplasmic α chain protein. Centrocyte-like cells proliferating deeper in the mucosa have mainly cytoplasmic α chain protein. It is likely that these centrocyte-like cells, stimulated by microbial antigens, differentiate into the plasma cells that secrete the α chain protein characteristic of the disease. Genetic analyses have revealed that cellular proliferations are monoclonal, even in early lesions.144,145 Thus, it can be proposed that in a way somewhat analogous to H. pylori–associated gastric MALT, lymphocytes in intestinal MALT may be stimulated by infectious agents, in particular Campylobacter jejuni,141 and proliferate in response. The lymphocytic response becomes monoclonal and initially depends on the presence of antigen. However, with time, the malignant cells acquire additional genetic changes, causing them to lose their dependence on antigen persistence. This loss of antigen dependence is associated with the development of more aggressive clinical features.

Pathology

Gross lesions are generally confined to the proximal small intestine, with adenopathy of adjacent mesenteric nodes.146 Although some patients have thickening of mucosal folds only, others have a generalized thickening of the bowel wall, discrete masses, nodules, or polypoid lesions. Although grossly only the proximal bowel wall is involved, histologically the disease is characterized by a dense mucosal and submucosal cellular infiltrate that extends continuously throughout the length of the small intestine. Various pathologic staging systems have been proposed (Table 29-5).146,147 In early-stage disease, the cellular infiltrate is composed of benign-appearing plasma cells or lymphoplasmacytic cells. However, as noted, various studies assessing immunoglobulin gene rearrangements or light chain restriction have suggested that even the earliest infiltrate is monoclonal. This early infiltrate broadens villi and shortens and separates crypts, but epithelial cells remain intact. A histologic variant, the follicular lymphoid type,

455

456

Section IV  Topics Involving Multiple Organs Table 29-5  Pathologic Staging Systems for Immunoproliferative Small Intestinal Disease World Health Organization* (a) Diffuse, dense, compact, and apparently benign lymphoproliferative mucosal infiltration (i) pure plasmacytic ( ii) mixed lymphoplasmacytic (b) As in (a), plus circumscribed “immunoblastic” lymphoma, in either the intestine and/or mesenteric lymph nodes (c) Diffuse “immunoblastic” lymphoma with or without demonstrable, apparently benign, lymphoplasmacytic infiltration Salem et al146 Stage 0: Benign-appearing lymphoplasmacytic mucosal infiltrate (LPI), no evidence of malignancy Stage I: LPI and malignant lymphoma in either intestine (Ii) or mesenteric lymph nodes (In), but not both Stage II: LPI and malignant lymphoma in both intestine and mesenteric lymph nodes Stage III: Involvement of retroperitoneal and/or extra-abdominal lymph nodes Stage IV: Involvement of noncontiguous nonlymphatic tissues Unknown or inadequate staging Galian et al147

STAGE

SMALL INTESTINE: SITE I

A

Mature† plasmacytic infiltration of lamina propria‡, with no or limited disorganization of general lymph node architecture; inconstant and variable villus atrophy Atypical plasmacytic or lymphoplasmacytic infiltrate, with presence of more or less atypical immunoblast-like cells, extending at least to the submucosa; subtotal or total villus atrophy Lymphomatous proliferation invading the whole depth of intestinal wall

B

C

MESENTERIC LYMPH NODES: SITE IIA

OTHER ABDOMINAL AND RETROPERITONEAL LYMPH NODES: SITE IIB

OTHER LYMPH NODES: SITE III

OTHER SITES: SITE IV

Infiltrate cytologically similar to that in site I

Atypical plasmacytic or lymphoplasmacytic infiltrate, with presence of more or less atypical immunoblast-like cells; total or subtotal obliteration of nodal architecture§

Infiltrate cytologically similar to that in site I

Lymphomatous proliferation with total obliteration of nodal architecture‡

Lymphomatous proliferation similar to that in site I

*Alpha-chain disease and related small-intestinal lymphoma: A memorandum. Bull WHO 1976; 54:615-24. † Rare cells may show an immature pattern. ‡ Limited and superficial extensions to submucosa may be observed. § Some sinuses, especially in the peripheral area, may still be recognizable. Modified from Fine KD, Stone MJ. Alpha-heavy chain disease, Mediterranean lymphoma, and immunoproliferative small intestinal disease: A review of clinicopathological features, pathogenesis, and differential diagnosis. Am J Gastroenterol 1999; 94:1139-52.

has been described in some patients (see Fig. 29-7). This variant features a diffuse involvement of the mucosa, with lymphoid follicle-like structures. As the disease progresses to intermediate and late stages, the villi are further broadened and may become completely effaced, crypts are fewer, and the immunoproliferation extends more deeply. Atypical lymphoid cells infiltrate the benign-appearing plasma cells and lymphoplasmacytic cells. With time, the process evolves into overt lymphoma. Mesenteric lymph nodes are enlarged in early lesions, with preserved architecture, although follicles may be shortened by a histologically benign-appearing lymphocytic or plasmacytic infiltrate. As the disease progresses, the lymph node may acquire a more dysplastic appearance.

Clinical Features

Patients usually present with diarrhea, colicky abdominal pain, anorexia, and significant weight loss, with a duration of symptoms from months to years. The diarrhea initially may be intermittent but becomes voluminous and

foul-smelling as malabsorption develops. About half of patients have fever. Physical examination reveals evidence of malnutrition, digital clubbing, and peripheral edema. Late physical manifestations are ascites, hepatosplenomegaly, an abdominal mass, and peripheral lymphadenopathy. Endoscopy may reveal thickened mucosal folds, nodules, ulcers, or evidence of submucosal infiltration, rendering the intestine immobile, tender, and indistensible. Small bowel barium radiographs show diffuse dilation of the duodenum, jejunum, and proximal ileum, with thickened mucosal folds. Patients are frequently anemic because of vitamin deficiencies, and the erythrocyte sedimentation rate is elevated in one third of cases. The circulating lymphocyte count is low, and measures of humoral and cellular immunity are impaired. Stool examination frequently reveals Giardia lamblia infestation. As noted, C. jejuni has been implicated in a high percentage of patients by PCR assay, DNA sequencing, fluorescence in situ hybridization, and immunohistochemical studies on intestinal biopsy specimens.141 Serum IgG

Chapter 29  Gastrointestinal Lymphomas and IgM levels may be high or low; IgA levels are usually low or undetectable. The characteristic and unique laboratory abnormality is the presence of the α chain protein.148 This 29- to 34-kd protein is a free α1 heavy chain with an internal deletion of the variable (VH) and CH1 regions. It is devoid of light chains and thus corresponds to the Fc portion of the α1 subunit of IgA. The α chain protein amino terminal contains sequences that are not homologous to any known immunoglobulin sequence. These changes are often the result of insertions or deletions, usually involving the VH-JH and CH2 regions,149 but the source of inserted genetic material is unknown. The α chain production migrates as a broad band within the α2 and β regions on serum protein electrophoresis. In addition to electrophoresis, the protein can be detected by immunoelectrophoresis or immunoselection (the most sensitive and specific methods)149 in serum, urine, saliva, or intestinal secretions. Detection of α chain protein from these sources is more likely in patients with early disease than in patients with more advanced disease, but, regardless of stage, α chain protein can be detected in tissue sections in most cases of IPSID by immunofluorescence or immunoperoxidase staining of plasma or lymphoma cells.148 It has been postulated that chronic antigenic stimulation of the intestinal IgA secretory apparatus results in expansion of several plasma cell clones. Eventually, a structural mutation occurs in a particular clone, resulting in an internal deletion of part of the α heavy chain. This leads to an inability to make light chains and results in secretion of α chain protein rather than intact IgA.149,150

Diagnosis and Staging

Because the more malignant-appearing histology may be present only in deeper layers of the intestine, endoscopic biopsy alone is often considered an inadequate evaluation; staging laparotomy is therefore strongly recommended by some authors to allow full-thickness intestinal biopsy and biopsy of mesenteric lymph nodes.151 However, it should be noted that some investigators do not routinely perform laparotomies; instead, upper and lower endoscopy, small bowel series, bone marrow biopsies, and fine-needle aspiration of enlarged lymph nodes are performed.152 One of the staging systems may then be applied (see Table 29-5). More advanced disease, poor performance status, and comorbid illnesses portend a worse prognosis.

Treatment

Because of the relative rarity of this lymphoma, no large trials investigating therapy have been carried out.152,153 Patients often require intensive nutritional support.154 Patients with early disease (e.g., Salem stage 0 disease; see Table 29-5) are generally treated with antibiotics for six months or more. The two most commonly used regimens are tetracycline alone and a combination of metronidazole and ampicillin. Response rates have ranged from 33% to 71%139; in one study, the complete response rate was 71%, with disease-free survival of 43% at five years.152 In patients who do not significantly improve by six months or who do not achieve complete remission by 12 months, or who have advanced disease at presentation, chemotherapy should be given. Most investigators recommend anthracyclinecontaining regimens such as CHOP.155,156 For example, one investigator has reported a complete response of 67% and a survival of 58% at 3.5 years in patients treated with antibiotics, total parenteral nutrition, and anthracycline-based combination chemotherapy.156 However, good results have been reported with non–anthracycline-containing regimens as well; in one report, 56% of patients with advanced

disease were free of disease at five years.152 Finally, total abdominal radiotherapy has been used in a small number of patients but, on the basis of current data, it is difficult to assess its proper role.157

ENTEROPATHY-TYPE INTESTINAL T CELL LYMPHOMA

Enteropathy-type T cell lymphoma occurs as a complication of celiac disease (see Chapter 115).158 Malignant transformation of intraepithelial T cells leads to an aggressive malignancy, causing most patients to die within a few months of diagnosis.159,160 Treatment of celiac disease with a glutenfree diet may decrease the risk of this malignancy.161

Epidemiology

Enteropathy-type T cell lymphoma is a rare malignancy and comprises only 0.5% of all NHL cases.162-164 Celiac disease has a prevalence of 0.5% to 1% in the United States and Europe158,165,166 and is more common in whites compared with African Americans and Asians. In a cohort of more than 1000 patients with celiac disease, the most common cause of death was NHL.167 The diagnosis of lymphoma is usually made concomitantly with or shortly after the diagnosis of celiac disease,168 although the two conditions are commonly diagnosed simultaneously, especially in patients who have a long history of malabsorption. Adherence to a gluten-free diet for more than five years appears to reduce the risk of developing lymphoma.161 Males are affected twice as often as females.160 The peak incidence occurs in the sixth or seventh decade of life.169

Cause and Pathogenesis

Enteropathy-type intestinal T cell lymphoma occurs in patients with adult celiac disease.164 As discussed in Chapter 104, celiac disease is characterized by a hereditary sensitivity to gluten.170 Gluten peptides are presented by celiac disease–specific HLA-DQ2 and HLA-DQ8 positive antigenpresenting cells and thus elicit an immune response in which gluten-specific intraepithelial lymphocytes damage intestinal epithelium. Intraepithelial T cells in celiac disease have a normal immunophenotype (CD3+/CD8+) and are polyclonal.171,172 Malignant transformation of intraepithelial T cells results in a monoclonal population of intraepithelial T cells that have an abnormal phenotype.173-176 Monoclonal populations of intraepithelial T cells in celiac mucosa may result in any one of several interrelated processes.176,177 The first condition is refractory celiac disease, a condition in which patients lose responsiveness to a gluten-free diet.178 The second condition, ulcerative jejunitis, is characterized by inflammatory jejunal ulcers and unresponsiveness to a gluten-free diet.179 The third condition is enteropathy-type intestinal T cell lymphoma, an aggressive malignancy of the small intestine.174,175 In patients with any of these three conditions, uninvolved mucosa adjacent to the lesions can contain monoclonal T cells containing the same rearranged T cell receptor genes.180 In addition, patients with ulcerative jejunitis can subsequently develop enteropathy-type intestinal T cell lymphoma, in which the same clone is isolated in the jejunitis and the subsequent lymphoma. Thus, these three conditions have come to be considered to represent a spectrum of disorders mediated by monoclonal intraepithelial T cells. Comparative genomic hybridization studies have shown recurrent chromosomal gains in enteropathy-type intestinal T cell lymphoma at chromosomes 9q, 7q, 5q, and 1q and recurrent losses at 8p, 13q, and 9p. A gain at 9q is the most common, seen in 58% of cases examined.181 Another study

457

458

Section IV  Topics Involving Multiple Organs Clinical Features

Figure 29-8.  Photomicrograph of enteropathy-type intestinal T cell lymphoma in a patient with celiac disease. Mesenteric fat of the small bowel wall is involved with a monomorphic population of small-tointermediate-sized irregular T lymphocytes. Cells were positive for CD2, CD3, and CD7, and negative for CD5. T cell gene rearrangement studies were positive (i.e., showed a clonal band indicating a clonal T-cell process). (Courtesy of Dr. Edward Lee, Washington, DC.)

has shown that loss of heterozygosity at chromosome 9q21 is a frequent finding.182 In addition, one study has suggested that gain of chromosome 1q may be an early event in the genesis of enteropathy-type intestinal T cell lymphoma lymphomagenesis.183

Pathology

Tumors typically occur in the jejunum but may occur in other sites of the small intestine. Lymphoma may occur in single or multiple sites. Grossly, the lymphomas commonly appear as ulcerating lesions, with circumferential involvement of the small bowel.160 Lesions may also appear as nodules, plaques, or strictures, but large masses are uncommon. Mesenteric lymph nodes are often enlarged, either because of tumor involvement or of edema and reactive changes. Distant sites, especially the bone marrow or the liver, are sometimes involved. Histologically, the lymphoma is generally characterized by large, highly pleomorphic cells with numerous, bizarre, multinucleated forms.184,185 However, cells sometimes have a more immunoblastic, anaplasti, or, less commonly, small cell appearance (Fig. 29-8). Eosinophils and other inflammatory cells are often present. Histologic features often vary within the same tumor and between multiple lesions of the same patient. Uninvolved mucosa usually has the typical appearance of celiac disease, with villous atrophy, crypt hyperplasia, plasmacytosis in the lamina propria, and an increase in intraepithelial lymphocytes. However, the enteropathy may be subtle in some cases, with only an increase in the intraepithelial lymphocytes. Immunophenotyping typically shows that the malignant cells are CD3+, CD7+, CD4−, CD5−, and contain cytotoxic granules recognized by the antibody TIA-1.173 Surface CD3 may be negative but cytoplasmic CD3e chains are usually detectable.174 Cells are CD8+ in a minority of cases and a small percentage of cases coexpress CD8 and CD56.186 CD103 is commonly positive. Cases with anaplastic morphology are CD30+. Genotypic studies have shown monoclonal T cell receptor gene rearrangements.180,187,188 As noted, monoclonal T cell populations can also be detected in mucosa not involved by lymphoma.

Patients may have a history of documented celiac disease, with the time to development of lymphoma varying widely. However, at least half of patients have celiac sprue diagnosed at the same time as the lymphoma. The most common symptoms at presentation are abdominal pain, weight loss, diarrhea, or vomiting. Less common symptoms may include fever, night sweats, and small bowel obstruction or perforation. It is rare for patients to have palpable abdominal masses or peripheral lymphadenopathy, but extraintestinal sites of involvement may include the liver, spleen, thyroid, skin, nasal sinus, and brain.158 In one series, 68% of patients had anemia and 25% of patients had elevations of the serum lactate dehydrogenase level.160 Hypoalbuminemia is a common finding at presentation. Diagnosis is usually made by endoscopic biopsies or fullthickness, laparoscopic small bowel biopsies, but approximately 20% of patients have been diagnosed with barium studies and small bowel biopsies. Traditionally, patients were staged with CT and bone marrow biopsies, but 18F-FDG PET appears to be more sensitive and specific than CT in differentiating enteropathy-associated intestinal T cell lymphoma from refractory celiac disease.189 In the two largest recent clinical studies, the Ann Arbor staging system was used to stage patients (see Table 29-2). In these two combined studies, 9% of the patients had Ann Arbor stage IE disease, 62% had stage IIE disease, and 29% had stage IV disease.160

Treatment

No large controlled trials of therapy for enteropathy-type intestinal T cell lymphoma have been reported. Thus, standard treatment is not well defined. Typically, patients are treated with a combination of surgery and chemotherapy.160 Surgery involves removal of as much tumor as is feasible. Intensive chemotherapy is then administered, with the most common regimens being ones that contain anthracyclines such as CHOP. There is no evidence for superiority of any particular chemotherapy regimen. In one study, in which 24 patients received chemotherapy, small bowel perforation developed in 4 patients and in 3 of these the perforation occurred soon after receiving chemotherapy and was fatal.160 Nutritional status is commonly poor, requiring parenteral nutrition. Because of poor nutritional and performance status, less than 50% of patients are able to complete the prescribed treatment regimen. In one study, the overall response rate to chemotherapy was 58%, with a complete remission rate of 42%.160 However, relapse occurs at a median of six months from the time of diagnosis in approximately 80% of patients, usually in small bowel sites. Various salvage regimens have been tried for patients with relapsed disease, but few relapsed patients have survived. The actuarial one- and five-year survival rates in one study were 39% and 20%, respectively, with one- and five-year failurefree survival rates of only 19% and 3%, respectively. Another series from the United Kingdom found that survival at 2.5 years was only 13%.168 Thus, the prognosis for this lymphoma is poor. There are a few reports of patients treated with high-dose chemotherapy followed by autologous stem cell transplantation, usually with poor outcomes.190-192 Therapy with novel agents are needed. Alemtuzumab (Campath), an anti-CD52 monoclonal antibody, has been used to treat refractory celiac disease.193 Conceivably, earlier diagnosis may improve the outcome. The diagnosis should be considered for patients who present in midlife with celiac disease and for those who have clinical deterioration after having been stable on a gluten-free diet.

Chapter 29  Gastrointestinal Lymphomas UNCOMMON SMALL INTESTINAL LYMPHOMAS Natural Killer Type T Cell Intestinal Lymphoma

Extranodal NK T cell lymphoma, nasal type, is a distinct pathologic entity in the World Health Organization classification of hematolymphoid malignancies.96 Very rare cases of intestinal NK cell lymphomas have been described.125 Most of the cases reported have not involved patients with celiac sprue or sensitivity to gluten.123,124 Optimal management of this very rare disorder has not been determined. Most patients undergo resection, with some receiving adjuvant chemotherapy.122

OTHER SITES NHL less commonly occurs in other sites of the GI tract, including the oral pharynx, esophagus, liver, pancreas, biliary tree, colon, and rectum. Signs and symptoms reflect the site of presentation. Because of the relative rarity of these disorders, the literature is fairly limited. Therefore, definitive conclusions cannot be reached about the optimal management of these more unusual GI lymphomas. Standard principles of lymphoma management dictate diagnostic procedures, staging, prognostic assessment, and treatment. As is the case for all lymphomas, histology and stage guide treatment. Waldeyer’s ring lymphomas are usually diffuse large cell lymphomas, but other histologies may be present instead.194,195 Endoscopy and imaging of the remainder of the GI tract should be included in the staging workup, because lymphomatous involvement in other sites may accompany Waldeyer’s ring involvement. Ann Arbor stage I or II diffuse large cell lymphoma is managed with combined anthracycline-based chemotherapy and/or local radiotherapy.115,196 Primary hepatic lymphoma (PHL; see Chapter 35 and 94) is more common in men and has a median age of approximately 50 years.197,198 PHL can present as a single, large, multilobulated mass or as single or multiple nodules. The histology is usually diffuse large B cell, but MALT lymphoma (extranodal marginal B cell lymphoma) has been reported as well (see Table 35-2). Rare cases of T cell hepatic lymphoma have been reported. Diagnosis is usually by needle biopsy. Because of the rarity of the disease, optimal therapy is uncertain. Long-term disease-free survival has been reported after resection, but multiagent chemotherapy is probably most appropriate for diffuse large B cell lymphoma. Less aggressive chemotherapy may be appropriate for lymphomas with marginal zone histology. The antiCD20 monoclonal antibody, rituximab, has activity in B cell lymphomas and thus may have a role in the treatment of these diseases. An association of hepatitis C and hepatic and splenic marginal zone lymphoma has been established, and response of the lymphoma to hepatitis C treatment has been documented199,200; whether there may be an association of other hepatitis viruses and hepatic lymphomas is unknown. As discussed in Chapter 60, the pancreas is a rare site of lymphoma.201-203 Patients have a clinical presentation similar to that of pancreatic adenocarcinoma, with abdominal pain and obstructive jaundice; chylous ascites has also been reported. Histology is usually diffuse large B cell, and therapy consists of anthracycline-containing combination chemotherapy (e.g., the CHOP regimen) with rituximab. Patients with biliary obstruction may require a biliary drainage procedure before being treated with chemotherapy to avoid excessive chemotherapy-related toxicity.

Primary colorectal lymphomas (see Chapter 123) most commonly involve the cecum,204,205 with high- or intermediate-grade histology. Most colorectal lymphomas are Ann Arbor stage IE or IIE. Again, therapy is dictated by histology and stage. Resection is the standard therapy, with adjuvant chemotherapy given for patients with aggressive histology.

IMMUNODEFICIENCY-RELATED LYMPHOMAS POST-TRANSPLANTATION LYMPHOPROLIFERATIVE DISORDERS

The post-transplantation lymphoproliferative disorders (PTLDs)206-208 complicate 0.8% to 20% of cases of those with solid organ transplants (see Chapter 34), with the incidence being highest in heart-lung transplant recipients. PTLDs are also seen in bone marrow transplant recipients, particularly in patients receiving T cell–depleted allografts. PTLD results from proliferation of Epstein-Barr virus (EBV)–transformed B cell clones that have developed in part because of immunosuppression. The histologic appearance of PTLD is highly variable, with lesions being polymorphic or monomorphic; the histology may represent infectious mononucleosis, aggressive NHL, or plasmacytoma.209 Lesions may be polyclonal, oligoclonal, or monoclonal. The clinical presen­ tation also varies greatly, with some patients having a syndrome resembling infectious mononucleosis and some having a more lymphoma-like presentation, with nodal or extranodal disease. Involvement of extranodal areas is common, with the GI tract being a common site. The literature regarding the treatment of PTLD suffers from a lack of prospective trials and lack of standardized histologic classification. The treatment approach varies, but usually consists initially of withdrawal of immunosuppression210; anthracycline-based chemotherapy is then reserved for patients who fail to respond to this first maneuver. Surgical or radiation therapy may cure patients with localized disease. Other treatments have included acyclovir or gan­ ciclovir directed at EBV and interferon-α. Monoclonal antibodies to B cells such as rituximab have activity in this disorder, and donor leukocyte infusions are frequently used for patients with PTLD that develops after allogeneic bone marrow transplantation.211

HUMAN IMMUNODEFICIENCY VIRUS– ASSOCIATED NON-HODGKIN’S LYMPHOMA

The risk of developing NHL is markedly increased in patients with HIV (see Chapter 33), and development of lymphoma is considered an AIDS-defining condition. These malignancies are B cell neoplasms,212,213 with most cases having small noncleaved cell or diffuse large cell histology. EBV is implicated in about half of non–central nervous system HIV-related lymphomas. HIV-associated NHL typically has an aggressive presentation, with rapidly growing disease and prominent B symptoms. The GI tact is a common site, including unusual sites such as the anus and rectum. Historically, chemotherapy has been poorly tolerated and lower dose chemotherapy regimens have been used.214,215 However, patients with higher CD4+ T cell counts (as is more commonly seen in view of the standard current usage of highly active antiretroviral therapy) may be more able to tolerate full-dose chemotherapy regimens and may have a better prognosis than has been seen in previous studies.216

459

460

Section IV  Topics Involving Multiple Organs Primary effusion lymphoma (PEL) is a rather recently described clinicopathologic entity associated with the herpesvirus HHV-8 (Kaposi’s sarcoma–associated virus).217,218 Histology shows a distinctive morphology that bridges large cell immunoblastic lymphoma and anaplastic large cell lymphoma.219 Tumor cells show monoclonal immunoglobulin gene rearrangements but typically lack B cell–associated antigens. HHV-8 is detectable by PCR assay. Patients are usually HIV-positive, but the syndrome has been reported in HIV-negative patients. Patients present with malignant effusions in the pleural or peritoneal cavity, which remain localized to the body cavity of origin. Disease progression is rapid, with survival of only a few weeks to months. Optimal therapy has not been defined.

KEY REFERENCES

Bertoni F, Zucca E. State-of-the-art therapeutics: Marginal-zone lymphoma. J Clin Oncol 2005; 23:6415-20. (Ref 57.) Ferrucci PF, Zucca E. Primary gastric lymphoma pathogenesis and treatment: What has changed over the past 10 years? Br J Haematol 2007; 136:521-38. (Ref 14.) Gale J, Simmonds PD, Mead GM, et al. Enteropathy-type intestinal T-cell lymphoma: Clinical features and treatment of 31 patients in a single center. J Clin Oncol 2000; 18:795-803. (Ref 160.) Koch P, del Valle F, Berdel WE, et al. Primary gastrointestinal nonHodgkin’s lymphoma: I. Anatomic and histologic distribution, clinical features, and survival data of 371 patients registered in the German Multicenter Study GIT NHL 01/92. J Clin Oncol 2001; 19:3861-73. (Ref 15.) Koch P, del Valle F, Berdel WE, et al. Primary gastrointestinal nonHodgkin’s lymphoma: II. Combined surgical and conservative or conservative management only in localized gastric lymphoma—results

of the prospective German Multicenter Study GIT NHL 01/92. J Clin Oncol 2001; 19:3874-83. (Ref 109.) Koch P, Probst A, Berdel WE, et al. Treatment results in localized primary gastric lymphoma: Data of patients registered within the German multicenter study (GIT NHL 02/96). J Clin Oncol 2005; 23:7050-9. (Ref 110.) Lecuit M, Abachin E, Martin A, et al. Immunoproliferative small intestinal disease associated with Campylobacter jejuni. N Engl J Med 2004; 350:239-48. (Ref 141.) Liu H, Ye H, Ruskone-Fourmestraux A, et al. T(11; 18) is a marker for all stage gastric MALT lymphomas that will not respond to H. pylori eradication. Gastroenterology 2002; 122:1286-94. (Ref 42.) Nagai S, Mimuro H, Yamada T, et al. Role of Peyer’s patches in the induction of Helicobacter pylori–induced gastritis. Proc Natl Acad Sci U S A 2007; 104:8971-6. (Ref 19.) Parsonnet J, Hansen S, Rodriguez L, et al. Helicobacter pylori infection and gastric lymphoma. N Engl J Med 1994; 330:1267-71. (Ref 26.) Romaguera JE, Medeiros LJ, Hagemeister FB, et al. Frequency of gastrointestinal involvement and its clinical significance in mantle cell lymphoma. Cancer 2003; 97:586-91. (Ref 128.) Wotherspoon AC, Doglioni C, Diss TC, et al. Regression of primary lowgrade B-cell gastric lymphoma of mucosa-associated lymphoid tissue type after eradication of Helicobacter pylori. Lancet 1993; 342:575-7. (Ref 20.) Wundisch T, Thiede C, Morgner A, et al. Long-term follow-up of gastric MALT lymphoma after Helicobacter pylori eradication. J Clin Oncol 2005; 23:8018-24. (Ref 32.) Zucca E, Bertoni F, Roggero E, Cavalli F. The gastric marginal zone B-cell lymphoma of MALT type. Blood 2000; 96:410-19. (Ref 18.) Zucca E, Bertoni F, Roggero E, et al. Molecular analysis of the progression from Helicobacter pylori–associated chronic gastritis to mucosaassociated lymphoid-tissue lymphoma of the stomach. N Engl J Med 1998; 338:804-10. (Ref 24.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

30 Gastrointestinal Stromal Tumors (GISTs) George D. Demetri

CHAPTER OUTLINE Incidence  461 Location  461 Clinical Features  462 Pathology and Molecular Pathobiology  462 Unique Clinicopathologic Subtype of Sarcomas (Mesenchymal Neoplasia)  462 Uncontrolled Kinase Activation: Primary Molecular Pathogenesis  462 Benign Gastrointestinal Stromal Tumors  464 Diagnosis  464 Diagnostic Approach to Stromal Tumors by Site  465 Imaging Studies  466

Recent research advances have greatly elucidated the pathogenesis of gastrointestinal stromal tumors (GISTs) at the molecular level, and translation of this understanding into highly effective, molecularly targeted therapies has followed in rapid succession. GISTs has evolved rapidly from an arcane disease known only to a few investigative teams worldwide into a proof of concept model for cancer medicine, personalizing highly selective therapies to the dysregulated cell signaling that contributes to the abnormal behavior of these neoplastic cells.

INCIDENCE It is difficult to obtain accurate data regarding the true incidence of GISTs. This is because of referral bias, which concentrates GIST cases with a worse prognosis and a more malignant behavior in academic cancer centers, and also the lack of definitive diagnostic techniques before the molecular definitions of GISTs in 1998 and later. Before 2000, the number of new GIST cases in the United States had been underestimated and underreported. However, it is now recognized that the true incidence of GISTs is much higher than previously believed. Much of this is to the result of an increased recognition of GISTs; new molecularly targeted drugs, such as imatinib mesylate (Gleevec), have become available. A population-based study to assess the incidence of GISTs using the most up-to-date criteria has estimated an annual incidence of approximately 15 cases/million.1 This has increased the estimated incidence of GISTs in the United States alone to approximately 5000 new cases/year.2,3 However, even this figure may underestimate the actual incidence because many smaller

Role of Biopsy  468 Differential Diagnosis  468 Treatment  468 Advanced Gastrointestinal Stromal Tumors: Metastatic, Surgically Unresectable, or Recurrent Disease  468 Primary Localized Disease  472 Special Considerations  473 Familial Disease  473 Pediatric Patients  473 Relationship to Other Genetic Syndromes Predisposing to Neoplasms  473

lesions, including those detected incidentally by many gastroenterologists on endoscopy performed for other reasons, are almost certainly not revealed in these incidence figures. Not all GIST cases will prove to be life-threatening, because many patients who present with GISTs have limited disease with lesions that are small (smaller than 1 cm) and fully curable with optimal surgery as first-line therapy. It remains a question as to whether there are any true GIST lesions that might be able to be observed without resection. GISTs occur predominantly in adults, with a median incidence in the fifth decade of life, but these tumors can occur across the age spectrum, from infants to older adults. The incidence has been reported to be slightly higher in men than women. Recent studies of surgical or autopsy specimens of stomach samples resected for non-neoplastic disease have documented a remarkably high incidence of occult microscopic GIST lesions, in the range of 20% to 35%.4,5

LOCATION Most GISTs (60% to 70%) arise in the stomach, 20% to 30% originate in the small intestine, and less than 10% in the esophagus, colon, and rectum (see later). GISTs can also occur in extraintestinal sites in the abdomen or pelvis such as the omentum, mesentery, or retroperitoneum.6-8 However, data on sites of GIST origin are somewhat limited. Large studies before 1999 may have included other subtypes of mesenchymal neoplasms because these series were not selected by immunohistochemical or molecular genetic markers specific for true GISTs.

461

462

Section IV  Topics Involving Multiple Organs CLINICAL FEATURES The clinical presentation of patients with GISTs can vary tremendously on the basis of the anatomic location of the primary lesion, as well as other factors, such as tumor size and presence or absence of symptomatic metastases. For most GIST patients, the initial detection of GISTs may be an incidental finding or result from evaluation of nonspecific symptoms. Symptoms from GISTs are noted usually only after tumors have reached a large size or have impinged on a specific anatomic region (e.g., constricting gastric outflow). Most symptomatic patients present with tumors larger than 5 cm in maximum dimension. Symptoms at presentation may include a palpable abdominal mass or swelling, abdominal pain, nausea, vomiting, anorexia, and early satiety. Although it has been reported that up to 40% of GIST patients present with acute hemorrhage into the gastrointestinal (GI) tract or peritoneal cavity from tumor rupture, it is clear that such reports are dependent on referral bias of patients with large GIST lesions or multifocal disease. The vast majority of GIST metastases at presentation are intra-abdominal, with metastases to the liver, omentum, or peritoneal cavity.6 Metastatic spread to lymph nodes and to other regions via lymphatics is rare; most lesions thought to be nodal metastases by imaging studies simply represent metastatic deposits of tumor nodules in the omentum or peritoneum rather than true lymphatic spread of the disease.

PATHOLOGY AND MOLECULAR PATHOBIOLOGY UNIQUE CLINICOPATHOLOGIC SUBTYPE OF SARCOMAS (MESENCHYMAL NEOPLASIA)

Although relatively rare compared with epithelial cancers, GISTs represent a form of sarcoma that comprises approximately 1% to 3% of all malignant GI tumors. As noted, the diagnostic criteria for GISTs before 1999 were somewhat subjective, controversial, and perhaps even a bit confusing, with a great deal of room for interobserver variation. The term GIST was initially a purely descriptive term applied by Mazur and Clark in 1983 to define intra-abdominal tumors that were not carcinomas (i.e., nonepithelial) and that did not exhibit features of smooth muscle or nerve cells.9 The morphology of the cells was the dominant feature driving the diagnostic scheme. However, pathologists subsequently recognized that there was no completely clear differential expression of muscle or nerve antigenic markers when careful immunohistochemical analyses were performed on samples identified as GIST lesions by cellular morphology. The expression of differentiation antigens used as markers for muscle cells (e.g., smooth muscle actin) and nerve cells (e.g., S100) was noted to vary widely in GI mesenchymal lesions, leading to interesting hypotheses about whether GIST lesions from different patients were attempting to recapitulate distinct myogenic or neural programs of differentiation. To accommodate these empirical observations, it was proposed that approximately one third of GIST lesions differentiated along smooth muscle lineages, another third were neurogenic in origin, and the final third lacked any detectable lineage-specific markers (null phenotype) by immunohistochemical analysis.10-12 Nomenclature for GISTs had also been confusingly complex before the availability of molecular- and

mechanism-based diagnostic tests. Because of limited understanding of the disease before 1999, GISTs were most commonly diagnosed as leiomyomas or leiomyosarcomas because of the histologic resemblance to these smooth muscle neoplasms. Other terms that had often been applied to a GIST included benign leiomyoblastoma and, recognizing some of the neural characteristics, plexosarcomas13 or gastrointestinal autonomic nerve tumors (GANTs).14 All these terms are now recognized as referring to GISTs. Insightful studies by several pathology groups have noted that the panoply of tumors lumped together as smooth muscle tumors of the GI tract were likely not simply leiomyosarcomas nor benign leiomyomas; a subset of these tumors originating in the bowel wall had several unique histologic features, probably representing a totally different diagnostic group altogether.15,16 Additionally, clinical oncologists had noted that putative leiomyosarcomas of the GI tract had a completely different prognosis than true leiomyosarcomas arising in other parts of the body (e.g., the uterus). The leiomyosarcomas of the GI tract were extraordinarily resistant to standard chemotherapy regimens, further supporting the notion that these lesions represent a different form of cancer entirely. Immunohistochemical analysis of GISTs in the early 1990s attempted to find specific markers that might distinguish GISTs from other spindle cell tumors of the GI tract, such as schwannomas and sarcomatoid carcinomas. There was some initial enthusiasm for the CD34 antigen as such a marker; however, this antigen is also expressed by hematopoietic stem cells and by vascular and myofibro­ blastic cells. Also, the sensitivity and specificity of CD34 are low because only approximately half of GIST cases express CD34, and other smooth muscle, myofibroblastic (e.g., desmoid), or Schwann cell tumors can also express CD34. Therefore, CD34 is not a reliable marker to distinguish true GISTs from other neoplasms.17,18 It is clear that before approximately 1999, there were no objective, reproducible, and clearly defined criteria for the diagnosis and classification of GISTs, and it is likely that several types of epithelioid and spindle cell tumors were included in the clinical diagnostic category of GISTs. Similarly, many true GIST cases had various different diagnostic labels, such as leiomyoblastomas and GANTs. This makes the interpretation of published GIST series accumulated before the year 2000 difficult, given the heterogeneity that was reflected by the diagnostic term GIST before the widespread use of specific kinase-directed diagnostic and molecular markers for this disease.

UNCONTROLLED KINASE ACTIVATION: PRIMARY MOLECULAR PATHOGENESIS

A critical advance in the understanding of GISTs at a mole­ cular level occurred in the late 1990s, with the recognition that these tumors exhibited some histopathologic similarities with the pacemaker cells of the gut known as the interstitial cells of Cajal (ICCs).19 ICCs are normally present in the myenteric plexus and serve to coordinate gut peristalsis by assisting the linkage of smooth muscle cells of the bowel wall with the autonomic nervous system (see Chapters 96 to 98). GIST cells and ICCs have certain ultrastructural features in common, such as the combination of neural and muscle phenotypes. The molecular pathogenesis of GISTs was advanced further by a key observation made by Hirota and colleagues20 in 1998. This group was studying the role played by the KIT receptor tyrosine kinase (RTK) in cell growth and development. Expression of the KIT RTK can be detected by immu-

Chapter 30  Gastrointestinal Stromal Tumors (GISTs)

A

B

Figure 30-1.  A, High-power photomicrograph of a typical spindle cell gastrointestinal stromal tumor (GIST). The cells are monomorphic, have abundant pale, eosinophilic, fibrillary cytoplasm, and lack mitotic activity. B, KIT immunostaining. This medium-power photomicrograph of a spindle cell GIST shows diffuse and strong cytoplasmic immunoreactivity for KIT. The entrapped muscle fibers from the bowel wall are negative by CD117 immunostaining for KIT (A, Hematoxylin and eosin, ×100; B, CD117 immunostain, ×100). (Courtesy of Dr. Brian P. Rubin, Cleveland, Ohio.)

nohistochemical staining for the CD117 antigen as a marker of the KIT protein. In normal cell signaling, KIT binds its ligand, known as stem cell factor (SCF) or Steel factor; ligand binding brings together two molecules of KIT with subsequent activation of a signaling cascade. A homodimeric complex is formed with these two KIT receptors, leading to cross-phosphorylation of critical tyrosine residues in the intracellular domains of KIT, which activate signal transduction pathways downstream of KIT. The net physiologic effect of normal ligand-induced KIT activation is the controlled stimulation of cell proliferation and enhanced cell survival; therefore, uncontrolled activation could theoretically lead to neoplastic growth and transformation of cells. Hirota and colleagues20 recognized this potential mechanism and provided the critical confirmation of this theory at the cellular and molecular levels. This same team, in a project led by Nishida, also expanded these observations to familial GISTs, with germline carriage of an activating mutation in the KIT gene encoding the RTK.21 This elegant work supported some of the key biologic similarities between GIST and ICC cells22-25 because both cell types had been shown to express the KIT RTK. It is likely that the cells of GISTs and normal ICCs share a common precursor cell.26,27 The KIT RTK and SCF play essential roles in the development and maintenance of normal ICCs, as well as other cells, including melanocytes, erythrocytes, germ cells, and mast cells. KIT expression is noted in the vast majority (>95%) of GISTs, but KIT is not expressed by true smooth muscle tumors of the GI tract nor by stromal tumors at other anatomic locations, such as endometrial stromal tumors. Although the origin of the neoplastic cells of GISTs remains a matter of active investigation, some data suggest that GISTs originate from CD34-positive stem cells residing within the wall of the gut, which can then differentiate incompletely toward the ICC phenotype.27-29 GISTs characteristically exhibit expression of CD117 by immunohistochemical assays (>95% of lesions but, importantly, not 100%, because there are true KIT-negative GIST cases).30 Levels of expression of CD117 (KIT) are generally diffuse and strong in the spindle cell GIST subtype (Fig. 30-1). In contrast, in the epithelioid subtype, CD117 expression is typically focal and weakly positive in a dot-like pattern (Fig. 30-2). As noted, CD34 expression is neither sensitive nor specific for GIST because this antigen can also

Figure 30-2.  High-power view of an epithelioid gastrointestinal stromal tumor (GIST) showing epithelioid cytomorphology, fibrillary cytoplasm, and a lack of mitotic activity (Hematoxylin and eosin, ×200). (Courtesy of Dr. Brian P. Rubin, Cleveland, Ohio.)

be noted in desmoid tumors, and approximately 30% to 40% of GIST lesions are negative for CD34.21,26 True leiomyosarcomas express two smooth muscle markers, smooth muscle actin (SMA) and desmin, but fail to express CD117. Schwannomas are usually positive for the neural antigen S100 but are also negative for CD117. Normal mast cells and ICCs in the surrounding stromal tissues serve as ideal positive internal controls because these normal cells strongly express CD117. Activating mutations in the KIT gene were identified in five of six cases of human GISTs originally analyzed by Hirota and colleagues,20 with evidence that the mutations resulted in uncontrolled, ligand-independent activation of the KIT kinase. Genetically engineered cells harboring the mutant overactive KIT proteins were tumorigenic in nude mice, serving as proof of concept that the malignant phenotype was directly induced by the aberrant signaling pathways associated with uncontrolled KIT kinase activation. The oncogenic potential of mutant, uncontrollably active KIT in the pathogenesis of GISTs in humans has also been

463

464

Section IV  Topics Involving Multiple Organs supported by the identification of familial syndromes with an autosomal dominant inheritance pattern and an abnormally high incidence of GISTs, usually occurring as multiple foci within any affected individual (see later).21,31,32 Genetic analysis of such kindreds reveals that they harbor germline-activating KIT mutations, similar to the mutations that were first described in sporadic cases of GISTs. With analyses using optimal tumor specimens and sophisticated technology, it has become clear that KIT mutations can be detected in more than 90% of GIST cells.33-35 Constitutive activation of the KIT kinase enzymatic function has been reported to characterize every GIST sample analyzed by immunoblotting technique, even in cases in which there are no detectable mutations in the KIT gene. The mechanisms whereby nonmutated (wild-type) KIT is maintained in an uncontrollably phosphorylated state are poorly understood and are likely to be fertile ground for future research, with therapeutic potential. Importantly, the vast majority of GIST cells at initial presentation demonstrate only a single site of mutation in the KIT gene; complex genetic changes in untreated GISTs at initial presentation are vanishingly rare. Gain of function mutations have been identified most commonly in exon 11 of KIT (up to 70% of cases), an exon that encodes the intracellular juxtamembrane domain of the KIT protein. Mutations in the KIT gene locus have also been described in other regions including (in decreasing order of prevalence) exon 9 (the KIT extracellular domain), exon 13 (kinase domain), and exon 17 (kinase domain).33-35 Structural biology studies have revealed the mechanism whereby normal (wild-type) KIT is kept in an autoinhibited con­ formation until ligand binds; mutational changes in conformation interfere with this autoinhibition and lead to a structural basis for the aberrant activation of the KIT kinase function.36,37 Another key advance in the understanding of GISTs has been the recognition that signaling through other uncontrolled kinases in addition to KIT could drive the neoplastic phenotype of GIST cells. Specifically, it is now recognized that approximately 5% of GIST cells are not through activation and aberrant signaling of the KIT receptor, but rather through mutational activation of the structurally related kinase known as the platelet-derived growth factor receptoralpha (PDGFRA).38,39 The definitive diagnostic criteria of uncommon CD117negative lesions that are nevertheless truly GISTs are currently somewhat obscure. GIST lesions can be heterogeneous in the expression of CD117, even within a single mass. It is possible that a needle biopsy could yield cells consistent with a GIST yet be CD117-negative simply by sampling bias alone. However, expert pathologists can also define a rare subset of GISTs (<5% of cases overall) that have no CD117 expression; these are most likely dependent on an alternative kinase such as PDGFRA.38 Molecular analyses of the KIT and PDGFRA genotypes may be useful to define with certainty the group of rare patients with CD117-negative GISTs in the future. Now, it is important to note that the diagnosis of a GIST should be made on the grounds of morphologic, clinicopathologic, and immunohistochemical data, as well as possibly molecular analysis if there is any ambiguity from the other pathologic assessments. KIT mutations have also been documented in small GISTs (<1 cm in greatest dimension)40; such lesions are most often detected incidentally (e.g., during upper endoscopy for reflux symptoms) and may appear morphologically benign. These findings support the hypothesis that activating mutations in the KIT proto-oncogene represent an early event in

the transformation from a normal precursor cell into a GIST lesion. Because lesions in familial GISTs (see later) may not present clinically until the second or third decade of life, or even much later, it is likely that second hits are necessary to attain a more aggressive malignant phenotype. The other key signaling steps that confer a more malignant phenotype to GIST cells remain obscure. However, the unique aspects of the signaling cascades in GISTs are being actively elucidated, and these appear to differ from KIT signaling in hematologic cancers. For example, the STAT5 pathway of leukemic cells is not typically activated in GISTs, whereas STAT1 and STAT3 are activated at a high level.33

BENIGN GASTROINTESTINAL STROMAL TUMORS

The literature before 2000 was somewhat confusing about whether mutational status of KIT could distinguish between so-called benign and malignant GISTs. With the recognition that KIT mutations can be found in even the smallest GIST,40 there is now consensus that KIT genotype alone cannot account for differences between GISTs that may behave in an indolent manner (and which, when small, may be curable by optimal resection alone) versus those that are clearly aggressive and malignant by all functional definitions. It is important to note that a well-differentiated benign cell morphology alone should not provide any reassurance that an individual GIST lesion will pursue a benign clinical course. Consensus was reached at a meeting held at the National Cancer Institute (NCI) among pathologists with expertise in GISTs. This consensus defined the two most reliable prognostic factors for behavior of a primary GIST as the size of the primary tumor and the number of mitoses, reflecting the proliferative activity of the cells.2 Other factors, such as the specific histologic subtype (epithelioid vs. spindle cell; see Figs. 30-1A and 30-2), the degree of cellular pleomorphism, and patient age may have some contribution to prognosis but are most likely to play a minor role in determining the clinical outcome. Recurrence and survival rates have also been reported to correlate with the location of the primary GIST lesion, with small bowel tumors showing a somewhat worse prognosis. Most important, the consensus panel emphasized that no GIST lesion should ever be considered completely benign. This is because morphologically benign lesions have a finite capacity to recur and metastasize, occasionally several years after initial presentation of the primary disease. Nonetheless, further research is necessary to define the true risks of tiny GIST lesions (<1 cm) because these are far more common than previously appreciated.

DIAGNOSIS The revolutionary changes in the scientific understanding, diagnostic evaluation, and therapeutic management of GIST patients (discussed later) has led several professional organizations to develop consensus-based (and, whenever possible, evidence-based) clinical practice guidelines. The National Comprehensive Cancer Network (NCCN) has developed extensive publically accessible guidelines to assist clinicians in the management of GIST patients (see the GIST and Soft Tissue Sarcoma Guidelines on www. nccn.org). Also, the European Society of Medical Oncology (ESMO) has published expert-driven clinical practice guidelines.41 Given the rapid progress in this field, these guidelines are reviewed at least annually to ensure that current information is available to assist clinicians in pro-

Chapter 30  Gastrointestinal Stromal Tumors (GISTs) viding the most up-to-date care and accurate information to patients. The diagnostic evaluation of a suspected or proven GIST is similar to that of other GI neoplasms. The most important element is to keep GISTs in the suspected differential diagnosis of any mass lesion noted throughout the length of the GI tract, as well as in extraintestinal sites of the abdomen and pelvis. As for any GI evaluation, the site of lesion may determine which diagnostic tools are most appropriate for the patient.

DIAGNOSTIC APPROACH TO STROMAL TUMORS BY SITE Esophageal Tumors

GIST lesions may arise within the esophagus, although this is a rare presentation for larger lesions. Most esophageal GIST lesions are noted incidentally during upper endoscopy performed for some other unrelated symptom or disorder, such as reflux esophagitis. Esophageal GIST lesions may be small (only a few millimeters in size) in this location and may be resected using endoscopic techniques.42 Margins may be involved if a lesion, unsuspected as a GIST and thought to be benign, is simply popped out using an endoscopic procedure. It remains unclear whether watchful waiting with serial endoscopic follow-up is appropriate for any patient with small GIST lesions (<1 cm in maximal dimension). As noted, histopathology showing putatively benign GIST cells cannot be viewed reassuringly because histology does not perfectly predict the malignant behavior of GISTs. A careful risk-based assessment that takes into account other aspects of the tumor, as well as patient-specific factors (e.g., age, comorbidities, patient preferences), must be performed. It will be important for prospective research to accumulate a larger objective database regarding the outcomes of such patients so that medical decision making can be based on solid evidence. Although a small esophageal GIST lesion is probably a single primary, it is appropriate to take a careful family history and perform computed tomography (CT) scanning of the abdomen and pelvis to ensure that no other lesions are present in the patient’s family, especially if the disease is first detected in a young person. Some series have attempted to distinguish between clinical outcomes of patients with GISTs and those with leiomyosarcomas or other mesenchymal neoplasms, such as leiomyomas.43

Gastric Tumors

The most common primary site for GISTs is the stomach. Most GIST lesions are submucosal, rather than mucosal, without overlying ulceration (Fig. 30-3). This explains why many GIST masses may only be visualized on endoscopy as a subtle, smooth protrusion with overly normal mucosa.44 Also, this submucosal localization can make diagnostic biopsy through an endoscope difficult. It is not uncommon for superficial biopsies to reveal only normal mucosa, whereas deeper biopsies or histopathology from a definitive resection would show the true underlying GIST cells.

Small Intestinal Tumors

The second most common site for GISTs to arise is in the small intestine. These lesions have been reported to have a worse prognosis overall, for unclear reasons. One possibility is that small intestinal lesions simply remain asymptomatic for a longer time before the disease is brought to medical attention and detected. If the differences were solely because of this lead time bias, small intestinal GISTs should present

Figure 30-3.  Spindle cell gastrointestinal stromal tumor (GIST) of the stomach. The lesion is well circumscribed and does not invade the muscularis mucosa. Invasion of the muscularis mucosa is considered an adverse prognostic factor (Hematoxylin and eosin, ×50). (Courtesy of Dr. Brian P. Rubin, Cleveland, Ohio.)

with significantly larger lesions than other primary sites, which is often the case. GIST masses in the small intestine tend to arise in the jejunum, followed by the ileum and the duodenum. The large lesions may be highly vascularized and present significant risks of bleeding, even with only a biopsy. Because complete surgical resection is the treatment of choice for localized GISTs, there is some controversy over whether any preoperative biopsy is necessary or whether that simply represents an extra risk for the patient. This is a challenging subject because other disease entities enter into the differential diagnosis of a large abdominal mass involving the small intestine and mesentery. In general, clinical practice guidelines have suggested that resection may be performed without antecedent biopsy if a GIST is strongly suspected and if surgery can be accomplished without significant risk of morbidity to the patient. If only radical surgery leading to significant functional impairment could remove the lesion, it may be in the patient’s best interest to consider a preoperative biopsy to establish the diagnosis of a GIST. The prognosis of GISTs involving the small intestine is related to the adequacy of resection.6,45 In one series of 50 GIST patients involving the small intestine, 70% could be completely resected, with median overall survival longer than five years for patients with localized or locally advanced disease. Patients who underwent complete resections exhibited a five-year overall survival rate of 42%, while those whose lesions could not be completely resected had a five-year overall survival rate of only 8%.28

Colorectal Tumors

GIST lesions are rare in the colon and rectum, accounting for approximately 5% of GIST cases. Perirectal and rectal GISTs present unique management challenges. As in other sites, small GIST lesions in the rectum may present as small, hard nodules less than 1 cm in diameter found incidentally

465

466

Section IV  Topics Involving Multiple Organs during a clinical examination. However, much larger tumors can ulcerate, perhaps mimicking a rectal adenocarcinoma. Diagnosis can be challenging because the epithelioid or mixed cell variants of GISTs can also be misclassified as adenocarcinomas, especially if small biopsies are confounded by severe inflammatory changes or associated abscess formation. GIST lesions in this location present similar challenges to those in the small intestine; specifically, if nonmutilating surgery can achieve negative margins for a tiny GIST (<1 cm), that is probably a reasonable way to proceed. However, many GIST lesions in the rectal region will prove more challenging to resect, and surgery may only be feasible with significant functional morbidity. In such cases, it may be best to consider the patient unresectable without unacceptable risk and to consider preoperative therapy with imatinib to decrease the tumor bulk, followed by complete resection after maximal tumor shrinkage has occurred. In limited data available from older series, recurrence of rectal GIST has been common, with relatively poor prognosis. It remains to be seen whether newer approaches using multidisciplinary therapy, including imatinib, might improve these outcomes for patients with better functional results.

A

IMAGING STUDIES Endoscopic Ultrasonography

Endoscopic ultrasonography (EUS) is a useful technology for evaluating possible GIST lesions because of their submucosal localization. The paradigm of a GIST lesion visualized by EUS is that of a hypoechoic mass contiguous with the fourth (muscularis propria) or second (muscularis mucosae) layers of the normal gut wall. In one study,46 the features most predictive of so-called benign GI tumors were regular margins, tumor size 3 cm or smaller, and a homogeneous echogenicity pattern. Multivariant analysis identified the presence of cystic spaces and irregular margins as independent predictors of malignant potential. A second study identified tumor size larger than 4 cm, irregular extraluminal borders, echogenic foci larger than 3 mm, and cystic spaces larger than 4 mm as factors that correlated with malignant behavior in GIST.47

Computed Tomography and Magnetic Resonance Imaging

CT is the most effective way to image primary lesions in the stomach because the oral contrast will outline masses and gastric thickening (Fig. 30-4). Differential diagnosis with inadequate gastric distention can be a challenge, especially when monitoring for recurrence following surgery. CT is also essential to stage the extent of disease completely and accurately. For measurable GIST lesions, it is particularly useful to perform CT with noncontrast image acquisition, as well as assess early and late images following the administration of intravenous contrast. Magnetic resonance imaging (MRI) is inferior to CT for visualizing a gastric GIST because of uncontrollable movement of the gastric wall and surrounding tissues. MRI scanning can be useful for the assessment of liver metastases because some GIST lesions can be fully isodense to normal tissues and thus invisible against surrounding hepatic parenchyma on CT scans. Baseline imaging is critical for GIST, because endoscopic imaging alone may only reveal a small fraction of the underlying tumor. Additionally, imaging patterns (particularly tumor density on CT imaging) can be interpreted qualitatively to assess the impact of targeted therapy using metrics

B Figure 30-4.  A, Computed tomograph scan showing a 4- to 5-cm exophytic gastric gastro­intestinal stromal tumor (GIST) arising from the greater curvature of the stomach (arrow). B, Gross photograph of the 3.5 × 4.5 × 4 cm tumor after it was resected and cut open. Histology showed a spindle cell GIST that was positive for KIT immunoreactivity. (Courtesy of Dr. Jay N. Yepuri and Dr. Christopher Bell, Dallas, Tex.)

other than tumor size. Tumors that become more hypodense on CT imaging have a more favorable antineoplastic response to targeted therapies, such as the kinase inhibitors imatinib or sunitinib.48

Positron Emission Tomography

One of the most impressive aspects of GIST diagnostic imaging is the use of 18F-fluorodeoxyglucose (18F-FDG)– positron emission tomography (PET) scans to add complementary information to that obtained by conventional anatomic imaging (Fig. 30-5). Although CT or MRI scanning can assess the size of GIST lesions accurately, the functional imaging of GIST with 18F-FDG–PET can provide useful additional information that can assist clinicians in the management of GIST patients. The actual mechanisms responsible for the high-level avidity of GISTs for the 18F-FDG tracer used most commonly in PET imaging are not yet known; however, it is likely that there is a direct connection between signaling through the overactive KIT RTK and glucose transport proteins. In this way, one could explain the rapid changes in PET imaging associated with inhibition of KIT signaling by pharmacologic means.49,50 Large GIST lesions

Chapter 30  Gastrointestinal Stromal Tumors (GISTs)

A

B

C

Figure 30-5.  Positron emission tomography (A, B) and computed tomography scans (C) scans in a patient with a gastrointestinal stromal tumor metastatic to the liver, before (left) and after (right) treatment with imatinib mesylate. (Courtesy of Dr. A. Van den Abbeele, Boston, Mass.; and modified from Demetri GD, Benjamin RS, Blanke CD, et al. NCCN Task Force report: management of patients with gastrointestinal stromal tumor [GIST]—update of the NCCN clinical practice guidelines. J Natl Compr Canc Netw 2007; 5[Suppl 2]:S1-S29).

467

468

Section IV  Topics Involving Multiple Organs can demonstrate centers with predominantly cystic or low attenuation characteristics noted on CT or MRI scans. It is clear by 18F-FDG–PET scans that the internal mass of large GIST lesions can often be viewed as metabolically quiescent, likely because of the endogenous necrosis of large lesions in their central portions. Even though GIST lesions can be vascular, the internal portion can nonetheless represent a confluent mass of necrotic material, with the more viable aspects of the GIST pushing out toward the edges of the lesion. Much of the added value of 18F-FDG–PET imaging in serial imaging can also be obtained by qualitative assessment of tumor density obtained via CT imaging.48 However, occasionally, metastatic GIST lesions in the omentum can be subtle and easy to overlook on CT scans because small lesions could blend into the folds of the bowel walls and be difficult for even the most experienced radiologist to detect. 18F-FDG–PET imaging can detect lesions approximately 1 cm or larger in size without difficulty because neither the normal bowel nor omentum will take up the 18F-FDG tracer with excess avidity.

ROLE OF BIOPSY

GIST lesions can be highly vascularized, which may present an unacceptable risk for biopsy. Additionally, percutaneous biopsies, even if only using fine-needle aspiration techniques, may impose a risk of tumor rupture, tumor cell seeding along the biopsy tract, or spreading tumor cells via peritoneal or mesenteric contamination. Therefore, many surgeons recommend that preoperative biopsy not be performed if resection is planned to minimize risk to the patient. However, biopsy must be performed in cases of unresectable GIST to make the diagnosis and justify pre­ operative administration of imatinib therapy (see later).

DIFFERENTIAL DIAGNOSIS

GIST was originally described as a monomorphic spindle cell neoplasm. However, it is now clear that GIST can exhibit a wide variety of histologic appearances, ranging from an epithelioid form with large round cells (see Fig. 30-2) to the spindle cell form (see Fig. 30-1), and even a subset of GIST lesions with mixed histology. The spindle cell GIST variant is far more common, representing approximately 70% of cases. The epithelioid, or round cell, pattern represents most of the remaining 30% and may contain an admixture of spindle cells. The epithelioid subset generally was previously diagnosed as leiomyoblastomas, although some may have been mistaken for poorly differentiated carcinomas. The differential diagnosis of GI tract neoplasms that appear to be mesenchymal in origin includes GISTs (80% of the time), but there are definitely true smooth muscle neoplasms of the GI tract, including true leiomyomas and leiomyosarcomas (approximately 15%). Schwannomas account for the remaining 5%. Therefore, the differential diagnosis is complex and requires expert pathology review, as well as adequate and appropriately processed and fixed diagnostic tissues. As noted, expression of KIT is not limited to GIST cells. Normal ICCs and mast cells express CD117 and depend on KIT for normal growth and development. A relatively limited number of other tumors may also express immunohistochemically detectable CD117. These include certain subsets of soft tissue sarcomas, including Ewing’s sarcoma and angiosarcoma, as well as other neoplasms, such as occasional small cell lung cancers, melanomas, desmoid tumors, seminomas, ovarian carcinomas, mastocytomas, neuroblastomas, adenoid cystic carcinomas, and rare subsets of lymphoma and acute myeloid leukemia.22,23,51,52 It is also relevant

to note that expression of the CD117 antigen does not imply the activation of the KIT target, nor does it correlate necessarily with any KIT gene mutation. The same CD117 antigen is expressed by cells harboring normal (wild-type) KIT as those that have activating KIT mutations. Also, expression of KIT protein does not necessarily mean that the protein is involved in the pathogenesis of that specific cancer. In all these regards, GIST was a special example of a disease in which expression correlates universally with kinase activation, and this activation is now validated as being etiologically relevant to the neoplastic behavior of GIST cells.

TREATMENT ADVANCED GASTROINTESTINAL STROMAL TUMORS: METASTATIC, SURGICALLY UNRESECTABLE, OR RECURRENT DISEASE Ineffectiveness of Systemic and Locoregional Conventional Chemotherapy

Often, new approaches to disease management in cancer medicine are first performed in patients with advanced disease. This has been true in the clinical development of molecularly targeted therapies for patients with GISTs. Therefore, this section will first discuss the management of patients with advanced disease and then its relevance for patients with potentially curable early-stage, limited GISTs. There was universal consensus that treatment of advanced disease represented a pressing unmet medical need before the advent of molecularly targeted therapy. Efforts of medical oncologists to treat GISTs with conventional cytotoxic chemotherapy were universally futile. The rates of benefit or objective antitumor response to various chemotherapy agents for patients with GIST or abdominal leiomyosarcomas were routinely reported to be 0% to 4%.49,53 Some investigators attempted to improve on the dismal results with chemotherapy by administering the drugs via an intraperitoneal route.54 However, because GISTs rarely remain confined to the peritoneal surfaces with the hematogenous dissemination of metastases to the liver and other intra-abdominal locations, and because most of the life-threatening complications of GIST arise from hepatic involvement or from bulk disease affecting the omentum, this intraperitoneal approach was not particularly promising. On the basis of these disappointing results, conventional cytotoxic chemotherapy has generally been regarded as useless for the treatment of patients with GIST. The mechanisms that might explain the high levels of resistance to chemotherapy exhibited by GIST may result, in part, by the expression of increased levels of Pglycoprotein (the product of the MDR-1 gene) and the multidrug resistance protein in GISTs and other intra-abdominal sarcomas. In one study evaluating the differences in outcome between GISTs and leiomyosarcomas, significantly higher levels of expression of P-glycoprotein (38% vs. 13%) and multidrug resistance protein-1 (35% vs. 13%) were demonstrated in the GIST cells.55 It has been postulated that these cellular efflux pumps may prevent chemotherapy from reaching effective intracellular concentrations in the target GIST cells. There are some uncontrolled data regarding the potential to control metastatic GIST for a limited time by locoregional

Chapter 30  Gastrointestinal Stromal Tumors (GISTs) techniques, such as hepatic resection, hepatic artery embolization, or chemoembolization. Although a subset of patients with metastatic GISTs involving the liver have demonstrated antitumor responses and a somewhat limited progression-free survival following chemoembolization, the benefits are generally measured in months rather than years; this has not been viewed as a particularly promising strategy for the management of most GIST patients.56,57 Radiotherapy plays little, if any, role in the management of patients with metastatic GIST. There are few cases in which radiotherapy has been carefully studied in this disease, most likely because the delivery of therapeutic doses of radiotherapy to the liver or the GI tract usually causes more morbidity than benefit. However, it is possible that targeting radiotherapy with newer techniques such as intensity modulated radiotherapy (IMRT) or proton beam irradiation might be used to palliate patients suffering from focal bleeding or pain from a specific lesion of treatment-refractory GIST. In general, however, the diffuse pattern of disease spread that characterizes GIST does not allow radiotherapy to function as an effective thera­ peutic modality for most patients with advanced disease. Similarly, surgery has traditionally not played a significant role in the management of patients with metastatic GIST; most patients with liver and peritoneal metastases from GIST are judged unresectable because of multifocal hepatic metastases or multiple sites of intra-abdominal metastatic disease. Clearly, for patients with metastatic or unresectable GISTs, the prognosis was dismal before the advent of molecularly targeted therapy. For patients with metastatic or recurrent GISTs or GI sarcomas (most of which were likely to have been a true GIST), most studies prior to the introduction of targeted therapy with kinase inhibition had documented poor survival rates, with fatal outcomes from disease progression generally occurring within two years from the date of first recurrence or metastasis.6,53,58

Imatinib Mesylate: First-Generation Selective Tyrosine Kinase Inhibitor Targeting KIT

The identification of the uncontrollably activated KIT receptor tyrosine kinase has provided a molecule critical to the pathobiology of GIST and represents an appealing therapeutic target for inhibition. It was serendipitous that a medication being developed for an entirely different purpose showed dramatic activity at inhibiting the KIT enzymatic action. The initial concept for this molecularly targeted approach came from the collaborative studies of researchers at CibaGeigy (now known as Novartis). These studies linked academia and the biopharmaceutical industry in efforts to develop small molecules for inhibition of tyrosine kinases such as the PDGF receptor (useful potentially to block restenosis of coronary stents) and the constitutively active kinase function of the BCR-ABL oncoprotein, which is a target critical to the pathogenesis of chronic myeloid leukemia (CML). A small molecule in the 2-phenylaminopyrimidine class was identified by Druker and colleagues at Ciba-Geigy, with potent inhibitory activity for ABL and the dysregulated BCR-ABL in vitro.59 Additional screening studies from the laboratories of Druker and Buchdunger and associates59-61 demonstrated that this agent, signal transduction inhibitor-571 (STI-571), subsequently called imatinib mesylate (now known by the commercial name of Gleevec in the United States and Glivec elsewhere), could also potently inhibit the tyrosine kinase activity of KIT and

platelet-derived growth factor receptor (PDGFR). Subsequent studies performed in a human mast cell leukemia cell line that harbored a KIT mutation similar to the mutations noted in GISTs documented that imatinib could inhibit both mutant and wild-type KIT protein.62 Laboratory experiments testing imatinib in human GIST cell lines with defined activating mutations of KIT have revealed dramatic evidence of anti-GIST activity from this agent. The addition of imatinib to cultured human GIST cells rapidly blocks the constitutive activation of KIT, arrests cell proliferation, and induces apoptosis in the tumor cells.63 By all criteria, therefore, the clinical development of imatinib has been promising as a treatment of GIST to target the fundamental molecular pathogenesis of this disease. The collaborative worldwide clinical development of imatinib as a molecularly targeted therapy of GIST proceeded at a dramatic pace since the earliest demonstration of the unprecedented activity of this agent (Table 30-1). The first clinical experience with imatinib in the treatment of GIST began in March 2000, with a single-patient pilot study of a woman in Helsinki, Finland, who had far advanced, heavily pretreated, and widely metastatic GIST. The case history of this patient documents the rapid response and sustained clinical benefit from imatinib dosing that this patient enjoyed for approximately three years.64 However, resistance to imatinib (see later) ultimately developed and the patient succumbed to metastatic GIST refractory to selective kinase inhibition. Nonetheless, on the basis of the dramatic and durable benefits in this patient, as well as the striking scientific rationale and strong preclinical data, other studies testing imatinib for GISTs were started. A multicenter United States–Finland collaborative study randomized patients with metastatic GISTs between two dose levels of the drug (400 or 600 mg of imatinib administered orally each day continuously, as long as the disease was stable or responding to therapy).49 The United States–Finland trial rapidly accrued 147 patients with metastatic GIST. Almost concurrently, a dose-finding study was also begun in Europe under the auspices of the European Organization for the Research and Treatment of Cancer (EORTC) Sarcoma Group to test imatinib for GISTs, as well as other forms of sarcomas.65 The maximal tolerated dose of imatinib identified in the EORTC dose-ranging phase I trial was judged to be 800 mg/day (given as 400 mg twice daily); at the higher dose level of 1000 mg daily (given as 500 mg twice daily), unacceptably severe, dose-limiting toxicities, such as nausea, vomiting, and severe edema, were reported. The results from these two trials in the United States and Europe confirmed the exceptional activity of imatinib in controlling metastatic GIST, inducing objective responses in most patients, providing control of symptoms, and prolonging the survival of these patients in comparison with historical series. The results from these wholly independent trials were remarkably concordant (see Table 30-1), further confirming the dramatic effectiveness of imatinib in the treatment of metastatic GISTs. Most patients with advanced GISTs treated with imatinib exhibited objective responses (almost all partial responses), and an additional subset of patients experienced objectively stable disease associated with imatinib dosing. The median time to objective response was more than three months, although some patients experienced dramatic disease regressions within one week after starting imatinib oral dosing. There were no significant differences in response rates or duration of disease control between the 400 and 600 mg daily dose levels of imatinib in the United States–Finland trial, although the study was

469

470

Section IV  Topics Involving Multiple Organs Table 30-1  Clinical Studies of Imatinib Mesylate in Patients with Metastatic or Unresectable Gastrointestinal Stromal Tumors (GISTs) REFERENCE

NO. OF PATIENTS

DAILY IMATINIB DOSE (mg)

RESULTS

64 65

1 40 (36 GIST)

400 400-1000

49

147

400 or 600

66

51 (27 GIST)

800

73

946

400 vs. 800

72

746

400 vs. 800

Major response, durable for >2 yr Partial remissions in 19/36 (53%) with GIST, with additional minor responses in 6/36 (17%) Total clinical benefit rate = 70% No responses in non-GIST tumors Partial remissions in 97/147 (66%), with additional minor responses and durable stable disease in 25/147 (17%) Total clinical benefit rate = 83%, with no differences between doses Complete remissions in 4%, partial remissions in 67%, with additional minor responses and durable stable disease in 18% Total clinical benefit rate = 89% Complete remissions in 5%, partial remissions in 45%, with additional minor responses and durable stable disease in 32% Total clinical benefit rate = 82%, with no differences between doses Complete remissions in 2%, partial remissions in 46%, with additional minor responses and durable stable disease in 26% Total clinical benefit rate = 74%, with no differences between doses

relatively underpowered to detect any differences.49 Imatinib was well tolerated overall in both studies. On the basis of these trials, the U.S. Food and Drug Administration (FDA) approved the use of imatinib for the treatment of metastatic or unresectable GIST in early 2002. Approval in Europe and the rest of the world followed quickly thereafter. In a subsequent trial, the EORTC group went on to expand its exploration of imatinib in GISTs and other forms of sarcomas. In this trial, the high levels of antitumor activity against GISTs were again confirmed, whereas there was no demonstrable benefit for patients with other forms of soft tissue sarcomas.66 This work supports the hypothesis that specific molecular targeting of a signaling pathway crucial to GIST cells can explain the extraordinary activity of imatinib in this disease. Without such a target to inhibit (as in sarcomas other than GISTs), imatinib treatment does not have major anticancer activity. In particular, it is relevant to note that the activation of the PDGFR signaling pathway in dermatofibrosarcoma protuberans (DFSP) has been targeted successfully using imatinib as well, with evidence of clinical benefit for DFSP patients.67-69 With reference to the molecular targeting of kinase inhi­ bition in GISTs, correlative molecular studies performed in conjunction with the United States–Finland trial have documented differences in the activity of imatinib on the basis of the genotype of the GIST lesions treated. Specifically, patients whose GISTs harbored KIT mutations in exon 11 (the most common molecular subtype) had higher rates of objective response and more durable disease control over time with continuous dosing of imatinib than patients whose disease had exon 9 KIT mutations or no detectable KIT mutations at all.70 Imatinib-sensitive PDGFRA mutations can explain the clinical benefit seen in certain GIST patients whose disease does not harbor any KIT mutations.70 One of the more impressive aspects of this work has been the outstanding tolerability of imatinib overall. It is fortu-

nate that normal physiologic processes (e.g., hematopoiesis) that depend on the normal receptor-ligand signaling through the KIT receptor do not fail with fatal consequences when KIT function is blocked by imatinib therapy. The adverse effects of imatinib are generally mild (grade 1 or 2 by the NCI common toxicity criteria) and include edema (in approximately 74% of patients), especially notable in the loose subcutaneous tissues of the facial periorbital region, diarrhea (45%), myalgia or musculoskeletal pain (40%), skin rashes (30%), and headache (25%). Myelotoxicity has been much less common in GIST patients than in patients with CML. Nonetheless, GIST patients treated with imatinib can occasionally exhibit severe cytopenias and, because of this risk, should be monitored carefully. The most worrisome adverse events observed in the imatinib treatment of patients with advanced GIST have included hemorrhages from abdominal or GI sites in approximately 5% of GIST patients. These hemorrhagic events were likely related to bleeding from bulky tumor masses, which may have been induced by the potent and rapid antitumor effects of imatinib. There were no deaths on the initial trial directly attributable to the study drug.49 Most of the side effects of imatinib therapy become more mild over time, suggesting that some sort of tachyphylaxis mechanism may be present. For example, the edema associated with imatinib therapy of GIST often improves with continued dosing over time, although diuretics may be used judiciously and are often effective at managing this side effect. Counseling GIST patients on the use of a low-salt diet may also be highly effective for controlling this side effect of treatment. Nausea with imatinib administration is usually mild and self-limited; for most patients, taking the daily dose with food and dividing the dose may be useful. Muscle cramps, frequently in the calves, are usually transient and self-limited; many patients have noted that increased fluid intake can help alleviate the frequency or severity of the muscle cramping. In general, imatinib is reasonably well

Chapter 30  Gastrointestinal Stromal Tumors (GISTs) tolerated and patients have been able to comply with longterm dosing without an excessively negative impact on their functional status. Another fascinating aspect of this work was the finding that imatinib could rapidly and dramatically affect the 18FFDG tracer uptake by tumors as imaged by functional imaging with PET scans (see Fig. 30-5). The decreases in tumor avidity for 18F-FDG on PET scanning could be detected as early as 24 hours following a single dose of imatinib. PET imaging could detect the biological activity of imatinib far earlier than changes in anatomic measures on CT scanning. The findings of PET scans were also highly reliable, correlating both with beneficial response to imatinib and documenting progressive disease in the small subset of patients with primary resistance to imatinib. These data indicate that functional imaging of GIST with 18F-FDG–PET scanning represents a useful diagnostic modality for early-response assessment with imatinib therapy.50,71 This paradigm demonstrates that PET imaging may be a useful tool for future drug development efforts because the signal of drug activity can be detected rapidly and reliably following effective target inhibition. The optimal dose of imatinib in advanced GISTs remains uncertain. Although there were no documented benefits to the higher dosage of 600 mg/day in the United States– Finland trial, there were a few patients who regained disease control when crossed over from the lower dose (400 mg daily) to the higher dose level. Therefore, there might be some marginal benefit to be obtained from dose escalation of imatinib in a subset of patients whose disease progresses while taking lower doses of imatinib. Structural studies of the different mutant kinase isoforms may shed light on these mechanisms because certain mutations may prove more dose sensitive than others. To explore more definitively whether there is a clinically significant dose response above the lowest recommended daily dose of 400 mg, two large phase III randomized studies have been conducted. In both studies, patients with advanced metastatic and/or unresectable GISTs were randomized to receive imatinib at 400 mg or 800 mg daily. Patients were allowed to cross over from the lower dose to the higher dose if progression of disease occurred at the lower dose. These studies were considered to be adequately powered to determine whether this twofold difference in imatinib dose would translate into significant clinical benefits as measured by improved response rates, duration of disease control, or survival for patients with advanced unresectable or metastatic GISTs. The most recent updates on these studies have been published,72,73 and together these studies have treated and studied almost 1700 GIST patients worldwide. Importantly, because these trials shared the same fundamental designs at inception, a meta-analysis has now been conducted based on a pooled analysis of all primary data. Although there was no survival difference documented in either trial between these two dose levels, there was an interesting, although subtle, difference observed in terms of duration of disease control (see Table 30-1). Although the North American Sarcoma Intergroup trial demonstrated only a favorable trend in the duration of disease control associated with the higher dose of imatinib, the European-led study noted a modest but statistically significant benefit in favor of the higher dose arm for progression-free survival; however, neither showed any difference in overall survival. In the combined meta-analysis, it is clear that the benefit of the higher dose imatinib was limited solely to the subset of GIST patients whose tumors harbored mutations in KIT exon 9, encoding a mutation in the extracellular domain that promotes dimerization of the

kinase.74 This has led to the recommendation by some experts to use the higher dose of imatinib as first-line therapy for advanced GIST in patients whose tumors have documented exon 9 KIT mutations. Such modest benefit must, however, be balanced by the additional toxicities, because the higher dose of imatinib was associated with a greater incidence of adverse effects and led to a greater number of dose reductions for toxicity in these large studies. The optimal duration of imatinib for patients with metastatic GIST has been defined as lifelong therapy on the basis of current evidence.75 A randomized study in France that discontinued imatinib therapy in GIST patients once a maximal response was reached has found that the disease rapidly recurs following cessation of imatinib dosing.76 For optimal management of metastatic disease, medical oncologists, surgeons, radiologists, and nuclear medicine imaging experts must all collaborate closely to determine the best course of action for any given patient. This important message has been emphasized in the Task Force Report on GIST Clinical Practice Guidelines of the NCCN.75 For example, disease that is initially judged as unresectable may become amenable to surgical excision following a major response induced by imatinib therapy. Most centers recommend surgical resection of such patients because it is feared that residual GIST lesions likely harbor cells with kinase inhibitor resistance mutations, which evolve into clinical appearance of GIST resistance to imatinib and subsequent progression of treatment-refractory disease. Resistance to imatinib may be primary and manifest as rapid progression of disease despite initial imatinib dosing, although this appears in far less than 20% of patients (see Table 30-1). Alternatively, clonal evolution of GIST may occur over time, with the emergence of resistant disease after more than one or two years of durable response and disease control by imatinib. Several mechanisms of resistance to imatinib in GIST have been described,77 and these are overall similar to the resistance mechanisms that have been described in imatinib-resistant CML.78 It is unclear what role should ideally be played by other modalities (e.g., surgical resection, radiofrequency ablation [RFA], chemoembolization, or other locoregional approaches) for managing metastatic GIST once imatinib has achieved the optimal effect or following the development of resistance to imatinib with oligoclonal or widespread systemic progression. Certainly, many metastatic GIST lesions may remain controlled on imatinib, although limited clonal progression appears as the first sign of resistance to imatinib.79,80 It may be feasible in such patients to resect the resistant clonal growth while maintaining control over most of the disease by continuation of imatinib dosing. In fact, several nonrandomized surgical series have suggested benefit to resection of resistant clones for patients whose GISTs have not developed widespread systemic resistance to kinase inhibitor therapy.81,82 The value of such surgical strategies will be tested in future randomized trials to assess whether early surgical intervention will improve disease control compared with kinase inhibitor therapy alone.

Gastrointestinal Stromal Tumors Resistant to Imatinib: Development of Sunitinib

To address resistance, other kinase inhibitors with varying target specificities have been developed to control GISTs that have become refractory to imatinib. Sunitinib (Sutent) is the first of these that has been FDA-approved for GIST, as well as by worldwide regulatory authorities. Sunitinib inhibits multiple receptor tyrosine kinases including KIT;

471

472

Section IV  Topics Involving Multiple Organs PDGFRs (alpha and beta); vascular endothelial growth factor receptors (VEGFRs)-1, -2, and -3; FMS-like tyrosine kinase-3 receptor (FLT3); macrophage colony-stimulating factor receptor (CSF-1R); and glial cell–line derived neurotrophic factor receptor (RET, rearranged during t ransfection). Sunitinib demonstrated important antineoplastic activity in the first phase I clinical trial,83 and subsequently this drug was proven to have potent activity in a definitive international prospective randomized, placebo-controlled clinical trial.84 In molecular analysis of GIST samples from the initial phase I and II trial, primary and secondary mutations in KIT and PDGFRA were shown to affect the treatment outcomes with sunitinib in patients with imatinib-resistant GISTs.85 Sunitinib was effective for treatment of GISTs of all KIT and PDGFRA genotypes prior to imatinib exposure, and sunitinib showed particular efficacy for those with a wild-type genotype, a primary KIT exon 9 mutation, or secondary KIT mutations in exon 13 or 14. Structural biology analyses of these mutated kinases have provided insight as to how sunitinib can inhibit the kinase function when mutations encode certain amino acid changes that induce steric hindrance to the binding and inhibition of imatinib.86 Importantly, the incidence of secondary mutations (especially exon 17 KIT mutations, which confer resistance to imatinib and sunitinib) was higher in patients whose GIST was initially sensitive to imatinib. In the phase III trial, which led to the regulatory approval of sunitinib, 312 patients with metastatic or surgically unresectable GIST following failure of imatinib caused by resistance or intolerance were randomized to receive sunitinib, 50 mg/day (n = 207), or placebo (n = 105) on a dosing schedule with four weeks of drug dosing followed by a two-week period off drug.84 The primary endpoint of the study was disease control as assessed by time to progression (TTP). The trial was unblinded early when a planned interim efficacy analysis showed that sunitinib was associated with a significant improvement in median TTP of more than fourfold compared with placebo. The median progression-free survival (PFS) for patients receiving sunitinib was significantly greater than for those receiving placebo (sunitinib, 24.1 weeks. vs. placebo, 6.0 weeks). In the initial analysis, sunitinib also significantly improved overall survival (OS; hazard ratio, 0.49; 95% confidence interval [CI], 0.29 to 0.83); at the time of the interim analysis, the median OS had not been reached in the group receiving sunitinib. Although improved disease control was demonstrated with sunitinib, the objective rates of response were very low, even though significantly more patients treated with sunitinib had an objective response compared with placebo (6.8% vs. 0%). Sunitinib was reasonably well tolerated, with the most common adverse events being fatigue, diarrhea, abdominal pain, and nausea. Patients on sunitinib also experienced a greater incidence of skin abnormalities, including palmarplantar erythrodysesthesia (hand-foot syndrome), oral cavity mucosal irritation, and, with longer exposure, a relatively high incidence of hypothyroidism.87 Certain patients also exhibited cardiac dysfunction, which was in general reversible with temporary discontinuation of sunitinib dosing.88 Thus, sunitinib appears to have unique activity for the management of imatinib-resistant GIST as second-line therapy, but the powerful inhibition of a number of other kinase signaling pathways can induce other unpleasant or medically relevant adverse effects that require close monitoring and possibly adjustment of dosing. Additional studies are evaluating a lower continuous daily dose regimen of sunitinib, which appears to have similar activity with perhaps somewhat improved tolerability.89 Clinical trials

are also evaluating several other kinase inhibitors, including the second-generation selective tyrosine kinase inhibitor nilotinib, the multitargeted kinase inhibitor sorafenib, and the combination of imatinib plus RAD001, an inhibitor of the serine-threonine kinase mTOR. In each case, a certain level of clinical efficacy has been demonstrated in limited phase II trials, and larger trials are warranted and necessary to define more precisely whether specific molecular subtypes of GIST might benefit from these treatment strategies. Given the evolution of several different kinase mutations in individual patients with imatinib- and sunitinib-resistant GISTs,90,91 new strategies are needed to control the polyclonal resistance that emerges following failure of first-line therapy. Because mutated kinase proteins appear to maintain dependence on the protective chaperone function of the heat shock protein 90 (Hsp90), inhibitors of Hsp90 such as IPI-504 are being tested in clinical trials as a consequence of promising data from the laboratory.92 The early phase I-II trial experience with the Hsp90 inhibitor IPI-504 has been encouraging,93 and these data have now been translated into other clinical trials to test the value of this novel molecular target for GISTs.

PRIMARY LOCALIZED DISEASE

Definitive expert surgery remains the mainstay of treatment for patients with primary localized GISTs (early-stage GISTs). However, with the advent of highly effective drug therapy, current guidelines recommend that surgical resection of GISTs be undertaken as the first intervention only if there is an acceptably low risk of functional deficit or morbidity from the surgery. If a large GIST were to be detected, it might be judicious to consider such a lesion unresectable without causing unacceptable risk for morbidity. In this case, preoperative administration of imatinib should be considered, because trials have shown that such neoadjuvant administration can be effective at diminishing the size of tumors and thereby facilitating effective surgical intervention.94 This is the specific clinical situation in which early assessment of therapeutic response by 18F-FDG–PET scanning could prove valuable to confirm that the patient’s disease is exhibiting the desired response to imatinib. This should minimize the risk of disease progression that otherwise might put the patient at risk for further growth and invasion into surrounding vital structures. Following maximal response (usually occurring within three to six months), definitive surgery could be performed. The surgical approach to GIST resection of primary disease must take into account the specific growth and behavior characteristics of this disease. GISTs rarely involve the locoregional lymph nodes, and so extensive lymph node exploration or resection is rarely indicated. GIST lesions are highly vascularized and often exhibit a fragile pseudocapsule; therefore, surgeons should be careful to minimize the risk of tumor rupture, which might subsequently increase the risk of peritoneal dissemination.95 The margins of resection from the tumor specimen should be carefully oriented and examined, and biopsy samples from several different areas of the tumor should be evaluated by the surgical pathologist. The natural history of early-stage primary GIST has been examined in studies from single-institution referral centers. These are certainly prone to selection bias, and it is clear in this evolving field that many early-stage GIST patients have likely been managed by physicians of multiple specialties, including gastroenterology and general surgery. However, one of the larger GIST series from a referral cancer center evaluated 200 patients followed prospectively at the Memorial Sloan-Kettering Cancer Center6; 80 of these patients

Chapter 30  Gastrointestinal Stromal Tumors (GISTs) (40%) had primary disease managed with complete surgical resection. This latter group, with primary resected GISTs, demonstrated five-year disease-specific survival rates of only 54%, supporting the fact that GISTs, as seen at such an academic referral center, can exhibit a high risk for recurrence and ultimately prove to be a life-threatening disease. On multivariate analysis, large tumor size (>10 cm) was the only factor that reduced disease-specific survival. In an earlier study of 191 so-called GI leiomyosarcomas (of which a sizable proportion were likely to have been true GIST), investigators at the M.D. Anderson Cancer Center reported that smaller tumor size (<5 cm), complete surgical resection without tumor rupture, and low histologic grade of tumor were significant favorable prognostic factors. The propensity of GISTs to recur was also confirmed by these data because only 10% of these patients were disease-free on long-term follow-up.95 Anatomic location of the primary tumor also appears to be an important prognostic factor for primary localized GIST, in addition to tumor size and the proliferation rate of the tumor (as measured by the number of tumor cells in mitosis).96 Assessing the risk of recurrence will be exceedingly important to advise patients and to make reasonable judgments about the potential value of adjuvant systemic therapy with a kinase inhibitor.

yet been noted because of the very short follow-up period.97 It is certainly possible that a longer duration of adjuvant therapy might lead to even better outcomes, and this is being tested in the other European trials. Given the potential toxicities and costs of imatinib, it is important to decide what constitutes sufficient clinical activity to justify the universal administration of imatinib in the adjuvant setting. Although imatinib has powerful activity in prolonging the survival of patients with recurrent and metastatic GISTs, it is possible that earlier administration of this systemic therapy in the adjuvant setting might not change the natural history of this disease sufficiently to affect overall survival. However, it is possible that with longer duration of dosing in patients at highest risk of recurrence, such trials may document important clinical benefits. These trials in progress will generate crucial evidence on which to base optimal medical practice in the future. Nonetheless, for the moment, the adjuvant activity of imatinib seems clear. It merits a thoughtful discussion of possible risks and benefits in all patients with resected GISTs who are at moderate to high risk of disease recurrence by current risk classification systems.

Adjuvant Therapy for Patients with Early-Stage Gastrointestinal Stromal Tumors

FAMILIAL DISEASE

To date, there have only been a limited number of case reports and small series that have investigated the role of adjuvant treatment using conventional modalities, such as radiotherapy after surgical resection. As noted, radiotherapy does not appear to have an important role in the treatment of GIST, with only minimal activity seen at doses that are safe to administer, given the toxicity to small bowel and other intra-abdominal structures. Because cytotoxic chemotherapy is also not associated with disease control or objective responses in metastatic GISTs, there have only been small series of patients who have received adjuvant systemic or intraperitoneal chemotherapy, and these data have not clearly identified any benefits. The standard of care after complete surgical resection of GISTs was therefore obser­ vation alone before the availability of imatinib. Because there is now effective medical treatment for advanced GISTs, it is important that all GIST patients undergo regular surveillance following resection. In this way, any recurrent disease could be detected and treated at the earliest stage, thereby hopefully avoiding complications that might stem from treatment of large bulky disease (e.g., intratumoral hemorrhages). It is now clear that the administration of imatinib in the postresection (adjuvant) setting has the potential to delay tumor recurrence, especially for patients who present with very large tumors and who are likely at very high risk of disease recurrence and metastatic spread. The activity of adjuvant imatinib in GIST patients who are at moderate to high risk of recurrence has been investigated in two large multicenter trials conducted by several groups, including the American College of Surgeons Oncology Group (ACOSOG) and several European cooperative oncology groups. In the initial adjuvant trials of ACOSOG, imatinib (or placebo) was administered for an arbitrary one year’s duration. In the randomized trial Z9001, administration of imatinib following resection of primary limited GIST significantly prolonged recurrence-free survival compared with placebo (98% vs. 83% free of recurrence at one year; hazard ratio 0.35), although no overall survival benefit has

SPECIAL CONSIDERATIONS GIST rarely can be associated with familial inheritance patterns in which several members of a kindred have the disease.20,98-100 In several of these families, KIT mutations have been reported. Additional characteristics of affected family members include cutaneous lesions such as hyperpigmentation or skin lesions that resemble the clinical appearance of urticaria pigmentosa. These skin pigmentation abnormalities are no doubt caused by the effect of the mutationally activated KIT kinase function on melanocyte growth and development. The mechanisms whereby such pigmentation disorders remain focal, rather than disseminated, may provide clues as to why GIST lesions may take decades to appear in these rare familial cases. These cases have tended to be autosomal dominant germline mutations, and the GIST lesions seen in affected members tend to be multifocal. Familial GIST has also been associated with germline mutations in certain subunits of the succinate dehydrogenase genes SDHB, SDHC, and SDHD.101

PEDIATRIC PATIENTS

Although very rare, GIST can affect pediatric patients. The disease seems to have a very different molecular profile, however, because KIT mutations are rarely, if ever, demonstrable in GIST patients younger than 18 years.102-104 The presence of constitutive KIT signaling is nonetheless a common feature of pediatric GIST, although the mechanism of KIT activation in the absence of mutations remains obscure. Wild-type GIST, especially in pediatric patients, appears to exhibit high levels of the insulin-like growth factor-1 receptor, which may provide another therapeutic target for intervention using investigational new inhibitory monoclonal antibodies or small molecules.105

RELATIONSHIP TO OTHER GENETIC SYNDROMES PREDISPOSING TO NEOPLASMS

Several other syndromes that predispose to the development of neoplasms have been described in association with GISTs. One of the more widely known is the Carney triad, which includes GISTs (often multifocal) in addition to pul-

473

474

Section IV  Topics Involving Multiple Organs monary chondromas and extra-adrenal paragangliomas.106 A variant of this syndrome, known eponymously as the Carney-Stratakis syndrome, has been described, with only GISTs and familial paragangliomas being present.107 Additionally, a linkage between neurofibromatosis type I (NF1) and an increased incidence of GISTs has been widely noted.108,109 Molecular analysis of GIST lesions arising from patients with NF1 disease has documented that these GISTs do not harbor detectable mutations in the KIT gene. It is unclear whether GISTs that arise in the setting of a genetic predisposition syndrome have the same response to imatinib as sporadically occurring GISTs.

KEY REFERENCES

Blanke CD, Rankin C, Demetri GD, et al. Phase III randomized intergroup trial assessing imatinib mesylate at two dose levels in patients with unresectable or metastatic gastrointestinal stromal tumor expressing the KIT receptor tyrosine kinase: S0033. J Clin Oncol 2008; 26:626-32. (Ref 72.) Blay JY, Le Cesne A, Ray-Coquard I, et al. Prospective multicentric randomized phase III study of imatinib in patients with advanced gastrointestinal stromal tumors comparing interruption versus continuation of treatment beyond 1 year: The French Sarcoma Group. J Clin Oncol 2007; 25:1107-13. (Ref 76.) Choi H, Charnsangavej C, Faria SC, et al. Correlation of computed tomography and positron emission tomography in patients with metastatic gastrointestinal stromal tumor treated at a single institution with imatinib mesylate: proposal of new computed tomography response criteria. J Clin Oncol 2007; 25:1753-9. (Ref 48.) Debiec-Rychter M, Sciot R, LeCesne A, et al. KIT mutations and dose selection for imatinib in patients with advanced gastrointestinal stromal tumours. Eur J Cancer 2006; 42:1093-103. (Ref 74.) DeMatteo RP, Maki RG, Singer S, et al. Results of tyrosine kinase inhibitor therapy followed by surgical resection for metastatic gastrointestinal stromal tumor. Ann Surg 2007; 245:347-52. (Ref 82.)

Demetri GD, Benjamin RS, Blanke CD, et al. NCCN Task Force report: Management of patients with gastrointestinal stromal tumor (GIST)— update of the NCCN clinical practice guidelines. J Natl Compr Canc Netw 2007; 5(Suppl 2):S1-29. (Ref 75.) Eisenberg BL, Harris J, Blanke CD, et al. Phase II trial of neoadjuvant/ adjuvant imatinib mesylate (IM) for advanced primary and metastatic/recurrent operable gastrointestinal stromal tumor (GIST): Early results of RTOG 0132/ACRIN 6665. J Surg Oncol 2009; 99:42-7. (Ref 94.) ESMO Guidelines Working Group; Blay JY, Le Cesne A. Gastrointestinal stromal tumors: ESMO clinical recommendations for diagnosis, treatment and follow-up. Ann Oncol 2007; 18(Suppl 2):ii27-9. (Ref 41.) Heinrich MC, Maki RG, Corless CL, et al. Primary and secondary kinase genotypes correlate with the biological and clinical activity of sunitinib in imatinib-resistant gastrointestinal stromal tumor. J Clin Oncol 2008; 26:5352-9. (Ref 85.) Liegl B, Kepten I, Le C, et al. Heterogeneity of kinase inhibitor resistance mechanisms in GIST. J Pathol 2008; 216:64-74. (Ref 91.) McWhinney SR, Pasini B, Stratakis CA; International Carney Triad and Carney-Stratakis Syndrome Consortium. Familial gastrointestinal stromal tumors and germ-line mutations. N Engl J Med 2007; 357:1054-6. (Ref 101.) Miettinen M, Lasota J. Gastrointestinal stromal tumors: pathology and prognosis at different sites. Semin Diagn Pathol. 2006; 23:70-83. (Ref 96.) Tarn C, Rink L, Merkel E, et al. Insulin-like growth factor 1 receptor is a potential therapeutic target for gastrointestinal stromal tumors. Proc Natl Acad Sci U S A 2008; 105:8387-92. (Ref 105.) Verweij J, Casali PG, Zalcberg J, et al. Progression-free survival in gastrointestinal stromal tumours with high-dose imatinib: Randomised trial. Lancet 2004; 364:1127-34. (Ref 73.) Wardelmann E, Merkelbach-Bruse S, Pauls K, et al. Polyclonal evolution of multiple secondary KIT mutations in gastrointestinal stromal tumors under treatment with imatinib mesylate. Clin Cancer Res. 2006; 12:1743-9. (Ref 90.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

31 Gastrointestinal Carcinoid Tumors (Gastrointestinal Neuroendocrine Tumors) and the Carcinoid Syndrome Kjell Öberg

CHAPTER OUTLINE Clinical Features  476 Pathology  476 Molecular Genetics  478 Classification and Subtypes  478 Esophagus  478 Stomach  478 Pancreas  479 Duodenum and Ampulla of Vater  479 Small Intestine  480 Appendix  480 Colon  481 Rectum  481 The Carcinoid Syndrome  482

In 1888, Lubarsch described what is now recognized to be the entity of carcinoid tumors when he reported the autopsy finding of two patients with multiple tumors in the distal ileum.1 The term carcinoid was introduced by Oberndorfer in 1907 in his description of a class of malignant tumors that behaved less aggressively than the more common adenocarcinomas of the gastrointestinal (GI) tract.2 The traditional classification of carcinoids, based on their embryonic origin into foregut, midgut, and hindgut car­ cinoids, has been gradually abandoned.3 A tumor biology– based classification system introduced by the World Health Organization (WHO) in 2000 has greater applicability, although more recent amplification and revision of this system by the European Neuroendocrine Tumor Society (TNM classification) appears likely to become the standard (Table 31-1).4,5 Carcinoids arise from cells of the diffuse neuroendocrine system and can arise almost anywhere within the gastroin­ testinal tract.6 Carcinoids, also called gastrointestinal neu­ roendocrine tumors (GI NETs), are related to medullary carcinoma of the thyroid, pheochromocytoma and pancre­ atic neuroendocrine tumors. GI NETs synthesize bioactive

Pathophysiology  483 Carcinoid Crisis  484 Diagnosis  484 Biochemical Markers  484 Tumor Localization  485 Treatment  486 Surgery  486 Hepatic Artery Embolization and Chemoembolization  488 Chemotherapy  488 Somatostatin Analogs  489 Interferon-α  489 New Biologic Agents  489 Peptide Receptor Radionuclide Therapy  489

amines and peptides, including neuron-specific enolase (NSE), 5-hydroxytryptamine (5-HT, or serotonin), and 5-hydroxytryptophan (5-HTP). They also secrete peptides such as chromogranin A, pancreatic polypeptide, calcito­ nin, tachykinins (neurokinin A and substance P), and various growth factors, such as transforming growth factor-β (TGF-β), platelet-derived growth factor (PDGF), endocrine growth factor (EGF), fibroblast growth factor (FGF), and the vascular endothelial growth factor (VEGF) family of growth factors, including their receptors.7-9 Of clinical relevance is the observation that several different genes and genetic divergences related to tumor development are evident. GI NETs can be sporadic (nonfamilial) or part of a familial syndrome, such as von Hippel-Lindau syndrome or neuro­ fibromatosis (NF). GI NETs comprise 0.5% of all malignan­ cies and, as shown in Figure 31-1, their incidence has increased substantially over the last several decades.10,11 In this chapter, we will use the more familiar term carcinoid, although the term GI NET is the most appropriate, but not as well established. Carcinoid tumors occur most frequently in the GI tract (67%), with the bronchopulmonary system being the second

475

476

Section IV  Topics Involving Multiple Organs Table 31-1  TNM Classification for Endocrine Tumors* A. Gastric Tumors TX Primary tumor cannot be assessed T0 No evidence of primary tumor Tis In situ tumor, dysplasia (<0.5 mm) T1 Tumor invades lamina propria or submucosa and is ≤1 cm T2 Tumor invades muscularis propria or subserosa or is >1 cm T3 Tumor penetrates serosa T4 Tumor invades adjacent structures NX Regional lymph nodes cannot be assessed N0 No regional lymph node metastases N1 Regional lymph node metastases MX Distant metastases cannot be assessed M0 No distant metastases M1 Distant metastases B. Tumors of the Duodenum, Ampulla, and Proximal Jejunum TX Primary tumor cannot be assessed T0 No evidence of primary tumor T1 Tumor invades lamina propria or submucosa and is ≤1 cm T2 Tumor invades muscularis propria or is >1 cm T3 Tumor invades pancreas or retroperitoneum T4 Tumor invades peritoneum or other organs NX Same as for gastric N0 Same as for gastric N1 Same as for gastric MX Same as for gastric M0 Same as for gastric M1 Same as for gastric C. Tumors of Distal Jejunum and Ileum TX Primary tumor cannot be assessed T0 No evidence of primary tumor T1 Tumor invades mucosa or submucosa and is ≤1 cm T2 Tumor invades muscularis propria or is >1 cm T3 Tumor invades subserosa T4 Tumor invades peritoneum/other organs NX Same as for gastric N0 Same as for gastric N1 Same as for gastric MX Same as for gastric M0 Same as for gastric M1 Same as for gastric *European Neuroendocrine Tumor Society. T, primary tumor (for any T, an “m” is added for multiple tumors; N, regional lymph nodes; M, metastases). Adapted from Solcia E, Kloppel G, Sobin L. Histological typing of endocrine tumours. In: Verlag S, editor. World Health Organization Histological Classification of Tumours. 2nd ed. New York: Springer; 2000. p 38; and Rindi G, Kloppel G, Alhman H, et al. TNM staging of foregut (neuro)endocrine tumors: A consensus proposal including a grading system. Virchows Arch 2006; 449:395-401.

most common location (25%), followed by considerably less frequent locations, such as the ovaries, testes, and hepatobi­ liary system.12 The most common location in the GI tract is the ileum (17%). The overall incidence of carcinoid tumors is difficult to determine, because it appears likely that most tumors remain asymptomatic. An autopsy study has esti­ mated the annual incidence to be 8.4/100,000 people.13 The SEER database from 1973 to 2004 indicated an annual inci­ dence of 2.0 to 2.5/100,000/year, with a 3.5% annual increase over this time period.10 Ileal tumors have specifically increased in prevalence—in white males (by 274%), black males (500%), white females (213%), and black females (286%), respectively. A recent evaluation of the SEER data­ base comprising 35,825 cases in the United States has indi­ cated an annual incidence of 5.25/100,000/year for all types of neuroendocrine tumors and a prevalence of 35/100,000.14 The annual incidence of gastrointestinal carcinoids was 2.53/100,000. Overall, the incidence of gastrointestinal car­ cinoids is higher in African Americans (4.5/100,000) com­

pared with white Americans (2.5/100,000) with striking differences in certain sites, particularly the rectum. As a result of the increasing incidence and prevalence, gastro­ intestinal neuroendocrine tumors represent a substantial clinical problem. The increasing incidence is probably mostly to the result of the introduction of better diagnostic tools (imaging and immunohistochemical), as well as greater awareness by pathologists and clinicians. There are no known environmental risk factors for carcinoids.

CLINICAL FEATURES An early and accurate diagnosis is often delayed by four or five years because most small intestinal carcinoids (Fig. 31-2) are small, initially asymptomatic, or misdiagnosed as conditions such as allergy or irritable bowel syndrome. Pro­ gressive growth of the tumor may cause vague abdominal discomfort because of intermittent intestinal obstruction. As the tumor gets larger, invading the intestinal wall and occluding the gut lumen, emergency clinical presentations of an acute abdomen (intussusception, obstruction, perfora­ tion, bleeding) may arise because of local tumor mass effect or tumor-induced fibrosis.15-18 The frequency of intestinal obstruction that is secondary to a gastrointestinal NETassociated mesenteric fibrosis ranges from 42% to 66%.18,19 Another characteristic feature of these patients is vascular elastosis (thickening of the vessel wall), resulting in is­ chemic changes of the gut. The cause of this condition is unclear, but a local effect of serotonin or other tumor prod­ ucts, such as TGF-a, have been postulated to have a direct trophic effect on smooth muscle and fibroblasts in the vessel wall. Both mesenteric fibrosis and elastosis may cause abdominal angina.15 A rare cutaneous manifestation of ileal carcinoids is a fibrotic, scleroderma-like condition mostly affecting the lower extremities.20 As the disease advances, gastrointestinal carcinoids frequently metastasize locally to mesenteric lymph nodes and to the liver. A classic carcinoid syndrome is relatively uncommon (10% to 15%), typically consisting of diarrhea, cutaneous flushing, bronchoconstric­ tion and right-sided heart failure.21,22 As many as 15% to 25% of gastrointestinal carcinoids exhibit a synchronous or metachronous association with other tumors, usually ade­ nocarcinomas of the colon.23-25 This may reflect the activity of growth factors produced by the carcinoid tumor. A rela­ tively large percentage of GI NETs are multicentric; for example, up to 33% of carcinoids in the small intestine are multicentric.24-26

PATHOLOGY The most common histopathologic type of intestinal carci­ noid is the enterochromaffin (EC) cell carcinoid, accounting for more than 98% of cases. The EC cell carcinoid is defined by its argentaffin staining properties, serotonin production, and typical pleomorphic secretory granules. Morphologi­ cally, EC cell carcinoids are characterized by medium-sized tumor cells arranged in an organoid pattern and showing only mild to moderate atypia. Tumor necrosis is absent and the mitotic rate is low (<2 mitoses/10 high-powered field [HPF]). Distinct growth patterns have been described in gastrointestinal carcinoids: (1) nodular or insular pattern; (2) trabecular pattern; (3) acinar and tubular pattern; and (4) atypical solid pattern. Mixed patterns may be seen. EC cell carcinoids of the jejunum and ileum predominantly display

Chapter 31  Gastrointestinal Carcinoid Tumors and the Carcinoid Syndrome Lung and bronchus Small intestine Rectum Other non-epithelial skin Stomach Appendix Cecum Pancreas Sigmoid colon Rectosigmoid junction

1.4

Incidence (per 100,000)

1.2 1.0 0.8 0.6 0.4 0.2 0 1970

1975

1980

1985

1990

1995

Year

2000

2005

Figure 31-1.  Incidence of different subtypes of neuroendocrine tumors, 1970 to 2005, from the Surveillance Epidemiology and End Results (SEER) data base.10,11 Note the significant increase in most subtypes of neuroendocrine tumors since the 1970s. (From Modlin IM, Oberg K, Chung DC, et al. Gastroenteropancreatic neuroendocrine tumours. Lancet Oncol 2008; 9:61-72.)

Figure 31-2.  Midgut carcinoid in the small intestine. The size is typical for primary small intestinal carcinoids, and the tumor is subepithelial in location.

an insular growth pattern (93%; Fig. 31-3) and, less fre­ quently, mixed insular and glandular (5%) or trabecular (2%) growth patterns. EC cell carcinoids invade the submu­ cosa, muscularis propria, or mesentery. Invasion of lym­ phatics and veins, or perineural growth, can be seen. EC cell carcinoids may produce multiple hormones. Most of the tumors (>90%) secrete serotonin and tachykinins (substance P, neurokinin A). Production of other hormones such as gastrin, glucagon, cholecystokinin, calcitonin, somatostatin, or adrenocorticotropic hormone (ACTH) can be demon­ strated in less than 5% of cases. Neuroendocrine markers (proteins associated with large dense core vesicles or synaptic-like microvesicles) are abundantly expressed in EC cell carcinoids and can be used to confirm the neuroen­ docrine phenotype of the these tumors.24,25,27 Chromogranin A, synaptophysin, synaptic vesicle protein 2, neuronspecific enolase, and Leu7 are expressed by 92% to 100% of jejunoileal carcinoid tumors. Expression of the vascular monoamine transporters 1 and 2 (VMAT-1 and -2) has been demonstrated in more than 90% of tumors and is related to the production and storage of serotonin in EC cell carci­ noids. Other markers identified in EC cell carcinoids include cytokeratins 8 and 18, carcinoembryonic antigen (CEA), and

Figure 31-3.  Histopathology of small intestinal carcinoid with characterisitic insular growth pattern (Hematoxylin and eosin, ×80).

prostatic acid phosphatase. High expression of the intestinal transcription factor CDX2 has been demonstrated in EC cell carcinoids and may become a useful marker of intestinal origin.6,27-33 Carcinoids (GI NETs) arising in the duodenum and jejunum include gastrin, somatostatin, and EC cell tumors and gangliocytic paraganglioma. Gastrin cell tumors (gastri­ nomas) are the most common in the duodenum, might be nonfunctioning or functioning (Zollinger-Ellison syndrome [ZES]), and associated with the multiple endocrine neopla­ sia type I (MEN-I) syndrome (see Chapter 32). Tumors are mainly located in the first and second parts of the duode­ num. A small proportion of tumor cells may produce other hormones in addition to gastrin, such as cholecystokinin (CCK), somatostatin, pancreatic polypeptide, neurotensin, and insulin. Morphologically, tumor cells are uniform, with rounded nuclei and abundant cytoplasm. The growth pattern is usually trabecular or cribriform. Necroses are absent and the mitotic rate is low. Most of the tumor cells

477

478

Section IV  Topics Involving Multiple Organs are positive for neuroendocrine markers, such as chromo­ granin A, synaptophysin, Leu7, and NSE.34-36 Somatostatin cell tumors (somatostatinomas) represent the second most frequent histopathologic type in the duo­ denum, accounting for 15% to 27 % of duodenal GI NETs (see Chapter 32). Somatostatin cell tumors are preferentially localized to the periampullary region. They are identified by their content of somatostatin and typical large electrondense secretory granules. A subset of tumor cells may also contain calcitonin, pancreatic polypeptide, and ACTH. Morphologically, tumors are characterized as a mixture of tubular-glandular, insular, and trabecular growth patterns and the presence of psammoma bodies. Tumor cells are uniform and the mitotic rate is low.34,35,37,38 Gangliocytic paraganglioma represent the third most fre­ quent histopathologic type in the duodenum and account for 6% to 9% of duodenal carcinoids. The tumors are pref­ erentially located in the periampullary region and are not associated with familial syndromes. Gangliocytic paragan­ gliomas are identified by their characteristic morphology, which includes a mixture of three different cell types— spindle cells, epithelial cells, and ganglion cells. Spindle cells are arranged in fascicles and stain positive for S-100. Epithelial cells represent the endocrine component and are positive for pancreatic polypeptide, somatostatin, and chromogranin A. Ganglion cells represent the neural component and are positive for synaptophysin and NSE. Gangliocytic paraganglioma are regarded as benign hamar­ tomatous tumors, derived from the ventral pancreatic primordium.39-42 Endocrine tumors of the stomach are composed of cells with features similar to those of the normal endocrine cell counterparts. The most representative cell types are the histamine-producing enterochromaffin-like (ECL) cell in the corpus and fundus and the gastrin-producing G cell in the antrum, representing about 50% of all endocrine cells at both sites. The other endocrine cell types present through­ out the stomach are somatostatin-producing (D), ghrelinproducing (P), and serotonin-producing enterochromaffin (EC) cells. All cell types may be found in gastric tumors, but preneoplastic lesions are known to be comprised mostly of ECL cells. Antral G cell hyperplasia is often observed in chronic atrophic gastritis (see Chapter 51), but it is not con­ sidered a preneoplastic lesion. More than 90% of gastric endocrine tumors are well-differentiated tumors–carcino­ mas (WDET/C), according to the WHO classification. More than 90% of the tumors occur in the corpus-fundus area and are composed of ECL cells, with only rare EC, G, or P cell tumors. Poorly differentiated neuroendocrine carcinomas of the GI tract are highly malignant tumors.43-50 In general, the cytosol markers NSE and PGP 9.5, and the small synaptic-like vesicle marker synaptophysin, are expressed in the poorly differentiated tumors in contrast to welldifferentiated neuroendocrine tumors. Chromogranin A and tissue-specific hormones are generally absent or sparse in poorly differentiated neuroendocrine carcinomas. Hor­ mones and chromogranin A are stored in large dense core vesicles, which are rarely observed in poorly differentiated gastric and intestinal tumors.

MOLECULAR GENETICS Studies have shown that development of various types of GI NETs might involve different genes associated with distinct abnormalities, including point mutations, dele­ tions, methylations, and chromosomal losses and gains

(see Chapter 3). The menin gene, a tumor suppressor gene, encodes a protein of 610 amino acids; mutations of this protein cause most cases of MEN-I and a smaller proportion of sporadic and gastric duodenal endocrine tumors.51 Menin is mainly a nuclear protein, but in dividing cells, it interacts in the cytoplasm with several proteins that control tran­ scription, regulation of genome stability, and cell division. Small intestinal carcinoid tumors show deletions on chro­ mosome 18.52,53 Recent studies have also demonstrated over­ expression of the neoplasia-related genes NAP1L1 (mitotic regulation), MAGE D2, and MTA1. These genes are thought to regulate the malignant potential of these tumors and their propensity to metastasize.54 Hindgut neuroendocrine tumors express receptors for TGF-α and EGF.51

CLASSIFICATION AND SUBTYPES Carcinoid tumors were previously classified according to their embryologic region of origin into foregut, midgut, or hindgut tumors. This classification has now largely been abandoned and a WHO classification system is generally accepted. It has been recently updated by the European Neuroendocrine Tumor Society to a TNM and grading system (see Table 31-1).4,5

ESOPHAGUS

Carcinoid tumors of the esophagus are rare.55 However, in a series from Japan, esophageal carcinoids constituted 27% of all GI carcinoids, more frequent than ileal and appendi­ ceal carcinoids.42 The male-to-female ratio was 3.3, indicat­ ing a male preponderance for this type of carcinoid tumor. Dysphagia is the most common presenting symptom and is usually localized to the distal portion of the esophagus. Esophageal carcinoids may occur in conjunction with ade­ nocarcinomas arising from Barrett’s esophagus.56

STOMACH

A trend for an increased incidence of gastric neuroendo­ crine tumors ha been noted in surgical and endoscopic series. In the Surveillance, Epidemiology, and End Results (SEER) database, there was an increase in gastric neuroen­ docrine tumors, from 2.4% to 8.7% of all GI carcinoids, between 1950 and 1999.12 In mainly endoscopic series, much higher relative incidences were reported, ranging from 11% to 41% of all gastrointestinal neuroendocrine tumors.57 However, the increased relative incidence of gastric neuroendocrine tumors must be put into the context of a wider incremental trend for all types of gastrointestinal well-differentiated neuroendocrine tumors (carcinoids). Gastric neuroendocrine tumors are currently classified as well-differentiated and poorly differentiated lesions on the basis of the differentiation status of the tumor cells. The most common types are tumors derived from histamineproducing ECL cells in the corpus-fundus region, followed by gastrin producing G cell tumors of the antrum.57-59 Rare types of gastric carcinoid are those producing gastrin, ghrelin, and serotonin.43,44 Three clinicopathologic subtypes of ECL cell tumors are recognized (Table 31-2). The type I ECL tumor is associated with diffuse corpus-restricted chronic atrophic gastritis (see Chapter 51). Type II is associated with MEN-I, ZES, and hypertrophic gastropathy. Type III tumors are sporadic are not associated with any distinctive gastric pathology.43,58,59 Types I and II tumors share in common hypergastrinemia, whereas type III tumors are independent of any overt hor­ monal imbalance. Type I ECL tumors account for the largest

Chapter 31  Gastrointestinal Carcinoid Tumors and the Carcinoid Syndrome Table 31-2  Types of Gastric Carcinoid Tumors PARAMETER

TYPE I

TYPE II

TYPE III

Relative frequency (%) Gender Cell of origin Number of tumors Size of tumors Underlying gastric histology Antral G cell hyperplasia Hypergastrinemia ECL cell hyperplasia Treatment

80 F>M ECL One or more Small (<1 cm) AMAG Present Present Present <1 cm, endoscopic removal; 1-2 cm, endoscopic or surgical removal; >2 cm, surgical removal Rare

≈6 F=M ECL One or more Small (<1 cm) Hypertrophic gastropathy Absent Present Present <1 cm, endoscopic removal; 1-2 cm, endoscopic or surgical removal; >2 cm, surgical removal Uncommon

≈14 M>F EC (usually) One Large (>2 cm) Normal, nonspecific Absent Absent Absent Surgery with removal of regional lymph nodes Common

Excellent

Excellent

Variable

Metastases to liver, lymph nodes Survival

AMAG, autoimmune metaplastic atrophic gastritis (see Chapter 51); EC, enterochromaffin; ECL, enterochromaffin-like.

fraction of well-differentiated neuroendocrine tumors of the stomach, are especially prevalent in older women, and are associated with antral G cell hyperplasia. Frequently, mul­ tiple and multicentric lesions are present and are generally small and limited to the mucosa or submucosa. Metastases are rare and survival is excellent.45 Type II ECL tumors are rare and account for only 6% of gastric carcinoids. Type II tumors arise in adult patients of both genders who have hypergastrinemia, hypertrophic hypersecretory gastropa­ thy, and ECL cell hyperplasia. The ECL tumors are often multiple, multicentric, small in size, and limited to mucosa and submucosa. Despite metastases to local lymph nodes, the survival is excellent and tumor-related death is rare.57 Type III ECL cell tumors are usually single, isolated growths arising in the stomach, without any significant underlying gastric pathology. In one series, they accounted for 14% of all gastric carcinoids. They are more common in men, usually in their sixth decade, and without hypergastrinemia and gastrin-dependent ECL cell hyperplasia. The tumor size may be significantly larger than in types I and II (mean, 3.2 cm), with invasion of the stomach wall and metastases in more than 50% of patients.57 Predictors of malignancy in well-differentiated gastric carcinoids include size, histologic grading, mitotic count and Ki67 index, and p53 overexpression. Poorly differentiated endocrine tumors of the stomach are aggressive, usually large carcinomas and develop with no site predilection in the stomach. They usually occur in patients in the sixth to seventh decade of life. These poorly differentiated tumors are often metastatic at time of diagnosis; their prognosis is invariably poor.

PANCREAS

The differentiation between pancreatic carcinoid and other pancreatic neuroendocrine tumors is primarily a matter of definition (see Chapter 32 for a more detailed discussion). Maurer and colleagues have defined a pancreatic carcinoid as a tumor with typical histologic features of an NET along with evidence of increased serotonin metabolism.60 Using this definition, they found only 29 cases in the literature between 1966 and 1996. In a more recent publication from Japan, Soga found 156 cases of pancreatic carcinoids among 11,343 cases of neuroendocrine tumors (1.4%).61 These tumors were characterized by a high metastatic rate (66%), a large tumor size (averaging almost 7 cm), and a relatively high incidence of the carcinoid syndrome (23%). Serotonin was detected by immunohistochemical methods in 93% of cases. The five-year survival rate was extremely low (29%)

Table 31-3  Types of Duodenal and Ampullary Carcinoid Tumors cell of origin/ TUMOR TYPE

RELATIVE FREQUENCY (%)

Gastrin (G) cell* Somatostatin (D) cell Gangliocytic paraganglioma EC, others

50-60 15-77 6-9 <5

*Functional (Zollinger-Ellison syndrome) or nonfunctional. EC, enterochromaffin.

when compared with small intestinal carcinoid (82%). Pan­ creatic carcinoids, like other pancreatic tumors, tend to present later when compared with carcinoids in other loca­ tions. Abdominal pain, diarrhea, and weight loss are the most common presenting symptoms.

DUODENUM AND AMPULLA OF VATER

About 4% of all gastrointestinal carcinoids occur in the duodenum and duodenal carcinoids account for 11% of all intestinal carcinoids. The relatively frequency of duodenal carcinoids increased from 3.6% to 16% of all intestinal carcinoids from 1973 to 2002. The annual incidence of duodenal carcinoid is 0.07/100,000. It is higher in males than in females and is higher in blacks than in whites.10-12 Duodenal carcinoids present at a mean age of 48 to 62 years. Most duodenal carcinoids give rise to symptoms related to local growth, such as obstruction, jaundice, abdominal pain with or without pancreatitis, gastrointestinal bleeding, nausea, and vomiting.35,62 A minority of patients with duo­ denal carcinoids (<10%) present with symptoms and/or signs of hormone overproduction. ZES occurs in approxi­ mately 10% of patients with duodenal carcinoids (see Chapter 32). The carcinoid syndrome may occur (4%) and, in rare cases, Cushing’s syndrome and acromegaly may be seen. Duodenal carcinoid tumors are mainly located in the first and second parts of the duodenum. An especially high fre­ quency of tumors (25%) is found in the periampullary region. The size of the primary tumor ranges from 0.1 to 4 cm, with an average size of 1.2 to 1.8 cm. Gastrin (G) cell tumors represent the most frequent histo­ pathologic type, accounting for 50% to 60% of all duodenal carcinoids (Table 31-3). G cell tumors can be nonfunction­ ing or functioning (Zollinger-Ellison syndrome) and may be

479

480

Section IV  Topics Involving Multiple Organs associated with the MEN-I syndrome.36,62 Somatostatin cell tumors represent the second most frequent histopathologic type (15% to 27%).38 Somatostatin cell tumors are preferen­ tially localized to the periampullary region. They are identi­ fied by their content of somatostatin, but a subset of tumor cells may also contain calcitonin, pancreatic polypeptide, and ACTH. Gangliocytic paragangliomas represent the third most fre­ quent histopathologic type and account for 6% to 9% of duodenal carcinoids.39,40,62 The tumors are preferentially located in the periampullary region and are not associated with familial syndrome. The tumors are regarded as benign hamartomatous tumors, derived from the ventral pancreatic primordium. EC cell tumors of the duodenum are rare, and have the same histopathologic features as ECL cell tumors in the lower jejunum and ileum. The overall five-year survival rate for duodenal carcinoids is reported to be 84%.10-12 However, survival rates vary with histopathologic type, extent of disease, presence of hor­ monal syndrome, and genetic background. In patients with Zollinger-Ellison syndrome, duodenal localization of the primary tumor carries a much better prognosis than pancre­ atic localization (see Chapter 32), with 10-year survival rates of 94% and 55%, respectively.10-12,62

SMALL INTESTINE

Forty-four percent of all gastrointestinal carcinoids arise in the small intestine. The ileum is the most frequent location for gastrointestinal carcinoids.10-12,14 The annual incidence of intestinal carcinoids is 0.63/100,000 and 0.04 and 0.31/100,000 for jejunal and ileal carcinoids, respectively. The incidence of intestinal carcinoids has increased fourto fivefold during the past 30 years. Small intestinal car­ cinoids result from malignant transformation of EC cells. A con­siderable percentage (17% to 29%) of the patients develop noncarcinoid GI and non-GI tumors, mainly adenocar­cinomas of the gastrointestinal tract, suggesting a common pathogenic mechanism for carcinoid and non­ carcinoid tumors.23 Familial occurrence of intestinal car­ cinoid is rare, although a few families have been reported. Jejunal and ileal carcinoids are also rarely associated with MEN-I.27 Patients with carcinoid tumors of the jejunum and ileum have a mean age of 55 to 63 years at the time of diagnosis.15,21 Usually, there is a delay of four to five years for diagnosis of this tumor.17 The most frequent presentations are abdominal pain (16% to 41%) and episodic small bowel obstruction (24% to 41%), which are usually of long duration. Diarrhea, gastrointestinal bleeding, flushing, and weight loss are presenting symptoms in 4% to 30%. The carcinoid syndrome is encountered in 5% to 18% of patients.16,63

Carcinoid tumors occur all along the jejunum and ileum, but increase in frequency distally, with their highest fre­ quency in the terminal ileum.12,13 There is a single primary tumor in 74% of patients, whereas multiple tumors occur in 26% of patients.26 Clonality studies have indicated that multiple tumors are generated by the metastasis of a single primary tumor to different locations in the intestine.64 The size of the primary tumor ranges from 0.3 to 5.5 cm (mean, 2.5 cm). Intestinal carcinoids are usually associated with pronounced fibrosis and tissue scarring, which may cause intestinal obstruction and infarction. The desmoplastic stromal reaction surrounding intestinal carcinoids has been attributed to secretion of hormones and growth factors by tumor cells. Small intestinal carcinoids may produce multiple hormones. Most secrete serotonin and tachykinins (substance P and neurokinin A). Production of other hor­ mones, such as gastrin, glucagon, cholecystokinin, calcito­ nin, somatostatin, and ACTH, can be demonstrated in less than 5% of cases.7,63 Data from the SEER registry for period 1973 to 2002, sum­ marized in Table 31-4, reveal an overall five-year survival rate for intestinal carcinoids of 63% and a five-year survival rate of 84%, 72%, and 43% for localized, regional, and distant (metastatic) disease, respectively.10,11,16 However, five-year survival rates of 69% have been reported for patients with intestinal carcinoids metastatic to the liver who were given active interventional treatment.65 Risk factors for mortality include distant metastases, carcinoid syndrome, and female gender.66 Histopathologic markers shown to be associated with poor prognosis are a solid growth pattern and a Ki67 index above 1%.67

APPENDIX

The appendix is one of the most common sites in the GI tract for the development of a carcinoid tumor. The fre­ quency of carcinoid tumors in nonselected series of appen­ dectomy specimens ranges from 0.2% to 0.9%, but more recent data have suggested that the overall trend is toward a decreased prevalence.68,69 It is reported that appendiceal carcinoids occurs twice as often in women as in men. This observation may partly reflect the frequency of incidental appendectomy during gynecologic and gallbladder surgery but the female-to-male ratio is 2 : 1, even after correction for these factors. Carcinoid tumors of the appendix have been classified based on their histogenesis, cell types, and histologic pattern, as well as their clinical behavior. All typical appen­ diceal carcinoid tumors are clinically silent and are discov­ ered incidentally during surgery performed for symptoms of acute appendicitis or during incidental appendectomy. Most typical appendiceal carcinoids occur in the tip of the organ. They are usually small (<1 cm) and rarely exceed

Table 31-4  Five-Year Survival Rate (%) for Small Intestinal Neuroendocrine Tumors According to Tumor Stage* SITE Small intestine Duodenum Jejunum Ileum Overlapping lesion Site not specified

ALL

LOCALIZED

REGIONAL spread

DISTANT spread

UNSTAGED

64 64 58 66 73 60

74 73 73 77 80 71

73 66 62 75 84 73

45 32 43 49 57 42

55 50 60 71 NA 45

*SEER, 1973-2004. NA, not available. Adapted from National Cancer Institute. Surveillance Epidemiology and End Results (SEER) database, 2007. Available at http://seer.cancer.gov; National Cancer Institute. CRISP database, 2007. Available at http://crisp.cit.nih.gov; and Modlin IM, Champaneria MC, Chan AK, Kidd M. A three-decade analysis of 3,911 small intestinal neuroendocrine tumors: The rapid pace of no progress. Am J Gastroenterol 2007; 102:1464-73.

Chapter 31  Gastrointestinal Carcinoid Tumors and the Carcinoid Syndrome although the staining is often focal and always much less extensive than in classical carcinoid tumors. The most common and distinctive feature is mucin-containing gobletshaped epithelial cells arranged in round or oval clusters.73,74 The overall five-year survival rate for different subtypes of goblet cell carcinoids is 42%, with a three-year survival rate of only 17% for the most malignant subtypes.

2 cm in diameter. Metastases are rare in typical appendiceal carcinoid.68,69 The overall prognosis for EC cell appendiceal tumors is also favorable, with a 10-year disease specific survival of more than 98%. The incidence of metastases is less than 1% if tumor size is between 1 and 2 cm. Thus, appendectomy alone is likely to be adequate treatment for the most typical appendiceal carcinoid tumors that are smaller than 2 cm.70,71 Conversely, right hemicolectomy should be considered in selected patients with a carcinoid more than 2 cm in diameter, with a tumor that extends into the muscularis propria, and those with a positive resection margin, associated perforation, or evident lymph node metastases. Typical appendiceal carcinoids have the best prognosis of all types of carcinoids, which probably reflects the anatomic site and biologic behavior of the tumor itself; early detection and removal are beneficial. In contrast to typical carcinoid tumors of the appendix, goblet cell carcinoids of the appendix (Fig. 31-4) have a mixed phenotype, with partial neuroendocrine differentia­ tion and intestinal-type goblet cell morphology.72 The mean age at diagnosis is 50 years, with a range of 29 to 80 years. Goblet cell carcinoids constitute a family of appendiceal carcinoids with various morphologic features and malig­ nant behavior. Some of them are histologically aggressive neoplasms, such as adenocarcinomas of the GI tract. The tumors demonstrate neuroendocrine differentiation, with positive staining for chromogranin and synaptophysin,

A

C

COLON

Colonic carcinoids account for about 5% of all carcinoid tumors.10,11 The average age at diagnosis is 65 years, similar to other colonic cancers. There is a female preponderance. These tumors occur more commonly on the right side of the colon. Colonic carcinoids also tends to present as a larger lesion than most other carcinoids, with an average diameter of approximately 5 cm. Symptoms of colonic carcinoids are usually seen in bulky, advanced lesions. They may cause malaise, anorexia, and weight loss before localizing symp­ toms are evident. Advanced lesions may cause abdominal pain or colonic obstruction. Approximately one in three patients present with distant metastatic disease.75-77 The overall five-year survival rate is 42%.

RECTUM

The rectum is the second most frequent GI site for NETs (27.4%). Rectal NETs, however, only comprise 1% to 2% of all rectal tumors and exhibit the most benign clinical

B

Figure 31-4.  Goblet cell carcinoid of the appendix. A, The wall of the appendix is involved with a goblet cell carcinoid tumor. (Hematoxylin and eosin, ×40.) B, Mucin-positive goblets cells are seen within the wall of the appendix. (Mucin stain, ×20.) C, Positive immunostaining of a goblet cell carcinoid for chromogranin A. (×250.) (A, B, Courtesy of Dr. Pamela Jensen, Dallas, Tex.; C, courtesy of Dr. Laura H. Tang and Dr. David Klimstra, New York.)

481

482

Section IV  Topics Involving Multiple Organs profile of neuroendocrine tumors, possibly reflecting their early diagnosis by endoscopic examination. The annual increase in incidence of rectal NETs in the United States over the last 30 years is 8.6%. It has increased from 0.12/100 000 in 1973 to 0.93/100 000 in 2004.10,11 The reason for this rapid increase is unknown, but an increased awareness and wider use of endoscopy are believed to play a major role. The average age at diagnosis of rectal carcinoids is 48 to 52 years,75 with an equal gender distribution. Approximately 50% of patients with rectal NETs are asymptomatic at presentation. Clinical symptoms, when present, include weight loss, constipation, and changes in bowel habits, including diarrhea. Rectal pain may occur as component of late presentation.75,78,79 Rectal NETs often stain for glucagon, pancreatic polypeptide, serotonin, peptide YY, and somatostatin.24 Overall, rectal carcinoids fall into two groups, small solitary tumors, measuring less than 1 cm and larger lesions with the possibility of metas­ tases. Rectal carcinoids present with metastases in only 4% to 18% of cases. Rectal tumors more than 1 cm in diam­ eter metastasize more often than smaller tumors, and those larger than 2 cm have a high rate of metastases (60% to 80%). The five-year survival rate for rectal carcinoids is 87% for localized disease, 41% for regional disease, and 25% for distant (metastatic) disease.12,16

THE CARCINOID SYNDROME The classic carcinoid syndrome is relatively uncommon in patients with carcinoid tumors, occurring in 10% to 15% of these patients. Symptoms include diarrhea, cutaneous flushing, bronchoconstriction, and right-sided heart failure. Flushing attacks occur in 23% to 65% of patients with car­ cinoid syndrome at their initial presentation and in 63% to 78% at some time during the disease course. The typical flush is the sudden appearance of a deep red erythema of the upper part of the body, primarily the face and neck. Flushes are often associated with an unpleasant feeling of warmth, occasionally with lacrimation, itching, palpitation, facial or conjunctive edema, and diarrhea (Fig. 31-5). Flushes may be spontaneous or precipitated by stress, alcohol, intake of foods such as cheese and spicy meals, exercise, or by injection of agents such as catecholamines, calcium, or pentagastrin. Flushes may be brief, lasting 2 to 5 minutes, especially initially, or they may be prolonged and last for hours, especially later in the course of the disease.80,81 Flushes are usually seen with carcinoid tumors of the ileum or jejunum, but can also occur in some patients with gastric or duodenal tumors. The flush associated with gastric carcinoids is also reddish, but is patchy in distribu­ tion over the neck and face and is sometimes referred to as geographic flushing. It is often provoked by food intake. The flushing is frequently associated with pruritus, often related to release of histamine from the ECL cell gastric carcinoid. The classic flushing seen in ileojejunal carcinoids is related to the release of amines such as serotonin and peptides such as bradykinin and tachykinins (neurokinin A, substance P).80-83 Diarrhea is present in 32% to 73% of patients initially and in 67% to 84% at some time during the disease course. Diarrhea usually occurs with flushing, but it may also occur alone. The diarrhea is described as watery and the number of stools ranges from 2 to 20 daily; 60% of patients have a fecal output of less than 1 L/day. Steatorrhea is present in approximately 60% of cases. Abdominal pain may be present with the diarrhea or be present independently. Diar­

Figure 31-5.  Typical carcinoid flushing involving the face and neck in a patient with small intestinal carcinoid and the carcinoid syndrome.

rhea is usually related to the secretion of serotonin in classic midgut carcinoids; it can also be related to the secretion of calcitonin and prostaglandins.84-86 Carcinoid heart disease occurs in about 50% of patients with the classic carcinoid syndrome.87 In a retrospective study of 200 patients, carcinoid heart disease was diagnosed approximately 1.5 years after the diagnosis of a carcinoid tumor or carcinoid syndrome.88 Carcinoid heart disease is a major cause of morbidity and mortality.80,89 Today, carci­ noid heart disease is rare, occurring in about 3% to 4% of patients with carcinoid syndrome. The striking decrease of this manifestation has been suggested to be related to the use of somatostatin analogs and interferon-α, which block the release of serotonin and tachykinins from the tumor. Typically, the carcinoid syndrome occurs when hepatic metastases from a primary gastrointestinal EC cell tumor is evident and the breakdown of hormonally active tumor products by the liver is impaired, allowing these products to reach the systemic circulation. In patients with the car­ cinoid syndrome, those with the highest levels of tachyki­ nins and serotonin in plasma and of 5-hydroxyindoleacetic acid (5-HIAA) in the urine are more likely to develop cardiac valvulopathy, indicating that one or more of these hormon­ ally active tumor byproducts are involved in the pathogen­ esis of carcinoid heart disease.90-92 Serotonin has been found to modulate cell proliferation in valvular subendocardial cells and human heart valves have been shown to express mRNA for the 5-HT 1B, 1D, 2A, and 2B receptors.93,94 Fen­ fluramine, a serotonergic drug once used as an appetite suppressant, was withdrawn from the market in 1997 because it induced a valvular heart disease similar to that seen in the carcinoid syndrome.95 TGF-b, known to affect cell growth and differentiation and to stimulate fibroblasts to produce extracellular matrix protein, is expressed in coronary heart disease lesions and is up-regulated by sero­

Chapter 31  Gastrointestinal Carcinoid Tumors and the Carcinoid Syndrome tonin.96,97 Another growth factor that may contribute to drive the fibrotic process is connective tissue growth factor (CTGF). CTGF is secreted by the EC cell and acts in concert with TGF-b to drive the overproduction of collagen.98 The cardiac disease in carcinoid syndrome is caused by fibrosis involving the endocardium, primarily on the right side of the heart, although left-sided lesions can also occur. Fibrous plaques are seen on the endocardial surface of the valvular cusps and in the cardiac chambers, as well as on the intima of the great veins and arteries, areas exposed to the highest concentration of tumor products. Proliferation of myofibro­ blasts with collagen deposition is present in all plaques.87,99 These fibrous deposits tend to cause constriction of the tricuspid and pulmonic valves, resulting in regurgitation (Fig. 31-6). Transthoracic echocardiography (TTE) is a cornerstone for establishing the diagnosis of carcinoid heart disease, and evidence of cardiac involvement can be identified by TTE in 45% to 77% of patients with carcinoid syndrome.87,90 Left-sided valvular pathology occurs in 10% to 15%, whereas myocardial metastases occur in less than 5%.100 In patients with significant right heart failure, valve replace­

ment is the treatment of choice. The use of bioprosthetic valves has now become more common and widely accepted. The advantages of bioprosthetic valves compared with mechanical ones include a better hemodynamic profile and no need for anticoagulation treatment, which represents a considerable risk for bleeding during hepatic artery embolization or surgical procedures.101 However, when using a bioprosthetic valve, there is a risk for a new fibrotic process, perhaps involving another valve replacement within 10 years. Other clinical manifestations in patients with carcinoid syndrome are wheezing or asthma-like symptoms, occurring in 3% to 18% of patients, and pellagra-like skin lesions with hyperkeratosis and pigmentation in 2% to 5% of cases.80 Rarely reported are rheumatoid arthritis, changes in mental state, and visual changes during flushing caused by vasospasm. A variety of noncardiac problems caused by increased fibrous tissue have been reported, including retroperitoneal fibrosis leading to a ureteral obstruction, Peyronie’s disease of the penis, intra-abdominal fibrosis, and occlusion of the mesenteric artery or vein. Sexual dys­ function is a common complaint by men with carcinoid syndrome.80,98

PATHOPHYSIOLOGY

Figure 31-6.  Carcinoid heart disease in a patient with small intestinal carcinoid tumor with liver metastases. Note the fibrotic thickening of the cusps of the pulmonary valve and also of the endocardium.

N

NH2 Tryptophan

Tryptophan hydroxylase

HO

–CH2–CH–COOH –

–

–CH2–CH–COOH

Symptoms of carcinoid syndrome were originally attributed to the secretion of 5-HT (serotonin) by the tumor. In a large review of 748 cases of carcinoid syndrome, 92% had increased circulating serotonin levels.102 Patients may develop a typical or atypical type of carcinoid syndrome. In patients with typical carcinoid syndrome, the conversion of tryptophan to 5-HTP by tryptophan hydroxylase is the ratelimiting step (Fig. 31-7).80,81 Once formed, 5-HTP is rapidly converted to 5-HT in the tumor by aromatic l-amino acid decarboxylase (l-Dopa decarboxylase); it is stored in the neurosecretory tumor granules or released into the vascular compartments, where most is taken up and stored in platelet granules. A small amount of 5-HT remains in the plasma. Most 5-HT in the circulation is converted by monoamine oxidase and aldehyde dehydrogenase to 5-HIAA, which is excreted in large amounts in the urine (see Fig. 31-7).103 The biochemical steps illustrated in Figure 31-7 are the typical

N

NH2

5-Hydroxytryptophan

Aromatic-L-amino acid decarboxylase

HO

–C

N 5-Hydroxytryptamine (serotonin) Monoamine oxidase HO N 5-Hydroxyindoleacetaldehyde Aldehyde dehydrogenase HO N 5-Hydroxyindoleacetic acid (5-HIAA)

Figure 31-7.  Synthesis of serotonin (5-hydroxytryptamine [5-HT]). 5-HT synthesis involves enzymatic steps for its production from tryptophan and for its degradation to 5-hydroxyindoleacetic acid (5-HIAA), which is then excreted into the urine.

483

484

Section IV  Topics Involving Multiple Organs pathways for EC cells tumors, which characteristically secrete large amounts of serotonin. Some carcinoid tumors cause an atypical carcinoid syndrome and are thought to be deficient in l- Dopa decarboxylase; thus, they cannot convert 5-HTP to 5-HT and, as a consequence, 5-HTP is secreted into the bloodstream.104 Plasma serotonin levels are normal in these patients, but urinary 5-HT levels are usually ele­ vated because some of the 5-HTP can be decarboxylated in the kidney and excreted as serotonin. The exact role of serotonin in causing the flushing in carcinoid syndrome remains unclear. Antagonists against serotonin receptor subtypes typically have no effect on the flushing. The precise mediator of the flushing in patients with carcinoid syndrome may differ, depending on the tumor type. In patients with gastric carcinoid tumors, the red, patchy, pruritic flush is thought to be caused by hista­ mine, because this type of flushing can be prevented by the use of histamine H1 and H2 receptor antagonists.105 In addi­ tion to serotonin, other candidates for mediators of flushing include the tachykinins (substance P, neuropeptide K, neurokinin A) and bradykinin.106,107 Agents that block the release of tachykinins attacks, such as somatostatin analogs, may prevent flushing attacks. Patients with carcinoid syndrome often have increased colonic motility, with a shortened colonic transit time and possibly a secretory or absorptive alteration. Serotonin may be predominantly responsible for the diarrhea in some patients through its effect on gut motility and intestinal electrolyte and fluid secretion.85,86 Serotonin receptor antag­ onists (especially 5-HT3 receptor antagonist) such as ondan­ setron relieve the diarrhea.

Table 31-5  Factors That Interfere with Determination of Urinary 5-HIAA FOOD

DRUG

Factors That May Produce False-Positive Results Avocado Acetaminophen Banana Acetanilid Chocolate Caffeine Coffee Fluorouracil Eggplant Guaifenesin Pecan l-Dopa Pineapple Melphalan Plum Mephenesin Tea Methamphetamine Walnuts Methocarbamol Methysergide maleate Phenmetrazine Reserpine Salicylates Factors That May Produce False-Negative Results None Corticotropin (ACTH) p-Chlorophenylalanine Chlorpromazine Heparin Imipramine Isoniazid Methenamine mandelate Methyldopa Monoamine oxidase inhibitors Phenothiazines Promethazine 5-HIAA, 5-hydroxyindoleacetic acid; l-Dopa, l-dihydroxyphenylalanine.

CARCINOID CRISIS

Carcinoid crisis may occur in a number of situations, but most commonly in the setting of surgical or anesthetic stress. The effect of the anesthetic agent or manipulation of the tumor intraoperatively may elicit exacerbation of the carcinoid symptoms, including profound flushing, hyper- or hypotension, tachyarrythmias, hyperglycemia, and refrac­ tory bronchospasm.108,109 Patients with known carcinoid syndrome should receive somatostatin analogs before surgery. The use of octreotide before invasive procedures is important to prevent carcinoid crisis. A bolus dose of 250 to 500 µg of octreotide should be given subcutaneously within one to two hours before the surgery. This should be followed by an IV infusion of 50 to 200 µg/hour during the procedure. The postoperative dose should be 50 to 200 µg/ hour for 24 hours, followed by resumption of the preopera­ tive treatment schedule. In patients with hypotension, it is important to realize that most pressor substances are ineffective or might aggravate hypotension by increasing serotonin and peptide release from the tumor. A combina­ tion of fluid replacement and intravenous octreotide is therefore recommended in this situation. Sometimes, gluco­ corticoid administration might even further improve the hypotension.9

DIAGNOSIS BIOCHEMICAL MARKERS Specific Markers

Carcinoid tumors can synthesize and secrete serotonin, tachykinins, prostaglandins, catecholamines, and hista­ mine. The breakdown product of serotonin, 5-HIAA, is excreted by the kidney and can be measured after a 24-hour

urine collection. Restriction of the intake of serotonin-rich food is important for standardization of reference levels (Table 31-5). Serotonin concentrations can also be deter­ mined in platelets and platelet-rich plasma. Elevated 24-hour urinary 5-HIAA levels have 23% sensitivity and 100% specificity in predicting the presence of a midgut carcinoid tumor. Measurement of urinary 5-HIAA excretion for diagnosis of carcinoid tumors is the predominant biochemical analytic procedure. In some studies, platelet serotonin levels were more sensitive than urinary 5-HIAA levels. Urinary serotonin levels are not affected by the patient’s diet, as are urinary 5-HIAA levels. Elevation of urinary 5-HIAA levels can occur in intestinal malabsorption and in a number of other conditions (see Table 31-5).110-112 Tachykinins (substance P, neurokinin A and neuropeptide K) can also be measured in the plasma.106 Histamine can be measured in gastric carcinoids and is measured as histamine metabolites in the urine (N-methylhistamine, 1-methylhistamine [MH], and 1-methylimidazole acetic acid [MIAA]).

Nonspecific Markers

Neuroendocrine cells have vesicles containing peptide hor­ mones, biogenic amines, and neurotransmitters. These ves­ icles store and release acidic, soluble secretory proteins also known as granins. The granin family consists of the classic granins—chromogranin A (CgA), chromogranin B, and sec­ retogranin II (sometimes called chromogranin C)—as well as secretogranin III (1B1075), secretogranin IV (HISL-19), secretogranin V (7B2), and secretogranin VI (NESP55).113,114 Chromogranin A is an acidic glycoprotein of 439 amino acids, with a molecular mass of 48 kd. CgA is released together with hormones on stimulation from normal endo­ crine cells and is also released from neuroendocrine tumors.

Chapter 31  Gastrointestinal Carcinoid Tumors and the Carcinoid Syndrome Various assays for measurements of intact CgA and its dif­ ferent cleavage products are currently available. Several CgA-related peptides, or cleavage breakdown products, have been identified in human tissues. Tumors can also release different molecular forms of CgA. CgA levels are increased in most patients with metastatic gastrointestinal carcinoid tumors. A significant correlation between serum CgA levels and tumor mass has been demonstrated. False elevated serum CgA levels have also been documented in patients with impaired renal and liver function because of decreased metabolism and clearance. Chronic atrophic gas­ tritis and inflammatory bowel disease are other conditions with elevated serum CgA levels.113-115 NSE is the neuron-specific isomer of the glycolytic enzyme 2-phospho-d-glycerate hydroxylase or enolase. This isomer is present in neurons and neuroendocrine cells and therefore can serve as a biochemical marker for tumors derived from these cells. Serum NSE levels are frequently elevated in patients with several types of neuroendocrine tumors.116 Like CgA, NSE is a general neuroendocrine marker that cannot differentiate between different subtypes of neuroendocrine tumors. Elevated NSE levels are, however, associated with poor tumor differentiation (high-grade tumors). Human chorionic gonadotrophin (HCG) is a glycoprotein hormone synthesized during pregnancy by the trophoblastic cells of the placenta. HCG consists of an alpha and beta subunit. The beta subunit is specific for HCG, whereas the alpha subunit is also common to the other hormones of the glycoprotein family (luteinizing hormone [LH], folliclestimulating hormone [FSH], and thyroid-stimulating hormone [TSH]). Ectopic secretion of alpha subunits is frequently encountered in neuroendocrine tumors, par­ ticularly in patients with poorly differentiated carcinoid tumors.117

TUMOR LOCALIZATION

A number of techniques have been used to determine the location of the primary tumor and tumor stage, including the following: GI endoscopy; barium radiography; ultraso­ nography; endoscopic ultrasonography; computed tomo­ graphy; magnetic resonance imaging; angiography-guided selective venous sampling for various hormones; various forms of radionuclide scanning—radiolabeled somatostatin receptor scintigraphy (SRS), positron emission tomography (PET), and iodinated metaiodobenzylguanidine (MIBG) scanning; and capsule endoscopy. Endoscopy plays a key role in the management of gastric carcinoid tumors, with its ability to identify, biopsy, and even resect the lesions. The advent of endoscopic ultra­ sound (EUS) has expanded the role of endoscopy in the management of gastric carcinoids through its ability to provide high-resolution ultrasound images of lesions within and adjacent to the GI tract.118 Endoscopy and EUS can be used to identify rectal and colonic carcinoid tumors. Video capsule endoscopy119 should be used with caution in patients with evidence of small bowel obstruction or a history of small bowel resection. Computed tomography (CT) and magnetic resonance imaging (MRI) have a crucial role to play in the management of patients with neuroendocrine tumors. The introduction of multidetector CT (MDCT) has dramatically altered the performance of CT and its display.120 The most common types of carcinoid tumors usually present as small primary tumors; CT imaging typically demonstrates a secondary feature. Liver metastases are the most frequent findings, followed by tumor-associated desmoplastic fibrosis around the primary tumor and lymph node metastases. MRI has not

been widely used for the detection of primary midgut NETs. The best sequence for demonstrating the primary tumor has been the post–gadolinium T1-weighted fat-suppressed image. Hindgut and gastric carcinoids are usually diagnosed at endoscopy, although barium studies may demonstrate a filling defect. CT and MRI can be used to stage lymph node disease and distant (metastatic) spread as part of the preop­ erative plan. The clinical workup of patients with gastric carcinoid tumors always includes a CT or MRI examination, both for initial workup and for follow-up during different treatment modalities. Gastrointestinal carcinoid tumors usually possess highaffinity receptors for somatostatin (80% to 100% of cases). Somatostatin receptors are present in the primary tumor and metastases. Five subtypes of somatostatin receptor, SST1 to SST5, have been described. Octreotide, a synthetic soma­ tostatin analog, binds with high affinity to SST2 and SST5 receptors and with lower affinity to the SST3 receptor. Studies have shown that almost all carcinoid tumors (80% to 90%) possess SST2 receptors, and 50% to 60% have SST5 receptors.121 Indium-111 diethylenetriaminepentaace­ tic acid defenyl alanyl octreotide (111In-DTPA octreotide) is a standard agent for localizing carcinoid tumors using radio­ nuclide scanning. Somatostatin SRS scanning can image the tumor in 73% to 89% of patients with carcinoids (Fig. 31-8).121-123 Numerous studies have demonstrated that SRS has high sensitivity for the localization of gastrointestinal carcinoid tumors, especially the extent of metastatic spread (see Chapter 32). In general, SRS has excellent specificity, but somatostatin receptors may be expressed in other con­ ditions, which can lead to false-positive results. Examples included granulomas in sarcoidosis and tuberculosis, lym­ phoma cells, and thyroid diseases. One way to assess the importance of SRS in the management of GI carcinoid tumors is to determine its ability to alter the clinical man­ agement of these patients. The results of eight different studies have documented the ability of SRS to alter clinical management in 21% to 53% of patients with NETs.124-126 PET using the 18F-labeled glucose analog deoxyglucose (FDG) can be combined with CT (PET-CT) and MRI (PETMRI) to image tumors.127 FDG-PET has not been shown to be useful for most neuroendocrine gastrointestinal tumors, except for the subset of tumors with a high proliferation rate and poor cell differentiation.128 Based on the amine and precursor uptake and decarboxylation (APUD) diethylene­ triaminepentaacetic acid concept, 11C-labeled and 18F-labeled l-Dopa have been used to visualize NETs by PET. In a

Figure 31-8.  Somatostatin receptor scan of a patient with a small intestinal carcinoid. The scan shows multiple liver metastases and a cluster of lymph node metastases.

485

486

Section IV  Topics Involving Multiple Organs comparative study in 17 patients with gastrointestinal carcinoids, 18F–l-Dopa PET showed a higher sensitivity (65%) than FDG-PET (29%) and SRS (57%), but morpho­ logic imaging by CT and MRI was found to be most sensitive (73%).129 PET with 11C–5-HTP showed significantly higher sensitivity than that of 11C–l-Dopa. In 38 consecutive patients with various NETs, 11C–5-HTP PET visualized tumors in 95% of patients, SRS in 84%, and CT in 79%.130 More tumors were detected by 11C–5-HTP-PET than SRS and CT. The size of the surgically removed PET-positive tumors ranged from 5 mm to 3 cm in diameter (Fig. 31-9). PET with 68Ga-labeled octreotide, with its 68-minute halflife, is available from a generator that produces the radio­ pharmaceutical independently of on an onsite cyclotron. 68 Ga-labeled octreotide has been used in a few published reports for PET imaging of NETs of the GI tract. 68Ga–DOTA (1,4,7,10-tetraaxacyclododecane–1,4,7,10-tetraacetic acid)TOC PET (Fig. 31-10) has demonstrated a higher sensitivity than SRS in GI carcinoid tumors.131 Technetium-labeled somatostatin analogs have also been evaluated. 68Ga-labeled octreotide has advantages, with fast pharmacokinetics and rapid tumor accumulation, allowing PET to be performed approximately 1 hour after injection.

Figure 31-9.  11C–5-hydroxytryptophan (5-HTP) positron emission tomograph (PET) scanning of a patient with a 5-mm duodenal gastrin-producing tumor.

Several preparations have been synthesized, such as 68 Ga–DOTA-TOC, 68Ga–DOTA-NOC, and 68Ga–DOTA-TATE. Another potential improvement in somatostatin receptor imaging with PET is the better quantification of tumors and normal tissue uptake, thereby allowing the possibility for adequate dosimetry before peptide receptor radionuclide treatment (see later).127

TREATMENT Figure 31-11 presents a treatment algorithm for patients with the more common carcinoid tumors—gastric, small intestinal, appendiceal, and rectal.

SURGERY

Surgery is the only form of curative therapy for carcinoid tumors. Unfortunately, most symptomatic patients are not candidates for curative treatment. In these individuals, the focus of therapy is palliation of symptoms and facilitating medical treatment.132 In the case of appendiceal tumors smaller than 1 cm and without metastases (the majority), a simple appendectomy is sufficient.133,134 In tumors of the appendix 2 cm or larger, a right hemicolectomy is the opera­ tion of choice. With rectal carcinoid tumors smaller than 1 cm, local resection is usually adequate and results in cure. The depth of invasion is also an important prognostic factor and should also be assessed in all tumors. If no invasion of the muscu­ laris propria is present for rectal carcinoid tumors smaller than 2 cm, local resection is adequate.135 For a rectal tumor larger than 2 cm, an abdominoperineal resection or a low anterior resection with primary anastomosis is recom­ mended by some surgeons.135 In patients with midgut carcinoid tumors, malignancy is independent of size.136 Therefore, a wide en bloc resection of the adjacent lymph node–bearing mesentery is recom­ mended for all small intestinal carcinoid tumors. If the midgut carcinoid is 2 cm or larger, a full-scale cancer opera­ tion should be carried out. In patients with type 1 or 2 gastric carcinoids, lesions smaller than 1 cm can be removed endoscopically.137 For

Figure 31-10.  68Ga-DOTA octreotide scanning of a patient with a small intestinal carcinoid. The scan demonstrates multiple liver metastases and two mesenteric lymph node metastases.

Chapter 31  Gastrointestinal Carcinoid Tumors and the Carcinoid Syndrome A

Gastric carcinoid

Type 1

Type 2

Type 3

Size <1 cm

Size >2 cm

Size <1 cm

Size >2 cm

Endoscopic excision

Resection

Endoscopic excision

Resection

B

Resection with lymph node resection

Confirmed diagnosis of small intestinal carcinoid

Localized disease

Metastatic disease

Surgery

Follow up

Low proliferating index (Ki 67 <3%)

High proliferating index (Ki 67 >10%)

SMS±IFN−α

Cytotoxic therapy

STZ+5FU or STZ+DOX Cisplatinum+Etoposide Temozolomide+Cap

Recurrence

Progression:

177Lu-DOTA-octreotate

(PRRT*) or Temozolomide+Cap+Bev

*PRRT can be considered when a high content of SSTR is present (Grade 3 or 4)

C

Typical appendiceal carcinoid

Size <2 cm*

Appendectomy

Atypical carcinoid Goblet cell carcinoid Adenocarcinoid

Size ≥2 cm*

Right sided hemicolectomy

Right sided hemicolectomy plus lymph-node resection combined with chemotherapy

*1-2 cm tumors extending to muscularis propria or located close to the base of appendix might require right hemicolectomy rather than just appendectomy Figure 31-11.  Algorithm for the treatment of carcinoid subtypes. A, Gastric. B, Small intestinal. C, Appendiceal.

487

488

Section IV  Topics Involving Multiple Organs D

Rectal carcinoid

Size <1 cm

Size >2 cm

Size 1-2 cm without infiltration or regional metastases

Endoscopic resection

Anterior resection Mesorectal excision Lymph node resection

Lymph node involvement Figure 31-11, cont’d. D, Rectal. The algorithm for small intestinal carcinoid refers to treatment of metastatic disease, the most common presentation of this subtype. Bev, bevacizumab; Cap, capecitabine; DOX, doxorubicin; 5FU, 5-fluorouracil; IFN-α, interferon-α; PRRT, peptide receptor radionuclide therapy; Lu-DOTA, lutetium177-DOTA octreotate; SMS, somatostatin analog; SSTR, somatostatin receptor(s); STZ, streptozotocin.

type 1 or 2 lesions between 1 and 2 cm in size, there is no general agreement on treatment, with some recommending endoscopic and others recommending surgical treatment. Type 3 gastric carcinoids tend to be larger and more aggres­ sive so excision and regional lymph node resection are recommended.7 Resection of hepatic metastases may be beneficial (or sometimes curative) in selected patients. For patients with symptomatic tumors treatable by hepatectomy, resection is the treatment of choice if tumors are completely resectable. The role of cytoreductive or debulking surgery in patients from whom all tumors cannot be removed is unclear. No prospective randomized trials have addressed this question, but there are a number of retrospective analyses suggesting that such an approach should be considered in selected cases. In a study of 314 patients with midgut carcinoid tumors who underwent surgery mainly to remove the primary tumor and debulk mesenteric metastases, it was concluded that surgery provided considerable symptomatic relief and improved survival.136 Resection of liver metasta­ ses may relieve clinical symptoms, an effect that may last several months. It is recommended that if more than 90% of imaged tumors can be safely removed, resection should be considered.137 Local ablative therapies can also be used in the liver to control tumor growth. In several series, radio­ frequency ablation resulted in symptom reduction and improved overall survival.132,138-140 The precise role of liver transplantation has yet to be determined. In a review of 103 patients with malignant NETs who underwent liver transplantation, the two- and five-year survival rates were 60% and 47%, respectively. However, recurrence-free survival was less than 24%. Mul­ tivariate analyses have identified age older than 50 years and transplantation combined with Whipple’s resection as adverse prognostic factors.141,142

HEPATIC ARTERY EMBOLIZATION AND CHEMOEMBOLIZATION

Transcatheter arterial chemoembolization (TACE) is an excellent method of treatment of nonresectable metastases from endocrine tumors of digestive origin and yields accept­ able results in terms of symptom control and tumor response. Chemoembolization and other ablative therapies should be largely reserved for patients with advanced neuroendocrine tumors not amenable to curative surgery. TACE can be com­ bined with systemic chemotherapy (see later) in selected patients—namely, those with hepatic metastases or other extrahepatic metastatic disease. In multiple series, chemo­ embolization with doxorubicin (Adriamycin) and strep­

tozotocin, or both in combination, gave symptomatic responses in 67% to 100% of patients with midgut carci­ noid tumors.143,144 Urinary 5-HIAA levels decreased more than 50% in 50% to 90% of patients. Inclusion criteria for chemoembolization vary but, in general, include patients with unresectable disease occupying less than 50% of the hepatic volume, a patent portal vein, near-normal liver function, a total serum bilirubin less than 2 mg/dL, and no contraindication to angiography. An uncommon model is to a mix a cytotoxic drug (e.g., doxorubicin) with iodized oil (Lipiodol), which is injected into the branches of the hepatic artery, distal to the origin of the gastroduodenal artery. This is followed by embolization with a gelatin sponge (2- to 3-mm particles or microspheres [embos­ pheres]), which are placed distally in the distribution of the hepatic artery until a marked decrease in blood flow is observed. Usually, one lobe of the liver is embolized at a time and the interval between sessions varies from one to three months. The need for a repeated session should be guided by the individual patient’s response.145 More than three or four embolizations produces a diminishing effect, depending on revascularization from the diaphragm. The most common side effects of the TACE procedure include nausea, vomiting, pain, and increased serum liver amino­ transferase levels. Procedure-related complications are rare, but may be life-threatening and include liver failure, chole­ cystitis, renal failure, and carcinoid crisis.

CHEMOTHERAPY

Chemotherapy for metastatic small intestinal carcinoid tumor has in general been disappointing. Single-agent therapy with doxorubicin, 5-fluorouracil (5-FU), dacarba­ zine, actinomycin D, cisplatin, etoposide, streptozotocin, or carboplatin has low tumor response rates of 0% to 30%.146 Furthermore, the duration of responses (when they occur) are short, usually less than one year. Combinations of che­ motherapeutic agents for metastatic carcinoid tumors has not been shown to have any clear advantaged compared with single-agent chemotherapy. For example, a combina­ tion of streptozotocin with 5-FU or doxorubicin produced response rates of 0% to 40%.147,148 Newer cytotoxic regimens with temozolomide have not improved response rates in patients who had small intestinal carcinoids, with low pro­ liferation indices (Ki67 index of 1% to 2%).149 However, in patients with high proliferation indices (Ki67 index > 15%), cytotoxic treatment (e.g., with etoposide and cisplatin) may sometimes produce a significant antitumor response, but such combination therapies should be reserved for advanced tumors with evidence of progression late in the disease course.150,151

Chapter 31  Gastrointestinal Carcinoid Tumors and the Carcinoid Syndrome SOMATOSTATIN ANALOGS

Somatostatin and its analogs are effective in controlling carcinoid-related symptoms, particularly the carcinoid syn­ drome, in up to 60% to 80% of patients. They inhibit syn­ thesis and release of tumor-produced amines and peptides and also block their effect on target tissues.152 The two clini­ cally available somatostatin analogs, octreotide and lanreo­ tide, bind to somatostatin receptor subtypes 2 and 5. They are able to control flushing and diarrhea in 50% to 80% of patients.153 Octreotide has immediate-release activity and is used at subcutaneous doses of 100 to 1000 µg, three times daily. However, long-acting prolonged-release formulations are now commonly used, such as Sandostatin LAR or Lan­ reotide Autogel. The recommended doses are 10 to 30 mg of Sandostatin LAR and 60 to 120 mg of Lanreotide Autogel once a month.154-156 Some patients develop side effects with somatostatin analog treatment (see later). Therefore, it is recommended to begin therapy with the immediate-release form of octreotide at a dose of 100 µg, two or three times daily, for four to five days before switching to a long-acting formulation. In general, these somatostatin analogs have a weak tumor­ icidal effect, and a decrease in tumor size is seen in only 3% to 8% of patients.154,157 However, somatostatin analogs appear to have a tumoristatic effect, stabilizing the growth of metastatic disease. In various studies, 30% to 60% of patients with metastatic disease have demonstrated tumor stabilization during treatment with somato­statin analogs. Several trials comparing octreotide and lanreotide have suggested a similar efficacy between the two analogs in the control of diarrhea and flushing.155,158 Treatment of nonfunc­ tioning tumors (i.e., those without any hormone-related symptoms) with somatostatin analogs is controversial. Side effects of somatostatin analog treatment are mild, with steatorrhea, borborygmus, flatulence, and abdominal pain. A few cases of hypocalcemia and bradycardia, impaired glucose tolerance, and cholelithiasis have been reported.154

INTERFERON-α

Interferon-α was introduced in 1981 for the treatment of small intestinal carcinoid tumors. In multiple studies in more 600 patients worldwide, the subjective and bioche­ mical response rates were 40% to 60%, with a significant tumor reduction in 10% to 15%.153,159,160 Recombinant interferon-α (Intron A, Roferon A) can be given in a dose of 3 to 5 MU subcutaneously three to five times per week, or as pegylated interferon (Pegintron A, Pegasys) in a dose of 80 to 120 µg once a week. The interferon dose should be individually titrated and can be monitored by monitoring the white blood cell count, which should be around 3000/mm3. Treatment with interferon-a is associated with significant side effects, including flu-like symptoms in 80% of patients; fatigue in 60% to 70%; anemia, leukopenia, and thrombocytopenia in 15% to 30%; and autoimmune thyroid disease in approximately 15%.160,161 Because of their separate tumoristatic effects and ability to control symptoms, combinations of somatostatin analogs and interferon-α have been evaluated in a number of patients with malignant carcinoid syndrome, either alone or in com­ bination with other agents. The objective response rates include mostly stabilization of tumor disease in up to 60% of patients, with reduced risk of tumor progression in patients treated with a combination.162-165

NEW BIOLOGIC AGENTS

Molecular targeted treatment has recently been applied in the treatment of gastrointestinal carcinoid tumors. These

agents include VEGF inhibitors (bevacizumab) and mTOR inhibitors (RAD001). The number of patients treated is still small, but significant antitumor responses have been obtained in 10% to 20% of patients.166 The precise role of the new agents has yet to be determined with randomized controlled trials. It is also likely that they should not be used as single agents, but in combination with cytotoxic and other agents.

PEPTIDE RECEPTOR RADIONUCLIDE THERAPY

Peptide receptor radionuclide therapy (PRRT) is a new treat­ ment modality for patients with inoperable or metastasized NETs.167 Most NETs overexpress receptors for somatostatin, mainly receptor subtype 2. Somatostatin receptor scintigra­ phy was introduced in the late 1980s and the radiopharma­ ceutical agent 111In-DTPA octreotide became the gold standard for NET staging. Radiolabeled somatostatin analogs were first tried in 1992 using high doses of 111In-octreotide. Thereafter, somatostatin peptides with higher receptor affinity were developed and conjugated with DOTA, a chelator that allowed stable labeling with the pure, highenergy beta emitter yttrium-90 and the medium-energy beta emitter lutetium-177.168,169 The overall tumor response rates with 90Y-DOTA octreotate are a partial remission in 6% to 29% and disease stabilization in 50% to 88%. For 177 Lu-DOTA octreotate, the partial remission response is 25% to 30% and the rate of disease stabilization is 30% to 40%. Time to disease progression has been very favorable using these agents, with a median overall survival from the date of first diagnosis of 10.5 years for 177Lu-DOTA octreo­ tate.170 The treatment seems to be safe, with less than 5% severe adverse events, including myelodysplasia and renal failure, mainly occurring in patients previously treated with cytotoxic agents. The precise role of PRRT in the treat­ ment algorithm will hopefully be clarified by forthcoming studies.

KEY REFERENCES

Akerstrom G, Hellman P. Surgery on neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007; 21:87-109. (Ref 132.) de Herder WW. Biochemistry of neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007; 21:33-41. (Ref 9.) Eriksson B, Oberg K. Summing up 15 years of somatostatin analog therapy in neuroendocrine tumors: Future outlook. Ann Oncol 1999; 10(Suppl 2):S31-8. (Ref 154.) Kloppel G, Anlauf M. Epidemiology, tumour biology and histopatho­ logical classification of neuroendocrine tumours of the gastrointesti­ nal tract. Best Pract Res Clin Gastroenterol 2005; 19:507-17. (Ref 27.) Kwekkeboom DJ, de Herder WW, Kam BL, et al. Treatment with the radiolabeled somatostatin analog (177 Lu-DOTA 0,Tyr3)octreotate: Toxicity, efficacy, and survival. J Clin Oncol 2008; 26:2124-30. (Ref 170.) Modlin IM, Lye KD, Kidd M. A 5-decade analysis of 13,715 carcinoid tumors. Cancer 2003; 97:934-59. (Ref 12.) Modlin IM, Lye KD, Kidd M. A 50-year analysis of 562 gastric carci­ noids: Small tumor or larger problem? Am J Gastroenterol 2004; 99:23-32. (Ref 48.) Oberg K. Chemotherapy and biotherapy in the treatment of neuroendo­ crine tumours. Ann Oncol 2001; 12(Suppl 2):S111-14. (Ref 160.) Oberg K, Eriksson B. Nuclear medicine in the detection, staging and treatment of gastrointestinal carcinoid tumours. Best Pract Res Clin Endocrinol Metab 2005; 19:265-76. (Ref 122.) Oberg K, Stridsberg M. Chromogranins as diagnostic and prognostic markers in neuroendocrine tumours. Adv Exp Med Biol 2000; 482:329-37. (Ref 115.) Rindi G, Kloppel G, Alhman H, et al. TNM staging of foregut (neuro) endocrine tumors: a consensus proposal including a grading system. Virchows Arch 2006; 449:395-401. (Ref 5.) Rockall AG, Reznek RH. Imaging of neuroendocrine tumours (CT/MR/ US). Best Pract Res Clin Endocrinol Metab 2007; 21:43-68. (Ref 120.) Ruszniewski P, Malka D. Hepatic arterial chemoembolization in the management of advanced digestive endocrine tumors. Digestion 2000; 62(Suppl 1):79-83. (Ref 144.)

489

490

Section IV  Topics Involving Multiple Organs Solcia E, Kloppel G, Sobin L. Histological typing of endocrine tumours. In: Verlag S, editor. World Health Organization histological classifica­ tion of tumours. 2nd ed. New York: Springer; 2000. p 38. (Ref 4.) Sundin A, Garske U, Orlefors H. Nuclear imaging of neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007; 21:69-85. (Ref 127.) Tang L, Shia J, Soslow RA, et al. Pathologic classification and clinical behavior of the spectrum of goblet cell carcinoid tumors of the appen­ dix. Am J Surg Pathol 2008; 32:1429-43. (Ref 74.)

Yao JC, Hassan M, Phan A, et al. One hundred years after “carcinoid”: Epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States. J Clin Oncol 2008; 26:3063-72. (Ref 14.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

32 Endocrine Tumors of the Pancreas and Gastrointestinal Tract Robert T. Jensen and Jeffrey A. Norton

CHAPTER OUTLINE General Considerations  491 Historical Aspects  491 Prevalence and Incidence  492 Origin and Histologic Features  492 Classification  493 Pathophysiology  493 Molecular Pathogenesis  494 Multiple Endocrine Neoplasia  494 Other Inherited Syndromes Associated with Pancreatic Endocrine Tumors  495 Insulinomas  495 Definition  495 Pathophysiology and Pathology  495 Clinical Features  496 Diagnosis and Differential Diagnosis  496 Treatment  498 Gastrinomas  498 Definition  498 Pathophysiology and Pathology  499 Clinical Features  501 Diagnosis and Differential Diagnosis  502 Treatment  503 Glucagonomas  505 Definition  505 Pathophysiology and Pathology  506 Clinical Features  506 Diagnosis and Differential Diagnosis  507 Treatment  508 VIPomas  508 Definition  508 Pathophysiology and Pathology  509 Clinical Features  509 Diagnosis and Differential Diagnosis  510 Treatment  510

GENERAL CONSIDERATIONS HISTORICAL ASPECTS

In 1927,1 five years after the discovery of insulin, the first pancreatic hormone-producing tumor syndrome was described in a patient with a metastatic islet cell tumor and hypoglycemia, and whose tumor extracts had hypoglycemic effects. Numerous other pancreatic endocrine tumors (PETs) have since been described (Table 32-1), with the description

Somatostatinomas  511 Definition  511 Pathophysiology and Pathology  511 Clinical Features  512 Diagnosis and Differential Diagnosis  512 Treatment  513 GRFomas  513 Definition  513 Pathophysiology and Pathology  513 Clinical Features  513 Diagnosis and Differential Diagnosis  514 Treatment  514 PPomas and Nonfunctioning Pancreatic Endocrine Tumors  514 Definition  514 Pathophysiology and Pathology  514 Clinical Features and Diagnosis  514 Treatment  515 Other Pancreatic Endocrine Tumors  515 Tumor Localization  515 Management of Metastatic Pancreatic Endocrine Tumors  519 Tumor Biology, Prognostic Factors, and Survival  519 Chemotherapy  520 Surgical Treatment  521 Hepatic Artery Embolization and Chemoembolization  521 Radiofrequency Ablation  521 Somatostatin Analogs  521 Interferon-α  521 Liver Transplantation  522 Somatostatin Receptor-Directed Radiotherapy  522 Possible New Treatments  522

of Zollinger-Ellison syndrome in 1955,2 the Verner-Morrison syndrome caused by a diarrheogenic-producing tumor in 1958,3 glucagonoma syndrome by Mallinson in 1974,4 the somatostatinoma syndrome in 1977,5,6 and GRFomas (pancreatic tumors secreting growth hormone-releasing factor) in 1982.7,8 PETs secreting adrenocorticotropic hormone (ACTH; ACTHomas) are also included because 4% to 16% of cases of ectopic Cushing’s syndrome are caused by ACTHsecreting PETs.9-11 PETs causing the carcinoid syndrome (see Chapter 31),12 tumors secreting renin and causing

491

492

Section IV  Topics Involving Multiple Organs Table 32-1  Pancreatic Endocrine Tumors HORMONE CAUSING SYMPTOMS

ANNUAL INCIDENCE (NEW CASES/YR/ MILLION POPULATION)

Insulinoma Gastrinoma, Zollinger-Ellison syndrome

Insulin Gastrin

1-2 0.5-1.5

VIPoma, Verner-Morrison syndrome, WDHA, pancreatic cholera Glucagonoma

VIP

0.05-0.2

Glucagon

0.01-0.1

Somatostatinoma GRFoma

Somatostatin GRF

Rare Unknown

ACTHoma

ACTH

PET causing hypercalcemia

PTHrP

Uncommon (4%-16% of all ectopic Cushing’s syndrome) Rare

PET causing carcinoid syndrome

Serotonin, tachykinins

<1% all carcinoids

PET secreting renin PET secreting erythropoietin PET secreting luteinizing hormone (LH) PPoma, nonfunctional PET

Renin Erythropoietin LH

Rare Rare Rare

None

1-2

TUMOR OR SYNDROME

PRIMARY SYMPTOMs OR SIGNs (%)

RATE OF MALIGNANCY (%)

Hypoglycemic symptoms (100) Abdominal pain (76), diarrhea (65), dysphagia-pyrosis (10-31) Diarrhea (100), flushing (20)

<10 60-90

Dermatitis (70-90), weight loss (66-96), diarrhea (15) Diarrhea (40-90) Symptoms of acromegaly, abdominal pain Ectopic Cushing’s syndrome

50-80

Symptoms secondary to malignant tumor, hypercalcemia Carcinoid syndrome (see Chapter 31; diarrhea, flushing) Hypertension Polycythemia Masculinization (female), loss of libido (male) None

84%

>60

>70 >30 >95% (pancreatic)

77% Unknown 100% Unknown >60

ACTH, corticotropin; GRF, growth hormone–releasing factor; PET, pancreatic endocrine tumor; PP, pancreatic polypeptide; PTHrP, parathyroid hormone–related protein; VIP, vasoactive intestinal peptide; WDHA, watery diarrhea hypokalemia achlorhydria.

hypertension,13 luteinizing hormone (LH) resulting in masculinization or changes in libido,14 erythropoietin resulting in polycythemia,15 and parathyroid hormone-related protein (PTHrP), resulting in hypercalcemia,16 have also been described. PETs secreting calcitonin17 are proposed to cause a distinct syndrome with diarrhea. However, too few cases have been well described to include this syndrome. Fur­ thermore, other causes of hypercalcitonemia, such as medullary thyroid cancer, are only associated with diarrhea in 25% to 42% of patients.18 Other functional hormonal syndromes have been described, with nonpancreatic, primarily intra-abdominal neuroendocrine and nonneuroendocrine tumors, including secretion of GLP-2 (glucagon-like peptide-2), causing intestinal villous hypertrophy (enteroglucagonomas) and secretion of GLP-1 causing hypoglycemia and delayed GI transit,19 and intestinal and ovarian tumors secreting peptide YY resulting in altered intestinal motility and constipation.20 PETs secreting ghrelin have been described21 but were not associated with acromegaly, increased serum growth hormone levels, or increased insulin growth factor-1 (IGF-1) concentrations. PETs are classified as functional if associated with a clinical syndrome caused by hormone release by the tumor or nonfunctional if not associated with a clinical syndrome caused by hormone release (see Table 32-1). In the nonfunctional category are included nonfunctional PETs (NF-PETS), which have the histologic characteristics of a PET but no associated elevation in plasma hormone levels or clinical syndrome, as well as PETs that release pancreatic poly­ peptide (PPomas), ghrelin, neurotensin (neurotensinomas), or other peptides that do not cause a distinct clinical syndrome.9,22,23

PREVALENCE AND INCIDENCE

PETs account for 1% to 10% of tumors arising in the pancreas.22,24,25 The overall prevalence of functional PETs is low, reported to be approximately 10/million (1/100,000). In contrast, the prevalence of PETs in autopsy studies is higher, 0.5% to 1.5%.26 The annual incidence of PETs is reported at 1 to 4 cases/million/year.9 Nonfunctional PETs account for 14% to 30% of all PETs in most studies, but they are as high as 60% to 80% in some studies.27 Insulinomas and gastrinomas occur with an equal annual incidence of 0.5 to 3 cases/million.28-30 VIPomas are 12.5% as common and glucagonomas 6% as common as insulinomas and gastrinomas. Somatostatinomas are very rare31,32 and the incidence of GRFomas, PETs secreting renin, erythropoietin, or LH or PETs causing hypercalcemia is unknown (see Table 32-1).

ORIGIN AND HISTOLOGIC FEATURES

PETs are often called islet cell tumors but it is unproven that they originate from the pancreatic islets.9,33,34 These tumors frequently contain ductular structures, produce hormones not normally present in the adult pancreas, such as gastrin and vasoactive intestinal peptide (VIP), and may produce multiple hormones.35 It has been suggested that these tumors represent a dedifferentiation of an immature stem cell. The finding of the ductular structures in many PETs and the budding off of endocrine cells from ductules during ontogenesis of the pancreas has led to the suggestion these tumors are ductular in origin.36 It was originally proposed that PETs might originate from cells that are part of the diffuse neuroendocrine cell system.9,34,37,38 These cells share certain cytochemical prop-

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract

A

Although some have suggested that elevations of the alpha or beta subunit of HCG or of chromogranin A may be indicative of malignancy, this is not established.9,28 Furthermore, some studies have reported that serial measurements of chromogranin levels may be useful for monitoring tumor growth39-42 or as a marker of survival.43 In an occasional patient, a second clinical hormone tumor syndrome may be present initially or develop with time.9,28 Whereas one study reported that this occurred in 7% of all patients with PETs during a three-year follow-up,9 another study35 has reported that this is a rare occurrence, occurring at a rate of 2/100 followed over ten years. However, there appears to be a high incidence of the development of Cushing’s syndrome in patients with a functional PET, especially in those with gastrinomas.44,45

CLASSIFICATION

B Figure 32-1.  A, Hemorrhagic and cystic pancreatic endocrine tumor (PET, a gastrinoma) in the tail of the pancreas (arrow), measuring 2 cm in diameter. B, Histology of gastrinoma, which is similar to other PETs. This highly vascular tumor is composed of tubules of bland endocrine cells.

erties (amine precursor uptake and decarboxylation) and the tumors have been called APUDomas.33 Ultrastructurally, the cells often have electron-dense granules and produce multiple regulatory hormones and amines, neuron-specific enolase, synaptophysin, and chromogranin A or C. These cells are thought to give rise to carcinoid tumors, medullary carcinomas of the thyroid, melanomas, and pheochromocytomas, and there are marked similarities in the histology of these tumors and PETs. Histologically, PETs consist of a relatively homogeneous sheet of small round cells with uniform nuclei and cytoplasm (Fig. 32-1). Mitotic figures are uncommon.33 Malignancy can be determined only by metastases or invasion and cannot be predicted by light microscopic or ultrastructural studies.20,28 At present, controversy still surrounds the exact cell of origin for PETs, with some recent studies suggesting they originate from islet cells and others supporting a ductular origin.34,36 Most PETs produce multiple gastrointestinal hormones, which can be localized by immunocytochemical methods.9,33 In many studies, most functional and nonfunctional PETs had cells immunoreactive to peptides that were not causing clinical symptoms.35 It is unclear why usually only one or no clinical syndrome is seen, despite the immunoche­ mical occurrence of multiple hormones.28 A functional PET syndrome should be diagnosed only if the appropriate clinical symptoms are present, not only based on immunocytochemistry. PETs frequently produce chromogranins or the alpha or beta subunit of human chorionic gonadotropin (HCG), which can be localized by immunocytochemistry or by documenting elevated circulating levels.9,28,33 Chromogranins are water-soluble acidic glycoproteins that are present in almost all endocrine or neuronal tissues.26 Plasma chromogranin A levels are elevated in more than 90% of patients with various PETs and carcinoid tumors (see Chapter 31).

PETs are classified clinically according to the functional syndrome produced (see Table 32-1). Although clinical syndromes have been attributed in some studies to patients with PETs who had elevated plasma levels of neurotensin or pancreatic polypeptide (PP), the existence of these syndromes has not been established with certainty.9,35,46 PETs can be associated with four different inherited disorders: multiple endocrine neoplasia type I (MEN-I), von HippelLindau disease, tuberous sclerosis, and neurofibromatosis-1 (von Recklinghausen’s disease).47-49 These association are important to recognize because family screening may be needed and because these PETs may have a different natural history (see later).28 It has been proposed9,33 that the terms PET and carcinoid tumor be replaced by the term neuroendocrine tumor (NET) and a new classification based on clinical and morphologic categories be used (see Chapter 31). This WHO classification classifies all NETs into well-differentiated endocrine tumors or carcinomas, poorly differentiated endocrine tumors, or mixed exocrine-endocrine tumors, which better allows comparisons of NETs from the pancreas and other gastrointestinal sites. The well-differentiated NETs of the pancreas are specifically divided into well-differentiated tumors and well-differentiated endocrine carcinomas. The welldifferentiated endocrine tumors are further divided into those with benign behavior (confined to pancreas, non­ aggressive, <2 cm in size, functional or nonfunctional), and uncertain behavior (confined to pancreas, >2 cm in size, or with angioinvasion, functional or nonfunctional). The well-differentiated endocrine carcinomas of the pancreas show low-grade malignancy, with gross invasion and/or metastases and can be functional or nonfunctional. In this chapter, the term PET will be retained because of its widespread use. Recently, for the first time, a TNM classi­ fication has been proposed for PETs and for other GI neuroendocrine tumors (see Table 31-1).50 This TNM classification is based on the level of tumor invasion, tumor size, tumor extent, and with grading using the mitotic index or the proliferative index, Ki-67. The importance of these classification systems for predicting prognosis will be discussed in a later section.

PATHOPHYSIOLOGY

In patients with functional PETs, the symptoms caused by the inappropriately released hormone are usually responsible for the initial manifestations of the disease. In general, only late in the course of the disease do symptoms from the tumor itself, such as abdominal pain, become prominent. In various series, in a small percentage of the patients with a functional syndrome in whom no tumor was found preoperatively and at the time of surgery, hyperplasia of the

493

494

Section IV  Topics Involving Multiple Organs pancreatic islets was regarded as a possible cause of the disease.9,28 Beta cell hyperplasia or nesidioblastosis, which is a subtype of beta cell hyperplasia consisting of the proliferation of islet cells from pancreatic ducts, is reported to be a cause of hypoglycemia and hyperinsulinemia in a number of infants and newborns. Recently, this condition has been recognized in adolescents and adults and occurs in 5% of patients with hyperinsulinism.51,52 It has been suggested that Zollinger-Ellison syndrome and Verner-Morrison syndrome are caused in up to 10% of cases by hyperplasia of pancreatic ducts, producing gastrin and VIP, respectively. However, this concept has not been substantiated by immunocytochemical studies and thus is not generally accepted.

MOLECULAR PATHOGENESIS

Until recently, the molecular pathogenesis of neuroendocrine tumors (NETs) (carcinoids and PETs) was largely unknown.53,54 Numerous studies have demonstrated that in contrast to most common nonendocrine tumors (e.g., colonic or pancreatic adenocarcinoma), mutations in common oncogenes (e.g., ras, fos, myc, src, jun) and common tumor suppressor genes (e.g., p53, retinoblastoma gene) are uncommon in most NETs (carcinoids, PETs).26,38,55 Recent studies have provide evidence that alterations in the MEN-I gene, p16/MTS1 tumor suppressor gene, DPC4/Smad 4 gene, amplification of the HER-2/neu proto-oncogene, increased expression of growth factors and/or their receptors (endothelial growth factor [EGF], hepatocyte growth factor, platelet-drived growth factor [PDGF]), and deletions of a possible unknown tumor suppressor gene on chromosome 1 or 3p may all be important.26,38,53-57 Alterations in the MEN-I gene occur in up to one third of sporadic (i.e., noninherited) PETs58 and alterations in the p16/MTS1 gene occur in 50% to 92% of PETs and thus may be particularly important. Alterations in the MEN-I gene are discussed in the next section. Genomic-wide allelotyping and comparative genomic hybridization studies have demonstrated that chromosomal losses (especially on 1p, 3p, 3q, 6q, 9q, 12q) and chromosomal gains (especially in 7q, 17q, 17p, 20q) frequently occur in PETs and carcinoids, but their frequency varies markedly in these two gastrointestinal (GI) neuroendocrine tumors, providing evidence that they have a different molecular pathogenesis. Gene expression profiling using microarrays has identified a large number of genes that are altered in PETs in comparison to normal islets or between PETS with different degrees of aggressiveness.26,59-62 At present, it is not clear which of the different genes altered in the different studies will have important prognostic implications or provide important insights into their pathogenesis that will lead to new treatments.

MULTIPLE ENDOCRINE NEOPLASIA There are three well-established MEN syndromes that can be distinguished by the presence or absence of PETs; medullary thyroid carcinoma, parathyroid disease, pheochromocytoma and a specific phenotype.47,49,63 Each of these syndromes—MEN type I, MEN type IIa, and MEN type IIb— has autosomal dominant inheritance. MEN-I, or Wermer’s syndrome, is considered in detail later; it is characterized by hyperparathyroidism and PETs without the presence of medullary thyroid carcinoma, pheochromocytoma, or unusual phenotype. MEN-IIa, or Sipple’s syndrome, is characterized by bilateral medullary thyroid carcinoma and pheochromocytomas (in 20% to 40%); when they occur,

they are bilateral in 70%, with hyperparathyroidism in 17% but without the occurrence of PETs or a specific phenotype. MEN-IIb includes bilateral medullary thyroid carcinoma, which often appears at an early age and appears to be more aggressive than these tumors in patients with MEN-IIa. Pheochromocytomas, when they occur, are bilateral in 70%. Parathyroid disease is seldom present in MEN-IIb, and patients have a characteristic phenotype, with multiple mucosal neuromas, frequently marfanoid habitus, puffy lips, prominent jaw, pes cavus, and medullated corneal nerves, but no PETs. The genetic defect in MEN-I is located on the long arm of chromosome 11 and is caused by mutations in a 10-exon gene encoding for a 610–amino acid protein, MENIN, a nuclear protein that interacts with the AP1 transcription factor, Jun D, nuclear factor κB (NF-κβ), pem, SMAD3, RPA2 (a DNA processing factor), FAN CD2 (a DNA repair factor), nucleoside diphosphate kinase, NM23β, and various cytoskeleton-associated proteins.64 The development of MEN-I endocrine tumors conformed to Knudsen’s9 two-hit model theory of neoplasm, with an inherited mutation in one chromosome unmasked by a somatic deletion or mutation of the other normal chromosome, thereby removing the suppressor effect of the normal gene. In PETs from patients without MEN-I, up to 90% have loss of heterozygosity on chromosome 11 and 27 % to 39% have mutations in the MEN-I gene.26,53,54,58 This suggests that sporadic PETs share a similar tumorigenesis to PETs that occur in patients with MEN-I, which principally involves deletion of a tumor suppressor gene. The MEN-II syndromes are caused by alterations in the pericentromeric region of chromosome 10 in the RET proto-oncogene, which is a 21-exon gene encoding for a tyrosine kinase receptor. Mutations in a cysteine-rich extracellular portion of the receptor primarily cause MEN-IIa, whereas mutations in the gene region encoding the intracellular catalytic core of the tyrosine kinase domain cause MEN-IIb. In patients with MEN-I, hyperparathyroidism is the most common clinical abnormality, occurring in 78% to 97% (Table 32-2).9,28,47-49 Functional PETs are the second most common clinical abnormality, occurring in 81% to 82% of patients. Gastrinomas occur in 54%, whereas insulinomas, glucagonomas, and VIPomas occur in 18%, 3%, and 3% of patients, respectively.65 Nonfunctional PETs and PPomas may be the most common PET in patients with MEN-I because they are almost always found in histologic studies.66 However, large nonfunctional PETs causing symptoms occur in only 0% to 13% in various series. Many patients without a functional PET do not routinely undergo surgical exploration,56,67,68 and imaging studies routinely miss most small PETs smaller than 1 cm.26,69 Therefore, the true incidence of asymptomatic PETs in these patients is unknown. Pathology studies9,47,48,66 have demonstrated that in almost every patient with MEN-I, the pancreas demonstrates diffuse microadenomatosis, with or without larger tumors. With immunocytochemistry, PP is most frequently seen followed by glucagon and insulin, with gastrin rarely found. These results are consistent with clinical studies that have demonstrated that gastrinomas in more than 80% of patients with MEN-I and Zollinger-Ellison syndrome are located in the duodenum.28,67,68,70,71 MEN-I is present in 20% to 25% of patients with gastrinomas, 4% of patients with insulinomas, 13% to 17% of patients with glucagonomas, 33% of patients with GRFomas, 9% of patients with VIPomas, and 7% with somatostatinomas.9,27,28,47,48,72-78 Characteristically, hyperparathyroidism is the initial manifestation of MEN-I, usually presenting in the third decade of life, followed by the development of a PET

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract Table 32-2  Frequency of Features of Patients with Multiple Endocrine Neoplasia Type I

feature Hyperparathyroidism Pancreatic endocrine tumors   Nonfunctional or PPomas   Gastrinomas   Insulinomas   Glucagonomas   VIPomas   Somatostatinomas   GRFoma Pituitary tumors   Prolactin-secreting   Growth-hormone secreting   Cushing’s syndrome Adrenal tumors   Cortical adenomas   Hyperplasia, carcinoma (uncommon) Carcinoid tumors   Gastric (ECLoma)   Lung   Thymic Skin tumors   Angiofibromas > collagenoma > café-au-lait macules > lipomas CNS tumors—meningiomas, ependymomas, schwannomas Smooth muscle tumors— leiomyomas, leiomyosarcomas Thyroid tumors—adenomas

AVERAGE FREQUENCY (RANGE), % OF ALL PATIENTS 97 (78-100) 81-82 80-100 (microscopic), 0-13 (symptomatic) 54 (20-61) 18 (7-31) 3 (1-6) 1 (1-12) 0-1 <1 54-65 (15-100) 15-46 6-20 16 27-36 (symptomatic, <2%)

7-35 (symptomatic, <5%) 0-8 0-8 40-100 (angiofibromas) 88, 72, 38, 34 (symptomatic, <1%) 0-8, 0-1 (symptomatic, <1%) 1-7 (symptomatic, <1%)

of patients with VHL, a PET is seen which is usually asymptomatic and nonfunctional (>98%), although occasional insulinomas and VIPomas are described. The mean age at diagnosis of a PET in VHL is 29 to 38 years. Most patients have a single PET (67% to 70%). Malignant PETs occurs in 8% to 50% of VHL patients with PETs, with liver metastases in 9% to 37%. NF-1 is caused by a defect on chromosome 17q11.2 encoding for a 2845–amino acid protein, neurofibromin, which functions as a ras signaling cascade inhibitor.9,81 In various series, from 0% to 10% of NF-1 patients develop a carcinoid tumor, usually in the periampullary region (54%) of the duodenum (see Chapter 31).78,82 Most of these duodenal carcinoids are somatostatinomas by immunocytochemistry, but they rarely produce the somatostatinoma syndrome.47 NF-1 has rarely been associated with Zollinger-Ellison syndrome and insulinomas.46 NF-1 accounts for 48% of all duodenal somatostatinomas and approximately 25% of all ampullary carcinoid tumors. These tumors frequently (63%) show psammoma bodies histologically, and metastases to liver and or lymph nodes occur in 30%. Tuberous sclerosis is caused by mutations in the 1164– amino acid protein, hamartin (TSC-1), or the 1807–amino acid protein, tuberin (TSC-2).47,83 These two proteins are important in regulating the PI3K signaling cascade (see Chapter 3) and also for regulation of the small GTPase, Rheb, which play important roles in the regulation of protein translation and synthesis, growth, and proliferation, as well as maintenance of cellular energy levels. A few cases of nonfunctional and functional PETs (insulinomas and gastrinomas) have been reported in patients with tuberous sclerosis.26,46

0-10 (0-30; symptomatic, <1%)

CNS, central nervous system; ECL, enterochromaffin-like cell; GRF, growth hormone-releasing factor; PP, pancreatic polypeptide; VIP, vasoactive intestinal peptide. From references 28, 46-49, and 65.

in the fourth to fifth decade.49 It is important to recognize whether a patient has MEN-I because patients with and without MEN-I differ in their clinical presentation, in the possibility of surgical cure, and in the clinical and diagnostic approach to the tumor.56,67,68 Patients with MEN-I may develop more than one PET over time so long-term followup will differ from that of a patient without MEN-I. Screening of other family members will be indicated in patients with MEN-I, whereas it will not be in patients with sporadic disease. In some PETs, the presence of the hypercalcemia caused by the hyperparathyroidism may affect release of the hormones by the tumor.79

OTHER INHERITED SYNDROMES ASSOCIATED WITH PANCREATIC ENDOCRINE TUMORS Three inherited phacomatoses have an increased occurrence of PETs—von Hippel-Lindau disease (VHL), von Recklinghausen’s disease (neurofibromatosis-1 [NF-1]), and Bourneville’s disease (tuberous sclerosis).26,46,47 VHL is caused by a defect on chromosome 3p25 encoding for a 232–amino acid protein, pVHL, that forms a complex with a number of proteins, including elongin B and C as well as Cullin 2, which regulate ubiquitin-dependent proteolysis of large cell proteins.80 VHL mutations result in altered transcriptional regulation, resulting in pathologic changes in angiogenic, growth, and mitogenic factors. In 10% to 17%

INSULINOMAS DEFINITION

Insulinomas are insulin-secreting tumors that primarily originate in the pancreas and cause symptoms as a result of hypoglycemia (Table 32-3).

PATHOPHYSIOLOGY AND PATHOLOGY

Insulinomas almost always (98.2%) occur in or are attached to the pancreas.9,26,29,30 An occasional insulinoma presenting as a carcinoid tumor has been reported in the duodenum, ileum, and lung, but truly ectopic insulinomas are rare (1% to 3%).84 Insulinomas are evenly distributed in the pancreas, with approximately one third in the pancreatic head, body, and tail.9,26,84-86 Insulinomas are usually small. In one large series, 5% were less than 0.5 cm, 34% were 0.5 to 1 cm, 53% were 1 to 5 cm, and only 8% were more than 5 cm.85 Insulinomas are usually solitary, with multiple tumors occurring in only 2% to 13% of cases.9,26,29,30 If multiple insulinomas are found, MEN-I should be suspected.47 Insulinomas are generally well encapsulated, firmer than normal pancreas, and highly vascular. Only 5% to 16% of insulinomas are malignant.87 Malignant tumors are generally larger, averaging 6 cm in one series, and 5% of patients have metastases at presentation.88 Metastases are usually to the liver (47%), regional lymph nodes (30%), or both. Among adults with hyperinsulinism and pancreatic islet cell disease, histologic studies have shown a solitary insulinoma in 86% of cases, adenomatosis in 5% to 15%, nesidioblastosis in 4%, and islet hyperplasia in 1%.87 Adenomatosis consists of multiple macroadenomas or

495

496

Section IV  Topics Involving Multiple Organs Table 32-3  Frequency of Symptoms and Signs in Patients with Insulinoma finding Any Time During Clinical Course Neuropsychiatric symptoms (loss of consciousness, confusion, dizziness, diplopia) Confusion or abnormal behavior Obesity Amnesia or coma Cardiovascular symptoms, palpitations, tachycardia Convulsions (grand mal) Gastrointestinal symptoms (hunger, vomiting, pain) During First Attack Neuroglycopenic symptoms   Visual disturbances (diplopia, blurred vision)   Confusion   Altered consciousness   Weakness   Transient motor defects, hemiplegia   Dizziness   Fatigue   Inappropriate behavior   Speech difficulty   Headache   Seizure   Syncope   Difficulty concentrating or thinking   Paresthesias   Memory loss   Lethargy   Stupor   Amnesia   Ataxia   Disorientation   Mental change Adrenergic symptoms   Sweating   Tremulousness   Hunger, nausea   Palpitations

FREQUENCY (%) 92 80 52 47 17 12 9

59 51 38 32 29 28 27 27 24 23 23 21 19 17 15 12 12 8 4 4 4 43 23 12 10

From references 85-87.

microadenomas and occurs especially in patients with MEN-I. A second diffuse lesion is nesioblastosis, a condition in which islet cells bud off from ductular structures and are mixed with lobular elements. This condition previously was reported almost exclusively in infants and children, but has been recognized in 5% of adults and adolescents with hyperinsulinism.9,51,89 Diffuse islet cell hyperplasia, which consists of excessive and diffuse pro­ liferation of beta cells in the islets, has been reported in adults.85 Of 1137 cases of organic hyperinsulinism, only 6% had diffuse islet cell hyperplasia and another 0.6% had both an insulinoma and diffuse islet cell disease. At present, it is unclear whether many of these cases were in fact nesioblastosis because in most cases appropriate immunofluorescence staining methods were not applied. Insulin is synthesized and stored in beta cells of the pancreatic islets.29,30 Insulin is synthesized in the rough endoplasmic reticulum as preproinsulin, from which proinsulin is liberated and transferred to the Golgi of the cell.90 Pro­ insulin consists of a 21–amino acid alpha chain and a 30– amino acid beta chain connected by a 33–amino acid connecting peptide (C-peptide). In secretory granules, a protease excises the C-peptide and thus, when secretion occurs,

the C-peptide and the double-stranded insulin molecule are released in equimolar amounts. Small amounts of intact proinsulin remain in granules and are also released; this can be detected in the plasma. Proinsulin contains the alpha and beta chains of insulin and, because most insulin antibodies used in radioimmunoassays recognize moieties on these chains, they also recognize proinsulin. Normal subjects have less than 25% of their total serum insulin as proinsulin, whereas over 90% of patients with insulinomas have an elevated proportion of proinsulin relative to total insulin.91

CLINICAL FEATURES

Insulinomas can occur at any age but are rare in adolescents, usually occurring in patients between 20 to 75 years. A large majority occur between the ages of 40 to 45 years and 60% of patients are women.9,26,29,30 Symptoms are caused by hypoglycemia (see Table 32-3) characteristically associated with fasting and thus more frequently occur when a meal is delayed, missed, or before breakfast. Symptoms may also occur during exercise. In one study,87 26% of patients had symptoms during or after an overnight fast, 27% had symptoms prior to lunch or dinner, 8% had symptoms only after a missed meal, 29% had symptoms only before lunch or dinner, and only 9% were uncertain about when their symptoms occurred. The hypoglycemia with fasting or exercise, which is characteristic of insulinomas, differs temporally from hypoglycemia, which occurs after meals (postprandial hypoglycemia). Postprandial hypoglycemia can be caused by a number of other unrelated conditions and is increasingly being reported in patients after various gastric bypass surgery procedures for obesity.92 Most symptoms of insulinomas (82% to 92% of patients)85,93 are caused by neuroglycopenia, because glucose is the main source of energy for the brain. Neuroglycopenic symptoms include somnolence, visual disturbances, irritability, abnormal behavior, confusion, amnesia, paresthesias, stupor, drowsiness, coma, and seizures. Symptoms of hypoglycemia can also be caused by catecholamine release (adrenergic symptoms) and include anxiety, palpitations, weakness, fatigue, headache, tremor, and sweating. Coma occurs in up to 53% of patients and convulsions in 12%. In one study of symptoms of a first attack, 49% of patients initially had both neuroglycopenic and adrenergic symptoms, 38% had neuroglycopenic symptoms only, 12% had adrenergic symptoms only, and 1% had no symptoms. Of the neuroglycopenic symptoms, visual disturbances (57%), confusion (51%), and altered consciousness (38%) are the most common. Of the adrenergic symptoms, sweating (43%) and tremulousness (23%) are the most common. Patients frequently learn to avoid symptoms by eating frequently and obesity may result. In one study,87 40% of patients with organic hypoglycemic were overweight. The average duration of neuroglycopenic symptoms prior to diagnosis is often prolonged, being more than three years in 25% of patients and more than five years in 20%.

DIAGNOSIS AND DIFFERENTIAL DIAGNOSIS

The key to establishing the diagnosis of insulinoma is suspecting by clinical history—that the symptoms could be caused by hypoglycemia—and establishing the relationship of the symptoms to fasting.9,29,30,89 Whipple’s triad, published in 1938, and long used as diagnostic criteria for insulinoma, was based on this association, consisting of characteristic hypoglycemia symptoms, the presence of hypoglycemia (blood sugar < 50 mg/dL), and relief of symptoms following glucose ingestion. Unfortunately, these symptoms are not specific for insulinoma.87

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract Table 32-4  Causes of Spontaneous Hypoglycemia Postprandial (Reactive) Hypoglycemia Functional: recognizable anatomic lesion Alimentary hyperinsulinism, usually secondary to previous gastric surgery such as Billroth gastrectomy Mild diabetes mellitus (usually type 2) Idiopathic Caused by specific hepatic enzyme deficiencies Hereditary fructose intolerance (in infants, children) Galactosemia (in infants, children) Familial fructose and galactose intolerance (rare) Fasting Hypoglycemia Organic hyperinsulinism: specific anatomic lesion present Pancreatic islet disease Insulinoma—single or multiple; benign or malignant Microadenomatosis with or without macroscopic adenomas Hyperplasia Nesidioblastosis Nonpancreatic tumors Severe congestive heart failure Severe renal insufficiency in patients with non–insulin-dependent diabetes mellitus Hepatic enzyme deficiencies or decreased hepatic glucose output (primarily in infants, children) Glycogen storage diseases Glycogen synthetase deficiencies Other enzyme deficiencies (e.g., fructose 1,6-diphosphatase deficiency) Endocrine hypofunction Anterior pituitary (in infants, children usually) Adrenocortical (Addison’s disease) Diffuse acquired liver disease Ethanol Severe malnutrition Sepsis Exogenous agents (factitious) Sulfonylureas, biguanides Insulin Ingestion of ackee fruits (hypoglycine) Other drugs (aspirin, pentamidine) Functional hypoglycemia with no persistent anatomical defect Autoantibodies to insulin receptor Spontaneous autoimmune anti-insulin antibody syndrome Transient hypoglycemia of infancy From references 9, 29, 30, 87, and 93.

Organic hypoglycemia is generally defined as a fasting blood glucose level lower than 40 mg/dL. In healthy individuals, after an overnight fast, plasma glucose values usually do not decrease below 70 mg/dL.87 After an overnight fast, only 53% of patients with insulinoma have a blood glucose level below 60 mg/dL and only 39% below 50 mg/dL. However, if a blood glucose determination is combined with a concomitant plasma insulin level, this insulin level will be inappropriately elevated in 65% of patients. Hypoglycemia can be classified as a fasting hypoglycemia or postprandial (reactive) hypoglycemia, of which there are a number of different causes (Table 32-4). The distinction can usually be made by a careful clinical history. Because a single overnight fasting blood sugar level, even when combined with a simultaneous plasma insulin level, does not establish the presence of fasting organic hypoglycemia in more than 35% of patients with organic hyperinsulinism, an extended fast is done with blood glucose, plasma insulin, and C-peptide levels measured at three- to six-hour intervals.29,30,89,93 Traditionally, a 72-hour fast is performed, although one study has proposed that a 48-hour fast is sufficient.94 If at any point during the fast the patient becomes symptomatic, plasma insulin and glucose values should be determined before intravenous glucose is given

and the test stopped. Within 24 hours of starting the fast, 75% to 80% of patients with an insulinoma will have symptoms and a blood sugar level lower than 40 mg/dL, by 48 hours in 90% to 98%, and by 72 hours in almost 100%.9,86 In nonobese normal subjects, serum insulin concentrations decrease to lower than 6 µU/mL when blood glucose levels decrease to lower than 40 mg/dL, and the ratio of plasma insulin (in µU/mL) to glucose (in mg/dL) remains lower than 0.3. The test is considered positive for insulinoma if the plasma insulin-to-glucose ratio is above 0.3. In some normal obese subjects, because of hyperinsulinemia caused by insulin resistance, the fasting plasma insulin to glucose ratio may be above 0.3 and therefore may mimic the pattern in insulinoma. In these patients, the fasting glucose is normal and above 0.3 and does not decrease to lower than 55 mg/dL with fasting, as occurs in patients with insulinomas. Two additional important aspects of the diagnosis of insulinoma should be remembered. First, until recently, most the serum insulin levels were determined by radioimmunoassays that in many cases, had some cross-reactivity with proinsulin. Currently, insulin-specific radioimmunoassays (either immunoradiometric assay [IRMA] or immunochemoluminescence), which have no cross-reactivity with proinsulin, are being increasingly used instead of the classic radioimmunoassay.91,93,95 Using these newer assays, insulin levels below 6 µU/mL can be found in more than half of patients with hypoglycemia caused by insulinomas. In a recent study91 of 33 patients with insulinomas and 67 controls, the most sensitive and specific criteria for diagnosis of insulinoma during a 72-hour fast was the combination of a fasting glucose level lower than 45 mg/dL with an elevated proinsulin level. In an occasional patient with classic symptoms of an insulinoma, insulin levels and the insulin-to-glucose ratio remain within the normal range when assessed by the new insulin IRMA assay.96 Such patients can be secreting proinsulin, and serum proinsulin levels should be assessed using a specific proinsulin radioimmunoassay. A number of conditions (see Table 32-4) in addition to insulinoma can cause fasting hypoglycemia, including organic hyperinsulinism caused by pancreatic islet disease, factitious use of excessive insulin or hypoglycemic agents, and autoantibodies against the insulin receptor or insulin.89,93 To differentiate these conditions, measurements of plasma proinsulin, C-peptide, antibodies to insulin, and plasma sulfonylurea levels are carried out.9,29,30,87 The plasma proinsulin level is elevated in 80% to 90% of patients with insulinoma to more than 22% of the plasma insulin level.87,91,95 In patients with surreptitious use of insulin or oral hypoglycemic agents, the proinsulin level is decreased or normal. C-peptide is released in equimolar quantities to insulin into the plasma; thus, it would be expected to be proportionately elevated with insulin in insulinomas. However, it has a much longer plasma half-life than insulin and thus the plasma level shows less fluctuation than insulin. The measurement of C-peptide has proven useful in differentiating organic hypersecretion of insulin, such as in patients with insulinoma from patients surreptitiously using insulin because commercial insulin preparations contain no C-peptide. In insulinoma, the characteristic findings are an elevated or normal plasma C-peptide concentration, whereas in patients surreptitiously using insulin the plasma insulin level will be high and the C-peptide level low. The C-peptide level does not differentiate patients surreptitiously taking oral hypoglycemic agents from patients with insulinomas, in that both have low blood sugar, elevated insulin, and elevated C-peptide levels.9 Various pro-

497

498

Section IV  Topics Involving Multiple Organs vocative tests with tolbutamide, leucine, and glucagon and secretin and suppression tests have all been described for the diagnosis of insulinoma; however, each had its limitations and are currently seldom used.

TREATMENT

Treatment of insulinoma consists of two different approaches. Initial treatment is directed at controlling the symptoms of hypoglycemia; then, after tumor localization studies, at a possible surgical cure. Tumor localization of all PETs involves similar approaches (see later). For the 5% to 13% of patients with metastatic insulinoma, chemotherapy or other therapies directed at the tumor itself may need to be considered.88,97 This latter group will be considered in a later section on treatment of metastatic PETs.

Medical Therapy

Hypoglycemia is controlled in most patients by a combination of diet and pharmacologic therapy. The use of appropriately timed feedings with a bedtime snack or midmorning, midafternoon, or 3 am snack may be sufficient to control all symptoms.29,30,87 It is generally advised not to restrict intake to rapidly absorbed carbohydrates because their ingestion may occasionally stimulate insulin secretion from the tumor. More slowly absorbed forms of carbohydrates, such as starches, bread, potatoes and rice, are preferable. During a hypoglycemic episode, however, rapidly absorbable forms of carbohydrate such as fruit juice with glucose or sucrose are preferable. Occasional patients with severe hypogly­ cemia may require the use of a continuous intravenous infusion of glucose together with an increase in dietary carbohydrates. A number of drugs have been reported to combat the hyperinsulinemia. Diazoxide, which is a nondiuretic benzothiazide analog, has potent hyperglycemic effects.29,30,98 It directly inhibits insulin release from beta cells through stimulation of α-adrenergic receptors and also has an extrapancreatic hyperglycemic effect that enhances glycogenolysis.87 The major side effects of diazoxide are sodium retention (47%) and even edema, GI symptoms such as nausea, and hirsutism. Addition of a diuretic such as trichlormethiazide, a benzothiadiazide derivative, can correct the edema as well as augment the hyperglycemic effect of diazoxide. The GI side effects can be reduced by taking the diazoxide with a meal. Diazoxide therapy should be initiated with 150 to 200 mg given in two to three divided daily doses and, if not effective, increased to a maximum of 600 to 800 mg/day. Side effects are dose-related and may limit the ability to reach maximal doses. Approximately 60% of patients will respond to diazoxide. Patients have been treated for over 20 years with diazoxide, demonstrating that it can be used long-term.88 Verapamil, propanolol, phenytoin, and glucocorticoids have been reported effective in occasional patients, but they may have only minor hyperglycemic effects and their successes are anecdotal.9 The long-acting somatostatin analog, octreotide, has controlled hypoglycemia in a number of cases of insulinoma.9,99-102 This compound has a half-life of 100 minutes, 33 times longer than that of native somatostatin, and thus can be self-administered two to four times daily subcutaneously. Octreotide controls symptoms and hypoglycemia in 40% to 60% of patients. Somatostatin analogs are thought to act primarily on PETs such as insulinomas by interacting with high-affinity somatostatin receptors on the tumor. Five subtypes of somatostatin receptors exist and octreotide (or lantreotide) has high affinity for subtypes 2 and 5, which have been shown to be present on PETs.103 The response rate of insulinomas to octreotide is likely lower than that of

other PETs because they frequently possess low levels of somatostatin receptors, whereas the other PETs possess high densities of receptors, with high affinity for these analogs, in 80% to 90% of cases. Octreotide is generally well tolerated. It is usually given in a starting dose of 50 µg two or three times daily and can be increased to dosages as high as 1500 µg/day. A rather recent significant advance is the availability of long-acting depot forms of somatostatin analogs (octreotide long-acting release [octreotide LAR] and lanreotide slow-release [lanreotide SR] or lanreotide autogel) that last two to four weeks.104 After a 30-mg dose, octreotide LAR maintains a plasma octreotide level of 1 ng/mL or higher for 25 days and therefore can be given monthly. The main side effects of octreotide treatment include gastro­ intestinal symptoms such as bloating and abdominal cramping. Long-term side effects include malabsorption, cholelithiasis, and, in an occasional patient, worsening of glucose tolerance. In addition to improving symptoms, octreotide decreases plasma insulin levels in 65% of patients. Because somatostatin analogs also decrease glucagon and growth hormone secretion, occasionally their administration may worsen the hypoglycemia.9

Surgical Therapy

Detailed tumor localization studies are important because insulinomas are frequently small and uniformly distributed throughout the pancreas and thus can be difficult to find.29,30,105 In addition, in the uncommon patient with metastatic disease, unnecessary surgery can be avoided. Localization methods will be dealt with in a separate section (see later). All authorities recommend that if metastatic disease in the liver is not present (>90% of cases), surgical exploration be indicated.29,30,84,106 A careful surgical exploration should be done using the results of the tumor localization studies by a group experienced in treating such patients and with expertise in the use of intraoperative ultrasound.84 In most studies, 70% to 97% of all patients are cured by surgery.9,26 Failure to localize an insulinoma at surgery presents a difficult problem and the role of blind distal pancreatectomy is controversial. Because insulinomas are equally distributed in the pancreas, a pancreatectomy distal to the superior mesenteric vessels gives only a 50% chance of success. To decrease the possibility of a negative laparotomy, the use of intraoperative ultrasound and preoperative venous sampling from portal venous tributaries or hepatic veins for insulin levels after intra-arterial calcium administration should detect tumors in almost every case105,107,108 and therefore obviate the need for a blind resection. Because they are almost invariably intrapancreatic and usually benign, insulinomas are increasingly being resected successfully using a laparoscopic approach if the insulinoma can be localized preoperatively.109

GASTRINOMAS DEFINITION

Zollinger-Ellison syndrome (ZES) is caused by ectopic secretion of gastrin by a PET (i.e., gastrinoma), which causes excessive gastric acid secretion, characteristically causing peptic disease (often severe) and/or gastroesophageal reflux disease (GERD) (Table 32-5). This disease was first described in 1955 by Zollinger and Ellison, surgeons at Ohio State University, in two patients with extreme acid hypersecretion and intractable peptic ulcer disease caused by a non– beta cell tumor of the pancreas that recurred with any

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract Table 32-5  Frequency of Clinical and Laboratory Features in Patients with Zollinger-Ellison Syndrome (%) Source of Study FEATURES Clinical Features Male gender Mean age at onset (yr) Mean duration of symptoms (yr) Abdominal pain Diarrhea Heartburn Nausea Vomiting Bleeding MEN-I History of confirmed peptic ulcer Abdominal perforation Esophageal stricture Laboratory Feature Fasting hypergastrinemia Positive secretin test (>120-pg/ml increase) BAO ≥ 15 mEq/hr (no gastric surgery) BAO ≥ 5 mEq/hr (previous gastric surgery)

NIH

LITERATURE

56% 41 5.2 75% 73% 44% 30% 25% 24% 22% 71% 5% 4%

44%-70% 41-53 3.2-8.7 26%-98% 17%-73% 0%-56% 8%-37% 26%-51% 8%-75% 10%-48% 71%-93% 5%-18% 4%-6%

99% 94%

96%-100% 94%

93%

44%-100%

93%

43%-100%

BAO, basal acid output; MEN-I, multiple endocrine neoplasia type I; NIH, National Institutes of Health. From references 46, 48, 65, 72, 113, 142, and 144.

gastric surgical procedure less than a total gastrectomy.2 Most clinicians now generally use gastrinoma and ZES synonymously, although it is important to remember that some pathologists continue to use the term gastrinoma to designate any tumor containing gastrin by immunohistochemistry.110 Gastrin can be detected by immunohistochemical staining in 50% of ovarian cancers, some bronchogenic carcinomas, acoustic neuromas, pheochromocytomas, colo­ rectal cancers, and other PETs; however, except for an occasional ovarian tumor, hypergastrinemia is not seen.111 Therefore, essential to establishing the clinical diagnosis of ZES gastrinoma, it is necessary to demonstrate fasting hypergastrinemia that is inappropriate.28,46,65,71

PATHOPHYSIOLOGY AND PATHOLOGY

Almost all the symptoms listed in Table 32-5 are caused by the gastric acid hypersecretion and only late in the disease course are symptoms such as pain caused by the gastrinoma per se, and then only in a small percentage of patients.28,72 Peptic ulcer disease, GERD, and diarrhea, the most frequent clinical manifestations, all disappear when gastric acid hypersecretion is controlled surgically, medically, or by nasogastric suction.18,26,110,113 The increased parietal cell mass caused by parietal cell hyperplasia driven by hypergastrinemia results in an increased maximal acid output; the hypergastrinemia also results in increased basal acid output. These are characteristic findings in ZES.113 High serum gastrin levels stimulate the growth of the gastric mucosa, resulting in large gastric folds (Fig. 32-2), with not only parietal cell hyperplasia but also proliferation of gastric enterochromaffin-like cells (ECL cells), which secrete histamine.9,112-115 The proliferation of ECL cells results in ECL cell hyperplasia, which can lead to the development of gastric carcinoid tumors (ECLomas), especially in patients with ZES and MEN-I (see Table 31-2).114-118 Recent studies have

demonstrated that in sporadic ZES patients and patients with MEN-I–ZES, more than 99% show various degrees of ECL hyperplasia,114,119 with the changes generally being much more advanced in patients with MEN-I–ZES. Furthermore, no gastric carcinoids were seen in sporadic cases, whereas 23% of patients with MEN-I–ZES had type 2 gastric carcinoids. The increased gastric acid secretion results in diarrhea because of direct damaging effects of acid on the small intestinal mucosa; furthermore, the low pH inactivates lipase and can precipitate bile acids. There is no evidence that the chronically high levels of serum gastrin directly contribute to the pathogenesis of the diarrhea by altering intestinal secretion or motility, as was originally proposed. Experimental evidence demonstrates that gastrin-like peptides can stimulate growth and development of colorectal cancers.116,118 Whether gastrin-like peptides are important to growth or development of colorectal tumors in humans is controversial. Increased proliferative rates of colonocytes and rectal mucosal cells have been reported in patients with ZES.9,116 However, epidemiologic studies in hypergastrinemic patients with pernicious anemia or ZES have not shown an increased risk of colorectal cancer.111,120 Gastrinomas were originally reported to be non–beta islet cell tumors or neuroendocrine tumors of the pancreas.2,28,121 However, in most current series, more than 50% of gastrinomas are located in the duodenum and duodenal gastrinomas outnumber pancreatic gastrinomas by two- to five-fold in different series.46,56,65,68,71,110,122-124 Within the duodenum, gastrinomas are more frequent proximally (D1, 56%; D2, 32%; D3, 6%; and D4, 6%). Within the pancreas, gastrinomas show a pancreatic head-to-body-to-tail ratio of approximately 1 : 1 : 2. The increased identification of duodenal gastrinomas accounts for the fact that 60% to 90% of gastrinomas are now found in the “gastrinoma triangle.”124-127 This is an area formed by the junction of the cystic and common bile ducts posteriorly, the junction of the second and third parts of the duodenum inferiorly, and the junction of the pancreatic neck and body medially. Gastrinomas originate in a nonduodenal-nonpancreatic abdominal location in 2% to 24% of patients in different series, including in the ovary, liver and biliary tract, jejunum, mesentery, renal capsule, omentum, and pylorus.128 Lymph node primary gastrinomas are reported in up to 11% of sporadic cases of ZES (i.e., non–MEN-I).110,129,130 The issue of lymph node primary gastrinomas remains controversial, even though patients with sporadic ZES have remained cured, with follow-up to 20 years postresection of only lymph node(s) containing gastrinoma. Pathology studies reported endocrine cell nests in lymph nodes of patients without gastrinomas that could possibly give rise to gastrinomas. Two extra-abdominal locations have been reported to be the primary site of the gastrinoma in 0.5% of patients—namely, the cardiac intraventricular septum and non–small cell lung cancer.131,132 At presentation, 70% of 221 National Institutes of health (NIH) patients with ZES had localized disease, with a primary tumor only in 36% (literature range, 23% to 51%) or with a primary tumor and lymph node metastases in 28% (literature range, 8% to 61%).110 In these 221 patients, no primary tumor was located in 13% (literature range, 7% to 48%). This failure to locate a primary tumor is unlikely to be the result of some other entity such as nesioblastosis causing hypergastrinemia and mimicking ZES, because in our surgical series at the NIH, gastrinoma was found in the last 81 patients, However, 16% of these patients only had positive lymph nodes without a primary tumor found and were not cured by resection, suggesting that a small primary

499

Symptoms such as those of peptic ulcer disease or severe GERD; chronic diarrhea signs such as prominent gastric folds on endoscopy (see left upper panel) or a radiographic study Suspect ZES (Exclude retained antrum syndrome by history) Measure fasting serum gastrin (FSG)

Normal

Not elevated (<2% ZES pts)

Elevated (>98% ZES pts)

ZES unlikely

ZES possible

(If strong clinical suspicion)

• Repeat FSG off drug (H2R antagonist × 48 hrs) (PPI × 1 week) • Measure fasting gastric pH

Secretin test and BAO Normal FSG 100 90 80 70 60

NIH (n = 309)

50

pH ≤2

pH >2

ZES possible

ZES unlikely <1%

FSG >10-fold increased

FSG elevated (1–9.9-fold)

FSG normal

ZES present

ZES likely

ZES unlikely <1%

Do secretin test Measure BAO

40 Literature (n = 2,229)

30 20

Secretin test: ≥120 pg/mL gastrin increase BAO ≥15 mEq/hr (no previous gastric surgery) ≥5 mEq/hr (previous gastric surgery)

10 0 0

1

10

100 1000 >1000

ZES present

Fasting serum gastrin (fold normal)

1000

Secretin (2 CU/kg)

900 600 300 0 0

5

10

15

20

Time (min/postinjection)

120 Basal gastric acid output (mEq/hr)

ZES

Serum gastrin concentration (pg/ml)

Section IV  Topics Involving Multiple Organs

Cumulative percent

500

No gastric surgery

110

110

100

100

90

90

80

80

70

70

60

60

50

50

40

40 30

30

20

20

10

10

0

Previous gastric surgery

n = 205

0

n = 30

Figure 32-2.  Algorithm for the diagnosis of Zollinger-Ellison syndrome (ZES). Right upper panel, Typical ZES patient with a positive secretin test result (i.e., >120-pg/mL increase in fasting gastrin level). Right lower panel, Marked elevation in basal acid output (BAO) seen in ZES patients, with or without previous gastric acid–reducing surgery. Mean + SEM (standard error of the mean) values for patients with ZES, with or without previous gastric acid– reducing surgery, are shown. Dotted horizontal lines show criteria of >15 mEq/hr or >5 mEq/hr proposed to distinguish patients with ZES with or without previous gastric acid–reducing surgery, respectively, from patients without ZES. Left upper panel, Prominent gastric folds found on endoscopy in 92% of ZES patients, as compared with a normal subject. Left lower panel, Fasting serum gastrin levels from 309 National Institutes of Health (NIH) ZES patients and 2229 ZES patients from the literature. Fasting gastrin levels are expressed as a multiple (fold) of the upper limit of normal on the horizontal axis and as the cumulative percentage of patients with the indicated gastrin level on the vertical axis. Only 0.3% of NIH patients and 3.4% of all patients in the literature had normal values; 60% of patients had a <10-fold serum gastrin increase.28,72,112,113,142,144 CU, clinical units; FSG, fasting serum gastrin; GERD, gastroesophageal reflux disease; H2R, histamine H2-receptor; PPI, proton pump inhibitor.

tumor was missed.129 Of the NIH patients, 17% had metastatic disease to the liver (literature range, 13% to 53%), with 3% (literature range, 4% to 14%) having liver metastases only found, and 14% (literature range, 4% to 23%) having lymph node metastases only. In older studies, gastrinomas were malignant in 60% to 90% of patients28,133 based on the occurrence of lymph node or hepatic metastases. These data suggest that all gastrinomas should be considered potentially malignant.110 As with other PETs, only the presence of metastases or gross invasion of normal tissues remains the generally accepted criterion for malignancy.45 At present, it is not known how frequently, if at all, hepatic metastases develop without

lymph node involvement. Studies have demonstrated that bone metastases also occur more frequently than previously reported in patients with advanced tumors; in one study,134 115 patients with ZES were examined prospectively with serial bone scans, octreotide scans, and magnetic resonance imaging (MRI) of the spine and bone. No bone metastases were seen in patients without liver metastases. Of all patients, 7% had bone metastases and 31% had liver metastases. The most frequent areas of bone involvement were the pelvis (75%), scapula, and ribs. Two large NIH studies68,135 have supported the proposal of older studies121 that gastrinomas demonstrate two general growth patterns, an aggressive growth pattern in 25% and

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract

No liver mets (n = 158)

100

Probability of survival (%)

90

Liver mets

80

PRIMARY GASTRINOMA LOCATION

Primary location Lymph Pan node Duod mets

P < 0.02

P = NS

Pan P < 0.0001

Duod

70

0

60

10

20

30

40

50

60

70

80

90 100

Percent

50

Single liver lobe mets (n = 14) P = 0.0004

40 30 20

Diffuse liver mets (n = 27)

10

PRIMARY GASTRINOMA SIZE

Primary size Lymph node mets

≤1 cm 1.1-2.9 cm ≥3 cm

Liver mets

≤1 cm 1.1-2.9 cm ≥3 cm

NS

P < 0.0001

0 0

5

10

15

20

0

10

20

30

40

50

60

70

80

90 100

Percent

Years since diagnosis

Figure 32-3.  Effect of gastrinoma extent on survival and of primary gastrinoma location and size on development of lymph node or liver metastases (mets). Left panel, Survival in 199 patients with Zollinger-Ellison syndrome, with or without liver metastases. Right panel, top, Percentage of patients with pancreatic (n = 41) or duodenal (n = 42) primary gastrinomas who developed lymph node or liver metastases. Right panel, bottom, Percentage of patients with primary gastrinoma <1 cm (n = 45), 1.1 to 2.9 cm (n = 32), or ≥3 cm (n = 41) in diameter who developed lymph node or liver metastases. (Left panel, Adapted from Yu F, Venzon DJ, Serrano J, et al. Prospective study of the clinical course, prognostic factors and survival in patients with longstanding Zollinger-Ellison syndrome. J Clin Oncol 1999; 17:615-30; right panels, adapted from Weber HC, Venzon DJ, Lin JT, et al. Determinants of metastatic rate and survival in patients with Zollinger-Ellison syndrome: A prospective long-term study. Gastroenterology 1995; 108:1637-49.)

a nonaggressive growth pattern in 75%. The 10-year survival in patients with tumors showing aggressive growth was only 30%, whereas it was 96% in those with tumors without aggressive growth. Aggressive disease was defined by the initial presence (19% of patients) or the development (5% of patients) of liver metastases. The presence and extent of liver metastases was the most important prognostic factor (Fig. 32-3).25 Particularly important predictors of the presence of liver metastases, but not lymph node metastases, were the presence of a pancreatic as opposed to a nonpancreatic gastrinoma and size larger than 3 cm. Most duodenal gastrinomas are small (92% < 1 cm) compared with only 8% of pancreatic gastrinomas. The number of large duodenal gastrinomas or small pancreatic gastrinomas was not sufficient to determine whether gastrinoma size and location are independent predictors of the development of liver metastases. The percentage of pancreatic and duodenal gastrinomas with lymph node metastases was similar. Therefore, the development of lymph node metastases was independent of primary gastrinoma location, as it was of primary gastrinoma size. The difference in biological behavior of duodenal and pancreatic gastrinomas has been used to support the hypothesis that they have different origins.136 Another important finding from the natural history studies of gastrinomas that has important clinical applications is that even in patients with gastrinomas that have metastasized to the liver, the tumor growth rate is highly variable. In one study of 19 patients with liver metastases with ZES receiving no treatment, the metastatic tumor showed no growth in 26% of patients over a 29-month period, showed slow growth in 32%, and showed rapid growth in 42%.137 In patients with rapid tumor growth, 62% died during followup, whereas none died in the no-growth or slow-growth group. Therefore, the rate of growth of the metastases is an

important prognostic factor and not all patients with liver metastases require equally aggressive antitumor treatment. The development of bone metastases or ectopic Cushing’s syndrome caused by release of ACTH by the gastrinoma is associated with a poor prognosis.25,68,110,135 Survival in patients developing bone metastases or ectopic Cushing’s syndrome was 1.9 ± 0.4 years and 1.7 ± 0.4 years, respectively, after their diagnosis.

CLINICAL FEATURES

The principal clinical features of ZES are summarized in Table 32-5.72 There is a slight male predominance, with a mean age of onset of 41 years and a mean delay in diagnosis of five to six years. Abdominal pain primarily caused by peptic ulcer disease remains the most frequent early symptom, although increasingly diarrhea or GERD symptoms are reported early in the course of the disease. The abdominal pain that patients with ZES initially develop is clinically indistinguishable from that seen in patients with other forms of peptic ulcer disease.28,110 Later, the symptoms may become persistent, refractory to antisecretory medications, or associated with complications, which may suggest the diagnosis (Table 32-6). Diarrhea initially may be the only symptom in 9% to 10% of patients and occurs with abdominal pain in 28% to 56%.46,65,72 Symptoms of GERD are being increasingly recognized, with up to 31% presenting with GERD as the initial manifestation and 49% to 61% of patients having symptoms of GERD or esophageal lesions at initial evaluation.138 Most patients today with ZES have a typical duodenal ulcer at diagnosis, and 18% to 29% have no ulcer at diagnosis. This is an important difference from older studies, in which more than 90% of patients with ZES presented with peptic ulcers, and multiple ulcers or ulcers in atypical locations were frequent.112,133 In contrast to older

501

502

Section IV  Topics Involving Multiple Organs Table 32-6  Historical Features, Symptoms, and Signs That Suggest Zollinger-Ellison Syndrome Historical Feature In Patients with PUD or GERD Diarrhea (>50%) Weight loss (20%) Long history (>5 yr) of persistent symptoms (>50%) Bleeding, perforation, penetration (15%) No helicobacter pylori (PUD: 10%-50%) Family history of PUD or gerd Endocrinopathy (25%) Refractoriness to treatment In Patients with Persistent Diarrhea Malabsorption that is unexplained Abdominal pain (55%) Esophageal disease/symptoms (45%) Weight loss (15%) Lack of response to disease-specific treatments (gluten-free or lactose-free diet, antibacterial treatment) Family history of endocrinopathies or PUD (25%) Secretory diarrhea Signs Prominent gastric folds on UGI endoscopy or X-ray (90%) Multiple peptic ulcers or ulcers in unusual locations (1%-5%) Esophageal stricture caused by PUD (3%-4%) Complicated PUD (gastric outlet obstruction, 10%; perforation, 7%; recurrent bleeding, 5%-10%) GERD, gastroesophageal reflux disease; PUD, peptic ulcer disease; UGI, upper gastrointestinal; ZES, Zollinger-Ellison syndrome. From reference 72.

studies, in which up to 100% of patients presented with or developed a complication of advanced acid peptic disease (bleeding, penetration, esophageal stricture, perforation, obstruction), at present less than 30% of patients develop these complications, even though 71% have a confirmed history of peptic ulcer disease. Patients presenting with duodenal gastrinomas do not different clinically from patients with pancreatic tumors.124 In the 25% of patients with ZES who have MEN-I syndrome, the clinical presentation is similar to that of those with sporadic disease (see Table 32-5).47,48,67,72 Some important clues that should suggest MEN-I are the following: (1) a history of nephrolithiases and/or renal colic is much more frequent in patients with ZES with MEN-I than with sporadic ZES (47% vs. 4%); (2) patients with MEN-I present with ZES at a younger age (34 vs. 43 years); and (3) 72% have a family history of endocrinopathies. In a review of 107 patients with ZES with MEN-I prospectively studied at the NIH and 1007 cases from the literature,48 88% to 94% developed hyperparathyroidism, 31% to 60% pituitary disease, 6% to 30% various carcinoids (gastric, bronchial, thymic), and 6% to 16% other functional PETs, each of which may mask the diagnosis of ZES and be the presenting feature of the syndrome.

DIAGNOSIS AND DIFFERENTIAL DIAGNOSIS

Despite the widespread availability of gastrin radioimmunoassays and the publication of more than 3300 articles on ZES, as well as most physicians’ general awareness of the features of ZES, the diagnosis of ZES continues to be delayed four to six years after onset of symptoms (see Table 32-5).46,48,72 This likely occurs for a number of reasons. ZES is an uncommon cause of peptic disease (1 to 3 new cases/ million/year). Initially, it is frequently clinically indistinguishable from patients with peptic ulcer disease (2300 cases/million/year), and GERD, which can affect 3% to 4%

of the population.28,67,112 In contrast to older studies, the initial symptoms of abdominal pain or GERD symptoms that most ZES patients currently exhibit are generally indistinguishable from those seen in peptic disease. Also, the characteristics of the ulcer itself (single, duodenal in location) do not distinguish it from those typically seen in patients with duodenal ulcer disease. Furthermore, the widespread use of proton pump inhibitors (PPIs) is complicating and delays the diagnosis of ZES. One study139 has analyzed the number of patients referred and diagnosed with ZES in two well-established referral centers (Università La Sapienza, Rome, and the NIH) before and after the introduction of PPIs. The widespread use of PPIs decreased referrals of patients with possible ZES by 62%, decreased by 40% the number of cases of ZES diagnosed, and was associated with a two- to six-fold increase in patients with a false diagnosis of ZES referred to the U.S. center. This false diagnosis occurred because chronic treatment with PPIs causes hypergastrinemia in 80% to 100% of all patients with peptic ulcer disease or GERD and it frequently reaches five times normal levels, a range seen in 60% of patients with ZES. The use of histamine H2 receptor antagonists often suggests the diagnosis of ZES because they often do not control acid hypersecretion in ZES with the conventional doses used to treat routine peptic ulcer disease or GERD. Thus, treatment failure with these drugs leads to a suspicion of the diagnosis of ZES.9 In contrast, PPIs mask the diagnosis of ZES because they control symptoms in most patients with ZES with conventional doses used in the treatment of peptic ulcer disease or GERD, and therefore a treatment failure rarely occurs with PPIs.71,140 There are a number of distinctive clinical and laboratory features that should suggest the diagnosis of ZES in a patient with acid peptic disease or diarrhea (see Table 32-6). A few important points will be briefly discussed. Diarrhea is now infrequent in patients with peptic disease or GERD because higher doses of antacids are rarely used, so its presence should lead to a suspicion of ZES because it occurs in 73% of ZES patients (see Tables 32-5 and 32-6). Diarrhea alone can be the presenting symptom in up to 27% of patients with ZES.28,72,112 MEN-I is a cause of ZES in 20% to 25% of patients and is characterized by endocrine abnormalities in multiple endocrine glands (parathyroid > pancreas > pituitary). Therefore, anyone with a personal or family history of endocrinopathies, or laboratory evidence of endocrinopathies, should be suspect of having ZES.46,48 Helicobacter pylori is an important cause of peptic ulcer disease (present in 80% to 100%); however, it is present in only 10% to 50% of patients with ZES.141 Prominent gastric folds on upper gastrointestinal (UGI) endoscopy should suggest the diagnosis of ZES (see Fig. 32-2). Whereas enlarged gastric folds were reported in only 13% to 30% in older studies of patients with ZES, an NIH prospective study72 involving 261 patients, in which all patients underwent UGI endoscopy, reported this finding in 94% of cases. This endoscopic finding particularly contrasts with the loss of gastric folds reported in many patients with acid hyposecretory disorders, resulting in fasting hypergastrinemia (e.g., atropic gastritis, pernicious anemia; see Chapter 51), and therefore should be an important clue to the possible presence of ZES. The diagnosis of ZES requires the demonstration of acid hypersecretion in the presence of hypergastrinemia (see Fig. 32-2).28,46,65,112,113,124 Therefore, to diagnose ZES, assessments of fasting serum gastrin levels and acid secretion are required. At the NIH, over 99% of 309 ZES patients studied had fasting hypergastrinemia and, in 2229 cases from the literature, 97% had fasting hypergastrinemia.142 Hence, if

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract the fasting serum gastrin level is normal, especially in repeated determinations, the diagnosis of ZES is very unlikely. There are two specific exceptions to this general conclusion. First, in MEN-I patients after parathyroidectomy, correction of hyperparathyroidism may result in normalization of fasting gastrin levels in patients with ZES.67,79,143 Second, after gastrinoma resection, the fasting serum gastrin level may be normalized, even though the patient is not cured.144 In these latter two special cases and the rare case in which ZES is suspected clinically even though the fasting gastrin level is normal, a secretin provocative test should be performed. Because PPIs can elevate fasting serum gastrin levels up to three- to five-fold, a range that overlaps with gastrin levels in 60% of patients with ZES, it is difficult to diagnose ZES when the patient is taking PPIs.9,46,71,139 Therefore, if the fasting gastrin level is elevated while taking a PPI, the fasting gastrin level should be performed after stopping the PPI for at least one week (see Fig. 32-2). However, abruptly stopping PPIs can lead to the rapid development of peptic complications in a small percentage of ZES patients. In one prospective study of acid secretion in 235 ZES patients and review of 984 cases from the literature,113 99% of patients had a fasting gastric pH lower than 2; therefore, gastric pH should be determined when the fasting gastrin is repeated (see Fig. 32-2) to exclude physiologic hypergastrinemia (i.e., caused by hypochlorhydria or achlorhydria). The most common causes of physiological hypergastrinemia include atrophic gastritis, H. pylori infection, pernicious anemia, the use of potent acid suppressant drugs (particularly PPIs), chronic renal failure, and following gastric acid-reducing surgery.28,46,65,71 In patients with fasting serum gastrin levels less than 10-fold increased (i.e., usually <1000 pg/mL) and gastric pH lower than 2, which includes 60% of those with ZES, other conditions such as gastric outlet obstruction, H. pylori infection, rarely renal failure or short bowel syndrome, antral G cell hyperfunction or hyperplasia, and retained gastric antrum syndrome can mimic ZES and need to be considered. In such patients, secretin test and basal acid output (BAO) tests should be performed.142,144,145 Historically three different gastrin provocative tests have been used for the diagnosis of ZES—secretin, calcium, and standard meal gastrin provocation tests.28,112,144,145 Both the secretin and calcium provocation studies are based on the finding that gastrinomas release an exaggerated amount of gastrin in response to these agents, which is likely caused by the presence of specific receptors for these agents on tumor cells.146,147 The secretin test is the only one widely used at present (see Fig. 32-2). Until recently, the most widely used criterion was an increase of 200 pg/mL or more following secretin.144,145 However, a recent study144 of 293 NIH ZES patients and 537 ZES patients from the literature, demonstrates that this criterion (200 pg/mL or more) does not have the highest sensitivity and specificity and therefore should be replaced. A gastrin increase of 120 pg/mL or more after secretin has a significantly higher sensitivity than the 200-pg/mL or more criterion (94% vs. 83%), with no loss of specificity (100%); thus, it is the criterion that is currently recommended.144 In a hypergastrinemic patient with gastric pH lower than 2, no false-positive secretin tests have been reported and, as noted, 94% of patients with ZES will have a 120-pg/mL or more increase in fasting gastrin with secretin provocation.28,46,65,144,145 The BAO is elevated in more than 90% of patients with ZES; it is more than 15 mEq/hr in patients without previous gastric surgery and more than 5 mEq/hr

in patients with previous acid-reducing surgery (see Fig. 32-2 and Table 32-5).113 Gastric acid secretion is now uncommonly measured. However, a recent study has reported that it can be measured during UGI endoscopic evaluation, which may facilitate its more general use.148 The results of a prospective study of gastric acid secretion in 235 patients with ZES and a comparison with 984 ZES cases from the literature provided some important insights into when ZES should be suspected.113 The BAO did not differ between patients with duodenal or pancreatic gastrinoma.124 BAO was slightly higher in patients with sporadic ZES than in those with ZES–MEN-I in one study, but not in another.67 The mean BAO for patients without previous gastric surgery was 42.3 mEq/hr (range, 1.6 to 118 mEq/hr), with 94%, 90%, and 86% having a BAO more than 10, 15, or 18 mEq/hr, respectively, criteria proposed for diagnosis of ZES in different studies (see Fig. 32-2). In patients with previous gastric acid-reducing surgery, the mean BAO was 27.6 mEq/ hr (range, 5.9 to 103 mEq/hr); 100% and 73% exceeded 5 mEq/hr and 14.4 mEq/hr, criteria proposed for diagnosis in different studies (see Fig. 32-2). These results, as well as those from 984 cases from the literature,113 confirm the reports of older, smaller studies28 that most ZES patients have marked acid hypersecretion. Therefore, prompt and effective treatment of the gastric acid hypersecretion is essential for the management of these patients46,65,124 (see later).

TREATMENT

ZES requires immediate and long-term control of the gastric hypersecretion and then treatment directed against the gastrinoma.28,46,56,65,71,122,124 Treatment for the gastric hypersecretion is required because almost every patient with ZES has acid hypersecretion and in more than 50% of cases it is more than five times normal.113 Older studies have demonstrated that if the acid hypersecretion is not controlled, complications of peptic disease almost invariably occur and can frequently be life-threatening.112,133 Furthermore, immediate treatment is needed for the hypersecretion because life-threatening complications can occur rapidly. Currently, even though most patients receive some form of gastric antisecretory treatment prior to the diagnosis of ZES, 24% develop bleeding, bowel perforations develop in 6% to 7%, and 8% to 10% develop GERD severe enough to result in strictures.9,72 Treatment also needs to be directed against the gastrinoma because its removal may result in cure,68 and 60% to 90% of gastrinomas were malignant in older studies.25,133 With increased ability to control the gastric hypersecretion medically, the natural history of the gastrinoma is becoming one of the major determinants of longterm survival.45,135

Pharmacologic Treatment of Gastric Acid Hypersecretion

Acid hypersecretion in patients with ZES can now be controlled with oral antisecretory drugs both acutely and long term in all patients, with the exception of the rare patient who will not or cannot take oral medications regularly.28,46,65,71,110,124,149 PPIs, because of their long duration of action (i.e., >48 hrs) and potency, allow once or twice daily dosing in most patients (>95%) and are now the drugs of choice.140,150 Histamine H2 receptor antagonists are also effective; however, higher than conventional doses are required (mean dosages—cimetidine, 3.6 g/day; ranitidine, 1.2 g/day; famotidine, 0.25 g/day).110 All PPIs have been shown to be effective in ZES (omeprazole, lansoprazole, pantoprazole, esomeprazole, rabeprazole).149-152 PPIs are usually started with a dosage equivalent

503

504

Section IV  Topics Involving Multiple Organs to 60 mg/day of omeprazole, except in patients with complicated ZES–MEN-I, severe GERD, or previous Billroth II resection, for whom an initial dose equivalent to 60 mg twice daily of omeprazole is recommended.138 Sufficient antisecretory drug needs to be given to reduce acid hypersecretion to lower than 10 mEq/hr in patients without previous gastric acid-reducing surgery (<5 mEq/hr in patients with previous gastric acid-reducing surgery) for the hour prior to the next dose of drug. This degree of suppression allows peptic disease and GERD lesions to heal and also prevents their recurrence. Although initial lower doses of PPIs are frequently used, because of the desirability of rapidly controlling the acid hypersecretion in ZES patients, these higher doses are recommended. This recommendation was supported by a study153 that demonstrated that a low initial dose of omeprazole (i.e., 20 mg/day) controlled acid hypersecretion within 24 hours in only 68% of ZES patients. Patients with previous Billroth II resection or moderate to severe GERD with ZES require greater acid suppression, and antisecretory drugs should be increased to control symptoms and heal all mucosal lesions, which frequently require suppression to 1 to 2 mEq/hr or even lower.154 Patients with ZES and MEN-I with hyperparathyroidism also require higher doses of antisecretory drugs. Studies have demonstrated that parathyroidectomy for hyperparathyroidism in MEN-I–ZES patients reduces the fasting gastrin level, reduces the BAO, and increases the sensitivity to a given dose of antisecretory drug, which facilitates medical management.67,79,143 Acid secretory control in especially important for MEN-I–ZES patients, because a recent study has reported155 that these patients, compared with those with sporadic ZES, have a three-fold higher rate of esophageal strictures, five-fold higher rate of Barrett’s esophagus, and eight-fold higher incidence of dysplasia, which was thought in part caused by inadequate control of the acid hypersecretion early in these patients’ course. Occasional patients with ZES require parenteral control of the gastric acid hypersecretion caused by surgery, vomiting, chemotherapy, or gastric outlet obstruction. Parenteral histamine H2 receptor antagonists (cimetidine, ranitidine)9,46 and PPIs (omeprazole, pantoprazole)156-158 have been reported to be effective. Several IV PPI preparation are approved in the United States and are effective at controlling acid hypersecretion in patients with ZES—for example, the usual dose of pantoprazole IV is 80 mg given every 8 hours by 15-minute IV infusion. Parenteral histamine H2 receptor antagonists are effective; however, continuous infusions of relatively high doses are required to control hypersecretion (mean dose, 1 mg/kg/hr of ranitidine). Long-term oral antisecretory treatment with high doses of histamine H2 receptor antagonists or PPIs has remained effective for more than 10 years without the development of tachyphylaxis that limited effectiveness in most patients.28,46,65,110,140,149 Because antisecretory drug requirements can change with time, it is recommended that patients with ZES have their acid secretory control checked after six months initially, after starting histamine H2 receptor antagonists, or yearly after starting PPIs. A recent study has reported that this can be reliably performed during UGI endoscopic studies.148 High doses of cimetidine can cause antiandrogen side effects in males, and high doses of some histamine H2 receptor antagonists can interfere with cytochrome P450 metabolizing enzymes or result in central nervous system (CNS) symptoms in patients with hepatic or renal failure. Long-term treatment with PPIs rarely has dose-related side effects, although potential side effects have been raised because of their effects on nutrient absorption.159,160 In studies of long-term PPI treatment in ZES

patients, some patients developed decreased serum vitamin B12 levels, but not decreased body iron stores.161 This likely occurs because the absorption of protein-bound vitamin B12 in foods require the presence of acid and pepsin and therefore potent acid suppressants such as PPIs can interfere with vitamin B12 absorption. At present, it is unclear if long-term PPI treatment in patients with ZES will result in clinically significant vitamin B12 deficiency. In addition, long-term PPI treatment may interfere with calcium absorption. A recent epidemiologic study has reported that long-term high-dose PPI usage is associated with an increased incidence of hip fractures.159 There is no evidence that long-term treatment with PPIs increases the rate of development of gastric carcinoids in patients with or without ZES or other tumors.116,118,162 These observations are important because the risk of developing gastric carcinoids on PPIs in animal studies is related to the severity of the hypergastrinemia. Hence, the chronic use of PPIs, by causing physiologic hypergastrinemia, could contribute to the extent of hypergastrinemia in ZES patients.114-118,163 In a detailed study of 106 patients with sporadic ZES, many of whom had been treated long-term with PPIs, no gastric carcinoids were found, even though over 99% had ECL hyperplasia and 7% had dysplasia.114 In contrast, in a similar study in MEN-I–ZES patients, 53% had advanced ECL changes and 23% had gastric carcinoids, demonstrating that gastric carcinoids are more than 70 times more common in MEN-I– ZES.119 However, in this latter study, there was no evidence that the use of PPIs increased the rate of development of gastric carcinoids, but was instead best correlated with the extent of hypergastrinemia and presence of the MEN-I. Furthermore, animal and cell biology studies have provide evidenced for a possible role of gastrin-related peptides in growth and development of other tumors, including colorectal cancer.164 Although studies in ZES patients9 have shown increased proliferative rates of colonic mucosal cells, epidemiologic studies have not shown an increased association of colorectal cancer and hypergastrinemia.120,162 Therefore, in ZES patients there is no evidence that chronic hypergastrinemia, whether caused only by the gastrinoma or possibly contributed to by chronic use of PPIs, increases the development and/or growth of colorectal cancer.

Surgical Treatment of Gastric Hypersecretion

Total gastrectomy, the only effective means to treat the gastric acid hypersecretion seen in ZES for many years, is now rarely used.2,28,56,65,71,133,165 At present, total gastrectomy is relatively safe, with an overall mortality of 5.8% in 248 cases since 1980 and 2.4% for elective cases.47 However, the long-term morbidity of total gastrectomy in ZES remains unclear.110 Vagotomy decreases gastric acid hypersecretion in patients with ZES.9,28,110,166 Parietal cell vagotomy decreased BAO by 41% and decreased the dose of oral histamine H2 receptor antagonist by 40%. Long-term, 36% of patients were able to discontinue antisecretory drugs and 80% continued to have a decrease in BAO. It has been proposed that at the time of laparotomy for possible cure, a parietal cell vagotomy should be performed to decrease antisecretory drug requirements in the 70% of patients who are not cured long term.68 Parathyroidectomy in patients with ZES–MEN-I with hyperparathyroidism decreases fasting gastrin levels, decreases gastrin increases after secretin injection, decreases BAO, and increases sensitivity to antisecretory drugs.28,67,79,110,143 This latter observation is particularly important because patients with ZES–MEN-I with hyperparathyroidism can be relatively more resistant to PPIs, and

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract higher and more frequent doses are required in these patients.140,150 This may lead to a significantly increased expense for antisecretory drugs in these patients, which can affect compliance.

Surgical Treatment of Gastrinoma

Because gastrinomas are malignant in 60% to 90% of cases, detailed tumor imaging is essential to determine tumor extent and location of the primary tumor.26,28,68,69,110,122 Localization methods and results are discussed in a later section. Most but not all authorities recommend that if diffuse metastatic disease to the liver is not present and the patient does not have MEN-I, surgical exploration is indicated.9,28,56,68,71,110,122,167 In a large prospective study from the NIH68 involving 151 patients with ZES, patients with sporadic ZES (i.e., no MEN-I), gastrinomas were found in 92%; 51% were disease-free immediately after tumor resection and 34% were disease-free at 10 years. This cure rate is higher than the 0% to 20% rate reported in many earlier studies and is, in large part, the result of finding increased numbers of small duodenal gastrinomas.125,126 It is essential to perform a routine duodenotomy at surgery, which iden­ tifies 20% to 25% more duodenal tumors than other commonly used methods.110,122, A prospective NIH study123 has demonstrated that the routine use of duodenotomy increases the short- and long-term cure rates (Fig. 32-4). Two NIH studies have demonstrated that routine surgical resection, primarily in sporadic ZES patients, not only decreases the rate of development of liver metastases, but also increases survival.168,169 In patients with MEN-I and ZES, the role of surgery for cure of the gastrinoma is controversial.47,56,67,68,71,170 This controversy has occurred because the disease-free rate for ZES in patients with MEN-I is very low (0% to 5%) because most these patients have multiple duodenal gastrinomas and frequently have lymph node metastases at surgery.171 Studies have shown that gastrinomas and other pancreatic endocrine tumors can show aggressive growth in 15% of the patients, with hepatic metastases in 23%, and the rate of progression of these tumors is becoming an important determinant of the long-term survival of these patients.67,170,172

DuodX (n = 94 pts)

P = 0.0029

No DuodX (n = 68 pts) 0

A

5

10

Years post surgery

DEFINITION

Glucagonomas are PETs that secrete excessive amounts of glucagon and cause a distinct syndrome characterized by a specific dermatitis (migratory necrolytic erythema), weight loss, glucose intolerance, and anemia (Table 32-7). Although

15

Surgery (n = 160)

100 90 80 70 60 50 40 30 20 10 0

P <0.0001

No surgery (n = 35)

0

B

LIVER METASTASES FREE FROM FIRST NIH VISIT Percent free of liver metastases

100 90 80 70 60 50 40 30 20 10 0

GLUCAGONOMAS

DISEASE-RELATED SURVIVAL FROM ONSET

Percent surviving

Percent disease-free

DISEASE-FREE SURVIVAL PER OPERATION

The role of surgery is controversial in part because it is unclear whether early surgical resection will alter survival in MEN-I–ZES; without aggressive resections, it does not cure these patients. Some recommend routine exploration in all patients with MEN-I–ZES, others recommend no exploration and still others recommend exploration only if lesions larger than 2 to 3 cm are seen.28,173 Although Whipple resection can cure a high proportion of MEN-I–ZES patients, it is not generally recommended because of the excellent long-term prognosis of unoperated patients with small PETs (<2 cm) and because of the side effects of Whipple resections. Studies have shown that patients with PETs smaller than 2 cm with MEN-I–ZES had a 100% survival rate at 15 years,170 which is similar to results in MEN-I patients with nonfunctional PETs smaller than 2 cm who had no increase mortality compared with MEN-I patients without PETs.174 At present, we recommend surgical exploration only if lesions larger than 2 cm are imaged in MEN-I–ZES patients. At surgery, we do not recommend routine distal pancreatectomy; this will rarely result in cure but because larger lesions are associated with more frequent liver metastases, it may decrease the liver metastases rate.9,175 A small percentage of patients (10% to 15%) with ZES develop limited metastatic disease in the liver and a number of studies recommend surgical resection in these patients.28,122,128,176,177 Although this will result in cure in only a small percentage of patients, it has been proposed that it may extend survival, though there are no prospective studies that establish this. In patients with widely metastatic disease, various antitumor treatments are frequently used and are dealt with in a later section on treatment of advanced disease.

10

20

30

Years from onset

40

Surgery (n = 160)

100 90 80 70 60 50 40 30 20 10 0

P <0.0001

No surgery (n = 35)

0

C

5

10

15

20

25

30

Years from first NIH visit

Figure 32-4.  A, Effect of routine duodenotomy (DuodX) on disease-free survival of Zollinger-Ellison syndrome patients (pts). B, Effect of routine surgical exploration on disease-free survival. C, Effect of routine surgery on likelihood of being free of liver metastases. NIH, National Institutes of Health. (Adaped from Norton JA, Alexander HR, Fraker DL, et al. Does the use of routine duodenotomy (DUODX) affect rate of cure, development of liver metastases or survival in patients with Zollinger-Ellison syndrome (ZES)? Ann Surg 2004; 239:617-26; and Norton JA, Fraker DL, Alexander HR, et al. Surgery increases survival in patients with gastrinoma. Ann Surg 2006; 244:410-9.)

505

506

Section IV  Topics Involving Multiple Organs Table 32-7  Frequency of Clinical Features and Laboratory Abnormalities in Patients with Glucagonoma PARAMETER

FREQUENCY (%)

Clinical Features Dermatitis Diabetes mellitus, glucose intolerance Weight loss Glossitis, stomatitis, cheilitis Diarrhea Abdominal pain Thromboembolic disease   Venous thrombosis   Pulmonary emboli Psychiatric disturbance Laboratory Abnormalities Anemia Hypoaminoacidemia Hypocholesterolemia Renal glycosuria

54-90 22-90 56-96 29-40 14-15 12 12-35 4-24 11 Low 33-85 26-100 80 Unknown

From references 4, 9, 73-76, 181, 182, and 184.

Mallinson and colleagues4 specifically established the association of the rash with glucagon-producing tumors of the pancreas when they reported nine cases in 1974, the disease had been described earlier by others, although the association with glucagon release was not appreciated. In 1942, Becker and associates described the association of a PET with a skin rash.178 In 1966, McGarvan and coworkers reported a patient with an elevated fasting glucagon level, dermatitis, diabetes, and a PET.179 In 1973, Wilkinson and colleagues80 described the rash as necrolytic migratory erythema.

PATHOPHYSIOLOGY AND PATHOLOGY

In contrast to insulinomas, most glucagonomas are large at the time of diagnosis, with the average size between 5 and 10 cm (range, 0.4 to 35 cm).73,74,76,181 Similar to other PETs except insulinoma, 50% to 80%9,182,183 of glucagonomas had evidence of metastatic spread or invasion. The most common site of metastatic spread was to the liver (43% to 82%), with lymph nodes (38%), bone, and mesentery less commonly.9 Most glucagonomas occur in the pancreas (>97%); however, a glucagonoma associated with the typical clinical syndrome was found in the proximal duodenum. Glucagonomas usually occur as a single tumor, although 10% to 12% of patients in one series had multiple tumors or diffuse involvement by a single mass.181 The pathophysiology of the glucagonoma syndrome is related to the known actions of glucagon. Glucagon stimulates glycogenolysis, gluconeogenesis, ketogenesis, lipolysis, and insulin secretion, as well as having inhibitory effects on pancreatic and gastric secretion and on gut motility.184 Hyperglycemia results from the increased hepatic glycogenolysis and gluconeogenesis. Because glucagon also increases secretion of insulin, which prevents lipolysis and maintains normal free fatty acid concentrations, ketonemia usually does not develop.9 The weight loss has been attributed to the known catabolic effects of glucagon.181 However, severe anorexia and adipsia are seen in rats with transplanted glucagonomas and studies have suggested that the tumor is producing a novel anorectic substance. It is not clearly established that the skin rash is caused by the hyperglucagonemia per se because numerous patients have been given large doses of glucagon over extended periods and the skin rash did not develop. However, in two studies, pro-

longed hyperglucagonemia caused by glucagon administration did cause a typical skin rash.185,186 It is possible that hypoaminoacidemia, which develops in 80% to 90% of glucagonoma patients,182 or essential fatty acid deficiency may be involved in the genesis of the rash because, if these metabolic derangements are corrected, the dermatitis may improve without changing plasma glucagon levels. The similarity of the skin lesions to those seen in patients with zinc deficiencies has resulted in trials of zinc, with some responses. However, in some patients, the rash has resolved with rehydration and glucose solution; therefore, there may be differing contributing factors in different patients. The severe hypoaminoacidemia is believed secondary to the hyperglucagonemia, and glucagon infusions can alter amino acid metabolism. This conclusion is further supported by findings that total pancreatectomy or somatostatin administration, which decreases plasma glucagon levels, increases plasma amino acid levels, whereas glucagon administration decreases plasma amino acid levels. The role of glucagon per se in causing the thromboembolic phenomenon is not clear. Glucagon is known to affect coagulation parameters; however, the relationship of this to the thromboembolic events is not known. The anemia may be caused by the glucagon excess because prolonged treatment with a longacting glucagon preparation decreases erythropoiesis in animals. Immunocytochemical and histologic studies of glucagonomas show results typical of PETs. Glucagon is one of the most commonly seen peptides in immunocytochemical studies of PETs, but in many cases it is not associated with any syndrome. In one series of 1366 autopsy cases, a frequency of 0.8% adenomas was reported and all contained glucagon-producing cells.9 In glucagonomas, as with other PETs, multiple GI hormones are frequently seen on immunocytochemical studies.73 The morphology of most glucagon-producing tumors demonstrates no histologic features that distinguish them from other PETs. Even though the tumors are usually malignant, mitotic figures and nuclear atypia are uncommon. With electron microscopy, readily identifiable A granules, typical of those seen in normal alpha cells, are recognizable in PETs; they stain positive for glucagon and yet do not cause the glucagonoma syndrome, whereas in patients with a glucagonoma syndrome, atypical granules are usually seen.181

CLINICAL FEATURES

Glucagonomas usually occur in middle-aged or older adults.9,73,76,181,182 No cases have been reported in individuals younger than 19 years; only 16% of all cases occur in those younger than 40 years, with most cases occurring in individuals 50 to 70 years old. Glucagonomas occur slightly more commonly in women. Cutaneous lesions are one of the most common manifestations of the disease, occurring in 53% to 90% (Fig. 32-5 and see Table 32-7).74-76 Cutaneous lesions often precede the diagnosis of the syndrome for long periods, with a mean of six to eight years in one study and a maximum of 18 years.187 In 70% of patients, the skin lesion is the presenting sign of the disease and has been reported up to three years before the PET is found.188 Skin lesions may wax and wane and may be misdiagnosed as pemphigus foliaceous, pemphigoid, vasculitis, acrodermatitis enteropathica, psoriasis, herpes, seborrheic or contact dermatitis, eczema, pellagra, or even a chemical burn.189 Excellent descriptions of the typical rash, necrolytic migratory erythema associated with glucagonoma, have been published.180,181 Characteristically, the skin lesion starts as an erythematous area typically at periorofacial or intertriginous areas, such as the groin, buttocks, thighs or perineum,

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract

Figure 32-5.  Necrolytic migratory erythema in a patient with glucagonoma characterized by rapidly eroding superficial blisters. The lesions are usually localized to the buttocks, groin, perineum, elbows, hands, feet, and perioral area. (Courtesy of Dr. Carl Grunfeld, San Francisco, Calif.)

and then spreads laterally. The lesions subsequently become raised, with superficial central blistering. The top of the bullae frequently detach or rupture, leaving eroded areas that crust. The lesions tend to heal in the center while the edges continue to spread, with a crusting well-defined edge. Healing is associated with the development of hyperpigmentation. This entire sequence characteristically takes one to two weeks and while some new lesions are developing, others are healing. Therefore, a mixed pattern of erythema, bullous formation with epidermal separation, crusting, and hyperpigmentation, together with normal skin, can occur. The histopathology can be as varied as the clinical presen­ tation.9 In its classic form, early lesions demonstrate a superficial spongiosis and necrosis, with subcorneal and midepidermal bullae. Fusiform keratinacytes with pyknotic nuclei are often seen, as are mononuclear inflammatory infiltrates.4,180 This characteristic histologic pattern is best seen in an early lesion. Glossitis or angular stomatitis is reported to occur in 34% to 68% of patients.73,182 In addition, some patients develop a nail dystrophy, with brittleness and crumbling of the nails. In a review187 of 13 cases of necrolytic migratory erythema, it was reported that patients frequently present with noncharacteristic clinical and histologic features and the diagnosis of glucagonoma was delayed. Glucose intolerance with or without frank diabetes mellitus occurs in 22% to 90% of cases (see Table 32-7).9,73-76,181,182 In one series, 42% of patients required oral hypoglycemic agents and 24% required insulin. The onset of the diabetes mellitus preceded the diagnosis of the glucagonoma by up to 10 years in one study, with an average time of five years. A number of patients with hyperglucagonemia do not have diabetes mellitus. Furthermore, although correlations between plasma glucagon levels and changes in plasma glucose have been described in a few cases, this correlation was not easily demonstrated when a number of proven cases were analyzed. Tumor resection and normalization of blood glucagon may not result in normalization of the glucose

intolerance.82 In some patients, removal of the glucagonoma improved glucose tolerance.4 In various patients with glucagonomas, plasma insulin concentrations are normal or elevated, but circulating plasma insulin and glucagon levels are not correlated. The relationship of the diabetes mellitus to the hyperglucagonemia remains unclear. Hypoaminoacidemia occurs in 26% to 100% of patients with a clinical glucagonoma4,9,73 and essential fatty acid deficiencies are also reported. Plasma concentrations of amino acids are frequently less than 25% of normal, with glycogenic amino acids most affected, whereas branched chain amino acids are reported to be less affected.181 The degree of hypoaminoacidemia may vary with the intensity of the disease. The hypoaminoacidemia is reported to be associated with the pathogenesis of the skin rash (see earlier). Weight loss is a prominent feature of the glucagonoma syndrome, occurring in 56% to 96% of patients (see Table 32-7).9,73-76,181 A number of observations suggest that the weight loss is a unique aspect of the syndrome. Weight loss is seen even in patients with small tumors without metastatic spread.4,9,182 The weight loss is often associated with anorexia and may be profound, with a mean weight loss of 20 kg in 44 cases, with a maximal loss of 30 kg. Venous thromboembolism is common in patients with glucagonoma, occurring in 4% to 35% of patients in various series.9,74-76,182 Venous thrombosis occurred in 24% of patients and pulmonary emboli in 12%. This complication is thought to be related to the glucagonoma syndrome because it is not seen as frequently in other PETs. Anemia occurs in 33% to 85% of cases. The anemia is usually normocytic and normochromic (in 73% of cases in one study) and may be severe, but usually is not. When serum iron, folate, and vitamin B12 levels have been measured, they are usually normal. The anemia is reported to respond to successful tumor therapy. Psychiatric symptoms, although mentioned in several reports,4 appear to be relatively infrequent. Depression is the most commonly reported abnormality. Whether psychiatric disturbances occur with increased frequency compared with other comparable debilitating illnesses is unclear. Abdominal pain without specific identifying characteristics is reported in approximately 12% of patients. Diarrhea, weight loss, and hepatomegaly are more frequent in patients with metastatic disease.73 Diarrhea is reported in 14% to 15% of patients181 with severe steatorrhea. Diarrhea is reported far more commonly than constipation (15% vs. 4% in one study)9 and the cause of the diarrhea remains unclear. Jejunal biopsies have been reported to be normal or to show hypertrophic folds. It remains possible that other hormones may also be secreted by the tumor that could contribute to the diarrhea, but this possibility has not been extensively studied. Other laboratory findings reported to be abnormal in patients with glucagonoma include the presence of renal glycosuria and hypocholesterolemia. Renal glycosuria may occur early, and may represent a direct renal effect of glucagon. Hypocholesteremia was reported in 80% of cases.

DIAGNOSIS AND DIFFERENTIAL DIAGNOSIS

Glucagonomas are usually suspected because of the skin rash, although occasionally the diagnosis is suspected in a patient with a pancreatic mass with weight loss or diabetes.73,74,76,181,187 The skin lesion is most frequently confused with pemphigus foliaceous, although a number of dermatologic lesions have also been misdiagnosed. A number of cases of patients with a typical necrolytic migratory erythema have been described who do not have glucagonoma.9,182 Necrolytic migratory erythema is therefore not

507

508

Section IV  Topics Involving Multiple Organs pathognomonic for glucagonoma. The rash can be seen in myelodysplastic disorders,190 short bowel syndrome,191 hepatitis B infections, malnutrition,192 cirrhosis, celiac disease, other malignancies, malabsorptive syndromes, nutritional deficiencies, and inflammatory bowel diseases.193 In some studies, up to 20% of patients with glucagonomas have Zollinger-Ellison syndrome; 13% to 17% have MEN-I whereas 0% to 3% of MEN-I patients develop glucagonomas.47,194 Once the diagnosis is suspected, it can be confirmed by demonstrating an increase in plasma glucagon concentration. In most laboratories, the upper limit of normal for fasting glucagon concentration using antibody 30 K of Unger is 150 to 200 pg/mL. In one large review, only two patients had a plasma glucagon level of 200 to 500 pg/mL, four cases had 500 to 1000 pg/mL, and 52 cases more than 1000 pg/mL.181 These results are in close agreement with another study9 in which the mean plasma glucagon concentration in 73 cases of glucagon was 2110 pg/mL, with a range of 550 to 6600 pg/mL, with 30% between 500 and 1000 pg/mL and the remaining 70% more than 1000 pg/mL. However, in some studies,74-76,195 patients with the glucagonoma syndrome with only mildly elevated plasma glucagon levels have been described; in one recent study,76 25% of patients had plasma glucagon levels lower than 500 pg/ mL. Hyperglucagonemia is reported to occur in cirrhosis, chronic renal insufficiency, diabetic ketoacidosis, prolonged starvation, acute pancreatitis, acromegaly, hypercorticism, septicemia, severe burns, severe stress (trauma, exercise), celiac disease, and familial hyperglucagonemia and with danazol therapy and hepatic disease.9,179,180,182 It has been reported that the plasma glucagon level in these conditions does not exceed 500 pg/mL. It therefore has been recommended that a plasma glucagon concentration more than 1000 pg/mL is diagnostic of glucagonoma. The one reported exception to this is patients with cirrhosis, in whom the plasma level of glucagon can be more than 1000 pg/mL. Because a necrolytic migratory erythematous-like rash has been reported in patients with hepatic disease, diagnostic confusion could occasionally result. However, in some studies, no overlap in plasma glucagon levels was found between patients with cirrhosis (or any other of the conditions listed) and the values seen in patients with glucagonomas. Various provocative tests have been described, such as the use of secretin to cause a paradoxical increase in glucagon release or a mixed or carbohydrate-rich meal.9 However, at present, none of these tests are sufficiently reliable to differentiate glucagonoma with normal plasma glucagon elevations from these other conditions. Patients with familial hyperglucagonemia have been described. These patients are asymptomatic and can be distinguished from patients who have the glucagonoma syndrome. In addition, fractionation of the plasma glucagon immunoreactivity gives a different pattern than that seen in normals or patients with glucagonoma; specifically, an increased percentage of the high molecular weight big plasma glucagon peak is described in familial cases.181

TREATMENT

Because glucagonomas are generally malignant and it is not possible to predict in a given patient when metastases may develop, surgical resection should be considered in all patients, if feasible. Similar to the considerations with other malignant functional PETs, initial treatment can be directed at controlling the symptoms, restoring nutritional status, and controlling the hyperglycemia while tumor localization studies are being performed and possible surgical curative

resection considered. Tumor localization studies are dealt with in a separate section.

Medical Treatment

Preoperative medical control of symptoms is important in these patients, because they are generally poor operative risks. The catabolic effects of glucagon, combined with glucose intolerance and diabetes mellitus (see Table 32-7), can markedly affect the nutritional status of these patients. They have an increased incidence of venous thromboembolism, increasing the postoperative risk. To improve the metabolic status of these patients prior to surgery, blood transfusions in those with severe anemia and a period of extended hyperalimentation are recommended. Parenteral nutrition with restoration of plasma amino acid levels and/or essential fatty acid levels to normal is also reported to have an excellent effect in healing the dermatitis.9 The long-acting somatostatin analog, octreotide, has been useful in controlling symptoms in patients with glucagonoma.74-76,100-102,181 The rash improved with octreotide treatment in 54% to 90% of patients, with complete disappearance in up to 30%. Octreotide generally improved the symptoms of weight loss, abdominal pain, and diarrhea as well. Diarrhea sometimes even resolves. Diabetes mellitus was not improved with octreotide treatment. Plasma glucagon levels decreased in 80% to 90% of patients but only decreased into the normal range in 10% to 20% of patients with octreotide treatment. In most studies, 100 to 400 µg/ day of octreotide was used.9,183 However, now that longacting formulations of octreotide (octreotide LAR) and lanreotide (Autogel or lanreotide SR) are available, it is likely these will be predominantly used in the future. It has been reported that octreotide LAR and lanreotide SR are effective at reducing symptoms and serum glucagon levels in a few patients with glucagonoma.196 In some patients, with continued treatment, the dosage had to be increased to continue to control symptoms.

Surgical Treatment

Approximately 50% to 90% of patients with glucagonomas have metastases at the time of diagnosis.9,73,74,74-76,197 Surgical resection has been successful in a number of cases.180-182 The exact percentage of cases that can be cured is unknown; however, with the syndrome currently recognized relatively late in most patients, it is likely that fewer than 20% of patients can be cured. A number of operated patients developed recurrence with elevated plasma glucagon elevations after what was thought to be complete tumor removal.82 However, even if a patient eventually develops a postoperative recurrence or if surgical debulking is performed, an extended disease-free interval may be attained. A number of studies have reported a benefit to patients even if surgical debulking only can be done. In patients with widely metastatic disease in whom surgical debulking is not possible, various chemotherapeutic agents are frequently used and will be dealt with in a latter section on treatment of advanced disease.

VIPOMAS DEFINITION

The VIPoma syndrome is caused by a neuroendocrine tumor, usually pancreatic in location (in adults) that secretes excessive amounts of VIP, which causes a syndrome characterized by extreme secretory diarrhea, hypochlorhydria, and hypokalemia (Table 32-8). The syndrome was described

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract Table 32-8  Frequency of Symptoms, Signs, and Laboratory Findings in Patients with VIPoma Syndrome FEATURE

FREQUENCY (%)

Symptom or Sign Secretory diarrhea Volume depletion Weight loss Abdominal cramps, colic Flushing Laboratory Findings Hypokalemia Hypochlorhydria Hypercalcemia Hyperglycemia

89-100 44-100 36-100 10-63 14-33 67-100 34-72 41-50 18-100

VIPoma, vasoactive intestinal peptide secreting pancreatic endocrine tumor. From references 77, 199, and 202-205.

in 1957 by Priest and Alexander198 and by Verner and Morrison in 19583 and is commonly also called the VernerMorrison syndrome (see Table 32-1). The absence of gastric acid secretion in these patients was later noted.9 Because of the resemblance of the diarrhea fluid to that seen in cholera, the term pancreatic cholera was proposed and the acronym WDHA (watery diarrhea, hypokalemia and achlorhydria) was created.199 VIP was long suspected as the mediator of this syndrome and the ability of VIP to produce secretory diarrhea in humans at blood levels seen in VIPomas was confirmed in 1983.200

PATHOPHYSIOLOGY AND PATHOLOGY

In adults, 80% to 90% of VIPomas are pancreatic in location, with rare cases caused by VIP-producing inte­ stinal carcinoids, bronchial carcinomas, or pheochromocy­ tomas.199,201-204 Extrapancreatic VIPomas are reported in the retroperitoneum, liver, esophagus, and small intestine. VIPomas are usually large solitary tumors. In two series,77,202 only 2% of tumors were multiple. In the pancreas, 42% to 75% occur in the pancreatic tail.201,203-205 In various series, 29% to 78% of the VIPomas had metastases at the time of diagnosis or surgery, which is comparable to the 63% to 90% malignancy rate reported with gastrinomas, glucagonomas, and somatostatinomas. In children younger than 10 years, and rarely in adults (5% of cases), the VIPoma syndrome is caused by a ganglioneuroma or ganglioneuroblastoma. These tumors are extrapancreatic and are less often malignant (10% of cases) than pancreatic VIPomas. By immunocytochemistry, VIP was detected in 57% to 100% of VIPomas.77,201 PP is found in 34% to 53% of VIPomas, glucagon in 19% to 27%, somatostatin in 10% to 46%, insulin in 5%, and gastrin in 0% to 23%; 45% secrete multiple hormones. VIP-producing tumors also elaborate peptide histidine methionine (PHM-27), a 27–amino acid peptide that shares with VIP a common precursor peptide (prepro-VIP–PHM-27). PHM-27–like immunoreactivity has also been found in the plasma and tumor of patients with VIPomas.9 On conventional microscopy, VIPomas show the typical microscopic features of PETs. Mitoses are uncommon, seen in only 12%. On electron microscopy, a mixture of cells are seen usually in the same tumor, with 90% having cells with a few scattered, inconspicuous secretory granules, 89% having a few small agranular cells, and 52% have some well-differentiated endocrine cells, with welldeveloped granules. The secretory granules are small (120 to 180 nm) and resemble those of the D1 or P cells

of the normal gut. Histologic and electron microscopic studies do not allow VIPomas to be clearly differentiated from some other PETs; however, the presence of immunoreactive VIP is strongly suggestive of VIPoma, because this is uncommonly found in other PETs (10/104 in one study). It is now clear that VIP is the major mediator of the VIPoma syndrome.9,199 In early studies, plasma VIP infusions in humans did not produce the syndrome. However, in subsequent studies, plasma VIP is usually elevated in the VIPoma syndrome.202 Also, a continuous infusion of VIP for 10 hours in normal human subjects to achieve plasma levels similar to those seen in patients with the VIPoma syndrome produced watery diarrhea in 6 to 7 hours.200 The ability of VIP to produce diarrhea is consistent with its known actions in the intestine. Receptors for VIP have been identified on intestinal epithelial cells. VIP stimulates rat intestinal electrolyte and fluid secretion in animals, stimulates chloride secretion with increased short circuit current, and activates adenylate cyclase and cyclic AMP in intestinal cells, which leads to intestinal secretion. PHM-27–like immunoreactivity, found in 92% of VIPomas, can induce intestinal chloride secretion and thus could contribute to the pathogenesis of the diarrhea. However, PHI (the porcine equivalent of PHM) is 32-fold less potent than VIP and, because VIP is always, present it is likely the important peptide in most cases. The pathogenesis of the severe hypokalemia is likely primarily caused by fecal loss but also may be contributed to by the secondary hyperaldosteronism that results from VIP stimulation of renin release.9 The mechanism of the hypercalcemia is unclear, and may be partially the result of the ability of VIP to stimulate bone osteolytic activity. The hyperglycemia has been attributed to the glycogenolytic effect of VIP on the liver. The flushing, which is seen in 14% to 28% of patients with VIPoma, has been attributed to the known potent vasodilatory effects of VIP. The fact that only a minority of patients with the VIPoma syndrome develops flushing despite high plasma VIP levels has been attributed to the fact that prolonged VIP infusions result in a gradual loss of flushing, suggesting tachyphylaxis. The pathogenesis of the hypo- or achlorhydria, which frequently occurs in patients with the VIPoma syndrome, is not entirely clear, but has been attributed to the known inhibitory effect of VIP on gastric acid secretion.

CLINICAL FEATURES

The mean age for adults at the time of diagnosis is 42 to 51 years, with a range of 32 to 81 years.77,199,202-205 There is a female predominance in some studies, but not others. In children, the mean age is 2 to 4 years old, with a range from 10 months to 9 years. The cardinal features of the VIPoma syndrome are the presence of severe secretory diarrhea (89% to 100%) associated with hypokalemia (67% to 100%) and volume depletion (44% to 100%; see Table 32-8). The diarrhea may be episodic and, in two studies,199,203 was intermittent in 53% to 54% of patients initially. The diarrhea is typically large in volume, with all patients having more than 1 L and most more than 3 L/day.9,18 A stool volume of less than 700 g/day has been proposed to rule out the diagnosis of VIPoma. The diarrheal fluid is described as having the appearance of weak tea198 and to persist during fasting.9 Only 10% of patients have less than five bowel movements/day. Gross steatorrhea is usually not present and, in one study, none of 52 patients with VIPomas had 24-hour fecal fat excretion exceeding 15 g/day.202 Weight loss is usually present, with a range from 7 to 27 kg in one study. Flushing occurs in

509

510

Section IV  Topics Involving Multiple Organs 14% to 33% of patients, is usually present in the head or trunk area, and is characteristically erythematous. Laboratory studies typically demonstrate hypokalemia (67% to 100%) and, to a lesser degree, hypercalcemia (41% to 50%) and hyperglycemia (18% to 50%; see Table 32-8). The hyperkalemia is often severe, below 2.5 mmol/L at some time in 93% of patients.202 The hypercalcemia is usually not extreme. The hyperglycemia is usually mild.199 Tetany has been occasionally reported.9 Tetany may occur and has been attributed to hypomagnesemia resulting from the diarrhea. If measured, hypochlorhydria occurs in 34% to 72% of cases.

making a diagnosis of VIPoma in a patient with diarrhea because other conditions, such as prolonged fasting, inflammatory bowel disease, small bowel resection, radiation enteritis, or chronic kidney disease can occasionally elevate VIP levels.212 In a rare patient, intestinal perfusion studies may be helpful in the differential diagnosis.199 Net secretion of electrolytes and water occurs in VIPomas instead of a net absorption in perfused small intestinal segments. This method is reported to be particularly helpful in differentiating VIPomas from surreptitious laxative ingestion because the latter group of patients shows normal perfusion results.

DIAGNOSIS AND DIFFERENTIAL DIAGNOSIS

TREATMENT

The diagnosis of the VIPoma syndrome requires the demonstration of an elevated plasma concentration of VIP and establishment of the presence of a large-volume secretory diarrhea. The volume of the diarrhea9,18 should suggest the diagnosis because in 70% to 85% of patients, the diarrhea is more than 3 L/day and is never below 700 mL/day.199,202 Despite the severity of the diarrhea, it may be present for long periods prior to the correct diagnosis, with a delay in diagnosis after onset of 32 months in one study and varying from two months to 14 years in various studies. A large number of possible causes for the diarrhea can be excluded by fasting the patient because in patients with VIPomas, the diarrhea persists during fasting. The diarrhea fluid should be characteristic of a secretory diarrhea, wherein the stool electrolytes can account for all the stool water osmolality:

The first objective is the replenishment of fluid and electrolyte losses to correct the profound hypokalemia, volume depletion, and acidosis that are usually present (see Table 32-8). The patients may require 5 L/day of fluid or more9 and over 350 mEq/day of potassium.9,213 Renal failure associated with the potassium deficiency may occur in these patients and can be a cause of death.3 Furthermore, heart failure, perhaps related to chronic hypokalemia, has been reported. Therefore, during volume replacement, fluid and electrolyte requirements should be carefully monitored. The diarrheal output should be controlled by medical therapy (see next section). Once the fluid and electrolyte abnormalities are corrected, patients should undergo imaging studies and other appropriate evaluation to establish the diagnosis (see earlier).

(Sodium + potassium ) × 2 ≅ measured osmolality

In the past, numerous drugs were reported to control, to varying degrees, the diarrheal output in small numbers of VIPoma patients, including prednisone (60 to 100 mg/day), clonidine, indomethacin, phenothiazines, lithium, propanolol, metoclopramide, loperamide, lidamidine, angiotensin II, and norepinephrine.9,199,202 It has been proposed that these agents primarily enhance sodium absorption in the proximal small intestine or inhibit secretion.9 Currently, long-acting somatostatin analogs such as octreotide or lanreotide are the agents of choice.18,100-102,203 Octreotide will control short- and long-term diarrhea in 78% to 100% of patients with VIPoma.77,202 In two reviews, octreotide completely abolished diarrhea in 10% of patients in one study and in 65% in the other,9 and relieved the diarrhea in 90% to 95% of patients. In one study, octreotide continued to be effective at six months in all patients, whereas in another study octreotide was effective long term in 56% and 22% required an increase in dosage. In one study,199 responses were short-lived in 17% of patients. In nonresponsive patients or in patients whose symptoms recur, the administration of glucocorticoids along with octreotide has proven effective in a small number of cases. Plasma VIP concentrations decreased in 80% to 89% of patients taking octreotide. The changes in plasma VIP concentration with octreotide treatment did not always mirror the clinical responses. In one review,213a of the 13 patients whose diarrhea was abolished by octreotide, plasma levels returned to normal in only 15%, decreased but not to normal ranges in 46%, decreased and then rose to pretreatment levels in 31%, and did not change in 8%. In 25% of patients who did not respond to octreotide, plasma VIP levels did not change. This discrepancy in extent of clinical response and degree of change in plasma VIP levels may be partially explained by the observation that multiple forms of VIP may be elevated in the plasma, and only the form coinciding with the native peptide may disappear after treatment with

Other diseases can cause a chronic secretory diarrhea with large volumes and give rise to a syndrome called the pseudo-VIPoma syndrome, with most of the clinical features of the VIPoma syndrome. Occasionally, these patients have gastrinoma,28,72 chronic laxative abuse,206 sprue,207 AIDS,208,209 and, in some cases, secretory diarrhea of unknown origin (see Chapter 15).210 The diagnosis of gastrinoma can be excluded by measuring fasting serum gastrin and, if needed, gastric acid secretory rate.28 Patients with surreptitious use of diuretics or laxatives can be very difficult to detect; therefore, in a patient with unexplained chronic diarrhea, screens for laxatives should be performed. However, with the standard methods used to detect laxatives, many abusers may be missed and can only be detected using advanced methods, such as gas liquid chromatography analysis combined with mass spectroscopy determination.211 To differentiate these other conditions from VIPomas, a reliable measurement of plasma VIP concentrations is required. The fasting plasma VIP level variations in most laboratories is 0 to 190 pg/mL. In one study, the mean value for 29 patients with VIPomas was 956 pg/mL, with the lowest value of 225 pg/mL; in a second large study, the mean value was 675 pg/mL, with the highest value of 53 pg/mL seen in normals and the lowest value of 160 pg/mL seen in a VIPoma patient. In this study,202 the mean values for patients with pancreatic VIPomas or ganglioneuromas were 702 and 539 pg/mL, respectively. In a literature review of 29 cases,204 the VIP values ranged from 100 to 7200 pg/mL, with a mean value of 632 pg/mL. With current VIP radioimmunoassays, the sensitivity is reported to be 88% and specificity 100%.199 VIP levels are reported to fluctuate in some patients and thus it is important to perform the study while the patient is having diarrhea. Elevated VIP levels alone should not be the sole basis for

Medical Treatment

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract octreotide, with the remaining VIP immunoreactivity representing nonbiological fragments. In patients with unresectable or metastatic VIPomas, long-term treatment with octreotide and/or chemotherapy will need to be considered. Continuous use of octreotide decreases annual medical costs for a VIPoma patient by 50%.9 There are only anecdotal studies on long-term treatment of patients with VIPomas with the new, longacting depot forms of somatostatin (octreotide LAR and lanreotide SR).101,102,104,203

somatostatin-containing tumor (see Chapter 31) and the term somatostatinoma syndrome will refer to the presence of a somatostatinoma with the accompanying clinical syndrome caused by ectopically released somatostatin.

PATHOPHYSIOLOGY AND PATHOLOGY

In four reviews, 46% to 75% of the somatostatinomas were located in the pancreas.9,32,78,216 The distribution ratios of the tumors in the pancreatic head-to-body-to-tail were 11 : 0 : 3, 14 : 2 : 5, 5 : 1 : 3, and 2.3 : 1 : 2. Tumors not in the pancreas arose from the duodenum (in 90% of cases in one study) or from the duodenum-ampulla (91%), jejunum (5%), or cystic duct (5%) in a second study. This distribution in the upper gastrointestinal tract may be a consequence of the large number of somatostatin-producing D cells in this region. Tumors were usually solitary (90% to 96%) and in two studies varied from 1.5 to 10 cm in diameter (mean, 4.9 and 3.6 cm). The average size of pancreatic tumors was more than duodenal somatostatinomas (5.1 vs. 2.4 cm). In various series,217 43% to 90% of all tumors had evidence of metastatic spread, primarily to proximal lymph nodes. In two of these studies,32,78 the malignancy rate of 50% was the same for pancreatic and duodenal somatostatinomas; however, in another study, 92% of pancreatic somatostatinomas had metastases whereas 69% of intestinal somatostatinomas had metastasized. Metastases occur to the liver and lymph nodes (31%) and to bone less frequently (4%). In one review, liver metastases (40% vs. 11%) and bone metastases (6% vs. 0%) occurred more frequently in pancreatic somatostatinomas, whereas the rate of metastases to lymph nodes was similar (25% vs. 35%). With duodenal somatostatinomas, the occurrence of lymph node metastases correlated with the primary tumor size. Duodenal somatostatinomas with lymph node metastases were significantly larger than those without metastases (2.9 vs. 1.4 cm; P < 0.05). Using a cutoff of 2 cm in diameter, diagnostic accuracy for malignancy was 78%, with a specificity of 87% and sensitivity of 63%. With light microscopic studies, most tumors appeared as well-differentiated tumors with varying degrees of fibrous septa.32 Histologically, a specific feature of duodenal somatostatinoma is the presence of psammoma bodies, which are rarely found in pancreatic somatostatinomas or other types of duodenal carcinoid tumors.9,31,78,82 Electron microscopic studies have reported that the secretory granules are typical of those in D cells in 52% to 89% of tumors. Immunocytochemical analysis has demonstrated somatostatin-like immunoreactive (SLI) material in all tumors (by definition) and 10% to 33% contained insulin, 22% to 27% calcitonin, 8% to 13% gastrin, and 9% gluca-

Surgical Treatment

After performing imaging studies to localize the primary VIPoma and determine its extent, surgical cure should be considered for all patients without metastatic disease.9,203-205 In one series, surgical resection of a pancreatic VIPoma relieved all symptoms in 33% of patients202 and in other series 30% were cured. Attempted curative surgical resection was only possible in about one third of patients. Surgical resection with complete control of all symptoms was possible in 78% of all patients with VIP-producing ganglioneuroblastomas.202 For patients with limited metastatic disease to the liver, radiofrequency ablation may be helpful,214 whereas in patients with advanced metastatic disease, debulking surgery has been recommended (see later).

SOMATOSTATINOMAS DEFINITION

Somatostatinomas are neuroendocrine tumors that usually originate in the pancreas or intestine, which may release large amounts of somatostatin and cause a distinct clinical syndrome (the somatostatinoma syndrome) characterized by diabetes mellitus, gallbladder disease, diarrhea, and weight loss (Table 32-9). In 1977, the first two cases of somatostatinoma were described by Ganda and colleagues5 and Larsson and associates.6 Steatorrhea and hypochlorhydria were added as additional features.9,215 Somatostatinomas are one of the least common PETs (see Table 32-1). This definition of the somatostatinoma syndrome is not uniformly used in the literature. In fact, in many reports, the term somatostatinoma is used to mean an endocrine tumor possessing somatostatin immunoreactivity, with no requirement for an accompanying functional syndrome. For example, in a large literature,78 173 cases were found, of which only 17 had the clinical somatostatinoma syndrome. Therefore, somatostatinoma will be used here to refer to a

Table 32-9  Frequency of Clinical and Laboratory Findings in Patients with Somatostatinoma or Somatostatinoma Syndrome (%)* Somatostatinoma FEATURE Clinical Finding Diabetes mellitus Gallbladder disease Diarrhea Weight loss Laboratory Finding Steatorrhea Hypochlorhydria

PANCREATIC

INTESTINAL

SOMATOSTATINOMA SYNDROME

95 94 66-97 32-90

21 43 11-36 20-44

95 68 37 68

83 86

12 17

47 26

*Somatostatinoma is a pancreatic endocrine tumor containing somatostatin by immunocytochemistry that can occur with (in 11%) or without (in 89%) the somatostatinoma syndrome, which is caused by ectopically released somatostatin. From references 9 and 78.

511

512

Section IV  Topics Involving Multiple Organs gon. Twenty-six percent of somatostatinomas were producing multiple GI hormones. In one study,217 all the duodenal somatostatinomas were pure somatostatinomas, whereas the pancreatic tumors were mixed and also stained positive for a number of other hormones (gastrin, insulin, glucagon, ACTH). Extraction of tumors and analyses for somatostatinlike immunoreactive material have demonstrated that somatostatin-14 is usually not the predominant form. Somatostatin-28 and larger forms predominate, and this heterogeneity is thought to reflect incomplete processing of precursors. The pathophysiology of the somatostatinoma syndrome can be explained by the known actions of somatostatin. Somatostatin exists as a tetradecapeptide (SS-14) as well as a larger form, SS-28.9 Somatostatin is found throughout the gastrointestinal tract, especially in gastric and duodenal D cells. It has a largely inhibitory action,100-102 inhibiting the release of numerous gastrointestinal hormones and inhibiting basal and stimulated acid secretion, but stimulating pancreatic secretion and intestinal absorption of amino acids, sugars, and calcium. Somatostatin also has stimulatory and inhibitory effects on intestinal motility and inhibitory effects on gallbladder contraction. The development of diabetes mellitus is likely secondary to the inhibitory action of somatostatin on insulin release, possibly by replacement of functional pancreatic tissue by the tumor.31,32 Gallbladder disease may be a result of somatostatin inhibition of gallbladder emptying, as demonstrated by the occurrence of cholelithiases or biliary sludge in patients taking octreotide. Diarrhea and steatorrhea, likely caused by the ability of somatostatin to inhibit pancreatic secretion of enzymes and bicarbonate, gallbladder motility, and absorption of lipids,9,100-102 occur also in patients treated with high doses of octreotide. The hypochlorhydria is likely secondary to the known ability of somatostatin to inhibit gastric acid secretion. The weight loss may be secondary to the malabsorption but there may be other causative factors not identified.

CLINICAL FEATURES

The mean age of patients with somatostatinoma is 45 to 54 years,32,216,217 with most patients 40 to 60 years old (see Table 32-9). In one series,78 women with pancreatic tumors were older than those with duodenal tumors (55 vs. 49 years). For patients with pancreatic tumors, 66% were females, whereas for those with intestinal somatostatinomas, 43% were females in one series9 but, in another series,78 somatostatinomas occurred equally in both genders. The youngest and oldest patients were 21 and 91 years. In regard to symptoms with somatostatinomas, it is important to distinguish between symptoms likely caused by the PET itself and those caused by ectopic somatostatin release (somatostatinoma syndrome). In one large literature review,78 only 11% of all 173 cases of somatostatinomas were associated with the specific symptoms associated with the somatostatinoma syndrome; in a second histologic study, it was 0% (0 of 42; see Table 32-9).32 Overall, 93% of patients with somatostatinomas had symptoms, with abdominal pain (40%), weight loss (26%), jaundice (23%), diarrhea (18%), nausea and vomiting (16%), and the detection of an abdominal tumor or hepatomegaly (22%) being the most common. These symptoms, in general, were caused by the PET and not, in most cases, by the ectopic release of somatostatin. Diabetes mellitus was present in 55% of patients with somatostatinomas in some series. However, in others, the percentage differed in patients with pancreatic tumors (95%) and those with intestinal tumors (21%; see Table

32-9). In patients with the somatostatinoma syndrome, 95% had diabetes mellitus.78 The diabetes was mild in most cases31,32 and could be controlled with oral hypoglycemic agents or small doses of insulin. Gallbladder and biliary tract disease occurred in up to 65% of patients in some series,9 including cholelithiasis in 35%, a massively dilated gallbladder without evidence of cholelithiasis in 10%, and obstructive jaundice caused by local tumor invasion in 10%. In other reports, gallbladder disease was present in 94% of patients with pancreatic tumors and 43% of those with intestinal tumors. Biliary calculi were present in 68% of patients with the somatostatinoma syndrome (see Table 32-9).78 Diarrhea and steatorrhea were reported in 18% to 92% of patients in some studies (see Table 32-9). Steatorrhea and diarrhea were reported in 83% and 92% of patients with pancreatic tumors and 12% and 10% to 38% of patients with intestinal tumors.78 Diarrhea characteristically consisted of 3 to 10 frequent, foul-smelling stools/day, with 20 to 76 g/day of steatorrhea.9 In some cases, the time course and severity of the diarrhea and steatorrhea paralleled that of the disease in that it worsened when metastases occurred and improved with successful tumor resection. Diarrhea was present in 37% and steatorrhea in 47% of patients with the somatostatinoma syndrome. Hypochlorhydria was found to be present in 70% of patients with somatostatinoma78 and in 86% of patients with pancreatic tumors and 17% with intestinal tumors in another study.9 Basal and stimulated gastric acid secretion was depressed. Hypochlorhydria was present in 26% of patients with the somatostatinoma syndrome (see Table 32-9). Weight loss ranged from 9 to 21 kg and was reported in one in three patients with pancreatic tumors and one in five patients with intestinal tumors. Mild to moderate anemia was reported in 15% to 67% of patients, with hemoglobin values of 10 to 13 g/dL. The somatostatinoma syndrome was more frequently associated with pancreatic tumors than duodenal somato­ statinomas (18.5 vs. 2.5%).78 Half of patients with somato­ statinoma had other endocrinopathies in one review,9 including MEN-I and MEN-II. In one series,9 20% of patients had hypoglycemic attacks and were diagnosed as having insulinomas. Somatostatinomas are rare in patients with MEN-I, occurring in only 0.65%.47,194

DIAGNOSIS AND DIFFERENTIAL DIAGNOSIS

Somatostatinomas are generally not diagnosed until a neuroendocrine tumor is discovered.31,32,217 Almost half all duodenal neuroendocrine tumors contain somatostatin,124 but most somatostatinomas (>98%) do not produce the somatostatinoma syndrome and the presenting symptoms such as abdominal pain, weight loss, jaundice, and diarrhea78,218 are not specific for a somatostatinoma. The symptoms characteristic of the somatostatinoma syndrome, similar to those of glucagonomas, are less pronounced than gastrinomas or insulinomas and probably are not detected until patients develop high somatostatin blood levels, which is late in the course of the disease when the tumor is large. In most cases, somatostatinomas are found at the time of laparotomy for cholecystectomy or during gastrointestinal imaging studies for various nonspecific complaints, such as abdominal pain or diarrhea. High plasma somatostatin concentrations have been reported with tumors outside the pancreas or intestine, such as patients with medullary thyroid carcinoma, small cell lung cancer, bronchial oat cell carcinoma, pheochromo­ cytoma, and other catecholamine-producing extra-adrenal paraganglionomas.9

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract The diagnosis of somatostatinoma is usually established by identifying that a tumor is a PET and contains increased numbers of somatostatin-containing D cells by immunocytochemistry. Increasingly, because of the widespread use of endoscopy and fine-needle cytology, somatostatinomas are being diagnosed from duodenal biopsy specimens or using cytology.124,219 The diagnosis of somatostatinoma syndrome requires the demonstration of elevated plasma levels of somatostatin, because somatostatin is frequently found by immunocytochemical studies, especially in duodenal tumors, without elevated plasma somatostatin levels.9,31,32,78,82 Modest somatostatin elevations should be interpreted with caution because they can also occur in nonendocrine disorders. Although plasma levels are usually elevated in pancreatic somatostatinomas, in duodenal or small intes­ tinal tumors, plasma levels often may be inconclusive or normal.216 The diagnosis of somatostatinoma when plasma somatostatin concentrations are only marginally elevated or normal may require the development of specific provocative tests. Arginine and tolbutamide stimulate plasma somatostatin increases; however, because arginine is a wellestablished stimulant of somatostatin release from normal D cells, arginine testing will likely not differentiate tumors from a normal response. Tolbutamide also may not be useful because even though it stimulates somatostatin release in animals, tolbutamide is reported not to change plasma somatostatin concentrations in normal volunteers. At present, the key to diagnosing the somatostatinoma syndrome is to be aware of its clinical features and to consider measuring plasma somatostatin levels in selected patients with diabetes (e.g., diabetics without a family history of diabetes), with gallbladder disease with a pancreatic mass, or with a history of unexplained diarrhea.215 Duodenal somatostatin-containing neuroendocrine tumors are being increasingly associated with von Recklinghausen’s disease (NF1).9,31,32,47,78,82,220 Such somatostatinomas resemble sporadic duodenal somatostatinomas in that they rarely are associated with symptoms of the somatostatinoma syndrome (0% to 2%), elevated plasma somatostatin levels are infrequent, and they frequently contain psammoma bodies (37% to 66%). NF1 is found in a higher proportion of duo­ denal than pancreatic somatostatinomas (43% vs. 17%) and somatostatinomas in patients with NF1 are less frequently malignant (47% vs. 69%).

TREATMENT

Patients with somatostatinomas may be severely malnourished and require correction of their nutritional deficiencies, which may require hyperalimentation. The diabetes mellitus is usually mild and hyperglycemia can be controlled with oral hypoglycemic agents or low doses of insulin.9

Medical Treatment

When four patients with a somatostatinoma were treated with the somatostatin analog octreotide, fasting plasma somatostatin concentrations decreased in three patients and symptoms caused by the somatostatinoma syndrome (diarrhea and diabetes) were improved in two patients.

Surgical Treatment

Surgery has been performed in 50% to 90% of patients with somatostatinomas.217 In one series, 65% of patients underwent successful resection,9 but the percentage cured was not stated, as was the case in the other series. Although an occasional patient might be cured, in most series because of the late diagnosis, cure is not possible. Five-year survival in patients without metastases was 100%,78 which was sig-

nificantly better than the 33% to 60% five-year survival seen in patients with metastases. In a number of patients treated with a combination of surgical resection and cytotoxic chemotherapy or radiofrequency ablation,31,32 60% of patients were alive six months to five years after diagnosis, but it is not clearly established whether debulking surgery extends survival. If imaging studies demonstrate possibly resectable tumor, current results suggest these patients would benefit from surgical resection.

GRFOMAS DEFINITION

GRFomas are tumors frequently originating in the gastro­ intestinal tract that secrete large amounts of growth hormone-releasing factor (GH-RF), which causes acromegaly. GRFomas were first described in 19827,8 and approximately 50 cases have been reported.

PATHOPHYSIOLOGY AND PATHOLOGY

GRFomas can originate in the lung (≈50%), with most in the right lung, pancreas (≈30%), small intestine (8% to 10%), or, rarely, in the adrenal gland.9,31,32,221,222,222 Most pancreatic GRFomas originate in the pancreatic tail. Multiple pancreatic GRFomas occurred in 30% in one series, generally in patients with MEN-I.47 GRFomas occur in less than 1% of MEN-I patients. GRFomas in the pancreas are generally large (>6 cm), ranging in size from 1 to 25 cm. Metastases were present in 33% to 39% of all GRFomas in two series, in 30% of patients with pancreatic GRFomas, and in two of three cases of intestinal GRFomas. Metastases were to regional lymph nodes and less frequently to the liver. In one series, there was no relationship among tumor size, plasma GH-RF levels, and the presence of metastases, with the three largest tumors unassociated with metastatic disease or invasion. Approximately 40% of pancreatic GRFomas occur in patients with gastrinomas and in 40% Cushing’s syndrome was also present. On light microscopic studies, typical features of a PET are seen, with trabecular or solid nests and sheets of uniform tumor cells. In electron microscopic studies, tumor cells containing 100- to 250-nm secretory granules are seen.7,9,221 Immunochemical studies demonstrated GH-RF–immunoreactive (GH-RF–IR) material in all tumors examined, with 10% to 80% of cells possessing GH-RF. GH-RF–IR was seen in 31% of various PETs in one study (n = 45)222 and in another study, in which no patients had acromegaly, GH-RF was present in 23% of PETs by immunocytochemistry and in 37% using radioimmunoassay. The known actions of GH-RF account for the clinical features of the syndrome. GH-RF is a 44–amino acid peptide7 that is a potent stimulant of the release of growth hormone from the anterior pituitary; therefore, patients present with acromegaly.

CLINICAL FEATURES

Patients are from 15 to 66 years old, with an average age of 38 to 39 years.9,221 Patients with intestinal GRFomas were younger, with two of the three patients younger than 20 years. A female predominance (73%) is seen for all GRFomas, as well as patients with only pancreatic GRFomas (78%). The clinical features fall into three categories: (1) acromegalic features caused by the excess of GH-RF; (2) clinical syndromes caused by hormones other than GH-RF; and (3) local symptoms caused by mass effects. Acromegalic features were indistinguishable from those of patients with

513

514

Section IV  Topics Involving Multiple Organs classic acromegaly and included enlargement of hands and feet, facial changes, skin changes, headache, and peripheral nerve entrapment.31,32 The duration from the onset of the acromegalic changes to the diagnosis was 5.3 years in patients with pancreatic GRFomas and six years for all GRFomas in another study. The syndromes caused by other hormones were caused by the presence of gastrinoma, Cushing’s syndrome or to hyperinsulinemic hypoglycemia. MEN-I was present in 16% of patients and hyperprolactinemia was observed in 70% of patients with GRFomas as compared with 50% of patients with somatotrophic adenomas.

DIAGNOSIS AND DIFFERENTIAL DIAGNOSIS

The diagnosis should be suspected in any patient with acromegaly without a pituitary adenoma but associated with hyperprolactinemia, with a paradoxical growth hormone response to thyrotropin-releasing hormone (TRH) or during an oral glucose tolerance test, or associated with an abdominal mass.31,32,221,222 Thus, the diagnosis should also be suspected in any patient with a pancreatic or intestinal tumor who develops clinical features of acromegaly. Because up to 33% of patients with pancreatic GRFomas have MEN-I, 40% have Cushing’s syndrome caused by an ACTH-producing pancreatic tumor, and 40% have Zollinger-Ellison syndrome, it should be particularly suspected in these patients. Of all patients with acromegaly, GRFomas are an uncommon cause, responsible for no cases in 177 consecutive patients with acromegaly. The diagnosis is established by demonstrating elevated plasma growth hormone levels (usually >5 µg/L in men and >10 µg/L in women) and the demonstration of elevated plasma GH-RF levels. In normal subjects and patients with acromegaly not caused by a GRFoma, plasma GH-RF–IR are within the normal to low-normal range, which in most laboratories is from 50 to 100 pg/mL.221,222 The lowest GH-RF–IR level in a patient with a proven GRFoma reported was 300 pg/mL. It has thus been suggested that a plasma GRF level higher than 300 pg/mL is strongly suggestive of the presence of a GRFoma.9 Plasma insulin-like growth factor I (IGF-I) is also elevated in patients with GRFomas.

TREATMENT

Tumor localization studies should be performed to evaluate the extent of disease (see later). In patients without metastatic disease to the liver, surgical resection of the GRFoma should be carried out. Prior to surgery and in those patients with nonresectable lesions, various agents may be helpful to reduce plasma growth hormone levels. Even though dopamine agonists such as bromocriptine are widely used in patients with classic acromegaly, they are able to reduce plasma GRF levels in only 25% of patients with GRFomas. Octreotide is the agent of choice.31,32,100-102,221,223,224 Treatment in a small number of cases suggests the long-acting forms of somatostatin (octreotide LAR or lanreotide SR) will also be effective. In most cases, but not all, octreotide significantly suppressed or normalized growth hormone and IGF-I levels and, in some cases, this was associated with pituitary shrinkage.9 The suppression of growth hormone secretion by octreotide was mainly caused by suppression at the pituitary level because plasma GH-RF levels never become undetectable. Surgical resection should be directed at the primary tumor, not the pituitary. Surgery has resulted in regression of the GRFoma syndrome in a small number of cases.7 The actual number of patients who achieve longterm cure is unknown.

PPOMAS AND NONFUNCTIONING PANCREATIC ENDOCRINE TUMORS DEFINITION

A PPoma is a tumor, usually of the pancreas, that secretes excessive amounts of pancreatic polypeptide (PP). The clinical symptoms are caused by local effects of the tumor itself, not by the actions of PP.225 Strictly speaking, a nonfunctioning PET is a tumor of the pancreas that has typical histologic features of a PET, is not associated with elevated plasma levels of any known peptide, and whose symptoms are entirely caused by the local effects of the tumor itself. The term nonfunctional PET is most widely used to indicate a PET occurring in a patient in whom there are no clinical symptoms caused by hormone overproduction. However, such nonfunctional PETs frequently secretes peptides. This would include PPomas and PETs secreting neurotensin, HCG subunits, ghrelin, chromogranin, or neuron-specific enolase.

PATHOPHYSIOLOGY AND PATHOLOGY

Except in patients with MEN-I,47,174 NF-PETs are usually large.9,22,225,226 In one series, the mean size was 4 cm.23 They are usually solitary tumors, except in patients with MENI.66,227,228 Sixty percent of sporadic NF-PETs occur in the pancreatic head. The malignancy rate in sporadic NF-PETs varies from 38% to 92% in different series. Histologically, NF-PETs are similar and cannot be differentiated from other PETs, even by immunocytochemistry. In one series of 30 NF-PETs, 50% had insulin-like immunoreactivity (IR), 30% glucagon IR, 43% PP IR, 13% somatostatin IR, and only 13% produced none of these peptides. Elevated plasma levels of chromogranin A and B are found in 69% to 100% of patients with these tumors, neuron-specific enolase in 31%, PP in 50% to 75%, α-HCG in 40%, and β-HCG in 20%. An elevated plasma PP level in a patient with a pancreatic tumor is suggestive of a PET because none of 53 patients with adenocarcinoma of the pancreas had elevated plasma PP levels. Infusions of PP into animals and humans have shown this peptide to have numerous biological effects, including the following: a net secretory effect on water and electrolytes in the small intestine; inhibitory effects on fluid, electrolyte, and enzyme secretion by the pancreas; effects on esophageal, gastric, intestinal, and gallbladder motility; and metabolic effects, such as inhibition of somatostatin or insulin release. In various studies,225 patients with PPomas have been reported to have symptoms attributed to elevated plasma levels of PP, including persistent watery diarrhea, diabetes mellitus, weight loss, decreased gastric acid secretion, peptic disease, flushing, and acute psychosis.9 Furthermore, the plasma PP level is frequently elevated in other symptomatic PETs. However, the symptoms of these patients do not differ from those without elevation of plasma PP levels; thus, it is now generally agreed that plasma elevations of PP are not associated with specific symptoms.35 At present, it is unclear why patients who have elevated plasma levels of PP do not have specific symptoms.

CLINICAL FEATURES AND DIAGNOSIS

Typically, the patient with a NF-PET is 40 to 60 years of age. These tumors occur approximately equally in both genders. The median delay in the time from diagnosis to the first symptoms varies from 0.5 to 2.7 years. In different studies, 36% to 56% of patients with NF-PETs presented with abdominal pain, 27% to 40% with jaundice, 28% to 46% with weight loss, and 8% to 40% with abdominal

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract mass.9,22,23,225,226,229,230 In 16% to 35% of patients, the tumors were found incidentally at surgery and the remaining patients presented with a variety of symptoms caused by the tumor mass. The main diagnostic challenge is differentiating the NF-PET from a nonendocrine pancreatic tumor and also in determining whether the tumor is associated with a symptomatic tumor syndrome (e.g., insulinoma, glucagonoma, gastrinoma). Elevated plasma levels of PP do not establish the diagnosis of a PPoma even when a pancreatic mass is present. Plasma PP levels are reported to be elevated in 22% to 71% of patients with functional PETs in various studies,28,225 as well as in nonpancreatic carcinoid tumors. In one large study,9 plasma PP levels exceeded a value of 1000 pg/mL in 45% of patients with various PETs, with this elevation in 32% of gastrinomas, 21% of insulinomas, 57% of glucagonomas, 74% of VIPomas, 33% of somatostatinomas, and 45% of carcinoid tumors. Furthermore, elevated plasma levels of PP can occur in other situations such as old age, after bowel resection, with alcohol abuse, during certain infections, in chronic noninfective inflammatory disorders, acute diarrhea, chronic renal failure, diabetes, chronic relapsing pancreatitis, hypoglycemia, or after eating. To increase the specificity of an elevated plasma PP level for a pancreatic tumor, an atropine suppression test has been proposed. In one study of 48 patients with elevated plasma PP levels, atropine (1 mg IM) did not suppress the levels in any of the 18 patients with PETs, but did suppress the level by 50% or more in all patients without tumors. Somatostatin receptor scintigraphy (SRS), discussed in detail in the following section on tumor localization, has also been shown to be useful in distinguishing pancreatic adenocarcinoma from a NF-PET.22,100-103,230,231

TREATMENT

With NF-PETs, treatment needs to be directed only at the tumor itself because no hormonal syndrome is present. Tenyear survival is better in patients with smaller tumors of 3 cm or smaller at presentation (80%) than in larger tumors (40%) in patients who were asymptomatic when the tumor was discovered (70% vs. 50% if symptomatic) and if no metastases were present (75% vs. 25% if there were metastases).9,23 Overall survival is 30% to 63% at five years, with a median survival of 72 months.22 In one series, a diagnosis of a NF-PET was never made preoperatively. Of 25 cases in this series, a Whipple pro­ cedure was done in five patients (20%), partial or total pancreatectomy in 25%, and tumor excision in 10%. The remaining patients had a biopsy only. The survival rates were 60% at three years and 44% at five years.9 In eight other studies, curative resection was attempted in 26% to 79% of patients with NF-PETs, with a five-year survival of 44% to 63% and a median survival varying from 2.5 to 4.8 years.226,229 The cure rate of these tumors at present is generally low. In one study,23 75% of patients undergoing resection were alive at the end of the study (mean follow-up, 4.4 years), which was better than the 47% survival in those not undergoing any surgical resection.

ing intestinal motility; and stimulating jejunal and ileal fluid, electrolyte, pancreatic protein, and bicarbonate secretion. Clinical features of patients with a PET and possible neurotensinomas include hypokalemia, weight loss, diabetes mellitus, cyanosis, hypotension, and flushing. The existence of a specific neurotensinoma syndrome has been questioned. In one study of patients with gastrinomas,35 those with or without an elevated neurotensin level did not clinically differ. Patients with PETs with Cushing’s syndrome (ACTHoma) have been reported.10,11,44,232 In various reviews, 4% to 16% of ectopic Cushing’s syndrome cases were caused by a PET.10 Cushing’s syndrome was reported in 19% of patients with gastrinoma and MEN-I; in these patients, the disease was of pituitary origin and was mild. Cushing’s syndrome occurs in 4% to 5% of sporadic gastrinoma cases; in these patients, Cushing’s syndrome was severe, caused by ectopic ACTH production, usually occurred with metastatic tumors, responded poorly to chemotherapy, and was associated with a poor prognosis. In a large prospective study,45 the development of Cushing’s in patients with ZES was shown to be an independent predictor of poor survival, with patients having a mean survival of 1.7 years after its onset. Cushing’s syndrome as the only manifestation of a PET occurs occasionally and may precede any other hormonal syndrome. In every case in one series, ectopic Cushing’s syndrome caused by a PET only occurred in the presence of metastatic disease. Hypercalcemia caused by a PET secreting PTH-rP or by an unknown hypercalcemic substance that mimics the action of PTH and causes hyperparathyroidism has been reported.9,16,233 The tumors are usually large and metastatic to the liver by the time of diagnosis, although in some cases resection of the PET resulted in cure or remission. PETs causing the carcinoid syndrome are rare but have been well described (see Chapter 31). Pancreatic carcinoid tumors are usually large and 67% to 88% are malignant.9,12,183 The carcinoid syndrome is present in 23% to 65% of these patients and in 23% in one review of 156 pancreatic carcinoids in the literature. Even though foregut carcinoids, which include pancreatic PETs,170 may lack Dopa decarboxylase, the enzyme that converts 5-hydroxytryptophan to serotonin (5-hydroxytryptamine), 84% of patients with PETs causing the carcinoid syndrome have increased urinary 5-hydroxyindoleacetic acid (5-HIAA) levels, which can be used for their detection. There is only one case described of a malignant PET secreting renin, which resulted in hypertension, although renin-secreting juxtaglomerular cell tumors, Wilms’ tumors, and ovarian tumors have been described.13 Similarly, there is only one reported case of a malignant PET secreting erythropoietin resulting in polycythemia.15 Ectopic release of erythropoietin has also been reported with pheochromocytomas, renal cell cancers, posterior fossa tumors, hemangioblastomas, and Wilms’ tumors. PETs secreting LH have been described, which can result in masculinization in females and loss of libido in males.9,14 PETs may contain ghrelin IR material and secrete ghrelin234 but, as noted in the historical section earlier, these do not seem to produce a distinct clinical syndrome.

OTHER PANCREATIC ENDOCRINE TUMORS In a few patients with PETs secreting the peptide neurotensin, a neurotensinoma syndrome has been proposed. Neurotensin is a 13–amino acid peptide, originally isolated from bovine brain, that has a number of biological effects, including causing tachycardia, hypotension, and cyanosis; affect-

TUMOR LOCALIZATION It is essential for the correct management of patients with PETs that extent (stage) of the tumor and often the site of the primary tumor be established (see Chapter

515

516

Section IV  Topics Involving Multiple Organs Table 32-10  Sensitivities of Imaging Methods for Localization of Pancreatic Endocrine Tumors (%)* Other Pancreatic Endocrine Tumors IMAGING STUDY Abdominal ultrasonography CT MRI Arteriography Selective venous sampling   Portal venous sampling   Post-intraarterial calcium Somatostatin receptor scintigraphy Endoscopic ultrasonography Intraoperative ultrasonography

Insulinomas

PRIMARY

LIVER METASTASES

30 (0-64) 31 (16-60) 10 (0-25) 60 (41-81)

22 42 27 70

(6-70) (33-100) (21-100) (35-100)

44 (14-76) 70 (35-100) 80 (67-100) 71 (33-86)

80 (63-96) 88 54 (15-56) 81 (57-92) 89 (80-100)

71 (17-94) ND 70 (58-77) 70 (40-100) 91 (80-100)

NA NA 93 (88-100) NA NA

*Values shown are mean values (and range). CT, computed tomography; NA, not applicable; ND, no data available. From references 9, 26, 28-30, 56, 69, 107, 242, and 253.

31).22,30,69,71,110,181,199,235 The ability of standard imaging studies, such as computed tomography (CT), ultrasonography (US), and MRI to localize the PET depends on the tumor size. CT and MRI localize less than 10% of PETs smaller than 1 cm, 30% to 40% of tumors 1 to 3 cm, and more than 50% of PETs larger than 3 cm in diameter.9,26,28,69 PETs are hypervascular tumors and the ability to localize different PETs appear to be influenced more by tumor size and location than PET type. Insulinomas are usually small (<1 cm) at the time of diagnosis, as are most duodenal gastrinomas, whereas most other PETs present late in their course and are large (>4 cm).29,46,58,65,71,73-77,135 Furthermore, insulinomas, nonfunctioning PETs, glucagonomas and VIPomas in adults are almost always located in the pancreas, whereas somato­ statinomas, gastrinomas, and VIPomas in children frequently occur extrapancreatically. Table 32-10 lists the sensitivities of the various localization studies to detect insulinomas, as well as the primary tumor and liver metastases of other PETs. The results with insulinomas are presented separately because they are almost always benign, small (<1 cm) at presentation, and entirely within the pancreas.84,85 Because there are no systematic localization studies of the less common PETs (all but gastrinomas, insulinomas, nonfunctioning PETs; see Table 32-1), and because gastrinomas closely resemble the other less common PETs in biological behavior and in imaging results, the results from gastrinomas were primarily used to determine the values in Table 32-10 for the other PETs. Overall, US, CT, and MRI are not very sensitive in localizing a primary tumor, being positive in 10% to 40% of cases.236,237 Of the standard imaging studies, selective abdominal angiography238 is the most sensitive for localizing the primary tumor, identifying 60% of small PETs such as insulinomas and 70% of the other PETs. Of the standard imaging studies, CT is generally recommended as the initial test because of its generally availability. For detecting metastatic disease to the liver, the sensi­ tivity of US is 44%; CT, 70%; MRI, 80%; and selective angiography, 71% (see Table 32-10). Studies9,26,69,236,237 have shown that improvements in MRI and CT with contrast greatly improves their sensitivity for the detection of metastatic disease to the liver in patients with PETs. Metastatic liver lesions are much easier seen on short TI inversion recovery (STIR) images by MRI compared with CT. Overall, even with these improvements, 40% to 60% of small primary tumors are missed by standard imaging studies and 5% to 30% of patients with metastatic disease to the liver will be missed.

SRS is becoming the procedure of choice for most PETs.56,69,239-241 PETs as well as a number of other tumors, including CNS tumors, lymphomas, breast cancer, and small cell lung cancer frequently possess increased densities of somatostatin receptors and can be imaged using SRS (see Chapter 31).65,100-103,242 For SRS, both 111In-DTPA-DPhe1 octreotide and 123I-Tyr3 octreotide have been used. 111 In-DTPA-DPhe1 octreotide (OctreoScan) is approved for use in the United States and has the advantage of a longer half-life (2.8 days), allowing longer imaging times and an easier labeling method using chelation instead of oxidative methods. In addition, this radiopharmaceutical agent is primarily excreted in the urine instead of bile, which allows tumors in the upper abdomen to be visualized better. Using in vitro autoradiography, 88% of carcinoids, 100% of gastrinomas, nonfunctional PETs, and glucagonomas, but only 67% of insulinomas (and no pancreatic adenocarcinomas), possessed somatostatin receptors. These results are consistent with SRS results from a large combined series,103 which detected carcinoid tumors in 89% of patients, gastrinomas in 77%, insulinomas in 53%, nonfunctional PETs in 83%, glucagonomas in 100%, and VIPomas in 80%. Furthermore, SRS frequently identifies hepatic and extrahepatic lesions not seen by other imaging modalities.241-244 The usefulness of SRS has been compared with that of other imaging modalities to localize gastrinomas or insulinomas.9,68,69,100-103,241,242 In patients with insulinomas, SRS detects only 10%, whereas CT, US, and MRI detect insulinomas in 20% of patients. In patients with ZES, which resembles other PETs except insulinomas in the frequency of the presence of somatostatin receptors, SRS was positive in 75% of patients. In one study involving 80 patients,242 SRS had the highest sensitivity for the primary tumor (58%) and was more sensitive than all the conventional imaging studies combined. SRS is now the method of choice to localize metastatic disease.241-244 Figures 32-6 and 32-7 demonstrate the enhanced sensitivity of SRS over conventional tumor localization modalities. In Figure 32-6, in a patient with a malignant PET, CT was negative, whereas SRS detected liver and lymph node metastases. In Figure 32-7, in a patient with a malignant gastrinoma and glucagonoma, the bone scan showed a questionable left scapula–rib metastasis, whereas the SRS showed scapula, spine, and pelvic bone metastases. SRS has the advantage of identifying unsuspected metastatic foci and is more sensitive than the other modalities for detecting liver or bone metastases (see Table 32-10).103,134,245 It has been found that the use of SRS after conventional imaging studies changes the clinical

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract CT R

L

SRS

Liver

R

L

Spleen Metastases Tumor in lymph nodes

Kidneys Figure 32-6.  Comparison of computed tomography (CT) (top) and somatostatin receptor scintig­raphy (SRS; bottom) results in a patient with a metastatic pancreatic endocrine tumor (PET). This PET was secreting pancreatic polypeptide, chromogranin A, and gastrin. The CT scan is negative, whereas the SRS image shows numerous liver and lymph node metastases. These results demonstrate the greater sensitivity of SRS compared with conventional imaging studies for localizing lymph node or liver metastases (ultrasonography, CT, MRI), as has been shown in a number of studies.100-102,241,242,246

management in 21% to 53% of patients.243,246,247 To obtain optimum results with SRS, single photon emission CT (SPECT) must be performed in addition to planar imaging. SRS can give false-positive localization for neuroendocrine tumors and is reported to be positive in certain thyroid disorders, granulomatous disease, accessory spleen, wound infections, and various arthritides. In one study,248 SRS had a false-positive rate of 12% for localizing a gastrinoma. However, when the clinical context is carefully considered, the percentage of patients in whom false-positive SRS localization altered management was 3%. Insulinomas have been found to overexpress receptors for GLP-1; a recent study249 has demonstrated that a radiolabeled GLP-1 analog can detect occult insulinomas not localized by other imaging modalities. Studies have demonstrated that endoscopic ultrasound (EUS) is a sensitive method to detect primarily PETs located in the pancreas (see Table 32-10).239,240 EUS localized an insulinoma in 81% (range, 57% to 92%) of patients with insulinomas and was superior to conventional imaging studies and almost as sensitive as calcium provocative testing (88%). In patients with gastrinomas,9,56,69 EUS localizes a gastrinoma in 67% to 70% (range, 40% to 100%) and

is comparable to angiography. In comparison to SRS, EUS localized 90% of insulinomas versus 54% for SRS, whereas for gastrinomas both EUS and SRS localized the tumor in 70% of patients. EUS is particularly sensitive for identifying PETs in the pancreas (>85%)239,240 and, in some studies, is clearly more sensitive than SRS, especially for insulinomas. EUS in MEN-I patients can detect PETs not seen by other modalities and may prove particularly useful for serial assessments to determine changes in the size of small lesions,174,228,250,251 which are frequently not routinely resected (see surgical section). Figure 32-8 shows the ability of EUS to identify an insulinoma in the pancreatic tail. EUS, especially of the pancreas, requires considerable expertise, whereas SRS can be performed in most nuclear medicine departments. Furthermore, EUS6 may yield false-positive results, although the false-positive rate is not clear, especially for tumors outside the pancreas. Cytologic confirmation of a functional PET is rarely needed, but EUS-guided fine-needle aspiration for cytology may sometimes be helpful in distinguishing a nonfunctional PET from another pancreatic tumor.252 EUS also provides information on adjacent lymph nodes, but provides no information on more distant sites of metastases, which may affect the surgical approach.56 Neither SRS nor EUS appear to identify small extrapancreatic PETs, especially duodenal gastrinomas.56,69 Some studies recommend SRS be used in combination with EUS in patients with gastrinomas to improve sensitivity.99 Functional localization of PETs by determining the site of the maximal hormonal gradient using selective venous blood sampling still remains a useful technique in some cases.105,107,108 Originally, this approach involved considerable expertise because transhepatic catheterization of portal venous tributaries was required. Furthermore, procedurerelated complications occurred in 20% of patients. A simplified method has been described.253 By using various secretagogues, such as secretin for gastrinomas or calcium for insulinomas, selective intra-arterial injection during angiography with hepatic venous hormone sampling has helped localize gastrinomas and insulinomas. This latter procedure is easier to perform than portal venous sampling, has fewer complications, and has equal to greater sensitivity and thus has replaced portal venous sampling. During this procedure the secretagogue is injected selectively into various arteries (superior mesenteric, splenic, right and left hepatic, gastroduodenal arteries) and, when the selected artery supplies the area of the PET, there is a sharp increase in the hepatic venous hormone concentration with the secretagogue injection. Figure 32-9 shows the ability of calcium infusion to localize an insulinoma accurately to the pancreatic body, a location that was suspected from the MRI study. In a comparative study in insulinomas, which are frequently smaller than 1 cm in diameter and difficult to localize, the intra-arterial calcium test with selective hepatic venous sampling was positive in 88% of patients, US in 9%, CT in 17%, MR in 43%, selective angiography in 36%, and portal venous sampling in 67%. Other studies support the conclusion that the intra-arterial calcium test with hepatic venous insulin sampling is a highly sensitive method to localize PETs. Furthermore, the intra-arterial calcium test may allow differentiation of the cause of the hypoglycemia and whether it is caused by nesidioblastosis or insulinoma.254 This is becoming an increasingly important distinction with the increased occurrence of hypoglycemia after gastric bypass surgery for obesity, which is primarily caused by nesidioblastosis,92 but on occasion can be caused by an insulinoma.255 Calcium infusion may also increase the release of hormones from VIPomas, PPomas, glucagonomas,

517

518

Section IV  Topics Involving Multiple Organs BONE SCAN

R

SRS

L

R

L

Liver mets

Figure 32-7.  Comparison of the ability of bone scanning (left panel) and somatostatin receptor scintigraphy (SRS; right panel) to localize bone metastases (mets) in a patient with a metastatic pancreatic endocrine tumor (PET). The PET was secreting both glucagon and gastrinoma. This patient had bone metastases in the lumbar spine, left pelvis, and left scapula. The SRS (right panel) demonstrates the metastases in each area (solid arrows), whereas the bone scan (left panel) shows only a questionable metastasis in the left scapula (solid arrow). The primary tumor in the pancreatic tail is shown by the dotted arrow on the SRS. Liver mets are seen on SRS. These results demonstrate the greater sensitivity of SRS in detecting bone metastases in patients with malignant PETs.134

A

B

Figure 32-8.  Endoscopic ultrasonography (EUS) for localization of pancreatic endocrine tumors (PETs). A, The EUS balloon is in the stomach. EUS shows a sonolucent pancreatic endocrine tumor (insulinoma) 2 cm in diameter (three arrows) near the splenic vein (small arrow, labeled V) B, EUS of the pancreatic neck region showing a 0.7-cm pancreatic endocrine tumor (gastrinoma). (Courtesy of Dr. Norman Thompson, Ann Arbor, Mich.)

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract tumors that cannot be identified by other methods68,122,123,125,126 and its use has been shown to increase the disease-free survival rate (see Fig. 32-4). Newer modalities are being used that may increase the sensitivity for detection and localization of primary and metastatic tumors in patients with PETs. An example is the use of hybrid scanning, particularly combining CT and SRS,57,258,259 and the use of scanning, either alone or combined with MRI or CT. SRS alone does not allow exact localization of a lesion within a given region, such as the distinction between a gastrinoma in the duodenum versus the pancreatic head. However, SRS is increasingly being combined with CT (SRS-CT hybrid scanners), which allows better definition of the exact location. Similarly, numerous recent studies, primarily from Europe, have shown that PET scanning for PETs using 11C-5-hydroxytryptophan or 68Galabeled somatostatin analogs may have improved sensitivity over SRS alone or conventional imaging studies (see Chapter 31).241,260-263 In a number of studies,260,264 PET scanning was reported to have greater accuracy and/or sensitivity than SRS alone. At present, PET is not approved for this use in the United States and has limited availability.

T

MRI

INTRA-ARTERIAL CALCIUM GLUCONATE INJECTION 1000

Plasma insulin level (µM/mL)

Splenic artery 800

MANAGEMENT OF METASTATIC PANCREATIC ENDOCRINE TUMORS

600

400

200

Hepatic artery GDA SMA

0 Pre

20

40

60

90

Seconds (postinjection) Figure 32-9.  Localization of an insulinoma by MRI (top) and by intraarterial calcium injection and hepatic venous sampling for insulin concentrations (bottom). Top, On the MRI scan, a probable insulinoma (T) was seen in the pancreatic body, an area supplied by the splenic artery. Bottom, Ca gluconate (10%; 0.025 mEq Ca/kg) was selectively injected into the superior mesenteric, gastroduodenal, common hepatic, and splenic arteries. Venous samples were collected prior to and 30, 40, 60, and 90 seconds postinjection and assayed for insulin concentration. A significant (>50%) increase in hepatic venous insulin concentration occurred at all times after calcium injection into the splenic artery but not after calcium injection into the other vessels. At surgery, an insulinoma was found in the pancreatic body. GDA, gastroduodenal artery; SMA, superior mesenteric artery.

GRFomas, or somatostatinomas, so that a similar approach may be useful with these tumors.9 At surgical exploration a number of procedures are helpful in localizing different PETs. Intraoperative ultrasonography (IOUS) is useful, particularly for localizing intrapancreatic PETs, and will identify some insulinomas not found by other means.256,257 It is essential for the surgeon to perform a Kocher maneuver to palpate the pancreatic head carefully as well as mobilize the pancreatic tail to allow its careful palpation. Lastly, in patients with gastrinomas, a routine duodenotomy is essential to detect small duodenal

The treatment of all metastatic PETs is considered together because in most aspects it is similar for each of the tumors (see Chapter 31). Metastatic PETs are relatively slowgrowing compared with other more common malignant adenocarcinomas.45,135,265,266 The long-term natural history of most functional PETs is not known, because until recently effective treatment for the clinical syndrome with functional PETs was not available, and therefore patients often died of complications of the hormonal excess rather than the tumor. However, this is changing with the recent availability of agents such as long-acting somatostatin analogs. In contrast, with nonfunctional tumors and gastrinomas, for which effective therapy for the gastric hypersecretion has existed for more than 30 years, the natural history of the malignant tumor itself has been assessed. Because of the similar biological behavior of all PETs, the assessment of these latter tumors will likely also provide insights into the natural history of all these less common malignant PETs.28,135 Insights into PET biology and natural history have identified prognostic factors that determine survival.25 An awareness of these factors is essential in planning the type and timing of treatment of advanced disease.

TUMOR BIOLOGY, PROGNOSTIC FACTORS, AND SURVIVAL

With increased ability to control the hormone excess state, the survival of patients with functional PETs is increasingly being determined by the tumor’s biology and natural history of the PET’s growth pattern.25,45,54,265 Information has mainly been obtained from studies of the natural history and prognostic factors determining survival in patients with nonfunctional PETs22 and from patients with gastrinomas because of the ability to control the gastric hypersecretion medically in these patients for a number of years.45 Studies have demonstrated that PETs grow at different rates in different patients.25,45,135,137,266 As noted, in approximately 25% of patients with gastrinomas followed long term, the gastrinoma demonstrates aggressive growth, whereas in the remaining 75% growth is indolent or no

519

520

Section IV  Topics Involving Multiple Organs Table 32-11  Prognostic Factors Associated with Decreased Survival in Patients with Various Pancreatic Endocrine Tumors Female gender Absence of MEN-I syndrome Liver metastases Extent Growth Lymph node metastases Bone metastases Incomplete tumor resection Nonfunctional tumor Ectopic Cushing’s syndrome (gastrinomas) Depth of tumor invasion Tumor size Histologic features High nuclear atypia Poor tumor differentiation High growth indices (Ki-67 index > 2%, PCNA expression) Capsular invasion Vascular or perineural invasion Necrosis Flow cytometric features (i.e., aneuploidy) Laboratory findings Elevated serum chromogranin A (in some studies) Elevated serum gastrin level (gastrinomas) Lack of progesterone receptors Ha-Ras oncogene or p53 overexpression Molecular biological features High HER2/neu gene expression (gastrinomas) High 1q loss of heterozygosity (gastrinomas) Increased EGF or IGF receptor expression (gastrinomas) Chromosomal instability CGH findings (loss = 1p, 3p, 3q, 6q, 9q, 12q; gains = 7q, 17q, 17p, 20q) Gene microarray studies may also have prognostic potential (see text). EGF, epidermal growth factor; IGF, insulin-like growth factor; MEN-I, multiple endocrine neoplasia type I; PCNA, proliferating cell nuclear antigen; PET, pancreatic endocrine tumor. From references 40, 45, 53-56, 100-102, 135, 265, 266, 268, and 314-317.

growth occurs. Similarly, aggressive growth occurred in only 40% of patients with liver metastases and all deaths occurred in this subset of patients. At present, the molecular basis for this difference in growth remains unclear. In different PETs, prognostic factors have been defined (Table 32-11). The most important prognostic factor in all studies was the development of liver metastases.77,267,268 In one large study involving 221 patients with gastrinomas, the 15-year survival for all patients was 90%, without liver metastases was 96%, and with liver metastases was 26%.45 The development of liver metastases, their extent (one lobe, both, diffuse), presence of bone or lymph node metastases, larger primary tumor size, primary location (pancreatic gastrinomas have a worse prognosis than duodenal), various histologic features, laboratory, flow cytometric, and molecular features all have predictive value. Recently, microarray analysis53,59,60,62 has been used to attempt to identify genes that contribute to aggressive growth of PETs. In these studies, a large number of genes have been found to be up-regulated or down-regulated in association with aggressive growth behavior, but none have been shown to be clinically useful for management of patients. Because of the variable growth rate of different PETs and the importance of prognostic factors in patient management, there have been a number of proposed classification systems for PETs and, for the first time, a TMN classification50 (see Chapter 31). Some been shown to have prognostic value.269-272

Two of the most important clinical outcomes are the development of bone metastases and ectopic Cushing’s syndrome, both of which were independent predictors of poor survival, with a mean survival from their onset of less than two years.25,45 Most would agree that treatment directed at metastatic disease that is increasing in size is indicated, as well as treatment to prevent the development of metastases. However, there is currently no agreement about what type of therapy is most appropriate for patients with metastatic disease, when therapy should be started, and even the efficacy of various therapies. Chemotherapy,273-276 debulking (cytoreductive) surgery, with or without chemotherapy,56,176,177 hepatic arterial embolization with or without chemoembolization,277-279 hormonal therapy with longacting somatostatin analogs,100-102,183,280 interferon-α, liver transplantation,183,281-283 and targeted radiotherapy using radiolabeled somatostatin analogs284 have all been reported to be useful in some cases (see Chapter 31).

CHEMOTHERAPY

Most studies of chemotherapy in metastatic PETs have included mixtures of patients with functional PETs, occasionally nonfunctional PETs, and in some cases patients with carcinoid tumors.28,273,274 Results from these studies are limited in a number of ways. There are often small numbers of cases. Some studies have suggested that responsiveness to chemotherapy is equal in the different PETs, but others have suggested there may be important differences. For example, up to 93% of metastatic glucagonomas are reported to respond to dacarbazine (DTIC) including some complete remissions, whereas DTIC produces a low response rate in other PETs. Similarly, up to 90% of VIPomas respond to streptozotocin, whereas only 5% to 40% of metastatic gastrinomas respond to streptozotocin, with no complete responses. The current recommendation is the combination of streptozotocin and doxorubicin.273-275 This recommendation is based on two studies from the Eastern Cooperative Oncology Group (ECOG) published in 1980286 and 1992.285 The study in 1980 demonstrated that streptozotocin plus 5fluorouracil (5-FU) is more effective than streptozotocin alone. The 1992 study demonstrated that streptozotocin plus doxorubicin causes tumor regression in 69% of patients, which is significantly better than the 45% with streptozotocin alone or 30% with chlorozotocin. Furthermore, the patients treated with streptozotocin and doxorubicin had a significantly better survival. Streptozotocin is a glycosamine nitrourea compound originally derived from a Streptomyces species, and in preclinical studies was found to have cytotoxic effects on pancreatic islets. Streptozotocin was found to have clinical effectiveness against a PET in 19689 and since then has been used as the initial agent either alone or in combination with other agents for treating metastatic PETs. In various series, streptozotocin alone produces an objective tumor response in 36% to 62%.286 In contrast, other agents, such as doxorubicin, DTIC, tubercidin, etoposide, and carboplatin, have generally had a lower response rate of 6% to 33%. Streptozotocin causes nausea and vomiting in almost all patients and transient dose-related renal dysfunction, including proteinuria (40% to 50%) and a decrease in creatinine clearance, as well as abnormalities in hepatic function, leukopenia, and thrombocytopenia in 6%. In the 1992 ECOG study, nine patients developed renal failure and seven required dialysis. The nausea and vomiting can now be controlled in almost all patients using 5-HT3 receptor antagonists such as ondansetron. Chlorozotocin is structurally closely related to streptozotocin but causes less nausea and vomiting. When given alone or combined with

Chapter 32  Endocrine Tumors of the Pancreas and Gastrointestinal Tract 5-fluorouracil, it gives similar results to those seen with streptozotocin. Because of the limited effectiveness of single agents, various combinations have been investigated.273,274 The combination of streptozotocin and 5-fluorouracil was found to be more effective than streptozotocin alone. However, in a later study, streptozotocin plus 5-FU was less effective than streptozotocin plus doxorubicin.285 In the 1980 ECOG study,286 the 42 patients treated with streptozotocin had a 36% response rate, with 12% showing a complete response, whereas with streptozotocin plus 5-FU, 63% demonstrated a response, with 33% having a complete response. Response rates in this study between different functional tumors or between functional and nonfunctional tumors did not differ. In more recent prospective studies, the response rate with streptozotocin plus 5-FU was significantly lower in patients with metastatic gastrinomas to the liver. No patient had a complete response and there was no difference in survival for responders and nonresponders. At present, the difference in response rates between the early study and more recent ones remains unexplained. Streptozotocin has been used in combination with other agents such as doxorubicin or tubercidin in small numbers of cases, with response rates reported from 20% to 100%. Streptozotocin combined with 5-FU, with doxorubicin, with both these agents, with 5-FU plus tubercidin, or with doxorubicin and cisplatin has been used in different studies.9,273-275,285,286 Only the combination of streptozotocin and doxorubicin are established as superior to streptozotocin and 5-FU. The combination of etoposide and cisplatin has been evaluated in patients with PETs and carcinoid tumors.273,274,287,288 In one study,287 12 of 18 anaplastic neuroendocrine tumors, 2 of 14 PETs, and 0 of 13 of metastatic carcinoid tumors demonstrated partial to complete regression. In a second study,288 only 1 of 12 patients with a welldifferentiated neuroendocrine tumor responded, whereas 42% of 41 patients with a poorly differentiated tumor showed an objective tumor response. Hematologic toxicity occurred in 60% and there was one treatment-related death.

SURGICAL TREATMENT

Systematic removal of all resectable tumor (debulking or cytoreductive surgery) has been recommended, if possible, for all PETs, including gastrinomas, VIPomas, glucagonomas, and somatostatinomas.128,176,177,181,199,276,289 In various studies in patients with advanced PETs, cytoreductive surgery is reported to result in occasional cures, five-year survivals of 75% to 80% in resected patients, and increased survival over those not undergoing resection.56 Unfortunately, such resection is possible in only a small proportion of patients (5% to 15%). Even though this approach is recommended, it is not clear whether such an approach actually increases survival. This approach may be required in patients with symptomatic PETs in whom octreotide or the use of chemotherapy alone cannot reduce plasma hormone levels sufficiently to control symptoms.

HEPATIC ARTERY EMBOLIZATION AND CHEMOEMBOLIZATION

Hepatic artery embolization with or without postocclusion chemotherapy has been used successfully in patients with metastatic PETs to the liver (see Chapter 31).277-279 Because the liver derives only 20% to 25% of its blood supply from the hepatic artery (the rest comes from the portal vein), and because most PETs are vascular with an arterial supply, hepatic artery embolization has been possible. In some studies, 68% to 100% of treated patients have symptomatic

improvement. Chemotherapy using doxorubicin or other chemotherapeutic agents in iodized oil, combined with gelatin or sponge particles, has been reported to improve symptoms in 66% to 100% of patients and decrease tumor size and/or hormone levels in 37% to 100% of patients. Almost all patients report abdominal pain, nausea, vomiting and fever, usually lasting three to ten days, with severe complications occurring in 10% to 15%, including hepatic failure, acute renal failure, infection, and death. With the availability of SRS, it is now possible to assess the extent of metastatic disease easily. In a patient with diffusely metastatic disease to the liver, with minimal or no bone metastases, in whom hormone symptoms cannot be controlled by octreotide, chemotherapy, or other medical treatments, this hepatic artery therapy should be considered.

RADIOFREQUENCY ABLATION

Radiofrequency thermal ablation (RFA) works by converting RF waves into heat, which results in cellular destruction at temperatures higher than 60°C.279,290,291 RFA can be applied to liver metastases if their number is limited (usually less than five) and if they are not too large (usually <3.5 cm diameter), RFA can be performed laparoscopically or at the time of surgical exploration. Response rates are high (usually >80%), have lasted more than two years and the procedure-related morbidity is low (<15% to 20%). These results suggest that RFA may be a useful approach in patients with unresectable metastatic PETs, particularly patients with functional PETs, which are difficult to control medically.

SOMATOSTATIN ANALOGS

Except for insulinomas, more than 90% of all PETs possess somatostatin receptors and these receptors may mediate many of the effects of somatostatin analogs on these tumors.100-103,241 The effect of somatostatin analogs on tumor growth function and size has been examined in a number of studies.280,292,293 Somatostatin analogs cause a decrease in PET tumor size in fewer than 17% of patients. However, stabilization of metastatic disease is seen in 30% to 80% of patients. Tumor stabilization is also reported with depot long-acting formulations of somatostatin analogs. Studies have demonstrated that somatostatin analogs can induce apoptosis in neuroendocrine tumors, which may contribute to its tumoristatic effects.294 At present, it is unclear whether this stabilization will lead to increased survival.

INTERFERON-α

Interferon-α has been reported to be effective at controlling symptoms in a number of patients with PETs.101,102,274,281,292 In a review9 of 322 patients with various neuroendocrine tumors treated with interferon-α, 43% of patients had a biochemical response (i.e., a >50% decrease in hormone levels) and 12% had a decrease in tumor size, with a mean duration of 20 months (range, 2 to 96 months). Disease stabilization is seen in 75% to 80% of patients with metastatic neuroendocrine tumors. Studies295 have demonstrated that interferon-a can induce an increase in bcl-2 expression in NETs, which may contribute to the tumoristatic effect by stabilizing cells at the G0 phase of the cell cycle (see Chapter 3). These results suggest that interferon-a, similar to somatostatin analogs, may possibly extend survival by decreasing tumor growth rate. Interferon-α has been used in combination with somatostatin analogs,296-298 including randomized, prospective studies, to compare its antigrowth effect with either agent alone. The results did not support the results of nonrandomized studies, which had shown an enhanced antitumor action of the combination over

521

522

Section IV  Topics Involving Multiple Organs each drug alone. A number of methodologic issues have been raised with these studies, and thus it remains controversial whether this drug combination is better than either agent alone.

LIVER TRANSPLANTATION

Liver transplantation has been carried out in a small number of patients with metastatic PETs.281-283,299 In a review of 103 cases of patients with malignant NETs who underwent liver transplantation (including 48 PETs), the five-year survival rate was 45%. However, the recurrence-free survival rate was low (<24%).9 Recent series and reviews have suggested that liver transplantation should be restricted to young patients with metastases limited to the liver with a previously resected primary PET who have failed other therapies to control hormonal and/or tumor symptoms.

SOMATOSTATIN RECEPTOR-DIRECTED RADIOTHERAPY

Recently, there has been an attempt to use radiolabeled somatostatin analogs for antitumor treatment, especially in Europe (see Chapter 31 for more detailed discussion).279,284,300-305 At present, this form of treatment is not approved in the United States.

POSSIBLE NEW TREATMENTS

Some studies, generally small series and in preliminary form, have reported effects of a number of newer novel agents in patients with progressive malignant NETs (both PETs and carcinoids) and have suggested that a number have some activity against PETs. These include growth factor inhibitors directed against IGF-1, transforming growth factor-α, platelet-derived growth factor (PDGFR), epidermal growth factor, or vascular endothelial growth factor (VEGF), or inhibitors of their receptors (imatinib–platelet-derived growth factor receptor, gefitinib–epidermal growth factor receptor [EGFR]), inhibitors of mTor (temsirolimus, everolimus), angiogenesis inhibitors (SU11248-EGFR, c-kit, PDGFR, endostatin, bevacizumab-VEGFR), VEGF tyrosine kinase inhibitors (sunitinib, vatalanib, sorafenib) and a DTIC-related compound, temozolomide.303,306-312 In one study involving 107 patients (66 PETs, 41 carcinoids)313 using sunitinib, which has activity against a number of tyrosine receptors, including VEGFR-1, VEGFR-2, VEGFR-3, and PDGFR, 17% of patients showed a decrease in tumor size and 68% showed stabilization. In a study with temsirolimus, an mTor inhibitor, the response rate was 7%, and

in another study with another mTor inhibitor, everolimus, it was 15%.308,309,312 A number of these agents, particularly sunitinib, various mTor inhibitors, and bevacizumab have shown sufficient activity that larger studies are planned, either alone or in combination with other agents.

KEY REFERENCES

de Herder WW, Niederle B, Scoazec JY, et al. Well-differentiated pancreatic tumor/carcinoma: Insulinoma. Neuroendocrinology 2006;84:183-8. (Ref 30.) Duerr EM, Chung DC. Molecular genetics of neuroendocrine tumors. Best Pract Res Clin Endocrinol Metab 2007; 21:1-14. (Ref 53.) Gibril F, Jensen RT. Diagnostic uses of radiolabelled somatostatin receptor analogues in gastroenteropancreatic endocrine tumors. Dig Liver Dis 2004; 36:S106-20. (Ref 69.) Jensen RT, Berna MJ, Bingham MD, Norton JA. Inherited pancreatic endocrine tumor syndromes: Advances in molecular pathogenesis, diagnosis, management and controversies. Cancer 2008; 113(Suppl): 1087-43. (Ref 47.) Jensen RT, Niederle B, Mitry E, et al. Gastrinoma (duodenal and pancreatic). Neuroendocrinology 2006; 84:173-82. (Ref 71.) Kloppel G. Tumour biology and histopathology of neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007; 21:15-31. (Ref 33.) Kwekkeboom DJ, de Herder WW, Kam BL, et al. Treatment with the radiolabeled somatostatin analog (177 Lu-DOTA 0,Tyr3)octreotate: Toxicity, efficacy, and survival. J Clin Oncol 2008; 26:2124-30. (Ref 303.) McLean AM, Fairclough PD. Endoscopic ultrasound in the localisation of pancreatic islet cell tumours. Best Pract Res Clin Endocrinol Metab 2005; 19:177-93. (Ref 239.) Metz DC, Jensen RT. Gastrointestinal neuroendocrine tumors: Pancreatic endocrine tumors. Gastroenterology 2008; 135:1469-92. (Ref 26.) Modlin IM, Oberg K, Chung DC, et al. Gastroenteropancreatic neuroendocrine tumours. Lancet Oncol 2008; 9:61-72. (Ref 57.) Oberg K, Eriksson B. Endocrine tumours of the pancreas. Best Pract Res Clin Gastroenterol 2005; 19:753-81. (Ref 183.) Oberg K, Kvols L, Caplin M, et al. Consensus report on the use of somatostatin analogues for the management of neuroendocrine tumors of the gastroenteropancreatic system. Ann Oncol 2004; 15:966-73. (Ref 100.) Que FG, Sarmiento JM, Nagorney DM. Hepatic surgery for metastatic gastrointestinal neuroendocrine tumors. Adv Exp Med Biol 2006; 574:43-56. (Ref 176.) Rindi G, Kloppel G, Alhman H, et al. TNM staging of foregut (neuro) endocrine tumors: A consensus proposal including a grading system. Virchows Arch 2006; 449:395-401. (Ref 50.) Sundin A, Garske U, Orlefors H. Nuclear imaging of neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007; 21:69-85. (Ref 241.) Yao JC. Molecular targeted therapy for carcinoid and islet-cell carcinoma. Best Pract Res Clin Endocrinol Metab 2007; 21:163-72. (Ref 312.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

33 Gastrointestinal Consequences of Infection with Human Immunodeficiency Virus C. Mel Wilcox

CHAPTER OUTLINE Odynophagia and Dysphagia  524 Evaluation and Management  525 Diarrhea  526 Evaluation and Management  526 Abdominal Pain  530 Evaluation and Management  530

For the 15 years of the acquired immunodeficiency syndrome (AIDS) epidemic that antedated effective antiviral therapy for the human immunodeficiency virus (HIV), the world witnessed an explosion of cases typically manifested by opportunistic infections (e.g., Pneumocystis jiroveci pneumonia) and neoplasms (e.g., Kaposi’s sarcoma). Early in the epidemic, the focus of attention was on characterizing these disorders, and when effective, using prophylactic antimicrobial therapy. In 1995 the concept of highly active antiretroviral therapy (HAART) was born, and the face of the epidemic changed overnight. HAART decreases viral replication and, consequently, circulating HIV. In some patients HIV becomes undetectable in the blood. Associated with a reduction in viral load, there is substantive improvement in immune function that can be assessed by objective measures such as an increase in the CD4 lymphocyte count and clinically by a decrease in opportunistic infections (OIs), as well as improved survival.1-3 With immune reconstitution provided by HAART, both primary and secondary prophylaxis against a variety of OIs also may be discontinued.4 The current focus of management thus centers around viral control rather than prevention and treatment of opportunistic infections. Although dramatic changes have been witnessed as a result of HAART, the global epidemic continues, ensuring that this disease will be present for years to come.5 With the immune reconstitution associated with HAART, there also has been a shift to the management of chronic diseases, as well as drug side effects. Hepatitis C virus (HCV) infection is highly prevalent in HIV-infected patients. Because of HAART, chronic liver disease has assumed increasing importance, as evidenced by reports demonstrating that end-stage liver disease, most often a result of HCV, is a leading cause of hospitalization and death.6 Similarly,

Anorectal Disease  531 Evaluation and Management  531 Gastrointestinal Bleeding  531 Evaluation and Management  532 Hepatomegaly and Abnormal Biochemical Liver Tests  532 Evaluation and Management  534

HIV-infected patients responding to HAART who have gastrointestinal (GI) complaints are more likely to have druginduced side effects or nonopportunistic gastrointestinal infections, shifting management strategies back to disorders prevalent in normal hosts.7 Because patients are generally approached on the basis of clinical presentation, accordingly, this chapter is organized primarily around symptom diagnosis (diarrhea, odynophagia and dysphagia, anorectal disease, abdominal pain, GI bleeding, jaundice, and hepatomegaly). Specific HIV-related disorders (limited to HIV-1) and their treatments are presented within the context of their most common associated symptoms. In addition, the relevant effect of HAART in relation to these symptom complexes and diseases is discussed. Throughout this chapter when referring to patients with AIDS, we specify those patients with a CD4 count of less than 200/mm3 who are at risk for or who have developed opportunistic disorders. Generally these are patients who are not receiving or have failed HAART. Some patients responding to HAART may have an absent HIV viral load but may still have a CD4 count less than 200/mm3 depending on the CD4 level when HAART was initiated. However, these are different patients than those with a similar CD4 count, but high viral load, and who remain at risk for AIDSrelated complications. Although HAART has dramatically altered the occurrence of GI complications, many of the same principles of management established before HAART remain applicable. In general, the approach to investigating GI symptoms in the patient with AIDS parallels that of non–HIV-infected patients. Several general points must be considered when evaluating GI symptoms in AIDS: 1. Clinical signs and symptoms infrequently suggest a specific diagnosis.

523

CD4 lymphocyte count per mm3

524

Section IV  Topics Involving Multiple Organs 300 Lymphoma Kaposi’s sarcoma Tuberculosis 150

75

Thrush Candida esophagitis Herpes simplex virus esophagitis Idiopathic esophageal ulcer Microsporidiosis Cryptosporidiosis Mycobacterium avium complex Cytomegalovirus disease

0 Time (years) Figure 33-1.  Time line of opportunistic disorders based on CD4 lymphocyte count. (From Wilcox CM, Saag MS. Gastrointestinal complications of HIV infection: Changing priorities in the HAART era. Gut 2008; 57: 861-70.)

2. GI symptoms in a patient on HAART are most often drug induced or nonopportunistic in etiology. 3. Risk stratification for an opportunistic disorder may be predicted on the basis of the extent of immunocom­ promise (i.e., CD4 count >200/mm3 favors common bacteria and other nonopportunistic diseases; CD4 count <100/mm3 favors cytomegalovirus [CMV], fungi, Mycobacterium avium complex [MAC], and unusual protozoa) (Fig. 33-1). 4. In AIDS, GI pathogens are usually part of a systemic infection (e.g., CMV, MAC). Thus, identification of a pathogen outside the gut in the appropriate clinical setting may negate GI evaluation. 5. Although evaluation should proceed from less invasive to more invasive and should be dictated by the severity and acuity of symptoms, early endoscopy in selected settings is key. 6. Multiple infections are common. 7. Evidence of tissue invasion should be sought as a hallmark of pathogenicity. 8. Without improvement of immune function (HAART), recurrence of OIs is almost uniform, necessitating maintenance of antimicrobial therapy. 9. Treatment of all opportunistic disorders should include HAART. The natural history of opportunistic diseases can be favorably altered by HAART.

ODYNOPHAGIA AND DYSPHAGIA Before the era of HAART, esophageal complaints (dyspha­ gia and odynophagia) were commonly reported to occur in at least one third of patients during the course of HIV disease. Because of HAART, the incidence of esophageal disease has fallen, and the number of patients with diseases not unique to AIDS, such as gastroesophageal reflux, has risen.7 Candida albicans, the most frequent esophageal infection in AIDS, frequently coexists with other disorders in this setting. Although most cases of Candida occur in the setting of AIDS, Candida esophagitis may occur during primary HIV infection as a result of transient immuno­ suppression.8 Oral thrush often predicts concurrent esopha-

Figure 33-2.  Cytomegalovirus and herpes simplex virus esophagitis. Diffuse circumferential ulceration is seen on this endoscopic view and the gastrointestinal junction is seen in the distance. In patients with acquired immunodeficiency syndrome, multiple pathogens are frequently found. (From Wilcox CM. Atlas of Clinical Gastrointestinal Endoscopy. Philadelphia, Pa: Saunders; 1995. p 28.)

gitis; however, the absence of thrush does not exclude the possibility of esophageal candidiasis. Overall, the positive and negative predictive values of thrush for Candida esophagitis for patients naïve to HAART are 90% and 82%, respectively.9 Patients with esophageal candidiasis generally complain of substernal dysphagia; odynophagia, when present, is usually not severe. Definitive diagnosis is established by upper endoscopy, which reveals either focal or diffuse plaques in association with mucosal hyperemia and fri­ ability; well-circumscribed ulcer(s) suggests an additional process. Biopsies show desquamated epithelial cells with typical-appearing yeast forms; fungal invasion is usually present only in the superficial epithelium. Although CMV is the most commonly identified pathogen in AIDS, its association with esophageal disease is less frequent than Candida. CMV causes mucosal ulceration; thus patients with CMV esophagitis complain of odynophagia or substernal chest pain, characteristically severe.10 Dysphagia is much less common than in patients with Candida esophagitis and is rarely the primary complaint. Fever is rare. Generally, upper endoscopy reveals extensive ulcerations that are large and deep, although the endoscopic pattern is variable (Fig. 33-2).11 Candidal coinfection is common. Mucosal biopsies characteristically demonstrate viral cytopathic effect in mesenchymal and/or endothelial cells in the granulation tissue. As is typical for gut involvement with CMV, characteristic inclusions may be absent, necessitating confirmation by immunohistochemical stains. Biopsy of granulation tissue in the ulcer base provides the highest yield for viral cytopathic effect, whereas viral culture is less sensitive and cytologic brushings are unhelpful.12 A syndrome of nonspecific (idiopathic, aphthous) esophageal ulceration is common (Fig. 33-3).10 The clinical presentation and endoscopic appearance are indistinguishable from CMV. Criteria for diagnosis of idiopathic ulcers include the following: (1) endoscopic and histopathologic ulcer; (2) no evidence of viral cytopathic effect by both routine histol-

Chapter 33  Gastrointestinal Consequences of Infection with Human Immunodeficiency Virus

Figure 33-3.  Human immunodeficiency virus–associated idiopathic ulcers. Multiple well-circumscribed ulcerations throughout the esophagus are evident in this endoscopic view. The ulcers have a punched-out appearance, with normal-appearing intervening mucosa. The ulcers seem to be raised resulting in a heaped-up appearance. (From Wilcox CM. Atlas of Clinical Gastrointestinal Endoscopy. Philadelphia, Pa: Saunders; 1995. p 75.)

ogy and immunohistochemical studies; and (3) no clinical or endoscopic evidence of reflux disease or pill-induced esophagitis. As with CMV, these ulcers occur in late-stage disease, with most patients having a CD4 count less than 50/mm3. However, they have also been described in patients with the acute HIV seroconversion syndrome. The pathogenesis of these ulcers remains unknown, but mucosal HIV infection does not appear to be causative. In contrast with other immunocompromised hosts, herpes simplex virus (HSV) esophagitis is infrequent in AIDS.10 In immunocompetent patients, esophagitis is usually due to HSV type 1; however, AIDS patients may have esophagitis due to either type 1 or type 2 herpes. The disease is similar to herpetic infections of other mucous membranes in that the pathogenetic features follow a predictable sequence: discrete vesicles form, then shallow ulcers, which finally coalesce into regions of diffuse shallow ulceration. It is during this late stage of diffuse esophagitis that most patients with herpes are evaluated. In contrast with CMV esophagitis and idiopathic ulcer, these ulcers tend to be shallow; large, deep ulcers are rare (Fig. 33-4). Biopsies and cytologic brushings taken from the margin of the ulcers (the sites of active viral replication) are most likely to show epithelial cell invasion and nuclear changes typical of herpes infections. Viral cultures of biopsy specimens are usually positive.12 Isolated cases of esophagitis/ulcerations in AIDS owing to bacteria (actinomycosis, MAC, Rochalimaea henselae); fungi (Histoplasma, Mucormycosis, Torulopsis, Pneumo­ cystis); parasites (Leishmania, cryptosporidiosis); and pills (zidovudine [AZT], didanosine [ddI]) all have also been reported. Esophageal neoplasms also have been described in AIDS patients, including Hodgkin’s and non-Hodgkin’s lymphoma, histiocytic lymphoma, Kaposi’s sarcoma, squamous cell carcinoma, and adenocarcinoma.

Figure 33-4.  Herpes simplex virus esophagitis. Endoscopic view demonstrates diffuse erythema which surrounds multiple whitish plaques, which represent shallow ulceration. Islands of normal-appearing esophageal mucosa are still present.

Table 33-1  Differential Diagnosis of Dysphagia and Odynophagia in Patients with AIDS Candida albicans* Cytomegalovirus* Idiopathic ulcerations* Herpes simplex Histoplasma capsulatum Mycobacterium avium complex Cryptosporidium spp. Neoplasm: Kaposi’s sarcoma, lymphoma, squamous cell carcinoma, adenocarcinoma Gastroesophageal reflux disease* Pill-induced esophagitis *More frequent cause. AIDS, acquired immunodeficiency syndrome.

EVALUATION AND MANAGEMENT

A specific cause of esophageal complaints in the AIDS patient cannot be made on the basis of symptoms or physical examination alone (Table 33-1). Nevertheless, a few generalizations may be made. The presence of oral thrush associated with mild to moderate dysphagia without odynophagia is likely caused by Candida esophagitis. In contrast, the patient with severe odynophagia without dysphagia or thrush is more likely to have ulcerative esophagitis (viral, idiopathic). The patient complaining of substernal burning and regurgitation is most likely to have gastroesophageal reflux disease. Intermittent solid and liquid dysphagia may be related to a motility disturbance.13 Endoscopy with biopsy is the only means of establishing a specific etiology for the cause of dysphagia and odyno­ phagia. Conventional barium swallow radiography in the patient with esophageal complaints is not worthwhile, although it may reveal typical features of Candida, CMV, or herpes. In addition, the radiographic appearance of an esophageal ulcer cannot adequately distinguish etiology,

525

526

Section IV  Topics Involving Multiple Organs thus mandating subsequent endoscopy and biopsy for a definitive diagnosis. Multiple mucosal biopsies are preferred over brush cytology for ulcerated lesions.12 Given the preponderance of Candida infection, an empirical approach to the management of esophageal symptoms is reasonable in most patients with AIDS. Patients with dysphagia and/or odynophagia who also have oral thrush should be treated empirically with fluconazole 100 mg/day after a 200-mg loading dose.14 Itraconazole or fluconazole suspensions are effective alternatives. If symptoms persist despite a 1-week empirical trial, endoscopy with biopsy should be performed in preference to the initiation of other empirical trials or escalation of the dose of fluconazole. Narcotics are appropriate for the patient with severe pain until specific treatment for the underlying cause can be initiated. Relapse of Candida esophagitis is invariable unless immune function is improved with antiretroviral treatment. Despite chronic prophylaxis, relapse frequently occurs often due to antifungal resistance. This is much less of a problem now that patients are receiving HAART. In the absence of HAART, caspofungin and posaconazole may be effective for patients with resistance to antifungal drugs.15 CMV and HSV infection should be treated similarly to other gut involvement with these viruses (see following). Idiopathic ulcers respond in more than 90% of patients to oral glucocorticoids (e.g., 40  mg prednisone per day initially, tapered over 4 weeks).16 The basis for glucocorticoid efficacy is unknown; infectious causes should be assiduously excluded before administering gluocorticoids in this setting. Thalidomide is also highly effective and may be curative when prednisone fails.16 The main side effects of thalidomide are somnolence, rash, and neuropathy. The devastating teratogenic effects mandate its use to be limited to men.

DIARRHEA Before HAART, diarrhea occurred in up to 90% of patients during the course of HIV disease, especially those from developing countries. In the era of HAART, diarrhea is a less frequent complaint and etiologically is now most often drug-induced (antiretroviral therapy) or is caused by dis­ orders unrelated to HIV infection.17 Alterations in the mucosal immune system in AIDS predispose to intestinal infections, may lead to untreatable chronic infection by organisms that typically cause self-limited infection in healthy hosts (e.g., Cryptosporidium), and may contribute to a more virulent clinical course of otherwise common enteric infections (e.g., Salmonella, Shigella, Campylo­ bacter). Despite the vast spectrum of protozoal, viral, bacterial, and fungal organisms that cause diarrhea in the patient with AIDS, a differential diagnosis can be developed on the basis of the clinical presentation and degree of immunodeficiency (Table 33-2).

EVALUATION AND MANAGEMENT

Protozoa account for the most prevalent class of diarrheal pathogens in most series,18 largely because many of these infections can lead to chronic diarrhea and are refractory to treatment. Cryptosporidium, a cause of self-limited diarrhea in healthy hosts, remains the most frequent protozoa identified in HIV-infected patients worldwide.18 Clinical presentation and outcome are related to the degree of immunocompromise and the subtype of organism.19 Stool testing of asymptomatic individuals shows a high carriage rate

Table 33-2  Differential Diagnosis of Diarrhea in Patients with AIDS Protozoa Microsporidium* Cryptosporidium spp.* Isospora belli Toxoplasma spp. Giardia lamblia Entamoeba histolytica Leishmania donovani Blastocystis hominis Cyclospora spp. Pneumocystis jiroveci Bacteria Clostridium difficile Salmonella spp.* Shigella spp.* Campylobacter jejuni* Mycobacterium avium complex Mycobacterium tuberculosis Small bowel bacterial overgrowth Vibrio spp. Viruses Cytomegalovirus* Herpes simplex Adenovirus spp. Rotavirus spp. Norovirus HIV? Fungi Histoplasmosis Coccidioidomycosis Cryptococcosis Candidiasis Penicillium marneffei Neoplasms Lymphoma Kaposi’s sarcoma Idiopathic “AIDS enteropathy” Drug Induced HIV protease inhibitors Pancreatic Disease Pancreatic insufficiency Chronic pancreatitis Infectious pancreatitis (CMV, MAC) Drug-induced pancreatitis (e.g., pentamidine) *More frequent cause. AIDS, acquired immunodeficiency syndrome; CMV, cytomegalovirus; HIV, human immunodeficiency virus.

because some subtypes appear to be less pathogenic.20 The small bowel is the most common site of infection, although the organisms can be recovered in all regions of the gut, as well as in biliary and respiratory epithelium. Diarrhea is typically severe, with stool volumes of several liters per day not uncommon. Borborygmi, nausea, and weight loss are frequently associated symptoms; right upper quadrant pain suggests biliary tract involvement (see later). The pathogenesis of this infection is uncertain. The diagnosis of intestinal cryptosporidiosis is most often made by acid-fast stain of the stool, where the organisms appear as bright red spherules, similar in size to red blood cells. The sensitivity of stool testing varies and depends on the burden of organisms, character of the stool (formed versus liquid), and primary site of infection. Stool antigen detection and polymerase chain reaction (PCR) markedly increase sensitivity of stool testing. Cryptosporidia may be identified in small bowel or rectal biopsies even when the stool examination is negative.21

Chapter 33  Gastrointestinal Consequences of Infection with Human Immunodeficiency Virus Table 33-3  Treatment of Infectious Causes of Diarrhea in Patients with AIDS pathogen Protozoa Cryptosporidia Cyclospora spp. Isospora belli Microsporidia Viruses Cytomegalovirus Herpes simplex Bacteria Salmonella, Shigella, Campylobacter spp. Clostridium difficile Small intestinal bacterial overgrowth Mycobacterium tuberculosis Mycobacterium avium complex Fungi Histoplasmosis Coccidioidomycosis Cryptococcosis

TREATMENT

DURATION (DAYS)

Paromomycin, azithromycin, nitazoxanide Trimethoprim-sulfamethoxazole or ciprofloxacin Trimethoprim-sulfamethoxazole or ciprofloxacin or pyrimethamine Albendazole (Encephalitozoon intestinalis) Metronidazole, atovaquone, fumagillin (not available in United States)†

14-28 14-28 14-28 14-28

Ganciclovir Foscarnet Cidofovir Acyclovir

14-28* 14-28* 14-28* 5-10*

Fluoroquinolone (e.g., ciprofloxacin) Vancomycin, metronidazole Metronidazole, ciprofloxacin Isoniazid, rifampin, pyrazinamide, ethambutol Multidrug regimens for symptomatic infection (see text)

10-14* 10-14 10-14 9-12 mo 9-12 mo

Amphotericin B; then itraconazole Amphotericin B; then fluconazole Amphotericin B; then fluconazole

28 28 28

*Duration of therapy dictated by immune reconstitution with highly active antiretroviral therapy. † Molina JM, Tourneur M, Sarfati C, et al. Fumagillin treatment of intestinal microsporidiosis. N Engl J Med 2002; 346:1963. AIDS, acquired immunodeficiency syndrome.

Specific antimicrobial treatment of cryptosporidial infection remains disappointing. Numerous antimicrobial agents have been tested, most without significant effect (Table 33-3). Nitazoxanide and azithromycin have been most recently evaluated with mixed results, and combination therapies have also been used. Nitazoxanide can lead to cryptosporidial oocyte clearance in HIV-seronegative patients but has not shown efficacy in HIV-infected patients.22 Although not pathogen-specific, currently the most effective therapy for cryptosporidia is HAART, in which improvement of immune function results in a clinical remission of diarrhea and clearance of cryptosporidia from the stool and on small bowel biopsy.23 For patients failing HAART and/or in whom antimicrobial therapy is ineffective, symptomatic treatment should include fluid support, antidiarrheal agents, and occasionally narcotics such as tincture of opium to control the diarrhea. Isospora belli is a sporozoan, which, like Cryptospori­ dium, is a cause of chronic diarrhea in untreated patients with HIV infection. The disease is rare in the United States, but it is more frequent and endemic in developing countries such as Haiti. The organism may be identified by acid-fast stain of the stool or duodenal secretions or on mucosal biopsy. This infection can be effectively treated with antibiotics, specifically trimethoprim-sulfamethoxazole and ciprofloxacin.24 Microsporidia emerged as common intestinal infections in AIDS, but their prevalence has markedly fallen in the HAART era.25 Intestinal and hepatobiliary disease may be caused by two species of microsporidia: Enterocytozoon bieneusi and Encephalitozoon intestinalis. The reported prevalence of microsporidia without HAART varied from 15% to 39%.18,21,26 Typical symptoms include watery, nonbloody diarrhea of mild to moderate severity usually without associated crampy abdominal pain. Weight loss is common, although not to the degree observed with Cryptosporidium. Infection is associated with severe immunodeficiency with median CD4 counts of infected individuals of less than

Figure 33-5.  Endoscopic biopsy specimen of small bowel microsporidiosis. This thin plastic section demonstrates shedding of an epithelial cell containing microsporidial oocysts. (From Gazzard BG. Diarrhea in human immunodeficiency virus antibody-positive patients. Semin Gastroenterol 1991;2:3.)

100/mm3.26 As with infection from cryptosporidia, the pathogenesis of disease remains poorly defined. The organism incites little tissue inflammation and is rarely associated with villous atrophy and cell degeneration. Microsporidia can be discerned by light microscopy when tissue is embedded in plastic or paraffin (Fig. 33-5). Staining of embedded mucosal biopsies with Brown-Brenn, Gram stain, or modified Masson trichrome stain is superior to routine hematoxylin and eosin staining.27 E. intestinalis can usually be differentiated from E. bieneusi by its larger size and infection of lamina propria macrophages; electron microscopy is definitive. Stool staining techniques are only moderately sensitive, while small bowel biopsies are generally positive. No effective therapy is available for E. bieneusi,

527

528

Section IV  Topics Involving Multiple Organs whereas albendazole is effective for E. intestinalis.28 As with the treatment of cryptosporidia, HAART is the best therapy, resulting in resolution of diarrhea with loss of this pathogen from stool and on small bowel biopsy.23 For unclear reasons, infections by the protozoa Giardia lamblia and Entamoeba histolytica are not consistently seen with increased frequency or virulence in AIDS.29 However, in a study from Taiwan, where E. histolytica is endemic, amebic colitis was identified as a common cause of diarrhea.30 The nonpathogenic Entamoeba dispar is mor­ phologically similar to E. histolytica and can only be distinguished by more specific stool or enzyme-linked immunosorbent assay tests.29 Blastocystis hominis, Endoli­ max nana, and Entamoeba coli are nonpathogenic protozoa that are seen more commonly in men who have sex with men and are often found in association with other protozoal parasites. Rare cases of enteric leishmaniasis, Pneumocystis jiroveci infection, and toxoplasmosis have been reported. Helminths, particularly Strongyloides stercoralis and Ascaris lumbricoides, are uncommon pathogens.31 Patients may present with abdominal pain, diarrhea, and eosinophilia. The clinical syndrome and recurrence rate associated with these parasites do not appear to be altered in the setting of HIV infection. Viral infection of the large bowel, and rarely the small bowel, is an important cause of diarrhea in HIV infection. CMV is the most common viral cause of diarrhea and the most frequent cause of chronic diarrhea in patients with AIDS and multiple negative stool tests.32 This infection characteristically occurs late in the course of HIV infection when the CD4 lymphocyte count falls below 100/L (see Fig. 33-1). Infection is most common in the colon (Fig. 33-6), but concomitant disease in the esophagus, stomach, or small bowel may be observed. Isolated small bowel disease typically results in abdominal pain rather than diarrheal illness. The pathogenesis has not been totally elucidated. Infection of vascular endothelial cells is common, suggesting a role

Figure 33-6.  Cytomegalovirus colitis. Endoscopic photograph of the sigmoid colon showing edema and diffuse subepithelial hemorrhage typical for cytomegalovirus. This endoscopic appearance is similar to that of idiopathic ulcerative colitis.

for mucosal ischemia; true histopathologic evidence of vasculitis is rare. An important role for local proinflammatory cytokine activation has been suggested.33 The clinical manifestations of enteric CMV infection vary greatly and include asymptomatic carriage, nonspecific symptoms of weight loss and fevers, and focal enteritis/ colitis including appendicitis or diffuse ulcerating hemorrhagic involvement with bleeding or perforation. As a result, patients can present with one of several constellations of symptoms, including abdominal pain; peritonitis; watery, nonbloody diarrhea; or hematochezia.34 The most common presentation, however, is abdominal pain asso­ ciated with chronic diarrhea. Although the endoscopic spectrum is variable, the hallmark of CMV enteritis/colitis is subepithelial hemorrhage and mucosal ulceration (see Fig. 33-6).34 The diagnosis of GI CMV infection is best established by demonstrating viral cytopathic effect in tissue specimens.35 The inclusions may be atypical in appearance or few in number, requiring immunostaining and/or in situ hybridization for confirmation.35 Cultures for CMV are usually positive when inclusions are present, but they are less sensitive and specific for CMV infection than histopathologic identification. If inclusions are few in number and are demonstrable in tissue that appears macroscopically normal, the patient should be considered to have CMV colonization rather than true CMV infection. A number of effective therapies are available for the treatment of CMV (see Table 33-3). The most commonly used agent is ganciclovir, an acyclovir derivative, which is effective in approximately 75% of cases.36 Ganciclovir requires daily intravenous administration for several weeks, depending on the location and severity of disease. Valganciclovir, an oral analog of ganciclovir, has excellent GI absorption and efficacy for CMV retinitis, but has not been well studied for induction therapy in GI disease. Valganciclovir has become widely used as preemptive therapy in the transplant setting and has shown promise as first-line therapy.37 An alternative of equivalent efficacy to ganciclovir is foscarnet, a pyrophosphate analog that inhibits viral replication. In contrast to ganciclovir, it has the advantage of being less marrow suppressive, although renal insufficiency and disturbances in mineral metabolism (hypocalcemia, hypomagnesemia, hypophosphatemia) are frequent. Foscarnet tends to be less well tolerated than ganciclovir. Although rarely used, cidofovir is the newest agent for CMV. Like ganci­ clovir and foscarnet, it must be given intravenously, and similar efficacy rates have been reported for retinal disease. Anecdotal experience suggests it to be effective for GI disease caused by CMV. Because of its long half-life (2 weeks), cidofovir can be given once weekly, which may be particularly advantageous for some patients. The main side effect is nephrotoxicity that may be irreversible. Because of the severe immunodeficiency required for the development of CMV disease, recurrences are common following withdrawal of therapy; however, immune reconstitution with HAART will negate the need for long-term suppressive therapy.3 At the time of diagnosis of GI CMV infection, all patients should have an ophthalmologic examination to exclude CMV retinitis because this site of infection requires close follow-up to ensure remission, thereby preventing blindness. Although widely used in the transplant setting, the role of CMV antigenemia or DNA concentrations by PCR to predict subsequent disease and guide the use of preemptive therapy remains less well defined.38 A number of other viruses (e.g., Norwalk, adenovirus), as well as novel enteric viruses (astrovirus, picobirnavirus)

Chapter 33  Gastrointestinal Consequences of Infection with Human Immunodeficiency Virus

A

B

have been identified in symptomatic and asymptomatic patients, but their overall contribution to diarrheal disease in AIDS is small.39 The role of HIV itself as a diarrheal pathogen is limited. Although HIV can be identified within gut tissue in some patients with AIDS, the virus has been confined to lamina propria macrophages and enterochromaffin cells, and not epithelial cells. An idiopathic AIDS enteropathy has been proposed to account for the diarrhea in AIDS patients who lack an identifiable pathogen and may reflect indirect effects of HIV on enteric homeostasis. With improvements in diagnostic techniques, greater awareness of the spectrum of diarrheal pathogens in AIDS, recognition of the importance of drugs as additional causes and use of panendoscopy with biopsy for patients with negative stool tests, a dimin­ ishing fraction of patients have truly “idiopathic diarrhea.” Although a variety of morphologic and functional abnormalities of the small bowel have been shown in HIV-infected patients, their role in causing or contributing to GI symptoms is likely small.40 Institution of protease inhibitors has been shown to improve chronic unexplained diarrhea.41 Infections by enteric bacteria are more frequent and more virulent in HIV-infected individuals compared with healthy hosts. Salmonella, Shigella, and Campylobacter have higher rates of bacteremia and antibiotic resistance. Diagnosis is straightforward because the organisms usually can be grown from stool samples (see Chapter 107). These enteric infections typically present with high fever, abdominal pain, and diarrhea that may be bloody. Abdominal pain can be severe, mimicking an acute abdomen. As noted, bacteremia is common, and parenteral antibiotics should be administered empirically in severely ill patients when these infections are suspected pending results of stool and blood cultures; a fluoroquinolone such as ciprofloxacin may be a particularly attractive choice for empirical therapy and if organisms are multiply resistant. Diarrhea due to Clostridium difficile has emerged as the most common bacterial pathogen, not because it is an OI, but rather because antibiotic use is far greater and hospital-

Figure 33-7.  Intestinal Mycobacterium avium complex. A, Hematoxylin and eosin staining of a small bowel biopsy specimen shows marked thickening of the villi with a cellular infiltrate. B, High-power view with acid-fast staining shows numerous macrophages filled with mycobacteria.

ization more frequent in this population than in healthy hosts.42 The clinical presentation, response to therapy, and relapse rate are no different than in immunocompetent patients.43 Diagnosis rests on standard assays of stool for C. difficile enterotoxin. Treatment with metronidazole or vancomycin is generally effective (see Chapter 108). Small bowel bacterial overgrowth (see Chapter 102) is uncommon in AIDS patients,44 and its role in causing diarrhea appears limited. Mycobacterial involvement of the bowel either by Myco­ bacterium tuberculosis or MAC may lead to diarrhea, abdominal pain, and, rarely, obstruction or bleeding in patients with late-stage AIDS. Although M. tuberculosis infection appears to be symptomatic in all cases, a large number of patients with MAC have an asymptomatic GI infection. Duodenal involvement is most common and may be suspected at endoscopy by the presence of yellow mucosal nodules, often in association with malabsorption, bacteremia, and systemic infection. Diagnosis of GI MAC infection is best made by endoscopic biopsy; fecal acid-fast smear is much less sensitive than culture. The organism is readily seen on biopsy specimens with acid-fast staining, and the number of organisms is often striking (Fig. 33-7). Blood culture positivity may suggest the diagnosis. Affected patients have severe malabsorption and weight loss in association with blunting of villi and suffusion of macrophages with mycobacteria. As is typical of MAC infection, in AIDS there is a poorly formed inflammatory response and granulomas are rarely present. Response to antibiotic therapy is variable and depends in part on the extent of immunocompromise; however, eradication is rarely achieved. Multidrug regimens are required for therapy including combinations of amikacin, ethambutol, rifampin, clarithromycin, and ciprofloxacin, which can reduce, but not eradicate, MAC organisms. As with other OIs, institution of HAART in these patients may improve immune function, hasten clinical resolution of the infection, prevent relapse such that long-term antimicrobial therapy will be unnecessary, and enhance survival.3,45

529

530

Section IV  Topics Involving Multiple Organs Although extrapulmonary M. tuberculosis is characteristic of AIDS, luminal GI tract involvement remains infrequent but when present usually involves the ileocecal region or colon.46 Fistula formation, intussusception, and perforation, as well as peritoneal and rectal involvement, also have been reported. Tuberculous involvement of the gut in HIV is most commonly found in developing countries. In contrast with MAC, M. tuberculosis infections in AIDS generally respond to multidrug antituberculous therapy with clinical and microbiologic cure, although treatment may be difficult due to drug interactions.47 Infections caused by mycobacteria (e.g., MAC lymphadenitis) and viruses (e.g., CMV uveitis) have been described following institution of HAART.48 This immune reconstitution syndrome results in an exuberant inflammatory response directed toward previously quiescent or incubating pathogens, resulting in paradoxical exacerbations of these infections. Fungal infections of the gut have been recognized in AIDS. GI histoplasmosis has been most commonly described and occurs in the setting of disseminated infection, often in association with pulmonary and hepatic histoplasmosis. It may manifest as a diffuse colitis with large ulcerations and diarrhea, as a mass, or as serosal disease in association with peritonitis.49 The diagnosis of disseminated histoplasmosis may be suspected in a patient with high fever and markedly elevated serum LDH.50 The diagnosis is established by fungal smear and culture of urine, infected tissue or blood; histoplasmosis antigen assay may provide supportive evidence. The infection is often managed initially by liposomal amphotericin B administration. Long-term suppressive therapy with itraconazole has been used successfully, whereas a response to HAART mitigates the need for longterm suppressive therapy. Rare cases of systemic cryptococ­ cosis and coccidioidomycosis with gut involvement also have been described. A peculiar fungal infection due to Penicillium marneffei has been reported from Southeast Asia that can cause colitis and chronic diarrhea.51 With the advent of HAART, drug-induced diarrhea has become an increasingly important and a frequent cause of antiretroviral drug discontinuation.52 The most common agents associated with diarrhea are the protease inhibitors, with nelfinavir having the highest rate.53 Generally the diarrhea is mild to moderate in severity and is not associated with weight loss. The mechanism(s) for diarrhea due to these agents is poorly understood. Symptomatic therapies are generally effective. A suggested approach to the evaluation of diarrhea is outlined in Table 33-4.

ABDOMINAL PAIN The exact frequency of abdominal pain in patients with AIDS is unknown, but like other GI complications of AIDS, the prevalence and etiology have been altered by HAART. In most patients with AIDS, abdominal pain, when severe, is directly related to HIV and its consequences. However, the physician must consider not only the manifestations of OIs and neoplasms but also the more common causes of abdominal pain in the general population.54 The differential diagnosis of abdominal pain in AIDS, presented in Table 33-5, is organized by the site of origin of the pain. For each organ system, a list of potential complications with their likely causes is offered. In some instances causes are listed because of their known ability to produce symptoms by involving a particular organ. Table 33-5 does not include non–AIDS-specific diagnoses that have assumed more importance in the era of HAART. Table 33-6 defines

Table 33-4  Evaluation of Diarrhea in Patients with AIDS In all patients Stool specimen for bacterial culture: For Salmonella, Shigella, and Campylobacter spp.; Clostridium difficile toxin Stool smear for fecal leukocytes, ova and parasite examination (at least 3-6 specimens), and acid-fast stain If patient has rectal bleeding, tenesmus, or fecal leukocytes Flexible sigmoidoscopy or colonoscopy with biopsy of mucosa for histopathology, viruses, protozoa Cultures of rectal tissue for bacteria (especially for Campylobacter spp.); viruses (optional) If diarrhea and weight loss persist and above evaluation is negative Upper endoscopy with small bowel mucosal biopsy AIDS, acquired immunodeficiency syndrome.

Table 33-5  Differential Diagnosis of Abdominal Pain in Patients with AIDS* ORGAN Stomach Gastritis Focal ulcer Outlet obstruction Mass Small Bowel Enteritis Obstruction Perforation Colon Colitis Obstruction Perforation Appendicitis Liver, Spleen Infiltration Biliary tract Cholecystitis Papillary stenosis Cholangitis Pancreas Pancreatitis Tumor Mesentery, Peritoneum Infiltration

CAUSES CMV,† Cryptosporidia CMV,† PUD Cryptosporidia, CMV, lymphoma, PUD Lymphoma, KS, CMV Cryptosporidia,† CMV, MAC Lymphoma,† KS CMV,† lymphoma CMV, enteric bacteria,† HSV Lymphoma,† KS, intussusception CMV,† lymphoma, HSV KS,† Cryptosporidia, CMV Lymphoma,† CMV, MAC CMV,† Cryptosporidia,† Microsporidia CMV,† Cryptosporidia,† KS CMV† CMV,† KS, pentamidine, ddI Lymphoma, KS MAC,† Cryptococcus spp., KS, lymphoma, histoplasmosis, tuberculosis, coccidioidomycosis, toxoplasmosis

*The differential diagnosis does not include non–AIDS-specific conditions. † More frequent diagnosis. AIDS, acquired immunodeficiency syndrome; CMV, cytomegalovirus; ddI, didanosine; HSV, herpes simplex virus; KS, Kaposi’s sarcoma; MAC, Mycobacterium avium complex; PUD, peptic ulcer disease.

abdominal pain in terms of the four most common pain syndromes, their most likely causes, and the diagnostic methods indicated. Generally the duration and severity of symptoms dictate the urgency of evaluation.

EVALUATION AND MANAGEMENT

(see Table 33-6) As with any patient, the history is helpful in localizing the origin of abdominal pain. Associated symptoms and signs should suggest the particular organ involved, and the quality and duration of the abdominal pain may implicate specific diseases. Generally, the same workup as for a

Chapter 33  Gastrointestinal Consequences of Infection with Human Immunodeficiency Virus Table 33-6  Evaluation of Abdominal Pain Syndromes in Patients with AIDS SUSPECTED DIAGNOSIS

DIAGNOSTIC APPROACH

Dull pain, diarrhea, mild nausea, vomiting Acute, severe pain, with peritoneal irritation

Infectious enteritis

Stool culture, O&P; sigmoidoscopy

Perforation, infectious peritonitis

Right upper quadrant pain, abnormal liver biochemical tests

Cholecystitis, cholangitis, hepatic infiltrates, cholangiopathy Intestinal obstruction

Abdominal plain films, surgical consultation, US or CT, paracentesis if ascites is present, laparoscopy US or CT, MRC, ERCP, liver biopsy

SYNDROME

Subacute pain, severe nausea and vomiting

Abdominal plain films, CT, small bowel series, endoscopy, barium enema

AIDS, acquired immunodeficiency syndrome; CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; MRC, magnetic resonance cholangiography; O&P, ova and parasites; US, ultrasonography.

patient without AIDS should be initiated. Abdominal ultrasonography and CT scanning are useful early in the assessment of abdominal pain and may highlight regions of disease not suspected clinically. In the patient with acute pancreatitis, drug-induced causes must be considered.55 Management of abdominal pain falls broadly into surgical versus nonsur­gical options. Indications for surgical intervention in AIDS patients are the same as for patients without AIDS. All tissue specimens must be submitted for viral and fungal culture and for pathologic examination, and enlarged mesenteric nodes should undergo biopsy. Laparoscopic surgery will provide a less invasive alternative to laparotomy in selected patients.56 The nonsurgical management of abdominal pain is determined by the clinical evaluation.

ANORECTAL DISEASE The frequency of anorectal disease among homosexual AIDS patients is higher than in other AIDS patients. Common findings in HIV-infected patients include peri­ rectal abscesses, anal fistulas, perianal HSV, idiopathic ulcerations, and infectious proctitis, but lymphoma, ulcerations due to CMV, tuberculosis, and histoplasmosis may also be seen (Table 33-7). The frequency of anorectal squamous cell carcinomas is strikingly higher in homosexual men than in other members of the population, and the risk increases as HIV disease advances. These neoplasms result from human papillomavirus (HPV) infections acquired through sexual contact, particularly HPV types 16 and 18. Morphologic studies have documented histologic progression, often in the same lesion, from a benign lesion, condyloma acuminatum, to marked anal dysplasia or squamous cell carcinoma. Cytologic specimens of the anal canal, similar to Papanicolaou smears, are increasingly used for screening and have high predictive value for dysplasia.57

Table 33-7  Differential Diagnosis of Anorectal Disease in Patients with AIDS Infections Bacteria Chlamydia trachomatis* Lymphogranuloma venereum Neisseria gonorrhoeae* Shigella flexneri Mycobacterium tuberculosis Protozoa Entamoeba histolytica Leishmania donovani Viruses Herpes simplex* Cytomegalovirus* Fungi Candida albicans Histoplasma capsulatum Neoplasms Lymphoma* Kaposi’s sarcoma Squamous cell carcinoma Cloacogenic carcinoma Condyloma acuminatum Other Idiopathic ulcers* Perirectal abscess, fistula* *More frequent diagnosis. AIDS, acquired immunodeficiency syndrome.

EVALUATION AND MANAGEMENT

In HIV-infected patients and patients with AIDS, physical examination should include careful inspection of the skin and mucous membranes, as well as palpation of the lymph nodes. Visual inspection of the anus for ulcers, fissures, and masses should precede digital examination. Palpation of the perianal area and buttocks for abscess should be performed. The presence of severe pain on rectal examination strongly suggests ulcerative disease, hemorrhoids, or neoplasms. Palpation of the anal canal may reveal masses or fissures not otherwise evident. All patients with anorectal symptoms should have anoscopy and sigmoidoscopy (rigid or flexible) with mucosal biopsy. Evaluation under general anesthesia may be necessary when pain is severe. Specimens should be evaluated for evidence of neoplasm or infection; when appropriate, they should be examined with bacterial (including gonococcal and chlamydial), viral, and fungal cultures. CT scan may define the extent of disease if a neoplasm is identified. Healing of anorectal disease following surgical or medical therapy will largely be determined by the stage of HIV infection. HIV-positive patients without AIDS have favorable outcomes following anorectal surgery, with acceptable wound healing, whereas patients with AIDS are more likely to have a poor outcome. The survival of patients with squamous cell cancer has improved in the HAART era.58

GASTROINTESTINAL BLEEDING GI bleeding in AIDS is as likely to arise from sources not unique to AIDS as from OIs or neoplasms. Infections and neoplasms seen exclusively with AIDS can rarely cause GI bleeding (Table 33-8). Studies have found that the causes of upper GI bleeding are most frequently due to disorders not linked to AIDS including peptic ulcer, whereas in contrast, the most common cause of lower GI bleeding is

531

532

Section IV  Topics Involving Multiple Organs Table 33-8  Differential Diagnosis of Gastrointestinal Bleeding in Patients with AIDS (Excluding Non–AIDS-Specific Diagnoses)

Table 33-9  Differential Diagnosis of Hepatomegaly and Abnormal Biochemical Liver Tests in Patients with AIDS

Esophagus Candida spp.* Cytomegalovirus* Herpes simplex Idiopathic ulcer Stomach Cryptosporidiosis Cytomegalovirus* Kaposi’s sarcoma* Lymphoma Small Intestine Kaposi’s sarcoma* Lymphoma* Cytomegalovirus Salmonella spp. Cryptosporidia Colon Cytomegalovirus* Entamoeba histolytica Campylobacter jejuni Clostridium difficile Shigella spp. Idiopathic ulcerations Kaposi’s sarcoma* Lymphoma

Hepatic Parenchymal Disease Infection Mycobacterium avium complex† Cytomegalovirus Hepatitis C† Bacillary peliosis hepatis Mycobacterium tuberculosis† Cryptococcus spp. Hepatitis B, D Pneumocystis jiroveci Microsporidia Drug* Neoplasm Lymphoma Kaposi’s sarcoma Biliary Disease Cholangitis Cytomegalovirus† Cryptosporidia† Microsporidia Neoplasm Lymphoma† Kaposi’s sarcoma

*More frequent diagnosis. AIDS, acquired immunodeficiency syndrome.

CMV colitis.59 As in any other patient, ulceration, from any cause, is the most common pathologic lesion; thus, disorders causing ulcer (e.g., CMV, HSV) are the most common etiologies. Enteric pathogens including Campylobacter, Shigella, Salmonella, E. histolytica, and Chlamydia may cause rectal bleeding from ulceration or colitis. Other pathogens such as MAC, microsporidia, and cryptosporidia almost never cause bleeding because mucosal infection does not typically result in ulceration. Enteric lymphoma (e.g., Burkitt’s) or Kaposi’s sarcoma lesions may ulcerate and bleed spontan­eously, although most enteric Kaposi’s sarcoma lesions are asymptomatic.

EVALUATION AND MANAGEMENT

The evaluation of GI bleeding in a patient with AIDS parallels the approach taken in otherwise healthy patients (see Chapter 19). Endoscopy is preferred in all patients, especially those with severe immunodeficiency, given the likelihood of opportunistic diseases that generally require mucosal biopsy for diagnosis and because endoscopic therapy for hemostasis can be performed. Appropriate initial management of severe GI bleeding due to AIDS-related diseases does not require a specific diagnosis, and treatment parallels any other patient (see Chapter 19). Specific therapies for the underlying disease necessarily depend on the results of mucosal biopsy and/or microbiologic studies.

HEPATOMEGALY AND ABNORMAL BIOCHEMICAL LIVER TESTS Hepatomegaly, a frequent finding in AIDS, is usually asso­ ciated with one or more liver chemistry test abnormalities. As with other organ systems, the spectrum and clinical

*Especially sulfonamides, protease inhibitors. † More frequent diagnosis. AIDS, acquired immunodeficiency syndrome.

manifestations of hepatobiliary disease in patients with HIV evolves as immunocompromise advances, and HAART has altered the frequency, manifestation, and outcome of a number of these diseases (Table 33-9). Hepatobiliary disease can be broadly classified into either hepatic parenchymal abnormalities, biliary abnormalities, or a combination of the two. Currently, parenchymal abnormalities are most often related to viral hepatitis and drug-induced disease. In the era of HAART, liver disease has assumed much greater importance as a cause of morbidity and mortality6 and now represents one of the most frequent non–HIV-related causes of death.60 Drug-induced liver injury has emerged as the most prevalent cause of liver test abnormalities and is related to the increasing array of antiretroviral medications. Use of other prescription (or nonprescription) drugs, as well as herbal remedies, should also be considered a cause of abnormal liver test results in the HIV-infected patient.61 Before HAART, drug hepatotoxicity was most commonly due to sulfonamides, and the increased frequency of adverse reactions to these medications is well recognized in AIDS.62 The protease inhibitors are the most common causes of abnormal liver tests, with ritonavir cited most often.63 The mechanisms of liver injury include drug allergy or idiosyncratic reactions to drugs and exacerbation of underlying viral hepatitis. The major risk factors for drug-induced hepatotoxicity include coexistent viral hepatitis, older age, and greater rise in CD4 cells after HAART.64-66 The liver test abnormalities usually follow a hepatocellular pattern; jaundice (primarily indirect hyperbilirubinemia) is uncommon but has been observed most frequently with indinavir. The lactic acidosis syndrome, characterized by marked hepatomegaly, steatosis, metabolic lactic acidosis, and liver failure, is now well recognized. The pathogenesis of the syndrome is due to impaired mitochondrial DNA synthesis from the nucleoside reverse transcriptase inhibitors such as zidovudine, dideoxyinosine (ddI), and stavudine.67 An associated myopathy, peripheral neuropathy, and pancre-

Chapter 33  Gastrointestinal Consequences of Infection with Human Immunodeficiency Virus atitis may occur as well. Risk factors for the syndrome are unknown. The liver tests typically show a hepatocellular pattern but can be normal or minimally increased. Hepatic steatosis is evident on imaging of the liver. Although reversal has occurred in some patients following drug withdrawal, most patients have worsening disease and death. Liver transplantation is curative. MAC is consistently the most frequent specific hepatic finding in AIDS in late-stage HIV disease.68 The pathologic hallmark of the infection is the presence of poorly formed granulomas containing acid-fast bacilli within foamy histiocytes. Organisms may be observed in the absence of granulomas and can be cultured from liver biopsy in the absence of infected histiocytes. In developing countries, M. tubercu­ losis is the most common OI involving the liver. M. tuber­ culosis, in contrast to MAC, may occur before HIV-infected patients are profoundly immunocompromised. Tuberculosis is commonly extrapulmonary (≈80%) in patients with HIV infection.69 Hepatic disease as part of miliary tuberculosis has been noted. Rarer manifestations include tuberculous abscesses and bile duct tuberculomas.46 The diagnosis of hepatic tuberculosis is made by culture of the organism from liver tissue obtained by percutaneous or laparoscopic biopsy. PCR may allow earlier diagnosis. As with MAC, typical-appearing mycobacteria can be observed by appropriate staining of biopsy specimens. CMV is an uncommon liver pathogen most often found at autopsy. However, it rarely is a cause of clinical hepatitis or cause of other hepatic symptoms. Typical viral inclusions are usually identified in Kupffer cells but can sometimes be seen in hepatocytes or sinusoidal endothelial cells or in association with granulomas. Clinical manifestations and histologic features of viral hepatitis from hepatitis B virus (HBV), hepatitis C virus (HCV), or hepatitis D virus and hepatitis A virus (HAV) are altered in the presence of HIV coinfection but in remarkably different ways for each virus. Clinical and autopsy studies in AIDS patients have reported up to a 90% seroprevalence of hepatitis B markers indicating past or present infection.70,71 More recent studies suggest lower rates, perhaps partly due to use of HBV vaccines.72 Concurrent HIV and HBV infections lead to alterations of HBV antigen-antibody display, viral replication, and clinical consequences. Several reports have described reappearance of hepatitis B surface antigen (HBsAg) in HIVinfected patients previously thought to be immune to HBV, as indicated by the presence of anti-HBs. Recurrence of HBsAg may arise from either reinfection or reactivation with advanced immunodeficiency. In addition, there is an accelerated loss of naturally acquired anti-HBs even in those patients who remain HBsAg negative. With loss or reduction in immunity to HBV, there is an increased pre­ valence of hepatitis B e antigen (HBeAg) expression, elevated mean levels of deoxyribonucleic acid (DNA) polymerase, and increased titers of antihepatitis B core antigen.73 Acquisition of the chronic carrier state is also much more likely in the HIV-infected patient, especially if infection occurs when immunodeficiency is more advanced. Thus a larger proportion of patients with HIV and HBV infections have a chronic carrier state, with highly infectious serum and body fluids, compared with those who are HIV negative. Although HIV infection leads to more prevalent chronic HBV carriage, it appears to attenuate the severity of biochemical and histologic liver disease. The mechanism for reduced hepatitis B virus-related liver injury following HIV infection is not certain but has been attributed to a diminution in lymphocyte-mediated hepatocellular injury as a

result of HIV effects on lymphocytes. In those patients without serologic evidence of past or present HBV and HIV infection, the efficacy of vaccination is related to the stage of immunocompromise.71 HBV has no independent effect on survival for patients with HIV.74 Conversely, the institution of HAART in a chronic carrier of HBV can have catastrophic consequences following immune reconstitution. Patients may develop an acute flare of hepatitis that can be severe, leading to fulminant hepatic failure. However, the proportion of coinfected patients who develop an acute hepatitis B flare following use of HAART appears to be low.75 It is believed that reconstitution of immune function with HAART leads to production of antibody that is directed to infected hepatocytes as in the normal host. Seroconversion to anti-HBe and/or anti– hepatitis B surface antigen (HBs) may also be observed. Inclusion of lamivudine, which has potent antiviral effects on HBV, in the HAART regimen may reduce the likelihood of an acute flare of hepatitis B. Also the development of escape mutants during long-term lamivudine therapy may precipitate acute hepatitis. Current recommendations are to include two drugs that are active against HBV to prevent emergence of resistant variants. These observations suggest that all patients who are to receive HAART therapy should be screened for active or past HBV infection. Vaccination should be considered in all eligible patients but is less effective, especially in those most immunosuppressed. Treatment options for HBV infection in the setting of HIV have been summarized.76 The consequences of HIV infection on delta hepatitis (hepatitis D) appear similar to those of HBV, although far fewer patients have been studied.77 Case reports and small case series have reported higher serum titers of HAV ribonucleic acid (RNA), as well as more prolonged viremia and higher serum aminotransferase levels among HIV-infected patients. Despite these observations there are no data to suggest that acute HAV infection in HIV-infected patients leads to more severe hepatic disease or worse outcomes compared with non–HIV-infected patients. Hepatitis A vaccine is safe in HIV-positive patients, although less immunogenic.71 The prevalence of HCV infection in those with HIV infection depends in large part on the risk group evaluated. Prevalence is highest in injection drug users (52% to 89%) and hemophiliac patients with HIV,78 whereas in homosexual men and non-drug users, the prevalence is much lower, ranging from 1% to 11%.79 Unlike HBV, the clinical course of HCV worsens as HIV-related immunocompromise advances. Studies in large cohorts of hemophiliac patients have demonstrated dramatic increases in HCV RNA levels with progressive HIV disease, associated with aspartate aminotransferase (AST) elevations and hepatomegaly.80-82 Coinfected patients also have a higher rate of active cir­ rhosis on biopsy and an accelerated course to clinical cirrhosis and liver failure. Factors that predict fibrosis and progression to cirrhosis in coinfected patients include older age at infection, higher serum alanine aminotransferase (ALT) levels, higher inflammatory activity, alcohol consumption of more than 50  g/day and CD4 count less than 500 cells/mm3.83,84 Steatohepatitis also may play a role.85 The mechanism for this more rapid disease course is unknown but has been similarly recognized in other immunocompromised patients. However, as HIV-infected patients are now living longer as a result of HAART, HCV-induced liver disease and its consequences (e.g., hepatocellular cancer) are assuming significant clinical relevance. Recent studies show HCV-related cirrhosis and its complications to be a common cause of hospitalization and cause of death.6,60

533

534

Section IV  Topics Involving Multiple Organs Like hepatitis B, HCV does not cause progression of HIV disease. The effect of HAART on hepatitis C viral dynamics and liver injury is emerging. Some studies have found attenuation of disease,86 whereas others documented exacerbations reflected by increases in serum aminotransferases.87 Hepatitis C viral load has also been variably affected. The effect of HAART on the natural history of HCV-coinfected patients has been contradictory.88,89 Interferon therapy for HIV/HCV coinfected patients is less effective than in otherwise healthy individuals, particularly if the CD4 count is very low (<200/mm3). Combination therapy of peginterferon plus ribavirin results in sustained virologic response rates of 14% to 36% for HCV genotype 1 infection and 43% to 73% for HCV genotypes 2 and 3.90 Guidelines for therapy have been proposed.76 Fungal infections of the liver are not unusual when immunocompromise is advanced. Histoplasmosis, cryp­ tococcosis, and coccidioidomycosis of the liver may be observed in patients with disseminated fungal disease, predominantly but not exclusively in regions of high prevalence of the organism.91 Candida infection of the liver is rare, in contrast to its high prevalence in mucosal sites. Kaposi’s sarcoma, which is caused by infection with human herpesvirus 8 (HHV-8), is most often found at postmortem or incidentally at liver biopsy but may occasionally cause aminotransferase elevations or jaundice. Hepatic involvement by non-Hodgkin’s lymphoma may be the index manifestation of AIDS and the primary site of the neoplasm. This tumor in the AIDS patient tends to be more aggressive histologically and clinically, spreading rapidly to extranodal sites, making liver involvement more likely.92 The lesions are typically focal and may be large.93 The prognosis is determined largely by the extent of underlying immunocompromise and Karnofsky performance score rather than the lymphoma itself. Improvements in survival have been demonstrated in those receiving HAART.94 Isolated cases of P. jiroveci hepatitis have been described and are attributable to the use of inhaled pentamidine, which fails to protect extrapulmonic sites from this opportunistic pathogen. The liver may also be the site of infection by the protozoa Cryptosporidium, Microsporidium, or Dicrocoelium dentriticum or by other multicellular organisms including leishmania.95 Bacillary peliosis hepatis, caused by either Bartonella henselae or Bartonella quintana, is a systemic infection that may be associated with fever, skin lesions, abdominal pain, and lytic bone lesions.96 Liver tests usually show a disproportionate elevation of the serum alkaline phosphatase. Liver biopsies demonstrate regions of a myxoid stroma in association with granular purple material, which with Warthin-Starry stain or electron microscopy reveal clumps of organisms. Treatment with either erythromycin (orally, or in severe cases, intravenously), tetracycline, minocycline, or a cephalosporin are reportedly effective, although prolonged or lifelong therapy is necessary. Biliary tract involvement in AIDS may result in marked liver test abnormalities and right upper quadrant symptoms; jaundice is unusual. A syndrome resembling sclerosing cholangitis with papillary stenosis is well recognized and has been termed AIDS cholangiopathy. Patients characteristically develop significant upper abdominal pain in association with marked elevation of serum alkaline phosphatase, as well as minimal elevations of bilirubin, AST, and ALT. Currently, biliary tract disease is as frequently due to non– HIV-related diseases (e.g., bile duct stones) as diseases linked to AIDS.

Figure 33-8.  Endoscopic retrograde cholangiopancreatography in a patient with acquired immunodeficiency syndrome cholangiopathy. Papillary stenosis is present (arrow).

Ductular changes consist of papillary stenosis alone, sclerosing cholangitis-like lesions alone, a combination of the two, or long extrahepatic strictures. Most series have found papillary stenosis with intrahepatic disease as the most common findings (Fig. 33-8). Ultrasonography or CT detects ductular abnormalities, usually dilatation, in most of those with cholangiographically proven disease, implying that a negative imaging study does not definitively exclude the diagnosis. The etiology in most cases is due to infection of the duodenal and biliary epithelium with Cryp­ tosporidium, CMV, or Microsporidium.97 For patients with predominantly papillary stenosis, sphincterotomy results in a symptomatic improvement in most patients; serum alkaline phosphatase may continue to rise, however, probably reflecting progression of associated intrahepatic disease. In some patients eradication of the infecting pathogen results in improvement of the radiographic abnormalities.98 Survival in AIDS cholangiopathy is linked to severity of immunodeficiency.99 Other less common causes of biliary tract disease in AIDS include primary bile duct lymphoma, epithelial angiomatosis, and lymphomatous nodal obstruction of the biliary tree, Kaposi’s sarcoma, and biloma. In addition, chronic pancreatitis or choledocholithiasis may lead to biliary obstruction, although their incidence is not clearly increased in HIV infection. Acalculous cholecystitis has also been described in AIDS patients, presenting as severe abdominal pain and, occasionally, peritonitis. This syndrome is usually caused by a specific infection, most frequently CMV, but also from microsporidia, cryptosporidia, and I. belli.100 Laparoscopic cholecystectomy is the treatment of choice.

EVALUATION AND MANAGEMENT

The initial decision in evaluating the AIDS patient with jaundice, hepatomegaly, or both is to determine whether the

Chapter 33  Gastrointestinal Consequences of Infection with Human Immunodeficiency Virus findings are due to intrahepatic or extrahepatic disease. Simultaneous disease in both sites must also be considered. A history of mild jaundice, often in association with fever and constitutional symptoms, is more consistent with intrahepatic disease, whereas symptoms of deep jaundice associated with pain of relatively acute onset suggest extrahepatic disease. Careful review of medications, both prescription and nonprescription, is essential. Because the clinical history and the finding of symptomatic hepatomegaly are nonspecific, further evaluation is always necessary. Elevations of serum ALT and/or AST are common, but neither the pattern nor the extent of elevation of these tests appears to correlate with specific findings in the liver. Nevertheless, some generalizations can be made. Significant elevation of the aminotransferases favors a drug-induced or viral cause. In contrast, marked elevation of alkaline phosphatase correlates statistically with the presence of MAC infection in the liver in AIDS when extrahepatic obstruction is absent. Ultrasonography, CT, and MR cholangiography (MRC) should be used early because they are especially useful in identifying ductal dilation, gallbladder pathology, and focal hepatic lesions.101 The indications for liver biopsy for the patient with suspected intrahepatic disease are limited. Biopsy is appropriate when symptomatic, treatable disease of the liver is anticipated, and when a specific diagnosis of hepatic disease is necessary. Although a specific diagnosis is likely in most patients, liver biopsy rarely identifies a previously undiagnosed infection, suggesting that the liver is rarely the site of disease not manifest elsewhere. This observation underscores the importance of reserving liver biopsy for those circumstances in which less invasive diagnostic methods such as blood cultures and bone marrow biopsy have not yielded a diagnosis.102 Biopsy may play a role in the treatment decision for HCV therapy as in the normal host. Focal lesions identified by abdominal imaging can be sampled under ultrasonography or CT guidance. Use of transjugular liver biopsy may be indicated in selected settings such as hemophilia. Specific infections or neoplasms are usually evident on tissue sections of appropriately stained biopsy material. An extrahepatic cause for jaundice is suggested on CT or ultrasonography by the presence of dilated ducts or other biliary and/or pancreatic abnormalities. Once extrahepatic obstruction is recognized, the possibility of papillary stenosis associated with AIDS cholangiopathy must be considered, as well as the possibility of choledocholithiasis or other disorders, depending on the imaging studies. Further

testing when indicated may include endoscopic retrograde cholangiopancreatography (ERCP) if CT, ultrasonography, or MRC demonstrates extrahepatic biliary ductal dilation. Bile duct, ampullary, and duodenal biopsy specimens or bile and/or biliary cytology (with appropriate staining) collected during ERCP can be examined for the presence of viruses, protozoa, or neoplastic cells.

KEY REFERENCES

Call SA, Heudebert G, Saag M, et al: The changing etiology of chronic diarrhea in HIV-infected patients with CD4 cell counts less than 200 cells/mm3. Am J Gastroenterol 2000; 95:3142-7. (Ref 17.) Ko WF, Cello JP, Rogers SJ, et al: Prognostic factors for the survival of patients with AIDS cholangiopathy. Am J Gastroenterol 2003; 98:2176-81. (Ref 99.) Monkemuller KE, Call SA, Lazenby AJ, Wilcox CM. Declining pre­ valence of opportunistic gastrointestinal disease in the era of com­ bination antiretroviral therapy. Am J Gastroenterol 2000; 95:457-62. (Ref 7.) Novoa AM, de Olalla PG, Clos R, et al: Increase in the non-HIV-related deaths among AIDS cases in the HAART era. Curr HIV Res 2008; 6:77-81. (Ref 60.) Puoti M, Torti C, Ripamonti D, et al: Severe hepatotoxicity during combination antiretroviral treatment: Incidence, liver histology, and outcome. J Acquir Immune Defic Syndr 2003; 32:259-67. (Ref 66.) Rockstroh JK, Bhagani S, Benhamou Y, et al: European AIDS Clinical Society (EACS) guidelines for the clinical management and treatment of chronic hepatitis B and C coinfection in HIV-infected adults. HIV Medicine 2008; 9:82-8. (Ref 76.) San-Andrés FJ, Rubio R, Castilla J, et al: Incidence of acquired immunodeficiency syndrome–associated opportunistic diseases and the effect of treatment on a cohort of 1115 patients infected with human immunodeficiency virus, 1989-1997. Clin Infect Dis 2003; 36:117785. (Ref 2.) Sanchez TH, Brooks JT, Sullivan PS, et al: Bacterial diarrhea in persons with HIV infection, United States, 1992-2002. Clin Infect Dis 2005; 41:1621-7. (Ref 42.) Tanaka PY, Hadad DJ, Barletti SC, et al. Bone marrow biopsy in the diagnoses of infectious and non-infectious causes in patients with advanced HIV infection. J Infect 2007; 54:362-6. (Ref 102.) Weber R, Sabin CA, Friis-Müller N, et al. Liver-related deaths in persons infected with the human immunodeficiency virus: the D:A:D study. Arch Intern Med 2006; 166:1632-41. (Ref 6.) Wilcox CM: Etiology and evaluation of diarrhea in AIDS: A global perspective at the millennium. World J Gastroenterol 2000; 6:177-186. (Ref 32.) Wilcox CM, Alexander LN, Clark WS, et al: Fluconazole compared with endoscopy for human immunodeficiency virus-infected patients with esophageal symptoms. Gastroenterology 1996; 110:1803-9. (Ref 14.) Wilcox CM, Schwartz DA, Clark WS: Esophageal ulceration in human immunodeficiency virus infection: Etiology, response to therapy, and long-term outcome. Ann Intern Med 1995; 123:143-9. (Ref 10.) Full references for this chapter can be found on www.expertconsult.com.

535

CHAPTE R

34 Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation Anne M. Larson and George B. McDonald

CHAPTER OUTLINE Complications of Solid Organ Transplantation  537 Kidney and Kidney/Pancreas Transplantation  539 Liver Transplantation  540 Heart, Lung, and Heart/Lung Transplantation  540 Intestinal Transplantation  541 A Problem-Oriented Approach to Diagnosis in Solid Organ Transplant Recipients  541

Transplantation of a solid organ is an immunologic mirror of the transplantation of allogeneic hematopoietic cells. Transplanted organs can be rejected by the patient in whom they are placed, whereas allogeneic hematopoietic cells can damage the organs of their recipient. There are similarities in the intestinal and hepatic complications of these transplant procedures, particularly with regard to infections and the side effects of immunosuppressive drugs. However, there are extreme differences in the patient populations being transplanted, in the preparation for transplant, and in the degree and length of immunosuppression. For this reason, this chapter presents separate problem-oriented approaches to the complications of solid organ and hematopoietic cell transplantation.

COMPLICATIONS OF SOLID ORGAN TRANSPLANTATION Gastrointestinal complaints after solid organ transplant (SOT) are reported in 20% to 35% of recipients, with a frequency as high as 60% reported in India.1 Most of the problems relate to graft dysfunction, adverse effects of medications, opportunistic infections, or malignancy (Table 34-1).1-3 Infectious complications remain a major source of morbidity and mortality following SOT, particularly within the first six months. During the first month following

Complications of Hematopoietic Cell Transplantation  544 Evaluation of Gastrointestinal and Liver Problems before Transplantation  544 Problems from Transplant through Day 200  545 Problems in Long-Term Transplant Survivors  553

SOT, infections include those present prior to transplant (e.g., urinary tract infection), those related to technical complications of the procedure itself (e.g., biliary sepsis), or those transmitted with the allograft. Opportunistic viral, fungal, and parasitic infections are more likely to develop after the first month, with herpesvirus infections being the most common (Fig. 34-1). Several noninfectious complications can mimic infection (see Table 34-1). Cytomegalovirus (CMV) is the predominant viral pathogen occurring within the first year after SOT, with the intestine and hepatobiliary tracts major sites of infection (see Fig. 34-1). Factors predisposing to CMV infection include the type of immunosuppression used, that is, use of antilymphocyte antibody in addition to conventional immunosuppression or maintenance mycophenolate mofetil (MMF) therapy, and the recipient’s risk of infection.4,5 CMVnegative recipients who received a CMV-positive graft are at the greatest risk of primary CMV infection.6 The peak incidence is generally four to six months after transplantation, with fever, malaise, myalgia, and occasionally cough and minor elevations of serum alanine aminotransferase (ALT).7 CMV-deoxyribonucleic acid (DNA) or antigen is generally detected in the bloodstream, but CMV can be recovered from intestinal biopsy tissue in the absence of detectable virus in the bloodstream. Either post-transplant antiviral prophylaxis or preemptive therapy with either ganciclovir or valgan­ciclovir significantly reduces the risk of CMV disease.8-11 Valganciclovir should not be used in the

537

Medications Obstruction Uremia Dialysis Pancreatitis Hepatitis Cholecystitis Gastroparesis

GERD ± peptic stricture Pill esophagitis Thoracostomy tube

Noninfectious Causes Intestinal obstruction Pseudo-obstruction Narcotic bowel syndrome Diverticular disease Ischemic colitis Acute pancreatitis Intestinal motility disorder (intestinal transplant) Biliary leak (LT) (Acute GVHD)

CMV Clostridium difficile H. pylori–related ulcers Perforation with abscess, peritonitis Acute cholecystitis (Viral pancreatitis) (VZV)

ABDOMINAL PAIN

NSAID, idiopathic gastroduodenal ulcers GERD Diverticula (especially KT) Ischemic colitis (especially KT) Biliary or Roux-en-Y anastomotic bleeding (LT) Liver biopsy (hemobilia) Variceal bleeding (Acute GVHD)

CMV Fungal infection (Candida, molds) H. pylori–related ulcers EBV-LPD C. difficile (HSV esophagitis)

GASTROINTESTINAL BLEEDING

Promotility drugs MMF Sorbitol colitis Ischemic colitis Mg++ salts Antibiotic-associated diarrhea

CMV C. difficile EBV-LPD (G. lamblia) (Cryptosporidia) (Microsporidia) (Rotavirus) (Strongyloides) (Enteric bacterial pathogens)

DIARRHEA

Lymphoma Skin cancer Colon cancer Recurrent hepatocellular carcinoma Lung cancer

EBV-LPD MALT lymphoma (H. pylori-related) (Kaposi’s sarcoma)

MALIGNANCY

Drug toxicity Vascular injury (LT) Nodular regenerative hyperplasia Biliary tract disease Recurrent hepatocellular carcinoma

Sepsis-related cholestasis (cholangitis lenta) Herpes viruses (CMV, HSV, VZV, EBV) HBV HCV

HEPATOBILIARY PROBLEMS

*Conditions that are in the literature but rarely seen are in parentheses. CMV, cytomegalovirus; EBV-LPD, Epstein-Barr virus–lymphoproliferative disease; GERD, gastroesophageal reflux disease; GVHD, graft-versus-host disease; HBV, hepatitis B virus; HCV, hepatitis C virus; HSV, herpes simplex virus; KT, kidney transplant; LT, liver transplantation; MALT, mucosa-associated lymphoid tissue; MMF, mycophenolate mofetil; NSAID, nonsteroidal anti-inflammatory drug; VZV, varicella-zoster virus.

CMV HSV Helicobacter pylori (VZV) (Giardia lamblia) (Cryptosporidia) (Rotavirus) EBV-LPD

Candida albicans Other fungal species CMV HSV (VZV) (Mycobacterium tuberculosis) (Parasites)

ANOREXIA, NAUSEA, VOMITING

Infections

ESOPHAGEAL SYMPTOMS

Table 34-1  Causes of Intestinal and Hepatobiliary Problems in Solid Organ Transplant Recipients*

538 Section IV  Topics Involving Multiple Organs

Chapter 34  Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation

A

B A

C

D

Figure 34-1.  Endoscopic photographs of gastrointestinal tract infections following solid organ transplantation. A, Distal esophageal ulcerations caused by cytomegalovirus. B, Duodenal ulceration caused by herpes simplex virus, showing a deep, irregular ulcer surrounded by edematous mucosa. C, Colon mucosa in cytomegalovirus infection, showing focal ulceration (arrow) and depressed ulceration and intramucosal hemorrhage in surrounding mucosa. D, Colon mucosa in cytomegalovirus infection showing diffuse mucosal friability and ulceration.

setting of liver transplantation, however, because there is a higher rate of tissue-invasive disease.12 In this setting, ganciclovir is recommended. Herpes simplex virus (HSV) is the second most commonly seen viral infection and characteristically represents reactivation of latent virus within the recipient, typically two to four weeks after transplant. HSV has tropism for squamous epithelium (nose, mouth, esophagus) but can involve the intestine and liver if patients are not receiving pro­ phylaxis (see Fig. 34-1B). Other herpesvirus infections— Epstein-Barr virus (EBV), varicella-zoster virus (VZV), and human herpesvirus 6 (HHV-6)—are less common. MMF immunotherapy may increase the risk of VZV dissemination. Fungal infections usually develop after the first month post-transplant, particularly among patients who have discontinued fungal prophylaxis. The most common fungi are candidal species (Candida albicans, Candida tropicalis), but molds such as Aspergillus and Zygomycetes are emerging as pathogens.13 Less common infections (Nocardia, Pneumocystis, Toxoplasma; parasites such as Strongyloides) also may occur after the first month. Once beyond the first six months following SOT, opportunistic infections occur less frequently, but recipients remain at risk for community-acquired infections. Post-transplant lympho­ proliferative disease continues to be a problem for SOT recipients, who require continued high-level immune suppression. B and T cell lymphomas can be seen (Fig. 34-2).

KIDNEY AND KIDNEY/PANCREAS TRANSPLANTATION

Many of the serious infections reported in kidney transplant (KT) recipients are now uncommon because of more intense

B Figure 34-2.  Computed tomographic findings in lymphoproliferative disease following solid organ transplant. A, Retroperitoneal mass (arrows) caused by an Epstein-Barr virus–positive B cell lymphoma following liver transplantation. B, Distal small intestinal mass (arrows) following renal transplantation, caused by a T cell lymphoma. The mass was causing intestinal obstruction, as evidenced by the dilated loops of small intestine proximal to the mass.

surveillance, prophylaxis, and preemptive treatment of viral and fungal infection. However, if untreatable lifethreatening infection should develop, immunosuppressive drugs can be discontinued and the patient maintained on dialysis. This option is unavailable to recipients of other organs. Gastrointestinal complications are among the most pre­ valent complications post KT, seen in up to 50% of patients, and correlate with patient long-term survival.14-16 It has been reported that KT patients who experience gastroesophageal reflux disease (GERD) or dyspepsia have an increased risk of graft loss and death, the mechanism of which is unclear.17 Graft pancreatitis and graft duodenitis generally occur early after kidney/pancreas transplant (KPT) and may lead to intra-abdominal infection.18,19 The frequency of hepatitis C virus (HCV) or hepatitis B virus (HBV) infection ranges from 5% to 66% of KT and KPT recipients, depending on country of origin. The effect of HCV on patient and graft outcomes remains controversial.20 Many have shown outcomes to be inferior in patients who are chronically infected with either HCV or HBV.21-24 HBV antiviral therapy has improved clinical outcome, but HCV antiviral therapy with interferon alpha and ribavirin cannot be used in the post-KT setting because of increased risk of allograft rejection. Cirrhotic patients who undergo KT have a significantly worse 10-year survival (about 20% to 30%). Gastrointestinal CMV infection is seen in about 7% of KT and KPT recipients, with pancreas recipients at greater risk due to higher levels of immunosuppression.25 About 4% develop intestinal fungal infections, most often with

539

540

Section IV  Topics Involving Multiple Organs candidal species. HSV infection post KT is generally asymptomatic and self-limited, presenting as stomatitis, mononucleosis, hepatitis, or pneumonia.26 Cholecystitis is seen in KT recipients, and the incidence is higher among diabetic patients.27 Traditionally gastrointestinal hemorrhage occurred in up to 20% of KT recipients and had a high mortality.28 Many KT recipients with gastroduodenal ulcers have no history of gastroduodenal disease. Approximately 50%, however, have complaints of dyspepsia, and about 30% are colonized with Helicobacter pylori.29 With decreased use of glucocorticoids and use of proton pump inhibitors, ulcer formation and hemorrhage are rare.3 Renal recipients are at particular risk for the development of intestinal ischemia compared with other SOT recipients. However, the incidence is low (<5%) and the etiology is multifactorial.30 Recipients with polycystic kidney disease more often develop intestinal ischemia and obstruction.31 Intestinal ischemia in this setting carries a high mortality. Ischemia should be considered in KT recipients with abdominal pain, particularly older patients (>40 years of age) who have received a cadaveric kidney.30

A

B

LIVER TRANSPLANTATION

Gastrointestinal complications unique to orthotopic liver transplant (OLT) are generally related to the surgery itself, that is, hemorrhage, hepatic arterial stenosis or thrombosis, biliary tract dysfunction, bowel perforation, bowel obstruction, and gastrointestinal bleeding.32 Hepatic artery thrombosis presents with a spectrum of consequences, ranging from mildly elevated liver enzymes to fulminant hepatic failure. Post-OLT, the biliary tree receives its entire blood supply from the hepatic artery, thus loss of flow results in bile duct necrosis and leakage with development of bilomas and abscesses (Fig. 34-3A and B). Gradual loss of hepatic arterial flow can result in ductopenia, which is indistinguishable from ductopenic rejection. Portal vein thrombosis can lead to hepatic ischemia and severe hepatic dysfunction if it occurs early in the post-transplant course; later, signs of portal hypertension develop. Rarely, hepatic vein thrombosis and inferior vena cava thrombosis/stenosis can create a Budd-Chiari–like syndrome. Biliary leakage and stricture formation, generally at the anastomotic site, are the most common biliary abnormalities seen following OLT (see Fig. 34-3A and B).33 Anastomotic strictures generally occur within two to six months post OLT, but can occur in the newly transplanted patient as well. Strictures and leaks in patients with duct-to-duct anastomoses are often amenable to endoscopic therapy, whereas those with choledochojejunostomies may require percutaneous or surgical correction. The incidence of biliary cast syndrome has decreased to 5% to 20%, and generally occurs within the first year post OLT.34 Clinical factors associated with development of biliary casts include hepatic ischemia and biliary strictures. Endoscopic and percutaneous therapy is successful in up to 70%, but surgical intervention may be required, and mortality is reported at 10% to 30%.35 CMV hepatitis is more severe in OLT recipients than in recipients of other organs.6,36 Patients often have elevations in serum aminotransferases, which can be confused with rejection, and therefore liver biopsy is essential for differentiation. The diagnosis can usually be confirmed by detection of CMV in the bloodstream. Asymptomatic low-level CMV viremia does not require antiviral therapy.37 Liver transplant recipients more often develop invasive fungal infections than other SOT recipients, with a high mortality. In the absence of prophylaxis, intestinal colonization with

C Figure 34-3.  Hepatobiliary imaging following liver transplantation. A, Endoscopic retrograde cholangiogram showing an ischemic stricture of the bile duct (arrow). B, Endoscopic retrograde cholangiogram showing a bile leak (arrowhead) at the biliary anastomosis (arrow). C, Magnetic resonance cholangiogram of the intrahepatic biliary system showing recurrent sclerosing cholangitis in the liver graft. The arrow points to a stricture, with upstream biliary dilation.

Candida is nearly universal post OLT, and Candida accounts for the majority of all invasive fungal infections following OLT.38 A serum galactomannan assay is useful for detecting mold infections.39,40 There is a risk for recurrence of the underlying liver disease following OLT, including HCV, HBV, autoimmune hepatitis, nonalcoholic steatohepatitis (NASH), primary biliary cirrhosis (PBC), and primary sclerosing cholangitis (PSC) (see Fig. 34-3C).41-43 Recurrence of HCV in the liver allograft is nearly universal, with 75% developing signs of liver damage and 25% progressing to cirrhosis within 5 years, which leads to increased graft loss.44-46 HBV recurrence may be prevented with the use of hepatitis B immunoglobulin (HBIG) and antiviral medications. PBC recurs in about 26% of patients post-liver transplant.41

HEART, LUNG, AND HEART/LUNG TRANSPLANTATION

Up to half of heart (HT), lung (LT), and heart-lung transplantation (HLT) recipients experience gastrointestinal complications, with up to 20% requiring surgery.47,48 The most common complications include diarrhea, GERD, dyspepsia, nausea and vomiting, abdominal pain, pancreatitis, herpesvirus infections (especially CMV), cholelithiasis, ulcers, and hepatobiliary disease.47-49 GERD and gastro­ paresis are particularly problematic after LT or HLT and

Chapter 34  Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation may be related to medications and vagal nerve injury during the operation.48,50,51 Symptomatic gastroparesis has been described in 25% of LT recipients and up to 80% in HLT recipients.52,53 The course is often waxing and waning, suggesting a neuropathic, infectious (CMV), or medication-induced etiology, but ultimately there is partial or complete remission.52,54 Recipients with GERD and/or gastroparesis are at particular risk for the development of obliterative bronchiolitis, which significantly threatens the longevity of LT recipients.51,52 Proton pump inhibitors can be used to help control reflux; however, if reflux disease is unremitting, laparoscopic fundoplication may be successful.55-57 LT recipients may develop giant gastric ulcers (>3 cm in diameter) that occur despite routine use of acid suppression. These ulcers carry significant morbidity and mortality, and are more often associated with bilateral LT, high-dose nonsteroidal anti-inflammatory drugs (NSAIDs) after transplant, acute rejection requiring high-dose glucocorticoids, and cyclosporine immunosuppression. For this reason, some authors believe NSAIDs should not be used in the post-transplant setting. Recipients of LT and HT more often develop CMV infection (15% to 25%) than other SOT recipients. Generally CMV infection presents as pneu­ monitis, but gastrointestinal CMV infection remains a major cause of morbidity (see Fig. 34-1). LT and HLT recipients have the highest incidence of fungal infection in the SOT setting in which Aspergillus, not Candida species, predominates. Patients undergoing LT for cystic fibrosis experience a unique set of gastrointestinal complications.58 Pancreatic insufficiency, a marker for severe cystic fibrosis, is common. Cystic fibrosis–induced secondary biliary cirrhosis can complicate absorption of immunosuppressive medications such as cyclosporine. If severe liver disease is detected prior to LT, lung-liver transplant should be considered. Distal intestinal obstruction syndrome occurs in about 20% and is similar to the incidence in the nontransplant setting. Cystic fibrosis patients also may experience cholecystitis, peptic ulcer disease, and GERD. Primary HCV infection following HT leads to significantly decreased one- and three-year survival. However, acquisition of HBV following HT does not appear to affect survival, at least up to five years.59,60

INTESTINAL TRANSPLANTATION

Most complications are related to underlying diseases, graft rejection, intestinal ischemia, and anastomotic leaks. Bacterial and fungal infections are common, often associated with mucosal disruption following surgery, but a source may not be identifiable. Two types of malignancy related to intense immunosuppression have been reported, EBV lymphoproliferative disease (EBV-LPD) and de novo cancers of nonlymphomatous origin.61,62 Surveillance for EBV DNA and preemptive treatment reduces the frequency of lymphoproliferative disease (see following). Altered intestinal motility and anorexia have been reported.

the rapidity with which disease can progress. GERD is the most common cause of heartburn and midchest pain, particularly following lung transplantation (see lung transplant earlier), but viral and fungal esophagitis may underlie these symptoms, particularly after chemoprophylaxis has been discontinued. Candidal esophagitis is seen with particular frequency in those with diabetes; other risk factors include use of broad-spectrum antibiotics, high-dose immunosuppression, and the presence of a Roux-en-Y anastomosis in liver transplant recipients. Severe necrotizing fungal esophagitis can lead to perforation, which can have a fatal outcome in up to one third of patients. Odynophagia, dysphagia, or hematemesis should lead to consideration of an esophageal infection; herpesviruses (CMV, HSV) and fungal species (Candida) are responsible for the largest proportion, but unusual organisms can be seen.63 Dysphagia secondary to pill esophagitis may develop in SOT recipients, caused by antibiotics, antivirals, potassium chloride, bisphosphonates, and NSAIDs. Esophageal strictures following severe esophageal infection have been reported and may present a long time after eradication of the organism. Anorexia, nausea, and/or vomiting are common following SOT, particularly early in the post-transplant course.15,47,48 These symptoms are often related to herpesvirus infections or to medications (including immunosuppressive drugs), and thus, endoscopic evaluation is necessary for diagnosis in most patients. Tacrolimus (Prograf) is a macrolide lactone that can cause nausea, abdominal pain, and diarrhea, often leading to anorexia, food aversion, and weight loss. These side effects are dose dependent and can be managed with dose reduction or, more rarely, drug discontinuation. Sirolimus (Rapamune), a newer macrolide immunosuppressant, has a GI side effect profile similar to tacrolimus. MMF (CellCept) is an inhibitor of nucleic acid synthesis with gastrointestinal side effects of nausea, vomiting, and diarrhea, often requiring dosing modifications. A formulation of mycophenolic acid delayed-release tablets (Myfortic) appears to have significantly fewer gastrointestinal side effects with similar therapeutic efficacy.64 Less common causes of anorexia and nausea include pancreatitis, cholecystitis, or cystitis. Rarely following SOT, graft-versus-host disease (GVHD) presents with fever, skin rash, and gastrointestinal symptoms, particularly nausea, vomiting, and diarrhea.65,66 Endoscopic evaluation with biopsy is essential if GVHD is suspected and skin lesions are absent, recognizing that other conditions such as viral infections and drug reactions can have a GHVD-like histologic pattern.67 Symptomatic gastroparesis is frequently seen in the setting of lung transplant but is less often reported in the setting of other solid organ transplant.52 CMV and VZV may rarely involve intestinal neural plexuses, leading to intestinal dilation or gastroparesis. H. pylori infection may be associated with symptomatic dyspepsia, gastritis, and gastroduodenal ulceration, but there is no relationship between the use or degree of immunosuppression and H. pylori colonization; its incidence is similar to that seen in the nontransplant setting.2

Diarrhea and Constipation A PROBLEM-ORIENTED APPROACH TO DIAGNOSIS IN SOLID ORGAN TRANSPLANT RECIPIENTS Upper Gastrointestinal Symptoms and Signs

The approach to SOT patients with esophageal or gastric symptoms is influenced by a high frequency of nonspecific symptoms as harbingers of serious infection (for example, CMV infection presenting as nausea and vomiting) and by

Colonic and small bowel complications (diverticulitis, ischemic colitis, malignancy, and infections) have been reported to occur following all types of SOT. Early in the posttransplant setting, infections predominate. Diarrhea is commonly infectious and may be accompanied by fever,68 abdominal pain (46%), nausea (32%) and vomiting (22%).69,70 The microbes usually responsible are CMV and Clostridium difficile, but the literature describes a wide range of organisms in SOT recipients, particularly when

541

542

Section IV  Topics Involving Multiple Organs they are cared for in infection-endemic areas (for example, adenovirus, rotavirus, coxsackievirus, bacterial enteric pathogens, enterohemorrhagic Escherichia coli, Yersinia enterocolitica, Giardia lamblia, Candida species, cryptosporidium, Enterocy­tozoon bieneusi, Isospora belli, and Strongyloides stercoralis).70 Diagnosis can be made by examination of stool specimens in nearly all cases; the exceptions are CMV, certain parasites, and EBV-LPD. Small intestinal involvement with CMV often causes profuse watery diarrhea with protein-losing enteropathy, particularly if the diagnosis is delayed.71 Colonic involvement may appear as an inflammatory colitis resulting in bloody diarrhea and is often associated with fever, abdominal distention, and pain.72 Diagnosis of CMV may require mucosal biopsy, particularly if blood specimens are negative for CMV DNA or antigen. C. difficile infection may present with a more severe course post SOT, and patients with fulminant colitis and toxic megacolon require prompt surgical intervention to prevent perforation and peritonitis.70 Signs of colitis may be subtle due to concomitant immunosuppression. Only about 70% of patients respond to treatment with metronidazole; persistent and more severe cases require oral vancomycin. Recurrence may develop in up to 20% of cases.73 The use of probiotics (e.g., Saccharomyces boulardii) in SOT recipients remains controversial because there have been reports of yeast dissemination and infection in the immunocompromised host.74 Intestinal fungal infections can be seen in up to 25% of SOT recipients. In the absence of prophylaxis, intestinal fungal overgrowth and diarrhea can result from antibiotic use or intestinal dysmotility. Common parasitic infections also must be considered in an immunocompromised host, particularly in areas of high endemicity. The protozoa are a much less frequent cause of acute diarrhea post SOT. Microsporidia (E. bieneusi) is a more rarely reported cause of chronic diarrhea, perhaps reflecting the fact that it is often not sought out in the postSOT setting. Clinically, patients with this infection experience fatigue, intermittent diarrhea, and weight loss. There are no clearly effective therapies for E. bieneusi. Symptoms of colitis or toxic megacolon are most often associated with infection, but in up to 20% of cases, no clear etiology can be found.70,75 Early recognition, diagnosis, and treatment of colitis can decrease disease-associated mortality. Eosinophilic colitis with diarrhea has been reported with the use of tacrolimus and cyclosporine. Histologically this is characterized by eosinophilic colonic infiltrates and peripheral eosinophilia. Elevated serum immunoglobulin E may be present in some patients. Drug-related diarrhea is seen in up to a third of SOT patients, most commonly with tacrolimus or sirolimus.70,76 MMF causes watery diarrhea in up to 30% of patients, and may require dose reduction or discontinuation. The mechanism of MMF-induced diarrhea is unclear. There have also been reports of altered tacrolimus metabolism and absorption in patients suffering from MMF-induced or other causes of chronic diarrhea. Antithymocyte globulin (ATG) and anti–T cell antibody (OKT3) therapies are associated with diarrhea, which predictably lasts for three to four days and resolves spontaneously. Most cases of immunosuppressantinduced diarrhea can be managed with dose manipulation, but some are so severe that discontinuation of the immunosuppressant is required. Diarrhea also can be caused by magnesium-containing preparations prescribed to correct renal magnesium wasting and by antibiotics prescribed either prophylactically or therapeutically. Noninfectious diarrhea has been reported to increase the risk of graft loss and mortality.76

Constipation is seen in less mobile recipients who are receiving certain medications (e.g., narcotics, calcium- and aluminum-containing antacids, anticholinergics). This is generally responsive to increased patient mobility, decreased use of narcotics, use of methylnaltrexone,77 and therapy with laxatives and senna.

Abdominal Pain

Abdominal complications are common, affecting up to 30% of patients following SOT.78 Symptoms may be mild despite the presence of life-threatening complications. All patients with abdominal pain should be aggressively evaluated, with particular attention to whether the patient requires urgent surgery or a specific medical treatment. Most recipients with abdominal pain will not need surgery. The intra-abdominal conditions presenting with pain that require urgent surgery are abscess, perforation, severe colitis, appendicitis, intestinal obstruction, intestinal ischemia, and acute cholecystitis. These disorders may appear in the early post-transplant period. Immunosuppression may mask symptoms and suppress the host response, leading to a delay in diagnosis and increase in mortality. Most transplant patients with acute appendicitis have right lower quadrant pain. Overall, intestinal perforation occurs in less than 5% of SOT recipients, although the incidence may be slightly higher in the setting of lung transplant.14,79 Perforation may occur spontaneously without clear etiology, but it is associated with colon diverticula in up to two thirds of cases (particularly renal transplant recipients) and ischemia in 15%. Perforation, especially of a diverticulum, carries a mortality of up to 55%.30,72 Risk factors for the development of colonic perforation include diverticular disease, immunosuppression (particularly glucocorticoids), CMV infection, fungal infections (e.g., mucormycosis), unrecognized lymphoma (EBV-LPD), colon cancer, and ischemia.2,72 Abdominal radiographs and helical computed tomography (CT) scans can confirm the presence of perforation, but may not reveal its source before surgery. Diverticular perforation is especially common after renal transplant, often leading to abscess formation and fistulization, sometimes without causing severe pain or findings of peritonitis. Pretransplant colonic screening for diverticulosis in patients less than 50 years of age has not been shown to predict post-transplant colonic perforations. SOT recipients also are at increased risk for the development of cholelithiasis.80 Factors related to gallstones include cyclosporine, obesity, and cystic fibrosis as an underlying disorder. Abdominal pain is frequently associated with tissue-invasive CMV disease. Although generally producing a diffuse pattern of mucosal edema, CMV may also cause focal ulceration, perforation, high-grade stricture, and intestinal obstruction (see Fig. 34-1). The first manifestation of disseminated VZV infection is often severe abdominal pain related to intestinal pseudo-obstruction and visceral neuropathy. Early treatment of both CMV and VZV infection results in improved survival. Abdominal pain may also be a manifestation of transplant-related complications that do not usually have a dire outcome. Pain has been reported with oral tacrolimus, sirolimus, and MMF. Abdominal pain is seen in up to 19% of patients taking MMF and can significantly limit its use.81 The etiology of MMF-induced pain been postulated to involve local irritant and inflammatory effects as well as interference with rapidly dividing intestinal cells, a hypothesis supported by a study showing fewer gastrointestinal complications with delayed-release mycophenolic acid (Myfortic), than with MMF.64 Narcotic-induced ileus is

Chapter 34  Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation common after surgery. Care must be taken to rule out an infectious etiology such as CMV or VZV, both of which can involve the intestinal nerve plexuses.2 Noninfectious pseudo-obstruction often can be managed conservatively with nasogastric decompression, vigorous correction of electrolyte imbalance, and withdrawal of opiates. Opioidrelated gut symptoms can also be blocked with the use of methylnaltrexone while not interfering with central pain relief.77 Neostigmine can be safely used for treatment of intestinal pseudo-obstruction in the transplant setting.72 Surgical intervention may be required in the setting of massive colon dilation. Acute pancreatitis has been reported in 1% to 2% of renal transplant recipients, up to 6% of liver transplant recipients, and up to 18% of heart transplant recipients; it may have a fatal outcome.82 Acute pancreatitis is associated with CMV infection, hypercalcemia, cholelithiasis, biliary manipulation, malignancy, recent alcohol ingestion, and medications such as azathioprine, cyclosporine, tacrolimus, and glucocorticoids. Treatment of pancreatitis in the post-transplant setting is identical to that in the non-transplant setting, except for the need to exclude viral infection and some immune suppressive medications. Pneumatosis intestinalis may be discovered during abdominal imaging after SOT as an incidental finding, but can also be a manifestation of life-threatening intestinal ischemia or infection with a gas-forming organism. Pneumatosis intestinalis is associated with CMV infection, C. difficile colitis, and sepsis and can be seen in patients receiving glucocorticoid therapy. The majority of patients require no specific intervention, and the gas collections resolve spontaneously unless caused by ischemia or an infection with a clostridial organism

Gastrointestinal Bleeding

When gastrointestinal bleeding occurs, it is often secondary to infectious ulcers. Noninfectious causes of hemorrhage include NSAID gastroduodenal ulcers, diverticular bleeding, anastomotic bleeding, and ischemic colitis. The current incidence of gastroduodenal ulcer disease in the transplant population is about 5%, with perforation rates of less than 1%.3 Prophylaxis with histamine (H2)-receptor antagonists or proton pump inhibitors decreases the occurrence of ulcer disease in this population; both therapies are equally effective.3 Patients infected with H. pylori prior to transplantation are more likely to develop peptic ulcer disease following transplant.83 In the absence of effective antiviral prophylaxis, viral ulcerations are the most common cause of intestinal bleeding. HSV-associated esophageal ulcers may present with severe bleeding even in the absence of eso­ phageal symptoms. CMV can lead to ulceration throughout the entire intestinal tract. Although CMV esophageal ulcers are usually shallow (see Fig. 34-1A), ulcers elsewhere can be deep, erode into vessels, and lead to severe bleeding. CMV can also cause diffuse inflammation similar to that seen in idiopathic inflammatory bowel disease (see Fig. 34-1C and D). VZV and EBV are much less often associated with gastrointestinal (GI) bleeding. Although EBV itself does not cause mucosal ulceration, EBV-LPD can form mucosal tumors that can ulcerate and bleed (see Fig. 34-2). Massive bleeding has been reported in the setting of invasive fungal infection.

Gastrointestinal Malignancy

Post-transplant lymphoproliferative disorders (PTLDs), lymphoid proliferations, or lymphomas associated with EBV infection (EBV-LPD) occur in 1% to 20% of transplant recipients.84 Although most PTLDs are of B cell origin, T

cell lymphoma has been reported. EBV reactivation generally presents in the early post-transplant setting as a mononucleosis-like syndrome with diffuse adenopathy and fever; detection of EBV DNA in the bloodstream may allow preemptive therapy, with lower doses of immune suppression or treatment with rituximab.85 PTLD manifesting later than a year after transplant is more insidious, often pre­ senting with extranodal disease or visceral involvement. Gastrointestinal PTLD can present with diarrhea, intestinal obstruction (see Fig. 34-2B), bleeding, or perforation. Mucosa-associated lymphoid tissue-type (MALT) lymphomas have also been reported in the post-transplant setting.86 Fortunately, they often respond to reduction in immunosuppression, antibiotics (if associated with H. pylori), surgery, or chemotherapy. The risk of cancer in long-lived transplant recipients is higher than in the general population, particularly for lymphomas, skin cancers, colorectal and anal cancers, and Kaposi’s sarcoma.87 Patients who underwent liver transplant for cirrhosis secondary to primary sclerosing cholangitis are at high risk for the development of colonic dysplasia and diffuse colon cancer related to underlying ulcerative colitis.88 If severe colonic dysplasia is discovered, colectomy can be performed safely as early as 10 to 12 weeks following transplant.

Hepatobiliary Complications

Drug-induced hepatotoxicity can be problematic after a transplant because this diagnosis is often one of exclusion. Azathioprine hepatotoxicity presents as an elevation in serum aminotransferases in up to 10% of recipients; injury is generally cholestatic, with centrilobular hepatocyte damage. Azathioprine is being used less often following organ transplantation. A less common presentation is the slow insidious development of sinusoidal obstruction syndrome (veno-occlusive disease), which often manifests as portal hypertension, usually regressing following withdrawal of the drug. Cyclosporine- or tacrolimus-induced cholestasis can occur when blood levels are high. Sirolimus has been reported to cause dose-dependent elevations in serum aminotransferases. Transplant recipients are exposed to numerous other pharmacologic agents that alone, or in combination, can produce cholestasis, fatty liver, hepatitis, or a mixed histologic picture. Bacterial sepsis can have profound effects on liver function, with severe cholestasis the most common finding (a syndrome called cholangitis lenta).89 CMV infection may lead to elevations in hepatic enzymes with either a cholestatic or hepatocellular picture. CMV hepatitis is more frequent and severe in liver transplant recipients, compared with recipients of other organs.6 VZV and HSV infection can lead to hepatitis and fulminant liver failure.26 EBV hepatitis is seen in 2% to 3% of patients after SOT but is generally mild. Primary or recurrent disease with either HCV or HBV can lead to liver disease in the post-transplant setting. These viruses may be transmitted to the recipient by any solid organ from the donor. Immunosuppression leads to a marked increase in HCV titers and in some cases, to aggressive hepatic disease post-transplant, with progression to cirrhosis within 3 to 10 years.90 The results of treatment of hepatitis C in the post-transplant setting with interferon-α (INF-α)–based regimens are disappointing. Sustained virologic clearance can be achieved in 10% to 30% of post-LT recipients, but use of interferon-based therapies is limited because of side effects.91 HCV can be successfully treated in renal transplant recipients, but because the rate of renal graft failure related to INF-α is unacceptable, treatment

543

544

Section IV  Topics Involving Multiple Organs should not be attempted. It is unknown whether interferon alpha therapy for HCV in HT or LT recipients carries an increased risk of graft failure and its use is not recommended.92 Chronic HBV carriers (hepatitis B surface antigen-positive recipients) may develop a hepatitis flare following transplant, but disease often responds to antiviral agents. Vascular injury associated with liver transplantation may lead to liver dysfunction. Nodular regenerative hyperplasia (with subsequent portal hypertension) and peliosis hepatis have both been reported following renal transplantation. Organ transplant recipients, particularly following LT, are at high risk for biliary tract disease. Presentation includes acalculous cholecystitis, gallbladder sludge, thickened gallbladder wall, dilated bile ducts, or cholelithiasis.80 Gallbladder and biliary disease necessitating cholecystectomy occurs in about 1% to 6% of transplant recipients. Emergent cholecystectomy in the post-transplant setting carries a high mortality (29%).80 However, pretransplant screening for gallstones and prophylactic cholecystectomy remain controversial. The etiology of biliary tract disease is multifactorial, including obesity, use of total parenteral nutrition, fasting, biliary strictures, and medications. Cyclosporine is excreted in the bile, where it may precipitate and has been implicated in an increased incidence of cholelithiasis and cholangitis.27,93 Some centers recommend that biliary calculi be removed prior to transplantation or immediately on discovery after transplantation, but this recommendation is not universal. Patients who have undergone liver transplantation for hepatocellular carcinoma are at risk for tumor recurrence in the graft, particularly if the lesions were multiple or large prior to transplant. PTLD may also involve the liver.

COMPLICATIONS OF HEMATOPOIETIC CELL TRANSPLANTATION Hematopoietic cell transplantation (HCT) uses one of three sources of hematopoietic and immune cells: bone marrow, peripheral blood stem cell, or cord blood.94 Transplanted cells can be one’s own (autologous transplant), from an identical twin (syngeneic transplant), or from another person (allogeneic transplant). Allogeneic cells can come from a sibling who is human-leukocyte-antigen (HLA) matched with the recipient, or from another family member, or from an HLA-matched unrelated donor, or from an HLAmismatched unrelated donor (as with cord blood donors). HCT differs from solid organ transplant in three important ways: (1) the indication for HCT often involves a potentially fatal malignancy or inborn error of metabolism; (2) prep­ aration for HCT requires either high-dose myeloablative therapy or intense immune suppression, resulting in extreme susceptibility to infection and, with some preparative regimens, organ damage; and (3) recipients of allogeneic donor cells commonly develop acute and chronic GVHD. HCT patients face combined morbidity from the toxicity of chemotherapy drugs, infections, acute and chronic GVHD, and recurrent malignancy.94

EVALUATION OF GASTROINTESTINAL AND LIVER PROBLEMS BEFORE TRANSPLANTATION Ulcers and Tumors in the Gastrointestinal Tract

Mucosal ulcerations may bleed profusely when platelet counts drop after HCT, and in immunocompromised patients, ulcers may have an infectious etiology (e.g., CMV,

HSV or fungal infection) that requires specific antimicrobial treatment.95 Intestinal ulcerations should be healed before the start of conditioning therapy. CMV, Entamoeba histolytica and C. difficile are causes of colonic ulceration that may mimic idiopathic inflammatory bowel disease. Selected patients with ulcerative colitis and Crohn’s disease have undergone allogeneic and autologous HCT without complications of bleeding, perforation, or dissemination of microorganisms.96,97 The presence of fecal occult blood should prompt colonoscopy and upper endoscopy before HCT, especially in patients older than 50 years. Endoscopic biopsy may be required for staging some forms of lymphoma with a predilection for gut involvement, such as mantle cell lymphoma.

Diarrhea

Patients with diarrhea should be investigated for organisms that may cause morbidity during the period of immuno­ suppression after HCT (e.g., E. histolytica, strongyloides, G. lamblia, cryptosporidia, clostridial infections, CMV, rotavirus, adenovirus).98 Cryptosporidiosis may be resistant to therapy in an immunosuppressed patient,99 but restoration of normal immunity after allogeneic HCT can result in clearance of Cryptosporidia.100 Similarly, protracted diarrhea related to immune dysregulation can be treated with allogeneic HCT, for example, by restoration of T regulatory cells in children with immundysregulation polyendocrinopathy enteropathy X-linked (IPEX) syndrome.101 Typhlitis is a syndrome of cecal edema, mucosal friability, and ulceration in neutropenic patients, often associated with polymicrobial sepsis; its cause is usually an intestinal clostridial infection, particularly with Clostridium septicum.102 After treatment, the risk of post-HCT typhlitis is no different than that of other patients.

Perianal Pain

Pain near the anal canal in a granulocytopenic patient is assumed due to bacterial infection until proved otherwise. Extensive supralevator and intersphincteric abscesses may be present without being apparent on external examination. CMV infection of anal glands may lead to fistula formation. Perineal HSV infection causes painful skin ulcerations. Perianal infections must be treated before HCT.

Fungal Liver Infections

Diagnosis depends on liver imaging (using high-resolution CT or magnetic resonance imaging [MRI]) in conjunction with circulating fungal biomarkers (galactomannan and glucan assays),103,104 polymerase chain reaction (PCR), or culture of liver biopsy material. Therapy with newer antifungal drugs (caspofungin or azole drugs) should be continued through HCT until engraftment is established, which can effect resolution of intractable fungal liver abscesses.105-107

Viral Hepatitis in Allogeneic Hematopoietic Cell Transplant Donors

Donors who are viremic with HBV or HCV will transmit virus to their recipients.108 When two equally HLA-matched donors are available, the uninfected donor is preferred. If the more suitable donor has chronic hepatitis B, it may be possible to prevent passage of virus by treating that donor.109,110 HBV persisting in donor peripheral blood stem cells may have to be eliminated to prevent passage.110,111 HBsAg negative/anti-HBc–positive donors can be used if their serum and peripheral blood stem cells are HBV DNA negative. A donor who is naturally anti-HBs positive may be the preferred donor if the recipient is HBsAg positive or

Chapter 34  Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation anti-HBc positive because adoptive transfer of immunity can lead to clearance of virus.112 If a donor is infected by HCV and if time permits, treatment of the donor prior to harvest of donor cells may render them nonviremic, and much less likely to transmit infection.113 If HCV is transmitted, the acute phase of HCV infection may cause elevated liver enzymes at two to three months post HCT, after recovery of T cell function; after 10 years the outcome is no different than in transplant recipients without hepatitis C infection.114 In the long term, HCV-infected transplant recipients are at risk for development of cirrhosis and hepatocellular carcinoma.115

Chronic Liver Disease in Candidates for Hematopoietic Cell Transplantation

The risks faced by patients with fibroinflammatory liver disease include fatal sinusoidal obstruction syndrome following some myeloablative regimens and fulminant hepatitis B. In the absence of antiviral prophylaxis, fatal fulminant hepatitis B develops in approximately 15% of hepatitis B-infected HCT recipients.108 There is a 35% risk of postHCT reactivation of HBV in patients with isolated anti-HBc antibodies, usually during treatment for acute GVHD.116 Severe hepatitis B has been seen in anti–HBc/anti–HBspositive patients and in a patient with occult hepatitis B.117 Prior to HCT, liver biopsy should be considered if there is a clinical suspicion of cirrhosis or extensive fibrosis, as these are relative contraindications to transplant even with reduced intensity conditioning regimens.118 A reduced intensity conditioning regimen may allow congenitally immunodeficient children with chronic liver disease to be transplanted successfully, with resolution of liver disease.119

Recent Liver Dysfunction in Candidates for Hematopoietic Cell Transplantation

Patients who come to HCT following recent chemotherapy or radiation therapy that damaged the liver may be at significant risk from additional liver insults after HCT.120 These patients must be carefully evaluated with regard to the risk posed by liver-toxic conditioning regimens, particularly those containing cyclophosphamide (CY) or total body irradiation.121-123 Imatinib (Gleevec) and gemtuzumab ozogamicin (Mylotarg) deserve special mention in this context. Iminatib mesylate (and similar drugs) may cause acute hepatocellular necrosis and multiacinar collapse, with eventual healing by focal fibrosis124,125; we have successfully given CY-based myeloablative conditioning to patients who recovered from the acute injury but who had patchy fibrosis on biopsy. Gemtuzumab ozogamicin causes sinusoidal liver injury in 3% to 15% of patients126 and is a risk factor for fatal sinusoidal obstruction syndrome (SOS) if given in proximity to a liver-toxic myeloablative regimen.126,127

Gallbladder and Bile Duct Stones

HCT candidates with asymptomatic gallstones (incidentally discovered during CT or ultrasonography) do not require operative intervention.128 Patients with symptomatic cholelithiasis or stones in the common duct are at risk for biliary sepsis after HCT.

Iron Overload

HCT candidates with diseases such as thalassemia, aplastic anemia, and chronic leukemia or lymphoma may come to HCT with marked hepatic siderosis. The amount of liver iron can be accurately determined by iron-specific MRI (FerriScan or T2*).129 In patients with extreme iron over-

load, effective pre-HCT iron chelation therapy improves post-HCT survival.130 Although some studies suggest an association between excess tissue iron stores and regimenrelated toxicity, others have failed to demonstrate this. In most patients the quantitation of tissue iron stores and a decision about iron mobilization can be deferred until after recovery from HCT.

PROBLEMS FROM TRANSPLANT THROUGH DAY 200 Anorexia, Nausea, and Vomiting

Myeloablative conditioning therapy makes most patients nauseated and anorexic,131 findings associated with delayed gastric emptying.132 Serotonin-antagonist drugs are very effective in relieving symptoms during chemotherapy. Mucositis caused by myeloablative conditioning therapy may lead to oral mucosal swelling, pain, and in severe cases sloughing of pharyngeal and esophageal epithelium, intense gagging, an inability to swallow, vomiting, retrosternal pain, and airway obstruction. Opioid therapy is effective in relieving pain but can lead to gastric stasis and intestinal ileus with worse anorexia and vomiting. Methylnaltrexone, a peripheral mu-opioid receptor antagonist, can block gut opioid symptoms while allowing pain relief.77 Appetite and food intake may remain poor for up to three weeks after myeloablative therapy, an effect that is mediated by cytokines that affect appetite (interleukin [IL]-2, IL-6, tumor necrosis factor-α [TNF-α]).131 Some regimens, notably those that contain very high-dose melphalan or multiple alkylating agents, may cause unusually severe intestinal mucosal necrosis and anorexia. An early sign of acute GVHD is loss of appetite, often followed by nausea and vomiting. Before day 20, these symptoms overlap with effects of conditioning therapy. After day 20, more than 80% of patients with intractable anorexia, nausea, or vomiting will have gastric and duodenal GVHD as the sole explanation.133 Endoscopy shows edema of the gastric antral and duodenal mucosa, patchy erythema, and bilious gastric fluid, and histology demonstrates epithelial cell apoptosis and drop-out, often with localized lymphocytic infiltrates (Fig. 34-4).134 However, there is a false-negative rate with histology because of the patchiness of GVHD lesions. Immunosuppressive therapy using a 10-day course of prednisone 1 mg/kg/day plus oral beclomethasone dipropionate 8 mg/day is an effective therapy that avoids prolonged systemic immunosuppression and results in better outcomes.135,136 Recipients of auto­ logous grafts may also develop a syndrome of anorexia, nausea, and vomiting that is associated with diffuse gastric edema and erythema.137 Gastric histology shows typical GVHD. Symptoms respond to a 10-day course of prednisone 1 mg/kg/day. Endoscopy also serves to rule out intestinal infection with herpesviruses, bacteria, and fungi. CMV infection of the esophagus and upper intestine accounted for a third of patients with unexplained nausea and vomiting during the pre-ganciclovir era. These infections are usually diagnosed between days 50 and 150,138 but can present earlier if patients have activated CMV before transplant.95 Gut CMV infections can be seen in the absence of evidence of CMV in the bloodstream. HSV esophagitis may present similarly in patients not receiving prophylactic acyclovir. Intestinal CMV and HSV infections are now rare. Fungal esophagitis may cause anorexia but not the incessant vomiting often seen with GVHD or herpesvirus infections. Studies of gastric emptying and myeloelectric activity in HCT patients have shown that symptoms of nausea and

545

546

Section IV  Topics Involving Multiple Organs

A

B

C

D

E

F

Figure 34-4.  Endoscopic and histologic manifestations of acute graft-versus-host disease (GVHD) of the gastrointestinal tract. A, Esophagus: desquamation of squamous epithelium of the distal esophagus in severe GVHD. B, Stomach: diffuse mucosal edema and erythema in the gastric antrum in moderately severe GVHD. C, Small intestine: mucosal edema, focal bleeding, and ulceration in severe GVHD. D, Rectum: focal apoptosis (arrow) caused by GVHD in crypt epithelium (Hematoxylin and eosin, alcian blue). E, Colon: the bottom of a colon crypt (arrows) is missing epithelial cells; apoptotic debris is admixed with mucus at the base of the crypt (Hematoxylin and eosin, alcian blue). F, Small intestine: gross and histologic findings in fatal GVHD. Autopsy photograph of a small intestinal segment opened to reveal sloughing of the mucosa (top); histology shows complete absence of epithelial cells, a lymphoid infiltrate, and submucosal edema (bottom) (Hematoxylin and eosin).

vomiting are frequently accompanied by retention of radionuclide meals and disordered electrical activity.139 Pro­ motility agents such as metoclopramide, domperidone, and low-dose erythromycin are occasionally useful, but in patients with persistent symptoms, endoscopic evaluation should precede empirical promotility therapy. Anorexia and vomiting may also be manifestations of central nervous system (CNS) disease; other neurologic signs and symptoms usually dominate the clinical picture in patients with CNS disorders. In patients with a history of glucocorticoid exposure, adrenal insufficiency may also cause upper gut symptoms. Oral medications such as calcineurin inhibitors, MMF, trimethoprim-sulfamethoxazole, itraconazole, posaconazole, voriconazole, imatinib and similar drugs, and highdose opioids also cause nausea and vomiting.106,107 Mycophenolic acid causes fewer GI symptoms than MMF.64 Parenteral infusions of fat, glucose, and amino acids reduce food intake, slow gastric emptying, and cause nausea. Even after total parenteral nutrition has been stopped, appetite suppression may linger for one to three weeks.140

Jaundice, Hepatomegaly, and Abnormal Liver Biochemical Tests

Development of jaundice following HCT is an ominous prognostic sign, with increased nonrelapse mortality in patients whose total serum bilirubin exceeds 4 mg/dL.141 Fortunately the frequency of severe liver injury after HCT

is much lower than it was 10 years ago because of less frequent use of drugs that cause sinusoidal injury and because of prophylaxis of cholestatic injury with ursodiol.142-144 There are multiple causes of jaundice after HCT (Table 34-2). Sinusoidal Obstruction Syndrome (Veno-occlusive Disease) Some myeloablative conditioning regimens may damage hepatic sinusoids, leading to hepatomegaly, fluid retention and weight gain, and elevated serum bilirubin.145 Individual variability in cyclophosphamide metabolism, irradiation dose, use of gemtuzumab ozogamicin, and preexisting liver inflammation and fibrosis are risk factors.121,123,126 Reduced intensity regimens rarely affect hepatic sinusoids.118 A clinical diagnosis of SOS may suffice if typical signs develop before day 20 post transplant, but Doppler ultrasonography, measurement of the wedged hepatic venous pressure gradient, and liver histology may be needed in difficult cases.146-148 Initial histologic changes of SOS are dilation of sinusoids, extravasation of red cells through the space of Disse, necrosis of periventricular hepatocytes, and widening of the subendothelial zone in central veins (Fig. 34-5A).145 The later stages are characterized by extensive collagenization of sinusoids and venules (see Fig. 34-5B). More than 80% of patients with SOS recover completely. A poor prognosis correlates with the degree of bilirubin elevation and weight gain, higher serum aminotransferase enzyme (AST and ALT) values, higher wedged hepatic venous pressure

Table 34-2  Liver Diseases after Hematopoietic Cell Transplantation DISEASE

FREQUENCY

TIMING

DIAGNOSIS

TREATMENT

PREVENTION

SOS (veno-occlusive disease)

0%-20% (regimen dependent)

Onset before day 20

Typical clinical features Imaging WHVPG, histology (Figure 34-5) Note atypical presentations (acute hepatitis, anasarca)

None proven. Defibrotide successful in ≈40% of patients with severe SOS

Assess patient risk Choose non–liver toxic agents for conditioning therapy

Cholestasis of sepsis (cholangitis lenta)

Common in neutropenic patients

Following sepsis or neutropenic fever (usually before day 30)

Exclude other causes of cholestasis Inferential diagnosis

Treat underlying infection

Infection prophylaxis or expectant treatment Ursodiol

Acute GVHD

≈20% of allograft recipients Rare after autograft

Days 15-50

Confirm GVHD in skin, intestine Exclude other causes of cholestasis Histology (Figure 34-5C)

Glucocorticoids (2 mg/kg/day of methyl prednisolone or prednisone) Ursodiol

Optimal donor selection Complete GVHD prophylaxis T cell depletion protocols Ursodiol

Acute viral hepatitis

Uncommon when prophylaxis is used against herpes viruses, HBV

HSV, days 20-50 Adenovirus, days 30-80 VZV, days 80-250 HBV and HCV, during immune reconstitution

Pretransplant serology and PCR results Isolation of virus from other sites (stool and urine for adenovirus) PCR of serum for specific viruses Liver histology, PCR, immunostains (Figure 34-5)

HSV, VZV: acyclovir Adenovirus: cidofovir HBV: nucleos(t)ide analogs (see Chapter 78)

HSV and VZV infection: acyclovir prophylaxis for all patients If patient is at risk for HBV infection: lamivudine or other nucleos(t)ide analog, choose HBV immune donor

Fungal abscess

Rare when prophylaxis is used

Day 10-60

Hepatic pain, fever Liver imaging Serum fungal antigen

Antifungal drugs (vary with organism)

Pretransplant screening Fluconazole prophylaxis for all patients

Drug-induced liver injury

Common

Days 0-100

Clinical evidence

Discontinue drug

None

Ischemic liver disease

Confined to patients with septic or hemorrhagic shock or respiratory failure

Days 0-30

Clinical evidence

Restore cardiac output

Early treatment of sepsis, bleeding

Biliary obstruction

Transient biliary sludge common Stones and chloromas rare

Days 15-60

History, examination Biliary ultrasonography

Ursodiol to increase bile-salt dependent bile flow Papillotomy ± stent if obstruction persists

None

Idiopathic hyperammonemia

Rare

Days 10-50

Venous blood ammonia

None proven

Unknown

Chronic hepatitis C

Formerly common

After day 80

HCV RNA in serum Elevations of serum AST, ALT after immune reconstitution

Peg interferon-α plus ribavirin after full immune reconstitution

Screen hematopoietic cell donors

Iron overload

Very common

Pretransplant Long-term follow-up after transplant

Transferrin saturation Bone marrow iron quantitation Liver iron quantitation Magnetic resonance imaging

May not be necessary Phlebotomy, chelation if iron burden is very high (see text)

Avoid medicinal iron supplements

Chronic GVHD

Common after allografts

After day 80

Prior acute GVHD history Chronic GVHD in other organs Consistent serum ALT, alkaline phosphatase levels Histology (Figure 34-5D)

Immunosuppressive drug therapy Ursodiol

Screening for chronic GVHD at day 80

ALT, alanine aminotransferase; AST, aspartate aminotransferase; GVHD, graft-versus-host disease; HBV, hepatitis B virus; HCV, hepatitis C virus; HSV, herpes simplex virus; PCR, polymerase chain reaction; RNA, ribonucleic acid; SOS, sinusoidal obstruction syndrome (veno-occlusive disease); VZV, varicella-zoster virus; WHVPG, wedged hepatic venous pressure gradient.

548

Section IV  Topics Involving Multiple Organs

A

B

C

D

E

F

Figure 34-5.  Histology of liver diseases following hematopoietic cell transplantation. A, Sinusoidal obstruction syndrome (SOS; also known as venoocclusive disease of the liver) at 23 days post-transplant: high-power view of hemorrhage into the space of Disse, hepatocyte necrosis, disrupted sinusoids, and subendothelial edema of the central vein (arrow). The lumen of the vein is patent (Hematoxylin and eosin). B, SOS caused by gemtuzumab ozogamicin: high-power view of zone 3 of the acinus showing extensive sinusoidal fibrosis, hepatocyte dropout, and a patent central vein (Masson trichrome stain). C, Acute graft-versus-host disease (GVHD) at day 82 post-transplant: portal area containing abnormal small bile ducts (arrows) with epithelial cell dropout, cytoplasmic eosinophilia, and vacuolization (Hematoxylin and eosin). D, Chronic GVHD at day 184 post-transplant: high-power view of a portal area, with damaged small bile ducts (arrows) infiltrated by lymphocytes (Hematoxylin and eosin). E, Varicella-zoster virus hepatitis: Low-power view of confluent necrosis of hepatocytes (arrows, pointing to pale area) adjacent to hepatocytes that are normal (periodic acid–Schiff). F, Adenovirus hepatitis: low-power view of a focal area of confluent hepatocyte necrosis (arrows, pointing to basophilic area) with remnants of hepatocytes that contain intranuclear inclusions typical of adenovirus (“smudge cells”), best seen at higher power (Hematoxylin and eosin).

gradient, development of portal vein thrombosis, and multiorgan failure. Treatment of severe SOS is unsatisfactory; the best results are with intravenous defibrotide (25 mg/kg/ day), a porcine oligonucleotide that has effects on microvascular endothelial cells.127 SOS can be prevented by identifying patients at risk and altering the transplant regimen using (1) conventional, not transplant therapy; (2) a reducedintensity conditioning regimen118; (3) a myeloablative regimen that does not contain cyclophosphamide (e.g., targeted busulfan-fludarabine for allogeneic HCT149,150 or BEAM (BCNU [carmustine], etoposide, arabinoside [cytarabine], melphalan) for autologous HCT151); (4) modification of CY-based regimens.152,153 Patients with cirrhosis may decompensate even after a reduced-intensity allograft.118 Cholestatic Liver Diseases Cyclosporine inhibits canalicular bile transport and commonly causes mild increases in serum bilirubin. Sepsisassociated cholestasis is an important contributor to hyperbilirubinemia in the weeks after HCT, mediated by endotoxins and cytokines such as IL-6 and TNF-α.89 Many drugs used after HCT have been associated with cholestatic liver disease. Acute GVHD is the most common cause of severe cholestatic injury, related initially to cytokines such as IL-6 and later to destruction of small bile ducts caused by alloreactive T cells.154 Hepatic GVHD usually follows cutaneous and/or intestinal GVHD, and is heralded by a

gradual rise in serum bilirubin, alkaline phosphatase, and aminotransferases. In allograft recipients on minimal immunosuppression or after donor lymphocyte infusion, GVHD may present as acute hepatitis.155,156 A cholestatic condition identical to GVHD occurs rarely in autologous HCT reci­pients. Characteristic liver biopsy findings in GVHD include lymphocytic infiltration of small bile ducts with nuclear pleomorphism and epithelial cell dropout (see Fig. 34-5C and D). Because these patients are frequently pancytopenic, inflammatory infiltrates may be minimal. In advanced cases of hepatic GVHD, it may be difficult to identify small bile ducts because they have been destroyed. Only 30% of patients with liver GVHD have resolution of liver abnormalities after initial immunosuppressive treatment. Pro­phylactic ursodiol reduces the frequency of cholestasis in general and GVHD-related cholestasis specifically, compared with placebo, and should be given routinely through day 80 in allograft recipients.142 More than 50% of patients with acute liver GVHD will develop chronic GVHD. Acute Hepatocellular Injury Severe hepatic injury (serum ALT >1500 U/L) is now mostly related to zone three necrosis from circulatory causes such as SOS and hypoxic hepatitis, and not to infection. However, acute hepatitis caused by HSV, VZV, adenovirus or HBV can be fatal after HCT.108,157,158 Except for sporadic cases of

Chapter 34  Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation adenovirus hepatitis, infections by these other viruses have become rare because of antiviral prophylaxis and preemptive treatment. Hepatic infections caused by CMV and HCV are seldom severe.114 Despite acyclovir, HHV-6 and HHV-8 reactivation have been associated with the development of fever, rash, and hepatitis in HCT recipients.159 Other non­ infectious causes of acute hepatocellular injury include a hepatitic presentation of GVHD and drug toxicity.155,156 When there is uncertainty about the cause of rising serum ALT, DNA blood tests for herpesviruses, adenovirus, and HBV; transvenous measurement of the wedged hepatic venous pressure gradient; and liver biopsy are indicated (see Fig. 34-5). If acyclovir is not being given, it should be started empirically, particularly if the patient presents with abdominal complaints typical of VZV infection (see Fig. 34-5E).160 Adenovirus hepatitis should be suspected if the patient has concomitant pulmonary, renal, bladder, or intestinal symptoms (see Fig. 34-5F); the most effective treatment is cidofovir when given early in the course of infection.161,162 Fulminant hepatitis B may develop during immune reconstitution in patients at risk, but can be prevented with prophylactic lamivudine or adefovir.108,163 If severe hepatitis B reactivation does occur, usually because a diagnosis of HBV was not made prior to HCT,117 antiviral therapy with the most potent antiviral drug available should be initiated immediately, although progression to fatal liver failure is not uncommon.164 Fulminant hepatitis B has also occurred after discontinuation of prophylactic antiviral therapy; all HBV-infected patients, particularly those with high pretransplant HBV DNA levels, should be monitored following antiviral drug withdrawal.165,166 HCV infections are seldom severe; asymptomatic elevations of serum ALT are commonly seen from days 60 to 120, frequently coinciding with the tapering of immunosuppressive drugs.114 Therapy directed at chronic HCV infection should be considered once the patient has ceased all immunosuppressive drugs and has no evidence of active GVHD.167 About a third of HCV-infected transplant survivors will develop cirrhosis over 20 to 30 years. Fungal and Bacterial Infections Antifungal prophylaxis has significantly reduced the incidence of invasive fungal disease in HCT recipients, particularly in those requiring ongoing immunosuppression for treatment of GVHD.106,168,169 If invasive fungal disease does occur, infection with resistant Candida species or molds is likely. Signs are fever and tender hepatomegaly, with increased serum alkaline phosphatase levels. Highresolution CT or MRI may demonstrate multiple fungal abscesses, and serological tests for fungal antigens may be useful for diagnosis. Antifungal drugs and return of neu­ trophil function after HCT can lead to resolution of previously treatment-refractory mold infection.105 Bacterial liver abscesses are rare in HCT recipients, probably because of the high use of systemic antibiotics; however, latent mycobacterial infection may reactivate within the liver with prolonged immunosuppressive therapy. Disseminated bacille Calmette-Guérin (BCG) infection with liver involvement has been reported. Disseminated clostridial infection and gallbladder infection with gas-producing organisms may lead to air in the liver and biliary system. Gallbladder and Biliary Disease Gallbladder sludge (calcium bilirubinate) is almost universally present in HCT patients. Although sludge is usually asymptomatic, passage through the bile duct may cause epigastric pain, nausea, and elevated serum liver enzymes. Endoscopic papillotomy is rarely indicated. Biliary sludge

may be a cause of acute “acalculous” cholecystitis, acute pancreatitis, and bacterial cholangitis.170,171 Acute cholecystitis is uncommonly seen in HCT recipients and is frequently acalculous. Cholecystitis in this setting may also be due to leukemic relapse with gallbladder involvement or infection by CMV or fungi. Diagnosis is difficult because of the high frequency of gallbladder abnormalities on ultrasonography following HCT. Pericholecystic fluid, gall­ bladder wall necrosis, or localized tenderness suggests cholecystitis. A radionuclide bile excretion study, with morphine infusion to enhance gallbladder filling, can be useful; nonvisualization of the gallbladder suggests cholecystitis. Biliary obstruction is a rare event, caused by a variety of disorders (e.g., lymphoblastic infiltration of the bile duct and gallbladder in EBV-LPD; CMV-related biliary disease; dissecting duodenal hematoma complicating endoscopic biopsy; inspissated biliary sludge; and leukemic relapse [chloroma] in the head of the pancreas).170,171 Therapeutic endoscopic cholangiopancreatog­ raphy may be needed if there is cholangitis or per­sistent obstruction. Malignant Disorders EBV-LPD was commonly seen two to four months after HCT, particularly in recipients of HLA-mismatched T cell– depleted grafts and after potent anti–T cell therapies. Liver involvement occurred in more than 50%, manifested by abnormal serum alkaline phosphatase and massive hepa­ tosplenomegaly. EBV-LPD is now infrequent because of EBV-DNA surveillance and preemptive treatment with rituximab. Idiopathic Hyperammonemia and Coma A syndrome of hyperammonemia and coma has been described in patients who received high dose chemotherapy, including conditioning for HCT.172 Patients present with progressive lethargy, confusion, weakness, incoordination, vomiting and hyperventilation. The diagnosis is confirmed when the plasma ammonia exceeds 200 µmol/L and there is no evidence of liver failure. This syndrome is rare, but is associated with a high mortality. Its pathogenesis involves the unmasking of a latent genetic disorder similar to ornithine transcarbamylase deficiency.173,174

Gastrointestinal Bleeding

Bleeding that does not require transfusion is very common after HCT, particularly when platelet counts are low. Causes include retching-induced trauma to the esophageal or gastric mucosa (Fig. 34-6A), mucosal injury from conditioning therapy, peptic esophagitis, C. difficile colitis, anal fissures and hemorrhoids, and mild acute GVHD. The incidence of severe GI bleeding, particularly in patients with adequate platelet counts, is less than 1% because of effective prophylaxis against viruses, fungi, and acute GVHD.175 Mortality from severe intestinal bleeding, however, remains at 40%.175,176 The most common cause of severe bleeding is refractory acute GVHD, which can result in bleeding from extensive ulceration in the small intestine and cecum (see Fig. 34-4). In some patients with GVHD, bleeding may appear to be coming from specific areas of the mucosa, but when such patients are operated on or come to autopsy, diffuse rather than focal mucosal ulceration is the rule. Ulcers in the stomach or duodenum that develop after HCT are usually caused by acute GVHD or CMV infection, but with preemptive ganciclovir therapy, bleeding CMV ulcers have become rare.175 Gastric ulcerations also may be caused by infection by VZV, bacteria (phlegmonous gastritis), or EBV (lymphoproliferative disease) (see Chapter 51).

549

550

Section IV  Topics Involving Multiple Organs

A

B

C

D

E

Figure 34-6.  Uncommon gastrointestinal problems in hematopoietic cell transplant recipients. A, Esophagus: barium contrast radiograph of an intramural hematoma that occupies one wall of the esophagus from the aortic arch to the lower esophagus. The red line approximates the normal esophageal contour. This hematoma was caused by retching in a thrombocytopenic patient. B, Stomach: diffuse oozing of blood through the mucosa of the gastric antrum in gastric antral vascular ectasia (GAVE). When the blood was washed off, the underlying mucosa was not ulcerated, yet blood reappeared. C, High-power view of an antral biopsy specimen illustrating the features of GAVE: capillary dilation, thrombosis, and fibromuscular hyperplasia (Hematoxylin and eosin). D, Duodenum: linear ulceration with yellow exudates (arrows) caused by Rhizopus infection in a transplant patient receiving immunosuppressive therapy for graft-versus-host disease (GVHD). The surrounding mucosa is abnormal because of GVHD. E, Colon: sigmoid colon in adenovirus colitis showing diffuse mucosal edema, ulceration, and hemorrhage.

Gastric antral vascular ectasia, is also a cause of severe upper intestinal bleeding in HCT recipients who received oral busulfan as part of conditioning therapy.177,178 Diffuse areas of hemorrhage are seen in the gastric antrum and proximal duodenum, but the underlying mucosa is intact (see Fig. 34-6B). Histology is diagnostic, revealing abnormal dilated capillaries, thromboses, and fibromuscular hyperplasia in the lamina propria (see Fig. 34-6C). Endoscopic laser therapy is the treatment of choice to control bleeding, but multiple laser treatments may be required to obliterate ectatic lesions.178 Rare gastroduodenal causes of bleeding post HCT include ulcers caused by molds (see Fig. 34-6D), Dieulafoy lesions, Curling (stress) ulcers, duodenal biopsy sites, adenovirus colitis (see Fig. 34-6E), and C. septicum infection (typhlitis).179 There is no effective therapy for mucosa that is diffusely oozing blood other than raising the platelet count and treating the underlying condition. In GVHD, re-epithelialization of ulcerated intestinal mucosa occurs very slowly (see Fig. 34-4F). Focal bleeding lesions, especially those caused by mucosal infection, can be treated with endoscopic cautery, heater probe, or epinephrine injection provided platelet counts are adequate. Unless the underlying disease process is eliminated, these endoscopic methods will not cure the bleeding problem. Attempts to resect large segments of diffusely bleeding intestine involved with GVHD have not been successful.

Dysphagia

Mucositis, acid-peptic esophagitis, and pill esophagitis are the leading causes of dysphagia. Infections of the esophagus have largely disappeared because of antiviral and antifungal prophylaxis. Desquamation of oropharyngeal epithelium caused by conditioning therapy may lead to pain on initiating a swallow and inability to move a bolus past the cricopharyngeus. Rarely, non-healing esophageal ulcerations, strictures, and dysphagia result from conditioning therapy. The abrupt onset of severe retrosternal pain, hematemesis, and painful swallowing suggests a hematoma in the wall of the esophagus, a result of retching when platelet counts are very low (see Fig. 34-6A).180 Endoscopy is relatively contraindicated because many intramural hematomas represent contained perforations. The course of intramural hematomas is one of slow resolution over one or two weeks. In patients with severe GVHD, esophageal edema, erythema, and a peeling epithelium lead to ulcerations (see Fig. 34-4A).181 Pill esophagitis occurs after ingestion of medications that might be used after HCT, such as phenytoin, foscarnet, captopril, oral bisphosphonates, ascorbic acid, ciprofloxacin, clindamycin, and oral potassium chloride.

Diarrhea (see Table 34-3) Diarrhea caused by mucosal damage from myeloablative conditioning therapy is seldom severe, usually resolving by

Chapter 34  Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation Table 34-3  Causes of Diarrhea after Hematopoietic Cell Transplantation CAUSE

FREQUENCY

DIAGNOSIS

SEVERITY

TREATMENT

Myeloablative conditioning therapy

Common

Exclude infection, hyperacute GVHD

Usually mild, can be severe after some regimens

Self-limited; octreotide useful in severe cases

Acute GVHD

Common after allografts; occurs in ≈10% of autografts

Association with skin and liver GVHD. Exclude infection. Mucosal histology in problematic cases

Ranges from mild to intractable, highvolume diarrhea

Immunosuppressive drugs, usually prednisone initially

Now uncommon

CMV antigen or DNA in blood; viral culture, immunohistology of mucosal biopsy Adenovirus DNA in blood; viral culture, immunohistology of mucosal biopsy Research test only ELISA of stool

Potentially fatal if not detected early

Ganciclovir or foscarnet

Serotype dependent; may be rapidly fatal

Cidofovir

Self-limited Serotype dependent; can be severe Usually fatal when lymphomatous intestinal involvement develops

None None

Viral infections CMV

Adenovirus

Sporadic

Astrovirus, norovirus Rotavirus

Unusual Rare

EBV-LPD

Now rare

EBV DNA in blood; mucosal biopsy

Common

Toxin and antigen in stool

Usually mild to moderate

Sporadic Rare except in endemic areas

Clinical syndrome of typhlitis Stool, blood culture

Potentially fatal Potentially fatal

Oral vancomycin preferred to metronidazole Imipenem, oral vancomycin Based on organism sensitivities

Rare Rare

Stool EIA Stool microscopy, PCR

Can be protracted Often protracted

Metronidazole Recovery of immunity

Rare

Potentially fatal

Rare

Stool microscopy, antigen, DNA; possibly serum EIA Stool microscopy

Potentially fatal

Metronidazole or tinidazole, followed by paromomycin Ivermectin

Common Common Common

Clinical diagnosis Clinical diagnosis Clinical diagnosis

Dose dependent Diet dependent Medication dependent

Reduce dose; switch to IV Mg++ Disaccharide dietary restriction Restore flora

Common Unusual Common

Clinical diagnosis Clinical diagnosis Clinical diagnosis

Dose dependent Dose dependent Dose dependent

Unusual

Clinical diagnosis

Medication dependent

Reduce dose; switch to IV Mg++ Reduce dose; loperamide Reduce dose; substitute mycophenolic acid formulation; loperamide Discontinue

Bacterial infections Clostridium difficile Clostridium septicum Enteric pathogens Parasitic infections Giardia lamblia Cryptosporidium parvum Entamoeba histolytica Stronglyoides stercoralis Osmotic diarrhea Oral magnesium salts CHO malabsorption Antibiotic use Medication related Oral magnesium salts Tacrolimus Mycophenolate mofetil Metoclopramide

Rituximab when detected early; withdrawal of immunosuppressive drugs

CHO, carbohydrate; CMV, cytomegalovirus; DNA, deoxyribonucleic acid; EBV-LPD, Epstein-Barr virus–lymphoproliferative disease; EIA, enzyme immunoassay; GVHD, graft-versus-host disease; IV, intravenous; PCR, polymerase chain reaction.

days 12 to 15. Cytarabine-containing regimens, high-dose melphalan, and multiple alkylating regimens cause more severe, protracted diarrhea. Intravenous octreotide and oral loperamide at (4 mg by mouth every six hours) may be effective for severe diarrhea associated with conditioning therapy. Acute GVHD is the most common cause of diarrhea after day +15.176,182 The onset of diarrhea can be sudden, with daily stool volumes in excess of 2 L in severe cases. The diarrheal fluid is watery and green, with ropy strands of mucoid material that reflect transmucosal protein loss. In an allografted patient with skin and liver abnormalities typical of acute GVHD, this diarrheal syndrome is almost diagnostic of intestinal GVHD, particularly when there is falling serum albumin and negative stool studies for infection. In GVHD, abdominal imaging (CT, positron-emission

tomography [PET]-CT, or focused ultrasonography) may reveal intestinal edema, but this finding does not differen­ tiate acute GVHD from CMV infection.148,183,184 Pneumatosis intestinalis, which may be associated with GVHD or CMV enteritis, may be seen by plain x-ray, CT, or MRI. A definitive diagnosis of GVHD in problematic cases requires mucosal biopsy. In mild cases, the gastroduodenal and rectosigmoid mucosa are grossly normal, but moderately severe GVHD causes diffusely edematous and erythematous mucosa (see Fig. 34-4).134 Severe GVHD may lead to ulcerations and large areas of mucosal sloughing in the stomach, small intestine and colon (see Fig. 34-4F). Even when the endoscopic appearance is normal, biopsies often reveal intestinal crypt cell necrosis and apoptotic bodies diagnostic of acute GVHD (see Fig. 34-4D). In severe cases of GVHD,

551

552

Section IV  Topics Involving Multiple Organs whole crypts are destroyed, then adjacent crypts, and finally whole segments of intestinal mucosa (see Fig. 34-4E and F). Other histologic findings that support the diagnosis of GVHD include pericapillary hemorrhage,185 infiltrating neutrophils,186 or eosinophils.187 The use of capsule endoscopy for diagnosis of GVHD can provide visual inspection of the small intestine that cannot be seen with routine endoscopy, but does not allow biopsy.188 The negative predictive value of a capsule endoscopy examination of the small intestine appears to be high and is thus useful for excluding more severe acute GVHD.188 Bleeding often accompanies diarrhea in patients with mucosal ulceration.175 Successful treatment of acute GVHD with immunosuppressive therapy results in a dramatic reduction in stool volume, with resolution of accompanying symptoms of abdominal pain, nausea, and vomiting. The management of patients whose diarrhea and other symptoms of intestinal GVHD persist after 7 to 14 days of immunosuppressive therapy is unsatisfactory because the rate of failure of secondary therapy is high.189 Overall prognosis in patients with GVHD can be estimated by use of the acute GVHD Activity Index.190 In allograft recipients, infectious causes of diarrhea are far less common than GVHD, accounting for only 10% to 15% of diarrheal episodes.176,182 In countries where intestinal parasitism and bacterial contamination of water are endemic, the spectrum of infections may be wider.98 C. difficile colitis is usually a relatively mild, treatable disease when diagnosed at the onset of diarrhea, but the recent emergence of more virulent strains of C. difficile has changed the natural history of this infection. The inappropriate use of proton pump inhibitors for marginal indications191 increases the risk of C. difficile colitis twofold.192 Other relatively common causes of infectious diarrhea include astrovirus, norovirus, rotavirus, CMV, and adenovirus.99,182,193,194 Some serotypes of adenovirus cause necrotizing enteritis and rapidly fatal multiorgan failure involving the gut, liver, lungs, and kidneys (see Fig. 34-6E).157,158,195 There should be a sense of urgency in identifying adenovirus as a cause of enteritis, as early treatment with cidofovir appears to be effective.161,162 CMV is the only cause of enteritis after HCT that requires an intestinal biopsy for diagnosis.182 Otherwise the negative predictive value of a stool examination (including PCR) for other viruses, bacteria, fungi, and parasites is high. Watery diarrhea secondary to intestinal parasite infection (Cryptosporidium, G. lamblia, and E. histolytica) and mycobacteria infection are rare outside of endemic areas.100,194,196,197 Diarrhea may also result from carbohydrate malabsorption (particularly in patients on antibiotics), oral magnesium salts, tacrolimus (a motilin agonist), metoclopramide, and MMF.198,199 MMF gut toxicity can be addressed by switching to mycophenolic acid.64

Abdominal Pain

It is extremely important to distinguish abdominal pain as an indicator of a rapidly progressive, fatal condition from illnesses with a benign natural history that require only conservative management. The causes of abdominal pain after HCT are listed in Table 34-4. The illnesses that may progress rapidly include intestinal perforation, some infections (e.g., typhlitis caused by C. septicum, adenovirus, and VZV), gallbladder necrosis, liver abscess, and acute GVHD presenting only as abdominal pain.200 Fortunately, these disorders are far less common than intestinal pseudoobstruction, hepatic pain related to SOS, multisystem acute GVHD, and hemorrhagic cystitis. Intestinal perforation may develop in the setting of lysis of a transmural lymphoma or metastatic carcinoma shortly after conditioning therapy, or later on, from CMV ulcers or diverticular perforation. Perforation may present with only mild to moderate abdominal

pain and pneumoperitoneum on plain abdominal x-ray. Dilation of the bowel in the absence of a mechanical obstruction is the most common cause of moderate to severe abdominal pain. Most patients with pseudo-obstruction have an underlying intestinal disease, such as enteritis from conditioning therapy, GVHD, or infection, but frequently the acute presentation is related to increasing use of muopioid medications. Pseudo-obstruction is more frequent among patients with lymphoma, a result of intestinal neuropathy from repeated use of vincristine (see Chapter 120). To allow pain relief without affecting colon motility, methylnaltrexone can be used.77 Alternatively, colon distention may decrease after switching from a mu-opioid to a kappaopioid agonist (for example, to butorphanol). Neostigmine (2 mg intravenously) has been successfully used in patients with acute colonic pseudo-obstruction after HCT.201 In visceral VZV infection, abdominal distention, severe pain, fever, and rising serum ALT may precede cutaneous manifestations by up to 10 days.160 In rare instances, a skin rash never develops. Acyclovir should be started on clinical suspicion while serum is analyzed by PCR for VZV DNA.160 More severe acute intestinal GVHD may present with nausea, anorexia, periumbilical crampy abdominal pain, and diarrhea. The sudden onset of intestinal edema (see Fig. 34-4B and C) can cause a rigid abdomen with rebound tenderness preceding the development of a skin rash or diarrhea. The decision to treat a patient empirically with prednisone when definitive evidence of GVHD is not at hand can be difficult, but when the pretest probability of GVHD is high (e.g., an HLA-mismatched or unrelated donor; engraftment; a nascent skin rash) and that of perforation or infection low, treatment should be started while GVHD is sought by endoscopic mucosal biopsy (see Fig. 34-4). Pancreatitis is an uncommon cause of abdominal pain in HCT patients, but in a study of autopsied patients the prevalence of acute pancreatitis was 28%.202 Symptoms of pancreatitis had been absent in many of the patients who were found to have florid pancreatitis at autopsy, suggesting that symptoms had been masked by immunosuppressive drugs. Patients with low platelet counts or prolongation of blood clotting may rarely bleed into the retroperitoneum, abdominal wall, or intra-abdominal viscera, particularly after duodenal biopsy, causing significant pain. Intestinal infections presenting with significant pain are listed in Table 34-4. Typhlitis (C. septicum infection) occurs in granulocytopenic patients but is not common after HCT. Symptoms include fever, right lower quadrant pain, nausea and vomiting, diarrhea, occult blood in stool, and shock.102 Typhlitis is usually diagnosed using imaging studies; laparotomy is rarely necessary.203 If typhlitis is a possibility, imipenem and oral vancomycin therapy should be started along with coverage for luminal bacteria and fungi.204

Perianal Pain

Perianal pain after HCT can be caused by an anal fissure, a thrombosed external hemorrhoid, cellulitis related to tissue maceration, fistulas, and abscesses. In patients with granulocytopenia, infections in the perineum or perianal spaces are usually polymicrobial, arising either from anal crypts or from tears in the anal canal. After HCT these infections can be difficult to recognize because they may not produce abscesses but rather a spreading cellulitis. Extensive supralevator and intersphincteric abscesses may be present without being apparent on external examination. CT, MRI, or endoscopic ultrasonography can give a clear view of the anatomy involved, particularly if there is pus present.205 When antibiotics covering anaerobic and aerobic bacteria are given to patients with incipient perianal infection, far fewer patients require surgical drainage than in the past.

Chapter 34  Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation Table 34-4  Causes of Acute Abdominal Pain after Hematopoietic Cell Transplantation CAUSE

FREQUENCY

DIAGNOSIS

SEVERITY

TREATMENT

Sinusoidal obstruction syndrome (SOS)

Now uncommon

Tender hepatomegaly, weight gain, jaundice

Potentially fatal

See section on SOS (see Table 34-2)

Intestinal damage from conditioning therapy

Unusual cause of pain

Examination, imaging

Can be protracted after some regimens

None

Colon pseudo-obstruction

Common, particularly in patients with prior vincristine exposure and current mu-opioid therapy. May also occur with GVHD and VZV or CMV infection.

Distention, tympany, abdominal radiographs

When medication related, usually resolves; when a sign of GVHD or viral infection, may be severe

Reduce opioid, anticholinergic drug exposure; rule out treatable underlying causes; consider methylnaltrexone or kappa agonist opioid; neostigmine if persistent and severe

Hemorrhagic cystitis

Common after cyclophosphamide and with viral bladder infection

Suprapubic pain, hematuria, viral cultures (JC/BK virus or adenovirus)

Can be protracted with viral infection

Urologic therapy, antiviral drugs if appropriate

Acute GVHD

Common, particularly with more severe GVHD

Evaluate skin, intestinal symptoms, serum bilirubin level Intestinal imaging (CT, ultrasound); mucosal biopsy

Potentially fatal

Immunosuppressive drug therapy

Biliary pain

Unusual

RUQ/epigastric localization; gallbladder sludge, edema, gas; biliary dilation on ultrasonography

Passage of sludge is usually self-limited; necrotic gallbladder requires surgery

Persistent biliary obstruction requires stent placement; surgery for gallbladder necrosis

Pancreatitis

Unusual

Serum lipase

Usually self-limited, but pancreatic necrosis may occur

Address biliary, infection, and medication causes

Hematomas

Rare; can be seen after duodenal biopsy

Examination, abdominal imaging, endoscopy

Can be protracted

Restoration of platelet counts; intestinal obstruction may require surgery

Intestinal infection

Unusual

Diagnostic and imaging tests for clostridial infection, VZV, CMV, adenovirus, molds

Potentially fatal if not treated (especially C. septicum, viral, or mold infection)

Treat organism identified

Intestinal perforation

Rare

Plain films, CT

Potentially fatal

Surgery; identification of underlying cause (CMV ulcer, intestinal tumor necrosis, diverticula)

Liver abscess/bacterial infection

Rare (usually fungal)

Liver imaging (MRI preferred), examination, serum fungal antigen detection

Potentially fatal if not treated

Appropriate antifungal, antimycobacterial therapy

Intestinal infarction

Rare (usually disseminated Aspergillus infection)

Intestinal imaging, examination, chest film, galactomannan EIA

Uniformly fatal

Antifungal drugs active against Aspergillus

EBV-lymphoproliferative disease

Rare with surveillance for EBV DNA in serum

Abdominal imaging, endoscopy

Usually fatal once tumor masses form

Rituximab when detected early; withdrawal of immunosuppressive drugs

CMV, cytomegalovirus; CT, computed tomography; DNA, deoxyribonucleic acid; EBV, Epstein-Barr virus; EBV-LPD, Epstein-Barr virus–lymphoproliferative disease; EIA, enzyme immunoassay; GVHD, graft-versus-host disease; JC/BK; polyomaviruses; MRI, magnetic resonance imaging; RUQ, right upper quadrant; VZV, varicella-zoster virus.

PROBLEMS IN LONG-TERM TRANSPLANT SURVIVORS Liver Disease Caused by Graft-Versus-Host Disease Long-term survivors of transplant who have hepatic GVHD usually have other evidence of chronic GVHD, a pleomorphic immune disorder characterized by oral and ocular sicca; ulceration in squamous epithelium of the skin, mouth,

esophagus, and vagina; subcutaneous fibrosis; contractures and myositis; immunodeficiency; and other manifestations of immune dysregulation. Liver involvement in these patients, like gut mucosal inflammation, is considered to be a protracted form of acute GVHD because clinical and histologic changes in these organs are identical to those in patients with acute GVHD that occurs in the months

553

554

Section IV  Topics Involving Multiple Organs following HCT.206 Patients with isolated elevations of serum alkaline phosphatase or ALT related to GVHD should be followed closely, but may not require high-dose immunosuppressive therapy. By the time that jaundice develops in patients with chronic GVHD, liver biopsy shows extensive damage to small bile ducts and, in severe cases, ductopenia. In patients receiving no, or tapering doses of, immunosuppression, liver GVHD may also present as an acute hepatitis, with abrupt elevations of serum aminotransferase levels to more than 2000 U/L.155 A similar presentation can be seen after donor lymphocyte infusion (DLI).156 In patients presenting with an acute hepatitis, blood samples for viral DNA or ribonucleic acid (RNA) or liver biopsy are essential in excluding acute viral hepatitis due to a herpesvirus (HSV or VZV) or a hepatitis virus and to make a definitive diagnosis of hepatic GVHD (see Fig. 34-5D and E). A serum autoantibody test for CYP1A2 may prove diagnostically useful in the diagnosis of hepatitic GVHD because this enzyme appears to be a target antigen in GVHD.207 Immunosuppressive drug treatment of chronic GVHD is successful in 50% to 80% of patients with extensive multiorgan disease, but refractory chronic GVHD is a fatal illness. The addition of ursodeoxycholic acid (15 mg/kg/day) may result in biochemical improvement.208 Ductopenic GVHD is potentially reversible if ongoing immunologic destruction of biliary epithelium ceases, but this process may take months before resolution of jaundice.155 Liver transplantation, including living-donor transplantation from the original stem cell donor,209 has been performed for patients with liver failure due to chronic hepatic GVHD, although frequently there are contraindications to this approach.210

Chronic Viral Hepatitis and Cirrhosis

HCV infection in HCT survivors almost always results in chronic hepatitis.114,115 In the first 10 years of HCV infection after HCT, there is little liver-related morbidity.114 However, a third of patients transplanted before the 1990s will develop cirrhosis related to chronic HCV infection over a 20- to 40-year time frame.115 The reasons for more rapid progression of fibrosis after HCT may be related to conco­ mitant liver involvement with GVHD, immunosuppression, and iron overload.211 Iron overload is particularly severe in thalassemic patients who have undergone HCT.212 Patients with chronic HCV should be offered therapy with com­ bination pegylated INF-α plus ribavirin, unless there are contrain­dications.167,213 Pegylated interferons, with their longer half-lives, should be administered with caution because some HCT patients experience rapid falls in platelet and granulocyte counts. INF-α may also activate chronic GVHD. Liver transplantation should be considered in any HCT survivor with incipient liver decompensation; in some cases, the original allogeneic cell donor can be a partial liver donor.214 The prevalence of chronic HBV infection among HCT survivors varies widely depending on the country. Patients who remain viremic or HBsAg positive after HCT are at risk of flares of hepatitis B at times of reduction of immunosuppression, such as during tapering or cessation of treatment for chronic GVHD. All long-term survivors with chronic hepatitis B should be regularly monitored to assess the need for antiviral therapy. Hepatitis B e antigen and antibody status should be determined in all patients, and liver enzymes monitored every 6 to 12 months. The need for antiviral treatment with an oral nucleoside or nucleotide analog (for instance, entecavir, telbivudine, adefovir, tenofovir) is based on the ALT and HBV DNA levels and the severity of hepatic fibrosis.215 These levels may change over time, emphasizing the need to continually reassess patients

who are not on treatment. As in the peritransplant period, patients with viremia or positive HBsAg should receive an antiviral agent whenever they receive immunosuppressive or cytotoxic therapy.216

Hemosiderosis

Iron overload is caused by a combination of multiple red cell transfusions (e.g., for thalassemia) and dyserythropoiesis leading to increased iron transport by the intestine. After HCT, iron accumulation stops and body iron stores fall slowly over time.217 The consequences of extreme iron overload in HCT survivors are primarily those of cardiac, pituitary, and endocrine pancreatic dysfunction. Iron overload may be an important cofactor in liver disease in longterm survivors of HCT and should be part of a screening panel.218,219 In the past, liver biopsy with liver iron determination was required, but increasingly noninvasive methods (e.g., MRI129) are being used to provide assessments of liver iron concentration and distribution. Patients with liver iron content greater than 15,000 µg/g dry weight should be treated aggressively with phlebotomy and chelation; when liver iron content is 7000 to 15,000 µg/g dry weight, phlebotomy is indicated; when liver iron content is less than 7000 µg/g dry weight, treatment is indicated only if there is evidence of liver disease.220 Mobilization of iron from heavily overloaded patients improves cardiac function, normalizes serum ALT levels, and results in improved liver histology.219-222

Hepatic Drug Toxicity

Drug-induced liver injury may be related to drugs in common use by transplant survivors, including antihypertensive drugs, lipid-lowering agents, hypoglycemic agents, NSAIDs, antidepressants, antibiotics, and herbal preparations.223 Some drug reactions may result in chronic liver disease.224

Fungal Liver Infections

Fungal abscesses can recur after apparently successful antifungal therapy when high-dose immunosuppressive drugs are started for GVHD. Oral, nonsterile herbal remedies contaminated by molds may lead to liver abscesses in immunosuppressed HCT survivors.225

Liver Cancer

Compared with the general population, patients who survive more than 10 years post HCT have an eightfold risk of developing a new solid malignancy. The risk of hepatocellular carcinoma is particularly elevated.226 Transplant survivors with risk factors for hepatocellular carcinoma (HCV or HBV infection, obesity, diabetes, low platelet count) should be screened at yearly intervals (see Chapter 94).227,228 Chronic hepatitis C may also be a risk factor for development of lymphoma229 and other lymphoproliferative disorders230 after transplant.

Other Hepatobiliary Disorders

There is a higher than expected incidence of gallstones and stone-related biliary problems after HCT than in an agematched population, probably related to earlier formation of biliary sludge. Chronic cyclosporine dosing also may lead to gallstones and biliary symptoms. Patients who have experienced SOS from either chemotherapy or a myeloablative conditioning regimen may rarely develop hepatic nodularity caused by atrophy of zone three and hypertrophy of zone one hepatocytes, without fibrosis.231 This process (nodular regenerative hyperplasia) is usually clinically silent unless portal hypertension develops, manifested by variceal bleed-

Chapter 34  Gastrointestinal and Hepatic Complications of Solid Organ and Hematopoietic Cell Transplantation ing, ascites, splenomegaly, and thrombocytopenia, but with preserved liver function. Focal nodular hyperplasia (nodules readily seen by liver imaging, with characteristic central scars) has been described as an incidental finding in 12% of a cohort of HCT survivors, up to 14 years after transplant.232 These lesions are likely the result of sinu­ soidal liver injury caused by myeloablative conditioning regimens.

Esophageal Disorders

Some patients with extensive chronic GVHD have eso­ phageal desquamation, webs, submucosal fibrous rings, bullae, and long, narrow strictures in the upper and midesophagus.233-236 Although the most common symptom is dysphagia, some patients present with insidious weight loss, retrosternal pain, and aspiration of gastric contents. The diagnosis is made by barium contrast radiography and endoscopy,236 which should be done with caution because perforations have been reported.233 Tight strictures are difficult to dilate safely, but failure to dilate strictures may lead to progressive esophageal narrowing. Esophageal involvement can be prevented by prompt treatment of chronic GVHD at its early stages. Therapy with proton pump inhibitors should be considered if there is uncontrolled acid reflux. Myasthenia gravis may also complicate chronic GVHD, with dysphagia as its presenting complaint.237 Sporadic cases of fungal and rarely viral esoph­ agitis may occur in patients with chronic GVHD on immunosuppressive and antibiotic therapy. Esophageal strictures may be sequelae of earlier herpesvirus infection or mucositis. Squamous cell carcinoma of the esophagus has been reported in HCT survivors, usually with concomitant chronic GVHD of the oropharynx.238

Gastrointestinal Disorders

The incidence of diarrhea falls sharply after day 100 except in patients who have received allografts following reduced intensity conditioning therapy239 and in those whose acute GVHD has never resolved. Patients with protracted acute GVHD, however, often have symptoms that wax and wane with intensity of immunosuppressive therapy for up to 15 years after HCT, with each exacerbation similar to the presenting signs of GVHD that occurred earlier after HCT (satiety, poor appetite, nausea, episodic diarrhea, and weight loss).240,241 The endoscopic and histologic appearance of intestinal mucosa is identical to that seen in acute GVHD (see Fig. 34-4). Use of oral beclomethasone dipro­ pionate can be effective in treating patients with protracted acute GVHD involving the GI tract.242 Before the introduction of more effective immunosuppressive drugs, chronic GVHD resulted in extensive collagen deposition in submucosal and subserosal areas of the intestinal tract, resulting in refractory malabsorption243; this process has not been seen in recent years. There are sporadic cases of C. difficile, CMV,138 and rarely G. lamblia and cryptosporidiosis in longterm survivors. Chronic intestinal viral infection can be seen in patients who remain on immunosuppressive drugs, including rotavirus, norovirus, and adenovirus. Intestinal diseases in donors have been reported in their recipients,

such as idiopathic inflammatory bowel disease and celiac disease.244

Pancreatic Disease

Acute pancreatitis has been described in transplant survivors, usually related to passage of gallstones or sludge. Cyclosporine and tacrolimus also may cause pancreatitis.245-247 Diarrhea, steatorrhea, and weight loss secondary to pancreatic insufficiency have developed in some HCT survivors.248,249 The cause is unclear; previous pancreatic necrosis, prolonged glucocorticoid exposure, and tacrolimus toxicity are the leading possibilities.202 Transient pancreatic insufficiency has been noted in patients with gut and liver GVHD.250 Chronic GVHD and extreme iron overload also may contribute to pancreatic damage.

KEY REFERENCES

Abu-Elmagd K, Reyes J, Bond G, et al. Clinical intestinal transplantation: A decade of experience at a single center. Ann Surg 2001; 234:404-16. (Ref 61.) Appelbaum FR, Forman SJ, Negrin RS, Blume KG. Thomas’ Hemato­ poietic Cell Transplantation. 4th ed. Oxford, UK: Wiley-Blackwell Publishing; 2009. (Ref 94.) Assi MA, Pulido JS, Peters SG, et al. Graft-vs.-host disease in lung and other solid organ transplant recipients. Clin Transplant 2007; 21:1-6. (Ref 66.) Berenguer M. What determines the natural history of recurrent hepatitis C after liver transplantation? J Hepatol 2005; 42:448-56. (Ref 46.) Berkowitz N, Schulman LL, McGregor C, Markowitz D. Gastroparesis after lung transplantation. Potential role in postoperative respiratory complications. Chest 1995; 108:1602-7. (Ref 52.) Hockenbery DM, Cruickshank S, Rodell TC, et al. A randomized, placebo-controlled trial of oral beclomethasone dipropionate as a prednisone-sparing therapy for gastrointestinal graft-versus-host disease. Blood 2007; 109:4557-63. (Ref 135.) Hogan WJ, Maris M, Storer B, et al. Hepatic injury after nonmyeloa­ blative conditioning followed by allogeneic hematopoietic cell transplantation: A study of 193 patients. Blood 2004; 103:76-82. (Ref 118.) Lau GK, Suri D, Liang R, et al. Resolution of chronic hepatitis B and anti-HBs seroconversion in humans by adoptive transfer of immunity to hepatitis B core antigen. Gastroenterology 2002; 122:61424. (Ref 112.) McDonald GB, Slattery JT, Bouvier ME, et al. Cyclophosphamide metabolism, liver toxicity, and mortality following hematopoietic stem cell transplantation. Blood 2003; 101:2043-8. (Ref 121.) Ponec RJ, Hackman RC, McDonald GB. Endoscopic and histologic diagnosis of intestinal graft-vs.-host disease after marrow transplantation. Gastrointest Endosc 1999; 49:612-21. (Ref 134.) Ponticelli C, Passerini P. Gastrointestinal complications in renal transplant recipients. Transplant Int 2005; 18:643-50. (Ref 16.) Ruutu T, Eriksson B, Remes K, et al. Ursodeoxycholic acid for the prevention of hepatic complications in allogeneic stem cell transplantation. Blood 2002; 100:1977-83. (Ref 142.) Small LN, Lau J, Snydman DR. Preventing post-organ transplantation cytomegalovirus disease with ganciclovir: A meta-analysis comparing prophylactic and preemptive therapies. Clin Infect Dis 2006; 43:869-80. (Ref 11.) St Pierre TG, Clark PR, Chua-anusorn W, et al. Noninvasive measurement and imaging of liver iron concentrations using proton magnetic resonance. Blood 2005; 105:855-61. (Ref 129.) van Burik JA, Lawatsch EJ, DeFor TE, Weisdorf DJ. Cytomegalovirus enteritis among hematopoietic stem cell transplant recipients. Biol Blood Marrow Transplant 2001; 7:674-9. (Ref 138.) Full references for this chapter can be found on www.expertconsult.com.

555

CHAPTE R

35 Gastrointestinal and Hepatic Manifestations of Systemic Diseases Rajeev Jain and Dwain L. Thiele

CHAPTER OUTLINE Rheumatologic and Collagen Vascular Diseases  557 Rheumatoid Arthritis  558 Adult-Onset Still’s Disease  559 Progressive Systemic Sclerosis  560 Systemic Lupus Erythematosus  560 Polymyositis and Dermatomyositis  561 Mixed Connective Tissue Disease  561 Sjögren’s Syndrome  561 Polyarteritis Nodosa and Other Vasculitides  562 Behçet’s Disease  563 Seronegative Spondyloarthropathies (Reactive Arthritides)  563 Marfan’s and Ehlers-Danlos Syndromes  563 Familial Mediterranean Fever  563 Oncologic and Hematologic Diseases  564 Metastases  564 Paraneoplastic Syndromes  564 Hematologic Malignancies  564 Systemic Mastocytosis  567 Myeloproliferative and Myelophthisic Disorders  568 Dysproteinemias  568 Coagulation Disorders  569 Red Blood Cell Dyscrasias  570 Endocrine Diseases  572 Diabetes Mellitus  572 Thyroid Disease  575 Adrenal Disease  576

Numerous systemic and extraintestinal diseases have gastrointestinal and hepatic manifestations. Because it is impossible to discuss each entity in great detail in a single chapter, we endeavor here to emphasize frequently encountered diseases and those that may be of particular interest to the reader because of recent developments. For the sake of clarity, some diseases that result in similar manifestations are presented in tabular form. Some topics are taken up in detail in other chapters. The reader is referred to these chapters for a more complete discussion. Although not always a manifestation of systemic disease, nodular

Pituitary Disease  576 Parathyroid Disease  576 Disorders of Lipid Metabolism  577 Hyperlipoproteinemias and Dyslipidemias  577 Abetalipoproteinemia  577 Tangier Disease  577 Neutral Glycosphingolipidoses  577 Renal Diseases  578 Neurologic Diseases  578 Neurogenic Abdominal Pain  578 Gastrointestinal Complications of Acute Head Injury and Stroke  579 Gastrointestinal Problems after Spinal Cord Injury  579 Diseases of the Autonomic Nervous System  580 Extrapyramidal Disorders  580 Multiple Sclerosis  581 Neuromuscular Disorders  581 Pulmonary Diseases and Problems in Patients Who Require Critical Care  581 Intensive Care Unit Patients and Septic Patients  581 Cardiovascular Diseases  584 Infiltrative Diseases  584 Amyloidosis  584 Granulomatous Liver Disease  588 Sarcoidosis  588 Other Infiltrative Disorders  590 Nodular Disorders of the Liver  590

regenerative hyperplasia of the liver is also discussed in this chapter.

RHEUMATOLOGIC AND COLLAGEN VASCULAR DISEASES Rheumatologic diseases encompass a wide variety of clinical syndromes and are frequently associated with gastrointestinal abnormalities (Table 35-1). In addition,

557

558

Section IV  Topics Involving Multiple Organs Table 35-1  Gastrointestinal Manifestations of Rheumatologic Diseases DISEASE

ABNORMALITY/ASSOCIATION

CLINICAL MANIFESTATIONS

Rheumatoid arthritis

Temporomandibular arthritis Esophageal dysmotility Visceral vasculitis Amyloidosis

Impaired mastication Dysphagia, GERD Abdominal pain, cholecystitis, intestinal ulceration and infarction Pseudo-obstruction, malabsorption, protein-losing enteropathy, intestinal ulceration and infarction, gastric outlet obstruction Variceal hemorrhage Enteritis, diarrhea, fever, eosinophilia, megacolon Dysphagia, GERD, stricture, Barrett’s esophagus Gastric retention, GERD Constipation, pseudo-obstruction, malabsorption, intussusception, volvulus, pneumatosis intestinalis Hemorrhage, stasis, bacterial overgrowth Mesenteric thrombosis, infarction, pancreatic necrosis Calcific pancreatitis, pancreatic exocrine insufficiency Dysphagia, reflux GI ulceration, intestinal infarction, intussusception, pancreatitis, pneumatosis intestinalis Oral fissures, oropharyngeal dysphagia Dysphagia

Scleroderma

SLE Sjögren’s syndrome

Polymyositis-dermatomyositis MCTD

Portal hypertension (Felty’s syndrome) Gold enterocolitis Esophageal dysmotility Gastroparesis Intestinal fibrosis and dysmotility Pseudodiverticula Arteritis (rare) Pancreatitis Esophageal dysmotility Mesenteric vasculitis Desiccation of membranes Esophageal webs Gastric lymphoid infiltrates Pancreatitis Primary biliary cirrhosis Skeletal muscle dysfunction Dysmotility Mesenteric vasculitis (rare) Dysmotility

PAN

Mesenteric vasculitis (rare) Mesenteric vasculitis

CSS

Mesenteric vasculitis Eosinophilic gastritis Mesenteric vasculitis

Henoch-Schönlein purpura Kohlmeier-Degos disease Cogan’s syndrome Wegener’s granulomatosis Cryoglobulinemia Behçet’s disease Reactive arthritis Familial Mediterranean fever Marfan/Ehlers-Danlos syndromes

Mesenteric vasculitis Mesenteric vasculitis (infrequent) Crohn’s disease Mesenteric vasculitis Mesenteric vasculitis (rare) Mucosal ulcerations Ileocolonic inflammation Serositis/peritonitis, amyloidosis, PAN, Henoch-Schönlein purpura Defective collagen

Abdominal pain, pancreatic exocrine insufficiency Jaundice, hepatic failure, variceal hemorrhage Aspiration, impaired glutition Dysphagia, GERD, gastroparesis, constipation, diverticula GI ulceration, perforation, pneumatosis intestinalis Dysphagia, GERD, stricture, gastroparesis, bezoars, pseudoobstruction Ulceration, perforation, pancreatitis Cholecystitis, appendicitis, intestinal infarction, pancreatitis, perforation, strictures, mucosal hemorrhage, submucosal hematomas Hemorrhage, ulceration, intestinal infarction, perforation Gastric masses Intussusception, ulcers, cholecystitis, hemorrhage, intestinal infarction, appendicitis, perforation Hemorrhage, ulceration, intestinal infarction, malabsorption Hemorrhage, ulceration, intestinal infarction, intussusception Bloody diarrhea, abdominal pain, fissures, fistulas Cholecystitis, appendicitis, ileocolitis, intestinal infarction Intestinal infarction, ischemia Hemorrhage, perforation, pyloric stenosis Complications as in rheumatoid arthritis Usually asymptomatic Abdominal pain, fever, dysmotility Megaesophagus, hypomotility, diverticula, megacolon, malabsorption, perforation, arterial rupture

CSS, Churg-Strauss syndrome; GERD, gastroesophageal reflux disease; GI, gastrointestinal; MCTD, mixed connective tissue disease; PAN, polyarteritis nodosa; SLE, systemic lupus erythematosus.

the medications used to treat these diseases often produce gastrointestinal and hepatic toxicity. This section focuses on the more common abnormalities that may be encountered by the gastroenterologist.

RHEUMATOID ARTHRITIS

Approximately 0.8% of adults worldwide are affected with rheumatoid arthritis (RA), which is a chronic, inflammatory autoimmune disease primarily targeting the synovial tissues with systemic manifestations. Oropharyngeal symptoms may occur in patients with RA as a result of xerostomia, temporomandibular joint (TMJ) arthritis, cervical spine abnormalities, and laryngeal involvement.1 Esophageal dysmotility, characterized by low-amplitude peristaltic waves, has been described in the proximal, middle, and distal esophagus with reduced lower esophageal sphincter (LES) pressure.1,2 Rheumatoid vasculitis typically occurs

in the setting of severe RA with rare gastrointestinal manifestations such as ischemic cholecystitis or appendicitis, ulceration, pancolitis, infarction, or intra-abdominal hemorrhage due to a ruptured visceral aneurysm.3,4 Other gastrointes­tinal complications of RA include amyloidosis (discussed later) and malabsorption. Felty’s syndrome–RA, spleno­megaly, and leukopenia have been associated with severe infections, portal hypertension, and variceal hemorrhage.5

Hepatic Abnormalities

Abnormal liver function tests, especially elevations of serum alkaline phosphatase of hepatobiliary origin,6-8 are commonly observed in patients with RA. In one large series of patients with RA,6 18% had elevated levels of serum alkaline phosphatase and 11% were found to have hepatomegaly. Fluctuations in serum alkaline phosphatase levels

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases have also been reported to correlate with disease activity.6,8 However, degrees of alkaline phosphatase elevations are usually modest, with the mean level being less than twofold abnormal.6 Furthermore, other clinical signs of liver disease are usually absent, and liver biopsy and autopsy studies have not revealed any consistent or specific findings, with the most common abnormalities being fatty change, Kupffer cell hyperplasia, and mild mononuclear cell infiltration of the portal tracts or rare parenchymal foci of hepatocyte necrosis.6,9-11 Periportal fibrosis is also present in a small minority of cases.11 Determination of etiology of hepatic dysfunction in patients with active rheumatoid arthritis is complicated by the fact that many of the agents commonly used as therapy for this disease have known potential for liver injury.7,10-12 In a small subset of patients with RA and/or Sjögren’s syndrome, antimitochondrial antibodies are present along with the biochemical and histologic features of primary biliary cirrhosis.13-15 The incidence of primary biliary cirrhosis or autoimmune hepatitis appears to be much higher in patients with Sjögren’s syndrome alone than in those with Sjögren’s RA.16,17 Because chronic hepatitis C and RA are relatively common diseases of adults, it is not surprising that these entities are found concurrently in some patients. However, in addition, it has been noted that 75% of individuals with chronic hepatitis C infection develop rheumatoid factors,18 and a subset of these rheumatoid factor–positive individuals develop essential mixed cryoglobulinemia that may be manifested in part by development of arthralgias.19,20 Liver disease in such individuals is often asymptomatic and biochemical abnormalities modest or even absent.19,20 Thus some individuals with essential mixed cryoglobulinemia associated with chronic hepatitis C infection may instead be labeled as having RA. However, anticyclic citrulinated peptide (CCP) antibodies are rarely found in subjects with chronic hepatitis C and nonspecific rheumatologic manifestations and thus anti-CCP antibodies appear to be reliable markers of RA.21 Most rheumatic disease patients with progressive liver disease have concomitant chronic viral or autoimmune hepatitis.22 In patients with concomitant hepatitis B infection, the intermittent use of tumor necrosis factor (TNF) inhibitors or other immunosuppressive therapy may be associated with severe flares of hepatitis B23-25 and prophylactic use of antiviral therapy should be considered.25 TNF inhibitor therapy in RA also has been associated with flares of severe liver disease, with characteristics of autoimmune hepatitis.26 Perhaps the most distinctive association between RA and hepatic abnormalities is seen in another subset of patients who develop splenomegaly and neutropenia (Felty’s syndrome). Felty’s syndrome is associated with an even higher incidence of hepatomegaly and liver function test abnormalities than seen in uncomplicated RA.27,28 However, there is little correlation between serum hepatic enzyme abnormalities and histopathologic findings.27,28 Nevertheless, more than half of patients with this syndrome have been found to have hepatic histologic abnormalities that range from sinusoidal lymphocytosis and portal fibrosis to the more distinctive picture of nodular regenerative hyper­ plasia, which has been reported on multiple occasions in patients with Felty’s syndrome27-30 and in one small prospective series was found to be present in 5 of 18 (28%) patients. Hepatic encephalopathy or other manifestations of liver failure have not been reported in patients with Felty’s syndrome, and nodular regenerative hyperplasia but portal hypertension and esophageal variceal hemorrhage may occur.28-30

Gastrointestinal Abnormalities

The most common gastrointestinal problems encountered in patients with RA are due to drug therapy with nonster­ oidal anti-inflammatory drugs (NSAIDs), glucocorticoids, and disease-modifying antirheumatic drugs (DMARDs). NSAIDs are most commonly associated with upper gastrointestinal complications such as perforation, ulcers, and bleeding (see Chapters 52 and 53). Less commonly recognized complications of NSAIDs include pill esophagitis (Chapter 45), small bowel ulceration (Chapter 115), strictures of the small and large intestine, and exacerbations of diverticular disease and inflammatory bowel disease.31 Significant risk factors for the development of serious upper gastrointestinal events in patients with RA include NSAIDs therapy, age older than 65 years, history of peptic ulcer disease, glucocorticoid therapy, and severe RA.32,33 In patients with RA, the use of certain selective cyclooxygenase-2 inhibitors results in a lower incidence of gastrointestinal complications than that seen with nonselective NSAIDs.34,35 Helicobacter pylori and NSAIDs are independent and possibly synergistic risk factors for peptic ulceration. As such, chronic NSAID users who develop ulcers should be assessed for H. pylori infection and undergo eradication therapy when infection is present.32,33 Although hypergastrinemia has been reported in patients with RA, the incidence of peptic ulcers is no greater than that seen in patients with osteoarthritis.36 As reviewed in Chapter 53, NSAID-associated gastric and duodenal ulcers can be prevented with misoprostol, high-dose histamine (H2) blockers, and proton pump inhibitors.37 Once identified, ulcers may be treated successfully using proton pump inhibitors despite continued NSAID therapy. In the subgroup of patients with a history of bleeding ulcers, therapy with cyclooxygenase-2 inhi­ bitors rather than NSAIDs may be cost effective and less expensive than combining an NSAID with a proton pump inhibitor.38,39 Synthetic DMARDs such as gold and penicillamine are rarely used because of toxicity and marginal efficacy.40 Gold, parenteral as well as oral forms, has been associated with diarrhea, enterocolitis, toxic megacolon, and death. The onset of gold colitis usually occurs within several weeks after the start of therapy and is manifested by nausea, vomiting, diarrhea, and fever. Although the colon is most commonly involved, gold-induced gastrointestinal toxicity may affect the esophagus, stomach, and small bowel, with 25% of patients developing a peripheral eosinophilia.41 Treatment includes dose reduction or discontinuation of gold, antidiarrheals, glucocorticoids, cromolyn sodium, or the chelating agent dimercaprol.41,42 Leflunomide, a synthetic DMARD that inhibits pyrimidine synthesis, can cause diarrhea in up to 32% of patients.43 It may also cause severe hepatic toxicity (see Chapter 86). Biologic DMARDs, which inhibit the action of TNF-α (infliximab, etanercept, and adalimumab) or interleukin-1 (IL-1; anakinra), have not shown significant gastrointestinal adverse effects but may cause hepatic toxicity on occasion (see later).

ADULT-ONSET STILL’S DISEASE

Adult-onset Still’s disease, the adult form of juvenile RA, often has gastrointestinal manifestations such as weight loss, sore throat, hepatosplenomegaly, elevated aminotransferases, and abdominal pain, in addition to fever.44 In contrast to the lack of significant hepatic dysfunction in classic rheumatoid arthritis, adults with Still’s disease present with features of mild hepatitis in the majority of cases and life-threatening acute liver failure in exceptional cases.45-49

559

560

Section IV  Topics Involving Multiple Organs Variable degrees of aminotransferase and alkaline phosphatase elevations are typically observed in such patients during symptomatic disease flares. Liver biopsies usually reveal moderate portal mononuclear cell infiltration with occasional evidence of focal hepatocyte necrosis.46 Biopsies obtained in patients with jaundice and biochemical evidence of severe hepatitis have been found to have interface and lobular hepatitis with lymphoplasmocytic inflammation reminiscent of autoimmune hepatitis.49 Most cases of severe hepatitis have been observed in patients previously treated with salicylates or other NSAIDs,46,49 but liver enzyme abnormalities are also commonly noted prior to therapy. Some patients with severe hepatitis have been reported to respond to immunosuppressive therapy,49 whereas others required liver transplantation or have died of liver failure.45,47-49 Although severe hepatitis is a rare complication of adult-onset Still’s disease, liver failure appears to be the most common cause of death related to this disease.45

PROGRESSIVE SYSTEMIC SCLEROSIS

Progressive systemic sclerosis (PSS, scleroderma) is a multisystem disorder characterized by obliterative small vessel vasculitis and proliferation of connective tissue with fibrosis of multiple organs. Patients with limited cutaneous involvement frequently display findings of the CREST syndrome (calcinosis, Raynaud’s phenomenon, esophageal dysmotility, sclerodactyly, and telangiectasias). Gastrointestinal manifestations occur in up to 90% of patients with PSS.50 Gastrointestinal tract involvement can occur from the mouth to the anus. Atrophy and fibrosis of the perioral skin may limit mandibular motion. The periodontal ligament may become hypertrophic, and the gingivae, tongue papillae, and buccal mucosa may become friable and atrophic, resulting in impaired sensation and taste. The esophagus is the most frequently involved gastro­ intestinal organ.50 On pathology, atrophy of smooth muscle layers and intimal proliferation of arterioles is seen in the distal esophagus.51 Dysphagia occurs as a result of impaired esophageal motility, and gastroesophageal reflux disease (GERD) is related to hypotensive LES pressures, impaired esophageal clearance of acid, and reduced acid-neutralizing capacity due to xerostomia with reduced saliva production.52 The incidence of esophagitis approaches 100% in patients with severe cutaneous involvement.53 The extent of hypomotility varies from occasional uncoordinated contractions to complete paralysis.54 The severity of esophageal dysmotility correlates with the development of interstitial lung disease.55,56 Stricture formation from GERD may contribute to dysphagia, affecting approximately 8% of patients.57 Upper gastrointestinal hemorrhage has been reported from esophageal ulcers, rare esophageoatrial fistulas, and esophageal telangiectasia.58,59 An increased risk of infectious esophagitis with Candida (see Chapter 45) has been attributed to esophageal dysmotility and concomitant immunosuppressive therapy.60 Severe esophagitis typically responds to proton pump inhibitors but may require higher doses for maximal effect.61 A neuropathic achalasia-like syndrome has also been reported.62 The prevalence of Barrett’s metaplasia of the esophagus in patients with PSS ranges between 7% and 38%.63,64 A recent cohort study reported an increased incidence of carcinoma of the tongue and esophagus in patients with PSS.65 Gastric involvement most commonly leads to gastroparesis but other manifestations may include dyspepsia, exacerbation of GERD, or gastric hemorrhage from gastric antral vascular ectasia (GAVE, watermelon stomach). Delayed

gastric emptying has been shown using radionucleotide scintigraphy or radiopaque pellets, with cutaneous electrogastrography demonstrating bradygastria and decreased amplitude of electrical activity.56,66,67 Prokinetic agents such as metoclopramide and erythromycin may increase LES pressures and improve gastric emptying in some patients with PSS.57 The pathologic changes in the small bowel of PSS patients consist of smooth muscle atrophy and deposition of collagen in submucosal, muscular, and serosal layers. Small bowel hypomotility is present in as many as 88% of cases.68 In the early stages of the disease, hypomotility is caused by neuropathic involvement, which may be more responsive to prokinetic agents. In advanced cases, hypomotility is more likely a result of “myopathic” and “fibrotic” changes.57 The interdigestive migrating motor complex (IMMC) is frequently absent or markedly diminished in amplitude in PSS patients with symptoms of intestinal dysmotility.69 Small bowel radiographic abnormalities are present in about 60% of PSS patients, but they may not correlate with symptoms. The duodenum is often dilated, especially in its second and third portions, often with prolonged retention of barium.70 Typically the jejunum is dilated and foreshortened because of mural fibrosis, but valvulae conniventes of normal thickness give rise to an accordion-like appearance. Pneumatosis cystoides intestinalis, pseudoobstruction, pseudodiverticula, sacculations, intussusception, acquired intestinal lymphangiectasia, and small bowel volvulus have been noted.71-73 Symptoms of small intestinal PSS include bloating, borborygmi, anorexia, nausea, and vomiting. Rarely, thrombosis of large mesenteric arteries with extensive bowel necrosis may occur.74 Malabsorption with steatorrhea is present in as many as a third of PSS patients68 and is caused by bacterial overgrowth (see Chapters 101 and 102). Although antibiotic therapy can be effective in these patients, d-xylose malabsorption is often incompletely reversed, suggesting that collagen deposition in PSS may also contribute to malabsorption.75 Although often disappointing, the use of prokinetic agents such as metoclopramide may be effective in some cases. Octreotide in low doses and erythromycin also may provide sustained relief from nausea, abdominal pain, and bloating in some patients with pseudo-obstruction.76 Delayed colonic transit and impaired anal sphincter function are frequently found in constipated patients with PSS.77,78 Cisapride (a drug no longer available in the United States) accelerates colonic transit,79 but refractory cases may require surgery.80 Colonic stricture, volvulus, and bleeding from mucosal telangiectasias have been reported.81,82 Widenecked diverticula can be seen, especially in the antimesenteric border of the transverse and descending colons. Rectal prolapse worsens anal sphincter function, aggravating fecal incontinence in patients with PSS.83 Rectal bleeding can occur from vascular ectasia.84 Pancreatic exocrine secretion is depressed in a third of patients with PSS, and idiopathic calcific pancreatitis has been reported.85 In addition, arteritis resulting in ischemic pancreatic necrosis has been described in patients with PSS.86,87 Gallbladder motility is not altered in PSS.88

SYSTEMIC LUPUS ERYTHEMATOSUS

Systemic lupus erythematosus (SLE) is a multisystem disease characterized by immune system abnormalities and the production of autoantibodies with tissue damage. Gastrointestinal symptoms are common in patients with active SLE. Oral ulcers (one of the criteria used to diagnose SLE) are most commonly seen in the buccal mucosa, hard palate,

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases and vermilion border.89 In SLE, dysphagia (1% to 13% of patients) and GERD (11% to 50% of patients) poorly correlates with esophageal manometric abnormalities such as hypoperistalsis.90 Dysphagia is typically related to GERD or peptic stricture, with one report of esophageal epidermolysis bullosa acquisita.91 Malabsorption of d-xylose, steatorrhea, hyperplastic gastropathy, and protein-losing enteropathy have been described (see Chapter 28); the latter can be steroid responsive.92,93 Lupus peritonitis is a diagnosis that can be made only after other causes have been carefully excluded. Pneumatosis cystoides intestinalis may be an isolated benign condition or may accompany lupus vasculitis or necrotizing enterocolitis.94,95 One of the most devastating complications of lupus is gastrointestinal vasculitis. Affecting only 2% of patients, it has a fatality rate of more than 50%.96 Common sequelae include ulceration, hemorrhage, perforation, and infarction.97-99 Pancreatitis,100,101 gastritis, hemorrhagic ileocolitis resembling inflammatory bowel disease, and intussusception also have been reported. Although occasional case reports have documented polyarteritis-like changes on visceral arteriograms (described later), the typical pathologic changes are seen in the small vessels of the bowel wall rather than the medium-sized vessels of the bowel wall.94 Computed tomography (CT) scan may help establish the diagnosis of ischemic bowel disease in SLE if there are at least three of the following five CT findings: (1) bowel wall thickening, (2) target sign (a thickened bowel wall with peripheral rim enhancement or an enhancing inner and outer rim with hypoattenuation in the center), (3) dilatation of intestinal segments, (4) engorgement of mesenteric vessels, and (5) increased attenuation of mesenteric fat.102 Because visceral angiography is not routinely helpful, the diagnosis is difficult to establish. The role of endoscopy or upper gastrointestinal series in the diagnosis of lupus vasculitis is not well defined. The diagnosis currently rests on clinical judgment, findings on CT scans, and occasionally from surgical specimens when exploratory laparotomy is undertaken to rule out acute surgical emergencies.103 Treatment of abdominal lupus-induced vasculitis with glucocorticoids has been largely unsatisfactory. Although a controlled clinical trial comparing cyclophosphamide with glucocorticoids has not been performed, anecdotal reports of dramatic responses to intravenous cyclophosphamide are promising.94 Some investigators have suggested that cyclophosphamide be considered early in patients who have not shown significant improvement shortly after high-dose glucocorticoids are started. Patients with SLE have a 25% to 50% incidence of abnormal liver biochemical tests during the course of their disease, but clinically significant liver disease is rare.104 Abnormal liver tests are commonly associated either with medication use or with mild, predominantly lobular hepatitis associated with periods of SLE activity.104,105 Despite the shared association with antinuclear antibodies, the typical histologic and clinical features of autoimmune hepatitis are rarely observed in adult patients with SLE.104,106 Concurrent SLE and autoimmune hepatitis occur more frequently in pediatric patients.106 In addition, SLE patients with anticardiolipin antibodies or lupus anticoagulants may have thrombotic events in the liver leading to Budd-Chiari syndrome or nodular regenerative hyperplasia manifested by complications of portal hypertension.104,107

POLYMYOSITIS AND DERMATOMYOSITIS

Polymyositis is a syndrome characterized by weakness, high serum levels of striated muscle enzymes (creatine kinase,

aldolase), and electromyographic (EMG) or biopsy evidence of an inflammatory myopathy. When accompanied by a characteristic violaceous rash on the extensor surfaces of the hands and periorbital regions, the disease is termed dermatomyositis. The primary gastrointestinal symptoms are due to involvement of the cricopharyngeus, resulting in nasal regurgitation, tracheal aspiration, and impaired deglutition.108 Involvement is not limited to skeletal muscle fibers. Disordered esophageal motility, impaired gastric emptying, and poorly coordinated small intestinal peristalsis have been noted.109 Malabsorption, malnutrition, and pseudoobstruction rarely occur.110 Pathologically, edema of the bowel wall, muscle atrophy, fibrosis, and mucosal ulcerations or perforation due to vasculitis may be seen at any level of the gut. Symptoms include heartburn, bloating, constipation, and gastrointestinal hemorrhage. Pneumoperitoneum, pneumatosis intestinalis, colonic dilation, and pseudodiverticula also may be seen. Perforations of the esophagus and of duodenal diverticula have been described as rare complications.111,112 In middle-aged to older adult patients, dermatomyositis and possibly polmyositis are associated with an increased prevalence of malignancy.113 The possibility that gastrointestinal symptoms may be the resultl of an underlying malignancy should be considered when evaluating these patients (see Chapter 22).

MIXED CONNECTIVE TISSUE DISEASE

Mixed connective tissue disease (MCTD) is a syndrome with overlapping features of PSS, polymyositis, and SLE, often in the presence of high levels of antibody directed against ribonucleoprotein. Upper gastrointestinal symptoms are seen in most patients.114 Abnormalities include diminished esophageal peristalsis (48%), esophageal stricture (6%), abnormal gastric emptying (6%), and gastric bezoar (2%).114 Small intestinal and colonic involvement includes dilation of proximal bowel, slow transit, intestinal pseudoobstruction, diverticulosis, and, rarely, intestinal vasculitis. Pancreatitis also has been reported.114 Unlike PSS, the esophageal motility disturbances seen in MCTD appear to improve with the administration of glucocorticoids.

SJÖGREN’S SYNDROME

Sjögren’s syndrome (SS), occurring alone (primary SS) or in association with systemic autoimmune rheumatic diseases (secondary SS), is characterized by lymphocytic tissue infiltration of lacrimal and salivary glands with the clinical findings of keratoconjunctivitis sicca and xerostomia. As reviewed in Chapter 22, excessive dryness of the mouth and pharynx leads to oral symptoms of soreness, adherence of food to buccal surfaces, fissuring of the tongue, and periodontal disease.115 Dysphagia, reported by up to three quarters of patients with SS, can result from esophageal dysmotility and a lack of saliva; however, symptoms do not correlate with manometry or salivary secretion.116-118 Mild atrophic antral gastritis can be seen in 25% of patients with primary SS, but 31% were infected with H. pylori.119 Older studies that reported higher rates and greater severity of gastritis did not control for H. pylori infection. GAVE can occur in patients with SS and is responsive to fulguration therapy.120 A triad of sclerosing cholangitis, chronic pancreatitis, and SS has been reported in eight patients.121 Pancreatic exocrine function is frequently impaired.122 In primary SS, 7% of patients have positive antimitochondrial antibodies and among patients with primary biliary cirrhosis, clinical manifestations of SS are common (see Chapter 89).115

561

562

Section IV  Topics Involving Multiple Organs POLYARTERITIS NODOSA AND OTHER VASCULITIDES

Polyarteritis nodosa (PAN) is a necrotizing vasculitis of small and medium-sized muscular arteries, frequently with visceral involvement (Fig. 35-1). A characteristic feature of this condition is the finding of aneurysmal dilatations up to 1 cm in size seen on visceral angiography (Fig. 35-2). Abdominal complications occur in up to 50% of patients and carries a poor prognosis.123 Other clinical features of PAN include fever, myalgia, arthralgia of the large joints, mononeuritis multiplex, and livedo reticularis. Mesenteric visceral arteriograms are abnormal in up to 80% of patients

Figure 35-1.  Sigmoidoscopic examination in an intravenous drug user who presented with a foot drop and fecal occult blood. Although not shown, the mucosa in the rectum and proximal sigmoid colon was entirely normal. A sural nerve biopsy confirmed the diagnosis of polyarteritis nodosa.

Figure 35-2.  Celiac arteriogram in a patient with polyarteritis nodosa and hepatitis B surface antigenemia. Multiple saccular and fusiform aneurysms, as well as arterial tapering and beading, are seen throughout the celiac artery and its branches, especially the hepatic artery. (Courtesy Connie Wofsy, MD.)

with gastrointestinal involvement, with the superior mesenteric artery most commonly involved.123 Organ damage resulting from ischemia frequently underlies symptoms. The most common gastrointestinal manifestation is abdominal pain with other common symptoms including nausea, vomiting, and gastrointestinal bleeding.123 Bowel infarction and perforation, aneurysmal rupture, and acute cholecystitis are common causes of acute abdomen in PAN.123 Rarely, PAN can present as acalculous cholecystitis secondary to isolated vasculitis of the gallbladder.124 Pancreatitis,125 appendicitis,126 hemobilia,127 solitary biliary strictures,128 and hepatic infarcts129 also have been reported to complicate PAN. The frequency of hepatitis B infection in PAN has declined from more than 30% to less than 10% because of improved screening of the blood supply and vaccination against hepatitis B.130 Because of the frequent association with hepatitis B infection and potential association with hepatitis C infection (see Chapters 78 and 79), patients with clinical manifestations of PAN should be assessed for evidence of hepatitis B or C infection. Churg-Strauss syndrome (CSS, allergic granulomatous angiitis) is a small to medium-sized vessel vasculitis characteristically associated with eosinophilia, asthma, sinusitis, and rhinitis. Abdominal pain is the most common gastrointestinal symptom.131 Preceding the vasculitic phase of CSS, patients may present with an eosinophilic gastroenteritis associated with abdominal pain, nausea, vomiting, diarrhea, and bleeding with an absolute eosinophil count of greater than 1500 cells/mm3 (see Chapter 27).132 Additional gastrointestinal manifestations of CSS include pancreatitis, cholecystitis, ascites, small intestinal ulcerations, and perforation.131,133,134 Colonic involvement may present with multiple ulcers or obstruction.134,135 Henoch-Schönlein purpura (HSP) is a systemic vasculitis characterized by nonthrombocytopenic purpura, arthralgias, renal disease, and colicky abdominal pain. Although the disease is frequently seen in children and adolescents, adults of any age may be affected. Colicky abdominal pain and gastrointestinal bleeding are seen in two thirds of cases.136 Colonoscopic and endoscopic findings in bleeding patients include erosive duodenitis, small aphthous ulcerations, and petechial colonic lesions.137 In patients who undergo CT scan, common findings include bowel-wall thickening, dilated intestinal segments, mesenteric vascular engorgement, and regional lymphadenopathy.138 Other reported gastrointestinal complications of HSP include protein-losing enteropathy, esophageal and ileal structures, gastric and small bowel perforations, bowel infarction, pancreatitis, appendicitis, cholecystitis, intramural hematomas, and intussusception.139 Malignant atrophic papulosis (Kohlmeier-Degos disease) is a rare vasculitis that causes nausea, vomiting, bleeding, malabsorption, bowel ischemia, and perforation.140 Scattered on the skin are red papules that become hypopigmented atrophic scars (see Fig. 22-13). Cogan’s syndrome is characterized by nonsyphilitic interstitial keratitis, audiovestibular symptoms, and large-vessel vasculitis that may involve the gut. Gastrointestinal manifestations include abdominal pain, diarrhea, hepatomegaly, and splenomegaly.141 Crohn’s disease has been reported in association with this rare condition.142 Wegener’s granulomatosis, a systemic vasculitis characterized by pulmonary, sinus, and renal involvement, less commonly affects the gut.143 Inflammatory ileocolitis with hemorrhage, gangrenous cholecystitis, and bowel infarction have been reported.144 Wegener’s granulomatosis may mimic Crohn’s disease with granulomatous gastritis or ileitis.145,146

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases Mixed immunoglobulin (IgG-IgM) cryoglobulinemia characterized by the triad of purpura, arthralgia, and asthenia may complicate chronic hepatitis C infection (see Chapter 79) and a variety of immune diseases, including inflammatory bowel disease, celiac disease, and postintestinal bypass syndrome. Cryoglobulinemia may cause severe visceral vasculitis with diarrhea, ischemia, and perforation of the small or large intestine.147

BEHÇET’S DISEASE

Behçet’s disease is an idiopathic inflammatory disorder characterized by oral aphthous ulcers, genital ulcers, uveitis, and skin lesions, with gastrointestinal involvement occurring in up to 50% of patients.148 As in Crohn’s disease, ulceration may occur throughout the alimentary tract, with the ileocecal region most commonly affected. Differentiating Behçet’s from Crohn’s disease can be difficult because of similarities in gastrointestinal symptoms, endoscopic findings, histology, and extraintestinal manifestations. Involvement of the esophagus includes ulcers (Fig. 35-3), varices, and perforation.149 In the stomach, which is infrequently involved, aphthous ulcers may be seen. The typical intestinal involvement in Behçet’s disease includes “punched-out” ileocecal ulcerations, the most common finding on colo­noscopy. Additional manifestations of Behçet’s disease include abdominal pain, diarrhea, bleeding, perforation, and fistulas (perianal, rectovaginal, and enteroenteric).150 Hepatic or portal vein thrombosis may occur in patients with Behçet’s, and this syndrome should be included in the differential diagnosis of patients presenting with Budd-Chiari syndrome.151,152 Medical therapy of the gastrointestinal lesions of Behçet’s disease includes mesalamine, gluco­corticoids, immunomodulators such as azathioprine and 6-mercaptopurine, infliximab, and thalidomide.153,154 Surgical intervention is associated with a high rate of recurrence, with nearly 50% requiring repeat surgery.155

Figure 35-3.  Aphthous ulcerations of the esophagus on an esophagogram in a patient with Behçet’s disease. (Courtesy of the Radiology Learning Center, University of California School of Medicine, San Francisco, Calif.)

SERONEGATIVE SPONDYLOARTHROPATHIES (REACTIVE ARTHRITIDES)

The term seronegative spondyloarthropathy is used to describe an interrelated group of inflammatory disorders that include ankylosing spondylitis, reactive arthritis (formerly called Reiter’s syndrome), and psoriatic arthritis. The term has also been used to describe the enteropathic spondylitis associated with Crohn’s disease and ulcerative colitis.156 These disorders are characterized by the absence of rheumatoid factor, an association with human leukocyte antigen-B27 (HLA-B27), and inflammation at the site of bony insertion of ligaments and tendons (enthesitis). There is a high prevalence of clinically silent inflammatory colon lesions in patients with these seronegative spondylo­ arthropathies.157 Capsule endoscopy may yield more small bowel abnormalities than ileocolonoscopy.158 Conversely, 22% of patients with inflammatory bowel disease have evidence of a seronegative spondyloarthropathy, with ankylosing spondylitis most commonly seen.159 Although infliximab has been shown to induce remissions in some patients with ankylosing spondylitis as well as in Crohn’s disease, the effect of infliximab on gastrointestinal inflammatory lesions in typical seronegative spondyloarthropathies has not yet been studied.

MARFAN’S AND EHLERS-DANLOS SYNDROMES

Owing to defective collagen synthesis, patients with Marfan’s or Ehlers-Danlos syndrome develop skin fragility, megaesophagus, small intestine hypomotility, giant jejunal diverticula, bacterial overgrowth, and megacolon.160 Mesenteric arterial rupture and intestinal perforation also can occur.161

FAMILIAL MEDITERRANEAN FEVER

Familial Mediterranean fever (FMF) is an autosomal recessive inherited disease characterized by recurrent selflimiting attacks of fever, joint pain, and abdominal pain. Acute attacks typically last three to five days. FMF is most commonly seen in people of Mediterranean origin including Sephardic Jews, Arabs, Turks, Greeks, Italians, and Armenians, although FMF has been described in Cubans, and Belgians. The gene responsible for FMF in Mediterranean patients, designated MEFV, has been mapped to chromosome 16, which encodes a 781-amino acid protein called pyrin or marenostrin.162 Gastrointestinal symptoms, typically manifest as episodic abdominal pain, are seen in 95% of patients, and this may be the presenting symptom in as many as 50% of cases.163 Abdominal pain may be diffuse or localized and may range from mild bloating to acute peritonitis with boardlike rigidity, rebound tenderness, and air-fluid levels on upright radiographs. The acute presentation may be confused with acute appendicitis, cholecystitis, or pelvic inflammatory disease, whereas relapsing and remitting attacks may be confused with other diseases such as porphyrias. Small bowel obstruction from adhesions may occur as a consequence of recurrent sterile peritonitis or as a result of previous exploratory surgery. In patients with obstruction due to adhesions, abdominal attacks without other typical symptoms (arthralgias, fever) should tip off the clinician to consider an obstruction.164 The diagnosis of FMF is based on validated clinical criteria including fever, serositis, location of pain, and response to colchicine.165 In FMF, the long-term prognosis is poor in patients who develop nephrotic syndrome and chronic kidney disease from amyloid A deposition163 (amyloidosis is discussed later in this chapter). Prophylactic colchicine has been shown to reduce the frequency of attacks, prevent amyloi-

563

564

Section IV  Topics Involving Multiple Organs dosis, and avoid renal failure.166 Vasculitis in the form of HSP, PAN, protracted febrile myalgia, or Behçet’s is encountered in 3% of FMF patients.163

ONCOLOGIC AND HEMATOLOGIC DISEASES METASTASES

Metastasis to the gut can occur by direct invasion from adjacent organs, by intraperitoneal seeding, or by hematogenous or lymphatic spread. About 20% of all patients with nongastrointestinal malignancies have metastases to the gastrointestinal tract, the most common of which are breast, lung, and ovarian cancers and melanoma (Fig. 35-4).167 Patterns of metastases are not random but reflect the location and histologic type of the primary tumor. The esophagus is most frequently affected by direct extension from tumors arising from adjacent structures (bronchus and stomach). The stomach is a particularly common site of breast cancer metastases, and the small intestine can be involved by tumor extension from the stomach, pancreas, biliary system, kidney, or retroperitoneum. The pancreas is usually an asymptomatic site of metastasis, with the most common primary tumors being lung, gastrointestinal, and renal.168 The ileum may be affected by cancers arising in the colon or pelvis. Metastases to the gut typically begin in the serosa or submucosa and produce intraluminal lesions that can lead to obstruction, submucosal polypoid masses that can result in intussusception or ulcerated mucosal lesions. The most common presenting clinical condition in patients with metastatic lesions to the gut is small bowel obstruction. Lobular breast cancer, malignant melanoma, and non–small cell lung cancer are the most common neoplasms to cause small bowel obstruction from isolated metastases.169 In addition, pain, fever, ascites, gastrointestinal bleeding, and perforation have been described. Metastases to the gastrointestinal tract may be difficult to diagnose. Barium contrast studies may reveal extramural masses, mucosal ulcerations, or a rigid stomach with the appearance of linitis plastica. CT may be helpful in determining the primary tumor, in tumor staging, and in detecting large serosal implants. Small bowel metastases,

however, are detectable radiographically in only 50% of cases.170 When feasible, surgical resection should be used to treat gastrointestinal metastases that result in obstruction, perforation, or significant hemorrhage. If a solitary bowel metastasis is the only evident site of disseminated malignancy, segmental bowel resection should be performed, offering a small chance for cure. In aggressive resections of melanoma metastases, the mesenteric nodes draining the involved segment of bowel should be resected because they frequently contain tumors.171

PARANEOPLASTIC SYNDROMES

Paraneoplastic syndromes affecting the gut include the hormonal effects of carcinoid tumors, vasoactive intestinal polypeptide-secreting tumors (VIPomas), gastrinomas, and somatostatinomas (see Chapters 31 and 32), as well as the gastrointestinal effects of hypercalcemia (constipation, nausea, and vomiting). A watery diarrhea syndrome with elevated serum immunoreactive VIP has been described accompanying nonpancreatic tumors such as bronchogenic carcinomas, ganglioneuromas, pheochromocytomas, and a rare mastocytoma.172 Elevated serum levels of somatostatin, calcitonin, gastrin, and corticotropin also have been reported in pheochromocytoma.173 Paraneoplastic gastrointestinal dysmotility may occur in some patients with occult or established malignancy and specific serum antibodies. Clinically the patient may present with pseudoachalasia, gastroparesis, intestinal pseudoobstruction, or constipation. Intestinal pseudo-obstruction (see Chapter 120) is most frequently associated with small cell carcinoma of the lung but has been described with other tumors such as squamous cell lung carcinoma, lymphoma, melanoma, and cancers of the kidney, breast, and prostate.174-176 Patients with paraneoplastic intestinal pseudo-obstruction characteristically suffer from consti­ pation and obstipation and from symptoms of intestinal obstruction. In addition, dysphagia, gastroparesis, early satiety, autonomic insufficiency, and peripheral neuropathy have been described.177 The onset of symptoms may precede the discovery of the primary tumor by several years. The gastrointestinal pathology in this syndrome is confined to the myenteric plexus, in which an inflammatory lymphocytic infiltrate is variably seen accompanying neuronal degeneration.178 Cross-reacting autoantibodies found in the sera of these patients bind to the primary tumor cells and to neural cells in the myenteric plexus, resulting in inflammation and destruction of the myenteric plexus.179 In the setting of pseudo-obstruction, detection in the serum of circulating antineuronal nuclear antibodies (ANNA-1 or anti-Hu), type 1 Purkinje cell antibodies (PCA-1), or N-type calcium channel binding antibodies should suggest a paraneoplastic process and prompt further evaluation for an underlying malignancy.177 ANNA-1 are postulated to induce neuronal apoptosis leading to gut dysmotility.180 Although the symptoms of paraneoplastic pseudo-obstruction may resolve with successful treatment of the primary tumor, persistence of gastrointestinal symptoms despite effective anticancer treatment is more common. Attempts to alleviate the symptoms of pseudo-obstruction with prokinetic agents have been disappointing.

HEMATOLOGIC MALIGNANCIES Figure 35-4.  Endoscopic view of an ulcerated metastatic melanoma lesion involving the second portion of the duodenum in a young man who presented with upper gastrointestinal bleeding.

Liver involvement during hematologic malignancies is only rarely life threatening or a source of great morbidity. Nevertheless, the liver is a major component of the reticuloendothelial system, and thus it is not surprising that malignant

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases Table 35-2  Involvement of the Liver in Patients with Hematologic Malignancies Frequency of Liver Infiltration (%)

Hodgkin’s lymphoma Non-Hodgkin’s lymphoma Hepatosplenic γδ T cell lymphoma Multiple myeloma Leukemia   ALL   AML   CLL   HCL   LGLL

Clinical Evaluation

Postmortem

8-14

55

16-57 80

52 100

30-40

40-50

— — — 100 75-100

>95 75 98 100 —

OTHER NOTABLE HISTOLOGIC ABNORMALITIES (FREQUENCY) Portal lymphocytic infiltrates (32%), granulomas (9%-25%), steatosis (11%), hemosiderosis (9%), idiopathic cholestasis (<5%) Portal lymphocytic infiltrates (20%-25%), steatosis (7%) Predominantly sinusoidal infiltrates Amyloidosis (10%), light-chain deposition, extramedullary hematopoiesis — — — Angiomatous lesions (64%) —

ALL, acute lymphocytic leukemia; AML, acute myelogenous leukemia; CLL, chronic lymphocytic leukemia; HCL, hairy cell leukemia; LGLL, large granular lymphocyte leukemia. Data abstracted from references 182-185, 187, 188, 190, 191, 193, 197, 213, 217, 240-242.

infiltration of the liver commonly occurs in such diseases.181 As detailed in Table 35-2, the frequency of malignant infiltration varies from less than 10% to nearly 100% depending on the nature of the underlying hematologic malignancy. In addition to histologic and biochemical abnormalities related to malignant infiltration, a variety of other hepatic abnormalities are observed in a significant fraction of such patients. Many of these abnormalities are related to toxicity of pharmacologic or radiation therapies or to the secondary opportunistic or transfusion-related infections commonly observed in such patients. In addition, a variety of non­ specific histologic abnormalities of uncertain etiology such as steatosis, fibrosis, hemosiderosis, and nonspecific portal lymphocytic infiltrates are observed commonly in treated as well as untreated patients. Other hepatic manifestations relatively unique to selected malignancies also may occur. Such notable paraneoplastic manifestations include granuloma formation or development of pronounced intrahepatic cholestasis in patients with Hodgkin’s disease and deposition of amyloid in patients with multiple myeloma.

Hodgkin’s Lymphoma (see Chapter 29) As detailed in Table 35-2, malignant infiltration of the liver is observed in only a minority of patients with untreated Hodgkin’s disease.182,183 However, autopsy series have noted hepatic involvement in as many as 55% of patients,184 suggesting that hepatic involvement increases with disease progression. Although Reed-Sternberg cells have been reported in only 8% of liver biopsies at the time of initial evaluation, fully a third of specimens exhibit nonspecific mononuclear cell infiltrates in portal tracts and approximately 10% to 25% have noncaseating hepatic granulomas not associated with malignant histiocytes or infectious etiologies.182,185 Moderate elevations of serum alkaline phosphatase activity are often observed, especially in febrile patients or patients with advanced stage disease.186 Although such elevations almost invariably appear related to elevations of the hepatic fraction of serum alkaline phosphatase activity,186 not all patients with elevated alkaline phosphatase levels are found to have tumor infiltration of the liver.182,186 All patients with hepatic involvement have been reported to have splenic involvement,182 but again the presence of splenic infiltration does not invariably imply liver involvement.

Although Hodgkin’s disease may involve extrahepatic bile ducts or lymph nodes in the porta hepatis and cause extrahepatic obstruction, multiple reports describe an additional syndrome of idiopathic intrahepatic cholestasis unrelated to hepatic infiltration, extrahepatic obstruction, or other identifiable causes.187-189 The degree of cholestasis is often disproportionate to apparent tumor load.187,189 However, cholestasis has been reported to resolve with response to systemic therapy,187 although in other cases this syndrome has been associated with intractable, fatal liver damage. Recently, progressive loss of small intrahepatic bile ducts has been documented in some of these patients,189 suggesting that this syndrome may be caused by destruction of bile duct epithelial cells either by direct effects of tumor cells invading the intrahepatic bile ducts or by indirect effects of cytokines released from lymphoma cells. As liver involvement with Hodgkin’s disease defines a patient as having stage IIIE or IV disease, correct interpretation of causes of abnormal liver biochemistries in patients with this disease is often of significant importance in determining prognosis and therapy. Numerous studies have noted the superiority of laparotomy or peritoneoscopy to blind percutaneous liver biopsy in detecting hepatic involvement with Hodgkin’s disease. Presumably this relates to the relatively small volume of tissue obtained at percutaneous liver biopsy and the difficulty in finding diagnostic Reed-Sternberg cells in the liver. Laparoscopy provides a diagnostic yield equal to that obtained at laparotomy, and laparoscopy with or without laparoscopic splenectomy has become the standard approach to diagnostic staging in the majority of patients.183,190

Non-Hodgkin’s Lymphoma (see Chapter 29) Lymphoma involves the gastrointestinal tract either as the primary site or secondarily from systemic lymphomas. Lymphomas may affect any organ and must be included in the differential diagnosis of any gastrointestinal symptom, especially in patients with advanced acquired immunodeficiency syndrome. As noted in Table 35-2, the frequency of liver involvement at initial clinical staging is significantly higher in patients with non-Hodgkin’s lymphomas than in those with Hodgkin’s disease. When evaluated by percutaneous liver biopsy, 16% to 26% of patients with non-

565

566

Section IV  Topics Involving Multiple Organs Hodgkin’s lymphomas are found to have liver infiltration191 with significantly higher percentages found to have hepatic involvement when evaluated by laparoscopy.192 In both Hodgkin’s and non-Hodgkin’s lymphomas, the majority of infiltrative lesions are portal in location.193 Although the overall frequency of hepatic involvement appears similar in different histologic types of lymphoma,191 primary hepatic lymphoma is an unusual variant that occurs more often in diffuse large cell lymphomas of B cell origin than in T cell or non–B, non–T cell lymphomas.194,195 In contrast to secondary lymphomatous involvement of the liver that is often only detected by histologic evaluation, patients with primary lymphoma are commonly found to have evidence of mass lesions on CT, magnetic resonance imaging, or other hepatic imaging procedures that may mimic primary or metastatic carcinoma.194-196 Some reports have suggested an association between primary hepatic lymphomas and immunosuppression or chronic viral hepatitis, but such comorbid conditions are noted in only a minority of cases.194,195 Recently, hepatosplenic γδT cell lymphoma has been recognized as a distinct lymphoma entity.197 This extremely rare form of lymphoma occurs most frequently in young males who present with hepatosplenomegaly secondary to diffuse hepatic sinusoidal and splenic sinus infiltration with clonal populations of T cell receptors, gamma delta (γδTCR) expressing cells. Lym­phadenopathy is absent, but bone marrow involvement common at the time of presentation and cytogenetic analysis commonly reveals an isochromosome 7q and trisomy 8.19 The most common liver test abnormality reported in patients with non-Hodgkin’s lymphoma is a moderately elevated serum alkaline phosphatase. Overall liver test abnormalities are poorly predictive of the presence or absence of lymphomatous infiltration of the liver.192 This likely relates in part to the presence of a variety of nonspecific histologic abnormalities191,193 including portal lymphocytic infiltrates, hemosiderosis, and steatosis that may be associated with liver test abnormalities in patients without hepatic involvement. Other patients with lymphomatous liver infiltrates may have normal liver tests. Noncaseating granulomas also have been found in the portal tracts of patients with non-Hodgkin’s lymphoma though at a much lower frequency than observed in Hodgkin’s disease.198 Extrahepatic obstruction secondary to nodal involvement in the porta hepatis also may occur,192 and in some cases bile duct involvement may mimic the features of cholangiocarcinoma.199 Percutaneous liver biopsies have been found to be of value in detecting hepatic involvement with lymphoma, and if such specimens are properly processed, immunotyping can be performed to better characterize the phenotype of the malignant cells.200 However, quantity of tissue obtained appears very important in determining diagnostic sensitivity, with biopsy at laparotomy being superior to either blind percutaneous or laparoscopic biopsies in obtaining a diagnosis of hepatic infiltration by non-Hodgkin’s lymphoma.192

Leukemia

Approximately 10% of patients with leukemia suffer significant gastrointestinal complications, either from the leukemia itself or as the result of chemotherapy (Table 35-3).201 Examination of autopsy specimens reveals gastrointestinal involvement in less than 15% of all patients with leukemia.202 Acute myelogenous leukemia is the type most likely to affect the gut. Lesions result from four major causes: leukemia cell infiltration, immunodeficiency, coagulation disorders, and drug toxicities. Radiologically, leukemic

Table 35-3  Gastrointestinal Complications of Leukemia Leukemic Invasion of the Bowel and Related Structures Intussusception Adynamic ileus Mucosal ulceration Perforation, hemorrhage Hepatosplenomegaly Splenic infarction, rupture Portal hypertension Ascites, variceal hemorrhage, encephalopathy Biliary and pancreatic duct obstruction Protein-losing enteropathy Pneumatosis intestinalis Watermelon rectum Immunodeficiency Necrotizing enterocolitis (typhlitis) Increased susceptibility to common infections Appendicitis, wound infections, perirectal abscess, sepsis Opportunistic infections Esophageal or hepatic candidiasis, mucositis Herpes infections (HSV < CMV); protozoa Pseudomembranous colitis Coagulation Defects Intramural hemorrhage Hemorrhagic necrosis, obstruction Gastrointestinal hemorrhage Drug Toxicity Mucositis Nausea and vomiting Ileus, megacolon Bowel necrosis Pancreatitis Complications Associated with BMT (see Chapter 34) SOS of the liver Ascites, encephalopathy, hepatic failure Graft-versus-host disease Hemorrhage, malabsorption, strictures, webs Cholestatic liver disease, protein-losing enteropathy Lymphoproliferative syndromes EBV-associated B cell proliferative disease B cell lymphoma BMT, bone marrow transplantation; CMV, cytomegalovirus; EBV, Epstein-Barr virus; HSV, herpes simplex virus; SOS, sinusoidal obstruction syndrome.

lesions assume many forms. Infiltration of the bowel may produce polypoid masses (chloromas), plaquelike thickenings, ulcers, and diffuse masses. Esophageal filling defects with clot and debris have been described.203 Gastric mucosal folds can assume a “brainlike” deeply convoluted appearance resembling adenocarcinoma. Diffuse intestinal leukemoid polyposis may produce obstruction, hemorrhage, or intussusception. Immunodeficiency and immunocytopenia may lead to agranulocytic ulcers with bacterial invasion and bleeding. Coagulation defects can produce intramural hematomas and hemorrhagic necrosis of the bowel. Clinical syndromes are myriad. Common oral symptoms (see Chapter 22) are gingival bleeding, hypertrophy, inflammation, and focal ulcerations. Oral mucositis (stomatitis) is a severe inflammatory condition seen in the setting of recent chemotherapy, radiation therapy, or bone marrow transplantation. Treatment consists of appropriate antifungal, antiviral, or antibacterial therapy as well as viscous lidocaine and systemic analgesia. Esophageal lesions, usually caused by candidiasis or herpesviruses, may cause odynophagia, dysphagia, or bleeding (see Chapter 45). Gastric acid hypersecretion with peptic ulcers has been reported in a patient with hyperhistaminemia secondary to basophilic granulocytic leukemia.204 Massive gastrointestinal hemorrhage may result from infec-

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases tious lesions, agranulocytic ulcers, or primary leukemic lesions of the gastrointestinal tract. The treatment of bleeding gastric and colonic leukemic lesions with radiation therapy has occasionally met with success and has been advocated by some investigators.205 A dire complication, seen in 5% of patients with acute leukemia and 3% of those with chronic leukemia, is the development of an acute abdomen. Acute appendicitis, abdominal abscesses, and perforation are noted with increased frequency. Necrotizing ileocecal enterocolitis and leukemic typhlitis are relatively infrequent but lifethreatening problems in neutropenic leukemia patients. Typhlitis (i.e., inflammation of the cecum) complicates 6.5% of cases of acute myeloid leukemia and 4.6% of cases of acute lymphoblastic leukemia.206 Typhlitis typically manifests after induction chemotherapy and is usually preceded by neutropenia.207 Rarely, typhlitis can be the presenting manifestation of acute leukemia.208 Although the cause of this condition is not entirely clear, multiple factors such as chemotherapy, radiotherapy, neutropenia, and altered gastrointestinal flora are implicated in its pathogenesis.207 Cecal superinfection with fungi and with cytomegalovirus also has been associated with typhlitis. Patients usually present with fever, severe right lower quadrant pain, and occasionally an acute abdomen. Bloody diarrhea accompanies typhlitis in 35% of patients.209 The diagnosis can be inferred indirectly by the finding of symmetrical cecal thickening on abdominal ultrasonography or CT.207 Bowel wall thickness greater than 10 mm is associated with a 60% mortality.210 Most patients with leukemic typhlitis can be managed conservatively with the administration of intravenous fluids, packed red blood cells, and, as needed, granulocyte colonystimulating factor (G-CSF), platelets, and broad-spectrum antibiotics. On rare occasions surgery may be required if dire complications arise. Pseudomembranous colitis may complicate leukemia even in the absence of antibiotic therapy.211 Other rare complications are listed in Table 35-3. Proctologic problems can include stercoral ulcers, neutropenic ulcers, and perirectal abscesses (see Chapter 125). At initial presentation, hepatomegaly is present in the majority of patients with acute lymphocytic leukemia (ALL) and in a significant minority of patients with acute myelogenous leukemia (AML). In patients with advanced stages of these acute leukemias, incidence of liver involvement has been reported in more than 95% of cases of ALL and in about three quarters of patients with AML at autopsy.212 The hemorrhagic complications of these acute leukemias rarely permit histologic evaluation of the liver in patients with early, active disease. Therefore, it is difficult to discern the relative contributions of leukemic infiltrates, extra­medullary hematopoiesis, other infectious or toxic complications of these diseases, or the therapies employed to the development of hepatomegaly and/or liver test abnormalities. In patients with leukemias that run more chronic courses, sufficient numbers of patients have been biopsied to indicate that hepatic involvement is far more commonly detected on histologic evaluation than initially indicated by clinical or laboratory assessment.213-217 In an autopsy series, 98% of patients with chronic lymphocytic leukemia (CLL) were found to have leukemic infiltration consisting predominantly of portal infiltrates that usually left the hepatic limiting plates intact.213 However, in some cases, leukemic infiltrates were observed to bridge adjacent portal tracts and were associated with hepatocellular necrosis, bridging necrosis and occasionally pseudolobule formation. In contrast to the predominantly portal pattern of hepatic infiltra-

tion during CLL, liver involvement during hairy cell leukemia (HCL), large granular lymphocyte leukemia (LGLL), or the adult T cell leukemia/lymphoma syndrome associated with human T-lymphotropic virus type 1 (HTLV1) infection is usually characterized by diffuse sinusoidal infiltration or a mixed pattern of portal and sinusoidal involvement.214-217 As in the case of CLL, all or nearly all patients with HCL including some without hepatomegaly or liver test abnormalities demonstrate hepatic infiltration on histologic evaluation.215,216 Although infiltration by HCL occasionally may be missed on conventional histologic evaluation, use of tartrate-resistant acid phosphatase staining215 and immunotyping by staining with monoclonal antibodies against lymphocyte cell surface markers200 has been reported to enhance diagnostic sensitivity and specificity. HCL has also been associated with angiomatous lesions in the liver created by disruption of the sinusoidal wall, creation of wide areas of communication between the sinusoidal lumen and space of Disse, and replacement of the sinusoidal cell lining by tumor cells in direct contact with hepatocytes.216

SYSTEMIC MASTOCYTOSIS

Systemic mastocytosis, a clonal disorder of the mast cell– progenitor associated with activating mutations in the c-kit gene, is characterized by a dense infiltrate of mast cells in extracutaneous tissue (bone marrow, spleen, liver, lymph nodes, and gastrointestinal tract) (Fig. 35-5).218 The classic dermatologic finding of urticaria pigmentosa may be seen with or without systemic involvement (see Chapter 22). The typical symptoms of mastocytosis (pruritus, flushing, tachycardia, asthma, headache) are believed to result from the release of histamine and prostaglandins (e.g., PGD2) from mast cells.219 Heparin is also released from mast cells and may contribute to a bleeding diathesis.220 Eighty percent of patients have gastrointestinal symptoms that include nausea, vomiting, diarrhea, and abdominal pain.221 Hepatomegaly, portal hypertension, splenomegaly, and ascites may occur frequently.222 These symptoms can be precipitated or provoked by heat, alcohol, aspirin, anticholinergics, NSAIDs, and contrast media.219 Hyperhistaminemia produces gastric hypersecretion in more than 40% of cases,223 and secretion may be as marked as in Zollinger-Ellison syndrome.224 Gastric hyperacidity correlates with the degree of histaminemia and with the presence of acid-peptic disease.223,224 Duodenal ulceration or duodenitis has been reported in more than 40% of cases.223 Gastrointestinal hemorrhage from peptic ulcers and from bleeding esophageal varices has been reported.224,225 Diarrhea has been reported in as many as 60% of cases, and minimal fat malabsorption occurs in some cases.220,223 Decreased absorption of d-xylose and vitamin B12 is also found in patients with mastocytosis.223 The cause of diarrhea is unclear. Some diarrheal symptoms (but not malabsorption) respond to H2-receptor antagonists, but there is no clear correlation between stool output and the degree of plasma histaminemia or gastric acidity.223 It is presumed that diarrhea and malabsorption are the result of morphologic changes in the absorptive mucosa. Jejunal biopsy specimens may show large numbers of mast cells in the lamina propria, muscularis mucosa, and submucosa, with normal villi or mild villous atrophy.226 The colon may also be involved (see Fig. 35-5). Endoscopy may reveal urticaria-like mucosal lesions, thickened gastric folds, and edematous mucosa, whereas colonoscopy has shown purple pigmented lesions.227 Small bowel radiographic abnormalities include bull’s-eye lesions resembling metastases, edema, thickened folds, and a nodular mucosal pattern.228 Abdominal ultrasound and CT may show hepa-

567

568

Section IV  Topics Involving Multiple Organs

A

B Figure 35-5.  Systemic mastocytosis involving the colon. A, An interstitial infiltrate of mast cells with pale cytoplasm is present. (Hematoxylin and eosin, ×100). B, Mast cell tryptase immunohistochemical stain highlights the interstitial infiltrate (×400). The patient also had bone marrow involvement. (Courtesy Imran Shahab, MD, Dallas, Tex.)

tosplenomegaly, adenopathy, thickening of the omentum and the mesentery, and ascites.229 H1-receptor and H2receptor antagonists, anticholinergics, oral disodium cromoglycate, and glucocorticoids have been used successfully to relieve the diarrhea and abdominal pain of mastocy­ tosis.219 Imatinib mesylate, a tyrosine kinase inhibitor, is not effective in systemic mastocytosis because a conformational change associated with the most common mutation (Asp816Val) interferes with drug binding.218

MYELOPROLIFERATIVE AND MYELOPHTHISIC DISORDERS

Because the liver is a major site of extramedullary hematopoiesis, hepatomegaly or mild liver test abnormalities secondary to extramedullary hematopoiesis may be appreciated in any of a variety of myeloproliferative disorders or marrow infiltrating malignancies. Benign or malignant proliferations of histiocytes (macrophages) or dendritic cells may be complicated by hepatomegaly or jaundice related to diffuse infiltration of hepatic sinusoids by erythrophagocytic histiocytes, development of peliosis hepatis or intrahepatic, or extrahepatic invasion of bile ducts and portal tracts by histiocytes

or Langerhans’ (dendritic) cells.230-232 Erythrophagocytosis may be a manifestation of malignant histiocytosis or represent a reactive benign histiocyte proliferation in patients with advanced T cell lymphomas.197 Thus assessment of involved tissues for malignant cells of T cell or more rarely B cell origin should also be included in the diagnostic evaluation of cases of uncertain origin. Liver biopsies are commonly abnormal in untreated patients with Langerhans cell histiocytosis (formerly termed histiocytosis X). The most common abnormality is mild mononuclear cell infiltration of the portal tracts.231 However, portal triaditis associated with periportal fibrosis, cirrhosis, or extrahepatic cholangiographic evidence of sclerosing cholangitis also may be seen and in some patients may lead to severe cholestatic liver disease.231,232 Primary myelofibrosis (PMF) is a myeloproliferative disease characterized by bone marrow fibrosis with pro­ gressive anemia and splenomegaly. Portal hypertension, which occurs in 7% of patients with PMF, results from increased portal venous flow and from infiltration of the liver by foci of extramedullary hemato­poiesis.233 Massive gastrointestinal hemorrhage complicates 5% of cases and most often is due to bleeding esophageal varices. Extramedullary hematopoiesis can involve the esophagus, stomach, and small bowel leading to abdominal pain and hemorrhage.234 Increased thrombotic complications have been associated with PMF, polycythemia vera, and essential thrombocytosis.235 Splenic infarction can cause left upper quadrant abdominal pain. As many as 50% of patients with hepatic vein thrombosis, or the Budd-Chiari syndrome, have an overt myelodysplastic syndrome.236 One study237 suggests that 80% of patients with hepatic vein thrombosis may have latent myeloproliferative abnor­malities without overt disease (see Chapter 83).

DYSPROTEINEMIAS

Multiple myeloma or plasma cell tumors may directly involve the gastrointestinal tract with amyloidosis or with local infiltration by plasmacytomas. Twenty-one percent of patients with amyloidosis have multiple myeloma.238 As with gastrointestinal involvement by amyloidosis from other causes (see later), bowel wall infiltration and dysmotility underlie most clinical symptoms. Primary extramedullary plasmacytoma account for 3% to 5% of all plasma cell dyscrasias with gut involvement noted from the oral cavity to anus with manifestations including dysphagia, hemorrhage, pseudo-obstruction, and polyposis.239 Hepatomegaly and abnormalities of liver biochemistries are commonly observed in patients with multiple myeloma.240 In up to half of patients with hepatic histologic evaluation, either diffuse sinusoidal or portal infiltration or, less commonly, nodule formation in the liver by malignant plasma cells has been observed.193,240,241 The frequency of jaundice has ranged from 0% to 30% in series of patients with hepatic infiltration by multiple myeloma.240,241 Ascites formation or, more rarely, esophageal varices have been reported to complicate the course of 10% to 35% of patients with massive hepatic infiltration.200,240 Portal hypertension secondary to tumor infiltration appears to be the cause in most patients, although other causes including congestive heart failure, dissemination of myeloma cells into the peritoneal cavity or development of tuberculous peritonitis also have been noted. In addition to direct malignant infiltration and development of such nonspecific hepatic abnormalities as hemosiderosis or portal lymphocytic infiltrates, multiple myeloma is complicated in about 10% of patients by deposition of amyloid or non–amyloid-containing IgG light chain deposits in the space of Disse.200,240-242 Extramedullary hemato­

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases poiesis also may contribute to hepatomegaly or liver test abnormalities in these patients.240 Clinical staging and follow-up of patients with multiple myeloma are largely based on assessment of marrow, osseous, and serum or urinary abnormalities, and thus histologic evaluation of the liver is only occasionally considered. As discussed in the section dealing with amyloidosis later in this chapter, potential diagnostic benefits of liver biopsy must be weighed against concerns regarding bleeding complications. Waldenström’s macroglobulinemia, a neoplasm characterized by malignant proliferation of lymphocytes producing IgM, presents with hepatomegaly or splenomegaly in one third of patients.243 Gastrointestinal IgM deposition may occur in an infiltrative pattern characterized by diffuse infiltration of the bowel wall with neoplastic cells similar to the pattern seen in immunoproliferative diseases. More commonly, acellular macroglobulin is deposited predominantly in the tips of the villi, the interstitium, and the lacteals, leading to lymphangiectasia.244 Small intestinal mucosal IgM deposits may stain weakly with periodic acid–Schiff, simulating the microscopic appearance of Whipple’s disease. Gastric involvement may present with epigastric pain or bleeding, whereas small intestinal disease can present with steatorrhea, diarrhea, protein-losing enteropathy, pseudo-obstruction, or occult bleeding.244 The rare plasma cell proliferative disorder, γ heavy-chain disease, has been associated with abdominal pain, weight loss, and gastric infiltration from malignant plasma cells.245 α Heavy-chain disease, an immunoproliferative small intestinal disease (IPSID), is a mucosa-associated lymphoidtissue lymphoma characterized by infiltration of the bowel wall resulting in malabsorption and protein-losing enteropathy (see Chapter 29). IPSID is mostly seen in the Mediterranean basin, Middle East, Far East, and Africa, with a recent study suggesting that Campylobacter jejuni may be a causative agent.246

Table 35-4  Causes of Intramural Hematomas in the GI Tract Platelet Deficiency Idiopathic thrombocytopenia Thrombotic thrombocytopenic purpura Hemolytic-uremic syndrome Leukemia (see Table 35-3) Hypersplenism Platelet Dysfunction Glanzmann’s thrombasthenia Bernard-Soulier syndrome Coagulation Defects Hemophilia von Willebrand’s disease Dysfibrinogenemia Disseminated intravascular coagulation Hepatic failure Pharmacotherapy Heparin Warfarin Streptokinase/urokinase Tissue plasminogen activator Trauma Blunt abdominal trauma Endoscopic manipulation Forceful vomiting (esophageal hematoma) Vasculitis Polyarteritis nodosa Henoch-Schönlein purpura Ehlers-Danlos syndrome GI, gastrointestinal.

COAGULATION DISORDERS

In hemophiliac individuals, acute abdominal pain can be a manifestation of spontaneous intra-abdominal hemorrhage. Gastrointestinal bleeding may occur from varices related to chronic liver disease secondary to hepatitis C acquired from transfused blood products. von Willebrand’s disease, heparin or warfarin therapy, hepatic failure, qualitative or quantitative platelet defects, and other bleeding diatheses may result in gastrointestinal hemorrhage or intramural hematomas (Table 35-4). Radiologically, intramural bleeding can be recognized by thickened mucosal folds, rigidity, luminal narrowing (Fig. 35-6), and intragastric masses. Intestinal obstruction and intussusception may result. Hemolytic-uremic syndrome (HUS) consists of a triad of acute kidney injury, microangiopathic hemolytic anemia, and thrombocytopenia without the consumption of humoral clotting factors through defibrination. In children, idiopathic, sporadic, and epidemic cases have variously been described. In adults, HUS occurs in conjunction with complications during childbirth or chemotherapy, with mitomycin C being the most common implicated agent.247 More commonly, adult HUS is preceded by a mild diarrheal illness. Enteric pathogens associated with the HUS prodrome (“HUS colitis”) include Shigella, Salmonella, Yersinia, Campylobacter, and the “hemorrhagic” 0157:H7 strain of Escherichia coli (see Chapter 107).248-252 Undercooked hamburger is the most common vector for 0157:H7 infection with apple juice, radish sprouts, and sausages also implicated in the spread of this infection.248 Several studies suggest that antibiotic therapy of E. coli 0157:H7 with antibiotics increases the risk of development of HUS in children

Figure 35-6.  Computed tomography (CT) scan demonstrating an intramural hematoma of the third portion of the duodenum in a patient treated with warfarin. The contrast-filled duodenal lumen (shown anterior and just to the left of the spine) is circumferentially narrowed by a submucosal infiltrate with the same CT density as that of blood. The hematoma resolved with conservative measures. (Courtesy J. C. Ryan, MD.)

and adults; however, this assertion has been challenged in a meta-analysis.253 Empirical therapy of diarrhea with antimicrobial agents may be appropriate for certain subsets of patients, such as those who are quite clearly at high risk of invasive infections.254 Once HUS appears, colonic involvement is common owing to microangiopathic thrombosis of submucosal vessels and intramural hemorrhage.255 Pancreatitis has also been described.256 Radiographic

569

570

Section IV  Topics Involving Multiple Organs abnormalities include mucosal irregularities, intestinal dilation, filling defects, bowel wall edema, and findings that may resemble those of idiopathic ulcerative colitis, or vascu­litis.257 Because HUS is usually self-limited, therapy consists of hemodialysis and supportive gastrointestinal care. Severe complications may include hemoperitoneum, transmural bowel necrosis with perforation, or colonic stricture.258,259 Thrombotic thrombocytopenic purpura (TTP) is an idiopathic disorder consisting of thrombocytopenia, microangiopathic hemolytic anemia (without significant consumption of clotting factors), fever, renal insufficiency, and profound neurologic dysfunction. Compared with HUS, central nervous system (CNS) symptoms predominate in TTP and renal failure is less severe than in HUS—20% of patients have nonspecific abdominal complaints. The bleeding diathesis of TTP can lead to gastrointestinal hemorrhage, but TTP may also cause thrombosis of intestinal vessels that resembles HUS, clinically and pathologically. Acute colitis, cholecystitis, and pancreatitis have been described.260,261 Plasmapheresis allows 90% of patients with TTP to survive an episode without permanent organ damage.255 Venous thromboembolism (VTE), comprising deep venous thrombosis (DVT) and pulmonary embolism, may be the first manifestation of malignancy. Trousseau’s syndrome refers to migratory superficial thrombophlebitis or DVT occurring in patients with occult cancer. In patients with VTE without a predisposing factor, the incidence of previously undiagnosed cancer was 6% at baseline and 10% from baseline to one year.262 In patients with newly diagnosed VTE, an extensive screening strategy for occult malignancy, which included abdominal imaging and colonoscopy, was superior to a limited screening strategy of medical history, physical examination, and basic blood work in the detection of cancer.263

RED BLOOD CELL DYSCRASIAS Sickle Cell Disease

Sickle cell disease is an autosomal recessive disorder of hemoglobin structure that is characterized by chronic hemolytic anemia and recurrent episodes of vascular occlusion leading to ischemia and distal tissue infarction in multiple organs. Eight percent of African Americans are heterozygous for the hemoglobin S trait and homozygotes comprise 0.2% of African Americans. Patients with sickle cell anemia and other hemoglobinopathies may develop splenic infarction and liver disease (see following), likely from ischemic injury due to intrasinusoidal sickling and impairment of intrahepatic blood flow and delivery of oxygen to hepatocytes.264,265 Chronic anemia due to hemolysis is typically present and predisposes to an indirect-reacting bilirubin elevation and to the formation of pigmented gallstones.266 Patients with other hereditary defects involving red blood cell cytoskeletal proteins, hereditary spherocytosis, and hereditary elliptocytosis, also have diminished red blood cell survival, leading to an increased incidence of pigmented gallstones. Transfusions are frequently used in the therapy for sickle cell anemia, and therefore such patients who were transfused prior to 1992 are at increased risk for hepatitis C.267 Multitransfused teenage and adult patients with sickle cell anemia also have been found to have degrees of excess hepatic iron stores that are comparable to that noted in thalassemia major.268-270 Sickle cell crisis, an acute manifestation of this disease, is characterized by severe skeletal pain and fever. Abdomi-

Figure 35-7.  Sickled red blood cells in a Kupffer cell (erythrophagocytosis) in a hepatic sinusoid of patient with sickle cell disease. (Hematoxylin and eosin, ×200.)

nal pain is also commonly present, and it is important to distinguish vaso-occlusive crises from surgical conditions such as cholecystitis, bowel infarction, appendicitis, and pancreatitis. Abdominal pain from vaso-occlusive crises tends to be more diffuse and associated with remote pain such as limb and chest pain. The pain of vaso-occlusive crises is typically relieved with hydration and oxygen within 48 hours.271 Sickle cell hepatopathy, a syndrome characterized by severe hyperbilirubinemia out of proportion to degree of ongoing hemolysis, is a rare complication of sickle cell anemia that can be associated with coagulaopthy and encephalopathy, leading to death from acute liver failure. Because treatment with exchange transfusions is associated with improved outcome,272 it is important to distinguish this complication of sickle cell anemia from other causes of liver disease that are common in this patient population.273 When histologic evaluation of the liver has been performed in patients with sickle cell anemia at autopsy, cholecystectomy or diagnostic percutaneous liver biopsy, dilated sinusoids, erythrophagocytosis by Kupffer cells (Fig. 35-7), and varying degrees of parenchymal atrophy in the central zones of the liver have been observed frequently.265,274-277 In association with hepatic sinusoids engorged by phagocytosed, sickled red blood cells, adjacent areas of ischemic necrosis have been reported in patients with acute episodes of jaundice, right upper quadrant pain, fever, and leukocytosis thought to be secondary to intrahepatic sickle cell crises or sickle cell hepatopathy.264,265,274,276,278 Accumulation of collagen or thin basement membranes within the space of Disse,274 peri-sinusoidal fibrosis,276 and an apparently high incidence of cirrhosis in patients with sickle cell anemia264 has suggested that recurrent ischemic injury secondary to intrahepatic sickling may also be a cause of chronic liver disease. Although early reports suggested that viral hepatitis was an unusual cause of acute or chronic liver disease in such patients,274,275 studies conducted in the 1980s and 1990s suggested that hepatitis B276,277 and hepatitis C267 are common infections in patients with sickle cell anemia and may account for many episodes of acute or chronic liver disease previously attributed to sickle cell hepatopathy. More recent studies of adult sickle cell anemia patients presenting with acute hepatic dysfunction have found that viral or autoimmune hepatitis plays a role in about a third of such epi-

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases sodes.273 Of special note, intrasinusoidal sickling and Kupffer cell erythrophagocytosis (see Fig. 35-7) almost invariably are found in all patients with sickle cell disease irrespective of apparent cause of liver disease or degree of serum alanine aminotransferase (ALT) elevation.277 Intrasinusoidal sickling was also found in two liver biopsies performed after recovery from acute viral hepatitis. To some extent the presence of nonphagocytosed sickled red blood cells in hepatic sinusoids could be attri­buted to the fact that formalin fixation was noted to induce irreversible sickling of red blood cells in patients with hemoglobin SS or sickle cell disease.277,279 In addition, Omata and colleagues277 have suggested that Kupffer cell erythrophagocytosis may reflect the role of Kupffer cells in clearance of sickled red cells in functionally asplenic patients with sickle cell disease. Thus, assessment of degree of intrasinusoidal sickling or even Kupffer cell erythro­phagocytosis is a poor indicator of possible ischemic liver injury in patients with sickle cell disease. In contrast, other features of vascular insufficiency such as acute ischemic necrosis, sinusoidal dilatation and perisinusoidal fibrosis appear to be more specific markers of vascular injury in patients with symptomatic liver dysfunction in the absence of viral hepatitis or other causes for hepatocellular injury.276 Diggs265 reported in 1965 that 10% of patients presenting with acute sickle crises were jaundiced. More recent assessment of prevalence of liver disease in patients with sickle cell anemia have found persistent abnormalities of one or more liver enzyme tests in 24% of patients followed for sickle cell anemia.279 In addition, 48 of 72 (67%) patients without other biochemical evidence of liver disease had total serum bilirubin levels of greater than 2 mg/dL (greater than 34 µmol/L). Thus laboratory abnormalities suggesting possible liver disease are relatively common in patients with sickle cell disease and frequently lead to diagnostic evaluations as detailed in Figure 35-8. Of note, however, a high rate of complications has been reported in association with liver biopsies performed during acute sickling crises280 and thus invasive diagnostic procedures should be used judiciously in this patient population. In sickle cell disease patients without liver disease, hyperbilirubinemia is exclusively unconjugated or indirect and only uncommonly exceeds levels of 4.5 mg/dL (77 µmol/L). In more severely jaundiced patients, higher serum lactic dehydrogenase levels are also seen and suggest higher rates of hemolysis.279 However, in the setting of acute viral hepatitis or other causes of liver dysfunction, extreme levels of hyperbilirubinemia consisting of relatively equal amounts of direct and indirect bilirubin are observed.279,281 Although early reports suggested that a total serum bilirubin level of greater than 25 mg/dL (greater than 428 µmol/L) was a grave prognostic sign,265,282 relatively benign courses have been noted in patients with extreme degrees of hyperbilirubinemia during the course of acute viral hepatitis or presumed intrahepatic sickle cell crises.279,281,283 The degree of serum ALT or aspartate aminotransferase (AST) elevation in patients with sickle cell anemia and acute viral hepatitis is similar to that observed in other patients with acute viral hepatitis, with most symptomatic patients having elevations more than 10-fold the upper limit of normal.275,281 However, among patients with jaundice and other symptoms such as fever, leukocytosis and intense right upper quadrant pain attributed to intrahepatic sickling or sickle cell hepatopathy, serum AST and ALT values often have been found to be only modestly elevated272,275,282,283 although in other cases, elevations in excess of 15 times the upper limit of normal have been

Elevated: Indir Bili, LDH Normal: Dir Bili, ALT, AP/GGTP

Elevated: Dir Bil., ALT and/or AP/GGTP

Dx: Hemolysis

Evidence of sickle crisis? Yes Treat sickle crisis No

Additional liver testing unnecessary

Positive

Normal liver tests

Hepatitis serologies

Ultrasonography

Negative Dx: Viral hepatitis

Persistently abnormal liver tests

Dilated ducts/ gallstones Murphy's sign

Dx: Gallstone disease Drug Hepatotoxicity Fe/TIBC/Ferritin ANA/SMA/LKM1 Ceruloplasmin

Specific Dx

Diagnostic uncertainty Consider liver biopsy and/or cholangiography Figure 35-8.  Algorithm for the evaluation of abnormal liver biochemical test levels in patients with sickle cell disease. ALT, alanine aminotransferase; ANA, anti-nuclear antibodies; AP, alkaline phosphatase; Dir Bil, direct bilirubin; Dx, diagnosis; Fe, iron; GGTP, gamma glutamyl transpeptidase; Indir Bili, indirect bilirubin; LDH, lactate dehydrogenase; LKM1, liverkidney microsomal antibodies type 1; SMA, smooth muscle antibodies; TIBC, total iron binding capacity.

noted.274 Thus, in patients with jaundice and prominent ALT elevations, both acute viral hepatitis and ischemic injury related to sickle cell disease itself must be considered as possible etiologies. In addition, coincidental causes of liver disease such as autoimmune hepatitis284,285 have been reported in patients with sickle cell anemia and clinically apparent liver disease that was initially incorrectly ascribed to complications of sickle cell disease. Thus when evaluating liver disease in these patients, care must be taken to consider the full spectrum of possible etiologies. Although the majority of patients with acute hepato­ cellular dysfunction thought secondary to either viral hepatitis or intrahepatic sickle cell crises recover following supportive care, cases of acute hepatic failure have been reported.264,275,278,286 Most such patients have presented with right upper quadrant pain, jaundice, modest aminotransferase elevations (less than 10-fold elevated) and progressive coagulopathy, and at postmortem examination have had histologic findings suggesting that the initiating cause of liver failure was centrilobular necrosis secondary to vascular complications of sickle cell disease. Recovery from severe cholestasis and coagulopathy has been reported after

571

572

Section IV  Topics Involving Multiple Organs exchange transfusions.287 It also has been suggested that sickle cell disease may be a predisposing factor to the development of fulminant hepatic failure in children with acute viral hepatitis.286 Thus even in cases of liver failure with an apparent viral etiology, aggressive measures directed at reversal or prevention of intrahepatic sickling may be warranted. In all sickle cell disease patients with direct hyperbili­ rubinemia and especially those with right upper quadrant pain, fever, or leukocytosis, acute cholecystitis with or without choledocholithiasis must be investigated as possible primary or contributing causes. A number of reports indicate that bile duct stones are found in a significant fraction of sickle cell disease patients undergoing cholecystectomy for symptomatic biliary tract disease.288 However, other studies have noted no objective evidence of choledocholithiasis or acute or chronic cholecystitis in many sickle cell disease patients undergoing cholecystectomy for presumed symptomatic biliary tract disease. This has led to speculation that events related to intrahepatic ischemia might better explain the signs and symptoms of liver disease in many of these patients.276 A significant rate of operative and anesthetic complications have been reported in patients with sickle cell disease.276,288,289 Thus cholecystectomy is not recommended in asymptomatic patients with cholelithiasis. However, several reports have noted that in sickle cell patients with recurrent bouts of right upper quadrant pain and jaundice, a marked decrease in such symptomatic episodes is noted after cholecystectomy.276,288,289 Thus in recurrently symptomatic patients with gallstones in whom there is difficulty in distinguishing between cholecystitis and intrahepatic crisis, cholecystectomy is recommended.289 However, in such patients special attention should be directed toward minimizing risks of anoxic injury during surgery by preoperative transfusion of red blood cells and expansion of intravascular volume and by intra- and postoperative oxygen therapy. Chronic liver disease related to chronic viral hepatitis or to post-transfusional iron overload is being recognized with increased frequency in adult patients with sickle cell disease.267-269,290,291 Ferritin levels should be monitored in sickle cell patients who are recipients of multiple red cell transfusions and iron chelation therapy considered for those with evidence of significant iron overload.269 In recipients of chelation therapy and in the general sickle cell disease patient population, zinc deficiency related to excess renal losses is prevalent and via effects on ornithine transcarbamylase may potentiate hyperammonemia. Thus zinc therapy should be considered in sickle cell disease patients with evidence of hepatic encephalopathy.292

Thalassemia

Patients with thalassemia typically develop hepatomegaly from extramedullary hematopoiesis, with CT revealing well-defined hypodense lesions that enhance in the portovenous phase of contrast injection.293 These patients can also develop iron overload due to multiple transfusions, with the resulting end-organ dysfunction in the liver, gonads, and pancreas. The early parenteral use of the ironchelating agent desferoxamine, in an amount proportional to the iron load, has been shown to be effective in halting the progression of fibrosis in patients with thalassemia major.294 However, deferiprone, an orally acting ironchelating agent, has not been as effective in reducing the body iron burden and may worsen the fibrosis.295 Further studies are needed to better define the role of orally active iron chelators in thalassemic patients with iron overload.

ENDOCRINE DISEASES (see Table 35-5) DIABETES MELLITUS

Gastrointestinal symptoms are more prevalent in diabetic patients in comparison to the general population and have a negative effect on the quality of life.296,297 These symptoms are independently associated with poor glycemic control and peripheral neuropathy.298 Autonomic dysfunction in diabetics (diabetic autonomic neuropathy [DAN]) can manifest itself in one or more organ systems. The pathogenesis of DAN is related to hyperglycemia, neurovascular insuf­ ficiency, autoimmune damage, and neurohormonal growth factor deficiency.299 In type 1 diabetes, enteric neurotransmission may be modulated by a functional IgG autoantibody that acts as an agonist at the L-type calcium channels of smooth muscle of the colon.300 Constipation, abdominal pain, nausea, vomiting, dysphagia, diarrhea, and fecal incontinence are symptoms of enteric DAN that are more commonly seen in older patients with long-standing type 1 diabetes, poor glucose control, and symptoms of cardiovascular or peripheral neuropathy.301 Although motility disturbances are common in these patients, they do not correlate well with the presence or severity of symptoms. This suggests that other manifestations of DAN may play a role in the development of symptoms.

Esophageal Dysfunction (see Chapter 42)

Diabetic patients have been shown to have numerous esophageal motility abnormalities (although these may be clinically silent) such as hypotensive lower esophageal pressure, decreased amplitude of contractions, and simultaneous prolonged aperistaltic contractions in the body of the esophagus (see Chapter 42).301 Esophageal dysmotility in diabetes has been attributed to DAN mediated by vagal nerve dysfunction but recent motor nerve conduction studies suggest a motor neuropathy.302 Esophageal scintigraphy has demonstrated prolonged esophageal transit time.303 GERD is seen more frequently in diabetics and associated with the car­ diovascular autonomic dysfunction, increased body mass index, disease duration, and poor glycemic control.304,305 Odynophagia in a diabetic patient should suggest possible Candida infection (see Chapter 45).306

Gastric Dysfunction

Abnormal gastric motility (see also Chapter 48) results in disordered gastric emptying, or gastroparesis diabeticorum (GD), which affects as many as 30% to 60% of diabetic patients.307 In this disorder, the normal physiology of gastric emptying, largely under the control of the vagus nerve, is grossly disturbed. Liquid emptying may be normal, but solid emptying is frequently delayed. There is an increased frequency of postcibal antral dysrhythmias such as antral tachygastria. Phase 3 contractions of the IMMC, which normally stimulate antral contractions, are frequently absent, resulting in poor antral expulsion of indigestible solids, predisposing to bezoars. Furthermore, maintenance of the gastroduodenal pressure gradient, as well as receptive relaxation of the stomach, is abnormal. Prolonged pyloric contractions (pylorospasm) may cause functional resistance to gastric outflow.301 The pathophysiology of these motor disturbances is unclear. Hyperglycemia can cause delayed gastric emptying in diabetic patients as well as in normal volunteers.308 As noted, vagal parasympathetic function, which is involved in gastric emptying, may not be entirely normal. High plasma levels of the gut peptide motilin are reported in patients with GD.309 Because motilin stimulates the initiation of phase 3 activity, the elevation of this

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases Table 35-5  Gastrointestinal Manifestations of Endocrine Diseases DISEASE

ABNORMALITY/ASSOCIATION

GASTROINTESTINAL MANIFESTATIONS

Hyperthyroidism

Lymphocytic mucosal infiltrates Accelerated intestinal transit Ulcerative colitis Minor histologic changes in liver Rare chronic hepatitis (with thyroiditis) Impaired LES function Gastric hypomotility Decreased intestinal transit

Superficial gastritis, steatorrhea Diarrhea Bloody diarrhea Aminotransferase(s) elevation, mild indirect hyperbilirubinemia Increased ALP, aminotransferase(s) Reflux esophagitis Bezoars Constipation, fecal impaction, volvulus, pseudo-obstruction, rectal prolapse, diarrhea, steatorrhea (bacterial overgrowth)

Hypothyroidism

Liver biochemical test abnormalities Primary biliary cirrhosis Cronkhite-Canada syndrome Celiac disease Familial polyendocrine failure MCT

Increased serum calcitonin MEN-IIA, MEN-IIB

Adrenal insufficiency

Corticosteroid deficiency Familial polyendocrine failure

Pheochromocytomas

Increased plasma catecholamines Cholelithiasis MEN-IIA Increased pituitary ACTH Increased pituitary GH Adrenal insufficiency/hypothyroidism Increased serum calcium Peptic ulceration Pancreatitis MEN-I Familial polyendocrine failure

Hypercortisolism (Cushing’s disease) Acromegaly Panhypopituitarism Hyperparathyroidism

Hypoparathyroidism

Diabetes mellitus

Malabsorption Intestinal lymphangiectasia Esophageal dysmotility Esophageal candidiasis Gastroparesis Small intestinal dysmotility Impaired intestinal fluid reabsorption Colonic dysmotility Intestinal ischemia Pancreatic disease Cholelithiasis Sclerosing cholangitis Hepatic steatonecrosis Hepatocellular carcinoma Diabetic radiculopathy Familial polyendocrine failure Celiac disease

Features of hepatic cirrhosis Intestinal polyps Diarrhea, steatorrhea Esophageal candidiasis, adrenal insufficiency, hypogonadism, diabetes mellitus, hypothyroidism Watery diarrhea (increased intestinal secretion due to calcitonin?) Pheochromocytoma (see below), mucosal neuromas, ileus, megacolon Nausea, vomiting, anorexia, diarrhea, malabsorption Esophageal candidiasis, hypothyroidism, hypogonadism, diabetes mellitus, hypoparathyroidism Paralytic ileus, megacolon Biliary pain, cholecystitis MCT (see above) Gastric ulceration Colorectal polyps As for adrenal insufficiency and hypothyroidism Constipation, nausea, vomiting Bleeding, abdominal pain, perforation Acute pancreatitis Gastrinoma, VIPoma, others Esophageal candidiasis, hypothyroidism, hypogonadism, diabetes mellitus, adrenal insufficiency Diarrhea, steatorrhea Protein-losing enteropathy Dysphagia, reflux esophagitis Odynophagia, dysphagia Nausea, vomiting, gastric outlet obstruction, bezoars Bacterial overgrowth, malabsorption, diarrhea “Diabetic” diarrhea Constipation, megacolon, fecal incontinence Ischemic colitis, bowel infarction Acute pancreatitis, pancreatic carcinoma Biliary sepsis Biliary obstruction, sepsis Abnormal liver biochemical tests, hepatic fibrosis 2.5-fold increased risk Abdominal pain Candidiasis, hypothyroidism, hypogonadism, hypoparathyroidism, adrenal insufficiency Diarrhea, steatorrhea

ACTH, adrenocorticotropic hormone; ALP, alkaline phosphatase; GH, growth hormone; LES, lower esophageal sphincter; MCT, medullary carcinoma of the thyroid; MEN, multiple endocrine neoplasia; VIP, vasoactive intestinal polypeptide.

peptide in diabetic patients with GD may, in part, be compensatory. This is consistent with the observation that the treatment of GD with prokinetic agents is associated with a fall in plasma motilin levels.309 Epigastric discomfort, nausea, vomiting, pyrosis, early satiety and weight loss are symptoms associated with GD that typically are most severe postprandially. Only abdominal bloating or fullness has been shown to be an independent predictor of delayed gastric emptying.310 Markedly delayed gastric emptying may make the regulation of blood glucose levels difficult. Hyperglycemia further impairs

gastric emptying and may accelerate the onset of diabetic ketoacidosis, particularly when it is associated with severe vomiting. Although many diabetics have abnormal gastric emptying, few develop overt clinical symptoms. Furthermore, an occasional patient may have symptoms suggestive of GD but little or no delay in gastric emptying. The diagnosis of GD should be strongly suspected from the history. Physical examination may reveal gastric dilation with a succussion splash. A saline load test is not a sensitive test in GD because liquid emptying is frequently normal. The usual method for diagnosis is exclusion of

573

574

Section IV  Topics Involving Multiple Organs structural lesions by esophagogastroduodenoscopy or by standard barium radiographic examination. Food remnants may be noted in stomach. Such studies should be followed by more quantitative measurements of the degree of delay. Radiolabeled scintigraphy is the preferred way to confirm the diagnosis and to quantify the response to therapy.311 When interpreting these studies, it should be noted that anticholinergics, tricyclic antidepressants, benzodiazepines, and ganglionic-blocking agents may contribute to delayed emptying in these patients. In July 2006, the U.S. Food and Drug Administration (FDA) approved the SmartPill Capsule (SmartPill, Buffalo, NY) for evaluation of gastric emptying by measurement of luminal pressure, pH, and temperature data after ingestion of a radiotelemetry capsule. The management of GD requires a multimodal approach. Most importantly, glycemic control should be optimized. Dietary changes include a low-fat, low-fiber soft diet with frequent small meals. Sometimes patients may require a mostly liquid diet, at least temporarily. Antiemetics and prokinetics are the two primary classes of medical therapy for GD. Antiemetics such as promethazine or prochlorperazine can be given orally or in suppository form but longterm use is limited because of side effects. Other treatments for nausea and vomiting include scopolamine patch, 5-hydroxytryptamine (5-HT3) receptor antagonists such as odansetron, dronabinol, or low-dose tricyclic antide­ pressants to modify visceral hypersensitivity.312 Prokinetic agents, which increase gastric motor activity, are frequently used to treat GD. Metoclopramide (10 to 20 mg, 30 minutes before each meal and at bedtime) and domperidone (10 to 30 mg four times a day also given 30 minutes before meal and bedtime) are dopamine antagonists that increase antral contractions and decrease receptive relaxation of the proximal stomach.312 Metoclopramide crosses the blood-brain barrier whereas domperidone does not. Domperidone is not approved for use by the FDA—40% of patients cannot tolerate metoclopramide because of CNS side effects, and 5% of patients taking domperidone develop symptoms of hyper­ prolactinemia such as gynecomastia in men and breast enlargement and lactation in women.312 A third prokinetic drug is cisapride which, although efficacious in GD, has been severely restricted due to life-threatening proarrhythmic cardiac side effects. The macrolide antibiotic erythromycin, a motilin agonist, has been found to be effective in accelerating gastric emptying in GD but data regarding symptomatic relief are limited.313 Erythromycin can be given orally (125 mg two or three times daily) or intravenously (200 mg over 5 to 10 minutes every eight hours) but its use limited by nausea and abdominal cramping as well as loss of effectiveness over time secondary to downregulation of the motilin receptor.312 In a small pilot study, the infusion of ghrelin, an appetite stimulating hormone synthesized within the stomach, was shown to improve gastric emptying.314 Small case series suggested that endoscopic therapy with injection of botulinum toxin into the pyloric sphincter resulted in improvement in both gastric emptying and subjective symptoms; however, a double-blind randomized controlled trial showed no significant benefit.315 In severe or refractory cases, a venting gastrostomy and feeding jejunostomy tube can be placed. Surgical therapy had been limited to partial or complete gastric resection in medically refractory cases, with often disappointing results. Enterra (Medtronic, Minneapolis, Minn), a gastric electrical stimulator (GES) device, was approved in 2000 for treatment of refractory gastroparesis. At laparoscopy or laparotomy, two electrodes are placed into the muscularis propria of the greater curvature 10 cm from the pylorus and attached to a

neurostimulator placed subcutaneously in the abdominal wall. The GES device delivers high-frequency (12 cycles per minute [CPM]), low-energy pacing, with 81% and 63% reduction in vomiting frequency at 6 and 12 months, respectively, in patients with GD.316 The most common significant complication was pacemaker hardware infection, which was seen in 5% of patients and required GES device removal. In long-term follow-up of up to five years, GES device placement has been shown to improve glycemic control and nutritional parameters, enhance quality of life, and decrease health care costs.316 Acute erosive gastritis is common in diabetic ketoacidosis and is frequently accompanied by bleeding. A postulated association between diabetes and H. pylori has been called into question.317 The incidence of duodenal ulcer in diabetes is lower than expected. Autoimmune chronic gastritis and gastric atrophy also may be seen with long-standing diabetes. In type 1 diabetes, 15% to 20% of patients have serologic evidence antiparietal cell antibodies, and this subset of patients has an increased prevalence of autoimmune gastritis with pernicious anemia, iron deficiency anemia, hypochlorhydria, and hypergastrinemia.318

Diabetic Diarrhea

Diarrhea is a common symptom of autonomic neuropathy, affecting 3.7% of diabetic patients, predominantly those with type 1 diabetes (see also Chapter 15).319 A common cause of diarrhea in diabetic patients is drug therapy, but diarrhea or increased stool frequency may occur because of coexistent celiac disease, pancreatic insufficiency, bacterial overgrowth, consumption of artificial sweeteners, islet cell tumors or fecal incontinence.320 Metformin, a biguanide derivative with structural similarity to 5-HT3–receptor agonists, is associated with diarrhea, usually with the initiation of treatment but also late in therapy.321-323 Extended release metformin is less likely to cause diarrhea than immediate release metformin and may be an alternative to discon­ tinuing metformin therapy.324 Acarbose and miglitol are α-glucosidase inhibitors that competitively inhibit the breakdown of oligo- and disaccharides to monosaccharides in the small intestinal brush border. Thirty percent of patients treated with acarbose develop abdominal discomfort, flatulence, and diarrhea.325 Acarbose may cause diarrhea by an increase in colonic butyrate production, which increases prostaglandin E production leading to water and electrolyte loss.326 True diabetic diarrhea frequently affects patients with dyspeptic symptoms and GD, and it appears to be more common in men than in women. It may be particularly troublesome at night. Diabetic diarrhea occurs mostly in patients with poorly controlled type 1 diabetes who also have evidence of diabetic peripheral and autonomic neuropathy. Associated steatorrhea is common and does not necessarily imply a concomitant gastrointestinal or pancreatic disease. The diarrhea is often intermittent and painless, and it may alternate with periods of normal bowel movement or with constipation. The pathogenesis of diabetic diarrhea is unclear. Marked abnormalities are found in the motor pattern of the small intestine. Phase 3 contractions during the IMMC are shorter, and phase 2 activity of the stomach and upper small intestine is abnormal. No significant differences between diabetic patients and control subjects, however, can be observed in mouth-to-cecum or whole-gut transit times. In patients treated with prokinetic agents, fasting IMMC and fed motor patterns in the small intestine may be normalized, but the symptomatic improvement of diarrhea is no better than with placebo.327 Sympathetic denervation of the gut is common

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases in diabetic patients with autonomic neuropathy. Because adrenergic nerves normally stimulate intestinal absorption of fluids and electrolytes, decreased intestinal absorption, rather than intestinal dysmotility, may underlie the pathogenesis of diabetic diarrhea. The management of diabetic diarrhea is difficult, but strict control of blood glucose levels may help. Because gastroi­ ntestinal adrenergic function is impaired in autonomic neuropathy, adrenergic agonists may stimulate intestinal absorption of fluids and electrolytes. In addition, they may partially correct the motility disturbances of DAN. The α2-adrenergic agonist clonidine (0.1 to 0.6 mg twice daily) may be successful in the therapy of diabetic diarrhea, presumably by reversing the peripheral adrenergic resorptive abnormalities. Because the antihypertensive effects of clonidine are mediated through the CNS, diabetic patients with severe autonomic neuropathy may not necessarily experience worsening of preexisting postural hypotension during therapy. Clonidine does not alter diabetic control or renal function. If the medication needs to be withdrawn, it should be done slowly to avoid “rebound” hypertension. The long-acting somatostatin analog octreotide (50 to 100 µg subcutaneously, twice daily) may be used in the treatment of refractory diabetic diarrhea.328 It may, however, predispose to intestinal bacterial overgrowth owing to decreased small bowel transit time, and it may aggravate steatorrhea by inhibiting pancreatic exocrine function. Symptomatic measures that may be used include the prescription of codeine sulfate (30 mg every six to eight hours), diphenoxylate with atropine (Lomotil), or loperamide. In some patients, psyllium hydrophilic mucilloid may be helpful.

Fecal Incontinence

A troublesome symptom of DAN is fecal incontinence (see Chapter 17). Incontinence often coincides with the onset of diabetic diarrhea, but in most cases the total stool volume is normal. Steatorrhea is present in as many as 30% of cases.329 Autonomic dysfunction is thought to be respon­sible for the impairment of normal internal anal sphincter resting tone and reflexive internal sphincter relaxation. Primary management is empiric including antidiarrheal therapy and biofeedback training, whereas more severe cases may benefit from surgery or sacral nerve stimulation.330 In some patients incontinence remits spontaneously.

Constipation and Megacolon

The colon is frequently involved in diabetes mellitus. The most common gastrointestinal complaint of diabetics is constipation (see Chapter 18), related in some cases to autonomic neuropathy.331 Occasionally, severe constipation with megacolon may be encountered. Rarely, chronic intestinal pseudo-obstruction may result.332 High-fiber diets have not proved to be of great benefit, and anorectal myectomy has not been adequately evaluated. Complications of severe constipation include stercoral ulcer, perforation, volvulus, and anal overflow diarrhea. Treatment is aimed at symptomatic relief with enemas, laxatives, and cathartics.

Unexplained Abdominal Pain

Diabetic radiculopathy or diabetic plexus neuropathy of thoracic nerve roots may cause otherwise unexplained upper abdominal pain in patients with diabetic neuropathy. Pain may be associated with anorexia and weight loss, which mimics intra-abdominal malignancy. The diagnosis may be strengthened by an abnormal EMG of the anterior

abdominal wall muscles when compared with an EMG of thoracic paraspinal muscles.333

Biliary Tree and Liver

Cholelithiasis, cholecystitis, and cholangitis are thought to occur more frequently in diabetic patients. Lithogenic bile composition and stasis of bile in the gallbladder may contribute to stone formation in patients with diabetes, and it is generally thought that diabetic patients have an increased incidence of cholelithiasis. As with infections in general, hepatobiliary sepsis tends to be more severe in diabetic patients. In addition to severe bouts of cholecystitis and ascending cholangitis, unusual infections with gasproducing organisms and rare abscesses due to Yersinia enterocolitica have been reported.334,335 An increased incidence of sclerosing cholangitis also has been reported in diabetic patients.336 Despite the increased severity of cholecystitis and cholangitis in these patients, however, it is not recommended that diabetic patients with asymptomatic gallstones undergo prophylactic cholecystectomy. The most prominent hepatic complication of type 2 diabetes is nonalcoholic fatty liver disease, discussed in Chapter 82.

Pancreatic Disease

The prevalence of acute pancreatic disease (see Chapter 58) and pancreatic insufficiency (see Chapter 59) is increased in patients with diabetes. Acute pancreatitis is twice as frequent in young, type 1 diabetic patients. Acute pancreatitis causing diabetic ketoacidosis (DKA) has a particularly serious prognosis, with a high mortality rate.337 In the setting of DKA, nonspecific elevations (less than three times the upper limits of normal) in serum amylase and lipase occur in 16% to 25% of cases.338 The incidence of clinically apparent chronic pancreatitis, however, is not increased. Diabetes is a risk factor pancreatic cancer (see Chapter 59) and associated with increased mortality rate.339,340 Diabetes of new onset may also be an early sign of pancreatic cancer.

THYROID DISEASE Hyperthyroidism

Hyperthyroidism may underlie a number of important gastrointestinal symptoms, owing to its own effects on almost all organs of the gastrointestinal system. In addition, these symptoms sometimes occur in the absence of the cardinal features of hyperthyroidism (“apathetic” hyperthyroidism). Apathetic thyrotoxicosis may present with protracted abdominal pain, recurrent vomiting (thyrotoxic vomiting), marked weight loss, and altered bowel habits. Patients affected by thyroid storm may display a constellation of symptoms involving high fever, marked tachycardia, agitation, and delirium along with intestinal manifestations that include acute abdominal pain, vomiting, jaundice and severe diarrhea. Even in the absence of overt congestive heart failure or thyroid storm, jaundice, mild aminotransferase elevations (less than 250 IU/L) and prolonged prothrombin times may be observed.341 Hyperthyroidism clearly affects gastrointestinal motility. Excess thyroid hormone may cause myopathy, resulting in dysfunction of the striated muscles of the pharynx and the cervical esophagus. This is a potential mechanism that may explain dysphagia.342 Dysphagia is a rare manifestation of hyperthyroidism and can be readily reversible with correction of the thyrotoxic state.343 More than 25% of hyperthyroid patients have mild to moderate diarrhea. Intestinal transit time inversely correlates with thyroid hormone levels,344 whereas gastric emptying is not significantly increased with the hyperthyroid

575

576

Section IV  Topics Involving Multiple Organs state.344,345 Although hypermotility is the most likely explanation for diarrhea, thyroid hormone itself can induce secretory diarrhea by increasing intracellular cyclic adenosine monophosphate, akin to the actions of cholera toxin and VIP. Hyperthyroid-associated diarrhea (and steatorrhea, when present) readily responds to treatment with propylthiouracil. Treatment with propylthiouracil can lead to a euthyroid state with concomitant normalization of orocecal transit times and relief of gastrointestinal symptoms.346 The relationship between transient hyperthyroidism and hyperemesis gravidarum during pregnancy is discussed in Chapter 38. Infrequently, hyperthyroidism may coexist with ulcerative colitis. Hyperthyroidism may intensify the symptoms of ulcerative colitis, and it may impair the response to therapy.347

Hypothyroidism

Hypothyroidism is most commonly caused by an autoimmune mechanism or as a consequence of therapy for hyperthyroidism. It is occasionally seen in association with other diseases such as ulcerative colitis,347 pernicious anemia,348 and primary biliary cirrhosis.349 Hypothyroidism is seen in approximately 20% of patients with primary biliary cirrhosis.349 Rarely, celiac disease or diabetes mellitus is also associated with autoimmune thyroiditis. Hypothyroidism is associated with hypomotility of the gastrointestinal tract. Disturbances of esophageal peristalsis and LES function resulting in reflux and esophagitis may be seen with severe hypothyroidism. Replacement therapy can normalize sphincter tone and restore peristalsis. Hypothyroidism also may result in gastric and intestinal hypomo­ tility. In rare instances, phytobezoars may form and result in gastrointestinal obstructions.350 Severely impaired colonic motility may manifest with constipation, obstipation, sigmoid volvulus, rectal prolapse, fecal impaction, and rarely megacolon. Hypothyroidism can be a cause of ileus and should be considered as an etiology of pseudoobstruction.351 Diarrhea, although rare in hypothyroidism, may be due to bacterial overgrowth from bowel hypomo­ tility. Antibiotic treatment can result in resolution of diarrheal symptoms.352 Myxedema has also been found in association with Cronkhite-Canada syndrome (see Chapter 122).353

Medullary Carcinoma of the Thyroid

Medullary carcinoma of the thyroid (MCT) is a calcitoninproducing tumor of the C cells of the thyroid gland. Diarrhea is seen in one third of patients with MCT. Diarrhea may occur presumably due to the effects of high circulating calcitonin on the gut.354 MCT also may produce VIP and prostaglandins that contribute to diarrhea. Decreased colonic transit time due to as yet unknown humoral agents may also underlie the diarrhea of MCT.355 MCT is also associated with multiple endocrine neoplasia (MEN) syndromes IIA and IIB. These syndromes can be complicated by hyperparathyroidism and pheochromocytomas and in MEN-IIB with mucosal neuromas.

ADRENAL DISEASE

Adrenal insufficiency, or Addison’s disease, is associated with gastrointestinal symptoms or pathology in more than half of cases. A constellation of symptoms including anorexia, weight loss, nausea, vomiting, diarrhea, and abdominal pain may be present. Patients with Addison’s disease may also present with chronically elevated aminotransferase levels.356 Cyclical vomiting in children may rarely be due to adrenal insufficiency.357 Malabsorption and

diarrhea seen in some patients with Addison’s disease are apparently due to functional defects in enterocytes that can be readily reversed with the administration of glucocorticoids. Atrophic gastritis, achlorhydria, and pernicious anemia may be present in association with autoimmune Addison’s disease. Pheochromocytomas are tumors arising from the adrenal medulla and chromaffin tissue that secrete high levels of catecholamines, leading to hypertension. The humoral effects of high circulating levels of catecholamines may result in ileus or pseudo-obstruction.358 Gastrointestinal manifestations of pheochromocytoma also include ischemic colitis, diarrhea, acute abdominal pain, and, rarely, gastrointestinal bleeding.359 For unclear reasons, pheochromocytoma is associated with an increased incidence of cholelithiasis. Some patients also have MEN-IIA or MEN-IIB syndrome (see earlier).

PITUITARY DISEASE

Pituitary disorders infrequently affect the gastrointestinal tract, except in association with MEN-I syndrome (see Chapter 32). Hypercortisolism, caused by the inappropriate secretion of corticotropin in Cushing’s disease, may be associated with an increased incidence of gastric ulceration when concomitant NSAIDs are used.360 Panhypopituitarism may present with addisonian crisis, with hypotension, nausea, vomiting, abdominal pain, and diarrhea. The excessive secretion of pituitary growth hormone with concomitant elevation of insulin-like growth factor I results in acromegaly. The incidence of adenomatous colonic polyps may be increased in patients with acromegaly and the adenomas tend to be larger, multiple, and right-sided.361 The risk of colon cancer is approximately two-fold higher in acromegaly, but screening recommendations have been contentious.362

PARATHYROID DISEASE Hyperparathyroidism

Gastrointestinal problems are common in patients with hyperparathyroidism.363 Most common complaints are constipation, diffuse abdominal discomfort, or nausea and vomiting. A minority (5% to 15%) have peptic ulcer disease, and a small percentage (1% to 2%) develop pancreatitis. Remission of pancreatitis after parathyroidectomy has been reported.364 Severe pancreatitis may also occur immediately following parathyroidectomy has been reported.365 Gastrointestinal symptoms associated with hypercalcemia include nausea, vomiting, anorexia, and abdominal pain. Some patients with hyperparathyroidism have MEN-I or MEN-II syndrome (see earlier).

Hypoparathyroidism

Hypoparathyroidism with hypocalcemia may be associated with malabsorption and mild to moderate steatorrhea. Constipation, and in rare instances, even pseudoobstruction may be important gastrointestinal disturbances in this disease. In the familial polyendocrine failure syndrome (candidiasis, endocrinopathy, or polyendocrine autoimmune disease, type 1), patients have hypoparathyroidism, adrenal insufficiency, hypogonadism, and, in many cases, diabetes mellitus. From 4% to 29% also have malabsorption. Varying degrees of gastric atrophy with antiparietal cell antibodies, hypochlorhydria, autoimmune hepatitis, dental enamel hypoplasia, and severe oral and esophageal candidiasis are also seen.366 Intestinal lymphangiectasia with protein-losing enteropathy also has been reported in association with malabsorption and hypoparathyroidism.367

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases Idiopathic hypoparathyroidism may also coexist with celiac disease likely due to autoimmune reactivity. When this occurs, a gluten-free diet may lead to the disappearance of parathyroid immunoreactivity.368

increased endogenous cholesterol synthesis has also been reported.374 The intestinal mucosa may appear yellowish on endoscopy, reflecting the presence of mucosal lipid.375 Therapy consists of substituting medium-chain for longchain triglycerides.

TANGIER DISEASE

DISORDERS OF LIPID METABOLISM HYPERLIPOPROTEINEMIAS AND DYSLIPIDEMIAS

In familial hyperchylomicronemia (type I phenotype), the plasma is lactescent, with marked elevation of chylomicrons and triglycerides due to a deficiency of lipoprotein lipase. Manifestations can appear early in life and include recurrent episodes of abdominal pain, fever, peritonitis, and pancreatitis (see Chapter 58). In most patients the cause of recurrent attacks of pain is not known.369 Patients with familial hyperbetalipoproteinemia (type IV phenotype) suffer from premature atherosclerosis, hyperuricemia, and attacks of pancreatitis that generally occur when plasma triglyceride values are greater than 2000 mg/dL.370 The hyperlipidemia may mask elevated plasma amylase values. Type IV patients also have an increased incidence of cholelithiasis and cholecystitis.371 Patients with familial hyperlipoproteinemia (type V phenotype) are prone to bouts of abdominal pain, with or without pancreatitis. Exacerbation of endogenous hypertriglyceridemia by diabetes, diet, alcohol, or medications can also cause pancreatitis.370

ABETALIPOPROTEINEMIA

Abetalipoproteinemia is an autosomal recessive disorder characterized by acanthotic erythrocytes, serum lipid abnormalities, ataxia, atypical retinitis pigmentosa, and steatorrhea.372 The typical laboratory feature is complete absence in plasma of all lipoproteins containing apolipoprotein B: chylomicrons, low-density lipoprotein (LDL), and verylow-density lipoprotein (VLDL). The histologic appearance of the small intestine after an overnight fast is marked by mucosal epithelial cells loaded with lipid droplets (Fig. 35-9).373 By contrast, the submucosa and lamina propria show practically no lipid, and the lymphatics are empty. The villi are normal in length and configuration. Mild steatorrhea with onset during the first 2 years of life is seen (see Chapter 101). Cholesterol malabsorption with

Figure 35-9.  Small intestinal biopsy specimen obtained from a patient with abetalipoproteinemia. The biopsy specimen is notable for the accumulation of lipid droplets within the intestinal epithelial cells.

Tangier disease is an autosomal recessive disorder characterized by accumulation of cholesterol esters in macrophages in tonsils, thymus, lymph nodes, marrow, liver, and the gut. Tangier disease is caused by a mutation in the adenosine triphosphate-binding cassette protein, ABCA1, which mediates the efflux of excess cellular sterol to apolipoprotein A-I (apo A-I), a step leading to the formation of beneficial high-density lipoprotein (HDL).376 These patients have very low levels of plasma cholesterol and HDLs, owing to a lack of apo A-I. The gene encoding apo A-I is normal in Tangier disease, but a defect in post-translational processing results in rapid degradation of apo A-I.376 The striking clinical findings include yellow-orange “streaked” tonsils in 80% of cases, hepatosplenomegaly, and peripheral neuropathy. Patients may have diarrhea without steatorrhea. Colonoscopy reveals orange-brown mucosal spots throughout the colon and rectum, and laparoscopy reveals similar yellow patches on the surface of the liver due to cholesterol esters in hepatic reticuloendothelial cells.377

NEUTRAL GLYCOSPHINGOLIPIDOSES

Fabry’s disease is an X-linked disorder of glycolipid metabolism due to the deficiency or absence of the enzyme α-galactosidase A, resulting in globotriaosylceramide deposition in many tissues and subsequent organ dysfunction. Impaired motility is the prominent gastrointestinal abnormality.378 Electron microsopic examination of biopsy specimens from the small intestine and rectum reveals large sphingolipid-filled vacuoles in the ganglion cells of Meissner’s plexus within smooth muscle cells of the muscularis mucosa and within endothelial cells lining the blood vessels. Mucosal enterocytes are normal.378 The prevalence of gastrointestinal symptoms was 52%, with abdominal pain and diarrhea being the most frequent symptoms.379 Delayed gastric emptying, bacterial overgrowth, increased fecal bile acid loss, cholelithiasis, and jejunal diverticulosis with perforation have been documented. Successful treatment of the gastrointestinal component of this disorder with metoclopramide and tetracycline has been reported. Glycolipid deposition in small vessels can induce severe vasculitis and thrombosis, resulting in ischemic bowel lesions. Ileal perforation also has been reported.380 Thirty percent to 60% of obligate carrier women and 60% of men have nonspecific gastrointestinal symptoms that can improve with agalsidase alfa enzyme replacement therapy.381 Gaucher’s disease is a rare, usually autosomal recessive deficiency of the enzyme acid β-glucosidase resulting in the deposition of glucosylceramide in the cells of the reticuloendothelial system, including the liver and spleen. In the adult form of the disease gastrointestinal complications predominate, including hepatosplenomegaly, hepatic cirrhosis, ascites, and esophageal variceal hemorrhage.382,383 Niemann-Pick disease is a rare autosomal-recessive disease with a predilection for Ashkenazi Jews. Types A and B result from defects in sphingomyelinase. In type C Niemann-Pick disease, mutation in Niemann-Pick protein C results in defective transport of cholesterol across the lysosomal membrane.384 Sphingomyelinase deficiency results in the deposition of sphingomyelin and cholesterol in the liver and spleen, the CNS, and the lungs and skin. Gastrointestinal complications include hepatosplenomegaly and liver

577

578

Section IV  Topics Involving Multiple Organs failure. Rare adult survivors of Niemann-Pick disease have been reported.385

RENAL DISEASES Upper gastrointestinal tract symptoms are common in patients with chronic kidney disease (CKD) who require peritoneal dialysis (PD) or hemodialysis (HD). Anorexia, singultus (hiccups), nausea, vomiting, epigastric pain, and heartburn are common manifestations of azotemia. Delayed gastric emptying is common in CKD.386 Although the prevalence of peptic ulcer is only 2%, which is not significantly different from that in the general population, gastritis, duodenitis, and mucosal erosions are commonly seen.387 A variety of data suggests that neither hyperacidity, hypergastrinemia, nor H. pylori plays major roles in the pathogenesis of uremic gastropathy, although these data have recently been called into question.388 Impaired mucosal cytoprotection has been postulated but not proved. Also seen on esophagogastroduodenoscopy are esophagitis, Brunner’s gland hyperplasia, gastric fold thickening, and nodular duodenitis (which resolves after renal transplantation) and angiodysplasia.389-391 GERD may be related to the absence of H. pylori infection, amyloidosis, and PD, which increases intra-abdominal pressure.392 In controlled studies, the incidence of gallstones in CKD is similar to healthy controls.393 It is possible that angiodysplastic lesions in the upper and lower gastrointestinal tract are no more common in patients with CKD than in the general population but are discovered more frequently because of their greater tendency to bleed. Angiodysplasia in renal failure is acquired lesions formed by repeated episodes of submucosal venous outflow obstruction resulting in incompetent precapillary sphincters with subsequent arteriovenous communication.394 Angiodysplastic lesions are much more likely to bleed in patients with CKD than in patients with normal renal function,395 perhaps because of uremic platelet dysfunction. In a series of CKD patients with upper gastrointestinal hemorrhage, gastric ulcer (37%) and duodenal ulcer (23%) were the two most common bleeding lesions, but angiodysplasia of the upper gastrointestinal tract was the cause of bleeding in 13%. In contrast, angiodysplasia was only responsible for 1.3% of upper intestinal tract bleeding in control patients. In CKD patients with recurrent hemorrhage, angiodysplasia was the most frequent cause of bleeding. Angiodysplasia as a cause of bleeding was most closely associated with the duration of renal failure and the need for hemodialysis.396 In CKD, peptic lesions may be managed successfully with standard medical treatments in appropriate “renal” doses (see Chapter 53), and angiodysplasia may be treated with laser, electrocoagulation, or surgery (see Chapters 19 and 36). Small intestinal complications of CKD include ileus, ulceration, and nonocclusive ischemic bowel disease.397-399 Diarrhea may occur secondary to bacterial overgrowth related to abnormal small intestinal motility.400 In addition, exocrine pancreatic insufficiency has been documented in a number of hemodialysis patients.401 The cause of the condition is not known, but patients may improve clinically with pancreatic enzyme replacement. Patients with CKD appear more likely to develop colonic perforation from ruptured diverticula, fecalomas (secondary to the use of aluminum-containing antacids or barium), or cecal ulcers that may bleed profusely.402 Life-threatening hemorrhage from rectal ulcers also has been reported.403 Colonic intussusception and ileus are also encountered in

CKD. The diarrhea experienced by some patients with CKD appears to be related to abnormal bile acid metabolism.404 Ischemic colitis in patients receiving HD tends to be more right-sided in anatomic distribution, which is associated with poor outcome.405 See Chapter 34 for a discussion of gastroin-testinal and hepatic complications associated with renal transplantation.

NEUROLOGIC DISEASES Because of the importance of nerves and neurotransmitters on gastrointestinal function (see Chapter 1), it is not surprising that neurologic diseases are frequently associated with gastrointestinal symptoms. Some of the more common disorders affecting the CNS (brain and spinal cord), cranial nerves, autonomic nervous system, neuromuscular junction, and musculature are presented in Table 35-6 and in the following sections.

NEUROGENIC ABDOMINAL PAIN

(see Chapters 10 and 11) Neurogenic causes of abdominal pain originate in either the CNS or peripheral nervous system. Central neurogenic abdominal pain can result from abdominal migraines. Although classic migraine headaches are often associated with nausea and vomiting, abdominal migraine is a migraine variant characterized by recurrent gastrointestinal symptoms, including vomiting and epigastric pain.406 It is often seen in children and is not always associated with headache. The pathophysiology of the gastrointestinal symptoms is unclear, and some have questioned the validity of this controversial diagnosis. In one Greek study, most pediatric patients with abdominal migraine had evidence of esophagitis, gastritis, or duodenitis at endoscopy.407 A more recent study of 53 children with abdominal migraine and no underlying gastrointestinal pathology demonstrated symptomatic improvement or successful prophylaxis using standard antimigraine therapies such as propranolol or cyproheptadine.408 Adults with migraine do not tend to report abdominal pain,409 although some investigators believe that a subset of adult patients with recurrent nonorganic abdominal pain may suffer from abdominal migraine. Unfortunately, the lack of a precise definition of abdominal migraine has made research in this area difficult. Abdominal epilepsy is an uncommon cause of central neurogenic abdominal pain. One retrospective study describes the spectrum and clinical course of 10 patients with abdominal epilepsy.410 These patients, each with temporal lobe electroencephalographic (EEG) abnormalities, experienced a variety of paroxysmal gastrointestinal symptoms that included periumbilical and right upper quadrant pain, bloating, and diarrhea. All of the patients also experienced CNS symptoms such as headaches, confusion, dizziness, syncope, or blindness, occurring daily in association with the gastrointestinal complaints. Anticonvulsant therapy resulted in resolution of gastrointestinal and CNS disturbances. Even patients with classic epileptic seizure disorders may experience viscerosensory auras (vague epigastric sensations and nausea). Although they can occur in patients with temporal lobe mass lesions, epigastric auras are more frequently associated with hippocampal sclerosis. These epigastric auras can be unpleasant or even debilitating. Unfortunately, 25% of patients who are rendered seizure free after anteromedial temporal lobe resection con-

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases Table 35-6  Gastrointestinal Manifestations of Neuromuscular Diseases location/DISEASE Cerebrum/Cerebellum Cerebrovascular accident Multiple sclerosis Cerebral palsy Migraine headache “Abdominal epilepsy”/viscerosensory auras Pseudotumor cerebri Brainstem/Cranial Nerve Cerebrovascular accident, other brainstem disorders Multiple sclerosis Pseudobulbar palsy Diphtheria Spinal Cord/Peripheral Nerve Spinal cord injury/transection Amyotrophic lateral sclerosis Charcot-Marie-Tooth syndrome Tabes dorsalis Poliomyelitis Alcoholic and amyloid neuropathy Extrapyramidal/Autonomic Parkinson’s disease Huntington’s chorea Familial dysautonomia (Riley-Day syndrome) Shy-Drager syndrome Chagas’ disease Paraneoplastic neuropathy Ganglioneuromatosis Diabetic neuropathy Neuromuscular Junction Myasthenia gravis Muscle Disease Stiff-man syndrome Oculopharyngeal muscular dystrophy Mitochondrial neurogastrointestinal encephalomyopathy Duchenne’s muscular dystrophy Familial visceral myopathy Myotonic dystrophy Polymyositis/dermatomyositis

GASTROINTESTINAL MANIFESTATIONS Oropharyngeal dysphagia, gastroparesis, constipation, peptic ulceration, anorectal dysfunction Oropharyngeal dysphagia, gastroparesis, constipation, anorectal dysfunction Oropharyngeal dysphagia Nausea, vomiting, abdominal pain (abdominal migraine), peptic ulceration Abdominal pain, bloating, diarrhea Nausea, vomiting Oropharyngeal dysphagia, dysgeusia Oropharyngeal dysphagia, dysgeusia Oropharyngeal dysphagia Oropharyngeal dysphagia Gastroparesis, constipation, incontinence, megacolon, ileus, autonomic dysreflexia Oropharyngeal dysphagia, ileus Oropharyngeal dysphagia, delayed gastric emptying Abdominal pain crises, diarrhea Ileus, gastric atony, megacolon Esophageal and gastrointestinal dysmotility Oropharyngeal dysphagia, constipation, fecal incontinence Oropharyngeal dysphagia, gastroparesis, constipation Esophageal dysmotility, vomiting crises, gastric atony, diarrhea, megacolon Postprandial orthostatic hypotension, esophageal dysmotility, achlorhydria, constipation Achalasia, megaesophagus, megaduodenum, megacolon Achalasia, megacolon, intestinal pseudo-obstruction Constipation, megacolon See Table 35-5 Oropharyngeal dysphagia, autoimmune hepatitis, primary biliary cirrhosis Oropharyngeal dysphagia Oropharyngeal dysphagia Dysmotility, achalasia, malabsorption, pseudo-obstruction, diarrhea Oropharyngeal dysphagia, gastric atony, malabsorption, megacolon, pseudo-obstruction Dysphagia, pseudo-obstruction Oropharyngeal dysphagia, esophageal dysmotility, gastric atony, megacolon, pseudoobstruction, volvulus, gallbladder dysfunction See Table 35-1

tinue to have persistent epigastric auras.411 Although epilepsy with gastrointestinal symptoms is uncommon, an EEG should be considered in the diagnostic workup of patients with unexplained paroxysmal gastrointestinal complaints associated with CNS symptoms. Peripheral neurogenic abdominal pain emanating from peripheral nerves or spinal nerve roots is usually intermittent and sharp. It is not associated with food intake or abdominal distention and is usually easy to diagnose. Possible causes include infections such as herpes zoster (pain may precede the rash), and syphilis (gastric crisis in tabes dorsalis); inflammatory conditions such as PAN; metabolic disturbances such as diabetic neuropathy (see earlier); toxic ingestions such as lead poisoning; and nerve root impingement due to osteoarthritis, tumors, or herniated disks.

GASTROINTESTINAL COMPLICATIONS OF ACUTE HEAD INJURY AND STROKE

Increased intracranial pressure from any cause may lead to episodic projectile vomiting, which may precede other signs. In addition to direct effects of CNS injury on the control of oropharyngeal muscle movement during swallowing, acute head trauma and stroke (cerebrovascular

accidents) are associated with a high incidence of upper gastrointestinal tract pathology.412 In one prospective study, acute gastrointestinal erosive lesions were found at endoscopy in 75% of patients. Erosive gastritis was seen in 69% of cases, gastric ulcer in 23%, esophagitis in 11%, and duodenal inflammation in 8%.413 Most lesions were present within a week of injury. There was no correlation between glucocorticoid administration and development of upper gastrointestinal tract lesions. Although these lesions fall within the spectrum of stress gastropathy (see Chapters 52 and 53), an additional pathogenetic mechanism may play a role in their development. Serum gastrin levels are elevated in patients with head injury,414 presumably through a direct neurogenic reflex, and gastric hypersecretion has been reported.415 Because these injuries are largely preventable, patients with acute neurologic injury due to trauma, severe strokes, neurosurgery, or some other condition should receive prophylactic antiulcer therapy (see Chapter 53).

GASTROINTESTINAL PROBLEMS AFTER SPINAL CORD INJURY

The effect of spinal cord injuries on the gastrointestinal tract depends on the level of the lesion. Delayed gastric emptying is seen in patients with cervical spinal cord injuries.416 In

579

580

Section IV  Topics Involving Multiple Organs the early postinjury period, severe gastric stasis, gastric dilation, and ileus are often present.417 Nasogastric suction is frequently required. Promotility agents such as metoclopramide can be effective because the enteric nervous system and smooth muscle layers are intact. A frequent problem in the first weeks after injury is peptic ulceration, although the mechanism is unclear. Upper gastrointestinal hemorrhage is more common with cervical cord injuries, with the use of oral anticoagulants, or when there is respiratory distress.418 Ulcer perforation and peritonitis may not be detected initially because of myelopathy involving sensory fibers or because of concomitant glucocorticoid therapy. When ulcer surgery is required, gastric resection or simple closure of the perforation is sufficient; truncal vagotomy is not performed because of the risk of severe gastric retention.419 Pancreatitis, another early complication of spinal injuries, also may be related to the effects of glucocorticoids.417,420 Autonomic dysreflexia, a life-threatening condition, sometimes affects patients whose lesion lies above the fifth thoracic root, the upper level of greater splanchnic flow. The pathophysiology involves an abnormal autonomic reflex that is initiated by fecal impaction or bladder distention and leads to severe hypertension and tachycardia.421 If untreated, seizures, subarachnoid hemorrhage, and stroke may result. Routine bladder catheterization and avoidance of constipation are preventive. Patients often face a different set of problems in the months to years following permanent spinal cord damage. With chronic loss of function, patients with quadriplegia are more likely to have gut complications than are patients with paraplegia. The incidence of gastroesophageal reflux is increased.417 There may be decreased bioavailability of orally administered drugs owing to impaired gastric emptying.422 Secondary amyloidosis involving the gastrointestinal tract is more common (see later), especially when chronic pyelonephritis and renal failure complicate spinal cord disease.417,423 Other complications include cholelithiasis, the superior mesenteric artery syndrome (see Chapters 14 and 36), hemorrhage due to solitary colonic ulcer, and the precocious appearance of diverticulosis.78,417 Many patients with spinal cord injury have marked impairment of their bowel function. Fecal incontinence and urgency can have a significant effect on quality of life.424,425 Chronic constipation plagues many patients with spinal cord injury. Damage to neurons in the spinal cord eliminates the sensation of rectal fullness and the voluntary control of defecation. There may be decreased splanchnic outflow, impairing the coordination of intestinal and colonic motility. Prolonged transit time may be explained by decreased colonic activity, colonic contractions, and intraluminal pressure. Further, the gastrocolic reflex after feeding may be regionally diminished in the rectosigmoid.426 Fortunately, the lower motor neurons of the second, third, and fourth sacral roots, which provide the sensory and motor fibers for the defecation reflex, are usually intact. Rehabilitation of bowel function is individualized to each patient’s disability. Physical exercise, adequate fluid intake, and stool softeners are prescribed, as for constipation of any origin. Most patients can learn to distend the rectum digitally on a regular schedule to initiate the defecation reflex. Stimulatory laxatives such as bisacodyl suppositories occasionally are necessary. It is not clear which patients, if any, may benefit from promotility agents. In the setting of major bowel dysfunction, colostomy may be a viable option.427 Patients with spinal cord injuries often have subtle symptoms and signs of colorectal emergencies that may lead to a delayed diagnosis and increased morbidity and mortality.428

DISEASES OF THE AUTONOMIC NERVOUS SYSTEM

Congenital and neurodegenerative diseases of the autonomic nervous system may affect the gastrointestinal tract and are listed in Table 35-6. These include familial dysautonomia, or the Riley-Day syndrome. Patients present at birth with feeding difficulties, poor temperature control, and motor incoordination. Common gastrointestinal symptoms include dysphagia, gastroesophageal reflux, and abnormal swallowing reflexes. A large percentage of patients require feeding gastrostomy or fundoplication. Lower gastrointestinal tract symptoms are less common, but patients may develop diarrhea as a result of decreased motility and bacterial overgrowth.429 Idiopathic autonomic neuropathy is a relatively uncommon acquired cause of autonomic neuropathy.429 It can affect sympathetic as well as parasympathetic function (pandysautonomia), or it may be limited to a parasympathetic deficit (cholinergic dysautonomia). The onset of symptoms can be acute or subacute. Cholinergic dysautonomia usually affects patients in their teens, whereas pandysautonomia manifests in early middle age.430 The etiology of this condition is unknown, but most cases develop after nonspecific viral infections. The syndrome also has been seen in association with mononucleosis,431 Stevens-Johnson syndrome,432 and herpes zoster or herpes simplex infections.433 In nearly 70% of cases, the initial manifestation of the disease is in the gastrointestinal tract. The most common symptoms are due to excessive cholinergic activity, such as diarrhea, hyperhidrosis, and hypersalivation. Although a large proportion of these patients has manometric abnormalities, complications of motility disorders such as bacterial overgrowth are usually not seen.429 A poorly understood autonomic condition is postprandial orthostatic hypotension, which occurs commonly in older adults and in patients with autonomic dysfunction. Direct gastrointestinal symptoms are infrequent, but abdominal pain and nausea may occur after meals. Other clinical manifestations include postprandial presyncope and syncope. The diagnosis is established by demonstrating a 20-mm Hg or more decrease in systolic blood pressure after a standardized test meal.434 Secondary causes of autonomic neuropathy include the porphyrias (variegate, acute intermittent, hereditary coproporphyria), discussed in Chapter 76, diabetes mellitus (discussed earlier), paraneoplastic autonomic neuropathy, amyloidosis (discussed following), and Chagas’ disease (see Chapter 109).

EXTRAPYRAMIDAL DISORDERS

Huntington’s chorea is a hereditary neurodegenerative basal ganglia disease characterized by chorea, dementia, and emotional changes. Dysphagia is a common symptom that may lead to fatal complications from aspiration pneumonia.435 Gastroparesis and constipation have been reported. Parkinson’s disease is frequently associated with gastrointestinal symptoms. Oropharyngeal dysphagia and drooling can be particularly distressing. Abnormalities are found in the oral, pharyngeal, and esophageal stages of deglutition. Poor voluntary control of the tongue results in lingual hesitancy, poor bolus formation, and delayed transit into the pharynx. This defect in bolus propulsion persists in the poorly contracting pharynx. Food is often retained in the valleculae, and tracheal aspiration frequently occurs even in asymptomatic patients. Incomplete upper esophageal sphincter relaxation can be seen in 21% of patients.436 Manometric studies suggest that between 61% and 73% of patients have abnormal esophageal function, including

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases aperistalsis, simultaneous or ineffective contractions, and decreased LES tone.437 Treatment of swallowing difficulties in patients with Parkinson’s disease should include optimization of therapy directed at the tremor and associated depression and treatment of underlying gastroesophageal reflux if it exists. Levodopa has beneficial effects on the oral and pharyngeal aspects of swallowing and should be taken at mealtimes. Anticholinergics may impair esophageal motility, and excessive dosing of levodopa itself may cause nausea. Botulinum toxin injection of the cricopharyngeal muscle leads to significant short-term improvement in swallowing and weight gain.438 Finally, smaller and more frequent meals of soft food and posterior spoon placement may help in the oral phase of swallowing. Delayed gastric emptying, constipation due to delayed colonic transit, and external sphincter dysfunction also can be seen in Parkinson’s disease.437

MULTIPLE SCLEROSIS

Multiple sclerosis (MS) is associated with fecal incontinence and constipation in up to 70% of patients (see Chapters 17 and 18).439 Impaired external and internal anal sphincter function contributes to fecal incontinence in MS.440 Defecography in MS patients with intractable constipation can demonstrate rectal intussusception, rectal outlet obstruction, and failure of the puborectalis and anal sphincter muscles to relax appropriately.441,442 Biofeedback training may be beneficial in patients with limited disability and a nonprogressive disease course.443 Cranial nerve involvement in MS may lead to oropharyngeal dysphagia.

NEUROMUSCULAR DISORDERS

Degenerative diseases of peripheral motor neurons can present with gastrointestinal symptoms. In amyotrophic lateral sclerosis, bulbar dysfunction leads to oropharyngeal dysphagia with complications of malnutrition and aspiration. Patients often need gastrostomy or jejunostomy placement; a forced vital capacity of less than 50% is associated increased postprocedural mortality.444 Charcot-Marie-Tooth degenerative peripheral neuropathy also has been reported to cause pharyngeal dysphagia and abnormal gastrointestinal motility with delayed gastric emptying and disordered motility of the esophageal body.445 Motor end-plate disorders such as myasthenia gravis, an autoimmune disorder of the neuromuscular junction, are frequently associated with oropharyngeal dysphagia. Autoimmune hepatitis and primary biliary cirrhosis also may be associated with myasthenia gravis.446,447 Inherited muscular dystrophies, including myotonic dystrophies (see later), are generally believed to involve only skeletal muscles. However, several muscular (and neuromuscular) disorders have been associated with motor disturbances of the gastrointestinal system. Duchenne’s muscular dystrophy (DMD) is an X-linked recessive disease that is the most common neuromuscular disease of childhood. Patients with DMD may experience nausea and vomiting, as well as abdominal distention, constipation, pseudo-obstruction, and gastric dilation. Even when gastrointestinal symptoms are absent, dysfunction of smooth muscle in the upper gastrointestinal tract is detectable.448 Patients with DMD as well as those with oculopharyngeal muscular dystrophy may experience cervical dysphagia. Also, gastroparesis and intestinal pseudo-obstruction have been reported in DMD.449 Mitochondrial neurogastrointestinal encephalomyopathy (MNGIE) is a rare autosomal recessive mitochondrial myopathy defined by the constellation of peripheral neuropathy, opthalmoparesis, and gastrointestinal dysmotility; muscle

biopsy reveals histologic features of a mitochondrial myopathy.450 Gastrointestinal symptoms include dysmotility, diarrhea, pseudo-obstruction, achalasia, and malabsorption. A possible variant of MNGIE may be the “3-A” syndrome of familial achalasia, alacrima, and ACTH (corticotropin) sensitivity, characterized by postural hypotension, achalasia, decreased sweating and tears, and denervation hypersensitivity of the pupils.451 Stiff-man syndrome manifests as symmetrical stiffness and painful spasm of the axial musculature. Dysphagia, perhaps due to spasm of the cricopharyngeus and upper esophagus, and delayed gastric emptying have been reported.452 Myotonic dystrophy, an autosomal dominant disease of striated and smooth muscle, is considered a rare cause of gastrointestinal dilation and abnormal peristalsis. Gastrointestinal problems can be the presenting feature in 28% of patients with common symptoms including abdominal pain, dysphagia, emesis, diarrhea, and fecal incontinence.449

PULMONARY DISEASES AND PROBLEMS IN PATIENTS WHO REQUIRE CRITICAL CARE Chronic obstructive pulmonary disease is associated with peptic ulcers. The etiology of peptic ulceration remains obscure, but the risk appears to increase with the amount of cigarettes smoked per day.453 Patients with chronic lung disease secondary to α1-antitrypsin deficiency also appear to be at increased risk for peptic ulceration.454 Hepatic complications of α1-antitrypsin deficiency are discussed in Chapter 73. Gastrointestinal and hepatobiliary complications are common in patients with cystic fibrosis (see Chapter 55). Cardiac complications of severe chronic obstructive pulmonary disease are discussed later.

INTENSIVE CARE UNIT PATIENTS AND SEPTIC PATIENTS

Mechanical ventilation in the intensive care unit (ICU) in patients with serious medical illness or in postoperative patients is associated with significant gastrointestinal and hepatobiliary complications. Gastric and small intestinal motility is delayed in ICU patients, which may lead to intolerance of enteral nutrition and increased risk of aspiration.455 Acute stress-induced gastropathy is a common finding in ICU patients (see Chapters 52 and 53), and the propensity for clinically significant hemorrhage is increased in patients on mechanical ventilation and in those with significant coagulopathies.456 The successful prophylaxis of stress gastropathy in the ICU setting is discussed in Chapter 53. Ischemic colitis may be seen following cardiac arrest or hypotensive episodes. Acalculous cholecystitis is also commonly seen in severely ill ICU patients and manifests as acute abdominal pain or abdominal sepsis. The diagnosis and therapy of acalculous cholecystitis are discussed in Chapter 67. In 2003, a novel coronavirus (SARS-CoV) led to the epidemic of severe acute respiratory syndrome (SARS). SARSCoV replication has been documented in the small and large intestines.457 Diarrhea can be a presenting symptom in 15% to 20% of patients with SARS.457,458 The SARS-CoV ribonucleic acid (RNA) can be detected in the stool up to 10 weeks after the onset of symptoms.457 In patients requiring hospitalization, approximately 50% will have diarrhea and laboratory evidence of anemia and elevated alanine aminotransferase levels.458

581

582

Section IV  Topics Involving Multiple Organs Hepatic Dysfunction During Systemic Infection

Structural and functional hepatic abnormalities occur commonly during systemic infections. In many spirochetal, rickettsial, viral, and mycobacterial infections, hepatic abnormalities can be attributed to direct involvement of the liver by a multiorgan or systemic infection.459 However, in a variety of bacterial infections, intrahepatic cholestasis is observed in the absence of evidence of direct invasion of the liver by the infectious agent. Jaundice is a wellrecognized complication of severe bacterial infections in neonates,460-462 but is noted much less frequently in adults with bacteremia. In one survey of bacteremia in adult patients, only 7 of 1150 patients (0.6%) were noted to develop jaundice in the absence of evidence of primary hepatic or biliary disease.463 However, the incidence of subclinical liver dysfunction detected only by laboratory testing is much higher.459,464,465 Moreover, in some diseases such as lobular pneumonia attributed to Streptococcus pneumoniae infection, the frequency of hepatic dysfunction has been noted to be much higher with clinically apparent jaundice noted in 15% to 30% of patients, and biochemical abnormalities are nearly universal.459 Although most early reports of cholestasis and liver enzyme test abnormalities in patients with lobular pneumonia involved illness attributed to pneumococcal infection, similar clinical, biochemical and histologic patterns of hepatic abnormalities have been reported in patients with pneumonia attributed to infection by Klebsiella pneumoniae or other bacterial agents. Similarly, cholestasis has been noted in bacteremic patients infected with a wide variety of gram-positive and gram-negative organisms.459,464-466 The primary site of bacterial infection in such patients is also quite variable, with jaundice or other markers of hepatic dysfunction reported not only in patients with pneumonia but also in cases of pyelonephritis, diverticulitis, appendicitis, endocarditis, and pulmonary, soft tissue, abdominal, or pelvic abscesses.459,463-466 In some series, 40% to 90% mortality rates were noted in patients who developed jaundice in the setting of extrahepatic bacterial infections,463,466 although hepatic failure appeared to play little or no direct role in such deaths. Such observations seem to indicate that hepatic dysfunction is related to severity of underlying disease, and thus development of jaundice may be a sign of poor prognosis. However, among the subset of patients who become clinically jaundiced in the course of extrahepatic bacterial infections, the level of serum bilirubin does not appear to be different in survivors versus nonsurvivors.466 Moreover, in patients with pneumococcal pneumonia, the development of jaundice does not seem to correlate with a poor prognosis.459 Patients with jaundice in the setting of generalized sepsis usually exhibit a cholestatic pattern of biochemical and histologic abnormalities. In the majority of patients, serum alkaline phosphatase levels rise to only one to three times the upper limit of normal. In rare patients, 5- to 10-fold elevations of serum alkaline phosphatase levels with similar levels of abnormality in serum gamma glutamyl transpeptidase have been noted.464,465 Serum aminotransferase elevations tend to be modest.464-466 Peak serum bilirubin levels typically range from 5 to 10 mg/dL (86 to 171 µmol/L) with a significant component of conjugated hyperbilirubinemia invariably noted.459,464-466 However, serum bilirubin levels of 10 to 20 mg/dL (171 to 342 µmol/L) have been reported in 30% of patients, with cholestasis ascribed to extrahepatic infection,466 and levels as high as 30 to 50 mg/dL (513 to 855 µmol/L) have occasionally been noted.459 Levels of serum enzymes and bilirubin may appear discrepant with deeply jaundiced patients often having normal or near

Figure 35-10.  Liver biopsy specimen obtained from a septic patient with marked hyperbilirubinemia and normal serum alkaline phosphatase levels with the histologic picture of “cholangitis lenta.” Bile is inspissated in proliferated periportal bile ductules (arrows). The interlobular bile ducts in the portal tract are normal in appearance without bile stasis or injury. (Hematoxylin and eosin, ×25.)

normal alkaline phosphatase levels whereas some nonicteric patients may have prominent elevations of serum alkaline phosphatase and gamma glutamyl transpeptidase.464-466 Jaundice and liver enzyme abnormalities usually develop within several days of onset of bacteremia and resolve, albeit slowly, following adequate treatment of the underlying infection. Coagulopathy related to hepatic dysfunction is not a feature of this syndrome although hypoalbuminemia and hyperglobulinemia are common in icteric and nonicteric patients.464,466 Pruritus is usually absent, even in deeply jaundiced individuals, but mild hepatomegaly is frequently noted on physical examination.466 In patients with cholestasis related to extrahepatic infections, liver histology usually reveals minimal or no histologic evidence of hepatocyte necrosis. Central and midzonal bile stasis is apparent in the majority of liver biopsies obtained from jaundiced patients. In some cases a more striking picture of acute cholangiolitis termed “cholangitis lenta” (Fig. 35-10) is noted.467 In these cases, portal tracts are surrounded by dilated cholangioles containing deeply stained bile thrombi. Neutrophils are usually present within and around these dilated cholangioles but do not involve ducts within the portal tracts, thus presenting a picture distinct from that seen with extrahepatic obstruction. However, in the majority of cases, histologic findings are much less distinctive and may include a variety of relatively nonspecific findings such as mild portal mononuclear cell infiltrates, either mild or occasionally extensive fatty change and parenchymal foci of cell dropout, Kupffer cell hyperplasia, and inflammatory infiltrates.463-466 The diversity of bacteria and extrahepatic sites of infection that have been implicated in cases of cholestasis has long suggested that the factors that initiate hepatic dysfunction are humoral and common to most if not all forms of bacterial infection. Circulating endotoxin has been observed to markedly impair basolateral and canalicular bile acid and organic anion transport and to induce transcriptional downregulation of multiple hepatocyte membrane transporters.468,469 Endotoxin and other bacterial cell wall components stimulate release of proinflammatory cytokines such as TNF, IL-1, and IL-6 that have been implicated as mediators

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases of the cholestatic effects of these bacterial products.469-472 TNF administration in humans causes frequent mild and occasionally severe hepatotoxicity.473 Of note, elevated levels of IL-6 have been observed in close association with paraneoplastic cholestasis syndromes in patients with renal cell carcinoma (Stauffer’s syndrome) or other malignancies such as Hodgkin’s disease that are known to secrete cytokines such as IL-6, IL-1, or TNF.189,474 Thus proinflammatory cytokines appear to be the final common mediators of cholestasis associated with a variety of inflammatory diseases.469,475 A number of observations suggest that other factors also play a contributing role in development of jaundice in patients, with cholestasis associated with severe extrahepatic infections. In patients with lobar pneumonias, jaundice occurs more frequently in patients with a history of alcoholism, and also appears to be especially prevalent in black mens.459 These observations suggest that preexisting liver disease and hemolysis associated with glucose-6phosphatase deficiency may play a role in development of jaundice in many of these patients.459 Because many patients with jaundice associated with severe bacterial infection have life-threatening illnesses and often have concurrent evidence of prerenal azotemia,466 it is likely that decreased hepatic perfusion in patients with septic shock and decreased urinary excretion of conjugated bilirubin play a significant role in determining the extend of hyperbilirubinemia in such patients. Because cholestasis induced by systemic or severe extrahepatic infections rarely contributes to patient morbidity or mortality, the major effect of this condition relates to potential errors in diagnosis or to inappropriate treatments directed at suspected primary hepatic or biliary diseases. In patients presenting with fever and jaundice in the absence of prominent serum aminotransferase elevations or symptoms suggestive of primary hepatic or biliary tract disease, it is important to include bacterial sepsis and other extrahepatic bacterial infections such as lobar pneumonia, pyelonephritis, appendicitis, or diverticulitis in the differential diagnosis. Because the biochemical features of cholestasis associated with systemic infections cannot readily be distinguished from those seen in patients with extrahepatic obstruction or hepatic abscesses,464 ultrasonographic evaluation of the liver and biliary tree is almost always indicated. However, in such patients, careful evaluation for bacteremia and extrahepatic sites of bacterial infection should be included in the initial evaluation. When the presence of an extrahepatic or systemic infection has been established in a patient with a biochemical, clinical, and ultrasonographic picture consistent with intrahepatic cholestasis, additional radiologic and endoscopic procedures seeking to exclude intrahepatic or biliary tract disease are often best deferred until after initial response to appropriate antimicrobial therapy can be assessed.468 In patients with mild jaundice and liver enzyme abnormalities, significant improvement in these abnormalities is usually apparent within the first week of therapy.459,464 However, in those with more severe cholestasis, resolution often evolves slowly over a period of weeks.

Postoperative Cholestasis

Mild abnormalities of liver function are common after surgery476 and likely relate to multiple factors influencing hepatic function in the perioperative setting (Table 35-7). Frank jaundice is much less common in the postoperative setting476,477; however, occasionally severe cases of cholestasis develop in the absence of extrahepatic obstruction or obvious causes of hepatic parenchymal injury.478,479 In addi-

Table 35-7  Factors Contributing to Postoperative Jaundice Increased Bilirubin Production Destruction of transfused erythrocytes Hemolysis second to preexisting conditions (e.g., G6PD deficiency, hemoglobinopathies) Hemolysis second to mechanical prostheses Resorption of hematomas Hepatocellular Injury Ischemic hepatitis Drug- or anesthetic-induced hepatotoxicity Viral hepatitis Extrahepatic Biliary Obstruction Bile duct ligation Choledocholithiasis Postoperative pancreatitis Extrinsic mass compression of bile duct Intrahepatic Cholestasis Sepsis, pyogenic abscess Drug-induced cholestasis Total parenteral nutrition Preexisting Abnormalities in Bilirubin Metabolism or Secretion Chronic liver disease Gilbert’s syndrome G6PD, glucose-6 phosphate dehydrogenase.

tion, approximately half of cirrhotic patients become icteric in the postoperative period.476,477 The syndrome of postoperative cholestasis is characterized by development of jaundice 2 to 10 days after a prolonged, complicated operative procedure. Serum bilirubin levels may rise to 10 to 40 mg/dL (171 to 684 µmol/L) with associated mild to occasionally prominent alkaline phosphatase abnormalities and generally less prominent rises in serum aminotransferases.476,478,479 Although in early reports of this syndrome there was a 50% or greater mortality rate due to the severity of the underlying diseases,478,479 the course of the liver disease appeared to be benign in that no patients developed encephalopathy or other evidence of hepatic insufficiency, and jaundice resolved within two to three weeks of onset in surviving patients.476,478,479 Histologic evaluation of liver biopsy or postmortem specimens in such patients usually shows only features of intrahepatic cholestasis, although Kupffer cell erythrophagocytosis and centrilobular congestion have been noted in some specimens.478,479 Such episodes of benign postoperative cholestasis have almost invariably been preceded by long, complicated operative procedures associated with periods of hypotension and large volume blood transfusions. Thus it is likely that increased bilirubin loads due to premature destruction of transfused erythrocytes or resorption of hematomas and the effects of transient ischemia or passive hepatic congestion play a major role in pathogenesis of this syndrome.476,478,479 In addition, many of the reported cases of postoperative cholestasis have been noted to have peritonitis or other infectious complications during the postoperative period, suggesting that cholestatic responses to severe bacterial infections also play a prominent role in many cases of postoperative cholestasis.479 Because of the multitude of factors that may influence liver function in the perioperative period (see Table 35-7), determination of the etiology of postoperative jaundice is rarely straightforward. Severe ischemic injury to the liver or hepatotoxicity due to drugs or anesthetics such as halothane or structurally similar agents is generally readily distinguishable from causes of cholestasis by the presence of much more dramatic elevations of serum aminotransferases.479-481 However, in patients with cholestatic biochemi-

583

584

Section IV  Topics Involving Multiple Organs cal abnormalities, exclusion of etiologies such as sepsis, bacterial abscess, extrahepatic obstruction, or acalculous cholecystitis482 that would indicate need for therapeutic intervention requires additional bacteriologic and radiologic investigation. It is also often worthwhile to review preoperative liver function tests when available as preexisting liver disease or Gilbert’s syndrome may contribute significantly to postoperative jaundice.476 In patients with delayed onset of jaundice or liver enzyme abnormalities, etiologies such as hepatic abnormalities associated with parenteral nutrition,483,484 wound infections, toxicity of drugs administered in the postoperative period, or development of the multiple organ failure syndrome485,486 should be considered.

CARDIOVASCULAR DISEASES Congestive heart failure, cor pulmonale, or constrictive pericarditis can result in intestinal malabsorption (see Chapter 98) or protein-losing enteropathy (see Chapter 28). Cardiac ascites can also complicate right-sided congestive heart failure. Many such patients have a high serum ascites albumin gradient (SAAG) with high ascites protein content as a consequence of the high systemic venous pressures of the liver and peritoneal cavity (see Chapter 91). Ischemic “low-flow” hepatopathy may also result (see Chapter 83). Ishemic heart disease is a risk factor for ischemic colitis.487 Fibromuscular dysplasia, a nonathersclerotic, noninflammatory vascular disease, usually affects the renal and carotid arteries and may affect the mesenteric arteries, causing abdominal angina and acute intestinal ischemia (see Chapter 36).488 Colonic angiodysplasia has been associated with aortic stenosis and gastrointestinal bleeding (Heyde’s syndrome) but the association has remained controversial (see Chapter 36).489 An acquired type 2A von Willebrand syndrome, characterized by the proteolysis of the largest multimers of von Willebrand factor, has been implicated as the cause of bleeding with resolution after valve replacement.490,491 Similarly, hypertrophic obstructive cardiomyopathy has been associated with angiodysplasia in the duodenum, ileum, and colon.492

INFILTRATIVE DISEASES AMYLOIDOSIS

Amyloidosis is a disorder of protein metabolism that leads to extracellular deposition of insoluble proteinaceous material consisting of three components: (1) a nonfibrillar glycoprotein, serum amyloid P (SAP), that is present as a minor component in every form of amyloid deposit, (2) a fibrillar protein that varies in different forms of the disease (Table 35-8), and (3) glycosoaminoglycans, predominantly of the heparin sulfate and dermatan sulfate type, that are non­ covalently associated with the fibrillar proteins. Amyloid deposits appear homogeneous and amorphous under the light microscope and when stained with Congo red produce a green birefringence when viewed with a polarizing microscope. Despite these relatively uniform and specific staining characteristics, amyloid deposits are produced by a variety of diseases of diverse etiology and may be present in single or multiple organs.

Classification

Original attempts to classify amyloidosis were based on apparent etiology and perceived differences in organ involvement. The most popular such classification493 subdivided amyloidosis into (1) primary amyloidosis, a syndrome having no apparent preceding or coexisting disease, (2) secondary amyloidosis, a syndrome associated with a variety of chronic inflammatory diseases, (3) localized or tumorforming amyloidosis, (4) familial amyloidosis and (5) amyloidosis associated with multiple myeloma. However, most cases of primary or idiopathic amyloidosis are now known to be due to deposition of immunoglobulin light chains produced by occult B lymphocyte dyscrasias494 whereas a minority of such cases represent sporadic cases of hereditary systemic amyloidosis.495,496 Furthermore, it is apparent that numerous mutations in at least seven distinct serum proteins can produce the clinical syndrome of hereditary systemic amyloidosis.494,496,497 Thus accurate determination of the etiology of amyloidosis requires precise identification of the fibrillar protein present in amyloid deposits. Early descriptions emphasized apparent differences in organ involvement in various amyloid syndromes with primary amyloidosis found to principally involve the tongue, heart,

Table 35-8  Classification of Systemic Amyloidosis Syndromes AMYLOID PROTEIN

PRECURSOR OF FIBRIL PROTEIN

CLINICAL SYNDROME

AA

Serum amyloid A protein (SAA)

AL

Immunoglobulin light chains

AH Aβ2M ATTR

Heavy chain of IgG-1 Plasma β2-microglobulin Normal plasma transthyretin or genetic variants (multiple)

ACys

Genetic variant Leu68Gln of cystatin C

AGel AApoAI AApoAII

Genetic Genetic Genetic A-II Genetic Genetic

variant Asp187Asn or Asp187Tyr of gelsolin variant of apolipoprotein A-I variants stop78Gly or stop78Ser of apolipoprotein

Reactive (secondary) amyloidosis associated with chronic inflammatory diseases Amyloidosis associated with occult immunocyte dyscrasia (“primary”), multiple myeloma, macroglobulinemia or monoclonal gammopathy Same as above Hemodialysis associated amyloidosis Senile systemic amyloidosis, autosomal dominant familial amyloid polyneuropathy Hereditary cerebral hemorrhage with amyloidosis, Icelandic type Familial amyloid polyneuropathy (Finland) Non-neuropathic systemic amyloidosis (Ostertag type) Non-neuropathic systemic amyloidosis (Ostertag type)

variants Ile56Thr, Trp64Arg or Asp67His of lysozyme variants of fibrinogen

Non-neuropathic systemic amyloidosis (Ostertag type) Non-neuropathic systemic amyloidosis (Ostertag type)

ALys AFib

List compiled from references 494, 496, 497. IgG, immunoglobulin G.

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases gastrointestinal tract, muscle, nerves and skin with secondary amyloidosis reported to predominantly involve the liver, spleen, kidneys, and adrenal glands. However, it is now clear that the liver is commonly involved and may indeed be the presenting site of amyloid deposition in patients with primary amyloidosis.498 Moreover, the gastrointestinal tract is also an early site of involvement in secondary amyloidosis.499,500 As the basic biochemical and molecular properties of amyloid deposits have come to be better understood, classification systems based on the nature of the fibrillar protein component, the underlying disease pathogenesis, and the extent of organ involvement have become more useful.16,494,497 Table 35-8 details the nature of fibrillar protein found in amyloid deposits present in the liver and/or other gastrointestinal organs involved in various forms of systemic amyloidosis. In the classification system used in Table 35-8, the letter A is used to designate amyloid fibril protein and is modified by a second letter or letters to indicate the specific fibrillar protein. Thus the amino terminal fragment of immunoglobulin light chains found in the majority of amyloid deposits in both “primary” and multiple myeloma– associated cases is designated AL and the amyloid A component found in secondary amyloidosis is designated AA. Despite the apparent diversity of proteins that may serve as precursors for amyloid fibrils, these fibrils in all forms of amyloidosis share a similar ultrastructural morphology. They are 7- to 15-nm diameter rigid, twisting nonbranching fibrils that take up Congo red dye from alkaline alcoholic salt-saturated solutions and then display strong apple green

uniaxial positive birefringence when viewed in polarized light. This common staining pattern is thought to reflect either a common intermolecular packing motif shared by all amyloid fibrils501 or a common secondary structure, giving rise to antiparallel β-pleated sheets arranged with their long axis perpendicular to the long axis of the fibril.16,494 Other proteins with repeating β-sheet motifs are also insoluble and highly proteinase resistant, and thus the common structural motif of the amyloid proteins may explain similarities in both histologic staining patterns and patterns of pathologic involvement in amyloidosis caused by highly disparate precursor protein abnormalities. The chief sites of intestinal amyloid deposition are the blood vessel walls (producing ischemia and infarction), the muscle layers of the intestine (causing dysmotility), and the muscularis mucosa (impairing absorption).502 The mucosa itself is infiltrated only with massive deposition. Direct-pressure damage to cells in the myenteric plexus and visceral nerve trunks also has been demonstrated.503 Amyloidosis may cause gastrointestinal symptoms extending from the mouth to the anus (Table 35-9),500 although intestinal amyloidosis can be asymptomatic. Macroglossia and TMJ arthritis may cause drooling and difficulties with mastication.504 Gastrointestinal dysmotility can lead to dysphagia, gastroparesis, constipation, megacolon, or pseudo-obstruction.505 Diarrhea may occur due to intestinal amyloid deposition, steatorrhea from pancreatic disease, bile acid malabsorption, or rapid intestinal transit.506 Functional or mechanical gastric outlet obstruction may be due to dysmotility or to the presence of an antral amy-

Table 35-9  Symptoms, Signs, and Clinical Manifestations of Amyloidosis LOCATION

SYMPTOMS

PHYSICAL SIGNS

CLINICAL MANIFESTATIONS

Mouth

Mouth fullness Bleeding Toothache Paresthesias Dry mouth

Macroglossia Reduced tongue mobility and induration Nodular lesions of tongue and buccal mucosa Oral hemorrhagic bullae

Dysphonia Difficulty with mastication and deglutition Sicca syndrome Jaw claudication Upper airway obstruction Sleep apnea

Esophagus

Heartburn Waterbrash Dysphagia

Stomach

Nausea Vomiting Epigastric pain Anorexia Abdominal fullness Bleeding Weight loss

Succussion splash Cachexia

Erosions and ulcerations Amyloid nodules Gastroparesis Gastric outlet obstruction

Small intestine

Diarrhea Constipation Abdominal pain Bleeding Weight loss

Cachexia

Intestinal ischemia/bleeding Pseudo-obstruction Malabsorption Obstruction

Colon

Diarrhea Constipation Abdominal pain Bleeding

Cachexia

Colonic ischemia/bleeding Pseudo-obstruction Fecal incontinence Volvulus Amyloid polyposis

Liver

Jaundice Abdominal pain

Hepatomegaly Portal hypertension

Elevated serum alkaline phosphatase Focal intrahepatic mass Spontaneous hepatic rupture

Spleen

Abdominal pain

Splenomegaly

Functional hyposplenism Spontaneous splenic rupture

Pancreas

Diarrhea/steatorrhea Abdominal pain

Esophageal dysmotility Esophagitis

Exocrine pancreatic insufficiency Pancreatitis

Modified with permission from Stone MJ, Guirl MJ. Amyloidosis. In: Johnson LR, editor. Encyclopedia of Gastroenterology. San Diego: Elsevier; 2004. p 59.

585

586

Section IV  Topics Involving Multiple Organs

Figure 35-12.  Hepatic biopsy specimen showing extracellular amyloid diffusely present in the space of Disse (arrows). (Congo red, ×200.)

Figure 35-11.  Small bowel series showing symmetrical, sharply demarcated thickening of the valvulae conniventes throughout the small intestine, producing a uniform appearance characteristic of amyloidosis. (Courtesy R. H. Marshak, MD, New York, NY.)

loidoma.507 Gastrointestinal bleeding (presumably from increased capillary fragility),508 intestinal infarction (secondary to massive amyloid deposition in blood vessels),509 and protein-losing enteropathy510 have also been seen. On endoscopy, the findings are nonspecific, with descriptions of erythema, erosions, plaque-like mucosa, large duodenal folds, mucosal friability, and polypoid protrusions.511 Radiographic abnormalities include esophageal dysmotility and the appearance of diminished or rigid gastric rugal folds on upper gastrointestinal tract series. Small intestinal radiographs may reveal thickening of the valvulae conniventes (Fig. 35-11) and ulceration, and CT can show diffuse bowel thickening, intramural bowel hemorrhage, hepatomegaly with decreased attenuation, and ascites.512,513 Colonic imaging studies may show multiple filling defects, ulceration due to ischemia, or narrowing and rigidity, especially in the sigmoid colon and the rectum. Diagnosis is established by histologic demonstration of amyloid protein in involved tissues. Abdominal fat pad biopsy is a safe and simple procedure that will be positive in 85% of patients with primary AL amyloidosis.514 If AL amyloidosis is present in a patient undergoing endoscopic evaluation, biopsies from the rectum and small intestine will yield a diagnosis in 75% and 83% of patients, respectively.515 Biopsies containing amyloid will stain with Congo red, but rarely globular depositions can be seen in the lamina propria.516 There is no specific therapy for primary amyloidosis. Prokinetics such as metoclopramide may be of some benefit in patients with dysmotility. Surgical decompression has been used for intractable colorectal symptoms.517 In the liver, as in other organs, amyloidosis gives rise to amorphous, hyaline extracellular deposits in the walls of arteries and arterioles with lesser involvement of portal or hepatic veins. In the liver the space of Disse also serves as a major site of amyloid involvement. Three basic patterns of hepatic amyloid deposits have been described: (1) exten-

Figure 35-13.  Hepatic biopsy specimen showing globular deposition of amyloid in numerous Kupffer cells in the hepatic sinusoids occasionally indenting adjacent hepatocytes. (Congo red, ×100.)

sive space of Disse and sinusoidal intralobular or parenchymal involvement (Fig. 35-12), (2) vascular and periportal involvement, and (3) a mixture of parenchymal and periportal patterns of involvement.518,519 An additional unusual histologic presentation of amyloid deposition in the liver is the presence of oval globular deposits, 5 to 40 µm in diameter, in the space of Disse or portal triads of patients without typical nonglobular parenchymal or vascular involvement (Fig. 35-13).520 Patients with predominantly sinusoidal or parenchymal involvement may present with massive hepatomegaly occasionally associated with ascites and on light microscopy are found to have amyloid deposits that distort and compress the normal hepatocyte plates, often leaving little of the normal hepatic parenchyma. Other patients may present only with infiltration of the portal blood vessel walls with amyloid deposits that spare the hepatic parenchyma. Such patients typically have less prominent or no apparent hepatomegaly.

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases Although prior reports suggested that primary AL forms of amyloidosis were more likely to be associated with vascular involvement in the liver with secondary AA amyloidosis more commonly associated with parenchymal infiltration, more recent reports have found frequent vascular and relatively rare parenchymal involvement in AA forms of amyloidosis as well as near universal parenchymal infiltration in AL amyloidosis.518,519,521 Although statistically significant differences in patterns of amyloid deposition may be seen in AL versus AA amyloidosis,521 the degree of overlap is such that histologic characterization has limited value in ascribing an etiology to individual cases of hepatic amyloidosis.518,519 Instead modern diagnostic efforts tend to focus on more specific immunohistochemical techniques that attempt to identify the nature of the amyloid fibril proteins along with clinical evaluation of the patient for potential underlying diseases or amyloid fibril protein precursor abnormalities. Commercially available antibodies to serum amyloid A and β2-microglobulin have been reported to be nearly universally successful in staining AA and Aβ2M amyloid deposits, respectively,16 and antibodies against other known amyloid fibril proteins are available. However, DNA analysis is mandatory for the diagnosis of hereditary amyloidosis.494,496 AL fibril proteins most commonly derive from the amino terminal variable region of immunoglobulin light chains with λ light chains more frequently the source than κ light chains.494 Perhaps for these reasons, commercially available antisera stain AL deposits in only about half of cases16 because they are reactive predominantly with invariant epitopes on κ and λ chains, and less reactive with variable regions. However, in AL amyloidosis, careful investigation of serum and urine samples usually reveals the source of the monoclonal light-chain fibril protein.498

Clinical and Laboratory Features of Hepatic Amyloidosis

Systemic amyloidosis usually presents after age 40 and is somewhat more common in men than women. The most common symptoms are fatigue, weight loss, and edema with each of these symptoms reported by 40% to 70% of patients. Weight loss is noted by more than half of patients and is often severe.498 Edema is usually associated with congestive heart failure and is rarely associated with ascites or other stigmata of portal hypertension.493,498 In addition to such nonspecific systemic complaints, approximately 25% of patients with systemic amyloidosis report paresthesias. which is even more common in certain familial forms of amyloidosis.16,493 Other complaints may include cough or dyspnea secondary to pulmonary involvement, purpuric or papular lesions of the skin, carpal tunnel syndrome, orthostatic hypotension secondary to autonomic neuropathy and gastrointestinal bleeding, diarrhea, or malabsorption secondary to gastrointestinal involvement.16,493 Presenting complaints referable to hepatic involvement are uncommon and may include right upper quadrant discomfort related to hepatomegaly or, much more rarely, severe cholestasis, hepatic encephalopathy, or intractable ascites. Severe cholestatic presentations appear to be largely limited to patients with advanced AL amyloidosis.522 Hepatomegaly is noted on physical examination in 4% to 50% of patients with the more common forms of systemic amyloidosis (AL and AA) and is the most common hepatic manifestation of AL amyloidosis.498 Elevations of serum alkaline phosphatase values and hypoalbuminemia are common findings in patients with hepatic amyloidosis, with abnormalities of serum aminotransferases and elevations of

serum bilirubin levels being less common.493 Of interest, at initial diagnosis 80% of patients with hepatic AL amyloidosis have hypercholesterolemia, often but not always associated with the nephrotic syndrome.498 Abnormal prothrombin times may also be present.523 However, there is poor correlation between degree of abnormality in liver function tests and extent of hepatic amyloid deposition. In addition, patients without hepatomegaly or abnormal liver function tests may prove to have hepatic involvement on histologic evaluation.

Diagnosis of Systemic Amyloidosis with Hepatic Involvement

In patients with hepatomegaly and mild liver test abnormalities that develop in the setting of known monoclonal gammopathies, a chronic inflammatory disease or a constellation of systemic signs or symptoms typical of systemic amyloidosis, hepatic amyloid deposits should be suspected. Because there is no laboratory test capable of making a specific diagnosis of amyloidosis, a histologic diagnosis is required to confirm these suspicions. Liver biopsy has a high diagnostic yield in systemic amyloidosis. However, previous reports of significant hemorrhagic complications following needle biopsy of the liver or other organs with amyloid deposits523 suggest that when possible diagnostic biopsies should be limited to sites accessible to local control of bleeding. Patients with hepatic amyloidosis almost invariably have involvement of other organ sites that are equally or more amenable to diagnostic biopsies. Needle aspiration of abdominal subcutaneous fat and endoscopic biopsy of stomach, duodenum, or rectum are alternative approaches that have been recommended as having high diagnostic yield with lesser rates of life-threatening hemorrhage.16,499 Hepatic dysfunction is rarely a source of great morbidity in systemic amyloidosis, and the degree or type of hepatic involvement has little direct bearing on therapeutic decisions. Moreover, scintigraphy using radiolabeled SAP has proven to have high sensitivity and specificity in demonstrating liver involvement during systemic amyloidosis.524 Thus it is rarely necessary to resort to liver biopsy to confirm infiltration of the liver in patients with amyloid involvement demonstrated in other organs. However, in many cases of hepatic amyloidosis, patients present initially with hepatomegaly or liver test abnormalities and lack other clinical findings suggestive of systemic amyloidosis. In such cases the differential diagnosis usually includes many disorders not likely to be manifested during histologic evaluation of subcutaneous or rectal tissues. In other cases of systemic amyloidosis, attempts to obtain a histologic diagnosis from extrahepatic sites may prove unproductive. In such situations, liver biopsies may be appropriate if coagulation test results are in an acceptable range and patients have not had prior evidence of bleeding dyscrasias. One review of bleeding manifestations in 100 patients with amyloidosis noted that all patients with hemorrhagic complications of diagnostic procedures had history of bleeding disorders,523 and a recent series reports a 4% bleeding rate with no fatalities following liver biopsy in 98 patients with hepatic AL amyloidosis.498 In earlier reports, episodes of bleeding induced by diagnostic or therapeutic procedures were not found to correlate with abnormalities of coagulation tests but were associated with other bleeding problems such as intracutaneous hemorrhage (ecchymoses), gastrointestinal bleeding, hematuria, or hemoptysis523 that preceded the diagnostic procedure. These findings suggest that although coagulation abnormalities complicate amyloidosis, other factors such as amyloid infiltration of blood

587

588

Section IV  Topics Involving Multiple Organs vessel walls likely also play a major role in the propensity to hemorrhagic complications.

Treatment and Prognosis

Patients with AL hepatic amyloidosis have a median survival of less than one year,498 and for most such patients this is an incurable and inevitably progressive disease. The majority of deaths are related to cardiac or renal complications or, in the case of multiple myeloma, to progression of the underlying malignancy.16,498 Morbidity and mortality are rarely determined by extent of hepatic involvement. However, despite the grim prognosis reported in patients presenting with advanced cardiac or renal involvement, patients with initial hepatic manifestations may survive for prolonged intervals. Thus careful thought must be given to additional therapeutic approaches that may further enhance the prognosis in these patients. In general, therapy is directed at management of renal, cardiac, or other organ complications and, when possible, reduction of the amount of amyloid precursor protein to prevent or slow the rate of additional amyloid deposition. In selected patients, renal or cardiac transplantation has been performed with apparent increased survival and improved quality of life.16,525 Regression of AL amyloidosis with reduction in organomegaly and improvement in organ function have been reported in selected patients following cytoreductive therapy and stem cell transplantation to eliminate the abnormal B or plasma cell clone.526 In contrast to the generally poor response of AL amyloidosis to therapeutic maneuvers, AA amyloidosis has more frequently been noted to benefit from therapies designed to prevent initial amyloid deposition or disease progression. Colchicine therapy decreases symptoms and prevents amyloid deposition in patients with familial Mediterranean fever and appears to benefit patients who have already developed amyloidosis.527 In other cases of AA amyloidosis secondary to Crohn’s disease, juvenile rheumatoid arthritis or chronic infections such as tuberculosis or leprosy, disease progression appears to cease following initiation of specific therapy for the underlying disease.17,528 Finally, liver transplantation has proven effective in halting progression or inducing modest improvement in neurologic symptoms and regression of systemic amyloid deposits in patients with ATTR (type I familial amyloid neuropathy) amyloidosis.529 Liver transplantation has also been successfully used in other forms of genetic amyloidosis including AApoAI, ALys and AFib amyloidosis.16,494,525 The salutary effects of liver transplantation in ATTR (and potentially in AApoAI and AFib) amyloidosis appear to relate to the rapid disappearance of variant transthyretin (formerly prealbumin) from the circulation following replacement of the liver, the predominant site of synthesis of this serum protein.530,531 Of additional import, because ATTR amyloidosis rarely involves the liver, has never been associated with liver failure and usually does not manifest with extrahepatic complications until the third decade or later in life, ATTR liver explants are considered suitable for sequential or “domino” transplantation into older recipients or into recipients with primary hepatic malignancies.529,531,532 In contrast, although patients with AApoAI hereditary amyloidosis also achieve sufficient reduction in levels of amyloidogenic protein levels after orthotopic liver transplantation to achieve significant clinical benefit, their livers are a site of amyloid deposition and thus are not suitable for domino transplant.533

GRANULOMATOUS LIVER DISEASE

Granulomas are focal accumulations of inflammatory cells that invariably include mononuclear phagocytes and form

in reaction to a variety of foreign agents. Because of its rich blood supply and large number of reticuloendothelial cells, the liver is a common site for granuloma formation. Foreign body granulomas that include indigestible particulate matter such as starch, silicone or mineral oil and lipogranulomas, that form in the setting of hepatic steatosis, represent limited macrophage responses without apparent involvement of other components of the immune system.534 In contrast, epithelioid and lymphophistiocytic granulomas form under the influence of interferon-γ and other T-helper 1 cytokines that cause differentiation of monocytes and histiocytes into epitheliod cells and recruit CD8+ cytotoxic T cells as well to the periphery of the focus of inflamation.535 Such granulomas form in response to a host of infectious or systemic inflammatory diseases and also may be seen as part of immune responses to drugs, toxins, or neoplastic processes.534-536 Hepatic granulomas are most frequently found near portal tracts but may be found in the lobule and are commonly associated with elevations of serum alkaline phosphatase. Hepatomegaly, right upper quadrant pain, fever, and weight loss are other nonspecific features of granulomatous disease of the liver.534,537 In some cases, distinctive features such as presence of mineral oil or other particulate matter within the granuloma or the formation of distinctive fibrin-rings in Q fever provides insight into the etiology. However, in most cases of granulomatous liver disease, a broad spectrum of potential etiologies as summarized in Table 35-10 must be considered.534,537,538 In such cases a careful investigation of other features of the liver biopsy that includes special stains helpful in diagnosis of fungal, mycobacterial, and other distinctive infectious entities (Warthin-Starry stain for cat-scratch disease, periodic acid–Schiff (PAS) stains for Whipple’s disease) is indicated. Drug history and toxin exposure must be reviewed and additional workup may include chest radiography; cultures for bacteria (including brucella and mycobacteria) and fungi; serologic testing for Q fever, brucella, syphilis, and viral hepatitis; a tuberculin skin test; and an antimitochondrial antibody test.

SARCOIDOSIS

Sarcoidosis is a systemic disease of uncertain etiology that is characterized by the presence of granulomas in multiple organs. Hepatic involvement is not a significant source of morbidity in most patients, yet in some reports as many as 80% to 95% of North or South American patients with sarcoidosis prove to have hepatic granulomas detected during liver biopsy.539,540 As summarized in Table 35-11, hepatic involvement in sarcoidosis may be associated with a number of disparate clinical syndromes. In most of these patients, hepatic granulomas appear incidental to a disease primarily involving the lung or other organs.541 However, because of the highly variable clinical picture of sarcoidosis, liver test abnormalities, the presence of noncaseating epithelioid granulomas in histologic sections of liver biopsy specimens and/or symptoms apparently related to the presence of hepatic involvement may be among the earliest manifestations of this systemic disease.537,541,542 Alternatively, liver biopsy results may prove quite helpful in confirming the diagnosis of sarcoidosis in patients with suspected sarcoid involvement of other organ sites. Sarcoid granulomas (Fig. 35-14) are found scattered diffusely throughout the hepatic lobule with some increased frequency in the portal tracts or periportal areas.539,541,542 They are made up of epithelioid cells, sometimes with multinucleated giant cells and surrounding lymphocytes or macrophages. Only rarely do hepatic granulomas contain laminated concretions (Schaumann’s bodies), asteroid

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases Table 35-10  Causes of Hepatic Granulomas INFECTIOUS Bacterial Diseases Tuberculosis Disseminated Mycobacterium avium complex Brucellosis Tularemia Listeriosis Lepromatous leprosy Disseminated BCG Syphilis (secondary) Rickettsiosis Q fever Viruses Cytomegalovirus Epstein-Barr virus Fungal Diseases Histoplasmosis Coccidioidomycosis Cryptococcosis Parasitic Diseases Toxoplasmosis Schistosomiasis Visceral larva migrans Fascioliasis Hepatic capillariasis Ascariasis

NEOPLASTIC

MEDICATIONS

MISCELLANEOUS

Hodgkin’s lymphoma Non-Hodgkin’s lymphoma Renal cell carcinoma

Allopurinol Carbamazepine Chlorpropamide Diltiazem Gold Halothane Hydralazine Methyldopa Nitrofurantoin Penicillin Phenylbutazone Phenytoin Procainamide Quinidine Quinine Sulfonamides

Sarcoidosis Primary biliary cirrhosis Berylliosis Talc Whipple’s disease Inflammatory bowel disease Wegener’s granulomatosis Lymphomatoid granulomatosis Idiopathic

BCG, bacille Calmette-Guérin.

Table 35-11  Hepatic Involvement in Sarcoidosis Most Common Incidental hepatic granulomas in active pulmonary, cutaneous, or ocular disease Common Hepatic granulomas, fever, and weight loss with or without extrahepatic disease Rare Severe intrahepatic cholestasis Portal hypertension secondary to cirrhosis, extensive granulomas, or nodular hyperplasia

bodies, or calcium oxylate crystals. Although frank caseation is not seen, some central granular necrosis of granulomas may occur. Sarcoid granulomas are typically small and not detected on radiographic studies. However, sarcoid granulomas may cluster, forming large aggregates, and may be surrounded by significant amounts of fibrosis and/or inflammation541,542 leading occasionally to the appearance on ultrasound, CT, or magnetic resonance images of multiple 0.1- to 3-cm nodules.541,543,544 In addition to granulomas, patients with hepatic sarcoidosis commonly have varying degrees of Kupffer cell hyperplasia and mononuclear cell infiltration in the portal tracts and the hepatic lobules.539,542 In patients with clinical manifestations of portal hypertension or liver disease, additional vascular and cholestatic lesions have been observed. Granulomatous phlebitis of portal and hepatic veins has been observed in association with multiple foci of parenchymal fibrosis or with diffuse nodular regenerative hyperplasia.545,546 In patients with cholestasis, ductopenia, bile duct lesions similar to those of primary biliary cirrhosis, and periductal fibrosis reminiscent of primary sclerosing cholangitis have been observed. Finally, among patients with clinical evidence of liver disease, cirrhosis has been noted in 6% and lesser degrees of fibrosis have been observed in an additional 15%.545

Clinical Features of Hepatic Sarcoidosis

Figure 35-14.  A liver biopsy specimen demonstrating multiple granulomas in a patient with sarcoidosis (Masson trichrome, ×80.) (Courtesy of Edward Lee, MD, Washington, DC.)

Liver test abnormalities characterized by elevations of serum alkaline phosphate levels with or without less prominent serum aminotransferase abnormalities are noted in approximately one in three patients with sarcoidosis.537,547 Most sarcoidosis patients with liver test abnormalities have no signs or symptoms of liver disease and during initial follow-up, improvement in liver enzyme abnormalities is observed in half or more of patients irrespective of use of immunosuppressive therapy.547 Small subsets of patients with hepatic sarcoidosis present with jaundice or pruritus

589

590

Section IV  Topics Involving Multiple Organs as a manifestation of chronic intrahepatic542,548,549 or extrahepatic550 cholestasis or present with complications of portal hypertension.542,551,552 In patients with prominent intrahepatic cholestasis, the histologic evolution of the disease suggests progressive destruction of bile ducts by granulomas leading to progressive depletion of interlobular bile ducts, periportal fibrosis and development of “biliary” cirrhosis reminiscent of the abnormalities observed in late stages of primary biliary cirrhosis.548,549 Other patients with apparent biliary sarcoidosis and cholestasis have been observed to have cholangiographic abnormalities typical of sclerosing cholangitis that resolve following glucocorticoid therapy. Most patients with cholestatic liver disease in the setting of systemic sarcoidosis have been found to lack antimitochondrial or antineutrophil cytoplasmic antibodies.549,550 However, rare patients with multiorgan granulomas and antimitochondrial antibodies who appear to fulfil common diagnostic criteria for sarcoidosis and primary biliary cirrhosis have been reported.552,553 Patients with sarcoidosis who present with complications of portal hypertension may have either typical histologic features of cirrhosis or only extensive granulomas or nodular regenerative hyperplasia. In some of these latter patients, measurements of wedged hepatic vein pressure have been normal, suggesting that granulomatous infiltration551 or granulomatous phlebitis leading to obliteration of portal vein branches may lead to development of presinusoidal portal hypertension.546 Patients have been observed to develop portal hypertension while on steroid therapy, and in patients presenting with jaundice as a manifestation of severe intrahepatic cholestasis or with complications of portal hypertension, no consistent improvement on glucocorticoid therapy has been noted.542,548,551,552,554 Indeed, many authors have reported that the severity of glucocorticoidinduced side effects tends to exceed the therapeutic benefit achieved in this group of patients.542,549,552 Successful transplantation of the liver has been performed in such patients with advanced hepatic disease; however, intrahepatic as well as extrahepatic sarcoidosis may recur following liver transplantation.555,556 Another group of patients with hepatic sarcoidosis has been reported to present with systemic symptoms including fever and weight loss with markedly abnormal serum alkaline phosphatase levels but without jaundice or other complications of chronic liver disease. Such patients tend not to have prominent radiographic or symptomatic evidence of pulmonary sarcoidosis with confirmation of multiorgan granulomatous disease usually made by biopsy of skin, lymph nodes, or conjunctiva.537,557 Such patients with symptomatic, predominantly extrapulmonary sarcoidosis are indistinguishable clinically from a group of patients characterized as having idiopathic granulomatous hepatitis without evidence of extrahepatic involvement.540,558-560 The relative proportion of patients with hepatic granulomas who are classified as having systemic sarcoidosis versus idiopathic granulomatous hepatitis varies greatly from series to series537,538,540,557 and may depend on referral patterns, length of follow-up, or persistence in pursuit of evidence of extrahepatic involvement. In evaluating such patients, it is important to exclude infectious or other defined causes of granulomas (reviewed in Chapter 82) and to observe for the possibility of spontaneous remission.537,558 However, in patients with prolonged symptomatic courses and no defined etiology for hepatic granulomatous disease other than presumed sarcoidosis or idiopathic granulomatous hepatitis, symptomatic responses to glucocorticoid therapy have been reported to be gratifying.537,538,558-560 Although relapses are common in patients who are only treated for

brief intervals with glucocorticoids, most such patients again experience remission of systemic symptoms and exhibit some improvement in liver enzyme tests when retreated with glucocorticoids or other immunosuppressive agents.537,538,558-561 Such patients rarely have significant fibrosis on initial or follow-up biopsies and lack manifestations of portal hypertension.537,560 It is unclear whether this relatively benign course is simply characteristic of this variant of sarcoidosis or relates to the benefits of immunosuppressive therapy. Clinical gastrointestinal tract involvement in sarcoidosis is unusual. When present, symptoms result from granulomatous infiltration of the affected organ. Hilar and mediastinal lymph node enlargement may cause dysphagia.562 The most common luminal gastrointestinal site affected by sarcoidosis is the stomach. Histologic differentiation from Crohn’s disease, tuberculosis, or secondary syphilis can be difficult. There may be diffuse ulceration and antral narrowing, or the disease can resemble linitis plastica or Ménétrier’s disease.562 Seventy-five percent of patients with symptomatic gastric sarcoidosis present with pain and 25% present with bleeding. Gastric outlet obstruction may occur. Half of patients require surgery for bleeding or suspected malignancy; two of three patients improve symptomatically with glucocorticoid therapy.562 The healing of a sarcoid ulcer with antacids has been reported.563 The small intestine may be involved indirectly in sarcoidosis via mesenteric lymph node enlargement. Dilated lacteals seen on a small intestine biopsy specimen are evidence of possible lymphatic obstruction. Malabsorption and protein-losing enteropathy have been reported.564 Colonic sarcoidosis is rare but reported findings include friable mucosa, nodular hyperplasia, obstructing lesion mimicking carcinoma, and polyposis.565,566 Other rare hepatobiliary complications of sarcoidosis are Budd-Chiari syndrome and obstructive jaundice caused by hepatic hilar lymphadenopathy or strictures of the bile ducts.567 Sarcoidosis involving the pancreas appears to be rare. The clinical presentation may be similar to that of pancreatic cancer, acute pancreatitis, or pancreatic insufficiency.568-570

OTHER INFILTRATIVE DISORDERS

Eosinophilic infiltration of the gastrointestinal mucosa characterizes eosinophilic gastroenteritis (see Chapter 27), the hypereosinophilic syndrome, CSS, systemic gold toxicity, and PAN. Langerhans cell granulomatosis (histiocytosis X, eosinophilic granuloma) also may infiltrate the gastrointestinal tract. Small vessel hyalinosis is a rare familial syndrome consisting of diarrhea, rectal bleeding, malabsorption, and protein-losing enteropathy, combined with poikiloderma, hair graying, and cerebrovascular calcifications.571 Pathologically, basement membrane–like deposits can be seen in the subepithelial space of intestinal capillaries, arterioles, and small veins.

NODULAR DISORDERS OF THE LIVER Nodular hepatocellular lesions have been classified into broad categories that include either those nodules composed of regenerative hepatocytes or those containing dysplastic cells.572,573 Dysplastic or neoplastic lesions are discussed in Chapter 94. Nodular lesions in the liver created by regenerative changes include focal nodular hyperplasia (also discussed in Chapter 94), lobular or segmental hyperplasia, and regenerative nodules either associated with

Chapter 35  Gastrointestinal and Hepatic Manifestations of Systemic Diseases fibrous septa and cirrhosis or not associated with fibrous septa or cirrhosis. Lobar or segmental hyperplasia represents diffuse enlargement of a lobe or portion of a lobe and is usually associated with developmental anomalies (see Chapter 71) or with atrophy, necrosis, or fibrosis of other lobes. For instance, in the Budd-Chiari syndrome, the caudate lobe is often hyperplastic because its hepatic drainage may be preserved when the main hepatic veins are occluded.572,574 In addition, in patients with cirrhosis, the mean percentage of liver volume occupied by the caudate lobe and the lateral segment of the left lobe tends to increase.575 Developmental abnormalities of the liver include anomalous lobulations or projections from the right (Riedel’s lobe) or left lobe of the liver, accessory lobes or ectopic tissue with or without pedicles connecting to the liver, and hypoplasia or absence of a hepatic lobe (reviewed in reference 576). Uncommon liver lobulations may be perceived as abdominal or perigastric masses and raise concerns about neoplastic disease. Liver scans often clarify the benign nature of these malformations. Regenerative nodules have been classified by histologic criteria as either monoacinar nodules that contain only a single portal tract, or multiacinar nodules that contain two or more portal tracts.572,573 In cirrhotic livers, these nodules are surrounded by fibrous septa and are usually referred to as cirrhotic nodules. In contrast, nodular regenerative hyperplasia is a distinct regenerative abnormality of the liver that occurs in the absence of cirrhosis. The unifying characteristic of all cases of nodular regenerative hyperplasia appears to be the presence of obliterative lesions in small portal veins, or more rarely, hepatic veins.572-574,577,578 Obstruction of portal venous blood supply is associated with ischemia and atrophy followed by hyperplasia of acini with preserved arterial blood flow.577 In autopsy series, a strong association with increased patient age has been noted. For these reasons, it has been proposed that this pattern of regeneration represents a secondary, nonspecific adaptation to alter blood flow to the liver and therefore appears not to represent a single, specific entity. Nodular regenerative hyperplasia is composed of multiple, monoacinar nodules without fibrous septa that usually involve most of the liver. Nevertheless, this diagnosis may be difficult to establish based on findings in small-needle biopsies.572 As illustrated in Figure 35-15, reticulin stains are especially useful in identifying the unique structural features of nodular regenerative hyperplasia. Nodular regenerative hyperplasia was originally described as a rare lesion in Felty’s syndrome28,29 (see Rheumatoid Arthritis section) or in patients with various hematologic disorders associated with portal or hepatic vein thrombosis.579 Many such cases were discovered during the evaluation of patients with complications of portal hypertension. However, nodular regenerative hyperplasia has been found in 0.7% to 2.6% of autopsies, with only a minority of cases having evidence of portal hypertension.580-582 Thus it appears that this lesion occurs more frequently than suspected by clinical manifestations. Although mild to moderate elevation of serum alkaline phosphatase is often noted in patients with nodular regenerative hyperplasia, radiologic evaluations often suggest apparently normal liver. In some cases well-defined nodules that appear hypodense on CT or display abnormal echogenicity on ultrasonography583-585 may be noted. This pseudotumoral appearance on ultrasonographic or CT images can be clarified by magnetic resonance imaging, which usually reveals subtle focal lesions with dynamic behavior similar to that of normal liver parenchyma.584 Rarely, monoacinar nodules form confluent masses in the perihilar area that may be many centimeters

Figure 35-15.  Nodular regenerative hyperplasia in a patient with polycythemia vera and portal hypertension. The liver biopsy specimen is finely nodular because of nodular proliferation of hepatocytes compressing liver plates at the periphery of nodules (arrows). Fibrosis is not increased. (Masson trichrome, ×5.)

in diameter. This syndrome was formerly known as partial nodular transformation and is often associated with highgrade obstruction of medium-sized or large portal veins.572 Nodular regenerative hyperplasia has been described in association with Felty’s syndrome, primary hypogammaglobulinemia or other immunologically mediated diseases such as systemic lupus erythematosus, progressive systemic sclerosis, sarcoidosis, or polymyalgia rheumatica, and has been reported in patients with polycythemia vera, agnogenic myeloid hyperplasia, or a variety of other hematologic disorders.546,581,586 An association with azathioprine use587 and thioguanine therapy588 or other chemotherapeutic agents has been reported. Although cases of nodular regenerative hyperplasia occurring after orthotopic liver transplantation were in the past largely ascribed to azathioprine use, this lesion is also observed after liver transplantation in patients never exposed to azathioprine.589 Finally, a histologically identical lesion has been found to occur in early stages of primary biliary cirrhosis.580,590 In these patients, nodular transformation is reported to be focal rather than diffuse in the majority of cases. However, this lesion may be associated with evidence of portal hypertension in patients with primary biliary cirrhosis who have not yet developed cirrhosis.580,590 Although hepatic function remains normal in patients with nodular regenerative hyperplasia, complications of portal hypertension such as variceal hemorrhage, hypersplenism or, more rarely, ascites develop in some patients.580-582 Thus when the diagnosis is made, implicated medications should be discontinued, and treatment of associated conditions should be considered. Variceal hemorrhage should be managed initially with endoscopic band ligation or sclerotherapy, with selective shunt surgery an option in refractory cases as in other cases of portal hypertension with preserved hepatic function as detailed in Chapter 90.

KEY REFERENCES

Abell T, Lou J, Tabbaa M, et al. Gastric electrical stimulation for gastroparesis improves nutritional parameters at short, intermediate, and long-term follow-up. JPEN J Parenter Enteral Nutr 2003; 27:277-81. (Ref 316.)

591

592

Section IV  Topics Involving Multiple Organs Ahn H, Li CS, Wang W. Sickle cell hepatopathy: clinical presentation, treatment, and outcome in pediatric and adult patients. Pediatr Blood Cancer 2005; 45:184-90. (Ref 272.) Briere JB. Budd-Chiari syndrome and portal vein thrombosis associated with myeloproliferative disorders: Diagnosis and management. Semin Thromb Hemost 2006; 32:208-18. (Ref 236.) Carrier M, Le Gal G, Wells PS, et al. Systematic review: The Trousseau syndrome revisited: Should we screen extensively for cancer in patients with venous thromboembolism? Ann Intern Med 2008; 149:323-33. (Ref 262.) Devaney K, Goodman ZD, Epstein MS, et al. Hepatic sarcoidosis. Clinicopathologic features in 100 patients. Am J Surg Pathol 1993; 17:1272-80. (Ref 545.) Ebert EC. Gastrointestinal manifestations of Henoch-Schonlein purpura. Dig Dis Sci 2008; 53:2011-19. (Ref 139.) Ebert EC. Gastrointestinal manifestations of Behcet’s disease. Dig Dis Sci 2008; 54:201-7. (Ref 148.) Ebert EC, Hagspiel KD, Nagar M, Schlesinger N. Gastrointestinal involvement in polyarteritis nodosa. Clin Gastroenterol Hepatol 2008; 6:960-6. (Ref 130.) Flobert C, Cellier C, Berger A, et al. Right colonic involvement is associated with severe forms of ischemic colitis and occurs frequently in

patients with chronic renal failure requiring hemodialysis. Am J Gastroenterol 2000; 95:195-8. (Ref 405.) Geier A, Fickert P, Trauner M. Mechanisms of disease: mechanisms and clinical implications of cholestasis in sepsis. Nat Clin Pract Gastroenterol Hepatol 2006; 3:574-85. (Ref 469.) Irving KS, Sen D, Tahir H, et al. A comparison of autoimmune liver disease in juvenile and adult populations with systemic lupus erythematosus—a retrospective review of cases. Rheumatology (Oxford) 2007; 46:1171-3. (Ref 106.) Jensen RT. Gastrointestinal abnormalities and involvement in systemic mastocytosis. Hematol Oncol Clin North Am 2000; 14:579-623. (Ref 221.) Leung WK, To KF, Chan PK, et al. Enteric involvement of severe acute respiratory syndrome–associated coronavirus infection. Gastroenterology 2003; 125:1011-17. (Ref 457.) Pfeiffer RF. Gastrointestinal dysfunction in Parkinson’s disease. Lancet Neurol 2003; 2:107-16. (Ref 437.) Sellin JH, Chang EB. Therapy insight: Gastrointestinal complications of diabetes—pathophysiology and management. Nat Clin Pract Gastroenterol Hepatol 2008; 5:162-71. (Ref 320.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

36 Vascular Lesions of the Gastrointestinal Tract Lawrence J. Brandt and Charles S. Landis

CHAPTER OUTLINE Vascular Lesions  593 Angioectasia  593 Hereditary Hemorrhagic Telangiectasia (Osler-Weber-Rendu Disease)  600 Progressive Systemic Sclerosis  601 Gastric Antral Vascular Ectasia (Watermelon Stomach) and Portal Hypertensive Gastropathy  601 Portal Colopathy and Enteropathy  602 Dieulafoy’s Lesion  602

Through the widespread use of endoscopy and angiography, as well as advances in imaging techniques such as computed tomographic angiography (CTA) and magnetic resonance angiography (MRA), vascular lesions of the gastrointestinal (GI) tract are being increasingly characterized. Vascular lesions are a common cause of GI hemorrhage and may be solitary or multiple, benign or malignant, isolated or part of a syndrome or systemic disorder (Table 36-1). It is important at the outset to understand the nomenclature for the commonest lesions. Vas and its derivative vascular are Latin words meaning vessel; the Greek equivalent is angeion. Ectasia is a word of Greek derivation that refers to the process whereby a blood vessel becomes dilated or lengthened; the resulting lesion also can be referred to as an ectasia. Telangiectasia is the lesion resulting from dilation of the terminal aspect (tele) of a vessel. Angiodysplasia is used as a general term to describe the lesion or process whereby a badly formed (dys, “bad”; plasis, “molded”) vessel develops. An arteriovenous malformation is a congenital lesion, whereas an angioma is a neoplasm. This chapter discusses the more important vascular lesions that cause GI bleeding and that are representative of the spectrum of vascular lesions of the GI tract.

VASCULAR LESIONS ANGIOECTASIA

Angioectasia (AE) of the colon is a distinct clinical and pathologic entity.1-3 It is the most common vascular abnormality of the GI tract and probably the most frequent cause of recurrent or chronic lower intestinal bleeding in persons older than 60 years of age.4 AEs are probably acquired with aging, and there does not appear to be a gender predominance. In contrast to congenital or neoplastic vascular lesions of the GI tract, AEs are not associated with lesions of the skin or other viscera. However, when patients with

Hemangiomas  604 Blue Rubber Bleb Nevus Syndrome  605 Congenital Arteriovenous Malformations  606 Klippel-Trenaunay and Parkes Weber Syndromes  606 Abdominal Aortic Aneurysm  606 Mycotic Aneurysms  607 Paraprosthetic Enteric and Aortoenteric Fistulas  608 Superior Mesenteric Artery Syndrome  608 Celiac Axis Compression Syndrome  608

vascular lesions of the colon are aggressively studied with angiography or enteroscopy, concomitant lesions may be seen in the small intestine in approximately 10% of patients.3,5,6 AEs almost always are confined to the cecum or ascending colon, usually are multiple rather than single, and usually are smaller than 10 mm in diameter. They are seldom identified by the surgeon at operation or by the pathologist using standard histologic techniques, but usually they can be diagnosed by angiography; colonoscopy (Figs. 36-1 and 36-2); or, as recently shown, helical CTA.7 The roles of computed tomography (CT) and magnetic resonance imaging (MRI) for vascular lesions of all types are evolving but are certain to increase as these sophisticated modes of diagnosis become more widely available; it is also clear that conventional angiography at present is more important for therapy than for diagnosis. To determine the precise nature of a vascular lesion, histologic examination, with or without injection studies of the vasculature, is necessary. In one report in which histologic confirmation of vascular lesions was not performed, AEs reportedly occurred distal to the hepatic flexure in 46% of patients8; review of tissue sections from supposed AEs in the small bowel or left colon revealed histologic changes different from those of AEs in the right colon (personal review by S. J. Boley and L. J. Brandt). Bleeding from cecal AEs was first shown in 1961 by intraoperative angiography and has since become well recognized, especially after the introduction of selective angiography and colonoscopy for identifying the source of intestinal bleeding (see Chapter 19).8 In older literature, AEs and diverticulosis were considered the two most common causes of severe lower GI hemorrhage in older adults; however, more recent publications have cited AEs and diverticulosis to be responsible for 3% to 37% (mean: 10%) and 15% to 55% (mean: 30%) of major lower intestinal bleeding episodes, respectively (see Chapters 19 and 117).9 The problem of attributing bleeding to one or the other

593

594

Section IV  Topics Involving Multiple Organs

Figure 36-1.  Endoscopic image of an angioectasia (AE) in the ascending colon. This AE has a typical coral reef-like pattern of small vessels distorting the mucosa and submucosa. A tortuous submucosal vein, which is the earliest stage in the development of an AE, probably is present among the linear vessels intersecting the ectasia.

Table 36-1  Vascular Lesions of the Gastrointestinal Tract Primary Vascular Lesions Aneurysms of the aorta and its branches Angioectasia (angiodysplasia, vascular ectasia) Arteriovenous malformation Blue rubber bleb nevus Capillary phlebectasia Dieulafoy’s lesion Glomus tumor Hemangioma Hemangiomatosis Hemangioendothelioma Hemangiopericytoma Hemangiosarcoma Hemorrhoids Kaposi’s sarcoma Diseases and Syndromes with Vascular Lesions Blue rubber bleb nevus syndrome Ehlers-Danlos syndrome Hereditary hemorrhagic telangiectasia (Osler-Weber-Rendu disease) Klippel-Trenaunay or Parkes Weber syndrome Kohlmeier-Degos syndrome Marfan’s syndrome Pseudoxanthoma elasticum Scurvy Systemic sclerosis (scleroderma, CREST) Turner’s syndrome von Willebrand’s disease Systemic Disorders Associated with Vascular Lesions Portal hypertension Congestive gastropathy and colopathy GAVE (watermelon stomach) Spider angioma Varices Renal failure Gastrointestinal telangiectasias Watermelon stomach Vasculitis (e.g., polyarteritis nodosa) Iatrogenic lesions Radiation telangiectasia CREST, calcinosis, Raynaud’s phenomenon, esophageal dysmotility, sclerodactyly, telangiectasia; GAVE, gastric antral vascular ectasia.

Figure 36-2.  Endoscopic images of multiple angioectasias (AEs) in the ascending colon of an older adult patient who presented with recurrent bouts of lower intestinal bleeding. AEs can be single or, as shown here, multiple and of various shapes and sizes. Draining veins are seen adjacent to the AEs.

cause, when bleeding from the lesion is not demonstrated by endoscopy or by extravasation of contrast material on radiologic imaging studies, is compounded by the frequency and coexistence of these disorders without bleeding in people older than 60 years of age. The prevalence of diverticulosis is estimated to be as high as 50% in the population older than age 60; mucosal and submucosal AEs of the right colon can be found by injection studies of colons removed at surgery in more than 25% and 50%, respectively, of patients in this age range without evidence of bleeding.1,10 In large series of colonoscopic examinations, AEs have been seen in 0.2% to 2.9% of nonbleeding persons and 2.6% to 6.2% of patients evaluated specifically for occult blood in the stool, anemia, or hemorrhage.3,11-12 In a patient being studied for GI bleeding, in whom the site of active bleeding is unproven, the only basis for determining that an identified ectasia or diverticulosis is responsible for bleeding is the indirect evidence provided by the patient’s course after ablation or resection of the suspected lesion. It is unusual for AEs found incidentally to bleed, and an AE, even in a patient with a history of bleeding, cannot be assumed to be the cause.13 Bleeding from AEs typically is recurrent and low grade, although approximately 15% of patients present with massive hemorrhage. The nature and degree of bleeding

Chapter 36  Vascular Lesions of the Gastrointestinal Tract

A

C

B

Figure 36-3.  A, A specimen of resected colon that has been injected with silicone rubber, but not cleared (see text for details). Stereomicroscopy reveals the honeycomb-like pattern of normal colon crypts. B, Coral reef appearance of an angioectasia (AE) in an injected, but not cleared, colon. The normal crypts are seen surrounding the AE. C, Injected, cleared, and transilluminated colon, showing a mucosal ectasia surrounded by normal crypts with ectatic venules leading to a large, distended, tortuous underlying submucosal vein. (A and B, From Mitsudo S, Boley SJ, Brandt LJ, et al. Vascular ectasias of the right colon in the elderly: a distinct clinical entity. Hum Pathol 1979; 10:589; C, from Boley SJ, Sammartano RJ, Adams A, et al. On the nature and etiology of vascular ectasias of the colon: Degenerative lesions of aging. Gastroenterology 1977; 72:650, with permission.)

frequently vary in the same patient with different episodes: Patients may have bright red blood, maroon stools, or melena on separate occasions. In 20% to 25% of episodes, only tarry stools are passed, and in 10% to 15% of patients, bleeding is evidenced solely by iron deficiency anemia, with stools that are intermittently positive for occult blood.4 This spectrum reflects the varied rate of bleeding from the ectatic capillaries, venules, and arteriovenous communications, depending on the developmental stage of the lesions (see later). In more than 90% of instances, bleeding stops spontaneously. In 1958 E. C. Heyde described what is still a controversial association of AEs, GI bleeding, and aortic stenosis; aortic valve replacement had even been recommended for “Heyde’s syndrome” when bleeding could not be managed adequately. Numerous reports of Heyde’s syndrome appeared in the literature; subsequent analysis14 and many studies,15 however, failed to support the association. This association recently has been suggested again16 in a retrospective study in which the frequency of aortic stenosis was 31.7% in patients with “AVMs” compared with 14% in the general population; severe aortic stenosis was also more likely in the group with intestinal vascular lesions. The additional postulate has been offered that deficiencies of the largest forms of von Willebrand factor multimers (von Willebrand syndrome, type 2A) result in hemostatic abnormalities that may predispose preexisting AEs to bleed.17 Preoperative deficiency of these multimers reverses after aortic valve replacement,18 but the general recommendation to replace the aortic valve to control bleeding from AEs seems premature now that a variety of transendoscopic and angiographic means are available to ablate AEs.

Pathology

Histologic identification of AEs is difficult unless special techniques are used.1 Although usually less than one third of lesions are found by routine pathologic examination, almost all can be identified by injecting the colonic vasculature with silicone rubber, dehydrating the cells with increas-

ing concentrations of ethyl alcohol, clearing the specimen by immersing it for 24 hours in a bath of methyl salicylate, and then viewing the specimen by dissecting stereomicroscopy (Fig. 36-3).1 In a study using these methods, surgically resected colons were analyzed and found to have one or more mucosal AEs measuring 1 mm to 1 cm in diameter. AEs were usually multiple, and in this study, all were located within the cecum and ascending colon; the most distal one was 23 cm beyond the ileocecal valve.1 Microscopically, mucosal AEs consist of ectatic, distorted, thin-walled venules, capillaries, and arterioles, vessels that are lined by endothelium, and, infrequently, a small amount of smooth muscle. The earliest abnormality is the presence of dilated, tortuous, submucosal veins (Fig. 36-4A), often in areas where mucosal vessels appear normal. More extensive lesions show increasing numbers of dilated and deformed vessels traversing the muscularis mucosa and involving the mucosa (see Fig. 36-4B and C) until, in the most severe lesions, the mucosa is replaced by a maze of distorted, dilated vascular channels (see Fig. 36-4D). Enlarged arteries and thick-walled veins occasionally are seen in advanced lesions, in which the dilated arteriolarcapillary-venular unit has become a small arteriovenous fistula because of loss of prearteriolar sphincter function. Large thick-walled arteries are more typical of congenital arteriovenous malformations.

Pathogenesis

The previously described studies using injection and clearing techniques indicated that AEs are acquired lesions associated with aging and that they represent a unique clinical and pathologic entity.1 That AEs are common lesions associated with aging is supported by their frequent identification at colonoscopy in older adults and in injected colons resected from older patients with no history of bleeding.1,11 Boley postulated that the likely cause of AEs is partial, intermittent, low-grade obstruction of submucosal veins at the site where these vessels pierce the muscular layers of the colon1 (Figs. 36-5 and 36-6). He further suggested that

595

596

Section IV  Topics Involving Multiple Organs

A

B

C

D

Figure 36-4.  Histopathology of angioectasia (AE). A, Large distended veins filling the submucosa with a few dilated venules in the overlying mucosa. This appearance is the hallmark of an early AE. The black material in the lumen of the vessels is Microfil. B, A more advanced AE lesion in which dilated tortuous veins in the submucosa extend into the mucosa. C, A further stage in the development of an AE lesion in which the ectatic vessels are disrupting and replacing the mucosa. D, A late stage of AE shows total disruption of the mucosa with replacement by ectatic vessels. Only one layer of endothelium separates the lumen of the cecum from those of the dilated vessels. (Hematoxylin and eosin stain, ×50.) (From Boley SJ, Sammartano RJ, Adams A, et al. On the nature and etiology of vascular ectasias of the colon: Degenerative lesions of aging. Gastroenterology 1977; 72:650, with permission.)

Figure 36-5.  Vasa rectum and accompanying vein traversing the cecal muscularis propria. Compression of the vein is the functional anatomic explanation for intermittent, partial low-grade venous obstruction. (Elastin-von Gieson, ×50.) (From Boley SJ, Sammartano RJ, Adams A, et al. On the nature and etiology of vascular ectasias of the colon: Degenerative lesions of aging. Gastroenterology 1977; 72:650, with permission.)

repeated episodes of transiently elevated pressure during muscular contraction and distention of the cecum over many years conceivably result in dilation and tortuosity of the submucosal vein and, later, of the venules and capillaries of the mucosal units that drain into it. Finally, he postulated that the capillary rings dilate, the precapillary sphincters lose their competency, and a small arteriovenous fistula is produced. The latter is responsible for the “earlyfilling vein,” which was the original angiographic hallmark of this lesion (Fig. 36-7). Prolonged increased flow through the arteriovenous fistula can then produce alterations in the arteries supplying the area and in the extramural veins that drain it. This developmental concept of the cause of AEs was based on the finding of (1) a prominent submucosal vein, either in the absence of any mucosal lesion, or underlying only a minute mucosal AE supplied by a normal artery; (2) dilation of the veins, starting where they traverse the muscularis propria (see Fig. 36-5); and (3) previous studies showing that venous flow in the bowel may be diminished by increases in colon motility, intramural tension, and intraluminal pressure.19 Following this logic, the prevalence of AEs in the right colon can be attributed to the greater tension in the cecal wall compared with that in other parts of the colon, according to LaPlace’s principle: T ∝ pDP (where T is tension, D is diameter, and P is intraluminal pressure). An alternative concept for the development of AEs is based on the demonstration that AEs have been shown to express vascular endothelial growth factor (VEGF) and its receptors along the endothelial lining in surgical specimens from patients who have undergone colectomy for

Chapter 36  Vascular Lesions of the Gastrointestinal Tract

A

B

C

D

E

Figure 36-6.  Proposed concept of the development of cecal angioectasia. A, Normal state of vein perforating muscular layers. B, With muscular contraction or increased intraluminal pressure, the vein is partially obstructed. C, After repeated episodes over many years, the submucosal vein becomes dilated and tortuous; this is the stage that accounts for the slowly emptying vein on mesenteric angiography. D, Later, the veins and venules draining into the abnormal submucosal vein become similarly involved. E, Ultimately, the capillary ring becomes dilated, the precapillary sphincter becomes incompetent, and a small arteriovenous communication is present through the ectasia; this stage accounts for the early filling vein seen on mesenteric angiography. (From Boley SJ, Sammartano RJ, Adams A, et al. On the nature and etiology of vascular ectasias of the colon: Degenerative lesions of aging. Gastroenterology 1977; 72:650, with permission.)

Diagnosis and Management

A

B Figure 36-7.  Angiography of angioectasia (AE). A, Superior mesenteric artery arteriogram from a patient with AEs shows two densely opacified, slowly emptying, dilated tortuous cecal veins (arrows). Note the late visualization of the ileocolic vein after other veins have cleared. B, Arterial phase of the same arteriogram shows two vascular tufts (thick arrows) and two early-filling veins (each shown by a pair of thin arrows). (From Boley SJ, Sprayregen S, Sammartano RJ, et al. The pathophysiologic basis for the angiographic signs of vascular ectasias of the colon. Radiology 1977; 125:615, with permission.)

recurrent bleeding20; this indicates a proli­ferative phase of angio­genesis. VEGF and VEGF receptor 1 have been shown to be up-regulated by hypoxia21 and therefore a role has been suggested for hypoxia in the patho­genesis of AEs. Further research still is needed to clarify the pathophys­ iology of AEs.

Management of bleeding AEs consists of three phases: (1) diagnosis; (2) conversion of an emergency situation to an elective one by control of acute bleeding; and (3) definitive treatment of the AE by colonoscopic ablation or surgical removal. The diagnostic approach to colonic AEs is essentially the same as that for lower intestinal bleeding in general and includes radionuclide bleeding scans; colonoscopy; angiography; and, to exclude the small intestine as a site of bleeding, push enteroscopy and wireless capsule endoscopy (WCE). WCE is a relatively new diagnostic technique that enables visualization of the entire small intestine, and rarely the very proximal colon where colonic AEs are found. WCE is particularly useful for evaluating patients with obscure and occult GI bleeding.22 WCE has been shown to be superior to push enteroscopy in the evaluation of patients with small intestinal bleeding, yielding a diagnosis in 55% to 75% of patients with obscure bleeding that required transfusions, most common of which was angiodysplasia.23,24 Radionuclide scans are used to determine whether a patient is actively bleeding and, if so, to localize the site (see Chapter 19). Although angiography previously had been the principal means of identifying AE as the source of bleeding, colonoscopy currently is the preferred method. Helical CTA is a relatively new, sensitive, specific, and well-tolerated technique to diagnose colonic AEs, although prospective studies comparing CTA with other imaging techniques are necessary.7 The endoscopist’s ability to diagnose the specific nature of a vascular lesion is limited by the similar appearance of different types of lesions. AEs, spider angiomas, hereditary hemorrhagic telangiectasia, angiomas, the focal hypervascularity of radiation colitis, ulcerative colitis, Crohn’s disease, ischemic colitis, certain infections (e.g., syphilis, Pneumocystis), hyperplastic and adenomatous polyps, and malignancies, including lymphoma and leukemic infiltrations, can all, on occasion, resemble each other (Table 36-2). Because traumatic and endoscopic suction artifacts may resemble vascular lesions, all lesions must be evaluated on insertion of the colonoscope, rather than during withdrawal. Pinch biopsy samples of vascular lesions obtained during endoscopy usually are nonspecific; therefore, the risk of performing biopsies of these abnormalities is not justified. Because the appearance of vascular lesions is influenced by a patient’s blood pressure, blood volume, and state of hydration, such lesions may not be evident in those with

597

598

Section IV  Topics Involving Multiple Organs 1

2

3

4

Figure 36-8.  Endoscopic images showing progressive changes in the appearance of a cecal angioectasia (AE) after the administration of naloxone. Subtle at first, the AE becomes a pale and then deep red, fan-shaped obvious vascular lesion. (From Brandt LJ, Spinnell M. Ability of naloxone to enhance the colonoscopic appearance of normal color vasculature and colon vascular ectasias. Gastrointest Endosc 1999; 49:79.)

Table 36-2  Lesions That May Be Confused with Angioectasias on Endoscopy Vascular Lesions Arteriovenous malformations Angiomas Phlebectasias Spider angiomas Telangiectasias Varices Venous stars Nonvascular Lesions Trauma Polyps Adenomatous Hyperplastic Lymphoid Neoplasms Leukemic infiltration Lymphoma Colitis Infectious Inflammatory bowel disease Ischemic Radiation

severely reduced blood volumes or shock; thus accurate evaluation may not be possible until red cell and volume deficits are corrected. Meperidine also may diminish the prominence of some vascular abnormalities (e.g., AEs and the telangiectasias of hereditary hemorrhagic telangiectasia); use of meperidine, therefore, should be minimized and its effects reversed by naloxone so that vascular lesions can be detected accurately. Such a masking effect does not seem to occur with fentanyl. Naloxone has been shown to enhance the appearance of normal colonic vasculature in approxi-

mately 10% of patients and cause existing AEs to appear (2.7%) or increase in size (5.4%) (Fig. 36-8).25 For these reasons, naloxone is an important adjunctive medication for patients undergoing endoscopic evaluation for lower intestinal bleeding. Cold water lavage of the colon, as is sometimes done to cleanse the luminal surface of debris during colonoscopy, also may cause underlying AEs to disappear transiently.26 Angiography is used to determine the site and nature of lesions during active bleeding and can identify some vascular lesions even after bleeding has ceased. The three reliable angiographic signs of AEs are a densely opacified, slowly emptying, dilated, tortuous vein; a vascular tuft; and an early-filling vein (see Fig. 36-7).27 A fourth sign, extravasation of contrast material, identifies the site of bleeding when bleeding volume is at least 0.5 mL/minute but is not specific for AE. The slowly emptying vein (see Fig. 36-7A) persists late into the venous phase, after the other mesenteric veins have emptied. Vascular tufts (see Fig. 36-7B) are created by the ectatic venules that join the mucosal AE and the submucosal vein. They are seen best in the arterial phase; are usually located at the termination of a branch of the ileocolic artery; appear as small candelabra-like or oval clusters of vessels; and still are seen in the venous phase communicating with a dilated, tortuous, intramural vein. The earlyfilling vein is seen in the arterial phase within four or five seconds of injection (see Fig. 36-7B); it is not a valid sign of AE if vasodilators such as papaverine or tolazoline (Priscoline) have been used to enhance the study. When the lesion is bleeding, intraluminal extravasation of contrast material usually appears during the arterial phase of angiography and persists throughout the study. Extravasation identifies the site of active bleeding, but in the absence of other signs of AEs, it suggests another cause for the bleeding. Management of incidental (nonbleeding) AEs detected by colonoscopy is expectant. The natural history of colonic AE

Chapter 36  Vascular Lesions of the Gastrointestinal Tract is benign in healthy, asymptomatic people, and the risk of bleeding is small.13,28,29 In such cases, endoscopic therapy is not warranted.30 Bleeding can be controlled endoscopically or angiographically in most patients, thereby avoiding the morbidity and mortality of emergency operation. In decades past, intraarterial embolization and vasopressin were used to control upper and lower GI bleeding, respectively. Vasopressin, given via an angiographic catheter placed into the feeding splanchnic vessel, arrested hemorrhage successfully from AE in more than 80% of patients in whom extravasation was demonstrated. Now, superselective microcoil embolization has largely replaced intra-arterial vasopressin infusion for the treatment of lower intestinal hemorrhage.31 Such embolization is highly effective and safe but complicated by ischemic events in approximately 5% of cases.32 Vasopressin still is recommended, however, when intestinal lesions are diffuse throughout the bowel or when superselective catheterization is not possible.32 Hormonal therapy, using estrogens in combination with progestins, has been used to treat patients with a variety of vascular lesions of the GI tract, in an attempt to reduce or terminate bleeding. The mechanisms by which such agents work are not known, although procoagulant effects and endothelial injury are popular theories. Although one long-term observational study33 showed that combination hormonal therapy stopped bleeding in patients with occult GI bleeding of obscure origin (likely to have resulted from small bowel angiodysplasia), current studies do not support the use of these agents to prevent rebleeding from GI angiodysplasia.34 It is likely that hormonal therapy affects different vascular lesions differently and that vascular lesions in the small intestine may respond differently to such treatment than the same lesions in the colon; no study of hormonal therapy has been done for known colonic AEs. A novel therapy for AEs, and perhaps other vascular lesions in the gastrointestinal tract, is the use of antiangiogenic factors. Thalidomide was developed in the 1950s as a sedative, sleeping pill, and antiemetic for pregnant women, but it soon became notorious for causing phocomelia and other malformations in the newborn.34 In 1994, D’Amato and colleagues reported that thalidomide inhibited VEGF and basic fibroblast growth factor-mediated angiogenesis, which led to further characterization and subsequent clinical applications of its antiangiogenic activity.35,36 Recent data suggest the mechanism for its antiangiogenic effect is related to reduced expression of integrin genes and resulting decreased cell-cell surface interactions and response to angiogenic cytokines.37 Several case reports and case series have described the successful use of thalidomide to treat life-threatening or refractory bleeding from intestinal AEs and Crohn’s disease with refractory bleeding.38-42 After treatment with thalidomide for three months, substantial reductions in the number, size, and color intensity of AEs were observed by WCE.39 Of the available antiangiogenic biologic therapies, most information regarding clinical efficacy and toxicity is available for bevacizumab (Avastatin), a humanized monoclonal antibody against VEGF that is effective against colon and renal cancers and that also has a strong antiangiogenic activity.43 Curiously, dose-dependent nasal and GI bleeding is observed in up to 59% of patients during treatment, possibly caused by a loss of vascular integrity as a result of bevacizumab-induced endothelial-cell shedding in highly regenerative mucosal tissues with active angiogenesis. It is unclear why some antiangiogenic substances like bevacizumab cause mucosal bleeding and others like thalidomide

do not; this disparity effect may be related to the phase of angiogenesis that is antagonized, or might reflect a particular strong antiangiogenic activity. Although VEGF-based antiangiogenic therapy is a promising therapy, the issue of aggravation of bleeding from vascular lesions needs further study. A more detailed understanding of the angiogenic cascade and how antiangiogenic substances act within it will be needed to resolve this issue. Neodymium:yttrium-aluminum-garnet (Nd:YAG) laser3,6,44,45; endoscopic sclerosis11; monopolar46 and bipolar47 electrocoagulation; heater probe47; and, recently, hemoclips in combination with cautery,48 endoscopic band ligation,49 and argon plasma coagulation (APC)50 have been used to ablate vascular lesions throughout the GI tract and can be used to control active bleeding (Fig. 36-9). Control of bleeding has been obtained with a variety of endoscopic thermal means in 47% to 88% of cases,3 and no technique has been established as superior to the next.11 Severe delayed bleeding occurs in 5% of patients with colonic AEs after thermal therapy.46 Recurrent bleeding from colonic AEs appears to be reduced after these therapies, but more than one treatment session is usually necessary.47 Rebleeding can be expected to increase with time after the procedure and has been seen in 28% to 52% of patients over a follow-up period ranging from 15 to 36 months.3 In preparation for endoscopic ablation of vascular lesions, aspirin and aspirin-containing drugs, other nonsteroidal anti-inflammatory agents, anticoagulants, and antiplatelet agents should be withdrawn at least one week to 10 days before the procedure, if possible. During colonoscopy, care should be taken not to distend the cecum fully because the wall would be further thinned and the risk of perforation increased. Aspiration of some luminal gas just before thermal therapy adds a measure of safety. Again, use of meperidine should be minimized and, if it is used, its effects should be reversed by naloxone before treatment in order to detect vascular lesions more accurately. Right hemicolectomy is indicated when AE has been identified by either colonoscopy or angiography and when therapy by either or both of these two modalities is unsuccessful, cannot be performed, or is unavailable. The presence or absence of diverticulosis in the left colon does not alter the extent of colonic resection in this circumstance; only the right half of the colon is removed, but it is important that the entire right half of the colon be removed to ensure that no AEs are left behind. If the site of bleeding (and its cause) is not identified, and bleeding recurs or is continuous, recent experience suggests that a subtotal colectomy is appropriate surgical therapy. Older literature emphasized that the morbidity and mortality rates of a right hemicolectomy (which would remove all bleeding AEs and 50% to 70% of bleeding diverticula) followed by a left hemicolectomy (if bleeding recurred postoperatively) were less than the morbidity and mortality rates of a subtotal colectomy. More recent literature suggests that morbidity and mortality rates of a subtotal colectomy are not statistically different from those accompanying a “blind” hemicolectomy, that is, when the bleeding site is not identified.51,52 In one surgical series and review, mortality for subtotal colectomy was 0% to 40% with a rebleeding rate of 0% to 8%, and mortality and rebleeding rates for a directed limited colectomy were 2% to 22% and 0% to 15% respectively51; in contrast, the mortality and rebleeding rates for a blind limited colectomy were 20% to 57% and 35% to 75%, respectively. In another surgical series, frequency of bowel movements after limited colectomy was 2.4 per day, a number not substantially different from the 3.5 bowel movements per day documented after subtotal colectomy.52

599

600

Section IV  Topics Involving Multiple Organs

A

B

C

Figure 36-9.  Endoscopic images showing angioectasia and diverticulosis of the colon. A (upper), A single angioectasia (AE) nestled among diverticula in the ascending colon of an older adult man with lower intestinal bleeding. AEs and diverticula are probably the two most common causes of major recurrent lower intestinal bleeding in older adults, therefore finding them together in the same patient is not unusual. A (lower), The lesion after treatment with argon plasma coagulation. B (upper), Multiple AEs in the ascending colon. B (lower), AEs after treatment with a heater probe. Not every vessel must be ablated for the bleeding to resolve. C (upper), A solitary AE. C (lower), The heater probe is being used to ablate the lesion.

HEREDITARY HEMORRHAGIC TELANGIECTASIA (OSLER-WEBER-RENDU DISEASE)

This autosomal dominant familial disorder is characterized by telangiectasia of the skin and mucous membranes, as well as recurrent GI bleeding.53-55 The pathogenesis may relate to mutations of the endoglin (ENG) and activin receptor–like kinase 1 (ALK-1) genes, which have an im­ portant role in determining the properties of endothelial cells during angiogenesis (see later).56 Lesions typically are noticed in the first few years of life, and recurrent epistaxis in childhood is characteristic of the disease. By age 10, about half of patients have had some GI bleeding. Severe hemorrhage is unusual before the fourth decade and has a peak incidence in the sixth decade. In most patients, bleeding presents as melena; bright red blood per rectum and hematemesis are less frequent. Hematochezia in a patient with hereditary hemorrhagic telangiectasia (HHT) suggests bleeding from a source other than telangiectasia. Bleeding is chronic and may be severe; patients may receive more than 60 transfusions in a lifetime. A family history of the disease has been reported in 80% of patients with HHT but is less common in those who bleed later in life. Telangiectasias usually are present on the lips, oral and nasopharyngeal membranes, tongue, and periungual areas; lack of involvement of these sites casts suspicion on the diagnosis (Fig. 36-10). The clinical diagnosis of HHT currently requires the presence of at least three of four relevant clinical criteria. These so-called Curaçao criteria include epistaxis (spontaneous and recurrent nosebleeds); telangiectases (multiple at characteristic sites, e.g., lips, oral cavity, fingers or nose); visceral lesions (e.g., pulmonary, hepatic, cerebral, spinal, or

GI vascular malformations); and family history (a firstdegree relative with HHT).57 The clinical diagnosis of HHT can be confirmed by molecular genetic analysis. In most cases, HHT is caused by mutations in one of the two known HHT genes. Mutations of the ENG lead to type 1 HHT.58 The ENG gene is located on chromosome 9q34.1 and encodes for endoglin, a type III transforming growth factor-β (TGF-β) receptor. Type 2 HHT is attributed to mutations of the activin A receptor type II-like 1 gene (also termed activin receptor-like kinase-1 (ALK1),59 which codes for the ACVRL1 protein, a type I TGF-β receptor. Both receptors are members of the TGF-β receptor family, are expressed predominantly on vascular endothelium, and play essential roles in maintaining vascular integrity. Evidence for the existence of two other as yet unidentified HHT genes has been reported.60,61 Despite genotypic heterogeneity in HHT, the clinical expression of the different HHT genotypes appears to be the same. HHT and nonhereditary intestinal AE (see earlier) are characterized by increased production of VEGF. High serum levels of VEGF, which also correlate with severity of bleeding, are found in patients with HHT.62,63 Vascular involvement of the liver is common in HHT and frequently is asymptomatic; hepatic manifestations during the course of the disease are seen in 8% to 31% of patients. Typical clinical presentations of liver involvement are highoutput heart failure resulting from arteriovenous shunting, portal hypertension, and biliary tract disease.64,65 Serious com­plications, including liver failure necessitating liver transplantation, have been reported. Telangiectasias occur in the colon but are more common in the stomach and small bowel, where they also are more apt to cause major bleeding.

Chapter 36  Vascular Lesions of the Gastrointestinal Tract

A

D

B

C

Figure 36-10.  Telangiectasias of hereditary hemorrhagic telangiectasia. A, Multiple telangiectasias on the nose and lips. Telangiectasias of varying size and shape in the proximal gastric body (B), antrum (C), and duodenal bulb (D). (From Wilcox CM. Atlas of Clinical Gastrointestinal Endoscopy. Philadelphia, Pa: WB Saunders; 1995. p 123.)

Telangiectasias are seen easily on endoscopy, although in the presence of severe anemia, blood loss, or hypotension, they transiently may become less obvious or even invisible; after correction of blood volume and blood pressure, they become prominent again. Evaluation by conventional angiography or newer techniques such as helical CTA66 and MRA may be unrevealing or demonstrate arteriovenous communications, conglomerate masses of abnormal vessels, phlebectasia, and aneurysms.67 Angiography may be misleading when it demonstrates multiple vascular abnormalities because some of these lesions may be in the mesentery rather than in the bowel and are not potential sites of GI blood loss. Grossly, the telangiectasias are the size of millet seeds and typically appear as cherry red, smooth hillocks. Pathologically the major changes involve the capillaries and venules, but arterioles also may be affected. Lesions consist of irregular, ectatic, tortuous blood spaces lined by a delicate single layer of endothelial cells and supported by a fine layer of fibrous connective tissue. No elastic lamina or muscular tissue is present in these vessels, so they cannot contract; this property may explain why the lesions tend to bleed. Arterioles show intimal proliferation and commonly have thrombi in them, suggesting vascular stasis. In contrast to the thinned venules of AEs, venules are abnormally thick in HHT; have prominent, well-developed longitudinal muscles; and apparently play a major role in regulating blood flow in telangiectasias.66 Many forms of treatment have been recommended for telangiectasias, including estrogens,68 aminocaproic acid,69 endoscopic thermal ablation,6,45 and resection of involved bowel. Endoscopic ablation, including the use of the APC and thermal contact devices, is most promising when lesions are within reach of the endoscope and not too

diffuse. Endoscopic therapy may be performed during active bleeding or between bleeding episodes and has reduced the need for emergency bowel resection. Long-term follow-up studies are necessary to evaluate the ultimate efficacy of the various forms of therapy. Bevacizumab, the humanized monoclonal antibody against VEGF discussed earlier, was used to treat a 47-yearold woman with HHT and severe liver involvement. Significant clinical improvement was observed three months after initiating treatment with reversal of cholestasis, resolution of cardiac failure and ascites, and improvement in nutritional status. A marked reduction in liver vascularity and liver volume also were seen over a six-month interval.70

PROGRESSIVE SYSTEMIC SCLEROSIS

(see also Chapter 35) Vascular lesions are a prominent feature of progressive systemic sclerosis, especially in the calcinosis, Raynaud’s phenomenon, esophageal dysmotility, scleroderma, and telangiectasia (CREST) variant.71 Sites most frequently involved by these telangiectasias are the hands, lips, tongue, and face, but gastric, intestinal, and colorectal lesions have been reported. These tiny lesions may be the source of occult or clinically significant bleeding and are best treated, if possible, by endoscopic thermal ablation.72

GASTRIC ANTRAL VASCULAR ECTASIA (WATERMELON STOMACH) AND PORTAL HYPERTENSIVE GASTROPATHY

Gastric antral vascular ectasia (GAVE), or watermelon stomach, describes a vascular lesion of the gastric antrum that consists of tortuous, dilated vessels radiating outward

601

602

Section IV  Topics Involving Multiple Organs

Figure 36-11.  Endoscopic appearance of watermelon stomach, also referred to as gastric antral vascular ectasia (GAVE).

from the pylorus-like spokes of a wheel and resembling the dark stripes on the surface of a watermelon.73 This lesion may cause acute hemorrhage and chronic occult bleeding. Its cause is unknown, although it has been proposed that gastric peristalsis causes prolapse of the loose antral mucosa with consequent elongation and ectasia of the mucosal vessels (Fig. 36-11).73 GAVE also has been thought to result from delayed gastric emptying, as well as from humoral factors such as hypergastrinemia, prostaglandin E2, 5-hydroxytryptamine (serotonin) produced by neuroendocrine cells, and vasoactive intestinal polypeptide (VIP). GAVE is seen particularly in middle-aged or older women and in association with achlorhydria, atrophic gastritis, cirrhosis, and the CREST syndrome, as well as after bone marrow transplantation.71,74 The association with cirrhosis and portal hypertension in approximately 40% of reported cases of GAVE suggests that this lesion may be caused by portal hypertension or hepatic veno-occlusive disease.75 Microscopic features of GAVE include dilated capillaries with focal thrombosis, dilated and tortuous submucosal venous channels, and fibromuscular hyperplasia of the muscularis mucosa. Some researchers believe that GAVE and portal hypertensive gastropathy (PHG) are different manifestations of the same pathogenetic process, whereas others view them as separate entities with distinct clinical and histologic features. Recent evidence suggests that these are distinct entities.76 The pathophysiology of GAVE does not appear to involve portal hypertension because it does not respond to therapies directed at reducing portal pressure. Given case reports of GAVE resolution after liver transplantation, however, it is possible that liver insufficiency may play a role in its pathophysiology.77,78 PHG is characterized endoscopically by three patterns: (1) fine red speckling of the mucosa; (2) superficial reddening, especially on the tips of the gastric rugae; and, most commonly, (3) the presence of a mosaic pattern with red spots (snakeskin appearance) in the gastric fundus or body (Fig. 36-12). Histologically, the stomach in PHG contains dilated, tortuous, irregular veins in the mucosa and submucosa, sometimes with intimal thickening, usually in the absence of significant inflammation.76 Limited therapeutic options are available for treating GAVE. Estrogen-progesterone has been tried79 and appears

to have some efficacy. Successful use of tranexamic acid, an antifibrinolytic agent,80,81 and thalidomide82 also have been reported. Transjugular intrahepatic portosystemic shunting (TIPS) does not appear to be effective for GAVE.83 This is not surprising given that the pathophysiology of GAVE may not be related to portal hypertension.78,84 Antrectomy has been used for patients in whom pharmacologic and endoscopic therapies have failed.85 Portal hypertension in a patient with GAVE and GI bleeding makes the bleeding more difficult to manage; bleeding is usually greater and more resistant to treatment in the presence of portal hypertension.86 Several case reports have detailed reversal of GAVE after liver transplantation,77 however, the data are insufficient to recommend this therapy unless the patient is otherwise a liver transplant candidate. Iron therapy and blood transfusions were the mainstays of medical treatment for GAVE, and antrectomy often was required in severe cases before development of transendoscopic thermal ablation techniques. All of the available endoscopic therapies have been used successfully to ablate GAVE, and now antrectomy is rarely required.87 TIPS offers another modality when GAVE is associated with portal hypertension or when bleeding resulting from PHG is not controlled by transendoscopic coagulation therapy (see also Chapters 19 and 90). The initial management of PHG is with iron supplementation and nonselective beta blockers. Propranolol is the nonselective beta blocker that was investigated in the classic randomized controlled trial evaluating the role of beta blockade in preventing recurrent bleeding in severe PHG.88 In this study, patients who received propranolol had a significantly lower rebleeding rate at 12 months (35% versus 62%) and at 30 months (48% versus 93%) compared with patients taking a placebo. If a patient is refractory to beta blocker therapy, however, shunt therapy (TIPS or shunt surgery) is indicated. TIPS placement or shunt surgery has been effective in almost all cases.83,89-92 The choice between TIPS or surgery should be made based on local expertise. Somatostain analogs such as octreotide, which are established as effective treatment for acute variceal bleeding, also have been shown to be effective for bleeding from PHG. In two studies evaluating the effect of these vasoactive drugs in acute hemorrhage from PHG,93,94 bleeding was successfully treated in all patients who received somatostatin or octreotide. Vasopressin and its analog, terlipressin, also have been tried, but results have been mixed.93,95

PORTAL COLOPATHY AND ENTEROPATHY

Portal colopathy is the term used to describe vascular manifestations of portal hypertension in the colon. Mani­ festations include hemorrhoids, varices, and spider-like telangiectasias (Fig. 36-13A and B). Mucosal lesions of portal colopathy typically resemble those seen in PHG and may have a diffuse, colitis-like appearance, including granularity, erythema, telangiectases, and friability. Varices and spider-like telangiectases also may be seen in the small intestine, warranting the term portal enteropathy. Histologic changes of portal colopathy and enteropathy are similar to those of portal gastropathy.96 The lesions of portal colopathy and enteropathy are amenable to the same thermal therapies used for GAVE and PHG.97

DIEULAFOY’S LESION

This vascular lesion is an unusual cause of massive GI hemorrhage, usually from the stomach, but sometimes from the small or large bowel (Fig. 36-14).98 It is twice as common in men as in women and presents at a mean age of 52 years.

Chapter 36  Vascular Lesions of the Gastrointestinal Tract

A

B

C Figure 36-12.  Portal hypertensive gastropathy. A, Mild disease is manifested by prominence of the areae gastricae, with areas of erythema and subepithelial hemorrhage. This appearance is not pathognomonic and may be noted with other disorders that induce mucosal edema, such as Helicobacter pylori gastritis. B, Severe gastropathy with diffuse subepithelial hemorrhage in a snakeskin pattern. C, Low power photomicrograph showing prominent edema of the lamina propria with multiple congested blood vessels. No histologic evidence of gastritis is seen. (From Wilcox CM. Atlas of Clinical Gastrointestinal Endoscopy. Philadelphia, Pa: WB Saunders; 1995. p 109.)

A

B

Figure 36-13.  Endoscopic images showing two examples of portal colopathy. A, A solitary lesion that resembles a spider-like telangiectasia or angioectasia is seen in the rectosigmoid. B, Patchy foci of erythema in the descending colon of a patient with cirrhosis and portal hypertension.

603

604

Section IV  Topics Involving Multiple Organs

A

B

Figure 36-14.  Endoscopic images of a Dieulafoy’s lesion. A, Arterial bleeding (spurting) just distal to the gastroesophageal junction. B, The bleeding point was a small defect without endoscopic evidence of ulceration. (From Wilcox CM. Atlas of Clinical Gastrointestinal Endoscopy. Philadelphia, Pa: WB Saunders; 1995. p 122.)

The vascular abnormality is the presence of arteries of persistently large caliber in the submucosa and, in some instances, the mucosa, typically with a small, overlying mucosal defect. Dieulafoy called the lesion “exulceratio simplex” because he thought it was the initial stage of a gastric ulcer. This lesion also has been called an atherosclerotic aneurysm, an inaccurate term because the caliber of the artery’s walls is uniform throughout and shows no unusual degree of arteriosclerosis. It is believed that focal pressure from these large “caliber-persistent” vessels thins the overlying mucosa, leading to erosion of the exposed vascular wall and resulting hemorrhage. Massive hema­ temesis or melena typically is not preceded by any GI tract symptoms and usually is followed by intermittent and severe bleeding over several days. The most common site of bleeding is 6 cm distal to the cardioesophageal junction, where the arteries are largest, but many lesions have been reported in extragastric locations, including the esophagus, small bowel, rectum,98 and even outside the GI tract in the bronchus, presenting as hemoptysis. It may be difficult to find a Dieulafoy lesion in a patient with upper GI bleeding because the overlying mucosal defect may be small and hidden between the gastric rugae, and the caliberpersistent vessel may constrict and retract after the bleeding episode. The mortality rate for older adult patients with this lesion has been high, nearly 80% before diagnostic endoscopy was available and more than 20% in cases reported between 1970 and 1986. The high mortality rate resulted from the inability to localize the bleeding site and the frequent need for emergency gastric surgery. Current angiographic and endoscopic techniques used to localize and treat bleeding lesions have led to an improvement in 30-day mortality rates, now reported to be 13%. Therapeutic approaches to bleeding Dieulafoy’s lesions include injection therapy, heater probe, laser, APC, band ligation, and hemoclip placement.99

HEMANGIOMAS

Considered by some to be true neoplasms, hemangiomas generally are thought to be hamartomas because most are present at birth. Hemangiomas are the second most common vascular lesion of the colon and may occur as solitary or

multiple lesions limited to the colon or part of diffuse GI or multisystem angiomatoses. Hemangiomas may be classified as cavernous, capillary, or mixed types. Most are small, ranging from a few millimeters to 2 cm, but larger lesions occur, especially in the rectum. Bleeding from colonic hemangiomas usually is slow, producing occult blood loss with anemia or melena. Hematochezia is less common, except with large cavernous hemangiomas of the rectum, which may cause massive hemorrhage. The diagnosis is best established by endoscopy, including enteroscopy because roentgen studies, including angiography, frequently are normal. The diagnosis of cavernous hemangioma of the rectum often can be suggested on plain films of the abdomen by the presence of phleboliths and displacement or distortion of the rectal air column (Fig. 36-15). On barium enema, the affected rectal lumen typically shows narrowing and rigidity, scalloping of the rectal wall, and widening of the presacral space (see Fig. 36-15). Endoscopically, one sees elevated plum-red nodules or vascular congestion; ulcers and proctitis also may be present. Angiography can demonstrate these lesions but seldom is necessary to establish the diagnosis. Hemangiomas are well circumscribed but not encapsulated. Grossly, cavernous hemangiomas appear as polypoid or mound-like reddish purple lesions on the mucosa. Histologically, numerous dilated, irregular blood-filled spaces are seen within the mucosa and submucosa and sometimes extend through the muscular wall to the serosal surface. The vascular channels are lined by flat endothelial cells with flat or plump nuclei, and their walls are composed of fibrous tissue. Capillary hemangiomas are plaque-like or moundlike reddish purple lesions composed of a proliferation of fine, closely packed, newly formed capillaries separated by little stroma. The endothelial cells are large, usually hypertrophic, and in some areas may form solid cords or nodules with ill-defined capillary spaces. Small hemangiomas that are solitary or few in number and can be approached endoscopically are locally ablated. Most large or multiple lesions require resection of either the hemangioma alone or the involved segment of colon. Local measures to control massive bleeding from cavernous hemangioma of the rectum usually are effective only temporarily. Embolization and surgical ligation of major feeding

Chapter 36  Vascular Lesions of the Gastrointestinal Tract

B

A Figure 36-15.  Two examples of cavernous hemangioma of the rectum. A, Plain film of the pelvis reveals a soft tissue mass with foci of calcification in abnormal vascular channels. This appearance of pelvic phleboliths in a child is pathognomonic for a cavernous hemangioma. B, A barium enema film shows the characteristic phlebolith pattern outside the colon, with scalloping of the bowel lumen caused by pressure from the vascular lesion.

vessels also have been used, but ultimately excision of the rectum often is required.100 In a particular entity known as diffuse intestinal hemangiomatosis, numerous lesions, usually of the cavernous type, involve the stomach, small bowel, and colon; hemangiomas of the skin or soft tissues of the head and neck frequently are present. The occurrence of bleeding or anemia in childhood typically leads to the diagnosis, which usually is made after various studies including endoscopy, barium series, scintigraphy with 99mTc-labeled red blood cells,101 and contrast-enhanced CT. Angiographic findings can be normal despite the number of lesions. Surgical intervention may be required for continuous, slow bleeding or for intussusception. At operation, all identifiable lesions should be excised either through enterotomies or by limited bowel resections. Intraoperative endoscopy may be helpful in finding small lesions. Repeated operations may be necessary to control blood loss.102 Semaxanib, a small-molecule inhibitor of VEGF receptor 2, recently has been used for treatment of hemangioblastoma in patients with von Hippel-Lindau disease (vHLD). In vHLD, a loss of von Hippel-Lindau protein results in an accumulation of hypoxia-inducible factor and subsequently, excessive production of VEGF.103 Recent studies have reported regression or stabilization, with improvement of macular edema, in patients with hemangioblastoma treated with semaxanib, which suggests inhibition of VEGF103,104 and a role for VEGF inhibitors in the management of hemangiomas.

BLUE RUBBER BLEB NEVUS SYNDROME

In 1860, an association among cutaneous vascular nevi, intestinal lesions, and GI bleeding was described, and almost a century later this constellation of findings was named blue rubber bleb syndrome by Bean to distinguish it from other cutaneous vascular lesions (Fig. 36-16). Although the GI tract is most frequently involved, other sites may be affected, including the eyes, nasopharynx, parotid glands, lungs, liver, spleen, heart, brain, skeletal muscles, urinary

Figure 36-16.  Fingertip lesion in a patient with the blue rubber bleb nevus syndrome.

bladder, and penis. Orthopedic abnormalities may be present, and calcification, thrombosis, and consumptive coagulopathy (with thrombocytopenia) may occur within the lesions.105 A familial history is infrequent, although a few cases of autosomal dominant transmission have been reported106 and one analysis has identified a responsible locus on chromosome 9. The lesions are distinctive: they are blue and raised, vary from 0.1 to 5 cm in diameter, and have a wrinkled surface. Characteristically, the contained blood can be emptied by direct pressure, leaving a wrinkled sac. Lesions may be single or numerous and are usually found on the trunk, extremities, and face. They may involve any portion of the GI tract but are most common in the small bowel. In the colon they are more common distally. They are detected infrequently by barium or angiographic studies and are seen best by endoscopy. Originally the lesions were thought to be hemangiomas, but they are now considered to be venous malformations. Resection of the involved segment of bowel

605

606

Section IV  Topics Involving Multiple Organs

Figure 36-17.  Mesenteric angiogram in a patient who presented with recurrent lower intestinal bleeding. The angiogram shows a complex, racemose configuration of vessels in a large congenital arteriovenous malformation involving the superior and inferior mesenteric arterial circulations.

is recommended for recurrent hemorrhage. Endoscopic laser coagulation may be dangerous because these lesions may involve the full thickness of the bowel wall; successful sclerotherapy and band ligation of GI tract lesions have been reported.

CONGENITAL ARTERIOVENOUS MALFORMATIONS

Arteriovenous malformations (AVMs) are embryonic growth defects and are considered to be developmental anomalies. Although AVMs are found mainly in the extremities, they may occur anywhere in the vascular tree. In the colon they may be small and resemble AEs or they may involve a long segment of bowel. The most extensive lesions typically are in the rectum and sigmoid. Histologically, AVMs are persistent congenital communications between arteries and veins located primarily in the submucosa. Characteristically, there is “arterialization” of the veins (i.e., tortuosity, dilatation, and thick walls with smooth muscle hypertrophy and intimal thickening or sclerosis). In long-standing AVMs, the arteries are dilated with atrophic and sclerotic degeneration. Angiography is the primary means of diagnosis (Fig. 36-17). Early-filling veins in small lesions and extensive dilatation of arteries or veins in large lesions are typical. Patients with significant bleeding from large AVMs should undergo resection of the involved segment; transendoscopic therapy may be beneficial for smaller lesions.

KLIPPEL-TRENAUNAY AND PARKES WEBER SYNDROMES

In its initial description, the Klippel-Trenaunay syndrome consisted of (1) a vascular nevus involving the lower limb; (2) varicose veins limited to the affected side and appearing at birth or in childhood; and (3) hypertrophy of all tissues of the involved limb, especially the bones.107 Subsequently, a variety of vascular lesions associated with the hypertrophic limb were described and some authors now divide the

syndrome into two: Klippel-Trenaunay and Parkes Weber; the former is a pure low-flow condition, whereas the latter is characterized by arteriovenous fistulas. Several genetic defects in the regulation of the angiogenic factor VG5Q have been shown in patients with this syndrome.108 The cause of bony elongation is controversial, but one theory invokes in utero venous hypertension and stasis.107 Edema of the involved leg is common, and if the thigh is involved, a variety of lymphatic abnormalities are usually present (e.g., chylous mesenteric cysts, chyloperitoneum, protein-losing enteropathy; see Chapters 28 and 37). Symptomatic GI involvement is rare. In the largest series, the most common GI symptom was hematochezia, reported by only 6 of 588 patients.107 GI bleeding may be recurrent and mild or severe and usually is caused by a rectal hemangioma, localized rectovaginal varices resulting from obstruction of the internal iliac system, or portal hypertension with varices. Bleeding may be intensified by consumption coagulopathy, which may occur within the smaller sinusoids of the vascular lesion. Physical examination is diagnostic, and various imaging techniques are used to define the anatomy and plan surgical repair.109 Most recently, MRA has been used for diagnosis and to detect arteriovenous shunting.110 Endoscopic thermal ablation therapy is useful in controlling hemorrhage and preventing or minimizing recurrent GI bleeding, especially when the lesions are relatively well localized (L. Brandt, personal experience).

ABDOMINAL AORTIC ANEURYSM Approximately 95% of abdominal aortic aneurysms (AAAs) are atherosclerotic in origin, but other factors such as genetic predisposition are also important; less common causes include trauma, vasculitis, infection, and congenital abnormalities. Eighty-five percent of affected people are men. Familial clustering of AAAs has been noted in 15% to 20% of cases, and in some families an abnormality has been identified on chromosome 16111; defects in procollagen III in patients with Ehlers-Danlos syndrome type IV and altered gene expression causing abnormalities of the elastin and collagen content of aneurysms have been shown in other families.111 Most AAAs are asymptomatic. The most common symptom is epigastric pain, often radiating through to the back; severe pain may presage rupture. On physical examination, a pulsatile epigastric mass may be palpable. Distinguishing an aneurysm from an overlying abdominal mass with transmitted pulsations may be difficult on physical examination and is best done by imaging studies. A bruit may be present, but unless recent in onset, it is usually of no diagnostic help. Abdominal plain films may show a soft tissue mass with peripheral calcification in the region of the abdominal aorta. With large aneurysms, erosion of the lumbar vertebrae or displacement of surrounding viscera, including bowel, kidneys, and ureters, may be seen. Because plain film studies are not sufficiently sensitive to establish the presence or size of an aneurysm, ultrasonography, CT, and MRI have become the standard means of evaluation. These procedures are simple, safe, and accurate in the diagnosis and sizing of aneurysms. Ultrasonography is less sensitive than CT in determining the extent of the aneurysmal process, but it is useful for following changes in the size of the aneurysm.112 CT and MRI are used preoperatively to demonstrate aortic and vascular anatomy. Preoperative angiography is not used as frequently as in the past and is most appropriate

Chapter 36  Vascular Lesions of the Gastrointestinal Tract in patients with evidence of peripheral vascular disease, severe hypertension, symptoms of chronic mesenteric ischemia (see Chapter 114), if thoracic or iliac artery involvement is suspected and, in cases of horseshoe or pelvic kidneys, to demonstrate renal artery anatomy. Angiography is not used to estimate the size of the aneurysm because intraluminal laminated thrombus limits delineation of the entire lumen. The major complication of AAAs is rupture, which is heralded by the sudden onset or worsening of pain in the abdomen, flank, or back; pain may be present for several weeks and is attributable to “leakage” that precedes overt rupture. Pain may be exacerbated by lying recumbent and relieved by sitting or leaning forward. In one series, only 14% of patients referred for treatment of rupture had been known to have an aneurysm previously.113 Severe abdominal pain also may be seen with aortic dissection as the splanchnic vessels become compromised and acute intestinal ischemia develops. The consensus among vascular surgeons is that the most important predictor of rupture is the size of the aneurysm. The risk of rupture for small aneurysms is negligible and in one series was reported to be 0% at five years114; the five-year risk for aneurysms that are 5.5 to 5.9 cm in size is about 20% to 25%; for a 6-cm aneurysm, the risk is 35% to 40%; and for those larger than 7 cm, the risk is 75%.111 Other risk factors for rupture include hypertension and the presence of chronic obstructive pulmonary disease. AAAs most commonly rupture into the retroperitoneal tissues that surround the aorta. Less commonly, the aneurysm may communicate with the peritoneal cavity, in which case shock develops rapidly. Patients whose aneurysm ruptures into the small intestine, usually the third or fourth portions of the duodenum, typically present with massive GI bleeding; bleeding may be intermittent because a clot alternately forms and is dislodged from the eroded bowel or fistulous opening. Indeed, many of these patients will have a “herald bleed” followed by massive hemorrhage several hours or days later.115 Endoscopy is the most sensitive method for diagnosing this complication. Rarely, abdominal aneurysms rupture into the inferior vena cava; if so, a loud bruit can be heard. Operative management of an AAA usually consists of replacement of the aneurysm with a prosthetic graft, which may be done via laparotomy through a retroperitoneal or groin incision and use of an endovascular graft. In elective cases, preoperative angiography is useful for demonstrating additional vascular disease (e.g., stenosis or occlusion of the splanchnic arteries) and, by allowing planned vascular reconstruction, may help avoid postoperative bowel ischemia. The mortality of aneurysm repair in good-risk patients is 1% to 4%115; mortality increases sharply to 34% to 85% when surgery is done as an emergency for rupture or impending rupture.113,116,117 Increasingly, endovascular aneurysm repair (EVAR) is being used as an alternative to open repair of an AAA. Recent studies have shown that both can be performed safely in patients treated for elective infrarenal AAAs. EVAR has the perioperative advantages of reduced blood loss, and reduced length of intensive care unit and hospital stay118,119; however, concerns have been raised about endovascular leaks and late rupture.115,120 Aneurysms larger than 5 cm, symptomatic aneurysms, or enlarging aneurysms of any size should be treated electively in good-risk patients. Patients who cannot tolerate an open operation may sometimes still be treated with an endovascular graft.120-122 Patients with asymptomatic and nonexpanding aneurysms that are 4 to 5 cm in diameter are best treated conservatively because rupture of such small AAAs is rare. Aneurysms that are not treated surgically should be

followed by ultrasound every three to six months. The growth rate of AAAs is variable and has been less in recent studies than in older ones. Study of the growth rate of small aneurysms (average initial size of 4 cm) in a large population of well-studied patients revealed that over an average of 3.3 years, 58.4% of patients had no change or a decrease in aneurysm size, 25.3% had an expansion between 0.1 and 0.25 cm, 12.6 had an increase of greater than 0.25 cm, and only 3.7% had an enlargement of more than 0.5 cm.111 On average, the growth rate of an AAA is 0.35 cm per year.

MYCOTIC ANEURYSMS Mycotic aneurysms of the aorta and splanchnic vessels are rare. They were so-named by Sir William Osler because their appearance reminded him of fungi (mykes, fungus). In the past, mycotic aneurysms were most commonly caused by septic emboli from bacterial endocarditis. Today the main risk factor is intravenous drug use. Other important risk factors include contiguous spread from adjacent infectious processes, arterial manipulation, and immunocompromise (e.g., alcoholism, diabetes mellitus, chemotherapy, and treatment with glucocorticoids). Salmonella (especially Salmonella choleraesuis) and Staphylococcus are the most common infecting organisms. The celiac artery (CA) is most often affected, followed by the superior mesenteric artery (SMA) and inferior mesenteric artery (IMA). Early in the course, symptoms of mycotic aneurysms are nonspecific. Later, fever, chills, and abdominal pain are typical. Diagnosis is by imaging the vasculature: mycotic aneurysms typically are lobulated and saccular and affect the upper abdominal aorta (Fig. 36-18). The destructive process can develop quickly, leading to rapid expansion and rupture. Treatment is surgical, usually with resection of the aneurysm and vascular reconstruction.123

Figure 36-18.  Spiral computed tomographic scan of the abdomen revealing an irregular pseudoaneurysm arising from a calcified left common iliac artery. Contrast material is seen in the clot and extends outside the wall of the vessel and into the adjacent bowel. (Courtesy Dr. Alla Rozenblitt, Bronx, New York.)

607

608

Section IV  Topics Involving Multiple Organs PARAPROSTHETIC ENTERIC AND AORTOENTERIC FISTULAS An uncommon but potentially catastrophic complication of aortic aneurysmectomy and other procedures in which vascular prostheses are placed in the retroperitoneum or abdomen is the formation of a fistula between the graft and the adjacent bowel, usually the third or fourth portion of the duodenum (Fig. 36-19).124,125 The frequency of this complication is between 0.6% and 2.35%. Such fistulas develop as early as 21 days postoperatively, but in most cases, they are delayed beyond two years; in one case, an interval of 14 years was documented. This complication is thought to result from local conditions at the time of, or subsequent to, graft placement, including infection, damage to the duodenum or its blood supply during the dissection, and subsequent erosion of the duodenal wall by the graft. Newer surgical techniques, including the use of nonabsorbable sutures and antibiotics, strict hemostasis, and covering of suture lines with retroperitoneal tissue and peritoneum, as well as the increasing use of endovascular grafts, may reduce the frequency of fistula formation. Patients with aortoenteric fistulas present with upper or lower GI bleeding that if untreated may be massive and rapidly fatal. Upper GI endoscopy is the procedure of choice for aiding diagnosis by excluding other obvious lesions, but CT imaging is helpful.126 The most important clue to the diagnosis is awareness of the possibility in a patient with GI bleeding who has had an aortoiliac artery graft. Prompt diagnosis and expedient surgical repair are essential for survival.

SUPERIOR MESENTERIC ARTERY SYNDROME

25 degrees, the SMA impinges on the duodenum, thereby leading to gastric and intestinal obstruction, a condition referred to as Wilkie’s syndrome or the superior mesenteric artery syndrome (Fig. 36-20).127,128 The latter term may be confusing because the condition is not one of vascular insufficiency. Symptoms may be acute or chronic and typically include epigastric pain, vomiting, and early satiety. The syndrome has been associated with immobilization in a body cast; rapid growth in children; and marked, rapid weight loss in adults, particularly young women with an eating disorder (see Chapter 8). Rarely, anatomic anomalies predispose to the condition, including a high ligament of Treitz or low origin of the SMA. Barium studies may show an abrupt cutoff in the third portion of the duodenum with dilatation proximally, particularly when the patient is supine. Treatment approaches have included small feedings or a liquid diet. Modern imaging techniques such as CTA and MRA can provide noninvasive and detailed anatomic information that can be used in diagnosing the condition and planning surgical approaches.129,130 Symptoms typically improve after restoration of lost weight or removal of a body cast. Surgery is necessary only rarely. Duodenojejunostomy may relieve the symptoms and has been performed for this condition laparoscopically.131

CELIAC AXIS COMPRESSION SYNDROME Whether celiac axis compression syndrome (CACS) is a cause of GI ischemia has been a subject of controversy ever since the description of postprandial pain and an epigastric bruit in a patient in whom angiography showed narrowing of the CA caused by compression of a fibrotic celiac ganglion.132 After release of the artery, the murmur and

The third portion of the duodenum is cradled in an angle of about 45 degrees formed by the root of the SMA and the wall of the aorta. When this angle is narrowed to less than

Figure 36-19.  Endoscopic view of the third portion of the duodenum, where part of an aortic graft is seen. Patients with aortic graft-enteric fistulas typically present with gastrointestinal bleeding, abdominal pain, and fever because the graft usually has become infected by the time it erodes into the gastrointestinal tract.

Figure 36-20.  Film from an apper gastrointestinal and small bowel followthrough in a patient with superior mesenteric artery syndrome. The patient had symptoms compatible with gastric outlet obstruction, and on this film the second and third portions of the duodenum are markedly dilated. (Courtesy Dr. Ellen Wolf, Bronx, New York.)

Chapter 36  Vascular Lesions of the Gastrointestinal Tract

Figure 36-21.  Film from a lateral flush aortogram showing typical compression of the origin of the celiac axis with some poststenotic dilatation. The study is from a patient with no gastrointestinal complaints related to this finding. (From Boley SJ, Brandt LJ, Veith FJ. Ischemic disorders of the intestines. Curr Probl Surg 1978; 15:1.)

postprandial pain disappeared. Since that description, compression of the CA by the median arcuate ligament of the diaphragm and the celiac ganglion has been identified but is not well understood. A major difficulty in determining the validity of CACS as an entity, also sometimes referred to as Dunbar syndrome, arises from the different criteria used by various investigators to define it.133,134 At the least, clinical features that should be present to diagnose CACS include postprandial epigastric pain, diarrhea, weight loss, and an abdominal bruit that intensifies with expiration. Compression of the CA is demonstrated by lateral aortography or selective studies of the CA. Endoscopic ultrasound, CTA, and MRA are noninvasive means of demonstrating the anatomy and compression of the CA.135 Compression by the crural fibers of the diaphragm or the celiac ganglion produces a smooth, asymmetrical narrowing of the superior aspect of the celiac axis and displaces it toward the SMA (Fig. 36-21). These findings are shown best during expiration. The clinical significance of narrowing of the CA on angiography has been questioned because it occurs with equal frequency in patients in whom intestinal angina is suspected, in those with GI diseases not primarily characterized by pain, and in those with miscellaneous problems that do not involve the alimentary tract. Because the anatomic lesion that forms the basis for the syndrome is narrowing of the major artery to the upper abdominal viscera, the pain most frequently has been attributed to ischemia. This concept has persisted, despite clinical and experimental evidence that isolated compromise of the CA is almost always compensated by collateral circulation from either the SMA or the IMA. A popular alternative theory to the ischemic origin of the pain in CACS is that the pain arises in the celiac ganglion itself, possibly secondary to pressure or throbbing by the

compressed artery. The increased splanchnic blood flow and dilation of the artery that accompany the ingestion of food may explain the relationship of pain to meals. Operative approaches to CACS include division of the median arcuate ligament, with or without gangliectomy, or arterial reconstruction or bypass. Laparoscopy has been successful in releasing the compression.135 Results of operations for CACS have varied as much as have the criteria used to diagnose them. In the largest study of the long-term results of patients treated for CACS, Evans found that 83% of patients were asymptomatic 6 months after a decompression procedure, but only 41% remained asymptomatic 3 to 11 years later.136 Furthermore, no correlation existed between the presenting symptoms and the results of surgery and no clinical patterns emerged to identify those patients who might benefit from surgery. Additionally, of 12 patients treated nonoperatively, 9 remained free of pain at the time of Evans’s report. The controversy concerning CACS continues. A small number of patients who have otherwise unexplained abdominal pain not helped by standard regimens are relieved by some aspect of the operations performed for celiac axis compression.137 If surgery is performed only in patients who fulfill the criteria previously described, unnecessary procedures should be kept to a minimum.

KEY REFERENCES

Azuma H. Genetic and molecular pathogenesis of hereditary hemorrhagic telangiectasia. J Med Invest 2000; 47:81-90. (Ref 56.) Bauditz J, Lochs H, Voderholzer W. Macroscopic appearance of intestinal angiodysplasias under antiangiogenic treatment with thalidomide. Endoscopy 2006; 38:1036-9. (Ref 41.) Boley SJ, Brandt LJ. Vascular ectasias of the colon. Dig Dis Sci 1986; 31:26S-42S. (Ref 4.) Boley SJ, Sammartano RJ, Adams A, et al. On the nature and etiology of vascular ectasias of the colon. Degenerative lesions of aging. Gastroenterology 1977; 72:650-60. (Ref 1.) Brandt LJ, Spinell MK. Ability of naloxone to enhance the colonoscopic appearance of normal colon vasculature and colon vascular ectasias. Gastrointest Endosc 1999; 49:79-83. (Ref 25.) Carey EJ, Leighton JA, Heigh RI, et al. A single-center experience of 260 consecutive patients undergoing capsule endoscopy for obscure gastrointestinal bleeding. Am J Gastroenterol 2007; 102: 89-95. (Ref 23.) Chahwan S, Comerota AJ, Pigott JP, et al. Elective treatment of abdominal aortic aneurysm with endovascular or open repair: The first decade. J Vasc Surg. 2007; 45:258-62. (Ref 118.) Foutch PG, Rex DK, Lieberman DA. Prevalence and natural history of colonic angiodysplasia among healthy asymptomatic people. Am J Gastroenterol 1995; 90:564-7. (Ref 29.) Junquera F, Feu F, Papo M, et al. A multicenter, randomized, clinical trial of hormonal therapy in the prevention of rebleeding from gastrointestinal angiodysplasia. Gastroenterology 2001; 121:1073-9. (Ref 34.) Junquera F, Quirga S, Saperas E, et al. Accuracy of helical computed tomographic angiography for the diagnosis of colonic angiodysplasia. Gastroenterology 2000; 119:293-9. (Ref 7.) Kuo WT, Lee DE, Saad WE, et al. Superselective microcoil embolization for the treatment of lower gastrointestinal hemorrhage. J Vasc Interv Radiol 2003; 14:1503-9. (Ref 32.) Ripoll C, Garcia-Tsao G. Treatment of gastropathy and gastric antral vascular ectasia in patients with portal hypertension. Curr Treat Options Gastroenterol. 2007;10:483-94. (Ref 88.) Savastano S, Teso S, Corra S, et al. Multislice CT angiography of the celiac and superior mesenteric arteries: Comparison with arteriographic findings. Radiol Med (Torino) 2002; 103:456-63. (Ref 130.) Shurafa M, Kamboj G. Thalidomide for the treatment of bleeding angiodysplasias. Am J Gastroenterol 2003; 98:221. (Ref 38.) Vincentelli A, Susen S, Le Tourneau T, et al. Acquired von Willebrand syndrome in aortic stenosis. N Engl J Med 2003; 349:343-9. (Ref 18.) Full references for this chapter can be found on www.expertconsult.com.

609

CHAPTE R

37 Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm Stephan G. Wyers and Jeffrey B. Matthews

CHAPTER OUTLINE Anatomy and Physiology  611 Gross Anatomy  611 Microscopic Anatomy  612 Blood Supply and Innervation  612 Physiology  612 Surgical Peritonitis  612 Causes and Pathogenesis  612 History and Physical Examination  614 Laboratory Tests and Imaging  614 Diagnosis  615 Treatment  615 Prognosis  616 Peritonitis of Other Causes  616 Primary Peritonitis  616 Peritonitis with Continuous Ambulatory Peritoneal Dialysis  616 Tuberculous Peritonitis  616 Peritonitis Associated with Acquired Immunodeficiency Syndrome  617 Chlamydia Peritonitis  617 Fungal and Parasitic Peritonitis  617 Starch Peritonitis  617

Secondary peritonitis is often referred to as surgical peritonitis because the many and varied disease processes which present with peritonitis frequently require procedural intervention for treatment. Primary or spontaneous bacterial peritonitis is discussed in Chapter 91. In addition, this chapter discusses the primary disease processes affecting the peritoneum, mesentery, omentum, and diaphragm. Primary disease processes of these structures are often diagnosed late due to the often nonspecific and vague symptoms related to them.

ANATOMY AND PHYSIOLOGY GROSS ANATOMY

The peritoneum is a membrane covered by a single sheet of mesothelial cells, with an estimated area of 1.7 m2, similar

Rare Causes of Peritonitis  618 Intra-Abdominal Adhesions  618 Peritoneal Tumors  618 Tumors Metastatic to the Peritoneum  618 Pseudomyxoma Peritonei  619 Mesothelioma  619 Pelvic Lipomatosis  619 Benign Peritoneal Cysts  619 Diseases of the Mesentery and Omentum  619 Hemorrhage  620 Tumors  620 Inflammatory and Fibrotic Conditions  620 Infarction of the Omentum  621 Epiploic Appendagitis  621 Diseases of the Diaphragm  621 Hernias and Eventration  621 Tumors  621 Hiccups  621 Laparoscopy in the Evaluation of Peritoneal Diseases  622 General Considerations  622 Evaluation of Ascites of Unknown Origin  622 Staging Laparoscopy  622

to the total body surface area. The structure of the peritoneum is sealed in men and open to the exterior via the ostia of fallopian tubes in women. Usually the peritoneal space contains a few milliliters of sterile peritoneal fluid that may act as part of the local defense against bacteria, as well as a lubricant. The peritoneum is divided into parietal and visceral components. The parietal peritoneum covers the anterior, lateral, and posterior abdominal walls; the inferior surface of the diaphragm; and the pelvis. A large portion of the surface of the intraperitoneal organs (stomach, jejunum, ileum, transverse colon, liver, and spleen) is covered by visceral peritoneum, whereas only the anterior aspect of the retroperitoneal organs (duodenum, left and right colon, pancreas, kidneys, and adrenals) is covered by visceral peritoneum. The intraperitoneal organs are suspended by thickened bands of peritoneum, or abdominal ligaments. The nine ligaments and two mesenteries identified by Meyers and colleagues are the

611

612

Section IV  Topics Involving Multiple Organs coronary, gastrohepatic, hepatoduodenal, falciform, gastrocolic, duodenocolic, gastrosplenic, splenorenal, and phrenicocolic ligaments, the transverse mesocolon, and the small bowel mesentery.1 These ligamentous structures, which are apparent at laparotomy, as well as on computed tomography (CT), subdivide the abdomen into interconnected compartments. Familiarity with the anatomy can be used to predict the route of spread of disease; for example, the gastrohepatic and gastrocolic ligaments allow a gastric tumor to spread to the liver and colon. The spread of infection within the peritoneal cavity is governed by the site of infection, the sites of fibrinous and fibrous adhesions, intraperitoneal pressure gradients, and the position of the patient. After leakage of visceral contents, dependent recesses (e.g., paracolic gutters, pelvis, lesser sac, and subhepatic and subphrenic spaces) tend to become sites of abscess formation. For instance, patients with perforated peptic ulcer disease may present with right lower quadrant pain secondary to the dependent nature of the right lower quadrant and the right paracolic gutter. A common practice before modern imaging and percutaneous drainage methods was to place the patient in a semirecumbent position (Fowler’s position) to encourage pooling of contaminated fluids within the pelvis, in order to palpate the resultant abscess and drain it through the rectum. The mesentery is defined as a membranous bilayer of peritoneum that attaches an organ to the body wall. An omentum is a fold of peritoneum that connects the stomach with adjacent organs of the peritoneal cavity. The greater omentum spreads from the greater curvature of the stomach to the transverse colon. The lesser omentum, which joins the lesser curvature of the stomach to the liver, is called the gastrohepatic omentum. The right edge of the lesser omentum is the hepatoduodenal ligament, and the opening posterior to this (the epiploic foramen of Winslow) is the only connection between the greater and lesser peritoneal sacs.

MICROSCOPIC ANATOMY

The word peritoneum is derived from the Greek perimeaning “around” and tonos, meaning “a stretching,” therefore a “stretching around.” Mesothelium is of mesodermal origin. An interesting phenomenon is the potential for these cells to be phagocytic. Mesothelial cells are covered by microvilli in their apical surface and are joined by intercellular gaps that allow rapid absorption of fluid and particulate matter from the peritoneal cavity. The peritoneum can regenerate after injury or surgery. In animal models of abdominal wall hernias repaired with composite mesh grafts a functional neoperitoneum covers the graft in 7 to 14 days.2

BLOOD SUPPLY AND INNERVATION

The visceral peritoneum is supplied by the splanchnic blood vessels, and the parietal peritoneum by intercostal, subcostal, lumbar, and iliac vessels. The visceral peritoneum is supplied by nonsomatic nerves, whereas the parietal peritoneum is supplied by somatic nerves. Therefore, visceral pain is poorly localized, diffuse, and vague (see Chapter 10). Visceral pain is caused by stretching, distention, torsion, and twisting. The visceral peritoneum does not produce pain when it is cut or burned. When visceral pain fibers of midgut structures are stimulated, a vague periumbilical discomfort results because the visceral pain fibers enter the spinal cord at the same level as the T10 dermatome somatic fibers (see Chapters 10 and 11). This sensation is, therefore, experienced as discomfort in the dermatomal distribution. Likewise, visceral stimulation

from foregut structures produces epigastric (T8 distribution) discomfort, and visceral stimulation in the hindgut produces suprapubic (T12) discomfort. Parietal (somatic) pain fibers are activated by such stimuli as cutting, burning, and inflammation. This type of pain is sharply localized. A good example of this process is appendicitis. Early in the disease process the patient experiences periumbilical discomfort secondary to distention of the appendiceal lumen, and this progresses to localized right lower quadrant pain and tenderness as the inflammation becomes transmural and stimulates the parietal peritoneum.

PHYSIOLOGY

Particles, solutes, and fluids are absorbed from the peritoneal cavity by two different routes. Substances smaller than 2 kd may be absorbed through peritoneal mesothelial venous pores and are directed to the portal circulation.3 Particles larger than 3 kd are absorbed through peritoneal mesothelial lymphatics, entering the lymphatic thoracic duct and from there the systemic circulation.4 This last route of absorption plays an important role in controlling abdominal infections because it has a huge capacity for absorption. The anatomic structure of these large channels between the peritoneal cavity and the diaphragmatic vessels and the negative pressure of the thorax during inspiration make this mechanism extremely effective in the removal of bacteria and cells. The large surface area and semipermeability of the peritoneal membrane can be exploited therapeutically in peritoneal dialysis of patients with kidney failure and in rewarming hypothermic patients with peritoneal lavage.

SURGICAL PERITONITIS Secondary (surgical) peritonitis is a result of an inflammatory process in the peritoneal cavity secondary to inflammation, perforation, or gangrene of an intra-abdominal or retroperitoneal structure. Surgical intervention is typically required to treat these processes, although antibiotics often are useful while the process resolves (e.g., uncomplicated diverticulitis). If untreated, secondary peritonitis will, in most cases, lead to septic shock and death.

CAUSES AND PATHOGENESIS

Secondary peritonitis has numerous causes. The diagnosis is based on history, physical examination, radiographic studies, and operative exploration. History and physical examination are very important in secondary peritonitis, and a good history and physical examination can often obviate further studies. Some of the more common causes of secondary peritonitis include perforated peptic ulcer disease, appendicitis, diverticulitis, acute cholecystitis, and postsurgical complications. Other nonbacterial causes of peritonitis include leakage of blood into the peritoneal cavity due to rupture of a tubal pregnancy, ovarian cyst, or aneurysmal vessel. Blood is highly irritating to the peritoneum and may cause abdominal pain similar to that found in septic peritonitis. Bile leakage into the peritoneal cavity also can cause signs and symptoms of peritonitis, especially when there is also bacterial contamination of the bilious contents. However, pure bile in the abdomen can be surprisingly asymptomatic. Large bilomas may have minimal symptoms. Bacteria reach the peritoneal cavity by a variety of pathologic processes: transmural inflammation with luminal obstruction (see Chapter 119), perforation of the gastrointestinal (GI) tract, and ischemia (see Chapter 114). The initial

Chapter 37  Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm inoculum of bacteria is determined by the normal flora in the involved portion of the GI tract (see Chapter 102).

Flora

Although the flora of the gut, especially of the large bowel, is diverse and extensive, the numbers of types of organisms rapidly decrease after leakage of gut contents into the peritoneal cavity.5 Aerobes such as Escherichia coli and enterococci and anaerobes such as Bacteroides fragilis and Clostridium organisms predominate. A recent study of infections associated with ruptured colonic diverticulitis reported anaerobes only in 15% of cases, aerobic bacteria only in 11%, and mixed aerobic and anaerobic flora in 74%; cultures from peritoneal abscesses detected anaerobic bacteria in 18%, aerobes alone in 5%, and mixed aerobic and anaerobic flora in 77%.6 In addition to bacteria, the presence of fungi in intra-abdominal infection has recently been more frequently recognized and may have clinical significance. For instance, a positive fungal culture is quite common in perforated peptic ulcer disease and may adversely affect outcome.7 On the basis of an animal model of monomicrobial and polymicrobial peritonitis with various combinations of bacteria, it is apparent that (a) E. coli is the organism most often responsible for death from this form of iatrogenic peritonitis, at least in part because of its ability to cause bacteremia, and (b) that combinations of anaerobes and facultative organisms lead to abscess formation.8 As stated, 77% of bacterial cultures from peritoneal abscesses are polymicrobial.6 Other adjuvant substances, such as devitalized tissue, mucus, bile, hemoglobin, and barium, can act synergistically with microorganisms to increase mortality in surgical peritonitis through their ability to interfere with phagocytosis and killing of bacteria. These considerations form the basis for the treatment of surgical peritonitis, which is described later. The peritoneal cavity possesses several lines of defense against bacterial infection (Table 37-1). Peritonitis results when these are overwhelmed.

Peritoneal Clearance of Bacteria

Once bacteria enter the peritoneal cavity, clearance of the offending microorganisms begins immediately. Within 6 minutes of intraperitoneal inoculation of bacteria in dogs, bacteria can be cultured in thoracic lymph, indicating

Table 37-1  Peritoneal Defense Mechanisms Against Bacteria Removal Mechanisms Peritoneal clearance of bacteria through the diaphragm via the thoracic duct Leukocyte-Attracting Mechanisms Microvilli of the mesothelial cell ICAM-1 (CD 54) and VCAM-1 (CD 106) Killing Mechanisms Macrophages Neutrophils Opsonins Complement C3b Immunoglobulin G Fibronectin Mast cell–derived leukotrienes Sequestration Mechanisms Fibrin trapping of bacteria Formation of fibrinous adhesions Omental loculation of foci of inflammation ICAM, intercellular adhesion molecule; VCAM, vascular cell adhesion molecule.

passage of organisms through the diaphragm. Twelve minutes later, bacteremia may be evident. This clearance mechanism is probably important in survival because blockade of the thoracic duct in an animal model of peritonitis decreases bacteremia episodes4 but increases mortality and induces liver necrosis. This appears to be directly related to the amount of endotoxin to which the liver is exposed.9 Decades before it was known that the diaphragm was the predominant site of clearance of bacteria, Fowler, in 1900, proposed his head-up, pelvis-down position for prevention of absorption of toxins from infected peritoneal cavities. In the preantibiotic era, documentation of the delayed clearance of bacteria from experiments in infected dogs in the head-down position confirms the wisdom of this positioning for patients with peritonitis.

Killing Mechanisms

In addition to mechanisms of bacterial clearance through the diaphragm, intraperitoneal defense mechanisms include cellular and humoral responses (see Chapter 2). Macrophages and neutrophils are attracted to the peritoneal cavity, and in this setting, microvilli of the mesothelial cells play a significant role in leukocyte migration into the peritoneal cavity by providing the needed substrates for their adhesion, namely intercellular adhesion molecule-1 (ICAM-1, or CD 54), and vascular cell adhesion molecule-1 (VCAM-1, or CD 106).10 The degree of cellular recruitment may be a key factor in a patient’s survival because a prolonged peritoneal inflammatory response has been observed to be adversely correlated with survival in an animal model of peritonitis.11 Humoral antibacterial agents, such as complement factors, fibronectin, and globulins, are released into the peritoneal cavity. These opsonins coat bacteria and render them recognizable as foreign; then they are entrapped and killed by phagocytes.12

Sequestration Mechanisms

Sequestration mechanisms include fibrin trapping of bacteria, fibrinous adhesions, and omental loculation of foci of infection (see Table 37-1).13 It has been known since 1950 that bacteria are more readily destroyed on a surface than in a liquid medium. The microscopic and macroscopic networks of surfaces provided by fibrin and the omentum assist phagocytes in locating, trapping, ingesting, and killing bacteria. The volume of peritoneal fluid in which infection develops has a remarkable effect on mortality; 20% of rats inoculated with E. coli diluted in 1 mL of saline die, whereas 75% of rats inoculated with the same number of viable bacteria but diluted in 30 mL of saline die.14 This phenomenon explains in part the risk of development of spontaneous bacterial peritonitis in relation to the ascitic fluid total protein concentration.15 The more voluminous the ascitic fluid, the lower the concentration of proteins and opsonins, the less efficient the trapping of bacteria, and the higher the risk of an uncontrolled infection (see Chapter 91). Patients undergoing chronic ambulatory peritoneal dialysis may be vulnerable to peritonitis because of dilution of opsonins by dialysis fluids. Bacterial contamination in the peritoneal cavity and the subsequent response of immune cells such as neutrophils and macrophages lead to an inflammatory response including the release of cytokines. The systemic inflammatory response syndrome (SIRS) is marked by fever, a hyperdynamic cardiovascular response, muscle protein breakdown,16 and respiratory failure. If the underlying cause is treated by surgical intervention, antibiotics, or the body’s own defense mechanisms, these processes can be thwarted

613

614

Section IV  Topics Involving Multiple Organs or reversed. However, if the process goes unchecked, multisystem organ failure and death will result. In addition, even if the underlying cause is treated, the inflammatory response can lead to multisystem organ failure and death if the treatment is delayed or the inflammatory response is particularly vigorous. Patients with severe peritonitis may have a higher mortality from a shift from type 1 to type 2 T-helper cells leading to greater immunosuppression.17 When treating peritonitis or operating within the abdomen, the clinician’s goal is to minimize or eliminate inflammation. For instance, laparoscopic operations may induce less of a systemic inflammatory response than their open counterparts.18,19 In addition, laparoscopy differs from laparotomy in regard to peritoneal macrophage response,20 less cortisol release,21 and less reduction in natural killer (NK) cell subsets.22 Laparoscopic operations may well confer an immunologic advantage over conventional open operations.23 The additional benefits of smaller incisions, less tissue trauma, decreased postoperative pain, and shorter recovery are driving a trend to laparoscopic operations over open operations even in acute settings.

HISTORY AND PHYSICAL EXAMINATION

Clinical history and careful physical examination are the key factors in making a timely diagnosis of surgical peritonitis. In general, the sooner the diagnosis is made, the better the prognosis. Abdominal pain is the hallmark of peritonitis. The exact details of the onset of pain can be helpful in drawing attention to the affected organ (see Chapter 10). The pain’s character, location, area of radiation, change over time, and provocative and palliative factors are key pieces of information in assisting with the diagnosis. Peritoneal inflammation is typically associated with ileus, and therefore nausea and vomiting are common symptoms. The ability of the clinician to elicit an accurate history of abdominal pain and peritoneal signs is limited in patients with neurologic and immunologic compromise. The pain of peritonitis can be reduced or even absent in older adult patients. Infants and children may be incapable of furnishing any history or cooperating with the physical examination. Notoriously difficult patients to assess for secondary peritonitis include emergency room patients under the influence of alcohol or illicit drugs, trauma patients with central nervous system or spinal cord injuries, and sedated and ventilated intensive care unit (ICU) patients. Analgesics typically will not relieve the pain of peritonitis on examination but may relieve some discomfort as related to the history of present illness. In fact, it has been shown that early provision of analgesia to patients with undifferentiated abdominal pain does not affect diagnostic accuracy.24 Diabetic patients have deficits in neurologic and immune function. Patients receiving immunosuppressive and antiinflammatory drugs, such as glucocorticoids and chemotherapeutic drugs, may have blunted perception of pain and minimal signs of peritoneal irritation. Patients with cirrhosis and ascites may show no pain during episodes of spontaneous bacterial peritonitis unless the parietal peritoneum becomes involved with the inflammatory process (see Chapter 91). On examination, the patient with surgical peritonitis is usually immobile because any movement acutely worsens the pain. Fever of 100° F or higher is typical, as is tachycardia, which may be in part secondary to pain. Hypotension is usually a late finding accompanying sepsis. Fever is a basic endogenous mechanism to help fight infection. In fact, the increase in body temperature that is usually found during bacterial infections, including peritonitis, seems to be essential for optimal host defense against bacteria.25 The

absence of percussible hepatic dullness suggests the presence of free air in the peritoneal cavity. Exquisite tenderness to percussion should lead to very gentle palpation. Overly vigorous palpation of a very tender abdomen may cause patients such pain that they are subsequently unable to cooperate for the remainder of the examination. Palpation should begin farthest from the area that the patient identifies as the source of the most pain. Palpation of a truly boardlike abdomen is so impressive to the examiner that it cannot be forgotten. Lesser degrees of rigidity must be compared with this extreme end of the spectrum. Voluntary guarding in the presence of mild tenderness may be misinterpreted as rigidity by the inexperienced examiner if the patient is anxious and palpation too vigorous. It is usually not necessary to check for rebound tenderness to palpation if rebound tenderness is noted during auscultation or percussion. Often, the presence of rebound tenderness can be inferred if the patient’s pain is exacerbated when the bed or stretcher is jarred. Peritoneal signs signify inflammation of the parietal peritoneum secondary to an intra-abdominal process. Peritoneal signs consist of rebound tenderness, involuntary guarding and extreme tenderness on palpation. Peritonitis can be diffuse, such as that associated with perforated ulcer, or localized, such as that of diverticulitis confined to the left lower quadrant. Significant septic processes may be confined to the pelvis by overlying bowel and omentum with a resulting absence of peritoneal signs in the anterior abdominal wall. Therefore, careful rectal and pelvic exams are essential in order to detect pelvic peritonitis. The presence of iliopsoas and obturator signs (described in Chapter 116) can be helpful in detecting retroperitoneal or pelvic inflammation and abscesses. Repeated physical examinations by the same examiner will provide evidence of progressive peritoneal irritation. The evolution of the physical exam over time provides additional information for diagnosis and evaluation of response to initial conservative therapy. This, together with laboratory tests and imaging procedures described below, will indicate the need for surgical intervention.

LABORATORY TESTS AND IMAGING

The most common laboratory sign of peritonitis in an immunocompetent patient is an increased white blood cell count with left shift. The presence of circulating juvenile forms (e.g., bands) is a reflection of an increasing demand of white cells from the bone marrow. A low white blood cell count in the course of a bacterial infection associated at times with gram-negative septicemia may indicate the presence of an exhausted bone marrow, with a poorer prognosis. In addition, metabolic acidosis, hemoconcentration, and prerenal azotemia may be present. Free air may be detected on upright chest radiograph or on upright or decubitus abdominal films, but this finding may be only 60% sensitive in detecting gut perforation.26 The absence of free air should not delay surgical intervention in an otherwise appropriate clinical setting. Ultrasonography can be helpful in demonstrating abscesses, bile duct dilatation, and large fluid collections. CT scan of the abdomen and pelvis, generally with both oral (occasionally rectal) and intravenous contrast, is increasingly preferred as the most sensitive and specific imaging modality for acute abdominal pain. Multidetector CT scanners are capable of imaging the entire abdomen and pelvis in a single breathhold. The axial images are of extremely high resolution and can be reconstructed in coronal, sagittal, and three dimensional sets of images.27 CT is much more sensitive than plain films for the detection of free air, and with multidetector CT it is possible to visualize the actual site of perforation.28

Chapter 37  Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm Although CT images are increasingly accurate and the images compelling, they should not delay surgical consultation, resuscitation, and operation in a patient with suspected peritonitis.

DIAGNOSIS

The diagnosis of surgical peritonitis is suspected on the basis of history, physical examination, and laboratory and imaging tests and is confirmed at laparotomy or laparoscopy when purulent fibrinous peritonitis is found. In those patients whose history and physical examinations are unreliable, CT and peritoneal lavage are extremely valuable in confirming the diagnosis of surgical peritonitis. CT is less invasive but peritoneal lavage can be performed quickly in hemodynamically unstable patients. Peritoneal lavage is performed by inserting a catheter under sterile conditions into the peritoneal cavity and infusing 1 L of normal saline. If the effluent contains more than 500 white blood cells (WBCs) per cubic millimeter, an amylase or bilirubin level greater than the corresponding serum value, or bacteria on Gram stain, there is approximately a 90% likelihood of surgical peritonitis.29 Laparotomy is usually indicated in this setting. Finally, diagnostic laparoscopy is extremely accurate in making the diagnosis of surgical peritonitis and many of the underlying diseases can be dealt with laparoscopically, avoiding the need for laparotomy.30

TREATMENT

The following two principles in the management of surgical peritonitis cannot be overemphasized. First, not all patients with peritonitis require surgery. For example, a patient with localized left lower quadrant peritonitis secondary to diverticulitis can be managed with bowel rest and intravenous antibiotics alone. Another patient with the same clinical presentation and findings of a diverticular abscess on CT scan can be successfully treated with antibiotics and percutaneous drainage (see Chapter 26). Second, the absence of peritonitis does not exclude the possibility of surgical emergency. The classic example of this clinical situation is early acute mesenteric ischemia with abdominal pain out of proportion to findings on physical examination findings. Likewise, a complete mechanical small bowel obstruction without peritoneal signs, an indication of perforation or vascular compromise, still requires operation. For most cases of secondary peritonitis fluid resuscitation and antibiotic therapy followed by urgent laparotomy or laparoscopy are the mainstays of treatment. Fluid resuscitation is guided by frequent monitoring of physiologic parameters in an ICU, including blood pressure (by arterial line if shock is present), heart rate, central venous pressure or pulmonary capillary wedge pressure, and urine output. Hematocrit, WBC, electrolytes, glucose, creatinine, and blood gases should also be monitored. Hypovolemia, hypotension, metabolic acidosis, hypoxia, and hemoconcentration from loss of plasma into the peritoneal cavity are expected. Glucocorticoids have been shown not to provide benefit in the setting of septic shock.31 The patient should be aggressively fluid resuscitated to treat intravascular fluid depletion secondary to movement of fluid out of the vascular space. Pressors are generally to be avoided, if possible, and surgical intervention should be pursued when indicated as soon as the patient is hemodynamically stable for operation.

Antibiotics

Antibiotic therapy is required before, during, and after surgical intervention. The type of bacteria causing secondary peritonitis depends in part on the normal flora of the part of the GI tract that is the source of sepsis and in part on the

clinical setting. In community-acquired peritonitis, susceptible gram-negative bacilli, strict anaerobic bacteria, and enterococci are typically found. In general, antibiotics directed against the most likely pathogens should be chosen. For instance, colonic processes require coverage for gramnegative aerobes and anaerobes. In animal models, antibiotics directed against gram-negative enteric aerobic organisms minimize mortality, and drugs effective against anaerobes prevent abscess formation.32 It has been shown in experimental models of peritonitis that there is synergism between aerobic and anaerobic bacteria.33 The coverage of all potential organisms is not necessary.34 The flora of surgical peritonitis simplifies with time, even before initiation of antibiotics. Killing certain key species may change the microenvironment sufficiently to prevent growth and allow killing of other flora. If a Candida species is cultured from the peritoneal cavity in a patient with secondary peritonitis, this organism should be treated if the patient is in septic shock or is immunocompromised despite being hemo­ dynamically stable. Hemodynamically stable immunocompetent patients with secondary peritonitis do not need treatment for Candida.35 A variety of antibiotic regimens have been proposed using the following classes of antibiotics alone or in combination: second-generation cephalosporins, third-generation cephalosporins, broad-spectrum beta-lactams, fluoroquinolones and metronidazole, and aminoglycosides with clindamycin or metronidazole. Many controlled trials of antibiotic regimens show equivalency. For example, it has been shown that monotherapy with a broad-spectrum beta-lactam is as effective as combination therapy with a beta-lactam and an aminoglycoside.36 Data-supported guidelines regarding optimal treatment have been hampered by suboptimal study design and nonuniform efficacy criteria in the controlled trials that have been performed. A recent Cochrane review of 40 randomized trials involving 16 different regimens showed no difference in mortality.37 The specific antibiotics chosen should take into account other considerations such as the avoidance of toxicities, the sensitivity profile of cultured organisms, the ease and route of administration, and cost. The availability of broad-spectrum antibiotics, including betalactams, fluoroquinolones, and third- and fourth-generation cephalosporins, makes it unnecessary to use aminoglycosides with their potential nephrotoxicity in patients with compromised renal function.36 The failure to clear secondary peritonitis after an appropriate course of antibiotic therapy or the recurrence of peritonitis is termed tertiary peritonitis. Nosocomial infections occurring in patients after long periods of hospitalization may include infections with multiresistant Pseudomonas, Enterobacter, Enterococcus, Staphylococcus, and Candida species. The development of multiple organ dysfunction syndrome (MODS) after an initial operation should prompt an aggressive search for inadequate source control and abscesses, involving repeat CT scans, percutaneous or operative drainage, and culture of persistent fluid collections, in addition to antimicrobial therapy.38

Surgical Intervention

Antibiotics help treat or prevent fatal bacteremia but do not cure most patients with surgical peritonitis unless operative intervention is also undertaken. Neither free leakage of gut contents nor large abscesses can be sterilized by antibiotics alone in the absence of drainage. Surgical intervention should occur as soon as possible after the patient is stabilized and resuscitated and antibiotics have been given. Laparotomy remains the gold standard for definitive diagnosis and mainstay of therapy in surgical peritonitis.

615

616

Section IV  Topics Involving Multiple Organs However, a recent review confirms the success of an increasing number of laparoscopic procedures for some forms of peritonitis.30 With either laparoscopic or conventional open operations, the aims of surgical treatment include source control, peritoneal decontamination, and prevention of recurrent infection. Repeat laparotomy and laparostomy, in which the abdomen is left open, are useful tools when control of the source of infection is not possible at the initial operation.38 Surgical re-exploration may be undertaken for the following reasons: (1) tenuous control of the source of infection, (2) reassessment of bowel viability, (3) inadequate or poor drainage, (4) hemodynamic instability, (5) infected pancreatic necrosis or diffuse fecal peritonitis at the initial operation, (6) reassessment of a tenuous anastomosis, and (7) the development of intra-abdominal hypertension (abdominal compartment syndrome). This syndrome is described in more detail in Chapter 10. An abdominal compartment syndrome results when the closure of the abdomen at either the level of the fascia or skin causes intra-abdominal pressure to rise to a degree that impairs respiratory, hepatic, and renal function.39 Preoperative and postoperative fluid and nutritional support are crucial to prompt wound healing and survival. Peritonitis has been compared with a 50% total body surface area burn, and even a calorie intake of 3000 to 4000 kcal per day may not achieve a positive nitrogen balance. Inability to achieve positive nitrogen balance may, however, be secondary to accelerated proteolysis11 and negative nitrogen balance associated with pathologic proteolysis will not be treated by any amount of caloric intake. This proteolysis may only be thwarted with treatment of the septic process and recovery of the patient. The enteral route of nutrition is preferred over parenteral. Placement of a feeding jejunostomy tube at the initial operation is prudent in these critically ill patients.

PROGNOSIS

Despite the modern approach to the diagnosis and treatment of surgical peritonitis, mortality remains high in certain subgroups of patients, especially older adult patients and patients who suffer multiple organ failure before the development of peritonitis.40 In general, peritonitis-related mortality may be as low as 14%,41 with appendicitis and perforated duodenal ulcer at the low end of the spectrum (10%) and postoperative (tertiary) peritonitis at the high end (as high as 50%).40

PERITONITIS OF OTHER CAUSES (see Table 37-2) PRIMARY PERITONITIS

Spontaneous bacterial peritonitis (SBP), or peritonitis without a known surgical source, is the most common cause of primary peritonitis. This occurs predominantly in patients with cirrhosis and ascites and is discussed in Chapter 91. Primary peritonitis may also occur in patients with ascites due to nephrotic syndrome.42 Primary peritonitis in the absence of cirrhosis or nephrosis is much less common and usually occurs in children. Primary peritonitis is treated without surgical intervention, using antibiotics directed against the offending organism.

PERITONITIS WITH CONTINUOUS AMBULATORY PERITONEAL DIALYSIS

Continuous ambulatory peritoneal dialysis (CAPD) is a common treatment of end-stage kidney disease.43 Bacterial

Table 37-2  Causes of Nonsurgical Peritonitis Spontaneous bacterial peritonitis (see Chapter 91) Chronic ambulatory peritoneal dialysis Mycobacterium tuberculosis Acquired immunodeficiency syndrome-associated Chlamydia trachomatis Neisseria gonorrhoeae (Fitz-Hugh–Curtis syndrome) Rare causes Polyarteritis nodosa Systemic lupus erythematosus Scleroderma Familial Mediterranean fever

peritonitis develops in this setting about 1.4 times per patient-year of treatment.44 The most common isolates in patients treated with CAPD are Staphylococcus epidermidis and other skin flora.45 Other pathogens, such as fungi or Mycobacterium tuberculosis, are less frequent. The most probable explanation for this high incidence of infection is inadvertent contamination of the indwelling catheter. Even with better patient education regarding sterile technique, peritonitis in this group of patients is a major source of morbidity and the largest single cause of patient failure on CAPD.46 New technical maneuvers47 or special management of insertion site48 may decrease the incidence of infections in these patients. Abdominal pain and tenderness are found in about 75% of patients, but fever is found in only about one third.49 A consistent feature is cloudy effluent, noted in 98%.50 The diagnosis is suspected on the basis of signs and symptoms and is confirmed by a fluid WBC count greater than 100 neutrophils/mm3 or the presence of organisms on Gram stain. Treatment should be started immediately without waiting for the culture results, similar to the empiric treatment of patients with cirrhosis and neutrocytic ascites.50 Initial treatment of suspected CAPD peritonitis should cover the most frequently isolated bacteria. Vancomycin and second- or third-generation cephalosporins are good options. The intraperitoneal route of administration is probably the most effective.50 The sensitivity of the organism isolated determines the subsequent antibiotic choice. Most of these patients are successfully treated on an outpatient basis without stopping dialysis. Prompt treatment ensures survival; however, recurrent infection is common and may lead to catheter removal or scarring of the peritoneum and poor dialysis exchange. Addition of heparin to the dialysis bag in cases of peritonitis may decrease the formation of fibrin and thereby the incidence of postinfection adhesions. However, these infections often require removal of the catheter if they do not respond to antibiotic treatment. Repeated infections lead to sclerosing encapsulating peritonitis (abdominal cocoon syndrome) and loss of surface area for effective dialysis.

TUBERCULOUS PERITONITIS

The number of patients with tuberculous peritonitis has increased in recent years, due in part to the development of this disease in patients with acquired immunodeficiency syndrome (AIDS), with a high rate of multiresistant strains of M. tuberculosis.51 Noncirrhotic patients with this form of peritonitis usually have ascites with a high protein content, low glucose concentration, and a low serum-to-ascites albumin gradient of less than 1.1 g/dL.52 Patients almost always have an elevated ascitic fluid WBC count and a lymphocytic predominance. The algorithm in evaluation of patients with

Chapter 37  Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm high-lymphocyte-count ascites includes cytologic evaluation of the fluid and consideration of laparoscopy.52 Patients with lymphocytic ascites and fever usually have tuberculosis, whereas afebrile patients usually have malignancyrelated ascites. Cancer is the cause of lymphocytic ascites about 10 times more frequently than is tuberculosis (see Chapter 91). If peritoneal carcinomatosis is present, the cytologic findings are positive more than 90% of the time, and the laparoscopy can be avoided.53 If the cytology is negative, laparoscopy is performed and is nearly 100% sensitive in detecting tuberculous peritonitis. Tuberculous peritonitis may also appear in a miliary form or as a pelvic mass with high serum levels of CA125, making the diagnosis difficult to distinguish from metastatic ovarian cancer.54 Adenosine deaminase levels are typically elevated in the ascitic fluid in tuberculous ascites, and this finding can help differentiate tuberculous peritonitis from other causes of peritonitis and ascites. A six-month treatment course consisting of isoniazid, rifampin, and pyrazinamide for the first eight weeks, followed by isoniazid and rifampin for the next four months, is considered adequate.55 More antituberculous drugs may be necessary, depending on local susceptibility testing. More than half of patients with tuberculous peritonitis in the United States have underlying cirrhosis, usually alcohol related,56 whereas in Third World countries, peritoneal tuberculosis usually occurs in the absence of cirrhosis. The presence of cirrhosis affects the results of ascitic fluid tests, including reducing the sensitivity of adenosine deaminase to only 30% (see Chapter 91).56 Furthermore, ascites in tuberculous peritonitis may diminish or disappear with diuretics, but fever usually persists, as does a high ascitic fluid leukocyte count. Antituberculous therapy must be supervised carefully by public health personnel, as well as physicians. Erratic treatment leads to emergence of resistant strains.

PERITONITIS ASSOCIATED WITH ACQUIRED IMMUNODEFICIENCY SYNDROME (see Chapter 33)

Patients with AIDS may develop peritonitis from many different pathogens: bacteria (monomicrobial or polymicrobial); viruses (cytomegalovirus, herpes, and others) and fungal organisms (Histoplasma, Cryptococcus, and Coccidioides); parasites (Pneumocystis jiroveci, Trypanosoma cruzi); and mycobacteria (M. tuberculosis and Mycobacterium avium-intracellulare). Also, neoplastic lesions, such as Kaposi’s sarcoma and non-Hodgkin’s lymphoma, may metastasize to the peritoneum. Like other forms of peritonitis, the common features of presentation are abdominal pain; anorexia; fever; and ascites, which typically has a high protein content. The diagnosis of a rare form of peritonitis with one of these organisms sometimes leads to a diagnosis of AIDS57 in a human immunodeficiency virus (HIV)positive patient. The treatment of these opportunistic infections involving the peritoneum is generally pharmacologic (e.g., antibiotics, amphotericin B, ganciclovir) unless bowel involvement has led to gut perforation, which may occur with cytomegalovirus, for instance. Also, laparotomy may be indicated for obstructive symptoms, as with lymphoma. Bowel resection is required in this instance. With advances in highly active antiretroviral therapy (HAART) there has been a three-fold reduction in the mortality rate from AIDS from 1995 to 2002,58 a decline in the prevalence of opportunistic GI disease,59 and a major decrease in the number of operations for AIDS-related surgical illness.58

Figure 37-1.  Laparoscopic photograph of perihepatitis (Fitz-Hugh–Curtis syndrome) showing adhesions on the surface of the liver. (From Frumovitz MM, eMedicine.com, Inc., 2004.)

CHLAMYDIA PERITONITIS

Fitz-Hugh–Curtis syndrome, or perihepatitis (Fig. 37-1), was formerly most commonly associated with Neisseria gonorrhoeae. However, in recent years Chlamydia is increasingly implicated in perihepatitis.60 Chlamydia perihepatitis occurs only in women, owing to seeding of bacteria into the peritoneal cavity from the fallopian tubes. Symptoms presenting in these patients include inflammatory ascites, pain in the right upper abdominal quadrant, fever, and a hepatic friction rub. If there is enough ascitic fluid to be clinically detectable, it has an elevated white cell count with a predominance of neutrophils and a high protein content, even in excess of 9 g/dL.60 Laparoscopy is very helpful in confirming the diagnosis, revealing “violin strings” and “bridal veil” adhesions from the abdominal wall to the liver. Doxycycline is usually curative. Also, these adhesions may be an incidental finding during laparoscopy or laparotomy for another reason. In this situation, no treatment is required.

FUNGAL AND PARASITIC PERITONITIS

Fungal peritonitis can be due to gut perforation, especially perforation of the upper gastrointestinal tract. It can also be a complication of acquired immunodeficiency (see Chapter 33). Fungal peritonitis may be limited to the pelvis in cases of gynecologic dissemination; this may be treated with fluconazole.61 The most common isolate is Candida spp., probably because routine blood culture media can detect Candida. Although infrequent, fungal peritonitis has been described in patients undergoing chronic ambulatory peritoneal dialysis.62 Although rare in the United States, peritoneal histoplasmosis, coccidioidomycosis, and cryptococcal infection are increasing in frequency in the setting of acquired immunodeficiency. Schistosomiasis, pinworms, ascariasis, strongyloidiasis, and amebiasis also may involve the peritoneal cavity (see Chapters 109 and 110).

STARCH PERITONITIS

Years ago, approximately 1 of 1000 patients who underwent laparotomy developed fever and migratory abdominal pain two to three weeks postoperatively due to contamination of the peritoneum by glove powder starch. This is much less frequent today63 probably because starch has been replaced by other more inert substances. Glove powder is known to be a source of formation of abdominal granulomas.64 Glove powder granulomas also may mimic peritoneal carcinoma-

617

618

Section IV  Topics Involving Multiple Organs tosis. These lesions should be biopsied and sent for frozen section if the etiology is in question and if the results could change the operative procedure.65 Starch peritonitis is a difficult diagnosis to make, and a high index of suspicion is required. Treatment is nonoperative, and glucocorticoids may be of benefit.66

RARE CAUSES OF PERITONITIS

Connective tissue diseases lead to peritonitis as a manifestation of serositis in approximately 5% of patients with lupus and approximately 10% of patients with polyarteritis and scleroderma.67 Treatment of the underlying disease usually controls the serositis (see Chapter 35). Familial Mediterranean fever is an autosomal recessive hereditary disease that affects the peritoneum, as well as other serous membranes. It is more frequently found in patients of Ashkenazi Jewish, Armenian, and Arabic ancestry. It is an aseptic form of recurrent peritonitis; no infectious agent has been observed to be related to this disease. Patients usually present with sporadic episodes of abdominal pain and fever, and synovitis and pleuritis may also be present. Treatment with colchicine appears to prevent attacks and can prevent fatal renal amyloidosis (see Chapter 34).68

INTRA-ABDOMINAL ADHESIONS

The aftermath of secondary peritonitis and the surgery to correct it is the variable formation of intra-abdominal adhesions, abnormal fibrous bands between peritoneal surfaces that are usually separate. Adhesions may be congenital, but the vast majority is acquired as result of peritoneal injury. Intraperitoneal foreign bodies such as suture material, clips, and mesh also contribute to adhesion formation. Intraabdominal adhesions can be a considerable source of morbidity and mortality. They are the most common cause of small bowel obstruction (Chapter 119). Adhesions are a leading cause of secondary infertility in women, accounting for 15% to 20% of cases. Pelvic adhesions may be a source of chronic lower abdominal and pelvic pain. Adhesions may preclude peritoneal dialysis or intraperitoneal chemotherapy should they be necessary. Extensive adhesions may preclude laparoscopic procedures and have been shown to increase blood loss, operative time, and risk of enterotomy in reoperative surgery. These patients are then at increased risk for postoperative complications and prolonged hospital stay. The socioeconomic cost of adhesive disease is considerable.69 Formidable effort has been devoted to the prevention of adhesion formation. Tissue damage, hemorrhage, and inflammation in the peritoneal cavity lead to fibrin deposition on the peritoneal surfaces allowing adjacent surfaces to adhere in this sticky matrix. Various strategies for prevention of adhesion formation include reduction of peritoneal injury, inhibition of the inflammatory response, prevention of fibrin formation, promotion of fibrinolysis, prevention of collagen deposition, and barrier separation of the peritoneal surfaces. Although various experimental strategies in animal experiments have reduced the number and severity of adhesions, few of these have translated into clinical practice. The preponderance of evidence in human as well as animal studies shows decreased adhesions at the incision sites and at the operative site in laparoscopic surgery compared with open surgery.70 Seprafilm, a hyaluronic-carboxymethylcellulose membrane, has been shown in human trials to reduce intra-abdominal adhesions after general surgical procedures but there has been no demonstrable reduction in bowel obstruction. The risk of abscesses and anastomotic leaks was increased.71

PERITONEAL TUMORS TUMORS METASTATIC TO THE PERITONEUM

Metastatic cancer is by far the most common peritoneal tumor (Fig. 37-2). Although it is frequently assumed that tumors cause ascites only when malignant cells line the peritoneal cavity (i.e., peritoneal carcinomatosis), extraperitoneal tumors, including massive liver metastases, hepatocellular carcinoma with or without cirrhosis, malignant lymph node obstruction as in lymphoma, and Budd-Chiari syndrome with or without inferior vena cava obstruction, are associated with ascites.53 Ascitic fluid characteristics often allow their distinction,53 which is important because each may require different treatment (see Chapter 91 for details of pathogenesis and ascitic fluid analysis). Tumors that preferentially metastasize to the peritoneum include adenocarcinomas of the ovary, stomach, colon, breast, pancreas, and lung, as well as lymphoma and other sarcomas (see Fig. 37-2).

Clinical Features

Ascites usually appears in patients as evidence of advanced disease of a known tumor with a large burden, rather than as a primary manifestation of cancer. Weight loss, abdominal pain, and early satiety are common. Ascites in a middleaged woman without risk factors for liver disease may be the first manifestation of peritoneal spread of an ovarian cancer; the prognosis in this situation is better than that of nonovarian cancer (see later). Patients with malignancy-related ascites of recent onset usually tolerate its presence poorly, probably because of less compliance of the abdominal wall compared with patients with cirrhosis who have chronic ascites. As the malignancy progresses, the fluid component tends to be replaced by solid tumor, leading to bowel obstruction. Some common myths about peritoneal carcinomatosis are that the cytology is insensitive and that the fluid is frequently bloody (see Chapter 91).

Treatment

Paracentesis Therapeutic paracentesis for symptomatic palliation is the mainstay of treatment for the majority of patients with peri-

Figure 37-2.  Intraoperative photograph of peritoneal carcinomatosis. (From Free Picture Gallery, 2001. www.laparoscopyhospital.com/gallery6. htm.)

Chapter 37  Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm toneal carcinomatosis. The recommendation to use diuretics for treatment was based largely on supposition rather than hard data. A study of ascitic fluid volume and blood volume in patients with peritoneal carcinomatosis who lost weight taking large doses of diuretics demonstrated that the weight was lost at the expense of blood volume, not ascitic fluid volume. The characteristics of the ascitic fluid may help direct diuretic use. In general, ascites with a high serum-ascites albumin gradient (≥1.1g/dL) responds to diuretics.72 Therefore, in cancer patients, diuretics should be reserved for those with edema or some specific indication other than peritoneal carcinomatosis.

Lymphatic or extraperitoneal spread of the tumor is rare. Presenting symptoms and signs include painless abdominal distention and an ovarian mass; mucin may accumulate intraperitoneally many years after resection of an ovarian mass.81 Definitive diagnosis is made when the jelly-like material is encountered at laparotomy or laparoscopy. However, there are often characteristic findings on CT that can suggest the diagnosis preoperatively.82 Cytoreductive surgery with intraperitoneal hyperthermic perfusion is effective current treatment for pseudomyxoma peritonei and has acceptable morbidity and mortality rates.83 The largest series of patients reports an operative mortality of 2%, morbidity of 40%, and a median survival of 13 years.84 Unfortunately, recurrence usually causes bowel obstruction, malnutrition, and death. The optimal treatment for mucinous cystadenocarcinoma of the appendix in the presence of pseudomyxoma peritonei is right hemicolectomy with aggressive tumor debulking.85

Surgery and Intraperitoneal Chemotherapy Because the usual response to routine therapy is poor, new treatments have been suggested, such as peritonectomy combined with hyperthermic antiblastic perfusion.73 Recent results show that cytoreductive surgery and intraperitoneal hyperthermic chemoperfusion can be performed with success.74 In some instances this approach allows better survival for patients with extensive carcinomatosis who were no longer responsive to traditional therapies. The rationale for application of intraperitoneal chemotherapy is that its use would allow larger local concentrations of drugs delivered to tumor cells, and the increased temperature makes the chemotherapeutic agents more effective. Mitomycin C is a commonly used chemotherapeutic drug for this purpose. Hyperthermic intraperitoneal chemoperfusion may be most useful when complete tumor cytoreduction is possible or in cases with positive cytology or gross perforation.75 Other treatment options that are under investigation include gene therapy76 and the use of angiogenesis inhibitors to reduce the ability of the peritoneal tumor to spread.77 Also, systemic “antidotes” to certain chemotherapy drugs (e.g., leucovorin or methotrexate) could be administered to reduce toxicity further.

Sixty-five percent to 70% of mesotheliomas arise in the pleura, and 25% in the peritoneum.86 Most peritoneal mesotheliomas are malignant, associated with asbestos exposure, and detected 35 to 40 years after initial exposure. The families of asbestos workers are also at risk. Diagnosis is usually made at laparotomy or laparoscopy, but occasionally diagnostic malignant mesothelial cells are found on ascitic fluid analysis. Serum osteopontin levels may help distinguish pleural mesothelioma from asbestosis without mesothelioma.87 Distinction of mesothelioma from peritoneal carcinomatosis of unknown primary may be difficult, even at autopsy. Classic treatments for localized peritoneal mesothelioma include cytoreductive surgery, hyperthermic intraoperative or postoperative intraperitoneal chemo­therapy, and immunotherapy.88 Mesothelioma is a nearly uniformly lethal neoplasm with a median survival of only six months.

Treatment of Ovarian Cancer

PELVIC LIPOMATOSIS

The results of treatment for ovarian cancer are the most encouraging. Seventy-five percent of patients with epithelial ovarian cancer present with advanced (stage III/IV) disease. Cytoreductive surgery and chemotherapy (cisplatinum and paclitaxel) have led to long-term survival. The degree of cytoreduction that can be achieved surgically correlates with improved five-year survival (52% for microscopic residual versus 29% for macroscopic <1 cm residual).78 Experimental approaches are emerging with bevucizumab, an inhibitor of vascular endothelial growth factor (VEGF) receptors, to control ascites formation and tumor growth.79

Prognosis

Prognosis is very poor in general for patients with peritoneal involvement with metastatic cancer.80 In one large study, only 70% of patients survived one month, 25% survived 3 months, 12% survived 6 months, and 4% survived longer than one year after diagnosis.80 Their course involves recurrent and progressive bowel obstruction, malnutrition, and wasting before death.

PSEUDOMYXOMA PERITONEI

Pseudomyxoma peritonei represents a rare (≈2 in 10,000 laparotomies) and special case in metastatic peritoneal tumors.81 Seventy-five percent of these patients are women between 45 and 75 years of age. This tumor causes gelatinous implants on the peritoneum. The sites of origin of the tumor are ovary and appendix. Its degree of malignant potential is variable; about 50% of patients live 5 years.81

MESOTHELIOMA

Fat deposits normally found in the perirectal and perivesical spaces may develop nonmalignant overgrowth and are recognized as a distinct clinicopathologic entity, pelvic lipomatosis. It occurs predominantly in African American men (male-to-female ratio 18 : 1) between 20 and 60 years of age89 and may cause hypertension; proliferative cystitis, urinary tract obstruction, and occasionally gastrointestinal symptoms. The abnormal proliferation of fat is accompanied by varying degrees of fibrous reaction. Transrectal ultrasonography and CT are important in diagnosis, particularly in differentiating pelvic lipomatosis from liposarcoma. The disease does not progress in most patients; however, in some, urinary obstruction requires diversion.

BENIGN PERITONEAL CYSTS

Benign peritoneal cysts are rare. Benign cystic mesotheliomas occur in adult women, are manifested by pain, and recur after resection. Benign cystic lymphangiomas affect young men, present as mass lesions, and seldom recur after resection.

DISEASES OF THE MESENTERY AND OMENTUM Diseases of the mesentery and omentum (in decreasing order of frequency) include hemorrhage, tumors, inflammatory and fibrotic conditions, and infarction. Abscesses are covered in Chapter 26.

619

620

Section IV  Topics Involving Multiple Organs HEMORRHAGE

Mesenteric and retroperitoneal bleeding and their complications are usually due to trauma or anticoagulants. In rare cases, aneurysms of the splanchnic arteries may rupture, leading to intraperitoneal hemorrhage. Traumatic hematomas may or may not require surgical intervention, depending on the site of the lesion and whether the trauma was blunt or penetrating.90 Intraperitoneal bleeding may be a consequence of a previous surgical procedure, such as cytoreductive surgery for gynecologic cancer.91 A special case of spontaneous hemoperitoneum is found in patients with cirrhosis and hepatocellular carcinoma. Symptoms usually are pain and those from mass effects of the hematoma such as symptoms of intestinal obstruction. Diagnosis depends on a high index of suspicion and ultrasonography or CT, which demonstrates the collection of blood. An ultrasound-guided fine-needle aspiration may help in confirming the diagnosis. Treatment consists of discontinuation of anticoagulants (in those being so treated) and reversal of anticoagulation. In others treatment is dictated by the local or systemic symptoms of hemorrhage. In certain cases angiographic embolization may help treat intraperitoneal hemorrhage.92

TUMORS

Tumors originating in the mesentery and omentum are rare and include soft tissue tumors (e.g., cysts, fibromas, sarcomas, desmoids) and tumors specific to this site, such as Castleman’s disease and leiomyomatosis peritonealis disseminata. Most tumors are large when detected in this site because of the large potential space in which they can grow. They may also be detected incidentally when an imaging study is performed for an unrelated reason. These typically present with nonspecific symptoms such as abdominal discomfort or low-grade obstructive symptoms.

Mesenteric Cysts

Mesenteric cysts are probably the most uncommon among these rare tumors.93 A review of the English-language literature revealed only 139 such lesions as of 1986.93 They occur in children and adults. Symptoms include pain in 58% and abdominal distention in 50%. Some cases may present with fever and chills, and others are asymptomatic, discovered incidentally and misdiagnosed before laparotomy.94 These are typically large (13 cm), fluid-filled (≈2000 mL) lesions and, despite their size, are malignant in only 3% of cases and cause death in only 2% of cases.93 They are usually cured by complete excision. If a small mesenteric cyst is found incidentally at laparotomy, it does not need to be resected. The treatment of choice for a complication (i.e., cyst rupture or hemorrhage) is excision, and this has been performed laparoscopically.95

Solid Tumors

Solid tumors appear to be next in decreasing order of frequency. Among mesenteric tumors, two thirds are benign, including fibromas, xanthogranulomas, lipomas, leiomyomas, capillary and cavernous hemangiomas, neurofibromas, and mesenchymomas. The malignant tumors include hemangiopericytomas, fibrosarcomas, liposarcomas, leiomyosarcomas, and malignant mesenchymomas. Solid tumors of the omentum are remarkably similar in histologic type and prevalence of malignancy.96 Typical of mesenteric and omental tumors, symptoms and signs include pain and distention with large lesions. Treatment is surgical resection. Prognosis is generally fair: about 18% of patients die of the tumor, overall, and the rate of five-year survival for patients with malignant tumors is only 21%.97

Needle biopsy may be attempted with these tumors, although laparoscopy or laparotomy may be required for diagnosis as well as treatment.

Multifocal Leiomyomas (Leiomyomatosis Peritonealis Disseminata)

Multifocal leiomyomatous tumors are even less common, can be malignant, and can mimic peritoneal carcinomatosis. They may appear together with other leiomyomatous lesions98 or endometriosis.99 These lesions consist of small, rubbery nodules and appear to be hormone sensitive, developing sometimes during pregnancy or estrogen therapy and regressing with hormone withdrawal. These tumors can cause abdominal pain or GI bleeding

Castleman’s Disease

Castleman’s disease, giant lymph node hyperplasia, is rare. There is considerable heterogeneity in the disease but it is classified in unicentric and multicentric forms.100 Castleman’s disease is associated with a variety of other autoimmune diseases. The central lymph nodes of the mesentery and mediastinum are more frequently involved in the unicentric form. In the unicentric form of the disease surgical removal of the mass is successful and prognosis is good. The multicentric form is treated with systemic therapies with variable success. The prognosis is considerably worse, with patients at risk for conversion to frank lymphoma.101

INFLAMMATORY AND FIBROTIC CONDITIONS

This subset of diseases of the mesentery and retroperitoneum is the most confusing, in part because of their rarity and because of overlapping clinical and histologic features. At least a dozen terms are used to describe the three basic diseases: retractile mesenteritis, mesenteric panniculitis, and retroperitoneal fibrosis. To add to the confusion, some cases have been reported with different names. These diseases could easily represent different aspects of the same spectrum of inflammation and scarring of these structures. Retractile mesenteritis was the name used in the first description of these diseases. This entity represents the fibrotic end of the spectrum and has been known as sclerosing mesenteritis, multifocal subperitoneal sclerosis, fibromatosis, and desmoid tumor.102 The inflammatory end of the spectrum has been called mesenteric panniculitis, mesenteric lipodystrophy, lipogranuloma of the mesentery, liposclerotic mesenteritis, mesenteric Weber-Christian disease, and systemic nodular panniculitis.103 There have been attempts to subclassify this disease into diffuse, single, and multiple forms and to suggest an association with lymphoma.104 Overlapping names such as sclerosing lipogranuloma, the well-documented progression and conversion of mesenteric panniculitis to retractile mesenteritis over a 12-year period, and the concurrence of sclerosing mesenteritis and retroperitoneal fibrosis indicate that these are simply stages of one basic underlying process. Although mesenteric panniculitis and retractile mesenteritis are usually manifested by abdominal pain, symptoms of gut obstruction, and a mass lesion,104 cases associated with prolonged high-grade fever and autoimmune hemolytic anemia without abdominal symptoms have been described.105 Retractile mesenteritis and mesenteric panniculitis are always idiopathic, but retroperitoneal fibrosis has a cause approximately 30% of the time, including drugs, malignancy, trauma, or inflammation.106 Most of the reported cases have been drug induced (methysergide, ergotamine). The process of fibrosis may lead to ureteral or vascular obstruction.

Chapter 37  Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm

Figure 37-3.  Computed tomography scan of the abdomen in a 63-year-old man with nausea, periumbilical pain, and a 10-pound weight loss. The scan on the left shows a soft tissue mass in the retroperitoneum encasing the aorta (arrow). Open biopsy of the mass showed inflammation and fibrosis with no evidence of tumor, which is compatible with retroperitoneal fibrosis. Symptoms resolved, and the mass regressed dramatically on glucocorticoid therapy. The scan on the right was taken nine months after diagnosis and therapy. (Courtesy Jeffrey H. Phillips, MD, Dallas, Texas.)

Histologically, retractile mesenteritis and mesenteric panniculitis can have inflammation with lymphocytes and neutrophils, fat necrosis, fibrosis, and calcification.102 In contrast, only mesenteric panniculitis has multinucleate giant cells, cholesterol clefts, lipid-laden macrophages, and lymphangiectasia.102 Retroperitoneal fibrosis consists of dense connective tissue, with or without inflammation.

Diagnosis and Treatment

These diseases have usually been diagnosed at laparotomy or autopsy in the past; however, noninvasive techniques such as CT scan (Fig. 37-3) or MRI may assist in preoperative diagnosis.107,108 Radiologic findings suggestive of mesenteric panniculitis have been found in 0.6% of patients in a large series of abdominal CT scans. There was a female predominance and an association with malignancy in 34 of 49 patients with radiologic features of mesenteric panniculitis.108 Retroperitoneal fibrosis is more common in men and typically causes the ureters to deviate medially on radiographic evaluation. Treatment may be necessary in patients with retractile mesenteritis if it obstructs the intestine. Treatment is usually surgical, but administration of progesterone has been reported to down-regulate fibrogenesis.109 The prognosis of patients with retroperitoneal fibrosis seems to be better than in the past. Successful treatment of this entity with immunosuppressives, such as azathioprine with steroids, has been reported.110 In other cases, ureterolysis may be required.

INFARCTION OF THE OMENTUM

Infarction of the omentum occurs when a portion of the omentum twists around a narrow vascular pedicle.111 If a diagnosis by imaging techniques (such as CT scan or MRI) is achieved preoperatively,112 laparoscopic resection of the necrotic mass is curative.113 However, the diagnosis is difficult and often delayed.

EPIPLOIC APPENDAGITIS

Epiploic appendagitis (primary inflammation of the colonic epiploic appendices) is an entity that is occasionally seen and is often confused with the diagnosis of appendicitis.

The diagnosis of epiploic appendagitis requires a high index of suspicion. It typically presents with right lower quadrant abdominal pain. However, constitutional symptoms such as nausea, vomiting, and anorexia are less frequent and the pain tends to have a more sudden onset. The patient can typically locate the exact location of the pain with one finger, and the point of tenderness and pain tends to be more localized and slightly more cephalad than in appendicitis. Epiploic appendagitis can be diagnosed by CT scan, and the treatment is nonoperative if the diagnosis can be made. Often the diagnosis is made intraoperatively while operating for presumed appendicitis, in which case appendectomy should be performed, and the epiploic appendage may be removed or left intact.

DISEASES OF THE DIAPHRAGM HERNIAS AND EVENTRATION

Diaphragmatic hernias consist of herniation of an abdominal organ through the diaphragm into the thorax and are discussed in detail in Chapter 24. Eventration is not a true hernia but consists of a localized weakness in the dome of the diaphragm that can lead to bulging of abdominal viscera into the thorax. This is usually an incidental finding on chest films, but large eventrations can cause shortness of breath by loss of lung volume on the affected side and mediastinal shift to the unaffected side. Symptomatic patients can be surgically corrected with thoracoscopic plication of the diaphragm.

TUMORS

Diaphragmatic tumors are usually of connective tissue origin and may be benign or malignant or may consist of simple cysts.114,115 They are detected by screening chest films or in evaluation of pleuritic chest pain.

HICCUPS (see Chapter 12)

Hiccups are quick inhalations that follow abrupt rhythmic involuntary contractions of the diaphragm and closure of

621

622

Section IV  Topics Involving Multiple Organs the glottis. When they last only a few minutes, they are considered a form of physiologic myoclonus.116 For hiccups of longer duration, home remedies include breath holding, sudden fright, rebreathing from a paper bag, eating dry granulated sugar, and drinking cold liquids. Intractable hiccups can last weeks or even longer, can be familial, and are usually due to diaphragmatic irritation, gastric distention, thoracic or central nervous system irritation or tumors, hyponatremia, or other metabolic derangements. Treatment includes pharmacologic agents, noninvasive phrenic nerve stimulation, or rarely phrenic nerve crushing. Drugs that have been reported to be successful include chlorpromazine, metoclopramide, quinidine, phenytoin, valproic acid, baclofen, sertraline, gabapentin, and nifed­ ipine. The implantation of breathing pacemakers that control the diaphragmatic excursions may be an interesting approach for treatment of chronic hiccups.116 Postoperative hiccups after abdominal surgery may be due to subphrenic abscess or other sources of diaphragmatic irritation such as acute gastric dilatation, and this should be considered before assuming a more benign cause.

LAPAROSCOPY IN THE EVALUATION OF PERITONEAL DISEASES GENERAL CONSIDERATIONS

Diagnostic laparoscopy, as first described by Kelling in 1901, is a safe and effective means of evaluating the abdominal cavity. It allows direct visualization of the liver surface, peritoneal lining, and mesentery for directed biopsies. (See Figs. 37-2 and 37-4 for illustrations of peritoneal carcinomatosis and lymphoma.) Ascitic fluid can be collected easily. Although less invasive imaging techniques such as CT have reduced its necessity, laparoscopy continues to have a role in the evaluation of liver and peritoneal diseases. Although generally well tolerated, possible complications include prolonged abdominal pain, vasovagal reaction, viscus perforation, bleeding (either from biopsy sites or

within abdominal wall), splenic laceration, ascites fluid leakage, and postlaparoscopy fever.117 It has been suggested that abdominal insufflation during laparoscopy could increase bacterial translocation, making the practice of laparoscopy dangerous in certain clinical settings, such as septic peritonitis.118 These observations, however, are not uniformly accepted.119 The adverse hemodynamic consequences of abdominal insufflation can be overcome in the vast majority of patients with aggressive resuscitation and careful anesthetic management. Despite these concerns, laparoscopy is becoming a common technique used in patients requiring operation for diseases causing peritonitis. A laparoscopic approach has been effective in treating perforated gastroduodenal ulcer120 and has been advocated as the treatment of choice for patients with appendicitis.121 Laparoscopic cholecystectomy is safe and effective treatment of acute cholecystitis,122 and laparoscopic colectomy can be performed for acute diverticulitis.123 Evidence-based guidelines for the application of laparoscopic operation in surgical peritonitis have been developed.30

EVALUATION OF ASCITES OF UNKNOWN ORIGIN (see Chapter 91)

Clinical presentation, conventional laboratory examinations, and ascitic fluid analysis identify the cause of ascites in the majority of patients; however, conventional para­ centesis occasionally fails to make a diagnosis. In these instances diagnostic laparoscopy affords direct and sensitive technique for obtaining specimens for histology and culture. In the United States, occult cirrhosis and peritoneal malignancy account for the majority of cases.117 In studies from Asian countries, peritoneal malignancy is also the most common cause of unexplained ascites, but tuberculous peritonitis accounts for an increasing number of cases.124 In patients with HIV, peritoneal involvement may result from a variety of opportunistic infections and neoplasms (see earlier section and Chapter 33). Non-Hodgkin’s lymphoma (Fig. 37-4) accounts for the majority of these peritoneal lesions revealed by laparoscopy, but M. tuberculosis, M. avium-intracellulare, and P. jiroveci are often revealed.125

STAGING LAPAROSCOPY

Figure 37-4.  Laparoscopic appearance of non-Hodgkin’s lymphoma in a patient infected with the human immunodeficiency virus. The raised, white, irregular plaques are present over the visceral and parietal peritoneal surfaces. (From Jeffers LJ, Alzate I, Aguilar H, et al. Laparoscopic and histologic findings in patients with the immunodeficiency virus. Gastrointest Endosc 1994; 40:160.)

Laparoscopy has found increasing utility in the staging of malignant solid tumors of the gastrointestinal tract. Diagnostic laparoscopy coupled with laparoscopic ultrasound, peritoneal fluid cytology, and biopsy allow for improved selection of patients that will benefit from larger, definitive operations for curative intent. In hepatocellular carcinoma, the use of diagnostic laparoscopy and laparoscopic ultrasonography demonstrates that 25% to 33% of patients with potentially resectable disease are found to be unresectable and can be spared unnecessary laparotomy.126 Staging laparoscopy and laparoscopic ultrasonography are commonly used to select patients under consideration for hepatic resection of metastatic colorectal cancer. In an extensive laparoscopic staging procedure for pancreatic cancer, Conlon and associates identified 36% of patients with metastatic disease after an initial negative CT.127 The overall accuracy of staging laparoscopy in this series of patients for selection of patients for pancreatic resection was 98%. In a large review of 420 patients with upper gastrointestinal malignancy, laparoscopic staging prevented unnecessary laparotomy in 5% of esophageal cancers and 20% of patients with a tumor of the gastroesophageal junction.128 The finding of metastatic disease on staging laparoscopy in esophageal and gastric cancers will not necessarily obviate the need for palliative operations.

Chapter 37  Surgical Peritonitis and Other Diseases of the Peritoneum, Mesentery, Omentum, and Diaphragm KEY REFERENCES

Bridda A, Mencarelli R, Frego M. Peritoneal mesothelioma: A review. Med Gen Med 2007; 9:32. (Ref 88.) D’Angelica M, Spiros P, Hiotis HJK, et al. Laparoscopic staging for liver, biliary, pancreas, and gastric cancer. Curr Prob Surg 2007; 44:228-69. (Ref 126.) Dunn DL, Barke RA, Knight NB, et al. Role of resident macrophages, peripheral neutrophils, and translymphatic absorption in bacterial clearance from the peritoneal cavity. Infect Immun 1985; 49:257-64. (Ref 13.) Gutt CN, Oniu T, Schemmer P, et al. Fewer adhesions in laparoscopic surgery? Surg Endosc 2004; 18:898-906. (Ref 70.) Higgins PM, Aber GM. Idiopathic retroperitoneal fibrosis: An update. Dig Dis 1990; 8:206-22. (Ref 100.) Leschka S, Alkadhi H, Wildermuth S, Marincek B. Multi-detector computed tomography of acute abdomen. Eur Radiol 2005; 15:2435-47. (Ref 27.) Lorber B, Swenson RM. The bacteriology of intra-abdominal infections. Surg Clin North Am 1975; 55:1349-54. (Ref 5.) Meyers MA, Oliphant M, Berne AS, Feldberg MAM. The peritoneal ligaments and mesenteries: Pathways of intra-abdominal spread of disease. Radiology 1987; 163:593-604. (Ref 1.)

Novitsky YW, Litwin DEM, Callery MP. The net immunologic advantage of laparoscopic surgery. Surg Endosc 2004; 18:1411-19. (Ref 23.) Ordonez CA, Puyana JC. Management of peritonitis in critically ill patients. Surg Clin North Am 2006; 86:1323-49. (Ref 38.) Runyon BA, Morrissey R, Hoefs JC, Wyle F: Opsonic activity of human ascitic fluid: A potentially important protective mechanism against spontaneous bacterial peritonitis. Hepatology 1985; 5:634-7. (Ref 12.) Saltzman DJ, Williams RA, Gelfand DV, Wilson SE: The surgeon and AIDS: Twenty years later. Arch Surg 2005; 140:961-7. (Ref 58.) Sauerland S, Agresta F, Bergamaschi R, et al. Laparoscopy for abdominal emergencies: Evidence based guidelines of the European Association for Endoscopic Surgery. Surg Endosc 2006; 20:14-29. (Ref 30.) Wong PF, Gilliam AD, Kumar S, et al. Antibiotic regimens for secondary peritonitis of gastrointestinal origin in adults. Cochrane Database Syst Rev Issue 2, 2005. (Ref 37.) Yan TD, Black D, Savady R, Sugarbaker PH. A systematic review on the efficacy of cytoreductive surgery and perioperative intraperitoneal chemotherapy for pseudomyxoma peritonei. Ann Surg Oncol 2006; 14(2):484-92. (Ref 84.) Full references for this chapter can be found on www.expertconsult.com.

623

CHAPTE R

38 Gastrointestinal and Hepatic Disorders in the Pregnant Patient Silvia Degli Esposti and John F. Reinus

CHAPTER OUTLINE Gastrointestinal and Hepatic Function in Normal Pregnancy  625 Gastrointestinal Function  625 Hepatic Function and Liver Biochemical Tests  626 Drug Safety in Pregnant Patients  626 Endoscopy During Pregnancy  626 Imaging and Radiation Exposure During Pregnancy  627 Gastrointestinal Disorders and Pregnancy  628 Nausea, Vomiting, and Hyperemesis Gravidarum  628 Gastroesophageal Reflux Disease  628 Gastric and Duodenal Ulcer Disease  629 Inflammatory Bowel Disease  629 Appendicitis  630 Biliary and Pancreatic Disorders and Pregnancy  630 Gallstone Disease  630 Pancreatitis  630

The unique physiologic milieu of pregnancy affects the function of many organ systems, not least among them the digestive tract. Pregnancy may promote symptoms of gastrointestinal dysfunction, alter the natural history of new or preexisting digestive diseases, and result in disorders that occur only during gestation. Conversely, gastrointestinal and hepatic diseases can adversely affect the clinical course of pregnancy and its outcome.

GASTROINTESTINAL AND HEPATIC FUNCTION IN NORMAL PREGANCY GASTROINTESTINAL FUNCTION

The gastrointestinal tract undergoes dramatic modifications during pregnancy. Intra-abdominal organs must move to accommodate uterine growth, hormonal and humoral factors alter motility, and the immunologic adaptation to pregnancy affects response to disease. Heartburn, nausea, abdominal cramps, and altered bowel habits, the most common gastrointestinal symptoms of pregnant women, are caused by normal physiologic changes in gut motility. These symptoms usually are transitory and easily treated with conservative measures. However, it may be a challenge to distinguish among symptoms of altered motility and

Hepatic Disorders Unique to Pregnancy  631 Cholestasis of Pregnancy  631 Liver Disease of Preeclampsia  632 Hemolysis, Elevated Liver Enzymes, and Low Platelet Count (HELLP) Syndrome  632 Hepatic Rupture, Hematoma, and Infarct  634 Acute Fatty Liver of Pregnancy  634 Usual Hepatic Disorders and Pregnancy  636 Viral Hepatitis  636 Chronic Liver Disease and Portal Hypertension  637 Wilson Disease  637 Autoimmune Liver Diseases  637 Hepatic Neoplasia and Mass Lesions  637 Hepatic Vein Thrombosis (Budd-Chiari Syndrome)  637 Pregnancy after Liver Transplantation  638

those that signal the onset or worsening of problems that require immediate medical attention. The amplitude and duration of esophageal muscle contractions in pregnant and nonpregnant women are similar.1 In the distal esophagus, the velocity of peristaltic waves has been found to decrease by approximately one third during pregnancy, but remains within the normal range.2 In contrast, resting lower esophageal sphincter tone pro­ gressively declines during gestation, most likely as con­ sequence of inhibition of smooth muscle contraction by progesterone.2-4 This effect coupled with increased abdominal pressure during gestation is responsible for the gastroesophageal reflux symptoms that occur in 70% of pregnant women.5 The effects of pregnancy on gastric motility are unclear. Delayed gastric empting has been demonstrated by some authors, especially during delivery,6 whereas no effect on gastric emptying has been found by others.7 Pregnant women appear to have normal basal and stimulated gastric secretion.8 Transit time of intestinal contents is prolonged during gestation; delayed small-bowel transit is most pronounced during the third trimester and is associated with slowing of the migratory motor complex.9,10 Colonic transit time is prolonged in pregnant animals. Progesterone is thought to have a direct inhibitory effect on smooth muscle cells that slows motility.11 A role for endogenous opioids also has been sug-

625

626

Section IV  Topics Involving Multiple Organs gested.12 Together, these changes often result in mild physiologic constipation. The absorptive capacity of the small intestine increases during pregnancy to meet the metabolic demands of the fetus; increased absorption of calcium, amino acids, and vitamins has been demonstrated.13-16 Animal experiments have revealed pregnancy-induced increases in smallintestinal weight and villous height in conjunction with mucosal hypertrophy.17,18 The activity of some brush border enzymes increases during lactation and then decreases after weaning.19,20 During pregnancy the maternal immune system must adapt to the presence of the fetus. Adaptive changes can influence the response to infection and modulate the course of underlying autoimmune disease. Although many pregnancy-induced immunologic changes remain obscure, it has been proven that during pregnancy there is a shift from cellular to humoral responses with down-regulation of Th1 and up-regulation of Th2 cytokines. Pregnancy modulates natural killer (NK) cell cytotoxicity and induces T-regulatory cells that affect the maternal immune response.21,22 Unfortunately, we still do not understand the affects of pregnancy on the mechanisms responsible for autoimmune diseases such as Crohn’s disease and autoimmune hepatitis well enough to allow us to predict clinical events. Pregnancy causes an alteration in bile composition, including cholesterol supersaturation, decreased chenodeoxycholic acid and increased cholic acid concentrations, and an increase in the size of the bile acid pool.23 These changes are associated with greater residual gallbladder volumes in the fasting as well as fed states. Sex-steroid hormones may inhibit gallbladder contraction in pregnant women, promoting precipitation of cholesterol crystals and stone formation.24,25

HEPATIC FUNCTION AND LIVER BIOCHEMICAL TESTS

During pregnancy, maternal blood volume increases progressively until, by the thirtieth week of gestation, it is 50% greater than normal, remaining so until confinement.26 This volume expansion, attributed to the effects of steroid hormones and elevated plasma levels of aldosterone and renin, is responsible for dilution of some blood constituents, such as red blood cells (physiologic anemia). Total serum protein concentration diminishes 20% by mid-pregnancy, largely as a result of a reduced serum albumin level. A reciprocal relationship between falling serum albumin and rising serum alpha-fetoprotein concentrations in pregnant women has been proposed.27 Despite increases in maternal blood volume, the levels of many serum proteins measured to assess hepatic injury are unchanged or even increased during gestation. Progesterone causes a proliferation of smooth endoplasmic reticulum, whereas estrogens promote formation of rough endoplasmic reticulum and associated protein synthesis. Pregnant women synthesize at an accelerated rate the products of the cytochrome P-450 gene superfamily and other proteins, including coagulation factors, binding globulins, and ceruloplasmin. Maternal serum alkaline phosphatase levels normally are elevated during the third trimester of pregnancy, largely due to placental production; for this reason, measurement of alkaline phosphatase in pregnant women is only of clinical use early in gestation. Alterations in maternal concentrations of plasma proteins may persist for several months postpartum.

DRUG SAFETY IN PREGNANT PATIENTS Patients and physicians tend to withhold pharmacologic treatment with medications during pregnancy because they fear harming the fetus. Nevertheless, avoidance of medical interventions may adversely affect the mother’s health and the pregnancy outcome. Having stated this, no medication or other therapeutic intervention can be considered definitely safe during pregnancy. Indeed, the placenta is not a reliable barrier to the passage of most drugs, the distribution of a drug within the fetal compartment cannot be accurately predicted, and data on long-term effects of in utero fetal drug exposure are practically impossible to collect. The necessity of any proposed therapy should be discussed with the patient, and known and unknown risks of treatments must be carefully evaluated. The U.S. Food and Drug Administration (FDA) categorizes drugs based on their potential fetal toxicity during pregnancy (Tables 38-1 and 38-2). The FDA classification, however, is of limited practical value because it is based on very few data. A recent publication of the American College of Physicians28 makes recommendations concerning drug therapy in pregnant women.

ENDOSCOPY DURING PREGNANCY It is estimated that 20,000 pregnant women undergo endoscopy each year.29 Recommendations concerning endoscopy in this setting are largely based on expert opinion and

Table 38-1  Food and Drug Administration Categories of Fetal Risk from Medicines* CATEGORY

CRITERIA

A

Adequate well-controlled studies in pregnant women have not shown an increased risk of fetal abnormalities. Animal studies have revealed no evidence of harm to the fetus; however, there are no adequate and well-controlled studies in pregnant women. Or Animal studies have shown an adverse effect, but adequate and well-controlled studies in pregnant women have failed to demonstrate a risk to the fetus. Animal studies have shown an adverse effect and there are no adequate and well-controlled studies in pregnant women. Or No animal studies have been conducted and there are no adequate and well-controlled studies in pregnant women. Adequate well-controlled or observational studies in pregnant women have demonstrated a risk to the fetus. The benefits of therapy, however, may outweigh its potential risk. Adequate well-controlled or observational studies in animals or pregnant women have demonstrated evidence of fetal abnormalities. The use of the product is contraindicated in women who are or may become pregnant.

B

C

D

X

*U.S. Food and Drug Administration. FDA Consumer Magazine. May-Jun 2001; 35:3.

Chapter 38  Gastrointestinal and Hepatic Disorders in the Pregnant Patient Table 38-2  Food and Drug Administration Categories of Fetal Risk for Some Medications Used to Treat Gastrointestinal and Hepatic Diseases MEDICATION

INDICATION

FDA CATEGORY

Adalimumab Adefovir Amoxicillin Azathioprine Balsalazide Benzodiazepines Bismuth Budesonide Certolizumab pegol Cimetidine Clarithromycin Cyclosporine Entecavir Esomeprazole Famotidine Fentanyl Infliximab Interferon alpha Lamivudine Lansoprazole Meperidine Mesalamine Methotrexate Metoclopramide Metronidazole Nizatidine Olsalazine Omeprazole Ondansetron Pantoprazole Penicillin Prednisone Prednisolone Propofol Rabeprazole Ranitidine Ribavirin Sucralfate Sulfasalazine Telbivudine Tenofovir Tetracycline Thalidomide

IBD Hepatitis B Infection IBD; autoimmune hepatitis IBD Sedation Helicobacter pylori infection IBD IBD GERD; PUD Infection Transplantation Hepatitis B GERD; PUD GERD; PUD Sedation, analgesia IBD Hepatitis B, C Hepatitis B GERD; PUD Sedation, analgesia IBD IBD Nausea; GERD Infection GERD; PUD IBD GERD; PUD Nausea, vomiting GERD; PUD Infection IBD; autoimmune hepatitis IBD; autoimmune hepatitis Sedation GERD; PUD GERD; PUD Hepatitis C PUD IBD Hepatitis B Hepatitis B Antibiotic Sedative; myeloma

B C B D C D C B B B C C C B B C B C C B B B X B B* B C C B B B B C B B B X B B B B D X

*After first trimester only. B, no evidence of risk in humans; C, possible risk; D, evidence of fetal risk; X, definite fetal risk; GERD, gastroesophageal reflux disease; IBD, inflammatory bowel disease; PUD, peptic ulcer disease.

case reports. Although the safety of endoscopy during pregnancy has not been completely established, it is performed routinely if there is a clear indication.30 Pregnant women have undergone upper gastrointestinal tract endoscopy, colonoscopy, sigmoidoscopy, endoscopic retrograde colangiopancreatograpy (ERCP), and percutaneous gastroscopy safely.31 In addition to general contraindications to endoscopic procedures, specific contraindications during pregnancy include imminent or threatened delivery, ruptured membranes, placental abruption, and pregnancy-induced hypertension.32 Several precautions should be observed to avoid complications when endoscoping a pregnant patient.32 Given the extreme sensitivity of the fetus to maternal hypoxia, preg-

Table 38-3  Fetal Effects of Radiation During Gestation* GESTATIONAL AGE (DAYS)

EFFECTS OF RADIATION

0-9 13-50 51-280

Death Teratogenesis Growth restriction Growth restriction CNS abnormalities Possible cancer risk

*Data from reference 38. CNS, central nervous system.

nant women should receive supplemental oxygen with continuous blood-saturation monitoring. When the fetus is capable of surviving outside the uterus, usually around 24 weeks of gestation, external fetal heart monitoring before, during, and after invasive procedures is advisable to enable prompt delivery if fetal distress occurs. In the second and third trimesters, the supine position and external abdominal pressure should be avoided because resulting compression of the vena cava and aorta may cause hypotension and placental hypoperfusion. ERCP should be performed only with therapeutic intent and by expert endoscopists; every effort should be made to avoid fetal radiation.31 Sedation with meperidine (pregnancy category B), which crosses the blood-brain barrier more slowly than fentanyl (pre­ gnancy category C) and morphine (pregnancy category C), is preferred, although fentanyl may be superior during lactation because it is poorly excreted in breast milk. Sedation with benzodiazepines (pregnancy category D) should be avoided, especially during the first trimester, because diazepam has been reported to cause fetal malformations.33,34 Extensive experience with propofol (pregnancy category B) is lacking, and its high lipid solubility is a reason for concern.35 Patients are advised to avoid breast-feeding and to discard breast milk for 24 hours after a procedure requiring sedation.32

IMAGING AND RADIATION EXPOSURE DURING PREGNANCY The National Commission on Radiation Protection recommends limiting exposure to ionizing radiation during pregnancy to less than 5 cGy.36,37 The potential for radiation damage to the fetus is determined by dose and gestational age at the time of exposure (Table 38-3). Computed tomography (CT) should be performed only when its potential benefits clearly outweigh its risks and should be done, if possible, after completion of organogenesis.38 Helical CT may be associated with less fetal radiation exposure than plain CT. Magnetic resonance imaging (MRI) often is a superior alternative to CT; MRI without contrast has not been associated with adverse pregnancy outcomes and magnetic fields are not con­ sidered harmful to living organisms.39 There is a theore­ tical risk of thermal injury to the fetus from MRI in early pregnancy, and MRI is not recommended during the first 12 weeks of gestation. Gadolinium crosses the placenta and its safety in pregnant women has not been formally assessed. Ultrasonography is widely used and safe during pregnancy.

627

628

Section IV  Topics Involving Multiple Organs GASTROINTESTINAL DISORDERS AND PREGNANCY NAUSEA, VOMITING, AND HYPEREMESIS GRAVIDARUM (see Chapter 14)

Sixty percent to 70% of pregnant women report having some nausea in their first trimester, and more than 40% report vomiting.40 Onset of these symptoms typically is in the fourth to sixth week of gestation, with a peak occurrence in the eighth to twelfth week and resolution by week 20. Although nausea and vomiting may vary from mild to severe, most affected individuals still are able to obtain adequate oral nutrition and hydration, in some cases by eating frequent small meals of dry starchy foods. Severe persistent vomiting demanding medical intervention, or hyperemesis gravidarum, is less common, occurring in 2% or less of all pregnancies.41,42 Hyperemesis is associated with fluid, electrolyte, and acid-base imbalances; nutritional deficiency; and weight loss, and is defined by the presence of ketonuria and a 5% decrease from prepregnancy weight. Although the prognosis of hyperemesis gravidarum is generally favorable, severe untreated disease may lead to significant maternal and fetal morbidity. Symptoms usually begin at weeks 4 to 5 and improve by weeks 14 to 16 of gestation. In up to 20% of affected patients, however, vomiting persists until delivery.43 Reported risk factors for hyperemesis include a personal or family history of the disorder, a female fetus or multiple gestation, gestational trophoblastic disease, fetal trisomy 21, hydrops fetalis, and maternal Helicobacter pylori infection.44 The etiology of hyperemesis gravidarum is likely multifactorial, including contributions by hormonal changes, gastrointestinal dysmotility, H. pylori infection, and psychosocial factors. Pregnancy-related hormones, specifically human chorionic gonadotropin (HCG) and estrogen, have been implicated as important causes of hyperemesis.45 Symptoms worsen during periods of peak HCG concentrations, and conditions associated with higher serum HCG levels, for example, multiple gestation, trophoblastic disease, and trisomy 21, are associated with an increased incidence of hyperemesis.46 Elevated serum estrogen concentrations, as seen in obese patients, also have been asso­ ciated with this disorder.47 Estrogen and progesterone are thought to cause nausea and vomiting by altering gastric motility and slowing gastrointestinal transit time.48 Other hormones implicated in the pathogenesis of hyperemesis include the thyroid hormones and leptin.49 Abnormal thyroid function test results are found in two thirds of patients with hyperemesis gravidarum.50 The alpha-subunit of HCG has thyroid-stimulating hormone (TSH)–like activity that suppresses TSH release and causes a slight rise in free thyroxine (T4) levels51 but, despite these findings, this gestational transient thyrotoxicosis is not associated with unfavorable pregnancy outcomes and usually does not require treatment. The role of H. pylori as an etiology of hyperemesis is controversial. Several studies have found H. pylori infection to be significantly associated with the disorder,44,52 whereas others could not establish any relationship between the two conditions.53 A small study of the effect of H. pylori eradication in pregnant patients with vomiting showed symptomatic improvement.54 Vomiting in patients with hyperemesis gravidarum often is triggered by olfactory, and even auditory and visual stimuli. A pregnancy-unique quantification of nausea and emesis (PUQE score) can be used to evaluate the number of hours of nausea and the number of episodes of emesis and retching per day in affected women, and is helpful in tailor-

ing therapy.55 Hospital admission for intravenous fluid and electrolyte replacement and, sometimes, nutritional support is indicated when affected individuals develop hypotension, tachycardia, ketosis, weight loss, or muscle wasting. Abnormal laboratory test results in such patients include hypokalemia, hyponatremia, and ketonuria. Hyperemesis is associated with slight increases in serum aminotrans­ ferase and bilirubin levels in 25% to 40% of cases. Hyperamylasemia is seen in a quarter of affected patients due to excessive salivary gland production stimulated by prolonged vomiting.56 Severe hyperemesis gravidarum is associated with poor maternal and fetal outcomes. In a study of more than 150,000 singleton pregnancies, infants born to women with hyperemesis who had gained less than 7 kg of weight during pregnancy were more likely to have low birth weights, be premature and small for gestational age, and to have low Apgar scores.41 Severe, albeit rare, maternal complications of hyperemesis include Mallory-Weiss tears with upper gastrointestinal bleeding, Boerhaave’s syndrome, Wernicke’s encephalopathy with or without Korsakoff’s psychosis, central pontine myelinolysis, retinal hemorrhage, and spontaneous pneumomediastinum.57 Lastly, severe depression after elective termination of pregnancy has been reported.58 Given the potential for morbidity and mortality in hyperemesis gravidarum, affected individuals should be treated aggressively. Obstetric management should be overseen, if possible, by physicians qualified in maternal-fetal medicine. The goals of therapy are maintenance of adequate maternal fluid intake and nutrition as well as symptom control. Patients should be advised to eat multiple small meals as tolerated and to avoid an empty stomach, which may trigger nausea. Also avoidance of offensive odors, separation of solid and liquid foods, and consumption of a high-carbohydrate diet may be helpful.59 Antiemetics and antireflux medications are first-line pharmacologic therapy for outpatients who have failed dietary modifications (see Table 38-2). Phenothiazines (chlorpromazine, prochlorperazine), the dopamine antagonist metoclopramide, and pyridoxine (vitamin B6) have proven bene­ ficial in this setting.60 Extensive data show lack of teratogenesis and good fetal safety for many of these drugs.61,62 Treatment with ondansetron (pregnancy category B), a 5-hydroxytryptamine-3 (5-HT3) receptor antagonist, should be considered in patients who do not respond to the above measures. The safety of ondansetron therapy during pregnancy is supported by a recent controlled trial,63 case reports, and widespread clinical experience. Failure of oral medical therapy can be managed in the home setting with intravenous fluid replacement, medications, and multivi­ tamins. As many as 50% of pregnant patients treated through central intravenous catheters, including those inserted peripherally, have catheter-related complications,64 most likely as result of the relative hypercoagulable state and increased susceptibility to infections seen in pregnant women. Enteral feeding through a nasoenteric tube or even total parenteral nutrition is sometimes required to maintain maternal nutrition.65

GASTROESOPHAGEAL REFLUX DISEASE

(see Chapter 43) At least as many women experience pyrosis as nausea during pregnancy; by the end of the third trimester, 50% to 80% of pregnant patients have had new, or an exacerbation of preexisting, heartburn.66,67 Pyrosis, however, rarely is accompanied by overt esophagitis or its complications.68

Chapter 38  Gastrointestinal and Hepatic Disorders in the Pregnant Patient Pregnant women with heartburn also may have regurgitation and, as already mentioned, nausea and vomiting, as well as atypical symptoms, such as persistent cough and wheezing. Symptoms usually develop toward the end of the second trimester, persist until delivery, and may be predictive of recurrent gastroesophageal reflux disease (GERD) later in life.66 Risk factors for reflux include multiparity, older maternal age, excessive weight gain, and reflux complicating a prior pregnancy.5,66,69 The pathogenesis of GERD in pregnant women is related to the effects of gestational hormones on esophageal motility, lower esophageal sphincter muscle tone, and gastric emptying. Compression of the stomach and increased intraabdominal pressure caused by the enlarging uterus also contribute to development of this disorder. Esophagogastroduodenoscopy (EGD) is rarely required for the assessment of pregnant women with symptoms of GERD.70 There are no data assessing the use of 24-hour ambulatory pH monitoring in this setting, and use of a barium esophagram is undesirable because it entails fetal x-ray exposure. Thus evaluation of suspected GERD in a pregnant woman depends on the clinical experience and judgment of the physician and requires due consideration of the patient’s history and all potential, reasonable causes for the patient’s present symptoms. Mild reflux symptoms often can be controlled by modifications of diet and lifestyle. Liquid antacids and sucralfate (FDA category B) often are prescribed as first-line pharmacologic therapy.71 Alginic acid (FDA category B) also is effective.72 Magnesium-containing antacids should be avoided during the late third trimester because they theoretically may impair labor. Ranitidine (FDA category B) remains the treatment of choice for patients who have persistent heartburn despite liquid antacid therapy.73 Proton pump inhibitors should be reserved for the most refractory cases, given their more recent introduction to the market. A recent meta-analysis found no significant risk of fetal malformations in babies exposed to proton pump inhibitors in utero.74 Omeprazole is a pregnancy class C drug because it has caused fetal toxicity in animals; all other available proton pump inhibitors are pregnancy category B drugs. The promotility agent, metoclopramide, has not been used extensively to treat GERD during gestation, although it is used during obstetric anesthesia.

GASTRIC AND DUODENAL ULCER DISEASE

(see Chapters 52 and 53) Case studies and retrospective series suggest that the incidence of peptic ulcer disease (PUD) is lower in pregnant women than in nonpregnant individuals.75,76 This impression, if it is valid, may be related to decreased use of nonsteroidal anti-inflammatory drugs (NSAIDs) by cautious patients or possibly to increased use of antacid medications to treat nausea or heartburn. It is conceivable, but equally unproven, that gestational steroids promote gastrointestinal mucosal cytoprotection. PUD is likely underdiagnosed during pregnancy, given the reluctance of physicians to perform diagnostic tests on pregnant women. Gastric acid secretion and the natural history of H. pylori infection, as far as we know, are not altered by gestation. The dyspeptic symptoms that often accompany pregnancy, especially nausea, vomiting, and heartburn, may make diagnosis of PUD in this setting difficult. Because PUD is exceedingly common in the population as a whole, physicians who care for pregnant women should be vigilant for its occurrence in their patients. A trial of empirical acid suppression may be useful in women with suspected PUD, both as a diagnostic and a therapeutic maneuver, and is

thought to be safe.77-80 First-line therapies include ranitidine and sucralfate. In confusing cases, diagnostic EGD is indicated (see earlier). Patients with H. pylori infection may be given antibiotics during pregnancy or after delivery.

INFLAMMATORY BOWEL DISEASE

(see Chapters 111 and 112) Physicians who treat patients with inflammatory bowel disease (IBD) are likely to encounter the disorder in pregnant women. The majority of cases of IBD first present in women younger than age 30 years, the years of peak fertility.81,82 Ulcerative colitis and Crohn’s disease may be more common in women than in men; some authors report women to have an approximately 30% greater risk than men of developing IBD.82 There is controversy regarding the effects of IBD on fertility. Pregnancy rates in IBD patients may be spuriously low because of self-image problems that result in sexual avoidance and voluntary childlessness.83 Fear of IBD in offspring and fear of fetal malformation secondary to maternal therapy are often cited as major causes of childlessness by affected women.84 Female fertility itself, however, does not appear to be impaired by uncomplicated IBD.85,86 A notable exception is fertility in ulcerative colitis patients treated with total colectomy and ileoanal J-pouch anastomosis.87,88 A recent meta-analysis found a three-fold increase in the risk of infertility in IBD patients who had undergone this procedure.88 Infertility in these individuals most likely is caused by pelvic adhesions and fallopian tube scarring. Potential infertility should be discussed with patients of childbearing age who are considering this operation. Male fertility is impaired by sulfasalazine treatment, which causes decreased sperm counts that usually return to normal within six months of discontinuing the drug.89 An initial presentation of IBD during pregnancy is unusual; when IBD first develops in a pregnant woman, it most often does so during the first trimester.90,91 Cases of this type are no more severe than those in nonpregnant individuals. Likewise, pregnancy does not appear to increase the severity of, or morbidity due to, preexisting IBD; disease activity prior to conception seems to be the most important factor determining the cause of the illness during gestation.92 Some authors have suggested that pregnancy might even have a beneficial effect on the disease.93 Disease activity does appear to determine the effect of IBD on pregnancy outcome, although pregnancy is not contraindicated in patients with even the most severe disease. The goals of the treating physician are thus to minimize IBD symptoms and morbidity prior to conception. Most experts agree that during gestation affected patients should continue optimized prepregnancy therapy to avoid possible flares resulting from medication withdrawal. Exacerbations of IBD that do occur during pregnancy should be managed aggressively because they may result in fulminant colitis and have serious consequences, including hemorrhage, perforation, sepsis, fetal demise, and premature labor. Treatment of fulminant colitis is the same as in nonpregnant individuals, namely high-dose glucocorticoids, intravenous antibiotics, cyclosporine, and salvage biological therapies (see Table 38-2). The indications for bowel surgery likewise are the same as in nonpregnant patients, although bowel surgery is associated with premature labor as well as maternal and fetal mortality.94,95 A Turnbull-Blowhole colostomy to achieve colonic decompression and fecal diversion may be safer than total colectomy.96 Synchronous cesarean section and subtotal colectomy have been advocated for patients with fulminant colitis after 28 weeks of gestation.97 IBD patients are at risk for poor pregnancy outcomes, even

629

630

Section IV  Topics Involving Multiple Organs if they have mild or inactive disease.98 Major complications include premature birth, low-birth-weight and smallfor-gestational-age infants, and increased cesarean section rates.99 The risk of fetal malformations in this setting is unclear.100 The majority of IBD patients require several medications to remain symptom-free. Limited reliable safety data are available on most commonly used IBD drugs, thus it is important to carefully review the possible risks and known benefits of therapy with patients before conception. Potentially teratogenic drugs should be discontinued before conception, if at all possible. Methotrexate and thalidomide (pregnancy category X) are known teratogens and abortifacients, and should be used with caution in patients of childbearing age. The optimum period of abstinence from these medications before conception is unknown; a minimum of six months is recommended. The 5-aminosalicylates (all pregnancy category B except osalazine, which is pregnancy category C) are widely used during pregnancy to treat mild IBD. A prospective study of pregnant patients treated with mesalamine, as well as a large case series, did not show any increased risk of teratogenicity from this therapy.101,102 Azathioprine and its metabolite, 6-mercaptopurine (pregnancy category D), also are commonly used as maintenance treatments in patients with mild to moderate IBD. They both cross the placenta and are excreted in breast milk; however, data concerning use of these agents in the transplant setting have failed to confirm the teratogenicity that had been seen in animal studies.103 A study of pregnant IBD patients treated with 6-mercaptopurine failed to demonstrate an increase in preterm delivery, spontaneous abortion, congenital abnormalities, or childhood neoplasia.104 Based on these data and extensive experience with this drug and its metabolites in pregnant women, experts concur that their discontinuation before or during pregnancy is not advisable. Instead, dose reduction and careful monitoring of metabolite levels in the mother are recommended.105 Glucocorticoids (pregnancy category C) have been used for decades to treat pregnant patients with moderate to severe IBD as well as other more common glucocorticoidresponsive diseases, such as asthma. Early reports suggested an increased risk of congenital malformations in the infants of treated mothers.102 Subsequent prospective studies and substantial experience with drugs in this class confirm that the risk of malformations secondary to their use is extremely low. Glucocorticoid treatment during pregnancy is, however, associated with other complications including maternal glucose intolerance, macrosomia, and fetal adrenal suppression.106 Prednisolone (pregnancy category C) is more efficiently metabolized by the placenta than other glucocorticoids and may pose less risk of adrenal suppression.107 Adverse outcomes have not been reported after use of oral budesonide (pregnancy category C) during pregnancy.108 Many organ transplant recipients have been treated with cyclosporine (pregnancy category C) without reports of significant teratogenicity. At present, there are very few data concerning the safety of treatment with other immunomodulatory agents in pregnant women. These compounds are immunoglobulins and as such are capable of reaching the fetal compartment, especially during the third trimester, when they are actively transported across the placenta. High levels of infliximab (pregnancy category B) were detected in infants exposed to the drug in utero.109 Post-marketing registries of safety data and small case series have not identified an increased incidence of fetal malformations or miscarriage in women treated during pregnancy with infliximab or adalimumab.110 Experts have suggested that therapy with antibodies against tumor necrosis factor-α (TNF-α) be discontinued early in the third trimester to avoid

significant fetal exposure until better data on the safety of these agents are available; when necessary glucocorticoids may be substituted.87

APPENDICITIS (see Chapter 116) Suspected acute appendicitis is the most common nonobstetric indication for exploratory laparotomy in pregnant women.111,112 Appendicitis complicates approximately 1 in 1500 pregnancies, and may develop at any time during the course of gestation.112 Diagnosis may be difficult because the enlarging uterus displaces the cecum cephalad, altering the location of pain caused by appendiceal inflammation, and resulting in increasingly delayed detection as pregnancy progresses.113 Late diagnosis of an inflamed appendix is responsible for complications that are associated with excess maternal and fetal morbidity and mortality.114 During all three trimesters of pregnancy, right lower quadrant pain is the most common presenting symptom of appendicitis.115 In addition to pain, affected individuals frequently complain of nausea, but this symptom often is difficult to interpret during gestation. Graded-compression ultrasonography is the diagnostic test of choice for pregnant patients suspected of having appendicitis.115 Helical CT also has been reported to be helpful in this setting.112 Pregnant patients with appendicitis during any trimester may be treated with laparoscopic appendectomy,116 although potential interference by the gravid uterus may be a relative contraindication to this procedure during the third trimester.117 Appropriate supportive care can prevent fetal loss associated with appendiceal perforation.118

BILIARY AND PANCREATIC DISORDERS AND PREGNANCY GALLSTONE DISEASE (see Chapter 65)

Pregnant women tend to form gallstones because of changes in gallbladder function and bile composition (see earlier). Gallstones frequently are noted during gestation when ultrasonographic examination is performed to evaluate the fetus119; the prevalence of gallstones in asymptomatic pregnant women is reported to be between 2.5% and 12%. Despite this high prevalence, the incidence of acute cholecystitis is not increased by pregnancy. Cholecystitis is probably more common in the postpartum period than during gestation.120 Other complications of cholelithiasis, including choledocholithiasis and pancreatitis (discussed later), also are rare in pregnant women. Initial conservative management of suspected gallstonerelated disease with intravenous fluids, analgesia, and antibiotics has been recommended to reduce maternal and fetal morbidity incident to surgery.121 A more aggressive operative approach, however, may be associated with superior outcomes. Open cholecystectomy in the first trimester of pregnancy can precipitate abortion, whereas in the third trimester it can induce premature labor. Many experts believe that laparoscopic cholecystectomy is the preferred approach when surgery is indicated in cases of acute cholecystitis, even near term when the uterus is very large.122-125 Endoscopic extraction of bile duct stones with minimal use of fluoroscopy and appropriate maternal shielding is acceptable when necessary to treat choledocholithiasis in pregnant women.125

PANCREATITIS (see Chapter 58) Acute pancreatitis is uncommon during gestation, occurring once in every 1066 to 3300 pregnancies.126,127 Most cases are due to gallstones and present during the third trimester or

Chapter 38  Gastrointestinal and Hepatic Disorders in the Pregnant Patient the puerperium. The hypertriglyceridemia normally seen in pregnant women may be more severe in persons with familial hyperlipidemia, predisposing them to develop pancreatitis on this basis.128 The clinical characteristics of acute pancreatitis during gestation are similar to those in nonpregnant women.

HEPATIC DISORDERS UNIQUE TO PREGNANCY Pregnant women may develop liver diseases that are etiologically related to gestation or one of its complications. As a rule, these disorders become clinically evident during the third trimester or just after delivery. They may be severe, even life threatening, but affected individuals are expected to survive with prompt diagnosis and appropriate management. Liver diseases unique to pregnancy are associated with increased fetal morbidity and mortality.

CHOLESTASIS OF PREGNANCY

Cholestasis of pregnancy is a form of intrahepatic cholestasis associated with pruritus, elevated serum bile acid levels, and the findings of bland cholestasis on liver biopsy.129 This disorder may have a variable course, making it difficult to diagnose.130 It nevertheless has serious implications for fetal well-being, and cases must be identified as promptly as possible.131 Cholestasis of pregnancy usually presents in the third trimester, but may be seen earlier in gestation, even in the first trimester. Its first and most characteristic symptom is pruritus, and as a result patients may be referred to a dermatologist for initial evaluation. As in other forms of cholestasis, the pruritus of cholestasis of pregnancy is most severe in the skin of the palms and soles and experienced most intensely at night. Only 10% to 25% of affected individuals later develop jaundice. Elevated serum bile acid levels (>10 µmol/L) confirm the presence of cholestasis; some patients with the disorder also have bilirubinuria and even mild hyperbilirubinemia.132 Serum alkaline phosphatase concentrations are modestly increased, but gamma glutamyl transpeptidase (GGTP) levels are normal or only marginally elevated.132 The latter pattern of test results is atypical of adult cholestasis, but is seen in pediatric patients with progressive familial intrahepatic cholestasis, as in Byler’s syndrome.133 Serum aminotransferase levels are elevated in affected women, sometimes to values of 1000 U/L or higher, making it difficult, on occasion, to distinguish cholestasis of pregnancy from hepatitis.134 Symptoms and laboratory test abnormalities of patients may wax and wane. Intense cholestasis is associated with steatorrhea that usually is subclinical but can cause fat-soluble vitamin deficiencies, most notably deficiency of vitamin K. Improvement of symptoms and laboratory test results begins with delivery of the infant, and usually, although not invariably, is prompt and complete. Rare patients experience prolonged cholestasis that may be indicative of underlying biliary tract disease, such as primary biliary cirrhosis or sclerosing cholangitis.135,136 Women with ordinary cholestasis of pregnancy have no residual hepatic defect after resolution of the disorder, but they are at increased risk for development of gallstones, cholecystitis, and pancreatitis.137 In addition, 60% to 70% of affected indivi­ duals develop cholestasis during subsequent pregnancies (although recurrent episodes may be less severe than the initial one) or with use of oral contraceptives. The risk of recurrence with subsequent pregnancy is increased by interval cholecystectomy.138

Cholestasis of pregnancy has serious implications for fetal well-being. There are many reports of increased frequencies of fetal distress, unexplained stillbirth, and need for premature delivery in the babies of women with this disorder.139 Fetal hypoxia and meconium staining have been reported at delivery in 19% of Swedish women with cholestasis of pregnancy.140 These complications were shown to correlate with maternal bile acid levels greater than 40 µmol/L. Although the risk to the fetus may be reduced by close monitoring of affected mothers, it cannot be eliminated completely.141-144 Planned early elective delivery as soon as the fetal lungs have matured has been recommended for this reason. As discussed in Chapter 64, a number of the molecular mechanisms of bile formation have been elucidated in recent years, resulting in a more sophisticated understanding of many cholestatic disorders.145 Mutations of the MDR3 (ABCB4) gene are likely responsible for approximately 15% of cases of cholestasis of pregnancy.146-148 The MDR3 gene product is a phospholipid flippase that translocates phosphatidylcholine from the inner to the outer leaflet of the canalicular hepatocyte membrane where it is solubilized by bile acids to form mixed micelles. There is, however, no relationship of cholestasis of pregnancy to human leukocyte antigen (HLA) type.149 Environmental and hormonal factors likely also contribute to development of cholestasis in pregnant women. In Chile and Scandinavia, where cholestasis of pregnancy is common, the disorder occurs most often during colder months. The incidence of cholestasis of pregnancy in Chile has declined, possibly due to a fall in mean plasma selenium levels.150 An increased sensitivity to the cholestatic effects of exogenous estrogen has been demonstrated in family members, including male relatives, of patients who develop cholestasis while pregnant.151 Therapeutic or experimental administration of estrogen compounds to susceptible women can precipitate the disorder.152,153 Similarly, progesterone therapy during gestation is associated with development of cholestasis.154 The finding that ursodeoxycholic acid alters the metabolism of progesterone may explain its therapeutic effect in this setting.155 It is possible that some women with cholestasis of pregnancy have inherited an enhanced sensitivity to estrogen or a variation in the metabolism of progesterone that causes cholestasis in response to a variety of stimuli, including some medications and dietary factors. The differential diagnosis of cholestasis of pregnancy includes other cholestatic disorders such as primary biliary cirrhosis, primary sclerosing cholangitis, benign recurrent intrahepatic cholestasis, viral hepatitis, toxic liver injury, and bile duct obstruction. Liver biopsy specimens of affected individuals reveal bland changes typical of cholestasis due to a variety of etiologies, but biopsy usually is not necessary to make the diagnosis. It is important to remember that pregnancy may exacerbate a preexisting subclinical cholestatic disorder. For example, a family of sisters with progressive liver disease who also developed recurrent severe cholestasis of pregnancy was described in 1997.135 Management of cholestasis of pregnancy is primarily palliative. Ursodeoxycholic acid is helpful in relieving symptoms, may reduce fetal complication rates,146 and is well tolerated by mother and fetus.131,156,157 Studies of treated individuals have demonstrated a change in the bile acid content of maternal serum and amniotic fluid, as well as increased placental bile acid transport.158-160 Most investigators have prescribed a conventional dose (15 mg/kg/day), although one report suggests that a higher dose (20 to 25 mg/ kg/day) is more effective.153 Treatment with bile-acid binders such as cholestyramine157 and guar gum also may relieve

631

632

Section IV  Topics Involving Multiple Organs symptoms,161 but it is important to keep in mind that therapy with these agents worsens steatorrhea and resultant fat-soluble vitamin deficiencies.162 Administration of Sadenosyl-l-methionine (SAMe) to patients with cholestasis of pregnancy has had mixed therapeutic results131,163-165; use in combination with ursodeoxycholic acid may increase its benefit.166 A short course of oral dexamethasone (12 mg/day for 7 days) has been reported to reduce itching and serum bile acid levels in persons with this disorder167 but also was associated with clinical deterioration in one case.168 Sedatives, such as phenobarbital, may relieve itching in cholestasis patients, but may adversely affect the fetus. Exposure to ultraviolet B light has been suggested as therapy in this setting. As in other cholestatic syndromes, no treatment is always and completely effective in women with cholestasis of pregnancy, with the usual exception of delivery.

LIVER DISEASE OF PREECLAMPSIA

Preeclampsia is a disease of unclear etiology that remains difficult to define and, on occasion, to diagnose.169 In broad terms, preeclampsia is a form of pregnancy-related hypertension that is associated with damage and dysfunction of one or more maternal organs, possibly including the liver, that may produce severe, even life-threatening, complications and affect pregnancy outcome. Preeclampsia complicates 3% to 10% of pregnancies, occurring in the second half of pregnancy or the puerperium and most commonly, but not exclusively, in primiparous women or women with multiple gestations.170 Usual criteria for making the diagnosis include a sustained blood pressure of 140/90 mm Hg or greater after the twentieth week of pregnancy in a previously normotensive woman, accompanied by proteinuria (≥300 mg/24 hr), which is approximately equivalent to a urine protein concentration of 30 mg/dL (“1+ dipstick”) on random testing.169 Many patients also are hyperreflexic and have edema. Liver disease is recognized as a common and potentially ominous complication of preeclampsia. The HELLP syndrome (hemolysis, elevated liver enzymes, low platelet count), first described by Weinstein in 1982,171 is the most usual form of preeclamptic liver disease and is presumed to underlie development of hepatic hematoma, rupture, and infarction in some women with this disorder.172-174 Recent evidence, however, suggests that there are different preeclampsia phenotypes and that HELLP syndrome may be a

distinct genetic and clinical entity.175 Although preeclampsia is common in patients with acute fatty liver of pregnancy (AFLP) and may play a role in the pathogenesis of this disorder, AFLP usually is not classified as a preeclamptic liver disease.176

HEMOLYSIS, ELEVATED LIVER ENZYMES, AND LOW PLATELET COUNT (HELLP) SYNDROME

HELLP syndrome is seen in up to 12% of women with severe preeclampsia.177-179 In addition to the diagnostic abnormalities of hemolysis, elevated serum aminotransferase levels, and thrombocytopenia in conjunction with hypertension and proteinuria, patients with typical HELLP syndrome frequently have complaints of chest, epigastric, and right upper quadrant abdominal pain (Table 38-4). These symptoms often are accompanied by nausea, vomiting, headache, and blurred vision in varying combinations. Some pregnant patients, however, may present with an asymptomatic fall in the platelet count during observation for preeclampsia, or initially have no hypertension or proteinuria.180 Other women may complain of malaise, suggesting the diagnosis of a viral syndrome.181 Most affected individuals seek treatment after the twenty-seventh week of gestation, but up to 11% may do so earlier. It is important to note that presentation of HELLP syndrome after delivery, despite absence of signs of preeclampsia at delivery, occurs in up to 30% of cases.177,182 The diagnosis of HELLP syndrome is based on an assessment of the clinical circumstances and features of the illness at the time of presentation; there is no single diagnostic test that confirms the presence of the disorder.179,183 Hemolysis in patients with HELLP is mild. Fragmented red-blood cells (schistocytes) are seen on smears, and the serum lactate dehydrogenase (LDH) level is elevated. Serum aminotransferase levels also are elevated, sometimes minimally and other times greater than 1000 U/L in association with laboratory signs of cholestasis.177,184 Serum bilirubin levels often are elevated, but in most patients to low levels compatible with the finding of hemolysis. Elevated serum levels of glutathione S-transferase alpha,185 d-dimer,186 tissue polypeptide antigen (TPA),187 and fibronectin188 have been described in persons with HELLP syndrome, and these tests may have some use in predicting the presence or severity of liver disease.

Table 38-4  Clinical Characteristics of Hemolysis, Elevated Liver Enzymes, and Low Platelet Count (HELLP) Syndrome Presenting Symptom Abdominal pain (right upper quadrant; epigastric) Nausea or vomiting Headache Bleeding Jaundice Laboratory Test Level (Normal Value) Serum aspartate aminotransferase (<40 U/L) Serum bilirubin (<1 mg/dL) Platelets (>125 × 103/mm3) Maternal Complications Disseminated intravascular coagulation Abruptio placentae Acute kidney injury Hepatic subcapsular hematoma Death

Percent Affected 65 36 31 9 5 Median (Range) 249 (70-633) 1.5 (0.5-25) 57 (7-99) Percent Affected 21 16 8 1 1

From Sibai BH, Ramadan MK, Usta I, et al. Maternal morbidity and mortality in 442 pregnancies with hemolysis, elevated liver enzymes and low platelets (HELLP) syndrome. Am J Obstet Gynecol 1993; 169:100-6.

Chapter 38  Gastrointestinal and Hepatic Disorders in the Pregnant Patient

Figure 38-1.  Histology of hemolysis, elevated liver enzymes, low platelet count (HELLP) syndrome. The portal triad on the left of the figure (the horizontal arrow points to an interlobular bile duct in the portal triad) is surrounded by pockets of hemorrhage (vertical arrows) and by an area of fibrin deposition (to the left of the portal triad).

Cross-sectional abdominal imaging, especially CT and MRI, may be useful in making the diagnosis of HELLP syndrome and detecting cases of intrahepatic hemorrhage and infarction. Imaging should be performed in patients with complaints of severe abdominal pain, neck or shoulder pain, or a sudden drop in blood pressure. One report documented abnormal findings in 45% of such patients.172 Liver biopsy specimens demonstrate periportal hemorrhage, intrasinusoidal fibrin deposition and irregular areas of liver cell necrosis with mild reactive hepatitis, findings characteristic of preeclampsia (Fig. 38-1). If steatosis is present, it is modest and does not have the appearance of the extensive pericentral microvesicular fat accumulation that occurs in patients with AFLP (discussed later). There is little if any correlation between the severity of liver biopsy lesions and laboratory test abnormalities in patients with HELLP syndrome; mild thrombocytopenia and mild increases in serum aminotransferase levels do not connote insignificant liver damage.189 Liver biopsy, however, rarely is necessary to make a diagnosis in these patients, and possibly may precipitate development of intraparenchymal hepatic hematoma or contained hepatic rupture. Although most pregnant women with low platelet counts and preeclampsia have HELLP syndrome, the differential diagnosis includes other causes of thrombocytopenia, including immune thrombocytopenic purpura, thrombotic thrombocytopenic purpura,190 and the antiphospholipid antibody syndrome.191 Elevated serum aminotransferase levels in patients with preeclampsia are most frequently misdiagnosed as being caused by viral hepatitis.192 A diagnosis of AFLP also should be considered in patients with clinical findings of HELLP syndrome, but AFLP usually is associated with signs of more significant liver disease and possibly liver failure, albeit with lower serum amino­ transferase levels, and is not necessarily associated with thrombocytopenia. Any hypothesis concerning the pathophysiology of preeclampsia must explain the known characteristics of the disorder, in particular the inappropriately high systemic vascular resistance and low plasma volume seen in affected individuals. Preeclampsia follows, and presumably is a consequence of, an abnormality of placental formation in which a failure of both trophoblast invasion of the uterine lining

and dilation of the spiral arteries result in the physiologic inability to increase uteroplacental perfusion appropriately as gestation progresses.193 Female relatives, including the mothers of patients with preeclampsia, often have a history of the disorder, and evidence exists in some populations for its inheritance as either an autosomal recessive trait or an autosomal dominant trait with variable penetrance.175,194,195 Women with a circulating procoagulant, for example factor V Leiden or anticardiolipin antibody, are at risk for developing early and severe preeclampsia.196,197 There is convincing evidence that development of this disorder is mediated by excessive production of soluble fms-like tyrosine kinase 1 (sFlt1), a potent antagonist of vascular endothelial growth factor (VEGF) and placental growth factor (PlGF), and of soluble endoglin (sEng), an inhibitor of capillary formation.198,199 Pregnant rats that overexpress both sEng and sFlt1 develop proteinuria, severe hypertension, laboratory findings of HELLP syndrome, and intrauterine growth restriction.200 Women with preeclampsia similarly produce excessive amounts of sFlt1 and sEng.199 The events responsible for abnormal production of these factors in this setting have yet to be elucidated. Preeclampsia frequently is diagnosed in patients with AFLP. Although the latter disorder is not formally classified as a hepatic complication of preeclampsia, these two disease entities may share a common pathogenesis. Women with long-chain 3-hydroxyacyl-CoA dehydrogenase (LCHAD) deficiency, a cause of defective intramitochondrial beta oxidation of fatty acids, have been reported to develop preeclampsia and HELLP syndrome while pregnant with an affected fetus.201,202 Abnormal mitochondrial fatty-acid metabolism also may play a role in the pathogenesis of AFLP (see following). Other studies, however, have not demonstrated LCHAD deficiency in women with HELLP syndrome.203 Additional hypotheses advanced to explain the pathogenesis of preeclampsia include lipid peroxidation and oxidative stress in response to an unknown primary insult, endothelial dysfunction,188,204 abnormal fluidity of the endothelial cell membrane, abnormal cell permeability to calcium, or inheritance of a molecular variant of angiotensin known to be associated with hypertension.205 Interestingly there is no animal model for the human syndrome of preeclampsia. Furthermore, the histologic characteristics of this condition are unique and are not similar to those of any other known liver disease in humans or animals. The clinical abnormalities that characterize HELLP syndrome usually resolve rapidly after childbirth.206 Transient postpartum diabetes insipidus has been reported in women with the disorder.207 Rarely, HELLP syndrome becomes gradually worse prior to delivery with subsequent development of postpartum liver failure, sepsis, consumptive coagulopathy, and, rarely, even death.208 In the absence of appropriate supportive therapy and expedited delivery, affected patients may progress to renal failure, hepatic hematoma, and hepatic rupture. Neither serum aminotransferase levels nor platelet counts are predictive of outcome in women with HELLP syndrome.209 The disorder can recur during subsequent pregnancies but usually does not.210,211 Management of HELLP syndrome is primarily supportive; patients should be treated in an intensive care setting prior to delivery, preferably by an obstetrician qualified in the practice of maternal-fetal medicine. Some affected patients may have a decline in serum aminotransferase levels and a rise in platelet counts with supportive care.212 Under such circumstances, a delay in delivery may be appropriate in cases of fetal immaturity, but the fetus usually fails to grow in the setting of preeclampsia. Patients with severe pre-

633

634

Section IV  Topics Involving Multiple Organs eclampsia and HELLP syndrome may require antepartum platelet transfusions and hemodialysis. Plasmapheresis after delivery has been advocated by some authorities but has not been proven to alter disease outcome.182,213 Glucocorticoid therapy also has been used in this setting. Many affected women receive glucocorticoids before delivery, not as disease treatment, but to promote fetal lung maturity214; glucocorticoid therapy of HELLP syndrome per se has not been evaluated in a controlled clinical trial.215-217 Orthotopic liver transplantation may be appropriate treatment for some HELLP-syndrome patients,218,219 but early diagnosis and prompt delivery almost always make this and other extreme therapeutic measures unnecessary in these individuals. Full recovery without sequelae is anticipated for the vast majority of affected patients.

HEPATIC RUPTURE, HEMATOMA, AND INFARCT

Spontaneous rupture of the liver may complicate preeclampsia and HELLP syndrome, usually in the third trimester of pregnancy close to term or in the early postpartum period. Patients with this often fatal disorder present with abdominal distention and pain, and cardiovascular collapse.220,221 In contrast to other patients with preeclampsia, women with spontaneous hepatic rupture tend to be older and to have had multiple previous pregnancies. Diagnosis is made by signs of liver rupture on ultrasonography, CT, or MRI in conjunction with aspiration of blood on para­ centesis.172,222 Imaging studies often show that affected patients have a partially contained subcapsular hematoma (Fig. 38-2).223 Management must be aggressive, with rapid delivery of the fetus by the obstetrician and repair of the liver, preferably by an experienced liver surgeon. Post­ operatively, patients have a protracted course that may include disseminated intravascular coagulation and hepatic failure. A few patients with hepatic rupture have undergone

orthotopic liver transplantation after emergent hepatectomy and interval portosystemic shunt as a temporizing measure while a donor graft is sought.219,224,225 Rarely, a hemodyna­ mically stable patient with a ruptured liver can be successfully treated without surgery.226 Survivors of hepatic rupture may have uneventful subsequent pregnancies,227 but recurrence of hematoma and rupture also have been reported.228 Some pregnant women with preeclampsia, HELLP syndrome, and abdominal pain have contained subcapsular hematoma. In this circumstance, patients can be observed with serial CT imaging and managed without surgery.172,228 Experts have recommended angiographic embolization of hepatic artery branches supplying blood to the affected portion of the liver in such cases. Hepatic hematoma and rupture complicating preeclampsia presumably result from extravasation of blood from one or several microscopic areas of periportal hemorrhage under Glisson’s capsule. Periportal hemorrhage is a typical pathologic finding in the livers of patients with preeclampsia and HELLP syndrome.229 The capsule is believed to be stretched and torn away from the surface of the liver by the enlarging hematoma. Ultimately, the capsule ruptures, allowing the liver surface to bleed freely into the peritoneal cavity. Necrotic hepatic infarcts also may complicate preeclampsia. Affected individuals present with fever, leukocytosis, anemia, and marked elevation of serum aminotransferase levels,172,174,226 and in the most severe cases develop multiorgan failure, including liver failure. Cross-sectional imaging demonstrates confluent hepatic infarcts. Needle aspiration of these areas yields blood and necrotic tissue; immediately adjacent liver parenchyma contains periportal hemorrhage and fibrin deposition typical of preeclampsia and HELLP syndrome. Hepatic infarction sometimes is associated with the presence of a hypercoagulable condition such as factor V Leiden or antiphospholipid antibody.230

ACUTE FATTY LIVER OF PREGNANCY

Figure 38-2.  Subcapsular hepatic hematoma in a patient with preeclampsia. This coronal section of a T1-weighted magnetic resonance imaging scan demonstrates the subcapsular clot or hemorrhage (horizontal arrows) adjacent to the liver (vertical arrow). (From Barton JR, Sibai BM. Hepatic imaging in HELLP syndrome [hemolysis, elevated liver enzymes, and low platelet count]. Am J Obstet Gynecol 1996; 174:1820.)

AFLP is a form of microvesicular fatty liver disease unique to human gestation that presents late in pregnancy, often as fulminant hepatic failure with sudden onset of coagulopathy and encephalopathy in a woman without a prior history of liver disease.231,232 AFLP is diagnosed on the basis of typical clinical and pathologic manifestations in approximately 1 of 6700 third-trimester pregnancies,233 but it also is recognized that subclinical cases exist.234 The pathophysiologic mechanisms underlying development of this disorder are unknown, although at least some patients with AFLP have an inherited LCHAD deficiency that also affects the fetus.235,236 AFLP presents late in pregnancy; in the majority of cases, symptoms develop between 34 and 37 weeks of gestation, although cases beginning as early as 19 to 20 weeks of gestation have been reported. Rarely, the onset of AFLP is after delivery. Initial symptoms usually include nausea and vomiting, often associated with abdominal pain. Pruritus may be an early complaint; overlap with cholestasis of pregnancy has occurred but is rare.237 Patients with AFLP frequently are confused, and have pregnancy-related complications, such as premature labor, vaginal bleeding, and decreased fetal movement. The disorder is most common in primiparous women and in women with multiple gestations.238 Affected individuals have a much greater than expected number of male fetuses (2.7 : 1).239 Of note is the fact that preeclampsia, an accompanying diagnosis in 21% to 64% of cases,233,240 also is associated with first pregnancies, twin pregnancies, and male fetuses.

Chapter 38  Gastrointestinal and Hepatic Disorders in the Pregnant Patient On laboratory evaluation, women with AFLP often have prolonged prothrombin times and decreased serum fibrinogen levels as well as leukocytosis. Their serum aminotransferase levels usually are moderately elevated (≈750 U/L), but rarely may be very high or even normal. Jaundice is common but not invariable. Initial blood tests often reflect renal dysfunction with elevations of the serum creatinine, uric acid, and blood urea nitrogen levels. The course of AFLP is quite variable. Hypoglycemia and hyperammonemia occur and should be suspected when at-risk patients exhibit signs of altered central nervous system function. Other complications of liver failure, including ascites, pleural effusion, acute pancreatitis, respiratory failure, renal failure, and infection may develop in patients with AFLP; vaginal bleeding or post–cesarean section bleeding is common in these individuals. Transient diabetes insipidus sometimes is seen241; more rarely, affected patients have myocardial infarction242 or pulmonary fat emboli.243 Diagnosis of AFLP almost always is based on the appearance of typical clinical features of the disorder, including laboratory test results, during the later stages of pregnancy. Hepatic imaging is not reliable in confirming the presence of AFLP,244 but plays a crucial role in identifying hepatic hematoma, rupture and infarction. Liver biopsy usually is unnecessary to make the diagnosis, but histologic results may be pathognomonic and therefore useful if the obstetrician has reservations about delivery. However, transjugular sampling of liver tissue may be necessitated by coagulopathy. The histologic hallmark of AFLP is microvesicular fatty infiltration of the liver that is most prominent in hepatocytes surrounding central veins (zone three) and spares those surrounding portal areas (Fig. 38-3). Microvesicular steatosis of this type has a relatively homogeneous appearance on light microscopy and may be difficult to discern on examination of ordinary hematoxylin and eosin–stained specimens. To confirm the diagnosis, special techniques must be used; frozen tissue may be stained for fat with oil-red O, or electron microscopy can be used to examine a glutaraldehydefixed specimen. Plans must be made prior to the biopsy for appropriate handling of the liver tissue. Other histologic findings in affected patients can be misleading; they may include lobular disarray suggestive of viral hepatitis and biliary ductular proliferation and inflammation sugges-

Figure 38-3.  Histology of acute fatty liver of pregnancy. The perivenular hepatocytes are pleomorphic and vacuolated, and there is lobular disarray. Large fat droplets are not seen.

tive of cholangitis.234,245 Patients with AFLP do not have the periportal hemorrhage and fibrin deposition seen in the livers of individuals with preeclampsia and HELLP syndrome. The differential diagnosis in suspected cases of AFLP includes those causes of acute hepatic failure not associated with pregnancy, especially viral hepatitis and toxic liver injury. Uncommon types of viral hepatitis, such as hepatitis E and herpes simplex hepatitis, may be more severe in pregnant than in nonpregnant individuals.246,247 These viral agents can be identified by appropriate serologic tests. A more difficult problem is distinguishing AFLP from other liver diseases that complicate pregnancy, particularly the preeclamptic liver diseases, namely HELLP syndrome and hemorrhagic or ischemic liver injury. For example, patients with AFLP may develop preeclampsia and disseminated intravascular coagulation with attendant thrombocytopenia, thereby meeting the diagnostic criteria for HELLP syndrome. Fortunately it is not usually necessary to distinguish among these various diagnoses because AFLP, HELLP syndrome, and preeclampsia are treated by expedited delivery of the infant. It is, however, of crucial importance to recognize hepatic hematoma and rupture rapidly. The pathogenesis of AFLP, like that of preeclampsia, has not been elucidated. Initially AFLP was thought to be caused by exposure to a toxin; microvesicular steatosis of the liver is known, for example, to be caused by treatment with sodium valproate or intravenous tetracycline. Despite an intensive search, however, no toxin that might be responsible for development of AFLP has been identified. Because of the coincidental occurrence of preeclampsia and AFLP in many patients, the disorder has been considered by some experts to be a severe form of preeclamptic liver disease.234,248,249 Arguing against this conclusion are the absence of the usual histologic features of preeclampsia in liver biopsy specimens from patients with AFLP and the absence of the usual clinical features of preeclampsia in many patients with AFLP. There is a well-established association between AFLP and inherited defects in beta oxidation of fatty acids.202,236,250 This connection is empirically supported by similar clinical and histologic findings in patients with AFLP and those with Jamaican vomiting sickness, a liver disease caused by a toxin in unripe akee fruit that disables intramitochondrial beta oxidation of fatty acids. Maternal liver disease (HELLP or AFLP) has been reported in 62% of the mothers of infants with defects of fatty-acid oxidation.236 AFLP may develop regardless of maternal genotype if the fetus is deficient in LCHAD and carries at least one allele for the G1528C LCHAD mutation.235 Another beta oxidation defect, carnitine palmitoyltransferase I deficiency, also has been associated with AFLP.251 Prenatal genetic diagnosis based on chorionic villus sampling has proved to be feasible and accurate in pregnant members of affected families.252,253 Not every investigator, however, has been able to confirm the association between AFLP and beta oxidation defects,254 and other as yet unknown mechanisms may play a role in the pathogenesis of this disorder. Patients with AFLP should be managed in an intensive care setting, preferably by obstetricians qualified in the practice of maternal-fetal medicine in cooperation with other appropriate specialists. Early diagnosis and prompt delivery of the infant are imperative to minimize maternal and fetal morbidity and mortality. Affected individuals may be very ill post-partum until the physiologic defects responsible for their clinical abnormalities resolve and the livers recover. Supportive care may include infusion of blood products, mechanical ventilation, hemodialysis, and antibi-

635

636

Section IV  Topics Involving Multiple Organs otic therapy. Hepatic encephalopathy is treated as indicated by measures intended to evacuate feces and bacteria from the colon. Infusion of concentrated glucose solution may be required to treat or prevent hypoglycemia. Although many patients with AFLP have disseminated intravascular coagulation and depressed antithrombin III levels, treatment with heparin or antithrombin III is not recommended.255 Patients with diabetes insipidus may be managed with 1-deamino8-d-arginine-vasopressin (DDAVP).241 Some individuals with liver failure secondary to AFLP require emergency orthotopic liver transplantation as a potentially life-saving measure.256,257 Most affected women, however, recover completely with appropriate supportive care. Persistent or even increasing hyperbilirubinemia and multiple complications after delivery do not necessarily indicate the need for liver transplantation. Survival of patients with AFLP has been reported to be 100% with prompt diagnosis, delivery of the infant, and intensive care.233,240,258 Infants of affected women have perinatal mortality rates of less than 7%; the surviving baby may have LCHAD deficiency and develop nonketotic hypoglycemia and obtundation. Recurrence of AFLP has been documented, particularly in women with LCHAD deficiency.259,260 In cases of AFLP the mother, father, and child should be tested for the G1528C LCHAD mutation.235

USUAL HEPATIC DISORDERS AND PREGNANCY VIRAL HEPATITIS

Viral hepatitis is the most common form of liver disease worldwide, and frequently affects women of childbearing age, either as an acute infection or as a chronic disease. Hepatitis A does not appear to alter the normal course of pregnancy, nor does pregnancy appear to influence the natural history of hepatitis A. Acute and chronic viral hepatitis of other types, however, may have implications for maternal well-being as well as the outcome of gestation.

Hepatitis E Virus (see Chapter 80)

Hepatitis E virus infection occurs in nonindustrialized nations, usually as an epidemic disease during the monsoon season in central and south Asia and India. Hepatitis E is rare in the West. Cases have been reported in travelers returning to the United States from endemic areas, particularly Mexico. Acute hepatitis E during the third trimester of pregnancy is a cause of fulminant hepatic failure and has a mortality rate of up to 20%.181 Maternal hepatitis E virus infection also has been associated with intrauterine fetal death.261,262 The risks of intrauterine death and abortion in any trimester are greater in pregnant women with hepatitis E than they are in their uninfected counterparts. Maternalfetal transmission of hepatitis E resulting in symptomatic neonatal hepatitis has occurred263; no known therapy prevents vertical transmission of this virus. Pregnant women should avoid traveling to endemic areas during monsoon season and outbreaks of the disease.

Herpes Simplex Virus (see Chapter 81)

Subclinical hepatitis associated with primary herpes simplex virus infection is common. In pregnant or immunosuppressed individuals, this virus may cause severe liver disease. Infection during pregnancy, particularly the third trimester, can result in fulminant hepatic failure.247 Affected individuals are obtunded and usually anicteric with elevated serum aminotransferase levels and coagulopathy. They may have subtle oropharyngeal or genital herpetic

lesions. Encephalopathy may result from herpes encepha­ litis. The diagnosis of herpes simplex virus infection can be confirmed by serologic testing. Liver biopsy specimens from affected patients usually demonstrate characteristic intracytoplasmic inclusion bodies and areas of focal hemorrhage. Treatment with acyclovir is effective and appears to prevent viral transmission to the fetus.247

Hepatitis B and D Viruses (see Chapter 78) As a rule, maternal hepatitis B virus infection does not influence the course of pregnancy, nor does pregnancy alter the natural history of maternal hepatitis B infection. Most women of childbearing age with chronic hepatitis B are healthy virus carriers with a very low risk of developing complications of their disease. The importance of hepatitis B during pregnancy is related to its role in the perpetuation of chronic infection through vertical transmission: maternal-fetal transmission of hepatitis B virus is responsible for most cases of chronic hepatitis B worldwide, especially in Southeast Asia and Africa.264 Mothers with a reactive serum test for hepatitis B e antigen have more circulating virus and higher rates of perinatal transmission than do mothers who have undetectable serum hepatitis B e antigen and a reactive serum test for hepatitis B e antibody,265 although the latter individuals can still be a source of neonatal infection.266 Without treatment, 90% of infants born to hepatitis B e antigen–positive mothers and 10% of infants born to hepatitis B e antigen–negative mothers develop hepatitis B virus infection. The infants of mothers with a reactive serum test for hepatitis B surface antigen should receive hepatitis B immunoglobulin at birth and hepatitis B vaccine during the first day of life and at ages one and six months.267 Women with chronic hepatitis B are not treated with interferon during pregnancy.268 There is a paucity of data on the use of oral agents in this setting; treatment of hepatitis B with nucleotide and nucleoside analogs generally is guided by experience in pregnant patients with human immunodeficiency virus (HIV) infection.269 Telbivudine and tenofovir are pregnancy category B drugs and are therefore preferred. Lamivudine is a pregnancy category C drug, but is thought to be associated with a low risk of complications270 and has been reported to reduce the incidence of neonatal vaccination failure.271 Data concerning the safety of entecavir and adefovir in pregnant women classified as pregnancy category C drugs are insufficient to allow any conclusions. Hepatitis D virus infection requires simultaneous acute or chronic hepatitis B virus infection. There is no evidence that pregnancy changes the natural course of hepatitis D. Prevention of vertical transmission of hepatitis D is best accomplished by vaccination of the mother against infection with hepatitis B virus, or appropriate therapy of existing maternal hepatitis B prior to pregnancy in conjunction with vaccination and administration of immunoglobulin to the infant. A case report has documented prevention of vertical transmission of hepatitis B and D viruses by this management.272 Hepatitis C Virus (see Chapter 79) Chronic hepatitis C virus infection does not appear to affect the outcome of pregnancy,273 and there are no convincing data to suggest that pregnancy alters the natural history of hepatitis C infection. Vertical transmission of hepatitis C virus is uncommon274,275 unless maternal serum virus titers are unusually high, as sometimes occurs in patients with HIV coinfection.276 Serum levels of hepatitis C virus ribonucleic acid (RNA) greater than or equal to 1019 copies per mL have been associated with vertical transmission in as

Chapter 38  Gastrointestinal and Hepatic Disorders in the Pregnant Patient many as 36% of cases.277 The incidence of perinatal infection does not seem to be related to whether the baby is delivered vaginally or by cesarean section.277 Although hepatitis C viral RNA can be detected in breast milk,278 breastfeeding is not considered to be a risk factor for neonatal infection.279 Vertical transmission of hepatitis C is not prevented by treatment of the infant with immunoglobulin.272 Women with hepatitis C are not treated with interferon and ribavirin during pregnancy; ribavirin is a well established teratogen (pregnancy category X).

CHRONIC LIVER DISEASE AND PORTAL HYPERTENSION (see Chapter 90)

Women with chronic liver disease and resultant cirrhosis often have anovulatory menstrual cycles or are amenorrheic and therefore unlikely to become pregnant. Portal hyper­ tension, ascites, and compensatory dilation of submucosal esophageal veins connecting the portal circulation and the azygos vein can occur in pregnant women with noncirrhotic portal hypertension aggravated by increased circulating blood volume. Even in the absence of pathologic causes of portal hypertension, these esophageal venous collaterals may become engorged during gestation due to normal circulatory changes, including increased blood flow and compression of the inferior vena cava by the enlarging uterus, and may be seen on endoscopy. Enlarged veins of the latter type do not bleed spontaneously. The effect of pregnancy on the risk of variceal bleeding in women with underlying portal hypertension is unknown. Furthermore, the most appropriate method for treating esophageal varices during pregnancy is controversial. β-Adrenergic receptor antagonists are tocolytic, but these drugs do not prevent normal labor in chronically treated pregnant patients. Use of beta blockers as primary prophylaxis against variceal bleeding in pregnant women has not been formally evaluated. Vasopressin and octreotide infusions to treat acute variceal hemorrhage may cause uterine ischemia and induce premature labor. Some authors have suggested prophylactic band ligation, portosystemic shunt procedures, and cesarean section to decrease the risk of bleeding from varices during gestation. Ascites and hepatic encephalopathy are no more common in pregnant than in nonpregnant women with chronic liver disease, and when they occur are managed in the customary manner.

WILSON DISEASE (see Chapter 75) Wilson disease in women of childbearing age is associated with amenorrhea and infertility. Treatment of affected individuals to remove excess copper may result in resumption of ovulatory cycles and a subsequent pregnancy. Pregnant patients must remain on medication to treat Wilson disease because discontinuation of therapy can cause sudden copper release, hemolysis, acute liver failure, and death.280 D-penicillamine is potentially teratogenic in humans,281 but has been used safely during pregnancy at doses necessary for copper chelation.282 Similarly, trientine is teratogenic in animals but appears to be safe in humans as treatment for copper overload. Zinc is not teratogenic, and some experts favor its use during pregnancy as therapy for Wilson disease for this reason.283 AUTOIMMUNE LIVER DISEASES

(see Chapters 88 and 89) Autoimmune diseases of all types, including autoimmune hepatitis, are more common in women than in men. In women, classic (type 1) autoimmune hepatitis typically

presents around the expected time of menarche but is associated with amenorrhea. Immunosuppressive therapy is highly effective in controlling the disease in most patients; treated women who subsequently conceive a child should continue taking immunosuppressive medications during pregnancy. The doses of azathioprine prescribed as part of standard treatment regimens are not thought to be teratogenic. Occasionally autoimmune liver disease will worsen during the postpartum period when the physiologic immunosuppression of pregnancy resolves. For this reason, affected patients should have frequent measurements of serum aminotransferase levels for approximately six months after delivery. Primary biliary cirrhosis is much more common in postmenopausal women than it is in their fertile counterparts. Pregnant women with primary biliary cirrhosis may experience an exacerbation of pruritus284 that can be ameliorated by treatment with ursodeoxycholic acid,285 although the safety of this therapy during pregnancy has not been formally proven.

HEPATIC NEOPLASIA AND MASS LESIONS

(see Chapter 94) Mass-like defects of the hepatic parenchyma may be discovered during pregnancy, usually as an incidental finding on ultrasonography. Benign liver lesions found commonly in women of childbearing age include adenomas, focal nodular hyperplasia, and hemangiomas. Hepatic adenomas are associated with oral contraceptive use and may enlarge during pregnancy; growing lesions can bleed and rupture into the abdominal cavity. Focal nodular hyperplasia and hemangiomas in pregnant patients also have been reported to hemorrhage. Women known to have a benign hepatic nodular defect should be evaluated with serial ultrasonography to measure mass size and look for evidence of intralesional bleeding. Hepatocellular carcinoma occurs almost exclusively in persons with chronic liver disease and may present in the absence of cirrhosis in young people with chronic hepatitis B virus infection. At-risk patients should have standard screening for liver cancer during pregnancy. It must be borne in mind that maternal serum alpha-fetoprotein levels always are modestly elevated during normal pregnancy,286 and can further increase in cases of fetal Down syndrome, neural tube defects, and hydatidiform mole, thus limiting the positive predictive value for diagnosing hepatocellular carcinoma. Hepatic fibrolamellar carcinoma has been reported to occur in pregnant women.287 Fibrolamellar carcinoma is a slow-growing cancer usually found in young adults; the median age of affected persons is 25 years.288 Unlike typical primary liver cancer, this neoplasm has no known association with cirrhosis or chronic liver disease and is not a cause of increased serum alpha-fetoprotein levels. Although hepatic metastases are approximately 30 times more common than primary liver cancer, most malignant tumors are rare in women of childbearing age.

HEPATIC VEIN THROMBOSIS (BUDD-CHIARI SYNDROME) (see Chapter 83)

Pregnancy is a predisposing factor for the development of venous thrombosis. Hepatic vein thrombosis may occur in association with HELLP syndrome289 and with preeclampsia in women who have antiphospholipid antibody.290 Pregnant women who develop hepatic vein thrombosis should be evaluated for the presence of antiphospholipid antibody

637

638

Section IV  Topics Involving Multiple Organs and other circulating procoagulants, for example factor V Leiden.291

PREGNANCY AFTER LIVER TRANSPLANTATION

After successful orthotopic liver transplantation, women of childbearing age may become pregnant and deliver normal infants.103,292 Delaying pregnancy until the second post-transplant year may be associated with a lower risk of prematurity. Transplant patients should continue immunosuppressive therapy during gestation; treatment regimens used to prevent graft rejection have not been associated with teratogenicity. Adverse effects of these medications, however, including hypertension and hyperglycemia, may increase the incidence of fetal distress and preeclampsia in liver transplant recipients who are pregnant. In rare instances, pregnancy has been complicated by organ rejection.

KEY REFERENCES

ACOG Committee on Obstetric Practice. Guidelines for diagnostic imaging during pregnancy. Committee opinion no: 299. Obstet Gynecol 2004; 104:647-51. (Ref 37.) ACOG Committee on Practice Bulletins—Obstetrics. Diagnosis and management of preeclampsia and eclampsia. No.: 33. Obstet Gynecol 2002; 99:159-67. (Ref 169.) Francella A, Dyan A, Bodian C, et al. The safety of 6-mercaptopurine for childbearing patients with inflammatory bowel disease: A retrospective cohort study. Gastroenterology 2003; 124:9-17. (Ref 104.)

Kanal E. Pregnancy and safety of magnetic resonance imaging. Magn Reson Imaging Clin N Am 1994; 2:309-17. (Ref 39.) Keely E, Barbour LA, Lee RV, editors. Medical care of the pregnant patient. 2nd ed. Philadelphia: ACP Press; 2008. Chapter 3. (Ref 28.) Lammert F, Marschall HU, Glantz A, et al. Intrahepatic cholestasis of pregnancy: Molecular pathogenesis, diagnosis and management. J Hepatol 2000; 33:1012-21. (Ref 129.) Mahadevan U. American Gastroenterological Association Institute technical review on the use of gastrointestinal medication in preg­ nancy. Gastroenterology 2006; 131:283-311. (Ref 268.) Mahadevan U, Sandborn WJ, Li DK, et al. Pregnancy outcomes in women with inflammatory bowel disease: A large community-based study from Northern California. Gastroenterology 2007; 133:1106-12. (Ref 98.) McKay DB, Josephson MA. Pregnancy in recipients of solid organs— Effects on mother and child. N Engl J Med 2006; 354:1281-93. (Ref 103.) Medical radiation exposure of pregnant and potentially pregnant women. Bethesda (MD): National Council on Radiation Protection & Measurements; 1977. Report No: 54. (Ref 36.) Mofenson LM; Centers for Disease Control and Prevention, U.S. Public Health Service Task Force. Recommendations for use of antiretroviral drugs in pregnant HIV-infected women for maternal health and interventions to reduce perinatal HIV transmission in the United States. MMWR Recom Rep 2002; 51(RR-18):1-38. (Ref 269.) Nikfar S, Abdollahi M, Moretti ME, et al. Use of proton pump inhibitors during pregnancy and rates of major malformations: A meta-analysis. Dig Dis Sci 2002; 47:1526-9. (Ref 74.) Qureshi WA, Rajan E, Adler DG, et al. ASGE guideline: Guidelines for endoscopy in pregnant and lactating women. Gastrointest Endosc 2005; 61:357-362. (Ref 30.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

39 Radiation Injury Brian G. Czito and Christopher G. Willett

CHAPTER OUTLINE Molecular Mechanisms of Radiation-Induced Gastrointestinal Damage  639 Apoptosis  639 Role of Cytokines  639 Esophagus  640 Incidence and Clinical Features  640 Treatment and Prevention  641 Stomach  641 Incidence and Clinical Features  641 Treatment and Prevention  642 Small Intestine  642 Incidence and Clinical Features  642 Treatment and Prevention  645

Early and late gastrointestinal injury may occur following irradiation of thoracic, abdominal, and pelvic malignancies of gastrointestinal and non-gastrointestinal origin. Toxicity to the gastrointestinal tract frequently limits radiation doses that can be delivered for many tumor types. As with most other toxicities associated with radiotherapy, gastrointestinal side effects are categorized into two broad types: early or acute reactions, such as diarrhea and nausea, experienced during and soon following the completion of a course of therapy; and late or chronic reactions, such as ulceration, stricture formation, and bowel obstruction, that can arise months to years after the course of radiation therapy. The incidence and severity of radiation-induced morbidity depend on total radiation dose, radiation fraction size, treatment volume, treatment techniques, and the presence or absence of other treatment modalities including systemic chemotherapy and surgery. A grading system for adverse events (severity, one to five) has been developed.1 This chapter discusses the early and late responses of the esophagus, stomach, small and large intestine, rectum, anus, and liver to radiation and combined radiation-chemotherapy treatment regimens.

MOLECULAR MECHANISMS OF RADIATIONINDUCED GASTROINTESTINAL DAMAGE Radiation injury is the phenotype of a complex set of interactions between multiple cytokines and molecular pathways. Stromal injury with subsequent progressive fibrosis is the most significant component of radiation injury. The risk of fibrosis likely increases when surgery or chemotherapy is combined with radiotherapy. In addition, the volume of tissue irradiated, total dose, and dose per fraction influence the development and severity of radiation-induced fibrosis.

Large Intestine  647 Incidence and Clinical Features  Treatment and Prevention  649 Anus  650 Incidence and Clinical Features  Treatment  650 Liver  650 Incidence and Clinical Features  Treatment  651 Therapeutic Techniques to Reduce

647 650 650 Toxicity  651

APOPTOSIS

In animal studies, a rapid increase in the rate of programmed cell death (apoptosis) of intestinal crypts cells can be observed after exposure to low-dose radiation (1 to 5 cGy). The rate of apoptosis is radiation dose dependent and reaches a plateau at 1 Gy. Radiation exposure also increases expression of the tumor suppressor gene P53 in the stem cell region. The apoptosis induced by radiation is dependent on the presence of P53. Additionally, the rate of spontaneous and radiation-induced apoptosis is significantly increased in animals lacking bcl-2, suggesting a protective effect of bcl-2 against radiation-induced apoptosis.2 It is therefore postulated that P53 promotes apoptosis after irradiation and that bcl-2 protects the mucosa. The expression of higher levels of bcl-2 may explain the increased tolerance of the colorectal mucosa to radiation as compared with the small intestine (see following).

ROLE OF CYTOKINES

Ionizing radiation activates the translation of the gene coding for transforming growth factor-β (TGF-β) in the intestines. TGF-β is a potent fibrogenic and proinflammatory cytokine, leading to hyperplasia of connective tissue mast cells and leukocyte migration into the intestinal wall. TGF-β promotes intestinal fibrosis by stimulating the expression of collagen and fibronectin genes and the chemotaxis of fibroblasts. The extracellular matrix of the intestine is also increased as TGF-β inhibits its degradation. The increased expression of TGF-β is especially enhanced in areas with histopathologic changes consistent with radiation damage: areas with mucosal ulceration, mucosal and serosal thickening, inflammatory cell infiltrates, and vascular sclerosis.3,4 TGF-β exists in three isoforms: TGF-β1, TGF- β2, and TGFβ3. All three isoforms are overexpressed in the early postradiation phase. However, only isoform β1 remains elevated six months after radiation exposure. In the first 2 weeks after radiation, TGF-β1 messenger RNA is increased in epithelial cells, fibroblasts of the submucosa and subserosa, vascular

639

640

Section IV  Topics Involving Multiple Organs endothelial cells, and smooth muscle cells of the intestinal wall. However, at 26 weeks, the expression of TGF-β1 of epithelial cells returns to baseline level but TGF-β1 expression remains elevated in vascular endothelial cells, fibroblasts, and smooth muscle cells.5 Compared with control mouse intestine, the TGF-β1 immunoreactivity or overexpression is substantially increased in areas of radiationinduced acute and late bowel injury.6,7 In addition, pathologic examination of bowel specimens from patients undergoing surgery for radiation enteropathy showed an increased TGF-β in areas with vascular sclerosis and fibrotic areas of the serosa and muscularis propria as compared with patients who have surgery for other causes.8 Neutralizing antibodies to TGF-β and gene therapy using decorin (a natural TGF-β inhibitor) have been shown to suppress or reverse fibrosis in preclinical models.9 Epidermal growth factors, interleukins, and tumor necrosis factors are also being investigated for their effects in chronic radiation injury.10 Another cytokine implicated in the development of radiation injury is connective tissue growth factor (CTGF). The CTGF expression is increased in intestinal radiation fibrosis associated with chronic radiation injury.11 CTGF is found commonly in the extracellular region surrounding the area of active fibrosis or neovascularization. TGF-β1 may induce CTGF, which in turn functions as a mediator of intestinal radiation fibrosis by sustaining the activation of fibrogenesis in the irradiated gastrointestinal tract. Mechanisms underlying the pathogenesis of radiation-induced gastrointestinal damage remain an active area of investigation.

ESOPHAGUS INCIDENCE AND CLINICAL FEATURES

Early and late injury of the esophagus often occurs following irradiation of malignances of the thorax and upper abdomen (e.g., esophageal/gastroesophageal junctional carcinomas and lung carcinomas). The normal esophageal mucosa undergoes continuous renewal or turnover. Acute esophageal side effects are believed to be primarily related to radiation damage to the basal epithelial layer, manifested histologically by vacuolization, resulting in epithelial thinning followed by denudation (Fig. 39-1). These changes manifest clinically as dysphagia, odynophagia, and sub­ sternal discomfort, usually occurring within two to three weeks following initiation of radiation therapy. Patients may describe a sudden, sharp, severe chest pain radiating to the back. As treatment progresses, pain may become constant and may not necessarily be related to swallowing. The symptoms may be confused with Candida esophagitis, which may occur in conjunction with radiation esophagitis. Concurrent chemotherapy exacerbates these toxic effects. Endoscopically, mucositis and ulceration may be observed. Perforation and bleeding are rare in the acute phase.12 After treatment completion, basal proliferation returns and regeneration occurs, usually within three weeks.13 Following recovery from acute injury, late effects such as benign stricture, persistent ulceration and fistula formation may occur months to years following treatment. These effects are believed primarily due to inflammation and scar formation within the esophageal muscle. The connective tissues surrounding the esophagus may also exhibit severe fibrosis over time.14 Small vessel telangiectasias may be seen endoscopically. Histologic studies of the esophagus in previously irradiated patients have demonstrated epithelial thickening, chronic inflammation, fibrosis of the submucosa

Figure 39-1.  Acute radiation-induced esophageal injury. Acute esophageal ulceration with abundant fibroblasts is seen. (Hematoxylin and eosin, ×400.)

and muscularis propria, and (rarely) chronic ulceration. Complete epithelial recovery from radiation effects may take 3 to 24 months.15 Late effects often manifest as dysphagia due to stricture as well as altered motility due to fibrosis or muscular damage, possibly with accompanying nerve injury. Fistula formation is unusual and radiation dose dependent. Barium swallow examination may show strictures as well as disruption of peristalsis at the level of the irradiated esophagus with repetitive and nonperistaltic waves above and below the irradiated region. Abnormal peristalsis has been reported at 1 to 3 months following treatment completion, whereas most strictures occur 4 to 8 months following treatment completion. Late effects are usually not seen until 3 months following completion of radiation therapy, with a median time to onset of 6 months in some series.16,17 Development of radiation-related late complications is dose related. The TD5/5 (i.e., dose at which 5% of patients will develop complications at five years) has been estimated to be 60 Gy when one third of the length of the esophagus is irradiated.18 Other series have reported complication rates of less than 1% to 30% with doses of approximately 60 Gy, with benign stricture as the primary complication. Seaman and Ackerman treated patients to doses of 60 to 75 Gy, resulting in severe esophagitis and stricture formation in some patients. They concluded that the upper limit of dose tolerance of the esophagus was 60 Gy given at 10 Gy per week.19 Other studies reported that patients receiving 60 Gy had late complication rates of 1.2% to 18%. With contemporary radiation doses of 50 Gy, late complication rates have been observed in 2% of patients or less.20,21 Brachytherapy (the temporary insertion of a radioactive source into or adjacent to a tumor) has also been used as a technique for radiation dose escalation in esophageal cancer. Gaspar and colleagues reported the results of a phase I/II study examining the role of brachytherapy in addition to external beam radiation therapy in the treatment of esophageal cancer. The one-year actuarial fistula formation rate was 18%, and the authors recommended caution in the use of this approach, particularly in conjunction with concurrent chemotherapy.22 In contrast, other authors have reported much lower rates of fistula formation with brachytherapy.23 The length of esophagus being irradiated may not closely

Chapter 39  Radiation Injury correlate with the incidence of esophagitis after radiotherapy.24,25 However, the intensity of cancer treatment such as use of concurrent chemotherapy with radiation therapy increases the rate of acute esophagitis.26 Maguire and colleagues evaluated 91 patients treated with radiation therapy for non–small cell lung cancer and found that the percent esophageal volume and surface area treated to greater than 50 Gy predicted late esophageal toxicity. Patients who had preexisting gastroesophageal reflux disease and esophageal erosions secondary to tumor were at increased risk for late toxicity. Hyperfractionation (multiple daily radiation treatments) was also associated with increased acute toxicity.27 Singh and associates studied patients with non–small cell lung cancer who received conformal daily radiation therapy with or without concurrent chemotherapy; they found that a maximal esophageal “point” dose of 69 Gy (radiation therapy alone) and 58 Gy (with concurrent chemotherapy) predicted significant toxicity. Twenty-six percent of patients receiving concurrent chemoradiotherapy developed grade 3 or higher esophageal toxicity, whereas only 1.3% of patients who received radiation therapy alone experienced this degree of toxicity.28 Ahn and colleagues found that the most powerful predictor of late esophageal toxicity in 254 patients treated for non–small cell lung cancer was the severity of acute esophageal toxicity. Severe acute toxicity was predicted by the use of twice daily radiation, older age, increasing nodal stage and a variety of dosimetric parameters. The overall incidence of late toxicity was 7%, with a median and maximal time to onset of 5 and 40 months, respectively.29 The Radiation Therapy Oncology Group, using standard radiation therapy techniques, reported grade 3 or higher acute esophageal toxicity in 34% of patients treated with concurrent hyperfractionated radiation therapy and chemotherapy versus 1.3% of patients treated with standard thoracic radiotherapy alone.30 Based on these and other data, it is clear that the addition of concurrent chemotherapy to radiation therapy increases the incidence of esophageal toxicity.

TREATMENT AND PREVENTION

The treatment and prevention of radiation-induced esophagitis have come under increased attention with the use of aggressive combination chemotherapy and radiation therapy regimens. The treatment of acute esophagitis is based on the grade of symptoms experienced by the patient. Treatment interruptions may ease the symptoms of acute esophagitis, but may also compromise the patient’s cancer treatment. Treatment interruption is reserved for severe cases. The management of acute esophagitis usually includes symptomatic management such as topical anesthetics (including viscous lidocaine-based regimens), oral analgesics (including anti-inflammatory agents and narcotics), gastric anti­ secretory drugs (histamine [H2] blockers, proton pump inhibitors), promotility agents (such as metoclopramide), and treatment of superimposed infection (candidiasis). Dietary modification, including bland foods as well as pureed or soft foods and soups, can help patients maintain food and liquid intake. Other modifications include avoidance of smoking, alcohol, coffee, spicy or acidic foods or liquids, chips, crackers, fatty foods, and indigestible foods. A study of dietary modifications and pharmacologic prophylaxis for radiation-induced esophagitis reported decreased toxicity and fewer treatment interruptions. It was recommended to drink between meals and to eat six smaller meals per day, consisting of semisolid food, soup, highcalorie supplements, purees, puddings, milk, and soft breads.31 Additionally, ingestion of hot or cold foods should be avoided if possible; instead foods and liquids should be

at room temperature. In severe cases, feeding tube placement may be required. Radioprotective chemical agents have been investigated as a means of mitigating radiation-induced normal tissue toxicity. The best-studied of these is amifostine, an organic thiophosphate. This agent is a scavenger of free radicals and serves as an alternative target to nucleic acids for alkylating or platinum agents.32 In one trial, patients treated with chemotherapy and radiation therapy for non–small cell lung cancer were randomized to receive amifostine or no drug. Amifostine did not significantly reduce grade 3 or higher esophagitis in these patients. However, patient self-assessments suggested a significantly lower incidence of acute esophagitis in those who received amifostine.33 Other trials have demonstrated a protective effect,34,35 whereas others have not confirmed this.36 Larger, randomized and placebo-controlled trials are needed to determine the ultimate efficacy of amifostine in preventing radiation injury of the esophagus. The management of late esophageal radiation stricture consists of endoscopic dilatation (often serially), usually resulting in symptomatic improvement. Additionally, longterm use of gastric antisecretory drugs as well as prokinetic agents such as metoclopramide have been recommended to decrease gastroesophageal (GE) reflux effects. Dilatations in advanced stricture can cause esophageal rupture and therefore their use should be approached cautiously. Uncommonly, tube feedings may be required for patients with significant weight loss or for those only able to take in liquids. Surgical intervention may be required for patients who develop perforation or fistula. Finally, it is important to note that the clinical symptoms associated with late radiation injury are often difficult to distinguish from those caused by recurrent (or new) cancer. Therefore, patients with strictures or ulcerations should also be evaluated to differentiate chronic radiation changes from cancer recurrence.

STOMACH INCIDENCE AND CLINICAL FEATURES

The stomach may be damaged following irradiation of the upper abdomen for cancer, including esophageal-GE junctional, gastric, and pancreatic carcinomas. Radiation to the stomach in animals using a very high single dose of irra­ diation results in erosive and ulcerative gastritis. A slightly lower single dose (23 Gy) results in gastric dilatation and gastroparesis, with replacement of the normal gastric mucosa by hyperkeratinized squamous epithelium. With even lower doses, gastric obstruction occurring months after irradiation was observed, with an atrophic gastric mucosa and intestinal metaplasia seen in surviving animals.37 Studies in which serial gastric biopsies were obtained following irradiation of patients for peptic ulcer disease noted coagulation necrosis of chief and parietal cells with mucosal thinning, edema, and chronic inflammatory infiltration.17,38 In addition, gastric acid production decreased after relatively low doses of gastric irradiation. In the past, radiotherapy had been used to decrease acid production in patients with peptic ulcer disease. Even with a relatively low dose of 18 Gy delivered in 10 fractions, approximately 40% of ulcer patients had a 50% reduction in gastric acid secretion that lasted for a year or more.39 Clinically, radiation-induced gastritis may occur within a week of starting radiotherapy, with microscopic changes including edema, hemorrhage, and exudation. Histologic

641

642

Section IV  Topics Involving Multiple Organs changes may also include disappearance of cytoplasmic details and granules in parietal and chief cells as early as one week into therapy. Cell damage and subsequent cell death are often seen first in the depths of glands, followed by thinning of the gastric mucosa.40 Additional mucosal changes include deepening of the glandular pits and proliferation of cells in the glandular neck. Loss of glandular architecture and thickening of the mucosa can be seen by the third week of radiotherapy. Approximately three weeks after completing radiotherapy, histologic recovery may be seen. Signs of recovery of early radiation injury to the stomach include re-epithelialization and fibrosis. Symptoms of acute radiation injury of the stomach consist primarily of nausea and vomiting, dyspepsia, anorexia, abdominal pain, and malaise. These are more common with the concurrent administration of chemotherapy. Radiationinduced nausea and vomiting may occur within the first 24 hours following treatment. It is estimated that approximately half of patients receiving upper abdominal radiation will experience emesis within two to three weeks following initiation of irradiation.41 Late effects of gastric irradiation have been classified into four categories: (1) acute ulceration (occurring shortly after completion of radiation therapy); (2) gastritis with smoothened mucosal folds and mucosal atrophy on endoscopy accompanied by radiographic evidence of antral stenosis (1 to 12 months following irradiation) (see Chapter 51); (3) dyspepsia, consisting of vague gastric symptoms without obvious clinical correlate (6 months to four years following irradiation); and (4) late ulceration (averaging 5 months after irradiation).17,42 The TD5/5 for treatment of the entire stomach has been estimated to be 50 Gy. Large studies of upper abdominal irradiation have suggested that prior abdominal surgery as well as using a higher radiation dose per fraction of radiation may increase the risk of late effects.43 Studies from Walter Reed Army Medical Center delivering abdominal radiation using now antiquated techniques in testicular cancer patients have suggested that higher radiation doses lead to an increasing risk of late ulceration and perforation, with ulceration occurring in approximately 6% of patients treated to 45 to 50 Gy, 10% of patients treated to 50 to 60 Gy, and 38% of patients treated to greater than 60 Gy. No ulceration was seen in patients receiving less than 45 Gy. In this series, symptomatic gastritis occurred approximately 2 months following radiation completion, with ulcer formation occurring at a median of 5 months. Six of 233 patients (3%) required surgery for ulcer hemorrhage or pain related to ulcer disease, almost all of whom had received doses of greater than 50 Gy.17,45 Other studies of patients treated with radiation therapy for Hodgkin’s lymphoma or for testicular, gastric, or cervical cancer have established tolerance limits for gastric irradiation.43-46 These studies delivered doses of 40 to 60 Gy. Patients who received doses greater than 50 Gy experienced gastric ulceration and gastric ulcer–associated perforation at rates of 15% and 10%, respectively. If indicated, the dose to the entire stomach with conventionally administered radiation therapy is limited to 45 to 50 Gy, with an estimated 5% to 10% risk of severe radiation toxicity. Where appropriate, reduced field boosts can be given to treat to doses up to 55 Gy with acceptable toxicity. As in the esophagus, combining chemotherapy with radiation therapy decreases the tolerance of the gastric mucosa to radiation therapy. 5-Fluorouracil (5-FU) is the most common chemotherapy agent delivered concurrently with radiation therapy in the management of gastrointestinal tumors. This agent can be given in an adjuvant or neoadjuvant setting or as “definitive” therapy for gastroesophageal

junction, gastric, pancreatic and biliary cancers. 5-FU is a radiation sensitizer, but has historically been given safely with radiation therapy at doses of 45 to 50 Gy without substantial increases in toxicity. Newer systemic agents have been shown to increase acute gastric toxicity when delivered with radiotherapy, including taxanes, gemcitabine, and epidermal growth factor inhibitors. These regimens remain the subject of investigation in the treatment of abdominal malignancies.

TREATMENT AND PREVENTION

Acute symptoms of gastric radiation in toxicity are treated with antiemetics (5-hydroxytryptamine-3 [5-HT3] antagonists, phenothiazines, metoclopramide, glucocorticoids, benzodiazepines, antihistamines, or anticholinergics), as well as consumption of a light meal prior to delivery of radiation therapy. Randomized trials of prophylactic 5-HT3 inhibitors have shown efficacy compared with placebo in preventing radiation-induced nausea and vomiting.47 A randomized trial of 211 patients receiving upper abdominal radiation compared the 5-HT3 inhibitor ondansetron given twice daily, with or without dexamethasone delivered daily for the first five fractions of treatment. Patients receiving dexamethasone showed a trend toward improved complete control of nausea (50% versus 38%) and significant improvement in complete control over emesis. The authors concluded that the addition of dexamethasone resulted in modest improvement in protection against radiationinduced emesis.48 Narcotic and non-narcotic agents are often used for pain. Additionally, it is generally recommended that patients be placed on antacid medications, including proton pump inhibitors. Careful nutritional support along with antiemetic therapy is essential for patients undergoing radiotherapy to the abdomen. Acute symptoms generally resolve within one to two weeks following completion of radiation therapy. Late gastritis-related symptoms are often treated with antacids (H2 antagonists, proton pump inhibitors, sucralfate, and so on). These may be used on a long-term basis to avoid late ulceration. With more severe complications of bleeding, ulceration, gastric outlet obstruction, fistula formation, or perforation, patients may require endoscopic therapeutic approaches or rarely surgical intervention with partial gastrectomy.

SMALL INTESTINE Small bowel injury, or radiation enteritis, is the primary treatment-limiting toxicity in the radiotherapeutic management of abdominal and pelvic malignancies. The small bowel can be damaged during radiation treatment of malignances of the stomach, pancreas, rectum, and anus, as well as during treatment of gynecologic malignancies. The first case of radiation enteropathy was described in 1897.49

INCIDENCE AND CLINICAL FEATURES

The epithelium of the gastrointestinal tract has a high proliferative rate, making it susceptible to radiation- and chemotherapy-induced mucositis. The intestinal lining is normally replaced every three to five days, reflecting this high cellular turnover rate. Irradiation of intestinal mucosa primarily affects the clonogenic intestinal stem cells within the crypts of Lieberkühn (cells that provide, via selfreplication and eventual maturation, replacement cells in the intestinal villi). Stem cell damage, either as a primary result of radiation damage or as a result of radiation-

Chapter 39  Radiation Injury

Figure 39-2.  Microabscesses and radiation-related fibroblasts. Submucosal reaction shows large, bizarre radiation fibroblasts that have both cytomegaly and nucleomegaly. Smooth muscle cells also have reactive changes. Microabscesses composed of excess neutrophils infiltrate the stroma. (Hematoxylin and eosin, ×400.) (Courtesy Dr. Robin Amirkahn, Dallas, Tex.)

induced microvascular damage, leads to a decrease in cellular reserves for the intestinal villi, resulting in mucosal denudement, shortened villi, and decreased absorptive area, with associated intestinal inflammation and edema. Histologic changes are seen within hours of irradiation. Within two to four weeks, an infiltration of leukocytes with crypt abscess (microabscess) formation can be seen (Fig. 39-2). Ulceration may also occur. This acute injury can result in impaired absorption of fats, carbohydrates, proteins, bile salts, and vitamin B12, with associated loss of water, electrolytes, and protein. Additionally, impaired ileal bile salt absorption increases loads of conjugated bile salts entering the colon, which are in turn deconjugated by colonic bacteria, resulting in intraluminal water retention with resultant diarrhea. Furthermore, impaired digestion of lactose may occur following radiation, leading to increased bacterial fermentation with associated flatulence, distention, and diarrhea, possibly accompanied by bacterial overgrowth. There is also evidence of acutely altered gut motility following radiation therapy.50 Patients with acute radiation enteritis experience diarrhea, abdominal cramping or pain, nausea and vomiting, anorexia, and malaise. Radiation-induced diarrhea often appears during the third week of a fractionated radiation course, with reported rates of 20% to 70%.51 Acute radiation enteropathy with diarrhea may be seen in some patients after delivery of doses of 18 to 22 Gy using conventional fractionation, and in most patients receiving doses of 40 Gy. The symptoms and pathologic findings typically subside and spontaneously disappear two to six weeks following completion of radiation therapy.52 However, growing evidence suggests that patients who develop acute small intestine toxicity may be at higher risk for chronic effects.53 Histologic changes of chronic toxicity to the small intestine include progressive occlusive vasculopathy with foam cell invasion of the intima and hyaline thickening of the arteriolar walls, as well as collagen deposition and fibrosis, often in the submucosal layer of the bowel wall. The small bowel becomes thickened with development of telangiectasias, whereas the vessel walls of small arterioles are obliterated, causing ischemia (Fig. 39-3).54 As the vasculopathy

Figure 39-3.  Submucosal arteriole in chronic radiation enteropathy. Radiation-induced changes include thickening of vessel walls and subin­ timal hydropic change and fibrosis. Vessel wall thickening results in luminal narrowing/occlusion and subsequent tissue ischemia. (Hematoxylin and eosin, ×200.) (Courtesy Dr. Robin Amirkahn, Dallas, Tex.)

Figure 39-4.  Small intestinal submucosal fibrosis following radiation therapy. The patient presented with small intestinal obstruction due to this stricture. (Hematoxylin and eosin, ×40.) (Courtesy Dr. Robin Amirkahn, Dallas, Tex.)

progresses, mucosal ulceration, necrosis, and occasionally perforation of the intestinal wall can be seen, resulting in fistula and abscess formation. Lymphatic damage results in constriction of the lymphatic channels, which contributes to mucosal edema and inflammation.55 Histologically, the mucosa atrophies, with atypical hyperplastic glands and intestinal wall fibrosis (Fig. 39-4).15 As the ulcers heal, there can be fibrosis with narrowing of the intestinal lumen with subsequent stricture formation and even obstruction with dilatation of the proximal bowel. Bacterial overgrowth may be an indirect complication arising from stasis in a dilated loop of bowel proximal to the stricture. Although the affected segments of intestine and serosa appear thickened with areas of telangiectasias,56 it should be noted that even if the gut appears normal, patients can still be at risk of spontaneous perforation.57

643

644

Section IV  Topics Involving Multiple Organs Table 39-1  Clinical Complications of Chronic Radiation Enteritis or Proctitis* complication

LESION(S)

SYMPTOMS

Obstruction

Stricture

Infection

Abscess

Fistulization

Fistula

Bleeding

Ulceration

Malabsorption

Small bowel damage

Constipation, nausea, vomiting, postprandial abdominal pain Abdominal pain, fever, chills, sepsis, peritonitis Fecal, vaginal, or bladder discharge; pneumaturia Rectal pain, tenesmus, rectal bleeding, anemia Diarrhea, steatorrhea, weight loss, malnutrition, cachexia

*From Girvent M, Carlson GL, Anderson I, et al: Intestinal failure after surgery for complicated radiation enteritis. Ann R Coll Surg Engl 2000; 82:198.

Chronic radiation enteritis can cause significant morbidity. This complication tends to be progressive, with an onset at least 6 months after radiotherapy. Late radiation injury to the small intestine occurs at a median of 8 to 12 months following radiation therapy, though it can appear years later.17 There are numerous clinical manifestations of the chronic phase of radiation enteritis (Table 39-1). These include malabsorption and diarrhea, with more rapid transit times occurring in the affected bowel. Rarely, chronic malnutrition may develop, resulting in anemia and hypoalbuminemia. There can be bleeding from ulceration and pain and bloating from strictures, as well as fevers from abscess. Fibrosis and vasculitis of the bowel may lead to dysmotility, stricture, and malabsorption.58,59 Malabsorption and other complications may require surgical intervention and parenteral alimentation (discussed in Chapters 4 and 5 and following). Patients with severe chronic radiation enteritis have a poor long-term prognosis with a mortality rate of approximately 10%.60-66 The overall incidence of chronic radiation enteritis has not been precisely defined. Retrospective series suggest an incidence of 5% to 15%; however, these studies often include a large number of patients who were lost to follow-up or died between the end of radiation therapy and the completion of the study. A review of randomized trials of adjuvant radiation therapy for rectal cancer shows severe long-term complications as low as 1.2% and as high as 15%.67 In older series of radiotherapy for abdominal cancers, symptoms of chronic bowel dysfunction were present in many patients, although such symptoms are often multifactorial and influenced by other treatment modalities including surgery and chemotherapy.68,69 However, it appears that advances in radiation therapy treatment techniques have reduced chronic small intestinal toxicity rates. Certain factors have been found to predispose patients to radiation toxicity to the small intestine. Women, older patients, and thin patients may have a larger amount of small bowel in the pelvic cul-de-sac, which can increase the probability of radiation injury in the treatment of pelvic malignancies.70 Patients with a history of pelvic inflammatory disease or endometriosis also appear to be at higher risk of radiation complications.71,72 Patients who have had previous abdominal surgery can develop adhesions that decrease the mobility of the small bowel, thereby allowing it to be consistently exposed to fractionated radiation therapy.73,74 In addition, patients with prior pelvic surgery may have an increase in the amount of small bowel within the pelvis, allowing increased exposure during pelvic irradiation. In a

series published by Eifel and associates, the risk of small bowel complications was significantly higher in women who had undergone a previous laparotomy.75 With modern treatment approaches (allowing direct visualization of the volume of bowel in treatment field) and use of improved treatment techniques (discussed later), as well as a shift toward neoadjuvant approaches, this risk factor may not be as relevant as in prior years. Patients with diabetes, hypertension, and cardiovascular disease also have an increased risk of pre-existing vascular damage or occlusion.76 These comorbid conditions are compounded by the pathologic changes of chronic radiation injury, which include vasculopathy and ischemia, predisposing these patients to radiation-related small bowel toxicity. Patients with collagen vascular disease and inflammatory bowel disease also have a higher risk of acute as well as chronic radiation-induced injury. Patients with these diseases may have pathologic changes which include transmural fibrosis, collagen deposition, and inflammatory infiltration of the mucosa. The late effects induced by radiation therapy to the small bowel are likely additive to these preexisting changes, and studies have shown that these patients have a lower gastrointestinal tolerance to radiation therapy.77-79 Patients whose inflammatory bowel disease or nonmalignant systemic disease is quiescent or well controlled appear to fare better than patients with active disease. Studies have also addressed the effect of radiation dose on occurrence of small bowel toxicity. Volume of the treatment field, volume of irradiated small bowel, total radiation dose, fraction size, treatment time, and treatment technique all influence small bowel tolerance. The TD5/5 for small volumes of small bowel has been estimated to be 50 Gy. Patients can generally receive 45 to 50 Gy in 1.8 to 2 Gy daily fractions to a pelvic field without a significant rate of toxicity.80 For postoperative patients, radiation to 45 to 50 Gy in five weeks is associated with an approximate 5% incidence of small bowel obstruction requiring surgery, whereas at doses greater than 50 Gy, the incidence rises to as high as 25% to 50%.70 Doses greater than 2 Gy per fraction in the postoperative setting also increase the risk of toxicity.81 At radiation doses of 70 Gy or greater, the incidence of toxicity rises precipitously.82 A study of different treatment techniques to minimize the effect of pelvic radiation on the small bowel showed that irradiating smaller volumes of bowel yielded less toxicity.83 In addition, treating patients in the prone position with external compression and bladder distention decreased side effects, likely from exclusion of portions of the small bowel from the radiation field. Another study treating postoperative patients with pelvic radiation therapy noted less small bowel toxicity by placing patients in the decubitus position.84 Studies have also analyzed dose-volume parameters associated with acute small bowel toxicity in patients undergoing treatment with 5-FU–based chemoradiation therapy for rectal cancer.85,86 These found strong correlations between acute toxicity and the amount of small bowel irradiated at each dose level analyzed. Another analysis evaluating rectal cancer patients treated preoperatively with chemoradiotherapy also showed a strong correlation between the occurrence of severe diarrhea and irradiated small bowel volume, surmising that limiting the volume of small bowel receiving greater than 15 Gy may significantly improve treatment tolerance.87 These and other studies imply that attention to detail in radiation planning with the use of modern treatment techniques are important considerations in patient treatment. The combination of radiation and chemotherapy (e.g., 5-FU) increases the risk of small bowel toxicity. In a

Chapter 39  Radiation Injury Table 39-2  Pathophysiologic Features of Patients with Late Radiation Enteropathy

Table 39-3  Therapeutic Options for Patients with Late Radiation Enteropathy

PATHOPHYSIOLOGIC FEATURE

UNDERLYING PATHOPHYSIOLOGIC FEATURE

CLINICAL SYMPTOM OR SIGN

Mucosal dysfunction

Lactose intolerance Vitamin B12 deficiency Steatorrhea

Stricture or blind loop syndrome with bacterial overgrowth

Diarrhea

Intestinal dysmotility

Bloating Constipation Diarrhea

Abnormal bile acid recirculation

Cholerrheic diarrhea

From Hauer-Jensen M, Wang J, Denham J. Bowel injury: Current and evolving management strategies. Semin Radiat Oncol 2003; 13:357-371.

randomized trial delivering 40 to 48 Gy using parallel, opposed fields to patients with rectal cancer, the incidence of severe small bowel complications was significantly higher in patients who received chemotherapy and radiation therapy than in patients who received radiation therapy alone. There were two treatment-related deaths (4%) in the combination therapy arm.88,89 In other trials in which multiple field radiation techniques were used, bolus 5-FU and radiation therapy showed no increase in chronic toxicity when chemotherapy and radiation therapy were combined. There was, however, a mild increase in acute diarrhea symptoms.83,90 The use of continuous infusional 5-FU as opposed to bolus 5-FU combined with radiation therapy also has been studied. Continuous 5-FU with radiation to 50.4 Gy in 1.8-Gy fractions was associated with more acute diarrhea, but no significant increase in chronic or severe small bowel toxicity as compared with bolus 5-FU therapy.91 Capecitabine, a prodrug that is converted to 5-FU in the tumor, also appears to enhance acute diarrhea rates when combined with radiation therapy.92 Although the addition of concurrent chemotherapy increases the acute toxicity of external radiotherapy (e.g., diarrhea, bowel frequency, cramping), its contribution to late bowel toxicity is poorly defined.17 There is investigation in integration of novel chemotherapeutic and “targeted” agents with radiation therapy in the neoadjuvant therapy of gastrointestinal cancers. Data from phase I and phase II trials using novel agents such as oxaliplatin, irinotecan, and epidermal growth factor receptor inhibitors suggest that the addition of these agents may significantly increase grades 3 and 4 gastrointestinal toxicity rates relative to conventional neoadjuvant chemoradiotherapy regimens, further emphasizing the importance of careful radiation planning to maximize normal tissue sparing in these patients.93-95 Diagnosis of chronic enteropathy is often a clinical one. The cause of symptoms can be variable from patient to patient. Therefore, individualization of diagnostic and therapeutic approaches is indicated. General diagnostic evaluation procedures and therapeutic options are displayed in Tables 39-2 and 39-3. Consultation with the treating radiation oncologist should be requested if the clinical presentation is consistent with radiation enteritis. Review of the patient’s previous radiation treatment record will reveal the total dose, fractionation, volume of treatment, and other radiation parameters. Analysis of the treatment plan may show areas of high dose, especially if the patient had an

THERAPEUTIC OPTIONS

Nutritional deficits

Correction of specific deficits Low-fat diet Lactose-free diet Elemental diet Total parenteral nutrition

Intestinal dysmotility (increased or decreased)

Loperamide Octreotide Prokinetic agents

Bile acid malabsorption

Bile-salt binding agent

Bacterial overgrowth

Antibiotics

From Hauer-Jensen M, Wang J, Denham J. Bowel injury: Current and evolving management strategies. Semin Radiat Oncol 2003; 13:357-371.

intracavitary implant or brachytherapy. Lesions encountered at endoscopy or x-ray studies are usually localized in the area of high dose. Ulceration of the mucosa, thickening of jejunal folds, and thickening of the intestinal loops are radiologic signs that suggest radiation damage to the small bowel (Fig. 39-5). Faster intestinal transit and reduced bile acid and lactose absorption are observed in patients with chronic radiation enteritis.96 These effects may be improved after the administration of loperamide. Antibiotics are indicated if there is small bowel bacterial overgrowth syndrome (see Chapter 102).97,98

TREATMENT AND PREVENTION

The management of acute radiation small bowel toxicity should be based on the severity of symptoms. Most cases of acute radiation enteritis are self-limited, requiring only supportive treatment. Diarrhea, nausea, vomiting, and abdominal cramping are treated symptomatically. Antidiarrhea medications such as loperamide, diphenoxylate with atropine, anticholinergic agents, and opiates can be used. Antiemetic agents may also be effective. A low-fat, lactose-free diet also may improve symptoms. A study of oral sulcralfate in patients receiving pelvic irradiation showed a decrease in frequency and improvement in consistency of bowel movements. In this study, not only were acute symptoms improved, but also chronic symptoms a year after completion of radiation seemed to be improved.99 Cholestyramine to treat diarrhea from bile acid malabsorption has shown some benefit,100 and treatment with anti-inflammatory agents has decreased some symptoms.101 Intractable diarrhea during the combined-modality treatment may require hospital admission for administration of parenteral fluids and electrolytes. Patients who are refractory to conventional anti­diarrhea medications may benefit from administration of a synthetic somatostatin analog such as octreotide. The management of chronic radiation enteritis remains a major challenge because of the progressive evolution of the pathophysiology, including obstructive endarteritis and fibrosis. In addition, there is variation in clinical manifestations of patients with chronic radiation enteropathy. The treatment should be as conservative as possible because of the diffuse nature of the process and the high morbidity associated with surgery. However, surgical intervention is indicated in intestinal obstruction, perforation, fistulas, and severe bleeding (discussed following). Chronic effects of diarrhea are managed symptomatically with low-residue diet. Fiber supplementation (e.g., Meta­ mucil, Citrucel) has shown benefit in some cases. In the

645

646

Section IV  Topics Involving Multiple Organs

A

B

C

Figure 39-5.  Radiographic evidence of radiation injury of the intestine. A, In early injury, edema may cause separation of intestinal loops, lead to thickening and straightening of mucosal folds, and impart a spiked appearance (arrows) to the small bowel mucosa. B, Severe radiologic abnormalities of the rectosigmoid colon are evident when this barium enema examination was performed 2 months after the patient underwent radiation therapy for cervical carcinoma. Subacute radiation injury of the colon may present radiologically as edematous, occasionally ulcerated mucosa, with asymmetrical areas of narrowing suggestive of Crohn’s colitis or recurrent tumor (arrows). C, Late radiation change in the colon with stricture formation (arrow) after a cumulative dose of approximately 55 Gy (5500 cGy).

rare setting of malnutrition related to chronic radiation injury, total parenteral nutrition (TPN) can improve clinical outcome, and methylprednisolone may add to the effects of TPN.72 However, the five-year survival rate for patients undergoing TPN ranges from 36% to 54%.62,102 It has been estimated that overall mortality rates associated with chronic radiation enteropathy are approximately 10%.103 Endoscopic techniques are sometimes required for diagnosis of bleeding small intestinal ulcers. Newer endoscopic techniques, including double-balloon enteroscopy and capsule endoscopy may help facilitate this diagnosis. Additionally, the double-balloon enteroscope method may allow therapeutic intervention in certain situations, including coagulation of small bowel telangiectasias. Significant bleeding refractory to endoscopic intervention may be managed surgically. Small bowel obstruction is generally managed conservatively with bowel rest and tube decompression. If the obstruction is severe or chronic, bowel resection or lysis of adhesions may be required, although the need for surgery is relatively uncommon. It is difficult to perform surgery for chronic radiation enteritis because of the diffuse fibrosis and alterations in the intestine and mesentery. The risk of anastomotic leak is high if the anastomosis is performed using irradiated bowel.103 The risk of leak can be lowered if at least one limb of the anastomosis is previously unirra­ diated.104 However, it may be difficult to distinguish between

the area of normal tissue and the irradiated portion of the intestines by gross examination during surgery, even when the fresh tissue is sent for frozen section. Another method the surgeon can use to circumvent this technical difficulty is to create the anastomosis with unirradiated colon. The accuracy in localizing injured bowel may be improved by intraoperative endoscopic examination, which can detect radiation-induced mucosal injury.105 Limited resection of the diseased intestine is the goal, but if the lesion is too diffuse, a bypass procedure may be attempted. If feasible, resection of the affected bowel results in a superior outcome than an enteric bypass procedure. However, extensive surgical resection of the diseased intestines may lead to short bowel syndrome (see Chapter 103) and increase the need for total parenteral nutrition. Because of the progressive evolution of the fibrosis, the patient may require additional surgery. Surgical bypass of the injured bowel may be associated with a blind loop syndrome, and the patient still may be at risk for perforation, bleeding, abscess, and fistulas due to the persistence of the affected bowel. Bypass procedures should be performed when resection is not possible or as a temporary management before resection at a later date. Surgery should be performed by an experienced team familiar with the management of radiation enteritis. Perforations and fistulae are best managed surgically. It should be noted that many patients with chronic small bowel radiation toxicity are nutritionally

Chapter 39  Radiation Injury depleted and are more susceptible to anastomotic leakage and dehiscence after surgery. The postoperative mortality of these patients may be significant and must be taken into consideration before a decision to proceed with surgery is made. A recent approach to treatment of chronic radiation enteritis is the application of hyperbaric oxygen.106,107 The rationale for hyperbaric oxygen is the creation of an oxygen gradient in hypoxic tissue that stimulates the formation of new blood vessels. Neoangiogenesis improves the blood supply and decreases the ischemia and necrosis responsible for severe complications. In a retrospective study of 36 patients with severe radiation enteritis refractory to medical management, improvement of clinical symptoms was reported in two thirds of the patients treated with hyperbaric oxygen.108 Hyperbaric oxygen may be helpful in management of bleeding due to chronic radiation enteritis in patients who are not controlled with conservative measures such as formalin and laser therapy.109,110 A large clinical series of 65 consecutive patients with chronic radiation enteritis (rectal and small bowel), primarily manifested as chronic bleeding, were treated with hyperbaric oxygen. Response rates for rectal and more proximal sites were 65% and 73%. The response rate for bleeding was 70% and for other symptoms (pain, diarrhea, weight loss, fistula, and obstruction) was 58%. The authors concluded that hyperbaric oxygen therapy resulted in healing or clinically significant improvement in two thirds of patients with chronic radiation enteritis.111 The optimal application of hyperbaric oxygen treatments in chronic enteritis remains a topic of ongoing investigation. Because chronic radiation enteritis is complex and rarely curable, prevention is key and measures to decrease its incidence are imperative. Pancreatic enzymes can exacerbate acute intestinal radiation toxicity,113 and reducing pancreatic secretion with a synthetic somatostatin receptor analog such as octreotide may reduce early and delayed radiation enteritis in animal studies.114 One of the major risk factors for injury is previous abdominopelvic surgery, which leads to the prolapse of the small intestines into the pelvis, exposing them to more radiation. Anticipation for the need of radiation and chemotherapy before or after surgery requires close collaboration among surgical, radiation, and medical oncologists. If gross residual tumor is found unexpectedly at surgery, outlining the tumor bed with surgical clips to facilitate postoperative treatment planning and surgical techniques to keep the small intestine outside the pelvis (e.g., omentoplasty or polyglycolic mesh) may significantly decrease the rate of complications (see following). Postoperative bowel adhesions may also increase the volume of bowel irradiated compared with normal small intestine, which is usually mobile and can move out of the radiation field. If radiation therapy is anticipated after surgery, attempts should be made at the time of surgery to displace the bowel outside the radiation field.115 One simple technique is the surgical placement of a polyglycolic, biodegradable mesh that moves the intestines out of the pelvis.116,117 This procedure has minimal morbidity and it does not significantly increase operating time. It also does not require a second operation to remove the mesh because it is absorbed three to four months after surgery. Magnetic resonance imaging (MRI) can be used after surgery to verify the position of the mesh, the small bowel, and eventual disappearance of the mesh. A reduction of 50% of the volume of the small bowel exposed to the radiation has been demonstrated with placement of a mesh during surgery, allowing a higher dose of radiation to be given postoperatively where indicated.118,119 Other techniques such as pelvic

reconstruction, omentoplasty, and transposition of the colon may also significantly decrease the volume of bowel exposed to radiation therapy.119-122 Radiation therapy technique is critical in reducing the rate of complications. The use of only anterior and posterior fields for pelvic radiation should be avoided if possible because of the high dose and large volume of bowel irradiated. A higher rate of operative mortality was reported in trials using this technique preoperatively for rectal cancers.123,124 As discussed, the toxicity of radiation therapy correlates with the volume of small bowel irradiated.125 In many patients, treatment in the prone position with a special “belly board” allows the displacement of the small intestines out of the radiation field.126,127 Patients should be instructed to maintain a full bladder during the radiation session, which further displaces the intestines out of the pelvis.74 Three-dimensional treatment planning optimizes the treatment technique by facilitating more accurate dose distributions. A three-dimensional treatment algorithm ensures the sparing of excessive radiation dose to normal tissues by the judicious use of multiple fields to the target volume from multiple geometries.128 In gynecologic brachytherapy, appropriate packing to displace the rectum and bladder away from the radioactive sources will decrease the risk of complications. In addition, newer treatment techniques such as intensity-modulated radiotherapy (IMRT) use sophisticated planning techniques to avoid critical structures (discussed following). Preventive therapeutic strategies also include investigation of antioxidants, free-radical scavengers and other cytoprotectant agents, cytokine modification, enterotrophic (growth-promoting) strategies, novel anti-inflammatory agents, modulators of intraluminal contents, modulators of endothelial dys­ function, as well neuroimmunomodulators to prevent this complication.103 Treatment of radiation enteritis is often only partially successful. Management is patient specific and should be as conservative as possible because of the relentless progression of the disease. Further understanding of the mechanism of fibrosis and the interaction of the molecular events controlling apoptosis and fibrosis may assist in the identification of the patient at risk for radiation complications and in the development of new therapeutic approaches. Careful radiation planning is critical in preventing acute and chronic radiation enteritis.

LARGE INTESTINE INCIDENCE AND CLINICAL FEATURES

Acute and chronic injury of the large intestine is similar to injury of the small intestine just discussed. Acutely, there is a decrease in the stem cell mitotic rate, resulting in a depletion of precursor cells required to replenish the epithelium as it normally sheds. Acute injury can be accom­ panied by superficial mucosal erosions and lamina proprial hemorrhage. There is also a thickening of the mucosa with proliferation of fibroblasts (Fig. 39-6).129 Late changes include vascular fibrosis with associated ischemia and formation of telangiectasias which can be a source of bleeding (Fig. 39-7). Late radiation bowel changes can lead to fluid and electrolyte malabsorption, obstruction, chronic proctitis, and fistula formation. Ischemic changes also include ulceration (Fig. 39-8), perforation, and fistulae.56 Bowel wall fibrosis may also occur, causing decreased motility, stricture and compliance.130 A decrease in rectal compliance may reduce the ability of the rectum to act as a reservoir, leading to fecal frequency, urgency, and incontinence.

647

648

Section IV  Topics Involving Multiple Organs

Figure 39-6.  Acute radiation injury to the rectum. Rectal mucosa with superficial mucosal erosion and focal lamina propria hemorrhage are seen. (Hematoxylin and eosin, ×100.) (Courtesy Dr. Robin Amirkahn, Dallas, Tex.)

Figure 39-7.  Typical findings of radiation colitis in a patient treated for prostate cancer. Top panels: Endoscopic view of the rectum reveals the characteristic fine tortuosity and curling of the new vessels. Lower panels: These demonstrate superficial burns from argon plasma coagulation, which was used to stop this patient’s bleeding. It is not necessary to ablate the lesions completely, but merely to cause mucosal and submucosal fibrosis, thereby entrapping the vessels in the scarring process. (Courtesy of Lawrence J. Brandt, MD, Bronx, New York.)

Acute colitis from radiation therapy manifests clinically as diarrhea, cramping, tenesmus, urgency, incontinence, and less commonly as mucoid or bloody rectal discharge. These symptoms can result from rectal inflammation, edema, and spasm. Symptoms often begin two to three weeks into treatment, and usually resolve within several weeks to three months following radiation completion. A relationship between the incidence of acute and chronic radiation injury is uncertain.131,132 Chronic changes appear within six months to two years and beyond following completion of radiation therapy, with symptoms similar to acute

Figure 39-8.  Rectal mucosa after radiotherapy. Residual but malformed rectal crypts and flat, regenerating surface mucosa in the region of a radiation-induced rectal ulcer are seen. Note fibrosis and inflammation of the lamina propria. (Hematoxylin and eosin, ×100.) (Courtesy Dr. Robin Amirkahn, Dallas, Tex.)

injury. Patients may present with tenesmus, bleeding, lowvolume diarrhea, rectal pain, and occasionally low-grade obstruction or fistulae.133 Patients can develop a pancolitis that mimics inflammatory bowel disease. It should be remembered that pelvic irradiation is a risk factor for development of rectal cancer.134 The large intestine is less radiosensitive than the small intestine. This may be partially explained by the fact that higher doses of radiation are often delivered to smaller volumes of the rectum compared with small intestine (i.e., focal “collateral” rectal irradiation in prostate cancer therapy, focal collateral rectal irradiation in high-dose gynecologic malignancy implants, and so on). The rectum is also a readily accessible organ by endoscopy, allowing early diagnosis as well as intervention, possibly preventing symptomatic progression. Series have reported the risk of serious late rectal complications is 5% or less when less than 80 Gy is delivered.17 Radiation injury of the large intestine occurs most frequently in the rectum due to its location adjacent to the prostate, bladder, cervix, uterus, and ovaries, exposing it to collateral radiation dose with treatment of these organs.15 Acute rectal injury is often self-limited, but the incidence of chronic radiation proctitis is increasing with increased use of pelvic radiation therapy and radiation dose escalation.135,136 As is true with the other sites, the incidence of large bowel toxicity is associated with radiation dose, volumes treated, and the use of concurrent chemotherapy. In an early study of radiation therapy for testicular cancer, radiation injury to the colon occurred in 18% of patients after 45 Gy and in 37% of patients after 60 to 64 Gy.137 The treatment of rectal cancer commonly uses doses of 45 to 54 Gy, whereas treatment of prostate and cervical cancer uses higher doses, ranging from 60 to 80 Gy. The incidence of severe rectosigmoid toxicity in cervical cancer patients was 4% or less for patients receiving doses below 80 Gy, and 13% for doses greater than 95 Gy.138 The treatment of prostate cancer with doses of 60 to 70 Gy has been associated with an incidence of severe proctitis below 8%.139 Radiation doses of 60 to 70 Gy for anal cancer yield an incidence of severe rectal toxicity of 5% or less.140-142 Treatment using conformal radiation, three-field, and four-field techniques further decreases the risk of rectal toxicity.138,143 A trial of

Chapter 39  Radiation Injury conformal versus conventional radiation for prostate cancer reported less radiation proctitis (5% versus 15%, respectively).144 The use of IMRT (discussed later) may further improve this rate, as demonstrated by a recent series treating of prostate cancer patients to doses greater than 80 Gy. Of 772 patients, 4.5% developed acute grade 2 rectal symptoms and no patient developed acute grade 3 or above rectal symptoms. Eleven patients (1.5%) developed late grade 2 rectal bleeding and four patients (0.1%) experienced late grade 3 rectal toxicity requiring either transfusion or laser cauterization. No late grade 4 rectal complications were observed.145 Combining chemotherapy with radiation therapy increases toxicity rates. A study of 5-FU plus cisplatin and radiation therapy for cervical cancer showed a combined incidence of severe small and large bowel toxicity of 18%.146 In contrast, a combination of 5-FU and mitomycin C with radiation doses of 40 to 55 Gy in the treatment of anal cancer was associated with a less than 5% risk of severe rectal complications.147 Multiple trials have combined 5-FU– based chemotherapy and radiation therapy as neoadjuvant and adjuvant treatment for rectal cancer.88,91,148,149 The toxic effect of combined chemotherapy and radiation has varied, from no significant increase in toxicity to a 24% incidence of severe diarrhea and a 25% incidence of chronic bowel injury.150 Because of the increase in toxicity seen with single or opposed-only radiation fields, the use of conformal and multifield techniques is necessary when using combination therapy. The increasing use of neoadjuvant chemoradiotherapy has also raised the concern of increased postoperative complications in these patients. However, a large randomized trial showed a significant reduction in the rates of acute and chronic gastrointestinal toxicity in patients treated neoadjuvantly versus patients treated adjuvantly.149 In contrast to small bowel injury, previous abdominopelvic surgery does not appear to predispose the rectum to radiation injury, likely due to the fact it is not otherwise mobile. However, because of the similarity of vascular changes seen with small bowel radiation injury, a history of diabetes, hypertension, cardiovascular disease, or peripheral vascular disease may predispose large bowel to radiation toxicity.76-79 Patients with collagen vascular disease and inflammatory bowel disease also have an increased propensity for large bowel radiation toxicity (as discussed in the section on small bowel toxicity).

TREATMENT AND PREVENTION

Management of large bowel radiation toxicity is based on symptom control. Acute toxicity is treated with antimotility agents, such as loperamide or diphenoxylate with atropine, and a low-residue diet. Opiates and anticholinergics may also be of benefit. Glucocorticoid-containing suppositories may be helpful in the management of patients with anorectal inflammation. Colonoscopy should be avoided if possible because of the potential risk of perforation associated with the friability of the rectal mucosa during radiation.151 For chronic diarrhea due to decreased rectal compliance, stool softeners or fiber supplements can help. As in acute proctitis, steroid suppositories may be beneficial. The benefit of steroid retention enemas is unclear.152 Sucralfate enemas, also controversial, may be used for ulcerative lesions. Short-chain fatty acids, which nourish the colonic mucosa, have been studied in acute and chronic radiation proctitis. Initial relief of symptoms can be seen, but symptoms recur shortly after stopping treatment.153 Hyperbaric oxygen has been used to stabilize bleeding related to telangiectasias, but this treatment is not widely available and requires many treatments before any effect is seen.154,155

Nonetheless, a randomized trial in patients with refractory chronic radiation proctitis reported that hyperbaric oxygen therapy significantly improved healing.156 Treatment of colorectal ulcerations associated with bleeding is initially endoscopic, with the use of coagulation techniques, such as argon plasma coagulation. Bleeding due to radiation proctopathy is usually minor and often controlled endoscopically with conservative measures such as cauterization of the telangiectasias with laser treatment (see Fig. 39-7),157-160 or by application of formalin.161-164 Sucralfate enemas may alleviate radiation proctopathy by forming a protective complex with the rectal mucosa. It also increases the local levels of fibroblast growth factors and prostag­ landins. Sucralfate enemas appear to be helpful in chronic proctopathy but their benefit is unclear during the acute period.165-167 Short-chain fatty acid enemas may be also helpful for management of chronic hemorrhagic radiation proctopathy by inhibiting the inflammatory response including the NF-kB pathway168,169 (see Chapter 2). Strictures can also be endoscopically dilated. For patients who have refractory bleeding, stricture, perforation, or fistulae, surgical management may (rarely) be necessary. Management of a pelvic fistula (e.g., vaginal or bladder fistula) is complex and requires fecal diversion before the corrective surgery. A thorough radiographic investigation such as barium enema, small bowel follow-through, or enteroclysis to delineate the extent of the fistula should be performed before surgery. Patients with fistulas may present with additional challenges such as electrolyte imbalance, malnutrition, and infections. Many surgical techniques have been described to repair fistulas, but the corrective surgery is best done when the patient is medically stable and enough time has elapsed after the surgical diversion. This allows healing and decreased inflammation of the affected tissues.170,171 Prevention of large bowel toxicity from radiation has been studied. Prostaglandins have been investigated as a potential radioprotector. Prostaglandin E2 and prostaglandin analogs display radiation protection in animal studies.172-175 Misoprostol suppositories also have been shown to reduce symptoms of acute radiation enteritis in patients undergoing radiation therapy for prostate cancer.176 However, a randomized placebo-controlled trial from Germany in patients with prostate cancer undergoing irradiation found that significantly more patients experienced rectal bleeding in the misoprostol group.177 Amifostine is a sulfhydryl compound that is converted intracellularly to an active metabolite, which in turn binds to free radicals and protects the cell from radiation injury.178 Amifostine has been investigated for the prevention of chronic radiation enteritis. Amifostine has demonstrated protection of the small as well as the large intestines in preclinical studies.179 Amifostine has also been shown to reduce the incidence of early and delayed radiotherapeutic injuries at several anatomic sites. In one randomized study the late effects of radiation were significantly reduced in the group receiving parenterally administered amifostine.180 However, the median follow-up was quite short (24 months), and longer follow-up is necessary to confirm the benefits of the medication given the incidence of late complications increases with time. Another randomized trial evaluated 205 patients with pelvic malignancies who received radiation therapy, alone or with intravenous amifostine. Patients receiving amifostine experienced a significantly lower incidence of grades 2 and 3 bladder and lower gastrointestinal tract toxicity, with no significant difference between the two groups in tumor response to treatment.181 There is also evidence to suggest that intrarectal application of amifostine

649

650

Section IV  Topics Involving Multiple Organs may reduce the risk of proctitis in patients undergoing radiotherapy for prostate cancer.182 A preclinical study showed a possible role for anti– TGF-β1 interventions to reduce delayed radiation fibrosis and enteropathy.183 Many special diets and nutrients such as fiber, elemental diets, short-chain fatty acids, and amino acids such as glutamine may reduce radiation toxicity to the intestine. However, consistent clinical results were not observed.184-189 Preventive therapy must have high efficacy, low toxicity, low cost, and not protect the tumor from radiation therapy. Unfortunately, no currently available therapy fulfills all of these objectives. As described previously, careful radiation planning and delivery are of paramount importance.

ANUS

size greater than 2 Gy results in a high incidence of anal toxicity.194 Patients with HIV and anal cancer who are treated with combined chemotherapy and radiation therapy have an increased risk of acute and late anal toxicity.204

TREATMENT

Treatment of acute toxicity is primarily supportive, including proper skin care, dietary modifications, pain medications and topical steroid medications, with breaks in radiation treatment if severe. The effects are self-limited, and usually resolve within weeks of conclusion of therapy. Treatment for chronic toxicity with anal stricture and stenosis includes sphincter dilatation. Rare patients can require colostomy for severe symptoms. Small studies of hyperbaric oxygen therapy have shown efficacy in treating chronic anorectal ulcers.205 There is also a report of oral vitamin A therapy as a treatment for anorectal ulceration, but confirmatory studies are lacking.191

INCIDENCE AND CLINICAL FEATURES

The anal canal is typically spared from significant radiation exposure except in irradiation for anal, low rectal, and gynecologic cancers. The primary acute toxicity from anal cancer irradiation is diarrhea from the exposure of the large bowel to irradiation. However, damage to the anus itself can occur in the form of acute desquamation or ulceration, with later development of ulcers, strictures, anorectal fistulae, and incontinence.190 The primary data on anal toxicity from radiation therapy come from studies using radiation or chemoradiotherapy for the treatment of anal cancer itself. Anal toxicity manifests as mucosal edema and friability.191 These changes are often exacerbated by diarrhea that occurs from rectal toxicity. Chronically, anal fibrotic changes may evolve. Clinically, anal toxicity presents initially as a perianal skin reaction that ranges from minimal skin changes to moist desquamation and erythema, as well as diarrhea. These changes are self-limited and usually resolve within a few weeks of completion of therapy. Acute toxicity can lead to an interruption of therapy, although this may be less common with modern radiation treatment techniques.147,191,192 The incidence of acute toxicity is high, and is increased with concurrent chemotherapy delivery or use of a large dose per fraction.142,193-195 Phase III studies and series of patients treated with combined chemotherapy and radiation therapy have noted an incidence of skin toxicity of grade 3 or above in 26% to 78% of patients using doses of 45 to 60 Gy in 1.8- to 2.25-Gy fractions.147,190,195-199 A study using conformal radiation techniques (50.4 Gy in 1.8-Gy fractions) combined with chemotherapy reported grade 3 (or higher) acute skin toxicity in only 20% of patients.200 In a recent multi-institutional experience of anal cancer patients treated with IMRT-based chemoradiotherapy, grade 3 skin toxicity was seen in 38% of patients with no grade 4 toxicity observed, comparing favorably to the results of previous randomized trials.192 Further studies of advanced radiation therapy techniques are needed to confirm this improvement. Late anal toxicity occurs within months to years following completion of therapy. The most common late complication is anorectal ulceration. Patients also may develop anal stricture or stenosis, incontinence, anal pain, or anorectal fistulae.190,194,195,201 There does not appear to be an increase in the occurrence of chronic anal toxicity with the addition of chemotherapy to radiation therapy.147,199,201 Doses of 45 to 60 Gy in fractions of 1.8 to 2 Gy are considered safe, resulting in chronic grade 3 (or higher) toxicity rates of zero to 22%.140,141,193,195,202,203 Doses greater than 65 Gy or fraction

LIVER INCIDENCE AND CLINICAL FEATURES

Radiation-induced liver disease (RILD) is seen in approximately 5% of patients when the whole liver radiation dose reaches 30 to 35 Gy at 2 Gy per fraction.206,207 The pathologic lesion in RILD is central vein thrombosis at the lobular level (veno-occlusive disease), which results in marked sinusoidal congestion leading to lobular hemorrhage and secondary injury to surrounding hepatocytes.208 Fibrin deposition in the central veins is thought to be the cause of the venoocclusive injury. It is unknown what stimulates the fibrin deposition, but there are hypotheses that suggest that TGF-β is increased in the setting of exposure to radiation, and this in turn stimulates fibroblast migration to the site of injury, causing fibrin and collagen deposition. Foci of necrosis are found in the affected portion of the lobules.209 Severe acute hepatic toxicity may progress to fibrosis, cirrhosis, and liver failure. Radiation-induced liver disease is a clinical syndrome consisting of anicteric hepatomegaly, ascites, and elevated liver enzymes, particularly serum alkaline phosphatase, which may be elevated out of proportion to serum aminotransferase activity or serum bilirubin. RILD occurs typically between two weeks to four months after completion of radiation therapy. Patients note fatigue, weight gain, increased abdominal girth, and occasionally right upper quadrant pain. Abdominal imaging with computed tomography (CT) scan or MRI can be used in diagnosis. RILD can progress to a chronic phase in which patients can develop increasing fibrosis and liver failure. Recent studies have emphasized the effect of the liver volume irradiated in addition to dose.208 Although radiation hepatopathy can occur after doses of 35 to 40 Gy to the entire liver, significantly higher doses can be given with few clinical complications if sufficient normal liver is spared. Studies by Lawrence and colleagues report that if less than 25% of the normal liver is treated with radiation therapy there may be no upper limit on dose associated with radiation hepatopathy.208 Estimates of the hepatic irradiation doses associated with a 5% risk of RILD for uniform irradiation of one third, two thirds, and the whole liver are 90 Gy, 47 Gy, and 31 Gy, respectively. Combining chemotherapy and radiation can increase liver damage, particularly if the chemotherapeutic agents are hepatotoxic. Chlorambucil, busulfan, and platinum drugs are used with radiation in bone marrow transplantation and are hepatotoxic agents.

Chapter 39  Radiation Injury In contrast, fluoropyrimidines do not seem to increase radiation-related hepatotoxicity.207,210

TREATMENT

RILD is often fatal once it occurs because there are no effective treatments.

THERAPEUTIC TECHNIQUES TO REDUCE TOXICITY Gastrointestinal toxicity is a significant obstacle in the radiotherapeutic management of many malignancies, resulting in patient morbidity and impeding tumor control by limiting the timely delivery (and intensity) of radiation dose. Avoidance of normal tissue with delivery of therapeutic target doses is a primary goal of the radiation oncologist. As discussed, different techniques may be implemented to decrease volumes of nontarget gastrointestinal tissues treated, including the use of multiple treatment fields to avoid “hot-spots,” treating in the prone position, use of a belly board or false table-top, as well as treating the patient with a full bladder to displace bowel out of the radiation field. In the past, radiation therapy plans were based on twodimensional (2D) planning, in which treatment fields were defined using x-ray images and known anatomic landmarks. With improvements in imaging and computer capabilities, three-dimensional (3D) treatment planning became available in the 1980s. An advanced form of 3D planning, intensity-modulated radiation therapy (IMRT) has now been implemented in clinical practice.211,212 IMRT is a potentially significant advance in achieving these goals. As opposed to conventional “static” fields, IMRT uses the principle of multiple “fields-within-fields” that more accurately conform radiation dose to target tissues while sparing normal structures. IMRT requires target tissues and normal organs are accurately defined. Dose constraints are assigned to these organs along with a desired (prescription) dose to the target volume(s). “Inverse planning,” whereby computer search algorithms establish multiple (and sometimes unconventional) beam or field designs is then performed, attempting to meet the prescribed target dose and normal tissue dose constraints. Individual fields are treated with multiple, small “beamlets” rather than one uniform beam, and each beam delivers a different dose (intensity) to the different parts of the target. This allows close conformation of radiation dose to the shape of the target and preferential sparing of normal surrounding organs from the high-dose areas. Collectively, early clinical results in varying cancers using IMRT-based chemoradiotherapy have shown significant decreases in treatment-related toxicities, with cancerrelated outcomes similar to conventional radiotherapy approaches. For example, Mundt and associates showed a marked improvement in small bowel dosimetry for patients with gynecologic malignancies treated with IMRT compared with conventional 3D planning. An experience of 36 patients with gynecologic malignancies treated with intensity-modulated whole pelvic radiotherapy were compared with outcomes of 30 patients treated at the same institution with 3D conformal radiotherapy. Patients were well matched with respect to demographic and treatment factors. Signifi-

cantly lower rates of chronic gastrointestinal toxicity were seen in the IMRT group, with only 11% of women treated with IMRT experiencing grades 1 to 3 toxicity (0% grade 3) versus 50% in the non-IMRT group.213 In a different series, Salama and colleagues reported on 53 patients with anal carcinoma treated with IMRT-based chemoradiotherapy. The median radiation doses to the pelvis and the primary disease were 45 and 52 Gy, res­pectively. Fifteen percent of patients experienced acute grade 3 gastrointestinal toxicity with no grade 4 toxicity observed, comparing favorably to observed rates of severe gastrointestinal toxicity in contemporary trials using conventional radiation planning.192 This is especially notable given the significantly higher pelvic doses delivered in patients receiving IMRT in this series. These techniques will require further demonstration of meaningful clinical benefit in patient outcomes to further solidify their routine use in clinical practice, with cooperative group trials underway.

KEY REFERENCES

Ahn SJ, Kahn D, Zhou S, et al. Dosimetric and clinical predictors for radiation-induced esophageal injury. Int J Radiat Oncol Biol Phys 2005; 61:335-47. (Ref 29.) Clark R, Tenorio L, Husse J, et al. Hyperbaric oxygen treatment for chronic refractory radiation proctitis: A randomized and controlled double blind cross-over trial with long-term follow-up. Int J Radiat Oncol Biol Phys 2008. March 12 [Epub ahead of print]. PMID 18342453. (Ref 156.) Coia LR, Myerson RJ, Tepper JE. Late effects of radiation therapy on the gastrointestinal tract. Int J Radiat Oncol Biol Phys 1995; 31:1213-36. (Ref 17.) Dawson LA, Ten Haken RK, Lawrence T. Partial irradiation of the liver. Semin Radiat Oncol 2001; 15:240-6. (Ref 208.) Gunnlaugsson A, Kjellen E, Nilsson P, et al. Dose-volume relationships between enteritis and irradiated bowel volumes during 5-fluorouracil and oxaliplatin based chemoradiotherapy in locally advanced rectal cancer. Acta Oncol 2007; 46:937-44. (Ref 87.) Hauer-Jensen M, Wang J, Denham JW. Bowel injury: Current and evolving management strategies. Semin Radiat Oncol 2003; 13:357-71. (Ref 103.) Hille A, Schmidberger H, Hermann RM, et al. A phase III randomized, placebo-controlled, double-blind study of misoprostol rectal suppositories to prevent acute radiation proctitis in patients with prostate cancer. Int J Radiat Oncol Biol Phys 2005; 63:1488-93. (Ref 177.) Horiot JC, Aapro M. Treatment implications for radiation-induced nausea and vomiting in specific patient groups. Eur J Cancer 2004; 40:979-87. (Ref 47.) Johnston M, Robertson G, Frizelle F, et al. Management of late complications of pelvic radiation in the rectum and anus. Dis Colon Rectum 2003; 46:247-59. (Ref 136.) Marshall GT, Thirlby RC, Bredfeldt JE, et al. Treatment of gastrointestinal radiation injury with hyperbaric oxygen. Undersea Hyperb Med 2007; 34:35-42. (Ref 112.) Movsas B, Scott C, Langer C, et al. Randomized trial of amifostine in locally advanced non-small-cell lung cancer patients receiving chemotherapy and hyperfractionated radiation: Radiation therapy oncology group trial 98-01. J Clin Oncol 2005; 23:2145-54. (Ref 33.) O’Rourke IC, Tiver K, Bull C, et al. Swallowing performance after radiation therapy for carcinoma of the esophagus. Cancer 1988; 61:2022-6. (Ref 16.) Willett CG, Ooi CJ, Zietman AL, et al. Acute and late toxicity of patients with inflammatory bowel disease undergoing irradiation for abdominal and pelvic neoplasms. Int J Radiat Oncol Biol Phys 2000; 46:9958. (Ref 79.) Wong RK, Paul N, Ding K, et al. 5-Hhydroxytryptamine-3 receptor antagonist with or without short-course dexamethasone in the prophylaxis of radiation induced emesis: A placebo-controlled randomized trial of the National Cancer Institute of Canada Clinical Trials Group (SC19). J Clin Oncol 2006; 24:3458-64. (Ref 48.) Full references for this chapter can be found on www.expertconsult.com.

651

CHAPTE R

40 Complications of Gastrointestinal Endoscopy Michael B. Kimmey

CHAPTER OUTLINE Preparation of the Patient for Endoscopy  653 History and Physical Examination  653 Antibiotic Prophylaxis  654 Management of Anticoagulant and Antiplatelet Drugs  654 Informed Consent  654 Complications of Sedation  654 Infectious Complications  655 Other General Complications  656 Electrosurgery  656 Miscellaneous Complications  656 Timing and Severity of Complications  656 Medicolegal Considerations  656 Complications of Upper Endoscopy  657 Respiratory Problems  657 Complications of Topical Anesthesia  657 Hemorrhage  657 Perforation  657 Complications of Dilation  657 Complications of Endoscopic Hemostasis  657

Gastrointestinal (GI) endoscopy plays an important role in the management of patients with GI disorders. These proce­ dures are used to diagnose GI diseases, obtain tissue to confirm a diagnosis, and provide specific therapy. All procedures, whether medical or surgical, have inher­ ent risks. The risk of the procedure must always be weighed against its anticipated benefit. Patients as well as physicians must assume that there is some risk involved when they undertake any procedure including GI endoscopy. Compli­ cations of GI endoscopy are sometimes unavoidable; none­ theless, strict attention to detail and knowledge of potential complications and their risk factors can minimize their occurrence. Preparation of the patient and planning of the endoscopic procedure can reduce the risks of the procedure. This chapter first reviews preparation of the patient for endos­ copy, and then complications related to sedation for pro­ cedures, as well as other general complications common to all endoscopic procedures. Complications of specific endoscopic procedures are then detailed. Whenever pos­ sible, discussion focuses on how complications can be avoided. The incidence of specific endoscopic complications is difficult to assess and depends on a number of variables that are not always comparable across studies. Reports are often retrospective, which may underestimate the actual number

Complications of Enteral Access Procedures  658 Complications of Other Therapeutic Procedures  658 Complications of Small Bowel Endoscopy  658 Double Balloon Enteroscopy  658 Capsule Endoscopy  658 Complications of Colonoscopy and Sigmoidoscopy  658 Hemorrhage  658 Perforation  659 Postpolypectomy Coagulation Syndrome  660 Complications Related to Colon Preparation  660 Other Complications  660 Complications of ERCP  660 Pancreatitis  660 Hemorrhage  660 Perforation  661 Cholangitis  661 Other Complications  661 Complications of Endoscopic Ultrasonography  661 Complications of Fine-Needle Aspiration  661

of complications compared with prospective studies. Imme­ diate complications are more easily assessed than delayed complications, which may be undetected, unrecognized, or unreported. Variations in patient populations and disease severity can also affect reported complication rates. In rec­ ognition of these limitations, this chapter includes only an estimate of average rates of specific complications.

PREPARATION OF THE PATIENT FOR ENDOSCOPY HISTORY AND PHYSICAL EXAMINATION

A medical history should be performed prior to endoscopy. Experience with endoscopic procedures and sedation should be discussed, with special attention to intolerance or an adverse outcome of a previous procedure. A review of chronic medications and alcohol use may predict intoler­ ance to sedation or the need for larger than usual doses of sedatives or the use of additional drugs to augment sedation. A limited physical examination should be performed with examination of the heart, lungs, and abdomen and an assess­ ment of the patient’s airway. Preprocedure assignment of an American Society of Anesthesiology (ASA) five-category physical status classification is also required by hospital

653

654

Section IV  Topics Involving Multiple Organs Table 40-1  Endoscopic Procedures for Which Antibiotic Prophylaxis Is Recommended* Upper Endoscopy PEG placement Control of variceal bleeding† ERCP Bile duct obstruction when drainage may be incomplete (e.g., intrahepatic biliary strictures) Liver transplant recipient with biliary obstruction Pancreatic pseudocyst EUS-FNA Pancreatic cysts Mediastinal cysts Intramural cysts *Antibiotics are not recommended for prevention of endocarditis and in patients with vascular grafts or prosthetic joints. † Antibiotics recommended at time of hospital admission for all cirrhotic patients with gastrointestinal bleeding. ERCP, endoscopic retrograde cholangiopancreatography; EUS-FNA, endoscopic ultrasonography–fine-needle aspiration; PEG, percutaneous endoscopic gastrostomy.

endoscopy departments and most accredited ambulatory endoscopy centers.

ANTIBIOTIC PROPHYLAXIS

A recognition of the increased risk for infections associated with certain endoscopic procedures and patient risk factors allows the use of prophylactic antibiotics in select situa­ tions (Table 40-1).1 These antibiotics are intended to prevent local complications such as cellulitis around a percutane­ ous endoscopic gastrostomy (PEG) tube site following its placement or infection of a pancreatic cyst following endoscopic ultrasonography (EUS)–guided fine-needle aspiration (FNA). Other infections may be the consequence of procedure-related bacteremia. Whereas patients with cardiac valve abnormalities may be at an increased risk for endocarditis from bacteremia, the role of prophylactic antibiotics for these patients has recently been reevaluated. The risk of endocarditis following endoscopy is very low and is likely less than the risk of endocarditis from ordinary daily activities such as teeth brushing. For these reasons, the American Society for Gastrointestinal Endo­ scopy and the American Heart Association no longer recommend routine use of prophylactic antibiotics prior to endoscopy.1,2

MANAGEMENT OF ANTICOAGULANT AND ANTIPLATELET DRUGS

Patients taking warfarin are at increased risks for bleeding following polypectomy, endoscopic sphincterotomy, balloon dilation, percutaneous gastrostomy, and EUS-FNA aspiration.3 Warfarin should be held before these procedures so that the prothrombin time (or international nor­malized ratio, INR) can return to normal and be restarted within 1 week after the procedure. In patients with mech­anical heart valves and other situations at high risk for thromboembo­ lism, use of unfractionated or low-molecular-weight heparin should be substituted for warfarin except during the 12 hours before and after the procedure.3-5 Drugs that affect platelet function such as aspirin, nonste­ roidal anti-inflammatory drugs, ticlopidine, clopidogrel, and other newer agents have not been clearly shown to increase bleeding complications from endoscopic proce­ dures.3,5,6 Studies have been underpowered to detect small

but potentially clinically significant effects on bleeding complication rates, however. The risk of bleeding following a specific procedure should be weighed against the risk of thromboembolic complications if antiplatelet therapy is discontinued in an individual patient.

INFORMED CONSENT

Written informed consent should be obtained before the performance of any endoscopic procedure.7 This consent should be obtained by the endoscopist personally and wit­ nessed by another health care worker or family member whenever possible. The components of the informed consent process include a discussion of the benefits and alternatives to the procedure, as well as a discussion of the known risks of the procedure. The severity and frequency of complica­ tions influence the informed consent discussion; however, patient perception of risk is highly variable.8 The benefits and risks of sedation should always be included in the informed consent process if sedation is used. Written infor­ mation about procedure-specific risks should be provided to the patient in advance of the procedure whenever possible.

COMPLICATIONS OF SEDATION Sedation is used for most endoscopic procedures to allow a more comfortable experience for the patient and in some cases to allow a calm and still working environment for the endoscopist. Moderate (or conscious) sedation is used most commonly for GI endoscopy. By using a combination of a benzodiazepine and narcotic administered intravenously, the patient can be monitored by an assistant who is perform­ ing interruptible tasks.9 Deep sedation, usually achieved through the use of intravenous propofol, is being increas­ ingly used in the United States, but risks are higher than with conscious sedation, and special monitoring and train­ ing are advisable. According to the ASA, deep sedation can be administered by non-anesthesiologists, but personnel who can rescue the patient from general anesthesia should be present.10 Several large series have demonstrated that propofol can be administered safely under the gastroenterologist’s supervision; however, additional training and monitoring with capnography are strongly recommended.11-14 Recovery times following propofol use are shorter than those follow­ ing traditional conscious sedation with midazolam and fentanyl.15 Although most endoscopy can be performed safely under conscious sedation administered or supervised by the endoscopist, there are some situations in which having the assistance of an anesthesiologist or nurse anesthetist to administer deep sedation or general anesthesia can reduce the risk to the patient.16 Patients with a history of being dif­ ficult to sedate may benefit from deep sedation. This often includes alcoholic patients and those who are on high doses of narcotics. Patients with hemodynamic instability and respiratory compromise should also have special monitor­ ing of sedative drug use. Cardiorespiratory complications, usually attributed to sedation, are the most common complications of GI endos­ copy. Survey data suggest that approximately half of endo­ scopic complications are in this category.17,18 The reported frequency of cardiac and respiratory complications of endoscopy is between 2 and 5 per 1000 procedures with approximately 10% of these complications resulting in death.19 Cardiopulmonary complications as high as 11 per

Chapter 40  Complications of Gastrointestinal Endoscopy Table 40-2  Side Effects of Medications Used for Sedation, Analgesia, and Reversal AGENT

COMMON SIDE EFFECTS OF CLASS

Benzodiazepines

Respiratory depression, hypotension, cardiac arrhythmias, headache, confusion, nystagmus

Diazepam Midazolam Narcotics (Opiates) Meperidine Fentanyl Topical Anesthetics Lidocaine Benzocaine Miscellaneous Agents Propofol

AGENT-SPECIFIC SIDE EFFECTS

Respiratory depression, hypotension, urinary retention

Phlebitis and thrombosis at IV site Amnesia Myoclonus, seizures, nausea and vomiting

Hypersensitivity reactions, methemoglobinemia

Droperidol Diphenhydramine Promethazine Reversal Agents Flumazenil* Naloxone†

Respiratory depression and arrest, hypotension, bradycardia, hyperlipidemia Sedation, extrapyramidal effects, prolonged QT interval, cardiac arrest Sedation, nausea, dry mouth Sedation, orthostatic hypotension, extrapyramidal effects, hemolytic anemia Vasodilation, headache, seizures Hypotension, cardiac arrhythmias, narcotic withdrawal

*Reverses effects of benzodiazepines. † Reverses effects of opiate narcotics. IV, intravenous.

1000 procedures have been reported with propofolmediated sedation.20 Respiratory complications of sedation are most com­ monly due to hypoventilation. Combinations of benzodiaz­ epine and narcotics are known to produce more respiratory depression than use of either agent alone. The routine use of pulse oximetry allows more judicious titration of sedative medications but does not detect significant hypercarbia.21 The latter can be detected with capnography,22 but this is still not used routinely in most endoscopic facilities. Airway assessment is also important before endoscopy for the safety of the upper endoscopic procedure and for the ability to provide respiratory support should hypoventilation occur. Risk factors for airway compromise include diffi­ culties with previous anesthesia or sedation, obesity, a small mouth or lower jaw, and a history of stridor or sleep apnea.9,10 Hypotension is also usually due to medications. Narcotics in particular cause peripheral venous dilation and reduced cardiac preload, which in the fasting volume-depleted patient can lead to significant hypotension. This problem is usually responsive to intravenous fluid boluses, reason enough to require intravenous access during endoscopy done with sedation.9 Vasovagal reactions are the most common cause of cardiac arrhythmia during endoscopy. These reactions have been reported to occur in 16% of colonoscopies but can also occur with endoscopy of the upper GI tract.23,24 Reducing painful stimuli and suctioning air from the bowel are gener­ ally sufficient to reverse the vagally mediated bradycardia and hypotension; however, reversal with atropine is required for persistent bradycardia with hypotension in approxi­ mately one third of cases.23 Self-limited ventricular arrhyth­ mias may be seen in up to 20% of older adult patients undergoing upper endoscopy, particularly if they have elec­ trocardiographic changes suggesting ischemia.25 Careful monitoring of the patient during endoscopy helps detect cardiorespiratory complications at an early stage so

that specific action can be taken. Observation of the patient by a qualified assistant who is not actively involved in performing the procedure can detect apnea and loss of consciousness.9 Intermittent blood pressure readings and continuous pulse oximetry are useful adjuncts and are rec­ ommended for patients receiving sedation.9,26 Although con­ tinuous monitoring of the electrocardiogram is also advisable for patients with a history of cardiac arrhythmias, routine monitoring of the electrocardiogram is not required for all patients. The endoscopist should be familiar with specific side effects of the medications used for sedation during endos­ copy (Table 40-2). In addition to commonly used benzodi­ azepines and narcotics, other drugs are sometimes used to augment sedation. Reversal agents for benzodiazepines (flu­ mazenil) and narcotics (naloxone) are useful agents when oversedation occurs. Patients should be observed in the endoscopy unit follow­ ing a procedure until they are conscious and their vital signs have returned to baseline.9,10 Scales are available to assist staff in objectively quantifying discharge criteria.27 Because sedating medications can have subtle effects on higher level mental functions for hours after administration, it is advis­ able to have the patient accompanied by another individual on discharge and to recommend that the patient not drive or operate machinery until the day following the procedure.9

INFECTIOUS COMPLICATIONS Infections such as with Pseudomonas species or hepatitis C can be introduced by the endoscope, and are not specific to the procedure type or the patient’s underlying disease. Whereas these exogenous infections are infrequent, they merit discussion because they are usually avoidable.28 Current endoscopes are highly specialized and expensive

655

656

Section IV  Topics Involving Multiple Organs pieces of equipment that are designed to be reused. Because the GI tract is not sterile, high-level disinfection between uses is deemed to be sufficient for preventing transmission of infectious organisms between patients.29,30 The process of high-level disinfection includes mechanical cleaning of the working channels and exterior of the endoscope, followed by soaking in disinfectant solutions such as glutaraldehyde or orthophthalaldehyde, and then thorough rinsing and drying of the instruments. Most instances of documented transmission of infection can be traced to a failure in one of the recommended steps of endoscope reprocessing.28,30,31 The estimated pre­ valence of transmission of infectious organisms by endo­ scopes is 1 in 1.8 million procedures.30,31 This number may underestimate the frequency of this problem, however, because some infections may not be detected or reported. Prospective studies have not demonstrated transmission of hepatitis C by properly performed gastrointestinal endoscopy.32,33 High-level disinfection kills most viruses and bacteria that could contaminate endoscopes. Common blood-borne pathogens such as human immunodeficiency virus and hepatitis B and C viruses are readily inactivated by the highlevel disinfection process. Although prions such as the Jakob-Creutzfeldt agent may not be inactivated by high-level disinfection, transmission of these agents by endoscopy has not been reported.28 Infectious organisms may also be transmitted by endo­ scopic accessories or by contaminated needles and drugs used for sedation. Most endoscopic accessories are designed for single use and should be discarded following use. Reusable accessories should be sterilized after use to avoid transmitting infectious organisms.29 Outbreaks of hepatitis C following endoscopy have been traced to improper sterile technique and contamination of multidose vials of sedative medications. Sterile single-use needles and intravenous tubing should always be used. Unused medication should be discarded after each procedure, and use of multidose medication vials should be discouraged.

OTHER GENERAL COMPLICATIONS ELECTROSURGERY

Electrosurgery is the use of high-frequency electrical current to produce the controlled application of heat to tissue. Elec­ trosurgery is used frequently in GI endoscopy to coagulate blood vessels and remove or cut tissue without inducing bleeding. Current electrosurgical generators are carefully engineered to promote patient and operator safety. Compli­ cations are infrequent but can be a result of equipment failure, operator error, or patient factors. Electrosurgical generators are reliable devices that have long useful lives. Routine electrical maintenance and safety checks are mandatory, however. Operator error can be reduced by familiarity with the equipment being used. A large number and variety of electrosurgical generators exist, and they differ in their effect on tissue even with the same energy settings.34 The dispersive electrode or ground pad should be placed in firm contact with a large skin surface to avoid skin burns. Electrosurgery with monopolar devices causes a spark of energy when the device is activated. If sufficient concentra­ tions of explosive gases are present, the spark may trigger an explosion. This has been reported when electrosurgery has been performed in the poorly prepared colon or follow­

ing mannitol-containing laxatives in which high concentra­ tions of hydrogen and methane may be present.35 Patients who have implanted electrical devices such as cardiac pacemakers and implanted defibrillators require special considerations. In the first instance, electrocautery can inhibit cardiac pacemaker function. This can be mini­ mized by placing the dispersive electrode well away from the pacemaker on the patient’s thigh or buttock and by using brief bursts of electrosurgical output. In the second instance, electrosurgery can cause unwanted activation of implanted cardiac defibrillators (ICDs), resulting in patient movement during endoscopic surgery and even resulting in electrical shock to the endoscopist. This is best avoided by deactivat­ ing these ICDs before electrosurgery. An external defibrilla­ tor should be immediately available if the implanted device is deactivated, and the patient’s cardiac rhythm must be continuously monitored.

MISCELLANEOUS COMPLICATIONS

Abdominal distention is usually caused by air insufflation with upper endoscopy or colonoscopy but can also be due to gas administered during laser endoscopy or argon plasma coagulation. This can cause significant discomfort and even lead to vasovagal reactions (see earlier). Care should be taken to suction air before withdrawal of the endoscope. Although the prevalence of minor complications of endoscopy has not been well established, these problems should not be trivialized. Sore throats after upper endos­ copy, pain, infections or phlebitis at intravenous catheter sites, and prolonged recovery from the effects of sedation can affect a patient’s quality of life.36

TIMING AND SEVERITY OF COMPLICATIONS Endoscopic complications can occur during the procedure itself or be delayed. Recognition of immediate complica­ tions allows prompt management of the problem and may improve the outcome of the patient. The endoscopist should have a high index of suspicion for complications of endo­ scopic procedures and investigate all potential problems. The value of a negative investigation should not be under­ estimated because this is often reassuring to the endoscopist as well as the patient. Delayed procedural complications can be caused by late occurrence of the problem or the delayed presentation or recognition of an early complication. For example, bleeding from a colonic ulcer after polypectomy may not occur until one to two weeks after colonoscopy.37 Asking the patient to watch for bloody or melenic stools and to inform the endoscopist promptly should they occur should be part of the postprocedure discharge instructions. Attrib­ uting abdominal distention to intracolonic air after a colonoscopy and not suspecting colonic perforation may lead to delayed recognition of an immediate life-threatening complication.

MEDICOLEGAL CONSIDERATIONS Complications of endoscopy are one of the most frequent causes of malpractice suits against gastroenterologists.38 Malpractice can occur if the physician does not meet his or her obligation to the patient or if the care provided does not meet the standard of care. The occurrence of a complication

Chapter 40  Complications of Gastrointestinal Endoscopy does not mean malpractice was committed if the procedure was properly performed and the patient was informed of potential complications of the procedure.

COMPLICATIONS OF UPPER ENDOSCOPY RESPIRATORY PROBLEMS

It is remarkable that respiratory complications of upper endoscopy are infrequent, given that the endoscope is passed through the oropharynx. Stridor, reflecting upper airway compromise, occurs rarely during endoscopy, usually in patients with small upper airways due to con­ genital anomalies, prior surgery, or radiation. Use of smallcaliber endoscopes in selected patients may help reduce this problem. Patients with neuromuscular weakness such as occurs with amyotrophic lateral sclerosis may also have symptoms and signs of upper airway obstruction during endoscopy. Nasal administration of positive-pressure venti­ lation during upper endoscopy may increase procedural safety in these patients.39 Aspiration pneumonia is another complication of upper endoscopy that occurs in approximately 1 in 10,000 proce­ dures.17 Careful attention to oral suctioning and selective use of endotracheal intubation for endoscopy in at-risk patients such as the obtunded patient with upper GI bleed­ ing may reduce this complication. Elective endotracheal intubation of patients with suspected variceal hemorrhage who are not encephalopathic may increase the risk of aspi­ ration pneumonia, however.40

COMPLICATIONS OF TOPICAL ANESTHESIA

Topical anesthesia is commonly used during upper endos­ copy to ease the performance of endoscopy and improve patient tolerance.41 Lidocaine and especially benzocaine sprays very rarely cause methemoglobinemia, which should be considered when oxygen saturation detected by pulse oximetry falls during or following an endoscopy in which topical anesthesia was administered.42,43 Clues to methemo­ globinemia are a normal respiratory rate and arterial oxygen content (Po2) despite clinical cyanosis and low oxygen saturation on pulse oximetry. The condition can be fatal but is reversible after intravenous administration of methylene blue.

HEMORRHAGE

Bleeding complications of upper endoscopy are most com­ monly seen during therapeutic procedures such as dilation or enteral access procedures (see later). Bleeding caused by passage of the endoscope through the oropharynx or from a Mallory-Weiss tear at the gastroesophageal junction has been reported but is rare44,45 and usually stops spontane­ ously unless the patient has a coagulopathy. Bleeding from mucosal biopsies, even when a large cup forceps is used, is also infrequent but can lead to either intraluminal hemor­ rhage or intramural hematomas.46 Biopsies should not be done in patients with significant prolongations in the pro­ thrombin time (INR >2) or with severe thrombocytopenia (<20,000).17,44

PERFORATION

Perforation of the upper GI tract during diagnostic endos­ copy has been estimated to occur in 3 to 5 out of 10,000 procedures and is usually associated with therapeutic as opposed to diagnostic procedures.17,47 The most common site of perforation from upper endoscopy is in the orophar­

ynx or cervical esophagus. Patients with Zenker’s divertic­ ula, proximal esophageal strictures and cancers, and those with large cervical osteophytes are at increased risk of per­ foration. The presence of crepitus in the neck soft tissues, fever, and chest or neck pain following endoscopy should prompt an investigation for perforation with chest and neck radiographs; a pharyngoesophagogram with water-soluble contrast; and, if necessary, neck and chest computed tomog­ raphy (CT) scan. Pharyngeal perforations that are not recog­ nized at the time of endoscopy may manifest after a few days or weeks with a retropharyngeal abscess, which should be drained surgically. Most perforations within the neck can be managed con­ servatively in conjunction with an otolaryngologist and use of broad-spectrum intravenous antibiotics. Intrathoracic perforations can also be managed conservatively with anti­ biotics and nasoesophageal suction if the perforation is small and contained to the tissues immediately surrounding the esophagus. When there is communication with the pleural space, thoracotomy is usually recommended. The recent availability of removable polyethylene esophageal stents may change this recommendation, especially in patients with malignancy or increased risks for surgery because placement of this type of stent may allow esopha­ geal perforations to seal without surgery.48

COMPLICATIONS OF DILATION

Dilation anywhere in the GI tract increases the risk of com­ plications compared with diagnostic endoscopy. The great­ est risk is perforation, with significant bleeding being much less common. The type of dilator used, whether a guidewire bougie or through the scope balloon, does not appear to significantly affect risk.49 Malignant, radiation-induced, and lye strictures in the esophagus are more likely to perforate than peptic strictures. Balloon dilation of pyloric and duo­ denal strictures appears to carry a greater risk of perforation than dilation of strictures at surgical anastomoses. Although never proved, the practice of gradual dilation over multiple endoscopic sessions may carry a lower risk of perforation. Strictures should be dilated to a diameter that results in symptom resolution, and not necessarily to the size of the uninvolved lumen. Cerebral air embolism complicated by stroke and seizure has been reported following esophageal dilation and other endoscopic procedures.50

COMPLICATIONS OF ENDOSCOPIC HEMOSTASIS

Multipolar and heater probe therapy of bleeding peptic ulcers carries a risk of perforation of approximately 1%.51 This risk increases to 4% when a second treatment session is required within 48 hours for hemostasis of recurrent bleeding.52 Injection therapy with epinephrine can cause tachycardia and in some cases ischemic ulceration at the site of injection (see Chapters 10 and 53). Variceal sclerotherapy has been largely replaced by vari­ ceal band ligation because of similar efficacy but fewer complications with ligation (see Chapter 90). Sclerotherapy causes esophageal ulcerations that can bleed significantly in 6% of patients and lead to delayed esophageal strictures in up to 20% of patients.53 Other complications of sclero­ therapy include esophageal perforation in 2% to 5% of patients, pleural and pericardial effusions, mediastinitis, and paralysis.54 These complications carry a high mortality, largely because of the presence of severe underlying liver disease in these patients. Band ligation leads to fewer ulcer­ ations and delayed strictures than sclerotherapy and per­ foration in less than 1% of cases.53,55

657

658

Section IV  Topics Involving Multiple Organs COMPLICATIONS OF ENTERAL ACCESS PROCEDURES

Endoscopy is used to place a variety of tubes into the upper GI tract for the delivery of enteral nutrition (see Chapter 5). Endoscopic nasoenteric tube placement ensures delivery of the feeding tube into the small intestine and is associated with usually minor, self-limited complications in 10% of cases.56 Epistaxis is the most common complication, occur­ ring in 2% to 5%. Proximal migration out of the small intestine occurs in 15%, and tube clogging in up to 20% of cases.57 PEG involves direct puncture of the stomach, through the abdominal wall, under endoscopic control (see Chapter 5). Complications occur in 1.5% to 4% of procedures.57 The most frequent complication of this procedure is infection at the site of tube entry, which occurs in up to 30% of cases. These infections are usually minor and can be treated with antibiotics but can occasionally be severe and lead to necrotizing fasciitis that requires surgical débridement. As mentioned, preprocedure antibiotics have been shown to reduce the frequency of minor wound infections after PEG placement.58,59 Bleeding during PEG placement is usually minor and selflimited but can require endoscopic hemostasis in 1% of cases, usually as a result of puncture of an artery in the gastric wall.60 Other rare complications of PEG procedures include premature tube dislodgment that can lead to peri­ tonitis and inadvertent puncture of the liver and colon, the latter leading to formation of a gastrocolic fistula. The “buried bumper syndrome” occurs when the external bolster of the PEG tube remains too tight and causes migration of the internal bolster (or bumper) into the gastric wall.61 When PEG tubes are placed in patients with head and neck or esophageal cancer, seeding of the PEG site with tumor implants may rarely occur by either local or hematogenous routes.62,63 Aspiration pneumonia in patients receiving enteral feedings can be either due to aspiration of oropharyngeal bacteria or from gastroesophageal reflux of enteral feedings. The former mechanism is probably more common.64 However, GE reflux can be responsible for pneumonia, espe­ cially in patients with a history of documented aspiration, overt vomiting or regurgitation, reduced level of conscious­ ness, neuromuscular or structural problems of the orophar­ ynx, prolonged supine position, and high gastric residual volumes.64 Placement of feeding tubes directly into the proximal jejunum is being increasingly performed. Complications are similar to those of PEG, with the addition of rare cases of small bowel obstruction that can be caused by large inter­ nal bolsters or small intestinal volvulus.65

COMPLICATIONS OF OTHER THERAPEUTIC PROCEDURES

Expandable metal stents are used to treat malignant eso­ phageal strictures (see Chapter 46) and are associated with complications in approximately 10% to 20% of cases.66 Perforation, chest pain, and gastroesophageal reflux are the most common complications. Delayed complications include distal or proximal migration and food impaction. Endoscopic removal of foreign bodies from the upper GI tract (see Chapter 25) is associated with complications in up to 8% of cases.67 Aspiration pneumonia, bleeding, and esophageal perforation can be avoided by attention to guidelines on the types of foreign bodies that should be removed endoscopically and the timing of the intervention.68

COMPLICATIONS OF SMALL BOWEL ENDOSCOPY Complications of enteroscopy are similar to those of upper endoscopy. Push enteroscopy with an overtube adds the risk of pharyngeal or esophageal perforation from advance­ ment of the overtube. This complication has become less frequent with the use of more flexible overtubes.

DOUBLE BALLOON ENTEROSCOPY

Double balloon enteroscopy is a new technique for visua­ lization of larger extents of the small bowel by using sequen­ tially inflated balloons on the shaft of the endoscope and an accompanying overtube. The endoscope has a channel that allows application of therapeutic techniques such as polypectomy, injection, and thermal coagulation. As with upper endoscopy, complication rates are higher with thera­ peutic procedures (4%) compared with diagnostic pro­ cedures (1%).69 An unexpected complication of double balloon enteroscopy is acute pancreatitis. This has been reported in about 1 in 500 procedures and is hypothesized to be due to increased intraluminal pressures in the duode­ num between the inflated balloons on the endoscope and overtube.69,70

CAPSULE ENDOSCOPY

The primary complication of capsule endoscopy is retention of the capsule above an unsuspected small intestinal stric­ ture. This occurs in 1% to 2% of capsule procedures.71 These strictures can be quite focal and not detected by small intestinal radiography. A dissolvable, radiopaque, radiofre­ quency emitting capsule has been proposed for detecting small bowel strictures prior to capsule ingestion.72 Other rare complications of capsule endoscopy include aspiration of the capsule into the trachea.73

COMPLICATIONS OF COLONOSCOPY AND SIGMOIDOSCOPY The overall risk of complications of diagnostic colonoscopy is approximately 0.3%.74-78 The risk is higher (2%) when polypectomy is performed. The main complications of colo­ noscopy are perforation, bleeding, and postpolypectomy syndrome. Colon preparation regimens can also cause com­ plications in some patient populations (see later). The risks of sigmoidoscopy are approximately two-fold lower than colonoscopy but include the same types of complications seen with colonoscopy.79

HEMORRHAGE

The most common cause of bleeding during or following colonoscopy is a polypectomy. This occurs in approxi­ mately 1.5% to 3% of patients undergoing polypectomies, with approximately an equal distribution of immediate and delayed bleeding.80 Delayed bleeding usually occurs within two weeks following polypectomy and is more common when large polyps are removed in patients with uncon­ trolled hypertension.37 Immediate bleeding usually can be stopped at the time of colonoscopy by holding a snare around a polyp stalk for five minutes, injecting dilute epinephrine solution, or placing a hemostatic clip on the bleeding vessel (Fig. 40-1). Delayed bleeding is best diag­ nosed by repeat colonoscopy, which allows treatment using the same modalities that are used with immediate bleeding.

Chapter 40  Complications of Gastrointestinal Endoscopy

A

B

Figure 40-2.  A patient with iron deficiency anemia attributed to angioectasia in the cecum was treated four days previously with argon plasma coagulation. The patient developed hematochezia and then underwent bowel cleansing before repeat colonoscopy, which revealed active bleeding from a vessel in the base of an ulcer (arrow) at the site of the previous treatment.

Delayed bleeding also can occur from causes other than polypectomy. Ulcerations caused by treatment of angioec­ tasia in the colon with argon plasma coagulation or multi­ polar probes can bleed (Fig. 40-2; also see Chapters 19 and 36). Biopsy sites are rare sources of bleeding following colo­ noscopy, even when multiple biopsies are taken during surveillance for dysplasia in patients with inflammatory bowel disease.81

PERFORATION

Colonic perforation is the most feared complication of colo­ noscopy but is fortunately infrequent, occurring in 0.1% to 0.9% of colonoscopies.75-79,82 Polypectomy approximately doubles the risk of perforation over the background perfora­ tion rate with diagnostic colonoscopy.77 Perforations can be caused by excessive air pressure (barotrauma), by tearing of the antimesenteric border of the colon from excessive pressure on colonic loops, and at the

Figure 40-1.  A, This polyp stalk was believed to be the cause of hematochezia in a patient who had undergone colonoscopy and polypectomy five days previously. B, Although no active bleeding was encountered during the second procedure, a hemostatic clip was placed in an effort to reduce the risk of further bleeding.

Figure 40-3.  Pericolonic fat is seen through the wall of the sigmoid colon, which was perforated during colonoscopy.

sites of electrosurgical applications. Barotrauma occurs most often in the cecum, where colonic diameter is greatest and therefore tension on the colonic wall is highest. Colonic tears occur most frequently in the sigmoid colon, where loops are most frequent. Large tears can sometimes be rec­ ognized during the colonoscopy and should lead to immedi­ ate operative intervention (Fig. 40-3). Perforation can also occur at polypectomy sites or following the application of argon plasma coagulation or thermal probes for the treat­ ment of bleeding diverticula or angioectasia.83 Other uses of colonoscopy as a treatment modality also increase the risk of colonic perforation. Colonoscopic decompression of acute colonic pseudo-obstruction (see Chapter 120) is associated with perforation in 2% of cases.84 Placement of colonic stents to relieve malignant obstruction (see Chapter 123) may cause perforation in up to 5% of cases.85 Balloon dilation of colonic strictures in patients with Crohn’s disease (see Chapter 111) carries a 10% risk of perforation.86 Perforation should be suspected in the patient who com­ plains of abdominal or shoulder pain following colonos­

659

660

Section IV  Topics Involving Multiple Organs copy or when abdominal distention is excessive and unremitting. If perforation is not recognized early, progres­ sive abdominal pain and fever may ensue due to bacterial peritonitis. Upright abdominal and chest radiographs usually reveal free air under the diaphragm. Abdominal CT scan should be obtained if abdominal radiographs are unrevealing and if there is a high clinical suspicion of perforation. Treatment of colonoscopic perforation depends on the severity of the perforation and the condition of the patient. Large tears and patients with signs of peritonitis require operative treatment. Earlier detection may allow primary repair without formation of a colostomy. Stable patients with microperforations caused by barotrauma or electrocau­ tery can sometimes be managed without surgery with bowel rest and parenteral antibiotics.87,88 Careful observation of the patient by the physician in conjunction with a surgeon is advisable in this situation.

POSTPOLYPECTOMY COAGULATION SYNDROME

Full-thickness electrosurgical burns following polypectomy may cause localized abdominal pain, fever, and leukocyto­ sis without free intra-abdominal air on imaging studies. Localized peritoneal signs may be present. Patients usually present from 1 to 5 days following colonoscopy, and symp­ toms resolve in 2 to 5 days.89 Management depends on the severity of symptoms. Patients with mild pain and little fever can be managed with oral antibiotics as outpatients. Patients with more severe pain and fever should be observed in the hospital with bowel rest, intravenous antibiotics, and frequent physical examinations and radiographs to detect perforation.90

COMPLICATIONS RELATED TO COLON PREPARATION

Oral administration of sodium phosphate and polyethylene glycol electrolyte solutions to clean the colon before colo­ noscopy can cause problems in susceptible patients. Hyper­ phosphatemia, hypocalcemia, and acute kidney injury are most likely to occur with sodium phosphate preparations in patients with renal insufficiency, older adults, and patients taking angiotensin-converting enzyme inhibitors and angiotensin receptor blockers.91-94 Sodium phosphate solutions can cause intravascular volume depletion in patients with congestive heart failure, acute kidney injury, and cirrhosis. Polyethylene glycol–containing solutions are better tolerated in these situations but may also cause fluid shifts; patients with severe underlying diseases should have a more gradual bowel preparation and be monitored closely.95,96

OTHER COMPLICATIONS

Other rare complications of colonoscopy include splenic rupture or hemorrhage, intra-abdominal bleeding caused by mesenteric vessel rupture, and acute appendicitis.97 Chemi­ cal colitis due to inadequate rinsing of disinfectant solu­ tions has been reported.98 Death from colonoscopy is rare, occurring in less than 1 in 16,000 procedures.90

COMPLICATIONS OF ERCP Endoscopic retrograde cholangiopancreatography (ERCP) is one of the most rewarding endoscopic procedures but also is one of the most dangerous. Appropriate training, ade­

Table 40-3  Risk Factors for Post-ERCP Pancreatitis PATIENT FACTORS

PROCEDURAL FACTORS

Young age Female gender Suspected sphincter of Oddi dysfunction Previous post-ERCP pancreatitis Recurrent pancreatitis

Number of injection attempts Pancreatic duct injection Pancreatic sphincterotomy Balloon dilation of biliary sphincter Difficult or failed cannulation Precut sphincterotomy

ERCP, endoscopic retrograde cholangiopancreatography. From Freeman ML, Guda NM: Prevention of post-ERCP pancreatitis: A comprehensive review. Gastrointest Endosc 2004; 59:845.

quate ongoing experience, and good clinical judgment are requisites to avoiding complications.99 A striking observa­ tion about post-ERCP complications is that the patients who are at the highest risk for complications, and especially the more severe complications, are the ones who are least likely to benefit from the procedure.100,101

PANCREATITIS

The most common complication of ERCP is acute pancre­ atitis (see Chapter 58), which occurs in 2% to 25% of cases.101-103 The risk factors for post-ERCP pancreatitis have been well defined and include patient and procedural factors (Table 40-3).101,104,105 Pancreatitis severity can range from mild, resulting in two to three days of hospitalization, to severe, requiring surgery or even causing death. Pharmacologic approaches to preventing post-ERCP pancreatitis have been disappointing.105 Using alternative imaging techniques such as magnetic resonance cholangio­ pancreatography (MRCP) or endoscopic ultrasonography in patients with a low probability of requiring endoscopic therapy may be the best way to avoid this complication. When ERCP is performed in high-risk patients such as a young woman with a small bile duct and suspected sphinc­ ter of Oddi dysfunction, placing a temporary pancreatic stent can significantly reduce the risk of post-ERCP pancre­ atitis.106,107 Whether the use of pure cutting current for sphincterotomy reduces the risk of post-ERCP pancreatitis is controversial.108-110 Treatment of post-ERCP pancreatitis is supportive and similar to treatment of other causes of acute pancreatitis (see Chapter 58). There is no established role for repeat ERCP or pancreatic stenting in this setting.

HEMORRHAGE

Bleeding complications of ERCP are usually secondary to sphincterotomy and occur in 1% to 2% of cases.100,102,103 Risk factors for postsphincterotomy bleeding include coagu­ lopathy, use of anticoagulants within 72 hours following the procedure, cholangitis, precut sphincterotomy, and low case volume of the endoscopist.100 Bleeding seen at the time of the sphincterotomy is also predictive of delayed bleeding and should be treated aggressively. Treatment of delayed postsphincterotomy hemorrhage initially should be directed at transfusion of blood products and correction of coagulopathy. Repeat endoscopy with use of epinephrine injection, thermal probes, and hemostatic clips is effective for stopping bleeding in most cases.111 When endoscopic therapy fails, angiographic therapy or surgery should be considered depending on the condition and comorbidities of the patient.

Chapter 40  Complications of Gastrointestinal Endoscopy PERFORATION

Another complication of ERCP is perforation of the upper GI or biliary tract. This occurs in approximately 0.5% of cases and can be caused by guidewires, periampullary perforation from sphincterotomy, or the endoscope at sites remote from the ampulla.112 Guidewire perforations gener­ ally occur in the biliary tree and usually do not create significant bile leaks if distal obstruction is relieved with sphincterotomy or biliary stenting. Perforations remote from the ampulla are often large and require surgical repair. These perforations should be suspected in patients with marked abdominal pain and distention following ERCP and diagnosed with upright abdominal and chest radiographs. Periampullary perforations are usually contained to the retroperitoneum surrounding the ampulla and can be diag­ nosed with abdominal CT scan. If recognized promptly, the majority can be treated with nasogastric suction and intra­ venous antibiotics.113 Percutaneous or surgical drainage is required if CT scan documents an enlarging abscess despite conservative therapy.

CHOLANGITIS

Ascending cholangitis following ERCP occurs in less than 1% of cases and is usually caused by injecting contrast into an obstructed biliary tree and then not providing adequate biliary drainage by removing all stones or placing a biliary stent.102,114 Patients with complex biliary strictures at the hepatic hilum due to cholangiocarcinoma have an increased risk of cholangitis. Care should be taken to inject contrast only into bile ducts that can be subsequently drained with a stent.115 Prophylactic antibiotics have not been shown to reduce the risk of cholangitis following ERCP116,117 and their use is declining.118 Current guidelines recommend prophy­ lactic antibiotics only for patients with biliary obstruction when it is anticipated that drainage might be incomplete (see Table 40-1).1 Treatment of cholangitis includes antibiotics and provi­ sion of biliary drainage. Repeat ERCP for stone removal or stent placement is usually required. Percutaneous biliary drainage should be undertaken if ERCP cannot relieve the biliary obstruction.

OTHER COMPLICATIONS

Acute cholecystitis occurs in less than 0.5% of ERCPs. Cystic duct obstruction by gallbladder stones and bile duct stents that occlude the cystic duct orifice are the usual causes. Pancreatic infection can also occur in up to 8% of cases when contrast is injected into obstructed pancreatic ducts or pseudocysts and drainage is not provided.119 A plan for pseudocyst drainage, either endoscopic, percutaneous, or surgical, should always be made before undertaking ERCP in patients with a known pancreatic pseudocyst (see Chapter 61).120,121

COMPLICATIONS OF ENDOSCOPIC ULTRASONOGRAPHY Diagnostic EUS carries the same risks of sedation, bleeding, and perforation as diagnostic endoscopy. The frequency of complications related to EUS has been reported to be from 0.1% to 0.3%.122,123 The addition of medical ultrasound carries no additional known risks to the patient. Because most ultrasound endoscopes are forward oblique viewing rather than forward viewing, care must be taken in passing the endoscope through the oropharynx and strictures.124 The risk of perforation of malignant esophageal strictures was as high as 24% with older endoscopes that were larger

and had a blunt tip. Perforation using current instruments is much less common. It is advisable to serially dilate tight esophageal strictures before EUS or to use special smallcaliber, wire-guided instruments in this setting.125

COMPLICATIONS OF FINE-NEEDLE ASPIRATION

The addition of FNA to EUS has introduced the potential for additional complications of hemorrhage, infection, and pancreatitis. Bleeding due to EUS-FNA is rare but can be lethal if a major vessel is lacerated.126 This complication has not been reported with linear array ultrasonography endo­ scopes that are used for FNA, however. Infection is primarily a risk when pancreatic or mediasti­ nal cysts are aspirated. This has been reported to occur in up to 14% of cases.127 The use of prophylactic antibiotics and aspiration of all fluid from the cyst with one pass of the needle have been advocated as ways to reduce the risk of infection.128 Prophylactic antibiotics are not recommended prior to FNA of solid lesions because bacteremia is rare, even following lower GI tract procedures.129 Pancreatitis following EUS-FNA of the pancreas has been reported in less than 1% of cases.130 Care should be taken to find a needle path away from the main pancreatic duct when undertaking FNA of a pancreatic lesion.

KEY REFERENCES

American Society of Anesthesiologists. Task Force on Sedation and Analgesia by Non-Anesthesiologists. Practice guidelines for sedation and analgesia by non-anesthesiologists. Anesthesiology 2002; 96:1004-17. (Ref 10.) American Society for Gastrointestinal Endoscopy. Guideline on the management of low-molecular-weight heparin and nonaspirin anti­ platelet agents for endoscopic procedures. Gastrointest Endosc 2005; 61:189-94. (Ref 5.) American Society for Gastrointestinal Endoscopy. Guideline on antibi­ otic prophylaxis for GI endoscopy. Gastrointest Endosc 2008; 67:7918. (Ref 1.) American Society for Gastrointestinal Endoscopy. Guideline on infec­ tion control during GI endoscopy. Gastrointest Endosc 2008; 67:78190. (Ref 30.) Ciancio A, Manzini P, Castagno F, et al. Digestive endoscopy is not a major risk factor for transmitting hepatitis C virus. Ann Intern Med 2005; 142:903-9. (Ref 32.) Cobb WS, Heniford T, Sigmon LB, et al. Colonoscopic perforations: incidence, management, and outcomes. Am Surg 2004; 70:750-8. (Ref 88.) Cotton PB, Connor P, Rawls E, Romagnuolo J. Infection after ERCP, and antibiotic prophylaxis: a sequential quality-improvement approach over 11 years. Gastrointest Endosc 2008; 67:471-5. (Ref 118.) Evans LT, Saberi S, Kim HM, et al. Pharyngeal anesthesia during sedated EGDs: Is “the spray” beneficial? A meta-analysis and systematic review. Gastrointest Endosc 2006; 63:761-6. (Ref 41.) Freeman ML, Nelson DB, Sherman S, et al. Complications of endoscopic biliary sphincterotomy. N Engl J Med 1996; 335:909-18. (Ref 100.) Khurana A, McLean L, Atkinson S, Foulks CJ. The effect of oral sodium phosphate drug products on renal function in adults undergoing bowel endoscopy. Arch Intern Med 2008; 168:565-7. (Ref 94.) Ko CW, Riffle S, Shapiro JA, et al. Incidence of minor complications and time lost from normal activities after screening or surveillance colonoscopy. Gastrointest Endosc 2007; 65:648-56. (Ref 36.) Levin TR, Zhao W, Conell C, et al. Complications of colonoscopy in an integrated health care delivery system. Ann Intern Med 2006; 145:880-6. (Ref 77.) McClave SA, Chang W-K. Complications of enteral access. Gastrointest Endosc 2003; 58:739-51. (Ref 57.) Qadeer MA, Vargo JJ, Khandwala F, et al. Propofol versus traditional sedative agents for gastrointestinal endoscopy: A meta-analysis. Clin Gastroenterol Hepatol 2005; 3:1049-56. (Ref 14.) Tarnasky PR, Palesch YK, Cunningham JT, et al. Pancreatic stenting prevents pancreatitis after biliary sphincterotomy in patients with sphincter of Oddi dysfunction. Gastroenterology 1998; 115:1518-24. (Ref 106.) Full references for this chapter can be found on www.expertconsult.com.

661

CHAPTE R

41 Anatomy, Histology, Embryology, and Developmental Anomalies of the Esophagus John D. Long and Roy C. Orlando

CHAPTER OUTLINE Anatomy and Histology  665 Musculature  665 Innervation  665 Circulation  667 Mucosa  667 Submucosa  667 Embryology  668 Developmental Anomalies  668

ANATOMY AND HISTOLOGY The esophagus acts as a conduit for the transport of food from the oral cavity to the stomach. To carry out this task safely and effectively, the esophagus is constructed as an 18- to 26-cm long hollow muscular tube with an inner “skinlike” lining of stratified squamous epithelium (Fig. 41-1). Between swallows the esophagus is collapsed, but the lumen distends up to 2 cm anteroposteriorly and 3 cm laterally to accommodate a swallowed bolus. Structurally, the esophageal wall is composed of four layers: innermost mucosa, submucosa, muscularis propria, and outermost adventitia; unlike the remainder of the gastrointestinal tract, the esophagus has no serosa.1,2 These layers are depicted anatomically and as viewed by endoscopic ultrasonography in Figure 41-2.

MUSCULATURE

The muscularis propria is responsible for carrying out the organ’s motor function. The upper 5% to 33% is composed exclusively of skeletal muscle, and the distal 50% is composed of smooth muscle. In between is a mixture of both types.3 Proximally, the esophagus begins where the inferior pharyngeal constrictor merges with the cricopharyngeus, an area of skeletal muscle known functionally as the upper esophageal sphincter (UES) (Fig. 41-3A). The UES is contracted at rest and thereby creates a high pressure zone that prevents inspired air from entering the esophagus. Below the UES the esophageal wall comprises inner circular and outer longitudinal layers of muscle (see Fig. 41-2A). The esophageal body lies within the posterior mediastinum

Esophageal Atresia and Tracheoesophageal Fistula  668 Congenital Esophageal Stenosis  671 Esophageal Duplications  671 Vascular Anomalies  672 Esophageal Rings  672 Esophageal Webs  674 Heterotopic Gastric Mucosa (Inlet Patch)  675

behind the trachea and left mainstem bronchus and swings leftward to pass behind the heart and in front of the aorta.1 At the T10 vertebral level the esophageal body leaves the thorax through a hiatus located within the right crus of the diaphragm (see Fig. 41-1). Within the diaphragmatic hiatus the esophageal body ends in a 2- to 4-cm length of asymmetrically thickened circular smooth muscle known as the lower esophageal sphincter (LES) (see Fig. 41-3B).4 The phrenoesophageal ligament, which originates from the diaphragm’s transversalis fascia and inserts on the lower esophagus, contributes to fixation of the LES within the diaphragmatic hiatus. This positioning is beneficial because it enables diaphragmatic contractions to assist the LES in maintenance of a high-pressure zone during exercise. The LES is contracted at rest, creating a high-pressure zone that prevents gastric contents from entering the esophagus. During swallowing, the LES relaxes to permit the swallowed bolus to be pushed by peristalsis from the esophagus into the stomach.

INNERVATION

The esophageal wall is innervated by parasympathetic and sympathetic nerves; the parasympathetics regulate peri­ stalsis through the vagus nerve (Fig. 41-4). The cell bodies of the vagus nerve originate in the medulla. Those located within the nucleus ambiguus control skeletal muscle, and those located within the dorsal motor nucleus control smooth muscle. Medullary vagal postganglionic efferent nerves terminate directly on the motor endplate of skeletal muscle in the upper esophagus, whereas vagal preganglionic efferent nerves to smooth muscle in the distal esopha-

665

666

Section V  Esophagus Distance from incisors

Stratified squamous Lamina Muscularis epithelium propria mucosae

Mucosa

Longitudinal folds

Cervical esophagus

UES Trachea Submucosa Adventitia

40 cm

Aorta

Submucosal Inner circular layer gland with Outer longitudinal layer duct

A

Muscularis propria

Thoracic esophagus

Right crus of diaphragm LES

Abdominal esophagus

B

Figure 41-1.  The esophagus, approximately 20 cm in length, originates in the neck at the level of the cricoid cartilage, passes through the chest, and ends after passage through the hiatus in the right crus of the diaphragm by joining the stomach below. On barium esophagogram, adjacent structures may indent the esophageal wall, including the aortic arch, left mainstem bronchus, left atrium, and diaphragm. LES, lower esophageal sphincter; UES, upper esophageal sphincter. (Modified from LiebermannMeffert D. Anatomy, embryology, and histology. In: Pearson FG, Cooper JD, Deslauriers J et al, editors. Esophageal Surgery. 2nd ed. Philadelphia, Pa: Churchill Livingstone; 2002. p 8.)

gus terminate on neurons within Auerbach’s (myenteric) plexus, located between the circular and longitudinal muscle layers.3 A second neuronal sensory network, Meissner’s plexus, located within the submucosa, is the site of afferent impulses within the esophageal wall. These are transmitted to the central nervous system through vagal parasympathetic and thoracic sympathetic nerves. Sensory signals transmitted via vagal afferent pathways travel to the nucleus tractus solitarius within the central nervous system (see Fig. 41-4); from there nerves pass to the nucleus ambiguus and dorsal motor nucleus of the vagus nerve, where their signals may influence motor function.5 Pain sensation arising from the esophagus is typically triggered by stimulation of chemoreceptors in the esopha-

Figure 41-2.  Cross-sectional and endoscopic ultrasonographic anatomy of the esophagus. A, The anatomic layers within the wall of the esophagus are depicted. B, An endoscopic ultrasonographic image depicting the pattern of light and dark rings created by echoes from the different layers. (A, Interface between lumen and mucosa; B, mucosa; C, submucosa; D, muscularis propria; E, adventitia.) Note that A, C, and E are hyperechoic and B and D are hypoechoic. (A, Modified from Neutra MR, Padykula HA. The gastrointestinal tract. In: Weiss L, editor. Histology, Cell and Tissue Biology. 5th ed. New York, NY: Elsevier Science; 1983. p 664.)

geal mucosa or submucosa and/or mechanoreceptors in the esophageal musculature.6 Central perception then occurs when these impulses are transmitted to the brain by sympathetic and vagal afferents. Sympathetic afferents travel through the dorsal root ganglia to the dorsal horn of the spinal cord, and vagal afferents travel through the nodose ganglia to the nucleus tractus solitarius in the medulla. Information from sympathetic/spinal afferents then proceeds via the spinothalamic and spinoreticular pathways to the thalamus and reticular nuclei before transmission to the somatosensory cortex for pain perception and limbic system for pain modulation. Information from vagal afferents in the medulla also travels to the limbic system and frontal cortex for pain modulation. Furthermore, because the esophageal neuroanatomic pathways overlap with those of the heart and respiratory system, in clinical practice it may be difficult to discern the organ of origin for some chest pain syndromes.6

Chapter 41  Anatomy, Histology, Embryology, and Developmental Anomalies of the Esophagus UES Inferior constrictor

UES

Thyroid cartilage Cricoid cartilage

Trachea

MUCOSA

Cricopharyngeus

Proximal esophagus

A LES Pleura Diaphragm

veins. The submucosal venous anastomotic network of the distal esophagus is important because it is where esophageal varices emerge in patients with portal hypertension.1-3 The lymphatic system of the esophagus is also segmental; the upper esophagus drains to the deep cervical nodes, the midesophagus to the mediastinal nodes, and the distal esophagus to the celiac and gastric nodes. However, these lymphatic systems are also interconnected by numerous channels, accounting for the spread of most esophageal cancers beyond the region at the time of their discovery.

A Ring Phrenoesophageal ligament Sling fibers

Squamocolumnar junction Peritoneum

B Figure 41-3.  A, Anatomic detail of the upper esophageal sphincter (UES) and its relationship to adjacent structures. B, Anatomic detail of the lower esophageal sphincter (LES) and its relationship to the diaphragm, phrenoesophageal ligament, and squamocolumnar junction. (A, Modified from AGA Clinical Teaching Project. Esophageal disorders: Upper esophageal sphincter anatomy, slide 14, American Gastroenterological Association, 1995; B, modified from Kerr RM. Hiatal hernia and mucosal prolapse. In: Castell DO, editor. The Esophagus. Boston, Mass: Little, Brown & Company; 1992. p 763.)

CIRCULATION The arterial and venous blood supply to the esophagus is segmental. The upper esophagus is supplied by branches of the superior and inferior thyroid arteries, the midesophagus by branches of the bronchial and right intercostal arteries and descending aorta, and the distal esophagus by branches of the left gastric, left inferior phrenic, and splenic arteries.1-3 These vessels anastomose to create a dense network within the submucosa that probably accounts for the rarity of esophageal infarction. The venous drainage of the upper esophagus is through the superior vena cava, the midesophagus through the azygos veins, and the distal esophagus through the portal vein by means of the left and short gastric

On endoscopy the esophageal mucosa appears smooth and pink. Furthermore, the esophagogastric junction can be recognized by the presence of an irregular white Z-line (ora serrata) demarcating the interface between the lighter esophageal and the redder gastric mucosae. On biopsy, histology shows the esophageal mucosa to be lined by a nonkeratinized, stratified squamous epithelium (Fig. 41-5). This multilayered epithelium consists of three functionally distinct layers: stratum corneum, stratum spinosum, and stratum germinativum. The most lumen-oriented stratum corneum acts as a permeability barrier between luminal content and blood by having layers of pancake-shaped glycogen-rich cells connected laterally to each other by tight junctions and zonula adherens and having their intercellular spaces filled with a dense matrix of glycoconjugate material.7 The middle layer of stratum spinosum contains metabolically active cells with a spiny shape. The spiny shape is due to the numerous desmosomes connecting cells throughout the layer. Furthermore, this same desmosomal network maintains the structural integrity of the tissue. The basal layers of stratum germinativum contain cuboidal cells that occupy 10% to 15% of the epithelium’s thickness and are uniquely capable of replication.2 Consequently, basal cell hyperplasia, defined as basal cells occupying more than 15% of epithelial thickness, is common in gastroesophageal reflux disease, reflecting an increased rate of tissue repair2 (see Chapter 43). The esophageal epithelium contains a small number of other cell types including argyrophilic endocrine cells, melanocytes, lymphocytes, Langerhans cells (macrophages), and eosinophils. Neutrophils are not present in healthy epithelium.2 Below the epithelium is the lamina propria, a loose network of connective tissue within which are blood vessels and scattered lymphocytes, macrophages, and plasma cells (see Fig. 41-5). The lamina propria protrudes at intervals into the epithelium to form rete pegs or dermal papillae. Normally these protrude to less than 50% of the epithe­ lium’s thickness; when greater, it also is a recognized marker of gastroesophageal reflux disease.2 The muscularis mucosae is a thin layer of smooth muscle that separates the lamina propria above from the submucosa. Its functions are unclear.

SUBMUCOSA

The submucosa comprises a dense network of connective tissue, within which are blood vessels, lymphatic channels, neurons of Meissner’s plexus, and esophageal glands (see Fig. 41-2A). These glands, which vary as to number and distribution along the esophagus, consist of cuboidal cells organized as acini.8 They produce and secrete a lubricant, mucus, and factors such as bicarbonate and epidermal growth factor that are important for epithelial defense and repair. The secretions from these glands pass into tortuous collecting ducts that deliver them to the esophageal lumen.

667

668

Section V  Esophagus Cortical stimuli

Sensory pathways Motor pathways Nucleus solitarius Dorsal vagal nucleus Nucleus ambiguus Inspiratory center Medulla Phrenic nucleus Phrenic nerve

Pharynx

Vagus nerve

Vagus nerve

+Ach

Esophagus Figure 41-4.  Neural pathways of the esophagus. Extrinsic innervation is provided principally by the vagus nerve. Afferent vagal pathways carry stimuli to the nucleus solitarius, and efferent pathways originating in the dorsal vagal nucleus mediate esophageal peristalsis and lower esophageal sphincter relaxation. Ach, acetylcholine; NO, nitric oxide; VIP, vasoactive intestinal peptide. (From Mittal RK, Balaban DH. The esophagogastric junction. N Engl J Med 1997; 336:924.)

Myenteric plexus

Crural diaphragm

Phrenic nerve +Ach

–VIP –NO +Ach

Lower esophageal sphincter

larynx, trachea, bronchi, and lungs of the respiratory tract develop from a common tube.3 By gestational week 4, this tube, composed of endoderm, develops a diverticulum on its ventral surface that is destined to become the epithelium and glands of the respiratory tract (Fig. 41-6A to D). This diverticulum subsequently elongates, becomes enveloped by splanchnic mesenchyme (future cartilage, connective tissue, and smooth muscle), and buds off to become the primitive respiratory tract. Concomitantly, the lumen of the dorsal tube, the primitive foregut, fills with proliferating (ciliated-columnar) epithelium. By week 10, vacuoles appear and subsequently coalesce within the primitive foregut to reestablish the lumen. By week 16, the columnar epithelium lining the primitive foregut and future esophagus is replaced by stratified squamous epithelium, a process that is complete by birth.

Figure 41-5.  Esophageal epithelium. The human esophagus as shown on this biopsy specimen is lined by nonkeratinized stratified squamous epithelium. The cells of the surface (top) are long and flat and have a small nuclear-to-cytoplasmic ratio that contrasts with the cells of the basal layer (bottom), whose density, cuboidal shape, and large nuclear-to-cytoplasmic ratio account for their prominence. A subpopulation of these basal layer cells appears to have properties of esophageal stem cells.6a Rete pegs or dermal papillae containing elements of the lamina propria normally extend into the epithelium about one half the distance to the lumen. (Courtesy of Pamela Jensen, MD, Dallas, Tex.)

EMBRYOLOGY A brief review of the embryology of the upper digestive system is presented as a guide to understanding the origin of many of the developmental anomalies discussed in this chapter. In the developing fetus, the oropharynx and esophageal components of the gastrointestinal tract and the

DEVELOPMENTAL ANOMALIES Congenital anomalies of the esophagus are relatively common (1 in 3000 to 4500 live births) and are due to either transmission of genetic defects or intrauterine stress that impedes fetal maturation. Esophageal anomalies are common in premature infants, and 60% have other anomalies, reflected by the term VACTERL (formerly VATER), a mnemonic for the association of anomalies of the vertebral, anal, cardiac, tracheal, esophageal, renal, and limb systems. Common specific defects include patent ductus arteriosus, cardiac septal defects, and imperforate anus.9

ESOPHAGEAL ATRESIA AND TRACHEOESOPHAGEAL FISTULA

Esophageal atresia and tracheoesophageal fistulas are the most common developmental anomalies of the esophagus (Fig. 41-7). The former results from failure of the primitive

Chapter 41  Anatomy, Histology, Embryology, and Developmental Anomalies of the Esophagus Lung bud Primitive common upper digestive and respiratory tract

Foregut

A A

B

B Trachea

Tracheoesophageal septum Esophagus

Bronchial bud

D Lung

C

C

D

Stomach

Figure 41-6.  Developmental stages in the formation of separate respiratory and digestive systems. These systems are derived from a common tube of endoderm during embryogenesis. A, Single primitive tube. B, Formation of a lung bud in the fourth week. C, Elongation of the dorsal tube (primitive foregut) and lung bud and formation of a tracheoesophageal septum by four to six weeks. D, Separation of the primitive foregut from the tracheobronchial tree at six weeks.

E

Figure 41-7.  Esophageal atresia (A) and tracheoesophageal fistulas (B through E) are the most common developmental anomalies of the esophagus. In the most common tracheoesophageal fistula, the trachea communicates with the distal segment of the atretic esophagus (B). The next most common type is the H-type tracheoesophageal fistula, in which the trachea communicates with an otherwise normal esophagus (C). Tracheoesophageal fistulas in which the trachea communicates with both upper and lower segments of an atretic esophagus (D) or only the upper segment of an atretic esophagus (E) are rare. (Modified from The nonneoplastic esophagus. In: Fenoglio-Preiser CM, editor. Gastrointestinal Pathology. An Atlas and Text. 2nd ed. Philadelphia, Pa: Lippincott-Raven; 1999. p 31.)

Table 41-1  Genes Associated with Defects in Tracheoesophageal Development GENE

SPECIES

MOLECULAR FUNCTION

FOREGUT PHENOTYPES AND MALFORMATIONS

Foxf1 Gli2 & Gli3 Nkx2.1 Noggin RAR & 2

Mouse Mouse Mouse Mouse Mouse

Shh CHD7 FL13 MID1 MYCN SOX2 17q22-q23.3 deletion 22q11.2 deletion

Mouse Human Human Human Human Human Human Human

Forkhead family transcription factor Hedgehog pathway transcription factor Homeodomain-containing transcription Secreted BMP antagonist Retinoic acid receptor, nuclear hormone receptor superfamily Secreted hedgehog family ligand Chromodomain-containing DNA binding factor Hedgehog pathway transcription factor Microtubule-binding ubiquitin ligase Basic helix-loop-helix transcription factor HMG-box transcription factor Unknown Unknown

Narrow esophagus or TEF, abnormal lungs EA/TEF, abnormal lungs Esophagus and lung do not separate, abnormal lungs EA/TEF in 60% of heterozygotes Failure of foregut separation and abnormal cartilage development EA/TEF, abnormal lung development EA/TEF in some patients EA/TEF rare, predominantly tracheoesophageal clefts EA/TEF rare, predominantly tracheoesophageal clefts EA/TEF variable EA common, TEF rare EA/TEF rare EA/TEF rare

BMP, bone morphogenic protein; EA, esophageal atresia; HMG, high mobility group; TEF, tracheoesophageal fistula. Adapted from Que J, Choi M, Ziel JW, et al. Morphogenesis of the trachea and esophagus: Current players and new roles for noggin and Bmps. Differentiation 2006; 74:422-37.

foregut to recanalize and the latter from failure of the lung bud to separate completely from the foregut. Although the mechanisms are unclear, esophageal atresia and tracheoesophageal fistulas may result from genetic defects, such as those reported in mouse and humans (Table 41-1).10 In addition, experimental administration of the anticancer drug, adriamycin, into mouse or rat embryos commonly results in esophageal atresia and tracheoesophageal fistulas; and these

defects may be accompanied by other anomalies that comprise the VACTERL group.11-12 Esophageal atresia occurs as an isolated anomaly in only 7% of cases; the rest are accompanied by a form of tracheoesophageal fistula (distal-type fistula comprises 89% [see Fig. 41-7B] and the H-type fistula [see Fig. 41-7C] comprises 3% of cases).13 In isolated atresia, the upper esophagus ends in a blind pouch and the lower esophagus connects to the

669

670

Section V  Esophagus stomach (see Fig. 41-7A). The condition is suspected prenatally by the development of polyhydramnios (due to the inability of the fetus to swallow and so absorb amniotic fluid) or at birth by the regurgitation of saliva and a scaphoid (gasless) abdomen. Isolated atresia without a distal tracheoesophageal fistula results in a gasless abdomen because no pathway exists for inspired or swallowed air to enter the bowel. Furthermore, and as an indicator of high gastrointestinal obstruction, esophageal atresia results in the rapid onset of choking, coughing, and regurgitation on first feeding (Table 41-2). Once suspected, the diagnosis can be confirmed by failure to pass a nasogastric tube into the stomach and by a concurrent chest radiograph with air contrast in the upper esophageal segment (the air being introduced through a catheter positioned within the upper esophageal segment). In some instances, injection of 1  mL of barium into the obstructed segment helps with the diagnosis. As mentioned, esophageal atresia usually is associated with a tracheoesophageal fistula, most often the distal type (see Fig. 40-7B).13 Thus, the atretic upper esophagus ends in a blind pouch and the trachea communicates with the distal esophageal segment. The clinical presentation with this configuration is usually similar to isolated esophageal atresia with the additional risk of aspiration pneumonia from refluxed gastric contents entering the trachea through the fistula (see Table 41-2). Nonetheless, distinction between an isolated atresia and one associated with a distal tracheoesophageal fistula is straightforward because the com­munication between the trachea and the esophagus results in a gas-filled abdomen, as shown on plain radiographs. In some instances, confirmation of the type of configuration is obtained by esophagography with or without bronchoscopy. The three less common types of tracheoesophageal fistula are when the atretic upper esophagus communicates with the trachea, when both upper and lower segments of the atretic esophagus communicate with the trachea, and when an H-type fistula communicates with the trachea in a nonatretic esophagus (see Figs. 41-7E, D, and C, res­ pectively). Because these types have in common the communication between upper esophagus and trachea, they all manifest clinically with signs and symptoms of recurrent

(aspiration) pneumonia (see Table 41-2). Distinguishing among types, however, should not be difficult. Esophageal atresia accompanied by proximal tracheoesophageal fistula presents in infancy as recurrent pneumonia, and the presence or absence of bowel gas on a plain radiograph indicates whether an accompanying distal tracheoesophageal fistula exists. In contrast, in those with an H-type tracheoesophageal fistula without esophageal atresia, the diagnosis can be delayed until childhood or, at times, adulthood. Diagnosis of a suspected H-type fistula is usually made by esopha­ gography, but this may be difficult owing to the small size of some communications.14 In such cases, detection may be improved by ingestion of methylene blue and searching by bronchoscopy for the blue-stained fistula site. Treatment of esophageal atresia and tracheoesophageal fistulas is surgical, and the choice of procedure depends on the distance between the upper and lower esophageal segments. Short gaps permit end-to-end anastomosis, as do some long gaps after lengthening of the upper segment by either bougienage or intraoperative myotomy.13 If approximation of the two segments is not possible, the colon is interposed. The results of surgical correction of esophageal atresia are excellent when it exists as an isolated anomaly, with overall outcome determined by the gravity of accompanying genetic anomalies and by the birth weight of the infant.15 Long-term survival after successful repair of isolated esophageal atresia has steadily increased over the years and now approaches 90%. Notably, patients who survive until adulthood are at increased risk of developing gastroesophageal reflux disease (GERD),16-18 with the increased risk of GERD being due to abnormalities of esophageal motility and impaired acid clearance.19 In the largest follow-up to date, these patients, as adults, were found to have high rates of dysphagia (52%), reflux symptoms (63%), esophagitis (58%), Barrett’s metaplasia (11%), and strictures (42%)18; and anywhere from 6% to 45% require surgical fundoplication. Unfortunately, 15% to 30% of Nissen operations in these patients fail, usually resulting in reoperation.20 However, despite the increased incidence of Barrett’s metaplasia, a study of 272 such patients in Finland did not find an increased risk of esophageal or nonesophageal cancer.21

Table 41-2  Clinical Aspects of Esophageal Developmental Anomalies ANOMALY

AGE AT PRESENTATION

PREDOMINANT SYMPTOMS

DIAGNOSIS

TREATMENT

Isolated atresia

Newborns Newborns

Esophageal stenosis

Infants to adults

Duplication cyst

Infants to adults

Vascular anomaly

Infants to adults

Esophageal ring

Children to adults

Dysphagia

Esophageal web

Children to adults

Dysphagia

Esophagogram* Plain film: gasless abdomen Esophagogram* Plain film: gas-filled abdomen Esophagogram* Bronchoscopy† Esophagogram* Endoscopy† EUS* MRI/CT† Esophagogram* Angiography† MRI/CT/EUS Esophagogram* Endoscopy† Esophagogram* Endoscopy†

Surgery

Atresia + distal TE fistula H-type TE fistula

Regurgitation of feedings Aspiration Regurgitation of feedings Aspiration Recurrent pneumonia Bronchiectasis Dysphagia Food impaction Dyspnea, stridor, cough (infants) Dysphagia, chest pain (adults) Dyspnea, stridor, cough (infants) Dysphagia (adults)

Infants to adults

*Diagnostic test of choice. † Confirmatory test. ‡ Primary therapeutic approach. § Secondary therapeutic approach. CT, computed tomography; EUS, endoscopic ultrasonography; MRI, magnetic resonance imaging; TE, tracheoesophageal.

Surgery Surgery Dilation‡ Surgery§ Surgery Dietary modification‡ Surgery§ Dilation‡ Endoscopic incision§ Bougienage

Chapter 41  Anatomy, Histology, Embryology, and Developmental Anomalies of the Esophagus CONGENITAL ESOPHAGEAL STENOSIS

Esophageal stenosis is a rare anomaly, occurring in only 1 in every 25,000 live births.22 The stenotic segment varies from 2 to 20 cm in length and is usually located within the middle or lower third of the esophagus (Fig. 41-8A). The precise cause of the congenital stenosis is not entirely clear. Some patients (17% to 33%) have other associated anomalies, the most common being esophageal atresia (see Fig. 41-8B) and tracheoesophageal fistula.22 When resected, many stenotic walls contain tracheobronchial remnants (TBRs), which are sequestered respiratory tissue (hyaline cartilage, respiratory epithelium), suggesting its origin is incomplete separation of lung bud from primitive foregut.23 In other cases, stenosis results from fibromuscular hypertrophy associated with damage to the myenteric plexus with loss of the muscle-relaxing nitrinergic neural elements. A third subtype, a membranous diaphragm, is limited to the mucosa and does not involve the muscle layers.24 Although tight stenoses are symptomatic in infancy, most stenoses present with dysphagia and regurgitation in childhood when more solid food is ingested (see Table 41-2). The stenosis is best demonstrated by esophagography, which may reveal either an abrupt or tapered stricture. Dilatation of the esophagus proximal to the stenosis is commonly noted. Endoscopy may be of value by demonstrating normal mucosa in the stenotic region, helping to exclude an acquired cause for the stenosis. Endoscopic ultrasound (EUS) may show thickening of single or multiple layers of the esophageal wall and may demonstrate cartilaginous structures.25 Some patients improve after endoscopic-guided bougienage, although endoscopists should approach esophageal dilation carefully in these patients because chest pain and mucosal tears are commonly reported. Problematic stenoses require surgical resection of the involved segment. In general, congenital stenoses caused by TBRs rarely improve with bougienage, and therefore identification of this subtype by EUS may identify a group in need of surgical therapy.25 One novel surgical approach to this lesion is circular myectomy, a technique that involves stripping of the esophageal

A

B

muscle layers containing the TBRs with preservation of the mucosal layer. This has the advantage of avoiding many of the potential complications associated with primary repair and end-to-end esophageal anastomosis.26

ESOPHAGEAL DUPLICATIONS

Congenital duplications of the esophagus occur in 1 in 8000 live births.2 They arise as epithelial-lined outpouchings off the primitive foregut and evolve to produce either cystic or tubular structures that do not communicate with the esophageal lumen. Cysts account for 80% of the duplications and are usually single fluid-filled structures.2 They may be found attached to the esophagus or to the tracheobronchial tree and are usually located within the right posterior inferior mediastinum. Some cysts are discovered while asymptomatic, manifesting as a mediastinal mass on a chest radiograph or a submucosal lesion on an esophagogram (Fig. 41-9A). Others manifest with symptoms from compression of structures adjacent to the tracheobronchial tree (coughing, stridor, tachypnea, cyanosis, wheezing, or chest pain) and of structures adjacent to the esophageal wall (dysphagia, chest pain, or regurgitation) (see Table 41-2).27 The diagnosis of an esophageal duplication cyst is supported by the demonstration of a cystic mass on computed tomography (CT), magnetic resonance imaging (MRI), or EUS (see Fig. 41-9B).28 Benign duplication cysts are anechoic by EUS, and for cysts in which the appearance is hypoechoic, fine-needle aspiration and cytologic evaluation of cyst contents may exclude malignancy. This approach may be particularly helpful in asymptomatic patients in whom the cysts are discovered incidentally on imaging or endoscopy.28 However, only surgical excision for pathologic assessment can exclude a cystic neoplasm. Surgical excision is also favored because it has low morbidity.29 Rarely, large duplication cysts can manifest with acute life-threatening respiratory symptoms. In this circumstance, emergent decompression can be achieved by radiologic or endoscopically guided needle aspiration. The tubular esophageal duplication is far less common than its cystic counterpart. It is usually located within the

Figure 41-8.  Congenital esophageal stenosis. A, Barium esophagogram with a tapered narrowing in the distal esophagus and dilatation of the proximal esophagus. B, Barium esophagogram with an abrupt narrowing in the midesophagus (large arrows). The small arrow indicates the site of a previous repair for esophageal atresia. (A and B, From Usui N, Kamata S, Kawahara H, et al. Usefulness of endoscopic ultrasonography in the diagnosis of congenital esophageal stenosis. J Pediatr Surg 2002; 37:1744.)

671

672

Section V  Esophagus

A

C a

A

S

B

Figure 41-9.  Esophageal duplication cyst. A, Barium esophagogram shows extrinsic compression of the wall of the esophagus. B, Endoscopic ultrasonographic image shows the distortion of the esophageal wall created by the hypoechoic cyst (C) and the cyst’s relationship to other hypoechoic areas created by the aorta (A), azygos vein (a), and spine (S). (A, Courtesy of David Ott, MD, Winston-Salem, North Carolina; B, from Kimmey MB, Vilman P. Endoscopic ultrasonography. In: Yamada T, editor. Atlas of Gastroenterology. 3rd ed. Philadelphia, Pa: Lippincott Williams & Wilkins; 2003. p 1044.)

esophageal wall, parallels the true esophageal lumen, and, in contrast to duplication cysts, communicates with the true lumen at either or both ends of the tube.27 Tubular duplications usually cause chest pain, dysphagia, or regurgitation in infancy, and the diagnosis is established by esophagography or endoscopy. Reconstructive surgery is indicated for patients who are symptomatic.27

VASCULAR ANOMALIES

Intrathoracic vascular anomalies are present in 2% to 3% of the population. Only rarely do they produce symptoms of esophageal obstruction despite evident vascular compression on an esophagogram. In infancy, most intrathoracic vascular anomalies manifest as respiratory symptoms from compression of the tracheobronchial tree. Later in childhood or adulthood, however, these same abnormalities can produce dysphagia and regurgitation, owing to esophageal compression (see Table 41-2). Dysphagia lusoria is the term given for symptoms arising from vascular compression of the esophagus by an aberrant right subclavian artery.30 The right subclavian artery in this circumstance arises from the left side of the aortic arch and courses from the lower left to the upper right side posterior to the esophagus (Fig. 41-10A). In 20% of cases the artery courses anterior to the esophagus.31 It is estimated that arteria lusoria is present in 0.7% of the general population on the basis of autopsy studies. Typically the diagnosis is established by barium esophagogram, which shows the characteristic pencil-like indentation at the level of the third and fourth thoracic vertebrae (see Fig. 41-10B).30 Confirmation is by CT, MRI, arteriography, or EUS.31 Given

the frequency with which such lesions are asymptomatic, endoscopy or esophageal manometry may be desirable to exclude other causes of dysphagia. During endoscopy the right radial pulse may diminish or disappear from instrumental compression of the right subclavian artery. Esophageal manometry has demonstrated a high-pressure zone at the location of the aberrant artery.32 Symptoms usually respond to simple modification of the diet to meals of soft consistency and small size. When necessary, surgery relieves the obstruction by reanastomosing the aberrant artery to the ascending aorta.32

ESOPHAGEAL RINGS

The distal esophagus may contain two “rings,” the A and B (Schatzki’s) ring, that demarcate anatomically the proximal and distal borders of the esophageal vestibule (Fig. 41-11A). The A (muscular) ring is located at the proximal border. It is a broad (4 to 5 mm) symmetrical band of hypertrophied muscle that constricts the tubular esophageal lumen at its junction with the vestibule. In this location the A ring, which is covered by squamous epithelium, corresponds to the upper end of the lower esophageal sphincter.33 The A ring is rare, and because it varies in caliber on esophagography depending on the degree of esophageal distention, it is generally asymptomatic. Occasionally an A ring is found in association with dysphagia for solids and liquids (see Table 41-2).33 Symptomatic A rings can be treated by passage of a 50-French mercury-weighted esophageal dilator or by injection of botulinum toxin.34 The B ring, otherwise known as the mucosal or Schatzki’s ring, is very common, and found in 6% to 14% of subjects

Chapter 41  Anatomy, Histology, Embryology, and Developmental Anomalies of the Esophagus Right common carotid artery

Esophagus Left common carotid artery Left subclavian artery

Figure 41-10.  Dysphagia lusoria. A, Anatomic configuration of an aberrant right subclavian artery (lusorian artery) as it courses behind the esophagus from the aortic arch toward the right shoulder. B, Barium esophagogram shows the characteristic diagonal indentation of the eso­ phageal wall at the level of the third and fourth thoracic vertebrae. Dysphagia in association with such aberrant vasculature is known as dysphagia lusoria. (A, From Janssen M, Baggen MG, Veen HF, et  al. Dysphagia lusoria: Clinical aspects, manometric findings, diagnosis, and therapy. Am J Gastroenterol 2000; 95:1411; B, courtesy of David Ott, MD, Winston-Salem, North Carolina.)

Lusorian artery

Ascending aorta

Trachea Descending aorta

A

A

B

B

Figure 41-11.  Esophageal ring. A, Barium esophagogram of lower esophageal B ring. The distal B ring, or Schatzki’s ring, is a ring of mucosa localized to the squamocolumnar junction. Below the B ring is a hiatal hernia. The hernia is visualized as a small sac between the B ring above and the diaphragm below. B, Endoscopic view of Schatzki’s ring. (A, Courtesy of David Ott, MD, Winston-Salem, North Carolina; B, courtesy of John D. Long, MD, WinstonSalem, North Carolina.)

673

674

Section V  Esophagus having a routine upper gastrointestinal series.35 A recent review of more than 10,000 upper endoscopies found a Schatzki ring in 4% of cases.36 On barium study it is always found in association with a hiatal hernia and is recognized as a thin (2-mm) membrane that constricts the esophageal lumen at the junction of the vestibule and gastric cardia (see Fig. 41-11A). The Schatzki ring has squamous epithelium on its upper surface and columnar epithelium on its lower surface and so demarcates the squamocolumnar junction. The ring itself is composed of only mucosa and submucosa; there is no muscularis propria. Schatzki’s rings can be congenital or acquired, and a relationship to gastroesophageal reflux disease is likely (see Chapter 43).35 Most B rings are asymptomatic, yet when the diameter of the esophageal lumen is narrowed to less than or equal to 13 mm, rings commonly are the cause of intermittent dysphagia for solids or unheralded acute solid-food impactions (see Table 41-2).37 Identification of symptomatic rings on esophagography or endoscopy is generally not difficult (see Fig. 41-11B), although attention should be paid to adequately distend the distal esophagus.35 In some instance, the obstructing ring is best demonstrated radiographically by its ability to trap a swallowed marshmallow or a barium tablet. Asymptomatic B rings require no treatment, and those producing dysphagia are effectively treated by passage of either a single, large (≥50-French), mercury-weighted dilator or a series of such dilators of progressively larger diameter.38 Early studies reported that 32% of patients required repeat dilation after 1 year.35 More recent studies report much lower rates (13%), perhaps due to the more routine use of both larger dilators and a short course of postdilation antireflux therapy.39 In a randomized placebo-controlled study of 44 patients with symptomatic Schatzki’s rings, maintenance therapy with omeprazole resulted in a 40% reduction in the need for redilation after a mean follow-up of 35 months.40 Symptomatic rings that are refractory to dilation have been successfully treated by endoscopic means using electrocautery incision.41 A randomized, controlled trial of standard bougie dilation versus electrocautery incision for

A

symptomatic Schatzki rings has demonstrated that the two therapies have comparable initial success rates but that endoscopic incision had a longer duration of symptom resolution.42 A syndrome in which multiple esophageal rings are found has been described.43 The term “corrugated ringed esophagus” generally implies a condition in which the muscular rings are concentric and evenly spaced over a long segment of the esophagus, usually starting proximally and extending, in some instances, the entire length of the organ. These rings may not be apparent on barium studies but on endoscopy persist despite maximal air insufflation. Another defining characteristic of the corrugated ringed esophagus is that dilation is difficult with mucosal tears and perforations being common. For this reason, it is recommended that dilation in patients with solid-food dysphagia be limited to a maximum diameter of a 40-French bougie. Support for the corrugated ringed esophagus being a congenital anomaly is based on the large male predominance (75% of reported cases) and history of solid-food dysphagia, often with impactions, beginning in early childhood.43 Several studies have proposed that many of these cases are a manifestation of eosinophilic esophagitis (see Chapter 27), and a few studies suggest an association with GERD. The former is readily diagnosed by esophageal biopsy showing greater than 15 to 20 eosinophils per high-power field and the latter by response to acid suppression therapy.

ESOPHAGEAL WEBS

Esophageal webs are developmental anomalies characterized by one or more thin horizontal membranes of stratified squamous epithelium within the upper (cervical) esophagus and midesophagus. Unlike rings these anomalies rarely encircle the lumen but instead protrude from the anterior wall, extending laterally but not to the posterior wall (Fig. 41-12A and B). Webs are common in the cervical esophagus and are best demonstrated on an esophagogram with the lateral view. In up to 5% of cases they are identified in an asymptomatic state, but when they are symptomatic they

B

Figure 41-12.  Esophageal webs. A, Barium esophagogram of a cervical esophageal web seen on the lateral view as a thin membrane protruding from the anterior esophageal wall. Webs, unlike rings, often incompletely encircle the esophageal lumen. B, Endoscopic view of a cervical esophageal web. (A, Courtesy of David Ott, MD, Winston-Salem, North Carolina; B, courtesy of John D. Long, MD, Winston-Salem, North Carolina.)

Chapter 41  Anatomy, Histology, Embryology, and Developmental Anomalies of the Esophagus

A

B

Figure 41-13.  Inlet patch. A, Endoscopic view of heterotopic gastric mucosa in the cervical esophagus (“inlet patch”). B, Microscopic view of an inlet patch shows glandular epithelium with parietal cells (right) adjacent to normal esophageal squamous epithelium (left). (A, From Avidan B, Sonnenberg A, Chejfec G, et al. Is there a link between cervical inlet patch and Barrett’s esophagus? Gastrointest Endosc 2001; 53:717; B, courtesy of Pamela Jensen, MD, Dallas, Tex.)

cause dysphagia for solids (see Table 41-2).44 Webs are fragile membranes and so respond well to esophageal bougienage with mercury-weighted dilators. An association among cervical esophageal webs, dysphagia, and iron deficiency anemia in adults has been described as the Plummer-Vinson or Paterson-Kelly syndrome.44 The syndrome, although uncommon, occurs primarily in women. Recent reports have shown an association between Plummer-Vinson syndrome and celiac disease.45 It is an important syndrome because it identifies a group of patients at increased risk for squamous carcinoma of the pharynx and esophagus.44 Correction of iron deficiency in PlummerVinson syndrome may result in resolution of the associated dysphagia as well as disappearance of the web.44

HETEROTOPIC GASTRIC MUCOSA (INLET PATCH)

The inlet patch refers to the appearance on endoscopy of a small (0.5 to 2 cm) distinctive, velvety red island of heterotopic gastric mucosa amid a lighter pink squamous mucosa, generally localized immediately below the upper esophageal sphincter (Fig. 41-13A). When sought, an inlet patch is found in up to 10% of endoscopies, and biopsy specimens reveal gastric fundic- or antral-type mucosa (see Fig. 41-13B).46 The fundic-type mucosa contains chief and parietal cells and thus in some specimens retains the capacity for acid secretion.47 Similar to gastric mucosa in the stomach, the inlet patch may be infected with Helicobacter pylori.48 However, inlet patches are usually asymptomatic and unassociated with disease and thus require no treatment. A possible association with globus pharyngeus was suggested in a recent pilot study in which this symptom was improved after ablation of inlet patches using argon plasma coagulation.49 In rare instances, an inlet patch is found in association with an esophageal web or stricture50 or ulcer, the latter resulting in bleeding or perforation.46 In addition, although adenocarcinoma arising in an inlet patch is a rare complication, a total of 24 such cases have been reported in the literature.46

KEY REFERENCES

Achildi O, Grewal H. Congenital anomalies of the esophagus. Otolaryngol Clin North Am 2007; 40:219-44. (Ref 3.) Atmatzidis K, Papaziogas B, Pavlidis T, et al. Plummer-Vinson syndrome. Dis Esophagus 2003; 16:154-7. (Ref 44.) Cioffi U, Bonavina L, De Simone M, et al. Presentation and surgical management of bronchogenic and esophageal duplication cysts in adults. Chest 1998; 113:1492-6. (Ref 29.) Deurloo JA, Ekkelkamp S, Schoorl M, et al. Esophageal atresia: Historical evolution of management and results in 371 patients. Ann Thorac Surg 2002; 73:267-72. (Ref 13.) Hirano I, Gilliam J, Goyal RK. Clinical and manometric features of the lower esophageal muscular ring. Am J Gastroenterol 2000; 95:43-9. (Ref 33.) Jalil S, Castell DO. Schatzki’s ring. A benign cause of dysphagia in adults. J Clin Gastroenterol 2002; 35:295-8. (Ref 35.) Janssen M, Baggen MGA, Veen HF, et al. Dysphagia lusoria: Clinical aspects, manometric findings, diagnosis, and therapy. Am J Gastroenterol 2000; 95: 411-16. (Ref 30.) Mittal RK, Balaban DH. The esophagogastric junction. N Engl J Med 1997; 336:924-32. (Ref 4.) Que J, Choi M, Ziel JW, et al. Morphogenesis of the trachea and esophagus: Current players and new roles for noggin and Bmps. Differentiation 2006; 74:422-37. (Ref 10.) Sgouros SN, Vlachogiannakos J, Karamanolis G, et al. Long-term acid suppressive therapy may prevent the relapse of lower esophageal (Schatzki’s) rings: A prospective, randomized, placebo-controlled study. Am J Gastroenterol 2005; 100:1929-34. (Ref 40.) Skandalakis JE, Ellis H. Embryologic and anatomic basis of esophageal surgery. Surg Clin North Am 2000; 80:85-155. (Ref 1.) Takamizawa S, Tsugawa C, Mouri N, et al. Congenital esophageal stenosis: Therapeutic strategy based on etiology. J Pediatr Surg 2002; 37:197-201. (Ref 25.) Taylor ACF, Breen KJ, Auldist A, et al. Gastroesophageal reflux and related pathology in adults who were born with esophageal atresia: A long-term follow-up study. Clin Gastroenterol Hepatol 2007; 5:7026. (Ref 18.) Von Rahden BHA, Stein HJ, Becker K, et al. Heterotopic gastric mucosa of the esophagus: Literature-review and proposal of a clinicopathologic classification. Am J Gastroenterol 2004; 99:543-51. (Ref 46.) Wills JC, Hilden K, DiSario JA, Fang JC. A randomized, prospective trial of electrosurgical incision followed by rabeprazole versus bougie dilation followed by rabeprazole of symptomatic esophageal (Schatzki’s) rings. Gastrointest Endosc 2008; 67:808-13. (Ref 42.) Full references for this chapter can be found on www.expertconsult.com.

675

CHAPTE R

42 Esophageal Neuromuscular Function and Motility Disorders Peter J. Kahrilas and John E. Pandolfino

CHAPTER OUTLINE Esophageal Motor and Sensory Function  677 Oropharynx and Upper Esophageal Sphincter  677 The Pharyngeal Swallow  679 Esophagus  680 Esophagogastric Junction  682 Esophageal Sensation  685

The esophagus is a hollow muscular tube with a sphincter at each end joining the hypopharynx above to the stomach below. It ultimately functions to transport food and fluid between these endpoints, otherwise maintaining itself empty. Despite this simplicity of function, the control mechanisms of the esophagus are far from simple. One complexity of it is that the esophagus encompasses the anatomic and physiologic transition between two distinct regions: the oropharynx and the gut. The oropharynx is constructed of striated muscle, controlled by the cerebral cortex and medulla, and capable of precise tactile sensation. The distal esophagus is composed of smooth muscle, controlled by the vagus nerve and enteric nervous system, and only capable of notoriously imprecise sensation. Although there is a gradual transition between these endpoints, motor function in the oropharynx and esophageal body are quite distinct. With that in mind, the ensuing discussion includes selected aspects of pharyngeal, gastric, and diaphragmatic function that are inextricably entwined with esophageal function. Similarly, in the interest of presenting swallowing as the complex, highly integrated motor response that it is, the discussion of swallowing encompasses the oropharyngeal swallow and central nervous system control mechanism, as well as strictly esophageal aspects of swallowing.

ESOPHAGEAL MOTOR AND SENSORY FUNCTION OROPHARYNX AND UPPER ESOPHAGEAL SPHINCTER

The oral cavity and pharynx are critically involved in the task of ingesting food. Within the oral cavity, the lips, teeth, hard palate, soft palate, mandible, floor of the mouth, and tongue serve functions in masticating, containing, and forming food into a bolus suitable for transfer to the pharynx.

Esophageal Motility Disorders  686 Epidemiology  686 Pathogenesis  687 Clinical Features  691 Differential Diagnosis  692 Diagnostic Methods  693 Treatment  700

The pharynx is a hollow cavity separated into three segments (Fig. 42-1): nasopharynx, oropharynx (oral pharynx), and hypopharynx (laryngeal pharynx). The nasopharynx extends from the base of the skull to the distal edge of the soft palate. Although not part of the alimentary tract, muscles in the nasopharynx contribute to elevating the soft palate and sealing the nasopharynx during swallowing, preventing nasopharyngeal regurgitation. The oropharynx extends from the soft palate to the base of the tongue. The inferior margin of the oropharynx is demarcated by the valleculae anteriorly and the mobile tip of the epiglottis posteriorly. The hypopharynx extends from the valleculae to the inferior margin of the cricoid cartilage and includes the upper esophageal sphincter (UES). Musculature of the soft palate, tongue, and pharynx all participate in swallowing. These intrinsic muscles collapse and shorten the pharyngeal lumen and then expel its contents into the esophagus. Additionally, extrinsic muscles that elevate and pull the pharynx forward alter the shape of the pharynx and seal the airway during swallowing. Within the nasopharynx, the levator veli palatini, tensor veli palatini, and palatoglossus act to raise and tense the soft palate and uvula, sealing the nasopharynx from the oropharyngeal cavity. Posteriorly, the stylohyoid, styloglossus, palatopharyngeus, stylopharyngeus, and digastric posterior elevate the larynx, while the geniohyoid, mylohyoid, digastric anterior, and thyrohyoid pull the larynx forward and contribute to UES opening. The intrinsic muscles of the pharynx are the superior, middle, and inferior pharyngeal constrictors (see Fig. 42-1). The constrictors overlap and insert into a collagenous sheet, the buccopharyngeal aponeurosis. The superior pharyngeal constrictor arises from the pterygoid hamulus, pterygomandibular raphe, mandible, and tongue; passes posteromedially; and inserts to the posterior raphe. The middle constrictor arises from the hyoid bone and stylohyoid ligament, passes posteromedially, and inserts in the posterior

677

678

Section V  Esophagus

Soft palate Hard palate Oral cavity Tongue

Mylohyoid ms. Thyrohyoid membrane Vocal cord Transverse arytenoid ms. Cricothyroid membrane Cricoid cartilage

Lateral pterygoid plate Oral pharynx Buccinator Valleculae Digastric (ant. belly) Hyoid bone Epiglottis Laryngeal pharynx (hypopharynx) Mylohyoid Stylohyoid Esophagus Hyoid bone Thyroid cartilage Cricothyroid membrane Cricoid cartilage

A

Digastric (post. belly) Superior constrictor Styloid process Styloglossus Stylohyoid ligament Glossopharyngeus Stylopharyngeus Middle constrictor Hyoglossus Thyrohyoid membrane Inferior constrictor Cricopharyngeus Esophagus

B

Figure 42-1.  Anatomy of the pharynx. A, Sagittal representation of the pharynx showing the musculoskeletal structures involved in swallowing. Note that the esophagus is collapsed and empty at rest. In the course of a swallow the laryngeal inlet is sealed and the mouth of the esophagus is opened by a highly coordinated muscular activity. B, Cutaway diagram of the musculature of the pharynx. Note that the hyoid bone is positioned as a fulcrum and is instrumental in directing anterior, superior traction forces critical to closing the larynx and opening the esophageal inlet during swallow. ms., muscle. (Reprinted from Kahrilas PJ, Frost F. Disorders of swallowing and bowel motility. In: Green D, editor. Medical Problems of the Chronically Disabled. Rockville, Md.: Aspen Publishers, 1990. pp 11-37.)

median raphe. The inferior constrictor is composed of the thyropharyngeus (superior part) and the cricopharyngeus (inferior part). The thyropharyngeus arises from the thyroid cartilage, passes posteromedially, and inserts in the median raphe. The cricopharyngeus has superior and inferior components, each of which arise bilaterally from the sides of the cricoid lamina; the superior fibers course posteromedially to the median raphe and the inferior fibers loop around the esophageal inlet without a median raphe. Killian’s triangle, a triangular area of thin muscular wall is formed posteriorly between these components of the cricopharyngeus and is the most common site of origin for pharyngeal pulsion diverticuli. The pharynx also contains five single or paired cartilages: the epiglottic, arytenoid, cuneiform, corniculate, and cricoid (see Fig. 42-1). The spaces formed between the lateral insertion of the inferior constrictor and the lateral walls of the thyroid cartilage are the pyriform sinuses that end inferiorly at the cricopharyngeus muscle, separating the pharynx from the esophagus. The larynx and trachea are suspended in the neck between the hyoid bone superiorly and the sternum inferiorly. A number of muscles, categorized as the laryngeal strap muscles, contribute to this suspension and, together with the intrinsic elasticity of the trachea, permit the larynx to be raised and lowered. The hyoid bone also serves as the base for the tongue that rests on it. Laryngeal movement is crucial to the successful enactment of the swallow response because the laryngeal inlet is closed and physically removed from the bolus path in the course of a swallow. Failure to achieve this synchronized laryngeal movement can result in aspiration. The pharyngeal muscles are densely innervated with motor fibers coming from nuclei of the trigeminal, facial, glossopharyngeal, and hypoglossal nuclei as well as nucleus

ambiguus and spinal segments C1 to C3. The innervation of the major pharyngeal muscles is as follows: mylohyoid, tensor veli palatini, and anterior digastric muscles (trigeminal nerve); stylohyoid and posterior part of the digastric (facial nerve); stylopharyngeus (glossopharyngeal nerve); levator veli palatini, palatopharyngeus, salpingopharyngeus, thryroarytenoid, arytenoid, pharyngeal constrictors, and cricopharyngeus (vagus nerve); thyrohyoid, geniohyoid, and tongue (hypoglossal nerve).1 Nucleus ambiguus is the vagal nucleus responsible for innervation of the striated muscle of the pharynx, larynx, and esophagus.2 All motor neurons within the nucleus ambiguus seem to participate in swallowing with those innervating the esophagus situated rostrally and those innervating the larynx more caudally.3 The muscular components of the UES are the cricopharyngeus, adjacent esophagus, and adjacent inferior constrictor (see Chapter 41, Fig. 41-3). The cricopharyngeus contributes the zone of maximal UES pressure, which is about 1 cm in length.4 The closed sphincter has a slit-like configuration with the cricoid lamina anterior and the cricopharyngeus making up the lateral and posterior walls. Not surprisingly, resting UES pressure is markedly asymmetrical with greatest values anteriorly and posteriorly.5 Neural input via vagal trunks originating in nucleus ambiguus to the UES is required for maintenance of high resting pressure and for the coordination of relaxation with swallowing.6 Cessation of motor neuron firing, or administration of curare, causes relaxation, whereas increased spike activity increases tone. Vagal transection abolishes contractile activity in the cricopharyngeus and inferior pharyngeal constrictor muscles.7 Manometric evaluation of UES function is difficult because it is a short, complex anatomic zone that moves

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders mm Hg 100

A

Pharyngeal contraction

cm 0

B

90

5

80

Proximal striated muscle segment

70 60 50

10

UES opening UES opening

15

40

Common cavity

30 25 20 15 10 5 0 -5 -12

20

25 tLESR

LES relaxation

30

35 0

3

6

9

12

Time (sec)

briskly during swallowing. Furthermore, UES pressure measurement is heavily influenced by recording methodology due to its marked asymmetry and to the fact that pharyngeal stimulation by the measurement device stimulates sphincter contraction. Thus, it is not surprising that there is great variability in reported “normal” UES pressure ranges and it is currently impossible to define a meaningful normal range.8 UES relaxation during swallowing also poses substantial recording challenges, making for great variability in technique and interpretation among centers. However, high-resolution manometry using solid state technology has become clinically available, and this technology permits accurate tracking of UES relaxation and intrabolus pressure changes during swallowing (Fig. 42-2). The main function of the UES is to maintain closure of the proximal end of the esophagus unless opening is required for either swallowing or belching.9 It constitutes an additional barrier to refluxed material entering the pharynx and prevents air from entering the esophagus by contracting in synchrony with inspiration. Inspiratory augmentation is most evident during periods of low UES pressure and can be exaggerated in individuals experiencing globus sensation.10 Balloon distention of the esophagus stimulates UES contraction with the effect being more pronounced with more proximal balloon positions.11 However, when the distension pattern of gas reflux is simulated using a cylindric bag or rapid air injection into the esophagus, UES relaxation rather than contraction occurs.4 Belch-induced UES relaxation is also associated with glottic closure.12 Stress augments UES pressure13 and anesthesia14 or sleep15 virtually eliminates it. Neither experimental acid perfusion of the esophagus16 nor spontaneous gastroesophageal acid reflux alters continuously recorded UES pressure in either normal volunteers15 or in individuals with peptic esophagitis.16

THE PHARYNGEAL SWALLOW

The oral phase of swallowing is largely voluntary and highly variable. Disorders of the oral phase of swallowing occur

15

18

Figure 42-2.  High-resolution manometric analysis of upper esophageal sphincter (UES) function. High-resolution manometry can quantify the completeness of lower esophageal sphincter (LES) relaxation, intrabolus pressure, and the duration of UES opening without movement artifact. A, UES opening during swallowing. UES relaxation pressure during swallowing is typically less than 10 mm Hg. UES closure occurs as the pharyngeal contraction wavefront propagates through to the proximal esophagus. B, UES opening during belching. During a transient LES relaxation (tLESR), a common cavity develops that leads to UES relaxation and gas venting. Note the lack of pharyngeal activity and the esophageal shortening during the tLESR.

with many conditions characterized by global neurologic dysfunction such as head trauma, cerebral tumors, or chorea. Detailed discussion of these conditions can be found in texts on swallow evaluation and therapy.17,18 The pharyngeal swallow is the complex coordinated contraction that transfers oral contents into the esophagus. A typical individual swallows about 600 times a day without giving significant thought or effort to the activity.19 Afferent sensory fibers capable of triggering the pharyngeal swallow travel centrally via the internal branch of the superior laryngeal nerve (from the larynx) and via the glosso­ pharyngeal nerve (from the pharynx).20 These sensory fibers converge before terminating in the medullary swallow center.21 Although understood physiologically as the patterned activation of motor neurons and their corresponding motor units, swallowing is clinically evaluated in mechanical terms and best evaluated by videofluoroscopic or cinera­ diographic analysis. The pharyngeal swallow rapidly re­ configures pharyngeal structures from a respiratory to an alimentary pathway and then reverses this reconfiguration within one second. The pharyngeal swallow response can be dissected into several closely coordinated actions: (1) nasopharyngeal closure by elevation and retraction of the soft palate, (2) upper esophageal sphincter (UES) opening, (3) laryngeal closure, (4) tongue loading (ramping), (5) tongue pulsion, and (6) pharyngeal clearance. Precise co­ ordination of these actions is an obvious imperative and to some degree the relative timing of these events is affected either by volition or by the volume of the swallowed bolus (Fig. 42-3).22 The most fundamental anatomic reconfiguration required to transform the oropharynx from a respiratory to a swallow pathway is to open the inlet to the esophagus and seal the inlet to the larynx. These events occur in close synchrony. UES opening occurs by laryngeal elevation and anterior traction via the hyoid axis.23,24 The mechanical determinants of laryngeal vestibule closure are laryngeal elevation and anterior tilting of the arytenoid cartilages against the base

679

680

Section V  Esophagus Glossopalatal junction opening

Velopharyngeal junction closure

Laryngeal vestibule closure

UES opening

Time (sec) 0

0.2 Pharyngeal reconfiguration

–0.4' Volume dependent 1 mL = 0.0s 20 mL = 0.2s

–0.2'

0'

Pharyngeal clearance and offset of reconfiguration

Figure 42-3.  Time line showing the coordination and volume-induced modifications in the timing of events within a pharyngeal swallow. Each horizontal bar depicts the period during which one of the oropharyngeal valves is in its swallow configuration as opposed to its configuration during respiration. Events at the onset and offset of pharyngeal reconfiguration bear a fixed time relationship to each other regardless of swallow bolus volume. The stereotypy of these phases is demonstrated by referencing onset events from time 0, counting forward, and offset events from time 0′, counting forward or backward, respectively. This timing scheme defines the volume dependent, middle portion of the time line (shaded) that has a value of 0 for 1-mL swallows and 0.2 second for 20-mL swallows. Thus, the alteration in the timing of the swallow response with larger-volume swallows occurs by prolonging the persistence of pharyngeal reconfiguration without changing the synchrony of events at onset or offset. (Reprinted from Kahrilas PJ, Lin S, Chen J, Logemann JA. Oropharyngeal accommodation to swallow volume. Gastroenterology 1996; 111:297-306; with permission.)

of the epiglottis.17 Thus, analyzing the efficacy of either of these events inevitably focuses on laryngeal elevation.25 UES relaxation occurs at roughly the same degree of laryngeal elevation regardless of swallow volume, and precedes sphincter opening by about 0.1 second. It is critical to recognize the distinction between UES relaxation and UES opening. UES relaxation occurs due to cessation of excitatory neural input while the larynx is elevating. Once the larynx is elevated, UES opening results from traction on the anterior sphincter wall caused by contraction of the suprahyoid and infrahyoid musculature that also results in a characteristic pattern of hyoid displacement.23,24 The two main determinants of bolus transport out of the oropharynx are the action of the tongue and of the pharyngeal constrictors. Tongue motion varies substantially with swallow conditions and propels most of the bolus into the esophagus prior to the onset of the pharyngeal contraction with larger volume swallows.26 On the other hand, the pro­ pagated pharyngeal contraction has similar propagation and vigor regardless of bolus volume.27 However, the pro­ pagated pharyngeal contraction is more involved with the process of clearance than of bolus propulsion; it strips the last residue from the pharyngeal walls. Upper esophageal sphincter closure coincides with the arrival of the pro­ pagated pharyngeal contraction as evident by the fixed time relationship between these events.24 However, the contractile activity of the sphincter has an added dimension as well, exhibiting increased electromyographic activity during laryngeal descent.28 The magnitude of this postdeglutitive contraction is further augmented by either sphincteric or proximal esophageal distention resulting in a grabbing effect such that the sphincter and laryngeal descent complement each other to clear residue from the hypopharynx.29 This clearing function probably acts to minimize the risk of post-swallow aspiration by preventing residual material from adhering to the laryngeal inlet when respiration resumes.

ESOPHAGUS

The esophagus is a 20- to 22-cm muscular tube with a wall composed of skeletal and smooth muscle (see Chapter 41, Fig. 41-1). The proportion of each muscle type is species dependent, but in man, the proximal 5% is striated, the middle 35% to 40% is mixed with an increasing proportion of smooth muscle distally, and the distal 50% to 60% is entirely smooth muscle.30 The bundles of the outer (longitudinal) muscle arise from the cricoid cartilage receiving slips from the cricopharyngeus and pass dorsolaterally to fuse posteriorly about 3 cm distal to the cricoid cartilage. This arrangement results in a posterior triangular area devoid of longitudinal muscle, Laimer’s triangle. Distal to Laimer’s triangle the longitudinal muscles form a continuous sheath of uniform thickness around the esophagus.31 The adjacent, inner muscle layer is formed of circular or, more precisely, helical muscle also forming a sheath of uniform thickness throughout the length of the esophageal body. The overlapping helices exhibit decreasing degree of helicity moving distally ranging from 60 degrees in the proximal esophagus to nearly 0 degrees in the most distal esophagus.32 The extrinsic innervation of the esophagus is via the vagus nerve (see Chapter 41). Fibers innervating the striated muscle are axons of lower motor neurons with cell bodies situated in nucleus ambiguus, whereas the smooth muscle esophagus is innervated by the dorsal motor nucleus of the vagus.33,34 Efferent nerve fibers reach the cervical esophagus by the pharyngoesophageal nerve35 and histologic studies show that vagal efferents synapse directly on striated muscle neuromuscular junctions.36,37 The vagus nerves also provide sensory innervation. In the cervical esophagus this is via the superior laryngeal nerve with cell bodies in the nodose ganglion, whereas in the remainder of the esophagus sensory fibers travel via the recurrent laryngeal nerve or, in the most distal esophagus, via the esophageal branches of the vagus. Histologic studies demonstrate many free nerve endings in

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders the mucosa, submucosa, and muscular layers.2,38 Additionally, a few encapsulated structures resembling spindles have been described in humans. These vagal afferents are strongly stimulated by esophageal distention. The esophagus also contains an autonomic nerve network, the myenteric plexus, located between the longitudinal and circular muscle layers.39 Myenteric plexus neurons are sparse in the proximal esophagus and their function in that region is unclear because the striated muscle is directly controlled by somatic motor fibers. On the other hand, the thoracic esophagus receives innervation from preganglionic neurons in the dorsal motor nucleus of the vagus that then synapse in myenteric plexus ganglia, relay neurons between the vagus and the smooth muscle. The ganglia of the myenteric plexus are more numerous in the distal esophagus than in the striated muscle region, but throughout, they are still far less dense and smaller than in other regions of the gut.40,41 A second nerve network, the submucosal or Meissner’s plexus, is situated between the muscularis mucosa and the circular muscle layer, but this is exceedingly sparse with few ganglia in the human esophagus.41

Properties of Esophageal Peristalsis

The esophagus does not normally exhibit spontaneous contractions and its intraluminal pressure closely reflects pleural pressure, becoming negative during inspiration. However, swallowing or focal distention initiates peristalsis. Primary peristalsis is that which is initiated by a swallow and traverses the entire length of the esophagus; secondary peristalsis can be elicited in response to focal esophageal distention with air, fluid, or a balloon, beginning at the point of distention.42 The mechanical correlate of peristalsis is a stripping wave that milks the esophagus clean from its proximal to distal end. The propagation of the stripping wave corresponds closely with that of the manometrically recorded contraction such that the point of the inverted “V” seen fluoroscopically at each esophageal locus occurs concomitantly with the upstroke of the pressure wave. The likelihood of achieving complete esophageal emptying from the distal esophagus is inversely related to peristaltic amplitude such that emptying becomes progressively impaired with peristaltic amplitudes of 30 mm Hg or less.43 Another essential feature of peristalsis is deglutitive inhibition. A second swallow, initiated while an earlier peristaltic contraction is still progressing in the proximal esophagus, causes rapid and complete inhibition of the contraction induced by the first swallow.44 If the first peristaltic contraction has reached the distal esophagus, it may proceed distally for a few seconds after the second swallow, but its amplitude then diminishes until it disappears.45 Deglutitive inhibition in the distal esophagus is attributable to hyperpolarization of the circular smooth muscle and is mediated via inhibitory ganglionic neurons in the myenteric plexus. Deglutitive inhibition can be demonstrated experimentally in the esophagus by creation of an artificial high-pressure zone with an intraluminal balloon.46 The artificial highpressure zone is created by distending the esophageal lumen with a balloon and recording intraluminal pressure between the balloon and the esophageal wall. Once the high pressure zone is established in the normally flaccid tubular esophagus, deglutitive inhibition is evident by relaxation of the artificial high pressure zone commencing concurrently with the swallow. The physiologic control mechanisms governing the striated and smooth esophageal musculature are distinct. The striated muscle of the esophagus receives exclusively excitatory vagal innervation and its peristaltic contraction results from sequential activation of motor units in a craniocaudal

sequence. These fibers release acetylcholine and stimulate nicotinic cholinergic receptors on the motor endplates of the striated muscle cells. Physiologic evidence of this arrangement was provided by an ingenious series of experiments using the nerve suture technique.2 In these experiments, the vagal branch innervating the esophagus was severed and the central end anastomosed to the peripheral end of the also severed spinal accessory nerve. Thus, after a period of nerve regeneration, the vagal branch effectively innervates the sternocleidomastoid and trapezius muscles. Occurrence of excitatory vagal discharges can then be surmised from the contractile activity of these readily accessible muscles. Nerve suture experiments demonstrated several properties of vagal control of esophageal striated muscle: (1) vagal efferent fibers exhibit no spontaneous discharge but fire in spike bursts during primary or secondary peristalsis; (2) once activated, vagal fibers innervating different levels of the esophagus fire sequentially, demonstrating peristaltic programming by the medullary swallow center; (3) peristaltic vagal motor discharges are potentiated by stimulation of afferent fibers from the esophagus (designed to mimic the effect of a bolus being pushed ahead of the contraction); (4) peristaltic vagal motor discharges are stronger during primary than secondary peristalsis; and (5) vagal motor fibers are inhibited during the pharyngeal stage of deglutition or after distention of a proximal esophageal segment supporting the notion that deglutitive inhibition has a central origin. Thus, there is substantial evidence that peristalsis in the striated muscle esophagus is controlled by the swallowing center of the medulla in much the same way as is the oropharyngeal musculature. The vagus nerves also exhibit control of primary peristalsis in the smooth muscle esophagus. Deviation of a swallowed bolus at the level of the cervical esophagus (thereby eliminating the potential for bolus-initiated afferent feedback) does not eliminate the primary peristaltic contraction in the distal esophagus.47 Furthermore, primary peristalsis of the smooth muscle persists even after curarization. Because curarization paralyzes the oropharyngeal and cervical esophagus, the persistence of distal peristalsis in these experiments strongly suggests that it is triggered by the medullary swallowing center that can elicit the entire motor sequence of primary peristalsis without receiving afferent feedback. The mechanism of vagal control of the smooth muscle esophagus is more complex than that of the striated muscle because vagal fibers synapse on myenteric plexus neurons rather than directly on muscle cells. Experimentally, vagal stimulation either excites or inhibits esophageal musculature depending on the stimulation parameters used.48,49 In the esophagus of the opossum, vagal or swallow-induced stimulation causes depolarization with superimposed spikes on longitudinal muscle but an initial hyperpolarization followed by depolarization and spike burst on circular muscle.50 With swallowing initiated by superior laryngeal nerve stimulation, the response characteristics of single nerve fibers participating in smooth muscle peristalsis could be divided into two groups.51 Activity of short latency vagal fibers correlated temporally with the onset of the deglutitive inhibition while the activity of long latencies fibers was tem­ porally correlated with the onset of contraction at each esophageal locus. Thus, activity of neurons in the dorsal motor nucleus of the vagus reflects several pro­perties of primary peristalsis in the smooth muscle esophagus including deglutitive inhibition and both the speed and vigor of peristaltic contraction. Control of peristalsis may also arise in the myenteric plexus. Stimulation of decentralized vagal efferents evokes

681

682

Section V  Esophagus peristalsis similar to that seen with swallowing and this is obliterated with transection across the smooth muscle esophagus, suggesting that an intact intramural neural myenteric plexus is necessary for peristaltic propagation. In contrast, transection across the striated muscle in the proximal esophagus does not inhibit peristaltic progression across the transection site or distally.47 Further evidence supporting the potential autonomy of peripheral mechanisms is that distention anywhere within the smooth muscle esophagus will elicit secondary peristalsis despite extrinsic denervation.52 Regardless of central or ganglionic control, esophageal smooth muscle contraction is ultimately elicited by ganglionic cholinergic neurons. Swallow-induced peristalsis is highly atropine sensitive, can be augmented by cholinergic agonists, and is inhibited by acetylcholinesterase.53,54 Less clear, however, are the control mechanisms for the direction and velocity of the peristaltic wavefront. Nerve conduction studies indicate that neural stimuli initiated by swallowing propagate with a speed of 5 to 6 m per second and therefore reach the ganglionic neurons along the length of the esophagus essentially simultaneously.51 However, the latency between the arrival of the vagal stimulus and muscle contraction progressively increases moving aborally. In humans, the latent period is two seconds in the proximal smooth muscle esophagus and five to seven seconds just proximal to the lower esophageal sphincter (LES). The in vitro correlate of this is that when electrically stimulated, distal esophageal muscle strips exhibit longer latencies to contraction than do strips from the proximal esophagus.55,56 The latency gradient can be changed by varying vagal stimulation parameters or by pharmacologic manipulation suggesting it to be the result of an interaction between the initial inhibition and subsequent excitation of esophageal smooth muscle.57 The current hypothesis is that peristaltic direction and velocity result from a neural gradient along the esophagus, wherein excitatory ganglionic neurons dominate proximally and inhibitory ganglionic neurons dominate distally. This organization is supported by the demonstration with pressure topography plotting of two subsegments within the smooth muscle segment, the first of which is strongly reactive to stimulation with cholinergic drugs.58,59 The primary inhibitory neurotransmitter (formerly referred to as the nonadrenergic, noncholinergic transmitter) is nitric oxide (NO) produced from l-arginine by the enzyme NO synthase in myenteric neurons.60,61 NO synthase inhibitors reduce the latency to contraction in vivo in response to swallowing.62,63 In addition to NO neurons, there is also evidence for a role of vasoactive intestinal polypeptide (VIP)–containing neurons in the initial inhibition.64,65 Sympathectomy of the esophagus has no apparent effect on peristalsis.66,67 On the other hand, bilateral vagotomy results in paralysis of the striated muscle segment. Severing only the afferent nerve supply to the striated muscle abolishes secondary peristalsis while leaving primary peristalsis intact, highlighting the role of central programming in the latter and the necessity of afferent sensory signals in the former.2 Recordings from the cervical esophageal vagal afferents show these to be highly sensitive to intraluminal distention, implicating them as the sensory basis for secondary peristalsis. In the smooth muscle segment of the esophagus, vagal cooling or vagotomy reduces the amplitude of primary peristalsis68,69 but does not affect secondary peristalsis.70 A tool introduced in the study of esophageal physiology is high-resolution manometry or, more precisely, highresolution esophageal pressure topography. High-resolution

esophageal pressure topography allows the imaging of esophageal contractile activity as a continuum not only in time but also along the length of the esophagus. Figure 42-4 illustrates the conversion from conventional manometric study to high-resolution esophageal pressure topography. Note how high-resolution esophageal pressure topography reveals that the vigor and propagation of peristalsis along the length of the esophagus is not seamless. Rather, there is a distinct transition zone between the striated and smooth muscle segments characterized by the minimal peristaltic amplitude, a slight delay in progression, and an increased likelihood of failed transmission.71 Detailed modeling studies suggest this to be the transition point between two distinct contraction waves governing the proximal and distal esophagus respectively.72 The topographic analysis also reveals a segmental characteristic of peristaltic progression through the smooth muscle esophagus with two distinct contractile segments separated by a pressure trough followed by the LES, which contracts with vigor and persistence quite dissimilar to the adjacent smooth muscle esophagus.73 The longitudinal muscle of the esophagus also contracts during peristalsis with the net effect of transiently shortening the esophagus by 2 to 2.5 cm.74 Similar to the pattern of circular muscle contraction, longitudinal muscle contraction is propagated distally as an active segment at a rate of 2 to 4 cm per second.75 The segment of contracting longitudinal muscle precedes, but overlaps with the contracting segment of circular muscle. Thus, within a given esophageal segment, the contraction of the longitudinal and circular muscle are slightly out of phase with each other. Propulsive force occurs in the zone of overlap as the delayed circular muscle contraction “catches up” with the distal longitudinal muscle contraction.75 Central mechanisms also control the contractions of esophageal longitudinal muscle. Swallowing induces peristaltic sequences with gradual activation of longitudinal muscle progressing from orad to caudad. This progression is associated with a progressive increase in latency similar to that seen with the circular smooth muscle esophagus.76 However, unlike the responses observed in the circular muscle, stimulation of decentralized vagal efferent fibers causes simultaneous contractions in the longitudinal muscle layer, suggesting this muscle layer is free of inhibitory neuron control.57

ESOPHAGOGASTRIC JUNCTION

The anatomy of the esophagogastric junction (EGJ) is complex (see Chapter 41). The distal end of the esophagus is anchored to the diaphragm by the phrenoesophageal membrane that inserts circumferentially into the esophageal musculature close to the squamocolumnar junction (SCJ). The esophagus then traverses the diaphragmatic hiatus and joins the stomach in almost a tangential fashion. Thus, there are three significant contributors to the EGJ high-pressure zone: the LES, the crural diaphragm, and the muscular architecture of the gastric cardia that constitutes the distal aspect of the EGJ high-pressure zone. The LES is a 3- to 4-cm segment of tonically contracted smooth muscle at the distal extreme of the tubular esophagus. Surrounding the LES at the level of the SCJ is the crural diaphragm, most commonly bundles of the right diaphragmatic crus forming a teardrop-shaped canal about 2 cm long on its major axis (Fig. 42-5).77,78 Elegant anatomic studies suggest that the component of the EGJ high-pressure zone distal to the SCJ is largely attributable to the opposing sling and clasp fibers of the middle layer of gastric cardia musculature.77,79,80 In this region, the lateral wall of the esopha-

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders Pressure scale (mm Hg) Striated muscle

Conventional sensor #1

40

150

0 40

Conventional sensor #2

0 100

Conventional sensor #3 cm

0 100

2nd trough Conventional sensor #5

0 20

Sleeve Sleeve sensor 0

Esophagus

Seconds

Aorta L1

Right crus of diaphragm

0

LES

Left crus of diaphragm

Figure 42-5.  Drawing of the most common anatomy of the diaphragmatic hiatus in which the muscular elements of the crural diaphragm derive from the right diaphragmatic crus. The right crus tendon arises from the anterior longitudinal ligament overlying the lumbar vertebrae. Once muscular elements emerge from the tendon, two flat muscular bands form and cross each other in scissor-like fashion, form the walls of the esophageal hiatus, and decussate with each other anterior to the esophagus. (Modified from Jaffee BM. Surgery of the esophagus. In: Orlando RC, editor. Atlas of Esophageal Diseases. 2nd ed. Philadelphia: Current Medicine, Inc; 2002. pp 221-42.)

13

mm Hg

3rd trough

mm Hg

0 40

mm Hg

Smooth muscle

mm Hg

Smooth muscle

Conventional sensor #4

mm Hg

1st trough, transition zone

mm Hg

0

Figure 42-4.  Topographic depiction of esophageal peristalsis using high-resolution manometry. The left panel shows a schematic drawing of placement of a monitoring system with five manometric side holes spaced 4-cm apart and a 6-cm sleeve sensor placed just distal to the last manometric port. The horizontal arrows to the right panel point to the corresponding data tracings obtained from each combined manometry/impedance or sleeve recording site. The isocontour plot clearly demonstrates that progression through the esophagus is not seamless. The proximal striated and distal smooth muscle esophageal contractile segments are separated by a transition zone (first pressure trough). The smooth muscle esophagus is also divided into two distinct contractile segments separated by a second pressure trough. The region of the lower esophageal sphincter (LES) is also distinguished by a separate contractile segment that contracts with vigor and persistence dissimilar to the adjacent smooth muscle segment and is separated from the adjacent smooth muscle by a third pressure trough. Representative line tracings are displayed in a conventional ma­nometry mode to illustrate the conceptual conversion from line tracings to high-resolution pressure topography.

gus meets the medial aspect of the dome of the stomach at an acute angle, defined as the angle of His. Viewed intraluminally, this region extends within the gastric lumen, appearing as a fold that has been conceptually referred to as a flap valve because increased intragastric pressure forces it closed, sealing off the entry to the esophagus.81,82 Physiologically, the EGJ high-pressure zone is attributable to a composite of both the LES and the surrounding crural diaphragm. Concurrent fluoroscopy and manometry, aided by the placement of an endoclip at the SCJ, localized the EGJ high-pressure zone as extending 1 to 1.5 cm proximal to the SCJ and about 2 cm distal to it.83 Manometric and intraluminal ultrasound studies suggest that axial asymmetry of the pressure profile is attributable to the varying thickness of the muscularis propria, whereas the radial pressure asymmetry results from asymmetrical compression by the surrounding crural diaphragm.83,84 Resting LES tone ranges from 10 to 30 mm Hg relative to intragastric pressure with considerable temporal fluctuation. The mechanism of LES tonic contraction is likely a property of the muscle itself and of nerves affecting the sphincter. This conclusion is supported by the observation that pressure within the sphincter is minimally affected by the elimination of neural activity by close intra-arterial injection of tetrodotoxin.85 Myogenic LES tone varies directly with membrane potential86 and superimposed electrical spike activity that leads to an influx of Ca++. Resting membrane potential of the LES is slightly greater (i.e., less negative) than that of the adjacent esophagus.87 Modulation of spike activity and membrane potential are not completely understood; however, it is believed that spike activity may be regulated by K+- and Ca++-activated Cl− channels.88

683

684

Section V  Esophagus Sphincter tone may be maintained by inositol phosphatemediated continuous release of intracellular calcium.89,90 Inositol phosphate concentrations are higher in the LES than in adjacent esophagus. Apart from myogenic factors, LES pressure is also modulated by intra-abdominal pressure, gastric distention, peptides, hormones, foods, and many medications. Large fluctuations of LES pressure occur with the migrating motor complex (MMC); during phase III, LES pressure may exceed 80 mm Hg. Lesser fluctuations occur throughout the day with pressure decreasing in the post-prandial state and increasing during sleep.91 As mentioned, superimposed on the myogenic LES contraction, input from vagal, adrenergic, hormonal, and mechanical influences will alter LES pressure. Vagal influence is similar to that of the esophageal body with vagal stimulation activating excitatory and inhibitory myenteric neurons.49 Dogs provide an interesting model for studying this because they have an entirely striated muscle esophagus except for a smooth muscle band at the LES. Vagal fibers innervating the dog LES are of two types: (1) spontaneously active fibers that exhibit a sudden increase with swallowing, abruptly cease firing when the peristaltic contraction arrives, and then resume a spontaneous rate; and (2) spontaneously active fibers that cease activity with swallowing and resume normal activity when the bolus reaches the stomach.92 Thus, the LES pressure at any instant reflects the balance between excitatory and inhibitory neural input and altering the pattern of vagal discharge can result in a swallow-mediated LES relaxation. Data on humans suggest that, similar to the dog, basal LES tone is partially generated by cholinergic input.93 The excitatory and inhibitory intramural neurons are acetylcholine sensitive with nicotinic and muscarinic receptors.94 At the LES, the excitatory neurons release acetylcholine, whereas the inhibitory neurons mainly use NO as a neurotransmitter. Adrenergic influences on LES pressure are complex and mostly mediated through actions on the myenteric neurons.34,95 Sympathetic fibers from the stellate and proximal thoracic ganglia follow the splanchnic nerve, form a recurrent loop through the celiac ganglion, and then synapse on both the excitatory and inhibitory myenteric neurons expressing α-adrenergic receptors. Adrenergic stimulation increases LES pressure by activating excitatory neurons and reducing inhibitory neuron activity. Adrenergic action on the esophageal body is the opposite of that on the LES, with direct inhibition of the muscle and inhibition of the excitatory myenteric neurons. The crural diaphragm is also a major contributor to EGJ pressure (see Fig. 42-5). Even after esophagogastrectomy, with removal of the smooth muscle LES, a persistent EGJ pressure of about 6 mm Hg can be demonstrated during expiration.96 During inspiration there is substantial augmentation of EGJ pressure attributable to crural diaphragm contraction. Experimentally, the inspiratory augmentation of EGJ pressure can be increased even more with enhanced respiratory effort or, conversely, can be eliminated by manual ventilation. The augmentation of LES pressure observed during sustained inspiration corresponds both temporally and quantitatively with the augmentation of crural electromyographic activity, and this augmented EGJ pressure can obscure intrinsic LES relaxation induced by esophageal distention.97 Crural diaphragm contraction is also augmented during abdominal compression, straining, or coughing.98 On the other hand, during esophageal distention, vomiting, and belching, electrical activity in the crural diaphragm is selectively inhibited despite continued respiration demonstrating a control mechanism independent of

the costal diaphragm.99,100 This reflex inhibition of crural activity is eliminated with vagotomy.

Lower Esophageal Sphincter Relaxation

The neural mediation of LES relaxation has been studied extensively.94,101,102 LES relaxation can be triggered by distention from either side of the esophagogastric junction or by swallowing.103 Relaxation induced by esophageal distention is modulated by bolus volume and is unaffected by vagotomy, demonstrating it to be an intramural process. Relaxation is, however, antagonized by tetrodotoxin proving that it is mediated by postganglionic nerves.104 Deglutitive LES relaxation is mediated by the vagus nerve, which synapses with myenteric plexus inhibitory neurons. Ganglionic transmission is through nicotinic and muscarinic acetylcholine receptors and can be blocked by a combination of hexamethonium (i.e., nicotinic blocker) and atropine (i.e., muscarinic blocker). Current evidence implicates NO as the main neurotransmitter in the postganglionic neurons responsible for LES relaxation. NO is produced by NO synthase from the precursor amino acid l-arginine. Neuronal NO synthase is a soluble cytosolic enzyme and has been identified in neurons of the myenteric plexus, co-localizing with VIP, which may be a second inhibitory neurotransmitter in the LES as well as in the esophageal body.105,106 NO is released with neural stimulation in the esophagus, LES, and stomach.60,107-109 In the LES, NO has a marked inhibitory effect, and multiple in vitro and in vivo studies have shown that NO synthase inhibitors block neurally mediated LES relaxation. Although the evidence implicating NO as the main inhibitory transmitter facilitating LES relaxation is very convincing, NO may not work alone. VIP-containing neurons have been demonstrated in the submucosal plexus and VIP relaxes the LES by direct muscle action.110-114 Electrical stimulation of LES muscle strips also causes LES relaxation and release of VIP into the muscle bath. Furthermore, VIP antiserum partially reduces LES relaxation evoked by vagal or field stimulation.65,114 It is thought that VIP acts on NO synthase containing neural terminals as a prejunctional neurotransmitter, facilitating the release of NO and on gastric muscle cells to stimulate production of NO by the muscle.115-119 In addition to VIP, there is some evidence that peptide histidine isoleucine (PHI) in the cat and, to a lesser extent, calcitonin gene-related peptide (CGRP) in the opossum may participate as inhibitory neurotransmitters.114,120 Like VIP, PHI and CGRP relax the LES by a direct action on the muscle.114,120,121 PHI is of interest because it is derived from the same precursor as VIP and coexists with VIP in the same neurons.122 The notion that multiple neurotransmitters may interact to produce LES relaxation may resolve inconsistencies and discrepancies that follow from the assumption that any one neurotransmitter is uniquely responsible for LES relaxation. Reports of the co-localization of NO synthase, VIP, pituitary adenylate cyclase activating peptide (PACAP), CGRP, and galanin in myenteric neurons of the distal esophagus support this concept.123-125

Transient Lower Esophageal Sphincter Relaxation

During rest the EGJ must contain gastric juice, but also be able to transiently relax and permit gas venting of the stomach without allowing reflux of gastric juice and food. These functions are accomplished by prolonged LES relaxations that occur transiently without swallowing or peristal-

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders

UES relaxation

Length along the esophagus

5 10

Pressure (mm Hg) 30

Gastroesophageal common cavity

20

15 Proximal esophageal clip

20 25 30 35

10 5.0 cm

7.0 cm

0

SCJ clip Abdominal Onset of strain tLESR 4490

4495

9.5 cm

4500

4505

4510

–10 4515

Time (sec) Figure 42-6.  Fluoroscopic visualization of movement of two endoclips in the esophagus (one placed at the squamocolumnar junction [SCJ] and the other placed 10 cm proximal to the SCJ) during a transient lower esophageal sphincter relaxation (tLESR), recorded in a high-resolution esophageal pressure topography format. The manometric recording spans from the pharynx to the stomach and, in this instance, the tLESR is associated with an abdominal strain and a “microburp” evident by the brief upper esophageal sphincter (UES) relaxation and abrupt depressurization of the esophagus with gas venting. When the clip data are imported into the isocontour plots, it is evident that the SCJ clip excursion mirrors movement of the esophagogastric junction (EGJ) high-pressure band. Esophageal shortening is most prominent in the distal portion of the 10-cm segment isolated by the endoscopic clips, as seen from the approximately 5-cm movement of the distal SCJ clip concurrent with only minimal movement of the proximal clip. Note also the absence of crural diaphragm contractions during the span of the tLESR.

sis. These transient LES relaxations (tLESRs) are thought to be an important mechanism in the pathogenesis of GERD (see Chapter 43). tLESRs are a complex reflex distinguishable from swallow-induced relaxation in several ways: (1) a prolonged (more than 10 seconds) LES relaxation independent of a pharyngeal swallowing, (2) contraction of the distal esophageal longitudinal muscle causing esophageal shortening, (3) absence of synchronized peristalsis, and (4) crural diaphragm inhibition, which is not the case with swallow-induced relaxation (Fig. 42-6).126,127 tLESRs occur most frequently in the postprandial state during gastric accommodation attributable to vagally mediated receptive relaxation of the fundus (see Chapter 48). In the setting of the completely relaxed EGJ during tLESRs, even the minimal gastroesophageal pressure gradients observed with gastric distention (3 to 4 mm Hg) are sufficient to facilitate gas venting of the stomach (belching). Thus, tLESRs are the physiologic mechanism of belching.128,129 Proximal gastric distention is the major stimulus for tLESR. Distention stimulates mechanoreceptors (intraganglionic lamellar endings) in the proximal stomach activating vagal afferent fibers projecting to the nucleus of the solitary tract.127,130,131 The afferent and efferent neural pathways responsible for swallow and non-swallow LES relaxations have been compared in the mouse. The afferent arm of swallow-induced relaxation lies in the pharyngeal and superior laryngeal nerves, with the central neural circuit in the medullary subnuclei.132-135 Non–swallow-induced relaxations, in contrast, are initiated through gastric afferents in the subdiaphragmatic vagus and activate neurons in the caudal part of the dorsal motor nucleus.136 The efferent limb of both swallow and non-swallow LES relaxations lies in the preganglionic vagal inhibitory pathway to the LES. Both types of relaxation can be blocked by bilateral cervical vagotomy, cervical vagal cooling, or NO synthase inhibi-

tors.127,137 Vagal outflow from the dorsal motor nucleus completely inhibits both the LES and the crural diaphragm, an important distinction from swallow-induced LES relaxation, which is not associated with concomitant inhibition of the crural diaphragm. tLESRs triggered by gastric distention likely use NO and cholecystokinin (CCK) as neurotransmitters, as evidenced by increased tLESR frequency after intravenous CCK and blockade by either NO synthase inhibitors or CCK-A receptor antagonists.138-141 The CCK (and fatty meal) augmentation of tLESR frequency is mediated through CCK-A receptors.142,143 Muscarinic receptor involvement in the tLESR pathway is suggested by inhibition of tLESRs by atropine.144-147 Finally γ-aminobutyric acid (GABA)B agonists, such as baclofen, inhibit tLESRs (see Chapter 43), possibly by acting on peripheral receptors and receptors located in the dorsal motor nucleus of the vagus.148-153

ESOPHAGEAL SENSATION

The human esophagus can sense mechanical, electrical, chemical, and thermal stimuli. These stimuli are perceived as chest pressure, warmth, or pain with substantial overlap in perception among stimuli.154,155 Esophageal sensation is carried via both the vagal and spinal afferent nerves. The associated vagal neurons are located in the nodose and jugular ganglia while the corresponding spinal neurons are located in thoracic and cervical dorsal root ganglia. Vagal afferents to the upper one third of the esophagus are carried in the superior laryngeal nerve, whereas those to the remainder of the esophagus and LES are carried in vagal branches.156 The spinal afferents are contained in the thoracic splanchnic nerves projecting onto the lower cervical to upper lumbar spinal segments.156,157 Compared with vagal afferent fibers, relatively little is known about esophageal spinal afferents, but spinal pathways are thought to be primarily

685

686

Section V  Esophagus nociceptive. Supportive of that concept, prolonged acid perfusion produces esophageal hypersensitivity to distention via spinal sensitization.158,159 Esophageal sensations are usually perceived substernally; in the instance of pain, radiation to the midline of the back, shoulders, and jaw is very analogous to cardiac pain. These similarities are likely due to convergence of sensory afferent fibers from the heart and esophagus in the same spinal pathways, even to the same dorsal horn neurons in some cases.160 Vagal sensory endings in the esophagus consist of free nerve endings, intraganglionic laminar endings (IGLEs) within the myenteric ganglia, and intramuscular arrays (IMAs) within the muscularis propria. Labeling studies demonstrated the densest innervation of free endings between the muscularis mucosa and muscularis propria along the entire length of the esophagus.161 Electron and confocal microscopy reveal that most vagal afferents terminating in the myenteric ganglia do so in specialized laminar structure that encapsulates myenteric ganglia (IGLEs).162 One vagal afferent axon may end in several IGLEs. Combined electrophysiologic and tracer studies have demonstrated that the majority of the tension-sensitive esophageal afferents emanate from IGLEs.163 These endings detect passive and active tension of hollow viscera. In addition to IGLEs, another specialized vagal axonal ending primarily found in the longitudinal and circular smooth muscle forms a branching array parallel to the muscle fibers (IMAs).164 Although more ubiquitous in other parts of the gut, these are essentially restricted to the LES in the esophagus.165 IMAs maintain a close network with interstitial cells of Cajal and it appears that ICCs serve a trophic function.166 Functionally, IMAs appear to be stretch-sensitive endings, sensitive to changes in the muscle length.167 With IGLE and IMA sensory endings concentrated deeply within the muscularis propria beneath a relatively impermeable mucosa, it seems unlikely that intraluminal acid can directly stimulate them. However, these afferents easily respond to chemical mediators such as 5-hydroxytryptamine or alpha, beta-methylene adenosine triphosphate (ATP) as well as to mucosally applied bile or capsaicin,168,169 suggesting that these chemicals induce the release of some endogenous substance that in turn excites the muscle afferents. Supportive of that concept, muscle afferents have been shown to be sensitive to the selective purinergic P2X3 agonist alpha, beta-methylene ATP163,170 and immuno­ histochemical studies have documented the presence of P2X3 receptors in IGLEs implicating direct activation of purinergic P2X2 and P2X3 receptors as an initiating sensory event.171,172 With respect to free nerve endings, acid can excite eso­ phageal vagal and spinal afferents by activating two protongated channels: transient receptor potential vanilloid-1 (TRPV1) and acid-sensing ion channels (ASICs).173-176 Capsaicin, a derivative of chili pepper, excites afferent fibers by activating the TRPV1 channels, which can also initiate a positive feedback loop of increased (neurogenic) inflammation via the release of neuropeptides and inflammatory substances.177,178 ASICs are the other major receptor class that are sensitive to acid, although it is doubtful that acid is their natural ligand; ASICs are probably more involved in mechanotransduction.176 Owing to its significance in the pathogenesis of reflux disease, there has been substantial interest in modulating the tLESR reflex (see Chapter 43). The current concept is that vagal afferent endings terminating in IGLEs located in the proximal stomach are primarily responsible for initiating the reflex, which is then mediated through the medulla and back to the esophagus and diaphragm via vagal efferent

and the phrenic nerves.179 Pharmacologic and physiologic studies have demonstrated that the mechanotransduction properties of tension-sensitive vagal afferent fibers can be attenuated by the GABAB receptor agonist baclofen, thereby reducing the frequency of tLESR.180 Glutamate receptors are also present in vagal and spinal sensory afferent fibers, and metabotropic glutamate receptor antagonists (especially mGluR5 antagonists) have been shown to inhibit tLESR.153

ESOPHAGEAL MOTILITY DISORDERS A working, albeit restrictive, definition of an esophageal motility disorder is an esophageal disease attributable to neuromuscular dysfunction that causes symptoms referable to the esophagus, most commonly dysphagia, chest pain, or heartburn. Employing this definition, there are relatively few firmly established primary esophageal motility disorders: achalasia, distal esophageal spasm (DES), and gas­ troesophageal reflux disease (GERD). GERD is clearly the most prevalent among the group and, fittingly, it is addressed in detail elsewhere in this text (see Chapter 43). Esophageal motility disorders also can be secondary phenomena in which case esophageal dysfunction is part of a more global disease: pseudoachalasia, Chagas disease, and scleroderma. Dysphagia attributable to pharyngeal or UES dysfunction can be included in a discussion of esophageal motor disorders, but this is usually as a manifestation of a more global neuromuscular disease process. The major focus of this chapter is on the primary esophageal motility disorders, particularly achalasia. However, mention is made of the secondary motility disorders and proximal pharyngoesophageal dysfunction when important unique features exist.

EPIDEMIOLOGY

Estimates of the prevalence of dysphagia among individuals older than 50 years of age range from 16% to 22%181,182 with most of this related to oropharyngeal dysfunction. Within health care institutions, it is estimated that up to 13% of hospitalized patients and 60% of nursing home residents183 have feeding problems, most of which are attributed to oropharyngeal dysfunction as opposed to esophageal dysfunction. Most oropharyngeal dysphagia is related to impaired neuromuscular function; the prevalence of the most common anatomic etiology, Zenker’s diverticulum (discussed in Chapter 23), is estimated to range from a meager 0.01% to 0.11% of the U.S. population, with a peak incidence in men between the seventh and ninth decades.184 The consequences of oropharyngeal dysphagia are severe: volume depletion, malnutrition, aspiration, choking, pneumonia, and death. In fact, mortality of nursing residents with dysphagia and aspiration can be as high as 45% over one year.185 As the U.S. population continues to age, oropharyngeal dysphagia will become an increasing problem associated with complex medical and ethical issues. Achalasia is the most easily recognized and best-defined motor disorder of the esophagus. The annual incidence of achalasia is about 1/100,000 population in the United States and Europe,186,187 affecting both genders equally and usually presenting between ages 25 and 60.188 Because achalasia is a chronic condition, its prevalence greatly exceeds its incidence, with prevalence estimates in Europe ranging from 7.1/100,000 in Wales to 13.4/100,000 in Ireland.189 Reports of familial clustering of achalasia raise the possibility of genetic predisposition. Achalasia has been reported in one pair of monozygotic twins,190 in siblings,191 and in children

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders Table 42-1  Affected Phases, Manifestations, and Typical Disease Conditions Causing Oropharyngeal Dysphagia affected phase OF THE OROPHARYNGEAL SWALLOW Nasopharyngeal closure Laryngeal closure UES opening Tongue loading and bolus propulsion Pharyngeal clearance

manifestation OF DYSFUNCTION

associated disease(s)

Nasopharyngeal regurgitation Nasal voice Aspiration during bolus transit

Myasthenia gravis

Dysphagia Post-swallow residue in hypopharynx, aspiration Diverticulum formation Sluggish misdirected bolus Post-swallow residue in hypopharynx Aspiration

Stroke Traumatic brain injury Cricopharyngeal bar Hypopharyngeal diverticulum Parkinson’s disease Parkinson’s disease Surgery-induced defects Cerebral palsy Polio or post-polio Oculopharyngeal dystrophy Stroke

UES, upper esophageal sphincter.

of affected parents.192 However, a genetic determinant for achalasia is not strong.193 Emphasizing this point, a survey of 1012 first-degree relatives of 159 achalasic patients identified no affected relatives.194 Familial adrenal insufficiency with alacrima is a rare genetic achalasia syndrome. This condition is inherited as an autosomal recessive disease that manifests itself with the childhood onset of autonomic nervous system dysfunction including achalasia, alacrima, sinoatrial dysfunction, abnormal pupillary responses to light, and delayed gastric emptying.195 It is caused by mutations in AAAS, a gene which encodes a protein known as ALADIN. No population-based studies exist on the incidence or prevalence of esophageal motility disorders other than achalasia. Thus, the only way to estimate the incidence or prevalence of spastic disorders is to examine data on populations at risk and reference the observed frequency of spastic disorders to the incidence of achalasia which, as detailed earlier, is about 1 per 100,000 population. Doing so, the prevalence of DES is similar to that of achalasia (or much lower if more restrictive diagnostic criteria are used). Populations at risk for motility disorders are patients with chest pain or dysphagia, so it is among these patients that extensive manometric data have been collected. Mano­ metric abnormalities are prevalent among these groups,196-205 but in most cases the manometric findings are of unclear significance.197

PATHOGENESIS Oropharyngeal Dysphagia

Obstructing lesions of the oral cavity, head, and neck can cause dysphagia, and must be excluded before diagnosing a neuromuscular disorder. Structural abnormalities may result from trauma, surgery, tumors, caustic injury, congenital anomalies, or acquired deformities. The most common structural abnormalities of the hypopharynx associated with dysphagia are hypopharyngeal (Zenker’s) diverticula and cricopharyngeal bars. If the etiology of oropharyngeal dysphagia is not readily apparent after initial evaluation for anatomic disorders, evidence of functional abnormalities should be sought. Primary neurologic or muscular diseases involving the oropharynx are often associated with dysphagia. Thus, whereas esophageal dysphagia usually results from esophageal diseases, oropharyngeal dysphagia frequently results from neurologic or muscular diseases, with oropharyngeal dysfunction being

just one pathologic manifestation. Although the specifics vary from disease to disease, the net effect on swallowing can be analyzed according to the mechanical description of the swallow outlined earlier. Table 42-1 summarizes the mechanical elements of the swallow along with the manifestation and consequence of dysfunction and provides representative pathologic conditions in which they are likely encountered. Neurologic examination may indicate cranial nerve dysfunction, neuromuscular disease, cerebellar dysfunction, or an underlying movement disorder. Functional abnormalities can be attributable to dysfunction of intrinsic musculature, peripheral nerves, or central nervous system control mechanisms. Of note, contrary to popular belief, the gag reflex is not predictive of pharyngeal swallowing efficiency or aspiration risk. The gag reflex is absent in 20% to 40% of normal adults.206 Evident in Table 42-1, oropharyngeal dysphagia is frequently the result of neurologic or muscular diseases. Neurologic diseases can damage the neural structures requisite for either the afferent or efferent limbs of the oropharyngeal swallow. Virtually any neuromuscular disease can cause dysphagia. Because there is nothing unique to neurons controlling swallowing, their involvement in disease processes is usually random. Furthermore, in most instances, functions mediated by adjacent neuronal structures are concurrently involved. The following dis­ cussion focuses on neuromuscular pathologic processes most commonly encountered. These entities are also discussed in Chapter 35.

Stroke

Aspiration pneumonia has been estimated to inflict a 20% death rate in the first year after a stroke, and 10% to 15% each year thereafter.207 It is usually not the first episode of aspiration pneumonia, but the subsequent recurrences over the years that eventually causes death.208 The ultimate cause of aspiration pneumonia is dysphagia leading to aspiration that can occur by at least three mechanisms: absence or severe delay in triggering the swallow, reduced lingual control, or weakened laryngopharyngeal musculature.17 Conceptually, these mechanisms can involve motor or sensory impairments. Cortical strokes are less likely to result in severe dysphagia than brainstem strokes.209 Cortical strokes are also more likely to demonstrate neurologic recovery. Of 86 consecutive patients who sustained an acute cerebral infarct, 37 (43%) experienced dysphagia when evaluated within four days of the event. However, 86% of

687

688

Section V  Esophagus these patients were able to swallow normally two weeks later,209 with recovery resulting from contralateral areas taking over the lost function.210 Failure to recover swallowing function was more likely among patients incurring larger strokes or patients who have had prior infarcts.

Poliomyelitis

Most cases of poliomyelitis involve only the spinal cord. However, the fatality rate from bulbar polio far exceeds that of spinal disease, primarily a consequence of respiratory depression. Bulbar poliomyelitis is also associated with dysphagia. In one analysis of the persistent sequelae of bulbar poliomyelitis, 28 of 47 patients (60%) had recurrent or continued involvement of the pharynx 17 or more months after their acute illness.211 Speech and swallowing dysfunction result from weakness of the pharyngeal musculature.212 Neurologists have also observed an increasing number of patients with new paretic symptoms traceable to their remote polio infection 30 to 40 years earlier. The new, slowly progressive post-polio muscular atrophy may occur in muscles that were clinically unaffected by the acute illness.211 One investigation studied 13 patients with postpolio dysphagia and demonstrated palatal, pharyngeal, and laryngeal weakness.213 More than half of the patients evaluated demonstrated silent aspiration.

Amyotrophic Lateral Sclerosis

Amyotrophic lateral sclerosis (ALS) is a progressive neurologic disease characterized by degeneration of motor neurons in the brain, brainstem, and spinal cord. Specific symptoms are dependent on the locations of affected motor neurons and the relative severity of involvement. When the degenerative process involves the cranial nerve nuclei, swallowing difficulties ensue. Oropharyngeal dysfunction characteristically begins with the tongue and progresses to involve the pharyngeal and laryngeal musculature. Patients experience choking attacks, become volume depleted and/ or malnourished, and incur aspiration pneumonia. The decline in swallowing function is progressive and predictable, invariably leading to gastrostomy feeding. Patients often die as a consequence of their swallowing dysfunction in conjunction with respiratory depression.214

Parkinson’s Disease

Although only 15% to 20% of patients with Parkinson’s disease complain of swallowing problems, more than 95% have demonstrable defects when studied videofluoroscopically.215 This disparity suggests that patients compensate in the early stages of the disease and complain of dysphagia only when it becomes severe. Abnormalities include repetitive lingual pumping prior to initiation of a pharyngeal swallow, piecemeal swallowing, and oral residue after the swallow. Patients may also exhibit a delayed swallow response and a weak pharyngeal contraction, resulting in vallecular and pyriform sinus residue. Recent data suggest this to be related to the combination of incomplete UES relaxation and a weakened pharyngeal contraction.215

Tumors

Medullary or vagal tumors are potentially debilitating with respect to swallowing. Astrocytomas are the most common tumor subtype affecting adults whereas medulloblastomas are the most common type encountered in children.216 The morbidity of these tumors is often substantially increased as a result of the relative inaccessibility of the medulla to surgery. Unilateral lesions of the vagus can result in hemiparesis of the soft palate and pharyngeal constrictors, as well as of the laryngeal musculature. Surgical manipulation

of this region can even result in complete loss of the pharyngeal swallow response.217 The recurrent laryngeal nerves can be injured as a result of thyroid surgery, aortic aneurysms, pneumonectomy, primary mediastinal malignancies, or metastatic lesions to the mediastinum. Owing to its more extensive loop in the chest, the left recurrent laryngeal nerve is more vulnerable than the right to involvement with mediastinal node malignancy. Unilateral recurrent laryngeal nerve injury results in unilateral adductor paralysis of the vocal cords. This defect can result in aspiration during swallowing because of impaired laryngeal closure. It is rare, however, to have any primary pharyngeal dysfunction resultant from recurrent laryngeal nerve injury.218

Oculopharyngeal Dystrophy

Oculopharyngeal muscular dystrophy is a syndrome characterized by progressive dysphagia and ptosis. Historically, afflicted patients reaching age 50 typically died of starvation resulting from pharyngeal paralysis.219 The disease is now known to be a form of muscular dystrophy and is inherited as an autosomal dominant disorder with occurrences clustered in families of French-Canadian descent. Genetic studies of an afflicted family indicate linkage to chromosome 14, perhaps involving the region coding for cardiac alpha or beta myosin heavy chains.220 Oculopharyngeal dystrophy affects the striated pharyngeal muscles and the levator palpebrae. Although other forms of muscular dystrophy occasionally affect the pharyngeal constrictors, this is rarely a dominant manifestation. The first symptom of oculopharyngeal dystrophy is usually ptosis that slowly progresses and eventually dominates the patient’s appearance. Dysphagia may begin before, be concomitant with, or occur after ptosis. The dominant functional abnormalities are of a weak or absent pharyngeal contraction with hypopharyngeal stasis.219 Dysphagia is slowly progressive, but may ultimately lead to starvation, aspiration pneumonia, or asphyxia.

Myotonic Dystrophy

Myotonic dystrophy is a rare disorder characterized by prolonged contraction and difficulty in relaxation of affected skeletal musculature. Recent investigations suggest that even though only half of the patients complain of dysphagia, pharyngeal and esophageal motor abnormalities can be universally demonstrated. The pattern of abnormality is of a weakened pharyngeal contraction, absent peristalsis in the striated muscle esophagus, and diminished or absent peristalsis in the smooth muscle segment of the esophagus. Myotonia has not been demonstrated in any part of the esophagus.29 Thus, the risk of aspiration in this disease is similar to other forms of muscular dystrophy. Aspiration can occur during the swallow due to poor pharyngeal clearance combined with concurrent weakness of the laryngeal elevators or after the swallow when the substantial pharyngeal residue might fall into the reopened airway.

Myasthenia Gravis

Myasthenia gravis is a progressive autoimmune disease characterized by high circulating levels of acetylcholine receptor antibody and destruction of acetylcholine receptors at neuromuscular junctions. Musculature controlled by the cranial nerves is almost always involved, particularly the ocular muscles. Dysphagia is prominent in more than a third of cases and, in unusual instances, can be the initial and dominant manifestation of the disease.17 In mild cases, dysphagia may not be evident until after 15 to 20 minutes of eating. Classically, manometric studies reveal a progressive deterioration in the amplitude of pharyngeal contrac-

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders

Zenker’s diverticulum Cricopharyngeus

Cricopharyngeus

Esophagus

Trachea

Figure 42-7.  Radiograph of a small Zenker’s diverticulum filled with barium. Although the point of herniation is midline and posterior at Killian’s dehiscence, the diverticulum migrates laterally in the neck as it enlarges because there is no potential space between the posterior pharyngeal wall and the vertebral column. Other sites of herniation can occur (see text).

Figure 42-8.  Cricopharyngeal bar in a patient with oropharyngeal dys­ phagia. The posterior indentation of the barium column is caused by a noncompliant cricopharyngeus muscle. (Courtesy of Dr. Richard Gore, Evanston, Illinois.)

tions with repeated swallows. Peristaltic amplitude recovers with rest or following the administration of 10 mg edrophonium chloride, an acetylcholinesterase inhibitor. In more advanced cases, the dysphagia can be profound and associated with nasopharyngeal regurgitation and nasality of the voice, even to the extent of being confused with bulbar ALS or a brainstem stroke.221

geal (Zenker’s) diverticula demonstrated structural changes that would decrease UES compliance and opening.223 The cricopharyngeus samples from these patients had “fibroadipose tissue replacement and (muscle) fiber degeneration.” Thus, although the muscle relaxes normally during a swallow, it cannot distend normally, resulting in the appearance of a cricopharyngeal indentation, or bar, during a barium swallow (Fig. 42-8). Diminished sphincter compliance necessitates increased hypopharyngeal intrabolus pressure to maintain trans-sphincteric flow through the smaller UES opening. The increased stress on the hypopharynx from the increased intrabolus pressure may ultimately result in diverticulum formation.

Hypopharyngeal (Zenker’s) Diverticulum and Cricopharyngeal Bar

Hypopharyngeal diverticulum and cricopharyngeal bars are closely related disease entities in that it is a crico­ pharyngeal bar that can result in diverticulum formation. Zenker’s diverticulum (Fig. 42-7), is discussed in Chapter 23. Zenker’s diverticulum originates in the midline posteriorly at Killian’s dehiscence, a point of pharyngeal wall weakness between the oblique fibers of the inferior pharyngeal constrictor and the transverse cricopharyngeus muscle (see Fig. 42-6).222 Other locations of acquired pharyngeal diverticula include (1) the lateral slit separating the cricopharyngeus muscle from the fibers of the proximal end of the esophagus through which the recurrent laryngeal nerve and its accompanying vessels run to supply the larynx; (2) at the penetration of the inferior thyroid artery into the hypopharynx; and (3) at the junction of the middle and inferior constrictor muscles. The unifying theme of these locations is that they are sites of potential weakness of the muscular lining of the hypopharynx through which the mucosa herniates, leading to a “false” diverticulum. The best-substantiated explanation for the development of diverticula is that they form as a result of a restrictive myopathy associated with diminished compliance of the cricopharyngeus muscle. Surgical specimens of cricopharyngeus muscle strips from 14 patients with hypopharyn-

Achalasia

Achalasia is characterized by impaired LES relaxation with swallowing, and aperistalsis in the smooth muscle esophagus. The resting LES pressure is elevated in about 60% of cases. If there are nonperistaltic, spastic contractions in the esophageal body, the disease is referred to as vigorous achalasia or, more recently, spastic achalasia.224 These physiologic alterations result from damage to the innervation of the smooth muscle segment of the esophagus (including the LES). Proposed neuroanatomic changes responsible for achalasia include loss of ganglion cells within the myenteric (Auerbach’s) plexus, degeneration of the vagus nerve, and degeneration of the dorsal motor nucleus of the vagus. Of these three possibilities, only the loss of ganglion cells is well substantiated. Several observers report fewer ganglion cells and ganglion cells surrounded by mononuclear inflammatory cells in the smooth muscle esophagus of achalasics.225 One report additionally noted ganglion cell degeneration extending into the proximal stomach in half of 34 specimens analyzed.226 The degree of ganglion cell loss parallels the duration of disease such that ganglion cells are

689

690

Section V  Esophagus almost absent in patients afflicted for 10 or more years.227 A morphologic study of 42 esophagi resected from patients with advanced achalasia reported reduced numbers of ganglion cells and inflammation within the myenteric plexus in all cases.228 The ultimate cause of ganglion cell degeneration in achalasia is gradually being unraveled, with increasing evidence pointing toward an autoimmune process attributable to a latent infection with human herpes simplex virus 1 (HSV-1) in genetically susceptible individuals.229,230 Immunohistochemical analysis of the myenteric plexus infiltrate in achalasia patients revealed that the majority of inflammatory cells are either resting or activated cytotoxic T cells.231 In addition, immunoglobulin M (IgM) antibodies and evidence of complement activation have been demonstrated within myenteric ganglia.232 Antibodies against myenteric neurons have been repeatedly shown in serum of achalasia patients,233,234 especially in patients with HLA DQA1* 0103 and DQB1* 0603 alleles.235 The trigger for initiating the autoimmune response leading to the development of achalasia is suspected to be a viral infection, but studies implicating varicella zoster or measles virus have been contradictory.232,236,237 However, an elegant recent study provided strong evidence implicating HSV-1 as the culprit.230 T cells of achalasia patients exhibited clonal expansion within the myenteric plexus of the LES, and were activated by HSV-1 antigens, but not by cytomegaloviral, adenoviral, or enteroviral antigens. Furthermore, HSV-1 antibodies and HSV-1 deoxyribonucleic acid (DNA) were isolated in 84% and 63% of achalasic patients, respectively, potentially implicating HSV-1 in the majority of achalasia cases. Interestingly, HSV-1 was also detected in LES tissue from nonachalasic organ donors, suggesting that the development of achalasia is dependent on both the virus and a genetic predisposition as indicated by the specific HLA associations. Achalasia may also be associated with degenerative neurologic disorders such as Parkinson’s disease. Patients with both achalasia and Parkinson’s disease were noted to have Lewy bodies (intracytoplasmic hyaline or spherical eosinophilic inclusions) in the degenerating ganglion cells of the myenteric plexus.238 Physiologic studies in individuals with achalasia also suggest dysfunction consistent with postganglionic denervation of esophageal smooth muscle. Such damage can affect excitatory ganglion neurons (cholinergic), inhibitory ganglion neurons (NO ± VIP), or both. Consider first the excitatory ganglion neurons. Muscle strips from the circular layer of the esophageal body of achalasic patients contract when directly stimulated by acetylcholine but fail to respond to ganglionic stimulation by nicotine, indicating a postganglionic excitatory defect. However, it is likely that loss of excitatory innervation is variable among achalasic patients. Partial preservation of the postganglionic cholinergic pathway is suggested by the observations that an achalasic patient’s LES pressure increases after administration of the acetycholinesterase inhibitor, edrophonium, and decreases after administration of the muscarinic antagonist, atropine.239 These observations are crucial to understanding why botulinum toxin may have therapeutic benefit in achalasia (see section on treatment). Regardless of excitatory ganglion neuron impairment, it is clear that inhibitory ganglion neuron dysfunction is an early manifestation of achalasia. These neurons mediate deglutitive inhibition (including LES relaxation) and the sequenced propagation of esophageal peristalsis; their absence offers a unifying hypothesis for the key physiologic abnormalities of achalasia, namely, impaired LES relaxation and aperistalsis. Inhibitory ganglion neurons use NO as

a neurotransmitter, and patients with achalasia have been shown to lack NO synthase in the gastroesophageal junction.240 VIP may be a co-transmitter in these neurons and immunohistochemical studies have demonstrated a marked reduction of VIP-staining neurons in achalasic individuals.112 A multitude of evidence supports impaired physiologic function of post-ganglionic inhibitory innervation in the smooth muscle esophagus of achalasic patients. Muscle strips from their LES do not relax in response to ganglionic stimulation as they do in normal controls241 and CCK, which normally stimulates the inhibitory ganglion neurons, thereby reducing LES pressure, paradoxically increases the LES pressure in achalasics.242 Impaired inhibitory innervation of the smooth muscle esophagus above the LES is more difficult to demonstrate because of the absence of resting tone in this region. However, in a clever experiment, Sifrim and colleagues used an intraesophageal balloon to create a high-pressure zone in the tubular esophagus that then relaxed with the onset of deglutitive inhibition. This deglutitive relaxation in the esophageal body was absent in early, nondilated cases of achalasia.243

Distal Esophageal Spasm

The term “diffuse esophageal spasm” and our present concept of this entity dates to Fleshler’s 1967 description of a “clinical syndrome characterized by symptoms of substernal distress or dysphagia or both, the roentgenographic appearance of localized, nonprogressive waves (tertiary contractions), and an increased incidence of nonperistaltic contractions recorded by intraluminal manometry.”244 Because only the smooth muscle esophagus is affected, the entity was subsequently more precisely labeled “distal esophageal spasm.”245,246 Clearly, distal esophageal spasm is a disorder of peristalsis. However, in most afflicted patients, the esophagus retains the ability to propagate normal peristaltic contractions the majority of the time suggesting that the neuromuscular pathology is more subtle than with achalasia. Partly because of this fact, the criteria for diagnosing DES remain variable and confusing.245 The neuromuscular pathology responsible for DES is unknown and there are no known risk factors or other conditions associated with DES. Furthermore, because neither the esophageal muscularis propria or myenteric plexus is readily accessible for biopsy and patients with spastic disorders of the esophagus rarely undergo esophageal surgery, only a paucity of pathologic material has been available for analysis. The most striking reported pathologic change is diffuse muscular hypertrophy or hyperplasia in the distal two thirds of the esophagus. Muscular thickening of up to 2 cm has been reported in patients with clinical and manometric evidence of DES.247 However, there are other welldocumented cases of spasm in which esophageal muscular thickening was not found at thoracotomy248 and still other instances of patients with muscular thickening not associated with DES symptoms.249 Similarly, little evidence of neuropathology has been reported; diffuse fragmentation of vagal filaments, increased endoneural collagen, and mitochondrial fragmentation have been described, but the significance of these findings is unclear.250 Despite the absence of defined histopathology, physiologic evidence implicates myenteric plexus neuronal dysfunction in spastic disorders of the esophagus. During peristalsis, vagal impulses reach the entire smooth muscle segment of the esophagus simultaneously and activate myenteric plexus neurons between the longitudinal and circular muscle layers.51 Ganglionic neurons then intervene between the efferent vagal fibers and the smooth muscle,

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders belonging to either an inhibitory population that hyperpolarizes the muscle cell membrane and inhibits contraction or to an excitatory population that depolarizes the membrane, thereby prompting contraction. Thus, the instantaneous activity of the musculature at each esophageal locus is determined by the balance between these controlling influences from the myenteric plexus. Experimental evidence suggests heterogeneity among patients with spastic disorders, such that some primarily exhibit a defect of inhibitory interneuron function, whereas in others the defect is of excess excitation. Two in vivo experiments implicate a defect of myenteric plexus inhibitory interneuron function in the genesis of simultaneous contractions in the distal esophagus. In one, the propagation of a swallow-induced contraction was timed in normal subjects and in a group of patients with a simultaneous contraction in the distal esophagus.251 Within the proximal esophagus the two groups exhibited similar contraction propagation, consistent with this timing being the result of the sequenced activation of motor units by vagal efferent nerves programmed within the medullary swallow center. However, once entering the smooth muscle segment, the patients’ contractions diverged from those of the normal subjects, resulting in a simultaneous contraction in the distal esophagus. The distal esophageal contractions were otherwise normal, but the progressive delay of initiation of the contraction at more distal loci, a function attributable to increasing dominance of inhibitory interneurons in the distal esophagus, was absent. Furthermore, if these patients swallowed twice within a five-second interval, there was no deglutitive inhibition of the first peristaltic contraction within the smooth muscle esophagus, as is observed in normal subjects. A second experiment demonstrating impaired deglutitive inhibition in DES comes from work using an artificial high-pressure zone within the distal esophagus. Patients with motor disorders characterized by rapidly propagating or simultaneous contractions exhibited only partial relaxation of the artificial high-pressure zone, proportional to the impairment of propagation velocity.243 Taken together these findings strongly suggest that one potential neuropathologic process in DES is a selective, intermittent dysfunction of myenteric plexus inhibitory interneurons. A second group of patients in the analysis of Behar and Biancani had normal propagation latency but exhibited frequent spontaneous distal esophageal contractions. These patients had significantly longer and higheramplitude contractions at each locus within the distal esophagus.251 Patients with peristaltic disorders characterized by excess excitation demonstrate heightened sensi­ tivity to stimulation with cholinergic agents,112,252 the cholinesterase inhibitor edrophonium,253 pentagastrin,254 and ergonovine.255 An electromyographic correlate of this excitability is found from bipolar ring electrode recordings from the distal esophagus.256 Whereas normal individuals uniformly exhibited spiking activity prior to each esophageal contraction, DES patients exhibited spike-independent spontaneous esophageal contractions. The preceding discussion suggests that the physiologic abnormalities of patients with spastic disorders are heterogeneous, but all are characterized by an imbalance between the excitatory and inhibitory influences on the esophageal smooth muscle. The suggestion of an impairment of the pathway of deglutitive inhibition is particularly interesting in that it places DES in a pathophysiologic continuum with achalasia, consistent with documented case reports of patients undergoing this evolution.257 Furthermore, there are marked similarities between spastic achalasia and DES,

both characterized by rapidly propagated contractions in the distal esophagus, with the only differences being involvement of the LES and constancy of the disorder in vigorous achalasia. Similar to achalasia, the simultaneous contractions typifying DES impair bolus transit through the esophagus, potentially explaining the associated dysphagia.258

CLINICAL FEATURES

Dysphagia is a fundamental symptom of esophageal motility disorders. Esophageal, as opposed to oropharyngeal, dysphagia is suggested by the absence of associated aspiration, cough, nasopharyngeal regurgitation, dry mouth, drooling, pharyngeal residue following swallow, or co-occurring neuromuscular dysfunction (e.g., weakness, paresthesia, slurred speech). The associated conditions of heartburn, esophagopharyngeal regurgitation, chest pain, odynophagia, or intermittent esophageal obstruction suggest esophageal dysphagia. However, an important limitation of the patient history with esophageal dysphagia is that a patient’s identification of the location of obstruction is of limited accuracy. Specifically, a distal esophageal obstruction caused by an esophageal ring or achalasia often is perceived as cervical dysphagia, such that patients correctly localize distal dysfunction only 60% of the time.259 Because of this subjective difficulty in distinguishing proximal from distal lesions within the esophagus, an evaluation for cervical dysphagia should encompass the entire length of the esophagus. Another important consideration in patient management is that esophageal motility disorders are much less common than mechanical or inflammatory etiologies of dysphagia, such as tumors, strictures, rings, and peptic, pill-induced, or infectious esophagitis. Historical points suggestive of a motor disorder are difficulty with solids and liquids as opposed to only solids, which is more suggestive of mechanical obstruction. However, the functional consequences of mechanical or inflammatory disorders can exactly mimic those of primary motility disorders. Thus, as with the evaluation of oropharyngeal dysphagia, an esophageal motility disorder should be considered as an etiology for dysphagia only after exclusion of other more common diagnoses by endoscopic, histologic, or radiographic examination.

Achalasia

Clinical manifestations of achalasia may include dysphagia, regurgitation, chest pain, hiccups, halitosis, weight loss, aspiration pneumonia, and heartburn. All patients have solid food dysphagia; the majority of patients also have variable degrees of liquid dysphagia. The onset of dysphagia is usually gradual, with the duration of symptoms averaging two years at presentation.188 The severity of dysphagia fluctuates, but eventually plateaus. With long-standing disease there is progressive esophageal dilatation, and regurgitation becomes frequent when large amounts of food and fluid are retained in the dilated esophagus. The regurgitant is often recognized as food that has been eaten hours, or even days, previously. It tends to be nonbilious, non-acid, and mixed with copious amounts of saliva. Patients often fail to recognize the slimy mucoid regurgitant as saliva, being unfa­ miliar with its normal consistency. Chest pain is a frequent complaint early in the course of achalasia, occurring in approximately two thirds of patients.260 Its etiology is unknown, but is speculated to be related to the occurrence of esophageal spasm (more recently, spasm of longitudinal muscle) or to the process of esophageal dilatation associated with disease progression. Treatment of achalasia (discussed later) is less effective in relieving chest pain than it is in relieving dysphagia or regurgitation. However, unlike

691

692

Section V  Esophagus dysphagia or regurgitation, chest pain may spontaneously improve or disappear over time.260 An estimated 10% of people with achalasia have bronchopulmonary complications as a result of regurgitation and aspiration; in some instances, it is these complications rather than dysphagia that prompts them to seek medical care.261 Another interesting, but fortunately rare, symptom of achalasia is airway compromise and stridor as a result of the dilated esophagus compressing the membranous trachea in the neck.262 This is hypothesized to result from dysfunction of the belch reflex.263 It is paradoxical that many achalasic patients complain of heartburn, even after the onset of dysphagia.264 Although gastroesophageal reflux may be a common sequela of the treatments for achalasia, it seems physiologically inconsistent to simultaneously have dysphagia from impaired LES relaxation and reflux from excessive LES relaxation. In support of this skepticism, ambulatory 24-hour esophageal pH studies of achalasic patients have only shown periods of esophageal acidification caused by the bacterial fermentation of retained food in the esophagus rather than discrete gastroesophageal reflux events.265 Furthermore, prolonged LES recordings have shown nearly a complete absence of transient LES relaxations in achalasics.266 However, there are occasional exceptions to this, evident from a welldocumented case of an achalasic patient with intact transient LES relaxation despite the absence of deglutitive LES relaxation.267 The cause of the heartburn often reported by patients with achalasia is unknown.

Distal Esophageal Spasm

The major symptoms of DES are dysphagia and chest pain. Weight loss is rare. Dysphagia is usually intermittent and sometimes related to swallowing specific substances such as red wine or liquids at extreme hot or cold temperature. In some instances, patients experience episodes of esophageal obstruction while eating that persists until relieved by emesis. Esophageal chest pain is very similar in character to angina and is often described as crushing or squeezing in character, radiating to the neck, jaw, arms, or midline of the back. Pain episodes may last from minutes to hours, but continued swallowing is not always impaired. The mechanism producing esophageal pain is poorly understood. Recent data suggest that it may be related to sustained contraction of esophageal longitudinal muscle.268 Chest pain is also prevalent in patients subsequently found to have manometric abnormalities that are insufficient to establish a diagnosis of achalasia or DES. Among such individuals, there is a high prevalence of reflux and of psychiatric diagnoses, particularly anxiety and depression.197 Evidence also suggests a lower visceral pain threshold in this group, and symptoms of irritable bowel syndrome (Chapter 118) may be seen in more than 50% of these patients.269

DIFFERENTIAL DIAGNOSIS

The history is crucial in the evaluation of dysphagia. Major objectives of the history are to differentiate oropharyngeal dysphagia from esophageal dysphagia, xerostomia (hyposalivation), or globus sensation. All are frequently confused with each other. Globus sensation, in particular, is frequently confused with dysphagia. Unlike dysphagia, which occurs only during swallowing, globus sensation is prominent between swallows. Patients relate the nearly constant sensation of having a lump in their throat or feeling a foreign object caught in their throat. In some instances globus is

associated with reflux symptoms and in others with substantial anxiety. It is the linkage with anxiety that led to the older nomenclature, “globus hystericus.” Unfortunately, studies have failed to define an objective anatomic or physiologic cause for globus and we are left with the crucial data being in the history; globus sensation persists regardless of the act of swallowing.

Achalasia

The differential diagnosis of achalasia includes other esophageal motility disorders, with functional attributes overlapping those of achalasia and diseases of distinct pathophysiology that duplicate the functional consequences of achalasia. With respect to other motility disorders, there are many similarities between DES and achalasia, especially the subtype of spastic achalasia. In fact, the only distinction between these entities is the demonstration of incomplete LES relaxation in vigorous achalasia. Thus, some have speculated that DES and vigorous (spastic) achalasia may represent early disease and subsequently evolve into full-fledged achalasia.257 Testing this hypothesis, a report on a prospective cohort of patients diagnosed with esophageal spasm between 1992 and 2003 revealed that achalasia was subsequently diagnosed in only one.270 Given that rarity and the possibility of the case initially being misdiagnosed, it seems reasonable to conclude that at most only a small minority of DES cases are part of the continuum with achalasia. With respect to other diseases that duplicate the functional consequences of idiopathic achalasia, the main considerations are Chagas disease, pseudoachalasia associated with malignancy or infiltrative diseases, or various surgical procedures, as discussed following.

Chagas Disease

Esophageal involvement in Chagas disease (see Chapter 109), which is endemic in areas of central Brazil, Venezuela, and northern Argentina, can be indistinguishable from idiopathic achalasia. An estimated 20 million South Americans are infected. Due to immigration, about 500,000 people in the United States are believed infected. Chagas disease is spread by the bite of reduvid (kissing) bug that transmits the parasitic protozoan, Trypanosoma cruzi. An acute septicemic phase of the illness follows that varies in severity from going unnoticed to being fatal.271 The chronic phase of the disease develops up to 20 years after infection and results from destruction of autonomic ganglion cells throughout the body, including the heart, gut, urinary tract, and respiratory tract. Chronic cardiomyopathy with conduction system disturbances and arrhythmias is the most common cause of death. Within the digestive tract, the organs most commonly affected are the esophagus, duodenum, and colon. The severity of esophageal dysfunction is directly proportional to the degree of intramural ganglion cell loss. Abnormal peristalsis is first detectable after 50% of ganglion cells are destroyed, whereas esophageal dilatation occurs only after 90% are destroyed. Paralleling this, the initial dysfunction is confined to the esophageal body, with LES dysfunction occurring late in the course of the disease.271 The most obvious clinical distinction between idiopathic achalasia and esophageal involvement in Chagas disease is evidence of additional organ involvement (megaureter, cardiomyopathy, megaduodenum, megacolon, megarectum) in Chagas disease. With respect to esophageal pathology, the two are otherwise indistinguishable. The diagnosis of Chagas disease is made in the acute phase by visualizing the parasite in a blood smear. In the chronic phase, the diagnosis is confirmed by serologic tests using complement fixation or

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders polymerase chain reaction. The treatment of the achalasia syndrome in Chagas disease is similar to that for idiopathic achalasia (discussed later). Treatment of the infection itself is of limited efficacy in the acute phase and of no proven efficacy with chronic disease.

Pseudoachalasia

Neither the radiographic nor the manometric features of achalasia are specific for idiopathic achalasia or achalasia associated with Chagas disease. Tumor-related pseudoachalasia accounts for up to 5% of cases with manometrically defined achalasia. Pseudoachalasia is more likely in older age groups (sixth decade and beyond), in patients with recent onset of symptoms (within the past year), and in those with early weight loss in excess of 7 kg.188 However, even though these criteria make pseudoachalasia more likely, they still have a poor predictive value.272 Tumor infiltration (especially carcinoma in the gastric fundus) can completely mimic the functional impairment seen with idiopathic achalasia.273 It is because of this potential pitfall that a thorough anatomic examination including endoscopy should be done as part of the diagnostic evaluation of every new case of achalasia. A clue to the presence of pseudoachalasia on endoscopic examination is of more than the slightest resistance of passage of the endoscope across the gastroesophageal junction. In idiopathic achalasia, the endoscope should pop through with only gentle pressure required. If suspicion of pseudoachalasia is high, endoscopic biopsy, computerized tomography, magnetic resonance imaging, or endoscopic ultrasound should be considered for further evaluation, depending on the individual circumstances. Adenocarcinoma of the gastroesophageal junction accounts for more than one half of pseudoachalasia cases, with myriad other tumors and miscellaneous conditions accounting for the remainder. Within the spectrum of malignancies, pancreatic, hepatoma, lung (small cell or non–small cell), esophageal squamous cell, prostate, and lymphoma cases have been reported.188 These tumors produce an achalasia syndrome by infiltrating the wall of the esophagus at the gastroesophageal junction, causing a malignant obstruction at the LES with proximal esophageal dilatation.273 Similarly, pseudoachalasia has also been reported to result from esophageal infiltration by amyloid,274 eosinophilic gastroenteritis,275 and sarcoidosis.276 Although often speculated in the literature, it is less certain, and certainly much less common, that an achalasic syndrome occurs as a paraneoplastic syndrome without direct tumor stenosis of the gastroesophageal junction.188 Postsurgical Dysphagia is common following fundoplication in the early postoperative period and patients are often advised to consume soft diets for the first two to four weeks. Dysphagia that persists beyond two to four weeks should be evaluated with an upper endoscopy or barium esophagogram to assess the integrity of the wrap and evaluate for possible para­ esophageal hernia. Subjects without an overt mechanical disruption should be evaluated with manometry to assess peristaltic function, LES pressure, and LES relaxation to determine whether the wrap is too tight or an underlying motility disorder, such as achalasia, exists. Diagnosing achalasia in the context of fundoplication can be difficult, because aperistalsis and impaired LES relaxation can be seen in both entities. In order to distinguish mechanical obstruction due to an obstructive fundoplication (or crural repair) from achalasia, one can administer amyl nitrite

during manometry and observe the effect on the EGJ highpressure zone. The mechanical effect of a fundoplication is less affected by the smooth-muscle relaxing effects of the amyl nitrite than the hypertensive sphincter of a person with achalasia.277 Bariatric surgery, especially laparoscopic adjustable gastric banding, can be complicated by development of a pseudoachalasia syndrome. A recent report examined 121 patients a year after this procedure and found that 14% of them had esophageal dilatation in excess of 3.5 cm. Affected patients developed an achalasia-type syndrome with dysphagia and vomiting.278 This form of pseudoachalasia is usually, but not always, reversible with removal of the gastric band.279

Distal Esophageal Spasm

The pain associated with DES can closely mimic that of angina pectoris. Given the potentially fatal consequences of the latter, this must always be considered carefully in the differential diagnosis. Features suggesting an esophageal, as opposed to a cardiac, etiology of chest pain include (1) prolonged, nonexertional pain; (2) pain that interrupts sleep; (3) meal-related pain; (4) relief with antacids; and (5) additional accompanying esophageal symptoms such as heartburn, dysphagia, or regurgitation. However, even these characteristics occasionally exhibit overlap with cardiac pain. Furthermore, even within the spectrum of esophageal diseases, neither chest pain nor dysphagia is specific for DES because both symptoms are also characteristic of common esophageal disorders including peptic or infectious esophagitis. Hence, only after these more common diagnostic possibilities have been excluded by appropriate radiographic evaluation, endoscopic evaluation, and in some instances, a therapeutic trial of antisecretory medications, should DES be considered as the etiology of the still unexplained symptoms.

DIAGNOSTIC METHODS Endoscopy

Upper endoscopy should be the first test for evaluating new onset dysphagia because it combines the ability to detect most structural causes of dysphagia with the ability to obtain biopsies. The increasing recognition of eosinophilic esophagitis (Chapter 27) as a confounding clinical entity has increased the potential value of biopsies when performing upper endoscopy in the evaluation of dysphagia.280 The endoscopist should have a very low threshold for obtaining multiple (preferably five) esophageal mucosal biopsy specimens to evaluate for eosinophilic esophagitis even with a normal appearing esophageal mucosa.281 Additionally, should a stricture or mucosal ring be detected, dilation can be accomplished in the same session. However, even though upper endoscopy is an excellent tool for evaluating dysphagia, it has substantial limitations in assessing extraluminal structures and abnormal esophageal motility. It also has the potential to miss subtle obstructing lesions, such as webs and rings.

Contrast Imaging

Contrast studies of the oropharynx and esophagus are useful in assessing dysphagia after endoscopy if the latter was inconclusive or instead of it, if endoscopy is not readily available. Videofluoroscopy is particularly useful for a functional evaluation of the oropharyngeal phase of swallowing following an examination for anatomic explanations. Frequently referred to as a modified barium swallow,

693

694

Section V  Esophagus

Figure 42-9.  Characteristic barium swallow findings in three cases of advanced idiopathic achalasia. Note esophageal dilatation with an air-fluid level (left radiograph) and the tapering at the gastroesophageal junction. Radiographic findings can be subtle in the early phases of the disease. The example on the right was taken from a timed barium swallow examination indicating that barium was retained within the dilated esophagus for five minutes.

Logemann has described a protocol composed of a series of swallow tasks.17 Images are obtained in a lateral projection, framed to include the oropharynx, palate, proximal esophagus, and proximal airway. These images are then evaluated with respect to four major categories of oropharyngeal dysfunction: (1) inability or excessive delay in initiation of pharyngeal swallowing; (2) aspiration; (3) nasopharyngeal regurgitation; and (4) residue of the ingestate within the pharyngeal cavity after swallowing. Furthermore, the procedure allows for evaluation of the efficacy of various compensatory dietary modifications, postures, and swallowing maneuvers in compensating for observed swallowing dysfunction. A barium esophagogram can also provide useful information regarding UES function, peristalsis, and bolus clearance through the EGJ. With advanced cases of achalasia (Fig. 42-9) the findings are somewhat obvious and it is only necessary clinically to differentiate between primary and secondary etiologies. However, with good technique, normal peristalsis can also be verified with 91% to 95% specificity.282,283 Peristalsis is best evaluated in the prone position so that clearance does not occur by gravity. In the prone position, the primary peristaltic wave manifests as an inverted “V” (∧), the peak of which represents the tail of the bolus. Luminal closure at the tail of the bolus corresponds closely to the leading edge, or upstroke, of the peristaltic wavefront as recorded manometrically (Fig. 42-10). Peristaltic abnormalities are inferred by retrograde escape of the bolus through the peristaltic wavefront resulting in incomplete esophageal emptying. Normally the EGJ will become widely patent when the bolus reaches this area and impaired relaxation can be inferred when either a smooth tapering is noted at the EGJ or bolus transit across the EGJ is impeded. Alternatively, fluoroscopy will occasionally demonstrate spastic contractions, evident by a corkscrew appearance (Fig. 42-11).

Esophageal Manometry (High-Resolution Esophageal Pressure Topography)

Esophageal manometry is a test in which intraluminal pressure sensors, either water perfused or solid state, are positioned within the esophagus to quantify the contractile characteristics of the esophagus and segregate it into functional regions. The concept of high-resolution esophageal manometry is to use a sufficient number of pressure sensors within the esophagus such that intraluminal pressure can be monitored as a continuum along the length of the esophagus, much as time is viewed as a continuum in line tracings of conventional manometry such as those in Figure 42-10. When high-resolution manometry is coupled with sophisticated algorithms to display the manometric data as pressure topography plots, esophageal contractility is visualized with isobaric conditions among sensors indicated by isocoloric regions on the pressure topography plots. Figure 42-12 depicts a normal swallow in a high-resolution esophageal pressure topography plot encompassing both sphincters (UES and LES) and the intervening esophagus. The relative timing of sphincter relaxation and segmental contraction as well as the position of the transition zone are all readily demonstrated. The manometric evaluation of deglutitive EGJ relaxation is probably the most important measurement made during clinical esophageal manometry. Incomplete EGJ relaxation is an essential feature in the diagnosis of achalasia and achalasia is not only the best-defined esophageal motor disorder, but also the one with the most specific treatments. Despite this cardinal significance, there was no convention for defining incomplete deglutitive EGJ relaxation with conventional manometry. Furthermore, numerous potential confounding factors exist including crural diaphragm contraction during relaxation, deglutitive esophageal shortening, hiatal hernia, sphincter radial asymmetry, and move-

mm Hg

Side-hole sensor #2 Ω#2

4000

0 40

100 4000

3.3 s

Ω#2

Side-hole sensor #1

1.5 s

40

Ω#1

Impedance rings Ω#1

mm Hg

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders

0

100

5.0 s

Sleeve Sleeve sensor

mm Hg mm Hg

8.3 s

100

40

4000 Ω#5

Side-hole sensor #5 Ω#5

100 4000

6.6 s

0

0

100 12.0 s

20 mm Hg

Side-hole sensor #4

0 100

Ω#4

Ω#4

Ω#3

Side-hole sensor #3

Ω#3

4000

mm Hg

100

0

0

13 Seconds Figure 42-10.  Representative physiologic data, modified to illustrate the relationship among videofluoroscopic, manometric, and impedance representations of esophageal peristalsis. Schematic drawing of placement of a combined manometry/intraluminal impedance monitoring system with five manometric side holes spaced 4 cm apart and a 6-cm sleeve sensor placed just distal to the last manometric port. The impedance rings (Ω) are also spaced 4 cm apart with the rings straddling the manometric ports. The arrows to the right panel point to the corresponding data tracings obtained from each combined manometry/impedance or sleeve recording site. The right panel illustrates the concurrent videofluoroscopic, manometric, and multichannel intraluminal impedance recordings of a 5-mL diatrizoate (Renografin) swallow that was completely cleared by one peristaltic sequence. Representative tracings from the videofluoroscopic sequence overlaid on the combined manometric/impedance tracing show the distribution of the bolus at the times indicated by the vertical arrows. At each recording site, the black line intersecting the pressure scale (mm Hg) on the left represents the manometric tracing and the blue line intersecting the impedance scale in ohms (W) on the right represents the impedance recording tracing. Bolus entry at each combined manometry/impedance recording site is signaled by a subtle increase in pressure (intrabolus pressure) and a sharp decrease in impedance. In this example, the bolus propagates past W#4 rapidly, as indicated by an abrupt reduction in impedance in W#2, W#3, and W#4 at time 1.5 seconds. Luminal closure and hence the tail of the barium bolus (inverted “V”) is evident at each recording site by the upstroke of the peristaltic contraction and an increase in recorded impedance. Hence, at 5 seconds, the peristaltic contraction was beginning at side-hole sensor #3, corresponding to an increase in impedance and the tail of the barium bolus at the same esophageal locus. Finally, after completion of the peristaltic contraction (time 12 seconds), all diatrizoate was in the stomach.

ment of the recording sensor relative to the EGJ.8 With high-resolution esophageal pressure topography, this situation is greatly improved. A study comparing criteria for detecting impaired deglutitive EGJ relaxation within the deglutitive relaxation window (see Fig. 42-12) in a large group of patients and control subjects concluded that the optimal measure for quantifying deglutitive relaxation was the integrated relaxation pressure (IRP), with normal being defined as 15 mm Hg or less.284 Conceptually, the IRP is the average EGJ pressure for the four seconds of greatest relaxation within the relaxation window. This single measure of deglutitive EGJ relaxation exhibited 98% sensitivity and 96% specificity for distinguishing well-defined achalasia patients from control subjects and patients with other diagnoses.284 Apart from improving the sensitivity of manometry in the detection of achalasia, high-resolution esophageal pressure topography has also defined a clinically relevant subclassification of achalasia.224 A diagnosis of achalasia requires both aperistalsis and impaired deglutitive EGJ relaxation.

In its most obvious form this occurs in the setting of eso­ phageal dilatation with negligible pressurization within the esophagus (Fig. 42-13A). However, despite there being no peristalsis, there can still be substantial pressurization within the esophagus. In fact, a very common pattern encountered is achalasia with esophageal compression and panesophageal pressurization (see Fig. 42-13B). The other, less common pattern is of spastic achalasia in which there is a spastic contraction within the distal esophageal segment (see Fig. 42-13C and D). In a series of 99 consecutive patients with newly diagnosed achalasia, 21 had the classical pattern in Figure 42-13A, 49 had the panesophageal pressurization pattern of Figure 42-13B, and 29 had the spastic achalasia pattern of Figure 42-13C.224 Logistic regression analysis found panesophageal pressurization (see Fig. 42-13B) to be a predictor of response to treatment, whereas spastic achalasia (see Fig. 42-13C) and pretreatment esophageal dilatation were predictive of a poorer response to treatment. Following the analysis of the EGJ, a swallow is further categorized by the characteristics of the distal esophageal

695

Section V  Esophagus contraction. This analysis is largely facilitated by the generation of a pressure topography plot highlighting the 30 mm Hg isobaric contour. Under circumstances of normal deglutitive EGJ relaxation, the 30 mm Hg pressure threshold reliably delineates the wavefront of the peristaltic contraction.285 Contractile front velocity (CFV) is calculated from the 30 mm Hg isobaric contour plots by calculating the slope of the line connecting the 30 mm Hg isobaric contour at the proximal margin of the first subsegment and the distal margin of the second subsegment (see Fig. 42-12). A CFV of more than 8 cm/second is indicative of a spastic contraction.285,286 Although the CFV is easily definable in the circumstance of normal EGJ relaxation, it is more complex when EGJ relaxation is impaired (Fig. 42-14). With impaired EGJ relaxation, there is compartmentalized pressurization between the contractile front of the distal esophageal contraction and the EGJ with a high intrabolus pressure residing between the two. In such instances, the slope of the 30 mm Hg isobaric contour is no longer indicative of the CFV but now indicates increased intrabolus pressure as a result of functional obstruction at the EGJ. In such circumstances, the algorithm for computing CFV defaults to computing the slope of an isobaric contour line of magnitude greater than the EGJ relaxation pressure (e.g., 50 mm Hg in a given patient) so as to consistently represent the timing of luminal closure (see Fig. 42-14).

Figure 42-11.  Corkscrew esophagus on barium esophagogram in a patient with symptomatic distal esophageal spasm.

Swallow 0 mm Hg

UES

>110

5

S1, Proximal esophageal segment

Figure 42-12.  High-resolution esophageal pressure topography spanning from the pharynx (locations 0 to 2 cm) to the stomach (locations 29 to 35 cm) of a normal subject with normal peristalsis and normal esophagogastric junction (EGJ) relaxation. The transition zone, demarcating the end of the proximal esophageal segment S1 (striated muscle) and the beginning of the distal esophageal segment S2 (smooth muscle), is readily identified as a pressure minimum. Note that the distal esophageal segment, in fact, has three subsegments (S2, S3, S4) within it, each with an identifiable pressure peak. S4, the lower esophageal sphincter (LES), contracts at the termination of peristalsis and then descends back to the level of the crural diaphragm as the period of swallow-related esophageal shortening ends. The onset of the deglutitive relaxation window is at the onset of upper esophageal sphincter (UES) relaxation, whereas the offset is 10 seconds later. The spatial domain within which EGJ relaxation is assessed (the eSleeve range) is user defined, spanning at least 6 cm (in this example, labeled 0 and 6 cm), depending on the extent of esophageal shortening after the swallow. The contractile front velocity (CFV) is the slope of the line connecting points on the 30 mm Hg isobaric contour at the proximal margin of S2 and the distal margin of S3.

Distal esophageal segment

S2 15 Contractile front velocity (CFV) = 15 cm/5s = 3 cm/s

20

25

S3

70

50

S4 (LES)

30

0 cm EGJ

30

Deglutitive EGJ relaxation window

10

6 cm 0s

35

90

Center of transition zone

10 Location along lumen (cm)

696

0

5

Crural diaphragm contractions 10 Time (sec)

10 s 15

eSleeve range 20

–10

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders ACHALASIA WITH COMPRESSION

Length along the esophagus (cm)

CLASSIC ACHALASIA 1

1 5 UES Relaxation

10

mm Hg 150 125

15

100

20

4s IRP = 42 mmHg

25

1.5s

2.5s

30 35

10

75

20

50

25

25 0 –10

30

Length along the esophagus (cm)

SPASTIC ACHALASIA (2D)

5

20

200

30

100 30 0

CFV = 25 10 cm/s

35 0

C

3

6

9

1

5 6

7 8 9 10

SPASTIC ACHALASIA (3D) mm Hg

5

>300 100

10 15

50 25

20

10

25

5 0

30 35

0

12 15 18 21

Time (sec)

50

Time (sec)

mm Hg

300

75 25 0 –10

B

15

100

35

400

10

125 30 mm Hg isobaric contour

0 1 2 3 4

Time (sec)

1

UES Relaxation

15

0 1 2 3 4 5 6 7 8 9 10 11

A

mm Hg 150

5

D

3

6

9

12

15

18

21

Time (sec)

Figure 42-13.  Achalasia subtypes are distinguished by three distinct manometric patterns of esophageal body contractility (panels A-C). In classic achalasia (panel A), there is no significant pressurization within the body of the esophagus and impaired esophagogastric junction (EGJ) relaxation. The integrated relaxation pressure (IRP) was 42 mm Hg in this example. Panel B represents a swallow from a patient with the “achalasia with compression” subtype exhibiting rapid panesophageal pressurization of the fluid column trapped between the sphincters as the esophagus shortens at seven to eight seconds. Panel C illustrates a pressure topography plot typical of spastic achalasia. Although this swallow is also associated with rapidly propagated pressurization, the pressurization is attributable to an abnormal lumen-obliterating contraction. A three-dimensional rendering of these same pressure data (panel D) illustrates the peaks and valleys of that spastic contraction (brown vs. red); this swallow would likely appear as a rosary-bead or corkscrew pattern on fluoroscopy. (Modified from Pandolfino JE, Kwiatek MA, Nealis T, et al. Achalasia: A new clinically relevant classification by high resolution manometry. Gastroenterology 2008; 135:1526-33.)

Apart from a rapid CFV, other common abnormalities of the distal esophageal contraction are weak or absent peristalsis. In such instances, the 30 mm Hg isobaric contour is either discontinuous or absent, reflective of either focal or diffuse hypotensive contraction within the distal segment. Each swallow is thus characterized as normal (intact 30 mm Hg isobaric contour and a CFV < 8 cm/second), hypotensive (3 cm or greater defect in the 30 mm Hg isobaric contour), or absent (complete failure of contraction with no pressure domain > 30 mm Hg). Weak or absent peristalsis is a risk factor for impaired bolus clearance, but, whether impaired bolus clearance occurs depends on the balance between the severity of weakness and the magnitude of outflow resistance at the EGJ.287 Once swallows are characterized by the integrity of deglutitive EGJ relaxation and normality of the CFV, the distal esophageal contraction is further analyzed for the vigor of contraction using a newly developed measure for high-resolution esophageal pressure topography, the distal contractile integral (DCI). The DCI integrates the

length, vigor, and persistence of the two subsegments of the distal esophageal segment contraction, expressed as mm Hg·s·cm. A DCI value greater than 5000 mm Hg·s·cm is considered elevated.288 Adopting the nomenclature “nutcracker esophagus” from conventional manometry, this is the high-resolution manometry criterion defining hyper­ tensive peristalsis and was seen in 9% of a 400-patient series.285 However, there was substantial heterogeneity as to the locus of the hypertensive contraction within this group, potentially involving either or both of the subsegments within the distal esophageal contraction. Similarly, the LES can exhibit a hypertensive postdeglutitive contraction, defined as exceeding 180 mm Hg. Furthermore, one particularly interesting subgroup, defined by having a higher threshold DCI (>8000 mm Hg·s·cm), exhibited repetitive high-amplitude contractions and was clinically distinguishable by the uniform association with dysphagia or chest pain. Similar to DES, this “spastic nutcracker” pattern is very rare, found in only 12 (3%) of this 400-patient series (Fig. 42-15).

697

Section V  Esophagus HIGH IBP, NORMAL CFV (at 50 mm Hg IBC)

Length along the esophagus (cm)

0 Pressure (mm Hg) 150

5 10

120 15

90

20

60 30 mm Hg isobaric contour

25

A

30 0

50 mm Hg isobaric contour

30 35

0

5

10 Time (sec)

15

20

RAPID CFV, SPASTIC CONTRACTION Figure 42-14.  Differentiating increased intrabolus pressure (IBP) from a rapidly pro­ pagated spastic contraction (bottom). A, Swallow with functional obstruction at the eso­phagogastric junction (EGJ). Note that the 30 mm Hg isobaric contour (IBC) line deviates quickly from the 50 mm Hg isobaric contour line (arrows). In this case the contraction front velocity (CFV) was normal, reflecting the propagation velocity of the 50 mm Hg isobaric contour rather than the 30 mm Hg isobaric contour. B, Swallow with rapid CFV attributable to spasm. EGJ relaxation is normal and the 30 and 50 mm Hg isobaric contours parallel each other, indicating that no compartmentalized esophageal pressurization has occurred. The entire distal esophagus is contracting simultaneously. (Modified from Pandolfino JE, Ghosh SK, Rice J, et al. Classifying esophageal motility by pressure topography characteristics: A study of 400 patients and 75 controls. Am J Gastroenterol 2008; 103:27-37.)

0 Length along the esophagus (cm)

698

Pressure (mm Hg) 150

5 10

120 15 20 25

0

B

Following analysis of individual swallows by the criteria outlined, the component results are synthesized into a global manometric diagnosis by the criteria detailed in Table 42-2. Patients with normal EGJ relaxation, normal CFV, and a DCI less than 5000 mm Hg·s·cm are normal. The abnormalities encountered are described in specific functional terms with the intent that these then be interpreted within the clinical context of the patient. The classification detailed in Table 42-2 represents an incremental update on the Chicago classification,289 the task of a newly convened international working group focused on the standardization of the performance and interpretation of high-resolution esophageal pressure topography studies.

Intraluminal Impedance Measurement

60 30

50 mm Hg isobaric contour

30 35

90

30 mm Hg isobaric contour

Intraluminal impedance monitoring was described more than a decade ago as a method to assess intraluminal bolus transit without using fluoroscopy. The technique uses an intraluminal catheter with multiple, closely spaced pairs of

0

5

10

15

20

Time (sec)

metal rings (see Fig. 42-10). An alternating current is applied across each pair of adjacent rings and the resultant current flow between the rings is dependent on the impedance of the tissue and luminal content between the rings. Impedance decreases when the electrodes are bridged by liquid and increases when they are surrounded by air thereby providing data on the direction, content, and completeness of bolus transit. Validation data suggest that liquid bolus entry at the level of an electrode pair is indicated by a 50% drop in impedance and return of the impedance tracing to 50% of baseline correlates with the passage of the tail of the bolus on fluoroscopy, also indicated by the contractile upstroke noted on manometry (see Fig. 42-10). Validation studies of intraluminal impedance measurement against videofluoroscopy have shown excellent concordance in ascertaining bolus transit, reporting agreement in 97% (83/86) of swallows analyzed.290 Intraluminal impedance measurement has also recently been combined with manometry to assess the efficacy of

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders

Length along the esophagus (cm)

0 Pressure (mm Hg)

5

150

10

120 30 mm Hg

15

90

20

60

25 30 35

0

5

10

15

20

25

Table 42-2  The Chicago Classification of Distal Esophageal Motility Disorders With Normal EGJ Relaxation (Mean Integrated Relaxation Pressure <15 mm Hg)

DISORDER

CRITERIA

Aperistalsis

100% of swallows with absent peristalsis

Hypotensive peristalsis   Intermittent

30

  Frequent

0

Hypertensive peristalsis

30

Time (sec) Figure 42-15.  An uncommon spastic variant of hypertensive peristalsis (spastic nutcracker) identifiable by a distal contractile integral value greater than 8000 mm Hg⋅s⋅cm. In this example, the contraction did not meet contractile front velocity criteria for spasm (8 cm/s). The contraction has a spastic component that occurs after the wavefront propagates to the esophagogastric junction. Clinically these patients uniformly experience chest pain and dysphagia.

Spastic nutcracker Distal esophageal spasm (DES)

With Impaired EGJ Relaxation (Mean Integrated Relaxation Pressure ≥15 mm Hg)

DISORDER Achalasia Classic achalasia

esophageal emptying as a function of distal peristaltic amplitude. In a receiver operating characteristic (ROC) analysis of a large number of swallows, a 30 mm Hg cutoff had 85% sensitivity and 66% specificity for identifying incomplete bolus transit.291 With diminishing peristaltic amplitudes, the sensitivity progressively decreased and the specificity progressively increased. This study illustrates the complementary nature of manometry and impe­ dance testing in assessing esophageal function and may develop into a valuable clinical tool for the assessment of dysphagia.

Sensory Testing

Esophageal sensory nerves play a key role in determining symptoms of esophageal neuromuscular diseases because the esophagus is sensitive to a variety of stimuli including mechanical (elicited by luminal distention or highamplitude contractions), chemical (acid or other consti­ tuents of reflux), and temperature.292 Typically the visceral input is not perceived consciously. However, some patients may experience symptoms attributed to hyperalgesia (exaggerated pain perception) or allodynia (perception of pain to a stimulus that is usually not painful).155,293 Esophageal symptoms may be described as burning, pressing, pricking, or heat sensations. Nevertheless, these symptoms are not specific to a given stimulus, and substantial overlap in perception among stimuli is common. Although the precise mechanism by which an esophageal stimulus causes pain or the perception of dysphagia is unclear, methodologies devised to evoke or stimulate pain by simulating physiologic events are available to assess the possible relationship between ongoing symptoms and suspected causes. These tests typically use forms of distention studies (balloon, barostat, impedance planimetry, or volume challenges) or direct mucosal stimulation (chemical, electrical, or thermal). Balloon distention studies have shown that esophageal distention can provoke chest pain and that

More than 30% of swallows with peristaltic defects ≥3 cm in 30 mm Hg pressure isocontour 70% or more of swallows with peristaltic defects ≥3 cm in 30 mm Hg pressure isocontour Normal CFV, mean DCI >5000 and <8000 mm Hg⋅s⋅cm or LES after-contraction >180 mm Hg Normal CFV, mean DCI >8000 mm Hg⋅s⋅cm (Fig. 42-15) Normal EGJ relaxation and spasm (CFV >8 cm/s) with ≥20% of swallows (see Fig. 42-14B)

Achalasia with esophageal compression Spastic achalasia

Functional EGJ obstruction*

CRITERIA Impaired EGJ relaxation and aperistalsis (see Fig 42-13A) Impaired EGJ relaxation, aperistalsis, and panesophageal pressurization with ≥20% of swallows (see Fig. 42-13B) Impaired EGJ relaxation, aperistalsis, and spasm (CFV >8 cm/s) with ≥20% of swallows (see Fig. 42-13C and D) IBP >30 mm Hg compartmentalized between the peristaltic wavefront (normal or nutcracker) and EGJ (see Fig. 42-14A)

*May represent an achalasia variant. CFV, contractile front velocity; DCI, distal contractile integral; EGJ, esophagogastric junction; IBP, intrabolus pressure; LES, lower esophageal sphincter.

patients with esophageal chest pain tend to have lower threshold volumes for both first perception and first pain perception compared with controls.294,295 Combining impedance planimetry with balloon distention allowed other investigators to correlate biomechanical properties of the esophagus with the generation of chest pain.296,297 Results of those studies suggested that the tension-strain curve in chest pain patients was shifted to the left when compared with controls, a finding consistent with reduced compliance of the esophageal wall.269 The standard test of chemosensitivity is the Bernstein test wherein 0.1 normal hydrochloric acid is perfused in the esophagus to reproduce chest pain or heartburn. Typically, acid infusion is alternated with saline perfusion in a blinded fashion to increase the objectivity of the test, but no standardized protocol exists. Beyond the standard Bernstein perfusion test to assess esophageal sensitivity to acid, newer probes have been devised to test esophageal responsiveness to thermal challenges and transmucosal electrical nerve stimulation. However, although these tools have unquestionably been useful in improving our understanding of the

699

700

Section V  Esophagus interaction between peripheral receptors and central pain perception, their clinical utility remains limited owing to the lack of protocol standardization and the somewhat cumbersome nature of the studies. Currently, use of these devices is limited to subspecialty centers and further refinement will be required before mainstream clinical use can be advocated.

TREATMENT Oropharyngeal Dysphagia

Management of oropharyngeal dysphagia is focused on four specific issues: (1) identification of an underlying systemic disease, (2) characterization of a disorder amenable to surgery or dilation, (3) identification of specific patterns of dysphagia amenable to swallowing therapy, and (4) assessment of aspiration risk. Identifying Underlying Disease A potential outcome of the evaluation is the identification of an underlying neuromuscular, neoplastic, or metabolic disorder that dictates specific management. For example, dysphagia can be the presenting symptom in patients with myopathy, myasthenia, thyrotoxicosis, motor neuron disease, or Parkinson’s disease. In each instance, managing the underlying disease requires a specific treatment. Whether or not treatment of the underlying disorder improves swallowing function depends on the natural history of the specific disease and whether effective treatment exists. Disorders Amenable to Surgery The most common surgical treatment for oropharyngeal dysphagia is cricopharyngeal myotomy but the efficacy of myotomy in neurogenic or myogenic dysphagia is variable. Most series evaluating the efficacy of myotomy in these circumstances are uncontrolled and lack validated or even specific outcome measures. Thus, although an overall favorable response rate in excess of 60% is reported in this literature, there are no validated criteria for patient selection. Theoretically, the functional limitation faced by patients with neurogenic or myogenic dysphagia is of weak pharyngeal propulsion and the potential benefit of myotomy in that circumstance is less obvious than in the case of obstruction at the level of the cricopharyngeus.298 Patterns of Oropharyngeal Dysphagia Amenable to Swallow Therapy Identifying potential treatments for oropharyngeal dysphagia begins with definition of the aberrant physiology as categorized in Table 42-1. This is best accomplished with a videofluoroscopic swallowing study that first characterizes a patient’s swallow dysfunction and then proceeds to test the effectiveness of selected compensatory or therapeutic treatment strategies. Compensatory treatments include postural changes, modifying food delivery or consistency, or the use of prosthetics. For instance, head turning can eliminate aspiration or pharyngeal residue by favoring the more functional side in patients with hemiparesis.17 Similarly, diet modifications can reduce the “difficulty” of the swallow. Therapeutic strategies are designed to alter the physiology of the swallow, usually by improving the range of motion of oral or pharyngeal structures using voluntary control of oropharyngeal movement during a swallow. Depending on the severity of the impairment, level of motivation, and global neurologic integrity, defective elements of the swallow can be selectively rehabilitated. For a detailed description of the techniques and limitations of swallow therapy, the reader is referred to treatises on the topic.17,299

Evaluating Aspiration Risk Oropharyngeal dysphagia is responsible for an estimated 40,000 deaths a year due to aspiration pneumonia.300 Videoflouroscopy is considered the most sensitive test for detecting aspiration, reportedly detecting instances not evident by bedside evaluation in 42% to 60% of patients. However, despite the logical association between deglutitive aspiration and the subsequent development of pneumonia, this sequence is not inevitable. In fact, available data suggest that radiographic aspiration has a positive predictive value of only 19% to 68% and a negative predictive value of 55% to 97% for pneumonia.300 Nonetheless, the balance of evidence suggests that detection of aspiration is a predictor of pneumonia risk, and that its detection dictates that compensatory swallowing strategies, non-oral feeding or corrective surgery be instituted. Whether non-oral feeding eliminates the risk of aspiration is controversial. In one study of 22 patients with radiographic aspiration, pneumonia and death were more frequent among patients who received feeding tubes.185 This suggests that aspiration of oral secretions may be the essential element in pneumonia risk and has led some to consider procedures such as tracheostomy to protect the airway.

Hypopharyngeal (Zenker’s) Diverticulum and Cricopharyngeal Bar

The treatment of hypopharyngeal diverticulum is cricopharyngeal myotomy with or without a diverticulectomy (see Chapter 23). Cricopharyngeal myotomy reduces both the resting sphincter tone and resistance to flow across the UES. A study found that the compliance of the sphincter following diverticulectomy with myotomy was restored to normal following surgery, as indicated by normal hypopharyngeal intrabolus pressure during swallowing.301 Good or excellent results are reported in 80% to 100% of Zenker’s patients treated by transcervical myotomy combined with diverticulectomy or diverticulopexy.299 There are instances in which a limited procedure would be adequate, but a definitive approach to the problem of pulsion diverticula should generally involve myotomy and diverticulectomy. Diverticulectomy alone risks recurrence because the underlying stenosis at the level of the cricopharyngeus is not remedied. Similarly, myotomy alone may not solve the problem of food accumulation within the diverticulum, with attendant regurgitation and aspiration. Small diverticula may, however, disappear spontaneously following myotomy. A more recent trend is to treat Zenker’s diverticula via either rigid or flexible endoscopy. With both techniques, the principle is to divide the septum between the lumen of the diverticulum and the lumen of the esophagus. The division allows food and liquid to flow out of the diverticulum distal to the cricopharyngeus (which was within the septum) rather than to accumulate within the diverticulum. In the case of rigid endoscopy the procedure is performed under general anesthesia with a stapling device. In the case of flexible endoscopy the procedure is performed under light sedation with a needle knife, argon plasma coagulation, or hot biopsy forceps. Controlled trials have not been done comparing the two procedures, but a recent summary of 376 reported cases treated with flexible endoscopic methods found treatment to result in clinical resolution in 43% to 100% of cases among series.184 Whether a cricopharyngeal bar in the absence of a diverticulum requires treatment is less clear. Certainly, if dysphagia is present and combined fluoroscopic/manometric analysis demonstrates reduced sphincter opening in conjunction with elevated upstream intrabolus pressure,

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders there is good rationale for treatment. One recent uncontrolled series suggests that in this scenario dilation with a large-caliber bougie may be efficacious in relieving dys­ phagia and this approach is certainly a reasonable treatment option prior to myotomy.302

Achalasia

Because the underlying neuropathology of achalasia cannot be corrected, treatment is directed at compensating for the poor esophageal emptying and preventing complications. In practical terms this amounts to reducing LES pressure so that gravity promotes esophageal emptying. Peristalsis is not restored with therapy. LES pressure can be reduced by pharmacologic therapy, forceful dilation, or surgical myotomy. Pharmacologic treatments, on the whole, are not very effective, making them more appropriate as temporizing maneuvers than definitive therapies. The definitive treatments of achalasia are disruption of the LES either surgically (Heller myotomy) or with a pneumatic dilator. Which of these is the optimal approach remains an issue of debate given the paucity of randomized controlled trials with accepted criteria for assessing efficacy. A further limitation of the previous studies is failing to stratify patients by disease severity or, as more recently defined, by disease subtype.224 High-resolution esophageal pressure topography allows the subtyping of achalasia into three distinct patterns: I, classic achalasia; II, achalasia with compression; and III, spastic achalasia (see Fig. 42-13). From a conceptual vantage point types I and II represent a continuum, with type II representing early disease before the progression of esophageal dilatation characteristic of type I. Type III, on the other hand, is a subtype characterized by spasm of the distal esophagus. The significance of these disease subtypes is in how differently they responded to therapy, be it botulinum toxin injection, pneumatic dilation, or Heller myotomy. In a series of 99 new cases of achalasia, the overall treatment response was 56% with type I, 96% with type II, and only 29% with type III. The literature pertinent to achalasia treatment is mainly composed of numerous uncontrolled case series using a variety of qualitative endpoints as indications of efficacy. As noted, there is also minimal standardization as to the criteria for defining achalasia, the disease severity included in one series versus another, or the technical details of how pneumatic dilation or Heller myotomy are performed. Furthermore, some series were collected prospectively, some retrospectively, and some a combination. Given all of these limitations, there is little merit to embarking on a detailed comparison of outcomes between techniques. The existing treatment data are summarized next. Pharmacologic Therapy Smooth muscle relaxants such as nitrates or calcium channel blockers, administered sublingually immediately prior to eating can relieve dysphagia in achalasia by reducing the LES pressure. Amyl nitrite,303 sublingual nitroglycerin, theophylline, and β2-adrenergic agonists304 have also been tried. The largest reported experience has been with isosorbide dinitrate (Isordil) and nifedipine.305 Isosorbide dinitrate, 5 to 10 mg sublingually before meals, reduces LES pressure by 66% for about 90 minutes, with the degree of dysphagia relief paralleling the magnitude of the LES response over the 19-month trial.306 Side effects, particularly headache, are common. Placebo-controlled trials have not been reported. Calcium channel blockers (diltiazem, nifedipine, verapamil) reduce LES pressure by 30% to 40% for more than

an hour.306,307 The largest clinical experience in achalasia has been with nifedipine (Procardia). Nifedipine, 10 mg sublingually (capsules are crushed in the mouth) administered before meals (30 to 40 mg per day) was studied in 29 patients with early achalasia (prior to esophageal dilatation) in a placebo-controlled trial. Nifedipine was significantly better that placebo (which had no benefit), with good results in 70% of achalasic patients followed for 6 to 18 months.305 However, subsequent placebo-controlled crossover trials have found only minimal benefit with nifedipine.308 Side effects of nifedipine include flushing, dizziness, headache, peripheral edema, and orthostasis. Sildenafil (Viagra) is another smooth muscle relaxant that can decrease LES pressure in patients with achalasia by blocking phosphodiesterase type 5, the enzyme that destroys cyclic guanosine monophosphate induced by NO. A double-blind placebo controlled trial found that 50 mg of sildenafil significantly reduced LES pressure and relaxation pressure when compared with placebo.309 The effect peaked at 15 to 20 minutes after administration and persisted for less than one hour. Although conceptually appealing, the practicality of using sildenafil clinically is limited by its cost that is rarely, if ever, covered by health care insurance. Botulinum Toxin Injection The initial landmark study of botulinum toxin in achalasia reported that intrasphincteric injection of 80 units of botulinum toxin decreased LES pressure by 33% and improved dysphagia in 66% of patients for a six-month period.310 Botulinum toxin irreversibly inhibits the release of acetylcholine from presynaptic cholinergic terminals, effectively eliminating the neurogenic component of LES pressure. However, because this inhibitory effect is eventually reversed by the growth of new axons, botulinum toxin is not a long-lasting therapy. The technique involves injecting divided doses of botulinum toxin into four quadrants of the LES with a sclerotherapy catheter. Side effects are rare, but include chest discomfort for several days and rash. Although many patients initially experience a good response, there is minimal continued efficacy at one year.311-313 Repeat injection can be effective for a reasonable subset of patients, but the injection leads to a local inflammatory reaction and fibrosis, ultimately limiting this strategy. Doses greater than 100 units do not have increased efficacy.314 Studies comparing botulinum toxin injection to pneumatic dilation suggest that the expense of repeated injection outweighs the potential economic benefits of added safety, unless the patient’s life expectancy is minimal.315 Thus, this option is mainly reserved for older adults or frail individuals who are poor risks for definitive treatments. Pneumatic Dilation Therapeutic dilation for achalasia requires distention of the LES to a diameter of at least 3 cm to produce a lasting reduction of LES pressure, presumably by partially disrupting the circular muscle of the sphincter. Dilation with an endoscope, standard bougies (up to 60 French), or with throughthe-scope balloon dilators (up to 2 cm) provides very temporary benefit at best. Only dilators specifically designed to treat achalasia achieve adequate diameter for lasting effectiveness. The basic element of an achalasia dilator is a long, noncompliant, cylindrical balloon that can be positioned across the LES fluoroscopically (Rigiflex dilator) or endoscopically (Witzel dilator) and then inflated to a characteristic diameter in a controlled fashion using a handheld manometer. There is general agreement that pneumatic dilation can be done on an outpatient basis with the patient

701

702

Section V  Esophagus under conscious sedation. The technique of pneumatic dilation is variable among practitioners in terms of patient preparation, parameters of balloon inflation, and postdilation monitoring. In patients with substantial esophageal retention, it is useful to impose a liquid diet for one or more days prior to the procedure. Reported balloon inflation periods range from several seconds to five minutes.316 Although there is minimal methodologic consistency among authors, a cautious approach of beginning with a smalldiameter dilator (3 cm) and progressing to larger diameters (3.5 and 4 cm) only when the smaller dilator proved ineffective is fairly universal. As for inflation pressures, these are of minimal relevance with modern noncompliant balloon dilators because they do not distend beyond their specified diameter regardless of inflation pressure. Hence, it is simply necessary to observe under fluoroscopy that the balloon is properly positioned to capture the LES, observed as the “waist” of the hourglass-shaped balloon silhouette and that the waist fully effaces as the inflation proceeds. As for technical details of the procedure other than balloon diameter, there is minimal evidence that they influence outcome. The major complication of pneumatic dilation is esophageal perforation (see Chapter 40); mortality is fortunately rare.317 The reported incidence of esophageal perforation consequent from pneumatic dilation ranges between 1% and 5%261,316 with a global average of 3%. Because most perforations are readily evident or at least suspected within an hour of the procedure, patients should be observed closely for signs of an esophageal leak for at least two hours after pneumatic dilation. Alternatively, some practitioners routinely obtain a fluoroscopic examination of the esophagus following pneumatic dilation to ensure that perforation has not occurred. Usually, water-soluble contrast is given first, followed by barium. If a perforation appears small and contained or intramural, conservative management in the hospital consisting of close observation while maintaining the patient on nothing per mouth status and administering intravenous antibiotics is appropriate.261 If a perforation is substantial, or if worsening pain and fever occur during observation of what was thought to be a small perforation, surgical repair should be pursued expediently. Patients with a perforation from pneumatic dilation that is recognized and promptly treated surgically (within six to eight hours) have outcomes comparable with those of patients undergoing elective Heller myotomy.318 The best predictor of efficacy following a pneumatic dilation is the postdilation LES pressure; neither sphincter relaxation nor peristaltic function is significantly changed. A postdilation LES pressure less than 10 mm Hg is associated with prolonged remission, whereas a postdilation LES pressure greater than 20 mm Hg predicts that little benefit will occur from the procedure.319 In instances of an unsatisfactory result, it is reasonable to perform a subsequent dilation within a matter of weeks using an incrementally larger dilator. If the benefit of dilation persisted for a year or more, it is neither unusual nor dangerous to repeat pneumatic dilation as necessary. The clinical efficacy of dilation has been reported to range from 32% to 98%.311 Patients having a poor initial result or rapid recurrence of symptoms have diminished likelihood of responding to additional dilations.311 Subsequent response to surgical myotomy is not influenced by the history of previous dilations.261 Heller Myotomy Current surgical procedures for treating achalasia are variations on the esophagomyotomy described by Heller in 1913 consisting of an anterior and posterior myotomy performed

through either a laparotomy or a thoracotomy.311 Subsequently, this was modified to an anterior myotomy via thoracotomy. The appeal of myotomy is that it offers a more predictable method of reducing LES pressure than does pneumatic dilation.320 Although clearly efficacious, open Heller myotomy is associated with considerable morbidity related to thoracotomy, which led most patients to pursue pneumatic dilation as the initial intervention. However, adoption of the laparoscopic approach for achalasia surgery has led many practitioners to reconsider this. Published series of the efficacy of Heller myotomy in treating achalasia report good to excellent results in 62% to 100% of patients, with persistent dysphagia troubling less than 10% of patients.311 Recent studies suggest that a laparoscopic approach is associated with similar efficacy, reduced morbidity, and shorter hospital stay when compared with myotomy via thoracotomy, laparotomy, or thoracoscopy.311,321-325 The overall mortality from Heller myotomy is less than 2%. Historically, postmyotomy reflux in achalasic patients could be particularly severe, making this a hotly disputed detail of the surgical technique.326 However, with the broad use of proton pump inhibitors, reflux is usually easily controlled, making these complications very unlikely. Thus, laparoscopic Heller myotomy combined with a partial fundoplication (Toupet or Dor) has become the preferred surgical procedure for achalasia. An unsatisfactory result following Heller myotomy can result from incomplete myotomy, scarring of the myotomy, functional esophageal obstruction from the antireflux component of the operation, paraesophageal hernia, or severe esophageal dilatation. Heller Myotomy versus Medical Treatment Although pharmacologic therapy is simple and safe, it is increasingly clear that this should be reserved for use as a temporizing measure while more definitive therapy is being considered. Thus, practically speaking, the therapeutic choice is between pneumatic dilation and laparoscopic Heller myotomy as the primary therapy for achalasia. However, there are as yet no prospective controlled trials comparing these treatments. One controlled trial compares pneumatic dilation to myotomy via thoracotomy. That study reported 95% symptom resolution with myotomy and 51% symptom resolution in the dilation group, but the study was criticized for the methodology of pneumatic dilation used.327 Most case series report symptom resolution in approximately 70% of patients with pneumatic dilation, substantially higher than the 51% reported in the controlled trial, but still substantially lower than that reported in uncontrolled series of laparoscopic Heller myotomy (85% to 91%). Furthermore, although laparoscopic Heller myotomy is invasive, its morbidity and mortality are low. On the other hand, pneumatic dilation has a reported perforation rate that averages 3%, an incidence that probably exceeds that sustained by clinicians with substantial experience. Even though these patients do well if the perforation is recognized and addressed promptly, they may require a thoracotomy. Thus, it appears that cogent arguments can be made for each of these therapies and likely one should assess the local skills available, as well a patient preference, in selecting the most appropriate initial therapy. Treatment Failures Persistent dysphagia after treatment suggests treatment failure and should be evaluated with some combination of endoscopy, esophageal manometry, and fluoroscopic imaging. Endoscopy may detect esophagitis, stricture, para-

Chapter 42  Esophageal Neuromuscular Function and Motility Disorders esophageal hernia, or anatomic deformity. Manometry may be useful to quantify residual LES pressure, with values exceeding 10 mm Hg arguing for further therapy targeting the LES. Fluoroscopy is useful to identify anatomic problems as well as to evaluate esophageal emptying by using a timed barium swallow, a standardized method of measuring the height of the esophageal barium column one and five minutes after ingestion.328 In some instances these evaluations will lead to further intervention. In the case of a patient not previously operated on this could potentially be either repeat dilation or Heller myotomy. In patients who have already undergone myotomy, detection of an excessively short myotomy or functional esophageal obstruction from the antireflux component of the surgery usually requires reoperation, but pneumatic dilation can be pursued as an alternative. Reoperation, in general, is less effective than an initial operation for any indication in achalasia.329 Occasionally patients fail to respond to optimally performed dilation or myotomy and require alternative approaches. In extremely advanced or refractory cases of achalasia, esophageal resection with gastric pull-up or interposition of a segment of transverse colon or small bowel may be the only surgical option.330 Indications for this intervention include unresolvable obstructive symptoms, starvation, chronic aspiration, cancer, and perforation during dilation. Although excellent long-term functional results can be achieved, the reported mortality rate of this surgery is about 4%, consistent with the mortality rate of esophagectomy done for other indications. Risk of Squamous Cell Cancer Numerous series report cases of squamous cell carcinoma developing in the achalasic esophagus (see Chapter 46).331 The relative risk of developing squamous cell cancer has been estimated to be 33-fold relative to the non-achalasic population.332 The pathogenesis of the carcinoma is obscure, but stasis esophagitis is the likely precipitating factor. The tumors develop many years after the diagnosis of achalasia and usually arise in a greatly dilated esophagus, often in the middle third of the esophagus. Symptoms attributable to the cancer can be delayed, and the neoplasms are often large and advanced at the time of detection. These considerations raise the issue of surveillance endoscopy in achalasic individuals to detect early squamous cell cancer. However, an elegant analysis of a database encompassing the entire Swedish population of 1062 achalasic patients suggests that after discounting incident carcinomas, the overall odds ratio of squamous cell cancer for these people compared with age-matched controls was 17, corresponding to a 0.15% incidence of squamous cell cancer among the achalasic subjects.333 The authors calculated that if surveillance endoscopy was done annually, 406 examinations would need to be done in men and 2220 in women before one potentially treatable tumor was found. However, even that calculation is optimistic given that detection of a small cancer in a massively dilated esophagus with retained food and stasis esophagitis is far from ensured. Given these considerations, a surveillance program is currently not the standard of practice.

Distal Esophageal Spasm

Despite the dogma of treatment with smooth muscle relaxants, minimal controlled data exist regarding pharmacologic therapy of DES. Long-term studies are not available, and the entire basis for this therapy is anecdotal. Furthermore, most instances of esophageal chest pain are attributable to reflux rather than DES, and reflux symptoms will likely be made

worse by treating with smooth muscle relaxants. Uncontrolled trials of small numbers of DES patients report clinical response to nitrates,334 calcium channel blockers,335 hydralazine,336 botulinum toxin,337 and anxiolytics. The only controlled trial showing efficacy was with the anxiolytic trazodone, suggesting that reassurance and control of anxiety are important therapeutic goals.338 Also consistent with that conclusion, success has been reported using behavioral modification and biofeedback.339 Although the rationale for dilation is unclear, use of bougie dilators has been suggested as a therapy for dys­ phagia or chest pain in patients with spastic disorders. However, in the only controlled trial of this therapy, dilation with an 8-mm “placebo” dilator was as effective as an 18-mm “therapeutic” dilator in producing transient symptom relief.340 Alternatively, pneumatic dilation has been used in DES patients with severe dysphagia. In one practitioner’s experience, 45% of DES patients noted relief from pneumatic dilation, compared with 80% of achalasic patients.261 In another series of nine patients with DES and LES dysfunction treated with pneumatic dilation, dysphagia but not chest pain was improved during 37 months of observation.341 However, it is not clear that the patients who benefited by pneumatic dilation in these series would not be more properly categorized as spastic achalasia, emphasizing the need for accurate manometric classification.224 If dysphagia becomes so severe in DES that weight loss is observed or if pain becomes unbearable, surgical therapy consisting of a Heller myotomy across the LES with proximal extension of the incision up the distal esophagus to include the involved area of spasm or even esoph­ agectomy should be considered.186,342 However, there are no controlled studies of these procedures in welldefined DES patients and the indication is, fortunately, extremely rare.

Esophageal Hypersensitivity

Therapies for esophageal motor disorders have traditionally centered on improving esophageal contractility and emptying. However, the efficacy of these therapies is very limited except in the instance of achalasia. More recently, there has been a paradigm shift with the realization that minor manometric findings formerly interpreted as indicative of symptomatic hypercontractile conditions were often an epiphenomenon indicative of hypersensitivity syndromes. Hence, there is substantial interest in developing treatments directed at reducing esophageal hypersensitivity, and a number of pharmacologic and behavioral therapies have been identified with the potential to modulate pain perception and improve esophageal symptoms associated with swallowing. Pharmacologic Treatments Numerous neuropeptides and pharmacologic agents can reduce chemical and mechanical visceral sensitivity suggesting a possible role in the treatment of esophageal pain syndromes. Although data on these agents specific to eso­ phageal motor disorders are sparse, there is substantial literature focused on the treatment of noncardiac chest pain with or without motor abnormalities, and it is reasonable to generalize these findings to the treatment of esophageal hypersensitivity (see Chapter 12). Antidepressants are the most common medications prescribed for visceral pain modulation or chest pain of esophageal origin. Among antidepressants, the tricyclic antidepressants (TCAs) are the best studied. The mechanism of action for this therapeutic benefit is unknown

703

704

Section V  Esophagus because these agents act centrally as well as peripherally and have multiple receptor targets (acetylcholine, histamine, α-adrenergic). In a randomized placebo-controlled study, imipramine at a dose of 50 mg nightly was shown to be effective in reducing chest pain in patients with normal coronary angiograms.343 Similar results have been reported with other TCAs, and treatment with these agents at doses lower than those used for mood altering effects is common. Typical starting doses for TCAs (amitriptyline, nortriptyline) are 10 to 25 mg at bedtime with escalation of 10- to 25-mg increments to a target of 50 to 75 mg.344 Low-dose trazodone also has been used to treat noncardiac chest pain associated with esophageal dysmotility.338 In a double-blind placebo-controlled study in patients with noncardiac chest pain, the group taking 100 to 150 mg of trazodone had significant symptomatic improvement and less residual distress related to their esophageal symptoms. Esophageal motor function was not altered. Recent data also support the effectiveness of selective serotonin reuptake inhibitors (SSRIs) in the treatment of esophageal hypersensitivity. Intravenous citalopram at a dose of 20 mg was studied in a randomized, double-blinded, crossover study and found to significantly reduce both chemical (acid perfusion) and mechanical (balloon distention) esophageal sensitivity.345 Although clinical trials are not yet available, mechanistic studies assessing other SSRIs have also yielded encouraging results. Along similar lines, there has been a substantial interest in developing serotonin (5-HT) medications.293,346 5-HT3 antagonists and 5-HT4 agonists have been the most extensively studied given their effects on gut motility and as treatments for nausea. Unfortunately, several of these medications have proven to have unacceptable risks related to cardiac dysrhythmias or gut ischemia that led to their withdrawal. Theophylline has shown promising effects in the treatment of noncardiac chest pain, presumably by adenosine receptor blockade. In a recent placebo-controlled doubleblind study, sensory and biomechanical properties of the esophagus were assessed using impedance planimetry in 16 patients with esophageal hypersensitivity.347 Chest pain thresholds increased after intravenous theophylline and the esophageal wall was shown to relax and become more distensible. In a parallel study using oral theophylline and placebo in 24 chest pain patients there was a significant reduction in chest pain episodes, chest pain duration, and chest pain severity in the theophylline group.347 Although limited, these are very promising results for patients

with symptoms thought to be attributable to mechanical hypersensitivity. Nonpharmacologic Treatments Although the link between esophageal hypersensitivity, psychological factors, and psychiatric abnormalities is unclear, therapy focused on reassurance, behavioral modification, and relaxation techniques may be helpful. These therapies will most likely benefit patients with comorbidities such as panic disorder, generalized anxiety, and depression. However, it is also possible that therapies using controlled breathing, relaxation techniques, or hypnotherapy may benefit patients with hypersensitivity by diverting mental attention and reducing hypervigilance for visceral stimuli. Well-performed prospective trials are necessary to define the clinical role of these therapies.

KEY REFERENCES

Behar J, Biancani P. Pathogenesis of simultaneous esophageal contractions in patients with motility disorders. Gastroenterology 1993; 105:111-18. (Ref 251.) Cook IJ, Kahrilas PJ. AGA technical review on management of oropharyngeal dysphagia. Gastroenterology 1999; 116:455-78. (Ref 299.) Kahrilas PJ, Clouse RE, Hogan WJ. American Gastroenterological Association technical review on the clinical use of esophageal manometry [comment]. Gastroenterology. 1994; 107:1865-84. (Ref 197.) Logemann J. Evaluation and treatment of swallowing disorders. Austin, Tex: Pro-Ed Inc; 1998. (Ref 17.) Massey BT, Dodds WJ, Hogan WJ, et al. Abnormal esophageal motility. An analysis of concurrent radiographic and manometric findings. Gastroenterology 1991; 101:344-54. (Ref 258.) Ott DJ, Richter JE, Chen YM, et al. Esophageal radiography and manometry: Correlation in 172 patients with dysphagia. AJR Am J Roentgenol 1987; 149:307-11. (Ref 282.) Pandolfino JE, Ghosh SK, Rice J, et al. Classifying esophageal motility by pressure topography characteristics: A study of 400 patients and 75 controls. Am J Gastroenterol 2008; 103:27-37. (Ref 285.) Pandolfino JE, Kwiatek MA, Nealis T, et al. Achalasia: A new clinically relevant classification by high-resolution manometry. Gastroenterology 2008; 135:1526-33. (Ref 224.) Sifrim D, Janssens J, Vantrappen G. A wave of inhibition precedes primary peristaltic contractions in the human esophagus. Gastroenterology 1992; 103:876-82. (Ref 46.) Vaezi MF, Richter JE, Wilcox CM, et al. Botulinum toxin versus pneumatic dilatation in the treatment of achalasia: A randomised trial. Gut 1999; 44:231-9. (Ref 312.) Vela MF, Richter JE, Wachsberger D, et al. Complexities of managing achalasia at a tertiary referral center: Use of pneumatic dilatation, Heller myotomy, and botulinum toxin injection. Am J Gastroenterol 2004; 99:1029-36. (Ref 329.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

43 Gastroesophageal Reflux Disease Joel E. Richter and Frank K. Friedenberg

CHAPTER OUTLINE Epidemiology  705 Health Care Impact  707 Pathogenesis  707 Antireflux Barriers  707 Mechanisms of Reflux  708 Hiatal Hernia  710 Esophageal Acid Clearance  710 Gastric Factors  713 Clinical Features  713 Classic Reflux Symptoms  713 Extraesophageal Manifestations  714 Differential Diagnosis  715 Associated Conditions  715 Diagnosis  715 Empirical Trial of Acid Suppression  715 Endoscopy  716 Esophageal Biopsy  716 Esophageal pH Monitoring  717

Gastroesophageal reflux disease (GERD) is a consequence of the failure of the normal antireflux barrier to protect against frequent and abnormal amounts of gastroesophageal reflux (GER; i.e., gastric contents moving retrograde effortlessly from the stomach to the esophagus). GER itself is not a disease but rather a normal physiologic process. It occurs multiple times each day, especially after large meals, without producing symptoms or mucosal damage. In contrast, GERD is a spectrum of disease usually producing symptoms of heartburn and acid regurgitation. Most patients have no visible mucosal damage at the time of endoscopy (nonerosive GERD), whereas others have esophagitis, peptic strictures, or Barrett’s esophagus. Symptoms may include chest pain or evidence of extraesophageal manifestations such as pulmonary, ear, nose, or throat symptoms. GERD is a multifactorial process and one of the most common diseases of mankind. It greatly affects health care, contributing to the expenditure in the United States of nearly 12 billion dollars per year for antacid medications.

EPIDEMIOLOGY Although GERD is widely reported to be one of the most prevalent diseases of the gastrointestinal tract, prevalence data are based primarily on estimates rather than actual data. Furthermore, estimates differ depending on whether the analysis is based on symptoms (usually heartburn) or signs of disease (i.e., esophagitis).

Barium Esophagogram  719 Esophageal Manometry  719 Clinical Course  719 Nonerosive Reflux Disease  719 Erosive Reflux Disease  720 Complications  720 Hemorrhage, Ulcers, and Perforation  720 Peptic Esophageal Strictures  720 Barrett’s Esophagus  720 Treatment of Uncomplicated Disease  720 Nonprescription Therapies  721 Prescription Medications  721 Maintenance Therapies  723 Surgical Therapy  724 Endoscopic Therapy  725 Treatment of Complications  725 Chest Pain and Extraesophageal Manifestations  725 Peptic Esophageal Strictures  726

On the basis of symptoms, GERD is common in Western countries. In a nationwide population-based study by the Gallup Organization in the United States, 44% of the respondents reported heartburn at least once a month.1 More convincing data were obtained from a mailing of 2200 validated self-report questionnaires to a predominantly white population living in Olmsted County, Minnesota.2 The prevalence of heartburn and acid regurgitation in the past year was 42% and 45%, respectively. Symptoms that occurred at least weekly were reported by 20% of respondents, with an equal gender distribution across all ages. Most subjects reported their heartburn as being moderately severe, with a duration of 5 years or more, and only 5.4% had seen a physician for their reflux symptoms within the past year. More varying prevalence rates for symptomatic GERD have been reported from Europe, ranging from 5% in Switzerland to 27% in Finland.3 In contrast, the true prevalence of esophagitis is very difficult to define because healthy subjects rarely undergo upper endoscopy. Studies suggest that 7% of Americans have erosive esophagitis, whereas European studies identify prevalence rates ranging from 2% to 10%.4 GERD affects nearly equal proportions of men and women, but a male predominance occurs in esophagitis and Barrett’s esophagus.4 Increasing age is an important factor in the prevalence of GERD complications, probably the result of cumulative acid injury over time to the esophagus (Fig. 43-1).5,6 The prevalence of GERD only recently has been studied in multiracial populations. In a cross-sectional survey among employees at a Houston Veterans Affairs hospital,

705

Section V  Esophagus 50

400

400

350

350

300

300

250

250

200

200

Symptom severity

25 Esophagitis

15 10 5 0

100

>70

Age (years)

Figure 43-1.  Relationship of gastroesophageal reflux disease (GERD) symptom severity or presence of esophagitis with age in a large population study. Although the severity of GERD symptoms decreased with older age, the prevalence of esophagitis increased with age. (From Johnson DA, Fennerty MB. Heartburn severity underestimates erosive esophagitis severity in elderly patients with gastroesophageal reflux disease. Gastroenterology 2004; 126:660-4.)

the prevalence of heartburn was similar (23% to 27%) across ethnic groups including African Americans, Hispanics, Asians, and whites. However, African Americans had significantly less esophagitis than whites (24% versus 50%) for the same severity of symptoms (weekly or more).7 A study from Boston reviewed endoscopic reports from nearly 2500 consecutive patients, finding complicated GERD in 12% of white patients, 3% of African American patients, and 2% of Asian patients.8 The prevalence of GERD is relatively low among residents of Africa and Asia. For example, a cross-sectional study in Singapore reported prevalence rates for reflux symptoms of 7.5% in Indians, 0.8% in Chinese, and 3% in Malays.9 There have been exceptions such as the remarkable increase in the frequency of reflux symptoms seen in Japan and Singapore.10 An endoscopic, population-based study from South Korea encompassing more than 25,000 individuals, found the prevalence of erosive esophagitis to be 8%, whereas nonerosive reflux disease occurred in 4% of examined individuals.11 More than 90 % of subjects with erosions had mild disease, consistent with previous endoscopic studies from Asia. A recent systematic review that summarized trend data from longitudinal population-based studies performed in Asia failed to demonstrate an increase in prevalence over the past decade.12 Possible reasons for the lower GERD prevalence include low dietary fat; low body mass index (BMI); and lower gastric acid output, possibly related to Helicobacter pylori infection.11,13 The prevalence of GERD has been increasing in Western countries over the past 30 years.14 El Serag and Sonnenberg observed opposing time trends in the prevalence of peptic ulcer disease and GERD in the United States. Rates of duodenal ulcer fell between 1970 and 1995, while the prevalence of GERD and esophageal adenocarcinoma rose significantly (Fig. 43-2).15 The authors speculated that the decreasing prevalence of H. pylori may be playing a con-

50

Non-white men

50

Non-white men

0

0 1975–79

61–70

100

1970–74

41–50 51–60

150

1990–95

31–40

150

1985–89

21–30

White men

1980–84

<21

White men

1970–74

20

1990–95

Rate per 10,000 hospitalizations

30

1985–89

35

1980–84

40

1975–79

45

Patients (%)

706

Duodenal ulcer Reflux disease Figure 43-2.  Opposing time trends in the rates of hospitalization for duodenal ulcer and gastroesophageal reflux disease. Hospitalization rates were analyzed using the computerized database from the U.S. Department of Veterans Affairs. (Modified from El-Serag HB, Sonnenberg A. Opposing time trends of peptic ulcer and reflux disease. Gut 1998; 43:327.)

tributory role to the increasing prevalence of GERD in these regions. Recent data suggest that many patients with H. pylori–induced gastritis have involvement of the antrum and corpus, decreasing parietal cell mass, reducing acid secretion, and elevating gastric pH.13 This may have a protective effect on the esophageal mucosa in patients susceptible to GERD. An additional explanation for an increased prevalence of GERD in Western populations is the epidemic increase in obesity.16 In obese individuals (defined as a body mass index ≥30), epidemiologic studies suggest the prevalence of GERD is considerably higher than in the nonobese population.4,5,16 Jacobson and associates looked at the participants in the Nurses’ Health Study and found a nearly linear increase in the adjusted odds ratio for reflux symptoms for each BMI stratum.17 Interestingly, even for those participants with a normal BMI (22.5 to 24.9 kg/m2), the risk was elevated relative to a control group having a BMI in the range of 20 to 22.4 kg/m2.17 A study from the Houston VA Medical Center found a linear relationship between BMI and weekly symptoms of heartburn or regurgitation.18 Data from large population-based studies in England and Germany have been similar.19,20 A Norwegian study suggested that the odds of developing GERD were higher in obese subjects, and the risk was greater in obese women compared with male participants.21 In contrast, a nationwide case-control study from Sweden, consisting primarily of older adult men, failed to find an association between obesity and GERD.22

Chapter 43  Gastroesophageal Reflux Disease Central adiposity, as measured by the waist-to-hip ratio, may be more important than BMI in the pathogenesis of GERD. A large study from the Kaiser Permanente health system found a significant relationship between increased abdominal diameter and reflux symptoms independent of BMI.23 Similarly, El-Serag and colleagues found that the relationship between increasing BMI and increased acid exposure in the distal esophagus was primarily explained by the subject’s waist circumference.24 Obesity appears to be associated with complications related to long-standing GERD such as erosive esophagitis, Barrett’s esophagus, and esophageal adenocarcinoma.18,25-27 In a Swedish case-control study, researchers identified an association between esophageal adenocarcinoma and an individual’s BMI 20 years prior to the development of malignancy.28 Another case-control study demonstrated that central adiposity rather than BMI was associated with the presence of Barrett’s esophagus, particularly long-segment disease.29 Several mechanisms have been proposed to explain the association between obesity and GERD. These include an increased prevalence of esophageal motor disorders, diminished lower esophageal sphincter pressure, increased prevalence of hiatal hernia, and increased intragastric pressure (particularly with central obesity).30 In addition, visceral fat is metabolically active and produces a variety of cytokines including interleukin-6 (IL-6) and tumor necrosis factoralpha (TNF-α), which may affect the function of the lower esophageal sphincter. Along with environmental factors, the epidemiology of GERD may be affected by genetics. Family clustering of GERD and its complications, especially Barrett’s esophagus, have been reported.31,32 Especially exciting are the observations from two large case-control studies of twins from the United States and Sweden33,34 suggesting that genetic liability for GERD, as defined by frequent symptoms, is in the range of 30% to 45%. Although one group defined a locus on chromosome 13 associated with severe pediatric GERD,35 this has not been confirmed by other pediatric researchers36 and not yet evaluated in adults. The genetic mechanisms are unknown but may be related to a smooth muscle disorder associated with hiatal hernia, reduced lower esophageal sphincter (LES) pressure, and impaired esophageal motility.32

HEALTH CARE IMPACT Although rarely a cause of death, GERD is associated with considerable morbidity and complications, such as esophageal ulceration (5%), peptic stricture (4% to 20%), and Barrett’s esophagus (8% to 20%).5 Not surprisingly, the burden of GERD on health care is great. In 2004, GERD was by far the most common digestive disease diagnosis during ambulatory care visits, constituting 17.5% of all GI diagnoses. There were 6 or more outpatient visits with a GERD diagnosis listed per 100 people in the United States.37 GERD was the 10th most common inpatient GI diagnosis, with an estimated total number of discharges of 95,000 per year, a two day median length of stay, and median charges of $8060.38 In summary, GERD was the second most costly GI disease in 2004, behind liver disease, with total direct and indirect costs of nearly $12.6 billion. More than 60 million prescriptions for the treatment of GERD were estimated to be filled at retail pharmacies in 2004, representing 48% of all prescriptions for GI disorders and more than 50% of their costs. The large majority of prescriptions and their costs

were for proton pump inhibitors, which were the five most commonly prescribed and costliest GI medications.37 In 2004, two PPIs (lansoprazole and esomeprazole) were the second and fourth, respectively, top selling drugs of all classes in the United States. A recent economic survey from Germany reported that 6% of individuals with established GERD missed at least one day of work per year due to this disorder. Sixty-one percent of these patients visited their physician at least once in the previous year and 2% were hospitalized specifically for GERD.39 They estimated direct and indirect costs of approximately $600 per patient per year. Furthermore, GERD as a chronic disease significantly impairs quality of life. Compared with other chronic medical conditions, this impairment is similar to, or even greater than, that from arthritis, myocardial infarction, heart failure, or hypertension.40 GERD comorbidities are common and include irritable bowel syndrome and psychological distress in 36% and 41% of patients, respectively.41 These comorbidities potentiate the negative effect on quality of life seen with GERD, and affect the response to treatment with proton pump inhibitors.

PATHOGENESIS The pathogenesis of GERD is complex, resulting from an imbalance between defensive factors protecting the esophagus (antireflux barriers, esophageal acid clearance, tissue resistance) and aggressive factors refluxing from the stomach (gastric acidity, volume, and duodenal contents).

ANTIREFLUX BARRIERS

The first tier of the three-tiered esophageal defense against acid damage, the antireflux barriers, is an anatomically complex region including the intrinsic lower esophageal sphincter (LES), diaphragmatic crura, the intra-abdominal location of the LES, the phrenoesophageal ligaments, and the acute angle of His (Fig. 43-3). The LES involves the distal 3 to 4  cm of the esophagus and at rest is tonically contracted.42 It is the major component of the antireflux barrier, being capable of preventing reflux even when completely displaced from the diaphragmatic crura by a hiatal hernia.43 The proximal portion of the LES is normally 1.5 to 2  cm above the squamocolumnar junction, whereas the distal segment, about 2 cm in length, lies within the abdominal cavity. This location maintains gastroesophageal competence during intra-abdominal pressure excursions. Resting LES pressure ranges from 10 to 30 mm Hg with a generous reserve capacity because only a pressure of 5 to 10  mm  Hg is necessary to prevent GER.44 The LES maintains a high-pressure zone by the intrinsic tone of its muscle and by cholinergic excitatory neurons.45,46 There is considerable diurnal variation in basal LES pressure; it is lowest after meals and highest at night, and large increases occur with phase III of the migrating motor complex. It is also influenced by circulating peptides and hormones, foods (particularly fat), as well as a number of drugs (Table 43-1) (see also Chapter 48). The LES lies within the hiatus created by the right crus of the diaphragm and is anchored by the phrenoesophageal ligaments, which insert at the level of the squamocolumnar junction (see Fig. 43-3). Developmentally, the crural diaphragm arises from the dorsal mesentery of the esophagus and is innervated separately from the costal diaphragm. It is inhibited by esophageal distention, vomiting, and during transient LES relaxations (tLESRs), but not during swallow-

707

708

Section V  Esophagus Table 43-1  Modulators of Lower Esophageal Sphincter Pressure INCREASE LES PRESSURE

DECREASE LES PRESSURE

Hormones/peptides

Gastrin Motilin Substance P

Secretin Cholecystokinin Somatostatin VIP

Neural agents

α-Adrenergic agonists β-Adrenergic antagonists Cholinergic agonists

α-Adrenergic antagonists β-Adrenergic agonists Cholinergic antagonists

Foods

Protein

Fat Chocolate Peppermint

Other factors

Histamine Antacids Metoclopramide Domperidone Cisapride Prostaglandin F2α Baclofen

Theophylline Prostaglandins E2 and I2 Serotonin Meperidine Morphine Dopamine Calcium channel blockers Diazepam Barbiturates

LES, lower esophageal sphincter; VIP, vasoactive intestinal peptide.

MECHANISMS OF REFLUX Transient Lower Esophageal Sphincter Relaxations

Lower esophageal sphincter

Crural diaphragm

Costal diaphragm

Phrenoesophageal ligament Intra-abdominal esophagus Angle of His

Figure 43-3.  Anatomy of the gastroesophageal junction illustrating the major elements of the antireflux barrier.

ing. The crural diaphragm provides extrinsic squeeze to the intrinsic LES, contributing to resting pressure during inspiration and augmenting LES pressure during periods of increased abdominal pressure, such as with coughing, sneezing, or bending.47 Crural contractions impose rhythmic pressure increases of about 5 to 10 mm Hg on the LES pressure recording. During deep inspirations and some periods of increased abdominal straining, these changes may lead to pressures of 50 to 150 mm Hg.48 The oblique entrance of the esophagus into the stomach creates a sharp angle on the greater curve aspect of the gastroesophageal junction, the angle of His. This angle has been shown in cadavers to create a flap valve effect that contributes to gastroesophageal junction competency.49

tLESRs are the most frequent mechanism for reflux in patients with healthy sphincter pressures. Figure 43-4 illustrates a transient LESR and highlights differences from swallow-induced LESRs. tLESRs occur independently of swallowing, are not accompanied by esophageal peri­ stalsis, persist longer (>10 seconds) than swallow-induced LESRs, and are accompanied by inhibition of the crural diaphragm.50 tLESRs account for nearly all reflux episodes in healthy subjects and 50% to 80% of episodes in GERD patients, depending on the severity of associated esoph­ agitis (Fig. 43-5).51 However, one study suggests that low basal LES pressure, rather than tLESRs, may be the primary mechanism of GER in patients with nonreducible hiatal hernias.52 tLESRs are not always associated with GER. In normal subjects 40% to 60% of tLESRs are accompanied by reflux episodes, compared with 60% to 70% in GERD patients.45,51,53 Possible factors determining whether reflux occurs include abdominal straining, presence of a hiatal hernia, degree of esophageal shortening, and duration of tLESRs. The dominant stimulus for tLESR is distention of the proximal stomach by either food or gas,54,55 which is not surprising given that a tLESR is the mechanism of belching. More varying stimuli are fat, stress, and subthreshold (for swallowing) stimulation of the pharynx.49 Various drugs may impair tLESRs including cholecystokinin A (CCK-1) receptor antagonists, anticholinergic drugs, morphine, somato­ statin, nitric oxide inhibitors, 5-hydroxytryptamine (5-HT)3 antagonists, and γ-aminobutyric acid (GABAB) agonists.56 Evidence indicates that tLESRs are mediated through vagal pathways.54 Gastric distention activates mechanoreceptors (intraganglionic lamellar endings) adjacent to the gastric cardia, sending signals to the brainstem center via vagal afferent pathways.57 The structured sequence of motor events including LESR, crural diaphragm inhibition, and secondary esophageal peristalsis suggests that this process occurs in a programmed manner, probably controlled by a pattern generator within the vagal nuclei. The motor arm is the vagus nerve sharing common elements with swallowinduced LESR.56

Chapter 43  Gastroesophageal Reflux Disease 7 Distal Esophageal 4 pH 1 + Submandibular 0 EMG –

Sw

mm Hg

Esophageal body

12 cm above LES 9 cm above LES 6 cm above LES

3 cm above LES

50

0 50

0 50

0 50

0 70

Sw tLESR

LES 0 40 Gastric 0

1 minute Figure 43-4.  Example of a transient lower esophageal sphincter (LES) relaxation (tLESR) on an esophageal manometry study. LES pressure is referenced to gastric pressure, which is indicated by the horizontal dashed line. Note that the tLESR persisted for almost 30 seconds, whereas the swallow-induced LES relaxation (LESR) to the right (Sw) persisted for only 5 seconds. Also note the absence of a submandibular electromyographic (EMG) signal during the tLESR, which indicates absence of a pharyngeal swallow. Finally, the associated esophageal motor activity is different in the two types of LESR: the swallow-induced relaxation is associated with primary peristalsis, whereas the tLESR is associated with a vigorous, repetitive “off contraction” throughout the esophageal body. (From Kahrilas PJ, Gupta RR. Mechanisms of reflux of acid associated with cigarette smoking. Gut 1990; 31:4.)

Swallow-Induced Lower Esophageal Sphincter Relaxations

About 5% to 10% of reflux episodes occur during swallowinduced LESRs. Most episodes are associated with defective or incomplete peristalsis.53 During a normal swallowinduced LESR, reflux is uncommon because (1) the crural diaphragm does not relax, (2) the duration of LESR is relatively short (5 to 10 seconds), and (3) reflux is prevented by the oncoming peristaltic wave (see Fig. 43-4). Reflux during swallow-induced LESRs is more common with a hiatal hernia. This may be due to the lower compliance of the esophagogastric junction in hernia patients, permitting it to open at pressures equal to or lower than intragastric pressure, thereby allowing reflux of gastric juices accumulating in the hiatal hernia.58,59

Hypotensive Lower Esophageal Sphincter Pressure

GER can occur in the context of a hypotensive LES by either strain-induced or free reflux.44,51 Strain-induced reflux

occurs when a relatively hypotensive LES is overcome and “blown open” by an abrupt increase in intra-abdominal pressure from coughing, straining, or bending over. This type of reflux is unlikely when the LES pressure is greater than 10  mm  Hg. Free reflux is characterized by a fall in intraesophageal pH without an identifiable change in intragastric pressure, usually occurring when LES pressure is less than 5 mm Hg. Reflux due to a low or absent LES pressure is uncommon. Mostly it occurs in patients with severe esophagitis and may account for up to 25% of reflux episodes (see Fig. 43-5); it rarely occurs in patients without esophagitis.45,51,60 The mechanisms responsible for idiopathic low LES pressure (i.e., not part of a systemic disease such as scleroderma) are poorly understood. The presence of a hiatal hernia reduces the pressure measured in the LES due to losing the intrinsic support of the crural diaphragm.44 Some LES weakness may be secondary to esophagitis impairing the excitatory cholinergic pathways to the LES. Induction of experimental esophagitis in cats attenuates the

709

Section V  Esophagus

100 90 Percentage of reflux episodes

710

Weakened and shortened LES

80 70

Loss of diaphragmatic support for the LES

60 50 40

Retention of gastric fluid in hernial sac

30 Control GERD Mild esophagitis Severe esophagitis

20 10 0

Swallow- Absent basal Straining induced LES pressure LESR Figure 43-5.  Proportion of reflux episodes in control subjects and in patients with gastroesophageal reflux disease (GERD) occurring by the following mechanisms: transient lower esophageal sphincter relaxation (tLESR), swallow-induced lower esophageal sphincter relaxation (LESR), absent basal LES pressure, and straining in the presence of low LES pressure. (From Holloway RH. The anti-reflux barrier and mechanisms of gastro-oesophageal reflux Ballieres Clin Gastroenterol 2000; 14:681.) tLESR

release of acetylcholine and lowers LES pressures—changes that are reversible on healing of the esophagitis.60 However, healing of esophagitis in humans is rarely accompanied by an increase in LES pressure.61

HIATAL HERNIA

The contribution of the hiatal hernia to GERD is controversial. Opinion has shifted widely from one that virtually equated hiatal hernia with reflux disease to one that denied it a causal role. Epidemiologic and physiologic data confirm the importance of the hiatal hernia in patients with more severe esophagitis, peptic stricture, or Barrett’s esophagus.62 Hiatal hernia occurs in 54% to 94% of patients with reflux esophagitis, a rate strikingly higher than that in the healthy population.63 Two studies have also found that in individuals with reflux symptoms, the presence of hiatal hernia confers a significantly increased risk of erosive esophageal injury.64 The hiatal hernia impairs LES function through several mechanisms, as well as impairing esophageal acid clearance (Fig. 43-6).65 Reflux is worse in patients having a “nonreducible” as opposed to a “reducible” hiatal hernia. Nonreducing hernias are those in which the gastric rugal folds remain above the diaphragm between swallows.62 Statistical modeling has revealed a significant interaction between hiatal hernia and LES pressure, such that the likelihood of GER is increased as basal LES pressure decreases, an effect substantially amplified by the presence of a hernia and as the hernia size increases.43 Displacement of the LES from the crural diaphragm into the chest reduces basal LES pressure and shortens the length of the high-pressure zone primarily due to the loss of the intra-abdominal LES segment.62 Hiatal hernia eliminates the increase of LES pressure that occurs during straining and increases tLESRs during gastric distention with gas.65,66 Large, nonreducible hernias also impair esophageal acid clearance because of an increased tendency for reflux

Loss of intraabdominal LES segment

Stretching and rupture of the phrenoesophageal ligament Widened diaphragmatic hiatus

Figure 43-6.  Schematic diagram showing the effect of a hiatal hernia on the antireflux barrier. LES, lower esophageal sphincter.

to occur from the hernia sac during swallow-induced LESRs.53 Finally, an alteration of esophagogastric junction compliance, especially in GERD patients with hiatal hernia, has been identified.59 For the same degree of intragastric pressure, the esophageal junction opens at a lower pressure and the cross-sectional area is greater and more symmetrical as intragastric pressure increases. These changes in compliance simulated a 10-fold increase in air and 6-fold increase in liquid reflux across the esophageal junction. The etiology of a hiatal hernia remains unclear. Familial clustering of GERD suggests the possibilities of an inherited smooth muscle disorder. Animal studies propose that reflux itself causes esophageal shortening promoting the development of a hiatal hernia.67 Other studies find an association with obesity68 and heavy lifting,69 raising the possibilities that over time chronic intra-abdominal stressors may weaken the esophageal hiatus, causing the development of a hiatal hernia. This theory is attractive as it helps to reconcile the increased prevalence of hiatal hernias as the population grows older.63

ESOPHAGEAL ACID CLEARANCE

The second tier against reflux damage is esophageal acid clearance. This phenomenon involves two related but separate processes: volume clearance, which is the actual removal of the reflux material from the esophagus, and acid clearance, which is the restoration of normal esophageal pH following acid exposure through titration with base from saliva and esophageal gland secretions. Although the competency of the antireflux barrier determines the frequency and volume of GER, esophageal acid clearance determines the duration of acid exposure to the mucosa and probably the severity of mucosal damage.

Volume Clearance

Esophageal peristalsis clears acid volume in the upright and supine positions but is inoperative during deep rapid-

Chapter 43  Gastroesophageal Reflux Disease clearance time and is most beneficial in patients with aperistalsis (i.e., scleroderma).72

Bolus volume (mL)

15

Salivary and Esophageal Gland Secretions

10

5

Pressure (mm Hg)

Esophageal pH

0 8 4 0

140

0 Acid bolus 0

DS

DS 1

DS

DS 2

DS

DS 3

Time (minutes) Figure 43-7.  Relationship between esophageal peristalsis, distal esophageal pH, esophageal emptying, and esophageal acid clearance during an acid clearance test performed with radiolabeled 0.1  N hydrochloric acid. Bolus volume within the esophagus is derived from scintiscanning over the chest. Note that, although all but 1 mL of the infused fluid is cleared from the esophagus by the first peristaltic contraction, the distal esophageal pH remains low for at least 3 minutes. Stepwise increases in distal esophageal pH occur with subsequent swallows. DS, dry swallow. (From Helm JF, Dodds WJ, Pek LR, et al. Effect of esophageal emptying and saliva on clearance of acid from the esophagus. N Engl J Med 1984; 310:284.)

Saliva is the second essential factor required for normal esophageal acid clearance. Compared with gastric acid, saliva is a weak base with a pH of 6.4 to 7.8.73 Although saliva is ineffective in neutralizing large acid volumes (5 to 10  mL), it easily neutralizes the small amount of acid remaining in the esophagus after several peristaltic contractions (see Fig. 43-7).70 The importance of saliva is supported by observation that increased salivation induced by oral lozenges or bethanechol significantly decreases acid clearance time. In contrast, suction aspiration of saliva markedly prolongs acid clearance, despite the presence of normal peristaltic contractions.73 Modulation of salivation may contribute to GERD. Decreased salivation during sleep is the reason that nocturnal reflux episodes are associated with markedly prolonged acid clearance times.74 Xerostomia (see Chapter 22) is associated with prolonged esophageal acid exposure and esophagitis.75 Cigarette smoking promotes GER. Originally attributed to nicotine’s effect on lowering LES pressure, cigarette smokers also have prolonged esophageal acid clearance times due to hyposalivation.76 Finally, the esophagosalivary reflex is impaired in patients with reflux esophagitis and individuals with strictures.77 This is a vasovagal reflex demonstrated by perfusing acid into the esophagus, which stimulates salivation. This reflex explains the symptoms of water brash (copious salivation) observed in some reflux patients. In addition to saliva, the aqueous bicarbonate-rich secretions of the esophageal submucosal glands dilute and neutralize residual esophageal acid.78 Acid refluxing into the esophageal lumen stimulates these glands and helps neutralize the acid, even if swallowing does not occur.79

Tissue Resistance

eye-movement (REM) sleep. Helm and colleagues70 showed that one or two primary peristaltic contractions completely clear a 15-mL fluid bolus from the esophagus (Fig. 43-7). Primary peristalsis is elicited by swallowing. Secondary peristalsis, initiated by esophageal distention from acid reflux, is much less effective in clearing the refluxate, thus offering only an ancillary protective role. Peristaltic dysfunction (i.e., failed peristaltic contractions and hypotensive [<30 mm Hg] peristaltic contractions that incompletely empty the esophagus) increases in frequency with the severity of esophagitis. Kahrilas and colleagues71 found that the prevalence of peristaltic dysfunction rose from 25% in individuals with mild esophagitis to more than 50% in patients with severe esophagitis. Whether esophagitis per se leads to peristaltic dysfunction or whether an underlying smooth muscle motility disorder of the esophagus predisposes to the development of reflux disease is not clear. Animal studies have found that esophageal dysmotility associated with active esophagitis is reversible, but esophageal dysmotility associated with stricture or extensive fibrosis is irreversible.60 Clinical observations suggest that impaired motor function does not revert to normal following either effective medical or surgical therapies.61 Gravity contributes to bolus clearance when reflux occurs in the upright position. At night when supine, this mechanism is not operative unless the head of the bed is elevated. This important lifestyle change markedly improves acid

Although clearance mechanisms minimize acid contact time with the epithelium, even healthy subjects have acid reflux during the day and sometimes at night. Nevertheless, only a few subjects experience symptomatic GER and even fewer suffer GERD. This is due to a third tier for esophageal defense, known as tissue resistance. Conceptually, tissue resistance can be subdivided into pre-epithelial, epithelial, and postepithelial factors, which act together to minimize mucosal damage from the noxious gastric refluxate.80 The pre-epithelial defense in the esophagus is poorly developed. There is neither a well-defined mucous layer nor buffering capacity by the surface cells to secrete bicarbonate ions into the unstirred water layer. This results in a lumento-surface pH gradient in the esophagus of only 1 : 10, in contrast with the stomach and duodenum, where the gradient can range from 1 : 1000 to 1 : 10,000.81 The epithelial defenses consist of structural and functional components. Structural components include the cell membranes and intercellular junctional complexes of the esophageal mucosa. As reviewed in Chapter 41, this structure is a 25- to 30-cell-thick layer of nonkeratinized squamous epithelium functionally divided into a proliferating basal cell layer (stratum basalis), a midzone layer of metabolically active squamous cells (stratum spinosum), and a 5- to 10-cell-thick layer of dead cells (stratum corneum). The esophageal mucosa is a relatively “tight” epithelium that resists ionic movement at the intercellular, as well as the cellular, level as the result of tight junctions and the matrix of lipid-rich glycoconjugates in the intercellular space.82 Luminal acid attacks the epithelial defenses by damaging

711

712

Section V  Esophagus A

B

C

D

Figure 43-8.  Micrographs of esophageal epithelium, obtained during transmission electron microscopy. Normal subjects (A) do not have dilated intercellular spaces. In contrast, patients with “bile” reflux (B), nonerosive gastroesophageal reflux disease (GERD) (C), and erosive gastroesophageal reflux disease (D) have dilated intercellular spaces (irregular white spaces). This dilation appears to be the earliest cellular marker of GERD and is independent of the degree of esophagitis. (From Calabrese C, Fabbri A, Bortolotti M, et al. Dilated intercellular spaces as a marker of oesophageal damage: Comparative results in gastro-oesophageal reflux disease with or without bile reflux. Aliment Pharmacol Ther 2003; 18:525.)

the intercellular junction, allowing hydrogen ions to enter and acidify the intercellular space. As documented by transmission electron microscopy, the intercellular spaces expand and eventually the buffering capacity of this space is overwhelmed, leading to acidification of the adjacent cytosol via the basolateral membrane.80 The functional components of tissue resistance include the ability of the esophageal epithelium to buffer and extrude hydrogen ions. Intracellular buffering is accomplished by negatively charged phosphates and proteins, as well as bicarbonate ions. When the mucosal buffering capacity is exceeded and intracellular pH falls, the epithelium has the capacity to actively remove or neutralize H+. This is possibly by the action of two transmembrane proteins, one an Na+/H+ exchanger and the other an Na+-dependent Cl−/HCO3− exchanger.83,84 After reflux-induced cell acidification, these transporters restore the intracellular pH to neutrality by exchanging H+ for extracellular Na+ or by exchanging Cl− for extracellular HCO3−, respectively. Additionally, esophageal cells contain within their membrane an Na+-independent Cl−/HCO3− exchanger that extrudes HCO3− from the cytoplasm when the intracellular pH is too high.83 When the epithelial cells are no longer able to maintain intracellular pH, they lose their ability to volume regulate, edema occurs, balloon cells develop, and cell death follows. Additional contributors to the epithelial defense include salivary epidermal growth factor, transforming growth factor-α, and prostaglandin E2. These factors enhance epithelial cell turnover, enhance esophageal mucin production, and modulate bicarbonate secretion.85

Data suggest that dilated intercellular spaces are the earliest markers of esophageal epithelial cellular damage (Fig. 43-8). These alterations arise with exposure to acid and pepsin during gastroesophageal reflux, but the exact pathway of damage to the intercellular junctions remains unclear and seems to be multifactorial.86 Other noxious contents of the refluxate, such as bile acids, are harmful, and dilated intercellular spaces can be induced by acute psychological stress.87 Dilated intercellular spaces can be assessed quantitatively with electron microscopy (EM), but they also are recognizable with light microscopy (LM). In studies by Calabrese and colleagues,88,89 all controls had intercellular spaces less than 1.69 µm. Symptomatic patients had a mean intercellular space value and a mean value of the maximum dilated intracellular space at least three times greater than controls. Statistical differences were not observed between esophagitis patients and nonerosive GERD patients. The authors speculated that increased paracellular permeability could partly explain the development of heartburn in the absence of overt esophagitis. This hypothesis is supported by the presence of sensory neuron receptors within the intercellular space, only a few cell layers from the esophageal lumen.90 Importantly, aggressive acid inhibition with proton pump inhibitors leads to complete resolution of the dilated intercellular spaces in nearly all patients over three to six months. These changes correlated closely with the resolution of heartburn.89 The postepithelial defense is provided by the esophageal blood supply. Blood flow delivers oxygen, nutrients, and bicarbonate and removes H+ and CO2, thereby maintaining

Chapter 43  Gastroesophageal Reflux Disease normal tissue acid-base balance. Blood flow to the esophageal mucosa increases in response to the stress of luminal acid.91 Cellular injury also stimulates cell proliferation, which results in thickening of the basal cell layer of the epithelium. Unlike the stomach, where superficial mucosal injury can be repaired in hours, the esophagus repairs itself more slowly, over days to weeks. Acid suppression with proton pump inhibitors has been shown to reverse the characteristic histologic changes of esophageal reflux including basal layer thickening and dilation of intercel­ lular spaces.77,92

GASTRIC FACTORS

Gastric factors (volume and components of the gastric refluxate) are potentially important in the production of reflux esophagitis. Gastric acidity determines the degree of potential mucosal damage of the refluxate. Increases in gastric volume augment the rate of tLESRs, making more gastric contents available for reflux.

Gastric Acid Secretion

Acid and activated pepsin are the key ingredients of the gastric refluxate producing esophagitis. In animal studies, acid alone causes minimal injury at a pH of less than 3, primarily by protein denaturization. However, acid combined with even small amounts of pepsin disrupts the mucosal barrier, resulting in increased H+ permeability, histologic changes, and hemorrhage.93 Supporting these animal studies, clinical series find that the degree of esophageal injury, from nonerosive GERD to Barrett’s esophagus, parallels the increase in the frequency and duration of acid reflux with a pH of less than 4.94,95 Conversely, per­ fusing the esophagus of animals with the pepsin solution of pH 4 to 7.5 produces minimal mucosal disruption or change in mucosal permeability.93 These observations are the cornerstone of acid inhibition therapy for the treatment of GERD. Overall, gastric acid secretion is normal in patients with GERD. For example, Hirschowitz96 compared the gastric acid secretion of 115 patients with esophagitis with more than 500 age-, gender-, and disease-matched controlled subjects without esophagitis. The average fasting basal and maximum secretions of both acid and pepsin were the same in both groups, and esophagitis severity was not related to any of these factors. On the other hand, local distribution of acid rather than total gastric secretion may be more relevant to the pathogenesis of GERD. Data suggest that the gastroesophageal junction may escape the buffering effect of meals, remaining highly acidic (median pH 1.6) compared with the body of the stomach (pH 4.7). This “proximal pocket of acid” extends from the cardia into the distal esophagus and could account for the high prevalence of disease in this location.97 A recent study confirmed these findings, demonstrating a pH less than 4 for 26.5% of the time over 24 hours when measured at the lower esophageal sphincter.98 H. pylori infection, especially with the cagA+ virulent strain, is a “biological antisecretory agent” that lowers gastric acidity, thereby possibly protecting from the development of severe esophagitis and Barrett’s esophagus.99,100 Acid output may be decreased by several mechanisms: (1) the associated severe corpus gastritis, which, over time, progresses to multifocal atrophic gastritis (see Chapter 51); (2) increased gastric alkaline (bicarbonate) secretion, which returns to normal after H. pylori eradication101; and perhaps (3) production of ammonia by the bacteria itself.102 After eradication of H. pylori, the corpus mucosa can regenerate to normal, increasing acid secretion and potentiating reflux

in susceptible patients, possibly contributing to the reports of esophagitis after eradication of this organism.103,104 The consequence of returning the stomach to a healthy state is unknown but may be an underlying factor in the increasing prevalence of severe GERD, Barrett’s esophagus, and even adenocarcinoma in Western populations.13,102

Duodenogastric Reflux

Along with acid and pepsin, duodenal contents may be injurious to the esophageal mucosa. Animal studies demonstrate that conjugated bile acids produce their greatest injury in the presence of acid and pepsin, whereas trypsin and the deconjugated bile acids are damaging in a more neutral environment.105 These experiments suggest that duodenogastric reflux into the esophagus predisposes to complications of GERD.106,107 However, the accurate measurement of duodenogastric reflux is difficult. Traditionally, this phenomenon was defined indirectly by measuring the esophageal pH greater than 7 (i.e., “alkaline reflux”).107 However, on the basis of newer technology that accurately measures bilirubin, the most common pigment of bile, independent of pH, we now know that this technique is inaccurate.108 These studies show that acid and bile reflux increase in parallel across the spectrum of GERD, suggesting a synergistic role in the development of esophagitis and its complications.109,110 Additionally, aggressive acid suppression with proton pump inhibitors decreases both acid and duodenogastric reflux by decreasing the volume of gastric contents available to reflux into the esophagus.110

Delayed Gastric Emptying

The importance of delayed gastric emptying in the pathogenesis of GERD is controversial. Early studies found delay in the gastric emptying of solids in up to 50% of reflux patients.111 However, methodologic problems may have invalidated these studies. More recent investigations found only a 6% to 38% incidence of delayed gastric emptying, regardless of the severity of the esophagitis.112,113 Nevertheless, delayed gastric emptying is a major factor contributing GERD in some groups such as diabetic patients with autonomic peripheral neuropathy.

CLINICAL FEATURES CLASSIC REFLUX SYMPTOMS

Heartburn is the classic symptom of GERD, with patients generally reporting a burning feeling, rising from the stomach or lower chest and radiating toward the neck, throat, and occasionally the back.114 It occurs postprandially, particularly after large meals or after ingesting spicy foods, citrus products, fats, chocolates, and alcohol. The supine position and bending over may exacerbate heartburn. Recent studies have suggested that sleep deprivation as well as psychological or auditory stress may lower the threshold for symptom perception.115-117 Nighttime heartburn may cause sleeping difficulties and impair next-day function.118 When heartburn dominates a patient’s complaints, it has high specificity (89%) but low sensitivity (38%) for GERD as diagnosed by 24-hour esophageal pH testing.119 GERD is usually diagnosed symptomatically by the occurrence of heartburn two or more days a week, although less frequent symptoms do not preclude the disease.120,121 Although an aid to diagnosis, the frequency and severity of heartburn do not predict the degree of esophageal damage.5

713

714

Section V  Esophagus Heartburn symptoms can arise from acid reflux, weakly acidic reflux, bile reflux, and mechanical stimulation of the esophagus. The receptor that mediates the sensation of heartburn during acid perfusion has not been identified, although the capsaicin or vanilloid receptor 1 (TRPV1) is a leading candidate. TRPV1 is a cation channel that is expressed by sensory neurons, and its activation by heat, acid pH, or ethanol may trigger burning pain.122-126 Weakly acidic reflux, as detected by combined pH and impedance technology, appears to produce symptoms when there is a large proximal extent reached by the refluxate, large reflux volumes, and prolonged acid-clearance times.127 The mechanism that underlies bile-acid–induced esophageal symptoms is unknown. Bile acids are postulated to induce the release of intracellular mediators via damage to lipid membranes.128 In addition to acid-induced and bile-acid–induced esophageal damage, pepsin can cause direct damage to the esophageal mucosa, leading to dilated intercellular spaces and increased esophageal mucosal permeability.129 Esophageal distention and sustained esophageal contractions are other mechanisms proposed to explain the symptom of heartburn. Balaban and associates used high-frequency endoluminal ultrasound to demonstrate a correlation between spontaneous chest pain or chest pain induced by edrophonium chloride and sustained esophageal longitu­ dinal muscle contractions.130 Esophageal hyperalgesia (lowered pain threshold) and psychological comorbidity have also been postulated to contribute to heartburn symptoms.131 Other common symptoms of GERD are acid regurgitation and dysphagia. The effortless regurgitation of acidic fluid, especially after meals and worsened by stooping or the supine position, is highly suggestive of GERD.119 Among patients with daily regurgitation, LES pressure is usually low; many have associated gastroparesis, and esophagitis is common, making this symptom more difficult to treat medically than classic heartburn. Dysphagia is reported by more than 30% of individuals with GERD.132 It usually occurs in the setting of long-standing heartburn with slowly progressive dysphagia for solids. Weight loss is uncommon, because patients have good appetites. The most common causes are a peptic stricture or Schatzki’s ring, but other etiologies include severe esophageal inflammation alone, peristaltic dysfunction, and esophageal cancer arising from Barrett’s esophagus (see Chapter 44). Less common symptoms associated with GERD include water brash, odynophagia, burping, hiccups, nausea, and vomiting.133 Water brash is the sudden appearance in the mouth of a slightly sour or salty fluid. It is not regurgitated fluid, but rather secretions from the salivary glands in response to acid reflux.73 Odynophagia may be seen with severe ulcerative esophagitis. However, its presence should raise the suspicion of an alternative cause of esophagitis, especially infections or injury from impacted pills (see Chapter 45). Some patients with GERD are asymptomatic. This is particularly true in the older adults, perhaps because of decreased acidity of the reflux material in some or decreased pain perception in others.5 Many older adult patients present first with complications of GERD because of longstanding disease with minimal symptoms. For example, up to one third of patients with Barrett’s esophagus are insensitive to acid at the time of presentation.134

EXTRAESOPHAGEAL MANIFESTATIONS

GER may cause a wide spectrum of conditions including noncardiac chest pain, asthma, posterior laryngitis, chronic cough, recurrent pneumonitis, and even dental erosion.135

Some of these patients have classic reflux symptoms, but many are “silent refluxers,” contributing to problems in making the diagnosis. Furthermore, it may be difficult to establish a causal relationship even if GER can be documented by testing (e.g., pH studies) because individuals may simply have two common diseases without a causeand-effect relationship.

Chest Pain

GER-related chest pain may mimic angina pectoris, having a squeezing or burning quality; being in a substernal location; and radiating to the back, neck, jaws, or arm. It frequently is worse after meals, can awaken the patient from sleep, and may worsen during emotional stress. Heavy exercise, even treadmill testing, may provoke GER.136 Refluxrelated chest pain may last for minutes to hours, often resolves spontaneously, and may be eased with antacids. The majority of patients with GERD-induced chest pain have heartburn symptoms.137 Multiple studies since the mid-1990s identify GER, rather than spastic motility disorders, as the most common esophageal cause of noncardiac chest pain.138 The mechanism for GERD-related chest pain is poorly understood and is probably multifactorial, related to H+ ion concentration, volume, and duration of acid reflux; secondary esophageal spasm; and prolonged contractions of the longitudinal muscles.139

Asthma and Other Pulmonary Disorders

The prevalence of GERD in asthmatics is estimated between 34% and 89%, depending on the group of patients studied and how GERD is defined (e.g., symptoms or 24-hour pH monitoring).140 Symptomatic GERD is an important comorbid condition in asthma patients, being associated with greater asthma severity.141 GERD should be considered in asthmatics who present in adulthood, those without an extrinsic (allergic) component, and those not responding to bronchodilators or glucocorticoids.142 Up to 30% of patients with GERD-related asthma have no esophageal complaints. Other pulmonary diseases associated with GERD include aspiration pneumonia, interstitial pulmonary fibrosis, chronic bronchitis, and bronchiectasis. In addition, preliminary data suggest that GER may worsen the course of obstructive sleep apnea in a subset of patients.143 Proposed mechanisms of reflux-induced asthma include aspiration of gastric contents into the lungs with secondary bronchospasm and activation of a vagal reflex from the esophagus to the lungs causing bronchoconstriction. Animal144 and human145 studies report bronchoconstriction after esophageal acidification, but the response is mild and inconsistent. On the other hand, intratracheal infusion of even small amounts of acid induces profound and reproducible bronchospasm in cats.144 The reflux of acid into the trachea as compared with the esophagus alone predictably caused marked changes in peak expiratory flow rates in asthmatic patients.146 Although both mechanisms may trigger reflux-induced asthma, patients with severe asthma probably suffer from intermittent microaspiration.

Ear, Nose, and Throat Diseases

GERD may be associated with a variety of laryngeal symptoms and signs, of which “reflux laryngitis” is the most common.147,148 These patients present with hoarseness, globus sensation, frequent throat clearing, recurrent sore throat, and prolonged voice warm-up. Ear, nose, and throat signs attributed to GERD include posterior laryngitis with edema and redness (Fig. 43-9), vocal cord ulcers and granulomas, leukoplakia, and even carcinoma. These changes are

Chapter 43  Gastroesophageal Reflux Disease ASSOCIATED CONDITIONS

Figure 43-9.  Characteristic laryngeal findings of “reflux laryngitis” in a 31-year-old man with hoarseness whose symptoms and signs resolved after proton pump inhibitor treatment for three months. Black arrows: bilateral erythema of medial arytenoid walls. White arrows: red streaks on the true vocal folds. Reflux changes in the larynx are usually confined to the posterior portion nearest the upper esophageal sphincter (bluish gray slit behind the arytenoid complex).

usually limited to the posterior third of the vocal cords and interarytenoid areas, both in proximity to the upper esophageal sphincter. Animal studies find that the combination of acid, pepsin, and conjugated bile acids is very injurious to the larynx.149 Human studies report that proximal esophageal acid exposure, especially while sleeping, is significantly increased in patients with laryngeal symptoms and signs.150 GERD has been postulated to be a leading cause of chronic cough (after sinus problems and asthma).151 GER increases the cough sensitivity reflex (i.e., reduces the cough threshold) in patients with chronic cough.152 However, the im­portance of this association in humans has not been shown in treatment studies, which, in sum, have not demonstrated a superiority of proton pump inhibitors over placebo.153 Dental erosion, the loss of tooth structure by nonbacterial chemical processes, can be caused by GER in healthy subjects and patients with bulimia.154 Microaspiration of gastric contents is the most likely etiology of these complaints.

DIFFERENTIAL DIAGNOSIS Symptoms associated with GERD may be mimicked by other esophageal and extraesophageal diseases including achalasia, Zenker’s diverticulum, gastroparesis, gallstones, peptic ulcer disease, functional dyspepsia, and angina pectoris. These disorders usually can be identified by failure to respond to aggressive proton pump inhibitor (PPI) therapy and appropriate diagnostic tests. Although GERD is the most common cause of esophagitis, other etiologies (pills, infections, or radiation esophagitis) need to be considered in difficult-to-manage cases, older individuals, or immunocompromised patients.

Several medical and surgical conditions discussed elsewhere in this book can predispose to GERD. The most common is pregnancy, in which 30% to 80% of women complain of heartburn, especially in the first trimester (see Chapter 38). Pregnancy increases the risk for reflux by reducing LES pressure due to the effects of estrogen and progesterone and possibly mechanical factors from the gravid uterus.155 Although symptoms may be severe, esophagitis is uncommon and this type of situational GERD is cured with childbirth. Up to 90% of patients with scleroderma have GERD due to smooth muscle fibrosis causing low LES pressure and weak or absent peristalsis (see Chapter 35). Severe disease is common, with up to 70% of patients having esophagitis. Many patients have peptic strictures, and Barrett’s esophagus and carcinoma of the esophagus have been reported.156 Acid hypersecretion and increased gastric volume are the major factors causing GERD in patients with the Zollinger-Ellison syndrome (see Chapter 32). In these patients, the esophagitis and complications are more difficult to treat than the ulcer disease.157 After Heller myotomy for achalasia, 10% to 20% of patients may develop GERD158 (see Chapter 42). In patients undergoing laparoscopic gastric banding for morbid obesity, de novo symptoms of GERD may develop postoperatively (see Chapter 7).30 Finally, prolonged nasogastric intubation may cause reflux esophagitis, in part because acid tracks orad along the tube and because the tube mechanically interferes with the LES barrier function.159

DIAGNOSIS A large number of tests are available for evaluating patients with suspected GERD. Many times these tests are unnecessary because the classic symptoms of heartburn and acid regurgitation are sufficiently specific to identify reflux disease and begin medical treatment. However, this is not always the case, and clinicians must decide which tests to choose so as to make a diagnosis in a reliable, timely, and cost-effective manner depending on the information desired (Table 43-2).160

EMPIRICAL TRIAL OF ACID SUPPRESSION

An empirical trial of acid suppression is the simplest and most definitive method for diagnosing GERD and assessing its relationship to symptoms. Unlike other tests that only suggest an association (e.g., esophagitis at endoscopy or positive symptom index on pH testing), the response to antireflux therapy ensures a cause-and-effect relationship between GERD and symptoms. With the advent of PPIs, it has become the first test used in patients with classic or atypical reflux symptoms without alarm complaints. Symptoms usually respond to a PPI trial in one to two weeks. If symptoms disappear with therapy and then return when the medication is discontinued, GERD has been established. In empirical trials for heartburn, the initial PPI dose was high (e.g., omeprazole 40 to 80  mg/day), usually given for at least two weeks, and a positive response was defined as at least 50% improvement in heartburn. Using this approach, the PPI empirical trial has a sensitivity of 68% to 83% for determining the presence of GERD.161,162 In noncardiac chest pain, Fass and colleagues163 found that a seven-day trial of omeprazole 40 mg in the morning and 20 mg at night had

715

716

Section V  Esophagus Table 43-2  Diagnostic Tests for Gastroesophageal Reflux Disease Tests for Reflux Intraesophageal pH monitoring (catheter or catheter-free system) Ambulatory impedance and pH monitoring (nonacid reflux) Barium esophagogram Tests to Assess Symptoms Empirical trial of acid suppression Intraesophageal pH monitoring with symptom analysis Tests to Assess Esophageal Damage Endoscopy Wireless capsule endoscopy Esophageal biopsy Barium esophagogram Tests to Assess Esophageal Function Esophageal manometry Esophageal impedance

a sensitivity of 78% and specificity of 86% for predicting GERD when compared with traditional tests. Likewise, Ours and colleagues164 found omeprazole 40  mg twice daily for two weeks to be a very reliable method for identifying acidrelated cough. An empirical PPI trial for diagnosing GERD offers many advantages: the test is office based, easily done, relatively inexpensive (especially with use of over-the-counter [OTC] PPIs), available to all physicians, and avoids many needless procedures. For example, Fass and colleagues163 showed a savings of greater than $570 per patient due to a 59% reduction in the number of diagnostic tests performed for noncardiac chest pain. Disadvantages are few, including a placebo response and uncertain symptomatic endpoint if symptoms do not totally resolve with extended treatment.

ENDOSCOPY

Upper endoscopy is the standard for documenting the presence and extent of esophagitis and excluding other etiologies for the patient’s symptoms. However, only 20% to 60% of patients with abnormal esophageal reflux by pH testing have esophagitis at endoscopy. Thus, the sensitivity of endoscopy for GERD is poor, but it has excellent specificity at 90% to 95%.165 The earliest endoscopic signs of acid reflux include edema and erythema, but these findings are nonspecific and dependent on the quality of endoscopic visual images.166 More reliable signs are friability, granularity, and red streaks. Friability (easy bleeding) results from the development of enlarged capillaries near the mucosal surface in response to acid. Red streaks extend upward from the esophageal junction along the ridges of the esophageal folds.167 Erosions develop with progressive acid injury, characterized by a shallow break in the mucosa with a white or yellow exudate surrounded by erythema. Typically, erosions begin at the gastroesophageal junction, occurring along the tops of esophageal mucosal folds, where acid injury is most prone, and they may be single or multiple. Erosions can also be caused by nonsteroidal anti-inflammatory drugs, heavy smoking, and infectious esophagitis.165 Ulcers reflect more severe esophageal damage, being deeper into the mucosa or submucosa and either isolated along a fold or surrounding the esophageal junction. Multiple classification systems for esophagitis have been proposed; some are confusing and none has worldwide acceptance. In Europe the most popular scheme is the Savary-Miller classification.168 The most

Table 43-3  Endoscopic Grading Systems for Esophagitis Savary-Miller Classification Grade 0 Not applicable Grade I Single, erosive, or exudative lesion on one longitudinal fold Grade II Multiple erosions on more than one longitudinal fold Grade III Circumferential erosions Grade IV Ulcer, stricture, or short esophagus, isolated or associated with grades I through III Grade V Barrett’s esophagus ± grades I through III Los Angeles Classification Grade A One or more mucosal breaks confined to folds, ≤5 mm Grade B One or more mucosal breaks >5 mm confined to folds but not continuous between tops of mucosal folds Grade C Mucosal breaks continuous between tops of two or more mucosal folds but not circumferential Grade D Circumferential mucosal break

thoroughly evaluated esophagitis classification is the Los Angeles (LA) system, which is gaining acceptance in the United States and Europe (Table 43-3) (Fig. 43-10A to D).169 Esophageal capsule endoscopy for the evaluation of reflux symptoms has thus far been disappointing. The capsule is 11 by 26  mm and acquires video images at 14 frames per second. After swallowing images are transmitted to a portable receiver via digital radiofrequency. In one study, compared with standard upper endoscopy, the capsule has a sensitivity of 50% for erosive esophagitis, 54% for the presence of a hiatal hernia, and 79% for the presence of Barrett’s esophagus.170 As mentioned, most patients with GERD are treated initially with PPIs and without endoscopy. The important exception is the patient experiencing “alarm” symptoms: dysphagia, odynophagia, weight loss, and gastrointestinal bleeding. Here endoscopy should be performed early to diagnose complications of GERD (e.g., strictures) and to rule out other entities such as infections, ulcers, cancer, or varices. Current guidelines suggest the major role of endoscopy is to diagnose and treat GERD complications, especially peptic strictures, and to define Barrett’s esophagus.171 Using this rationale, the majority of patients with chronic GERD need only one endoscopy while on therapy.

ESOPHAGEAL BIOPSY

Like endoscopy, the role of esophageal biopsies in evaluating GERD has evolved over the years. Microscopic changes of reflux may occur even when the mucosa endoscopically appears normal.172 These classic changes of basal cell hyperplasia and increased height of the rete peg, both representing increased epithelial turnover of the squamous mucosa, are sensitive but not specific histologic findings for GERD.173 Acute inflammation characterized by the presence of neutrophils and often eosinophils (Fig. 43-11) is very specific for esophagitis; however, the sensitivity is low, in the range of 15% to 40%.174 Thus, there is little value for histologic examination of normal-appearing squamous mucosa to either confirm or exclude pathologic acid reflux.175 In patients with classic esophagitis, biopsies are usually not taken unless necessary to exclude neoplasm, infection, pill injury, or bullous skin disease. Therefore, the current primary indication for esophageal biopsies is to determine the presence of Barrett’s epithelium.171 When this diagnosis is suspected, biopsies are mandatory and best done when esophagitis is healed (see Chapter 44).

Chapter 43  Gastroesophageal Reflux Disease

A

B

C

D

Figure 43-10.  Endoscopic photographs of the four grades of esophagitis (A to D) using the Los Angeles classification system as outlined in Table 42-3.

ESOPHAGEAL pH MONITORING

Figure 43-11.  Histopathology of gastroesophageal reflux disease. Inflammatory cells (eosinophils and neutrophils) are interspersed between squamous epithelial cells. (Courtesy Edward Lee, MD, Washington, D.C.)

Ambulatory intraesophageal pH monitoring is the standard test for establishing pathologic reflux.176,177 For catheterbased pH testing, the probe is passed nasally, positioned 5 cm above the manometrically determined LES, and connected to a battery-powered data logger capable of collecting pH values every four to six seconds. An event marker is activated by the patient when symptoms, meals, and body position changes occur. Patients are encouraged to eat normally and engage in regular daily activities, with monitoring carried out for 18 to 24 hours. Reflux episodes are defined by a pH drop of less than 4. Conventionally measured parameters include percent of total time when pH is less than 4; percent of time, upright and supine when pH is less than 4; total number of reflux episodes; duration of longest reflux episode; and number of episodes greater than five minutes.176 The percent of total time pH is less than 4 is the most reproducible measurement for GERD, with reported upper limits of normal ranging from 4% to 5.5%.176 Ambulatory pH testing discerns positional variations in

717

Section V  Esophagus Physiologic reflux pattern M

M S

M

S M

8

pH

6 4 2

04:00 PM

10:00 AM

10:00 PM

04:00 AM

10:00 AM

A Upright reflux pattern M

S

M

SM

8

pH

6 4 2

11:00 AM

05:00 PM

11:00 PM

05:00 AM

11:00 AM

B Combined reflux pattern M

S

M S

8 6 pH

718

4 2

11:00 AM

05:00 PM

11:00 PM

05:00 AM

11:00 AM

C Figure 43-12.  Common patterns of 24-hour esophageal pH monitoring. A, The physiologic pattern of gastroesophageal reflux (GER) seen in healthy subjects. Reflux is noted after meals (M) but not during sleep (S). A reflux episode is defined as a pH drop to less than 4. B, Upright reflux pattern with extensive GER during the day but not at night. These patients have frequent symptoms, but esophagitis is uncommon. C, Combined pattern with GER during the day and at night. Most of these patients have esophagitis.

GER, meals, and sleep-related episodes and helps relate symptoms to reflux events (Fig. 43-12). A critical limitation of esophageal pH monitoring is that there exists no absolute threshold value that reliably identifies GERD patients. Studies comparing patients with endoscopic esophagitis who underwent pH tests report sensitivities from 77% to 100% with specificities from 85% to 100%.178 However, these patients with esophagitis rarely need pH testing; rather, patients with normal endoscopy and suspected GERD might benefit most from this test. Unfortunately, data on these patients are less conclusive, with considerable overlap between controls and nonerosive refluxers.177 Other drawbacks of pH testing include possible equipment failure, pH probe missing reflux events because the probe is buried in a mucosal fold, and false-negative studies due to dietary or activity limitations from poor tolerability of the nasal probe.178 A major advance in esophageal pH testing has been the development of a catheter-free system.179 This system uses

a wireless pH capsule that is affixed to the esophageal mucosa with a delivery system that drives a small needle into the epithelium. The capsule then transmits pH data to a portable receiver using radiofrequency signals. Catheterfree testing is rapidly becoming the preferred method of pH testing because monitoring can be extended beyond 24 hours and limitations on normal daily activities and meals are negligible.179 Ambulatory esophageal pH monitoring is the only test that records and correlates symptoms with reflux episodes over extended periods of time. However, because only 10% to 20% of reflux episodes are associated with symptoms, different statistical analyses have evolved, attempting to define a significant association between symptoms and reflux episodes, including the symptom index, symptom sensitivity index, and symptom association probability.146 Unfortunately, no studies have defined the accuracy of these symptom scores in predicting response to therapy. Therefore, pH testing can define an association between

Chapter 43  Gastroesophageal Reflux Disease complaints and GER, but only treatment trials address the critical clinical issue of causality. Clinical indications for ambulatory pH monitoring are established.171 Before fundoplication, pH testing should be done in patients with normal endoscopy to ensure the presence of pathologic acid reflux. After antireflux surgery, persistent or recurrent symptoms warrant repeat pH testing. In these situations, pH monitoring is performed with the patient off antireflux medications. Esophageal pH testing is particularly helpful in evaluating patients with reflux symptoms who are resistant to treatment and who have normal or equivocal endoscopic findings. For this indication, pH testing is usually done on PPI therapy to define two populations: those with and those without continued abnormal acid exposure times. The group with persistent GER needs intensified medical therapy, whereas patients with symptoms and good acid control have another etiology for their complaints. Finally, ambulatory pH testing may help in defining patients with extraesophageal manifestations of GERD. In this situation, pH testing is often done with additional pH probes in the proximal esophagus or pharynx.147 Initially most of these studies were done off antireflux medications to confirm the coexistence of GERD; however, this does not guarantee symptom causality. Therefore, one approach is to first treat aggressively with PPIs, reserving pH testing for those patients not responding after 4 to 12 weeks of therapy.164 A relatively new method of evaluating GERD has been combined impedance and acid testing, which allows the measurement of acid and nonacid (volume) reflux. Nonacid reflux is measured by the detection of a retrograde bolus of ion-rich fluid in the esophagus. Refluxates that are a mixture of liquid and air are also readily detected. In a large group of normal subjects, roughly 40% of reflux episodes were either weakly acidic (pH 4 to 6.5) or alkaline (nadir esophageal pH during episode >6.5).180 In a multicenter study using impedance, 37% of patients experienced continued reflux symptoms despite twice-daily PPI therapy due to nonacid reflux.181 These patients would have been interpreted as negative for reflux had they been studied using conventional pH only. Another study using 24-hour ambulatory pH impedance found a temporal relationship between symptoms and nonacid reflux in 4.1% and 16.7% of subjects off and on PPI therapy, respectively.182 Regurgitation and cough were the most prevalent symptoms associated with nonacid reflux.

BARIUM ESOPHAGOGRAM

The barium esophagogram is an inexpensive, readily available, and noninvasive esophageal test. It is most useful in demonstrating anatomic narrowing of the esophagus and assessing the presence and reducibility of a hiatal hernia. Schatzki’s rings, webs, or minimally narrowed peptic strictures may only be seen with an esophagogram, being missed by endoscopy, which may not adequately distend the esophagus. Giving a 13-mm radiopaque pill or marshmallow along with the barium liquid can help to identify these subtle narrowings.183 The barium esophagogram allows good assessment of peristalsis and is helpful preoperatively in identifying a weak esophageal pump.184 The ability of barium esophagogram to detect esophagitis varies, with sensitivities of 79% to 100% for moderate to severe esophagitis, whereas mild esophagitis is usually missed.185,186 Barium testing also falls short when addressing the presence of Barrett’s esophagus. The spontaneous reflux of barium into the proximal esophagus is very specific for reflux, but it is not sensitive. Provocative maneuvers (e.g., leg lifting, coughing, Valsalva, or water siphon) can elicit stress reflux

and improve the sensitivity of the barium esophagogram, but some argue that these maneuvers also decrease its specificity.185,186

ESOPHAGEAL MANOMETRY

Esophageal manometry allows assessment of LES pressure and relaxation, as well as peristaltic activity, including contraction amplitude, duration, and velocity. However, esophageal manometry is generally not indicated in the evaluation of the uncomplicated GERD patient because most have a normal resting LES pressure.51 (It is an integral component of pH testing to accurately define LES location; see earlier.) Esophageal manometry to document adequate esophageal peristalsis is traditionally recommended before antireflux surgery.187 If the study identifies ineffective peristalsis (low amplitude or frequent failed peristalsis),188 then a complete fundoplication may be contraindicated. However, this assumption has recently been challenged by several studies finding that reflux control was better and dysphagia no more common in patients with weak peristalsis after a complete, as opposed to a partial, fundoplication.189 An improvement of traditional manometry, combining it with impedance testing, is helping to clarify this controversy. Using this technique, a study found that less than 50% of patients with ineffective peristalsis had a significant delay in esophageal bolus transit measured by impedance.190 Therefore, potentially only these patients with a significant physiologic defect in motility will require a modified fundoplication.

CLINICAL COURSE The clinical course of GERD depends to a great extent on whether the patient has erosive or nonerosive disease. There is controversy as to whether GERD exists as a spectrum of disease severity or as a categorical disease in three distinct groups, including Barrett’s esophagus. Patients tend not to cross over from one group to another; in follow-ups ranging from six months to longer than 22 years, less than 25% of patients with nonerosive disease evolved over time to having esophagitis, nearly all to LA grade A/B disease, or to having complications of GERD.191-194

NONEROSIVE REFLUX DISEASE

Early studies from tertiary referral centers suggested that the majority of GERD patients had esophagitis.195 However, studies carried out in community practices reveal that up to 70% of GERD patients had a normal endoscopic examination.196-198 Endoscopy-negative GERD patients are more likely to be female, younger, thin, and without hiatal hernia, and they have a higher prevalence of functional GI disorders.199 Despite their mild mucosal damage, these patients demonstrate a chronic pattern of symptoms with periods of exacerbation and remission.200 Nonerosive GERD is suspected in the patient with typical reflux symptoms and a normal endoscopy and confirmed by the patient’s response to antisecretory therapy. Esophageal pH testing identifies three distinct subsets of nonerosive GERD patients. First are the patients with excessive acid reflux who usually respond to PPI therapy. Second are the patients with normal reflux parameters but a good correlation between symptoms and acid reflux episodes. This group represents 30% to 50% of nonerosive GERD patients and has “functional heartburn.”200 These patients probably have heightened esophageal sen­ sitivity to acid and are less likely to respond to antireflux therapy.201 The third group is characterized by normal acid exposure times and poor symptom correlation. Whether

719

720

Section V  Esophagus they truly represent a subset of nonerosive GERD is questionable.

EROSIVE REFLUX DISEASE

Patients with erosive esophagitis tend to be male, older, and overweight and are more likely to have hiatal hernias.199 The clinical course of these patients with erosive esophagitis is more predictable and associated with complications of GERD. Longitudinal studies have shown that up to 85% of patients with erosive GERD, on no maintenance reflux therapy, will relapse within six months of stopping PPI therapy, and the relapse rate is highest in patients with more severe grades of esophagitis (see Table 43-3).202,203 Several studies confirm that erosive esophagitis patients are prone to reflux complications, including ulcers, strictures, and Barrett’s esophagus. In a Finnish study, 20 patients with erosive GERD treated with lifestyle changes, antacids, and prokinetic drugs were followed for a median of 19 years. Fourteen patients continued to have erosions, and six new cases of Barrett’s esophagus were detected.192 In another more recent European study193 over two years, patients with LA grade C/D esophagitis developed Barrett’s esophagus at a rate of 5.8% compared with only 1.4% for LA grade A/B and 0.5% for nonerosive GERD. However, these data must be contrasted with a two-year U.S. trial in which no patient with erosive esophagitis developed Barrett’s esophagus204 and in another study in which stricture was reported in only 0.9% of 957 patients over 7.6 years of symptom driven antireflux treatment.194

COMPLICATIONS HEMORRHAGE, ULCERS, AND PERFORATION

GERD-related non-cancer deaths are rare (0.46 per 100,000 persons). The most common fatal causes are hemorrhagic esophagitis, aspiration pneumonia, ulcer perforation, and rupture with severe esophagitis.205 Major hemorrhage and esophageal perforation are usually associated with deep esophageal ulcers or severe esophagitis.206 Esophageal per-

Figure 43-13.  Classic peptic stricture demonstrated by barium esophagogram (A) and endoscopy (B). The film shows a large hiatal hernia (HH) common to all GERD strictures. Dark arrow points to short thick fibrous stricture with mul­tiple pseudodiverticula (white arrows). Although not seen on barium examination, the endoscopic view also demonstrates circumferential esophagitis (Los Angeles grade D). GERD, gastroesophageal reflux disease.

forations are very rare in the PPI era but can result in mediastinitis and death. Clinically important hemorrhage has been reported in 7% to 18% of GERD patients207 and may result in iron deficiency anemia.

PEPTIC ESOPHAGEAL STRICTURES

Strictures occur in 7% to 23% of patients with untreated reflux esophagitis, and are especially seen in older men.208 They may be linked to chronic nonsteroidal anti-inflammatory drug (NSAID) use.209 Stricture formation is complex, starting as reversible inflammation with edema, cellular infiltration, and vascular congestion, progressing to collagen deposition and ending in irreversible fibrosis. As dysphagia progresses, heartburn often decreases, reflecting the stricture acting as a barrier to further reflux. Dysphagia is usually limited to solids. Unlike malignant strictures, patients with peptic strictures have a good appetite, alter their diet, and lose little weight. Peptic strictures are smooth-walled, tapered, circumferential narrowings in the lower esophagus, usually less than 1 cm long but occasionally extend to 8 cm (Fig. 43-13). In these unusual cases, the clinician should suspect a predisposing condition, such as Zollinger-Ellison syndrome, or another condition such as pill esophagitis or a stricture from prolonged nasogastric intubation.208 A mid- to upper esophageal stricture should raise concern for Barrett’s esophagus or malignancy. Although once controversial, today a Schatzki’s ring is considered a forme fruste of an early peptic stricture.210 All stricture patients should undergo endoscopy, at least initially, to confirm the benign nature of the lesion and, if necessary, take biopsies to exclude cancer and Barrett’s esophagus.

BARRETT’S ESOPHAGUS (see Chapter 44) TREATMENT OF UNCOMPLICATED DISEASE The rationale for GERD therapy depends on a careful definition of specific aims. In patients without esophagitis, the therapeutic goals are to relieve reflux symptoms and prevent

HH

B

A

Chapter 43  Gastroesophageal Reflux Disease frequent symptomatic relapses. In patients with esophagitis, the goals are to relieve symptoms and heal esophagitis while preventing further relapses and complications.

NONPRESCRIPTION THERAPIES

Although GERD is common, many sufferers do not seek medical care, instead choosing to change their lifestyles and self-medicate with OTC antacids, histamine-2 receptor antagonists (H2RAs), and even PPIs. These observations have led to the “iceberg” model of the GERD population. The vast majority of heartburn suffers are invisible because they self-medicate and do not seek professional help; only those at the tip of the iceberg, typically patients with severe symptoms or reflux complications, are seen by physicians.211

Lifestyle Modifications

Selective lifestyle changes, carefully explained to the patient, should be part of the initial management plan and are especially helpful in those with mild, intermittent complaints. These include elevating the head of the bed, avoiding tight-fitting clothes, losing weight if overweight, restricting alcohol and smoking, making dietary changes, refraining from lying down after meals, and avoiding bedtime snacks. Physiologic studies show that these maneuvers enhance esophageal acid clearance, decrease acid reflux–related events, or ease heartburn symptoms.212 Headof-the-bed elevation can be done by using 6- to 8-inch blocks or a foam wedge under the mattress to elevate the upper torso. Eating several hours before retiring and avoiding bedtime snacks keeps the stomach empty at night, thereby decreasing nocturnal reflux episodes. Avoiding tight-fitting clothes and losing weight are interventions aimed at reducing the incidence of reflux by the “abdominal stress” mechanism. Targeted weight loss may be helpful, whereas discrete periods of weight gain can be associated with exacerbation of reflux symptoms.30 Cessation of smoking and alcohol reduction is valuable because both agents lower LES pressure, reduce acid clearance, and impair intrinsic squamous epithelial protective functions.76,211 Reducing meal size and avoiding fats, carminatives, and chocolate reduces reflux frequency by decreasing episodes of tLESRs, as well lowering LES pressure.211 Additionally, some patients complain of heartburn after citrus drinks, spicy foods, tomato-based products, coffee, tea, or cola drinks. Stimulation of gastric acid secretion or esophageal sensitivity to low pH (or perhaps hyperosmolar solutions) may account for these symptoms.213 However, indiscriminate food prohibition should be avoided but rather tailored to individual sensitivity to better promote compliance. Finally, patients should avoid, if possible, drugs that lower LES pressure (see Table 43-1) or promote localized esophagitis, such as certain bisphosphonates (see Chapter 45). How good are the clinical studies assessing the efficacy of these commonly prescribed lifestyle changes? In an evidence-based review,211 studies of smoking, alcohol, chocolate, fatty foods, and citrus products had sound phy­ siologic data that their intake can adversely effect symptoms or promote reflux on esophageal pH tests. However, there was little convincing evidence that cessation of these pro­ ducts predictably improved reflux symptoms. Only elevation of the head of the bed, left lateral decubitus positioning, and weight loss were associated with GERD improvement in case-controlled studies.211

Over-the-Counter Medications

These drugs are used in treating mild, infrequent heartburn symptoms triggered by lifestyle indiscretions. Antacids

increase LES pressure but work primarily by buffering gastric acid, albeit for short periods. Heartburn symptoms are rapidly relieved, but patients need to take antacids frequently, usually 1 to 3 hours after meals. Gaviscon, containing alginic acid and antacids, mixes with saliva to form a highly viscous solution that floats on the gastric pool, acting as a mechanical barrier. Both antacids214 and Gaviscon215 are more effective than placebo in relieving symptoms induced by a heartburn-promoting meal. However, they do not heal esophagitis, and long-term trials suggest symptom relief in only 20% of patients.216,217 OTC H2RAs are available at doses usually one half the standard prescription dose. Although onset of relief is not as rapid as antacids, the OTC H2RAs relieve symptoms for 6 to 10 hours.218 Therefore, they are particularly useful when taken before potentially refluxogenic activities. Like antacids, OTC H2RAs are ineffective in healing esophagitis.218 OTC combinations of antacids and H2RAs are available. The long-term safety and efficacy of PPIs led the U.S. Food and Drug Administration (FDA) to approve omeprazole at full dose (20 mg) for OTC use in 2003. Drug labeling suggested daily use for only two weeks and recommended physician follow-up for persistent symptoms. Despite initial “real world” concerns of abusing this drug, early actual-use data support that consumers accurately self-select if OTC omeprazole is appropriate for use, comply with a two-week regimen, and seek physician care for longer-term management of frequent heartburn.219

PRESCRIPTION MEDICATIONS

Patients with frequent heartburn, esophagitis, or complications usually see a physician and receive prescription medications. Prokinetic drugs attempt to correct the GERD-related motility disorders associated with GERD. However, the most clinically effective drugs for short- and long-term reflux treatment are acid suppressive drugs.

Prokinetic Drugs

Until recently three prokinetic drugs were available for treating GERD: bethanechol, a cholinergic agonist; metoclopramide, a dopamine antagonist; and cisapride, a serotonin (5-HT4) receptor agonist that increases acetylcholine release in the myenteric plexus. These drugs improve reflux symptoms by increasing LES pressure, acid clearance, or gastric emptying. However, none alters tLESRs, and their effectiveness decreases with disease severity.220 Current prokinetics provide modest benefit in controlling heartburn but have unreliable efficacy in healing esophagitis unless combined with acid-inhibiting drugs.220 Prokinetic drugs are limited by their side effect profiles. Bethanechol commonly causes flushing, blurred vision, headaches, abdominal cramps, and urinary frequency. Metoclopramide, which crosses the blood-brain barrier, has a 20% to 50% incidence of fatigue, lethargy, anxiety, and restlessness and rarely causes tremor, parkinsonism, dystonia, or tardive dyskinesia, especially in older patients. Side effects may be decreased by reducing the dosing regimen to twice a day, taking a larger single dose before dinner or at bedtime, or using a sustained release tablet. Domperidone, another dopamine antagonist not crossing the blood-brain barrier, has fewer side effects but is not available in the United States. Cisapride was the best prokinetic drug for treating GERD but was withdrawn from the U.S. market because of reports of serious cardiac dysrhythmias (ventricular tachycardia, ventricular fibrillation, torsades de pointes, and QT prolongation) with associated cardiac arrest and deaths related to possible drug interactions.221

721

Section V  Esophagus Regulating the frequency of tLESRs is an attractive target for GERD treatment because of its pivotal role in most reflux episodes. Potential agents to modify this vagovagal reflex include atropine, morphine, CCKA (CCK-1) receptor antagonists, NO synthase inhibitors, and γ-aminobutyric acid (GABAB) agonists.222 Only the latter category of drug, baclofen, has been extensively studied in humans and found to be a potent inhibitory neurotransmitter in the central nervous system, antagonizing the release of neurotransmitters from vagal nerve afferents. The frequency of tLESRs, especially after meals, is decreased, reducing exposure time for acid and duodenal reflux. This correlates with improvement of acute and chronic symptoms in GERD patients.223,224 Baclofen needs to be titrated upward slowly (5  mg three or four times daily initially and increased as needed over 10 days to 40 to 60  mg per day). Side effects including drowsiness, nausea, and lowering of the threshold for seizures require discontinuation in up to 20% of patients. Potential applications could be in patients with nonerosive GERD or as adjunct therapy in patients with persistent symptoms on PPIs possibly related to nonacid duodenal reflux.224

that the overall esophagitis healing rates with H2RAs rarely exceeded 60% after up to 12 weeks of treatment, even when higher doses were used.226,227 Healing rates differ in individual trials depending primarily on the severity of esophagitis being treated: grades I and II esophagitis heal in 60% to 90% of patients, whereas grades III and IV heal in only 30% to 50% despite high-dose regimens.227 Although PPIs are more effective than H2RAs (Fig. 43-15; discussed following), nocturnal gastric acid breakthrough while on PPI therapy may cause reflux symptoms in some patients. H2RAs given at bedtime successfully eliminated this problem in one study, suggesting a new indication for H2RAs in the PPI era.228 However, this study used only a single evening dose and did not account for the tolerance that frequently develops to H2RAs over weeks to months.229 This tolerance impairs the effectiveness of chronic nocturnal dosing of H2RAs to eliminate nocturnal acid breakthrough,230 but suggests a useful role in as-needed medications in situations in which lifestyle indiscretions may promote nocturnal complaints. The H2RAs are very safe with a side effect rate (most of which are minor and reversible) of about 4%.225 Serum concentrations of phenytoin, procainamide, theophylline, and warfarin are higher after the administration of cimetidine and, to a lesser degree, ranitidine, whereas these interactions are not reported with the other two H2RAs.

Histamine-2 Receptor Antagonists (H2RAs)

(see also Chapter 53) These drugs (cimetidine, ranitidine, famotidine, and nizatidine) are more effective in controlling nocturnal, as compared with meal-related, acid secretion because the parietal cell is stimulated postprandially by gastrin acting via histamine and by acetylcholine (see Chapter 49).225 The four H2RAs are equally effective when used in proper doses, usually twice a day before meals. GERD trials find that heartburn can be significantly decreased by H2RAs, when compared with placebo, although symptoms are rarely abolished (Fig. 43-14). A comprehensive meta-analysis found

Proton Pump Inhibitors (see also Chapter 53) PPIs inhibit meal-stimulated and nocturnal acid secretion to a significantly greater degree than H2RAs231 but rarely make patients achlorhydric. After oral ingestion, acid inhibition is delayed because PPIs need to accumulate in the parietal cell secretory canaliculus to bind irreversibly to actively secreting proton pumps.232 Therefore, the slower a PPI is cleared from plasma, the more it is available for delivery to the proton pumps. PPIs should be taken before

100

100

(3) 80

(17) (13)

60

(12)

PPI (10) (10)

40 (3)

H2RA

(27) (4)

(22)

PPI

60

(23)

40 (2)

(25) (25) (9)

H2RA (5)

20

20

0

(5)

(8) Placebo

0 0

A

(2) (26)

80 Percentage of patients healed

Percentage of patients free of heartburn

722

1–2

3–4 Weeks

0

6–8

B

2

4

6

8

12

Weeks

Figure 43-14.  A, Symptom relief time curve over 8 weeks for a proton pump inhibitor (PPI) or histamine-2 receptor antagonist (H2RA) corrected for patients free of heartburn at baseline. By week 2, more patients treated with a PPI were asymptomatic compared with those treated with a H2RA, even after a much longer duration of treatment with the H2RA. B, Esophagitis healing time curve for PPI, H2RA, and placebo over 12 weeks. By 4 weeks treatment with a PPI healed esophagitis in more patients than the other two drug classes over 12 weeks, implying a substantial therapeutic gain. The numbers of studies included for each time point and treatment are shown in parentheses. (Data based on meta-analysis from Chiba N, Gara CJ, Wilkinson JM, Hunt RH. Speed of healing and symptom relief in grade II to IV gastroesophageal reflux disease: A meta-analysis. Gastroenterology 1997; 112:1798.)

Chapter 43  Gastroesophageal Reflux Disease

Wrap Esophagus

A

Wrap

140˚± 20 Esophagus

Esophagus

B

180˚

C

Wrap

Figure 43-15.  The most common surgical fundoplications used during antireflux surgery. A, The most popular worldwide is the 360-degree Nissen fundoplication. B, An anterior wrap (e.g., Thal, Dor) is commonly used to prevent gastroesophageal reflux after a Heller myotomy for achalasia. The experience with this repair is limited in patients with classical gastroesophageal reflux disease. C, The posterior wrap (Toupet) is popular in patients with poor esophageal motility because postoperative dysphagia is less frequent than after other operations. This is a 220- to 250-degree wrap. (From Oelschlager BK, Eubanks TR, Pellegrini CA. Hiatal hernia and gastroesophageal reflux disease. In: Townsend CM, Beauchamp RD, Foshee JC, et al, editors. Sabiston Textbook of Surgery: The Biological Basis of Modern Surgical Practice. 18th ed. Philadelphia: Saunders; 2007.)

the first meal of the day, when most proton pumps become active. Because not all pumps are active at any given time, a single PPI dose will not inhibit all pumps. A second dose, if necessary, can be taken before the evening meal. PPIs (omeprazole, lansoprazole, rabeprazole, pantoprazole, and esomeprazole) have superior efficacy compared with H2RAs on the basis of their ability to maintain an intragastric pH greater than 4 from 10 to 14 hours daily compared with approximately 6 to 8 hours daily with the H2RAs.233,234 PPIs are superior to H2RAs in completely relieving heartburn symptoms in patients with severe GERD, usually within one to two weeks (see Fig. 43-15A).227 PPI therapy has been shown in a Cochrane meta-analysis to be superior to placebo and H2RAs in nonerosive GERD and for undiagnosed reflux symptoms in primary care, although the effect is 20% to 30% lower than in patients with esophagitis.235 Controlled studies and a large metaanalysis report complete healing of even severe ulcerative esophagitis after eight weeks in more than 80% of patients taking PPIs compared with 51% on H2RAs and 28% receiving placebo (see Fig. 43-15B).227,236-239 In another recent Cochrane review involving 4064 patients in 26 trials,240 PPIs were superior to H2RAs in healing esophagitis at four to eight weeks (risk ratio, 0.47) with a number-to-treat of 3. In patients not healing initially, prolonged therapy with the same dose or an increased PPI dose usually resulted in 100% healing.241 Until recently, therapeutic efficacy among PPIs was similar. However, large studies have found the newest PPI esomeprazole 40  mg superior to omeprazole 20  mg and to lansoprazole 30  mg in healing esophagitis.242,243 A meta-analysis of 10 randomized clinical trials244 comparing esomeprazole to all other PPIs found the therapeutic advantage is minimal with LA grade A/B esophagitis (number-to-treat 50 and 33, respectively), and greater with severe LA grade C/D esophagitis (number-to-treat 14 and 8, res­pectively). This superiority is related to higher systemic bioavailability and less interpatient variability with esomeprazole. Several PPIs are available in the United States for intravenous use.245

PPIs are well tolerated, with headaches and diarrhea described as the most common side effects. Increased fasting serum gastrin levels are reported with all the PPIs, but the elevations generally do not exceed the normal range for gastrin and return to normal values within one to four weeks of drug discontinuation. Omeprazole decreases the clearance of diazepam and warfarin due to competition for the cytochrome P-450 isoenzyme P2C19.246 The four newer PPIs have minimal or no important drug-drug interactions.

MAINTENANCE THERAPIES

GERD may be a chronic relapsing disease, especially in patients with low LES pressure, severe grades of esophagitis, and difficult-to-manage symptoms.217 After esophagitis is healed, recurrence within six months of stopping medication occurs in more than 80% of patients with severe esophagitis and in 15% to 30% of those with milder esophagitis.202,247 Cochrane reviews have identified the superiority of PPIs over H2RAs in maintaining the remission of esophagitis over 6 to 12 months.248 Among 10 randomized trials, the relapse rate for esophagitis was 22% on PPIs versus 58% with H2RAs, with a number-to-treat of 2.5. The FDA has approved all the PPIs, sometimes at one half the acute dose, for maintenance therapy, but only ranitidine 150  mg twice a day among the H2RAs has maintenance indications for mild esophagitis. Many clinicians now place their patients with severe disease (daily symptoms, severe esophagitis, or complications) on chronic PPI therapy indefinitely. The efficacy of this approach is supported by open, compassionate use data primarily from the Netherlands and Australia.249 In a study of 230 patients with severe esophagitis healed with 40  mg omeprazole, all subjects remained in remission for up to 11 years. More than 60% were maintained on omeprazole 20  mg a day, whereas higher doses of 60  mg or more were necessary in only 12% of patients, confirming a lack of tolerance to PPIs. Relapses were rare (one per 9.4 years of follow-up), strictures did not occur, and Barrett’s esophagus did not progress.245

723

724

Section V  Esophagus Although PPIs offer the best symptom relief and esophagitis healing, many patients do well long term on lesser treatments after having their complaints initially alleviated with PPIs. Using this “step-down approach,” a recent Veterans Affairs study reported that 58% of 71 patients on chronic PPIs could be switched to H2RAs and/or prokinetics or taken off medication completely.250 Younger age and severe heartburn symptoms predicted PPI requirement. Overall, this approach saved money for the health care system. A similar study by the same investigators found that 80% of patients using multiple-dose PPIs could be stepped down to single-dose PPI, remaining symptom free for six months with considerable cost savings.251 Hence the adage “once on a PPI, always on a PPI” is not true. The initial safety concerns with the PPIs was the potential to develop stomach or other cancers with long-term use. This was initiated by reports of omeprazole producing hypergastrinemia and gastric carcinoid tumors in rats, changes also subsequently demonstrated with chronic ranitidine therapy and subtotal resection of the gastric fundus.252 However, rats have a high density of enterochromaffin-like (ECL) cells and an exaggerated gastrin response to achlorhydria; chronic omeprazole therapy in other species with lower densities of ECL cells (mice, dog, man) has not caused carcinoid tumors. Another review found no evidence of an increased risk of colon cancer with chronic PPI use.253 Fundic gland polyps are the most common gastric polyp found at endoscopy. Their association with chronic PPI use has been a topic of debate since these drugs were first described. A recent study evaluated 599 patients of whom 322 used PPIs and 107 had fundic gland polyps.254 Longterm PPI use was associated with up to a four-fold increase in the risk of fundic gland polyps. Low-grade dysplasia was found in one fundic gland polyp. Etiologically, these polyps seem to arise because of parietal cell hyperplasia and parietal cell protrusions resulting from acid suppression. Recent studies confirm that chronic acid suppression may be associated with an increased risk of community-acquired pneumonias and enteric infections. In a large Scandinavian population-based study,255 the adjusted relative risk for pneumonia among current PPI users, compared with those who stopped using PPIs, was 1.89. Current users of H2RAs had a 1.63-fold increased risk of pneumonia compared with those who stopped. A significant positive dose-response relationship was observed in the PPI users. Likewise, a recent systematic review found an increased risk of enteric infections with acid suppression.256 The correlation was stronger with Salmonella, Campylobacter, and other enteric infections, compared with Clostridium difficile, and greater with PPI compared with H2RA therapy. Chronic high-dose use of PPIs may also affect the absorption of calcium and vitamin B12. A nested case-controlled study from the United Kingdom among 13,556 patients found that the risk of hip fractures increased with chronic PPI use over one year (adjusted odds ratio, 1.44), especially in those patients receiving high-dose PPIs (adjusted odds ratio, 2.65). A smaller but still significant risk was observed in chronic H2RA users.257 A large Canadian study258 reached similar conclusions, but found the risk for hip fractures became apparent after five years of treatment (adjusted odds ratio, 1.62) and after seven years for all osteoporotic fractures (adjusted odds ratio, 1.92). The mechanism for this association is unknown and cause-and-effect is not proven as yet. It has been suggested that if PPIs cause osteoporosis, they may interfere with insoluble calcium absorption or possibly inhibit the osteoclastic proton transport system, potentially reducing bone resorption. PPIs could retard the

absorption of vitamin B12 by decreasing gastric acidity; reducing the release of cobalamin from dietary protein; or by promoting small bowel bacterial overgrowth, thereby increasing luminal cobalamin consumption. However, cohort and case-control studies have not shown a convincing link between PPI use and vitamin B12 deficiency.259 Finally, a study suggested that patients on long-term omeprazole who are infected with H. pylori develop atrophic gastritis, a precursor to gastric adenocarcinoma, at a more rapid rate than noninfected patients.260 Nevertheless, a subsequent FDA panel determined that the available data were insufficient for recommending screening and treatment of H. pylori infection in patients on long-term PPI therapy.261

SURGICAL THERAPY

Only surgical fundoplication can correct the physiologic factors contributing to GERD and potentially eliminate the need for long-term medications. Antireflux surgery reduces GER by increasing basal LES pressure, decreasing episodes of tLESRs, and inhibiting complete LESR.262 This is done by reducing the hiatal hernia into the abdomen, reconstructing the diaphragmatic hiatus, and reinforcing the LES.263 Before laparoscopic surgery, the three most common operations were the Nissen fundoplication, Belsey Mark IV repair, and Hill posterior gastropexy. Since the introduction of minimally invasive surgery, the two most popular procedures, performed laparoscopically through the abdomen are the Nissen 360-degree fundoplication and the Toupet partial fundoplication (see Fig. 43-15).264 The former is a superior operation with better long-term durability, but it causes more postoperative dysphagia and gas bloat symptoms.265,266 The typical hospital stay is 1 to 2 days, and many patients return to normal activity in 7 to 10 days. Patients with more severe disease and a short esophagus suggested by a large nonreducible hernia, tight stricture, or long-segment Barrett’s esophagus will require a Collis lengthening procedure to create a 3- to 5-cm neoesophagus that allows the fundoplication to be placed in the abdomen under minimal tension.267 The popularity of antireflux surgery has undergone an explosion since the advent of the laparoscopic operation. The numbers of antireflux operations performed in the United States nearly tripled from 11,000 per year in 1985 (open surgery) to a peak of nearly 32,000 in 1999 but have leveled off at around 24,000 cases per year (11 cases per 100,000 adult population).268,269 A systematic review identified six randomized controlled trials involving 449 patients that compared open and laparoscopic fundo­ plication.270 There was no significant difference in recurrence rates between the procedures, and laparoscopic fundoplication was associated with lower operative morbidity (number-to-treat to prevent complication, 8) and shorter hospital stay. In the PPI era, symptom resolution on treatment helps predicate the success of antireflux surgery for classic as well as atypical symptoms.271 Antireflux surgery is a reasonable option in (1) the healthy patient with typical or atypical GERD symptoms well controlled on PPIs desiring alternative therapy because of drug expense, poor medication compliance, or fear of unknown long-term side effects; (2) patients with volume regurgitation and aspiration symptoms not controlled on PPIs; and (3) recurrent peptic strictures in younger patients.265 Patients recalcitrant to PPI therapy may well have another etiology for their complaints (e.g., pill esophagitis, gastroparesis, functional heartburn) and should be approached cautiously with surgery.

Chapter 43  Gastroesophageal Reflux Disease Testing must be done before antireflux surgery. Endoscopy is necessary to exclude stricture, Barrett’s esophagus, and dysplasia or carcinoma. A barium esophagogram can help define a nonreducible hiatal hernia, a shortened esophagus, and poor esophageal motility. Esophageal manometry, possibly combined with impedance if available, will identify ineffective esophageal peristalsis and previously misdiagnosed achalasia or scleroderma. Twenty-four-hour pH testing is necessary in patients with nonerosive GERD or those with esophagitis not responding to PPI therapy. Gastric analysis and gastric emptying studies may be indicated in select patients. Careful testing will result in modification of the original operation or an alternative diagnosis in approximately 25% of patients.187 Antireflux surgery relieves reflux symptoms and reduces the need for stricture dilation in more than 90% of patients,272 but Barrett’s esophagus rarely regresses and the risk of developing esophageal cancer is unchanged.273 Older studies found antireflux surgery superior to antacids, H2RA, and prokinetic therapy,204 but not PPI therapy, especially when dose titration is permitted.274 Mortality is rare (<1%) after antireflux surgery, but new postoperative complaints occur in up to 25% of patients including dysphagia, gas bloat, diarrhea, and increased flatus.272 Most symptoms improve over 1 year, but persistent complaints suggest too tight a wrap, a displaced fundoplication, or inadvertent damage to the vagus nerve. Successful antireflux surgery does not guarantee a permanent cure. Best surgical results are obtained by experienced surgeons in high-volume centers who report long-term symptom recurrence in only 10% to 15% of patients.265 However, many operations are performed in lower-volume community hospitals where the results are more variable. A study from community hospitals in a midsize Midwest city275 found that 32% of patients had symptom relapses after two years and 7% required repeat surgery. In contrast, a review of more than 3000 patients undergoing antireflux surgery at VA medical centers between 1990 and 2001 was more reassuring.269 Postoperative dysphagia was recorded in 19.4% of patients, 6.4% required esophageal dilations, and a repeat antireflux operations was needed in only 2.3% of patients. There seems to be increasing consensus from long-term follow-up studies that 25% to 62% of patients are back on some type of acid suppressive medication 5 to 15 years after their antireflux surgery.269,275-277 However, the evidence for recurrent GERD by pH testing is infrequent. Tertiary specialized centers are seeing an increased rate of fundoplication failures. The most common reason for failure are herniation of the intact fundoplication into the chest, “slipped” fundoplication with a recurrent hiatal hernia possibly due to a short esophagus, paraesophageal hernia through an intact fundoplication, too tight a fundoplication, and malpositioned fundoplication usually on the cardia of the stomach. Total breakdown of the fundoplication is now rare.278 Revisional antireflux surgery needs to be performed by very experienced surgeons and can be done laparoscopically, although many prefer an open approach. Reoperation has increased morbidity and mortality compared with the initial operation.

ENDOSCOPIC THERAPY

A variety of endoscopic techniques for the treatment of GERD have been developed as alternatives to antisecretory therapy or antireflux surgery.279-281 These techniques include the delivery of radiofrequency energy to the gastroesophageal junction (Stretta), injection of bulking agents (Enteryx), or implantation of a bioprosthesis (Gatekeeper) into the LES,

and suture plication of the proximal gastric folds (EndoCinch endoscopic plication system). Studies to date have primarily enrolled PPI-dependent patients without severe esophagitis or large hiatus hernia. All of these techniques decrease reflux symptoms, improve quality of life, and decrease the need for antisecretory medications. However, physiologic studies are much less impressive, with LES pressure rarely increasing, pH normalizing in only 30% of patients, and even mild esophagitis infrequently healing.279-281 Controlled studies with Stretta,282 Enteryx,281 and the Plication283 system likewise show a decrease in heartburn symptoms and improved quality of life after the active therapy compared with the sham-treated group after three to six months. Only the Plication study showed a significant decrease in pH values, but only by 18%, whereas the other techniques observed no change in pH or LES parameters.283 Most studies of endoscopic therapy have only limited follow-up information on a relatively small number of patients. The durability of these techniques beyond one to two years remains unclear and seems to gradually decrease over time. The cost-effectiveness of these techniques is difficult to define. Most importantly, safety issues have haunted these procedures, especially when used in the broader community of gastroenterologists. Chest pain, bleeding, esophageal perforations, mediastinitis, and at least eight deaths to date have been attributed to these endoscopic techniques. Serious adverse events, including deaths, led to the voluntary withdrawal of Enteryx by the manufacturer in September 2005 and suspension of the Gatekeeper clinical program in late 2005. A recent AGA Institute medical position statement recommended that “current data suggest that there are no definite indications for endoscopic therapy for GERD at this time.”284

TREATMENT OF COMPLICATIONS CHEST PAIN AND EXTRAESOPHAGEAL MANIFESTATIONS

The efficacy of acid suppression, especially by PPI treatment, in the extraesophageal presentations of GERD is variable. There are two systemic reviews,285,286 both suggesting that patients with noncardiac chest pain respond to PPIs better than placebo. These reports identified eight randomized controlled trials that assessed 321 patients, with a pooled relative risk for continued chest pain after PPI therapy compared with placebo of 0.54, with a number-totreat of 3. Systematic reviews and randomized trials, however, do not support the efficacy of aggressive acid suppression, particularly with PPIs, in other extraesophageal disorders, such as chronic cough,287 asthma,288 or ear, nose, and throat disorders.289 Some patients definitely respond, usually over several months, but the heterogeneous etiology of these groups and the lack of pretreatment predictors of response confound any treatment approach to these patients. Sleep disturbances may occur in up to 75% of patients with GERD, impairing quality of life. In a large multicenter study, patients with GERD-associated sleep disturbances and nighttime heartburn were randomized to two doses of esomeprazole (40 mg and 20 mg) or placebo for four weeks.290 GERD-related sleep disturbances resolved in significantly more patients on esomeprazole 40  mg (73.7%) or 20  mg (73.2%) than those who received placebo (41.2%). These changes were associated with improved sleep quality and daytime productivity.

725

726

Section V  Esophagus PEPTIC ESOPHAGEAL STRICTURES

Dysphagia in patients with esophageal strictures and rings is related to stricture diameter and severity of esophagitis.291 When the esophageal lumen diameter is less than 13 mm, dysphagia is common and esophageal dilation is required. Simple short strictures can be dilated by blind peroral passage of rubber Hurst (rounded ends) or Maloney (tapered ends) mercury-filled dilators of increasing sizes (16 to 60 French, 3 French equals 1 mm). Complicated longer, tighter, or more irregular strictures will require bougienage over a guidewire using hollow-centered, Savary, plastic-covered polyvinyl dilators or balloon (Gruentzig) dilators.292 The extensive use of PPIs has markedly impacted our treatment of peptic strictures and esophageal rings. PPIs are superior to H2RAs in relieving the symptoms of heartburn and dysphagia experienced by stricture patients while reducing the frequency of repeat dilations and the cost for treating these patients.293 Several studies in community and veterans hospitals note an approximate 33% decline in the incidence of recurrent strictures. The timeline for this decrease parallels the marked increase in PPI use since 1995.294 Another study convincingly shows that in patients with symptomatic Schatzki’s rings, maintenance PPI therapy after bougienage markedly decreases future relapses of the rings.295 In a randomized study, 30 patients with symptomatic rings without esophagitis were dilated and randomized to placebo or omeprazole 20 mg per day. In the treated group, one patient relapsed after 13 months, whereas seven patients relapsed on placebo after a mean of 20 months.

KEY REFERENCES

American Lung Association Asthma Clinical Research Centers, Mastronarde JG, Anthonisen NR, et al. Efficacy of esomeprazole

for treatment of poorly controlled asthma. N Engl J Med 2009; 360:1487-99. (Ref 288.) Dent J. Microscopic esophageal mucosal injury in nonerosive reflux disease. Clin Gastroenterol Hepatol. 2007; 5:4-16. (Ref 172.) DeVault KR, Castell DO: Updated guidelines for the diagnosis and treatment of gastroesophageal reflux disease. Am J Gastroenterol 2005; 100:190. (Ref 171.) Falk GW, Fennerty MB, Rothstein RI. AGA Institute technical review on the use of endoscopic therapy for gastroesophageal reflux disease. Gastroenterology 2006; 131:1351-66. (Ref 284.) Fass R. Erosive esophagitis and non-erosive reflux disease. Comparison of epidemiologic, physiologic and therapeutic characteristics. J Clin Gastroenterol 2007; 41:131-7. (Ref 200.) Friedenberg FK, Xanthopoulos M, Foster GF, Richter JE. The association between gastroesophageal reflux disease and obesity. Am J Gastroenterol 2008; 103:2111-22. (Ref 30.) Gralnek IM, Dulai GS, Fennerty MB, Speigel BMR. Esomeprazole versus other PPIs in erosive esophagitis: A meta-analysis of rando­mized clinical trials. Clin Gastroenterol Hepatol 2006; 4:1452-58. (Ref 244.) Hirano I, Richter JE. ACG practice guidelines: Esophageal reflux testing. Am J Gastroenterol 2007; 102:668-85. (Ref 176.) Kaltenback T, Crockett S, Gerson LB. Are lifestyle measures effective in patients with gastroesophageal reflux disease? Am J Gastroenterol 2006; 101:2128-38. (Ref 211.) Mittal RK, Balaban DH: The esophagogastric junction. N Engl J Med 1997; 336:924-32. (Ref 47.) Mittal RK, Holloway RH, Penagini R, et al. Transient lower esophageal sphincter relaxation. Gastroenterology 1995; 109:601-610. (Ref 54.) Moayyedi P, Talley N. Gastroesophageal reflux disease. Lancet 2006; 367:2086-100. (Ref 240.) Serag HB. Time trends for gastroesophageal reflux disease: A systematic review. Clin Gastroenterol Hepatol 2007:17-26. (Ref 14.) van Malenstein H, Farré R, Sifrim D. Esophageal dilated intercellular spaces (DIS) and nonerosive reflux disease. Am J Gastroenterol. 2008; 103(4):1021-8. (Ref 86.) Watson DI. Laparoscopic treatment of gastro-oesophageal reflux disease. Best Pract Res Clin Gastroenterol 2004; 18:19-35. (Ref 265.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

44 Barrett’s Esophagus Stuart Jon Spechler and Rhonda F. Souza

CHAPTER OUTLINE Diagnosis  727 Epidemiology  727 Pathogenesis  728 Molecular Biology of Neoplasia  729 Dysplasia  729

Barrett’s esophagus is the condition in which an abnormal columnar epithelium that is predisposed to malignancy replaces the stratified squamous epithelium that normally lines the distal esophagus.1 The condition is named for Norman Barrett, an Australian surgeon who drew attention to the columnar-lined esophagus in 1950.2 Barrett’s esophagus is a consequence of chronic gastroesophageal reflux disease (GERD), which damages the esophageal squamous epithelium and causes it to heal through a metaplastic process in which columnar cells replace reflux-damaged squamous cells. The columnar-lined esophagus causes no symptoms, and the condition has clinical importance only because it is a risk factor for esophageal adenocarcinoma, a tumor whose frequency has increased more than six-fold over the past several decades.3

Management  730 Treatment of Gastroesophageal Reflux Disease  730 Endoscopic Surveillance for Dysplasia  730 Treatment of Dysplasia  731 Recommendations  732

epithelium, which is composed almost exclusively of mucus-secreting cells, also is metaplastic, has malignant predisposition, and can be considered diagnostic of Barrett’s esophagus.5,6 This debate remains unresolved. Barrett’s esophagus can be further categorized as longsegment (when the metaplastic epithelium extends at least 3 cm above the GEJ) or short-segment (when <3 cm of metaplastic epithelium lines the esophagus).7 Another more recently proposed system for categorizing Barrett’s esophagus, the Prague C and M criteria, identifies the circumferential (C) and the maximum extent (M) of Barrett’s metaplasia.8 Data suggest that the cancer risk in Barrett’s esophagus may vary with the extent of the metaplastic lining. However, the clinical value of the proposed clas­ sification systems has not been established and presently patients with any extent of Barrett’s metaplasia are managed similarly.

DIAGNOSIS Barrett’s esophagus is diagnosed by endoscopic examination, and two criteria must be fulfilled. First, the endoscopist must ascertain that columnar-appearing epithelium lines the distal esophagus. Second, biopsy specimens of that columnar-appearing epithelium must show evidence of metaplasia, which is a change from one adult cell type to another. To ascertain that columnar-appearing epithelium lines the distal esophagus, the endoscopist first must locate the gastroesophageal junction (GEJ, which is recognized as the most proximal extent of the gastric folds), and then determine that columnar-appearing epithelium extends above the GEJ into the esophagus (Fig. 44-1). Endoscopically, columnar epithelium has a reddish color and velvetlike texture that can be distinguished readily from normal esophageal squamous epithelium, which is pale and glossy. There is disagreement among experts regarding the histologic type of epithelium required to confirm that there is evidence of metaplasia in the esophagus.4 Virtually all would agree that the finding of an intestinal-type epithelium with goblet cells (which has been called intestinal meta­ plasia, specialized intestinal metaplasia, or specialized columnar epithelium) is clear evidence of metaplasia. Most published studies on Barrett’s esophagus have used intestinal metaplasia as a requisite diagnostic criterion. However, some authorities argue that gastric cardiac-type

EPIDEMIOLOGY Barrett’s esophagus typically is discovered during endoscopic examinations performed for the evaluation of GERD symptoms in middle-aged and older adults.9 The average age at the time of diagnosis is approximately 55 years. The condition is rare in children younger than age 10 and virtually nonexistent in children younger than age 5.10 White men predominate in most series and, for unknown reasons, Barrett’s esophagus is uncommon in black and Asian populations. Among adult patients who have endoscopic examinations because of GERD symptoms, long-segment Barrett’s esophagus is found in 3% to 5%, whereas 10% to 20% have short-segment Barrett’s esophagus.1 In the general adult population of Western countries, the prevalence of Barrett’s esophagus (predominantly short-segment) is between 1.6% and 6.8%.11,12 Published estimates on the annual incidence of cancer in patients with long-segment Barrett’s esophagus have ranged from 0.2% to 2.9%, but it has been shown that many of those estimates were based on older, small studies that suffered from publication bias. Modern, larger studies, which are less susceptible to such bias, suggest that the risk of cancer in the general population of patients with Barrett’s esophagus is approximately 0.5% per year.13

727

728

Section V  Esophagus Table 44-1  Proposed Physiologic Abnormalities Contributing to Gastroesophageal Reflux Disease in Patients with Barrett’s Esophagus* ABNORMALITY

POTENTIAL CONSEQUENCES

Extreme LES hypotension Ineffective esophageal motility

Gastroesophageal reflux Defective clearance of refluxed material Reflux of highly acidic gastric juice Esophageal injury caused by reflux of bile acids and pancreatic enzymes Delayed healing of reflux-damaged esophageal mucosa Loss of sensation to refluxed caustic material and resulting failure to initiate therapy

Gastric acid hypersecretion Duodenogastric reflux Decreased salivary secretion of epidermal growth factor Decreased esophageal pain sensitivity

*See Chapter 43 for detailed discussion of these abnormalities. LES, lower esophageal sphincter.

Figure 44-1.  Endoscopic photograph of Barrett’s esophagus. The arrows mark the gastroesophageal junction (GEJ), which is identified endoscopically as the most proximal extent of the gastric folds. The reddish color and velvet-like texture of the Barrett’s epithelium contrast sharply with the pale and glossy appearance of the esophageal squamous epithelium. Note that Barrett’s columnar epithelium extends well above the GEJ to line the distal esophagus.

The epidemiology of esophageal adenocarcinoma is similar to that of Barrett’s esophagus. GERD is strongly associated with both conditions and, like Barrett’s esophagus, esophageal adenocarcinoma affects white men predominantly.9,14 Obesity, especially with central adiposity, pre­ disposes to both Barrett’s esophagus and esophageal adenocarcinoma,14,15 and the dramatic rise in the frequency of obesity in the United States has paralleled a similar rise in the prevalence of Barrett’s cancer. The mechanisms underlying these associations with obesity are not clear, but may relate to the fact that central adiposity predisposes to GERD, perhaps by increasing intra-abdominal pressure16 (see also Chapters 43 and 46). Obesity also is associated with elevated serum levels of pro-proliferative hormones such as insulin-like growth factor I (IGF I) and leptin, and with decreased levels of the antiproliferative hormone adiponectin, factors that may contribute to carcinogenesis in Barrett’s esophagus. It has been proposed that the declining frequency of infection with Helicobacter pylori in Western populations also may be contributing to the rising frequency of esophageal adenocarcinoma (see Chapter 46). A number of studies have suggested that H. pylori infection may protect against the development and neoplastic progression of Barrett’s esophagus, perhaps because, in a subset of patients, this infection may prevent GERD by decreasing gastric acid secretion.17 Other factors that appear to protect against the development of esophageal adenocarcinoma include the use of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs),18,19 and the consumption of a diet high in fruits and vegetables.18 Although cigarette smoking and alcohol consumption are very strong risk factors for squamous cell carcinoma of the esophagus, cigarette smoking only modestly increases the risk for esophageal adenocarcinoma and alcohol does not appear to affect that risk at all.18

PATHOGENESIS Patients with long-segment Barrett’s esophagus often have severe GERD (see Chapter 43). Table 44-1 lists some physi-

ologic abnormalities that have been reported in Barrett’s patients, and suggests how those abnormalities might contribute to GERD severity. Individual patients may exhibit any, all, or none of those abnormalities, and their prevalence in Barrett’s esophagus is disputed. For example, some investigators have described normal gastric acid secretion in patients with long-segment Barrett’s esophagus.20 In addition, many patients with short-segment Barrett’s esophagus have no GERD symptoms and no endoscopic signs of esophagitis. Indeed, one large study has suggested that shortsegment Barrett’s esophagus may affect approximately 5% of adults, irrespective of the presence of GERD symptoms.11 Studies have shown that even in healthy volunteers, the very distal esophagus can be exposed to acid for more than 10% of the day.21 Such acid exposure can damage the esophagus directly and indirectly when nitrite (generated from dietary nitrate) reacts with acid to produce nitric oxide. High concentrations of nitric oxide in the distal esophagus have been observed in patients with GERD who have ingested nitrate.22 The progenitor cells that give rise to Barrett’s metaplasia are not known. The prevailing hypothesis is that metaplasia results when GERD damages the esophageal squamous epithelium, thereby exposing multipotential stem cells in the basal layers to gastric juice, which stimulates their abnormal differentiation into columnar cells. Two other candidates for Barrett’s progenitor cells include stem cells in the ducts of the esophageal submucosal glands and circulating bone marrow stem cells.23 Genes that appear to play a key role in the squamous-to-columnar metaplasia of Barrett’s esophagus include certain Cdx genes, which are known to mediate the differentiation of intestinal epithelial cells, and the gene encoding bone morphogenetic protein (BMP)-4, which also is involved in columnar cell differentiation.24 Reflux esophagitis appears to upregulate the expression of these genes by the squamous epithelium. Barrett’s epithelial cells appear to be more capable of resisting reflux-induced esophageal injury than the native squamous epithelial cells. Unlike squamous cells, for example, Barrett’s cells secrete mucins and express the tight-junction protein claudin 18, features that render the epithelium more resistant to acid-peptic attack.25,26 Unfortunately, Barrett’s epithelium also is predisposed to neoplasia.

Chapter 44  Barrett’s Esophagus MOLECULAR BIOLOGY OF NEOPLASIA During carcinogenesis, Barrett’s epithelial cells accumulate a series of genetic and epigenetic alterations that endow the cells with the physiological attributes of malignancy (see Chapter 3). Those include self-sufficiency in growth signals, insensitivity to anti-growth signals, evasion of apoptosis, limitless replicative potential, sustained angiogenesis, and the abilities to invade adjacent structures and to metastasize (Fig. 44-2).27 Numerous genetic alterations have been described during the neoplastic progression of Barrett’s esophagus. Although a single such alteration may have multiple disparate effects, conceptually it can be useful to classify the alteration according to the major physiologic cancer attributes that it endows (see Fig. 44-2).27 For example, the expression of oncogenes (e.g., cyclin D1, K-ras), growth factors (e.g., transforming growth factor-α [TGF-α]), and growth factor receptors (e.g., epidermal growth factor receptor [EGFR]) enable Barrett’s cells to acquire self-sufficiency in growth signals. Insensitivity to antigrowth signals occurs primarily through the inactivation of tumor suppressor genes (e.g., TP53 and p16). Inactivation of TP53 also enables cells to evade apoptosis. Reactivation of the enzyme telomerase, which enables the cells to replace telomeres needed for cell division, can endow the cells with limitless replicative potential.18 Neoplasms can increase their vascular supply by secreting angiogenic factors such as vascular endothelial growth factor (VEGF).18 Finally, for neoplastic cells to invade and metastasize, they must dissociate themselves from surrounding cells by disrupting cell adhesion proteins such as the cadherins and catenins, and by degrading the extracellular matrix through the secretion of enzymes such as matrix metalloproteases (MMPs).18 During carcinogenesis, Barrett’s epithelial cells display genetic instability manifested as gains or losses in segments

Genetic Instability Metaplasia

Dysplasia

Cancer

Cancer attributes Self-sufficiency in growth signals

↑ Cyclins ↑ TGF-α and EGFR

Insensitivity to anti-growth signals

Inactivation of p16 Inactivation of TP53

Limitless replicative potential

Telomerase reactivation

Tissue invasion and metastasis

DYSPLASIA Before neoplastic Barrett’s cells become malignant, some of the same genetic alterations that endow the physiologic attributes of malignancy also cause morphologic changes in the tissue that the pathologist recognizes as dysplasia (Fig. 44-3). Dysplasia (also called intraepithelial neoplasia) can be viewed as the histologic expression of genetic alterations that favor unregulated cell growth.30 Dysplasia is recognized by cytologic and architectural abnormalities in esophageal biopsy specimens that include (1) nuclear changes such as enlargement, pleomorphism, hyperchromatism, stratification, and atypical mitoses; (2) loss of cytoplasmic maturation; and (3) crowding of tubules and villiform surfaces. Dysplasia is categorized as low-grade or high-grade depending on the degree of histologic abnormalities, with more pronounced abnormalities assumed to reflect more severe genetic damage and greater potential for carcinogenesis. Pathologists have difficulty distinguishing low-grade dysplasia in Barrett’s esophagus from reactive changes caused by reflux esophagitis, and inter-observer agreement for the diagnosis of low-grade dysplasia may be less than 50%. Interobserver agreement is better (approximately 85%) for high-grade dysplasia, but there is substantial disagreement among pathologists in distinguishing high-grade dysplasia from intramucosal carcinoma (see Chapter 46). Dysplasia in Barrett’s esophagus often causes no endoscopically apparent abnormalities, and dysplasia can be patchy in its extent and severity. These factors contribute to the substantial problem of biopsy sampling error in iden-

↑ K-ras

Evasion of apoptosis

Sustained angiogenesis

of chromosomes, which alter the cells’ deoxyribonucleic acid (DNA) content. Aneuploidy is the condition in which there is abnormal cellular DNA content, and aneuploid cells are at increased risk for neoplastic progression.28 Aneuploidy can be detected by flow cytometry and by fluorescence in situ hybridization (FISH), and has been proposed as a biomarker for neoplastic progression in Barrett’s esophagus, as have a number of the genetic alterations discussed in the preceding paragraph.29 Although there have been some promising preliminary studies, molecular biomarkers are not yet ready for routine clinical use in patients with Barrett’s esophagus.

↑ VEGF ↓ Cadherin-catenin MMP expression

Figure 44-2.  Molecular biology of neoplasia in Barrett’s esophagus. Some of the major genetic alterations that have been identified during carcinogenesis in Barrett’s esophagus are shown. The histologic stage at which each genetic change has been recognized is also depicted. These DNA alterations allow the cells to acquire the physiologic attributes of malignancy. EGFR, epidermal growth factor receptor; MMP, matrix metalloprotease; TGF, transforming growth factor; VEGF, vascular endothelial growth factor.

Figure 44-3.  Dysplasia in Barrett’s esophagus. This biopsy specimen taken during endoscopic surveillance shows low-grade dysplasia in the gland at 11 o’clock and high-grade dysplasia in the gland in the center of the photomicrograph.

729

730

Section V  Esophagus tifying dysplasia. Although endoscopists traditionally have used a four-quadrant biopsy sampling system (essentially a random sampling technique) to find dysplasia in Barrett’s esophagus, this system can miss areas of dysplasia and even cancer. In series of patients who had esophagectomies because endoscopic examination with biopsies revealed high-grade dysplasia in Barrett’s esophagus, for example, a number of studies have found that invasive cancer is present in 30% to 40% of the resected esophagi.31 However, a critical review of those studies suggests that 13% is a more accurate estimate of the frequency of invasive cancer in this situation.32 Researchers have tried to develop endoscopic techniques for recognizing dysplasia and early cancer in Barrett’s esophagus including chromoendoscopy, autofluorescence endoscopy, magnification endoscopy, narrow band imaging, optical coherence tomography, Raman detection methods, and confocal laser endomicroscopy.33 There have been some promising preliminary results (see Chapter 46) but presently none of the techniques has provided sufficient clinical information to justify its routine application in clinical practice. The overall incidence of cancer development in patients with Barrett’s esophagus is approximately 0.5% per year. A recent study suggests that patients who have non-neoplastic Barrett’s esophagus develop low-grade dysplasia at the rate of 4.3% per year, and high-grade dysplasia at the rate of 0.9% per year.34 Few meaningful data are available on the natural history of low-grade dysplasia in Barrett’s esophagus, but a recent study that included 156 patients with low-grade dysplasia found that they developed cancer at an incidence of 0.6% per year.35 For patients with high-grade dysplasia, however, the cancer risk is substantially higher. A recent critical review estimates that the rate of cancer development for patients with high-grade dysplasia in Barrett’s esophagus is 4% to 6% per year.30

MANAGEMENT TREATMENT OF GASTROESOPHAGEAL REFLUX DISEASE

The general approach to the treatment of GERD for patients with Barrett’s esophagus is very similar to that recommended for patients who have GERD without Barrett’s esophagus (see Chapter 43). One important difference, however, is that modern authorities generally advocate initial and maintenance therapy with a proton pump inhi­ bitor (PPI) for patients with Barrett’s esophagus, irrespective of symptoms and signs of esophagitis. This practice is based on indirect evidence suggesting that acid reflux promotes carcinogenesis in Barrett’s metaplasia, and that aggressive control of acid reflux may interfere with carcinogenesis.35 For patients with Barrett’s esophagus, the elimination of heartburn by antisecretory therapy should not be construed as evidence that acid reflux has normalized. Studies using esophageal pH monitoring have shown that Barrett’s patients may be rendered asymptomatic by PPIs given in dosages that fail to normalize esophageal acid exposure. In one such study of 48 patients with Barrett’s esophagus, for example, 24 of them had persistently abnormal acid reflux during PPI therapy that had abolished their GERD symptoms.36 It has been proposed that patients with long-segment Barrett’s esophagus might be unusually resistant to PPIs, but another study has suggested that this problem is not due to gastric resistance to the antisecretory effects of PPIs.37 In that study, patients with long-segment Barrett’s esophagus treated with

high doses of esomeprazole exhibited normal degrees of gastric acid suppression, but up to 23% still had abnormal esophageal acid exposure. This suggests that the so-called PPI resistance of patients with Barrett’s esophagus is a consequence of their profound reflux diathesis. Some surgeons have proposed that fundoplication might be more effective than antisecretory therapy for preventing cancer in Barrett’s esophagus, but a number of high quality studies on this issue refute that contention. These studies include one randomized controlled trial of medical and surgical GERD therapies, studies using large patient databases, and some meta-analyses.38-40 Available data suggest that antireflux surgery should not be performed solely for cancer prevention in patients with Barrett’s esophagus.

ENDOSCOPIC SURVEILLANCE FOR DYSPLASIA

Although a number of medical societies recommend regular endoscopic surveillance for patients with Barrett’s esophagus, some authorities have questioned the wisdom of these surveillance programs with arguments that can be summarized as follows: (1) Endoscopy is expensive, (2) endoscopy has risks and adverse emotional and financial consequences, (3) the absolute risk of cancer for patients with Barrett’s esophagus (0.5% per year) is small and therefore the large majority of patients derive no benefit from endoscopic surveillance, and (4) there is no proof, in the form of a randomized controlled trial, that endoscopic surveillance for Barrett’s esophagus has any effect on patient survival. Proponents of endoscopic surveillance counter the aforementioned arguments as follows41: (1) the concept that surveillance for Barrett’s esophagus can prevent deaths from esophageal adenocarcinoma seems reasonable, (2) no proof of efficacy in the form of a randomized controlled trial is likely to become available in the foreseeable future, (3) a number of observational studies suggest that surveillance is beneficial, (4) virtually all of the published computer models on this issue suggest that surveillance can be beneficial, and (5) the risks of endoscopy for otherwise healthy individuals with Barrett’s esophagus are very small, and no study has shown an overall survival disadvantage for patients in surveillance programs. The potentially adverse emotional and financial consequences of establishing a diagnosis of Barrett’s esophagus are regrettable, but less so than the failure to prevent an esophageal cancer. Therefore, the proponents argue, it is ethically wrong for physicians to forgo the potentially life-saving practice of performing endoscopic surveillance for Barrett’s esophagus while awaiting the results of a definitive study that may never appear. There have been numerous debates at medical meetings and in medical journals regarding the utility of surveillance for Barrett’s esophagus, with no clear winners. This issue is likely to remain contentious for the foreseeable future. Likewise, whether certain individuals should be screened for Barrett’s esophagus is also controversial (see Chapter 46). Some authorities also recommend a program of expectant management with intensive endoscopic surveillance (i.e., endoscopic examinations every three to six months) for patients with high-grade dysplasia in Barrett’s esophagus, withholding more invasive treatments (discussed following) until biopsy specimens reveal adenocarcinoma.30,42 Although this practice has been endorsed as a management option by the American College of Gastroenterology, few published data directly support the safety and efficacy of intensive surveillance for high-grade dysplasia. Available studies show that intensive endoscopic surveillance generally is safe, but even patients who are compliant with intensive surveillance programs can develop incurable cancers.30

Chapter 44  Barrett’s Esophagus TREATMENT OF DYSPLASIA

Because of difficulties in verifying the diagnosis of lowgrade dysplasia and its low rate of progression to cancer, invasive therapies generally are not recommended for patients with low-grade dysplasia in Barrett’s esophagus. For patients found to have high-grade dysplasia, there are several proposed management options, as discussed following. All of these options are associated with substantial risks and uncertain benefits.

Esophagectomy

Esophagectomy is the most definitive and most hazardous of the treatments for dysplasia in Barrett’s esophagus. The average hospital stay for open esophagectomy is approximately two weeks, and 30% to 50% of patients develop at least one serious postoperative complication such as pneumonia, myocardial infarction, and wound infection.43 A number of series describe high operative mortality rates that in some series can exceed 20%. In addition to death and short-term complications, esophagectomy can be accompanied by substantial long-term morbidity including profound weight loss and dysphagia. Interest in endoscopic therapies for dysplasia has been driven largely by the perception that esophagectomy has unacceptably high rates of mortality and morbidity. However, recent data suggest that for patients with Barrett’s esophagus those rates have been exaggerated. For example, mortality rates for esophagectomy are inversely related to the frequency with which the operation is performed. In a study of data from the Dutch National Medical Registry, the mortality rates for esophagectomy were 12.1%, 7.5%, and 4.9% at centers performing 1 to 10, 11 to 20, and more than 50 esophagectomies per year, respectively.44 Furthermore, estimates of mortality rates for esophagectomy generally have been based on series of patients with esophageal cancer who are often older adults and debilitated. Mortality rates for younger and otherwise healthy patients with dysplasia in Barrett’s esophagus may be substantially lower, especially when the operation is performed by experienced surgeons in a high-volume center. Finally, published data do not confirm the perception that most patients have an unacceptable quality of life after esophagectomy. In a study in which quality of life questionnaires were administered to 199 patients before and after esophagectomy, for example, quality of life was found to decline substantially immediately after the operation, but to return to baseline values within 2 years.45 Thus, the esophagectomy option still warrants serious consideration, especially for young and fit patients.

Endoscopic Therapies (see also Chapter 46) There are two general types of endoscopic therapies available for the treatment of Barrett’s esophagus: (1) endoscopic ablative therapy, which uses thermal energy (e.g., delivered by laser, electrocoagulation, argon plasma coagulation, the HALO360 System, BÂRRX Medical, Sunnyvale, Calif., cold nitrogen gas) or photochemical energy (photodynamic therapy) to ablate the Barrett’s epithelium; and (2) endoscopic mucosal resection (EMR), in which a diathermic snare or endoscopic knife is used to remove a segment of Barrett’s epithelium, usually down to the submucosa. After these endoscopic treatments, patients are given potent antireflux therapy (usually PPIs) so that the ablated mucosa can heal with the regrowth of normal esophageal squamous epithelium rather than with the regeneration of more Barrett’s epithelium. The ablative therapies destroy metaplastic tissue, but do not provide a pathology specimen by which to judge the depth of neoplastic invasion and the complete-

ness of the ablation. In contrast, EMR provides large tissue specimens that can be examined by the pathologist to determine the character and extent of the mucosal abnormality and, for neoplastic lesions, the depth of involvement and the adequacy of resection. When reviewing literature on endoscopic therapies for dysplasia, it is important to consider the issue of follow-up duration. Patients treated for carcinomas traditionally have been deemed cured if they have no evidence of the cancer at five years, because it is assumed that any cancer stem cells that survived the treatment would have become clinically manifest within that time. It often takes considerably longer than five years for dysplasia to progress to invasive cancer, however. Five years after the treatment of dysplasia, therefore, it is not appropriate to conclude that a patient who has not developed cancer has been “cured.” Unfortunately, many studies on endoscopic treatments for dysplasia in Barrett’s esophagus include considerably less than five years of follow-up, a factor that severely limits the conclusions that can be drawn regarding efficacy.

Endoscopic Ablative Therapies

(see also Chapter 46 and Fig. 46-14) An ideal ablative technique would inflict an injury deep enough to destroy all of the abnormal epithelium, but not so deep as to cause serious complications like esophageal hemorrhage, perforation, and stricture formation. So far, none of the ablative therapies has achieved this ideal, and all have been associated with serious complications. In addition, the procedures often leave behind residual foci of metaplastic epithelium.46 Partially ablated Barrett’s epithelium can heal with an overlying layer of squamous epithelium that “buries” metaplastic tissue (with its neoplastic potential) and hides it from the endoscopist. Even after apparent complete ablation, furthermore, Barrett’s metaplasia may recur over time. To date, photodynamic therapy (PDT) has been the most extensively studied of the endoscopic ablative treatments for dysplasia in Barrett’s esophagus (see also Chapter 46 and Fig. 46-14). For PDT, patients are given a systemic dose of a light-activated chemical that is taken up by the esophageal cells. The esophagus is then irradiated using a low-power laser that activates the chemical, which transfers that acquired energy to molecular oxygen. This results in the formation of singlet oxygen, a toxic molecule that destroys the abnormal cells and their vasculature. In a multicenter randomized trial of PDT using porfimer sodium for high-grade dysplasia in Barrett’s esophagus, 138 patients were treated with PDT plus omeprazole 20 mg twice daily, and 70 received omeprazole 20 mg twice daily alone.47,48 No dysplasia was found on repeat endoscopy with biopsy in 77% of the patients treated with PDT as compared with 39% of the patients who received omeprazole alone (P < 0.0001). During up to 5 years of follow-up, 15% of the PDT patients developed cancer, compared with 29% of those treated with omeprazole alone (P = 0.027). There was no procedure-related mortality, but 69% of the patients who received PDT developed photosensitivity reactions and 36% developed esophageal strictures. Although this study documents the superiority of PDT over PPI alone for eradicating dysplasia and preventing cancer in Barrett’s esophagus, the frequency of serious complications is disconcerting, as is the fact that 15% of the patients who received PDT developed cancer nevertheless. There has been much recent interest in the HALO360 system (BÂRRX Medical, Sunnyvale, Calif), which uses a balloon-based array of closely spaced electrodes to deliver

731

732

Section V  Esophagus radiofrequency energy to ablate the esophageal mucosa. This system was designed with the intent of inflicting a uniform, circumferential thermal injury whose depth is controlled by a generator, which can vary the power, density, and duration of the energy applied. There is also a smaller, endoscope-mounted radiofrequency catheter ablation device (the HALO90 ablation catheter) that can be used for the focal ablation of metaplasia that remains behind after treatment with the HALO360 system. In an uncontrolled study, 142 patients with high-grade dysplasia in Barrett’s esophagus were treated with the HALO360 system.49 There were no serious adverse events reported during 229 total ablation sessions, although 1 patient was found to have developed an asymptomatic esophageal stricture on follow-up endoscopic examination. At least one postablation endoscopy was performed during a median follow-up period of 12 months in 92 patients, only 9 of whom (10%) had high-grade dysplasia found in followup esophageal biopsy specimens. However, persistent lowgrade dysplasia was found in another 9 patients (10%), and 42 patients (46%) had residual foci of nondysplastic intestinal metaplasia in the esophagus. The preliminary results of a randomized sham-controlled trial of radiofrequency ablation for patients with dysplasia in Barrett’s esophagus have been presented in abstract form.50 The trial included 64 patients with low-grade dysplasia and 63 with high-grade dysplasia who were randomized to receive either radiofrequency ablation with the HALO360 system or sham ablation. Twelve months after the procedure, no dysplasia was found in esophageal biopsy specimens for 80% of the HALO-treated patients who had high-grade dysplasia at baseline, and for 11% of the patients who received sham treatment (P < 0.001). Among the patients with low-grade dysplasia at baseline, 90% of the HALO-treated patients had no dysplasia at 12 months, compared with 37% of the sham-treated patients (P < 0.001). Complications were few and easily managed. Five patients developed esophageal strictures that resolved with dilation. One patient experienced upper gastrointestinal bleeding, and 2 developed chest pain following the procedure that resulted in overnight hospitalizations. These reports document the feasibility of eradicating neoplastic Barrett’s epithelium with endoscopic ablation, but none has established the long-term benefit of the techniques for cancer prevention. Endoscopic ablation is expensive, and all of the treatments entail inconvenience and risks. Without histologic examination of the resected esophagus or durations of follow-up well beyond five years, it is not yet possible to verify claims that dysplasia and cancer indeed are “eliminated” by endoscopic ablation.

Endoscopic Mucosal Resection (see also Chapter 46) EMR commonly is performed using a “suck and cut” method in which the endoscopist elevates the dysplastic area by injecting fluid into the submucosa, after which the elevated mucosa is suctioned into a cap that fits over the tip of the endoscope.51 A polypectomy snare is then deployed around the suctioned area to remove it. A recent variation on this technique is the “band and snare” method that uses a ligating device, similar to that used for endoscopic variceal ligation, which deploys elastic bands around the suctioned mucosal segment without the requirement for prior submucosal fluid injection.52 The banded segment is removed using a polypectomy snare (see Fig. 46-11). Available reports on limited (noncircumferential) EMR describe few serious complications and virtually no procedure-related mortality. However, esophageal stricturing occurs frequently if EMR is used to remove the entire

circumferential extent of Barrett’s epithelium in a single endoscopic session.53 If the EMR specimen shows that there is no submucosal invasion and the margins of the specimen are free of neoplastic cells, then the patient may be cured and an esophagectomy is unlikely to show residual tumor.54 However, limited data suggest that a single cap-assisted EMR leaves neoplastic cells behind in the large majority of cases. The long-term data that are available on the efficacy of EMR are limited but impressive (see also Chapter 46). Ell and colleagues performed EMR on 100 patients with early adenocarcinomas in Barrett’s esophagus (tumor diameter <20 mm, well-differentiated histology, no invasion of lymphatics or blood vessels, and no evidence of metastases, submucosal invasion, or lymph node involvement).55 There were no serious complications, and the calculated five-year survival rate was an extraordinary 98%. However, recurrent or metachronous cancers were found in 11% of the patients during a mean follow-up period of 37 months. The recurrent tumors were treated successfully with more endoscopic therapy, but this high recurrence rate shows that EMR often leaves behind cells with neoplastic potential. A study from the Mayo Clinic compared long-term survivals in patients with high-grade dysplasia who were treated either with esophagectomy or with a combination of EMR and PDT.56 There was no statistically significant difference in survival for patients treated with either of the therapies, even though 6.2% of the patients treated with PDT and EMR were found to have a metachronous esophageal cancer during the follow-up period. Another recent report describes the long-term results of endoscopic therapies in 349 patients who had high-grade dysplasia or mucosal adenocarcinoma in Barrett’s esophagus.57 Endoscopic treatments included EMR alone for 279 patients, PDT alone for 55, EMR and PDT combined for 13, and argon plasma coagulation alone for 2 patients. Serious complications of endoscopic therapy occurred in 5% of cases (important bleeding in 2 patients, esophageal stricture in 15 patients). During a mean follow-up of 64 months, a complete remission (defined as complete elimination of the neoplastic lesion and at least one follow-up endoscopy showing no neoplasia) was achieved in 97%. However, metachronous neoplasms were found during the follow-up period in 21%. The calculated five-year survival rate was 84%, and none of the deaths were from esophageal cancer. Available data show that EMR for dysplasia and early cancers in Barrett’s esophagus is safe in experienced hands, and five-year survival rates are excellent. However, most reports have come from only a handful of highly specialized centers, and it is not clear that those results can be duplicated in community practice settings. Furthermore, recurrent neoplasms develop frequently after endoscopic therapy, especially if the residual Barrett’s epithelium is not eradicated, mandating close surveillance.

RECOMMENDATIONS

No management strategy for patients with Barrett’s esophagus has been verified by studies demonstrating that the strategy prolongs life by preventing deaths from esophageal cancer. The bulk of the indirect evidence available suggests that acid suppression with PPIs may reduce the cancer risk in Barrett’s esophagus, and we recommend that Barrett’s patients should be treated with a PPI using whatever dose is needed to control GERD symptoms and to maintain the healing of esophagitis. We do not routinely recommend esophageal pH monitoring to verify that esophageal acid exposure has been normalized by PPI therapy. We also feel

Chapter 44  Barrett’s Esophagus that antireflux surgery should not be prescribed solely as a cancer preventive strategy. Finally, we feel that it is appropriate to consider the prescription of low-dose aspirin for its cancer-preventive and cardioprotective effects, especially for patients who are not at high risk for complications of NSAID treatment. Because these patients should already be taking a PPI, the risks of aspirin causing serious gastrointestinal toxicity in average risk individuals should be minimal. The management strategy that has been endorsed by the American College of Gastroenterology is arguably the most complete and widely followed of the published guidelines to date for the management of patients with Barrett’s esophagus.58 Their guidelines are as follows: • Patients with Barrett’s esophagus should have regular surveillance endoscopy to obtain esophageal biopsy specimens. GERD should be treated prior to surveillance to minimize confusion in the interpretation of dysplasia caused by inflammation. • For patients who have had two consecutive endoscopies that show no dysplasia, surveillance endoscopy is recommended at an interval of every three years. • If dysplasia is noted, another endoscopy should be performed with extensive biopsy sampling (especially from areas with mucosal irregularity) to look for invasive cancer, and the histology slides should be interpreted by an expert pathologist. • For patients with verified low-grade dysplasia after extensive biopsy sampling, yearly surveillance endoscopy is recommended. • For patients with verified high-grade dysplasia, treatment should be individualized. If there is accompanying mucosal irregularity, the area should be evaluated by EMR, primarily to determine if there is submucosal invasion (for more details, see Chapter 46). Individualization of treatments for high-grade dysplasia is based initially on considerations of age, comorbidities, and life expectancy. For example, endoscopic therapy (ablative or EMR), or even no therapy, may be preferable to esophagectomy for an older and infirm patient who has a very limited life expectancy. The extent of the Barrett’s metaplasia is also an important factor when considering endoscopic therapy. Is there short-segment Barrett’s esophagus with a single focus of dysplasia that can easily be removed endoscopically, or is there a long-segment of Barrett’s epithelium with multiple neoplastic areas that make endoscopic treatments difficult and less effective? In the latter situation, a young and otherwise healthy patient may be better served by esophagectomy than by endoscopic therapy. It is also important to consider the patient’s preferences. Is the patient willing to accept the need for long-term endoscopic surveillance and the possibility of recurrence that accompanies the endoscopic treatments for neoplasia in Barrett’s esophagus? Available data suggest that following EMR, residual, nonneoplastic Barrett’s epithelium should be eradicated to prevent the recurrence of neoplasia. It is not yet clear whether that eradication should be effected by extended EMR or by endoscopic ablation. If all of the Barrett’s epithelium is to be removed by EMR, it seems prudent not to perform circumferential EMR in a single endoscopic session because this practice has been associated with esophageal stricture formation. For endoscopic treatment of long-

segment Barrett’s esophagus, the combination of using EMR for the initial staging or treatment of dysplasia along with an ablation procedure to eradicate the remaining Barrett’s epithelium seems most reasonable. Ablation should be delayed for approximately two months (the ideal interval is not known) after the initial EMR to allow the mucosal wound to heal. The preferred ablation procedure is disputed. PDT with porfimer sodium is the most extensively studied of the ablation techniques to date, but this procedure is associated with substantial inconvenience and frequent serious side effects. Preliminary data on radiofrequency ablation with the HALO system suggest that this technique may have similar efficacy to PDT but with less patient inconvenience and fewer side effects. Far more data are needed before dogmatic recommendations can be made regarding the choice of ablation procedure for the eradication of Barrett’s epithelium.

KEY REFERENCES

Ell C, May A, Pech O, et al. Curative endoscopic resection of early esophageal adenocarcinomas (Barrett’s cancer). Gastrointest Endosc 2007; 65:3-10. (Ref 55.) Iijima K, Henry E, Moriya A, et al. Dietary nitrate generates potentially mutagenic concentrations of nitric oxide at the gastroesophageal junction. Gastroenterology 2002; 122:1248-57. (Ref 22.) Lagergren J, Bergstrom R, Lindgren A, Nyren O. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999; 340:825-31. (Ref 14.) Overholt BF, Wang KK, Burdick JS, et al on behalf of the International Photodynamic Group for High-Grade Dysplasia in Barrett’s Esoph­ agus. Five-year efficacy and safety of photodynamic therapy with Photofrin in Barrett’s high-grade dysplasia. Gastrointest Endosc 2007; 66:460-8. (Ref 48.) Pech O, Behrens A, May A, et al. Long-term results and risk factor analysis for recurrence after curative endoscopic therapy in 349 patients with high-grade intraepithelial neoplasia and mucosal adenocarcinoma in Barrett’s oesophagus. Gut 2008; 57:1200-6. (Ref 57.) Pohl H, Welch HG. The role of overdiagnosis and reclassification in the marked increase of esophageal adenocarcinoma incidence. J Natl Cancer Inst 2005; 97:142-6. (Ref 3.) Prasad GA, Wang KK, Buttar NS, et al. Long-term survival following endoscopic and surgical treatment of high-grade dysplasia in Barrett’s esophagus. Gastroenterology 2007; 132:1226-33. (Ref 56.) Rex DK, Cummings OW, Shaw M, et al. Screening for Barrett’s esophagus in colonoscopy patients with and without heartburn. Gastroenterology 2003; 125:1670-7. (Ref 11.) Shaheen NJ, Crosby MA, Bozymski EM, Sandler RS. Is there publication bias in the reporting of cancer risk in Barrett’s esophagus? Gastroenterology 2000; 119:333-8. (Ref 13.) Souza RF, Krishnan K, Spechler SJ. Acid, bile and CDX: The ABCs of making Barrett’s metaplasia. Am J Physiol Gastrointest Liver Physiol 2008; 295:G211-18. (Ref 24.) Souza RF, Morales CP, Spechler SJ. Review article: A conceptual approach to understanding the molecular mechanisms of cancer development in Barrett’s oesophagus. Aliment Pharmacol Ther 2001; 15:1087-100. (Ref 28.) Souza RF, Spechler SJ. Concepts in the prevention of adenocarcinoma of the distal esophagus and proximal stomach. CA Cancer J Clin 2005; 55:334-51. (Ref 18.) Spechler SJ. Barrett’s esophagus. N Engl J Med 2002; 346:836-42. (Ref 1.) Spechler SJ. Dysplasia in Barrett’s esophagus: Limitations of current management strategies. Am J Gastroenterol 2005; 100:927-35. (Ref 30.) Wang KK, Sampliner RE; Practice Parameters Committee of the American College of Gastroenterology. Updated guidelines 2008 for the diagnosis, surveillance and therapy of Barrett’s esophagus. Am J Gastroenterol 2008; 103:788-97. (Ref 58.) Full references for this chapter can be found on www.expertconsult.com.

733

CHAPTE R

45 Esophageal Disorders Caused by Medications, Trauma, and Infection David A. Katzka

CHAPTER OUTLINE Medication-Induced Esophageal Injury  735 Mechanisms  735 Clinical Features and Diagnosis  736 Prevention, Treatment, and Clinical Course  736 Specific Medications  737 Esophageal Injury from Nasogastric and Other Nonendoscopic Tubes  739 Esophageal Injury from Penetrating or Blunt Trauma  739 Esophageal Tears and Hematomas  740 Mallory-Weiss Syndrome  740 Boerhaave’s Syndrome  740 Spontaneous Esophageal Hematoma  741

MEDICATION-INDUCED ESOPHAGEAL INJURY Medication-induced esophageal injury may occur at any age and with a variety of commonly used medications. Nevertheless, medication-induced esophageal injury is most likely underdiagnosed in clinical practice for several reasons. First, initial consideration of common and more serious problems such as an acute coronary syndrome or pulmonary embolism may occur due to the severe chest pain, often pleuritic in nature, that may be associated with pill-induced esophagitis. Second, patients may be assumed to be having a severe episode of acid reflux, a far more common condition than a medication-induced esophageal ulceration. Third, several of the medications that may cause medication-induced esophagitis are over-the-counter medications (e.g., nonsteroidal anti-inflammatory drugs [NSAIDs]) or may have been taken safely for years (e.g., tetracycline) without injury and therefore not considered by patients to be a possible contributor to their symptoms. Fourth, because it is not routinely reported or recognized, medication-induced esophageal injury is often considered an uncommon entity.1,2 As a result, medication-induced esophageal injury often may not be considered. This can be problematic because recognition of this entity might result in failure to discontinue the offending agent or to give the patient proper instruction in avoiding future injury. It may also lead to extensive and erroneous evaluation and treatment of other conditions. This chapter provides a detailed overview of medication-induced esophageal injury, with

Esophageal Infections in the Immunocompetent Host  741 Candida albicans  741 Herpes simplex  742 Human papillomavirus  742 Trypanosoma cruzi  743 Mycobacterium tuberculosis  743 Treponema pallidum  743 Other Infections That Rarely Involve the Esophagus  743 Acute Esophageal Necrosis  743

particular attention to suspecting this entity both by its symptoms and by the medications that are potentially culpable.

MECHANISMS

Medications may cause esophageal injury through several mechanisms. These can initially be divided into those that cause direct injury to esophageal mucosa because of their caustic nature or by facilitation of injury through another mechanism such as induction of acid reflux (e.g., calcium channel antagonists). When medications directly damage the esophageal mucosa, it may be through one of four known mechanisms: (1) production of a caustic acidic solution (e.g., ascorbic acid and ferrous sulfate); (2) production of a caustic alkaline solution (e.g., alendronate); (3) creation of a hyperosmolar solution in contact with esophageal mucosa (e.g., potassium chloride); and (4) direct drug toxicity to the esophageal mucosa (e.g., tetracycline). For many medications, the mechanism of esophageal injury does not fall into any of these known categories. Other factors may influence the toxicity of the pill, particularly contact time, pills coated with gelatinous material,3 sustained release formulations, and a wax matrix form of the drug.4 Cellulose fiber and guar gum pills may swell and lodge in the esophagus, causing complete obstruction because of their water-absorbing capacity. It is commonly assumed in medication-induced esophagitis that injury is predisposed by an anatomic or motility disorder of the esophagus or that the medication was taken incorrectly, in either case allowing for prolonged exposure

735

736

Section V  Esophagus of the medication to esophageal mucosa. For example, studies have shown that patients with left atrial enlargement,5 esophageal strictures,6 esophageal dysmotility,7 and esophageal diverticula8 (either Zenker’s or epiphrenic diverticula) have greater risk of pill injury. Similarly, in the patient with normal esophageal function, the site of druginduced injury most commonly occurs where there are areas of normal hypomotility or extrinsic compression such as in the trough zone of the esophagus (where the smooth and skeletal muscle overlap) or at the level of the aortic or left bronchial impression on the esophagus.9,10 These locations of relative stasis allow for a pill, when taken incorrectly, to cause injury. However, any part of the esophagus may be involved. Methods of taking a medication incorrectly that predispose to injury include ingesting a pill without enough water or assuming a recumbent position or sleeping immediately after pill ingestion, or both. The latter two factors are particularly problematic, by eliminating the help of gravity in esophageal transit and by reducing saliva production and frequent swallowing, which occur normally while awake. Importantly, however, many, if not most, patients who suffer pill-induced esophageal injury presumably have normal esophageal function and do not neces­ sarily ingest their medication in a faulty manner. That pill-induced esophageal injury can occur under “normal” conditions is supported by data demonstrating prolonged radiographic retention of capsules in the esophagus by normal subjects even when taken with water in the upright position.3,11

CLINICAL FEATURES AND DIAGNOSIS

Patients typically note an acute onset of chest pain, which may radiate over the central chest and to the back. The pain is commonly accentuated with inspiration and may be accompanied by severe odynophagia, even to small sips of liquids. Some patients may complain of a severe acute onset or heartburn-type symptoms. This set of symptoms associated with a potentially injurious medication taken incorrectly (particularly just before bedtime without enough

A

water) strongly suggests the diagnosis. If objective confirmation of the diagnosis is necessary, endoscopy or radiography can be used. Endoscopy is felt to be more sensitive, although trials comparing the two have not been performed. Findings range from discrete ulcers to diffuse severe esophagitis with pseudomembranes, as may be seen with bisphosphonates12 or with sodium polystyrene sulfonate suspension (Kayexalate), in which the appearance may mimic candidal esophagitis.13 Occasionally, severe inflammatory reactions causing stenoses and tumor-like appearances may occur.14,15 Similar findings may be seen radiographically, particularly when double-contrast radiography is used.16,17 The range of findings described on esophagography may also include solitary or multiple ulcers; small or large ulcerations; ulcers with punctate, ovoid, linear, serpiginous, or stellate collections of barium; confluent ulcers; or areas of normalappearing mucosa separating ulcers (Fig. 45-1).9 The occurrence of multiple esophageal septa has also been described.18 Rarely, severe complications of medication-induced injury may occur. These may include esophagorespiratory fistula, esophageal perforation, hemorrhage secondary to ulceration, and chronic stricture formation.

PREVENTION, TREATMENT, AND CLINICAL COURSE

No specific treatments have been shown to be beneficial in altering the course of medication-induced injury. Treatment is aimed at symptom control, prevention of superimposed injury from acid reflux, maintenance of adequate hydration, and removal of the offending medication. Symptom control may be achieved topically by local anesthetics such as viscous lidocaine solution. Occasionally, narcotics are necessary. Prevention of superimposed reflux is best achieved with a twice-daily proton pump inhibitor, although no data clearly suggest that prevention of acid reflux hastens symptomatic or pathologic improvement of pill-induced injury. For patients who have severe odynophagia prohibiting adequate oral intake, intravenous hydration may be necessary

B

Figure 45-1.  A, Esophageal ulceration secondary to tetracycline, with arrow pointing to area of ulcerations demonstrated by barium esophagography. B, Endoscopic image of tetracycline-induced esophageal burn. (A, Courtesy Dr. Marc Levine, Philadelphia, Penn.)

Chapter 45  Esophageal Disorders Caused by Medications, Trauma, and Infection for a few days. Removal of the cause of injury is self-evident, although this is not always easily achieved. This is particularly true in clinical situations in which there may not be an adequate substitute such as in aspirin prophylaxis for cardiovascular disease, bisphosphonates for severe osteoporosis, or high-dose NSAIDs for pain from chronic inflammatory arthritides. No data address the question of whether rechallenge with a pill that induced prior esophagitis poses higher risk of recurrent injury if the pill is taken with better caution, with the possible exception of bisphosphonates. It is also unclear if patients with a theoretical underlying risk (e.g., esophageal dysmotility) have even greater risk of esophagitis with rechallenge. In the absence of stricture formation or catastrophic presentation, most patients have clinical resolution of symptoms within two to three weeks, and radiographic resolution has been described in 7 to 10 days.16 Because no treatment has been proven effective, it is hoped that proper administration of potentially injurious medications will help avoid occurrence of esophageal injury. On the basis of the sometimes normally slow transit of medications through the esophagus, particularly for gelatin capsules and larger tablets,3 the following recommendations are made: (1) medications should be swallowed with at least 8 ounces of a clear liquid; (2) patients should remain upright for at least 30 minutes following ingestion of the medication; (3) in patients with potential underlying increased risk for pill-induced injury (e.g., inability to follow the previous instructions, poor esophageal motility, anatomic compromise of the esophageal lumen), one should search for alternative safer medications or carefully weigh the risks and benefits of this medication against the disease for which this medication is necessary.

Table 45-1  Medications Commonly Associated with Esophagitis or Esophageal Injury Antibiotics Clindamycin Doxycycline Penicillin Rifampin Tetracycline Antiviral Agents Nelfinavir Zalcitabine Zidovudine Bisphosphonates Alendronate Etidronate Pamidronate Chemotherapeutic Agents Bleomycin Cytarabine Dactinomycin Daunorubicin 5-Fluorouracil Methotrexate Vincristine Nonsteroidal Anti-inflammatory Drugs Aspirin Ibuprofen Naproxen Other Medications Ascorbic acid Ferrous sulfate Lansoprazole Multivitamins Potassium chloride Quinidine Theophylline

SPECIFIC MEDICATIONS

Several broad categories of medication types cause esophageal injury. These include antibiotics, antivirals, NSAIDs, specific antiarrhythmic drugs, vitamins, and miscellaneous isolated drugs from varied categories.

Antibiotics (Table 45-1)

Tetracycline, doxycycline, and their derivatives are by far the most common causes of pill-induced esophagitis, with almost as many cases reported as all other cases combined.10 Its commonality of injury may be more a reflection of how frequently the drug is used than a strong propensity of tetracycline to produce such injury. This relatively low incidence of esophageal ulceration from tetracycline for all users is suggested by a lack of any cases of esophageal injury seen in a recent survey of 491 Gulf War veterans treated with doxycycline.2 The mechanism of injury is felt to be corrosive damage because tetracycline dissolved in water produces a solution with a very low pH.11 Symptoms typically last several days to several weeks. Ulcerations may vary in appearance but are typically small and superficial, located in the mid-esophagus just above the aortic arch or left mainstem bronchus9 with a burn-like appearance (see Fig. 45-1). Stricture formation is uncommon. Injury from other antibiotics is uncommon and mostly documented in case reports. These include clindamycin,19,20 rifampin,21 and penicillin,22 but the incidence is still exceedingly low given their common use. If a history is compatible with pill-induced esophageal injury, any antibiotic currently being used should be considered a possible culprit, although rare. Antiviral agents, particularly those used for treatment of human immunodeficiency virus, also have been reported to

cause medication-induced esophageal injury. These include zalcitabine,23 zidovudine,24 and nelfinavir.25

Bisphosphonates

The most rapidly emerging category of medication-induced esophagitis over the past decade has been those injuries secondary to bisphosphonates used to treat osteoporosis. This class of medications has in fact become the most prevalent cause of medication-induced esophagitis. To date, injury has been reported mostly with alendronate12,26-32 but also with etidronate33 and pamidronate.34 Although the overall incidence of injury is probably small (fewer than 100 cases reported)10 when considering the millions of patients using the medication, injury can be serious and even fatal. Unfortunately reflux-type symptoms are common and can be difficult to distinguish from medication-induced mucosal injury. Risedronate has low potential for causing esophageal injury, if at all.35 Part of this might be explained by the rapid esophageal transit and subsequently minimal contact time of the drug with esophageal mucosa.36 In one study prospectively following 255 patients treated with risedronate and undergoing endoscopy 8 and 15 days later, no patients developed esophageal ulceration. This study also underscored the overall safety of bisphosphonates in general in that only 3 of 260 patients receiving alendronate developed esophageal ulceration.37 Diagnosis is best made endoscopically, with marked exudates and inflammation seen. Biopsies show an intense inflammatory exudate and granulation tissue that may contain polarizable crystals and multinucleated giant cells.38

737

738

Section V  Esophagus Stricture formation occurs in up to one third of patients,10 and life-threatening hemorrhage30 and esophageal perforation28 have been reported. Patients who sustain injury are described commonly to take the bisphosphonate not in accordance with directions (i.e., in the upright position with at least 8 ounces of beverage, remaining upright for at least 30 minutes). Still, as with other pill-induced esophagitides, patients taking the medication correctly may sustain esophageal injury. One question frequently answered anecdotally, but not clearly addressed scientifically, is whether patients with a history of gastroesophageal reflux disease (GERD) should avoid bisphosphonates. Furthermore, if GERD is a risk factor, it is unclear what degree of reflux constitutes risk. The decision should weigh the severity of osteoporosis and risk of fracture against the risk of esophagitis. Patients with GERD that predisposes to stasis such as those with stricture or severe ineffective esophageal motility should be particularly cautious.

Nonsteroidal Anti-inflammatory Drugs

NSAIDs are another common cause of pill-induced esophageal injury. Similar to the other common causes or medication-induced esophageal injury, they occur in a small fraction of all NSAID users. Aspirin, naproxen, indomethacin, and ibuprofen account for the majority of cases,10 but most other NSAIDs have been reported to cause esophageal injury in case reports. Not surprisingly, hemorrhage, which may be severe,39 is a common complication of these esophageal ulcers, especially when compared with other medication causes of esophagitis. Bronchoesophageal fistula has also been reported.40 Notably, it is over-the-counter use of NSAIDs that is most commonly associated with injury,41 in keeping with their more commonly used venue. In a study of 1122 patients hospitalized for gastrointestinal bleeding, any dose of aspirin including a low dose was associated with increased risk of developing esophagitis.42 Other studies have also identified NSAIDs in general as a risk factor for erosive esophagitis.43 Whether the esophagitis in these studies is all directly due to these medications or whether they act synergistically with reflux-induced injury is unclear, although one study has suggested that aspirin makes the esophageal mucosa more sensitive to acid and pepsin.44

Other Medications Commonly Associated with Pill-Induced Injury

Potassium chloride (KCl) pills have been associated with esophageal injury. Injury can be severe, as documented by reports of esophageal stricture formation45,46 or of perforation into the left atrium,47 bronchial artery,48 or mediastinum.49 Patients who sustain esophageal injury from KCl commonly report associated conditions such as cardiac, including left atrial, enlargement, or prior cardiac surgery.50-52 Whether these processes truly predispose to pill stasis and injury because of extrinsic esophageal compression by the heart is unclear, because patients using KCl have a high prevalence of cardiac disease. Quinidine is another cardiac medication with the potential for severe esophagitis.15 Endoscopically, quinidine may be associated with anything from mild ulceration to a marked inflammatory response with edema suggesting carcinoma.14,15 Ferrous sulfate,53 theophylline,54,55 oral contraceptives,56 ascorbic acid,22 and multivitamins57 have caused esophageal ulceration. Numerous other medications have been reported to cause esophageal ulceration in single case reports. Examples include sildenafil,58 pheny­ toin,11 warfarin,59 glyburide,60 lansoprazole,61 valproic

acid,62 chlorazepate,63 captopril,64 foscarnet,65 and throat lozenges.66,67

Chemotherapy-Induced Esophagitis

Dactinomycin, bleomycin, cytarabine, daunorubicin, 5fluorouracil, methotrexate, vincristine, and chemotherapy regimens used in hematopoietic stem cell transplantation may cause severe odynophagia as a result of oropharyngeal mucositis, a process that can also involve the esophageal mucosa.68 Esophageal damage is unusual in the absence of oral changes. Although mucositis is self-limited in most cases, some patients have oral and esophageal damage that persists for weeks to months. Chemotherapy that is given months after thoracic irradiation to the esophagus, particularly doxorubicin, may cause a “recall” esophagitis. Vinca alkaloid drugs are neurotoxic, and dysphagia may complicate vincristine therapy.69

Esophageal Injury from Variceal Sclerotherapy

For many years variceal sclerotherapy was the mainstay of therapy for endoscopic control of esophageal variceal bleeding. Although it is still an accepted form of therapy, it has been largely replaced by several other methods including intravenous administration of octreotide, variceal banding, and transvenous intrahepatic portosystemic shunts. Nevertheless, its continued use by some physicians, as well as the occurrence of complications that may persist for several years, compel the gastroenterologist to recognize its various forms of potential esophageal injury. Complications from variceal sclerotherapy can be divided into two main categories: gross structural injury and esophageal motility change. There is a wide range of gross injury from variceal injection. Injection of sclerosant into and around varices causes necrosis of esophageal tissues and mucosal ulceration; the risk is related to the number of injections and the amount of sclerosant. Small ulcers appear within the first few days after sclerotherapy in virtually all patients; larger ulcers develop in roughly one half of patients. Other complications include intramural esophageal hematoma,70 strictures,71 and perforation.72 Strictures occur in approximately 15% of patients undergoing sclerotherapy71,73,74 and are usually amenable to Savary or balloon dilation. Unusual manifestations of sclerotherapy with deep needle penetration include pericarditis, esophageal-pleural fistula, and tracheal obstruction due to compression by an intramural hematoma.75,76 One case of squamous cell carcinoma of the esophagus was attributed to a course of variceal sclerotherapy five years earlier.77 Several studies have demonstrated abnormal esophageal motility after completed courses of sclerotherapy. These abnormalities may be related to wall injury or vagal dysfunction.78 Specific motility abnormalities include delay in esophageal transit and decreased amplitude and coordination of esophageal contractions.74,79 There is debate over whether these changes are reversible, with different studies demonstrating worsening74 or resolution80 of motility abnormalities over four weeks’ time. Whether these studies reflect the effects of irreversible fibrosis or reversible inflammatory neuropathy, respectively, is unclear. One potential consequence of motility dysfunction is the occurrence of pathologic gastroesophageal reflux, as documented by abnormal esophageal pH monitoring81 and by abnormal scintigraphy and barium studies after sclerotherapy.74 Other studies have also shown abnormal reflux following sclerotherapy that correlated with esophageal dysmotility, and this did not occur in patients undergoing band ligation.79 Furthermore, the amount of sclerosant injected paravariceally appears to correlate with increased acid reflux.81

Chapter 45  Esophageal Disorders Caused by Medications, Trauma, and Infection The only agent that has been shown effective in preventing postsclerotherapy strictures and in healing ulcers is sucralfate, either alone or in combination with antacids and cimetidine.82,83 Acid suppressive therapy alone, with either H2 receptor antagonists or proton pump inhibitors, has not been shown to be effective in preventing or healing postsclerotherapy ulcers or strictures.84,85

ESOPHAGEAL INJURY FROM NASOGASTRIC AND OTHER NONENDOSCOPIC TUBES Nasogastric tubes have long been recognized as a potential source of esophageal injury and stricture formation (Fig. 45-2). The putative mechanism is gastroesophageal reflux. In patients undergoing elective laparotomy, recent data have demonstrated an esophageal pH of less than 4 for nearly 9 of the first 24 hours compared with less than one half hour in a control group without tube placement.86 One study demonstrated an increase in acid exposure even in normal volunteers undergoing nasogastric tube placement.87 When strictures occur, they are characteristically long, narrow, and difficult to manage endoscopically. Whether general use of potent acid-suppressing therapies has decreased the incidence of these strictures is unknown. Respiratory luminal devices have also been reported as potential sources of esophageal trauma. Esophageal laceration with use of a Combitube,88 tracheoesophageal fistula with a cuffed tracheal tube,89 and esophageal perforation from a thoracostomy tube90 or transesophageal echocardiography probes91 have been reported. More recently, several authors have reported the occurrence of an atrial-esophageal fistula complicating cardiac radiofrequency ablation procedures.92-96 This serious and often fatal complication has been described to occur anywhere from 10 days to five weeks after ablation. The initial presentation includes fever and neurologic abnormalities, the latter as a result of air emboli to the brain from the esophagus through the fistula into the left heart. Laboratory studies may reveal leukocytosis and positive blood cultures.

A

B

Imaging studies may reveal gas bubbles in the left atrium. Although generally fatal due to sepsis and upper gastrointestinal (GI) hemorrhage, a recent report describes one patient who survived with surgical repair of the fistula,95 suggesting that prompt recognition may reduce mortality.

ESOPHAGEAL INJURY FROM PENETRATING OR BLUNT TRAUMA Noniatrogenic traumatic injury to the esophagus may occur through either penetrating or, less commonly, blunt injuries. Blunt trauma resulting in esophageal perforation is exceedingly rare; most cases have occurred in the cervical esophagus after motor vehicle accidents from the steering wheel97 or seat belt.98 Penetrating injuries to the esophagus are usually caused by gunshot or knife wounds, although cer­vical esophageal perforation secondary to cervical spine surgery has been well recognized.99 In general, injuries from penetrating wounds are divided into those of the cervical and lower esophagus. Perforation of the cervical esophagus may be diagnosed initially by the finding of extramural air on radiographic studies such as lateral views of the neck or computed tomography. Gastrografin contrast studies confirm the diagnosis, although this test is not always possible in patients with severe traumatic injuries. Although routine endoscopy is relatively contraindicated in these patients, intraoperative endoscopy may be a valuable diagnostic tool for the diagnosis of perforation.100 Cervical esophageal penetrating injuries are usually associated with concurrent tracheal, carotid, or spinal injury. One area of debate in management of these injuries is whether surgical exploration is necessary in all patients. The concern in waiting is the development of sepsis, airway compromise, or tracheoesophageal fistulae,101 estimated to occur in approximately 4% of penetrating esophageal wounds102 and particularly in those patients undergoing tracheostomy for tracheal damage. Another downside of watchful waiting is the contamination of a previously sterile field. This may eliminate the option of primary closure and

Figure 45-2.  A, Nasogastric tube–induced stricture demonstrated by barium esophagography. B, Endoscopic appearance of a tight nasogastric tube–induced stricture. (A, Courtesy Dr. Marc Levine, Philadelphia, Penn.; B, courtesy Dr. Gregory G. Ginsberg, Philadelphia, Penn.)

739

740

Section V  Esophagus necessitate a two-step procedure, first with performance of a diverting cervical esophagostomy before definitive repair. As a result, some investigators continue to recommend an aggressive multimodal surgical approach.99 In contrast, a recent study of 17 patients with cervical esophageal injury from knife or gunshot wounds suggested that conservative management with enteral feeding and antibiotics may allow for nonoperative healing.103 A consensus seems to be that in those patients with contained small luminal cervical per­ foration, without sepsis and without the need for surgical exploration for other injuries, a conservative approach may be tried.104 For penetrating trauma to the more distal esophagus, similar principles apply but with some important differences. First, although diagnosis is often made by finding extraesophageal air on either chest or computed tomography (or finding extravasation on contrast study of the esophagus), endoscopy may be performed, particularly for those unstable patients in whom contrast esophagography is not practical.105 Some investigators feel that endoscopy should be the diagnostic test of choice.106 Second, as opposed to a more contained perforation that occurs in the neck as dictated by its close tissue planes, perforation in the more distal esophagus may extend farther into the mediastinum and pleura. Also, there is a threat of coexistent injury to the aorta. Third, because of the segmental and often variable esophageal blood supply (particularly in the distal esophagus), simple closure of a perforation is not often adequate due to wound ischemia and consequent leakage.107 As a result, esophageal resection with esophagogastric anastomosis is often necessary in these patients.105 Fourth, because access to the esophagus through the mediastinum is so much more difficult than access through the neck, the consequences of perforation into the mediastinum, pleura, or aorta can be more devastating, and the surgery required for distal esophageal perforation may be much more extensive. As a result, the decision whether to operate is far more difficult. Despite these caveats, surgery is not only recommended in most patients107,108 but must be performed in a timely fashion because there is significantly higher morbidity and mortality when surgery is delayed beyond 1 to 12 hours.109 There may be a role for conservative management, with antibiotics and nasogastric tube placement bypassing the perforation, in only a select group of patients. Finally, although metallic stents have been used successfully for nonoperative management of other causes of esophageal perforation,110 their role in managing traumatic perforation of the esophagus has not been studied and should not be considered at this time.

ESOPHAGEAL TEARS AND HEMATOMAS MALLORY-WEISS SYNDROME (see also Chapter 19)

The Mallory-Weiss syndrome was originally described by Doctors Kenneth Mallory and Soma Weiss in 1929, who described patients with lacerations of the gastric cardia due to forceful vomiting.111 The laceration is felt to result from shearing forces on the gastroesophageal junction and proximal stomach as it herniates through the diaphragm because of high intra-abdominal pressures due to forceful vomiting.112,113 In accordance with Laplace’s law, this shearing force has its greatest effect when there is a hiatal hernia, thus exposing a relatively large-volume dilated sac to high wall tension. It is not surprising that the majority of patients who sustain a Mallory-Weiss tear have a hiatal hernia.114

Although most tears will occur within 2  cm of the gastroesophageal junction, the likelihood of a more distal tear in the proximal portion of the stomach is increased when a larger hiatal hernia is present. Any bodily action that results in an abrupt increase in intra-abdominal pressure and gastric herniation may cause a Mallory-Weiss tear. Such actions include forceful coughing, straining, retching during endoscopy, transesophageal echocardiography, and cardiopulmonary resuscitation.112,115-120 Other factors that predispose to tearing include alcohol and aspirin use.113,121-122 Most patients present with hematemesis, but some present with melena alone. Although a classic history includes vomiting or retching followed by hematemesis, up to a third of patients do not have an antecedent history of vomiting; hematemesis is their presenting symptom.121,122 Typically, one laceration is seen at the time of endoscopy, most commonly along the lesser curve of the cardia, although more than one tear may occur in up to 10% of patients.114 Bleeding is typically self-limited, but may be massive in up to 10% of patients114 and even fatal.123 Endoscopy is also important not for diagnosing a tear but for ruling out other upper gastrointestinal lesions that are found in more than a third of patients during the initial endoscopic evaluation. Such lesions include peptic ulcers, gastritis or gastropathy, erosive esophagitis, esophageal varices, and gastric outlet obstruction. Treatment for Mallory-Weiss tear has usually been supportive because of the self-limited nature of the bleed, along with attempts to reduce retching and vomiting. More recently, several methods of endoscopic therapy have been used. Injection of epinephrine and polidocanol has been shown to significantly reduce bleeding and transfusion requirement and to shorten the hospital stay.124 Endoscopic band ligation also has been shown to be efficacious125,126 and in one trial was equivalent to injection therapy.127 Endoscopic clip placement has also been suggested as a therapeutic alternative for controlling bleeding from MalloryWeiss tears.128 For patients with persistent bleeding despite endoscopic therapy, angiographic embolization through the left gastric artery may be used.129 The need for surgical intervention is rare.

BOERHAAVE’S SYNDROME

A more extreme version of an esophageal tear that occurs in response to an acute increase in intra-abdominal pressure and accentuation of the intragastric-to-intrathoracic pressure gradient is Boerhaave’s syndrome. In this syndrome, a transmural tear with perforation occurs. The perforation specifically occurs at the margin of the contact between “clasp” and oblique esophageal fibers.130 Similar to the Mallory-Weiss syndrome, preceding symptoms such as severe vomiting and retching, abdominal straining, blunt trauma, and coughing may precipitate this perforation.114 In addition to acute pressure changes at the gastroesophageal junction, some investigators have postulated that an abnormal esophageal mucosa may predispose to perforation. These conditions include reflux esophagitis,131 Barrett’s esophagitis with ulceration,132 infectious esophagitis,133,134 and eosinophilic esophagitis.135 The clinical presentation is often catastrophic with shock and sepsis due to a large esophageal perforation. Because of the acute presentation of severe chest pain, it is often confused with acute cardiac or pulmonary events, dissecting aortic aneurysm, or pancreatitis,136 often leading to a delay in diagnosis and greater morbidity and mortality. Diagnosis is suggested by subcutaneous emphysema with crepitus and radiographic findings of pneumomediastinum and a left pleural effusion (that may contain salivary amylase, errone-

Chapter 45  Esophageal Disorders Caused by Medications, Trauma, and Infection ously suggesting pancreatitis) or even a frank empyema. Perforation of the esophagus may be confirmed by esophageal contrast studies using Gastrografin. Management is surgical repair and drainage, although successful nonoperative treatment has been reported with placement of a selfexpandable plastic stent.137

SPONTANEOUS ESOPHAGEAL HEMATOMA

Spontaneous esophageal hematoma is a rare entity in which an abrupt bleed occurs between the mucosa and muscularis propria of the esophageal wall, often for a long length of the esophagus. The term spontaneous is somewhat of a misnomer in the sense that several underlying factors have been identified that may predispose to hematoma formation. These include use of aspirin,138,139 underlying coagulopathy or use of anticoagulant,140,141 abrupt increases in the intrabdominal-to-intrathoracic pressure gradient such as may occur with forceful vomiting, coughing or sneezing,142 and foreign body ingestion.143 Not all cases have an obvious predisposing factor, however, and do in essence present spontaneously.144 One third of patients classically present with a triad of retrosternal chest pain, dysphagia, and hematemesis and 50% present with at least two of these symptoms.144 As in Boerhaave’s syndrome, there is often a delay in diagnosis because of the symptomatic overlap with more common cardiopulmonary catastrophes.145 Inter­ estingly there may be a predisposition to this syndrome in middle-aged women, although this association is not uniform.146-148 Diagnosis can be made by several means. Computed tomography (CT) of the chest demonstrates a diffusely thickened esophagus and sometimes a “double barrel” appearance with obliteration of the esophageal lumen.149 Magnetic resonance imaging also may be an accurate means of making the diagnosis.150 Endoscopically, obliteration of the esophageal lumen is seen with visualization of a long, deep, friable, blue submucosal mass with or without a visible tear.143 Sometimes it may be difficult to distinguish hematoma from an esophageal malignancy.151 Conservative treatment is the mainstay of treatment, maintaining the patient without oral intake and monitoring hemodynamic status143; it usually takes up to several weeks to fully heal. Progress is monitored on repeated CT or endoscopy, usually at one-week intervals. The need for surgical intervention is rare.

ESOPHAGEAL INFECTIONS IN THE IMMUNOCOMPETENT HOST (Table 45-2) Esophageal infections are most common in immunocompromised patients such as those infected with human immunodeficiency virus (see Chapter 33) and those receiving chemotherapy or immunosuppressive therapies, particularly for hematologic malignancies or following organ transplantation (see Chapter 34). Nevertheless, there are some esophageal infections that occur in immunocompetent hosts. These include infections that (1) are more typically associated with immunodeficiency but are occasionally seen in patients with intact immune systems; (2) occur in patients with underlying esophageal diseases, particularly with those associated with prolonged stasis of luminal content; and (3) involve the esophagus because of a localized area of esophageal immune compromise such as with the use of inhaled topical steroids for respiratory disorders. The types of organisms found in these situations tend to be few in number, with Candida the dominant organism.

Table 45-2  Esophageal Infections in the Immunocompetent Host Typically associated with immunodeficiency Herpes simplex Candida albicans Mycobacterium tuberculosis Associated with esophageal stasis (e.g., achalasia, scleroderma) Candida albicans Associated with use of glucocorticoid inhalers Candida albicans Other esophageal infections Trypanosoma cruzi Treponema pallidum Human papillomavirus

CANDIDA ALBICANS

Candidal organisms are the most common esophageal infection in the immunocompetent host. Although several species of Candida have been implicated in esophageal infection, including Candida tropicalis, Candida albicans accounts for the vast majority. In one large series of 933 patients in India with dysphagia or odynophagia, 56 were found to have candidal esophagitis of varying severity.152 How many patients had clear motility disorders or Candida as a commensurate rather than a pathologic organism is not totally clear because Candida colonization of the esophagus in healthy ambulatory adults has a reported prevalence of approximately 20%.153 Although candidal esophagitis may occur rarely without a clear underlying mechanism, one should generally assume a predisposing condition, even in the immunocompetent host. The conditions that most predispose to candidal infection in the esophagus are those associated with severe stasis such as achalasia or scleroderma (Fig. 45-3). In achalasia, infection seems related to severity, with those patients who have long-standing disease with marked esophageal dilation most at risk. These infections can be very difficult to treat medically until effective achalasia therapy, and therefore drainage of the esophagus, is provided. Candida is seen less often in scleroderma with esophageal involvement than in achalasia but, similarly, is usually seen in those patients with esophageal dilation and poor peristalsis. One risk factor for candidal infection in scleroderma might be acid suppression, as suggested by one study of patients with systemic sclerosis, in which the prevalence of Candida esophagitis was 44% (21 of 48 patients) for those on no acid suppression, compared with 89% (16 of 18 patients) among those on potent acid suppressive therapy.154 Topical glucocorticoids (contained in inhalers for treatment of asthma) may lead to oropharyngeal and esophageal candidiasis in otherwise healthy adults.155 Likewise, candidal esophagitis has been described and must be considered in patients with eosinophilic esophagitis treated with swallowed flu­ ticasone.156,157 Other medical illnesses that predispose to fungal esophagitis albeit via impaired immune mechanisms include diabetes mellitus, adrenal insufficiency, alco­holism, and advanced age.158 Also a rare condition known as esophageal intramural pseudodiverticulosis of the esophagus may be associated with candidal infection.159 Diagnosis of candidal esophagitis can be made by its endoscopic appearance, with a characteristic white pseudomembranous or plaque-like appearance adherent to esophageal mucosa. Confirmation can be made by brushing the lesion followed by cytology or biopsy, in which inflammation, hyphae, and masses of budding yeast are seen (not usually seen with

741

742

Section V  Esophagus

Figure 45-3.  A, Achalasia with candidal infection demonstrated by barium esophagography. B, Endoscopic photograph of a dilated esophagus with debris and Candida plaques (arrow) in a patient with achalasia. (A, Courtesy Dr. Marc Levine, Philadelphia, Penn.)

A

B

colonization alone). The entity of the “black esophagus” has also been described with candidal esophagitis.160 Although not as sensitive as endoscopy, candidal esophagitis may be diagnosed by double-contrast barium esopha­ gography.161 The characteristic findings are discrete plaque-like lesions oriented longitudinally, with linear or irregular filling defects with distinct margins produced by trapped barium. Occasionally mass-like lesions and strictures may be seen. Treatment for most patients who have fungal esophagitis and have no immunologic deficiencies is with oral fluconazole or a topical antifungal agent. Fluconazole pills (100 to 200  mg/day) are commonly used because they are more convenient than topical therapy. The advantage of nonabsorbable topical agents is that they are virtually devoid of adverse effects and drug-drug interactions. Clotrimazole, a nonabsorbable imidazole, is well tolerated when delivered as a 10-mg buccal troche dissolved in the mouth five times daily for one week. Nystatin, a nonabsorbable polyene with a different mechanism of action and less palatability than clotrimazole, is also effective when used at a dose of one or two troches (each containing 200,000 units) four or five times daily for up to 14 days.

virus (HSV).160,165 Histologic stains of HSV-infected epithelial cells demonstrate multinucleated giant cells, ballooning degeneration, “ground glass” intranuclear Cowdry type A inclusion bodies, and margination of chromatin. Immunohistologic stains using monoclonal antibodies to HSV antigens or in situ hybridization techniques may improve the diagnostic yield in difficult cases by identifying infected cells that lack characteristic morphologic changes. HSV also may be cultured from esophageal tissue, which is more sensitive than routine histology or cytology. Most patients have self-limited disease paralleling concordant nasolabial herpes, if present, but upper gastro­ intestinal bleeding and perforation have been reported.162 Treatment for herpetic esophagitis is the same as other herpes simplex infections in the immunocompetent host, such as prompt initiation of a 7- to 10-day course of orally administered acyclovir or valacyclovir. Occasionally, severe odynophagia necessitates initial treatment with intravenous acyclovir, 250 mg/m2 every eight hours and then changing to oral therapy when the patient can take oral medication. Given the relative rarity of esophageal involvement, however, no outcome data exist specifically on treating esophageal herpes simplex infection.

HERPES SIMPLEX

HUMAN PAPILLOMAVIRUS

Herpes simplex esophagitis has been described in the immunocompetent host162-164 and can represent either primary infection or, more commonly, a reactivation of latent virus in the distribution of the laryngeal, superior cervical, and vagus nerves. All ages are affected, and oropharyngeal lesions are found in only one in five cases. Severe odynophagia, heartburn, and fever are the dominant symptoms. Nausea, vomiting, and chest pain also may occur. The endoscopic appearance is characterized by diffuse friability; ulceration; and exudates, mostly in the distal esophagus. Classically, the earliest esophageal lesions are rounded 1- to 3-mm vesicles in the mid- to distal esophagus, the centers of which slough to form discrete circumscribed ulcers with raised edges. These lesions can also be appreciated radiographically. The appearance of a “black esophagus” has also been reported with herpes simplex

Human papillomavirus (HPV) is a small double-stranded deoxyribonucleic acid (DNA) virus that infects squamous epithelium of healthy individuals, producing warts and condylomata. The virus can be sexually transmitted. Esophageal infections with HPV are typically asymptomatic. HPV lesions are most frequently found in the mid- to distal esophagus as erythematous macules, white plaques, nodules, or exuberant frond-like lesions.166 In one patient a papilloma developed at a sclerotherapy injection site.167 The diagnosis is made by histologic demonstration of koilocytosis (an atypical nucleus surrounded by a ring), giant cells, or immunohistochemical stains. Treatment is often not necessary, although large lesions have required endoscopic removal. Other treatments such as those using systemic interferon-α (IFN-α), bleomycin, and etoposide have yielded varying results.168 One patient had numerous lesions in the

Chapter 45  Esophageal Disorders Caused by Medications, Trauma, and Infection esophagus and upper airway that were unresponsive to all forms of therapy and eventually fatal.169 HPV infection has been implicated as a risk factor for squamous cell carcinoma, particularly carcinoma of the uterine cervix. An association between HPV and squamous cell carcinoma of the esophagus has been demonstrated by polymerase chain reaction (PCR) or in situ DNA hybridization in esophageal tumor specimens from South Africa, northern China, and Alaska.170 In contrast, HPV DNA was not found in or near esophageal squamous cell carcinomas from the continental United States, Europe, Japan, or Hong Kong.171,172

TRYPANOSOMA CRUZI (see also Chapter 109) Chagas’ disease is the result of progressive destruction of mesenchymal tissues and nerve ganglion cells throughout the body by Trypanosoma cruzi, a parasite endemic to South America. Abnormalities of the heart, esophagus, gallbladder, and intestines are the clinical consequence. Esophageal manifestations may appear 10 to 30 years after the acute infection and typically include difficulty swallowing, chest pain, cough, and regurgitation. Nocturnal aspiration is common. Esophageal manometric recordings are identical to findings in achalasia, although the lower esophageal sphincter (LES) pressure is lower in Chagas’ disease.173 Manometric abnormalities of the esophagus can be found in asymptomatic seropositive patients.174 The putative mechanism is the development of antimuscarinic receptor antibodies in response to the infection.175 A chagasic esophagus may be responsive to nitrates; balloon dilation; or, ultimately, myectomy at the gastroesophageal junction.176 Patients who have intractable symptoms or pulmonary complications secondary to megaesophagus may be candidates for esophagectomy.177 Those with long-standing stasis due to Chagas’ disease often have hyperplasia of esophageal squamous epithelia and are at increased risk for esophageal cancer. MYCOBACTERIUM TUBERCULOSIS

Most reports of esophageal Mycobacterium tuberculosis infections are from areas of endemic tuberculosis. Esophageal manifestations of tuberculosis are almost exclusively a result of direct extension from adjacent mediastinal structures, but there are well-documented cases of primary esophageal tuberculosis.178,179 The clinical presentation of secondary esophageal tuberculosis is quite different from those of most other causes of infectious esophagitis. Specifically, dysphagia is often accompanied by weight loss, cough, chest pain, and fever. Subsequent complications include bleeding, perforation, and fistula formation.179 Choking on swallowing may be indicative of an underlying fistula between the esophagus and respiratory tract. Other radiographic findings include displacement of the esophagus by mediastinal lymph nodes and sinus tracts extending into the mediastinum. Endoscopy is often necessary to confirm active tuberculosis; caution is advised to prevent infection of medical staff by aerosolized tubercle bacilli. Endoscopic findings include shallow ulcers, heapedup lesions mimicking neoplasia, and extrinsic compression of the esophagus.180 Lesions should be biopsied and brushed thoroughly, and specimens should be obtained for acid-fast stain, mycobacterial culture, and PCR, in addition to routine studies. When extrinsic compression is the only esophageal manifestation of tuberculosis, the diagnosis must be confirmed by bronchoscopy, mediastinoscopy, or transesophageal fine-needle aspiration cytologic evaluation.181 Surgery is sometimes required to repair fistulas, perforations, and bleeding ulcers.

TREPONEMA PALLIDUM

Syphilis, which became increasingly prevalent in the United States in the 1990s, can rarely cause esophageal disease in immunocompetent individuals. Earlier literature described gummas, diffuse ulceration, and strictures of the esophagus in tertiary syphilis.182 The diagnosis of syphilitic esophagus should be considered when a patient has an inflammatory stricture and other evidence of tertiary syphilis. Histologic evaluation may show perivascular lymphocytic infiltration; however, specific immunostaining should be done if this diagnosis is a possibility.

OTHER INFECTIONS THAT RARELY INVOLVE THE ESOPHAGUS

Rare viral infections that might involve the esophagus in the immunocompetent adult include herpes zoster and EpsteinBarr virus,183 both of which may produce ulceration. Rare fungal infections of the esophagus include blastomycosis, presenting as an esophageal mass,184 and histoplasmosis, through direct extension of mediastinal adenopathy similar to tuberculosis.185

ACUTE ESOPHAGEAL NECROSIS Acute esophageal necrosis (black esophagus) is a rare disorder that is poorly understood. Ischemia is thought to play a role in its pathogenesis.186

KEY REFERENCES

Cummings JE, Schwiekert RA, Saliba WI, et al. Brief communications: Atrial-esophageal fistulas after radiofrequency ablation. Ann Int Med 2006; 144:572-4. (Ref 92.) D’Avila A, Ptaszek LM, Yu PB, et al. Left atrial-esophageal fistula after pulmonary vein isolation. Circulation 2007; 115:e432-3. (Ref 95.) Donta ST, Engel CC Jr, Collins JF, et al. Benefits and harms of doxycycline treatment for Gulf War veterans’ illness. Ann Intern Med 2004; 141:85-94. (Ref 2.) Famularo G, De Simone C: Fatal esophageal perforation with alendronate. Am J Gastroenterol 2001; 96:3212-13. (Ref 28.) Grudell ABM, Mueller PS, Viggiano TR. Black esophagus: Report of six cases and review of the literature, 1963-2003. Dis Esophag 2006; 19:105-10. (Ref 160.) Gurvits GE, Shapis A, Lau N, et al. Acute esophageal necrosis: A rare syndrome. J Gastroenterol 2007; 42:29-38. (Ref 186.) Kato S, Yamamoto R, Yoshimitsu S, et al. Herpes simplex esophagitis in the immunocompetent host. Dis Esophag 2005; 18:340-4. (Ref 163.) Kikendall JW. Pill-induced esophageal injury. In: Castell DO, Richter JE, editors. The esophagus. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2004. p 572. (Ref 10.) Konikoff MR, Noel RJ, Blanchard C, et al. A randomized, double-blind, placebo-controlled trial of fluticasone propionate for pediatric eosinophilic esophagitis. Gastroenterology 2006; 131:1381-91. (Ref 156.) Macedo G, Azevedo F, Ribeiro T. Ulcerative esophagitis caused by etidronate. Gastrointest Endosc 2001; 53:250-1. (Ref 33.) Manning BJ, Winter DC, McGreal G, et al. Nasogastric intubation causes gastroesophageal reflux in patients undergoing elective laparotomy. Surgery 2001; 130:788-91. (Ref 86.) Nagri S, Hwang R, Anand S, Kurz J. Herpes simplex esophagitis presenting as acute necrotizing esophagitis (“black esophagus”) in an immunocompetent patient. Endoscopy 2007; 39:E169. (Ref 165.) Reed AR, Michell WL, Krige JE. Mechanical tracheal obstruction due to an intramural esophageal hematoma following endoscopic variceal sclerotherapy. Am Surg 2001; 67:690-2. (Ref 75.) Schaefer ET, Fitzgerald JF, Molleston JP, et al. Comparison of oral prednisone and topical fluticasone in the treatment of eosinophilic esophagitis: A randomized trial in children. Clin Gastroenterol Hepatol 2008; 6:165-73. (Ref 157.) Straumann A, Bussmann C, Zuber M, et al. Eosinophilic esophagitis: Analysis of food impaction and perforation in 251 adolescent and adult patients. Clin Gastroenterol Hepatol 2008; 6:598-600. (Ref 135.) Full references for this chapter can be found on www.expertconsult.com.

743

CHAPTE R

46 Tumors of the Esophagus Ananya Das

CHAPTER OUTLINE Carcinoma  745 Epidemiology  745 Etiology and Risk Factors (Squamous Cell Cancer)  747 Etiology and Risk Factors (Adenocarcinoma)  748 Family History and Genetic Factors  749 Molecular Biology  749 Pathology  751 Clinical Features  752 Diagnosis  753 Screening and Surveillance  756 Staging  756 Prevention  760 Treatment  760 Other Malignant Epithelial Tumors  767 Squamous Cell Carcinoma Variants  767 Small Cell Carcinoma  767 Malignant Melanoma  767

Esophageal tumors can be divided into benign and malignant tumors based on their biological behavior and into epithelial and nonepithelial tumors based on histopathology (Table 46-1).

CARCINOMA EPIDEMIOLOGY

A vast majority of all esophageal cancers are malignant epithelial tumors (carcinomas). The two most common types of esophageal carcinomas are squamous cell cancer and adenocarcinoma (Figs. 46-1 and 46-2). Malignant tumors of the esophagus are one of the commonest types of cancer. More than half a million patients were newly diagnosed with esophageal cancer worldwide in 2007 alone. Globally the incidence of esophageal cancer is sixth and ninth among cancers in men and women, respectively, and is the fifth and ninth leading causes of cancer deaths.1 The American Cancer Society estimated that approximately 16,470 new esophageal cancer cases would be diagnosed in the United States in 2008 and that 14,280 patients would die from esophageal cancer.2 In the United States from 2001 to 2005, the median age at diagnosis for cancer of the esophagus was 69 years. Approximately 3% of all patients with esophageal cancer were diagnosed in individuals younger than age 45; 36% between 45 and 64; 29% between 65 and 74; approximately 30% older than age 75. In 2005 the age-adjusted incidence rate of esophageal cancer in the United States for all races and both sexes was 4.3 per 100,000 and varied from 2.9 in Hispanics to 4.9 in blacks. The age-adjusted incidence rates were higher in men than women for all races (overall, 7.5 versus 1.8 per 100,000 in

Benign Epithelial Tumors  768 Squamous Papilloma  768 Adenoma  768 Inflammatory Fibroid Polyp  768 Malignant Nonepithelial Tumors  768 Lymphoma  768 Sarcoma  768 Metastatic Carcinoma  768 Benign Nonepithelial Tumors  768 Gastrointestinal Stromal Tumor  768 Granular Cell Tumor  769 Fibrovascular Polyp  769 Hamartoma  769 Hemangioma  769 Lipoma  770

men and women, respectively). The highest incidence rate in the United States was in black men. Based on data from 2003 to 2005, it is estimated that 0.5% of all men and women born today in the United States will be diagnosed with cancer of the esophagus at some time during their lifetime (lifetime risk, 1 in 198). The overall five-year relative survival rates of patients with esophageal cancer by race and sex, reported for the period from 1996 to 2004 from 17 geographic areas were 16.5% for white men, 17.6% for white women, 9.2% for black men, and 12.9% for black women.3 Squamous cell cancer is the commonest type of esophageal carcinoma worldwide. There are marked geographic variations in the incidence of different types of squamous cell esophageal cancer, with more than 80% of these cancers occurring in developing countries. The highest incidence of esophageal cancers, with incidence rates greater than 100 per 100,000, is in the “Asian esophageal cancer belt,” extending from northern Iran through the central Asian republics to north-central China. The intermediate risk regions, with incidence rates from 20 to 50 cases per 100,000, are in parts of east and southeast Africa (e.g., eastern Kenya, Zimbabwe, and the Transkei region of South Africa), in southeastern South America (southern Brazil, Uruguay, Paraguay, northern Argentina), and in certain parts of western Europe such as northern France and Switzerland. The rest of the world including the United States is considered a low-incidence area, with rates below 10 per 100,000 of the population.1 In high-incidence areas, there are unexplained shifts in incidence rates within countries and regions. For example, within the esophageal cancer belt, the Chinese counties with the highest cancer rates are located in the north-central provinces of Shanxi and Henan, whereas in central Asia the high-risk areas are in parts of

745

746

Section V  Esophagus

Figure 46-1.  A, Distal esophagectomy specimen. The normal white squamous epithelium on the left side of the specimen has been replaced by metaplastic Barrett’s epithelium on the right side. Note the small, irregular ulcer (arrows), which represents a focus of adenocarcinoma. B, Resected specimen of distal esophagus and proximal stomach in a patient with Barrett’s esophagus. Note the large, irregular, ulcerating adenocarcinoma, which involves a small portion of proximal stomach. (Courtesy Edward Lee, MD, Washington, D.C.)

B

A

A

B

Figure 46-2.  A, Histopathology of invasive squamous cell carcinoma. (Hematoxylin and eosin, ×40) B, Histopathology of invasive adenocarcinoma (Hematoxylin and eosin, ×40). (Courtesy Giovanni De Petris, MD, Scottsdale, Ariz.)

Table 46-1  Classification of Esophageal Tumors EPITHELIAL TUMORS

NONEPITHELIAL TUMORS

Malignant Squamous cell carcinoma Adenocarcinoma Adenocarcinoma of the esophagogastric junction Verrucous carcinoma Carcinosarcoma Small cell carcinoma Malignant melanoma Benign Squamous papilloma Adenoma Inflammatory fibroid polyp

Malignant Lymphoma Sarcoma, including malignant GIST Metastatic carcinoma Benign GIST Leiomyoma Granular cell tumor Fibrovascular tumor Hemangioma Hamartoma Lipoma

GIST, gastrointestinal stromal cell tumor.

Turkmenistan and Kazakhstan. In northern Iran there is quite a dramatic difference in the relatively small area of the Caspian littoral; the relatively dry regions east of the Caspian Sea have a high incidence with a much lower incidence in the more humid western parts.4 In Western countries, squamous cell cancers are more common in blacks. In low-incidence areas such as the United States, there is a male preponderance in the incidence of squamous cell cancers. However, such gender

specificity is lost in areas with a high incidence of squamous cell cancers, such as China, where women are nearly equally affected. Persons with low socioeconomic status as well as unskilled manual workers are at almost two-fold higher risk of developing squamous cell cancers. Living without a lifepartner increases the risk of squamous cell cancers, even after adjustment of confounding variables.5 Although the incidence of squamous cell esophageal cancer has decreased over the past two decades in most Western countries and in parts of Asia, including certain high-risk areas of China, there has been a disturbing upward trend in the incidence of esophageal and gastroesophageal junctional adenocarcinoma in the United States and in northern Europe including Denmark, Finland, Norway, Sweden, Scotland, and Switzerland.6-8 In the 1960s squamous cell esophageal cancers comprised approximately 90% of all esophageal cancers. However, because of an alarming rise in the incidence of esophageal adenocarcinoma, esophageal adenocarcinoma is now the predominant type of esophageal carcinoma in the United States. This reversal pattern has also been recently noted in some European countries such as Denmark and Scotland. In a longitudinal study of annual incidence of adenocarcinoma of the esophagus reporting from 43 tumor registries in North America, Europe, and Australia since 1960, the average increase in annual incidence ranged from 15% to 42% in European countries, 23.5% in Australia, and 20.6% in the United States.9

Chapter 46  Tumors of the Esophagus ETIOLOGY AND RISK FACTORS (SQUAMOUS CELL CANCER) Dietary and Nutritional Factors

Drinking from non–tap water has been shown to be an independent risk factor for squamous cell cancer of the esophagus in high-incidence areas of China. Exposure to polycyclic aromatic hydrocarbons commonly present in soot extract of coal-burning stoves, a domestic cooking device used in poorly ventilated kitchens throughout the esophageal cancer belt, has been implicated as a risk factor of squamous cell cancer, particularly in women who do not smoke. An association with dietary intake of N-nitroso compounds has been implicated in the pathogenesis of squamous cell cancer, particularly in high-incidence areas. N-nitroso compounds are derived from reduced dietary nitrates, often by ubiquitous fungal toxins, and certain food material are particularly susceptible to be contaminated by N-nitroso compounds, such as smoked pickles and a breadlike food called qocho or kocho in Ethiopia.10 It has been hypothesized that in black South Africans, the rising incidence of squamous cell cancer could be partly related to a recent dietary shift to maize from sorghum.11 Fusarium fungi, which grow sparsely on sorghum, grow freely on maize, producing fumonisins, which reduce nitrates to nitrites and synthesize cancer-producing nitrosamines. N-nitroso compounds are mutagenic in animal models and act by inducing alkyl adducts in deoxyribonucleic acid (DNA).12 Chewing betel-quid with areca nut and use of gutka (pan masala, a dry powdered mixture of areca nut, catechu, lime, unspecified spices, and flavoring agents), which is very common in Southeast Asia, has been proposed to have a synergistic carcinogenic effect in association with alcohol and tobacco.13 In areas of the world endemic for squamous cell cancers, drinking very hot beverages has been shown to have carcinogenic attributes, probably by causing chronic esophageal damage related to repeated thermal injury. Certain dietary factors such as selenium have been shown to be protective against squamous cell cancers. Low levels of selenium have been associated with squamous cell cancers in high-risk areas of China, and selenium supplementation has been shown to be protective in controlled trials from China.14,15 Similarly, zinc deficiency, which potentiates the carcinogenic effects of nitrosamines as well as causes overexpression of the cyclo-oxygenase pathway in esophageal carcinogenesis, has been recognized as an independent risk factor of squamous cell cancers.16 Evidence suggests that low dietary folate intake and impaired folate metabolism due to functional polymorphisms in folate-metabolizing pathways, particularly polymorphisms in the 5,10-methylenetetrahydrofolate reductase (MTHFR) gene, play an important role in the causation of many gastrointestinal cancers, including esophageal cancer. In a recent meta-analysis of seven case-control studies, relative cancer risks for individuals in the highest compared with the lowest category of dietary folate intake were 0.66 for esophageal squamous cell carcinoma and 0.50 for esophageal adenocarcinoma.17 Also other investi­ gators have shown that the MTHFR 677TT gene variant, which is associated with reduced enzyme activity, was associated with an increased risk of esophageal squamous cell carcinoma.18 There are clearly unknown interactions of genetic, nutritional, and environmental elements that may be synergistically important in the etiology of squamous cell esophageal cancers. For example, in northern Iran, the traditionally semi-nomadic Turkoman population has one of the highest

incidences of squamous cell cancers. Intensive investigation of well-known nutritional and environmental risk factors has failed to identify a specific risk factor that could explain the very high incidence of squamous cell cancers in this population. Prevalence of cigarette smoking and alcohol consumption (discussed following), which are major risk factors for squamous cell cancers in Western populations, was on a negligible scale among men in this population and practically absent among women. Foodstuffs had very few carcinogens such as nitrosamines, aflatoxins, or polycyclic hydrocarbons. It has been postulated that a very restricted diet (low in animal protein, fruit, or fresh vegetables), along with consumption of dry, coarse bread contaminated with sharp seeds and silica fibers, frequent use of hot tea, and the regular smoking or swallowing of opium dross (“sukhtheh”), along with potential genetic predispositions, may act synergistically by an unspecified mechanism to impart a very high risk for developing squamous cell esophageal cancer in this population. Coffee drinking and increased intake of fruit, fish, and white meat has a protective effect on squamous cell cancers. In contrast, red meat, salted meat, and meat boiled at high temperature may increase the risk of squamous cell cancers.19,20

Alcohol and Tobacco

Alcohol and tobacco use, particularly in the form of pipe, cigar, or cigarette smoking, is a dominant risk factor of squamous cell cancers in low-incidence areas, such as the United States. On the other hand, in high-incidence areas, tobacco or alcohol does not seem to play a dominant causative role, although in Asian countries, where smoking is becoming increasingly popular, particularly among men, its relative importance as a risk factor is increasing. Epidemiologic studies suggest that the amount of alcohol consumed is more important than type of alcohol consumed, and the most prevalent alcoholic beverage consumed in a particular region tends to be the one with the highest risk of squamous cell esophageal cancers in that population.21 Multiple epidemiological studies carried out in regions with different incidences of squamous cell cancers suggest that specific polymorphisms in genes encoding for alcohol metabolizing enzymes may determine individual susceptibility to the carcinogenic role of alcohol.22,23

Preexisting Diseases of the Esophagus

Several underlying esophageal diseases are known to increase risk of subsequent development of squamous cell cancers of the esophagus. There is a fairly well-established association of esophageal cancer with achalasia. In a recent epidemiologic study from Sweden involving 2896 patients with achalasia, excess risks for squamous cell carcinoma and adenocarcinoma of the esophagus were observed, predominantly in men. There was no association with esophagomyotomy.24 Esophageal strictures caused by ingestion of lye, a caustic corrosive agent, are associated with a very high risk of squamous cell cancer, which develops three to five decades after the initial event. In achalasia and lye strictures, relative stasis, stagnation of food, and chronic inflammation have been implicated, but no definite carcinogenic mechanisms have been established. Tylosis, particularly the inherited form, which manifests in an autosomal dominant fashion with hyperkeratosis of palms and soles, has been associated with squamous cell cancer. For an affected family member, the estimated lifetime risk of esophageal cancer varied from 40% to 92% by the age of 70 (see Chapter 22). Although rare, the carcinogenetic events in tylosis has received recent attention in that a

747

748

Section V  Esophagus tylosis esophageal cancer (TOC) gene locus has been mapped to chromosome 17q25 by linkage analyses. Interestingly, loss of heterozygosity at this same locus, which contains the promoter sequence of the cytoglobin gene, has been detected in a significant number of patients with sporadic squamous cell cancers.25 The association of chronic mucocutaneous candidiasis with squamous cell carcinoma of the oral cavity and esophagus has been described in patients with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED).26 A rare syndrome of iron deficiency anemia, dysphagia, and postcricoid esophageal web known as Plummer-Vinson syndrome in the United States and Patterson-Kelly syndrome in the United Kingdom has been reported to have an asso­ ciation with squamous cell cancers (see Chapter 41). The anecdotally reported association of squamous cell cancers with partial gastrectomy has not been substantiated. Patients with history of squamous cell cancer of the upper aero-digestive tract have an increased risk of synchronous or metachronous squamous cell carcinoma of the esophagus, most likely due to shared risk factors. In prospective controlled studies, 3% to 14% of patients with squamous cell cancers of head and neck area developed synchronous or metachronous squamous cell cancers of the esophagus.27,28 Radiation therapy following mastectomy moderately increases the risk of squamous cell esophageal cancer in the upper and middle thirds of the esophagus, starting approximately 5 years after exposure, with the risk persisting after 10 years; no similar increase in the risk of eso­phageal adenocarcinoma has been reported. This finding appears to be a function of the portals used for postmastectomy radiation therapy, which typically do not expose the lowest third of the esophagus, where adenocarcinoma commonly arises.29 Given that human papillomavirus (HPV) has been implicated in the etiopathogenesis of squamous cell cancers of the oropharynx, its association with squamous cell cancers of the esophagus has been studied. Patients with squamous cell cancers of the esophagus from the United States, Europe, and Japan are infrequently HPV positive. On the other hand, in certain high-incidence areas, such as China and South Africa, HPV DNA can be demonstrated in a significant proportion of patients with squamous cell cancer of the esophagus. The interactive effect of genetic, environmental, and dietary risk factors in determining geographic susceptibility to the oncogenetic potential of chronic HPV infection needs further study.30 There is no clear association of esophageal cancer with chronic infection with other viruses such as herpes simplex virus or Epstein-Barr virus.

ETIOLOGY AND RISK FACTORS (ADENOCARCINOMA) Dietary and Nutritional Factors

Increased intake of cereal fiber has been shown to have a dose-dependent inverse association with gastric adenocarcinoma and, to a lesser extent, with distal esophageal adenocarcinoma, but not with squamous cell cancer of the esophagus. Diets high in fiber, beta-carotene, folate, and vitamins C, E, and B6 may be protective, whereas diets high in cholesterol, animal protein, and vitamin B12 may be associated an with increased risk of esophageal adenocarcinoma.31 Antioxidant intake appears to be protective against esophageal adenocarcinoma but not gastric adenocarcinoma.32 Carbonated soft drinks may have an inverse relationship, but drinking tea and coffee is not associated with esophageal adenocarcinoma.33

Alcohol and Tobacco

Numerous epidemiologic studies, including prospective studies, suggest that association between alcohol and tobacco use is less consistent with esophageal adenocarcinoma than with squamous cell cancers. In general, smoking is considered a moderate risk factor for esophageal adenocarcinoma, whereas alcohol use has no association with esophageal adenocarcinoma. Relative to persons who had never smoked, current cigarette smoking was found in a prospective study to be associated with increased risk of squamous cell cancers (hazard ratio [HR], 9.27), esophageal adenocarcinoma (HR, 3.70), adenocarcinoma of the gastric cardia (HR, 2.86), and noncardia gastric adenocarcinoma (HR, 2.04).34

Obesity

Pooled results from observational studies support a positive association between an increased body mass index (BMI) greater than 25 kg/m2 and the risk for esophageal adenocarcinoma and possibly for adenocarcinoma of the cardia. In one study, increasing abdominal waist size was associated with an increased risk of esophageal adenocarcinoma, independent of BMI.35 Most recent published information support the hypothesis that obesity, in particular, abdominal obesity, contributes to GERD, which may in turn increase the risk of Barrett’s esophagus (see Chapters 43 and 44).36-38 In contrast to adenocarcinoma, in a large population-based prospective study of Chinese men, a low BMI was associated with an increased risk of squamous cell carcinoma of the esophagus.39

Gastroesophageal Reflux Disease

Barrett’s esophagus, which is a complication of gastroesophageal reflux disease (GERD), is one of the strongest risk factors for esophageal adenocarcinoma (see Chapter 44). However, there is controversy whether GERD without Barrett’s esophagus is by itself a risk factor for esophageal adenocarcinoma. Many patients with esophageal adeno­ carcinoma do not report preexisting GERD symptoms and only a minority of patients with GERD develop Barrett’s esophagus. However, it is also thought that chronic mucosal inflammation in patients with recurrent GERD may predispose to esophageal adenocarcinoma. A large populationbased Swedish study showed that in persons with recurrent symptoms of reflux, as compared with people without such symptoms, the odds ratio for esophageal adenocarcinoma was 7.7 and was 2 for adenocarcinoma of the cardia. The more frequent, more severe, and longer-lasting the symptoms of reflux, the greater the risk.40 A recent Australian study suggested that obesity and GERD may play a synergistic role in the pathogenesis of esophageal adenocar­ cinoma.41 Overall, epidemiologic evidence suggests that chronic, severe GERD may predispose susceptible individuals to esophageal adenocarcinoma.

Barrett’s Esophagus (see also Chapter 44) The reported prevalence of Barrett’s esophagus in the general population varies widely from 0.9% to 4.5%.42,43 Such fivefold variations in prevalence are mostly related to the type of population studied and also the definition of Barrett’s esophagus that is used. For example, experts disagree whether histopathologic demonstration of intestinal metaplasia is a requirement for diagnosing Barrett’s esophagus.44,45 Recent studies have suggested an increasing incidence of Barrett’s esophagus.46,47 The reported annual risk esophageal adenocarcinoma in patients with Barrett’s esophagus varies from 0.2% to 2% (see Chapter 44). In more recent studies with long-term

Chapter 46  Tumors of the Esophagus follow-up of patients with histologically confirmed longsegment Barrett’s esophagus, the annual risk of developing high-grade dysplasia (HGD) or adenocarcinoma is approximately 1%.48,49 Although the annual incidence of cancer in those with low-grade dysplasia (LGD) was 0.6% to 1.3%, the incidence rate in patients with HGD is almost 10 times higher.50,51 In a meta-analysis of 236 patients with Barrett’s esophagus and HGD, esophageal adenocarcinoma was reported in 69 patients over 1241 patient-years of follow-up, with a weighted incidence rate of 6.58 per 100 patientyears.52 In another recently published meta-analysis, the overall pooled risk for developing adenocarcinoma in all patients with Barrett’s esophagus was 6.1 per 1000 personyears and was just 4.1 per 1000 person-years when early incident cancers and HGD at baseline were excluded.53 The risk of esophageal adenocarcinoma in patients with short segment Barrett’s esophagus or with specialized intestinal metaplasia of the esophagogastric junction is not known, but may be less compared with long-segment Barrett’s esophagus. However, it is important to note that these two entities are several-fold more common than long-segment Barrett’s esophagus and could account for a large number of patients with gastroesophageal junctional adenocarcinoma. In addition to segment length and dysplasia, hiatal hernia, central obesity, and possibly smoking may be other factors favoring progression from nondysplastic Barrett’s esophagus to adenocarcinoma.54,55

Helicobacter pylori Infection

Epidemiologic studies show that as the prevalence of Helicobacter pylori infection has decreased in Western societies, the prevalence of GERD, Barrett’s esophagus, and distal esophageal and gastroesophageal junctional adenocarcinoma has rapidly increased. Although there are no data to suggest that H. pylori plays any role in esophageal mucosal resistance to acid-induced injury, in esophageal acid clearance, or in competence of the lower esophageal sphincter, it has been suggested that eradication of H. pylori in patients with corpus-predominant gastritis may exacerbate acid reflux. It has been postulated that the alkaline ammonia produced by H. pylori colonizing the cardiac mucosa could protect the neighboring distal esophageal squamous mucosa from damage by acidic reflux. There appears to be a negative association between H. pylori infection and GERD symptoms, although an association between successful H. pylori therapy and development of new or recurrent reflux symptoms has not been established.56,57 In a recent meta-analysis, patients with Barrett’s esophagus and adenocarcinoma, but not those with squamous cell esophageal cancer, were less likely to be infected with H. pylori, particularly cagApositive strains.58 The clinical implication of such inverse correlations is unclear and will require careful evaluation, given that H. pylori infection is considered an important etiologic factor for gastric cancer.

Diverse Preexisting Conditions

Patients with acid hypersecretory states or conditions associated with severe GERD (e.g., scleroderma) may be at increased risk for esophageal adenocarcinoma. In a population-based study from Sweden, cholecystectomy was associated with a moderately increased risk of sub­ sequent esophageal adenocarcinoma, but not squamous cell cancer. Increased duodeno-gastroesophageal reflux and a toxic effect on bile acids and bile salts on the esophageal mucosa was thought to be the mechanism. A case control study failed to identify an association between diabetes and esophageal adenocarcinoma.59 GERD is common in premature infants and in infants who are small for gestational age,

and epidemiologic data suggest that preterm birth and low birth weight may be risk factors for development of esophageal adenocarcinoma. Down syndrome is not associated with esophageal cancer.60

Drugs

There is conflicting evidence regarding whether long-term use of drugs that relax the lower esophageal sphincter, such as anticholinergics, β-adrenergic agonists, theophylline or aminophylline, and benzodiazepines increase risk of esophageal adenocarcinoma.61 A few epidemiologic studies and a recent meta-analysis reported a protective effect of aspirin use (and to a lesser extent nonsteroidal anti-inflammatory drug [NSAID] use) with both types of esophageal cancer, an effect that appears to be dose dependent.62

FAMILY HISTORY AND GENETIC FACTORS

In areas with a high incidence of squamous cell esophageal cancer, family history is a strong risk factor for this disease. Gene expression profiling studies in these families have found consistent ribonucleic acid (RNA) expression patterns.63 Furthermore, the possibility of an esophageal squamous cell cancer (ESCC) susceptibility gene has been considered based on frequent allelic loss on chromosome 13 in these patients.64 Familial aggregation of Barrett’s esophagus, and related esophageal and junctional adenocarcinoma, has been described and a study reported that in up to 7% patients with Barrett’s esophagus, a familial aggregation can be confirmed (see Chapter 44). It has been suggested that familial Barrett’s esophagus is likely a complex genetic disorder consistent with a major mendelian autosomal dominant gene with relatively high penetrance.65,66 Individual variations in risk of developing esophageal cancer of any type may be partly explained by the presence of specific variant alleles (polymorphisms) of different genes that are present in a significant proportion of the population and likely play important roles in the multistep process of carcinogenesis (see Chapter 3). Such genetic polymorphisms that may increase susceptibility to esophageal cancer have been described in genes involved in alcohol metabolism, folate metabolism, carcinogen metabolism, DNA repair, cell cycle control, and oncogenes.67

MOLECULAR BIOLOGY

Similar to other cancers, esophageal cancers of both types develop through a multistep progressive process, which at the cellular level is reflected by disorders of the control of cell proliferation, differentiation, and controlled cell death (apoptosis) (see Chapter 3). There has been increasing interest in understanding the molecular mechanisms of these carcinogenetic events, not only for understanding etiologic factors but also for potential therapeutic targeting and determining prognosis. A simplified approach to understand the extensive recent research on the putative molecular mechanisms of esophageal carcinogenesis is to describe the known aberrations in terms of phenotypic features of cancer cells such as autonomous growth, attenuation of antiproliferative pathways, disordered apoptosis, unlimited replication, and factors promoting neoangiogenesis, invasion, and metastasis.

Autonomous Growth (Growth Factors)

Increased expression of epidermal growth factor (EGF) and transforming growth factor-α (TGF-α) has been demonstrated in esophageal adenocarcinoma and has been associated with invasive disease and possibly poorer outcome.68,69

749

750

Section V  Esophagus Furthermore, increased concentrations of TGF-β in azygos vein blood and increased expression of connective tissue growth factor (CTGF) and endoglin, a member of the TGF-β receptor family, have been correlated with angiogenesis, angiolymphatic invasion, and metastasis in patients with esophageal cancer.70 On the other hand, increased expression of SMAD has been associated with better survival in patients with squamous cell esophageal cancer.71 C-erbB2, a proto-oncogene coded by the HER2/neu gene and sharing a significant homology with EGF receptor, is amplified in patients with Barrett’s esophagus and adenocarcinoma, is clinically associated with a poorer outcome, and may play an early role in disease invasiveness.72,73 HER2/neu gene overexpression may promote carcinogenesis through multiple pathways, such as inhibition of apoptosis and enhanced cell proliferation, increased mitogen activated protein kinase (MAPK) activity and increased matrix metalloproteinase (MMP) activity, and by potent induction of vascular endothelial growth factor (VEGF). The restriction point (R-point) is a critical gatekeeping checkpoint that controls G1- to S-phase transformation in the cell cycle (see Chapter 3). Cyclin D and E are proteins that are important regulators of the R-point and serve as the final step in many proliferation cascades. CCND1, the gene that encodes cyclin D1, is associated with an increased risk for esophageal adenocarcinoma.74 Also cyclin B has been shown to be overexpressed in esophageal squamous cell cancer, with possibly with a negative prognosis.75

Attenuation of Antiproliferative Pathways

Attenuation of antiproliferative cellular signals is a hallmark of cancer cells. Most cellular antiproliferative pathways converge on the retinoblastoma protein (Rb) pathway, which plays an important antiproliferative role by having a regulatory role on transition of cells from G0 to S phase (see Chapter 3). Loss of heterozygosity (LOH) of Rb protein and its functional repression by hypermethylation of p16, a tumor suppressor gene, has been described in patients with esophageal adenocarcinoma.76,77 Expression of p21, another cell cycle regulator that is activated after DNA damage by ionizing radiation, has been correlated with responsiveness of esophageal cancers to chemoradiotherapy.78 Although mutations of the APC gene is uncommon in patients with esophageal cancer, allelic deletion of 5q where APC resides, as well as functional repression of APC by hypermethylation of the promoter region of this gene, is quite common in patients with esophageal adenocarcinoma.79 In fact, plasma levels of hypermethylated APC DNA may be a biomarker of esophageal cancer.80

Disordered Apoptosis

Programmed cell death, or apoptosis, is an important defense mechanism against cancer and is regulated by TP53 and bcl-2 family of genes and their proteins (see Chapter 3). Alterations in TP53 in patients with esophageal cancer are seen in up to 50% to 99% of patients and are usually associated with a more aggressive form of tumor.81 Similarly, the bcl-2 family of proteins (Bcl-2, Bcl-xl, Bax) maintains an intricate balance of pro- and antiapoptotic factors in the cellular milieu and may play an important yet unclear role in the progression of non-dysplastic Barrett’s esophagus to high-grade dysplasia and adenocarcinoma.82 Nuclear factor kappa B (NF-κB), an antiapopotic factor, is expressed in 60% of patients with esophageal adenocarcinoma and has prognostic significance.83 The protein kinase Akt (also known as protein kinase B), a recently described serine-threonine kinase, is an important mediator of growth and antiapoptotic signals in esophageal

adenocarcinoma. Increased Akt activation in the basal epithelium has been associated with metaplastic progression from squamous epithelium to nondysplastic Barrett’s esophagus, and eventual progression to high-grade dys­ plasia and adenocarcinoma.84 The hormone leptin and short-term acid exposure activate Akt in esophageal adenocarcinoma cells in vitro.85

Unlimited Replication

Telomeres are composed of several thousand repeats of short six-base-pair sequence elements and are located at the ends of chromosomes. At each cell replication the telomeres are shortened so that telomere length serves as counters of cycles of replication. After a finite number of cell replications, the telomeres are short enough to trigger the cell to exit from G1 to G0 phase, thus arresting further replication. Cancer cells, which are characterized by ability for limitless replication, achieve this ability partly by increased expression of telomerase, a ribonucleoprotein reverse transcriptase that counteracts shortening of telomeres. Increased expression of telomerase has been associated with progression of metaplastic epithelium to dysplasia to adenocarcinoma in patients with Barrett’s esophagus.86

Factors Promoting Neoangiogenesis, Invasion, and Metastasis

The ability of cancer cells to invade and disseminate is determined by their ability to disrupt intercellular adhesion as well as by altering the delicate balance between extra cellular matrix degrading proteases, such as urokinase-type plasminogen activator (uPA) of the serine protease system, MMP, and antiproteinases, such as tissue inhibitory metalloproteinase (TIMP) (see Chapter 3). Not unexpectedly, esophageal cancer, one of the most aggressive cancers with respect to invasiveness and propensity to metastasis, is associated with molecular abnormalities related to cell-cell adhesion molecules (CAMs), such as E-cadherins, integrins, and CD44 transmembrane glycoproteins, and overexpression or altered localization of uPA, the cysteine protease, cathepsin B (CTSB) protein, MMP and TIMP. Evidence suggests that many of these may be useful prognostic markers.87,88

Alteration of the Cyclooxygenase Pathway

Cyclooxygenases (COX) are key enzymes in the prostaglandin metabolism. COX-2, an inducible enzyme, has been shown to be an important mediator of tumorigenesis and angiogenesis. Selective inhibition with COX-2 inhibitors induces apoptosis and reduces angiogenesis. Activation of the COX-2 pathway can inhibit apoptosis by lowering the intracellular level of arachidonic acid, a fatty acid precursor of prostaglandins, which is known to promote apoptosis. A second mechanism that can be targeted by COX-2 inhibitors is prostaglandin-E2 (PGE2) dependent, and appears to be mediated by increase in intracellular Bcl-2 as well as intracellular cyclic adenosine monophosphate (cAMP), both of which suppress apoptosis. The angiogenetic potential of the COX-2 pathway occurs through increased expression of VEGF, increased endothelial survival by bcl-2 and Akt signaling, induction of MMP, EGF receptor–mediated angiogenesis, and suppression of interleukin 12 (IL-12) expression. Studies have shown that patients with esophageal cancer who express a high level of COX-2 in their tumors are more likely to have metastatic disease at presentation, to have more local tumor recurrences after definitive treatment, and to have a worse survival than patients not overexpressing COX-2.89,90

Chapter 46  Tumors of the Esophagus Microsatellite Instability

Several investigators have reported microsatellite instability (MSI) in less than 20% of esophageal and gastroesophageal junctional adenocarcinomas. Early unconfirmed evidence, however, suggests that MSI may be an early event in the progression of intestinal metaplasia to esophageal adenocarcinoma.91

Chromosomal Abnormalities

Defects in mitotic checkpoint genes lead to aneuploidy, or an abnormal number of chromosomes. Although a vast majority of esophageal cancers have aneuploidy, the exact gene defects are not characterized. DNA aneuploidy detected by flow cytometry has been considered to be a marker of progression of intestinal metaplasia to esophageal adenocarcinoma.92 A host of specific chromosomal abnormalities have been described in patients with esophageal cancer, and recent studies based on a comparative genomic hybridization technique using different types of esophageal cancer and gastric cardiac adenocarcinoma may lead to identification of genes important for disease-specific pathogenesis.93

Molecular Events in Neoplastic Progression of Barrett’s Esophagus

Research has focused on the complex molecular events that lead to development of esophageal adenocarcinoma in the setting of Barrett’s esophagus.94 It is postulated that the initial turn-on event, which is related to chronic exposure to noxious stimuli, such as acid reflux and bile salt exposure, is the activation of the inflammatory cascade mediated by cytokines such as tumor necrosis factor-α (TNF-α) and IL-1b. This cytokine activation eventually leads to ectopic expression of CDX1 and CDX2 homeobox proteins that play a major role in the development of intestinal epithelium in the embryonic stage and in this setting cause intestinal metaplasia to develop.95-97 With the development of metaplastic intestinal epithelium, multiple poorly characterized genetic (e.g, aneuploidy, loss of heterozygosity, mutation) and epigenetic events (e.g., hypermethylation) set off the metaplasia to dysplasia to adenocarcinoma sequence, which eventually leads to genetic instability with clonal expansion of cells with genetic errors. The exact sequence of this multistep process is not known, but several early events in this transition sequence have been identified.98 Abnormalities of p16 and possibly overexpression of cyclin D1 are relatively early events that lead to loss of control of cell cycle regulatory genes. Telomerase may also play in early role. Prevalence of TP53 damage such as loss of heterozygosity and mutation increases with advancing grade of dysplasia. Similarly, progression of dysplasia is characterized by decreased expression of E-cadherin and membranous β-catenin expression. Growth signals such TNF-α and COX-2 expression also increase with progression of grades of dysplasia and warrant further evaluation as potential biomarkers.

Future Molecular Directions

Exciting developments in the diverse field of genome analysis technology, stem cell signaling pathways, microarray analysis, and proteomics will hopefully lead to clearer understanding of the pathobiology of carcinogenesis and the diverse but interrelated molecular events responsible for development and progression of dysplasia involved into the eventual development of esophageal cancers. Such developments will also likely lead to maturation of our evolving knowledge on a host of potential molecular biomarkers, prognostic indicators, and therapeutic targets for chemopre-

ventive and chemotherapeutic strategies for incorporation into standard clinical practice.

PATHOLOGY Squamous Cell Cancer

Esophageal squamous cell cancer typically develops by progression from premalignant (dysplastic) precursor lesions (Fig. 46-2). Many pathologists describe dysplasia using a two-tiered system: low-grade dysplasia, which includes mild and moderate dysplasia; and high-grade dysplasia, which includes severe dysplasia and cancer in situ.99 In the World Health Organization (WHO) classification the term “intraepithelial neoplasia” is preferred over “dysplasia,” although dysplasia is by far more popular.100 Dysplasia includes both cytologic and architectural abnormalities of varying severity and extent. Low-grade dysplasia often involves the basal part of the epithelium and high-grade dysplasia usually involves the entire epithelial layer and, in a small proportion of patients, dysplasia extends into the ducts of the esophageal glands, simulating stromal invasion or spread in a horizontal pagetoid pattern.101 Dysplasia is often multifocal, supporting a field defect hypothesis in the pathogenesis of esophageal squamous cancer. Cytologic changes are typically seen as coarse chromatin, increased nuclear to cytoplasmic ratio, nuclear hyperchromasia, nuclear pleomorphism, and mitotic figures. Architectural changes of dysplasia include disorganization, loss of polarity, overlapping nuclei, and lack of surface maturation. The presence of these features is often used to distinguish dysplasia from non-neoplastic (reactive) processes typically seen with esophageal mucosal inflammation. Approximately 50% to 60% of squamous cell esophageal cancers occur in the middle third of the esophagus, 33% involve the distal esophagus, and 10% occur in the proximal esophagus. The gross appearance of squamous cell esophageal cancers ranges from a flat, relatively inconspicuous lesion to elevated, plaque-like, polypoid, ulcerated, exophytic, and infiltrative lesions. Superficial squamous cell esophageal cancers invade the submucosa but not the muscularis propria, whereas advanced cancers usually invade the muscularis propria and beyond, either with an infiltrative pattern of small isolated tumor nests or in an expansile pattern of a solid advancing sheet of tumor cells, often with a prominent peripheral lymphocytic infiltration. Approximately one third of squamous cell esophageal cancers are well differentiated, whereas two thirds are either moderately or poorly differentiated, without any features of keratinization. Squamous cell cancers are aggressive and local lymph nodal metastasis occurs early, partly related to the presence of lymphatic channels in the esophageal lamina propria. Because of the prognostic importance of lymph nodal metastasis and its effect in deciding management options, early squamous cell cancer has been categorized into six subcategories based on depth of tumor infiltration. Intramucosal carcinomas are divided into three groups (m1, m2, and m3), and carcinomas invading the submucosa also are divided into three groups (sm1, sm2, and sm3) (Fig. 46-3). Data from Japan suggest that although m1 and m2 early squamous cell cancers do not usually have nodal metastasis, m3 disease has up to an 8% incidence of nodal metastasis, and this incidence progressively increases as the tumor infiltrates the submucosa, with sm1, sm2, and sm3 tumors having incidences of nodal involvement of 17%, 28%, and 49%, respectively. Similarly the incidence of vascular invasion progressively increases, with m1 tumors having

751

752

Section V  Esophagus m1 m2 m3

sm1 sm2

sm3 Mucosa Submucosa

Muscularis

Figure 46-3.  A schematic diagram of depth of invasion of a superficial esophageal neoplastic lesion, as evaluated in a resected mucosal specimen. Depths m1, m2, and m3 denote invasion limited to the epithelial layer, lamina propria, and muscularis mucosae, respectively. Similarly, sm1, sm2, and sm3 denote superficial, intermediate, and deep submucosal invasion, respectively.

none, whereas almost 90% of sm3 tumors show vascular involvement.102 Invasion of local structures, such as mediastinal pleura, the trachea, the bronchi, and the aorta as well as distant metastases to liver, lung, bone, and other sites may be present in more than a third of these patients at presentation and signifies a poor prognosis. Relatively rapid invasion of the tumor into neighboring structures in the mediastinum has been attributed to the absence of a true serosal layer in the esophageal wall.103

Adenocarcinoma

Compared with squamous cell esophageal cancer, a larger body of information exists related to the pathobiology of dysplasia in Barrett’s esophagus and its progression to esophageal adenocarcinoma. Grossly, esophageal dysplasia in Barrett’s esophagus is usually flat and inconspicuous. Dysplasia may be described as low grade or high grade and generally follows the same patterns of cytological or architectural changes described earlier with squamous cell dysplasia. There is considerable interobserver variation in the interpretation and grading of dysplasia and adenocarcinoma among pathologists.104 Furthermore, progression of dysplasia from low grade to high grade is a morphologic continuum and thus may not follow identifiable features of orderly progression for definitive categorization.105 “Adenoma-like” dysplastic changes show minimal crypt distortion and mild cytological abnormalities. This is in contrast to the much less common “nonadenoma-like” dysplasia that typically shows crowded glands containing cuboidal cells with a high nuclear to cytoplasmic ratio, round nuclei with irregular contours, vesicular chromatin, and prominent nucleoli.106 Goblet cells are usually depleted in dysplastic epithelium, and even the surrounding nondysplastic mucosa may show a decrease in goblet cells.107 Many pathologists often use the term “indefinite for dysplasia” when true dysplasia cannot be accurately differen­tiated from reactive changes in the presence of esophageal inflammation. Although lack of surface maturation is normally considered a cardinal feature of dysplasia, this has been recently debated.108 Adjunct immunohistochemical markers (e.g., proliferating cell nuclear antigen (PCNA) and Ki67, cyclin D1, and TP53) have been used

to improve the accuracy of diagnosis of dysplasia in this setting.109,110 Recent studies have reported the utility of immunostaining for alpha-methylacyl-CoA racemase (AMACR) for differentiating non-dysplastic epithelium from low-grade and high-grade dysplasia and also from adenocarcinoma.111,112 In one study, AMACR staining was negative in all cases of Barrett’s esophagus considered negative for dysplasia, whereas 38% of cases of low-grade dysplasia, 81% of cases of high-grade dysplasia, and 72% of cases of adenocarcinoma were positive.111 Another recent study also reported a high negative predictive value for AMACR immunostaining.112 Almost all esophageal adenocarcinomas arise in the setting of Barrett’s esophagus and typically occur in the distal third of the esophagus, including the esophagogastric junction (see Chapter 44). Adenocarcinomas unrelated to Barrett’s esophagus are extremely rare and usually arise from foci of gastric heterotopia in the cervical esophagus (inlet patch; see Chapter 41). Esophageal adenocarcinoma may have a flat, inconspicuous appearance or can be polypoid, ulcerated, or infiltrative. The majority are well or moderately differentiated, usually comprising cystic or tubular glands in solid nests and irregular clusters and often in a cribriform pattern with considerable stratification. In poorly differentiated carcinomas, the tumor cells infiltrate the esophageal wall with sheets of poorly formed glands with a prominent desmoplastic stroma. Signet ring cells and bizarre pleomorphic tumor cells may be present. One of the common problems encountered by pathologists is dis­ tinguishing a tumor of the gastric cardiac from an esophageal adenocarcinoma. In these instances, correlation of the biopsy site with endoscopic anatomic landmarks is crucial. The incidence of lymph node metastasis in esophageal adenocarcinoma is related to depth of tumor infiltration and appears to be equal or less than in squamous cell esophageal cancer. Whereas nodal disease is rare with adenocarcinomas limited to the esophageal mucosa, the rate of nodal metastases for tumors invading into the submucosa is reported to be 27% to 41% for all patients, and 67% to 78% for those with tumors infiltrating the deep submucosa.113-115 Involvement of celiac and perihepatic lymph nodes is more common with esophageal adenocarcinoma than squamous cell carcinoma because of the more common occurrence of these former tumors at or near the gastroesophageal junction.

CLINICAL FEATURES

Patients with squamous cell cancer and adenocarcinoma of the esophagus present with similar symptoms. The most common is progressive dysphagia, often accompanied by disproportionate weight loss. The esophagus can distend up to a diameter of 40 mm in normal adults; solid food dysphagia can occur at luminal diameter of 25 mm, but at or below a diameter of 13 mm dysphagia is always present. Dysphagia initially can be subtle and reported by patients as a sensation of transient sticking of food. Odynophagia usually coincides with presence of an ulcerated tumor. Many patients with early esophageal cancer are asymptomatic. Some may present with iron deficiency anemia. Chest pain or pain radiating to the back is an ominous symptom, often indicating invasion into periesophageal structures. People with esophageal adenocarcinoma are almost eight times as likely to report at least weekly symptoms of reflux or regurgitation as control subjects; no association of GERD symptoms and squamous cell cancer was reported.40 Regurgitation of undigested food material proximal to the area of obstruction and symptoms related to aspiration pneumonia are often present at advanced stages of the disease. Hoarse-

Chapter 46  Tumors of the Esophagus ness can result from recurrent laryngeal nerve involvement by the tumor per se or metastatic lymph nodes. Esophagorespiratory fistula develops in approximately 5% to 15% of all patients with advanced esophageal cancer and is associated with a particularly poor prognosis. Fistula usually manifests with intractable cough and recurrent pneumonia. Uncommon sites of fistulae from esophageal carcinoma include extension to the aorta, pleura, pericardium, and mediastinum. Hematemesis can be due to local hemorrhage from an ulcerated tumor; rarely, exsanguination can occur from development of an aortoesophageal fistula. Many patients have symptoms related to metastases in the lung, liver, bone, or brain, either at presentation or during the course of the disease. The signs of esophageal cancer are nonspecific and commonly include anemia, malnutrition, and signs of recent weight loss of varying degree. Clinical signs of aspiration pneumonia may be present. Cervical lymphadenopathy and enlargement of the liver due to presence of multiple metastatic lesions are present in a minority of these patients.

DIAGNOSIS Laboratory Tests

Laboratory abnormalities in patients with esophageal cancer are nonspecific and most often show anemia and hypoalbuminenia. Anemia can have features of iron deficiency or anemia of chronic disease. Hypercalcemia due to osteolytic metastasis or, less commonly, as a response to circulating parathyroid hormone–related protein (in squamous cancer) may be present in up to a third of patients. Hypercalcemia is more common in patients with squamous cell cancers than adenocarcinomas and may indicate advanced disease and a poor prognosis. There are no specific serologic markers for esophageal cancer.

Imaging Tests

Chest Radiography Chest radiographs in patients with esophageal cancer may show aspiration pneumonia, lung metastases, and pleural effusion. A dilated fluid-filled esophagus may be seen. Evidence of mediastinal widening and tracheal deviation, a thickened retrotracheal stripe, pneumomediastinum, and signs of esophagorespiratory fistula or mediastinal abscess can often be appreciated. Contrast Esophagography The role of contrast esophagography in diagnosis of esophageal cancer has diminished over the years with the increasing use of endoscopy for primary diagnosis. Esophagography should be performed only when this test is likely to affect decision making. Double-contrast barium radiographs often show early cancers as small polypoid lesions, plaque-like lesions, or focal irregularity of the wall. Advanced cancers commonly appear as areas of irregular luminal narrowing, ulceration, and stricture, with abrupt shoulders. Esophagogastric junctional cancers typically extend into the gastric cardia, and for this reason the fundus and cardia of the stomach should always be included in an optimal contrast radiographic study of the esophagus. To evaluate a suspected esophagorespiratory fistula, contrast radiographs are very useful for delineation of the anatomy prior to endoscopic stenting. Such contrast radiographic studies should be performed using barium instead of a hyperosmolar contrast agent such as meglumine ditrizoate because of the risk of pulmonary edema with the hyperosmolar contrast agent (Fig. 46-4).

Figure 46-4.  This barium esophagogram in a patient complaining of cough and dysphagia demonstrates complete esophageal obstruction and a tracheoesophageal fistula. Note the presence of ingested barium in the airways.

Computed Tomography Esophageal cancer typically appears on computed tomography (CT) scan as irregular thickening of the esophageal wall, with a dilated esophageal lumen proximal to the area of the malignant stricture and often as an intraluminal mass. Obliteration of the fat plane between the tumor and adjacent structures, perforation or a sinus tract into the mediastinum, fistula to the tracheobronchial tree, or enlarged mediastinal lymph nodes can be seen in patients with locally advanced tumor. The primary role of CT in esophageal cancer man­ agement is for staging, particularly to exclude distant metastatic disease to the lung and liver and to assess for mediastinal invasion. Three-dimensional multidetector CT imaging and CT esophagography appear to be promising techniques in early reports, reproducing a contrast radiograph–like image with fairly accurate depiction of length of the tumor and its location.116 Endoscopy and Biopsy During endoscopy, the location of the esophageal tumor, its relation to anatomic landmarks such as the upper esophageal sphincter and the esophagogastric junction, and the degree of luminal obstruction are typically assessed. Also endoscopic visualization combined with endoscopic biopsy provides an accurate assessment of malignant (and premalignant) lesions of the esophagus. A Paris classification has been proposed for endoscopic assessment of superficial neoplasia of the gastrointestinal tract, including the esophagus, but its clinical acceptance has been limited particularly in North America.117 Advanced esophageal cancers appear endoscopically as fungating, friable, often ulcerated mass lesions occupying some or all of the luminal circumference, usually with indistinct margins (Fig. 46-5). Less frequently, both types of cancers (squamous and adenocarcinoma) can present with a submucosal infiltrative pattern and may not have a prominent luminal component. If the esophago­ gastric junction is involved by such a tumor, the clinical presentation is often suggestive of peseudoachalasia (see Chapter 42). Typically a higher number of biopsies translates to a higher accuracy rate. Tumors with a submucosal infiltrative pattern may require an aggressive biopsy technique to obtain deeper tissue, using a bite-on-bite technique

753

754

Section V  Esophagus or endoscopic ultrasonography (EUS)-guided fine-needle aspiration of submucosal tissue. Brush cytology is a useful and quick technique to increase the yield in establishing a diagnosis of esophageal cancer, particularly with the availability of advanced cytologic techniques such as fluorescence in situ hybridization (FISH).118

Endoscopic Detection of Dysplasia and Early Cancer

Numerous endoscopic techniques and newer imaging technologies have evolved over the past decade for endoscopic detection of early or superficial esophageal cancer. Conventional Chromoendoscopy Chromoendoscopy with topical application of stains or pigments has been used for improved visualization of subtle mucosal pathology. Lugol’s iodine, when sprayed on the esophageal mucosa during endoscopy, stains the

Figure 46-5.  Advanced squamous cell carcinoma of the esophagus that is nearly completely occluding the lumen.

Figure 46-6.  Images of nondysplastic Barrett’s esophagus using a high-resolution endoscope without (A) and with (B) narrow band imaging. In the lower panels (C, D) an area of early intramucosal cancer in the background of high-grade dysplasia associated with Barrett’s esophagus is shown. Note the irregular and distorted pit and vascular pattern in the area of high-grade dysplasia or intramucosal cancer compared with the nondysplastic Barrett’s esophagus (A, B), which has a regular pit and vascular pattern.

glycogen-rich squamous epithelium black, dark brown, or green-brown, whereas abnormal (neoplastic or inflamed) glycogen-depleted squamous epithelium does not pick up the stain, providing good visual contrast. Lugol’s iodine increases detection of dysplasia and early squamous cell cancers compared with endoscopic visualization alone, and this method may be particularly useful in screening patients at high risk for squamous cell cancer.119 Lugol’s iodine may also be useful in identifying residual islands of Barrett’s epithelium after endoscopic mucosal ablative therapy and development of neosquamous epithelium. For endoscopic detection of high-grade dysplasia and early adenocarcinoma in the setting of Barrett’s esophagus, a number of chro­ moendoscopic techniques, based on application of vital staining with dyes such as methylene blue, toluidine blue, and cresyl violet and contrast stains such indigo carmine, have been used. The largest experience is with methylene blue–based chromoendoscopy, although the reported results of chromoendoscopy in this setting are quite variable, leading to controversy on the utility of this technique in clinical practice.120 A meta-analysis concluded that targeted biopsies after methylene blue–based chromoendoscopy is not superior to random biopsy in detecting specialized intestinal metaplasia and dysplasia of any grade in patients with Barrett’s esophagus who are undergoing endoscopic surveillance and thus cannot be recommended for routine clinical practice.121 Chromoendoscopy using acetic acid, which dissolves the superficial mucus layer by breaking the glycoproteins’ disulfide bonds and reversibly acetylates cellular proteins, has been used to highlight the mucosal vasculature pattern in patients with Barrett’s esophagus.122 Electronic Chromoendoscopy Electronic chromoendoscopy by spectral manipulation of white light, such as narrow band imaging using narrow bandwidth filters or the use of postprocessing spectral estimation technique, has been used during esophageal endoscopy to highlight the surface texture such as pit pattern, as well as the mucosal capillary vascular pattern (Fig. 46-6).123,124 Although initial reports of higher accuracy for

A

B

C

D

Chapter 46  Tumors of the Esophagus

A

C

B

D

early detection of dysplasia and early esophageal cancer were encouraging,125-128 a later report questioning the effect of electronic chromoendoscopy in this setting has dampened the initial enthusiasm.129 High-Resolution Endoscopic Imaging High-resolution endoscopes with pixel densities of approximately 850,000 (compared with approximately 100,000 to 300,000 in conventional endoscopes) and magnification endoscopes, which can optically zoom the image from 1.5 to 150 times, have been used in prospective studies to increase the yield of detection of early esophageal cancers and foci of high-grade dysplasia, often performed in conjunction with chromoendoscopy.130 Fluorescent-aided confocal endomicroscopy, which uses principles of confocal imaging, has been made commercially available, as an endoscope and as an endoscopic probe. Initial experience in early esophageal neoplasia has been encouraging, particularly for subsurface imaging (Fig. 46-7).131 One of the limitations of confocal endomicroscopy is the requirement of a fluorescent dye administered either topically or parenterally for fluorescent contrast. Endocytoscopy is based on the technology that incorporates optical contact microscopy on a probe-based platform for real-time microscopic examination of the esophageal epithelial cells at a magnification ranging from 450 to 1100 times.132,133 EUS and optical coherence tomography (OCT) are endoluminal cross-sectional imaging techniques. The former uses acoustic waves and the latter coherent light waves to quantitatively measure the degree of back-scattering from the tissue to image tissue. High-resolution EUS using catheter miniprobes has limited accuracy for detecting invasive adenocarcinoma in patients with Barrett’s esophagus and high-grade dysplasia or intramucosal carcinoma.134,135 The utility of EUS seems mostly to be for staging esophageal cancer rather than early detection. Several investigators have reported the utility of OCT in detecting areas of highgrade dysplasia and early esophageal cancer; further techni-

Figure 46-7.  Confocal endomicroscopic images (Pentax EC-3870CIK, Pentax of America, Montvale, N.J.) after intravenous administration of 10% fluorescein sodium. A, Columnar epithelium with (dark arrow) and without goblet cells. B, Goblet cells (black arrows) and narrow lumen of a Barrett’s gland (white arrow). C, High-grade dysplasia with loss of basal border (black arrows) of dysplastic cells. D, Disorganized architecture of a gland consistent with invasive cancer (black arrow). The white arrow shows a nondys­ plastic Barrett’s gland with a single layer of columnar epithelium.

cal developments, such as availability of ultrahigh-resolution and Doppler-enabled OCT, as well as optical coherence microscopy, which marries features of OCT and confocal microscopy for 3-D imaging, may further enhance the usefulness of endoscopic OCT.136-138 Spectroscopic Imaging Tissue fluorescence is characterized by absorption of incident light by tissue fluorophores and re-emission of absorbed light at a different wavelength. Autofluorescence of neoplastic tissue is substantially different from normal tissue, a principle that has been used for imaging early esophageal cancers and areas of high grade dysplasia, with limited success. Spectroscopic assessment, including lightscattering spectroscopy and fluorescence spectroscopy, have been evaluated. A recent study showed that trimodal spectroscopy, which combines fluorescence (providing information of tissue chemistry and reflectance) and lightscattering spectroscopy (providing morphologic information such as nuclear size and density), were complementary and provided better results in detecting high-grade dysplasia than either technique alone.139 Raman spectroscopy, which assesses energy shifts due to interaction with different molecules, provides a fingerprint of molecular composition of neoplastic tissue and has been used to study the molecular progression of Barrett’s epithelium to carcinoma, with acceptable accuracy.140 Thus, a host of novel technologies are emerging for both structural and functional imaging of early esophageal neoplasia, which may be useful for early diagnosis, screening, and surveillance of esophageal cancers. However, there is insufficient evidence to choose one technique over another or to recommend routine use of any of the emerging imaging technologies in clinical practice. It is likely that a combination of those imaging modalities that allow “broad-field” evaluation and those that permit high-magnification imaging of suspicious areas will provide the most effective endoscopic imaging tool.

755

756

Section V  Esophagus SCREENING AND SURVEILLANCE

Esophageal cancer fulfills some of the accepted criteria for initiating a targeted screening program in that it is a common disease; the epidemiology, natural history, and risk factors are reasonably well known; and a latent period or asymptomatic stage exists during which appropriate treatment is likely to improve the outcome, given that late diagnosis is generally associated with a dismal prognosis. However, none of the available tests to screen for esophageal cancer fits the criteria of an ideal screening test. Endoscopic screening is accurate and safe, but it is expensive and relatively invasive and its acceptance by the population at risk is not high. Also, there is no convincing evidence that endoscopic screening is effective in reducing mortality or morbidity related to esophageal cancer. Most of the published information on screening for Barrett’s esophagus suggests that reflux symptoms are an insensitive marker of Barrett’s esophagus. Also, only a small proportion of patients with Barrett’s esophagus eventually develops esophageal adenocarcinoma (see Chapter 44). Likewise, only a minority of patients diagnosed with adenocarcinoma were known to have Barrett’s esophagus.141,142 Moreover, a diagnosis of Barrett’s esophagus, made by screening, may have long-term emotional and financial implications.143,144 On the other hand, studies using decision analysis models suggest that endoscopic screening for esophageal adenocarcinoma in high-risk groups, such as white men with GERD symptoms of long duration may be cost-effective.145-147 Published professional guidelines, such as the American College of Gastroenterology guidelines, do not recommend screening for Barrett’s esophagus in the general population and recommend that the use of screening in selective populations at higher risk should be individualized. Screening techniques for esophageal cancer are nonendoscopic and endoscopic. Nonendoscopic techniques include cytology and molecular screenings. Cytologic techniques use a modification of a balloon-based technique or a spongebased technique, using polyurethane mesh compressed inside dissolvable gelatin capsule and is based on the principle of scraping the esophageal mucosa to obtain exfoliative cytology samples. In areas of China and elsewhere with a very high incidence of squamous cell cancer, longterm follow-up studies have demonstrated that cytologic screening techniques may be a viable option for low cost nonendoscopic screening.148 Experience with the use of balloon-based exfoliative cytologic techniques in the area of Barrett’s esophagus is very limited. There is no reliable molecular screening test for esophageal cancer, although a number of molecular markers are being studied for potential application as a screening test in these patients. There has been interest in detecting the plasma level of p16 and hypermethylated APC and also the level of Mcm5 in gastric aspirates, each of which are markers of early epithelial carcinogenesis; small studies have reported a positive predictive value for esophageal cancer.80,149 Endoscopy is by far the most common technique used for screening for esophageal cancer. Endoscopy is performed in conjunction with endoscopic biopsies, often targeted by conventional or electronic chromoendoscopy, or other emerging techniques as described in the preceding section. Studies have shown unsedated, ultrathin esophagoscopy to be tolerable, safe, and accurate in screening in patients for Barrett’s esophagus.150 Esophageal capsule endoscopy and its modifications such as string capsule endoscopy or tethered capsule endoscopy also has been used for screening for Barrett’s esophagus.151,152

Surveillance refers to testing persons who are known to have cancer precursor lesions, such as metaplasia or dysplasia. Although the theoretical argument for surveillance to detect advanced dysplasia and early cancers to allow an appropriate timely intervention to improve outcomes is attractive, the evidence supporting this approach is limited and often contradictory.153,154 Randomized controlled trials have not been performed and are probably impossible to conduct. Some recent retrospective and case-control studies suggest that surveillance endoscopy can detect early-stage cancers and improve survival. In contrast, other studies suggest that the majority of patients with Barrett’s esophagus do not benefit from endoscopic surveillance.155-157 Available studies have been limited by lack of current knowledge on the natural history of dysplasia, as well as lack of agreement among pathologists in terms of diagnosis of dysplasia.104 Economic analyses, with their inherent limitation of uncertainty and assumptions, suggest that although surveillance of patients with nondysplastic Barrett’s esophagus may not be cost-effective, the periodic surveillance of patients with dysplastic Barrett’s esophagus is cost-effective.145,147

STAGING Staging System

Staging of esophageal cancer is not only important for planning appropriate stage-specific therapy but also provides crucial information regarding prognosis. For example, Surveillance, Epidemiology, and End Results (SEER) data from 2005 showed that one-year survival in patients with local, regional, and distant (metastatic) disease were 68%, 54%, and 28%, respectively. The five-year survival for esophageal cancers confined to the mucosa (T1m, N0, M0) may approach 90%.158 The TNM staging system for esophageal cancer is outlined in Table 46-2. The prefixes “c,” and “p” can be used to designate whether the information for a given TNM stage grouping is based solely on clinical information or includes surgical pathology findings, respectively. Some investigators also use a “u” prefix to specify the use of EUS in determining the cancer stage.

Endoscopic Staging

Standard endoscopic evaluation of esophageal cancer can provide very useful information on the extent of tumor. For example, the ability to lift a nodular lesion with submucosal injection of saline suggests early mucosal disease. Endoscopic mucosal resection of early esophageal cancer and pathologic examination of the resected specimen provide excellent information on the depth of tumor invasion and has been used for accurate pathologic staging of early esophageal squamous cell cancer as well as early adenocarcinoma complicating high-grade dysplasia in Barrett’s esophagus.159 At the other extreme, esophageal cancers that are 5 cm or greater in length, or are sufficiently stenotic to prevent passage of a customary endoscope, are likely to be T3 or higher-stage lesions.160

Computed Tomography Staging

CT is typically the initial staging modality once a diagnosis has been established by endoscopy. CT is valuable in detecting metastatic disease in the liver, lungs, and periaortic lymph nodes. CT has a reasonable accuracy in detecting invasion of mediastinal structures in locally advanced tumors and has accuracy rates of up to 90% in detecting aortic, tracheobronchial, and pericardial invasion (Fig. 46-8). Despite improvements in CT technology, the constitu-

Chapter 46  Tumors of the Esophagus Table 46-2  American Joint Committee on Cancer Staging System for Cancers of the Esophagus Tumor Node Metastasis Definitions Primary Tumor (T) TX: Primary tumor cannot be assessed T0: No evidence of primary tumor Tis: Carcinoma in situ (T1a, or T1m) T1: Tumor invades lamina propria or submucosa (T1b, or T1sm) T2: Tumor invades muscularis propria T3: Tumor invades adventitia T4: Tumor invades adjacent structures Regional Lymph Nodes (N) NX: Regional lymph nodes cannot be assessed N0: No regional lymph node metastasis N1: Regional lymph node metastasis Distant Metastasis (M) MX: Distant metastasis cannot be assessed M0: No distant metastasis M1: Distant metastasis Tumors of the lower thoracic esophagus: M1a: Metastasis in celiac lymph nodes M1b: Other distant metastasis Tumors of the midthoracic esophagus*: M1a: Not applicable M1b: Nonregional lymph nodes and/or other distant metastasis Tumors of the upper thoracic esophagus: M1a: Metastasis in cervical nodes M1b: Other distant metastasis AJCC Stage Groupings Stage 0 Tis, N0, M0 Stage I T1, N0, M0 Stage IIA T2, N0, M0 T3, N0, M0 Stage IIB T1, N1, M0 T2, N1, M0 Stage III T3, N1, M0 T4, any N, M0 Stage IV Any T, any N, M1 Stage IVA Any T, any N, M1a Stage IVB Any T, any N, M1b *For tumors of the midthoracic esophagus, use only M1b because tumors with metastasis in nonregional lymph nodes have an equally poor prognosis as those with metastasis in other distant sites. Modified from American Joint Committee on Cancer (AJCC) Cancer Staging Manual. 6th ed. New York: Springer; 2002. p 91.

ent layers of the esophageal wall cannot be easily differentiated from each other, which explains the poor accuracy (50% to 60%) of CT in assessing tumor stage. CT has a sensitivity of 50% to 79% in detecting nodal disease with reported specificities of 25% to 67%. Accuracy of CT in assessment of periesophageal abdominal lymph nodes is superior to detection of thoracic adenopathy. CT (even helical CT) is insensitive for detection of involvement of celiac lymph nodes in esophageal cancer. In a prospective trial, positive and negative predictive values for helical CT assessing celiac lymph nodes were only 67% and 77%, respectively, using the gold standard of EUS with fine-needle aspiration (FNA).161 Although magnetic resonance imaging (MRI) can easily delineate the margins of the air-filled esophagus from the surrounding mediastinal

fat, MRI does not offer any significant advantage over CT, even with the use of experimental endoscopic MRI techniques.162

Positron Emission Tomography Staging

Fluorine 18 fluoro-deoxyglucose (FDG) positron emission tomography (PET) is being increasingly used for staging esophageal cancer. PET has the ability to image the entire body and thus provide information on both locoregional and metastatic disease (Fig. 46-9). A recent review reported that FDG PET is more sensitive and specific than CT for detecting distant metastases in patients with esophageal carcinoma. However, the sensitivity of FDG PET for detecting locoregional disease is low and inferior to EUS due to a number of factors, including (1) limitation of spatial resolution with currently available PET scanners; (2) masking of adjacent involved lymph nodes due to FDG accumulation in the primary tumor; and (3) low FGD uptake by poorly differentiated esophageal and EG junctional adenocarcinomas.163

Endoscopic Ultrasonography Staging

EUS is the cornerstone in the pretreatment staging evaluation of esophageal cancer. EUS images of the normal esophageal wall at typical scanning frequencies (7.5 to 12, or even 20 MHz) show characteristic delineation of the different echo layers of the esophageal wall. EUS is the best imaging modality for tumor staging, with an overall accuracy for tumor staging of up to 85% to 90% in the hands of experienced endosonographers (Fig. 46-10).164 Accuracy seems to be the lowest (≈ 80%) for T2 cancers. In addition to imaging the esophageal wall itself, EUS also obtains high-resolution views of the periesophageal tissue, including the celiac trunk, the left lobe of the liver, the left adrenal gland, the thoracic and proximal abdominal aorta, the azygos vein, the right and left pleura, the left atrium, and the subcarinal region. The overall accuracy of EUS for defining the presence of nodal metastasess ranges between 65% and 86%. Lymph nodes that are hypoechoic, round, and 1 cm or larger, with clearly demarcated borders, are more likely to be malignant than nodes that are elongated and hyperechoic with poorly demarcated borders, which are most often reactive nodes. The higher the number of criteria a lymph node acquires, the more likely it is to be malignant, although all four of these criteria are found only in a minority of malignant nodes. The addition of EUS FNA to EUS imaging increases the accuracy of nodal assessment, and more liberal use of EUS-guided FNA has been advocated.165 Adjunctive analytic techniques, such as digital image analysis and elastography may enhance the ability of EUS in differentiating benign from malignant periesophageal lymph nodes. There is no difference in staging accuracy between curved linear array and radial scanning echoendoscopes. Highfrequency catheter-based ultrasound probes that provide higher resolution images but with limited depth of imaging, may be particularly useful to assess the feasibility of endoscopic mucosal resection by differentiating T1m cancers from T1sm cancers (see Table 46-2). However, the reported accuracy of these probes in demarcating mucosa from the submucosa has been variable. Whether stenotic esophageal cancers should be dilated to enable complete EUS examination is debated, with some investigators reporting a high rate of perforation when dilations were performed prior to EUS. If the additional information obtained by performing a complete EUS evaluation does not lead to a significant change in management, the risk of dilating stenotic cancers may not be justified. A

757

758

Section V  Esophagus

Figure 46-8.  A, Computed tomography (CT) image showing circumferential thickening of the wall of the upper third of the esophagus with transmural invasion and nearly complete occlusion of the esophageal lumen (top center). B, Paratracheal nodal involvement is also seen. Although CT is effective for the detection of distant metastases, it does not allow accurate local tumor staging.

A

slimmer ultrasonic esophagoprobe may be useful in staging highly stenosing esophageal cancers.

Other Staging Modalities

Minimally invasive procedures such as laparoscopy and mediastinoscopy have received considerable attention as preoperative staging modalities in patients with esophageal cancer and may be useful in selected patients. Other investigations such as cervical ultrasonography, bronchoscopy, or lymphangiography have no role in routine preoperative workup of esophageal cancer.

Effect of Staging Approach on Outcome

There have been few analyses of the effect of EUS-based staging on outcomes in patients with esophageal cancer. In a prospective study of 204 patients pretreatment, EUS predicted survival in esophageal cancer based on initial Tcategory and N-category; celiac nodes and lymphadenopathy at EUS were shown to be important predictors of survival.166 Another study assessed the clinical impact of EUS on esophageal cancer in 107 patients with esophageal cancer and reported improved survival in patients who were evaluated with EUS for staging.167 In an analysis of the SEER-Medicare–linked database, receipt of EUS was independently associated with improved survival, possibly because of improved stage-appropriate management in a cohort of 2830 patients with esophageal cancer.168 Studies have also evaluated the effect on patient outcome of FDG PET–based staging, with divergent results.169,170 There are other possible explanations for the increased survival associated with receipt of EUS and perhaps PET. The concept of stage migration, also known as the Will Rogers phenomenon, is well known in oncology and occurs with more accurate staging of cancer.171 If a population of patients with cancer is staged more accurately by any newer staging modality, then some patients with subtle advanced disease will be upstaged; this upstaging will appear to improve survival in patients with early and advanced cancers. A decision analysis study favored a strategy of initial staging of esophageal cancer by PET and EUS FNA over a strategy based on CT and EUS FNA in terms of higher effectiveness, although the PET-based strategy was more expensive.172

B

Restaging

There has been resurgence of interest in restaging of esophageal cancer after preoperative chemoradiotherapy (neoadjuvant therapy). Currently, there are no reliable biomarkers or histopathologic or morphologic predictors of response to preoperative chemoradiotherapy. In clinical practice, in order to decide whether to attempt curative surgical resection, response to preoperative chemoradiotherapy is assessed by different imaging modalities. Although commonly performed for restaging, CT has a poor ability to discriminate responders from nonresponders; in a metaanalysis, the overall sensitivity and specificity of CT in predicting pathological response varied from 33% to 55% and 50% to 70%, respectively.173 EUS is one of the common modalities for restaging esophageal cancer after chemoradiotherapy. Similar to other morphologic imaging modalities, EUS suffers from its inability to discriminate viable tumor from post-chemoradiotherapy fibrosis and scarring. Primarily for this reason, studies of EUS in restaging of esophageal cancer have consistently yielded poor accuracy for T category and investigators often relied more on surrogate markers of tumor volume, such as maximal cross-sectional areas or tumor thickness during assessment of response to neoadjuvant therapy.174 Patients with recalcitrant nodal disease after chemoradiotherapy are particularly likely to have a poor long-term outcome. EUS, particularly with its ability to sample suspected lymph nodes by FNA, is considered a useful modality in this regard, with an overall sensitivity and specificity of approximately 50% to 100% and 36% to 100%, respectively, with the lower estimates derived from studies in which EUS FNA was not used. FDG PET is increasingly being used for restaging of esophageal cancer. This imaging modality has two basic advantages over EUS: the ability to image the entire body in one examination for evaluation of metastatic disease, and, when combined with CT, the ability to provide structural and metabolic information with high reproducibility and internal validity. Serial quantitative measurement of tumor FDG uptake (by standardized uptake value, or SUV) provides an objective means to monitor response to chemoradiotherapy and has been correlated with survival in these patients. Given that almost half of patients with esophageal cancer have a minimal or suboptimal response to chemoradiother-

Chapter 46  Tumors of the Esophagus

CT Transaxials

CT Coronals

PET Coronals

Fused Coronals

PET Transaxials

CT Sagittals

PET Sagittals

Fused Sagittals

Fused Transaxials

Figure 46-9.  Computed tomography (CT)/positron emission tomography (PET) scans with fluorine-18 fluorodeoxyglucose in a patient with metastatic adenocarcinoma of the esophagogastric junction. Multiple areas of uptake are seen in the axial skeleton, soft tissues, and lymph nodes. CT, PET, and fused images are shown. (Courtesy Dr. Michael Roarke, Scottsdale, Ariz.)

apy, there has been considerable interest in predicting response early in the course of chemoradiotherapy in order to prevent continued toxic, futile, and costly treatment in nonresponders. In a prospective study it was shown that performing FDG PET two weeks after initiation of systemic chemotherapy, a metabolic response in the tumor can be objectively identified by a reduction of SUV compared to baseline, and this response predicted an eventual clinical

response to completed chemoradiotherapy with a high degree of accuracy.175 There have been only a few studies that have directly compared EUS and FDG PET head to head in patients undergoing chemoradiotherapy. In a prospective study of 41 consecutive patients, it was shown that PET CT (using a hybrid PET CT scanner) was superior to EUS and CT scan in evaluating nodal status after chemoradiotherapy. PET CT

759

760

Section V  Esophagus

Normal Wall

B

A

Figure 46-10.  Endoscopic ultrasonography (EUS) staging images of esophageal tumors using a radial scanning echoendoscope: A, A T1 lesion is observed as a hypoechoic thickening of the mucosal layer adjacent to the normal-appearing wall pattern. B, A T2 lesion is seen as a hypoechoic mass invading into but not through the muscularis propria. C, A T3 lesion is seen as a hypoechoic mass (inferior) with an irregular margin extending into the periesophageal fat tissue and disrupting the normal wall layer pattern which is seen at the upper half of the image. D, The T4 lesion seen here is a circumferential hypoechoic mass with direct extension in the wall of the thoracic aorta (two to three o’clock) and the right mainstem bronchus (six to eight o’clock), denoting unresectability. E, A typical malignant-appearing lymph node is greater than 1 cm in diameter, hypo­ echoic, and round and has sharply demarcated borders. F, EUS-guided fine-needle aspiration performed using a linear array echoendoscope allows cytopathologic sampling to confirm lymph node metastasis.

C

D

F

E

was also the more accurate modality for identifying nodal status.176 In a systematic review of published information, the overall joint sensitivity-specificity (Q-point) values for FDG PET was reported to be 85% which was no different than EUS at 86% and significantly better than CT at 54%.173 Thus, subject to local availability of resources and expertise, it appears that for restaging of esophageal cancer patients after chemoradiotherapy, conventional CT, particularly the single detector variety, has no role and FDG PET and EUS FNA are likely to provide complementary information. For example, if hybrid PET CT shows suspected nodal disease or residual primary tumor, this finding should be confirmed by EUS FNA. On the other hand, it is not known whether, in a patient with a completely negative PET CT, additional evaluation with EUS FNA will improve accuracy of assessment of restaging evaluation.

selenium supplementation and use of COX inhibitors have received attention as strategies for chemoprevention of esophageal cancer. Unfortunately, two randomized trials have reported rather disappointing results. In one trial of 100 patients with Barrett’s dysplasia, administration of 200 mg of celecoxib twice daily for 48 weeks did not prevent progression of the dysplasia to adenocarcinoma.177 In a second randomized placebo-controlled trial involving 238 Chinese patients with mild or moderate esophageal squamous dysplasia, selenomethionine (200 µg/day) and/or celecoxib 200 mg twice daily did not prevent esophageal squamous carcinogenesis in high-risk subjects. Selenomethionine did result in a nonsignificant trend toward dysplasia regression (43% vs. 32%) and less dysplasia progression (14% vs. 19%) compared with no selenomethionine.178

PREVENTION

The management of esophageal cancer has been evolving and has become more individualized and stage specific. Primary treatment modalities comprise resection of the

Based on data from epidemiologic studies and supported by insight in the molecular pathobiology of esophageal cancers,

TREATMENT

Chapter 46  Tumors of the Esophagus primary tumor, chemotherapy, and radiotherapy. In many patients with early esophageal cancer, endoscopic resection has become a reasonable option, with good outcome data. In most patients who have locally advanced esophageal cancer, combined modality therapy (chemotherapy plus surgery or chemotherapy and radiation therapy plus surgery) has gained popularity. For advanced tumors suitable for palliative treatment, the emphasis has shifted from surgical intervention to nonsurgical modalities, such as endoscopic palliation or chemoradiotherapy. A concise summary of the available treatment-specific modalities for different tumor stages is provided below.

Early Esophageal Cancers

Early esophageal cancers include carcinoma in situ (Tis) or cancers that invades the mucosa (T1a or T1m) or submucosa (T1b or T1sm) but without nodal disease (i.e., T1, N0, M0; see Table 46-2). Endoscopic Therapy Esophageal resection has been the standard of care for early esophageal cancer. However, evidence has accumulated that early esophageal cancers with m1 or m2 depth of invasion have little or no risk of nodal disease. In such patients, endoscopic resection may be an attractive alternative to surgery. One endoscopic resection technique is similar to saline-assisted polypectomy using a simple polypectomy snare. A second technique (lift-and-cut) uses a doublechannel endoscope and a biopsy forceps to lift up the lesion prior to polypectomy. A third and common technique is the cap-assisted technique, which uses a transparent cap attached to the endoscope to suction the target lesion (typically after submucosal injection of saline below the lesion to form a pseudopolyp), which is then resected using a prelooped snare. A fourth ligate-and-cut (band-and-cut) technique uses a banding device to suction the target lesion, followed by application of a rubber band ligation prior to polypectomy. This particular technique has become popular

A

B

C

D

with the introduction of a commercially available endoscopic mucosal resection (EMR) kit, which is a modified variceal band ligation device that allows multiple resections with a single endoscopic intubation of the esophagus (Figs. 46-11 and 46-12). The latter two techniques have been compared in a controlled trial and were similar in terms of size of the resected tissue specimens and complications.179 To date, the largest series using EMR in the setting of Barrett’s adenocarcinoma reported 100 consecutive patients with early esophageal adenocarcinoma. These patients underwent 144 resections (maximum, 3 resections) using the ligate-and-cut technique; a complete local remission was achieved within 1.9 months (range, 1 to 18 months) in 99 of the 100 patients. During a mean follow-up period of 36.7 months, recurrent or metachronous carcinomas were detected in 11% of the patients, but successful repeat treatment with endoscopic resection was possible in each case. The calculated five-year survival rate was 98%. Minor bleeding, which was easily controlled with endoscopic therapy, occurred in 11% of patients.158 Other investigators have reported similar success with endoscopic resection of early esophageal cancer, including in patients with squamous cell cancers.180,181 The ideal lesion for EMR is a solitary, nodular lesion, less than 20 mm in diameter, limited to the mucosa and with a suitable macroscopic appearance (macroscopic types I, IIa, IIb, and IIc by the Paris classification). En-bloc resection is always preferred to piecemeal resection, the latter of which increases the risk of local recurrence in addition to making a his­topathologic interpretation of complete resection more difficult. Endoscopic submucosal dissection (ESD) is becoming popular in Japan and involves a deeper and larger resection of the esophageal wall by dissecting the submucosal connective tissue just beneath the target lesion from the underlying muscle layer using the hook knife or other electrocautery devices. ESD is usually reserved for larger lesions with evidence of some degree of submucosal invasion. ESD also requires high level of expertise on the part of the endosco-

Figure 46-11.  A schematic diagram showing different techniques of endoscopic mucosal resection. A, The simplest technique, similar to saline-assisted polypectomy using a polypectomy snare. B, Lift-and-cut technique, through a double-channel endoscope, using a biopsy forceps to lift the lesion prior to snare polypectomy. C, Cap-assisted technique, in which a transparent cap is attached to the tip of the endoscope with a snare that is pre-looped inside the cap. With gentle suction the lesion is sucked into the cap to create a pseudopolyp. The snare is closed, suction released, the closed snare is pushed out of the endoscope, and the pseudopolyp is cut with electrocautery. D, Ligate-and-cut (band-and-cut) technique, which uses a banding device to suction the target lesion, followed by application of a rubber band prior to polypectomy.

761

762

Section V  Esophagus

A

B

C Figure 46-12.  A, Endoscopic image of a nodular superficially ulcerated lesion at the gastroesophageal junction, which, on biopsy, was an adenocarcinoma in the setting of Barrett’s esophagus. B, An endoscopic mucosal resection (EMR) was performed using the ligate-and-cut technique (see Fig. 46-11D). C, Photomicrograph of the EMR specimen showed adenocarcinoma (Hematoxylin and eosin, ×2). The inset is a magnified view (Hematoxylin and eosin, ×40) of the area of the mucosa indicated by the square in which an intramucosal adenocarcinoma (m1; see Fig. 46-3) could be identified by the presence of fused glands with severe cytological atypia. (Courtesy Giovanni De Petris, MD, Scottsdale, Ariz.)

pist.181 In a comparative trial, ESD using the hook knife had a higher en-bloc and curative resection rate than EMR using a transparent cap, but with a similar complication rate.182 Endoscopic resection is often combined with other forms of endoscopic ablative therapy such as photodynamic therapy (PDT), laser ablation, thermal ablation with argon plasma coagulation (APC), or radiofrequency ablation. Combination therapy is particularly applicable for larger lesions, multifocal lesions, or for treatment of flat high-grade dysplasia in patients with Barrett’s esophagus. Experience with EMR combined with other endoscopic ablative therapy is limited, but may be a viable alternative to surgery.183 PDT has been studied as single-modality therapy. In a recent report, a complete response could be achieved in all 31 patients with early adenocarcinoma and Barrett’s esophagus treated with 5-aminolevulinic acid–based PDT, but local recurrence occurred in almost one third of patients.184 Surgical Therapy Although recent studies have reported excellent outcomes with endoscopic therapy for early esophageal cancers, there is a relative paucity of long-term follow-up data. Thus, the role of endoscopic therapy for early esophageal cancer is still a matter of controversy, particularly in the United States. Also, excellent outcomes after endoscopic therapy of early esophageal cancer are typically seen in specialized centers treating a small number of patients; it is unclear whether such excellent outcomes can be achieved in community practice.185 In a recent study of early esophageal

cancer that analyzed data from a large population-based cancer registry, cancer-free survival was similar in patients undergoing endoscopic therapy or surgical resection.186 Obviously, management decisions for treatment of early esophageal cancer should be individualized. Although younger, otherwise healthy patients with larger mutifocal lesions with submucosal invasion may benefit from surgery, selected older patients with limited disease and multiple comorbidities may be more appropriate candidates for endoscopic therapy. When surgical resection is considered for treatment of early esophageal cancer without submucosal involvement, less morbid surgical techniques such as vagal-sparing esophagectomy, which has fewer postoperative complications, may be considered.187 Radiation Therapy and Chemotherapy The role of radiotherapy, with or without chemotherapy, as definitive treatment of early esophageal cancer has been evaluated. In a series of 141 Japanese patients with superficial esophageal cancer who were treated with external beam radiotherapy, the three-year survival rates were 90% and 70% for tumors limited to the mucosa and submucosa, respectively, and were comparable with reported surgical outcomes.188 It is likely that concurrent chemotherapy may be effective in decreasing local recurrence, particularly in tumors with submucosal invasion. Addition of high-dose brachytherapy may decrease the rate of local recurrence, but at a higher risk of radiation toxicity.189 Some investigators have reported excellent results using a combination of endo-

Chapter 46  Tumors of the Esophagus scopic resection followed by chemoradiotherapy or even using chemoradiotherapy after EMR for management of nodal recurrence.190,191 Until controlled studies are conducted, radiotherapy (with or without concurrent chemotherapy) should be reserved for those patients with early esophageal cancer who cannot undergo endoscopic resection due to contraindications to this procedure (e.g., esophageal varices or severe cervical spinal disease) in whom endoscopic procedures may be more risky.

Locally Advanced Cancer

Surgical Therapy Different surgical techniques have been used for primary management of esophageal cancer, the choice of which depend on factors such as tumor location, stage, extent of lymphadenectomy, replacement conduit planned, use of neoadjuvant therapy, and the preference of the surgeon. Overall, 30% to 59% of patients with esophageal cancer are candidates for potentially curative resection, with five-year survival rates ranging from 15% to 24%. However, the operative mortality and perioperative morbidity rates (e.g., cardiopulmonary events, infection, anastomotic leaks or strictures) vary from 4% to 10% and 26% to 41%, respectively.192 Transhiatal esophagectomy and Ivor-Lewis transthoracic esophagectomy are the most common techniques used in the United States. Modifications of transthoracic esophagectomy, such as transthoracic total esophagectomy with node dissection and cervical esophagogastric anastomosis (the three-field technique) and left thoracoabdominal incision with gastric pull-up into the left chest, are also used frequently. These techniques have their own advantages and disadvantages. For example, a complete mediastinal exploration and lymphadenectomy is not possible with the trans­ hiatal approach. With the transthoracic technique, there is greater likelihood of a troublesome intrathoracic anastomotic leakage and a higher incidence of postoperative bile reflux. In a prospective randomized trial, transhiatal esophagectomy was associated with lower morbidity than transthoracic esophagectomy with extended en-bloc lymphadenectomy.193 However, a recent update of this cohort showed a 14% five-year survival advantage in patients with distal esophageal adenocarcinomas if they had been randomized to the transthoracic approach.194 Despite this finding, an earlier meta-analysis of published trials, and a more recent population-based outcomes study found that transthoracic and transhiatal resections yielded similar long-term survival rates.195,196 Published data also suggest that patients undergoing esophagectomy in high-volume hospitals may have less perioperative morbidity, lower mortality, and improved early clinical outcomes, although there are conflicting data on the effect of hospital volume on longterm survival.197,198 Radiotherapy Radiation therapy has been a cornerstone of treatment of esophageal cancer. Modern radiotherapy techniques such as three-dimensional conformal radiotherapy [3D-CRT] and intensity modulated radiotherapy [IMRT] have reduced toxicity from radiotherapy. Radiotherapy alone can result in long-term survival in patients with esophageal cancer, and reports of five-year survival of 21% to 59% have been published in selected patients with early disease. In a recent Chinese trial, 269 patients with resectable squamous esophageal cancer type were randomized to surgery versus latecourse accelerated hyperfractionated conformal radiotherapy alone; the five-year survival rates were similar: 36.9% versus 34.7%, respectively.199

Combined Chemoradiotherapy More commonly, combined modality treatment with chemotherapy and radiotherapy is used. This approach is based on the argument that the antitumor effect is additive (related to radiation sensitization) and is based on outcome data from randomized control trials. Combined chemoradiotherapy can be definitive therapy or preoperative (neoadjuvant) therapy. Chemoradiotherapy can be used concurrently or administered sequentially. In a Cochrane review of randomized trials, concomitant (but not sequential) chemoradiotherapy provided a significant reduction in mortality at one and two years. Combined chemoradiotherapy provided an absolute risk reduction for mortality by 9% and 4% after one year and two years, respectively, with a 12% reduction in local recurrence rate.200 However, there was significant toxicity related to such treatment. Thus, it appears that combined concurrent chemoradiotherapy, particularly using cisplatinum-based chemotherapy, is superior to radiotherapy alone in the setting of definitive therapy for locoregional disease (without surgical intervention), but better outcomes come with a cost of higher toxicity. Also, use of higher dosage of radiotherapy did not increase survival, but added to toxicity.201 Given the poor long-term results with surgical resection alone, investigators have studied the effect of adding preoperative chemoradiotherapy on long-term survival. This neoadjuvant approach carries the theoretical promise of improved local tumor control and eradication of micrometastasis. Several randomized controlled trials have been published comparing neoadjuvant chemoradiotherapy with surgery alone in patients with localized cancer of the thoracic esophagus or esophagogastric junction. Although it appears that neoadjuvant chemoradiotherapy, particularly administered concurrently, may impart a modest survival advantage, the results from the various trials are not uniform. Two meta-analyses have addressed this issue. The first showed that compared with surgery alone, neoadjuvant chemoradiation followed by surgery improved three-year survival and reduced locoregional cancer recurrence.202 The overall rate of resection attempt was reduced, but when resection was attempted, a higher rate of complete (R0) resection was achieved. There was also a trend toward increased treatment complications with chemoradiotherapy, and benefits were restricted to those undergoing concurrent chemoradiotherapy administration. In a second, more recent meta-analysis, pooled results from 10 randomized trials comparing neoadjuvant chemoradiotherapy with surgery alone in more than 1000 patients showed that the relative risk for all-cause mortality with neoadjuvant chemoradiotherapy versus surgery alone was 0.81 (95% CI 0.70 to 0.93), resulting in a 13% survival difference at two years. Results were similar in patients with squamous cell or adenocarcinoma of the esophagus.203 In order to consolidate the modest gains achieved by neoadjuvant therapy in this situation, several investigators have attempted intensification of the preoperative regimen by including cycles of induction chemotherapy prior to concurrently administered chemoradiotherapy. Also, other chemotherapeutic drugs with better radiosensitizing potential (such as paclitaxel, carboplatin, and oxaliplatin) have been studied, with some initial encouraging results. However, the strategy of intensification of the preoperative regimen has not been proven in randomized controlled trials. The concern remains whether the increased complications seen with such aggressive regimens will be justified by a corresponding increase in longterm survival. Few trials have compared neoadjuvant chemotherapy (without radiation) with surgery alone. In the same meta-

763

764

Section V  Esophagus analysis referred to earlier, neoadjuvant chemotherapy was favored over surgery alone, albeit with a lower reduction in mortality compared with neoadjuvant chemoradiotherapy.203 Several issues remain unanswered in the management of locally advanced esophageal cancer and esophagogastric junctional cancers. It is not known if there is any difference in outcome in patients treated with primary chemoradiotherapy or with neoadjuvant chemoradiotherapy followed by surgery. Experience from few controlled studies suggests that patients treated with neoadjuvant chemoradiotherapy followed by surgery had better locoregional control and needed fewer palliative procedures, but had similar overall survival and quality of life compared with those treated with primary (definitive) therapy.204,205 It is important to note that these conclusions are based on studies in which most patients had squamous cell cancer; in patients with adenocarcinoma, the case for the necessity of surgery after chemoradiotherapy is stronger. Salvage esophagectomy in patients with recurrence after primary (definitive) chemoradiotherapy may be a feasible option in a minority of patients, but such a strategy is plagued by two problems: first, difficulty in identifying recurrent disease in these patients and second, the increased risk of perioperative complications in patients undergoing salvage esophagectomy compared with primary esophagectomy or a planned esophagectomy after combined modality treatment.

Locally Advanced Unresectable Cancer

Up to two thirds of patients with esophageal cancer present with tumors that are considered unresectable or inoperable because of distant metastatic disease (M1b), presence of extraregional nodal disease (M1a), T4 tumors, high surgical risk due to comorbid candidates or advanced age, or unwillingness to undergo a major surgical procedure. In such patients, palliative surgical procedures are no longer recommended, primarily because nonsurgical interventions are either superior or at least equivalent to surgery in achieving survival and quality of life. It is generally accepted that patients without distant metastases but with otherwise inoperable or unresectable tumors are best treated with chemoradiotherapy. Such an approach provides long-term progression-free survival in a minority of patients and, in the vast majority, provides sustained relief of dysphagia, a major determinant of quality of life and an important outcome in these patients. Superiority of more aggressive regimens, based on addition of induction chemotherapy or brachytherapy to chemoradiotherapy regimens, is not proven and has not been studied in a controlled fashion. Metastatic Cancer Chemotherapy Similar to other advanced gastrointestinal cancers, systemic chemotherapy has a role in palliative treatment in patients with advanced, unresectable esophageal cancer, particularly those with distant metastatic disease. Although a large volume of published information exists in this area evaluating a large number of chemotherapy regimens, there is no convincing evidence that one particular regimen is superior to others. Based on randomized trials, it appears that a combination of cisplatinum, 5-fluorouracil (5-FU), and either epirubicin (an anthracyclic antineoplastic agent) or docitaxel is appropriate for first-line therapy. Capecitabine, the orally administered fluoropyrimidine, has attracted attention as a replacement agent for 5-FU infusion, because it eliminates the need for central venous access and an ambulatory infusion pump. Recent evidence suggests that capecitabine and oxaliplatin are as effective as 5-FU and cisplatin.206 Irinotecan-containing regimens have also been

studied recently in advanced esophageal and esophagogastric junctional cancers. Biological agents such as cetuximab (a monoclonal antibody against the EGF receptor), erlotinib (an orally active tyrosine kinase inhibitor), and bevacizumab (which targets VEGF) have been studied in a preliminary fashion as single agents or parts of combination chemotherapeutic regimens. In the recent Southwest Oncology Group (SWOG) trial, patients with unresectable or metastatic esophagogastric junctional adenocarcinoma had a reasonable response to erlotinib, with a very attractive toxicity profile.207 Experience with second-line chemotherapy in patients who fail initial chemotherapy regimens has been generally disappointing. Endoscopic Palliative Treatment Palliation of dysphagia, restoration or preservation of swallowing, and maintenance of nutritional status are the priority concern in planning palliative therapy and are the most important determinants of quality of life. Although systemic chemotherapy and radiation therapy (external beam or brachytherapy) can provide relief of dysphagia in the majority of patients, such relief is usually temporary. Therefore, most patients with unresectable esophageal or esophagogastric junctional cancers eventually become candidates for endoscopic palliative therapy for sustained relief of dysphagia. Endoscopic therapy for palliation of dysphagia can be divided into two categories: (1) techniques that ablate neoplastic tissue, and (2) techniques such as dilation and stent placement that displace neoplastic tissue. Endoscopic tissue ablative techniques include contact thermal, noncontact thermal, cytotoxic injection, and photodynamic therapies. The electrosurgical tumor probe (BICAP tumor probe, ACMI Circon, Santa Barbara, Calif) was often used in the past for thermal contact ablation of bulky esophageal tumors and could establish esophageal patency, with improvement in dysphagia, in up to 80% of patients. However, this technique has been largely abandoned because of an unacceptable rate of life-threatening complications, such as tracheoesophageal fistula, esophageal perforation, and delayed hemorrhage in up to 20% of patients. It is interesting to note this contact thermal ablation device is the precursor of the radiofrequency ablation balloon device, which has recently become a popular form of endoscopic therapy for patient with dysplastic Barrett’s esophagus. Laser Therapy.  Laser photoablation using neodymium: yttrium-aluminum-garnet (Nd:YAG), potassium titanyl phosphate (KTP), and argon lasers has been used extensively in the palliation of malignant dysphagia associated with esophageal cancers (Fig. 46-13). The bulk of the experience has been attained with the Nd:YAG laser in the noncontact mode. The endoscopist usually passes the laser fiber through the accessory channel of the endoscope and with the tip of the fiber placed about 10 mm from the tumor tissue, a combination of tissue vaporization and coagulation necrosis is achieved with a typical Nd:YAG laser setting of 40 to 100 W, used either in a pulsed or continuous mode. The ablation is typically started at the distal margin of the tumor, moving proximally in a circumferential manner. However, if there is complete luminal obstruction, the laser ablation is started at the proximal extent of the tumor in order to try to carefully bore through the obstructive tumor tissue. In the majority of patients, more than one session is required to establish reasonable luminal patency, and the procedure is typically repeated at 48- to 72-hour intervals. Chest pain, odynophagia, low-grade fever, and leukocytosis

Chapter 46  Tumors of the Esophagus

A

B

C

Figure 46-13.  Laser therapy of esophageal cancer. A focal fungating adenocarcinoma (A) can be debulked for palliation of dysphagia with a noncontact laser (B), resulting in tumor coagulative necrosis and vaporization, with reconstitution of the esophageal lumen (C).

A

B

may occur after the procedure. Follow-up endoscopy with a “touch-up” session in three to four weeks is often required, depending on the response to laser ablation. Reasonable improvement in dysphagia can be achieved in up to 70% of patients, which typically lasts three to six weeks and rarely beyond 12 months in absence of retreatment. Complication rates are approximately 4% in experienced hands, with a 1% procedural mortality rate. Polypoid, fleshy, noncircumferential lesions in the body of the esophagus are easier to treat with laser ablation. Anastomotic recurrences and tumor overgrowth in a previously placed endoprosthesis can also be treated by laser photoablation. Argon Plasma Coagulation.  APC has been used for palliation of malignant dysphagia, with limited success. Because the depth of tissue ablation achieved by APC is shallow, APC is somewhat ineffective in ablating large bulky tumor causing luminal obstruction. However, APC acting as a noncontact thermal device may have a role in controlling slow surface hemorrhage from friable tumor masses and in treating tumor in-growth in esophageal metal stents. Photodynamic Therapy.  PDT has a role in palliation of dysphagia in patients with advanced esophageal cancer (Fig. 46-14). Comparative trials suggest that PDT may be better than endoscopic laser ablation in terms of fewer perforations and better tumor response, particularly in proximal tumors and long tumors. Photosensitivity is a significant problem in up to 20% of patients. PDT has also been used for ablation of tumor in-growth after stent placement.

Figure 46-14.  A, Photodynamic therapy of high-grade dysplasia in Barrett’s esophagus. This patient was not a candidate for esophagectomy. B, Forty-eight hours after photodynamic therapy there is intense, circum­ ferential, superficial tissue destruction with sharply demarcated borders delineating the proximal extent of laser light exposure.

Cytotoxic Injection Therapy.  Endoscopic injection of cytotoxic agents such as mitomycin adsorbed onto activated carbon particles and injection of chemical sclerosants such as polidocanol, ethanol, and sodium morrhuate to induce tumor necrosis, volume reduction, and hemostasis has been reported in small case series, but with limited success. Injection therapy does not appear to have superiority over other ablative methods. Dilation.  Esophageal dilation provides temporary palliation of malignant dysphagia and is often used as an adjunctive treatment combined with other forms of palliation such as stent placement or ablative therapy. Most patients derive initial benefit from dilation therapy to a luminal diameter that allows advancement from a liquid to a soft diet (i.e., 12 mm). However, dilation can be complicated by perforation in up to 10% of cases. Moreover, the improvement in dysphagia is usually incomplete and temporary. Blind passage of Maloney dilators is not recommended in complex malignant strictures due to a higher risk of perforation. Polyvinyl dilators (such as the Savary-Gilliard or American type) and hydrostatic through-the-scope (TTS) balloons are commonly used. Esophageal Stent Placement.  Use of esophageal stents for palliation of malignant dysphagia has increased rapidly and compares favorably with other endoscopic palliative techniques. Self-expanding covered metallic stents are the most effective palliative therapy for tracheoesophageal fistula. However, palliative stent placement in patients with

765

766

Section V  Esophagus

Figure 46-15.  Stent therapy of an esophageal fistula. A, This patient with a circumferential esophageal carcinoma, previously treated with chemoradiotherapy, developed an esophagomediastinal fistula, seen infe­ riorly. B, Placement of a covered self-expanding metallic stent achieved long-term symptomatic palliation.

A

esophageal cancer is not always followed by improvement of nutritional parameters.208 The rigid plastic esophageal stents have been replaced by self-expanding metal stents (SEMS) because of their familiarity, ease of placement, improved palliation of dysphagia, and better safety profile. The covered SEMS are the only effective practical treatment available for management of esophagorespiratory fistula (Fig. 46-15). SEMS are typically placed as an outpatient procedure under conscious sedation. The patient is placed in left lateral or supine position and an initial endoscopic examination is carefully performed to estimate the lesion’s location, length, configuration, luminal diameter, and relationship to the upper esophageal sphincter and esophagogastric junction. A recent contrast esophagogram is often helpful to plan the placement of the SEMS. Esophageal dilation is often performed prior to stent placement, but may not be always recommended or required. Most endoscopists use fluo­ roscopic guidance for placement, although one study questioned the utility of such a practice.209 Accurate marking of the tumor margins in relation to the distal and proximal end of the SEMS, and consideration of the stent foreshor­ tening are critical to optimal placement. A post-placement barium contrast study should be performed routinely in patients with esophagorespiratory fistula to document effective sealing of the fistula. Most patients can be started on oral intake immediately after the endoscopic procedure. Patients should have individualized dietary instruction to prevent stent occlusion by food impaction. Antiemetic and antitussive drugs may be needed if the patient experiences significant postprocedure retching, coughing, or hiccups, both to relieve those symptoms and prevent stent dislodgement. A variety of esophageal SEMSs are commercially available, most of which are covered and available in different lengths and diameters. The Wallstent II (Microvasive, Boston Scientific, Inc., Natick, Mass) consists of a bilayer chromium alloy tubular mesh coated with a polyurethane sleeve between two mesh tubes. The stent can be recaptured and repositioned when up to 50% of the stent has been deployed. The Wallstent is delivered on an introducer catheter over a previously placed guidewire and deployed by withdrawing a translucent outer sheath under fluoroscopic guidance. The Flamingo Wallstent is a popular modification of this covered stent, with an exaggerated proximal flange to reduce the risk of distal stent migration. The Ultraflex stent (Microvasive, Boston Scientific, Inc., Natick, Mass) consists of a knitted nitinol wire tube. The

B

stent is constrained on the introducer catheter by a spiral retention suture. Unraveling the suture deploys the stent. One unique advantage of this stent is that models are available with either distal or proximal release systems, allowing precise deployment. The delivery device is 5 to 7 mm in diameter and allows insertion of the endoscope alongside the stent delivery device for concomitant endoscopic and fluoroscopic observation during placement and deployment. The Z-stent (Cook Medical, Inc., Winston-Salem, N.C.) is made of stainless steel wires shaped in a Z configuration and releases without shortening to a diameter of 18 mm. The version preferred by many endoscopists, because of a lower risk of stent migration, has uncovered segments in the proximal and distal sections. A model designed with a latex “wind-sock” extending from the distal opening of the stent (Dua stent) and intended to prevent gastroesophageal reflux has been marketed for esophagogastric junction and distal esophageal cancers, tumors in which the stent must bridge the esophagogastric junction. Another commercially available antireflux stent (FerX-Ella; Dr Karel Volenec, Ella CS, Hradec Králové, Czech Republic) has been used to minimize reflux of gastric contents. Clinical experience with these antireflux stents is limited, and randomized studies with small numbers of patients failed to show a difference in reflux symptoms or global quality of life.210,211 Another introduced version of the Z-stent (Evolution Controlled-Release Stent, Cook Medical, Inc., Winston-Salem, N.C.) allows more controlled stepwise release and better ability to recapture and reposition the stent. A covered hybrid esophageal metallic stent has been marketed (Allimax-E, Alveolus, Charlotte, N.C.). Some experts believe that covered metal stents are more likely to migrate compared with the bare metal uncovered stent. A doublelayered stent, the Niti-S stent (Taewong Medical, Seoul, South Korea), with an inner polyurethane layer (to prevent tumor in-growth) and an outer uncovered nitinol wire tube (to allow the mesh of the stent to embed itself in the esophageal wall to minimize stent migration) has been marketed; initial experience with this stent has been promising.212 The Polyflex (Boston Scientific, Inc., Natick, Mass) is a completely coated self-expanding nonmetallic (plastic) stent introduced for palliation of malignant and benign strictures. Advantages are its lower cost and potential of removability. Published series have shown the Polyflex stent to be safe and effective in relieving malignant dysphagia. In a randomized prospective trial of 101 patients with unresectable esophageal cancer, there was no difference

Chapter 46  Tumors of the Esophagus between the Polyflex and Ultraflex stents in palliating dysphagia. However, Polyflex stents had a higher rate of complications such as tumor overgrowth, hyperplastic granulomatous reaction, food bolus impaction, and stent migration and may be more technically difficult to place.213 There is little comparative information on the utility of different types of SEMS. A randomized trial comparing Flamingo, Ultraflex, and Z-stents failed to show superiority of one stent over another.214 In another randomized study comparing Ultraflex, Niti-S, and Polyflex stents, all three stents adequately palliated malignant dysphagia, although Polyflex stents had a higher rate of migration.215 The decision to place a particular type of stent should be individualized and is typically determined by assessment factors such as tumor length, location, and configuration; the patient’s per­ formance status and overall prognosis; plans for further palliative therapy, such as chemoradiotherapy; and local expertise. Available information reports good results with SEMS in palliating dysphagia and patients with esophagorespiratory fistula, with 70% to 100% reported success rates in sealing the fistula.216 Complications occur in 30% to 40% of patients with SEMS, although most are minor. Common complications include tumor in-growth or tumor overgrowth (5% to 20%), stent migration (10%), chest pain, procedure-related perforation, food bolus impaction, bleeding, foreign body sensation, and reflux esophagitis. Also, aspiration of gastric contents is commonly seen in patients with a stent placed across the esophagogastric junction. Brachytherapy Brachytherapy administered intraluminally by a nasoesophageal applicator has gained attention for palliation of malignant dysphagia. In a randomized trial, single-dose brachytherapy was more effective than metal stent placement for palliation of dysphagia in patients with unresectable esophageal cancer, although dysphagia improved more rapidly in patients who were treated with the Ultraflex SEMS. The brachytherapy group also had fewer complications than the stent group, and costs were similar.217 A prognostic score based on age, gender, WHO performance score, tumor length, and presence of metastasis has been suggested to identify patients with a poor prognosis in whom stent placement is at least equivalent to brachytherapy.218 In a Markov model analysis using current levels of reimbursements, brachytherapy (and not SEMS) was shown to be the most cost-effective intervention for palliation of malignant dysphagia.219 In a randomized trial of 53 patients with advanced esophageal cancer from China, a SEMS loaded with radioactive iodine (125I) seeds yielded better palliation of malignant dysphagia and better survival than a conventional SEMS.220 Nutritional Therapy Most patients with esophageal cancer are nutritionally compromised and need specific nutritional management for improvement of functional status before and after surgery, during chemoradiotherapy, and as an adjunct to other palliative measures. Success of endoscopic therapy in palliation of dysphagia does not necessarily translate into improved nutritional status, and many of these patients need establishment of enteral access by endoscopic, surgical, or radiologic procedures for maximizing nutritional intake. A surgical feeding jejunostomy should be performed in all patients undergoing esophagectomy; percutaneous endoscopic gastrostomy (PEG) is contraindicated in those patients who may be candidates for gastric pull-up surgery. In many patients with esophagectomy, or those with a

SEMS placed across the esophagogastric junction, feeding the patient through a PEG tube may lead to recurrent aspiration, and in these patients a PEG with a jejunal feeding tube extension or even direct percutaneous endoscopic jejunostomy may be preferred. Some patients with markedly compromised nutritional status may benefit for short-term parenteral nutrition, especially during the perioperative period (see Chapters 4 and 5). Analgesics Many patients with terminal advanced esophageal cancer may suffer from pain from tumor ulceration or local invasion or from metastatic disease and need narcotic analgesic medications for control of pain. A pain medicine or palliative care consultation may be appropriate in such cases. Multidisciplinary Care It is clear that appropriate management of patients with esophageal cancer needs input from many providers from different specialties. A multidisciplinary team approach with a tailored treatment plan for each patient with esophageal cancer has been shown to result in improved staging, lower operative mortality, and improved five-year survival when compared with a group of patients managed by surgeons who were working independently.221

OTHER MALIGNANT EPITHELIAL TUMORS (see Table 46-1) SQUAMOUS CELL CARCINOMA VARIANTS

A variant of squamous cell carcinoma, verrucous carcinoma, is characterized by an exophytic papillary growth and microscopically is composed of moderately differentiated squamous cells with a fibrous stroma with minimal cytologic atypia, prominent acanthosis, hyperkeratosis, swollen rete pegs, and coexisting inflammation. These tumors tend to be slow growing and have low metastatic potential, with a favorable prognosis. A commoner variant of squamous cell carcinoma is basaloid cancer, which is a bulky, ulcerative stricturing tumor, often with multiphasic differentiation and a worse prognosis. Another variant is the carcinosarcoma or polypoid cancer, with an exophytic growth pattern and aggressive biological behavior. These tumors, thought to arise from mesenchymal metaplasia of malignant squamous cells, may be solitary or multiple. They occur more commonly in men, appearing in middle or advanced age. Adenoid cystic carcinoma is a rare form of tumor presenting grossly as a submucosal nodule and with both ductal and basaloid cells and usually with a nonaggressive clinical course.

SMALL CELL CARCINOMA

The esophagus is the most common extrapulmonary site of small cell carcinoma, accounting for 1% to 4% of all esophageal neoplasms. Early metastasis is common with small cell tumors, most commonly to the periesophageal and mediastinal lymph nodes and to the liver. These patients have very poor prognosis, with only a 10% one-year survival. Surgical resection is appropriate if preoperative staging evaluation excludes extranodal metastasis.

MALIGNANT MELANOMA

Primary esophageal melanoma is rare and is estimated to account for 0.1% of esophageal tumors. Primary esophageal melanomas usually arise as polypoid tumors. When they grow large and ulcerate, bleeding and odynophagia are the

767

768

Section V  Esophagus presenting symptoms. Early metastasis to lymph nodes, liver, and lung is common, conferring a generally poor survival rate.

BENIGN EPITHELIAL TUMORS SQUAMOUS PAPILLOMA (Fig. 46-16) Squamous papillomas are usually small, white or pink, sessile or polypoid benign tumors that are histologically composed of finger-like projections of lamina propria covered by hyperplastic squamous epithelium. Mucosal biopsy or polypectomy is safe and usually diagnostic. Although their pathogenesis is not known, GERD and HPV have been implicated. Malignant transformation of a squamous papilloma is very rare. ADENOMA

True esophageal adenomatous polyps, arising within segments of Barrett’s esophagus, are rare. Although benign, they are considered dysplastic, with malignant potential. Endoscopic resection using standard snare or mucosectomy techniques is usually curative.

INFLAMMATORY FIBROID POLYP

Inflammatory fibroid polyps, also known as inflammatory pseudopolyps and eosinophilic granulomas are rare nonneoplastic polyps, often related to chronic inflammation from GERD. They are commonly seen in the distal esophagus or at the esophagogastric junction. Endoscopic resection is both diagnostic as well as curative.

MALIGNANT NONEPITHELIAL TUMORS (see Table 46-1)

LYMPHOMA

Primary lymphoma of the esophagus (usually of B cell origin) is extremely uncommon, except in patients with acquired immunodeficiency syndrome (AIDS). More

common is lymphomatous involvement of the esophagus due to extrinsic compression or direct invasion from involved mediastinal lymph nodes. Presenting symptoms include dysphagia and weight loss, and esophageal fistulas may occur. Therapy is dependent on symptoms, disease stage, and patient performance status and usually involves chemoradiotherapy.

SARCOMA

Malignant mesenchymal esophageal tumors are uncommon and represent about 5% of all gastrointestinal sarcomas. Leiomyosarcomas are the most common and can be difficult to distinguish from leiomyomas. Others include malignant gastrointestinal stromal tumors (see following and Chapter 30), rhabdomyosarcoma, fibrosarcoma, liposarcoma, fibrous histiocytoma, and choriocarcinoma. Tumor characteristics include spindle-shaped smooth muscle cells, high mitotic rates, local invasion, and, infrequently, distant metastasis. Most patients exhibit dysphagia. Superficial mucosal biopsy specimens are often nondiagnostic, and EUS-guided FNA is usually required for a definitive diagnosis. Cytologic needle aspiration specimens, however, cannot reliably exclude malignancy. Treatment is based on surgical resection, often followed by radiotherapy. Kaposi’s sarcoma (KS) has been reported in the esophagus usually concomitant with oral and skin lesions in patients with AIDS. When symptomatic, KS can present with dysphagia, odynophagia, and, rarely, gastrointestinal hemorrhage. In symptomatic patients, endoscopic laser ablation or injection sclerotherapy may be attempted for hemorrhage or obstruction.

METASTATIC CARCINOMA

Metastatic carcinoma to the esophagus is unusual, with melanoma and breast cancer the two cancers that most frequently metastasize to the esophagus. Radiographic and endoscopic studies typically demonstrate compression, with overlying normal mucosa. EUS is useful in distinguishing extrinsic from intrinsic involvement and in detecting local lymphadenopathy. EUS-guided FNA has proved useful in confirming the suspected diagnosis. Palliation of dysphagia, if present, is usually the only treatment available.

BENIGN NONEPITHELIAL TUMORS GASTROINTESTINAL STROMAL TUMOR

Figure 46-16.  Endoscopic image of an asymptomatic esophageal polyp with a long stalk. Endoscopic biopsy showed the polyp to be a squamous papilloma.

(see also Chapter 30) Gastrointestinal stromal tumors (GISTs) are rare benign tumors of the esophagus and account for approximately 1% of all GISTs (Fig. 46-17). Most esophageal GISTs arise from the distal third of the esophagus and are solitary. Most GISTs express CD117 antigen in contrast to leiomyomas or other spindle-cell tumors of the gastrointestinal tract, which are typically CD117-negative. Histopathologic evaluation reveals that these tumors are firm, round, gray or yellow, unencapsulated, and composed of spindle-shaped cells with cigar-shaped elongated nuclei. The risk of malignant transformation of GISTs is variable and depends on the size of the GIST and the mitotic rate. Lesions less than 2 cm in size with a mitotic rate of less than five per 50 highpower fields are considered to have a very low risk of malignancy. Approximately half of patients with esophageal GISTs have symptoms such as dysphagia, odynophagia, or reflux symptoms. Barium contrast studies, if performed, show a smooth crescent-shaped defect in the contour of the esophageal lumen, without mucosal abnormality. On endoscopy,

Chapter 46  Tumors of the Esophagus

A

B

C

the mucosa is usually intact and the mass appears as a rounded, smooth raised submucosal lesion protruding into the esophageal lumen. EUS is the most accurate tool for diagnosing GISTs and distinguishing them from other submucosal lesions. During EUS, GISTs are seen to be typically arising from the fourth wall layer (the muscularis propria), are hypoechoic and homogeneous, and have sharply demarcated margins. GISTs arising from muscularis mucosae have been reported. EUS-guided FNA and CD117 (c-KIT) immunostaining of the aspirate establishes the diagnosis, but FNA may not be necessary if typical EUS findings are present. It should be noted that neither EUS nor aspiration cytology can accurately distinguish benign from malignant GISTs preoperatively. Management of an esophageal GIST should take into consideration the presence or absence of symptoms and the risk of eventual malignant transformation. Most small, incidentally detected GISTs can be followed conservatively. Surgery should be considered in symptomatic patients, in those in whom the diagnosis is uncertain (particularly if there is suspicion of malignant transformation), and in patients in whom the lesion shows interval growth during surveillance endoscopy. Surgical excision, particularly using a minimally invasive approach, has been reported. Endoscopic resection may be feasible in a minority of patients with small GISTs arising from the muscularis mucosae. Administration of orally active tyrosine kinase inhibitors such as imatinib mesylate (Gleevec) and sunitinib (Sutent) has become an option in patients with resected GISTs as an adjuvant therapy and in patients with unrespectable GIST as a cytoreductive therapy (see Chapter 30).222

GRANULAR CELL TUMOR

Figure 46-17.  Esophageal gastrointestinal stromal tumor (GIST). A, Endoscopic image of a submucosal esophageal mass. B, Endoscopic ultrasonography demonstrates the lesion, which appears as a hypoechoic mass (arrows) arising from the muscularis propria. C, The GIST was removed by minimally invasive video-assisted thoracoscopic surgery. (Courtesy, Dawn Jarosjewski, MD, Scottsdale, Ariz.)

Granular cell tumors are submucosal neoplasms that are thought to originate from cells of neural origin. The esophagus is one of the more common gastrointestinal sites for these tumors. Endoscopically they appear as submucosal nodules, pinkish tan, and rubbery. On EUS they are hypoechoic to isoechoic and arise within the submucosal layer. Diagnostic tissue usually can be confirmed with endo-

scopic biopsy samples obtained using the bite-on-bite technique. Management options include observation, endoscopic resection, and surgery. Granular cell tumors larger than 4 cm or those that exhibit growth should be considered potentially malignant.

FIBROVASCULAR POLYP

Large, benign fibrovascular polyps occur most commonly on the upper third of the esophagus, near the cricopharyngeus muscle (Fig. 46-18). They may contain a mixture of fibrovascular tissue, adipose cells, and stroma, but are uniformly covered by squamous epithelium. Although most are asymptomatic, bizarre symptoms of polyp regur­ gitation and asphyxiation have been reported. Barium esophagography and endoscopy are usually sufficient for diagnosis, but MRI can help determine the origin of these polyps and plan for surgery. The latter is recommended for large symptomatic polyps. EUS-assisted endoscopic snare resection of fibrovascular polyps may be feasible for smaller lesions.

HAMARTOMA

Hamartomas of the esophagus are uncommon and are benign developmental tumors consisting of disorganized and excessive focal growth of mature normal cells. On pathologic examination, the mass can contain various elements including cartilage, bone and bone marrow, adipose and fibrous tissue, and smooth and skeletal muscle. Esophageal hamartomas may grow to a large size as long pedunculated polyps. Most occur in the upper esophagus and show obstructive symptoms and, less commonly, present with hematemesis. Surgical excision is required for symptomatic lesions.

HEMANGIOMA

Hemangiomas are extremely uncommon benign esophageal tumors and include cavernous hemangiomas (the vast majority) and capillary hemangiomas. Hemangiomas appear nodular, are blue to red, and are soft and pliable when probed with a closed biopsy forceps. Classically, pressure from the forceps causes the lesion to blanch. Common

769

770

Section V  Esophagus

Figure 46-18.  Fibrovascular polyp of the eso­ phagus. A, Barium esophagogram showing a lobulated filling defect in the upper third of the esophagus. B, Closer view, showing polyp with a long stalk. (Courtesy Mark Feldman, MD, Dallas.)

A

symptoms include hemorrhage and dysphagia. The differential diagnosis includes KS. Therapy has traditionally been surgery, but endoscopic therapies may be considered.

LIPOMA

Esophageal lipomas, the most uncommon of all gastrointestinal lipomas, are encapsulated lesions composed of welldifferentiated adipose tissue, generally arising in the submucosa and often growing intraluminally in a pedunculated fashion. Lipomas with long pedicles can produce laryngeal obstruction and asphyxiation. On endoscopy, lipomas classically have smooth and normal-appearing overlying mucosa and a yellowish tint; they may have superficial central ulceration. Typically, when grasped with biopsy forceps, these lesions tend to “tent” and when palpated with a closed biopsy forceps, they indent or “cushion.” EUS is diagnostic and classically reveals a homogeneous hyperechoic lesion with smooth outer margins, arising from the submucosa. Endoscopic or surgical resection is recommended for symptomatic lesions.

KEY REFERENCES

Atherfold PA, Jankowski JA. Molecular biology of Barrett’s cancer. Best Pract Res Clin Gastroenterol 2006; 20:813-27. (Ref 98.) Coorley DA, Kerlikowske K, Verma R, Buffler P. Protective association of aspirin/NSAIDs and esophageal cancer: A systematic review and meta-analysis. Gastroenterology 2003; 124:47-56. (Ref 62.) Das A, Singh V, Fleischer DE, Sharma VK. A comparison of endoscopic treatment and surgery in early esophageal cancer: An analysis of surveillance epidemiology and end results data. Am J Gastroenterol 2008; 103:1340-5. (Ref 186.) Gebski V, Burmeister B, Smithers BM, et al. Survival benefits from neoadjuvant chemoradiotherapy or chemotherapy in oesophageal carcinoma: A meta-analysis. Lancet Oncol 2007;8:226-34. (Ref 203.)

B

Guo JH, Teng GJ, Zhu GY, et al. Self-expandable esophageal stent loaded with 125I seeds: Initial experience in patients with advanced esophageal cancer. Radiology 2008; 247:574-81. (Ref 220.) Lagergren J, Bergstrom R, Lindgren A, Nyren O. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999 18; 340:825-31. (Ref 40.) Larsson SC, Giovannucci E, Wolk A. Folate intake, MTHFR polymorphisms, and risk of esophageal, gastric, and pancreatic cancer: A meta-analysis. Gastroenterology 2006; 131:1271-83. (Ref 17.) Ngamruengphong S, Sharma VK, Das A. Diagnostic yield of methylene blue chromoendoscopy for detecting specialized intestinal metaplasia and dysplasia in Barrett’s esophagus: A meta-analysis. Gastrointest Endosc 2009; 69:1021-8. (Ref 121.) Rastogi A, Puli S, El-Serag HB, et al. Incidence of esophageal adeno­ carcinoma in patients with Barrett’s esophagus and high-grade dysplasia: A meta-analysis. Gastrointest Endosc 2008; 67:394-8. (Ref 52.) Stein HJ, Feith M, Bruecher BL, et al. Early esophageal cancer: pattern of lymphatic spread and prognostic factors for long-term survival after surgical resection. Ann Surg 2005; 242:566-73. (Ref 115.) Steyerberg EW, Homs MY, Stokvis A, et al. Stent placement or brachytherapy for palliation of dysphagia from esophageal cancer: A prognostic model to guide treatment selection. Gastrointest Endosc 2005; 62:333-40. (Ref 218.) Takubo K, Aida J, Sawabe M. Early squamous cell carcinoma of the oesophagus: The Japanese viewpoint. Histopathology. 2007; 51:73342. (Ref 102.) Wang KK, Sampliner RE, Practice Parameters Committee of the American College of Gastroenterology. Updated guidelines 2008 for the diagnosis, surveillance and therapy of Barrett’s esophagus. Am J Gastroenterol 2008; 103:788-97. (Ref 44.) Westerterp M, van Westreenen HL, Reitsma JB, et al. Esophageal cancer: CT, endoscopic US, and FDG PET for assessment of response to neoadjuvant therapy—Systematic review. Radiology 2005; 236:84151. (Ref 173.) Wolfsen HC, Crook JE, Krishna M, et al. Prospective, controlled tandem endoscopy study of narrow band imaging for dysplasia detection in Barrett’s esophagus. Gastroenterology 2008; 135:24-31. (Ref 128.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

47 Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum M. Gaith Semrin and Michael A. Russo

CHAPTER OUTLINE Embryology and Anatomy of the Stomach  773 General Considerations  773 Vascular Supply and Drainage; Lymphatic Drainage  775 Gastric Innervation  776 Tissue Layers of the Stomach  776 Microscopic Anatomy  776 Anatomy of the Duodenum  779 General Considerations  779

EMBRYOLOGY AND ANATOMY OF THE STOMACH GENERAL CONSIDERATIONS

The stomach, as a J-shaped dilation of the alimentary canal, is continuous with the esophagus proximally and the duodenum distally. It functions primarily as a reservoir to store large quantities of recently ingested food, thus allowing intermittent feedings, initiating the digestive process, and releasing its contents in a controlled fashion downstream to accommodate the much smaller capacity of the duodenum. The stomach volume ranges from about 30 mL in a neonate to 1.5 to 2 L in adulthood. The stomach is recognizable in the fourth week of gestation as a dilation of the distal foregut (Fig. 47-1).1 As the stomach enlarges, the dorsal aspect grows more rapidly than the ventral aspect, thus forming the greater curvature. Additionally, during the enlargement process the stomach rotates 90 degrees around its longitudinal axis, orienting the greater curvature (the dorsal aspect) to the left and the lesser curvature (ventral aspect) to the right. The combined effects of rotation and ongoing differential growth result in the stomach lying transversely in the mid and left upper abdomen. The events also explain the vagal innervation of the stomach: the right vagus nerve innervating the posterior stomach wall (the primordial right side) and the left vagus nerve innervating the anterior wall (the primordial left side).

Vascular Supply and Drainage; Lymphatic Drainage  779 Duodenal Innervation  779 Microscopic Anatomy  779 Congenital Anomalies of the Stomach and Duodenum  780 Gastric Anomalies  780 Duodenal Anomalies  785

The final location of the stomach is variable owing in part to its two-point fixation at the gastroesophageal and gastroduodenal junctions, allowing for considerable mobility. The gastroesophageal junction generally lies to the left of the 10th thoracic vertebral body, 1 to 2 cm below the diaphragmatic hiatus. The gastroduodenal junction lies at L1 and generally to the right of the midline in the recumbent fasted individual. The gastroduodenal junction of a distended upright adult may be considerably lower. The left-sided and caudal greater curvature may extend below the umbilicus depending on the degree of distention, position, and gastric peristaltic phase. The greater curvature forms the left lower stomach border, whereas the lesser curvature forms the right upper border. Posteriorly, portions of the pancreas, transverse colon, diaphragm, spleen, and apex of the left kidney and adrenal gland bound the stomach. The posterior wall of the stomach actually comprises the anterior wall of the omental bursa, or lesser peritoneal sac. Anteriorly the liver bounds the stomach, whereas the inner aspect of the anterior abdominal wall bounds the anterior left lower aspect. The stomach is completely invested by peritoneum, except for a small bare area at the gastroesophageal junction. This peritoneum passes as a double layer from the lesser curvature to the liver as the gastrohepatic portion of the lesser omentum and then hangs down from the fundus and greater curvature as the greater omentum, extending to the

773

774

Section VI  Stomach and Duodenum Pharynx (cranial part of foregut) Celiac trunk

Aortic arches

Septum transversum Spinal cord Superior mesenteric artery (SMA) Inferior mesenteric artery

Brain

Midgut Heart

Cloaca (caudal part of hindgut)

A

Yolk stalk Dorsal mesentery Esophagus

Dorsal abdominal wall

Stomach

Dorsal aorta

Spleen Dorsal mesogastrium

Proximal part of stomach

Celiac trunk Foregut artery (SMA) Ventral mesogastrium

Dorsal artery

Pancreas

C

B Aorta

Duodenum

Esophagus

Pancreas

Posterior abdominal wall

Spleen Greater curvature of stomach

Liver

Omental foramen Duodenum

Omental bursa (area indicated by broken line)

D

Stomach Right gastro-omental artery

E

Greater omentum

Stomach

Dorsal aorta

Level of section on right

Omental bursa

Omental bursa Omental foramen

Greater omentum

F

Greater omentum

Dorsal mesogastrium

Stomach

G

Plane of section on right

Dorsal abdominal wall Greater omentum

Figure 47-1.  Development of the stomach and duodenum and formation of the omental bursa (lesser sac) and greater omentum. A, Median section of a 28-day embryo. B, Anterolateral view of a 28-day embryo. C, Embryo about 35 days old. D, Embryo about 40 days old. E, Embryo about 48 days old. F, Lateral view of the stomach and greater omentum of an embryo at about 52 days. The transverse section shows the omental foramen and omental bursa. G, Sagittal section showing the omental bursa and greater omentum. The embryology of the duodenum is discussed further in Chapters 55 and 96. (From Moore KL, Persaud TVN. The developing human. 7th ed. Philadelphia: WB Saunders; 2003. p 258.)

Chapter 47  Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum transverse colon (as the gastrocolic ligament), spleen (as the gastrosplenic ligament), and diaphragm (as the gastrophrenic ligament). The stomach is divided into four regions that can be defined by anatomic or histologic landmarks (Fig. 47-2).2 Anatomically the cardia is a small ill-defined area of the stomach immediately adjacent to its junction with the esophagus. This region of the stomach has been the focus of intense investigation. Controversy exists as to the nature, location, extent, and even existence of cardiac mucosa. The fundus projects upward, above the cardia and gastroesophageal junction. This dome-shaped area of the stomach is its most superior portion and is in contact above with the left hemidiaphragm and to the left with the spleen. The body, or corpus, the largest portion of the stomach, is located immediately below and continuous with the fundus. The incisura angularis, a fixed, sharp indentation two thirds of the distance down the lesser curvature, marks the caudal aspect of the gastric body (Fig. 47-3). The gastric antrum Lower esophageal sphincter Pylorus

Fundus Oxyntic gland mucosa

Body

Antrum Pyloric gland mucosa Figure 47-2.  Anatomic regions of the stomach. The line is drawn from the incisura angularis along the lesser curvature to an indistinct border between the gastric body and antrum along the greater curvature. (From Johnson LR. Gastrointestinal physiology. 6th ed. St Louis: Mosby; 2001. p 76.)

Figure 47-3.  Film from an upper gastrointestinal series demonstrating the incisura angularis (arrow) on the distal lesser curvature. (Courtesy of James W. Weaver, MD.)

extends from its indistinct border with the body to the junction of the pylorus with the duodenum. These gross anatomic landmarks correspond roughly with the mucosal histology because antral mucosa (pyloric gland mucosa) actually extends from an area on the lesser curvature somewhat above the incisura. The pylorus (pyloric channel) is a tubular structure joining the duodenum to the stomach and contains the palpable circular muscle, the pyloric sphincter. The pylorus is somewhat mobile owing to its enclosure between the peritoneum of the greater and lesser omenta but is generally located 2 cm to the right of midline at L1. Corresponding motor and secretory functions of these regions of the stomach are discussed in detail in Chapters 48 and 49.

VASCULAR SUPPLY AND DRAINAGE; LYMPHATIC DRAINAGE The arterial blood supply to the stomach is derived from branches of the celiac artery—common hepatic, left gastric, and splenic arteries—that form two arterial arcades situated along the lesser curvature and the lower two thirds of the greater curvature. The lesser curvature is supplied from above by the left gastric artery and from below by the right gastric artery, a branch of the common hepatic artery or gastroduodenal artery (which is a branch of the common hepatic artery). The greater curvature below the fundus is supplied from above by the left gastroepiploic artery (a branch of the splenic artery) and from below by the right gastroepiploic artery (a branch of the gastroduodenal artery). The right and left gastroepiploic arteries usually terminate by anastomosing, thus completing the greater curvature arterial arcade; occasionally they end without anastomosis. The arterial supply to the gastric fundus and left upper aspect of the greater curvature is via the short gastric arteries, which arise from the splenic artery. The venous drainage of the stomach generally accompanies the arterial supply, emptying into the portal vein or one of its tributaries, the splenic or superior mesenteric veins. The left and right gastric veins drain the lesser curvature of the stomach. The left gastric vein is also known as the coronary vein. The right and left gastroepiploic veins drain the inferior aspect and a portion of the greater curvature of the stomach. The right gastroepiploic vein and several more distal veins become the gastrocolic veins, eventually ter­ minating in the superior mesenteric vein. There is no gas­ troduodenal vein. The left gastroepiploic vein becomes the splenic vein and later receives the short gastric veins, thus draining the fundus and upper great curvature of the stomach. Most of the lymphatic drainage of the stomach eventually reaches the celiac nodes after passing through intermediary lymph nodes. Lymphatic channels anastomose freely in the gastric wall, with lymphatic flow directed through one-way valves into one of four groups of nodes. The inferior gastric region drains into subpyloric and omental nodes, then the hepatic nodes, terminating in the celiac nodes. The splenic or superior aspect of the greater curvature lymph initially drains into pancreaticosplenic nodes and then into celiac nodes. The superior gastric or lesser curvature region lymph drains into the left and right gastric nodes adjacent to their respective vessels and terminates in the celiac nodes. The hepatic or pyloric portion of the lesser curvature lymph drains into the suprapyloric nodes, then into the hepatic nodes, and finally into the celiac nodes.

775

776

Section VI  Stomach and Duodenum Microvilli

GASTRIC INNERVATION The autonomic innervation of the stomach stems from the sympathetic and parasympathetic nervous systems delivered via a complex tangle of nerves coursing along the visceral arteries. The gastric sympathetic innervation is derived from preganglionic fibers arising predominantly from T6 to T8 spinal nerves, which synapse within the bilateral celiac ganglia to neurons whose postganglionic fibers course through the celiac plexus along the vascular supply of the stomach. Accompanying these sympathetic nerves are afferent paintransmitting fibers from the stomach and motor fibers to the pyloric sphincter. The parasympathetic innervation is via the right and left vagus nerves, which form the distal esophageal plexus, and gives rise to the posterior and anterior vagal trunks near the gastric cardia. The trunks contain preganglionic parasympathetic fibers, as well as afferent fibers from the viscera. Both trunks give rise to celiac and hepatic branches before continuing on within the lesser omentum slightly to the right of the lesser curvature as the anterior nerve of Latarjet and the posterior nerve of Latarjet. These nerves give rise to multiple gastric branches to the stomach wall, where the preganglionic fibers synapse with the ganglion cells in the submucosal (Meissner’s) and myenteric (Auerbach’s) plexuses. From these plexuses, postganglionic fibers are distributed to secretory components including cells and glands and to motor components such as muscle.

TISSUE LAYERS OF THE STOMACH The luminal surface of the gastric wall forms thick, longitudinally oriented folds, or rugae, that flatten with dis­ tention. Four layers make up the gastric wall: mucosa, submucosa, muscularis propria, and serosa. Mucosa lines the gastric lumen, appearing as a smooth, velvety bloodfilled lining. The mucosa of the cardia, antrum, and pylorus is somewhat paler than that of the fundus and body. It is within the gastric mucosa that most of the functional secretory elements of the stomach are located (see Chapter 49). The submucosa, immediately deep to the mucosa, provides the dense connective tissue skeleton of collagen and elastin fibers. Lymphocytes, plasma cells, arterioles, venules, lymphatics, and the submucosal plexus are also contained within the submucosa. The third tissue layer, the muscularis propria, is a combination of three muscle layers: inner oblique, middle circular, and outer longitudinal. The inner oblique muscle fibers course over the gastric fundus, covering the anterior and posterior aspects of the stomach wall. The middle circular fibers encircle the body of the stomach, thickening distally to become the pyloric sphincter. The outer longitudinal muscle fibers course primarily along the greater and lesser curvatures of the stomach. The final layer of the stomach is the transparent serosa, a continuation of the visceral peritoneum.

MICROSCOPIC ANATOMY The gastric mucosal surface is composed primarily of a simple layer of columnar epithelial cells 20 to 40  mm in height. These surface mucous cells (Fig. 47-4), which are similar throughout the stomach, contain basally located

Junctional complex Mucous granules

Rough endoplasmic reticulum

Golgi complex

Nucleus Macula adherens Basement lamina

Mitochondria

Figure 47-4.  Schematic representation of a surface mucous cell.

nuclei, prominent Golgi stacks, and dense cytoplasm with especially apically dense mucin-containing membranebound granules. The cells secrete mucus in granules that are released via exocytosis, apical expulsion, and cell exfoliation. The primary role of mucus, along with bicarbonate, is luminal cytoprotection from “the elements”: acid, pepsin, ingested substances, and pathogens. Cellular renewal time for a gastric surface mucous cell is approximately three days. The surface epithelial lining is invaginated by gastric pits, or foveolae, that provide the gastric glands access to the gastric lumen, with a ratio of one pit to four or five gastric glands. The gastric glands of different anatomic regions of the stomach are lined with different types of specialized epithelial cells, allowing for differentiation of these regions by type of gastric gland (see Fig. 47-2). The first region, the cardia, is a small transition zone from esophageal squamous epithelium to gastric columnar epithelium. The cardia has been a controversial histologic area of discussion, with theories suggesting that its presence is pathologic. However, recent observations concluded that cardiac mucosa develops during gestation and is present at birth.3 The cardiac glands have a branched and tortuous configuration and are populated by mucous, endocrine, and undifferentiated cells. There is a gradual transition from cardiac glands to the second region, the acid-secreting segment of the stomach. This region encompasses the gastric fundus and body and contains the parietal (or oxyntic or fundic) glands. Parietal, chief (also known as peptic), endocrine, mucous neck, and undifferentiated cells compose the oxyntic glands. The final region, corresponding to the antrum and pylorus, contains the pyloric glands, composed of endocrine cells, including gastrin-producing G cells and mucous cells. By far the most numerous and distinctive gastric glands are the oxyntic glands (Fig. 47-5), responsible for the secretion of acid, intrinsic factor, and most gastric enzymes. These fairly straight and simple tubular glands are closely associated in the areas of gastric fundus and body. A typical gland is subdivided into three areas: the isthmus (where surface mucous cells predominate), the neck (where parietal and mucous neck cells predominate),

Chapter 47  Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum Gastric gland

MSC

MNC

ECL cell

D cell PC

CC

Figure 47-5.  Schematic representation of an oxyntic (gastric) gland, with mucous surface cells (MSC), mucous neck cells (MNC), enterochromaffinlike (ECL) cells, somatostatin containing D cells, parietal cells (PC), and chief cells (CC). (From Lloyd KCK, Debas T. The peripheral regulation of gastric acid secretion. In Johnson LR, et al, editors. Physiology of the gastrointestinal tract, vol 2. 3rd ed. New York: Lippincott-Raven; 1994.)

and the base (where chief cells predominate, along with some parietal and mucous neck cells). Endocrine cells, somatostatin-containing D cells, and histamine-secreting enterochro­maffin-like (ECL) cells are scattered throughout the oxyntic epithelium. The principal cell type of the oxyntic gland is the parietal cell (Fig. 47-6), responsible for the oxyntic mucosal secretion of 3 × 106 hydrogen ions per second, at a final hydrochloric acid (HCl) concentration of around 150  mmol/L. Parietal cells bulge into the lumina of the oxyntic glands and, as the primary hydrogen secretors, have ultrastructural characteristics different from other gastric cells: large mitochondria, microvilli lacking in glycocalyx, and a cytoplasmic canaliculi system in contact with the lumen. In the nonsecreting parietal cell, a cytoplasmic tubulovesicular system predominates and short microvilli line the apical canaliculus. In the secreting state, the tubulovesicular system disappears, leaving an extensive system of intracellular canaliculi containing long microvilli. Mitochondria occupy approximately 30% to 40% of the secreting parietal cell volume, providing energy required for acid secretion across apical microvilli (see Fig. 47-6). The so-called proton pump—the H+,K+-ATPase—resides in the apical microvillus membrane, as does carbonic anhydrase. The apical H+,K+ATPase functions as the proton translocator in gastric acid secretion (see Chapter 49). Acid secretion begins within 5 to 10 minutes of stimulation. Additionally, parietal cells are the site of intrinsic factor secretion via membraneassociated vesicle transport. Closely associated with parietal cells are mucous neck cells, which appear singly, close to parietal cells or in groups of two or three in the oxyntic gland neck or isthmus. Mucous neck cells differ from their surface counterparts in their synthesis of acidic, sulfated mucus rather than the

Secretory canaliculus Tubulovesicles Microvilli

Nucleus Mitochondria

A

B

Figure 47-6.  Parietal cell. A, Electron photomicrograph. B, Schematic. (A and B, from Johnson LR. Gastrointestinal physiology. 6th ed. St Louis: Mosby; 2001. pp 78, 79.)

777

778

Section VI  Stomach and Duodenum neutral mucus. Additionally, mucous neck cells have basal nuclei and larger mucous granules around the nucleus rather than apically located granules. Function of the two cell types appears different in that surface mucous cells are cytoprotective, whereas the mucous neck cell functions as a stem cell precursor for surface mucous, parietal, chief, and endocrine cells. Chief cells, also known as zymogen cells, predominate in deeper layers of the oxyntic glands. These pyramidshaped cells play a role in synthesis and secretion of pepsinogens I and II. The cytoplasm of chief cells has prominent basophilic staining owing to abundance of ribosomes; these ribosomes are either free in cytoplasm or in association with an extensive endoplasmic reticulum system. Zymogen granules lie in the apical cytoplasm; their contents are released into the gastric lumen following fusion of the limiting membrane of the granule with the luminal membrane. Once in the lumen, pepsinogens are converted to pepsin. A variety of endocrine, or enteroendocrine, cells are scattered among the cells of the oxyntic glands. These cells vary in location, being either open or closed relative to the gastric lumen. Open endocrine cells have apical membranes containing receptors; these open cells discharge their contents by basilar exocytosis into the bloodstream, thus exerting an endocrine effect. The closed endocrine cells contain several processes that terminate near its target cells, constituting the so-called paracrine effect. The oxyntic gland model of the closed cell is the D cell, which secretes somatostatin via long processes reaching ECL, parietal, and chief cells. Enteroendocrine cell types have also been classified by their granular staining with silver or chromium. Those cells containing granules that reduce silver without pretreatment are called argentaffin cells. Argentaffin cells that stain with potassium dichromate are enterochromaffin (EC) cells; most of these contain serotonin. Cells with granules staining with silver only in the presence of a reducing agent are called argyrophilic, or ECL cells. Located primarily in the oxyntic glands, ECL cells are the only enteroendocrine cells containing histamine.

The final region of the stomach encompasses the antrum and pylorus and contains extensively coiled antral glands composed of endocrine and epithelial cells. The epithelial cells are predominantly mucous cells, and there are small numbers of pepsinogen II–secreting oxyntic cells. Although also small in number, gastrin-secreting (G) cells play a vital physiologic role and are the prototype of the open enteroendocrine cell. These cells, which occur either singly or in small clusters in the mid- to deep sections of antral glands (Fig. 47-7A), contain a basilar cytoplasm densely packed with gastrin-containing secretory granules (see Fig. 47-7B). Gastrin release is stimulated by gastric distention, vagal stimulation, dietary amino acids, and peptide, with rapid appearance of the hormone into the bloodstream in the postprandial period (see Chapter 49). The apical or luminal surface of the G cell is narrowed into small microvilli thought to contain receptors responsible for amino acid and peptide stimulation of gastrin release. Significant quantities of gastrin are also secreted into the gastric lumen; gastrin is a known gastric growth and differen­ tiation factor, mediated through upregulation of heparinbinding epidermal-like growth factor (HB-EGF) in gastric parietal cells.4,5 Antral enteroendocrine D cells found in close association with G cells manufacture somatostatin, a potent inhibitor of gastrin secretion. The D cells are also present in small numbers in oxyntic glands. Somatostatin is thought to inhibit acid secretion through paracrine (direct action on ECL and perhaps parietal cells or indirect action on G cells) or endocrine effects (direct action on parietal cells) (see Chapter 49 for more details). Immediately deep to the basement membrane of the gastric mucosa epithelial layer lies the lamina propria, which contains a variety of leukocytes (polymorphonuclear leukocytes, plasma cells, lymphocytes, eosinophils), mast cells, fibroblasts, and endocrine-like cells. A few lymphatic channels course through the lamina propria. Additionally, the mucosal capillary plexus lies in the lamina propria and forms a venule plexus, which communicates with the venules in the muscularis mucosa. These venules eventually empty into veins of the submucosa.

Microvilli Junctional complex

Desmosome Secretory granules

Rough endoplasmic reticulum

Desmosome

Mitochondria

A

B

Figure 47-7.  Gastrin (G) cells. A, Scattered G cells (pink) are evident in pyloric glands on this photomicrograph (immunoperoxidase stain). B, Schematic representation of a G cell.

Chapter 47  Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum ANATOMY OF THE DUODENUM GENERAL CONSIDERATIONS

The duodenum is the most proximal section of the small intestine and is continuous proximally with the pylorus and distally with the jejunum. It forms a C-shaped loop around the head of the pancreas. In the adult the length of the duodenum is approximately 30 cm (12 inches, hence its name duodenum) and is subdivided into four sections (commonly termed the first, second, third, and fourth parts), whose borders are delineated by angular course changes. The first part of the duodenum is about 5  cm in length and courses rightward, upward, and backward from the pylorus. The proximal portion of the first part of the duodenum is also referred to as the duodenal bulb or cap. Loosely attached to the liver by the hepatoduodenal portion of the lesser omentum, the first part moves in response to movement by the pylorus. The gastroduodenal artery, the common bile duct, and the portal vein lie posterior, whereas the gallbladder lies anterior to the first part of the duodenum. The second part of the duodenum is 7 to 10  cm in length, coursing downward parallel and in front of the hilum of the right kidney and to the right in contact with the pancreatic head. Slightly inferior to the midpoint of the second part of the duodenum on the posteromedial wall, the nipple-like major duodenal papilla marks the location of the ampulla of Vater, through which the pancreatico­ biliary ducts empty into the duodenum. On the same wall 2  cm proximal to the major papilla, there may be a minor duodenal papilla that forms the opening for the accessory pancreatic duct. The third part of the duodenum is about 10 cm in length and courses transversely from right to left, crossing the midline anterior to the spine, aorta, and inferior vena cava. The superior mesenteric artery and vein course anterior to the third part of the duodenum generally to the right of midline. The fourth and final section of the duodenum is 5  cm long and courses upward to the left of the aorta to reach the inferior border of the pancreas. The junction between the duodenum and the jejunum (duo­denojejunal flexure) is fixed posteriorly by the ligament of Treitz. The duodenal wall is composed of outer longitudinal and inner circular muscle layers. As is the case with the remainder of the small intestine, the luminal surface is lined with mucosa, forming circular folds known as the plicae circulares or valvulae conniventes. An exception to this is the duodenal bulb, distinguished radiographically and endoscopically by its smooth, featureless mucosa. The first few centimeters of the duodenum are shrouded by anterior and posterior elements of the peritoneum. The remainder of the duodenum lies posterior to the peritoneum and thus is retroperitoneal. The duodenum develops during the fourth week of gestation from the distal foregut, proximal midgut, and the adjacent splanchnic mesenchyme. The junction of the foregut and midgut occurs in the second part of the duo­ denum, slightly distal to the major papilla. As the stomach rotates, so too does the duodenum, thus developing a C-shaped configuration. During weeks five and six of embryologic development, the duodenal lumen is temporarily obliterated owing to proliferation of its mucosal lining. During the following weeks, luminal vacuolization and degeneration of some of the proliferating cells result in recanalization of the duodenal lumen. Epithelium and glands develop from embryonic endoderm, whereas connective tissue, muscle, and serosa are derived from mesoderm.

VASCULAR SUPPLY AND DRAINAGE; LYMPHATIC DRAINAGE The arterial supply to the duodenum is based on its embryonic origin in that branches of the celiac trunk (as derived from foregut) supply the proximal duodenum, whereas the distal duodenum (as derived from midgut) is supplied by branches of the superior mesenteric artery. From the celiac trunk arises the common hepatic artery, from which arises the gastroduodenal artery. The gastroduodenal artery in turn branches into the superior pancreaticoduodenal artery, which gives off anterior and posterior branches to the duodenum. These branches anastomose with analogous branches of the inferior pancreaticoduodenal artery, a branch of the superior mesenteric artery. The venous drainage corresponds to the arterial supply, with the superior pancreaticoduodenal veins coursing between the duodenum and pancreatic head to enter the portal vein. Likewise, both anterior and posterior inferior pancreaticoduodenal veins empty into either a jejunal vein or directly into the superior mesenteric vein. The duodenal lymphatic drainage also corresponds to the vascular supply. Small anterior and posterior duodenal lymph channels drain into the pancreaticoduodenal nodes. From these nodes lymph drains superiorly into the hepatic nodes or inferiorly into superior mesenteric nodes located at the origin of the superior mesenteric artery.

DUODENAL INNERVATION As in the case in the stomach, duodenal innervation is provided by the sympathetic and parasympathetic nervous systems. The preganglionic sympathetic nerves course through the celiac and superior mesenteric ganglia, with postganglionic neurons entering the duodenal intramural plexuses. Afferent fibers accompany the sympathetic neurons, primarily carrying fibers for visceral pain sensation. Parasympathetic fibers, supplied by the hepatic branch of the anterior vagus nerve and the mesenteric nerves, synapse with Meissner’s and Auerbach’s plexuses in the duodenal wall.

MICROSCOPIC ANATOMY Microscopically the duodenum differs dramatically from the gastric mucosa with the change from gastric glands and pits to a mucosa lined with villi surrounded by crypts of Lieberkühn and submucosa with characteristic Brunner’s glands. A single layer of epithelial cells provides the interface between the duodenal lumen and mucosa in the areas of both villi and crypts. Deep to this epithelial layer are contained absorptive cells, Paneth cells (which secrete lysozyme and other host defense factors), mucous cells, and endocrine cells. The villi in the proximal duodenum have a distorted appearance thought to be related to gastric acid. In contrast, the villi of the distal duodenum are tall, slender, and very regular, similar to those in the jejunum. The ratio of villi to crypts in the distal duodenum is 4 : 1 or 5 : 1, again similar to the ratio in the jejunum. Within the submucosa of the duodenum are located the branched Brunner’s glands, which secrete an alkaline and clear mucus containing bicarbonate, epidermal growth factor, and pepsinogen II. Brunner’s glands

779

780

Section VI  Stomach and Duodenum are most numerous in the proximal duodenum and decrease in number distally. Rather than emptying into the duodenum through their own duct system, they empty into the duodenum through adjacent intestinal glands.

CONGENITAL ANOMALIES OF THE STOMACH AND DUODENUM The congenital anomalies of the stomach and duodenum are summarized in Table 47-1.

GASTRIC ANOMALIES

Congenital anomalies of the stomach are among the least frequently encountered malformations of the gastrointestinal tract. These lesions may present during the neonatal period or later in life, depending on the degree of gastric outlet obstruction.

Gastric Atresia

Gastric atresias generally occur in the antrum or pylorus in one of three forms: complete segmental defect, segmental defect bridged by a remnant of a fibrous cord, or a membrane (also called a web, diaphragm, or septum). These lesions are uncommon, with a reported incidence of 1 to 3 per 100,000; membranes comprise the majority. Membranes consist of gastric mucosa, submucosa, and muscularis mucosa. In contrast, the fibrous cord generally lacks mucosal elements but contains normal serosal and muscular layers. Membranes may be complete (totally obstructive) or incomplete (perforate). For the sake of clarity, incomplete gastric membranes, which are by definition not atresias, are also considered here. Pathogenesis The cause of these lesions remains unknown, but the timing of a contrary developmental event may determine the type of atresia. For example, if there is fusion of redundant endo-

derm before eight weeks’ gestation (before muscle layer development), then discontinuity of gastric wall musculature would result in a segmental defect with or without a fibrous cord. On the other hand, if redundancy occurs after eight weeks’ gestation, when muscle layers are complete, a simple membrane develops. An alternative mechanism— focal ischemia at a critical time in development—has been proposed. Finally, failure of recanalization of the gastric lumen following temporary obstruction from mucosal proliferation has been proposed as a cause but is not a viable explanation because obstruction or recanalization does not occur in the stomach (unlike the esophagus and duo­ denum).6 Total epithelial detachment of gastric mucosa, associated with α6β4 integrin expression deficiency at the junction of epithelial cells and lamina propria, has been noted in a child with pyloric atresia.7 Genetic factors also are important. In addition to a fami­ lial form (autosomal recessive), gastric atresia is also asso­ ciated with Down syndrome and junctional epidermolysis bullosa. In the case of epidermolysis bullosa–pyloric atresia– obstructive uropathy association, mutations in the α6 and β4 integrin subunits of the hemidesmosome have been noted.8,9 Junctional epidermolysis bullosa, also known as the JEB– pyloric atresia syndrome,10 is the only type that has been described. Other associated anomalies are malrotation, atrial septal defect, absent gallbladder, tracheoesophageal fistula, vaginal atresia, and absent extrahepatic portal vein.11 In addition, gastric atresia may be associated with multiple intestinal atresias and immunodeficiency.12 Clinical Features and Diagnosis In proximal gastrointestinal obstruction, polyhydramnios is commonly noted during pregnancy. Newborn infants with any variant of gastric atresia have signs of gastric outlet obstruction including onset of forceful, nonbilious emesis following the first feeding. There may be drooling and respiratory distress. The abdomen is generally scaphoid unless gastric distention is present. When diagnosis is delayed,

Table 47-1  Anomalies of the Stomach and Duodenum AGE AT PRESENTATION

SYMPTOMS AND SIGNS

TREATMENT

Infancy

Nonbilious emesis

Any age

Failure to thrive, emesis

Gastroduodenostomy, gastrojejunostomy Incision or excision, pyloroplasty

Rare

Infancy

Emesis, malnutrition

Infancy

Nonbilious emesis

Gastric duplication

United States, 3 : 1000 (range, 1-8 : 1000 in various regions); male/female, 4 : 1 Rare male/female, 1 : 2

Any age

Gastric diverticulum Gastric teratoma Gastric volvulus

Rare Rare Rare

Any age Any age Any age

Abdominal mass, emesis, hematemesis; peritonitis if ruptured Usually asymptomatic Upper abdominal mass Emesis, refusal to feed

Duodenal atresia or stenosis Annular pancreas Duodenal duplication Malrotation and midgut volvulus

1 : 20,000

Newborn

1 : 10,000 Rare Rare

Any age Any age Any age

ANOMALY

INCIDENCE

Gastric, antral, or pyloric atresia Pyloric or antral membrane Microgastria

3 : 100,000, when combined with webs As above

Pyloric stenosis

Bilious emesis, upper abdominal distention Bilious emesis, failure to thrive Gastrointestinal bleeding, pain Bilious emesis, upper abdominal distention

Continuous-drip feedings or jejunal reservoir pouch Pyloromyotomy

Excision or partial gastrectomy Usually unnecessary Resection Reduction of volvulus, anterior gastropexy Duodenojejunostomy or gastrojejunostomy Duodenojejunostomy Excision Reduction, division of bands, possibly resection

Chapter 47  Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum severe metabolic acidosis, dehydration, and shock occur; prolonged gastric distention may result in perforation. Abdominal radiographs demonstrate gaseous distention of the stomach and a gasless intestine. Upper gastrointestinal contrast study shows complete obstruction of the stomach, generally at the level of the antrum or pylorus. The type of lesion (such as a membrane) can be determined only via surgical exploration. In the usual case of incomplete antral and pyloric membranes, the age of presentation depends on the degree of obstruction; symptoms may therefore develop at any age from infancy to adulthood. These lesions, except for the presence of membrane perforations, are identical to the membranes of gastric atresia. Luminal narrowing occurs secondary to the malformation itself as well as from local inflammation and edema. The primary symptom is vomiting, which in infants or children may result in failure to establish normal weight gain. In older children and adults, the symptoms may mimic those of peptic ulcer disease, with nausea, epigastric pain, and weight loss. Diarrhea has been observed, but its physiologic basis is unknown. The abdominal radiograph is typically normal, although gastric distention may be noted. Occasionally prenatal diagnosis may be suggested by ultrasonographic findings of dilated stomach without polyhydramnios. Definitive diagnosis is established by contrast radiography, ultrasonography, or upper endoscopy. Contrast radiography demonstrates the membrane as a thin, circumferential filling defect in the antrum or pylorus. Careful observation shows contrast material with delayed passage through a central defect in the membrane; overall gastric emptying is delayed. Ultrasonography may demonstrate segmentation of the antrum, whereas on upper endoscopy a small, fixed opening in the antrum or pylorus may be evident, surrounded by a mucosa free of folds. Treatment Following patient stabilization with fluids and gastric decompression, definitive treatment is surgical. Complete or incomplete antral membranes are treated by simple excision. Pyloric membranes require pyloroplasty. The presence of a concomitant duodenal atresia has been described (also known as windsock diaphragm), and its presence or absence is verified by passage of a catheter distally into the duodenum intraoperatively. Endoscopic therapy using a snare, papillotome, laser, or dilation via balloon also has been described. In cases involving atretic gap, gastroduodenostomy is considered curative. An alternative approach is pyloric sphincter reconstruction via longitudinal pyloromyotomy, followed by end-to-end anastomosis of cul-de-sacs of gastric and duodenal mucosa.13 Gastrojejunostomy is not recommended in children because of the risk of marginal ulcer.

Microgastria

Microgastria is an extremely rare congenital anomaly of the caudal part of the foregut. A small, tubular or saccular, incompletely rotated stomach is associated with a megaesophagus. Varying degrees of the anomaly occur owing to arrested development during the fifth week of gestation in differentiation of the greater curvature of the stomach; neither rotation nor fusiform dilation of the stomach occurs.14 A localized vascular insufficiency has been postulated to lead to the development of microgastria after the eighth week of gestation.15 The etiology is unknown. Fortunately normal histology is preserved. Microgastria may occur as an isolated anomaly but more commonly is in association with duodenal atresia; nonrotation of the

midgut; ileal duplication; hiatal hernia; asplenia; partial situs inversus; or renal, upper limb (microgastria–limb reduction anomaly), cardiac, pulmonary, skeletal, or spinal anomalies. In isolation, microgastria is not lethal, but other associated anomalies may be. It has been suggested that microgastria in association with limb reduction defects and central nervous system anomalies has a genetic basis, with an autosomal recessive pattern of inheritance.16 Chromosome analysis is normal. Clinical Features and Diagnosis The infant typically presents with postprandial vomiting and malnutrition. There may also be diarrhea (a result of rapid gastric emptying) and dumping syndrome. Respiratory symptoms including respiratory distress at birth and stridor, as well as recurrent pulmonary infections, have been reported. Anemia may occur owing to iron deficiency because decreased gastric acid secretion may preclude adequate iron absorption and cobalamin (vitamin B12) deficiency due to hyposecretion of intrinsic factor. Prenatal ultrasonography may detect a small stomach and poly­ hydramnios. Contrast radiography shows the megaesophagus and tubular or small stomach. The lower esophageal sphincter is poorly defined, and gastroesophageal reflux is usually severe. Treatment The medical management of microgastria includes frequent small-volume feedings or continuous-drip feedings into the stomach. An alternative is nocturnal drip feedings via jejunostomy to supplement oral intake. The surgical creation of a double-lumen Roux-en-Y pouch anastomosed to the greater curvature of the stomach has been described. This Hunt-Lawrence jejunal pouch has allowed normal growth and development and prevented reflux and dumping syndrome.17

Gastric Diverticulum (see also Chapter 23) A gastric diverticulum is the rarest type of gastrointestinal diverticulum. The true congenital diverticulum contains all gastric tissue layers and is located on the posterior wall of the cardia. The intramural (or partial) diverticulum projects into but not through the muscular layer, most commonly located along the greater curvature of the antrum. The false (or pseudo-) diverticulum is formed by mucosal and submucosal herniation through a defect in the muscular wall and lacks muscularis propria. Familial occurrence has not been described for any of these lesions. Clinical Features and Diagnosis Most congenital gastric diverticula are asymptomatic and are incidental findings on radiography or endoscopy, or at autopsy (see Chapter 23). Size varies from 1 to 11 cm. Contrast radiography shows a rounded, well-delineated mobile pouch, often with an air-fluid level. Emptying of the diverticulum may be delayed. On endoscopy, the diverticulum is seen as a well-delineated opening; distention by the scope may reproduce symptoms. Unfortunately, both upper gastrointestinal radiologic studies and endoscopy may miss the diagnosis due to the typical location at the gastroesophageal junction. Symptoms may be epigastric or lower chest pain, indigestion, bleeding, or nonbilious emesis. The differential diagnosis includes an acquired gastric diverticulum found in association with pancreatitis, gastric outlet obstruction, trauma, ulcer disease, or malignancy. Hiatal hernia and hypertrophic gastric folds may mimic a diverticulum on contrast studies. Radiology cannot distinguish between congenital and acquired diverticula.

781

782

Section VI  Stomach and Duodenum Treatment In the case of an incidentally discovered proximal gastric diverticulum, treatment is unnecessary. If symptoms are thought to be consistent with the diagnosis, the diverticulum may be amputated or invaginated. Because of the risk of malignancy associated with distal gastric diverticula, surgical treatment by amputation, invagination, or segmental resection has been recommended. Laparoscopic resection following gastroscopic localization has been described.

errors in separation of notochord and endoderm, persistence of embryonic diverticula, and persistence of vacuoles within the epithelium of the primitive foregut.18 Most duplications occur in women (65%) and are detected during infancy or childhood (80%) with no familial tendency. Aside from concurrent duplications, vertebral anomalies are the second most commonly linked abnormality.19 Carcinomas arising in congenital duplications have been described in adults.

Gastric Duplication

Clinical Features and Diagnosis The clinical presentation of gastric duplication depends on factors such as size, location, and communicating structure (if any). Symptoms and signs vary and may include colic; abdominal mass; epigastric pain; failure to gain appropriate weight; vomiting; occult or frank upper or lower gastrointestinal bleeding secondary to peptic ulceration, the latter occurring via erosion into the colon; hematobilia via a communication with intrahepatic bile duct; respiratory distress or hemoptysis (perforated cyst fistulized to lung)20; pyloric obstruction; peritonitis secondary to rupture; pancreatitis; pancreatic pseudocyst; and acute abdomen. In early infancy symptoms may mimic those of hypertrophic pyloric stenosis. Diagnosis is suggested by an abdominal radiograph showing displacement and extrinsic compression of gastric lumen. Contrast radiography may demonstrate the duplication via a mass effect on the stomach (Fig. 47-8A) or the cyst may be imaged directly when there is communication

Approximately 20% of all gastrointestinal duplications are gastric. Duplication of the stomach can occur in isolation, as a triplication (two gastric duplications in one individual), or with duplications of other structures in the gastrointe­ stinal tract such as the esophagus or duodenum. Location is contiguous with the stomach, generally along the greater curvature or posterior wall and contains all layers of the gastric wall. Gastric and pancreatic mucosa lining the duplication are most clinically significant secondary to potential complications such as peptic ulcer disease and pancreatitis. Because the duplication rarely communicates with the stomach, a tubular, fusiform, or spherical cystic mass develops. Infrequently there may be a connection to the colon, pancreas, or pancreatic duplication; the connection may be the result of an acquired fistula from a penetrating peptic ulcer within the duplication. Several embryologic defects have been proposed as etiologies for duplications including

Antrum

GB

A

B

C

D

Figure 47-8.  Gastric duplication in a 12-year-old boy with a 1-year history of vomiting and intermittent abdominal pain. Physical examination and laboratory studies were normal. A, Film from an upper gastrointestinal series shows an extrinsic mass displacing and compressing the antrum and duodenal C loop. B, Ultrasonographic image shows a hypoechoic mass behind the gastric antrum and medial to the gallbladder. C, Computed tomography scan shows a circumferential soft-tissue thickening displacing and narrowing the antrum. D, Intraoperative picture of the gastric duplication after dissection of the stomach and before resection.

Chapter 47  Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum with the gastrointestinal tract. Ultrasonography (see Fig. 47-8B) including prenatal ultrasonography,20 computed tomo­graphy (CT) scan (see Fig. 47-8C), magnetic resonance cholangiopancreatography (MRCP), Tc-99m pertechnetate, and endoscopy with endoscopic ultrasonography (EUS) may also demonstrate the lesion. Peristalsis identified by EUS in a juxtaenteric cyst is specific for a duplication cyst and may be considered as a diagnostic feature of this condition.21 Treatment Surgical excision is considered optimal therapy (see Fig. 47-8D). Laparoscopic resection has been described.22 When complete excision is not possible, as may be the case when cyst and viscus have a common muscle layer, debulking, cyst-gastrostomy, or partial gastrectomy may be necessary. Additionally, mucosectomy or mucosal surface ablation should be considered because the development of malignancy in enteric duplications has been documented in adults.23

Gastric Teratoma

Gastric teratomas are benign neoplasms of the stomach that occur almost exclusively in men. Gastric teratomas are rare, comprising only 1% of all childhood teratomas. These tumors may have their origins in pluripotential cells and contain all three embryonic germ cell layers. They are almost always diagnosed during infancy owing to their large size. Most are located along the greater curvature of the stomach and are extragastric, although intramural extension has been reported.24 The immature type (containing yolk cell tumor, germinoma, and embryonal carcinoma) may infiltrate regional structures—omentum, regional lymph nodes, left lobe of the liver—whereas the mature tissue form does not. In virtually all cases, gastric teratoma is an isolated finding and is not associated with other tumors or malformations.25 Clinical Features and Diagnosis The typical patient is a male infant with an abdominal mass; mean age at presentation is 3.2 months.26 Vomiting may be present from intrinsic compression and gastrointestinal bleeding due to transmural growth and disruption of gastric mucosa. Polyhydramnios may be noted prenatally secondary to gastric obstruction by the mass. The newborn infant with a teratoma may be delivered prematurely or have respiratory distress on the basis of increased abdominal pressure. Delivery may be difficult, putting the infant at risk for injuries such as shoulder dystocia. Gastric teratoma associated with gastric perforation, mimicking meconium peritonitis, has been described.27 Noncontrast radiography demonstrates characteristic calcifications. Ultrasonography demonstrates solid and cystic areas, and CT or magnetic resonance imaging confirms the diagnosis and evaluates regional infiltration.26 Treatment Tumor excision with primary gastric repair is the procedure of choice and is curative. Partial or total gastrectomy is required for intramural tumor extension. Malignant transformation to adenocarcinoma has been reported,28 as well as premalignant changes,29 and peritoneal gliomatosis has been observed. Fortunately even those cases with malignant histologic features or extension into adjacent tissues have an excellent prognosis.25 In the case of im­mature gastric teratomas, a serum alpha fetoprotein level may be useful, especially because of the possibility of recurrence or metastasis and the need for adjuvant chemotherapy.30

Gastric Volvulus

Gastric volvulus (see Table 47-1) is discussed in greater detail in Chapter 24.

Infantile Hypertrophic Pyloric Stenosis

Infantile hypertrophic pyloric stenosis (IHPS) is a form of gastric outlet obstruction caused by hypertrophy of circular muscle surrounding the pyloric channel. Correction of IHPS is the most common abdominal operative procedure during the first 6 months of life. Because the muscular hypertrophy and obstruction tend to be an evolving process during the postnatal period, IHPS is arguably not a true congenital defect.31 The etiology of IHPS remains the subject of speculation. A localized lack of nitric oxide synthase, an enzyme associated with smooth muscle relaxation, or abnormal neuronal innervation associated with decreased muscle neurofilaments, nerve terminals, synaptic vesicle protein, and neural cell adhesion molecule32 has been implicated. However, anatomic studies cannot determine whether nitric oxide synthase deficiency is a primary or secondary event,33 and nitric oxide synthase deficiency is only notable in a subset of cases.34 Pacemaker cells that regulate gastrointestinal motility, the interstitial cell of Cajal, are observed only near the submucosa in IHPS instead of throughout the pylorus.35,36 Epidermal growth factor (EGF), EGF receptor, and heparin-binding EGF-like growth factor are markedly increased in smooth muscle cells in IHPS,37 but their triggers are unknown. The incidence of IHPS in the United States is approximately 3 in 1000 live births but varies among ethnic groups and regions from 1 to 8 in 1000 live births. Incidence is highest among whites (especially northern Europeans), whereas incidence is lower among African Americans and Africans and lowest among Asians. Men outnumber women by a ratio of 4 : 1 or 5 : 1. Familial clustering of IHPS is widely recognized, with autosomal dominant forms reported.38 Approximately 50% of identical twins are affected, leading credence to the roles of genetic and environmental factors. Male relatives of affected women are more likely to develop IHPS, such that siblings and offspring of affected women are more likely to develop IHPS than are relatives of affected men. Others at increased risk are first-born male infants, especially those with high birth weights or born to professional parents. IHPS also occurs in association with Turner’s syndrome, trisomy 18, Cornelia de Lange syndrome, esophageal atresia, Hirschsprung’s disease, phenylketonuria, and congenital rubella syndrome. Multiple reports have described an association between early macrolide exposure in infants, including exposure through breast milk,39 and the development of IHPS.40-42 In addition, a recent decline in the incidence of IHPS that parallels that of sudden infant death syndrome has been observed and coincides with the recommendation of supine infant sleeping position. This has led to the hypothesis that posture may contribute to the underlying etiology.43 Clinical Features and Diagnosis Infants with IHPS are typically asymptomatic until 3 to 4 weeks of age, although a small number may present as early as the first week of life. Initially infants present with mild spitting, which progresses to projectile vomiting following feedings. Vomiting may be so forceful as to exit through the nostrils, as well as the mouth. Emesis may contain “coffee ground” material or small amounts of frank blood but is rarely bilious. Early in the course the infant remains hungry following vomiting episodes but, with time, loses interest in feeding and may present wasted and severely dehydrated.

783

784

Section VI  Stomach and Duodenum Decreased urinary and stool output accompanies dehydration. Marked metabolic alkalosis develops secondary to chloride loss in the vomitus. Infants may be misdiagnosed with formula allergy or gastroesophageal reflux. On physical examination the infant with IHPS may appear wasted and dehydrated, but the extent is variable and related to severity and duration of symptoms. The classic physical signs are a palpable pyloric mass and visible peristaltic waves. The palpable “olive” is most easily felt in a wasted patient, immediately following emesis or aspiration of the stomach. The location of the olive varies from the level of the umbilicus to near the epigastrium. The pyloric mass is palpable in 70% to 90% of affected infants, depending on the experience and patience of the examiner. Emptying the stomach by nasogastric tube placement and palpation of the stomach with the infant in the prone position may enhance sensitivity. Peristaltic waves are best observed during feeding of the naked infant while the infant is cradled in the mother’s left arm. Many infants appear jaundiced due to an indirect hyperbilirubinemia related to dehydration and perhaps malnutrition. When the presentation is typical and the olive palpated, no studies are necessary. However, in the minority of infants with projectile vomiting, definitive diagnosis requires radiologic studies. Noncontrast radiography demonstrates a distended stomach with paucity of gas beyond the stomach. Diagnosis is confirmed by abdominal ultrasonography of the pylorus, which has supplanted contrast radiography as the diagnostic study of choice for IHPS. Because dehydration may affect the pyloric ultrasound measurements, ensuring an adequate hydration status may be prudent before ultrasonographic evaluation.44 The length of the hypertrophied canal is variable and may range from as little as 14 mm to more than 20 mm. The numeric value for the lower limit of muscle thickness has varied in reports in the literature, ranging between 3 and 4.5  mm. This appears as a characteristic sonolucent “donut” (Fig. 47-9). Many consider the numeric value less important than the overall morphology of the canal and real-time observations. A negative ultrasonographic study hinges on an unequivocal diagnosis of a normal pyloric ring and a distensible antropyloric portion of the stomach.45 However, when the differential diagnosis includes IHPS, gastroesophageal reflux, or other upper gastrointestinal disorders, contrast radiography may be the

Figure 47-9.  Abdominal ultrasonographic image in a 1-month-old infant with idiopathic hypertrophic pyloric stenosis demonstrating the sonolucent “donut” of pyloric hypertrophy on cross-section. Crossbars measure an abnormal (7 mm) muscle thickness. (Courtesy Jeanne Joglar, MD.)

appropriate first test. Contrast radiography must be performed carefully, and gastric contents should first be aspirated. The infant is given barium by nipple and imaged in a semiprone position. Characteristic findings include an elongated narrow pylorus with the appearance of a “double channel.” There is also indentation of the adjacent antrum and duodenum by the pyloric mass producing the so-called shoulders (Fig. 47-10). The most common abnormality that mimics IHPS is pylorospasm. Diagnosis of IHPS by endoscopy has been described in which the pylorus appears as a cauliflower-like narrowing, through which a 7.8-mm (external diameter) endoscope cannot be passed.46 However, another report on endoscopic diagnosis has refuted these claims.47 Endoscopy is also potentially beneficial to evaluate for eosinophilic gastroenteritis (see Chapter 27), which has been linked to pyloric stenosis.48 Treatment The initial therapy for IHPS is fluid and electrolyte replacement to correct dehydration and hypochloremic metabolic alkalosis. Depending on severity, fluid and electrolyte repletion can usually be accomplished within 24 hours. Definitive therapy is the Ramstedt pyloromyotomy, which entails a longitudinal incision through the hypertrophied pyloric muscle down to the submucosa on the anterior surface of the pylorus. After spreading the muscle, the intact mucosa bulges through the incision to the level of the incised muscle. An alternative operation is the pyloric traumamyoplasty. With this procedure the pylorus is grasped with a Babcock clamp that disrupts the hypertrophied circular muscles in two places.49,50 Laparoscopic pyloromyotomy with its improved cosmetic results and reduced need for analgesics is becoming increasingly popular.51,52 Although infants may continue to vomit for the first few days postoperatively, persistent vomiting is suggestive of inadequate surgery. Nonoperative therapy consists of the use of anticho­ linergic medications53 and paste-consistency feedings until such time that the muscle hypertrophy resolves.54 Because

Figure 47-10.  Film from an upper gastrointestinal series in a 1-month-old infant with idiopathic hypertrophic pyloric stenosis demonstrating an elongated pylorus and antral and duodenal “shoulders” secondary to a mass effect. (Courtesy Marcia Pritchard, MD.)

Chapter 47  Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum of the high failure rate, the prolonged recovery period (compared with surgery), and the low risk of pyloromy­ otomy, the nonoperative approach is rarely used in the United States. The prognosis following surgery is excellent. The infant resumes normal growth and development. Although divergent gastric emptying rates have been observed many years following treatment of IHPS, gastric emptying by scintigraphy using radiolabeled liquids or solids was similar in patients treated surgically or conservatively versus controls.46

Adult Hypertrophic Pyloric Stenosis

Hypertrophic pyloric stenosis rarely occurs in adults. There are approximately 200 cases described in the literature. When HPS occurs in adults, its anatomic features are identical to the infantile type. In adults pyloric thickening is generally associated with peptic ulcer disease, hypertrophic gastopathy,55 or carcinoma. In a few cases, no etiology is determined; it is therefore unknown whether these are missed infantile cases or whether the hypertrophy occurred later in life. There is a family history of IHPS in some cases of adult HPS, thus again suggesting a role for genetic predisposition or missed infantile cases. In addition, 80% of adult HPS occur in men. The resected pylorus demonstrates normal mucosa and marked circumferential thickening of the muscularis propria.56 Microscopically there are variable degrees of inflammatory changes or edema, and degenerative changes in the ganglion cells of the myenteric plexus have been reported.55 Clinical Features and Diagnosis Symptoms of adult HPS are similar to those observed in infancy: nausea, mild vomiting, early satiety, and epigastric pain, especially after eating. In contrast with the infantile form, the physical examination may not be helpful because the pyloric mass is difficult to palpate in adults. On contrast radiography, the elongated narrow pylorus is again apparent; gastric emptying is delayed, and the stomach may be dilated. Ultrasonography is the screening procedure of choice; it is generally considered abnormal if the pylorus is 1  cm or more thick with persistent elongation of more than 2 cm.55 Upper endoscopy is indicated to differentiate idiopathic HPS from carcinoma or chronic peptic ulcer disease. Treatment Traditionally, surgical pyloromyotomy or resection of the involved region has been considered the procedure of choice. Because of the risk of a small focus of carcinoma, surgical resection of the pylorus has been recommended. Endoscopic balloon dilation has also been efficacious in the management of HPS, but a high postprocedure recurrence rate—80% within the first 6 months—has been reported.57 Additionally, palliation of pyloric stenosis caused by gastric cancer using an endoscopically placed stent has been described.58

DUODENAL ANOMALIES Duodenal Atresia and Stenosis

Duodenal atresia and stenosis are congenital defects characterized by complete and partial obstruction of the duodenum, respectively. Atresias occur in various anatomic configurations including a blind-ending pouch with no connection to the distal duodenum (least common), a pouch with a fibrous cord connecting to the distal duodenum, or a complete membrane obstructing the lumen (most common).

Perforate membranes are also a cause of duodenal stenosis. All three lesions occur with greatest frequency near the ampulla of Vater, with most lesions (80%) occurring distal to this landmark. The overall incidence of the three anomalies combined is about 1 per 200,000 live births with a slight predilection for girls. The etiology of these lesions may relate to failure to recanalize the duodenal lumen by vacuolization at 8 to 10 weeks’ gestation. This is distinct from atresia or stenosis of the jejunum and ileum, which are caused by vascular accidents in utero.59 Duodenal stenosis has been observed in sonic hedgehog (shh)–mutant mice, thus adding to our understanding that mutations in signaling pathways may play a role in this malformation.60 In two series of more than 100 cases,61,62 more than 50% of affected patients had associated congenital defects including pancreatic defects; intestinal malrotation with congenital bands; esophageal atresia; Meckel’s divertic­ ulum; imperforate anus; congenital heart disease; central nervous system lesions; renal anomalies; and, rarely, biliary tract anomalies. Trisomy 21 is strongly associated with duodenal atresia/stenosis/web in that anywhere from 25% to more than 50% of cases occur in infants and children with this chromosomal anomaly. Familial association is rare, although isolated case reports suggest a possible genetic association.63,64 A report of father and son with periam­ pullary obstruction due to duodenal stenosis and annular pancreas (in the father) and segmental duodenal atresia (in the son) serves as a reminder that with increased survival of affected infants, a genetic basis may be realized in the future.65 Clinical Features and Diagnosis The diagnosis of duodenal atresia may be suspected prenatally when ultrasonography demonstrates gastric and proximal duodenal dilation and polyhydramnios. Polyhydramnios is present in 33% to 50% of cases of duodenal atresia. The absence of gastric and proximal duodenal dilation in the presence of polyhydramnios does not exclude the diagnosis because intrauterine emesis may limit pre­ obstructive dilation. High-frequency transvaginal transducers used in ultrasonography may overdiagnose intestinal dilation, so longer scanning is recommended once obstruction is suspected.66 The infant with duodenal atresia is often born preterm and has early feeding intolerance characterized by vomiting and upper abdominal distention. Emesis is usually bilious because most lesions occur distal to the entry of the bile duct into the duodenum. Nonbilious emesis is seen in 15% to 20% of cases secondary to more proximal obstruction. Any child with trisomy 21 and vomiting (especially bilestained) requires further evaluation for duodenal stenosis. Duodenal stenosis or a partial membrane may present at any age, depending on the degree of obstruction. Infants and children present with vomiting, failure to gain weight adequately, and/or aspiration. Vomiting may be intermittent and of variable severity such that symptomatic lesions may remain undiagnosed for months to years. Occasionally, diagnosis is delayed until adulthood. Noncontrast radiographs of the infant with duodenal obstruction classically demonstrate the presence of air in the stomach and in the first portion of the duodenum—the “double-bubble” sign (Fig. 47-11). The absence of air beyond the second bubble should be interpreted as probable duodenal atresia. Contrast radiography is generally effective in demonstrating atresias, stenosis, membranes, and other anomalies resulting in external compression of the duodenum (Fig. 47-12). In addition, normal or abnormal rotation and fixation of the bowel can be assessed. Competence of

785

786

Section VI  Stomach and Duodenum

Figure 47-11.  Anteroposterior and lateral noncontrast films of an infant with duodenal atresia demonstrate the “double-bubble” sign. (Courtesy Marcia Pritchard, MD.)

The operation has evolved from a side-to-side anastomosis to a proximal transverse to distal longitudinal (“diamond shaped”) anastomosis. Associated malrotation is corrected with a Ladd procedure. A catheter is passed into the distal duodenum to investigate for a second obstruction, which occurs in about 3% of cases. Membranes may be excised without anastomosis if the membrane was an isolated finding. Balloon dilation has been described for membranous duodenal stenosis.67,68 Endoscopic laser resection of membranes has been reported; unfortunately, subsequent scar formation has resulted in stenosis and the need for surgery.69 Late complications continue to plague patients even following surgical repair: motility problems, megaduodenum, gastroesophageal reflux unresponsive to medications, gastropathy, adhesions, and peptic ulcer disease occur months to years following primary repair.62 Approximately 12% of patients required revision or another intra-abdominal surgery over a 30-year follow-up period.70 Two teenagers have presented with choledochal cyst.61 Megaduodenum proximal to the obstruction, with abnormal peristalsis, is a common long-term issue, but most patients are asymptomatic. For symptomatic patients with megaduodenum, bowel plication may be indicated.71

Annular Pancreas

b

d a

c

Figure 47-12.  Upper gastrointestinal contrast film of an infant with a duodenal membrane/web demonstrates dilatation (a) from the pylorus (b) to the third portion of the duodenum proximal to the webbed segment of duodenum (without contrast) (c) and a normal caliber fourth portion of the duodenum (d). (Courtesy Korgun Koral, MD.)

the ampullary sphincter has been noted to be compromised in few reported cases. Reflux of contrast medium through the ampulla poses risks of developing cholangitis and pancreatitis. Occasionally upper endoscopy is useful in diagnosing or defining a duodenal stenosis or membrane. Treatment A newborn infant suspected of duodenal obstruction should have a nasogastric tube placed for decompression, and correction of fluid and electrolyte abnormalities should be instituted. The surgical approach in the past was duodenojejunostomy, but now duodenoduodenostomy is preferred.62

Annular pancreas is an unusual congenital malformation characterized by a thin ring of pancreatic tissue, most often encircling the second portion of the duodenum, contributing in variable degrees to obstruction (see also Chapter 55). The lesion may present in the neonatal period, in childhood, or adulthood. It is the most common congenital anomaly of the pancreas presenting in children. Some cases remain asymptomatic and are discovered as an incidental finding during endoscopic retrograde cholangiopancreatography (ERCP) or at autopsy. The anomalous tissue is histologically normal and contains a moderately sized pancreatic duct. The pancreatic tissue may penetrate the muscularis of the duodenal wall or remain distinct from the duodenum. Several hypotheses exist regarding embryologic origin of annular pancreas. Evidence appears to favor Lecco’s 1910 hypothesis that the ventral pancreatic anlage becomes fixed to the duodenal wall before rotation during the fifth week of gestation. With subsequent growth and fusion of the dorsal and ventral anlagen, a partial (75%) or complete (25%) ring of pancreatic tissue is formed.6 Incidence of the disorder is approximately 1 in 100,000 live births, but this figure does not account for cases found during adulthood, during ERCP (when it is usually noted as an incidental finding), or at autopsy. The true incidence may be as high as 1 per 250 live births. In infancy the incidence is equal in male and female infants. In adulthood men outnumber women by 2 : 1. Infant and childhood cases are associated with other congenital anomalies in an estimated 40% to 70% of cases, including trisomy 21, duodenal atresia, cardiac defects, anorectal malformations, Meckel’s diverticulum, tracheoesophageal fistula, malrotation, genitourinary malformation, and situs inversus.72 In adults, compared with children, it is more common to have malrotation, duodenal web, Shatzki’s ring, duodenal atresia, tracheoesophageal fistula, and genitourinary abnormalities. Also adults are at increased risk of pancreatobiliary neoplasia. In 13 adults with annular pancreas, 6 had pancreatobiliary neoplasia, including 2 with adenocarcinoma of the pancrease, 2 with ampullary adenoma, and 1 with adenocarcinoma of the gallbladder.72 The finding of an annular pancreas in a woman and her child, as well as in two successive generations, suggests a possible hereditary link.73

Chapter 47  Anatomy, Histology, Embryology, and Developmental Anomalies of the Stomach and Duodenum Clinical Features and Diagnosis Annular pancreas produces symptoms when tissue obstructs the duodenum or biliary tree. Controversy exists as to whether the annular pancreas actually plays a role in obstruction. The abnormally located pancreatic tissue is a visible indicator of an underlying duodenal abnormality that can range from minimal duodenal stenosis to atresia.74 Infants may present with high-grade obstructive symptoms and signs such as bilious emesis and upper abdominal distention indistinguishable from duodenal atresia or malrotation with midgut volvulus. During childhood intermittent bilious emesis and failure to thrive are common presenting symptoms, whereas during adulthood the most common symptom is abdominal pain. Other symptoms and signs in adults include nausea, vomiting, gastric outlet obstruction, pancreatitis, pancreatolithiasis, pancreas divisum, pancreatic mass, gastric or duodenal ulcer, or biliary obstruction resulting in jaundice. In the adult, development of symptoms peaks in the third to fifth decades. Noncontrast radiologic studies of the infant may demonstrate the double-bubble sign identical to that seen in duodenal atresia (see Fig. 47-10). Contrast radiography should be done to ensure that the obstruction is not due to midgut volvulus, a surgical emergency. In adults, transabdominal ultrasound, EUS, CT, or magnetic resonance pancreatography may diagnose annular pancreas. ERCP may demonstrate ductular structures consistent with annular pancreas, but in some cases it may not be technically feasible owing to duodenal obstruction proximal to the major ampulla. Endoscopic ultrasonography is especially useful when prior gastric resection or duodenal obstruction precludes ERCP; in addition, a mass may be staged or undergo fine-needle aspiration at the time of EUS. The ability to evaluate for mass is a new consideration, given reports of ampullary carcinoma in association with annular pancreas; hence, jaundice should not be attributed to annular pancreas until carcinoma is ruled out.75 Magnetic resonance pancreatography, which allows spatial resolution of the entire pancreaticobiliary tree, can identify the annulus and the duct within that surrounds the duodenum. Finally, intraoperative diagnosis at laparotomy is not unusual. Treatment The preferred operative therapy for annular pancreas includes duodenoduodenostomy or duodenojejunostomy. Prognosis postoperatively is excellent with either, and postoperative deaths among infants are generally due to associated anomalies. Division or dissection of the pan­ creatic tissue is not recommended owing to the high risk of complications, including pancreatitis, pancreatic fistula, and incomplete relief of symptoms due to intrinsic duo­ denal narrowing. An annular pancreas identified at the time of organ procurement has been transplanted along with a long segment of duodenum with good results, so that annular pancreas can be considered suitable for transplantation.76

Duodenal Duplication Cysts

Duodenal duplication cysts are a rare anomaly, totaling only 7% of gastrointestinal duplications. Most commonly located posterior to the first or second portion of the duodenum, these spherical or tubular cysts generally do not communicate with the duodenal lumen but do share blood supply with the duodenum. Three histologic criteria for duodenal duplication cysts exist: gastrointestinal mucosa, a smooth muscle layer in the wall, and an association with the duodenal wall. The mucosa is typically duodenal, but in 15% of cases there is gastric mucosa, and very rarely, pancreatic

tissue is found. Men and women are affected equally. Several embryologic theories have been postulated but none explain the diversity of anatomic varieties.77 Clinical Features and Diagnosis Duplications may be clinically silent for years before presentation. Presenting signs and symptoms of these cysts are typically that of partial gastric outlet obstruction include vomiting, decreased oral intake, periumbilical tenderness, and abdominal distention. Conversely, an asymptomatic mass found on physical or radiologic examination may be noted first. If heterotopic gastric mucosa is present, bleeding or perforation may be the initial presenting sign. In the neonate duodenal obstruction due to a large duplication cyst has been reported. Infected duodenal duplication cyst has been noted as well.78 Pancreatitis may recur if the cyst compresses or is in communication with the pancreatic duct. Finally, jaundice and duodenojejunal intussusception resulting in small bowel obstruction have been described.79 Noncontrast, as well as contrast, radiography may demonstrate obstruction or compression effect, but in general, findings are nonspecific and only suggestive. Abdominal ultrasonography may show unilocular cystic structure with echogenic mucosa surrounded by thin hypoechoic halo of muscle layer.78 Peristaltic waves through the cyst may be evident on ultrasound. Antenatal ultrasound scans are more often identifying suspected cysts. CT may demonstrate an encapsulated, noncommunicating cyst posterior to the duodenum. On ERCP, a compressible periampullary mass may be seen. Treatment Surgical therapy should be individualized in accordance with the anatomy of the cyst. Because of potential complications, early neonatal resection, even for asymptomatic cysts, has been advocated.80 Operations that have been performed including local excision and cystojejunostomy. Mucosal stripping of the common muscular wall and resection coupled with removal of free walls has been recommended.81 This, however, may be complex because of the proximity of the cyst to the papilla and biliary-pancreatic confluence. Endoscopic drainage, as well as removal, has been successful in adult and pediatric cases. Invasive carcinoma has been reported in an adult with duodenal duplication cyst, so endoscopic drainage without resection may require reconsideration.

Malrotation and Midgut Volvulus

The entity of malrotation and volvulus are described in detail in Chapters 24 and 96. Refer to Table 47-1 and the aforementioned chapters for additional information.

KEY REFERENCES

Carachi R, Azmy A: Foregut duplications. Pediatr Surg Int 2002; 18:371. (Ref 18.) Çorapçioglu F, Ekingen G, Sarper N, Güvenç BH. Immature gastric teratoma of childhood: A case report and review of the literature. JPGN 2004; 39:292-4. (Ref 30.) Fine J-D, Johnson LB, Weiner M, Suchindran C. Gastrointestinal complications of inherited epidermolysis bullosa: Cumulative experience of the National Epidermolysis Bullosa Registry. JPGN 2008; 46:14758. (Ref 10.) Gupta DK, Srinivas M, Dave S, et al. Gastric teratoma in children. Pediatr Surg Int 2000; 16:329. (Ref 26.) Hellan M, Lee T, Lerner T. Diagnosis and therapy of primary hypertrophic pyloric stenosis in adults: Case report and review of literature. J Gastrointestinal Surg 2006; 10:265-9. (Ref 55.) Johnson LR. Gastrointestinal physiology. 6th ed. St. Louis: Mosby; 2001. p 75. (Ref 2.)

787

788

Section VI  Stomach and Duodenum MacMahon B. The continuing enigma of pyloric stenosis of infancy: A review. Epidemiology 2006; 17:195-201. (Ref 43.) Merrot T, Anastasescu R, Pankevych T, et al. Duodenal duplications. Clinical characteristics, embryological hypotheses, histological findings, treatment. Eur J Pediatr Surg 2006; 16:18-23. (Ref 77.) Moore KL, Persaud TVN. The developing human. 7th ed. Philadelphia: WB Saunders; 2003. p 258. (Ref 1.) Naik-Mathuria B, Olutoye O. Foregut abnormalities. Surg Clin North Am 2006; 86:261-84. (Ref 50.)

Ng WT, Lee SY: Hypertophic pyloric stenosis, congenital or not congenital: A critical overview. Pediatr Surg Int 2002; 18:563. (Ref 31.) Okoye BO, Parikh DH, Buick RG, et al. Pyloric atresia: Five new cases, a new association, and a review of the literature with guidelines. J Pediatr Surg 2000; 36:1242. (Ref 11.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

48 Gastric Neuromuscular Function and Neuromuscular Disorders Kenneth L. Koch

CHAPTER OUTLINE Electrophysiologic Basis of Gastric Neuromuscular Function  789 Extracellular Slow Waves and Plateau and Action Potentials  789 Intracellular Electrical Recordings from Gastric Smooth Muscle Cells  790 Interstitial Cells of Cajal  791 Nervous System Innervation  794 Gastric Neuromuscular Activity during Fasting  794 Gastric Neuromuscular Activity after a Meal  794 Regulation of Gastric Neuromuscular Activity after a Meal  798 Gastric Sensory Activities  801 The Stomach and the Regulation of Food Intake, Hunger, and Satiety  802 Developmental Aspects of Gastric Neuromuscular Function  803 Assessment of Gastric Neuromuscular Function  803 Gastric Emptying Rates  803 Gastric Contractions  804

Gastric neuromuscular function refers to the motility or motor activities of the stomach. The three major neuromuscular activities of the stomach are (1) receptive relaxation of the fundus; (2) recurrent peristaltic waves of the corpus and antrum; and (3) antral peristaltic waves coordinated with antropyloroduodenal coordination. These major neuromuscular activities of the stomach accomplish three key functions: (1) to receive the ingested food that we eat (receptive relaxation); (2) to mill (triturate) the ingested foodstuffs into a nutrient suspension termed chyme; and (3) to empty the chyme from the stomach into the duodenum in a highly regulated fashion in order to maximize further digestion and absorption of the nutrients. These critical gastric neuromuscular activities and related functions are complexly modulated by the central nervous system (CNS), the parasympathetic nervous system (PNS) and sympathetic nervous systems (SNS), the interactions of the CNS and the activity of the enteric nervous system (ENS), the interstitial cells of Cajal (ICCs) that regulate the frequency of contractions and organize peristaltic waves, and the host of neurotransmitters that ultimately regulate the contraction and relaxation of gastric smooth muscle.

Gastric Myoelectrical Activity  804 Gastric Relaxation, Accommodation, and Volume  805 Histopathologic Studies in Gastric Neuromuscular Disorders  805 Neuromuscular Disorders of the Stomach  806 Gastroparesis  806 Gastric Neuromuscular Dysfunction Associated with Other Gastrointestinal Disorders  809 Dumping Syndrome and Rapid Gastric Emptying  810 Diagnosis  810 Symptoms  810 Physical Examination  811 Standard Tests  811 Specialized Noninvasive Tests  811 Treatment  812 Drug Therapy  812 Electrical Therapy  812 Endoscopic Therapy  814 Diet Therapy  814

ELECTROPHYSIOLOGIC BASIS OF GASTRIC NEUROMUSCULAR FUNCTION EXTRACELLULAR SLOW WAVES AND PLATEAU AND ACTION POTENTIALS

The stomach is a sophisticated complex sphere of smooth muscle organized into circular, longitudinal, and oblique muscle layers. Gastric myoelectrical activity, termed pace­ setter potentials or slow waves, regulate, control, and pace gastric smooth muscle contractions1,2 (the term slow waves is used in this chapter). In the normal human stomach the slow waves occur at approximately 3 cycles per minute (cpm) or between 2.5 and 3.7 cpm.3,4 From the pacemaker region on the greater curvature of the stomach, between the fundus and the proximal corpus, slow waves propagate circumferentially and migrate distally toward the pylorus every 20 seconds at a velocity of approximately 14 mm/ second in the distal antrum (Fig. 48-1).5,6 The gastric slow waves originate from the interstitial cells of Cajal.7,8 The depolarization upstroke of the slow wave reduces the threshold for circular smooth muscle contraction; and, in the appropriate situation, the amplitude of the circular

789

790

Section VI  Stomach and Duodenum Cutaneous electrodes 3 cpm EGG waves 500 µV 60 s Fundus

A Pacemaker region

Corpus

Serosal electrodes 3 cpm slow waves

B

Slow wave

C

Pylorus

(Pacesetter potentials)

D Antrum Duodenum 1 mV 60 s

Figure 48-1.  Gastric electrical activity recorded from electrodes positioned on the serosa of the stomach from the fundus (A) to the antrum (D). Slow waves originate in the pacemaker region located at the juncture of the fundus and the corpus on the greater curvature. Note the fundus does not have slow wave activity (electrode A). Slow waves propagate circumferentially and migrate distally to the pylorus approximately every 20 seconds, or 3 cycles per minute (cpm) (dotted lines with arrowheads). The myoelectrical activity of the slow wave can be recorded with cutaneous electrodes. The summed gastric myoelectrical activity recorded from electrodes positioned on the abdominal surface in the epigastrium is termed an electrogastrogram (EGG), and the normal rhythm is 3 cpm. (Modified from Koch KL. Electrogastrography. In: Schuster M, Crowel M, Koch, KL, editors. Atlas of Gastrointestinal Motility. Ontario, Canada: BC Decker; 2002. pp 185-201.)

smooth muscle contraction increases with the onset of the plateau potentials and action potentials.9,10 The aborad propagation of slow waves linked to plateau potentials (with or without action potentials) is the electrophysiologic basis of gastric peristaltic waves (Fig. 48-2). Thus, the slow waves linked with plateau or action potentials propagate through the corpus and antrum and create moving “ring contractions” that resolve in the antrum or at the pylorus in a terminal antral contraction. The pylorus provides an electrical barrier between the 3 cpm slow wave of the distal antrum and the 12 to 13 cpm slow wave of the duodenum.

INTRACELLULAR ELECTRICAL RECORDINGS FROM GASTRIC SMOOTH MUSCLE CELLS

Intracellular recordings from smooth muscle cells from the different regions of the stomach (fundus to the mid-corpus to the terminal antrum) illustrate the electrophysiologic characteristics that distinguish these regions (Fig. 48-3).1 Key features are (1) regional differences in resting mem­ brane potential, which range from −48 to −75 millivolts (mV); (2) regional differences in threshold for contraction that vary from −52 to −40 mV; and (3) the occurrence of plateau potentials with or without spike potentials.1 The fundic smooth muscle cells are unique because their resting membrane potential lies at or above the threshold for con­ traction (−50 mV), a situation that promotes sustained smooth muscle contraction and ongoing fundic tone. Inhibi­ tory vagal input to the fundus increases during swallowing and results in decreasing muscle tone associated with “receptive relaxation” and the accommodation of swal­ lowed foodstuffs.11,12 Fundic muscle tone decreases in pro­

portion to the intensity and duration of the inhibitory neural discharge. In contrast to the fundus, intracellular recordings from the corpus indicate a lower resting membrane potential of −60 mV. The rapid upstroke depolarization in these cells is followed by a plateau potential that slowly returns to the baseline resting electrical potential (see Fig. 48-3F). The plateau potentials are associated with circular muscle con­ traction activity in the corpus and antrum.1 The plateau potential may be accompanied by action potentials in the corpus and antrum (see Fig. 48-3G and H). Extrinsic stimuli such as release of acetylcholine or stretch of the stomach wall increases the amplitude and duration of the plateau potential and the occurrence of action potentials, resulting in contractions of varying force, as seen in the muscle of the terminal antrum. Depending on the excitatory neural stimuli and the amplitude of plateau potentials and the number of action potentials, peristaltic contraction waves of the circu­ lar muscle layer vary from very-low-amplitude contractions to high-amplitude lumen-occluding contractions. At the pylorus, the plateau potentials have long durations and superimposed action potentials that result in closure of the pyloric sphincter in conjunction with the terminal antral contraction (see Fig. 48-3H and I).1 The membrane potential and the force of smooth muscle contraction also distinguish the fundus, corpus, and antrum (Fig. 48-4). The resting membrane potential of the fundus is approximately −50 mV and produces the sustained contrac­ tion and the resting tone of the fundus.1 This fundic tone ensures a sensitive response to excitatory or inhibitory stimuli for relaxation or contraction of the fundus. Recep­ tive relaxation during ingestion of food is accomplished by these electrophysiologic attributes of smooth muscle in the

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders Cutaneous 3 cpm EGG waves 500 µV 60 s A

Fundus

Pacemaker region Corpus

3 cpm slow waves plus plateau or action potentials Slow wave (Pacesetter potentials)

B

C

Pylorus

Plateau potentials

D Antrum Duodenum

Action potentials Moving peristaltic contraction

1 mV 60 s

Figure 48-2.  Gastric slow waves linked with plateau potentials (or action potentials), the electrophysiological basis of gastric peristaltic waves. The plateau and action potentials occur during circular muscle contractions. Peristaltic waves originate in the pacemaker area. The frequency (3 cycles per minute [cpm]) and propagation velocity (approximately 14 mm/second) of the gastric peristaltic waves are controlled by the slow wave, which leads the contraction from the proximal corpus to the distal antrum, as shown at electrodes A through D. The solid black lines and arrows indicate the circumferential and distal propagation of the peristaltic wave, which forms a ring contraction (small arrow), indicating a moving peristaltic contraction. Peristaltic contractions occur three times per minute, the frequency of the gastric slow wave. The increased myoelectrical activity of the plateau potentials and action potentials linked with the slow wave results in increased amplitude of the EGG signal (thick black lines). The fundus does not participate in the gastric peristaltic contractions. (Modified from Koch KL. Electrogastrography. In: Schuster M, Crowel M, Koch, KL, editors. Atlas of Gastrointestinal Motility. Ontario, Canada: BC Decker; 2002. pp 185-201.)

fundus. In contrast, the resting membrane potentials of the corpus and antrum are −60 to −70 mV, respectively. In the presence of plateau potentials or action potentials the membrane potential reaches −45 mV or less and smooth muscle contraction occurs. If the plateau potentials have higher amplitude, then contractions of larger amplitude or force occur. When the plateau potential and action poten­ tials are linked to the propagating slow waves in the antrum, then the moving ring contractions of the gastric peristaltic “waves” are formed (see Fig. 48-2). In conjunction with terminal antral contractions, the pyloric sphincter contraction prevents emptying of gastric content into the duodenum and results in retention of solid foodstuffs in the stomach. Thus, the peristaltic waves asso­ ciated with terminal antral and pyloric sphincter contrac­ tion produce little or no emptying of the gastric contents from the stomach into the duodenum. In contrast, if the pylorus remains open during the gastric peristaltic wave, then an aliquot of nutrient chyme is emptied into the duodenum.

INTERSTITIAL CELLS OF CAJAL

ICCs are the “pacemaker cells” for the smooth muscle appa­ ratus of the gastrointestinal tract.7,8,13,14 ICCs originate from c-Kit–positive mesenchymal cell precursors.15 ICCs in the stomach are located in submuscular, intramuscular, myen­ teric, and subserosal layers of the gastric wall.16,17 Figure 48-5 shows the anatomic relationships between the ICCs in the myenteric plexus (MY-ICCs), the intramuscular ICCs

(IM-ICCs), the enteric neurons, and the circular smooth muscle cells. MY-ICCs are located between the circular and longitudinal muscle layers of the stomach and are the ICCs responsible for the generation of the slow waves (see Fig. 48-5). These ICCs spontaneously generate slow waves that are conducted into adjacent smooth muscle cells and cause depolarization and contraction of the smooth muscle by activating voltage-dependent, dihydropyridine-sensitive (L-type) calcium channels (see Fig. 48-5).18-20 Increased amplitude of the plateau potential correlates with increased amplitude of smooth muscle contraction (Fig. 48-6). The slow waves propagate circumferentially and distally through the ICC network via gap junctions and entrain more distal ICCs with slower intrinsic frequencies to the higher slow wave frequency, the pacemaker frequency. ICCs are also located within the layers of the circular smooth muscle (IM-ICCs), where they integrate and coordi­ nate the spread of the slow wave and the smooth muscle contraction initiated by the MY-ICCs.10 Slow waves are not regenerated in the smooth muscle cells because the ion channels needed to generate and propagate slow waves are not expressed by gastric smooth muscle. In the corpus and antrum the MY-ICCs form a continuous lattice-like network of interconnections that extend from the pacemaker region circumferentially and aborally to the pylorus. The MY-ICCs establish the dominant pacemaker frequency, and IM-ICCs carry the slow wave into the circu­ lar smooth muscle bundles to coordinate circumferential and aboral propagation of the contraction wave.

791

Section VI  Stomach and Duodenum Fundus

–18 –48

A Very orad corpus

–21 –51

B Orad corpus

–26 –56

C Mid orad corpus

–30

D

1.5 Force of contractions (gF)

792

Fundus

1.0

Corpus 0.5 Resting potential

Antrum

–60 0.0

Caudad corpus –23

E

–63 Orad antrum –49

F

–69 Orad terminal antrum –27

G

–70

–75 –70

–60

–50

–40

–30

Membrane potential (mV) Figure 48-4.  Relationship between membrane potential in millivolts (mV) and force of contraction in grams (gF) in gastric smooth muscle from the fundus, corpus, and antrum. The resting membrane potential in fundic smooth muscle is approximately −50 mV, a potential that generates muscle contraction and “resting tone” in the fundus. The resting membrane potential in the antrum smooth muscle is −70 mV, almost 30 mV below the threshold for smooth muscle contraction. When the resting membrane potential reaches −40 or −35 mV, the steep slope of the voltage-contraction curve is observed in the corpus and antrum. (From Szurszewski JH. Electrophysiological basis of gastrointestinal motility. In: Johnson LR, editor. Physiology of the Gastrointestinal Tract. 2nd ed. New York, NY: Raven Press; 1986. p 383.)

Caudad terminal antrum –28

H

–71 Pylorus

I

10 sec 0 –75

10 sec Figure 48-3.  Intracellular electrical recordings from smooth muscle from the fundus to the pylorus (A through I). Resting membrane potential in millivolts (mV) is shown on the vertical axis and time is shown on the horizontal axis. Distinctive electrical characteristics in each region are shown: A, Spontaneous electrical activity in the fundic smooth muscle is absent. B through E, The resting membrane potential is less negative in the smooth muscle in the corpus compared with the antrum (F through H). Spontaneous upstroke depolarization is also recorded in the corpus and antrum, as well as in the pylorus (panel I). The upstroke depolarization in the smooth muscle is initiated by interstitial cells of Cajal (see text). The upstroke depolarization is followed by the plateau potential and repolarization (D through I). The upstroke depolarization and plateau potentials are associated with contraction of the smooth muscle. Action potentials are superimposed on the plateau potentials in the terminal antrum and pylorus (G through I) and are associated with increased amplitude of smooth muscle contraction. (From Szurszewski JH. Electrophysiological basis of gastrointestinal motility. In: Johnson LR, editor. Physiology of the Gastrointestinal Tract. 2nd ed. New York, NY: Raven Press; 1986. p 383.)

The ICCs have innate rhythmicity that is based on their unique metabolism and fluxes in intracellular and extracel­ lular calcium.19,21 The most active area of depolarization and repolarization of the ICCs is in the pacemaker area of the stomach located between the fundus and the proximal corpus (see Fig. 48-1). The depolarization and repolariza­

tion of the ICCs is regenerated and propagated through the network of ICCs in a migrating wave front that moves from the pacemaker region on the greater curve through the corpus and antrum to the pylorus, as shown in Figures 48-1 and 48-2, proscribing the pathway of gastric peristaltic con­ tractions. Excitatory inputs (e.g., cholinergic stimuli, stretch) to the MY-ICC results in opening of calcium channels and depolarization of the smooth muscle cells with IM-ICC acti­ vation to coordinate the contractions of the circular muscle cells in time and space. Thus, the ICC networks provide the control of frequency and propagation velocity for the circu­ lar muscle contractions that comprise gastric peristalsis waves. The fundus of the stomach lacks slow waves (see Figs. 48-1 and 48-3A). The IM-ICCs in the fundus have a role in mechanoreception and act as sensory cells with intercon­ nections to the vagal afferent neurons that innervate the fundus.22,23 Fundic IM-ICCs are also innervated by inhibi­ tory vagal neurons that regulate tone in the fundus.23 Thus, the ICCs also participate in the relaxation of fundic tone that occurs during accommodation.24 Loss of ICCs in the antrum is associated with gastroparesis in patients with diabetes mellitus.25 Patients with diabetic gastroparesis and severe loss of ICCs have more gastric elec­ trical dysrhythmias (tachygastria), more upper gastrointes­ tinal symptoms, and poorer response to gastric electrical stimulation compared with patients with normal numbers of ICCs.26 Electrical dysrhythmias and gastroparesis were recorded in rodents with experimental diabetes.27 Interrup­ tion of ICC pathways from nondiabetic mechanisms also results in gastric dysrhythmias and ectopic pacemakers that are similar to gastric dysrhythmias found in patients with diabetes.28,29

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders Active propagation of slow waves in ICC network

Spontaneous activation of pacemaker current

Interstitial cell network in pacemaker region (MY-ICC)

Electrotonic conduction of slow waves via gap junctions

Smooth muscle cells

Depolarization and activation of Ca+ channels in SMC

L-type

Intramuscular ICC (IM-ICC) Enteric motor neuron

Varicose terminals

Neural input to IM-ICC conditions responses of smooth muscles to slow waves

IM-ICCs and MY-ICCs are electrically coupled to smooth muscle cells via gap junctions Figure 48-5.  Relationships among interstitial cells of Cajal (ICCs), smooth muscle cells in the circular muscle layer, and motor neurons of the enteric nervous system. ICCs in the region of the myenteric plexus (MY-ICCs) are pacemaker cells and spontaneously generate slow wave depolarizations. Slow waves conduct to adjacent smooth muscle cells via low-resistance junctions (gap junctions) as shown by the curved arrow. Depolarization of smooth muscle cells leads to activation of L-type calcium channels, Ca2+ entry, and contraction of the smooth muscle cells. Thus, slow waves naturally organize the contractile pattern of gastric smooth muscles into a series of phasic contractions. Smooth muscle cells do not possess the ionic mechanisms necessary to regenerate slow waves, so the amplitude of slow waves decreases as slow waves conduct from smooth muscle cell to smooth muscle cell in a muscle bundle. Active propagation of slow waves from the dominant (i.e., highest frequency) pacemaker along the greater curvature of the gastric corpus to the pyloric sphincter requires a continuous coupled network of MY-ICCs. SMC, smooth muscle cell. A second class of ICCs lies within circular layers of smooth muscle bundles and are known as intramuscular ICCs (IM-ICCs). The IM-ICCs appear to be important in mediating neurotransmission because they form very close, synaptic connections with the varicose terminals of enteric motor neurons (short arrows). IM-ICCs are also electrically coupled via gap junctions to smooth muscle cells. Postjunctional neural responses can be conducted from IM-ICCs to muscle bundles. Thus, stimulation of excitatory enteric neurons leads to depolarization of IM-ICCs; and the depolarization, conducting through gap junctions to the MY-ICCs network, causes positive chronotropic effects on the frequency of gastric slow waves. Depolarization of IM-ICCs in response to excitatory nerve stimulation also increases the contractile responses of smooth muscle cells to slow wave depolarizations initiated by MY-ICCs. Stimulation of inhibitory enteric neurons causes hyperpolarization and stabilization of membrane potential and tends to inhibit contractile responses to slow wave depolarization. (Figure courtesy of KM Sanders.)

TENSION

0.5 g

A MP Plateau phase

Upstroke

–40

mV

–70

B

RMP

Figure 48-6.  Relationships between smooth muscle contraction (tension) and membrane potential (MP). In these intracellular recordings from antral smooth muscle, the upstroke potential is the rapid depolarization (upstroke) event, followed by the plateau phase. The plateau potentials are associated with contraction of the smooth muscle cell, as shown in panel A. Note that an increase in the amplitude of the plateau potentials (red line in panel B) is associated with greater contractility (tension). During redepolarization to the resting membrane potential (RMP), the contraction resolves. (Modified from Sanders KM, Ordog T, Koh SD, Ward SM. Properties of electrical rhythmicity in the stomach. In: Koch KL, Stern RM, editors. Handbook of Electrogastrography. New York, NY: Oxford Press; 2004. pp 13-36.)

793

794

Section VI  Stomach and Duodenum NERVOUS SYSTEM INNERVATION

As reviewed earlier in Chapter 47, neurons of the ENS populate the stomach wall from the fundus to the pylorus.30 These neurons are located in the myenteric plexuses between the circular muscle and the longitudinal muscle layers (see Fig. 48-5). Neurons of the ENS are also located in submucosal and subserosal plexuses. The ENS provides local reflex circuits within the gastric wall: (a) sensory affer­ ent neurons located in the mucosa linked to (b) interneurons in the myenteric plexus that are linked to (c) efferent neurons that innervate the smooth muscle and glands to perform the gastric secretomuscular functions.31,32 Release of excitatory neurotransmitters such as acetylcholine and substance P stimulates smooth muscle contractions, whereas inhibitory neurotransmitters such as nitric oxide and vaso­ active intestinal polypeptide inhibit contractions. These enteric neural circuits within the gastric wall are pro­ grammed to modulate peristaltic contractions (in conjunc­ tion with ICC activity described earlier) by sequential inhibition of the distal smooth muscle segment and contrac­ tion at the immediate proximal segment of the stomach wall.33,34 Serotonin in the bowel wall has a primary role in initiating and controlling peristaltic events.32 Neurons of the ENS are located in proximity to the MY-ICCs and IM-ICCs (see Fig. 48-5).35 The ENS neurons provide additional control and modulation of contraction and relaxation of the gastric smooth muscle via cholinergic excitation and nitrergic inhibitory neurotransmission. Neurons of the ENS form gap junctions with MY-ICCs and IM-ICCs and provide an additional layer of neural control that integrates slow wave activity and smooth muscle activ­ ity. Thus, postganglionic excitatory and inhibitory neurons innervate MY-ICCs to modulate gastric neuromuscular con­ traction and relaxation and provide chronotropic effects on the slow waves. The PNS and SNS modulate gastric neuromuscular activ­ ity. The vagus nerve provides the PNS input for the stomach, although approximately 80% of vagal fibers are afferent neurons. The afferent neurons are responsive to moment-tomoment contraction and relaxation (tone) of the stomach wall.36 Efferent activity of the vagus nerve increases the release of acetylcholine, which increases the amplitude of gastric contractions and stimulates secretion of gastric acid and pepsin. The SNS innervates gastric smooth muscle with neurons that travel with the splanchnic vasculature. SNS activity generally elicits inhibitory action on the smooth muscle via effects on the myenteric neurons of the ENS.37 The release of various hormones, ranging from cholecys­ tokinin (CCK) to gastrin, affects the neuromuscular activity of the stomach. Gut hormones produce their effect on smooth muscle, ICCs, ENS as well as vagal efferent or affer­ ent functions. These effects will be characterized under fasted and fed conditions and discussed in more detail following.

GASTRIC NEUROMUSCULAR ACTIVITY DURING FASTING

In the fasting state, electrical and contractile events of the corpus or antrum occur in a highly regular pattern termed the migrating myoelectrical (or “motor”) complex, or MMC.38,39 The three phases of the MMC, as described by changes in intraluminal contractions, recur approximately every 90 to 120 minutes. Phase 1 is a period of quiescence wherein little or no contractile activity is recorded. Phase 1 is followed by phase 2 during which random, irregular contractions occur. Phase 3 of the MMC is a burst of regular, high-amplitude phasic contractions that last from 5 to 10

minutes (Fig. 48-7A). Phase 3 contractions are also termed the “activity front.” The activity front migrates from the antrum to the ileum, a journey of 90 to 120 minutes’ dura­ tion. The three phases of the MMC occur regularly in the small intestine, whereas approximately 50% of the phase 3 activity fronts originate in the stomach and then migrate through the small intestine.40 The MMCs that originate in the stomach or duodenum travel through the small intestine and terminate in the distal ileum. If fasting continues, then another phase 3 activity front reappears in the antrum or duodenum at the 90- to 120-minute interval. The highamplitude, three-per-minute contractions of phase 3 that develop in the distal antrum empty nondigestible, fibrous foodstuffs that remain in the stomach after a meal. Cyclic contractile activity associated with the onset of phase 3 also has been identified in the lower esophageal sphincter, the sphincter of Oddi, and the gallbladder. The phase 3 contractions correlate with rapid eye movement (REM) sleep and are related to a larger system of biological clocks.38,41 MMCs develop after vagotomy, indicating that nonvagal mechanisms initiate and sustain MMC neuromus­ cular activity. Motilin is released during the intense phase 3 contractions that occur in the proximal duodenum.42 Motilin appears to be important for phase 3 activity of the MMC because administration of a motilin-neutralizing anti­ body abolishes the phase 3 contractions.

GASTRIC NEUROMUSCULAR ACTIVITY AFTER A MEAL

Three basic gastric neuromuscular activities occur during and after ingestion of solid foods: (1) receptive relaxation to accommodate the ingested food; (2) trituration of the ingested solid food by recurrent corpus-antral peristaltic waves to produce chyme; and (3) antral peristalsis with antropyloroduodenal coordination to empty chyme in small aliquots into the duodenum in a controlled manner for optimal digestion and absorption of the nutrients.

Gastric Neuromuscular Response to the Ingestion of Solid Foods

The neuromuscular work of the stomach in mixing, milling, and emptying food depends upon the physical characteris­ tics, volume, and the fat, protein, and carbohydrate content of the ingested food. For example, almost 240 minutes of neuromuscular work is required to empty approximately 95% of a 255-kcal low-fat, egg substitute sandwich.43 In contrast, 35 minutes of gastric neuromuscular work is required to empty almost 70% of a 20-kcal 500-mL soup broth meal that was consumed in four minutes.44 Each meal requires its own specific time to be emptied, an emptying time achieved by the distinct gastric neuromuscular “work” elicited by the specific food. Figure 48-8 illustrates gastric neuromuscular activity required to receive, mix, and empty a solid meal. The spectrum of gastric work extends from fundic relaxation to gastric peristalsis to antropyloroduode­ nal coordination, the work that is needed to produce chyme and empty it into the duodenum. Ingestion of food abolishes the fasted state as regular three per minute gastric peristalsis begins in the corpus and antrum to mix the food; and in the fed state, a pattern of continuous small bowel con­ tractions with short runs of peristalsis over distances of 2 to 4 cm optimize digestion and absorption of nutrients (see Fig. 48-7B). Solid food delivered from the esophagus into the fundus is associated with receptive relaxation of the fundus, the “work” of fundic muscle relaxation. As the fundic smooth

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders Pressure cm H2O 1

Antral

3

Antral

5

Antral

2

Duo

4

Duo

6

Duo

Pressure cm H2O

Phase 3

100

100 Antral 0 100 Antral Antral 0 Duo 2 100 Antral 0 Antral 2 100 Duo 0 100 Duo 1 Duo 0 100 Duo 3 Duo 0 Antral 1

0 100 0 100 0 100 0 100 0 100 0

Tasks Event Meal Supine Symptom EPS-1

Tasks Event Meal Supine Symptom EPS-1

1/09:50

1/09:50 Minutes

A

1/10:00

Eat

2/08:20

B

2/08:30 Minutes

2/08:40

Figure 48-7.  Antroduodenal motor activity in a healthy subject. A, Fasted state: Phase 3 contractions in the antrum (antral) and duodenum (Duo). Intraluminal contractions in the antrum (channels 1, 3, and 5) and the duodenum (channels 2, 4, and 6) are shown. A phase 3 activity front with three-minute antral peristaltic contractions lasting almost six minutes is noted in channels 1, 3 and 5. The phase 3 activity front propagates distally and migrates past the duodenal recording ports. The frequency of contractions in the duodenum is approximately 11 or 12 per minute, the same as the frequency of the duodenal slow wave. After completion of the phase 3 contractions, the quiescence of phase 1 and lack of contractions are seen in the antrum. B, Fed state: The subject ingested a standard liquid meal (Ensure). Contractions of variable amplitude are seen in the antrum and a series of relatively lowamplitude, irregular contractions are noted in the duodenum, all of which represent the fed state and are in marked contrast to phase 3 activity during fasting shown in panel A. (Modified from Koch KL. The stomach. Manometry. In: Schuster M, Crowell M, Koch KL, editors. Atlas of Gastrointestinal Motility. Ontario, Canada: BC Decker; 2002. pp 135-150.)

Fundic relaxation

Antral peristalsis – emptying

Fundic contraction – emptying

Pyloric resistance

Corpus – antral filling/mixing Duodenal resistance Antropyloroduodenal coordination Figure 48-8.  Spectrum of gastric neuromuscular work after ingestion of a solid meal. To receive the ingested solid foods and accommodate the volume of food without increasing intragastric pressure, the fundic smooth muscle relaxes (receptive relaxation). The fundus then contracts to empty the ingested food into the corpus and antrum for trituration and emptying. Recurrent corpus-antral peristaltic waves mill the solids into chyme, which is composed of 1- to 2-mm solid particles suspended in gastric juice. Antral peristaltic waves, indicated by the ring-like indentation in the antrum, empty 2 to 4 mL of the chyme through the pylorus and into the duodenal bulb at the slow wave frequency of three peristaltic contractions per minute. Antropyloroduodenal coordination indicates efficient emptying of chyme through the pylorus, which modulates flow of the chyme by varying sphincter resistance. Contractions in the duodenum also provide resistance to emptying. (Modified from Koch KL. Physiological basis of electrogastrography. In: Koch KL, Stern RM, editors. Handbook of Electrogastrography. New York, NY: Oxford Press; 2004. pp 37-67.)

795

Section VI  Stomach and Duodenum 500 0 min

450 400 Intragastric volume (mL)

796

30 min

60 min

180 min

240 min

Fundus

350

Antrum

300 250

120 min

200 150 100 50 0 –30–20–10 0 10 20 30 40 50 60 70 80 90 100110120 Time after the meal (min)

Figure 48-9.  Gastric accommodation of the fundus and proximal stomach in a healthy volunteer after a test meal. The intragastric volume, measured with a barostat balloon, increases from approximately 200 mL to approximately 450 mL during the 20 minutes after the meal is ingested. As the meal is emptied, the volume within the stomach slowly decreases over the two-hour postprandial period. Relaxation of the proximal stomach and accommodation of the meal volume reflect vagal-mediated receptive relaxation. (From Tack J, Piessevuax H, Coulie B, et al. Role of impaired gastric accommodation to a meal in functional dyspepsia. Gastroenterol­ogy 1998; 115:1346-52.)

muscle relaxes, larger amounts of solid or liquid food are accommodated in the fundus and proximal corpus with little or no increase in intraluminal pressure. Liquids, in contrast, are immediately distributed throughout the antrum and corpus (emptying of liquids is discussed in the next section). Relaxation of the fundus occurs before the work of trituration in the corpus-antrum and is a vagal nerve– mediated event that requires nitric oxide.45,46 Figure 48-9 shows an example of the changes in intragastric volume during relaxation of the fundus and proximal corpus in response to a caloric meal.47 Relaxation of the fundus and the stimulation of mechanoreceptors (stretch), mediated through IM-ICCs in the fundic wall, activate vagal afferent neurons and vagovagal reflexes. These reflexes involve the nucleus of the tractus solitarius and efferent neurons from the dorsal motor nucleus of the vagus. Vagal excitatory neurons are inhibited and the vagal inhibitory neural trans­ mitters nitric oxide and vasoactive intestinal peptide (VIP) are released to accomplish receptive relaxation. Other factors influence the muscle tone of the fundus. Antral distention, duodenal distention, duodenal acidifica­ tion,48 intraluminal perfusion of the duodenum with lipid or protein, and colonic distention all decrease fundic tone through various reflexes. The gastric reflex is mediated through an arc initiated by capsaicin-sensitive afferent vagal nerves and is mediated by 5-hydroxytryptamine-3 (5-HT3), gastrin-releasing peptide (GRP) and CCKA receptors.49 Solid foods labeled with technetium are accommodated initially in the fundus and proximal corpus, and by obtain­ ing frequent scintigraphic images, the distribution of the labeled solid meal can be followed over four hours using scintigraphic methods.43 Figure 48-10 shows that imme­ diately after ingestion of this solid meal, the food is accom­ modated and the majority of the meal is retained in the

Figure 48-10.  Gastric emptying of an egg sandwich. One-minute scin­ tigraphic images of a radiolabeled 255-kcal substitute egg meal in the stomach at time 0, 30, 60, 120, 180, and 240 minutes after ingestion are shown. The yellow and pink areas indicate regions of the stomach with higher isotope counts and more food than the other regions. Note the persistence of portions of the meal in the fundus at 120 minutes after ingestion. The meal is slowly redistributed from the fundus to the antrum for trituration and emptying. Only a small amount of the meal remains in the stomach by 240 minutes, and most of the labeled eggs are in the small intestine.

fundus and proximal corpus. Subsequently, contractions of the fundus press portions of the food into the corpus and antrum for trituration. This early postprandial period of accommodation and trituration, that occurs before gastric emptying of the nutrients, is termed the lag phase. The lag phase may last from 45 to 60 minutes for solid foods, but the duration of the lag depends on the components of the meal, the thoroughness of chewing the food, and the time required to ingest the meal. For the 255-kcal egg substitute test meal that is ingested in a 10-minute period, the lag phase is 30 to 45 minutes. Once portions of the meal have been triturated into 1- to 2-mm particles suspended in gastric juice, the linear phase of gastric emptying of the chyme begins. Recurrent gastric peristaltic waves mix saliva, acid, and pepsin with the chewed food and then mill the food to produce chyme. The normal peristaltic waves occur every 20 seconds, generated by 3 cpm slow waves linked to plateau and action poten­ tials. In healthy subjects approximately 60% of the egg substitute meal is emptied in two hours and more than 95% is emptied at four hours (Fig. 48-11).43 During the linear phase of gastric emptying, each peristal­ tic wave empties from 3 to 4 mL of chyme into the duode­ num.50,51 Movement of chyme into the duodenum is usually, but not always, pulsatile due to the systole-like effect of antral peristaltic waves.51 The volume of chyme delivered into the duodenum by each peristaltic wave is modulated by the configuration of the peristaltic wave (e.g., depth of contraction, length of the peristaltic wave) pressure within the stomach, and resistance to flow provided by the pyloric sphincter and duodenal contractions.52,53 The gastric peristaltic wave delivers a larger stroke volume when the pylorus and the duodenum are relaxed to receive the aliquot of chyme, but the overall rate of calories delivered each minute to the duodenum is consistent at approximately 3 to 4 kcal/min.54

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders 120 110 100 90

% gastric retention

80 70 60 50 40 30 20 10 0 0

30

60

90

120

150

180

210

240

Time (min) Figure 48-11.  Solid phase gastric emptying curve for 123 subjects after ingestion of a 255-kcal substitute egg meal, the same meal shown in Figure 48-10. Note that only approximately 15% of the eggs are emptied in the first 45 minutes, the lag phase of gastric emptying of this meal. At 90 minutes approximately 50% of the meal has been emptied and 50% is retained. By 240 minutes more than 95% of the meal has been emptied. (Modified from Tougas G, Eaker EY, Abell TL, et al. Assessment of gastric emptying using a low fat meal: Establishment of international control values. Am J Gastroenterol 2000; 95:1456-62.)

As time elapses after ingestion of the meal, the chewed food is continually redistributed from the fundus to the antrum for trituration (see Figs. 48-8 and 48-10). Some gastric peristaltic waves end at various points in the antrum and others end with a terminal antral contraction associated with closure of the pylorus that prevents the emptying of larger food particles or indigestible solids. These terminal antral and pyloric contractions result in retention of the solid particles in the corpus and antrum. In this manner, solid food particles that require further trituration are retained and subjected further to the milling effects of the recurrent peristaltic waves (see Figs. 48-7B and 48-8). The intragastric pressure and gastric intraluminal pH values recorded after a healthy subject ingested an egg substi­ tute meal are shown in Figure 48-12. During the first 2 hours after the meal the amplitude of intraluminal contractions varies from 10 to 40 mm Hg. Approximately three and a half hours after the solid meal was ingested, high amplitude con­ tractions (>65 mm Hg) occur just before the pH increases from 1 to 6 as the motility/pH capsule is emptied from the acidic antrum into the alkaline environment of the duodenum. After the digestible components of the meal are emptied, strong antral contractions (phase 3–like contractions) empty the capsule from the stomach into the duodenum.55 Thus, fibrous and indigestible materials are emptied by high-amplitude, antral contractions, whereas the digestible nutrients in the chyme are emptied earlier by the lower-amplitude peristaltic waves during the linear phase of emptying.56 The pylorus modulates the rate of gastric emptying by several mechanisms. Increased pyloric tone and isolated pyloric pressure waves prevent gastric emptying and promote retention of food for further milling. Pyloric con­ tractions associated with terminal antral contractions are common during the lag phase when trituration is occurring. Once the linear phase of emptying begins, the numbers of isolated pyloric contraction waves diminish as chyme is available for emptying via the gastric peristaltic waves.

250

10

225

9

205 pH

Pressure (mm Hg)

185

8 7

165

6

145 125

5

pH

105

Pressure

pH

4

85

3

65 45

2

25

1

0:03

0:28

0:57

1:26

1:55

2:24

2:52

3:21

3:50

0 4:19 4:30

Time (hours) Figure 48-12.  Gastric contractions and intraluminal gastric pH recordings during the emptying of a 255-kcal egg substitute meal, the same meal shown in Figure 48-10, recorded with an ambulatory capsule pH and motility device. pH is shown on the right vertical axis and pressure in mm Hg is shown on the left vertical axis. The pH increases to approximately 3 for the first 45 minutes as gastric acid is buffered by the meal. The pH then gradually decreases to 1 and remains near 1 at about 3 hours after ingestion of the meal. Stomach contractions are generally of low amplitude, less than 10 mm Hg after ingestion of the meal. At approximately 3 hours and 40 minutes after the meal, the recorded pH increases abruptly to 7 and then decreases and remains stable at around 6. Prior to the abrupt increase in pH, there is a series of clustered, high-amplitude antral contractions (pressure). These antral contractions empty the capsule from the antrum (pH 1) into the duodenum, where the pH is 6 or more. The contractions that occurred during the 3 hours and 50 minutes required to empty the meal document the neuromuscular work required to triturate and empty this meal in a healthy subject.

797

Section VI  Stomach and Duodenum SOLID PHASE MEAL

% gastric retention

100

Fundus

75

50

25

A

B

Antru

m

Figure 48-13.  Three-dimensional ultrasound images from the stomach before and after a healthy subject ingested a 500-mL soup meal. A, In the fasted state, intragastric volume is approximately 38 mL. B, Ten minutes after ingestion of the meal the stomach volume is 350 mL. Note that the antrum, corpus, and fundus are now distended, indicating the marked relaxation of the smooth muscle required to accommodate this volume of liquid. (Modified from Gilja OH, Detmer PR, Jong JM, et al. Intragastric distribution and gastric emptying assessed by three-dimensional ultrasonography. Gastroenterology 1997; 113:38-49.)

Gastric Neuromuscular Response to the Ingestion of Liquids

0

us

rp

Co

The gastric neuromuscular activity required to mix and empty liquids from the stomach is distinctly different com­ pared with the emptying of solid foods.55-59 Figure 48-13 shows three-dimensional (3-D) ultrasound images of the stomach in a healthy subject during the fasting state and 10 minutes after the subject ingested 500 mL of soup.58 The intragastric volume was approximately 40 mL during fasting and increased almost nine-fold to 350 mL 10 minutes after ingestion of the meal, indicating the remarkable relaxation of the smooth muscle of the antrum and corpus (in addition to the fundic relaxation) that was required to accommodate this liquid volume. (In contrast, solid meals are initially accommodated and retained primarily in the fundus and proximal stomach, as shown in Figs. 48-9 and 48-10.) Once accommodated, nutrient liquids are emptied into the duo­ denum in a controlled but more rapid rate compared with solid foods, which require trituration. Noncaloric liquid meals empty without the lag phase in a curve described as monoexponential emptying (Fig. 48-14).60 Caloric-dense liquids on the other hand are retained for longer periods in the antrum and are emptied slower than noncaloric liquids. Liquids are emptied from the stomach by a combination of (a) pressure gradients between the stomach and the duode­ num that produce flow of liquid into the duodenum; (b) antral peristaltic contractions that produce a pulsatile pattern of emptying of liquids from the antrum into the duodenum; and (c) duodenogastric reflux events that modify gastric emptying results.51,52 From a gastric myoelectrical activity viewpoint, ingestion of water until the point of full­ ness induces a brief “frequency dip” followed by normal 3 cpm activity on the electrogastrogram (EGG) (Fig. 48-15).61 The rate of gastric emptying of liquids is influenced by the volume, nutrient content, viscosity, and osmolarity of the ingested liquid.53,54,59,62 These factors affect the neuro­ muscular activity of the stomach, which ultimately pro­ duces the rate of emptying. These factors are discussed below.

0

60

A

120

180

240

Time (min) LIQUID PHASE MEAL 100

% gastric retention

798

75

50

25

0 0

B

60

120

180

240

Time (min)

Figure 48-14.  Emptying of a mixed liquid and solid meal in healthy subjects, who ingested 300 mL of radiolabeled water with two radiolabeled eggs and toast. A, Emptying rate for the solid phase of the meal. A short lag phase is noted before the linear phase of emptying, and by 60 minutes approximately 55% of the meal is emptied (45% is retained). The lag phase may be shortened if the subject has taken a relatively long time to eat the meal or the solids require little trituration. B, Empyting rate for the liquid phase of the meal. Approximately 80% of the water is emptied (20% is retained) at 60 minutes, as the liquid is rapidly distributed throughout the antrum and corpus. This is considered a monoexponential liquid emptying curve. (Modified from Maurer AH, Parkman HP, Knight LC, Fisher RS. Scintigraphy. In: Schuster M, Crowel M, Koch, KL, editors. Atlas of Gastrointestinal Motility. Ontario, Canada: BC Decker; 2002. pp 171-184.)

REGULATION OF GASTRIC NEUROMUSCULAR ACTIVITY AFTER A MEAL Gastric emptying rates are regulated to achieve a consistent, regular presentation of chyme to the duodenum in order to optimize secretion of pancreatic enzymes and bile appropriate for digestion of the contents of the chyme. Various gastric emptying rates are achieved by variations in the neuromuscular armamentarium of the stomach: fundic relaxation and contraction; the characteristics of gastric peristaltic contractions; temporary suspension of 3 cpm slow waves and the onset of gastric dysrhythmias; the coor­

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders PERCENTAGE DISTRIBUTION OF EGG POWER 1–2.5 cpm (bradygastria) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

2.5–3.75 cpm (normal)

3.75–10.0 cpm (tachygastria)

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

OBL

10

20

30

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

OBL

Minutes

10

20

30

10.0–15.0 cpm (duodenal – resp.) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

OBL

10

Minutes

20

30

OBL

Minutes

10

20

30

Minutes

RUNNING SPECTRAL ANALYSIS EGG RHYTHM STRIPS

Power Time (1 minute between lines)

B1

60s

B Resp.

500 µV Water load A Resp.

A1

0

3

60s

6

9

12

15

Frequency (cpm) Figure 48-15.  Running spectral analysis of the electrogastrogram (EGG) signal and EGG rhythm strips before and after ingestion of a water-load in a healthy subject. The X-axis shows the frequencies in the EGG signal in cycles per minute (cpm). The Y-axis indicates time, and the peaks (or Z-axis) indicate the power of the frequencies contained in the EGG signal. The baseline EGG rhythm strip (A) shows 3 cpm activity. The regularity and the amplitude of the 3 cpm EGG signal is increased (B) after the subject ingested 750 mL of water (water-load arrow) over a 5-minute period. The running spectral analysis shows relatively low power 3 cpm peaks at baseline (A1). After ingestion of the water load the peaks initially disappear (the frequency “dip”) and then 3 cpm peaks emerge and are prominent until the end of the 30-minute recording (B1). This is a normal gastric myoelectrical response to the filling of the stomach with water and the subsequent emptying of the water. The four graphs in the red insert show the percentage distribution of EGG power in the four relevant frequency ranges during baseline (BL) and the 10, 20, and 30 minutes after ingestion of the water by the subject (green lines). Normal ranges are shown by blue lines. Note the initial decrease in the percentage of normal EGG activity (2.5-3.75 cpm) 10 minutes after ingestion of the water (the frequency dip), followed by increased percentages in the 3 cpm normal range 20 and 30 minutes after ingestion of the water. Resp., respiratory activity. (Modified from Koch KL. Physiological basis of electrogastrography. In: Koch KL, Stern RM, editors. Handbook of Electrogastrography. New York, NY: Oxford Press; 2004. pp. 37-67.)

dination of antropyloroduodenal contractions and duodenal contractions; or resistances that promote duodenogastric reflux. The attributes of a specific meal stimulate the appropriate gastric neuromuscular responses that affect the rate of gastric emptying. Table 48-1 lists gastric neuro­ muscular factors, meal-related factors, and other factors that modulate the rate of gastric emptying. The rate of gastric emptying is decreased by the temporary occurrence of gastric dysrhythmias, modulation of the amplitude and the propagation distances of antral contractions, enhanced contractions of the pylorus, and reduced antropyloroduo­ denal coordination.

Meal-related factors that affect gastric emptying include the digestible components of the solids and liquids, fat content (nutrient density), viscosity, acid content, volume, and indigestible foodstuffs. For example, foods with high fat content empty slower than foods with high protein or carbohydrate content. Triglycerides are mixed with gastric lipase during the initial intragastric phases of digestion (Chapter 49) and are broken down to fatty acids and monoor diglycerides before emptying into the duodenum.63 The duodenum is exquisitely sensitive to diet-derived fatty acids. Longer chain fatty acids (longer than C12) exposed to the mucosa of the duodenum result in release of CCK.

799

800

Section VI  Stomach and Duodenum Table 48-1  Factors That Modulate the Gastric Emptying Rate

Factors Gastric Neuromuscular Tachygastria Decreased fundic accommodation Increased fundic accommodation Antral hypomotility Pylorospasm Antroduodenal dyscoordination Meal-Related Factors Volume Increased acidity Increased osmolarity Nutrient density: fat > protein > CHO Tryptophan Undigestible fibers Small Intestinal Factors Fatty acids in duodenum Fatty acids in ileum Colonic Factors Constipation, IBS Other Factors Hyperglycemia Hypoglycemia Illusory self-motion (vection)

Proximal

Effect on Rate of Gastric Emptying (Symptoms) Delay (nausea) Acceleration (early satiety) Delay (prolonged fullness) Delay (prolonged fullness) Delay (right upper quadrant pain) Delay (prolonged fullness) Proportional to meal volume Delay Delay Delay

Middle

Distal 0.0

2.5

A

5.0

7.5

10.0

7.5

10.0

Time (min)

Proximal

Middle

Delay Delay

Distal

Delay (“duodenal tasting,” “duodenal brake”) Delay (“ileal brake”)

0.0

2.5

5.0

Delay

B

Delay Acceleration Delay

Figure 48-16.  Electrical recordings from electrodes secured to the mucosa of the proximal, middle, and distal antrum in a healthy subject. A, 3 cycle per minute (cpm) electrical slow waves in the proximal, middle, and distal electrode leads. The slow waves are propagated in an aborad direction as indicated by the dotted lines. B, Disruption of propagation and the onset of a 5- to 6-cpm tachygastria in the distal lead during hyperglycemia (glucose clamping), with a blood glucose level of 240 mg/dL. (Modified from Coleski R, Hasler WL. Coupling and propagation of normal dysrhythmic gastric slow waves during acute hyperglycemia in healthy humans. Neurogastroenterol Motil 2009; 21:492-99.

CHO, carbohydrate; IBS, irritable bowel syndrome.

CCK relaxes fundic tone, decreases antral contraction, and increases pyloric tone, all of which result in delay in gastric emptying. In contrast, short- and medium-chain fatty acids (shorter than C12) do not have these neuromuscular effects on gastric emptying rates.54,64 CCK released from the duodenum also activates CCKA receptors on the vagal affer­ ent neurons with synapses in the nucleus tractus solitar­ ius.65 Neurons from the nucleus tractus solitarius ascend to the periventricular nucleus that participate in mechanisms of satiation, and descending vagal efferent neurons from the dorsal motor nucleus of the vagus inhibit gastric emptying and maintain fundic relaxation. The sensitivity of the duo­ denal mucosa to fat and other nutrients led to the concept of duodenal tasting and duodenal brake, sensorimotor events that modulate gastric emptying of nutrients.66-68 Monosaccharides in the duodenum stimulate the release of incretins such as glucagon-like polypeptide-1 (GLP-1), which promotes insulin secretion to match increasing post­ prandial blood glucose levels and decreases antral contrac­ tions.69-71 In order to harmonize the relationships between glucose absorption, glycemia, and insulin secretion, the gastric emptying of carbohydrates is highly regulated.72 Hasler and colleagues showed that hyperglycemia decreases antral contractions and increases gastric dysrhythmias, a “physiologic” gastric dysrhythmia that decreases the rate of gastric emptying (Fig. 48-16).6,73 Glucagon infusions also induce bradygastrias.74 Hyperglycemia increases fundic compliance and decreases sensations related to fundic distention.75 Blood glucose levels greater than 220 mg/dL result in decreased antral contractions, decreased gastric emptying, and induced gastric dysrhythmias,6,73 all of which are gastric neuromuscular activities that reduce gastric emp­ tying and reduce further exposure of the duodenum to nutri­ ents. Hypoglycemia on the other hand increases gastric contractility and emptying.76

Time (min)

The interaction between nutrients in the lumen and the regulation of the rate of gastric emptying continues in the later postprandial period as digestion and absorption of nutrients occur throughout the small intestine. For example, if diet-derived fatty acids or carbohydrates reach the lumen of the ileum, the so-called ileal break is activated and gastric emptying is delayed. Infusion of nutrients into the lumen of the ileum delays gastric emptying,77 an enterogastric reflex mediated in part by peptide YY, CCK, and GLP-1 (see Chapter 1).61,68,77 Regulation of stomach emptying also is achieved by vagus nerve and splanchnic nerve activity that modulates the neu­ romuscular activities of the stomach described earlier. Vagal afferent nerves “monitor” neuromuscular function in the stomach moment by moment, and interactions between afferent vagal nerve activity and the nucleus tractus soli­ tarius and synapses with the efferent vagal nerve output from the dorsal motor nucleus produce an ongoing inter­action of CNS excitatory and inhibitory effects on the stomach. Gastric emptying is delayed during stress. Corticotropin-releasing factor (CRF) plays a role in the mediation of stress and inhibits gastric emptying through central dopamine1 and 2 and vasopressin (AVP) pathways in the periventricular nucleus.78 Other factors that affect the rate of gastric emptying not already mentioned include rec­ tocolonic distention, nausea and vomiting of pregnancy, and vection-induced motion sickness.79 Stimulation of various areas in the CNS affects gastric neuromuscular func­ tion. Illusory self-motion (vection) induces antral hypomo­

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders tility, tachygastria, and decreased gastric emptying.79,80 A series of studies using the experience of illusory self-motion, a unique CNS sensory stimulation, showed that the onset of nausea was associated with tachygastria and increased levels of plasma vasopressin.80,81 Gender affects the gastric emptying rate of a standard meal. Gastric emptying is significantly slower in healthy women compared with men.43,82,83 Gender differences in gastric emptying rates may be related to fluctuations in sex hormones, but phases of the menstrual cycle (variations in estradiol and progesterone concentrations) have not shown consistent relationships with emptying measurements.84 The rate of gastric emptying increases as body mass index rises, a relationship that may be relevant to the onset and maintenance of obesity.

GASTRIC SENSORY ACTIVITIES Free nerve endings in the stomach act as polymodal sensory receptors that respond to light touch or pressure, acid, and other chemical stimuli. Afferent neurons within the stomach

CNS perceptions Discomfort, Nausea Pain

are termed intrinsic primary afferent neurons, or IPANs.85 Cell bodies of IPANs reside in the submucosal or the myen­ teric plexus areas of the stomach wall. IPANs may be acti­ vated by serotonin release from local enterochromaffin cells.32,86 The afferent information in the IPANs is used in local reflexes and provides input to vagal and splanchnic afferent neurons for vagovagal and spinal reflexes, respec­ tively, to subserve transmission of visceral sensory informa­ tion to CNS centers. Vagal afferent neurons whose cell bodies reside in the nodose ganglia connect with the nucleus of the tractus solitarius and second-order neurons connect with higher center of the hypothalamus, and some inputs reach the cortex, where they are consciously perceived as visceral sensations (stomach emptiness or fullness) or symptoms such as nausea or abdominal pain (Fig. 48-17). From the SNS, splanchnic or spinal primary afferent neurons in the gastric wall mediate pain sensations. Cell bodies of these neurons lie in the dorsal horn of the spinal cord with second-order neurons that ascend via the spino­ thalamic and spinoreticular tracks in the dorsal columns. Sensory neurons are thin, myelinated A-delta or unmyelin­ ated C fibers. Spinal afferents include a population of unmy­ elinated C fibers. Capsaicin-sensitive unmyelinated fibers

Visceral perceptions Discomfort, Nausea Pain

Dorsal column

Dorsal root ganglion A-delta fiber

Spinothalamic tract

C fiber

To CNS

IML

Motor n.

Somatic nerves

Vagus nerve

Splanchnic nerves SPINAL CORD

Somatic perceptions Discomfort Pain

SKIN

Celiac ganglia Vertebral ganglia T5-T9

Visceral and motor reflexes

C fiber A-delta fiber Afferent n.

A-delta fiber C fiber Pacemaker region

Efferent n.

Afferent n. Efferent n. Gastric dysrhythmias Bradygastria Tachygastria

STOMACH

60 s

Figure 48-17.  Afferent and efferent neural connections between the stomach and central nervous system (CNS). The vagus nerve contains afferent nerves with A-delta and C pain fibers with cell bodies in the nodose ganglia with connections to the nucleus tractus solitarius (not shown). Low threshold mechano- and chemoreceptors stimulate visceral sensations such as stomach emptiness or fullness and symptoms such as nausea and discomfort. These stimuli are mediated through vagal pathways and become conscious perceptions of visceral sensations if sensory inputs reach the cortex. The splanchnic nerves also contain afferent nerves with A-delta and C fibers that synapse in the celiac ganglia with some cells bodies in the vertebral ganglia (T5-T9). Interneurons in the white rami in the dorsal horn of the spinal cord cross to the dorsal columns and spinothalamic tracts and ascend to sensory areas of the medulla oblongata. These splanchnic afferent fibers are thought to mediate high-threshold stimuli for visceral pain. In contrast to visceral sensations, somatic nerves such as from the skin carry sensory information via A-delta and C fibers through the dorsal root ganglia and into the dorsal horn and then through dorsal columns and spinothalamic tracts to cortical areas of somatic representation. Changes in gastric electrical rhythm, excess amplitude contractions, or stretch on the gastric wall are peripheral mechanisms that elicit changes in afferent neural activity (via vagal and/or splanchnic nerves) that may reach consciousness to be perceived as visceral perceptions (symptoms) emanating from the stomach. IML, intermediolateral nucleus; n., nerve.

801

802

Section VI  Stomach and Duodenum contain neuropeptides such as CGRP, VIP, somatostatin, substance P, and neurokinin A. These fibers are considered to be the primary route of transmission for various pain stimuli from the gut to the CNS. These nerve fibers may respond to inflammatory mediators that also awaken “silent” nociceptive fibers.86 In addition to interacting with IPANs, vagal afferent axons have multiple connections with the enteric neurons and innervate the circular muscle fiber bundles via connections with ICCs.87 Vagal afferent neurons are also sensitive to chemostimuli via mucosal neurons and mechanosensitive neurons and ICCs in the muscle layers. CCK receptors on vagal afferent neurons are primarily activated by physio­ logic mechanical and chemical stimuli from the stomach during fasting and fed conditions. These vagal afferents mediate the sensory response to intraluminal acid and fat. Acid may have a direct action on the nerve endings themselves.88 Nausea is a common sensation that is often attributable to stomach dysfunction. During the illusion of self-motion, gastric dysrhythmias develop as healthy individuals report nausea.89 Plasma vasopressin levels increase in the subjects who develop nausea, but do not increase in those who experience no nausea.90 This brain-gut, gutbrain interaction during illusory self-motion illustrates the temporal relationships between the onset of gastric dysrhythmias in the periphery and acute, severe nausea experience of the subject. On the other hand, mechanical or physical distention of the antrum, but not the fundus, using a balloon induces nausea sensations and gastric dysrhythmias in healthy individuals.91 These studies show that gastric dysrhythmias originate in the antrum in humans and that stretch of the antral wall is another mechanism that elicits gastric dysrhythmias and nausea sensations from the stomach. Distention of the gastric antrum and corpus by the water-load test (rather than a balloon) also elicits the gastric dysrhythmias and nausea in susceptible individuals.61

THE STOMACH AND THE REGULATION OF FOOD INTAKE, HUNGER, AND SATIETY Hunger is a basic human drive, a stressful condition that is eliminated or reduced by the ingestion of food. Hunger is also described as an uncomfortable “emptiness” of the stomach. The ingestion of food elicits relaxation of the stomach musculature (receptive relaxation) and accommo­ dation of the physical volume of the meal; as these gastric neuromuscular events occur, hunger disappears and the comfortable, postprandial sensations of stomach fullness are experienced. The volume of food ingested suppresses hunger and stim­ ulates the sense of fullness more than the calorie content of the meal.92-94 Infusion of nutrients into the stomach induces a greater intensity of fullness or satiety compared with infu­ sion of the same nutrients into the duodenum.67 The sup­ pression of hunger is greater when nutrients are taken by mouth, indicating that CNS, oropharyngeal, and gastric neu­ romuscular factors are integrated to produce the comforts of normal postprandial stomach fullness.95 Healthy individuals usually eat until they are reasonably full. The physiologic attributes of fullness are not com­ pletely known, but the physical stretch on the stomach walls induced by the volume of food ingested and the gastric juice secreted are responsible, in part, for the sense

of postprandial fullness.93,96 Subjects experience a dramatic change from the sensation of stomach emptiness at baseline to the sensation of stomach fullness after ingesting water over a five-minute period. The average volume of water ingested to achieve fullness is 600 mL of water; in contrast, patients with functional dyspepsia ingest, on the average, 350 mL of water on average and feel full, indicating a dis­ turbance in stomach wall relaxation and/or wall tension.61 Similarly, fullness and satiety can be achieved by ingesting a nutrient drink until achieving maximum tolerated satiety.97 The presence of acid or nutrients in the duodenum or an elevated blood glucose level decreases the stomach wall tension.98,99 The ingestion of a solid meal initially elicits fundic relax­ ation, and little emptying of the food occurs during the lag phase. Sensations of fullness continue during the lag phase when the food is being triturated. Once the linear phase of gastric emptying begins, there is a progressive perception of decreasing stomach fullness and increasing stomach empti­ ness over time. Four or five hours after a solid meal, the stomach is indeed empty and the healthy individual feels hungry once again. The physiologic mechanisms of hunger and satiety (and stomach emptiness and fullness) are under intense investi­ gation. In the fasting state plasma motilin levels increase during the phase 3 of the MMC, but correlations between the sensation of hunger and increases of motilin or onset of phase 3 have not been described. As discussed in other chapters, ghrelin is a 28 amino acid peptide secreted from endocrine cells of the oxyntic glands in the gastric fundus.100 Ghrelin levels increase in the plasma during fasting (hunger) and stimulate food intake, probably acting via vagal afferent nerves.101 Orexins or appetite-stimulating peptides are syn­ thesized by neurons in the lateral hypothalamus, promote food intake, and stimulate gastric contractility (in the rat) by actions on the dorsal motor nucleus of the vagus with projections to the gastric fundus and corpus.102 After inges­ tion of food, ghrelin levels decrease103 and are profoundly suppressed after gastric bypass surgery.104 Ghrelin also has promotility effects on the stomach and is being evaluated for the treatment of gastroparesis.105,106 Other hormones are candidates for important roles in the sensation of fullness or satiety, and these hormones are released after the ingestion of meals. CCK is released from the duodenal mucosa exposed to fatty acids as described previously. CCK receptors participate in fullness and nausea sensations elicited by intraduodenal lipid and gastric dis­ tention.107,108 Leptin is synthesized in the stomach and released after food ingestion; circulating leptin reduces food intake via CNS regulation of the arcuate nucleus.71,109 GLP-1 enhances fullness after a standard meal, reduces antral motility, and increases gastric volume.71,110 Apolipoprotein A-IV is released from the small intestine during absorption of triglycerides and decreases food intake and gastric motil­ ity, in part, via CCK and vagal afferent pathways.71,111 Polypeptide-YY (PYY) is released from the ileocolonic area after meals and is an important mediator of the “ileal brake” effect112 and appetite suppression.113,114 The brain and these gut hormones are clearly linked in the regulation of food intake and the regulation of gastric neuromuscular activity that produces stomach empty­ ing.71,114 The cephalic phase of gastric physiology is well known but has not been re-explored for many years. The sight, smell, and taste of food stimulate central vagal effer­ ent activity that increases gastric acid secretion, gastric contractility, and increases 3 cpm gastric myoelectrical activity.115-117 Sham feeding, during which the subject chews

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders

Meal

Meal

Sham feeding

Time in 64 second intervals

1

A

3

5

7

9

11

13

Sham feeding

15

Frequency (cpm)

Time in 64 second intervals

1

B

3

5

7

9

11

13

15

Frequency (cpm)

Figure 48-18.  Gastric myoelectrical response to sham feeding with tasty food and “disgusting” food. A, Running spectral analysis (RSA) of the electrogastrogram (EGG) signal recorded while a healthy subject chewed a warm hot dog and spit it out into a paper bag (sham feeding). The increase in amplitude of the peaks in the normal 3 cycle per minute (cpm) range during sham feeding is the normal response. “Meal” indicates the actual ingestion of a hot dog. B, RSA of the EGG recorded while a healthy subject chewed a cold tofu dog and spit it out (sham feeding). The subject felt “disgusted” during the sham feeding effort. Note the lack in increase of 3 cpm peaks during sham feeding the tofu dog compared with (A). The subject then ingested a warm hot dog on a bun at “Meal”; note the subsequent increase in peaks at 3 cpm. (Modified from Stern RM, Crawford HE, Stewart WR, et al. Sham feeding. Cephalic-vagal influences on gastric myoelectric activity. Dig Dis Sci 1989; 34:521-7.)

and spits out the test meal rather than swallowing it, elicits the cephalic-vagal reflex. Sham feeding a warm hot dog on a bun elicits enhanced 3 cpm EGG activity, whereas sham feeding a cold tofu dog, a food that the subjects considered disgusting, resulted in blunted or no increase in the 3 cpm myoelectrical activity (Fig. 48-18).117 Thus, sensory and emotional attributes of food during the cephalic phase of ingestive behavior also affect the neuromuscular activity of the stomach.

DEVELOPMENTAL ASPECTS OF GASTRIC NEUROMUSCULAR FUNCTION Gastric peristalsis appears between 14 and 23 weeks of gestation. Grouped or clustered peristaltic waves are evident by 24 weeks.118 The neuroregulatory mechanisms respon­ sible for the coordination of antropyloroduodenal motility in gastric emptying are well developed by 30 weeks of gesta­ tion.119 EGG recordings show normal 3 cpm activity in preterm infants delivered at 35 weeks that are similar to EGG signals recorded in full-term infants.120,121 On the other hand EGG recordings from premature infants (less than 35 weeks’ gestation) showed considerable tachygastria.120 Gastric myoelectrical activity matures further over the first 6 to 24 months of life and achieves full adult values by the end of the first decade.121,122 The development of ICCs has been studied intensely because of the interest in gastric electrical rhythmicity, smooth muscle contractions, and gastric dysrhythmias. Labels for the tyrosine kinase receptor (c-Kit) and the avail­ ability of knock-out mice lacking c-Kit have led to increased understanding of the development of ICCs.123 The ICCs dem­ onstrate differential development, with c-Kit expression on ICCs in the MY-ICCs developing before birth, whereas ICCs in the deep muscular plexus (IM-ICCs) develop after birth.124 ENS and ICC networks are not fully developed and are poorly coupled at birth, but progressive maturity of gastric

rhythmicity and contractility occur during perinatal devel­ opment.120,121 The ENS and ICCs in the deep muscular plexus are closely related, whereas ICCs in the myenteric plexus can develop normally in the absence of the enteric nervous system.124 Loss of ICCs in the pylorus is associated with loss of the inhibitory neural activity that may contrib­ ute to the development of pyloric stenosis in infants (see Chapter 47).125

ASSESSMENT OF GASTRIC NEUROMUSCULAR FUNCTION GASTRIC EMPTYING RATES

Clinical tests are available to assess the neuromuscular functions of the stomach, including emptying, contractions, and electrical rhythm, and they are discussed following.

Scintigraphy

Test meals labeled with radioisotope are available to assess the rate of gastric emptying. The seminal solid-phase gastric emptying protocol was a multinational study that used a 255-kcal technetium-99m (99mTc)–labeled egg substitute with bread and jam as the standard meal.43 Scans were obtained for 1 minute immediately after ingestion of the meal and at 30 minutes, 60 minutes, 120 minutes, 180 minutes, and 240 minutes in 123 healthy individuals. Delayed gastric emptying was defined as greater than 60% retention of the meal at 120 minutes and greater than 10% retention of the meal in the stomach 240 minutes after ingestion (see Figs. 48-10 and 48-11). The four-hour empty­ ing test was superior to the two-hour test because almost 20% of patients with suspected gastroparesis had normal emptying at two hours, but abnormal emptying at four hours.126 Pitfalls in the scintigraphic method for solid-phase gastric emptying studies include improper binding of the isotope with the test meal, which results in rapid or normal empty­

803

804

Section VI  Stomach and Duodenum ing and continuance of medications that may stimulate (e.g., metoclopramide) or inhibit (e.g., narcotics, anticholin­ ergic agents) gastric smooth muscle contractions. These medications should be stopped five to seven days before all gastric neuromuscular tests, if possible. Radiation exposure for the subject occurs with the scintigraphic tests, and mul­ tiple tests in the same subject are not advisable. Liquidphase gastric emptying tests can be performed with indium 111-diethylenetriaminepentaacetic acid (111In-DTPA) 99mTclabeled water or other liquids. Patients with unexplained nausea symptoms may have altered emptying of liquid meals, even if solid phase emptying is normal.127,128

Capsule Technology

Gastric emptying time of test meals is obtained from a small capsule that measures intraluminal pH and contractions (see Fig. 48-12). The capsule is swallowed with a standard test meal. During the postprandial period, measurements of luminal pH and contractions are transmitted to a receiver worn by the subject. In healthy subjects the capsule is emptied from the stomach into the duodenum approxi­ mately five hours after ingestion of the egg substitute meal. Emptying of the capsule correlated with 90% emptying of the technetium-labeled egg substitute solid meal. The test had very good sensitivity and specificity in detecting gastroparesis.83

Breath Tests

Breath tests indirectly reflect gastric emptying of solid and liquid test meals. The solid meals are labeled with C13 and include C13 octanoic acid, C13 acetate, or C13 Spirulina pla­ tensis. The C13 octanoic acid breath test has been performed in many experimental protocols and is used widely in Europe for research and clinical studies.129,130 The C13labeled food is emptied from the stomach and absorbed in the small intestine. The labeled nutrients are metabolized in the liver to C13O2, excreted in the lungs, and detected in breath samples. Breath samples are collected at 45, 90, 120, 150, and 180 minutes after the meal in the C13 Spirulina test. C13 is a stable isotope with no radiation risks. The C13 breath tests are generally comparable to scintigraphy.131 Pitfalls include spurious results in patients with malabsorp­ tive conditions, liver diseases or lung diseases that may preclude normal oxidation and excretion of the C13-labeled foods.

Ultrasonography

Transabdominal ultrasonographic techniques are used to measure antral diameter and antropyloroduodenal func­ tion.132,133 Three-dimensional ultrasound methods show the intragastric distribution of the test meal and regional varia­ tions in gastric volume (see Fig. 48-13).58 The technique is ideal with a liquid meal, but solids can also be identified and gastric emptying rates can be determined. The clinical application is limited by the high level of expertise required by the ultrasound operators.

Computed Tomography and Magnetic Resonance Imaging

These two techniques have been used to measure gastric emptying and demonstrate intragastric distribution of test meals. Computed tomography (CT) and magnetic resonance imaging (MRI) technologies offer unique anatomic and func­ tional views of the stomach in the fasting and postprandial periods.134 Sequential antral contractions can be visualized. Because of expense and availability, these techniques are not used in clinical practice.

GASTRIC CONTRACTIONS Antroduodenal Manometry

Antroduodenal manometry is an invasive technique wherein a water-perfused multilumen catheter is placed either through the nose or the mouth and advanced to a position where the proximal catheter ports are in the distal antrum and the distal ports are in the duodenum.40 Placement of the catheter requires endoscopic or fluoroscopic aid. The recordings typically last for several hours in order to record phases 1, 2, and 3 of the MMC and several more hours to record postprandial contractions after the subject ingests a test meal (see Fig. 48-7A and B). Antroduodenal manometry testing is not only invasive but also time intensive and requires extensive assistant or physician time for perfor­ mance of the test and interpretation of the data. Intraluminal manometry catheters detect only lumen-occluding contrac­ tions.136 Intraluminal pressure transducer devices fail to record almost 50% of contractions in the corpus and antrum because the majority of postprandial peristaltic waves are not lumen-occluding contractions. Manometry catheters positioned in the duodenum can detect patterns of neuro­ pathic or myopathic dysfunction.

Capsule Technology

The ingestible capsule described previously measures contractions of the stomach wall. After a standard test meal, irregular contractions occur at one to three per minute and are consistent with manometric recordings from the antrum.137 Several minutes of sustained high-amplitude, antral contractions occur prior to the emptying of the capsule into the duodenum (see Fig. 48-12) and some pat­ terns are consistent with phase 3–like contractions recorded by antroduodenal manometry. In patients with gastropare­ sis, capsule studies showed a decreased motility index during the 20 to 30 minutes before emptying of the capsule, but a normal motility index in the 10 minutes before the capsule was emptied into the duodenum. These studies suggest that normal terminal antral contractions may be maintained even in patients with gastroparesis, whereas antral contractility required for trituration of digestible foodstuffs is abnormal.138 Determination of the unique gastric contraction patterns after ingestion of a variety of common foods is possible using the capsule motility device.

GASTRIC MYOELECTRICAL ACTIVITY

Electrogastrography (EGG) refers to the recording methods used to noninvasively measure gastric myoelectrical activity using electrodes positioned on the abdominal surface.139,140 The EGG signal summates the ongoing gastric myoelectrical activity. The EGG reflects the slow wave activity during fasting and the summation of slow wave activity linked to plateau and action potential activity during the postprandial period (see Figs. 48-1 and 48-2).140 In response to a water load or a nutrient load, the amplitude of the EGG signal increases in the normal 2.5 to 3.7 cpm range, as determined by visual and computer analysis (e.g., an increase in the percentage of EGG power or in the post­ prandial power ratio in the normal frequency range) (see Fig. 48-15).60,140,141 Pitfalls for recording and analyzing EGGs include failure to identify artifact in the signal and harmonics in the computer analyses.142 Changes in the EGG frequency and amplitude are key measures. After ingestion of most solid or liquid meals, a so-called frequency dip occurs in the first 10 to 15 minutes after the meal. The frequency dip reflects changes resulting from marked gastric relaxation and accommodation of the test meal related to the volume or the temperature of the

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders TACHYGASTRIA 500 µV

A

60 s BRADYGASTRIA (LOW-AMPLITUDE) 500 µV

60 s

B

BRADYGASTRIA (HIGH-AMPLITUDE)

500 µV

sures after meals are not useful. The barostat balloon was designed to measure changes in tone (or gastric relaxation) and volume in the more spherical areas of the proximal stomach.153 Barostat methods involve a large, collapsed thin-walled balloon that is mounted on a catheter and passed through the mouth into the stomach. Intraballoon pressure is maintained with infused air during the baseline or fasting period with the balloon slightly distended. A test meal is then ingested. As the fundus and proximal stomach relax in response to the meal, more air is concomitantly infused into the balloon to maintain the established baseline intraballoon pressure (see Fig. 48-9).154 The volume of air that is infused to maintain baseline pressure is measured and is an estimate of the increased gastric volume that occurs as the proximal stomach relaxes. Barostat studies demonstrate abnormalities in fundic relaxation in almost 30% of patients with functional dys­ pepsia.155 The failure of fundic relaxation correlates with early satiety. Failure of fundic relaxation has also been recorded in patients with gastroparesis of diverse causes.156,157 Because the barostat method is invasive and uncomfortable for patients, these studies have been limited to the research laboratory.

Scintigraphy and Other Tests

60 s

C

NORMAL 1000 µV

D

60 s

Figure 48-19.  Gastric dysrhythmias recorded with electrogastrogram (EGG) methods are shown. A, Tachygastria, an abnormally rapid signal at 6 cycles per minute (cpm), shown by dots. B and C, Bradygastria at low- or high-amplitude 1 cpm wave, respectively. The one-per-minute waves are indicated by the solid curved lines and the smaller waves in the EGG signal represent respiratory activity. D, Normal 3 cpm EGG signal is identified by dots.

meal.143,144 Several minutes after ingestion of the meal, the frequency of the EGG signal returns to the middle of the normal 2.5 to 3.7 cpm range (see Fig. 48-15). Gastric dysrhythmias are associated with symptoms of nausea in subjects with motion sickness,89 nausea and vom­ iting of pregnancy,145,146 functional dyspepsia,60,147 and gas­ troparesis.148-150 Gastric dysrhythmias include 0.5 to 2.5 cpm signals termed bradygastrias and 3.7 to 10 cpm signals termed tachygastrias (Fig. 48-19). Recordings that have com­ binations of tachygastria and bradygastrias are termed mixed or nonspecific gastric dysrhythmias.151 Gastric myo­ electrical activity can also be recorded from serosal elec­ trodes placed during surgery or with mucosal electrodes placed during endoscopy.6,74,152

GASTRIC RELAXATION, ACCOMMODATION, AND VOLUME Barostat Tests

Due to the spherical shape of the fundus and the proximal stomach, manometric catheters to record intraluminal pres­

Excessive or poor fundic relaxation in response to liquids and solids can be demonstrated with scintigraphy, ultra­ sound, and MRI. Single photon emission computed tomog­ raphy (SPECT) is a method that outlines the gastric wall before and after ingestion of a meal to determine changes in volume of the stomach. This method requires intravenous injection of 99mTc-pertechnetate to outline the gastric wall. The accommodation response can be identified with SPECT.158,159

Non-Nutrient Liquid and Nutrient Drink Tests

Non-nutrient liquids (the water-load test) and nutrient drink tests are used to assess overall gastric volume or gastric capacity and visceral sensations such as nausea, stomach fullness, or satiety in response to ingestion of these liquids,60,97,160 often in conjunction with measures of gastric myoelectrical activity, accommodation, or emptying. In the water-load test, water is consumed over a 5-minute period until the subject feels full. In the typical caloric drink test, subjects drink 150 mL of the liquid (e.g., Ensure) every 5 minutes until maximum tolerated satiety is achieved, an endpoint that requires almost 30 minutes and the con­ sumption of 800 to 1000 mL of the nutrient drink.161 In many healthy subjects, nausea and gastric dysrhythmia are evoked by the satiety drink test. Subjects with functional dyspepsia or gastroparesis ingest much smaller volumes of water or nutrient drink and report fullness, indicating impaired relaxation and accommodation of the stomach.60,97

HISTOPATHOLOGIC STUDIES IN GASTRIC NEUROMUSCULAR DISORDERS

Efforts to define the histopathologic basis of gastric neuro­ muscular disorders have provided basic knowledge for the evolving field of neurogastroenterology. Most of the fullthickness specimens from the gastric wall have been har­ vested during the placement of gastric electrical stimulation devices or the placement of jejunostomy feeding tubes in patients with severe gastroparesis. Mucosal biopsies do not typically contain smooth muscle, ICCs, or ENS neurons, but an endoscopic biopsy technique has been described that provides a full-thickness specimen that contains elements of circular muscle, ENS neurons, and ICCs.162 Histologic

805

806

Section VI  Stomach and Duodenum abnormalities in smooth muscle, neurons of the ENS, number or location of ICCs, and distribution of key neural or muscular receptors are possible underlying mechanisms of disordered gastric smooth muscle relaxation, peristaltic contraction, and gastric slow wave activity. Loss of ICCs (MY-ICCs) was reported in patients with type 1 and type 2 diabetic gastroparesis.25,163,164 Interestingly, similar loss of ICCs was shown in diabetic mice, and the ICCs were restored with intense insulin therapy, suggesting that the ICCs are not entirely destroyed in diabetes but dedifferentiate into immature myoblasts during prolonged hyperglycemia.164 Damage to ICCs, ENS, and smooth muscle can also be seen in inflammatory and neoplastic condi­ tions.165-168 Other studies in humans have shown fibrosis of smooth muscle layers but intact myenteric plexus and vagus nerve in patients with diabetic gastroparesis, loss of ICCs but minimal smooth muscle fibrosis, and inflammatory T lymphocyte infiltration in myenteric neurons.169-171 Pitfalls with neurohistologic studies include tissue sampling of variably affected regions of the stomach, with patchy distri­ bution abnormalities of ICCs, ENS neurons, or smooth muscle. Nevertheless, results from histochemical studies are needed to provide new directions for understanding the neuromuscular dysfunction of the stomach and to stimulate ideas for novel therapeutic approaches.

NEUROMUSCULAR DISORDERS OF THE STOMACH Gastric neuromuscular disorders encompass a continuum of electrical and contractile dysfunction.172 At one end of the spectrum are gastric dysrhythmias, which are subtle electri­ cal disturbances associated with mild to severe nausea symptoms (Fig. 48-20). Abnormalities in relaxation of the

Vagal afferent nerve Impaired fundic relaxation Abnormal fundic emptying

Pacemaker region Weak 3 cpm rhythm

Pylorospasm

Gastric dysrhythmias Tachygastria Bradygastria 60 s

Dilated gastric antrum Antral hypomotility Gastroparesis Figure 48-20.  Spectrum of gastric neuromuscular disorders. Gastric neuromuscular disorders range from abnormal fundic relaxation and emptying to gastric dysrhythmias and antral hypomotility and gastroparesis. Pyloric sphincter dysfunction, as well as duodenal dysfunction, antroduodenal dyscoordination, and vagal neurohypersensitivity, may all be present in some patients with gastric neuromuscular disorders. cpm, cycles per minute. See text for details. (Modified from Koch KL, Stern RM. Functional disorders of the stomach. Semin Gastrointest Dis 1996; 7:185-95.)

fundus are associated with early satiety. At the severe end of the spectrum, antral hypomotility and profound gastro­ paresis are associated with prolonged postprandial fullness, vomiting, bloating, weight loss, and malnutrition that may require enteral or parenteral nutritional support. Patients with gastroparesis may also have gastric dysrhythmias, a dilated antrum, poor fundic relaxation, and gastric hyper- or hyposensitivity due to vagal or sphlanchnic nerve dysfunc­ tion.172,173 Clinical tests currently approved by the U.S. Food and Drug Administration (FDA) to assess gastric neuromus­ cular function are the scintigraphy tests to measure the rate of gastric emptying, the capsule motility device to measure gastric emptying, gastric pH, and contraction patterns of the stomach, and electrogastrography devices to measure gastric myoelectrical activity before and after provocative test meals. These tests provide objective assess­ ments of different aspects of the neuromuscular activity of the stomach in health and disease. Results of gastric empty­ ing and gastric myoelectrical activity tests provide objective diagnoses of gastric dysrhythmias and gastroparesis.

GASTROPARESIS

Gastroparesis means “paralysis” of the stomach as defined by the delayed rate of emptying of a standard test meal. Approximately 90% of the patients with gastroparesis have either diabetic, postsurgical, or idiopathic gastroparesis, but the less common forms of obstructive and ischemic gastro­ paresis are reversible. Major categories of gastroparesis are described below.

Diabetic Gastroparesis

Fifty percent of patients with long-standing type 1 diabetes mellitus develop gastroparesis. These patients often have diabetes for more than 10 years, erratic and elevated glucose levels, peripheral neuropathy, nephropathy, and cardiovascular disease.174-176 In a minority of patients the gastroparesis is the initial complication of their diabetes. Upper GI symptoms, however, may be minimal and non­ specific. This is not surprising because patients with diabe­ tes often have ischemic heart disease with minimal chest discomfort or cholecystitis with minimal or no right upper quadrant pain. One important manifestation of gastric emptying dysfunc­ tion in patients with diabetes is erratic glucose control, especially with unexpected hypoglycemic episodes in the postprandial period. Because the patient may have no per­ ception of delayed emptying of food, the usual insulin doses are administered before meals. Insulin levels increase in the blood, but if gastric emptying is delayed, then nutrient delivery into the duodenum is delayed. Thus, plasma glucose levels decrease in response to the insulin treatment and symptomatic hypoglycemia develops unex­ pectedly in the postprandial period because of delayed emptying. Acute hyperglycemia (>220 mg/dL) is associated with tachygastrias and delayed gastric emptying in healthy sub­ jects and in patients with diabetes.73,175,177-179 Hyperglycemia is also associated with loss of ICCs,25,163 antral hypomoti­l­ ity,180 isolated pyloric contractions,181 gastric dysrhyth­ mias,6,177,178 and impaired prokinetic action of prokinetic drugs like erythromycin.182 Relatively minor increases in the blood sugar level, even elevations within the physiologic range, delay gastric emptying in normal volunteers and diabetic patients.179 Acute hyperglycemia produced by glucose clamp methodology elicits fullness, decreased antral contractility, blunts the contractile pyloric response to intraduodenal lipid infusion, and modifies upper GI sensations.183,184

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders RUNNING SPECTRAL ANALYSIS Power

EGG rhythm strips

60 s 1 mV

60 s

3

6

9

Frequency (cpm)

Power

Time (1 min between lines)

Time (1 min between lines)

1 mV

0

A

RUNNING SPECTRAL ANALYSIS

12

15

EGG rhythm strips B

B1

60 s

500 µV

0

B

3

6

Waterload (250 mL)

A

A1

9

60 s

12

15

Frequency (cpm)

Figure 48-21.  Obstructive and idiopathic gastroparesis. A, Running spectral analysis (RSA) of the electrogastrogram (EGG) recording from a patient with gastroparesis due to mechanical obstruction at the pylorus secondary to chronic peptic ulcer disease. Note the persistent high-amplitude 3 cycles per minute (cpm) waves in the EGG rhythm strips and the uniform and unvarying peaks at 3 cpm in the RSA, findings that would not be expected in a patient with gastroparesis due to electrical and contractile dysfunction. B, RSA and EGG rhythm strips in a patient with idiopathic gastroparesis. Note that the EGG rhythm strips show a 7 to 8 cpm tachygastria before (B) and after (A) the water-load test. The RSA shows multiple peaks in the 7 to 8 tachygastria range and few peaks in the normal 3 cpm range. This patient had electrical and contractile abnormalities of the stomach as documented by the tachygastria and gastroparesis. (Modified from Brzana RJ, Koch KL, Bingaman S. Gastric myoelectrical activity in patients with gastric outlet obstruction and idiopathic gastroparesis. Am J Gastroenterol 1998; 93:1803-9.)

Gastric neuromuscular abnormalities documented in patients with chronic type 1 diabetes include abnormal intragastric distribution of food,185 reduced receptive relaxation and accommodation,186 reduced incidence of the antral component of the MMC, antral dilation, postprandial antral hypomotility,176 and electrical dysrhythmias.148,177,187 Loss of the antral phase 3 contractions results in poor emp­ tying of fibrous debris in the stomach and is the neuromus­ cular basis for the formation of bezoars (see Chapter 25). Thus, the progressive neuromuscular dysfunction of the diabetic stomach reflects the effects of chronic hyperglyce­ mia and superimposed, intermittent hyperglycemia epi­ sodes that occur repeatedly over many years. Gastric smooth muscle dysfunction is another mechanism of delayed gastric emptying in some patients with diabetes. In rats with diabetes gastric smooth muscle contractility is reduced in response to electrical stimulation.188 The inhi­ bitory effect of hyperglycemia on stem cell factor has a role in the reduction in ICCs and smooth muscle contractility.189 Pylorospasm as a cause of gastroparesis was documented in one study.190 Some patients with pylorospasm have a form of obstructive gastroparesis; that is, the neuromuscular function of the corpus and antrum is intact, but persistent functional obstruction at the pylorus (e.g., pylorospasm) causes delayed emptying. These patients have normal or increased 3 cpm myoelectrical activity, a discordant finding in patients with gastroparesis that also suggests the possibil­ ity of gastric outlet obstruction (Fig. 48-21A).149 In type 2 diabetes mellitus the incidence of gastroparesis ranges from 30% to almost 50%.191 Patients with type 2 diabetes differ significantly from patients with type 1 dia­ betes in that they have insulin resistance, the diabetes appears later in life, and the hyperglycemia has been present longer before diagnosis compared with patients with type 1 diabetes.193 Almost 20 million people in the United States have type 2 diabetes mellitus, and many have unsuspected gastroparesis.192 Gastroparesis in patients with type 2 diabe­ tes presents with subtle dyspepsia-like symptoms, but a minority of patients may present acutely with nausea, vom­

iting, and severe gastroparesis. Gastric dysrhythmias have been recorded in up to 75% of patients with type 2 diabetes with an average hemoglobin A1C of 8.2.193 In the early stages of type 2 diabetes mellitus gastric emptying may be accelerated.194 Gastroparesis has been also described in experimental diabetes, specifically in nonobese diabetic mice and in the type 2–like diabetes db/db mouse. Studies by Watkins and associates showed that neuronal nitric oxide synthase (nNOS) was reduced in the diabetic mouse stomach and was associated with antropylorospasm and gastroparesis.195 Treatment with insulin or sildenafil restored NOS and was associated with reversal of the delay in gastric emptying. Hypomotility of the fundus and hypercontractility of the pylorus were found in db/db mice.196 In other studies of diabetic mice the loss of ICCs was associated with electrical dysrhythmias, delayed gastric emptying, and reduced neu­ rotransmission in gastric smooth muscle.197 Chronic hyper­ glycemia also leads to glycosolation metabolic end products that interfere with neural function and smooth muscle contraction.

Postsurgical Gastroparesis

Gastroparesis occurs in a subset of patients undergoing subtle or radical stomach operations that range from vagot­ omy to fundoplication to antrectomy. Truncal vagotomy produces complex effects on the neuromuscular function of the stomach. After vagotomy, the fundus fails to relax nor­ mally after meals, resulting in rapid filling of the antrum.198 Vagotomy is also associated with gastric dysrhythmias,3 decreased antral contractions, and poor antropyloroduode­ nal coordination.199 Truncal vagotomy performed during ulcer operations required a pyloroplasty to reduce sphinc­ teric resistance to outflow because antral contractions were weak and peristalsis was disrupted.200 Most patients recover from the effects of vagotomy, but in patients undergoing extensive resection of the antrum and corpus, prolonged symptoms and chronic gastric neuromuscular dysfunction are likely.

807

808

Section VI  Stomach and Duodenum During fundectomy and lower esophageal resection for esophageal cancer, the vagus nerves are transected and the fundic reservoir is lost. Although pyloroplasty is performed to facilitate gastric emptying, the loss of the fundus and variable amounts of the corpus (pacemaker region) often leads to chronic nausea, gastric dysrhythmias, and gastro­ paresis. Antral resections with Billroth I, Billroth II, or Roux-en-Y gastrojejunostomy are performed to treat gastric tumors or peptic ulcer disease, but these operations may lead to profound neuromuscular dysfunction of the stomach.201,202 Critical amounts of the corpus-antrum (including unknown amounts of the gastric wall containing the pacemaker region) required for normal gastric neuro­ muscular activity may be resected and the efficacy of tritura­ tion by the corpus and antrum may be markedly reduced. Ingested food is retained in the remnant fundus and fails to empty into the corpus203; the corpus fails to mix and empty gastric contents even though the anastomosis is widely patent. Liquids empty poorly into the duodenal or jejunal segments if no peristaltic contractions are present. The Roux-en-Y gastroenterostomy operation may result in the Roux syndrome in which postprandial pain, bloating, and nausea develop. Delayed gastric emptying is due to “func­ tional obstruction” by the Roux limb as the neuromuscular dyssynchrony within the Roux limb prevents emptying of the stomach.204,205 After vagotomy and antrectomy, a minor­ ity of patients develop the dumping syndrome described below. Fundoplication is commonly performed to treat gastro­ esophageal reflux disease that fails to respond to medical therapy (see Chapter 43). Postfundoplication gastroparesis and early satiety, bloating, prolonged fullness, and nausea occur in a minority of patients.206 These patients have altered fundic relaxation, delayed gastric emptying, and gastric dysrhythmias, possibly on the basis of vagal nerve injury during or after the fundoplication procedure.207-209 Because gastric emptying studies are infrequently per­ formed before this operation, it is not known how many of the patients already had gastroparesis before the fundoplica­ tion. In some patients the fundoplication results in rapid filing of the antrum (due to poor relaxation of the fundus) and the rate of gastric emptying is increased.210

Ischemic Gastroparesis

Chronic mesenteric ischemia may cause ischemic gastropa­ resis.211 These patients present with chronic symptoms of gastroparesis. Ischemic gastroparesis is distinct from acute mesenteric ischemia, which presents as an abdominal catas­ trophe with an acute abdomen and gangrenous small intes­ tine (see Chapter 114). Chronic mesenteric ischemia is usually due to progressive atherosclerosis or hyperplasia of the intima of the arteries of the celiac, superior mesenteric, or inferior mesenteric artery. Collaterals of these obstructed arteries form over time so that neuromuscular function of the stomach is preserved, at least for some time. Bypass graft surgery or dilatation of the stenotic arteries results in resolu­ tion of symptoms, eradication of gastric dysrhythmias, and reversal of gastroparesis.211 Thus, ischemic gastroparesis is a reversible form of gastroparesis and should be suspected in patients with gastroparesis, weight loss, and a history of peripheral vascular disease, cerebral vascular disease, or myocardial infarction. An abdominal bruit is present in approximately 50% of patients. There are other uncommon forms of mesenteric vascular compromise (e.g., the median arcuate ligament syndrome) that may result in decreased blood flow to the stomach.212 Release of the arcuate ligament and restoration of blood flow has been associated with improvement in gastric emptying.

On the other hand, superior mesenteric artery syndrome is not accepted as a cause of mechanical obstruction that leads to gastroparesis, nausea, and vomiting.

Obstructive Gastroparesis

Obstructive gastroparesis refers to mechanical obstruction at the pylorus, duodenum, or postduodenal area by tumor, chronic peptic ulcer or inflammation, rings, or webs.149 These patients have documented gastroparesis, but the EGG signal is normal and the amplitude of the 3 cpm EGG wave after the water-load test is increased.149 Figure 48-21A shows an example of high-amplitude 3 cpm EGG waves in a patient with pyloric outlet obstruction due to chronic peptic ulcer disease and fixed narrowing of the pylorus. In these patients the smooth muscle, ENS, and ICCs of the antrum are intact, but the recurrent gastric peristaltic waves meet sustained resistance at the point of obstruction (the pylorus) and the emptying of solid food is delayed. In contrast, patients with idiopathic gastroparesis have bradygastria, tachygastria, or mixed gastric dysrhythmias indicating electrical and con­ tractile dysfunction (see Fig. 48-21B). Surgical correction of the obstruction with Billroth I or Billroth II gastrojejunos­ tomy is necessary to correct obstructive gastroparesis, although some patients may respond to balloon dilation of strictures (see Chapter 53). In patients who have had prolonged obstruction, the stomach may dilate, smooth muscle contractions are weak, and gastric dysrhythmias are present. A more subtle type of gastric outlet obstruction occurs in pylorospasm. The sustained pyloric “spasm” or tone may cause right upper quadrant abdominal pain and prevents normal gastric peristaltic waves from empting chyme into the duodenum. Thus, the rate of emptying is delayed and gastroparesis is present. These patients also have normal 3 cpm EGG rhythm because the neuromuscular apparatus of the corpus-antrum is intact. Dilatation of the pylorus with a 20-mm balloon for two minutes decreased postprandial symptoms and improved the rate of gastric emptying.213 Finally, some patients have normal 3 cpm EGG signals and poor gastric emptying on the basis of “electromechanical dissociation.” In these situations it is possible that MY-ICCs that generate slow waves may be normal, but IM-ICCs and/ or smooth muscle dysfunction is present.

Idiopathic Gastroparesis

Almost one third of patients with gastroparesis have idio­ pathic gastroparesis.214 In many of these patients an acute febrile illness preceded the diagnosis of gastroparesis by many months.215,216 These “herald” illnesses are frequently described as flu-like with fever and nausea and vomiting. Patients often date the onset of their nausea from that point. Norwalk virus, herpes simplex virus, and EpsteinBarr virus infections have been documented in patients with sudden onset gastroparesis and normal immune systems, whereas other patients are immunocompro­ mised.217 Postviral gastroparesis may resolve completely over one to two years. It is unclear if the offending virus affects ICCs, smooth muscle or the ENS of the stomach, or vagal or splanchnic nerves, but these cases are analogous to postviral cardiomypathy or postviral neuropathy. Degen­ eration and loss of ICCs has been reported in patients with severe idiopathic gastroparesis.165,218 Loss of ICCs in knock­ out mice is associated with gastric dysrhythmias.219 Gastric dysrhythmias are common in patients with idiopathic gastroparesis (see Fig. 48-21B).149 Other historical clues to the genesis of idiopathic gastroparesis are exposures to (1) multiple courses of antibiotics (for example, treatments for chronic ear or sinus infections), (2) anesthetic agents for a

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders variety of common operations, and (3) food poisoning–like illnesses of unknown cause. Idiopathic gastroparesis is diagnosed in patients who have delayed gastric emptying and gastric dysrhythmias, but no primary causes of gastroparesis such as diabetes, ischemia, or gastric surgery. Many other diseases and dis­ orders, including intestinal pseudo-obstruction (discussed later), autonomic nervous system abnormalities, thyroid disease, or CNS diseases, may cause or are associated with gastroparesis. If these disorders are identified, then the gas­ troparesis is considered secondary to these diseases.

GASTRIC NEUROMUSCULAR DYSFUNCTION ASSOCIATED WITH OTHER GASTROINTESTINAL DISORDERS Functional Dyspepsia

Functional dyspepsia (FD) symptoms include epigastric dis­ comfort, early satiety, fullness, nausea, and vomiting in the setting of normal upper endoscopy and routine laboratory tests. FD is divided further into an epigastric pain syndrome and postprandial distress syndrome, the latter comprising 80% of the FD patients.220 One of the dominant symptoms that patients with FD have is recurrent and unexplained nausea. This is a debilitating symptom with a large differ­ ential diagnosis (Table 48-2) and these disease categories should be considered in the evaluation of the patient. If the patient has one of these disorders, then FD is not the diagnosis. The pathophysiology of the FD symptoms remains an area of intensive investigation (see Chapter 13). The FD symp­ toms may be caused by one of several neuromuscular dis­ orders of the stomach as shown in Figure 48-20. Gastroparesis is found in 17% to 40% and gastric dysrhythmias in 40% to 55% of patients with FD symptoms.60,147 If secondary causes for gastroparesis have been excluded, then these patients have idiopathic gastroparesis and/or gastric dys­ rhythmias, not FD.

Table 48-2  Differential Diagnosis of Chronic Nausea and Vomiting Mechanical gastrointestinal tract obstruction Pylorus, small intestine, colon Mucosal inflammation Peritoneal irritation Carcinomas Gastric, ovarian, renal, bronchogenic Metabolic/endocrine disorders Diabetic mellitus, hypothyroidism, hyperthyroidism, adrenal insufficiency, uremia Medications Anticholinergics, narcotics, l-dopa, progesterone, calcium channel blockers, digitalis, nonsteroidal anti-inflammatory agents, antidysrhythmia agents, lubiprostone Gastroparesis Obstructive Diabetic Ischemic Postsurgical Miscellaneous, including pseudo-obstruction Idiopathic Gastric dysrhythmias Tachygastria, bradygastria, mixed Central nervous system disorders Tumors, migraine, seizures, stroke Psychogenic disorders Anorexia nervosa, bulimia nervosa

The rate of gastric emptying is normal in the majority of patients with FD, but gastric dysrhythmias may be present in these patients. In one series of patients with FD, 35% had gastric dysrhythmias and normal gastric emptying.63 Improvement in gastric dysrhythmia and FD symptoms was reported in patients treated with cisapride,221,222 and similar results were reported in children and adults with dys­ pepsia.223,224 Thus, gastric dysrhythmias themselves are the cause of some of the FD symptoms and are a therapeutic target. Gastric neuromuscular disorders such as abnormalities in gastric accommodation or gastric hypersensitivity to disten­ tion may account for FD symptoms in other patients.225-230 Thus, a variety of neuromuscular dysfunctions in different regions of the stomach may be present in patients with FD symptoms.231 Exposure of the duodenal mucosa to acid is also a proposed mechanism of FD symptoms.232 Increasing data that link certain dyspepsia symptoms such as nausea or bloating or discomfort to distinct neuromuscular abnor­ malities of the stomach will aid in improved diagnosis and treatment of these patients.

Gastroesophageal Reflux Disease

Gastroesophageal reflux disease (GERD) typically presents with heartburn and regurgitation, but many patients have additional symptoms of early satiety, postprandial fullness, and nausea (see Chapter 43). Approximately 30% of patients with GERD also have FD and represent an “overlap syn­ drome.”233 Furthermore, 20% to 30% of patients with GERD have delayed gastric emptying.234,235 Twenty-five percent of patients with GERD and dyspepsia had delayed gastric emp­ tying and almost 70% had gastric dysrhythmias.236 The FD-like symptoms do not resolve when the GERD symptoms are treated with proton pump inhibitors (PPIs) because the gastric neuromuscular disorder is not treated by PPIs. Lower esophageal sphincter abnormalities, including transient lower esophageal sphincter relaxations, may con­ tribute to inefficient gastric emptying and gastric dysrhyth­ mias reported in these patients.237 In response to a meal, the fundus relaxes more in GERD patients compared with controls.238 GERD patients also tended to retain solids and liquids in the proximal stomach compared with control subjects,239 an abnormality that may stimulate additional transient lower esophageal sphincter relaxations. Radiofrequency ablation (RFA) treatment at the lower esophageal sphincter region in patients with heartburn and dyspepsia improved heartburn, gastric emptying rates, and gastric dysrhythmias.240,241 These data suggest in some patients with GERD and gastroparesis, treatments of GERD by RFA or fundoplication improves the gastric electrical rhythms and the rate of gastric emptying.

Constipation, Irritable Bowel Syndrome, and Pseudo-Obstruction

Gastroparesis and gastric dysrhythmias after the water-load test have been reported in 20% to 30% and 60% of patients with constipation-predominant irritable bowel syndrome, respectively.242,243 These patients represent another overlap syndrome of gastrointestinal neuromuscular disorders (see Chapter 118). GERD, gastroparesis/gastric dysrhythmias, and irritable bowel syndrome are all present in some patients and indicate a disorder of diffuse gastrointestinal neuromuscular dysfunction.233 Patients with intestinal pseudo-obstruction syndromes often have GERD, gastroparesis, small bowel dysmotility, and colonic inertia and reflect the severest form of general­ ized GI neuromuscular disorders.244,245 Pseudo-obstruction

809

810

Section VI  Stomach and Duodenum secondary to scleroderma commonly involves the esopha­ gus and stomach but may also cause small bowel dysmo­ tility and subsequent bacterial overgrowth.246 Intestinal pseudo-obstruction may be due to idiopathic degenerative or inflammatory processes involving the smooth muscle or enteric nervous system, as discussed in Chapter 120. A variety of neurologic diseases may also involve the stomach or other regions of the digestive tract and cause pseudo-obstruction–like symptoms. These neurologic disorders include spinal cord injuries, head injuries, amyo­ trophic lateral sclerosis, myasthenia gravis, a variety of muscular dystrophies, and Parkinson’s disease. Chapter 35 discusses stomach neuromuscular dysfunction and auto­ nomic neuropathies.

bowel transit and poor absorption of the ingested nutrients lead to an osmotic form of diarrhea.260 Patients with no history of gastric operations may also have rapid gastric emptying (dumping syndrome),259 includ­ ing patients with FD.260 Rapid gastric emptying is present when more than 30% of the meal (a low-fat egg substitute meal) is emptied in 30 minutes, or more than 70% is emptied at 60 minutes.128 Idiopathic abnormally rapid emptying is diagnosed in patients with no history of gastric operations or other causes. Rapid gastric emptying is also associated with early stages of type 2 diabetes mellitus (see earlier), Zollinger-Ellison syndrome, and the variety of gastric sur­ geries described earlier. Details of these entities is reviewed under the postsurgical syndromes in Chapter 53.

Miscellaneous Conditions

Patients with FD symptoms and cirrhosis with portal hyper­ tension,247 chronic kidney disease,248 chronic pancreatitis,249 or advanced human immunodeficiency virus (HIV) infec­ tions250 may have underlying gastroparesis. Patients with the Rett syndrome, which includes lack of development, autistic behavior, ataxia, and dementia in young girls, fre­ quently have failure to thrive and significant gastroparesis and esophageal contraction abnormalities.251 Patients with a variety of cancers may have local and systemic effects on the stomach neuromuscular apparatus that results in gastro­ paresis.252,253 The neoplastic neuropathic syndromes and the effects of cytotoxic chemotherapy and radiation may result in gastroparesis and affect the patient’s nutrition and intra­ vascular volume status. H. pylori infection does not affect the rate of gastric emp­ tying. On the other hand, abnormal gastric myoelectrical activity has been reported in patients with H. pylori infec­ tion, and the dysrhythmias disappeared after eradication of H. pylori.254 Gastric emptying is delayed in patients with anorexia nervosa,255 but is normal in patients with bulimia.256 Cyclic vomiting syndrome (CVS) is an unusual entity in that days of profound and unremitting nausea and vomiting (that requires hospitalization) are followed by many days or months with virtually no GI symptoms.257 CVS occurs in adults as well as in infants.257,258 When patients are well, EGG abnormalities are present and some of these patients have gastroparesis.

DUMPING SYNDROME AND RAPID GASTRIC EMPTYING

Dumping syndrome occurs in some patients who have had vagotomy and pyloroplasty or Billroth I or Billroth II gas­ trojejunostomy.259 In these patients the ingested foods are not accommodated and retained in the fundus (because of poor fundic relaxation) and the foods are not normally tritu­ rated because the corpus and antrum have been resected. Thus, liquid and solid nutrients are rapidly emptied or “dumped” into the duodenum or jejunum. Symptoms of dumping syndrome include nonspecific abdominal discom­ fort, bloating, and nausea that may precipitate vomiting. These symptoms are usually experienced in the first hour after ingestion of foods and can mimic symptoms of gastro­ paresis. Sweating and lightheadedness, however, may occur and be followed by abdominal cramps and diarrhea that occur two to four hours after the meal and are clues to the dumping syndrome. Early symptoms are due to the disten­ tion of the small bowel, whereas symptoms that occur later are due to rapid absorption of carbohydrates and hypergly­ cemia that is poorly matched in time with secretion of insulin. This mismatch of plasma glucose and insulin results in symptomatic hypoglycemia. The rapid small

DIAGNOSIS SYMPTOMS

Most affected patients have few upper GI symptoms when they are fasting. However, the ingestion of meals stimulates the disordered gastric neuromuscular apparatus, and early satiety, prolonged epigastric fullness, nonspecific epigas­ tric discomfort, mild to severe nausea, and vomiting are experienced.233 These are also the symptoms associated with functional dyspepsia (Chapter 13). Vomitus that con­ tains undigested, chewed food is strong evidence for gas­ troparesis. Prolonged postprandial fullness, weight loss, and female gender are predictive factors for gastropare­ sis.261 By adjusting their diet, patients learn to reduce these postprandial symptoms and carry on for months or years before they seek medical attention or their physicians rec­ ognize the possibility of gastroparesis. The Gastroparesis Cardinal Symptom Index is a useful validated question­ naire to quantify gastroparesis symptoms.262 Persistent nausea is one of the most noxious symptoms of the gastric neuromuscular disorders. A thorough review of the causes of nausea and vomiting is required (Chapter 14) and an appropriate differential diagnosis should be considered (see Table 48-2).263 In the evaluation of unexplained nausea and vomiting, gastric neuromuscular dysfunction must be distinguished from esophageal diseases and rumination syndrome. Occult GERD may present as unexplained nausea because these patients report little or no heartburn.264 Regurgitation is the gentle delivery of gastric content into the esophagus and pharynx (and the content is sometimes reswallowed), whereas vomiting is the forceful ejection of gastric contents from the mouth. Rumination refers to the effortless return of ingested liquids and solid foods into the mouth without burning, bitter taste, or nausea. Patients with rumination have impaired gastric accommodation and a more sensitive relaxation of the lower esophageal sphincter pressure in response to gastric distention.265-267 Rumination occurs in healthy adolescents and young adults, but was previously recognized among children with neural and developmental disorders. Abdominal pain, in contrast to the abdominal discomfort of bloating and nausea, occurs infrequently in patients with gastric neuromuscular disorders. Any recurrent abdominal pain syndrome should be worked up extensively because the nausea and vomiting may be secondary to the specific cause of the pain (e.g., burning epigastric pain associated with nausea and vomiting may be due to active peptic ulcer disease). Once the peptic ulcer disease is diagnosed and treated, the pain and nausea disappear. A specific pain syndrome may suggest diagnoses such as cholecystitis,

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders pancreatitis, or sphincter of Oddi dysfunction. On the other hand, the epigastric discomfort or pain in some patients may originate from excessive muscle tone of the fundus, high-amplitude antral contractions, pylorospasm, or hyper­ sensitivity of the stomach.268 Recurrent retching and vomit­ ing may result in abdominal pain, which is a result of abdominal muscle and rib tenderness or the abdominal wall syndrome.269,270

PHYSICAL EXAMINATION

The general examination may be normal or reveal signs of volume depletion, weight loss, and poor nutrition. Inspec­ tion of dentition may show erosion of enamel associated with chronic GERD or bulimia. Abdominal examination may detect masses, organomegaly, and areas of tenderness. Abdominal distention and succession splash may be present. Auscultation over the epigastrium may detect bruits that indicate stenoses of the celiac or superior mesenteric arter­ ies. Tenderness that is well localized at healed incisions and persists when the anterior abdominal muscles are con­ tracted (Carnett’s sign) suggests an abdominal wall syn­ drome.269,270 Neurologic examination may reveal nystagmus, facial weakness, ataxia, or other abnormalities.

STANDARD TESTS

Common causes of nausea and vomiting and dyspepsia symptoms are excluded by normal upper gastrointestinal (GI) series, CT of the abdomen and head, routine laboratory studies, and upper endoscopy.271 In patients with dyspep­ sia, gastric cancers are detected in less than 1% of the patients undergoing upper endoscopy.272 Most patients seen by gastroenterologists for dyspepsia or chronic nausea have already received acid-suppression therapy with PPIs, but symptoms persist. In such patients, gastric neuromuscular disorders should be considered.

SPECIALIZED NONINVASIVE TESTS

An objective diagnosis of neuromuscular disorders of the stomach can be established by the results of gastric empty­ ing tests and EGG. Gastric emptying tests and EGG are

complementary in defining different aspects of gastric neuromuscular disorders (Table 48-3).147,271 By combining the results of gastric emptying tests and EGG, four patho­ physiologic categories of gastric neuromuscular function are defined and rational treatment approaches can be designed. The categories are (1) gastroparesis with gastric dysrhyth­ mia; (2) normal gastric emptying with gastric dysrhythmia; (3) normal gastric emptying with normal gastric electrical rhythm; and (4) gastroparesis with normal gastric electrical rhythm. The EGG testing device is cleared by the FDA, but the combination of EGG and gastric emptying testing is infrequently performed. Nevertheless, the four categories provide a conceptual framework for understanding the spectrum of gastric neuromuscular disorders and providing an approach to therapy (see Table 48-3). Category 1 patients have severe neuromuscular dysfunc­ tion with gastroparesis and gastric dysrhythmias such as tachygastria. Among patients with “functional dyspepsia,” 17% to 32% have gastric dysrhythmias and gastropare­ sis.60,147,271 Patients in category 1 often have more severe symptoms, require many drugs, and may require venting gastrostomies, enteral feeding, and gastric pacing, as dis­ cussed following. Among patients with functional dyspepsia, 40% to 60% have gastric dysrhythmias despite normal gastric emptying. Such category 2 patients had a significantly better response to the prokinetic agent cisapride than patients with normal EGG recordings.221 In patients with normal gastric myoelectrical activity and normal gastric emptying (category 3), the nausea symptoms are likely due to visceral hypersensitivity or nongastric causes. Nongastric diagnoses should be considered in this patient group. Atypical GERD may cause nausea, and a 24-hour pH study will confirm the relationship.264 A CCKstimulated gallbladder emptying study may document gall­ bladder dysfunction in the absence of cholelithiasis. If postprandial abdominal discomfort and disturbed bowel function components are present, then irritable bowel syn­ drome should be considered. Central nervous system causes of nausea also should be assessed.

Table 48-3  Categories of Gastric Neuromuscular Disorders and Treatment Approaches Based on Gastric Electrical and Emptying Test Results CATEGORY 1

CATEGORY 2

CATEGORY 3

CATEGORY 4

Test Results Gastric dysrhythmia and gastroparesis Diagnosis Severe gastric myoelectrical contractile disorder Treatment Nausea/vomiting diet† Prokinetic therapy Anti-nauseant therapy G tube/J tube Total parenteral nutrition Acustimulation Endoscopic therapies* Gastric electrical stimulation Gastric pacemaker

Test Results Gastric dysrhythmia and normal emptying Diagnosis Gastric myoelectrical disorder

Test Results Normal gastric rhythm and emptying Diagnosis Visceral hypersensitivity Nongastric causes

Test Results Normal gastric rhythm and gastroparesis

Treatment Nausea/vomiting diet† Prokinetic therapy Anti-nauseant therapy

Treatment Nausea/vomiting diet† Antidepressant therapy Drugs for fundic/antrum relaxation (Further workup for nongastric causes)

Treatment Surgery for obstruction Nausea/vomiting diet† Prokinetic therapy Anti-nauseant therapy

Diagnosis Mechanical obstruction Electro-contractile dissociation

*See Table 48-4. † See Table 48-5. G tube, gastrostomy tube; J tube, jejunostomy tube. Modified from Koch KL. Nausea and Vomiting. In: Wolfe MM, editor. Therapy of Digestive Disorders. 2nd ed. Philadelphia, Pa: Elsevier; 2006. pp 1003-17.

811

812

Section VI  Stomach and Duodenum Patients who have gastroparesis and normal or highamplitude 3 cpm EGG signals (category 4) may have mechanical obstructions of the stomach and duodenum that are reversible with operation.149 Alternatively, patients with a normal-amplitude 3 cpm pattern and gastroparesis may have a form of “electromechanical dissociation,” requiring medical therapy.

TREATMENT By grouping patients on the basis of the gastric emptying and myoelectrical results, the pathophysiologic findings can help in the understanding of symptoms and assist in devel­ oping an approach to treatment (see Table 48-3). If delayed gastric emptying is confirmed, then the major causes of gastroparesis should be reviewed to specifically define the cause of the gastroparesis (see Table 48-2). The reversible causes of gastroparesis, gastric outlet obstruction, and chronic mesenteric ischemia, should be excluded. If gastric emptying is normal, then gastric dysrhythmias and gastric accommodation disorders may be the neuromuscular disor­ ders that are relevant to the symptoms. Treatment of gastric neuromuscular disorders is problematic. Scant specific ther­ apies are available to address specific pathophysiologic dis­ orders. Treatments listed in Table 48-4 reflect the broad but limited armamentarium that ranges from prokinetic agents to diet counseling and electrical therapies.

DRUG THERAPY Prokinetic Agents for Corpus-Antrum

Drugs with prokinetic effects on gastric contractility and gastric dysrhythmias are usually prescribed for patients in categories 1 and 2 (see Table 48-4). Patients who have gastroparesis and tachygastria have severe electrical and contractile abnormalities of the stomach. The treatment includes prokinetic, antinauseant therapies, and dietary counseling.271 Erythromycin is a macrolide antibiotic and motilin-like molecule that increases gastric emptying by stimulating strong phase 3–like antral contractions.273,274 Erythromycin, however, often increases nausea and vomiting symptoms. Metoclopramide, a substituted benzamide related to pro­ cainamide, is a useful prokinetic antiemetic. However, metoclopramide also causes depression, extrapyramidal side effects, and irreversible tardive dyskinesia.275,276 Dom­ peridone is a dopamine antagonist that decreases nausea, corrects gastric dysrhythmias, and increases gastric empty­ ing rates.277,278 Domperidone may be obtained through a new drug application process with the FDA. Cisapride and tega­ serod are 5-HT4 agonists and were not approved for gastric emptying disorders, but these drugs increase gastric empty­ ing rates and decrease dyspepsia symptoms in some patients.279-281 Cisapride and tegaserod were withdrawn from the market but are available through special FDA programs.

Prorelaxant Agents for Fundus and Pylorus

Drugs that relax the fundus are few. Sumitriptan, a 5-HT1 antagonist, decreases fundic tone but was not better than placebo in reducing symptoms in patients with FD.282 Trials of dicyclomine or calcium channel blockers to decrease fundic tone have not been reported. Botulinum toxin relaxes the pyloric sphincter pressure and is described below.

Antinauseant Therapy

Nonspecific approaches to treating nausea and vomiting from gastric neuromuscular disorders include 5-HT3 sero­ tonin antagonists such as ondansetron and granisetron (see Table 48-4). These agents, as well as the phenothiazines and antihistamines such as promethazine, dimenhydrinate, and cyclizine, are often used for these symptoms, but there are no controlled trials in patients with gastric neuromuscular disorders. Lorazepam or alprazolam or other antianxiety medications reduce nausea in some patients.283 An uncon­ trolled trial of tricyclic antidepressants such as amitripty­ line alleviated nausea in approximately 70% of patients with unexplained nausea.284

ELECTRICAL THERAPY Acustimulation

Acustimulation (mild electrical stimulation of acupuncture points) reduces nausea of pregnancy, nausea due to chemo­ therapy agents, postoperative nausea, and the nausea of motion sickness.285,286 These treatment methods have not been systematically studied in patients with gastric neuro­ muscular disorders.

Gastric Electrical Therapies

Three different methods are being investigated to treat gastroparesis: (1) gastric electrical stimulation (GES) with high-frequency (e.g., 12 cpm) and short duration (300 microsecond) stimulation; (2) gastric pacing with lowfrequency (e.g., 3 cpm) and long duration (300 millisecond) stimulation; and (3) sequential neural electrical stimulation with multiple pairs of electrodes positioned on the corpus-antrum. Gastric Electrical Stimulation GES at 12 cpm during a 12-month period of continuous treatment significantly decreased nausea and vomiting in patients with refractory, idiopathic, diabetic gastroparesis or postsurgical gastroparesis.287-289 Stimulating the stomach at 12 cpm (four times the normal slow wave frequency) resulted in improvement in nausea and vomiting in patients with gastroparesis.290 No placebo-controlled studies are available to document these benefits of GES. Gastric empty­ ing rates and gastric dysrhythmias did not improve in those patients reporting improvement in symptoms. Vagal affer­ ent nerve stimulation and CNS changes in response to the GES have been proposed as the mechanisms of benefit. Uncontrolled studies from several centers indicate that 50% to 70% of patients treated GES report a decrease in their chronic nausea and vomiting symptoms. Complications include a 10% incidence of infections in the subcutaneous pocket where the device is positioned. There is also a small incidence of fracture of the electrode wires. Gastric Pacing Lin and coleagues291 studied low-frequency gastric stimula­ tion using a 3 cpm stimulus to pace or entrain the normal slow wave rhythm in patients with gastroparesis. Gastric pacing seeks to stimulate three gastric peristaltic contrac­ tions per minute and improve gastric emptying. Electrodes were positioned on the serosal surface of the stomach during surgery in 13 patients. Stimulation at a frequency up to 10% higher than the normal 3 cpm, at an amplitude of 4mA and a pulse width of 300 milliseconds was used to entrain the slow waves. Gastric dysrhythmias were eradicated in some patients. In a similar study of 9 additional patients, electri­ cal stimulation was used to entrain the slow wave and

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders Table 48-4  Drug and Nondrug Therapies Used to Treat Symptoms in Patients with Gastric Neuromuscular Disorders THERAPY

MECHANISMS/SITES OF ACTION

DOSAGE

ADVERSE EFFECTS

Prokinetic Therapy Macrolides Erythyromycin

Motilin receptor agonist

125-250 mg four times daily

Nausea, diarrhea abdominal cramps, rash

Dopamine (D2) receptor antagonist; 5-HT3-receptor antagonist; 5-HT4 receptor agonist D2-receptor antagonist (peripheral)

5-20 mg before meals and at bedtime

Extrapyramidal symptoms, dystonic reactions, anxiety, depression, hyperprolactinemia, tardive dyskinesia Hyperprolactinemia, breast tenderness, galactorrhea

5-HT4 receptor agonist

5-20 mg before meals

Partial 5-HT4 receptor agonist

2-6 mg three times daily

Cardiac dysrhythmias, diarrhea, abdominal discomfort Diarrhea, abdominal pain

Muscarinic antagonist see Endoscopic Therapies

10-20 mg before meals

Drowsiness, dry mouth

5-HT3 receptor antagonist

4-8 mg twice daily, either orally or intravenously 2 mg once daily

Headache, increased liver enzymes

Substituted Benzamides Metoclopramide

Domperidone* Serotonin Agonists Cisapride* Tegaserod* Prorelaxant Therapies Dicyclomine Botulinum toxin (Botox) Anti-nauseant Therapy Serotonin Antagonists Ondansetron Granisetron Phenothiazines Prochlorperazine Antihistamines Promethazine Dimenhydrinate Cyclizine Butyrophenones Droperidol Antidepressants Amitriptyline Nortriptyline Benzodiazepines Lorazepam Alprazolam Electrical Therapies Acustimulation Acupressure/Acupuncture Gastric electrical stimulation Gastric pacing Endoscopic Therapies Botox injection into pylorus Balloon dilation of pylorus Radiofrequency ablation at LES Diet Therapies Gastroparesis diet High-protein drinks Gastrostomy Jejunostomy Total parenteral nutrition

5-HT3 receptor antagonist

10-20 mg before meals and at bedtime

Headache, increased liver enzymes

Central nervous system (CNS) sites

5-10 mg three times daily

Hypotension, extrapyramidal symptoms

CNS, H1 receptor antagonist H1 receptor antagonist H1 receptor antagonist

25 mg twice daily 50 mg four times daily 50 mg four times daily

Drowsiness Drowsiness Drowsiness

Central dopamine receptor antagonist

2.5-5 mg intravenously every 2 hours

Sedation, hypotension

CNS sites CNS sites

25-100 mg at bedtime 10-75 mg at bedtime

Constipation Constipation

CNS sites CNS sites

0.5-1 mg four times daily 0.25-0.5 mg three times daily

Drowsiness, lightheadedness Drowsiness, lightheadedness

Spinal/vagal afferents? Endorphins ?Vagal afferents Control dysrhythmias, improve gastric emptying

Variable NA 12 cpm, 330 milliseconds, 5 mÅ 3 cpm, 300 milliseconds, 4 mÅ

Local tenderness Pocket infection Pocket infection

Relax pyloric muscle Stretch pyloric muscle Improve GEPG, improve gastric myoelectrical activity

25-50 units per quadrant 20 mm balloon, 2 min NA

None Perforation Transient dysphagia

Diet based on gastric emptying physiology Decreases gastric dysrhythmias Venting paretic stomach Enteral nutritional support Bypass paretic stomach

see Table 46-5

None

Unknown As needed As needed As needed

None See Chapter 5 See Chapter 5 Sepsis, thrombosis of central veins

*Compassionate clearance use only. CNS, central nervous system; D2, dopamine2; 5-HT, 5-hydroxytryptamine; GEPG, gastroesophageal pressure gradient; H1, histamine1; LES, lower esophageal sphincter; NA, not applicable.

tachygastria was converted to normal 3 cpm rhythms in 2 patients.292 After one month of gastric pacing treatment, gastric emptying was significantly improved, symptoms of gastroparesis were significantly reduced, and 8 of 9 patients no longer required jejunostomy tube feedings. These pacing

devices require more electrical energy compared with GES; thus, battery life is a major limitation at this time. Adverse events due to gastric pacing devices are discomfort at the site of electrical stimulation and fracture or dislodgement of the electrodes.

813

814

Section VI  Stomach and Duodenum Table 48-5  Diet for Nausea and Vomiting in Patients with Gastric Neuromuscular Disorders DIET Step 1: Sports Drinks and Bouillon For severe nausea and vomiting: Small volumes of salty liquids, with some caloric content to avoid volume depletion Multiple vitamin Step 2: Soups If Step 1 is tolerated: Soup with noodles or rice and crackers Peanut butter, cheese, and crackers in small amounts Caramels or other chewy confection Ingest above foods in at least six small-volume meals/day Multiple vitamin Step 3: Starches, Chicken, Fish If Step 2 is tolerated: Noodles, pastas, potatoes (mashed or baked), rice, baked chicken breast, fish (all easily mixed and emptied by the stomach) Ingest solids in at least six small-volume meals/day Multiple vitamin (liquid or dissolvable)

GOAL

AVOID

1000-1500 mL/day in multiple servings (e.g., 12, 120-mL servings over 12-14 hr) Patient can sip 30-60 mL at a time to reach approximately 120 mL/hr

Citrus drinks of all kinds; highly sweetened drinks

Approximately 1500 calories/day to avoid volume depletion and maintain weight (often more realistic than weight gain)

Creamy, milk-based liquids

Common foods that patient finds interesting and satisfying and that provoke minimal nausea/vomiting symptoms

Fatty foods that delay gastric emptying; red meats and fresh vegetables that require considerable trituration; pulpy fibrous foods that promote formation of bezoars

Modified from Koch KL. Nausea and vomiting. In Wolfe MM, editor. Therapy of Digestive Disorders. 2nd ed. Philadelphia: Elsevier; 2006. pp 1003-17.

Sequential Neural Electrical Stimulation Sequential neural electrical gastric stimulation is gastric pacing that uses a microprocessor to sequentially activate a series of electrodes that encircle the distal two thirds of the stomach. The stimulation sequence induces propagated contractions that cause a forceful emptying of the gastric content.293

soups containing noodles or rice. Milk-based creamy soups are avoided. A dissolvable multiple vitamin is taken daily. Step 3 emphasizes starches and chicken and turkey breast. These solid foods require less gastric work to mix and empty compared with fresh vegetables or red meats. Fried and greasy foods are avoided because fats delay gastric emptying.

ENDOSCOPIC THERAPY

Nutraceuticals

Endoscopic therapies refer to drug or device therapies deliv­ ered to the relevant regions of the stomach via endoscopes. The injection of botulinum toxin into the pylorus to decrease sphincter pressure and to improve gastric emptying and nausea and vomiting in patients with gastroparesis pro­ duced results similar to placebo injections.293-296 Balloon dilation of the pylorus improves gastric emptying in patients with gastroparesis and normal 3 cpm EGG patterns.213 RFA treatment applied to the lower esophageal sphincter (LES) area in patients with GERD and dyspepsia improve gastric dysrhythmias and emptying.240,241

DIET THERAPY Dietary Counseling

Many patients with acute or chronic nausea and vomiting from gastric neuromuscular disorders do not know what to eat and may benefit from dietary counseling.297 In the Nausea/Vomiting (Gastroparesis) Diet, liquid and solid foods that are easy for the stomach to mix and empty are prescribed.271 The Nausea/Vomiting Diet is based on gastric emptying principles and is a three-step diet that requires minimal neuromuscular work of the stomach as the diet is advanced (Table 48-5). Step 1 is primarily electrolyte solutions that are con­ sumed in small amounts over a 24-hour period in order to avoid volume depletion. Liquids require less gastric neuro­ muscular work to empty than solid foods. If patients tolerate step 1, then step 2 may be tried next. Step 2 diet includes

Liquid protein meals with or without ginger decrease nausea associated with motion sickness and gastric dysrhythmias, nausea of pregnancy, and delayed nausea after chemother­ apy.298-301 These protein-based meal therapies have not been formally evaluated in patients with gastroparesis or gastric dysrhythmias. A rationale for the nutraceutical approach to treatment of nausea in patients with gastric neuromuscular disorders has evolved and deserves further study.

Other Approaches to Nutritional Support

For patients with chronic nausea and vomiting, percutane­ ous endoscopic gastrostomy (PEG) tubes may be placed for venting gastric contents in order to avoid frequent vomiting episodes and thereby improve quality of life.302 The venting PEG does not treat the underlying gastric neuromuscular disorder, but it allows the patients to empty the stomach rather than suffer repeated episodes of emesis and discom­ fort. Medications and some nutritional liquids may be toler­ ated when given through the gastrostomy tube. Jejunal feeding tubes for enteral nutrition may be needed to provide basic caloric support for patients with severe nausea and vomiting from gastric neuromuscular disorders. A PEG with J-tube extension usually fails because a single vomiting episode may propel the extension tube into the stomach. A J-tube placed endoscopically or surgically is required (see Chapters 4 and 5).303 Total parenteral nutrition (TPN) via central intravenous catheters should be avoided if at all possible because of the frequent development of line sepsis and occasional venous thrombosis.

Chapter 48  Gastric Neuromuscular Function and Neuromuscular Disorders KEY REFERENCES

Abell T, McCallum R, Hocking M, et al. Gastric electrical stimulation for medically refractory gastroparesis. Gastroenterology 2003; 125:421-8. (Ref 287.) Brzana RJ, Koch KL, Bingaman S. Gastric myoelectrical activity in patients with gastric outlet obstruction and idiopathic gastroparesis. Am J Gastroenterol 1998; 93:1803-9. (Ref 149.) Camilleri M. Integrated upper gastrointestinal response to food intake. Gastroenterology 2006; 131:640-58. (Ref 54.) De Giorgio R, Sarnelli G, Corinaldesi R, Stranghellini V. Advances in our understanding of the pathology of chronic intestinal pseudoobstruction. Gut 2004; 53:1549-52. (Ref 245.) Hinder RA, Kelley KA. Human gastric pacesetter potential: Site of origin, spread, and response to gastric transection and proximal vagotomy. Am J Surg 1997; 133:29-33. (Ref 3.) Huizinga JD. Physiology and pathophysiology of the interstitial cell of Cajal: From bench to bedside II. Gastric motility: Lessons from mutant mice on slow waves and innervation. Am J Physiol 2001; 281:G112934. (Ref 17.) Ladabaum U, Koshy SS, Woods ML, et al. Differential symptomatic and electrogastrographic effects of distal and proximal human gastric distension. Am J Physiol 1998; 275:G418-24. (Ref 91.) Murray CD, Martin NM, Patterson M, et al. Ghrelin enhances gastric emptying in diabetic gastroparesis: a double-blind, placebocontrolled, crossover study. Gut 2005; 54:1693-8. (Ref 105.)

Ordog T. Interstitial cells of Cajal in diabetic gastroenteropathy. Neuro­ gastroenterol Motil 2008; 20:8-18. (Ref 164.) Owyang C, Hasler WL. Physiology and pathophysiology of the intersti­ tial cells of Cajal: From bench to bedside. VI. Pathogenesis and thera­ peutic approaches to human gastric dysrhythmias. Am J Physiol 2002; 283:G8-15. (Ref 29.) Sanders KM, Koh SD, Ward SM. Interstitial cells of Cajal as pacemakers in the gastrointestinal tract. Annu Rev Physiol 2006; 68:307-43. (Ref 2.) Schulze K. Imaging and modeling of digestion in the stomach and the duodenum. Neurogastroenterol Motil 2006; 18:172-83. (Ref 53.) Stern RM, Jokerst MD, Levine ME, Koch KL. The stomach’s response to unappetizing food: Cephalic-vagal effects on gastric myoelectrical activity. Neurogastroenterol Motil 2001; 13:151-4. (Ref 117.) Tougas G, Eaker EY, Abell TL, et al. Assessment of gastric emptying using a low fat meal: Establishment of international control values. Am J Gastroenterol 2000; 95:1456-62. (Ref 43.) Woods SC. Gastrointestinal satiety signals. I. An overview of gastroin­ testinal signals that influence food intake. Am J Physiol Gastrointest Liver Physiol 2004; 286:G7-13. (Ref 114.) Full references for this chapter can be found on www.expertconsult.com.

815

CHAPTE R

49 Gastric Secretion Mitchell L. Schubert and Jonathan D. Kaunitz

CHAPTER OUTLINE Functional Anatomy  817 Acid Secretion: Paracrine, Hormonal, Neural, and Intracellular Regulation  821 Histamine  821 Gastrin  821 Acetylcholine  822 Somatostatin  823 Miscellaneous Peptides  823 Parietal Cell Intracellular Pathways  823 Proton Pump Inhibitors  824 Integrated Response to a Meal: Interplay of Neural, Paracrine, and Hormonal Pathways  825 Helicobacter pylori–Induced Perturbations in Acid Secretion  826

The stomach is an active reservoir that stores, grinds, and slowly dispenses partially digested food to the intestine for further digestion and absorption. Its main secretory function is the production of hydrochloric acid. Gastric acid secre­ tion is present on the first day of life and increases as infants become more mature.1 By two years of age, acid secretion is similar to that of adults, when corrected for body weight.2 Most studies indicate that the rate of acid secretion changes little after the second decade of life unless there is coexi­ sting disease of the acid-secreting glandular mucosa such as infection with Helicobacter pylori (HP) or atrophic gastritis.3-5 Acid facilitates the digestion of protein by converting the proenzyme pepsinogen to the active proteolytic enzyme pepsin. It also facilitates the absorption of iron, calcium, vitamin B12, and certain medications (e.g., thyroxin) as well as prevents bacterial overgrowth, enteric infection, and possibly community-acquired pneumonia.6-18 The stomach also secretes lipase, intrinsic factor, elec­ trolytes (e.g., HCO3−, K+, and Cl−), and mucins in addition to a variety of neurocrine, paracrine, and hormonal agents (Fig. 49-1). Neurocrine agents are released from nerve ter­ minals and reach their targets via synaptic diffusion (e.g., acetylcholine [ACh], gastrin-releasing peptide [GRP], and vasoactive intestinal peptide [VIP]). Paracrine agents are released in proximity to their targets and reach them via diffusion (e.g., histamine and somatostatin). Hormones are released into the circulation and reach their targets via the bloodstream (e.g., gastrin). Gastric mucosal integrity depends on a delicate balance between secretion of aggressive (e.g., acid and pepsin) and defensive (e.g. bicarbonate and mucin) substances (Fig. 49-2).19 When mucosal defense mechanisms are over­ whelmed, ulceration may occur. In order to reap the benefits of acid without untoward effects, gastric exocrine and endocrine secretion must be precisely regulated. This is

Measurement of Gastric Acid Secretion  827 Indications for Secretory Testing  827 Methods for Measuring Acid Secretion  827 Basal Acid Output  827 Maximal Acid output and Peak Acid Output  827 Sham Feeding–Stimulated Acid Output  828 Meal-Stimulated Acid Output  828 Diseases Associated with Increased Gastric Acid Secretion  828 Pepsinogen Secretion  829 Gastric Lipase Secretion  829 Intrinsic Factor Secretion  829 Bicarbonate Secretion  830 Mucus Secretion  831

accomplished by a highly coordinated interaction among a multitude of neural, paracrine, and hormonal pathways.

FUNCTIONAL ANATOMY The stomach consists of three anatomic (fundus, corpus or body, and antrum) and two functional (oxyntic and pyloric gland) areas (Fig. 49-3). The oxyntic gland area, the hall­ mark of which is the oxyntic (oxys, Greek for acid) or parietal cell, comprises 80% of the organ (fundus and corpus). The pyloric gland area, the hallmark of which is the G or gastrin cell, comprises 20% of the organ (antrum). The human stomach contains approximately 1 × 109 parietal cells and 9 × 106 gastrin cells.20 There is debate as to whether the cardia, a transition zone of 0 to 9 mm between the squamous mucosa of the esophagus and the oxyntic mucosa of the stomach, exists as a normal anatomic structure or develops as a result of abnormal reflux. Autopsy and endoscopic studies suggest that cardiac mucosa is absent in more than 50% of the general population.21 Gastric anatomy is discussed in greater detail in Chapter 47. The glandular area is organized in vertical tubular units that consist of an apical pit region, an isthmus, and the actual gland region that forms the lower part of the unit; the latter consists of a neck and a base (Fig. 49-4). The progeni­ tor cell of the gastric unit, located in the isthmus, gives rise to all gastric epithelial cells. In the oxyntic gland area, the mucus-producing pit cells migrate upward from the pro­ genitor cell toward the gastric lumen. Acid-secreting pari­ etal cells migrate downward to the middle and lower regions of the gland; those at the bases are less active acid secretors. The turnover time for parietal cells is 54 days in mice and 164 days in rats.20 Chief (zymogenic) cells predominate at the base and secrete pepsinogen and leptin22; the latter is

817

Section VI  Stomach and Duodenum H+

ECL cell (Histamine)

Gastrin cell

M

+

H+

H+

Parietal cell oo

d

ve

ss

el H2

Bl

818

3

+

K-2

CC

ACh

Neurocrine pathway

+

SST − R2

D cell (SST)

Hormonal pathway

Paracrine pathway

Figure 49-1.  Neural, hormonal, and paracrine pathways directly regulating parietal cell acid (H+) secretion. Left: Acetylcholine (ACh), released from postganglionic intramural neurons within the oxyntic mucosa, stimulates the parietal cell directly via M3 receptors coupled to release of intracellular calcium. Center: Gastrin, released from antral G cells, travels in the circulation to reach the parietal cell and directly activates cholecystokinin-2 (CCK-2) receptors coupled to release of intracellular calcium. Right: Histamine, released from oxyntic enterochromaffin-like (ECL) cells, diffuses to the parietal cell and directly activates histamine (H2) receptors coupled to generation of cyclic adenosine monophosphate (cAMP). Somatostatin (SST), released from oxyntic D cells, diffuses to the parietal cell and directly activates SSTR2 receptors coupled to inhibition of acid secretion. Ach, gastrin, and SST also have important indirect actions affecting acid secretion not shown here. +, stimulatory; −, inhibitory.

OFFENSE Toxins

Steroids NSAIDs Microbes

ROS

H+ Pepsin

Mucus HCO3−

Phospholipids Figure 49-2.  Gastroduodenal offense and defense. Mucosal integrity depends on a delicate balance between aggressive and defensive factors. When mucosal defense mechanisms are overwhelmed, ulceration may occur. CA, carbonic anhydrase; CGRP, calcitonin gene-related peptide; H2S, hydrogen sulfide; HCO3−, bicar­ bonate; NO, nitric oxide; NSAIDs, nonsteroidal anti-inflammatory drugs; ROS, reactive oxygen species.

Mucosal blood flow

Neutrophils Cytokines

H+ Pepsin

H+ Pepsin

H+ Pepsin

H+ Pepsin

Mucus HCO3−

Mucus HCO3−

Mucus HCO3−

Mucus HCO3−

Epithelial integrity

Regulatory peptides Traditional growth factors Cytokines

Nerves (NO; CGRP) Myofibrolasts Immune cells DEFENSE

H2S CA

Prostaglandins

Chapter 49  Gastric Secretion also present in parietal cells.23 Several distinct neuroendo­ crine cell types are contained within the gland but only some of their products have been assigned physiologic func­ tions (see Chapter 1). The cells include enterochromaffin (EC) cells (atrial natriuretic peptide [ANP], serotonin, and adrenomedullin), enterochromaffin-like (ECL) cells (hista­ mine), D cells (somatostatin and amylin), and A-like or Gr cells (ghrelin and obestatin).24-29 ECL cells constitute 66% of the neuroendocrine cell population in rats and 30% in

Esophagus Fundus

ia

rd

Ca

Corpus (body) Angularis incisura Lesser curve

Antrum Greater curve Duodenum G cell (gastrin)

Parietal cell (H+)

Figure 49-3.  Functional gastric anatomy. The stomach consists of three anatomic (fundus, corpus or body, and antrum) and two functional (oxyntic and pyloric gland) areas. The hallmark of the oxyntic gland area is the parietal cell. The hallmark of the pyloric gland area is the G or gastrin cell.

humans. Somatostatin-containing D cells possess neurallike cytoplasmic processes that terminate in the vicinity of parietal and ECL cells (see Fig. 49-4). The functional correlate of this anatomic coupling is a tonic paracrine restraint exerted by somatostatin on acid secretion directly as well as indirectly by inhibiting histamine secretion (Fig. 49-5).30-32 Somatostatin-containing D cells are also present in the pyloric gland area (see Figs. 49-4 and 49-5); in this region they exert a tonic paracrine restraint on gastrin secretion from G cells.33,34 The pyloric gland also contains EC cells (ANP and serotonin), A-like or Gr cells (ghrelin and obestatin), and endocrine cells containing orexin.26,35,36 The stomach is innervated by a neural network, the enteric nervous system (ENS), that contains intrinsic neurons, that is, neurons whose cell bodies are contained within the gastric wall (Figs. 49-6 to 49-8). The ENS, the third division of the autonomic nervous system (the other two being the sympathetic and parasympathetic), is often referred to as the “little brain” because it contains as many neurons as the spinal cord, approximately 106, and can function autonomous of central input (see Fig. 49-6).37 Nevertheless, the ENS receives input from and sends input to the central nervous system via sympathetic and parasym­ pathetic neurons. In rats and guinea pigs, most of the intrin­ sic neural innervation of the stomach originates in the myenteric plexus, located between the circular and longitu­ dinal muscle layers; the submucosal plexus in these species, adjacent to the mucosal layer, contains only a small number of neurons. Humans, in contrast, have a clearly defined submucosal plexus that regulates gastric secretion and con­ tains a variety of neurotransmitters (see Figs. 49-7 and 49-8). It should be noted that the vagus nerve is predominantly afferent, containing 80% to 90% afferent fibers and 10% to 20% efferent fibers. The efferent fibers arise from the dorsal motor nucleus of the brainstem. They are preganglionic and do not directly innervate parietal or neuroendocrine cells but rather synapse with postganglionic neurons of the ENS. The postganglionic neurons contain a variety of transmitters including ACh, GRP, nitric oxide (NO), VIP, and pituitary

OXYNTIC GLAND

Surface mucous cell Parietal cell Mucous neck cell Enterochromaffin-like cell (histamine)

PYLORIC GLAND

GASTRIC PIT (FOVEOLUS) ISTHMUS (PROGENITOR ZONE)

Mucous neck cell

NECK G cell (gastrin)

D cell (somatostatin) Chief cell (pepsinogen) Enterochromaffin cell (ANP)

Surface mucous cell

BASE

D cell (somatostatin)

Enterochromaffin cell (ANP) Figure 49-4.  Gastric gland anatomy. Somatostatin-containing D cells contain cytoplasmic processes that terminate in the vicinity of acid-secreting parietal and histamine-secreting enterochromaffin-like cells in the oxyntic gland area (fundus and corpus) and gastrin-secreting G cells in the pyloric gland area (antrum). The functional correlate of this anatomic coupling is a tonic paracrine restraint on acid secretion by somatostatin that is exerted directly on the parietal cell as well as indirectly by inhibiting histamine and gastrin secretion. ANP, atrial natriuretic peptide. (From Schubert ML, Peura DA. Control of gastric acid secretion in health and disease. Gastroenterology 2008; 134:1842-60.)

819

Section VI  Stomach and Duodenum H+ D cell (SST)

−

Gastrin SSTR2 cell

D cell (SST)

Parietal SSTR2 − cell

H2

− R2 SST

Antrum

+ Fundus

820

ECL cell (Histamine)

Figure 49-5.  Model illustrating the inhibitory actions of somatostatin (SST) on gastric acid secretion in the oxyntic gland area (fundus and body) and the pyloric gland area (antrum). SST-containing D cells are structurally and functionally coupled to their target cells: parietal, enterochromaffin-like (ECL), and gastrin cells. SST, acting via SSTR2 receptors, tonically restrains acid secretion. This restraint is exerted directly on the parietal cell as well as indirectly by inhibiting histamine secretion from ECL cells and gastrin secretion from G cells.

AUTONOMIC NERVOUS SYSTEM

Sympathetic division

Vagus nerve (preganglionic) Parasympathetic division

Enteric Division Myenteric plexus Submucosal plexus

Figure 49-6.  The autonomic nervous system consists of three divisions: sympathetic, parasympathetic, and enteric. The enteric division consists of the myenteric plexus, which regulates motility and the submucosal plexus, which regulates secretion. Although the enteric division can function auto­nomously, it receives input from and sends projections to the other divisions.

ACh

ACh

ACh

ACh

ACh

Intramural gastric nerves (postganglionic)

ENTERIC NERVOUS SYSTEM Myenteric plexus

Circular muscle ACh

GRP

NO

VIP

PACAP

Target cell G/D/ECL/parietal

Longitudinal muscle

Mucosa

Submucosal plexus

Figure 49-7.  The enteric nervous system contains intrinsic neurons, the cell bodies of which are contained within the gastric wall. The myenteric plexus, which innervates the circular and longitudinal muscle layers, regulates motility. The submucosal plexus, which innervates the mucosa, regulates secretion.

Figure 49-8.  Functional neural anatomy. The vagus nerve contains preganglionic neurons that synapse with postganglionic neurons within the wall of the stomach and that are part of the enteric nervous system. The postganglionic neurons contain a variety of transmitters including acetylcholine (ACh), gastrin-releasing peptide (GRP, or mammalian bombesin), nitric oxide (NO), vasoactive intestinal peptide (VIP), and pituitary adenylate cyclase–activating by the parietal cell polypeptide (PACAP). The postganglionic neurons regulate acid secretion by the parietal cell directly and/or indirectly by modulating the secretion of gastrin from G cells, somatostatin from D cells, and possibly histamine from enterochromaffin-like (ECL) cells.

TR

Somatostatin −

M

3

HISTAMINE

Histamine, produced in ECL cells by decarboxylation of L-histidine by histidine decarboxylase (HDC), stimulates the parietal cell directly by binding to H2 receptors coupled to activation of adenylate cyclase and generation of cAMP (see Fig. 49-9).43 Histamine also stimulates acid secretion indirectly by binding to H3 receptors coupled to inhibition of somatostatin release from oxyntic D cells, thus resulting in stimulation of histamine release and acid secretion (see Fig. 49-10).44,45 Gastrin, PACAP, VIP, and ghrelin stimulate, whereas somatostatin, CGRP, prostaglandins, peptide YY (PYY), and galanin inhibit histamine secretion.46,47 As dis­ cussed following, gastrin also exerts a direct proliferative effect on ECL cells. ACh has no direct effect on histamine secretion.48-50

Gas trin

AC

Ca++ Parietal cell

ATP

+

K-2

Parietal cells secrete hydrochloric acid at a concentration of approximately 160 mM or pH 0.8. Acid is thought to gain access to the lumen via channels in the mucus layer created by the relatively high intraglandular hydrostatic pressures generated during secretion, about 17 mm Hg.40 Acid facilitates the digestion of protein and absorption of iron, calcium, and vitamin B12 as well as prevents bacterial overgrowth, enteric infection, and possibly community acquired pneumonia.6-14,18,41 However, when levels of acid (and pepsin) overwhelm mucosal defense mechanisms, ulcers occur. To prevent such damage, gastric acid must be precisely regulated and produced according to need. This is accomplished by a highly coordinated interaction among a number of neural, hormonal, and paracrine pathways. These pathways can be activated directly by stimuli origi­ nating in the brain or reflexively by stimuli originating in the stomach after ingestion of a meal such as mechanical stimulation (e.g., distention) or chemical stimulation (e.g., protein and acid). The principal stimulants of acid secretion are (1) ACh, released from postganglionic enteric neurons (neurocrine), (2) gastrin, released from antral G cells (hormonal), and (3) histamine, released from oxyntic ECL cells (paracrine) (see Fig. 49-1; Fig. 49-9). These agents interact with specific G protein–binding receptors (M3, CCK-2, and H2, respectively) that are coupled to two major signal transduction pathways: intracellular calcium in the case of gastrin and ACh, and adenylate cyclase or adenosine 3′,5′-cyclic monophosphate (cAMP) in the case of histamine (see Fig. 49-9). There is evidence for potentiation (or synergism) between histamine and either ACh or gastrin, probably as a result of postrecep­ tor interaction between the two signaling pathways.42 The main inhibitor of acid secretion is somatostatin, released from oxyntic and antral D cells (paracrine) (see Figs. 49-1, 49-5, and 49-9). Each of these agents acts directly on the parietal cell as well as indirectly by modulating the secre­ tion of neuroendocrine cells (Fig. 49-10).

n

T SS

CC

ACID SECRETION: PARACRINE, HORMONAL, NEURAL, AND INTRACELLULAR REGULATION

Ga str i

H2

R2

+

+

Histamine +

− AC h

ECL cell

SS

D cell

CCK-2

adenylate cyclase–activating polypeptide (PACAP) (see Fig. 49-8).38 In the stomach, afferent nerve fibers containing calcitonin gene–related peptide (CGRP) are of extrinsic origin, that is, the cell bodies are located outside the stomach wall.39 Postganglionic neurons of the ENS regulate acid secretion directly, as is the case for ACh, and/or indirectly by modulating the secretion of gastrin from G cells, soma­ tostatin from D cells, and possibly histamine from ECL cells (see Fig. 49-8).

2

Chapter 49  Gastric Secretion

cAMP

H+,K+ATPase

Ca++

K+

Acid Figure 49-9.  Model illustrating parietal cell receptors and transduction pathways. The principal stimulants of acid secretion at the level of the parietal cell are histamine (paracrine), gastrin (hormonal), and acetylcholine (ACh; neurocrine). Histamine, released from enterochromaffin-like (ECL) cells, binds to H2 receptors that activate adenylate cyclase (AC) and generate adenosine 3´,5´-cyclic monophosphate (cAMP). Gastrin, released from G cells, binds to cholecystokinin-2 (CCK-2) receptors that activate phospholipase C (not shown) to induce release of cytosolic calcium (Ca++). Gastrin stimulates the parietal cell directly and, more importantly, indirectly by releasing histamine from ECL cells. ACh, released from intramural neurons, binds to M3 receptors that are coupled to an increase in intracellular calcium. The intracellular cAMP- and calcium-dependent signaling systems activate downstream protein kinases, ultimately leading to fusion and activation of H+,K+-ATPase, the proton pump. Somatostatin, released from oxyntic D cells, is the principal inhibitor of acid secretion. Somato­ statin, acting via the SSTR2 receptor, inhibits the parietal cell directly as well as indirectly by inhibiting histamine release from ECL cells. +, stimulatory; −, inhibitory.

GASTRIN

Gastrin, the main stimulant of acid secretion during meal ingestion, is produced in G cells of the gastric antrum and, in much lower and variable amounts, in the proximal small intestine, colon, and pancreas. Gastrin is synthesized as a large precursor molecule of 101 amino acids, which is con­ verted to progastrin (80 amino acids) by cleavage of the N-terminal signal peptide. Progastrin is further processed to yield peptides with C-terminal glycine, that is, G34gly and G17gly. The final processing step involves amidation to yield G34amide and G17amide. The plasma half-life of G34amide is 30 minutes and that of G17amide is three to seven minutes; they are metabolized primarily by the kidney and, in addition, by the intestine and liver.51,52 Conse­ quently, most gastrin in the circulation during fasting is G34, whereas after a meal it is G17. In patients with renal insuffiency as well as massive small bowel resection, fasting blood levels of G17 and G34 are elevated.53,54 It should be noted that the commercially available test substance penta­ gastrin is not a naturally occurring peptide but rather is a manufactured analog that contains the biologically active C-terminus sequence Trp-Met-Asp-Phe-NH2. Gastrin and cholecystokinin (CCK) belong to the same family of peptides and possess an identical carboxylterminal pentapeptide sequence (-Gly-Trp-Met-Asp-PheNH2). Two main classes of gastrin/CCK receptors have been characterized: CCK1 (formerly CCK-A) and CCK-2 (formerly CCKB or CCKB/gastrin). CCK1 receptors are specific for CCK, whereas CCK-2 receptors recognize both CCK and gastrin with high affinity. Gastrin, acting via CCK-2 receptors that

821

822

Section VI  Stomach and Duodenum VAGUS

CGRP

VIP ACh +

ACh GRP

−

+

D cell (SST) +

−

+

ACh

+

G cell (GASTRIN)

+ +

PARIETAL cell H2

−

+

+

−

−

D cell (SST)

+

H3 −

ECL cell (Histamine)

HP (Chronic) Acid HP (Acute)

ACh

ANTRUM

−

FUNDUS

Figure 49-10.  Model illustrating the neural, paracrine, and hormonal regulation of gastric acid secretion. Efferent vagal fibers synapse with intramural gastric cholinergic (ACh) and peptidergic (gastrin-releasing peptide [GRP] and vasoactive intestinal peptide [VIP]) neurons. In the fundus (oxyntic mucosa), ACh neurons stimulate acid secretion directly as well as indirectly by inhibiting somatostatin (SST) secretion, thus eliminating its restraint on parietal cells and histamine-containing enterochromaffin-like (ECL) cells. In the antrum (pyloric mucosa), ACh neurons stimulate gastrin secretion directly as well as indirectly by inhibiting SST secretion, thus eliminating its restraint on gastrin-containing G cells. GRP neurons, activated by luminal protein, also stimulate gastrin secretion. VIP neurons, activated by low-grade gastric distention, stimulate SST and thus inhibit gastrin secretion. Dual paracrine pathways link SST-containing D cells to parietal cells and to ECL cells in the fundus. Histamine released from ECL cells acts via H3 receptors to inhibit SST secretion. This serves to accentuate the decrease in SST secretion induced by cholinergic stimuli and thus augments acid secretion. In the antrum, dual paracrine pathways link SST-containing D cells to gastrin cells. Release of acid into the lumen of the stomach restores SST secretion in both the fundus and antrum; the latter is mediated via release of calcitonin gene-related peptide (CGRP) from extrinsic sensory neurons. Acute infection with Helicobacter pylori (HP) also activates CGRP neurons to stimulate SST and thus inhibit gastrin secretion. In duodenal ulcer patients who are chronically infected with HP, the organism or cytokines released from the inflammatory infiltrate inhibit SST and thus stimulate gastrin (and acid) secretion.

activate phospholipase C to induce release of intracellular calcium, stimulates the parietal cell directly and, more importantly, indirectly by releasing histamine from ECL cells (see Figs. 49-9 and 49-10).55,56 Gastrin regulates the secretion and synthesis of histamine in a biphasic manner. The first phase involves release of stored histamine. The second phase relates to the replenishment of histamine stores and involves an increase in HDC activity followed by an increase in HDC gene transcription.57 H2 receptor, HDC, and CCK-2 receptor knockout mice manifest decreased acid secretion, especially in response to gastrin.58-60 Although amidated gastrins had been thought to be the only forms with biological activity, glycine-extended gas­ trins may regulate the capacity of the parietal cell to respond to secretagogues, release histamine from ECL cells, and stimulate proliferation of colonic mucosa and colorectal cancers.61,62 ACh, GRP, secretin, β2/β3-adrenergic agonists, calcium, protein, and alcoholic beverages produced by fer­ mentation stimulate, whereas somatostatin, galanin, and adenosine inhibit gastrin secretion. In addition, at least two negative-feedback pathways, mediated via release of soma­ tostatin, regulate gastrin secretion. The first is activated by luminal acidity and involves sensory CGRP neurons (see Fig. 49-10). Low intragastric pH (high intragastric acidity) activates CGRP neurons that, via an axon reflex, stimulate somatostatin and thus inhibit gastrin secretion.63-65 Con­ versely, when intragastric pH rises (low intragastric acidity), for example, by administering antisecretory medications such as proton pump inhibitors (PPIs) or by developing gastric atrophy, somatostatin secretion is inhibited and patients develop hypergastrinemia. There is some evidence that bacterial overgrowth induced by hypochlorhydria may also contribute to hypergastrinemia.66 The second negativefeedback pathway involves a paracrine action whereby

gastrin directly stimulates somatostatin and thus attenuates its own secretion (see Fig. 49-10).67 Gastrin also functions as a trophic hormone to stimulate mucosal proliferation. CCK-2 receptors have been localized to the progenitor zone in oxyntic glands, and chronic hyper­ gastrinemia induces proliferation of ECL and parietal cells directly as well as indirectly via the autocrine or paracrine action of growth factors such as heparin-binding epidermal growth factor, amphiregulin, transforming growth factor-α, metalloproteinases, and regenerating islet-derived 1.68,69 Rats rendered hypergastrinemic by a PPI demonstrate a five-fold increase in the number of ECL cells and a 1.5-fold increase in the number of parietal cells.70 Gastrin acts directly on ECL cells to induce hyperplasia, dysplasia, and eventually neoplasia (carcinoids).71 In con­ trast to rodents, humans rarely develop carcinoid tumors in response to hypergastrinemia unless other factors are present such as chronic active gastritis or gastrinoma asso­ ciated with multiple endocrine neoplasia type 1 (see Chapter 31).72 Because ECL cells contain somatostatin subtype 2 receptors (SSTR2), somatostatin scintigraphy with 111indiumdiethylenetriamine pentaacetic acid [111In-DTPA]octreotide is the preferred imaging method to detect carcinoid tumors (see Chapter 31).73,74

ACETYLCHOLINE

ACh, released from postganglionic neurons whose cell bodies are located primarily in the submucosal (Meissner’s) plexus, stimulates the parietal cell directly as well as indirectly by inhibiting somatostatin secretion (see Fig. 49-10). The parietal cell muscarinic receptor is of the M3 subtype.75,76 Like CCK-2 receptors, M3 receptors are coupled to activation of phospholipase C with generation of inositol trisphosphate and release of intracellular calcium

Chapter 49  Gastric Secretion

SOMATOSTATIN

The main inhibitor of acid secretion is somatostatin. Soma­ tostatin is synthesized from a 92-amino acid preproso­ matostatin precursor molecule that is processed to yield somatostatin-14 and somatostatin-28. Somatostatin-14 is predominantly found in stomach, pancreatic islets, and enteric neurons, whereas somatostatin-28 is the major form in small intestine. The half-life of somatostatin-14 is 1 to 3 minutes, whereas the half-life of somatostatin-28 is about 15 minutes. In the stomach, somatostatin cells are closely coupled to their target cells (gastrin cells in the antrum, or parietal and ECL cells in the fundus/body) either directly via cytoplas­ mic processes or indirectly via the local circulation.28,79 The functional correlate of this anatomic coupling is a tonic restraint exerted by somatostatin on acid secretion from the parietal cell, histamine secretion from the ECL cell, and gastrin secretion from the G cell (see Figs. 49-5 and 49-10).30,31,34,80,81 Removing this restraint (i.e., disinhibition or elimination of the influence of an inhibitor), by activation of cholinergic neurons, is an important physiologic mecha­ nism for stimulating acid secretion (see Fig. 49-10). In the stomach, the actions of somatostatin are mediated primarily via the somatostatin subtype 2 receptor (SSTR2).82-84 Gastrin, GRP, VIP, PACAP, β2/β3-adrenergic agonists, secretin, ANP, adrenomedullin, amylin, adenosine, and CGRP stimu­ late, whereas ACh and interferon-γ inhibit somatostatin secretion. As mentioned, an increase in luminal acidity acts to attenuate acid secretion via a pathway involving release of somatostatin in the antrum and the fundus. The change in gastric somatostatin secretion can be demonstrated over a range of pH 3 to pH 5, which is within the range observed after ingestion of a meal (Fig. 49-11).85

MISCELLANEOUS PEPTIDES

Ghrelin, the natural ligand for the growth hormone secreta­ gogue receptor, is present in greatest concentrations in gastric oxyntic mucosa and is localized to A-like (or Gr) cells.86,87 Lesser amounts are present in the antrum, small intestine, and colon (see Chapter 1). Plasma ghrelin concen­ trations increase before meals and decrease postprandi­ ally.88 It is postulated that ghrelin triggers premeal hunger and promotes feeding. Its suppression after Roux-en-Y gastric bypass may, in part, contribute to weight loss.89 Most studies report that exogenously administered ghrelin stimu­ lates acid secretion.90,91 The stimulatory effect appears to involve the vagus nerve and histamine release because it is abolished by vagotomy and is associated with an increase in HDC messenger ribonucleic acid (mRNA).92,93 Orexin-A, derived from propro-orexin by posttranslational processing, is co-localized with gastrin in human pyloric mucosa.91,94 Intracerebroventricular and peripherally administered orexin-A stimulate gastric acid secretion.95 In rats equipped with gastric fistulas, an orexin receptor 1 antagonist inhibits basal and pentagas­ trin-stimulated acid secretion, implying that endogenous orexin-A stimulates acid secretion.94,95 ANP, CCK, secretin, neurotensin, glucagon-like peptide 1 (GLP-1), glicentin, oxyntomodulin, peptide YY, adreno­

180 160 140 Somatostatin secretion % of basal

(see Fig. 49-9).77 Alcoholic beverages produced by fermen­ tation stimulate gastric acid secretion and the effect may be mediated via activation of M3 receptors.78 ACh also stimulates acid secretion indirectly by activating M2 and M4 receptors on D cells coupled to inhibition of somato­ statin secretion, thus removing the tonic restraint exerted by this peptide on gastrin, ECL, and parietal cells (see Fig. 49-10).

120 100

_

ADD H+

ADD HCO3

80 60 40 20 0 4.8

4.4

4.0

3.6

3.2

pH luminal perfusate Figure 49-11.  Relationship between luminal pH and gastric somatostatin secretion. In isolated mouse stomach, addition of bicarbonate (HCO3−) to neutralize basal acid secretion or HCl to augment luminal acidity causes a corresponding change in somatostatin secretion. (From Schubert ML, Edwards NF, Makhlouf GM. Regulation of gastric somatostatin secretion in the mouse by luminal acid: A local feedback mechanism. Gastroen­ terology 1988; 94:317-22).

medullin, amylin, glucose-dependent insulinotropic poly­ peptide (GIP), leptin, epidermal growth factor, and interleukin-1β (IL-1β) inhibit acid secretion. The effect of each, except perhaps for IL-1β, is mediated via release of somatostatin.24-26,96 The term enterogastrone has been used to describe the intestinal factor or factors responsible for inhibiting acid secretion in response to nutrients in the intestine. Prime candidates include CCK, secretin, neuro­ tensin, GLP-1, glicentin, and oxyntomodulin because they are present in intestinal mucosa, released into the circula­ tion in response to luminal nutrients, and capable of inhibit­ ing acid secretion at “physiologic” concentrations.97-101 Although it is likely that enterogastrone activity represents the combined influence of several of these peptides, the strongest evidence favors CCK. CCK, produced in I cells in the proximal small intestine, is released by luminal protein and fat. The acid-inhibitory response to intraduodenal fat is abolished by pretreatment with a CCK-1 receptor antagonist in dogs, and the response is blocked in CCK-1 receptor knockout mice as well.102-105

PARIETAL CELL INTRACELLULAR PATHWAYS

In parietal cells, acid secretion is increased by activation of intracellular cAMP- and calcium-dependent signaling pathways that activate downstream protein kinases, ulti­ mately leading to fusion and activation of H+,K+-ATPase (the proton pump) with concomitant activation of luminal mem­ brane conductances for K+ and Cl− (see Fig. 49-9; Fig. 49-12). The H+,K+-ATPase actively pumps out H+ against a tremen­ dous concentration gradient (cell interior pH 7.4 or 40 nM; acid secreted at pH 0.8 or 160 million nM) in exchange for luminal K+. The energy required comes from adenosine triphosphate (ATP) produced by the parietal cell’s extensive

823

Section VI  Stomach and Duodenum HCO3− CI− 7

AE2

H+ + HCO3− 3Na+

H2CO3 CA CO2 + H2O

2K+

H+,K+-ATPase α-Subunit β-Subunit

CI− SL C2 6A

824

2CI− CI−

Tubulovesicle

Na+ NKCCI

K+

K+ Resting parietal cell

H+,K+ATPase

H+

K+

CI−

HCI Figure 49-12.  Model illustrating ion transport in the parietal cell. Acid secretion requires a functional H+,K+-ATPase as well as apical K+ and Cl− channels and basolateral transporters and/or channels for K+, Cl−, and HCO3−. Acid is produced from the hydration of CO2 to form H+ and HCO3−, a reaction catalyzed by cytoplasmic carbonic anhydrase (CA). In the presence of luminal K+, H+,K+-ATPase pumps H+ into the lumen in exchange for K+. Luminal K+ channels (KCNE2/KCNQ1 and ROMK [not shown]) recycle K+ across the luminal membrane. The source of intracellular K+ is the basolateral Na+,K+-ATPase and the sodium-2 chloride potassiumcotransporter-1 (NKCC1). For each H+ secreted, a HCO3− exits the cell across the basolateral membrane via the anion exchanger (AE2, or Slc4a2). Concurrently with H+, Cl− is extruded across the luminal membrane via an apical chloride channel. The sources of intracellular Cl− are AE2, NKCCl, and the SLC26A7 channel.

mitochondrial network. H+,K+-ATPase consists of an αsubunit that carries out the catalytic and transport function of the enzyme and contains sequences responsible for apical membrane localization106 as well as a β-subunit, which is heavily glycosylated, and protects the enzyme from degra­ dation and is necessary for trafficking to and from the luminal membrane.107 In the resting unstimulated state, H+,K+-ATPase activity is sequestered within cytoplasmic tubulovesicles. On stimula­ tion, there is a dramatic morphologic transformation as these vesicles fuse with the apical plasma membrane, result­ ing in a 6- to 10-fold increase in the membrane and the formation of the canalicular system (Fig. 49-13). Transloca­ tion of the H+,K+-ATPase into the canalicular membrane together with the presence of luminal K+ activates the enzyme.108 On cessation of secretion, the H+,K+-ATPase is retrieved from the apical membrane and the tubulovesicular compartment is reestablished. The precise mechanisms regulating trafficking are not known, but data suggest that it involves actin-based microfilaments, small GTPases, docking/fusion proteins, ezrin, and clathrin.109-111 Acid secretion requires not only a functional H+,K+ATPase but also apical K+ and Cl− channels and basolateral HCO3− and Cl− exchangers. Acid is produced from the hydra­ tion of CO2 to form H+ and HCO3−, a reaction catalyzed by cytoplasmic carbonic anhydrase (see Fig. 49-12). Because the H+,K+-ATPase is unable to pump H+ into the lumen without a parallel uptake of K+, sufficient quantities of K+ must be delivered to the lumen. This K+ recycling is accom­ plished by luminal KCNE2/KCNQ1 and ROMK (KCNJ1)

Stimulated parietal cell

Figure 49-13.  Model illustrating translocation and activation of H+,K+ATPase. In the resting state, H+,K+-ATPase is sequestered within cyto­ plasmic tubulovesicles and is inactive. On stimulation, the tubulovesicles move to and fuse with the apical membrane, forming an extensive canalicular system. Translocation of H+,K+-ATPase into the canalicular membrane together with the presence of luminal K+ activates the enzyme.

Benzimidazole Imidazole Benzene H N

Pyridine O

CH2 S Methylsulphinyl bridge N

N

Figure 49-14.  Structure of proton pump inhibitors (PPIs). PPIs consist of two heterocyclic moieties, a benzimidazole ring, and a pyridine, connected by a methylsulphinyl bridge. PPIs are weak bases (pKa 4-5) that accumulate and activate in acidic spaces within the body that have a pH less than 4. Once activated within the parietal cell canaliculus, the PPI binds covalently with certain cysteine residues within the α-subunit of the inserted H+,K+-ATPase.

potassium channels. KCNQ1 is a voltage-activated K+ channel, which, when modified by the small regulatory subunit, KCNE2, becomes voltage insensitive, constitutively open, and acid activated.108,112,113 ROMK may be regulated by the cystic fibrosis transmembrane conductance regulator (CFTR).108,112-114 The concentration of K+ in gastric juice (8 to 12 mM) exceeds plasma K+ by two- to four-fold. For each H+ secreted, an HCO3− ion exits the cell across the basolateral membrane via the anion exchanger 2 (AE2), Slc4a2 (see Fig. 49-12).115 As a result of this HCO3−/Cl− exchange, the pH within the parietal cell remains only slightly alkaline during acid secretion.116 Rapid entry of HCO3− from parietal cells into blood has been referred to as the alkaline tide. Some of this HCO3− may be taken up and secreted by surface epithelial cells. Concurrently with H+, Cl− is extruded across the luminal membrane via an apical chloride channel, the precise identity of which is not known. The sources of intra­ cellular Cl− are the basolateral anion exchanger 2 (AE2), sodium-2 chloride potassium-cotransporter-1 (NKCC1), and the SLC26A7 channel (see Fig. 49-12).117,118

PROTON PUMP INHIBITORS

Current PPIs (e.g., omeprazole, lansoprazole, rabeprazole, pantoprazole, and esomeprazole) consist of two heterocy­ clic moieties, a pyridine and a benzimidazole ring, con­ nected by a methylsulphinyl group (Fig. 49-14). They are weak bases (pKa 4 or 5) that concentrate in acidic spaces within the body that have a pH less than 4. The pKa of a

Chapter 49  Gastric Secretion disulfide bonds by reducing agents such as glutathione (15 hours in rat) could also play a role.123

PPI

PPI

PPI+ +

+

H ,K - Sulfenamide+ ATPase S-S-Sulfenamide+

Blood (pH 7.4)

Cytosol (pH 7.3)

Secretory canaliculus (pH <2.0)

Figure 49-15.  Model illustrating the mechanism of action of proton pump inhibitors (PPIs). PPIs reach the parietal cell from the bloodstream, diffuse through the cytoplasm, and accumulate in the acid environment of the secretory canaliculus. In the canaliculus, the PPI becomes protonated and trapped as a sulfenic acid followed by dehydration to a sulfenamide. The sulfenamide binds covalently by disulfide bonds to one or more cysteines of the H+,K+-ATPase to inhibit the enzyme. Whereas all PPIs bind to cysteine 813, omeprazole also binds to cysteine 892, lansoprazole to cysteine 321, and pantoprazole to cysteine 822.

molecule refers to the degree of willingness of the com­ pound to accept or donate a proton and is based on a loga­ rithmic scale such that a compound with a pKa of 5 is 10-fold more basic than a compound with a pKa of 4. When a compound is in an environment with a pH equal to its pKa, half the molecules will be protonated and half will be non-protonated. If a compound with a pKa of 4 is placed in an environment with a pH less than 1, more than 99.9% of the molecules will be protonated. PPIs are membrane per­ meable in the nonprotonated form and relatively imperme­ able in the protonated form. In blood (pH 7.4), PPIs are predominantly nonprotonated and thus pass readily into cells (time to reach peak plasma concentration, ≈2 hours; elimination half-life, ≈1 hour), diffuse through the cyto­ plasm, and then become protonated and trapped, probably as sulfenamides or sulfenic acid, in the acid environment of the secretory canaliculus (Fig. 49-15).119 As a consequence of the basic pKa of PPIs and “ion trapping,” the concentra­ tion of PPIs in the secretory canaliculus is 100,000- to 1,000,000-fold higher than in the blood. The sulfenamide rapidly reacts with cysteines on the luminally exposed α-subunit of the H+,K+-ATPase to form a covalent (electro­ chemical) disulfide bond (see Fig. 49-15).120 Whereas all PPIs bind to cysteine 813, omeprazole also binds to cysteine 892, lansoprazole to cysteine 321, and pantoprazole to cys­ teine 822. Because only the apical-membrane inserted H+,K+-ATPase is susceptible to blockade by PPIs and an acid environment (pH < 4) is necessary for trapping and convert­ ing the PPI to its reactive species, the potency of PPIs is decreased when administered during the basal (fasting) state or when acid secretion is inhibited.121,122 Because most pumps are inserted with breakfast, it is recommended that PPIs be taken 30 minutes to 1 hour before the first meal. If greater inhibition is needed, an additional dose should be taken before dinner. Recovery from inhibition of acid secre­ tion occurs by de novo synthesis of pump protein (54 hours in rat). It has been postulated that reduction of the cysteine

INTEGRATED RESPONSE TO A MEAL: INTERPLAY OF NEURAL, PARACRINE, AND HORMONAL PATHWAYS

Stimuli originating inside and outside the stomach converge on gastric intramural efferent neurons that are the primary regulators of acid secretion. The effector neurons comprise cholinergic and peptidergic (i.e., GRP and VIP) neurons. Although nitric oxide and PACAP neurons are present in gastric mucosa their precise physiologic roles as regulators of acid secretion are not known. The effector neurons act on target cells directly as well as indirectly by regulating secretion of gastrin, somatostatin, and possibly histamine (see Figs. 49-8 and 49-10). During the basal state, acid secretion is maintained at an economically low level by the continuous inhibitory restraint exerted by somatostatin on the ECL cell (histamine) and parietal cell (acid) in the fundus/body and on the G cell (gastrin) in the antrum (see Figs. 49-5, 49-9, and 49-10). During ingestion of a meal, maximal secretion may be achieved by removing the inhibitory influence of somato­ statin while at the same time directly stimulating acid and gastrin secretion. This is accomplished, in large part, by activation of intramural cholinergic neurons (see Fig. 49-10). The thought, sight, smell, and taste of food contributes up to 50% of total postprandial acid secretion.124,125 Anticipa­ tion of a meal activates central neurons whose input is relayed via the vagus nerve to gastric intramural cholinergic neurons in oxyntic as well as pyloric mucosa. The compo­ nents of the central nervous system include the dorsal motor nucleus of the vagus, the nucleus tractus solitarius, and the hypothalamus. In the fundus/body, ACh, released from cho­ linergic neurons, stimulates the parietal cell directly, as well as indirectly, by eliminating the inhibitory paracrine influence of somatostatin on parietal and ECL cells.44,126 The resultant increase in histamine stimulates acid secretion directly via H2 receptors on the parietal cell and indirectly via H3 receptors that mediate suppression of somatostatin secretion (see Fig. 49-10).44,127 Thus, histamine, acting via H3 receptors, amplifies the ability of secretagogues to stimu­ late acid secretion by suppressing somatostatin secretion. The net effect of cholinergic neurons is suppression of all paracrine inhibitory influence (i.e., somatostatin) and enhancement of paracrine stimulatory influences (i.e., his­ tamine acting via H2 receptors) on parietal cells. There is some evidence that PACAP, a member of the glucagon/VIP superfamily of regulatory peptides, may participate in the regulation of acid secretion, but its precise physiologic role is not known.26,128,129 PACAP is present in gastric mucosal nerves and capable of releasing histamine from ECL cells and somatostatin from D cells. The net effect of exogenous PACAP on acid secretion has been reported to be either stimulation or inhibition, depending on the relative contri­ butions of released histamine and somatostatin in each preparation.128,130,131 In the antrum, cholinergic neurons activated by antici­ pation of the meal stimulate gastrin secretion directly as well as indirectly by suppressing somatostatin secretion (see Fig. 49-10).30,31,34,132-144 In physiologic concentrations, gastrin stimulates parietal cells directly and, more importantly, indirectly by enhancing the secretion of histamine.145,146 As the meal enters the stomach, the same cholinergic neurons are further activated mechanically by high disten­ tion and chemically by protein components of the

825

826

Section VI  Stomach and Duodenum food.141,142,147,148 Initially, the ingested meal acts as a buffer of secreted acid. The decrease in acidity (i.e., increase in pH) further inhibits somatostatin secretion and thus increases gastrin secretion. Luminal protein activates GRP neurons that directly stimulate gastrin secretion (see Fig. 49-10).79,142 It should be noted that suppression of somatostatin secretion permits an optimal gastrin response. As the meal empties the stomach, a number of paracrine and neural pathways are activated to restore the inhibitory influence of somatostatin in the fundus/body and antrum and hence restrain acid secretion (see Fig. 49-10). First, a stimulatory paracrine pathway linking gastrin to antral somatostatin cells is activated that acts to restore antral somatostatin secretion after release of gastrin.67 Second, there is less activation of cholinergic neurons by anticipa­ tion of the meal as well as by protein and distention. Third, as distention decreases, VIP neurons are preferentially acti­ vated that stimulate somatostatin secretion.141 Fourth, as the buffering capacity of the meal is lost, antral and fundic somatostatin cells (via sensory CGRP neurons) are exposed to the full stimulatory effect of luminal acid. Fifth, entero­ gastrones released from the small intestine, such as CCK, stimulate somatostatin secretion. The resultant increase in antral and fundic somatostatin secretion attenuates gastrin and acid secretion and restores the basal interdigestive state. This state is marked by the continuous restraint exerted on G (gastrin), ECL (histamine), and parietal (acid) cells by contiguous somatostatin cells (see Figs. 49-9 and 49-10). A decrease in this restraint is sufficient to again initiate acid secretion.

HELICOBACTER PYLORI–INDUCED PERTURBATIONS IN ACID SECRETION

(see also Chapter 50) Acute infection with H. pylori (HP) results in hypochlor­ hydria,149-152 whereas chronic infection results in either hypo- or hyperchlorhydria (Fig. 49-16). Appreciation of the pathways discussed earlier provides some insight into the mechanisms whereby HP colonizes the stomach and may lead to ulceration. The decrease in acid secretion during acute HP infection is thought to facilitate survival of the organism and its colo­ nization of the stomach.153 The mechanism whereby HP inhibits acid secretion is multifactorial and includes (1) direct inhibition of the parietal cell (and perhaps ECL cell) by a constituent of the bacterium (e.g., vacuolating cyto­ toxin, lipopolysaccharide, or acid-inhibitory factor) and (2) indirect inhibition of parietal cell function as a result of changes in cytokines as well as hormonal, paracrine, and neural regulatory mechanisms.154-157 HP itself inhibits human H+,K+-ATPase α-subunit gene expression.158 It also elicits secretion of at least two cytokines, IL-1β and tumor necrosis factor-α, that directly inhibit parietal cell secre­ tion.159 In preliminary studies we have shown that HP acti­ vates CGRP sensory neurons coupled to stimulation of somatostatin and thus inhibition of gastrin, histamine, and acid secretion.160 Chronic infection with HP may be associated with either decreased or increased acid secretion depending on the severity and distribution of gastritis (see Fig. 49-16).161 Most patients chronically infected with HP manifest a pangastri­ tis and produce less than normal amounts of acid.162 Reduced acid secretion, initially, is due to functional inhibition of parietal cells by either products of HP itself or, more likely, products of the inflammatory process, as discussed earlier for acute infection163,164; this is usually reversible on eradica­ tion of the organism.165-167 In such patients, HP may be pro­

Acute infection

Antral gastritis

15%

Increase SST

Decrease SST

Decrease gastrin

Increase gastrin

Pangastritis

85%

Decrease acid Increase acid Decrease acid Figure 49-16.  Model illustrating the consequences of Helicobacter pylori (HP) infection on gastric acid secretion. Acute HP infection is associated with an increase in somatostatin (SST) and thus a decrease in gastrin and acid secretion. Chronic infection may be associated with either decreased or increased acid secretion depending on the severity and distribution of gastritis. Most patients chronically infected with HP manifest a pangastritis and exhibit decreased acid secretion. A minority of chronically infected patients manifest an antral-predominant gastritis; these patients, who are predisposed to duodenal ulcer, produce increased amounts of acid as a result of reduced antral SST secretion and a subsequent reciprocal increase in gastrin secretion. (From Schubert ML, Peura DA. Control of gastric acid secretion in health and disease. Gastroenterology 2008; 134:1842-60.)

tective against gastroesophageal reflux disease (GERD), Barrett’s esophagus, and esophageal adenocarcinoma, as well as augment the antisecretory effect of PPIs.168,169 Con­ versely, rebound acid hypersecretion may occur in HPeradicated patients when long-term (i.e., >8 weeks) use of PPIs is discontinued; this may unleash or exacerbate GERD, particularly in patients with large hiatal hernias.170,171 Acid hypersecretion persists at least eight weeks and appears to be due to hypergastrinemia-induced increases in parietal and ECL cell masses.172 The reason the phenomenon does not occur in HP-positive individuals who stop taking PPIs may be due to the fact that HP as well as the cytokines produced by the inflammatory infiltrate inhibit acid secre­ tion and thus mask the rebound. With time, atrophy of oxyntic glands with loss of parietal cells may occur in patients chronically infected with HP, resulting in irrevers­ ible achlorhydria. Autoimmune gastritis is an inflammatory disorder of the oxyntic mucosa often associated with antiparietal cell autoantibodies directed against H+,K+-ATPase with sub­ sequent loss of parietal cells.173 H+,K+-ATPase is a major autoantigen in a subset of patients infected with HP and these antibodies may play a role in the subsequent develop­ ment of atrophic gastritis. It is postulated that antibodies are acquired due to molecular mimicry between HP lipopoly­ saccharide and H+,K+-ATPase, both of which contain Lewis epitopes.174 About 10% to 15% of patients chronically infected with HP have antral-predominant inflammation. These patients, who are predisposed to duodenal ulcer (see Chapters 50, 51, and 52), produce increased amounts of acid as a result of reduced antral somatostatin content and elevated basal and stimulated gastrin secretion (see Fig. 49-16).175-177 Gastrin stimulates histamine secretion from fundic/body ECL cells and induces ECL hyperplasia.178 The mechanism by which somatostatin secretion is decreased is not known but may

Chapter 49  Gastric Secretion involve cytokines induced by the inflammation and/or the production of Na-methyl histamine, a selective H3-receptor agonist, by HP.179,180 One may speculate that the H3-receptor agonist could diffuse across the antral mucosa to interact with H3 receptors on antral somatostatin cells, causing inhibition of somatostatin secretion, and, thus, stimulation of gastrin secretion.45 In addition, IL-8 and platelet activat­ ing factor are upregulated in HP-infected mucosa and are capable of stimulating gastrin release from isolated G cells.181,182

MEASUREMENT OF GASTRIC ACID SECRETION INDICATIONS FOR SECRETORY TESTING

Gastric secretory testing assesses the basal and maximal capacity of the stomach to produce acid. Clinically, its utility has diminished but it may assist in the diagnosis and management of patients with hypergastrinemia (e.g., gastri­ noma) and in the diagnosis of incomplete vagotomy in patients with postoperative recurrent ulcer. Demonstrating fasting acid secretion or an acidic fasting gastric pH excludes achlorhydria as a cause of elevated fasting serum gastrin concentration. Patients with gastrinoma (Zollinger-Ellison syndrome; ZES) demonstrate hypergastrinemia with ele­ vated basal acid output (see Chapter 32).

METHODS FOR MEASURING ACID SECRETION

Aspiration of gastric juice is the most widely used method for measuring acid secretion in humans. Traditionally, this is performed by positioning a nasogastric tube into the most dependent portion of the stomach of a fasted individual. Proper positioning may be verified fluoroscopically or by recovery of more than 90 mL after injection of 100 mL water. Gastric juice is collected by suction. When the tube is pro­perly positioned, only 5% to 10% of gastric juice escapes collection and enters the duodenum. Neutralization by bicarbonate and diffusion of tiny amounts of acid back into the mucosa result in a small underestimation of the true rate of secretion. More recently, an endoscopic technique has been described to measure acid secretion in patients with gastrinoma. In this technique, all gastric contents are aspirated and discarded and then a single 15-minute sample of gastric juice is collected under direct endoscopic visualization.183 The H+ concentration in a sample of gastric juice can be determined by one of two methods. First, the specimen

can be titrated in vitro with a base (e.g., NaOH). The millimoles (mmol) of base needed to titrate a volume of gastric juice to an arbitrary pH endpoint (e.g., 7) represents the “titratable” acidity in mmol per liter of the sample. The other method is to measure the pH of the sample with an electrode. Because pH electrodes measure H+ activity and not concentration, it is necessary to convert activity to concentration using a table of activity coefficients for H+ in gastric juice.184 Once the H+ concentration of the sample in mmol per liter is determined by either of these methods, it is multiplied by the volume of the sample in liters to determine the acid output during the collection period (e.g., mmol per hour or mmol per kilogram of body weight per hour).

BASAL ACID OUTPUT

Basal acid output (BAO) estimates resting acid secretion in the absence of intentional and avoidable stimulation. It is expressed as the sum of the measured acid output, expressed as mmol H+ per hour, for four consecutive 15-minute periods. The upper limit of normal for BAO is about 10 mmol H+ per hour in men and 5 mmol H+ per hour in women (Table 49-1).185 BAO fluctuates from hour to hour in the same person. The lowest BAO occurs between 6 and 11 am and the highest occurs between 2 and 11 pm. Variation is also related to cyclic gastric motor activity with increased BAO in late gastric phase III (migrating motor complex).186

MAXIMAL ACID OUTPUT AND PEAK ACID OUTPUT

Maximal acid output (MAO) and peak acid output (PAO) estimate the acid secretory response to an exogenous secre­ tagogue, usually pentagastrin (6 µg/kg subcutaneous or intramuscular or 6 µg/kg/hr continuous intravenous infu­ sion). Pentagastrin is a manufactured analog of gastrin that contains its biologically active C-terminus sequence. Pos­ sible side effects include flushing, nausea, abdominal pain, dizziness, and palpitations. MAO is the sum of acid output of four consecutive 15-minute collection periods, and PAO is calculated by multiplying by two the sum of the two highest outputs recorded in the four test periods. The expected range for MAO is 5 to 50 mmol H+ per hour and for PAO is 10 to 60 mmol H+ per hour. MAO and PAO are higher in men and in smokers; they correlate with pari­ etal cell mass (i.e., the total number of parietal cells). Typical results for MAO in healthy subjects and in disease are shown in Table 49-1.

Table 49-1  Typical Results of Gastric Secretory Testing in Health and Disease Basal Acid Output (mmol H+/hr)

Normal subjects   Men   Women Duodenal ulcer   Men   Women Gastric ulcer   Men   Women Gastrinoma   Both sexes

Maximal Acid Output (mmol H+/hr)

Average

Range

Average

Range

2.5 1.5

0-10 0-5

25 15

7-50 5-30

5.0 3.0

0-15 0-15

40 30

15-60 10-45

1.5 1.0

0-8 0-5

20 12

5-40 3-25

10-90

65

30-120

40

827

828

Section VI  Stomach and Duodenum SHAM FEEDING–STIMULATED ACID OUTPUT

The cephalic phase of acid secretion whereby the smell, sight, and thought of appetizing food, transmitted via the vagus nerve, stimulates acid secretion can be studied by sham feeding. Sham feeding, in which foods are chewed then spit out, increases acid secretion to about 50% of PAO. Thought and taste appear to play more important roles than sight and smell. Cholinergic and GRP neurons are involved because the response can be abolished by atropine or a selective GRP antagonist.187

MEAL-STIMULATED ACID OUTPUT

Continuous intragastric titration is primarily a research tool used to measure acid secretion in response to food in the stomach.188,189 It measures the cephalic and gastric phases of acid secretion. A double-lumen tube is placed in the most dependent part of the stomach and a homogenized meal buffered to pH 5.5 or 5 is infused into the stomach. Small volumes of gastric contents are sampled from one lumen, the pH is measured, and the contents are returned to the stomach. The second lumen is used to infuse sodium bicar­ bonate to maintain gastric pH at the meal pH. The amount of bicarbonate required to keep the pH of gastric contents constant is a measure of the postprandial acid secretory response. Rates of gastric acid secretion after eating increase rapidly and approach the PAO.

DISEASES ASSOCIATED WITH INCREASED GASTRIC ACID SECRETION Duodenal ulcer patients, as a group, manifest increased basal and stimulated gastrin and acid production (see Table 49-1).190 It is recognized that most cases of duodenal ulcer are due to infection with HP (Chapter 52) and that this infec­ tion is responsible for the perturbations in acid secretion observed in these patients. Pentagastrin-stimulated PAO, an indicator of functional parietal cell mass, is increased in HP-infected duodenal ulcer patients as is GRP-stimulated peak acid output, an indicator of the stomach’s functional response to endogenous gastrin.177,191,192 Suppression of somatostatin secretion by the infection may be the root cause for these changes (see Fig. 49-10; Fig. 49-16). Eradica­ tion of HP restores somatostatin as well as basal and stimu­ lated gastrin and acid secretion, over time, to normal in most individuals, thus providing a permanent cure for duo­ denal ulcer disease.176,177,191,193-195 In contrast to duodenal ulcer, gastric ulcer patients, as a group, exhibit normal or decreased basal and stimulated acid production (see Table 49-1), even though they too are often infected with HP. This suggests that altered gastric mucosal defense may be of primary pathophysiologic importance. Gastric ulcers have been classified according to their location and concomitant association with duodenal ulcer.196 Type I ulcers occur in the gastric body and are generally characterized by low acid secretion. These find­ ings may reflect a greater degree and more generalized mucosal inflammation of the oxyntic mucosa with reduced functional parietal cell mass. Type II ulcers occur in the antrum and are characterized by low, normal, or high acid secretion. Type III ulcers occur within 3 cm of the pylorus, commonly accompany duodenal ulcer, and are characterized by high acid output. Type IV ulcers occur in the gastric cardia and are characterized by low acid secretion.197 Accordingly, the more distant a gastric ulcer is from the pylorus the more likely acid secretion will be low.

A number of uncommon conditions are marked by gastric acid hypersecretion and subsequent ulceration (see Chapter 52). In patients with systemic mastocytosis, high histamine levels, as a consequence of increased numbers of mast cells, continuously stimulate parietal cells to secrete acid.198 When a portion of gastric antrum is retained in the afferent remnant after antrectomy with Billroth II anastomosis, it is bathed in alkaline secretions leading to decreased soma­ tostatin secretion, hypergastrinemia, increased acid produc­ tion, and anastomotic ulceration.65,85,199 Acid hypersecretion also can result from chronic hypercalcemia of any cause because calcium directly stimulates gastrin secretion from G cells and acid secretion from parietal cells.200,201 The best characterized acid hypersecretory condition is ZES,202-204 as discussed in Chapter 32. The BAO is almost always higher than 15 mmol/hr and the BAO/PAO ratio is usually 0.6 or greater (see Table 49-1). Gastrin, synthesized by the tumor, is secreted into the bloodstream, where it binds to CCK-2 receptors on acid-producing parietal and histamine-containing ECL cells to induce secretion as well as proliferation. The clinical correlate of the proliferation is rugal hypertrophy with prominent gastric folds. Diagnosis and treatment of gastrinoma are discussed in detail in Chapter 32. The basis of the secretin test to diag­ nose gastrinoma is that normally somatostatin cells in the antrum tonically restrain gastrin secretion from G cells. Secretin stimulates the G cell directly and, at the same time, inhibits the G cell indirectly by stimulating somatostatin secretion; the effect of the latter usually dominates and gastrin is not stimulated to a significant degree (Fig. 49-17).

+

Gastrin +

CCK-2 ECL cell (histamine) +

Secretin + + SST −

CCK-2 Parietal H2 cell

H+ Fundus and body (oxyntic mucosa)

G

Antrum (pyloric mucosa) Figure 49-17.  Zollinger-Ellison syndrome (ZES). Model illustrating the action of gastrin in oxyntic mucosa and secretin in pyloric mucosa of stomach in patients with ZES. Gastrin, synthesized and secreted by the gastrinoma into the bloodstream, acts via cholecystokinin-2 (CCK-2) receptors on acid-secreting parietal and histamine-secreting enterochromaffinlike (ECL) cells to increase acid secretion and induce cell proliferation. In the antrum, exogenous secretin (i.e., secretin stimulation test [see text and Chapter 32]) stimulates gastrin secretion directly and concomitantly inhibits gastrin secretion by stimulating somatostatin (SST) secretion resulting in little or no gastrin release. Because the gastrinoma does not contain functionally coupled SST cells, the effect of secretin in ZES patients is solely to stimulate gastrin secretion from the tumor. (From Hung PD, Schubert ML, Mihas AA. Zollinger-Ellison syndrome. Curr Treat Options Gastroenterol 2003; 6:163-70.)

Chapter 49  Gastric Secretion Because the gastrinoma does not contain functionally coupled somatostatin cells, the effect of secretin is solely stimulation of gastrin secretion from the tumor.205-207 Almost all gastrinomas contain somatostatin receptors and soma­ tostatin receptor scintigraphy (SRS) using [111In-DPTADphe1]-octreotide is considered the initial localization study of choice, with a 71% sensitivity and 86% specificity for primary tumors and 92% detection for metastatic disease.208,209 PPIs are the antisecretory therapies of choice and are able to control acid secretion and prevent complica­ tions in most patients with ZES.210

PEPSINOGEN SECRETION Pepsinogens, which belong to a family of enzymes called gastric aspartic proteases, are inactive polypeptide proen­ zymes known as zymogens. They are synthesized primarily in chief cells but also in mucous neck cells. Pepsinogens are converted in the gastric lumen by gastric acid to pepsins, which contain two active-site aspartate residues. Once this reaction begins, pepsins can autocatalyze the conversion of pepsinogens to pepsins.211 Pepsins are optimally active at pH 1.8 to 3.5, reversibly inactivated at pH 5, and irreversibly denatured at pH 7. Gastric acid not only provides an optimum pH for peptic activity but itself denatures dietary protein, making it more susceptible to peptic hydrolysis. Thus, acid and pepsin work in concert to promote digestion of dietary protein. As discussed, partially digested protein stimulates gastrin and thus acid secretion.142 More recent data suggest that pepsins may also be important for killing ingested bacteria.17,212 Pepsinogens have been electrophoretically separated into seven isozymogens. The five fractions (pepsinogen 1 to pep­ sinogen 5) that migrate toward the anode most rapidly at pH 5 are similar immunologically and are referred to as group I pepsinogens (PGI; old term, pepsinogen A).213 PGI is expressed in chief and mucous cells of the oxyntic mucosa. Migrating slightly behind the PGIs are two immu­ nologically similar isozymogens, pepsinogen 6 and pep­ sinogen 7, that are referred to as group II pepsinogens (PGII; old term, pepsinogen C). PGII, which represents approxi­ mately 20% of total pepsin content, is expressed in oxyntic and pyloric mucosa as well as in duodenal Brunner’s glands. The most important physiologic stimulant for pepsinogen secretion is ACh released from intramural cholinergic neurons. ACh, acting via both M1 and M3 muscarinic recep­ tors on chief cells, induces an increase in cytosolic calcium.214 Calcium, in turn, activates cytosolic kinases, phosphatases, and nitric oxide synthase that induce pep­ sinogen secretion.215 Other agents capable of stimulating pepsinogen secretion from chief cells via the calcium signal­ ing pathway include CCK, gastrin, and GRP.216-219 Agents that increase cAMP within chief cells, such as isoprotere­ nol, secretin, VIP, and histamine also augment pepsinogen secretion as do agents that activate tyrosine kinase such as epidermal growth factor and transforming growth factor-α.218 Inhibitors of pepsinogen secretion include somatostatin, neuropeptide Y, PYY, and IL-1β. Optimal pepsinogen secre­ tion requires a functional sodium-2 chloride potassium cotransporter-1 (NKCC1), a basolateral cotransporter respon­ sible for chloride uptake into secretory epithelia,118 includ­ ing the chief cell (and the parietal cell; see Fig. 49-12). It has been postulated that loss of the flushing action of neutral gastric secretion may impair pepsinogen secretion. Serum levels of PGI correlate with maximal acid output. A linear correlation exists between the loss of chief cells in

patients with oxyntic atrophy and serum PGI220,221; a serum PGI/PGII ratio of 2.5 or less has been used as a noninvasive test to detect gastric mucosal atrophy.5,222 Serum PGI is also increased in humans treated with PPIs, but the precise mechanism is not known.223 Both PGI and PGII are filtered and metabolized by the kidney, but serum PGI concentra­ tion is increased more than PGII concentration in patients with renal insufficiency.224,225

GASTRIC LIPASE SECRETION Gastric lipase, secreted by chief cells of the oxyntic mucosa, helps initiate the digestion of dietary triglycerides by hydro­ lyzing them to free fatty acids, diglycerides, and 2monoglycerides. The properties of gastric lipase are quite distinct from those of pancreatic lipase. Gastric lipase has a pH optimum of 4.5 to 5.5 (compared with 6.5 to 7.4 for pan­ creatic lipase) and does not require colipase. Furthermore, protection from peptic proteolysis by an N-glycosylated asparagine at residue 308 permits gastric lipase to retains its full activity in acidic gastric juice (pH 2) despite a high gastric juice peptic activity.226,227 Stimulants and inhibitors of gastric lipase are similar to those for pepsinogen secretion. Aging has been reported to decrease gastric lipase secretion, but data are controver­ sial.228 In humans, increasing the amount of lipid in the diet causes a corresponding increase in gastric lipase secretion into gastric juice.229 The amount of gastric lipase secreted after a meal is small relative to the amount of pancreatic lipase. However, the specific activity of gastric lipase is equal to or greater than that of pancreatic lipase. Thus, gastric lipase is capable of digesting 10% to 25% of dietary triglyceride.230 In patients with chronic pancreatitis, gastric lipase secretion is increased three- to four-fold and can partly, but incompletely, compensate for loss of pancreatic lipase (Fig. 49-18).231 A feedback mechanism exists whereby fat in the small intestine inhibits gastric lipase secretion by a humoral mechanism, with GLP-1 a prime candidate to be the mediator.232 Orlistat (tetrahydrolipstatin) is a lipophilic compound derived from lipstatin produced by Streptomyces toxitricini and is a potent inhibitor of all lipases.233 Orlistat reacts covalently with the catalytic serine residue of lipases.234 Orlistat decreases lipid digestion by about 35% and has been used clinically to induce modest weight loss.235 Adverse gastrointestinal events include loss of lipid-soluble vitamins, accelerated gastric emptying, fecal urgency, abdominal discomfort, flatus with discharge, and fecal urgency (see Chapter 6).236,237

INTRINSIC FACTOR SECRETION Intrinsic factor (IF), a 50-kd glycoprotein secreted by human parietal cells and, to a lesser degree, chief cells, is necessary for the absorption of cobalamin (vitamin B12).238 Although all stimulants (e.g., gastrin, histamine, and ace­ tylcholine) and inhibitors (e.g., somatostatin) of gastric acid secretion discussed previously have similar effects on IF secretion, the secretion of IF is not coupled to acid secre­ tion. PPIs, for example, inhibit acid but have no significant effect on basal or stimulated IF secretion.239 Several reports indicate, however, that chronic use of PPIs may result in low serum levels of cobalamin, probably as a result of impaired acid-facilitated release of cobalamin from

829

Section VI  Stomach and Duodenum 365% of mean control values

2% of mean control values

food.240,241 The recommended daily allowance of cobalamin is 2 µg/day and total body stores are 2.5 mg. Cobalamin deficiency due to antisecretory therapy is therefore rare because acid secretion is not completely suppressed and because the body contains relatively large stores.10,242,243 The delivery of cobalamin from food to tissues begins with release of cobalamin from dietary protein by the pHdependent activity of pepsin, followed by binding of cobala­ min to two binding proteins that are secreted into gastric juice: IF and haptocorrin (R binder).244,245 Haptocorrin is also secreted in saliva and bile. Haptocorrin binds cobalamin more avidly than IF in the acidic stomach and, therefore, most cobalamin initially becomes attached to haptocorrin. In the duodenum, cobalamin is released from haptocorrin by pancreatic trypsin and the free cobalamin then binds to IF. IF-cobalamin complexes are resistant to pancreatic pro­ teolysis and eventually bind to a specific receptor on the distal ileal mucosa. This receptor, cubilin, is expressed in clefts between microvilli and mediates endocytosis of the IF-cobalamin complex.246 Once within the ileal enterocyte, IF is degraded by lysosomal enzymes and cobalamin binds to transcobalamin. The cobalamin-transcobalamin complex is released into the circulation and from there enters cells by receptor-mediated endocytosis. Once within cells, cobalamin is dissociated from its transport protein and converted to its active forms, methylcobalamin and 5-deoxyadenosyl cobalamin. The active forms serve as coenzymes for methionine synthase and methylmalonylcoenzyme A mutase, enzymes involved in methylation of homocysteine to methionine and the catabolism of branchedchain amino acids and odd-chain fatty acids in mitochon­ dria, respectively.247 When radiolabeled cobalamin is administered orally after a large dose of nonradioactive cobalamin is given parenter­ ally, patients with IF deficiency excrete much lower amounts of radioactive cobalamin in a 24-hour urine collection than do normal controls (Schilling test, part I). If IF is administered orally together with radioactive cobalamin to IF-deficient patients, urinary radioactive cobalamin excre­ tion normalizes (Schilling test, part II). In addition to IF

50 40 30 20 10 0

80 Lipolytic activity (%) in duodenal contents

60

Human pancreatic lipase outputs (mg)

Figure 49-18.  Relative outputs of human gastric lipase (left graph) human pancreatic lipase (middle graph), and lipolytic activity in duodenal contents (right graph) in healthy volunteers (violet columns) and patients with chronic pancreatitis with exocrine insufficiency (CP; beige columns). Note that despite the increase of human gastric lipase output in patients with CP, the overall magnitude of the rise is insufficient to fully correct the reduction in duodenal lipolytic activity. (From Carrière F, Grandval P, Renou C, et al. Quantitative study of digestive enzyme secretion and gastrointestinal lipolysis in chronic pancreatitis. Clin Gastroenterol Hepatol 2005; 3:28-38.)

70

31% of mean control values

600

80 Human gastric lipase outputs (mg)

830

500 400 300 200 100

70 60 50 40 30 20 10 0

0 Healthy volunteers Patients with CP

deficiency, cobalamin malabsorption may result from achlorhydria or hypochlorhydria (reduced peptic hydroly­ sis of cobalamin from food protein), bacterial overgrowth (cobalamin competed for by bacteria), pancreatic insuffi­ ciency (impaired tryptic cleavage of haptocorrin-cobalamin complex), ileal receptor defect (cubilin mutation), and ileitis or ileal resection (absent IF-cobalamin absorptive site).248-250 Secretion of IF far exceeds the amount necessary for cobalamin absorption. Thus, in most patients with hypo­ chlorhydria, continued IF secretion in low amounts is suf­ ficient to prevent cobalamin deficiency. Patients with pernicious anemia, however, which affects 2% of indivi­ duals older than the age of 60, do develop cobalamin defi­ ciency.251 The pathology involves a chronic inflammatory, mainly lymphocytic, infiltrate of the oxyntic mucosa accom­ panied by loss of parietal and chief cells. The pathogenesis involves proinflammatory TH1 CD4 T cells directed toward the α- and β-subunits of the parietal cell H+,K+-ATPase, as well as circulating antibodies directed against H+,K+-ATPase or IF.252 Chronic infection with HP may play a primary role in the immunologic response. It is proposed that HP-induced inflammation results in breakdown of tolerance for self antigens such as H+,K+-ATPase in genetically susceptible individuals or that antibodies are acquired due to molecular mimicry between HP lipopolysaccharide and H+,K+-ATPase, both of which contain Lewis epitopes.173,174 Antibodies directed against H+,K+-ATPase are found in 90% of patients with pernicious anemia but the incidence of these anti­ bodies may decrease to about 55% to 80% with progression of autoimmune gastritis presumably because of disappear­ ance of HP and the loss of antigenic drive.252-254

BICARBONATE SECRETION Regulation of gastric bicarbonate (HCO3−) secretion has been studied much less intensively than has duodenal HCO3− secretion. The precise function of gastric HCO3− secretion is

Chapter 49  Gastric Secretion uncertain given the overwhelming simultaneous secretion of H+. Nevertheless, HCO3− secretion has been implicated in the formation of a protective pre-epithelial alkaline layer (see Fig. 49-2). Measurement of gastric HCO3− secretion has been impeded by the presence of considerable H+ secretion, necessitating the use of potent antisecretory compounds or measurement methods unaffected by the presence of acid. One means of measuring bulk HCO3− secretion is with the use of inline pH and CO2 electrodes, in which HCO3− concentration is calcu­ lated using the Henderson-Hasselbach equation.255,256 An electrophysiologic method has been used to measure alka­ line secretion in the lumen of individual gastric glands isolated from the frog Rana esculenta.257 Gastric HCO3− secretion is an energy-dependent process. The finding that there is virtually no change in gastric electrical potential difference during HCO3− secretion sug­ gests that HCO3− transport takes place via an electroneutral ion exchange mechanism, probably an exchange of HCO3− for Cl− at the luminal surface.258 Although several candidate anion exchangers, solute carriers, and anion transporters have been localized to gastric surface cells including AE2, AE4, Slc4a2, Slc4a4, PAT1, and Slc26a6, there is little evidence that gastric surface cells actually secrete HCO3−.259-263 The cell responsible for HCO3− secretion is not known. The source of some of the HCO3− secreted during H+ secretion may actually be the parietal cell. As discussed previously, for each H+ secreted, a HCO3− ion exits the parietal cell across the basolateral membrane via the anion exchanger 2 (AE2), Slc4a2 (see Fig. 49-12).115 This HCO3− may alkalinize the blood that perfuses surface epithelial cells, be taken up by sodium bicarbonate cotransporters (NBC1 and NBC2), and then be secreted by epithelial cells in an effort to protect them from luminal acid.264 The pari­ etal cell, however, may not be the only source of HCO3− for surface cells, as marked inhibition of gastric H+ secretion by PPIs does not significantly diminish gastric HCO3− secre­ tion in patients with duodenal ulcer.265 Prostaglandin E2 analogs stimulate gastric HCO3− and mucus secretion266,267 and blockade of endogenous prosta­ glandin synthesis reduces gastric HCO3− secretion.268 In rats and mice, the prostaglandin E receptor subtype involved in stimulation of gastric HCO3− secretion is EP1.269,270 Gastric mucosal prostaglandin synthesis and HCO3− secretion decline in older adults.271,272

MUCUS SECRETION A firmly adherent viscous mucus gel overlies the gastric surface. It is composed of 95% water and 5% extensively cross-linked mucin glycoproteins that are products of MUC genes.273,274 Ultrastructural studies reveal alternating layers of two distinct mucin classes, MUC5AC (secreted by surface and pit area epithelial cells) and MUC6 (secreted from neck and gland cells).275-278 Observations that the gastroprotective compound, geranylgeranylacetone (GGA), increases MUC6 expression in rat gastric mucosa279 and that lafutidine, a gastroprotective compound used clinically in Japan, increases mucus thickness and mucin content in humans280 suggest that the mucus gel may contribute to mucosal defense.273,281,282 Furthermore, electron microscopy studies demonstrate that HP accumulates within and disrupts the MUC5AC-enriched gel layer.277,283,284 MUC5AC secretion is increased in first-degree relatives of subjects with gastric cancer infected with HP, suggesting that mucus secretion may be a marker for a more severe inflammatory response to the organism.285

Mucus gel thickness can be measured continuously and noninvasively in living rodents by alternately focusing between fluorescently labeled surface epithelial cells and the mucus gel surface, as delineated by carbon particles or by fluorescent microspheres.286-288 A pH gradient at the gastric mucosal surface has been observed in a variety of species including humans.289-292 In most cases, the gradient is relatively alkaline at the tissue surface and gradually more acidic at distances further from the surface. The gradient is due to active bicarbonate secre­ tion and is considered a defense against luminal acid.290,293 Although most measurements using microelectrodes have reported values near neutrality at the gastric mucosal surface, more recent measurements using pH-sensitive fluo­ rescent dyes with ex vivo confocal microscopy report a surface pH near pH 4 in guinea pigs and frogs.294,295 The observation that the steady-state surface pH gradient extends beyond the thickness of the mucus gel layer suggests that the unstirred layer formed at the interface between the mucus and the aqueous lumen or the interface between the epithelial surface and the luminal contents may also play a role in mucosal defense by restricting mixing of molecules within the unstirred layer.283,293,296 Although it has been suggested that gastric mucus may physically retard diffusion of protons, the evidence for this is equivocal. In vitro measurements of proton and bicar­ bonate diffusion have yielded diffusion coefficients that are up to 10 times slower in isolated mucus compared with saline solution (≈0.5 − 2 × 10−5 cm2/sec).287,297 However, these diffusion coefficients in mucus are comparable to those found in saline for ions such as Na+, K+, and Cl−. Even accepting the slowest reported diffusion of protons, one-third of protons will theoretically diffuse 80  µm or greater within 10 seconds, a distance greater than the average mucus thickness in rats and mice.281,290,296 There are also concerns regarding diffusion measurements per­ formed in vitro as nondestructive removal of the tightly adherent mucus is extremely difficult, mucus structure and function is sensitive to environmental conditions, and measurement instruments may produce artifactual unstirred layers.298,299 Trefoil factor (TFF) peptides are cosecreted with mucins.300 These are 7- to 12-kd peptides sharing a common structure of three internal disulfide bonds that yield the signature “trefoil” structure of three internal loops and are designated TFF1, TFF2, and TFF3. TFFs are markedly pepsin resistant and able to survive intact in the gastric lumen301; the con­ centration of TFF2 in rat gastric mucus has been estimated at 10 µm.302,303 The distribution of each trefoil peptide in the normal GI tract is distinct. TFF1 is stored in gastric pit and surface mucous cells, TFF2 is present in gastric gland mucous cells, and TFF3 is stored in intestinal goblet cells.303-307 The relative abundance of TFF1 and TFF2 expression is reciprocally regulated by gastrin, H+,K+ATPase, and inflammation.308 The localization and coordinated secretion of trefoil peptides with mucins suggest that they too may be involved in mucosal defense.305,309,310 In support of this notion (1) increased TFF1 expression is observed after administration of gastroprotective agents311; (2) addition of TFF2 to mucin solutions significantly increases viscosity and elasticity in vitro and in vivo312-313; (3) chronic treatment with PPIs increases TFF2 concentration in gastric secretions coincident with promoting repair and preventing injury in response to luminal noxious agents314; and (4) TFF2 -/- knockout mice exhibit shortened gastric glands with decreased epithelial migration, increased net acid secretion, and a four-fold increase in number

831

832

Section VI  Stomach and Duodenum of lesions after a 12 hour-exposure to the nonselective cyclooxygenase (COX) inhi­bitor indomethacin.315

Acknowledgment

The authors thank Mary Beatty-Brooks for the artwork and Mark Feldman, John Walsh, Andrew Soll, and Gabriel Makhlouf, pioneers in the field of gastric exocrine and endocrine secretion, for their encouragement, support, and friendship.

KEY REFERENCES

Carrière F, Grandval P, Renou C, et al. Quantitative study of digestive enzyme secretion and gastrointestinal lipolysis in chronic pancrea­ titis. Clin Gastroenterol Hepatol 2005; 3:28-38. (Ref 231.) Flemström G, Isenberg JI. Gastroduodenal mucosal alkaline secretion and mucosal protection. News Physiol Sci 2001; 16:23-8. (Ref 258.) Fossmark R, Johnsen G, Johanessen E, Waldum HL. Rebound hyperse­ cretion after long-term inhibition of gastric acid secretion. Aliment Pharmacol Therapeut 2005; 21:149-54. (Ref 172.) Gillen D, Wirz AA, Neithercut WD, et al. Helicobacter pylori infection potentiates the inhibition of gastric acid secretion by omeprazole. Gut 1999; 44:468-75. (Ref 169.) Hatlebakk JG, Katz PO, Camacho-Lobato L, Castell DO. Proton pump inhibitors: Better acid suppression when taken before a meal than without a meal. Aliment Pharmacol Therapeut 2000; 14:1267-72. (Ref 122.)

Heitzmann D, Warth R. No potassium, no acid: K+ channels and gastric acid secretion. Physiology 2007; 22:335-41. (Ref 108.) Jain RN, Samuelson LC. Differentiation of the gastric mucosa: Role of gastrin in gastric epithelial cell proliferation and maturation. Am J Physiol Gastrointest Liver Physiol 2006; 291:G762-5. (Ref 68.) Prinz C, Zanner R, Gratzl M. Physiology of gastric enterochromaffin-like cells. Ann Rev Physiol 2003; 65:371-82. (Ref 46.) Manela FD, Ren J, Gao J, et al. Calcitonin gene-related peptide modu­ lates acid-mediated regulation of somatostatin and gastrin release from rat antrum. Gastroenterology 1995; 109:701-6. (Ref 65.) Moss SF, Legon S, Bishop AE, et al. Effect of Helicobacter pylori on gastric somatostatin in duodenal ulcer disease. Lancet 1992; 340:9302. (Ref 176.) Schubert ML, Edwards NF, Makhlouf GM. Regulation of gastric soma­ tostatin secretion in the mouse by luminal acid: A local feedback mechanism. Gastroenterology 1988; 94:317-22. (Ref 85.) Wang TC, Dockray GJ. Lessons from genetically engineered animal models I. Physiological studies with gastrin in transgenic mice. Am J Physiol Gastrointest Liver Physiol 1999; 277:G6-11. (Ref 60.) Whited KL, Thao D, Lloyd KCK, et al. Targeted disruption of the murine CCK1 receptor gene reduces intestinal lipid-induced feedback inhibi­ tion of gastric function. Am J Physiol Gastrointest Liver Physiol 2006; 291:G156-62. (Ref 104.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

50 Helicobacter pylori David A. Peura and Sheila E. Crowe

CHAPTER OUTLINE Epidemiology  833 Pathogenesis  834 Colonization and Virulence Factors  834 Host Response to Infection  835

Helicobacter pylori are unique bacteria ideally suited to live in the acidic environment of the human stomach. Their spiral shape and multiple unipolar flagella allow them to move freely through the gastric mucous layer, where they remain protected from low gastric pH.1 Organisms produce large amounts of urease, an enzyme that hydrolyzes urea to alkaline ammonia and CO2. This permits the bacteria to further control the pH of their microenvironment. Urease is also the basis of clinical diagnostic tests (urea breath test and rapid urea biopsy tests) for infection. H. pylori remain difficult and tedious to culture because they grow slowly and require specialized culture media and a controlled microaerophilic environment. When they gain access to a human host, H. pylori recognize and attach to various gastric epithelial surface receptors, thereby chronically colonizing the mucosa, disrupting cell function, inciting an intense local inflammatory and systemic immune response, and altering acid secretory physiology.2,3 The ultimate clinical manifestations of H. pylori infection include gastric and duodenal ulcer, gastric mucosa–associated lymphoid tissue (MALT) lymphoma, and adenocarcinoma; yet most infected individuals remain asymptomatic for life despite developing chronic histologic gastritis.1,2,4,5 What factors determine why some develop disease in response to infection and others do not remains a mystery, but host genetics, bacterial characteristics, and environmental features undoubtedly can influence clinical outcome.5,6 Research focusing on epidemiology, pathogenesis, management, and prevention of H. pylori infection and associated clinical conditions continues to be fueled by the tremendous worldwide prevalence of infection (especially in less-developed countries), the huge health and economic burden imposed by ulcer disease and gastric cancer, and heightened prominence and awareness of the bacteria accompanying Drs. Robin Warren and Barry Marshall’s receipt of the 2005 Nobel Prize in Physiology or Medicine for seminal contributions to the field.7

EPIDEMIOLOGY Helicobacter pylori infection remains one of the most common chronic bacterial infections in humans. Estimates suggest that more than 50% of the world’s population is infected with the bacterium and genetic sequence analysis proposes that humans have been infected for

Conditions Arising from Infection  838 Diagnosis  839 Treatment  841

more than 58,000 years at a time when they first migrated from Africa.8 While H. pylori have been demonstrated worldwide in individuals of all ages, infection is more common and acquired at an earlier age in developing countries compared with industrialized nations.9,10 In developing nations, the majority of children become infected before the age of 10, and during early childhood spontaneous elimination of bacteria and subsequent reinfection is quite common. Infection persists in older children and adults so that in the developing areas of the world H. pylori prevalence can reach more than 80% by age 50. In developed countries, such as the United States, young children can also acquire H. pylori, but usually before age 5.11 Spontaneous clearance often occurs and there is less chance of reinfection; thus, persistent childhood infection is much less frequently seen than in less-developed countries.9 In fact, serologic evidence of H. pylori is uncommon in children before age 10, but rises to 10% in adults between 18 and 30 years of age and further increases to 50% in those 60 or older.9 This increased prevalence of infection with age was initially thought to represent continuing acquisition throughout adult life. However, new adult infection and reinfection are quite uncommon, especially in developed countries. Epidemiologic evidence supports childhood-acquired infection even in developed nations, so the frequency of H. pylori infection for any age group in any locality reflects that particular birth cohort’s rate of bacterial acquisition early in life.9 In the United States, within any age group, infection appears to be more common in blacks than whites.12 Also Hispanic immigrants and their first-generation children are more likely to harbor H. pylori than their second-generation relatives.13 These differences probably relate to factors early in life that are linked to acquiring infection. The risk of acquiring H. pylori is associated with living conditions and the family’s socioeconomic status during one’s childhood.9 Housing density, crowded conditions in the home, number of siblings, sharing a bed, and lack of hot or running water have been linked to higher rates of infection. In Japan, the rapidly declining prevalence of H. pylori appears to parallel the nation’s postwar economic progress and improvement in hygiene and sanitation. Of the Japanese born before 1950, more than 70% are infected compared with 45% born between 1950 and 1960 and 25% born between 1960 and 1970.14 Presently, childhood infection in Japan is rare. Predictions based on a similar declining prevalence in the United States suggest that the organism could

833

834

Section VI  Stomach and Duodenum eventually become extinct here and in other areas of the world, which would affect infection-related illness.15 Twin studies support genetic susceptibility to H. pylori infection because monozygotic twins who were raised in different households have a greater concordance of infection than dizygotic twins also raised separately.16 However, twins growing up together have a higher concordance of H. pylori status than twins growing up separately, suggesting childhood environmental factors are also important for acquisition. Humans appear to be the major reservoir of H. pylori. However, domestic cats and captive primates and sheep can harbor these organisms,10 although it is possible that these animals actually acquired H. pylori from a human source. In the case of cats, isolation of viable bacteria from saliva and gastric juice suggests at least the possibility of transmission to humans.10 Especially in developing countries, contaminated water might serve as an environmental source of bacteria because the organism can remain viable for several days in water.17 Bacterial deoxyribonucleic acid (DNA) can be found in samples of municipal water from endemic areas of infection but whether viable H. pylori are present remains to be proven.5 In countries where infection is common, children who drink untreated stream water, eat uncooked vegetables, or swim in rivers and streams are more likely to harbor the bacteria, providing further indirect evidence of an environmental source of organisms. Person-to-person transmission of bacteria from fecal-oral, oral-oral, or gastro-oral exposure seems the most probable explanation for infection.5,10 Within-family clustering of infection (often with genetically identical strains of H. pylori) supports person-to-person transmission.9 Also infected individuals more often have infected spouses or children than uninfected individuals. Support for siblingto-sibling transmission comes from studies reporting that likelihood of infection correlated with number of children in the household and that younger children were more apt to be infected if older siblings were also infected.9 Motherto-child transmission is also quite likely.11,18 Fecal-oral transmission of bacteria is a possible mechanism by which H. pylori gain access to the human host. The bacterium can be cultured from diarrheal stools and vomitus, suggesting the potential for transmission.19 Exposure to an infected family member during an acute gastrointestinal illness, especially with vomiting, appears to be a risk factor for subsequent infection.20 How frequently bacteria are transmitted through oral-oral contact is not known. Although organisms can be identified in dental plaque and saliva, the prevalence is low and it is questionable if the mouth serves as a source or reservoir for H. pylori.21 Also dentists and oral hygienists who continually have occupational exposure to dental plaque and oral secretions do not have a higher prevalence of H. pylori.22 In developed countries, spousal transmission of infection also appears to be uncommon. Infected gastric secretions can serve as a source of bacterial transmission. Iatrogenic infection has occurred during the use of a variety of inadequately disinfected gastric devices, endoscopes, and endoscopic accessories.10 Also gastroenterologists and nurses appear to be at greater risk for acquiring H. pylori, presumably due to occupational contact with infected gastric secretions.23 Mandated universal precautions, standardized equipment disinfection, and use of video-endoscopes that reposition the instrument channel away from the mouth should reduce such iatrogenic and occupational transmission. Natural transmission could occur through contact with infected vomitus during

an acute illness20 or with regurgitated material from an infected child. Such contact could explain the higher concordance of maternal/child H. pylori infection and the presumed child-to-child transmission that occurs in an infant daycare setting.24

PATHOGENESIS Specific genetic or phenotypic factors in infectious agents have been implicated as single causal factors in a variety of infectious diseases and associated outbreaks. However, H. pylori infection alone appears insufficient to fully explain the spectrum of diseases that is associated with chronic infection. Research over the past quarter century suggests that the pathogenicity of H. pylori depends on bacterial and host factors in addition to less well-defined environmental conditions. Virulence of this infectious pathogen is based on bacterial properties that allow colonization and adaptation to the gastric environment and a host response that contributes to the host physiologic and histologic changes.

COLONIZATION AND VIRULENCE FACTORS One of the interesting aspects of this pathogen is how it confers disease when the organism resides, for the most part, in the lumen. Studies describing the genome of two distinct strains of H. pylori have helped to advance our understanding of the ecology of the organism and the potential gene expression patterns that can affect the pathogenesis of disease.25,26 Importantly, bacterial genes expressed in gastric mucosa differ from the pattern observed in vitro,27 whereas exposure of the bacterium to low pH increases its expression of genes encoding proteins involved in the motility apparatus and of genes encoding urease.28 The urease helps the organism adapt to the gastric milieu as it retains optimal function at two different pH values: usually pH 7.2 and pH 3.29 H. pylori is further adapted to the gastric pH by producing the molecular machinery required to migrate rapidly to a more favorable environment below the mucus layer. Motility is one of the few H. pylori characteristics shown to be necessary for successful colonization of the host. H. pylori show a strict tropism for the gastric mucosa or intestinal sites in which there is gastric metaplasia. The corollary is also true, as H. pylori do not colonize epithelium in the stomach that has undergone intestinal metaplastic change, possibly due to the production of antimicrobial factors that select against colonization. This possibility is supported by the fact that H. pylori rarely colonize the deeper portions of the gastric glandular mucosa, where O-glycans that impair H. pylori growth are found.30 This concept is also supported by another study that shows H. pylori decreases the expression of the antibacterial molecule, secretory leukocyte protease inhibitor,31 thereby removing an element of the host response that would be detrimental to the persistent infection. Another important factor that controls colonization is the expression of receptors on host cells that allow H. pylori to bind. Lewis (Le) antigens expressed by host cells serve as a receptor for bacterial binding.32 Specific bacterial gene pro­ ducts, most notably BabA, act as the bacterial ligand for the Leb receptor.33 The Leb receptor may be located in the gastric mucus because strains bearing BabA also bind to the mucin MUC5AC.34 Some studies suggest that the babA2

Chapter 50  Helicobacter pylori genotype is more frequently associated with inflammation, duodenal ulcer and gastric cancer.2 However, observations that binding of H. pylori to epithelial cells freshly isolated from human gastric biopsy specimens is unaffected by the expression of Le antigen, that infection is not increased in subjects with the Leb phenotype, and that individuals who do not express Leb can clearly be infected with H. pylori cast doubt on the true role of Leb and BabA.2 Additionally, in one report, the majority of strains infecting the individuals studied did not induce ulcers or cancer, despite expressing the babA2 gene.35 H. pylori also bind to the molecular complex of invariant chain and class II human leukocyte antigens (HLAs) expressed on the surface of gastric epithelial cells.36 Class II major histocompatibility complex (MHC) molecules, with their expression increased by infection, were the first epithelial cell receptor for H. pylori demonstrated to directly affect signaling in host cells. Binding of urease to epithelial cells via class II MHC was sufficient to induce apoptosis.37 More recently, the gastric trefoil protein TFF1 was shown to serve as a receptor for H. pylori.38 This molecule is predominantly expressed in the gastric mucosa and found in association with gastric mucus. In genetically engineered mice with TFF1 deficiency, a spontaneous, antral adenoma develops, suggesting that this molecule provides an element of control over gastric epithelial cell growth.39,40 The pathogen-associated molecular receptors (PAMPS) have also been examined for their role in binding of H. pylori to the host epithelial cells. The Toll-like receptors (TLRs) are a family of PAMPS, with an apparent different specificity for various bacterial molecules.41 For example, TLR4 is able to recognize the LPS of many bacteria, with cytokines and H. pylori particles increasing the expression of TLR4.42 H. pylori LPS stimulates monocytes and gastric epithelial cell responses via TLR4.2 Other PAMPs, including TLR2, are also activated by highly purified H. pylori LPS and have even been described as being more important than TLR4.43,44 These receptors may bind bacterial products, and thereby enhance bacterial binding and cell signaling.42 TLR5 binds bacterial flagellins and, similar to TLR2, induces a signaling response that can trigger acute inflammation. H. pylori produce flagellin that binds TLR5 and activates a response in vitro.43 There is disagreement on which TLR is stimulated by H. pylori LPS or whether H. pylori flagellin is ever able to bind TLR5,45 and this may reflect the different conditions used for bacterial culture and LPS or flagellin purification. Further studies are needed to establish the significance of H. pylori binding to TLRs in the pathogenesis of infection. After H. pylori migrate to the gastric epithelium, the organism attaches to host cells and may damage them in order to obtain nutrients from the subsequent inflammatory exudate or transudate. A key interaction between the bacteria and gastric epithelium involves a segment of bacterial DNA referred to as the cag pathogenicity island (cag PAI). Genes within the cag PAI encode proteins that provide a type IV secretion apparatus (i.e., cagE) that allows bacterial macromolecules to translocate into the host cell (i.e., cagA).26,46 cag PAI plays an important role in the pathogenesis of gastritis in humans26,47 because H. pylori bearing the cag PAI are associated with increased interleukin-8 (IL-8) expression and inflammation in gastric mucosal biopsy specimens and increased IL-8 expression and apoptosis in vitro.48 Infecting gerbils with mutated strains lacking cagE reduces the severity of gastritis and the development of gastric ulcers, intestinal metaplasia, and gastric cancer compared with gerbils infected with the wild-type strain.48,49 Human studies in which duodenal ulceration occurred

more frequently in children carrying strains expressing cagE associated with higher levels of gastric IL-850 corroborate animal and in vitro studies. All strains of H. pylori possess the vacA gene, with more than half expressing the vacuolating cytotoxin (VacA), which attaches to epithelial cells via an interaction with protein-tyrosine phosphatases.51 Although the majority of the VacA is secreted, some may remain on the surface of the bacteria and serve as a ligand for bacterial attachment via this protein-tyrosine phosphatase receptor. Several studies have examined the structure and function of VacA and its association with disease.26,52,53 For example, mice deficient in protein-tyrosine phosphatase beta do not develop ulceration when exposed to VacA.54 Different vacA alleles have been classified in the 5′ signal region (s-region) and the middle region (m-region) of the vacA gene.52 The s-region is present as s1 (which can be further distinguished as s1a, s1b, s1c) or s2, whereas the m-region is present as m1 or m2. Production of VacA is designated by the allelic combination s1/m1 and s1/m2. Specific vacA alleles (s1 and m1) are associated with disease55 and the induction of epithelial cell apoptosis.56 The interaction between VacA and its receptor(s) appears to be important in the pathogenesis of gastroduodenal disease, whether it serves as a ligand for bacterial attachment or as a secreted virulence factor.

HOST RESPONSE TO INFECTION Increasing evidence suggests that the host response to H. pylori infection is an intrinsic component of the patho­ genesis of gastrointestinal disease (Chapters 29, 52, and 54). The possibility that the host response may play a direct role in gastric cancer is supported by the observation that heterogeneity in the regions of the genome that control the magnitude of inflammation is associated with cancer linked to H. pylori infection.57 Polymorphisms in the regions controlling IL-1β58 were shown to be associated with an increased incidence of hypochlorhydria and gastric cancer. This seminal observation has been replicated in other studies in which IL-1β polymorphisms were associated not only with gastric cancer but also a decrease in recurrence of duodenal ulcer.59 An increase in IL-1 may not only drive inflammation but also lead to a physiologic state known to precede gastric cancer development because IL-1 potently inhibits gastric acid secretion. Increased gastric IL-1, more severe gastritis, gastric atrophy, and greater colonization with H. pylori strains have been associated with gastric cancer.2 Other genes that regulate the magnitude of the inflammatory response, including IL-10, tumor necrosis factor-α (TNF-α), and IL-8, have also been associated with the sequence of events leading to cancer.60,61 It is apparent that epithelial cells play an integral part in the host response to H. pylori infection as well as being the target of infection. The epithelial response to infection is complex, as it is driven by several variables: bacterial virulence factors; the signaling linked to specific receptors that recognize the bacterial components; and the local milieu of hormones, neurotransmitters, immune or inflammatory cytokines and mediators, as well as stromal factors. The epithelial cell responses include changes in epithelial cell morphology (the hummingbird phenotype),62 disruption of the tight junctional complexes,63 the production of cytokines,46 increased epithelial cell proliferation, increased rates of epithelial cell death via apoptosis, and the induction of numerous genes associated with the stress encountered in response to infection.2 The best overview of the

835

836

Section VI  Stomach and Duodenum epithelial cell response can be ascertained from the broad changes in the expression of hundreds of genes demonstrated with high-throughput gene expression systems.2 Detailed analyses of the effects of infection, different cell lines and various inflammatory mediators have yet to be carried out over the wide range of time points that would be relevant to the conditions epithelial cells face in vivo. The expression of genes in epithelial cells stimulated with H. pylori is regulated by transcription factors that are controlled by a series of signaling mechanisms. While many transcription factors are likely activated by infection, the most studied are nuclear factor-kappa B (NF-κB) and activator protein-1 (AP-1) which regulate the expression of a wide variety of proinflammatory cytokines and cellular adhesion molecules in response to infection or the local cytokine milieu. H pylori activates NF-κB in gastric epithelial cells, both in vitro and in vivo in patients with H. pylori gastritis; gastric epithelial cell NF-κB activity is markedly enhanced, correlating with the intensity of neutrophil infiltration and IL-8 protein levels.46 This pathway is of particular interest given the recent report that polymorphisms in the IL-8 gene lead to increased mucosal IL-8 expression, inflammation, and other premalignant changes associated with gastric cancer. H. pylori infection appears to activate NF-κB and AP-1 in gastric epithelial cell lines through various signaling mechanisms including mitogen-activated protein (MAP) kinases.47,64 The MAP kinase cascades regulate a wide range of cell functions, including proliferation, inflammatory responses, and cell survival. For example, cag PAI-positive H. pylori activate the ERK, JNK, and p38 MAP kinase pathways, and ERK and p38 regulate IL-8 production in gastric epithelial cells.65,66 An inhibitor of p38α MAP kinase, FR167653, reduced both neutrophil infiltration and gastric mucosal injury in H. pylori–infected Mongolian gerbils.67 One of the novel concepts emerging is the role of oxidative stress in regulating gene expression during H. pylori infection. Apurinic-apyrimidinic endonuclease-1 (also referred to as redox factor-1) plays a key role in the regulation of redox-sensitive signaling and is expressed in gastric epithelial cells during infection with H. pylori,68 contributing to activation of AP-1 and NF-κB required for the host response to infection, including IL-8 production.69 However, the role it and other redox-regulatory molecules play in the pathogenesis of diseases associated with H. pylori has yet to be clearly defined. Oxidation of DNA by reactive oxygen species such as hydroxyl radicals are thought to play a causal role in malignant transformation through the induction of DNA damage. Oxidative DNA damage is increased in gastric epithelial cells by H. pylori infection. There is growing interest in the role of antioxidants in disease prevention or treatment because infection is associated with decreased levels of a tissue antioxidant scavenger, vitamin C. Moreover, there is evidence that diets high in antioxidants70 or “nutraceuticals” of the isothiocyanate group, such as sulforaphane,71 can antagonize oxidative stress and protect the host from gastric cancer, perhaps by decreasing inflammation and attenuating bacterial load. CagA protein translocates into the host cell cytoplasm where it is tyrosine phosphorylated by host Src kinases, and through other interactions, regulates epithelial cell morphology.72-76 CagA in strains from distinct geographic populations appears to be phosphorylated in a different manner resulting in different effects on intracellular signaling.77,78 It is intriguing to speculate that heterogeneity in the CagA protein may lead to distinct effects on the host response that account for some of the geographic differences in disease. Although phosphorylation of the CagA protein may be important, it is not the only mechanism whereby this

molecule regulates the host response. A phosphorylationindependent effect on gene transcription, which is also attributed to the CagA protein, also has been reported.79 Moreover, greater than 30% of host gene expression is altered independently of the phosphorylation of CagA, at least in intestinal epithelial cell lines.80 Outer inflammatory protein A (OipA) is another bacterial product that induces epithelial cell IL-8 production81 and strains that express OipA are associated with increased bacterial density, mucosal IL-8 levels, and neutrophil infiltration, as well as more severe clinical consequences.82 Peptidoglycan has been reported to translocate into gastric epithelial cells via the type IV secretion system encoded by the cag PAI. Once inside the cell these bacterial compounds are recognized by nucleotide-binding oligomerization domain-1 (NOD1), thereby providing a novel mechanism of bacterial sensing.83 Binding to NOD1 can lead to activation of NF-κB and the subsequent expression of various host genes encoding proinflammatory molecules. As discussed in Chapter 49, gastric acid secretion is a major function of the gastric mucosa that is regulated by a variety of neural, endocrine, and immune factors.3 Elevated fasting and meal- or hormone-stimulated levels of gastrin are well documented in H. pylori infection, and there is evidence that gastrin expression is regulated by bacterial factors and cytokines. Expression of somatostatin, an acidinhibitory peptide, is diminished in infected individuals as is duodenal bicarbonate secretion. The net effect of H. pylori infection on acid secretion is complex and varies depending on the duration and distribution of infection and presence of mucosal atrophy. Secretion of mucus is also affected by H. pylori infection with decreased amounts of mucus and gastric mucosal hydrophobicity; these abnormalities reverse after eradication of infection. Epithelial barrier function is altered during H. pylori infection as a consequence of both direct effects of H. pylori and the accompanying inflammatory response that collectively increase epithelial cell proliferation and programmed cell death.2 Infection with H. pylori results in a unique inflammatory response in which infection persists despite the recruitment and activation of T and B lymphocytes, phagocytic cells, and other immune cell populations. Whereas several epithelial cell responses to H. pylori have been described earlier, they do not appear to fully account for the magnitude of the inflammatory response to an organism that resides predominantly in the lumen. Some bacteria may infect epithelial cells, and significant amounts of bacterial material may “leak” around epithelial cells and reach the lamina propria, where it can activate underlying phagocytes, including neutrophils and macrophages. One bacterial factor is the H. pylori neutrophil-activating protein (HP-NAP). This 150-kd decamer protein promotes neutrophil adhesion to endothelial cells and stimulates chemotaxis of monocytes and neutrophils, NADPH oxidase complex assembly at the plasma membrane, and the subsequent production of reactive oxygen intermediates (ROIs).84 In the inflammatory environment present during H. pylori gastritis, TNF-α and interferon-γ (IFN-γ) can prime neutrophils and potentate the effects of HP-NAP. When cells undergo apoptosis and die, they are removed by phagocytes. This response occurs in the digestive tract as epithelial cells migrate toward the lumen and undergo apoptosis, providing another means whereby the host can sample the antigenic challenges facing epithelial cells. Viral antigens are presented to T cells when infected apoptotic epithelial cells overlie the Peyer’s patch. Engulfment of H. pylori infected epithelial cells by phagocytes may also be an important mechanism by which H. pylori can activate

Chapter 50  Helicobacter pylori the host response, and it has been shown that macrophages bind and then engulf gastric epithelial cells that undergo apoptosis due to infection.2 Recruitment and activation of macrophages and neutrophils cause the release of other inflammatory mediators. Increased expression of inducible nitric oxide synthase (iNOS) is observed in the gastric mucosa during infection with H. pylori.2 Nitric oxide (NO) and superoxide (O2−), which may be produced by infiltrating neutrophils, react to form peroxynitrite (ONOO−), a potent oxidant and reducing agent. Although these products have potent antimicrobial effects, uncontrolled or inappropriate production could play a role in the gastric mucosal damage observed during H. pylori infection. The catabolism of urea by urease provides CO2, which rapidly neutralizes the bactericidal activity of the peroxynitrate by reacting with it to form ONO-OCO2. Urease may favor bacterial colonization by neutralizing some host responses but this also enhances the nitration potential of ONOO− and may favor mutagenesis of host cell DNA. Cytokines secreted by epithelial cells complement those released in the lamina propria. For example, neutrophils are not only activated by IL-8 but also by chemokines such as ENA-7885 and Gro-α,86 which can derive from the epithelium, the adjacent myofibroblasts or the macrophages within the lamina propria. Cytokine induction in macrophages includes induction of TNF-α and IL-6 by urease,87 whereas heat shock protein 60 induces IL-6.88 Intact bacteria can induce the production of chemokines that recruit T cells,89 as well as IL-1290,91 and IL-18,92 two cytokines that favor the selection of Th1 cells. Thus, intact bacteria or bacterial factors trigger a broad cytokine response within the lamina propria. As adaptive responses develop, different T lymphocyte helper (Th) cell subsets emerge, with characteristic patterns of cytokine secretion. Th1 cells promote cell-mediated immune responses through the production of IFN-γ and TNF-α, whereas Th2 cells produce IL-4, IL-5, IL-10, and transforming growth factor-β (TGF-β). Th2 cells can promote mucosal IgA or IgE responses, as well as diminish the inflammation caused by Th1 cytokines. Previous studies suggest that the infected gastric mucosa is preconditioned to favor Th1 cell development.90 One possible hypothesis to account for this tendency is that infection selectively blocks Th2 development. H. pylori can interfere with STAT6 activation by IL-4,93 which could impair Th2 development, and IL-12 and IL-18 induced in response to infection may positively select for the predominant Th1 response. Other cytokines that may enhance Th1 responses, such as IL-23 and IL-27, have yet to be studied in human tissue. T cell activation by H. pylori infection may contribute to more severe inflammation and gastroduodenal diseases. Increased levels of biologically active IL-17, a cytokine produced by activated CD4+ T lymphocytes, are found in the mucosa of H. pylori–infected patients.94,95 IL-17, in turn, induces IL-8 expression by gastric epithelial cells, thereby enhancing neutrophil recruitment. Activation of transcription factors by IL-17 may also contribute to the increased levels of numerous other proinflammatory cytokines and enzymes observed during H. pylori infection, such as IL-1β, TNF-α, and cyclooxygenase-2 (COX-2). IFN-γ and TNF-α produced by Th1 cells can increase the expression of many genes in the epithelium, including IL-8. These cytokines also enhance bacterial binding36 and may contribute to enhanced bacterial load.96 In animal models Th1 cells increase epithelial cell apoptosis36 as well as inflammation, atrophy, and dysplasia.97 TNF-α, IFN-γ, and IL-1β upregulate gastric mucosal Fas antigen expression.98 Since Th1

cells express higher levels of Fas ligand (FasL) than Th2 cells, the relative increase in Th1 cells during H. pylori infection may induce epithelial cell death through Fas-FasL interactions.98,99 This notion is substantiated by the observation that proton pump H+,K+-ATPase–specific Th1 cells in the gastric mucosa kill target cells via Fas-FasL interactions and may act as effector cells in autoimmune gastritis.100 Because Th1 cells cannot clear H. pylori, some other T cell subset may have to be stimulated in order to confer immunity. Studies in animal models indicate that protective immunity was induced by vaccines for Helicobacter spp. via Th cells other than Th1 cells, possibly including Th2 cells. The anti-inflammatory cytokines associated with Th2 cells or other regulatory subsets of Th cells can attenuate the pathogenic effects of Th1 cells.101 More direct evidence suggests that IL-4 can decrease gastritis, an effect that may be mediated by the release of somatostatin.97 As gastric responses can be modified by Th2 cells, the role of other T cell subsets, such as regulatory T cells (Treg), in the pathogenesis of disease associated with H. pylori infection is being addressed. Depletion of Treg in neonatal mice leads to autoimmune gastritis,102 and infection with H. pylori alleviates autoimmune gastritis induced in neonatal mice.103 This suggests that infection may stimulate a subset of antiinflammatory T cells that impair excessive inflammation, which could otherwise lead to the spontaneous clearance of the organism, an effect that appears to occur in the human mucosa in response to H. pylori infection.104 Antibodies in the gastrointestinal tract are normally of the immunoglobulin A (IgA) isotype, which are highly adapted for mucosal protection, conferring protective immunity without activating complement and stimulating deleterious amounts of inflammation. During infection with H. pylori, the number of IgA producing cells increases. IgG and IgM are also detected, along with activated complement. It has been suggested that the level of autoantibodies in humans correlates with the severity of gastritis.2 Local immune complexes contribute to gastroduodenal inflammation and tissue damage during infection and may contribute to autoimmune gastritis. Monoclonal antibodies that recognize H. pylori cross-react with human and murine gastric epithelial cells.105,106 Adoptive transfer of these antibodies to recipient mice induces gastritis,105 as does the transfer of B cells that recognize heat shock proteins from individuals with maltoma.107 Anti-Le antibodies have been described in humans and occur independently of the Le phenotype of the host but they do not appear to be autoreactive. Autoantibodies induced in mice may recognize different targets within the gastric mucosa and even though they may cross-react with human gastric tissue, autoantibodies induced in humans may have a completely different specificity. With few exceptions, infection with H. pylori persists for the life of the host unless there is some intervention with antibiotics. This observation has led to investigations as to whether immunity is impaired by immunologic avoidance or tolerance. Several bacterial factors, including catalase and urease, antagonize innate host responses. Production of the enzyme arginase by H. pylori inhibits NO production and may favor bacterial survival,108 whereas virulent strains of H. pylori have also been shown to alter mucus production109 and phagocytosis.110 The VacA toxin can impair antigen presentation by macrophages by inhibiting the Iidependent pathway, which is mediated by newly synthesized class II MHC molecules.111 Moreover, H. pylori express antigenic molecules that mimic host molecules, such as Lewis antigens that theoretically could stimulate T cells to release cytokines that inhibit autoimmune reactions.

837

838

Section VI  Stomach and Duodenum However, the cytokine profile associated with H. pylori infection is not one that would be expected to occur in a tolerant environment. For example, IL-4, IL-10, and TGF-β (which could mediate an anti-inflammatory effect) are not expressed to the same levels as proinflammatory cytokines such as IFN-γ and TNF-α.2 Because the infected gastric mucosa is characterized by chronic active inflammation, tolerance, if it has occurred, may favor persistent infection even though it cannot prevent the chronic inflammatory response. An alternative possibility is that rather than the T cell responses being inappropriate, they lack a level of coordination necessary to achieve immunity. Under certain circumstances, bacteria can inhibit the growth of T cells and actually induce their apoptosis.112-114 This process has been demonstrated using cell lines, and attributed to very specific patterns of gene expression, including the induction of FasL and Fas on T cells, which enables apoptosis.98,114 The studies by Wang and associates suggest this can occur directly,114 but peptides from H. pylori may activate monocytes to produce oxygen radicals which impair the expression of CD3 zeta chain of the T cell receptor complex and the induction of T cell apoptosis.113 The loss of T cells by apoptosis as an explanation is further supported by the presence of apoptotic T cells in the gastric mucosa,112 as well as the expression of Fas and FasL on biopsy specimens.98 More recently, VacA has been shown to impair the growth of a T cell line, independently of any effect on apoptosis, by inhibiting nuclear translocation of a transcription factor, NFAT, that is required for IL-2 production.115,116 However, these events are complex, with VacA toxin impairing IL2–dependent expansion in T cells freshly isolated from human peripheral blood, but not their survival or ability to produce IL-2.117 H. pylori can also impair T cell function directly due to arginase118 as well as indirectly, as a consequence of inducing hydrogen peroxide in adjacent macrophages that leads to T cell apoptosis.113 These observations support a model in which H. pylori interferes with normal T cell activation in several ways. Despite the evidence that H. pylori can impair T cell responses, some antigen-specific T cell responses are found in the gastric mucosa.100 Nonetheless, it is still possible that a combination of the effects of H. pylori and the cytokine milieu disrupt the coordination required for the development of an effective, antigenspecific T cell response.

CONDITIONS ARISING FROM INFECTION Infection of the human stomach by H. pylori leads to gastritis, which initially affects the superficial layers of the mucosa (see Chapter 51). In some instances the infection is short lived, but typically the infection results in a unique pattern of gastritis, so-called chronic active gastritis, which is essentially a lifelong condition. As discussed, chemokines induced by infection lead to a persistent acute inflammatory infiltrate with neutrophils and other cells (active inflammation) coexisting with cells characteristic of chronic inflammation (lymphocytes, macrophages). In spite of this robust immune response bacteria persist in most infected humans, and the factors that result in clearance of some cases of acute infection remain largely unknown.2 Such insights might have augmented the efforts of the past two decades to develop vaccines to prevent chronic H. pylori infection and enhance eradication that have been unfor­ tunately, largely unsuccessful. Most chronically infected

individuals are asymptomatic with somewhere between 10% and 15% going on to develop peptic ulcer disease, gastric adenocarcinoma, and lymphoma of the gastric mucosa–associated lymphoid tissue. As discussed in the previous section, factors that contribute to whether an individual develops these complications of chronic infection include bacterial factors, although these are not yet well defined; host factors including genes that regulate the host immune response (cytokines and their receptors); and environmental factors. As mentioned, H. pylori organisms colonize only gastrictype epithelium within the human host and may colonize tissues outside the stomach when there is gastric metaplasia of the esophagus or duodenum, or in a Meckel’s diver­ ticulum. The pattern of colonization within the stomach appears to be an important determinant of H. pylori disease manifestations. It is unclear exactly what leads to duodenal ulcers associated with H. pylori infection, but it is thought that hyperacidity associated with antral colonization leads to gastric metaplasia of the duodenum, which can then become colonized, leading to duodenal ulcer in some instances. Distal gastric infection may also present with erosions and small ulcers in the gastric antrum, similar in appearance to lesions associated with anti-inflammatory drug use. Gastric ulcers and gastric adenocarcinoma occur more often when there is proximal colonization of the stomach (pan-gastritis), which results in injury to the gastric glands, leading to atrophic gastritis and associated hypoor achlorhydria (see Chapter 54). Precursor lesions of gastric cancer, including atrophic gastritis, intestinal metaplasia, and dysplasia,119 result from infection although most individuals with intestinal metaplasia in North American populations do not have evidence of infection.120 As such, there are no general recommendations to screen individuals with intestinal metaplasia in the United States. The patterns of infection-associated gastritis and resultant effects on acid secretion are thought to account for the reported differences of the impact of infection and eradication on pre-existing and new symptoms of gastroesophageal acid reflux.121 Those with antral-predominant infection have hyperacidity, which may promote acid reflux that improves after eradication therapy. In contrast, those with pangastritis do not apparently suffer from heartburn but after clearance of infection there is a slow return of secretory function, which may be associated with the development of heartburn. Perhaps even more contentious is the inverse association of H. pylori infection with Barrett’s esophagus and esophageal adenocarcinoma (see Chapters 43, 44, and 46). This may be a related to acid exposure but studies indicate that Cag A–positive strains are more strongly associated with a reduced frequency of Barrett’s esophagus and esophageal adenocarcinoma than Cag A–negative strains. This negative association seems to be greatest in Eastern countries.122 Of note, gastric cancer remains the second major cause of cancer death in the world. The burden of risk of gastric cancer is considered largely attributable to H. pylori infection, with cag PAI–bearing strains having a higher association with gastric cancer than cag PAI–negative strains.123 Given the burden of gastric cancer worldwide, the risk of infection seemingly outweighs the benefits in terms of the development of proximal gastrointestinal tract cancer. Over the past two decades a large number of associations with nongastric diseases and H. pylori infection have been reported including Raynaud’s, scleroderma, idiopathic urticaria, acne rosacea, migraines, thyroiditis, and GuillainBarré syndrome, but the data supporting an association

Chapter 50  Helicobacter pylori for this group of conditions are weak or nonexistent.124 Associations that have somewhat better levels of evidence for an association include coronary artery disease, immune thrombocytopenic purpura,125,126 and iron deficiency anemia127 and for the latter two conditions, eradication of infection may be considered when other treatments have failed. The proposed mechanisms leading to these various conditions range from systemic immune reactions, crossreactivity of bacterial and host proteins, and events secondary to gastric mucosal injury.

DIAGNOSIS The National Institutes of Health (NIH) proposed consensus recommendations for H. pylori testing applicable to the United States in 1994.128 H. pylori management in other parts of the world is generally concordant with the United States, but there are some exceptions related to regional differences in clinical practice and specific infection-related disease prevalence (gastric cancer and premalignant gastric histology).129,130 The American College of Gastroenterology published updated U.S. guidelines in 2007 that recommend testing for H. pylori only if a clinician is prepared to treat a patient with a positive test result.131 Specific indications for testing include patients with active or documented history of uncomplicated or complicated peptic ulcer, early gastric cancer, or gastric MALT lymphoma. Testing for H. pylori is often recommended in younger patients with uninvestigated dyspepsia and no “alarm features” (i.e., early satiety, unexplained weight loss, dysphagia, recurrent vomiting, family history of gastric cancer)132 and in patients with functional dyspepsia (symptoms and negative endoscopy).133 However, the clinical and cost benefits of H. pylori in the setting of dyspepsia remain controversial (see Chapter 13), especially in regions where prevalence of infection is relatively low and gastroesophageal reflux disease (GERD) as a cause of symptoms is high.134 Testing for infection prior to starting nonsteroidal anti-inflammatory drugs (NSAIDs) may reduce subsequent ulcers, but this not generally recommended or often done in the United States, where prevalence of H. pylori is low.131 Also there is no general recommendation to test asymptomatic persons, with the possible exception of those with a family history of gastric cancer,135 particularly individuals of Asian, Eastern European, or Mesoamerican descent, for whom the risk of gastric malignancy is highest. Occasionally immune thrombocytopenic purpura126,136 and refractory iron deficiency anemia127 respond to eradication of infection, so decisions to test for H. pylori in these conditions are made on a case-by-case and regional basis. Finally a potpourri of conditions is linked to H. pylori infection, but causal evidence and biologic plausibility are sufficiently weak or lacking in these clinical situations to recommend testing.124 Indications for testing and treatment for H. pylori are summarized in Table 50-1. There are endoscopic and nonendoscopic means to diagnose infection, and techniques can directly (histologic demonstration of organisms, presence of bacterial antigen in the stool, culture) or indirectly (using urease or an antibody response as a marker of bacteria) detect H. pylori.137,138 The appropriate method to choose depends on the clinical situation, population prevalence, and pretest probability of infection as well as test availability and cost. In addition, recent use of antibiotics or proton pump inhibitors can influence results of certain tests.131 The advantages and dis-

Table 50-1  Indications for Testing and Treatment of Helicobacter pylori Infection Supported by evidence Active peptic ulcer disease (gastric or duodenal ulcer) Confirmed history of peptic ulcer (not previously treated for H. pylori infection) Gastric MALT-lymphoma (low grade) Following endoscopic resection of early gastric cancer Uninvestigated dyspepsia (if H. pylori population prevalence high) Controversial Functional dyspepsia GERD Persons using NSAIDs, especially when first initiating NSAID treatment Unexplained iron deficiency anemia or immune thrombocytopenic purpura Populations at higher risk of gastric cancer (e.g. Asians, Eastern Europeans, Mesoamericans) GERD, gastroesophageal reflux disease; MALT, mucosa-associated lymphoid tissue; NSAIDs, nonsteroidal anti-inflammatory drugs. Adapted from references 129 and 131.

advantages of commonly used diagnostic tests for H. pylori are summarized in Table 50-2. Performing endoscopy solely to diagnosis H. pylori infection is not appropriate; there are three methods—biopsy urease test, histology, and (less often) culture—to identify the organism during an otherwise indicated endoscopic procedure. The choice of method depends on the clinical situation, cost, and test accuracy.131 Guidelines propose initially using a biopsy urease test because the method is quick, easy to perform, relatively inexpensive, and generally accurate. Gastric biopsy material is tested for urease activity by placing several pieces of tissue in a medium containing urea and a pH reagent. Bacterial urease hydrolyzes urea-liberating ammonia, producing an alkaline pH and a resultant color change of the test medium.137 Test results are often positive within minutes to hours. Several urease test kits are commercially available based on the methodology described here, differing only with regard to medium (agar gel or membrane pad) and testing reagents.137 These test kits are generally inexpensive but there are added costs associated with obtaining gastric tissue samples, for example, up-coding diagnostic esophagogastroduodenoscopy (EGD), to EGD with biopsy. Nevertheless, biopsy urease testing is less expensive than histology so one proposed cost-saving measure is to obtain specimens for histology but delay sending them to the laboratory pending urease test results. Specificity of the urease tests is 95% to 100% with falsepositive tests uncommon.137,139 Although reported sensi­ tivity of urease tests is 90% to 95%, accuracy can be negatively affected by blood in the stomach,140 and current or recent use of medications such as antibiotics, bismuthcontaining compounds, or acid inhibitors, especially PPIs.141 Therefore, a negative urease test does not necessarily exclude H. pylori infection in an individual taking anti­ secretory medication, a common scenario in patients referred for endoscopy. Testing samples from multiple regions of the stomach or stopping offending medication and delaying endoscopy for several weeks may improve test sensitivity in such patients. Evaluation of gastric mucosal histology is generally not necessary to diagnose H. pylori, but it can provide infor­ mation regarding the activity and severity of mucosal inflammation (see Fig. 51-3C). Histology can also detect metaplasia, dysplasia, and neoplasia.137 In addition to biopsying “clinically suspicious” areas, taking multiple biopsies

839

840

Section VI  Stomach and Duodenum Table 50-2  Diagnostic Tests for Helicobacter pylori NONENDOSCOPIC TESTs

ADVANTAGES

DISADVANTAGES

Serology (qualitative or quantitative immunoglobulin G [IgG]) Urea breath test (13C or 14C)

Widely available, inexpensive, good NPV

Poor PPV if HP prevalence is low, not useful after treatment Availability and reimbursement inconsistent, accuracy affected by PPI and antibiotic use, small radiation dose with 14C test Fewer data available for polyclonal test, accuracy affected by PPI and antibiotic use

Stool antigen test

Identifies active infection, accuracy (PPV, NPV) not affected by H. pylori prevalence, useful both before and after treatment Identifies active infection; accuracy (PPV, NPV) not affected by H. pylori prevalence; useful both before and after treatment (monoclonal test)

ENDOSCOPIC TESTs

ADVANTAGES

DISADVANTAGES

Histology

Excellent sensitivity and specificity, especially with special and immune stains; provides additional information about gastric mucosa Rapid results, accurate in patients not using PPIs or antibiotics, no added histopathology cost Specificity 100%, allows antibiotic sensitivity testing Excellent sensitivity and specificity, permits detection of antibiotic resistance

Expensive (endoscopy and histopathology costs), interobserver variability, accuracy affected by PPI and antibiotic use Requires endoscopy, less accurate after treatment or in patients using PPIs Difficult and tedious to perform; not widely available; expensive Not widely available; technique not standardized; expensive

Rapid urease test Culture Polymerase chain reaction (PCR) assay

NPV, negative predictive value; PPI, proton pump inhibitor (see Table 50-3); PPV, positive predictive value. Adapted from Chey WD, Wong BC. American College of Gastroenterology guideline on the management of Helicobacter pylori infection. Am J Gastroenterol 2007; 102:1808-25.

Figure 50-1.  Photomicrograph of a gastric mucosal biopsy specimen, stained with the Genta stain, from a patient with Helicobacter pylori gastritis. The bacteria are well seen. (Courtesy of Hala El-Zimaity, Houston, Tex.)

and sampling lesser and greater curvatures of gastric antrum and body are important, especially when looking for evidence of multifocal atrophic gastritis and/or intestinal metaplasia (see Chapter 51). Histologic examination had been considered the gold standard for identifying infection, with reported sensitivity and specificity as high as 95% and 98%, respectively.142 However, the distribution and density of organisms can vary within the stomach resulting in sampling error, particularly in patients taking antisecretory medications.131,137 Detecting organisms can be difficult when standard hematoxylin and eosin staining is used alone, but is less of an issue when processing tissue with special stains such as Giemsa, silver, or Genta (Fig. 50-1) or specific immune stains.138,142,143 H. pylori are difficult to culture because the organism is fastidious, slow growing, and requires specialized media

and growth environment.137,138 In fact the initial isolation of H. pylori occurred by happenstance when plated cultures incubated over a long holiday weekend. When culturing for H. pylori, tissue should be obtained before biopsy forceps become contaminated with formalin and placed in a container with only a few drops of saline to preserve the specimen during transport to a local or offsite microbiology facility.138 Although culture is not generally recommended, in those with refractory disease culture with antibiotic sensitivity testing can guide subsequent treatment, although in vitro sensitivity testing does not always predict clinical treatment outcome.138,144 Most often nonendoscopic tests are used to diagnose H. pylori infection, and serology remains the most popular method used, although use of other noninvasive methods that can detect active infection has increased. Infection incites a systemic immune response, and enzyme-linked immunosorbent assay (ELISA) technology can detect IgG antibodies to a variety of bacterial antigens in serum samples.137,138 Tests for IgA and IgM class antibodies are less reliable and not recommended.137 Office-based kits that test whole blood can provide results within 30 minutes and permit “point of service” testing. Although serology is inexpensive, noninvasive and ideally suited to a primary care setting, the prevalence of H. pylori in the population being tested influences its accuracy.131 The sensitivity of serology is generally quite high (90% to 100%) but specificity is variable (76% to 96%), especially if prevalence of H. pylori is low. Therefore, in places where infection is less common (most areas of the United States), the negative predictive value of serology is high. On the other hand, the corresponding positive predictive value is poor, suggesting most often positive results are actually falsely positive.131 So it is best to confirm positive serology results with another method such as a stool antigen or urea breath test before starting treatment or to use a test that detects active infection in the first place. Conversion of a positive serology to negative after treatment suggests bacterial cure, but in most instances serology remains positive for months to years even after successful treatment of infection.145 This “serologic scar” effectively precludes use of serology to confirm

Chapter 50  Helicobacter pylori 13C-urea 13CO 2 (µmol)

Positive urea breath test

NH2 H2O 2NH3 + 13CO 2

+ 13C

O

NH2 Urease Breath 13CO 2

Negative urea breath test 2 1 Time (hour)

Blood Figure 50-2.  The urea breath test. (From Walsh JH, Peterson WL. Drug therapy: The treatment of Helicobacter pylori infection in the management of peptic ulcer disease. N Engl J Med 1995; 333:984.)

bacterial eradication after treatment, a practice that is unfortunately still quite common in the primary care setting even though better tests to confirm eradication are more widely available. The urea breath test (UBT) detects active H. pylori infection and so it is useful for making the primary diagnosis, confirming the accuracy of serology, and documenting successful treatment.131 UBT relies on bacterial hydrolysis of orally administered urea tagged with a carbon isotope, either 13C or 14C (Fig. 50-2). Hydrolysis generates ammonia and tagged CO2, which can be detected in breath samples.137,138 The 13C test is best for children and pregnant women because it uses a nonradioactive isotope, whereas the radiation dose with the 14C test is 1 microCi138 equivalent to one day of background radiation exposure. The specificity of UBT is more than 95%131,137; therefore, false-positive results are uncommon. The sensitivity of the test is 88% to 95% with false-negative results reported in patients taking antisecretory therapy such as PPIs,131,141 bismuth, or antibiotics. To improve accuracy, antibiotics should be stopped at least four weeks and PPIs at least one week before breath testing.131 UBT is not accurate in patients who have had gastric resective surgery. An immunoassay that detects the presence of bacterial antigens in stool of infected patients is an alternative nonendoscopic method to diagnose active H. pylori infection as well as confirm eradication following treatment. Overall sensitivity and specificity of the stool test are comparable to the UBT (94% and 97%, respectively).131,138,140 A rapid H. pylori stool antigen test is available that permits testing during a clinic visit but it is slightly less accurate than a traditional laboratory based stool test.146 The sensitivity of stool testing is negatively affected by PPIs, bismuth, and antibiotics, which can decrease bacterial load, so similar precautions as described earlier for UBT are appropriate when using stool tests.131,138,141 Polymerase chain reaction is a sensitive method to detect H. pylori in gastric mucosal biopsies, but it is not practical for routine clinical diagnosis. It is, however, used for research purposes to identify bacteria when ordinary culture is difficult, as when testing stool or drinking water in a community setting, to type organisms during epidemiologic or transmission studies or for “real time” antibiotic resistance testing of tissue.138,147 Current recommendations for testing are as follows. A stool antigen assay or UBT is the preferred noninvasive

method for initial diagnosis of H. pylori because it can detect active infection. Serology is only useful to exclude H. pylori infection, and positive serology results should be confirmed by a test for active infection before starting treatment. Endoscopic biopsy is suitable for patients undergoing a diagnostic endoscopy who are found to have an abnormality such as an ulcer or for those requiring endoscopy to follow-up a gastric ulcer or suspected MALT lymphoma. Biopsy urease testing can be used in patients not taking a PPI or antibiotics when histopathology is not clinically necessary. When clinically indicated it is appropriate to confirm successful eradication of infection with either a UBT or stool antigen test. These tests should not be performed sooner than four to six weeks after completion of treatment because earlier testing might yield false-negative results. Also medications that could affect test results such as PPIs should be discontinued at least one week prior to testing to improve accuracy. Post-treatment endoscopy with biopsy is only necessary if a repeat procedure is clinically indicated to follow up complicated ulcer disease or other mucosal abnormality, but this should be delayed for at least four to six weeks after therapy. Sampling multiple areas of the stomach is important to avoid missing persistent infection because of density and distribution of bacteria by prior antibiotics and concomitant antisecretory medications. Serology is not useful for follow-up because the test remains positive in most patients for months or even years after infection is gone.

TREATMENT Because there is currently no “pylori specific” or single antibiotic available to cure infection, treatment requires combining several medications. Recommended regimens usually include two antibiotics dosed several times daily for 7 to 14 days along with acid-suppressive medication.131,148 Attempts to simplify regimens or shorten treatment duration generally reduce effectiveness. Compliance can be a problem because taking multiple medications is difficult, and minor medication-related side effects are frequent. Treatment success can vary among countries and even within regions of countries, possibly related to antibioticresistant organisms that are more common than previously appreciated.148,149 Despite these concerns, treatment regimens are available that cure H. pylori infection in more 75% of individuals.131,148,150,151 After cure, annual adult reinfection especially in developed countries is uncommon, probably less than 1%. Higher rates of reinfection are reported, but these often include cases that actually represent recrudescence of the original infection that failed to clear during antibiotic treatment.152 Reinfection tends to be higher in children especially after spontaneous clearance of a primary infection, and it is reported to be higher in adults living in areas of the world with high H. pylori prevalence.9,11,153 Commonly used treatment regimens are summarized in Table 50-3.129,131,148,150,151 Triple therapy, composed of two antibiotics, clarithromycin (500 mg twice daily) and amoxicillin (1 g twice daily) along with a PPI (e.g., omeprazole 20 mg twice daily, lansoprazole 30 mg twice daily, pantoprazole 40 mg twice daily, rabeprazole 20 mg twice daily, or esomeprazole 40 mg every day) for 7 to 14 days, is currently the most popular initial treatment for H. pylori. PPI triple therapy consistently cures more than 80% of infections, especially if organisms are sensitive to clarithromycin

841

842

Section VI  Stomach and Duodenum Table 50-3  First-Line Treatment of Helicobacter pylori Infection* TREATMENT REGIMEN* †

DURATION

ERADICATION RATE

COMMENTS

PPI , clarithromycin 500 mg, amoxicillin 1000 mg (each twice daily)

10-14 days

70%-85%

Macrolide resistance affects eradication success; not appropriate for penicillin allergic individuals or those who have received a clarithromycin regimen in the past

PPI†, clarithromycin 500 mg, metronidazole 500 mg (each twice daily)

10-14 days

70%-85%

Appropriate for penicillin-allergic individuals who have not received a clarithromycin-containing regimen in the past

PPI†, amoxicillin 1000 mg (each twice daily)   followed by PPI†, clarithromycin 500 mg, tinidazole 500 mg (each twice daily)

5 days 5 days

90%

Appears highly effective despite clarithromycin resistance; limited experience to date in the United States

Bismuth subsalicylate 525 mg, metronidazole 500 mg, tetracycline 500 mg (each four times daily)   plus PPI† or H2RA (twice daily)

10-14 days

75%-90%

Inexpensive but complicated regimen; consider in penicillin allergic individual or if clarithromycin resistance is suspected; can be used for retreatment (see Table 50-4)

*Note that not all of these regimens are currently approved by the U.S. Food and Drug Administration (FDA). † Lansoprazole 30 mg, pantoprazole 40 mg, rabeprazole 20 mg, omeprazole 20 mg, or esomeprazole 40 mg (esomeprazole can be taken once daily). H2RA, histamine H2-receptor antagonist; PPI, proton pump inhibitor. Adapted from references 131, 150, and 151.

and longer treatment duration (14 or 10 days vs. 7 days) is used. Metronidazole (500 mg twice daily) can be substituted for either amoxicillin or clarithromycin, but this is appropriate only for penicillin-allergic or macrolide-intolerant individuals because metronidazole resistance is common and can reduce treatment success.129,131,148,150 A 10-day sequential regimen (a PPI and amoxicillin 1 g, each given twice daily for the first 5 days, followed by the PPI, clarithromycin 500 mg, and tinidazole 500 mg, each given twice daily for the remaining 5 days) improved overall eradication rates compared with standard PPI triple therapy (89% vs. 77 %), but was particularly better for clarithromycin-resistant bacteria (89% vs. 29%).154 A pooled analysis of studies evaluating sequential therapy confirmed its superior efficacy especially with macrolide-resistant bacterial strains.155 Such results are encouraging, although most experience with this treatment is geographically limited to Mediterranean countries. However, there is no reason to expect different efficacy in other regions.131,154 Although used more than a decade ago, dual regimens consisting of a single antibiotic (amoxicillin or clarithromycin) and a PPI are no longer recommended because eradication is significantly less than with three drug regimens.150 Bismuth-based therapy, which combines a bismuth salt, metronidazole 500 mg and tetracycline 500 mg each given four times a day, and daily acid suppression (usually a PPI every day) for two weeks was actually one of the first therapies used to treat H. pylori. Although it remains effective (more than 80% eradication), the number of daily pills and associated frequent minor side effects negatively affect tolerability and compliance. For that reason, in the United States it is usually reserved as a second-line or retreatment regimen.129,131,148,150 A combination capsule that contains bismuth subcitrate 140 mg, metronidazole 125 mg, and tetracycline 125 mg is available in the United States and Canada, simplifying bismuth-based treatment. In a comparative study, patients treated with three combination capsules four times daily and PPI twice daily for 10 days had comparable H. pylori eradication with those treated with traditional PPI triple therapy (88% vs. 83%).156 Short course (1 to 7 days)

bismuth-based treatment157 has been evaluated, but consistent long-term cure of infection has not been confirmed, so abbreviated treatment cannot be recommended.150 Initial treatment of H. pylori infection fails in up to 25% of patients as a result of an infection with antibioticresistant organisms, poor compliance with medication, and patient demographics such as younger age, smoking, prior antibiotic use, and underlying condition (functional dyspepsia vs. peptic ulcer).149,158,159 A review of various retreatment regimens (Table 50-4) reported eradication rates of 46%, 70%, 80%, and 76 % percent for PPI dual therapies, PPI triple therapies, ranitidine bismuth citrate–based triple therapy and bismuth-based therapy, respectively.160 Ranitidine bismuth citrate is no longer available in the United States. When two new antibiotics are used during retreatment, cure of infection appears to be superior compared with when only one new antimicrobial is used. One more recently recommended “rescue therapy” includes a PPI, levofloxacin 250 mg, and amoxicillin 1 g, all given twice daily for 10 days. This combination cures infection in up to 80% of patients who have failed one or more prior treatment attempts. Less well studied, but reported to be 85% effective when used as retreatment, is a combination of PPI and amoxicillin 1 g, each twice daily, along with rifabutin 300 mg every day for 10 days. A lower dose of rifabutin (150 mg) appears to be less effective. Successful retreatment with regimens substituting furazolidone for metronidazole has also been reported, but furazolidone is no longer commercially available in the United States. As initial treatment for H. pylori, a 10- to 14-day course of standard PPI triple therapy described previously (PPI, amoxicillin and clarithromycin) is recommended but a 10-day sequential regimen would be an appropriate alter­ native, especially if clarithromycin-resistant infection is suspected (see following). If infection persists after this treatment, bacteria are likely resistant to clarithromycin. Therefore, retreatment should be with one of the PPI triple regimens noted earlier that incorporates a different combination of medications or a bismuth-based therapy for 14 days. Subsequent courses of treatment if necessary should also incorporate different antibiotic combinations when

Chapter 50  Helicobacter pylori Table 50-4  Rescue Treatment for Persistent Helicobacter pylori Infection* REGIMEN

DURATION

ERADICATION RATE

COMMENTS

Bismuth subsalicylate 525 mg, metronidazole 500 mg, tetracycline 500 mg (each four times daily)   plus PPI† or H2RA (twice daily)

14 days

70%

Inexpensive but complicated regimen, so compliance should be emphasized; less effective as retreatment than as initial therapy; full dose of metronidazole and two weeks of treatment appear necessary

PPI†, amoxicillin 1000 mg, levofloxacin 250 mg (each twice daily)

10-14 days

57%-91%

Limited data from the United States

PPI† amoxicillin 1000 mg, rifabutin 150 mg (each twice daily)

14 days

60%-80%

Expensive; adverse hematologic events and drug interactions possible. Limited data from the United States

*Note that not all of these regimens are currently approved by the U.S. Food and Drug Administration (FDA). † Lansoprazole 30 mg, pantoprazole 40 mg, rabeprazole 20 mg, omeprazole 20 mg, or esomeprazole 40 mg (esomeprazole can be taken once daily). H2RA, histamine H2-receptor antagonist; PPI, proton pump inhibitor. Adapted from references 131, 150, and 151.

possible to lessen the effect of acquired antimicrobial resistance. Although selection of a treatment regimen based on antibiotic sensitivity testing might improve subsequent treatment results, this is not routinely recommended. Primary resistance to antibiotics commonly used to treat H. pylori varies widely throughout the world.138,144,148 In the United States resistance to metronidazole can be detected in up to 40% of stains, whereas clarithromycin resistance is approximately 11%. Resistance to tetracycline and amoxicillin is unusual, generally less than 1%.148,161,162 In the United States, clarithromycin and metronidazole resistance increase with age and are more common in women than in men. Clarithromycin resistance is more common in the mid-Atlantic and northeast regions of the country. Metronidazole resistance is more common in Hispanics and Asians. Antibiotic resistance significantly affects the success of PPI triple regimens but is less important with bismuthbased regimens.138,149 A bacterial point mutation(s) that prevents reduction of metronidazole to its active metabolite is responsible for drug resistance.138,144 Resistance to metronidazole appears to be a relative condition that can be overcome in most instances by using a higher dose (500 mg) or combining the drug with a bismuth preparation. On the other hand, clarithromycin resistance appears to be an absolute situation that cannot be easily overcome by increasing the macrolide dose. One of three bacterial point mutations within its conserved loop of 23S strand of ribosomal RNA (A2143G, A2142G, and A2142C) can interfere with ribosomal macrolide binding and lead to clarithromycin resistance.138,144 The A2143G mutation appears to have the greatest negative effect on treatment and is likely the major reason for PPI triple therapy failure. Testing for specific mutations is not clinically available, so if clarithromycin resistance is suspected or confirmed by culture, nonmacrolide regimens or sequential therapy are appropriate treatment options. Table 50-4 summarizes rescue treatments for persistent infection. Failed attempts at eradication generally result in secondary antibiotic resistance.148 Therefore, one can assume when treatment with clarithromycin or metronidazole-containing regimens is unsuccessful, specific drug resistance has emerged which should influence any subsequent choice of therapy.138,144 Treatment-related side effects can occur in as many as 50% of patients taking one of the treatment regimens

described previously, but generally these are mild and do not require discontinuation of therapy. Some of the more common side effects include taste alteration and gastrointestinal (GI) upset with metronidazole and clarithromycin and allergic reactions and diarrhea with amoxicillin. In addition, tetracycline should not be prescribed to children or pregnant women. Side effects of H. pylori treatment have recently been extensively reviewed elsewhere.150,151

KEY REFERENCES

Amieva MR, El-Omar EM. Host-bacterial interactions in Helicobacter pylori infection. Gastroenterology 2008; 134:306-23. (Ref 5.) Blaser MJ, Atherton JC. Helicobacter pylori persistence: Biology and disease. J Clin Invest 2004; 113:321-33. (Ref 26.) Chey WD, Wong BC. American College of Gastroenterology guideline on the management of Helicobacter pylori infection. Am J Gastroenterol 2007; 102:1808-25. (Ref 131.) Correa P, Houghton J. Carcinogenesis of Helicobacter pylori. Gastroenterology 2007; 133:659-72. (Ref 119.) Ernst PB, Peura DA, Crowe SE. The translation of Helicobacter pylori basic research to patient care. Gastroenterology 2006; 130:188-206. (Ref 2.) Gillen D, McColl KE. Gastroduodenal disease, Helicobacter pylori, and genetic polymorphisms. Clin Gastroenterol Hepatol 2005; 3:1180-6. (Ref 6.) Jafri NS, Hornung CA, Howden CW. Meta-analysis: Sequential therapy appears superior to standard therapy for Helicobacter pylori infection in patients naive to treatment. Ann Intern Med 2008; 148:923-31. (Ref 155.) Jodlowski TZ, Lam S, Ashby CR. Emerging therapies for the treatment of Helicobacter pylori infections. Ann Pharmacother 2008; 42:162139. (Ref 151.) Malfertheiner PF, Megraud, C, O’Morain, F et al. Current concepts in the management of Helicobacter pylori infection: The Maastricht III Consensus Report. Gut 2007; 56:772-81. (Ref 129.) Megraud F, Lehours P. Helicobacter pylori detection and antimicrobial susceptibility testing. Clin Microbiol Rev 2007; 20:280-322. (Ref 138.) Naumann M, Crabtree JE. Helicobacter pylori–induced epithelial cell signalling in gastric carcinogenesis. Trends Microbiol 2004; 12:29-36. (Ref 46.) Suerbaum S, Michetti P. Helicobacter pylori infection. N Engl J Med 2002; 347:1175-86. (Ref 4.) Talley NJ, Vakil N. Guidelines for the management of dyspepsia. Am J Gastroenterol 2005; 100:2324-37. (Ref 132.) Vakil N, Megraud F. Eradication therapy for Helicobacter pylori. Gastroenterology 2007; 133:985-1001. (Ref 148.) Versalovic J. Helicobacter pylori. Pathology and diagnostic strategies. Am J Clin Pathol 2003; 119:403-12. (Ref 137.) Full references for this chapter can be found on www.expertconsult.com.

843

CHAPTE R

51 Gastritis and Gastropathies Edward L. Lee and Mark Feldman

CHAPTER OUTLINE Classification  845 Chronic Gastritis  846 Helicobacter pylori Gastritis  846 Environmental Metaplastic Atrophic Gastritis  846 Autoimmune Metaplastic Atrophic Gastritis  847 Carditis  848 Infectious Gastritides  849 Viruses  849 Bacteria  849 Fungi  851 Parasites  852 Granulomatous Gastritides  852 Sarcoidosis  853 Xanthogranulomatous Gastritis  853 Distinctive Gastritides  853 Collagenous Gastritis  853 Lymphocytic Gastritis  853 Eosinophilic Gastritis  854 Miscellaneous Forms of Gastritis  855 Gastritis in Inflammatory Bowel Disease (Crohn’s and Ulcerative Colitis)  855

Patients, clinicians, endoscopists, and pathologists have different concepts of what gastritis is. Some think of it as a symptom complex, others as a description of the endoscopic appearance of the stomach, and still others use the term to describe microscopic inflammation of the stomach. This third definition of gastritis is used in this chapter. There is not a close relationship between the presence of microscopic inflammation (histologic gastritis) and gastric symptoms (epigastric pain, nausea, vomiting, bleeding). The correlation between microscopic and gastroscopic abnormalities is also poor.1-2 In fact, most patients with histologic gastritis are asymptomatic and have normal gastroscopic findings. Certain disorders of the gastric mucosa including erosive processes and hyperplastic disorders may be associated with little or no inflammation (gastritis). These conditions collectively are referred to as reactive and hyperplastic gastropathies, respectively. By the earlier definition, a gastric biopsy must be obtained to be able to diagnose gastritis. Every biopsy represents an excellent opportunity for the clinician and pathologist to communicate to correlate clinical data, endoscopic findings, and pathology. Errors may occur when the pathologist attempts to diagnose biopsies without clinical input. It is important for the pathologist to become familiar with the range of normal gastric biopsy findings because many gastrointestinal biopsies obtained endoscopically show normal mucosa.3 Indications for gastroscopic biopsies include gastric erosion or ulcer, thick gastric fold(s), gastric polyp(s) or

Gastritis Cystica Profunda  855 Gastric Graft-versus-Host Disease  856 Allergic Gastritis  856 Reactive Gastropathies (Acute Erosive Gastritis)  856 Medications and Toxins  856 Alcohol  856 Portal Hypertensive Gastropathy  857 Cocaine  857 Stress  857 Radiation  857 Bile Reflux  857 Ischemia  858 Prolapse  858 Linear Erosions in a Hiatal Hernia (Cameron Lesions)  858 Aging Gastropathy  858 Hyperplastic Gastropathies  858 Ménétrier’s Disease and Hyperplastic, Hypersecretory Gastropathy  858 Zollinger-Ellison Syndrome  859 Differential Diagnosis of Gastritis and Gastropathy  859 Treatment and Prevention of Gastritis and Gastropathy  859

mass(es), and diagnosis of Helicobacter pylori infection. A set of five biopsies should be taken from patients in whom clinical or endoscopic findings are suspicious for one of the forms of chronic gastritis (discussed later). Preferred sites for this set of biopsies are shown in Figure 51-1. The location of the biopsy sites should be identified for the pathologist on an accessioning form.

CLASSIFICATION 2-7 There is no universally accepted classification of gastritis. The Sydney system was an attempt to unify terminology for endoscopic and histologic gastritis and gastropathy, and it was updated in 1995.4 However, the complexity of the Sydney system precluded widespread use. Failure to obtain adequate numbers of biopsies from various regions of the stomach (see Fig. 51-1) often prevents accurate classification and often precludes a thorough assessment of the distribution of gastritis.5 In this chapter we use a combination of classifications of gastritis by four experts: Rubin,2 Genta,4 Appelman,6 and Montgomery.7 The keystone of the mentioned classification is the fact that H. pylori and nonsteroidal anti-inflammatory drugs (NSAIDs) are the most common causes of gastritis and reactive gastropathies (acute erosive gastritis), respectively. The chapter outline provides an etiology-based classification of gastritis and gastropathies.

845

846

Section VI  Stomach and Duodenum CHRONIC GASTRITIS 7-9 Most forms of chronic gastritis are clinically silent. Their importance relates to the fact that these gastritides are risk factors for other conditions such as peptic ulcer disease, gastric polyps, and benign and malignant gastric neoplasms.8,9 Three types of chronic gastritis are recognized (Figs. 51-2 and 51-3). Biopsies from the antrum and the incisura are useful for diagnosing H. pylori infection with its diffuse antralpredominant gastritis. However, biopsies from the gastric body mucosa may be more diagnostic for H. pylori infection in some patients treated with proton pump inhibitors. Environmental metaplastic atrophic gastritis (also called multifocal atrophic gastritis, or MAG) is patchy and involves the antrum and body mucosa and sometimes, but not always, is associated with H. pylori infection.7 The diagnosis of autoimmune metaplastic atrophic gastritis, also called diffuse corporal atrophic gastritis (DCAG) and type A gastritis, can be confirmed with multiple biopsies from the gastric body that show atrophy and biopsies from the antrum that do not show atrophy. In most cases biopsies are obtained at the time of endoscopy.

HELICOBACTER PYLORI GASTRITIS 2,6,7,10-19

H. pylori gastritis (HPG) is caused by infection of the antral mucosa with H. pylori.2,6,7 In the United States H. pylori

gastritis is seen mainly in low socioeconomic and immigrant populations,7 and there is no increased risk of gastric cancer. Most patients with HPG are asymptomatic. In most cases the antrum appears normal to the endoscopist; some patients with active disease in the antrum may demonstrate red streaks. Radiographic differences between antral gastritis due to H. pylori and not due to H. pylori have been described; thickened gastric folds, especially in a polypoid configuration, and enlarged areae gastricae favor H. pylori as the cause, whereas antral erosions favor causes other than H. pylori.10 Histologically, a diffuse, chronic inflammatory infiltrate which includes numerous lymphocytes and plasma cells expands the lamina propria and epithelium (see Fig 51-3).7 The presence of acute inflammatory cells is best designated an active gastritis and not acute gastritis. Additional microscopic changes include injury to the surface and foveolar epithelium with loss of apical mucin and reactive nuclear changes and erosions.6,11 Lymphoid follicles with germinal centers are characteristic of an infection with H. pylori.3,12 H. pylori organisms lie in the superficial mucous layer along the mucosal surface and within the gastric pits. Although the organisms can be seen in routine hematoxylin and eosin–stained tissue when numerous organisms are present, special stains are useful when few organisms are present. Stains that may be used to highlight the organisms are Giemsa stain, Warthin-Starry silver stain, Gram stain, and immunocytochemical stains.13-15 Helicobacter heilmannii spiral bacteria are a less frequent cause of active gastritis.16-18 The organisms originally known as Gastrospirillum hominis are longer than H. pylori and have multiple spirals.16,17 A topographic study of H. pylori density and distribution and the com­parison of biopsy sites for the histopathologic diagnosis of H. pylori conclude that two antral biopsy specimens, one from the lesser and one from the greater curvature, have close to 100% sensitivity for detecting H. pylori infection (see Fig. 51-1).19 Biopsy specimens from the corpus increase the diagnostic yield if extensive intestinal metaplasia is present in the antrum.19

ENVIRONMENTAL METAPLASTIC ATROPHIC GASTRITIS 4,7,20-32

Figure 51-1.  Gastric biopsy protocol. Blue and black symbols represent sites from which gastric mucosal biopsies should be obtained. Biopsies from the antrum (greater and lesser curvature) and from the incisura are useful for diagnosing Helicobacter pylori infection. Biopsies from the gastric body (greater and lesser curvature) are useful for diagnosing autoimmune metaplastic atrophic gastritis. Biopsies from the antrum and body in combination are useful for diagnosing environmental metaplastic atrophic gastritis.

Environmental metaplastic atrophic gastritis (EMAG), also called mutifocal atrophic gastritis (MAG), is characterized by the involvement of the antrum and body with mucosal atrophy and intestinal metaplasia.4,7,20-22 Atrophic gastritis involving the body may be associated with pseudopyloric metaplasia, in which the mucosa resembles antral mucosa but stains for pepsinogen I (PGI), a proenzyme expressed in body mucosa.23 Gastroscopy may show a pale mucosa, shiny surface, and prominent submucosal vessels,24 and magnify-

Figure 51-2.  Topographic patterns of chronic, nonspecific gastritis. The darkest areas in the schematics of autoimmune metaplastic atrophic gastritis and environmental metaplastic atrophic gastritis represent areas of focal atrophy and intestinal metaplasia.

Diffuse corporal atrophic gastritis

Diffuse antral gastritis

Multifocal atrophic gastritis

Chapter 51  Gastritis and Gastropathies

A

B

C

D

E Figure 51-3.  Chronic nonspecific gastritides. A and B, Normal mucosal biopsy from the gastric body/fundus and antrum, respectively. C, Diffuse antral gastritis. The glands show an infiltrate of neutrophils, in addition to an increase in inflammatory cells in the lamina propria. This lesion is typically associated with Helicobacter pylori infection. D, Multifocal atrophic gastritis with intestinal metaplasia. Note several glands lined by goblet cells (arrow). The biopsy specimen is from the gastric body, and similar changes were present in the antrum. E, Diffuse corporal atrophic gastritis in a man with pernicious anemia. The gland in the lower left is lined by goblet cells. Nests of enterochromaffin-like cells are also visible (arrows).

ing endoscopy is much more sensitive in detecting atrophy.25 The pathogenesis of EMAG is multifactorial. H. pylori plays an important role and has been identified in about 85% of patients with EMAG. EMAG can occur early in life in H. pylori–infected individuals living in developing countries.23 Genetic and environmental factors, especially diet, are also important. Certain population groups are predisposed to EMAG including African Americans, Scandinavians, Asians, Hispanics, Central and South Americans, Japanese, and Chinese. In patients with EMAG, intestinal metaplasia is a risk factor for dysplasia and gastric cancer, usually the intestinal type (see Chapter 54).2,4,7,22,26-31 Inflammation in EMAG destroys gastric epithelial cells, and eventually the atrophic glands are replaced by metaplastic epithelium.4,7,22 In some

cases, especially in patients living in the Pacific basin, metaplastic gastric cells are ciliated, probably due to environmental factors that are more prominent in the Pacific than the Atlantic Ocean basins.32 Because criteria for gastric atrophy among pathologists are debated, intestinal metaplasia is the most reliable marker of atrophy. Intestinal metaplasia of the gastric mucosa can be classified into three types as described in Chapter 54, where their possible associations with the intestinal type of gastric cancer are discussed.

AUTOIMMUNE METAPLASTIC ATROPHIC GASTRITIS 6,7,33-55

Autoimmune metaplastic atrophic gastritis (AMAG), also called diffuse corporal atrophic gastritis (DCAG), is an auto-

847

848

Section VI  Stomach and Duodenum immune destruction of body/fundic glands. AMAG is relatively uncommon, accounting for less than 5% of all cases of chronic gastritis. Endoscopic features of AMAG include effacement of the gastric folds and a thin body/fundic mucosa. AMAG is the pathologic process in patients with pernicious anemia, an autoimmune disorder usually occurring in patients of northern European or Scandinavian background.33 Patients with AMAG exhibit achlorhydria or hypochlorhydria, hypergastrinemia secondary to low or absent gastric acid with antral G-cell hyperplasia, and low serum PGI concentrations, and they often have circulating antibodies to parietal cell antigens and to intrinsic factor.6,7,33 Incomplete (colonic) intestinal metaplasia (type III) may occur in AMAG and be a risk factor for gastric carcinoma in areas of the world that experience a higher incidence of gastric carcinoma than in the United States.34 Metaplastic intestinal Paneth cells in AMAG appear to secrete an antibacterial peptide of the alpha-defensin family, human defensin 5 (HD-5), a peptide not produced in the normal stomach.35 HD-5 could help the atrophic stomach against invasion by indigenous bacterial flora that overgrow in the anacidic stomach (see Chapter 49). Metaplastic pancreatic acinar cells are also a feature of autoimmune gastritis.36 Atrophic glands with extensive intestinal metaplasia are confined to the body/fundic mucosa. Early in the course of this disease, atrophy may be focal and the preserved islands of relatively normal oxyntic mucosa may appear polypoid endo­scopically or radiologically.37 Rarely, AMAG progresses to diffuse (complete) atrophy. Hypergastrinemia, a con­sequence of achlorhydria, is associated with an increase in enterochromaffin-like cell hyperplasia and gastric carcinoid tumors. Cases of gastric carcinoids and simul­ taneous gastric cancer have been described.38 Gastric car­ cinoid tumors are discussed further in Chapters 31 and 54. In one study from Italy, half of 150 patients with AMAG had antibodies to H. pylori and another 25% had H. pylori in their oxyntic mucosa in addition to having antibodies against H. pylori.39 Thus, H. pylori could have contributed to three quarters of the cases of AMAG. Recent studies suggest a role for H. pylori in the early pathogenesis of autoimmune gastritis; evidence of infection early in the course of the disease in individuals with parietal cell antibodies is frequent.40 If gastric atrophy and achlorhydria develop, the incidence of H. pylori infection then decreases. Among 267 H. pylori–infected patients with dyspepsia, 65 had AMAG. Compared with the 202 patients without AMAG, the atrophics were older, more likely to have antibodies against cagA and vacA, more likely to consume alcohol and coffee, more likely to be taking sedative medicines, and less likely to have anxiety.41 Whether H. pylori results in AMAG thus appears to depend on length of infection, as well as bacterial, dietary, and emotional factors. With regard to bacterial factors promoting atrophy, it appears that cagA+/vacA+ H. pylori are more likely to cause AMAG. These H. pylori are often the s1m1 vacA subtype that also express Lewis blood group antigens X and Y.42 Lewis antigens may help camouflage H. pylori because these antigens are also present on human gastric epithelial cells. It has been suggested that when antibodies to Lewis antigens from H. pylori develop, they cross-react with antigens on epithelial cells such as the H+,K+-ATPase on parietal cells, resulting in autoimmune chronic gastritis.43 Based on uncontrolled studies from Tokyo,44 eradication of H. pylori often leads to a decrease in the amount of gastric atrophy and intestinal metaplasia, whereas failed eradication attempts accomplish neither of these endpoints. Antibodies to parietal cell antigens, most notably the proton pump (H+,K+-ATPase) are frequently present in auto-

immune gastritis.45 These antibodies are frequently detected in patients with various autoimmune diseases including type 1 diabetes mellitus46 and thyroid diseases (Graves’, Hashimoto’s), explaining the association of these conditions with pernicious anemia. The risk of AMAG is increased three- to five-fold in type 1 diabetic individuals, and some authors have suggested screening type 1 diabetics with gastroscopy and mucosal biopsy.47 One in eight patients with chronic hepatitis C treated with interferon-α develops antibodies to parietal cells and to thyroid tissue, and these antibodies recede after therapy is stopped48; the clinical significance of these findings in the stomach is yet to be elucidated. A proportion of the CD4+ lymphocytes present in the chronic inflammatory infiltrate within the gastric mucosa proliferate in response to H+,K+-ATPase, and most CD4+ cells secrete Th1 cytokines such as tumor necrosis factor-α (TNF-α); provide help for B cell immunoglobulin production; and enhance perforin-mediated cytotoxicity, as well as Fas ligand–mediated apoptosis.45 These factors in combination may contribute to gland destruction in autoimmune gastritis. An interesting animal model of autoimmune gastritis has been developed in mice in which CD4+ T cells target the β subunit of the H+,K+-ATPase.49 The risk of gastric adenocarcinoma in patients with AMAG is unclear. One recent study suggested a cancer risk of slightly more than 1% per year,50 which would favor periodic endoscopic screening for individuals known to have AMAG. However, other investigators have found cancer much less often and have questioned the costeffectiveness of cancer screening by endoscopy in AMAG.51,52 The importance of incomplete intestinal metaplasia (type III) as a predictor of gastric cancer also has been questioned.53 Thus, at what intervals AMAG patients should be screened, if at all, remains a matter of debate.54 Molecular events involved in the sequence from AMAG to intestinal metaplasia are beginning to be clarified. For example, the expression of the intestinal transcription factor CDX2 precedes expression of other intestinal-specific genes such as CDX1, alkaline phosphatase, MUC2, HD-5, and sucrase-isomaltase55 and thus may be an early trigger of the metaplastic process that precedes dysplasia and carcinogenesis.

CARDITIS 56,57 There has been recent attention to inflammation of the small rim of cardiac glands at the proximal portion of the stomach.56 The pathogenesis of carditis is currently con­ troversial.57 Inflammation of this gland area has been attributed to H. pylori gastritis, EMAG, AMAG, gastroesopha­ geal reflux disease, and other factors. Likewise, atrophy in this area, often accompanied by intestinal metaplasia, has been proposed to be a precursor of adenocarcinoma of the gastroesophageal junction (see Chapters 42 and 44). Der and associates56 reported on 141 patients in whom the cardiac mucosa could be identified in endoscopic biopsies. In this endoscopy population, all biopsies exhibited acute and/or chronic carditis. Nearly 80% of them had no evidence of H. pylori infection on simultaneous biopsies from the gastric body and antrum. H. pylori was present in 20 patients, 17 of whom had pan-gastritis and 15 of whom had H. pylori carditis. The severity of chronic carditis was related directly to 24-hour acid exposure of the lower esophagus, whereas acute carditis was related to H. pylori infection.

Chapter 51  Gastritis and Gastropathies INFECTIOUS GASTRITIDES 58-123 VIRUSES Cytomegalovirus60-64

Cytomegalovirus (CMV) is a human herpesvirus that may affect the esophagus, stomach, small bowel, colon, rectum, anus, liver, and gallbladder. CMV infection may occur in an immunocompetent patient.57 However, gastrointestinal CMV infection usually occurs in the immunocompromised patient. Eosinophilic gastroenteritis with cytomeglovirus infection has been reported in an immunocompetent child.62 Patients with malignant disease, immunosuppression (especially due to steroid therapy), transplants, and acquired immunodeficiency syndrome (AIDS) may experience lifethreatening CMV infections. Patients with CMV infection of the stomach may experience epigastric pain, fever, and atypical lymphocytosis. Upper gastrointestinal tract radiographic studies may reveal a rigid and narrowed gastric antrum suggestive of an infiltrating antral neoplasm. Endoscopic studies may reveal a congested and edematous mucosa of the gastric antrum, covered with multiple ulcerations, suggestive of gastric malignancy, submucosal antral mass, or gastric ulcer (Fig. 51-4). A hypertrophic and/or polypoid type of gastritis resembling Ménétrier’s disease with a similar type of protein-losing gastropathy has been described.58,59 Examination of biopsy specimens shows inflammatory debris, chronic active gastritis, and enlarged cells with CMV inclusion bodies indicative of an active infection (see Fig. 51-4). “Owl-eye” intranuclear inclusions are the hallmark of CMV infection in routine hematoxylin and eosin histologic preparations and may be found in vascular endothelial cells, mucosal epithelial cells, and connective tissue stromal cells. Multiple, granular, basophilic, cytoplasmic inclusions may also be present. Usual treatment with intravenous ganciclovir or foscarnet is of uncertain value (see Chapter 33).

Other Herpesviruses65-70

Gastric involvement with herpes simplex and varicellazoster virus is rare. Infected individuals experience the infection at an early age, and the virus remains dormant until reactivation. Activation has been related to radiation therapy, chemotherapy, lymphoma, and cancer. The typical immunocompromised patient may experience nausea, vomiting, fever, chills, fatigue, cough, and weight loss. An

A

acute abdomen caused by varicella-zoster virus–induced gastritis after autologous peripheral blood stem cell transplantation in a patient with non-Hodgkin’s lymphoma has been reported.67 Barium-air double-contrast radiographs show a cobblestone pattern, shallow ulcerations with a ragged contour, and an interlacing network of crevices filled with barium that corresponds to areas of ulceration. Upper gastrointestinal endoscopy reveals multiple, small, raised, ulcerated plaques or linear, superficial ulcers in a crisscrossing pattern, giving the stomach a cobblestone appearance. Grossly, the ulcers are multiple, small, and of uniform size. Microscopically, cytologic smears and biopsy specimens show numerous single cells and clumps of cells, with ground-glass nuclei and eosinophilic intranuclear inclusion bodies surrounded by halos. Brush cytology and biopsies should be performed at the time of endoscopy. Brush cytology has the advantage of sampling a wider area of mucosa because biopsies may not be representative. Treatment with acyclovir is reasonable but of unproven value. Human herpesvirus 7, a cause of roseola, is frequently present in the gastric mucosa but does not appear to cause gastritis.68 Epstein-Barr virus (EBV) may cause an acute gastritis with lymphoid hyperplasia.69 There is little evidence that EBV causes chronic gastritis.70

Measles71

Rare cases of morbilliform gastritis with giant cells of the Warthin-Finkeldey type have been described.71

Enterovirus72

Recently, it has been proposed that some patients with chronic fatigue syndrome are chronically infected with a noncytopathic, noncytolyic enterovirus that can be detected by immunostaining or by reverse transcriptase–polymerase chain reaction (PCR) techniques using gastric biopsy samples.72 Confirmatory studies are awaited.

BACTERIA Helicobacter pylori (see earlier and Chapter 50)

Phlegmonous (Suppurative) and Emphysematous Gastritis73-82 Phlegmonous gastritis is a rare bacterial infection of the submucosa and muscularis propria of the stomach. Acute necrotizing gastritis (gangrene of the stomach) is a rare, often fatal disease that is now thought to be a variant of phlegmonous gastritis. It has been suggested that acute

B

Figure 51-4.  Cytomegalovirus (CMV) gastritis. A, Low-power histopathology of CMV gastritis. An acute inflammatory infiltrate is present in the lamina propria. Glandular destruction and reactive glands are present. Cystic glands are also evident. B, High-power view of the cystic area deep in the mucosa shown in A. Several cytomegalic cells with the typical intranuclear and intracytoplasmic inclusions of cytomegalovirus are present.

849

850

Section VI  Stomach and Duodenum necrotizing gastritis begins as phlegmonous gastritis, producing primary necrosis and gangrene. Acute necrotizing gastritis and phlegmonous gastritis have been associated with a recent large intake of alcohol; upper respiratory tract infection; AIDS and other immunocompromised states; and an infected peritoneojugular venous shunt. Fulminant and fatal gas gangrene of the stomach in a healthy, live liver donor has been reported.76 Patients typically present with acute upper abdominal pain, peritonitis, purulent ascitic fluid, fever, and hypotension. Preoperative diagnosis is possible with plain film, ultrasonography, or computed tomography (CT), and gastroscopy with or without biopsy and culture of gastric contents may establish the diagnosis. Grossly, the stomach wall appears thick and edematous with multiple perforations, and the mucosa may demonstrate a granular, green-black exudate. Microscopically, the edematous submucosa reveals an intense polymorphonuclear infiltrate and numerous gram-positive and gramnegative organisms, as well as vascular thrombosis. The mucosa may demonstrate extensive areas of necrosis. The mortality rate of phlegmonous gastritis is close to 70%, probably because it is so often misdiagnosed and because treatment is initiated too late. The definitive treatment is resection or drainage of the stomach, combined initially with large doses of systemic broad-spectrum antibiotics directed against the most common organisms such as streptococci, Escherichia coli, enterobacters, other gramnegative bacilli, and Staphylococcus aureus. Emphysematous gastritis is a variant of phlegmonous gastritis in which the infection in the gastric wall is due to gas-forming organisms such as Clostridium welchii. Emphysematous gastritis associated with invasive gastric mucormycosis has been reported.77 Predisposing factors are gastroduodenal surgery, ingestion of corrosive materials, gastroenteritis, or gastrointestinal infarction. Radiographic studies (plain films, CT) show gas bubbles conforming to the contour of the stomach, often in the form of cystic gas pockets.82

Mycobacteria83-85

Gastric infection with Mycobacterium tuberculosis is a rare entity that usually occurs in association with pulmonary tuberculosis. Patients typically present with abdominal pain, nausea and vomiting, gastrointestinal bleeding, fever, and weight loss. Gastric tuberculosis associated with anemia has been reported.85 Gastric tuberculosis may be associated with gastric outlet obstruction or with bleeding from a tuberculous gastric ulcer. Radiographic studies reveal an enlarged stomach with narrowed, deformed antrum with prepyloric ulcerations. Upper endoscopy demonstrates ulcers, masses, or gastric outlet obstruction. Grossly, the stomach may demonstrate multiple small mucosal erosions, ulcers, an infiltrating mass (hypertrophic) form, a sclerosing inflammatory form, acute miliary dissemination, and pyloric obstruction either by extension from peripyloric nodes or by invasion from other neighboring organs. Biopsies show necrotizing granulomas with the presence of acid-fast bacilli, best demonstrated with Kinyoun acid-fast stain. Treatment is discussed in Chapter 107. Although Mycobacterium avium complex (MAC) is a common opportunistic bacterial infection among patients with AIDS, the stomach is rarely involved. Gastric MAC may be associated with a chronic gastric ulcer refractory to conventional antiulcer therapy. Patients may present with fever, night sweats, anorexia, weight loss, diarrhea, abdominal pain, chylous ascites, severe gastrointestinal hemorrhage, or chronic gastric ulcer. Serial CT scans of the abdomen may show mesenteric lymphadenopathy.

Endoscopy may show a chronic gastric ulcer, a coarsely granular duodenal mucosa, or fine white duodenal nodules. Microscopically, the gastric mucosa demonstrates numerous foamy histiocytes containing many acid-fast bacilli. Treatment of MAC is difficult and is discussed in Chapter 33.

Actinomycosis86-87

Primary gastric actinomycosis is a rare, chronic, progressive, suppurative disease characterized by formation of multiple abscesses, draining sinuses, abundant granulation, and dense fibrous tissue. Abdominal actinomycosis is more common and has a predilection for the terminal ileum, cecum, and appendix. The presenting symptoms include fever, epigastric pain, epigastric swelling, abdominal wall abscess with fistula, and upper gastrointestinal bleeding. Radiographic studies frequently suggest a malignant tumor or an ulcer. Endoscopy is suggestive of a circumscribed and ulcerated gastric carcinoma. Grossly, the resected stomach demonstrates a large, ill-defined, ulcerated mass in the wall of the stomach that measures up to 4 cm. Microscopically, multiple abscesses show the infective agent, Actinomyces israelii, a gram-positive filamentous anaerobic bacterium that normally resides in the mouth. A biopsy of a mass containing pus or a biopsy of a draining sinus may reveal actinomycosis. If the disease is recognized only by histologic examination, the prognosis is good. Prolonged (6- to 12-month) high-dose antibiotic treatment with penicillin or amoxicillin is indicated.

Syphilis88-94

The incidence of syphilis in the United States increased 34% from 13.7 to 18.4 cases per 100,000 persons between 1981 and 1989. Several case reports and small series emphasize the importance of the gastroenterologist and pathologist remaining alert to the protean manifestations of syphilis and familiar with the histopathologic pattern of the disease. Gastric involvement in secondary or tertiary syphilis is rarely recognized clinically, and its diagnosis by examination of endoscopic biopsy specimens has been reported infrequently. The features of syphilis in the stomach should be recognized because they can provide a window of opportunity for effective antibiotic therapy before the disease progresses and causes permanent disability. Patients typically present with symptoms of peptic ulcer disease, and the most common gastric complaint is upper gastrointestinal tract bleeding. Other diseases that may mimic gastric syphilis include benign ulcer disease, gastric carcinoma, gastric lymphoma, tuberculosis, and gastric Crohn’s disease. Gastric syphilis in the setting of human immunodeficiency virus (HIV) has been reported.92 The acute gastritis of early secondary syphilis produces the earliest radiologically detectable sign of the disease. Radiographs show a nonspecific gastritis with diffusely thickened folds that may become nodular with or without detectable ulcers. Strictures in the mid-stomach (“hourglass” stomach) may be present (Fig. 51-5A). Endoscopy shows numerous shallow, irregular ulcers with overlying white exudate and surrounding erythema (see Fig. 51-5B). The surrounding mucosa also demonstrates a nodular appearance. Gastroscopy may also demonstrate prominent, edematous gastric folds. Grossly, the stomach may be thickened and contracted and may show multiple serpiginous ulcers. Partial gastrectomy specimens may show compact, thick, mucosal rugae and numerous small mucosal ulcers. Microscopically, biopsies show severe gastritis with dense plasma cell infiltrate in the lamina propria, varying numbers of neutrophils and lymphocytes, gland destruction, vasculitis, and granulomas.

Chapter 51  Gastritis and Gastropathies

B

A

C

Figure 51-5.  Gastric syphilis (luetic gastritis). Film from an upper gastrointestinal series (A) showing a stricture in the mid-stomach (hourglass stomach), with antral deformity. Endoscopic appearance before (B) and 4 weeks after (C) penicillin therapy in another patient with gastric syphilis.

Warthin-Starry silver stain or modified Steiner silver impregnation stain reveals numerous spirochetes. Serum Venereal Disease Research Laboratory (VDRL) and Treponema immunofluorescence studies may be positive, and the Treponema pallidum gene may be detected by the PCR. Treatment with penicillin is highly effective (see Fig. 51-5C).

Other Bacteria95,96

Because approximately 25% of patients with chronic gastritis have no current or past evidence of infection with H. pylori or other Helicobacters such as H. heilmannii, other bacteria have been sought. One gram-negative bacillus, Acinetobacter lwoffi, is a common commensal that is normally not pathogenic in humans but has been proposed to cause gastritis in a manner analogous to H. pylori.84 A case of transient gastritis caused by the gram-positive enterococcus has also been described.85

FUNGI Candidiasis97,98

Fungal contamination of gastric ulcers with Candida species is not uncommon. Data from some studies suggest that fungal colonization in patients with gastric ulcers and chronic gastritis have little clinical significance, whereas others suggest that fungal infection aggravates and per­ petuates gastric ulceration. Endoscopically, gastric ulcers associated with Candida albicans tend to be larger in diameter and are more often suspected to be malignant than typical gastric ulcers. Diffuse superficial erosions may be noted. Fungal colonization of the gastrointestinal tract is frequent in patients with underlying malignancy and in immu-

nocompromised patients who have been treated with antibiotics or glucocorticoids but may occur also in immunocompetent patients. Symptoms are nonspecific. Massive growth of yeast organisms in the gastric lumen (yeast bezoar) is a potential complication of gastric surgical procedures, usually for peptic ulcer disease. Candida infection of the stomach may occur in alcoholic patients who ingest corrosive chemicals such as concentrated sulfuric acid and thiocyanates. Radiologic studies show tiny aphthoid erosions, which represent the earliest detectable radiographic change in gastric candidiasis. Aphthoid ulcers progress to deep linear ulcers. Grossly, the gastric mucosa demonstrates tiny aphthous erosions; widespread punctate, linear ulcerations; or gastric ulcers. Microscopically, the layer of necrotic fibrinoid debris demonstrates yeasts or pseudohyphae. The organisms can be seen in the hematoxylin and eosin stain; however, special stains such as periodic acid–Schiff-diastase stain or Gomori methenamine silver stain may be required. Treatment is usually not necessary, but if symptomatic candidiasis is suspected, fluconazole is reasonable but of unproven efficacy.

Histoplasmosis99-101

Progressive disseminated histoplasmosis is rare, occurring most frequently in the very young or the older adult or in those with immunodeficiency. Although disseminated histoplasmosis can involve any portion of the gastrointe­ stinal tract, gastric involvement is rare. Hypertrophic gastric folds or a mass that mimics a gastric carcinoma may be associated with gastric histoplasmosis or disseminated histoplasmosis has been reported.100 Radiographic studies may demonstrate an annular infiltrating lesion of the

851

852

Section VI  Stomach and Duodenum stomach, and endoscopy may demonstrate enlarged and reddened gastric folds. Biopsy specimens show an intensive infiltration of macrophages containing Histoplasma cap­ sulatum. Gastric histoplasmosis has also been associated with a fatal hemorrhage from a gastric ulcer. Treatment with intravenous amphotericin B is appropriate.

Phycomycosis102-104

Gastric phycomycosis (also called zygomycosis or mucormycosis) is a rare and highly lethal fungal infection. Phy­ comycosis usually affects the paranasal sinuses, central nervous system, or lungs and is rarely confined to the gastrointestinal tract. Risk factors include malnutrition, immunosuppression, antibiotic therapy, and acidosis, usually diabetic ketoacidosis. A nosocomial outbreak of gastric mucormycosis due to contamination of wooden tongue depressors by Rhizopus microspores has been reported. Most patients presented with upper gastrointe­ stinal bleeding.103 Gastric phycomycosis can be classified as invasive or noninvasive (colonization). The former is characterized by deep invasion of the stomach wall and by blood vessel involvement with the fungus. Abdominal pain is the most frequent presenting complaint. In the noninvasive type, the fungus colonizes the superficial mucosa without causing an inflammatory response. Grossly, surgical specimens from affected patients reveal hemorrhagic necrosis involving the mucosa and gastric wall. Microscopically, nonseptate 10- to 20-µm hyphae branched at right angles are present in the tissue and they infiltrate into blood vessel walls. Treatment is resection of the affected necrotic portion of the stomach. Unfortunately, invasive gastric phycomycosis is almost always fatal. 105-106

Aspergillosis

Acute aspergillus gastritis in a case of fatal aplastic anemia and a fatal case of pseudo-membranous gastritis due to this organism have been reported.

PARASITES (see also Chapters 109 and 110) Cryptosporidiosis107-110

Cryptosporidiosis may rarely involve the stomach. Gastric outlet obstruction and antral stricture due to cryptosporidiosis have been reported in patients with AIDS and diarrhea. A case of cryptosporidiosis-associated erosive gastritis in a patient with HIV infection also has been reported. Also, cryptosporidiosis associated with the immunocompromised state and small cell lung cancer has been reported.110

kis. Hundreds of cases of anisakiasis have been diagnosed in Japan, and the number of reported cases in the United States has also increased. The parasite may migrate into the wall of the stomach, small intestine, or colon. Typically, patients present with sporadic epigastric pain or have no symptoms at all. Gastric perforation due to chronic gastric aniskiasis has been reported.120 Misdiagnosis is common. Some patients may experience a mild peripheral eosinophilia. Endoscopy may show firm, yellowish submucosal masses with erosions.119 Radiographic studies may reveal notched-shadow defects suggestive of a gastric tumor. Grossly, the stomach demonstrates multiple erosive foci with hemorrhage and small 5- to 10-mm gastric lesions in the stomach wall. Microscopically, sections of the stomach show a marked eosinophilic granulomatous inflammatory process with intramural abscesses and granulation tissue. The eosinophilic abscess may contain a small worm measuring 0.3 mm in diameter, which can be identified as the larval form. The diagnosis may be confirmed by a serodiagnostic test for human anisakiasis on the patient’s serum when the larvae may no longer be detectable by endoscopy.

Ascariasis121,122

Although gastric ascariasis is rare, patients have experienced chronic, intermittent gastric outlet obstruction caused by roundworms (Ascaris lumbricoides) inhabiting the stomach. Gastric ascariasis has also been associated with upper gastrointestinal hemorrhage with endoscopic examination showing several Ascaris worms in the stomach and duodenum.

Hookworm123

Endoscopic discovery and capture of Necator americanus in the stomach has been reported.123

GRANULOMATOUS GASTRITIDES A variety of granulomatous diseases affect the stomach. Crohn’s disease is the most common (Fig. 51-6) and is discussed later and also in Chapter 111. The differential diagnosis of granulomatous gastritis also includes sarcoidosis, as well as rarer conditions such as xanthogranulomatous

Strongyloidiasis111-115

The stomach is rarely affected by Strongyloides stercoralis. However, the organisms may colonize the intact gastric mucosa and may be associated with a bleeding peptic ulcer. S. stercoralis hyperinfection has been associated with cimetidine therapy in an immunosuppressed patient and was diagnosed by endoscopic gastric biopsy. Diagnosis can be made by endoscopic biopsy, examination of stools, examination of duodenal aspirate, and examination of peripheral smear with elevated eosinophil count. A histologic diagnosis of strongyloidiasis must be taken into consideration when examining gastric and duodenal biopsies in immunocompromised patients.113 Disseminated strongyloidiasis can be rapidly fatal. Treatment is discussed in Chapter 110.

Anisakiasis116-120

Invasive anisakiasis may occur after the ingestion of raw marine fish containing nematode larvae of the genus Anisa-

Figure 51-6.  Histopathology of granulomatous gastritis in a patient with Crohn’s disease. A noncaseating granuloma is present within the lamina propria. (Hematoxylin and eosin, ×200.)

Chapter 51  Gastritis and Gastropathies gastritis, foreign bodies,105 lymphoma,106 Whipple’s disease (see Chapter 106),107 Langerhans cell histiocytosis (gastric eosinophilic granuloma),108 granulomatous vasculitis109 (Churg-Strauss syndrome), and chronic granulomatous disease of childhood.110 An isolated, idiopathic granulo­ matous gastritis also occurs.111 Some of these latter cases may evolve to Crohn’s disease or sarcoidosis over time. Other cases of “idiopathic” granulomatous gastritis appear to be due to H. pylori infection and may resolve, albeit slowly, following appropriate antibiotic therapy.112,113 Idiopathic granulo­matous gastritis can be associated with gastric cancer.114

SARCOIDOSIS 124-137

Gastrointestinal manifestations of sarcoidosis are uncommon (see Chapter 35). Sarcoidosis is a systemic disease, frequently involving the lungs, lymph nodes, skin, and eyes. Diagnosis of sarcoidosis of the stomach cannot be made with confidence in the absence of disease in other organs. The stomach (usually the antrum) is the most common part of the gastrointestinal tract affected in sarcoidosis, being involved in 10% of cases.137 Affected patients, usually in the third to fifth decades of life, typically present with epigastric pain, nausea, vomiting, and weight loss. Occasionally they present with massive hemorrhage. Gastric sarcoidosis may cause pyloric outlet obstruction, achlorhydria, and pernicious anemia. Radiographically, gastric sarcoidosis may mimic the diffuse form of gastric carcinoma (“linitis plastica”) or Ménétrier’s disease. Endoscopy may reveal a narrow distal half of the stomach with multiple prepyloric ulcers or erosions, atrophy, thick gastric folds with a diffuse cobblestone appearance, or normal mucosa associated with microscopic granulomas. Surgical specimens of patients with gastric sarcoidosis show a thickened stomach wall with foci of erosions and ulcers. Microscopically, mucosal biopsies show multiple noncaseating granulomas. However, the presence of noncaseating granulomas in gastrointestinal tissue is a nonspecific finding and special stains should be performed to rule out infections, especially tuberculosis. In some cases it may be difficult to differentiate gastric sarcoidosis from gastric Crohn’s disease or from isolated, idiopathic granulomatous gastritis. Glucocorticoid therapy is the cornerstone of treatment for gastric sarcoidosis. Subtotal gastric resection is reserved for patients with obstruction and severe hemorrhage.

XANTHOGRANULOMATOUS GASTRITIS 138-139

Xanthogranulomatous gastritis is characterized by inflammation of the gastric wall by foamy histiocytes, inflam­ matory cells, multinucleated giant cells, and fibrosis. The destructive inflammatory process may extend into adjacent organs and simulate a neoplasm. Xanthogranulomatous gastritis has been associated with xanthogranulomatous cholecystitis. Xanthogranulomatous gastritis with pseu­ dosarcomatous changes mimicking a neoplasm has been reported.139

DISTINCTIVE GASTRITIDES COLLAGENOUS GASTRITIS 140-148

Subepithelial fibrosis has been reported in the colon (collagenous colitis), small bowel (collagenous sprue), and stomach (collagenous gastritis). Collagenous gastritis is a rare form of gastritis, and fewer than 20 cases have been

1

2 Figure 51-7.  Histopathology of collagenous gastritis (top, Hematoxylin and eosin, ×200; bottom, Masson trichrome, ×400). The subepithelial thickening of the collagen band is apparent. (From Wang HL, Shah AG, Yerian LM, et al. Collagenous gastritis: An unusual association with profound weight loss. Arch Pathol Lab Med 2004; 128:229.)

reported in the literature. Collagenous gastritis may be associated with collagenous colitis, lymphocytic colitis, and celiac disease. Patients may experience intermittent epigastric abdom­ inal pain, hematemesis, hematochezia, anemia, diarrhea, hypotension, or weight loss. Collagenous gastritis in a Korean child who had been receiving treatment for refractory iron deficiency anemia has been reported.144 Upper gastrointestinal barium radiography may demonstrate an abnormal mucosal surface with a mosaic-like pattern in the body of the stomach, corresponding to mucosal nodularity. Endoscopy may reveal multiple diffusely scattered, discrete submucosal hemorrhages; erosions; and nodularity of the body of the stomach along the greater curvature. Biopsy specimens from the body and antrum of the stomach reveal a patchy, chronic, superficial gastritis, focal atrophy, and focal deposition of collagen in the subepi­ thelial region of the lamina propria, which measures from 20 to 75 mm thick (Fig. 51-7). Tiny erosions of the surface epithelium are present, and the inflammatory infiltrate consists of mainly plasma cells, intraepithelial lymphocytes and eosinophils, together with marked hypertrophy of muscularis mucosae.144 Little is known about the etiology, natural history, and proper treatment of this rare condition.

LYMPHOCYTIC GASTRITIS 149-157

Lymphocytic gastritis is characterized by a dense lym­ phocytic infiltration of surface and pit gastric epithelium (Fig. 51-8A). Lymphocytic gastritis is related to an endoscopic form of gastropathy known as varioliform gastritis.

853

854

Section VI  Stomach and Duodenum

A

B

Figure 51-8.  Histopathologic examples of two distinctive forms of gastritis: lymphocytic gastritis (A) and eosinophilic gastritis (B). A, High-power view of the antral mucosa shows numerous dark-staining mononuclear cells with striking intraepithelial lymphocytosis. B, Numerous eosinophils are noted within the lamina propria and within the walls and lumina of the gastric glands. The patient also had peripheral blood eosinophilia. (Hematoxylin and eosin.) (Courtesy Pamela Jensen, MD, Dallas, Tex.)

Lymphocytic gastritis is also seen in H. pylori infection and in celiac disease. Recent findings provide compelling evidence that lymphocytic gastritis may occur as a manifestation of celiac disease, and thus the lymphocytic infiltration of celiac disease may affect gastric epithelial mucous cells. Lymphocytic gastritis in untreated celiac disease may be associated with functional changes such as increased permeability. Gastric biopsies from 10 of 22 patients with diarrhea or malabsorption and small bowel changes characteristic of celiac disease showed striking lymphocytic gastritis. Following institution of a gluten-free diet, lymphocytic gastritis resolves after approximately two years. Lymphocytic gastritis has also been attributed to an atypical host immune response to H. pylori. H. pylori eradication treatment in patients with lymphocytic gastritis causes significant improvement in the gastric intraepithelial lymphocytic infiltrate, corpus inflammation, and dyspeptic symptoms. H. pylori may be the cause of some cases of protein-losing hypertrophic lymphocytic gastritis. The disease may resolve clinically, endoscopically, and pathologically with therapeutic eradication of H. pylori in some patients. The relationship between lymphocytic gastritis and gastric lymphoid hyperplasia, which also is associated with H. pylori, is not yet clear.154 Patients with gastric lymphoma have a significantly increased prevalence of lymphocytic gastritis due to H. pylori. Because intraepithelial lymphocytes are speculated to have a role in the regulation of normal mucosal inflammatory reaction, they may also participate in the pathogenesis of mucosal lymphoma. In a 10-year follow-up study of lymphocytic gastritis, the patients with lymphocytic gastritis also appeared to have a significant increase in the grade of intestinal metaplasia in the corpus mucosa. In another study, lymphocytic gastritis was more prevalent in patients with gastric adenocarcinoma (16 of 30 cases [12.3%]) than in unselected patients undergoing endoscopy (0.8% to 2.5%). Lymphocytic gastritis with anasarca and venous thrombosis at the confluence of the splenic and mesenteric veins has been reported.153 Endoscopy in lymphocytic gastritis shows thick mucosal folds, nodularity, and aphthous erosion, historically known as varioliform gastritis. Gastric biopsies show expansion of the lamina propria by an infiltrate of plasma cells, lymphocytes, and rare neutrophils. These findings

may be seen in the antral mucosa only, body mucosa only, or in antral as well as body mucosa. The surface and superficial pit epithelium shows a marked intraepithelial infiltrate with T lymphocytes, with flattening of the epithelium and loss of apical mucin secretion. Quantification of epithelial lymphocytes revealed 46.5 lymphocytes per 100 epithelial cells in lymphocytic gastritis, compared with 3.5 lymphocytes per 100 cells in normal controls and 5.1 lymphocytes per 100 cells in disease controls including patients with H. pylori gastritis. The immunohistochemistry profile of lymphocytic gastritis in celiac disease and H. pylori infection and the interplay between infection and inflammation has been reported. Only intraepithelial lymphocytes positive for CD3 and CD8 were increased significantly in celiac disease patients with or without H. pylori infection.155

EOSINOPHILIC GASTRITIS 158-164

Eosinophilic gastrointestinal disorders are characterized by marked tissue eosinophilia in the absence of known causes for eosinophilia or other gastrointestinal disorders. These disorders include eosinophilic esophagitis, eosinophilic gastritis, eosinophilic enteritis, and eosinophilic colitis. Eosinophilic gastroenteritis is a rare condition of unknown etiology characterized by peripheral eosinophilia, eosinophilic infiltration of the gastrointestinal tract, and gastrointestinal symptomatology. It is discussed in detail in Chapter 27. The gastric mucosa is frequently involved, and thus eosinophilic gastritis is one of the manifestations of eosinophilic gastroenteritis. Eosinophilic gastroenteritis is classified according to the layer of gastrointestinal tract involved (i.e., mucosal layer disease, muscle layer disease, and subserosal disease). Mucosal involvement may result in abdominal pain, nausea, vomiting, diarrhea, weight loss, anemia, protein-losing enteropathy, intestinal perforation, and iron deficiency anemia secondary to gastrointestinal blood loss. Patients with muscular layer disease generally have obstructive symptoms, and patients with subserosal eosinophilic infiltration develop eosinophilic ascites. Patients with gastric involvement frequently present with pyloric obstruction. Radiographic studies may demonstrate thickened mucosal folds, nodularity, or ulcerations. Endoscopy may reveal normal-appearing mucosa or hyperemic edematous mucosa with surface erosions or prominent gastric folds.

Chapter 51  Gastritis and Gastropathies Eosinophilic gastritis simulating gastric car­cinoma has been reported.161 Gastric mucosal biopsies are critical to the diagnosis and show marked eosinophilic infiltration, eosinophilic pit abscesses, necrosis with numerous neutrophils, and epithelial regeneration (see Fig. 51-8B). Abnormal eosinophilic infiltration is defined as at least 20 eosinophils per highpower field either diffusely or multifocally. A full-thickness surgical biopsy is necessary for the diagnosis of muscle layer disease. As discussed in Chapter 27, patients with disabling symptoms can be effectively treated with glucocorticoids (after other systemic disorders associated with peripheral eosinophilia have been excluded) or possibly with oral sodium cromoglycate. Surgical intervention may be required in patients with obstructive complications or refractory disease. Collagenous colitis associated with eosinophilic gastritis in a four-year-old girl has been reported.

MISCELLANEOUS FORMS OF GASTRITIS GASTRITIS IN INFLAMMATORY BOWEL DISEASE (CROHN’S AND ULCERATIVE COLITIS) 165-177

Crohn’s disease is the most common disease associated with granulomatous gastritis.111 Crohn’s disease involving the stomach is uncommon, however, and almost always occurs together with lower intestinal disease (see Chapter 111). Cases may be isolated to the stomach or the stomach and duodenum. The diagnosis of isolated Crohn’s disease of the stomach should be made with caution,165 and close followup is indicated for the subsequent development of Crohn’s disease elsewhere in the gastrointestinal tract or of other granulomatous diseases such as sarcoidosis. Symptoms are nonspecific and include nausea and vomiting, epigastric pain, anorexia, and weight loss. Radiologic studies show antral fold thickening, antral narrowing, shallow ulcers (aphthae), or deeper ulcers. Involvement of the stomach from adjacent ileal or colonic disease segments is best visualized by radiologic examination. Endoscopy allows better visualization of mucosal defects and is characterized by reddened mucosa, irregularly shaped ulcers, and erosions in a disrupted mucosal pattern. Nodular lesions occur and often reveal the presence of erosions on the top of nodules. An atypical cobblestone pattern may be associated with the nodules surrounded by fissure-like ulceration. In contrast with peptic ulcers, the ulcerations and erosions of Crohn’s disease are frequently serpiginous or longitudinal, rarely round or oval. Ulcerations or erosions associated with Crohn’s disease of the stomach typically are most commonly located in the antrum and the prepyloric region. The microscopic features of surgical specimens of gastric Crohn’s disease can be, but are not always, similar to those in the ileum or colon (see Chapter 111). They include granulomas, transmural chronic inflammation, ulcers, and marked submucosal fibrosis (see Fig. 51-6). Granulomas may be present in endoscopically normal antral mucosa. In the past few years it has been recognized that although Crohn’s patients may have granulomatous gastritis, H. pylori–associated gastritis (HAG) or focal active antral gastritis is even more common.166-175 The majority of pediatric inflammatory bowel disease (IBD) patients (Crohn’s or ulcerative colitis [UC]) have HAG and/or focal active gastritis, although the latter is more prevalent in Crohn’s patients

than in UC patients.174 The focal active gastritis is not due to H. pylori and is accompanied by macrophages in the center of the focal lesion and mast cells at its periphery.169 It is as yet unclear whether the type of gastritis identified in pediatric IBD patients can reliably distinguish Crohn’s colitis from UC (except for granulomatous gastritis, which favors Crohn’s)168 and even less is certain in adult IBD patients. Therapy of gastritis in Crohn’s disease should be driven by gastric symptoms and not solely by demonstration of gastritis on mucosal biopsy. Double-blinded, randomized, controlled clinical trials of pharmacologic agents are lacking in gastric and gastroduodenal Crohn’s disease. Proton pump inhibitors should be the first therapy for symptomatic patients.176,177 The effectiveness of glucocorticoids, immunosuppressive medications, and anti–TNF-α drugs such as infliximab has not been clearly demonstrated. Gastric outlet obstruction refractory to medical therapy can be treated by gastroenterostomy, ideally laparoscopically. Treatment of Crohn’s disease is discussed in more detail in Chapter 111.

GASTRITIS CYSTICA PROFUNDA 178-182

Gastritis cystica profunda is a rare complication of partial gastrectomy with gastrojejunostomy for benign peptic ulcer disease and typically occurs at the site of the gastroenterostomy. Gastritis cystica profunda may also develop in the unoperated stomach, and chronic atrophic gastritis may be a risk factor for it. Radiography and endoscopy typically demonstrate multiple exophytic gastric masses that simulate a malignancy. Endoscopic ultrasound may assist in the diagnosis. Grossly, the gastric mucosal surface demonstrates multiple nodules and exophytic masses. On section, the gastric wall is thick and multiple cysts are present. Microscopically, the mucosa is characterized by foveolar hyperplasia, and cystic glands extend through a disrupted muscularis mucosae into the submucosa and, rarely, the muscularis propria (Fig. 51-9). Gastritis cystica profunda may be associated with gastric stump adenocarcinoma. Removal of this lesion by snare polypectomy after submucosal injection to elevate the lesion has been reported.179 The disease may also coexist with gastric inverted hyperplastic polyp, and the latter may in fact be a variant of gastritis cystica profunda.182 Gastritis cystica profunda, if

Figure 51-9.  Gastritis cystica profunda. Note the cystic dilatation of numerous gastric glands that extend through the muscularis mucosae (arrow), simulating a gastric carcinoma.

855

856

Section VI  Stomach and Duodenum present, should lead to a thorough examination for a synchronous or metachronous gastric cancer, although exact recommendations for surveillance interval are not clear.

GASTRIC GRAFT-VERSUS-HOST DISEASE 183-187 (see also Chapter 34)

Graft-versus-host disease (GVHD) may affect any portion of the gastrointestinal tract; therefore, upper as well as lower gastrointestinal tract biopsies may provide diagnostic information not evidenced in biopsy from a single site. GVHD most often occurs after allogeneic bone marrow transplantation and rarely after solid transplantation. GVHD occurs in an acute and chronic form. Acute GVHD occurs between post-transplant days 21 and 100, whereas chronic GVHD occurs after day 100. The gastrointestinal tract is commonly affected in acute GVHD, especially the small and large intestines, and to a lesser extent the stomach and esophagus. Gastric GVHD is characterized by nausea, vomiting, and upper abdominal pain without diarrhea. Stomach biopsies may be necessary to diagnose GVHD in patients with upper gastrointestinal symptoms but no diarrhea and normal rectal biopsy specimens. The basic pathologic lesion consists of necrosis of single cells (apoptotic bodies) in the crypts of the large and small intestinal mucosa and in the neck region of the gastric mucosa. The necrosis consists of an intraepithelial vacuole filled with karyorrhectic debris and fragments of cytoplasm.

ALLERGIC GASTRITIS 188-190

Whether allergies to certain foods can lead to gastritis and whether the gastritis correlates with upper gastrointestinal symptoms are unclear (see Chapter 9). Children with food allergy as diagnosed by an open elimination challenge test have no higher incidence of gastritis than children without food allergy.188 An exception may be infants who are allergic to cow’s milk protein, in whom hematemesis and endoscopic signs of gastritis are common.189

REACTIVE GASTROPATHIES (ACUTE EROSIVE GASTRITIS) The gastric mucosa may be damaged by a variety of agents or factors that do not produce a significant inflammatory infiltrate. Because of the paucity of inflammatory cells, the mentioned lesions are best referred to as reactive gastropathies, as opposed to acute erosive gastritis. The gastric mucosa in patients who experience a reactive gastropathy demonstrates a spectrum of hemorrhages, erosions, and ulcers. Erosions and ulcers are frequently multiple, and the base of these lesions often stains dark brown owing to exposure to acid. Grossly, most erosions and acute ulcers appear as welldefined hemorrhagic lesions 1 to 2 mm in diameter. If the insult is severe, the mucosa between the lesions is intensely hemorrhagic. Microscopically, an erosion demonstrates superficial lamina propria necrosis. An acute ulcer is an area of necrosis that extends to the muscularis mucosa. Foveolar hyperplasia, the reactive epithelial changes secondary to regeneration of the mucosa (Fig. 51-10), is often associated with glands with atypical nuclei that can be misdiagnosed as dysplasia or carcinoma. The diagnosis of neoplasia in a background of necrosis, cellular debris, and granulation tissue should be made with utmost caution. The biopsy procedure itself may induce tissue hemorrhage; thus, subepithelial hemorrhage should involve more

Figure 51-10.  Histopathology of foveolar hyperplasia. The gastric pits show an elongated, corkscrew appearance.

than one fourth of a biopsy specimen to be considered significant.17

MEDICATIONS AND TOXINS 191-201

Aspirin (even in low daily or less-than-daily doses) and NSAIDs (including cyclooxygenase [COX]-2 selective inhibitors) are the most common causes of reactive gastropathy (see Chapter 52). Oral iron therapy may rarely cause mild endoscopic abnormalities consisting of erythema, small areas of subepithelial hemorrhage, and erosions.191 Oral potassium chloride may also be associated with endoscopic erosions.192 Endoscopic petechiae, erosions, and erythema have been associated with long-term fluoride ingestion.195 Bisphosphonates for osteoporosis or Paget’s disease can also cause gastric erosions, although their clinical significance is uncertain.196,197 These drugs exacerbate gastric damage from NSAIDs such as naproxen as well. Various intensive cancer chemotherapy drugs given to children with leukemia, lymphoma, or solid tumors are associated with a hemorrhagic or erosive gastropathy and histologic evidence of inflammation, but cause and effect have not been clearly established in these ill individuals.198 Reactive gastric epithelial atypia and gastric ulceration may be associated with hepatic arterial infusion chemotherapy for metastatic disease to the liver.199,200 The marked epithelial atypia that results may erroneously be interpreted as carcinoma and lead to unnecessary surgery. Toxin ingestion of heavy metals such as mercury sulfate poisoning may cause an erosive or ulcerative gastropathy with hematemesis.201 Corrosive gastric injuries are discussed in Chapter 25.

ALCOHOL 202-209

After acute alcohol ingestion, subepithelial hemorrhages are seen frequently at endoscopy, typically without prominent mucosal inflammation on biopsy specimens (Fig. 51-11).202 Gastric biopsy specimens obtained from patients with chronic alcoholism have shown a higher prevalence of chronic antral gastritis due to H. pylori, with almost complete normalization of histologic findings after treatment.203,204 The combined effects of alcohol and the NSAID ibuprofen were associated with more gastric mucosal damage by endoscopic assessment than with either agent alone. The combination of alcohol and aspirin also caused more damage in the stomach than either agent alone, though not to a signifi-

Chapter 51  Gastritis and Gastropathies ing the gastric antrum.212 GAVE does not respond as readily as PHG to measures that reduce portal pressure.212

COCAINE 215-217

Gastrointestinal hemorrhage due to erosion throughout the gastric fundus, duodenal bulb has been reported with Gastrointestinal hemorrhage or pyloric cocaine is well described.215-217

diffuse exudative body, antrum, and crack cocaine use. perforation due to

STRESS 218

Erosions of the gastric mucosa may occur rapidly after major physical or thermal trauma, shock, sepsis, or head injury. These are often referred to as stress ulcers and are discussed in Chapters 52 and 53.

RADIATION 218-219 (see also Chapter 39)

Figure 51-11.  Histopathology of alcoholic gastropathy. Hemorrhage is confined to the superficial portion of the mucosa, and there is a paucity of inflammatory cells.

cant degree.205 Alcohol appeared to be an acute triggering factor in 35% of patients admitted to an intensive care unit for massive upper gastrointestinal bleeding in Sweden.206 Chronic alcohol ingestion was related to an increased risk of chronic atrophic gastritis and hypochlorhydria in a study from Poland.207 Recent alcohol ingestion was also found to be a risk factor for gastric erosions and ulcers in cirrhotic patients referred for upper endoscopy.208 Other alcohols, besides ethanol, can injure the stomach. Even the topical application of isopropyl alcohol (rubbing alcohol) used to cool a child with fever has resulted in hemorrhagic gastropathy with hematemesis.209

PORTAL HYPERTENSIVE GASTROPATHY 210-214

As discussed in more detail in Chapters 19 and 90, gastric mucosal lesions are common in portal hypertension, occurring in up to 65% of cirrhotic patients, and represent an important cause of gastrointestinal blood loss.210 Portal hypertensive gastropathy was found in 33% of children after surgery for biliary atresia. Risk factors for portal hypertensive gastropathy were frequent endoscopic treatment of gastroesophageal varices, liver dysfunction, and hypersplenism.213 Biopsies show vascular ectasia and congestion in the mucosal layer without a significant degree of inflammatory infiltrate. Portal hypertensive gastropathy (PHG) is a risk factor not only for upper gastrointestinal bleeding but also for gastroduodenal erosions and ulcers in cirrhotic patients.208 It has been suggested that perturbations in the tissue levels of TNF-α, prostaglandins, endothelin, and nitric oxide/peroxynitrite participate in the vascular congestion and mucosal damage characteristic of PHG.211 As discussed in Chapter 90, lowering portal pressure pharmacologically or by creation of a portosystemic shunt effectively treats PHG and reduces bleeding. Some patients with cirrhosis and portal hypertension have gastric antral vascular ectasia (GAVE), which can bleed and is sometimes difficult to distinguish from PHG involv-

Radiation effects on the stomach depend on the cell kinetics of the gastric mucosa, as well as the dose of the radiation. The gastric mucosal response to radiation is unique, however, in that the most radiosensitive epithelial cells are the differentiated cells (parietal and chief cells) rather than the germinative cells in the mucous neck region. Radiation injury to the stomach can be classified into acute (less than six months) and chronic (more than one year) phases. It is thought that the tolerance level for radiation-induced gastric ulceration is approximately 4500 cGy. After a gastric dose of 5500 cGy or more, 50% of patients will develop clinical evidence of gastric ulcer formation.218 Radiation-induced gastric ulcers are usually solitary, from 0.5 to 2 cm in diameter, and located in the antrum. Massive hemorrhagic gastropathy requiring endoscopically administered therapy to control the bleeding has been reported.219

BILE REFLUX 220-231

Bile reflux into the stomach is common after partial gastrectomy with anastomosis to the duodenum (Billroth I) or jejunum (Billroth II) and after truncal vagotomy and pyloroplasty for peptic ulcer (see Chapter 53). It has even been reported after parietal cell vagotomy.220 Bile reflux gastropathy also may occur after cholecystectomy or sphincteroplasty, which allows the continuous exposure of bile to the duodenum with the potential for duodenogastric reflux. Occasionally, bile reflux gastropathy is observed in adult or pediatric patients who have not had surgery.221,223 Interleukins, particularly IL-8 and perhaps IL-6, may participate in the gastric damage.224,225 Bile reflux contributes to muscosal lesions in the stomach and may facilitate H. pylori colonization in the corpus region.222 Endoscopy shows swelling, redness, erosions, and bile staining of the gastric mucosa. Biopsy specimens show foveolar hyperplasia, dilated cystic glands, atypical glands that may be misdiagnosed as dysplasia or carcinoma, and a paucity of acute and chronic inflammatory cells. Gastric atrophy may result and increase the risk of carcinoma in the gastric stump (see Chapter 54). In fact, bile reflux into the unoperated stomach has been proposed to be a risk factor for intestinal metaplasia in the distal stomach, at the gastroesophageal junction (cardia), and in the distal esophagus (Barrett’s esophagus).226-228 Unfortunately, bile-diverting procedures performed because of severe bile gastropathy do not reverse gastric atrophy or intestinal metaplasia.229 It may be worthwhile, at the time of the original gastric surgery for gastric cancer or peptic ulcer, to construct a 30-cm Roux-en-Y limb176 or perform a 10- to 12-cm isoperistaltic jejunal interposition230 to prevent bile gastropathy and subsequent metaplastic changes.

857

858

Section VI  Stomach and Duodenum In selected, previously unoperated patients with primary bile gastropathy, surgery using a Roux-en-Y choledochojejunostomy without gastric resection has been successful.221 In unoperated patients with bile gastropathy following cholecystectomy, the proton pump inhibitor rabeprazole and sucralfate were equally effective in relieving symptoms and improving gastroscopic evidence of mucosal damage as compared with observation alone (no placebo was given).231 At present, lacking definitive studies, medical therapy should precede surgical therapy for bile gastropathy occurring in the unoperated stomach such as it does spontaneously or after cholecystectomy or biliary sphincterotomy.

ISCHEMIA 232-237

Chronic ischemic gastropathy may occur secondary to chronic mesenteric insufficiency and can be reversed after a revascularization operation.191,192 Chronic ischemic gastropathy, as well as chronic ischemic gastric ulcers, may also occur in association with atheromatous embolization.193,194 Athletes involved in intense physical activity, especially long-distance running, may experience recurrent ischemic gastropathy and chronic gastrointestinal bleeding with anemia.195,196

PROLAPSE 238-239

The mucosa of the gastric cardia may prolapse into the esophageal lumen during retching and vomiting.238 The resulting mechanical injury to the cardia has been proposed to be a cause of upper gastrointestinal hemorrhage, but this association has been questioned.239 Esophagoscopy may demonstrate the prolapsed gastric mucosa. The congested mucosa may show erosions and superficial ulcerations.

LINEAR EROSIONS IN A HIATAL HERNIA (CAMERON LESIONS) 240-241

Linear gastric erosions in a hiatal hernia are discussed in Chapters 19, 24, and 52.

AGING GASTROPATHY 242,243

Normal aging is associated with impaired gastric mucosal defense against injury in animals and humans.242 Two proteins have been implicated in aging gastropathy: PTEN (phosphatase and tensin homolog deleted on chromosome 10), which is overexpressed with aging, and survivin, which is underexpressed.243

HYPERPLASTIC GASTROPATHIES 244-251 MÉNÉTRIER’S DISEASE AND HYPERPLASTIC, HYPERSECRETORY GASTROPATHY

Hyperplastic gastropathy is a rare condition characterized by giant gastric folds associated with epithelial hyperplasia. Two clinical syndromes have been identified: Ménétrier’s disease and a variant of it referred to as hyperplastic, hypersecretory gastropathy, and Zollinger-Ellison syndrome, which is discussed in Chapter 32. Figure 51-12A and B demonstrates enlarged gastric folds in these conditions. Ménétrier’s disease is typically associated with proteinlosing gastropathy (see Chapter 28) and with hypochlor­ hydria, whereas the hyperplastic, hypersecretory variant is associated with increased or normal acid secretion and parietal and chief cell hyperplasia, with or without excessive gastric protein loss. Other more common conditions can also cause enlarged gastric folds or protein-losing gastropathy including gastric

neoplasm (lymphoma, carcinoma), granulomatous gastri­ tides, gastric varices, infectious gastritis (particularly H. pylori and CMV58), eosinophilic gastritis, and ZollingerEllison syndrome. The enlarged gastric folds in Ménétrier’s disease are due to foveolar cell hyperplasia, edema, and variable degrees of inflammation. Patients may present with weight loss, epigastric pain, vomiting, anorexia, dyspepsia, hematemesis, and positive fecal occult blood tests. Most patients with a clinical syndrome associated with hyperplastic gastritis showed histology typical for the syndrome; however, clinical-histologic concordance was not absolute.247 The mechanism responsible for the low gastric acid secretion is unclear, but it could be related to overexpression of transforming growth factor-α (TGF-α), a ligand for epidermal growth factor receptor (EGFR), a receptor tyrosine kinase.244 Although mucus hypersecretion is often seen in Ménétrier’s disease, abnormalities in mucins have not been consistent in a few cases that have been examined.245,246 Ménétrier’s disease may be associated with hypertrophic lymphocytic gastritis and a carcinoid-like syndrome due to increased gastric mucosal production of prostaglandin E2 concentrations. Ménétrier’s disease may be self-limited and may completely resolve in patients younger than 10 years of age and when it occurs in the postpartum period. The cause of Ménétrier’s disease of childhood may be infection with CMV and activation of TGF-α. The risk of developing carcinoma in association with Ménétrier’s disease is an open question, and some authors question whether there is a significant risk. Of the 200 cases of Ménétrier’s disease reported in the literature, 30 (15%) have been associated with carcinoma. The mucosa of patients with Ménétrier’s disease demonstrates irregular hypertrophic folds that involve the entire gastric body. The mucosa also demonstrates a swollen, spongy appearance subdivided by creases, creating a picture similar to cerebral convolutions. A polypoid variant of Ménétrier’s disease that resembles multiple hyperplastic gastric polyps has been described (see Chapter 54). Gastric resections from patients with Ménétrier’s disease typically show large polypoid gastric folds or large cerebriform gastric folds with antral sparing (see Fig. 51-12C). In the absence of a gastrectomy, a full-thickness gastric mucosal biopsy is required to adequately assess the gastric histology in patients with hyperplastic gastropathy. The predominant microscopic feature of Ménétrier’s disease and hyperplastic, hypersecretory gastropathy is foveolar hyperplasia with cystic dilation (see Fig. 51-12D). The parietal and chief cells may be decreased and replaced by mucous glands. Inflammation in hyperplastic gastropathies is variable and may be absent. The etiology of Ménétrier’s disease is unknown, although some cases have undoubtedly been infections with CMV or H. pylori. Genetic factors have recently been emphasized after the report of the disorder in identical twin men who presented at ages 29 and 35, respectively.247 Hyperplasia of surface mucous cells may be due to enhanced EGFR signaling in the gastric mucosa due to local overproduction of TGF-α.244 Ideal treatment of hyperplastic gastropathy is unclear because the condition is rare and controlled trials are lacking. Spontaneous resolution may occur, especially in children. It is likely that some cases, particularly in children, were actually cases of CMV gastritis (see earlier discussion). Ganciclovir has been used successfully in children with Ménétrier’s disease associated with CMV gastritis.249 H. pylori infection should be treated, if present, and the entire syndrome may resolve.250 Symptoms may improve with antisecretory agents (histamine-2 [H2] receptor antagonists,

Chapter 51  Gastritis and Gastropathies

A

B

C

D

Figure 51-12.  Radiologic and histopathologic examples of hyperplastic gastropathy with giant gastric folds. A, Zollinger-Ellison syndrome (barium radiograph). B, Ménétrier’s disease (barium radiograph). C, Total gastrectomy specimen in a patient with Ménétrier’s disease (right: body, revealing hyperplastic mucosa and cerebriform rugal folds; left: antrum, with relative sparing). D, Histopathology of Ménétrier’s disease showing enlarged folds with foveolar hyperplasia, cystically dilated glands, and minimal gastritis.

anticholinergic agents, proton pump inhibitors), especially if the patient has Zollinger-Ellison syndrome or normo­ gastrinemic hyperplastic, hypersecretory gastropathy. It has been suggested that H2 blockers and anticholinergics reduce gastric protein loss by strengthening intercellular tight junctions. Some patients with Ménétrier’s disease have responded to glucocorticoids, octreotide, antifibrinolytic agents, or monoclonal antibody against the EGFR.251 Partial or total gastric resection is reserved for severe complications such as refractory or recurrent bleeding, obstruction, severe hypoproteinemia, or cancer development.

ZOLLINGER-ELLISON SYNDROME (see Chapter 32) DIFFERENTIAL DIAGNOSIS OF GASTRITIS AND GASTROPATHY The most important disorders that can simulate gastritis and gastropathy are gastric polyps (non-neoplastic and neoplastic) and gastric neoplasms such as adenocarcinoma and lymphoma (see Chapter 54).252,253 Although CT criteria have been useful in distinguishing benign gastritis or gastropathy

from gastric malignancy,254 endoscopy and gastric biopsy with review by an expert pathologist are the most useful diagnostic procedures. B cell clonality using advanced PCR technology can also help distinguish gastric marginal zone lymphomas from chronic gastritis.255 At the other end of the spectrum, many patients with gastritis have a normal endoscopic appearance,256 so the differential diagnosis of gastritis also includes functional dyspepsia (see Chapter 13), in which case the gastric biopsy is usually normal.

TREATMENT AND PREVENTION OF GASTRITIS AND GASTROPATHY 257 The treatment of these disorders depends on the underlying etiology (if one can be identified). In countries where the incidence of H. pylori infection is declining, the prevalence of chronic gastritis will decline as well (see Chapter 50). It has been shown in a case-control study performed in a region of southeastern China with a very high prevalence of chronic gastritis and gastric cancer that ingestion of green tea reduced the risk of gastritis and gastric cancer by close to 50%.257

859

860

Section VI  Stomach and Duodenum KEY REFERENCES

Carlson AP, Chan JWH, Ketai LH, Demarest GB. Emphysematous gas­ tritis in a severely burned patient: Case report and literature review. J Trauma 2007; 62:765-7. (Ref 82.) Chia JKS, Chia AY. Chronic fatigue syndrome is associated with chronic enterovirus infection of the stomach. J Clin Pathol 2008; 61:43-8. (Ref 72.) Coffey RJ, Washington MK, Corless CL, Heinrich MC. Ménétrier disease and gastrointestinal stromal tumors: Hyperproliferative disorders of the stomach. J Clin Invest 2007; 117:70-80. (Ref 244.) D’Elios MM, Bergman MP, Azzurri A, et al. H+,K+-ATPase (proton pump) is the target autoantigen of Th1-type cytotoxic T cells in autoimmune gastritis. Gastroenterology 2001; 120:377-86. (Ref 45.) DeBlock CEM, DeLeeuw IH, VanGaal LF. Autoimmune gastritis in type 1 diabetes: A clinically oriented review. J Clin Endocrinol Metab 2008; 93:363-71. (Ref 47.) Fox JG, Wang TC. Inflammation, atrophy, and gastric cancer. J Clin Invest 2007; 117:60-9. (Ref 9.) Hummel M, Oeschger S, Barth TFE, et al. Wotherspoon criteria combined with B cell clonality analysis by advanced polymerase chain reaction technology discriminates covert gastric marginal zone lymphoma from chronic gastritis. Gut 2006; 55:782-7. (Ref 255.) Insko EK, Levine MS, Birnbaum BA, et al. Benign and malignant lesions of the stomach: Evaluation of CT criteria for differentiation. Radiology 2003; 228:166-71. (Ref 254.)

Kaur G, Raj SM. A study of the concordance between endoscopic gastritis and histological gastritis in an area with a low background prevalence of Helicobacter pylori infection. Singapore Med J 2002; 43:90-2. (Ref 1.) Marshall JK, Thabane M, James C. Randomized active and placebocontrolled endoscopy study of a novel protected formulation of oral alendronate. Dig Dis Sci 2006; 51:864-8. (Ref 196.) Redeen S, Peterson F, Jonsson KA, et al. Relationship of gastroscopic features to histological findings in gastritis and Helicobacter pylori infection in a general population sample. Endoscopy 2003; 35:94650. (Ref 256.) Ricuarte O, Gutierrez O, Cardona H, et al. Atrophic gastritis in young children and adolescents. J Clin Pathol 2005; 58:1189-93. (Ref 23.) Rubio CA, Nesi G, Zampi GC, et al. Gastric ciliated metaplasia. A study of 3406 gastrectomy specimens from dwellers of the Atlantic and the Pacific basins. J Clin Pathol 2005; 58:605-10. (Ref 32.) Tarnawski A, Pai R, Deng X et al. Aging gastropathy-novel mechanisms: Hypoxia, up-regulation of multifunctional phosphatase PTEN, and proapoptotic factors. Gastroenterology 2007; 133:1938-47. (Ref 243.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

52 Peptic Ulcer Disease Nimish Vakil

CHAPTER OUTLINE Epidemiology  861 Population-Based Studies  861 Time Trends  861 Risk Factors  862 Helicobacter pylori Infection  862 Aspirin and Nonsteroidal Anti-inflammatory Drugs  862 Other Ulcerogenic Drugs  862 Other Risk Factors  862 Pathogenesis  863 Gastric Acid Secretion  863 Helicobacter pylori Infection  863 Gastric and Duodenal Mucosal Defense Mechanisms  863 Helicobacter pylori-Negative, NSAID-Negative Ulcers  864 Other Causes of Ulcer Disease in the Stomach and Duodenum  864

An ulcer in the gastrointestinal (GI) tract may be defined as a break in the lining of the mucosa, with appreciable depth at endoscopy or histologic evidence of involvement of the submucosa. Erosions are breaks in the surface epithelium that do not have perceptible depth. The term peptic ulcer disease is used broadly to include ulcerations and erosions in the stomach and duodenum from a number of causes. This is because pepsin, which is proteolytic in acidic solution, plays a major role in causing the mucosal breaks regardless of the cause of the inciting agent (e.g., Helicobacter pylori, aspirin, or a nonsteroidal anti-inflammatory drug [NSAID]). Decades of research focused on the role of acid secretion and the effects of stress, personality type, and genetics in the pathogenesis of ulcer disease. The discovery of histamine-2 (H2) receptors1 and development of H2receptor antagonist drugs, and the subsequent development of proton pump inhibitor drugs led to major changes in the management of peptic ulcer disease. The discovery of H. pylori and its treatment led to dramatic changes in the prevalence and recurrence of peptic ulcer disease, transforming peptic ulcer from a chronic recurrent disease to a curable one.2 H. pylori infection remains an important cause of peptic ulceration in the developing world. In the developed world, the use of NSAIDs has emerged as a leading cause of peptic ulcer disease, especially in the aging population in whom these drugs are often prescribed. Through all of these developments, the role of acid and pepsin in the genesis and perpetuation of mucosal injury remains a unifying aspect of the pathogenesis of peptic ulcer disease.

EPIDEMIOLOGY The epidemiology of peptic ulcer disease has undergone a remarkable change in the past century. The incidence of

Gastrinoma with or without Multiple Endocrine Neoplasia Syndrome, Type I  864 Systemic Mastocytosis  864 Miscellaneous Disorders  864 Clinical Features  864 Diagnosis  864 Endoscopy  864 Contrast Radiography  865 Diagnostic and Management Strategies for Patients with Suspected Peptic Ulcer Disease  866 Complications  867 Hemorrhage  867 Penetration and Perforation  867 Obstruction  868 Treatment  868

duodenal ulcer and gastric ulcer has declined in parallel with the decline in H. pylori prevalence, likely as a result of improved sanitary conditions and a safer food and water supply. The risk of developing peptic ulcer disease and the risk of dying from peptic ulcer disease increased in successive cohorts born between 1840 and 1890 and then declined thereafter.3 A peak in the incidence of gastric ulcer in the first half of the 19th century and a subsequent peak in the incidence of duodenal ulcer in the second half of the 19th century remain unexplained, although a number of theories have been proposed, among them the widespread adoption of smoking after the commercial manufacture of cigarettes in a setting of widespread H. pylori infection.

POPULATION-BASED STUDIES

In northern Sweden, a random sample of 1001 people underwent upper GI endoscopy after filling out symptom questionnaires.4 The prevalence of peptic ulcer disease in this sample was 4.1%, with 20 gastric ulcers and 21 duo­ denal ulcers. In a prospective Danish study of 2416 subjects, the 11-year cumulative incidence of peptic ulcer was 2.9%: 1.6% for duodenal ulcer, 1.3% for gastric ulcer, and 0.04% for combined ulcers.5 In countries with a high prevalence of H. pylori infection, the ratio between duodenal and gastric ulcers may be quite different. In a case-control study from Shanghai, China, in which the prevalence of H. pylori infection was 76%, recurrent or new peptic ulcers occurred in 3.6% of the population over 2 years and 85% of the ulcers were duodenal.6

TIME TRENDS

There has been a significant decline in mortality from peptic ulcer disease over time in most age groups.7,8 A notable exception is older adults, in whom peptic ulcer bleeding remains a life-threatening condition. A study based on the U.S. National Discharge survey reported that from 1992 to 1999, the annual rate of hospitalization for peptic ulcer

861

862

Section VI  Stomach and Duodenum disease declined 20%, from 205/100,000 population to 165/100,000 population. Mortality also declined 22%, from 7.7/100,000 to 6/100,000, respectively. Sales of acidinhibitory drugs correlated with a decrease in peptic ulcer disease hospitalizations and mortality,9 although this correlation does not prove causality. The observed decline in ulcer-related mortality was probably related to improvements in the general health of the population and the availability of effective treatment for peptic ulcer disease. In an analysis of the Canadian Institute of Health Information database, the prevalence of acute nonvariceal upper GI bleeding (largely ulcer-related bleeding) declined by 31%, from 77/100,000 population to 53/100,000 population over the 10-year period from 1993 to 2003.10 The need for surgical intervention also declined but the mortality rate remained unchanged. In the United Kingdom, overall hospital admission rates for peptic ulcer disease have also declined, as has mortality, but admission rates for peptic ulcer hemorrhage and perforation have increased.11 Peptic ulcer bleeding is most often seen in older adults, with 68% of patients presenting older than age 60, and 27% older than age 80.12

RISK FACTORS The principal risk factors of peptic ulcer disease are H. pylori infection and NSAID use (Fig. 52-1). However, some patients with peptic ulcer disease have neither of these risk factors.

HELICOBACTER PYLORI INFECTION

(see Chapter 50) H. pylori is a gram-negative bacillus that is uniquely adapted to life in the stomach. It is a major cause of peptic ulcer disease and accounts for a large proportion of peptic ulcers in countries where H. pylori infection is highly prevalent such as in Asia. In the United States and in western Europe,

None known

None known

ZES, other NSAID use

H. pylori infection

ZES, other NSAID use

H. pylori infection

the original estimates that H. pylori infection was the cause of 90% or more of duodenal ulcers, and 60% or more of all gastric ulcers have been lowered by the declining prevalence of H. pylori infection.13 It is estimated that close to 70% of duodenal ulcers are related to H. pylori in Western populations. For example, in an analysis of patients who participated in H. pylori eradication trials in the United States, the proportion of patients with peptic ulcer disease and H. pylori infection was 73%.14 In Rochester, New York, the prevalence of H. pylori infection in patients with duodenal ulcer disease was 61%.15

ASPIRIN AND NONSTEROIDAL ANTIINFLAMMATORY DRUGS (see also Chapter 51)

Aspirin is increasingly used in the prevention of cardiovascular disease.16,17 Aspirin and clopidogrel are frequently used in combination in patients who have had ischemic cardiac events or in patients who have recently had a stent placed in the coronary arteries.18 NSAIDs are used by approximately 11% of the U.S. population on a regular basis.19 This frequency is likely to increase as the population ages. Regular use of NSAIDs increases the odds of gastrointestinal bleeding five- to six-fold compared with persons not taking NSAIDs.20 Serious ulcer-related complications occur in 1% to 4% of NSAID users and NSAID-related complications are thought to account for 100,000 admissions every year.21 NSAID users who also take aspirin are at an especially high risk for complications. In a population-based study from Denmark, the odds ratio for GI bleeding if a person was taking low-dose aspirin was 2.6, and this risk increased to 5.6 in patients who were also taking an NSAID.22 In Spain, the death rate attributed to NSAIDaspirin use was 15.3/100,000. Up to one third of all NSAIDaspirin–related deaths in that study were attributable to low-dose aspirin use.23 H. pylori and NSAIDs may have a synergistic role in causing peptic ulcer disease. In a meta-analysis, the odds ratios for the development of a peptic ulcer in patients with H. pylori infection or NSAID use were 4.05 and 2.99, respectively, but the odds ratio increased significantly to 15.4 if both factors were present.24 The risk factors for peptic ulcer disease among patients taking NSAIDS and their risk ratios are listed in the next chapter (see Table 53-1).

OTHER ULCEROGENIC DRUGS

Deep ulcers and perforations of the stomach and duodenum have been described in cocaine and methamphetamine users, presumably due to mucosal ischemia.25 Bisphosphonates have also been associated with gastroduodenal ulceration,26 although esophageal injury with bisphosphonates in clinically more of a concern (see Chapter 45). There is little if any risk for peptic ulcer disease in patients taking glucocorticoids.27 In combination with NSAIDS, however, glucocorticoids increase the risk of peptic ulcer disease above the risk with NSAIDs alone.28

OTHER RISK FACTORS

Duodenal Gastric Figure 52-1.  Pie charts depicting conditions associated with peptic ulcer disease. The percentages shown are rough approximations based on studies from Western countries. The relative contributions of Helicobacter pylori infection and NSAID use to peptic ulcer vary considerably among different populations and, within populations, vary with age and socioeconomic status. Also, the separation depicted in this figure is somewhat artificial because NSAID use and H. pylori infection often coexist. NSAID, non-steroidal anti-inflammatory drug; ZES, Zollinger-Ellison syndrome.

Smoking has been implicated in the pathogenesis of peptic ulcer disease for decades, but its importance as a risk factor has declined after the discovery of H. pylori. A populationbased study evaluated the risk factors for peptic ulcer disease in 2416 Danish adults who were interviewed between 1982 and 1994.29 As expected, H. pylori seropositivity was a significant risk factor for ulcer disease; smoking increased the risk of peptic ulcer only in H. pylori–infected subjects. A large body of literature suggests that smoking

Chapter 52  Peptic Ulcer Disease may predispose to peptic ulcer disease, but H. pylori infection remains a confounder that was not addressed in earlier studies. It is noteworthy that cigarette smoking does not increase the risk of recurrent ulceration once H. pylori has been eradicated, suggesting that smoking may only play a role in infected subjects.30 The role of alcohol remains uncertain. Alcoholic beverages stimulate gastric acid production. Moreover, direct application of high concentrations of alcohol to the gastric mucosa causes demonstrable mucosal injury. In the Danish study alluded to previously,29 intake of spirits increased the risk of peptic ulcer disease in H. pylori– infected patients. With regard to diet, ulcer prevalence rates differ considerably in the north of India, where the principal cereal in the diet is wheat, and in the south, where rice is the predominant cereal. However, many other potential confounders were not accounted for in these populations.31 An association between the ingestion of spicy foods and peptic ulcer disease is weak, at best.32 Emotional stress was proposed as a major cause of ulcer disease or as a precipitant of ulcer complications,33 and much was written about its relationship to peptic ulcer disease prior to the description of H. pylori. Personality types and psychological profiles have been proposed to be linked to ulcer, but much of this literature is confounded by the lack of information on H. pylori infection. Welldocumented descriptions of an increase in ulcer disease after natural calamities such as earthquakes suggest that emotional stress among those not physically injured may play a role in triggering overt manifestations of peptic ulcer disease, especially in individuals who may be otherwise predisposed to ulcer (e.g., patients infected with H. pylori).33,34

PATHOGENESIS GASTRIC ACID SECRETION

Gastric acid and pepsin have long been considered the principal inciting agents in the pathogenesis of peptic ulcer disease.35 The control of gastric acid secretion in health and disease has been reviewed in detail36 and is discussed in Chapter 49.

HELICOBACTER PYLORI INFECTION

(see also Chapters 50 and 51) Most patients who have a chronic infection with H. pylori have a pan-gastritis in which the body and the antrum are equally involved. Gastritis results in inhibition of acid secretion, and various mechanisms have been proposed.37,38 These include direct inhibition of the parietal cell by lipopolysaccharides or toxins secreted by H. pylori or an indirect effect through stimulation of cytokines caused by the inflammation.1 Due to the decreased output of acid, these individuals with pan-gastritis do not usually develop ulcer disease related to H. pylori infection.39 The reduced acid production in H. pylori–infected individuals has been suggested as a factor that may protect against the development of reflux disease and its complications such as Barrett’s esophagus (see Chapters 43 and 44). Another pattern of H. pylori infection affects approximately 10% to 20% of patients with chronic infection and consists of an antrum-predominant gastritis (see Chapters 50 and 51). In these individuals, through a series of steps

not well understood but possibly involving reduced somatostatin concentrations in the antrum, basal and mealstimulated acid secretion often increases. The increased acid output from the stomach results in increased acid delivery to the duodenum that can result in gastric metaplasia in the duodenal bulb. Some investigators believe that the metaplastic epithelium then becomes infected with H. pylori from the stomach, resulting in focal duodenitis, sometimes followed by ulcer formation. In contrast with patients with duodenal ulcer disease who often have increased acid secretion, patients with gastric ulcer typically have normal or decreased acid production, suggesting that the mechanism for the development of ulceration is a failure in the gastric mucosal protective mechanisms, described in more detail below. Acid antisecretory medication heals ulcers in patients with gastric ulcer and hypochlorhydria because acid changes the balance in favor of mucosal defense factors that restore mucosal integrity (see Chapter 53).

GASTRIC AND DUODENAL MUCOSAL DEFENSE MECHANISMS

As described in Chapter 49 the gastric mucosa has multiple defense mechanisms to protect it from digestion by acid and pepsin.40 These include the gastric surface epithelium,41-43 the mucus/phospholipid and bicarbonate barrier (mucus layer),44,45 epithelial cell renewal and regeneration,46,47 mucosal blood flow and the alkaline tide, and prostaglandin production.48-51 Prostaglandins also stimulate mucus, bicarbonate, and phospholipid production. Many prostaglandins also increase mucosal blood flow and stimulate epithelial regenerative processes. Inhibition of these protective effects by NSAIDs increases the likelihood of injury to the epithelium and decreases its ability to respond and regenerate. Cyclooxygenase-1 (COX-1) and COX-2 are the enzymes responsible for the synthesis of prostaglandins. COX-1 is expressed in the stomach and is responsible for the maintenance of the integrity of gastric epithelium and the mucous barrier. COX-2 is not expressed in the healthy stomach but is rapidly expressed in response to the cytokines generated by inflammatory processes. Conventional NSAIDs such as ibuprofen inhibit the COX-1 and the COX-2 enzymes. It is believed that COX-1 inhibition reduces prostaglandin synthesis, which leads to a reduction in mucosal blood flow, hypoxia, and a reduction in mucosal defense. Neutrophilendothelial interactions then occur as a result of the vascular disturbances and neutrophil activation. In experimental studies, COX-1 inhibition alone is not sufficient to cause ulceration.52 The inhibition of COX-1 up-regulates the expression of COX-2 that suppresses the neutrophil endothelial interaction that is stimulated by COX-1 inhibition. Inhibition of COX-1 and COX-2 enzymes is therefore important in the generation of gastric injury. Some studies have suggested that a prostaglandinindependent mechanism may also contribute to damage by NSAIDs. One such mechanism is increased leukotriene production that occurs because arachidonic acid metabolism shifts to the alternative 5-lipooxygenase pathway when the COX-1 pathway is inhibited.40 The duodenum, like the stomach, secretes bicarbonate that neutralizes acid arriving into the duodenum.53 Decreased duodenal bicarbonate secretion has been reported in patients with duodenal ulcer.54 A protective role has also been postulated for pancreatic juice but the results have been conflicting.

863

864

Section VI  Stomach and Duodenum HELICOBACTER PYLORI-NEGATIVE, NSAID-NEGATIVE ULCERS

As the prevalence of H. pylori infection declines in the United States, a growing proportion of patients with peptic ulcers who have no evidence of H. pylori infection and have no history of aspirin or NSAID use have emerged. Eradication of H. pylori generally cures peptic ulcer disease and prevents recurrence, but a small number of patients do not heal their ulcers or suffer an ulcer recurrence despite successful eradication of H. pylori. The exact prevalence of H. pylori–negative, NSAID-negative (idiopathic) peptic ulcers is unknown. The pathogenesis of H. pylori–negative, NSAIDnegative (idiopathic) ulcers remains uncertain. One theory that has been proposed is that some idiopathic ulcers may occur from H. pylori colonization of the duodenum.55 These patients may test negative with conventional endoscopic tests for H. pylori because these tests focus on detecting H. pylori in gastric biopsies. Crohn’s disease, lymphoma, and other disorders that cause ulcers in the stomach or duodenum (see next section) should be distinguished from H. pylori–negative, NSAIDaspirin–negative ulcers.

OTHER CAUSES OF ULCER DISEASE IN THE STOMACH AND DUODENUM GASTRINOMA WITH OR WITHOUT MULTIPLE ENDOCRINE NEOPLASIA SYNDROME, TYPE I

(see Chapter 32) Gastrin-secreting tumors are an important cause of acid hypersecretion due to the sustained drive for acid secretion. This disorder may result in multiple ulcerations in the stomach or duodenum that are refractory to conventional treatment and are often associated with a chronic diarrhea. One fourth of all patients with Zollinger-Ellison syndrome have an autosomal dominant disorder characterized by pancreatic endocrine tumors, hyperparathyroidism, and pituitary adenomas.56

SYSTEMIC MASTOCYTOSIS (see Chapter 32)

This is an uncommon condition in which multiple ulcers may occur in the stomach or duodenum with infiltration of the mucosa with mast cells that can be recognized histologically.57 Secretion of histamine by the mast cells is thought to result in the excessive stimulation of acid production through the histamine receptor.

MISCELLANEOUS DISORDERS

Ulcerations in the upper GI tract may be a manifestation of Crohn’s disease, which can cause ulcerations anywhere in the gastrointestinal tract (see Chapter 111). Associations between peptic ulcers and α1-antitrypsin deficiency, chronic lung disease, and chronic renal failure have been described. Many of these studies predate the routine evaluation for H. pylori in patients with peptic ulcer disease, and some of the observed associations may be questioned. For example, a recent study of α1-antitrypsin–deficient patients with duodenal ulcer found that all the patients were infected with H. pylori.58 Another study reported a higher than expected rate of ulcer recurrence in patients with renal failure following successful eradication of H. pylori.59

CLINICAL FEATURES The “classic” symptoms attributed to peptic ulcer disease included a burning pain the epigastrium relieved by antac-

ids. These assumptions were challenged by the advent of endoscopic examinations that made it clear that the same symptoms were often present in patients with no visible abnormalities at endoscopy (nonulcer dyspepsia). In patients with NSAID-related ulcer disease, pain is often absent. Thus, classic symptoms are neither sensitive nor specific for ulcer disease. The physical examination is usually normal in patients with uncomplicated peptic ulcer disease. Epigastric tenderness is neither sensitive nor specific for ulcer. A systematic review evaluated, for the patient presenting with upper abdominal symptoms, the utility of the clinical examination and various computer models in predicting if peptic ulcer disease is present.60 Some studies evaluated the accuracy of gastroenterologists, some evaluated primary care physicians, and one evaluated both.61-66 The studies included 4684 patients of whom 802 (17%) had peptic ulcer disease (Fig. 52-2). The positive likelihood ratio of accurately diagnosing peptic ulcer by clinical features was 2.9 for gastroenterologists, which was somewhat better than the performance of primary care physicians (likelihood ratio, 2.2). The negative likelihood ratio for accurately excluding peptic ulcer by clinical features was similar in gastroenterologists and primary care phy­sicians (0.62). Computer models for the prediction of peptic ulcer disease were also not highly accurate, with a positive likelihood ratio of 1.9 and a negative likelihood ratio of 0.34.

DIAGNOSIS ENDOSCOPY

Endoscopy is the current reference standard for diagnosis of peptic ulcer disease. Its main limitation is its high cost in some countries such as the United States. The decision to perform endoscopy in a patient suspected of having peptic ulcer disease is based on a number of factors. Patients presenting with complications of peptic ulcer disease such as bleeding need endoscopic evaluation to allow an accurate diagnosis and for the administration of endoscopic therapy (see Chapters 19 and 53). The presence of “alarm” features such as weight loss or recurrent vomiting may prompt concern for malignancy (Table 52-1). Although endoscopy is regarded as the standard for a diagnosis of peptic ulcer disease (Figs. 52-3 and 52-4), small ulcers may be missed at endoscopy.67 The proportion of ulcers that may be missed at endoscopy is uncertain because comparative studies of contrast radiography and endoscopy were performed many years ago and technologic innovations in the quality and resolution of endoscopes have

Table 52-1  Alarm Features in Patients with Suspected Peptic Ulcer Disease* Age older than 55 years with new-onset dyspepsia Family history of upper gastrointestinal cancer Gastrointestinal bleeding, acute or chronic, including unexplained iron deficiency Jaundice Left supraclavicular lymphadenopathy (Virchow’s node) Palpable abdominal mass Persistent vomiting Progressive dysphagia Unintended weight loss *These features should prompt an upper endoscopy and often other testing to establish a definitive diagnosis.

Chapter 52  Peptic Ulcer Disease Primary care physicians Heikkinen et al Hansen et al DDSG

Positive likelihood ratio (95% CI)

Negative likelihood ratio (95% CI)

2.1 (1.4-3.2) 2.1 (1.7-2.6) 2.2 (1.6-3.0)

0.76 (0.61-0.91) 0.58 (0.45-0.72) 0.54 (0.37-0.73)

Gastroenterologists Bytzer et al Fjosne et al DDSG Barenys et al

1.9 (1.6-2.3) 3.9 (3.3-4.5) 3.4 (2.3-4.8) 2.8 (2.2-3.6)

0.50 (0.41-0.60) 0.47 (0.40-0.53) 0.54 (0.38-0.70) 0.45 (0.37-0.54)

Computer models Barenys et al Bytzer et al Johannessen et al Mann et al Holdstock et al Numans et al

3.0 (2.5-3.8) 1.4 (1.2-1.5) 2.0 (1.8-2.2) 1.6 (1.2-1.9) 1.7 (1.6-1.9) 2.2 (1.9-2.5)

0.28 (0.22-0.36) 0.62 (0.46-0.80) 0.19 (0.12-0.30) 0.37 (0.18-0.72) 0.58 (0.29-0.49) 0.31 (0.22-0.44)

Combined

2.2 (1.9-2.6)

0.45 (0.38-0.53)

0.1 1.0 10.0 Positive likelihood ratio (95% CI)

0.1 1.0 10.0 Negative likelihood ratio (95% CI)

Figure 52-2.  Positive and negative likelihood ratios of different approaches to diagnosing peptic ulcer disease. Each square represents an individual study. The size of the square is a measure of the size of the study and the horizontal line through the square indicates a graphical representation of the 95% CI of that study. For the combined analysis, the diamond and vertical dashed line indicate the pooled positive or negative likelihood ratio, with the left and right ends of the diamond indicating the pooled 95% CI. CI, confidence interval; DDSG, Danish Dyspepsia Study Group. (Reproduced from Moayyedi P, Talley N, Fennerty B, Vakil N. Can the clinical history distinguish between organic and functional dyspepsia? JAMA 2006; 295:1566-76.)

Figure 52-4.  Endoscopic view of a duodenal ulcer in a patient with a positive rapid urease test for Helicobacter pylori. There was no history of nonsteroidal anti-inflammatory drug use.

Figure 52-3.  Endoscopic view of a clean based antral gastric ulcer in a patient taking a non­steroidal anti-inflammatory drug. Tests for infection with Helicobacter pylori were negative.

ulcers at initial endoscopy can be found to be malignant at a subsequent examination, and cancers so detected are typically at an early stage with the possibility of a curative resection.

CONTRAST RADIOGRAPHY

improved the yield of endoscopy. Biopsies should be taken from the edges of a gastric ulcer because of the risk of malignancy. In patients with gastric ulcers it has been customary to repeat the endoscopy after approximately eight weeks of treatment to ensure that the ulcer has healed (see Chapter 53). A small number of chronic, nonhealing gastric ulcers can be identified as being malignant at the repeat endoscopy even though the initial biopsies had revealed no evidence of a malignancy. Although conflicting reports on the utility and cost-effectiveness of performing a repeat endoscopy after approximately eight weeks in patients with benign-appearing gastric ulcers have appeared in the literature, recent studies support the use of a repeat endoscopy in this setting.68,69 Up to 4% of apparently benign gastric

Contrast radiography of the upper GI tract, also referred to as the barium meal or an upper GI series, can often demonstrate a peptic ulcer (Fig. 52-5). In historical studies, contrast radiography has performed as well as endoscopy in the diagnosis of ulcer disease when the contrast examinations are performed and interpreted by experts.67 Much has changed since then. Contrast radiography of the upper GI tract is now frequently performed by technicians, and the availability of personnel trained in the interpretation of barium studies has decreased. A further disadvantage to contrast radiography is radiation exposure, which can be substantial. Endoscopy offers the obvious advantage of allowing mucosal biopsy, which can be used to diagnose H. pylori infection and to rule out malignancy in chronic gastric ulcers. Contrast radiography therefore has limited utility in modern practice.

865

866

Section VI  Stomach and Duodenum Dyspepsia not due to GERD not associated with NSAID use

Figure 52-5.  Radiograph from an upper gastrointestinal series showing a benign gastric ulcer. Note the smooth, symmetrical folds radiating to the ulcer crater, which appears to project outside the lumen of the stomach. (Courtesy Mark Feldman, MD, Dallas, Tex.)

DIAGNOSTIC AND MANAGEMENT STRATEGIES FOR PATIENTS WITH SUSPECTED PEPTIC ULCER DISEASE Upper abdominal symptoms that raise the question of peptic ulcer disease are common in clinical practice and account for 2% to 5% of visits in family practice.70 These symptoms, consisting of pain or discomfort in the upper abdomen are frequently referred to as dyspepsia. Due to the high cost of subjecting all dyspeptic individuals to endoscopy, noninvasive alternative strategies have been proposed as an initial step in the management of suspected peptic ulcer disease (Fig. 52-6).71 Economic models suggested that these noninvasive strategies could reduce the costs of managing peptic ulcer disease substantially.72 Subsequent clinical trials demonstrated the clinical efficacy of the approach and the initial strategy proposed was to perform noninvasive testing for infection, followed by antimicrobial treatment directed against H. pylori if the patient tested positive.73,74 This “test and treat” strategy offers a potentially curative treatment for patients who have H. pylori related ulcer disease and may cure a small proportion of patients with nonulcer dys­pepsia. As the prevalence of H. pylori infection declines in Western populations, other empirical strategies can be considered. Because acid inhibition plays a significant role in the management of patients with upper GI symptoms, empirical treatment with a proton pump inhibitor has been proposed as an alternative to routine endoscopy. Cost-models have suggested that when the prevalence of H. pylori infection in the population falls to the 10% to 15% range, an empirical course of therapy with a proton pump inhibitor may be a reasonable initial strategy.75 A cost simulation model in the United States76 supported the empirical management of patients, as recommended in the American Gastroenterological Association dyspepsia guidelines (see Fig. 52-6).71 A randomized controlled clinical trial compared empirical testing and treatment for H. pylori with empirical proton pump inhibitor therapy and concluded that both were equally cost-effective in the initial man­agement of dyspeptic patients, with the choice of which strategy to be taken left to a discussion between patient and physician.77

Age >55 or alarm features present (See Table 52-1.)

Age ≤55 No alarm features (See Table 52-1.)

EGD

Test for Helicobacter pylori

Negative

Positive

PPI trial 4-6 weeks

Treat for H. pylori

Fails

Fails

PPI trial 4 weeks Fails

Reassurance Reassess diagnosis

Consider EGD Figure 52-6.  American Gastroenterological Association guideline for the management of dyspepsia. This is the current management approach for patients with suspected peptic ulcer disease. EGD, esophagogastroduodenoscopy; GERD, gastroesophageal reflux disease; NSAID, nonsteroidal anti-inflammatory drug; PPI, proton pump inhibitor. (Adapted from Talley NJ. American Gastroenterological Association medical position statement: evaluation of dyspepsia. Gastroenterology 2005; 129:1753-5.)

The incidence of upper GI malignancies rises with age, and thus current empirical management strategies are generally reserved for younger patients with upper abdo­ minal symptoms. The age beyond which endoscopy should become routine is controversial and to a substantial degree depends on the epidemiology of gastric cancer in the population under consideration. In Western populations, upper GI cancer is rare in young individuals and therefore an age cutoff of 50 or 55 years is used.78 Older patients presenting with new-onset upper abdominal symptoms suggestive of peptic ulcer disease should be referred for endoscopy. Guidelines in the United Kingdom have not set an age cutoff for early endoscopy, and empirical therapy is preferred at all ages. In Asia and eastern Europe, an earlier age cutoff may be reasonable because the risk of gastric cancer is substantially higher than in Western nations and setting the age threshold too high may lead to a delayed diagnosis of a treatable cancer.79 The principal concern of any empirical treatment strategy for treating dyspepsia is that an underlying malignancy will be missed. To increase the odds of detecting an underlying malignancy, an age-based strategy has been

Chapter 52  Peptic Ulcer Disease Study A

0.67 (0.22, 0.96)

Study B

0.71 (0.44, 0.90)

Study C

0.56 (0.45, 0.66)

Study D

0.83 (0.61, 0.95)

Study E

0.50 (0.16, 0.84)

Studies combined

0.67 (0.54, 0.83) 0.1

A

0.2

0.5

1

Sensitivity (95% confidence interval)

Study A

0.95 (0.93, 0.97)

Study B

0.68 (0.66, 0.69)

Study C

0.86 (0.85, 0.86)

Study D

0.39 (0.37, 0.41)

Study E

0.59 (0.57, 0.62)

Studies combined

0.66 (0.55, 0.79) 0.2

B

0.5 Specificity (95% confidence interval)

proposed in which younger dyspeptic patients are treated empirically and older dyspeptic patients receive early endoscopy. Although there is some controversy about the age at which routine endoscopy for dyspepsia should begin, an age cutoff of 50 to 55 years is proposed in recent guidelines.80 A systematic review and meta-analysis of alarm symptoms in predicting malignancy (as opposed to peptic ulcer and nonulcer dyspepsia) has raised questions about their accuracy.81 The sensitivity of alarm symptoms for serious underlying pathology in 15 studies including 57,363 patients ranged from 0% to 83%, with considerable heterogeneity between studies (Fig. 52-7). Specificity of alarm symptoms for malignancy varied from 40% to 83%. Although current management schemes still list alarm features as an indication for early endoscopy in patients with undiagnosed upper abdominal pain (see Fig. 52-4), further research is necessary in this area. The issue of evaluation of undiagnosed dyspepsia is also discussed in Chapter 13.

COMPLICATIONS HEMORRHAGE

Hemorrhage from a peptic ulcer occurs when the ulcer crater erodes a blood vessel. It is a major cause of morbidity

1

Figure 52-7.  Meta-analysis plots for the utility of clinical alarm features in the diagnosis of upper gastrointestinal malignancies. A, Sensitivity of the presence of any alarm feature in detecting upper gastrointestinal (GI) malignancy in five studies (A-E). When the studies were combined, the sensitivity was only 67%. B, Specificity of the presence of any alarm feature in detecting upper GI malignancy. The combined specificity was only 66%. (Reproduced from Vakil N, Moayyedi P, Fennerty MB, Talley NJ. Limited value of alarm features in the diagnosis of upper gastrointestinal malignancy: Systematic review and meta-analysis. Gastroenterology 2006; 131:390-401.)

in the United States with an annual cost that exceeds 2 billion dollars.82,83 GI hemorrhage is an important cause of morbidity and mortality in older adutls.84 The typical presentation is with melena or hematemesis although a small proportion of patients with bleeding may present with hematochezia. Bleeding peptic ulcers are discussed in more detail in Chapters 19 and 53. Strategies to prevent NSAIDrelated bleeding in patients at high risk have been developed85 and are also discussed in more detail in Chapter 53, as is the treatment of acute ulcers occurring in critically ill patients in intensive care units (stress ulcers).

PENETRATION AND PERFORATION 86-93

As an ulcer deepens it can burrow into adjoining structures (penetration) or rupture into the peritoneal cavity (perforation). Ulcer-related penetration and perforation are less common than upper GI bleeding as a complication of peptic ulcer disease and are more common in older adults. Ulcers located posteriorally in the duodenum may penetrate the pancreas, whereas ulcers located more anteriorally located ulcers may penetrate the liver or biliary tract.88-90 There are no typical presenting signs that identify a penetrating ulcer, but pain radiating to the back is often cited as a symptom of a posterior penetrating ulcer. Computed tomography (CT) may be helpful in establishing a diagnosis.91 Free ulcer perforation with peritonitis is a medical emergency. The patient typically presents with severe gener­

867

868

Section VI  Stomach and Duodenum alized abdominal pain and signs of peritonitis (see Chapters 10 and 37). CT usually establishes the diagnosis.92 Perforation has a high mortality, particularly in older adults. A population-based study from three Danish counties reported that among 2061 patients hospitalized with peptic ulcer perforation, 38% were current NSAID users.93 The 30-day mortality associated with ulcer perforation in this population was 25% overall, and 35% among current NSAID users.

OBSTRUCTION

Chronic scarring, usually in the antrum of the stomach, can lead to a disorder characterized by recurrent vomiting and a narrowing of the H. pylori channel, which is usually visualized endoscopically or on contrast radiography. The symptoms may be insidious, manifesting as reflux disease that is difficult to control or dyspepsia until the narrowing becomes pronounced when vomiting and early satiety become prominent features. Benign pyloric stenosis due to recurrent peptic ulcer disease was a common condition in the past but is rather rare now. It is therefore important to rule out a neoplastic process, which is a more common cause of gastric outlet obstruction than benign pyloric stenosis due to ulcer disease.94

TREATMENT Treatment of peptic ulcer disease and its complications is discussed in Chapter 53.

KEY REFERENCES

Aro P, Storskrubb T, Ronkainen J, et al. Peptic ulcer disease in a general adult population: The Kalixanda study: A random population-based study. Am J Epidemiol 2006; 163:1025-34. (Ref 4.) Black J, Duncan W, Durant D. Definition and antagonism of histamine H2 receptors. Nature 1972; 236:365-90. (Ref 1.) El-Omar EM. Mechanisms of increased acid secretion after eradication of Helicobacter pylori infection. Gut 2006; 55:144-6. (Ref 37.)

Laine L, Takeuchi K, Tarnawski A. Gastric mucosal defence and cytoprotection: Bench to bedside. Gastroenterology 2008; 135:41-60. (Ref 40.) Lewis JD, Bilker WB, Brensinger C, et al. Hospitalization and mortality rates from peptic ulcer disease and GI bleeding in the 1990s: Relationship to sales of nonsteroidal anti-inflammatory drugs and acid suppression medications. Am J Gastroenterol 2002; 97:25409. (Ref 9.) Marshall B, Warren J. Unidentified curved bacilli in the stomachs of patients with gastritis and peptic ulcer. Lancet 1984; 1:1311-15. (Ref 2.) Moayyedi P, Talley N, Fennerty B, Vakil N. Can the clinical history distinguish between organic and functional dyspepsia? JAMA 2006; 295:1566-76. (Ref 60.) Ofman JJ, Etchason J, Fullerton S, et al. Management strategies for Helicobacter pylori–seropositive patients with dyspepsia (clinical and economic consequences). Ann Intern Med 1997; 126:280-91. (Ref 72.) Rosenstock S, Jorgensen T, Bonnevie O, Andersen L. Risk factors for peptic ulcer disease: A population based prospective cohort study comprising 2416 Danish adults. Gut 2003; 52:186-93. (Ref 29.) Schubert M, Peura D. Control of gastric acid secretion in health and disease. Gastroenterology 2008; 134:1842-60. (Ref 36.) Sonnenberg A. Time trends of ulcer mortality in non-European countries. Am J Gastroenterol 2007; 102:1101-7. (Ref 7.) Spiegel BM, Vakil NB, Ofman JJ. Dyspepsia management in primary care (a decision analysis of competing strategies) Gastroenterology 2002; 122:1270-85. (Ref 75.) Talley NJ, Vakil N. Practice Parameters Committee of the American College of Gastroenterology. Guidelines for the management of dyspepsia. Am J Gastroenterol 2005; 100:2324-37. (Ref 80.) Targownik L, Nabalamba A. Trends in management and outcomes of acute nonvariceal upper gastrointestinal bleeding: 1993-2003. Clin Gastroenterol Hepatol 2006; 4:1459-66. (Ref 10.) Vakil N, Moayyedi P, Fennerty BM, Talley N. Limited value of alarm features in the diagnosis of upper gastrointestinal malignancy: Systematic review and meta-analysis. Gastroenterology 2006; 131:390401. (Ref 81.) Wilcox C, Allison J, Benzuly K, et al. Consensus Development Conference on the Use of Nonsteroidal Anti-Inflammatory Agents, Including Cyclooxygenase-2 Enzyme Inhibitors and Aspirin. Clin Gastroenterol Hepatol 2006; 4:1082-9. (Ref 21.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

53 Treatment of Peptic Ulcer Disease Francis K. L. Chan and James Y. W. Lau

CHAPTER OUTLINE Overview  869 Antisecretory and Acid-Neutralizing Agents  869 Mucosa-Protective Agents  871 Ulcers Associated with Helicobacter pylori Infection  872 Duodenal Ulcer  872 Gastric Ulcer  872 Role of Maintenance Therapy  872 Ulcers Associated with Nonsteroidal Anti-inflammatory Drugs  872 Active Ulcers  872 Ulcer Prophylaxis  873

Chapter 52 reviews the epidemiology, clinical features, diagnosis, and complications of peptic ulcer disease. This chapter focuses on its treatment.

OVERVIEW For more than a century, peptic ulcer was considered a chronic, incurable disorder characterized by frequent exacerbations and remissions. The discovery of the link between Helicobacter pylori and peptic ulcer by Marshall and Warren1 in the mid-1980s revolutionalized the concept and treatment of peptic ulcer. Now there is overwhelming evidence to support H. pylori infection as the most important cause of duodenal and gastric ulcers worldwide. Curing the infection not only heals peptic ulcer but also prevents ulcer relapse.2 Peptic ulcer is, in fact, a curable infectious disease. Beside H. pylori–related peptic ulcers, use of nonsteroidal anti-inflammatory drugs (NSAIDs) and low-dose aspirin is another major cause of peptic ulcer complications particularly among older adults.3 Co-therapy with antiulcer drugs and the replacement of conventional nonselective NSAIDs with NSAIDs selective for cyclooxygenase-2 (COX-2 selective NSAIDs) have become alternative treatments for patients who are at risk for peptic ulcer disease. Data suggest that COX-2 selective NSAIDs and some nonselective NSAIDs increase the risk of serious cardiothrombotic events. Prescribing NSAIDs therefore requires a careful assessment of individual patients’ gastrointestinal (GI) and cardio­ vascular risks. As discussed in the preceding chapter, due to the declining prevalence of H. pylori infection, the proportion of patients with H. pylori–negative, NSAID-negative idiopathic ulcers is growing. In the United States, the reported propor-

The Role of Helicobacter pylori Infection in Ulcer Disease Associated with NSAID Use  876 Recommendations for the Prevention of NSAID-Induced Ulcer Complications  876 Refractory Ulcers  877 Stress-Related Mucosal Injury  877 Treatment of Complications of Peptic Ulcer Disease  878 Hemorrhage  878 Perforation  883 Obstruction  884

tion of H. pylori–negative, NSAID-negative idiopathic ulcers is between 20% and 30%.4,5 It has been argued that as the incidence of H. pylori–related ulcers falls, a greater proportion of H. pylori–negative, NSAID-negative idiopathic ulcers will be seen.6 Of interest, a prospective cohort study has demonstrated a four-fold rise in the absolute incidence of idiopathic bleeding ulcers, and the risk of recurrent ulcer bleeding in these patients is high.7 Thus, long-term prophylaxis with antisecretory drugs for idiopathic bleeding ulcers is advisable (see later), although this recommendation is not evidence based.

ANTISECRETORY AND ACID-NEUTRALIZING AGENTS

Before the discovery of H. pylori as a causal factor in peptic ulcer disease, drugs designed to reduce gastric acidity were the mainstays of treatment. Antisecretory or antacid therapy is not routinely required for patients with uncomplicated H. pylori ulcers in whom the bacterium is successfully eradicated, but these classes of drugs do play an important role in promoting healing of large ulcers, preventing early recurrent bleeding after endoscopic therapy for bleeding ulcers, and reducing the risk of ulcer relapse associated with NSAIDs. Specific therapies for peptic ulcer are discussed in the following sections.

Antacids

Mechanisms of Action When Peterson and coworkers8 showed that a liquid antacid preparation of magnesium-aluminum hydroxide (approximately 1000 mmol HCl neutralizing capacity per day) was more effective than placebo for hastening the healing of duodenal ulcer, it was thought antacids promoted ulcer healing by neutralizing gastric acid. However, later studies showed that far smaller doses of antacids (neutralizing

869

870

Section VI  Stomach and Duodenum capacity 120 mmol HCl per day) had virtually identical efficacy for healing peptic ulcerations.9 The precise mechanisms by which antacids hasten the healing of peptic ulcerations are not clear. Although antacids have been shown to be superior to placebo in healing peptic ulcers, their efficacy in ulcer healing is limited. In one study, the healing rates for gastric ulcer after 6 weeks were 67% in the antacids group and 25% in the placebo group.9 Adverse Effects For the magnesium-containing agents, the most common side effect is diarrhea. In contrast, antacids that contain aluminum hydroxide primarily, and those that contain calcium, may cause constipation. All antacids must be used with caution, if at all, in patients who have chronic kidney disease. In such patients, magnesium-containing agents can cause hypermagnesemia; the use of calcium carbonate can cause hypercalcemia and alkalosis and further renal impairment (milk-alkali syndrome), and aluminum hydroxide antacids can cause aluminum neurotoxicity.10

Histamine-2 Receptor Antagonists

Mechanisms of Action Four histamine-2 (H2) receptor antagonists are available— cimetidine (Tagamet), ranitidine (Zantac), famotidine (Pepcid), and nizatidine (Axid). These compounds are competitive inhibitors of histamine-stimulated acid secretion, although famotidine appears to have some component of noncompetitive inhibition as well.11 All four agents suppress basal acid output as well as acid output stimulated by meals (see Chapter 49). Pharmacokinetics The H2 receptor antagonists are well absorbed after oral dosing, and their absorption is not affected by food. Peak blood levels are achieved within 1 to 3 hours after an oral dose. These drugs cross the blood-brain barrier and the placenta.12,13 After oral administration, cimetidine, ranitidine, and famotidine undergo first-pass hepatic metabolism, which reduces their bioavailability by 35% to 60%. In contrast, nizatidine does not undergo first-pass metabolism, and its bioavailability approaches 100% with oral dosing. When administered in the evening, the drugs are especially effective in suppressing nocturnal acid output.14 All four H2 receptor antagonists are eliminated by a combination of renal excretion and hepatic metabolism. After intravenous administration, in contrast, all four agents are eliminated principally through renal excretion. For cimetidine and famotidine it is recommended that the doses be cut in half in patients whose creatinine clearance is 15 to 30 mL/minute. For nizatidine and ranitidine, the dose should be halved if the creatinine clearance is less than 50 mL/minute. Dialysis does not remove substantial amounts of the H2 receptor antagonists, so dose adjustments for dialysis are not necessary. Liver failure has been found to prolong the half-life of cimetidine, but dose reductions are generally not needed for patients with hepatic failure unless it is accompanied by chronic kidney disease.11 Tolerance to the antisecretory effects of H2 receptor antagonists appears to develop quickly and frequently.15 The mechanisms that mediate tolerance to the antisecretory effects of H2 receptor antagonists are not entirely clear. Adverse Effects The H2 receptor antagonists are a remarkably safe and welltolerated group of agents. The overall incidence of side effects is less than 4%, and serious side effects are decidedly uncommon. One meta-analysis of randomized clinical trials

concluded that the overall rate of adverse effects reported for the H2 blockers did not differ significantly from that for placebo.16 Nevertheless, a number of untoward effects have been described, primarily in anecdotal reports and uncontrolled series. Cimetidine has weak antiandrogenic activity that occasionally can cause gynecomastia and impotence.17 Myelosuppression is an uncommon, presumably idiosyncratic side effect of the H2 receptor antagonists. In one large series of patients with bone marrow transplants, however, ranitidine was implicated as a possible cause of myelosuppression in 5%.18 The contribution of ranitidine to the bone marrow suppression in such patients is not clear, but pending further data, it seems prudent to avoid the use of H2 receptor antagonists in bone marrow transplant recipients. Drug Interactions Both cimetidine and ranitidine bind to the hepatic cytochrome P-450 (CYP) mixed-function oxidase system, and this binding can inhibit the elimination of other drugs that are metabolized through the same system, including theophylline, phenytoin, lidocaine, quinidine, and warfarin.19 Famotidine and nizatidine have no significant avidity for the CYP system, and these agents do not appear to have any important drug interactions.

Proton Pump Inhibitors

Mechanisms of Action The proton pump inhibitors (PPIs) are a class of drugs that decrease gastric acid secretion through inhibition of H+,K+-ATPase, the proton pump of the parietal cell (see Chapter 49). Five PPIs are used widely as antisecretory agents—omeprazole (Prilosec), esomeprazole (Nexium; the S optical isomer of omeprazole), lansoprazole (Prevacid), pantoprazole (Protonix), and rabeprazole (Aciphex). These agents are prodrugs that must be activated by acid to inhibit the H+,K+-ATPase. However, PPIs as prodrugs are acid-labile compounds that must be protected from degradation by stomach acid after oral administration.20 Pharmacokinetics The PPIs are well absorbed after oral dosing. Absorption of the enteric-coated agents may be erratic, and peak serum concentrations are not achieved until 2 to 5 hours after oral administration. Although the plasma half-life of the PPIs is short (about 2 hours), the duration of acid inhibition is long (about 24 hours) as a result of covalent binding of the active metabolite to the H+,K+-ATPase. All PPIs undergo significant hepatic metabolism. Because there is no direct toxicity from PPIs, dose adjustments are not required even in patients with significant renal or hepatic impairment. However, there are significant genetic polymorphisms for one of the CYP isoenzymes involved in PPI metabolism, CYP2C19. Approximately 3% of white persons and 15% of Asians are deficient in CYP2C19. This polymorphism has been shown to substantially raise plasma levels of omeprazole, lansoprazole, and pantoprazole but not those of rabeprazole.21,22 As a result of their requirement for concentration and activation in acidic compartments, the PPIs bind predominantly to those proton pumps that are actively secreting acid. Thus, the efficacy of the PPIs for inhibiting acid secretion is limited if they are administered during the fasting state, when only approximately 5% of the stomach’s proton pumps are active. With meal stimulation, in contrast, 60% to 70% of the proton pumps actively secrete acid. Thus, the PPIs are most effective if they are administered immediately before meals. For once-daily dosing, it is recom-

Chapter 53  Treatment of Peptic Ulcer Disease mended that the PPIs be taken immediately before breakfast.23 Unlike H2 receptor antagonists, tolerance to the antisecretory effects of PPI therapy has not been seen during shortterm investigations. Adverse Effects The PPIs are a remarkably safe and well-tolerated group of agents. The most commonly reported side effects are headache and diarrhea, yet the rate at which patients experience these symptoms does not differ significantly from that for patients treated with placebo.24 PPIs and other antisecretory agents cause hypergas­ trinemia by inhibiting gastric acid secretion (see Chapter 49). In addition to stimulating acid secretion, gastrin has been shown to have trophic effects on the GI entero­ chromaffin-like (ECL) cells. Female rats in which protracted hypergastrinemia has been induced by PPIs develop ECL cell hyperplasia and gastric carcinoid tumors.25 However, there are no reports of gastric carcinoid tumors attributable to PPIs in humans. Even in patients with Zollinger-Ellison syndrome who have severe hypergas­ trinemia, carcinoid tumors are uncommon and occur predominantly in patients with multiple endocrine neoplasia type I (MEN-I).26 Some data suggest that the long-term administration of PPIs to patients who are infected with H. pylori might accelerate the development of gastric atrophy.27 However, these early observations have not been conformed by subsequent studies.28 The U.S. Food and Drug Administration (FDA) advisory group concluded that the available data did not establish such an effect, and did not recommend routine testing for and treatment of H. pylori before initiation of PPI therapy.29 Data from observational studies have found that PPIs increase the risk of osteoporosis-related fracture. The strength of the association increases with increasing duration and dose of PPI therapy.30,31 The mechanisms underlying such an association are unknown. Drug Interactions The elevation of gastric pH induced by the PPIs can affect the absorption of a number of medications. However, this antisecretory action rarely has clinically important effects on drug pharmacokinetics, except when the PPIs are given with ketoconazole or digoxin.32 Ketoconazole requires stomach acid for absorption, and this drug may not be absorbed effectively after PPIs have inhibited gastric acid secretion. Conversely, an elevated gastric pH facilitates the absorption of digoxin, resulting in higher plasma levels of this agent. If a patient requires both PPI and antifungal therapy, it is recommended that an agent other than ketoconazole be chosen. For patients treated concomitantly with PPIs and digoxin, clinicians should consider monitoring plasma digoxin levels. Because the PPIs are metabolized by the CYP system, there is potential for them to alter the metabolism of other drugs that are eliminated by CYP enzymes. Among the available PPIs, omeprazole appears to have the greatest potential for such drug interactions and has been shown to delay the clearance of warfarin, diazepam, and phenytoin.33 Lansoprazole, pantoprazole, and rabeprazole do not appear to interact significantly with drugs metabolized by the CYP system. Even with omeprazole, however, clinically important drug interactions are uncommon. Evidence is accumulating that several PPIs may inhibit the activation of clopidogrel to its active metabolite, thus impairing the antiplatelet effect of clopidogrel, with adverse cardiovascular outcomes.

MUCOSA-PROTECTIVE AGENTS Sucralfate

Mechanisms of Action Sucralfate (Carafate) has demonstrated efficacy (similar to that of the H2 receptor antagonists) in healing duodenal ulcer when given in a dose of 1 g four times daily.34 The drug has demonstrated efficacy in the treatment of gastric ulcer as well, but sucralfate has not been approved by the FDA for this indication. Sucralfate is a complex metal salt of sulfated sucrose. Although the sucralfate molecule contains aluminum hydroxide, the agent has little acidneutralizing capacity. When exposed to gastric acid, the sulfate anions can bind electrostatically to positively charged proteins in damaged tissue, thereby forming a protective barrier that may prevent further acid-peptic attack. Other proposed beneficial effects of sucralfate are enhancement of mucosal prostaglandin levels, stimulation of mucus and bicarbonate secretion, binding of mucosa-irritating bile salts, binding of epidermal growth factors, and promotion of angiogenesis.34 Pharmacokinetics Less than 5% of the sucralfate administered is absorbed owing to its poor solubility.34 The drug is excreted in the feces. The high aluminum content causes a small but significant rise in serum and urine aluminum levels within 2 days. In patients with normal renal function, the minor amounts of aluminum absorption with short-term therapy are of no clinical significance. Toxicity and Drug Interactions Because of the lack of systemic absorption, sucralfate appears to have no systemic toxicity. The effect on the disposition of aluminum in the body has not been adequately studied in patients with chronic kidney disease. Sucralfate is best avoided in this population. The drug can bind to a number of medications, including phenytoin and warfarin, reducing their absorption. Important drug interactions appear to be rare, however, and can be avoided entirely if sucralfate is administered at a time separate from other medications.34

Bismuth

Mechanisms of Action Two colloidal preparations of bismuth have been most commonly used, colloidal bismuth subcitrate and bismuth subsalicylate (e.g., Pepto-Bismol). These agents have some efficacy in healing peptic ulcers, but the mechanisms underlying this therapeutic effect are not clear.35 The bismuth forms complexes with mucus that appear to coat ulcer craters. Effects on increasing mucosal prostaglandin synthesis and bicarbonate secretion also have been proposed, and bismuth has documented antimicrobial actions against H. pylori. Bismuth has been approved by the FDA for use in combination with other agents for the treatment of H. pylori infection (see Chapter 50). Pharmacokinetics Bismuth is largely unabsorbed and is excreted in the feces. Colonic bacteria convert bismuth subcitrate and bismuth subsalicylate to bismuth sulfide, which turns the stools black. Trace amounts of bismuth are absorbed in the upper GI tract. Absorbed bismuth is slowly excreted in the urine for three months or longer.35 Toxicity Short-term, standard-dose therapy with bismuth appears to carry little risk of toxicity. However, there is the potential

871

872

Section VI  Stomach and Duodenum for bismuth neurotoxicity if the agent is given for extended periods in high dosage, especially in patients with chronic kidney disease.35

Prostaglandin E Analogs

Mechanisms of Action Endogenous prostaglandins, including prostaglandin E2 (PGE2), regulate mucosal blood flow, epithelial cell proliferation, epithelial restitution, mucosal immunocyte function, mucus and bicarbonate secretion, and basal acid secretion.36 There is substantial evidence that the ulcerogenic effect of an NSAID correlates well with its ability to suppress prostaglandin synthesis.37 Misoprostol, a prostaglandin E1 analog, is the only prostaglandin analog approved by the FDA for the prevention of NSAID-induced ulcer disease. The drug not only enhances mucosal defense mechanisms but also inhibits gastric acid secretion. After binding to the prostaglandin receptor on the parietal cell, misoprostol inhibits gastric acid secretion in a dosedependent manner that is mediated through inhibition of histamine-stimulated cyclic adenosine monophosphate (cAMP) production.38 It has been shown that misoprostol significantly reduces nocturnal, basal, and meal-stimulated acid secretion at a standard therapeutic dose, although the effect is not as potent as that of other classes of antisecretory agents.39 Pharmacokinetics Misoprostol is well absorbed after oral administration. The plasma concentration peaks at about 30 minutes, with a serum half-life of approximately 1.5 hours. The drug has no effect on hepatic cytochrome P-450. Misoprostol metabolites are excreted in the urine, but dose reduction is unnecessary in patients with chronic kidney disease.40 Toxicity Dose-related diarrhea is the most common side effect, occurring in up to 30% of patients and limiting the usefulness of misoprostol.41 Diarrhea is related to prostaglandin-induced increases in intestinal water and electrolyte secretion or acceleration of intestinal transit time. Administration of misoprostol with food may reduce diarrhea. Prostaglandins stimulate uterine smooth muscle. Misoprostol is therefore contraindicated in women who may be pregnant.

ULCERS ASSOCIATED WITH HELICOBACTER PYLORI INFECTION The discovery of H. pylori and its role in peptic ulcer disease has revolutionized the approach to management. Before this discovery, annual ulcer recurrence rates were as high as 80%, often requiring long-term maintenance therapy for ulcer prevention. Now it is well established that curing H. pylori infection not only heals peptic ulcer but also prevents relapse. The following sections outline the management of peptic ulcers associated with H. pylori infection. The choice of diagnostic tests and treatment regimens for H. pylori infection is discussed in Chapter 50.

DUODENAL ULCER

Because H. pylori infection accounts for 70% or more of duodenal ulcers, one must test for the infection using one of the noninvasive tests recommended in Chapter 50. If the diagnosis of ulcer disease is made endoscopically, gastric biopsy specimens should be taken to detect H. pylori infection. If H. pylori infection is documented, the patient should

be treated with one of the regimens recommended in Chapter 50, irrespective of whether he or she has a history of NSAID use. There is good evidence that a course of H. pylori eradication therapy is sufficient to heal complicated and uncomplicated duodenal ulcers such that additional antisecretory therapy is usually not required. In a meta-analysis of 52 trials, it was found that the eradication of H. pylori alone was superior to use of an ulcer-healing drug (relative risk of ulcer, 0.66) and to no treatment (relative risk, 0.37).42 Follow-up endoscopic examination to ensure healing and testing to document H. pylori eradication after antibiotic therapy are not recommended routinely in patients with uncomplicated ulcers. However, noninvasive tests such as the urea breath test and fecal antigen test can be used to confirm H. pylori eradication in patients with ulcer complications.

GASTRIC ULCER

If H. pylori infection is documented, the patient should be treated with one of the regimens recommended in Chapter 50 regardless of whether he or she has a history of NSAID use. Whether antisecretory therapy is required after a course of H. pylori eradication therapy is controversial. It has been shown that 1 week of antibacterial therapy without acid suppression effectively heals gastric ulcers.42 In a metaanalysis of ulcer healing trials, treatment with H. pylori eradication therapy was not significantly different from treatment with an ulcer healing drug.43 For patients with large (>1.5 cm) or complicated gastric ulcers, however, additional antisecretory therapy has been shown to promote ulcer healing.44,45 Routine follow-up endoscopy is recommended to document ulcer healing, to exclude malignancy, and to confirm successful H. pylori eradication (see also Chapter 52).

ROLE OF MAINTENANCE THERAPY

After the eradication of H. pylori infection, there is little evidence that maintenance therapy with antisecretory agents is required, even for patients with complicated peptic ulcers.46,47 A meta-analysis showed that H. pylori eradication therapy was superior to no treatment in preventing recurrence of duodenal ulcer (relative risk, 0.19) or gastric ulcer (relative risk, 0.31).42 In another meta-analysis of H. pylori eradication therapy versus maintenance antisecretory therapy in prevention of recurrent ulcer bleeding, rebleeding occurred in 1.6% of the H. pylori eradication therapy group and 5.6% of the maintenance therapy group (odds ratio, 0.25; 95% confidence interval [CI], 0.08 to 0.76).48 Although some prospective trials reported that patients with duodenal ulcer had asymptomatic ulcer recurrences after eradication of H. pylori,49 these asymptomatic ulcers probably had little clinical significance.

ULCERS ASSOCIATED WITH NONSTEROIDAL ANTI-INFLAMMATORY DRUGS ACTIVE ULCERS Histamine-2 Receptor Antagonists There are limited data on the efficacy of H2 receptor antagonists in healing NSAID-associated ulcers. Current evidence suggests that conventional doses of H2 receptor antagonists effectively heal duodenal ulcers but are ineffective for gastric ulcers. In a multicenter study, the effects of ranitidine on ulcer healing were compared in a group of patients who had stopped NSAID therapy and another group who continued NSAID therapy. Gastric ulcers healed in 63% of

Chapter 53  Treatment of Peptic Ulcer Disease those still taking NSAIDs compared with 95% of those who had stopped. At 12 weeks, 79% of gastric ulcers and 92% of duodenal ulcers were healed in the group continuing NSAIDs, whereas all ulcers healed in those who had stopped taking NSAIDs.50 The ability of H2 receptor antagonists given in conventional doses to heal NSAID-associated ulcers also depends on the size of the ulcers. One early study reported that when NSAIDs were continued, 90% of gastric ulcers smaller than 5 mm healed after 8 weeks of cimetidine, whereas only 25% of ulcers larger than 5 mm healed.51

Proton Pump Inhibitors

Several large-scale studies have investigated the efficacy of PPIs for healing of NSAID-associated ulcers.52-55 Current evidence indicates that PPIs are superior to standard-dose H2 receptor antagonist therapy in healing NSAID-associated ulcers. In a large-scale randomized comparison of two doses of esomeprazole, 20 and 40 mg, and ranitidine, 150 mg twice daily, in patients who continue to take NSAIDs, ulcer healing at eight weeks was found in 85.7% of patients given esomeprazole 40 mg daily, in 84.8% of those given esomeprazole 20 mg daily, and in 76.3% of those given ranitidine 150 mg twice daily.52 In another study of 350 patients with NSAID-associated gastric ulcers who continued to use NSAIDs, ulcer healing at eight weeks was found in 69% of patients given lansoprazole 15 mg daily, in 73% of those given lansoprazole 30 mg daily, but in only 53% of those given ranitidine 150 mg twice daily.55

Misoprostol

In a randomized placebo-controlled trial in which patients continued NSAID therapy, misoprostol resulted in healing of gastric and duodenal ulcers in 67% of patients at eight weeks, compared with 26% of patients treated with placebo.56 However, misoprostol is not as effective as PPIs in healing NSAID-associated ulcers. One large-scale, randomized trial compared misoprostol 200 µg four times daily with omeprazole 20 or 40 mg daily in patients who continued NSAID treatment.53 After eight weeks, duodenal ulcers healed in 89% of patients receiving either dose of omeprazole and in 77% of those receiving misoprostol. Gastric ulcers healed in 80% of those receiving 40 mg of omeprazole, in 87% of those receiving 20 mg of omeprazole, and in 73% of those receiving misoprostol.

Sucralfrate

In a single-blind endoscopic study, sucralfate was significantly less effective than omeprazole in healing NSAIDassociated gastroduodenal ulcers.57

Adverse Role of Cyclooxygenase-2 Inhibitors in Ulcer Healing

There is good evidence that COX-2 inhibitors induce less gastric mucosal injury than conventional NSAIDs. However, animal experiments have consistently shown that COX-2, but not COX-1, is up-regulated in gastric ulcer.58,59 The administration of COX-2 inhibitors actually retards the healing of rodent gastric ulcers.60-62 These results suggest that prostaglandins generated by COX-2 contribute to restoring the integrity of gastric mucosa. A double-blind randomized trial of celecoxib on the healing of bleeding gastric ulcer found that after eight weeks, the ulcer healing rate was 65.7% in the celecoxib group and 80% in the placebo group.63 This finding indicates that treatment with COX-2 inhibitors such as celecoxib delays the healing of complicated gastric ulcers.

Recommendations

For patients in whom ulcers develop in association with the use of NSAIDs, it is recommended that NSAIDs should be discontinued if possible. Current evidence indicates that PPIs are more effective than H2 receptor antagonists, sucralfate, and misoprostol in healing NSAID-associated ulcers when continuous NSAID treatment is required. When NSAIDs can be discontinued, an H2 receptor antagonist is an effective alternative. Treatment with COX-2 inhibitors in patients with active ulcers who continue to require antiinflammatory therapy is not recommended. In the Maastricht III Consensus Report, eradication of H. pylori is advisable in patients who plan to start long-term NSAID therapy.2 The influence of H. pylori infection on the healing and relapse of NSAID-associated ulcer is discussed later.

ULCER PROPHYLAXIS

For many years an ulcer visible at endoscopy has been extensively used as a surrogate endpoint to assess the efficacy of prophylactic agents in preventing complications of NSAID-induced ulcers. An “endoscopic ulcer” has been arbitrarily defined as a circumscribed mucosal defect having a diameter of 5 mm or more with a perceivable depth.63 However, many studies have loosened this criterion to include flat mucosal breaks with a diameter of 3 mm or more as ulcers. The distinction between small ulcers and erosions is arbitrary and is prone to interobserver bias. The clinical relevance of these minor endoscopic lesions is uncertain. Although endoscopic findings roughly correlate with clinical outcomes in subjects at low to average risk for complications, current evidence indicates that the results of endoscopic studies cannot be generalized to high-risk patients.64 Because there are few prospective outcome trials to evaluate the true efficacy of prophylactic agents, clinical judgment relies on data largely using endoscopic endpoints.

Antacids

Antacids have no proven efficacy in the prevention of NSAID-induced ulcers. However, many clinicians still prescribe antacids as co-therapy for patients taking NSAIDs, both for symptom relief and prevention of ulcers. A casecontrol study showed that NSAID users receiving prophylaxis with antacids and H2 receptor antagonists had a more than two-fold increased risk of ulcer complications compared with those not taking these prophylactic agents.65 This finding was attributed to the possibility that antacids might have masked the dyspeptic symptoms, thereby creating a false sense of protection and raising the risk of silent ulcer complications. Co-prescription of antacids in patients taking NSAIDs who are at risk for ulcer should be discouraged.

Histamine-2 Receptor Antagonists

Several endoscopic studies investigated the efficacy of standard-dose H2 receptor antagonist therapy for the pre­ vention of NSAID-induced ulcers.66 A systematic review of randomized trials showed that H2 receptor antagonists significantly reduce the risk of endoscopic duodenal ulcers but not gastric ulcers.96 In contrast, it has been shown that double-dose famotidine significantly reduced the risk of both endoscopic duodenal and gastric ulcers. In one study, NSAID-related gastric ulcers developed in 20% of patients receiving placebo, in 13% of those receiving 40 mg of famotidine once daily, and in only 8% of those receiving 40 mg of famotidine twice daily.67 In another study, however, the gastric ulcer rates at 24 weeks were 41% and 19% in the groups receiving placebo and double-dose famotidine, respectively.68 The large discrepancy (8% vs.

873

874

Section VI  Stomach and Duodenum 19%) in the efficacy of double-dose famotidine between these two studies raises doubt about the true efficacy of this agent in preventing NSAID-induced gastric injury. To date there is no clinical outcome study to assess whether highdose H2 receptor antagonist therapy prevents NSAIDinduced ulcer complications.

Misoprostol

More than 20 randomized controlled trials assessed the efficacy of misoprostol in preventing NSAID-induced ulcers. A meta-analysis of the randomized trials indicated that all doses of misoprostol (400 to 800 µg per day) reduce the risk of NSAID-induced endoscopic ulcers.66 Only 800 µg per day of misoprostol is documented to reduce ulcer complications, however. In a large-scale, randomized double-blind trial in patients with rheumatoid arthritis who received NSAIDs, misoprostol (200 µg four times daily) significantly lowered the rate of GI complications, by 40% (0.95% in the placebo group vs. 0.57% in the misoprostol group). However, up to 30% of misoprostol-treated patients experienced GI upset, thus limiting its clinical use.69 Subsequent endoscopic studies suggested that lower doses of misoprostol, such as 200 µg two or three times per day, prevented NSAIDinduced endoscopic ulcers with fewer side effects.70 However, there is evidence that low-dose misoprostol therapy fails to prevent ulcer complications.71 Misoprostol has been found to be superior to H2 receptor antagonists for the prevention of NSAID-induced gastric ulcers. In one study, ranitidine (150 mg twice daily) was compared with misoprostol (200 µg four times daily) in long-term NSAID users.72 After four to eight weeks, about 1% of patients in each group demonstrated endoscopic duodenal ulcers. In contrast, gastric ulcers occurred in 5.7% of patients receiving ranitidine, compared with 0.6% of those receiving misoprostol.

Proton Pump Inhibitors

A systematic review of randomized controlled trials of PPIs for prophylaxis against NSAID-induced endoscopic ulcers found that PPIs significantly reduce the risk of endoscopic duodenal and gastric ulcers.66 The efficacy of PPIs has been compared with that of H2 receptor antagonists and misoprostol in patients who continued to receive NSAIDs. Two studies compared omeprazole 20 mg once daily with standard-dose ranitidine (150 mg twice daily) and halfdose misoprostol (200 µg twice daily) for six months.53,54 Omeprazole was found to be more effective than standarddose ranitidine but only comparable with half-dose misoprostol in preventing gastric ulcers. However, it should be noted that the superiority of omeprazole in preventing NSAID-related ulcer was due to a significant reduction of duodenal ulcers. A post hoc analysis revealed that most of the added protection attributable to omeprazole occurred among those with H. pylori infection. Another study compared high-dose misoprostol (200 µg four times daily) with two doses of lansoprazole (15 and 30 mg daily) for the prevention of ulcers in long-term NSAID users without H. pylori infection and with a history of gastric ulcer.73 Misoprostol was more effective than the two doses of lansoprazole in preventing gastric ulcer, but there was no practical advantage of misoprostol over lansoprazole because of the high withdrawal rate in the misoprostol group. In a headto-head endoscopic ulcer prevention study comparing two doses of pantoprazole with 20 mg/day of omeprazole in patients with rheumatoid arthritis receiving NSAIDs, the six-month probabilities of remaining ulcer free were 91%, 95%, and 93% for pantoprazole 20 mg, pantoprazole 40 mg, and omeprazole 20 mg, respectively.74

Two identical multicenter randomized clinical trials (RCTs) have been reported together. They compared esomeprazole (20 or 40 mg) with placebo in the prevention of ulcers in patients taking NSAIDs or COX-2 inhibitors over a six-month period. Patients in both studies were H. pylori negative, older than 60, and had a history of gastric or duodenal ulcer. Overall, the rates of ulcers were 17.0%, 5.2%, and 4.6% in the groups receiving placebo, esomeprazole 20 mg, and esomeprazole 40 mg, respectively.75 Whether PPIs can reduce the risk of NSAID-associated ulcer bleeding is largely based on observational studies and one randomized trial in high-risk patients. A large-scale case control study found that PPI therapy was associated with a significant reduction in risk of upper GI bleeding among chronic NSAID users (relative risk, 0.13; 95% CI, 0.09 to 0.19).76 One randomized trial compared long-term (six-month) omeprazole therapy with one week of H. pylori eradication therapy for the prevention of recurrent ulcer bleeding in H. pylori–infected patients with a recent history of NSAID-related ulcer bleeding who continued to use naproxen.77 Recurrent ulcer bleeding was seen in 18.8% of patients undergoing eradication therapy, compared with 4.4% of patients receiving omeprazole. In a randomized comparison of diclofenac plus omeprazole versus celecoxib for secondary prevention of ulcer bleeding in patients who either were H. pylori negative or had undergone H. pylori eradication,78 a similar proportion had recurrent bleeding in six months (6.4% in the combination therapy group compared with 4.9% of patients in the celecoxib group). These results indicate that omeprazole reduces but does not eliminate the risk of ulcer bleeding associated with NSAID use in very-high-risk patients. However, the following two important issues remain unresolved: first, the actual risk reduction achieved by PPI is unknown because of the lack of a placebo group, and second, there are no data on the efficacy of PPIs in preventing ulcer complications in low- or moderate-risk users of NSAIDs.

Role of Cyclooxygenase-2 Inhibitors in Ulcer Prevention

Consistent with the notion that inhibition of COX-2 spares the gastric mucosa, clinical trials using endoscopic ulcer as the endpoint have consistently shown that COX-2 inhibitors induced fewer ulcers than do conventional NSAIDs. Five COX-2 inhibitors have been evaluated in clinical trials: the sulfonamides celecoxib [Celebrex] and valdecoxib [Bextra] (parecoxib is a prodrug of valdecoxib), the methylsulfones rofecoxib [Vioxx] and etoricoxib, and the phenylacetic acid derivative lumiracoxib. Four large-scale clinical outcome studies—the Celecoxib Long-Term Arthritis Safety Study (CLASS),79 the Vioxx Gastrointestinal Outcomes Research Study (VIGOR),80 the Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET),81 and the Multinational Etoricoxib and Diclofenac Arthritis Long-term programme (MEDAL)82—evaluated the gastrointestinal safety of celecoxib, rofecoxib, lumiracoxib, and etoricoxib, respectively. In the CLASS trial, patients with osteoarthritis or rheumatoid arthritis were randomized to receive celecoxib versus diclofenac or ibuprofen in two substudies of identical design. On primary analysis, there was no significant difference in the incidence of ulcer complications between the celecoxib group and the nonselective NSAIDs group.79 Whether the failure of CLASS was due to flaws in the study design remains controversial. The VIGOR study compared rofecoxib with naproxen in patients with rheumatoid arthritis.80 Unlike CLASS, VIGOR demonstrated that treatment with rofecoxib significantly reduced clinical GI events

Chapter 53  Treatment of Peptic Ulcer Disease (combined endpoint of ulcer complications and symptomatic ulcers) by about 50%, compared with treatment with naproxen. Patients requiring low-dose aspirin were excluded from the VIGOR study, whereas 20% of patients in CLASS received low-dose aspirin concomitantly. The TARGET study compared lumiracoxib with naproxen or ibuprofen in patients with osteoarthritis using two substudies of identical design.81 Randomization was stratified for low-dose aspirin use and age. TARGET showed that treatment with lumiracoxib significantly reduced the incidence of ulcer complications compared with nonselective NSAIDs in the subgroup of patients not taking low-dose aspirin. Current evidence indicates that low-dose aspirin negates the GI mucosa-sparing effect of COX-2 inhibitors. The MEDAL program was a prespecified pooled analysis of data from three prospective trials. A total of 34,701 arthritic patients were treated with 60 or 90 mg of etoricoxib or 150 mg of diclofenac daily. Unlike the previous large-scale studies, the primary endpoint of this study was cardiothrombotic events. Therefore, patients on low-dose aspirin were enrolled and encouraged to receive PPI co-therapy. There was no between-group difference in terms of complicated GI events including bleeding, perforation, or obstruction. However, the overall incidence of uncomplicated GI events was significantly less with etoricoxib than with diclofenac. The reduction in uncomplicated GI events with etoricoxib is maintained in patients treated with PPIs and is also observed with regular low-dose aspirin use.82 In a systematic review of randomized trials of COX-2 inhibitors, COX-2 inhibitors produced significantly fewer gastroduodenal ulcers (relative risk, 0.26; 95% CI, 0.23 to 0.30) and ulcer complications (relative risk, 0.39; 95% CI, 0.31 to 0.5), as well as fewer withdrawals caused by GI symptoms when compared with nonselective NSAIDs.83 Current evidence indicates that COX-2 inhibitors prob­ ably are as effective as a combination of nonselective NSAIDs combined with a PPI in patients at risk for ulcers. In a double-blind randomized outcome trial of celecoxib and the combination of omeprazole and diclofenac in patients with a recent history of ulcer bleeding, approximately 5% of patients in the two treatment groups still had recurrent ulcer bleeding at 6 months.78 Although the two treatments were comparable in terms of the incidence of ulcer bleeding, a subsequent follow-up endoscopic study showed that 20% to 25% of patients receiving either treatment developed recurrent ulcers at 6 months.84 These findings suggest that neither treatment could eliminate the risk of recurrent bleeding in very-high-risk patients. Recently, a double-blind randomized trial compared celecoxib alone with combination of celecoxib and esomeprazole in patients with a history of NSAID-associated ulcer bleeding. All patients had a negative test for H. pylori infection before randomization. After a median follow-up of 13 months, 8.9% of the celecoxib-alone group had recurrent ulcer bleeding compared with none of the combined therapy group (P = 0.0004).85

Cardiovascular Risk of COX-2 Inhibitors and Nonselective NSAIDs

Despite the improved gastric safety profile of COX-2 inhibitors, the cardiovascular risk associated with this new class of NSAIDs has been the subject of much concern. In the VIGOR study, the incidence of acute myocardial events, although low, was four times higher among patients receiving rofecoxib than among patients receiving naproxen.80 Whether the observed difference in infarction rates between the two treatments was related to an antiplatelet property

of naproxen or to a thrombotic effect of rofecoxib was hotly debated. Further data regarding the cardiovascular hazard of COX-2 inhibitors were derived from two long-term studies of colon polyp prevention using rofecoxib (Adenomatous Polyp Prevention on Vioxx [APPROVE] study)86 and celecoxib (Adenoma Prevention with Celecoxib [APC] study).87 In the APPROVE study, interim data at 18 months indicated that patients who received 25 mg rofecoxib a day had double the risk of serious cardiovascular events compared with patients who received placebo.86 In September 2004, rofecoxib was voluntarily withdrawn from worldwide markets in light of this unexpected finding. In APC study, interim data at 33 months showed that the occurrence of serious cardiovascular events was significantly higher for celecoxib at the very high dose of 400 mg twice a day (hazard ratio, 1.9; 95% CI, 1 to 3.3).87 In addition, a randomized, placebo-controlled trial of parecoxib and valdecoxib in patients who had undergone coronary artery bypass surgery found an almost four-fold increased risk of myocardial infarction.88 Do COX-2 inhibitors as a class increase the risk of myocardial infarction? Results of these three placebo-controlled trials indicate that this is the case. Both polyp prevention trials, however, investigated by design supratherapeutic doses of rofecoxib and celecoxib for extended time periods.86,87 In the TARGET study, rates of myocardial infarction with lumiracoxib were lower than with ibuprofen but higher than with naproxen. Neither result was statistically significant because the trial was underpowered to detect a difference in cardiovascular outcomes between treatment groups.84 The MEDAL program was a prespecified pooled analysis of cardiothrombotic events from three trials in which patients with osteoarthritis or rheumatoid arthritis were randomly assigned to etoricoxib (60 mg or 90 mg daily) or diclofenac (150 mg daily). After an average treatment of 18 months, rates of cardiothrombotic events were similar between the two treatment groups.89 Emerging evidence suggests that not only COX-2 inhibitors but also nonselective NSAIDs, with the exception of full-dose naproxen (1000 mg a day), increase cardiothrombotic risk. In a meta-analysis of randomized trials of COX-2 inhibitors (data mostly derived from rofecoxib and celecoxib), all COX-2 inhibitors increased the cardiothrombotic risk compared with placebo (risk ratio, 1.42; 95% CI, 1.13 to 1.78). This was largely attributable to an increased risk of myocardial infarction, with little difference in other vascular outcomes. A dose-dependent increase in cardiothrombotic events was observed with celecoxib. Importantly, there was no significant difference in cardiothrombotic risk between COX-2 inhibitors and nonselective NSAIDs. Naproxen (500 mg twice daily) was the only exception.90,91 In a meta-analysis of observational studies, high-dose rofecoxib (>25 mg a day), diclofenac, and indomethacin were associated with an increase in cardiothrombotic events, whereas celecoxib did not significantly increase the cardiothrombotic risk, though an increased risk could not be excluded with doses greater than 200 mg a day.92 In February 2005, the FDA issued recommendations proposing new serious labeling warnings for valdecoxib and celecoxib with respect to increased cardiovascular risks. In April of the same year, the FDA took the further step of asking the manufacturer to remove valdecoxib from the market. All sponsors of marketed prescription NSAIDs have been asked to revise the labeling for their products to include a boxed warning highlighting the potential for increased risk of cardiovascular events in addition to the potential life-threatening GI bleeding associated with their use. Manufacturers of over-the-counter NSAIDs are also

875

876

Section VI  Stomach and Duodenum being asked to revise their labeling to provide more specific information about the potential cardiovascular and GI risks associated with their individual products and to remind patients of the importance of limited dose and duration of treatment of these products in accordance with the package instructions.93 A large-scale, randomized head-to-head comparison of COX-2 inhibitors and nonselective NSAIDs using predefined cardiothrombotic events as the primary endpoint is under way.

THE ROLE OF HELICOBACTER PYLORI INFECTION IN ULCER DISEASE ASSOCIATED WITH NSAID USE

Whether H. pylori infection influences the risk of ulcer in patients receiving NSAIDs is one of the controversial issues in peptic ulcer research. Factors such as the choice of H. pylori diagnostic test, a history of ulcer complication, concomitant use of antisecretory agents, prior exposure to NSAIDs, and the use of low-dose aspirin affect the outcome.94 A meta-analysis showed that H. pylori raised the risk of ulcer bleeding more than 6-fold in patients receiving long-term NSAIDs, whereas H. pylori and NSAIDs alone raised the risk 1.79-fold and 4.85-fold, respectively.95 An updated meta-analysis showed similar findings.96 Among patients who are about to start NSAID therapy, eradication of H. pylori reduces the subsequent risk of ulcer development.97,98 Two systematic reviews have consistently shown that eradication of H. pylori is superior to placebo in preventing peptic ulcers among NSAID users.96,99 However, eradication of H. pylori infection alone is not sufficient for the prevention of ulcer bleeding in NSAID users with high ulcer risk.77 It has been suggested that the eradication of H. pylori might retard healing of gastric ulcers.100 This was not confirmed by a prospective randomized trial using ulcer healing as the predefined endpoint.101 Currently, there is no evidence that curing H. pylori infection has any clinically important negative effect on the healing of NSAID-related ulcers. There is growing evidence that H. pylori increases the ulcer risk in patients receiving low-dose aspirin. Among patients with H. pylori infection and a history of ulcer bleeding who continued to use low-dose aspirin, a randomized trial found that successful eradication of H. pylori alone substantially reduced the risk of recurrent bleeding in six months.77 However, a later low-dose aspirin study suggested that co-therapy with a PPI after eradication of H. pylori was still required because of a high failure rate

of H. pylori eradication and because concomitant NSAID use is not uncommon in clinical practice.102

RECOMMENDATIONS FOR THE PREVENTION OF NSAID-INDUCED ULCER COMPLICATIONS Assessment of Gastrointestinal Risk

Before the cardiovascular hazards of COX-2 inhibitors and nonselective NSAIDs was a concern, prevention of NSAIDinduced ulcer complications had been based on assessment of individual patients’ GI risk factors (Table 53-1). In clinical practice, patients receiving NSAIDs can be stratified according to their levels of GI risk, as follows (Table 53-2): • Low risk: absence of risk factors • Moderate risk: presence of one or two risk factors • High risk: history of ulcer complications, multiple (three or more) risk factors, or concomitant use of lowdose aspirin, glucocorticoids, or anticoagulant therapy Because H. pylori infection raises the risk of ulcer complications in NSAID users, patients with a history of ulcer who require NSAIDs should be tested for H. pylori, and if present, the infection should be eradicated. The Maastricht III Consensus Guidelines also consider it advisable to test and treat for H. pylori infection in patients who are about to start regular NSAID therapy.2 Low-Risk Patients Patients without risk factors are at very low risk of ulcer complications with NSAID use (1% per year). Rational use of NSAIDs, including avoidance of high doses of NSAIDs and use of a less ulcerogenic NSAID (e.g., ibu­profen, diclo­ fenac) at the lowest effective dose is a cost-effective approach. Table 53-1  Risk Ratios for the Various Risk Factors for Ulcer Complications Induced by NSAIDs* risk factor

risk ratio

History of complicated ulcer Use of multiple NSAIDs (including aspirin), cyclooxygenase-2 [COX-2] inhibitor) High doses of NSAIDs Use of an anticoagulant History of uncomplicated ulcer Age > 70 years Helicobacter pylori infection Use of a glucocorticoid

13.5 9 7 6.4 6.1 5.6 3.5 2.2

*Not all NSAIDs pose the same risk. NSAIDs, nonsteroidal anti-inflammatory drugs.

Table 53-2  Recommendations for Reducing the Risk of Ulcers Associated with Nonsteroidal Anti-inflammatory Drugs (NSAIDs) as a Function of Gastrointestinal and Cardiovascular Risk GI Risk

Low CV risk High CV risk†

LOW*

MODERATE*

HIGH*

Use the least ulcerogenic NSAID at the lowest effective dose Naproxen plus either a PPI or misoprostol

NSAID plus either a PPI or misoprostol Naproxen plus either a PPI or misoprostol

COX-2 inhibitor plus a PPI, or misoprostol Avoid NSAIDs or COX-2 inhibitors; use alternative therapy.

*Low GI risk denotes absence of any risk factors (see Table 53-1); moderate GI risk denotes presence of one or two risk factors; high GI risk denotes presence of three or more risk factors, prior complicated ulcer, or concomitant use of low-dose aspirin or anticoagulants. All patients with a history of ulcer who require NSAIDs should be tested for H. pylori, and if infection is present, eradication therapy should be given (see Chapter 50). † High CV risk denotes the requirement for prophylactic low-dose aspirin for primary or secondary prevention of serious cardiovascular events. COX-2, cyclooxygenase-2; CV, cardiovascular; GI, gastrointestinal; PPI, proton pump inhibitor.

Chapter 53  Treatment of Peptic Ulcer Disease Moderate-Risk Patients Moderate-risk patients, a group that by definition includes the older adult population, account for the majority of cases of NSAID-induced ulcer complications. These patients should receive co-therapy with antiulcer agents (PPIs or misoprostol). Alternatively, substitution of a COX-2 inhibitor alone maybe as effective as the combination therapy, although no study has focused on the moderaterisk group. High-Risk Patients In general, NSAIDs should be avoided in these patients, not only because of the high risk of ulcer complications but also owing to the serious consequences of ulcer complications in the presence of comorbidities. Glucocorticoid therapy can be considered if short-term anti-inflammatory therapy is required for acute, self-limiting arthritis (e.g., gout), because glucocorticoids alone do not increase the risk of ulcer. If regular anti-inflammatory therapy is required for chronic arthritis, the combination of a COX-2 inhibitor and either misoprostol or a PPI probably offers the best GI protection, although this approach remains to be examined in prospective trials.85

Approach to Patients with High Cardiovascular Risk

After the withdrawal of rofecoxib, rational prescription of NSAIDs has become a clinical challenge. Not only GI but also cardiovascular risk factor must be assessed for the individual patient (see Table 53-2). For patients who do not have a history of coronary heart disease or ischemic stroke, identifying those with significant cardiovascular risk is not that straightforward. The American Heart Association (AHA) recommends that aspirin should be considered in all apparently healthy men and women whose 10-year risk for a cardiovascular event is 10% or above.103 We consider patients with arthritis to have significant cardiovascular risk if they are already on aspirin for secondary prophylaxis or if they require aspirin for primary prophylaxis according to the AHA guidelines. Among patients with low cardiovascular risk, prescription of NSAIDs can be based on the presence of GI risk factors only. Patients known to have high cardiovascular risk should receive low-dose aspirin irrespective of NSAID use. Because of the potential cardiovascular hazards of COX-2 inhibitors and some nonselective NSAIDs, patients with high cardiovascular risk should avoid using these drugs if possible. Ibuprofen has been found to attenuate the cardioprotective effect of aspirin, possibly through competitive binding to platelet COX-1.104,105 Concomitant use of ibuprofen and low-dose aspirin therefore should be avoided. If anti-inflammatory analgesics are deemed necessary in patients at high cardiovascular risk, evidence suggests that full-dose naproxen (500 mg twice daily) does not increase the cardiothrombotic risk. However, it remains uncertain whether the cardioprotective effect of naproxen will persist at lower doses or when naproxen is co-prescribed with low-dose aspirin. Although naproxen has an antiplatelet effect, we do not recommend using naproxen as a substitute for low-dose aspirin in patients with high cardiovascular risk. This is because naproxen has a weak antiplatelet effect and patients take NSAIDs only intermittently for pain relief. One major drawback of concomitant use of NSAIDs and low-dose aspirin is that the combination will markedly increase the risk of ulcer complications over that incurred with NSAIDs alone (see Table 53-1; Chapter 52). Thus, co-therapy with a PPI or misoprostol is necessary even if patients do not have other GI risk factors (see Table 53-2).

REFRACTORY ULCERS Most peptic ulcers heal within eight weeks of initiation of antisecretory therapy. Nevertheless, there is a small, but considerable minority of patients whose ulcers persist despite conventional treatment. Such ulcers can be considered refractory. There is no standardized definition for refractory peptic ulcer, making comparisons among studies difficult. For the patient whose peptic ulcer does not heal despite a trial of conventional therapy, the clinician should ask the following questions: 1. Has the patient complied with the prescribed treatment? 2. Has the patient received an H2 receptor antagonist or a PPI? 3. Is there H. pylori infection? If antibiotic therapy already has been prescribed, the patient should be tested to confirm that the infection has indeed been eradicated. If no attempt has been made to seek and eradicate H. pylori infection, it should be made now. False-negative test results for H. pylori should be considered if the patient is undergoing acid-suppressive therapy. 4. Is the patient still taking an NSAID? NSAID use may be surreptitious. A careful history regarding the use of over-the-counter NSAIDs including low-dose aspirin should be obtained, and NSAIDs should be stopped if possible. 5. Does the patient smoke cigarettes? If so, he or she should be counseled strongly to discontinue cigarettes. 6. Has the duration of ulcer treatment been adequate? Large ulcerations require a longer duration of therapy than small ulcers to heal. A large ulceration (e.g., >2 cm) probably should not be considered refractory until it has persisted beyond 12 weeks of antisecretory therapy. 7. Is there evidence of a hypersecretory condition? A family history of gastrinoma or multiple endocrine neoplasia type 1 or a personal history of chronic diarrhea, hypercalcemia due to hyperparathyroidism, or ulcers involving the postbulbar duodenum suggests a diagnosis of Zollinger-Ellison syndrome or other hypersecretory state (see Chapter 32). 8. Is the ulcer penetrating the pancreas, liver, or other organ? 9. Is the ulcer indeed peptic? Primary or metastatic neoplasms, infections (e.g., cytomegalovirus), cocaine use, eosinophilic gastroenteritis, and inflammatory bowel diseases (e.g., Crohn’s disease) can cause ulcerations of the stomach and duodenum that can mimic peptic ulcers (see Chapter 52). These disorders should be considered and excluded appropriately.

STRESS-RELATED MUCOSAL INJURY Stress-related mucosal injury is an illness of the critically ill who are typically cared for in intensive care units. Fortunately, only a small proportion of patients with stressrelated mucosal lesions have clinically overt bleeding. In a prospective study of more than 2000 patients admitted to intensive care units, only 1.5% experienced clinically important bleeding.106 Respiratory failure and coagulopathy were strong, independent risk factors for stress-related hemorrhage. Important bleeding occurred in 3.7% of the 847 patients who had one or both of these risk factors, whereas

877

878

Section VI  Stomach and Duodenum only 0.1% of 1405 patients without respiratory failure or coagulopathy experienced such bleeding. Thus, routine uses of stress ulcer prophylaxis in the intensive care unit is not recommended unless the patient has a coagulopathy or is receiving mechanical ventilation. The approach to treatment of actively bleeding stress ulcers is similar to that of bleeding peptic ulcers (see later). Endoscopic control of hemorrhage followed by adjuvant therapy with PPI infusion is the preferred treatment option. However, the risk of recurrent bleeding is high, and bleeding carries a high mortality due to multiorgan failure. Although intravenous H2 receptor antagonists and oral or intragastric sucralfate are widely used to prevent stress ulcers in critically ill patients, the true efficacy of these treatments remains controversial. Moreover, few studies have evaluated PPIs for stress ulcer prophylaxis. The majority of studies have demonstrated that enteral or intravenous administration of PPIs to critically ill patients elevates intragastric pH and consistently maintains pH at 4 or higher.107 To date, no large-scale, clinical outcome study has been conducted to assess the efficacy of PPIs for this condition, despite their widespread use. An early, widely quoted metaanalysis conducted by Cook and associates108 concluded that H2 receptor antagonists were more effective than placebo in reducing the incidence of clinically important GI bleeding. However, in a later meta-analysis performed by Messori and colleagues,109 trials of ranitidine showed no difference from placebo; moreover, the data available on sucralfate did not allow any conclusion to be drawn. The Messori meta-analysis differed from the Cook meta-analysis in two important aspects. First, Cook and associates108 included trials that used either overt bleeding or clinically important GI bleeding as an endpoint, whereas Messori and colleagues included only trials that used clinically important GI bleeding as an endpoint. Second, in assessing the effectiveness of H2 receptor antagonists, Cook’s group included five trials that used cimetidine and three trials with negative results that used ranitidine, whereas Messori’s group included only the trials of ranitidine. A separate analysis showed that cimetidine but not ranitidine signi­ficantly reduced the rate of bleeding. The differences in outcome between cimetidine and ranitidine were probably due to chance rather than to a genuine difference between these two H2 receptor antagonists. Despite these discrepancies, there was either a rising trend108 or a significant increase109 in the incidence of nosocomial pneumonia with ranitidine compared with sucralfate. Overall, the findings are based on small numbers of patients with variable study design, and thus, firm conclusions cannot be drawn. Currently, neither the FDA nor the European Medicines Evaluation Agency has given approval to the routine use of H2 receptor antagonists, sucralfate, or PPIs for stress ulcer prophylaxis.

TREATMENT OF COMPLICATIONS OF PEPTIC ULCER DISEASE HEMORRHAGE 100,111 Initial Management (see Chapter 19)

Consensus groups have recommended a multidisciplinary approach to the care of patients presenting with upper GI bleeding.112,113 A team should include both medical and surgical gastroenterologists with access to skills in endoscopic hemostasis, which has become the mainstay of treatment. Patients identified as being at high risk of continued or recurrent bleeding should be admitted to an intensive care unit.

Patients with acute upper GI bleeding should be assessed promptly on presentation. Resuscitation and volume restoration should take priority and should precede endoscopy. Features of liver cirrhosis should call attention to the possibility of bleeding from esophagogastric varices rather than an ulcer. This distinction has prognostic as well as management implications. Variceal hemorrhage carries a higher death rate. The possibility of variceal hemorrhage calls for specific measures prior to endoscopy, such as the use of vasoactive drugs (e.g., octreotide) and antibiotics (e.g., cefotaxime) as prophylaxis against spontaneous bacterial peritonitis (see Chapters 90 and 91). The routine use of a nasogastric tube to obtain a gastric aspirate and to lavage the stomach cannot be recommended because in the majority of cases, the diagnosis of bleeding from an upper GI source is obvious. The insertion of a nasogastric tube is associated with a small risk of pulmonary aspiration, particularly in patients with compromised airways. Gastric lavage prior to endoscopy is often unnecessary and ineffective because blood and blood clots tend to pool at the fundus during endoscopic examination with the patient in the left lateral position, whereas bleeding lesions are usually located along the lesser curvature, angular notch, distal stomach, or duodenal bulb. The finding of fresh blood in an “in-and-out” nasogastric intubation and aspiration indicates ongoing and often massive bleeding and may predict a poor outcome.

Risk Stratification

Bleeding stops in about 80% of patients presenting with acute upper GI bleeding. The remaining 20% constitute a high-risk group with substantial morbidity and mortality. It is therefore important to identify and direct appropriate care to patients at risk of continued or recurrent bleeding. For practical purposes, the management distinctions to be made are whether the patient is in need of urgent endoscopy and whether the patient is likely to have recurrent bleeding after initial endoscopic control. Some of the clinical predictors of increased risk for continued or recurrent bleeding are older age, shock, comorbid illnesses, low hemoglobin value, need for transfusion, and the finding of fresh blood in emesis or on rectal examination.106 Patients with signs of ongoing bleeding need urgent endoscopy with a view to securing hemostasis. Older adult patients tolerate blood loss poorly and are likely to have organ dysfunction consequent to bleeding; in such patients, the basis for early intervention should be more liberal. Other clinical predictors associated with higher mortality include the onset of bleeding in patients already hospitalized for other reasons. Several derived risk scores have been developed to aid physicians in clinical decisions.114-116 The Rockall and Baylor scores are composite systems consisting of two components, the pre-endoscopy and postendoscopy scores.114,115 The Rockall scoring system (see Chapter 19) was derived from data gathered from the National United Kingdom Audit. A score of 0 to 2 indicates an excellent prognosis, whereas a score of 9 or more is associated with a high risk of death. The Blatchford score, on the other hand, uses clinical parameters only and is calculated from patients’ hemoglobin and blood urea concentrations, pulse and systolic blood pressure on admission, the presence or absence of melena or syncope, as well as of evidence of cardiac or hepatic failure.116 Endoscopic stigmata of bleeding not only pinpoint the source of bleeding but are themselves prognostic. The commonly used nomenclature is a version modified from Forrest and Finlayson’s117 original description, as follows (Table 53-3):

Chapter 53  Treatment of Peptic Ulcer Disease Table 53-3  Frequency and Prognosis of Various Endoscopic Stigmata of Hemorrhage in Patients with Bleeding Peptic Ulcer* ENDOSCOPIC CHARACTERISTIC†

FREQUENCY (%)

FURTHER BLEEDING (%)

SURGERY (%)

MORTALITY (%)

42 20 17 17 18

5 10 22 43 55

0.5 6 10 34 35

2 3 7 11 11

Clean base (type III) Flat pigmentation (type IIc) Adherent clot (type IIb) Nonbleeding visible vessel (type IIa) Active bleeding (type I)

*Percentages are average figures taken from multiple studies118 and therefore do not add up to 100%. † Classification by Forrest et al117 is shown in parentheses (see text). Adapted from Laine L, Peterson WL. Bleeding peptic ulcer. N Engl J Med 1994; 331:717.

Figure 53-1.  Endoscopic appearance of a duodenal bulbar ulcer with a fresh adherent clot (Forrest type IIb).

Type I: Active bleeding: Ia: Spurting hemorrhage Ib: Oozing hemorrhage Type II: Stigmata of recent hemorrhage: IIa: Nonbleeding visible vessel IIb: Adherent clot IIc: Flat pigmentation Type III: Clean-base ulcers Reported prevalences of these endoscopic stigmata and their respective risks of recurrent bleeding have varied widely. This variation may be attributed to differences in visual interpretation among endoscopists and to varying definitions of recurrent bleeding. In one review,118 the rate of recurrent bleeding was less than 5% in patients with a clean ulcer base (type III) and increased to 10% in patients with a flat pigmentation (IIc), to 22% in those with an adherent clot (IIb), to 43% in those with a nonbleeding visible vessel (IIa), and to 55% in those with active bleeding, either spurting or oozing (type I) (see Table 53-3). Actively bleeding ulcers and ulcers with nonbleeding visible vessels (“protuberant discoloration”) warrant endoscopic therapy.119 Endoscopic therapy of ulcers with “adherent clots” has been controversial (Fig. 53-1). The definition of adherent clot varies with the vigor in endoscopic washing. Some endoscopists use targeted irrigation from a thermal probe. Some go to the extent of mechanical removal using a

Figure 53-2.  Endoscopic appearance of a gastric ulcer at the angularis with a flat spot (Forrest type IIc).

polypectomy snare. Two randomized controlled studies and a meta-analysis compared medical therapy with endoscopic treatment in patients with ulcers harboring “adherent clots” and concluded that clot removal followed by endoscopic treatment to the vessel underneath would reduce the risk of recurrent bleeding from around 30% to 5%.120-122 It is not often possible to distinguish a clot from a vessel. Indeed, it is logical to believe that for every clot there is an underlying artery. Johnston123 introduced the term sentinel clot, which is often used synonymously with visible vessel. An ulcer stops bleeding when an eroded artery is plugged by a clot, which varies in color. The “sentinel clot” can be contiguous with a larger overlying clot. With time, the ulcer heals, initially leaving a flat pigment to the ulcer base, and the vessel eventually disappears from the ulcer floor. This evolution of a bleeding vessel usually takes less than 72 hours. Ulcers with a flat spot or a clean base do not warrant endoscopic therapy (Fig. 53-2).

Endoscopic Therapy (see also Chapter 19) Early endoscopy is generally defined as endoscopic examination performed within 24 hours of the patient’s admission. In patients with signs of active bleeding, it is common sense that urgent endoscopy establishes diagnosis and offers possible intervention. Such an approach in high-risk patients is generally believed to improve outcome. Rando­

879

880

Section VI  Stomach and Duodenum mized controlled trials demonstrated that early endoscopy in patients at low risk enabled their early hospital discharge and resource utilization or even management as outpatients.124-126 Many low-risk patients can be scheduled for endoscopy the morning after admission. A small but significant portion of patients with major bleeding require urgent endoscopy and therapy. Two meta-analyses of endoscopic therapy showed signi­ ficant reductions in the rates of further bleeding, surgery, and, importantly, mortality.127,128 Endoscopic therapy can be divided into injection, thermal, and mech­anical methods. Injection Methods Endoscopic injection of diluted epinephrine into a bleeding peptic ulcer works by volume tamponade and local vasoconstriction, as blanching and edema of the mucosa are observed. It is an easy technique to learn and diluted epinephrine is non-tissue damaging and therefore safe to use. Diluted epinephrine, however, does not induce vessel thrombosis. Recurrent bleeding after injection with diluted epinephrine alone occurs in 20% to 30% of patients. In theory, the addition of a second agent to cause vessel thrombosis would further reduce the rate of recurrent bleeding. Various sclerosants (e.g., sodium tetradecyl sulfate, polidocanol, absolute alcohol) have been applied to the vessel after initial hemostasis with epinephrine. The addition of a sclerosant has not shown to further reduce rebleeding.129 Sclerosants damage tissue in a dose-dependent manner. Cases of gastric necrosis, some of them fatal, have been reported after sclerosant injection.130 Thrombin and fibrin, derived from both bovine and human sources, have been used as injection agents. A largescale European multicenter trial demonstrated a statistically significant lower rate of recurrent bleeding associated with repeated injections of fibrin sealant (a mixture of fibrin and thrombin) at scheduled daily endoscopic examinations in comparison with a single injection of epinephrine plus polidocanol.131 The close surveillance rather than action of the fibrin sealant per se might have accounted for the difference. A single injection of fibrin sealant was not superior to epinephrine-polidocanol injection. There are also concerns about transmission of viral agents and anaphylaxis with the use of products derived from pooled plasma. Other agents for endoscopic injection therapy are normal saline and hypertonic saline. No single solution is superior to another for hemostasis. Diluted epinephrine is the only remaining agent that is widely used for injection therapy because of its safety, both local and systemic, as well as its low cost and easy availability. Thermal Methods Thermal methods of endoscopic therapy are divided into contact and noncontact methods. Noncontact methods refer to the former use of laser photocoagulation and the current use of argon plasma coagulation. Laser therapy is no longer used because laser units are bulky and difficult to transport. In canine mesenteric artery models, Johnston and colleagues132 compared laser photocoagulation with contact thermal probes in hemostasis. The use of 3.2-mm contact probes consistently sealed arteries up to 2 mm in size. Laser probes were much less effective. The researchers introduced the term coaptive thermocoagulation and emphasized the need for compression of vessel walls. The two walls of an artery are pressed together by firm tamponade. This in itself stops blood flow and reduces the “heat-sink” effect. Heat energy is then generated, welding the arterial lumen. The commonly used contact thermal probes are the heater probe, which has a polytetrafluoroethylene (Teflon)-coated copper

tip with three water ports for targeted irrigation, and bipolar probes. Firm tamponade is the key to successful application of contact probes. At least in animal experiments, thermal methods are superior to injection therapy in achieving hemostasis. Comparative clinical trials of injection and thermal methods did not show any difference in clinical outcomes. Mechanical Methods Because surgical plication of the bleeding artery is considered the most definitive treatment to achieve hemostasis, mechanical methods such as the endoscopic application of a hemoclip come closer to what would otherwise be done at surgery. The tangential application of hemoclips in treating posterior duodenal bulbar ulcerations or their use with the endoscope in a retroflexed position for treatment of high lesser curvature ulcers can be technically difficult. In a meta-analysis that included 15 randomized studies that compared injection, thermocoagulation, and hemoclipping, successful application of hemoclips was superior to injection alone but comparable to thermocoagulation in producing definitive hemostasis.133 Combination Methods A combination therapy incorporating advantages of injection and thermal methods may represent a better approach than either method alone. Preinjection with diluted epinephrine allows a clear view of the bleeding vessel, making accurate thermocoagulation or application of a second modality possible. The benefit of combination therapy has been confirmed in two meta-analyses.134,135 In the first meta-analysis, the addition of a second modality significantly reduced the rate of recurrent bleeding from 18.4% to 10.6% and that of emergency surgery from 11.3% to 7.6%. The mortality rate decreased significantly from 5.1% to 2.6%. Eleven studies used injected substances such as a sclerosant, tissue adhesive, or thrombin; two added hemoclips; and three evaluated the added use of thermal devices. Findings of the meta-analysis suggest that a second modality should be added after injection of diluted epinephrine to bleeding peptic ulcers. The meta-analysis also confirmed that the rate of significant complications such as perforation and gastric wall necrosis was higher in the combined therapy group (6 of 558 patients) than in the epinephrine-alone group (1 of 560 patients). Furthermore, the improvement in prognosis seems to be more evident in ulcers with active bleeding (Forrest type I ulcers). Despite the large volume of published literature, the best endoscopic therapy for bleeding peptic ulcers remains undefined. It is becoming clear that injection with diluted epinephrine alone is inadequate, especially in ulcers with active bleeding. The most widely adopted method is probably the combination therapy of preinjection with diluted epinephrine followed either by thermocoagulation using a 3.2-mm contact probe or the use of hemoclip. Thermocoagulation or hemoclipping on its own may suffice if the bleeding lesion can be seen clearly for its accurate application. The critical determinant of the efficacy of endoscopic therapy is the size of the eroded artery in the ulcer base. Swain and associates136 studied gastrectomy specimens in patients who required emergency gastrectomy for bleeding gastric ulcers. The researchers suggested that bleeding from arteries larger than 1 mm could not be stopped by existing methods of hemostasis. Most studies on predictors of persistent or recurrent bleeding from ulcers find ulcer size of larger than 2 cm and ulcer location either high on

Chapter 53  Treatment of Peptic Ulcer Disease the lesser curvature of the stomach or in the superior or posterior duodenal bulb to be associated with poorer outcomes. These are the classic locations for ulcers that erode into major artery complexes, such as the left gastric artery and the gastroduodenal artery, respectively. Surgery remains the only definitive method of securing bleeding in these patients.

Antisecretory Therapy

The rationale for antisecretory therapy is based on the fact that both pepsin activity and platelet aggregation are pH dependent. An ulcer stops bleeding when a fibrin or platelet plug blocks the rent in a bleeding artery. When gastric pH exceeds 4, pepsin is inactivated, preventing enzymatic digestion of blood clots. A gastric pH of 6 or greater is critical for clot stability and hemostasis. Labenz and associates137 studied gastric pH in patients with peptic ulcers receiving either a high dose of omeprazole (intravenous bolus 80 mg, followed by 8 mg/hr) or a high-dose ranitidine infusion (intravenous bolus 50 mg, followed by 0.25 mg/kg/hr). The gastric pH was less than 6 just 0.1% of the time in patients with either gastric or duodenal ulcers treated by high-dose omeprazole, much less than with ranitidine (20% in duodenal ulcers and 46% in gastric ulcers). In another study that measured gastric pH over three days, the use of histamine receptor antagonists given either in high-dose intravenous infusion or in bolus form led to progressive loss of antisecretory effect over days two and three because of tolerance. To achieve a gastric pH consistently above 6, a high-dose proton pump infusion is required. The use of H2 receptor antagonists in the management of bleeding peptic ulcers has been evaluated in numerous clinical trials and summarized in meta-analyses. Patients with duodenal ulcer bleeding, who typically have a higher gastric acid output, do not benefit from the use of H2 receptor antagonists. A recent meta-analysis of 30 rando­ mized trials concluded that the use of H2 receptor antagonists would be of benefit only in patients with gastric ulcers (absolute risk reductions of 7.2%, 6.7%, and 3.2% in the rates of recurrent bleeding, surgery, and death, respectively).138 Strong evidence for the use of PPIs in patients with bleeding peptic ulcers comes from a clinical trial reported by Lau and associates,139 in which early endoscopy was used to triage patients with bleeding peptic ulcers; only those at high risk of recurrent bleeding (i.e., those who had actively bleeding ulcers or ulcers with nonbleeding visible vessels) were enrolled. After endoscopic thermocoagulation of the ulcers, patients were randomly assigned to receive a highdose omeprazole infusion or a placebo for 72 hours. The rate of recurrent bleeding in those who received the PPI infusion was 6.7% at day 30, compared with 22.5% in those who received placebo. In addition, the trial showed significant reductions in the need for further intervention, transfusion, and hospitalization as well as a trend in reducing the death rate in patients who received omeprazole. Similar benefits with intravenous esomeprazole given after successful endoscopic therapy in patients at high risk for recurrent ulcer bleeding have recently been reported.140 A Cochrane Systematic Review later concluded that the use of PPI therapy significantly reduces rates of recurrent bleeding and surgery but not mortality.141 In a subgroup analysis including studies that allowed initial endoscopic control, a significant reduction in mortality among Asians was seen in association with the use of a PPI. This supports the use of PPI as an adjunct to endoscopic therapy. The optimal dose to use and the routine of PPI administration

continue to be controversial. The authors advocate the use of early endoscopic triage with a strategy to treat actively bleeding ulcers and ulcers with nonbleeding vessels, followed by adjunctive use of a high-dose intravenous infusion of a PPI. Preemptive use of an intravenous PPI infusion prior to endoscopy was studied in a large-scale randomized study.142 Patients with overt signs of upper GI bleeding were randomized to receive either a high dose PPI infusion or placebo. In the cohort, 60% were found to be bleeding from a peptic ulcer during endoscopy. The study demonstrated that early PPI infusion downstaged bleeding stigmata in ulcers and thereby reduced the need for endoscopic therapy. In the PPI group there were fewer ulcers with active bleeding or with major stigmata observed the next morning during endoscopy. PPI infusion starts ulcer healing, and significantly more clean-based ulcers are seen the next day. The study has cost-saving implications with less endoscopic therapy required with the use of intravenous PPI. In patients awaiting endoscopy it is reasonable to start PPI therapy.

Emergency Surgery

Indications Effective endoscopic intervention and improved pharmacotherapy have greatly reduced the need for emergency ulcer surgery. Not so long ago, surgery was the only reliable means of stopping bleeding. The National United Kingdom Audit performed more than a decade ago revealed an operative rate of 12% among 2071 patients with bleeding peptic ulcers and an associated mortality rate of 24%.143 Endoscopic intervention had not been used in 78% of these patients. In the current literature, surgery is often defined as an outcome in clinical trials of endoscopic therapy. Despite its diminished role in the management of bleeding peptic ulcer, surgery remains important. A common indication for emergency surgery is failure of endoscopic therapy. The usual scenarios are as follows: (1) spurting hemorrhage could not be stopped by endoscopic means; (2) the bleeding point could not be seen because of heavy active bleeding; and (3) recurrent bleeding appeared after initial endoscopic control (although as discussed later it is not entirely clear how many endoscopic attempts should be made before endoscopic therapy is deemed to have failed). Timing The timing of surgery for ulcer bleeding has been a subject of intense debate. In the 1980s, when endoscopic therapy was not available, Morris and colleagues144 published the only prospective randomized study that compared early surgery with delayed surgery, if needed, in 140 patients with bleeding ulcers. In patients younger than 60 years, there was no death in either group, but the more aggressive early surgery policy led to an unacceptably high operation rate (52% compared with only 5% for the delayed surgery group). For those older than 60 years, the operation rate was 62% in the early group and 27% in the delayed group. There were three deaths in 48 patients (6%) in the early surgery group and seven deaths in 52 patients (13%) in the delayed group. On intention-to-treat analysis, the difference did not reach statistical significance. According to treatmentreceived analysis, difference in mortality in patients with gastric ulcers was statistically significant (0 deaths in 19 of the early group vs. 5 in 21 of the delayed group, P < 0.01). The trial has been criticized for its small sample size and the use of subgroup analysis. In patients assigned to delayed surgery, ongoing bleeding was allowed before surgical intervention. Nevertheless, the study clearly demon-

881

882

Section VI  Stomach and Duodenum strated that early intervention reduced blood loss in older adult patients and improved outcome. Endoscopic therapy has replaced surgery as the first treatment in bleeding ulcers. After initial endoscopic control, most clinicians would adopt an expectant policy. Most of the affected patients are elderly and have comorbid illnesses. The prospect of long-term cure with a powerful PPI and H. pylori eradication provides incentives for clinicians to adopt a conservative stance. There remains a subgroup of patients predicted to be at risk for recurrent bleeding on the basis of ulcer characteristics at endoscopy. Such characteristics include larger ulcers located at the high lesser curvature and posterior duodenal bulb. The team approach allows experienced gastroenterologists and GI surgeons to confer after initial endoscopic control of bleeding. It is possible that some of these “difficult” ulcers should be selected for safer early elective surgery. A clinical trial to answer this question is unlikely. Choice of Operation The type of operation to be undertaken at emergency surgery for ulcer bleeding is controversial. Some surgeons maintain that oversewing of ulcers alone combined with acidsuppression therapy is safer than “definitive” surgery using either gastrectomy or vagotomy. H. pylori eradication and PPIs have provided incentives for surgeons to perform the minimum. Two randomized studies that compared minimal with definitive surgery have been published.145,146 A United Kingdom multicenter study compared minimal surgery (oversewing the vessel or ulcer excision alone plus intravenous H2 receptor antagonist therapy) with a definitive ulcer surgery (vagotomy and pyloroplasty or partial gastrectomy) in patients with bleeding gastric or duodenal ulcerations. Of the 62 patients assigned to conservative treatment, 7 experienced rebleeding, of whom 6 died. Of the 67 patients who received conventional ulcer surgery, 4 had rebleeding and none died; in all rebleeding cases, vagotomy and oversewing of the ulcer had been performed. The overall mortality in this high-risk group of patients was similar in the two groups: 26% after minimal surgery and 19% after conventional surgery. The trial was terminated because of the high rate of fatal rebleeding in the conservative surgery group in comparison with the conventional surgery group.145 In the French Association of Surgical Research trial, patients with duodenal ulcers were randomly assigned to either oversewing plus vagotomy and drainage or partial gastrectomy.146 After oversewing and vagotomy, recurrent postoperative bleeding occurred in 10 of 60 patients (17%), in 6 of whom conversion to a Billroth II gastrectomy was required. Five of these 6 patients experienced duodenal stump dehiscence. In the group of 60 assigned to undergo partial gastrectomy, only 2 patients (3%) had rebleeding, both of whom recovered after conservative treatment. Of the 60 patients assigned to partial gastrectomy, Billroth I reconstruction was performed in 18, Billroth I reconstruction plus vagotomy in 6, Billroth II reconstruction in 20, and Billroth II reconstruction plus vagotomy in 16 patients. No duodenal leak occurred in 24 patients after Billroth I reconstruction. Among the 36 patients who received Billroth II reconstruction, duodenal stump leaks occurred in 8 (22%). The rate of duodenal stump leak in the overall gastrectomy group was therefore 8 in 60 (13%). When the results were analyzed on an intention-to-treat basis, and data from patients with duodenal leaks after reoperations for rebleeding in the oversewing and vagotomy groups were included, the duodenal leak rates were similar in the two groups (7 of 58 vs. 8 of 60, respectively). The researchers concluded that a more

aggressive approach would be warranted in the surgical treatment of duodenal ulcers. The two randomized studies just reviewed emphasize that simple oversewing with or without vagotomy is associated with a higher rate of recurrent bleeding. In patients with recurrent bleeding, the mortality is exceptionally high. Exclusion of an ulcer (see later) or, in the case of gastric ulcers, ulcer excision is important in preventing recurrent bleeding. It is inappropriate, however, to recommend partial gastrectomy in all cases. The decision to perform a gastrectomy has to be balanced against the risk of duodenal stump dehiscence. The choice of resection is determined to a large extent by anatomic and pathologic considerations. The choice of the proper surgical procedure for the individual patient with massive gastric or duodenal ulcer bleeding also rests heavily on the experience and the judgment of the surgeon. Expertise in the surgical management of bleeding ulcers remains an important integral feature of an upper GI bleeding team. Difficult Ulcers Effective endoscopic therapy has selected a group of difficult ulcers for surgery. Anatomic factors that predict failure with endoscopic therapy are size larger than 2 cm and location at the posterior duodenal bulb or lesser curvature of the stomach. An ulcer at the former location often erodes into the gastroduodenal artery complex, and an ulcer at the latter location often erodes into the main left gastric artery or its first-generation branches. The size of the artery well exceeds the limits of endoscopic hemostasis. These ulcers represent challenges to GI surgeons, and expertise is required in dealing with them. At surgery, a bulbar duodenal ulcer can be accessed via a longitudinal pylorotomy extending into the duodenum. Berne and Rosoff147 identified the confluence of several branches of the gastroduodenal artery in the vicinity of a bleeding posterior duodenal ulcer. Ligations above and below the bleeding artery are insufficient to ensure hemostasis. Berne and Rosoff suggested a U stitch in the center after ligations above and below. Many surgeons perform plications at four quadrants, and a few figure-of-eight stitches at varying angles along the course of the artery are often required. The longitudinal pylorotomy is then closed vertically as a Mikulicz-Heineke type of pyloroplasty.147 Whether the procedure should be completed with a truncal vagotomy is unclear, because powerful PPI therapy is now available. Recurrent bleeding occurs in 5% to 17% of cases after a vagotomy and pyloroplasty, often with a fatal outcome. To avoid this complication, many surgeons argue for excluding the duodenal ulcer by closing the duodenal stump distal to the ulcer. Some surgeons advocate end-toend gastroduodenostomy (a Billroth I type reconstruction), in which the gastric remnant is advanced over the ulcer crater and sutured to the normal duodenal mucosa distal to the ulcer. No attempt is made to dissect the posterior duodenal wall distal to the inferior border of the ulcer. Often the duodenum retracts distally, and the stump can be closed by suturing of the divided anterior duodenal wall onto the distal lip of the ulcer (Nissen’s method). Other techniques in dealing with a difficult duodenal stump include a side catheter duodenostomy and the technique of Roux-en-Y jejunoduodenal anastomosis. Ulcers located at the incisura angularis and lesser curvature can erode into the left gastric artery or its first branches. In larger chronic ulcers, ulcer resection is often necessary. A high gastric ulcer presents a special problem. Oversewing of the bleeding point through an anterior gastrotomy is a simple option for old and frail patients. For lower-risk

Chapter 53  Treatment of Peptic Ulcer Disease patients, a Pauchet operation or a sleeve resection technique can be used. The stomach is transected from the greater curvature side. On the lesser curvature side, a tongue of gastric tissue that is based distally and stretches proximally toward the cardia is excised, usually freehand, to include the ulcer. A Roux-en-Y reconstruction is often required, with a large portion of stomach excised. Closure of duodenal stump is not a concern here because the duodenum is not diseased.

surgery for duodenal ulcer perforation are sterile; this is especially true when surgery is undertaken soon after perforation because the initial peritonitis is chemical in etiology. Nevertheless, acid-suppressive agents raise gastric pH and allow bacterial overgrowth in the stomach. The yield from culture increases with time as secondary peritonitis sets in. Coliforms, streptococci (usually anaerobic), staphylococci, and Candida species are the common organisms isolated. In late disease with abscess formations, anaerobes can often be isolated.

Surgery Versus Endoscopic Retreatment After Recurrent Bleeding

Medical Management

Recurrent ulcer bleeding is a major adverse prognostic factor for morbidity and mortality. Physicians often perform a second endoscopic examination to confirm recurrent bleeding and to re-treat the bleeding ulcer. The avoidance of salvage surgery may be desirable in older adult patients. There is, however, a concern that patients with rebleeding could be worse off after yet another failed endoscopic attempt and episodes of hypotension. The choice between endoscopic re-treatment and surgery for recurrent bleeding after initial endoscopic control was addressed by Lau and colleagues148 in a randomized trial. In a cohort of 1169 patients with bleeding peptic ulcers treated by epinephrine injection followed by thermocoagulation, recurrent bleeding occurred in 8.7%. Ninety-two patients (mean age 65 years, 76% men) were randomly assigned to undergo either endoscopic re-treatment or surgery. Using intention-to-treat analysis, the endoscopic re-treatment and surgery groups did not significantly differ in mortality at 30 days (10% for re-treatment vs. 18% for surgery), duration of hospitalization (median 10 vs. 11 days, respectively), need for intensive care or length of stay in an intensive care unit (5 vs. 10 patients, respectively; median of 59 days for both), or units of blood transfused (median 8 vs. 7 units, respectively). Patients who underwent surgery were significantly more likely to have complications (16 vs. 7, respectively). Endoscopic re-treatment was able to control bleeding in three quarters of the patients. In those for whom endoscopic re-treatment failed, salvage surgery carried substantial mortality. In a regression analysis of a small subgroup of patients, ulcers 2 cm or larger and hypotension at rebleeding were two independent factors predicting failure with endoscopic re-treatment. Findings of this trial suggest that a selective approach can be adopted on the basis of the characteristics of the ulcer.148 Large chronic ulcers should probably be treated with expedited surgery at the time of rebleeding. Early elective surgery may have been more appropriate in these chronic ulcers after initial endoscopic control. Angiographic embolization of bleeding arteries to peptic ulcer is a nonoperative option. In a nonrandomized comparison to surgery, angiographic embolization carried a similar rate of recurrent bleeding (29% vs. 23%), need for further intervention (16% vs. 31%), and death (26% vs. 21%).149

PERFORATION

A short period of resuscitation after ulcer perforation is diagnosed or suspected is often desirable, with restoration of fluids, electrolytes, and, if needed, blood. A nasogastric tube and a urinary catheter should be in place. Pain should be relieved with opiates after a presumptive diagnosis is made. As described in Chapter 37, intravenous broadspectrum antibiotics should be administered parenterally, even though gastric juice contains few organisms. A majority of cultures of peritoneal fluid collected at the time of

Nonoperative management of ulcer perforation has been advocated. The regimen involves nasogastric suctioning, parenteral antibiotics, and intravenous fluids. Crofts and associates150 randomly assigned patients with the presumptive diagnosis of perforated ulcers to either conservative treatment or surgery. Of 40 patients assigned to conservative treatment, 11 showed no improvement in 12 hours and underwent operation. Three of these 11 patients were found to have perforated carcinomas, 2 gastric and 1 sigmoid. Morbidity and mortality rates were similar in the medical and surgical groups. This trial may have included lower-risk patients or patients who presented early, because the 5% overall mortality is low. Findings of the study highlight common objections to the use of nonoperative management error in diagnosis, uncertainty of site of perforation, and the possibility of a perforated gastric tumor. Older adult patients should be operated on early for the following reasons: (1) malignancy is more likely; (2) atrophy of the greater omentum makes spontaneous sealing less likely; (3) such patients often withstand sepsis and organ dysfunction poorly; and (4) early surgery leads to a better outcome.150 In a patient in whom the perforation is considered confined on clinical assessment, it would be prudent to verify this judgment with a contrast radiographic study. If sealing is confirmed, conservative treatment may be reasonable.

Surgical Management

Surgery is the usual approach to an ulcer perforation. At the time of surgery the surgeon has to treat peritoneal contamination, the perforation, and the ulcer. The controversies in the operative management have been the need for definitive ulcer operation, the choice of which definitive operation to perform, and, more recently, the choice between laparoscopic and open suture repair. Treatment also differs for duodenal and gastric perforations. For example, simple closure of a perforated duodenal or a juxtapyloric ulcer with the use of an omental patch is widely practiced. Whether to perform definitive ulcer surgery at the time of ulcer perforation was an argument that predated the era of H. pylori and PPIs. Relapse of ulcer disease with modern medical therapy is now uncommon and therefore is no longer a factor in the consideration as to the type of surgery to perform. There is now evidence that H. pylori eradication reduces relapse of ulceration after patch repair.151 In a randomized study enrolling patients with perforated duodenal ulcers after omental patch repair, patients received either a PPI alone for four weeks or quadruple anti–H. pylori therapy. After a year, ulcer relapse had occurred in 38% in the PPI group compared with only 5% in the H. pylori eradication group. Thus, ulcer relapse is uncommon after simple closure of a perforation and H. pylori eradication. These data would support the performance of simple closure alone in perforated ulcers.151 Gastric ulcers account for about 20% of perforated peptic ulcers. Epidemiologic data suggest a rising proportion of gastric ulcers among perforated ulcers, especially in older

883

884

Section VI  Stomach and Duodenum adult patients who use NSAIDs. Patients with perforated gastric ulcers are more likely to be older and to have significant comorbid illnesses, making their prognosis less favorable. As with perforated duodenal ulcers, there has been debate regarding the choice of surgery for perforated gastric ulcers. No prospective controlled trials comparing simple closure and definitive treatment of a perforated gastric ulcer have been performed. An attempt should be made to determine the type of gastric ulcer. The optimal treatment of an angular notch ulcer along the lesser curvature of the stomach should entail an antrectomy followed by a Billroth type I gastroduodenostomy. High lesser curvature ulcers near the posterior wall can be managed with the Pauchet procedure as mentioned. Simple closure should be considered in prepyloric ulcers, especially when they are small. The role of a vagotomy or indeed gastric resection in these ulcers is unclear. Although the role of H. pylori in perforated prepyloric or antral gastric ulcers after their patch closure is not known, it is probably similar to the organism’s role in perforated duodenal ulcer. The advocates for primary resection in perforated gastric ulcers argue that mortality rates after gastrectomy are not increased and that the rate of postoperative ulcer-related complications is reduced. The arguments for primary resection include the possibility that the ulcer is malignant. Malignancy is seen in approximately 6% of perforated gastric ulcers.152 In a retrospective series comprising 287 perforated gastric ulcers, death occurred in 21.5% of patients who underwent patch closure alone and in 24.3% of those who underwent gastrectomy.153 A patch closure is a far more popular operation for perforated gastric ulcers than resection. Three randomized trials that compared laparoscopic repair with open repair in perforated ulcers favored the use of laparoscopic technique.154-156 In one study, Siu and associates155 randomly assigned 121 patients without other ulcer complications to undergo one of the two repairs. Laparoscopic repair was quicker to perform and led to less postoperative pain, smaller analgesic requirement, and shorter hospitalization. The same researchers later reported the routine use of laparoscopic repair in a cohort of 172 patients with perforated ulcers, about 80% of which were duodenal ulcers (Fig. 53-3).157 Conversion to open surgery occurred in 37 patients (around 1 in 5) because of perforations larger than 10 mm, non-juxtapyloric gastric ulcer, or perforations that could not be identified at laparoscopy. Persistent leaks after patch repair occurred in 2 patients who required

laparotomy and gastrectomy. The overall mortality rate was 8.1%. Several series studied risk factors for mortality in patients undergoing surgery for perforated ulcers. Boey and asso­ ciates158 identified preoperative shock, major medical illnesses, and perforation for longer than 12 hours as important risk factors. Irvin159 identified two other risk factors—age older than 70 years and use of NSAIDs. Definitive surgery (vagotomy and gastrectomy) also led to a higher mortality in Irvin’s series. The deaths, however, occurred in older patients who had other, concomitant ulcer complications.

OBSTRUCTION Medical Management

Patients with obstructing peptic ulcers are often volume depleted. The loss of fluid, hydrogen ions, and chloride ions in the vomitus leads to hypochloremic, hypokalemic metabolic alkalosis. The patient should be volume resuscitated with normal saline followed by potassium replacement once urine output is adequate. In severely malnourished patients, parenteral nutrition should be considered. A nasojejunal tube can be often inserted at endoscopy to bypass the stenosis and then used for enteral feeding. If the patient also has a nasogastric tube, however, management of two tubes is difficult. Decompression of the stomach by a largebore, preferably Salem-sump nasogastric catheter is an integral part of the management. This serves the following purposes: relief of vomiting, monitoring of fluid loss, and decompression of the stomach so it can regain tone. A high-volume non–bile-stained aspirate distinguishes gastric outlet obstruction from a high small bowel obstruction. The use of an intravenous PPI effectively reduces gastric acid output, making fluid and electrolyte management easier. PPI therapy also initiates ulcer healing, ameliorates inflammatory edema, and assists in resolving obstruction, although randomized controlled outcome studies of this issue are lacking. Around half of patients improve during the period of nasogastric suctioning, volume resuscitation, rehydration, and acid suppression. Improvement is especially noticeable in patients with active ulceration and acute edema. Surgery is thus deferred until after an adequate trial of conservative management. Other factors that may influence the decision to proceed to surgery are chronicity, a history of previous ulcer complication, and the patient’s age and general medical condition. Many authorities argue for initial endoscopic dilation before surgery.

Endoscopic Management

Figure 53-3.  Laparoscopic view of a perforated duodenal ulcer (arrow) with fibrinous exudate on the adjacent peritoneum.

Endoscopic balloon dilation has been used successfully in patients with benign gastric outlet obstruction (Fig. 53-4). During endoscopic examination, the stenosis is traversed by means of a biliary-type guidewire with a flexible hydrophilic tip. A low-compliance balloon is then passed over the guidewire. The use of a balloon is preferred because its inflation produces an even radial force, which has a theoretical advantage over the longitudinal shearing force associated with the use of conventional dilators. The availability of through-the-scope balloons that can be passed via the small channel of an endoscope enables the dilation to be seen and monitored. The procedure is typically performed with fluoroscopic guidance. A regimen of gradual dilation over two or three sessions seems sensible. The largest diameter of stenosis at which symptoms occur is unclear. Many authorities recommend dilation to 15 mm, which is often associated with relief of symptoms. The presence of gastric atony also contributes to symptoms. The risk of

Chapter 53  Treatment of Peptic Ulcer Disease

A

B

perforation rises with the size of balloon. Almost all of the perforations in one series occurred after dilation with a 20-mm balloon.160 Several series describing the use of endoscopic balloon dilation in the treatment of benign gastric outlet obstruction have been published.160-162 They varied in case mix, methods of dilation, and duration of follow-up. Most reported immediate relief of obstruction in 78% to 100% of cases. Recurrences, however, are common. In a series of 41 patients with a strong ulcer diathesis, half of the patients experienced relapse with recurrent obstruction, hemorrhage, or active ulceration after a median follow-up of 39 months.160 In a later series involving a handful of patients in whom H. pylori status was better defined, balloon dilation followed by H. pylori eradication in those shown to be infected led to more sustained symptom relief.163 Results of these studies suggest that endoscopic balloon dilation produces better results in patients with acute inflammatory edema than those with chronic scarring and fibrosis.

Surgical Management

A variety of operations have been described for obstructing duodenal, pyloric channel, and prepyloric ulcers. They include truncal vagotomy together with either a drainage procedure (either gastrojejunostomy or pyloroplasty) or an antrectomy. In the unusual event of an obstructing prepyloric gastric ulcer, an antrectomy followed by a Billroth type I gastroduodenostomy is the procedure of choice. Surgical management principles for gastric outlet obstruction were established in an era when (1) this peptic complication was common, (2) recurrence was likely without definitive surgical therapy, (3) potent antisecretory therapy and endoscopic dilation techniques were not available, and (4) parenteral nutrition options were limited. Although patients often presented with substantial debility due to profound weight loss and electrolyte disturbances, expectant medical management of the outlet obstruction ran the risk of dangerously delaying surgery while the malnutrition progressed. Consequently, early operation often was advised. Today, gastric outlet obstruction from peptic ulcer disease is uncommon, ulcer recurrence is unlikely once H. pylori and NSAIDs have been eliminated, excellent antisecretory therapy can be offered, there are a number of endoscopic techniques for dilating stenosis, and total parenteral nutri-

Figure 53-4.  A through-the-scope dilation of an obstructed pylorus caused by an ulcer. The procedure was performed under fluoroscopic guidance. A dual-channel endoscope with a 3.7-mm therapeutic channel was used. A, The stricture was first traversed with a biliary-type guidewire (arrowhead). A through-the-scope balloon was passed over the guidewire across the stricture. A waist, representing the stricture (arrow), was observed (B) and was nearly abolished on balloon inflation (C).

C

tion is widely available (see Chapter 5). Therefore, an immediate decision regarding the need for surgery is generally not necessary for a patient who presents with gastric outlet obstruction. The problem can be managed with medical and endoscopic means in approximately 70% of cases, and only 30% eventually require one of the previously mentioned operations to bypass the gastric outlet obstruction.

KEY REFERENCES

Barkun A, Bardou M, Marshall JK. Nonvariceal Upper GI Bleeding Consensus Conference Group: Consensus recommendations for managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2003; 139:843-57. (Ref 113.) Chan FK, Hung LC, Suen BY, et al. Celecoxib versus diclofenac and omeprazole in reducing the risk of recurrent ulcer bleeding in patients with arthritis. N Engl J Med 2002; 347:2104-10. (Ref 78.) Chan FK, Wong VW, Suen BY, et al. Combination of a cyclo-oxygenase-2 inhibitor and a proton pump inhibitor for prevention of recurrent ulcer bleeding in patients at very high risk: A double-blind randomized trial. Lancet 2007; 369:1621-6. (Ref 85.) Kahi CJ, Jensen DM, Sung JJ, et al. Endoscopic therapy versus medical therapy for bleeding peptic ulcer with adherent clot: A meta-analysis. Gastroenterology 2005; 129:855-62. (Ref 122.) Kearney PM, Baigent C, Godwin J, et al. Do selective cyclo-oxygenase-2 inhibitors and traditional non-steroidal anti-inflammatory drugs increase the risk of atherothrombosis? Meta-analysis of randomized trials. BMJ 2006; 332:1302-8. (Ref 90.) Laine L, Peterson WL. Bleeding peptic ulcer. N Engl J Med 1994; 331:717-27. (Ref 118.) Lanas A, Garcia-Rodriguez LA, Arroyo MT, et al. Effect of antisecretory drugs and nitrates on the risk of ulcer bleeding associated with nonsteroidal anti-inflammatory drugs, antiplatelet agents, and anticoagulants. Am J Gastroenterol 2007; 102:507-15. (Ref 76.) Lau JY, Leung WK, Wu JC, et al. Omeprazole before endoscopy in patients with gastrointestinal bleeding. N Engl J Med 2007; 356:163140. (Ref 142.) Leontiadis GI, Sreedharan A, Dorward S, et al. Systematic reviews of the clinical effectiveness and cost-effectiveness of proton pump inhibitors in acute upper gastrointestinal bleeding. Health Technol Assess 2007; 11:1-164. (Ref 99.) McGettigan P, Henry D. Cardiovascular risk and inhibition of cyclooxygenase: A systematic review of the observational studies of selective and nonselective inhibitors of cyclooxygenase 2. JAMA 2006; 296:1633-44. (Ref 92.) Rockall TA, Logan RF, Devlin HB, Northfield TC. Incidence of and mortality from acute upper gastrointestinal haemorrhage in the United Kingdom. Steering committee and members of the National Audit of Acute Upper Gastrointestinal Haemorrhage. BMJ 1995; 311:222-6. (Ref 110.)

885

886

Section VI  Stomach and Duodenum Saeed ZA, Winchester CB, Michaletz PA, et al. A scoring system to predict rebleeding after endoscopic therapy of nonvariceal upper gastrointestinal hemorrhage, with a comparison of heat probe and ethanol injection. Am J Gastroenterol 1993; 88:1842-9. (Ref 115.) Sung JJY, Tsoi KK, Lai LH, et al. Endoscopic clipping versus injection and thermocoagulation in the treatment of non-variceal upper gastrointestinal bleeding: A meta-analysis. Gut 2007; 56:1364-73. (Ref 133.) Sung JJY, Barkum A, Kuipers EJ, et al. Intravenous esomeprazole for prevention of recurrent peptic ulcer bleeding. Ann Int Med 2009; 150:455-64. (Ref 140.)

Vergara M, Calvet X, Gisbert JP. Epinephrine injection versus epinephrine injection and a second endoscopic method in high risk bleeding ulcers. Cochrane Database Syst Rev 2007; 2:CD005584. (Ref 135.) Vergara M, Catalan M, Gisbert JP, Calvet X. Meta-analysis: Role of Helicobacter pylori eradication in the prevention of peptic ulcer in NSAID users. Aliment Pharmacol Ther 2005; 21:1411-18. (Ref 96.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

54 Adenocarcinoma and Other Tumors of the Stomach Julian A. Abrams and Timothy C. Wang

CHAPTER OUTLINE Epidemiology  887 Etiology and Pathogenesis  888 Helicobacter pylori Infection  889 Dietary Factors  892 Cigarette Smoking  892 Alcohol  892 Obesity  892 Inherited Predisposition  892 Genetics  893 Premalignant Conditions  895 Chronic Atrophic Gastritis  895 Intestinal Metaplasia  895 Dysplasia  895 Gastric Polyps  896 Previous Gastrectomy  896 Peptic Ulcer Disease  896 Ménétrier’s Disease  897 Screening and Prevention  897 Screening and Surveillance  897 Prevention  897

Gastric tumors are defined as benign or malignant based on their potential to metastasize. Gastric adenocarcinoma makes up the majority of malignant gastric tumors, and will be referred to as gastric cancer in this chapter.

EPIDEMIOLOGY Gastric cancer remains the second leading cause of cancer mortality in the world,1 although the overall incidence is declining.2 The incidence of gastric cancer in Western countries has decreased dramatically over the past century.3 For example, gastric cancer mortality has decreased 86% since 1950 in the United States, and the incidence of gastric cancer has diminished four-fold since 1930 to approximately 8 cases per 100,000 people.4,5 As recently as 1930, gastric cancer was the leading cause of cancer mortality in the United States for men and the third leading cause for women.6 Gastric cancer is now the seventh leading cause of cancer mortality in the United States.5 It was estimated that in 2008, approximately 21,500 Americans would be diagnosed with gastric cancer and 10,880 would die of it.7 There is great geographic variation in gastric cancer incidence, with the highest incidence rates in the Far East (Fig. 54-1). Eastern Europe and Central and South America also have high incidence rates, and the lowest incidence

Clinical Features  898 Diagnosis  899 Endoscopy  899 Upper Gastrointestinal Series  900 Computed Tomography Gastrography  900 Serum Markers  900 Classification and Staging  900 Methods of Staging  901 Prognosis and Treatment  902 Surgical Therapy  902 Endoscopic Mucosal Resection and Submucosal Dissection  902 Chemotherapy  903 Chemoradiation  903 Intraperitoneal Chemotherapy  904 Unresectable Disease  904 Gastric Lymphoma  904 Gastric Carcinoid Tumors  904 Gastrointestinal Stromal Tumors  904 Miscellaneous Tumors  904

rates are observed in North America, North Africa, South Asia, and Australia.6 Although gastric cancer was common in industrialized countries in the past, the latest epidemiologic data indicate that more than 60% of new cases of gastric cancer are in developing countries, reflecting a more rapid decline in developed countries. In the United States, the median age of diagnosis is 71 years, with the highest proportion (28%) diagnosed between the ages of 75 and 84.8 In Japan, a country with a high incidence of gastric cancer, the mean age of diagnosis is roughly a decade earlier, perhaps reflecting lead-time bias due to widespread screening. The incidence of gastric cancer in men is approximately twice that in women (Table 54-1).3 The incidence of gastric cancer in blacks in the United States is nearly double that in whites.5 Native Americans and Hispanics also have a higher risk of development of gastric cancer than whites.9 Lower socioeconomic status is associated with a much higher incidence of gastric cancer.3 In the United States, the distribution of gastric cancer within the stomach is 39% in the proximal third, 17% in the middle third, 32% in the distal third, and 12% involving the entire stomach.10 In contrast to the pattern seen with noncardia tumors, the incidence rates of gastric cardia cancer are rising.2,11 Dietary, environmental, and genetic risk factors for gastric adenocarcinoma are listed in Table 54-2, some of which are or may be protective.

887

888

Section VI  Stomach and Duodenum

Figure 54-1.  Worldwide incidence of gastric cancer in men. (From Parkin DM. International variation. Oncogene 2004; 23:6239-40.)

<5.9

<9.6

<14.6

<23.6

<70.0

Age standardized rate per 100,000

Table 54-1  Gastric Cancer Incidence and Mortality Rates per 100,000 US Population by Race from 2001 to 2005 INCIDENCE

All races White Black Asian/Pacific Islander American Indian/Alaska Native Hispanic

MORTALITY

Men

Women

Men

Women

11.3 10.0 17.4 18.6 16.8 15.5

5.5 4.7 8.9 10.5 7.7 9.5

5.7 5.0 11.5 10.1 9.9 8.7

2.9 2.5 5.5 5.9 5.2 4.9

US, United States. Data from Age-Adjusted SEER Incidence and US Death Rates and 5-Year Relative Survival Rates. National Cancer Institute, Bethesda, Md., 2008. Accessed at http://seer.cancer.gov/cgi-bin/csr/1975_2005.

ETIOLOGY AND PATHOGENESIS Gastric cancer can be subdivided using the Lauren classification into two distinct histologic subtypes with different epidemiologic and prognostic features (Fig. 54-2).12 The intestinal type of cancer is characterized by the formation of gland-like tubular structures with features reminiscent of intestinal glands. This type of gastric cancer is more closely linked to environmental and dietary risk factors, tends to be the predominant form in regions with a high incidence of gastric cancer, and is the form of cancer that is now declining worldwide. The diffuse type of cancer lacks glandular structure and consists of poorly cohesive cells that infiltrate the wall of the stomach. It is found at the same frequency throughout the world, occurs at a younger age, and is associated with a worse prognosis than the intestinal form. Extensive involvement of the stomach by the diffuse type can result in a rigid and thickened stomach, a condition referred to as linitis plastica. Adenocarcinoma of the stomach can also be classified into proximal tumors (gastric cardia and gastroesophageal junction) and distal tumors (fundus, body and antrum of the stomach). Distal tumors have been declining, whereas proximal tumors have been increasing (see Chapter 46).

It is now believed that the development of intestinal-type cancers occurs through a multistep process in which the normal mucosa is sequentially transformed into a hyper­ proliferative epithelium, followed by an early adenoma, late adenoma, and then carcinoma. In colon cancer, the evidence is strong that each step in the transition is associated with a specific gene mutation,13 but evidence that gastric cancer follows a comparable sequence of genetic events has been lacking. However, in gastric and colon cancer, it does appear that deoxyribonucleic acid (DNA) mutations are established over time in stem cells; in intestinal metaplasia these mutations spread through the stomach through a process involving crypt fission and monoclonal conversion of glands.14 The contention that the pathogenesis of intestinal-type gastric cancer is a multistep process is supported mainly by the observation that both atrophic gastritis and intestinal metaplasia are found in higher incidences in patients with intestinal-type cancer and in countries with a high incidence of gastric cancer (see Chapter 51).15 This multistep model of intestinal-type gastric cancer, developed in large part by Correa and colleagues,16,17 postulates that there is a temporal sequence of preneoplastic changes that eventually lead to the development of gastric cancer. A common feature of the initiation and progression

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach to intestinal-type gastric cancer is chronic inflammation. Helicobacter pylori infection is the primary cause of gastric inflammation and the leading etiologic agent for gastric cancer (see Chapter 50). In a subset of patients infected with H. pylori, the inflammatory process leads to the development of atrophic gastritis (with loss of glandular tissue) followed by progression to intestinal metaplasia, dysplasia, early gastric cancer, and, eventually, advanced gastric cancer (Fig. 54-3). The current view is that all stages prior to the development of high-grade dysplasia are potentially reversible, although this concept is still somewhat controversial, it has been supported by a number of studies in Table 54-2  Risk Factors Including Protective Factors for Gastric Adenocarcinoma Definite

Probable

Possible

Questionable

Helicobacter pylori infection Chronic atrophic gastritis Intestinal metaplasia Dysplasia* Adenomatous gastric polyps* Cigarette smoking History of gastric surgery (esp. Billroth II)* Genetic factors Family history of gastric cancer (first-degree relative)* Familial adenomatous polyposis (fundic gland polyps)* Hereditary nonpolyposis colorectal cancer* Peutz-Jeghers syndrome* Juvenile polyposis* High intake of salt Obesity (adenocarcinoma of cardia only) Snuff tobacco use History of gastric ulcer Pernicious anemia* Regular aspirin or NSAID use (protective) Low socioeconomic status Ménétrier’s disease High intake of fresh fruits and vegetables (protective) High ascorbate intake (protective) Hyperplastic and fundic gland polyps High intake of nitrates High intake of green tea (protective)

*Surveillance for cancer is suggested in patients with this risk factor. NSAID, nonsteroidal anti-inflammatory drug.

A

B

animal models.18,19 Unlike the situation observed with colon cancer, the precise genes involved in each step of this progression are still not defined. Furthermore, during endoscopy the premalignant stages of gastric cancer are not as readily identifiable as those of colon cancer, and many gastric carcinomas are very heterogeneous, containing a large percentage of stromal cells. These stromal cells, which include cancer-associated fibroblasts known to promote tumor growth, have been reported to show distinct genetic and epigenetic changes that may confound tumor analysis.20 This feature makes characterization of the timing of specific gene mutations in gastric cancer difficult at best. Currently the role of chronic inflammation in the diffuse type of gastric cancer remains to be clarified, and the similarities if any to the proposed pathway in Figure 54-3 for the intestinal type of cancer.

HELICOBACTER PYLORI INFECTION

H. pylori is a gram-negative microaerophilic bacterium that infects nearly half of the world’s population and is recognized as the primary etiologic agent for gastric cancer (see Chapter 50). Indeed, H. pylori has been classified as a class I (or definite) carcinogen by the International Agency for Research on Cancer (IARC), a branch of the World Health Organization (WHO). Infection with H. pylori has been found in every population studied, although the prevalence is higher in developing countries.21,22 The natural history of chronic H. pylori infection includes three possible outcomes23: (1) superficial gastritis, in which most patients remaining asymptomatic; (2) duodenal ulcer phenotype, which occurs in 10% to 15% of infected subjects; and (3) gastric ulcer/gastric cancer phenotype, which is the least common in the United States. In general, the risk for gastric cancer is dependent on the types of gastritis, and an increased risk is associated with a low acid state. H. pylori–induced duodenal ulcer disease is associated with a high gastric acid output as well as a reduced risk for developing gastric cancer.24 Patients with H. pylori– associated gastric ulcer disease exhibit low gastric acid output, and their ulcers are typically associated with preneoplastic changes of atrophic gastritis and metaplasia. Overall, studies suggest that H. pylori–infected patients are at risk for development of chronic atrophic gastritis at a rate of 1% to 3% per year of infection.17,25,26 Thus, those patients who are genetically predisposed to forming atrophic gas­

Figure 54-2.  Histopathologic types of gastric cancer. A, The intestinal type of gastric adenocarcinoma is characterized by the formation of tubular structures mimicking intestinal glands. B, The diffuse type of gastric adeno­carcinoma contains singly invasive tumor cells that frequently contain abundant mucin and that lack any glandular structure. (Courtesy of Rhonda K. Yantiss, MD, Boston, Mass.)

889

890

Section VI  Stomach and Duodenum Normal gastric mucosa

H. pylori Salt?

At which point does antimicrobial eradication of H. pylori interrupt this cascade?

Superficial gastritis

Chronic inflammation

Atrophic gastritis

Intestinal metaplasia and SPEM

β-carotene?

Recruitment of BM-derived stem cells

Higher gastric pH Bacterial overgrowth and nitrate reduction

Salt? N-nitroso carcinogens

Dysplasia

Chronic inflammation and reactive oxygen species

Carcinoma Figure 54-3.  Proposed Correa pathway of pathologic events in gastric adenocarcinoma. In well-differentiated, intestinal-type gastric cancer, histopathologic studies have indicated that chronic Helicobacter pylori infection progresses over decades through stages of chronic gastritis, atrophy, intestinal metaplasia, dysplasia, and cancer. The development of cancer has been attributed to alterations in DNA caused by chronic inflammation, recruitment and engraftment of bone marrow-derived cells, an imbalance between epithelial cell proliferation and apoptosis, and, in a milieu of atrophy and achlorhydria, gastric colonization by enteric bacteria with nitrate reductase activity, which facilitates the formation of carcinogenic nitrosamines. Corpus-predominant atrophy, or the loss of specialized glandular cell types such as parietal and chief cells, appears to be the critical initiating step in the progression toward cancer. BM, bone marrow; SPEM, spasmolytic polypeptide-expressing metaplasia (also known as pseudopyloric metaplasia, mucous metaplasia, or antralization). (From Fox JG, Wang TC. Inflammation, atrophy, and gastric cancer. J Clin Invest 2007; 117:60-9.)

tritis in response to H. pylori infection are predisposed to gastric cancer. Although Helicobacter infection is associated with both intestinal-type and diffuse-type adenocarcinomas, the mechanisms responsible for the formation of intestinal-type adenocarcinoma have been better studied and are focused on here. The association of H. pylori with mucosa-associated lymphoid tissue (MALT) lymphoma is discussed briefly at the end of this chapter and in more detail in Chapter 29. The increased risk of development of gastric adenocarcinoma due to H. pylori infection depends on multiple factors including the strain of bacteria, host genetic factors, the duration of infection, and the presence or absence of other environmental risk factors (e.g., poor diet, smoking, etc.). In a Japanese cohort of 1526 patients with peptic ulcer disease, nonulcer dyspepsia, and gastric hyperplasia, only those infected with H. pylori developed gastric adenocarcinoma during follow-up (2.9% vs. 0%, P < 0.001).27 Additional cohort studies from China and Taiwan have reported similar

findings.28,29 At least in Western countries, the association between H. pylori and gastric cancer appears to be confined to noncardia gastric tumors.30 Potential mechanisms for H. pylori–induced gastric carcinogenesis include host factors, bacterial factors, environmental factors, and interactions among all three factors. Our latest understanding suggests that a combination of a virulent bacterial strain, a genetically permissive host, and a favorable gastric environment are necessary for disease to occur. The most important factor appears to be the induction of chronic inflammation by H. pylori infection. Several aspects of the inflammatory milieu have been implicated as carcinogens; they include increased oxidative stress and the formation of oxygen free radicals leading to DNA damage, increased CD4+ T cells and myeloid cells, and elevated proinflammatory cytokine production, all leading to accelerated cell turnover, reduced apoptosis, and the potential for faulty or incomplete DNA repair.31 Indeed, recent studies suggest that animals with deficient DNA repair mechanisms display more severe gastric dysplasia after chronic infection with H. pylori.32 Thus, evidence to date clearly indicates that the most important cofactor in the induction of Helicobacter-related disease is the host immune response. Indeed, chronic inflammation has been linked to a large number of cancers. H. pylori infection leads to innate and adaptive immune responses (see Chapter 2). Initiation of the innate immune response to H. pylori is just beginning to be unraveled. Classically, the innate immune system consists of professional antigen presenting cells (APCs) such as macrophages, dendritic cells, and in some cases epithelial cells. Recent work supports a role for pattern recognition receptors (Toll-like receptors [TLRs]) in the initial response to Helicobacter colonization and the subsequent induction of the adaptive response. The most convincing evidence to date implicates TLR2 as the major TLR in Helicobacter species recognition.33 A role for TLRs 4, 5, and 9 remains more contro­versial.34-37 TLR4, along with CD14 and MD-2, serves as the receptor for Escherichia coli lipopolysaccharide (LPS) and probably H. pylori LPS and thus may be involved as well. Chronic inflammation appears necessary for the progression through atrophy to gastric cancer. Disease mechanisms are difficult to study in human infection, and therefore much of our understanding of the immune response to Helicobacter organisms comes from work performed in the mouse model of infection. Different inbred strains of mice respond to infection with varying degrees of disease susceptibility, and several knockout models have helped to elucidate the roles of individual components of the immune response in disease. The C57BL/6 mouse is a susceptible inbred strain, in which initial colonization of the antrum by bacteria later spreads to the body or corpus, leading to severe chronic inflammation and increases in apoptosis (programmed cell death) and proliferation. The alterations in cellular turnover are associated with a loss of parietal and chief cells (atrophy), intestinal metaplasia, and dysplasia, followed by invasive gastric adenocarcinoma in mice 14 to 22 months after infection.38,39 Genetic manipulation of the C57BL/6-susceptible murine strain has facilitated detailed study and has thus led to a deeper understanding of genetic factors that promote murine gastric cancer, and in particular, the role of the adaptive immune response. For example, gastric Helicobacter infection in mice deficient in lymphocytes, including mice with recombinase-activating gene (RAG) deficiency, severe combined immunodeficiency, or T cell deficiency, does not result in tissue damage, cell lineage alterations, or

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach the metaplasia-dysplasia-carcinoma sequence.40,41 Infection in B cell–deficient mice (which retain a normal T cell response) results in severe atrophy and metaplasia identical to that seen in infected wild-type mice.41 Taken together these studies underscore the crucial role of CD4+ T lymphocytes in orchestrating gastric neoplasia. How do CD4+ T lymphocytes contribute to gastric cancer progression? Susceptible mouse strains, such as C57BL/6, mount a strong helper T cell type 1 (Th1) interferon-γ (IFNγ), interleukin-12 (IL-12) type of immune response, whereas resistant strains, such as the BALB/c, have an opposite Th2 response (IL-4, IL-5).39,42,43 A Th2 response is associated with protection from mucosal damage despite the inability to eliminate bacterial colonization and in fact is often asso­ ciated with higher bacterial colonization rates. Mouse strains such as the C3H, which has a mixed Th1/Th2 cytokine profile, show intermediate disease, suggesting that cytokines within an immune response interact to form a continuum of disease rather than discrete disease states. More recently, Th17 cells, which express IL-17, have been shown to be an important component of H. pylori–induced gastritis. Although the composite immune milieu most likely dictates disease manifestations, studies are beginning to define the role of individual cytokines in the predisposition to disease. This is best illustrated in the IFN-γ knockout mice, in which a lack of IFN-γ protects infected mice from atrophy.39,43 On the other hand, mice lacking IL-10, a cytokine that acts to dampen an immune response, develop severe atrophic gastritis in response to infection.39-43 More recently, genetic murine models have illustrated the importance of the IL-6–IL-11 family of cytokines in the development of gastric cancer.44 Manipulation of the immune response within wild-type strains confirms the central role of the Th1/Th2 response in producing disease. For example, infection with the intestinal helminth Heligmosomoides polygyrus skews the immune response toward Th2 polarization and protects the C57BL/6 host from Helicobacter-induced atrophy and metaplasia.45 This mouse model mimics both the parasitic infection status and the paradoxical low gastric cancer–high H. pylori infection rates seen in areas of Africa, potentially explaining this apparent inconsistency. These observations in mice led to human studies in Africa and Latin America that confirmed that geographic regions with low gastric cancer rates had much higher Th2 relative to Th1 immune responses to H. pylori.46,47 In general, the increased Th2-type responses were found in areas where serum immunoglobulin E (IgE) levels were high and the prevalence of intestinal parasitism by helminths is greater than 50%. These findings further stress the importance of the host response to infection and suggest the possibility that manipulation of the genetically predetermined host cytokine profile in response to environmental challenges may lessen or exacerbate the disease process. There is a great deal of genetic diversity between strains of H. pylori owing to point mutations, insertions, deletions, and base-pair substitutions within its genome. Several strains may infect a single individual, and existing strains can undergo mutations and change over time.48,49 Despite this genetic diversity, several genes are recognized as risk factors for gastric carcinoma, including the cag pathogenicity island, the vacA gene, and the babA2 gene. The H. pylori genome is 1.65 million base pairs and codes for approximately 1500 genes, two thirds of which have been assigned biological roles.50 The function of the remaining one third of the genome remains obscure. Factors that contribute to carcinogenesis include those that enable the

bacteria to effectively colonize the gastric mucosa, those that incite a more aggressive host immune response, and those that affect host cell-growth signaling pathways. Motility toward epithelial cells of the stomach is a vital feature of H. pylori survival tactics. This function is ensured by several factors, including spiraling movement (FlaA and FlaB proteins), which are designed to navigate the thick gastric mucus and through efficient modifications of the extracellular matrix and mucus layer, thus decreasing viscosity and allowing bacterial penetration.51,52 In addition, H. pylori expresses a variety of genes that contribute to buffering of stomach acid in order to maintain a relatively neutral pH. This includes a urease gene cluster, consisting of seven genes, of which UreA/UreB complex (comprising the urease enzyme) codes for 10% of the protein of H. pylori and is vital for its survival. A significant proportion (e.g., ≈20%) of H. pylori organisms can be found adherent to the surface of gastric mucous cells. Occasionally H. pylori can also be found intracellularly, particularly in preneoplastic and neoplastic lesions.53 Adhesion of the bacteria to the epithelial layer is facilitated by a large family of 32 related outer-membrane proteins (Hop proteins) that include the adhesins. One of the bestcharacterized adhesins is BabA, which is encoded by the strain-specific gene babA2, a member of a highly conserved family of outer membrane proteins. BabA binds to the fucosylated Lewis B blood group antigen on gastric epithelial cells and forms a scaffold apparatus that allows bacterial proteins to enter host epithelial cells. Bacterial strains that possess the babA2 gene adhere more tightly to epithelial cells, promote a more aggressive phenotype, and are associated with a higher incidence of gastric adenocarcinoma.54 The cag pathogenicity island is approximately 40 kb and contains 31 genes. The terminal gene of this island, cagA, is often used as a marker for the entire cag locus. Compared with cagA-negative (cag−) strains, cag-positive (cagA+) strains are associated with more severe inflammation, higher degrees of atrophic changes, and a greater chance of progressing to gastric adenocarcinoma.55-58 The estimated risk has ranged from an odds ratio of 2 to as high as 28.4.23 However, many of the genes adjacent to cagA code for a type 4 secretion system (TFSS), often viewed as a molecule needle that injects bacterial proteins (such as CagA) into host cells. The remarkable finding that CagA is injected into host cells, where it is phosphorylated by Src and c-Abl kinases, has raised the possibility that CagA could directly promote growth, migration, and transformation. Indeed, transgenic expression of H. pylori CagA induces gastrointestinal (GI) and hematopoietic neoplasms in mice.59 Other genes within the pathogenicity island are also believed to be important for disease (cagE or picB, cagG, cagH, cagI, cagL, cagM) because they appear to be required for in vitro epithelial cell cytokine release, although they do not seem to have as great an effect on immune cell cytokine activation as cagA.60-62 These findings may explain the attenuated inflammatory response and lower cancer risk with cagA− strains in vivo.63-66 All strains of H. pylori carry the vacA gene, which codes for a pore-forming vacuolating toxin, but expression of vacA differs according to allelic variation. Approximately 50% of H. pylori strains express the vacA protein, which has been shown to be a very powerful inhibitor of T cell activation in vitro.67 Although vacA and cagA map to different loci within the H. pylori genome, the vacA protein is commonly expressed in cagA+ strains. There are various forms of vacA, and the s1m1 strains are highly toxigenic. Other bacterial virulence factors, such as cagE, may play a role in the modu-

891

892

Section VI  Stomach and Duodenum lation of apoptosis and the host inflammatory response, thereby contributing to disease manifestations. Indeed, “virulent strains” (cagA+, cagE+, and VacA+ s1m1) appear to be more potent inducers of proinflammatory mediators than “nonvirulent strains” (cagA−, cagE−, and VacA−), possibly explaining the higher association of cagA+ strains with gastric cancer.68

DIETARY FACTORS

Numerous dietary factors have been implicated as risk factors for gastric cancer. The decline in gastric cancer rates has coincided with the widespread use of refrigeration and the concomitant higher intake of fresh fruits and vegetables and lower intake of pickled and salted foods. Use of refrigeration for more than 10 to 20 years has been associated with a decreased risk of gastric cancer.17,69 Lower temperatures reduce the rate of bacterial, fungal, and other contaminants of fresh food, as well as the bacterial formation of nitrites. Additionally, high intake of highly preserved foods may be associated with increased gastric cancer risk,70 likely because of higher contents of salt, nitrates, and polycyclic aromatic amines.71 Much attention has been given to the effects of high nitrate intake. When nitrates are reduced to nitrite by bacteria or macrophages, they can react with other nitrogenated substances to form N-nitroso compounds that are known mitogens and carcinogens.72,73 In rats, N-nitroso compounds have been shown to cause gastric cancer.74 However studies trying to link N-nitroso exposure to gastric cancer risk have been inconclusive, perhaps reflecting the fact that nitrate intake does not necessarily correlate with nitrosation levels.75,76 A Swedish cohort study found a nearly two-fold increased risk of gastric cancer associated with high dietary nitrate intake.70 However, separate large cohort studies from Europe did not demonstrate an association between nitrate intake and risk of gastric cancer.77,78 Another factor implicated in the development of gastric cancer is a diet high in salt (pickled foods, soy sauce, dried and salted fish and meat). High salt intake has been associated with higher rates of atrophic gastritis in humans and animals in the setting of Helicobacter infection79 and increases the mutagenicity of nitrosated food in animal models.17 High salt diets are associated with a roughly twofold increased risk of gastric cancer.80,81 Cohort and casecontrol studies have also found an increased risk of gastric cancer associated with processed meat intake.70,82 Possible mechanisms include higher bacterial loads, up-regulation of H. pylori cagA expression, and increased cell proliferation and p21 expression.79,83,84 Epidemiologic studies have had inconsistent findings with regard to fruit and vegetable consumption and risk of gastric cancer. A cohort study from Japan found significantly decreased risks of gastric cancer associated with increased vegetable and fruit intake.85 A Swedish cohort study demonstrated a reduced risk of gastric cancer associated with high vegetable intake, but no association was seen with amount of fruit consumption.86 A large cohort study of nearly 500,000 adults in the United States and a separate nested case-control study from Europe failed to find an association between fruit and vegetable intake and gastric cancer risk.87,88 Other foods or dietary factors that have been implicated as potential risk factors for gastric cancer are high intake of fried food, foods high in fat, high intake of red meat, and aflatoxins.82,89-91 Diets with a high intake of fresh fish and antioxidants may be protective.90,92-94 However, there are insufficient data to make definitive conclusions regarding these factors.

CIGARETTE SMOKING

Tobacco has long been established as a carcinogen, and numerous epidemiologic studies have demonstrated an association between cigarette smoking and risk of gastric cancer.95 Several large cohort studies from Europe and Asia have reported a significantly increased risk of gastric cancer among smokers.96-98 A meta-analysis found that compared with never smokers, current smokers had a 1.5-fold increased risk of gastric cancer for the cardia as well as noncardia region.99 The authors also reported an increased association with greater amounts of smoking. Moist snuff is a smokeless tobacco product promoted as an alternative to cigarettes and has reportedly reduced levels of carcinogenic nitrosamines. However, results of a Swedish cohort study demonstrated a 1.4-fold increased risk of noncardia gastric cancer among regular snuff users.100 Snuff exposure also increases the rate of gastric carcinogenesis in H. pylori–infected mice.101

ALCOHOL

Multiple cohort and case-control studies from the United States and Europe have found no significant association between alcohol consumption and cardia or noncardia gastric cancer.98,102,103 A separate population-based casecontrol study in the United States also found no association between any alcohol use and risk of either cardia or noncardia gastric cancer, although a reduced risk was reported with wine consumption (cardia, odds ratio [OR] 0.6; 95% confidence interval [CI]: 0.5 to 0.9; noncardia, OR 0.7; 95% CI: 0.5 to 0.9).104

OBESITY

Obesity is a recognized risk factor for numerous gastrointestinal malignancies.105 Increased body mass index (BMI) appears to be associated with a mild to moderate increased risk of gastric cardia cancer but not for noncardia gastric cancer.106-110 Results of the National Institutes of Health– American Association of Retired Persons (NIH-AARP) Diet and Health Cohort Study demonstrated that marked obesity (BMI = 35 kg/m2) was associated with a significantly increased risk of gastric cardia cancer (hazard ratio, 2.46) but not with noncardia gastric cancer.106 A separate cohort study from the Netherlands also found an increased risk of cardia cancer with increasing BMI.107

INHERITED PREDISPOSITION

As is true for most malignancies, genetic and environmental factors play important roles in the pathogenesis of gastric cancer. Generally, intestinal-type gastric cancer is considered to be largely due to environmental causes (i.e., H. pylori infection), whereas diffuse gastric cancer is considered a primarily genetic malignancy. In the case of intestinal-type gastric cancer, however, assigning relative values to environmental and genetic contributions is complex, given that the major environmental factor, H. pylori, also tends to exhibit familial clustering. Overall, 10% of cases of gastric cancer appear to exhibit familial clustering,111 and family history is likely an independent risk factor for the disease even after controlling for H. pylori status.112,113 In a cohort study of relatives of patients with gastric cancer, siblings had a two-fold increased risk of gastric cancer, adjusted for H. pylori infection.114 In a case-control study from Japan, a positive family history was associated with a significantly increased odds of gastric cancer in women (OR, 5.1), but not in men.115 A study from Scandinavia showed that having a twin with gastric cancer conferred a markedly higher risk for the disease (hazard ratios, 9.9 for monozygotic twins and 6.6 for dizygotic

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach twins), leading the researchers to calculate that heritable factors accounted for 28% of gastric cancers, compared with 10% for shared environmental factors and 62% for nonshared environmental factors.116 Some of the familial clustering seen with intestinal-type gastric cancer may be related to genetic factors that play a role in the host immune response to H. pylori infection. Data from South Korea indicate that individuals with a family history of gastric cancer more frequently have H. pylori infection as well as associated atrophic gastritis or intestinal metaplasia.117 In a case-control study from Scotland, relatives of patients with gastric cancer had a higher prevalence of atrophy and hypochlorhydria, but a similar prevalence of H. pylori infection, compared with controls.118 The greater prevalence of atrophy was confined to those patients with H. pylori infection, suggesting the possibility these individuals were perhaps exhibiting a more vigorous immune response to H. pylori. In a number of model systems, the development of gastric atrophy has been linked to a strong Th1 immune response.43,45,119 Thus, it was postulated that candidate disease-susceptibility genes for gastric atrophy and cancer might be genes that participate in the innate and adaptive immune responses to H. pylori infection. Inflammation is modulated by an array of pro- and anti-inflammatory cytokines, and several genetic polymorphisms have been described that influence cytokine response. IL-1β is an important proinflammatory cytokine and a powerful inhibitor of acid secretion. Thus, the initial report in this area described in the setting of H. pylori infection an association between proinflammatory IL-1 gene cluster polymorphisms (IL-1B encoding IL-1β, and IL-1RN encoding its naturally occurring receptor antagonist, IL-1RA) and neoplastic progression. Individuals with the IL-1β-31*C or -511*T and IL-1RN*2/*2 genotypes were shown in the study to be at higher risk for development of H. pylori– dependent hypochlorhydria and gastric cancer.120 The increased risk of progression to cancer with these genotypes was in the two- to three-fold range compared with noninflammatory genotypes. The initial report was confirmed in other studies.121-125 Subsequently, Hwang and colleagues126 demonstrated that carriers of the IL-1β-511T/T genotype or the IL-1RN*2 allele had higher mucosal IL-1β levels than noncarriers, and they also confirmed the association between the -511T/T genotype and severe gastric inflammation and atrophy. The importance of IL-1β in carcinogenesis has now been demonstrated in a transgenic study, in which stomach-specific expression of human IL-1β in transgenic mice led to spontaneous gastric inflammation and cancer that correlated with early recruitment of myeloid-derived suppressor cells (MDSCs) to the stomach.127 Additional associations with gastric cancer risk have been reported for genetic polymorphisms in tumor necrosis factor-α (TNF-α) and IL-10. Proinflammatory genotypes of TNF-α and IL-10 each were associated with a two-fold higher risk of noncardia gastric cancer. When combined with proinflammatory genotypes of IL-1β and IL-1RN, patients with three or four high-risk genotypes showed a 27-fold greater risk of gastric cancer.128 Additional studies have shown that polymorphisms of the TLR4 gene also increase the risk of gastric cancer. Carriers of the TLR4+896G polymorphism had an 11-fold increased risk of hypochlorhydria, and significantly more severe gastric atrophy and inflammation.129 Accumulated evidence suggests that the genetic predisposition to gastric cancer may be largely determined by the TLR and cytokine responses to chronic Helicobacter infection. The best described form of hereditary gastric cancer is the diffuse gastric cancer that is seen in the presence of a

germline mutation in the gene CDH1, which encodes the cell adhesion molecule E-cadherin. A large New Zealand kindred was found to have a germline mutation in the E-cadherin gene, and similar mutations have been reported in several additional kindreds, all with diffuse-type gastric cancer.130-133 The age of onset of gastric cancer in individuals with CDH1 mutations is less than 40 years but can be highly variable, and the estimated lifetime risk of gastric cancer is close to 70%.134,135 Germline CDH1 mutations are also associated with familial lobular breast cancer.136,137 A small part of the familial clustering of gastric cancer can be attributed to other cancer syndromes (see Chapter 122). Patients with familial adenomatous polyposis (FAP) have a prevalence of gastric adenomas ranging from 35% to 100%, and their risk of gastric cancer is close to 10-fold higher than that of the general population.138 These cancers frequently arise from fundic gland polyps and develop at an early age.139,140 Patients with hereditary nonpolyposis colorectal cancer (HNPCC) syndrome have an approximately 11% risk of developing gastric cancer, pre­dominantly of the intestinal type, with a mean age at diagnosis of 56 years.141 Patients with juvenile polyposis also have a 12% to 20% incidence of gastric cancer.142,143

GENETICS Although atrophy and intestinal metaplasia correlate with gastric cancer risk, direct cell progression through these stages has not been conclusively shown. Indeed, gastric cancer most likely arises from stem or progenitor cells present within the gastric mucosa rather than directly from terminally differentiated metaplastic cells. Investigators have for several decades sought to unravel the mutations responsible for gastric cancer initiation and progression in an attempt to uncover a logical progression of acquired mutations akin to what is seen in colorectal cancer. However, gastric cancer does not follow a pattern like colorectal carcinoma progression, and there is no clear-cut linear sequence of mutations in gastric cancers. There is likely a need for a genome-wide analysis of somatic mutations in gastric cancer. Nevertheless, even if a genome-wide sequencing study is performed, the precise role, if any, that identified mutations play in initiating malignant transformation, rather than cancer progression, may not be clear. Aneuploidy is common in gastric cancer (seen in 60% to 75% of cases), but cytogenetic studies have failed to identify any consistent chromosomal abnormality. Comparative genomic hybridization studies have shown that chro­mosome arms 4q, 5q, 9p, 17p, and 18q exhibit frequent decreases in DNA copy number, whereas chro­ mosomes 8q, 17q and 20q often have increased DNA copy number.144 There is a consensus that TP53 is the most commonly mutated gene in gastric cancer (60% to 70% of gastric cancers) and that mutations in Ras, APC, and Myc are rare.145,146 Loss of heterozygosity (LOH) at the APC locus occurs more commonly. Another genetic abnormally found at high frequency includes deletion or suppression of the fragile histidine triad gene (FHIT) (60%), a tumor suppressor locus on chromosome 3p. Genes that normally inhibit entry into the cell cycle, such as p16 and p27, show diminished expression in nearly one half of gastric cancers.147-152 Absence of p27 expression is associated with a poorer prognosis.147,149 Absence of p16 expression is seen most commonly in poorly differentiated carcinomas but has no measurable impact on patient prognosis.153 Diminished

893

894

Section VI  Stomach and Duodenum expression of p16 and p27 occurs in the absence of detectable mutations and is believed to be secondary to hypermethylation of the respective genes.151 Many of these cancers show hypermethylation of a number of promoter regions, including the MLH1 promoter region, and show the highlevel microsatellite instability (MSI) phenotype (see Chapter 3). Multiple tumor suppressor genes have been shown to be methylated in gastric cancers. Emerging evidence suggests that epigenetic changes, including global hypomethylation and promoter hypermethylation, occurs quite early in gastric carcinogenesis. In addition, it appears that DNA methylation changes also occur in the tumor-associated stromal fibroblasts, suggesting an important role for the tumor microenvironment.20 Overexpressions or amplifications of a number of growth factor pathways have been described, including cyclooxygenase-2 (COX-2) (70%), hepatocyte growth factor/ scatter factor (HGF/SF) (60%), vascular endothelial growth factor (VEGF) (50%), c-met (50%), amplified in breast cancer-1 (AIB-1) (40%), and β-catenin (25%) (Table 54-3).154 Approximately 15% of gastric cancers have been reported to overexpress both epidermal growth factor (EGF) and EGF receptor (EGFR), consistent with an autocrine mechanism. Mutations in PIC3A, a gene that codes for a catalytic subunit of phosphotidylinositol 3-kinase (PI3K), has been found in up to 25% of gastric cancers analyzed.155 In addition, mutations in human protein tyrosine phosphatases (PTPs) were found by the same laboratory in 17% of gastric cancers, with the protein tyrosinase phosphatase–receptor type (PTPRT) the most frequently altered.156

Table 54-3  Genetic Abnormalities in Gastric Adenocarcinoma ABNORMALITIES

APPROXIMATE GENE FREQUENCY (%)

DNA aneuploidy Microsatellite instability Deletion/Suppression TP53 gene Fragile histidine triad gene (FHIT) Adenomatous polyposis coli (APC) gene LOH Deleted in colorectal cancer (DCC) gene LOH Decreased Expression Due to Hypermethylation p16 TFF1 p27 MLH1 E-cadherin Amplification/Overexpression Cyclooxygenase-2 (COX-2) Hepatocyte growth factor (HGF) Vascular endothelial growth factor (VEGF) c-Met Amplified in breast cancer-1 (AIB-1) Beta-catenin EGF/EGFR Mutations PI3K PTPRT

60-75 15-50 60-70 60 50 50 ≈50 ≈50 <50 15-20 50 70 60 50 50 40 25 15 25 17

DNA, deoxyribonucleic acid; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; LOH, loss of heterozygosity; MLH1, human mutL homolog 1; PI3K, phosphatidylinositol 3-kinase; PTPRT, protein-tyrosine phosphatase receptor-type; TFF1, human trefoil factor 1.

Gastric-specific tumor suppressor genes TFF1 (Trefoil factor 1) and RUNX3 (Runt-related transcription factor 3), which have now been identified and may represent “gatekeepers” of the gastric cancer pathway, are logical targets for further study.157,158 Loss of TFF1 has been described in around 50% of gastric carcinomas, and TFF1 knockout mice develop spontaneous gastric antral tumors. Mutations of TFF1 also have been described, and these enhance gastric cancer cell invasion through signaling pathways that include PI3-K and phospholipase-C.159 TFF1 expression is repressed by STAT-3, and activation of STAT-3 is also emerging as a key pathway that leads to gastric cancer.44 RUNX3 most likely suppresses gastric epithelial growth by inducing p21 and Bim, attenuating Wnt signaling, and is altered in 82% of gastric cancers.160 Investigations into these genes and their contributions to the gastric cancer phenotype will prove valuable to our understanding of disease progression. MSI in dinucleotide repeats secondary to defects in DNA mismatch repair genes, such as MLH1 and MLH2 (mutL homologs 1 and 2), have been implicated in the development of colorectal cancer, and in particular the HNPCC syndrome. Patients with HNPCC have an 11% incidence of gastric cancer, suggesting that MSI may also play a role in the development of gastric cancer.141 MSI is found in 15% to 50% of sporadic gastric cancers, with a higher prevalence in the intestinal type of cancer.161-166 Low-level microsatellite activity (e.g., MSI-low) can be found in 40% of areas of intestinal metaplasia in patients with gastric cancer166 and in 14% to 20% of adenomatous polyps.164,166,167 MSI-high (MSI-H) occurs in only 10% to 16% of gastric cancers. MSI is associated with the less frequent occurrence of TP53 mutations, well- to moderately well differentiated histology, and distal tumor location. Studies that have examined the effect of MSI on patient survival have shown inconsistent results.167,168 When the findings are taken together, it would appear that MSI does play a role in the pathogenesis of gastric cancer, likely before the development of intestinal metaplasia (see Fig. 54-3), and is most commonly due to methylation of the MLH1 promoter. The data regarding the genetics of diffuse gastric cancer are less complete. Mutations in the E-cadherin (CDH1) gene have been linked to the development of the diffuse type of gastric cancer. Several kindreds, families with hereditary diffuse gastric cancer (HDGC) have been found to carry a germline mutation in the CDH1 gene, all with diffuse-type cancer.130-132,169,170 Further evidence supporting a role for E-cadherin in the pathogenesis of gastric cancer comes from studies showing that suppression of E-cadherin expression occurs in 51% of cancers, with a higher percentage found in diffuse-type cancers.171 Furthermore, E-cadherin underexpression is associated with higher rates of lymph node metastases and reduced survival.172,173 The overall rates of CDH1 mutations in gastric cancer are low, however, with the decreased expression of E-cadherin seen in gastric cancer likely secondary to hypermethylation of the CDH1 promoter, which occurs in 50% of gastric cancers and 83% of diffuse-type gastric cancers.174 E-cadherin is a transmembrane protein that connects to the actin cytoskeleton through α- and β-catenins to establish cell polarity and mediates homophilic cellular interactions.175,176 Decreased expression of E-cadherin is believed to promote dissociation of cancer cells from their cell matrix, enhancing the migration and invasion of gastric cancer cells. Expression of α-catenin is also decreased or absent in 68% of gastric cancers.177 Therefore, E-cadherin appears to act as a tumor suppressor gene that may be important in the pathogenesis of diffuse gastric cancer.

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach Perhaps as important as the genetic alterations acquired during the progression to gastric adenocarcinoma, is in what target cells do these changes occur? In order for a cell to accumulate the quantity of genetic changes necessary for autonomous growth, it must be long lived. For these reasons, the current thinking is that a resident tissue stem cell is the target of genetic mutations and becomes the “cancer stem cell” capable of autonomous growth and with metastatic potential. Work from our laboratory offers a new model for the cancer stem cell. Bone marrow–derived stem cells are capable of homing to injured and inflamed peripheral organs and differentiating into organ-appropriate cell lineages.178-181 In an environment of inflammation and altered growth signaling, these cells can differentiate aberrantly and become dysplastic and neoplastic, and we have shown they constitute the majority of cells within in situ as well as invasive gastric adenocarcinoma lesions.182 Although much work needs to be done to understand these findings completely, they offer an exciting possibility for new approaches to understanding and treating gastric and other inflammatory mediated cancers.

PREMALIGNANT CONDITIONS CHRONIC ATROPHIC GASTRITIS

Chronic atrophic gastritis, which is defined as the loss of specialized glandular tissue in its appropriate region of the stomach, is an established early morphologic change that occurs along the sequence toward the development of gastric cancer.16,183 The presence of atrophic gastritis has an annual incidence of progression to gastric cancer of approximately 0.5% to 1%.184-187 The extent of atrophic gastritis within the stomach correlates with risk of progression to cancer.188-190 There are two forms of atrophic gastritis. The more common is multifocal atrophic gastritis (MAG), which is associated with H. pylori infection and more likely to be associated with metaplasia. The presence of H. pylori infection is associated with an approximately 10-fold increased risk of atrophic gastritis.191 There is considerable regional variation in the prevalence of atrophic gastritis in H. pylori– infected individuals, with a roughly 3-fold increase in Asian compared with Western countries.191,192 The second form of atrophic gastritis, corporal atrophic gastritis, is associated

with antiparietal cell and intrinsic factor antibodies. This form of atrophy is confined to the body and fundus. Corporal atrophic gastritis is associated with pernicious anemia and an increased gastric cancer risk, albeit not as high as that seen with H. pylori–induced multifocal atrophic gastritis, owing most likely to a lesser degree of inflammation.185,193 Mechanisms underlying the increased risk of gastric cancer in the setting of gastric atrophy may be related to low acid output (achlorhydria), which predisposes to gastric bacterial overgrowth (with non-Helicobacter organisms), greater formation of N-nitroso compounds, and diminished ascorbate secretion into the gastric lumen.194 Additionally, serum gastrin levels are increased in response to the reduced acid output. Gastrin is a known growth factor for gastric mucosal cells, and sustained elevations of serum gastrin may contribute to abnormal growth and increased risk of neoplastic progression.195,196

INTESTINAL METAPLASIA

Intestinal metaplasia (IM) can be subdivided into three categories, as classified by Jass and Filipe and as discussed in Chapter 51.197 Type I is the complete form of IM, containing Paneth cells, goblet cells that secrete sialomucins, and absorptive epithelium with well-defined brush borders. Type I metaplasia does not raise the risk of gastric cancer. Type II or incomplete metaplasia contains few absorptive cells, few columnar intermediate cells, and goblet cells that express sialomucins. Type III is intermediate between type I and type II and contains properties of both.198 It is estimated that the presence of type II or III IM is associated with a 20-fold increased risk of gastric cancer. Early gastric cancer develops in 42% of patients with type III IM within five years of follow-up, suggesting that IM represents a precursor lesion for the intestinal form of gastric cancer.199 However, whether cancer arises from areas of IM or whether IM simply represents a marker for higher gastric cancer risk remains unclear. As is the case with atrophic gastritis, the prevalence of IM in H. pylori–infected individuals is higher in Asia (≈40%) as compared with the West.191,192

DYSPLASIA

Histologic assessment of gastric dysplasia and adenocarcinoma is based on the 2000 Vienna classification (Table 54-4).200 The prevalence of gastric dysplasia ranges from as low as 0.5% in low-risk areas201 to 20% in high-risk areas.202

Table 54-4  Padova International Classification System for Gastric Dysplasia CATEGORY

DEFINITION

HISTOLOGIC DESCRIPTION

I

1.0 Normal 1.1 Reactive foveolar hyperplasia 1.2 Intestinal metaplasia

II

Indefinite for dysplasia

III

Noninvasive neoplasia

IV V

Suspicious for invasive cancer Invasive cancer

Normal gastric architecture with absent or minimal inflammatory infiltrates. The general architecture is well preserved, with evidence of hyperproliferative epithelium, enlarged nuclei, and mitotic figures. Type I. Closely resembles the morphology of the small intestine, with absorptive enterocytes, well-defined brush borders, and well-formed goblet cells. Type II. Incomplete metaplasia with irregular mucous vacuoles, absence of brush borders, and difficult-to-identify absorptive enterocytes. Cells secrete mainly sialomucins. Type III. Same as type II except cells secrete mainly sulfomucins. Inability to discern whether cells are neoplastic or non-neoplastic. Usually found in setting of inadequate biopsy specimens and presence of architectural distortion and nuclear atypia. Phenotypically neoplastic epithelium confined to glandular structures inside the basement membrane. Includes adenomas. Should be divided into “low grade” and “high grade.” Presence of neoplastic epithelium, but where invasion cannot be clearly identified. Invasive carcinoma.

Adapted from Rugge M, Correa P, Dixon M, et al. Gastric dysplasia: The Padova International Classification. Am J Surg Pathol 2000; 24:167.

895

896

Section VI  Stomach and Duodenum

Figure 54-4.  Histopathology of gastric dysplasia. Left, Low-grade dysplasia is characterized by a proliferation of neoplastic epithelial cells with nuclear pseudostratification and hyperchromasia in the absence of architectural changes. Right, High-grade dysplasia has more severe cytologic abnormalities with abnormal architectural features, including irregular fused or cribriform glands and papillae.

Prospective studies have shown that low-grade dysplasia may regress in up to 60% of cases, whereas it progresses to high-grade dysplasia in 10% to 20% of cases (Fig. 54-4).203-205 High-grade dysplasia rarely regresses, and is associated with an annual incidence of progression to gastric cancer of 2% to 6%.205-207 In a prospective cohort study from the Netherlands, the presence of high-grade dysplasia was associated with a markedly increased risk of progression to cancer (adjusted hazard ratio, 40.1).206 High-grade dysplasia is often associated with synchronous cancer and can be uni- or multifocal.208

GASTRIC POLYPS

The prevalence of gastric polyps in the general population is approximately 0.8% to 2.4%.209,210 Gastric polyps consist predominantly of fundic gland polyps (≈50%), hyperplastic polyps (≈40%), and adenomatous polyps (≈10%).210,211 The clinical course of fundic gland polyps is generally benign, and they are detected with increasing frequency in the era of proton pump inhibitor (PPI) use. In a series of 599 consecutive patients who underwent upper endoscopy, use of PPIs for 5 years or longer was associated with a significantly increased risk of fundic gland polyps (hazard ratio, 3.8).212 The rate of malignant transformation of fundic gland polyps is generally quite low (≈1%) and confined to polyps larger than 1  cm.213 One notable exception to the benign nature of fundic gland polyps is in patients with FAP. In this group the prevalence of fundic gland polyps ranges from 51% to 88%, with dysplasia present in more than 40% of cases.139,140 Hyperplastic polyps are almost always benign lesions. The rare hyperplastic polyps that undergo malignant transformation often contain areas of intestinal metaplasia or dysplasia and typically form a well-differentiated intestinaltype cancer.213 In contrast to fundic gland and hyperplastic polyps, gastric adenomas undergo malignant transformation at a high rate. Gastric adenomas that were followed by serial endoscopy with biopsy were documented to progress to dysplasia and then to carcinoma in situ, which developed within 4 years of follow-up in approximately 11% of cases.214 Endoscopic biopsy of gastric polyps to find dysplasia or carcinoma is associated with significant sampling error.215 Therefore, it is suggested that all adenomas (and

other polyps larger than 1 cm or those causing symptoms such as occult bleeding) be removed by endoscopic polypectomy. Decisions regarding surveillance intervals should be made on an individual basis.

PREVIOUS GASTRECTOMY

It has been reported by several groups that gastric surgery for benign conditions can predispose patients to a higher risk of gastric cancer, beginning 20 years after the surgery.216-219 The risk is greatest for those who underwent surgery before the age of 50 years, perhaps reflecting the long lag period necessary between the operation and the development of cancer.217 The cancers tend to occur at or near the surgical anastomosis on the gastric side; only rarely do they reside on the intestinal side of the anastomosis.220 Numerous theories have been proposed to explain the increased propensity for cancer to form at the surgical anastomosis site. They include hypochlorhydria resulting in gastric bacterial overgrowth, with increased production of nitrites, chronic enterogastric reflux of bile salts and pancreatic enzymes, which are potent gastric irritants, and atrophy of the remaining fundic mucosa secondary to low levels of antral hormones such as gastrin.17,221,222 The Billroth II operation predisposes to the development of cancer at a four-fold higher rate than a Billroth I procedure, suggesting that bile reflux may be a significant predisposing factor.217 H. pylori and associated intestinal metaplasia are found less frequently in postgastrectomy gastric cancers as compared to distal gastric cancers in the nonoperative stomach.223 It is unclear whether screening for gastric cancer in this population of patients in areas of low cancer incidence would be cost effective. With the advent of H. pylori eradication therapy as well as PPIs, the number of gastric resections for peptic ulcer disease has decreased dramatically, significantly reducing the effect of the postgastrectomy state as a risk factor for gastric cancer.

PEPTIC ULCER DISEASE

Large epidemiologic studies have demonstrated a consistently increased risk of gastric cancer in patients with a history of a gastric ulcer. In a cohort study of nearly 58,000 adults from Sweden who were followed for an average of nine years, a history of gastric ulcer was associated with a

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach significant 1.8-fold increased risk of gastric cancer.224 Interestingly, a history of duodenal ulcer was associated with a significant 40% decreased risk of gastric cancer. A separate case control study of more than 90,000 U.S. veterans found a similarly increased risk of noncardia gastric cancer (but not cardia cancer) in patients with a history of gastric ulcer, and a corresponding decreased risk in those with a history of duodenal ulcer.225 The reasons for these disparate cancer risks are not entirely clear.

been paid to the possibility of “chemoprevention” of gastric neoplastic lesions. The approach most studied has been H. pylori eradication, but consideration also has been given to supplementation with antioxidants, nonsteroidal antiinflammatory drugs (NSAIDs), and COX-2 antagonists. In this section, we first summarize the literature regarding possible beneficial effects of H. pylori eradication and then consider other strategies.

MÉNÉTRIER’S DISEASE (see Chapter 51)

The effect of eradicating H. pylori on the subsequent risk of gastric cancer is not entirely clear. There is little question that chronic inflammation in a variety of organ systems can lead to malignancy237 and that H. pylori eradication can reduce or alleviate gastric inflammation. Studies in patients have demonstrated that H. pylori eradication can lead to decreased oxidative stress and cell proliferation.238 In addition, limited studies involving eradication of gastric Helicobacter organisms in Mongolian gerbils suggest that eradication can partially reverse atrophy and metaplasia and inhibit progression to gastric cancer.239-242 Studies in mice confirm the reversibility of metaplasia and prevention of gastric cancer with early eradication. With later eradication, cancer progression was slowed and cancer mortality dramatically decreased.18 Nevertheless, with regard to published trials in humans, conclusive evidence that treatment of H. pylori infection prevents gastric cancer is lacking, in part because of the rare endpoint—gastric cancer—needed for these studies. One approach has been to examine intermediate biomarkers such as gastric atrophy and intestinal metaplasia, which are generally considered premalignant lesions. Thus, a number of studies have looked at the effect of H. pylori eradication on these intermediate biomarkers, and a majority have shown a beneficial effect in preventing progression of gastric disease.243-247 In a randomized placebo-controlled trial of 587 Chinese patients with H. pylori infection, assignment to eradication therapy was associated with a significantly reduced risk of progression of intestinal metaplasia (odds ratio, 0.63).248 However, a randomized placebo-controlled trial of Mexican adults did not demonstrate any benefit of H. pylori eradication in the prevention of histologic progression.247 In an open-label randomized controlled trial of patients with resected early gastric cancer, patients were assigned to receive either H. pylori eradication therapy or placebo. After three years of follow-up, H. pylori eradication was associated with a reduced odds of development of metachronous gastric cancer (OR 0.35; 95% CI: 0.16 to 0.78).249 In a retrospective analysis of a cohort of Japanese patients who were treated for H. pylori, successful eradication (as compared to persistent infection) was associated with a significant 80% reduced risk of gastric cancer.250 A prospective, randomized placebo-controlled trial sought to determine whether H. pylori eradication in a highrisk population in China would reduce the incidence of gastric cancer.246 Although no overall benefit was seen in the group receiving H. pylori eradication, there was a clear reduction in gastric cancer incidence in a post hoc subgroup analysis of H. pylori carriers who did not have precancerous lesions (gastric atrophy, intestinal metaplasia, or dysplasia) at study initiation. It is possible that some of the patients in the eradication arm may have passed a “point of no return,” when cellular alterations had sufficiently accumulated to promote cancer.251 Further randomized trials are needed, but the evidence to date supports the notion that early eradication of H. pylori may prevent or delay progression to gastric cancer in high-risk patients.

In a review of case reports, 15% of patients with Ménétrier’s disease had associated gastric cancer,226 including several cases that documented a progression from dysplasia to cancer.227,228 Because of the rarity of Ménétrier’s disease, it has been difficult to study its relationship with gastric cancer in any controlled fashion, and no recommendations regarding endoscopic surveillance can be made.

SCREENING AND PREVENTION SCREENING AND SURVEILLANCE

The majority of the literature regarding screening for gastric cancer comes from east Asia, where the prevalence of this disease is among the highest in the world.229 Since 1960, the Japanese have been performing mass screening, using upper GI barium studies followed by endoscopy if any suspicious lesions are found. Japanese researchers have reported a sensitivity of 66% to 90% and a specificity of 77% to 90% for this screening approach.230 Not surprisingly, studies from Japan have also shown that use of screening leads to a diagnosis of gastric cancer at earlier stages, with one study reporting that more than 50% of cases detected by screening were diagnosed as stage I.231 (Staging of gastric cancer is discussed later.) Long-term follow-up data from the Japanese Public Health Center cohort of more than 42,000 adults showed that subjects who underwent screening had a nearly 50% reduced risk of death from gastric cancer.232 A separate cohort study of 87,000 adults from Japan found a 25% to 35% risk reduction in death from gastric cancer among those who participated in gastric cancer screening.233 However, similar risk reductions were seen for death from all causes, casting a level of uncertainty on the true magnitude of benefit associated with screening with respect to preventing death from gastric cancer. The serum pepsinogen (PG) test is increasingly used to screen for patients at highest risk for having preneoplastic gastric lesions.234 As discussed in Chapter 49, the stomach produces two types of pepsinogen: PGI and PGII. In the presence of atrophic gastritis, production of PGI from oxyntic glands is reduced, whereas PGII production remains relatively constant. Therefore, low serum pepsinogen I levels (<70 mg/L) and a serum pepsinogen I/II ratio less than 3 are useful for the identification of patients with atrophic gastritis.229 A prospective cohort study found that patients with low serum PGI concentrations had a significantly elevated risk of gastric cancer.235 Screening for gastric cancer with upper endoscopy is likely cost-effective in moderate- to high-risk populations, such as older Asian men.236 However, in populations with a lower incidence of gastric cancer, screening is less likely to have the same degree of benefit.

PREVENTION

Given the lethal nature of gastric cancer and its link to chronic infection and inflammation, much attention has

Eradication of Helicobacter pylori

897

898

Section VI  Stomach and Duodenum In Western countries, gastric cancer prevention has not been extensively pursued due to the low prevalence of H. pylori and decreasing incidence of gastric cancer. However, a study by Parsonnet and colleagues252 suggested that screening and treatment of H. pylori infection would be cost-effective in the prevention of gastric cancer, parti­ cularly in high-risk populations, if it was assumed that treatment of H. pylori infection prevented 30% of attributable gastric cancers. Using a more conservative 10% reduction in gastric cancer risk, an analysis from the United Kingdom also concluded that H. pylori eradication was cost-effective.253

Antioxidants

Chronic inflammatory states such as H. pylori gastritis can result in the generation of free radicals derived from oxygen and nitrogen.254 These free radicals can promote carcinogenesis via numerous mechanisms, including direct DNA damage and inhibition of DNA repair mechanisms, inhibition of apoptosis, and activation of cellular proliferation pathways. Antioxidants, such as carotenoids and vitamins C and E, bind with reactive oxygen and nitrogen species to neutralize their damaging effects. Epidemiologic data support a relationship between increased antioxidant intake and reduced risk of gastric cancer.255-259 In a nested case-control study from Japan, low plasma β-carotene levels were associated with an increased risk of gastric cancer.257 A case-control study from Korea found that elevated nitrate-to-antioxidant intake ratios were associated with an increased risk of gastric cancer.258 In a Swedish cohort study, high levels of vitamin A, retinol, and α- and β-carotene intake were associated with a 50% risk reduction in gastric cancer.259 However, in a randomized placebo-controlled trial of antioxidants (either vitamin A, C, or E) in patients with precancerous gastric lesions (nonatrophic or atrophic gastritis, intestinal metaplasia, or dysplasia), antioxidant supplementation did not result in either reduced histologic progression or increased histologic regression.260 A randomized controlled trial in China also found no effect of combined vitamin C, E, and selenium supplementation on the prevalence of a combined endpoint of atrophic gastritis, intestinal metaplasia, dysplasia, or cancer.261 Based on these results as well as those of the Beta Carotene and Retinol Efficacy Trial, in which subjects who received β-carotene and vitamin A had an increased risk of lung cancer,262 antioxidant supplementation for the prevention of gastric cancer cannot yet be recommended.

Aspirin and Nonsteroidal Anti-inflammatory Drugs

Among other effects, aspirin and other NSAIDs inhibit cyclooxygenases. COX-1 is constitutively expressed in the GI tract. COX-2 expression is generally not observed in normal GI mucosa, but has been described in multiple epithelial malignancies, including gastric cancer.263,264 COX-2 expression is associated with aggressive cell growth in human as well as mouse models of cancer265-268 and has been found to be overexpressed in 70% of gastric cancers.269 In this setting, COX-2 could promote the growth of tumors, inhibit apoptosis, and increase angiogenesis. COX-2 expression has been reported to be elevated in preneoplastic lesions, including intestinal metaplasia and dysplasia, and COX-2 levels appear to diminish after H. pylori eradication.270 Multiple epidemiologic studies have demonstrated a consistent association between NSAID use and a reduced risk of gastric cancer.271-274 In a case-control study from Los Angeles County, NSAID use for more than 5 years was

associated with a significantly decreased risk of noncardia gastric cancer (OR, 0.61), and there was a significant doserelated effect.272 A nested case-control study using the General Practitioners Research Database found that longterm users of non-aspirin NSAIDs had a significantly reduced risk of gastric cancer (OR, 0.65), although there was no observed association between aspirin use and gastric cancer.274 A meta-analysis reported a significant association between any NSAID use and reduced risk of gastric cancer (OR, 0.78), with similar findings for both acetylsalicylic acid (ASA) and non-ASA NSAIDs.273 In a randomized controlled trial of H. pylori–negative patients with intestinal metaplasia, patients receiving the COX-2 selective inhibitor rofecoxib or placebo had no difference in the rate of regression of intestinal metaplasia after two years.275 This trial was limited by the relatively short follow-up period and use of premalignant endpoints. Further trials in high-risk patients are warranted to determine if NSAIDs are effective for gastric cancer prevention; at the present time they cannot be recommended due to their unproven efficacy and known side effects.

Green Tea

Green tea is widely consumed in Asian countries and is hypothesized to have protective effects against cancer of the upper digestive tract. Polyphenols and other metabolites present in green teas, such as epigallocatechin-3-gallate (EGCG) and other catechins, have a variety of antitumor effects, including induction of apoptosis, inhibition of tumor cell growth and proliferation, and reduction in COX-2 expression.276-278 EGCG also has antioxidant properties and may have anti-inflammatory properties as well.279,280 The majority of case-control studies have shown an inverse association between the risk of gastric cancer and the consumption of green tea.281-283 However, a populationbased prospective cohort study in northern Japan found no association between green tea consumption and the risk of gastric cancer.284 A separate cohort study from Japan reported a reduced risk of gastric cancer in women with high green tea consumption, but no change in risk among men.285 Thus, prospective controlled trials are needed, and at present, green tea cannot be recommended as chemoprevention for gastric cancer.

CLINICAL FEATURES Early gastric cancers are asymptomatic in up to 80% of cases. When symptoms occur they tend to mimic those of peptic ulcer disease. With advanced gastric cancer, the common symptoms are weight loss (≈60% of patients) and abdominal pain (≈50%).10 Other presenting symptoms include nausea, vomiting, anorexia, dysphagia, melena, and early satiety. Pyloric outlet obstruction can occur with tumors of the antrum and pylorus, and tumors of the cardia can cause dysphagia due to involvement of the lower esophageal sphincter and development of pseudoachalasia.286 Rarely, paraneoplastic syndromes occur, such as thrombophlebitis (Trousseau’s sign), neuropathies, nephrotic syndrome, and disseminated intravascular coagulation.287-289 Dermatologic paraneoplastic syndromes may occur uncommonly (see Chapter 22) and include hyperpigmented patches in the axilla (acanthosis nigricans) and the sudden onset of seborrheic dermatosis (senile warts) and pruritus (sign of Leser-Trélat).290 The physical examination is usually unremarkable. Cachexia and signs of bowel obstruction are the most

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach common abnormal findings. Occasionally it is possible to detect an epigastric mass, hepatomegaly, ascites, and lower extremity edema.291 Laboratory studies are generally unrevealing until the cancer reaches advanced stages. Anemia and a positive test result for fecal occult blood may occur from chronic bleeding of an ulcerated mass. Hypoproteinemia can occur in patients with weight loss. Liver enzyme values, particularly alkaline phosphatase, can be elevated secondary to hepatic metastases. Gastric cancer is metastatic at the time of diagnosis in 33% of cases.292 The most common sites of metastasis are the liver (40%) and peritoneum.293 Other sites of spread include the periumbilical area (Sister Mary Joseph’s nodule), left supraclavicular sentinel nodes (Virchow’s node), the pouch of Douglas (rectal shelf of Blumer), and the ovaries (Krukenberg’s tumor). Gastric cancer has also been reported to metastasize to the kidney, bladder, brain, bone, heart, thyroid, adrenal glands, and skin.291 There are reports of unusual presentations of metastatic disease, such as shoulder-hand syndrome from bone metastasis, diplopia and blindness from orbit and retinal metastases, and virilization due to Krukenberg’s tumors.294-297

for the detection of gastric cancer in patients with occult bleeding and a normal colonoscopy will vary based on the patient’s baseline risk of gastric cancer. Extensive involvement by the diffuse type of gastric cancer can manifest as a rigid and thickened stomach, also known as linitis plastica. In Japan and other areas of high gastric cancer prevalence, chromoendoscopy, magnification endoscopy, and narrow band imaging are used alone or in combination as aids in the detection of early gastric cancer (see Fig. 54-5B). Distinct irregular mucosal surface and vascular patterns have been found to correlate with the presence of dysplasia and carcinoma.300 There are also ongoing investigations into the utility of newer techniques such as autofluorescence and confocal microendoscopy for the diagnosis of early gastric neoplasia.301,302 A classification system has been developed for early gastric cancer based on endoscopic appearance,303 the purpose of which is to assess early lesions for risk of submucosal invasion as well as risk of lymph node spread (Fig. 54-6). The three types include superficial polypoid (type

DIAGNOSIS ENDOSCOPY

Esophagogastroduodenoscopy (EGD) is currently the pro­ cedure of choice for the diagnosis of gastric cancer (Fig. 54-5A). When a nonhealing gastric ulcer is found, at least six to eight biopsy specimens from the edge and base of the ulcer are recommended.298 As discussed in Chapter 13, the American Gastroenterological Association has recommended that an upper endoscopy be performed in patients who are older than 55 years with new-onset dyspepsia and in patients younger than 55 years who have “alarm” symptoms (weight loss, recurrent vomiting, dysphagia, evidence of bleeding, anemia).299 Dyspeptic patients in whom an empirical trial of proton pump inhibitors and eradication of H. pylori do not relieve symptoms should undergo prompt endoscopic evaluation as well. The basis for these recommendations is the low incidence of gastric cancer in individuals younger than 55 years. The yield of upper endoscopy

A

0–Ip

0–Is

Protruded, pedunculated

0–IIa

Protruded, sessile

0–IIb

Superficial, elevated

0–IIc Flat

Superficial shallow, depressed

0–III Excavated Figure 54-6.  Schematic representation of the major variants of type 0 neoplastic lesions of the stomach: polypoid (Ip and Is), nonpolypoid (IIa, IIb, and IIc), nonpolypoid and excavated (III). (From The Paris endoscopic classification of superficial neoplastic lesion: Esophagus, stomach, and colon: November 30 to December 1, 2002. Gastrointest Endosc 2003; 58:S3-43.)

B

Figure 54-5.  Endoscopic examples of gastric cancer. A, Ulcerated gastric adenocarcinoma mass lesion. B, Chromoendoscopic view of a superficial depressed gastric cancer, highlighted with indigo carmine (arrow). (From Toyoda H, Tanaka K, Hamada Y, et al. Endoscopic diagnosis of hypopharyngeal, esophageal and gastric neoplasm. Dig Endosc 2006; 18:S41-3.)

899

900

Section VI  Stomach and Duodenum 0-I), superficial flat/depressed (types 0-IIa to 0-IIc), and superficial excavated (type 0-III) lesions. The most commonly observed subtype is 0-IIc, the nonpolypoid depressed lesion.303 This classification system is used most often in Japan, where endoscopic mucosal resection and submu­ cosal dissection are frequently performed for early gastric neoplasia.

UPPER GASTROINTESTINAL SERIES

Upper GI series has been largely replaced by upper endoscopy as the initial test of choice for the diagnosis of gastric cancer. Barium studies have been reported to have 60% to 70% sensitivity and 90% specificity for the detection of advanced gastric cancer.304

COMPUTED TOMOGRAPHY GASTROGRAPHY

Computed tomography (CT) colonography has gained significant attention for its potential role as a screening modality for colon polyps and colon cancer (see Chapter 123). CT gastrography has also been studied for the diagnosis of early gastric cancer. In a study of 39 patients from South Korea with early gastric cancer, CT gastrography had a sensitivity of 73% to 76% and good interobserver reliability (R = 0.84).305 Only small studies have been performed thus far using this imaging modality, and CT gastrography cannot yet be recommended for screening outside of the research setting.

SERUM MARKERS

To date no reliable serum marker has been identified with sufficient sensitivity and specificity for the diagnosis of gastric cancer. Low serum pepsinogen I levels, low ratios of serum pepsinogen I to pepsinogen II, and hypergastrinemia have been reported in patients with atrophic gastritis and intestinal metaplasia, but the utility for the detection of gastric cancer has been mixed.306,307 In a screening study of 17,000 Japanese men, a positive pepsinogen test (defined as a serum pepsinogen I <50 µg/L and a pepsinogen I/II <3), in combination with upper GI series, identified gastric cancer in 0.28% of subjects (≈1 in 350), and 88% of these were early gastric cancers.308 Additionally, 89% of the cancers identified by the pepsinogen test alone were early gastric cancers. The major limitation of this test is the low specificity for the diagnosis of gastric cancer.309

Tis

Early gastric cancer T1 Advanced gastric cancer

T2

T1

T1

T2

Serum carcinoembryonic antigen (CEA) and carbohydrate antigen (CA) 19-9 have been extensively studied for the diagnosis of gastric cancer. The sensitivities of these markers is quite low for early gastric cancer,310 and elevated levels are levels also seen in other epithelial malignancies. These tumor markers are frequently elevated in recurrent gastric cancer, especially in patients who had elevated levels prior to surgical resection.311 More recent studies have identified, among others, transforming growth factor-β1, CA 72-4, tumor M2-pyruvate kinase, hepatocyte growth factor, and others as potential markers for the diagnosis of gastric cancer.312-315 However, larger studies are required to determine their clinical utility.

CLASSIFICATION AND STAGING Several classification systems further define gastric cancer and predict prognosis. As mentioned (see Fig. 54-2), gastric cancers can be subdivided into intestinal and diffuse types. Gastric cancer can also be divided into early and advanced lesions. Early gastric cancer is defined as a cancer that does not invade beyond the submucosa regardless of lymph node involvement. Early gastric cancer has a much higher prevalence in the Far East, especially Japan, and carries a very favorable prognosis, with five-year survival rates greater than 90% being reported in Asia and greater than 80% in Western countries.316-319 The most commonly used clinical staging classification system for gastric cancer is the TNM system, used by the International Union Against Cancer (UICC) and the American Joint Committee on Cancer (AJCC).320,321 In the TNM staging system, T (tumor) indicates the depth of penetration (Fig. 54-7): T1 denotes a tumor confined to the mucosa or submucosa, T2 denotes involvement of the muscularis propria, T3 denotes invasion through the serosa, and T4 denotes invasion through the serosa and into adjacent organs or structures. N (nodes) indicates the amount of lymph node invasion: N0 denotes no lymph node involvement, N1 denotes involvement of 1 to 6 regional lymph nodes, N2 denotes involvement of 7 to 15 regional lymph nodes, and N3 denotes involvement of more than 15 regional lymph nodes. M (metastasis) indicates the presence of

T3

T4

EUS appearance Mucosa (superficial)

Hyperechoic

Mucosa (deep)

Hypoechoic

Submucosa

Hyperechoic

Muscularis propria

Hypoechoic

Serosa

Hyperechoic

T3 Adjacent organ

T4 Figure 54-7.  Classification of gastric adenocarcinoma by depth of invasion (T classification). In the TNM classification, T denotes depth of invasion: Tis designates carcinoma in situ; T1 tumors are confined to the mucosa and submucosa; T2 tumors penetrate the muscularis propria but not the serosa; T3 tumors penetrate the serosa without involving contiguous structures; and T4 tumors penetrate the serosa and involve adjacent organs and tissues. In early gastric cancer, the disease is confined to the mucosa and submucosa (T1), regardless of nodal involvement. EUS, endoscopic ultrasonography.

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach metastases, with M0 denoting no metastases and M1 denoting distant metastases (Table 54-5).

METHODS OF STAGING

Accurate staging in gastric cancer is important for treatment decisions. Endoscopic ultrasound (EUS) is the best-studied modality for the staging of gastric cancer and remains the test of choice. However, improvements in image quality for both CT and magnetic resonance imaging (MRI) make these studies potential alternatives to EUS.

Endoscopic Ultrasonography

EUS allows the visualization of the five layers of the gastric wall. The superficial gastric mucosa is represented by an echogenic first layer, and the deeper mucosa by a hypoechogenic second layer; the submucosa is represented by an echogenic third layer, the muscularis propria as a hypoechogenic fourth layer, and the serosa as an echogenic fifth layer.322 Studies of the ability of EUS to determine T stage have reported an accuracy rate of 67% to 92%, with accuracy of determination of serosal involvement ranging from 78% to 100%.323 EUS is particularly useful for staging early gastric cancer, which can be potentially resected endoscopically (Fig. 54-8). In a study of 295 patients with early gastric cancer, high-frequency EUS was found to have 90% accuracy for differentiating between mucosal and submucosal tumor invasion.324 In terms of N staging, the rate of detection of perigastric lymph nodes with EUS is comparable to staging with CT, with a diagnostic accuracy ranging from 50% to 80%.325-329 Table 54-5  Clinical Staging of Gastric Cancer Based on the TNM Classification

Tis T1 T2 T3 T4

N0

N1

N2

N3

M1 (ANY N)

0 IA IB II IIIA

— IB II IIIA IV

— II IIIA IIIB IV

— IV IV IV IV

— IV IV IV IV

is, in situ; M, metastases; N, node involvement; T, tumor. From references 320 and 321. See text for further details.

A

A particular difficulty with N staging lies in the fact that many small lymph nodes can also harbor metastases, and thus understaging can occur. One study of 1253 lymph nodes in 31 patients with gastric cancer found that 55% of lymph nodes containing tumor were smaller than 5mm.330 EUS also has the ability to identify and biopsy submucosal lesions, such as gastric lymphomas and stromal tumors (discussed later and in Chapters 29 and 31). These lesions typically involve thickening of the submucosa and muscularis propria and may appear as gastric fold thickening on barium studies or endoscopy.

Computed Tomography

Advances in imaging technology have greatly improved the ability of CT to stage gastric tumors. Although not as extensively studied as EUS, multidetector row CT (MDCT), by which the wall of the stomach can be seen as three layers (an inner layer corresponding to the mucosa, an intermediate layer corresponding to the submucosa, and an outer layer of slightly higher attenuation corresponding to the muscularis propria and serosa), appears to have comparable accuracy to EUS in terms of T and N staging. The loss of fat planes between the gastric mass and an adjacent organ suggests tumor invasion. The accuracy of MDCT for overall T staging ranges from 77% to 89%, and discriminates serosal involvement with an accuracy of 83% to 100%.323 Accuracy with respect to N staging may be as high as 89% with MDCT.331,332 As with all other imaging modalities, CT has difficulty discerning lymph node metastases smaller than 5 mm. At present, the role of CT is mainly for the detection of distant metastases and as a complement to EUS for assessing regional lymph node involvement. It is not yet clear whether EUS or MDCT (or the combination) is superior for T and N staging in gastric cancer, and the underlying technology continues to evolve and improve.

Magnetic Resonance Imaging

MRI with gadolinium has also been used for gastric cancer staging. It is similar to CT in its advantages (ability to find distant metastases) and weaknesses (need for adequate gastric distention). The accuracy of MRI ranges from 90% to 93% for T staging and from 91% to 100% for N staging.323 However, given the small number of studies, MRI cannot

B

Figure 54-8.  Staging of gastric cancer using EUS. A, Endoscopic image showing a 25-mm protruding mass on the posterior wall of the antrum. Biopsy of the mass revealed an adenocarcinoma. B, EUS image of the lesion, showing the hypoechoic mucosal mass (arrow) with an intact submucosal layer, indicating that the lesion is an early (T1) gastric cancer. EUS, endoscopic ultrasonography. (From Kim JH, Song KS, Youn YH, et al. Clinicopathologic factors influence accurate endosonographic assessment for early gastric cancer. Gastrointest Endosc 2007; 66:901-8.)

901

902

Section VI  Stomach and Duodenum yet be advocated as the test of choice for staging gastric cancer.

Positron Emission Tomography

Positron emission tomography (PET) is not very useful for staging in gastric cancer, largely due to the fact that most gastric adenocarcinomas have low fluorodeoxyglucose uptake.333

Restaging after Neoadjuvant Treatment

The accuracy of restaging gastric cancer after neoadjuvant chemotherapy decreases considerably. EUS has less than 50% accuracy for T and N restaging, and similarly disappointing results have been reported for post-treatment staging with CT.334 However, the use of preoperative staging to assess response to chemotherapy may correlate well with both overall and disease-free survival.335

PROGNOSIS AND TREATMENT Overall, the five-year survival rate in the United States from gastric cancer is 24% compared with 64% for colon cancer.292 The TNM classification is used to stratify disease into four clinical stages (I through IV) to predict prognosis in patients treated with gastrectomy (Table 54-6).336 The survival data from Japanese studies are generally superior to those seen in Western countries, perhaps because of the preference in Japan for extended lymphadenectomy or because of less understaging than is found in Western countries.336

SURGICAL THERAPY

Surgical resection remains the primary curative treatment for gastric cancer. In addition, surgical resection often provides the most effective palliation of symptoms, particularly those of obstruction. In some cases, surgery is required for diagnosis, as in cases of nonhealing gastric ulcers with negative biopsy results or in cases with persistent pyloric outlet obstruction suggesting an antral carcinoma. Surgery should be attempted in most cases of gastric cancer. However, in the presence of extensive involvement of diffuse-type cancer (or linitis plastica), bulky metastatic disease, retroperitoneal invasion, or peritoneal carcinomatosis, or if the patient has severe comorbid illnesses, the prognosis may be sufficiently poor to make the value of resection questionable. Surgery, and laparoscopy in particular, can be useful in the staging of cancer. Laparoscopy can help identify primary tumor resectability, peritoneal deposits, and appropriate

Table 54-6  Five-Year Survival Rates (%) for Patients with Gastric Cancer Based on Clinical Staging STAGE IA IB II IIIA IIIB IV

UNITED STATES

JAPAN

GERMANY

78 58 34 20 8 7

95 75 46 48 18 5

86 72 47 34 25 16

From Hundahl S, Philips J, Menck H. The National Cancer Data Base Report on poor survival of U.S. gastric carcinoma patients treated with gastrectomy: 5th ed. American Joint Committee on Cancer staging, proximal disease, and the “different disease” hypothesis. Cancer 200; 88:921-32.

candidates for neoadjuvant therapy. Laparoscopic peritoneal lavage has been used to detect occult intraperitoneal free cancer cells. A positive peritoneal lavage correlates significantly with eventual development of overt peritoneal metastases.337 In general, total gastrectomy is performed for proximal gastric tumors and diffuse gastric cancer, and partial gastrectomy is reserved for tumors in the distal stomach. Large, randomized multicenter trials performed in France and Italy compared subtotal with total gastrectomy for adenocarcinoma of the antrum and found no differences in five-year survival rates or operative mortality.338,339 Some centers have argued for performing a complete splenectomy with gastrectomy. However, several retrospective and prospective studies found that concurrent splenectomy increased morbidity and had either no effect on or worsened survival.340,341 The extent of lymphadenectomy accompanying the gastrectomy has been a subject of debate for many years. The Japanese advocate a more extensive lymph node dissection (D2 resection) than their Western counterparts (D1 resection) and have higher published survival rates. A D2 resection entails resection of the nodes of the celiac axis and the hepatoduodenal ligament in addition to the perigastric lymph nodes taken in a D1 procedure. The differences in reported survival rates may reflect the fact that the Japanese have a much higher incidence of early gastric cancer, and the more extensive lymph node dissection performed in Japan may find more positive lymph nodes, making survival rates of Japanese patients with N0 staging appear to be higher than those of their potentially understaged Western counterparts. A large prospective multicenter Dutch study of more than 1000 patients reported no significant difference in five-year survival, with higher rates of postoperative death and complications for D2 lymphadenectomy than for the more conservative D1 lymphadenectomy.342 A British prospective study of 400 patients likewise showed no benefit from more extensive surgery; five-year survival rates were 35% for D1 resection and 33% for D2 resection.340 At present, data are insufficient to support extended lymph node resection in centers outside Japan. To prevent understaging, the current recommendation is a minimum D1 lymphadenectomy with removal of at least 15 nodes.343

ENDOSCOPIC MUCOSAL RESECTION AND SUBMUCOSAL DISSECTION

Advances in endoscopic techniques have permitted endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) to be used as curative therapies for select early gastric cancers (EGCs). This technique has been used widely for intestinal-type cancers in Japan, where studies have shown that only 3.5% of patients with EGCs smaller than 2 to 3 cm have lymph node involvement, making these lesions amenable to local therapy. Lesions larger than 4.5 cm have a greater than 50% chance of spread into the submucosa, are associated with “positive” nodes, and are therefore less likely to be endoscopically resectable.344 The following criteria have been suggested for choosing to perform EMR in gastric cancer: (1) the cancer is located in the mucosa and the lymph nodes are not involved, as indicated by EUS; (2) the maximum size of the tumor is less than 2 cm when the lesion is slightly elevated (type IIa) and less than 1 cm when the tumor is flat or slightly depressed (type IIb or IIc, respectively) without an ulcer scar (see Fig. 54-6); (3) there is no evidence of multiple gastric cancers

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach

A

B

C

D

E

F

Figure 54-9.  Endoscopic submucosal dissection for early gastric cancer (EGC). A, Conventional endoscopic view showing an EGC type 0-Ip at the angularis. B, Endoscopic view after indigo carmine spraying to enhance the lesion margin. C, Normal tissue surrounding the lesion marked with a needle knife. D, A circumferential cut around the marking dots made with an insulated-tip electrosurgical knife. E, Base after submucosal dissection of lesion, performed en bloc. F, The specimen stretched and pinned on a wood plate before immersion in formalin. (From Lee IL, Wu CS, Tung SY, et al. Endoscopic submucosal dissection for early gastric cancers: Experience from a new endoscopic center in Taiwan. J Clin Gastroenterol 2008; 42:42-7.)

or simultaneous abdominal cancers; and (4) the cancer is of the intestinal type.345 It is generally not possible to remove lesions larger than 1.5 to 2 cm en bloc by EMR. Endoscopic submucosal dissection is a technique developed in Japan and permits en bloc resection of larger early gastric cancers. With ESD, submucosal saline injection is performed, followed by the use of endoscopic electrosurgical knives to resect the entire tumor (Fig. 54-9).346 In addition to an R0 resection, ESD allows for more precise histopathologic assessment of depth of invasion and lymphovascular involvement, and permits appropriate assessment for risk of lymph node metastasis. If the preprocedure evaluation does not reveal regional lymph node involvement, much larger superficial lesions can be resected. Because experience with this technique has increased, en bloc resection rates are now reported to be greater than 90%, with local recurrence rates less than 3%.346 Due to the large size of some of the lesions being resected, the risk of perforation is relatively high (2% to 6%).347,348 However, perforations recognized early can generally be treated with closure using endoscopic clipping.346 No randomized clinical trials have yet been performed comparing surgery to endoscopic resection for early gastric cancer.

nosis.322 In such patients who have undergone potentially curative resection, recurrence rates are high. Unfortunately, gastric cancer appears to be fairly resistant to conventional chemotherapy. Numerous clinical trials have been performed evaluating the role of adjuvant chemotherapy following attempted curative resection for gastric cancer.349 The majority of the studies were inconclusive, but a series of meta-analyses of these trials suggested a 15% to 20% reduction in the risk of death in patients who received adjuvant chemotherapy.350,351 More recent randomized trials of cisplatin- or epirubicin-based adjuvant chemotherapy have largely failed to show a benefit.352,353 Neoadjuvant chemotherapy, on the other hand, does appear to benefit patients with resectable disease. In the United Kingdom MAGIC trial, 503 patients with gastric, gastroesophageal junction, or distal esophageal cancer were randomly assigned to undergo surgery alone or surgery with neoadjuvant epirubicin, cisplatin, and 5-fluorouracil (5-FU). Compared with surgery alone, the neoadjuvant group had significantly improved five-year survival (36% vs. 23%), progression-free survival (hazard ratio, 0.66), and overall survival (hazard ratio, 0.75).354 As a result, preoperative chemotherapy is now considered an acceptable treatment option for gastric cancer.

CHEMOTHERAPY

Combined chemoradiation after surgical resection appears to be effective at improving progression-free and overall survival in gastric cancer. The Intergroup Trial 0116 randomized 603 patients with gastric or gastroesophageal junc-

In Western countries, approximately 75% of patients with gastric cancer have disease that has spread to the perigastric lymph nodes or have distant metastases at the time of diag-

CHEMORADIATION

903

904

Section VI  Stomach and Duodenum tion cancer to undergo surgery alone or surgery followed by 5-FU, leucovorin, and radiation therapy. Subjects in the surgery-alone group had shorter median survival time (27 months vs. 36 months) and worse overall- and relapse-free survivals (hazard ratios, 1.35 and 1.52, respectively).355 Following publication of the results of this study, adjuvant chemoradiation became the standard of care in the United States, although the optimal chemotherapy regimen is not yet clear. Early studies of the use of neoadjuvant chemoradiation have also shown promising results.356

INTRAPERITONEAL CHEMOTHERAPY

Because systemic chemotherapy is ineffective for peritoneal metastasis, intraperitoneal (IP) chemotherapy can be considered in patients whose tumors are resected for cure but have a high likelihood of microscopic residual disease. In a randomized trial of 248 patients with gastric cancer, postoperative hyperthermic IP chemotherapy was associated with improved overall survival compared with surgery alone.357 The treatment benefits were most pronounced in patients with stage III and IV disease, serosal invasion, and lymph node metastases. Although a second clinical trial reported similar results,358 other studies have failed to demonstrate a benefit of hyperthermic IP chemotherapy.359,360 A meta-analysis of studies of IP chemotherapy for patients with resectable gastric cancer reported a significantly reduced risk of death in patients who received hyperthermic IP chemotherapy (hazard ratio, 0.6).361 At present, the use of hyperthermic IP chemotherapy should be confined to patients enrolled in clinical trials, especially in Western countries.

UNRESECTABLE DISEASE

Unfortunately, up to one third of patients with gastric cancer will have unresectable disease at the time of diagnosis.292 Chemotherapy for locally advanced gastric cancer without distant metastases can result in shrinking of the tumor to the point at which successful curative resection is possible.362,363 Even when curative surgery is not possible, chemotherapy has been shown both to improve survival as well as quality of life compared to best supportive care in this group of patients.364 Several randomized clinical trials have demonstrated efficacy of multidrug cisplatin-based regimens.365,366 A newer clinical trial using the EOX regimen (epirubicin, oxaliplatin, and capecitabine) was found to be noninferior to cisplatin-based regimens, and had a median survival of 11.2 months.367 The benefit of the EOX regimen is the substitution of oxaliplatin and capecitabine for cisplatin and 5-FU, respectively, resulting in greater convenience, ease of administration, and potentially fewer side effects. Patients with advanced gastric cancer of the distal antrum or pylorus are at risk for developing gastric outlet obstruction. Traditionally, surgical gastrojejunostomy was performed for relief of symptoms and to allow continued enteral nutrition. With the advent of endoscopic stents, duodenal stenting across the obstructing tumor has emerged as a nonsurgical alternative for palliation. The results of a literature review of studies evaluating gastrojejunostomy and stenting found no differences between rate of technical success (96% to 100%), early and late complications, and persistent symptoms.368 Recurrent obstructive symptoms were more common with stenting. Both gastrojejunostomy and endoscopic stenting are acceptable options for the relief of malignant gastric outlet obstruction. The decision should be based on the individual clinical scenario as well as the availability of appropriate surgical or endoscopic expertise.

GASTRIC LYMPHOMA Gastric lymphomas account for approximately 3% of all gastric malignancies, and can be subdivided into those in which the stomach is the primary site of involvement and those with disseminated nodal disease and secondary involvement of the stomach. More than 95% of gastric lymphomas are non-Hodgkin’s lymphomas. Gastric lymphoma is the most common form of extranodal non-Hodgkin’s lymphoma, accounting for more than 30% of all cases.369,370 The clinical manifestations of gastric lymphoma are similar to those of gastric adenocarcinoma. Patients tend to be asymptomatic in the early stages. With more advanced disease, patients may present with nonspecific complaints including abdominal pain, nausea, vomiting, anorexia, weight loss, bleeding, fever, night sweats, swelling, and diarrhea.371 On endoscopy, gastric lymphomas manifest in a variety of ways and can appear as fungating lesions, polypoid masses, thickened gastric folds, or ulcerating lesions. Because of frequent involvement of the submucosa, deep biopsy specimens, obtained by snare technique, mucosal resection, or use of jumbo forceps, are often needed for a pathologic diagnosis. Endoscopic ultrasonography is essential to document extent of disease and may be more useful than CT in the evaluation of perigastric lymph node involvement. Other imaging modalities, such as CT and MRI, may be of value in determining involvement of the liver and spleen as well as of distant lymph nodes. Gastric lymphoma is discussed in detail in Chapter 29.

GASTRIC CARCINOID TUMORS Gastric carcinoid tumors (Fig. 54-10) account for 7% of all GI carcinoids and 0.2% of all gastric neoplasms.372 In the stomach, the well-differentiated tumors are mainly of enterochromaffin-like (ECL) cell origin, with a small minority being of other endocrine cell types.373 Although gastric carcinoids often contain neuroendocrine peptides, carcinoid syndrome does not occur unless there is hepatic involvement. The large majority of gastric carcinoid tumors express somatostatin receptors, and somatostatin receptor scintigraphy is a useful test to evaluate extent of and spread of disease. Classification, staging, prognosis, and therapy of gastric carcinoids are discussed in detail in Chapter 31.

GASTROINTESTINAL STROMAL TUMORS Gastrointestinal stromal tumors (GISTs) are intra-abdominal nonepithelial (mesenchymal) tumors that can develop throughout the GI tract. Of all GISTs in the GI tract, 50% to 60% occur in the stomach (Fig. 54-11).374 They are discussed in detail in Chapter 30.

MISCELLANEOUS TUMORS Metastatic disease to the stomach can occur with melanoma and with primary tumors of the breast, lung, ovary, liver, colon, and testis, with breast cancer being the most common.375 Other rare malignant tumors that can involve the stomach are Kaposi’s sarcoma (see Chapter 33), myenteric schwannoma, glomus tumor, small cell carcinoma, and

Chapter 54  Adenocarcinoma and Other Tumors of the Stomach

A

B

C

Figure 54-10.  Pathologic specimen of a gastric carcinoid tumor. A, Large, ulcerated tumor (gross). B, Fixed specimen (gross). C, Low-power histologic section showing a large tumor ulcerated but confined to the wall of the stomach. (Hematoxylin and eosin.) (Courtesy of Edward Lee, MD, Washington, D.C.)

Figure 54-11.  Pathology of gastrointestinal stromal tumor (GIST). A, GISTs frequently have a heterogeneous cut surface with areas of hemorrhage. B, Histologically these tumors are composed of plump spindle cells with prominent cyto­ plasmic vacuoles enmeshed within a myxoid stroma (see also Chapter 30). (Hematoxylin and eosin.)

parietal cell carcinoma.376-379 Miscellaneous benign tumors can involve the stomach and include pancreatic rests, xanthelasma, and fundic gland cysts.

KEY REFERENCES

Cunningham D, Allum WH, Stenning SP, et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med 2006; 355:11-20. (Ref 354.) de Vries AC, van Grieken NC, Looman CW, et al. Gastric cancer risk in patients with premalignant gastric lesions: A nationwide cohort study in the Netherlands. Gastroenterology 2008; 134:945-52. (Ref 206.)

El-Omar EM, Carrington M, Chow WH, et al. Interleukin-1 polymorphisms associated with increased risk of gastric cancer. Nature 2000; 404:398-402. (Ref 120.) Fukase K, Kato M, Kikuchi S, et al. Effect of eradication of Helicobacter pylori on incidence of metachronous gastric carcinoma after endoscopic resection of early gastric cancer: An open-label, randomised controlled trial. Lancet 2008; 372:392-7. (Ref 249.) Hartgrink HH, van de Velde CJ, Putter H, et al. Extended lymph node dissection for gastric cancer: Who may benefit? Final results of the randomized Dutch gastric cancer group trial. J Clin Oncol 2004; 22:2069-77. (Ref 342.) Kaurah P, MacMillan A, Boyd N, et al. Founder and recurrent CDH1 mutations in families with hereditary diffuse gastric cancer. JAMA 2007; 297:2360-72. (Ref 133.)

905

906

Section VI  Stomach and Duodenum Lichtenstein P, Holm NV, Verkasalo PK, et al. Environmental and heritable factors in the causation of cancer—Analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 2000; 343:78-85. (Ref 116.) Macdonald JS, Smalley SR, Benedetti J, et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med 2001; 345:72530. (Ref 355.) McDonald SA, Greaves LC, Gutierrez-Gonzalez L, et al. Mechanisms of field cancerization in the human stomach: The expansion and spread of mutated gastric stem cells. Gastroenterology 2008; 134:500-10. (Ref 14.) Ohnishi N, Yuasa H, Tanaka S, et al. Transgenic expression of Helicobacter pylori CagA induces gastrointestinal and hematopoietic neoplasms in mouse. Proc Natl Acad Sci U S A 2008; 105:1003-8. (Ref 59.) Plummer M, Vivas J, Lopez G, et al. Chemoprevention of precancerous gastric lesions with antioxidant vitamin supplementation: A random-

ized trial in a high-risk population. J Natl Cancer Inst 2007; 99:13746. (Ref 260.) Soetikno R, Kaltenbach T, Yeh R, Gotoda T. Endoscopic mucosal resection for early cancers of the upper gastrointestinal tract. J Clin Oncol 2005; 23:4490-8. (Ref 345.) Tu S, Bhagat G, Cui G, et al. Overexpression of interleukin-1beta induces gastric inflammation and cancer and mobilizes myeloidderived suppressor cells in mice. Cancer Cell 2008; 14:408-19. (Ref 127.) Uemura N, Okamoto S, Yamamoto S, et al. Helicobacter pylori infection and the development of gastric cancer. N Engl J Med 2001; 345:784-9. (Ref 27.) Wong BC, Lam SK, Wong WM, et al. Helicobacter pylori eradication to prevent gastric cancer in a high-risk region of China: A randomized controlled trial. JAMA 2004; 291:187-94. (Ref 246.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

55 Anatomy, Histology, Embryology, and Developmental Anomalies of the Pancreas Bradley A. Barth and J. Steven Burdick

CHAPTER OUTLINE Anatomy  909 Ductal Structures  910 Circulation  911 Lymphatic Drainage  912 Innervation  912 Histology and Ultrastructure  912 Embryology  915 Signaling and Growth Factors  916

The pancreas was one of the last organs in the abdomen to receive the critical attention of anatomists, physiologists, physicians, and surgeons.1,2 It was first referred to as the “finger of the liver” in the Talmud, written between 200 bc and ad 200. Galen named it (though Ruphos, circa ad 100, should probably be credited2), and thought the pancreas served to support and protect blood vessels. Vesalius considered the organ a cushion for the stomach. Little further information was available until Wirsung demonstrated the pancreatic ducts of humans in 1642 and de Graaf discovered pancreatic secretion from the pancreatic fistula of dogs in 1664. The digestive action of pancreatic secretions was discovered almost 200 years later. Eberle in 1834, Purkinje and Pappenheim in 1836, and Valentin in 1844 observed the emulsification of fat, proteolytic activity, and digestion of starch, respectively, by pancreatic juice and extracts. Bernard subsequently demonstrated the digestive action of pancreatic juice on sugar, fats, and proteins, using secretions from pancreatic fistula preparations. Kuhne introduced the term enzyme and isolated trypsin in 1876. The concept of enzymes led shortly to the identification of pancreatic amylase and lipase. In 1889, Chepovalnikoff, a student of Pavlov, discovered enterokinase in the duodenal mucosa, an enzyme that is essential for activation of the proteolytic enzymes. Another of Pavlov’s students, Dolinsky, stimulated pancreatic secretion by instilling acid into the duodenum in 1895. This led to the discovery of secretin by Bayliss and Starling, which proved to be not an enzyme but the first hormone to be identified. The histologic structure of the pancreas was first described in 1869 by Langerhans. Shortly thereafter, Heidenhain3

Developmental Anomalies  917 Annular Pancreas  917 Pancreas Divisum  918 Ectopic Pancreatic Tissue  918 Pancreatic Agenesis  919 Congenital Cysts  919 Anomalous Pancreaticobiliary Unions  919

characterized the periodic postprandial changes that occurred in the histology of the canine pancreas. He found that as the granular regions of cells disappeared after feeding, the enzyme activity in pancreatic juice increased; he concluded that the granules contained the precursors of the digestive enzymes. Pancreatic disease was rarely recorded before the 19th century. Friedreich wrote the first systematic description of pancreatic diseases in 1875. The description of acute pancreatitis by Fitz in 1889 remains a classic. Although Fitz suggested surgery for pancreatitis, surgery for pancreatic neoplasms and other diseases did not become popular until the 1930s, when it did so mainly as a result of the work of Whipple and Brunschwig.

ANATOMY The pancreas is a soft, elongated, flattened gland 12 to 20 cm in length.4-6 The adult gland weighs between 70 and 110 g. The head lies behind the peritoneum of the posterior abdominal wall and has a lobular structure. The pancreas is covered with a fine connective tissue but does not have a true capsule. The head of the pancreas is on the right side and lies within the curvature of the duodenum. The neck, body, and tail of the pancreas lie obliquely in the posterior abdomen, with the tail extending as far as the gastric surface of the spleen (Fig. 55-1). The second and third duodenum curvatures lie around the head of the pancreas. The anterior surface of the head of the pancreas is adjacent to the pylorus, the first part of the

909

910

Section VII  Pancreas Portal vein

Hepatic artery

Bile duct

Splenic artery

S

Superior pancreaticoduodenal artery

G Tail P

Body Neck

L

I

C

Head

A V

Inferior pancreaticoduodenal artery

Superior mesenteric artery Superior Uncinate mesenteric vein process

A Celiac artery

Hepatic artery

K

Figure 55-2.  Normal anatomic relation of the pancreas with other intraabdominal structures as shown by computed tomography. The borders of the pancreas are indicated by arrowheads. The splenic vein is indicated by an arrow. A, aorta; C, vena cava; G, incidental gallstone; I, small intestine; K, left kidney; L, liver; P, portal vein; S, stomach; V, vertebra. (Courtesy of M.P. Federle, MD.)

Bile duct Splenic vein

B

Portal vein Superior mesenteric artery

Superior mesenteric vein

Figure 55-1.  Diagrammatic representations of the pancreas. A, Anterior view. B, Posterior view.

duodenum, and the transverse colon. The posterior surface abuts the hilus and medial border of the right kidney, the inferior vena cava and the right renal vessels, the right gonadal vein, and the right crus of the diaphragm. The uncinate process is a prolongation of pancreatic tissue of variable size and shape. It projects off the lower part of the head of the pancreas, extending upward and to the left. The uncinate process lies anterior to the aorta and inferior vena cava and is covered superiorly by the superior mesenteric vessels that emerge below the neck of the pancreas. There is much variation in the uncinate process, which may even be absent altogether. The neck of the pancreas is a constricted part of the gland extending from the head of the pancreas toward the left, joining the head with the body of the pancreas. It is 1.5 to 2.0 cm long and 3.0 to 4.0 cm wide. Posterior to the neck of the pancreas lies the confluence of the portal vein with the superior mesenteric and splenic veins. Anteriorly it is covered in part by the pylorus and peritoneum of the lesser sac. The neck extends to the right as far as the anterosuperior pancreaticoduodenal artery from the gastroduodenal artery. The body of the pancreas runs toward the left side, anterior to the aorta. It is retroperitoneal and held against the aorta by the peritoneum of the lesser sac. The anterior surface of the body is covered by peritoneum of the omental bursa that separates the stomach from the pancreas. The antrum and body of the stomach and the transverse mesocolon contact the body anteriorly. Posterior to the body of

the pancreas are the aorta, the origin of the superior mesenteric artery, the left crus of the diaphragm, the left kidney, the left adrenal gland, and the splenic vein. The midline part of the body overlies the lumbar spine, which makes this area of the pancreas most vulnerable to abdominal trauma. The body passes laterally and merges with the tail of the pancreas without a discernible junction point. The tail is relatively mobile, its tip usually reaching the hilus of the spleen. With the splenic artery and vein, the tail is contained between the two layers of the splenorenal ligament. The splenocolic ligament attaches the splenic flexure of the colon to the spleen and brings it near the tail of the pancreas. The relationship of the pancreas to important structures in the posterior abdomen is seen in Figure 55-2. The distal end of the common bile duct, the duodenum, and the head of the pancreas form a unit. The common bile duct is located to the right of the gastroduodenal artery in the posterior wall of the duodenum. The bile duct passes through the substance of the pancreatic head, usually to join with the main pancreatic duct for some distance to reach the duodenal papilla (Fig. 55-3A).

DUCTAL STRUCTURES

The main pancreatic duct (of Wirsung) begins near the tail of the pancreas. It is formed from anastomosing ductules draining the lobules of the gland. It courses left to right and is enlarged by additional ducts. Through the tail and body, the duct lies midway between the superior and inferior margins and slightly posterior. The main duct turns caudal and posterior on reaching the head of the pancreas. At the level of the major papilla, the duct turns horizontally to join usually with the common bile duct (see Fig. 55-3A). This short common segment is the ampulla of the bile duct, which terminates in the duodenal papilla. The relationship of the common bile duct and the duct of Wirsung at the papilla is complex. The ducts may open separately at the ampulla and have an interposed septum or may have a common channel. A common channel for bile

Chapter 55  Anatomy, Histology, Embryology, and Developmental Anomalies of the Pancreas Cystic duct

4 mm), and tail (2 to 3 mm) are generally accepted. However, studies have shown an increase in pancreatic duct size with age and pancreatic disease.16-18

Bile duct

CIRCULATION Pancreatic duct

Accessory papilla

Major papilla

A Cystic duct

Accessory papilla

B

Bile duct

Pancreatic duct

Major papilla

Figure 55-3.  Anatomic arrangement of the pancreatic duct system. A, The most common arrangement. Most of the pancreatic secretion empties into the duodenum along with bile through the major papilla. The proximal portion of the embryonic dorsal pancreatic duct remains patent in about 70% of adults and empties through the accessory papilla. B, Pancreas divisum. The embryonic dorsal and ventral ducts fail to fuse. Most of the pancreatic secretion empties through the accessory papilla. Only pancreatic secretions from the uncinate process and part of the head of the pancreas (which are derived from the embryonic ventral pancreas) drain through the duodenal papilla.

and pancreatic secretion is ordinarily formed by the absence of a septum between the biliary and pancreatic ducts as they approach the ampulla of Vater. In adults studied by endoscopic retrograde cholangiopancreatography (ERCP), the length of the common channel averages 4.5 mm, with a range of 1 to 12 mm.7,8 In various series, more than two thirds of patients had some degree of a common channel.9-13 In a large autopsy series, 74% of patients had a common channel, 19% had separate openings, and 7% had an interposed septum.9 The accessory pancreatic duct of Santorini is frequently present and usually communicates with the main duct (see Fig. 55-3A). The accessory duct lies anterior to the bile duct and usually drains into the minor papilla, which lies proximal to the ampulla of Vater in the second duo­ denum. The accessory duct is patent in 70% of autopsy specimens. In about 10% of individuals there is no con­ nection between the accessory duct and the main duct.14 A number of variations in the two pancreatic ducts may be encountered. The greatest diameter of the main pancreatic duct is in the head of the pancreas, and the duct gradually tapers, progressing to the tail of the pancreas. The main duct ranges from 3.1 to 4.8 mm in the head of the pancreas and tapers to 0.9 to 2.4 mm in the tail.15 Specific normal limits of pancreatic duct diameter in the head (4 to 5 mm), body (3 to

The pancreas has a rich circulation derived from branches of the celiac and superior mesenteric arteries.18,19 The head of the pancreas and surrounding duodenum are supplied by two pancreaticoduodenal arterial arcades. They are formed by the anterior and posterior superior pancreaticoduodenal arteries from the hepatic branch of the celiac artery that join a second pair of anterior and posterior inferior pancreaticoduodenal arteries. The gastroduodenal artery arises off the common hepatic branch of the celiac artery. It divides to form the anterior and posterior superior pancreaticoduodenal arteries. The anterosuperior pancreaticoduodenal artery lies on the surface of the pancreas. It provides branches to the anterior surface of the duodenum, proximal jejunum, and pancreas. The artery enters the substance of the pancreas and, on the posterior surface, joins the anteroinferior pancreaticoduodenal artery from the superior mesenteric artery. The anteroinferior pancreaticoduodenal artery arises from the superior mesenteric artery by the inferior margin of the pancreatic neck. The posteroinferior pancreaticoduodenal artery arises from the gastroduodenal artery. Its course is visible on the posterior surface of the pancreas, and branches may join with branches of the gastroduodenal artery or with a branch of the dorsal pancreatic artery. It passes posterior to the pancreatic portion of the bile duct. At the neck, the dorsal pancreatic artery usually arises from the splenic artery. From this, a right branch supplies the head and usually joins the posterior arcade. It also gives off one or two left branches that pass through the body and tail of the pancreas, often making connections with branches of the splenic artery and a more distal connection with the splenic or the left gastroepiploic artery. All major arteries lie posterior to the ducts. The course of the splenic artery is posterior to the body and tail and loops above and below the superior margin of the pancreas. It gives off the dorsal pancreatic artery, which usually joins one of the posterior superior arcades after giving off the inferior pancreatic artery. The caudal pancreatic artery arises from the left gastroepiploic artery or from a splenic branch at the spleen. It joins with branches of the splenic and great pancreatic arteries and other pancreatic arteries. In general, the venous drainage of the pancreas is similar to the arterial blood supply. It flows into the portal venous system, which is formed by the joining of the superior mesenteric and splenic veins at the confluence behind the neck of the pancreas. The portal vein lies behind the pancreas and in front of the inferior vena cava, with the common bile duct to the right and the hepatic artery to the left. The splenic vein originates at the hilum of the spleen and curves behind the tail of the pancreas and below the splenic artery, to the right along the posterior surface of the pancreas. The pancreatic veins drain the neck, body, and tail of the pancreas and join the splenic vein. The pancreaticoduodenal veins lie close to their corresponding arteries and empty into the splenic or portal veins. Because of the close anatomic relationship of the splenic vein with the pancreas, inflammatory or neoplastic diseases involving the pancreatic body and tail can lead to splenic vein occlusion. This in turn can result in retrograde venous drainage toward the splenic hilum and then, by way of flow through the short gastric and left gastroepiploic veins, can create gastric varices.

911

912

Section VII  Pancreas LYMPHATIC DRAINAGE

The lymphatics, in general, drain the surface network of lymph toward regional nodes and are formed near the larger blood vessels.20,21 The superior lymphatic vessels run along the upper border of the pancreas closely with the splenic blood vessels. Those on the left side of the body and tail empty into nodes in the splenic hilum. Those on the right side of the body and the pancreatic neck empty into nodes near the upper border of the head. They also receive tributaries from the anterior and posterior pancreatic surfaces. The inferior lymphatic vessels run with the inferior pancreatic artery. Those that drain the lower left side of the body and tail drain toward nodes in the splenic hilum. The remaining regions of the neck and body drain toward the right. Lymphatic vessel drainage of the pancreatic head is composed of an anterior system and a posterior system. These vessels generally occupy the grooves between the head of the pancreas and the duodenum, near the pancreaticoduodenal blood vessels. Each drainage system (anterior and posterior) also has superior and inferior drainage systems. In addition, a set of lymphatics also drains the upper portion of the head, lying on the superior border. The lymphatic drainage of the head of the pancreas and duodenum eventually flows into the celiac and superior mesenteric groups of pancreatic nodes and into the cisterna chyli. The lymphatics of the tail drain into splenic hilar nodes. The lymphatics of the body pass to the pancreaticosplenic nodes lying along the superior border, which drain into celiac nodes. Lymphatics of the upper head of the pancreas pass through subpyloric nodes. Inferiorly, lymphatics drain into retropancreatic and antepancreatic nodes, which then drain into superior mesenteric nodes.

INNERVATION

The visceral efferent innervation of the pancreas is through the vagi and splanchnic nerves by way of the hepatic and celiac plexuses. The efferent fibers of the vagi pass through these plexuses without synapsing and terminate in parasympathetic ganglia found in the interlobular septa of the pancreas. The postganglionic fibers innervate acini, islets, and ducts. The bodies of the neurons of the sympathetic efferent nerves originate in the lateral gray matter of the thoracic and lumbar spinal cord. The bodies of the postganglionic sympathetic neurons are located in the great plexuses of the abdomen. Their postganglionic fibers innervate only blood vessels. The autonomic fibers, both efferent and afferent, are located in proximity to the blood vessels of the pancreas. Little is known about the distribution of the visceral efferent fibers in humans. They probably run through the splanchnic nerves to the sympathetic trunks and rami communicantes and through spinal nerves and ganglia. The vagi also carry some visceral afferent fibers.

HISTOLOGY AND ULTRASTRUCTURE The pancreas is a compound, finely nodular gland that is grossly similar to but less compact than the salivary glands. It is surrounded by fine connective tissue but does not have a fibrous tissue capsule. The lobules are visible on gross examination and are connected by connective tissue septa that contain the blood vessels, nerves, lymphatics, and excretory ducts (constituting about 18% of this organ). The gland is a mixed exocrine (about 80%) and endocrine (about 2%) organ (Fig. 55-4). The endocrine portion consists of the islets of Langerhans, which are spherical clusters of

Figure 55-4.  Histologic section of human pancreas obtained at autopsy shows dense-staining acinar cells and a light-staining islet of Langerhans just left of the center of the field. A small duct is visible on the left side of the illustration (9 o’clock position). (Hematoxylin and eosin, ×140.)

light-staining cells scattered throughout the pancreas (see Chapter 1). The exocrine portion consists of numerous dark-staining acini composed of tubular and spherical masses of cells, which are the subunits of the lobule.22,23 Silicone casts of the duct lumen formed by retrograde injection indicate that the tubular portions of the acini are extensive and that the exocrine cells are arranged primarily as curved, branching tubules that anastomose and end blindly (Fig. 55-5).24 The lumen of the acinus is the origin of the secretory duct and contains centroacinar cells, which are unique to the pancreas. These cells are pale staining in histologic sections and smaller than the acinar cells. The lumen of the acinus leads into the intralobular ducts, which are covered by low columnar epithelial cells similar in appearance to the centroacinar cells. These ducts are nonstriated and anastomose to form the interlobular ducts, which are lined by a columnar epithelium (see Fig. 55-4). Goblet cells and occasional argentaffin cells also are present. The interlobular ducts anastomose to become the main pancreatic duct. The larger ducts have a somewhat thick wall consisting of connective tissue and elastic fibers. Acinar, ductal, and islet cells can be distinguished by monoclonal antibodies specifically reactive with these cell types.25-27 Acinar cells are tall, pyramidal, or columnar epithelial cells, with their broad bases on a basal lamina and their apices converging on a central lumen (Fig. 55-6). In the resting state, numerous eosinophilic zymogen granules fill the apical portion of the cell. The basal portion of the cells contains one or two centrally located, spherical nuclei and extremely basophilic cytoplasm. The Golgi complex lies between the nucleus and zymogen granules and can be seen as a clear, nonstaining region (see Fig. 55-6). The acinar cells undergo cyclic changes in morphology in response to feeding and digestion.28,29 After a large meal, the zymogen granule content of the cells is depleted. This apparently occurs by a decrease in the size and number of granules.28 After depletion of the granules, the Golgi apparatus may be observed at the apex of the cell and appears more extensive than in the resting state. The reductions in size and number of granules occur with a substantial increase in pancreatic enzyme secretion (see Chapter 56).

Chapter 55  Anatomy, Histology, Embryology, and Developmental Anomalies of the Pancreas

MV Z

L

Z G

GE Z

N M GE Figure 55-5.  Scanning electron micrograph of a silicone cast of the acinar area of human pancreas. The cast is formed by retrograde injection into the pancreatic duct and demonstrates the continuous, branching nature of the acinar pancreas. The diameter of the cast is greater than that of the original acinar lumen. ×300. (From Bockman DE, Boydston WR, Parsa I. Architecture of human pancreas: Implications for early changes in pancreatic disease. Gastroenterology 1983; 85:55. Copyright 1983 by the American Gastroenterological Association.)

Ergastoplasm

Zymogen granules

Centroacinar cells

Golgi complex

Figure 55-6.  Photomicrograph of a human acinus, showing acinar and centroacinar cells. The acinar cell ergastoplasm, Golgi complex, and zymogen granules can be easily identified. Formalin stain, osmium fixation. Epon-embedded section, toluidine blue stain; ×3200. (From Bloom W, Fawcett DWA. Textbook of Histology. 11th ed. Philadelphia: WB Saunders; 1986. Courtesy of Susumu Ito, MD, Boston, Mass.)

The subcellular structure of the acinar cells can be visualized at the electron-microscopic level (Fig. 55-7). The acinar cell has several short, slender microvilli about 0.2 µm in length that extend into the lumen of the acinus. The lumen typically contains flocculent electron-dense material, which presumably is the secreted digestive enzymes. Thin filaments form the axis of the microvilli as well as a network beneath the apical plasmalemma. These microfilaments apparently play a structural role because their disruption causes expansion of the acinar lumen and loss of microvilli.30 Adjacent cells are joined at the apical surface by electron-dense intercellular junctions. Tight junctions form a belt-like band around the apical end of the cell and are

CJ

Figure 55-7.  Electron micrograph of a human acinar cell. CJ, intercellular space; G, Golgi complex; GE, granular endoplasmic reticulum; L, lumen of acinus; M, mitochondria; MV, microvilli; N, nucleus; Z, zymogen granules. ×15,000. (Courtesy of Susumu Ito, MD, Boston, Mass.)

produced by the apposition of the external membrane leaflets of neighboring cells.31 These junctions prevent the reflux of secreted substances from the duct into the intercellular space. Gap junctions are distributed on the lateral cellular membranes and are formed by the apposition of larger, disk-shaped membrane plaques. They allow communication between cells. Below the junctions, the lateral cell borders are relatively straight and have a few, small interdigitations. Pancreatic lateral cell membranes display unique antigenic determinants that are not found on apical cell membranes.32 The nucleus usually is spherical, about 6 µm in diameter,33 with one or more nucleoli in the interior and patches of dense heterochromatin along the inner nuclear membrane. Numerous conspicuous nuclear pores are located at regions where the lightly stained euchromatin makes contact with the nuclear membrane. These pores presumably are the sites where messenger and transfer ribonucleic acids (RNAs) are transported out of the nucleus into the cytoplasm. Binucleate cells also are seen occasionally. Mitochondria are elongate cylindrical structures that may appear oval in cross-section and may contain welldeveloped cristae and many matrix granules. They occur throughout the cytoplasm, among the granular endoplasmic reticulum or zymogen granules, and adjacent to the baso­ lateral cell border. The cytoplasmic matrix occupies about 45% of the cell volume.34 Granular endoplasmic reticulum (see Fig. 55-7) occupies about 20% of the cell volume34,35 and fills most of the basal region of the acinar cells, although small amounts also occur in the apical region adjacent to and among the zymogen granules. This reticulum is composed of numerous parallel cisternal membranes covered with closely spaced attached ribosomes, giving the structures a granular appearance. On the basis of studies with laboratory animals, the ribosomes of the granular endoplasmic reticulum have been found to be the site of protein synthesis.36,37 The Golgi complex is located between the nucleus and the mass of zymogen granules present in the resting gland

913

914

Section VII  Pancreas (see Fig. 55-7). It consists of flattened, membranous saccules as well as small vesicles or vacuoles that contain flocculent electron-dense material. The Golgi saccules have been distinguished from other intracellular vesicles by enzyme cytochemistry and by immunohistochemistry using anti­ enzyme and antireceptor antibodies.38 The Golgi complex is believed to play an important role in the transport of secretory proteins and the formation of zymogen granules. The mechanisms by which these processes occur are still unresolved. The secretory granules of the pancreas usually are divided into two types, electron-lucent condensing vacuoles and electron-dense zymogen granules. The condensing vacuoles are typically seen in the vicinity of the Golgi complex and, on the basis of autoradiographic data,36,37 are believed to be precursors of the zymogen granules. They are membranebound vesicles slightly larger than zymogen granules and much less numerous, occupying only about 2% of the cytoplasm.35 Zymogen granules also are spherical, membranebound vesicles, slightly less than 1 mm in diameter,28,39,40 filled with electron-dense material that apparently represents the digestive enzymes (see Fig. 55-7). Studies of the chemical composition of the zymogen granules have shown that they contain about 12 to 15 different digestive enzymes, which make up about 90% of the granule protein.41-44 Each granule apparently contains the entire complement of secreted digestive enzymes, because labeled antibodies to several different enzymes have been located over single zymogen granules from different cells.45,46 Individual zymogen granules can differ markedly in the concentration of specific digestive enzymes contained within the granules.47 The digestive enzymes within the granules are not in solution or suspension but in a solid-state array, which exhibits specific binding between the enzymes themselves and between the enzymes and the granule membrane.48-51 Isolated zymogen granules are stable at slightly acid pH; at alkaline pH, they release their enzymes into solution.49,51,52 This behavior may account for the solubilization of digestive enzymes within the alkaline duct lumen. Along the basal surface of the acinar cells, but not extending between adjacent cells, is a thin basal lamina, below which are collagen fibers and a rich capillary network. Efferent nerve fibers, derived from the sympathetic and parasympathetic systems, penetrate the basal lamina and terminate adjacent to the acinar cells. The centroacinar cells (Fig. 55-8) and duct cells have electron-lucent cytoplasm containing few cytoplasmic organelles or specializations. They typically contain free ribosomes and small, round mitochondria. They contain virtually no granular endoplasmic reticulum and therefore are not active in protein synthesis for secretion. Farther down the ducts, the cells contain more mitochondria, but they are never associated with invaginations of the baso­ lateral surface as occur in the transporting ductal epithelium of the salivary glands. Both centroacinar and duct cells secrete bicarbonate and water. Carbonic anhydrase, the enzyme responsible for formation of bicarbonate, has been demonstrated in the epithelium.23 The islets of Langerhans number about 1 million in the human pancreas and consist of anastomosing cords of polygonal endocrine cells (see Fig. 55-4). Each islet is about 0.2 mm in diameter, much larger than an acinus, and separated from the surrounding exocrine tissue by fine connective tissue fibers, which are continuous with those of the exocrine gland. Each islet is surrounded and penetrated by a rich network of capillaries lined by a fenestrated endothelium. The capil-

A C A L A

Figure 55-8.  Electron micrograph of a centroacinar cell (C) and several acinar cells (A). Note the electron-lucent cytoplasm, scattered mitochondria, and lack of other membranous organelles in the centroacinar cell. L, lumen of the acinus. ×9000. (Courtesy of Susumu Ito, MD, Boston, Mass.)

Acini

Islet

Figure 55-9.  Schematic diagram of the insuloacinar portal system, illustrating the dual blood supply to the exocrine pancreas. (From Goldfine ID, Williams JA. Receptors for insulin and CCK in the acinar pancreas: Relationship to hormone action. Int Rev Cytol 1983; 85:1.)

laries are arranged in a portal system that conveys blood from the islets to acinar cells (Fig. 55-9).53-57 This insulaacinar portal system consists of afferent arterioles that enter the islet, form a capillary glomerulus, and leave the islet as efferent capillaries passing into the exocrine tissue. A parallel arterial system supplies blood directly to the exocrine pancreas (see Fig. 55-9), and yet this portal system permits the local action of islet hormones, especially insulin, on the exocrine pancreas.55-57 Acinar cells surrounding islets of Langerhans, termed peri-insular acini, are morphologically and biochemically different from acini situated farther away (tele-insular acini).58,59 Peri-insular acini have larger cells, nuclei, and zymogen granule regions58 and different ratios of specific digestive enzymes.59

Chapter 55  Anatomy, Histology, Embryology, and Developmental Anomalies of the Pancreas Although the acinar cell secretes several different digestive enzymes in the exocrine pancreas, each cell type in the endocrine pancreas appears to secrete a single hormone. The four major types of cells found are B cells, A cells, D cells, and pancreatic polypeptide (PP) cells.60,61 B cells (beta cells), the most numerous (50% to 80%), secrete insulin.60 A cells or alpha cells (5% to 20%) secrete glucagon. PP cells (10% to 35%) secrete pancreatic polypeptide. D cells (5%) secrete somatostatin. Other rare cell types occur in the islet.62 In humans the islets are subdivided into units, each of which exhibits a central aggregation of B cells surrounded by varying numbers of peripherally located cells that secrete the other hormones.

EMBRYOLOGY The pancreas first appears in embryos of about 4 mm in the fourth week of gestation.63,64 The outpouchings from the endodermal lining of the duodenum develop at this time: the ventral pancreas and the dorsal pancreas (Fig. 55-10A). The dorsal anlage grows more rapidly, and by the sixth week, it is an elongated nodular structure extending into the dorsal mesentery, within which its growth continues (see Fig. 55-10B). The ventral pancreas remains smaller and is carried away from the duodenum by its connection with the common bile duct. The two primordia are brought into apposition by uneven growth of the duodenum, and they fuse by the seventh week (see Fig. 55-10C). The tail,

body, and part of the head of the pancreas are formed by the dorsal component; the remainder of the head and the uncinate process derive from the ventral pancreas. These primitive relations are still distinguishable in the adult pancreas.64 Both of the primitive pancreata contain an axial duct. The dorsal duct arises directly from the duodenal wall, and the ventral duct arises from the common bile duct. On fusion of the ventral and dorsal components, the ventral duct anastomoses with the dorsal one, forming the main pancreatic duct (see Fig. 55-10D). The proximal end of the dorsal duct becomes the accessory duct of Santorini in the adult and is patent in 70% of specimens.65 The common outlet of the bile duct and pancreatic duct observed in most adults is the result of the common origin of the bile duct and the ventral pancreas. The pancreatic acini appear in the third month of gestation as derivatives of the side ducts and termini of the primitive ducts. The acini remain connected to the larger pancreatic ducts by small secretory ductules. The primitive pancreas is composed of relatively undifferentiated epithelial cells similar in morphology to duct cells. Mesenchymal tissue in which the gland grows provides the thin connective tissue capsule and divides the gland into lobes and lobules. Distinct differences in morphology, enzyme content, and secretory capacity exist between the embryonic pancreas and the adult pancreas.66-72 Early in development, the pancreas has no zymogen granules and little granular endoplasmic reticulum.66-70 In humans, the pancreas is composed of

Common duct Ventral pancreas

Gallbladder Gallbladder

A

Dorsal pancreas

Ventral pancreas

Common duct

B

Dorsal pancreas

Dorsal pancreas Accessory papilla Accessory duct Ventral pancreas

C

Anastomosis of ducts

Duodenal papilla

Main pancreatic duct

D

Head of pancreas

Figure 55-10.  Stages in the embryonic development of the pancreas. A, At about 4 weeks of gestation, dorsal and ventral buds are formed from the duodenum. B, At 6 weeks the ventral pancreas extends toward the larger dorsal pancreas. C, By about 7 weeks fusion of dorsal and ventral pancreas has occurred and ductular anastomosis is beginning. D, At birth the pancreas is a single organ, and ductular anastomosis is complete. (Modified from Arey LB. Developmental anatomy: A Textbook and Laboratory Manual of Embryology. 7th ed. Philadelphia: WB Saunders; 1974.)

915

916

Section VII  Pancreas undifferentiated epithelial cells at 9 weeks of gestation.66 During subsequent cell differentiation, the specific activity of the digestive enzymes increases some thousand-fold,71,72 and the granules increase in size and come to occupy most of the cytoplasm of the cells, including the basolateral regions.67-70 At 12 weeks of gestation in humans, zymogen granules are first seen on electron microscopy. The cells also contain a Golgi complex and granular endoplasmic reticulum in relatively small amounts. By 20 weeks of gestation, larger zymogen granules typical of the adult are seen.66 Each digestive enzyme has a characteristic rate of accumulation and increases in concentration at different times.71,72 At birth, the granules in laboratory animals are the largest normally found in the pancreas, being about six times the volume of the granules in adults.73,74 At about this time, the capacity for stimulated secretion is attained. Differentiation of the pancreas continues beyond birth with regard to the size of the zymogen granules as well as the enzyme content of the tissue.75,76 The endocrine pancreas differentiates on roughly the same time course as the exocrine portion,77,78 and the appearance of islet hormones also precedes the appearance of secretion granules in the cells. In humans, endocrine cells are first observed singly or in small clusters along the basolateral portion of undifferentiated acinar cells (9 to 10 weeks), but by 12 to 16 weeks, distinct islets in various stages of complexity can be observed. The development of islets in the PP-rich regions lags slightly behind that of the glucagon-rich regions.79 Insulin cell numbers increase continuously with age, whereas glucagon cell numbers increase during fetal life and then decrease in infants and adults. The number of somatostatin cells is elevated in fetal and infant stages, whereas PP cells are the least abundant cells during these stages.79

SIGNALING AND GROWTH FACTORS Pancreatic development is influenced by a variety of signaling and transcription factors. Signaling factors originate from the mesenchyme, surrounding tissues, and vascular structures. Insight into pancreatic regulation by these pathways offers insight into autocrine function, endocrine and exocrine insufficiency, pancreatic cancer, pancreatic cysts, and chronic pancreatitis with potential diagnostic and therapeutic options. Importantly these pathways offer explanations for maladies including annular pancreas, pancreas divisum, and pancreatic heterotopia (rests), to be discussed later. Hedgehog (Hh) proteins are signaling molecules that regulate various aspects of morphogenesis, including cellular proliferation and differentiation (see also Chapter 3).80 There are three mammalian hedgehog genes: sonic (Shh), desert (Dhh), and Indian (Ihh), which encode secreted proteins that elicit concentration-dependent responses from target cells.81 Hedgehogs are extracellular proteins that act as signaling ligands. The transmembrane receptors patched (Ptc) and smoothened (Smo) are regulated through this interaction.82 Ptc is a negative regulator of Smo function. When Hh binds to Ptc, Smo is released from the inhi­ bitory effects. This leads to the activation of transcriptional factors. Indian hedgehog protein stimulates pancreatic growth. In contrast, sonic hedgehog protein expression inhibits pancreatic development and is excluded from developing pancreatic tissue.83 Transgenic mice that overexpress sonic hedgehog demonstrate marked reductions in exocrine and

endocrine pancreatic tissue. Indian hedgehog, when blocked in animal models, leads to a reduction in pancreatic size and impaired endocrine cell development in mice. Furthermore, homozygous inactivation of Ihh in mice resulted in the development of an annular pancreas in 42% of the mice.84 Inactivation of sonic hedgehog has also been linked with annular pancreatic development.85 A model for pancreas divisum in mice includes genotypes heterozygous for the null alleles of Shh and Ihh and deficient Smad2.84,86 Shh mutations in humans are associated with gut malrotations and imperforate anus, and with annular pancreas.84,85 The sonic hedgehog gene also functions to inhibit ectopic pancreatic formation in the stomach, duodenum, and liver. Inhibition of hedgehog signaling in chick embryos leads to ectopic budding of pancreatic structures in the stomach and duodenum.87 This offers a potential explanation of pancreatic heterotopia (pancreatic rests). The role of undifferentiated cells and their control in pancreatic development again becomes relevant in adults when the same mechanisms influence repair, replacement, and regeneration. Uncontrolled proliferation seen in pancreatic neoplasms has been linked with abnormal hedgehog signaling.82,88 Sonic hedgehog is normally not found in the mature pancreas (as opposed to Indian and desert hedgehog, which are normally expressed).89 Pancreatic cancer is associated with expression of Shh and up-regulation of Ihh. Higher levels of expression of Shh are seen with advancing degrees of atypia. Pancreatic intraepithelial neoplasia (PanIN) type lesions are also seen in transgenic mice overexpressing Shh. PanIN lesions are associated with intraductal papillary mucinous neoplasms (IPMN) of the pancreas and can progress to adenocarcinoma (see Chapter 60). The oncogenic function of hedgehog proteins has been term a “clever trick” because it does not appear to be directly targeted to tumor cells but rather to nonmalignant stromal cells that, as part of the tumor microenvironment, support tumor growth.90,91 Inhibitors of hedgehog activity may be novel therapies for neoplasia; furthermore, expression of Shh offers a potential means of diagnosing pancreatic adenocarcinoma. Pancreatic duct primary cilia and hedgehog proteins interface in animal and human cell lines during both pancreatic development and homeostasis. Primary cilia are organelles that act as sensors in the extracellular environment and coordinate multiple signal transduction pathways. Defects in ciliary assembly have been associated with pancreatic cysts, pancreatitis, and diabetes.92 Regulation of the Hh pathway is complex, with Gli transcription factors, effectors of Hh signaling, having either activation or repressor function and localization of these modulating functions to the cilium and cell body. In the developing human pancreas, Hh was linked to a ciliary localization in pancreatic duct epithelial cells. PANC1 and CFPAC-1 are human cancer cells lines derived from an epithelioid carcinoma and a patient with a cystic fibrosis transmembrane regulator (CFTR) mutation 508, respectively. The cilia in the proliferating areas expressed the transmembrane receptors Ptc and Smo. The abnormal cilia have been linked with a marker of proliferative risk.93 Abnormal pancreatic cilia associated with pancreatitis and pancreatic ductal dilation have increased beta-catenin expression, which is associated with activation of the Wnt pathway. Abnormal Wnt signaling has also been noted in cystic kidney disease. Dorsal and ventral pancreatic buds develop from the endoderm of the foregut and appear as two independent thickenings or anlages.11 These two anlages are not identical, are asynchronous, and receive distinct signals from

Chapter 55  Anatomy, Histology, Embryology, and Developmental Anomalies of the Pancreas their surrounding tissues. In order to understand their development further, we must reacquaint ourselves with two terms—notochord and homeobox (Pdx1). Notochord is composed of cells derived from the mesoderm defining the primitive axis of the body, and is the center of development of the axial skeleton. Homeobox refers to a class of highly conserved DNA sequences encoding protein domains involved in binding to DNA. These occur in genes involved in the control of development in humans. The dorsal pancreatic bud develops in proximity to the notochord. As the notochord is displaced by the dorsolateral splanchnic mesenchyme, a variety of new signals emerge from the aorta. Shh activity along with fibroblast growth factor (FGF), transforming growth factor-β (TGF-β), and vascular endothelial growth factor (VEGF) allows homeobox transcription factor (Pdx1) expression to mark duodenal cells and promote the dorsal pancreatic bud. The ventral bud development does not depend on the notochord but instead is closely connected to liver organogenesis. The initial pancreatic endoderm from both areas consist of progenitors of endocrine and exocrine cells. The cells express the homeobox transcription factor Pdx1, which is necessary for pancreatic and duodenal differentiation after the formation of the ventral and dorsal pancreatic buds. Defects that occur in Pdx1-null mice include abnormalities in the gastroduodenal junction, submucosal Brunner’s glands, enteroendocrine cell numbers in the stomach and duodenum, and formation of peribiliary glands and mucin-producing cells in the gallbladder.94-97 The interplay between the morphogenesis and cytodifferentiation is less well understood.98 The availability of a pool of precursor cells involved in maintenance, replacement, and repair of cells offers the potential for hypo­ proliferative and proliferative disorders. The mesenchymal growth factors and pancreatic epithelium work through the Notch pathway to regulate exocrine and endocrine cell differentiation.99 There are four Notch genes in mammals, each of which encodes a transmembrane receptor. Binding of ligands leads to intracellular cleavage of the Notch receptor with activation of the intracellular domain, allowing interaction with the deoxyribonucleic acid (DNA)–binding protein RBP-Jh. This induces expression of the basic helix-loop-helix HES genes, which down-regulate expression of target genes. The pancreatic anlage epithelium expresses the Notch1 and HES genes. The anlage develops into a branching epithelium, which is believed to be the source of endocrine and exocrine stem cells. Notch2 is expressed in high levels in this branched epithelium.100 Notch signaling defect knockout mice have a marked endocrine cell differentiation with depletion of epithelial precursors and concomitant exocrine hypoplasia.101 The Notch pathway operates as a negative regulator of endodermal endocrine differentiation. The Notch pathway controls expression of HES1, a basic helix-loop-helix gene that represses positive basic helix-loop-helix genes. In HES1-deficient mice, biliary epithelium differentiates into pancreas-like endocrine and exocrine cells, forming acini and islet cells.102 Thus, biliary epithelium has the potential for pancreatic differentiation, and HES1 determines biliary organogenesis by preventing pancreatic differentiation. Notch signaling allows cells to remain in a nondifferentiated proliferative state. It likely has the role in adults of maintaining the undifferentiated pool of pancreatic stem cells that are involved in islet cell turnover, replacement, and organ repair.101 Up-regulation of the Notch pathway has been linked to metaplastic conversion of exocrine cells in adult pancreatic explants encoded with adenovirus-

activated Notch. The exocrine epithelium converted to ductal cells, and the precursor population expanded.

DEVELOPMENTAL ANOMALIES ANNULAR PANCREAS

Annular pancreas is a congenital anomaly in which a portion of the pancreas forms a thin band around the preampullary portion of the duodenum, leading to complete or partial obstruction (Fig. 55-11). The incidence of annular pancreas is estimated to be between 1 in 1000103-105 and 3 in 20,000,106-107 based on retrospective studies of patients undergoing abdominal imaging and autopsy, respectively. Strong evidence suggests that genetic factors are involved in the pathogenesis of annular pancreas. There are reports of annular pancreas occurring in siblings108 and identical twins.109 Moreover, annular pancreas is more common in patients with other congenital anomalies such as trisomy 21, cardiac defects, malrotation, genitourinary anomalies, and tracheoesophageal fistula.110,111 The role of Hh genes in pancreatic organogenesis is discussed earlier in this chapter. Shh and Ihh gene defects have shown 42% to 85% incidence of annular pancreas in mouse models.112,113 Although annular pancreas is often diagnosed prenatally or during infancy, it is erroneous to consider it solely a disease of infancy. A review by Zyromski and associates revealed a second peak of detection in the fourth through seventh decades of life.111 In this study, symptoms differed dramatically in adults diagnosed with annular pancreas compared to children. Pediatric patients not identified by prenatal ultrasound tended to present with vomiting,

Figure 55-11.  Annular pancreas, causing a duodenal stricture. Barium contrast upper gastrointestinal series demonstrating a mid-duodenal stricture (arrow) with proximal dilatation, findings compatible with annular pancreas. Annular pancreas was identified on computed tomography scan and confirmed during surgery. (Courtesy of Michael Federle, MD.)

917

918

Section VII  Pancreas

A

B

Figure 55-12.  Pancreas divisum as demonstrated by endoscopic retrograde pancreatography. A, The major (duodenal) papilla has been cannulated. The duct terminates without communicating with the main pancreatic duct. B, The minor (accessory) papilla has been cannulated, and the main pancreatic duct is filled. (Courtesy of Markus Goldschmiedt, MD, Richardson, Tex.)

typically nonbilious, and feeding intolerance. In adults, the majority presented with abdominal pain, pancreatitis, evidence of biliary obstruction, or with nausea, vomiting, and bloating.111 Diagnosis in children is suspected by findings on abdominal radiographs, ultrasonography, or upper gastrointestinal (GI) series. In adults, computed tomography (CT) scan, magnetic resonance cholangiopancreatography (MRCP), or endoscopic retrograde cholangiopancrea­ tography (ERCP) may be more commonly used. Not infrequently, the diagnosis of annular pancreas is made at laparotomy. Surgical duodenoduodenostomy appears to be an effective therapeutic approach, and is considered the treatment of choice in pediatric patients and in some adult patients. Complex pancreatic surgery is more likely to be required in adults compared with children.110,111 The risk of pancreaticobiliary neoplasia is significantly increased in adults with annular pancreas.111,114 Thus, ongoing cancer screening and surveillance should be considered in this population.

PANCREAS DIVISUM

Pancreas divisum (PD) results from a failure of the dorsal and ventral pancreatic ducts to fuse during embryogenesis (Fig. 55-12). Thus, the bulk of pancreatic exocrine secretions must drain through the relatively small dorsal duct of Santorini and minor papilla rather than the ventral duct of Wirsung and the major papilla. Pancreas divisum has been detected in 5% to 10% of the population in autopsy studies and in a similar percentage of patients undergoing ERCP.115-117 However, Cotton reported that 25.6% of adult patients undergoing ERCP for evaluation of unexplained recurrent pancreatitis were found to have pancreas divisum.118 Pancreas divisum can be diagnosed by ERP, endoscopic ultrasound, abdominal CT, or MRCP. While only a minority of patients with PD develop symptoms due to their altered anatomy, it is becoming clear that endoscopic therapy can relieve symptoms in some patients with acute recurrent and chronic pancreatitis (see Chapters 58 and 59). In a small randomized, controlled trial by Lans and colleagues, 90% of patients with pancreas divisum and acute pancreatitis experienced symptomatic improvement after undergoing dorsal duct stent placement compared with only 11% of control patients with pancreas

Figure 55-13.  Endoscopic view of a pancreatic rest in the gastric antrum of an eight-year-old boy. Note the central umbilication.

divisum and pancreatitis not stented.119 More recently, Gerke and associates reported that 73% of patients with pancreas divisum and acute recurrent pancreatitis had short-term improvement from minor papilla sphincterot­ omy. Clinical improvement was much less likely in patients with pancreatic-type pain rather than true acute recurrent pancreatitis.120

ECTOPIC PANCREATIC TISSUE

Ectopic (heterotopic) pancreatic tissue, often referred to as a pancreatic rest, occurs in 0.55% to 13.7% of the population according to autopsy material, the most common sites being the stomach (Fig. 55-13), duodenum, proximal jejunum, and ileum. Less frequently heterotopic pancreatic tissue can be found in the umbilicus, common bile duct, gallbladder, and Meckel’s diverticula.121 Although rarely clinically significant, ectopic pancreatic tissue has been associated with ulceration and bleeding, intussusception,122 and even adenocarcinoma.123 Treatment is surgical resection when the tissue is symptomatic, but whether to remove

Chapter 55  Anatomy, Histology, Embryology, and Developmental Anomalies of the Pancreas ectopic pancreatic tissue that is found incidentally remains controversial.

PANCREATIC AGENESIS

Pancreatic agenesis is a rare condition that can be complete or partial. Complete pancreatic agenesis is mostly fatal as infants are stricken with diabetes and malabsorption, as well as intrauterine growth restriction due to lack of insulin, an important intrauterine growth factor.124,125 Mice homozygous for a targeted mutation in the insulin-promoter-factor 1 gene are born without a pancreas.126 Partial pancreatic agenesis, or agenesis of the dorsal pancreas, may be less significant clinically due to the presence of some functioning pancreatic tissue. Also known as congenital short pancreas, agenesis of the dorsal pancreas has been associated with polysplenia and intestinal malrotation.127 Pancreatic agenesis should be suspected based on clinical findings, and confirmed with MRI. This condition is discussed in further detail, as well as its genetic basis, in Chapter 57.

CONGENITAL CYSTS

Congenital cysts of the pancreas are rare and may be diagnosed at any age, even prenatally. The cysts may be solitary or multiple, and are distinguished from pancreatic pseudocysts by the presence of an epithelial lining. The clinical presentation is variable, ranging from an incidental finding on an imaging study to an abdominal mass, vomiting, biliary obstruction, or acute pancreatitis. Multiple pancreatic cysts tend to occur in patients with associated anomalies and may be seen in systemic disorders such as von Hippel-Landau syndrome or polycystic kidney disease.128 A review of 15 cases of congenital pancreatic cysts revealed that the majority of patients presented before the age of two years, and associated anomalies were found in 30% of cases.129 Congenital pancreatic cysts are more often located in the body and tail of the pancreas than in the head, and surgical therapy consists of total excision when possible. Cysts in the pancreatic head may be addressed using endoscopic or surgical drainage procedures when necessary.

ANOMALOUS PANCREATICOBILIARY UNIONS

Anomalous pancreaticobiliary union (APBU) is a congenital malformation of the confluence of the pancreatic and bile ducts. A common channel for the bile and pancreatic fluid is formed by the absence of a septum between the ducts. When this malunion occurs, the influence of the sphincter of Oddi is lost, allowing reflux of pancreatic exocrine secretions into the biliary system. This places affected patients at increased risk for pancreatitis, choledochal cysts, and neoplasms that manifest in adulthood. A classification for pancreaticobiliary anomalous junction has been proposed by dividing it into three types: a pb type, in which the pancreatic duct appears to join the common bile duct; a bp type, in which the insertion of the bile duct is into the pancreatic duct; and a Y type, in which there is a long common channel measuring greater than 15 mm in length.130,131 The bp and

pb types have each been reported to be the most common type in large series.132,133 Approximately 15% of children with recurrent pancreatitis have this disorder.133 Choledochal cysts are frequent with this abnormality (94% and 100% in two series).132,133 The increased risk of biliary tract malignancy in patients with APBU is well documented. Malunions were seen in 62.5% of adults with gallbladder cancer, in 50% of patients with gallbladder adenomyomatosis, and in 33.3% of patients with common bile duct cancer.132 Patients with APBU have increased cellular proliferative activity of the gallbladder mucosa, even in early childhood.134 Although there is controversy concerning management of APBU among patients of various ethnic groups,135-137 given the cancer risk for the patient with a dilated bile duct associated with a pancreaticobiliary malunion, consideration for cholecystectomy, resection of the bile duct, and hepaticojejunostomy may be advised.138

KEY REFERENCES

Cano DA, Hebrok M, Zenker M. Pancreatic development and disease. Gastroenterology 2007; 132:745-62. (Ref 125.) Cano DA, Sekine S, Hebrok M. Primary cilia deletion in pancreatic epithelial cells results in cyst formation and pancreatitis. Gastroenterology 2006; 131:1856-69. (Ref 92.) Cotton PB. Congenital anomaly of pancreas divisum as cause of obstructive pain and pancreatitis. Gut 1980; 21:105-14. (Ref 118.) Fogel EL, Zyromski NJ, McHenry L, et al. Annular pancreas in the adult: Experience at a large pancreatobiliary endoscopy center. Gastrointest Endosc 2006; 63:308. (Ref 105.) Foo FJ, Gill U, Verbeke CS, et al. Ampullary carcinoma associated with an annular pancreas. JOP 2007; 8:50-4. (Ref 114.) Gerke H, Byrne MF, Stiffler HL, et al. Outcome of endoscopic minor papillotomy in patients with symptomatic pancreas divisum. JOP 2004; 5:122-31. (Ref 120.) Hebrok M, Kim SK, St. Jacques B, et al. Regulation of pancreas development by hedgehog signaling. Development 2000; 127:4905-13. (Ref 113.) Jimenez JC, Emil S, Podnon Y, Nguyen N. Annular pancreas in children: A recent decade’s experience. J Pediatr Surg 2004; 39:1654-7. (Ref 110.) Kapa S, Gleeson FC, Vege SS. Dorsal pancreas agenesis and polysplenia/ heterotaxy syndrome: A novel association with aortic coarctation and a review of the literature. JOP 2007; 8:433-7. (Ref 127.) Matsushita M, Takakuwa H, Matsubayashi Y, et al. Management of anomalous pancreaticobiliary union. Endoscopy 2005; 37:682. (Ref 136.) Nielsen S, Mollgård K, Clement CA, et al. Characterization of primary cilia and hedgehog signaling during development of the human pancreas and in human pancreatic duct cancer cell lines. Dev Dyn 2008; 237:2039-52. (Ref 93.) Ramalho-Santos M, Melton DA, McMahon AP. Hedgehog signals regulate multiple aspects of gastrointestinal development. Development 2000; 127:2763-72. (Ref 112.) Tashiro S, Imaizumi T, Ohkawa H, et al. Committee for registration of the Japanese study group on pancreaticobiliary maljunction: Retrospective and nationwide survey in Japan. J Hepatobiliary Pancreat Surg 2003; 10:345. (Ref 138.) Zyromski NN, Sandoval JA, Pitt HA, et al. Annular pancreas: Dramatic differences between children and adults. J Am Coll Surg 2008; 206:1019-25. (Ref 111.) Full references for this chapter can be found on www.expertconsult.com.

919

CHAPTE R

56 Pancreatic Secretion Stephen J. Pandol

CHAPTER OUTLINE Functional Anatomy  921 Composition of Exocrine Secretions  921 Inorganic Constituents  921 Organic Constituents  923 Functions of the Major Digestive Enzymes  923 Amylase  923 Lipases  924 Proteases  924 Digestive Enzyme Synthesis and Transport  924 Regulation of Protein Synthesis  925

As has been reviewed in Chapter 55, the pancreas is an exocrine organ as well as an endocrine organ. This chapter is devoted to the exocrine pancreas. The exocrine pancreas has been of considerable interest to physiologists and other scientists for quite some time; in fact, the first demonstration of a hormone action was in the pancreas around the turn of the 20th century.1 The pancreas has been the major organ used to demonstrate the mechanisms of synthesis and transport for exportable proteins2 as well as the signaling pathways involved in regulated protein secretion.3 This chapter presents a concise description of the current understanding of pancreatic physiology.

FUNCTIONAL ANATOMY The functional unit of the exocrine pancreas is composed of an acinus and its draining ductule (Fig. 56-1).4 The ductal epithelium extends to the lumen of the acinus with the centroacinar cell situated between the ductal epithelium and the acinus. The function of centroacinar is not established but may play a role in providing progenitor cells for pancreatic cell lineages. The ductule drains into interlobular (intercalated) ducts, which in turn drain into the main pancreatic ductal system. The acinus (from the Latin term meaning “berry in a cluster”) can be spherical or tubular, as shown in Figure 56-1, or can have some other irregular form.4 The acinar cells are specialized to synthesize, store, and secrete digestive enzymes. On the basolateral membrane are receptors for hormones and neurotransmitters that stimulate secretion of the enzymes.3 The basal aspect of the cell contains the nucleus as well as abundant rough endoplasmic reticulum for protein synthesis (Fig. 56-2, left). The apical region of the cell contains zymogen granules, the store of digestive enzymes. The apical surface of the acinar cell also possesses microvilli. Within the microvilli and in the cytoplasm underlying the apical plasma membrane is a filamentous actin meshwork that is involved in exocyto-

Cellular Regulation of Enzyme Secretion  925 Organ Physiology  926 Interdigestive Secretion  926 Digestive Secretion  926 Feedback Regulation  927 Pancreatic Secretory Function Tests  928 Direct Tests  928 Indirect Tests  929

sis of the contents of the zymogen granules.5,6 Secretion is into the lumen of the acinus. Tight junctions between acinar cells form a band around the apical aspects of the cells and act as a barrier to prevent passage of large molecules such as the digestive enzymes.7 The junctional complexes also provide for the paracellular passage of water and ions. Another intercellular connection between acinar cells is the gap junction. This specialized area of the plasma membrane between adjacent cells acts as a pore to allow small molecules (molecular weight 500 to 1000 Da) to pass between cells. The gap junction allows chemical and electrical communication between cells.3 For example, calcium signaling is coordinated between the cells of an acinus with effects on digestive enzyme secretion.8,9 The duct epithelium consists of cells that are cuboidal to pyramidal and contain the abundant mitochondria necessary for energy products needed for ion transport (see Fig. 56-2, right). The duct cells as well as the centroacinar cells contain carbonic anhydrase, which is important for their ability to secrete bicarbonate.10

COMPOSITION OF EXOCRINE SECRETIONS INORGANIC CONSTITUENTS

The principal inorganic components of exocrine pancreatic secretions are water, sodium, potassium, chloride, and bicarbonate (Fig. 56-3). The purposes of the water and ion secretions are to deliver digestive enzymes to the intestinal lumen and to help neutralize gastric acid emptied into the duodenum. Pancreatic juice secreted during stimulation with secretin is clear, colorless, alkaline, and isotonic with plasma. The flow rate increases from an average rate of 0.2 or 0.3 mL/minute in the resting (interdigestive) state to 4.0 mL/minute during postprandial stimulation. The total daily volume of secretion is 2.5 L. The osmolality of pancreatic juice is independent of flow rate. However, when the pancreas is stimulated by secretin (the major mediator

921

Section VII  Pancreas Endocrine

Acinus

150

Neurocrine

Na+ Ach GRP VIP Substance P

CCK secretin

Secretin

HCO3 – Concentration mM

922

100

50

Ach

Cl– K+ 0

Duct Figure 56-1.  Regulation of exocrine pancreatic secretion during the intestinal phase of digestion. Exocrine secretions are regulated by endocrine and neurocrine pathways. In addition, the endocrine and neurocrine mediators regulate secretion from the acinus and the duct differ. For the acinus, cholecystokinin (CCK), secretin, acetylcholine (Ach), gastrin-releasing peptide (GRP), vasoactive intestinal polypeptide (VIP), and substance P regulate secretion. (Signaling mechanisms [second messengers] are shown in Fig. 56-6.) Secretin and Ach are the major regulators of pancreatic bicarbonate secretion from the duct. (Transporters involved in ductal secretion are illustrated in Fig. 56-4.) (Adapted from Gorelick F, Pandol SJ, Topazian M. Pancreatic physiology, pathophysiology, acute and chronic pancreatitis. Gastrointestinal Teaching Project, American Gastroenterological Association, 2003.)

ACINAR CELL Lumen

1

2

3

4

Flow (mL/min) Figure 56-3.  Secretion of pancreatic bicarbonate and other electrolytes. With stimulation (i.e., a meal) there is an increase in the flow of pancreatic secretions. Furthermore, with rising flow rates there are dramatic changes in the concentrations of chloride and bicarbonate in pancreatic juice. The increase in bicarbonate concentration results in an alkaline secretion. The bicarbonate comes from ductal epithelial cells in the pancreas. In contrast to acinar cells, the ducts secrete a large volume of fluid with a high concentration of bicarbonate. The volume of secretion from the acinar cells is believed to be small compared with ductal secretion, so with increasing stimulation of the pancreas, the concentration of ions approaches that of the ductal secretions. Of note, the alkaline secretions from the pancreas, the biliary system, and the duodenal mucosa neutralize the acid secretion delivered to the duodenum from the stomach. This pH-neutral environment is important for optimal digestive enzyme and intestinal mucosal function. (Adapted from Gorelick F, Pandol SJ, Topazian M. Pancreatic physiology, pathophysiology, acute and chronic pancreatitis. Gastrointestinal Teaching Project, American Gastroenterological Association, 2003.)

DUCT CELL Lumen

Zymogen granules

Prominent endoplasmic reticulum

0

Numerous mitochondria

Protein synthesis Ion transport and secretion Figure 56-2.  Ultrastructure of exocrine cells. The ultrastructure of exocrine cells reflects their specialized function. The pancreatic acinar cell (left) and duct cell (right) are both polarized, with clearly defined apical (luminal), lateral, and basal domains. The pancreatic acinar cell has prominent basally located rough endoplasmic reticulum for synthesis of digestive enzymes and apically located zymogen granules for storage and secretion of digestive enzymes. The pancreatic duct cell contains numerous mitochondria for energy generation needed for its ion transport functions. (Adapted from Gorelick F, Pandol SJ, Topazian M. Pancreatic physiology, pathophysiology, acute and chronic pancreatitis. Gastrointestinal Teaching Project, American Gastroenterological Association, 2003.)

of the greater volume output), bicarbonate and chloride concentrations change reciprocally (see Fig. 56-3). Secretin stimulates secretion by binding to its receptor on the acinar cell and activating adenylate cyclase and increasing cyclic adenosine monophosphate (cAMP) in the ductal cell; acetylcholine does so by binding to its receptor and raising intracellular calcium concentrations. The initial events involve cAMP-dependent and Ca2+-dependent chloride (Cl−) channel activation on the luminal membrane as well as K+ channel activation on the basolateral membrane (Fig. 56-4).10-12 The cAMP-dependent Cl− channel is the cystic fibrosis transmembrane conductance regulator (CFTR).10-13 The activation of both channels by cAMP leads to Cl− secretion into the lumen. The higher chloride concentration in the lumen is coupled to a Cl−-HCO3− antiport, resulting in an exchange of Cl− for HCO3− in the lumen. Newer evidence also suggests a HCO3− channel on the apical surface involved in HCO3− secretion and that this HCO3− conductance is likely CFTR.11,12 On the basolateral surface of the duct cell are a Na+,H+ antiport and a Na+,HCO3− cotransport, as well as ATPases (Na+,K+-ATPase, H+-ATPase) and K+ channels. In combination, these transporters facilitate HCO3− secretion at the apical surface as well as maintain intracellular pH.14 Na+ and water are secreted into the ductal system to counter the electrical and osmotic forces resulting from the HCO3− secretion.

Chapter 56  Pancreatic Secretion Cl– HCO3–

HCO3– Cl–

Na+

Na+

Cl–

Ca2+

CFTR

H+ + HCO3–

cAMP Carbonic anhydrase

H2CO3 H2O

Na+ H+

Ca2+ K+

3Na+

cAMP

H+

CO2 HCO3–

Na+

2K+

Figure 56-4.  Ion transport by the pancreatic duct cell. This figure illustrates ion transport responsible for NaHCO3 secretion by the pancreatic duct cell. HCO3− is delivered for ultimate secretion by two mechanisms. In one, membrane-diffusible CO2 is catalytically converted to HCO3− and H+ by the action of carbonic anhydrase (CA), which hydrates CO2, thereby forming H2CO3, which then dissociates to HCO3− and H+. The duct cell is rich in CA. The other is Na+,HCO3− cotransport on the basolateral membrane. The HCO3− is then available for apical secretion by both the cystic fibrosis transmembrane conductance regulator (CFTR) and Cl−-HCO3− anion exchange. Na+ and H2O are delivered to the lumen through intercellular junctions. H+ is removed from the cell by a basolateral Na+,H+ antiport and a H+-ATPase to maintain a constant intracellular pH. Secretion is activated by increased permeability of apical Cl− and HCO3− channels and basolateral K+ channels through agonists (i.e., secretin and acetylcholine) that increase cellular cyclic AMP (cAMP) and calcium concentrations. (Adapted from Gorelick F, Pandol SJ, Topazian M. Pancreatic physiology, pathophy­siology, acute and chronic pancreatitis. Gastrointestinal Teaching Project, American Gastroenterological Association, 2003.)

Table 56-1  Pancreatic Acinar Cell Secretory Products Proenzymes*

Enzymes

Cationic trypsinogen Anionic trypsinogen Mesotrypsinogen Chymotrypsinogen (A, B) Kallireinogen Procarboxypeptidase A (1, 2) Procarboxypeptidase B (1, 2) Prophospholipase Proelastase Amylase Carboxylesterase Sterol esterase Lipase DNase RNase

*Proenzymes listed are stored in the pancreas and secreted into the duodenal lumen as inactive proenzyme forms. If these enzymes were active in the pancreas, they would digest the pancreatic gland. Other enzymes, such as amylase and lipase, are stored and secreted in their active forms. Adapted from Gorelick F, Pandol SJ, Topazian M. Pancreatic physiology, pathophysiology, acute and chronic pancreatitis. Gastrointestinal Teaching Project, American Gastroenterological Association, 2003.

ORGANIC CONSTITUENTS 15-17

The human pancreas has a large capacity for synthesizing protein (mostly digestive enzymes). Table 56-1 lists the major proteolytic, amylolytic, lipolytic, and nuclease digestive enzymes. Some of the enzymes are present in more than one form (e.g., cationic trypsinogen, anionic trypsinogen, and mesotrypsinogen). Enzymes that could digest the pancreas are stored in the pancreas and secreted into the pancreatic duct as inactive precursor forms. As illustrated in Figure 56-5, activation of these enzymes takes place in the intestinal lumen, where a brush-border glycoprotein peptidase, enterokinase, activates trypsinogen by removing

(by hydrolysis) an N-terminal hexapeptide fragment of the molecule (Val-Asp-Asp-Asp-Asp-Lys).17 The active form, trypsin, then catalyzes the activation of the other inactive proenzymes. In addition to the digestive enzymes, the acinar cell secretes a trypsin inhibitor, pancreatic secretory trypsin inhibitor (PSTI). This 56–amino acid peptide inactivates trypsin by forming a relatively stable complex with the enzyme near its catalytic site.18 The function of the inhibitor is to inactivate trypsins that are formed autocatalytically in the pancreas or pancreatic juice, thus preventing disorders such as pancreatitis.19

FUNCTIONS OF THE MAJOR DIGESTIVE ENZYMES AMYLASE

Human amylase is secreted by both the pancreas and salivary glands. These enzymes digest starch and glycogen in the diet. Human salivary and pancreatic amylases have identical enzyme activities. However, they differ in molecular weight, carbohydrate content, and electrophoretic mobility.20 Salivary amylase initiates digestion in the mouth and may account for a significant portion of starch and glycogen digestion because it is transported with the meal into the stomach and small intestine, where it continues to have activity. In the stomach, the amylase activity is protected from secreted gastric acid by buffering from the meal and by the protected alkaline environment of salivary and gastric mucus. The action of both salivary and pancreatic amylase is to hydrolyze 1,4-glycoside linkages at every other junction between carbon 1 and oxygen. The products of amylase digestion are maltose and maltotriose (two- and three-α-1,4linked molecules, respectively) and α-dextrins containing 1,6-glycosidic linkages, because 1,6-glycosidic linkages in

923

924

Section VII  Pancreas ides, diglycerides, and monoglycerides. Bile salts are also important for the full activity of this enzyme.24

PROTEASES

Trypsinogen En te ro kin Zymogens as e Trypsin Active Enterocytes enzymes Duodenum

Pancreas

Figure 56-5.  Site of zymogen activation. Trypsinogen, chymotrypsinogen, proelastase, procarboxypeptidase, and prophospholipase A2 are stored in the pancreas and secreted into the duodenal lumen as inactive proenzyme forms. Other enzymes, such as amylase and lipase, are stored and secreted in their active forms. The active forms of these latter enzymes have no effect on the pancreatic gland because it does not contain starch or triglyceride. Activation of the inactive proenzymes takes place in the duodenal lumen. There, the brush-border enzyme, enterokinase, converts secreted trypsinogen into trypsin. Trypsinogen and the other proenzymes are then converted to active forms by proteolytic cleavage by trypsin. (Adapted from Gorelick F, Pandol SJ, Topazian M. Pancreatic physiology, pathophysiology, acute and chronic pancreatitis. Gastrointestinal Teaching Project, American Gastroenterological Association, 2003.)

starch cannot be hydrolyzed by amylase. The brush-border enzymes complete hydrolysis of the products of amylase digestion to glucose. The final product, glucose, is transported across the intestinal absorptive epithelial cell by a Na+-coupled transport (see Chapter 100).21,22

LIPASES

The pancreas secretes three lipases: lipase (or triglyceride lipase), phospholipase A2, and carboxylesterase (see Table 56-1). The most important source of these lipases, in contrast to amylase, is the pancreas. Salivary (lingual) and gastric lipases also contribute to fat digestion but in a minor fashion (see Chapter 49). Pancreatic lipase hydrolyzes a triglyceride molecule to two fatty acid molecules released from carbons 1 and 3 and a monoglyceride with a fatty acid esterified to glycerol at carbon 2.16,23 Lipase binds to the oil-water interface of the triglyceride oil droplet, where it acts to hydrolyze the triglyceride. Bile acids as well as colipase are important for the full activity of lipase. Bile acids aid in the emulsification of triglyceride to enlarge the surface area for lipase to act on, and they form micelles with fatty acids and monoglyceride, which in turn remove these products from the oilwater interface. Colipase is believed to form a complex with lipase and bile salts. This ternary complex anchors lipase and allows it to act in a more hydrophilic environment on the hydrophobic surface of the oil droplet. Phospholipase A2 catalyzes the hydrolysis of the fatty acid ester linkage at carbon 2 of phosphatidylcholine.16 This cleavage leads to the formation of free fatty acid and lysophosphatidylcholine. Carboxylesterase has a broad specificity and will cleave cholesterol esters, lipid-soluble vitamin esters, triglycer-

The pancreas secretes a variety of proteases that are activated in the duodenum by trypsin. Trypsin, chymotrypsin, and elastase are endopeptidases that cleave specific peptide bonds adjacent to specific amino acids. Also contained in pancreatic juice are the carboxypeptidases, exopeptidases that cleave peptide bonds at the carboxyl terminus of proteins. The combined actions of the pancreatic proteases and pepsin from the stomach result in the formation of oligopeptides and free amino acids. The oligopeptides are further digested by brush-border enzymes (see Chapter 100). Free amino acids and oligopeptides are transported across the intestinal mucosa by a group of Na+- and H+-coupled transporters.25 It is interesting that only certain amino acids (mostly essential amino acids) can be measured in the lumen during digestion, indicating that the combined action of the proteases is not random and that the products result from the combined specificities of the individual proteases. These amino acids have greater effects on stimulating pancreatic secretion, inhibiting gastric emptying, regulating small bowel motility, and causing satiety. Thus, the specific pattern of protease actions leads to physiologic regulation of several organs in the gastrointestinal tract.

DIGESTIVE ENZYME SYNTHESIS AND TRANSPORT Synthesis of digestive enzymes takes place in the internal space of the rough endoplasmic reticulum (RER) (see Fig. 56-2, left).2,26 The mechanism for translation of the cell’s messenger RNA (mRNA) into exportable protein is explained by the signal hypothesis. The main feature of the hypothesis is that ribosomal subunits attach to mRNA and initiate synthesis of a hydrophobic “signal” sequence on the NH2terminal of nascent proteins. This complex then attaches to the outer surface of the endoplasmic reticulum, and the signal sequence targets the protein being synthesized into the lumen of the RER. Newly synthesized proteins can undergo modifications in the endoplasmic reticulum, including disulfide bridge formation, phosphorylation, sulfation, and glycosylation. Conformational changes resulting in tertiary and quaternary structures of the protein also take place in the endoplasmic reticulum. Processed proteins in the RER are transported to the Golgi complex where further post-translational modification (glycosylation) and concentration occur.27 Important conformational specificities are required for appropriate function and transport of proteins between compartments. These conformational specificities are facilitated by proteins called chaperones and foldases among others. When there is an increased demand on the RER for protein synthesis or when there is a cell stress that causes unfolding of proteins, the RER responds by increasing synthesis of chaperones and foldases, ceases synthesis of the proteins, and even degrades unfolded proteins using the mechanisms of the unfolded protein response (UPR).28 By keeping proteins in their appropriate conformation, the UPR maintains normal transport of proteins between compartments of the cell. The Golgi complex also serves the important function of sorting and targeting newly synthesized proteins into various cell compartments. Digestive enzymes are trans-

Chapter 56  Pancreatic Secretion ported to the zymogen granules.27 Lysosomal hydrolases are sorted to the lysosome.29 For this lysosomal pathway, mannose-6-phosphate groups are added to oligosaccharide chains on the protein during its presence in the cis-Golgi complex. The mannose-6-phosphate groups serve as a recognition site for a specific receptor. The interaction of the lysosomal enzyme mannose 6-phosphate with its receptor leads to formation of vesicles that transport this complex to the lysosome, delivering the enzyme. In the lysosome the enzyme dissociates from the receptor, which in turn cycles back to the Golgi complex. Secretion of the digestive enzymes occurs by exocytosis. Exocytosis consists of movement of the secretory granule to the apical surface, the recognition of a plasma membrane site for fusion, and the fission of the granule membraneplasma membrane site after fusion.2,30 Recent studies demonstrate roles for actin-myosin, SNARE (soluble Nethylmaleimide–sensitive factor attachment protein [SNAP] receptor) proteins, and guanosine triphosphate (GTP)binding proteins in these processes.30-38 Intracellular signals generated by agonist receptors as discussed following interact with these entities to mediate digestive enzyme secretion via exocytosis of zymogen granules.

REGULATION OF PROTEIN SYNTHESIS

The mechanisms involved in regulating expression of digestive enzymes in the exocrine pancreas have been partially elucidated. The investigations have addressed the following two questions: First, what accounts for the specific expression of digestive enzymes in the pancreas? Second, how do alterations in dietary nutrients change the synthesis of specific digestive enzymes? Genes for digestive enzymes such as amylase, chymotrypsin, and elastase contain enhancer regions in their 5′ flanking nucleotide sequences that regulate the transcription of their mRNAs, termed the pancreas consensus element (PCE).39,40 A transcription factor, pancreas transcription factor-1 (PTF-1), is present selectively in the exocrine pancreas, binds to this region, and is essential for expression of these digestive enzymes.41-44 Thus, PTF-1 represents at least one of the differentiation-regulated factors that accounts for digestive enzyme expression in the pancreas. Numerous studies have demonstrated that the relative synthesis rates of specific digestive enzymes change as a function of dietary intake. For example, a carbohydrate-rich diet results in an increase in synthesis of amylase and a decrease in that of chymotrypsinogen45; a lipid-rich diet enhances lipase expression46; and an alcohol-rich diet decreases amylase expression.47 The mechanisms responsible for this adaptation are only partially understood. The regulation occurs at the level of gene transcription in many of these conditions.47 Several studies have also demonstrated that amylase gene expression is regulated by insulin and diet.45

CELLULAR REGULATION OF ENZYME SECRETION The mechanism of neurohumoral stimulation of the acinar cell has been demonstrated with the use of in vitro preparations of dispersed acinar cells and acini from small animals. Studies involving the use of human tissue are few. With the use of radiolabeled ligands and specific antagonists, receptors for cholecystokinin (CCK), acetylcholine, gastrin-releasing peptide (GRP), substance P, vasoactive intestinal peptide (VIP), and secretin have been identified

Exocytosis

↑Ca2+ ↑cAMP

GRP VIP Secretin Figure 56-6.  Receptor-mediated secretion. Pancreatic acinar cell agonists that stimulate digestive enzyme secretion act through two separate pathways. In one pathway, agonists such as gastrin-releasing peptide (GRP), cholecystokinin (CCK), and acetylcholine (Ach) mediate secretion through increases in cellular calcium (Ca2+). In the other pathway, agonists such as vasoactive intestinal polypeptide (VIP) and secretin mediate secretion through increases in cellular cyclic AMP (cAMP). Of note, simultaneous increases in both cellular calcium and cAMP after stimulation with a combination of agonists have a synergistic effect on secretion. That is, the observed response is greater than would be expected from the additive responses of the individual agonists acting alone. (Adapted from Gorelick F, Pandol SJ, Topazian M. Pancreatic physiology, pathophysiology, acute and chronic pancreatitis. Gastrointestinal Teaching Project, American Gastroenterological Association, 2003.) CCK

Ach

in preparations from several species.48-50 Furthermore, the molecular structure for each of these receptor types has been elucidated from cloning and sequencing. Each is a G-protein–coupled receptor with seven hydrophobic domains believed to be membrane-spanning segments. The receptors are on the basolateral plasma membrane of the acinar cell. Receptors on acinar cells have been divided into two categories according to the mode of stimulus-secretion coupling (Fig. 56-6). In one category are VIP and secretin. The interaction of these agents with acinar cells leads to activation of adenylate cyclase and a rise in cellular cAMP, which in turn activates enzyme secretion through cAMPdependent protein kinase A.3,48 In the other category are acetylcholine, GRP, substance P, and CCK. The actions of these agonists include stimulating cellular metabolism of membrane phosphoinositides and raising intracellular free calcium concentrations ([Ca2+]i) from mobilization of intracellular stores.3,48 Specifically, the agonist-receptor interaction leads to a phospholipase C-mediated hydrolysis of phosphatidylinositol 4,5-bisphosphate to 1,2-diacylglycerol and inositol 1,4,5-triphosphate (IP3). IP3, in turn, releases calcium from endoplasmic reticulum stores.51 The calcium release into the cytosol causes a rapid rise in the concentration of free calcium that is necessary for the secretory response.52 Calcium release into the cytosol is also mediated by ryanodine receptors and signals interacting with the ryanodine receptor, such as calcium and fatty acidcoenzyme A (CoA) esters.53 Other molecules involved in intracellular calcium release are cyclic adenosine diphosphate (ADP)–ribose and nicotinic acid adenine dinucleotide phosphate (NAADP).54 The mechanism by which increases in [Ca2+]i mediate secretion is not established but involves calmodulindependent protein kinases and actin-myosin interactions, SNARE proteins, and GTP-binding proteins as discussed

925

926

Section VII  Pancreas earlier.3 The continued stimulation of enzyme secretion by these agents also depends on the influx of extracellular calcium.52 This influx is mediated by changes in nitric oxide and cyclic guanosine monophosphate (cGMP). Components of the influx channel have been determined.55 The intracellular mechanism of enzyme secretion may also be regulated by 1,2-diacylglycerol and protein kinase C3,56 as well as arachidonic acid.57 Specific phosphorylations and dephosphorylations of cellular proteins also occur with cAMP agonists and calcium-phosphoinositide agonists.3 The exact roles of these events in secretion are not established. The enzyme secretory response of the acinar cell to a combination of an agonist that acts through cAMP and an agonist that acts through changes in calcium is greater than the sum of the two individual responses. An example of such a combination is VIP or secretin with acetylcholine. The exact mechanism of this potentiated response is not known, but it probably functions physiologically so that significant quantities of secretion occur with a combination of small increases in individual agonists.

ORGAN PHYSIOLOGY Human exocrine pancreatic secretion occurs during the fasting (interdigestive) state and after ingestion of a meal (digestive). The interdigestive pattern of secretion begins when the upper gastrointestinal tract is cleared of food. In an individual who eats three meals per day, the digestive pattern begins after breakfast and continues until late in the day, after the evening meal is cleared from the upper gastrointestinal tract.

INTERDIGESTIVE SECRETION

The interdigestive pancreatic secretory pattern is cyclic and follows the pattern of the migrating myoelectric complex (MMC).58 The patterns recur every 60 to 120 minutes, bursts of enzyme secretion being temporally associated with the periods of increased motor activity in the stomach and duodenum (i.e., phases II and III). In addition to pancreatic enzyme secretion is increased secretion of bicarbonate and bile (secondary to partial gallbladder contraction) into the duodenum during phases II and III of the MMC (see Chapter 97). The underlying mechanism involves the cholinergic nervous system.59 Motilin and pancreatic polypeptide are also involved in regulation of the MMC.60,61 The pancreatic secretion during the interdigestive phase integral to the “housekeeping” function of the MMC.

DIGESTIVE SECRETION

Like gastric secretion, exocrine pancreatic secretion with ingestion of a meal is divided into three phases: cephalic, gastric, and intestinal. The vagal nerves mediate the cephalic phase of the exocrine secretion. The extent of cephalic stimulation of exocrine pancreatic secretion in humans has been evaluated through measurement of exocrine secretions stimulated by sham feeding (chewing and spitting out the food). One study indicated that sham feeding stimulated pancreatic enzyme secretion at up to 50% of the maximal secretory rate, with no increase in bicarbonate secretion when gastric secretions were prevented from entering the duodenum.62 When gastric secretions were allowed entry into the duodenum, the rate of pancreatic enzyme secretion rose to about 90% of maximal, and bicarbonate was also secreted. These results suggest that cephalic stimulation specifically stimulates

acinar secretion and that a low pH in the duodenum (from gastric acid) augments this secretion as well as causes bicarbonate secretion. Results of investigations of the mechanism of neuro­ transmission during cephalic stimulation are controversial. Acetylcholine is certainly a major neurotransmitter involved because cholinergic antagonists greatly reduce and in some cases abolish sham feeding–stimulated pancreatic secretion in humans.63 Nerve endings containing the peptides VIP, GRP, CCK, and enkephalins have been identified in the pancreas. Data supporting the role of these peptides in the cephalic phase of secretion are strongest for VIP and GRP.64,65 Both are released into the pancreatic venous effluent with vagal stimulation in animals. Furthermore, as discussed, acinar cells have receptors for GRP and VIP that mediate enzyme secretion. The ductal epithelium also responds to VIP with the secretion of water and bicarbonate.64 The gastric phase of pancreatic secretion results from meal stimuli acting in the stomach. The major stimulus is gastric distention, which causes predominantly secretion of enzymes with little secretion of water and bicarbonate. Balloon distention of either the gastric fundus or the antrum results in a low-volume, enzyme-rich secretion by way of a gastropancreatic vagovagal reflex.66 When gastric juice and contents of a meal enter the duodenum, a variety of intraluminal stimulants can act on the intestinal mucosa to stimulate pancreatic secretion through neural and humoral mechanisms. Three gastric processes— secretion of acid, pepsin, and lipase; digestion; and emptying—are tightly coupled to the mechanisms of the intestinal phase of pancreatic secretion. The intestinal phase begins when chyme first enters the small intestine from the stomach. It is mediated by hormones and enteropancreatic vagovagal reflexes. The major mediator of hydrogen ion–stimulated bicarbonate and water secretion is secretin. Secretin is released from the duodenal mucosa by gastric acid, with a pH threshold of 4.5.67 The quantity of secretin released as well as the volume of pancreatic secretion depend on the load of titratable acid delivered to the duodenum.67,68 Immunoneutralization of secretin with specific antisecretin antibody decreases meal-stimulated pancreatic volume and bicarbonate secretion by as much as 80%.69 The antisecretin antibody also inhibits meal-stimulated enzyme secretion by as much as 50%, suggesting that secretin also has a role in enzyme secretion, possibly by potentiating the action of agonists such as acetylcholine. If exogenous secretin is infused to reproduce the plasma concentrations of secretin during a meal, the pancreatic bicarbonate output is less than the bicarbonate output observed with a meal. The bicarbonate response to secretin also depends on cholinergic input because atropine partially inhibits the response stimulated by exogenous secretin.70 Thus, the complete meal-stimulated response results from a combination of mediators. During the intestinal phase, secretion of digestive enzymes is mediated by intraluminal fatty acids more than eight carbons in length, monoglycerides of these fatty acids, peptides, amino acids, and, to a small extent, glucose. The most potent amino acids for stimulating secretion in humans are phenylalanine, valine, methionine, and tryptophan.71 The response to peptides and amino acids is related to the total load perfused into the intestine rather than the concentration.72 The mediators of the enzyme secretory response from intestinal stimuli are neural and humoral (Fig. 56-7). Truncal vagotomy and atropine markedly inhibit the enzyme (and bicarbonate) responses to low intestinal loads of amino

Chapter 56  Pancreatic Secretion Dorsal vagal complex

Vagal efferent Food

Ach GRP VIP

I cell CCK

Vagal afferent

Enzymes Peptides, amino acids, fatty acids Circulation Figure 56-7.  Cholecystokinin (CCK) stimulates pancreatic enzyme secretion by neural and hormonal pathways. This diagram illustrates the several pathways mediating meal-stimulated pancreatic secretion that involve CCK. First, meal nutrients such as peptides, amino acids, and fatty acids delivered into the duodenum stimulate the local release of CCK from the CCK-containing I cell to the area around the basolateral surface of the I cell. The released CCK can activate vagal afferent neurons that transmit the signal to the dorsal vagal complex, where the sensory information is integrated and vagal efferents are activated. Vagal efferents synapse with neurons in the pancreatic ganglia. In turn, via the neurotransmitters acetylcholine (Ach), gastrin-releasing peptide (GRP), and vasoactive intestinal polypeptide (VIP), effector neurons arising in the pancreatic ganglia activate secretion by pancreatic parenchymal cells. In addition to activating the neural pathway, CCK released by the I cell enters the general circulation and may act as a hormone on the pancreatic acinar cells to cause secretion. (Adapted from Gorelick F, Pandol SJ, Topazian M. Pancreatic physiology, pathophysiology, acute and chronic pancreatitis. Gastrointestinal Teaching Project, American Gastroenterological Association, 2003.)

acids and fatty acids as well as infusion of physiologic concentrations of CCK.73-75 These results indicate that vagovagal reflexes mediate enzyme secretion and augment bicarbonate secretion stimulated by secretin. CCK is the major humoral mediator of meal-stimulated enzyme secretion during the intestinal phase. The circulating concentration of CCK is found to rise with a meal76 and the main circulating form is 58 amino acids in size (CCK58).77 CCK is released from the upper small intestinal mucosa by digestion products of fat and protein and, to a lesser extent, by starch digestion products.76 Further experiments indicate that CCK activates afferent neurons in the duodenal mucosa.74,75 These afferent neurons activate a vagovagal reflex that causes pancreatic enzyme secretion, as illustrated in Figure 56-7. The elements of the secretory mechanism during the intestinal phase of a meal previously described are activated when standard nutrients enter the duodenum. Investigations have been performed to determine whether administration of alternative nutrients or delivery of nutrients to the jejunum rather than the duodenum may result in less activation of the pancreatic secretory response. Such information is critical for strategies to provide nutrients to patients with pancreatic disorders such as acute or chronic pancreatitis. With acute pancreatitis, stimulation of the pancreas can exacerbate the severity of the disease; in chronic pancreatitis, stimulation of the pancreas can exacerbate pain (see Chapter 58 and 59). The research in this area has delineated effects of different types of nutrients as well as location of administration on the pancreatic secretory response. Studies in normal human volunteers compared the effect of equicaloric amounts of long-chain triglycerides and medium-chain triglycerides, infused into the jejunum, on plasma CCK levels, pancreatic secretion, and gallbladder contraction.78 Medium-chain triglycerides led to none of these responses, but long-chain

triglycerides stimulated each of the responses. Thus, with respect to jejunal administration, medium-chain triglycerides can be used to provide an energy source without significantly stimulating the pancreas. Another study in healthy volunteers compared the effects of duodenal infusions of a complex liquid diet with those of an elemental diet with low fat content.79 Noteworthy was the finding of significantly less pancreatic secretion with the elemental diet than with the complex liquid diet. In the clinical setting, the results of these studies indicate that the type of nutrient administered to the intestinal lumen is the most important factor in regulating the pancreatic secretory response.

FEEDBACK REGULATION

In animals as well as humans, diversion of pancreatic juice from the intestine results in augmented pancreatic secretion.80 The augmented enzyme secretion is mediated by a rise in circulating CCK.81 Both the increase in CCK and enzyme secretion following diversion of pancreatic juice can be inhibited by replenishment of intraluminal trypsin and other digestive enzymes.82 It is hypothesized that during a meal, when trypsin is occupied by meal proteins, pancreatic secretion is enhanced because trypsin is not available to the surface of the intestinal epithelial cells to cause feedback inhibition. After the meal, trypsin is free and inhibits intestinal CCK release and pancreatic enzyme secretion. Intraluminal CCK-releasing factors that mediate this effect of trypsin have been described. One is a protein, called monitor peptide, that is secreted by the pancreas.83 Another is called luminal CCK-releasing factor (LCRF).84 Both monitor peptide, and LCRF causes CCK release from the enteroendocrine CCK cell (I cell) into the blood. These releasing factors are likely mediators of the physiologic feedback mechanism for enzyme secretion. It is believed that the effects of trypsin described earlier are due to deg-

927

928

Section VII  Pancreas radation of the CCK-releasing factors by trypsin when trypsin is not bound to meal proteins. A similar feedback mechanism involving a secretinreleasing factor (an intraluminal peptide) has been described that regulates release of secretin from the secretincontaining enteroendocrine cell of the intestine.85,86

PANCREATIC SECRETORY FUNCTION TESTS Various tests have been devised to measure the secretory function of the pancreas in order to diagnose disorders such as chronic pancreatitis and pancreatic cancer (Table 56-2).87-100 The function tests fall into two general categories: direct and indirect. Direct tests of pancreatic secretory function involve collection of pancreatic secretions after intravenous administration of a secretagogue or a combination of secretagogues. Indirect tests of pancreatic secretory function include the measurement of pancreatic enzymes in duodenal samples after nutrient ingestion; the measurement of products of digestive enzyme action on ingested substrates; the measurement of pancreatic enzymes in the stool; and the measurement of the plasma concentration of hormones or other markers that are altered in pancreatic insufficiency states. Which pancreatic function test should be used depends on the clinical question and the characteristics and availability of the test. The exocrine pancreas has a very large functional reserve. Maldigestion and malabsorption do not

occur until the functional capacity as measured by CCKstimulated digestive enzyme secretion is reduced to 5% to 10% of normal.71,87 Thus, many tests relying on the conversion of an ingested substrate by digestive enzymes to a measurable product will be insensitive in pancreatic disease unless moderate to severe pancreatic insufficiency is present. Therefore, the measurement of duodenal digestive enzymes after the intravenous administration of pancreatic secretagogues provides the greatest sensitivity and specificity. The major drawbacks to the direct tests are the requirements for duodenal intubation and the fact that very few centers are proficient in performing the studies pro­ perly. Improved imaging techniques for diagnosing pancreatic disease have greatly decreased the use of the tests. On certain occasions, however, pancreatic function tests are necessary for diagnosing pancreatic disease. More recently, a direct test using endoscopy and rapid collection has been described.87,92-95 The pancreatic function tests are described here to allow the reader to appreciate the relative diagnostic usefulness of each test.

DIRECT TESTS

Direct tests provide a gold standard for measurement of pancreatic function. Stimulation of secretion has been described most commonly with secretin, CCK, or the two combined. The combination provides the complete information about acinar and ductular cell secretions. In fact, it is accepted that the direct tests are more sensitive than current imaging techniques in the diagnosis of early pancreatic disease.87 In the classic description, the stomach and

Table 56-2  Pancreatic Secretory Function Tests TEST

DESCRIPTION

Direct Tests Secretin

Measurements of volume and HCO3− secretion into the duodenum after IV secretin Cholecystokinin Measurements of duodenal outputs of amylase, trypsin, chymotrypsin, and lipase after IV cholecystokinin Secretin and Measurements of volume, cholecystokinin HCO3−, and enzymes after IV secretin and cholecystokinin } Indirect Test Requiring Duodenal Intubation Lundh test meal Measurement of duodenal trypsin concentration after oral ingestion of a test meal

Tubeless Indirect Tests Fecal fat Measurement of fat in the stool after ingesting meals with a known amount of fat Fecal chymotrypsin} Fecal elastase 1

6

Measurement of chymotrypsin or elastase 1 in the stool

6

Oral ingestion of NBT-PABA or fluorescein dilaurate with a meal, followed by measurements of PABA or fluorescein in serum or urine

NBT-PABA} Fluorescein dilaurate

IV, intravenous; NBT-PABA, N-benzoyl-l-tyrosyl-p-aminobenzoic acid.

6

ADVANTAGES

DISADVANTAGES

CLINICAL INDICATIONS

Provide the most sensitive and specific measurements of exocrine pancreatic function

Require duodenal intubation and intravenous administration of hormones; not widely available

Detection of mild, moderate, or severe exocrine pancreatic dysfunction

Does not require IV administration of hormones

Requires duodenal intubation, a test meal, and normal anatomy, including small intestinal mucosa; not widely available

Detection of moderate or severe exocrine pancreatic dysfunction when a direct test cannot be done (i.e., limited availability of test)

Provides a quantitative measurement of steatorrhea

Requires sufficient dietary fat intake and collection of stool; only detects severe pancreatic dysfunction Insensitive for detecting mild or moderate dysfunction

Detection of severe exocrine pancreatic dysfunction and steatorrhea

Do not detect mild or moderate dysfunction; results may be abnormal in patients with small intestinal mucosal disease

Detection of severe exocrine pancreatic dysfunction

Do not require IVs, tubes, or administration of oral substrates Provide simple measurements for severe pancreatic dysfunction

Detection of severe exocrine pancreatic dysfunction

Chapter 56  Pancreatic Secretion duodenum are intubated. The gastric intubation is required to remove gastric secretions that would interfere with the ability to measure water and bicarbonate secretion from the pancreas. Low pH may also alter pancreatic enzyme activity. The duodenal tube is used for infusion of a nonabsorbable marker and collection of pancreatic secretions. The use of a nonabsorbable marker such as cobalamin or polyethylene glycol (PEG) allows the quantitation of secretions without the need for complete aspiration of secretions.88 The direct function tests are based on the principle that maximal water, bicarbonate, and enzyme secretion are related to the functional mass of the pancreas.89 Historically, the secretin test (intravenous administration of secretin, with volume and bicarbonate measurement) first provided information about the function of the pancreas in various clinical settings. Administration of CCK and the mea­ surement of digestive enzyme secretion also have been used successfully to demonstrate pancreatic insufficiency. Because the combination of secretin and CCK admini­ stration provides stimulation of both functional units of the exocrine pancreas, this combination is most commonly used.88-91 CCK is best delivered by constant intravenous infusion. The dose for synthetic CCK-octapeptide (carboxyterminal octapeptide of CCK; Squibb, Princeton, NJ) is 40 ng/kg/hr. The dose of synthetic secretin (SecroFlo, Repligen, Needham, Mass) is 0.2 µg/kg injected over 1 minute. Measurements, corrected for percentage recovery of a nonabsorbable marker, are made for fluid and for bicarbonate and protein concentrations, and the activity of digestive enzymes during collections over 15-minute periods for 1 hour. Amylase, trypsin, chymotrypsin, and lipase are the digestive enzymes most commonly measured.88,90 In a study of 236 subjects undergoing continuous infusion of CCK-octapeptide (40 ng/kg/hr) and purified natural secretin (0.25 clinical unit [CU]/kg/hr), measurements of chymotrypsin concentration and peak bicarbonate output were used to distinguish subjects with chronic pancreatitis from those without organic disease.91 The test was 83% sensitive and 89% specific. False-positive results may occur in patients with celiac disease and diabetes mellitus. An adaptation of the direct secretory test to upper endoscopy also has been described.92-95 At the time of endoscopy, either secretin, CCK, or the combination is administered intravenously, and pancreatic secretions are collected via the endoscope positioned in the duodenum.

INDIRECT TESTS Lundh Test Meal96

The subject ingests a 300-mL liquid test meal composed of dried milk, vegetable oil, and dextrose (6% fat, 5% protein, and 15% carbohydrate). After ingestion of this meal, samples are aspirated from the intubated duodenum at intervals for measurement of digestive enzyme concentration. Usually only trypsin activity is measured; however, the additional determination of lipase or amylase may improve test sensitivity. The test is not valid in patients with mucosal disease (e.g., celiac disease) or altered gastroduodenal anatomy (e.g., following a vagotomy and drainage procedure or a Billroth II gastrectomy). Comparisons of the Lundh test meal with the secretin-CCK test show that the latter is more sensitive in detecting mild forms of pancreatic disease, whereas the tests are comparable in more advanced disease.88

Measurement of Fecal Fat

Steatorrhea occurs when stimulated lipase output drops to less than 5% to 10% of normal.71 Thus, measurement of fat

in the stools collected for 72 hours in a subject ingesting a diet adequate in fat intake (70 to 100 g/day) is considered an effective means of diagnosing steatorrhea. Normally 7% or less of ingested fat appears in the stool. A simple qualitative microscopic examination of a single stool for oil is almost as sensitive as quantitative measurements for fat.88 Because steatorrhea occurs only with advanced pancreatic disease, measurement of fecal fat is not useful in the diagnosis of mild or moderate disease, nor is steatorrhea specific for pancreatic disease (see Chapter 101).

Measurement of Fecal Chymotrypsin and Elastase 1

Chymotrypsin measurements in the stool have been used as indirect tests of pancreatic function for many years,97 especially to establish pancreatic insufficiency in patients with cystic fibrosis.88 These measurements are about 85% sensitive in advanced pancreatic dysfunction and relatively insensitive in mild to moderate disease. Pancreatic elastase 1 can be measured in the stool using a monoclonal antibody against human elastase 1. This test, like other indirect tests, is not sensitive in detecting mild to moderate pancreatic disease.98,99

Other Tests

The bentiromide test is an indirect tubeless test in which the synthetic peptide N-benzoyl-l-tyrosyl-p-aminobenzoic acid (NBT-PABA) is specifically cleaved by chymotrypsin to NBT (N-benzoyl-l-tyrosyl) and para-aminobenzoic acid (PABA). PABA is then absorbed in the intestine, conjugated in the liver, and excreted in the urine.88 The PABA metabolite can be measured in either serum or urine. Prior gastric surgery, small bowel disease, liver disease, and renal insufficiency may interfere with the measurements. So may the use of several drugs (acetaminophen, phenacetin, chloramphenicol, benzocaine, lidocaine, procaine, sulfonamides, sulfonylureas, and thiazides) and prior ingestion of certain foods (prunes and cranberries). A broad range of sensitivities has been reported for the NBT-PABA test.88 In patients with severe pancreatic insufficiency and malabsorption, the sensitivity is 80% to 90%. In those with mild to moderate impairment, test sensitivity is as low as 40%. In order to improve test specificity, several modifications of the NBT-PABA test have been devised. Administering free PABA on a separate day or giving carbon-14 (14C)–labeled PABA simultaneously with NBTPABA may identify patients in whom abnormal NBT-PABA test results are caused by mucosal disease of the small bowel.88 Another test similar to the NBT-PABA test is the fluorescein dilaurate (pancreolauryl) test. The principle underlying this test is the same as that for the NBT-PABA test. Fluorescein dilaurate is an ester, poorly soluble in water, that is hydrolyzed by pancreatic carboxylesterase into lauric acid and free water-soluble fluorescein. The fluorescein is readily absorbed into the intestine, partly conjugated in the liver, and excreted in the urine. Fluorescein dilaurate is given in the middle of a breakfast meal.88 Urine is collected for 10 hours after breakfast, and the fluorescein excreted in the urine is measured. Like the PABA metabolite, fluorescein can also be measured in the serum. To evaluate the subject’s absorption, conjugation, and excretion, the test is repeated two to three days later with free fluorescein. The recovery rate on both days is expressed as a ratio. Like the NBT-PABA test, the pancreolauryl test is highly sensitive and specific for advanced pancreatic disease and less so for mild or moderate disease. Several other indirect tubeless tests have been created in an effort to improve sensitivity for identifying milder forms

929

930

Section VII  Pancreas of exocrine pancreatic dysfunction. They include triglyceride and cholesteryl breath tests; H2 and CO2 breath tests; the dual-label Schilling test; and plasma measurements of pancreatic polypeptide and amino acids.88,100 However, none of these tests has been shown to have greater sensitivity than the indirect tubeless tests described previously. In addition, many of them require radioactive isotopes or expensive equipment, making their usefulness less desirable. The profusion of tests of pancreatic function suggests that the ideal test has not been developed. The direct tests remain the gold standard and should be used to detect mild and moderate pancreatic inefficiency. For detection of severe insufficiency resulting in steatorrhea, many of the tests described have adequate sensitivity.

Petersen OH, Tepikin AV. Polarized calcium signaling in exocrine gland cells. Annu Rev Physiol 2008; 70:273-99. (Ref 54.) Stevens T, Conwell DL, Zuccaro G Jr, et al. A prospective crossover study comparing secretin-stimulated endoscopic and Dreiling tube pancreatic function testing in patients evaluated for chronic pancreatitis. Gastrointest Endosc 2008; 67:458-66. (Ref 94.) Steward MC, Ishiguro H, Case RM. Mechanisms of bicarbonate secretion in the pancreatic duct. Annu Rev Physiol 2005; 67:377-409. (Ref 10.) Whitcomb DC, Gorry MC, Preston RA, et al. Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene. Nat Genet 1996; 14:141-5. (Ref 19.) Whitcomb DC, Lowe ME. Human pancreatic digestive enzymes. Dig Dis Sci 2007; 52:1-17. (Ref 17.) Williams JA. Intracellular signaling mechanisms activated by cholecystokinin-regulating synthesis and secretion of digestive enzymes in pancreatic acinar cells. Annu Rev Physiol 2001; 63:77-97. (Ref 3.)

KEY REFERENCES

Full references for this chapter can be found on www.expertconsult.com.

O’Keefe SJ, Lee RB, Anderson FP, et al. Physiological effects of enteral and parenteral feeding on pancreaticobiliary secretion in humans. Am J Physiol Gastrointest Liver Physiol 2003; 284:G27-36. (Ref 79.)

CHAPTE R

57 Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood David C. Whitcomb and Mark E. Lowe

CHAPTER OUTLINE Definitions and Terminology  931 Models of Pancreatitis as a Complex Disorder  932 Major Susceptibility Genes Causing Pancreatic Disease  933 Cationic Trypsinogen Gene Mutations  933 Anionic Trypsinogen Gene Mutations  934 Calcium-Sensing Receptor Gene Polymorphisms  934 Cystic Fibrosis Transmembrane Conductance Regulator Gene Mutations  935 Shwachman-Bodian-Diamond Gene Mutations  937 Modifier Gene  937 Pancreatic Secretory Trypsin Inhibitor Gene Mutations  937 Pancreatic Disease in Children  937 Etiology of Acute Pancreatitis  938 Clinical Features  939 Recurrent Acute Pancreatitis  939 Clinical Aspects of Genetic Disorders Affecting the Pancreas  939 Cystic Fibrosis  940

The pancreas is protected from most environmental insults because of its retroperitoneal location, sphincter-protected duct system, and filtered blood supply. It is also generally protected from toxic injury because it does not play a major role in xenobiotic detoxification or clearance of waste products. These observations are important because they point to genetic variations as the primary factors that determine the risk of developing pancreatic disease. Insight into the etiology of pancreatic diseases required fundamental information about the human genome, and this critical information is now available. Utilization of these resources unlocked the mysteries of pancreatic disease, but also demanded the development of new disease models in order to organize, classify, and integrate this new information. It has become increasingly clear that most common pancreatic diseases have a genetic basis, either because of severe mutations in major genes or because of the convergence of common factors with low independent risks that cause a high risk interaction and together determine which

Drs. M. James Lopez and Richard J. Grand contributed to this chapter in previous editions of the textbook.

Hereditary Pancreatitis  949 Familial Pancreatitis  952 Tropical Pancreatitis  952 Shwachman-Diamond Syndrome  953 Rare Genetic Syndromes with Pancreatic Pathology  954 Johanson-Blizzard Syndrome  954 Pearson’s Marrow-Pancreas Syndrome  955 Pancreatic Agenesis  955 Agenesis of the Dorsal or Ventral Pancreas  955 Other Syndromes  955 Isolated Enzyme Defects  955 Familial Metabolic Syndromes Associated with Recurrent Acute and Chronic Pancreatitis  956 Familial Hyperparathyroidism with Hypercalcemia  956 Familial Hyperlipidemia  956

organ will be susceptible to injury (e.g., the pancreas) and the type of response to injury (e.g., fibrosis). This chapter focuses on the genetic factors contributing to the etiology of pancreatic diseases. and on childhood disorders of the pancreas because major genetic mutations are usually recognized in children, whereas disorders with a stronger environmental component (tobacco smoking, excessive alcohol consumption, or gallstones) become important later in life.

DEFINITIONS AND TERMINOLOGY There are a growing number of pancreatic disorders with a known genetic basis. However, there is no consensus terminology to distinguish disorders of different etiology that have identical end-stage pathology. Acute pancreatitis represents an event triggered by sudden pancreatic injury that is followed by sequential inflammatory responses (see Chapter 58). Chronic pancreatitis, on the other hand, is a process that usually begins with recurrent acute pancreatitis and ends with immune-mediated destruction of the pancreas and widespread glandular fibrosis (see Chapter 59).1

931

932

Section VII  Pancreas Therefore, the terms acute pancreatitis and chronic pancreatitis describe syndromes with similar clinical and pathologic characteristics caused by multiple etiologies and varying mechanistic pathways.2 Subsets of clinically defined types of chronic pancreatitis are associated with specific genetic factors. Care must be taken to correctly define what is observed, especially if there is concern of insurance or other forms of discrimination. The Genetic Information Nondiscrimination Act (GINA) was signed into law in the United States in 2008. Although the act applies specifically to Medicare, it does nonetheless provide protection from noninsurability, rate hikes, genetic risk being considered a preexisting condition, and employment discrimination based on genetic information. Genetic counseling is recommended when genetic testing is being considered. Hereditary pancreatitis refers to recurrent acute or chronic pancreatitis in an individual from a family in which the pancreatitis phenotype appears to be inherited through a disease-causing gene mutation expressed in an autosomal dominant pattern. Individuals with pancreatitis who carry a gene mutation that causes autosomal dominant pancreatitis have hereditary pancreatitis. Familial pancreatitis refers to pancreatitis from any cause that occurs in a family with an incidence that is greater than would be expected by chance alone, given the size of the family and incidence of pancreatitis within a defined population. Familial pancreatitis may or may not be caused by a genetic defect. Tropical pancreatitis (TP) is a form of early age-onset, nonalcoholic chronic pancreatitis occurring in tropical regions3 that is often clustered among family members, and that has a complex genetic basis. TP is further subdivided into fibrocalculous pancreatic diabetes (FCPD) and tropical calcific pancreatitis (TCP) based on the presenting feature of diabetes with fibrosis (FCPD) or severe pain with fibrosis and calcifications (TCP).4 The majority of children previously classified as having early onset idiopathic chronic pancreatitis have identifiable genetic etiologies, but there is no consensus terminology to distinguish the clinical diagnosis from the underlying etiology, especially when there is significant overlap. The best-characterized gene mutation–associated pancreatic disorders follow classic mendelian inheritance patterns, which are recognized as autosomal dominant (e.g., hereditary pancreatitis) or autosomal recessive (e.g., cystic fibrosis [CF]) genetic disorders. Complex trait genetics refers to inherited traits that do not follow mendelian patterns of single-gene genetics. Most cases of chronic pancreatitis in children are complex genetic disorders (see following). Complex traits or disorders, by necessity, are established when multiple factors occur together for the phenotype to be expressed, and may involve two or more genes (polygenic disorders) or gene-environment interactions. Complex disorders differ from additive genetic effects in which the genetic effects at two separate loci are equal to the sum of their individual effects. In polygenic disorders, the mutant alleles from more than one gene cause a disease in a symbiotic fashion, when neither of the mutant genes alone is disease causing. Modifier genes are not disease-causing; rather, they alter a particular aspect of the disease process or confer unique phenotypic features to a genetic disorder.

MODELS OF PANCREATITIS AS A COMPLEX DISORDER New models are necessary to understand the relationship among risk factors, etiologies, and the pathology of pancre-

atic disorders.5,6 Chronic pancreatitis, for example, can no longer be approached from an allopathic perspective (germ theory of disease in which one etiologic agent is responsible for a specific disease), but from the perspective of personalized medicine in which multiple factors must be considered in designing patient-specific treatments. For historical reasons most knowledge about chronic pancreatitis was derived from linking clinical symptoms in patients with abdominal pain and loss of pancreatic exocrine and endocrine function to pathologic changes in the pancreas observed at autopsy. The goal was to find the agent that caused chronic pancreatitis. Because of this approach the major advances in the late 20th century were technical in nature, allowing physicians to obtain biopsies (or highresolution images) without an autopsies.7 Therapeutic interventions were directed at replacing lost function and attempting to control pain. However, the factors that determine which patient with known environmental risk factors (e.g., alcoholism and tobacco smoking) would progress to chronic pancreatitis whereas others did not, or why other patients without identifiable environmental risk factors progressed from a normal pancreas to chronic pancreatitis remained obscure. Resolving the multiple etiologies and mechanistic pathways could not be accomplished by examining end-stage pathology or by comparative gene expression profiles among patients with a normal pancreas and those with chronic pancreatitis because gene profiling defines abnormalities at a molecular level without discerning the proximal causes of the disease process. A new etiologic model of pancreatitis became necessary when it was discovered that the major genetic susceptibility factors for chronic pancreatitis actually caused acute pancreatitis through genetic defects that affected the ability of the pancreas to protect itself from premature trypsin activation.8 Trypsin was recognized as the critical molecule in pancreatitis because it is the master enzyme that controls activation of the other digestive enzymes inside the pancreas, and these enzymes cause tissue injury that triggers an inflammatory response. Most factors that increase susceptibility to pancreatitis disrupt a mechanism protecting the pancreas from trypsin-associated injury. Because trypsinogen is synthesized, stored, and transported almost exclusively by the pancreas, loss of mechanisms that protect the body from trypsin-associated injury specifically target the pancreas for recurrent injury. The classic example is hereditary chronic pancreatitis in which it has been observed that affected individuals experience recurrent acute pancreatitis that precedes chronic pancreatitis by a number of years.9-11 It is now recognized that a variety of genetic mutations exists that each increase the risk of pancreatitis through inadequate injury protection from trypsin. Furthermore, individuals with pancreatitis susceptibility genes only have occasional episodes of acute pancreatitis that are often triggered by identifiable environmental factors such as alcohol consumption or cigarette smoking. Because any number of stressors can trigger pancreatitis in ge­ netically susceptible individuals, these factors can be cate­ gorized as metabolic and environmental stressors. Some individuals who inherit pancreatitis susceptibility genes and other genetically intact individuals exposed to strong environmental stimuli never develop recognizable acute pancreatitis, suggesting that triggering recurrent acute pancreatitis requires at least one factor in both domains. Another observation is that not all patients carrying the hereditary pancreatitis gene who have recurrent acute pancreatitis develop chronic pancreatitis.11 Chronic pancreatitis is characterized as chronic inflammation and fibrosis, which are immune-mediated processes.1 Fibrosis is the product of activated pancreatic stellate cells that are driven by anti-

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood

Metabolic and environmental stresses

Inadequate injury protection

RAP

CATIONIC TRYPSINOGEN GENE MUTATIONS

CP No disease

nized that the major susceptibility genes and modifier genes appear in multiple clinically defined syndromes. Therefore, information on the major susceptibility genes and modifier genes is presented first, followed by a description of the recognized clinical syndromes.

No disease

Altered immune response

Figure 57-1.  Chronic pancreatitis (CP) as a complex disorder. Three general categories of risk factors contribute to the development of chronic pancreatitis. Genetic mutations in any number of susceptibility or modifier genes (e.g., PRSS1, CFTR, SPINK1) result in inadequate injury protection from prematurely activated trypsinogen, resulting in pancreatic injury and acute pancreatitis. Repeated exposure to metabolic and environmental stressors that potentially activate trypsinogen result in recurrent acute pancreatitis (RAP) in subjects with inadequate injury protection. The normal immune response to injury is complete healing. However, patients with environmental or genetic factors that alter the immune response to promote and accelerate fibrosis (e.g., tobacco smoking, excessive alcohol consumption) develop chronic pancreatitis. CFTR, cystic fibrosis transmembrane conductance regulator; PRSS1, protease serine 1; SPINK1, serine protease inhibitor, Kazal type 1. (Modified from Whitcomb DC. Advances in understanding the mechanisms leading to chronic pancreatitis. Nat Clin Pract Gastroenterol Hepatol 2004; 1:46-52.)

inflammatory cytokines including transforming growth factor-β (TGF-β) and interleukin-10 (IL-10).12-14 The inflammatory process is independent of the mechanism of injury and can be modified by genetic and environmental factors that influence the severity of fibrosis.8,15 If the normal response to pancreatic injury is recovery, then a complication of recurrent acute pancreatitis is extensive fibrosis (i.e., chronic pancreatitis). Factors that affect the immune system by accelerating fibrosis can be grouped together in the category of altered immune response to pancreatic injury. Figure 57-1 illustrates the complexity of chronic pancreatitis and the typical interaction of at least three different categories of risk factors before chronic pancreatitis develops. The study of pancreatic diseases, including chronic pancreatitis, must advance to predictive mathematical models that include all of the relevant risk factors and variables that are common in the human population, so that patient-specific treatments can be designed. This chapter focuses on the major genetic variables that will necessarily become part of the eventual predictive models, but that provide insight into specific disease syndromes that are already recognized.

MAJOR SUSCEPTIBILITY GENES CAUSING PANCREATIC DISEASE Genetic variants that increase susceptibility to pancreatic disease generally disrupt a major function or process within the pancreas and thereby target the pancreas rather than other organs. In these cases the effect of genetic mutations is so strong that small minor environmental stressors or other risk factors are necessary for a form of the syndrome to be manifest. Several well-recognized syndromes include pancreatic injury or dysfunction as a dominant feature in CF and Shwachman-Diamond syndrome. It is now recog-

The cationic trypsinogen gene (UniGene name: protease, serine 1; PRSS1) was identified as the first pancreatitisspecific susceptibility gene through genetic linkage studies in families with hereditary pancreatitis.9 Cationic trypsinogen is the major form of trypsinogen (≈65%) followed by anionic trypsinogen (PRSS2, ≈30%), and mesotrypsin (PRSS3, ≈5%). The trypsin molecule is formed by a single peptide that folds into an enzyme with an active site between two globular domains linked by a single connecting chain. An eight amino acid extension of the enzyme, called the trypsinogen activation peptide (TAP), maintains the enzyme as inactive trypsinogen until it is cleaved by enterokinase or another trypsin molecule (autoactivation). Cleavage of TAP allows a conformation change that activates trypsin. Trypsin is also susceptible to trypsin-mediated autolysis beginning at the arginine 122 (R122) site of the connecting chain. The connecting chain is further degraded by chymotrypsin C (CTRC),16 which was biochemically characterized as enzyme Y.17 Finally, the trypsinogen molecule also has two separate calcium binding pockets that play key roles in trypsin regulation by exposing the activation site and blocking the autolysis site, respectively. Trypsin plays a critical role in pancreatic physiology as the activator of the other pancreatic zymogens, a process that normally occurs within the duodenum where the zymogen activation cascade is initiated by the enterokinasestimulated conversion of trypsinogen to trypsin. Tryp­ sinogen activation and trypsin inactivation are primarily controlled by trypsin (autoactivation and autolysis), and the ambient calcium concentration serves as the switch between on and off (Fig. 57-2).8 Calcium binding to the first binding pocket that is formed by four aspartic acids within the TAP portion of trypsinogen facilitates trypsinogen activation by trypsin. Calcium binding to the second calcium binding pocket formed by a peptide loop in both trypsinogen and trypsin located adjacent to the autolysis loop, prevents exposure of the trypsin-sensitive R122 autolysis site and thereby prevents autolysis. Thus, physiologic regulation of trypsin activity is determined by cellular calcium, with increased cellular calcium facilitating activation and preventing inactivation and low cellular calcium levels limiting activation and permitting autolysis. The maintenance of low calcium concentrations within the acinar cells is critical to protecting them from premature trypsinogen activation. However, other protective mechanisms are used to limit trypsinogen activation within the high calcium concentrations of the pancreatic duct fluid. Acinar cell calcium can rise through neurohormonal hyperstimulation (which opens basolateral calcium channels and is linked to calcium tunnels transporting calcium to the acinar pole)18,19; high extracellular calcium concentrations and submaximal pancreatic stimulation20; bile acid reflux, which opens apical membrane calcium pathways21; and prolonged, high-dose alcohol consumption, which lowers the threshold for stimulation-induced acute pancreatitis,22 possibly through mitochondrial damage23 and other factors that regulate intracellular calcium.24 Any process that increases acinar cell calcium will predispose to acute pancreatitis through a calcium-dependent trypsinogen activation and stabilization mechanism.24 More than 20 mutations have been identified in PRSS1 that increase susceptibility to recurrent acute pancreatitis,25

933

934

Section VII  Pancreas CASR (EtOH)

R122

Trypsinogen

Ca + + CTRC PRSS1 +

(zymogen granule)

+ +

− S

+

secretion

+

+

−

N29

(TAP)

−

Trypsin

Injury

Trypsin degradation (acinar cell) Trypsin flushing (into duodenum)

CFTR AIR

* SPINK1

N34

SPINK1 Figure 57-2.  X-ray crystallography-based model of cationic trypsinogen (PRSS1) and pancreatic secretory trypsin inhibitor (SPINK1). The tryp­ sinogen molecule contains two globular domains (blue and yellow) joined by a connecting side chain (see top of drawing). Trypsinogen is activated to trypsin with cleavage of trypsinogen activation peptide (TAP), allowing a three-dimensional conformational change, opening of the specificity pocket (S), and high-efficiency enzyme activity at the active site (*). The location of the two major PRSS1 mutations (N29, R112) associated with hereditary pancreatitis are illustrated. Note the location of R112 in the side chain connecting the two (blue and yellow) globular domains. The SPINK1 mole­cular is shown bound to trypsin. The location of the major SPINK1 mutation associated with idiopathic and familial pancreatitis, N34, is illustrated. (Figure courtesy of Drs. Andrew Brunskill and William Furey.)

although the R122H and N29I mutations are most common. The locations of the R122H and N29I (N21I) mutations are shown in relationship to the active site in Figure 57-2 and in a mechanistic model in Figure 57-3. The mutations are clustered in regions associated with calcium-dependent trypsin regulation and may confer “gain-of-function” features by facilitating trypsinogen activation or retarding trypsinogen inactivation independent of cell calcium. Gainof-function mutations often result in an autosomal dominant inheritance pattern; only one of the two trypsinogen alleles must code for a super-functional trypsin in order to prematurely trigger the zymogen activation cascade and cause pancreatitis, thus manifesting the phenotype. Other trypsinogen mutations that are unrelated to calciumdependent trypsin regulation may predispose to recurrent pancreatitis by altering the activation or inactivation process normally regulated by pH or through interaction with other molecules,26 but the clinical relevance of these potential types of trypsinogen variants remains an area of investigation. The fact that the trypsinogen molecule has two calcium “switch” sites may explain why pancreatitis occurs only intermittently. The clinically important trypsinogen mutations and polymorphisms are discussed following. The importance of maintaining a tight regulation of cationic trypsinogen is further highlighted by additional genetic findings. First, the R122H and N29I mutations

Figure 57-3.  Diagram of trypsin control mechanisms in the pancreas (green arrows are positive and red lines are inhibitory influences). Trypsinogen (far left) is protected from activation within the zymogen granules. Trypsinogen activation (arrow to trypsin) is supported by elevated calcium (Ca++) and trypsin autoactivation (green dashed arrow from trypsin to the left). Trypsin degradation is inhibited by calcium, but facilitated by trypsin (solid green arrow from trypsin to the right) and chymotrypsin-C (CTRC). Mutations in the calcium-sensing receptor (CASR) or CTRC lead to excess trypsin and eventually pancreatic injury (yellow). After pancreatic injury and an acute inflammatory response (AIR), there is a dramatic upregulation of SPINK1, which inhibits trypsin’s effects on autoactivation and further damage. If trypsinogen (and trypsin) are secreted into the duodenum, then the pancreas is protected by removal of trypsin from the pancreas by a CFTR-dependent mechanism. Blue letters indicate that mutations in the corresponding genes are associated with an increased risk of pancreatitis. Dashed lines indicate a pathway contingent on the absence of inhibition. CFTR, cystic fibrosis transmembrane conductance regulator; PRSS1, protease serine 1; SPINK1, serine protease inhibitor, Kazal type 1.

appear to be gene conversions, in which a segment of deoxyribonucleic acid (DNA) from another similar gene replaces a similar segment of the gene of interest. There is now strong evidence that the “H” of PRSS1 R122H is a conversion mutation from Tryp 6,27 and the “I” of N29I is from PRSS2.28,29 These findings highlight the critical importance of the highfidelity regulatory mechanism of PRSS1 in preventing recurrent, premature trypsinogen activation or trypsin survival in low-calcium environments. Secondly, it has been reported that gene copy number variants are also associated with a risk of chronic pancreatitis in France.30,31 In this case a segment of the genome containing the trypsinogen genes is duplicated on one chromosome, and when combined with the expression of trypsinogen for the opposite allele gives an extra “dose” of trypsinogen—enough to increase the risk of pancreatitis.

ANIONIC TRYPSINOGEN GENE MUTATIONS

Anionic trypsinogen (UniGene name: protease, serine 2; PRSS2) is a form of pancreatic trypsinogen that is usually expressed at about half the amount as cationic trypsinogen, although this ratio may change in some cases.32,33 To date, no gain-of-function mutations have been identified. However, a loss-of-function mutation, PRSS2 G191R, is associated with protection from pancreatitis.34,35 The mutation introduces an arginine (“R”) into a surface loop of PRSS2, making it a target for trypsin-mediated degradation.

CALCIUM-SENSING RECEPTOR GENE POLYMORPHISMS

Regulation of intra-acinar cell calcium is critical for the prevention of pancreatic injury.24,36 The calcium-sensing

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood Carbohydrate side chain Anion channel

Extracellular Cell membrane

Transmembrane domains

1

2

3

4

5

6

7

8

9

10

11

Intracellular

12

c

N (Regulatory domains)

NBD1

(Second messengers)

ATP

R domain

NBD2

P P

ATP

PKA Figure 57-4.  CFTR structural domains. The CFTR molecule is a single peptide that forms a regulated anion channel through the apical cell membrane of the pancreatic duct cell. CFTR exists in at least two conformations (single channel and double channel). The molecule is positioned in the cell membrane by 12 transmembrane domains (numbered 1 through 12). There are at least three major regulatory domains including nucleotide binding domain 1 and 2 (NBD1, NBD2) and a regulatory domain (R domain). Several second-messenger systems interact directly with the regulatory domain, including ATP and PKA. Calcium, intracellular glutamate, and other second messenger systems or factors also regulate various aspects of CFTR (not shown). ATP, adeno­sine triphosphate; CFTR, cystic fibrosis transmembrane conductance regulator; PKA, protein kinase A.

receptor (CASR) is a membrane-bound member of the G-protein–coupled receptor superfamily.37 CASR plays an important role in calcium homeostasis, as is reflected in its expression by cells of the parathyroid gland and renal tubules that are involved in the calcium metabolism. CASR has been identified in human pancreatic acinar and ductal cells, as well as in various nonexocrine tissues,38 although its functional significance in the pancreas has not yet been determined. More than 100 functional mutations (40 activating and 72 inactivating) have been described in the CASR mutation database related to familial hypo­ calcuric hypercalcemia (FHH), neonatal severe primary hyperparathyroidism (NSPHT), autosomal dominant hypocalcaemia (ADH) and related hypercalcemic or hypocalcemic disorders.39 In 2003 Felderbauer40 investigated a kindred with familial pancreatitis and a serine protease inhibitor Kazal type 1 (SPINK1) gene mutation (see following for discussion of SPINK1 mutations). However, only two of these family members had chronic pancreatitis, and both were found to have a novel CASR 518T>C mutation that was linked to hypercalcemia. An association between additional CASR variants, with or without SPINK1 mutations, was subsequently identified in patients in India with TP,41 as well as in the United States in sporadic and alcoholic chronic pancreatitis in which the CASR mutation doubles and triples the relative risk, respectively.42 The finding of different CASR polymorphisms in different populations is intriguing, but it appears that the presumed mild hypercalcemia is a cofactor rather than an independent risk factor, as seen in animal models of hypercalcemia.20,43 At the present time, CASR mutations analysis is not offered for clinical decision making.

CYSTIC FIBROSIS TRANSMEMBRANE CONDUCTANCE REGULATOR GENE MUTATIONS

The cystic fibrosis transmembrane conductance regulator (CFTR) gene is the most important molecule for regulating pancreatic duct cell function. This includes generating the bicarbonate-rich pancreatic juice that helps neutralize the

acidic chyme coming from the stomach and flushing digestive enzymes out of the pancreas and into the duodenum. Both functions are accomplished by anion secretion through the CFTR located on the apical side of the pancreatic duct cells (see Chapter 56). The CFTR molecule becomes relevant to pancreatic diseases when its function, or regulation of its function, is altered by various gene mutations. The CFTR molecule forms a regulated ion channel expressed on epithelial cells in the respiratory system, sweat glands, the digestive tract mucosa, biliary epithelium, pancreatic duct cells, and other locations. The primary anions conducted through CFTR under physiologic conditions are chloride and, under some conditions, bicarbonate. The CFTR gene contains more than 4300 nucleotides, with 24 exons and three splice variants that code for a single protein of 1480 amino acids.44 The CFTR molecule has 12 membrane spanning domains, 2 nucleotide-binding domains (NBD1 and NBD2), and a regulatory domain (R domain) with multiple phosphorylation sites (Fig. 57-4). Although the regulation of CFTR is complex, many of the components that are relevant to the pancreas are now understood. CFTR-associated secretion is stimulated when the duct cell is stimulated by secretin or vasoactive intestinal peptide (VIP) acting on receptors that increase intracellular cyclic adenosine monophosphate (cAMP). The cAMP activates protein kinase A–mediated phosphorylation of various sites in the R domain, followed by increased anion conductance (e.g., chloride, bicarbonate) through the CFTR channel. The function associated with the individual phosphorylation sites on the R domain differs.45 Possible consequences of R domain phosphorylation include movement and insertion into the apical membrane, increased or decreased channel activity or specificity, or stabilization of other parts of the molecule such as NBD1. Duct cell stimulation by cholinergic agents or other agonists that increase intracellular calcium also potentiate anion secretion. Recent studies demonstrate that chloride conductance is regulated by cytoplasmic glutamate, while bicarbonate conductance is regulated by binding of ATP to NBD1 and NBD2,46 which may form a heterodimer to maximize adenosine triphosphatase (ATPase) activity.47 Structural studies of the CFTR protein

935

936

Section VII  Pancreas suggest that the molecule exists in two different conformations depending on the presence or absence of ATP binding to NBD1 and NBD2.48 If the basolateral membrane of the duct cell is nearly impermeable to chloride during bicarbonate secretion and if bicarbonate conductance through CFTR is limited because of an unfavorable CFTR conformation, then net ion transport across the basolateral and apical membranes would markedly be reduced49 and the duct cells would not be able to help flush digestive enzymes out of the duct. This would put the pancreas at risk of recurrent acute pancreatitis because prematurely activated digestive enzymes would not be removed from the pancreas. Major mutations in both CFTR alleles result in loss of CFTR function. The consequences include inability to adequately hydrate mucus and other macromolecules, leading to accumulation of viscid material and inspissated glands. This condition results in progressive organ destruction of the pancreas and respiratory system, and dysfunction of the liver, intestine, sweat glands, and other sites where epithelial cell secretion plays an important physiologic role. The pancreas incurs a double risk because much of its proteins are zymogens and trypsin activation will lead to recurrent injury and eventually destruction of the pancreas through progressive fibrosis. Trypsin-mediated injury and destruction of the pancreas in children with CF is consistent with this model because the pancreatic pathology in CF is pseudocyst formation and fibrosis rather than atrophy (as expected with duct obstruction).50 It appears that pancreatic gland injury in CF children roughly parallels the expression of trypsinogen in the developing acinar cells, which begins at 16 weeks’ gestation and gradually increases in concentration until birth and through the first six months of life when levels markedly rise.51,52 The resulting histology has many of the features of end-stage chronic pancreatitis that develops in children and adults, but also has striking expanded ducts that appear as multiple protein filled cysts (Fig. 57-5).

The overall clinical picture in an individual case depends on the nature of the combined CFTR mutations, the genetic background in which the defective genes operates (e.g., modifier genes), and environmental factors.50,53 About 70% of white patients with CF have a three-base pair deletion of the phenylalanine-coding codon 508 (?F508), although 1600 other mutations have been reported. Distinct mutations are common to certain ethnic groups, including African Americans who carry the 3120+1G>A mutation at a frequency of 12.3%54 and the R334W mutation in Hispanics, which is associated with CF with pancreatic sufficiency (PS) but recurrent acute pancreatitis.55 Patients with one severe CFTR mutation and one mild CFTR mutation (e.g., R117H or R334W) often have CF with pancreatic sufficiency.56,57 The reason may be that mutations such as R117H markedly reduce chloride conductance without affecting bicarbonate secretion.46,58 Reduction of chloride conductance would affect all of the epithelial cells in organs that use CFTR to transport chloride, but would have much less effect on the pancreatic duct because it uses CFTR to transport bicarbonate.49 On the other hand, CFTR mutations that specifically inhibit bicarbonate secretion put the pancreas at risk of recurrent pancreatitis without affecting the organs that use CFTR to transport chloride. The functional consequence of CFTR mutations depends on the combined effects of both CFTR alleles with the severity of the phenotype dependent on the mildest mutation.59 The most common CFTR mutations have been organized according to the effect on clinical phenotype (severe, mildvariable, borderline, benign) and the effect on CFTR protein structure and function (classes 1 to 5).50,53 (Examples are given in Table 57-1.) Class 1 to 3 mutations result in no functional protein and therefore are associated with a severe phenotype when combined with a second severe CFTR mutation. Class 4 and 5 CFTR mutations result in CFTR proteins with altered, but residual function, and are associated with mild-variable or borderline phenotypes. There is current debate as to whether additional classes are justified. If a class 4 or 5 mutation associates with a class 1 to 3 mutation, the phenotype is mild CF, with only a subset of organs affected.60 The resulting conditions are often called atypical CF.59 If a class 1 to 3 or some class 4 mutations are combined with wild-type CFTR or a benign polymorphism, the overall function of CFTR is reduced by up to 50%, but the phenotype is usually normal because more than 90% of overall CFTR function must be lost before clinical features of atypical CF are seen.59 Recurrent acute pancreatitis requires, at minimum, a partially functioning pancreas and is therefore seen in some cases of CF with pancreatic sufficiency and atypical CF. Many of the features of CF cannot be explained by variations in CFTR sequence. Instead, these features are caused Table 57-1  Classification of CFTR Mutations, Resulting Defect, and Associated Degree of Pancreatic Dysfunction

Figure 57-5.  Histology of the pancreas in cystic fibrosis (CF). This case is from an autopsy of a child with severe features of cystic fibrosis. There are no residual normal ducts or acini. Instead, dilated ducts and “cysts” with inspissated material dominate the image. Other cases of CF span the spectrum between this image and chronic pancreatitis seen with other forms of pancreatitis, with acinar atrophy, fibrosis, and chronic inflammation. Arrows demonstrate residual islets. (Image courtesy of Alysa Krasinska, MD.) (From Whitcomb DC. Cystic fibrosis–associated pancreatitis. In: Beger HG, Warshaw A, Buchler MW, et al, editors. The Pancreas: An Integrated Textbook of Basic Science, Medicine and Surgery. Oxford, England: Blackwell; 2008.)

CLASS 1 2 3 4 5

MUTATION (EXAMPLE) W1282X ΔF508 G551D R117H IVS8 5T (with 12 or 13 TG)

DEFECT

PANCREAtic DYSFUNCTION

Synthesis Maturation Activation Conductance Abundance

Severe Severe Severe Mild Mild

CFTR, cystic fibrosis transmembrane conductance regulator. Modified from references 50, 53, and 381.

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood by specific environmental factors or modifier genes.50,61 Environmental factors, such as bacterial colonization of the respiratory system, tobacco smoke,62 and poor nutritional status contribute to the severity of lung disease.63 The risk of liver disease appears to be independent of CFTR genotype; it is associated with meconium ileus.64 Careful con­ sideration must be given to patient with either classic CF symptoms, or atypical presentations resulting from less common combinations of genetic and environmental factors. In 1998 two groups65,66 demonstrated that CFTR mutations were also very common in idiopathic and alcoholic chronic pancreatitis, which indicates that some of the more than 1600 known CFTR gene sequence variants cause milder disease (atypical CF59), pancreas-specific injury or that CFTR gene mutations may be part of a more complex trait.1 In some cases, recurrent acute pancreatitis and chronic pancreatitis appeared to be associated with heterozygous CFTR genotypes. Because heterozygous CFTR mutations and polymorphisms are common in the general European and American population, and because the parents of CF children (obligate CFTR mutation carriers without CF) do not have an increased incidence of acute or chronic pancreatitis compared with the normal population,67 it is likely that a second factor that specifically targets the pancreas is required.1 In early onset idiopathic pancreatitis this second factor may be a SPINK1 mutation (discussed later), a potent environmental factor,68 or other complex genetic combinations.69-73

SHWACHMAN-BODIAN-DIAMOND GENE MUTATIONS

The Shwachman-Bodian-Diamond syndrome gene (SBDS) is a gene of unknown function that is mutated in most cases of Shwachman-Diamond syndrome (SDS).74,75 This gene has five exons and encodes a predicted protein of 250 amino acids. The genetic defect in most cases of SDS is caused by gene conversion between the normal SBDS gene with a nonfunctional pseudogene designated SBDSP.74 The SBDSP pseudogene DNA code is 97% identical to SBDS gene code. The differences between the SBDS gene and SBDSP pseudogene are critical nucleotide deletions and nucleotide changes that render the SBDS gene product nonfunctional. Fourteen distinct mutations were initially identified in these kindreds, with the most common being the conversion mutations 183-184TA→CT and 258+2T→C. Interestingly, most patients had compound heterozygous mutations, and no patient was homozygous for the common 183-184TA→CT mutation. Although the function of the gene is unknown, there is significant homology with genes in other species that regulate or facilitate mRNA utilization or metabolism.74 Recent studies suggest that the gene encodes a protein that interacts with ribosomal RNA and helps regulate the maturation of the 60s ribosomal subunit and may also stabilize the mitotic spindle.76-78 The genetic defect results in an acinar cell–specific defect with markedly reduced zymogen synthesis and pancreatic insufficiency rather than susceptibility to pancreatitis. The other clinical features of the SDS are discussed later.

MODIFIER GENE A modifier gene differs from a susceptibility gene in that it does not, by itself, increase the risk of developing a pathologic outcome. Here we will discuss SPINK1 polymor-

phisms, which modify the body’s response to recurrent pancreatic injury.

PANCREATIC SECRETORY TRYPSIN INHIBITOR GENE MUTATIONS

Pancreatic secretory trypsin inhibitor (PSTI, UniGene name: serine protease inhibitor, Kazal type 1; SPINK1) is a 56 amino acid peptide that specifically inhibits trypsin by physically blocking its active site (see Fig. 57-2). SPINK1 is synthesized by pancreatic acinar cells along with trypsinogen and it co-localizes with trypsinogen in the zymogen granules. In the mechanistic models of pancreatic acinar cell protection, SPINK1 acts as the first line of defense against prematurely activated trypsinogen in the acinar cell.9,32,79 SPINK1 is an acute phase reactant and concentrations in serum rise markedly with inflammation.80,81 Under normal conditions there is a great excess of potential trypsin compared with SPINK1 so that the inhibitory capacity of SPINK1 is limited, but with inflammation the expression of SPINK1 is increased dramatically.82 A number of SPINK1 mutations have been identified, but a high-risk haplotype (five polymorphisms that are inherited together) defined by SPINK1 N34S is by far the most common, being present in 1% to 4% of most populations throughout the world.79,83,84 Several other variants of the SPINK1 gene also have been described.25 Pancreatitis associated with SPINK1 gene mutations is associated with early onset recurrent acute and chronic pancreatitis in children,79 familial pancreatitis,83 and TP,85 and is often a feature of polygenic pancreatitis–associated genotype. Because SPINK1 is a specific trypsin inhibitor, and because expression of SPINK1 is normally very low,82 it follows that SPINK1 cannot be a significant inhibitor of prematurely activated trypsin until the acute inflammatory process has become well established. Thus, the effect of loss-of-function SPINK1 polymorphisms would only become apparent late in the course of acute pancreatitis, or in the case of recurrent premature trypsinogen activation, as is seen with PRSS1 or CFTR mutations. This hypothesis was tested using multiple meta-analyses on data that were classified by proximal etiology.86 This analysis suggested that the strongest effect of SPINK1 N34S was in pancreatitis etiologies that were linked with recurrent trypsin activation, but the effect was low in other etiologies (e.g., alcoholassociated pancreatitis). If this is mechanistically true, then up-regulation of mutant SPINK1 will fail to prevent trypsin-associated recurrent pancreatic injury. Furthermore, suppression of recurrent proinflammatory responses with macrophage-derived anti-inflammatory cytokines (e.g., TGF-β1) will drive the pancreatic stellate cells to produce fibrosis.7,14,87 The implication is that the discovery of a SPINK1 mutation in an unaffected individual is of minimal importance, whereas the effect of a SPINK1 mutation in driving fibrosis in someone with PRSS1 or CFTR mutations is very strong. A summary of the mechanisms for trypsin activation and inactivation is illustrated in Figure 57-3.

PANCREATIC DISEASE IN CHILDREN Whereas pancreatic disease in children was once considered uncommon, evidence suggests that the incidence is increasing. Although diagnostic modalities and physician awareness continue to improve, this does not appear to account for the increase.88 Acute pancreatitis occurs in all age groups, including infants.89 The most common causes

937

938

Section VII  Pancreas of acute pancreatitis in adults are excessive alcohol use and gallstones. These risk factors are less often seen in children, although biliary pancreatitis is now recognized in this age group. The majority of cases of recurrent acute and chronic pancreatitis in children have a structural or genetic basis. The genetic factors predisposing to acute pancreatitis appear to be similar to those associated with chronic pancreatitis and are discussed in detail in the following sections.

ETIOLOGY OF ACUTE PANCREATITIS

The primary causes of acquired pancreatitis in children are listed in Table 57-2. In a recent review of 1276 children with acute pancreatitis compiled from five studies, the most common causes were idiopathic (22.2%), association with systemic disease (20.8%), trauma (18.6%), structural (e.g., pancreas divisum) (10.6%), and medications (10.2%); gallstones, post–endoscopic retrograde cholangiopancrea­ tography (ERCP), familial, hypercalcemia, hyperlipidemia, diabetic ketoacidosis, and “other” causes made up the remaining etiologies.88 Growing evidence suggests that some of these cases occur in children with high-risk genetic alterations, especially pancreatic-specific combinations of SPINK1 and CFTR mutations. Genetic testing, discussed later in this chapter, is usually performed after recurrent

Table 57-2  Reported Causes of Acquired Pancreatitis in Children Trauma ERCP Medications Alpha-methyldopa*† Azathioprine* Azodisalicylate† Cimetidine*† Cytosine arabinoside* Didanosine Erythromycin Estrogen* Furosemide*† Glucocorticoids*† Intravenous lipid emulsion Isoniazid* Lamivudine L-Asparaginase 6-Mercaptopurine* Mesalamine* Metronidazole* Pentamidine* Pentavalent antimony Procainamide*† Rifampin Sulfasalazine*† Sulfonamides* Sulindac† Tetracycline* Valproic acid* Zalcitabine† Infections Ascariasis Coxsackie B virus Echovirus virus Enterovirus virus Epstein-Barr virus Hepatitits A virus

Herpes viruses Influenza A Leptospirosis Malaria Measles Mumps Mycoplasmosis Rabies Rubella Typhoid fever AIDS-associated infections (see Chapter 33). Cryptosporidia† Cytomegalovirus HIV† Mycobacterium avium complex† Mycobacterium tuberculosis† Toxoplasma gondii Biliary tract disease Pancreas divisum Metabolic Cystic fibrosis Hypercalcemia Hypertrigylceridemia Protein-caloric malnutrition Reye’s syndorme Familial Miscellaneous Congenital partial lipodystrophy Diabetic ketoacidosis Henoch-Schönlein purpura Juvenile tropical pancreatitis Kawasaki’s disease Perforated duodenal ulcer Systemic lupus erythematosus Idiopathic

*Class 1 drug: pancreatitis occurs with rechallenge(see Chapter 58). † Probable cause of pancreatitis. AIDS, acquired immunodeficiency syndrome; ERCP, endoscopic retrograde cholangiopancreatography; HIV, human immunodeficiency virus.

episodes and when other common causes have been excluded.

Trauma

Trauma is a frequent cause of acute pancreatitis despite the fact that the pancreas is well protected from minor injury by its retroperitoneal location. The trauma is usually blunt, associated with injuries to other abdominal viscera and becomes evident soon after the injury,90 although injury may apparently precede the manifestation or recognition of pancreatitis by several weeks. In such an instance, a precise relationship is unclear. Perhaps of more importance is that the possibility of injury to the pancreas is often not considered in a severely injured or battered child.91

Structural Abnormalities

Structural abnormalities are being recognized earlier as imaging techniques such as magnetic resonance imaging (MRI) and magnetic resonance cholangiopancreatography (MRCP) improve. Pancreas divisum is the most common anatomic aberration, although a wide variety of other structural abnormalities of the bile and pancreatic duct also have been observed (see Chapter 55).88 The widespread availability of MRCP has drastically reduced the use of diagnostic ERCP. Post-ERCP pancreatitis has been a significant cause of pancreatitis in several series,92,93 and this etiology is seen wherever ERCP is performed in children. ERCP remains invaluable for therapeutic intervention (see Chapter 61).

Biliary Tract Disease

Gallstone pancreatitis is less common in children than in adults and is probably a reflection of the relative infrequency of cholelithiasis before puberty. However, because almost 10% (as high as 30% in a single study) of children with pancreatitis in some series had cystic duct stones or common bile duct disease, this diagnosis must be considered, regardless of age.94 Little is known about the natural history of this disease in children.

Medications

Medications remain a frequent cause of acute pancreatitis in children, although the disease underlying the prescription must also be considered in the differential diagnosis.95-97 Recent studies identified valproate as the most frequent drug associated with pancreatitis in children, followed by L-asparaginase, prednisone, or multiple medications.92,93,98 The development of persistent abdominal pain in a child receiving any medication should suggest the possibility of drug-induced pancreatitis. This is confirmed only by documentation of pancreatic disease, improvement on drug withdrawal, and return of disease when the drug is reintroduced.

Infection

Infections, particularly with viruses, are a frequently associated with childhood pancreatitis; a partial list of putative agents appears in Table 57-2. Enteroviruses, particularly coxsackievirus and echovirus, have been documented by stool isolation and concomitant serum titer rise in up to 8% of adults with idiopathic acute pancreatitis. Only about half of virus isolations are associated with an antibody rise.99-101 Pancreatitis has been reported in children with EpsteinBarr virus infections, often appearing after an initial clinical improvement.102,103 Interstitial pancreatitis has been described in the congenital rubella syndrome.104 Pancreatitis in children is often attributed to mumps virus on the basis of abdominal pain and an elevated serum amylase

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood value, with parotitis or waxing mumps antibody titers or both.105 Confirmation by serum isoamylase or lipase determinations and by abdominal ultrasonography is lacking, however, and the frequency of this entity may be overestimated. Bacterial pancreatitis has been described in one patient,106 although in that patient there may have been other causes for development of pancreatitis, including antecedent hypotension. Mycoplasma pneumoniae infection, followed one to two weeks later by clinically apparent pancreatitis, has been seen in an estimated 8% of patients with this infection. Complement-fixation titers and serum immunoglobulin M values were elevated, and other causes of pancreatitis were absent.107 Typhoid fever often manifests with abdominal pain; pancreatitis has been suggested as one possible cause.108 Although uncommon in the United States, ascariasis is among the most frequent causes of pancreatitis in children in regions such as South Africa and India. Worms can be found within the pancreatic duct and can be vomited as the initial diagnostic clue. Malaria also has been reported to cause pancreatitis.109 Pancreatitis is 35 to 800 times more common in patients with acquired immunodeficiency syndrome (AIDS).110 This extremely high risk for acute pancreatitis is attributed to several factors (see Chapter 33). A number of medications that are frequently prescribed to human immunodeficiency virus (HIV)–infected patients are associated with pancreatitis (see Table 57-2), possibly due to direct toxicity to pancreatic acinar cells. In addition, immunodeficiency itself predisposes patients to pancreatic infection.

Acute Pancreatitis in Systemic Diseases

Acute pancreatitis is often seen in patients with severe systemic illnesses.88,93 Hemolytic uremic syndrome (HUS) was the most common cause of acute pancreatitis of all systemic diseases in two major studies.92,93 The mechanism is unknown and likely multifactorial, and uremia itself is a risk factor for pancreatic injury.111-113 Systemic lupus erythematosus has been reported in association with pancreatitis.90 Two cases of clinically significant pancreatitis have been documented in association with Kawasaki’s disease.114 Histologic changes have been known to occur in the pancreas during Reye’s syndrome115 but it is unclear if these changes were specific to the disease. Usually this complication has been signaled by hypotension and rapid clinical deterioration during the treatment of advanced illness. Acute pancreatitis after organ transplantation is also common (see Chapter 34).93 Acute pancreatitis should be considered in the intensive care unit when the child is not responding to other therapies or appears to have an unexplained acute inflammatory process. The evaluation of a child with acute pancreatitis should include measurement of serum calcium and triglyceride levels, and these causes must be addressed to prevent recurrence. Hypercalcemia during parenteral nutrition leading to pancreatitis was first described in a child; similar reports have followed.116 Other causes of pancreatitis were not apparent in these patients, although it has been suggested that the calcium content of the solutions infused may not have been the only factor involved in the development of hypercalcemia and pancreatic disease.116 Pancreatitis is occasionally observed in a number of other metabolic disorders, such as diabetic ketoacidosis92,93,117 and various inborn errors of metabolism.118

Acquired Metabolic Derangements

Multiple metabolic derangements are associated with the development of pancreatic disease in children. Perhaps the most common of these is protein-calorie malnutrition. In severely malnourished children, pancreatic enzyme secre-

tion is often compromised, whereas volume and bicarbonate secretion are preserved. Recovery of pancreatic function is said to occur more promptly after kwashiorkor than after marasmus, but in either case the pancreatic disease may contribute to malabsorption during convalescence. Vigorous early refeeding of malnourished children has been associated with the development of clinically significant pancreatitis. Malnutrition was considered a major contributing factor to TP, but this has now been questioned because TP is observed primarily in well-nourished patients.3,119

CLINICAL FEATURES

The diagnosis of pancreatitis is based on the syndrome of sudden onset of typical abdominal pain plus elevation of serum amylase or lipase to at least three times the upper limit of normal levels.120 The diagnosis of acute pancreatitis can be difficult because there is no readily available confirmatory test. Although there have been multiple attempts to determine the sensitivity and specificity of elevations in both enzymes in adults, the studies all suffer from the absence of a method to separately and absolutely document pancreatitis. It is clear that both enzymes can be normal when there is radiographic and clinical evidence of pancreatitis. Also, both enzymes can be elevated by other conditions unrelated to pancreatitis. The level of elevation is also not diagnostic, although the higher the level above the upper reference limit the more likely there is to be pancreatic inflammation. Levels just above the upper reference limits may still be secondary to pancreatitis, especially in patients presenting several days after the onset of symptoms. The pain is usually supraumbilical, worsens with eating, and may be accompanied by nausea, vomiting, and occasionally jaundice. A transient fever is often present. In infants and toddlers, vomiting, fever, irritability, and abdominal distention can be presenting symptoms.89 Laboratory diagnosis centers on elevated serum amylase and lipase values. Normal amylase values increase with age, which is explained perhaps by the delayed appearance of pancreatic isoamylase, which is usually not present before the age of 3 months and often not detected until the age of 11 months; even then it is not present at adult levels until the age of 10 years. Salivary isoamylase appears and matures much sooner. The serum amylase concentration may be normal, however, despite other evidence of pancreatitis. The evaluation and treatment of acute pancreatitis are covered in Chapter 58.

RECURRENT ACUTE PANCREATITIS

Recurrent acute pancreatitis is seen in about 10% of children after a first episode of acute pancreatitis.93,121 The most common diagnoses in patients with recurrent acute pancreatitis are structural abnormalities, idiopathic pancreatitis, or familial pancreatitis.93,121 A careful evaluation aimed at identifying or ruling out reversible causes should be undertaken to prevent further attacks and to reduce the risk for developing chronic pancreatitis and its complications.

CLINICAL ASPECTS OF GENETIC DISORDERS AFFECTING THE PANCREAS Several genes that are critical to pancreatic function manifest genetic variations and polymorphisms at variable frequencies among different populations and may lead to major pancreatic disorders (Table 57-3). The major clinical syndromes include CF caused by mutations in the CFTR

939

940

Section VII  Pancreas Table 57-3  Hereditary and Congenital Disorders of the Exocrine Pancreas DISORDER

Defective gene or protein

Exocrine Pancreatic Insufficiency Pancreas agenesis PDX1 or PTF1A (recessive) Cystic fibrosis CFTRsev/CFTRsev (recessive) Shwachman-Diamond SBDS (recessive) syndrome Johanson-Blizzard syndrome UBR1 (recessive) Pearson’s marrow-pancreas Mitochondrial DNA syndrome Isolated enzyme deficiency See text Pancreatitis Hereditary PRSS1 (autosomal dominant) Unknown Familial SPINK1/SPINK1 (autosomal recessive) CFTRsev or CFTRbl/SPINK1 (complex) Atypical cystic fibrosis CFTRsev/CFTRm-v (recessive) Tropical pancreatitis SPINK1/unknown (complex) Sporadic (risk factors) CTRC CASR Metabolic Hyperlipidemias Liproprotein lipase Apolipoprotein C-II Hyperparathyroidism CFTR, cystic fibrosis transmembrane conductance regulator; CFTRsev, severe CFTR mutations (classes 1 to 3); CFTRm-v, mild or variable CFTR mutations (class 4 or 5); CFTRbl, borderline CFTR mutations (class 4 or 5); CTRC, chymotrypsin-C; CASR, calcium-sensing receptor; UBR1, ubiquitin protein ligase E3 component n-recognin1; PDX1, pancrease and duodenal homeobox 1; PTF1A, pancreas specific transcription factor, 1a. SBDS, Shwachman-Bodian-Diamond syndrome gene; SPINK1, serine protease inhibitor, Kazal type 1.

gene, hereditary pancreatitis caused by mutations in the cationic trypsinogen (PRSS1) gene, and familial pancreatitis, usually caused by homozygous SPINK1 mutations or polygenic CFTR-SPINK1 genotypes. The last two disorders can also appear as sporadic pancreatitis in children. Disorders with a strong genetic basis typically have a younger age of onset than disorders requiring significant environmental exposure, such as alcoholic chronic pancreatitis. SDS is an uncommon pancreatic insufficiency syndrome without pancreatitis. Rare syndromes, including Johanson-Blizzard syndrome, Pearson’s marrow-pancreas syndrome, and other disorders are also recognized as having a genetic basis and are summarized in the subsequent section.

CYSTIC FIBROSIS

CF (OMIN 2197000122) is the most common lethal genetic defect of white populations and is seen in about 1 in 2500 to 1 in 3200 live births. The incidence of CF is about 1 in 15,000 African American and 1 in 31,000 Asian American newborns.123 Expected survival for typical CF children born up to the early 1900s was only a matter of months. Fortunately, with improved pulmonary care and nutrition, the prognosis has dramatically improved, with median survival extending beyond 36.5 years of age in 2006 and many patients living into their 50s. Although CF affects many organs, the primary focus of this chapter is on manifestations of CFTR gene mutations on the pancreas, with brief discussions of liver and intestinal problems that are also seen by the gastroenterologist.

Clinical Features

CF is diagnosed within the first year of life in more than 70% of patients and in more than 85% by age 5. However, 8% of cases remain undiagnosed until after the age of 10 years.123 A small percentage remain undiagnosed until early

adulthood. Median survival is more than 40 years with more than 95% living past age 15.60 Newborn screening for CF is done routinely in 46 of 50 states and sporadically in the others.89 The early clinical features are those of maldigestion or other pancreatic and intestinal manifestations of CFTR mutations, whereas the latter course is dominated by pulmonary complications. The presenting features during infancy include meconium ileus, malabsorption with frequent foul stools, failure to thrive, or rectal prolapse.52 Pulmonary function is normal in patients with CF at birth but accounts for much of the morbidity and almost all of the mortality associated with CF beyond the neonatal period. The severity of lung disease depends on known and unknown factors, including chronic infection with Pseudomonas aeruginosa and nutritional status, and probably the effect of unidentified modifier genes because severity of lung disease differs among patients with identical CFTR genotypes. The phenotype-genotype relationship between some childhood disorders and CFTR mutations is often striking, with severe CFTR mutations detected in more than 85% of all children presenting with pancreatic insufficiency and in the majority of infants presenting with meconium ileus.53,124 In older patients, presenting symptoms of CF may include pulmonary disease, nasal polyps, congenital bilateral absence of the vas deferens (CBAVD) with male infertility, liver disease, recurrent acute pancreatitis, or chronic pancreatitis,59,125 although the prevalence of CFTR mutations in patients who have these common disorders is low. Clinical features of CF are listed in Table 57-4 and frequencies of the various gastrointestinal (GI) manifestations of CF are listed in Table 57-5. The variations in CF presentation and clinical features reflect different combinations of CFTR gene mutations, modifier genes, and environmental factors. In cases where only a fraction of the organ systems that CF typically targets are affected, the condition is referred to as atypical cystic fibrosis (aCF). In these cases a careful family history may also provide important clues to the diagnosis of CF. Clinical confirmation of a CF diagnosis rests on the demonstration of elevated sweat chloride concentrations126 (see Table 57-4) or demonstration of an abnormal nasal bioelectric response in specific testing protocols,127 reflecting abnormal CFTR function. When performed appropriately, these tests are reliable. However, false-positive as well as false-negative results may be observed in newborns, in patients with malnutrition, in the presence of some medications, or if inadequate sweat is obtained (see Table 57-5).128 Thus, most experts insist on using the standardized methods performed at CF centers, who use these testing methods frequently. The consensus of a Cystic Fibrosis Foundation panel suggested a diagnosis of CF could be made by the presence of one or more characteristic clinical features, a history of CF in a sibling, or a positive newborn screening test result with confirmation by laboratory evidence of CFTR dysfunction. Furthermore, they suggest that either sweat chloride or nasal bioelectrical responses should be abnormal on two separate days before the diagnosis is confirmed by one of these methods.123 Genetic testing is also commercially available to confirm the clinical diagnosis (two severe mutations must be identified), but these results cannot always be interpreted apart from the clinical context and functional testing—especially in cases with atypical symptoms. Mutational screening of the entire CFTR gene should be considered in atypical cases.

Pancreatic Pathology

Of patients with CF, 85% to 90% present with evidence of exocrine pancreatic dysfunction. Although pancreatic dys-

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood Table 57-4  Clinical Manifestations of Cystic Fibrosis Upper respiratory Sinusitis Mucous membrane hypertrophy: nasal polyposis Lower respiratory Atelectasis Emphysema Infections Bronchitis Bronchopneumonia, bronchiectasis, lung abscess Respiratory failure, right-sided heart failure Gastrointestinal Gastroesophageal reflux Peptic ulcer disease Meconium ileus Volvulus Peritonitis Ileal atresia Distal intestinal obstruction syndrome Fecal masses Intussusception Obstruction Rectal prolapse Pancreatic Nutritional failure caused by pancreatic insufficiency Diabetes mellitus Calcification Maldigestion Vitamin deficiencies Loss of bile salts Steatorrhea and azotorrhea Hepatobiliary Mucus hypersecretion Gallstones, atrophic gallbladder Focal biliary cirrhosis Cirrhosis Portal hypertension Esophageal varices Hypersplenism Reproductive Females: increased viscosity of vaginal mucus, decreased fertility Males: sterility; absent ductus deferens, epididymis, and seminal vesicles Skeletal Retardation of bone age Demineralization Hypertrophic pulmonary osteoarthrophathy Ophthalmologic Venous engorgement Retinal hemorrhage Other Salt depletion through excessive loss of salt through skin Heat stroke Hypertrophy of apocrine glands

Table 57-5  Frequency of Gastrointestinal Manifestations in Cystic Fibrosis

ORGAN

COMPLICATION

Pancreas

Total achylia Abnormal glucose tolerance Partial or normal function Pancreatitis Diabetes Meconium ileus Rectal prolapse Distal intestinal obstruction syndrome Intussusception Pneumatosis intestinalis ? Mucosal dysfunction Fatty liver Focal biliary cirrhosis Portal hypertension Gallbladder abnormal, nonfunctional, or small Gallstones Bile duct strictures Gastroesophageal reflux Esophagitis

Intestine

Liver Biliary tract

Esophagus

ADULT patients (%)

85-90* 20-30

85-90* 20-30

10-15*

10-15*

1-2 4-7 10-25 1-2 3

2-3 4-7 —

‡

1

18

‡

1-2

‡

‡

7† 2-3 2-3 25

20-60 11-70

8† 1-20†

10-25 1-20† 80

‡

‡

5-20

‡

*The frequency depends on the genotype. † These percentages are largely from unpublished series of patients. ‡ Actual prevalence is unknown. From references 60 and 128.

concentrated in the shrinking pancreas.128 Calcification, although rare, may be apparent on radiographs. Ultrasonography, MRI, and computed tomographic (CT) scanning can document the progression of pancreatic disease in CF. Radiographically, the pancreas can appear normal, as incomplete or complete lipomatosis, as a cystic pancreas, as a macrocystic pancreas, or as an atrophic pancreas.129,130 The greatest sensitivity is provided either by MRI or CT scanning, but even with these methods the correlation of abnormalities with the degree of exocrine dysfunction is poor.129

Exocrine Pancreas Dysfunction

function in an infant with CF may initially appear minimal, it usually progresses to pancreatic exocrine failure. When severely affected, the pancreas is shrunken, cystic, fibrotic, and fatty.128 Histologically, hyperplasia and eventual necrosis of ductular and centroacinar cells, together with inspissated secretions, lead to blockage of pancreatic ductules and subsequently encroach on acini, causing flattening and atrophy of the epithelium (see Fig. 57-5). Cystic spaces are filled with calcium-rich eosinophilic concretions. A mild inflammatory reaction may be present around obstructed acini, and progressive fibrosis gradually separates and replaces the pancreatic lobules. The islets of Langerhans are spared in most cases until late in the process and are

all PATIENTS (%)

Pancreatic enzymes play a critical but partial role in the digestion and absorption of nutrients. Patients with CF are usually pancreatic insufficient (PI), a problem that is compounded by intestinal pathology, high-caloric demands, and poor appetite. Fat and protein maldigestion with fecal losses are the primary, pancreatic manifestations of CF, although there may be considerable variation in severity from one patient to another. Steatorrhea and azotorrhea are generally greater with pancreatic insufficiency than with mucosal malabsorption. Exocrine pancreatic insufficiency may be recognized only when the secretion of lipase and trypsin falls to less than 10% of normal.131 Most patients with CF exhibit this pattern of pancreatic insufficiency. Recurrent acute pancreatitis may complicate the course of CF in patients who do not experience complete loss of pan-

941

942

Section VII  Pancreas creatic function in infancy. Among patients with typical CF, the incidence of pancreatitis over a 30-year follow-up is less than 2%.60 Pancreatitis tends to be more problematic in older patients, with the reported incidence among patients older than 30 years being about 2.4%.132

Endocrine Pancreas Dysfunction

Glucose intolerance has been reported in 30% to 75% of patients with CF, and clinically significant diabetes mellitus occurs in up to 10% of young patients.133,134 The previously reported estimates of 1% to 2% incidence135 may have reflected younger patients and poor survival among patients with CF before recent advances in treatment.134 CF-related diabetes mellitus (CFRD) develops with increasing age. At 20 years of age, 30% of CF patients will require insulin and 40% require insulin by age 30.136 The development of CFRD differs in etiology and presentation from typical type 1 or type 2 diabetes mellitus and may reflect destruction of the islets of Langerhans137 similar to what is observed in other forms of chronic pancreatitis. However, the severity of the endocrine deficiency lags behind the exocrine deficiency because the islets are relatively spared until later in the course of pancreatic destruction (see Fig. 57-5). CFRD is associated with deterioration in both respiratory and nutritional status, the development of late microvascular complications, and increased mortality.134 No well-designed studies have addressed this significant disease complication. However, most experts recognize the need for a multidisciplinary team approach, utilization of a high-energy diet (>100% of the recommended daily intake), and appropriate adjustment of insulin doses.134 Overnight enteral feedings may be necessary to maintain adequate nutrition.

Treatment

Pancreatic Enzyme Supplements Treatment of maldigestion from pancreatic exocrine failure in CF rests on the delivery of active digestive enzymes to the proximal small intestine with meals. Numerous pancreatic preparations are available commercially, but enzyme activities vary considerably from one product to another, and reduced activity of lipase remains a problem for some patients.138 Enteric-coated minimicrospheres are now the preferred form of replacement because they protect the digestive enzymes from destruction by gastric acid (pH < 4) and are effective in treating steatorrhea.139 The size of the microspheres must be considered. If the majority of the spheres are too large (>1  mm), emptying of the spheres/ enzymes can be delayed until after food is well into the small intestine.140 The use of histamine-2 (H2) receptor blockers or proton pump inhibitors along with uncoated or enteric-coated pancreatic enzyme supplements also should be considered in patients with CF, especially because the pancreatic and duodenal bicarbonate transport systems are disrupted.141,142 However, even with optimized treatment, fat absorption may not return completely to normal. In large part, the inability to normalize fat absorption may reflect decreased uptake of fatty acids by the abnormal intestinal mucosa.143 In contrast with other forms of pancreatic insufficiency, bicarbonate secretion within the duodenum and biliary tree is also impaired in CF, resulting in a significantly lower than normal duodenal pH.144-146 Thus, without acid suppression, the uncoated enzymes are susceptible to inactivation by gastric acid, and enteric-coated products may not release their contents.147 The use of antacids containing calcium carbonate or magnesium hydroxide should be avoided because they may interfere with the pancreatic enzyme supplements.

Initial therapy for pancreatic exocrine insufficiency in CF includes pancreatic enzyme replacement at doses ranging from 500 to 2000 units of lipase activity per kilogram of body weight per meal, given just before a meal and with snacks.148 The amount is usually advanced to 1000 to 2500 units of lipase activity per kilogram, with final dosage depending on the age, the degree of pancreatic insufficiency, the amount of fat ingested, and the commercial preparation chosen. Adequacy of treatment is typically determined on clinical grounds. Frequent, bulky, and fatty stools; excessive bloating and flatus; and excessive appetite or inadequate growth velocity are signs of inadequate treatment. Pancreatic enzyme replacement is not without potential complications. Perioral and perianal irritation are common in infants, although less common with the microsphere preparations. Because of the high purine content of pancreatic extracts, hyperuricosuria may develop in some patients taking large doses of enzyme preparations.149 Powdered preparations of pancreatic extracts have caused immediate hypersensitivity reactions in parents of patients with CF.150,151 Colonic strictures and fibrosing colonopathy have been reported with very high-dose administration of pancreatic enzymes and have led to a withdrawal of all the high-dose formulations of enzymes.152,153 Fibrosing colopathy was first recognized in 1994152 and nearly disappeared by 1996.154 It usually develops as an ascending colon stricture causing intestinal obstruction and appears pathologically as postischemic ulceration with mucosal and submucosal fibrosis.152 Nearly all patients were younger than the age of 12, had prior GI surgery, had prior distal intestinal obstruction syndrome, and used H2-receptor antagonists, glucocor­ ticoids, and recombinant human deoxyribonuclease.153 However, the most striking risk was the use of high doses of lipase-containing enzyme supplements. Compared with daily doses of pancreatic enzyme supplements containing up to 2400 units of lipase per kilogram per day, the relative risk of fibrosing colonopathy was 10.9 with a daily dose of 2401 to 5000 units of lipase per kilogram per day and 199.5 for patients taking more than 5000 units per kilogram per day.153 However, because the cases and controls were taken from the same centers where a single brand of enzyme supplement was generally used, it was never conclusively determined whether the problem was primarily related to the lipase content or the acid-resistant coating of the many capsules that were ingested. Vitamin Supplements Vitamin deficiencies may develop as a consequence of fat maldigestion and malabsorption, and therefore patients with CF are at risk. Nearly half of all newly diagnosed CF patients have a deficiency of vitamins A, D, or E.155,156 Vitamin A deficiency in CF rarely manifests with clinical abnormalities.128 Vitamin D levels are dependent on sunlight exposure and intake, and the bone demineralization seen in older CF patients may be more of a reflection of general malnutrition128 and the effects of glucocorticoid treatment. Chronic vitamin E deficiency is associated with hemolytic anemia (predominantly in infants) and neuroaxonal dystrophy with prominent neuromuscular symptoms, although these clinical symptoms are rare.128 Vitamin K deficiency and the consequent coagulopathy may occur at any age. Its manifestation may vary from mildly increased bruisability or purpura to catastrophic intracranial hemorrhage in the neonatal period. CF patients who have hepatic involvement are particularly prone to coagulation abnormalities from vitamin K deficiency.157 All CF patients should receive a multivitamin preparation daily; many patients

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood require vitamin A, E, K and D supplements.158 However, frequent and serial monitoring of the serum concentrations of fat-soluble vitamins is important in children with CF because deficiencies, especially those of vitamin E, may occur during therapy.156

Intestinal Manifestations

There are myriad recognized gastrointestinal manifestations of CF (see Table 57-5). Although pancreatic failure (just reviewed) and meconium ileus (discussed following) initially dominate the clinical picture, these additional manifestations cause significant morbidity for many patients. Pathology The mucosal glands of the small intestine of patients with CF may contain variable quantities of inspissated secretions within the lumen but rarely have increased numbers of goblet cells. Brunner’s glands may show dilation, flattening of epithelial lining cells, and stringy secretions within their lumens. Severe alterations in the intestinal glands of the small bowel are found in meconium ileus.159 However, even in patients without meconium ileus, these findings are common and appear unrelated to the severity of GI symptoms or changes in other organs. The small intestinal mucosa in older CF patients often shows widely dilated crypts packed with mucus; the mucus frequently appears laminated or may extrude from a gaping crypt. Bulging goblet cells seem to crowd out the intervening columnar epithelium. Variable cellular infiltration may be present in the lamina propria. Mucus in CF is more abundant, stains more intensely, and contains more weak acid groups and protein. Mucus has increased fucosylation and sulfation and decreased sialylation.160-162 Characteristic changes of CF occur in the appendix. Increased numbers of goblet cells distended with mucus line dilated crypts. Eosinophilic casts of these crypts are extruded into the lumen of the appendix. The diagnosis of CF may be suspected on the basis of the histologic appearance of the appendix.163 Although chronic changes in the appendix are a common finding at autopsy, the incidence of acute appendicitis is apparently not increased in CF, inasmuch as only about 1.5% of patients in three large series were found to have appendicitis.164-166 The diagnosis of appendicitis in CF is often delayed and confused with distal intestinal obstruction syndrome, which results in a higher frequency of appendiceal perforation found at the time of diagnosis. The use of chronic antibiotics may also mask typical appendiceal signs.167,168 A smaller subset of patients present with chronic, intermittent pain and tenderness in the right lower quadrant, which results from appendiceal distention by inspissated mucus (but there are no findings of appendicitis on histologic examination). These symptoms are relieved by appendectomy.164 Appendicitis must be considered in all CF patients who have right lower quadrant abdominal pain. Radiographic Features Characteristic radiographic features of the intestine are frequently observed in CF. In approximately 80% of patients, thickened duodenal folds, nodular filling defects, mucosal smudging, and areas of dilation and redundancy are seen.169 The findings are not age related, and duodenal biopsies do not adequately explain the radiographic appearance. Similar changes occur in the more distal small bowel, including thickening and distortion of jejunal folds and variable dilation of intestinal loops from the jejunum to the rectum.170 Pneumatosis coli, a benign condition secondary to chronic pulmonary disease and fecal impaction, may be seen.

Functional Abnormalities Small bowel mucosal dysfunction in CF has been suggested by studies that demonstrate absorption defects that are apparently unexplained by exocrine pancreatic insufficiency or that persist after adequate pancreatic replacement therapy. Decreased activity of certain cytoplasmic peptide hydrolases in intestinal mucosa and reduced uptake of phenylalanine, isoleucine, and glycine have been found in CF patients in comparison with control subjects.171 Basal and stimulated duodenal bicarbonate secretion is largely dependent on functional CFTR, and patients with CF suffer several consequences of diminished duodenal bicarbonate secretion. The importance of CFTR in bicarbonate secretion was first demonstrated in CFTR-deficient knockout mice.172-175 The same abnormalities in duodenal bicarbonate secretion are also present in CF patients, which partially explains the lower postprandial pH (one or two units) in the proximal duodenum of CF patients compared with normal subjects.144,145 Therefore, CFTR-dependent duodenocyte bicarbonate secretion, and, likely, other mechanisms of alkaline secretion, are defective in CF and contribute to the inability to maintain normal proximal duodenal pH.146 Unlike the small bowel and the respiratory system, the CFTR defect in the colon cannot be compensated by any other chloride channel.176 Therefore, the defect in colonic function closely relates to the CFTR genotype. Lactase deficiency in patients with CF is not related to the disease entity per se but merely reflects a normal ethnic- and age-related phenomenon. Young children with CF often have elevated lactase values in comparison with agematched controls. This finding may be a consequence of pancreatic insufficiency with slower turnover of microvillus membrane hydrolases.177 Xylose absorption is normal in CF patients.178

Meconium Ileus

Meconium ileus is the presenting symptom in 10% to 20% of infants with CF and appears to be related, in part, to genotype.179,180 Meconium ileus rarely occurs in infants without CF but has been reported in infants with stenosis of the pancreatic duct or partial pancreatic aplasia, with Hirschsprung’s disease, and in infants with otherwise normal GI tracts as a familial occurrence or as an isolated incident. Pathology Uncomplicated meconium ileus characteristically demonstrates a narrow distal ileum with beaded appearance caused by waxy, gray pellets of inspissated meconium, beyond which the colon is unused.181 Proximally, the ileal wall is hypertrophied; it then becomes greatly distended with extremely sticky, dark green to black meconium. As many as half of the cases of meconium ileus are complicated by volvulus, atresia, or meconium peritonitis from extravasation of meconium into the peritoneal cavity after intestinal perforation; it may manifest clinically merely as intraabdominal calcifications, a meconium pseudocyst, generalized adhesive meconium peritonitis, or meconium ascites. Fetal volvulus and vascular compromise may cause atresia. Identification of the role of the meconium ileus modifier genes should clarify the pathophysiology of these various presentations.182 Radiologic Features Characteristic radiologic findings reveal unevenly distended loops of bowel with absent or scarce air-fluid levels.183 Small bubbles of gas trapped in the sticky meconium may be scat-

943

944

Section VII  Pancreas Table 57-6  Indications for the Sweat Test (Quantitative Pilocarpine Iontophoresis) Siblings with cystic fibrosis Chronic pulmonary symptoms Cough Recurrent respiratory infection Bronchitis Bronchiectasis Lobar atelectasis Failure to thrive (stunting of growth) Rectal prolapse Nasal polyposis Intestinal obstruction of newborn Meconium ileus Jaundice in early infancy Cirrhosis in childhood or adolescence Portal hypertension Adult males with aspermia or azoospermia Heat stroke Hypoproteinemia Hypoprothrombinemia

Table 57-7  Conditions Associated with Elevated Sweat Electrolyte Concentration Figure 57-6.  Meconium ileus. Radiograph from a barium enema study in an infant with meconium ileus demonstrating a microcolon as well as meconium in the distal ileum (arrows). Distended small bowel loops are also noted.

tered throughout the distal small bowel. Barium enema demonstrates a microcolon and may outline the obstructing meconium mass in the distal ileum (Fig. 57-6). Abdominal calcification reflects meconium peritonitis, and a meconium pseudocyst may displace loops of bowel. Clinical Features Meconium ileus classically manifests with signs of intestinal obstruction within 48 hours of birth in an infant who is otherwise well; complicated meconium ileus manifests earlier, and infants appear much sicker. Hydramnios is a common prenatal finding. A family history of CF is helpful in establishing the diagnosis. The increased frequency of meconium ileus in some families with histories of CF is strongly associated with a yet-to-be-identified modifier gene on chromosome 19.184 Dilated, firm, rubbery loops of bowel may be visible and palpable through the abdominal wall, particularly in the right lower quadrant. Sweat tests should be performed in all infants with meconium ileus, with jejunal or ileal atresia, or with volvulus (Tables 57-6 and 57-7). Results are likely to be positive in 30% of patients with meconium peritonitis and in 15% to 20% of those with atresia of the small intestine.183 Although occasional infants with meconium plug syndrome have CF, meconium plug syndrome and meconium ileus must be carefully differentiated. Treatment Meconium ileus was considered invariably fatal until 1948, when the first patients were successfully treated by surgery. More recent reports indicate a very low operative mortality, and long-term survival approaches 90% for uncomplicated meconium ileus.180 Various irrigating solutions have been used during the operation and postoperatively to dissolve and dislodge the abnormal meconium. N-acetylcysteine (Mucomyst), which reduces the viscosity of mucoprotein

Cystic fibrosis Ectodermal dysplasia Glycogen storage disease, type 1 Adrenal insufficiency Familial hypoparathyroidism Fucosidosis Pitressin-resistant diabetes insipidus Mucopolysaccharidosis Familial cholestasis syndrome Environmental deprivation syndrome Acute respiratory disorders (croup, epiglottitis, viral pneumonia) Chronic respiratory disorders (bronchopulmonary dysplasia, α1-antitrypsin deficiency)

solutions by cleaving disulfide bonds in the mucoprotein molecule, and polysorbate 80 (Tween 80), a mild industrial detergent and preservative, are now generally recognized as safe and effective. Nonoperative relief of obstruction with diatrizoate (Gastrografin or Hypaque) enemas is also possible and has virtually eliminated prolonged hospitalization and early respiratory complications for most infants with uncomplicated meconium ileus.165,180 However, watersoluble hypertonic enemas may cause dangerous fluid and electrolyte shifts, especially in small, sick infants and can cause colonic perforation. Complicated meconium ileus requires surgical therapy. A diagnostic barium enema should precede therapeutic Gastrografin enemas.180,183,185 Infants with CF and meconium ileus who survive beyond six months of age have the same prognosis as any patient with CF and do not tend to have more severe disease.

Distal Intestinal Obstruction Syndrome

Intestinal impaction and obstruction (distal intestinal obstruction syndrome) remain common and troublesome features in CF beyond the neonatal period. Pathogenesis Mechanisms other than inspissated intestinal sections and pancreatic achylia are probably operative in the patho­ genesis of the distal intestinal obstruction syndrome and

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood include undigested food residues; possible disturbances of motility; dilation of the bowel, leading to fecal stasis; and dehydration. Intussusception and, less frequently, volvulus may complicate the distal intestinal obstruction syndrome. The incidence of distal intestinal obstruction syndrome is estimated to be as high as 10% among CF patients, although more recent data indicate a prevalence of 3% or less.165,186-189 Distal intestinal obstruction syndrome may in fact be the presenting disease symptom. Clinical Features A spectrum of clinical conditions results from partial or complete obstruction of the bowel by abnormal intestinal contents, including (1) abdominal pain caused by constipation or fecal impaction, (2) palpable cecal masses that may eventually pass spontaneously, and (3) complete obstruction of the bowel by firm putty-like fecal material in the terminal ileum or right colon or both.186,189,190 Abdominal pain, typically recurrent and cramping in nature, is the most common symptom of the distal intestinal obstruction syndrome. This type of pain may be the only symptom, and it may persist for years before distinct obstructive symptoms occur. Insufficient doses or cessation of pancreatic enzyme replacement, recent or concomitant respiratory infection, and dietary changes have been incriminated as precipitating factors.191 Patients with inadequately controlled steatorrhea may be at higher risk for development of this problem.189 Frequently, however, symptoms occur without warning in patients receiving presumably adequate medical management. The distal intestinal obstruction syndrome should be suspected in any CF patient who has abdominal pain, a palpable mass in the right lower abdominal quadrant, or bowel obstruction. When no acute symptoms are present, the soft, indentable, nontender nature of the palpable fecal mass on examination of the abdomen may be a diagnostic aid. The plain radiograph of the abdomen characteristically shows the proximal colon and distal small bowel packed with bubbly appearing fecal material. The fecal bolus can be identified on barium enema but may have to be differentiated from a cecal neoplasm or appendiceal abscess. It is also important to consider the diagnosis of appendicitis in these patients. Treatment Once a surgical issue, uncomplicated distal intestinal obstruction syndrome now usually responds to medical management. A stepwise approach with therapeutic trials of more than one modality should be attempted in each patient before a consideration of surgery.181 Vigorous medical therapy includes regular oral doses of pancreatic enzymes and stool softeners, oral or rectal administration of 10% N-acetylcysteine, and Gastrografin enemas. Maintenance treatment with oral doses of N-acetylcysteine, increased doses of pancreatic enzymes, and lactulose has been used successfully to prevent recurrent episodes of the syndrome. Treatment of this disorder with balanced intestinal lavage solutions has also proven beneficial.187,192

Intussusception (see also Chapter 119)

Intussusception, most often ileocolic, is a complication of the distal intestinal obstruction syndrome reported in approximately 1% of patients with CF.165,193 Presumably a tenacious fecal bolus adherent to the intestinal mucosa acts as the lead point of the intussusception. Most of the patients present acutely with intermittent, severe, cramping abdominal pain, although some experience pain for several months before the diagnosis is recognized. Only 25% of the patients

note blood in their stools. Efforts should be made to reduce intussusceptions by using radiologic techniques. Intussusception has been reported as the presenting symptom of CF, and CF is a major cause of intussusception after infancy.

Rectal Prolapse (see also Chapter 124)

In the past, rectal prolapse in the setting of CF was quite common, with a frequency of about 20%. The Cystic Fibrosis Registry now reports this complication occurs in 1% to 2% of patients.194 CF accounts for about 11% of’ all cases of rectal prolapse.195 Onset of rectal prolapse is usually in the first few years of life, is often the presenting symptom of CF, and many times is recurrent. Patients in whom CF is diagnosed early in life are much less likely to experience rectal prolapse than those diagnosed later in life except when stools are voluminous. Additional factors thought to be responsible for the high rate of rectal prolapse in CF patients include frequent bowel movements, varying degrees of malnutrition, and increased intra-abdominal pressure secondary to coughing. Medical management is almost always successful, and adequate replacement of pancreatic enzymes typically results in rapid improvement. However, up to 10% of patients may require surgical correction.

Gastroesophageal Reflux

Up to 20% of patients with CF complain of heartburn or regurgitation.196 In adults, the incidence of gastroesophageal reflux symptoms may be as high as 80%. Although the overall incidence of esophagitis is not known, esophagitis has been documented in up to 50% of patients who have significant respiratory problems. Barrett’s esophagus has also been observed in numerous CF patients.197 It is important to recognize and treat gastroesophageal reflux in these patients, but it can be difficult because many of the complaints can be attributed to CF alone and are consequently ignored. Approaches to treatment should be the same as in any other patient population (see Chapter 43).

Cancer Risk

Until the early 1990s, the idea that CF could be associated with the subsequent development of cancer was controversial. One study including 712 patients found no increased risk,198 whereas a second of 412 individuals suggested an increased risk of pancreatic and small intestinal tumors.199 In a more recent, prospective study of 38,000 CF patients, these discrepancies appear to be resolved. The investigators documented an increase in tumors of the digestive tract, but did not observe an increase in the risk of cancer in relation to the general population for all types of cancer.200 Cancer tended to occur in the third decade of life and involved the esophagus, small and large intestines, stomach, liver, biliary tract, pancreas, and rectum. Their pathogenesis is uncertain, but an increased risk of pancreatic cancer has been seen in patients with chronic pancreatic inflammation from other causes including alcohol,201 hereditary pancreatitis,202 and TP.203 Indeed, pancreatic cancer arising in the context of chronic inflammation is being increasingly recognized.204 This heightened cancer risk should be kept in mind as the survival of persons with CF continues to increase. Adolescents and adults with unexplained complaints, especially relating to the abdominal organs, should be evaluated for occult malignancy.

Liver Disease

The frequency of hepatic abnormalities in CF has decreased since the 1950s, with newer surveys noting a prevalence of about 15%. According to older literature, hepatic involvement in CF varied from 20% to 50% of cases studied,

945

946

Section VII  Pancreas although only about 5% of CF patients developed cirrhosis and approximately 2% progressed to clinically apparent liver disease requiring treatment.132 More recent literature suggests that most patients with mild liver abnormalities do not progress and the high frequency of abnormal liver injury test results noted in infancy spontaneously resolves. Nevertheless, approximately 10% of patients develop some degree of cirrhosis, usually prior to or during puberty.205 Although no genotypic association between liver involvement and CF has yet been confirmed, a familial tendency to develop cirrhosis has been observed in some patients,206 and there is now strong evidence that liver disease is associated with a modifier gene. In addition, some risk factors may predispose patients to the development of biliary and liver problems. Such factors include neonatal liver disease, pancreatic insufficiency, and possibly human leukocyte antigen (HLA) class206,207 and meconium ileus.64 Malnutrition may also predispose patients to fatty liver and specific nutrient deficits (protein, fat-soluble vitamins, minerals, essential fatty acids, carnitine). However, in a recent longitudinal four-year study of 124 children with CF, despite the finding that 92% showed some evidence of liver abnormality (6% based on clinical exam, 42% based on elevated serum aminotransferase activity, and 35% based on an abnormal hepatic ultrasonography), liver abnormalities did not correlate with a decline in nutritional status.208 A study from Sweden, however, suggests that essential fatty acid deficiency is more common in CF patients with marked hepatic steatosis.205 Altered drug metabolism in CF209 is characterized mainly by increased hepatic clearance of drugs.210 Newer studies identify the frequency of liver involvement in CF as follows: palpable liver (11%), elevated levels of liver enzymes (2.4%), low serum albumin levels (7.4%), cirrhosis with portal hypertension (2.5%), fatty liver (7%), neonatal liver disease (6%), and palpable spleen (2.2%).194 A second study found cirrhosis in 28% of adults with CF, two thirds of whom had associated portal hypertension.211 The prevalence of liver abnormalities in patients with pancreatic sufficiency is markedly lower. Pathology Hepatic changes may be present at any age and may be progressive. Excessive biliary mucus associated with mild periportal inflammation and early fibrosis is common in infants less than 1 year of age. Focal biliary fibrosis, characterized by inspissated granular eosinophilic material in ductules, bile duct proliferation, chronic inflammatory infiltrates, and variable fibrosis, is uncommon in infants but is present in more than 20% of surviving children and adolescents (Fig. 57-7). In time, focal lesions coalesce in some patients and progress to multilobular biliary cirrhosis.212 Bile stasis within lobules is conspicuously rare beyond the neonatal period, even in advanced liver disease caused by CF. Cholestasis is not uncommon in neonates and young infants; it may be prolonged and associated with excessive biliary mucus and mild periportal changes. Approximately half the reported cases were associated with meconium ileus.213 Fatty liver, often independent of nutritional status, remains one of the most common hepatic abnormalities encountered in CF.212 Unexplained hemosiderin deposits in hepatocytes, as well as Kupffer cells, may be prominent in infants and persist beyond four to six months of age. Radiologic Features Although the plain film of the abdomen may suggest hepatomegaly, it should not be used for estimating the size and

Figure 57-7.  Liver biopsy specimen from a patient with cystic fibrosis. Note the periportal bridging fibrosis, cholangiolar proliferation, and marked steatosis. Masson’s trichrome stain. (From Pathology of the Liver. 4th ed. MacSween RNM, Burt AD, Portmann BC, et al, editors. Philadelphia: Churchill Livingstone, 2002.)

shape of the liver. It can help identify radiopaque gallstones, pancreatic calcification, or fecal retention. Upper GI series may indicate the presence of esophageal varices. Ultrasonography is the best method for identifying abnormalities of the liver in CF, and may well reveal valuable information regarding the liver parenchyma.214-216 Cirrhosis produces increased coarse echogenicity in many cases and, in some patients, an irregular liver margin. Fatty infiltration is associated with an increase in fine echoes within the liver, with marked attenuation of the ultrasound beam in comparison to normal. A dilated portal vein is indicative of portal hypertension.214,216 Enlarged hepatic veins may be seen as a consequence of congestive heart failure or poor venous outflow from the liver due to constriction of the inferior vena cava by an enlarged liver at or above the entrances of the hepatic veins. CT can also be valuable in assessing the liver parenchyma and is most useful before liver transplantation. The anatomy and morphology of the liver and spleen can be well visualized with MRI, and MRI easily demonstrates fatty infiltration of the liver. Further study is needed to define MRI’s value in identifying early changes of cirrhosis in CF patients. MRI with angiographic images can also be useful in assessing hepatic vasculature in the pretransplantation evaluation. Hepatobiliary scintigraphy with scanning agents derived from iminodiacetic acid (IDA) is the best functional test available for imaging bile flow and can provide valuable information about hepatocyte function, liver size, and the presence of gallbladder filling. Qualitative examination has incorporated the use of deconvolution analysis to measure the hepatic extraction fraction and the hepatic halfclearance time. The hepatic extraction fraction provides a quantitative measure of hepatocyte uptake of tracer to reflect hepatocyte function. The hepatic half-clearance time provides a quantitative measure of clearance of tracer from hepatocytes and bile flow through the ducts.217,218 Functional Abnormalities Tests of hepatic function in CF may be normal even in cases of overt cirrhosis.219 Serum enzymes reflecting hepatocellular injury may be moderately elevated and fluctuate over the course of the illness. Up to 20% of CF patients with

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood pancreatic insufficiency have elevated serum alanine aminotransferase (ALT) values and 40% to 50% of patients have intermittently increased serum aminotranferases. An elevated serum alkaline phosphatase activity is the next most common chemical abnormality indicative of hepatic involvement, and the high serum activities frequently noted in normal infants and children (mainly resulting from the bone isoenzyme) may conceal increased levels of the hepatic isoenzyme.220 Fasting bile acid levels are elevated in many CF patients, and this may be among the more sensitive measures of liver function in this disease. Prothrombin time (international normalized ratio [INR]) is usually normal but may become prolonged as a consequence of reduced dietary intake of vitamin K or suppression of bowel flora by antibiotics, independent of changes in other liver function tests. Hypoalbuminemia is noted in approximately 7% of patients as liver disease progresses. Bile acid metabolism is disturbed in patients with CF and exocrine pancreatic insufficiency.219,221,222 Fecal bile salt losses are high, often approaching those of patients with ileal resection (see Chapter 64). Pancreatic enzyme replacement improves fat digestion and absorption, thereby reducing fecal bile acid excretion and steatorrhea. The fractional turnover rate of the bile acid pool is increased and the total bile acid pool size diminished in the absence of pancreatic enzymes,222 whereas the biliary lipid composition and saturation index approach those of patients with cholelithiasis.221 Treatment with pancreatic supplements returns abnormal biliary lipid values toward normal. Clinical Features Evidence for liver disease in patients with CF is often subtle; a variety of symptoms can be the presenting complaint. Although there is a broad spectrum of liver disease in CF patients, there are three predominant forms: (1) neonatal cholestasis, manifesting with or without meconium ileus or intestinal atresia223; (2) fatty liver syndrome; and (3) cirrhosis, manifesting either as portal hypertension or as liver failure. Notably, asymptomatic increases in serum liver enzymes or abnormal ultrasonographic findings may be the only clinical manifestations. More commonly, hepatomegaly or splenomegaly is the initial indication of hepatic disease. Esophageal varices or ascites are additional manifestations of hepatic involvement in CF, which may precede evidence of functional impairment (hepatocellular failure) by many years. Examination of the liver and spleen by percussion and palpation should be performed at each clinic visit, and the size and character of these organs recorded. In patients with suspected liver involvement, the degree of liver dysfunction and injury should be assessed at least annually by tests of synthetic capacity and reflections of liver cell damage. Synthetic capacity of the liver is most readily measured through serum protein analysis (at least total protein and albumin) and prothrombin time (INR). Liver cell damage is reflected by increases in serum bile acids, bilirubin (direct and total), aspartate aminotransferase (AST) and ALT, and alkaline phosphatase. Elevated gamma glutamyl transpeptidase (GGT) may reflect liver damage when other enzyme levels are normal.220 Normal liver biochemistry tests at regular intervals have been proposed as a good negative predictor for liver disease.205 Patients with hepatomegaly, abnormal liver chemical test results, and/or abdominal pain, even those whose complaints are not focused in the right upper quadrant, should undergo an assessment of the status of the liver and biliary system. After routine laboratory screening, ultrasonography

is likely to be most valuable. Depending on the results, it may be appropriate to obtain scintigraphic studies, MRCP, or ERCP. Liver biopsy in CF patients should be undertaken when indicated by the clinical course. As with all patients with CF, nutritional status must be assessed regularly. Guidelines are available from the Cystic Fibrosis Foundation.224 Plasma carnitine levels also should be measured.225 Nutritional rehabilitation should be accomplished in all patients with liver and biliary disease to eliminate avoidable complications of malnutrition.226 Treatment The treatment of symptomatic liver disease in CF is a challenge and usually requires a team approach. Treatment of cholestasis is probably best accomplished with ursodeoxycholic acid (20  mg/kg/day), although controlled clinical trials have not yet confirmed that ursodeoxycholic acid can prevent the progression of liver disease. This agent also may improve liver function in patients with elevated serum aminotransferase levels who lack cholestasis. The benefits of ursodeoxycholic acid have been shown to be dose dependent, and, indeed, scintigraphically evident improvement in hepatobiliary excretory function has been observed with higher doses of this bile acid.218,227,228 Nutritional rehabilitation is required in patients in whom disease activity has produced malnutrition. Preventive nutritional management in patients with early liver involvement is indicated. Attention should be paid to the provision of adequate quantities of fat-soluble vitamins. Indeed, essential fatty acid deficiency has been suggested to contribute to liver damage in CF.229 In patients with cirrhosis, infections (spontaneous bacterial peritonitis and cholangitis) necessitate treatment with appropriate antibiotics. Encephalopathy should be treated with lactulose and antibiotics (e.g., neomycin, metronidazole, rifaximin). Portal hypertensive bleeding should be treated vigorously on diagnosis. Endoscopic banding (or sclerotherapy) of bleeding esophageal varices is the most effective and rapid form of therapy. Adrenergic beta blockers (e.g., propranolol, nadolol) have not yet been widely used in patients with liver disease and CF. Portosytemic shunts have been placed effectively in CF patients with portal hypertension. The same indications should be applied in CF as in any other disorder when deciding on shunt surgery. The distal splenorenal shunt is the procedure of choice. Prophylactic shunting for varices that have never bled is not recommended. If severe lung disease is not a contraindication for surgery and the clinical status of the patient is acceptable, end-stage liver disease in CF is an indication for liver transplantation.230

Gallbladder and Biliary System

The gallbladder and biliary tract are abnormal in approximately 25% of patients with CF, independent of age, clinical course, or hepatic pathology.214,231-233 “Microgallbladders” are found in 23% and stones or sludge in 8% of patients. Data from the Cystic Fibrosis Registry suggests that only about 2% of individuals with CF eventually require gallbladder surgery.194 Pathology Small gallbladders are commonly found, characteristically containing thick, colorless “white bile.” Mucus is present within the epithelial lining cells, and numerous mucusfilled cysts may exist immediately beneath the mucosa. The cystic duct may be atrophic or occluded with mucus. Obstruction of the hepatic or common ducts by mucous

947

948

Section VII  Pancreas plugs does not occur, but intraductal stones sometimes cause obstructive symptoms and predispose to cholangitis. Radiologic Features The plain file of the abdomen may demonstrate radiopaque gallstones. Ultrasound examination shows gallbladder size, content (sludge, gallstones, or bile), and wall thickness. It is excellent for determining dilation of the biliary tract and may help detect cholecystitis.214 Scintigraphy is valuable in delineating functional, biliary system abnormalities in patients with liver disease, although test results are usually normal in CF patients without liver disease. ERCP reveals bile duct abnormalities in some patients with liver disease, but the prevalence of lesions such as bile duct stenosis (approximately 1% to 10%) is probably lower than originally reported.232 It is said that these lesions are not apparent in patients with normal liver function. There are no studies reporting the use of MRCP in patients with CF although it seems reasonable that MRCP should replace diagnostic ERCP in this patient population as it has for others. MRCP should be considered in patients with one or more of the following: unexplained abdominal pain, evidence of biliary tract disease, an abnormal ultrasonogram showing intrahepatic bile duct abnormalities, and a cholangitis-like illness. MRCP may also reveal irregular filling defects throughout the biliary tree with cystic dilations of the intrahepatic bile ducts and intrahepatic cholelithiasis. Irregularities of the smaller proximal ducts have also been noted by ERCP, presumably caused by focal biliary cirrhosis and these abnormalities may be detected by MRCP as well. CT scanning is of additional value in the assessment of biliary abnormalities and may be useful prior to MRCP. Therapeutic ERCP should be reserved for patients with gallstone disease or biliary strictures amenable to stenting. Clinical Features Acute or recurrent episodes of abdominal pain, whether diffuse or localized to the right upper quadrant may be related to obstruction caused by sludge, to infection, or to common bile duct obstruction caused by pancreatic fibrosis or perhaps by sclerosing cholangitis. Jaundice and itching may indicate bile duct obstruction. Treatment In general, obstructive jaundice in infants with CF is not a surgical condition. Cholecystectomy is indicated in CF whenever clinical disease mandates it and lung disease permits. However, treatment with ursodeoxycholic acid may be indicated as the first form of therapy for gallstones in patients with severe lung disease. Endoscopic papillotomy may be indicated as a first invasive therapeutic maneuver.

Genital Abnormalities in Male Patients

The most striking changes in the male genital tract occur in the epididymis, the vas deferens, and the seminal vesicles. The rete testes are intact. Multiple sections of spermatic cord rarely show histologic patency at more than one level. In addition to these defects, there is a striking increase in abnormalities associated with testicular descent, such as inguinal hernia, hydrocele, and undescended testes. Approximately 97% of males are sterile as a result of these changes. These abnormalities are unique; they have not been noted in any other genetic disease. These defects may be found in male infants shortly after birth and may be useful in supporting the diagnosis of CF in atypical cases.

Evaluation of patients not clinically suspected of having CF, but who have congenital absence of the vas deferens have a high frequency of CFTR mutations.234,235 Up to 70% of men with the sole finding of congenital absence of the vas deferens have a detectable mutation in at least one allele of CFTR. Other work has suggested that alterations in transcription also may be associated with this defect, inasmuch as a mutation, 5T, which reduced functional messenger ribonucleic acid (mRNA) transcripts of wild-type CFTR, is found in high frequency in men with congenital absence of the vas deferens.234,235 This group of men, without other manifestations of CF, likely represents a very mild form of the disease. On the other hand, the rate of CFTR mutations in patients with primary testicular failure is not elevated, and CFTR gene mutation screening is not warranted for this condition.236

Nutritional Management of Patients with Cystic Fibrosis

In the routine clinical setting, the nutritional management of patients with CF is based on an assessment of nutritional requirements (see Chapter 4), considering age, height, weight and anthropometrics, and severity of lung disease, as well as anorexia, pancreatic insufficiency, and mucosal dysfunction.158,226,237 Ideally, a normal age-appropriate diet should be encouraged, with adequate pancreatic enzyme replacement therapy provided (with gastric acid suppression, if indicated) to achieve as normal a fat balance as possible.238 However, children with end-stage lung disease may require in excess of 150% of the recommended dietary allowances (RDA) for age for calories and protein to promote normal growth. High-calorie, high-fat, and liberal salt diets are also encouraged by many CF centers. Several studies have suggested that improved nutritional therapy improves or at least slows progression of the pulmonary disease.226 All centers now place an emphasis on nutritional intervention before severe malnutrition is evident. In 2005, the Cystic Fibrosis Foundation revised the nutrition classification guidelines to eliminate the use of percentage of ideal body weight (% IBW) to define nutritional failure based on recommendations from consensus committees.239-241 The guidelines were reviewed and updated in 2008.242 For children younger than 2 years of age, weightfor-length percentile should be maintained at or above the 50%. Up to 20 years, a body mass index (BMI) percentile equal to or greater than the 50th percentile was recommended. It was recommended that adult women should have a BMI greater than or equal to 22 and that adult men should maintain a BMI greater than or equal to 23. Patients whose growth does not meet these guidelines or who are unintentionally losing weight should undergo a careful nutritional assessment. Malnutrition in CF can result from a variety of factors that increase nutrient loss, reduce energy intake, and increase energy expenditure.237 Increased losses are primarily related to underlying pancreatic insufficiency but are also influenced by conditions such as poorly controlled diabetes mellitus, vomiting and/or regurgitation, excess intestinal mucus, and inadequate bile salt secretion. Energy intake can be affected both by disease complications and by psychosomatic issues.237 Severe respiratory symptoms can be accompanied by anorexia, nausea, and vomiting. GI symptoms or complications such as abdominal pain, gastroesophageal reflux with chest pain, anorexia, and vomiting can lead to reduced caloric intake. In some patients, clinical depression, physical fatigue, a disordered sense of smell (food is unappetizing), and altered body image can lead to reduced food intake. Increased energy expenditure also frequently accompanies the severe respiratory disease of CF and is likely related to variables

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood including chronic infections, fever, increased respiratory effort, and bronchodilator medications.237 The optimal dietary intake for a CF patient is greater than the RDA of healthy children and adults. Ideally, a normal diet for age should be encouraged, except that fat intake should represent 35% to 40% of ingested calories, balanced with adequate pancreatic replacement therapy.238,239 However, a daily calorie and protein intake between 110% to 200% of the standard RDA for age is necessary to promote normal growth, and children with endstage lung disease may require in excess of 150% of the RDA.239,243,12 Nutritional intervention begins with addition of high-calorie foods to the usual diet, and use of nutritional supplements.239 When these methods fail, enteral feedings should be started via nasogastric, gastrostomy, or jejunostomy tubes. The presence and severity of gastroesophageal reflux disease (GERD) symptoms may influence the decision on the preferred route. Standard formulas are usually well tolerated. Nocturnal infusion is encouraged to promote normal eating patterns during the day. Initially, 30% to 50% of the estimated caloric needs should be provided overnight. Very low-fat, elemental formulas may be used without enzyme supplements for patients with an enteral feeding tube, and should be given by continuous infusion.239 Pancreatic enzyme supplements taken orally in the usual premeal dose are recommended before all nocturnal enteral feedings if nonelemental formulas are used. Patients receiving enteral feedings should be monitored for carbohydrate intolerance on at least two separate nights by measuring blood glucose levels two to three hours into the feeding and at the end of the feeding. Insulin may be required to prevent hyperglycemia, with adjustment of the insulin dosage during acute illness, glucocorticoid therapy, or other changes in health status. A great deal has been written regarding defined-formula diets as supplements or replacement for food in patients with CF. Although there is no evidence that these definedformula diets are superior to a balanced diet in providing appropriate protein, energy, and essential nutrients, liquid formulas are easy to administer by tube and may provide added nutrients when infused at night. Nutritional status should be followed carefully, and therapy instituted early. Some adolescents learn to pass soft Silastic feeding tubes nightly in order to administer nasogastric feedings. Gastrostomy feedings may be preferred by some families and patients, especially in younger children, for chronic administration of enteral supplements. Currently, gastrostomy or jejunostomy feedings are instituted at the first sign of nutritional failure.226,244,245 Finally, in some cases, parenteral nutrition may be necessary, but it should be reserved for acute support with a return to some form of enteral nutrition as soon as possible.

Prognosis

More than 50% of patients with CF survive to 36 years of age.194,246 Most significant morbidity and mortality are related to chronic obstructive pulmonary disease.194 The relative influence of nutritional support, pancreatic enzyme replacement, and aggressive treatment of pulmonary disease in improving the quality and duration of life remain under study. Patients with intact pancreatic function have better pulmonary status than those with pancreatic insufficiency,179 which suggests that there is a heterogeneous form of the disease (consistent with evolving genetic information) and/or that survival is longer with better nutrition and treatment. As survival improves, the problems facing these patients will likely differ and will be begin to spill over into the

domain of caretakers predominantly focused on the issues of adults. These medical problems will include such entities as pancreatitis, continued difficulties with adequate nutrition, cirrhosis with portal hypertension, diabetes with its long-term complications, osteopenia, and reproductive issues, as well as all of the more common problems seen in childhood.247,248 In a report of the Cystic Fibrosis Registry, gallbladder disease (0.9% of patients), peptic ulcer disease (0.7%), pancreatitis (0.8%), and cirrhosis with portal hypertension (1.2%) were more common in adults than in children. Pulmonary disease is more severe in adults than in children and malnutrition continues to be a problem in about 35% of adults with CF.194 Increasingly, these patients will require evaluation for potential malignancies of the digestive tract, evaluation for liver disease, or for other complications that will necessitate the specialized attention of a gastroenterologist. It is possible that in the near future, the prognosis of these patients will be changed dramatically by new therapies aimed at recovering specific functions through either drug treatment or gene therapy.

HEREDITARY PANCREATITIS

Hereditary pancreatitis is a syndrome of recurrent acute pancreatitis often leading to chronic pancreatitis that develops in an individual from a family in which the pancreatitis phenotype appears to be inherited through a diseasecausing gene mutation expressed in an autosomal dominant pattern.9,249 The most common cause is a mutation in the cationic trypsinogen gene (PRSS1) that appears to cause a gain-of-function through altering the regulatory domains usually controlled by calcium. (The details the PRSS1 gene mutations were presented earlier [see Fig. 57-2].) The majority of documented kindreds with PRSS1 mutations are from the United States and Europe, with a few families in Japan and South America but none identified in southern Asia. Most but not all kindreds with autosomal dominant appearing inheritance pattern of pancreatitis have PRSS1 mutations. In two large representative studies11,250 19% and 35% of pancreatitis-affected patients in hereditary pancreatitis families had no identifiable PRSSI mutations, suggesting that other genes or factors may be responsible for the high risk of pancreatitis in these families.

Clinical Features

The phenotypic features of hereditary pancreatitis caused by PRSS1 mutations are confined to the pancreas because the pancreas is the primary site of trypsinogen expression. Several studies suggest that there are small differences in the clinical features of hereditary pancreatitis depending on the genotype, with PRSS1 R122H being slightly more severe than N29I and patients in whom no gene mutation can be identified.10,11,251 Although about 20 disease-associated PRSS1 mutations have been identified, only the PRSS1 R122H and N29I are seen frequently enough to determine modifier-independent characteristics. Acute Pancreatitis The primary clinical phenotypic feature is recurrent acute pancreatitis. The severity of attacks is variable, with severe cases resulting in all of the complications commonly seen in other forms of acute pancreatitis. The patients develop typical epigastric abdominal pain, nausea, and vomiting, with elevated serum amylase and lipase levels. Some families appear to have more severe attacks, with nearly 90% reporting more than five hospitalizations10; an unusually high incidence of major complications such as splenic vein thrombosis was seen in another family.252 However, in other families the phenotype may be mild with attacks of pain

949

Section VII  Pancreas characterized as “of nuisance value only.”253 In Europe, patients reported that the vast majority of attacks lasted less than seven days and on the average had 2 (R122H) or 1.4 (N29I) attacks per year. The hospital admission rate was significantly greater with the R122H mutation (0.33 per year) than with the N29I mutations (0.19 per year).11 An uncommon finding is prolonged, persistent, or smoldering acute pancreatitis in which the patient may remain hospitalized for weeks or months. This problem has not been systematically studied and there are no clear treatment recommendations. Hereditary pancreatitis affects both sexes equally.11 The age of symptom onset is much younger than in most other causes of acute pancreatitis. The first mutation-specific studies reported an age of onset before 5 years in 58% of subjects with R122H mutations,251 but in only 27% of subjects with N29I,10 suggesting that the N29I mutation was slightly milder with fewer hospitalizations. However, patients with N29I mutations had more surgery than R122H patients, although their average age was older. A multicenter European study (EUROPAC) of 418 subjects from 112 families found that the median age of onset for subjects was 10 years with PRSS1 R122H, 14 years for N29I, and 14.5 years for patients with no identified mutations.11 The penetrance of the phenotype in gene mutation carriers is incomplete. Disease penetrance has been consistently reported at approximately 80%.251,253-255 However, there is no maximal age of disease onset, so that the cumulative risk of disease symptoms may be higher, with 96% penetrance by age 50 years.11 The apparent incomplete penetrance and variable expression appear to be determined by genetic and environmental factors,254 with earlier age of onset and more severe clinical course seen in patients with multiple mutations (e.g., PRSS1 plus SPINK1).256,257 The treatment of an acute attack of pancreatitis is currently identical to the treatment of nonhereditary pancreatitis (see Chapter 58). The best approach is to prevention recurrent attacks as much as possible. Multiple small meals, avoidance of fatty meals, and use of antioxidants and vitamins has been used as one approach.2,258 A small open-label trial using antioxidants and vitamins appeared to reduce the number of days of pain attacks in one hereditary pancreatitis family.259 More definitive studies will be needed with antioxidants and other treatments before clear recommendations become available. Chronic Pancreatitis Chronic pancreatitis is a process that is initiated by attacks of acute pancreatitis and characterized by inflammatory destruction of the normal parenchyma and progressive fibrosis.1 Reasons that some individuals with recurrent acute pancreatitis progress to chronic pancreatitis and others do not remains unclear, although there is a clear relationship between the number of attacks and the degree of fibrosis. The onset of chronic pancreatitis is currently impossible to determine, but the time between the onset of symptoms (usually acute pancreatitis) and exocrine failure can be determined (Fig. 57-8), although the moment a patient transitions from nonfailure to failure is also poorly defined in patients and between studies. In the EUROPAC study, the cumulative risk of pancreatic exocrine failure was 2.0% at 10 years of age, 8.4% at 20 years, 33.6% at 40 years, and 60.2% at 70 years of age.11 Thus, the exocrine failure associated with end-stage chronic pancreatitis progressed at substantially different rates than the onset of symptoms, and in only in a subset of affected patients. An unexpected observation in Europe was that surgery was much more common in female than male hereditary

100

% of affected patients (cumulative)

950

75

50

25 First symptom Exocrine failure Endocrine failure 0 0

10

20

30

40

50

60

70

Age of onset Figure 57-8.  Age of symptom onset in patients with hereditary pancreatitis. Subjects with hereditary pancreatitis due to the PRSS1 R122H mutations usually have an early age of onset (before age 20), although first symptoms can occur at older ages. A subset of subjects develops exocrine and/or endocrine failure, usually 20 to 30 years after the first symptom. (From Howes N, Lerch MM, Greenhalf W, et al. Clinical and genetic characteristics of hereditary pancreatitis in Europe. Clin Gastroenterol Hepatol 2004; 2:252-61.)

pancreatitis patients by age 50 years (24.3% vs. 10.5%), and more common with the N29I mutation (compared with R122H or no mutation) by age 50 years (34.7% vs. 12.6% and 13.2%)11 as previously suggested in the United States.10 Currently, the treatment of chronic pancreatitis and associated complications is identical to the treatment of chronic pancreatitis from other etiologies (see Chapter 59). Diabetes Mellitus Multiple factors contribute to the development of diabetes in these patients. Certainly, the destruction of islet cells is a major factor, but glucose intolerance also reflects peripheral insulin resistance, diminished ability in some patients for the remaining islet cells to compensate for diminished islet cell mass, and other genetic and environmental factors that are commonly associated with type 1 and type 2 diabetes in the general population. Some of these factors appear to influence the age of onset of diabetes in subjects with hereditary pancreatitis (Hanck and Whitcomb, 2002 unpublished). In the EUROPAC study the cumulative risk of endocrine failure was 1.3% at age 10 years, 4.4% at 20 years, 8.5 % at 30 years, and 47.6% by age 50 years (see Fig. 57-6).11 The cumulative incidence of endocrine failure continue to increase after age 50 years, especially in subjects the N29I variant.11 Pancreatic Cancer As discussed in Chapter 60, there are many reports of an increased incidence of pancreatic cancer in patients with hereditary pancreatitis.202,260 Pancreatic cancer appears to develop about 30 to 40 years after the onset of pancreatitis.202,261 The estimated accumulated risk of pancreatic cancer by age 70 in these families is about 40%,202,260,262,263 although it was slightly lower in the EUROPAC study.11

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood The reason for the high incidence of pancreatic cancer is unknown. The PRSS1 gene does not appear to play a role in sporadic pancreatic cancer,264 and our current knowledge of trypsin biology provides no rationale for how trypsin may act as an oncogene or other cancer-related factor. Rather, the recurrent pancreatic injury caused by unregulated trypsinogen activation, and subsequent inflammatory response appears to provide an environment that is oncogenic in nature.265 Treatment options for pancreatic cancer are very limited and the prognosis remains poor despite some recent advances. The most effective approach is prevention, with the recognition that the early onset pancreatitis in hereditary pancreatitis is one of the strongest known risk factors for the development of pancreatic cancer. The best target is tobacco smoking because it is a well known risk factor for pancreatic cancer—doubling the risk.266-268 Doubling (or halving) of risk becomes critically important when multiplied to an underlying risk of pancreatic cancer that is 50-fold the risk of the population. Indeed, in hereditary pancreatitis the age- and sex-adjusted odds ratio is doubled by tobacco smoking and the median age of diagnosis of pancreatic cancer is 20 years earlier in the smokers.202,261 Although hereditary pancreatitis families are clearly at high cancer risk, no effective screening methods have been established.260,269 The consensus guidelines of an expert panel suggested that if an individual in the pancreatic cancer risk age range was contemplating pancreatic surgery, a total pancreatectomy should be considered.270 This is based on the difficulty of early detection of pancreatic cancer in an anatomically distorted gland. Newer techniques for early detection and early diagnosis are being evaluated,271 but no controlled clinical trial results in hereditary pancreatitis have yet been reported.

Diagnosis

Prior to 1996, the diagnosis of hereditary pancreatitis was based purely on clinical criteria, including examination of the family pedigree.255 The discovery of the cationic trypsinogen gene mutation R122H opened the door to molecular diagnosis.1,9 Availability of a genetic test revealed that the clinical impression of mild cases was often wrong,251 and also that between 0% to 19% of patients presumed to have idiopathic chronic pancreatitis had hereditary pancreatitis– causing trypsinogen mutations.272-275 The phenotypic features were also clarified, in part because of the difference is proportion of the R122H and N29I mutations in different populations.11 Genetic testing for all of the trypsinogen mutations is now commercially available (Ambry Genetics, Irvine, Calif). Indications and limitations are discussed next.

Genetic Testing

Before clinicians order any test, they must determine the purpose of testing, have the experience to understand and interpret the test results, and anticipate how the results will guide patient management. These considerations are especially true for genetic testing because a genetic test result remains unchanged throughout the life of the patient, has implications for future descendants and other family members, and may affect social and reproductive choices, employment, and insurability.276,277 Thus, the clinician must fully comprehend the implications of testing, be prepared to provide pre- and post-test counseling to the patient (or refer the patient to a genetic counselor), and ensure that informed consent is obtained before testing.276,277 Reasons for cationic trypsinogen mutation testing also vary, but generally include verification of a clinical suspi-

cion, helping a patient understand or validate his or her condition, and assisting individuals at assessing risk of pancreatitis and eventually pancreatic cancer.2,202 This information may also be useful in making life decisions to minimize risk of disease (e.g., reproduction, diet, smoking).276 Identification of an established pancreatitis-associated gene mutation can be valuable in expediting an expensive and prolonged evaluation of recurrent pancreatitis in children, and may preclude further evaluation of elusive causes of pancreatitis in adults. Interpretation of Genetic Testing The positive and negative predictive value of a genetic test in identifying specific mutations is almost perfect with properly applied modern techniques. Interpretation of test results and explanation of their meaning to the patient continue to be pivotal issues because the test result has implications for the patient as well as the patient’s extended family. The prognosis for these patients can be outlined in general terms from the clinical discussion earlier, noting that there is significant variability and the effectiveness of future treatments in preventing side effects is unknown. Finally, the mutation-positive individual has a 50% chance of passing on the mutation to each child. A positive test result in clinically unaffected person is interpreted as an increased risk of pancreatitis, with this risk possibly diminishing with age. A negative test result in a family with a known mutation in the PRSSI gene essentially eliminates the risk of this genetic form of pancreatitis. If a mutation has not been previously identified in the family, then a negative test result in an unaffected person is considered noninformative because one cannot distinguish whether the tested individual is free from genetic risk or whether he or she has inherited a different pancreatitis predisposing gene mutation.2 Up-to-date counseling information should always be available through the commercial genetic testing laboratory. Genetic Testing of Children The genetic testing of children raises unique issues. Unlike an adult patient, a child legally cannot provide informed consent. Thus, the decision for a child is essentially left to the parents or legal guardian. For children seven years and older, a parent or legal guardian may provide consent for genetic testing, although these older children should also provide assent, or agreement to the testing.278-280 The primary reason for testing children for cationic trypsinogen gene mutations is to assist in determining the cause of unexplained pancreatitis or to confirm suspected pancreatitis in a child at risk for hereditary pancreatitis, thereby limiting further investigations. The testing of purely asymptomatic children is strongly discouraged because currently there is no clear medical benefit in identifying carriers at a young age.276,281 Testing for the purpose of intervention with diet, medication, or surveillance for complications of a genetic disorder has been advocated.281 Because alcohol consumption, emotional stress, and fatty foods have been reported to precipitate pancreatitis attacks,253 and because smoking increases the risk of pancreatitis282-284 and pancreatic cancer,262,267,285 testing for the purpose of encouraging mutation-positive older children to avoid these excesses is advocated by some caregivers. However, the avoidance of fatty foods, alcohol, and tobacco represents excellent general health advice for all children and, therefore, provides no compelling reason for testing.276 In either case, the personal desires of older children to postpone testing or to proceed with testing to relieve their own anxieties and learn more about their own health must also be carefully

951

952

Section VII  Pancreas considered.280 Ownership of test results in children must be addressed.2

FAMILIAL PANCREATITIS

Familial pancreatitis refers to pancreatitis from any cause that occurs in a family with an incidence that is greater than would be expected by chance alone, given the size of the family and incidence of pancreatitis within a defined population.286 Therefore, familial pancreatitis may or may not be caused by a genetic defect. Hereditary pancreatitis (discussed previously) is therefore a subset of familial pancreatitis in which the pancreatitis phenotype follows an autosomal dominant inheritance pattern. Familial pan­ creatitis is also seen in association with other inherited conditions, such as hypertryglyceridemia and hyperparathyroidism (see following). The etiologies of pure familial pancreatitis are becoming evident and include autosomal recessive conditions as well as complex genetic disorders with gene-gene or gene-environmental interactions. Familial pancreatitis in which the initial symptoms occur at a young age (e.g., less than 20 years) is more likely associated with a strong genetic risk. The most common causes are autosomal recessive disorders associated with homozygous or compound heterozygous CFTR mutations (atypical CF) or SPINK1 mutations. The important distinction should be made through genetic testing because patients with atypical CF have other organ systems at risk.59 As noted, the diagnosis of any form of CF should be made with great caution and under rigorous protocol because of the implications for the patient and family in terms of the more intense and chronic medical interventions that will be required, as well as the significant social and financial considerations. That said, the appropriate diagnosis must be made. Early onset familial and idiopathic pancreatitis is also caused by a combination of heterozygous CFTR and SPINK1 mutations as part of a complex genetic trait. The median age of onset of pancreatitis is about 13 years of age with an equal number of men and women. The clinical course and treatment recommendations for this disorder remain the same as that of other types of chronic pancreatitis.

TROPICAL PANCREATITIS

TP is a form of early onset idiopathic chronic pancreatitis with unique epidemiological and clinical features.3,119 TP is generally characterized by recurrent abdominal pain, pancreatic calculi, and diabetes mellitus, occurring mostly among poor children and young adults inhabiting many developing nations.119 However, there is no consensus on diagnostic criteria to distinguish TP (including TCP and FCPD) from other forms pancreatitis and perspectives on the epidemiology, clinical characteristics, and prognosis are heavily influenced by referral bias (e.g., those presenting to gastroenterologists for pain, to surgeons for severe pain, or to endocrinologists for diabetes mellitus). Furthermore, the age of onset and features of the “typical” patient have changed over the past 50 years, possibly due to improvements in sanitation and better nutrition.3,119,287 The current consensus is that the age of onset is older than in earlier reports, the nutritional status is usually normal at the onset of symptoms, and the distribution is well outside of tropical regions (e.g., northern India).

Clinical Features

The most striking feature of TP is the strong propensity for diabetes mellitus to develop well before exocrine failure, marked calcifications in a grossly dilated main pancreatic duct, and in many patients, pancreatic atrophy. The severity

of diabetes appears to correlate with the degree of calcification, suggesting that this represents a clearly different form of pancreatitis and that the pathophysiologic mechanisms are linked. The dilated main pancreatic duct and pancreatic atrophy may be due, in part, to proximal duct obstruction by stones.119 The variation seen in gross pathologic reports spans from complete fatty replacement of the pancreas to severe atrophy resulting in a thin membrane—like a bag filled with stones—to chronic pancreatitis that is indistinguishable from cases of alcoholic or hereditary chronic pancreatitis. In tropical calcific pancreatitis (TCP), the presenting complaint is usually abdominal pain. Initially the abdominal pain episodes last for days and often are aggravated by small amounts of food so that patients refuse to eat. In the early stages, the bouts of pain are severe and are associated with vomiting.119 This pain is typical of recurrent acute pancreatitis. Some patients develop severe pain late in the course of the disease associated with an inflammatory mass in the head of the pancreas or other features. This characteristic type of pain is similar to B-type pain described by Ammann and colleagues288 in alcoholic chronic pancreatitis, and remains resistant to all but the most aggressive treatment including major surgery. Diabetes mellitus develops in about half of these patients by age 50 years, about 10 years after the initial onset of pain.119 In fibrocalculous pancreatic diabetes (FCPD), the patients are often first diagnosed after referral to an endocrinologist for diabetes. On investigation at this stage of the disease, the pancreas is often atrophic, with a grossly dilated main pancreatic duct that is filled with large, calcific stones, but the patient reports minimal pain. Although many patients with TCP appear to have diabetes from loss of pancreatic islets, patients with fibrocalculous pancreatic diabetes have preserved alpha-cell function.4 The destructive process is not associated with typical anti–beta-cell antibodies.289 Early pathology reports in severe cases of TP with marked pancreatic atrophy also noted marked islet cell hyperplasia (nesidioblastosis) composed primarily of beta cells, but it is unclear if the beta-cell mass remains insufficient (i.e., a few hyperplastic islets in the context of massive loss) or if there is a confounding deficiency or defect in insulin release. The diabetes is often brittle and difficult to manage, even though it is seldom associated with ketoacidosis.3 This latter problem may be related in part to erratic carbohydrate absorption that can be corrected with adequate oral enzyme replacement therapy.290

Etiology

Early attempts to determine the etiology of TP focused on protein and carbohydrate malnutrition and environmental factors such as eating cassava melon. Although there appears to be a weak epidemiologic association, these factors cannot be the primary cause because most individuals with TP are not malnourished and the effect of cassava is weak at best.291 There are some mechanistic similarities between TP in southern Asia and idiopathic chronic pancreatitis in Europe and North America. In both diseases, a significant fraction of subjects have SPINK1 mutations, and especially the N34S haplotype. Interestingly, this high-risk haplotype was only observed in a subset of children in Germany79 or families in North America,83 with the phenotype of the heterozygous and homozygous being identical.83 Because these mutations appear to lower the threshold for intrapancreatic trypsin activation, it appears that trypsin-related injury is a component of both of these disorders. The association between SPINK1 mutations and TP was first recognized in Bangladesh in 2001, including in fibrocalculous pancreatic diabe-

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood tes and TCP,85,292 a finding confirmed in India in 2002,293 and extended by additional studies.294-296 Of note, a subset of patients with diabetes mellitus but without evidence of exocrine pancreatic disease also had SPINK1 mutations in Bangladesh, a finding that does not appear in diabetes populations tested in the United States.295 This finding suggests that the primary lesion is recurrent acute pancreatitis, and there is a spectrum of complications, include varying degrees of fibrosis, calcification, diabetes mellitus, and pain, that defines the varying phenotypic features.

recurrent infections. Because of the variable expression of pancreatic, hematologic, and other features, the diagnosis in mild cases may be delayed.300,304 Several hundred families, most with a single affected member, have been identified.301 The discovery of the SBDS gene and genotype-phenotype studies have confirmed the variable phenotypic features that were previously described based on family studies. SDS remains the second most frequently recognized cause of pancreatic insufficiency in children after CF.52,128,298

Treatment

Clinical Features

The fundamentals of treatment of TP are similar to other forms of chronic pancreatitis. In earlier stages of disease, pancreatic duct decompression and pancreatic enzyme supplements have been used for symptomatic relief,119,290 but adequate studies to prove their effectiveness are lacking. Oral pancreatic enzyme supplements remain useful for treating maldigestion and can improve glycemic control in patients with diabetes mellitus, presumably through providing a predictable pattern of nutrient absorption.290 The problem of unrelenting pain in TP is similar to other forms of chronic pancreatitis in which surgery is indicated, but not always fully effective in pain control.3,119

SHWACHMAN-DIAMOND SYNDROME

SDS (OMIM 260400122) is a rare autosomal recessive disorder associated with mutations in the SBDS gene (see earlier) and characterized by exocrine pancreatic insufficiency, hematologic abnormalities such as cyclic neutropenia, skeletal defects, short stature, and normal sweat electrolytes.52,297-301 Among these features exocrine pancreatic insufficiency and cyclic neutropenia are present in most patients.300 Myelodysplastic syndromes and acute leukemias develop in up to a third of patients302,303 and numerous other features have been reported (Table 57-8). Severe cases present in infancy with malabsorption, failure to thrive, or

Table 57-8  Clinical Features of Shwachman-Diamond Syndrome Pancreatic

Exocrine pancreatic hypoplasia Steatorrhea

91%-100% 55%-88%

Hematologic

Neutropenia Anemia Thrombocytopenia Pancytopenia Leukopenia Elevated fetal hemoglobin Myelodysplastic syndromes Leukemia

88%-100% 42%-66% 24%-34% 44% 52% 80% 8%-33% 12%

Skeletal

Metaphyseal dysostosis Long-bone tubulation defects Short or flared ribs Thoracic dystrophy Others

44% * * 32% <5%

Growth Other

Short statue (normal growth velocity) Psychomotor delay Mental retardation Renal tubular dysfunction Diabetes mellitus Dental abnormalities Ichthyosis Hepatomegaly Abnormal liver biochemical tests Myocardial abnormalities

Common Common 33% * <5% * Reported <5% Common 50% (autopsy)

*Frequency unknown.

Pancreatic Insufficiency The clinical features of SDS usually become evident in the first year of life.52,300,304,305 Severe pancreatic insufficiency, steatorrhea, and failure to thrive are frequent presenting symptoms.300,304 A normal sweat-chloride concentration or other normal measures of CFTR function distinguish SDS from CF.52 Serial assessments of exocrine pancreatic function reveal persistent deficits of enzyme secretion, but nearly half of patients showed moderate age-related improvements (more than four years of age) leading to pancreatic sufficiency,300,304 with some pancreatic digestive enzymes (e.g., trypsin) improving more than others (e.g., amylase). The pancreas itself may be small or even of normal size, but the acinar cells appear to have undergone fatty replacement.52 The extensive lipomatous changes result in characteristic changes during abdominal imaging by CT scan (Fig. 57-9), MRI, or ultrasound.306-308 Bone Marrow Dysfunction Neutropenia-related infections are also an early problem and are severe in at least 85% of patients, occasionally leading to death.128,298 The neutropenia occurs in a cyclical fashion in two thirds of patients, and, when tested, the neutrophils appear to have impaired chemotaxis.302 However, one patient with severe neutropenia and recurrent infections was successfully treated with granulocyte colony-stimulating factor.309 Unidentified serum factors may also impair immune function.310 Common infections include otitis media, sinusitis, pneumonia, osteomyelitis, urinary tract infections, skin infections, and lymphadenitis.311 Thrombocytopenia and anemia are also frequently seen.

Figure 57-9.  Computed tomographic appearance of the pancreas in a patient with Shwachman-Diamond syndrome. Note that the pancreas retains a typical size and shape, but it is highly fatty and therefore appears as a very low-density structure. (Figure courtesy of Professor Peter Durie.)

953

954

Section VII  Pancreas Patients with hypoplasia of all three bone marrow cellular lines have the worst prognosis.300 Thus, the median age of survival for all patients with SDS is 35 years, but patients with pancytopenia have a median life expectancy of only 24 years. Pancytopenia appears with a mean age of onset of 6 years and occurs in 10% to 25% of patients.311 Up to a third of patients will develop myelodysplastic syndrome (MDS) and about 10% to 25% will develop acute myeloid leukemia (AML) or other leukemias.300,302,304 This is illustrated by a French study reporting that 8 of 71 persons with SDS developed MDS and/or leukemia over a 10-year period.312 Growth and Development The birth weight of children with SDS is typically low (2.9 ± 0.5  kg, 25th percentile300), and by six months of age the mean heights and weights are characteristically below the 5th percentile. Thereafter, growth velocity appears normal.300 The short stature is independent of nutrition.52 Some clinically evident skeletal abnormalities may be present. For example, metaphyseal chondrodysplasia and dysostosis may be evident radiologically in 44% of patients, especially in the femoral head and proximal tibia.52,128 Thoracic dystrophy, short flared ribs, and other skeletal abnormalities also have been described.52,300,313 Most patients remain below the 3rd percentile for height and weight although some adults reach the 25th percentile for height.52,128 Although men and women are probably affected equally, men with mild disease and short stature are more likely to undergo thorough investigation than women, leading to a mild ascertainment bias.301,304

Molecular Pathology

The molecular defect in SSD is a novel gene of unclear function called the Shwachman-Bodian-Diamond gene (SBDS) (see earlier). The pancreatic lesion appears to result in developmental failure of the pancreatic acini in utero.52 Macroscopically the pancreas appears fatty and may be small or normal size. The main pancreatic ducts and islets are normal. Microscopically there is extensive fatty replacement of the pancreatic acinar tissue.52 Likewise, the hematologic disorder appears to affect cellular development and involves the progenitor cells and bone marrow stroma needed to support hematopoiesis.303 The defect appears to affect multiple specific lineages, although the hematologic deficit can be cured by bone marrow transplantation,310 or improved by granulocyte colony-stimulating factor.309 The long bone abnormalities in about 40% of patients with SDS appear to involve protein processing in the rough endoplasmic reticulum of cartilage chondrocytes, but the exact defect remains obscure.309,314 Although a number of specific cytogenetic abnormalities have been reported,302,315-317 none has been consistently observed. A recent molecular genetics study demonstrated that the majority of patients that were classified as having SDS according to rigorous clinical criteria had compound heterozygous mutations of the SBDS gene and the absence of full-length SBDS protein in leukocytes. However, a subgroup of patients had no SDBS gene mutations and had full-length SBDS protein in leukocytes, suggesting that SDS is heterogeneous.318

Treatment

The treatment of pancreatic exocrine deficiency is more straightforward with SDS than with CF because bicarbonate secretion in the pancreas and duodenum is spared. Optimal pancreatic enzyme replacement (500 to 2000 units lipase activity per kilogram before each meal, and half as much with snacks) should be initiated with an expectation of

diminished steatorrhea and improved weight gain, but not necessarily enhanced growth.128 Fat-soluble vitamins, medium-chain triglycerides, and other high-calorie supplements may be needed, as discussed for CF. During periods of granulocytopenia, febrile episodes should be evaluated and treated with antimicrobial drugs. Anecdotal information suggests that granulocyte colonystimulating factor can be used in patients with severe neutropenia who have suppurative infections.309 In the case of those with recurrent respiratory infections, humoral immunologic defects should also be considered. Episodes of bleeding or severe anemia may necessitate transfusion. Hip disease should be monitored, with surgical intervention if progression occurs. The use of recombinant human growth hormone in this condition has not been systematically investigated, but anecdotal reports have shown efficacy in accelerating growth.

RARE GENETIC SYNDROMES WITH PANCREATIC PATHOLOGY JOHANSON-BLIZZARD SYNDROME

Johanson-Blizzard syndrome,319 or nasal alar hypoplasia, hypothyroidism, pancreatic achylia, and congenital deafness syndrome (OMIM 243800122) is a rare autosomal recessive syndrome linked to a mutation in the ubiquitin-ligase E3 (UBR1) gene.320 About 100 cases have been identified, and the syndrome appears to be an autosomal recessive disorder because it clusters in highly inbred families. It is characterized by pancreatic insufficiency and growth restriction with lipomatous transformation of the pancreas.128 Ubiquitin-ligase E3 belongs to the N-end rule pathway of an intracellular protein degradation system. Digestive serine proteases, which are known to trigger autodigestion during pancreatitis, were detected as specific substrates of UBR1 and are upregulated in the pancreas when UBR1 is defective due to loss-of-function mutations.320 The predominant and most consistent human phenotype, pancreatic exocrine insufficiency, is associated with an intrauterine pancreatitis, resulting in complete gland destruction. The murine knockout of UBR1 exhibits features that are very similar to the human phenotype, with restricted growth, craniofacial abnormalities, and pancreatic exocrine insufficiency. The acinar cells have an impaired cholecystokinin (CCK) receptor-stimulated excitation-secretion coupling and are more susceptible to experimental pancreatitis than wildtype animals and to a more severe systemic inflammatory response in the disease course. These findings are consistent with intracellular failure to degrade various cytosolic digestive enzymes. Patients with Johanson-Blizzard syndrome have preservation of ductular output of fluid and electrolytes, like patients with SDS and unlike patients with CF.321 They also have decreased acinar secretion of trypsin, colipase and total lipase, and low serum immunoreactive trypsinogen levels, consistent with a primary acinar cell defect.321 Histologically the pancreatic ducts and islets are preserved but are surrounded by connective tissue and a total absence of acini.322 In addition to pancreatic acinar cell defects, the syndrome is characterized by thyroid dysfunction, aplastic alae nasi, cardiac anomalies, genitourinary malformations, deafness, midline ectodermal scalp defects, dental anomalies, and imperforate or anterior anus.319,323 There are no skeletal or hematologic abnormalities characteristic of SDS.321 The inheritance is thought to be autosomal recessive, although no specific gene defect has been identified.

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood PEARSON’S MARROW-PANCREAS SYNDROME

Pearson’s marrow-pancreas syndrome (MIM 55700) is a rare, autosomal dominant mitochondrial DNA (mtDNA) breakage syndrome324 characterized by refractory sideroblastic anemia with vacuolization of marrow precursors and exocrine pancreatic dysfunction.325 Patients with this syndrome may have transfusion-dependent macrocytic anemia in infancy, but all of the bone marrow cell lines otherwise appear to be normal. The pancreatic insufficiency appears to be due to pancreatic fibrosis rather than fatty replacement of the acinar cells as in SDS, and appear more likely to be associated with diabetes mellitus.326 The syndrome is generally fatal in infancy, and patients who survive develop progressive involvement of many systems, including the liver, kidney, gut, and skin, all of which have abnormal mitochondria.128 The molecular defect in Pearson’s marrow-pancreas syndrome was initially identified as a 4977 base pair deletion of mtDNA encompassing portions of the genes coding for NADH dehydrogenase, cytochrome oxidase, and adenosine triphosphatase (ATPase).327 A variety of mitochondrial defects have been identified,328,329 and these deletions appear to be flanked with nucleotide repeats.329 The clinical features and severity of disease appear to correlate with the organ distribution and proportion of abnormal mtDNA.328,330 Patients who survive may develop features of other mitochondrial DNA deletion syndromes such as Kearns-Sayre syndrome.331 No specific treatment to correct these abnormalities is yet known.

PANCREATIC AGENESIS (see also Chapter 55)

Agenesis of the pancreas (OMIM 260370122) is extremely rare, with about 10 cases reported in the literature.332,333 In 1 case, genetic testing identified a homozygous nucleotide deletion in the PDX1 (pancreas and duodenal homeobox 1) gene, resulting in premature termination of the PDX1 gene production. Thus, absence of PDX1 is a probable cause of pancreatic agenesis.334 A mutation in a second pancreas development gene, pancreas transcription factor 1 alpha (PTF1A) is also associated with pancreatic agenesis, but in this case the clinical features include cerebellar agenesis as well as pancreatic endocrine and exocrine dysfunction.335 The role of PTF1A in normal cerebellar development was confirmed by detailed neuropathologic analysis of PTF1A-/- mice.335 This condition is extremely rare, but demonstrates the essential role of PTF1A in pancreatic development. The clinical features of pancreatic agenesis include intrauterine growth restriction (probably from insulin deficiency), insulin-dependent diabetes, and pancreatic exocrine insufficiency. The differential diagnosis of pancreatic agenesis includes transient diabetes mellitus of the newborn, pancreatic hypoplasia, CF, SDS, Johanson-Blizzard syndrome, and other rare disorders. However, in pancreatic agenesis, the profound endocrine and exocrine deficiencies persist; serum C-peptide and glucagon levels are undetectable; and the pancreas is absent on imaging studies. These children are managed as having type 1 diabetes mellitus (treated with insulin) and severe pancreatic maldigestion (treated with pancreatic enzyme supplementation). Survival is possible with proper diagnosis and treatment.333,334

AGENESIS OF THE DORSAL OR VENTRAL PANCREAS

Agenesis of the dorsal pancreas (OMIM 167755122) is also extremely rare, with at least 15 cases reported.336 A case of fatty replacement of the pancreatic body and tail was reported,337 which may be a variant of this disorder. Unlike

complete agenesis of the pancreas, patients may be asymptomatic or may present with bile duct obstruction or pancreatitis. Agenesis and hypoplasia of the ventral pancreas is also extremely rare.338 The lack of the ventral and dorsal pancreas has been described in patients carrying mutations in the gene-encoding hepatocyte nuclear factor 1 homeobox B (HNF-1B).339

OTHER SYNDROMES

Numerous rare syndromes have been identified that affect the pancreas. Examples include asplenia with cystic liver, kidney, and pancreas (Iverson’s syndrome, OMIM 208540122), which describes dysplasia (in the sense of disturbed development) of the kidney, liver, and pancreas without other diagnostic abnormalities.340 Histologically the pancreas has dilated, large irregular-shaped ducts surrounded by concentric loose mesenchyme and prominent areas of fibrosis and atrophy of its parenchyma.340 A similar autosomal recessive renal-hepatic-pancreatic dysplasia has been described.341

ISOLATED ENZYME DEFECTS

Isolated defects of pancreatic enzymes are rarely identified. Furthermore, some doubt about the phenotype of some of the earlier cases of isolated pancreatic enzyme deficiencies has arisen with later studies demonstrating that the “isolated” cases of trypsin deficiency reported by Morris and Fisher342 and Townes343 were actually patients with Johanson-Blizzard syndrome.323 These issues will likely be resolved when the molecular causes of the various syndromes are discovered.

Lipase and Colipase Deficiencies

Congenital absence of pancreatic lipase (OMIM 246600122) is a rare disorder accompanied by variable preservation of other enzymes.344-346 The human lipase gene has been cloned347 and is located on the long arm of chromosome 10.348 However, the cause of the enzyme deficiency is unknown. To date no mutations have been identified in the human lipase gene of patients suspected of having lipase deficiency.349 Both men and women are affected. The earliest and most characteristic manifestation of this disease seems to be the passage of stool with an unusual amount of readily separable oil, which is often responsible for soiling. Failure to thrive is only occasionally noted, and systemic manifestations are absent. Pancreatic lipase activity within duodenal content is low to absent. Trypsin and amylase activity have been somewhat diminished in some patients; however, other parameters of exocrine function, including colipase and phospholipase A activities, and bicarbonate and volume secretion, are usually normal. Any residual lipase activity has been presumed to be a result of lingual or gastric lipase activity. Bile salt metabolism in this disease has not been extensively investigated. In addition to its functional absence, no immunologically active lipase can be detected,344 suggesting either the complete absence of pancreatic lipase or the occurrence of a major structural change affecting both immunogenicity and function. The biochemical response to exogenous pancreatic enzyme therapy is suboptimal, and limitation of dietary fat is often necessary to avoid oily stools and incontinence. The colipase gene is located on the long arm of chromosome 6.350 Colipase deficiency has been described in male offspring of both consanguineous and nonconsanguineous marriages.351,352 These patients present with loose stools and steatorrhea; growth and development are normal. Colipase

955

956

Section VII  Pancreas activity is markedly reduced with otherwise normal pancreatic enzyme secretion. Fat absorption improves dramatically with the intraduodenal instillation of purified colipase. In studies of patients with pancreatic insufficiency associated with CF and SDS, steatorrhea occurrs only when lipase and colipase secretion are diminished to less than 2% and less than 1% of mean normal values, respectively.353

Carboxyl Ester Lipase Mutations

Two Norwegian kindreds with diabetes mellitus and exocrine pancreas dysfunction were found to have mutations in exon 11 of the carboxyl ester lipase gene (CEL).354 Exon 11 encodes a variable number of tandem repeats in the carboxyl-terminus of human CEL. Both kindreds had autosomal dominant diabetes mellitus typically diagnosed before the age of 40 years. All mutation carriers had low fecal elastase levels and all 10 subjects tested had a low coefficient of fat absorption and decreased fat-soluble vitamin levels. None of the family members had a history of acute pancreatitis. One family member had an atrophied and fibrotic pancreas at autopsy. The histology revealed pronounced fibrosis and mucinous metaplasia. A second study showed that affected children demonstrated MRI evidence of lipomatosis of the pancreas well before the usual age when diabetes presented in relatives.354

Enterokinase Deficiency

Although few reports of congenital absence of enterokinase have appeared since the original description in 1969,355 a familial nature was suggested by its documentation in siblings.356 These patients presented with malabsorption, hypoproteinemia, and severe growth restriction. Evaluation included normal amylase and lipase activities and very low trypsin activity in the duodenum, with normal concentrations of sweat electrolytes. Luminal trypsinogen could be activated by the addition of exogenous enterokinase. Small intestinal morphology and disaccharidase levels were normal. Congenital enterokinase deficiency is recognized in 1% to 2% of infants undergoing evaluation of suspected pancreatic insufficiency.356 The steatorrhea associated with enterokinase deficiency may be related to a deficiency of phospholipase, the activation of which requires trypsin, which in turn is activated by enterokinase. Patients with CF and SDS have increased intraluminal but normal mucosal enterokinase activity.356 Enterokinase levels may also be reduced with significant mucosal injury. However, even in untreated celiac disease, normal mucosal and normal intraluminal enterokinase activities have been reported.357

FAMILIAL METABOLIC SYNDROMES ASSOCIATED WITH RECURRENT ACUTE AND CHRONIC PANCREATITIS Several familial metabolic syndromes have recurrent acute and chronic pancreatitis as a common feature. The most common include syndromes that involve states of recurrent hypercalcemia or hyperlipidemia.

FAMILIAL HYPERPARATHYROIDISM WITH HYPERCALCEMIA

Hypercalcemia is associated with acute pancreatitis, possibly through trypsinogen activation43 and trypsin stabilization (see Fig. 57-3).358-360 The relationship between hypercalcemia and pancreatitis became apparent in 1957

when Cope and associates361 suggested that pancreatitis may be a diagnostic clue to hyperparathyroidism. Shortly thereafter the relationship between familial hyperparathroidism and chronic pancreatitis was noted because three of nine family members with hyperparathyroidism had chronic pancreatitis.362 This relationship has been questioned by some363 and verified by others364,365 but is now accepted.366-369 Research into the CASR gene suggests that a combination of hypercalcemia and SPINK1 mutation, alcohol, or other factors is required to target the pancreas for recurrent acute and chronic pancreatitis.40-42

FAMILIAL HYPERLIPIDEMIA

Familial hyperlipidemia is often associated with recurrent acute pancreatitis. Incidence proportions for associated pancreatitis are approximately 35% for familial type I, 15% for type IV, and 30% to 40% for type V.128 The type of pancreatitis is usually acute and recurrent, although pancreatic insufficiency has been reported with types I and V.370,371 The mechanism whereby high serum triglyceride levels lead to pancreatic injury is unknown, although the most popular and well-substantiated theory involves the intrapancreatic breakdown of excessive triglycerides by lipase and release of noxious free fatty acids.372 Although hypertriglyceridemia (typically >500  mg/dL373) is asso­ ciated with recurrent acute pancreatitis, the relationship between hypertryglyceridemia, or other hyperlipidemias, and chronic pancreatitis remains controversial. Conditions to consider include familial lipoprotein lipase deficiency374,375 and apolipoprotein C-II deficiency,376,377 both of which cause chronic hypertriglyceridemia and bouts of pancreatitis that segregate with the disease gene. Chronic pancreatitis was seen in an extended Dutch kindred of patients with genetically deficient lipoprotein lipase catalytic activity.378 Chronic pancreatitis was not recognized in the kindred with lipoprotein lipase deficiency reported by Wilson and coworkers.374 Cox and associates376 reported a kindred with apolipoprotein C-II deficiency with recurrent pancreatitis and chronic pancreatitis, although “chronic pancreatitis” was not defined (i.e., one of five pancreatitis patients had “malabsorption syndrome” and diabetes). One of three patients with apolipoprotein C-II deficiency syndrome reported by Beil and colleagues had pancreatic calcifications.377 DiMagno and colleagues367 noted that 5 out of 462 patients evaluated for chronic pancreatitis had preexisting hyperlipidemia, but hyperlipidemia was not listed as an etiology of chronic pancreatitis. Clinical series373 and reviews366,379 of this topic generally recognize only acute pancreatitis with hypertriglyceridema or do not discuss this issue,369 whereas others note that familial chylomicronemia syndromes lead to severe pancreatic insufficiency.375 Taken together, it appears that in the most severe, prolonged, and poorly controlled cases (e.g., genetic lipoprotein lipase deficiencies) who suffer with recurrent attacks of acute pancreatitis, chronic pancreatitis can develop. However, this appears to be a rare occurrence.

ACKNOWLEDGMENTS

The authors would like to acknowledge the contribution of Markus Lerch for contributing to the section on JohansonBlizzard syndrome, and Julia Greer for critical review and editing of this chapter.

KEY REFERENCES

Aoun E, Chang CC, Greer JB, et al. Pathways to injury in chronic pancreatitis: Decoding the role of the high-risk SPINK1 N34S haplotype using meta-analysis. PLoS ONE 2008; 3:e2003. (Ref 86.)

Chapter 57  Hereditary, Familial, and Genetic Disorders of the Pancreas and Pancreatic Disorders in Childhood Desmond CP, Wilson J, Bailey M, et al. The benign course of liver disease in adults with cystic fibrosis and the effect of ursodeoxycholic acid. Liver Int 2007; 27:1402-8. (Ref 228.) Durno C, Corey M, Zielenski J, et al. Genotype and phenotype correlations in patients with cystic fibrosis and pancreatitis. Gastroenterology 2002; 123:1857-64. (Ref 56.) Forsmark CE, Baillie J. AGA Institute technical review on acute pancreatitis. Gastroenterology 2007; 132:2022-44. (Ref 96.) Howes N, Lerch MM, Greenhalf W, et al. Clinical and genetic characteristics of hereditary pancreatitis in Europe. Clin Gastroenterol Hepatol 2004; 2:252-61. (Ref 11.) Kandula L, Khan S, Whitcomb DC, et al. Acute pancreatitis in association with Campylobacter jejuni-associated diarrhea in a 15-year-old with CFTR mutations: Is there a link? J Pancreas 2006; 7:482-5. (Ref 68.) Lowe ME, Rothbaum RJ. Cystic fibrosis. In: Johnson L, editor. Encyclopedia of Gastroenterology. Philadelphia: Elsevier; 2004. pp 529-41. (Ref 167.) Martin SP, Ulrich II CD. Pancreatic cancer surveillance in a high-risk cohort: Is it worth the cost? Med Clin North Am 2000; 84:739-47. (Ref 269.) Omary MB, Lugea A, Lowe AW, et al. The pancreatic stellate cell: A star on the rise in pancreatic diseases. J Clin Invest 2007; 117:50-9. (Ref 14.) Rowntree RK, Harris A. The phenotypic consequences of CFTR mutations. Ann Hum Genet 2003; 67:471-85. (Ref 50.)

Stallings VA, Stark LJ, Robinson KA, et al. Evidence-based practice recommendations for nutrition-related management of children and adults with cystic fibrosis and pancreatic insufficiency: Results of a systematic review. J Am Diet Assoc 2008; 108:832-9. (Ref 242.) Sutton R, Criddle D, Raraty MG, et al. Signal transduction, calcium and acute pancreatitis. Pancreatology. 2003; 3:497-505. (Ref 24.) Whitcomb DC. Inflammation and cancer V. Chronic pancreatitis and pancreatic cancer. Am J Physiol Gastrointest Liver Physiol 2004; 287:G315-19. (Ref 204.) Whitcomb DC. Mechanisms of disease: Advances in understanding the mechanisms leading to chronic pancreatitis. Nat Clin Pract Gastroenterol Hepatol 2004; 1:46-52. (Ref 7.) Whitcomb DC, Barmada MM. A systems biology approach to genetic studies of pancreatitis and other complex diseases. Cell Mol Life Sci 2007; 64:1763-77. (Ref 6.) Whitcomb DC, Lowe ME. Human pancreatic digestive enzymes. Dig Dis Sci 2007; 52:1-17. (Ref 17.) Wilschanski M, Durie PR. Patterns of GI disease in adulthood associated with mutations in the CFTR gene. Gut. 2007; 56:1153-63. (Ref 60.) Full references for this chapter can be found on www.expertconsult.com.

957

CHAPTE R

58 Acute Pancreatitis Scott Tenner and William M. Steinberg

CHAPTER OUTLINE Incidence and Burden of Disease  959 Definitions  959 Natural History  961 Pathogenesis  961 Pathophysiology  962 Predisposing Conditions  963 Obstruction  963 Alcohol, Other Toxins, and Drugs  964 Metabolic Disorders  965 Infections  966 Vascular Disease  966 Trauma  966 Post-ERCP  967 Postoperative  968 Hereditary and Genetic  968 Controversial Causes  968 Miscellaneous  969 Clinical Features  969 History  969 Physical Examination  969 Differential Diagnosis  970 Laboratory Diagnosis  970 Pancreatic Enzymes  970 Other Blood and Urine Tests  971 Standard Blood Tests  971 Diagnostic Imaging  971 Abdominal Plain Film  971 Chest Radiography  972 Abdominal Ultrasonography  972 Endoscopic Ultrasonography  972 Computed Tomography  972 Magnetic Resonance Imaging  972

INCIDENCE and BURDEN OF DISEASE The incidence of acute pancreatitis in England, Denmark, and the United States varies from 4.8 to 38 per 100,000 patients.1-3 However, estimates of incidence are inaccurate because the diagnosis of mild disease may be missed, and death may occur before diagnosis in 10% of patients with severe disease.4 Diseases of the pancreas (acute and chronic pancreatitis) accounted for 327,000 inpatient hospital stays, 78,000 outpatient hospital visits, 195,000 emergency department visits, and 531,000 physician office visits in 1998.5 The cost of pancreatic diseases (direct and indirect costs) was estimated to be 2.5 billion dollars in the year 2000.5 In the same year, there were 2834 deaths in the United States from acute pancreatitis, making it the 14th most common cause of deaths due to gastrointestinal (GI) diseases.6 Acute pancreatitis ranks as the second most common inpatient GI diag-

Distinguishing Alcoholic from Gallstone Pancreatitis  973 Predictors of Severity  973 Scoring Systems  974 Organ Failure  974 Peritoneal Lavage  975 Laboratory Markers  975 Computed Tomography  975 Chest Radiography  976 Treatment  976 General Considerations  976 Fluid Resuscitation  977 Respiratory Care  977 Cardiovascular Care  977 Metabolic Complications  977 Antibiotics  977 Endoscopic  978 Nutritional  978 Surgical Therapy  979 Other Therapeutic Agents of Possible or Questionable Efficacy  980 Local Complications  980 Pseudocyst  980 Necrotizing Pancreatitis and Abscess  981 Gastrointestinal Bleeding  982 Splenic Complications  982 Bowel Compression or Fistulization  982 Systemic Complications  982 Organ Dysfunction  982 Metabolic Disturbances  982 Coagulation Disorders  982 Fat Necrosis  982 Miscellaneous  982

nosis in the United States after cholelithiasis and acute cholecystitis and ahead of acute appendicitis.6 The incidence of acute pancreatitis appears to be increasing.7-9 As the population is becoming increasingly overweight, the incidence of gallstones, the most common cause of acute pancreatitis is rising. Unfortunately, during this same period, the overall mortality rate from acute pancreatitis has declined only gradually to approximately 5% to 10%.10-12

DEFINITIONS Many pancreatologists use the 1992 Atlanta Symposium definition of acute pancreatitis,13 which is an acute inflammatory process of the pancreas with variable involvement of other regional tissues or remote organ systems. Acute pancreatitis is best defined clinically by a patient presenting

959

960

Section VII  Pancreas Table 58-1  Atlanta Criteria for Severe Acute Pancreatitis13 Organ Failure a.  Shock: systolic blood pressure <90 mm Hg b.  Pulmonary insufficiency: Pao2 ≤60 mm Hg c.  Renal failure: serum creatinine >2 mg/dL d.  Gastrointestinal bleeding: >500 mL/24 hr Local Complications a.  Necrosis b.  Abscess c.  Pseudocyst Unfavorable Early Prognostic Signs a.  Ranson’s signs (see Table 58-2) b.  APACHE-II points

G P

Table 58-2  Ranson’s Prognostic Criteria14,215 NON-GALLSTONE PANCREATITIS (1974) At Admission Age >55 yr White blood cells >16,000/mm3 Blood glucose >200 mg/dL Serum lactate dehydrogenase >350 IU/L Serum aspartate aminotransferase >250 IU/L During Initial 48 hr Hematocrit decrease of >10 % Blood urea nitrogen increase of >5 mg/dL Serum calcium <8 mg/dL Arterial po2 <60 mm Hg Serum base deficit >4 mEq/L Fluid sequestration >6 L

GALLSTONE PANCREATITIS (1982) Age >70 yr >18,000/mm3 >220 mg/dL >400 IU/L >250 IU/L >10% >2 mg/dL

Figure 58-1.  Acute interstitial pancreatitis. Contrast-enhanced computed tomography shows diffuse swelling of the pancreas (P) with peripancreatic inflammatory changes (arrows). The pancreas was well perfused without evidence of necrosis. G, gallbladder.

<8 mg/dL NA >5 mEq/L >4 L

NA, not applicable.

with two of the following criteria12: (1) symptoms, such as epigastric pain, consistent with the disease; (2) a serum amylase or lipase greater than three times the upper limit of normal; or (3) radiologic imaging consistent with the diagnosis, usually using computed tomography (CT) or magnetic resonance imaging (MRI). Pancreatitis is classified as acute unless there are CT, MRI, or endoscopic retrograde cholangiopancreatography (ERCP) findings of chronic pancreatitis. Then pancreatitis is classified as chronic pancreatitis, and any episode of acute pancreatitis is considered an exacerbation of inflammation superimposed on chronic pancreatitis (see Chapter 59). Once the diagnosis is established, patients are classified as having mild or severe pancreatitis. Mild acute pancreatitis consists of interstitial (edematous) pancreatitis on imaging, minimal or no extrapancreatic organ dysfunction, and typically an uneventful recovery. Severe pancreatitis manifests as organ failure or local complications such as necrosis, abscess, or pseudocyst. The Atlanta criteria13 (Table 58-1) defines severity by the presence of organ failure or pancreatic necrosis on dynamic contrast–enhanced CT scan (Figs. 58-1 and 58-2). Other acceptable markers of severe pancreatitis include three or more of Ranson’s 11 criteria for non-gallstone pancreatitis (Table 58-2),14 and an Acute Physiology and Chronic Health Evaluation (APACHEII) score of greater than eight.15 It is important to use precise terms in describing the anatomic complications of acute pancreatitis. The ability to

G

Figure 58-2.  Acute pancreatic necrosis. Contrast-enhanced computed tomography demonstrates focal areas of decreased perfusion in the pancreatic parenchyma (arrows) with surrounding peripancreatic inflammation. The necrosis was estimated to involve less than 30% of the pancreas. G, gallbladder.

apply appropriate therapy depends on a clear understanding of these terms. An old term that should be used only sparingly is phlegmon. Although this term is often used by radiologists to describe an inflammatory mass, this term has carried different meaning to gastroenterologists, internists, and surgeons. Whereas patients with interstitial pancreatitis have a normally perfused gland, manifesting contrastenhanced CT as a normal bright appearance indicating flow

Chapter 58  Acute Pancreatitis throughout the gland, patients with necrotizing pancreatitis have greater than 30% of the gland not perfused, with low attenuation. Pancreatic necrosis consists of focal or diffuse nonviable pancreatic parenchyma and usually peripancreatic fat necrosis. Pancreatic necrosis can be sterile or infected. Peripancreatic necrosis describes necrotic fatty and stromal tissue around the pancreas. It is more important to surgeons because this is typically not appreciated on imaging. However, the presence of peripancreatic necrosis may delineate a more complicated course for patients with acute pancreatitis. An acute fluid collection is fluid located in or near the pancreas that lacks a definite wall and typically occurs early in the course of acute pancreatitis. On CT scan these collections appear as a low attenuation mass with poor margins and no capsule. It is very difficult to distinguish acute fluid collections in the pancreatic parenchyma from pancreatic necrosis. An acute fluid collection occurs in 30% to 50% of cases of acute pancreatitis and most resolve spontaneously.16 A pseudocyst is a fluid collection that persists for 4 to 6 weeks and becomes encapsulated by a wall of fibrous or granulation tissue. Pseudocysts are located adjacent to or off the body of the pancreas. At times these enzyme-rich fluid-filled sacks can be found distantly in the pelvis and chest. When a pseudocyst is located within the body of the pancreas, the cyst may contain necrotic pancreatic debris even when the pseudocyst is fluidappearing with low attenuation on CT. The term for a walled-off fluid-appearing pseudocyst-like structure involving the pancreas is walled off pancreatic necrosis (WOPN). A pancreatic abscess is a circumscribed intra-abdominal collection of pus occurring after an episode of acute pancreatitis or pancreatic trauma. It usually develops close to the pancreas and contains little pancreatic necrosis. Due to confusion of whether an abscess represents an infected pseudocyst or infected pancreatic necrosis, the term abscess should be used sparingly. Because of important differences in management, it is best to use the terms infected pseudocyst and infected necrosis. The term hemorrhagic pancreatitis should also be used with caution, and this term is not a synonym for necrotizing pancreatitis. Hemorrhage is more commonly associated with pseudoaneurysm, an erosion of peripancreatic blood vessels with hemoperitoneum. Unfortunately, hemorrhagic pancreatitis has more commonly been used to inappropriately describe necrotizing pancreatitis. Of all these terms, the most important distinction is that between pancreatic necrosis and pseudocyst. WOPN is pancreatic necrosis that has liquefied after five to six weeks.17 Similar to a pseudocyst, a wall develops. However, whereas a pseudocyst always contains fluid, pancreatic necrosis, even if walled off early, contains a significant amount of debris that only becomes liquefied after five to six weeks. No attempt should be made to drain WOPN early (less than four weeks) because the debris is typically thick, often with the consistency of rubber early in the course of the disease. After five to six weeks, WOPN can be treated similar to the fluid-filled pseudocyst and drained surgically, endoscopically, or percutaneously.

NATURAL HISTORY Acute pancreatitis appears to have two distinct stages. The first stage is related to the pathophysiology of the inflammatory cascade. This first phase usually lasts a week. During this phase, the severity of acute pancreatitis is related to extrapancreatic organ failure secondary to the patient’s sys-

temic inflammatory response elicited by acinar cell injury. Infectious complications are uncommon at this time. Fever, tachycardia, hypotension, respiratory distress, and leukocytosis are typically related to the systemic inflammatory response syndrome (SIRS). Multiple cytokines are involved, including platelet activating factor, tumor necrosis factor-α (TNF-α) and various interleukins (ILs) (see Chapter 2). During the first week the initial state of inflammation evolves dynamically with variable degrees of pancreatic and peripancreatic ischemia or edema to either resolution or to irreversible necrosis and liquefaction, or the development of fluid collections in and around the pancreas. The extent of the pancreatic and peripancreatic changes is usually proportional to the severity of extrapancreatic organ failure. However, organ failure may develop independent of pancreatic necrosis.17 Approximately 75% to 80%, of patients with acute pancreatitis have a resolution of the disease process (interstitial pancreatitis) and do not enter the second phase. However, in 25% of patients, a more protracted course develops, often related to the necrotizing process (necrotizing pancreatitis) lasting weeks to months. The mortality peak in the second phase is related to a combination of factors, including organ failure secondary to sterile necrosis, infected necrosis, or complications from surgical intervention.11,12,18-20 There are two peaks for mortality. Most studies in the United States and Europe reveal that about half the deaths occur within the first week or two, usually of multiorgan failure.19-21 Death can be very rapid. About one quarter of all deaths in Scotland occurred within 24 hours of admission and one third within 48 hours.21 After the second week of illness, patients succumb to pancreatic infection associated with multiorgan failure. Some studies in Europe report a very high late mortality rate from infection.22 Patients who are older and have comorbid illnesses have a substantially higher mortality rate than younger healthier patients. In those who survive their illness, severe pancreatic necrosis can scar the pancreas, resulting in a stricture of the main pancreatic duct with subsequent obstructive chronic pancreatitis and permanent diabetes and malabsorption.23

PATHOGENESIS The initial step in the pathogenesis of acute pancreatitis is conversion of trypsinogen to trypsin within acinar cells in sufficient quantities to overwhelm normal mechanisms to remove active trypsin (see Fig. 57-3). Trypsin, in turn, catalyzes conversion of proenzymes, including trypsinogen and inactive precursors of elastase, phospholipase A2 (PLA2), and carboxypeptidase, to active enzymes. Trypsin also may activate the complement and kinin systems. Active enzymes autodigest the pancreas and initiate a cycle of releasing more active enzymes. Normally small amounts of trypsinogen are spontaneously activated within the pancreas, but intrapancreatic mechanisms quickly remove activated trypsin. Pancreatic secretory trypsin inhibitor (PSTI, now called SPINK1) binds and inactivates about 20% of the trypsin activity. Other mechanisms for removing trypsin involve mesotrypsin, enzyme Y, and trypsin itself, which splits and inactivates trypsin. The pancreas also contains nonspecific antiproteases such as α1-antitrypsin and α2macroglobulin. Additional protective mechanisms are the sequestration of pancreatic enzymes within intracellular compartments of the acinar cell during synthesis and transport and the separation of digestive enzymes from lysosomal hydrolases as they pass through the Golgi apparatus,

961

962

Section VII  Pancreas which is important because cathepsin B activates trypsin from trypsinogen. Low intra-acinar calcium concentrations also prevent further autoactivation of trypsin. In experimental pancreatitis, activation of trypsin occurs within 10 minutes, and large amounts of trypsin24 and increased concentrations of trypsinogen activation peptide (TAP) accumulate within the pancreas.25,26 TAP is cleaved when trypsinogen is activated to trypsin, and concentrations of TAP in plasma, urine, and ascites correlate with the severity of the pancreatic inflammatory response, with the highest levels associated with acinar necrosis and intrapancreatic hemorrhage.27,28 Co-localization of pancreatic enzymes in lysosomes, followed by acinar cell injury, is an attractive hypothesis for the pathogenesis of acute pancreatitis, but the relevance of co-localization to the pathogenesis of acute pancreatitis is unclear. Activation of trypsinogen occurs before biochemical or morphologic injury to acinar cells, in association with co-localization of lysosomal enzymes, such as cathepsin B, and digestive enzymes, including trypsinogen within unstable vacuoles.27,28 Complete inhibition of pancreatic cathepsin B activity in vitro prevents trypsinogen activation induced by the cholecystokinin (CCK) analog cerulein,29 supporting the co-localization hypothesis. Thus, complete inhibition of cathepsin B may prevent or be a treatment for acute pancreatitis. However, enzyme co-localization may occur without inducing significant acinar cell injury.30 Two other features of experimental acute pancreatitis are early blockade of the secretion of pancreatic enzymes while enzyme synthesis continues and disruption of the paracellular barrier of acinar cells and intralobular pancreatic duct cells. The disruption facilitates the extravasation of pancreatic enzymes from acinar cells and from the duct lumen into interstitial spaces. This phenomenon may explain the rapid development of interstitial edema and the increase of pancreatic enzymes in the serum.31 As discussed in Chapter 57, the discovery of genetic mutations associated with hereditary pancreatitis also lends support to the hypothesis that intrapancreatic activation of pancreatic zymogens is central to the pathogenesis of acute pancreatitis.32-35 The mutant trypsin in hereditary pancreatitis (usually R122H or N29I mutation) causes trypsin to be resistant to lysis or causes premature trypsinogen activation (gain of function mutation) leading to autodigestion of the pancreas and episodes of acute pancreatitis.36,37 Mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene have also been implicated in pancreatitis (see Chapter 57). CFTR anion channel allows for chloride and bicarbonate secretion into the pancreatic ducts and thus allows flushing of the liberated enzymes and proenzymes into the duodenum. There are more than 1200 mutations that have been described for the CFTR gene. Some of these are considered severe and some mild. Homozygote severe mutations produce a viscid, concentrated, acidic pancreatic juice leading to ductal obstruction and pancreatic insufficiency in infanthood. Heterozygotes of minor or major mutations may lead to acute recurrent or chronic pancreatitis by altering acinar or ductal cell function (e.g., alteration of bicarbonate conductance). A third genetic abnormality associated with pancreatitis is mutation of the SPINK1 gene.38 As noted, SPINK1 protects the pancreatic acinar cell by inhibiting prematurely activated trypsin. Mutations of this gene presumably limit the activity of this protein, but the exact mechanism is unclear. The pathogenesis of gallstone-related pancreatitis is unknown (see Chapter 65). Factors that may initiate gallstone pancreatitis include reflux of bile into the pancreatic duct39,40 or obstruction of the pancreatic duct at the ampulla

from stone(s) or edema resulting from the passage of a stone.41 Reflux of bile into the pancreatic duct could occur when the distal bile and pancreatic ducts form a common channel and a gallstone becomes impacted in the duodenal papilla. Alternatively, bile could reflux into the pancreatic duct from the duodenum through an incompetent sphincter of Oddi injured by recent passage of a gallstone. Experimentally, reflux of bile into the pancreatic duct, particularly if infected or mixed with pancreatic enzymes, causes pancreatic injury. Mixtures of bile and pancreatic enzymes increase the permeability of the main pancreatic duct, which is associated with local parenchymal inflammation.42 The common channel theory is somewhat problematic because pancreatic duct pressure is invariably higher than bile duct pressure, making bile reflux into the pancreatic duct unlikely. Reflux of bile from the duodenum also is unlikely because pancreatitis does not occur in conditions with easily demonstrable reflux, such as after surgical sphincteroplasty or endoscopic sphincterotomy. A popular opinion for the mechanism of gallstone pancreatitis is that an impacted gallstone in the distal bile duct obstructs the pancreatic duct, which increases pancreatic pressure, thereby damaging ductal and acinar cells. Experiments in the opossum that support this theory are the observations that ligation of the pancreatic duct causes severe necrotizing pancreatitis41 and that decompression of the ductal system within three days prevents progression to acinar cell necrosis and severe inflammation.43

PATHOPHYSIOLOGY The pathophysiology of acute pancreatitis starts with local acinar injury that, if unchecked, leads to local inflammatory complications, a systemic response and sepsis. Pathophysiologic mechanisms include microcirculatory injury, leukocyte chemoattraction, release of pro- and anti-inflammatory cytokines, oxidative stress, leakage of pancreatic fluid into the region of the pancreas, and bacterial translocation to the pancreas and systemic circulation. The release of pancreatic enzymes damages the vascular endothelium, the interstitium, and acinar cells.43-45 Acinar injury leads to expression of endothelial adhesion molecules (e.g., VCAM-1), which further propagates the inflammatory response.46 Microcirculatory changes, including vasoconstriction, capillary stasis, decreased oxygen saturation, and progressive ischemia, occur early in experimental acute pancreatitis. These abnormalities increase vascular permeability and lead to edema of the gland (edematous or interstitial pancreatitis). Vascular injury could lead to local microcirculatory failure and amplification of the pancreatic injury. It is uncertain whether ischemia-reperfusion injury occurs in the pancreas.40 Reperfusion of damaged pancreatic tissue could lead to the release of free radicals and inflammatory cytokines into the circulation, which could cause further injury. In early stages of animal and human pancreatitis, activation of complement and the subsequent release of C5a play significant roles in the recruitment of macrophages and polymorphonuclear leukocytes.47-49 Active granulocytes and macrophages release proinflammatory cytokines in response to transcription factors such as nuclear factor κB (NF-κB). Proinflammatory cytokines include TNF, IL-1, IL-6, and IL-8, and platelet-activating factor (PAF). Proinflammatory cytokines frequently are followed by anti-inflammatory cytokines (IL-2, IL-10, IL-11) that attempt to down-regulate inflammation.48 Other mediators of inflammation include arachidonic acid metabolites

Chapter 58  Acute Pancreatitis (prostaglandins, PAF, and leukotrienes), nitric oxide, proteolytic and lipolytic enzymes, and reactive oxygen metabolites that overwhelm scavenging by endogenous antioxidant systems. These substances also interact with the pancreatic microcirculation to increase vascular permeability, which induces thrombosis and hemorrhage and leads to pancreatic necrosis. A recent study suggests that gene polymorphisms that affect acinar cell glutathione concentrations may lead to increased oxidant stress and more severe pancreatitis.50 Meanwhile, ischemia and severe inflammation of the gland can lead to disruption of the main and secondary pancreatic ducts, leading to local fluid accumulations within and surrounding the pancreas that can eventuate into pseudocysts.51,52 Some patients with severe pancreatic damage develop systemic complications, including fever, acute respiratory distress syndrome (ARDS), pleural effusions, renal failure, shock, myocardial depression, and metabolic complications. SIRS is common in patients with acute pancreatitis and is probably mediated by activated pancreatic enzymes (phospholipase, elastase, trypsin) and cytokines (TNF, PAF) released into the portal circulation from the inflamed pancreas.53 Cytokines reaching the liver activate hepatic Kupffer cells, which, in turn, induces hepatic expression and secretion of cytokines into the systemic circulation. These cause acute phase protein synthesis (C-reactive protein [CRP], IL-6) and may cause SIRS and damage to the kidneys, lungs, and other organs leading to multiorgan dysfunction and failure.54 ARDS may be induced by active phospholipase A (lecithinase), which digests lecithin, a major component of lung surfactant. Acute renal failure has been explained on the basis of hypovolemia and hypotension. Myocardial depression and shock are likely secondary to vasoactive peptides and a myocardial depressant factor. Metabolic complications include hypocalcemia, hyperlipidemia, hyperglycemia with or without ketoacidosis, and hypoglycemia. The pathogenesis of hypocalcemia is multifactorial and includes hypoalbuminemia (the most important cause), hypomagnesemia, calcium-soap formation, hormonal imbalances (e.g., involving parathyroid hormone, calcitonin, and glucagon), binding of calcium by free fatty acid–albumin complexes, intracellular translocation of calcium, and systemic exposure to endotoxin.55 Pancreatic infection (infected necrosis and infected pseudocyst) can occur from the hematogenous route or from translocation of bacteria from the colon into the lymphatics. Under normal circumstances bacterial translocation does not occur because there are complex immunologic and morphologic barriers. However, during acute pancreatitis, these barriers break down, which can result in local and systemic infection.56 Penetration of the gut barrier by enteric bacteria is likely due to gut ischemia secondary to hypovolemia and pancreatitis-induced arteriovenous shunting in the gut.57 Indeed, in canine experimental pancreatitis, luminal Escherichia coli translocate to mesenteric lymph nodes and distant sites.58 In feline experimental pancreatitis, enclosing the colon in impermeable bags prevents translocation of bacteria from the colon to the pancreas.59

PREDISPOSING CONDITIONS Many conditions predispose to acute pancreatitis to varying degrees (Table 58-3). This list will undoubtedly continue to grow, and the number of cases diagnosed as “idiopathic” will decrease as our understanding of the disease improves.

Table 58-3  Conditions That Predispose to Acute Pancreatitis Obstructive Gallstones Tumors Parasites Duodenal diverticula Annular pancreas Choledochocele Other Alcohol/other toxins/drugs Ethyl alcohol Scorpion venom Methyl alcohol Organophosphorous insecticides Drugs (see Table 58-4) Metabolic Hypertriglyceridemia Hypercalcemia Infectious Vascular Vasculitis Emboli to pancreatic blood vessels Hypotension Trauma Post-endoscopic retrograde cholangiopancreatography (ERCP) Postoperative Hereditary/familial/genetic Controversial Pancreas divisum Sphincter of Oddi dysfunction Miscellaneous Idiopathic

Gallstones and chronic alcohol abuse account for 70% of acute pancreatitis in the United States.

OBSTRUCTION Gallstones

The most common obstructive process leading to pancreatitis is gallstones, which cause approximately 40% of cases acute pancreatitis,60 although only 3% to 7% of patients with gallstones develop pancreatitis. Gallstone pancreatitis is more common in women than men because gallstones are more frequent in women.61 Acute pancreatitis occurs more frequently when stones are less than 5  mm in diameter (odds ratio, 4 to 5),62 because small stones are more likely than large stones to pass through the cystic duct and cause ampullary obstruction. Cholecystectomy and clearing the bile duct of stones prevents recurrence, confirming the cause-and-effect relationship.61

Biliary Sludge and Microlithiasis

Biliary sludge is a viscous suspension in gallbladder bile that may contain small (<3  mm) stones (i.e., microlithiasis).63 Because small stones can hide in biliary sludge, the two are commonly referred together as biliary sludge and microlithiasis. Biliary sludge is asymptomatic in most patients. It is usually composed of cholesterol monohydrate crystals or calcium bilirubinate granules.64 On ultrasonography, sludge produces a mobile, low-amplitude echo that does not produce an acoustic shadow and that layers in the most dependent part of the gallbladder. Sludge may result from functional bile stasis, such as that associated with prolonged fasting or total parenteral nutrition, or from mechanical stasis such as occurs in distal bile duct obstruction. In addition, the cephalosporin antibiotic ceftriaxone can complex with bile to form a sludge within

963

964

Section VII  Pancreas the biliary system when its solubility in bile is exceeded; this rarely causes stones65 and the sludge disappears after stopping the drug. Commonly, biliary sludge is associated with idiopathic acute pancreatitis. However, the association between biliary sludge and acute pancreatitis is unproved. There is no prospective, randomized study documenting that removing sludge or microcrystals by cholecystectomy prevents further attacks of pancreatitis. Nevertheless, results of two uncontrolled studies suggest that biliary sludge can lead to pancreatitis, and that cholecystectomy, papillotomy, or ursodeoxycholic acid therapy reduces recurrent attacks of acute pancreatitis.64,66 In these two studies, the incidence of biliary sludge in presumed idiopathic pancreatitis was 67% and 74%, respectively. However, other investigators have detected microlithiasis or sludge in less than 10% of patients with recurrent acute pancreatitis.67,68 Until prospective controlled studies clarify the proper treatment of sludge and microlithiasis, firm recommendations about therapy cannot be made. Choices include cholecystectomy, urso­ deoxycholic acid therapy, endoscopic sphincterotomy, or watchful waiting.

Tumors

Tumors, presumably by obstructing the pancreatic duct, can cause recurrent acute pancreatitis especially in individuals older than age 40 (see Chapter 60). The most common tumor that presents in this manner is intraductal papillary mucinous neoplasm (IPMN).69 Pancreatic adenocarcinoma can also present as acute pancreatitis in a small percentage of patients.70 Metastases from other primary tumors (lung, breast) to the pancreas have also caused pancreatitis.71 Large adenomas of the major papilla can likewise occasionally be the cause of obstructive pancreatitis.

Other Obstructive Causes

Other obstructive conditions that are rarely associated with acute pancreatitis are conditions discussed elsewhere in this text and include choledochoceles,72 duodenal diver­ ticula,73 annular pancreas,74 and parasites that obstruct the pancreatico-biliary system such as ascaris75 or clonorchis.76 Ascariasis obstructing the pancreatic duct represents the second most common cause of acute pancreatitis in Kashmir.60

ALCOHOL, OTHER TOXINS, AND DRUGS Ethyl Alcohol

Alcohol causes at least 30% of cases of acute pancreatitis,77 and alcohol is the most common etiology of chronic pancreatitis in developed countries. Interestingly, only 10% of chronic alcoholic patients develop chronic pancreatitis. The classic teaching is that alcohol causes chronic pancreatitis, and that alcoholic patients who present with clinically acute pancreatitis have underlying chronic disease.16 However, a few patients with alcohol-induced acute pancreatitis by clinical criteria do not have or progress to chronic pancreatitis, even with continued alcohol abuse.77,78 By contrast, a small percentage of chronic alcoholic patients develop attacks of acute pancreatitis that are indistinguishable from other forms of acute pancreatitis, but eventually develop chronic pancreatitis after 10 to 20 years of alcohol abuse. Early in the course of the disease, when attacks occur, the diagnosis of underlying chronic pancreatitis is difficult without tissue specimens because the diagnosis of chronic pancreatitis is usually made after definite signs of chronic pancreatitis appear (e.g., pancreatic calcification, exocrine and endocrine insufficiency, or typical duct changes by CT or ERCP). Most of the models described suggest possible mechanisms of alcohol-related injury,

including perturbations in exocrine function, changes in cellular lipid metabolism, induction of oxidative stress, and activation of stellate cells. However, the exact mechanism remains unclear and may be related to other factors. Bordalo and colleagues79 first proposed that alcohol was directly toxic to the acinar cell through a change in cellular metabolism. Alcohol produces cytoplasmic lipid accumulation within the acinar cells, leading to fatty degeneration, cellular necrosis, and eventual widespread fibrosis. Fatty acid ethyl esters, by-products of pancreatic ethanol metabolism, may be the key factor in this “toxic metabolic” change. Bordalo and colleagues suggested that alcohol produces a stepwise progression from fatty accumulation to fibrosis (direct toxic effects on cellular metabolism). The main limitation to this toxic-metabolic theory of alcohol toxicity is the lack of proof of the steatopancreatitis precursor to fibrosis seen in liver disease.80 Sarles80 emphasized the duality of acute and chronic pancreatitis; they were separate diseases with distinct pathogenesis. Whereas acute pancreatitis can be precipitated in patients with gallstones immediately, alcoholism requires years of toxin exposure. Alcohol modulates exocrine function to increase the lithogenicity of pancreatic fluid, leading to the formation of protein plugs and stones. Chronic contact of the stones with the ductal cells produces ulceration and scarring, resulting in obstruction, stasis, and further stone formation. Eventually, atrophy and fibrosis develop as a result of this obstructive process. Several studies have provided mechanisms in which alcohol promote stone formation, including the known precipitation of GP-2 (a Tamm-Horsfall–like protein),81 increased secretion and viscosity of pancreatic juice, and hypersecretion of enzymes and lactoferrin. In addition to these ethanol-mediated perturbations in pancreatic exocrine function, specific proteins have been implicated in stone formation. Pancreatic stones consist of a calcium carbonate crystalline lattice interspersed within a gel-like matrix formed of multiple fibrillar proteins and polysaccharides.82 In contrast to the stone theory, which is based on the de novo development of fibrosis without acute pancreatitis, the necrosis-fibrosis hypothesis envisions the development of fibrosis from recurrent, perhaps subclinical, acute pancreatitis. Inflammation and necrosis from the initial episodes of acute pancreatitis produce scarring in the periductular areas and scarring leads to obstruction of the ductules leading to stasis within the duct and subsequent stone formation. Support for this theory comes from histopathologic studies that revealed mild perilobular fibrosis in resolving acute pancreatitis, with marked fibrosis with ductal distortion occurring later. It is thought that a stepwise progression occurs to fibrosis from recurrent episodes of acute pancreatitis. Support for this theory is seen in a clinical study by Ammann and colleagues.83 In this study, 254 patients were prospectively followed after the first episode of alcoholic pancreatitis. There was a direct correlation between the frequency and severity of attacks to the rate of progression to chronic pancreatitis. Whitcomb and Schneider have proposed an interesting hypothesis for chronic pancreatitis which unifies and incorporates recent knowledge in an attempt to reconcile prior theories.84 In at-risk individuals, the pancreatic acinar cells are stimulated by alcohol. Fibrosis does not occur because a profibrotic cellular infiltrate is not yet present. A sentinel event occurs as trypsinogen is activated. This event results in a massive inflammatory response. Cytokines are then released and work with the activation of stellate cells in the late phase. The attraction and activation of stellate cells set the stage for the development of fibrosis. If the inciting factors

Chapter 58  Acute Pancreatitis are removed, then the pancreas returns to normal. If the inciting factor, alcohol, is not removed, the acinar cells continue to secrete cytokines in response to the oxidative stress and the stellate cells continue to be activated. This model, using a sentinel event also represents a time for disease modifying therapy, when such therapy becomes available. These theories and animal models do not completely explain the development of alcoholic pancreatitis in humans. More important, the development of acute pancreatitis in an alcoholic individual without obvious structural damage suggests the possibility that a patient with idiopathic pancreatitis, found to have a history of alcohol use, may be inappropriately labeled as having alcoholic acute pancreatitis. It is important for clinicians to recognize that alcohol has been better established as a cause of chronic pancreatitis than acute pancreatitis. Some level of skepticism is likely to be helpful in the evaluation of a patient suspected as having acute alcohol-induced pancreatitis in the absence of obvious structural damage to the pancreas.

Other Toxins

Methyl alcohol,85 organophosphorous insecticides,86 and the venom of the Trinidad scorpion87 have been reported to induce pancreatitis. The mechanism of the latter two is thought to be by hyperstimulation of the pancreas. Smoking increases the risk of alcoholic and idiopathic pancreatitis, but not gallstone pancreatitis.88

Drugs

Medications are an infrequent but an important cause of acute pancreatitis.89 More than 120 drugs have been implicated, mostly from anecdotal case reports. Many case reports suffer from a combination of inadequate criteria for the diagnosis of acute pancreatitis, failure to rule out more common causes, or a lack of a rechallenge with the medication. Drug-induced pancreatitis rarely is accompanied by clinical or laboratory evidence of a drug reaction, such as rash, lymphadenopathy, or eosinophilia. Although a positive rechallenge with a drug is the best evidence available for cause and effect, it is not proof. It is clear that many patients with idiopathic pancreatitis or microlithiasis have recurrent attacks of acute pancreatitis. Therefore, stopping and restarting a drug with recurrence of pancreatitis may be a coincidence and not cause and effect. Despite the lack of a rechallenge, a drug may be strongly suspected if there is a consistent latency among the case reports between initiating the drug and the onset of acute pancreatitis. Table 58-4 shows the drugs with the greatest evidence for causing acute pancreatitis, those with rechallenges or with a relatively predictable latency.89 There are several potential pathogenetic mechanisms of drug-induced pancreatitis. The most common is a hypersensitivity reaction. This tends to occur four to eight weeks after starting the drug and is not dose related. On rechallenge with the drug, pancreatitis recurs within hours to days. Examples of drugs that operate through this mechanism are 6-mercaptopurine/azathioprine, aminosalicylates, metronidazole, and tetracycline. The second mechanism is the presumed accumulation of a toxic metabolite that may cause pancreatitis, typically after several months of use. Examples of drugs in this category are valproic acid and didanosine. Drugs that induce hypertrigylceridemia (e.g., thiazides, isotretinoin, tamoxifen) are also in this category. Finally, a few drugs may have intrinsic toxicity wherein an overdose can cause pancreatitis (erythromycin, acetaminophen). There is no documentation that drugs can cause pancreatitis after years of use. Drug-induced pancreatitis tends to be mild and self-limited.

Table 58-4  Drugs Associated with Acute Pancreatitis* Acetaminophen Alphamethyldopa 5-Aminosalicylic acid compunds Sulfasalazine Azodisalicylate Mesalamine L-Asparaginase Azathioprine Benazepril Bezafibrate Cannabis Captoril Carbimazole Cimetidine Clozapine Codeine Cytosine arabinoside Dapsone Didanosine Dexamethasone Enalapril Erythromycin Estrogen Fluvastatin Furosemide Hydrochlorothiazide

Hydrocortisone Ifosfamide Interferon-α Isoniazid Lamivudine Lisinopril Losartan Meglumine 6-Mercaptopurine Methimazole Metronidazole Nelfinavir Norethindrone/mestrol Pentamidine Pravastatin Procainamide Pyritinol Simvastatin Sulfamethazole Sulfamethoxazole Stibogluconate Sulindac Tetracycline Trimethoprim-sulfamethoxazole Valproic acid

*Class 1 and class 2 drugs. For class 1 drugs: two or more case reports published, absence of other causes of acute pancreatitis, rechallenge documented in at least one report. For class 2 drugs: four or more case reports published, absence of other causes of acute pancreatitis, consistent latency in at least 75% of cases published. From Badalov N, Baradarian R, Iswara K, et al. Drug induced acute pancreatitis: An evidence based approach. Clin Gastroenterol Hepatol 2007; 101:454-76.

METABOLIC DISORDERS Hypertriglyceridemia

Hypertriglyceridemia is perhaps the third most common identifiable cause of pancreatitis after gallstones and alcoholism. Serum triglyceride concentrations greater than 1000  mg/dL (11  mmol/L) may precipitate attacks of acute pancreatitis. Patients may have lactescent (milky) serum owing to increased concentrations of chylomicrons.90 The pathogenesis of hypertriglyceridemic pancreatitis is unclear, but the release of free fatty acids by lipase may damage pancreatic acinar cells or capillary endothelium.91 Hypertriglyceridemia may cause up to 5% of cases of acute pancreatitis. The association between hypertriglyceridemia and acute pancreatitis is best defined in children with rare inherited disorders of lipoprotein metabolism and severe hypertriglyceridemia92,93 who develop acute pancreatitis in early childhood. These children are homozygous for lipoprotein lipase deficiency or, even less commonly, apoprotein-CII (APO-CII) deficiency. Acute pancreatitis develops in 35%, 15%, and 30% to 40% of patients with type I, IIb, and V hyperlipidemia, respectively. Lowering serum triglyceride levels to less than 200 mg/dL (2.2 mmol/L) can prevent pancreatitis. Most adults with hyperchylomicronemia have a mild form of genetically inherited type I or type V hyperlipoproteinemia and an additional acquired condition known to raise serum lipids (e.g., alcohol abuse, obesity, diabetes mellitus, hypothyroidism, pregnancy, estrogen91 or tamoxifen therapy, glucocorticoid excess, nephrotic syndrome, thiazide or beta blocker therapy). Typically three types of patients develop hypertriglyceridemia-induced pancreatitis. The first is a poorly controlled diabetic patient with a history of hypertriglyceridemia. The second is an alcoholic

965

966

Section VII  Pancreas patient with hypertriglyceridemia detected on hospital admission. The third (15% to 20%) is a nondiabetic, nonalcoholic, nonobese person who has drug- or diet-induced hypertriglyceridemia. Drug-induced disease is more likely to occur if there is a background of hypertriglyceridemia prior to drug exposure. Most persons who abuse alcohol have moderate, but transient, elevations of serum triglyceride levels. This condition is likely an epiphenomenon and not the cause of their pancreatitis94 because alcohol itself not only damages the pancreas but also increases serum triglyceride concentrations in a dose-dependent manner. For example, serum triglyceride concentrations greater than 227  mg/dL (2.5  mmol/L) occurred in 10%, 14%, and 20% of people who drank three to five, six to eight, or nine or more drinks per day, respectively.95 Alcoholic patients with severe hyperlipidemia often have a coexisting primary genetic disorder of lipoprotein metabolism. The clinical manifestations of hypertriglyceridemia-­ associated pancreatitis are similar to other causes of acute pancreatitis; abdominal pain, nausea, and vomiting are the major symptoms. However, the serum amylase and/or lipase level may not be substantially elevated at presentation (see later).

Hypercalcemia

Hypercalcemia, of any cause, is rarely associated with acute pancreatitis. Proposed mechanisms include deposition of calcium in the pancreatic duct and calcium activation of trypsinogen within the pancreatic parenchyma.96 The low incidence of pancreatitis in chronic hypercalcemia suggests that factors other than the serum calcium per se (e.g., acute elevations of serum calcium) are responsible for pancreatitis. Acute calcium infusion into rats leads to conversion of trypsinogen to trypsin, hyperamylasemia, and dosedependent morphologic changes of acute pancreatitis such as edema and acinar cell necrosis. Hypercalcemia due to hyperparathyroidism has been associated with pancreatitis. However, primary hyperparathyroidism causes less than 0.5% of all cases of acute pancreatitis, and the incidence of acute pancreatitis in patients with hyperparathyroidism varies from 0.4% to 1.5%.97 Rarely, pancreatitis occurs with other causes of hypercalcemia, including metastatic bone disease, total parenteral nutrition, sarcoidosis, vitamin D toxicity, and infusions of calcium in high doses perioperatively during cardiopulmonary bypass.

INFECTIONS

Many infectious agents may cause acute pancreatitis,76 but often published reports do not meet usual standards for the diagnosis of pancreatitis or the infection. Using modern criteria for diagnosis of pancreatitis, definite pancreatitis exists if there is surgical, autopsy, or radiologic evidence; probable pancreatitis exists if there is biochemical evidence (more than three times the elevation of serum lipase or amylase) plus characteristic symptoms; and possible pancreatitis exists if there is only asymptomatic biochemical evidence. The definite criterion for an infection causing pancreatitis is finding the organism in the pancreas or pancreatic duct by stain or culture. Probable criteria for infection are culture of the organism from pancreatic juice or blood or serologic evidence combined with a characteristic clinical or epidemiologic setting. The criterion for a possible infection is culture of the organism from other body sites or serologic evidence of infection. Using these criteria, definite pancreatitis has been associated with viruses (mumps, coxsackievirus, hepatitis B, cyto-

megalovirus, varicella-zoster, herpes simplex, Epstein-Barr, hepatitis A, and hepatitis C); the vaccine that contains attenuated measles, mumps, and rubella (MMR); bacteria (Mycoplasma, Legionella, Leptospira, Salmonella, tuberculosis, and brucellosis); fungi (Aspergillus and Candida albicans); and parasites (Toxoplasma, Cryptosporidium, Ascaris, Clonorchis sinensis). C. sinensis and Ascaris cause pancreatitis by blocking the main pancreatic duct. In acquired immunodeficiency syndrome (AIDS), infectious agents causing acute pancreatitis include cytomegalovirus, Candida, Cryptococcus neoformans, Toxoplasma gondii, and possibly Mycobacterium avium complex.76 An infectious agent should be suspected of causing acute pancreatitis if the characteristic syndrome caused by the infectious agent is present because this occurs 70% of the time.76 Because an infectious agent may be found in the pancreas without pancreatitis, routine search for an infection in idiopathic pancreatitis is not recommended because false-positive test results may result. In addition, it is unknown whether treating an infectious agent reverses pancreatic pathology.

VASCULAR DISEASE

Rarely, pancreatic ischemia causes pancreatitis. In most cases it is mild, but fatal necrotizing pancreatitis may occur. Ischemia may result from vasculitis (systemic lupus erythematosus)98 and polyarteritis nodosa,99 atheromatous embolization of cholesterol plaques from the aorta to the pancreas after transabdominal angiography,100 intraoperative hypotension,101 hemorrhagic shock,102 ergotamine overdose, and transcatheter arterial embolization for hepatocellular carcinoma. Also, ischemia is one possible explanation for pancreatitis after cardiopulmonary bypass. In pigs, cardiogenic shock induced by pericardial tamponade causes vasospasm and selective pancreatic ischemia due to activation of the renin-angiotensin system.103 Acute pancreatitis has occurred in long-distance runners, which may be on an ischemic basis.104

TRAUMA

Either penetrating trauma (gunshot or stab wounds) or blunt trauma can damage the pancreas.105 In most cases there is also injury to adjacent viscera. Laparotomy is essential in all cases of penetrating trauma to assess and treat all intraabdominal injuries, including those to the pancreas. Blunt trauma results from compression of the pancreas by the spine, such as in an automobile accident. In blunt trauma it is important to determine preoperatively whether there is injury to the pancreas because depending on the severity of pancreatic injury, it will be necessary to include the pancreas in the surgical plan. Secondly, even in the absence of serious injury to adjacent organs, surgery or endoscopic therapy may be necessary to treat a pancreatic ductal injury. The diagnosis of traumatic pancreatitis is difficult and requires a high degree of suspicion. Trauma can range from a mild contusion to a severe crush injury or transection of the gland; the latter usually occurs at the point where the gland crosses over the spine. Transection injury can cause acute duct rupture and pancreatic ascites. Clinically, it is impossible to determine on the basis of the characteristics of the abdominal pain and tenderness whether the pancreas has been injured as opposed to other intra-abdominal structures. Serum amylase activity may be increased in abdominal trauma whether or not the pancreas has been injured. Diagnosis is highly dependent on CT, MRI, or magnetic resonance cholangiopancreatography (MRCP), which may

Chapter 58  Acute Pancreatitis show enlargement of a portion of the gland caused by a contusion or subcapsular hematoma, pancreatic inflammatory changes, or fluid within the anterior pararenal space if there is ductal disruption. The CT may be normal during the first two days despite significant pancreatic trauma. If there is a strong clinical suspicion of pancreatic injury or if the CT or MRCP scan shows an abnormality, ERCP is required to define whether there is pancreatic duct injury. If the pancreatic duct is intact and there are no other significant intra-abdominal injuries, surgery is not required. However, if ERCP reveals duct transection with extra­ vasation of pancreatic fluid and there are no other intraabdominal injuries, stenting of the pancreatic duct may be successful.106 Serious injuries to the pancreas can be treated with appropriate débridement. Associated injuries to the duodenum or bile duct can be treated by biliary diversion, gastrojejunostomy, and feeding jejunostomy. External pancreatic fistulas occur in approximately one third of patients after surgery for pancreatic trauma. Octreotide may be beneficial after pancreatic injury.107 The prognosis in patients with pancreatic trauma is favorable if there is no serious injury to other structures (regional blood vessels, liver, spleen, kidney, duodenum, and colon). However, duct injuries can scar and cause a stricture of the main pancreatic duct resulting in obstructive chronic pancreatitis.

POST-ERCP

Acute pancreatitis is the most common and feared complication of ERCP, associated with substantial morbidity and occasional mortality. About 500,000 ERCPs are performed annually in the United States. Asymptomatic hyperamylasemia occurs after 35% to 70% of ERCPs.108 Acute pancreatitis occurs in 5% of diagnostic ERCPs, 7% of therapeutic ERCPs, and up to 25% in those with suspected sphincter of Oddi dysfunction or in those with a history of post-ERCP pancreatitis.109 About half the cases are moderate to severe in intensity. The mechanisms that lead to post-ERCP pancreatitis are complex and not fully understood. Rather than a single pathogenesis, post-ERCP pancreatitis is believed to be multifactorial, involving a combination of chemical, hydrostatic, enzymatic, mechanical, and thermal factors. Although there is some uncertainty in predicting which patients will develop acute pancreatitis following ERCP, a number of risk factors acting independently or in concert have been proposed as predictors of post-ERCP pancreatitis (Table 58-5).110-113 Identification of these risk factors for post-ERCP pancreatitis is essential to recognize cases in which ERCP should be avoided if possible, or in which protective endoscopic or pharmacologic interventions should be considered.

Table 58-5  Factors That Increase the Risk of Post-ERCP Pancreatitis Patient Related Young age, female gender, suspected sphincter of Oddi dysfunction, recurrent pancreatitis, history of post-ERCP pancreatitis, normal serum bilirubin Procedure Related Pancreatic duct injection, difficult cannulation, pancreatic sphincterotomy, precut access, balloon dilation Operator or Technical Related Trainee (fellow) participation, nonuse of a guidewire for cannulation, nonuse of a pancreatic duct stent in high-risk procedures ERCP, endoscopic retrograde cholangiopancreatography.

In general, the more likely a patient is to have an abnormal bile duct or pancreatic duct, the less likely the patient will develop post-ERCP pancreatitis. Cheng created a 160 variable database that prospectively evaluated more than 1000 patients from 15 centers in the midwestern United States.112 Their study emphasized the role of patient factors, including age, sphincter of Oddi dysfunction (SOD), prior history of post-ERCP pancreatitis, and technical factors, including number of pancreatic duct (PD) injections, minor papilla sphincterotomy, and operator experience. The patient most at risk of developing post-ERCP pancreatitis was a woman with suspected choledocholithiasis and normal bilirubin, who underwent a sphincterotomy and no stone was found. In this patient population, more than a quarter of patients (27%) developed post-ERCP pancreatitis. MRCP and endoscopic ultrasound, which do not cause pancreatitis, can provide useful information (perhaps as accurate as ERCP) in many of these cases and are preferred modalities in the initial evaluation of such patients. Early recognition of post-ERCP pancreatitis may be possible by evaluating serum amylase or lipase after the procedure.114,115 In a study that involved 231 patients, the two-hour serum amylase and lipase were more accurate than a clinical assessment in distinguishing nonpancreatitis abdominal pain from post-ERCP acute pancreatitis. Serum values greater than 276 IU/L for amylase and greater than 1000 IU/L for lipase obtained from serum two hours after the procedure had almost a 100% positive predictive value for postERCP pancreatitis.116 More recently, Ito and colleagues found that if the serum amylase was normal at three hours, only 1% of patients had post-ERCP pancreatitis compared with 39% if the amylase was greater than five times the upper limit of normal.117 A serum amylase or lipase alone should not guide a decision of whether a patient has postERCP pancreatitis. However, it appears that the tests may assist the clinician’s assessment of a patient with post-ERCP abdominal pain. Although there has been an interest in developing medications that can prevent post-ERCP pancreatitis, studies have failed to identify a medication worthy of widespread use. In terms of attenuating the inflammatory response, the most promising results have been seen with nonsteroidal antiinflammatory drugs (NSAIDs). Two clinical trials have been published evaluating the role of diclofenac in reducing the incidence of post-ERCP pancreatitis.116,118 Both trials placed patients on 100 mg of diclofenac by rectal suppository and both showed a reduction in the incidence of acute pancrea­ titis. However, another trial failed to show any benefit to diclofenec in preventing post-ERCP pancreatitis.119 It has been suggested that relaxation of the sphincter of Oddi following ERCP will promote pancreatic drainage and prevent acute pancreatitis. Several agents have been used in the effort to relax the sphincter of Oddi with the purpose of preventing post-ERCP pancreatitis. There have been three placebo-controlled randomized studies evaluating the use of nitroglycerin during ERCP, with negative results.120-122 Furthermore, trials of oral nifedipine,123,124 sprayed lidocaine,125 and injected botulinum toxin126 failed to demonstrated any benefit in the reduction of severity or incidence of post-ERCP pancreatitis. There have been several studies evaluating the role of glucocorticoids using a variety of agents including oral prednisolone or intravenous hydrocortisone or methylprednisolone, with essentially no benefit in reduction of severity or incidence of post-ERCP pancreatitis. Gabexate is a protease inhibitor with anti-inflammatory properties. The in vivo effect of gabexate on inhibiting circulating trypsin is greater than most other protease

967

968

Section VII  Pancreas inhibitors. In 1995, Messori and colleagues127 published a meta-analysis of five trials128-132 showing a statistically significant reduction in the incidence of complications in patients receiving gabexate after the development of postERCP pancreatitis. However, the trials were small and had a limited number of patients. Several additional trials have been published with conflicting results.133-137 Although data are conflicting, it appears that infusions of the drug would likely need to be started 1 to 2 hours pre-ERCP and continued for 12 hours following ERCP to show a beneficial effect of the drug. In patients with a low risk, the costs likely outweigh any benefit. Currently, gabexate is not available in the United States. Theoretically, inhibition of exocrine pancreatic secretion could prevent post-ERCP pancreatitis by inducing “rest” to a damaged gland. Although an attractive concept, there is little scientific basis to support this approach. Somatostatin and its synthetic octapeptide analog, octreotide, are potent inhibitors of pancreatic secretion. Although several trials of somatostatin have demonstrated an efficacy in reducing the incidence of post-ERCP pancreatitis,138-143 the majority of the studies do not support the routine use of this medication.144-150 Octreotide, the analog of somatostatin,151 was only effective in reducing post-ERCP hyperamylasemia, and did not reduce the incidence of post-ERCP pancreatitis. Pancreatic stent placement clearly decreases the risk of post-ERCP pancreatitis in high-risk patients.152 Placement of pancreatic duct stents has become a standard practice for patients who are thought to be at high risk for pancreatitis after the procedure (see Table 58-5).153 Pancreatic duct stent placement is effective presumably by preventing cannulation-induced edema that can cause pancreatic duct obstruction. Pancreatic sphincter hypertension is believed to be an important causative factor in post-ERCP pancreatitis and may explain the high risk of pancreatitis in patients with SOD. There is prolonged alleviation of ductal obstruction when pancreatic stents are placed. Typically, 3- to 5-French unflanged pancreatic stents are used in the following settings: SOD, difficult cannulation, biliary orifice balloon dilation, and precut sphincterotomy. In general, pancreatic duct stents are not beneficial in patients who undergo routine biliary sphincterotomy. In all reported studies, which cumulatively include 1500 high-risk patients undergoing ERCP, only 1 patient developed severe pancreatitis after a pancreatic duct stent had been placed.154 Aside from the obvious benefits in preventing post-ERCP pancreatitis regarding morbidity and mortality, prophylactic stent placement is a cost-effective strategy for the prevention of post-ERCP pancreatitis for high-risk patients.155 Guidewire cannulation, whereby the biliary or pancreatic duct is initially cannulated by a guidewire inserted through the catheter or sphincterotome, has been shown to decrease the risk of pancreatitis (see Table 58-5).156 In a study of 400 consecutive patients who underwent ERCP by a single endoscopist, randomized to initial cannulation with contrast versus initial cannulation by a guidewire under fluoroscopic control, pancreatitis rates were profoundly different. No cases of acute pancreatitis were seen in the guidewire group compared with eight cases in the standard contrast group (P < 0.001). Cannulation success rates between the standard contrast and guidewire techniques were comparable 98.5% versus 97.5%. A later study157 confirmed a decrease in post-ERCP pancreatitis in 300 patients prospectively randomized to guidewire cannulation compared with conventional radiocontrast. However, the decrease in post-ERCP pancreatitis appears to be related to a decreased need for precut sphincterotomy in patients undergoing guidewire cannulation.

POSTOPERATIVE

Postoperative pancreatitis can occur after abdominal or thoracic surgery.158 Pancreatitis occurs after 0.4% to 7.6% of cardiopulmonary bypass operations113,159 and after 6% of liver transplantations.160 Twenty-seven percent of patients undergoing cardiac surgery develop hyperamylasemia, and 1% develop necrotizing pancreatitis.113 Significant risks for pancreatitis after cardiopulmonary bypass are preoperative renal insufficiency, postoperative hypotension, and administration of calcium chloride perioperatively. Mortality from postoperative pancreatitis is said to be higher (up to 35%) than for other forms of pancreatitis. Contributors to morbidity and mortality from postoperative pancreatitis are delay in diagnosis, hypotension, medications (e.g., azathioprine/perioperative calcium chloride administration), and infections.

HEREDITARY AND GENETIC

Hereditary pancreatitis is an autosomal dominant disorder with variable penetrance.161-169 It is discussed in Chapter 57.

CONTROVERSIAL CAUSES Pancreas Divisum

Pancreas divisum is the most common congenital malformation of the pancreas occurring in 5% to 10% of the general healthy population (see Chapter 55). Controversy continues to surround the issue as to whether pancreas divisum with otherwise normal ductular anatomy is a cause of acute recurrent pancreatitis. The presumed mechanism of action in those who develop pancreatitis is that there is relative obstruction to the flow of pancreatic juice through the minor papilla. The arguments in favor of attributing pancreatitis to pancreas divisum are the following: (1) various series from ERCP referral centers show that patients referred with recurrent acute pancreatitis have a higher frequency of pancreas divisum than would be expected from the general population170; (2) multiple observational series report that performing endoscopic sphincterotomy or placing a stent across the minor papilla reduces the rate of recurrent pancreatitis156; and (3) there is one randomized controlled study suggesting that patients with pancreas divisum who are stented for one year have a lower frequency of attacks of pancreatitis than those not stented.171 The arguments against the association are the following: (1) there are studies to show that the rate of pancreatitis in pancreas divisum patients is the same as the general population172; (2) the observational reports are flawed in that follow-up was not long enough (usually only 1 to 2 years) and that recurrent acute pancreatitis is a disease of great variability163; (3) the single randomized study163 was flawed in that it was not blinded, was small (19 patients total), and its patients probably had chronic pancreatitis in that they had multiple pain attacks in between attacks of acute pancreatitis; (4) the risk of endoscopic therapy is considerable with a high rate of post-ERCP pancreatitis in patients with pancreas divisum,111,112 therefore, making the risk-benefit ratio of treating pancreas divisum endoscopically questionable; and (5) the rate of genetic abnormalities in patients with pancreas divisum and acute recurrent pancreatitis are either the same173 or higher161 than expected in the general population or population of patients with acute pancreatitis of other etiologies, suggesting a possible genetic source. For example, there appears to be a higher incidence of CFTR mutations in patients with pancreas divisum who develop acute pancreatitis.161 Therefore, it may not be the presence of pancreas divisum alone that predisposes to acute pancreatitis but other factors may be necessary to precipitate an attack.163

Chapter 58  Acute Pancreatitis Sphincter of Oddi Dysfunction (see Chapter 63)

SOD is also a controversial cause of pancreatitis. Investigators who study patients with recurrent acute pancreatitis report that SOD (usually defined as a basal pancreatic sphincter pressure > 40 mm Hg) is the most common abnormality discovered, occurring in approximately 35% to 40% of patients. The argument in favor of this entity as a cause of acute pancreatitis is the many observational series that report that endoscopic pancreatic sphincterotomy or surgical sphincteroplasty reduces or eliminates recurrent attacks of pancreatitis.174 The arguments against SOD as a cause of acute pancreatitis are (1) the lack of any prospective controlled blinded trials in the treatment of this disorder; (2) the short duration of follow-up in the observational reports; and (3) the high risk of pancreatitis (25% to 35%) associated with ERCP, sphincter of Oddi manometry, and pancreatic sphincterotomy in patients with suspected SOD. Furthermore, there is a relative dearth of data determining the normal range of pancreatic sphincter pressure.174

MISCELLANEOUS

Pancreatitis has been rarely associated with Crohn’s disease.164 A recent case control study from Denmark found a 4-fold increase in acute pancreatitis in patients with Crohn’s and a 1.5-fold increase in patients with ulcerative colitis. This has been attributed by some to the use of drugs such as aminosalicylates/sulfasalazine, azathioprine, or 6-mercaptopurine (see Table 58-4). Theories to support a relationship between idiopathic inflammatory bowel disease (IBD) and pancreatitis are that pancreatitis is an extraintestinal manifestation of IBD, that duodenal Crohn’s disease can cause obstruction to the flow of pancreatic juice, that granulomatous disease can involve the pancreas, or that there is an autoimmune process affecting the pancreas. Celiac disease165 has also been described in association with pancreatitis, but the relationship remains uncertain. It has been suggested that abnormalities in the normal barrier of the small bowel seen in patients with celiac disease may allow excessive absorption of amylase from the intestinal lumen, leading to hyperamylasemia. In the setting of abdominal pain in a patient with celiac disease, it is not uncommon to find elevations in the serum amylase in the absence of acute pancreatitis.166 Pancreatitis has been seen in patients after severe burns.167 A relationship of smoking with acute pancreatitis has been suggested. A Swedish case control study showed that there is a fourfold increased rate of acute pancreatitis in heavy smokers compared with nonsmokers.168 Autoimmune pancreatitis (discussed in the next chapter in more detail) typically presents as a mass or fullness in the pancreas. It is most commonly seen in older men, with biliary obstruction and an elevation of the serum immunoglobulin (IgG4) level. Occasionally, patients will present with signs and symptoms of chronic pancreatitis, such as stricturing of the main pancreatic duct, diabetes, exocrine pancreatic insufficiency, or as acute pancreatitis. Investigators have more recently described patients with autoimmune recurrent pancreatitis, especially in younger women often without the classic elevation of serum IgG4.169

CLINICAL FEATURES It is difficult to diagnose acute pancreatitis by history and physical examination because clinical features are similar to those of many acute abdominal illnesses (Table 58-6).

Table 58-6  Differential Diagnosis of Acute Pancreatitis Biliary colic Acute cholecystitis Perforated hollow viscus (e.g., perforated peptic ulcer) Mesenteric ischemia or infarction Intestinal obstruction Inferior wall myocardial infarction Dissecting aortic aneurysm Ectopic pregnancy

HISTORY

Abdominal pain is present at the onset of most attacks of acute pancreatitis, but the timing of abdominal pain is variable. Biliary colic may herald or progress to acute pancreatitis. Pain in pancreatitis usually involves the entire upper abdomen. However, it may be epigastric, in the right upper quadrant, or, infrequently, confined to the left side. Pain in the lower abdomen may arise from the rapid spread of pancreatic exudation to the left colon. Onset of pain is rapid but not as abrupt as that of a perforated viscus. Usually it is at maximal intensity in 10 to 20 minutes. Occasionally, pain gradually increases and takes several hours to reach maximum intensity. Pain is steady and moderate to very severe. There is little pain relief with changing position. Frequently, pain is unbearable, steady, and boring. Band-like radiation of the pain to the back occurs in half of patients. Pain that lasts only a few hours and then disappears suggests a disease other than pancreatitis, such as biliary colic or peptic ulcer. Pain is absent in 5% to 10% of attacks, and a painless presentation may be a feature of serious fatal disease.4 Ninety percent of patients have nausea and vomiting. Vomiting may be severe, may last for hours, may be accompanied by retching, and may not alleviate pain. Vomiting may be related to severe pain or to inflammation involving the posterior gastric wall.

PHYSICAL EXAMINATION

Physical findings vary with the severity of an attack. Patients with mild pancreatitis may not appear acutely ill. Abdominal tenderness may be mild, and abdominal guarding is absent. In severe pancreatitis, patients look severely ill and often have abdominal distention, especially epigastric, which is due to gastric, small bowel, or colonic ileus. Almost all patients are tender in the upper abdomen, which may be elicited by gently shaking the abdomen or by gentle percussion. Guarding is more marked in the upper abdomen. Tenderness and guarding are less than expected, considering the intensity of discomfort. Abdominal rigidity, as occurs in diffuse peritonitis, is unusual but can be present, and differentiation from a perforated viscus may be impossible in these instances. Bowel sounds are reduced and may be absent. Additional abdominal findings may include ecchymosis in one or both flanks (Grey Turner’s sign; Fig. 58-3) or about the periumbilical area (Cullen’s sign), owing to extravasation of hemorrhagic pancreatic exudate to these areas. These signs occur in less than 1% of cases and are associated with a poor prognosis. Rarely there is a brawny erythema of the flanks caused by extravasation of pancreatic exudate to the abdominal wall. A palpable epigastric mass may appear during the disease from a pseudocyst or a large inflammatory mass. The general physical examination, particularly in severe pancreatitis, may uncover markedly abnormal vital signs if

969

970

Section VII  Pancreas 65). The pain of a perforated ulcer is sudden, becomes diffuse, and precipitates a rigid abdomen; movement aggravates pain. Nausea and vomiting occur but disappear soon after onset of pain (see Chapter 52). In mesenteric ischemia or infarction, the clinical setting often is an older person with cardiac dysrhythmia or arteriosclerotic disease who develops sudden pain out of proportion to physical findings, bloody diarrhea, nausea, and vomiting. Abdominal tenderness may be mild to moderate, and muscular rigidity may not be severe despite severe pain (see Chapter 114). In intestinal obstruction, pain is cyclical, abdominal distention is prominent, vomiting persists and may become feculent, and peristalsis is hyperactive and often audible (see Chapter 119). Other conditions that enter into the differential diagnosis of acute pancreatitis are listed in Table 58-6. Figure 58-3.  Grey Turner’s sign. Ecchymoses in the flanks in a 24-year-old man with a two-day history of upper abdominal pain secondary to mild acute alcoholic pancreatitis. (Courtesy of Nadia Habal, MD, Dallas, Tex.)

LABORATORY DIAGNOSIS PANCREATIC ENZYMES

there are third-space fluid losses and systemic toxicity. Commonly, the pulse is 100 to 150 beats/minute. Blood pressure can be initially higher than normal and then lower than normal with third-space losses and hypovolemia. Initially the temperature may be normal, but within one to three days it may increase to 101°F to 103°F owing to the severe retroperitoneal inflammatory process and the release of inflammatory mediators from the pancreas.175 Tachypnea and shallow respirations may be present if the subdiaphragmatic inflammatory exudate causes painful breathing. Dyspnea may accompany pleural effusions, atelectasis, ARDS, or congestive heart failure. Chest examination may reveal limited diaphragmatic excursion if abdominal pain causes splinting of the diaphragm, or dullness to percussion and decreased breath sounds at the lung bases if there is a pleural effusion. There may be disorientation, hallucinations, agitation, or coma,176 which may be due to alcohol withdrawal, hypotension, electrolyte imbalance such as hyponatremia, hypoxemia, fever, or toxic effects of pancreatic enzymes on the central nervous system. Conjunctival icterus may be present due to choledocho­ lithiasis (gallstone pancreatitis) or bile duct obstruction from edema of the head of the pancreas, or from coexistent liver disease. Uncommon findings include subcutaneous nodular fat necrosis,177 thrombophlebitis in the legs, and polyarthritis. Subcutaneous fat necroses are 0.5- to 2-cm tender red nodules that usually appear over the distal extremities but may occur over the scalp, trunk, or buttocks. They occasionally precede abdominal pain or occur without abdominal pain, but usually they appear during a clinical episode and disappear with clinical improvement. If they occur over a joint, they may be confused with arthritis. Some physical findings point to a specific cause of acute pancreatitis. Hepatomegaly, spider angiomas, and thickening of palmar sheaths favor alcoholic pancreatitis. Eruptive xanthomas and lipemia retinalis suggest hyperlipidemic pancreatitis. Parotid pain and swelling are features of mumps. Band keratopathy (an infiltration on the lateral margin of the cornea) occurs with hypercalcemia.

DIFFERENTIAL DIAGNOSIS

The abdominal pain of biliary colic may simulate acute pancreatitis. It is frequently severe and epigastric, but it lasts for several hours rather than several days (see Chapter

In general, the diagnosis of acute pancreatitis relies on at least a three-fold elevation of amylase or lipase in the blood.178

Serum Amylase

In healthy persons, the pancreas accounts for 40% to 45% of serum amylase activity, and the salivary glands account for the rest. Simple analytic techniques can separate pancreatic and salivary amylases. Because pancreatic diseases increase serum pancreatic (P) isoamylase, measurement of P-isoamylase can improve diagnostic accuracy. However, this test is rarely used. The total serum amylase test is most frequently ordered to diagnose acute pancreatitis because it can be measured quickly and cheaply. It rises within 6 to 12 hours of onset and is cleared fairly rapidly from the blood (half-life, 10 hours). Probably less than 25% of serum amylase is removed by the kidneys. It is uncertain what other processes clear amylase from the circulation. The serum amylase is usually increased on the first day of symptoms, and it remains elevated for three to five days in uncomplicated attacks. The sensitivity of the serum amylase level for detecting acute pancreatitis is difficult to assess because an elevated amylase is often used to make the diagnosis. In mild attacks, other tests to confirm pancreatic inflammation are either not sensitive enough (pancreatic imaging studies, other biochemical markers) or not necessary (surgery). A limitation of serum amylase is that it is not 100% sensitive or specific. With respect to sensitivity, which is greater than 85%, the serum amylase may be normal or minimally elevated in fatal pancreatitis,4 during a mild attack or an attack superimposed on chronic pancreatitis (because the pancreas has little acinar tissue), or during recovery from acute pancreatitis. Serum amylase also may be falsely normal in hypertriglyceridemia-associated pancreatitis179 because an amylase inhibitor may be associated with triglyceride elevations. In this case, serial dilution of serum often reveals an elevated serum amylase. Hyperamylasemia is not specific for pancreatitis because it occurs in many conditions other than acute pancreatitis. In fact, one half of all patients with an elevated serum amylase may not have pancreatic disease.180 In acute pancreatitis, the serum amylase concentration is usually more than two to three times the upper limit of normal; it is usually less than this with other causes of hyperamylasemia.177 However, this level is not an absolute discriminator.

Chapter 58  Acute Pancreatitis Thus, an increased serum amylase level supports rather than confirms the diagnosis of acute pancreatitis. In addition, there are some individuals who have persistent hyperamylasemia without clinical symptoms. This has been reported due to macroamylasemia or pancreatic hyperamylasemia on a familial basis.174 Nonpancreatic diseases that cause hyperamylasemia include pathologic processes in organs (e.g., salivary glands, fallopian tubes) that normally produce amylase. Furthermore, mass lesions such as papillary cystadenocarcinoma of the ovary, benign ovarian cyst, and carcinoma of the lung, cause hyperamylasemia because they produce and secrete salivary-type isoamylase. Transmural leakage of pancreatic-type isoamylase and peritoneal absorption probably explain hyperamylasemia in intestinal infarction and in perforated viscus. Renal failure increases serum amylase up to four to five times the upper limit of normal due to decreased renal clearance of this enzyme.181 Patients on hemodialysis tend to have higher serum amylase levels than those on peritoneal dialysis. In patients with chronic kidney disease, there is no clear correlation between the creatinine clearance and serum levels of amylase, and about one third of patients with marked renal insufficiency have normal pancreatic enzyme levels. Chronic elevations of serum amylase (without amylasuria) occur in macroamylasemia. In this condition, normal serum amylase is bound to an immunoglobulin or abnormal serum protein to form a complex that is too large to be filtered by renal glomeruli and thus has a prolonged serum half-life.182 Macroamylasemia may complicate the diagnosis of pancreatic disease, but it has no other clinical consequence. The urinary amylase-to-creatinine clearance ratio (ACCR) increases from approximately 3% to approximately 10% in acute pancreatitis.183 However, even moderate renal insufficiency interferes with the accuracy and specificity of the ACCR. Other than to diagnose macroamylasemia, which has a low ACCR, urinary amylase and the ACCR are not used clinically. Macroamylasemia can also be measured directly using serum. Deliberate contamination of urine with saliva, as in Munchausen’s syndrome, can increase the urine amylase, with the serum amylase being normal. This situation can be excluded by measuring salivary amylase in the urine.

Serum Lipase

The sensitivity of serum lipase for the diagnosis of acute pancreatitis is similar to that of serum amylase and is between 85% and 100%.174 Lipase may have greater specificity for pancreatitis than amylase as serum lipase is normal when serum amylase is elevated as in salivary gland dysfunction, tumors, gynecologic conditions, and macroamylasemia. Serum lipase always is elevated on the first day of illness and remains elevated longer than does the serum amylase.184 Consequently some suggest combining lipase with amylase as a test for acute pancreatitis. However, we and others have found that combining enzymes does not improve diagnostic accuracy. Specificity of lipase can suffer from some of the same problems as those of amylase. In the absence of pancreatitis, serum lipase may increase less than two-fold above normal in severe renal insufficiency.185 With intra-abdominal conditions that resemble acute pancreatitis,186 lipase increases to levels less than three-fold above normal, presumably by absorption through an inflamed or perforated intestine. Rarely, a nonpancreatic abdominal condition such as small bowel obstruction can raise the amylase and lipase above three times normal. Some believe that serum lipase measurement is preferable to that of serum amylase because it is as sensitive as amylase measurement and more specific,

whereas others find no clear advantage of one over the other.5

Other Pancreatic Enzymes

During acute pancreatic inflammation, pancreatic digestive enzymes other than amylase and lipase leak into the systemic circulation and have been used to diagnose acute pancreatitis. They include PLA2, trypsin/typsinogen, car­ boxylester lipase, carboxypeptidase A, colipase, elastase, and ribonuclease. None, alone or in combination, are better than serum amylase or lipase, and most are not available on a routine basis.

OTHER BLOOD AND URINE TESTS

Many nonenzymatic proteins are overexpressed in acute pancreatitis. Pancreatitis-associated protein (PAP), a heat shock protein, is undetectable in the normal pancreas but markedly increases in acute pancreatitis, with PAP detectable in serum. The sensitivity of serum PAP or pancreaticspecific protein (PSP) is no better than that of conventional tests,187 but PAP and PSP are as accurate as serum amylase for the detection of acute pancreatitis. The methemalbumin level in the circulation increases in acute pancreatitis, but it also increases in serious intraabdominal conditions such as intestinal infarction, limiting its usefulness.

STANDARD BLOOD TESTS

The white blood cell count frequently is elevated, often markedly so in severe pancreatitis. The blood glucose also may be high and associated with high levels of serum glucagon. Serum aspartate transaminase (AST), alanine aminotransferase (ALT), alkaline phosphatase, and bilirubin also may increase, particularly in gallstone pancreatitis. Presumably, calculi in the bile duct account for these abnormalities. However, pancreatic inflammation may partially obstruct the distal bile duct in acute pancreatitis of other causes and cause abnormalities in liver biochemical tests. Nevertheless, aminotransferases may help distinguish biliary from alcoholic pancreatitis (see later).188 The erythrocyte mean corpuscular volume (MCV) has been also shown to help differentiate alcoholic from nonalcoholic acute pancreatitis.189 Alcoholic patients tend to have a higher MCV due to the toxic effects of alcohol on erythrocyte formation in the bone marrow. Serum triglyceride levels increase in acute pancreatitis, but also with alcohol use, uncontrolled diabetes mellitus, or defective tri­glyceride metabolism.

DIAGNOSTIC IMAGING ABDOMINAL PLAIN FILM

Findings on a plain radiograph range from no abnormalities in mild disease to localized ileus of a segment of small intestine (“sentinel loop”) or the colon cut-off sign in more severe disease. In addition, an abdominal plain film helps exclude other causes of abdominal pain, such as bowel obstruction and perforation. Images of the hollow GI tract on an abdominal plain radiograph depend on the spread and location of pancreatic exudate. Gastric abnormalities are caused by exudate in the lesser sac producing anterior displacement of the stomach, with separation of the contour of the stomach from the transverse colon. Small intestinal abnormalities are due to exudate in proximity to small bowel mesentery and include ileus of one or more loops of jejunum (the sentinel loop), of the distal ileum or cecum,

971

972

Section VII  Pancreas or of the duodenum. Generalized ileus may occur in severe disease. Besides ileus, other abnormalities of the hollow GI tract may be present. The descending duodenum may be displaced and stretched by an enlarged head of the pancreas. In addition, spread of exudate to specific areas of the colon may produce spasm of that part of the colon and either no air distal to the spasm (the colon cut-off sign) or dilated colon proximal to the spasm. Head-predominant pancreatitis predisposes to spread of exudate to the proximal transverse colon, producing colonic spasm and a dilated ascending colon. Uniform pancreatic inflammation predisposes spread of exudate to the inferior border of the transverse colon and an irregular haustral pattern. Exudate from the pancreatic tail to the phrenicocolic ligament adjacent to the descending colon may cause spasm of the descending colon and a dilated transverse colon. Other findings on plain radiography of the abdomen may give clues to etiology or severity, including calcified gallstones (gallstone pancreatitis), pancreatic stones or calcification (chronic pancreatitis with a bout of acute inflammation), and ascites (severe pancreatitis). Gas in the retroperitoneum may suggest a pancreatic abscess.

CHEST RADIOGRAPHY

Abnormalities visible on the chest roentgenogram occur in 30% of patients with acute pancreatitis, including elevation of a hemidiaphragm, pleural effusion(s), basal or plate-like atelectasis secondary to limited respiratory excursion, and pulmonary infiltrates. Pleural effusions may be bilateral or confined to the left side; rarely they are only on the right side.190 Patients with acute pancreatitis found to have a pleural effusion and/or infiltrate on admission are more likely to have severe disease.191 During the first 7 to 10 days, there also may be signs of congestive heart failure or acute respiratory distress syndrome. Pericardial effusion is rare.

ABDOMINAL ULTRASONOGRAPHY

Abdominal ultrasonography is used during the first 24 hours of hospitalization to search for gallstones, dilation of the bile duct due to choledocholithiasis, and ascites. If the pancreas is seen (bowel gas obscures the pancreas 25% to 35% of the time), it is usually diffusely enlarged and hypoechoic. Less frequently there are focal hypoechoic areas. There also may be ultrasonographic evidence of chronic pancreatitis, such as intraductal or parenchymal calcification and dilation of the pancreatic duct. Ultrasound is not a good imaging test to evaluate extrapancreatic spread of pancreatic inflammation or necrosis within the pancreas and consequently is not useful to ascertain severity of pancreatitis. During the course of acute pancreatitis, ultrasound can be used to evaluate progression of a pseudocyst (discussed later). Due to overlying gas, the diagnosis of cholelithiasis may be obscured during the acute attack but may be found after bowel gas has receded.

ENDOSCOPIC ULTRASONOGRAPHY

Usually endoscopic ultrasonography (EUS) is not helpful early in acute pancreatitis. Imaging of the pancreas during an attack of acute pancreatitis and weeks following an episode reveal signals that are not normal (typically hypoechoic) and indistinguishable from chronic pancreatitis and malignancy. However, after a month, especially in patients with idiopathic interstitial pancreatitis, EUS may help determine the presence of small tumors, pancreas divisum, and bile duct stones.192 EUS is equal to MRCP and ERCP but far more sensitive than either abdominal ultraso-

nography or CT in detecting common duct stones.193 In a patient with biliary pancreatitis, whose serum bilirubin is rising in the setting of biliary sepsis, ERCP should not be delayed by first performing EUS (see later). Although there has been some concern that ERCP can worsen pancreatitis in such settings, ERCP appears to be safe in acute pancreatitis if needed. One caveat is that the contrast instillation into the pancreatic duct could introduce infection into necrotic areas of the pancreas. For this reason, EUS might be the best method of evaluating the bile duct in a patient with necrotizing pancreatitis.194

COMPUTED TOMOGRAPHY

CT is the most important imaging test for the diagnosis of acute pancreatitis and its intra-abdominal complications.195 The three main indications for a CT in acute pancreatitis are to exclude other serious intra-abdominal conditions, such as mesenteric infarction or a perforated ulcer; to stage the severity of acute pancreatitis; and to determine whether complications of pancreatitis are present, such as involvement of the GI tract or nearby blood vessels and organs, including liver, spleen, and kidney.196 Helical CT is the most common technique. If possible, scanning should occur after the patient receives oral contrast, followed by intravenous contrast to identify any areas of pancreatic necrosis. If there is normal perfusion of the pancreas, interstitial pancreatitis is said to be present (see Fig. 58-1). Pancreatic necrosis manifested as perfusion defects after intravenous contrast may not appear until 48 to 72 hours after onset of acute pancreatitis (see Fig. 58-2). Contraindications to using intravenous contrast are a patient’s history of severe allergy (respiratory distress or anaphylaxis) or significant renal impairment (serum creatinine greater than 2  mg/dL). If severe renal impairment requires dialysis, intravenous contrast medium may be used.197 Hives or less severe allergic reactions with previous administration of iodinated contrast material are not absolute contraindications, but a nonionic contrast agent should be used, and 200 mg of hydrocortisone should be administered intravenously every six hours for four doses starting before the scan and 50 mg of diphenhydramine (Benadryl) should be given intramuscularly 30 minutes before the scan.198 It has been suggested that intravenous contrast media early in the course of acute pancreatitis might increase pancreatic necrosis because iodinated contrast medium given at the onset of pancreatitis increases necrosis in experimental rat acute pancreatitis.199 However, it did not do so in the opossum.200 Data in humans are conflicting. Two retrospective studies suggested that early contrast-enhanced CT worsened pancreatitis199 but this was not corroborated by a third retrospective study.198 The severity of acute pancreatitis has been classified into five grades (A to E) based on findings on unenhanced CT199 (see following). Although the presence of gas in the pancreas suggests pancreatic infection with a gas-forming organism, this finding can also accompany sterile necrosis (Fig. 58-4) with microperforation of the gut or adjacent pseudocyst into the pancreas.201 Moreover, the great majority of pancreatic infections occur in the absence of gas on CT scan.

MAGNETIC RESONANCE IMAGING

MRI provides similar information regarding the severity of pancreatitis as does CT. MRI is as good as CT in detecting necrosis and fluid collections. MRI is better than CT, but equal to EUS and ERCP in detecting choledocholithiasis202 The MRCP contrast agent gadolinium, previously thought

Chapter 58  Acute Pancreatitis

G P

Figure 58-4.  Acute necrotizing pancreatitis. Contrast-enhanced computed tomography that shows the pancreas (P) is surrounded by peripancreatic inflammation that contains bubbles of air (arrows) due to sterile necrosis. The patient was not clinically ill, and therefore an abscess was not considered likely. G, gallbladder.

Conventional transabdominal ultrasonography should be performed in every patient with a first attack of acute pancreatitis to search for gallstones in the gallbladder, common duct stones, or signs of extrahepatic biliary tract obstruction. However, bile duct stones are frequently missed by transabdominal ultrasonography, and most stones pass during the acute attack. ERCP is limited to patients with severe acute pancreatitis due to gallstones with persistent bile duct obstruction and to those in whom the stone could not be removed during surgery. In most patients with biliary pancreatitis the bile duct can be imaged with an operative cholangiogram at the time of laparoscopic cholecystectomy performed during the same admission. Although EUS is the most accurate method of detecting common duct stones and has been recommended for evaluating the common duct prior to cholecystectomy, it is rarely needed in this setting. MRCP is another noninvasive test that is highly accurate in determining if common duct stones are present. If a common duct stone is found at surgery, it is either removed at operation or endoscopically after surgery. Laparoscopic exploration of the bile duct is as safe and effective as postoperative ERCP in clearing stones from the common duct.209

PREDICTORS OF SEVERITY to be safe in patients with renal failure,203 can cause nephrogenic systemic fibrosis (NSF), which has raised concern.204 MRI is less accessible and more expensive than CT. MRI also requires the patient to remain still during capture of images, which typically is longer than with spiral CT. The use of intravenous secretin prior to MRCP allows a better visualization of the pancreatic ducts. This has been shown to be particularly useful in the evaluation of patients with idiopathic pancreatitis and recurrent pancreatitis.205

DISTINGUISHING ALCOHOLIC FROM GALLSTONE PANCREATITIS Differentiation between alcoholic and gallstone pancreatitis is important because eliminating these causes may prevent further attacks of pancreatitis. Alcoholic pancreatitis occurs more frequently in men approximately 40 years old. The first clinical episode usually occurs after 5 to 10 years of heavy alcohol consumption. By contrast, biliary pancreatitis is more frequent in women, and the first clinical episode is often after the age of 40 years. Recurrent attacks of acute pancreatitis suggest an alcoholic etiology, but unrecognized gallstones may cause recurrent pancreatitis. Among patients with acute biliary pancreatitis discharged from hospital without cholecystectomy, 30% to 50% develop recurrent acute pancreatitis relatively soon after discharge (average time to recurrent pancreatitis, 108 days).206 Thus, removing the gallbladder in biliary pancreatitis is imperative. Laboratory tests may help distinguish between these two disorders. The specificity for gallstone pancreatitis of ALT concentration greater than 150 IU/L (approximately a threefold elevation) is 96%; the positive predictive value is 95%, but the sensitivity is only 48%.205 The AST concentration is nearly as useful as the ALT, but the total bilirubin and alkaline phosphatase concentrations are not as helpful to distinguish gallstone pancreatitis from alcoholic and other etiologies. There are differing reports as to whether a high serum lipase-to-amylase ratio can differentiate alcoholic from other causes of pancreatitis.207,208

Predicting severity of pancreatitis early in the course of disease is critical to maximize therapy and to prevent and minimize organ dysfunction and complications. Unfortunately the management of patients with acute pancreatitis is complicated by the inability to distinguish mild from severe disease during the early stages. The definition of the severity of acute pancreatitis early in the course of disease (during the first week) is typically based on clinical rather than anatomic parameters. At admission, several potential risk factors of severity and measurements that may reflect severity should be documented including age, body mass index, elevated hematocrit, elevated blood urea nitrogen (BUN), and pleural effusions or infiltrates on admission chest radiograph. The height of elevation of the serum amylase and lipase does not correlate with severity. Obese patients with pancreatitis have a higher incidence of local complications,210 respiratory failure,211 severe acute pancreatitis,212 and death from sterile necrosis213 than do nonobese patients. Initially at presentation and over the first 48 hours, patients should be classified temporarily as having severe acute pancreatitis (and managed as such initially) based on the presence of SIRS or organ failure. SIRS is defined by two or more of the following four criteria: pulse greater than 90 beats/minute; rectal temperature less than 36°C or more than 38°C; white blood count less than 4000 or more than 12,000/mm3; and respirations greater than 20/minute or Pco2 less than 32 mm Hg. The presence of SIRS at admission and persistence of SIRS to 48 hours increases the morbidity and mortality rate. In one study, 25% of patients with persistent SIRS died from acute pancreatitis, 8% with transient SIRS, and less than 1% without SIRS.214 Although severity is now defined by the presence of organ failure or anatomic complications of acute pancreatitis, such as pancreatic necrosis, prospective systems using clinical criteria have been developed to determine severity in patients with acute pancreatitis. These systems include Ranson criteria (see Table 58-2) and APACHE score.13,14 Unfortunately these scoring systems (discussed following) are cumbersome, requiring multiple measurements. Addi-

973

974

Section VII  Pancreas tionally, the systems are not accurate until 48 hours after presentation.

SCORING SYSTEMS Ranson’s Score

Ranson and colleagues identified 11 signs that had prognostic significance during the first 48 hours. The original list14 was analyzed in patients who primarily suffered from alcoholic pancreatitis and was then modified 8 years later for those with gallstone pancreatitis (see Table 58-2).215 Higher Ranson’s scores predict more severe disease. In mild pancreatitis (scores < 2), the mortality is 2.5% and in severe pancreatitis (scores > 3) the mortality is 62%.216 Also, the higher the Ranson’s score the higher the incidence of systemic complications, necrosis, and infected necrosis. These lists continue to remain in wide use in both the United States and Europe.217 The Ranson criteria have several drawbacks. First, the list is cumbersome and there are two lists to follow depending on suspected etiology (see Table 58-2). Second, an accurate Ranson’s score takes 48 hours to compute and the criteria have not been validated beyond the 48-hour time limit. Third, not all laboratories measure all the parameters in routine blood tests (e.g., serum lactate dehydrogenase [LDH]). Fourth, the overall sensitivity of the Ranson criteria (using three signs as the cutoff) for diagnosing severe disease is only 40% to 88% and the specificity is only 43% to 90%. The positive predictive value is approximately 50% and the negative predictive value around 90%.218 Therefore, the best use of Ranson’s score is to exclude severe disease. The Imrie or Glasgow score219 is a slightly simplified list (eight criteria) that is used commonly in the United Kingdom. It has similar drawbacks to the Ranson score.

APACHE-II Scores

APACHE-II is another commonly used scoring system in the United States to predict severity. It has the advantage of being able to be used on a daily basis and has similar positive and negative predictive values as the Ranson score at 48 hours after admission. The APACHE-II system assigns points for 12 physiologic variables, for age, and for chronic health status, in generating a total point score. The 12 variables are temperature; heart rate; respiratory rate; mean arterial blood pressure; oxygenation; arterial pH; serum potassium, sodium, and creatinine; hematocrit; white blood cell (WBC); and Glasgow Coma Scale. APACHE-II scores on admission and within 48 hours help distinguish mild from severe pancreatitis and to predict death.218,220 Most patients survive if APACHE-II scores are 9 or less during the first 48 hours. However, patients with APACHE-II scores of 13 or more have a high likelihood of dying. At admission, sensitivity is 34% to 70%, and specificity is 76% to 98%. At 48 hours, sensitivity remains less than 50%, but specificity is close to 90% to 100%.220 Strong drawbacks are its complexity, its low sensitivity on admission, and the fact that at 48 hours the score is no better than other scoring systems.221 Like the Ranson criteria, the APACHE-II score has its highest value in predicting mild disease.

BISAP

The problem with scoring systems is that they are cumbersome, using multiple variables. As described above, accuracy in predicting morbidity and/or mortality of the most commonly used scoring systems, Ranson and APACHE, is typically not achieved until 48 hours. By this time, it is usually apparent that the patient has developed severe disease manifested by organ failure. In order to develop a simple scoring system for patients with acute pancreatitis

that would be useful within the first 12 hours from admission, the Pancreas Center at Brigham and Women’s Hospital performed a series of studies retrospectively and prospec­ tively.221a,221b The studies were performed on a large data­ base including almost 37,000 patients and more than 200 hospitals. After careful analysis, including a validation study, they determined that a simple system that included 5 variables could accurately determine severity early in the course of the disease. The scoring system, referred to as BISAP (Bedside Index for Severity in Acute Pancreatitis), also uses the first letter of each parameter for 1 point. The BISAP score provides a single point for 5 parameters: blood urea nitrogen (BUN) greater than 25 mg/dL, impaired mental status, systemic inflammatory response syndrome, age greater than 60, and/or the presence of a pleural effusion, for a possible total of 5 points. A BISAP score greater than 3 is associated with a seven- to twelve-fold increase in developing organ failure.221b Accurate, yet much easier to use, this new simple scoring system appears to be useful in the early identification of patients who are at risk of developing complications and mortality.221a

Blood Urea Nitrogen

Several prognostic scoring systems, including the Ranson criteria and BISAP, incorporate blood urea nitrogen (BUN) for the prediction of mortality in patients with acute pancreatitis. Hemoconcentration, as described above, has been shown to be an accurate predictor of necrosis and organ failure. Both BUN and the hematocrit or hemoglobin are routine laboratory tests that may provide information on changes in intravascular volume status. Either test may be used in monitoring the early response to initial fluid resuscitation. Wu and colleagues221c recently performed a large observational cohort study on data from 69 U.S. hospitals and found that BUN may be superior to hemoglobin (not hematocrit). For every 5  mg/dL increase in BUN during the first 24 hours, the age- and genderadjusted odds ratio for mortality increased by 2.2. Of multiple routine laboratory tests examined, BUN yielded the highest accuracy at 24 hours and 48 hours. Although further study is needed, this paper suggests that following serial BUN measurements would be the most valuable single routine laboratory test for predicting mortality in acute pancreatitis.

ORGAN FAILURE

There is considerable interest among pancreatologists in using organ failure to predict severity. The Atlanta criteria defined which organ systems are of importance: pulmonary, renal, and cardiovascular. However, these criteria did not attempt to quantitate or prognosticate using organ failure. It has been appreciated that multiorgan failure or persistent single organ failure has a greater associated mortality than transient single organ failure. Multisystem organ failure is defined as two or more organs failing on the same day, rather than one organ failing on one day and another failing on the subsequent day. Patients with multisystem organ failure or persistent organ failure have a much higher mortality rate (approaching 50%) compared with patients with single and transient organ failure.222 Persistent organ failure is defined as lasting greater than 24 hours regardless of intervention. Survival among patients with organ failure at admission has also been shown to correlate with the duration of organ failure. When organ failure is corrected within 48 hours, mortality is close to zero. When organ failure persists for more than 48 hours, mortality is 36%.223 The Marshall Scoring System224 for organ failure is commonly used by intensivists for patients admitted to an intensive

Chapter 58  Acute Pancreatitis car unit. Data have not yet been generated using this system to prognosticate mortality in acute pancreatitis. Studies are needed to determine if this scoring system improves on the Ranson and APACHE scoring systems.

PERITONEAL LAVAGE

Percutaneous recovery of any volume of peritoneal fluid with a dark color or recovery of at least 20 mL of free intraperitoneal fluid of dark color portends a significant mortality.225 The sensitivity of peritoneal lavage is 36% to 72%, and the specificity is greater than 80% to 100%.226 An advantage is that it can be used any time, but it has not gained wide acceptance because it is invasive.

severity, such as CRP, TAP is not a surrogate marker of inflammation. Normally trypsinogen is cleaved to trypsin in the intestinal lumen by the enzyme enterokinase. Premature intrapancreatic activation during acute pancreatitis results in the release of TAP. The degree of pancreatic necrosis and systemic inflammatory response or sepsis is directly related to TAP concentration. Elevated urinary TAP (>30  nmol/L) correlates with disease severity. The test can be applied within 12 hours of admission. The positive predictive value of an elevated TAP is 80% and the negative predictive value approaches 100%.

Procalcitonin

Because the degree of elevation of serum amylase and lipase does not distinguish mild from severe pancreatitis,219 other factors have been examined.

This propeptide is another acute-phase reactant that has been shown to differentiate mild from severe acute pancreatitis within the first 24 hours after symptom onset. A serum strip test has been developed that has a sensitivity of 86% and a specificity of 95% in detecting organ failure.234

Hematocrit

COMPUTED TOMOGRAPHY

LABORATORY MARKERS

A high hematocrit on admission, or one that fails to decrease after 24 hours of rehydration is thought to be a sign of hemoconcentration due to retroperitoneal fluid loss and thus a marker of severe disease.227 One study showed that a hematocrit greater than 44% had a sensitivity of 72% on admission and of 94% after 24 hours in detecting organ failure. The negative predictive value at 24 hours was 96%. Although one study from Germany found no correlation between admission hematocrit and organ failure, most investigators have found hematocrit to be important in the management of patients with acute pancreatitis.228 An elevated hematocrit (>44%) is a predictor for the development of necrosis. The hematocrit should be observed at admission for prognostic purposes and followed prospectively to assist in guiding the rate of intravenous hydration.12

C-Reactive Protein

CRP is an acute-phase reactant produced by the liver and is used extensively in Europe as a marker of severe pancreatitis. CRP is inexpensive to measure and readily available. The sensitivity for detecting severe disease is 60% to 100% (using cutoffs of 100 to 210 mg/L, or 10 to 21 mg/dL) and the specificity is 75% to 100%.229

Interleukin-6

IL-6 is an acute-phase-reactant cytokine that is produced by a variety of cells and induces hepatic synthesis of CRP. Several studies have shown that it is a reasonably good marker to differentiate mild from severe disease, but the test is not readily available.230

Polymorphonuclear Leukocyte Elastase

Polymorphonuclear leukocyte elastase rises very early (before CRP) in acute pancreatitis. High levels have been reported to differentiate severe from mild disease,231 but the test is not generally available.

Phospholipase A2

PLA2 is involved in the release of prostaglandins from cell membranes and degrades surfactant in the lung. It may play a role in the pulmonary dysfunction associated with acute pancreatitis. Levels of catalytic type II PLA2 have been reported to differentiate between mild and severe disease within 24 hours of admission.232

Urinary Trypsinogen Activation Peptide

Urinary TAP may serve as an early predictor of severity in patients with acute pancreatitis.233 Unlike other markers of

The finding of extensive fluid collections or extensive necrosis on CT has been correlated with severe disease. Balthazar reported that 5 of 37 (13.5%) patients who had grade D or E findings on CT died, as opposed to none of 51 who had grades B or C findings (Table 58-7).119 Using the CT severity index (CTSI score) (see Table 58-7), among those with a score of 0 to 6, 3 of 77 (3.8%) died, as compared with 2 of 11 (18%) with scores of 7 to 10. The CT grading scores correlate better with local complications (pseudocysts and abscesses) than with mortality. Among the 37 patients with a grade D or E score, 54% developed a local complication, whereas only 2 of 51 (3.9%) with grades A through C developed this problem.211 Thus, the data do not confirm that the CTSI is any more predictive than the grades A through E score. There is controversy in the literature as to whether the extent of necrosis on CT predicts organ failure. Two studies

Table 58-7  Computed Tomography (CT) Grading System of Balthazar and CT Severity Index (CTSI) Balthazar Grades Grade A

Normal pancreas consistent with mild pancreatitis Grade B Focal or diffuse enlargement of the gland, including contour irregularities and inhomogeneous attenuation but without peripancreatic inflammation Grade C Grade B plus peripancreatic inflammation Grade D Grade C plus associated single fluid collection Grade E Grade C plus two or more peripancreatic fluid collections or gas in the pancreas or retroperitoneum CTSI = Balthazar Grade Score Plus Necrosis Score* Balthazar grade score: A = 0 B = 1 C = 2 D = 3 E = 4 Necrosis score: Absence of necrosis = 0 Necrosis of up to 33% of pancreas = 2 Necrosis of 33% to 50% = 4 Necrosis of >50% = 6 *Highest attainable score = 10 (Balthazar grade E + necrosis >50%).

975

976

Section VII  Pancreas failed to show any correlation between these two events,223,235 whereas a third study found a strong correlation.236

CHEST RADIOGRAPHY

A pleural effusion documented within 72 hours of admission190,193 by chest radiography (or CT) correlates with severe disease.

TREATMENT (Fig. 58-5) GENERAL CONSIDERATIONS

Patients with acute pancreatitis require adequate intravenous hydration and adequate analgesia to eliminate or markedly reduce pain. The patient is usually on nothing

Early course: 0-72 hours Is there organ failure? No

Yes

Admission to medical/surgical floor NPO, IV hydration (250-400 cc/hr) Nasal oxygen Frequent evaluation of oxygen saturation Hematocrit daily Serum electrolytes daily Pain control

Admission to an ICU Same orders as for floor admission Central line placement Evaluate need for assisted ventilation Assess for bile duct obstruction If bilirubin rising, consider urgent ERCP

Later course: >72 hours Evidence of severe disease or organ failure? No

Yes

Early refeeding Evaluate for etiology If GS, early cholecystectomy If ETOH, address psychosocial issues If high serum TGs, medical therapy

Interstitial pancreatitis on CT without peripancreatic necrosis: Continue supportive care Observation

To ICU if patient not already there Observe for biliary sepsis; if present, consider emergency ERCP Enteral feedings (NJ or NG) CT to evaluate for necrosis

Pancreatic/peripancreatic necrosis on CT: Continue supportive care Enteral feedings If infection suspected, consider antibiotics

Late course: 7-28 days Patient improving? Yes

No

Consider oral refeeding

If on antibiotics, consider FNA of pancreas for culture and change of antibiotics If not on antibiotics and FNA negative, keep off antibiotics Beyond 28 days Patient improving?

Yes Consider refeeding If patient cannot tolerate feedings, consider necrosectomy

No Consider necrosectomy by endoscopic, radiologic, or surgical means

Figure 58-5.  Algorithm for the management of acute pancreatitis at various stages in its course. CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; ETOH, ethyl alcohol; FNA, fine-needle aspiration; GS, gallstones; ICU, intensive care unit; IV, intravenous; NG, nasogastric; NJ, nasojejunal; NPO, nil per os (nothing by mouth); TGs, triglycerides.

Chapter 58  Acute Pancreatitis per mouth until any nausea and vomiting have subsided. Abdominal pain can be treated with opiate analgesics, often by a patient-controlled anesthesia pump. Opiate dosing is monitored carefully and adjusted on a daily basis according to ongoing needs. Although morphine has been reported to increase sphincter of Oddi tone and increase serum amylase,237 its use to treat the pain of pancreatitis has not been shown to adversely effect outcome. Nasogastric intubation is not used routinely because it is not beneficial in mild pancreatitis. It is used only to treat gastric or intestinal ileus or intractable nausea and vomiting. Similarly, proton pump inhibitors or H2-receptor blocking agents69 are not beneficial and not used. The patient should be carefully monitored for any signs of early organ failure such as hypotension, pulmonary, or renal insufficiency by closely following vital signs and urinary output. Tachypnea should not be assumed to be due to abdominal pain; monitoring oxygen saturation and, if needed, blood gases is advised and oxygen supplementation is mandatory if there is hypoxemia. It cannot be overemphasized that any patient who exhibits signs of early organ dysfunction should be immediately transferred to intensive care monitoring as deterioration can be rapid and fatal.

FLUID RESUSCITATION

As the inflammatory process progresses early in the course of the disease, there is an extravasation of protein-rich intravascular fluid into the peritoneal cavity resulting in hemoconcentration. The decreased perfusion pressure into the pancreas leads to microcirculatory changes that lead to pancreatic necrosis. An admission hematocrit of more than 47% and a failure of the admission hematocrit to decrease at 24 hours have been shown to be predictors of necrotizing pancreatitis.238 The relationship of hematocrit to severity of pancreatitis implies that the opposite is also true. Early vigorous intravenous hydration for the purpose of intravascular resuscitation is of foremost importance. The goal is to decrease the hematocrit. Laboratory and clinical studies with intravenous dextran to promote hemodilution have suggested efficacy in preventing severe disease.239 Too often patients with acute pancreatitis are given suboptimal intravenous hydration. One of the markers of severity of pancreatitis defined by Ranson and colleagues is intravascular losses (“fluid sequestration”). Ranson and colleagues found that a sequestration of more than 6 L of fluids during the first 48 hours was an independent predictor of disease severity in nongallstone pancreatitis.14 If this amount of fluid (6 L) is added to the minimal intravenous fluid requirements of a 70-kg person during the first 48 hours, intravenous hydration should be at least 250 to 300 mL/hour for 48 hours. The rate of volume replacement is likely to be more important during the first 24 hours, when a rising hematocrit has been shown to correlate closely with severe disease. A study from the Mayo Clinic showed that patients with severe acute pancreatitis who do not receive at least one third of their initial 72-hour cumulative intravenous fluid volume during the first 24 hours after emergency department presentation are at risk for greater mortality than those who are initially resuscitated more aggressively.240 Maintaining adequate intravascular volume in patients with severe disease may require 5 to 10 L of fluid such as isotonic saline daily for the first several days (200 to 400 mL/ hour). Respiratory distress often suggests development of ARDS independent of intravascular volume status. However, in a patient with unclear cardiac output, a Swan-Ganz catheter can be useful to gauge fluid resuscitation and to avoid congestive heart failure.

RESPIRATORY CARE

Hypoxemia (oxygen saturation <90%) requires supplemental oxygen, ideally by nasal prongs or by face mask if needed. If nasal oxygen fails to correct hypoxemia or if there is fatigue and borderline respiratory reserve, endotracheal intubation and assisted ventilation are required early. It is important to use a Swan-Ganz catheter to determine whether hypoxemia is due to congestive heart failure (increased pulmonary artery wedge pressure) or to primary pulmonary damage (normal or low pulmonary artery wedge pressure). Due to the common and indolent nature of hypoxemia affecting patients with acute pancreatitis, current guidelines recommend the initial routine use of nasal cannula oxygen to all patients with acute pancreatitis.12 ARDS is the most serious respiratory complication of acute pancreatitis; it is associated with severe dyspnea, progressive hypoxemia, and increased mortality. It generally occurs between the second and seventh day of illness (but can be present on admission) and consists of increased alveolar capillary permeability causing interstitial edema. Chest radiography may show multilobar alveolar infiltrates. Treatment is endotracheal intubation with positive endexpiratory pressure ventilation, often with low tidal volumes to protect the lungs from volutrauma. No specific treatment will prevent or resolve ARDS. After recovery, pulmonary structure and function usually return to normal.

Cardiovascular Care

Cardiac complications of severe acute pancreatitis include congestive heart failure, myocardial infarction, cardiac dysrhythmia, and cardiogenic shock. An increase in cardiac index and a decrease in total peripheral resistance may be present and respond to infusion of crystalloids. If hypotension persists even with appropriate fluid resuscitation, intravenous dopamine may help maintain the systemic blood pressure. Dopamine does not impair the microcirculation of the pancreas as do other vasoconstrictors.

METABOLIC COMPLICATIONS

Hyperglycemia may present during the first several days of severe pancreatitis but usually normalizes as the inflammatory process subsides. Blood sugars fluctuate, and insulin should be administered cautiously. Hypocalcemia due to low serum albumin causes no symptoms and requires no specific therapy. However, reduced serum ionized calcium may occur and cause neuromuscular irritability. If hypomagnesemia coexists, magnesium replacement should restore serum calcium to normal. Causes of magnesium depletion include vomiting, loss of magnesium in the urine, or deposition of magnesium in areas of fat necrosis. Once the serum magnesium is normal, signs or symptoms of neuromuscular irritability may require administering intravenous calcium gluconate as long as the serum potassium is normal and digitalis is not being given. Intravenous calcium increases calcium binding to myocardial receptors, which displaces potassium and may induce a serious dysrhythmia.

ANTIBIOTICS

Antibiotics are not indicated in mild pancreatitis. However, pancreatic sepsis (infected necrosis and, less often, abscess) and nonpancreatic sepsis (line sepsis, urosepsis, or pneumonia) are major sources of morbidity and mortality in severe acute pancreatitis. Thus, it would seem logical to consider antibiotic prophylaxis to improve the outcome. The use of prophylactic antibiotics in acute pancreatitis to prevent complications is controversial. In the 1970s, controlled studies compared intravenous antibiotics to no

977

978

Section VII  Pancreas therapy in the treatment of mild acute alcoholic pancreatitis,241 with negative results. However, low-risk patients were studied (mild alcoholic disease with no mortality) and the wrong antibiotic (ampicillin) was used.242 In the 1980s, the bacteriology of infected pancreatic tissue was elucidated by analyzing either surgical specimens243 or fine-needle aspirations of the pancreas.242 The majority of organisms detected were gram-negative aerobic or anaerobic species (Escherichia coli, Enterobacter aerogenes, Pseudomonas aeruginosa, Proteus species, Klebsiella pneumoniae, Citrobacter freundii, and Bacteroides species), with occasional gram positives (Streptococcus faecalis, Staphylococcus aureus, Streptococcus viridans, Staphylococcus epidermidis) and rare fungi (Candida species). Studies in the early 1990s elucidated the appropriate antibiotics to use given these organisms and the level of penetration of antibiotics into necrotic pancreatic tissue.244 Imipenem, fluoroquinolones (ciprofloxacin, ofloxacin, pefloxacin), and metronidazole emerged as the drugs that achieved the highest inhibitory concentrations in pancreatic tissue, whereas aminoglycosides did not. Several randomized controlled (no placebo given), nonblinded clinical trials of prophylactic intravenous antibiotics in patients with severe pancreatitis were performed in Europe in the 1990s.245-248 One study added oral nonabsorbable antibiotics to the intravenous antibiotic regimen.249 Meta-analyses of the intravenous antibiotic trials showed that mortality of necrotizing pancreatitis was significantly reduced by antibiotics.250,251 Very little comparative information is known as to which antibiotic is the most effective. One study compared pefloxacin to imipenem in severe disease, showing imipenem to be significantly more effective in reducing pancreatic and extrapancreatic infection; however, mortality was unaffected.252 Likewise, little information is available on the length of treatment. One investigation compared one week of ciprofloxacin to three weeks and showed that longer treatment reduced the rates of pancreatic and nonpancreatic infection.253 The aforementioned studies have been criticized because they were not placebo-controlled, double-blinded studies. Furthermore, the use of prophylactic antibiotics in all patients with severe pancreatitis raises the concern that some patients will become superinfected with resistant organisms or fungi, which might lead to greater mortality in the future. Two studies have used double-blind randomized protocols to study this question. The first randomized 114 patients with severe acute pancreatitis to prophylactic ciprofloxacin plus metronidazole or to placebo. Patients suspected of having an infection were allowed open-label antibiotics. There were no differences in rates of infected necrosis or mortality. The only difference noted was that the group randomized to receive placebo eventually received more open-label antibiotics than the group randomized to receive the original antibiotics (46% vs. 28%).254 Dellinger and colleagues255 performed a multicenter, double-blind placebo-controlled randomized study in 32 centers in North America and Europe. One hundred patients were equally randomized to two groups, meropenem (1  g intravenously every eight hours) or placebo within 5 days of the onset of symptoms. Meropenem was continued for 7 to 21 days. This study demonstrated no significant difference between the treatment groups for pancreatic or peripancreatic infection, mortality, or requirement for surgical intervention. Based on these last two placebo-controlled studies, routine use of antibiotics is questionable in the absence of biliary sepsis or obvious pancreatic or peripancreatic infection. Although practice guidelines published prior to this latest paper recommended the use of prophylactic antibiotics in patients

with severe necrotizing pancreatitis,256 more recent guidelines12 state that prophylactic antibiotics should not be used for the purpose of preventing infection in patients with necrotizing pancreatitis.

ENDOSCOPIC

The question of early removal of a possibly impacted gallstone in improving the outcome of gallstone pancreatitis remains a controversial issue. There have been three randomized fully published studies comparing urgent ERCP plus sphincterotomy (for any retained stones) versus conventional treatment in the management of gallstone pancreatitis.257-259 The earliest study,257 a single-center investigation from England, found that urgent ERCP within 72 hours of admission improved the outcome (complications and mortality) of patients with severe, but not mild, acute gallstoneinduced pancreatitis. The second,258 another single-center study from Hong Kong, found that the urgent intervention group had a reduction in biliary sepsis and a trend toward lower mortality compared with the control group. The third and largest study from Germany259 included 22 centers and came to an opposite conclusion, namely, that there was a higher complication rate and a trend toward higher mortality in the urgently treated ERCP group compared with standard nonurgent therapy. Differences in the designs of these studies do not allow direct comparisons. However, there is consensus that severe acute gallstone pancreatitis with ascending cholangitis (jaundice and fever) is an indication for urgent ERCP (see Chapter 61). In a prospective study of biliary pancreatitis using ERCP, Uomo and colleagues51 noted a 30.5% incidence of main pancreatic duct leakage. It has been proposed that early endoscopic stenting of the main pancreatic duct in patients with this problem may shorten hospital stay and the need for subsequent necrosectomy. Lau and colleagues260 evaluated 144 patients with severe acute pancreatitis and found that the presence of a PD leak was significantly associated with the development of necrosis. Patients with a PD leak had a longer length of stay compared with the patients with acute pancreatitis with no PD leak. In this retrospective study, patients who underwent early ERCP and had a PD stent placed were less likely to have other more invasive interventions performed, such as placement of external drains. Although PD stents have been shown to be helpful in the management of late complications related to pancreatic duct disruption, such as fistulas and pseudocysts, it remains unclear that routine use of these stents will prevent late complications. Kozarek and associates found that PD stent placement was associated with polymicrobial contamination of the pancreas.261 Although the contamination was largely asymptomatic, the benefit of placing a stent in the PD early in the course of acute pancreatitis to treat a duct disruption must be weighed against the possibility of seeding sterile necrosis with organisms that could lead to infected necrosis. Until randomized studies are performed, it is not clear whether potential advantages of early pancreatic duct stenting outweigh the risks.

NUTRITIONAL

In general, intravenous feedings are continued until patients are able to tolerate liquids or solids. The timing of early feedings is unclear. One report suggested that early refeeding improved outcome and allowed early discharge.262 However, a meta-analysis of three studies showed that early refeeding prolonged hospitalization.263 The question of whether an elevated serum amylase or lipase should influence the clinician to prolong the time to refeeding has been addressed. One hundred sixteen patients with acute pancre-

Chapter 58  Acute Pancreatitis atitis were fed at the clinician’s discretion; 21% developed pain on refeeding 250  kcal/day.264 If the serum lipase was more than three-fold elevated, refeeding increased the clinical relapse rate compared with the group in which the lipase was less than three-fold elevated (39% vs. 16%). However, the corollary of this finding is that most patients with threefold elevated serum lipase levels do not have a recrudescence of their pain on refeeding. Once refeeding is begun, a low-fat diet appears to be comparable to a clear liquid diet as the initial meal.265 Formerly, total parenteral nutrition (TPN) was the standard of care for refeeding patients with severe acute pancreatitis. Sax and coworkers266 randomized 54 patients with mild pancreatitis to intravenous nutrition or to TPN. The TPN group had a greater number of septic complications and longer hospitalizations. McClave267 randomized 30 patients with mild to moderate pancreatitis to receive TPN or enteral feedings administered through a nasoenteric tube, beginning 48 hours after admission. The Ranson and APACHE scores and blood glucose levels normalized more quickly in the enteral group, and the length of hospitalization showed a trend toward a shorter stay in the enteral group. Windsor268 randomized 34 patients with either mild to moderate pancreatitis to either oral feedings or TPN or with severe pancreatitis to either enteral feedings via a nasoenteric tube or TPN. The group receiving oral or enteral feedings had shorter intensive care unit (ICU) stays and improved acute phase response markers and disease severity scores compared with the TPN group. Kalfarentzos269 randomized 38 patients with severe necrotizing pancreatitis to TPN versus nasoenteric feedings. The enteral group had fewer septic complications and fewer total complications, although hospital stay, ICU stay, and days until resumption of a regular diet were similar in the two groups. Thus, these studies demonstrate that enteral nutrition is cheaper and safer and is preferable in patients with severe acute pancreatitis. It is still unclear, however, when nutrition should be initiated and for how long it needs to be continued. Furthermore it is unclear if nasoenteric feedings are needed or if nasogastric or even oral feedings are similarly effective if the patient tolerates this modality. Along those lines, a UK group randomized 50 patients with severe pancreatitis to nasogastric versus nasoenteric tube feedings. No difference was seen in the ability to tolerate feedings, in markers of inflammation, or in morbidity or mortality between the groups.270 They also noted no differences in the length of hospital stay or complications in patients with severe acute pancreatitis.

SURGICAL THERAPY

Cholecystectomy is routinely performed in patients with gallstone pancreatitis, and a consensus conference suggested that in mild or severe gallstone pancreatitis, cholecystectomy should be performed as soon as the patient has recovered and the acute inflammatory process has subsided.271 A second potential role for surgery in pancreatitis is to debride pancreatic necrosis (necrosectomy) or drain a pancreatic abscess. With regard to pancreatic necrosectomy, the data are more complicated and evolving. Studies in the 1980s suggested that early necrosectomy (within the first week of hospitalization for severe disease) reduced mortality. However, in the only randomized study comparing early (within 72 hours of admission) to late necrosectomy (12 days or more after admission), the mortality with early operation was greater than with later débridement (56% vs. 27%).272 Some investigators have reported that it is important to differentiate sterile necrosis from infected necrosis by fine-needle aspiration of the pancreas. Sterile necrosis

S F

P

Figure 58-6.  Walled-off pancreatic necrosis. Non–contrast-enhanced computed tomography reveals a 5.4-cm pus-filled fluid collection (arrows) with the tip of an aspirating needle in its lumen. The abscess is anterior to the pancreas (P) and medial to the stomach (S). A right subhepatic fluid collection (F) is present.

can be managed nonoperatively because the mortality of this condition without surgery is less than 5%.273,274 Infected necrosis (as documented by fine-needle aspiration of the pancreas with Gram stain and cultures), on the other hand, has been historically regarded as an indication for surgical débridement because of the previous belief that infected necrosis treated medically has a nearly uniform fatal outcome.275,276 This has led to the recommendation that patients who are not improving on maximal medical therapy or who develop organ failure should have a fine-needle aspiration of the pancreas (Fig. 58-6). The finding of infection should then lead to a consideration of surgical intervention. However, surgical therapy of infected pancreatic necrosis carries a substantial mortality of 15% to 73%,275-278 especially when it is carried out early within the first few weeks of the disease. Thus, it is unclear if the higher mortality rates that have been reported in patients with infected necrosis are the result of the underlying disease or the early surgical intervention. It has been shown that delaying surgical débridement beyond the fourth week in patients with pancreatic necrosis is associated with a lower mortality rate.277 The concept that infected pancreatic necrosis requires prompt surgical débridement has been challenged by several reports of patients who have been treated by antibiotic therapy alone.278-280 A series of 28 patients with infected pancreatic necrosis was treated prospectively with antibiotics rather than urgent surgical débridement.281 Among 12 patients who eventually required elective surgical intervention, there were two deaths. Among the other 16 patients who were treated with long-term antibiotic therapy, there were also two deaths. Thus, 14 of 28 patients with infected necrosis were successfully treated with no surgical, endoscopic, or radiologic drainage. The types of necrosectomy operations that have been recommended include necrosectomy with closed continuous irrigation via indwelling catheters, necrosectomy with

979

980

Section VII  Pancreas closed drainage without irrigation, or necrosectomy and open packing. There have been no randomized prospective trials comparing these various procedures. All are generally considered to provide equal benefit in skilled surgical centers. Several additional procedures have been introduced that are less invasive than standard open surgical débridement of infected necrosis. These techniques have generally been reserved for patients with infected pancreatic necrosis who are too ill to undergo prompt surgical débridement (such as those with organ failure or serious comorbid disease) and include laparoscopic necrosectomy with placement of largecaliber drains under direct surgical inspection and percutaneous catheter drainage of infected necrosis (see Chapter 26). Other minimally invasive approaches, including endoscopic (see Chapter 61) or a combined endoscopic and laparoscopic approach may become more commonly used as skill and technology advance.

OTHER THERAPEUTIC AGENTS OF POSSIBLE OR QUESTIONABLE EFFICACY

Pancreatic protease inhibitors have been used to treat established severe acute pancreatitis and to prevent post-ERCP pancreatitis (see earlier section). Gabexate mesylate is the most widely studied pancreatic protease inhibitor. A metaanalysis of five clinical trials of gabexate mesylate in acute pancreatitis found no effect on the 90-day mortality rate, but a reduced incidence of complications.282 This agent is not available in the United States. Multiple trials of the antisecretory hormone somatostatin or its synthetic analog octreotide have failed to show convincing evidence of efficacy in the treatment of acute pancreatitis.283 The use of antiinflammatory cytokines has so far not shown efficacy. The largest experience has been with lexipafant, a PAF inhibitor. After initial promising reports,284 subsequent studies have not shown clear efficacy.285 Japanese investigators have suggested that pancreatic protease inhibitors and antibiotics can be better targeted to the affected regions in the pancreas with continuous regional arterial infusion (CRAI) into the celiac, splenic, inferior pancreaticoduodenal, and common hepatic arteries. Using CT, Anai and colleagues286 showed that with CRAI the contrast was distributed to the entire pancreas in six of nine patients with inflammation of the entire pancreas; in the remaining three patients, contrast material did not penetrate the entire area of pancreatic inflammation. Two later studies suggested that intra-arterial infusion of the protease inhibitor nafamostat mesylate plus imipenem reduces mortality when compared with intravenous infusion of the same agents.287,288 These studies warrant further investigation. Many other measures have been shown to be ineffective in randomized trials, including anticholinergics, glucagon, fresh frozen plasma, and peritoneal lavage.60,289 Infectious complications and associated mortality are major concerns in acute pancreatitis. Enteral administration of probiotics might prevent infectious complications, but convincing evidence is scarce. Although an early study suggested that probiotics might decrease infectious complications,290 other studies have suggested an increase in morbidity and mortality.291 In this more recent study, infectious complications occurred in 46 patients in the probiotics group (30%) and in 41 of those in the placebo group (28%). Twenty four patients in the probiotics group died (16%), compared with 9 (6%) in the placebo group (relative risk of death, 2.53; 95% confidence interval, 1.22 to 5.25). Nine patients in the probiotics group developed bowel ischemia (8 with fatal outcome), compared with none in the placebo group (P =

0.004). At the present time, probiotic prophylaxis should therefore not be administered in this category of patients.292

LOCAL COMPLICATIONS (Table 58-8) PSEUDOCYST

A pseudocyst may occur secondary to acute pancreatitis, pancreatic trauma, or chronic pancreatitis. It usually contains a high concentration of pancreatic enzymes and variable amounts of tissue debris. Most are sterile. Regardless of size, an asymptomatic pseudocyst does not require treatment.293,294 It is satisfactory to monitor the pseudocyst with abdominal ultrasonography. Pseudocysts can be complicated by infection, intracystic hemorrhage (Fig. 58-7), or rupture leading to pancreatic ascites. Further, pseudocysts can migrate into the chest or other unusual locations. In patients with known pseudocysts, new symptoms, such as abdominal pain, chills, or fever, should alert the clinician to the emergence of an infected pseudocyst or abscess. Treatment choices include surgical, radiologic, and endoscopic drainage. No randomized prospective trials have compared these methods. Surgical drainage of a pseudocyst is possible with a cystgastrostomy or cyst-duodenostomy if the pseudocyst wall is broadly adherent to the stomach or duodenum, respectively. Other procedures include a Roux-en-Y cyst-jejunostomy or pancreatic resection if the pseudocyst is in the tail. Surgical mortality is 6% or less.293-295 Pseudocysts recur after internal (surgical) drainage in 15% of cases; recurrence is more fre-

Table 58-8  Complications of Acute Pancreatitis Local Pseudocyst Sterile necrosis Infected necrosis Abscess Gastrointestinal bleeding Pancreatitis-related Splenic artery or splenic artery pseudoaneurysm rupture Splenic vein rupture Portal vein rupture Splenic vein thrombosis leading to gastroesophageal variceal bleeding Pseudocyst or abscess hemorrhage Postnecrosectomy bleeding Nonpancreatitis-related Mallory-Weiss tear Alcoholic gastropathy Stress-related mucosal gastropathy Splenic complications Infarction Rupture Hematoma Fistulization to or obstruction of small or large bowel Right-sided hydronephrosis Systemic Respiratory failure Renal failure Shock Hyperglycemia Hypocalcemia Disseminated intravascular coagulation Fat necrosis (subcutaneous nodules) Retinopathy Psychosis

Chapter 58  Acute Pancreatitis S

P

Figure 58-7.  Pancreatic pseudocyst with acute hemorrhage. Contrastenhanced computed tomography demonstrates a 10-cm pancreatic pseudocyst (P) containing high-density (45 Hounsfield units) material (arrows) representing acute blood. The pseudocyst is compressing the stomach (S). These findings were confirmed at surgery.

quent if the main pancreatic duct is obstructed downstream from the surgical anastomosis. For this reason, a preoperative ERCP is usually done to determine whether there is duct obstruction. If this is the case, a resection of the pseudocyst is preferred, is possible. Percutaneous catheter drainage is effective treatment to drain and close sterile as well as infected pseudocysts.146 As with surgical drainage, percutaneous catheter drainage may fail if there is obstruction of the main pancreatic duct downstream from the pseudocyst. Therefore, an ERCP is usually done before attempting catheter drainage (see Chapter 26). As discussed in Chapter 61, there are two endoscopic methods to decompress a pancreatic pseudocyst: (1) insertion of a stent through the ampulla directly into the pancreatic duct and then into the pseudocyst itself,296,297 or (2) an endoscopic cyst-gastrostomy or cyst-duodenostomy. With either method, the catheter is removed after three to four weeks if closure of the pseudocyst is seen by CT scan. Failure of percutaneous or endoscopic drainage of a pancreatic pseudocyst increases morbidity and prolongs hospitalization. However, most series show long-term resolution with successful endoscopic drainage of pseudocysts.298 There are several complications of endoscopic drainage of pseudocysts (see Chapter 61). The most important is bleeding; the risk of bleeding may be reduced if endoscopic ultrasonography is used to be certain that there are no large vessels in the drainage area. Infection may occur if the double-pigtail catheter becomes occluded. A nasocystic drain to irrigate the cyst may prevent this complication. An endoscopically placed stent in the pancreatic duct may induce ductal changes identical to those of chronic pancreatitis. For this reason a stent should be removed after several weeks. If a pseudocyst accompanies considerable pancreatic necrosis, endoscopic and percutaneous catheter drainage should be used very cautiously because neither technique can evacuate the underlying particulate necrotic material, although both are successful in eliminating the fluid of the pseudocyst itself. In this situation, surgical drainage may be preferred because necrotic debris can be retrieved before completing the cyst-enteric anastomosis.

NECROTIZING PANCREATITIS AND ABSCESS

Necrotizing pancreatitis is defined, in the absence of laparotomy or autopsy, by the presence of greater than 30% of nonenhancement of the pancreas on a contrast-enhanced CT scan. MRI is also accurate in establishing the diagnosis. The determination that a patient has pancreatic necrosis has clinical implications because the morbidity and mortality of necrotizing pancreatitis are higher than that of interstitial pancreatitis. Furthermore, treatment of patients with necrotizing pancreatitis differs from patients with interstitial pancreatitis. Although there is some controversy, the extent of necrosis may not be as important in determining the morbidity and mortality of necrotizing pancreatitis as was once thought. Patients with necrotizing pancreatitis may appear ill with organ failure or may appear well with no evidence of organ failure. Pancreatic necrosis may either be sterile or infected. For the first 10 to 14 days, pancreatic necrosis is typically considered sterile and managed conservatively. Often patients appear ill and require ICU care. The goal during this time is to maximize supportive care and prevent infection by providing enteral feeding and minimizing potential sources of infection, such as intravenous lines. At this time, the use of antibiotics to prevent infection has been shown to be questionable and not recommended (see earlier). Surgical débridement of sterile pancreatic necrosis has also been shown not be helpful in the vast majority of patients.299 Early surgical débridement is exceedingly difficult and avoided early within the first 4 to 8 days due to the cement-like nature of the necrosis.300 However, surgical intervention is often needed in patients with infected necrosis. Infection of necrosis typically occurs after the tenth day of hospitalization. Infection of the pancreatic necrosis is believed to occur from translocation of bacteria from the colon. This may help explain why enteral feeding, decreasing the pathogenic intestinal flora, prevents infection of necrosis. Infection of the pancreatic necrosis should be suspected when symptoms recur, especially signs of sepsis, with fever, pain, and leukocytosis. When infection is suspected, the diagnosis is readily established by CT-guided fine-needle aspiration (CT-FNA). This procedure is safe and effective in establishing the diagnosis.244 The Gram stain alone has a sensitivity of almost 95% if carefully examined in a fresh specimen. The procedure is also safe, rarely introducing infection into a sterile field in the abdomen. If negative, an aspiration can be repeated every four to seven days if infection continues to be suspected. In the past, the diagnosis of infected necrosis implied the urgent need for surgical débridement. This is no longer true. In a persistently ill patient with sepsis or organ failure found to have infected necrosis, surgical débridement should be strongly considered. However, in a stable patient with infected necrosis, maximal supportive care and the use of pancreatic-penetrating antibiotics should be provided. It is in these patients that antibiotics such as fluoroquinolones, metronidazole, and imipenem should be administered. Although the antibiotics will not likely clear the infection in most patients, the ability to allow time for the formation of a fibrous wall, creating WOPN (see Fig. 58-7), will lead to a more minimally invasive approach to draining the pancreatic necrosis. Whereas early débridement of pancreatic necrosis within the first four to five weeks of an attack will require surgery, WOPN can be treated lapro­ scopically,301 percutaneously, or endoscopically (see Chapter 61).302 The timing and method of débridement require a clear discussion between the surgeon and gastro-

981

982

Section VII  Pancreas enterologist, but should be left at the discretion of the pancreatic surgeon.

mentioned, mortality is substantially increased with increasing organ dysfunction especially if shock is involved.226

GASTROINTESTINAL BLEEDING

METABOLIC DISTURBANCES

GI bleeding may arise from lesions not related to the local inflammatory aspects of pancreatitis such as stress-induced mucosal gastropathy, Mallory-Weiss tear, or alcoholic gastropathy. Alternatively, bleeding can be due to the inflammatory aspects of the pancreatitis (see Table 58-8). The latter is thought to occur from the irritative effects of liberated activated enzymes on vascular structures or pressure necrosis of inflammatory debris or fluid collections on surrounding structures. Rupture of the splenic artery, splenic vein, or portal vein has been reported.303 High mortality is reported with these complications. Temporizing treatments with interventional radiologic techniques are used, followed by more definitive surgical ligation and resection. Acute and chronic inflammatory processes of the pancreas can lead to thrombosis of the adjacent splenic vein, which can lead to gastric varices, with or without esophageal varices. These varices can rupture, leading to massive bleeding. Treatment of this problem can be endoscopic with banding of varices or splenectomy, which is curative. Pseudocysts can be complicated by pseudoaneurysm formation, which can usually be seen by dynamic contrast-enhanced CT. If these bleed, arteriography with embolization is the treatment of choice. Rarely, bleeding into the pancreatic duct occurs (hemosuccus pancreaticus), but this typically occurs in chronic pancreatitis. Postnecrosectomy bleeding is common and can be caused by overly aggressive débridement or the placement or the use of noncompliant drainage tubes next to vascular structures.

SPLENIC COMPLICATIONS

Splenic complications of pancreatitis include intrasplenic pseudocysts, infarction and necrosis of the spleen, splenic rupture, and hematoma.304 Some of these complications can be life threatening and require emergency splenectomy (see Table 58-8).

BOWEL COMPRESSION OR FISTULIZATION

Pressure necrosis from inflammatory debris from the tail of the pancreas can obstruct or fistulize into the small or large bowel. The most common site is the left colon.305 Treatment is frequently surgical.

SYSTEMIC COMPLICATIONS ORGAN DYSFUNCTION

Organ dysfunction has been discussed extensively earlier in the chapter. Respiratory insufficiency is the most common systemic complication associated with pancreatitis. The causes are multifactorial and include pleural effusions, pneumonia, atelectasis, and ARDS. Oxygen supplementation, antibiotics, thoracentesis, and assisted ventilation may be necessary. Renal complications are due to hypovolemia causing prerenal azotemia or hypotension, leading to acute tubular necrosis. These are treated by increasing intravenous fluid administration in the former case or with hemofiltration or hemodialysis in the latter situation. Shock is usually caused by hypovolemia secondary to third-space losses, vomiting, and interstitial pancreatic edema. Other uncommon causes of hypotension include myocardial infarction and pericardial effusions. Fluid replacement in severe acute pancreatitis is best accomplished using central venous– or Swan-Ganz catheter–assisted monitoring. As

Hyperglycemia and hypocalcemia are common in severe disease. Hyperglycemia is usually transient and is due to insulin deficiency from presumed islet cell necrosis or hyperglucagonemia. It is uncommon for these complications to require aggressive treatment.

COAGULATION DISORDERS

Mild coagulation defects are common in acute pancreatitis as measured by elevated d-dimer levels in the blood.306 Fullblown disseminated intravascular coagulation (DIC) with a bleeding diathesis associated with a hypercoagulable state is very uncommon.

FAT NECROSIS

Fat necrosis occurs in subcutaneous tissue, bone, retroperitoneal tissue, peritoneum, mediastinum, pleura, and pericardium.191 Histologically fat cells are necrotic, associated with a diffuse inflammatory infiltration. The subcutaneous lesions are circumscribed tender red nodules that are adherent to the skin but are movable over deeper structures. Most commonly they are over the ankles, fingers, knees, and elbows. The lesions may drain through the skin. Rarely there is also necrosis of adjacent tendons or involvement of joints, particularly the metatarsal, interphalangeal, wrist, knee, and ankle joints. The lesions usually resolve after days to weeks.

MISCELLANEOUS

Pancreatic encephalopathy consists of a variety of central nervous system symptoms occurring in acute pancreatitis, including agitation, hallucinations, confusion, disorientation, and coma.198 A similar syndrome may be due to alcohol withdrawal, and other causes are possible, such as electrolyte disturbances (e.g., hyponatremia) or hypoxia. Purtsher’s retinopathy (discrete flame-shaped hemorrhages with cotton wool spots) can cause sudden blindness.307 It is thought to be due to microembolization in the choroidal and retinal arteries.

KEY REFERENCES

Badalov N, Baradarian R, Iswara K, et al. Drug induced acute pancreatitis: An evidence based approach. Clin Gastroenterol Hepatol 2007; 101:454-76. (Ref 89.) Besselink MG, Verwer TJ, Schoenmaeckers EJ, et al. Timing of surgical intervention in necrotizing pancreatitis. Arch Surg 2007; 142:1194201. (Ref 277.) Freeman ML, DiSario JA, Nelson DB, et al. Risk factors for post-ERCP pancreatitis: A prospective, multicenter study. Gastrointest Endosc 2001; 54:425. (Ref 109.) Gardner TB, Vege SS, Chari ST, et al. Fluid resuscitation in acute pancreatitis. Clin Gastroenterol Hepatol 2008; 6:1070-6. (Ref 240.) Gerzof SG, Banks PA, Robbins AH, et al. Early diagnosis of pancreatic infection by computed tomography-guided aspiration. Gastroenterology 1987; 93:1315. (Ref 242.) Isenmann R, Runzi M, Kron M, et al. Prophylactic antibiotic treatment in patients with predicted severe acute pancreatitis: A placebocontrolled, double blind trial. Gastroenterology 2004; 126:997. (Ref 254.) Kalfarentzos F, Kehagias J, Mead N, et al. Enteral nutrition is superior to parenteral nutrition in severe acute pancreatitis: Results of a randomized prospective trial. Br J Surg 1997; 84:1665. (Ref 269.) Mier J, Luque-deLeon E, Castillo A, et al. Early vs. late necrosectomy in severe necrotizing pancreatitis. Am J Surg 1997; 173:71. (Ref 272.) Neoptolemos JP, Carr-Locke DL, London NJ, et al. Controlled trial of urgent endoscopic retrograde cholangiopancreatography and endoscopic sphincterotomy versus conservative treatment for acute pancreatitis due to gallstones. Lancet 1988; 2:979. (Ref 257.)

Chapter 58  Acute Pancreatitis Opie EL. The etiology of acute hemorrhagic pancreatitis. Bull Johns Hopkins Hosp 1901; 12:182. (Ref 40.) Ranson JHC, Rifkind RM, Roses DF. Prognostic signs and the role of operative management in acute pancreatitis. Surg Gynecol Obstet 1975; 139:69. (Ref 14.) Runzi M, Niebel W, Goebell H, et al. Severe acute pancreatitis: Nonsurgical treatment of infected necrosis. Pancreas 2005; 30:195-9. (Ref 281.) Singh V, Wu BU, Maurer R, et al. A prospective evaluation of the Bedside Index of Severity in Acute Pancreatitis. Am J Gastroenterol 2009; 104:966-71. (Ref 221b.)

Tenner S, Sica G, Hughes M, et al. Relationship of necrosis to organ failure in severe acute pancreatitis. Gastroenterology 1997; 113:899903. (Ref 17.) Wu BU, Johannes RS, Sun S, et al. Early changes in blood urea nitrogen predict mortality in acute pancreatitis. Gastroenterology 2009; 137:129-35. (Ref 221c.) Full references for this chapter can be found on www.expertconsult.com.

983

CHAPTE R

59 Chronic Pancreatitis Chris E. Forsmark

CHAPTER OUTLINE Epidemiology  985 Pathology  986 Pathophysiology  987 Etiology  988 Alcohol  988 Tobacco  989 Tropical Pancreatitis  989 Genetic  989 Autoimmune Pancreatitis  990 Obstructive Chronic Pancreatitis  992 Miscellaneous  994 Idiopathic Chronic Pancreatitis  994 Clinical Features  995 Abdominal Pain  995 Steatorrhea  996 Diabetes Mellitus  997 Physical Examination  997

The traditional definition of chronic pancreatitis has been permanent and irreversible damage to the pancreas, with histologic evidence of chronic inflammation, fibrosis, and destruction of exocrine (acinar cell) and endocrine (islets of Langerhans) tissue (Fig. 59-1). As a consequence of this histologic damage, varying degrees of permanent and irreversible clinical, morphologic, and functional derangements occur. Chronic pancreatitis was distinguished from acute pancreatitis, characterized by complete recovery of the pancreas after the acute episode. This definition, using histologic criteria, is imperfect for a number of reasons. Most important, pancreatic tissue is rarely available to clinicians. Patients may have histologic evidence of chronic pancreatitis and have no symptoms or complications from the disease. In addition, the histologic features of chronic pancreatitis are often focal, such that a small biopsy, even if available, might miss the disease. Finally, the individual histologic features that are consistent with chronic pancreatitis are not unique, and may be seen in other conditions (such as normal aging). Alternatively, chronic pancreatitis has been defined based on clinical features (pain and exocrine or endocrine insufficiency) or on imaging techniques including ultrasonography (US), computed tomography (CT), endoscopic ultrasonography (EUS), magnetic resonance imaging (MRI), and endoscopic retrograde cholangiopancreatography (ERCP). Defining chronic pancreatitis on the basis of imaging studies is also imperfect because the morphologic changes detected by these modalities may take years to develop. Indeed, many of the findings of these imaging studies can be normal early in the clinical course. Early diagnosis of disease, at a time when some effective therapy might be administered, is therefore often impossible. Diagnostic cri-

Diagnosis  998 Tests of Pancreatic Function  999 Tests of Pancreatic Structure (Imaging)  1000 Diagnostic Strategy  1003 Treatment  1004 Abdominal Pain  1004 Maldigestion and Steatorrhea  1009 Diabetes Mellitus  1010 Complications  1010 Pseudocyst  1010 Gastrointestinal Bleeding  1011 Bile Duct Obstruction  1013 Duodenal Obstruction  1014 Pancreatic Fistula  1014 Malignancy  1014 Dysmotility  1014

teria that rely on imaging findings are therefore a mixture of diagnostic and staging criteria, determining presence and severity of disease. These staging systems also tend to lump together all etiologies, thereby obscuring differences that are important to clinicians. Finally, these staging systems make the assumption that acute pancreatitis and chronic pancreatitis are entirely separate entities, when in fact there is now abundant evidence documenting the evolution in many patients from acute to chronic pancreatitis (see Chapter 58). It is probably most accurate to think of the two conditions as separate ends of the same spectrum; acute pancreatitis is an identifiable event, whereas chronic pancreatitis is an ongoing process of variable tempo. More recent reviews recognize these shortfalls and try to emphasize the importance of etiology, the difficulty of obtaining pancreatic tissue, and the lack of sensitivity of currently available diagnostic tools.1 These proposals have some advantages in that they attempt to define the disease in a more clinically useful manner, classify disease on the basis of etiology, and include advances in technology (e.g., EUS, magnetic resonance cholangiopancreatography [MRCP], genetic analysis) as well as functional and structural damage to the pancreas in the staging criteria.

EPIDEMIOLOGY Chronic pancreatitis can be demonstrated in up to 5% of autopsies.2,3 Similar, though less pronounced, histologic features are seen even more commonly.4,5 Determining the prevalence of chronic pancreatitis from this type of autopsy

985

986

Section VII  Pancreas

Figure 59-1.  Histology of chronic pancreatitis. Note the destruction of acinar tissue with replacement by extensive fibrosis and relative sparing of pancreatic islets. (Hematoxylin and eosin.)

data is misleading because the patients from whom the data are taken may not have had clinical symptoms of chronic pancreatitis during life. Long-standing alcohol use, even in moderate amounts, can lead to histologic changes of chronic pancreatitis without symptoms of chronic pancreatitis.6,7 Similarly, aging, chronic kidney disease, and long-standing diabetes mellitus can induce histologic changes within the pancreas that are difficult to distinguish from those of chronic pancreatitis.4,5 Making a diagnosis solely on the basis of autopsy data will therefore overestimate the rate of clinically important (i.e., symptomatic) chronic pancreatitis. Estimates of annual incidence of chronic pancreatitis in several retrospective studies range from 3 to 9 cases per 100,000 population.8,9 The only prospective study, which was largely limited to alcoholic chronic pancreatitis, noted an annual incidence of 8.2 cases per 100,000 population and an overall prevalence of 27.4 cases per 100,000.10 A nationwide cross-sectional survey in Japan revealed an overall prevalence rate of 35.5 cases per 100,000 population with an estimated incidence of 14.4 per 100,000 and a maleto-female ratio of 3.5 : 1.11 A similar survey in France noted an annual estimated incidence of 7.7 per 100,000 population and a prevalence of 26 per 100,000.12 In most studies, alcohol abuse accounts for two thirds of all cases of chronic pancreatitis. These limited epidemiologic data demonstrate substantial geographic variation.13 The variation may partly be due to differences in alcohol consumption in different populations, but another part of the variation in incidence rates may merely reflect different diagnostic approaches and different diagnostic criteria. Chronic pancreatitis accounts for substantial morbidity and health care costs. Approximately 26,000 hospital admissions to non-federal hospitals with a first-listed diagnosis of chronic pancreatitis occur yearly; in more than 80,000 yearly admissions, chronic pancreatitis is listed as one of the discharge diagnoses.14,15 The prognosis of chronic pancreatitis is variable and is driven largely by the presence of ongoing alcoholism in persons with chronic alcoholic pancreatitis and equally by concomitant tobacco use. One can estimate prognosis from such features as need for medical care or hospitalization or from the development of complications, reduced quality of life, or mortality.

Data on the quality of life of patients with chronic pancreatitis16-19 document that the presence of abdominal pain and alcohol abuse (in those with alcoholic chronic pancreatitis) are the dominant negative influences on quality of life and that, not surprisingly, quality of life is substantially worse for such patients than for the general population. Mortality in patients with chronic pancreatitis is also substantially influenced by the presence of continued alcoholism. In one large multicenter study, the standardized mortality ratio was 3.6:1 (i.e., those with a diagnosis of any form of chronic pancreatitis died at 3.6 times the rate of age-matched controls), and older subjects, those who smoked, and those with alcoholic chronic pancreatitis had the most significant reduction in survival.20 Continuing alcohol use raised mortality risk by an additional 60%. Overall, 10-year survival in patients with chronic pancreatitis is about 70%, and 20-year survival about 45%. Similar mortality data were noted in an analysis from Japan.21 The cause of death in patients with chronic pancreatitis usually is not the pancreatitis itself but other medical conditions commonly associated with smoking, continued alcohol abuse, pancreatic carcinoma, and postoperative complications.

PATHOLOGY The different etiologies of chronic pancreatitis usually produce similar pathologic findings (see Fig. 59-1), particularly as the disease progresses. In early chronic pancreatitis the damage is varying and uneven. Areas of interlobular fibrosis are seen, with the fibrosis often extending to the ductal structures. Infiltration of the fibrotic area and lobules with lymphocytes, plasma cells, and macrophages is seen.22 The ducts may contain eosinophilic protein plugs. In affected lobules, acinar cells are surrounded and replaced by fibrosis. The islets are usually less severely damaged until very late in the course of the disease. Features of acute pancreatitis also may be seen, such as edema, acute inflammation, and acinar cell or fat necrosis. As the disease progresses, fibrosis within the lobules and between lobules becomes more widespread. The pancreatic ducts become more abnormal with progressive fibrosis, stricture formation, and dilation. The ductal protein plugs may calcify and obstruct major pancreatic ducts. Ductal epithelium may become cuboidal, may develop atrophy or squamous metaplasia, or may be replaced by fibrosis entirely. If special stains are performed, activated pancreatic stellate cells may be identified in close association with fibrosis. These histologic features are found in most forms of chronic pancreatitis. Many of these changes, in particular perilobular fibrosis and ductal metaplasia, are also commonly seen in patients of advanced age without chronic pancreatitis, and in patients with long-standing diabetes mellitus.4,5 Obstructive chronic pancreatitis (associated with obstruction of the main pancreatic duct by a tumor or stricture) can differ slightly, in that the histologic changes are limited to the gland upstream of the obstruction and protein precipitates and intraductal stones are not seen.23 Autoimmune chronic pancreatitis demonstrates a more robust lymphoplasmacytic infiltrate, including plasma cells, and these are usually positive when stained for immunoglobulin G subtype 4 (IgG4).24 Obstructive phlebitis affecting the major and minor veins and a whorled fibrosis pattern are also characteristic, a pattern termed lymphoplasmacytic sclerosing pancreatitis (LPSP). A variant pattern termed idiopathic duct-centric chronic pancreatitis is characterized

Chapter 59  Chronic Pancreatitis by additional neutrophilic infiltration. With time, the pattern may assume a more end-stage chronic pancreatitis appearance and become indistinguishable from other forms of chronic pancreatitis.

PATHOPHYSIOLOGY The pathophysiology of chronic pancreatitis remains incompletely understood. The pathophysiologic processes must ultimately account for the features of chronic pancreatitis, including loss of parenchymal cells, self-sustaining chronic inflammation, and fibrosis. Any proposed mechanism must therefore include explanations for cellular necrosis or apoptosis, initiation and maintenance of inflammatory cell activation, and fibrogenesis by pancreatic stellate cells. The pancreas, like all other organs, has a limited repertoire of responses to injury and although it is not clear that all of the various etiologies have identical pathophysiology, the end histologic result is similar. The study of mechanisms of disease is hampered by the difficulty of obtaining tissue in humans and the relative lack of animal models of chronic pancreatitis,25,26 as opposed to acute pancreatitis. Alcoholic chronic pancreatitis, being the most common form, has received the most attention. No single theory explains adequately why only about 10% of heavy alcohol users develop chronic pancreatitis. Alcohol is metabolized by the liver and the pancreas. In the liver the main end product of oxidative alcohol metabolism is acetaldehyde. In the pancreas, an alternative pathway produces fatty acid ethanol esters (FAEEs). Alcohol and its metabolites like FAEE, have direct injurious effects on pancreatic acinar cells. Increased membrane lipid peroxidation, a marker of oxidative stress and free radical production, can be seen in animal models and human alcoholic chronic pancreatitis.27-29 In addition FAEEs are able to induce sustained elevations in cytosolic calcium in acinar cells,29,30 a mechanism shared by other experimental causes of pancreatitis. Alcohol may also lead to pathologic increases in acinar cell sensitivity to physiologic stimuli such as cholecystokinin (CCK)31 and redirect CCK-mediated zymogen exocytosis to the basolateral rather than apical membrane.32 This basolateral secretion would place digestive enzymes in the interstitial space, where they could produce tissue damage. Chronic alcohol ingestion in animal models also alters expression of multiple genes in acinar cells, which could increase the sensitivity to physiologic stress33 and up-regulate the expression and activity of enzymes involved in cell death.34 Alcohol can promote the inflammatory responses involved in pancreatitis.35 These multiple effects of alcohol on the acinar cell are complemented by alcohol injury to ductal cells.36 Finally alcohol and its metabolites appear to stimulate the pancreatic stellate cell.36-38 These cells, as in the liver, appear to be the final common pathway for fibrosis. Pancreatic stellate cells are found in association with the acinar units. They are typically found in the periacinar space, with long cytoplasmic processes extending to the acini themselves, but are also present in smaller numbers in association with blood vessels and ducts. Quiescent pancreatic stellate cells are recognized by the presence of vitamin A lipid droplets in the cytoplasm. When activated, they assume a stellate or myofibroblastic appearance, express smooth muscle actin, and lose the lipid droplets. This activation is necessary for the cell to begin to secrete extracellular matrix and produce fibrosis within the gland. Activation of pancreatic stellate cells can occur by alcohol or one of its metabolites, but also occurs in response to both

inflammatory cytokines that are released following pancreatic acinar cell necrosis and to reactive oxygen species.36-38 In addition, growth factors (platelet-derived growth factor, transforming growth factor-β1), hormones, intracellular signaling molecules, transcription factors, and angiotensin II can activate pancreatic stellate cells.38 Activated pancreatic stellate cells are found in areas of extensive necrosis and inflammation in acute pancreatitis, in human as well as animal tissues. These activated pancreatic stellate cells produce autocrine factors that maintain the activated phenotype. In addition to their role in secretion and modulation of the extracellular matrix, pancreatic stellate cells can proliferate in response to stimulation, migrate to areas of inflammation, and participate in phagocytosis. Activation of pancreatic stellate cells is likely occurring via at least two mechanisms in alcoholic chronic pancreatitis: directly by alcohol and its metabolites and indirectly by cytokines induced by cellular necrosis.36 Chronic alcohol ingestion may produce chronic pancreatitis by additional mechanisms. Longtime alcohol use leads to the secretion of a pancreatic juice rich in protein and low in volume and bicarbonate. These characteristics favor the formation of protein precipitates, which are present early in the evolution of alcoholic chronic pancreatitis. These precipitates may calcify, leading to the formation of pancreatic ductal stones and producing further ductal and parenchymal injury upstream from these stones. In most patients, however, these protein precipitates and ductal stones do not appear to cause pancreatic injury but instead seem merely to be markers for the underlying pathophysiologic events. There have been several hypotheses for the pathophysiology of chronic pancreatitis that attempt to interweave these concepts into a coherent paradigm. One hypothesis focuses on the concept that ductal obstruction (from strictures or stones) is the cause rather than the effect of chronic pancreatitis. This hypothesis, the ductal obstruction hypothesis, is not consistent with most clinical and experimental evidence and with few exceptions (such as the rare condition of obstructive chronic pancreatitis) is not applicable to human chronic pancreatitis. A second paradigm, the toxicmetabolic hypothesis, focuses primarily on the role of alcohol and its metabolites (or other toxins) and their ability to damage the pancreas and activate pancreatic stellate cells. A third model that has been proposed is the necrosisfibrosis hypothesis, which holds that the occurrence of repeated episodes of acute pancreatitis with cellular necrosis or apoptosis eventually leads to the development of chronic pancreatitis as the healing process replaces necrotic tissue with fibrosis. This last hypothesis has significant supporting evidence from some natural history studies that document the more common development of chronic pancreatitis in patients with more severe and more frequent acute attacks of alcoholic pancreatitis.39,40 The concept that multiple clinical or subclinical attacks of acute pancreatitis lead to chronic pancreatitis is certainly being reinforced by observations in both animal models and in humans.25,29,35,36,41 It is not clear why only a small subset of chronic alcoholics develop chronic pancreatitis. Some of the many possible explanations might include the presence of important cotoxins, differences in the genetic or epigenetic background, or differences in the microenvironment in the pancreas. For example, tobacco use is one very important cofactor for the development of alcoholic (and other types) chronic pancreatitis.42,43 There are also unexplained racial differences in the rates of alcoholic chronic pancreatitis.44 Multiple mutations have been identified in several forms of chronic pancreatitis, suggesting a complex genetic background that provides the relative predisposition to develop chronic

987

988

Section VII  Pancreas pancreatitis (see Chapter 57). Mutations in the cystic fibrosis transmembrane conductance regulator (CFTR), cationic trypsinogen gene (PRSS1 gene), serine protease inhibitor Kazal type 1 (SPINK1, a trypsin inhibitor, formerly called PSTI or pancreatic secretory trypsin inhibitor), and many others have been identified.36,41,45-48 These mutations and the many more yet to be identified may be adequate in and of themselves to produce pancreatitis (at this point only PRSS1 satisfies that requirement) or may only predispose to disease. In the case of PRSS1, a number of families with hereditary pancreatitis (see Chapter 57) have been identified with mutations in the PRSS1 gene. These mutations appear to lead to a gain of function mutation in which trypsinogen, once activated to trypsin, is resistant to inactivation. The mutant trypsin might therefore be able to activate other proenzymes and to produce clinical or subclinical episodes of acute pancreatitis, ultimately leading to chronic pancreatitis in affected patients. This presumed pathophysiology provides support for the necrosis-fibrosis theory of chronic pancreatitis, in which repeated episodes of acute pancreatitis ultimately lead to chronic pancreatitis. Cystic fibrosis is associated with abnormalities of bicarbonate secretion by ductal cells and the eventual development of dilated pancreatic ducts, intraductal precipitates, and pancreatic atrophy (see Chapter 57). Several studies have documented an increase in rate of CFTR mutations in patients with presumed “idiopathic” pancreatitis.49-52 SPINK1 mutations have been found in idiopathic chronic pancreatitis, hereditary pancreatitis, tropical pancreatitis, and alcoholic pancreatitis.45,46,53-55 These mutations are discussed in detail in Chapter 57 but are useful to consider here as a basis for an evolving paradigm of pathogenesis This paradigm has been labeled SAPE—sentinal acute pancreatitis event—and conceptualizes a complex genetic background that provides the predisposition to disease.1,29,36,37,41,46,47 This background may include mutations in genes that code for digestive enzymes, protease-enzyme inhibitors, ion channels, and a variety of others including genes that affect the metabolism of environmental toxins (e.g., tobacco or alcohol), genes that have a role in inflammation or fibrosis, and others yet to be discovered. Although a few mutations may be sufficient to produce pancreatic damage in most or all who carry it, the majority of mutations identified to date are not that dominant. On this complex polygenic background is overlaid some environmental insult, such as chronic alcohol use, that applies physiologic stress to the acinar, ductal, and stellate cells. This stress may be insufficient to produce injury or may produce cellular necrosis or apoptosis. The initiating event for necrosis is the activation of digestive enzymes within the acinar cell, either by the toxic effect of the environmental insult or because of some underlying mutation that leads to excessive activation of trypsin (see Fig. 57-3). Inflammation follows the necrosis, and this necroinflammatory process may either progress or resolve. In some individuals, the situation may never progress beyond this stage. In others, continued cell metabolic and oxidative stresses or some other trigger could produce continuing or repeated acinar cell injury with necrosis. This process, as is believed to be the case in other organs, could be associated with the activation of stellate cells. The acute inflammatory response could resolve, or, alternatively, the inflammatory response might switch to an anti-inflammatory response, with resident macrophages producing cytokines and activated stellate cells laying down extracellular matrix, with the formation of fibrosis. Fibrosis would start a vicious circle by causing additional acinar cell ischemia and continuing to drive the inflammatory–anti-inflammatory cycle. This type of hypothesis could

theoretically explain many forms of chronic pancreatitis. This framework seems to fit all the developing experimental and clinical data and is a useful way in which to think about the pathophysiology of chronic pancreatitis: as a disease associated with a variety of different genetic predispositions, a variety of disease triggers, and a similar final common pathway producing pancreatic injury and fibrosis, ultimately with organ failure.

ETIOLOGY (Table 59-1) ALCOHOL

In Western countries, alcohol is the cause of at least 70% of all cases of chronic pancreatitis.8-10,14,56-60 The risk of alcoholic chronic pancreatitis increases logarithmically with rising alcohol use, but there is no true threshold value below which the disease does not occur.48,60-63 Therefore, in countries with widespread alcohol consumption, it may be difficult to determine with certainty whether alcohol contributed to or caused the disease. In nearly all patients, at least 5 years (and in most patients more than 10 years) of intake exceeding 150 g/day (a standard drink contains about 14  g of alcohol) are required before the development of Table 59-1  Classification of Chronic Pancreatitis Alcoholic Tropical Tropical calcific pancreatitis Fibrocalculous pancreatic diabetes Genetic Autosomal dominant Hereditary pancreatitis (PRSS1 mutations) Autosomal recessive or modifier genes CFTR mutations SPINK1 mutations Others Metabolic Hypercalcemia Hyperlipidemia Hypertriglyceridemia Lipoprotein lipase deficiency Apolipoprotein C-II deficiency Obstructive Benign pancreatic duct obstruction Traumatic stricture Stricture after severe acute pancreatitis Sphincter of Oddi stenosis Duodenal wall cyst Pancreas divisum Celiac disease Malignant pancreatic duct stricture Ampullary or duodenal carcinoma Pancreatic adenocarcinoma Intraductal papillary mucinous neoplasm Autoimmune Autoimmune pancreatitis Immunoglobulin G subtype 4 (IgG4)–related systemic disease Postnecrotic Idiopathic Early-onset Late-onset Asymptomatic pancreatic fibrosis Chronic alcoholism Old age Chronic kidney disease Diabetes mellitus Radiotherapy CFTR, cystic fibrosis transmembrane conductance regulator gene; PRSS1, protease serine 1 gene; SPINK1, serine protease inhibitor, Kazal type 1 gene.

Chapter 59  Chronic Pancreatitis chronic pancreatitis. Only 3% to 15% of heavy drinkers ultimately develop chronic pancreatitis, suggesting an important cofactor.56,61,62,64 Potential cofactors include genetic polymorphisms and mutations,27-29,47,48,64,65 a diet high in fat and protein,47,61,64 the type of alcohol or manner of ingestion,62 an associated relative deficiency of antioxidants or trace elements,66,67 and smoking.42,47,56,61,64,68,69 Of these, smoking appears to be the strongest association. In some studies, 90% of those who develop alcoholic chronic pancreatitis are also chronic smokers.64,68,69 Smoking also appears to predispose to more rapid development of pancreatic calcification.42,43,62,70,71 There are also racial differences in the risk for development of alcoholic pancreatitis, perhaps suggesting some difference in the ability to detoxify environmental toxins or alcohol or other genetic or epigenetic factors.44,47,62,64 Although the risk of alcoholic chronic pancreatitis is higher in blacks,44 data from self-reported surveys of alcohol use demonstrate that the proportion of blacks who drink alcohol or smoke is actually lower than that in whites.72 Many patients with alcoholic chronic pancreatitis have an early phase of recurrent attacks of acute pancreatitis, which may last five or six years, followed by the later development of chronic pain or exocrine or endocrine insufficiency. It has been believed, based on several large natural history studies, that most patients who present with their first attack of acute alcoholic pancreatitis have already developed histologic chronic pancreatitis. Up to 40% of patients presenting with acute alcoholic pancreatitis, however, do not progress to clinically identifiable chronic pancreatitis (calcification and exocrine or endocrine insufficiency) even with very long follow-up, and may not even have recurrent attacks.39,62,64,73-75 Based on autopsy studies6,7 and functional studies,76 however, it is likely these patients actually have histologic evidence of chronic pancreatitis, although they may not have developed a recognizable chronic clinical illness. Not all patients with alcoholic chronic pancreatitis present with acute episodes of pancreatitis. Less than 10% of patients present with exocrine or endocrine insufficiency in the absence of abdominal pain.57-59,64 Some present with chronic pain in the absence of antecedent acute attacks of pain. Cessation of alcohol use after the onset of alcoholic pancreatitis appears to diminish the rate of progression to exocrine insufficiency and endocrine insufficiency but does not halt it.77 Stopping alcohol does seem to reduce the chance of recurrent attacks of acute alcoholic pancreatitis in those who have not yet developed obvious chronic pancreatitis.75 The prognosis of alcoholic chronic pancreatitis is relatively poor,20,21 and mortality is generally greater than that seen in chronic pancreatitis of other etiologies. Quality of life is also substantially diminished. Pain generally continues for years, although it may spontaneously remit. Exocrine and/or endocrine insufficiency develops in many patients, although this process may take several years. In one large natural history study, exocrine insufficiency developed in 48% of patients at a median of 13.1 years after presentation, whereas endocrine insufficiency developed in 38% after a median of 19.8 years after presentation.58 Diffuse pancreatic calcifications developed in 59% at a median of 8.7 years after diagnosis. Other studies have noted more rapid and more frequent development of calcifications, exocrine insufficiency, and endocrine insufficiency.57

TOBACCO

Exposure to tobacco smoke can induce pancreatic damage in animals.78 As mentioned, smoking is common in patients

with alcoholic chronic pancreatitis and is associated with an increased risk for pancreatic calcifications, and smoking cessation after the clinical onset of chronic pancreatitis reduces the risk of subsequent calcifications.79 There is also evidence that smoking is an independent risk factor for chronic pancreatitis.80,81 Smoking is also associated with a much higher rate of secondary pancreatic cancer and overall mortality in patients with chronic pancreatitis.11,20

TROPICAL PANCREATITIS

Tropical pancreatitis is one of the most common forms of chronic pancreatitis in certain areas of southwest India. It has been reported from a number of other areas, including Africa, southeast Asia, and Brazil. The disease as initially described is essentially restricted to areas within 30 degrees of latitude from the equator. Tropical pancreatitis is generally a disease of youth and early adulthood, with a mean age at onset of 24 years.82-84 More than 90% of patients have the illness before 40 years. The overall prevalence from surveys in an endemic area (southern India) is 1 in 500 to 1 in 800 population.82,85 Tropical pancreatitis accounts for about 70% of all cases of chronic pancreatitis in southern India, whereas alcohol is a more dominant cause in the north. Surveys in Kerala province in southwest India have noted that there is a shift toward older age at presentation, less malnutrition, and less severe diabetes.86 These later surveys also note higher rates of alcohol abuse in this area and alcohol gradually becoming a more common cause of chronic pancreatitis. The disease classically manifests as abdominal pain, severe malnutrition, and exocrine or endocrine insufficiency. One striking feature is the propensity to diabetes, and endocrine insufficiency appears to be an inevitable consequence of tropical chronic pancreatitis (often classified as a specific cause of diabetes called fibrocalculous pancreatic diabetes). Steatorrhea is rare owing to a generally very low-fat intake. Pancreatic calculi develop in more than 90% of patients. The pathology is characterized by large intraductal calculi, marked dilation of the main pancreatic duct, and gland atrophy. The pathophysiology of tropical pancreatitis is unknown. As discussed in Chapter 57, a number of genetic mutations have been identified, with mutations in the SPINK1 and chymotrypsinogen genes being most common.87 Environmental triggers for the disease that have been proposed include protein-calorie malnutrition, deficiencies of trace elements and micronutrients coupled with oxidative stress (via xenobiotics, industrial pollutants, diet, or nutritional deficiency), cyanogenic glycosides present in cassava (tapioca—a main dietary component), viral and parasitic infections, and others.

GENETIC

Only one type of mutation appears sufficient to cause chronic pancreatitis: mutations in PRSS1 in families with hereditary pancreatitis. All other identified mutations (discussed in Chapter 57) and polymorphisms should be considered as cofactors, mutations that predispose to disease by increasing susceptibility, or as modifier genes, that increase the pace or severity of disease. It is certainly possible that mixtures of polymorphisms and mutations work together to determine the susceptibility to disease. The most commonly identified mutations occur in the PRSS1 (cationic trypsinogen), SPINK1 (trypsin inhibitor), and CFTR genes. Several studies have suggested that certain less severe CFTR gene mutations and SPINK1 mutations may be associated with idiopathic chronic pancreatitis. CFTR gene mutations have been identified in up to half of patients with

989

990

Section VII  Pancreas idiopathic chronic pancreatitis.46,49-52 This proportion is far greater than expected within this population. Analysis of these data has suggested that the combination of a more severe CFTR mutation on one chromosome with a mild mutation on the other is particularly associated with chronic pancreatitis (see Chapter 57). These studies examined only the most common CFTR mutations. There are now more than 1200 known CFTR mutations, and further studies are needed to define their contribution to idiopathic and other forms of chronic pancreatitis. SPINK1 mutations have also been identified with increased frequency in patients with unexplained chronic pancreatitis, including some mixed heterozygotes with combined CFTR and SPINK1 mutations. Genetic testing for these mutations is commercially available.

AUTOIMMUNE PANCREATITIS

Autoimmune pancreatitis refers to a distinct chronic inflammatory and sclerosing disease of the pancreas. The distinct characteristic of the disease is a dense infiltration of the pancreas, and often other organs, with lymphocytes and plasma cells (Fig. 59-2A), many of which express IgG4 on their surface. The autoimmune target of this IgG4 and the trigger for disease are unknown. However, recently a protein

expressed in pancreatic acinar cells, UBR2 (ubiquitinprotein ligase E3 component n-recognin 2) has been proposed to be the target of the antibody.87a Of interest, the antibodies also react to a protein of Helicobacter pylori, PBP (plasminogen-binding protein), suggesting a role for H. pylori infection in autoimmune pancreatitis. Fibrosis, sclerosis, and obliterative phlebitis are characteristically seen in association with the chronic inflammatory infiltrate. Although this inflammatory infiltrate is present in the pancreas, similar infiltrates may be seen in the bile duct, salivary glands, retroperitoneum, lymph nodes, kidney, prostate, ampulla, and occasionally other organs.24 An equally characteristic feature of this disease is the often rapid response to glucocorticoid therapy. Much of the information on this disease comes from a series of studies from Japan and other Asian countries. Early reports noted chronic pancreatitis characterized by the presence of autoantibodies, elevated levels of serum immunoglobulins, enlargement of the pancreas (diffuse or focal), pancreatic duct strictures, and pathologic features of a dense lymphocytic infiltrate.88 Studies from Japan note that up to 6% of all patients evaluated for chronic pancreatitis have autoimmune pancreatitis, and the overall prevalence is estimated to be 0.82

A

B

C

D

Figure 59-2.  Autoimmune pancreatitis. A, Histopathology of a pancreatic resection specimen demonstrating a robust lymphoplasmacytic infiltrate involving the larger pancreatic ducts. (Hematoxylin and eosin.) B, Cholangiogram demonstrating a smooth stricture involving the intrapancreatic portion of the bile duct. C, Computed tomography shows a dilated pancreatic duct without pancreatic parenchymal atrophy. D, Pancreatogram reveals a moderately dilated pancreatic duct with diffuse areas of irregularity and alternating areas of stenosis and dilatation. There is an area of more dominant stricture in the pancreatic head. (Courtesy of C. Mel Wilcox, MD, Birmingham, Ala.)

Chapter 59  Chronic Pancreatitis per 100,000 persons.24,88-92 The disease may occur in an isolated form or may be associated with extrapancreatic manifestations. The most common extrapancreatic conditions identified include biliary strictures, hilar lymphadenopathy, sclerosing sialadenitis, retroperitoneal fibrosis, and tubulointerstitial nephritis.24,88-93 Autoimmune pancreatitis may therefore be one manifestation of what has been termed IgG4-related sclerosing disease or IgG4-related systemic disease. There also appears to be some overlap with an unusual variant of Sjögren’s disease termed Mikulicz’s disease, in which a massive IgG4-positive mononuclear infiltrate is seen in the salivary and lacrimal glands.94 The disease is seen more commonly in men (2 : 1) and usually manifests in middle age. More than 85% of patients present after the age of 50 years. One of the most common presentations is painless obstructive jaundice due to obstruction of the intrapancreatic bile duct (see Fig. 59-2B). Jaundice may occur from compression of the bile duct by the enlarged pancreas or by infiltration of the biliary tree by the chronic inflammatory process. Additional symptoms may include weight loss, vomiting, and glucose intolerance.24 Although pain is not frequently present, abdominal and referred back pain may occur. These clinical features, coupled with imaging studies demonstrating diffuse or focal pancreatic enlargement (see Fig. 59-2C), often raise the suspicion of pancreatic adenocarcinoma. In studies in the United States in patients who underwent pancreatic resection for presumed pancreatic carcinoma but were found to have no malignancy in the resected specimen, up to 11% show evidence of autoimmune pancreatitis.95 Jaundice or cholestasis may also occur due to strictures of the biliary tree in other locations. A pattern similar to that seen in primary sclerosing cholangitis (PSC) is seen, with a predilection for involvement of the hilar region. The pattern may mimic not only PSC but also cholangiocarcinoma. The disease, unlike classic PSC, is not associated with inflammatory bowel disease and is steroid-responsive (see Chapter 68). Additional clinical manifestations include a sclerosing sialadenitis (usually presenting as bilateral symmetrical swelling of the salivary glands), retroperitoneal fibrosis (most commonly presenting as hydronephrosis due to entrapment of the ureters), tubulointerstitial nephritis, lymphadenopathy (particularly mediastinal, cervical, and abdominal), prostatitis, and sclerosing cholecystitis, interstitial pneumonia, pseudotumors of the liver, lung, and pituitary. Abdominal ultrasonography usually shows a diffusely enlarged and hypoechoic pancreas. The appearance on EUS is similar. CT most commonly reveals a diffusely enlarged sausage-shaped pancreas (see Fig. 59-2C) in which enhancement with the intravenous contrast agent is delayed and prolonged.24,91,96 Some patients may have a capsule-like lowdensity rim around the pancreas in delayed images. Focal swelling can also occur, mimicking a pancreatic mass.24,91,96-98 Additional CT findings, such as contiguous fibrosis and inflammation extending into the retroperitoneum or surrounding the retroperitoneal vessels, can also raise a suspicion of carcinoma. One of the major clinical challenges in such cases is differentiating these inflammatory pseudotumors from true tumors.97 MRI of the pancreas also may reveal this diffuse pancreatic enlargement, typically with decreased T1-weighted intensity and increased T2-weighted intensity. MRCP can be very helpful in identifying the biliary strictures and in visualizing the pancreatic duct. CT and MRI may also reveal less common findings of autoimmune pancreatitis including gland atrophy and pancreatic calcifications. Endoscopic ultrasonography may demonstrate a diffusely enlarged and hypoechoic gland. The use of fine-needle aspiration of the gland is usually not diagnos-

tic, although there are reports of Tru-cut needle biopsy being diagnostic.99 One of the hallmarks of autoimmune pancreatitis is diffuse or segmental irregularity and narrowing of the pancreatic duct (see Fig. 59-2D). The duct may be diffusely narrowed and thread-like, or may instead demonstrate alternating areas of stricture and normal caliber or dilated duct.24,91,93,98 MRCP is often able to identify the pancreatic duct abnormalities but may not be able to visualize the duct if it is thread-like and diffusely affected. ERCP is better able to visualize the pancreatic duct, but carries more risk and cost than MRCP. Some patients may have a more focal segmental or isolated area of pancreatic duct narrowing, in a pattern more suggestive of malignancy. Some studies in which a second ERCP has been performed note progression from a segmental form to diffuse form in the absence of glucocorticoid treatment.24,100 The pathologic hallmark in the pancreas is infiltration of inflammatory cells and fibrosis around medium-sized interlobular ducts. Obliterative phlebitis of medium and small veins and a whirling or storiform fibrosis of the pancreas, which may extend to the peripancreatic tissues, is characteristic. The inflammatory infiltrate is composed of T lymphocytes and plasma cells. Interstitial fibrosis with acinar cell atrophy is another common finding. Two histologic variants have been described.24 The first, LPSP, is defined by the histologic features noted earlier (Fig. 59-3). A second variant, termed idiopathic duct centric chronic pancreatitis (IDCP), is less common. This variant has a neutrophilic infiltrate with microabscesses, and obliterative phlebitis is rare relative to the aforementioned LPSP variant. End-stage disease may demonstrate atrophy and calcification in the gland, and at that point it may be indistinguishable from other forms of chronic pancreatitis. The disease may be suspected from the clinical and imaging features previously noted. Laboratory evaluation may reveal elevations in serum immunoglobulins,24 seen in from one half to two thirds of cases, especially in IgG4. In studies from Japan, an IgG4 level greater than 135 mg/dL has a sensitivity of 90% and a specificity of greater than 95% for autoimmune pancreatitis.88,91 A large study in the United States, using an IgG4 cut-off of 140 mg/dL, found a sensitivity of 76% and a specificity of 93%.101 This study demonstrated that patients with histologically proven autoimmune pancreatitis may have a normal serum level of IgG4, and that up to 10% of patients with pancreatic ductal adenocarcinoma may have elevations in IgG4 levels. A variety of autoanti­ bodies have been reported, including antinuclear antibodies (ANAs), antilactoferrin antibodies, anticarbonic anhydrase II antibodies, antismooth muscle antibodies (ASMs), rheumatoid factor (RF), and antimitochondrial antibody (AMA). These autoantibodies do not have the sensitivity of IgG4 and hence are inferior for diagnostic purposes. There are several proposals for diagnostic criteria for autoimmune chronic pancreatitis. The first widely used system102 proposed by the Mayo Clinic uses the mnemonic HISORt (histology, imaging, serology, other organ involvement, and response to steroid therapy). A consensus conference led to combined Japan-Korea diagnostic guidelines.91 These systems are presented in Table 59-2. Autoimmune chronic pancreatitis may progress rapidly, from the initial symptoms to end-stage chronic pancreatitis, within months. There is evidence that early therapy with glucocorticoids may prevent subsequent disease complications.24,103 Glucocorticoid therapy usually produces rather dramatic improvement with rapid resolution of both symptoms and radiographic abnormalities. There are no clear recommendations for glucocorticoid dose, although 30 to 40 mg of prednisone orally per day for four to eight weeks

991

992

Section VII  Pancreas

A

B

C

D

Figure 59-3.  Lymphoplasmacytic sclerosing pancreatitis, the most common form of autoimmune pancreatitis. A, Cuff-like periductal lymphoplasmacytic infiltration with normal surrounding pancreatic parenchyma. (Hematoxylin and eosin, ×20.) B, Prominent periductal infiltrate. (Hematoxylin and eosin, ×200.) C, Plasma cell-rich, mixed infiltrate around bile ducts. (Hematoxylin and eosin, ×200.) D, Another example of a cuff-like infiltrate with periductal fibrosis. (Hematoxylin and eosin, ×20.) (Courtesy of Dr. Pamela Jensen, Dallas, Tex.)

is reasonable.24 Repeat pancreatic imaging at four weeks is prudent, to assess for clinical response. Once a response is clear-cut, tapering of the prednisone dose at a rate of 5 mg per week is typical. Complete serologic response (normalization of serum IgG4) may not be apparent for several months, although decreases may be seen within four weeks. Prior to initiating glucocorticoid therapy, vigorous efforts should be made to exclude malignancy. In some cases this is not possible, and some possibility of underlying malignancy may remain. A typical scenario might be a middleaged man with obstructive jaundice and an enlargement of the pancreatic head, in whom ERCP and EUS with tissue sampling have been inconclusive. In this setting, there may be concern in delaying the ultimate diagnosis of malignancy for a therapeutic trial of glucocorticoids. Response to glucocorticoid therapy, however, is usually obvious within two to four weeks and such a trial is not unreasonable if followup is intense. Some patients in whom malignancy remains a possibility undergo laparotomy and often resection of the mass. Between 30% and 40% of patients experience a relapse after glucocorticoid therapy.24,104-106 Relapse may be managed by a repeat course of glucocorticoids followed by maintenance at a low dose of prednisone (e.g., 5 to 10  mg/day). Immunodulators, such as azathioprine, have been used in a few steroid-dependent patients with success.24 Glucocorticoid therapy may improve not only the structural abnor-

malities of the pancreas but also exocrine and endocrine pancreatic function (and salivary function if it is affected).107 Improvement in function is variable, depending on the level of fibrosis and atrophy already established when therapy is initiated. Clinical relapses after resection (e.g., pancreaticoduodenectomy) are rare.

OBSTRUCTIVE CHRONIC PANCREATITIS

Obstruction of the main pancreatic duct by tumors, scars, ductal stones, duodenal wall cysts, or stenosis of the papilla of Vater or the minor papilla can produce chronic pancreatitis in the parenchyma upstream of the obstruction. Obstruction of the pancreatic ducts may also be an important contributor to other forms of chronic pancreatitis (i.e., obstruction of small or large ductal branches by protein precipitates in alcoholic chronic pancreatitis). Obstructive chronic pancreatitis, however, refers to a distinct entity produced by a (generally) single dominant narrowing or stricture of the main pancreatic duct. It is believed that relief of the obstruction can lead to reversal of pancreatic damage, although there is no direct proof of this belief. A number of distinct entities can produce obstructive chronic pancreatitis. Acquired strictures of the main pancreatic duct can occur as a consequence of tumor obstruction from adenocarcinoma, islet cell tumor, intraductal papillary mucinous tumors, or ampullary neoplasms (see Chapter 60). Benign strictures may also develop after a severe attack of acute

Chapter 59  Chronic Pancreatitis Table 59-2  Diagnostic Criteria for Autoimmune Pancreatitis FEATURE

MAYO CLINIC HISORt CRITERIA

ASIAN CONSENSUS CRITERIA

Histology

Lymphoplasmacytic infiltration with fibrosis and abundant IgG4-positive cell infiltration

Imaging

At least one of the following: Periductal lymphoplasmacytic infiltrate with obliterative phlebitis and storiform fibrosis Lymphoplasmacytic infiltrate with storiform fibrosis with abundant IgG4-positive plasma cells (≥10/HPF) Typical: diffusely enlarged gland with delayed rim enhancement, diffusely irregular and attenuated pancreatic duct Other: focal pancreatic mass or enlargement, focal pancreatic duct stricture; pancreatic atrophy, calcification, or pancreatitis

Serology

Elevated serum IgG4 level

Other organ involvement

Hilar or intrahepatic biliary strictures, distal (intrapancreatic) bile duct stricture, parotid or lacrimal gland involvement, mediastinal lymphadenopathy, retroperitoneal fibrosis Resolution or marked improvement of pancreatic or extrapancreatic manifestations

Response to glucocorticoid therapy Diagnostic groups

Group A: diagnostic pancreatic histology Specimen demonstrating the full spectrum of LPSP or ≥10 IgG4-positive cells/HPF Group B: typical imaging and serology CT or MRI showing diffusely enlarged pancreas with delayed and rim enhancement Pancreatogram showing diffusely irregular pancreatic duct Elevated serum IgG4 level Group C: response to glucocorticoids Unexplained pancreatic disease after negative workup for other causes and Elevated serum IgG4 or other organ involvement confirmed by the presence of abundant IgG4-positive cells and Resolution or marked improvement in pancreatic or extrapancreatic manifestations with glucocorticoid therapy

Both of the following: Diffuse, segmental, or focal enlargement of the gland, with or without a mass or hypoattenuated rim Diffuse, segmental, or focal pancreatic duct narrowing, with or without stenosis of the bile duct One of the following: High levels of total IgG or IgG4 Detection of autoantibodies (ANA, RF) Not included in criteria Included as an optional criteria, but only in patients with both imaging criteria above with a “negative workup” for pancreatobiliary cancer Not applicable

ANA, antinuclear antibodies; CT, computed tomography; HISORt, histology, imaging, serology, other organ involvement, and response to glucocorticoid therapy; HPF, high-power field; IgG, immunoglobulin G; LPSP, lymphoplasmacytic sclerosing pancreatitis; MRI, magnetic resonance imaging; RF, rheumatoid factor. From references 91 and 102.

pancreatitis, particularly an episode associated with significant pancreatic necrosis and pseudocysts. Blunt and penetrating trauma to the pancreas can also lead to pancreatic duct strictures, most commonly in the midbody of the gland where the duct crosses the spine. Each of these processes can be associated with chronic pancreatitis in the gland upstream from the obstruction. The pathology in obstructive chronic pancreatitis is one of diffuse interlobular and intralobular fibrosis, usually equally and symmetrically distributed in the affected region. Pancreas divisum is a common normal variant, occurring in approximately 4% to 11% of the population (see Chapter 55). In some patients with this anomaly, the minor papilla may be inadequate to allow free flow of pancreatic juice into the duodenum, possibly causing acute episodes of pancreatitis. Pancreas divisum is not often considered a cause of chronic pancreatitis. Large natural history studies have failed to identify a clear link between pancreas divisum and either acute or chronic pancreatitis. Nonetheless, rare patients may present with pancreas divisum and substantial dilation of the dorsal pancreatic duct, associated with obstructive chronic pancreatitis. These rare patients have strictures of the minor papilla. Most analyses of the effectiveness of minor papilla endoscopic sphincterotomy and stenting, or minor papilla surgical sphincteroplasty, have

not specifically examined the response to these therapies in this rare subgroup of patients,108 and the response to these therapies even in the setting of acute relapsing pancreatitis is unknown. Endoscopic therapy or surgical therapy for patients with pancreas divisum, chronic pancreatic type pain, and a nondilated dorsal pancreatic duct is rarely, if ever, effective. Sphincter of Oddi dysfunction (see Chapter 63), like pancreas divisum, is most often proposed as a cause of acute or recurrent acute pancreatitis rather than chronic pancreatitis. Although abnormalities of sphincter of Oddi manometry may be seen in some patients with chronic pancreatitis,109,110 they are unpredictable and not likely to be playing a role in causing disease. Interestingly, one family in whom sphincter of Oddi dysfunction had been presumed to be causing chronic pancreatitis was actually found to suffer from a mutation of the trypsinogen gene.1 It is much more likely that abnormal sphincter of Oddi pressures are a consequence of chronic pancreatitis rather than a cause. The response to sphincter ablation in patients with chronic pancreatitis and sphincter of Oddi dysfunction is unpredictable but generally poor. Surgical textbooks have long cautioned against sphincteroplasty as therapy for chronic pancreatitis, recognition of the general lack of efficacy of sphincter ablation in this situation.

993

994

Section VII  Pancreas MISCELLANEOUS Recurrent or Severe Acute Pancreatitis

Recurrent episodes of acute pancreatitis of any etiology may eventually lead to the development of a chronic inflammatory response within the pancreas, the activation of pancreatic stellate cells, and chronic pancreatitis. This can also happen with even one severe attack, usually with associated pancreatic necrosis or the need for surgical débridement. One more common example of this is hypertriglyceri­ demia. Elevations of serum triglyceride values greater than 1000 mg/dL can produce acute pancreatitis. Many of these patients will have repeated clinical and subclinical attacks of acute inflammation, and ultimately develop chronic pancreatitis.111 In patients with any etiology of severe acute pancreatitis complicated by substantial pancreatic necrosis, chronic pancreatitis may develop (see Chapter 58). This process occurs most commonly in patients who undergo surgical necrosectomy. In one study, patients with necrotizing biliary pancreatitis who did not undergo necrosectomy had exocrine insufficiency and endocrine insufficiency less commonly than those who did (13% vs. 58% and 26% vs. 75%, respectively, for exocrine and endocrine insufficiency).112 Thus, even in the absence of necrosectomy, severe necrotizing gallstone pancreatitis may occasionally lead to chronic pancreatitis.113 Other studies have noted exocrine and endocrine pancreatic insufficiency commonly occur after necrosectomy.114 Persistent decreases in pancreatic function testing may be seen in up to 80% of patients after recovery from necrotizing pancreatitis. Those with acute alcoholic pancreatitis, complicated by significant necrosis, are most likely to develop features of chronic pancreatitis,115 likely because they usually already have underlying chronic pancreatitis at the time of their first severe attack. Residual strictures of the pancreatic duct are also not uncommon after severe acute pancreatitis, and they may also contribute to the development of chronic pancreatitis in this group of patients. In others, a prolonged and smoldering clinical course ultimately leads to chronic pancreatitis.

Asymptomatic Pancreatic Fibrosis

There are several situations in which histologic evidence of chronic pancreatitis, and specifically fibrosis, may be seen in the absence of clinical chronic pancreatitis. Two of these conditions have already been discussed. Older adults commonly develop histologic changes within the pancreas that resemble chronic pancreatitis.3,4,5 ERCP may also demonstrate changes consistent with chronic pancreatitis in these patients.116 These structural changes do not correspond to functional disturbances of pancreatic function or to clinical features of chronic pancreatitis.117 As mentioned, chronic alcoholics who do not have clinical chronic pancreatitis commonly have histologic evidence of chronic pancreatitis.6,7 The incidence of acute pancreatitis is increased in patients undergoing hemodialysis, and some evidence suggests that chronic pancreatitis may also be seen with greater frequency in this population. One ultrasonographic screening study of patients undergoing hemodialysis noted changes consistent with chronic pancreatitis in 20%.118 Autopsy studies also show changes consistent with chronic pancreatitis in the majority of patients,119 although very few of such patients had clinical chronic pancreatitis. Chronic renal failure appears to frequently produce asymptomatic pancreatic fibrosis.120 Diabetes mellitus can occur as a consequence of destruction of islet cells from long-standing chronic pancreatitis.

Up to 1% of all diabetes mellitus is believed to be due to chronic pancreatitis. In contrast, changes in pancreatic morphology and function are common in patients with longstanding diabetes. The pancreas is smaller than normal, particularly in patients with type 1 diabetes.121 The pancreatic duct is abnormal in 40% to 50% of diabetic patients when studied by ERCP, with abnormalities suggestive of chronic pancreatitis.122,123 Pancreatic function, as defined by fecal elastase measurements124 or by more formal direct pancreatic function testing,125 is abnormal in 40% to 50% of patients. The abnormalities are much more likely to be seen in type 1 than type 2 diabetes. The reason for this association is not clear, and indeed, it is not clear whether the diabetes is causing the changes in the pancreas or vice versa. Because insulin is a trophic factor for the exocrine pancreas, and because diabetes can produce microvascular angiopathy, insulin deficiency and long-standing diabetes together could explain the pancreatic damage. Although these measures of pancreatic structure and function may be present, they are usually not responsible for symptoms and these patients do not routinely merit treatment with pancreatic enzymes.126 Clinical evidence of chronic pancreatitis and exocrine insufficiency is very rare in diabetic patients. These observations may be combined to suggest that a wide variety of disease states and normal wear and tear (e.g., normal aging, chronic alcohol ingestion) on the pancreas may produce damage that by histologic criteria resembles chronic pancreatitis, but is not sufficient to cause symptoms and is not severe enough to cross the diagnostic threshold of labeling that patient as having this disease. This concept will be discussed further under “Diagnosis.”

IDIOPATHIC CHRONIC PANCREATITIS

Idiopathic pancreatitis accounts for 10% to 30% of all cases of chronic pancreatitis. However, many patients are probably mislabeled as having idiopathic disease. Given that there is no threshold of alcohol ingestion for alcoholinduced chronic pancreatitis, some of these patients certainly suffer from alcoholic chronic pancreatitis. Similarly, some of these idiopathic cases occur in patients with genetic abnormalities,45,52,55 and some of what we formerly labeled idiopathic is instead autoimmune. Interpreting the literature on idiopathic chronic pancreatitis is therefore difficult because most studies of this entity are probably dealing with cases with several different etiologies. Idiopathic chronic pancreatitis appears to occur in two forms, an early-onset type that manifests in the late second or third decade of life and a late-onset form that appears in the sixth or seventh decade of life.58,127 Early-onset idiopathic chronic pancreatitis has a mean age at onset of around 20 years. There appears to be an equal gender distribution,58 although one series found a male preponderance.127 Pain is the predominant feature of this disease, occurring in up to 96% of patients, a higher rate than in either alcoholic or late-onset chronic pancreatitis (see later). Pancreatic calcifications, exocrine insufficiency, and endocrine insufficiency (i.e., diabetes) are extremely rare at presentation (<10%) and develop very slowly thereafter. The mean time to calcification in this group is 25 years, to exocrine insufficiency 26 years, and to endocrine insufficiency 27.5 years.58 Complications of chronic pancreatitis (pseudocyst, abscess, biliary obstruction, and duodenal obstruction) occur in about 20% of patients, whereas surgery (primarily for abdominal pain) is ultimately needed in 60%. Thus, early-onset idiopathic chronic pancreatitis is a disease characterized by severe pain but much delayed development of structural (calcifications) or functional (exocrine or endocrine insufficiency) evidence of chronic

Chapter 59  Chronic Pancreatitis pancreatitis. The delay may make diagnosis quite difficult because most available diagnostic tools rely on these structural or functional abnormalities. Late-onset idiopathic chronic pancreatitis manifests less commonly as pain. In the best-documented series,58 only 54% of patients presented with pain, although 75% ultimately experienced pain. The median age of onset is 56 years, and the disease occurs equally in men and women. Exocrine insufficiency (22%) and endocrine insufficiency (22%) were present not infrequently at the time of diagnosis, ultimately occurring in 46% and 41% of cases, respectively. The median time to development of exocrine insufficiency and endocrine insufficiency was 16.9 and 11.9 years, respectively. Life-table analysis suggested that with very long follow-up (>30 years), exocrine insufficiency will ultimately develop in 75%, endocrine insufficiency in 50% to 60%, and diffuse pancreatic calcifications in 90% of patients. The disease therefore tends to be one of a comparatively painless course associated with the frequent development of pancreatic calcifications, exocrine insufficiency, and endocrine insufficiency. Aging itself can be associated with the development of structural changes within the pancreatic parenchyma and duct that are indistinguishable from those seen in late-onset chronic pancreatitis,117 so the distinction between normal aging and late-onset idiopathic chronic pancreatitis may not always be clear.

CLINICAL FEATURES ABDOMINAL PAIN

Abdominal pain is the most common clinical problem in patients with chronic pancreatitis, and the symptom that most detracts from quality of life. Severe pain decreases appetite and limits food consumption, contributing to weight loss and malnutrition. Chronic severe pain leads to a dramatic reduction in quality of life,16-19 loss of social functioning, and the all too frequent addiction to narcotic analgesics. Intractable pain is the most common reason for hospitalization and for surgery in patients with chronic pancreatitis. There is no single characteristic pain pattern. Pain is most commonly described as being felt in the epigastrium, often with radiation to the back. Pain is usually described as boring, deep, and penetrating and is often associated with nausea and vomiting. Pain may be relieved by sitting forward or leaning forward, by assuming the kneechest position on one side, or by squatting and clasping the knees to the chest. Pain may worsen after a meal and often is nocturnal. The natural history of abdominal pain varies and is difficult to predict. As an example, many patients with alcoholic chronic pancreatitis initially present with episodes of pain interspersed with periods of feeling relatively well. During these more acute episodes of pain, such a patient may be labeled as having acute pancreatitis. As time passes, pain may become more continuous, and the diagnosis of chronic pancreatitis more obvious. Some patients may present with the more gradual onset of constant abdominal pain, and some may have no pain. Once pain develops, it commonly changes over time in character, severity, and timing. Depending on the etiology of chronic pancreatitis, 50% to 90% of patients experience pain during the course of the disease.40,57-59,127-129 Some of these same observational studies document a decrease in pain over time in the majority of patients, although the timing and the magnitude of this decrease vary among the studies. In one study, pain relief appeared to

occur most commonly at the time of development of diffuse pancreatic calcifications, exocrine insufficiency, and endocrine insufficiency.57 Other studies have not found this same correlation, but many have noted a less pronounced tendency for pain to “burn out” over time.58,59,129-131 Pain seems to eventually decrease in at least half of patients. Some of the pain relief is due to surgery for pain or complications, but pain relief over very long follow-up is also seen in medically treated patients in approximately similar proportions.58,59,129,132,133 The pain pattern in an individual patient, however, is not accurately predictable, and the pain may worsen, stabilize, or improve over time. The judgment of therapeutic efficacy for any treatment for chronic pancreatitis must take into account this extremely varying natural history of pain. The proposed causes of pain in chronic pancreatitis are varied but can be distilled to the following two main mechanisms: (1) increased pressure and ischemia and (2) alterations in peripheral and central nociceptive nerves.

Increased Pressure and Ischemia

One commonly invoked mechanism of pain is tissue ische­ mia, driven by increased pressure within the pancreatic duct or parenchyma. Several lines of clinical and experimental evidence point to increased pressure within the pancreatic duct or parenchyma as being important in the genesis of pancreatic pain. Pancreatic ductal and tissue pressures are usually elevated in patients with chronic pancreatitis undergoing surgery for chronic pain.134-136 Elevations in pancreatic ductal pressure measured during ERCP have also been documented in some but not all patients with painful chronic pancreatitis.110,137,138 Surgical drainage of the pancreatic duct leads to an immediate reduction in pressure to normal levels and is associated with pain relief.134,135 However, one study of ERCP-measured pancreatic duct pressures found no difference in patients with chronic pancreatitis between those who had pain and those who did not,137 and one small study of endoscopic stenting found that a reduction in pressure after stenting was not correlated with pain relief.138 One would assume that increased pressure within the pancreatic duct would be related to obstruction of the pancreatic duct, either in the main duct or in side branches. The presence of a pancreatic duct stricture and upstream pancreatic duct dilation might be an accurate indicator of a group of patients with increased pressure and therefore pain. However, there is often not a clear-cut relationship between pancreatic duct strictures or ductal dilation and pain.40,57-59,130,137-139 In many studies, pancreatic duct dilation and strictures are equally common in those with and without pain. Nonetheless, patients with a dilated pancreatic duct or pancreatic duct stricture are most likely to experience pain relief from endoscopic or surgical therapy. The mechanism by which increased pressure could cause pain is speculative but is likely related to pancreatic tissue ischemia. In animal models of chronic pancreatitis, increased pancreatic pressure is associated with reductions in pancreatic blood flow, tissue oxygen tension, and interstitial pH.140 In these models pancreatic secretagogues lead to a further decrease in pancreatic blood flow (rather than the normally expected increase), decreased capillary filling, and worsening tissue ischemia. These observations are consistent with those seen in a compartment type syndrome. Small studies in humans with chronic pancreatitis undergoing surgery also demonstrate lower pancreatic tissue pH than patients without chronic pancreatitis.140 Pancreatic blood flow, measured at ERCP with the use of platinum electrodes, is also lower in patients with chronic pancreati-

995

996

Section VII  Pancreas tis compared with controls.141 Tissue ischemia that is worsened by secretory stimulation of the pancreas may therefore be the mechanism by which elevations in tissue pressure cause pain.

Alterations in Peripheral and Central Nociceptive Nerves

Morphologic studies in patients with chronic pancreatitis demonstrate increases in the diameter and number of intrapancreatic nerves, foci of inflammatory cells associated with nerves and ganglia, and damage to the perineural sheath.142,143 The disruption of the perineural sheath may allow inflammatory mediators to gain access to the neural elements. Regardless of the local events in and around the pancreas causing pain, perception of the pain message requires communication with the central nervous system. The innervation of the pancreas is complex, with both visceral somatic and autonomic nerves (see Chapter 55). Dendrites of the pancreatic nociceptive sensory afferents travel with sympathetic nerves from the pancreas and reach the celiac ganglia, although no synapse is made there. These dendritic fibers continue, bundled as the left and right greater splanchnic nerves, to the sympathetic trunk ganglia, before reaching the first cell body, located in the dorsal root ganglia in spinal cord segments T5 through T9-T10. Projections of these dorsal root neurons often traverse upward and downward for several spinal segments before entering the dorsal horn of the spinal cord. Afferent pain fibers may cross the midline in several of these connections, accounting for the midline perception of pancreatic pain. Axons from the first-order dorsal root ganglion cell bodies have two distinct pathways. Some project to the dorsal horn of the spinal cord and may release a variety of mediators including substance P, calcitonin gene-related peptide, and glutamate onto second-order neurons that project to the thalamus via the spinothalamic white matter columns. These may then synapse with third-order neurons that project to the somatosensory cortex (for cognitive integration of pain) and to the limbic system and hypothalamus (for affective and autonomic integration of pain). A second pathway for projections involves synapses within the same level of the spinal cord with sympathetic efferent cell bodies that project back down the splanchnic nerves to the celiac plexus, with second-order sympathetic neurons projecting back to the pancreas. Vagal afferents may also carry noxious stimuli from the pancreas (especially for stretch). Noxious stimulation of these pathways can occur through a variety of mechanisms. Pressure, ischemia, inflammation, heat, and other classic stimuli can activate these pathways. The accumulation of inflammatory mediators and nerve injury can sensitize the nerve, making it hyper-responsive.144 At the periphery is an increase in substance P and calcitonin gene-related peptide in interlobular and intralobular nerve bundles in patients with chronic pancreatitis.145 Both of these neurotransmitters are involved in pain transmission. The close spatial relationship between intrapancreatic nerves and inflammatory cells strongly supports the concept of neuroimmune interaction. Growth-associated protein 43 (GAP-43), a marker of neuronal plasticity and remodeling, is seen in pancreatic nerves infiltrated with lymphocytes, and its presence correlates with pain.146,147 Additional studies have demonstrated expression of nerve growth factor (NGF) and one of its receptors (TrkA) in patients with painful chronic pancreatitis and in animal models of chronic pancreatitis.147,148 NGF is one of the key molecules involved in sensitization.144 Endogenous proteases, like trypsin, can also activate and sensitize sensory neurons in the pancreas, a process mediated through the protease-

activated receptor-2 (PAR-2).149 Another activator of PAR-2 is tryp­tase, a mast cell product. Interestingly, mast cells are seen commonly in pancreatic tissue specimens from patients with chronic pancreatitis. The exact mechanisms by which the inflammatory cells and their products and intrapan­ creatic neurons interact in chronic pancreatitis remain to be clarified, although the data suggest that the production of sensitizing factors in the vicinity of pancreatic nerves alters sensory neuron form and function. In addition, there is substantial evidence from studies of other types of chronic pain that chronic peripheral nerve injury or inflammation leads to changes in nociceptive processing that involve both the spinal cord and central nervous system. Chronic pain can produce a centrally sensitized pain state in which elimination of the original source of pain does not relieve pain.144,150 In this situation, pain may occur in response to innocuous or physiologic stimuli (allodynia) or may respond in an exaggerated fashion to stimuli that are painful (hyperalgesia). These phenomena depend on changes at both the spinal cord level and the brain. One study monitoring EEG activity of patients with chronic pancreatitis noted increased areas of referred pain to electrical stimulation of the esophagus, stomach, or duodenum.151 The central nervous system reorganization and plasticity underlying hyperalgesia and allodynia are likely major factors limiting the effectiveness of treatments for pain. Nowhere is this fact made more obvious than in the patient who continues to have pancreatic pain after a total pancreatectomy. Pain is complex, and no single mechanism is likely to be present in all patients, implying that no single therapy will be effective.

Other Causes of Pain

In addition to the two main mechanisms noted, a variety of other contributors should be considered. Complications of chronic pancreatitis may cause pain in their own right. These complications include duodenal obstruction, bile duct obstruction, a pseudocyst, and secondary pancreatic carcinoma. These usually have specific therapy. Hyperstimulation of the pancreas via CCK has also been postulated as a potential cause of pain because serum levels of CCK may be elevated in chronic pancreatitis and stimulation of the pancreas by CCK could increase pressure within the gland or facilitate basolateral rather than apical secretin of enzymes. Reducing serum CCK levels is the proposed mechanism for pancreatic enzymes reducing pain.

STEATORRHEA

The human pancreas has substantial exocrine reserve. Steatorrhea does not occur until pancreatic lipase secretion is reduced to less than 10% of the maximum output.152 Steatorrhea is therefore a feature of far-advanced chronic pancreatitis, in which most of the acinar cells have been injured or destroyed, but may also be seen with complete blockage of the pancreatic duct. With advanced chronic pancreatitis, maldigestion of fat, protein, and carbohydrates will occur. Azotorrhea (protein maldigestion) also occurs when secretion of proteases is less than 10% of normal.152 Affected patients may present with diarrhea and weight loss. Some patients may note bulky foul-smelling stools or may even note the passage of frank oil droplets. Unlike in small bowel diseases associated with malabsorption, watery diarrhea, excess gas, and abdominal cramps are uncommon in the steatorrhea seen in patients with chronic pancreatitis. This difference may be due to better-preserved carbohydrate absorption and small bowel and colonic function in patients with chronic pancreatitis and exocrine insufficiency than in

Chapter 59  Chronic Pancreatitis those with small intestinal diseases such as celiac disease. Even when there is significant loss of fat in stool, most patients pass only three or four stools daily and some may pass only one. In general, fat maldigestion occurs earlier and is more severe than protein or carbohydrate maldigestion. There are several explanations for this phenomenon. First, fat digestion depends primarily on pancreatic lipase and colipase, although gastric lipase is able to hydrolyze up to 20% of dietary fat (see Chapter 49). Second, lipase output decreases earlier and more substantially as chronic pancreatitis progresses compared with the secretion of other pancreatic enzymes such as trypsin or amylase. Third, lipase is more sensitive to acid destruction than other pancreatic enzymes. As bicarbonate secretion decreases in chronic pancreatitis and duodenal pH drops, lipase in particular is inactivated. Fourth, in addition to lipase inactivation, the low duodenal pH also predisposes to precipitation of bile salts, thereby preventing the formation of mixed micelles and further interfering with lipid digestion and absorption. Fifth, lipase is more sensitive to digestion and degradation by pancreatic proteases than other digestive enzymes. The median time to development of exocrine insufficiency in chronic pancreatitis has been reported to be as low as 5.6 years,57 but most studies report longer duration of disease prior to development of steatorrhea. In one large natural history study, the median time to development of exocrine insufficiency was 13.1 years in patients with alcoholic chronic pancreatitis, 16.9 years in patients with lateonset idiopathic chronic pancreatitis, and 26.3 years in patients with early-onset idiopathic chronic pancreatitis.58 With very long follow-up, approximately 50% to 80% of patients with chronic pancreatitis eventually have exocrine insufficiency.57-59 Significant weight loss is uncommon despite maldigestion. Patients generally increase their caloric intake to compensate for stool losses. Weight is usually maintained despite the fact that the resting energy expenditure is generally increased in patients with chronic pancreatitis. Weight loss is most commonly seen during painful flares that prevent adequate oral intake because of pain, nausea, or vomiting. Weight loss may also occur as a result of the development of a concomitant disease such as small bowel bacterial overgrowth (SBBO) or pancreatic or extrapancreatic malignancy. Finally, weight loss may occur in patients who develop financial difficulties, suffer from chronic severe alcoholism, or lose social support because these may contribute to inadequate caloric and protein intake. Substantial weight loss should lead to an investigation of these potential causes. More subtle weight loss may be more common than is appreciated.153,154 Deficiencies of fat-soluble vitamins may develop in patients with chronic pancreatitis and steatorrhea.155,156 Significant vitamin D deficiency and osteopenia and osteoporosis occur in patients with chronic pancreatitis.154,157-160 These studies demonstrate osteopenia in 50% to 70% of patients and osteoporosis in up to 20% of patients with chronic pancreatitis and steatorrhea. Levels of the active forms of vitamin D [1,25-(OH)2-D3 and 25-(OH)D3] are generally below the normal range in these patients. Deficiencies of water-soluble vitamins and micronutrients are rare and generally seen only as a consequence of inadequate intake in chronic alcoholics. Despite the fact that vitamin B12 absorption requires intact pancreatic function to degrade R-factor from dietary cobalamin (see Chapter 49), vitamin B12 deficiency is uncommon in patients with chronic pancreatitis in the absence of ongoing alcohol abuse.

DIABETES MELLITUS

Like exocrine insufficiency, endocrine insufficiency is a consequence of long-standing chronic pancreatitis and is especially common after pancreatic resection and in tropical (fibrocalcific) pancreatitis. Islet cells appear to be relatively resistant to destruction in chronic pancreatitis (see Fig. 59-1).161,162 When diabetes develops, the mechanism is more complex than just a simple loss of beta cells due to the progressive destruction of islets.162 Infusion of glucagonlike peptide 1 (GLP-1) in patients with diabetes due to chronic pancreatitis stimulates release of endogenous insulin, confirming that some residual functional beta cells remain.162-164 In addition, levels of islet amyloid polypeptide (amylin) may be elevated in patients with diabetes due to chronic pancreatitis, a feature most commonly associated with insulin resistance in type 2 diabetes mellitus.165,166 Several studies in animal models and humans also demonstrate a loss of hepatic insulin receptor expression and an impairment in hepatic insulin receptor function, a change that can produce decreased hepatic glucose output, and that appears to be due to decreased pancreatic polypeptide secretion from islet cells.162 There is also a relative decrease in stimulated glucagon secretion from these damaged islets, although basal levels may remain normal.163 These various factors make diabetes due to chronic pancreatitis different than either type 1 or type 2 diabetes. About half of patients with chronic pancreatitis who develop diabetes will require insulin.163 Unlike type 1 diabetes, insulin-producing beta cells and glucagon-producing alpha cells are injured. This combination increases the risk of prolonged and severe hypoglycemia with overvigorous insulin treatment, owing to the lack of a compensatory release of glucagon.162,167,168 Diabetes mellitus appears to be nearly as common as steatorrhea in patients with far-advanced chronic pancreatitis. In one study the median times to development of diabetes in patients with alcoholic, late-onset idiopathic, and early-onset idiopathic chronic pancreatitis were 19.8 years, 11.9 years, and 26.3 years, respectively.58 Other studies have noted shorter median times of 6 to 10 years.57,129 Ultimately, 40% to 80% of patients with chronic pancreatitis have diabetes after long follow-up, depending on etiology.162,163 In one large cohort study, 83% of patients had diabetes 25 years after the clinical onset of chronic pancreatitis.169 In that study, risk factors for the development of diabetes included early onset of pancreatic calcifications and resection of the pancreatic tail. The latter risk has been seen in other studies170 and is probably explained by the observation that the islets may be concentrated in the body and tail of the pancreas. Microangiopathic complications are as common in patients with diabetes associated with chronic pancreatitis as in patients with type 1 diabetes with similar duration of disease.171,172

PHYSICAL EXAMINATION Very few of the physical examination findings are diagnostic or specific for chronic pancreatitis. Patients generally appear well nourished and demonstrate mild to moderate abdominal tenderness. In chronic alcoholics with advanced disease, weight loss and malnutrition may be more evident or one may see signs and stigmata of coexistent cirrhosis. Rarely, a palpable mass is found, indicating a pseudocyst. Jaundice may be seen in the presence of coexistent alcoholic liver disease or bile duct compression within the head of the pancreas. A palpable spleen may also rarely be found

997

998

Section VII  Pancreas in patients with thrombosis of the splenic vein as a consequence of chronic pancreatitis or in patients with portal hypertension due to coexistent chronic liver disease. In some patients with autoimmune pancreatitis, evidence of a coexistent autoimmune feature, such as salivary gland enlargement or lymphadenopathy, may be found.

DIAGNOSIS An impressive number and variety of diagnostic tests for chronic pancreatitis have been developed over the last few decades. This bewildering array is not only confusing but serves to point out that no single test is so accurate that it can replace all others. Many of these diagnostic tests have not been studied in a sufficiently wide spectrum of disease to define their sensitivity and specificity. These diagnostic tests are usually separated into those that are designed to detect abnormalities of pancreatic function (discussed in Chapter 56) and those that detect abnormalities of pancreatic structure (Table 59-3). Before considering each of these tests in more detail, it is useful to remember that in almost all patients, chronic pancreatitis is a slowly progressive disease. In the early stages within the pancreas, chronic inflammation, cellular necrosis and apoptosis, and activation of pancreatic stellate cells have all developed, but these features of chronic pancreatitis remain visible only on histology. With progressive fibrosis and loss and destruction of tissue, the disease becomes more evident. Abnormalities of pancreatic structure or function may take years or even decades to develop, or may not develop at all.1,57,58 All available diagnostic tests are most accurate in far-advanced disease, when obvious functional or structural abnormalities have developed. Conversely, to greater or lesser degrees, all diagnostic tests are less accurate in less advanced or early chronic pancreatitis. Functional abnormalities in chronic pancreatitis include exocrine insufficiency (maldigestion and steatorrhea), and endocrine insufficiency (diabetes mellitus). In addition, some diagnostic tests measure maximum stimulated secre-

Table 59-3  Available Diagnostic Tests for Chronic Pancreatitis* Tests of function

Tests of structure

Direct hormonal stimulation (with pancreatic stimulation by secretin or cholecystokinin or both):   Using oroduodenal tube†   Using endoscopy† Magnetic resonance cholangiopancreatography with secretin stimulation Fecal elastase Fecal chymotrypsin Serum trypsinogen (trypsin) Fecal fat Blood glucose level Endoscopic ultrasonography Endoscopic retrograde cholangiopancreatography Magnetic resonance imaging with magnetic resonance cholangiopancreatography Computed tomography Abdominal ultrasonography Plain abdominal film

*Ranked in estimated order of decreasing sensitivity for each category. † See text for explanations.

tory capacity of the pancreas, which becomes abnormal before there is failure of exocrine or endocrine function. Structural abnormalities that can be diagnostic include changes within the main pancreatic duct (dilation, strictures, irregularity, pancreatic ductal stones), side branches of the pancreatic duct (dilation, irregularity), or pancreatic parenchyma (lobularity of the gland, alterations in echogenicity, cysts, enlargement or atrophy, and others). Patients with alcoholic chronic pancreatitis, hereditary chronic pancreatitis, tropical pancreatitis, and late-onset idiopathic chronic pancreatitis are most prone to development of these abnormalities of function or structure, although the process may still take several years. These changes develop particularly slowly, and sometimes not at all, in patients with early-onset idiopathic chronic pancreatitis.58 The preceding observations have led to a general classification of chronic pancreatitis as either big-duct or small-duct disease. Big-duct disease implies substantial abnormalities of the pancreatic duct (generally, dilation of the main pancreatic duct visible on ultrasonography, CT, and pancreatography). Small-duct disease (also called minimal-change chronic pancreatitis) implies the absence of these findings (e.g., normal or equivocal findings on ultrasonography, CT, or ERCP). Similarly, big-duct disease is usually associated with functional abnormalities such as diabetes and steatorrhea, whereas small-duct disease is less commonly associated with exocrine or endocrine insufficiency. This distinction has both diagnostic and therapeutic implications. The diagnosis of big-duct disease is much simpler, the disease is usually due to alcohol abuse, and the treatment options focus more on decompressing the dilated pancreatic duct. The diagnosis of “small-duct” disease is much more difficult because imaging findings and functional parameters may be equivocal, the disease is more commonly idiopathic, and treatment options focus on medical therapy rather than surgical or endoscopic attempts to decompress the pancreatic duct. The determination of the sensitivity, specificity, and accuracy of any of these diagnostic tests requires that the test result be compared with some gold standard, a test that gives reliable and certain evidence as to the presence or absence of disease. In the case of chronic pancreatitis, this gold standard is pancreatic histology (see Fig. 59-1). Unfortunately, the histologic changes are not uniform throughout the gland2,22,23 so that findings in a biopsy specimen may be misleading. Even more important, obtaining pancreatic tissue carries risk and is seldom performed solely for diagnosis. Given the lack of a useful gold standard, one is often left with comparing a new diagnostic test with some substitute for the gold standard. One such substitute is long-term follow-up. Most series have not monitored patients diagnosed with early chronic pancreatitis or possible early chronic pancreatitis (patients in whom diagnostic tests are not unequivocally positive) for long enough to establish the presence or absence of chronic pancreatitis with certainty. The second potential substitute for the gold standard is some other diagnostic test, and in fact, new diagnostic tests are often compared with such modalities as ERCP, CT, and pancreatic function tests, and composites of these.173 In patients with chronic pancreatitis and far-advanced structural or functional abnormalities, little else can mimic these abnormalities, and essentially all diagnostic tests are accurate. The situation is quite different in patients with early or less advanced or minimal-change chronic pancreatitis, and even more so in patients with suspected or possible chronic pancreatitis, in whom these easily identifiable structural or functional abnormalities are lacking. In this

Chapter 59  Chronic Pancreatitis latter situation, only tests of maximum sensitivity have a chance of enabling a diagnosis, and the lack of a gold standard can lead to diagnostic confusion and difficult decision making. In addition to choosing a diagnostic test on the basis of sensitivity and specificity, clinicians must consider the availability, cost, and risk of each of these tests to maximize diagnostic information and minimize risk. These issues are discussed here in relation to each of the available diagnostic tools.

TESTS OF PANCREATIC FUNCTION

Tests of pancreatic function can be divided into those that directly measure pancreatic function by measuring the output of enzymes or bicarbonate from the pancreas and those that measure the released enzymes indirectly (through its action on a substrate or its level in blood or stool) (see also Chapter 56).

Direct Tests

Direct hormonal stimulation tests are believed to be the most sensitive function test for chronic pancreatitis.1,174-176 A few studies have compared the results of these hormonal stimulation tests with pancreatic histology,177-179 with overall sensitivities ranging from 67% to 88%. The largest study com­ pared histology with combined secretin-cholecystokinin testing in 108 patients.177 There was a linear correlation of stimulated bicarbonate output with histologic severity of chronic pancreatitis. Although mean peak bicarbonate concentration was in the normal range (>80  mEq/L) in 69 patients with normal or equivocal histology, mean bicarbonate concentration was 70, 63, and 50  mEq/L in those with mild, moderate, and severe histologic chronic pancreatitis, respectively. The overall sensitivity of hormonal stimulation testing in this study was 67%, with a specificity of 90% and overall accuracy of 81%. When the analysis was restricted to the 29 patients with moderate or severe histologic changes of chronic pancreatitis, the sensitivity of hormonal stimulation testing rose to 79%. In this same group of 29 patients, the sensitivity of ERCP was 66%. In comparisons with ERCP, direct hormonal stimulation tests appear to be on average somewhat more sensitive for the diagnosis of chronic pancreatitis. The values for sensitivity of pancreatic function testing range from 74% to 97%, with specificity ranging from 80% to 90%.174,175,180-185 In these studies the two tests agree in about three quarters of patients, although some studies note higher rates of concordance. Most studies also note a general correlation between increasing structural abnormalities and progressive abnormalities of hormone stimulation test results, although the relationship is not exact. Most of these studies also identify patients with discordant results—patients with abnormal ERCPs and normal hormonal stimulation test results as well as those with normal ERCPs and abnormal hormonal stimulation test results. In four studies the percentage of patients with an abnormal hormonal stimulation test result and a normal ERCP ranged from 3% to 20%.180-184 Two small studies have followed such patients whose diagnosis was based solely on an abnormal hormonal stimulation test result, and both found development of chronic pancreatitis on follow-up in 90% of patients.184,186 These data point out that direct pancreatic function testing appears to be able to identify a group of patients with chronic pancreatitis who have functional abnormalities of stimulated secretion but who do not (yet) have ERCP-identifiable structural abnormalities. Conversely, most of these studies also report patients with normal hormonal stimulation test results and abnormal ERCPs. This group of patients is generally less common,

averaging less than 10% in several studies.180-186 Long-term follow-up in a small group of such patients noted development of chronic pancreatitis in 0% to 26%.184,186 These studies point out that in situations in which results of the two tests disagree, hormonal stimulation testing appears to be somewhat more sensitive and specific than ERCP. Some experts have suggested that the pancreas has such reserve that 30% to 50% damage to the gland is necessary before direct pancreatic function tests yield reliably positive results. Despite their theoretical advantages, direct pancreatic function tests have a number of limitations.175,176 First, they have not been well standardized across institutions offering the test. Second, they are available at only a very few referral centers and so are not available to the majority of clinicians seeing patients with chronic pancreatitis. Third, it can be difficult for patients to tolerate unsedated placement of an oroduodenal tube for the hour or more required for the test. Fourth, accurate measurement of bicarbonate concentrations or enzyme output may be challenging. False-positive results have been reported in patients who have undergone Billroth II gastrectomy, in patients with diabetes, celiac disease, and cirrhosis, and in patients recovering from a recent attack of acute pancreatitis. A direct pancreatic function test is most useful in patients with presumed chronic pancreatitis in whom easily identifiable structural and functional abnormalities have not been demonstrated on more widely available diagnostic modalities such as CT (e.g., a patient with small-duct disease). This type of test is most useful in ruling out chronic pancreatitis in patients who present with a chronic abdominal pain syndrome suggestive of chronic pancreatitis, saving these patients the label of chronic pancreatitis with its negative repercussions and the risk of such diagnostic tests as ERCP. There are variations of direct pancreatic function tests that may be easier for patients to tolerate (by sedating them) and might be able to be made more widely available. One proposed variation is to collect pancreatic secretions at the time of ERCP by placement of a catheter directly in the pancreatic duct (the so-called intraductal secretin test). This test typically samples pancreatic output for only 15 minutes, to minimize the likelihood of ERCP-induced pancreatitis. It is not standardized and does not appear to be as accurate as standard direct pancreatic function testing,175,187,188 probably because of the rather brief collection time. Another variation of pancreatic function testing is to use sedation and a standard upper endoscope to take the place of the usual oroduodenal tube, with analysis of bicarbonate output using the regular hospital laboratory. This variation attempts to bypass the difficulties limiting the widespread application of standard direct pancreatic function tests, such as passage of the collection tube in unanesthetized patients and need for a dedicated laboratory to measure the bicarbonate concentration by back-titration. Unlike the intraductal secretin test, this endoscopic variation appears to be nearly, although not quite, as accurate as standard direct pancreatic function testing.189-191 The initial descriptions of this test used a 60-minute collection with timed aspirates of duodenal fluid every 15 minutes. This is a long time to keep a patient sedated and an endoscopy room occupied, although it may be possible to shorten the test and only collect samples at 30 and 45 minutes after secretin injection192 with only moderate loss of sensitivity. Another variation of the test is to measure lipase output rather than bicarbonate output, with secretin193 or CCK194 as the secretagogue. These variations appear to be less accurate than standard direct pancreatic function testing. Like traditional pancreatic function testing, endoscopic-based pancreatic function tests have been compared to alternative diagnostic

999

1000

Section VII  Pancreas tests like ERCP,195 with an overall agreement in 86% of patients and a high negative predictive value for endoscopic pancreatic function testing. The value of pancreatic function tests, whether traditional or endoscopic, lies in their high sensitivity and consequent ability to rule out chronic pancreatitis.196

Indirect Tests

The desire to develop indirect tests of pancreatic function is an outgrowth of the complexity, discomfort, and limited availability of direct pancreatic function testing. Indirect tests can generally measure pancreatic enzymes in blood or stool. Tests that measure the effect of pancreatic enzymes on an orally administered substrate with collection of metabolites in blood, breath, or urine (see Chapter 56) are of historical interest only. Serum Trypsinogen Serum trypsinogen (often called serum trypsin) can be measured in blood and provides a rough estimation of pancreatic function. Very low levels of serum trypsinogen (<20 ng/ mL) can be seen in patients with advanced chronic pancreatitis with steatorrhea.197 Serum trypsin is not decreased in patients with other forms of steatorrhea, but low levels of serum trypsinogen may be seen in patients with pancreatic ductal obstruction, including malignant obstruction. The test is available through commercial laboratories. Pancreatic Enzymes in Stool Low concentrations of chymotrypsin or elastase in stool can reflect inadequate delivery of these pancreatic enzymes to the duodenum. Both can be measured on random samples of stool. Fecal chymotrypsin is low in most patients with chronic pancreatitis and steatorrhea.174 Fecal chymotrypsin is available from commercial laboratories. False-positive results have been reported in other malabsorptive conditions (celiac disease, Crohn’s disease), in diarrheal diseases in which the stool is diluted, and in severe malnutrition. Because the test is usually normal in patients without steatorrhea, it is reliably positive only in advanced chronic pancreatitis. Fecal elastase has significant advantages over fecal chymotrypsin in that it is much more stable in passage through stool and is easier to measure. Levels less than 200 µg per gram of stool are considered abnormal. The test is reasonably accurate in more advanced chronic pancreatitis.198-201 Fecal elastase can be low in other diseases causing diarrhea, such as short bowel syndrome and small bowel bacterial overgrowth. Fecal elastase measurement is available through reference laboratories in the United States. Fecal Fat Excretion Maldigestion of fat occurs after 90% of pancreatic lipase secretory capacity is lost. The simplest evaluation of pancreatic lipase action is the measurement of fecal fat excretion during a 72-hour collection of stool. Although theoretically quite simple, the test is difficult to perform in practice. The patient must follow a diet containing 100  g/ day fat for at least three days before the test, and complete collection of the sample is difficult to achieve. In health, less than 7  g of fat (7% of the ingested dose) should be present in stool. Measuring fecal fat requires that the dietary content of fat be known exactly, which is impossible unless the patient is housed in a general clinical research center. A qualitative analysis of fecal fat can also be performed with a Sudan III stain of a random specimen of stool. The finding of more than six globules per high-power field is considered a positive result, but as with fecal fat excretion, the patient

must be ingesting adequate fat to allow measurable steatorrhea. Sudan III staining of stool is positive only in patients with substantial steatorrhea.

TESTS OF PANCREATIC STRUCTURE (IMAGING) Plain Abdominal Radiography

The finding of diffuse (but not focal) pancreatic calcifications on plain abdominal films is quite specific for chronic pancreatitis. Focal calcifications may be seen in cystic and islet cell tumors of the pancreas, and in peripancreatic vascular calcifications. Calcifications occur late in the natural history of chronic pancreatitis and may take from 5 to 25 years to develop.57,58 Calcifications are most common in alcoholic, late-onset idiopathic, hereditary, and tropical pancreatitis and far less common in early-onset idiopathic pancreatitis. Acceleration of the clinical course of chronic pancreatitis and subsequent calcifications are particularly common in patients who smoke.42,43,202,203 Calcifications are not static once they develop and may in fact wax and wane over time.204

Abdominal Ultrasonography

Ultrasonography has been widely studied as a diagnostic tool for chronic pancreatitis.205 This modality is limited in that the pancreas (and particularly the pancreatic head) cannot be adequately visualized in some patients owing to overlying bowel gas or body habitus. Ultrasonographic findings indicative of chronic pancreatitis include dilation of the pancreatic duct, shadowing pancreatic ductal stones, gland atrophy or enlargement, irregular gland margins, pseudocysts, and changes in the parenchymal echotexture (Table 59-4). Most studies suggest a sensitivity of 50% to 80% with a specificity of 80% to 90%.174 The true sensitivity and specificity may be different because most of these studies are older and did not use modern state-of-the-art equipment.

Table 59-4  Grading of Chronic Pancreatitis by Ultrasonography (US) or Computed Tomography (CT) GRADE

US OR CT FINDINGS

Normal

No abnormal findings on a good-quality study visualizing the entire gland One of the following:   Mild dilatation of the pancreatic duct (2-4 mm) in the body of the gland   Gland enlargement ≤2-fold normal One of the preceding findings plus at least one of the following:   Pancreatic duct dilatation (>4 mm)   Pancreatic duct irregularity   Cavities <10 mm   Parenchymal heterogeneity   Increased echogenicity of duct wall   Irregular contour of the head or body   Focal necrosis of parenchyma Mild/moderate features plus one or more of the following:   Cavity >10 mm   Intraductal filling defects   Calculi/pancreatic calcification   Ductal obstruction (stricture)   Severe duct dilatation or irregularity   Contiguous organ invasion

Equivocal

Mild-moderate

Severe

Adapted from Sarner M, Cotton PB: Classification of pancreatitis. Gut 1984; 25:756.

Chapter 59  Chronic Pancreatitis In one more recent study comparing transabdominal ultrasonography with CT, ERCP, and EUS, the accuracy of ultrasonography was 56%.206 In this study, some abnormality (such as changes in parenchymal echotexture) was noted on ultrasonography in 40% of patients who had a normal pancreas as defined by the other diagnostic tests. A large screening study of transabdominal ultrasonography in Japan encompassing 130,000 examinations found increased echogenicity, mild dilation of the pancreatic duct, small cystic cavities, and even ductal calcification in the absence of clinical features of chronic pancreatitis.207 The majority of these abnormalities could not be attributed to chronic pancreatitis and were instead attributed to aging. These studies would suggest that there is a large spectrum of ultrasonographic findings in normal individuals and that it can be difficult to distinguish normal (or age-related) variability from chronic pancreatitis if the visualized changes are mild. Thus, transabdominal ultrasonography is often not useful in the evaluation of patients with suspected chronic pancreatitis. The finding of a normal pancreas or moderate to marked changes of advanced chronic pancreatitis is generally definitive. Mild changes of chronic pancreatitis are less specific and must be interpreted in light of the clinical history and the patient’s age. Ultrasonography can be useful in screening for complications of chronic pancreatitis (e.g., pseudocyst or bile duct obstruction) and in evaluating for other conditions that might mimic the symptoms of chronic pancreatitis (i.e., biliary tract disease).

Computed Tomography

The overall sensitivity of CT for chronic pancreatitis is between 75% and 90%, with a specificity of 85% or more.208-210 CT is able to image the pancreas in essentially all patients and hence has a substantial advantage over ultrasonography. Table 59-4 outlines the diagnostic abnormalities seen on CT in chronic pancreatitis. Most studies of diagnostic CT in chronic pancreatitis have not used stateof-the-art CT technology. One study using a 64 multi-row detector CT demonstrated the ability to accurately image the pancreatic duct and an overall sensitivity or more than 80% and specificity of near 100% when compared with MRCP or ERCP.211 It is almost certain that modern multidetector scanners using a pancreas protocol have better sensitivity than these older studies suggest, although the magnitude of the greater sensitivity is not known. Like all diagnostic tests, CT is most accurate in advanced chronic pancreatitis after substantial structural changes have developed (Fig. 59-4). Although CT is more expensive than ultrasonography and exposes the patient to ionizing radiation, it is more sensitive and more specific.

Magnetic Resonance Imaging

MRI, coupled with MRCP is as accurate, and probably more so, than CT in patients with chronic pancreatitis.208-212 MRCP results agree with ERCP results in about 90% of cases.212-213 Agreement between MRCP and ERCP is less common in areas where the pancreatic duct is small (tail of pancreas and side branches) or when the ductal changes are more subtle.214 Improved visualization of the pancreatic duct can be achieved by administering secretin.212-216 In addition, signal intensity and arterial enhancement ratios can be obtained, using gadolinium as a contrast agent, which may improve the ability to image the gland.212,216 Finally, a qualitative or semi-quantitative assessment of fluid output from the pancreas to the duodenum can be made during MRCP after secretin injection, which may allow additional insights into pancreatic function.217-219 The accuracy of this type of pancreatic function test has been compared with the intra-

Figure 59-4.  Computed tomography scan demonstrating several large, densely calcified stones (arrows) within a markedly dilated pancreatic duct in long-standing “big-duct” chronic pancreatitis.

ductal secretin test217 and with low-sensitivity tests such as fecal elastase,220 but not with a more formal hormonalstimulation pancreatic function test. Some analyses suggest that just measuring volume after secretin stimulation, instead of bicarbonate concentration, is too inaccurate to be clinically useful.221 Advancements in MR image analysis will continue to improve the image quality of MRCP, which in the future will equal ERCP in accuracy. Like ERCP, however, MRCP will be inaccurate in patients without significant ductal abnormalities. Although MRI is widely available, not all centers have the capacity to perform highquality MRCP.

Endoscopic Retrograde Cholangiopancreatography

Pancreatography has been considered the most specific and sensitive test of pancreatic structure, and many consider it the de facto gold standard. It also has the advantage over all previously discussed tests in that therapy (e.g., pancreatic duct stenting or stone extraction) may be administered during its performance. The disadvantage, however, is that ERCP is the riskiest diagnostic test, with complications occurring in at least 5% of patients (in as many as 20% of certain subgroups) and a mortality rate of 0.1% to 0.5%. In most studies in patients with chronic pancreatitis, the sensitivity of ERCP is between 70% and 90%, with a specificity of 80% to 100%.173,174,181-185,222,223 Thus, chronic pancreatitis can exist in the absence of any visible changes within the pancreatic duct.177,222-224 The diagnostic features of chronic pancreatitis on ERCP are listed in Table 59-5. These were developed at an international consensus conference held more than 20 years ago.225 The diagnosis is based on abnormalities seen in the main pancreatic duct and the side branches. ERCP is highly sensitive and specific in patients with advanced disease. The appearance of a massively dilated pancreatic duct with alternating strictures (the chain-of-lakes appearance) is characteristic of the most advanced chronic pancreatitis (Fig. 59-5). Less dramatic pancreatographic changes are less definitive and specific (Fig. 59-6).222,223

1001

1002

Section VII  Pancreas

Figure 59-5.  Endoscopic retrograde pancreatogram showing a markedly dilated pancreatic duct with alternating strictures and dilatation. This “chain-of-lakes” appearance is diagnostic of chronic pancreatitis.

Table 59-5  Cambridge Grading of Chronic Pancreatitis on Endoscopic Retrograde Pancreatography GRADE

MAIN PANCREATIC DUCT

SIDE BRANCHES

Normal Equivocal Mild Moderate Severe

Normal Normal Normal Abnormal Abnormal with at least 1 of the following:   Large cavity (>10 mm)   Obstruction   Filling defects   Severe dilatation or irregularity

Normal <3 Abnormal ≥3 Abnormal ≥3 Abnormal ≥3 Abnormal

Adapted from Axon ATR, Classen M, Cotton PB, et al. Pancreatography in chronic pancreatitis: International definitions. Gut 1984; 25:1107-12.

The accurate interpretation of an ERCP requires a study of adequate quality (filled to the second generation of the side branches and without significant movement artifact) and the capability to obtain radiographic images of high resolution. Many pancreatograms do not meet these criteria for an adequate study.223 An underfilled pancreatic duct can appear to have an irregular duct margin (leading to overdiagnosis of chronic pancreatitis) or might not delineate changes within the inadequately filled side branches (leading to underdiagnosis of chronic pancreatitis). The pancreatic duct abnormalities characteristic of chronic pancreatitis can be seen in other conditions as well. The most common is the effect of aging on the pancreatic duct. Although pancreatic function is well preserved in aging, impressive abnormalities may develop in the pancreatic duct. They include focal or diffuse dilatation of the main pancreatic duct and its side branches, the development of cystic cavities, and even ductal calculi.117,222,223,226,227 In the large screening ultrasonography study mentioned previously, 50% of all calcification and more than 80% of ductal dilation and cystic lesions seen were believed to be

Figure 59-6.  Endoscopic retrograde pancreatogram demonstrating subtle changes limited to the side branches (arrows) in a patient in whom a direct pancreatic function (secretin) test indicated chronic pancreatitis. These subtle findings are generally not sufficient for a definitive diagnosis of chronic pancreatitis.

attri­butable to aging, not chronic pancreatitis.207 Temporary changes in the pancreatic duct may also occur after an episode of acute pancreatitis and may take months to resolve.222,223,225 Pancreatic carcinoma may produce changes within the pancreatic duct that resemble those of chronic pancreatitis. Finally, the placement of pancreatic duct stents can produce new abnormalities within the pancreatic duct that mimic chronic pancreatitis and that may not entirely resolve after stent removal.222,223,228,229 Pancreatic stents are being placed with much more frequency to prevent postERCP pancreatitis. These temporary, very small-caliber stents used for these purposes appear to less commonly produce these ductal changes.230 There is significant potential for substantial interobserver and intraobserver variability in the interpretation of ERCP.222,223 The initial consensus conference225 identified some abnormalities such as a dilated duct, abnormal duct contour, and abnormal side branches but did not define absolute criteria to differentiate normal from abnormal or normal variant. In one study, 74 postmortem pancreatograms were submitted to six experienced endoscopists.116 They were asked to judge whether the pancreatogram demonstrated chronic pancreatitis and the severity of the abnormalities. The pancreas was then examined for histologic correlation. All six endoscopists correctly identified the 5 subjects with chronic pancreatitis. In the remaining 69 subjects, there was no histologic evidence of chronic pancreatitis. Depending on the observer, between 42% and 98% of these pancreatograms were read as demonstrating chronic pancreatitis, largely based on mild abnormalities within the main duct and side branches. The mistaken interpretations were felt to be due to age-related changes within the pancreas. Another study attempted to estimate intraobserver variability by submitting 51 pancreatograms to four expert endoscopists on three separate occasions.231 Each endoscopist was consistent in his or her own three reports in 47% to 95% of cases (yielding a rate of intraobserver variability as high as 53%). Much of the intraobserver and interobserver variability in ERCP evaluations is related to the inter-

Chapter 59  Chronic Pancreatitis Table 59-6  Diagnosis of Chronic Pancreatitis on Endoscopic Ultrasonography Parenchymal abnormalities

Ductal abnormalities

Hyperechoic foci Hyperechoic strands Lobularity of contour Cysts Main duct dilatation Main duct irregularity Hyperechoic ductal walls Visible side branches Calcification

pretation of mild or subtle pancreatographic changes rather than dramatic abnormalities. This is the most substantial clinical problem related to ERCP as a diagnostic tool; subtle or minor abnormalities of the pancreatic duct are quite nonspecific and are not reliable markers of chronic pancreatitis. In the majority of patients, ERCP and direct pancreatic function tests reach similar conclusions. This fact is reassuring because ERCP is widely available and pancreatic function tests are not. When using ERCP, however, astute clinicians should remember that other conditions may mimic the ductal changes of chronic pancreatitis and that subtle changes of ductal contour are quite nonspecific.

Endoscopic Ultrasonography

EUS allows a highly detailed examination of pancreatic parenchyma and the pancreatic duct by overcoming the imaging problems in transabdominal ultrasonography (such as intervening gas in the bowel lumen). The diagnosis of chronic pancreatitis on EUS is based on the presence of abnormalities in the pancreatic duct and the parenchyma (Table 59-6). These features may be individually classified as none, minimal, moderate, or extensive but in practice are generally only graded as present or absent, and the total number of features is used as the score. The sensitivity and specificity of the test is determined by the threshold total score used to define chronic pancreatitis. A large range of threshold scores have been used, ranging from 1 to 6. Most studies have used the presence of 3 or more features to define a positive result.222 EUS has been compared with pancreatic histology in a limited number of patients. One study compared EUS features with histology in 71 patients who underwent surgical therapy for presumed chronic pancreatitis.232 Utilizing a cut-off of more than three EUS criteria, the sensitivity of EUS was 83% and the specificity was 57%. In a subgroup with more advanced histologic evidence of chronic pancreatitis by histology, the sensitivity was 83% and specificity 80%. No single EUS criteria predicted fibrosis better than any other, but there was a general correlation between the number of EUS criteria and the histologic severity of disease. Another study in 42 patients who underwent EUS prior to pancreatic surgery (largely for carcinoma) determined a cut-off of more than four EUS criteria had a sensitivity of 90% and specificity of 86%.233 A study using EUS-guided Tru-cut biopsy of the pancreas234 noted poor agreement between EUS and histology, but the histologic specimens obtained were small and likely to not be representative of the entire gland. EUS and ERCP agree in about 80% of patients.222,223,225-227 EUS has also been compared with standard and intraductal direct pancreatic function testing. The agreement between

these tests varies widely in these studies, ranging from 10% to 90%.222,223,235,238-240 This variability is partly related to the severity of disease. The agreement is best in those with more advanced disease. In one study239 the sensitivity of EUS for advanced chronic pancreatitis (classic findings on ERCP and an abnormal pancreatic function test) was good (sensitivity of 73% and specificity of 81% using more than three criteria), but the sensitivity for less advanced chronic pancreatitis was only 10%. In the majority of instances in which EUS findings disagree with results of other diagnostic tests, it is the EUS findings that are abnormal when the other tests are normal. This speaks to the question of the specificity of EUS. Specificity is determined by the cut-off chosen. Specificity is degraded by other conditions that can mimic the EUS findings of chronic pancreatitis. This may be a substantial conundrum, because EUS changes of chronic pancreatitis are frequently encountered. The edema associated with a recent episode of acute pancreatitis can make duct margins and intralobular septa more apparent, which will reduce specificity.222 Age-related changes can be observed in the pancreas by EUS that mimic the changes of chronic pancreatitis.241 Chronic alcohol and tobacco use can produce similar changes, in the absence of clinical chronic pancreatitis.222,242-244 In one study 39% of patients with unexplained dyspepsia had five or more EUS features of chronic pancreatitis, and 34% of controls had three or more features.245 It is unlikely that chronic pancreatitis is present or will develop in all or even most of these patients. The diagnosis of chronic pancreatitis by EUS is based on the interpretation of sonographic images. There is moderate interobserver variation in this interpretation.246 An international consensus conference met in 2007 to try to develop more reliable EUS criteria for chronic pancreatitis.247 This new system (called the Rosemont criteria) uses major and minor criteria and attempts to provide semiquantification of severity. Whether it is more specific or less prone to inter­ observer variability remains to be determined. What also remains to be determined is whether the presence of three or four (or fewer) EUS features, in the absence of corroborating information from other diagnostic tests, is adequate for a conclusive diagnosis of chronic pancreatitis to be made.196,222,248,249 Resolution of this issue will require very long follow-up of patients, given the lack of a useful gold standard. Other advances in EUS imaging including the use of contrast, digital image analysis, and EUS elastography are just beginning to be studied, but may improve specificity in the future. EUS is highly accurate in patients with advanced chronic pancreatitis (Fig. 59-7). An EUS score of greater than 5 is highly specific for chronic pancreatitis. A completely normal or near normal EUS (0 to 2 criteria) essentially rules out chronic pancreatitis.222,248,249 EUS scores of 3 or 4 should be considered indeterminate and should be interpreted in light of the patients clinical features and with recognition of the effect of alcohol, tobacco, age, and other associated conditions (diabetes, chronic kidney disease) that can mimic the changes of chronic pancreatitis.

DIAGNOSTIC STRATEGY

The diagnosis of chronic pancreatitis is most often suspected due to the presence of an abdominal pain syndrome, and less commonly due to a suspicion of exocrine (diarrhea, steatorrhea, weight loss) or endocrine insufficiency (diabetes mellitus). In the subgroup with suspected exocrine or endocrine insufficiency, the disease is likely to be longstanding, and most available diagnostic tests will be able to detect this relatively far-advanced disease. In this situation,

1003

1004

Section VII  Pancreas patients with substantial abnormalities of pancreatic structure (particularly a dilated pancreatic duct). Identifying patients with big-duct chronic pancreatitis and differentiating them from patients who lack these findings is very helpful in choosing therapy for pain.

Medical Therapy

Several modalities are available to help control pain.

Figure 59-7.  An endoscopic ultrasonographic image of the pancreatic body in a patient with chronic pancreatitis. The markers on the dilated pancreatic duct demonstrate hyperechoic margins, one of the diagnostic features of chronic pancreatitis. The parenchyma surrounding these markers demonstrates hyperechoic strands and foci, additional features of chronic pancreatitis.

simple and straightforward tests like serum trypsinogen, fecal elastase, or abdominal imaging with US or CT are reasonable and likely to be diagnostic. In those with pain as the primary complaint, a high-quality CT scan using multidetector technology and a pancreatic protocol or an MRI with MRCP are probably the best initial diagnostic tests. Many of these patients may have already had a routine abdominal CT (often through an emergency department) but have not had a high-quality CT (or MRI with MRCP), so it may be worthwhile to repeat them if the previous studies are of low quality. If the CT or MRI does not demonstrate chronic pancreatitis, then tests of maximum sensitivity should be used next. Hormonal stimulation direct pancreatic function testing, if available, could logically be used next if CT findings are nondiagnostic. If this type of testing (including the new variations using endoscopic-based methods) is not available, EUS should be considered the best alternative. ERCP should be used very rarely for diagnostic purposes, but if used great care should be taken to avoid overinterpretation of subtle pancreatographic changes seen on ERCP.

TREATMENT ABDOMINAL PAIN

Pain is the most common and most debilitating symptom of chronic pancreatitis as well as the one most often requiring medical care. The initial evaluation of pain should focus on identifying associated conditions for which specific therapy exists. These conditions can include pancreatic pseudocyst, duodenal (and possibly bile duct) compression, superimposed pancreatic carcinoma, and gastroparesis. In addition to identifying (or suggesting) the presence of specifically treatable complications, CT can be very useful in identifying

Analgesics The majority of patients with chronic pancreatitis require some form of analgesia. Some patients’ pain may be managed with acetaminophen or aspirin, but most require more potent narcotic agents. There is a risk of addiction to narcotics with the use of these agents. The risk of addiction is not defined but is estimated to be about 10% to 30%. The introduction of high-dose extended-release oral opioids has created a higher risk of both dependence and diversion of medications for illicit use. Nonetheless, narcotics should not be withheld in patients with severe pain out of a concern for possible addiction, because pain relief is the first priority. Strategies to minimize the risk of overuse of narcotics and addiction include having a single physician take responsibility for prescriptions, ongoing counseling, and regular clinic visits. The use of a pain clinic may or may not be helpful. Some pain management clinics treat all patients in the same way, increasing dosage and potency of narcotics until pain is relieved. Although this strategy may be useful in managing pancreatic carcinoma pain, it typically leads only to greater dependence in patients with chronic pancreatitis. Pain management clinics that focus on non-narcotic approaches are most useful in patients with chronic pancreatitis. If nonarcotic agents fail, as they often do, it is useful to begin with the least potent opioid agents, including such agents as the combination of 100 mg propoxyphene napsylate with 650 mg acetaminophen (Darvocet-N 100) or tramadol (Ultram). Tramadol is a dual-action analgesic, with mu-opioid agonistic and monoaminergic properties. High dosages of tramadol are equivalent to oral morphine in treating chronic pancreatitis, with fewer effects on gut motility.250 More potent narcotics are required in many patients. In this situation it is useful to gradually increase the dosage or potency while focusing the patient on the goal of control of pain to an acceptable level rather than complete relief of pain. Adjunctive agents can also be considered in patients who need more potent narcotics. Many of these patients are depressed, and coexistent depression lowers pain threshold. Chronic pancreatic pain can lead to abnormal spinal cord gating in nociceptive neurons, a centrally sensitized pain state, with hyperalgesia and allodynia. Tricyclic antidepressants can be useful adjuncts, not only by treating depression and potentially modulating central pain perception but also because they have direct effects on pain and potentiate the effect of narcotics. Other antidepressants such as serotonin reuptake inhibitors (SSRIs) and combined serotonin and norepinephrine reuptake inhibitors (e.g., duloxetine) may also have this effect. Finally, the a2d subunit voltage-gated N-type calcium channel inhibitors (gabapentin and pregabalin) are used in a variety of chronic pain states as an adjunct to narcotics. None of these adjunctive agents has been rigorously studied in patients with chronic pancreatitis but they are commonly used in patients with continued severe pain from chronic pancreatitis, in an attempt to improve pain control and minimize the dosage and potency of narcotics. Given the increasing recognition of neuropathic mechanisms of pain in chronic pancrea­ titis,251,252 using these types of agents appears reasonable.

Chapter 59  Chronic Pancreatitis Cessation of Alcohol and Tobacco Continued alcohol abuse hastens the development of pancreatic dysfunction, although even complete abstinence does not prevent progression.77 Also, continued alcohol abuse, along with smoking, increases mortality.20 There are therefore good reasons to encourage patients to stop drinking and smoking separate from any effect on abdominal pain. Most studies, but not all, have documented an apparent decrease in pain or painful relapses in patients who stop drinking alcohol. In a summary of these studies, pain continued in 26% of abstinent patients, compared with 53% of those who continued to drink.253 In a study of the natural history of chronic pancreatitis, continued alcohol abuse was associated with a higher risk of painful relapses.254 These data support that abstinence has some beneficial effect on pain, but the magnitude of the effect is probably only modest. Antioxidants Damage by free radicals has been proposed as one mechanism for pancreatic damage in alcoholic and other forms of chronic pancreatitis. Patients with chronic pancreatitis (particularly alcoholic) have evidence of oxidant stress and reduced antioxidant capacity.255 Oxidant stress is a strong activator of pancreatic stellate cells.29,37 There are now several small randomized trials of a mixture of antioxidants (selenium, beta-carotene, vitamin C, vitamin E, and methionine) that indicate that this therapy reduces the pain of chronic pancreatitis.256-259 The overall effect of antioxidants is very modest but the data now support their use in these patients and the therapy is risk-free. Pancreatic Enzyme Therapy Pancreatic secretion is under feedback control (see Chapter 56). The use of pancreatic enzymes to reduce pain is based on the ability of these agents to activate this feedback control system in a way to reduce pancreatic secretion. Delivering proteases to the duodenum or very proximal jejunum can suppress pancreatic secretion. This action is due to the ability of the proteases in this segment of small bowel to reduce CCK release, by destroying an intestinal CCKreleasing factor, which is one of the primary stimulants of CCK release. In patients with chronic pancreatitis, the lack of delivery of serine proteases to the duodenum could allow

more CCK-releasing factor to escape denaturing. As a result, one would expect higher levels of CCK-releasing factor within the duodenum and higher serum levels of CCK. Higher levels of circulating CCK would stimulate the pancreas to secrete, with this strong stimulation leading to pancreatic pain by raising pancreatic duct or tissue pressure or by forcing digestive enzymes into the interstitium if secretion is occurring against pancreatic ductal obstruction. The oral administration of pancreatic enzymes could restore normal feedback suppression of pancreatic secretion by providing active serine proteases in the duodenum, which could again denature the CCK-releasing factor, thereby reducing the hyperstimulation and relieving pain. As described in Chapter 56, pancreatic secretion of volume and bicarbonate is not controlled by the presence of proteases within the duodenum. Also, pancreatic secretion is under humoral as well as neural control. Therefore, suppressing pancreatic enzyme release by administering oral enzyme supplements is not likely to produce complete suppression of secretion, and the magnitude of the effect on secretion could vary from patient to patient. The presence of this feedback control system, which can control pancreatic enzyme secretion, is well documented in humans without chronic pancreatitis as well as in some patients with chronic pancreatitis.260 One marker of this disordered feedback system might be elevations in serum CCK in patients with chronic pancreatitis, particularly those with pain. Elevations in serum CCK are seen in only some patients with chronic pancreatitis. It is likely that this disordered feedback, as in all presumed causes of pain, is only important in a subgroup of patients. Thus far six randomized, prospective, double-blind trials have attempted to delineate the effectiveness of orally administered pancreatic enzymes to decrease pain in patients with chronic pancreatitis. Two studies using enzymes in non–enteric-coated (tablet) form reported a benefit.261,262 Four other studies using enteric-coated microsphere preparations showed no benefit.263-266 The difference between these studies may reflect patient selection but may also reflect the different choice of enzyme preparations (Table 59-7). The feedback-sensitive part of the small bowel appears to be the most proximal portion, and enteric-coated preparations may not release the majority of their proteases until they reach the more distal small bowel. Non–enteric-

Table 59-7  Enzyme Products for the Treatment of Chronic Pancreatitis

Nonenteric Coated* Viokase 8 Viokase 16 Enteric Coated Creon 5 Creon 10 Creon 20 Ultrase MT 12 Ultrase MT 18 Ultrase MT 20 Pancrease MT 10 Pancrease MT 16 Pancrease MT 20

LIPASE CONTENT (USP UNITS)

DOSe TO TREAT PAIN

DOSE TO TREAT MALDIGESTION AND STEATORRHEA†

8,000 16,000

8 with meals and at night 4 with meals and at night

8-12 with meals 4-6 with meals

5,000 10,000 20,000 12,000 18,000 20,000 10,000 16,000 20,000

NA NA NA NA NA NA NA NA NA

12-18 with meals 6-9 with meals 3-4 with meals 5-8 with meals 4-6 with meals 3-4 with meals 6-9 with meals 4-6 with meals 3-4 with meals

*These agents require co-treatment with a histamine-2 receptor antagonist or proton pump inhibitor. † These doses are for average-sized meals. One third this amount is usually taken with snacks. The dosage should be split equally immediately before the meal, during the meal, and immediately after the meal. 3 USP units ≅ 1 international unit (IU). NA, not applicable.

1005

1006

Section VII  Pancreas coated enzymes might therefore be needed for adequate delivery of serine proteases to the duodenum. A meta-­ analysis combining all these studies concluded that enzymes do not reduce pain,267 but the combined analysis methodology may not be appropriate in this instance given the proposed mechanism just discussed. The randomized trials assessing the effectiveness of pancreatic enzymes for pain did not include a large number of patients. Relatively high dosages of non–enteric-coated enzymes were used in the trials demonstrating pain reduction (equivalent to 64,000 USP units of lipase with meals and at night, which has translated to four to eight pills four times daily). Because these non–enteric-coated enzymes can be inactivated by gastric acid, the concomitant use of an agent to suppress gastric acid or neutralize acid is required. In the two studies that demonstrated effectiveness, patients with less advanced disease (small-duct chronic pancreatitis without steatorrhea), women, and patients with idiopathic chronic pancreatitis had the best response. These agents are safe and are often worth a trial in these types of patients. A trial of enzymes for pain is rarely successful in those with advanced or big-duct chronic pancreatitis (mainly, advanced alcoholic chronic pancreatitis). Octreotide Octreotide, the synthetic analog of the native hormone somatostatin, decreases pancreatic secretion and reduces circulating CCK levels. This agent therefore might reduce pain via the same mechanisms invoked for the use of enzymes for pain. In addition, octreotide has some direct antinociceptive effect separate from any effect on pancreatic enzyme secretion. Several small placebo-controlled studies have been performed to assess the role of octreotide and have reached different conclusions. The largest of these trials268 demonstrated that the rate of placebo response in patients with chronic pancreatitis can average 35% to 40%, serving to emphasize the importance of placebo-controlled trials.

Endoscopic Therapy

As is discussed in Chapter 61, the general goal of endoscopic therapy is to improve drainage of the pancreatic duct by relieving ductal obstruction. The potential application of endoscopic therapy is limited to a subgroup of patients with amenable pancreatic ductal anatomy. These are generally patients with big-duct chronic pancreatitis and advanced structural abnormalities of the pancreatic duct. For the most part, this has meant the patient with a dilated pancreatic duct who also has a single dominant stricture or an obstructing stone in the head of the pancreas, with dilation of the pancreatic duct upstream of the dilatation (see Fig. 61-6). It is important to remember, however, that some patients with pancreatic duct dilation have little or no clinical symptoms and that a dilated duct does not necessarily correlate with elevations in duct pressure. Strictures and calculi in the upstream body or tail of the gland are not generally amenable to endoscopic therapy. Specific endoscopic therapies that have been studied are pancreatic sphincterotomy, stent placement, and stone extraction. The individual contribution of each of these therapies is difficult to quantify because they are usually performed together. Pancreatic Duct Sphincterotomy Pancreatic duct sphincterotomy is generally required for larger-caliber pancreatic stent placement and for pancreatic duct stone extraction. This may be performed with a pulltype sphincterotome or with a needle-knife sphincterotome over a small-caliber pancreatic duct stent. Both techniques are used depending on local preferences, but one random-

ized trial269 suggests the needle-knife technique may be safer in patients at high risk for post-ERCP pancreatitis (see Chapters 58 and 61). Major papilla pancreatic sphincterotomy alone as a therapy would be applicable only in patients in whom long-standing cicatricial stenosis of the sphincter has produced obstructive chronic pancreatitis, a form of chronic pancreatitis that is exceedingly rare. Whether sphincter of Oddi dysfunction might also predispose to chronic pancreatitis is controversial,109,110 and at the moment there are no data to support the performance of sphincter of Oddi manometry and manometry-guided therapy in patients with chronic pancreatitis.270 Very rarely, patients with pancreas divisum present with marked upstream dilation of the dorsal pancreatic duct and chronic pancreatitis. Sphincterotomy of the minor papilla over a stent in this setting may be useful, but minor papilla sphincterotomy for chronic pain in the absence of pancreatic ductal dilation is ineffective.108,271 Stent Placement Stent placement in the pancreatic duct, discussed in detail in Chapter 61, is most often performed to dilate and bypass an obstructing stricture (Fig. 59-8). A number of retrospective case series of stent therapy from expert centers report endoscopic success (successful placement of the stent) in close to 90% and pain improvement in about one half to two thirds of patients.272,273 The high rates of endoscopic success reflect the high degree of expertise from these centers and the careful patient selection. In the largest multicenter report involving more than 1000 patients, 57% of patients with a single dominant stricture in the head of the pancreas who underwent stenting had significant improvement in pain at a mean follow-up of 4.9 years, with an additional 19% noting significant pain improvement but still requiring ongoing endoscopic therapy.274 Complications of stent therapy occur in about 20% of patients, with a mortality rate of 0.6%.272,273 The most commonly reported complications are clogging of stents (producing recurrent pain, attacks of acute pancreatitis, or pancreatic sepsis), stent migration (which may require surgical extraction), and ductal perforation. New stent-induced strictures of the pancreatic duct occur in these patients, but are generally not of clinical significance (unlike those who develop these strictures in a normal preexisting duct).228,229,274 One might assume that patients with pain relief after stent placement would be those with high pancreatic duct pressures and that stent therapy reduced this pressure. In one study that measured pain relief and pancreatic duct pressure after stenting, three of nine patients with normal pressure at the end of the stenting period still had pain, whereas none of four patients with continued high pressure in the pancreatic duct still had pain.138 In another study, pain improvement was similar whether pancreatic duct diameter (a surrogate marker for a decrease in pancreatic duct pressure) decreased or not after stenting and pain relief was not affected by occlusion of the stent.139 It is not clear that the response to stent therapy is predictable from measurements of intraductal pressure, and the mechanism of pain relief requires further study. The decision to remove the stents entirely is therefore most often based on symptoms rather than ductographic features. Even in expert centers, about one in four patients requires surgery for failure of endoscopic therapy.274 Symptoms recur in one third to one half of patients after an initial clinical response. Pancreatic Duct Stone Removal The endoscopic removal of pancreatic duct stones can be difficult and is possible in only a subset of patients. Multiple stones, stones in the body and tail of the gland, stones

Chapter 59  Chronic Pancreatitis

A

B

Figure 59-8.  Endoscopic retrograde cholangiopancreatogram in a patient with chronic pancreatitis. A, A dilated pancreatic duct with a single dominant stricture in the head of the pancreas and upstream dilatation are seen. B, A stent has been placed across the stricture.

in side branches, impacted stones, or stones behind a tight pancreatic duct stricture are generally not manageable with endoscopic techniques. The removal of large or impacted stones often requires lithotripsy, using extracorporeal shock-wave lithotripsy (ESWL) or intraductal instruments. There is no close correlation between the presence of pancreatic duct stones and pain, and many patients with pancreatic ductal stones have no pain. Most retrospective case series report success rates in carefully selected patients in whom endoscopic stone extraction seems feasible. In an analysis of 17 published studies encompassing 588 patients, duct clearance ranged from 37% to 100%,275 with an average of about 60%. A survey from Japan in 555 patients reported complete stone clearance in 73% of patients.276 Pain improvement or relief is seen in about 75% to 90% of patients who undergo this type of therapy.273,275-278 The rate of symptom improvement is thus greater than the rate of complete stone clearance. In a randomized trial comparing ESWL alone to ESWL followed by endoscopic removal of stones, pain relief was similar in both groups.279 In this trial, pain relapse during two years of follow-up was 38% in the ESWL arm and 45% in the ESWL plus endoscopy arm. Treatment costs were three times higher in the ESWL plus endoscopy group. These data suggest that ESWL may be able to reduce the size of stones to the point that they are not obstructing, or that there is some other effect on pancreatic pain separate from the ability to fragment pancreatic stones. Complications of lithotripsy are infrequent, occurring in less than 10%.275-279 Combined Endoscopic Therapy Although the endoscopic therapies just discussed were presented as separate endoscopic techniques, a combination of these therapies is usually needed to manage patients with chronic pancreatitis (see Chapter 61). In large case series, endoscopic therapy is successful in 65% of patients using an intention-to-treat analysis.272,274 There are now two randomized trials comparing endoscopic with surgical therapy for chronic pancreatitis. One trial randomized 72 patients to endoscopic therapy (pancreatic sphincterotomy, stent therapy, or stone removal) or surgical therapy (pancreatic duct drainage or pancreatic resection).280 At one year of follow-up, rates of pain relief were similar. However, at five

years of follow-up, partial pain relief or absence of pain was seen in 86% of the surgical group and 61% of the endoscopic group. In addition, the surgical group had gained more weight; rates of diabetes were similar. This trial has been criticized in that the endoscopic therapy may have been less aggressive than optimal (some patients only underwent pancreatic duct sphincterotomy) and the surgical therapy was more aggressive than might be typical (80% underwent pancreatic resection). A second randomized trial compared a more aggressive endoscopic approach (including ESWL as needed) to a more routine surgical procedure (pancreaticojejunostomy or modified Puestow operation).281 This trial was stopped early when only 39 patients were entered, due to better outcome in the surgical group. At a median follow-up of two years, patients randomized to surgery had a lower pain score and better physical health on quality of life measurement. Complete or partial pain relief was seen in 32% of the endoscopic group and 75% of the surgical group. These data suggest that surgical therapy is more effective and more durable than endoscopic therapy, a fact that should be discussed with patients who are considering these options. Many patients still choose endoscopic therapy because of a desire to avoid surgery. Only a subset of patients with chronic pancreatitis and specific ductal anatomy are candidates for endoscopic therapy. These therapies should be considered only in patients with amenable anatomy and only in centers with substantial expertise in these techniques. The endoscopic treatment of complications such as bile duct strictures and pseudocysts is discussed following, as well as in Chapters 61 and 70.

Surgical Therapy

Surgical therapy in chronic pancreatitis is most commonly considered for intractable abdominal pain for which medical therapy has failed. Other indications for surgery in these patients are complications involving adjacent organs or structures (duodenal, splenic venous, or biliary complications), failure of endoscopic or radiologic management for pseudocysts, internal pancreatic fistulas, and exclusion of malignancy despite an extensive evaluation. The surgical approaches for these complications of chronic pancreatitis are discussed later. Surgical options for pain are pancreatic

1007

1008

Section VII  Pancreas ductal drainage, resection of all or part of the pancreas, and both. The choice of surgical procedure depends in large part on the ductal anatomy, presumed pathogenesis of pain, and associated complications as well as local surgical preferences and expertise.282-285 Ductal drainage procedures are the least technically demanding and preserve the most pancreatic parenchyma. The rationale for these procedures is to relieve ductal obstruction and reduce pancreatic pressures, thereby relieving pain. Pancreatic ductal drainage procedures generally require dilation of the pancreatic duct to more than 6 to 7 mm, a diameter that allows relatively easy identification and anastomosis. This operation is therefore considered in patients with big-duct chronic pancreatitis. The most commonly performed procedure is the lateral pancreaticojejunostomy or Partington-Rochelle modification of the Puestow procedure. In this procedure the pancreatic duct is opened longitudinally and anastomosed to a defunctionalized limb of small bowel, which is connected with a Roux-en-Y anastomosis. This limb also can be used to decompress any coexisting pseudocysts. At the time of the operation, ductal strictures can be incised and stones present can be readily removed as needed. The procedure also can be performed in the absence of a dilated pancreatic duct (normal duct Puestow procedure), but the efficacy for relieving pain is believed to be less. The procedure can be performed laparoscopically, although experience with this technique is limited. The operative mortality for a modified Puestow procedure is extremely low.284-286 No randomized trials comparing a modified Puestow procedure with other surgical therapies have been conducted. Immediate pain relief is seen in approximately 80% of patients.282-286 With long-term follow-up, only half continue to experience pain relief. The explanation for this decline in effectiveness is unknown but may reflect closure of the anastomosis, pain originating in the undrained segments of the head of the pancreas, or the development of other sources of pain (neural inflammation, central nervous system sensitization, duodenal or bile duct obstruction, etc.). There is thus a tradeoff between the simplicity and low risk of this procedure and the gradual deterioration of results over time. Exocrine and endocrine functions are generally unaffected by this surgical procedure per se but appear to continue to deteriorate as in unoperated patients. In an attempt to overcome the modest early and sub­ stantial late failure rates of simple drainage procedures, approaches combining resection of the pancreas with drainage of the pancreatic duct have been developed. These have focused particularly on the head of the pancreas because this is felt to be the pacemaker of the disease by many surgeons. A routine longitudinal pancreaticojejunostomy does not completely decompress the ducts in the head of the gland, the duct of Santorini, and the small ducts draining the uncinate process. Some patients may have an associated inflammatory mass of the head of the pancreas, making drainage of the pancreatic duct within the head of the pancreas more difficult. Options to deal with this problem include resection of the head of the pancreas (pancreaticoduodenectomy [Whipple operation], duodenum-preserving Whipple operation, or duodenum-preserving pancreatic head resections) and combinations of ductal drainage with local resection of all or part of the pancreatic head. It should be noted that improved pain relief after these surgical procedures involving pancreatic resection may be partially explained by the denervation of visceral pancreatic afferent nerves during more extensive dissection rather than better drainage of the pancreatic ducts in the head of the pancreas.

Whipple resection or duodenum-preserving Whipple resection produces pain relief in 65% to 95% of patients.282-286 Whipple operations are generally considered in patients with disease limited to the head of the pancreas, particularly those with a large inflammatory mass of the pancreas in whom malignancy is also being considered. Associated biliary or duodenal obstruction, seen more commonly in these patients with inflammatory masses of the head of the pancreas, can also be treated at the time of the resection. These operations have higher morbidity and mortality than simple ductal drainage operations. Although the mortality in high-volume centers is less than 3%, early postoperative complications (primarily disruptions of normal motility and pancreatic duct leaks) can occur in up to half of cases.286 Surgical mortality is significantly higher if the inflammatory mass occludes or compresses major arteries or veins.287,288 Three procedures have been developed to resect all or part of the head of the pancreas without the disruptions of gastrointestinal (GI) physiology seen with traditional Whipple operations and to limit the amount of pancreatic tissue removed, hopefully better preserving exocrine and endocrine function. The duodenum-preserving pancreatic head resection (DPPHR), developed by Beger, is performed by resecting the pancreatic head but sparing the duodenum, and covering the site with a defunctionalized Roux-en-Y jejunal limb to allow drainage of pancreatic and biliary secretions.285,289 Modifications of this procedure were subsequently developed to avoid dissecting around the portal and superior mesenteric veins (and the associated bleeding risk) and to limit the amount of pancreatic tissue (in particular islet cells) that is removed. In one modification, developed by Frey, less of the head of the pancreas is cored out, leaving the bile duct and peripancreatic vessels undisturbed.285,290 This approach is coupled with a longitudinal incision of the pancreatic duct in the body and tail of the pancreas and the overlaying of a long jejunal anastomosis covering both the opened duct and the cored-out head. A third operation, termed the Berne procedure, uses a pancreatic head resection without longitudinal duct incision, but leaves a narrow layer of pancreatic tissue against the duodenum and retropancreatic vessels.287,291 There have been four randomized trials comparing one of the Whipple operations with DPPHR (either the Frey or Beger procedure).285,292-294 In short-term follow-up, these procedures appear to have equivalent efficacy in relieving pain, with more diabetes seen in those undergoing Whipple procedures. In long-term follow-up this advantage of a DPPHR is lost.294 Randomized trials comparing the Beger with the Frey operations show similar rates of postoperative complications, efficacy, and long-term quality of life.285,293 Postoperative complications are more common than with a simple modified Puestow procedure, but both short- and long-term pain relief is superior. In the United States not many surgeons are trained in these variations of DPPHR, and the modified Puestow is done most commonly. More substantial pancreatic resection is rarely performed. In some patients with disease limited to the body and tail of the pancreas, typically after trauma to the pancreatic duct in the body of the pancreas with upstream obstructive chronic pancreatitis, resection of the body and tail may be considered. In patients with a nondilated pancreatic duct, a small V-shaped excision can be performed on the ventral surface of the pancreas over the pancreatic duct with an overlying pancreaticojejunostomy (similar to a modified Puestow operation) with acceptable results.295 Total or neartotal pancreatectomy has been only rarely performed in the past owing to unacceptable complications of severe brittle diabetes. Later studies using concomitant islet cell auto-

Chapter 59  Chronic Pancreatitis transplantation note insulin independence overall in about 40% of patients, with pain relief in 80% to 90% of patients. The risk of postoperative diabetes is dependent on the yield of islet cells at the time of the pancreatectomy. Islet yields are reduced in those patients with previous pancreatic surgery.296,297 At the moment, total pancreatectomy should be regarded as a salvage operation for a small number of patients with overwhelming pain in whom all other options have failed. The complications occurring after surgery for chronic pancreatitis vary with the operation chosen. They include pancreatic fistula, wound infection, delayed gastric emptying, intra-abdominal abscess, pancreatitis, cholangitis, and bile leak.282-286 The preoperative and perioperative uses of octreotide may reduce the risk of these postoperative complications, particularly pancreatic fistula. In the evaluation of patients who have undergone surgery for chronic pancreatitis, it is important to remember that exocrine insufficiency and endocrine insufficiency can develop as a consequence of the surgery as well as the ongoing disease process. Exocrine insufficiency in particular may escape detection because symptoms may be mild. Steatorrhea can develop in 30% to 40% of patients undergoing simple drainage procedures and in up to two thirds of those undergoing pancreatic resections.282-286 The use of pancreatic enzyme supplements after pancreatic surgery leads to better absorption of nutrients and should be considered for most (or all) patients after surgery for chronic pancreatitis. The development of endocrine insufficiency after pancreatic surgery is also common but not invariable, and some series have even noted improvements in glucose tolerance in some patients after surgery. In general, however, diabetes mellitus still commonly occurs after surgery either as a consequence of pancreatic resection or from the ongoing ravages of the disease.

Nerve Blocks and Neurolysis

The celiac plexus transmits visceral afferent impulses from the upper abdominal organs, including the pancreas. The greater, lesser, and least splanchnic nerves travel from the celiac plexus and then pass through the diaphragm to reach the spinal cord. Attempts to block the transmission of nociceptive stimuli have met with limited success. Celiac plexus block is used rarely in patients with chronic pancreatitis owing to the short duration of action. Celiac plexus block (usually using a combination of a glucocorticoid and a longacting local anesthetic like bupivacaine) and celiac plexus neurolysis (using an injection of absolute alcohol) can be administered by CT- or EUS-guided techniques (see Chapter 61). Celiac plexus block under EUS guidance is safer, more effective, and more long-lasting than that delivered under CT guidance,298-300 but the effect of even EUS-guided celiac plexus block appears to be too transitory for long-term management. Celiac plexus neurolysis has generally been used only in pancreatic carcinoma, although better methods of identifying the celiac ganglia on EUS may allow better targeted injections. There are too few patients with chronic pancreatitis who have undergone celiac plexus neurolysis to be able to judge efficacy.301 EUS-guided celiac plexus block can be considered in occasional patients with intractable pain requiring increasing dosages of narcotics in whom no other therapy is possible as a way to at least temporarily halt an upward spiral of narcotic use. Interfering with nerve transmission through the splanchnic nerves can also block central perception of nociceptive inputs. This generally involves sectioning the greater splanchnic nerve on one or both sides. Thoracotomy was used in the past for this procedure, but most recently it is

performed through a thoracoscopic approach. This can be done unilaterally or bilaterally. Pain relief after thoracoscopic splanchnicectomy averages about 50% to 75% at one year and drops to 25% to 50% with longer follow-up. The lack of complete response might be explained by the multiple spinal levels that receive input from the splanchnic nerves and the tremendous variation in the number of splanchnic roots, which makes complete neurotomy difficult. Significant complications are rare.302,303 This therapy is an alternative in patients for whom medical therapy has failed and particularly in those with small-duct chronic pancreatitis in whom surgical ductal drainage procedures are not possible. Another approach to minimizing nociception focuses on the central nervous system. This has included both spinal cord stimulation304 and transcranial magnetic stimulation of pain centers in the brain.305,306 These are novel approaches but their overall effectiveness remains to be determined.

MALDIGESTION AND STEATORRHEA

As noted, although patients with chronic pancreatitis may maldigest fat, protein, and carbohydrates, it is the fat maldigestion that is the principal clinical problem. It has been estimated that 30,000  IU (or about 90,000 USP units) of lipase delivered to the intestine with each meal should be sufficient to eliminate steatorrhea.307 This corresponds to approximately 10% of the normal pancreatic output of lipase.152 Lower dosages of enzymes can improve but not completely correct steatorrhea. It would seem relatively straightforward to achieve this goal with the use of enzyme supplements, but a number of factors limit the effectiveness of commercially available enzyme supplements. Pancreatic enzyme supplements vary in enzyme content. The lipase content of commercially available preparations is usually described in USP units rather than IU, and ranges from 4000 to 20,000 USP units of lipase per pill or tablet. The actual amount of lipase in a given preparation may vary among lots and may be substantially different from the labeled activity. This, along with other regulatory concerns, has led the U.S. Food and Drug Administration (FDA) to require proof of efficacy and safety for these products. The mandated evidence is substantial and may require costly new studies, which means that generic forms of pancreatic enzymes will cease to be available. Much of the lipase may not reach the small bowel in an active form, being denatured by gastric acid or destroyed by proteases. Most commercially available enteric-coated enzyme preparations use a microsphere size that is too big to empty from the stomach in synchrony with the food. These enteric-coated microspheres may also not release their enzyme contents until they reach the distal jejunum or ileum, too distal for efficient digestion and absorption. Finally, the enzyme preparations are of relatively low potency, so many pills or tablets must be taken with each meal and snack. This requirement can have a major negative influence on compliance. These factors can all interfere with the effective treatment of steatorrhea. The goal of managing steatorrhea is to administer 30,000 IU (90,000 USP units) of lipase in the prandial and postprandial portions of each meal. Table 59-7 estimates the dosages needed to treat steatorrhea. If non–enteric-coated preparations are chosen, concomitant suppression of gastric acid with a histamine-2 (H2) receptor antagonist or proton pump inhibitor is necessary. The effectiveness of enzyme supplementation is generally gauged by clinical parameters, including improvement in stool consistency, loss of visible fat in the stool, and gain in body weight. Performing a 72-hour fecal fat analysis before the start of and during therapy, to

1009

1010

Section VII  Pancreas prove effectiveness, is rarely needed but can be considered in those who do not show the expected response. It is important to periodically evaluate for deficiencies of fat-soluble vitamins, particularly vitamin D, and to assess for the presence of osteopenia or osteoporosis with a bone mineral density test.153-159 Appropriate enzyme therapy improves nutritional status, body weight, and quality of life.308 There are several explanations for failure of enzyme therapy for steatorrhea. The most common is inadequate dose, generally due to patient noncompliance with the number of pills that must be taken. Changing to a more potent preparation to reduce the number of pills taken is often helpful. It is also appropriate to make sure that acid suppression has been prescribed and is being used by patients on non–enteric-coated preparations. The enzymes should be taken spread out over the course of the meal. If the desired effect is not achieved, having the patient eat more frequent, smaller meals may be helpful. It is occasionally useful to change from one formulation to another (e.g., changing from enteric-coated preparations to a combination of a non–enteric-coated preparation plus an agent to suppress acid) or to raise the dose higher than 90,000 USP units of lipase per meal if the response is still not satisfactory. If all these measures fail to achieve the desired effect, it is appropriate to search for alternative diagnoses that could also produce malabsorption, such as celiac disease and SBBO (discussed in Chapter 102), which may be a particular problem in these patients.309 The mechanism of SBBO in these patients is unknown but is likely related to abnormalities in small bowel motility (induced by the disease or by narcotic analgesics), the common use of proton pump inhibitor therapy, which facilitates bacterial overgrowth in the stomach, and possibly a decrease in the bactericidal capacity of pancreatic juice. Finally, if all these measures fail, one can replace dietary fat with medium-chain triglycerides, which do not require lipolysis (and hence lipase) for absorption. In light of the FDA requirement for efficacy studies and the expected absence of generic products, a series of new enzyme products utilizing more effective lipase products of fungal or bacterial origin or bioengineered human lipases are on the horizon.

DIABETES MELLITUS

Diabetes mellitus is an independent predictor of mortality in patients with chronic pancreatitis. Morbidity and mortality due to diabetes mellitus may occur from progressive microangiopathic complications or from more dramatic complications, such as treatment-induced hypoglycemia (in those with inadequate glucagon reserve). Ketoacidosis is distinctly unusual. Given the risk of treatment-induced hypoglycemia and the difficulty of close follow-up in patients who continue to abuse alcohol, therapy is usually directed at controlling urinary losses of glucose rather than on tight control of blood glucose value. Some patients show response to the use of an oral hypoglycemic, such as a sulfonylurea, a thiazolidinedione, or metformin. Insulin is often needed, however, and patients with chronic pancreatitis tend to have lower insulin requirements than patients with type 1 diabetes mellitus. Overvigorous attempts at tight control of blood glucose value are often associated with disastrous complications of treatment-induced hypoglycemia.167 Attempts at tight control of blood glucose value are indicated in one subgroup, however—patients with hyperlipidemic pancreatitis—in whom the diabetes is usually a primary illness and tight control of blood glucose makes control of serum lipids possible.111 In long-standing diabetes, appropriate monitoring for nephropathy, retinopathy, and neuropathy is indicated.

COMPLICATIONS PSEUDOCYST

Pseudocysts occur in about 25% of patients with chronic pancreatitis12,57-59,127,128 and are most common in alcoholic chronic pancreatitis.310 The most common symptom associated with a pseudocyst is abdominal pain, which occurs in the majority of symptomatic patients. Less common manifestations are a palpable mass, nausea and vomiting (due to compression of the stomach or duodenum), jaundice (due to compression of the bile duct), and bleeding. Some patients are asymptomatic. Elevations in serum lipase and amylase values are found in at least one half of patients, and a persistent elevation in serum lipase or amylase can be a clue to the presence of a pseudocyst. The diagnosis of pseudocyst is generally easily made through imaging studies, including US, CT, MRI, and EUS. The advantages of CT and MRI in this setting are visualization of the capsule of the pseudocyst, which can be used to gauge the maturity of the collection, and determination of the relation of the pseudocyst to the stomach and duodenum, which can be useful in the choice of therapy. MRI can also give some additional information on the character of the contents of the pseudocyst, in particular whether it is mainly fluid or a mixture of fluid and solid material. ERCP is usually not required for diagnostic purposes, even though around 70% of pseudocysts communicate with the pancreatic duct.311-313 ERCP is associated with an approximately 15% chance of infection of a previously uninfected pseudocyst, so this procedure should be undertaken only after antibiotics have been administered and therapy is imminent. The natural history of pseudocysts in chronic pancreatitis is not fully defined. Overall, complications of pseudocysts occur in 20% to 40% of cases. Complications include compression of large peripancreatic vessels, duodenum, stomach, or duodenum; infection; hemorrhage; and development of a fistula. Many pseudocysts will remain without symptoms or complications. Unlike fluid collections and pseudocysts associated with acute pancreatitis, those occurring in a background of chronic pancreatic resolve far less commonly. Despite that, treatment is not necessary in all patients. Patients who have mature pseudocysts smaller than 6  cm, minimal or no symptoms, no complications, and are reliable may be managed conservatively.311-313 Even larger pseudocysts that remain asymptomatic can be managed expectantly. Very large pseudocysts, an enlarging pseudocyst, and symptomatic or complicated pseudocysts require therapy. Also unlike the acute fluid collections associated with acute pancreatitis (see Chapter 58), pseudocysts occurring in the setting of chronic pancreatitis are generally mature at the time of their diagnosis and a delay in therapy is not needed to allow the pseudocyst capsule to mature. If there is a question, maturity can usually be appreciated at CT as the presence of a visible capsule around the collection. Therapy for symptomatic, complicated, or enlarging pseudocysts can be surgical, percutaneous, or endoscopic. Percutaneous tube drainage of pseudocysts is possible if a safe tract to the collection can be identified. Percutaneous drainage of pancreatic pseudocysts complicating chronic pancreatitis is discouraged owing to the widely held view that such cysts are frequently associated with ductal obstruction downstream from the fluid collection, making the risks of fistula formation along the tract and of pseudocyst recurrence or chronic fistula after removal of the tube unacceptably high. The long-term success of percutaneous drainage is still unknown but is certainly relatively low. Reaccumula-

Chapter 59  Chronic Pancreatitis tion of the collection after tube removal is the rule. Complications, which occur in less than 10% to 15% of cases, include bleeding, infection of the cavity, and formation of a draining fistula along the tube tract. Endoscopic therapy of pseudocysts is possible if the fluid collection can be accessed through the papilla or through the wall of the stomach or duodenum, and are discussed in Chapter 61 (see Figs. 61-1 to 61-3). The route chosen depends on the location of the pseudocyst. Transpapillary drainage is preferred for smaller pseudocysts in the head of the gland that communicate with the pancreatic duct. All others that are amenable to endoscopic therapy are better managed with endoscopic cystogastrostomy or cystojejunostomy, depending on their location. Success rates of 70% to 90% are routinely reported.310-314 Many centers have begun to use endoscopic therapy as first-line therapy. The complication rate is reported as about 10%. Most complications are related to transmural stent placement and include bleeding (which may occasionally be massive), perforation, and infection of previously uninfected collections. It seems possible to reduce the risk of bleeding and improve therapeutic success through the use of EUS to assess for large vessels between the gut lumen and the pseudocyst or of a direct EUS-guided puncture to avoid nearby vessels.314,315 Large channel therapeutic echoendoscopes are available for this purpose. Antibiotic coverage and readily available surgical backup are essential if endoscopic therapy is undertaken. A site is chosen with no intervening vessels. The pseudocyst is punctured with an EUS needle through which is passed a guidewire. The puncture site is dilated and one or more double-pigtail stents are placed. Not all pseudocysts are amenable to endoscopic therapy, although the number is rising because EUS-guided techniques allow drainage of pseudocysts that do not compress the stomach or duodenum. The long-term success rate of endoscopic therapy is not well defined but appears to be not as good as surgical techniques. Surgical therapy has been used most extensively and usually involves cyst decompression into a loop of small bowel or stomach, often coupled with a pancreatic ductal drainage procedure (e.g., modified Puestow procedure). Surgical therapy has a long-term success rate of 90% and an operative mortality of less than 3%.310-313 Although pseudocysts recur after surgery in only about 10% of cases, pain may return in up to one half with long-term follow-up. This is true of all therapies for pseudocysts, in that pain from the underlying chronic pancreatitis may also occur in the absence of a pseudocyst. Surgical therapy is also necessary in patients who experience severe complications of lessinvasive endoscopic or percutaneous treatments.316,317 Cystogastrostomy and cystojejunostomy can also be performed with laparoscopic techniques, although experience with these approaches remains limited.318 No prospective randomized trials comparing surgical with endoscopic therapy have been performed. It has been noted that failure of percutaneous drainage of pseudocysts is often associated with a stricture of the pancreatic duct downstream (toward the duodenum) from the pseudocyst or a significant disruption of the duct.316 These features predict prompt recurrence after the tubes are removed. Endoscopic therapy is prone to a similar problem unless these anatomic problems are dealt with. MRCP and ERCP can be used to identify patients with pancreatic duct strictures or major disruptions who are at increased risk of recurrence. Although not a routine practice at all centers, performance of ERCP in association with (immediately before or after) EUS-guided pseudocyst drainage is reasonable.311,312 The goal of the ERCP is to identify a pancreatic

Table 59-8  Cystic Collections within the Pancreas Pseudocyst (70%-90%) Cystic neoplasms (10%-15%)* Mucinous tumors Mucinous cystadenoma and cystadenocarcinoma Intraductal papillary mucinous neoplasm (formerly mucinous ductal ectasia) Serous tumors Serous cystadenoma Cystadenocarcinoma (rare) Solid pseudopapillary tumor Rare neoplasms Acinar cell cystadenocarcinoma Choriocarcinoma Teratoma Neoplasms that may occasionally appear cystic Islet cell tumors Pancreatic ductal adenocarcinoma True cysts (rare) Polycystic disease of the pancreas Isolated Associated with polycystic disease of the kidneys von Hippel-Lindau disease Simple true cyst Dermoid cyst Miscellaneous cystic lesions (very rare) Lymphoepithelial cyst Endometrial cyst Macrocyst associated with cystic fibrosis Retention cyst Parasitic cyst (Echinococcus granulosus or Taenia solium) *See Chapter 60.

duct stricture or large duct disruption and treat this with a bridging stent. This is felt to reduce the risk of pseudocyst recurrence after removal of the transenteric pseudocyst stents. Pseudocysts account for 90% of all cystic collections associated with the pancreas. A number of other cystic collections can mimic the appearance of a pseudocyst, in particular cystic neoplasms (Table 59-8; see also Chapter 60). Cystic neoplasms are often discovered when CT or ultrasonography is performed for vague abdominal symptoms or often for other reasons. The appropriate therapy of these lesions is resection (not drainage as for pseudocysts). Cystic neoplasms should be suspected when a mildly symptomatic or asymptomatic fluid collection is discovered in a patient (usually a middle-aged woman) with neither a history of nor risk factors for pancreatitis. The finding of internal septations or nodules within the wall of the collection on CT or EUS is also highly suggestive of a cystic neoplasm (Fig. 59-9). The features that can assist in differentiation of a pseudocyst from a cystic neoplasm and benign serous cystadenomas are summarized in Table 59-9 and in Table 60-9. EUS is particularly helpful in differentiating pseudocysts from cystic tumors, which involves using endosonographic appearance, cytologic material obtained with EUS-guided fine-needle aspiration or brush cytology, and analysis of cyst fluid for tumor markers, amylase, and mucin.319-322

GASTROINTESTINAL BLEEDING

GI bleeding in the setting of chronic pancreatitis may develop from a variety of causes. Some are not related to chronic pancreatitis, such as bleeding from a Mallory-Weiss tear, esophagitis, peptic ulcer disease, and varices from concomitant alcoholic cirrhosis. Others occur as a direct result of the pancreatitis, most notably bleeding from a pancreatic

1011

1012

Section VII  Pancreas Table 59-9  Features That Distinguish Pseudocyst from Mucinous Cystic Neoplasm and Serous Cystadenoma FEATURE

PSEUDOCYST

MUCINOUS CYSTIC NEOPLASM

SEROUS CYSTADENOMA

Gender distribution Age at presentation History of acute or chronic pancreatitis Symptoms

Male predominance 3rd to 5th decade Common

Female predominance 5th to 7th decade Rare

Female predominance 7th decade Rare

Abdominal pain

Usually asymptomatic

Imaging characteristics (CT or EUS)

Unilocular, homogeneous, no solid component, findings of associated chronic pancreatitis Common

Asymptomatic, less commonly abdominal pain Unilocular or multilocular with septations, mural nodules, occasional calcifications in wall Uncommon Thick and viscous, may not be able to be aspirated Low High (>192 ng/mL) Mucinous epithelial cells or malignant cells Resection

Thin, clear

Communication with pancreatic duct Fluid characteristics

Thin, clear to dark

Amylase level in fluid CEA level in fluid Cytology of fluid or wall

High Low Inflammatory cells

Therapy

Drainage

Microcystic, honeycomb appearance Uncommon

Low Low Cuboidal cells without mucin Observation

CEA, carcinoembryonic antigen; CT, computed tomography; EUS, endoscopic ultrasonography.

Pseudoaneurysm

Figure 59-9.  Computed tomography scan demonstrating a cystic neoplasm seen in the tail of the pancreas in an older adult woman with no history of pancreatic disease. The loculations and mural nodules seen within the cavity (arrow) are suggestive of a cystic neoplasm, rather than a pseudocyst (see Chapter 60).

pseudocyst, pseudoaneurysm, and portal or splenic vein thrombosis.

Pseudocyst

Bleeding may occur from the wall of a pseudocyst. Bleeding occurs from small vessels (venous, capillary, or arteriole) in the wall, which can lead to expansion of the pseudocyst and further rupture of these small vessels.323 Blood may remain in the pseudocyst or may reach the gut via a spontaneous pseudocyst decompression into the GI lumen or into the pancreatic duct (hemosuccus pancreaticus). Bleeding from small vessels in the wall of the pseudocyst is generally of low volume.

Pseudoaneurysms form as a consequence of enzymatic and pressure digestion of the muscular wall of an artery by a pseudocyst. The pseudoaneurysm may rupture either into the pseudocyst (converting the pseudocyst into a larger pseudoaneurysm) or directly into an adjacent viscus, peritoneal cavity, or pancreatic duct. Pseudoaneurysmal bleeding may complicate 5% to 10% of all cases of chronic pancreatitis with pseudocysts, although pseudoaneurysms may be seen in up to 21% of patients with chronic pancreatitis undergoing angiography.323 Pseudoaneurysms are also seen after pancreatic surgery. Many visceral arteries may be involved, the splenic artery being most common, followed by gastroduodenal or pancreaticoduodenal arteries. Once bleeding occurs, the mortality is at least 40% being related both to the severity of the blood loss and to the presence of coexisting conditions. Although death from a pseudocyst is rare, more than half the overall mortality of pseudocysts is due to hemorrhage. Bleeding from a pseudoaneurysm may be slow and intermittent or acute and massive. Common presentations are abdominal pain (due to the enlargement of the pseudocyst), unexplained anemia, and overt GI bleeding (if the blood can reach the gut lumen through the pseudocyst or through the pancreatic duct). In many cases, an initial self-limited bleed occurs (so-called sentinel bleed), followed hours or days later by a massive exsanguinating hemorrhage. The initial self-limited nature of the bleed may be due to transient tamponade of the bleeding within the confines of the pseudocyst. The presence of unexplained blood loss or any amount of GI bleeding in a patient with pancreatitis or a known pseudocyst should immediately raise the possibility of a pseudoaneurysm. If a pseudoaneurysm is suspected in the setting of upper GI blood loss, an urgent upper endoscopy should be undertaken. If no obvious bleeding site is seen, pseudoaneurysm formation should be considered. Rarely, blood may be seen issuing from the ampulla (hemosuccus pancreaticus), but the absence of this finding does not rule out pseudoaneurysm. The next step in the evaluation should be an emergency CT scan with intravenous contrast. The finding of highdensity material within a pseudocyst on the initial noncon-

Chapter 59  Chronic Pancreatitis

Figure 59-10.  Computed tomography scan demonstrating a pseudocyst containing a pseudoaneurysm (arrow) that is opacified following intravenous injection of contrast agent.

trast images is highly suggestive of a pseudoaneurysm as is a circular opacifying structure within the low-attenuation pseudocyst after the intravenous administration of contrast agent (Fig. 59-10). It is prudent to avoid oral administration of a contrast agent so that it will not interfere with angiography if required. In most centers, such a CT finding is followed immediately by angiography to define and embolize the pseudoaneurysm. Once a pseudoaneurysm has been identified, it should be treated whether or not it has caused bleeding. Angiographic embolization or stent-graft placement has largely replaced primary surgery, which is reserved for cases in which these therapies have failed.324

Variceal Bleeding from Splenic Vein Thrombosis

Variceal bleeding may complicate chronic pancreatitis because of either associated alcoholic cirrhosis or thrombosis of the splenic (and, less commonly, portal) vein (see Chapters 19 and 90). Thrombosis of the splenic vein is most common and produces a segmental or left-sided portal hypertension.323 Decompression of splenic venous outflow occurs through the short gastric veins to the coronary vein, producing prominent variceal channels in the gastric cardia and fundus. Depending on the venous anatomy, esophageal varices may also be produced, although they are generally smaller than the gastric varices. The natural history of gastric varices in this setting is not known, but the overall risk of bleeding is less than with esophageal varices due to cirrhosis. In one analysis, the risk of gastric variceal bleeding was only 4%.325 Therapy is therefore not required in the absence of bleeding. Should bleeding occur, splenectomy is curative. Endoscopic control of bleeding is sometimes possible with gastric varices, utilizing cyanoacrylate injection or other techniques (see Chapter 19).

BILE DUCT OBSTRUCTION

The distal bile duct is enclosed within the posterior portion of the head of the pancreas. Inflammatory and fibrotic conditions of the head of the pancreas, as well as pseudocysts in this location, can compress this intrapancreatic bile duct, leading to abnormal liver chemistry values, jaundice, biliary pain, or cholangitis. Symptomatic bile duct obstruction occurs in about 10% of patients. The ductal stricture can be suspected from cholestatic liver chemistry values, CT or ultrasonography findings of biliary ductal dilation, or

Figure 59-11.  A retrograde cholangiogram showing a smooth stricture of the bile duct (arrows) as it passes through the head of the pancreas in a patient with chronic pancreatitis.

both. ERCP characteristically demonstrates a long tapered stenosis of the distal bile duct (Fig. 59-11). The occurrence of cholangitis is an absolute indication for therapy. The presence of abnormal liver chemistry values or jaundice is not so straightforward because those most affected are alcoholic patients, and alcoholic (and other intrinsic) liver disease can also produce substantial abnormalities in liver chemistry values. The clinical, biochemical, and even radiologic features are not always sufficient to distinguish biliary stenosis from intrinsic liver disease.326 It is for this reason that liver biopsy is often necessary to enable the choice of therapy. The mere presence of a stenosis of the intrapancreatic bile duct, in the absence of symptoms or progressive abnormalities in liver chemistry values, can usually be followed conservatively. If there is a concern about the development of secondary biliary cirrhosis, a liver biopsy should be performed. In patients with increasing jaundice or biliary pain, in the absence of alternative explanations (i.e., intrinsic liver disease), therapy should be considered. Definitive therapy of bile duct obstruction usually requires surgical biliary bypass with choledochojejunostomy or choledochoduodenostomy. One study suggested that hepatic fibrosis due to chronic biliary obstruction may actually decrease after successful surgical decompression.327 Although endoscopic stent therapy for biliary obstruction due to chronic pancreatitis is generally temporarily effective (see Chapter 61), the long-term success is relatively low.328 Placement of one or more plastic stents to treat bile obstruction is relatively simple, but the long-term management is complicated by the need for multiple stent exchanges over many months to years, and stent migration and obstruction are common. The use of multiple parallel stents shows a better response than a single stent.329 Long-term endoscopic stent therapy in alcoholic patients is particularly difficult and is associated with high complication rates owing to missed appointments for scheduled stent exchanges.330 The use of uncoated permanent metallic stents is discouraged because of the high rate of ultimate stent occlusion, although the use of fully coated, removable metal stents may provide additional options in the future.

1013

1014

Section VII  Pancreas The development of a bile duct stenosis in a patient with chronic pancreatitis may also signal the development of a pancreatic malignancy. EUS and ERCP are useful in this setting to attempt to differentiate benign from malignant strictures.

DUODENAL OBSTRUCTION

Approximately 5% of patients with chronic pancreatitis experience symptomatic duodenal stenosis. The majority of such patients have chronic alcoholic pancreatitis. Fibrosis in the head of the pancreas, often associated with an inflammatory mass, is the most common explanation. Symptoms include nausea, vomiting, weight loss, and abdominal pain. Coexistent obstruction of the bile duct may occur. The diagnosis is best made with CT using oral contrast or an upper GI barium study, because the extent of duodenal stenosis is often underestimated at the time of endoscopy. Because the degree of stenosis may improve with resolution of some of the inflammation, a trial of medical therapy may be worthwhile. Surgical therapy is required for those in whom conservative management fails. The simplest approach is a bypass with a gastrojejunostomy, which may be performed with laparoscopic techniques. This may be coupled with drainage of the bile duct and/or pancreatic duct (lateral pancreaticojejunostomy). Resection of the head of the pancreas with a Frey, Beger, Berne, or Whipple procedure (see earlier) may be considered in centers with this expertise in select patients with a large inflammatory mass of the head of the pancreas, or in those in whom malignancy is also being considered.

PANCREATIC FISTULA External Fistulas

External pancreatic fistulas occur most commonly as a consequence of surgical or percutaneous therapy for chronic pancreatitis or pseudocyst. It has been estimated that perhaps half of such fistulas heal with complete bowel rest and parenteral hyperalimentation. The most common complications are abscess and bleeding. There is some evidence that the addition of octreotide, in a dosage of 100 µg subcutaneously every eight hours, can hasten closure of such fistulas. Successful medical treatment, even with octreotide, can take many weeks. The placement of an endoscopic stent across the site of ductal disruption is effective at closing the fistula rapidly (see Chapter 61). Up to 75% of pancreaticocutaneous fistulas may be effectively treated with endoscopic techniques,331,332 although this approach may need to be coupled with percutaneous drainage of intra-abdominal fluid collections. In patients in whom endoscopic therapy fails or cannot be performed, surgical treatment can involve pancreatic resection (if the fistula is in the tail) or a fistulojejunostomy, in which the fistula tract is “capped” with a defunctionalized limb of jejunum.

Internal Fistulas

Internal pancreatic fistulas occur mainly in the setting of chronic pancreatitis after rupture of a pseudocyst. The fluid may track to the peritoneal cavity (pancreatic ascites) or into the pleural space (pancreatic pleural effusion). Affected patients may not complain of symptoms of chronic pancreatitis but may instead note abdominal distention or shortness of breath, respectively. Although such fistulas invariably occur in advanced chronic pancreatitis (particularly alcoholic) there may not be a clear-cut history of recent pancreatitis. The diagnosis can be established through documentation of high levels of amylase within the respective fluid, typically more than 4000 U/L.

Conservative treatment, consisting of complete bowel rest, parenteral hyperalimentation, paracentesis or thoracentesis, and octreotide, is effective in some internal pancreatic fistulas.333 If the leak is in the body or head of the pancreas, a pancreatic duct stent covering the fistula site is highly effective (see Fig. 61-8).331,332 In some cases, merely bridging the ampulla with a short pancreatic duct stent may be enough to heal the fistula. Endoscopic therapy is less effective but still worthwhile if the leak is from the tail but is ineffective if the leak is present upstream from a complete blockage of the pancreatic duct (excluded pancreatic tail syndrome). In this situation, resection or surgical drainage of the pseudocyst is required, and ERCP or MRCP is used preoperatively to delineate the ductal anatomy for surgical planning.

MALIGNANCY

The risk of pancreatic cancer is higher with all forms of chronic pancreatitis (see Chapter 60). The lifetime risk for pancreatic cancer in patients with chronic pancreatitis is about 4%.334 The risk of pancreatic cancer is highest in patients with hereditary pancreatitis, and particularly those who smoke (see Chapter 57). At present, there is no completely reliable way to differentiate chronic pancreatitis alone from chronic pancreatitis complicated by adenocarcinoma. The symptoms and signs may be similar (abdominal pain, weight loss, jaundice). In the absence of widespread metastases, imaging studies such as CT, ultrasonography, and even ERCP may be unable to establish the diagnosis. The role of EUS is evolving, but finding a small hypoechoic tumor within a diseased gland with preexisting altered echotexture can be difficult. However, EUS is superior to CT for detection of coexistent malignancy particularly when the lesion is small. EUS also has the substantial advantage of allowing directed tissue biopsy of any suspicions lesions. Tumor markers may also be helpful in attempting to differentiate chronic pancreatitis from cancer. CA 19-9, the tumor marker most commonly used for pancreatic adenocarcinoma, is elevated in the serum in 70% to 80% of patients with adenocarcinoma of the pancreas.335 Biliary obstruction and cholangitis can also raise CA 19-9 levels. The use of any of these techniques for surveillance is not cost-effective in the general population of patients with chronic pancreatitis, although they may be useful in families with hereditary pancreatic cancer. In some patients, laparotomy is required to determine the presence or absence of coexistent pancreatic carcinoma. In those with a benign pseudotumor who undergo resection to rule out malignancy, a variant of autoimmune chronic pancreatitis is often found.95,97 Several reports have also called attention to the development of extrapancreatic cancer in association with chronic pancreatitis, which are more common than pancreatic malignancy in these patients. These cancers, particularly those of the upper digestive tract and lungs, are probably related to the effect of concomitant tobacco use.20,334 The incidence of extrapancreatic carcinomas in these reports varies between 4% and 12%.

DYSMOTILITY

Gastroparesis and antroduodenal dysmotility are seen in patients with chronic pancreatitis,336,337 possibly as a consequence of perigastric inflammation, hormonal changes associated with chronic pancreatitis (e.g., increases in plasma CCK), or a side effect of narcotic analgesics. Gastroparesis is clinically important because it may produce symptoms occasionally indistinguishable from those of the disease and

Chapter 59  Chronic Pancreatitis may interfere with the effective delivery of pancreatic enzymes.337 Gastroparesis should be considered in patients with early satiety, nausea, vomiting, and weight loss.

KEY REFERENCES

Anapurthy R, Pasricha PJ. Pain and chronic pancreatitis. Is it the plumbing or the wiring? Curr Gastroenterol Rep 2008; 10:101-6. (Ref 144.) Baillie J. Pancreatic pseudocysts (part I). Gastrointest Endosc 2004; 59:873-9. (Ref 311.) Baillie J. Pancreatic pseudocysts (part II). Gastrointest Endosc 2004; 60:105-13. (Ref 312.) Blondet JJ, Carlson AM, Kobayashi T, et al. The role of total pancreatectomy and islet autotransplantation for chronic pancreatitis. Surg Clin North Am 2007; 87:1477-501. (Ref 296.) Cahen DL, Gouma DJ, Nio Y, et al. Endoscopic versus surgical drainage of the pancreatic duct in chronic pancreatitis. N Engl J Med 2007; 356:676-84. (Ref 281.) Chowdhury RS, Forsmark CE. Review article: Pancreatic function testing. Aliment Pharmacol Ther 2003; 17:733-50. (Ref 175.) Collins D, Penman I, Mishra G, Draganov P. EUS-guided celiac block and neurolysis. Endoscopy 2006; 38:935-9. (Ref 299.) Dite P, Ruizicka M, Zboril V, Novotmy I. A prospective, randomized trial comparing endoscopic and surgical therapy for chronic pancreatitis. Endoscopy 2003; 35:553-8. (Ref 280.)

Frulloni L, Lunardi C, Simone R, et al. Identification of a novel antibody associated with autoimmune pancreatitis. N Engl J Med 2009; 361: 2135-42. (Ref 87a.) Gardner TB, Chari ST. Autoimmune pancreatitis. Gastroenterol Clin North Am 2008; 37:439-60. (Ref 24.) Gleeson FC, Topazian M. Endoscopic retrograde cholangiopancreatography and endoscopic ultrasonography for diagnosis of chronic pancreatitis. Curr Gastroenterol Rep 2007; 9:123-9. (Ref 222.) Khalid A, Brugge W. ACG Practice Guidelines for the diagnosis and management of neoplastic cysts. Am J Gastroenterol 2007; 102:233949. (Ref 321.) Van der Gaag MA, Gouma DJ, Van Gulik TM, et al. Review article: Surgical management of chronic pancreatitis. Aliment Pharmacol Ther 2007; 26:221-32. (Ref 283.) Wilcox CM, Varadarajulu S. Endoscopic therapy for chronic pancreatitis: An evidence-based review. Curr Gastroneterol Rep 2006; 8:10410. (Ref 272.) Witt J, Apte MV, Keim V, Wilson JS. Chronic pancreatitis: Challenges and advances in pathogenesis, genetics, diagnosis, and therapy. Gastroenterology 2007; 132:1557-73. (Ref 36.) Full references for this chapter can be found on www.expertconsult.com.

1015

CHAPTE R

60 Tumors of the Pancreas Ramon E. Jimenez and Carlos Fernández-del Castillo

CHAPTER OUTLINE Pancreatic Cancer  1017 Epidemiology  1017 Pathology  1018 Clinical Features  1020 Diagnosis  1020 Staging  1023 Treatment  1024

PANCREATIC CANCER Pancreatic cancer is the second most common gastrointestinal malignancy in the United States, and approximately 37,680 new cases are expected to occur in 2008.1 Despite its relatively low incidence compared with other malignancies, it represents the fourth leading cause of cancer death in men and women (34,290 deaths expected in 2008). Overall, pancreatic cancer carries an unfavorable prognosis. For all stages combined, the one- and five-year relative survival rates are 24% and 5%, respectively.1 Early detection, accurate preoperative staging, and better treatment options remain a challenge.

EPIDEMIOLOGY Incidence

Pancreatic cancer is very rare before the age of 45 years, but its occurrence rises sharply thereafter. It affects men more than women (ratio of 1.3 : 1), and is more common in blacks. The incidence in black men is 14.8 per 100,000, compared with 8.8 in the general population.2

Populations at Risk

Genetic as well as environmental factors have been found to be associated with the development of pancreatic cancer (Fig. 60-1). Table 60-1 summarizes some of the genetic syndromes associated with an increased risk of pancreatic cancer. One of the most prominent of these syndromes is hereditary pancreatitis, even though it accounts for only a small fraction of pancreatic cancer cases (see Chapter 57). Affected patients have an abnormal trypsin gene that is transmitted as an autosomal dominant trait; their risk for development of pancreatic cancer by age 70 years is estimated at 40%.3 Patients with other, nonhereditary forms of chronic pancreatitis also have a higher likelihood of pancreatic cancer (see Chapter 59). A multinational study found this risk to be 2% per decade, independent of the type of pancreatitis.4 Individuals with familial Peutz-Jeghers syndrome, discussed in Chapter 122, also have an increased risk of pancreatic cancer, with an impressive 132-fold increased risk over the general population.5 Germline mutations in p16 are

Cystic Tumors of the Pancreas  1027 Differential Diagnosis  1027 Diagnostic Imaging  1028 Mucinous Cystic Neoplasms  1029 Serous Cystadenomas  1029 Intraductal Papillary Mucinous Neoplasms  1030 Solid Pseudopapillary Tumors  1032 Other Nonendocrine Pancreatic Tumors  1033

observed in kindreds with the familial atypical molemalignant melanoma (FAMMM) syndrome, and these individuals are at risk for pancreatic cancer and melanoma.6 Lastly, patients with BRCA2 gene mutations (which pre­ dispose to hereditary breast cancer in women and men), also have a familial predisposition to pancreatic cancer.7 Other unknown genetic abnormalities are yet to be identified. In several population studies, 7% to 8% of patients with pancreatic cancer have a first-degree relative with the disease.8 Patients with three or more first-degree relatives with the disease, have a 32-fold increased risk over the general population; those with two first degree relatives have 6-fold increased risk; and those with one first-degree relative have a 2.3-fold increased risk.9 There are no specific recommendations for screening patients at risk for pancreatic cancer, because available techniques lack sensitivity for detection of very small lesions. The timing and frequency of such screening are also uncertain. The American Gastroenterological Association (AGA) suggests that screening should begin at age 35 in patients with hereditary pancreatitis. Patients with a history of familial pancreatic cancer should begin screening 10 years before the age at which pancreatic cancer has been first diagnosed in their relatives. The AGA also states that such screening is probably best done with spiral computed tomography (CT) and endoscopic ultrasonography (EUS).10,11

Environmental Factors

The most important environmental factor in pancreatic cancer, and possibly the only one that has been firmly established, is cigarette smoking.12 Multiple cohort and casecontrol studies have found that the relative risk for smokers of developing pancreatic cancer is at least 1.5.13-15 The risk may be particularly elevated in smokers who have homozygous deletions of the gene for glutathione S-transferase T1 (GSTT1), which is a carcinogen metabolizing enzyme.16 Furthermore, the risk rises with the amount of cigarette consumption, and the excess risk level returns to baseline by 15 years after cessation of the habit.15 The second most important environmental factor associated with pancreatic cancer appears to be dietary influences. A high intake of fat or meat has been linked to the development of this neoplasm,17 and a protective effect is ascribed to fresh fruits and vegetables.18 Reduced serum levels of

1017

1018

Section VII  Pancreas Table 60-1  Historical Features and Germline Genetic Alterations Associated with an Increased Risk of Pancreatic Cancer history

GENE

RELATIVE RISK

individual RISK BY AGE 70

None Breast cancer

None BRCA2 BRCA1 p16 (CDKN2A) Unknown PRSS1 STK11/LKB1 MLH1, MSH2, others FANC-C, FANC-G, others Palladin

1 3.5-10× 2× 20-34× 32× 50-80× 132× Unknown Unknown Unknown

0.5% 5% 1% 10%-17% 16% 25%-40% 30%-60% <5% Unknown Unknown

FAMMM ≥3 FDR with PC Hereditary pancreatitis Peutz-Jeghers syndrome HNPCC syndrome Young age onset PC Family X

FAMMM, familial atypical multiple mole melanoma syndrome; ≥3 FDR with PC, three or more first-degree relatives with pancreatic cancer; HNPCC, hereditary nonpolyposis colorectal cancer syndrome; PC, pancreatic cancer; X, a single family in which familial pancreatic cancer was studied. Modified from Hruban RH, Pitman MB, Klimstra DS. AFIP Atlas of Tumor Pathology. Tumors of the Pancreas. Washington, D.C.: American Registry of Pathology; 2007.

Heredity 10%

Smoking 30%

Table 60-2  World Health Organization Classification of Primary Tumors of the Exocrine Pancreas

Diet 20%

Benign disease 4%

Unknown 36% Figure 60-1.  Risk factors for pancreatic cancer. (Courtesy of Dr. Albert P. Lowenfels.)

lycopene, a carotenoid present in fruits, and selenium were found in subjects who subsequently had pancreatic cancer.19 Diabetes mellitus is very common in patients with pancreatic cancer. In most cases diabetes has been diagnosed within the preceding two years and may or may not be associated with obesity.20 Thus, recent onset of diabetes may help identify patients with pancreatic cancer, particularly in individuals older than 50 years. Destruction of the pancreas is unlikely to be sufficient to cause endocrine insufficiency in most patients with pancreatic cancer, and it has been proposed that higher production of islet amyloid polypeptide (amylin) by the tumor is responsible for the diabetogenic state. In fact, glucose tolerance frequently improves in patients who have undergone tumor resection.21

PATHOLOGY

As reviewed in Chapter 55, three different epithelial cell types are found in the normal pancreas: (1) acinar cells, which account for about 80% of the gland volume; (2) ductal cells, comprising 10% to 15%; and (3) endocrine (islet) cells, making up about 1% to 2%. More than 95% of the malignant neoplasms of the pancreas arise from the exocrine elements of the gland (ductal and acinar cells) and demonstrate features consistent with adenocarcinoma. Endocrine neoplasms account for only 1% to 2% of pancreatic tumors (see Chapter 32). Nonepithelial malignancies are exceedingly rare.22

I. Benign i. Serous cystadenoma ii. Mucinous cystadenoma iii. Intraductal papillary mucinous adenoma iv. Mature cystic teratoma II. Borderline (uncertain malignant potential) i. Mucinous cystic tumor with moderate dysplasia ii. Intraductal papillary mucinous tumor with moderate dysplasia iii. Solid-pseudopapillary tumor I II. Malignant i. Ductal adenocarcinoma ii. Osteoclast-like giant cell tumor iii. Serous cystadenocarcinoma iv. Mucinous cystadenocarcinoma (noninvasive or invasive) v. Intraductal papillary mucinous carcinoma (noninvasive or invasive) vi. Acinar cell carcinoma vii. Pancreatoblastoma viii. Solid-pseudopapillary carcinoma ix. Miscellaneous carcinomas Data from Hamilton SR, Aaltonen LA. World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of the Digestive System. Lyon, France: IARC Press; 2000.

The World Health Organization (WHO) has proposed a classification of pancreatic exocrine tumors that is widely used today (Table 60-2).23 Ductal adenocarcinoma accounts for 85% to 90% of pancreatic tumors.22 Autopsy series have shown that 60% to 70% of these tumors are localized in the head of the gland, 5% to 10% in the body, and 10% to 15% in the tail. On gross examination these tumors appear as firm masses with ill-defined margins blending into the surrounding pancreatic parenchyma. The average size of carcinomas in the head of the pancreas is 2.5 to 3.5 cm, compared with 5 to 7 cm for tumors in the body or tail. Differences in tumor size at presentation are related to the earlier development of symptoms and signs in proximal tumors than in distal neoplasms. Tumors in the head of the gland have a propensity for obstruction of the distal common bile duct and pancreatic duct. Anatomic obstruction of these structures results in jaundice and chronic obstructive pancreatitis. Pancreatic pathologic changes observed include duct dilatation and

Chapter 60  Tumors of the Pancreas fibrous atrophy of the pancreatic parenchyma. Some tumors can involve the duodenum or the ampulla of Vater. Extrapancreatic extension into the retroperitoneal tissues is almost always present at the time of diagnosis and can result in invasion of the portal vein or the superior mesenteric vessels and nerves. Neoplasms of the tail of the pancreas do not cause biliary or pancreatic duct obstruction. Extrapancreatic extension in distal tumors causes invasion of the spleen, stomach, splenic flexure of the colon, or left adrenal gland. In patients with advanced disease, metastases to the lymph nodes, liver, and peritoneum are common; the lung, pleura, and bone are less commonly involved.24 Microscopically, ductal adenocarcinomas are graded as well, moderately, or poorly differentiated.22 Well-differentiated tumors show irregular tubular neoplastic glands with mild cellular atypia, low mitotic activity, and significant mucin production. Loss of differentiation results from lack of cellular arrangement into glandular structures, increases in cellular atypia and mitotic figures, and cessation of mucin production. Some studies have demonstrated that histologic grading correlates with survival.25 Ductal adenocarcinomas elicit a strong desmoplastic reaction that is responsible for their hard consistency on gross inspection. In contrast with chronic pancreatitis, intraductal calcifications are only rarely found. Pancreatic ducts outside the area of neoplasia may demonstrate papillary hyperplasia or mucinous cell hypertrophy. The significance of these findings is unknown. Microscopic extension of tumor is often evident in lymphatic channels and perineural spaces. Metastasis of tumors in the head of the pancreas to first eschelon lymph nodes in the pancreaticoduodenal basin is common. Celiac and para-aortic lymph node involvement can be observed in locally advanced disease. Several immunohistochemical markers have shown diagnostic usefulness in mucin-producing tumors such as pancreatic adenocarcinoma. Among the better-known markers are MUC1, MUC3, MUC4, carcinoembryonic antigen (CEA), CA 19-9, DuPan 2, and CA 125.22 These markers are unable to differentiate between tumors of pancreatic and extrapancreatic origins, limiting their usefulness in the evaluation of liver metastases of unknown primary. However, they are particularly useful in separating neoplastic from nonneoplastic ductal changes and in distinguishing ductal from acinar or neuroendocrine tumors. Cytokeratins are other useful markers in differentiating among acinar, ductal, and islet cell tumors. Although all ductal adenocarcinomas stain for cytokeratins 7, 8, 18, and 19, most acinar and neuroendocrine tumors do not stain for cytokeratin 7.22 Progression of small pancreatic intraductal lesions to ductal adenocarcinoma has been described, similar to the adenoma-carcinoma model in colorectal cancer.26 These precursor lesions are referred to as pancreatic intraepithelial neoplasia (PanIN). PanINs are microscopic findings that are graded from 1 to 3, the last being equivalent to carcinoma in situ.

Molecular Pathology

Multiple combinations of genetic mutations are commonly found in pancreatic adenocarcinomas. Table 60-3 lists the most common. The accumulation of genetic alterations over time is thought to result in the development of pancreatic cancer. For example, progression of PanIN lesions from grade 1 to 3 has been linked to an increasing number of genetic mutations (Fig. 60-2).27 A long list of oncogenes and their products has been implicated in the pathogenesis of pancreatic cancer.28 Mutations in the K-ras gene are a hallmark of pancreatic adenocarcinoma and appear to be present in more than 90% of

Table 60-3  Genetic Mutations in Pancreatic Cancer GENE (CHROMOSOMAL REGION)

PERCENT OF TUMORS WITH GENETIC mutation

K-ras (12p)* p16 CDKN2A (9p)† PK53 (17p)2 SMAD4/DPC4 (18q)† AKT2 (19q)* MYB (6q)* AIB1 (20q)* BRCA2 (13q)† LKB1/STK11 (19p)† MKK4 (17p)† TGF-β-R1 (9q) or TGF-β-R2 (3p) RB1 (13q)

>90% >95% 50%-70% 55% 10%-20% 10% 10% 7%-10% <5% <5% <5% <5%

*Oncogenes. † Tumor suppressor genes. TGF-β, transforming growth factor-β. Modified from Hruban RH, Wilentz RE. The pancreas. In: Kumar V, Abbas AK, Fausto N, editors. Robbins and Cotran pathologic basis of disease. 7th ed. Philadelphia: Elsevier; 2005.

Normal duct

PanIN-1 Early Events PanIN-2

Telomere shortening* K-ras mutation p16INKa loss

PanIN-3

Later Events

IPMN

MCN

TP53 loss SMAD4/DPC loss (BRCA2/LKB1)*

Pancreatic Ductal Adenocarcinoma Figure 60-2.  Pancreatic precursor lesions and genetic events involved in pancreatic adenocarcinoma progression. Pictured are three known human pancreatic ductal adenocarcinoma (PDAC) precursor lesions: PanIN, MCN, and IPMN. The PanIN grading scheme is shown on the left; increasing grade (1 through 3) reflects increasing atypia, eventually leading to frank PDAC. The right side illustrates the potential progression of MCNs and IPMNs to PDAC. The genetic alterations documented in adenocarcinomas also occur in PanIN, and to a lesser extent MCNs and IPMNs, in an apparent temporal sequence, although these alterations have not been correlated with the acquisition of specific histopathologic features. The various genetic events are listed and divided into those that occur predominantly early or later in PDAC progression. Asterisks indicate events that are not common to all precursor lesions (e.g., telomere shortening and BRCA2 loss are documented in PanIN, and LKB1 loss is documented in a subset of PDACs and IPMNs). (From Hezel AF, Kimmelman AC, Stanger BZ, et al. Genetics and biology of pancreatic ductal adenocarcinoma. Genes Dev 2006; 20:1218-49.)

1019

1020

Section VII  Pancreas tumors. Studies in intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs) of the pancreas have shown that the frequency of K-ras mutations correlates with the extent of microscopic dysplasia within tumors.29,30 Evidence suggests that K-ras mutation may be an early genetic event in pancreatic carcinogenesis, but K-ras mutations may be detected even in the setting of chronic pancreatitis without frank neoplasia.31 Loss of function in several tumor suppressor genes has also been found in pancreatic tumors, notably p16 (also called INK4A), TP53, and DPC4 (also called SMAD4). The combination of p16 and K-ras mutations is uncommon in other human tumors and may be a molecular “signature” for pancreatic adenocarcinoma.32 Loss of p16 function is observed in 80% to 95% of sporadic pancreatic cancers; p16 has a critical role in regulation of the cell cycle, and disruption of its function can lead to accelerated cell division and growth.33 Germline mutations in p16 are associated with the FAMMM syndrome, which is characterized by a high incidence of melanoma and pancreatic cancer.34 The most commonly mutated tumor suppressor gene in human cancer, TP53, is also mutated in a high percentage of pancreatic tumors.28,35 Disruption of TP53 function has been linked to alterations in the cell cycle, regulation of transcription, deoxyribonucleic acid (DNA) repair, and apoptosis, leading to the rampant genetic instability that characterizes pancreatic cancer (see Chapter 3). Another tumor suppressor gene involved in pancreatic carcinogenesis is DPC4.36 The current evidence suggests that DPC4 is a key transcription factor involved in the regulation of transforming growth factor-β expression and subsequent growth inhibition.37 Therefore, disruption of DPC4 could have critical effects in cell cycle regulation and cell differentiation. Besides oncogenes and tumor suppressor genes, other genetic pathways have been implicated in pancreatic carcinogenesis. Up-regulation of epidermal growth factor (EGF) receptors and their ligands has been well established.38 Telomere shortening and dysfunction also has been found.39 More recently, abnormal activation of genes involved in pancreatic organogenesis has been described (see Chapter 55). One study suggests that the sonic hedgehog gene product is abnormally present early as well as late in pancreatic carcinogenesis.40

CLINICAL FEATURES

Most patients with pancreatic cancer experience symptoms late in the course of disease. The lack of early symptomatology leads to delays in diagnosis, and less than 20% of patients present with resectable masses.41 Tumors of the head of the pancreas produce symptoms earlier in the course of disease. In contrast, tumors of the distal gland are characterized by their “silent” presentation, with physical findings appearing only after extensive local growth or widely metastatic disease has developed. Clinical signs and symptoms can offer clues to the resectability of pancreatic tumors (Table 60-4).42 Jaundice is often the first sign that brings patients to medical attention, especially with tumors in the head of the pancreas. Present in more than 50% of patients, jaundice results from obstruction of the extrahepatic bile duct. In less than one third of patients, a palpable nontender gallbladder, can be found (Courvoisier’s sign). Patients with concomitant obstruction of the pancreatic duct may also show pancreatic exocrine insufficiency in the form of steatorrhea and malabsorption. Few patients who have pancreatic tumors but no jaundice present with resectable disease. Pain can be a major symptom in many patients with pancreatic cancer. Pain is primarily due to invasion of the celiac

Table 60-4  Demographics and Presenting Symptoms and Signs in Patients with Unresectable (Palliated) and Resected Pancreatic Cancer*

Demographics Age, average (yr) Men/women Race Symptoms and Signs (%) Abdominal pain Jaundice Weight loss Nausea/vomiting Back pain

PALLIATED (N = 256)

RESECTED (N = 512)

64.0 57%/43% 91% white

65.8 55%/45% 91% white

64 57 48 30 26

36* 72* 43 18* 2*

*P = 0.001 vs. palliated group. Modified from Sohn TA, Lillemoe KD, Cameron JL, et al. Surgical palliation of unresectable periampullary adenocarcinoma in the 1990s. J Am Coll Surg 1999; 188:658.

or superior mesenteric arterial plexus.43 The pain is of low intensity, dull, and vaguely localized to the upper abdomen. In advanced disease, pain may be localized to the middle and upper back. The pain also may be postprandial and lead patients to reduce their caloric intake, a situation that ultimately results in weight loss or cachexia. Other nonspecific symptoms include nausea, fatigue, anorexia, and weight loss. These symptoms may or may not be caused by tumor involvement of the duodenum causing partial obstruction. Less common symptoms include those of diabetes and pancreatitis. New-onset diabetes mellitus may herald pancreatic cancer and can be observed in 6% to 68% of patients.44 Acute pancreatitis is occasionally the first manifestation of pancreatic cancer,45 and the clinician must keep this fact in mind especially when dealing with an older adult patient who presents with acute pancreatitis of unclear etiology.

DIAGNOSIS Computed Tomography

Although transabdominal ultrasonography is frequently the first modality used in many patients with pancreatic cancer (because 50% of them present with jaundice), the method of choice for diagnosis and staging of pancreatic cancer is CT.46-48 The pancreas is ideally imaged by means of the thinsection, pancreatic protocol, helical CT.49,50 In large series, a correct diagnosis of pancreatic cancer using CT can be made in up to 97% of patients.47 Refinements in the CT resolution of peripancreatic blood vessels has rendered routine angiography obsolete for the evaluation of suspected pancreatic masses.46,47 The pancreatic protocol CT consists of dual-phase scanning using intravenous and oral contrast agents. The first, arterial (pancreatic) phase is obtained 40 seconds after administration of intravenous contrast agent. At this time maximum enhancement of the normal pancreas is obtained, allowing identification of nonenhancing neoplastic lesions (Fig. 60-3A). The second, portal venous phase is obtained 70 seconds after injection of intravenous contrast agent and allows accurate detection of liver metastases and assessment of tumor involvement of the portal and mesenteric veins (see Fig. 60-3B).

Chapter 60  Tumors of the Pancreas

A

Arterial phase

B

Portal venous phase

Figure 60-3.  Pancreatic protocol computed tomography. A, Arterial phase showing a nonenhancing lesion in the head of the pancreas (arrows). B, Venous phase showing a noninvolved fat plane around the portal vein (arrows).

Current CT criteria for unresectability of a pancreatic tumor are as follows: (1) distant metastasis (e.g., liver, peritoneum, other), (2) arterial encasement of the celiac axis or superior mesenteric artery, and/or (3) occlusion of the portal vein or superior mesenteric vein.46-48 With these criteria, CT has been shown to be almost 100% accurate in predicting unresectable disease.46 However, approximately 25% to 50% of patients predicted to have resectable disease according to these CT criteria are found to have unresectable lesions at laparotomy.46-48 These patients clearly do not benefit from surgical exploration, and their identification by preoperative imaging remains a challenge. The most common causes of unresectability of a pancreatic tumor are small peritoneal or liver tumor implants and vascular involvement by tumor. The advent of helical pancreatic protocol CT has helped improve the preoperative determination of surgical resectability, particularly in relation to vessel invasion. Later studies have shown that assessment of the degree of circumferential vessel involvement by tumor can help predict unresectability.51-53 Other efforts aimed at detecting small peritoneal and liver metastases beyond the resolution of CT have focused on the development of staging laparoscopy, as discussed later.

Endoscopic Retrograde Cholangiopancreatography

Since its introduction in 1968, endoscopic retrograde cholangiopancreatography (ERCP) has become a mainstay in the differential diagnosis of various tumors of the periampullary region.54 The majority of these tumors originate from the pancreas (85%), and less commonly from the distal bile duct (6%), ampulla (4%), or duodenum (4%). ERCP allows visualization of the pancreatico-biliary tree to distinguish benign (stones) from malignant causes of obstruction. A “double-duct sign,” representing strictures in biliary and pancreatic ducts, is classically found in many patients with pancreatic cancer (Fig. 60-4). Tissue sampling can also be obtained during ERCP. For tumors of the ampulla or duodenum, biopsy of mucosal lesions is readily obtained with endoscopic forceps. Tumors of the distal bile duct may also be sampled via brush biopsy for routine cytology or genetic analysis.55,56 As discussed earlier, CT allows for identification of pancreatic tumors in the majority of patients with pancreatic cancer, rendering ERCP unnecessary in most cases. In prac-

Figure 60-4.  Endoscopic retrograde cholangiopancreatogram showing strictures of the biliary (open arrow) and pancreatic (closed arrow) ducts in pancreatic cancer. The bile duct is markedly dilated proximal to the stricture.

tice, however, many patients with pancreatic cancer undergo ERCP, not for the purpose of diagnosis but rather for stenting of the biliary duct (see Chapters 61 and 70). Routine preoperative biliary duct stenting to relieve jaundice has not been shown to decrease postoperative morbidity and mortality, and the procedure can increase the likelihood of surgical infectious complications.57-59 Therefore, its practice in patients with resectable tumors (as assessed by CT) cannot be recommended, unless it is anticipated that surgery will delayed for more than two weeks. For patients with jaundice and unresectable or metastatic disease, endoscopic biliary stenting, preferably with expandable metal stents, offers excellent palliation.60

1021

1022

Section VII  Pancreas

Figure 60-5.  Endoscopic ultrasonographic image of pancreatic cancer. The figure shows the needle during biopsy of the tumor.

using gadolinium contrast enhancement can demonstrate vascular involvement by tumor, obviating conventional angiography. Unlike CT, MRI does not involve radiation and uses an iodine-free contrast agent with rare renal toxicity. Limitations of MRI are related to cost, availability, and clinicians’ familiarity with and predilection for CT imaging. Magnetic resonance cholangiopancreatography (MRCP) can also be obtained at the time of MRI.68 In a prospective, controlled study, MRCP was found to be as sensitive as ERCP in detecting pancreatic carcinomas.69 MRCP uses heavy T2-weighted images that emphasize fluid-containing structures such as ducts, cysts, and peripancreatic fluid collections. Images obtained are highly comparable with those obtained with ERCP and readily demonstrate pancreatic ductal obstruction, ectasia, and calculi. In contrast to ERCP, MRCP is noninvasive and does not require injection of contrast into the pancreatico-biliary tree, avoiding possible complications such as allergy, pancreatitis, or infection. However, no therapeutic or diagnostic intervention can be performed with MRCP.

Positron Emission Tomography

Endoscopic Ultrasonography

EUS may be the most accurate test for the diagnosis of pancreatic cancer.61 Several studies have shown that EUS has a higher sensitivity and specificity than CT for detecting pancreatic masses (Fig. 60-5).62,63 EUS also has been found to be more accurate than CT in assessing vascular invasion and predicting tumor resectability. Other advantages of EUS include accurate assessment of peripancreatic nodal disease, and allowance of tumor biopsy by fine-needle aspiration (FNA).64 Limitations of EUS are manifold. EUS is highly operatordependent, and it is estimated that experience with 100 such examinations is needed to be considered proficient.63 Imaging by EUS can be compromised by the presence of a biliary stent, which results in imaging artifacts and loss of tissue detail. Due to technical and anatomic constraints, imaging of the portal vein and splenic vein is generally superior to imaging of the superior mesenteric artery and vein.65 For this reason EUS may lack accuracy when assessing vascular invasion at the level of the superior mesenteric vessels. Lastly, EUS provides no information regarding metastatic disease, and a complementary CT or magnetic resonance imaging (MRI) scan is required for complete staging of disease. Today EUS has a prominent role in the evaluation of patients with suspected pancreatic cancer, but is not necessary for all patients. In patients with stage IV disease, it is an excellent option to obtain tissue diagnosis. For those with borderline resectable tumors on CT, it allows for stratification of patients for surgical resection versus neoadjuvant treatment.

Magnetic Resonance Imaging

MRI has been increasingly used in the evaluation of pancreatic tumors, and several groups have shown results that rival those of helical CT.66,67 In one study, pancreatic tumor detection was reported in 90% of patients for MRI versus 76% for helical CT.66 Optimal MRI resolution is obtained with T1-weighted images and the use of dynamic gadolinium enhancement. Tumors are viewed as low-signal masses against the high-signal background of normal pancreatic parenchyma. Pancreatic masses, ductal dilation, and liver metastasis can be demonstrated in exquisite detail. Additionally, MR angiography and MR venography techniques

Positron emission tomography (PET) is a noninvasive imaging tool that provides metabolic rather than morphologic information on tumors.70,71 This diagnostic method is based on greater use of glucose by tumor cells than by normal pancreatic parenchyma. The radioactive glucose analog fluorodeoxyglucose (FDG) F 18 is administered intravenously, followed by detection of FDG uptake by the PET scanner. The normal pancreas is not usually visualized by FDG-PET. In contrast, pancreatic carcinoma appears as a focal area of increased uptake in the pancreatic bed. Hepatic metastases appear as “hot spots” within the liver. Owing to the lack of anatomic detail, PET scanning is not a principal diagnostic modality for pancreatic cancer. However, FDG-PET can be helpful in differentiating benign from malignant pancreatic masses when morphologic data are equivocal.72 It can also be useful in assessing tumor recurrence after pancreatic resection, when scar tissue or postoperative changes may be difficult to differentiate from carcinoma. Finally, FDG-PET can be of benefit in assessing tumor response to neoadjuvant chemoradiation, which may lead to alteration in clinical management.

Percutaneous and Endoscopic Ultrasonography–Guided Aspiration Cytology

FNA cytology of the pancreas has been one of the major advances in the management of patients with pancreatic tumors. CT-guided biopsy has been used for more than 20 years and is regarded as a safe, reliable procedure, with a reported sensitivity of 57% to 96% and virtually no falsepositive results.73,74 Experience with EUS-guided FNA shows similar results.64 Whenever a patient is deemed to have unresectable or metastatic pancreatic cancer, CT- or EUS-guided FNA biopsy is indicated for histologic confirmation of disease, unless a palliative surgical procedure is required. Even if the diagnosis of chronic pancreatitis is reasonably eliminated, proof of malignancy will exclude other rare benign diseases of the pancreas, such as tuberculosis and sarcoidosis. Furthermore, FNA cytology can usually distinguish between adenocarcinoma and other pancreatic tumors, such as neuroendocrine neoplasms and lymphomas, which carry a better prognosis (see Chapters 29 and 32).75 In the hands of experienced surgeons, tissue diagnosis is not a prerequisite to proceeding with surgery in most patients with potentially resectable tumors. One must also

Chapter 60  Tumors of the Pancreas keep in mind that a negative FNA result cannot exclude malignancy even after repeated sampling; in fact, it is the smaller (i.e., more curable) tumors that are most likely to be missed by the needle.

Serum Markers

A wide variety of tumor markers have been proposed for pancreatic cancer, but currently the only one with any practical usefulness is CA 19-9. Although not suitable for screening, this marker is a valuable adjunct in pancr­ eatic cancer for diagnosis, prognosis, and monitoring of treatment.76 For diagnosing pancreatic cancer, the sensitivity and specificity of CA 19-9 vary with the threshold value set. One of the major caveats of this tumor marker is that in the presence of jaundice, and especially with cholangitis, very high values can be found in the absence of malignancy (false-positive results). In addition, patients with a Lewis blood group phenotype (-a-b) do not express the CA 19-9 antigen, resulting in false-negative results. Despite these shortcomings, one study reported a sensitivity and specificity of 86% and 87%, respectively, using a cut-off value of 37 U/mL.77 Initial reports of CA 19-9 response to chemotherapy indicated that a decrease in value greater than 20% during treatment correlated with improved survival in patients with locally advanced or metastatic disease.78 A more contemporary study using a larger patient population did not confirm these findings. Based on these latest results, CA 19-9 response to chemotherapy cannot be used as an independent prognostic factor for survival.79

STAGING

The American Joint Committee on Cancer (AJCC) staging system for pancreatic cancer is shown in Table 60-5.80 The system was last revised in 2002, and modifications were made to better identify unresectable (T4, stages III and IV) from resectable disease (T1-3, stages I and II). Despite these changes, several limitations of the staging system still exist. First, adequate evaluation of lymph node status cannot be performed without surgical intervention; this drawback may lead to understaging of locally advanced disease in patients who are not candidates for laparotomy. Second, the margins of resection, which carry great prognostic significance, are not taken into consideration when assigning clinical stage. Because of these and other shortcomings, the AJCC staging system has found limited clinical applicability. Staging of pancreatic cancer is predicated on the identification of three distinct patient groups. The first group involves those presenting with metastatic disease. Surgery is best avoided in these patients in view of their short survival, and chemotherapy is their principal treatment modality other than palliative care measures.81 The second group comprises patients who have locally advanced (unresectable) disease but no metastases. These patients can benefit from neoadjuvant chemoradiation and, according to their treatment response, may be candidates for surgical exploration or intraoperative radiation therapy.82-85 A third group consists of patients with resectable disease for whom surgical resection should be advocated. A minimal staging workup for a patient with pancreatic cancer should include a physical exam, and a CT or MRI of the abdomen and pelvis.80 As discussed, contemporary CT or MRI is extremely accurate in identifying unresectable disease.46,47,66 In selected cases, EUS may complement CT scanning by allowing further assessment of vascular invasion and tissue sampling. However, CT imaging fails to

Table 60-5  TNM System and American Joint Committee on Cancer (AJCC) Staging of Pancreatic Cancer Tumor (T) TX T0 Tis T1

Primary tumor cannot be assessed No evidence of primary tumor Carcinoma in situ (includes PanIN III) Tumor limited to the pancreas, ≤2 cm in greatest dimension T2 Tumor limited to the pancreas, >2 cm in greatest dimension T3 Tumor extends beyond the pancreas but without involvement of the celiac axis or the superior mesenteric artery T4 Tumor involves the celiac axis or the superior mesenteric artery (unresectable primary tumor) Lymph Node Metastases (N) NX Regional lymph nodes cannot be assessed N0 No regional lymph node metastasis N1 Regional lymph node metastasis Distant Metastases (M) MX Distant metastasis cannot be assessed M0 No distant metastasis M1 Distant metastasis AJCC staging Stage 0 Tis Stage IA T1 Stage IB T2 Stage IIA T3 Stage IIB T1 T2 T3 Stage III T4 Stage IV Any T

N0 N0 N0 N0 N1 N1 N1 N1 Any N

M0 M0 M0 M0 M0 M0 M0 M0 M1

PanIN, pancreatic intraepithelial neoplasia. Reprinted from Greene FL, Page DL, Fleming ID, et al. AJCC Cancer Staging manual. 6th ed. New York: Springer-Verlag; 2002.

correctly predict resectability in 25% to 50% of patients. In most cases, lesions missed are beyond the resolution of current radiologic imaging; they include small implants on the peritoneal surfaces of the liver, abdominal wall, stomach, intestine, or omentum. Additionally, micrometastases only detectable by peritoneal washings are also missed. Successful detection of such tumor dissemination depends on access to the peritoneal cavity and visual inspection, which at present can be achieved only by laparoscopy or laparotomy. Several large studies have documented the value of staging laparoscopy in the evaluation of patients with pancreatic cancer.86-90 The staging procedure consists of a simple diagnostic laparoscopy with biopsy of suspicious nodules and collection of peritoneal washings for cytologic analysis. Peritoneal cytology results are important for staging because patients with occult metastases detected by this method have been shown to carry a prognosis similar to that of patients with M1 disease (stage IV).91,92 More complicated staging procedures involving extended laparoscopic dissections of the peripancreatic bed and laparoscopic ultrasonography can also be added if necessary.89,90 Published data demonstrate that approximately 25% of patients in whom localized disease is demonstrated by CT also have unsuspected metastatic implants.87,88 A further 10% have positive peritoneal cytology results despite a lack of gross metastatic disease at laparoscopy. For most of these patients, laparoscopy has prevented unnecessary surgical explorations to assess tumor resectability.93 The combina-

1023

Section VII  Pancreas Suspicion of pancreatic cancer

Hepaticojejunostomy

Helical CT

No tumor seen

Pancreatic head tumor <2 cm

ERCP and/or EUS

Pancreatic head Tumor of body or tumor >2 cm tail of the pancreas

Laparoscopy with cytology of washings

+

Surgical exploration for resection

–

Figure 60-6.  Massachusetts General Hospital algorithm for the diagnosis and staging of pancreatic cancer. CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasonography.

tion of CT scan and staging laparoscopy has enhanced the identification of patients with metastatic, localizedunresectable, and resectable pancreatic cancer and has helped in stratifying patients to different treatment protocols.88 Our current algorithm for the diagnosis and staging of pancreatic cancer is shown in Figure 60-6. We recommend laparoscopy for all patients with tumors in the body and tail of the pancreas (in which the frequency of unsuspected metastases approaches 50%) and for patients with tumors in the head of the pancreas larger than 2 cm, because the yield of laparoscopy in lesions smaller than 2 cm is less than 10%.

TREATMENT Surgical Therapy

Pancreaticojejunostomy

Pancreatic stent

Surgical resection is the only potentially curative treatment for pancreatic cancer. Because of advanced disease at presentation, only about 15% to 20% of patients are candidates for pancreatectomy.41 Extrapancreatic disease, most commonly presenting as malignant ascites or metastasis to the liver, peritoneum, or periaortic lymph nodes, is an absolute contraindication to resection. Relative contraindications to resection include encasement or occlusion of the superior mesenteric vein (SMV) and portal vein (PV), or direct extension of disease to the celiac axis, superior mesenteric artery (SMA), vena cava (VC), or aorta. Vascular resection and reconstruction may be applied in very selected cases to allow resection of locally advanced disease.94 The most common operation for pancreatic cancer is the Whipple pancreaticoduodenectomy (Fig. 60-7), which removes primarily the head of the pancreas.95 In the past, total pancreatectomy was advocated as a better operation for pancreatic cancer.96 It was thought to provide a more radical resection and allowed avoidance of the pancreaticojejunal anastomosis, which can be a source of considerable morbidity and mortality. However, total pancreatectomy has not been shown to improve survival when compared with the more limited pancreaticoduodenectomy, and results in obligate exocrine insufficiency and brittle diabetes, which are difficult to manage.97-99 Other extensions to the standard Whipple procedure, such as addition of retroperitoneal lymphadenectomy, have shown no significant survival benefit (Fig. 60-8) and may result in additional

Duodenojejunostomy Gastrostomy tube Figure 60-7.  Diagram of the pylorus-preserving pancreaticoduodenectomy. A pancreatic stent is shown in the pancreatic duct. (From Jimenez RE, Fernandez-del Castillo C, Rattner DW, et al. Outcome of the pancrea­ ticoduodenectomy with pylorus preservation or with antrectomy in the treatment of chronic pancreatitis. Ann Surg 2000; 231:293.)

1.0 .9 .8 .7 Proportion surviving

1024

.6 .5

Radical (n = 145)

.4 .3

Standard (n = 140)

.2 .1 0.0 0

12

24

36

48

60

Months Figure 60-8.  Actuarial survival after pancreaticoduodenectomy with standard and radical lymphadenectomy. (From Yeo CJ, Cameron JL, Lillemoe KD, et al. Pancreaticoduodenectomy with or without distal gastrectomy and extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma. Part 2: Randomized controlled trial evaluating survival, morbidity, and mortality. Ann Surg 2002; 236:355.)

morbidity (longer hospital stays, increased rates of pancreatic fistula, and higher incidence of delayed gastric emptying).100-102 In the past, pancreaticoduodenectomy was associated with high morbidity and mortality rates. Many contemporary large series now consistently show mortality rates of less than 3%, with a concomitant decrease in complications

Chapter 60  Tumors of the Pancreas Table 60-6  Published Contemporary Series of Pancreaticoduodenectomies reference

N

DATES

106 105 107 103 108 104

272 437 516 489 564 331

2001-2004 1988-2004 1980-2002 1990-2000 1984-1999 1993-1999

(Table 60-6).103-108 Pancreatic fistula, the most common and dreaded complication after the Whipple procedure, is observed in only 5% to 10% of patients today. These changes have been attributed to the emergence of intensive care units, as well as advances in surgical technique, anesthesia, antibiotics, and interventional radiology. Improved outcomes have also been linked to the development of centers of excellence for pancreatic surgery, where surgeons can develop greater expertise. A recent study using the Medicare database showed a fourfold increase in mortality when comparing pancreaticoduodenectomy performed in hospitals with an average of less than one case per year to those hospitals performing more than five cases per year.109 A subsequent study suggested that long-term outcomes are likewise improved at high-volume centers.110 Despite improved outcomes after pancreatectomy, surgery is still underused in the United States for treatment of pancreatic cancer. A study on early stage pancreatic cancer by Bilimoria and colleagues using the National Cancer Data Base from 1995 to 2004, identified that 38.2% of patients with potentially resectable disease were not offered surgery.111 Surgery was less likely to be offered to patients who were older than 65 years, black, of low socioeconomic status, less educated, or had cancer involving the head of the pancreas. The study also reported differences in outcomes between patients with stage I disease who underwent pancreatectomy versus those not offered surgery (median survival 19.3 months versus 8.4 months, respectively). These data raise awareness of nihilistic attitudes still linked to pancreatic surgery, and disparities in availability of health care today. This information also presents an opportunity for quality improvement in the future. Ultimately, prognosis for pancreatic cancer remains poor, even after potentially curative surgery in appropriately selected patients. Five-year actuarial survival rates range from 10.5% to 25% and median survivals between 10.5 and 20 months.108,112-114 Analysis of five-year survivors demonstrate an actual five-year survival rate of approximately 10%.115-117 Among those variables that have been shown by multivariate analysis to be significant predictors of a better outcome are tumor size less than 3 cm, absence of lymph node metastases, negative resection margins, welldifferentiated tumors, and intraoperative blood loss of less than 750 mL.108,112-114,118

Palliative Procedures

Large published series of patients with pancreatic cancer show that only between 15% and 20% are resectable. Patients with unresectable disease require some form of palliation to relieve jaundice, duodenal obstruction, or pain. Operations for biliary bypass are very effective, and are often combined with gastrojejunostomy to alleviate duodenal obstruction and celiac plexus block for pain control. However, these operations are not exempt from risk in debilitated patients with pancreatic cancer. A large series

MORTALITY rate (%)

PANCREATIC FISTULA (%)

1 1.1 3.9 1 2.3 2.1

4 12.6 9 11 5 2.1

shows postoperative mortality of 3.1% and a complication rate of 22%, with a median survival of 6.5 months.42 Relief of jaundice can also be achieved by biliary stents placed percutaneously or endoscopically (see Chapters 61 and 70). Because these procedures are usually well tolerated and perfomed on an outpatient basis, they have become increasingly popular in the management of malignant biliary obstruction. In experienced hands, endoscopic stent placement has a success rate of more than 85%, with a 1% to 2% procedure-related mortality. Several randomized trials have demonstrated no difference in survival between patients palliated endoscopically versus surgically for obstructive jaundice.119,120 Two types of biliary stents are available: plastic and selfexpandable metallic stents. The plastic stents are preferred for short-term use, and require exchange every three months to prevent complications from stent occlusion or cholangitis. The metallic self-expandable stents (Wallstent) have improved long-term patency rates when compared with plastic stents, and are preferred for long-term application.60,121,122 In clinical practice, plastic stents are preferred for patients who are surgical candidates, whereas metal stents are used for unresectable patients. Traditionally, duodenal obstruction has been treated surgically via gastrojejunostomy. The efficacy of this intervention has been demonstrated in the setting of randomized controlled trials.123,124 More recently, reports of the use of expandable metallic stents to relieve duodenal malignant obstruction have shown success, and this modality may be used increasingly in the future.125,126 Pain in pancreatic cancer can be extremely distressing and may respond poorly to oral narcotics. Percutaneous or surgical chemical neurolysis with alcohol is an alternative palliative measure that can help in controlling pain or decreasing narcotic use. EUS can be used for this purpose as well (see Chapter 61). Randomized trials have shown that neurolysis of the celiac ganglion can offer relief to many patients.127,128 Lastly, radiation therapy may also be used for pain management in selected patients.

Chemotherapy and Radiation Therapy

Adjuvant Chemoradiation From 1980 to 2000, adjuvant chemoradiation was standard of care in the United States and many other countries after potentially curative pancreatic cancer resection. These recommendations were based on the results of a study conducted by the Gastrointestinal Tumor Study Group (GITSG) between 1974 and 1982.129,130 The study randomized 43 patients after surgery to either observation or chemoradiation. Chemoradiation consisted of 4000 cGy of external beam radiation with concurrent bolus of 5-fluorouracil (5-FU) as radiosensitizer. Median survival in the treated group was 20 months, which was significantly longer than the 11-month survival in the untreated group. The GITSG study was criticized for its small sample size and lack of

1025

Section VII  Pancreas statistical power. However, a separate study by the Norwegian Pancreatic Cancer Trial group showed similar results and supported a survival benefit for adjuvant treatment.131 In 1999 the results of a study by the European Organization for Research and Treatment of Cancer (EORTC) questioned the value of adjuvant chemoradiation in pancreatic cancer.132 Similar to the study design by the GITSG, the EORTC randomized patients after surgery to observation or chemoradiation (4000 cGy external beam radiation and concurrent 5-FU by continuous infusion). This time the study group was relatively large and consisted of 114 patients with pancreatic cancer. The median survival was 4.5 months longer in the treatment group than in the observation group (17.1 vs. 12.6 months, respectively), but this difference was not statistically significant. Likewise, the projected two-year survival was not significantly different between groups (37 vs. 23 months, respectively). Further controversy ensued after the results of a study by the European Study Group for Pancreatic Cancer (ESPAC-1 trial) were released in 2001.133,134 This is the largest adjuvant treatment trial for pancreatic cancer to date, recruiting 289 patients. The ESPAC-1 study had a complicated design that randomized patients after surgery to four groups: observation, chemotherapy, chemoradiation, or chemoradiation plus chemotherapy. The results did not show a difference in median survival between patients receiving chemoradiotherapy and those who did not (15.5 vs. 16.1 months) (Fig. 60-9). Even for patients with positive resection margins who are thought to be prime candidates for chemoradiation, this treatment did not have a survival impact. Results of the ESPAC-1 trial are validated by its large sample size, but have generated great controversy. Criticism points mainly to the study’s convoluted randomization scheme, interrupted course of radiation, and pooling of data for statistical analysis. Thus, the best available data today are equivocal regarding the use of chemoradiation in adjuvant fashion outside

of clinical trials. New chemoradiation trials using gemcitabine instead of 5-FU have generated significant interest. A phase II trial by Small and colleagues showed a disease control rate of 84.6%, and a one-year survival rate of 47% for patients with unresectable disease.135 It remains to be seen whether these results hold ground in the setting of a phase III trial. Chemotherapy for Metastatic Disease Patients presenting with stage IV disease (see Table 60-5) as well as those developing distant metastasis after attempted curative resection are candidates for chemotherapy. Chemotherapy in this situation is never curative, and its palliative benefit must be balanced against its potential toxic side effects. Assessment of tumor response to chemotherapy is based primarily on serial imaging studies, serum marker (CA 19-9) trends, and changes in tumor-related symptoms. Only two chemotherapy agents have been associated with survivals longer than five months in pancreatic cancer: 5-FU and gemcitabine. Up until the mid-1990s, 5-FU was the drug of choice for treatment of advanced or metastatic pancreatic cancer. Monotherapy with 5-FU has a poor objective response rate (0% to 10%) and an average median survival of about five months.136 Combination chemotherapy regimens containing 5-FU have increased objective response rates (15% to 40%) but have failed to show a survival advantage in randomized trials.137-140 Gemcitabine is a deoxycytidine analog that is thought to interfere with DNA synthesis and repair. A well-known randomized trial by Burris and colleagues compared weekly gemcitabine to weekly 5-FU infusion in previously untreated patients.141 Patients treated with gemcitabine had a slightly improved median survival compared with those treated with 5-FU (5.6 vs. 4.4 months, respectively). An impressive one-year survival advantage was also noted for the gemcitabine arm (18% vs. 2%, respectively) (Fig. 60-10).

100

100

90 Median survival (mo)

75

Survival duration

80 70 % patients surviving

% patients surviving

1026

50

No chemoradiation

60

6 months 9 months 12 months

GEM

5-FU

5.65

4.41

46% 24% 18%

31% 6% 2%

50 40

GEM

30 20

25

5-FU

P = 0.0025

10

Chemoradiation

0 0

0 0

12

24

36

48

60

Months Figure 60-9.  Survival after pancreatic resection with or without adjuvant chemoradiation. (From Neoptolemos JP, Stocken DD, Friess H, et al. A randomized trial of chemoradiotherapy and chemotherapy after resection of pancreatic cancer. N Engl J Med 2004; 358:1200.)

2

4

6

8

10

12

14

16

18

20

Survival time (months) Figure 60-10.  Survival in 126 patients with metastatic pancreatic cancer treated with gemcitabine (GEM) or 5-fluorouracil (5-FU). (From Burris HA III, Moore MJ, Andersen J, et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: A randomized trial. J Clin Oncol 1997; 15:2403.)

Chapter 60  Tumors of the Pancreas Importantly, improvement in disease-related signs and symptoms (pain, performance status, weight loss) were more likely in the gemcitabine group (23.8%) than in the 5-FU group (4.8%). Gemcitabine has been shown to have clinical benefit even in patients for whom 5-FU treatment has failed.142 On the basis of the results of these studies, gemcitabine is currently the standard of care for patients with metastatic pancreatic cancer. However, single-agent gemcitabine treatment is associated with poor objective response rates (5.4%).141 In an attempt to improve response rates and survival, multiple gemcitabine-based combination chemotherapy regimens have been developed. In these protocols, gemcitabine has been coupled with either cisplatin, fluorouracil, irinotecan, oxaliplatin, pemetrexed, or exatecan.143-148 Objective response rates have shown improvement with combination therapy, but this improvement has not translated into a survival benefit. Poor results with conventional chemotherapy have led to ongoing development of novel agents against pancreatic cancer. As opposed to the nonselective action of current cytotoxic drugs, these new agents are designed to target specific cellular pathways involved in tumor progression. Examples include inhibitors of farnesyltransferase, epidermal growth factor receptor, and matrix metalloproteinases (Table 60-7).149,150 These new biological agents show impressive results in both in vitro and animal studies.151 However,

efficacy in human trials has been more difficult to demonstrate.149,150,152,153

CYSTIC TUMORS OF THE PANCREAS Cystic tumors of the pancreas are relatively uncommon, accounting for less than 10% of pancreatic neoplasms.154 A recent review of 24,000 abdominal CT and MRI studies at one institution during an 8-year period, revealed that pancreatic cysts were present in 1.2% of patients, and 60% of these were likely to be cystic neoplasms.155 MCNs, serous cystadenomas, and IPMNs comprise more than 80% of the primary cystic neoplasms of the pancreas (Table 60-8). Accurate recognition of these lesions is important because of their ability to masquerade as pancreatic pseudocysts, and their high cure rate following surgical treatment. Contemporary trends in pancreatic surgery reveal that cystic neoplasms account for an increasing number of resections.103 A summary of some of the distinguishing features of the most common cystic tumors of the pancreas can be seen in Table 60-8 and is discussed later.

DIFFERENTIAL DIAGNOSIS

Patient evaluation after discovery of a cystic lesion of the pancreas should initially be directed toward exclusion of a

Table 60-7  Targeted Therapeutic Agents in Pancreatic Cancer TARGETED MOLECULE

agent(S)

MECHANISM OF ACTION

Farnesyltransferase Epidermal growth factor receptor

Tipifarnib (Zarnestra) Cetuximab (Erbitux) Trastuzumab (Herceptin) Gefitinib (Iressa) Erlotinib (Tarceva) Marimastat BAY 12-9566

Inhibit mutant K-ras function Block cell membrane receptors for growth factors

Matrix metalloproteinase

Block extracellular matrix degradation and tumor invasion

Table 60-8  Epidemiologic and Biological Characteristics of Pancreatic Cystic Neoplasms TYPE

SEX PREDILECTION

PEAK DECADE OF LIFE

% OF CYSTIC NEOPLASMS

Serous cystadenoma

Female

7th

32-39

Mucinous cystic neoplasm

Female

5th

10-45

Intraductal papillary mucinous neoplasm

Equal distribution

6th-7th

21-33

Solid pseudopapillary neoplasm

Female

4th

<10

Cystic endocrine neoplasm

Equal distribution

5th-6th

<10

Ductal adenocarcinoma with cystic degeneration Acinar-cell cystadenocarcinoma

Slight male predominance Male

6th-7th

<1

6th-7th

<1

MALIGNANT POTENTIAL AND NATURAL HISTORY Resection is curative; serous cystadenocarcinoma is extremely rare Resection is curative, regardless of degree of epithelial dysplasia; poor prognosis when invasive adenocarcinoma is present Excellent prognosis for lesions showing only adenomatous and borderline cytologic atypia; poor prognosis when invasive adenocarcinoma is present Indolent neoplasm with rare nodal and extranodal metastases; excellent prognosis when completely resected Similar to that of solid neuroendocrine neoplasm (see Chapter 32) Dismal prognosis, similar to that of solid adenocarcinoma Similar to that of solid type; aggressive neoplasm with slightly better prognosis than ductal adenocarcinoma

Modified from Brugge WR, Lauwers GY, Sahani D, et al. Cystic neoplasms of the pancreas. N Engl J Med 2004; 351:1218-26.

1027

1028

Section VII  Pancreas pancreatic pseudocyst.154,156 As opposed to cystic neoplasms (see Fig. 59-7), pseudocysts lack an epithelial lining and represent collections of pancreatic secretions that have extravasated from a duct disrupted by inflammation or obstruction (see Chapters 58 and 59). Patients with pseudocysts often have a history of acute or chronic pancreatitis, or abdominal trauma, whereas most of those with cystic tumors lack such antecedent factors. Radiographic characteristics that favor a diagnosis of pseudocyst over cystic neoplasms include lack of septae, loculations, solid components, or cyst wall calcifications on CT or MRI. For pseudocysts, ERCP or MRCP traditionally demonstrates communication between the cyst and the main pancreatic duct. Evaluation of pseudocyst fluid reveals high levels of amylase, which is atypical for cystic tumors. If a diagnosis of pancreatic pseudocyst can be ruled out, evaluation should subsequently focus on identifying those tumors that require surgical resection due to actual or potential malignancy. As opposed to ductal adenocarcinoma, cystic neoplasms with malignant potential are slow growing, and favorable prognoses have been reported even in the setting of malignant degeneration. Tumors with malignant potential include MCNs, IPMNs, solid pseudopapillary tumors (SPTs), and cystic islet cell tumors. Serous cystadenomas, in contrast, are almost universally benign; they represent approximately one third of all pancreatic cystic neoplasms (see Table 60-8). The growing popularity of CT and MRI in the evaluation of abdominal complaints has led to increased identification

of cystic lesions in the pancreas as incidental findings.155,157 In addition to being asymptomatic, many of these cysts are also very small (<2 cm). Depending on the clinical scenario, it may be prudent to follow some of these patients nonoperatively.155,158 However, malignant cyst transformation is likely in patients older than 70 years, patients with new symptoms, or patients showing cyst growth on serial imaging.155,157,158 Nonoperative management is not adequate in these patients, and they should proceed to pancreatic resection if possible.

DIAGNOSTIC IMAGING

The diagnostic examination of choice is a CT or MRI of the abdomen with intravenous contrast enhancement, which enables tumor localization and, sometimes, discrimination between pseudocysts and cystic neoplasms. ERCP is useful for the diagnosis of IPMNs and allows procurement of biopsy samples by brushings. EUS allows detailed characterization of the cyst wall identifying fine structures such as septa, papilla, or wall nodules (Fig. 60-11). Additionally, FNA of the cyst contents is possible with EUS. Our group and others have demonstrated the value of cyst fluid analysis in the evaluation of cystic neoplasms (Table 60-9).154 Data also suggest that PET can be useful in the evaluation of cystic tumors of the pancreas. This modality detects metabolic activity within pancreatic cysts, which is believed to correlate with the presence or absence of malignancy. In an Italian study of 56 benign and malignant cystic tumors, PET had a sensitivity of 94% and a specificity of 97% in detect-

Table 60-9  Cyst Fluid Analysis in Cystic Lesions of the Pancreas PARAMETER

PSEUDOCYST

SEROUS CYSTADENOMA

MCN-BENIGN

MCN-MALIGNANT

IPMN

Viscosity Amylase CEA CA 72-4 Cytologic findings

Low High Low Low Histiocytes

Low Low Low Low Cuboidal cells with glycogenrich cytoplasm

High Low High Intermediate Columnar mucinous epithelial cells with variable atypia

High Low High High Adenocarcinoma cells

High High High Intermediate to high Columnar mucinous epithelial cells with variable atypia

CEA, carcinoembryonic antigen; IPMN, intraductal papillary mucinous neoplasm; MCN, mucinous cystic neoplasm.

Figure 60-11.  A mucinous cystic neoplasm of the tail of the pancreas. Left: The cyst is seen clearly on the computed tomography (CT) scan (arrow). Right: The endoscopic ultrasonography (EUS) image demonstrates septae and loculations within the cyst not clearly seen on CT. Arrows denote the neoplasm delineated by EUS. EUS allows for sampling of the cyst fluid.

Chapter 60  Tumors of the Pancreas ing malignant neoplasms.159 Although PET is still not frequently used in the evaluation of pancreatic cystic neoplasms, it may have a role in the nonsurgical follow-up of pancreatic cysts.

MUCINOUS CYSTIC NEOPLASMS

MCNs are the most frequently encountered cystic tumors of the pancreas, accounting for 10% to 45% of tumors. Most tumors consist of thick-walled cysts with occasional septations that are filled with thick mucous or hemorrhagic material (Fig. 60-12).160 MCNs display a clinical and histologic spectrum ranging from clearly benign to frankly malignant tumors. Current pathologic classification distinguishes between benign, borderline, or malignant (cystadenocarcinoma) tumors based on their maximal degree of dysplasia. This classification scheme correlates with patient prognosis and suggests that all these tumors should be treated as premalignant lesions with eventual evolution to aggressive behavior if left untreated.161 For many years, MCNs have been misidentified as IPMNs and vice versa. This has led to confusion regarding the true clinical and pathologic features of MCNs. The International Association of Pancreatology put forward guidelines to accurately differentiate an MCN from IPMN.162 These guidelines require the histologic presence of ovarian-type stroma within the tumor to establish the diagnosis of MCN. Using this criterion, several large series of patients with MCN have been published in the English literature and are used herein to describe their clinicopathologic features.163-166 MCNs occur almost exclusively in women, and are confined to the distal pancreas (body and tail of the gland). They are never multifocal, occurring only in one location within the pancreas. Mean age at presentation is 50 years. Most patients complain primarily of abdominal pain or a palpable mass. Today, with the liberal application of CT scanning in medical evaluations, an increasing percentage of tumors are being identified as incidental findings (approximately 25% of patients are asymptomatic). An MCN should be suspected when a CT or MRI of the abdomen shows a cyst within the body or tail of the pancreas in a middle-aged woman. If MRCP is done, there should be no communication between the pancreatic duct and the cyst itself. EUS is indicated when clinical and imaging characteristics deviate from this classic pattern. EUS can identify septations and cyst wall nodules in more detail than MRI or CT, and allows cyst wall biopsy and cyst fluid aspiration for analysis. Cyst fluid analysis generally

Figure 60-12.  Resection specimen of a mucinous cystic neoplasm of the tail of the pancreas. Note the multiloculated cyst.

reveals thick and mucoid material, low fluid amylase, elevated tumor markers (CEA), and mucinous epithelial cells by cytology. These findings may help in differentiating MCN from serous cystadenomas (low fluid CEA), and IPMN (high fluid amylase) (see Table 60-9). The pathologic classification of tumors in the most recent published series included benign adenomas (72%), borderline neoplasms (10.5%), carcinoma in situ (5.5%), and invasive cancer (12%).163 Similar distribution of tumors can be found in other published series where ovarian stroma was used as a screening criteria. Malignant MCNs were significantly larger than their benign counterparts (80 vs. 45 mm, respectively), and more likely to harbor nodules within their walls. Histologic and mutational analysis suggests that benign tumors can evolve into malignant ones given enough time.29 Owing to the inherent potential for malignancy in MCNs, surgical resection is advocated for all of them. Distal pancreatectomy with or without splenectomy is the procedure of choice, given that most tumors are located in the body or tail of the gland. A laparoscopic approach is acceptable as long as spillage of cyst contents is avoided, which can lead to intraperitoneal tumor spread. Excision of lymph nodes beyond those immediately adjacent to the pancreas is not necessary, given that lymph nodes metastases are rarely, if ever, observed. Five-year survival rates are excellent for benign or borderline MCNs, where the reported cure rate is 100%.161,163,166 Given that MCNs are never multifocal, long-term surveillance is not required for patients with resected noninvasive tumors. For invasive MCNs, five-year survival rates range from 30% to 63% in resected tumors.161,163-166 In our experience, unresectable malignant tumors carry as bad a pro­ gnosis as unresected ductal adenocarcinoma. There is no reported experience on the use of adjuvant treatment for MCN.

SEROUS CYSTADENOMAS

Serous cystadenomas, formerly known as microcystic adenomas, are the second most common cystic tumor of the pancreas.167 The clinical presentation of serous cystadenomas is similar to that of MCNs, occurring mostly in women (75%) with a mean age of 62 years.168 Most (50% to 70%) occur in the body or tail of the pancreas.169 An association with von Hippel-Lindau disease also has been noted. Most patients present with vague abdominal pain or discomfort, but a significant number can present with a palpable mass when the tumor has attained a large size (10 to 25 cm). Increasingly, incidental asymptomatic tumors are being detected during evaluation for other unrelated conditions. Macroscopically, serous cystadenomas consist of wellcircumscribed pancreatic neoplasms that on cross section show numerous tiny cysts separated by delicate fibrous septa, giving them a honeycomb appearance (Fig. 60-13).170 The cysts are filled with clear watery fluid and are often arranged around a central stellate scar that may be calcified. The pathognomonic image by CT scan is that of a spongy mass with a central “sunburst” calcification, but this finding occurs in only 10% of patients (Fig. 60-14A).154 EUS may allow better resolution of the honeycomb structure than CT (see Fig. 60-14B). Macrocystic variants are known that may be indistinguishable from MCNs.171,172 Cyst fluid analysis characteristically reveals low viscosity, low levels of CEA, and negative cytology (see Table 60-9). Unlike MCNs, serous cystadenomas are considered benign tumors. Rare case reports of serous cystadenocarcinomas exist but constitute less than 3% of known cases.173,174 Surgi-

1029

1030

Section VII  Pancreas

A

B

C Figure 60-13.  Serous cystadenoma of the tail of the pancreas. A, Most of the pancreatic parenchyma has been replaced by a cystic neoplasm. B, Cut surface shows multiple cysts and a central fibrotic scar. C, A high-power photomicrograph shows cysts containing serous fluid and lined by bland cuboidal cells rich in glycogen. (Hematoxylin and eosin.)

Cystic areas

Tumor

D2

D1

A

B

Figure 60-14.  Serous cystadenoma of the body and tail of the pancreas. A, Computed tomography appearance. Note the spongy appearance and the calcification. B, Honeycomb appearance on endoscopic ultrasonography of a 4.5-by-4.8-cm serous cystadenoma. (Courtesy of Michael Nunez, MD, Dallas, Tex.)

cal resection is the treatment of choice for symptomatic lesions. Options for resection depend on tumor location and include distal pancreatectomy with or without splenectomy, Whipple procedure, middle pancreatectomy, or enucleation. Serous cystadenomas may be safely observed if asymptomatic. Observation carries the risk of continued growth, which may lead to complications such as hemorrhage, obstructive jaundice, pancreatic insufficiency, or gastric outlet obstruction. Recent data suggest that tumor size at presentation may be used to identify patients who may

benefit from surgical resection.168 Tumors larger than 4 cm in maximum diameter at presentation were found to have faster growth rates than their smaller counterparts. These investigators recommend using this measurement to select patients for surgical treatment.

INTRADUCTAL PAPILLARY MUCINOUS NEOPLASMS

Since their initial description in 1982, hundreds of cases of intraductal papillary mucinous neoplasms have been

Chapter 60  Tumors of the Pancreas Table 60-10  Pathologic Spectrum of Intraductal Papillary Mucinous Neoplasms in Four Series IPMN—PATHOLOGIC CLASSIFICATION Adenoma Borderline Carcinoma in situ Invasive carcinoma   Positive lymph nodes*

HOPKINS179

MGH/VERONA177

MSKCC178

MAYO181

10% 18% 34% 38% (51)

12% 28% 18% 42% (41)

10% 15% 27% 48% (33)

29% 20% 15% 35% (30)

*Numbers in parentheses indicate percentage of patients with invasive carcinoma who had positive lymph nodes. IPMN, intraductal papillary mucinous neoplasm; MGH, Massachusetts General Hospital; MSKCC, Memorial Sloan-Kettering Cancer Center.

Figure 60-16.  Computed tomography scan of an intraductal papillary mucinous neoplasm (IPMN). The IPMN is affecting the head of the pancreas.

Figure 60-15.  Histologic view of a malignant intraductal papillary mucinous neoplasm. A papillary tumor is growing within the pancreatic duct. Note the surrounding lakes of mucin in the pancreas. At the top and right of the picture is the duodenum, which is focally invaded by the mucinous tumor. (Hematoxylin and eosin, low power.)

reported in the literature.175 Increased awareness about this disease and its differentiation from chronic pancreatitis, coupled with refinements in diagnostic imaging, have led to an explosion of “new” cases being recognized and reported. A variety of terms have been applied in reference to these tumors and include mucinous ductal ectasia, intraductal mucin-producing tumor, intraductal cystadenoma, pancreatic duct villous adenoma, and intraductal papillary neoplasm. Since 1996, however, both the World Health Organization and the Armed Forces Institute of Pathology have uniformly referred to this entity as IPMNs.24,176 IPMNs represent papillary neoplasms within the main pancreatic duct showing mucin hypersecretion that often leads to duct dilation and chronic obstructive pancreatitis (Fig. 60-15). IPMNs are considered premalignant pancreatic lesions, and histologically their epithelium may demonstrate areas ranging from hyperplasia to carcinoma within a single tumor. Based on their histologic characteristics, IPMNs are generally classified into the following groups: benign (adenoma), borderline, or malignant.176 The malignant group is further subclassified into noninvasive (carcinoma in situ) and invasive carcinoma based on extension

beyond the basement membrane. Malignant neoplasms account for 60% of IPMNs in recently published series, and two thirds of these showed invasive carcinoma (Table 60-10). Lymph node metastases are observed in 33% to 51% of invasive tumors.177-179 Clinically, IPMNs occur with equal frequency in men and women, with a median age at diagnosis of about 65 years. Approximately 75% of patients are symptomatic, with abdominal pain and weight loss being the most common complaints. A history of recurrent pancreatitis is given by 20% of patients, and acute pancreatitis is found in about 25% at presentation. No constellation of symptoms allows preoperative identification of the malignant variants of IPMNs. However, patients with malignant neoplasms are more likely to be older and more likely to present with jaundice or new-onset diabetes.177,178 Evaluation of patients with IPMNs includes CT or MRI of the abdomen and MRCP and ERCP. CT often demonstrates dilation of the pancreatic duct with or without an associated cystic mass. Duct dilation is often so impressive that it may mimic MCNs on CT imaging (Fig. 60-16). Evaluation by ERCP typically shows a patulous ampulla of Vater with extruding mucus, a finding that is often pathognomonic for IPMNs. Other findings during pancreatography include main duct dilation, filling defects due to viscid mucus or tumor nodules (Fig. 60-17), and communication between cystic areas and the main pancreatic duct. Adequate treatment of IPMNs requires pancreatic resection, which successfully relieves symptoms and prevents tumor progression to invasive carcinoma. Simpler proce-

1031

Section VII  Pancreas 1.0

Noninvasive (n = 80)

.9 Cumulative survival rate

1032

.8 .7 .6

P < 0.01

Invasive (n = 60)

.5 .4 .3 .2 .1 0 0

Figure 60-17.  Endoscopic retrograde cholangiopancreatogram showing an intraductal papillary mucinous neoplasm. Multiple filling defects are seen in the proximal pancreatic duct (short arrows). Bile duct obstruction, treated with a stent, can also be seen (long arrow).

12

24

36

48

60

72

84

96

108 120

Survival (months) Figure 60-18.  Actuarial survival in patients with invasive and noninvasive intraductal papillary mucinous neoplasms. (From Salvia R, Fernandez-del Castillo C, Bassi C, et al. Main-duct intraductal papillary mucinous neoplasms of the pancreas: Clinical predictors of malignancy and long-term survival following resection. Ann Surg 2004; 239:678.)

Table 60-11  Types of Surgical Resections for the Treatment of IPMN TYPE OF RESECTION Whipple Distal pancreatectomy Total pancreatectomy

HOPKINS179

MGH/VERONA177

MSKCC178

MAYO181

71% 12% 15%

63% 17% 19%

66% 23% 10%

50% 27% 23%

IPMN, intraductal papillary mucinous neoplasm; MGH, Massachusetts General Hospital; MSKCC, Memorial Sloan-Kettering Cancer Center.

dures such as sphincterotomy may partially treat symptomatology but do not address the malignant potential of these tumors. Pancreaticoduodenectomy is the treatment of choice for most patients given the predominance of IPMNs in the head of the pancreas, but distal pancreatectomy is indicated for lesions in the body or tail of the gland (Table 60-11). Intraoperative frozen sections are necessary to confirm negative resection margins and may mandate extended resection if positive. When IPMNs involve the entire ductal system, total pancreatectomy is the only curative surgical option. Prognosis after resection of IPMN is excellent, with five-year disease specific survival of 75% or better (Fig. 60-18).177-180 Factors associated with worse outcome include elevated bilirubin, an invasive IPMN, lymph node meta­ stases, and vascular invasion.178 Disease recurrence in the pancreatic remnant is rarely seen after resection of noninvasive IPMNs. Reported recurrences for invasive tumors have ranged from 12% to 65%, most occurring within 3 years of resection.177,178,181 These figures underscore the importance of long-term surveillance and postoperative imaging for those with invasive tumors. Patients with recurrent disease localized to the pancreas can benefit from completion pancreatectomy.177

SOLID PSEUDOPAPILLARY TUMORS

SPTs of the pancreas are rare neoplasms first described in 1934. They account for less than 10% of the cystic tumors of the pancreas.154 Although these tumors can occur in both genders, women are more frequently affected than men (10 : 1 ratio). In general, this is a disease of young women in their 30s, with very few cases (5%) reported in adults older than 50 years.182 Several pediatric series can be found in the literature, and these patients (age range 1 to 18 years) account for at least 20% of all known cases (Fig. 60-19).183 The most common clinical presentation is abdominal pain, found in approximately 50% of patients.182 The second most common finding is a large abdominal mass, which is the principal complaint in 35% of patients. At least 15% of tumors are discovered as incidental findings, often during routine physical exams or during evaluation after trauma. As for other cystic lesions of the pancreas, evaluation with CT (or MRI) coupled with biopsy via endoscopic ultrasound usually provides sufficient information to direct clinical management. Most pseudopapillary tumors are found in the body and tail of the pancreas (60%). The tumors can be quite large at presentation, with 34% of patients having masses larger than 10 cm in diameter. Grossly, small tumors are found to

Chapter 60  Tumors of the Pancreas 6%

22%

72%

>51 years 1-18 years 19-50 years Figure 60-19.  Age distribution of patients with solid pseudopapillary tumors. (From Papavramidis T, Papavramidis S. Solid pseudopapillary tumors of the pancreas: Review of 718 patients reported in English literature. J Am Coll Surg 2005; 200:965.)

be relatively solid, whereas larger variants show significant cystic degeneration. Microscopically, a mixture of solid, pseudopapillary, and hemorrhagic pseudocystic areas are observed. Uniform neoplastic tumor cells are seen separated by vascular hyalinized stroma.184 Although most SPTs demonstrate benign behavior, these tumors are considered to be lesions of uncertain malignant potential. A large meta-analysis of published patient series worldwide found a 20% frequency of solid-pseudopapillary carcinoma.182 Current criteria for malignancy include perineural invasion, angioinvasion, invasion of adjacent tissues, or metastases.184 The most common site of metastasis is the liver, and metastatic disease is observed in 5% to 10% of patients at presentation. Despite their female predominance, no association to estrogen receptors has been identified in SPTs. Other common mutated genes in pancreatic carcinogenesis such as K-ras and TP53 are also absent in these tumors. Instead, SPTs show abnormal expression of the E-cadherin protein. Although normal cells show E-cadherin expression within the cell membrane, 100% of SPTs show complete loss of E-cadherin expression in the cells or abnormal localization of E-cadherin to the cell nucleus.185 These findings are not related to mutations in the E-cadherin gene, but rather to abnormal expression of partner molecules such as β-catenin and p120.186 This finding is of diagnostic use, and loss of cell-to-cell adhesion may help to explain the cystic degeneration observed in this neoplasm. Overall, SPTs are very slow-growing neoplasms. Complete resection is the treatment of choice, and resection of synchronous metastases is also recommended when possible.187 Prognosis is usually excellent with five-year survival around 95%. Prolonged survival has also been reported even in the presence of metastatic disease.182,188 No significant data are available about adjuvant treatment for SPTs.

OTHER NONENDOCRINE PANCREATIC TUMORS Other nonendocrine tumors of the pancreas include acinar cell carcinomas, lymphomas, and sarcomas. Acinar cell carcinomas are extremely rare, representing 1% to 2% of

pancreatic tumors. Clinical presentation may be indistinguishable from pancreatic ductal adenocarcinoma, and some variants of acinar cell carcinomas can show cystic components. A minority of patients (10% to 15%) may present with the lipase hypersecretion syndrome, with high serum lipase levels and peripheral fat necrosis observed clinically.189 The lipase hypersecretion syndrome is often present in the setting of liver metastasis and is considered an unfavourable prognostic factor. Overall, acinar cell carcinomas are considered an aggressive malignancy, and 50% of cases present with liver metastasis. However, prognosis appears to be better than for ductal adenocarcinoma. In a recent study of acinar cell carcinomas using the Surveillance, Epidemiology, and End Results (SEER) database, fiveyear survival was 22% for patients with unresectable disease and 72% for those with resectable disease.190 Pancreatic lymphoma (see also Chapter 29) represents less than 1% to 2% of all pancreatic malignancies, and less than 1% of all extranodal non-Hodgkin’s lymphomas.191 Whenever a large mass is identified in the pancreas without biliary obstruction, pain, or weight loss, the diagnosis should be contemplated. An elevated serum lactate dehydrogenase level supports a diagnosis of lymphoma. FNA biopsy with flow cytometry is highly accurate in establishing the diagnosis.192 Identification of pancreatic lymphoma is important because treatment is primarily nonsurgical, and prognosis is much better than for pancreatic adenocarcinoma. Treatment usually consists of a combination of chemotherapy and radiation therapy, and cure rates near 30% are reported in the literature.192 A few patients with small tumors, thought to represent carcinomas, have been treated with surgery alone and have had excellent survival.191 Pancreatic endocrine tumors are discussed in Chapter 32.

KEY REFERENCES

Balcom JH, Rattner DW, Warshaw AL, et al. Ten-year experience with 733 pancreatic resections: Changing indications, older patients, and decreasing length of hospitalization. Arch Surg 2001; 136:391-8. (Ref 103.) Bilimoria KY, Bentrem DJ, Ko CY, et al. National failure to operate on early stage pancreatic cancer. Ann Surg 2007; 246:173-80. (Ref 111.) Birkmeyer JD, Finlayson SR, Tosteson AN, et al. Effect of hospital volume on in-hospital mortality with pancreaticoduodenectomy. Surgery 1999; 125:250-6. (Ref 109.) Brugge WR, Lauwers GY, Sahani D, et al. Cystic neoplasms of the pancreas. N Engl J Med 2004; 351:1218-26. (Ref 154.) Burris HA, III, Moore MJ, Andersen J, et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: A randomized trial. J Clin Oncol 1997; 15:2403-13. (Ref 141.) Crippa S, Salvia R, Warshaw AL, et al. Mucinous cystic neoplasm of the pancreas is not an aggressive entity: Lessons from 163 resected patients. Ann Surg 2008; 247:571-9. (Ref 163.) Ductal adenocarcinoma. In: Hruban RH, Pitman MB, Klimstra DS, editors. AFIP Atlas of Tumor Pathology. Tumors of the Pancreas. Washington, D.C.: American Registry of Pathology; 2007. pp 111-64. (Ref 22.) Gress FG, Hawes RH, Savides TJ, et al. Role of EUS in the preoperative staging of pancreatic cancer: A large single-center experience. Gastrointest Endosc 1999; 50:786-91. (Ref 63.) Hezel AF, Kimmelman AC, Stanger BZ, et al. Genetics and biology of pancreatic ductal adenocarcinoma. Genes Dev 2006; 20:1218-49. (Ref 28.) Jimenez RE, Warshaw AL, Rattner DW, et al. Impact of laparoscopic staging in the treatment of pancreatic cancer. Arch Surg 2000; 135:409-14. (Ref 88.) Lowenfels AB, Maisonneuve P, DiMagno EP, et al. Hereditary pancreatitis and the risk of pancreatic cancer. International Hereditary Pancreatitis Study Group. J Natl Cancer Inst 1997; 89:442-6. (Ref 3.) Neoptolemos JP, Stocken DD, Friess H, et al. A randomized trial of chemoradiotherapy and chemotherapy after resection of pancreatic cancer. N Engl J Med 2004; 350:1200-10. (Ref 134.)

1033

1034

Section VII  Pancreas O’Malley ME, Boland GW, Wood BJ, et al. Adenocarcinoma of the head of the pancreas: Determination of surgical unresectability with thinsection pancreatic-phase helical CT. AJR Am J Roentgenol 1999; 173:1513-18. (Ref 52.) Salvia R, Fernandez-del Castillo C, Bassi C, et al. Main-duct intraductal papillary mucinous neoplasms of the pancreas: Clinical predictors of malignancy and long-term survival following resection. Ann Surg 2004; 239:678-85. (Ref 177.)

Yeo CJ, Cameron JL, Lillemoe KD, et al. Pancreaticoduodenectomy with or without distal gastrectomy and extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma, part 2: Randomized controlled trial evaluating survival, morbidity, and mortality. Ann Surg 2002; 236:355-66. (Ref 102.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

61 Endoscopic Treatment of Pancreatic Disease Lincoln E. Ferreira and Todd H. Baron

CHAPTER OUTLINE Early Acute Pancreatitis  1035 Local Complications of Acute Pancreatitis  1035 Acute Pancreatic Pseudocyst  1036 Organized Pancreatic Necrosis (Walled-Off Pancreatic Necrosis)  1037 Pancreatic Abscess  1038 Complications of Endoscopic Therapy of Pancreatic Fluid Collections  1038 Recurrent Acute Pancreatitis  1038

Since the first report of endoscopic retrograde cholan­ giopancreatography (ERCP) and endoscopic biliary sphinc­ terotomy in 1974 there have been numerous advances in ERCP techniques. Use of ERCP continues to evolve in the treatment of pancreatic disease. Less invasive diagnostic modalities including endoscopic ultrasonography (EUS), computed tomography (CT), and magnetic resonance chol­ angiopancreatography (MRCP) have largely replaced diag­ nostic ERCP. However, ERCP endoscopic therapy remains useful for the treatment of pancreatic diseases. This chapter reviews the endoscopic treatment of acute pancreatitis and its complications, as well as the endoscopic treatment of recurrent acute pancreatitis, chronic pancreatitis, pan­creatic cancer, and pancreatic cysts.

EARLY ACUTE PANCREATITIS When a patient presents with acute pancreatitis the role of endoscopy is limited to two situations: first, those patients with gallstone-induced pancreatitis (see Chapters 58 and 66) and second, to provide nutritional support via enteric feeding (see Chapter 5). Gallstone pancreatitis is caused by impaction of a stone within the common channel of the ampulla of Vater, usually transiently. ERCP and biliary sphincterotomy are used to improve the outcome of gall­ stone pancreatitis by removal of an impacted stone with relief of pancreatic ductal obstruction. Initial studies of patients with acute gallstone pancreatitis and choledocho­ lithiasis used urgent (within 72 hours of admission) ERCP and sphincterotomy (if a stone was identified). An improved outcome was seen only in patients with clinically severe acute pancreatitis.1 Studies now suggest that the improved outcome following ERCP and sphincterotomy in gallstone pancreatitis results from reduced biliary sepsis rather than improvement in pancreatitis.2,3 A meta-analysis showed that early ERCP in patients with predicted mild or severe acute

Chronic Pancreatitis  1039 Pancreatic Ductal Endotherapy  1039 Drainage of Chronic Pancreatic Pseudocysts  1039 Biliary Strictures  1039 Refractory Pain  1039 Pancreatic Duct Leaks  1040 Pancreatic Cancer  1041 Pancreatic Cysts  1041

biliary pancreatitis without concomitant acute cholangitis did not significantly reduce overall complications and mor­ tality.4 Similarly, results from a prospective randomized trial showed that ERCP and sphincterotomy could decrease morbidity in patients with gallstones pancreatitis and ampullary obstruction when the duration of obstruction did not exceed 28 hours.5 ERCP in patients with severe gallstone acute pancreatitis is best reserved for patients with sus­ pected biliary obstruction, based on hyperbilirubinemia and evidence of clinical cholangitis because it is unlikely that the ampulla is obstructed in the presence of a normal serum bilirubin.6,7 Other imaging modalities in patients with severe biliary pancreatitis such as EUS and MRCP can help select patients for ERCP when bile duct stones are identified.8,9 If bile duct stones are not identified during ERCP performed for acute gallstone pancreatitis, there are no data to guide whether an empirical biliary sphincterotomy should be performed. However, sphincterotomy can reduce the risk of recurrent acute pancreatitis and cholangitis prior to cholecystectomy.10 Evidence supports enteral feeding for patients with severe acute pancreatitis based on randomized prospective studies comparing total parenteral nutrition with enteral feeding (through a nasoenteric feeding tube placed beyond the liga­ ment of Treitz) instituted within 48 hours of illness onset.11 Lower cost and fewer infectious complications are seen with enteral feeding (see Chapter 5). There are a variety of endoscopic techniques for placing nasojejunal feeding tubes in the setting of acute pancreatitis12 including transnasal endoscopy.13

LOCAL COMPLICATIONS OF ACUTE PANCREATITIS Acute fluid collections, acute pancreatic pseudocysts, orga­ nized pancreatic necrosis, and pancreatic abscesses may arise as a result of acute pancreatitis.14 Acute fluid collec­

1035

1036

Section VII  Pancreas

Figure 61-1.  Illustration of the mechanism of formation of an acute pancreatic pseudocyst. Limited necrosis of the main pancreatic duct produces a leak with accumulation of amylase-rich fluid.

tions form early in the course of acute pancreatitis and usually resolve without therapy. Acute pseudocysts arise as a sequela of acute pancreatitis, require at least four weeks to form, and are devoid of significant solid debris. Acute pancreatic pseudocysts usually form as a result of limited pancreatic necrosis that produces a pancreatic ductal leak (Fig. 61-1). Alternatively, areas of pancreatic and peripan­ creatic fat necrosis may liquefy over time and become a pseudocyst.15 Despite the requirement of at least four weeks for a pseudocyst to form, it is important to realize that some patients with significant early acute pancreatic necrosis (>30%) may evolve the pancreatic and peripancreatic necro­ sis into a collection that radiographically resembles a pseu­ docyst.16 These collections contain significant solid debris, and endoscopic treatment of them using typical pseudocyst drainage methods often results in infectious complications because of contamination and inadequate removal of solid debris.17,18

A

B Figure 61-2.  Illustration of transpapillary drainage of a pancreatic pseudocyst. A, Pancreatogram shows a leak off a side branch of the main pancreatic duct (arrows). B, A pancreatic duct stent is in place across the leak.

ACUTE PANCREATIC PSEUDOCYST

Drainage of an acute pseudocyst is indicated for treatment of symptomatic pseudocysts that may or may not be infected19 and for progressive enlargement on imaging studies. Symptoms and signs from an acute pseudocyst include abdominal pain, often exacerbated by eating, weight loss, gastric outlet obstruction, obstructive jaundice, and pancreatic duct leakage, which may result in pancreatic ascites or pancreatic fistulae.20 Pseudocysts may be drained through the papilla (transpapillary), through the gastric or duodenal wall (transmurally), or by using a combination of the two.

Transpapillary Drainage

If the pseudocyst communicates with the main pancreatic duct, placement of a pancreatic duct stent with or without pancreatic sphincterotomy is effective, especially for smaller pseudocysts (<5 to 6 cm) that are not otherwise approach­ able transmurally.21 The proximal end of the stent (toward the pancreatic tail) may directly enter the pseudocyst, bridge the area of leak into the pancreatic duct upstream from the leak, or lie completely downstream to the leak. Bridging the leak is the preferred approach because it restores ductal continuity and appears to be more effective (Fig. 61-2).22,23 Transpapillary drainage avoids bleeding or perforation that may occur with transmural drainage. However, pancreatic stents may induce scarring of the main pancreatic duct.24

Transmural Drainage

There is no standardized approach to transmural pseudo­ cyst drainage. Transmural drainage is performed by entering the cyst using a needle without cautery or using a cautery device (e.g., needle knife). Some endoscopists believe EUS-guided drainage is mandatory prior to performing endoscopic transmural drainage to prevent bleeding and perforation.25 Although the superiority of EUS-guided versus non–EUS-guided drainage has not been demon­ strated,26 there are increasing data to support its routine use during transmural drainage,26-28 especially when non–EUSguided drainage fails.29 EUS-guided entry is successful in more than 95% of patients and with low complication rates.30-32 Non–EUS-guided entry also can be performed33 and, in the hands of experienced operators, successful trans­ mural entry has been reported in 91 of 94 patients in lesions as small as 3 cm and without endoscopically visible extrin­ sic compression.34 Once the pseudocyst is successfully entered, the transmural tract is balloon dilated to 8 to 10 mm in diameter to allow placement of one or two 10-French stents (Fig. 61-3).35 Following uncomplicated attempted endoscopic drain­ age, a follow-up CT is obtained four to six weeks after the procedure. The internal stents are endoscopically removed after documented radiographic pseudocyst resolution. Success rates, recurrence rates, and complication rates of

Chapter 61  Endoscopic Treatment of Pancreatic Disease

A

A

B Figure 61-4.  A, Illustration of walled-off (organized) pancreatic necrosis. B, Computed tomography scan of patient with organized pancreatic necrosis. Successful endoscopic therapy using direct necrosectomy was performed (see Fig. 61-5).

B Figure 61-3.  Transduodenal drainage of a pancreatic pseudocyst A, The pseudocyst was entered and a guidewire coiled into the cyst. A dilating balloon is inflated across the duodenal wall. B, Two double pigtail stents are placed into the cyst.

endoscopic drainage of pancreatic pseudocysts are variable, likely because of many reports included acute and chronic pseudocysts and pancreatic abscesses. In addition, some patients underwent transpapillary drainage and others underwent transmural drainage. Nonetheless, cumulatively successful drainage is achieved in approximately 75% to 90%, with complication rates of about 5% to 10% and pseudocyst recurrence rates of 5% to 20%.36,37

ORGANIZED PANCREATIC NECROSIS (WALLED-OFF PANCREATIC NECROSIS)

Pancreatic necrosis is nonviable pancreatic parenchyma, usually associated with peripancreatic fat necrosis. In the earliest form, this is detected on contrast-enhanced CT by demonstrating areas of nonenhancing pancreatic paren­ chyma. Pancreatic necrosis is frequently accompanied by major pancreatic ductal disruptions. Over the course of several weeks, the collection may continue to evolve and expand the initial area of necrosis and contains both liquid and solid debris (Fig. 61-4). The terms organized pancreatic necrosis20 and walled-off pancreatic necrosis (WOPN)38 have been used to differentiate this process from the early (acute phase) of pancreatic necrosis. The CT appearance of organized pancreatic necrosis may be mistaken as an acute pseudocyst.16 The indications for and timing of drainage of sterile WOPN are controversial. Endoscopic drainage cannot be performed until the process becomes organized, which

1037

1038

Section VII  Pancreas usually occurs several weeks after onset of pancreatitis. Indications for drainage of sterile WOPN are refractory abdominal pain, gastric outlet obstruction or failure to thrive (continued systemic illness, anorexia, and weight loss) at four or more weeks after the onset of acute pancre­ atitis.38 Because endoscopic drainage of WOPN is more tech­ nically difficult, carries a higher rate of complications, and tends to involve more severely ill patients than patients with acute pseudocyst, the decision to endoscopically inter­ vene when the process is sterile pancreatic necrosis must be carefully considered.39 Alternative management options to endoscopic drainage include nutritional support with parenteral or enteral jejunal feeding and nonendoscopic drainage methods such as percutaneous and surgical drain­ age. The management option selected is usually based on local expertise and severity of comorbid medical illnesses. Ideally, these patients are best managed by a multidisci­ plinary approach. Infected necrosis is an indication for drainage. Percutane­ ous fine-needle aspiration may be required to determine the bacteriologic status prior to intervention.40 Because of the need to evacuate solid material, the endo­ scopic approach to drainage of WOPN differs from drainage of pseudocysts. In general, the transpapillary approach alone is not adequate because it does not allow removal of solid debris. Therefore, the transmural drainage approach is used. The endoscopic approach has evolved. Initially nasal irrigation tubes were placed of alongside transmurally placed stents in order to lavage the solid debris.41 Sub­ sequent approaches used percutaneous endoscopic gastros­ tomy (PEG) tubes to provide a method of placing irrigation tubes into the necrotic cavity and avoid uncomfortable transnasal tubes.42,43 The most recent approach for removal of necrotic debris is to perform direct endoscopic debridement. This is per­ formed by dilating the transmural tract with large-caliber balloons (up to 20 mm) and passing a forward-viewing endoscope through the tract directly into the necrotic cavity (Fig. 61-5).44,45 Snares, grasping forceps, and other accesso­ ries are then be used to remove solid debris. This approach has been shown in a retrospective study to be superior to the irrigation approach.46 Nevertheless, repeat procedures are often needed to re-dilate the transmural tract, to debride residual necrotic material, and to attempt to treat underly­ ing pancreatic ductal disruptions in hopes of preventing a ductal disconnection. Drainage of pancreatic necrosis is associated with a higher complication rate and longer hospital stay.27,47 Patients with acute pseudocysts tend to have less severe ductal abnor­ malities and fewer recurrences.

PANCREATIC ABSCESS

When a broad definition of pancreatic abscess is taken to include infected pseudocysts or infected liquefied collec­ tions without significant solid debris (pancreatic necrosis), success rates following endoscopic drainage are high, although there are few series with small numbers of patient.47-49

COMPLICATIONS OF ENDOSCOPIC THERAPY OF PANCREATIC FLUID COLLECTIONS

Operator experience likely plays a role in the outcome following endoscopic drainage.50 Life-threatening bleeding or perforation may arise following attempted endoscopic drainage of pancreatic fluid collections. Infectious compli­ cations usually occur from inadequate drainage of fluid and

A

B Figure 61-5.  Direct endoscopic necrosectomy performed in the patient depicted in Figure 61-4. A, Endoscopic view from inside the necrotic cavity; an indwelling pigtail stent is seen with surrounding necrotic debris. B, Necrotic solid material being withdrawn from the cavity through the posterior gastric wall with a snare.

solid debris. Infection can usually be managed by additional endoscopic procedures and placement of percutaneous drains. Endoscopic therapy followed by complications may adversely alter the surgical outcome as compared to patients undergoing primary surgical therapy.51,52

RECURRENT ACUTE PANCREATITIS Recurrent acute pancreatitis presents a diagnostic and therapeutic challenge. In about 10% to 22% of such cases, it is not possible to establish the etiology of the disease.53 Multiple factors discussed in other chapters may be involved with episodes of recurrent acute pancreatitis including alcohol use, microlithiasis, sphincter of Oddi dysfunction (SOD), pancreas divisum, hereditary pancreatitis, gene mutations, cystic fibrosis, choledochocele, annular pan­ creas, anomalous pancreaticobiliary junction, ampullary

Chapter 61  Endoscopic Treatment of Pancreatic Disease lesions, pancreatic tumors, and autoimmune pancreatitis (AIP).49 EUS is considered an important tool for evaluating for possible causes of recurrent acute pancreatitis.54 EUS with high resolution allows detection of gallstones, sludge, and microlithiasis55 that can direct treatment toward cholecys­ tectomy or ERCP. In addition, a diagnosis of AIP can be suspected based on ultrasonographic findings and con­ firmed by EUS-FNA or Tru-cut biopsies.56 In patients with otherwise unexplained recurrent acute pancreatitis after exhaustive clinical and laboratory evaluation and EUS, measurement of pancreatic sphincter pressure by ERCP with pancreaticobiliary manometry can allow the diagnosis of pancreatic sphincter of Oddi dysfunction.57 Subsequent biliary and pancreatic sphincterotomy (when elevated sphincter pressures are identified) may prevent recurrent attacks of pancreatitis. Pancreas divisum is present in approximately 10% of the population (see Chapter 55) and results from failure of the dorsal and ventral pancreatic ducts to fuse.58 The role of pancreas divisum as a cause of pancreatitis is controversial, but it is believed that in a subset of patients the minor papilla produces functional obstruction to the flow of pan­ creatic secretions. Pancreas divisum can be diagnosed by CT, magnetic resonance imaging (MRI), or EUS. Secretin MRCP may predict which patients have functional minor papilla obstruction.59 Pancreas divisum is confirmed by cannulation of the minor papilla, which can be facilitated by intravenous administration of secretin. Endoscopic minor papilla sphincterotomy in properly selected patients without extensive changes of chronic pancreatitis can reduce or prevent further attacks of acute pancreatitis.60-62

CHRONIC PANCREATITIS A variety of endoscopic interventions can be performed in patients with chronic pancreatitis (Table 61-1).63

PANCREATIC DUCTAL ENDOTHERAPY

Pancreatic duct strictures and pancreatic duct stones fre­ quently coexist, cause obstruction to the main pancreatic duct, and may contribute to abdominal pain and episodes of acute pancreatitis superimposed on chronic pancreatitis. Endoscopic therapy of pancreatic duct strictures is per­ formed with balloon or catheter dilation followed by place­ ment of one or more plastic pancreatic stents (Fig. 61-6).64 Stents are exchanged and remain in place for a variable period of time. Pancreatic sphincterotomy and stone removal can rarely be successfully performed using standard biliary stone removal techniques because the stones are calcified and usually impacted within side branches and pancreatic duct strictures. Extracorporeal shock-wave lithotripsy (ESWL),

Table 61-1  Endoscopic Therapies for Chronic Pancreatitis Endoscopic pancreatic sphincterotomy Extraction of pancreatic duct stones Dilation of pancreatic duct strictures Dilation of biliary strictures Treatment of pancreatic duct leaks resulting in ascites or pleural effusions Drainage of pancreatic pseudocysts Placement of biliary and pancreatic stents

if available, can be used to fragment stones prior to or without endoscopic removal.65-67 Intraductal lithotripsy under pancreatoscopic guidance has also been used to fragment and remove obstructing stones.68 Pancreatic endo­ therapy has a high rate of initial technical success.69 In some series long-term clinical success has been achieved in two thirds of patients without need for surgery and with a significant reduction in annual rate of hospitalizations for pain.65,70 Two prospective randomized studies comparing endo­ therapy and surgery for patients with painful obstructive chronic pancreatitis have been performed.71,72 Endotherapy was associated with a higher number of procedures, and pain improvement was seen more often in the surgical groups. There were limitations to endotherapy in both of these studies, however, namely, lack of availability of ESWL in one study and less aggressive endoscopic stricture therapy in another.73 In cases in which transpapillary approaches are not possible for treatment of pancreatic duct obstruction because of impassable stones or strictures within the pancreatic head, EUS-guided transgastric pancreatic duct puncture can be performed. This can facilitate a rendezvous procedure74 or provide transgastric or transduodenal ductal stent placement.75

DRAINAGE OF CHRONIC PANCREATIC PSEUDOCYSTS

The mechanism of formation of a chronic pseudocyst is different than an acute pseudocyst. As discussed in Chapter 59, chronic pseudocysts arise as a sequela of chronic pan­ creatitis and downstream pancreatic ductal obstruction from fibrotic strictures or stones (Fig. 61-7).76 This results in a pancreatic ductal blowout (leak) and accumulation of pan­ creatic fluid. These collections do not contain solid debris and usually do not arise as a result of acute inflammatory processes. The endoscopic approach to chronic pancreatic pseudocysts is as described for acute pseudocysts earlier. The main difference is that the underlying ductal abnor­ malities may lead to recurrences if left untreated.77,78

BILIARY STRICTURES

The fibrosing process within the pancreatic head can encase the distal bile duct and result in formation of a biliary stric­ ture. Possible sequelae include hepatic fibrosis (secondary biliary cirrhosis) and cholangitis. Balloon dilation and endoscopic insertion of multiple plastic stents across the biliary stricture may result in nonsurgical resolution.79-81 Patients with noncalcific pancreatitis appear to respond better to biliary stricture dilation than those with calcific pancreatitis. Endoscopic insertion of biliary stents can be used for treatment of benign biliary obstruction due to chronic pancreatitis in several situations: (1) preoperative placement for relief of jaundice or cholangitis, (2) temporary placement when biliary obstruction occurs following recent acute pancreatitis superimposed on chronic pancreatitis, (3) long-term therapy of refractory strictures.79 Placement of covered self-expandable metal stents with removal at 3 months has resulted in resolution of biliary strictures that were due to chronic pancreatitis in nearly 80% of patients at a minimal follow-up of 12 months.82

REFRACTORY PAIN

Patients with refractory pain from chronic pancreatitis may benefit from EUS-guided celiac plexus block. Unfortunately, this effect is short-lived and requires additional therapy.83,84

1039

1040

Section VII  Pancreas

A

B

D C Figure 61-6.  Endoscopic treatment of a pancreatic duct stricture. A, Initial pancreatogram shows stricture in the head of the pancreas (arrow) with upstream ductal dilatation. B, Balloon dilation of the stricture (note waist). C, Placement of two side-by-side plastic stents. D, Follow-up pancreatogram shows improvement in the stricture.

PANCREATIC DUCT LEAKS

Figure 61-7.  Illustration of the mechanism of formation of a chronic pancreatic pseudocyst. Obstruction of the main pancreatic duct results in an upstream leak and pseudocyst formation.

Pancreatic duct leaks and pancreatic duct disruptions may occur as a sequela of acute or chronic pancreatitis (see Chapters 58 and 59) as well as after pancreatic surgery (Fig. 61-8) and trauma. Leaks can arise from the tail, body, or head of the pancreas. Fluid can then track laterally (toward the spleen or into the abdomen), medially toward the duodenum or bile duct, centrally into the lesser sac, or into the mediastinum or abdomen with resultant pleural effusions and ascites, respectively. External leaks (pancreatico-cutaneous fistulas) usually occur following pancreatic surgery and percutaneous drain­ age of pancreatic fluid collections. External fistulae are iden­ tified by output of high-amylase fluid. Most pancreatic leaks that occur after pancreatic surgery are already controlled by indwelling surgical drains. Many

Chapter 61  Endoscopic Treatment of Pancreatic Disease

A

of these leaks will close over time; endoscopic therapy is reserved for persistent or refractory leaks.85-87 In the absence of a surgical drain, endoscopic therapy is performed to treat symptomatic leaks. Internal leaks that are asso­ ciated with clinical deterioration or symptoms require intervention.88 Endoscopic intervention for pancreatic duct leaks is similar to that described for pancreatic pseudocysts and necrosis. In the setting of a large pancreatic fluid collection, transmural drainage of the collection may be undertaken, with or without concomitant transpapillary therapy. This will control the leak internally. In the absence of a pancreatic fluid collection, the treatment is transpapillary pancreatic duct stent placement to promote internal drainage.20,89 Transpapillary therapy is performed with the intent of crossing the site of the leak,22 though it is not feasible when leaks follow pancreatic tail resection or are located at the pancreatic tail, because the stent would be located outside the pancreatic duct. In this situation, the tip of the stent is positioned downstream to the site of leakage. In patients with external drains in whom endoscopic therapy is being performed to allow drainage removal, success depends on the size of the external drain as compared with the size of the internal stent. Downsizing, clamping, or removing the external drain after successful endoscopic stent placement can promote internal drainage and fistula closure. A variety of glues have been used to seal pancreatic fistula and leaks,90,91 though these glues are not U.S. Food and Drug Administration (FDA) approved and they have the potential for occlusion of the main pancreatic duct.

PANCREATIC CANCER B

C Figure 61-8.  Endoscopic treatment of a pancreatic duct leak due to chronic pancreatitis. A, Stricture in the head of pancreas (arrow). B, Side-branch leak upstream from the stricture (arrow). C, A pancreatic stent is placed across the leak site.

Endoscopic treatment of pancreatic cancer primarily involves palliation of malignant biliary obstruction via ERCP placement of transpapillary biliary stents (see Chap­ ters 60 and 70). In some patients placement of a stent into the pancreatic duct can relieve abdominal pain from pan­ creatic ductal obstruction.92 In addition, some pancreatic cancer patients with pancreatic ductal obstruction will develop pseudocysts or leaks; endoscopic pancreatic duct stent placement across the pancreatic duct stricture can be useful. Pancreatic head cancer can produce gastric outlet obstruction due to duodenal invasion by tumor in 15% to 20% of patients.93,94 Endoscopic placement of selfexpandable metal duodenal stents is an effective palliative method with results comparable to surgical bypass.95 Com­ bined palliation of malignant biliary and duodenal obstruc­ tion is also feasible.96 Pancreatic cancer pain can be treated by EUS-guided celiac plexus block using ethanol to allow permanent ablation of the ganglion. This treatment appears to be effective and sustainable.79

PANCREATIC CYSTS There are a wide variety of pancreatic cysts and cystic neo­ plasms, as discussed in Chapter 60. CT, MRI, and EUS allow differentiation and direction of management toward obser­ vation or surgery.97 Preliminary studies suggest that EUS alcohol injection into specific types of cysts may be a useful nonsurgical alternative.98,99

1041

1042

Section VII  Pancreas KEY REFERENCES

Baron TH. Treatment of pancreatic pseudocysts, pancreatic necrosis, and pancreatic duct leaks. Gastrointest Endosc Clin North Am 2007; 17:559-79. (Ref 20.) Barthet M, Lamblin G, Gasmi M, et al. Clinical usefulness of a treatment algorithm for pancreatic pseudocysts. Gastrointest Endosc 2008; 67:245-52. (Ref 21.) Brugge WR, Lewandrowski K, Lee-Lewandrowski E, et al. Diagnosis of pancreatic cystic neoplasms: A report of the cooperative pancreatic cyst study. Gastroenterology 2004; 126:1330-6. (Ref 97.) Cahen DL, Gouma DJ, Nio Y, et al. Endoscopic versus surgical drainage of the pancreatic duct in chronic pancreatitis. N Engl J Med 2007; 356:676-84. (Ref 72.) Cahen D, Rauws E, Fockens P, et al. Endoscopic drainage of pancreatic pseudocysts: Long-term outcome and procedural factors associated with safe and successful treatment. Endoscopy 2005; 37:977-83. (Ref 35.) Costamagna G, Bulajic M, Tringali A, et al. Multiple stenting of refrac­ tory pancreatic duct strictures in severe chronic pancreatitis: Longterm results. Endoscopy 2006; 38:254-9. (Ref 81.) Delhaye M, Arvanitakis M, Verset G, et al. Long-term clinical outcome after endoscopic pancreatic ductal drainage for patients with painful chronic pancreatitis. Clin Gastroenterol Hepatol 2004; 2:1096-106. (Ref 69.)

Dumonceau JM, Costamagna G, Tringali A, et al. Treatment for painful calcified chronic pancreatitis: Extracorporeal shock wave lithotripsy versus endoscopic treatment: A randomised controlled trial. Gut 2007; 56:545-52. (Ref 66.) Hookey LC, Debroux S, Delhaye M, et al. Endoscopic drainage of pancreatic-fluid collections in 116 patients: A comparison of etio­ logies, drainage techniques, and outcomes. Gastrointest Endosc 2006; 63:635-43. (Ref 27.) Kwan V, Loh SM, Walsh PR, et al. Minor papilla sphincterotomy for pancreatitis due to pancreas divisum. ANZ J Surg 2008; 78:257-61. (Ref 62.) Petrov MS, van Santvoort HC, Besselink MG, et al. Early endoscopic retrograde cholangiopancreatography versus conservative manage­ ment in acute biliary pancreatitis without cholangitis: A metaanalysis of randomized trials. Ann Surg 2008; 247:250-7. (Ref 4.) Seewald S, Groth S, Omar S, et al. Aggressive endoscopic therapy for pancreatic necrosis and pancreatic abscess: A new safe and effective treatment algorithm (videos). Gastrointest Endosc 2005; 62:92-100. (Ref 44.) Varadarajulu S, Noone TC, Tutuian R, et al. Predictors of outcome in pancreatic duct disruption managed by endoscopic transpapillary stent placement. Gastrointest Endosc 2005; 61:568-75. (Ref 23.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

62 Anatomy, Histology, Embryology, Developmental Anomalies, and Pediatric Disorders of the Biliary Tract Frederick J. Suchy

CHAPTER OUTLINE Embryology of the Liver and Biliary Tract  1045 Anatomy  1047 Bile Ducts  1047 Gallbladder  1049 Congenital Anomalies of the Extrahepatic Ducts  1050 Congenital Anomalies of the Gallbladder  1050 An Overview of Disorders of the Biliary Tract in Infants and Children  1050 Diagnosis  1051 Pediatric Disorders of the Bile Ducts  1052 Biliary Atresia  1052 Spontaneous Perforation of the Bile Duct  1056 Bile Plug Syndrome  1056

In this chapter, the embryologic and anatomic characteristics of the bile ducts and gallbladder are reviewed with a focus on information that is useful in diagnosing and treating biliary tract disease and in understanding the anomalies and congenital malformations of these structures. Biliary tract disease in infants and children is considered because many of the disorders that occur early in life are caused by abnormal morphogenesis or adversely affect the process of development.

EMBRYOLOGY OF THE LIVER AND BILIARY TRACT The human liver is formed from two primordia (Fig. 62-1): the liver diverticulum and the septum transversum.1 Proximity of cardiac mesoderm, which expresses fibroblast growth factors (FGFs) 1, 2, and 8, and bone morphogenetic proteins cause the foregut endoderm to develop into the liver.2 Surrounding mesoderm and ectoderm participate in the hepatic specification of the endoderm, and many tran-

Primary Sclerosing Cholangitis  1056 Choledochal Cysts  1058 Congenital Dilatation of the Intrahepatic Bile Ducts  1059 Nonsyndromic Paucity of the Interlobular Bile Ducts  1060 Syndromic Paucity of the Interlobular Bile Ducts (Alagille Syndrome, or Arteriohepatic Dysplasia)  1060 Medical Management of the Chronic Cholestasis  1062 Pediatric Disorders of the Gallbladder  1063 Cholelithiasis  1063 Calculous Cholecystitis  1064 Acalculous Cholecystitis  1065 Acute Hydrops of the Gallbladder  1065 Gallbladder Dyskinesia  1066

scription factors, such as cJun, retinoblastoma gene, and nuclear factor kB, play important roles as regulators of liver embryogenesis.3 The liver diverticulum forms through proliferation of endodermal cells at the cranioventral junction of the yolk sac with the foregut and grows into the septum transversum in a cranioventral direction.4 The earliest marker of mammalian hepatic differentiation is the endodermal expression of albumin, transthyretin, and alpha fetoprotein. Cells that express these markers are called hepatoblasts, and they differentiate into hepatocytes and epithelial cells of the bile ducts. Signaling mediated by the stress-activated protein kinase (SAPK)/Jun N-terminal kinase (JNK) pathway promotes hepatoblast proliferation as well as survival.5 This early change occurs on the eighteenth day of gestation and corresponds to the 2.5-mm stage of the embryo. The signaling molecules that elicit embryonic induction of the liver from the mammalian gut endoderm or induction of other gut-derived organs are being defined. The homeobox gene Hhex is essential for proper hepatoblast differentiation and bile duct morphogenesis. Members of the GATA, FOXA, ONECUT1, and hepatocyte nuclear factor (HNF)3/forkhead transcription factor families are also

1045

Section VIII  Biliary Tract Peritoneal cavity Hepatic diverticulum

Ventral mesentery Stomach region

Ventral mesentery (part of septum transversum)

Foregut

Hepatic cords Gallbladder

Midgut

Vitelline duct

Foregut Midgut

B

A

Bile duct

Liver

Stomach

Dorsal pancreas

Gallbladder Cystic duct

Duodenal loop Ventral pancreas

Dorsal pancreas

Stomach

Liver

C

Dorsal mesentery

Cystic duct Gallbladder Bile duct Foregut

Midgut

Dorsal and ventral pancreas

Fused

1046

D

Figure 62-1.  Stages in the embryologic development of the liver, gallbladder, extrahepatic ducts, pancreas, and duodenum. A, Four weeks. B and C, Five weeks. D, Six weeks. (From Moore KL. The developing human. Philadelphia, Pa.: WB Saunders; 1973.)

required for the formation and differentiation of gut endoderm tissues.3,4 The septum transversum consists of mesenchymal cells and a capillary plexus formed by the branches of the two vitelline veins. At the 3- to 4-mm stage, between the third and fourth weeks of gestation, the growing diverticulum projects as an epithelial plug into the septum transversum.5 The homeodomain transcription factors Hex and Prox1, expressed in the anterior endoderm and hepatic diverticulum, are required for migration of hepatoblasts into the septum transversum that precedes liver growth and morphogenesis.6,7 Another homeodomain protein, Hlx, is necessary for hepatoblast proliferation. At the 5-mm stage, a solid cranial portion (hepatic) and a hollow caudal portion of the diverticulum can be clearly distinguished. The large hepatic portion differentiates into proliferating cords of hepatocytes and the intrahepatic bile ducts. HNF4a expression drives further hepatocyte differentiation and epithelial transformation into the characteristic sinusoidal architecture.8 The smaller, cystic portion, which initially is a cord of epithelial cells, forms the gallbladder, bile duct, and cystic duct through a process of elongation and recanalization. The intrahepatic bile ducts develop from primitive hepatocytes around branches of the portal vein. Cholangiocytes are associated with the basement membrane throughout bile duct development, suggesting that cholangiocytes receive morphogenic signals from components of the extracellular matrix including laminin and type IV collagen.9,10 A ring of hepatocytes in proximity to the portal vein branches first transforms into bile duct–type cells. A second layer of primitive hepatocytes is similarly transformed and produces a

circular cleft around the portal vein that is lined on both sides by bile duct epithelial cells.11 This double-walled cylinder with a slit-like lumen, the ductal plate, can be detected at 9 weeks of gestation. Thus, the entire network of interlobular and intralobular bile ductules develops from the limiting plate. The transcription factors Hes1, HNF6, and HNF1β are required for gallbladder and bile duct development.6 The Notch and transforming growth factor-β (TGF-β) signaling pathways are activated in hepatoblasts surrounding the portal veins, allowing hepatoblasts to become cholangiocytes.5 In sections of the 10-mm embryo, many of the liver cords are traversed by double-walled canals that branch and morphologically are indistinguishable from bile capillaries of the adult. These structures differ from those of the adult in that they are bounded by six or more liver cells instead of two. The process of differentiation of bile ductular epithelial cells (cholangiocytes) from primitive hepatocytes has been documented in humans through the use of immunohistochemical staining with several anticytokeratin antibodies. During the phenotypic shift toward bile duct–type cells, hepatocytes first display increased reactivity for cytokeratins 8 and 18 and express cytokeratin 19 at 20 to 25 weeks of gestation.12 Cholangiocytemesenchymal cell interaction is important for the formation of bile ducts. During the transition from ductal plates to bile ducts, portal myofibroblasts significantly expand and surround newly formed bile ducts. Periportal connective tissue, corticosteroid hormones, and basal laminar components may play important roles in the differentiation of bile ducts. The ductal plate structure requires extensive remodeling through a process of reabsorption, possibly through apop-

Chapter 62  Pediatric Disorders of the Biliary Tract tosis, to yield the characteristic anastomosing system of biliary channels that surround the portal vein. Proteins that appear to have a role in the promotion of apoptosis, specifically Fas antigen and c-myc, are consistently detected in primitive intrahepatic ductal cells.5 Lewis antigen, which is expressed in damaged and apoptotic cells, is also present. Bcl-2 protein, an inhibitor of apoptosis, is not found in early stages of intrahepatic bile duct cell development but becomes detectable later. Computed three-dimensional reconstruction of the developing ductal plate has shown that the ductal plate remodeling process starts at the porta hepatis at approximately 11 weeks of gestation and progresses toward the periphery of the liver.12 The process is in large part completed at term, but even at 40 weeks of gestation, some of the smallest portal vein branches may not be accompanied by an individual bile duct and may still be surrounded by a (discontinuous) ductal plate. In ductal plate malformation, which occurs in biliary disorders such as congenital hepatic fibrosis and Caroli’s disease (see later), insufficient reabsorption of ductal plates can result in the formation of large dilated segments of a primitive bile duct that surrounds the central portal vein.12 The gallbladder and extrahepatic bile ducts start to develop from hepatic endodermal cells and hepatoblasts immediately after formation of the liver primordium. Foxf1 is critical for mesenchymal epithelial cell induction of gallbladder morphogenesis.6 In embryos 5 to 6 mm in length, the original hepatic diverticulum differentiates cranially into proliferating hepatic cords and bile ducts and caudally into the gallbladder. The cystic portion of the liver diverticulum is hollow initially, but the lumen is filled as cells continually migrate into it. A study in 1994 showed that the primitive extrahepatic bile duct maintains continuity with the ductal plate, from which intrahepatic bile ducts are eventually formed.9,10 Contrary to long-held concepts of biliary development, no “solid stage” of endodermal occlusion of the bile duct lumen is found at any stage of gestation. At 16 mm, the cystic duct and proximal gallbladder are hollow, but the fundus of the gallbladder is still partially obstructed by remnants of the epithelial plug. The gallbladder is patent by the third month of gestation. Further development, until birth, consists primarily of continued growth. The characteristic folds of the gallbladder are formed toward the end of gestation and are moderately developed in the neonate. Bile secretion starts at the beginning of the fourth month of gestation; thereafter, the biliary system continuously contains bile, which is secreted into the gut and imparts a dark green color to the intestinal contents (meconium).

allows selective exchange of materials between compartments. No major ultrastructural differences exist between cholangiocytes lining small and large bile ducts, but the functional properties of cholangiocytes are heterogeneous.15 For example, large, but not small, intrahepatic bile ducts are involved in secretin-regulated bile ductal secretion.16 Correspondingly, the secretin receptor and chloridebicarbonate exchanger messenger ribonucleic acids (mRNAs) have been detected in large, but not small, intrahepatic bile duct units.15 Bile secretion begins at the level of the bile canaliculus, the smallest branch of the biliary tree.17 Its boundaries are formed by a specialized membrane of adjacent apical poles of liver cells. The canaliculi form a meshwork of polygonal channels between hepatocytes with many anastomotic interconnections.17 Bile then enters the small terminal channels (the canals of Hering), which have a basement membrane and are lined partly by hepatocytes and partly by cholangiocytes.13 The canals of Hering provide a conduit through which bile may traverse the limiting plate of hepatocytes to enter the larger perilobular or intralobular ducts.18,19 These smallest of biliary radicles are less than 15 to 20 µm in diameter with lumens surrounded by cuboidal epithelial cells. At the most proximal level, one or more fusiform-shaped ductular cells may share a canalicular lumen with a hepatocyte; gradually, the ductules become lined by two to four cuboidal epithelial cells as they approach the portal canal.17 Bile flows from the central lobular cells toward portal triads (from zone 3 to zone 1 of the liver acinus) (see Chapter 71). The terminal bile ductules are thought to proliferate as a result of chronic extrahepatic bile duct obstruction.19 The interlobular bile ducts form a richly anastomosing network that closely surrounds the branches of the portal vein.20-22 These bile ducts (Fig. 62-2) are initially 30 to 40 µm in diameter and are lined by a layer of cuboidal or columnar epithelium that displays a microvillar architecture on its

ANATOMY BILE DUCTS

The adult human liver has more than 2 km of bile ductules and ducts. Quantitative computer-aided three-dimensional imaging has estimated the volume of the entire macroscopic duct system of human liver to be a mean of 20.4 cm.3,13 In these studies the mean internal surface of 398 cm2 is magnified approximately 5.5-fold by the presence of microvilli and cilia at the apical surface of cholangiocytes that play an important role in the regulation of cholangiocyte functions. These structures are far from being inert channels; they are capable of modifying biliary flow and composition significantly in response to hormones such as secretin.14,15 A general feature of bile ductules is their anatomic intimacy with portal blood and lymph vessels, which potentially

Figure 62-2.  Ultrastructure of an interlobular bile duct. The duct is lined by a layer of cuboidal epithelial cells, which are joined by tight junctions (long arrow) and demonstrate a microvillar architecture on their luminal surface (short arrow). (From Jones AL, Springer-Mills E. The liver and gallbladder. In: Weiss L, editor. Modern concepts of gastrointestinal histology. New York, NY: Elsevier; 1984. p 740.)

1047

1048

Section VIII  Biliary Tract Bile duct Spiral valves of cystic duct

Pancreatic duct

Neck Hartmann’s pouch

Bile duct sphincter

Common hepatic duct Bile duct

B

Body Fundus

Bile duct

Pancreatic duct

Pancreas Duodenal muscles

Ampulla of Vater

A

Pancreatic duct sphincter Sphincter of ampulla Junction of bile duct and pancreatic duct

Bile duct sphincter

Extraduodenal junction of bile duct and pancreatic duct

Pancreatic duct sphincter

C

Sphincter of ampulla Sphincter of Oddi

Figure 62-3.  Schematic representation of the gallbladder, extrahepatic biliary tract, and choledochoduodenal junction (A), with enlarged views of the junction of the bile duct and pancreatic duct (B) and the sphincter of Oddi (C). (From Lindner HH. Clinical Anatomy. East Norwalk, Conn.: Appleton & Lange; 1989, copyright McGraw-Hill.)

luminal surface.17 The cells have a prominent Golgi apparatus and numerous vesicles that likely participate in the exchange of substances among cytoplasm, bile, and plasma through the processes of exocytosis and endocytosis.17 These ducts increase in caliber and possess smooth muscle fibers within their walls as they approach the hilum of the liver. The muscular component may provide the morphologic basis for the narrowing of the ducts at this level, as observed on cholangiography.22 Furthermore, as the ducts become progressively larger, the epithelium becomes thicker, and the surrounding layer of connective tissue grows thicker and contains many elastic fibers. These ducts anastomose further to form the large hilar, intrahepatic ducts, which are 1 to 1.5 mm in diameter and give rise to the main hepatic ducts. The common hepatic duct emerges from the porta hepatis after the union of the right and left hepatic ducts, each of which is 0.5 to 2.5 cm long (Fig. 62-3).23,24 The confluence of the right and left hepatic ducts is outside the liver in approximately 95% of cases; uncommonly, the ducts merge inside the liver, or the right and left hepatic ducts do not join until the cystic duct joins the right hepatic duct.24 As the hepatic ducts leave the porta hepatis, they lie within the two serous layers of the hepatoduodenal ligament. This sheath of fibrous tissue binds the hepatic ducts to the adjacent blood vessels. In the adult, the common hepatic duct is approximately 3 cm long and is joined by the cystic duct, usually at its right side, to form the bile duct (or common bile duct).24 However, the length and angle of junction of the cystic duct with the common hepatic duct are variable. The cystic duct enters the common hepatic duct directly in 70% of patients; alternatively, the cystic duct may run ante-

rior or posterior to the bile duct and spiral around it before joining the bile duct on its medial side.23 The cystic duct may also course parallel to the common hepatic duct for 5 to 6 cm and enter it after running posterior to the first portion of the duodenum. In humans, the large intrahepatic bile ducts at the hilum (1- to 1.5-mm diameter) have many irregular side branches and pouches (150- to 270-µm diameter) that are oriented in one plane, corresponding anatomically to the transverse fissure.17 Smaller pouches of the side branches are also found. Many side branches end as blind pouches, but others, particularly at the hilum, communicate with each other. At the bifurcation, side branches from several main bile ducts connect to form a plexus. The functional significance of these structures is not known. The blind pouches may serve to store or modify bile, whereas the biliary plexus provides anastomoses, which may allow exchange of material between the large bile ducts. The anatomy of the hepatic hilum is particularly important to the surgeon. A plate of fibrous connective tissue in the hepatic hilum includes the umbilical plate that envelops the umbilical portion of the portal vein, the cystic plate in the gallbladder bed, and the Arantian plate that covers the ligamentum venosum.24 Histologic examination of the sagittal section of the hilar plate reveals abundant connective tissue, including neural fibers, lymphatic vessels, small capillaries, and small bile ducts. The bile ducts in the plate system correspond to the extrahepatic bile ducts, and their lengths are variable for every segment.24 Like the intestine, the cystic, common hepatic, and bile ducts possess mucosa, submucosa, and muscularis.22 The ducts are lined by a single layer of columnar epithelium.

Chapter 62  Pediatric Disorders of the Biliary Tract Mucus secreting tubular glands can be found at regular intervals in the submucosa, with openings to the surface of the mucosa. The bile duct is approximately 7 cm long, runs between layers of the lesser omentum, and lies anterior to the portal vein and to the right of the hepatic artery.24 The bile duct normally is 0.5 to 1.5 cm in diameter.19 The wall of the extrahepatic bile duct is supported by a layer of connective tissue with an admixture of occasional smooth muscle fibers. The smooth muscle component is conspicuous only at the neck of the gallbladder and at the lower end of the bile duct. The bile duct passes retroperitoneally behind the first portion of the duodenum in a notch on the back of the head of the pancreas and enters the second part of the duodenum. The duct then passes obliquely through the posterior medial aspect of the duodenal wall and joins the main pancreatic duct to form the ampulla of Vater (see Fig. 62-3).23 The mucous membrane bulge produced by the ampulla forms an eminence, the duodenal papilla. In approximately 10% to 15% of patients, the bile and pancreatic ducts open separately into the duodenum. The bile duct tapers to a diameter of 0.6 cm or less before its union with the pancreatic duct.24 As they course through the duodenal wall, the bile and pancreatic ducts are invested by a thickening of both the longitudinal and circular layers of smooth muscle (see Fig. 62-3) of the sphincter of Oddi.25 There is considerable variation in this structure, but it is usually composed of several parts: (1) the sphincter choledochus—circular muscle fibers that surround the intramural portion of the bile duct immediately before its junction with the pancreatic duct; (2) the sphincter pancreaticus, which is present in approximately one third of individuals and surrounds the intraduodenal portion of the pancreatic duct before its juncture with the ampulla; (3) the fasciculi longitudinales—longitudinal muscle bundles that span intervals between the bile and pancreatic ducts; and (4) the sphincter ampullae—longitudinal muscle fibers that surround a sparse layer of circular fibers around the ampulla of Vater.22 The sphincter choledochus constricts the lumen of the bile duct and thus prevents the flow of bile. Contraction of the fasciculi longitudinales shortens the length of the bile duct and thus promotes the flow of bile into the duodenum. The contraction of the sphincter ampullae shortens the ampulla and approximates the ampullary folds to prevent reflux of intestinal contents into the bile and pancreatic ducts. When both ducts end in the ampulla, however, contraction of the sphincter may cause reflux of bile into the pancreatic duct.25 The arterial supply of the bile ducts arises mainly from the right hepatic artery.20 An extraordinarily rich plexus of capillaries surrounds bile ducts as they pass through the portal tracts.20,26 Blood flowing through this peribiliary plexus empties into the hepatic sinusoids via the interlobular branches of the portal vein. The peribiliary plexus may modify biliary secretions through the bidirectional exchange of proteins, inorganic ions, and bile acids between blood and bile. Because blood flows in the direction (from the large toward the small ducts) opposite to that of bile flow, the peribiliary plexus presents a countercurrent stream of biliary-reabsorbed substances to hepatocytes. The intrahepatic arteries, veins, bile ducts, and hepatocytes are innervated by adrenergic and cholinergic nerves. In the autonomic nervous system, there are a number of regulatory peptides such as neuropeptide tyrosine (NPY), calcitonin gene-related peptide, somatostatin, vasoactive intestinal polypeptide (VIP), enkephalin, and bombesin. NPY-positive nerves present in extrahepatic bile ducts may serve to regulate bile flow by autocrine or paracrine mechanisms.

An abundant anastomotic network of blood vessels from branches of the hepatic and gastroduodenal arteries supplies the bile duct.22,26 The supraduodenal portion of the duct is supplied by vessels running along its wall inferiorly from the retroduodenal artery and superiorly from the right hepatic artery. Injury to these blood vessels can result in bile duct stricturing.23 The lymphatic vessels of the hepatic, cystic, and proximal portions of the bile duct empty into glands at the hilum of the liver.22 Lymphatics draining from the lower portion of the bile duct drain into glands near the head of the pancreas.

GALLBLADDER

The gallbladder (see Fig. 62-3) is a storage reservoir that allows bile acids to be delivered in a high concentration and in a controlled manner to the duodenum for the solubilization of dietary lipid (see Chapter 64).22,27 It lies in a fossa on the undersurface of the right lobe of the liver.27 This distensible pear-shaped structure is 3 cm wide and 7 cm long in the adult and has a capacity of 30 to 50 mL.27 The gallbladder has a thin muscular layer with the smooth muscle cells largely oriented around the circumference of the gallbladder. The absorptive surface of the gallbladder is enhanced by numerous prominent folds. The gallbladder is covered anteriorly by an adventitia that is fused with the capsule of the liver. On its posterior aspect and at the apex, it is covered by the visceral peritoneum. The portions of the gallbladder are the fundus, body, infundibulum, and neck.22 The anterior portion of the fundus is located at the level of the right lateral border of the musculus rectus abdominis and the ninth costal cartilage. The posterior aspects of the fundus and body lie close to the transverse colon and duodenum, respectively. Thus, with perforation of the gallbladder, gallstones can readily penetrate these structures.27,28 The infundibulum is an area of tapering between the gallbladder body and neck. Hartmann’s pouch is a bulging of the inferior surface of the infundibulum that lies close to the neck of the gallbladder. Gallstones can become impacted in Hartmann’s pouch, thereby obstructing the cystic duct and producing cholecystitis.27 Extensive inflammation in Hartmann’s pouch can lead to obstruction of the adjacent common hepatic duct (Mirizzi’s syndrome). The gallbladder is connected at its neck to the cystic duct, which empties into the bile duct (see Fig. 62-3).27 The cystic duct is approximately 4 cm long and maintains continuity with the surface columnar epithelium, lamina propria, muscularis, and serosa of the gallbladder. The mucous membrane of the gallbladder neck forms the spiral valve of Heister, which is involved in regulating flow into and out of the gallbladder. The gallbladder is supplied by the cystic artery, which usually arises from the right hepatic artery.27,29 The artery divides into two branches near the neck of the gallbladder: a superficial branch that supplies the serosal surface and a deep branch that supplies the interior layers of the gallbladder wall. Variations in the origin and course of the cystic artery are common.27 Because the cystic artery is an end artery, the gallbladder is particularly susceptible to ische­ mic injury and necrosis that result from inflammation or interruption of hepatic arterial flow. The cystic vein provides venous drainage from the gallbladder and cystic ducts and commonly empties into the portal vein and occasionally directly into the hepatic sinusoids.22,27 The lymph vessels of the gallbladder are connected with the lymph vessels of Glisson’s capsule. Subserosal and submucosal lymphatics empty into a lymph

1049

1050

Section VIII  Biliary Tract gland near the neck of the gallbladder.22 The sympathetic innervation of the gallbladder originates from the celiac axis and travels with branches of the hepatic artery and portal vein. Visceral pain is conducted through sympathetic fibers and is frequently referred to the right subcostal, epigastric, and right scapular regions. Branches of both vagus nerves provide parasympathetic innervation that likely contributes to the regulation of gallbladder motility.22 The gallbladder is lined by a mucosa that manifests multiple ridges and folds and is composed of a layer of columnar epithelial cells. The gallbladder wall consists of a mucosa, lamina propria, tunica muscularis, and serosa.30 The tunica muscularis is thick and invested with an interlocking array of longitudinal and spiral smooth muscle fibers. Tubuloalveolar glands are found in the region of the neck of the gallbladder and are involved in the production of mucus.27,30 The Rokitansky-Aschoff sinuses are invaginations of the surface epithelium that may extend through the muscularis.22 These structures can be a source of inflammation, most likely as a result of bacterial stasis and proliferation within the invaginations. The ducts of Luschka may be observed along the hepatic surface of the gallbladder and open directly into the intrahepatic bile ducts rather than into the gallbladder cavity. These structures are thought to represent a developmental anomaly, and when they are present in the gallbladder bed may be a source of a bile leak after cholecystectomy.27

CONGENITAL ANOMALIES OF THE EXTRAHEPATIC DUCTS Accessory bile ducts are aberrant ducts that drain individual segments of the liver; they may drain directly into the gallbladder, cystic duct, right and left hepatic ducts, or bile duct.23,31 In rare cases, the right hepatic duct may connect to the gallbladder or cystic duct. These anomalies must be recognized on cholangiography in order to prevent inadvertent transection or ligation of bile ducts during surgery. Complete duplication of the bile duct occurs rarely. In most cases, separate ducts drain the right and left hepatic lobes and open into the duodenum.23 Variation in the drainage and course of the cystic duct is common.23 Duplication of the cystic duct may also be encountered. The cystic duct is absent in most cases of agenesis of the gallbladder; rarely the duct alone may be absent, and the gallbladder empties directly into the common hepatic duct.

CONGENITAL ANOMALIES OF THE GALLBLADDER A number of structural anomalies of the gallbladder have been described.23,31 Most of these defects are of no clinical importance, but occasionally the abnormal gallbladder may be a predisposing factor for bile stasis, inflammation, and formation of gallstones. Gallbladder disease in an anomalous or a malpositioned gallbladder may cause diagnostic confusion. Agenesis of the gallbladder may be an isolated anomaly or occur in association with other congenital malformations.31 The abnormality has a frequency at autopsy of 0.04% to 0.13% and likely reflects a lack of development of the gallbladder bud or failure of the normal process of vacu-

olization. Incomplete vacuolization of the solid endodermal cord during development can result in congenital strictures of the gallbladder or cystic duct. Biliary atresia is commonly associated with an absent or atretic gallbladder. Hypoplasia of the gallbladder has been described, particularly in patients with cystic fibrosis. Ectopic tissues of foregut endodermal origin, including gastric, hepatic, adrenal, pancreatic, and thyroid tissues, may be found within the gallbladder wall. A double gallbladder is another rare malformation, which occurs in approximately 1 to 5 per 10,000 persons in the general population.31,32 The two gallbladders may share a single cystic duct, forming a Y-shaped channel, or each may have a distinct cystic duct that enters the bile duct separately.23 Vesica fellae triplex, or triplication of the gall­ bladder, is another rare congenital anomaly.33 Multiple gallbladders are usually discovered because of cholelithiasis, sludge, cholecystitis, or neoplasia. Bilobed gallbladders and gallbladder diverticula are other rare anomalies. A single gallbladder may be divided by longitudinal septa into multiple chambers, probably secondary to incomplete vacuolization of the solid gallbladder bud during morphogenesis.32 Diverticula and septations of the gallbladder may promote bile stasis and gallstone formation. Various malpositions of the gallbladder have been described.32 Rarely, the gallbladder lies under the left lobe of the liver, to the left of the falciform ligament. This defect likely results from migration of the embryonic bud from the hepatic diverticula to the left rather than to the right.23 Some researchers have proposed that the second gallbladder may develop independently from the left hepatic duct, with regression of the normal structure on the right. In other cases, a caudal bud that advances farther than the cranial bud may become buried within the cranial structure, creating an intrahepatic gallbladder. It is thought that if the caudal bud lags behind the movement of the cranial bud, a floating gallbladder results. In this setting, the gallbladder is covered completely with peritoneum and suspended from the undersurface of the liver by mesentery to the gallbladder or cystic duct; the gallbladder is abnormally mobile and prone to torsion. Rarely, gallbladders have been found in the abdominal wall, falciform ligament, and retroperitoneum.32 Several forms of “folded” gallbladders have been described. In one variant, the fundus appears to be bent, giving the appearance of a “Phrygian cap.”32 The gallbladder is usually located in a retroserosal position, and the anomaly is thought to result from aberrant folding of the gallbladder within the embryonic fossa. Aberrant folding of the fossa during the early stages of development can result in kinking between the body and the infundibulum of the gallbladder. Kinked gallbladders probably do not lead to clinical symptoms but may be a source of confusion in the interpretation of imaging studies.32

AN OVERVIEW OF DISORDERS OF THE BILIARY TRACT IN INFANTS AND CHILDREN Cholestatic liver disease results from processes that interfere with either bile formation by hepatocytes or bile flow through the intrahepatic and extrahepatic biliary tree. A number of these disorders result from defective ontogenesis as well as from a failure of postnatal adaptation to the extrauterine environment. Table 62-1 provides a list of disorders that affect the biliary tract and occur in both infants and older children and that are discussed later in the chapter.

Chapter 62  Pediatric Disorders of the Biliary Tract Table 62-1  Disorders of the Biliary Tract in Infants and Children Cholangiopathies Allograft rejection Bile duct obstruction resulting from pancreatic disease (inflammatory or neoplastic) Bile plug syndrome Biliary helminthiasis Caroli’s disease Choledochal cysts Cystic fibrosis Extrahepatic biliary atresia Graft-versus-host disease Idiopathic bile duct stricture (possibly congenital) Post-traumatic bile duct stricture Paucity of intrahepatic bile ducts (syndromic and nonsyndromic) Sclerosing cholangitis (neonatal, inflammatory bowel disease– associated, immunodeficiency-related) Spontaneous perforation of the bile duct Tumors intrinsic and extrinsic to the bile duct Disorders of the Gallbladder Acalculous cholecystitis Acute cholecystitis Acute hydrops of the gallbladder Anomalies Cholelithiasis Chronic cholecystitis Tumors Modified from Balistreri WF. Neonatal cholestasis: Lessons from the past, issues for the future. Semin Liver Dis 1987; 7:61-6.

DIAGNOSIS

Table 62-2  Relative Frequencies of Various Forms of Neonatal Cholestasis DISORDER Idiopathic neonatal hepatitis Extrahepatic biliary atresia α1-Antitrypsin deficiency Intrahepatic cholestatic syndromes (Alagille syndrome, Byler’s disease, others) Hepatitis (cytomegalovirus, rubella, herpes simplex virus, others) Choledochal cyst Bacterial sepsis Endocrinopathy (hypothyroidism, panhypopituitarism) Galactosemia Inborn errors of bile acid metabolism Other metabolic disorders

dice may be greater during the neonatal period than at any other time of life (see Table 62-1).35,36 Liver dysfunction in the infant, regardless of the cause, is commonly associated with bile secretory failure and cholestatic jaundice. Although cholestasis may be traced to the level of the hepatocyte or the biliary apparatus, in practice there may be considerable overlap among disorders with regard to the initial and subsequent sites of injury. For example, damage to the biliary epithelium often is a prominent feature of neonatal hepatitis that results from cytomegalovirus infection. Mechanical obstruction of the biliary tract invariably produces liver dysfunction and in the neonate may be associated with abnormalities of the liver parenchyma, such as giant cell transformation of hepatocytes. Whether giant cells, a frequent, nonspecific manifestation of neonatal liver injury, reflect the noxious effects of biliary obstruction or whether the hepatocytes and the biliary epithelium are damaged by a common agent during ontogenesis, such as a virus with tropism for both types of cells, is unknown. Furthermore, another common histologic variable that often accompanies neonatal cholestasis is bile ductular paucity or a diminution in the number of interlobular bile ducts.37 This finding may be of primary importance in patients with syndromic paucity of intrahepatic bile ducts but may also occur as an occasional feature of many other disorders, including idiopathic neonatal hepatitis, congenital cytomegalovirus infection, and α1-antitrypsin deficiency.38 Serial liver biopsies usually show a progressive decrease in the number of bile ductules per portal tract, with a variable amount of associated inflammation.

FREQUENCY 30-35 30 7-10 5-6 3-5 2-4 2 ≈1 ≈1 ≈1 ≈1

Modified from Balistreri WF. Neonatal cholestasis: Lessons from the past, issues for the future. Semin Liver Dis 1987; 7:61.

There is a particular emphasis on neonatal cholangiopathies and the unique aspects of biliary disease in the older child. The general features of the many cholestatic liver diseases of the neonate are similar, and a central problem of pediatric hepatology is differentiating intrahepatic from extrahepatic cholestasis (Table 62-2).34 The treatment of metabolic or infective liver diseases and the surgical management of biliary anomalies require early diagnosis. Even when effective treatment is not possible, infants and children with progressive liver disease benefit from optimal nutritional support and medical management of chronic liver disease before they are referred for liver transplantation. Because of the immaturity of hepatobiliary function, the number of distinct disorders that exhibit cholestatic jaun-

In most infants with cholestatic liver disease the condition appears during the first few weeks of life. Differentiating conjugated hyperbilirubinemia from the common unconjugated, physiologic hyperbilirubinemia of the neonate or the prolonged jaundice occasionally associated with breastfeeding is essential.39 The possibility of liver or biliary tract disease must be considered in any neonate older than 14 days with jaundice. The stools of a patient with wellestablished biliary atresia are acholic; however, early in the course of incomplete or evolving biliary obstruction, the stools may appear normal or only intermittently pigmented. Life-threatening but treatable disorders such as bacterial infection and a number of inborn errors of metabolism must be excluded. Furthermore, the success of surgical procedures in relieving the biliary obstruction of biliary atresia or a choledochal cyst depends on early diagnosis and surgery. The approach to the evaluation of an infant with cholestatic liver disease is outlined in Table 62-3. The initial assessment should establish promptly whether cholestatic jaundice is present and assess the severity of liver dysfunction. A more detailed investigation may be required and should be guided by the clinical features of the case. All relevant diagnostic tests need not be performed in every patient. For example, ultrasonography may promptly establish a diagnosis of a choledochal cyst in a neonate with jaundice and thus obviate the need to exclude infectious and metabolic causes of liver disease. Numerous routine and specialized biochemical tests and imaging procedures have been proposed to distinguish intrahepatic from extrahepatic cholestasis in infants and thereby avoid unnecessary surgical exploration.39,40 Standard liver biochemical tests usually show variable elevations in serum direct bilirubin, aminotransferase, alkaline phosphatase, and lipid levels. Unfortunately, no single test has proved to have satisfactory discriminatory value, because at least 10% of

1051

1052

Section VIII  Biliary Tract Table 62-3  Evaluation of the Infant with Cholestasis History and Physical Examination Details of family history, pregnancy, presence of extrahepatic anomalies, and stool color Tests to Establish the Presence and Severity of Liver Disease Fractionated serum bilirubin analysis Liver biochemical tests (AST, ALT, alkaline phosphatase, 5′ nucleotidase, gamma glutamyl transpeptidase) Tests of liver function (prothrombin time, partial thromboplastin time, coagulation factors, serum albumin level, serum ammonia level, serum cholesterol level, blood glucose) Tests for Infection Complete blood count Bacterial cultures of blood, urine, and other sites if indicated Viral cultures Paracentesis if ascites Metabolic Studies α1-Antitrypsin level and phenotype if level is reduced Metabolic screen (urine and serum amino acids, urine organic acids) Red blood cell galactose-1-phosphate uridyl transferase activity Serologic tests (HBsAg, TORCH, STS, EBV, others) Serum iron and ferritin levels Sweat chloride analysis Thyroid hormone, thyroid-stimulating hormone (evaluation of hypopituitarism as indicated) Urine and serum analysis of bile acids and bile acid precursors Urine for reducing substances Imaging Studies Ultrasonography of liver and biliary tract (first) Hepatobiliary scintigraphy MRCP Radiography of long bones and skull for congenital infection and of chest for lung and cardiac disease Percutaneous or endoscopic cholangiography (rarely indicated) Procedures Bone marrow examination and skin fibroblast culture for suspected storage disease Duodenal intubation to assess fluid for bile pigment Percutaneous liver biopsy (for light and electron microscopic examination, enzymologic evaluation) Exploratory laparotomy and intraoperative cholangiography ALT, alanine aminotransferase; AST, aspartate aminotransferase; EBV, Epstein-Barr virus; HBsAg, hepatitis B surface antigen; MRCP, magnetic resonance cholangiopancreatography; STS, serologic test for syphilis; TORCH, toxoplasmosis, rubella, cytomegalovirus, herpesvirus.

infants with intrahepatic cholestasis have bile secretory failure sufficient to lead to an overlap in diagnostic test results with those suggestive of biliary atresia.41 The presence of bile pigment in stools is sometimes cited as evidence against biliary atresia, but coloration of feces with secretions and epithelial cells that have been shed by the cholestatic patient may be misleading. Ultrasonography can be used to assess the size and echogenicity of the liver. Even in neonates, high-frequency, realtime ultrasonography usually can define the presence and size of the gallbladder, detect stones and sludge in the bile ducts and gallbladder, and demonstrate cystic or obstructive dilatation of the biliary system.42,43 Extrahepatic anomalies also may be identified. A triangular cord or bandlike periportal echogenicity (3 mm or greater in thickness), which represents a cone-shaped fibrotic mass cranial to the portal vein, appears to be a specific ultrasonographic finding in the early diagnosis of biliary atresia.42,43 The gallbladder “ghost” triad, defined as gallbladder length less than 1.9 cm, lack of smooth or complete echogenic mucosal lining with an indistinct wall, and irregular or lobular contour, has been proposed as additional criteria for biliary atresia.

Computed tomography provides information similar to that obtained by ultrasonography but is less suitable in patients younger than 2 years because of exposure to radiation, the paucity of intra-abdominal fat for contrast, and the need for heavy sedation or general anesthesia.44 Magnetic resonance cholangiopancreatography (MRCP), performed with T2-weighted turbo-spin echo sequences, is widely used to assess the biliary tract in all age groups. In a 1999 study, MRCP reliably demonstrated the bile duct and gallbladder in normal neonates. In some patients with biliary atresia, nonvisualization of the bile duct and demonstration of a small gallbladder have been characteristic MRCP findings.45 A more recent study found that MRCP is 82% accurate, 90% sensitive, and 77% specific for depicting extrahepatic biliary atresia. Contrary to previous reports, false-positive and false-negative findings occur with MRCP. Differentiation of severe intrahepatic cholestasis from biliary atresia may be difficult because the ability of MRCP to delineate the extrahepatic biliary tree depends on bile flow.46 The use of hepatobiliary scintigraphic imaging agents such as 99mTc iminodiacetic acid derivatives may be helpful in differentiating extrahepatic biliary atresia from other causes of neonatal jaundice.44 Unfortunately, a 1997 study showed that in 50% of patients who had a paucity of interlobular bile ducts but no extrahepatic obstruction, biliary excretion of radionuclide was absent.47 Twenty-five percent of patients who had idiopathic neonatal hepatitis also demonstrated no biliary excretion. Nevertheless, the modality remains useful for assessing cystic duct patency in patients with a hydropic gallbladder or cholelithiasis. Percutaneous transhepatic cholangiopancreatography may be of value in visualizing the biliary tract in selected patients,48 but the technique is more difficult to perform in infants than in adults because the intrahepatic bile ducts are small and because most disorders that occur in infants do not result in dilatation of the biliary tree. Endoscopic retrograde cholangiopancreatography (ERCP) may be useful in evaluating children with extrahepatic biliary obstruction and has been performed successfully in a small number of cholestatic neonates.49 Considerable technical expertise is required of the operator to complete this procedure in infants. Most neonates require general anesthesia for a satisfactory examination. Percutaneous liver biopsy is particularly valuable in evaluating cholestatic patients and can be undertaken in even the smallest infants with only sedation and local anesthesia.50 For example, a diagnosis of extrahepatic biliary atresia can be made on the basis of clinical and histologic criteria in 90% to 95% of patients. When doubt about the diagnosis persists, the patency of the biliary tree can be examined directly by a minilaparotomy and operative cholangiogram.

PEDIATRIC DISORDERS OF THE BILE DUCTS BILIARY ATRESIA

Biliary atresia is characterized by the complete obstruction of bile flow as a result of the destruction or absence of all or a portion of the extrahepatic bile ducts.51 As part of the underlying disease process or as a result of biliary obstruction, concomitant injury and fibrosis of the intrahepatic bile ducts also occurs to a variable extent. The disorder occurs in 1 in 10,000 to 15,000 live births and accounts for approximately one third of cases of neonatal cholestatic jaundice (see Table 62-2). It is the most frequent cause of death from

Chapter 62  Pediatric Disorders of the Biliary Tract liver disease and reason for referral for liver transplantation in children (approximately 50% of all cases).52 The cause of biliary atresia is unknown. The disease is not inherited, and there have been several reports of dizygotic and monozygotic twins discordant for biliary atresia.53 In a study of 461 patients in France, seasonality, time clustering, and timespace clustering could not be demonstrated.54 Reports of familial cases have been rare; in most, a detailed histologic description of the extrahepatic biliary tree was not provided to exclude narrowing or hypoplasia of the bile duct associated with severe intrahepatic cholestasis. In the multistate case-controlled National Birth Defects Prevention Study conducted between 1997 and 2002, babies born to nonHispanic black mothers were at greater risk than nonHispanic white mothers. Conception during the spring and low dietary intakes of vitamin E, copper, phosphorus, and beta tocopherol were additional risk factors.55 Several mechanisms have been proposed to account for the progressive obliteration of the extrahepatic biliary tree.56 There is no evidence that biliary atresia results from a failure in morphogenesis or recanalization of the bile duct during embryonic development. Clinical features support the concept that in most cases, injury to the biliary tract occurs after birth. There is little support for an ischemic or toxic origin of extrahepatic bile duct injury. Congenital infections with cytomegalovirus, rubella virus, human herpesvirus 6, and papillomavirus occasionally have been implicated.56 Reovirus type 3 has been implicated on the basis of the serologic evaluation of patients and immunolocalization of reovirus 3 antigens in a bile duct remnant of a patient with biliary atresia.57,58 The results of studies on the role of reovirus in biliary atresia have been contradictory. In a 1998 report, reovirus RNA was detected by reverse-transcriptase polymerase chain reaction methodology in hepatic or biliary tissues, or both, in 55% of patients who had biliary atresia and 78% of patients who had a choledochal cyst,59 compared with 21% of patients who had other hepatobiliary diseases and 12% of autopsy controls. Initial reports of the involvement of group C rotavirus in biliary atresia have not been confirmed.60 A significant increase in human leukocyte antigen (HLA) B12 has been found among patients with biliary atresia who had no associated anomalies.61 The HLA haplotypes A9, B5, A28, and B35 have been found more frequently. Oligonucleotide-based gene chip analysis of cRNA from livers of infants with biliary atresia has demonstrated a coordinated activation of genes involved in lymphocyte differentiation and inflammation.62 The finding of overexpression of osteopontin and γ-interferon indicates a potential role of type 1 T helper (Th1)–like cytokines in the pathogenesis. Biliary atresia is associated with oligoclonal expansions of CD4+ and CD8+ T cells within liver and extrahepatic bile duct remnant tissues, indicating the presence of activated T cells that react to specific antigenic stimulation.63 In a Rhesus rotavirus (RRV) murine model of biliary atresia, γ-interferon was particularly important in mediating bile duct injury.64 In other studies adoptive transfer of T cells from RRV-diseased mice into naïve syngeneic severe combined immunodeficient (SCID) recipients, at a time when viral infection could no longer be demonstrated, caused bile duct specific inflammation, possibly in response to bile duct autoantigens.65 Circulating markers of inflammation persist in biliary atresia and are largely unaffected by portoenterostomy (see later), with clear progressive elevation in both Th1 effectors interleukin (IL)-2 and interferon, some Th2 effectors (IL-4), as well as the macrophage marker (tumor necrosis factor-α [TNF-α]). Increased expression of soluble cell adhesion molecules, sICAM-1 and sVCAM-1, are also

found and likely reflect ongoing recruitment of circulating inflammatory or immunocompetent cells into target tissues.66 Whether this immune response is induced by a viral infection or reflects a genetically programmed response to an infectious or environmental exposure remains unknown. Extrahepatic anomalies occur in 10% to 25% of patients and include cardiovascular defects, polysplenia, malrotation, situs inversus, and bowel atresias.67,68 Some patients who have heterotaxia, including an infant with biliary atresia and polysplenia, have been found to have loss-offunction mutations in the CFC1 gene.69,70 This gene encodes a protein called CRYPTIC, which is involved in establishing the left-right axis during morphogenesis. In contrast, limited studies of infants with biliary atresia and heterotaxia have not found mutations in the INV gene, which is also involved in determining laterality.71 In a microarray analysis of liver tissue from infants with a so-called embryonic form of biliary atresia in which extrahepatic malformations and early onset of cholestatic jaundice occur, a unique pattern of expression of genes involved in chromatin integrity and function (Smarca-1, Rybp, and Hdac3) and overexpression of five imprinted genes (Igf2, Peg3, Peg10, Meg3, and IPW) was found, implying a failure to down-regulate embryonic gene programs that influence the development of the liver and other organs.72 Jagged1 (the gene defective in Alagille syndrome [see later]) missense mutations were identified in 9 of 102 patients with biliary atresia and were associated with a poor prognosis.73

Pathology

Histopathologic findings on initial liver biopsy specimens are of great importance in the management of patients with biliary atresia.51,52 Early in the course, hepatic architecture is generally preserved, with a variable degree of bile ductular proliferation, canalicular and cellular bile stasis, and portal tract edema and fibrosis (Fig. 62-4).52 The presence of bile plugs in portal triads is highly suggestive of large duct obstruction. Furthermore, bile ductules show varying injury to the biliary epithelium, including swelling, vacuolization, and even sloughing of cells into the lumen. Portal tracts may be infiltrated with inflammatory cells, and in approximately 25% of patients there may be giant cell transformation of hepatocytes to a degree observed more commonly in neonatal hepatitis. Bile ductules occasionally may assume a ductal plate configuration suggesting that the disease has interfered with the process of ductular remodeling that occurs during prenatal development.74 Biliary cirrhosis may be present initially or may evolve rapidly over the first months of life, with or without the successful restoration of bile flow.75 The morbid anatomic characteristics of the extrahepatic bile ducts in biliary atresia are highly variable. Kasai proposed a useful classification of the anatomic variants.76 Three main types have been defined on the basis of the site of the atresia. Type I is atresia of the bile duct with patent proximal ducts. Type II atresia involves the hepatic duct, with cystically dilated bile ducts at the porta hepatis. In type IIa atresia, the cystic and bile ducts are patent, whereas in type IIb atresia, these structures also are obliterated. These forms of biliary atresia have been referred to as “surgically correctable” but unfortunately account for less than 10% of all cases. Ninety percent or more of patients have type III atresia, involving obstruction of the common, hepatic, and cystic ducts, without cystically dilated hilar ducts. The entire perihilar area is in a cone of dense fibrous tissue. The gallbladder is involved to some extent in approximately 80% of patients. The type III variant has been characterized as noncorrectable, in that there are no patent

1053

1054

Section VIII  Biliary Tract

A

C

B

D

Figure 62-4.  Histology of the liver and extrahepatic bile duct in biliary atresia. A, Hepatocellular and canalicular cholestasis, multinucleated giant cells (arrow), and portal tract inflammation. (Hematoxylin and eosin, ×400.) B, Expanded potal tract with portal fibrosis, bile ductular proliferation (thin arrows), and bile plug in the bile duct (block arrow) (Masson trichrome, ×250). C, Proximal common hepatic duct in the porta hepatis with sloughing of biliary epithelium, concentric fibrosis of the bile duct wall, lymphocytic infiltration around the duct, and a narrowed but patent lumen. (Hematoxylin and eosin, ×150.) D, Remnant of a bile duct with complete obliteration of the lumen (arrow) and concentric fibrosis of the duct wall. (Hematoxylin and eosin, ×40.) (From Sokol RJ, Mack C, Narkewicz MR, Karrer FM. Pathogenesis and outcome of biliary atresia: Current concepts. J Pediatr Gastroenterol Nutr 2003; 37:4-21. Used with permission.)

hepatic or dilated hilar ducts that can be used for a biliaryenteric anastomosis. Complete fibrous obliteration of at least a portion of the extrahepatic bile ducts is a consistent feature found on microscopic examination of the fibrous remnant.76 Other segments of the biliary tree may demonstrate lumens with varying degeneration of bile duct epithelial cells, inflammation, and fibrosis in the periductular tissues (see Fig. 62-4). In most patients, bile ducts within the liver that extend to the porta hepatis are patent during the first weeks of life but are destroyed progressively, presumably by the same process that damaged the extrahepatic ducts and by the effects of biliary obstruction. In more than 20% of patients, concentric tubular ductal structures similar to those observed in ductal plate malformations are found, indicating that the disease process interfered with the normal remodeling of the biliary tract.

Clinical Features

Most infants with biliary atresia are born at term after a normal pregnancy and have a normal birth weight.56 Female infants are affected more commonly than male infants. The perinatal course is typically unremarkable. Postnatal weight gain and development usually proceed normally. Jaundice is observed by the parents or the physician after the period of physiologic hyperbilirubinemia. Prolonged jaun-

dice may be erroneously attributed to breastfeeding.77 The possibility of liver or biliary tract disease must be con­ sidered in any neonate older than 14 days with jaundice.78 The stools of a patient with well-established biliary atresia are acholic; however, early in the course the stools may appear normally pigmented or only intermittently pigmented. The liver is typically enlarged with a firm edge palpable 2 to 6 cm below the right costal margin.52 The spleen is usually not enlarged early in the course but becomes enlarged as portal hypertension develops. Ascites and edema are not present initially, but coagulopathy may result from vitamin K deficiency. Laboratory studies initially reveal evidence of cholestasis, with a serum bilirubin level of 6 to 12 mg/dL, at least 50% of which is conjugated.52 Serum aminotransferase and alkaline phosphatase levels are moderately elevated. Serum gamma glutamyl transpeptidase and 5′ nucleotidase levels are also elevated.

Treatment

When the possibility of biliary atresia has been raised by clinical, pathologic, and imaging findings, exploratory laparotomy and operative cholangiography are necessary to document the site of obstruction and to direct attempts at surgical treatment.79-81 Sometimes, frozen sections of the

Chapter 62  Pediatric Disorders of the Biliary Tract

Figure 62-5.  The Kasai operation for biliary atresia. A 35- to 40-cm Roux-en-Y anastomosis is made to the porta hepatis after surgical excision of the atretic extrahepatic biliary tree and a cone of fibrous tissue from the porta hepatis. Multiple small but patent bile ducts may be uncovered by this dissection and drained into the Roux loop. An enlarged depiction of the anastomosis of the jejunal loop to the porta hepatis is shown on the left. (Figure drawn and kindly provided by Dr. Frederick Ryckman, Cincinnati, Ohio.)

35–40 cm

Prognosis

The prognosis of untreated biliary atresia is extremely poor; death from liver failure usually occurs within 2 years.89 Of 88 patients in the Biliary Atresia Registry (Surgical Section,

100

80

% Survival

transected porta hepatis are obtained to evaluate the presence and size of ductal remnants; however, the surgeon should avoid transection of the biliary tree, which may be patent but small as a result of biliary hypoplasia or markedly diminished bile flow associated with intrahepatic cholestasis. Patent proximal portions of the bile ducts or cystic structures in the porta hepatis allow conventional anastomosis with a segment of bowel in approximately 10% of patients. In most patients who have obliteration of the proximal extrahepatic biliary tree, the preferred surgical approach is the hepatoportoenterostomy procedure developed by Kasai (Fig. 62-5).82,83 The distal bile duct is transected, and the fibrous bile duct remnant is dissected to an area above the bifurcation of the portal vein.84 The dissection then progresses backward and laterally at this level, and the fibrous cone of tissue is transected flush with the liver surface, thereby exposing an area that may contain residual, microscopic bile ducts. The operation is completed by the anastomosis of a Roux-en-Y loop of jejunum around the bare edge of the transected tissue to provide a conduit for biliary drainage. A number of modifications of the enteric anastomosis, most involving exteriorization of the Roux-en-Y loop with diversion of the bile to the skin, have been used in an effort to decrease the high frequency of postoperative ascending cholangitis84; however, there may be severe fluid and electrolyte losses from the stoma and eventually massive bleeding from peristomal varices. There also is little evidence that the frequency of postoperative bacterial cholangitis is reduced through the use of these procedures.85 Many surgeons perform the original Kasai operation to prevent these complications and to facilitate liver transplantation, if required later. Multiple attempts at reexploration and revision of nonfunctional conduits should be avoided.85 Adjuvant therapy with glucocorticoids and ursodeoxycholic acid as a choleretic agent is widely prescribed postoperatively,86,87 but in a prospective, doubleblind, randomized placebo-controlled trial, glucocorticoids did not reduce the need for liver transplantation after a Kasai portoenterostomy.88

60 Kasai 40

No Kasai

20

0 0

1

2

3

4 5

6

7

8

9 10 11 12 13 14 15 16

Years Figure 62-6.  Actuarial survival of 670 infants with extrahepatic biliary atresia who underwent Kasai’s portoenterostomy (blue line) and 88 who underwent no operation or only an exploratory laparotomy for biliary atresia (red line). The average length of follow-up was 5 years. The difference in survival between the groups is statistically different (P = 0.001). (From Karrer FM, Lilly JR, Stewart BA, et al. Biliary atresia registry: 1976 to 1989. J Pediatr Surg 1990; 25:1076-80.)

American Academy of Pediatrics) who had either no surgery or a simple exploratory laparotomy, only 1 patient survived for more than 3 years. In the same series, follow-up data from numerous pediatric surgeons and practice settings in the United States disclosed a 5-year actuarial survival rate of 48% among 670 patients who had a Kasai operation (Fig. 62-6).90 Several large series from Europe and Japan have demonstrated similar or slightly better results.91-94 In a 2003 report from the Japanese Biliary Atresia Registry, 1381 patients had been enrolled since 1989.91 Jaundice resolved in 57% of patients after the Kasai operation, and the overall

1055

1056

Section VIII  Biliary Tract 5- and 10-year survival rates were 75.3% and 66.7%, respectively. At the time of the report, 57 of 108 patients had survived for 10 years without liver transplantation. In a series of all patients with biliary atresia identified in France over a period of 10 years (1986 to 1996), the overall survival rate of those treated with the Kasai operation and, if necessary, liver transplantation was 68%.92 The 10-year actuarial survival rate in patients with their native livers was 29%, a figure similar to the 31% compiled from 750 published cases by the authors. Therefore, children with biliary atresia derive long-term benefit from the hepatic portoenterostomy procedure, although most have some persisting liver dysfunction. Progressive biliary cirrhosis may result in death from hepatic failure or the need for liver transplantation despite an apparently successful restoration of bile flow. Several factors have been found to contribute to the varying outcome after hepatic portoenterostomy. The age of the patient at the time of surgery is most critical.91,92,95 In several series, bile flow was reestablished in 80% to 90% of infants who were referred for surgery within 60 days of birth.91,96 Resolution of jaundice may still occur with diagnosis after 90 days of age, but long-term survival is compromised even in the era of liver transplantation.95 In a U.S. series, predictors of a poor outcome were white race, surgery at more than 60 days of age, cirrhosis on the initial liver biopsy specimen, totally nonpatent extrahepatic ducts, and absent ducts at the level of transection in the liver hilum.96 Independent prognostic factors for overall survival in the large French study were performance of the Kasai operation and age less than 45 days at surgery.92 Complete atresia of extrahepatic bile ducts and polysplenia syndrome were associated with a less favorable outcome. The experience of the surgical center was also important.92 A normal serum bilirubin level three months after surgery is predictive of long-term survival.97-99 Prehilar bile duct structures of at least 150 to 400 µm, particularly if lined with columnar epithelium, have not been consistently associated with a favorable prognosis.97,100 The quantity of the bile flow has been correlated with the total area of the biliary ductules identified in the excised porta hepatis specimen.101,102 The rate of progression of the underlying bile ductular and liver disease also limits survival.51,103 The disorder is not limited to the extrahepatic biliary tree and can be associated with progressive inflammation and destruction of the intrahepatic bile ducts and eventual cirrhosis.51 Recurring episodes of ascending bacterial cholangitis, which are most frequent during the first two years after surgery, can contribute to the ongoing bile duct injury and even lead to reobstruction.104 Cholangitis develops primarily in infants who have some degree of bile drainage, probably because of the access to ascending infection provided by patent bile ducts in the porta hepatis. Prophylactic oral antibiotics are often used to prevent recurrent cholangitis after a Kasai portoenterostomy, but controlled trials of this approach have not been done.105 Substantial hepatocyte injury, as indicated by lobular disarray and giant cell transformation on liver biopsy specimens, also has been associated with a poor outcome. The presence of the ductal plate malformation on liver biopsy specimens also predicts poor bile flow after hepatoportoenterostomy. Growth failure was associated with the need for transplantation or death by 24 months of age. Esophageal variceal hemorrhage alone is not an absolute indication for urgent liver transplantation in patients with good bile drainage and preserved liver synthetic function.106,107 Liver transplantation is essential in the management of children in whom portoenterostomy does not successfully

restore bile flow, referral is late (probably at 120 days of age or later), and end-stage liver disease develops eventually despite bile drainage.92,108,109 Biliary atresia accounts for 40% to 50% of all liver transplants performed in children. The portoenterostomy is thought to make liver transplantation more difficult technically as a result of intra-abdominal adhesions and the various enteric conduits that are encountered110; however, with the use of reduced-size liver allografts and living-related donors, one-year survival rates of more than 90% can be expected.108,111,112

SPONTANEOUS PERFORATION OF THE BILE DUCT

Spontaneous perforation of the bile duct is a rare but distinct cholestatic disorder of infancy.113 The perforation usually occurs at the junction of the cystic and bile ducts. The cause is unknown, but there may be evidence of obstruction at the distal end of the bile duct secondary to stenosis or inspissated bile.114 Congenital weakness at the site of the perforation and injury produced by infection also have been suggested. Clinical signs, including jaundice, acholic stools, dark urine, and ascites, typically occur during the first months of life.114 The infant also may experience vomiting and lack of weight gain. Progressive abdominal distention is a usual feature; bile staining of fluid within umbilical or inguinal hernias may be observed. Mild to moderate conjugated hyperbilirubinemia with minimal elevation of serum aminotransferase levels is typical. Abdominal paracentesis reveals clear bile-stained ascitic fluid, which usually is sterile. Ultrasonography reveals ascites or loculated fluid in the right upper quadrant; the biliary tree is not dilated. Hepatobiliary scintigraphy demonstrates the free accumulation of isotope within the peritoneal cavity.114 Operative cholangiography is required to demonstrate the site of the perforation.115 Surgical treatment may involve simple drainage of the bilious ascites and repair of the site of the perforation.114-116 If the perforation is associated with obstruction of the bile duct, however, drainage via a cholecystojejunostomy may be required.

BILE PLUG SYNDROME

A plug of thick, inspissated bile and mucus also may cause obstruction of the bile duct.117,118 Otherwise healthy infants have been affected, but the condition occurs more commonly in sick, premature infants who cannot be fed and require prolonged parenteral nutrition. The pathogenesis may involve bile stasis, fasting, infection, and an increased bilirubin load. The cholestasis associated with massive hemolysis, or the inspissated bile syndrome, may have been a variant of the bile plug syndrome but is now infrequent with the advent of measures to prevent and treat Rh and ABO blood group incompatibilities. The clinical presentation may resemble that of biliary atresia. Ultrasonography may show dilated intrahepatic bile ducts. Exploratory laparotomy and operative cholangiography usually are required for diagnosis. Simple irrigation of the bile duct is curative.116

PRIMARY SCLEROSING CHOLANGITIS

Primary sclerosing cholangitis (PSC) is an uncommon, chronic, progressive disease of the biliary tract characterized by inflammation and fibrosis of the intrahepatic and extrahepatic biliary ductal systems leading eventually to biliary cirrhosis.119-121 Only aspects of PSC that are of particular importance to infants and children are discussed

Chapter 62  Pediatric Disorders of the Biliary Tract here (see Chapter 68 for a detailed discussion of PSC). PSC is a pathologic process that occurs in the absence of choledocholithiasis or a history of bile duct surgery. Sclerosing cholangitis may uncommonly present in the neonatal period; it may present later with features of autoimmunity (primary sclerosing cholangitis), often in association with inflammatory bowel disease; or it may occur with other disorders, including Langerhans cell histiocytosis, immunodeficiency, psoriasis, and cystic fibrosis. In adults, carcinoma of the bile ducts must also be excluded; however, this complication has not been reported in children. PSC is associated with inflammatory bowel disease (most often, ulcerative colitis) in 70% of adult patients, and in approximately 50% to 80% of children with the disorder.120,122 A male preponderance has been reported in some, but not all, large series of children with PSC. More than 200 cases of PSC have been reported in children, and most of these have occurred since the mid1980s, presumably as a result of improvements in pediatric cholangiography. The onset of PSC has been reported in the neonatal period; neonates accounted for 15 of 56 cases in a 1994 series of children with the disorder.122 Cholestatic jaundice and acholic stools were observed within the first two weeks of life. The presenting features were virtually identical to those of extrahepatic biliary atresia; however, percutaneous cholecystography disclosed a biliary system that was patent but exhibited rarefaction of segmental branches, stenosis, and focal dilatation of the intrahepatic bile ducts. The extrahepatic bile ducts were involved in six of eight patients. Jaundice subsided spontaneously within six months, but later in childhood all patients had clinical and biochemical features consistent with biliary cirrhosis and portal hypertension. In contrast with PSC in adults and older children, PSC in neonates has not been associated with intestinal disease. Inflammatory bowel disease–associated PSC usually occurs in patients with ulcerative colitis, although cases have been reported in patients with Crohn’s disease.123 The bowel symptoms can precede, occur simultaneously with, or appear years after the diagnosis of PSC. As in adults, treatment of the bowel disease in infants, including colectomy, does not influence the progression of PSC. Celiac disease has also been associated with PSC.124 Lesions similar to those of PSC have been defined by cholangiography in Langerhans cell histiocytosis, but the process is caused by histiocytic infiltration and progressive scarring of portal tracts, with resulting distortion of intrahepatic bile ducts. Cholestasis can occur before the diagnosis of Langerhans cell histiocytosis has been established but most often is found later.125 Children with Langerhans cell histiocytosis may have involvement of multiple organs, with diabetes insipidus, bone lesions, skin lesions, lymphadenopathy, and exophthalmos. Chemotherapy does not affect the course of the biliary tract disease. Liver transplantation has been successful in several children who experienced progression to end-stage liver disease.126 In some children with a variety of immunodeficiencies, both cellular and humoral, sclerosing cholangitis appears to develop. Cryptosporidia and cytomegalovirus have been found concurrently in the biliary tract in some of these patients, as well as in adults with the acquired immuno­ deficiency syndrome (AIDS). 127,128 Treatment of the associated infection has no proven effect on the biliary tract disease. There is no definitive diagnostic test for PSC; the diagnosis is based on a combination of biochemical, histologic, and radiologic data. Typically, adult patients exhibit fatigue,

weight loss, pruritus, right upper quadrant pain, and intermittent jaundice. In children, the clinical presentation is more variable; the most common symptoms are abdominal pain, jaundice, and chronic diarrhea.120 Physical examination sometimes reveals hepatomegaly, which may be associated with splenomegaly, conjunctival icterus, and, rarely, ascites. The serum alkaline phosphatase level is often elevated in patients with PSC, and serum aminotransferase levels may be mildly elevated129; however, in a 1995 series, 15 of 32 patients had a normal alkaline phosphatase level on presentation.130 Hyperbilirubinemia is seen in less than half of pediatric patients. Serum autoantibodies, including antinuclear antibodies and smooth muscle antibodies, may be found in some patients.129 Antineutrophil cytoplasmic antibodies (ANCAs) may be detected. On liver biopsy specimens, the histologic findings may be suggestive of PSC but usually are not diagnostic. Characteristic concentric periductal (“onion skin”) fibrosis may be present later in the course of the disease, but more often, only neoductular proliferation and fibrosis are found.131 Differentiating PSC from autoimmune hepatitis, particularly in the presence of circulating non–organ-specific autoantibodies and hepatic features on liver biopsy specimens, may be difficult. In 25% to 30% of cases an overlap syndrome may occur in children with both hepatic and cholestatic serum liver test results and with histologic features of autoimmune hepatitis and PSC.119 Serologic findings include the presence of antinuclear, smooth muscle, and anti–liverkidney microsome type 1 (anti-LKM-1) antibodies and perinuclear ANCAs.132 The diagnosis of PSC is established by cholangiography.133 ERCP has been the method of choice for visualizing the intrahepatic and extrahepatic bile ducts122,134; however, a 2002 study of children demonstrated that MRCP was comparable to ERCP in correctly identifying changes of PSC in 13 cases and excluding abnormalities in 5.133 Irregularities of the intrahepatic and extrahepatic ducts can be found, including alternating strictures and areas of dilatation that produce a beaded appearance. Involvement of the intrahepatic bile ducts predominates in patients whose condition appears after the neonatal period. Occasionally, dominant strictures of the extrahepatic ducts or papillary stenosis is found. Small-duct PSC with a normal cholan­ giogram but histologic features of PSC rarely occurs in children.135 The prognosis of PSC in children is guarded.120 The clinical course of the disorder is variable but usually progressive. In a 1994 series of 56 children, the median survival time from onset of symptoms was approximately 10 years, similar to that reported in adults.122 In another study of 52 children, the median survival free of liver transplantation was 12.7 years.120 Analysis of survival factors at presentation indicates that older age, splenomegaly, and a prolonged prothrombin time predicted a poor outcome.130 The occurrence of jaundice after the neonatal period with a persisting serum bilirubin level of more than five times the upper normal value was also associated with a poor outcome. Hepatocellular carcinoma also may occur, but cholangiocarcinoma, an important complication of adult PSC, has not been reported in children. The treatment of PSC in children is unsatisfactory.120,136 No published reports of controlled trials have demonstrated convincingly that any medical therapy improves histologic characteristics and prolongs survival. Uncontrolled experience has suggested some benefit for immunosuppressive therapy with prednisone and azathioprine in patients with the overlap syndrome.119 Ursodeoxycholic acid therapy

1057

1058

Section VIII  Biliary Tract in adults and in a limited number of children has led to an improvement in clinical symptoms and in liver test abnormalities, but a long-term benefit of treatment on survival has not been demonstrated.121 Liver transplantation is an important option for patients who experience pro­ gression to end-stage liver disease, and long-term results in children appear to be good137; however, recurrence of PSC after transplantation has been reported in children.131

CHOLEDOCHAL CYSTS Incidence and Classification

Choledochal cysts are congenital anomalies of the biliary tract that are manifested by cystic dilatation of the extrahepatic and intrahepatic bile ducts.138,139 The incidence of choledochal cysts is 1 in 13,000 to 15,000 in Western countries and as high as 1 in 1000 in Japan.140 Choledochal cysts are not familial; female children are affected more commonly than male children. Cases have been described in utero and in older adult patients, but approximately two thirds of patients seek medical attention before the age of 10. The classification proposed by Todani and colleagues (Fig. 62-7) are cited frequently.141,142 Several varieties of type I cysts, accounting for 80% to 90% of cases, exhibit segmental or diffuse fusiform dilatation of the bile duct. Type II cysts consist of a true choledochal diverticulum. Type III cysts consist of dilatation of the intraduodenal portion of the bile duct, or choledochocele. Type IV cysts may be subdivided into type IVa, or multiple intrahepatic and extrahepatic cysts, and type IVb, or multiple extrahepatic cysts. The type IVb variant either is uncommon or may overlap with type I. Whether type V, or Caroli’s disease, which consists of single or multiple dilatations of the intra-

hepatic ductal system, should be viewed as a form of choledochal cyst is unsettled.142,143

Etiology

The cause of choledochal cysts has not been established.140 Congenital weakness of the bile duct wall, a primary abnormality of epithelial proliferation during embryologic ductal development, and congenital obstruction of bile ducts have been suggested. A relationship to other obstructive cholangiopathies, such as biliary atresia, has been proposed but not proved.144 Reovirus RNA has been detected by reversetranscriptase polymerase chain reaction methodology in hepatic or biliary tissues of 78% of patients who have choledochal cysts.59 A high frequency (40%) of an anomalous union of the pancreatic and bile ducts, which may allow reflux of pancreatic secretions into the biliary tree, has been described.145 This process may result in progressive injury to the developing ductal system, with subsequent weakness and dilatation. Choledochal cysts have also been found in some patients with autosomal recessive polycystic renal disease.146

Pathology

The cysts are composed of a fibrous wall; there may be no epithelial lining or a low columnar epithelium.140 Mild chronic inflammation may be present. Complete, inflammatory obstruction of the terminal portion of the bile duct is common in infants who have a choledochal cyst. Liver biopsy specimens in the affected neonate show typical features of large duct obstruction.140 Findings may mimic those observed in extrahepatic biliary atresia. Portal tract edema, bile ductular proliferation, and fibrosis may be prominent. A pattern of biliary cirrhosis may be observed

II

Ib

III

Ic

Ia

IV a

IV b

V

Figure 62-7.  Classification of choledochal cysts according to Todani and colleagues.141 Ia, common type; Ib, segmental dilatation; Ic, diffuse dilatation; II, diverticulum; III, choledochocele; IVa, multiple cysts (intra- and extrahepatic); IVb, multiple cysts (extrahepatic); V, single or multiple dilatations of the intrahepatic ducts (Caroli’s disease). (From Savader SJ, Benenati JF, Venbrux AC, et al. Choledochal cysts: Classification and cholangiographic appearance. AJR 1991; 156:327-31.)

Chapter 62  Pediatric Disorders of the Biliary Tract in older patients with long-standing biliary obstruction. Carcinoma of the cyst wall may occur by adolescence.147,148

Clinical Features

The infantile form of choledochal cyst disease must be distinguished from other forms of hepatobiliary disease of the neonate, particularly biliary atresia.140 Disease often appears during the first months of life, and as many as 80% of patients have cholestatic jaundice and acholic stools.149 Vomiting, irritability, and failure to thrive may occur. Examination reveals hepatomegaly and in approximately one half of patients a palpable abdominal mass. In a series of 72 patients diagnosed postnatally, 50 (69%) exhibited jaundice that was associated with abdominal pain in 25 or with a palpable mass in 3; 13 (18%) had abdominal pain alone, and 2 (3%) had a palpable mass alone. In a 2008 series, adults were more likely to exhibit abdominal pain (97% versus 63%, P < 0.001), and children were more likely to experience jaundice (71% versus 25%, P = 0.001).148 In older patients, epigastric pain may result from pancreatitis. Intermittent jaundice and fever may result from recurrent episodes of cholangitis. The classic triad of abdominal pain, jaundice, and a palpable abdominal mass is observed in less than 20% of patients.148 Spontaneous perforation of a choledochal cyst may occur, particularly when bile flow is obstructed. Progressive hepatic injury can occur during the first months of life as a result of biliary obstruction caused by poor bile flow, sludge, protein plugs, and stones composed of fatty acids and calcium.150

Diagnosis

The diagnosis of a choledochal cyst is best established by ultrasonography (Fig. 62-8).44 In fact, several reports have demonstrated that antenatal ultrasonography can be used to detect a choledochal cyst in the fetus. Sequential ultrasonographic examinations have allowed the study of the evolution of choledochal cysts during pregnancy. In the older child, percutaneous transhepatic cholangiography or ERCP may help define the anatomic features of the cyst; its site of biliary origin, including an anomalous arrangement of the pancreaticobiliary junction; and the extent of both extrahepatic and intrahepatic disease, including the presence of intraductal strictures and calculi.151 MRCP is being used increasingly to evaluate the extent of the cyst and defects

within the biliary tree and to detect an anomalous union of the pancreaticobiliary duct.152 MRCP was less effective than ERCP for detecting minor ductal abnormalities and small choledochoceles in adults.151 In practice, most pediatric surgeons rely on an operative cholangiogram to define the extent of intrahepatic and extrahepatic disease.140

Treatment

The preferred treatment is surgical excision of the cyst with reconstruction of the extrahepatic biliary tree.138,140 Biliary drainage is usually accomplished by a choledochojejunostomy with a Roux-en-Y anastomosis. Excision of the cyst reduces bile stasis and the risk of cholangitis and cholangiocarcinoma. Simple decompression and internal drainage should be done only when the complicated anatomic characteristics do not allow complete excision. Long-term follow-up is essential because recurrent cholangitis, lithiasis, anastomotic stricture, and pancreatitis may develop years after the initial surgery.139

CONGENITAL DILATATION OF THE INTRAHEPATIC BILE DUCTS

Nonobstructive saccular or fusiform dilatation of the intrahepatic bile ducts is a rare, congenital disorder.153,154 In the pure form, known as Caroli’s disease, dilatation is classically segmental and saccular and associated with stone formation and recurrent bacterial cholangitis. A more common type, Caroli’s syndrome, is associated with a portal tract lesion typical of congenital hepatic fibrosis (CHF).154 Dilatation of the extrahepatic bile ducts (choledochal cysts) also may be present. Renal disease occurs in both forms, renal tubular ectasia occurs with Caroli’s disease, and both conditions can be associated with autosomal recessive polycystic kidney disease (ARPKD) or, rarely, autosomal dominant polycystic kidney disease.155 Mutations in a polycystic kidney and hepatic disease 1 gene (PKHD1) have been identified in patients with ARPKD.156 The gene encodes a large protein (4074 amino acids) called fibrocystin to reflect the main resulting structural abnormalities in liver and kidney.157 The protein shares structural features with the hepatocyte growth factor receptor and appears to belong to a superfamily of proteins that are involved in the regulation of cell proliferation and of cellular adhesion and repulsion.158 Fibrocystin is localized to the primary cilia of renal epithelial cells and cholangiocytes, suggesting a link between ciliary dysfunction and cyst development.

Pathology

The intrahepatic cysts are in continuity with the biliary tract and lined by epithelium that may be ulcerated and hyperplastic.153 The cysts may contain inspissated bile, calculi, and purulent material. Liver biopsy specimens may reveal normal tissue or features of acute or chronic cholangitis.159 Portal tract edema and fibrosis may be present. In cases associated with CHF, findings associated with the ductal plate malformation can be expected; the lumen of the portal bile duct forms an epithelium-lined circular cleft surrounding a central vascularized connective tissue core, or a series of bile duct lumens are arranged in a circle around a central fibrous tissue core.160 Figure 62-8.  Ultrasonographic demonstration of a type I choledochal cyst in an infant with cholestasis. A large cystic mass in the right upper quadrant is shown on this transverse scan. The point of juncture of the cyst with the bile duct is delineated by an arrow.

Clinical Features

Patients usually seek medical attention during childhood and adolescence because of hepatomegaly and abdominal pain.155,156 The disorder appears in the neonate as renal

1059

1060

Section VIII  Biliary Tract tion. Therapy with ursodeoxycholic acid has been used successfully to dissolve intrahepatic stones.163 Cholangiocarcinoma may develop within the abnormal bile ducts.166 Portal hypertension and variceal bleeding may predominate in patients with CHF and Caroli’s disease.156 End-stage renal disease develops in some patients who have associated polycystic kidney disease. Liver transplantation is an option in patients who have extensive disease and frequent complications, including refractory cholangitis.167

NONSYNDROMIC PAUCITY OF THE INTERLOBULAR BILE DUCTS

Figure 62-9.  Cholangiographic findings in Caroli’s disease. Percutaneous cholangiography reveals multiple cystic lesions throughout a markedly enlarged liver. The cystic lesions are in continuity with the bile ducts. The extrahepatic bile ducts are normal. (From Kocoshis SA, Riely CA, Burrell M, Gryboski JD. Cholangitis in a child due to biliary tract anomalies. Dig Dis Sci 1980; 25:59-65.)

disease or cholestasis.156 The saccular or fusiform dilatation of bile ducts predisposes to stagnation of bile leading to the formation of biliary sludge and intraductal lithiasis. Fever and intermittent jaundice may occur during episodes of bacterial cholangitis. Hepatosplenomegaly is found in cases associated with CHF; affected patients may exhibit bleeding esophageal varices.153 The polycystic kidneys may be palpable. Liver biochemical tests may have normal results or show mild to moderate elevations of serum bilirubin, alkaline phosphatase, and aminotransferase levels.154 Liver synthetic function is well preserved, but repeated episodes of infection and biliary obstruction within the cystic bile ducts eventually may lead to hepatic failure. The maximal concentrating capacity is the most frequently abnormal renal function test finding; variable elevations of blood urea nitrogen and serum creatinine levels reflect the severity of the underlying kidney disease.156

Diagnosis

Ultrasonography, MRCP, and computed tomography are of great value in demonstrating the cystic dilatation of the intrahepatic bile ducts.161,162 Renal cysts or hyperechogenicity of papillae may be detected. Percutaneous or endoscopic cholangiography (Fig. 62-9) usually demonstrates a normal bile duct with segmental, saccular dilatations of the intrahepatic bile ducts.159 Rarely, the process may be limited to one lobe of the liver.

Prognosis and Treatment

The clinical course is often complicated by recurrent episodes of cholangitis156,159; sepsis and liver abscess may occur. The prognosis in the setting of persistent or recurrent infection is poor. Calculi frequently develop within the cystically dilated bile ducts and can complicate the treatment of cholangitis.163 Patients who have extensive hepatolithiasis may experience intractable abdominal pain. Removal of stones by surgery, endoscopy, or lithotripsy usually is not feasible.164 Hepatic resection is indicated for disease limited to a single lobe.165 Surgical drainage procedures generally are not effective and may complicate later liver transplanta-

A paucity of interlobular bile ducts may be an isolated and unexplained finding in infants and children with idiopathic cholestasis or a feature of a heterogeneous group of disorders that include congenital infections with rubella and cytomegalovirus and genetic disorders such as α1-antitrypsin deficiency and inborn errors of bile acid metabolism.168,169 Bile duct paucity has been observed in some cases of Williams and Noonan syndromes.170,171 Paucity of interlobular bile ducts has been defined as a ratio of the number of interlobular bile ducts to the number of portal tracts of less than 0.4.12,172 At least 10 portal tracts should be examined on a liver biopsy specimen to be confident that bile duct paucity is present. The structural abnormality has also been referred to as intrahepatic biliary atresia or intrahepatic biliary hypoplasia; however, these terms imply more insight into the pathogenesis of ductular paucity than currently prevails. Cases may arise from true biliary dysgenesis but more often result from active injury and loss of bile ducts.12,172 Bile duct paucity may occur without associated developmental anomalies and without a documented intrauterine infection or genetic disorder; however, this idiopathic form of nonsyndromic bile duct paucity is likely to be heterogeneous in cause with extremely variable clinical features and prognosis.169,173 Cholestasis typically develops early in infancy and may be associated with progressive liver disease.

SYNDROMIC PAUCITY OF THE INTERLOBULAR BILE DUCTS (ALAGILLE SYNDROME, OR ARTERIOHEPATIC DYSPLASIA)

Syndromic paucity of interlobular bile ducts (Alagille syndrome, or arteriohepatic dysplasia) is the most common form of familial intrahepatic cholestasis. This disorder is characterized by chronic cholestasis, a decreased number of interlobular bile ducts, and a variety of other congenital malformations.174 An autosomal dominant mode of transmission with incomplete penetrance and variable expressivity has been established from family studies.175 A partial deletion of the short arm of chromosome 20 was detected in some patients and led to the identification of the Alagille syndrome gene. Mutations in the jagged1 (JAG1) gene have been identified in approximately 94% of affected patients and include total gene deletions as well as protein truncating, splicing, and missense mutations.176 JAG1 encodes a ligand in the Notch signaling pathway that is involved in cell fate determination during development.177 Mutations in the gene encoding for the NOTCH2 receptor have been found in patients with Alagille syndrome who were negative for JAG1 mutations.178 There appears to be no phenotypic difference between patients with deletion of the entire JAG1 gene and those with intragenic mutations.179 The disorder may affect only one family member; such cases may represent spontaneous mutations of the JAG1 gene. Alternatively, it is possible that

Chapter 62  Pediatric Disorders of the Biliary Tract

A

B

C

Figure 62-10.  Facial appearance in syndromic paucity of the intrahepatic bile ducts. A, Infant. B, Child. C, Young adult. (See text for description.) (From Alagille D, Estrada A, Hadchouel M, et al. Syndromic paucity of interlobular bile ducts [Alagille’s syndrome or arteriohepatic dysplasia]: Review of 80 cases. J Pediatr 1987; 110:195-200.)

the variability in gene expression is so great that minimally affected family members are not diagnosed. A 1994 analysis of 33 families collected through 43 probands corroborated the autosomal dominant inheritance and concluded that the rate of penetrance is 94% and that 15% of cases are sporadic; however, expressivity was variable, and 26 persons (including 11 siblings) exhibited minor forms of the disease.180

Clinical Features

Chronic cholestasis of varying severity affects 95% of patients.181,182 Jaundice and clay-colored stools may be observed during the neonatal period and become apparent in most patients during the first 2 years of life. Intense pruritus may be present by 6 months of age.174 The liver and spleen are often enlarged. During the first years of life, xanthomata appear on the extensor surfaces of the fingers and in the creases of the palms and popliteal areas. Dysmorphic facies (Fig. 62-10) are usually recognized during infancy and become more characteristic with age.183 The forehead is typically broad, the eyes are deeply set and widely spaced, and the mandible is somewhat small and pointed, imparting a triangular appearance to the face. The malar eminence is flattened, and the ears are prominent. Extrahepatic anomalies have been described with this syndrome, but the phenotypic expression varies considerably. In a 1999 series of 92 patients, cholestasis occurred in 96%, cardiac murmur in 97%, butterfly vertebrae in 51%, posterior embryotoxon (mesodermal dysgenesis of the iris and cornea) in 78%, and characteristic facies in 96% of patients.182 Short stature is a regular feature but is only partially attributed to the severity of chronic cholestasis. Growth hormone insensitivity associated with elevated circulating levels of growth hormone– binding protein has been described in these patients.184 Mild to moderate mental retardation affects 15% to 20% of patients. Congenital heart disease occurs in most patients, and peripheral pulmonic stenosis is observed in approximately 90%.182,185 Systemic vascular malformations also may be present. Osseous abnormalities include a decreased bone age, variable shortening of the distal phalanges, and vertebral arch defects (e.g., butterfly vertebrae, hemiver­ tebrae, and a decrease in the interpedicular distance).

Ophthalmologic examination may reveal eye anomalies, including posterior embryotoxon, retinal pigmentation, and iris strands. Renal abnormalities and hypogonadism also have been described.182 Laboratory studies reveal an elevation of total serum bilirubin levels (usually 2 to 8 mg/dL) during infancy and intermittently later in life.182 Approximately 50% of the total serum bilirubin is conjugated. Serum alkaline phosphatase, gamma glutamyl transpeptidase, and 5′ nucleo­ tidase levels may be extremely high and correlate somewhat with the degree of cholestasis. Serum aminotransferase levels are mildly to moderately increased. Serum cholesterol levels may be 200 mg/dL or higher, and serum triglyceride concentrations may range from 500 to 1000 mg/dL. Total serum bile acid concentrations are markedly elevated, but the bile acid profiles in serum, urine, and bile do not differ qualitatively from those seen in other cholestatic disorders.

Pathology

The hallmark of this condition is a paucity of interlobular bile ducts.173 Paucity may be defined as a significantly decreased ratio of the numbers of interlobular portal bile ducts to portal tracts (<0.4).172 The histologic features during the first months of life may overlap with those of neonatal hepatitis, in that there can be ballooning of hepatocytes, variable cholestasis, portal inflammation, and giant cell transformation. Often the number of interlobular bile ducts is not decreased on initial liver biopsy specimens, but bile duct injury consisting of cellular infiltration of portal triads contiguous to interlobular bile ducts, lymphocytic infiltration and pyknosis of biliary epithelium, and periductal fibrosis may be evident.168,186 Serial biopsy specimens from an individual patient may initially show bile duct proliferation, followed later in life by a paucity of bile ducts (Fig. 62-11).187 Paucity of interlobular bile ducts is usually apparent after 3 months. Mild periportal fibrosis also may be present, but progression to cirrhosis is uncommon. The extrahepatic bile ducts are patent but usually narrowed or hypoplastic. Ultrastructural studies have demonstrated the accumulation of bile pigment in the cytoplasm near lysosomes and vesicles of the outer convex

1061

1062

Section VIII  Biliary Tract Prognosis and Treatment

Figure 62-11.  Histologic features of syndromic paucity of the interlobular bile ducts. A portal triad in the liver with a distinct artery and vein but with no bile duct is shown in this low-power photomicrograph. (Masson trichrome stain.) (From Portmann BC, Roberts EA. Developmental abnormalities and liver disease in childhood. In: Burt AD, Portmann BC, Ferrell LD, eds. MacSween’s Pathology of the Liver. 5th ed. London: Churchill Livingstone; 2007. p 167.)

space of the Golgi apparatus. The bile canaliculi most often appear to be structurally normal, but in some cases they may appear to be dilated with blunting and shortening of microvilli.186

Pathogenesis

The mechanisms involved in the pathogenesis of bile duct paucity and cholestasis are unsettled. Also unknown is how the hepatobiliary disease relates to the multiplicity of congenital anomalies found in other organ systems. Mice homozygous for the Jag1 mutation die of hemorrhage early during embryogenesis and exhibit defects in remodeling of the embryonic and yolk sac vasculature.188 The strong JAG1 expression during human embryogenesis, both in the vascular system and in other mesenchymal and epithelial tissues, implicates abnormal angiogenesis in the pathogenesis of Alagille syndrome and particularly the paucity of interlobular bile ducts. In human embryos JAG1 is expressed in the distal cardiac outflow tract and pulmonary artery, major arteries, portal vein, optic vesicle, otocyst, branchial arches, metanephros, pancreas, and mesocardium; around the major bronchial branches; and in the neural tube.189 All these structures are affected in Alagille syndrome. Many of the JAG1 mutations generate premature termination codons, and many of these mutations produce a truncated protein that exerts a dominant-negative effect on Notch signaling.190 Although a vascular basis for the anomalies in Alagille syndrome seems possible, the precise mechanisms leading to bile duct paucity remain unknown. Notch signaling has an important role in the differentiation of biliary epithelial cells and is essential for their tubular formation during intrahepatic bile duct development.191 There is evidence that a lack of branching and elongation of bile ducts during postnatal liver growth contributes to peripheral bile duct paucity and cholestasis.192 It is of great interest to note that profound cholestasis can occur in this disorder during the neonatal period even when the interlobular bile ducts are not decreased in number. By contrast, later in life, when cholestasis may be less severe as judged by clinical and biochemical criteria, interlobular bile ducts may be undetectable on liver biopsy specimens.

The clinical course is marked by varying degrees of cholestasis, sometimes worsened by intercurrent viral infections. Morbidity may result from pruritus, cutaneous xanthomata, and neuromuscular symptoms related to vitamin E deficiency. Treatment involves the provision of an adequate caloric intake, prevention or correction of fatsoluble vitamin deficiencies, and symptomatic measures to relieve pruritus. The long-term prognosis depends on the severity of the liver disease and associated malformations.185 Of 80 patients who had this disorder and who were followed by Alagille and associates, 21 patients died, but only 4 died as a result of liver disease.174 Partial external biliary diversion may be effective for treating severe pruritus and hypercholesterolemia in patients without cirrhosis who do not respond to medical therapy.193,194 In another series of 92 patients, the mortality rate was 17%.182 The factors that contributed significantly to mortality were hepatic disease or transplantation (25%), complex congenital heart disease (15%), and intracranial hemorrhage (25%). Hepatocellular carcinoma may occur.195 In a retrospective review of 268 patients, vascular anomalies such as intracranial aneurysms accounted for 34% of the mortality.196 On the basis of these studies, the 20-year predicted life expectancy is approximately 75% for all patients, 80% for those not requiring liver transplantation, and 60% for those who require liver transplantation. In a study of 168 patients with liver disease, actuarial survival rates with a native liver were 51% and 38% at 10 and 20 years, respectively, and overall survival rates were 68% and 62%, respectively.197 Neonatal chole­ static jaundice was associated with poorer survival with a native liver. Survival and candidacy for liver transplantation may be limited by the severity of associated cardiovascular anomalies.198 In a series of patients who underwent liver transplantation, a higher than expected mortality rate of 43% was attributed to cardiac disease or a previous Kasai procedure.199

MEDICAL MANAGEMENT OF CHRONIC CHOLESTASIS

Cholestatic liver disease in children adversely affects nutritional status, growth, and development, which all contribute to morbidity and mortality.200 In a child with chronic, and sometimes progressive, cholestatic liver disease, efforts should be directed to promoting growth and development and minimizing discomfort.201 Protein-energy malnutrition leading to growth failure is an inevitable consequence of chronic liver disease in 60% of children.201,202 Steatorrhea is common in children with cholestasis, as a result of impaired intraluminal lipolysis, solubilization, and intestinal absorption of long-chain triglycerides.203 Medium-chain triglycerides do not require solubilization by bile salts before intestinal absorption and thus can provide needed calories when administered orally in one of several commercial formulas or as an oil supplement.203 Significant morbid conditions resulting from fat-soluble vitamin deficiencies can be prevented in large part in cholestatic children.204 Because metabolic bone disease, manifesting as rickets and pathologic fractures, can result from vitamin D deficiency, vitamin D should be provided as D2 (5000 IU/day) or as 25-hydroxycholecalciferol (3 to 5 µg/kg/ day). Supplements of elemental calcium (50 to 100 mg/kg/ day) and phosphorus (25 to 50 mg/kg/day) also may be required. Bone mass can be reduced in cholestatic children even with normal serum 25-hydroxyvitamin D levels, possibly related to impaired insulin-like growth factor I production by the liver.205

Chapter 62  Pediatric Disorders of the Biliary Tract Xerophthalmia, night blindness, and thickened skin have been reported in patients who have a vitamin A deficiency. Oral supplements of vitamin A, 5000 to 25,000 IU/day, should be administered.204 Vitamin K deficiency and associated coagulopathy may be treated initially with an oral water-soluble supplement administered in doses of 2.5 to 5 mg twice weekly to as much as 5 mg daily. Children who do not respond or who have significant bleeding require intramuscular injections of vitamin K.206 Chronic deficiency of vitamin E may produce a disabling, degenerative neuromuscular syndrome characterized by areflexia, ophthalmoplegia, cerebellar ataxia, peripheral neuropathy, and posterior column dysfunction.204 The onset can be observed within the first 2 years of life. Because serum vitamin E levels may be elevated spuriously in the presence of hyperlipidemia, the ratio of serum vitamin E to total serum lipids is most useful in monitoring the patient’s vitamin E status; deficiency in a child less than 12 years old, for example, is indicated by a ratio of less than 0.6. The child may not respond to massive doses of standard vitamin E preparations (150 to 200 IU/kg/day). Therapy with intramuscular dl-alpha-tocopherol (50 mg/day) or the water-soluble form of vitamin E, d-alpha-tocopherol polyethylene glycol-1000-succinate (15 to 25 IU/kg/day), is effective.204 Xanthomata and pruritus may cause substantial discomfort. Pruritus may be observed by three months of age.207 The success of most therapies for pruritus depends on the presence of patent bile ducts that allow bile acids and other biliary constituents to reach the gut lumen. Biliary diversion has been used as a successful alternative to relieve intractable pruritus in some patients with intrahepatic cholestasis.207,208 The antibiotic rifampin, through up-regulation of pathways for biotransformation and biliary excretion, and the choleretic bile acid ursodeoxycholic acid are used for the treatment of pruritus with varying degrees of success.209,210 Because of evidence that a component of pruritus may be of central neurogenic origin mediated by the opiate receptor system, opioid receptor antagonists such as naltrexone have been effective in some patients with severe pruritus unresponsive to other agents; however, side effects, withdrawal symptoms, and the lack of experience in children limit the general use of these medications.211 The nonabsorbable anion exchange resin cholestyramine may be used to bind bile acids, cholesterol, and presumably other potentially toxic agents in the intestinal lumen.207 This medication may lower serum lipid levels and bind the substances involved in the pathogenesis of pruritus. A dose of 0.25 to 0.5 g/kg/day is administered before breakfast or in divided doses before meals to relieve severe pruritus and xanthomata.207 Cholestyramine is relatively unpalatable,

however, and carries modest risks for intestinal obstruction, caused by inspissation of the drug, and hyperchloremic acidosis. Colesevelam is a novel bile acid sequestrant that has superior bile acid–binding efficacy compared with cholestyramine and is taken in a more palatable tablet form.212 Its use in cholestatic liver disease has been limited. Pruritus also has been treated with exposure to ultraviolet B light.

PEDIATRIC DISORDERS OF THE GALLBLADDER CHOLELITHIASIS

Cholelithiasis is uncommon in otherwise healthy children and usually occurs in patients who have a predisposing condition.213,214 An ultrasonographic survey of 1570 persons (ages 6 to 19 years) detected gallstones in only two female subjects, ages 13 and 18 years.215,216 None of the persons in the study population had undergone cholecystectomy. The overall prevalence of gallstone disease was 0.13% (0.27% in female subjects). Most cases come to light near the time of puberty, but gallstones have been reported at any age, including during fetal life. Pigmented gallstones predominate in infants and children.213 The conditions associated with an increased risk of cholelithiasis are listed in Table 62-4. An underlying cause of cholelithiasis can be identified in more than one half of children with calculous cholecystitis. An in-depth discussion of the pathogenesis of gallstones can be found in Chapter 65; however, certain factors may assume greater importance during infancy and childhood.214,217 For example, an increased frequency of calculous cholecystitis is reported in sick premature infants, who often undergo a period of prolonged fasting without frequent stimulation of gallbladder contraction and who require periods of prolonged parenteral nutrition.217 Many of these patients have complicated medical courses that include frequent blood transfusions, episodes of sepsis, abdominal surgery, and use of diuretics and narcotic analgesics. Limited analyses of gallstones in such cases generally have shown the presence of mixed cholesterol-calcium bilirubinate stones.218 In the critically ill infant there may be a continuum from the common occurrence of an enlarged, distended gallbladder filled with sludge to the eventual development of cholelithiasis. As in adults, the incidence of gallstones is increased in children with disease or prior resection of the terminal ileum.218 In a 2007 series, 24% of 30 children with gallstones had calcium carbonate stones, previously considered rare.219 Black pigment gallstones occur commonly in patients who have chronic hemolytic disorders.220 These stones are

Table 62-4  Conditions Associated with Cholelithiasis in Children and Their Relative Frequencies According to Age* 0-12 MONTHS

1-5 YEARS

6-21 YEARS

Abdominal surgery Total parenteral nutrition Sepsis Bronchopulmonary dysplasia Hemolytic disease Intestinal malabsorption Necrotizing enterocolitis Hepatobiliary disease

Hepatobiliary disease Abdominal surgery Artificial heart valve Malabsorption

Pregnancy Obesity Hemolytic disease Abdominal surgery Intestinal malabsorption Hepatobiliary disease Total parenteral nutrition

*In approximate order of frequency. Modified from Friesen CA, Roberts CC. Cholelithiasis: Clinical characteristics in children: Case analysis and literature review. Clin Pediatr (Phila) 1989; 28:294-8.

1063

1064

Section VIII  Biliary Tract composed predominantly of calcium bilirubinate, with substantial amounts of crystalline calcium carbonate and phosphate. In sickle cell disease, the risk of gallstones increases with age and occurs in at least 14% of children younger than 10 years and 36% of those between 10 and 20 years.221 The polymorphism in the promoter of the uridine diphosphate (UDP)-glucuronyl transferase 1A1 (UGT1A1) gene that underlies Gilbert’s syndrome, a chronic form of unconjugated hyperbilirubinemia (see Chapter 20), appears to be a major genetic risk factor, increasing the frequency and leading to an earlier age of onset of gallstones in patients with sickle cell disease.222 The genetic factors that lead to cholelithiasis have not yet been defined in children. Polymorphisms in genes encoding the biliary cholesterol transporter ATP-binding cassette (ABC)G5/G8 and the phospholipid transporter ABCB4 have been limited to gallstone disease in adults. The nuclear receptor subfamily 1, group H, member 4 (NR1H4) gene, which encodes the nuclear bile salt receptor farsenoid X receptor (FXR), is another candidate gene for cholestrol gallstone susceptibility. Obstructive jaundice in infants also may be caused by brown pigment cholelithiasis.223,224 Brown pigment stones are composed of varying proportions of calcium bilirubinate, calcium phosphate, calcium palmitate, cholesterol, and organic material. Unconjugated bilirubin accounts for a large percentage of the total bile biliary pigments. In several cases, bile has had high β-glucuronidase activity and on culture grew an abundant population of several bacteria. Pigment gallstones are postulated to have formed spontaneously in these infants, who had bacterial infections of the biliary tract. Patients who have no identifiable cause of cholelithiasis are more likely to be female, older, and obese; have a family history of gallbladder disease; and have a greater likelihood of adult-like symptoms.218 Gallstones were detected in 10 of 493 obese children (2%; 8 girls, 2 boys).225 Cholesterol gallstones predominate in these patients. Insights into the pathogenesis of gallstones have been gained through careful studies of Pima Indians, who have an extraordinarily high prevalence of cholesterol gallstones. Highly saturated bile has not been detected among Pima Indians younger than age 13 years, but bile saturation increases significantly in both sexes during pubertal growth and development.226 In this population the sex-related difference in the size of the bile acid pool begins during puberty; young men show a significant rise in the size of the bile acid pool with age, whereas young women show only a slight rise. Because cholesterol gallstones are associated with smaller bile acid pools, the divergence in bile acid pool size between the two sexes also may account for the sexrelated difference in the frequency of gallstones, which begins during adolescence. Prolonged use of high-dose ceftriaxone, a third-generation cephalosporin, has been associated with the formation of calcium-ceftriaxone salt precipitates in the gallbladder. The process, also called biliary pseudolithiasis, is observed in 30% to 40% of children treated with the drug for severe infections.227 Patients may complain of abdominal pain and exhibit signs of intrahepatic cholestasis. Biliary sludge and gallbladder precipitates are found on ultrasonography.228 The problem generally resolves spontaneously with discontinuation of the drug.

Clinical Features

Most gallstones are found in the gallbladder.218 Children have a lower incidence of bile duct stones than adults. Most patients are asymptomatic; the gallstones are discovered

either incidentally during the investigation of another problem or during screening because the patient has a condition associated with a high risk of cholelithiasis.214 Patients may complain of intermittent abdominal pain of variable severity; the pain may be localized to the right upper quadrant in older children but is generally poorly localized in infants.229 The physical examination findings are usually unremarkable. Tenderness in the right upper quadrant suggests cholecystitis, as occurs when a stone migrates to the neck of the gallbladder and obstructs the cystic duct. Infants may exhibit irritability, cholestatic jaundice, and acholic stools.214 Liver biochemical test results are usually normal.214 Plain films of the abdomen may reveal calculi, depending on the calcium content of the stone. Ultrasonography is considered the most sensitive and specific imaging technique for the demonstration of gallstones. Hepatobiliary scintigraphy is a valuable adjunct; failure to visualize the gallbladder provides evidence of acute cholecystitis (see later).

Treatment

Cholecystectomy remains the treatment of choice in patients who have symptoms or a nonfunctioning gallbladder.214 Laparoscopic cholecystectomy is done frequently in children and infants as young as 10 months.230 Operative cholangiography and exploration of the bile duct may be indicated on the basis of clinical imaging and operative findings. If choledocholithiasis is demonstrated prior to laparoscopic cholecystectomy in the older child and adolescent, then endoscopic sphincterotomy and stone extraction may be done first. In asymptomatic patients without biochemical abnormalities (“silent gallstones”), management poses a more difficult problem. Epidemiologic studies and radiocarbon dating of gallstones in adults indicate a lag time of more than one decade between initial formation of a stone and development of symptoms.231 In patients who have underlying disorders such as hemolysis or ileal disease, cholecystectomy may be carried out at the same time as another surgical procedure. In addition, elective laparoscopic cholecystectomy is becoming the norm in children with chronic hemolytic anemias and asymptomatic cholelithiasis to prevent the potential complications of cholecystitis and choledocholithiasis.232 In cases associated with hepatic disease, severe obesity, or cystic fibrosis, the surgical risk of cholecystectomy may be substantial, and clinical judgment must be applied. In these cases, the patient should be counseled about the nature of the disease and the symptoms that may develop. Spontaneous resolution of cholelithiasis and even bile duct stones has been reported in infants. Because recurrence of lithiasis is rare in infants, cholecystectomy may not be required; however, patients with obstructive cholestasis are at risk for sepsis and cholangitis and should undergo surgery.214 Little experience has been reported in children with alternative therapies for gallstones such as medical dissolution with oral bile acid administration or shock-wave lithotripsy. Ursodeoxycholic acid therapy is of no value in the treatment of the predominantly pigment stones found in this age group. Furthermore, ursodeoxycholic acid failed to dissolve radiolucent gallstones in 10 children with cystic fibrosis.233

CALCULOUS CHOLECYSTITIS

Cholelithiasis may be associated with acute or chronic inflammation of the gallbladder (see also Chapter 65).214 Acute cholecystitis is often precipitated by impaction of a stone in the cystic duct. A progressive increase in pressure in the gallbladder secondary to fluid accumulation, the pres-

Chapter 62  Pediatric Disorders of the Biliary Tract ence of stones, and the chemical irritant effects of bile acids can lead to progressive inflammation, congestion, and vascular compromise. Infarction, gangrene, and perforation can occur. Proliferation of bacteria within the obstructed gallbladder lumen can contribute to the process and lead to biliary sepsis. Chronic calculous cholecystitis is more common than acute cholecystitis. It may develop insidiously or after several attacks of acute cholecystitis. The gallbladder epithelium commonly becomes ulcerated and scarred.

Clinical Features

The acute onset of right upper quadrant pain is a constant feature of acute cholecystitis.218 The pain may be poorly localized in infants. Nausea and vomiting are frequent. Children have a higher frequency of jaundice (50%) than do adults. The patient may appear acutely ill with shallow respirations and may be febrile, particularly if bacterial infection is superimposed. Guarding of the abdomen is common, and palpation usually elicits tenderness in the right upper quadrant. Murphy’s sign may be present. The onset of chronic cholecystitis is usually more indolent. The clinical course may be marked by recurrent episodes of upper abdominal discomfort. Older patients may experience intolerance to fatty foods. In one series, episodes of right upper quadrant pain developed in 64% of children with cholelithiasis and no cystic or bile duct obstruction and was most likely a consequence of chronic cholecystitis.234 Physical examination may yield negative findings or may disclose local tenderness over the gallbladder. In acute cholecystitis, the white blood cell count is often elevated with a predominance of polymorphonuclear leukocytes.214 Serum bilirubin and alkaline phosphatase levels may be increased. Serum aminotransferase levels may be normal, but high elevations, suggestive of hepatocellular disease, can occur early with acute obstruction of the bile duct. In patients with chronic cholecystitis, results of the complete blood count and liver biochemical tests are usually normal. In patients with an acute or chronic presentation, a plain film of the abdomen may demonstrate calcifications in the right upper quadrant.214 Abdominal ultrasonography is extremely useful in documenting the presence of stones in the gallbladder, may show thickening of the gallbladder wall, and may demonstrate dilatation of the biliary tract secondary to obstruction of the bile duct by a stone that has migrated from the gallbladder. MRCP may demonstrate similar findings but usually requires general anesthesia in infants and young children.235 Hepatobiliary scintigraphy rarely is necessary in the acutely ill patient but may be of value in demonstrating a malfunctioning gallbladder in patients with chronic cholecystitis.

Treatment

The acutely ill patient should be treated with intravenous fluids, analgesics, and broad-spectrum antibiotics. Cholecystectomy should be performed as soon as fluid deficits are corrected and infection is controlled.236 High-risk, acutely ill patients may benefit from percutaneous drainage via a transhepatic cholecystostomy. The results of surgery are excellent (see also Chapter 66). Care should be taken to exclude bile duct stones by operative cholangiography and, if necessary, exploration of the duct. Laparoscopic bile duct exploration for choledocholithiasis can be performed safely in children at the time of cholecystectomy and can clear all of the bile duct stones in most patients.237 Cholecystectomy is also the treatment of choice for patients with chronic calculous cholecystitis. Laparoscopic

cholecystectomy is the preferred approach for most patients.238,239

ACALCULOUS CHOLECYSTITIS

Acalculous cholecystitis is an acute inflammation of the gallbladder without gallstones (see also Chapter 67).240 The disorder is uncommon in children but has been associated with infection or systemic illness. Pathogens have included streptococci (groups A and B); Leptospira interrogans; gram-negative organisms such as Salmonella and Shigella species and Escherichia coli; and parasitic infestations with Ascaris species or Giardia lamblia. In immunocompromised patients, pathogens such as Isospora belli and cytomegalovirus, Cryptosporidium, Aspergillus, and Candida species should be considered. Acalculous cholecystitis may follow abdominal trauma and has been observed in patients with systemic vasculitis, including polyarteritis nodosa, systemic lupus erythematosus, and mucocutaneous lymph node (Kawasaki’s) disease; however, in these conditions, gallbladder distention without inflammation also may occur. Congenital narrowing or inflammation of the cystic duct or external compression by enlarged lymph nodes has been associated with the disorder in children. Clinical features of acute acalculous cholecystitis include right upper quadrant or epigastric pain, nausea, vomiting, fever, and occasionally jaundice.241 Right upper quadrant guarding and tenderness are present; a tender gallbladder is sometimes palpable. The findings may be less apparent in infants or critically ill patients because the presentation may be obscured by the underlying illness. Laboratory evaluation may reveal elevated serum levels of alkaline phosphatase and conjugated bilirubin. Leukocytosis may occur. Ultrasonography discloses an enlarged, thick-walled gallbladder that may be distended with sludge but contains no calculi.241 Many patients respond to nonoperative management with nasogastric suction, intravenous fluids, and antibi­ otics, with resolution of clinical and imaging finding. Cho­ lecystectomy will be required in cases associated with increasing gallbladder wall thickening and distension and with persistence of the nonshadowing echogenic materials or sludge in the gallbladder and pericholecystic fluid.240,241 The diagnosis is confirmed at laparotomy. The gallbladder is usually inflamed, and cultures of bile may yield positive results for the offending bacteria or contain parasites. The gallbladder may become gangrenous. Cholecystostomy drainage may be an alternative approach in critically ill patients. Some children may present with chronic symptoms of right upper quadrant pain and nausea or vomiting.240 The white blood cell count and results of liver biochemical tests are usually normal. Most patients demonstrate abnormal gallbladder function on radionuclide hepatobiliary scanning.242 These patients generally have chronic inflammation in the gallbladder and require cholecystectomy.

ACUTE HYDROPS OF THE GALLBLADDER

Acute noncalculous, noninflammatory distention of the gallbladder may be observed in infants and children.243,244 The gallbladder is not acutely inflamed, and cultures of the bile are usually sterile. The absence of gallbladder inflammation and generally benign prognosis distinguish acute hydrops from acute acalculous cholecystitis. There may be a generalized mesenteric adenitis of lymph nodes near the cystic duct without mechanical compression. A temporal relationship to other infections, including scarlet fever and leptospirosis, has been observed in some cases.245 Acute hydrops also has been associated with Kawasaki’s disease

1065

1066

Section VIII  Biliary Tract and Henoch-Schönlein purpura.246 Like acalculous cholecystitis, the disorder can occur in children on prolonged parenteral nutrition. In some cases, a cause is not identified. Acute hydrops is associated with the acute onset of cramping abdominal pain and often nausea and vomiting.246 Fever and jaundice may be present. The right upper quadrant is usually tender, and the distended gallbladder may be palpable. Liver biochemical test levels may be mildly elevated. The white blood cell count may be elevated. Some of these changes can be attributed to the associated disorders such as scarlet fever or Kawasaki’s disease. Ultrasonography reveals an enlarged, distended gallbladder without calculi. The diagnosis of acute hydrops is confirmed in many patients at laparotomy.246 Cholecystectomy obviously is required if the gallbladder appears gangrenous. Pathologic examination of the gallbladder wall usually shows edema and mild inflammation. Cultures of the bile are usually sterile. These benign findings have led some surgeons to treat acute hydrops by a simple cholecystostomy instead of a cholecystectomy246; however, the treatment of gallbladder hydrops frequently is nonsurgical with a focus on supportive care and management of the intercurrent illness. In most patients, particularly in children on total parenteral nutrition in whom enteral feeding has been initiated, the process subsides spontaneously. Ultrasonography has been useful in establishing the diagnosis and following the spontaneous resolution of gallbladder distention. The prognosis is excellent. Gallbladder function can be expected to return to normal in most cases.246

GALLBLADDER DYSKINESIA

Gallbladder dyskinesia is recognized as a cause of chronic abdominal pain in children. The diagnosis is suggested by the presence of postprandial abdominal pain, the absence of cholelithiasis, and an abnormal ejection fraction on cholecystokinin-stimulated hepatobiliary scintigraphy. Pain relief after cholecystectomy has been variable in several reports.247 Gallbladder ejection fractions of less than 35% to 50% have sometimes been considered abnormal and an indication for surgery (see Chapter 63). Gallbladder dyskinesia was the most common indication for surgery in 62 (58%) of the 107 children who underwent cholecystectomy in one series.248 In another published report of 51 children who underwent laparoscopic cholecystectomy for gallbladder dyskinesia after exclusion of more common gastrointestinal disorders, 27 of 38 (71%) patients available for follow-up experienced complete relief of symptoms.249 The presence of nausea, upper abdominal pain, and a gallbladder ejection fraction of less than 15% most reliably predicted benefit from cholecystectomy (posi-

tive predictive value of 93%). Histologic evidence of chronic cholecystitis was found in only 10 of 27 (41%) children with complete relief of symptoms and was not an independent predictor of a successful outcome. The presence of chronic inflammation in these patients suggests that they may have had a chronic acalculous cholecystitis rather than gallbladder dysmotility.

KEY REFERENCES

Adkins RB Jr, Chapman WC, Reddy VS. Embryology, anatomy, and surgical applications of the extrahepatic biliary system. Surg Clin North Am 2000; 80:363-79. (Ref 23.) Avisse C, Flament JB, Delattre JF. Ampulla of Vater. Anatomic, embryologic, and surgical aspects. Surg Clin North Am 2000; 80:201-12. (Ref 25.) Bezerra JA, Balistreri WF. Cholestatic syndromes of infancy and childhood. Semin Gastrointest Dis 2001; 12:54-65. (Ref 39.) Bogert PT, LaRusso NF. Cholangiocyte biology. Curr Opin Gastroenterol 2007; 23:299-305. (Ref 16.) Davenport M. Biliary atresia. Semin Pediatr Surg 2005; 14:42-8. (Ref 85.) Edil BH, Cameron JL, Reddy S, et al. Choledochal cyst disease in children and adults: A 30-year single-institution experience. J Am Coll Surg 2008; 206:1000-5. (Ref 148.) Emerick KM, Rand EB, Goldmuntz E, et al. Features of Alagille syndrome in 92 patients: Frequency and relation to prognosis. Hepatology 1999; 29:822-9. (Ref 182.) Feldstein AE, Perrault J, El-Youssif M, et al. Primary sclerosing cholangitis in children: A long-term follow-up study. Hepatology 2003; 38:210-17. (Ref 120.) Imamoglu M, Sarihan H, Sari A, Ahmetoglu A. Acute acalculous cholecystitis in children: Diagnosis and treatment. J Pediatr Surg 2002; 37:36-9. (Ref 241.) Kaechele V, Wabitsch M, Thiere D, et al. Prevalence of gallbladder stone disease in obese children and adolescents: Influence of the degree of obesity, sex, and pubertal development. J Pediatr Gastroenterol Nutr 2006; 42:66-70. (Ref 225.) Kerkar N, Norton K, Suchy FJ. The hepatic fibrocystic diseases. Clin Liver Dis 2006; 10:55-71. (Ref 154.) Lykavieris P, Hadchouel M, Chardot C, Bernard O. Outcome of liver disease in children with Alagille syndrome: A study of 163 patients. Gut 2001; 49:431-5. (Ref 197.) Ng VL, Balistreri WF. Treatment options for chronic cholestasis in infancy and childhood. Curr Treat Options Gastroenterol 2005;8: 419-30. (Ref 200.) Roskams TA, Theise ND, Balabaud C, et al. Nomenclature of the finer branches of the biliary tree: Canals, ductules, and ductular reactions in human livers. Hepatology 2004; 39:1739-45. (Ref 18.) Shneider BL, Mazariegos GV. Biliary atresia: A transplant perspective. Liver Transpl 2007; 13:1482-95. (Ref 108.) Sokol RJ, Mack C, Narkewicz MR, Karrer FM. Pathogenesis and outcome of biliary atresia: Current concepts. J Pediatr Gastroenterol Nutr 2003; 37:4-21. (Ref 52.) Stringer MD, Taylor DR, Soloway RD. Gallstone composition: Are children different? J Pediatr 2003; 142:435-40. (Ref 216.) Tanimizu N, Miyajima A. Molecular mechanism of liver development and regeneration. Int Rev Cytol 2007; 259:1-48. (Ref 5.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

63 Biliary Tract Motor Function and Dysfunction B. Joseph Elmunzer and Grace H. Elta

CHAPTER OUTLINE Anatomy and Physiology  1067 Gallbladder Dyskinesia  1067 Sphincter of Oddi Dysfunction  1067 Definition  1067 Epidemiology  1068 Clinical Features  1068 Classification  1069

ANATOMY AND PHYSIOLOGY The sphincter of Oddi (SO) is composed of layers of smooth muscle that are embedded in, but functionally separate from, the muscle of the duodenal wall and that serve as a 6 to 10 mm high pressure zone. Three portions of the SO are identified: a small segment (sphincter ampullae) that covers the common channel formed by the union of the bile and pancreatic ducts (when a common channel is present); a second small portion (sphincter pancreaticus) that surrounds the beginning of the main pancreatic duct; and the largest portion (sphincter choledochus) that covers the distal bile duct (Fig. 63-1). In addition, the fasciculi longitudinales are muscle bundles that span intervals between the bile and pancreatic ducts and promote the flow of bile into the duodenum during contraction (see Chapter 62). In humans, the SO functions primarily as a resistor, with tonic contraction that limits bile flow during the interdigestive period. It also serves as a pump, with phasic contractions that facilitate the flow of bile into the duodenum, perhaps serving a housekeeping function for the distal bile duct. The SO participates in the migrating motor complex, with motilin-induced increases in the frequency and amplitude of sphincter contractions shortly before and during bursts of intense duodenal contractions. The complex neurohormonal control of biliary motility involves sympathetic, parasympathetic, and enteric nerves. Almost every neurotransmitter in the enteric nervous system has been identified in the biliary tree.

GALLBLADDER DYSKINESIA Biliary dysmotility, or clinical problems related to abnormal biliary tract motility, can occur in either the gallbladder or the SO. Although gallbladder stasis clearly predisposes to sludge and stone formation, whether gallbladder dysfunction, or delayed emptying of the gallbladder in the absence of stones or sludge, causes biliary symptoms is unclear.

Diagnosis  1069 Treatment  1071 Sphincter of Oddi Dysfunction in Pancreatitis  1072 Idiopathic Acute Recurrent Pancreatitis  1072 Chronic Pancreatitis  1072 Failure of Response to Biliary Sphincterotomy  1072

Patients who experience typical biliary pain but have no evidence of gallstones may be studied with scintigraphic imaging of the gallbladder during intravenous infusion of cholecystokinin (CCK). Delayed gallbladder emptying has been reported to be predictive of pain relief after cholecystectomy,1,2 although this finding remains controversial.3,4 Delayed gallbladder emptying is more common in patients with functional bowel disorders than in control subjects.5 In many of these patients, symptoms are probably caused by the underlying functional bowel disorder, and the gallbladder dysmotility is incidental. The histologic diagnosis of chronic cholecystitis in resected gallbladders has been proposed as confirmation of a gallbladder source of symptoms,1,2 although this suggestion also has been disputed.6 A meta-analysis of 9 studies3 and a systematic review of 23 studies4 concluded that evidence that the gallbladder ejection fraction predicts symptomatic relief after cholecystectomy is lacking (see Chapter 67). A more recent, smaller meta-analysis of mostly retrospective data, however, demonstrated a statistically significant benefit to cholecystectomy in symptomatic patients with gallbladder dyskinesia.7 To complicate matters further, many patients with biliary pain, no stones on ultrasound (US), and normal CCK scintigraphy experience symptomatic improvement after cholecystectomy.8 Although many surgeons offer laparoscopic cholecystectomy to patients with biliary pain and delayed gallbladder emptying, further prospective clinical trials are necessary to determine definitively the role of CCK cholescintigraphy in the management of acalculous biliary pain.

SPHINCTER OF ODDI DYSFUNCTION DEFINITION

Sphincter of Oddi dysfunction (SOD) is a benign, noncalculous obstructive disorder that occurs at the level of the SO. The pathogenesis of SOD can be divided into two subtypes: stenosis, which results from passive obstruction at the SO

1067

1068

Section VIII  Biliary Tract caused by fibrosis, inflammation, or both; and dyskinesia, which results from intermittent obstruction caused by sphincter muscle spasm. These two mechanisms of functional obstruction at the SO are not mutually exclusive. The term ampullary stenosis has generally been used interchangeably with SO stenosis. Less precise terms for SOD, such as postcholecystectomy syndrome, biliary spasm, biliary dyssynergia, papillary stenosis, and sclerosing papillitis, also have been used in the medical literature, leading to confusion in nomenclature.

EPIDEMIOLOGY

The frequency of manometrically detected SOD in patients with an intact gallbladder has not been well studied. Elevated basal SO pressure has been reported in 40% of patients with gallbladder stones, with or without biliary

Longitudinal muscle of duodenum Circular muscle of duodenum Bile duct Pancreatic duct

Sphincter choledochus Sphincter pancreaticus

3 portions of the sphincter of Oddi

Sphincter ampullae Figure 63-1.  Anatomy of the sphincter of Oddi. Note the three portions of the sphincter of Oddi: the sphincter ampullae (surrounding the short common channel), the sphincter pancreaticus, and the sphincter choledochus (the largest portion).

pain (“biliary colic”).9 When elevated levels of liver enzymes were present, 40% of 25 patients with gallbladder stones but without bile duct stones had an elevated basal SO pressure. Ruffolo and coworkers reported that approximately 50% of 81 patients with biliary-type pain, an intact gallbladder, and no evidence of gallstones had delayed gallbladder emptying, SOD, or both.10 By contrast, no basal SO pressure elevation greater than 30 mm Hg was found in 50 asymptomatic volunteers without gallstones.11 The frequency of SOD in postcholecystectomy patients with persistent or recurrent biliary-type pain has been better studied but depends on the criteria for patient selection. Pain resembling preoperative biliary pain occurs in 10% to 20% of postcholecystectomy patients.12 The most common explanation is that the preoperative symptoms were not caused by gallstones. The most likely diagnosis in this group of patients is a functional gastrointestinal disorder such as irritable bowel syndrome or functional (nonulcer) dyspepsia. SOD has been reported in 9% to 14% of patients evaluated for postcholecystectomy pain.13 When other causes of postcholecystectomy pain have been excluded and SO manometry (SOM) (see later) has been performed in a more carefully screened group, the frequency of SOD is 30% to 60%.14 When these patients are classified by the Milwaukee classification for possible SOD (Table 63-1), the frequencies of elevated basal SO pressure are 86%, 55%, and 28% for patients with suspected types I, II, and III SOD, respectively (see later).

CLINICAL FEATURES

SOD is a possible cause of three clinical conditions: (1) persistent or recurrent biliary-type pain following chole­ cystectomy, (2) recurrent idiopathic (unexplained) pancreatitis, and (3) biliary-type pain in patients with an intact gallbladder but without cholelithiasis (the least studied and most controversial clinical association) (Table 63-2). SOD generally occurs spontaneously, but has also been described with increased frequency in patients who have undergone liver transplantation,15 have the acquired immunodeficiency syndrome,16 are chronic opium users,17 or have hyperlipidemia.18 Although biliary SOD has been diagnosed in all age groups, it is most common in middle-aged women. The female preponderance varies from 75% to 90%. The pain

Table 63-1  Modified Milwaukee Classification for Biliary Sphincter of Oddi Dysfunction

TYPE

CLINICAL CRITERIA

I

Biliary-type pain; serum ALT, AST, or AP level elevated on one occasion; and bile duct diameter ≥10 mm* Biliary-type pain One of the other two criteria for biliary type I SOD Biliary-type pain only

II III

PROBABILITY OF PAIN RELIEF (%) with SPHINCTEROTOMY IF SOM FINDING IS

APPROXIMATE FREQUENCY OF ABNORMAL SOM (%)

Abnormal

65-85

90-95

90-95

65

85

35

59

55-60

Normal

<10

ALT, alanine aminotransferase; AP, alkaline phosphatase; AST, aspartate aminotransferase; SOD, sphincter of Oddi dysfunction; SOM, sphincter of Oddi manometry. *Original Milwaukee criteria required cholangiography. Modified from Eversman D, Fogel EL, Rusche M, et al. Frequency of abnormal pancreatic and biliary sphincter manometry compared with clinical suspicion of sphincter of Oddi dysfunction. Gastrointest Endosc 1999; 50:637-41.

Chapter 63  Biliary Tract Motor Function and Dysfunction Table 63-2  Clinical Associations with Sphincter of Oddi Dysfunction Strong Probable Possible

Biliary-type pain postcholecystectomy Biliary-type pain in a patient with an intact gallbladder Idiopathic acute recurrent pancreatitis Acquired immunodeficiency syndrome– associated viral and protozoal infections After liver transplantation Chronic pancreatitis Hyperlipidemia Opium use

is typical of biliary pain; it is severe and occurs in the epigastrium or the right upper quadrant and may radiate to the back or right shoulder blade. The pain is episodic, lasts more than 30 minutes, and occurs at least once a year.19 Less than half of affected patients have abnormal liver bioche­ mical test results with or without the pain, although transient elevation of serum aminotransferase levels during attacks of pain supports the diagnosis of SOD.

CLASSIFICATION

Patients with suspected biliary SOD are classified into three categories based on diagnostic criteria, which have undergone modification, known as the Milwaukee classification system. This system is relevant in clinical practice because of its ability to predict the outcome of biliary sphincter ablation (see later). According to the modified Milwaukee classification system (see Table 63-1), type I SOD is diagnosed in patients with biliary-type pain, serum liver enzyme (aminotransferase or alkaline phosphatase) levels that are elevated (more than 1.1 times the upper limit of normal), and bile duct dilatation to a diameter greater than 9 mm. Type II SOD is defined as biliary-type pain and either elevated liver enzyme levels or a dilated bile duct. Type III SOD is defined as biliary-type pain without any of the other objective abnormalities. This classification system does not require that the liver enzyme elevations be timed with attacks of pain, although such an association may be a predictor of response to treatment. The original Milwaukee classification system incorporated more stringent criteria, including aminotransferase or alkaline phosphatase elevations greater than twice the upper limit of normal on two separate occasions, biliary dilatation greater than 12 mm, and delayed drainage (>45 minutes) of bile into the duodenum at cholangiography. The stringency of these criteria has been questioned,20-22 and their use in clinical practice has diminished. A similar classification system for possible pancreatic SOD has been proposed.23 Patients with pancreatic SOD type I have pancreatic-type pain, a serum amylase or lipase level of at least 1.1 times the upper limit of normal on one occasion, and pancreatic duct dilatation (>6 mm in the head and >5 mm in the body); those with pancreatic SOD type II have pain and one of the other criteria; and those with pancreatic SOD type III have pancreatic-type pain only. Several studies have demonstrated a high frequency (30% to 65%) of SO hypertension in patients with unexplained recurrent pancreatitis and a frequency of 50% to 87% in those with chronic pancreatitis.24 The usefulness of this classification system for pancreatic SOD has not been tested and will depend on studies of the symptomatic outcome following pancreatic sphincter ablation.

DIAGNOSIS Noninvasive Tests

Evaluation of patients in whom SOD is suspected is initiated with liver biochemical testing, serum amylase and lipase measurements, and often abdominal imaging with either abdominal ultrasonography or computed tomography (CT). Serum aminotransferase elevations, if present, are mild, less than three to four times the upper limit of normal. Physical examination findings are usually normal, although mild right upper quadrant or epigastric tenderness may be present. Standard evaluation, including routine endoscopic and imaging studies, and therapeutic trials for more common causes of abdominal pain, such as gastroesophageal reflux disease, fatty liver, and irritable bowel syndrome, have usually been undertaken as well. Noninvasive tests for SOD include biliary scintigraphy; fatty meal, CCK, or secretin-stimulated ultrasonography; secretin-stimulated magnetic resonance imaging; and secretin-stimulated endoscopic ultrasonography. Biliary scintigraphy can be used to assess flow of bile into the duodenum and has been proposed as a safe screening test before SOM. Although scintigraphy findings are usually positive in patients with dilated bile ducts and high-grade biliary obstruction, the modality lacks sufficient sensitivity in patients with lower-grade obstruction (Milwaukee classification types II and III SOD).25 After a lipid-rich (fatty) meal or intravenous administration of CCK, the bile duct may dilate under pressure if the SO is dysfunctional, and this change can be detected on transcutaneous ultrasonography. Compared with SOM in postcholecystectomy patients, fatty meal ultrasonography has a sensitivity of 21% and a specificity of 97% for SOD.26 Similarly, after stimulation by intravenously administered secretin, the pancreatic duct may dilate, and this test therefore can be used to assess pancreatic sphincter dysfunction. Compared with SOM, secretin-stimulated ultrasonography testing has a sensitivity of 88% and a specificity of 82% for SOD in patients with recurrent acute pancreatitis.27 Secretin-stimulated magnetic resonance pancreatography has also been used to assess pancreatic outflow obstruction in patients with idiopathic acute recurrent pancreatitis. Preliminary reports have shown high specificity but low sensitivity rates compared with SOM.28 Similarly, in one study secretin-stimulated endoscopic ultrasono­graphy showed limited sensitivity in the evaluation of patients with recurrent pancreatitis and manometrically proved SOD.29 Therefore, noninvasive diagnostic tests for SOD lack sufficient diagnostic accuracy to replace SOM and do not yet play a clear role in the diagnostic algorithm for this condition.

Invasive Tests

Patients in whom SOD is suspected have the highest complication rates for endoscopic retrograde cholangiopancreatography (ERCP). Such patients have a three-fold increase in the risk of post-ERCP pancreatitis, with absolute rates exceeding 25%.30 Although the prophylactic placement of a temporary pancreatic stent reduces the risk of post-ERCP pancreatitis,30,31 substantial morbidity and occasional mortality occur in patients who undergo ERCP for SOD. Therefore, ERCP with manometry should be reserved for persons who have severe or debilitating symptoms. Cholangiography is essential for excluding stones or tumors as the cause of biliary obstruction and associated symptoms. Although alternative biliary imaging methods, such as magnetic resonance cholangiopancreatography and endoscopic ultrasonography, are safer than ERCP for excluding stones, tumors,

1069

Section VIII  Biliary Tract

Technique.  SOM uses pressure recording equipment and infusion systems similar to those used for esophageal motility studies. The infusion rate is 0.25 mL per channel using a low-compliance pump. Performance of manometry requires a two-person approach, with one person stationed at the recorder. Triple-lumen 5-French catheters are available in long-nose and short-nose types. The long-nose catheter has the advantage in the bile duct of allowing several pull-throughs without loss of cannulation, although in tortuous pancreatic ducts the nose is occasionally too long for free cannulation. The three orifices are spaced 2 mm apart and are oriented radially (Fig. 63-2). The middle port is used for aspiration, which has been shown in a controlled trial to lower the risk of pancreatitis.38 This port also allows the cannulated duct to be identified easily by the color of the aspirate: yellow from the bile duct and clear from the pancreatic duct. The middle port can accept a 0.018-inch

350 300 mm Hg

Sphincter of Oddi Manometry SOM is usually performed during ERCP, although it can be done in the operating room or via a percutaneous approach. All drugs that relax the SO (nitrates, calcium channel blockers, and anticholinergics) or stimulate it (narcotics and cholinergics) should be avoided for 12 hours prior to manometry. Diazepam,33 meperidine (in a maximum dose of 1 mg/kg), and fentanyl (maximum dose of 1 mg/kg)34 do not affect basal SO pressure. Studies suggest that midazolam and droperidol lower the basal SO pressure in hypertensive sphincters,35,36 although they are still used for sedation by some experts in the field. Glucagon should also be avoided, although some authorities use it if necessary to achieve cannulation and wait at least 8 to 10 minutes until sphincter function is restored before measuring pressures. In vitro and clinical data suggest that propofol does not affect the SO in a clinically relevant fashion and can be used for sedation during SOM.37 No prospective trials have evaluated the effect of inhaled general anesthetics on the SO, although these agents are used routinely in patients undergoing SOM who are difficult to sedate.

guidewire, thereby allowing wire exchange in difficult cannulations. When SOM is clinically indicated, many experts begin ERCP with the manometry catheter, because duodenal motility is minimized if contrast medium has not been injected. The duodenal, or zero, pressure should be measured before cannulation and after completion of manometric measurements in the SO. The catheter is withdrawn across the SO at 1- to 2-mm intervals by means of a standard station pull-through technique (Fig. 63-3). Abnormalities of the basal SO pressure should be observed on at least two pull-throughs. Depending on the clinical indication, pancreatic SOM may be performed via the same technique. Abnormal basal sphincter pressures are usually concordant for the two ducts but may occur in only the biliary or pancreatic portion of the sphincter.39 Increased basal sphincter pressure is more likely to be confined to the bile duct in

5 seconds

250 200 150 100

Tip (distal)

50 0 350 300 mm Hg

and pancreas divisum, they cannot diagnose (or be used to treat) SOD. Occasionally, an intra-ampullary neoplasm may simulate SOD. If there appears to be excess tissue in the ampulla after endoscopic sphincterotomy, biopsy specimens of the area should be obtained.32

250 200

Middle

150 100 50 0 450 400 350

mm Hg

1070

300 250

Proximal SO phasic pressure

200 150 100 50 0

Figure 63-2.  Aspiration manometry catheter. Diagram of a 5-French sphincter of Oddi manometry (SOM) catheter (Wilson Cook model no. SOM-21-S-Lehman, Wilson Cook, Winston-Salem, North Carolina). The arrows depict the direction of fluid movement; outward flow represents perfusion, and inward flow represents aspiration.

Duodenum Sphincter of Oddi (SO) basal pressure

Phasic Bile duct duration pressure

Figure 63-3.  Station pull-through tracing obtained through a triple-lumen catheter during sphincter of Oddi (biliary) manometry. This tracing is abnormal and shows an elevated mean basal pressure of approximately 70 mm Hg as well as phasic contractions of the sphincter of Oddi.

Chapter 63  Biliary Tract Motor Function and Dysfunction persons with elevated serum liver enzyme levels and more likely to be confined to the pancreatic duct in patients with pancreatitis.40 For patients in whom the clinical indication for SOM is biliary pain and not idiopathic pancreatitis, and in whom biliary SOM produces normal findings, some authorities avoid pancreatic cannulation entirely to reduce the frequency of pancreatitis. Other experts advise studying both ducts in all patients, with the intention of performing dual sphincterotomies if either duct is hypertensive, regardless of whether the indication is biliary pain or idiopathic pancreatitis. Further studies are necessary to determine the preferable strategy. When the clinical indication for SOM is idiopathic recurrent pancreatitis and pancreatic sphincterotomy is contemplated, pancreatic manometry is mandatory. After the tracings are completed, glucagon may be administered intravenously to decrease duodenal motility, and additional meperidine or fentanyl may be given for sedation to facilitate subsequent contrast injection or endoscopic therapy. If a cholangiogram is desired, the aspirating port can be used for contrast injection. In patients with suspected SOD, placement of a temporary pancreatic stent lowers the risk of post-ERCP pancreatitis, regardless of whether or not manometric measurements are abnormal.30,41 The standard upper limit of normal for baseline biliary sphincter pressure is 35 to 40 mm Hg. Normal pancreatic sphincter pressures are accepted as the same as those for the bile duct, although reference data are more limited. The phasic wave frequency, propagation direction of waves, and amplitude of the waves can also be determined, although their clinical use has been abandoned. Additional pharmacologic maneuvers, such as provocation with CCK, are also of doubtful value at this time. Diagnostic Use.  A landmark randomized controlled study of patients with suspected type II biliary SOD conducted by Geenen and colleagues42 established that SOM predicts improvement in pain after endoscopic sphincterotomy. Patients with a basal SO pressure greater than 40 mm Hg had a clinical response rate of 91%, compared with a 25% rate in patients with a high basal pressure in whom a sham sphincterotomy was performed. For patients with a normal SO pressure, the response to sphincterotomy was only 42% and similar to that after the sham procedure (33%). These results were confirmed in a second controlled study of patients with type II SOD and an elevated sphincter pressure.43 In this study, clinical improvement was demonstrated in 11 of 13 patients treated with sphincterotomy, compared with 5 of 13 control subjects treated with sham sphincterotomy. There was no difference in pain improvement between sphincterotomy and sham sphincterotomy in patients with manometric abnormalities other than elevated basal SO pressure, namely, “tachyoddia” (increased phasic wave frequency), increased retrograde contractions, and para­doxical response to CCK. Despite the findings of these studies, the use of SOM as a diagnostic tool remains somewhat controversial. Some uncontrolled studies suggest that more easily measurable criteria, such as elevated liver enzyme levels and biliary dilatation, are superior in predicting a response to sphincter ablation.44 Other studies suggest that manometry is highly specific for diagnosing SOD but may lack sensitivity; lack of sensitivity may account for the 42% symptom response rate to sphincterotomy in patients with biliary SOD type II and normal manometric results. A lack of sensitivity also may explain the relatively low rate of abnormal SOM results (65% to 85%) in patients with type I biliary SOD, in whom

the response rate to sphincterotomy is greater than 90%.45 It is hypothesized that the relatively low frequency of sphincter hypertension in patients with type I biliary SOD is the result of a different pathogenesis of sphincter obstruction, namely, sphincter stenosis rather than sphincter hypertension. Another possible explanation for the insensitivity of SOM is that short-term observation of sphincter pressure may not detect intermittent spasm that is not occurring at the time of the procedure. In one study, results of a second SOM were abnormal in 5 of 12 (42%) persistently symptomatic patients with an initially normal SOM result.46 An additional problem with manometry is that it is a dif­ficult technique to perform, is not widely available, and has a success rate of only 75% to 92% in the most experienced hands. Manometrically proved SOD appears to be less common in patients with type III biliary SOD than in those with type II SOD, and the response to sphincter ablation is only 39% to 60%.47 A response to sphincter ablation in patients with type III biliary SOD and normal SOM findings is infrequent. Obviously, pain alone is a poor indicator of any specific motility disorder. Abnormal small bowel interdigestive motor activity48 and duodenal visceral hyperalgesia in response to duodenal (but not rectal) distention49 have been demonstrated in patients with type III biliary SOD. Placement of a pancreatic or biliary stent on a trial basis to predict a response to subsequent sphincterotomy and injection of botulinum toxin into the SO have been proposed as alternative methods of diagnosing SOD.50,51 Although preliminary data have suggested some utility to these approaches, the need for multiple procedures, with their attendant risks, and the risk of stent-induced pancreatic ductal damage have limited their widespread application.

TREATMENT Medical Therapy

Dietary and medical therapy for suspected or documented SOD has undergone minimal study. A low-fat diet is recommended for reducing pancreaticobiliary stimulation, although no data are available to substantiate this approach. Nifedipine, nitrates, octreotide, and antispasmodics have been shown to lower basal SO pressure, although consistent clinical outcomes data are lacking. Two short-term, placebocontrolled crossover studies showed that 75% of patients with suspected or documented SOD experienced statistically less pain with use of oral nifedipine.52,53 A more recent study, however, demonstrated that slow-release nifedipine provided no clinical benefit but increased cardiovascular side effects compared with placebo.54 A single prospective study of nitrates in SOD published in abstract form revealed reduction in pain, but therapy was also limited by side effects.55 Despite these limited data, however, in light of the benign nature of SOD, medical therapy should be attempted in all patients with suspected type III SOD and in patients with less severe type II SOD before sphincter ablation is offered. A trial of antispasmodics, such as hyoscyamine, or a low-dose tricyclic antidepressant (to reduce visceral hypersensitivity) is often attempted in these patients. Patients with type II SOD and more severe pain are less likely to respond to medical therapy and can be considered for an initial trial of endoscopic therapy.

Sphincterotomy

Historically, surgical biliary sphincterotomy and sphincteroplasty were used successfully to ablate the SO. Endo-

1071

1072

Section VIII  Biliary Tract scopic techniques have largely replaced open surgery for biliary as well as pancreatic sphincter ablation. Most data on endoscopic sphincterotomy relate to biliary sphincter ablation alone. The most common indication for SOM is biliary-type pain in a postcholecystectomy patient. If SOM findings are abnormal, relief of abdominal pain after sphincterotomy occurs in 90% to 95% of patients with type I biliary SOD, 85% of those with type II SOD, and 55% to 60% of those with type III SOD. When the SOM result is normal, pain relief after sphincterotomy still occurs in 90% to 95% of patients with type I biliary SOD. Because SOM findings may be misleading in these cases (they are normal in 14% to 35% of patients), SOM is not clinically indicated for patients with type I biliary SOD. Rather, endoscopic sphincterotomy can be performed empirically. Pain relief after sphincterotomy occurs in 35% to 42% of patients with type II biliary SOD and normal SOM results. Although this response rate is similar to that in sham-treated controls, a true clinical response likely occurs in a few patients. Sphincterotomy is clearly indicated in patients with type II SOD and abnormal SOM findings, but whether SOM is required to justify sphincterotomy in this group remains controversial. Some authorities advocate empirical biliary sphincterotomy in patients with type II SOD. This strategy has the advantage of providing therapy to those patients with normal SOM who respond to sphincterotomy, but the advantage comes at the expense of additional procedure-related complications, such as bleeding and perforation. A decision-analysis model revealed that a strategy of empirical sphincterotomy by an experienced biliary endoscopist in patients meeting criteria for type II SOD is cost-effective compared with a SOM-driven strategy.56 Further prospective studies are necessary to determine if the risks of empirical sphincterotomy in patients with type II SOD are offset by the clinical and economic benefits. In patients with type III biliary SOD and normal SOM findings, the clinical response rate to sphincterotomy is less than 10%. Therefore, an abnormal SOM result is mandatory before sphincterotomy is performed in a patient with type III SOD. Some therapeutic endoscopists do not offer ERCP with manometry to patients with clinical type III SOD because the risk-benefit ratio can be unfavorable. Others, however, offer this intervention to patients with a compelling clinical history, severe symptoms, and lack of response to medical therapy, provided that the risks of the procedure are well understood by the patient. Additional clinical outcomes data in patient with type III SOD are forthcoming. Few studies have addressed SOD in patients with biliarytype pain, an intact gallbladder, and no gallstones. Whether cholecystectomy is pathophysiologically responsible for the development of SOD or whether patients with SOD are simply more likely to have undergone cholecystectomy because of the nature of their symptoms is unclear. Cholecystectomy has been postulated to unmask preexisting subclinical SOD by removing the reservoir that serves to decompress the extrahepatic biliary system during SO spasm.57 Further, nerves that travel from the gallbladder to the SO via the cystic duct are severed during cholecystectomy, potentially leading to altered SO motility.58 Limited data, however, suggest that SOD occurs in patients with an intact gallbladder. Of patients with documented SOD and an intact gallbladder who are treated with sphincterotomy first, 43% have long-term pain relief; some additional patients eventually improve following cholecystectomy.59 Clearly, more information is needed on how to assess and treat this challenging group of patients.

SPHINCTER OF ODDI DYSFUNCTION IN PANCREATITIS IDIOPATHIC ACUTE RECURRENT PANCREATITIS

SOD has been found in 25% to 60% of patients with idiopathic acute recurrent pancreatitis. Recurrent attacks of pancreatitis appear to be prevented by pancreatic sphincterotomy in 60% to 80% of patients with manometrically proved pancreatic sphincter hypertension, although only uncontrolled studies have been conducted.60 Endoscopic pancreatic therapy carries a higher risk of acute and longterm complications, including post-ERCP pancreatitis and papillary restenosis; the latter can sometimes lead to the development of more frequent episodes of pancreatitis or chronic pain. For the patient with an intact gallbladder and unexplained pancreatitis, some authors advocate either biliary sphincterotomy, treatment with ursodeoxycholic acid, or empirical cholecystectomy, with the implication that microlithiasis is the cause (see Chapters 58 and 65).61 Others report that biliary sphincterotomy alone benefits only one third of these patients, whereas dual (biliary and pancreatic) sphincterotomies benefit 80%, thereby suggesting that pancreatic SOD plays an important role and that pancreatic sphincter therapy must be performed along with biliary sphincterotomy.62 More studies are required to determine the preferred approach (biliary, pancreatic, or dual sphincterotomies) and to clarify the rates of success and complications associated with each approach.

CHRONIC PANCREATITIS

SOD has been described in 50% to 87% of patients with chronic pancreatitis.24 Whether SOD is the result of chronic inflammation or independently plays a role in the pathogenesis of chronic pancreatitis is not known. Endoscopic pancreatic sphincterotomy improves pain scores in 60% to 65% of patients with pancreatic SOD, although controlled studies are not available.63,64 In some cases, pancreatic sphincterotomy must be performed to facilitate other therapeutic maneuvers, such as pancreatic ductal stone extraction and stricture dilation. The role of SOM in chronic pancreatitis remains unclear.

FAILURE OF RESPONSE TO BILIARY SPHINCTEROTOMY Possible explanations for a lack of response to biliary sphincterotomy in patients with SOD are listed in Table 63-3. Perhaps the most likely explanation is that the pain was not of pancreaticobiliary origin but was caused instead by altered gut motility or visceral hypersensitivity.49 Alternatively, the biliary sphincterotomy may have been inade-

Table 63-3  Possible Causes for Failure to Achieve Pain Relief after Biliary Sphincterotomy in Patients with Presumed Sphincter of Oddi Dysfunction* Inadequate initial sphincterotomy Non–pancreaticobiliary pain, especially functional gastrointestinal disease Occurrence of restenosis Residual pancreatic sphincter hypertension Subtle chronic pancreatitis with a normal pancreatogram *In alphabetical order.

Chapter 63  Biliary Tract Motor Function and Dysfunction quate, or restenosis may have occurred.65 The likelihood of clinical success of further biliary endoscopic treatment or of surgical sphincteroplasty in such cases is unknown. The role of residual pancreatic sphincter hypertension as a source of continuing pain in the absence of pancreatic abnormalities is unclear. Some experts advocate initial dual sphincterotomies to prevent this problem, although the reintervention rate for persistent or recurrent pain has not been different from that for historical controls in whom a single sphincterotomy (of one duct) was performed.66 Finally, some patients in whom SOD is suspected and who have shown no response to biliary sphincterotomy may have subtle chronic pancreatitis and normal pancreatographic findings. Endoscopic ultrasonography may demonstrate parenchymal changes consistent with chronic pancreatitis in these patients (see Chapter 59).67

KEY REFERENCES

Geenen JE, Hogan WJ, Dodds WJ, et al. The efficacy of endoscopic sphincterotomy after cholecystectomy in patients with sphincter of Oddi dysfunction. N Engl J Med 1989; 320:82-7. (Ref 42.) Guelrud M, Plaz J, Mendoza B, et al. Endoscopic treatment in type II pancreatic sphincter dysfunction. Gastrointest Endosc 1999; 41:398. (Ref 62.)

Rolny P, Geenen JE, Hogan WJ, et al. Clinical features, manometric findings and endoscopic therapy results in group I patients with sphincter of Oddi dysfunction. Gastrointest Endosc 1991; 37:252. (Ref 45.) Sherman S. Idiopathic acute pancreatitis: Role of ERCP in diagnosis and therapy. ASGE Clinical Update 2004; 12:1. (Ref 60.) Sherman S, Troiano FP, Hawes RH, et al. Sphincter of Oddi manometry: Decreased risk of clinical pancreatitis with use of a modified aspirating catheter. Gastrointest Endosc 1990; 36:462-6. (Ref 38.) Sherman S, Troiano FP, Hawes RH, et al. Frequency of abnormal sphincter of Oddi manometry compared with the clinical suspicion of sphincter of Oddi dysfunction. Am J Gastroenterol 1991; 86:586-90. (Ref 14.) Singh P, Das A, Isenberg G, et al. Does prophylactic pancreatic stent placement reduce the risk of post-ERCP pancreatitis? A meta-analysis of controlled trials. Gastrointest Endosc 2004; 60:544-50. (Ref 31.) Tarnasky PR, Hoffman B, Aabakken L, et al. Sphincter of Oddi dysfunction is associated with chronic pancreatitis. Am J Gastroenterol 1997; 92:1125-9. (Ref 24.) Testoni PA, Caporuscio S, Bagnolo F, et al. Idiopathic recurrent pancreatitis: Long-term results after ERCP, endoscopic sphincterotomy, or ursodeoxycholic acid treatment. Am J Gastroenterol 2000; 95:1702-7. (Ref 61.) Toouli J, Roberts-Thomson IC, Kellow J, et al. Manometry based randomized trial of endoscopic sphincterotomy for sphincter of Oddi dysfunction. Gut 2000; 46:98-102. (Ref 43.) Full references for this chapter can be found on www.expertconsult.com.

1073

CHAPTE R

64 Bile Secretion and the Enterohepatic Circulation Paul A. Dawson

CHAPTER OUTLINE Bile Acid Synthesis and Metabolism  1076 The Enterohepatic Circulation  1079 Hepatic Bile Acid Transport and Bile Secretion  1080 Bile Acid–Independent Bile Flow  1080 Cholehepatic Shunt Pathway  1080 Hepatic Bile Acid Transport  1081 Hepatic Sinusoidal Na+-Dependent Bile Acid Uptake  1081 Hepatic Sinusoidal Na+-Independent Bile Acid Uptake  1081 Hepatic Sinusoidal Bile Acid Efflux  1083 Canalicular Bile Acid Transport  1083 Intestinal and Renal Bile Acid Transport  1083 Molecular Mechanisms of Ileal and Renal Bile Acid Transport  1084

Bile formation is essential for intestinal lipid digestion and absorption, cholesterol homeostasis, and hepatic excretion of lipid-soluble xenobiotics, drug metabolites, and heavy metals. The process of bile formation depends on hepatic synthesis and canalicular secretion of bile acids, the predominant organic anions in bile, and maintenance of hepatic bile formation is essential for normal liver function. Most of the bile acids secreted by the hepatocyte have been secreted previously into the small intestine and have undergone enterohepatic cycling. As a result, disturbances in bile acid synthesis, biliary secretion, and intestinal absorption have profound effects on hepatic and gastrointestinal physiology. Identification of the hepatic, biliary, and intestinal bile acid transporters has advanced our understanding of genetic and acquired disorders of bile formation and secretion. Transporter gene mutations have been identified in disorders such as progressive familial intrahepatic chole­ stasis (PFIC) types 1 to 3, intrahepatic cholestasis of pregnancy (ICP), low phospholipid-associated cholelithiasis (LPAC), Dubin-Johnson syndrome, and primary bile acid malab­sorption (PBAM).1 In addition, developments in the nuclear and G-protein–coupled receptor fields have provided new insights into the regulation of bile acid synthesis, secretion, and enterohepatic cycling.2 This chapter reviews the current knowledge of the hepatic synthesis, biliary secretion, and enterohepatic circulation of bile acids. Bile is a complex, lipid-rich micellar solution that is isosmotic with plasma and composed primarily of water, inorganic electrolytes, and organic solutes such as bile acids, phospholipids (mostly phosphatidylcholine), cholesterol, and bile pigments (Table 64-1). The relative proportions of the major organic solutes in bile is illustrated in Figure 64-1. The volume of hepatic bile secreted is estimated to range

Disorders of the Enterohepatic Circulation  1084 Bile Acid Synthesis  1084 Membrane Transport of Bile Acids and Biliary Lipids  1085 Bile Acid Biotransformation (Deconjugation and Dehydroxylation)  1086 Bile Acid Circulation  1086 Bile Acid Therapy, Sequestrants, and Transport Inhibitors  1087 Bile Acid Therapy  1087 Bile Acid Sequestrants  1088 Bile Acid Transport Inhibitors  1088

from 500 to 600 mL per day, and bile acids are the dominant organic components. Actively secreted across the canalic­ ular membrane, bile acids induce the secretion of other biliary constituents. In healthy humans, canalicular secretion is efficient and remarkably concentrative; the intracellular monomeric concentration of bile acid is less than 10 mmol/L in the hepatocyte and more than 1000 mmol/L in canalicular bile. Bile acids travel down the biliary tree and are stored in the gallbladder. After a meal the gallbladder contracts and empties its contents into the duodenum, where bile acids facilitate absorption of cholesterol and fat-soluble vitamins. Bile acids are poorly absorbed in the proximal small intestine but are absorbed almost completely by the terminal ileum. The bile acids are returned to the liver in the portal circulation, actively absorbed at the hepatocyte sinusoidal membrane, and re-secreted into bile.3 The functions of bile acids in the liver and gastrointestinal tract are multiple.3,4 First, bile acids induce bile flow and hepatic secretion of biliary lipids (phospholipid and cholesterol). The vectorial movement of bile acids from blood into the bile canaliculus generates an osmotic water flow and is a major determinant of bile formation. Second, bile acids play an important role in the digestion of dietary fats and are essential for the intestinal absorption of cholesterol and fat-soluble vitamins. Bile acids promote intestinal absorption by solubilizing dietary lipids and lipid digestion products as mixed micelles to facilitate their aqueous diffusion across the intestinal mucosa. Fat-soluble vitamins (A, D, E, and K) are poorly absorbed in the absence of bile acid micelles, and disturbances in the secretion or enterohepatic cycling of bile acids lead to fat-soluble vitamin deficiency. Along with their major role in promoting dietary lipid

1075

1076

Section VIII  Biliary Tract Table 64-1  Composition of Hepatic Bile COMPONENT

CONCENTRATION

Electrolytes and minerals (mmol/L):   Sodium   Potassium   Chloride   Bicarbonate   Calcium   Phosphate   Magnesium Metals (µmol/L):   Iron   Copper Organic constituents (mmol/L):   Bile acids   Bilirubin (total)   Phospholipid (lecithin)   Cholesterol   Glutathione   Glucose   Urea   Protein (g/dL)

140-160 3-8 70-120 20-50 1-5 0-1.2 1-3 2-72 12-21 5-50 1-2 0.5-20.0 0.5-1.0 3-5 0.2-1.0 2.2-6.5 0.2-3.0

Values obtained from measurements of human bile are drawn from Albers CJ, Huizenga JR, Krom RA, et al. Composition of human hepatic bile. Ann Clin Biochem 1985; 22:129-32; Keulemans YC, Mok KS, de Wit LT, et al. Hepatic bile versus gallbladder bile: A comparison of protein and lipid concentration and composition in cholesterol gallstone patients. Hepatology 1998; 28:11-16; and Ho KJ. Biliary electrolytes and enzymes in patients with and without gallstones. Dig Dis Sci 1996; 41:2409-16.

Phospholipid Cholesterol Bilirubin Protein

Bile acid

Fecal bile acids

Biliary bile acids

Cholic

Chenodeoxycholic

Lithocholic Deoxycholic

Deoxycholic

Lithocholic and ursodeoxycholic

Figure 64-1.  Top, Typical solute composition of hepatic and gallbladder bile in healthy humans. Bile acids are the primary solute in bile, con­ stituting approximately 67% of bile. Phospholipid accounts for approximately 22%, cholesterol 4%, bilirubin 0.3%, and protein 4.5% of bile. Biliary bile acid composition is shown in the bottom left. Cholic acid, chenodeoxycholic acid, and deoxycholic acid constitute more than 95% of the biliary bile acids, and virtually all of the biliary bile acids are in conjugated form. The proportion of biliary lithocholic acid and ursodeoxycholic acid varies but rarely exceeds 5%. The majority of lithocholic acid in bile is present in sulfated form. When administered in therapeutic doses, the proportion of ursodeoxycholic acid in bile may rise to as much as 40%. Fecal bile acid composition is shown in the bottom right. Fecal bile acids are almost entirely unconjugated as a result of bacterial deconjugating enzymes in the colon and consist primarily of the dehydroxylated bile acids, deoxycholic acid, and lithocholic acid (see Figure 64-2).

absorption, bile acids may facilitate intestinal absorption of protein by accelerating protein hydrolysis by pancreatic proteases.5 Third, bile acids play a complex role in maintaining cholesterol homeostasis. On one hand, bile acids promote cholesterol intake by facilitating the intestinal absorption of biliary and dietary cholesterol. On the other hand, bile acids function through several mechanisms to promote cholesterol elimination from the body. Bile acids are water-soluble end products of cholesterol catabolism and a major route for the elimination of cholesterol via fecal excretion. Bile acids also promote hepatic secretion of cholesterol into bile by inducing bile flow and solubilizing biliary cholesterol, thereby enabling cholesterol to move from the hepatocyte to the intestinal lumen for elimination. Fourth, bile acids contribute to the antimicrobial defense mechanisms of the intestine through direct bacteriostatic actions of bile acid–fatty acid mixed micelles in the proximal small intestine and by inducing expression of antimicrobial genes in the distal small intestine.6,7 Fifth, bile acids act to prevent the formation of calcium gallstones and oxalate kidney stones.8 Sixth, bile acids can function as hormones to regulate the enterohepatic circulation as well as fat, glucose, and energy homeostasis by signaling through nuclear and G-protein–coupled receptors.2

BILE ACID SYNTHESIS AND METABOLISM Bile acids are synthesized from cholesterol in pericentral hepatocytes of the hepatic acini. In this process, cholesterol, a lipophilic compound, is converted into a water-soluble product. In humans the newly synthesized (primary) bile acids are cholic acid (CA), a trihydroxy-bile acid with hydroxy groups at the C-3, C-7, and C-12 positions, and chenodeoxycholic acid (CDCA), a dihydroxy-bile acid with hydroxy groups at the C-3 and C-7 positions (Fig. 64-2). The kinetics for bile acid synthesis and turnover in humans are summarized in Table 64-2. Hepatic bile acid synthesis involves two major pathways, the “classical” neutral pathway (cholesterol 7α-hydroxylase pathway) that favors CA biosynthesis and the “alternative” acidic pathway (oxysterol 7α-hydroxylase pathway) that favors CDCA biosynthesis.9 In the classical pathway, the enzyme cholesterol 7α-hydroxylase (cytochrome P-450 7A1 [CYP7A1]) converts cholesterol directly into 7α-hydroxycholesterol. In the alternative pathway, cholesterol must first be converted by C-24, C-25, or C-27 sterol hydroxylases into an oxysterol, the major species being 27-hydroxycholesterol, before conversion by the oxysterol 7α-hydroxylase (CYP7B1). The overall process of bile acid biosynthesis is complex, involving 17 different enzymes divided into two broad groups.9 The first group performs modifications to the sterol ring structure, whereas the second group modifies the sterol side chain. In the classical pathway, sterol ring modifications precede side chain changes, whereas in the alternative pathway the side chain modifications occur before or during changes to the sterol ring structure. Of the two major biosynthetic pathways, the classical (CYP7A1) pathway is quantitatively more important in humans. This conclusion is supported by the finding that bile acid production is decreased by almost 90% in an adult patient with an inherited mutation in the CYP7A1 gene.10 In contrast to adults, the alternative pathway may be dominant in neonates, as evidenced by the low expression of CYP7A1 in newborns and the finding of severe cholestatic liver disease in an infant with an inherited CYP7B1 mutation.11

Chapter 64  Bile Secretion and the Enterohepatic Circulation

HO

7

3

Cholesterol Liver

OH COOH

Primary HO

COOH

OH Cholic acid

HO

OH Chenodeoxycholic acid

Intestinal bacteria

OH

HO

Deoxycholic acid

COOH

COOH

COOH

Secondary

HO

HO

Lithocholic acid

O 7-oxo lithocholic acid Intestinal bacteria

Liver

Liver COOH

COOH –O SO 3

OH

HO Sulfolithocholic acid

Ursodeoxycholic acid

Figure 64-2.  Major primary and secondary bile acids and their sites of synthesis and metabolism. Secondary metabolism of bile acids includes 7α-dehydroxylation by the intestinal flora, deconjugation by the intestinal flora, epimerization of the 3α- and 7α-hydroxy groups by the intestinal flora, hepatic reduction of the 7-oxo derivative of chenodeoxycholic acid to 7-oxo lithocolic acid, and hepatic re-epimerization of 3β-hydroxy bile acids. Bile acids are also sulfated at the 3 or 7 positions by the liver and kidney. (Adapted with permission from Carey MC, Cahalane MJ. Enterohepatic circulation. In: Arias IM, Jakoby WB, Popper H, et al, editors. The Liver: Biology and Pathology. 2nd ed. New York, NY: Raven Press; 1988. p 591.)

Table 64-2  Pool Size and Kinetics of Individual Bile Acids in Healthy Subjects

BILE ACID Cholate Deoxycholate Chenodeoxycholate Lithocholate Total

POOL SIZE (mg)

FRACTIONAL TURNOVER RATE (DAYS−1)

hepatic SYNTHESIS (mg/day)

DAILY INPUT from primary bile acids (mg/day)

500-1500 250-800 500-1200 50-150 1300-3650

0.2-0.5 0.1-0.4 0.2-0.4 0.8-1.0 —

120-400 0 100-250 0 220-650

— 40-200 — 50-100 90-300

Values for measurements of bile acid pool size, turnover, synthesis, and daily input are drawn from Vlahcevic ZR, Miller JR, Farrar JT, et al. Kinetic and pool size of primary bile acids in man. Gastroenterology 1971; 61:85-90; Cowen AE, Korman MG, Hofmann AF, et al. Metabolism of lithocholate in healthy man. II. Enterohepatic circulation. Gastroenterology 1975; 69:67-76; Everson GT. Steady-state kinetics of serum bile acids in healthy human subjects: Single and dual isotope techniques using stable isotopes and mass spectrometry. J Lipid Res 1987; 28:238-52; Berr F, Pratschke E, Fischer S, et al. Disorders of bile acid metabolism in cholesterol gallstone disease. J Clin Invest 1992; 90:859-68; Hulzebos CV, Renfurm L, Bandsma RH, et al. Measurement of parameters of cholic acid kinetics in plasma using a microscale stable isotope dilution technique: Application to rodents and humans. J Lipid Res 2001; 42:1923-9.

1077

1078

Section VIII  Biliary Tract The rate-limiting step for the classical pathway is the enzyme CYP7A1. Bile acid feedback inhibition of CYP7A1 is well established experimentally; bile acid synthesis is decreased after administration of hydrophobic bile acids and increased by interruption of the enterohepatic circulation following ileal resection or administration of bile acid sequestrants.12 In contrast to the classical (CYP7A1) pathway, the alternative pathway for bile acid synthesis is not subject to regulation by bile acids but is controlled by cholesterol delivery to the mitochondria, the site of 27-hydroxylation.13 The molecular mechanisms responsible for the negative feedback regulation of the CYP7A1 pathway involve the liver and small intestine and have been elucidated. For the major pathway, bile acids act as ligands for the farnesoid X receptor (FXR) in ileal enterocytes to induce synthesis of an endocrine hormone, the fibroblast growth factor-19 (FGF19). FGF19 is secreted into the portal circulation and acts on hepatocytes through its cell surface receptor, a complex of the β-klotho protein and fibroblast growth factor-4, to repress CYP7A1 expression and bile acid synthesis.14-16 In a second pathway, excess bile acids in the hepatocyte activate FXR and indirectly suppress expression of CYP7A1 through a complex mechanism that involves the orphan nuclear receptor designated small heterodimer partner (SHP). SHP interferes with the activity of several nuclear receptors including hepatocyte nuclear factor 4α (HNF4α) and liver receptor homolog 1 (LRH-1), which are required for expression of the CYP7A1 gene.12,17 Finally, bile acids also inhibit CYP7A1 gene transcription in a SHP and FGFR4-independent fashion through activation of the c-Jun NH2-terminal kinase (JNK) pathway.18,19 These complex molecular titrations link bile acid synthesis to changes in intestinal as well as hepatic bile acid levels. The receptors and protein factors that participate in the regulation of bile acid synthesis and the enterohepatic circulation of bile acids are summarized in Table 64-3. Before secretion into the bile canaliculus, both CA and CDCA are N-acyl amidated with glycine or taurine, a process commonly termed conjugation. This conjugation enhances the hydrophilicity of the bile acid and the acidic strength of the side chain, in essence converting a weak acid (pKa ≈ 5.0 for the unconjugated bile acid) to a strong acid (pKa ≈ 3.9 for the glycine conjugate; pKa ≈ 2.0 for the taurine conjugate). The major function of conjugation is to decrease the passive diffusion of bile acids across cell membranes during their transit through the enterohepatic circulation. As a result, conjugated bile acids are absorbed only if a specific membrane carrier is present. Compared with unconjugated bile acids, conjugated bile acids are also more soluble at acidic pH and more resistant to precipitation in the presence of high concentrations of calcium. The net effect of conjugation is to maintain high intraluminal concentrations of bile acids down the length of the small intestine to facilitate lipid digestion and absorption. The importance of bile acid conjugation is illustrated by the finding that inherited defects in bile acid conjugation present with the clinical phenotype of fat-soluble vitamin malabsorption and steatorrhea.11,20 Most of the conjugated bile acids secreted into the small intestine are efficiently absorbed intact; however, bile acids are also metabolized during their passage down the intestine, by the endogenous bacterial flora. In the distal small intestine, the bacterial flora deconjugates a small fraction, ≈15%, of the bile acids. The unconjugated bile acids are absorbed passively or actively and returned to the liver, where they are efficiently reconjugated and mix with newly synthesized bile acids to be re-secreted into bile. This process of intestinal deconjugation and hepatic reconjuga-

Table 64-3  Proteins Involved in the Regulation of Bile Acid Synthesis and Enterohepatic Cycling PROTEIN (GENE) FXR (NR1H4) HNF4α (NR2A1) SHP (NR0B2)

LRH-1 (NR5A2) PXR (NR1I2) VDR (NR1I1) FGFR4 (FGFR4) β-klotho (KLB)

FGF19 (FGF19)

DESCRIPTION AND FUNCTION of protein IN BILE ACID METABOLISM Bile acid–activated nuclear receptor; regulation of bile acid synthesis, transport, and metabolism Nuclear receptor; positive regulator of cytochrome P-450 7a1 (CYP7a1) expression and hepatic bile acid synthesis Nuclear receptor; negative feedback regulation of hepatic bile acid synthesis by antagonizing HNF4α, LRH-1; regulation of bile acid transport and metabolism Nuclear receptor; positive regulator of CYP7a1 expression and hepatic bile acid synthesis Bile acid and xenobiotic-activated nuclear receptor involved in detoxification of secondary bile acids Vitamin D and bile acid–activated nuclear receptor; involved in detoxification of LCA Membrane receptor; negative feedback regulation of CYP7a1 and hepatic bile acid synthesis Membrane co-receptor associated with FGFR4; confers liver specificity to FGFR4-FGF19 pathway; negative feedback regulation of CYP7a1 and hepatic bile acid synthesis Protein growth factor; secreted by intestine in response to bile acids; regulates hepatic bile acid synthesis via FGFR4:β-klotho

FXR, farnesoid X-receptor; FGF19, fibroblast growth factor 19; FGFR4, fibroblast growth factor receptor 4; HNF4α, hepatocyte nuclear factor 4 alpha; LCA, lithocolic acid; LRH-1, liver receptor homolog 1; PXR, pregnane X-receptor; SHP, small heterodimer partner; VDR, vitamin D receptor.

tion is a normal part of bile acid metabolism. An additional bacterial modification is the epimerization of the 3α-hydroxy or 7α-hydroxy groups to their corresponding 3β- or 7β-hydroxy forms.21,22 The 7α-hydroxy group of CDCA is epimerized to form the 3α,7β-dihydroxy bile acid, ursodeoxycholic acid (UDCA). After being absorbed from the intestine, UDCA is conjugated to taurine or glycine in the liver and circulates as a minor component, normally less than 5%, of the bile acid pool. In addition to endogenous formation, UDCA is used as a therapeutic agent in various forms of cholestatic liver disease (see Chapters 76 and 89).23 A small fraction of bile acids secreted into the small intestine escapes absorption and passes into the large intestine, where deconjugation is almost complete. In the colon, the action of bacterial 7α-dehydroxylase converts CA to deoxycholic acid (DCA), a dihydroxy bile acid with hydroxy groups at the C-3 and C-12 positions, and converts CDCA to lithocholic acid (LCA), a monohydroxy bile acid with a hydroxy group at position C-3 (see Fig. 64-2).21 The colon absorbs about 50% of the DCA formed as well as a small fraction of the LCA. After returning to the liver, DCA is reconjugated and circulates with the primary bile acids. Hepatic conjugation of the circulating bile acids is extremely efficient, so virtually all the biliary bile acids (primarily CA, CDCA, DCA, and UDCA) are in conjugated form. Bacterial deconjugation and dehydroxylation in the colon are also efficient, so the feces contain mainly the unconjugated secondary bile acids DCA and LCA (see Fig. 64-1). The secondary bile acids can be metabolized further by various pathways, including hepatic re-epimerization of 3β-hydroxy bile acids, hepatic reduction of 7-oxo-

Chapter 64  Bile Secretion and the Enterohepatic Circulation circulation, and kidneys (Fig. 64-3). At a fundamental level, the enterohepatic circulation of bile acids can be considered to consist of a series of storage chambers (gallbladder, small intestine), valves (sphincter of Oddi, ileocecal valve), mechanical pumps (bile canaliculi, biliary tract, small intestine), and chemical pumps (hepatocyte, cholangiocyte, and ileocyte). Efficient intestinal reabsorption and hepatic extraction of bile acids permit an effective recycling and conservation mechanism that largely restricts bile acids to the intestinal and hepatobiliary compartments. During fasting, bile acids move down the biliary tract and are concentrated approximately 10-fold in the gallbladder. After an overnight fast, most of the bile acids are sequestered in the gallbladder, resulting in low levels of bile acids in the small intestine, portal vein, systemic circulation, and liver. In response to a meal, cholecystokinin is released from the intestinal mucosa and acts on the biliary tree to relax the sphincter of Oddi and stimulate gallbladder contraction. A concentrated solution of mixed micelles (bile acids, phospholipids, and cholesterol) is emptied from the gallbladder via the bile duct into the small intestine, where the micelles facilitate fat absorption by stimulating the action of pancreatic lipase on triglyceride, solubilizing the hydrolytic products, and delivering lipids to the mucosal surface. During the digestion of

lithocholate to CDCA or UDCA, and sulfation of bile acids by the liver or kidney (see Fig. 64-2). Additional minor pathways include hydroxylation and glucuronidation of bile acids. The conjugation of LCA with sulfate or glucuronide blocks intestinal absorption, and the modified LCA is rapidly lost from the circulating pool of bile acids. In humans, sulfation of LCA plays an important protective role, because unmodified LCA is intrinsically hepatoxic.24 The molecular mechanisms responsible for inducing these hepatoprotective modifications are being elucidated and involve induction of cytosolic sulfotransferases and cytochrome P-450 enzymes by the nuclear receptors pregnane X receptor (PXR) and constitutive androsterone receptor (CAR).25 For example, LCA, as well as other xenobiotic inducers such as rifampin, bind and directly activate PXR, thereby stimulating important mechanisms for hepatic detoxification and transport of bile acids.26

THE ENTEROHEPATIC CIRCULATION The anatomic components of the enterohepatic circulation are the liver, biliary tract, gallbladder, small intestine, portal venous circulation, and, to a lesser extent, colon, systemic Hepatocyte Na+ BA–

Cholangiocyte NTCP

OST α/β

BA–

BSEP

BA–

Na+ BA–

ATP

Renal proximal tubule cell ASBT

OST α/β

BA–

Na+ BA–

ASBT

Ileocyte

OST α/β BA–

Na+ BA– ASBT

Figure 64-3.  Enterohepatic circulation of bile acids showing the individual transport proteins responsible for bile acid (BA) transport across epithelia of various cells, including hepatocytes, cholangiocytes, ileocytes (ileal enterocytes), and renal proximal tubule cells. A−, anion; ASBT (gene symbol SLC10A2), apical sodium bile acid transporter; BSEP (ABCB11), bile salt export pump; NTCP (SLC10A1), Na+-taurocholate cotransporting polypeptide; OST, organic solute transporter. (Adapted with permission from Mosely RH. Bile secretion and cholestasis. In: Kaplowitz N, editor. Liver and Biliary Disease. 2nd ed. Philadelphia, Pa: Williams & Wilkins; 1996. p 194.)

1079

1080

Section VIII  Biliary Tract a large meal, the gallbladder remains contracted, and bile acids secreted by the liver bypass the gallbladder and empty directly into the duodenum. During this period, the bile acid concentration in the small intestine is approximately 5 to 10 mmol/L; micelle formation requires that the intraluminal bile acid concentration be greater than 1.5 mmol/L. During the interdigestive period, the sphincter of Oddi contracts and the gallbladder relaxes, causing a larger fraction of the secreted bile acid to enter the gallbladder for storage. This gallbladder relaxation is mediated in part by the ileal release of fibroblast growth factor 19 (FGF19).27 Therefore, the enterohepatic cycling of bile acids accelerates during digestion and slows between meals and during overnight fasting. This rhythm of bile acid secretion is maintained even after cholecystectomy.28 When the gallbladder is absent, bile acids are stored in the proximal small intestine. After ingestion of a meal, small intestinal contractions propel the stored bile acids to the distal ileum, where they are actively reabsorbed. The enterohepatic cycling of bile acids is extremely efficient; less than 10% of the intestinal bile acids escape reabsorption and are eliminated in the feces. Bile acids are absorbed from the small intestine predominantly by an active transport system restricted to the terminal ileum and, to a lesser extent, by passive absorption down the length of the intestine. Of all the conjugated anions secreted into bile, only bile acids are actively absorbed in conjugated form and undergo an enterohepatic circulation. In adult humans, the enterohepatic circulation maintains a bile acid pool size of 50 to 60 mmol per kg body weight, corresponding to approximately 2 to 4 g. The bile acid pool cycles two to three times per meal, resulting in 6 to 10 cycles per day, and the intestine may reabsorb between 10 and 30 g of bile acid per day. Approximately 0.2 to 0.6 g of bile acids escapes reabsorption and is eliminated in the stool each day. Hepatic conversion of cholesterol to bile acid balances fecal excretion, and this process represents an important route for elimination of cholesterol from the body. The kinetics of bile acid turnover in humans are summarized in Table 64-2. An enterohepatic circulation of bile acids is advantageous because it results in the accumulation of a large mass of detergent molecules that can be used repeatedly during digestion of a single meal or multiple meals throughout the day. The presence of an ileal active transport system and enterohepatic circulation dissociates hepatic bile acid secretion from bile acid synthesis, thereby improving the efficiency of intestinal absorption. Because secretion of bile acids induces hepatic bile flow, maintenance of the enterohepatic circulation also permits continuous secretion of bile. The dissociation of bile acid biosynthesis from intestinal delivery is also promoted by the presence of the gallbladder, because the availability of a concentrative storage reservoir permits bile acids to be delivered in a high concentration and controlled fashion to the duodenum. The ileal bile acid transporter and gallbladder are complementary rather than redundant, and they function together to conserve bile acids. In the presence of a gallbladder but the absence of an active ileal bile acid transporter, the secreted bile acids would be poorly reabsorbed. Emptying of the gallbladder contents would necessarily be followed by a refractory period during which the bile acid supply would not be sufficient to promote lipid digestion and absorption. The refractory period would last until hepatic synthesis could restore the bile acid pool. The existence of an ileal bile acid transporter and an enterohepatic circulation permits the bile acid pool to be used repeatedly during the digestion of a single meal.

HEPATIC BILE ACID TRANSPORT AND BILE SECRETION Bile formation by hepatocytes involves secretion of osmotically active inorganic and organic solutes into the canalicular lumen, followed by passive water movement. Canalicular bile formation has been studied using metabolically inert markers such as mannitol and erythritol and is traditionally divided into two components: bile acid–dependent bile flow (bile flow relating to bile acid secretion) and bile acid– independent flow (bile flow attributed to active secretion of inorganic electrolytes and other solutes). Hepatic adenosine triphosphate (ATP)–dependent carriers actively secrete bile acids into the canalicular lumen, from which they are too large to diffuse back across the paracellular junctions that line the canaliculi. Solutes such as the conjugated bile acids that are actively pumped across the canalicular membrane generate bile flow and are termed primary solutes. The list of primary solutes includes conjugated bilirubin, glutathione, heavy metals, and conjugates of various metabolites and xenobiotics. Water, plasma electrolytes, calcium, glucose, amino acids, bicarbonate, and other low-molecular-weight solutes that flow into the canaliculus in response to the osmotic gradient are termed secondary solutes. The choleretic activity of each primary solute is defined as the amount of bile flow induced per amount of solute secreted. The choleretic activity varies for different bile acid species and ranges from 8 to 30 microliters of bile flow induced per micromole of bile acid secreted. In humans most canalicular bile flow is generated by bile acid secretion; however, the secretion of other primary solutes by the hepatocyte and biliary epithelium also contributes to bile formation. Newly secreted hepatic canalicular bile is modified during its transit in the biliary tract via the action of ductule epithelial cells (cholangiocytes), and this ductular secretion may account for up to 40% of the volume of bile. The ductular modifications to hepatic bile include the absorption of solutes such as glucose, amino acids, and bile acids; the secretion of solutes such as bicarbonate and chloride; and the movement of water through specific channels (aquaporins) and paracellularly.29,30

BILE ACID–INDEPENDENT BILE FLOW

Canalicular bile flow is generated by active secretion of primary solutes in addition to bile acids. Reduced glutathione (GSH) and bicarbonate (HCO3−) constitute the major components of the bile acid–independent fraction of bile flow. The ATP-dependent canalicular secretion of GSH and GSH conjugates via the multidrug resistance–associated protein 2 (MRP2) plays a particularly important role. In addition to being secreted at high concentrations into bile, intraluminal catabolism of GSH by gamma glutamyl transpeptidase (GGTP) further raises the solute concentration and contributes to the osmotic driving force for canalicular bile formation. Besides the ATP-dependent secretion of organic anions into bile, hepatic and biliary ATP-independent secretion of bicarbonate via the HCO3-/Cl- anion exchanger AE2 contributes to the bile acid–independent bile flow.31 The majority of this HCO3- secretion occurs at the level of the bile duct epithelial cells in response to stimulation by a variety of hormones and neuropeptides, such as secretin and vasoactive intestinal peptide.32

CHOLEHEPATIC SHUNT PATHWAY

The term cholehepatic shunt was coined to describe the cycle whereby unconjugated dihydroxy bile acids secreted

Chapter 64  Bile Secretion and the Enterohepatic Circulation into bile are absorbed passively by cholangiocytes, returned to the hepatocyte via the periductular capillary plexus, and resecreted into bile. Absorption of the protonated unconjugated bile acid molecule generates a bicarbonate anion, resulting in a bicarbonate-rich choleresis. Premature absorption and resecretion of the bile acid also increases bile aciddependent bile flow. This cycle explains the hypercholeresis observed for unconjugated C-24 dihydroxy bile acids such as UDCA, unconjugated C-23 bile acid analogs such as norursodeoxycholate (norUDCA), and certain drugs such as the nonsteroidal anti-inflammatory drug sulindac.33 The physiologic significance of the cholehepatic shunt pathway for hepatic bile secretion is unclear, because as originally proposed, the cholehepatic shunt has only a passive absorption component. Unlike exogenously administered UDCA or norUDCA, which may be conjugated incompletely, endogenous bile acids are conjugated efficiently to taurine or glycine before their biliary secretion. As such, the majority of the biliary bile acid pool is ionized and unable to diffuse passively across the biliary epithelium. The finding that the biliary epithelium expresses the apical sodium-dependent bile acid transporter (ASBT; gene symbol SLC10A2) has provided a physiologic mechanism for cholehepatic shunting of conjugated bile acids.34 ASBT activity in cholangiocytes and cholehepatic shunting contribute to the choleresis and increase in biliary lipid secretion and bile flow associated with bile acid feeding.35 Because bile acids ultimately empty from the biliary tract into the small intestine, the quantitative contribution of this pathway to bile production under physiologic or pathophysiologic conditions remains to be determined. Indeed, the major function of the ASBT in biliary epithelium may be to permit cholangiocytes to sample biliary bile acid concentrations to activate cellular signaling pathways rather than to transport significant quantities of bile acids.36

HEPATIC BILE ACID TRANSPORT

Approximately 95% of the bile acids secreted into bile are derived from the recirculating pool. To maintain this process, liver parenchymal cells must transport bile acids efficiently from the portal blood into bile. This vectorial trans-hepatocellular movement of bile acids is a concen­ trative transport process that is driven by a distinct set of primary (ATP-dependent), secondary (Na+ gradient dependent), and tertiary (OH−- or HCO3−-dependent anion exchange) transport systems at the sinusoidal and canalicular plasma membranes.37,38 Bile acid flux through the liver and the number of participating hepatocytes vary. In the fasting state, bile acids are taken up predominantly by the periportal hepatocytes (the hepatocytes closest to the portal venules in the liver acinus), whereas during feeding, more hepatocytes in the liver acinus participate in bile acid uptake. In periportal hepatocytes, bile acid synthesis is repressed, but in pericentral (perivenous) hepatocytes, bile acids are actively synthesized. Therefore, periportal hepatocytes primarily absorb and secrete recirculating bile acids; pericentral cells predominantly secrete newly synthesized bile acids; and the transport of recirculating bile acids through periportal hepatocytes drives the majority of bile flow.39 The concentration of bile acids in the portal blood of healthy humans is 20 to 50 µmol/L. Uptake by the liver is typically expressed as fractional extraction, or first-pass extraction, and represents the percentage of bile acids removed during a single passage through the hepatic acinus. The fractional extraction of bile acids from sinusoidal blood ranges from 50% to 90% and remains constant irrespective

of systemic bile acid concentrations. The hepatic fractional extraction is related to bile acid structure and albumin binding and is largest (80% to 90%) for conjugated hydrophilic bile acids such as conjugated CA and smallest (50% to 60%) for unconjugated hydrophobic protein-bound bile acids such as CDCA. The concentration of total bile acids in the systemic circulation reflects this efficient hepatic extraction, averaging 2 to 5 µmol/L and 5 to 15 µmol/L in the fasting and fed states, respectively.40 Remarkable progress has been made in identifying the transport proteins that function to maintain the enterohepatic circulation of bile acids.37,38 The major bile acid carriers in the human hepatocyte, cholangiocyte, ileal enterocyte, and renal proximal tubule cell are shown in Figures 64-3 and 64-4, and the general properties of these carriers are listed in Table 64-4. Because of their importance for bile secretion, the bile acid transporters are highlighted; however, the hepatocyte sinusoidal and canalicular membranes also express specialized transport proteins for a wide spectrum of endogenous and exogenous compounds besides bile acids.41,42

HEPATIC SINUSOIDAL Na + -DEPENDENT BILE ACID UPTAKE

Hepatocellular uptake of bile acids occurs against an unfavorable electrochemical ion gradient, resulting in a 5- to 10-fold concentration gradient between the portal blood and hepatocyte cytosol. The uptake of conjugated bile acids at the sinusoidal (basolateral) membrane is mediated predominantly (>80%) by a secondary active Na+-dependent transport system. The driving force for the Na+-dependent uptake is generated by the basolateral Na+,K+-ATPase that maintains the prevailing out-to-in Na+ gradient. In contrast to conjugated bile acids, Na+-dependent uptake accounts for less than half of the uptake of unconjugated bile acids such as CA and UDCA. The Na+-taurocholate cotransporting polypeptide (NTCP; gene symbol SLC10A1) is responsible for the hepatocytic Na+-coupled uptake of all the major species of bile acids as well as certain drugs such as rosuvastatin.43,44 The presence of additional sinusoidal Na+-dependent bile acid transporters has been suggested, but support for the existence of such transport systems is equivocal. No inherited defect in NTCP has been described, but polymorphisms that are specific to different ethnic groups and that interfere with bile acid transport in vitro have been reported for the human NTCP gene.45 Unfortunately, clinical data for these subjects were not reported. In addition, although inherited disorders characterized by hypercholanemia have been described, mutations in genes other than NTCP, such as the tight junction proteins ZO-2 and claudin-1, were found to be responsible for these diseases.20,46 These findings leave open the possibility that an isolated NTCP gene defect may be asymptomatic because the liver also expresses Na+-independent bile acid transporters.

HEPATIC SINUSOIDAL Na + -INDEPENDENT BILE ACID UPTAKE

Unconjugated bile acids such as cholate are taken up predominantly by hepatic Na+-independent transport systems. Identification of Na+-independent hepatic transporters was originally confounded by the broad substrate specificity of these transporters, and the multiplicity of substrates could not be reconciled easily with the properties of known hepatic transport systems. The list of substrates that share

1081

1082

Section VIII  Biliary Tract Hepatocytes K+ K+

Na+,K+-ATPase K+ Na+ NTCP Na+

Na+-H+ exchanger

channel NPC1L1 Chol

OC+ MDR1 PL MDR3 MRP2 OA–

BA– OATP OA–, BA–

A–

Cl–

Cl– Cl– channel AE2

H+ BA– BA–

HCO3 – BA– ABCG5/G8 Chol

BSEP

FIC1

BA–

Na+

MRP2 MRP4 OSTα/β Na+ HCO3 – Na+-HCO3– symporter

Cholangiocytes

MRP3 OSTα/β

Cl–

ASBT BA– BA–

CFTR

BA– Cl– Na+

HCO3–

AE2

Figure 64-4.  Hepatocyte and cholangiocyte transporters important for bile acid secretion. At the sinusoidal membrane of hepatocytes, the Na+-taurocholate cotransporting polypeptide (NTCP; gene symbol SLC10A1) mediates the uptake of conjugated bile acids. Sodium-dependent uptake of bile acids through the NTCP is driven by an inwardly directed sodium gradient generated by the Na+,K+-ATPase, and the membrane potential is generated in part by a K+ channel. The Na+-independent bile acid uptake is mediated by the organic anion-transporting polypeptides OATP1B1 (SLCO1B1) and OATP1B3 (SLCO1B3). The sinusoidal membrane also contains a sodium-hydrogen exchanger and a sodium-bicarbonate cotransporter (symporter). At the canalicular membrane, bile acids are secreted via the bile salt export pump (BSEP; ABCB11), and sulfated or glucuronidated bile acids are secreted via the multidrug resistance– associated protein 2 (MRP2; ABCC2). The canalicular membrane also expresses ATP-dependent export pumps that transport phospholipid (multidrug resistance protein 3, MDR3; ABCB4), cholesterol and plant sterols (ABCG5/ABCG8), and drug metabolites (MDR1; ABCB1) into bile. The canalicular membrane expresses ATP-independent transporters, including the Niemann-Pick C1 Like 1 (NPC1L1) protein that imports sterols, chloride channels, and the chloride-bicarbonate anion exchanger isoform 2 (AE2) that secretes biocarbonate. FIC1 (ATP8B1) is a P-type ATPase mutated in progressive familial intrahepatic cholestasis type 1. Within the cholangiocytes of the large bile ducts, conjugated bile acids are absorbed by the apical Na+-dependent bile acid transporter (ASBT; SLC10A2). Bile acid may then exit at the basolateral surface into the hepatic arterial circulation via the heteromeric transporter OSTαOSTβ or an ATP-dependent carrier, MRP3 (ABCC3). Cholangiocytes also express a variety of other carriers important for modifying bile, including the cystic fibrosis transmembrane regulator (CFTR), AE2 for secretion of bicarbonate, the sodium-glucose cotransporter (SGLT1; SLC5A1 [not shown]), and numerous aquaporin isoforms to facilitate water movement. A−, anion; BA−, bile acid anion; Chol, cholesterol; OA−, organic anion; OC+, organic cation; PL, phospholipid. (Adapted with permission from Trauner M, Meier PJ, Boyer JL. Molecular pathogenesis of cholestasis. N Engl J Med 1998; 339:1217-27. Copyright 1998 Massachusetts Medical Society. All rights reserved.)

apparently common transport pathways included conjugated and unconjugated bile acids, sulfobromphthalein (BSP), cardiac glycosides and other neutral steroids, linear and cyclic peptides, and numerous drugs. This conundrum was resolved with the discovery of the organic aniontransporting polypeptide (OATP) gene family and the finding that multiple OATP family members contribute to the hepatic uptake of bile acids, organic anions, and drugs.47 The OATP-type transporters include more than 39 members identified in human, mouse, and rat tissues. The original Human Genome Organization (HUGO) Gene Nomenclature Committee designation for the OATP solute carriers was SLC21A, but confusion related to species differences led to the adoption and widespread acceptance of a new species-independent classification and nomenclature system designated OATP/SLCO.47 In contrast to wellunderstood Na+-cotransporters such as the NTCP, the

driving force for OATP-mediated solute uptake remains poorly understood and may vary among individual OATP family members. Potential transport mechanisms include facili­tative diffusion and electroneutral exchange that couple solute uptake to GSH efflux.48,49 The human OATP family members, OATP1B1 (gene symbol SLCO1B1; original protein name OATP-C) and OATP1B3 (gene symbol SLCO1B3; original protein name OATP8) are expressed primarily in liver, exhibit partially overlapping substrate specificities, and account for the majority of hepatic Na+-independent bile acid clearance. OATP2B1 (gene symbol SLCO2B1; original protein name OATP-B) is expressed in numerous human tissues including liver. OATP2B1 does not transport bile acids but rather functions along with OATP1B1 and OATP1B3 to transport organic anions and solutes such as BSP, bilirubin glu­ curonides, steroid metabolites (estradiol-17β-glucuronide,

Chapter 64  Bile Secretion and the Enterohepatic Circulation Table 64-4  Function of Transport Proteins Involved in Bile Formation and the Enterohepatic Circulation of Bile Acids TRANSPORTER (GENE SYMBOL)

LOCATION

Hepatocyte NTCP (SLC10A1) OATP1B1 (SLCO1B1) OATP1B3 (SLCO1B3) Na+,K+-ATPase BSEP (ABCB11) MDR3 (ABCB4) ABCG5 ABCG8 NPC1L1 FIC1 (ATP8B1) MRP2 (ABCC2)

Bile Acid–Dependent Bile Flow Basolateral membrane Basolateral membrane Basolateral membrane Basolateral membrane Canalicular membrane Canalicular membrane Canalicular membrane Canalicular membrane Canalicular membrane Canalicular membrane Bile Acid–Independent Bile Flow Canalicular membrane

OATP (1B1, 1B3, 2B1)

Basolateral membrane

MRP3 (ABCC3)

Sinusoidal Bile Acid Export Basolateral membrane

MRP4 (ABCC4) OSTα-OSTβ Cholangiocyte Aquaporin 1 (AQP1) Aquaporin 4 (AQP4) AE2 (SLC4A2) CFTR (ABCC7) ASBT (SLC10A2) Ileal Enterocyte ASBT (SLC10A2) NPC1L1 OSTα-OSTβ MRP3 (ABCC3)

FUNCTION Na+-dependent bile acid and xenobiotic uptake Na+-independent bile acid and xenobiotic uptake Na+-independent bile acid and xenobiotic uptake Secretion of 2 Na+ in exchange for 3 K+ ATP-dependent bile acid export ATP-dependent phosphatidylcholine export ATP-dependent sterol export ATP-dependent sterol export Sterol import ATP-dependent aminophospholipid flipping ATP-dependent transport of glucuronide, glutathione, and sulfate conjugates Na+-independent transport of organic anions, cations, and neutral steroids ATP-dependent export of bile acids and glucuronide conjugates ATP-dependent export of glutathione and bile acids Bile acid export

Basolateral membrane Basolateral membrane Ductular Secretion Apical membrane Basolateral membrane Apical membrane Apical membrane Apical membrane

Water transport Water transport HCO3− secretion in exchange for Cl− Cl− secretion Bile acid uptake (cholehepatic shunt)

Apical membrane Apical membrane Basolateral membrane Basolateral membrane

Na+-dependent bile acid uptake Sterol import Bile acid export Bile acid export

ABC, ATP-binding cassette; AE2, chloride-bicarbonate anion exchanger isoform 2; ASBT, apical Na+ bile acid transporter; BSEP, bile salt export pump; CFTR, cystic fibrosis transmembrane regulator; FIC1, P-type ATPase mutated in progressive familial intrahepatic cholestasis type 1; MDR, multidrug resistance protein; MRP, multidrug resistance–associated protein; NPC1L1, Niemann-Pick C1 Like 1; NTCP, Na+-taurocholate cotransporting polypeptide; OATP, organic anion transporting polypeptide; OST, organic solute transporter; SLC, solute carrier.

dehydroepiandrosterone-3-sulfate, and estrone-3-sulfate), arachidonic acid products (prostaglandin E2, thromboxane B2, and leukotriene C4), and a wide variety of drugs (such as rifampin, pravastatin, rosuvastatin, digoxin, and fexofenadine).50 Because the majority of hepatic bile acid uptake is sodium dependent, the major physiologic role of these broad-specificity solute carriers may be hepatic clearance of non–bile acid metabolites and xenobiotics. Indeed, inherited polymorphisms in the OATP genes are now recognized to influence drug metabolism and contribute to some forms of drug toxicity.51

HEPATIC SINUSOIDAL BILE ACID EFFLUX

At the basolateral membrane of hepatocytes, bile acid efflux is an important protective mechanism to reduce bile acid overload under cholestatic conditions.52 Hepatic sinusoidal bile acid export is mediated by the heteromeric organic solute transporter, OSTα-OSTβ, and by members of the MRP family including MRP3 (gene symbol ABCC3) and MRP4 (gene symbol ABCC4).53 This is an important part of the adaptive response to conditions of bile acid overload that also includes down-regulation of the major liver bile acid uptake transporters, NTCP, and members of the OATP family.52

CANALICULAR BILE ACID TRANSPORT

Canalicular secretion is the rate-limiting step in hepatic transport of bile acids from blood into bile. Whereas bile

acid concentrations within the hepatocyte are in the micromolar range, canalicular bile acid concentrations are more than 1000-fold higher, necessitating active transport across the canalicular membrane. Functional evidence of ATPdependent bile acid transport by the canalicular membrane existed since the early 1990s, and functional expression and characterization studies revealed that a novel ABC transporter closely related to the multidrug resistance protein (MDR1)/P-glycoprotein gene is the canalicular bile acid transporter. This 160-kd protein (originally named “sister of P-glycoprotein” [Spgp]) was shown to transport conjugated bile acids efficiently and was subsequently renamed the bile salt export pump (BSEP; gene symbol ABCB11).54 The role of BSEP as the major canalicular bile acid efflux pump was confirmed by the identification of ABCB11 mutations in patients with PFIC type 2, a hepatic disorder characterized by biliary bile acid concentrations less than 1% of normal (see Chapter 76).55,56

INTESTINAL AND RENAL BILE ACID TRANSPORT Bile acids enter the small intestine along with other biliary constituents and facilitate absorption of dietary lipids and fat-soluble vitamins. The intestinal absorption of bile acids is extremely efficient—less than 10% of the intestinal bile acids escape reabsorption and are eliminated in the feces.

1083

1084

Section VIII  Biliary Tract Bile acids are absorbed through a combination of passive absorption in the proximal small intestine and active absorption in the distal ileum. Numerous observations indicate that the terminal ileum is the major site of bile acid reabsorption. For example, there is little drop in intraluminal bile acid concentrations prior to the ileum, and bile acid malabsorption occurs after ileal resection. Studies using in situ perfused intestinal segments to measure bile acid absorption have demonstrated that ileal bile acid transport is a high-capacity system that is sufficient to account for the biliary output of bile acids, thus suggesting that the ileal active transport system is the major route of conjugated bile acid uptake. Only a small fraction of the glycine-conjugated bile acids are protonated in the duodenum, where the intraluminal pH becomes transiently acidic during digestion, and these nonionized bile acids may be absorbed by passive or facilitative diffusion. Bile acids also are deconjugated and dehydroxylated by the bacterial flora in the distal small intestine and colon. These unconjugated hydrophobic bile acids are weak acids and can be absorbed passively if they remain in solution. A fraction (10% to 50%, depending on the bile acid species) of the bile acids returning in the portal circulation escapes hepatic extraction and spills into the systemic circulation. The binding of bile acids to plasma proteins reduces glomerular filtration and minimizes urinary excretion of bile acids. In healthy humans, the kidney filters approximately 100 µmol of bile acids each day. Remarkably, only 1 to 2 µmol is excreted in the urine because of a highly efficient tubular reabsorption. Even in patients with cholestatic liver disease, in whom plasma bile acid concen­ trations are elevated, the 24-hour urinary excretion of nonsulfated bile acids is significantly less than the quantity that undergoes glomerular filtration. Subsequent studies have shown that bile acids in the glomerular filtrate are actively reabsorbed from the renal tubules, and this process contributes to the rise in serum bile acid concentrations in patients with cholestatic liver disease. As in the ileum, the renal proximal tubule epithelium expresses a Na+ gradient-driven transporter that functions as a salvage mechanism to conserve bile acids.

MOLECULAR MECHANISMS OF ILEAL AND RENAL BILE ACID TRANSPORT

Bile acids are transported actively across the ileal brushborder membrane by the well-characterized apical sodium bile acid transporter (ASBT; gene symbol SLC10A2). The relationship between the hepatic, biliary, ileal, and renal Na+–bile acid cotransport systems was resolved with the cloning of the bile acid carriers from those tissues. The liver and ileum express distinct, but related, Na+–bile acid cotransporters, NTCP and ASBT, whereas the ileal enterocyte, renal proximal tubule cell, and cholangiocyte all express the same Na+–bile acid cotransporter (ASBT).38 The inwardly directed Na+ gradient maintained by the basolateral Na+,K+-ATPase as well as the negative intracellular potential provide the driving force for ASBT-mediated bile acid uptake. The ASBT transports all the major species of bile acids but does not appear to transport any non–bile acid–related metabolites or drugs. The finding that ASBT gene mutations are responsible for PBAM, a disorder associated with intestinal bile acid malabsorption and steatorrhea, demonstrated that most intestinal bile acid absorption in humans is mediated by the ASBT.57 Little is known about the intracellular transport of bile acids in the ileal enterocyte. The ileal lipid-binding protein (ILBP, also called the ileal bile acid binding protein [IBABP];

gene symbol FABP6) is a member of the fatty acid binding protein family and a major ileal enterocyte cytosolic protein. ILBP is believed to be involved in the transcellular transport of bile acids; however, a mouse model lacking appreciable ILBP expression did not exhibit impaired intestinal bile acid absorption, indicating that ILBP is not essential for this process.58 The proteins responsible for bile acid export across the basolateral membrane of the ileal enterocyte, cholangiocytes, and renal proximal tubule cell have been identified. The heteromeric organic solute transporter, OSTα-OSTβ, is expressed on the basolateral membrane of ileal enterocytes, hepatocytes, cholangiocytes, and renal proximal tubule cells and transports bile acids as well as a variety of organic anions.59 Although no inherited defects have been reported for the OSTα or OSTβ genes in humans, targeted inactivation of the OSTα gene in mice resulted in impaired intestinal bile acid absorption and altered bile acid metabolism, indicating that OSTα-OSTβ is a major basolateral bile acid transporter.60,61

DISORDERS OF THE ENTEROHEPATIC CIRCULATION Cholestasis is defined as interruption of the normal process of bile formation and is classically subdivided into intrahepatic cholestasis, a functional defect in bile formation at the level of the hepatocyte, and extrahepatic cholestasis, an obstruction to bile flow within the biliary tract. Impaired hepatic transport of bile acids and other organic solutes is a prominent feature of both inherited and acquired forms of cholestatic liver disease. Disorders of the enterohepatic circulation are generally classified into the following four categories: (1) defects in bile acid formation (synthesis and conjugation), (2) defects in membrane transport of bile acids (uptake and secretion), (3) disturbances involving bacterial transformation (deconjugation and dehydroxylation), and (4) disturbances in movement through or between organs (bile acid circulation). The measurement of serum bile acid levels has only a limited diagnostic role for disorders of the enterohepatic circulation. Specific diagnosis of the rare inherited bile acid synthesis defects is made by analysis of body fluids (bile, blood, and urine) using fast atom bombardment–mass spectroscopy (FAB-MS) and gas chromatography–mass spectroscopy (GC-MS).11,62 These disorders present with markedly reduced or complete absence of CA and CDCA levels and greatly elevated concentrations of atypical bile acids in bile, serum, and urine. Measurement of serum bile acid concentrations, however, offers little, if any, benefit over conventional liver biochemical tests in the diagnosis or management of most forms of liver disease or bile acid malabsorption.

BILE ACID SYNTHESIS

Continuous bile acid synthesis from cholesterol is required to maintain the bile acid pool in the enterohepatic circulation. The maximal rate of bile acid synthesis is on the order of 4 to 6 g/day. Although inherited defects in biosynthesis are extremely rare, these disorders serve to illustrate the importance of bile acid synthesis for normal hepatic function. The effects of a cessation in bile acid synthesis include depletion of the bile acid pool by fecal excretion, loss of bile acid–dependent bile flow, decreased biliary excretion of cholesterol and xenobiotics, reduced intestinal absorption of cholesterol and fat-soluble vitamins, and accumulation of cytotoxic bile acid biosynthetic intermediates. Inherited defects in eight of the bile acid biosynthetic enzymes have

Chapter 64  Bile Secretion and the Enterohepatic Circulation been reported; because the side chain modification steps in bile acid biosynthesis occur in the peroxisome, disorders that disrupt peroxisome biogenesis such as Zellwegner syndrome also affect bile acid synthesis (see Chapter 76). The list of inborn errors includes cholesterol 7α-hydroxylase (CYP7A1), sterol 27-hydroxylase (CYP27A1), oxysterol 7α-hydroxylase (CYP7B1), 3β-hydroxy-Δ5-C27-steroid oxidoreductase (HSD3B7), Δ4-3-oxosteroid 5β-reductase (AKR1D1), 2-methylacyl-coenzyme A racemase (AMACR), D-bifunctional protein (HSD17B4), and the bile acid coenzyme A:amino acid N-acyl-transferase (BAAT).11,62 In general, genetic defects that affect early steps in the bile acid biosynthetic pathway cause disease in newborns, whereas the consequences of defects in later steps are varied. In some cases, a single enzyme defect is not sufficient to eliminate production of all bile acids, because multiple biosynthetic pathways exist.9 For example, cerebrotendinous xanthomatosis (CTX) is a rare, inherited disease caused by mutations in the mitochondrial enzyme sterol 27-hydroxylase (CYP27A1). In CTX, the alternative pathway of bile acid synthesis is blocked, and bile acid synthesis is diminished but not eliminated. Liver disease does not occur in patients with CTX, which is characterized by progressive neurologic disturbances, premature atherosclerosis, cataracts, and tendinous xanthomas. Bile acid therapy with hydrophobic bile acids such as CDCA has been reported to suppress the biochemical abnormalities in CTX and to slow progression of the disease. The most commonly reported defect in bile acid synthesis is 3β-hydroxy-C27-steroid oxidoreductase (HSD3B7) deficiency, and this enzyme defect affects both the classical and

alternative pathways for bile acid biosynthesis. The disease is characterized by progressive intrahepatic cholestasis and accumulation of abnormal bile acids. These unusual conjugates of dihydroxy- and trihydroxy-5-cholenoic acid are poorly transported by the BSEP and interfere with the canalicular secretion of other bile acids. Clinical manifestations include unconjugated bilirubinemia, jaundice, serum aminotransferase elevations, steatorrhea, pruritus, and poor growth. Disease progression varies but ultimately results in cirrhosis and hepatic failure in a high proportion of affected patients. Treatment with exogenous primary bile acids such as UDCA reverses the biochemical abnormalities and may be life-saving.11

MEMBRANE TRANSPORT OF BILE ACIDS AND BILIARY LIPIDS

A growing number of disorders have been found to be associated with mutations in the bile acid transporter or biliary organic solute transporter genes.1 The list of these disorders includes PFIC types 1 to 3, benign recurrent intrahepatic cholestasis (BRIC) types 1 and 2, LPAC, ICP, Dubin-Johnson syndrome, sitosterolemia, and PBAM and is summarized in Table 64-5. PFIC type 1 (PFIC1; formerly called Byler’s disease) manifests primarily as chronic intrahepatic cholestasis and coarse granular bile in patients with normal serum GGTP levels. The gene defect in patients with PFIC1 had been mapped to chromosome 18 in the same region where a similar but milder disease phenotype, BRIC, has been localized. A combined search for the two disease loci identified

Table 64-5  Disorders of the Enterohepatic Circulation Associated with Inherited Transporter Defects disorder

TRANSPORTER DEFECT

DEFECTIVE protein (gene)

Progressive Familial Intrahepatic Cholestasis (PFIC) Type 1 Aminophospholipids

FIC1 (ATP8B1)

Type 2

Bile acids

BSEP (ABCB11)

Type 3

Phosphatidylcholine

MDR3 (ABCB4)

Benign Recurrent Intrahepatic Cholestasis (BRIC) Type 1 Aminophospholipids

FIC1 (ATP8B1)

Type 2

Bile acids

BSEP (ABCB11)

LPAC

Phosphatidylcholine

MDR3 (ABCB4)

Intrahepatic Cholestasis of Pregnancy (ICP) ICP Phosphatidylcholine

MDR3 (ABCB4)

ICP

Bile acids

BSEP (ABCB11)

DJS Sitosterolemia PBAM

Organic anion conjugates Cholesterol, phytosterols Bile acids

MRP2 (ABCC2) ABCG5, ABCG8 ASBT (SLC10A2)

CHARACTERISTIC FEATURES Progressive cholestasis, elevated serum bile acids, pruritus, normal serum gamma glutamyl transpeptidase (GGTP), pancreatitis, intestinal malabsorption, hearing loss Progressive cholestasis, jaundice, giant cell formation, lobular and portal fibrosis, hepatobiliary malignancy, normal serum GGTP Cholestasis, extensive bile duct proliferation and periportal fibrosis, elevated serum GGTP Periodic attacks of cholestasis, elevated serum bile acids, pruritus, normal serum GGTP, malabsorption, hearing loss Periodic attacks of cholestasis, cholelithiasis, normal serum GGTP, hepatosplenomegaly Cholelithiasis, intrahepatic hyperechoic foci, biliary cirrhosis, elevated serum GGTP Cholestasis in third trimester of pregnancy, fetal loss and prematurity, elevated serum GGTP Cholestasis in third trimester of pregnancy, fetal loss and prematurity, normal serum GGTP Jaundice, benign conjugated hyperbilirubinemia Xanthomas, hypersterolemia, coronary artery disease Chronic diarrhea, steatorrhea, fat-soluble vitamin malabsorption

ABC, ATP-binding cassette; ASBT, apical Na+–bile acid transporter; BSEP, bile salt export pump; DJS, Dubin–Johnson syndrome; FIC1, P-type ATPase mutated in progressive familial intrahepatic cholestasis type 1; ICP, intrahepatic cholestasis of pregnancy; LPAC, low-phospholipid-associated cholelithiasis; MDR, multidrug resistance protein; MRP, multidrug resistance–associated protein; PBAM, primary bile acid malabsorption.

1085

1086

Section VIII  Biliary Tract a P-type ATPase, designated FIC1 (gene symbol ATP8B1), as the defective gene for both PFIC1 and BRIC type 1. P-type ATPases are distinct from ABC transporters and constitute a large family that includes ion pumps such as the Na+,K+ATPase, Ca2+-ATPase, and the copper-transporting Wilson disease gene product (see Chapter 75). An analysis of the spectrum of ATP8B1 mutations in patients with PFIC1 and BRIC revealed that mutation type and location in the gene generally correlate with clinical severity. More innocuous ATP8B1 missense mutations are common in BRIC type 1 (BRIC1), whereas nonsense, frame-shifting, and largedeletion mutations are more common in PFIC1.63 FIC1 functions as an aminophospholipid (phosphatidylserine) flippase, and activity requires the coexpression of CDC50 proteins that promote the trafficking of FIC1 to the plasma membrane.64 In humans, the FIC1 protein is expressed in many tissues, including the pancreas, small intestine, urinary bladder, stomach, and prostate. The expression and activity of FIC1 in many extrahepatic tissues may help explain the high frequency of diarrhea, pancreatitis, respiratory symptoms, and hearing loss in patients with PFIC1 disease. Whereas the mechanism responsible for cholestasis associated with FIC1 deficiency remains to be fully elucidated, results from the analysis of an ATP8B1 mutant mouse indicate that loss of aminophospholipid flippase activity makes the canalicular membrane more susceptible to damage from hydrophobic bile acids.65 In addition, defective bile acid signaling pathways may contribute to the pathophy­siology of PFIC1.66 Treatment options for PFIC1 and BRIC1 include supplementation with fat-soluble vitamins, partial external biliary diversion, and, for patients with PFIC1, liver transplantation.67 PFIC2 (also called BSEP disease) is associated with low bile acid secretion, progressive cholestasis, normal serum GGTP levels, lobular and portal fibrosis, hepatic giant cell transformation, and the lack of bile duct proliferation on examination of liver histology. The disease was mapped to chromosome 2q24, and the defective gene has been shown to encode the canalicular BSEP protein (ABCB11). BSEP protein could not be detected on the canalicular membrane in liver biopsy specimens from patients with PFIC2, therefore suggesting that mutations in ABCB11 impair the synthesis, cellular trafficking, or stability of BSEP protein.56 In addition to a progressive intrahepatic cholestatic phenotype, ABCB11 missense mutations were found in patients with a milder disease similar to BRIC.68 Unlike the form of BRIC that is associated with ATP8B1 mutations, patients with BRIC caused by ABCB11 mutations lack extrahepatic symptoms such as pancreatitis and are more likely to form gallstones. On the basis of these findings and consistent with the genetic classification of PFIC into subtypes, BRIC is classified as types 1 and 2, corresponding to ATP8B1- and ABCB11-mutation asso­ ciated forms, respectively. The ABCB11 mutations associated with milder disease may also confer an increased risk for the development of ICP and drug-induced cholestasis.69 Patients with PFIC2 present as infants with high serum bile acid levels, intractable pruritus, intestinal malabsorption, failure to thrive, and cholestasis, ultimately developing fibrosis and end-stage liver disease before adulthood.67 In addition to cholestasis, outcome data have confirmed that severe BSEP deficiency significantly increases the risk for hepatobiliary malignancy, including hepatocellular carcinoma and cholangiocarcinoma.56,70 The treatment of PFIC2 is similar to that for PFIC1, with liver transplantation as the ultimate therapy for patients with severe forms of the disease.

PFIC3 is quite different from the other PFIC subtypes. Serum GGTP levels are markedly elevated in these patients, and liver histologic examination shows extensive bile duct proliferation and portal and periportal fibrosis. The most common clinical presentations include jaundice, pale stools, pruritus, and hepatomegaly. The defect in PFIC3 lies in MDR3 (gene symbol ABCB4), a canalicular phosphatidylcholine (PC) transporter that belongs to the ABC transporter superfamily.71 In PFIC3, hepatic bile acid secretion is unimpaired, but PC transport is greatly diminished. In bile, PC normally forms mixed micelles with bile acids and acts to buffer the cytotoxic detergent properties of the bile acids. In the absence of biliary phospholipid, the bile acids are highly toxic and cause cholestatic liver damage. In addition to PFIC3, mutations in MDR3 have been associated with LPAC and ICP.1 PFIC3 has been treated with UDCA, with variable results, and with liver transplantation.67

BILE ACID BIOTRANSFORMATION (DECONJUGATION AND DEHYDROXYLATION)

Bile acid deconjugation normally begins in the distal small intestine and is mediated by spilling of colonic bacteria across the ileocecal valve. In patients with small intestinal stasis and bacterial overgrowth, deconjugation of bile acids also occurs in the proximal intestine. The unconjugated bile acids are absorbed passively, and extensive bacterial deconjugation can reduce the intraluminal concentration of bile acids and impair formation of micelles in the small intestine. Increased bacterial deconjugation can be detected indirectly by measuring levels of unconjugated bile acids in the systemic venous plasma. This test is rarely used, however, because bacterial overgrowth is detected more easily with a hydrogen breath test after ingestion of a glucose-containing meal (see Chapter 102). In the colon bile acids are first deconjugated and then 7α-dehydroxylated.22 In healthy humans, colonic bile acids eventually undergo nearly complete 7α-dehydroxylation to yield the secondary bile acids DCA and LCA.21 The conversion of CA to DCA and subsequent colonic absorption of DCA is increased in subjects with an elevated proportion of DCA in their bile acid pool. The underlying mechanism is related to prolonged colonic transit time that results in increases in the concentration of bacteria, 7α-dehydroxylase activity, and colonic pH and DCA solubility.72 Because accumulation of DCA in bile can contribute indirectly to cholesterol cholelithiasis, strategies to accelerate colonic transit or acidify the colonic luminal contents may be useful in preventing cholesterol gallstones in these subjects.

BILE ACID CIRCULATION Biliary Obstruction and Biliary Fistula

Biliary obstruction caused, for example, by a stone that obstructs the bile duct leads to retention of hepatic bile acids and, ultimately, hepatocyte necrosis or apoptosis. A portion of the retained bile acids is modified by sulfation, and both sulfated and unsulfated bile acids are regurgitated from hepatocytes back into the systemic circulation. Despite increased urinary excretion of bile acids, plasma concentrations of bile acids rise as much as 20-fold. When biliary obstruction is incomplete, bile acids continue to be secreted into the small intestine, efficiently absorbed by the ileum, and returned to the liver via the portal circulation. In this clinical condition, administration of bile acid sequestrants decreases the intestinal absorption of cytotoxic bile acids and may slow the progression of liver damage.73 With complete bile duct obstruction, bile acids are not secreted

Chapter 64  Bile Secretion and the Enterohepatic Circulation into the small intestine, and intestinal malabsorption of fat-soluble vitamins and steatorrhea results. Secondary bile acids are not formed, and fecal bile acid output diminishes. In the patient with a biliary fistula, bile acids are diverted from entering the small intestine. Because bile acid biosynthesis is controlled by negative feedback, bile acid synthesis rises markedly, up to 20-fold. Hepatic function is not impaired, although the flux of bile acids through the liver is decreased substantially because maximal bile acid synthesis (4 to 6 g/day) is less than the normal flux in the presence of an intact enterohepatic circulation (12 to 18 g/day). As in biliary obstruction, lower bile acid concentrations in the small intestine result in malabsorption of fat-soluble vitamins. Absorption of dietary fats, especially dietary triglycerides that contain longer-chain fatty acids, is also decreased.

Cholecystectomy

Despite removal of a major storage pool of bile acids, the overall effect of cholecystectomy on biliary secretion is small, and daily bile acid secretion is not altered substantially.28 In the absence of a gallbladder, the bile acid pool is stored in the small intestine during the fasting state. After ingestion of a meal, the bile acid pool moves to the terminal ileum, where it is actively absorbed and returned to the liver via the portal circulation. A change in the composition of the bile acid pool has been reported and is characterized by increased dehydroxylation of CA to DCA. In the small subset of patients with postcholecystectomy diarrhea, the movement of the bile acid pool to the small intestine may overwhelm the ileal transport system, thereby leading to bile acid malabsorption. The diarrhea is usually only transitory, and affected patients generally respond to administration of a bile acid sequestrant.

Ileal Resection

Resection of the terminal ileum causes intestinal bile acid malabsorption. If the resection is short (<100 cm), the effect on bile acid metabolism is minimal because greater biosynthesis balances increased fecal loss. With longer resections, hepatic bile acid synthesis rises more dramatically to compensate for the increased loss. The unabsorbed bile acids enter the colon in greater amounts and act to inhibit water absorption or induce secretion, thereby resulting in mild, watery diarrhea. Symptomatic response is obtained with administration of a bile acid sequestrant. When more than 100 cm of ileum is resected, including the ileocecal valve, bile acid secretion diminishes because the bile acid biosynthetic potential is well below the normal hepatic secretion rate. The bile acid pool becomes progressively depleted during the day, and intestinal fat malabsorption appears because of the lack of micelles and loss of intestinal mucosal surface. The increased fatty acid flux through the colon inhibits water absorption and results in severe diarrhea that responds poorly to bile acid sequestrants. If the diarrhea is of sufficiently large volume and accompanied by malabsorption of other nutrients, the patient may be diagnosed as having short bowel syndrome. Therapy is complex and has only limited success. In some patients, fecal weight and frequency are reduced by elimination of fat from the diet. Other therapies include bile acid replacement and administration of glutamine and growth factors (see Chapter 103).

Bile Acid Malabsorption and Diarrhea

The enterohepatic circulation conserves bile acids efficiently, thereby maintaining bile flow and adequate intralu-

minal bile acid concentrations for micellar solubilization and absorption of lipids. Impaired intestinal absorption of bile acids may contribute to the pathogenesis of a number of gastrointestinal disorders, including idiopathic chronic diarrhea, chronic ileitis, gallstone disease, postcholecystectomy diarrhea, Crohn’s disease, and irritable bowel syndrome. Symptomatic bile acid malabsorption results from failure of the active intestinal transport component and is most commonly caused by ileal resection, ileal disease such as Crohn’s disease, ileal bypass, and radiation enteritis. Bile acid malabsorption is associated also with conditions such as peptic ulcer surgery, chronic pancreatitis, celiac disease, diabetes mellitus, cystic fibrosis, and the use nonsteroidal anti-inflammatory drugs (which may cause intestinal ulcers and strictures), colchicine, olsalazine, and antineoplastic agents. Primary and idiopathic forms of bile acid malabsorption are rare disorders. Mutations in the ASBT gene cause congenital or primary bile acid malabsorption but are not responsible for a more common condition termed idiopathic bile acid malabsorption that describes patients with adultonset intestinal bile acid malabsorption, chronic diarrhea, and a morphologically normal ileum.57 Bile acid malabsorption permits higher concentrations of dihydroxy bile acids to reach the colon, where they alter water and electrolyte movement and lead to diarrhea. Water transport in the colon is critical for the regulation of intestinal fluid and electrolyte balance and is the ultimate determinant of diarrhea (see Chapter 15). Bile acids play a role in colonic water transport apparently by increasing chloride secretion in the perfused colon by an inositol 1,4,5-triphosphate and calcium–dependent mechanism. Trihydroxy bile acids have no effect, but dihydroxy bile acids induce net fluid secretion at high concentrations and block absorption of fluid and water at low concentrations. To induce net secretion, bile acids must (1) have an appropriate structure (hydrophobic dihydroxy bile acids, DCA or CDCA), (2) be present in high concentrations (>1.5 mM) in the aqueous phase, and (3) exist in an environment with the appropriate alkaline pH (fecal pH > 6.8). An additional consequence of bile acid malabsorption and steatorrhea is increased renal oxalate excretion. Under conditions of intestinal bile acid malabsorption, the concentration of long-chain fatty acids is elevated in the colon, where the fatty acids form insoluble calcium soaps. Consequently, less free calcium is available to precipitate dietary oxalate. The higher soluble oxalate concentrations in the lumen and increased colonic permeability promote hyper­ absorption of oxalate. The enteric hyperoxaluria contributes to the formation of kidney stones in affected patients, such as those with Crohn’s disease or short bowel syndrome. Therapeutic approaches to the fat malabsorption and diarrhea associated with bile acid malabsorption are discussed in Chapters 15 and 101.

BILE ACID THERAPY, SEQUESTRANTS, AND TRANSPORT INHIBITORS BILE ACID THERAPY

Bile acid therapy traditionally has been divided into two types: displacement and replacement. The goal of displacement therapy is to alter the composition of the bile acid pool to reduce either the cytotoxicity of endogenous bile acids or biliary cholesterol secretion, whereas the aim of replacement therapy is to correct a bile acid deficiency. More recently, the recognition that bile acids signal through nuclear and G-protein–coupled receptors to induce hepato-

1087

1088

Section VIII  Biliary Tract protective mechanisms and modulate fat, glucose, and energy homeostasis has renewed interest in therapeutic uses for bile acids or bile acid derivatives.2 With the exception of guggul lipid, however, an FXR modulator used in traditional Ayurvedic medicine and sold in the West as a food supplement, these agents are still in preclinical or clinical development. Because of its safety and lack of hepatotoxicity, UDCA is the most widely used form of bile acid therapy.23 After oral administration, UDCA accumulates and ultimately constitutes up to 40% of the circulating bile acid pool, displacing the more hydrophobic endogenous bile acids.23 UDCA was originally administered and approved by the U.S. Food and Drug Administration (FDA) for gallstone dissolution but is not commonly used today for that purpose because of the success of laparoscopic cholecystectomy. The FDA has also approved bile acid therapy for treatment of primary biliary cirrhosis, in which UDCA may delay the progression of liver fibrosis and improve survival (see Chapter 89). UDCA therapy may also have favorable effects in other conditions, such as ICP, viral hepatitis, and parenteral nutrition–associated cholestasis.23 A chemical analog of UDCA, norUDCA, has one less carbon atom in the side chain, is poorly amidated, and undergoes significant cholehepatic shunting. Therapy with norUDCA is effective in treating peribiliary fibrosis in the MDR2 knockout mouse, an animal model of primary sclerosis cholangitis, and is in preclinical development as a potentially new agent for cholestatic liver disease.74 Bile acid replacement therapy is used to treat rare inborn errors of bile acid biosynthesis. In affected patients, administration of a mixture of UDCA and CA suppresses the synthesis of cytotoxic bile acid precursors and restores the input of primary bile acids into the enterohepatic circulation.11 Bile acid replacement therapy is also indicated in patients with severe bile acid malabsorption or short bowel syndrome, in which a deficiency of bile acids in the proximal small intestine leads to impairment of micellar solubilization and intestinal fat malabsorption.75

BILE ACID SEQUESTRANTS

Bile acid sequestrants are positively charged polymeric resins that bind bile acids in the intestinal lumen to decrease the aqueous concentration and the efficiency of intestinal conservation of bile acids. In patients with mild bile acid malabsorption, bile acid sequestrants reduce diarrhea by lowering the concentration of free bile acids in the colon. Bile acid sequestrants also are used to decrease pruritus in patients with cholestasis, presumably by reducing the concentration of bile acids (or other anionic biliary constituents) in the systemic circulation.67 Efficacy is only moderate because of the weak bile acid–binding efficiency of the resins and poor patient adherence to therapy. In addition to the older preparations, cholestyramine and colestipol, more potent sequestrants with superior bile acid-binding properties have been developed. Colesevelam is a newer bile acid sequestrant and has been approved by the FDA for the treatment of hypercholesterolemia and for the improve-

ment of glycemic control in adults with type 2 diabetes mellitus.

BILE ACID TRANSPORT INHIBITORS

An alternative to luminal sequestration of bile acids with binding resins is direct inhibition of the ileal ASBT. Several potent ASBT inhibitors have been developed and evaluated in animal models. Although these agents are being targeted primarily for the treatment of hypercholesterolemia, they may also prove useful for blocking the inappropriate ileal conservation of bile acids that contributes to hepatocellular damage and pruritus in cholestatic liver disease.73

KEY REFERENCES

Ballatori N, Christian WV, Lee JY, et al. OSTalpha-OSTbeta: A major basolateral bile acid and steroid transporter in human intestinal, renal, and biliary epithelia. Hepatology 2005; 42:1270-9. (Ref 59.) Fickert P, Wagner M, Marschall HU, et al. 24-norUrsodeoxycholic acid is superior to ursodeoxycholic acid in the treatment of sclerosing cholangitis in Mdr2 (Abcb4) knockout mice. Gastroenterology 2006; 130:465-81. (Ref 74.) Hagenbuch B, Meier PJ. Organic anion transporting polypeptides of the OATP/SLC21 family: Phylogenetic classification as OATP/ SLCO superfamily, new nomenclature and molecular/functional properties. Pflugers Arch 2004; 447:653-65. (Ref 47.) Hamilton JP, Xie G, Raufman JP, et al. Human cecal bile acids: Concentration and spectrum. Am J Physiol Gastrointest Liver Physiol 2007; 293:G256-63. (Ref 21.) Hofmann AF, Hagey LR. Bile acids: Chemistry, pathochemistry, biology, pathobiology, and therapeutics. Cell Mol Life Sci 2008; 65:2461-83. (Ref 3.) Inagaki T, Choi M, Moschetta A, et al. Fibroblast growth factor 15 functions as an enterohepatic signal to regulate bile acid homeostasis. Cell Metab 2005; 2:217-25. (Ref 14.) Inagaki T, Moschetta A, Lee YK, et al. Regulation of antibacterial defense in the small intestine by the nuclear bile acid receptor. Proc Natl Acad Sci U S A 2006; 103:3920-5. (Ref 7.) Klomp LW, Vargas JC, van Mil SW, et al. Characterization of mutations in ATP8B1 associated with hereditary cholestasis. Hepatology 2004; 40:27-38. (Ref 63.) Knisely AS, Strautnieks SS, Meier Y, et al. Hepatocellular carcinoma in ten children under five years of age with bile salt export pump deficiency. Hepatology 2006; 44:478-86. (Ref 70.) Lundasen T, Galman C, Angelin B, Rudling M. Circulating intestinal fibroblast growth factor 19 has a pronounced diurnal variation and modulates hepatic bile acid synthesis in man. J Intern Med 2006; 260:530-6. (Ref 15.) Marschall HU, Wagner M, Zollner G, et al. Complementary stimulation of hepatobiliary transport and detoxification systems by rifampicin and ursodeoxycholic acid in humans. Gastroenterology 2005; 129: 476-85. (Ref 26.) Oude Elferink RP, Paulusma CC, Groen AK. Hepatocanalicular transport defects: Pathophysiologic mechanisms of rare diseases. Gastroenterology 2006; 130:908-25. (Ref 1.) Paulusma CC, Groen A, Kunne C, et al. Atp8b1 deficiency in mice reduces resistance of the canalicular membrane to hydrophobic bile salts and impairs bile salt transport. Hepatology 2006; 44:195-204. (Ref 65.) Strautnieks SS, Byrne JA, Pawlikowska L, et al. Severe bile salt export pump deficiency: 82 different ABCB11 mutations in 109 families. Gastroenterology 2008; 134:1203-14. (Ref 56.) Thomas C, Pellicciari R, Pruzanski M, et al. Targeting bile-acid signalling for metabolic diseases. Nat Rev Drug Discov 2008; 7:678-93. (Ref 2.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

65 Gallstone Disease David Q.-H. Wang and Nezam H. Afdhal

CHAPTER OUTLINE Definition and Types of Gallstones  1089 Epidemiology  1089 Risk Factors  1090 Composition and Abnormalities of Bile  1093 Physical Chemistry of Bile  1093 Hepatic Secretion of Biliary Lipids  1095 Pathophysiology  1097 Hepatic Hypersecretion of Biliary Cholesterol  1097 Rapid Cholesterol Nucleation and Crystallization  1098 Imbalance of Pronucleating and Antinucleating Factors  1098 Gallbladder Dysfunction  1099 Intestinal Factors  1100 Growth of Gallstones  1101 Genetics  1101 Pigment Stones  1102 Black Pigment Stones  1104 Brown Pigment Stones  1104

DEFINITION AND TYPES OF GALLSTONES Cholesterol cholelithiasis is one of the most prevalent and most costly digestive diseases in western countries. At least 20 million Americans (12% of adults) have gallstones.1-4 The prevalence of gallstones appears to be rising, and each year approximately one million new cases are discovered.5-7 Although many gallstones are “silent,” approximately one third eventually cause symptoms and complications.8 An estimated 700,000 cholecystectomies are performed for gallstone disease each year, and medical expenses for the treatment of gallstones exceeded $6 billion in the year 2000.2 In addition, complications of gallstones result in 3000 deaths (0.12% of all deaths) every year.1 On the basis of chemical composition and macroscopic appearance, gallstones are divided into three types: cholesterol, pigment, and rare stones.3,4,9 The majority (∼75%) of gallstones in the United States and Europe are cholesterol stones,8 which consist of cholesterol monohydrate crystals and precipitates of amorphous calcium bilirubinate, often with calcium carbonate or phosphate in one of the crystalline polymorphs. These stones are usually subclassified as either pure cholesterol or mixed stones that contain at least 50% cholesterol by weight. The remaining gallstones are pigment stones that contain mostly calcium bilirubinate, which are subclassifed into two groups: black pigment stones (∼20%) and brown pigment stones (∼4.5%). Rare Drs. Jeffrey D. Browning and Jayaprakash Sreenarasimhaiah contributed to this chapter in previous editions of this book.

Natural History  1104 Asymptomatic Stones  1104 Symptomatic Stones  1105 Stones in Patients with Diabetes Mellitus  1105 Diagnosis and Clinical Disorders  1105 Imaging Studies  1106 Biliary Pain and Chronic Cholecystitis  1112 Acute Cholecystitis  1113 Choledocholithiasis  1116 Cholangitis  1117 Uncommon Complications  1118 Emphysematous Cholecystitis  1118 Cholecystoenteric Fistula  1118 Mirizzi’s Syndrome  1119 Porcelain Gallbladder  1119

stones (∼0.5%) include calcium carbonate stones and fatty acid–calcium stones. Gallstones also are classified by their location into intrahepatic stones, gallbladder stones, and choledocholithiasis (bile duct stones). Intrahepatic stones are predominantly brown pigment stones. Gallbladder gallstones are mainly cholesterol stones, with a small group of black pigment stones. Bile duct stones are composed mostly of mixed cholesterol stones.

EPIDEMIOLOGY Investigations of gallstone prevalence are more common than those of gallstone incidence because of the nature of the statistical analyses. Prevalence is often defined as the number of cases of gallstones at any one point or period of time divided by the population at risk of forming stones. Incidence is usually defined as the number of new cases of gallstones occurring in a time period divided by the population at risk of forming stones. Therefore, the discovery of incidence requires that investigation for gallstones be performed at a minimum of two different times, that is, at the beginning and at the end of an interval of time. By contrast, prevalence rates can be obtained by sampling at only one point in time, for example, at ultrasonographic screening or necropsy. Although determining the true incidence of gallstones in a given population is not easy, a large study of the incidence of gallstones in the Danish population has been reported.10 The five-year incidence of gallstones was 0.3%, 2.9%, 2.5%,

1089

Section VIII  Biliary Tract 65.6

Women Men

40 Prevalence (%)

1090

30

20

10

0

Figure 65-1.  Prevalence rates of cholesterol gallstones by gender in 18 countries based on representative ultrasonographic surveys.

l nd an ia na sh m ia ia an rk zi ck ru es ly ay na ny cs ile ns la ap unis Chi de ndo Ind uss Ir ma Bra Bla Pe hit Ita rw nti ma ani Ch ica i a J T n la i R W er No rge Ger isp S. Th De ng d K S. A U. Am .H U. Ba nite e S tiv U U. Na

and 3.3% for Danish men, and 1.4%, 3.6%, 3.1% and 3.7% for Danish women ages 30, 40, 50, and 60, respectively. Women have a higher incidence than men at age 30 and 40, but the difference declines with increasing age. These incidence rates could reflect genetic and environmental factors in the specific populations studied because they are in accordance with estimated prevalence rates reported for Denmark and other populations.11 In a major Italian study, the incidence of gallstones was obtained at ten years’ follow-up in an originally gallstone-free cohort in the town of Sirmione.12 This study revealed that new cases of gallstones developed at a rate of 0.5% per year. Although age, female gender, parity, obesity, and hypertriglyceridemia were associated with gallstones in the cross-sectional prevalence study of Sirmione, multivariate analyses of risk factors for the formation of gallstones in the longitudinal study iden­tified only age and obesity to be risk factors. The differences in the incidence of gallstone formation among different populations are striking, suggesting that genetic factors play a key role in the pathogenesis of gallstones. Pathogenic factors are likely to be multifactorial and to vary among populations. Most relevant studies have found that the prevalence of gallstones in women ranges from 5% to 20% between the ages of 20 and 55 and from 25% to 30% after the age of 50. The prevalence in men is approximately one half that of women of the same age. Ultrasonographic screening or necropsy data are often used to estimate the prevalence of gallstone disease in different populations, as illustrated in Figure 65-1. Although ultrasonographic screening cannot be used to distinguish cholesterol from pigment stones, 70% to 80% of detected gallbladder gallstones are assumed to be cholesterol stones. In older studies of American Pima Indians, the prevalence of gallstones was investigated by oral cholecystography.13 The well-studied Pima Indians in southern Arizona display a high prevalence of gallstones, which occur in 70% of the women after the age of 25 years. Subsequently, real-time ultrasonography has been used for screening in nationally representative samples of civilian Mexicans, Hispanic White Americans, non-Hispanic white Americans, and nonHispanic black Americans of both genders ages 20 to 74. The cross-sectional prevalence rates of gallstones have been

found to be highest in certain tribes of Native Americans (e.g., Pima Indians), higher in Hispanic Americans than in whites, and lowest in black Americans.7 Figure 65-2 displays the world distribution of cholesterol gallstones. American Pima Indians are an extremely high-risk population. Other high-risk populations include Native American groups in North and South America and Scandinavians, of whom 50% develop gallstones by age 50. By contrast, African and Asian populations show the lowest risk of gallstones. Within a given population, first-degree relatives of index cases of persons with gallstones are 4.5 times as likely to form gallstones as matched controls. These epidemiologic investigations underscore the likely role of genetic predisposition in gallstone disease.

RISK FACTORS Age and Gender

Epidemiologic and clinical studies have reported that cholesterol gallstones occur infrequently in childhood and adolescence, and the prevalence of cholesterol gallstones increases linearly with age in both genders and approaches 50% at age 70 in women.14,15 Furthermore, elderly persons are at higher risk for complications of gallstones, and mortality from surgery is often unacceptably high in patients older than age 65. Cholesterol saturation of bile is significantly higher in elderly Swedes and Chilean women than in younger controls, and age correlates positively with an increased hepatic secretion rate of biliary cholesterol.16,17 In animals, aging has been shown to be associated with increased cholesterol gallstone formation as a result of increased biliary secretion and intestinal absorption of cholesterol, decreased hepatic synthesis and secretion of bile salts, and reduced gallbladder contractility (see later).18 Epidemiologic investigations have found, and clinical studies have confirmed, that at all ages, women are twice as likely as men to form cholesterol gallstones. The difference between women and men begins during puberty and continues through the childbearing years because of the effects of female sex hormones8 and differences between the sexes in how the liver metabolizes cholesterol in response to estrogen. Human and animal studies have suggested that

Chapter 65  Gallstone Disease

High Intermediate Low No data

estrogen increases the risk of cholesterol gallstones by augmenting the hepatic secretion of biliary cholesterol, thereby leading to an increase in cholesterol saturation of bile.19-22

Diet

Epidemiologic investigations have shown that cholesterol cholelithiasis is prevalent in populations that consume a Western diet that consists of high amounts of total calories, cholesterol, saturated fatty acids, refined carbohydrates, proteins, and salt, as well as a low amount of fiber. The incidence of cholesterol gallstones is significantly higher in North and South American as well as European populations than in Asian and African populations.3,23 Several clinical studies have found an association between the increased incidence of cholesterol gallstones in China and westernization of the traditional Chinese diet.24 In Japan, cholesterol cholelithiasis was once rare, but since the 1970s, the adoption of Western-type dietary habits has led to a markedly increased incidence.25

Pregnancy and Parity

Pregnancy is a risk factor for the development of biliary sludge and gallstones.26 During pregnancy, bile becomes more lithogenic as a result of increased estrogen levels, which result in increased cholesterol secretion and supersaturated bile. In addition, gallbladder motility is reduced, with a resulting increase in gallbladder volume and bile stasis. These alterations promote sludge and stone formation.27 Increased plasma levels of progestogen also reduce gallbladder motility. Because plasma hormone concentrations increase linearly with duration of gestation, the risk of gallstone formation is especially hazardous in the third trimester of pregnancy. Increasing parity is pro­ bably a risk factor for gallstones, especially in younger women.

Rapid Weight Loss

Rapid weight loss is a well-known risk factor for the formation of cholesterol gallstones.28 As many as 50% of obese

Figure 65-2.  Prevalence of cholesterol gallstones around the world.

patients who undergo gastric bypass surgery form biliary sludge and eventually gallstones within six months after surgery. Gallstones also develop in 25% of patients who undergo strict dietary restriction. Furthermore, 40% of these patients display symptoms related to gallstones within the same six-month period. The mechanisms by which rapid weight loss causes gallstone formation include increased hepatic secretion of biliary cholesterol during caloric restriction, increased production of mucin by the gallbladder, and impaired gallbladder motility. Gallstones could be prevented in this high-risk population by prophylactic administration of ursodeoxycholic acid (UDCA). UDCA in a dose of 600 mg/day has been reported to reduce the frequency of gallstones from 28% to 3% in obese patients on a verylow-calorie diet.29

Total Parenteral Nutrition

Total parenteral nutrition (TPN) is associated with the development of cholelithiasis and of acalculous cholecys­ titis. As early as three weeks after the initiation of TPN, biliary sludge often forms in the gallbladder because of prolonged fasting. In addition, the sphincter of Oddi may fail to relax, leading to preferential flow of bile into the gallbladder. Finally, approximately 45% of adults and 43% of children form gallstones after three to four months of TPN.30,31 Because patients receiving TPN often have serious medical problems and are not good candidates for abdom­ inal surgery, prophylactic treatment to prevent gallstones should be prescribed if no contraindication exists. Cholecystokinin (CCK) octapeptide administered twice daily via an intravenous line to patients on long-term TPN has proved to be safe and cost effective32 and should be used routinely in TPN-treated patients.

Biliary Sludge

Biliary sludge is a crucial intermediate stage in the pathogenesis of both cholesterol and pigment gallstones because it facilitates crystallization and agglomeration of solid cholesterol crystals, as well as precipitation of calcium bili­rubinate, and leads ultimately to the development

1091

1092

Section VIII  Biliary Tract into macroscopic stones.33,34 In addition, biliary sludge can induce acute cholecystitis, cholangitis, and acute pancrea­ titis. Furthermore, biliary sludge is associated with many conditions that predispose to gallstone formation, including pregnancy, rapid weight loss, spinal cord injury, long-term TPN, and treatment with octreotide.3 Although biliary sludge is reversible in most cases, it persists or disappears and reappears in 12% to 20% of affected persons and eventually leads to gallstones.35 Treatment of patients with persistent biliary sludge with UDCA decreases the frequency of clinical complications of biliary sludge.

Drugs

Estrogens Most, but not all, relevant studies have shown that the use of oral contraceptive steroids and conjugated estrogens in premenopausal women doubles the frequency of cholesterol gallstones.8,36 The administration of estrogen to postmenopausal women and of estrogen therapy to men with prostatic carcinoma have similar lithogenic effects.36,37 Therefore, estrogen has been proposed to be an important risk factor for the formation of cholesterol gallstones. In mice, the hepatic estrogen receptor α, but not β, plays a crucial role in the formation of cholesterol gallstones in response to estrogen.21 The hepatic estrogen receptor α, which is activated by estrogen, interferes with the negative feedback regulation of cholesterol biosynthesis by stimulating sterolregulatory element binding protein-2 (SREBP-2), with the resulting activation of the SREBP-2–responsive genes in the cholesterol biosynthetic pathway.22 These alterations result in increased secretion of newly synthesized cholesterol and supersaturation of bile, thereby predisposing to precipitation of cholesterol levels and formation of gallstones. In addition, estrogen leads to a decrease in plasma low-density lipoprotein (LDL) cholesterol levels and an increase in plasma high-density lipoprotein (HDL) cholesterol levels. The decrease in plasma LDL levels is a result of increased expression of the hepatic LDL receptor, which increases the clearance of plasma LDL. The increased uptake of LDL by the liver may also result in increased secretion of cholesterol into the bile. High levels of estrogen may impair gallbladder motility and consequently induce gallbladder hypomotility. Lipid-Lowering Drugs Lipid-lowering drugs may influence the formation of gallstones because they regulate key pathways in cholesterol and bile salt metabolism. Clofibrate is a lipid-lowering drug that has a significant association with gallstone formation. Clofibrate induces cholesterol supersaturation in bile and diminishes bile salt concentrations by reducing the activity of cholesterol 7α-hydroxylase (the rate-limiting enzyme in bile salt synthesis).38 The 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) reduce the biliary cholesterol saturation index, but their role in the prevention or therapy of gallstone disease needs to be further investigated.39 The potent cholesterol absorption inhibitor ezetimibe prevents the formation of cholesterol gallstones and facilitate the dissolution of gallstones in gallstonesusceptible C57L mice. Ezetimibe also may act as a potent biliary cholesterol-desaturating agent in patients with gallstones.40 Cholestyramine and nicotinic acid have no statistical association with gallstone formation. Octreotide The somatostatin analog octreotide increases the frequency of gallstones when administered to patients as treatment for acromegaly, with approximately 28% of treated acromegalic

patients forming gallstones. Acromegalic patients who are treated with octreotide display dysfunctional gallbladder motility, sluggish intestinal transit, and increased colonic deoxycholic acid formation and absorption41; all of these physiologic effects facilitate the formation of cholesterol gallstones. Ceftriaxone The third-generation cephalosporin ceftriaxone has a long duration of action, with much of the drug excreted in the urine. Approximately 40% of the drug, however, is secreted in an unmetabolized form into bile, where its concentration reaches 100 to 200 times that of the concentration in plasma and exceeds its saturation level in bile. Once the saturation level of ceftriaxone is exceeded, it complexes with calcium to form insoluble salts, thereby resulting in the formation of biliary sludge. Up to 43% of children who receive high-dose ceftriaxone (60 to 100 mg/kg/day) have been reported to form biliary sludge, and 19% of these patients experience biliary symptoms.42 The sludge usually disappears spontaneously after ceftriaxone is discontinued.

Lipid Abnormalities

Epidemiologic investigations have shown that plasma HDL cholesterol levels are inversely correlated with the frequency of cholesterol gallstones.43 By contrast, hypertri­ glyceridemia is positively associated with an increased frequency of gallstones.44 These seemingly independent variables are actually interrelated because high plasma triglyceride levels tend to increase with increasing body mass and are inversely correlated with plasma HDL levels. Interestingly, high plasma total and LDL cholesterol levels are not likely to be risk factors for the formation of gallstones.

Systemic Diseases

Obesity Obesity is a well-known risk factor for cholelithiasis. A large prospective study of obese women demonstrated a strong linear association between the body mass index (BMI) and frequency of cholelithiasis.45 In this study,45 the risk of gallstones was seven-fold higher in women with the highest BMI (>45 kg/m2) than in nonobese control women.47 Obesity is associated with increased hepatic secretion of cholesterol into bile, possibly because of higher levels of HMG-CoA reductase activity (the rate-limiting enzyme in cholesterol synthesis), which lead to high levels of cholesterol biosynthesis in the liver. Studies that compare the amounts of pronucleating and antinucleating factors in the gallbladder bile of obese and nonobese subjects have not been performed, nor have studies of gallbladder motility in obese persons. Diabetes Mellitus Patients with diabetes mellitus have long been assumed to be at an increased risk of gallstones because hypertriglyceridemia and obesity are risk factors for gallstones and are associated with diabetes mellitus and because gallbladder motility is often impaired in patients with diabetes mellitus. Proving that diabetes mellitus is an independent risk factor for gallstones has been difficult, however. Mice with hepatic insulin resistance induced by liver-specific disruption of the insulin receptor are markedly predisposed toward the formation of cholesterol gallstones because of increased expression of the biliary cholesterol transporters Abcg5 and Abcg8, with a resulting increase in hepatic sterol secretion, and decreased expression of the bile acid synthetic enzymes, particularly Cyp7b1, with a resulting lithogenic bile salt profile.46

Chapter 65  Gallstone Disease Diseases of the Ileum Disease or resection of the terminal ileum is considered a risk factor for gallstone formation. For example, intestinal bile salt absorption is often impaired in patients with Crohn’s disease, who are at increased risk of gallstones.47 The loss of specific bile salt transporters (e.g., ileal apical sodium/bile salt transporter) in the terminal ileum may result in excessive bile salt excretion in feces and a diminished bile salt pool size, presumably with a consequent increase in the risk of cholesterol gallstones. These changes may lead, however, to the formation of pigment gallstones because increased bile salt delivery to the colon enhances solubilization of unconjugated bilirubin, thereby increasing bilirubin concentrations in bile.48 Spinal Cord Injuries Spinal cord injuries are often associated with a high frequency of gallstones. The frequency of gallstones in patients with spinal cord injuries has been reported to be 31%, with an annual rate of biliary complications of 2.2%. Although the complication rate associated with gallstones in patients with spinal cord injuries is at least two-fold higher than the rate of gallstones in the general population, the relative risk is still low enough that prophylactic cholecystectomy is probably not justified. The mechanisms responsible for the association between spinal cord injuries and gallstone formation remain unclear. Gallbladder relaxation is impaired in these patients, but gallbladder contraction in response to a meal is normal. Therefore, the increased risk of gallstones is unlikely to be the result of biliary stasis alone.

COMPOSITION AND ABNORMALITIES OF BILE PHYSICAL CHEMISTRY OF BILE Chemical Composition of Bile

Cholesterol, phospholipids, and bile salts are the three major lipid species in bile; bile pigments are minor solutes. Cholesterol accounts for up to 95% of the sterols in bile and gallstones; the remaining 5% of the sterols are cholesterol precursors and dietary sterols from plant and shellfish sources. Concentrations of cholesteryl esters are negligible in bile; they account for less than 0.02% of total sterols in gallstones. The major phospholipids are lecithins (phosphatidylcholines), which account for more than 95% of total phospholipids; the remainder consists of cephalins (phosphatidylethanolamines) and a trace amount of sphingomyelin. Phospholipids constitute 15% to 25% of total lipids in bile. Lecithins are insoluble amphiphilic molecules with a hydrophilic, zwitterionic phosphocholine head group and hydrophobic tails that include two long fatty acyl chains. Biliary lecithins possess a saturated C-16 acyl chain in the sn-1 position and an unsaturated C-18 or C-20 acyl chain in the sn-2 position. The major molecular species of lecithins (with corresponding frequencies) in bile are 16 : 0 to 18 : 2 (40% to 60%), 16 : 0 to 18 : 1 (5% to 25%), 18 : 0 to 18 : 2 (1% to 16%), and 16 : 0 to 20 : 4 (1% to 10%). Lecithins are synthesized principally in the endoplasmic reticulum of the liver from diacylglycerol through the cytidine diphosphatecholine pathway. The common bile salts typically contain a steroid nucleus of four fused hydrocarbon rings with polar hydroxyl functions and an aliphatic side chain conjugated in amide linkage with glycine or taurine. In bile, more than 95% of bile salts are 5β,C-24 hydroxylated acidic steroids that are amide-linked to glycine or taurine in an approximate ratio of 3 : 1. Bile salts constitute approximately two thirds of the solute mass of normal human bile by weight.

The hydrophilic (polar) areas of bile salts are the hydroxyl groups and conjugated side chain of either glycine or taurine, and the hydrophobic (nonpolar) area is the ringed steroid nucleus. Because they possess both hydrophilic and hydrophobic surfaces, bile salts are highly soluble, detergent-like, amphiphilic molecules. Their high aqueous solubility is attributable to their capacity to self-assemble into micelles when a critical micellar concentration is exceeded. The primary bile salts are hepatic catabolic products of cholesterol and are composed of cholate (a trihydroxy bile salt) and chenodeoxycholate (a dihydroxy bile salt) (see Chapter 64). The secondary bile salts are derived from the primary bile salt species by the action of intestinal bacteria in the ileum and colon and include deoxycholate, ursodeoxycholate, and lithocholate. The most important of the conversion reactions is 7α-dehydroxylation of primary bile salts to produce deoxycholate from cholate and lithocholate from chenodoxycholate. Another important conversion reaction is the 7α-dehydrogenation of chenodeoxycholate to form 7α-oxo-lithocholate. This bile salt does not accumulate in bile but is metabolized by hepatic or bacterial reduction to form the tertiary bile salt chenodeoxycholate (mainly in the liver) or its 7β-epimer ursodeoxycholate (primarily by bacteria in the colon). Bile pigments are minor solutes and account for approximately 0.5% of total lipids in bile by weight. They are mainly bilirubin conjugates with traces of porphyrins and unconjugated bilirubin. In human bile, bilirubin monoglu­ curonides and diglucuronides are the major bile pigments. Other bile pigments are monoconjugates and diconjugates of xylose, glucose, and glucuronic acid and various homoconjugates and heteroconjugates of them. Proteins and elements are also found in bile. Albumin appears to be the most abundant protein in bile, followed by immunoglobulins G and M, apolipoproteins AI, AII, B, CI, and CII, transferrin, and α2-macroglobin. Other proteins that have been identified but not quantitated in bile include epidermal growth factor, insulin, haptoglobin, cholecystokinin, lysosomal hydrolase, and amylase. Elements detected in bile include sodium, phosphorus, potassium, calcium, copper, zinc, iron, manganese, molybdenum, magnesium, and strontium.

Physical States of Biliary Lipids

Cholesterol is nearly insoluble in water, and the mechanism by which cholesterol is solubilized in bile is complex because bile is an aqueous solution. The two main types of macromolecular aggregates in bile are micelles and vesicles. Bile salts are soluble in an aqueous solution because they are amphiphilic, in that they have both hydrophilic and hydrophobic areas. This unique property of bile salts is dependent on the number and characteristics of the hydroxyl groups and side chains as well as the composition of the particular aqueous solution. When bile salt concentrations exceed the critical micellar concentration, their monomers can aggregate spontaneously to form simple micelles. The simple micelles (∼3 nm in diameter) are small, disk-like, and thermodynamically stable aggregates that can solubilize cholesterol. They can also solubilize and incorporate phospholipids to form mixed micelles that are capable of solubilizing at least triple the amount of cholesterol compared with that solubilized by simple micelles. Mixed micelles (∼4-8 nm in diameter) are large, thermodynamically stable aggregates that are composed of bile salts, phospholipids, and cholesterol. Their size depends on the relative proportion of bile salts and phospholipids. The mixed micelle is

1093

1094

Section VIII  Biliary Tract a lipid bilayer with the hydrophilic groups of the bile salts and phospholipids aligned on the “outside” of the bilayer, interfacing with the aqueous bile, and the hydrophobic groups on the “inside.” Therefore, cholesterol molecules can be solubilized on the inside of the bilayer away from the aqueous areas on the outside. The amount of cholesterol that can be solubilized is dependent on the relative proportions of bile salts, and the maximal solubility of cholesterol occurs when the molar ratio of phospholipids to bile salts is between 0.2 and 0.3. Furthermore, solubility of cholesterol in mixed micelles is enhanced when the concentration of total lipids in bile is increased. Observations by quasi-elastic light-scattering spectroscopy and electron microscopy used to examine model and native biles reveal that in addition to micelles, vesicles solubilize cholesterol in bile. Biliary vesicles are unilamellar spherical structures that contain phospholipids, cholesterol, and little, if any, bile salts. Vesicles are substantially larger than either simple or mixed micelles (40 to 100 nm in diameter) but much smaller than liquid crystals (∼500 nm in diameter) that are composed of multilamellar spherical structures. Because vesicles are present in large quantities in hepatic bile, they presumably are secreted by hepatocytes. Unilamellar vesicles are often detected in freshly collected samples of unsaturated bile and are physically indistinguishable from those identified in supersaturated bile. Dilute hepatic bile, in which cholesterol crystals and gallstones never form, is always supersaturated with cholesterol because vesicles solubilize biliary cholesterol in excess of what could be solubilized in mixed micelles. Cholesterol-rich vesicles are remarkably stable in dilute bile, consistent with the absence of cholesterol crystallization in hepatic bile. The unilamellar vesicles can fuse and form large multilamellar vesicles (also known as liposomes or liquid crystals). Solid cholesterol crystals may nucleate from multilamellar vesicles in concentrated gallbladder bile. Vesicles are relatively static structures that are affected by several factors, including biliary lipid concentrations and the relative ratios of cholesterol, phospholipids, and bile salts. The relative concentrations of these three important lipids in bile are influenced by their hepatic secretion rates, which vary with fasting and feeding. For example, during fasting, biliary bile salt output is relatively low. As a result, the ratio of cholesterol to bile salts is increased, and more cholesterol is carried in vesicles than in micelles. By contrast, with feeding, biliary bile salt output increases and more cholesterol is solubilized in micelles than in vesicles. In addition, when the concentration of bile salts is relatively low, especially in dilute hepatic bile, vesicles are relatively stable, and only some vesicles convert to micelles. By contrast, with increasing bile salt concentrations in concentrated gallbladder bile, vesicles may transform or convert completely to mixed micelles. Because relatively more phospholipids than cholesterol can be transferred from vesicles to mixed micelles, the residual vesicles are remodeled and may be enriched in cholesterol relative to phospholipids. If the remaining vesicles have a relatively low cholesterol-to-phospholipid ratio (less than 1), they are relatively stable, but if the cholesterol-to-phospholipid ratio in vesicles is greater than 1, vesicles become increasingly unstable. These cholesterol-rich vesicles may transfer some cholesterol to less cholesterol-rich vesicles or to micelles or may fuse or aggregate to form larger (∼500 nm in diameter) multilamellar vesicles (i.e., liposomes or liquid crystals). Liquid crystals are visible by polarizing light microscopy as lipid circular droplets with characteristic birefringence in the shape of a Maltese cross. Liquid crystals

are inherently unstable and may form solid plate-like cholesterol monohydrate crystals, a process termed cholesterol nucleation. Therefore, nucleation of cholesterol monohydrate crystal results in a decrease in the amount of cholesterol contained in vesicles but not in micelles, and vesicles may serve as the primary source of cholesterol for nucleation. Under normal physiologic conditions, bile is concentrated gradually within the biliary tree so that the bile salt concentration approaches its critical micellar concentration. When this occurs, bile salts begin to modify the structure of phospholipid-rich vesicles that are secreted into bile by hepatocytes. These interactions signify the start of a complex series of molecular rearrangements that ultimately lead to formation of simple and mixed micelles. In supersaturated bile, two pathways result in the formation of cholesterol-rich vesicles from phospholipid-rich vesicles at the canalicular membrane. Because bile salts solubilize phospholipids more efficiently than cholesterol, cholesterol-rich vesicles may form when bile salts preferentially extract phospholipid molecules directly from phospholipid-rich vesicles. The alternative pathway is the rapid dissolution of phospholipid-rich vesicles by bile salts with the production of unstable mixed micelles that contain excess cholesterol. Structural rearrangements of these unstable micellar particles result in the formation of cholesterol-rich vesicles.

Phase Diagrams and Cholesterol Solubility in Bile

In the 1960s, Small and colleagues defined the maximal solubility (saturation) limits for cholesterol in model quaternary bile systems that consisted of varying proportions of cholesterol, phospholipids, bile salts, and water.49,50 The relative proportions (as molar percentages) of the three lipids in bile play a critical role in determining the maximal solubility of cholesterol. When the relative proportions of the three lipids at a fixed total lipid concentration are plotted in a triangular coordinate, the solubility of cholesterol for any given solute concentration can be determined.51 The triangular coordinate diagram also illustrates the physical phases of cholesterol in bile. For example, the phase diagram shown in Figure 65-3 is specific for a total lipid concentration of 7.5 g/dL, which is typical of human gallbladder bile.52,53 For hepatic bile, with a typical total lipid concentration of 3 g/dL, the phase boundaries would be different, with a smaller micellar zone, all phase boundaries shifted to the left, and an expanded two-phase zone on the right (i.e., region E in Figure 65-3). The effect of the total lipid content on cholesterol solubilization in the micellar zone explains why hepatic bile tends to be more saturated with cholesterol than is gallbladder bile in the same subject. Because hepatic bile contains cholesterol-phospholipid vesicles that are relatively stable, solid cholesterol monohydrate crystals never occur in hepatic bile. Equilibrium phase diagrams can also be used to predict the phases in which cholesterol crystals can be found at equilibrium. Although the equilibration process starts after bile is secreted from hepatocytes and flows into the biliary tree, the evolution to cholesterol monohydrate crystals occurs only in the gallbladder. For example, in unsaturated bile, all cholesterol can be solubilized in both simple and mixed micelles, and relative lipid compositions are located in the micellar zone of the phase diagram. By contrast, in supersaturated bile, cholesterol cannot be completely solubilized by simple and mixed micelles, and relative lipid compositions are located outside the micellar zone of the phase diagram. Under these circumstances, high vesicle cholesterol concentrations and high total lipid concentra-

Chapter 65  Gallstone Disease 100 0

20

rol st e ole

d(

Ch

lipi

3 Phases

ho

40

60

sp

o Ph

(% )

80

%)

60

40 D 20

A

B

C

E

2 Phases

80

2 Phases 1 Phase

0 100

80

60

100 40

20

0

Mixed bile salts (%) Figure 65-3.  Equilibrium phase diagram of cholesterol-phospholipid (lecithin)–mixed bile salt system (37°C, 0.15 M NaCl, pH 7.0, total lipid concentration 7.5 g/dL) showing positions and configuration of crystallization regions. The components are expressed in moles percent. The onephase micellar zone at the bottom is enclosed by a solid angulated line, and above it, two solid lines divide the two-phase zones from a central threephase zone. Based on the solid and liquid crystallization sequences present in the bile, the left two-phase and central three-phase regions are divided by dashed lines into regions A to D. The number of phases given represents the equilibrium state. The phases are cholesterol monohydrate crystals and saturated micelles for crystallization regions A and B; cholesterol mono­ hydrate crystals, saturated micelles, and liquid crystals for regions C and D; and liquid crystals of variable composition and saturated micelles for region E. Of note is that decreases in temperature (37°C→4°C), total lipid concentration (7.5 g/dL→2.5 g/dL), and bile salt hydrophobicity (3α,12α→3α,7α→3α,7α,12α→3α,7β hydroxylated taurine conjugates) progressively shift all crystallization pathways to lower phospholipid contents, retard crystallization, and reduce micellar cholesterol solubilities. These changes generate a series of new condensed-phase diagrams with an enlarged region E. (Reproduced with permission from Wang DQ, Carey MC. Complete mapping of crystallization pathways during cholesterol precipitation from model bile: Influence of physical-chemical variables of pathophysiologic relevance and identification of a stable liquid crystalline state in cold, dilute and hydrophilic bile salt-containing systems. J Lipid Res 1996; 37:606-30.)

tions in bile work together to produce the solid crystalline phase. Therefore, with typical physiologic lipid ratios, at equilibrium, cholesterol monohydrate crystals are present with saturated simple and mixed micelles or with saturated micelles plus vesicles that have become multilamellar liquid crystals. The final physical chemical state of bile is also influenced by the ratio of the concentration of bile salts to that of phospholipids and the overall hydrophilichydrophobic balance of both bile salt and phospholipid species. Within the micellar zone (see Fig. 65-3), bile is a visually clear, stable solution that is considered unsaturated because all cholesterol can be solubilized in thermodynamically stable simple and mixed micelles. At the boundary line of the micellar zone, bile is saturated because all the solubilizing capacity for cholesterol is utilized and no further cholesterol can be carried in micelles. Outside the micellar zone, bile is supersaturated because excess cholesterol cannot be solubilized by micelles51,54 and exists in more than one phase (micelles, liquid crystals, and solid monohydrate crystals); the solution is visually cloudy. Obviously,

relatively stable unilamellar cholesterol-phospholipid vesicles solubilize a significant proportion of cholesterol outside the micellar zone. The term metastable zone refers to the area in the phase diagram (above but near the micellar zone) in which bile is supersaturated with cholesterol but may not form solid cholesterol crystals even after many days. The diagram also suggests that when the quantity of cholesterol in bile exceeds that which can be solubilized by the available bile salts and phospholipids, solid cholesterol crystals precipitate in bile. Furthermore, the proportional distance outside the micellar zone directed along an axis joined to the cholesterol apex is often calculated as the cholesterol saturation index (CSI) (or lithogenic index).54 Therefore, the degree of saturation of bile with cholesterol can be quantitated. A CSI for a sample of bile can be estimated directly from the diagram or calculated by using a formula. The CSI is the ratio of the actual amount of cholesterol present in a bile sample to the maximal amount of cholesterol that can be dissolved in it. Bile that has a CSI of 1 is saturated; bile with a saturation index less than 1 is unsaturated; and bile with a saturation index greater than 1 is supersaturated. The degree of saturation can also be expressed as percent saturation by multiplying the saturation index by 100. For example, at the boundary of the micellar zone, bile is saturated, and the CSI is 100%. Supersaturated bile has a CSI above 100%, and unsaturated bile has a CSI below 100%. The CSI values are also useful for predicting the proportion of lipid particles and the metastable and equilibrium physical-chemical states in bile.

HEPATIC SECRETION OF BILIARY LIPIDS Source of Lipids Secreted in Bile

The supply of hepatic cholesterol molecules that can be recruited for biliary secretion depends on the balance of input and output of cholesterol and its catabolism in the liver (see also Chapter 72) (Fig. 65-4). Input is related to the amount of cholesterol (both unesterified and esterified) taken up by the liver from plasma lipoproteins (LDL > HDL > chylomicron remnants) plus de novo hepatic cholesterol synthesis. Output is related to the amount of cholesterol disposed of within the liver by conversion to cholesteryl ester (to form new very-low-density lipoprotein [VLDL] and for storage) minus the amount of cholesterol converted to primary bile salts. An appreciable fraction of cholesterol in bile also may be derived from the diet via apolipoprotein E–dependent delivery of chylomicron remnants to the liver. Under low or no dietary cholesterol conditions, bile contains newly synthesized cholesterol from the liver and preformed cholesterol that reaches the liver in several different ways. Approximately 20% of the cholesterol in bile comes from de novo hepatic biosynthesis, and 80% is from pools of preformed cholesterol within the liver. De novo cholesterol synthesis in the liver uses acetate as a substrate and is regulated mainly by the rate-limited enzyme HMG-CoA reductase. This enzyme can be up- or down-regulated depending on the overall cholesterol balance in the liver. An increase in the activity of this rate-limiting enzyme leads to excessive cholesterol secretion in bile. The major sources of preformed cholesterol are hepatic uptake of plasma lipoproteins (mainly HDL and LDL through their receptors on the basolateral membrane of hepatocytes). Consistent with their central role in reverse cholesterol transport, HDL particles are the main lipoprotein source of cholesterol that is targeted for biliary secretion. Under conditions of a diet high in cholesterol, dietary cholesterol reaches the liver through the intestinal lymphatic pathways as chylomicrons and then chylomicron remnants, after chylomicrons are

1095

1096

Section VIII  Biliary Tract BLOOD

LDL HDL

HEPATOCYTE Acetate LDLR

ABCB4

HMGCR Biosynthesis

Canalicular membrane

SR-BI

CMR Basolateral membrane

BILE

ACAT CH ester Esterification

CH

CYP7A1 VLDL ABCA1

PL

ABCG5/G8

CH BS

Catabolism

CYP27A1

Vesicle

ABCB11

BS

Figure 65-4.  Uptake, biosynthesis, catabolism, and biliary secretion of cholesterol at the hepatocyte level. The hepatic uptake of cholesterol is mediated by the low-density lipoprotein (LDL) receptor (LDLR) for LDL, by scavenger receptor class B type I (SR-BI) for high-density lipoprotein (HDL), and by the chylomicron remnant receptor (CMRR) for chylomicron remnants (CMR). Biosynthesis of hepatic cholesterol (CH) from acetate is regulated by the ratelimiting enzyme 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR). Part of the cholesterol is esterified by acyl-coenzyme A:cholesterol acyltransferase (ACAT) for storage in the liver. Some of the cholesterol is used for the formation of very-low-density lipoprotein (VLDL), which is secreted into the blood. The ATP-binding cassette (ABC) transporter ABCA1 may translocate, either directly or indirectly, cholesterol and phospholipids to the cell surface, where they appear to form lipid domains that interact with amphipathic α-helices in apolipoproteins. This interaction solubilizes these lipids and generates nascent HDL particles that dissociate from the cell. A proportion of cholesterol is used for the synthesis of bile salts via the “neutral” and the “acidic” pathways, as regulated by two rate-limiting enzymes cholesterol 7α-hydroxylase (CYP7A1) and sterol 27-hydroxylase (CYP27A1), respectively. Hepatic secretion of biliary cholesterol (CH), bile salts (BS), and phospholipids (PL) across the canalicular membrane is determined by three lipid transporters, ABCG5/G8, ABCB11, and ABCB4, respectively. A vesicle is shown in the canaliculus.

hydrolyzed by plasma lipoprotein lipase and hepatic lipase. The synthesis of new cholesterol in the liver is reduced and comprises only approximately 5% of biliary cholesterol. Overall, the liver can systematically regulate the total amount of cholesterol within it, and any excess cholesterol is handled efficiently. Although biliary phospholipid is derived from the cell membranes of hepatocytes, the composition of biliary phospholipid differs markedly from that of hepatocyte membranes. The membranes of hepatocytes contain phosphatidylcholines (lecithins), phosphatidylethanolamines, phosphatidylinositols, phosphatidylserines, and sphingomyelins. The major source of phosphatidylcholine molecules destined for secretion into bile is hepatic synthesis. A fraction of biliary phosphatidylcholines may also originate in the phospholipid coat of HDL particles. Approximately 11 g of phospholipids are secreted into bile each day in humans. More than 95% of bile salt molecules, after secretion into bile, return to the liver through the enterohepatic circulation by absorption mostly from the distal ileum via an active transport system (see also Chapter 64). Consequently, newly synthesized bile salts in the liver contribute only a small fraction (<5%) to biliary secretion and compensate for bile salts that escape intestinal absorption and are lost in feces. The fecal excretion of bile salts is increased when the enterohepatic circulation of bile salts is partially or completely interrupted by surgery, disease states, or drugs (e.g., bile salt-binding resins such as cholestyramine). Complete interruption of the enterohepatic circulation results in upregulation of bile salt synthesis in the liver, which restores bile salt secretion rates to approximately 25% of their usual values. Cholesterol from two sources serves as substrate for de novo bile salt synthesis—cholesterol that is newly syn-

thesized in the smooth endoplasmic reticulum and cholesterol that is preformed outside the smooth endoplasmic reticulum. The first step in this process is catalyzed by cholesterol 7α-hydroxylase. In the basal state, de novo bile salt synthesis uses principally newly synthesized cholesterol as substrate. When de novo cholesterol biosynthesis is suppressed by long-term therapy with a HMG-CoA reductase inhibitor, preformed cholesterol originating from plasma lipoprotein substitutes for newly synthesized cholesterol.

Biliary Lipid Secretion

Bile salts have been shown to stimulate secretion of vesicles, which are always detected in freshly collected hepatic bile.55,56 When cultured under specified conditions, rat hepatocytes form couplets with isolated “bile canaliculi” at the interface between adjoining cells. With the use of laser light-scattering techniques, vesicle formation can be observed within these bile canaliculi after exposure to bile salts. In addition, rapid fixation techniques and electronic microscopy have provided direct morphologic evidence of vesicle formation at the outer surface of the canalicular membrane.57,58 Most, if not all, bile salts are thought to enter canalicular spaces as monomers, whereas biliary phospholipids and cholesterol enter as unilamellar vesicles (see Fig. 65-4). A study of the molecular genetics of sitosterolemia (see Chapter 64) has shown that efflux of biliary cholesterol from the canalicular membrane is protein mediated. Two plasma membrane proteins—adenosine triphosphate (ATP)binding cassette (ABC) transporters ABCG5 and ABCG8— promote cellular efflux of cholesterol. The significance of this process for bile formation has been examined in genetically modified mice, in which overexpression of abcg5 and abcg8 in the liver was shown to increase the cholesterol

Chapter 65  Gallstone Disease content of gallbladder bile.59-63 Despite a reduced frequency of gallstones, the formation of gallstones is still observed in abcg5/g8 double-knockout mice, as well as in abcg5 or abcg8 single-knockout mice, challenged with a lithogenic diet.59-63 These findings strongly suggest the existence of an ABCG5/G8-independent pathway for hepatic secretion of biliary cholesterol and its role in the formation of cholesterol gallstones. In addition, scavenger receptor class B type I (SR-BI) is localized in sinusoidal, and possibly, canalicular, membranes of the hepatocyte, and in transgenic and knockout mice fed a chow diet, biliary secretion of cholesterol varies in proportion to hepatic expression of SR-BI and to the contribution of SR-BI to sinusoidal uptake of HDL cholesterol destined for secretion into bile.64,65 Deletion of the Abcb4 gene completely inhibits hepatic secretion of biliary phospholipids in mice,66 suggesting that ABCB4 could be responsible for the trans­ location, or “flip,” of phosphatidylcholine from the endoplasmic (inner) to ectoplasmic (outer) leaflet of the canalicular membrane bilayer and that the action of ABCB4 may form phosphatidylcholine-rich microdomains within the outer membrane leaflet. Although the ectoplasmic leaflet of the canalicular membrane is cholesterol- and sphingomyelin-rich and is relatively resistant to penetration by bile salts, bile salts may promote vesicular secretion of biliary cholesterol and phosphatidylcholine. Bile salts may partition preferentially into these areas to destabilize the membrane and release phosphatidylcholine-rich vesicles because detergent-like bile salt molecules within the canalicular space could interact with canalicular membrane. Mutations of the ABCB4 gene in humans result in the molecular defect underlying type 3 progressive familial intra­ hepatic cholestasis (see Chapter 76).67 Biliary bile salts include those that are newly synthesized in the liver and those that undergo enterohepatic cycling. The precise molecular mechanism of bile salt secretion is not known, although it probably involves ABCB11, a bile salt export pump (see Chapter 64).68-70 Although hepatic secretion of biliary bile salts directly affects cholesterolphospholipid vesicle secretion, whether bile salt secretion is coupled to cholesterol and phospholipid secretion at a molecular level is not known. The relationship between bile salt secretion and cholesterol secretion is curvilinear: At low bile salt secretion rates (usually less than 10 µmol/hr/ kg), more cholesterol is secreted per molecule of bile salt than at higher rates. Although bile salt secretion rates are not low in normal subjects, they could diminish during prolonged fasting, during the overnight period, and with substantial bile salt losses, as occur with a biliary fistula or ileal resection when the liver cannot compensate sufficiently by increasing bile salt synthesis. At high bile salt secretion rates, for example, during and after eating, biliary cholesterol saturation is less than that during interprandial periods. In laboratory animals, biliary secretion of organic anions does not influence bile salt secretion but does inhibit secretion of phospholipid and cholesterol into bile because organic anions bind bile salts within bile canaliculi and prevent interactions with the canalicular membrane.

PATHOPHYSIOLOGY As shown in Figure 65-5, at least five primary defects must be present simultaneously for cholesterol gallstone for­ mation: certain genetic factors, including LITH genes (see later); hepatic hypersecretion; gallbladder hypomotility; rapid phase transitions; and certain intestinal factors.

Genetic factors and LITH genes Intestinal factors

Hepatic hypersecretion

Gallbladder hypomotility

Cholesterol Gallstones Rapid phase transitions

Figure 65-5.  Venn diagram of five primary defects that must be present simultaneously in gallbladder bile for cholesterol crystallization to take place. The five defects are genetic factors and LITH (gallstone) genes, hepatic hypersecretion of cholesterol, gallbladder hypomotility, rapid phase transitions, and intestinal factors. The hypothesis proposed is that hepatic cholesterol hypersecretion into bile is the primary defect and is the outcome in part of a complex genetic predisposition. The downstream effects include gallbladder hypomotility and rapid phase transitions (see Fig. 65-3). A major result of gallbladder hypomotility is alteration in the kinetics of the enterohepatic circulation of bile salts (intestinal factors). The alterations in the intestinal factors result in increased cholesterol absorption as well as reduced bile salt absorption that lead to abnormal enterohepatic circulation of bile salts and diminished biliary bile salt pool size. Not only does gallbladder hypomotility facilitate nucleation, but it also allows the gallbladder to retain cholesterol monohydrate crystals. Although a large number of candidate Lith genes have been identified in mouse models, the identification of human LITH genes and their contributions to gallstones require further investigation.

HEPATIC HYPERSECRETION OF BILIARY CHOLESTEROL

Hepatic hypersecretion of biliary cholesterol plays a primary role in the pathogenesis of cholesterol gallstone formation (see later). By definition, supersaturated bile contains cholesterol that cannot be solubilized at equilibrium by bile salts and phospholipids. Cholesterol supersaturation could result from (1) excessive hepatic secretion of biliary cho­ lesterol; (2) decreased rates of bile salt or phospholipid secretion into bile with relatively normal cholesterol secretion; or (3) a combination of hypersecretion of cholesterol and hyposecretion of the solubilizing lipids. With the passage of time and in the presence of heterogeneous pronucleating agents, usually mucin gel (see later), cholesterol supersaturation leads to precipitation of cholesterol monohydrate crystals in gallbladder bile, followed by agglomeration and growth of the crystals into mature and macroscopic stones.

1097

1098

Section VIII  Biliary Tract RAPID CHOLESTEROL NUCLEATION AND CRYSTALLIZATION

Cholesterol nucleation and crystallization is a process by which solid plate-like cholesterol crystals precipitate from supersaturated bile. The crystals can be detected by polarizing light microscopy in a sample of bile previously rendered crystal-free (“isotropic”).71 Bile from patients with cholesterol gallstones and from controls is supersaturated with cholesterol, and the degree of cholesterol supersaturation is not a reliable predictor of gallstones. On the other hand, rapid in vitro nucleation and crystallization of cholesterol monohydrate crystals from the isotropic phase of gallbladder bile distinguishes the lithogenic bile of patients with cholesterol gallstones from cholesterol-supersaturated bile of non-gallstone control subjects.71 The phase diagram of cholesterol, phospholipids, and bile salts discussed earlier (see Fig. 65-3) is often used to study the phase transitions where metastable intermediates form. Five crystallization pathways can be identified on the basis of the bile salt-to-phospholipid ratio, total lipid concentration, bile salt species (hydrophilic and hydrophobic properties), temperature, and CSI.52,72 Furthermore, these crystallization pathways have been confirmed in fresh human and mouse gallbladder bile.52,72,73 In Figure 65-3, for the cholesterolphospholipid-mixed bile salt model system, the five distinct crystallization pathways are designated A to E, with each representing a different sequence of phase transitions, including an anhydrous cholesterol pathway and a liquid crystalline pathway leading to the formation of solid platelike cholesterol monohydrate crystals.52,72 Transient arc-like crystals appear in some of the pathways and are consistent with crystalline anhydrous cholesterol (see later).74,75 Why anhydrous cholesterol crystals should precipitate in an aqueous environment is unknown, but they are characteristic of the pathways that seem to originate from unilamellar, as opposed to multilamellar, vesicles. In these pathways, the critical nucleus may be a unilamellar vesicle that contains liquid anhydrous cholesterol molecules in its core, possibly reflecting internal nucleation. In essence, these early vesicular “nuclei” may already have initiated the nucleation cascade by the time bile enters the gallbladder. The current paradigm for nucleation and crystallization, based principally on observations from video-enhanced polarized light microscopy, suggests that biliary vesicles must fuse or at least aggregate to form crystalline cholesterol monohydrate. Because cholesterol nucleation and crystallization are apparently initiated in vesicles, the stability of the vesicle determines the stability of bile. Unstable vesicles can fuse, aggregate, and grow into multilamellar liquid crystalline structures (liposomes) in which cholesterol crystallizes out of solution. Furthermore, evidence from quasielastic light-scattering spectroscopy shows that nucleation of solid cholesterol crystals may occur directly from supersaturated micelles in conjugated deoxycholate-rich bile in vitro without an intervening vesicle or liquid crystalline phase. In bile with the lowest phospholipid contents (region A in Fig. 65-3), arc-like crystals with a density (d = 1.030 g/mL) consistent with anhydrous cholesterol appear first and evolve via helical and tubular crystals to form plate-like cholesterol monohydrate crystals (d = 1.045 g/mL).52,74,75 With higher phospholipid contents (region B), cholesterol monohydrate crystals appear earlier than arc-like crystals and other transitional crystals. With typical physiologic phospholipid contents (region C), early liquid crystals (d = 1.020 g/mL) are followed by cholesterol monohydrate crystals; subsequently, arc-like and other intermediate crystals appear. With still higher phospholipid contents (region D),

liquid crystals are followed by cholesterol monohydrate crystals only. At the highest phospholipid mole fractions (region E), liquid crystals are stable and no solid crystals form. Decreases in temperature (37°C→4°C), total lipid concentration (7.5 g/dL → 2.5 g/dL), and bile salt hydro­ phobicity (3α,12α→3α,7α→3α,7α,12α→3α,7β hydroxylated taurine conjugates) progressively shift all crystallization pathways to lower phospholipid contents, reduce micellar cholesterol solubilization, and retard crystallization.52,72 Cholesterol crystallization pathways and sequences in human gallbladder bile are identical to those of model bile samples matched for appropriate physical-chemical conditions, and in a physiologic state, three of the five sequences observed in model bile samples are found in human and mouse gallbladder bile.72 Most notably, the kinetics of all these phase transitions are faster in lithogenic human bile than in identically patterned model bile samples, most likely a result, in part, of the combined influences of increased levels of cholesterol, secondary bile salts, and mucin glycoproteins.53 In addition, biliary lipid, electrolyte, and protein factors may be important in stabilizing supersaturated bile. Nonprotein factors that retard cholesterol nucleation and crystallization include (1) a total lipid concentration less than 3 g/dL; (2) reduced hydro­ phobicity of the bile salt pool; (3) low bile salt-to-lecithin ratios; (4) low cholesterol-to-lecithin ratios in vesicles; and (5) low total calcium ion concentrations. The states opposite to these conditions accelerate nucleation and crystallization.76

IMBALANCE OF PRONUCLEATING AND ANTINUCLEATING FACTORS

More rapid crystallization of cholesterol in the bile of patients with gallstones implies that lithogenic bile may contain pronucleating agents that accelerate crystallization or that normal bile may contain antinucleating agents that inhibit crystallization. Furthermore, bile may contain both accelerators and inhibitors of crystallization, and imbalances between them can induce rapid crystallization in gallbladder bile in patients with cholesterol gallstones.77,78 Mucin was the first biliary protein shown to promote cholesterol crystallization.79 The epithelial cells of the gallbladder secrete mucin that normally serves as a protective layer over the mucosa. Mucin or mucin glycoproteins are large molecules that consist of a protein core and many carbohydrate side chains.80 An important property of mucin is its ability to form a gel phase in higher concentrations, and the gel has greatly increased viscosity compared with the sol (soluble) phase. Gallbladder mucins, a heterogeneous family of O-linked glycoproteins, are divided into two classes: epithelial and gel-forming mucins.81 The epithelial mucins, which are produced by mucin gene 1 (MUC1), MUC3, and MUC4, do not seem to form aggregates and are integral membrane glycoproteins located on the apical surface of epithelial cells.82-85 The gel-forming mucins, MUC2, MUC5AC, and MUC5B, which are secreted by specialized gallbladder mucin-producing cells, provide a protective coating on the underlying mucosa.82-85 They form disulfide-stabilized oligomers or polymers, a phenomenon that accounts for their viscoelastic properties. Mucins from different organs vary in carbohydrate side chain, protein composition, and charge but generally have similar pro­ perties. Mucins have hydrophilic domains to which many water molecules bind. They have an overall charge and are capable of binding other charged species such as calcium. Hydrophobic domains in the mucin molecule (on the nonglycosy­lated regions of the polypeptide core) allow

Chapter 65  Gallstone Disease binding of lipids such as cholesterol, phospholipids, and bilirubin. Evidence suggests that gallbladder mucins play an important role in the early stages of gallstone formation and are a potent pronucleating agent for accelerating cholesterol crystallization in native and model biles. Indeed, hypersecretion of gallbladder mucins is a prerequisite for gallstone for­mation, and increased amounts of gallbladder mucins are consistently observed in gallbladder bile of several animal models of gallstones.73,79,86 Mucins are also found within gallstones, where they act as a matrix for stone growth.87 The mucins in gallstones have been found to extend from the amorphous center to the periphery in either a radial or laminated fashion. Mucins are also a major component of sludge in the gallbladder, and sludge has been suggested to be a precursor of gallstones. Therefore, two roles in the formation of gallstones have been proposed for mucins (1) a pronucleating agent for the nucleation and crystallization of cholesterol from saturated bile and (2) a scaffolding for the deposition of crystals during the growth of stones. The synthesis of mucin glycoproteins that are secreted by the epithelium of the gallbladder and biliary ducts may be regulated by mucosal prostaglandins that are derived from arachidonic acid–containing biliary phospholipids.80 During the formation of gallstones, the gallbladder hypersecretes mucins, mostly as a result of stimulation by some components of saturated bile. Then, the carbohydrate groups of the polymers of mucins avidly bind water to form gels. The hydrophobic polypeptides in the core of mucin glyco­ proteins also can bind bilirubin and calcium in bile. The resulting water-insoluble complex of mucin glycoproteins and calcium bilirubinate provides a surface for nucleation of cholesterol monohydrate crystals and a matrix for the growth of stones. Secretion and accumulation of mucin in the gallbladder is controlled by multiple mucin genes. MUC1 mucin in the gallbladders of mice has been shown to be reduced with disruption of the Muc1 gene, with a consequent decrease in susceptibility to cholesterol gallstone formation.88 Also, gene expression of the gallbladder Muc5ac gel-forming mucin gene is significantly reduced in Muc1-knockout mice in response to a lithogenic diet. As a result, cholesterol crystallization and the development of gallstone formation are significantly retarded. These findings suggest that genegene interactions between the Muc1 and Muc5ac genes might affect mucin secretion and accumulation in the gallbladder. Furthermore, increased gallbladder epithelial Muc1 mucin enhances cholelithogenesis mostly by promoting gallbladder cholesterol absorption and impairing gallbladder motility in mice that are transgenic for the human MUC1 gene; this lithogenic mechanism is completely different from that associated with the gel-forming mucins.89 Collectively, these findings support the concept that inhibition of the secretion and accumulation of not only the gelforming mucins, but also the epithelial mucins in the gallbladder may completely prevent the formation of cholesterol gallstones. Many glycoproteins that bind reversibly to concanavalin A-Sepharose also speed up cholesterol crystallization.90 These glycoproteins include aminopeptidase N, immunoglobulins, α1-acid glycoprotein, phospholipase C, fibronectin, and haptoglobin. Other pronucleating agents are the amphipathic anionic polypeptide fraction/calcium-binding protein, albumin-lipid complexes, and group II phospholipase A2. Nonprotein components of bile also expedite cholesterol crystallization. Calcium bound to micelles and vesicles in bile may accelerate cholesterol crystallization by promoting fusion of cholesterol-rich vesicles. The precipita-

tion of calcium salts in bile that is supersaturated with calcium salts and cholesterol may lead to rapid crystallization of cholesterol, an effect that is enhanced by the presence of mucins. The rapidity of cholesterol crystal formation also varies in proportion to the deoxycholate content of bile and is related to the effect of deoxycholate on the equilibrium phase relationships of biliary lipids. The degree of cholesterol supersaturation of bile may also be a determinant of rapid crystallization of cholesterol. Several inhibitors of cholesterol crystallization have been identified, including apolipoproteins AI and AII, a 120-kd glycoprotein, a 15-kd protein, and secretory immunoglo­ bulin A and its heavy and light chains.91-93 Apolipoproteins AI and AII may prolong crystal detection time of supersa­ turated model bile. Apolipoproteins AI and AII are present in a fraction of human bile that may inhibit cholesterol nucleation and crystallization. Precholecystectomy treatment with UDCA for three months prolongs the crystal detection time of bile from patients with cholesterol gallstones, suggesting also that UDCA could be an antinucleating factor.52,94-96 UDCA may exert its effect by stabilizing vesicles, perhaps by enhancing the incorporation of apolipoprotein AI into (or onto) the vesicles. In addition, a potential antinucleating factor from normal human gallbladder bile is detected by lectin affinity chromatography and high performance liquid ion-exchange chromatography and found to be a slightly acidic glycoprotein with an apparent molecular size of 120 kd. The protein may inhibit crystal growth by attaching to the most rapidly growing microdomains on a crystal face and interfering with further solute attachment. Whether only one or several antinucleating factors exist and how they may inhibit the initiation of cholesterol crystal formation are uncertain, but unilamellar vesicles have been proposed to be the key sites of action. In summary, although many biliary proteins, besides mucin gel, have been proposed as either pronucleating or antinucleating factors influencing cholesterol nucleation and crystallization in bile, their in vivo role, if any, in the pathogenesis of cholesterol gallstone formation remains unclear. Furthermore, proteolysis of soluble biliary glycoproteins does not influence the detection time of cholesterol monohydrate crystals either in normal or abnormal gallbladder and hepatic bile, an observation that suggests that soluble biliary proteins may not be an important pathophysiologic factor in cholesterol crystallization.

GALLBLADDER DYSFUNCTION

Under normal physiologic conditions, frequent gallbladder contractions occur throughout the day. Between meals, the gallbladder stores hepatic bile (with an average fasting volume of 25 to 30 mL in healthy subjects). Following a meal, depending on the degree of neurohormonal response, the gallbladder discharges a variable amount of bile.97 Use of a combined method of cholescintigraphy and ultrasonography has demonstrated that after a meal, the gallbladder empties immediately and refills repeatedly.97 By contrast, an increased fasting gallbladder volume as well as incomplete emptying and high residual gallbladder volume are often observed in patients with cholesterol gallstones, whether they have tiny or large stones or simply lithogenic bile. In this group of patients with cholesterol gallstones and gallbladder motility abnormalities, gallbladder wall inflammation is usually mild and cannot account for the impaired dynamics of the gallbladder. Furthermore, the poor interdigestive gallbladder filling is consistent with delivery of a greater percentage of lithogenic bile from the liver directly into the small intestine, with augmentation of the enterohepatic effects of increased recycling and bile salt hydropho-

1099

1100

Section VIII  Biliary Tract bicity. This observation suggests that emptying and filling of the gallbladder are affected in patients with gallbladder hypomotility.97,98 Clinical investigations confirm that gallbladder hypomotility is associated principally with the formation of cholesterol gallstones, although a milder degree of gallbladder dysmotility, in the absence of an enlarged gallbladder in the fasting state and any gallbladder inflammation, is also found in patients with pigment gallstones.99 In patients with cholesterol gallstones, impaired gallbladder motility persists in the stone-free gallbladder following successful extracorporeal shock-wave lithotripsy and oral bile acid dissolution therapy.100,101 The degree of impairment of gallbladder emptying has been found to increase in proportion to the cholesterol content of gallbladder bile, even in healthy subjects without gallstones. These findings suggest that excess cholesterol molecules in the gallbladder wall may act as myotoxic agents. In vitro studies in which gallbladder function was compared in patients with cholesterol gallstones and control subjects have shown abnormalities in the binding of agonists such as CCK to plasma membrane CCK-1 receptors, alterations in contraction of isolated smooth muscle cells, and decreased contractility of isolated smooth muscle strips and whole gallbladder preparations. In particular, signal transduction in response to binding of agonists is impaired. Defects in contractility associated with cholesterol gallstones are reversible at an early stage and are attributable primarily to excess accumulation of biliary cholesterol in the membranes of gallbladder smooth muscle cells. This mechanism appears to explain why gallbladder emptying is impaired before gallstones are formed in animal models at a time when bile is supersaturated with cholesterol. In addition, the intracellular mechanisms of smooth muscle contraction in human gallbladder muscle cells from patients with cholesterol gallstones seem to be intact. These findings support the hypothesis that the absorption of cholesterol from the gallbladder lumen is associated with gallbladder smooth muscle dysfunction. This alteration may induce stiffening of sarcoplasmic membranes secondary to an increase in cholesterol content of the membranes. As a result, when CCK binds to its receptor on smooth muscle cells of the lithogenic gallbladder, G-proteins are not activated and gallbladder motility is impaired.102,103 Gallbladder hypomotility could precede gallstone formation. Gallbladder stasis induced by the hypofunctioning gallbladder could provide the time necessary to accom­ modate nucleation of cholesterol crystals and growth of gallstones within the mucin gel in the gallbladder.104,105 Furthermore, the viscous mucin gel that forms within the gallbladder may contribute to hypomotility by impairing gallbladder emptying mechanically, possibly at the level of the cystic duct. In particular, sludge contains calcium, pigment, bile salts, and glycoproteins and could serve as a nidus for nucleation and crystallization of cholesterol or precipitation of calcium bilirubinate. The high frequency of cholelithiasis in patients receiving long-term TPN highlights the importance of gallbladder stasis in the formation of gallstones.106 For example, 49% of patients with Crohn’s disease who are on TPN have gallstones, whereas Crohn’s disease alone leads to gallstones in 27% of patients. During TPN, the gallbladder does not empty completely because the stimulus (ingestion of meals) for the release of CCK is eliminated. As a result, bile stagnates and sludge develops in the gallbladder, thereby enhancing the formation of gallstones. Daily intravenous administration of CCK can completely prevent gallbladder dysmotility and eliminate the inevitable risk of biliary sludge and gallstone formation. In addition, slow emptying and increased volume of the

gallbladder, as measured by ultrasonography, occur during pregnancy and during administration of oral contraceptives, two conditions that predispose to the formation of gallstones (see earlier).20,21 The concentration of bile by the gallbladder increases cholesterol solubility; however, it also enhances cholesterol nucleation and crystallization in bile, thereby suggesting that increased concentration of bile is a contributing factor for gallstone formation.107,108 In addition to concentrating bile, the normal gallbladder can also acidify bile. Acidification increases the solubility in bile of calcium salts (e.g., bilirubinate and carbonate), which may be promoters of nucleation and crystallization. Therefore, defective acidi­ fication may have an effect on the formation of gallstones. Differential absorption rates of cholesterol, phospholipids, and bile salts by the gallbladder epithelial cells may reduce cholesterol saturation of bile in normal subjects; however, the gallbladder epithelium of patients with cholesterol gallstones loses the capacity for selective absorption of biliary cholesterol and phospholipids.109,110 Impaired lipid absorption by the gallbladder may contribute to gallstone formation by sustaining cholesterol supersaturation of bile during storage.111 The physical-chemical fate of cholesterol absorbed by the gallbladder may be similar to that which occurs during the development of an atherosclerotic plaque. In all likelihood, cholesterol molecules are absorbed continuously by the gallbladder mucosa from supersaturated bile,112 and the unesterified cholesterol molecules diffuse rapidly to the muscularis propria because the gallbladder lacks an intervening muscularis mucosae and submucosa. Because the gallbladder apparently does not synthesize lipoproteins for exporting cholesterol to plasma, excess unesterified cholesterol molecules are removable from gallbladder mucosa and muscle only by esterification and storage or back diffusion into bile.113 In the lithogenic state, back diffusion of cholesterol molecules into bile is blocked because gallbladder bile is continuously saturated. As a result, gallbladder mucosal acyl-coenzyme A:cholesterol acyltransferase (ACAT) esterifies most, but not all, cholesterol molecules. As in an atherosclerotic plaque, mucosal and muscle membranes apparently become saturated with cholesterol and coexist with stored cholesteryl ester droplets. Furthermore, the unesterified cholesterol molecules become intercalated within the membrane bilayer of muscle cells, a process that may alter the physical state of phospholipid molecules, as reflected by their increased rigidity. Consequently, gallbladder motility function is impaired because signal transduction in response to CCK is diminished markedly. In addition, excess cholesterol molecules absorbed from the lithogenic bile may be direct stimulants to proliferative and inflammatory changes in the mucosa and lamina propria of the gallbladder.97

Intestinal Factors

The high efficiency of intestinal cholesterol absorption correlates positively and significantly with the frequency of cholesterol gallstones in inbred mice, and gallstonesusceptible C57L mice display significantly higher intestinal cholesterol absorption than gallstone-resistant AKR mice.114 These observations suggest that high dietary cholesterol intake and high efficiency of intestinal cholesterol absorption are independent risk factors for cholesterol gallstone formation. Differences in the metabolism of chylomicron remnant cholesterol between C57L and AKR mice may account for lithogenic bile formation in the former, and the cholesterol absorbed from the small intestine provides an important source for biliary cholesterol hypersecretion in mice challenged by a lithogenic diet.115

Chapter 65  Gallstone Disease Altered intestinal motility also may have a role in gallstone formation. Delayed or impaired small intestinal transit is associated with enhanced intestinal cholesterol absorption, biliary cholesterol secretion, and gallstones in CCK-1 receptor-knockout mice.115 The association of impaired colonic motility with increased biliary deoxycholate levels is found in some patients with cholesterol gallstones. Evidence for a causal relation between impaired intestinal motility, deoxycholate formation, and bile lithogenicity comes from studies in humans and mice. Clinical studies have found that acromegalic patients treated with octreotide (a known risk factor for cholesterol gallstone disease [see earlier]) display prolonged colonic transit times, high levels of biliary deoxycholate concentration, and biliary cholesterol precipitation.116-119 Furthermore, higher levels of biliary deoxycholate are associated with increased amounts of gram-positive anaerobic bacteria and increased activity of 7α-dehydroxylase in the cecums of patients with cholesterol gallstones compared with control subjects who have no stones.120 Biliary deoxycholate and cholesterol concentrations can be lowered by antibiotic treatment that reduces fecal 7α-dehydroxylation activity. Compared with resistant mice, gallstone-susceptible mice also have high biliary levels of deoxycholate, which are associated with cholesterol supersaturation and gallstone formation.73,121 Chronic intestinal infection has been proposed to be a potential factor in cholesterol gallstone pathogenesis. A mouse study has shown that distal intestinal infection with a variety of enterohepatic Helicobacter species (but not Helicobacter pylori) is essential for nucleation and crystallization of cholesterol from supersaturated bile.122,123 These Helicobacter species also have been identified in the bile and gallbladder tissue of Chilean patients with chronic cholecystitis.124 Whether chronic intestinal infection has a direct pathogenic role in the formation of cholesterol gallstones requires further investigation. Patients with Crohn’s disease and those who have undergone intestinal resection or total colectomy have bile that is supersaturated with cholesterol and are prone to precipitation of cholesterol crystals and formation of gallstones.125 The enterohepatic circulation of bile salts is probably impaired in these patients so that biliary bile salt secretion is greatly reduced and the solubilization of cholesterol in bile is decreased. Moreover, Crohn’s disease might lead to impaired enterohepatic cycling of bilirubin so that biliary bilirubin levels and precipitation of calcium bilirubinate are increased, thereby providing a nidus for cholesterol nucleation and crystallization.48,126

Growth of Gallstones

Although cholesterol nucleation and crystallization is a critical stage in the formation of cholesterol gallstones, findings in patients who have cholesterol crystals but no gallstones in the gallbladder suggest that growth of cholesterol crystals into gallstones does not always follow crystallization. Stone growth may represent a second critical stage in the formation of gallstones that results from delayed emptying of the gallbladder. When multiple gallstones are found in the gallbladder, they often are equal in size, indicating that cholesterol crystallization for this family of stones occurred simultaneously and the stones grew at the same rate. By contrast, stones of unequal size could represent different generations. The amorphous material in the center of stones contains bilirubin, bile salts, mucin glycoproteins, calcium carbonate, phosphate, copper, and sulfur, which could have provided a required nidus for cholesterol nucleation and crystallization. Cholesterol crystals could assemble about this nidus. The formation of a nidus and

subsequent stone growth could be determined by mucins, other biliary proteins, and the cholesterol saturation of bile. The growth of stones is likely a discontinuous process that is punctuated by deposition of rings of calcium bilirubinate and calcium carbonate. Because cholesterol crystals often aggregate randomly in amorphous groupings and layer radially and concentrically, cholesterol stones consist of radially or horizontally oriented cholesterol crystals embedded within an organic matrix. In the outer portion of stones, cholesterol crystals are oriented perpendicularly to the surface.127 Throughout the formation of gallstones, mucins could provide a matrix on which the growth of gallstones occurs. Furthermore, concentric pigmented rings separate layers of cholesterol crystals that have different axial orientations. The chemical composition of these rings often resembles the center of gallstones, and the rings may reflect cyclic deposition of calcium bilirubinate, other calcium salts, and mucin glycoproteins.

GENETICS The evidence for a genetic component of cholesterol gallstone disease in humans is mostly indirect and based on geographic and ethnic differences, as well as on family and twin studies.13,128-135 A genetic predisposition is clearly present in the Pima and certain other North and South American Indians, who display the highest frequency rates (48%) of gallstones.13,128,129 By contrast, the overall frequency in other American (whites) and European populations is about 20%. The lowest rates (<5%) are observed in African populations and intermediate rates are found in Asian populations (5% to 20%), as shown in Figures 65-1 and 65-2. Although some independent risk factors, such as aging, gender, parity, obesity, some drugs, and rapid weight loss, for gallstone formation have been found (see later),18,22,46,136-138 none of these factors can explain the striking differences in incidence rates of gallstones among different populations, thereby suggesting a genetic contribution to the etiology of the disease. Gallstones are more frequent by a ratio of 3 : 1 in siblings and other family members of affected persons than in spouses or unrelated controls.130 Using ultrasonography to detect gallstones in first-degree relatives of index patients, Gilat and colleagues132 found a 21% frequency in firstdegree relatives compared with 9% in matched controls, and Sarin and coworkers133 also observed a frequency that was five times higher in relatives than in controls. Furthermore, cholesterol supersaturation is higher in fasting duodenal bile of older sisters of patients with cholesterol gallstone than in controls.134 Cholesterol synthesis rates, bile saturation levels, and gallstone frequency rates are also significantly higher on pair-wise correlations in monozygotic than in dizygotic male twins.135 Despite these observations, a mode of inheritance that fits a mendelian pattern cannot be shown in most cases. To examine the influence of the genetic factors more rigorously, a study of populations with different incidence rates of gallstones but living in the same environment should provide insights into genetic mechanisms of the disease. Unfortunately, intermarriages between two populations result in a rapid loss of the original genetic background within a few generations, thereby making such a study impossible. A large study of 43,141 twin pairs in Sweden, however, has provided conclusive evidence for the role of genetic factors in the pathogenesis of cholesterol gallstones.139 In this study, concordance rates were significantly

1101

1102

Section VIII  Biliary Tract higher in monozygotic twins than in dizygotic twins, with genetic factors accounting for 25% of the phenotypic variation between the twins. The first evidence that human gallstones might be caused by a single gene defect came from a study by Lin and colleagues,140 who reported that among 232 MexicanAmericans, a variant of the cholesterol 7α-hydroxylase (CYP7A1) gene was associated with gallstones in men but not in women. CYP7A1 is an attractive candidate gene because it encodes the rate-limiting enzyme in the “neutral” pathway for hepatic bile salt synthesis (see Chapter 64) and because bile salts are essential for forming bile and for keeping cholesterol molecules solubilized in simple and mixed micelles in bile. Furthermore, Pullinger and colleagues found a link between another single gene defect of CYP7A1 and cholesterol gallstones associated with hypercholesterolemia resistant to HMG-CoA reductase inhibitors in two male homoyzgotes.141 Missense mutations in the ABC transporter B4 (ABCB4) gene (formerly named multidrug resistance gene 3, MDR3), which encodes the phosphatidylcholine transporter in the canalicular membrane of hepatocytes, are the basis for a particular type of cholelithiasis.66,142 The disorder is characterized by intrahepatic sludge, gallbladder cholesterol gallstones, mild chronic cholestasis, a high cholesterol-tophospholipid ratio in bile, and recurrent symptoms after cholecystectomy.143-145 A defect in the ABC transporter B4 gene could constitute the basis for this highly symptomatic and recurrent form of gallstone disease. In patients with hepatolithiasis, a common disease in Asia (see Chapter 68), low expression levels of ABCB4 and phosphatidylcholine transfer protein occur together, with markedly reduced phospholipid concentrations in bile.146 Furthermore, HMG-CoA reductase activity is increased and CYP7A1 activity is reduced in patients with gallstones compared with controls. In this disorder, the formation of cholesterolrich intrahepatic stones could be induced by decreased biliary secretion of phospholipids in the setting of increased cholesterol synthesis and decreased bile salt synthesis. Because hypomotility of the gallbladder favors gallstone formation, the genes for CCK and the CCK-1 receptor (CCK1R), which regulate gallbladder motility, are attractive candidates.115,148 Genetic variation in CCK-1R is associated with gallstone risk, and an aberrant splicing of CCK-1R, which is predicted to result in a nonfunctional receptor, is found in a few obese patients with gallstones.148,149 A search for mutations or polymorphisms in the CCK-1R gene in patients with gallstones has been unsuccessful, however.150 Some studies have reported that certain polymorphisms of the apolipoprotein (APO) E and APOB genes and the cholesteryl ester transfer protein, all of which are involved in carrying cholesterol in the plasma, are associated with gallstones. The APOE polymorphisms are the most extensively studied polymorphisms in patients with gallstones, but reports concerning the protective role of the ε4 allele against gallstones have been inconsistent.151-155 The ε2 allele appears to protect against gallstones, and the degree of dietary cholesterol absorption in the intestine varies with the APOE isoform (ε4>ε3>ε2). Also, the fecal excretion of cholesterol tends to be higher in persons with the APOE2 phenotype than in those with the APOE3 or APOE4 phenotypes.156 In a study of polymorphisms at the APOB, APOAI, and cholesteryl ester transfer protein gene loci in patients with gallbladder disease, a polymorphism of the cholesteryl ester transfer protein gene, in relation to another HDL lowering factor, was found to be associated with cholesterol gallstones.157 Also, a link was found between the X+ allele of the APOB gene and an increased risk of cholesterol gall-

stones.158 More recently, a genome-wide association study in a large cohort of patients with gallstones from Germany159 and a linkage study in affected sibling pairs160 identified a common variant (D19H) of the sterol transporters ABCG5 and ABCG8 on the canalicular membrane of hepatocytes as a risk factor for gallstones. This variant is also a susceptibility factor for gallstones in Chilean Hispanics,159 and other ABCG8 variants (T400K, D19H, A632V, M429V, and C54Y) as well as ABCG5 variants (Q604E) may be important risk factors for gallstone formation in Chinese and Canadian white populations.161-163 Table 65-1 summarizes the major classes of candidate genes for cholesterol gallstones (or Lith genes).23 Some candidate genes have not yet been identified in humans, and their roles in cholelithogenesis need to be investigated further. In general, genes that contribute to cholesterol gallstone formation include those that encode (1) hepatic and intestinal membrane lipid transporters; (2) hepatic and intestinal lipid regulatory enzymes; (3) hepatic and intestinal intracellular lipid transporters; (4) hepatic and intestinal lipid regulatory transcription factors; (5) hepatic lipoprotein receptors and related proteins; (6) hormone receptors in the gallbladder; and (7) biliary mucins. Human analogs of these genes could be involved in human gallstone disease. Changes in the expression and function of one of several ABC transporters in the canalicular membrane may result in an alteration of bile compositions and may influence gallstone formation. In addition, mutations in genes that encode several lipoprotein receptors and related proteins that determine the uptake of HDL and LDL and in several intracellular proteins that transport biliary lipids through the cytosol of hepatocytes, as well as regulatory transcription factors that regulate hepatic cholesterol and bile salt metabolism and biliary lipid secretion, may cause the formation of cholesterol gallstones. Mutations in genes that affect CCK, the CCK-receptor, and the secretion and properties of mucin may also play a role in the pathogenesis of gallstones. The factors that regulate intestinal membrane lipid transporters, lipid regulatory enzymes, intracellular lipid transporters, and lipid regulatory transcription factors may influence the amount of cholesterol of intestinal origin contributing to the liver for biliary secretion. Direct evidence for the role of intestinal factors in mouse gallstones comes from a study of ACAT gene 2-knockout mice in which the lack of cholesteryl ester synthesis in the intestine significantly reduces intestinal cholesterol absorption and causes complete resistance to diet-induced cholesterol gallstones.164 One study found that the potent cholesterol absorption inhibitor ezetimibe prevents gallstones by effectively reducing intestinal cholesterol absorption and biliary cholesterol secretion and protects gallbladder motility function by desaturating bile in mice.165 Therefore, reduced cholesterol absorption or hepatic chylomicron remnant uptake may induce a decrease in biliary cholesterol secretion and saturation.

PIGMENT STONES Although the pathogeneses of black and brown pigment gallstones are not as well understood as that of cholesterol gallstones and each type of stone probably has a distinctive pathogenesis, both types of pigment stone result from abnormalities in the metabolism of bilirubin and are pigmented as a result of bilirubin precipitation.166-168 In general, the bile of patients with both types of pigment stones

Multiple ABCG8 p.D19H (rs1188753) p.R64W (rs4944) -75G>A, RFLP

ABCB11 ABCG5/G8 ARDB3 APOA1 APOB APOC1 AR CCK1R CETP CYP7A1 ESR2 LRPAP1

Apolipoprotein B

Apolipoprotein C1 Androgen receptor Cholecystokinin 1 receptor Cholesterol ester transfer protein

Cytochrome P450 7A1 Estrogen receptor 2 LRP-associated protein 1

+ (China) + (Greece) + (India)

− − − − − − − − − − −

− − − + − − + − + − −

+ (India) + (Greece) + (India) + (Finland)

−

+

+ (China, Poland)

+ (Germany Serbia, Chile, China) + (Germany) + (China, India)

−

−

+

−

Common polygenic

Familial oligogenic

Rare monogenic

Gallbladder hypomotility ↑ Biliary cholesterol secretion secondary to ↑ reverse cholesterol transport ↑ Biliary cholesterol secretion secondary to hepatic VLDL synthesis and ↑ intestinal cholesterol absorption ↑ APOC1 remnant-like particle cholesterol Gallbladder hypomotility Gallbladder and small intestinal hypomotility ↑ Hepatic cholesterol uptake from HDL catabolism ↑ Bile salt synthesis ↑ Cholesterol synthesis ↑ Hepatic cholesterol uptake from chylomicron remnants via LRP

↑ Biliary cholesterol secretion

↓ Biliary bile salt secretion

↓ Biliary phospholipid secretion

POTENTIAL MECHANISMS

*Countries where inheritance patterns were reported are shown in parentheses. HDL, high density lipoprotein; LRP, low-density lipoprotein receptor-related protein; OMIM, Online Mendelian Inheritance in Man; RFLP, restriction fragment length polymorphism; rs, restriction site; SNP, single nucleotide polymorphism; VLDL, very-low-density lipoprotein. Reproduced with slight modifications and with permission from Lammert F, Sauerbruch T. Pathogenesis of gallstone formation: Updated inventory of human lithogenic genes. In: Carey MC, Dité P, Gabryelewicz A, et al, editors. Future Perspectives in Gastroenterology (Falk Symposium 161). Dordrecht, Germany: Springer; 2008. pp 99-107.

Promoter SNP -204A>C c.1092+3607(CA)n Intron 5 insertion/deletion (rs11267919)

RFLP c.172(CAG)n RFLP RFLP

c.2488C>T, c.4154G>A

Multiple

ABCB4

ATP binding cassette transporter B4 ATP binding cassette transporter B11 ATP binding cassette transporters G5/G8 β3 Adrenergic receptor Apolipoprotein A1

GENE VARIANTS

GENE SYMBOL

PROTEIN

Inheritance Pattern*

Table 65-1  Human Cholesterol Gallstone (LITH) Genes and Gene Products That Have Been Identified as of 2008

Chapter 65  Gallstone Disease 1103

1104

Section VIII  Biliary Tract contains an excess of unconjugated bilirubin, analogous to the saturation of bile with cholesterol in patients with cholesterol stones.169 Also, both types of pigment stones are composed primarily of bile pigment and contain a matrix of mucin glycoproteins. In black stones, however, the pigment is predominantly an insoluble, highly cross-linked polymer of calcium bilirubinate, whereas in brown stones, the main pigment is monomeric calcium bilirubinate. The two types of pigment stones also differ in radiodensity, location within the biliary system, and geographic distribution.

BLACK PIGMENT STONES

Black pigment stones are formed in uninfected gallbladders, particularly in patients with chronic hemolytic anemia (e.g., β-thalassemia, hereditary spherocytosis, and sickle cell disease) and liver cirrhosis. The unconjugated bilirubin produced in increased amounts precipitates as calcium bilirubinate to form stones.170 This type of stone is composed of either pure calcium bilirubinate or polymer-like complexes consisting of unconjugated bilirubin, calcium bilirubinate, calcium, and copper. Mucin glycoproteins account for as much as 20% of the weight of black stones. A regular crystalline structure is not present. Under normal physiologic conditions, unconjugated bilirubin is not secreted into bile. Bilirubin glucuronides are hydrolyzed by endogenous β-glucuronidase, and unconjugated bilirubin constitutes less than 1% of total bile pigment, mostly because the activity of the enzyme is inhibited by β-glucaro-1,4-lactone in the biliary system.171,172 The unifying predisposing factor in the formation of black pigment stones is the hypersecretion of bilirubin conjugates (especially monoglucuronides) into bile. In the presence of hemolysis, secretion of these bilirubin conjugates increases ten-fold. Unconjugated monohydrogenated bilirubin is formed by the action of endogenous β-glucuronidase, which coprecipitates with calcium as a result of supersaturation. A 1% hydrolysis rate could give rise to high concentrations of unconjugated bilirubin that often greatly exceed the solubility of bilirubin in bile. A defect in acidification of bile also may be induced by gallbladder inflammation or the reduced buffering capacity of sialic acid and sulfate moieties in the mucin gel. The reduction in buffering capacity facilitates the supersaturation of calcium carbonate and phosphate that would not occur at a more acidic pH. Gallbladder motility defects are not observed in patients with black pigment stones, as inferred from in vitro experiments of human gallbladder muscles.

BROWN PIGMENT STONES

Brown pigment stones are composed mainly of calcium salts of unconjugated bilirubin, with varying amounts of cholesterol, fatty acids, pigment fraction, and mucin glycoproteins, as well as small amounts of bile salts, phospholipids, and residues. Brown pigment stones may be easily distinguished grossly from black pigment stones by their reddish brown to dark brown color and lack of brightness. Their shape is irregular or molded and occasionally spherical. Most of the stones are muddy in consistency, and some show facet formation. Brown pigment stones are either smooth or rough without any surface luster and are soft, fragile, and light in comparison with other gallstones. The cut surface is generally a stratified structure (lamellation) or is amorphous without the radiating crystalline structure seen in cholesterol stones. Almost invariably, brown pigment stones have a lamellated cross-sectional surface with calcium bilirubinate-rich layers alternating with calcium palmitate-rich layers. Brown pigment stones are formed not only in the gallbladder, but also commonly in other portions of the biliary

tree, especially in intrahepatic bile ducts. The formation of brown pigment stones requires the presence of structural or functional stasis of bile associated with biliary infection, especially with Escherichia coli.173 These stones are more common in areas such as Asia, where Clonorchis sinensis and roundworm infestations are prevalent, and parasitic elements have been considered to be kernels of brown pigment stone formation (see Chapters 68 and 82).174 Bile stasis predisposes to the bacterial infection as well as the accumulation of mucins and bacterial cytoskeletons in the bile ducts. Bile stasis may be caused by bile duct stenosis and bacterial infection caused by infestation by parasites such as Clonorchis sinensis, roundworms, and their ova.175 Additionally, bacterial infection and colonization in bile ducts by enteric bacteria are found commonly in patients with brown pigment stones. As the incidence of biliary infections has decreased in Asian populations prone to development of brown pigment stones, the ratio of cholesterol stones to pigment stones also has changed in these populations. The percentage of brown pigment stones in Japan has fallen from 60% to 24% since the 1950s, and similar changes have been reported from other Asian countries.176-178 Enteric bacteria produce β-glucuronidase, phospholipase A1, and conjugated bile acid hydrolase. Activity of β-glucuronidase results in the production of unconjugated bilirubin from bilirubin glucuronide; phospholipase A1 liberates palmitic and stearic acids from phospholipids; and bile acid hydrolases produce unconjugated bile salts from glycine or taurine-conjugated bile salts. Partially ionized saturated fatty acids, unconjugated bilirubin, and unconjugated bile salts may precipitate as calcium salts. Mucin gel can trap these complex precipitates and facilitate their growth into macroscopic stones. Figure 65-6 shows the postulated mechanisms underlying the formation of brown pigment stones. Under normal physiologic conditions, bilirubin in bile exists mainly as bilirubin glucuronide, which is soluble in aqueous media. Bile also contains β-glucuronidase of tissue origin, the activity of which is inhibited by glucaro-1,4-lactone, which is also formed in the liver. If infection with E. coli occurs, the concentration of bacterial β-glucuronidase increases significantly and exceeds the inhibitory power of glucaro-1,4-lactone. As a result, bilirubin glucuronide is hydrolyzed to produce unconjugated bilirubin and glucuronic acid; the former is water-insoluble and combines with calcium to form calcium bilirubin at its carboxyl radical, leading to the formation of brown pigment gallstones.

NATURAL HISTORY The natural history of gallstones typically is described in two separate groups of patients: those who have symptoms and those who are asymptomatic. Necropsy studies clearly show that the vast majority of patients with gallstones are asymptomatic and remain so. Ascertaining the true frequency of complications in persons with asymptomatic stones (as well as those with symptomatic stones) is critical to providing rational, cost-effective recommendations regarding therapy (see later). Unfortunately, the information available on the natural history of gallstones has been sparse and somewhat varied.179-181

ASYMPTOMATIC STONES

The study that changed our understanding of the course and appropriate therapy of gallstone disease was performed by Gracie and Ransohoff.179 They monitored 123 University of

Chapter 65  Gallstone Disease Phospholipids

Bilirubin glucuronides

Conjugated bile salts

b-Glucaro–1,4– lactone Bacterial b-glucuronidase Glucuronic acid

(–)

Phospholipase A1

Bile salt hydrolase

Endogenous b-glucuronidase

Free bilirubin

Free fatty acids

Free bile salts

Brown pigment stones

Calcium

Dead bacteria and parasites

Michigan faculty members for 15 years after they had been found to have gallstones on routine screening ultrasonography. At 5, 10, and 15 years of follow-up, 10%, 15%, and 18% of the patients, respectively, had become symptomatic, and none had experienced serious complications. The investigators suggested that the rate at which biliary pain develops in persons with asymptomatic gallstones is about 2% per year for five years and then decreases over time. Biliary complications developed in only three patients in this study, and all complications were preceded by episodes of biliary pain. Studies have suggested that biliary pain, not a biliary complication, is the initial manifesting symptom in 90% of people with previously asymptomatic gallstones.179 Therefore, in patients with asymptomatic stones, the frequency of complications is low, and prophylactic cholecystectomy is not necessary. Subsequent studies have reported slightly higher rates of biliary pain and complications in patients with initially asymptomatic gallstones,180 but only one was a long-term and prospective study.181 The Group for Epidemiology and Prevention of Cholelithiasis (GREPCO) in Rome reported the courses of 151 subjects with gallstones, 118 of whom were asymptomatic on entering the study. In those who were initially asymptomatic, the frequency of biliary pain was 12% at 2 years, 17% at 4 years, and 26% at 10 years, and the cumulative rate of biliary complication was 3% at 10 years.181

Mucin gel

Figure 65-6.  Proposed mechanisms for the pathogenesis of brown pigment stones. Under normal physiologic conditions, unconjugated bilirubin is not secreted into bile. Although modest hydrolysis of bilirubin glucuronides by endogenous b-glucuronidase occurs, unconjugated bilirubin constitutes less than 1% of total bile pigment, mostly because the activity of b-glucuronidase is inhibited by b-glucaro-1,4-lactone in the biliary system. The presence of excess bacterial β-glucuronidase, however, overcomes the inhibitory (-) effect of b-glucaro-1,4-lactone, which results in hydrolysis of bilirubin glucuronide into free bilirubin and glucuronic acid. Free bilirubinate combines with calcium to yield water-insoluble calcium bilirubinate. In addition, phospholipase A1 liberates free fatty acids such as palmitic acid and stearic acid from phospholipids, and bile salt hydrolases produce unconjugated bile salts from glycine or taurine-conjugated bile salts. Dead bacteria and parasites could act as nuclei that accelerate the precipitation of calcium bilirubinate. The mucin gel in the gallbladder can trap these complex precipitates and facilitate their growth into macroscopic stones.

fore, cholecystectomy should be offered to patients only after biliary symptoms develop. Depending on the patient, a reasonable alternative approach may be to observe the pattern of pain before deciding on therapy because up to 30% of patients with one episode of biliary pain do not have a recurrent episode. This approach is particularly useful in patients with a high operative risk.

STONES IN PATIENTS WITH DIABETES MELLITUS

Diabetic patients with incidental cholelithiasis were long considered to have an increased risk of serious complications even when the gallstones were asymptomatic. Subsequent studies have shown that the natural history of gallstones in diabetic patients follows the same pattern observed in nondiabetic persons. A prospective study of patients with insulin-resistant diabetes mellitus showed that after five years of follow-up, symptoms had developed in 15% of the asymptomatic patients.185 This frequency is roughly the same as that reported for nondiabetic patients. Moreover, the complication and mortality rates were comparable to those in studies of nondiabetic patients with gallstones. Therefore, prophylactic cholecystectomy is generally not recommended in patients with insulin-resistant diabetes mellitus and asymptomatic gallstones.

SYMPTOMATIC STONES

The natural history of symptomatic gallstones has a more aggressive course than that of asymptomatic stones. The U.S. National Cooperative Gallstone Study showed that in persons who had an episode of uncomplicated biliary pain in the year before entering the study, the rate of recurrent biliary pain was 38% per year.182 Other investigators have reported a rate of recurrent biliary pain as high as 50% per year in persons with symptomatic gallstones.183 As noted earlier, biliary complications also are more likely to develop in persons with symptomatic gallstones. The risk of biliary complications is estimated to be 1% to 2% per year and is believed to remain relatively constant over time.184 There-

DIAGNOSIS AND CLINICAL DISODERS The clinical manifestations of gallstones are shown schematically in Figure 65-7 and summarized in more detail in Table 65-2.186-190 Biliary pancreatitis is discussed in Chapter 58. Although the standard approach to asymptomatic gallstones is observation, some patients with asymptomatic gallstones may be at increased risk of complications and may require special consideration. An increased risk of cholangiocarcinoma and gallbladder carcinoma has been associated with certain disorders of

1105

1106

Section VIII  Biliary Tract Stone intermittently obstructing cystic duct, causing intermittent biliary pain (20%)

Stone impacted in cystic duct, causing acute cholecystitis (10%) 3

2

4

Asymptomatic stone (75%)

** * *

1

*

7

*

Long-standing cholelithiasis, resulting in gallbladder carcinoma (<0.1%) Figure 65-7.  Schematic depiction of the natural history and complications of gallstones. The percentages indicate the approximate frequencies of complications that occur in untreated patients, based on natural history data. The most frequent outcome is for the patient with a stone to remain asymptomatic throughout life (1). Biliary pain (2), acute cholecystitis (3), cholangitis (5), and pan­ creatitis (5) are the most common complications. Mirizzi’s syndrome (4), cholecystoenteric fistula (6), Bouveret’s syndrome (6), and gallbladder cancer (7) are relatively rare. (The sum of the percentages is >100% because patients with acute cholecystitis generally have had prior episodes of biliary pain.)

Stone in the cystic duct compressing or fistulizing into the bile duct, causing Mirizzi’s syndrome (<0.1%)

6

*

Stone eroding through gallbladder into duodenum, resulting in cholecystoenteric fistula (prerequisite for gallstone ileus) and leading in some cases to Bouveret’s syndrome (gastric 5 outlet obstruction) (<0.1%) Stone impacted in distal bile duct, causing jaundice, biliary-type pain, and risk of ascending cholangitis or acute biliary pancreatitis (5%)

the biliary tree and in some ethnic groups (such as Native Americans) (see Chapter 69). Risk factors include choledochal cysts, Caroli’s disease, anomalous pancreatic ductal drainage (in which the pancreatic duct drains into the bile duct), large gall­bladder adenomas, and porcelain gallbladder (see Chapters 62 and 67). Patients at increased risk of biliary cancer may benefit from prophy­lactic cholecystectomy. In particular, if abdominal surgery is planned for another indication, an incidental cholecystectomy should be performed. Pigment gallstones are common and often asymptomatic in patients with sickle cell disease. Prophylactic cholecystectomy is not recommended, but an incidental chole­ cystectomy should be considered if abdominal surgery is performed for other reasons. Some authorities recommend combined prophylactic splenectomy and cholecystectomy in young asymptomatic patients with hereditary spherocytosis if gallstones are present. Morbidly obese persons who undergo bariatric surgery are at high risk of complications of gallstones (see Chapters 6 and 7). These patients have a frequency of gallstones of greater than 30%. An incidental cholecystectomy is recommended at the time of surgery. Some investigators have proposed that patients with incidental cholelithiasis who are awaiting heart transplantation undergo a prophylactic cholecystectomy irrespective of the presence or absence of biliary tract symptoms because they are at increased risk of post-transplant gallstone complications.191 A retrospective study that addressed this issue in renal transplant recipients, however, concluded that complications of gallstones could be managed safely after symptoms emerged.192 Imaging studies play a central role in the diagnosis of gallstones and associated conditions. An understanding of the usefulness and limitations of each available imaging technique is essential to ensure efficient and cost-effective

evaluation of patients who present with symptoms related to gallstones and gallbladder disease.

IMAGING STUDIES

As shown in Table 65-3, a wide array of imaging techniques are available to evaluate the biliary tract.193-196 Each modality has its strengths and limitations, and the methods vary widely in relative cost and risk to the patient. With the possible exception of ultrasonography, none of the modalities should be ordered routinely in the evaluation of a patient with suspected gallstone disease; rather, the diagnostic evaluation should proceed in a rational, stepwise fashion based on the individual patient’s symptoms, signs, and results of laboratory studies. Notably absent from the list of imaging studies of the biliary tract is the plain abdominal film. Although useful on occasion for evaluating patients with abdominal pain, plain abdominal films are limited by a lack of sensitivity and specificity. Only 50% of pigment stones and 20% of cholesterol stones contain enough calcium to be visible on a plain abdominal film. Because 80% of gallstones in the Western world are of the cholesterol type, only 25% of stones can be detected by simple radiographs. Plain abdominal films have their greatest usefulness in evaluating patients with some of the unusual complications of gallstones, such as emphysematous cholecystitis, cholecystenteric fistula, and gallstone ileus, or in detecting a porcelain gallbladder (see later).

Ultrasonography

Since its introduction in the 1970s, ultrasonographic examination of the biliary tract has become the principal imaging modality for the diagnosis of cholelithiasis. Ultrasono­graphy requires only an overnight or eight-hour fast, involves no ionizing radiation, is simple to perform, and provides accurate anatomic information. It has the

Table 65-2  Common Clinical Manifestations of Gallstone Disease* BILIARY PAIN

ACUTE CHOLECYSTITIS

CHOLEDOCHOLITHIASIS

CHOLANGITIS

Pathophysiology

Intermittent obstruction of the cystic duct No acute inflammation of the gallbladder

Intermittent obstruction of the BD

Stone in the BD causing bile stasis Bacterial superinfection of stagnant bile Early bacteremia

Symptoms

Severe, poorly localized, epigastric or RUQ visceral pain growing in intensity over 15 minutes and remaining constant for 1-6 hours, often with nausea Frequency of attacks varies from days to months Gas, bloating, flatulence, and dyspepsia are not related to stones Mild to moderate epigastric/RUQ tenderness during an attack with mild residual tenderness lasting days Often findings are normal

Impacted stone in the cystic duct Acute inflammation of the gallbladder Secondary bacterial infection in ≈50% 75% of cases are preceded by attacks of biliary pain Visceral epigastric pain gives way to moderately severe, localized pain in the RUQ, back, right shoulder, or, rarely, chest Nausea with some vomiting is frequent Pain lasting >6 hours favors cholecystitis over biliary pain alone Fever, but usually to <102°F unless complicated by gangrene or perforation Right subcostal tenderness with inspiratory arrest (Murphy’s sign) Palpable gallbladder in 33% of patients, especially those having their first attack Mild jaundice in 20%; higher frequency in elderly Leukocytosis with band forms is common Serum bilirubin level may be 2-4 mg/dL and aminotransferase and alkaline phosphatase levels may be elevated even in absence of BD stone or hepatic infection Mild serum amylase and lipase elevations are seen even in absence of pancreatitis If serum bilirubin is >4 mg/ dL or amylase or lipase is markedly elevated, a BD stone should be suspected Ultrasonography Hepatobiliary scintigraphy Abdominal computed tomography

Often asymptomatic Symptoms (when present) are indistinguishable from biliary pain Predisposes to cholangitis and acute pancreatitis

Charcot’s triad (pain, jaundice, and fever) is present in 70% of patients Pain may be mild and transient and is often accompanied by chills Mental confusion, lethargy, and delirium suggest sepsis

Often findings are completely normal if the obstruction is intermittent Jaundice with pain suggests stones; painless jaundice and a palpable gallbladder favor malignancy

Fever in 95% RUQ tenderness in 90% Jaundice in 80% Peritoneal signs in 15% Hypotension and mental confusion coexist in 15% and suggest gram-negative sepsis

Elevated serum bilirubin and alkaline phosphatase levels are seen with BD obstruction Serum bilirubin level >10 mg/ dL suggests malignant obstruction or coexisting hemolysis A transient “spike” in serum aminotransferase or amylase (or lipase) levels suggests the passage of a stone

Leukocytosis in 80%, but remainder may have normal white blood cell count with or without band forms Serum bilirubin level >2 mg/ dL in 80% Serum alkaline phosphatase level is usually elevated Blood cultures are usually positive, especially during chills or fever spike; two organisms are grown in cultures from one half of patients

ERCP Endoscopic ultrasonography Magnetic resonance cholangiography Percutaneous THC

ERCP Percutaneous THC

Natural history is not well defined, but complications are more common and more severe than for asymptomatic stones in the gallbladder

High mortality rate if unrecognized, with death from septicemia Emergency decompression of the BD (usually by ERCP) improves survival dramatically

Stone removal at the time of ERCP followed in most cases by early laparoscopic cholecystectomy

Emergency ERCP with stone removal or at least biliary decompression Antibiotics to cover gramnegative and possibly anaerobic organisms and Enterococcus spp. Subsequent cholecystectomy

Physical findings

Laboratory findings

Usually normal Elevated serum bilirubin, alkaline phosphatase, or amylase levels suggest coexisting BD stones

Diagnostic studies (see Table 65-3 for details on imaging studies) Natural history

Ultrasonography Oral cholecystography Meltzer-Lyon test (see Chapter 67)

Treatment (see Chapters 66 and 70)

After the initial attack, 30% of patients have no further symptoms Symptoms develop in the remainder at a rate of 6% per year, and severe complications at a rate of 1% to 2% per year Elective laparoscopic cholecystectomy possibly with IOC ERCP for stone removal or BD exploration if IOC shows stones

50% of cases resolve spontaneously in 7-10 days without surgery Left untreated, 10% of cases are complicated by a localized perforation and 1% by a free perforation and peritonitis Laparoscopic cholecystectomy possibly with IOC if feasible; otherwise open cholecystectomy BD exploration or ERCP for stone removal if IOC shows stones

BD, bile duct; ERCP, endoscopic retrograde cholangiopancreatography; IOC, intraoperative cholangiography; MRC, magnetic resonance cholangiography; RUQ, right upper quadrant; THC, transhepatic cholangiography. *See Chapter 58 for a discussion of biliary pancreatitis.

1108

Section VIII  Biliary Tract Table 65-3  Imaging Studies of the Biliary Tract TECHNIQUE

CONDITION TESTED FOR

Ultrasonography

Cholelithiasis

Choledocholithiasis Acute cholecystitis

EUS

Choledocholithiasis

Oral cholecystography*

Cholelithiasis

Cholescintigraphy (hepatobiliary scintigraphy; hydroxyiminodiacetic acid or diisopropyl iminodiacetic acid scan)

Acute cholecystitis

ERCP

Choledocholithiasis

Cholelithiasis MRCP

Choledocholithiasis

CT

Complications of gallstones

FINDINGS/COMMENTS Stones manifest as mobile, dependent echogenic foci within the gallbladder lumen with acoustic shadowing Sludge appears as layering echogenic material without shadows Sensitivity rate >95% for stones >2 mm Specificity rate >95% for stones with acoustic shadows Rarely, a stone-filled gallbladder may be contracted and difficult to see, with a “wall-echoshadow” sign Best single test for stones in the gallbladder Stones are seen in BD in only ≈50% of cases but can be inferred from the finding of a dilated BD (>6 mm diameter), with or without gallstones, in another ≈25% of cases Can confirm, but not exclude, BD stones Ultrasonographic Murphy’s sign (focal gallbladder tenderness under the transducer) has a positive predictive value of >90% in detecting acute cholecystitis when stones are seen Pericholecystic fluid (in the absence of ascites) and gallbladder wall thickening to >4 mm (in the absence of hypoalbuminemia) are nonspecific findings but are suggestive of acute cholecystitis Highly accurate for excluding or confirming stones in the BD Concordance of EUS with the ERCP diagnosis ≈95%; many studies suggest slightly higher sensitivity rates for EUS than for ERCP Specificity rate ≈97% Positive predictive value ≈99%, negative predictive value ≈98%, accuracy rate ≈97% With experienced operators, EUS can be used in lieu of ERCP to exclude BD stones, particularly when the clinical suspicion is low or intermediate Considered for patients with a low-to-moderate clinical probability of choledocholithiasis Stones manifest as mobile filling defects in an opacified gallbladder Sensitivity and specificity rates exceed 90% when the gallbladder is opacified, but nonvisualization occurs in 25% of studies and can result from multiple causes other than stones Opacification of the gallbladder indicates patency of the cystic duct May be useful in the evaluation of acalculous gallbladder diseases such as cholesterolosis and adenomyomatosis (see Chapter 67) Assesses patency of the cystic duct Normal scan shows radioactivity in the gallbladder, BD, and small bowel within 30-60 minutes Positive result is defined as nonvisualization of the gallbladder with preserved hepatic excretion of radionuclide into the BD or small bowel Sensitivity rate is ≈95% and specificity rate is ≈90%, with false-positive results seen in fasted, critically ill patients With cholecystokinin stimulation, gallbladder “ejection fraction” can be determined and may help evaluate patients with acalculous biliary pain (see Chapter 67) Normal scan result virtually excludes acute cholecystitis ERCP is the standard diagnostic test for stones in the BD, with sensitivity and specificity rates of ≈95% Use of ERCP to extract stones (or at least to drain infected bile) is life-saving in severe cholangitis and reduces the need for BD exploration at the time of cholecystectomy Recommended for patients with a high clinical probability of choledocholithiasis When contrast agent flows retrograde into the gallbladder, stones appear as filling defects and can be detected with a sensitivity rate of ≈80%, but ultrasonography remains the mainstay for confirming cholelithiasis Rapid, noninvasive modality that provides detailed bile duct and pancreatic duct images equal to those of ERCP Sensitivity rate ≈93% and specificity rate ≈94%, comparable with those for ERCP Useful for examining nondilated ducts, particularly at the distal portion, which often is not well visualized by ultrasonography Adjacent structures such as liver and pancreas can be examined at the same time Recommended for patients with a low-to-moderate clinical probability of choledocholithiasis Not well suited for detecting uncomplicated stones, but excellent for detecting complications, such as abscess, perforation of the gallbladder or BD, and pancreatitis Spiral CT may prove useful as a noninvasive means of excluding BD stones; some studies suggest improved diagnostic accuracy when CT is combined with an oral cholecystographic contrast agent

BD, bile duct; CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasonography; MRCP, magnetic resonance cholangiopancreatography. *Performed infrequently nowadays.

Chapter 65  Gallstone Disease

GB

Stone

A

Acoustic shadowing

GB

Stones

B

Acoustic shadowing

Figure 65-8.  A, Typical ultrasonographic appearance of cholelithiasis. A gallstone is present within the lumen of the gallbladder (GB), casting an acoustic shadow. With repositioning of the patient, the stones will move, thereby excluding the possibility of a gallbladder polyp. B, Cholelithiasis in the setting of acute cholecystitis. Multiple gallstones can be seen within the gallbladder lumen with associated acoustic shadowing. In addition, the gallbladder wall is thickened (arrowheads). (Courtesy of Julie Champine, MD, Dallas, Tex.)

additional advantage of being portable and thus available at the bedside of a critically ill patient.194 The diagnosis of gallstones relies on the detection of echogenic objects within the lumen of the gallbladder that produce an acoustic shadow (Fig. 65-8A). The stones are mobile and generally congregate in the dependent portion of the gallbladder. Modern ultrasonography is able to detect stones as small as 2 mm in diameter routinely. Smaller stones may be missed or may be confused with biliary sludge (layering echogenic material that does not cast acoustic shadows).197 The sensitivity of ultrasonography for the detection of gallstones in the gallbladder is more than 95% for stones larger than 2 mm.198 The specificity is greater than 95% when stones produce acoustic shadows. Rarely, advanced scarring and contraction of the gallbladder around gallstones make it impossible to locate the gallbladder or the stones; this finding should also raise the possibility of gallbladder cancer. The contracted gallbladder filled with stones may give a “double-arc shadow” or “wall-echo shadow” sign, with the gallbladder wall, echogenic stones, and acoustic shadowing seen in immediate proximity. If the

gallbladder cannot be identified ultrasonographically, then a complementary imaging modality such as oral cholecystography or abdominal computed tomography (CT) is warranted. Ultrasonography is the standard for the diagnosis of stones in the gallbladder but is distinctly less sensitive for the detection of stones in the bile duct (previously termed common bile duct).199 Because of the proximity of the distal bile duct to the duodenum, luminal bowel gas often interferes with the ultrasonographic image, and the entire length of the bile duct cannot be examined.200 As a result, only approximately 50% of bile duct stones are actually seen on ultrasonography.194 The presence of an obstructing bile duct stone, however, can be inferred when a dilated duct is found. Now that endoscopic retrograde cholangiopancreatography (ERCP) has uncovered a rising frequency of falsely negative ultrasonograms, the upper limit of normal of the diameter of the bile duct has declined from 10 mm to 6 mm. Even so, inferring choledocholithiasis from a dilated bile duct on ultrasonography has a sensitivity of only 75%. Finally, ultrasonography is quite useful for diagnosing acute cholecystitis.201 Pericholecystic fluid (in the absence of ascites) and gallbladder wall thickening to more than 4 mm (in the absence of hypoalbuminemia) are suggestive of acute cholecystitis (see Fig. 65-8B). Unfortunately, in the critical care setting, these nonspecific findings are seen frequently in patients with no other evidence of gallbladder disease.201 A more specific finding is the so-called sonographic Murphy’s sign, in which the ultrasonographer elicits focal gallbladder tenderness under the ultrasound transducer. Eliciting a sonographic Murphy’s sign is somewhat operator dependent and requires an alert patient. Presence of the sign has a positive predictive value of greater than 90% for detecting acute cholecystitis if gallstones are present.202 Because it provides accurate anatomic localization of biliary tract abnormalities, ultrasonography may help localize other abdominal diseases, such as abscesses or pseudocysts, that may be in the differential diagnosis.

Endoscopic Ultrasonography

Endoscopic ultrasonography (EUS) is highly accurate for detecting choledocholithiasis. Inherently more invasive and more expensive than standard ultrasonography, EUS has the advantage of being able to visualize the bile duct from within the gastrointestinal lumen and is reported to be comparable to ERCP in this respect. Intraluminal imaging provides several advantages over transabdominal ultrasonography, including closer proximity to the bile duct, higher resolution, and lack of interference by bowel gas or abdo­minal wall layers (Fig. 65-9). In several studies, EUS had a positive predictive value of 99%, a negative predictive value of 98%, and an accuracy rate of 97% for the diagnosis of bile duct stones compared with ERCP.203,204 If bile duct stones are found on EUS, endoscopic removal of the stones is necessary, and it can be argued that ERCP should be the initial study if choledocholithiasis is strongly suspected. Nonetheless, several studies that compared EUS with ERCP have found both techniques to be accurate for confirming or excluding choledocholithiasis, with EUS having advantages in both safety and cost.205-207 EUS also has been found to be superior to magnetic resonance cholangiopancreatography (MRCP) (or simply magnetic resonance cholangiography [MRC]) in detecting the presence or absence of bile duct stones (see later). The major benefit of EUS in patients with a clinical suspicion of choledo­cholithiasis is the ability to avoid unnecessary ERCP and sphincterotomy, which is not without risk. There-

1109

1110

Section VIII  Biliary Tract

GB

5 min.

10 min.

15 min.

20 min.

25 min.

30 min.

BD

BD PD PV

Distal BD Conf

Stone

Figure 65-9.  Endoscopic ultrasonography, with a radial sector scanning endoscope, demonstrating choledocholithiasis. The bile duct (BD) is shown extending to the level of the gallbladder (GB) (top) and distally (middle and bottom pictures). The greatest diameter of the BD is 12 mm (middle), and the duct tapers distally to a diameter of 7 mm (bottom). Within the distal BD a gallstone is clearly visualized (bottom). Note the proximity of adjacent structures to the bile duct and the ease with which these structures are resolved by endoscopic ultrasonography. Conf, confluence of the portal and splenic veins; PD, pancreatic duct; PV, portal vein.

fore, EUS is currently considered an appropriate modality for excluding bile duct stones, especially if the pretest probability of finding stones is low to intermediate.

Oral Cholecystography

Once the mainstay of imaging studies of the gallbladder, oral cholecystography (OCG) now has limited application as a secondary approach to identifying stones in the gall-

35 min. 45 min. 60 min. Figure 65-10.  Cholescintigraphy demonstrating an obstructed cystic duct characteristic of acute cholecystitis. The gamma-emitting radioisotope diisopropyl iminodiacetic acid (DISIDA) is injected intravenously, rapidly (at 5 min) taken up by the liver, and excreted into bile (at 20 min). Sequential images show the isotope quickly entering the duodenum (at 45 min) and passing distally in the small intestine without ever being concentrated in the gallbladder. The failure of the gallbladder to be visualized as a hot spot within 30 to 60 minutes constitutes a positive result and implies obstruction of the cystic duct.

bladder.194 The ease, reliability, and rapidity with which it can detect stones, along with the absence of ionizing radiation, have made ultrasonography the imaging study of choice. In unusual cases in which the gallbladder cannot be identified ultrasonographically (as when the gallbladder is contracted and full of stones),208 OCG is complementary to ultrasonography for demonstrating cholelithiasis. Additionally, when medical dissolution of stones or lithotripsy is being considered, visualization of the gallbladder on OCG excludes cystic duct obstruction (see Chapter 66).209 Because of the time required to complete the test (48 hours), OCG is not useful in patients with suspected acute cholecystitis or other complications of gallstone disease. On occasion, OCG may detect unsuspected disease of the gallbladder, such as adenomyomatosis or cholesterolosis (see Chapter 67).

Cholescintigraphy

Cholescintigraphy (hepatobiliary scintigraphy) is a radionuclide imaging test of the gallbladder and biliary tract that is most useful for evaluating patients with suspected acute cholecystitis. By demonstrating patency of the cystic duct, cholescintigraphy can exclude acute cholecystitis rapidly (within 90 minutes) from the differential diagnosis in a patient who presents with abdominal pain.210,211 The procedure can be performed on an emergency basis in a nonfasting patient after intravenous administration of gamma-emitting 99mTc-labeled hydroxyl iminodiacetic acid (HIDA) or diisopropyl iminodiacetic acid (DISIDA), which is taken up rapidly by the liver and secreted into bile. As shown in Figure 65-10, serial scans after injection normally should show radioactivity in the gallbladder, bile

Chapter 65  Gallstone Disease duct, and small intestine within 30 to 60 minutes.167 In the past, imaging of jaundiced patients with this technique was limited, but use of DISIDA may allow imaging of the biliary tree in a patient with a serum bilirubin value as high as 20 mg/dL. An abnormal or “positive” scan result is defined as nonvisualization of the gallbladder with preserved excretion into the bile duct or small intestine. The accuracy of the test for detecting acute cholecystitis is 92%, superior to that for ultrasonography. False-positive results occur primarily in fasting or critically ill patients, in whom gallbladder motility is decreased. The reduction in gallbladder motility leads to greater water resorption, which results in a gelatinous bile. In critically ill patients, cholestasis and hepatocyte dysfunction result in reduced clearance of radionuclide imaging agents. Although nonvisualization of the gallbladder because of cystic duct obstruction is the hallmark of acute cholecystitis, pericholecystic hepatic uptake of radionuclide is a useful secondary sign.212 In some patients (e.g., those with chronic cholecystitis, liver disease, or choledocholithiasis), imaging of the gallbladder on a radionuclide scan is delayed for several hours, and scanning must be repeated in four or more hours to confirm absence of acute cholecystitis. This delay in visualization of the gallbladder is problematic in the acutely ill patient but has largely been overcome with the administration of intravenous morphine sulfate to patients in whom the gallbladder fails to be visualized within 60 minutes. Morphine raises the pressure within the sphincter of Oddi, thereby leading to the preferential flow of bile into the gallbladder if the cystic duct is not obstructed. Another scan is obtained 30 minutes after injection of morphine, and if the gallbladder is visualized, cystic duct obstruction and, hence, acute cholecystitis, is excluded. The gallbladder may not be visualized in approximately half of critically ill patients even after injection of morphine, thereby leading to falsepositive cholescintigraphy results. Although primarily a tool for evaluating acutely ill patients with suspected acute cholecystitis, cholescintigraphy after administration of CCK may be useful in identi­ fying patients with chronic acalculous biliary pain who are likely to benefit from empirical cholecystectomy (see Chapter 67). An additional important role for cholescintigraphy is the noninvasive and clear detection of bile leakage from the cystic duct as a complication of cholecystectomy (see Chapter 66).213

Endoscopic Retrograde Cholangiopancreatography

Endoscopic retrograde cholangiopancreatography (ERCP) is one of the most effective modalities for detecting choledocholithiasis. The technique of this procedure is discussed in more detail in Chapter 70. Stones within the bile duct appear as filling defects and can be detected with a sensitivity of approximately 95% (Fig. 65-11). Care should be taken to avoid inadvertent injection of air into the biliary tract214 because bubbles may mimic gallstones. The specificity of ERCP for the detection of bile duct stones is approximately 95%. The therapeutic applications of ERCP have revolutionized the treatment of patients with choledocholithiasis215 and other bile duct disorders (see Chapter 70). Because ERCP is invasive and is associated with potential complications such as pancreatitis and postsphincterotomy bleeding, however, the role of ERCP has been challenged by other modalities that exclude choledocholithiasis more safely and accurately in patients in whom the clinical suspicion of choledocholithiasis is not high. As the use of EUS and MRC has increased, the role of ERCP in the diagnosis of choledocholithiasis has changed considerably. A National Institutes

Figure 65-11.  An endoscopic retrograde cholangiogram demonstrating dilatation of the bile duct to 15 mm with multiple filling defects representing choledocholithiasis (arrows). (Courtesy of Steve Burdick, MD, Dallas, Tex.)

GB

Stone

Figure 65-12.  Abdominal computed tomography demonstrating emphy­ sematous cholecystitis with associated cholelithiasis. Pockets of gas (yellow arrow), resulting from a secondary infection with gas-forming organisms, are present within the wall of the gallbladder (GB). (Courtesy of Julie Champine, MD, Dallas, Tex.)

of Health consensus conference has recommended the use of ERCP only when the clinical probability for choledocholithiasis is high (i.e., when the need for therapeutic intervention is likely). For diagnosis of choledocholithiasis alone, EUS and MRC are equal in accuracy to ERCP.216

Computed Tomographic Cholangiography and Magnetic Resonance Cholangiography

In patients with cholelithiasis or choledocholithiasis, CT has been used principally for detecting complications such as pericholecystic fluid in acute cholecystitis, gas in the gallbladder wall suggesting emphysematous cholecystitis, gallbladder perforation, and abscesses (Fig. 65-12). Spiral CT cholangiography (CTC) with use of an oral cholecysto-

1111

1112

Section VIII  Biliary Tract

GB

Stone

BD

obstruction. The term chronic cholecystitis to describe biliary pain should be avoided because it implies the presence of a chronic inflammatory infiltrate that may or may not be present in a given patient. Indeed, the severity and frequency of biliary pain and the pathologic changes in the gallbladder do not correlate significantly.222 The most common histologic changes observed in patients with biliary pain are mild fibrosis of the gallbladder wall with a chronic inflammatory cell infiltrate and an intact mucosa. Recurrent episodes of biliary pain also can be associated with a scarred, shrunken gallbladder and RokitanskyAschoff sinuses (intramural diverticula). Bacteria can be cultured from gallbladder bile or gallstones themselves in about 10% of patients with biliary pain, but bacterial infection is not believed to contribute to the symptoms (see also Chapter 67).

Clinical Features

Figure 65-13.  Magnetic resonance cholangiography demonstrating choledocholithiasis. Within the bile duct (BD) are two filling defects repre­ senting gallstones. GB, gallbladder. (Courtesy of Charles Owen, III, MD, Dallas, Tex.)

graphic contrast agent has been studied for the detection of choledocholithiasis.217,218 Although CTC is still inferior to ERCP imaging for detecting bile duct stones, it may reveal other surrounding pathologic abnormalities.217 MRC is highly useful for imaging the bile duct and detecting gallstones. This modality is especially useful for detecting abnormalities in the most distal extrahepatic portion of the bile duct when the duct is not dilated; this region often is not well visualized by transabdominal ultrasonography.195 With the advent of laparoscopic cholecystectomy, an easy, quick, and, preferably, noninvasive method of excluding bile duct stones is needed. MRC permits the construction of a three-dimensional image of the bile duct with a high sensitivity for detecting bile duct stones (Fig. 65-13).219,220 In a systematic review that compared MRC with diagnostic ERCP for the detection of choledocholithiasis, MRC had a sensitivity of 93% and a specificity of 94%.221 CTC and MRC are noninvasive but, unlike ERCP, have no therapeutic application. They are most useful for excluding choledocholithiasis (either preoperatively or postoperatively) in patients who undergo cholecystectomy and in whom the probability of bile duct stones is believed to be low. ERCP is now reserved for patients with a higher probability of bile duct stones and thus the need for therapeutic intervention.

BILIARY PAIN AND CHRONIC CHOLECYSTITIS

Biliary pain is the most common manifesting symptom of cholelithiasis, and about 75% of patients with symptomatic gallstone disease seek medical attention for episodic abdominal pain. In patients who present with a complication of gallstones, such as acute cholecystitis, a history of recurrent episodes of abdominal pain in the months preceding the complication can often be elicited.

Pathogenesis

Biliary pain (conventionally referred to as biliary “colic,” a misnomer) is caused by intermittent obstruction of the cystic duct by one or more gallstones. Biliary pain does not require that inflammation of the gallbladder accompany the

Biliary pain is visceral in nature and, thus, poorly localized.223 In a typical case, the patient experiences episodes of upper abdominal pain, usually in the epigastrium or right upper quadrant (RUQ) but sometimes in other abdominal locations. Ingestion of a meal often can precipitate pain, but more commonly no inciting event is apparent. The onset of biliary pain is more likely to occur during periods of weight reduction and marked physical inactivity such as prolonged bed rest than at other times. The term biliary colic, used in the past, is a misnomer because the pain is steady rather than intermittent as would be suggested by the word “colic.” The pain increases gradually over a period of 15 minutes to an hour and then remains at a plateau for an hour or more before slowly resolving. In one third of patients, the onset of pain may be more sudden, and on rare occasions, the pain may cease abruptly. Pain lasting more than six hours suggests acute cholecystitis rather than simple biliary pain. In order of decreasing frequency, biliary pain is felt maximally in the epigastrium, RUQ, left upper quadrant, and various parts of the precordium or lower abdomen. Therefore, the notion that pain not located in the RUQ is atypical of gallstone disease is incorrect. Radiation of the pain to the scapula, right shoulder, or lower abdomen occurs in one half of patients. Diaphoresis and nausea with some vomiting are common, although vomiting is not as protracted as in intestinal obstruction or acute pancreatitis. Like patients with other kinds of visceral pain, the patient with biliary pain is usually restless and active during an episode. Complaints of gas, bloating, flatulence, and dyspepsia, which are common in patients with gallstones, are probably not related to the stones themselves. These nonspecific symptoms are found with similar frequencies in persons without gallstones. Accordingly, patients with gallstones whose only symptoms are dyspepsia and other nonspecific upper gastrointestinal tract complaints are not candidates for cholecystectomy. Physical findings are usually normal, with only mild to moderate gallbladder tenderness during an attack and perhaps mild residual tenderness lasting several days after an attack.

Natural History

Biliary pain is cause for concern but not alarm. Approximately 30% of patients who have an attack of classic biliary pain will experience no additional attacks over the next 24 months. Therefore, a reasonable approach would be to offer cholecystectomy to patients with recurring episodes of biliary pain.224 In the remaining 70%, the frequency of recurrent attacks varies widely from patient to patient, but the pattern remains relatively constant for an individual

Chapter 65  Gallstone Disease patient over time. In patients monitored after an initial attack of biliary pain, symptoms sufficient to warrant cholecystectomy develop on average at a rate of approximately 6% per year. The cumulative risk that symptoms that require therapy will develop in asymptomatic persons with gallstones who are followed up for five years is 7.6%.225 The probability that a patient with a history of biliary pain will experience a complication of gallstones that requires urgent surgical intervention is only 1% to 2% per year.224

Diagnosis

In a patient with uncomplicated biliary pain, laboratory parameters are usually normal. Elevations of serum bilirubin, alkaline phosphatase, or amylase levels suggest coexisting choledocholithiasis. In general, the first, and often the only, imaging study recommended in patients with biliary pain is ultrasound of the RUQ. Ultrasonography is rapid, noninvasive, highly sensitive, and highly specific for detecting stones in the gallbladder. Despite the impressive diagnostic accuracy of ultrasonography, a clinically important stone is occasionally missed and the correct diagnosis delayed because of the large number of patients who undergo ultrasonography for any reason.195 Given the relatively benign natural history of biliary pain, patients with suspected gallstones but a negative ultrasonography result can safely be observed, further diagnostic testing being reserved for those in whom symptoms recur.226 Oral cholecystography is generally viewed as a secondary imaging study of the gallbladder and is reserved for patients in whom medical dissolution therapy or lithotripsy of gallstones is planned (see Chapter 66). In such cases, patency of the cystic duct must be confirmed by OCG before therapy. On rare occasions, OCG may demonstrate small floating gallstones that were missed by ultrasonography.

ciated with abdominal pain, RUQ tenderness, fever, and leukocytosis is the hallmark of acute cholecystitis. In approximately 90% of cases, the underlying cause is obstruction of the outlet of the gallbladder by a gallstone in the cystic duct, the gallbladder neck, or Hartman’s pouch.229 In the remaining 10% of cases, cholecystitis occurs in the absence of gallstones (acalculous cholecystitis; see Chapter 67). Acute cholecystitis caused by gallstones is a disease of young, otherwise healthy women and generally has a favorable prognosis, whereas acute acalculous cholecystitis occurs more commonly in critically ill elderly men and is associated with high morbidity and mortality rates.

Pathogenesis

Patients with recurrent, uncomplicated biliary pain and documented gallstones are generally treated with elective laparoscopic cholecystectomy, as discussed in Chapter 66. Acute biliary pain improves with administration of meperidine, with or without ketorolac or diclofenac. Aspirin taken prophylactically has been reported to prevent gallstone formation as well as acute attacks of biliary pain in patients with gallstones, but long-term use of NSAIDs does not prevent gallstone formation.227,228

Acute cholecystitis generally occurs when a stone becomes embedded in the cystic duct and causes chronic obstruction, rather than transient obstruction as in biliary pain.229 Stasis of bile within the gallbladder lumen results in damage of the gallbladder mucosa with consequent release of intracellular enzymes and activation of a cascade of inflammatory mediators. In animal studies, if the cystic duct is ligated, the usual result is gradual absorption of the gallbladder contents without the development of inflammation230; the additional instillation of a luminal irritant, such as concentrated bile or lysolecithin, or trauma from an indwelling catheter is required to cause acute cholecystitis in an obstructed gallbladder. Phospholipase A is believed to be released by gallstoneinduced mucosal trauma and converts lecithin to lysolecithin. Although normally absent from gallbladder bile, lysolecithin is present in the gallbladder contents of patients with acute cholecystitis.231 In animal models, installation of lysolecithin into the gallbladder produces acute cholecystitis associated with increased protein secretion, decreased water absorption, and evidence of white blood cell (WBC) invasion associated with elevated production of prostaglandins E and F1α. Administration of indomethacin, a cyclooxygenase inhibitor, has been shown to block this inflammatory response. Studies of human tissue obtained at cholecystectomy have demonstrated enhanced prostaglandin production in the inflamed gallbladder. Additionally, administration of intravenous indomethacin and oral ibuprofen to patients with acute cholecystitis has been shown to diminish both luminal pressure in the gallbladder and pain.231 Supporting evidence for the role of prostaglandins in the development of acute cholecystitis comes from a prospective study in which patients who presented with biliary pain were randomized to receive diclofenac, a prostaglandin synthetase inhibitor, or placebo.232 Ultimately, acute cholecystitis developed in 9 of 40 patients who received placebo, whereas episodes of biliary pain resolved in all 20 patients who received diclofenac. These data suggest a chain of events in which obstruction of the cystic duct in association with one or more intraluminal factors damages the gallbladder mucosa and stimulates prostaglandin synthetase. The resulting fluid secretion and inflammatory changes promote a cycle of further mucosal damage and inflammation.232 Enteric bacteria can be cultured from gallbladder bile in approximately one half of patients with acute cholecystitis.233 Bacteria are not believed, however, to trigger the actual onset of acute cholecystitis.

ACUTE CHOLECYSTITIS

Pathology

Differential Diagnosis

The differential diagnosis of recurrent, episodic upper abdominal symptoms includes reflux esophagitis, peptic ulcer, pancreatitis, renal colic, diverticulitis, carcinoma of the colon, irritable bowel syndrome, radiculopathy, and angina pectoris. Usually, a carefully taken history assists in narrowing the differential diagnosis. For example, relief of pain with food, antacids, or antisecretory drugs suggests peptic ulcer, whereas pain of a cramping nature suggests an intestinal disorder. The pain of angina pectoris usually is precipitated by exercise and does not last for hours, and the pain of renal stones usually is associated with abnormal findings on urinalysis. Like biliary pain, irritable bowel syndrome is common in young women, but pain is associated with altered bowel habits. The pain of shingles or a radiculopathy from osteoarthritis occasionally may resemble biliary pain.

Treatment

Acute cholecystitis is the most common complication of gallstone disease. Inflammation of the gallbladder wall asso-

If examined in the first few days of an attack of acute cholecystitis, the gallbladder usually is distended and contains

1113

1114

Section VIII  Biliary Tract a stone embedded in the cystic duct.234 After the gallbladder is opened, inflammatory exudate and, rarely, pus are present. Later in the attack, the bile pigments that are normally present are absorbed, having been replaced by thin mucoid fluid, pus, or blood. If the attack of acute cholecystitis is left untreated for a long period but the cystic duct remains obstructed, the lumen of the gallbladder may become distended with clear mucoid fluid, a condition known as hydrops of the gallbladder. Histologic changes range from mild acute inflammation with edema to necrosis and perforation of the gallbladder wall. Surprisingly, the severity of histologic changes correlates little with the patient’s symptoms.234 If the gallbladder is resected for acute cholecystitis and no stones are found, the specimen should be carefully examined histologically for evidence of vasculitis or cholesterol emboli because these systemic disorders may manifest as acalculous cholecystitis (see Chapter 35).

Clinical Features

Approximately 75% of patients with acute cholecystitis report prior attacks of biliary pain (see Table 65-2).235 Often, such a patient is alerted to the possibility that more than simple biliary pain is occurring by the prolonged duration of the pain. If biliary pain has been constant for more than six hours, uncomplicated biliary pain is increasingly unlikely, and acute cholecystitis should be suspected. As inflammation in the gallbladder wall progresses, poorly localized visceral pain gives way to moderately severe parietal pain that localizes to the RUQ.235 Less commonly the back or rarely the chest may be the site of maximal pain. Nausea with some vomiting is characteristic of acute cholecystitis, but these symptoms almost invariably follow, rather than precede, the onset of pain. Vomiting is not as persistent or as severe as that with intestinal obstruction or acute pancreatitis. In some patients, the symptoms of acute cholecystitis are nonspecific, with only a mild ache and anorexia, whereas other patients may present with toxic manifestations, including fever and severe RUQ pain with guarding and localized rebound tenderness. In contrast to uncomplicated biliary pain, the physical findings can, in many cases, suggest the diagnosis of acute cholecystitis. Fever is common, but body temperature is usually less than 102°F unless the gallbladder has become gangrenous or has perforated (Fig. 65-14). Mild jaundice is present in 20% of patients with acute cholecystitis and 40% of elderly patients. Serum bilirubin levels usually are less than 4 mg/dL.236 Bilirubin levels above this value suggest the possibility of bile duct stones, which may be found in 50% of jaundiced patients with acute cholecystitis. Another cause of pronounced jaundice in patients with acute cholecystitis is Mirizzi’s syndrome, which is associated with inflammatory obstruction of the common hepatic duct (see later). The abdominal examination often demonstrates right subcostal tenderness, with a palpable gallbladder in one third of patients. A palpable gallbladder is more common in patients having a first attack of acute cholecystitis because repeated attacks usually result in a scarred, fibrotic gallbladder that is unable to distend. For unclear reasons, the gallbladder is usually palpable lateral to its normal anatomic location. A relatively specific finding of acute cholecystitis is Murphy’s sign.235 During palpation in the right subcostal region, pain and inspiratory arrest may occur when the patient takes a deep breath that brings the inflamed gallbladder into

GB Fluid collection

Figure 65-14.  Ultrasonography demonstrating a complex fluid collection adjacent to the gallbladder (GB), consistent with gallbladder perforation. (Courtesy of Julie Champine, MD, Dallas, Tex.)

contact with the examiner’s hand. The presence of Murphy’s sign in the appropriate clinical setting is a reliable predictor of acute cholecystitis, although gallstones should still be confirmed by ultrasonography.

Natural History

The pain of untreated acute cholecystitis generally resolves in 7 to 10 days.237 Not uncommonly, symptoms remit within 48 hours of hospitalization. One study has shown that acute cholecystitis resolves without complications in approximately 83% of patients but results in gangrenous cholecystitis in 7%, gallbladder empyema in 6%, perforation in 3%, and emphysematous cholecystitis in fewer than 1%.238

Diagnosis

Perhaps because it is so common, acute cholecystitis is often at the top of the differential diagnosis of abdominal symptoms and is actually overdiagnosed when clinical criteria alone are considered. In a prospective series of 100 patients with RUQ pain and tenderness and suspected acute cholecystitis, this diagnosis was correct in only two thirds of cases. The clinician must therefore use laboratory and imaging studies to confirm the presence of acute chole­ cystitis, exclude complications such as gangrene and per­ foration, and look for alternative causes of the clinical findings. Table 65-3 details the most common laboratory findings in acute cholecystitis.237 Leukocytosis with a shift to immature neutrophils is common. Because a diagnosis of bile duct stones with cholangitis usually is in the differential diagnosis, attention is directed to results of liver biochemical tests.236 Even without any detectable bile duct obstruction, acute cholecystitis often causes mild elevations in serum aminotransferase and alkaline phosphatase levels. As noted earlier, the serum bilirubin level may also be mildly elevated (2 to 4 mg/dL), and even serum amylase and lipase values may be elevated nonspecifically. A serum bilirubin value greater than 4 mg/dL or amylase value greater than 1000 U/L usually indicates coexisting bile duct obstruction or acute pancreatitis and warrants further evaluation. When the level of leukocytosis exceeds 15,000 cells/mm3, particularly in the setting of worsening pain, high fever (temperature > 102°F), and chills, suppurative cholecystitis

Chapter 65  Gallstone Disease (empyema) or perforation should be suspected, and urgent surgical intervention may be required. Such advanced gallbladder disease may be present even if local and systemic manifestations are unimpressive. Ultrasonography is the single most useful imaging study in acutely ill patients with RUQ pain and tenderness. It accurately establishes the presence or absence of gallstones and serves as an extension of the physical examination. Presence of sonographic Murphy’s sign, defined as focal gallbladder tenderness under the transducer, has a positive predictive value greater than 90% for detecting acute cholecystitis if gallstones are also present, the operator is skillful, and the patient is alert.239 Additionally, ultraso­ nography can detect nonspecific findings suggestive of acute cholecystitis, such as pericholecystic fluid and gallbladder wall thickening greater than 4 mm. Both findings lose specificity for acute cholecystitis if the patient has ascites or hypoalbuminemia.195,240 Because the prevalence of gallstones is high in the population, many patients with nonbiliary tract diseases that manifest as acute abdominal pain (such as acute pancreatitis and complications of peptic ulcer) may have incidental and clinically irrelevant gallstones. The greatest usefulness of cholescintigraphy in these patients is its ability to exclude acute cholecystitis and allow the clinician to focus on nonbiliary causes of the patient’s acute abdominal pain.188 A normal cholescintigraphy scan result shows radioactivity in the gallbladder, bile duct, and small intestine within 30 to 60 minutes of injection of the isotope. With rare exceptions, a normal result excludes acute cholecystitis caused by gallstones. Several studies have suggested that the sensitivity and specificity of scintigraphy in the setting of acute cholecystitis are approximately 94% each. Its sensitivity and specificity are reduced considerably, however, in patients who have liver disease, are receiving parenteral nutrition, or are fasting. These conditions can lead to a false-positive scan result, defined as the absence of isotope in the gallbladder in a patient who does not have acute cholecystitis. If a positive scan result is defined as the absence of isotope in the gallbladder, then a false-negative scan result would be defined as filling of the gallbladder with isotope in the setting of acute cholecystitis, a situation that virtually never occurs. Therefore, scintigraphy should not be used as the initial imaging study in a patient with suspected cholecystitis but rather should be used as a secondary imaging study in patients who already are known to have gallstones and in whom a nonbiliary cause of acute abdominal pain is possible.241 In an effort to reduce the occurrence of false-positive cholescintigraphy scan results, morphine can be administered to the patient if the gallbladder has failed to be visualized after 60 minutes. As discussed earlier, by increasing pressure within the sphincter of Oddi, administration of morphine directs bile into the gallbladder unless the cystic duct is obstructed. Additional scans obtained 30 minutes after the injection of morphine occasionally show filling of the gallbladder with isotope, thereby excluding cystic duct obstruction. Unfortunately, despite the use of morphine augmentation, cholescintigraphy continues to have a falsepositive rate of 60% in critically ill patients. The greatest usefulness of abdominal CT in patients with acute cholecystitis is to detect complications such as emphysematous cholecystitis and perforation of the gallbladder. At the same time, CT can exclude other intraabdominal processes that may engender a similar clinical picture. For example, abdominal CT is highly sensitive for detecting pneumoperitoneum, acute pancreatitis, pancre-

atic pseudocysts, hepatic or intra-abdominal abscesses, appendicitis, and obstruction or perforation of a hollow viscus. An abdominal CT scan usually is not warranted in patients with obvious acute cholecystitis, but if the diagnosis is uncertain or the optimal timing of surgery is in doubt, CT may be invaluable.

Differential Diagnosis

The principal conditions to consider in the differential diagnosis of acute cholecystitis are appendicitis, acute pancreatitis, pyelonephritis or renal calculi, peptic ulcer, acute hepatitis, pneumonia, hepatic abscess or tumor, and gonococcal or chlamydial perihepatitis. These possibi­ lities should be considered before a cholecystectomy is recommended. Acute appendicitis is the disease most often confused with acute cholecystitis because the initial diagnostic impression is based largely on localized right-sided abdo­ minal tenderness, which may be lower than expected in cholecystitis or higher than expected in appendicitis. In general, fever, leukocytosis, and tenderness progress more inexorably in appendicitis. Abdominal CT usually can distinguish these two entities (see Chapter 116). Acute pancreatitis also may be difficult to distinguish from acute cholecystitis on the basis of the history and physical examination alone. Generally, vomiting is a more prominent feature of acute pancreatitis, and affected persons are more uncomfortable in a supine position. Hyperamylasemia is more profound in pancreatitis than cholecystitis, and an elevated serum lipase value is more specific (see Chapter 58). Diseases of the right kidney may produce pain and tenderness that mimic findings in acute cholecystitis, but urinalysis and ultrasonography can usually differentiate renal disease from cholecystitis. The pain of an uncomplicated peptic ulcer is usually chronic and seldom confused with that of acute cholecystitis, but a perforated ulcer, at least initially, may mimic severe acute cholecystitis. Signs of generalized peritonitis or pneumoperitoneum strongly suggest a perforated viscus, necessitating emergency laparotomy (see Chapters 10 and 52). Pneumonia with pleurisy may cause abdominal pain and tenderness, but the pleuritic nature of the pain and the chest radiograph findings should suggest the correct diagnosis. In some instances, acute hepatitis, especially when caused by alcohol, may manifest as severe RUQ pain and tenderness, fever, and leukocytosis and may be confused diagnostically with acute cholecystitis. In such cases, careful assessment of the liver biochemical values over time in combination with ultrasonography or cholescintigraphy may exclude a diagnosis of acute cholecystitis. Rarely, liver biopsy may be warranted (see Chapters 77 to 81 and 84). Gonococcal perihepatitis (Fitz-Hugh–Curtis syndrome) produces RUQ pain and tenderness, which often overshadow any pelvic complaints, as well as leukocytosis. Nevertheless, adnexal tenderness is found on physical examination, and a Gram stain of the cervical smear should show gonococci (see Chapter 37). Hepatic abscesses and tumors usually can be differentiated from acute cholecystitis on the basis of ultrasonographic findings. Prior undiagnosed gallbladder perforation may manifest with fever from a subhepatic abscess. Pseudolithiasis due to ceftriaxone therapy, most often in children, has caused symptoms resembling those of acute cholecystitis, although the gallbladder is his­ tologically normal (see Chapters 82 and 94).

1115

1116

Section VIII  Biliary Tract Treatment

The patient in whom acute cholecystitis is suspected should be hospitalized. The patient is often hypovolemic from vomiting and poor oral intake, and fluid and electrolytes should be administered intravenously. Oral feeding should be withheld and a nasogastric tube inserted if the patient has a distended abdomen or is vomiting persistently. In uncomplicated cases of acute cholecystitis, antibiotics need not be given. Antibiotics are warranted if the patient appears particularly toxic or a complication such as perforation of the gallbladder or emphysematous cholecystitis is suspected. Antibiotics that cover gram-negative enteric bacteria are effective. Coverage with a single agent such as cefoxitin is appropriate in mild cases, but more severely ill patients should receive broad-spectrum coverage with ampicillin and an aminoglycoside or with a thirdgeneration cephalosporin and metronidazole. Definitive therapy of acute cholecystitis consists of cholecystectomy. The safety and effectiveness of using a laparoscopic approach in the setting of acute cholecystitis have been demonstrated (see Chapter 66).242

CHOLEDOCHOLITHIASIS

Choledocholithiasis is defined as the occurrence of stones in the bile ducts. Like stones in the gallbladder, stones in the bile ducts may remain asymptomatic for years, and stones from the bile duct are known to pass silently into the duodenum, perhaps frequently. Unlike stones in the gallbladder, which usually become clinically evident as relatively benign episodes of recurrent biliary pain, stones in the bile duct, when they do cause symptoms, tend to manifest as life-threatening complications such as cholangitis and acute pancreatitis (see Chapter 58). Therefore, discovery of choledocholithiasis generally should be followed by some type of intervention to remove the stones (see Chapter 70).

Etiology

Gallstones may pass from the gallbladder into the bile duct or can form de novo in the duct. Generally, all gallstones from one patient, whether from the gallbladder or bile duct, are of one type, either cholesterol or pigment. Cholesterol stones form only in the gallbladder, and any cholesterol stones found in the bile duct must have migrated there from the gallbladder. Black pigment stones, which are associated with old age, hemolysis, alcoholism, and cirrhosis, also form in the gallbladder and only rarely migrate into the bile duct. The majority of pigment stones in the bile duct are the softer brown pigment stones. These stones form de novo in the bile duct as a result of bacterial action on phospholipid and bilirubin in bile (see earlier).243 They are often found proximal to biliary strictures and are frequently associated with cholangitis. Brown pigment stones are found in patients with hepatolithiasis and recurrent pyogenic cholangitis (see Chapter 68).244 Fifteen percent of patients with gallbladder stones also have bile duct stones. Conversely, of patients with ductal stones, 95% also have gallbladder stones.245 In patients who present with choledocholithiasis months or years after a cholecystectomy, determining whether the stones were overlooked at the earlier operation or have subsequently formed may be impossible. In fact, formation of pigment stones in the bile duct is also a late complication of endoscopic sphincterotomy.246 In a study of the long-term consequences of endoscopic sphincterotomy in more than 400 patients, the cumulative frequency of recurrent bile duct stones was 12%; all the recurrent stones were of the brown pigment type, irrespective of the chemical composition of

the original gallstones. This observation suggests that sphincterotomy permits chronic bacterial colonization of the bile duct that results in deconjugation of bilirubin and precipitation of pigment stones. Stones in the bile duct usually come to rest at the lower end of the ampulla of Vater. Obstruction of the bile duct raises bile pressure proximally and causes the ducts to dilate. Pressure in the bile duct is normally 10 to 15 cm H2O and rises to 25 to 40 cm H2O with complete obstruction. When pressure exceeds 15 cm H2O, bile flow decreases, and at 30 cm H2O, bile flow stops. The bile duct dilates to the point that dilatation can be detected on either ultrasonography or abdominal CT in approximately 75% of cases. In the patient who has had recurrent bouts of cholangitis, the bile duct may become fibrotic and thus unable to dilate. Moreover, dilatation of the duct is sometimes absent in patients with choledocholithiasis because the obstruction is low-grade and intermittent.

Clinical Features

The morbidity of choledocholithiasis stems principally from biliary obstruction, which raises biliary pressure and diminishes bile flow. The rate of onset of obstruction, its extent, and the amount of bacterial contamination of the bile are the major factors that determine the resulting symptoms. Acute obstruction usually causes biliary pain and jaundice, whereas obstruction that develops gradually over several months may manifest initially as pruritus or jaundice alone.247 If bacteria proliferate, life-threatening cholangitis may result (see later). The physical findings are usually normal if obstruction of the bile duct is intermittent. Mild to moderate jaundice may be noted when obstruction has been present for several days to a few weeks. Deep jaundice without pain, particularly with a palpable gallbladder (Courvoisier’s sign), suggests neoplastic obstruction of the bile duct, even when the patient has stones in the gallbladder. With longstanding obstruction, secondary biliary cirrhosis may result, leading to physical findings of chronic liver disease. As shown in Table 65-2, results of laboratory studies may be the only clue to the presence of choledocholithiasis.248 With bile duct obstruction, serum bilirubin and alkaline phosphatase levels both increase. Bilirubin accumulates in serum because of blocked excretion, whereas alkaline phosphatase levels rise because of increased synthesis of the enzyme by the canalicular epithelium. The rise in the alkaline phosphatase level is more rapid than and precedes the rise in bilirubin level.249 The absolute height of the serum bilirubin level is proportional to the extent of obstruction, but the height of the alkaline phosphatase level bears no relation to either the extent of obstruction or its cause. In cases of choledocholithiasis, the serum bilirubin level is typically in the range of 2 to 5 mg/dL205 and rarely exceeds 12 mg/dL. Transient “spikes” in serum aminotransferase or amylase levels suggest passage of a bile duct stone into the duodenum. The overall sensitivity of liver biochemical testing for detecting choledocholithiasis is reported to be 94%; serum levels of gamma glutamyl transpeptidase are elevated most commonly but may not be assessed in clinical practice.249

Natural History

Little information is available on the natural history of asymptomatic bile duct stones. In many patients such stones remain asymptomatic for months or years, but available evidence suggests that the natural history of asymptomatic bile duct stones is less benign than that of asymptomatic gallstones.247,250

Chapter 65  Gallstone Disease Diagnosis

Ultrasonography actually visualizes bile duct stones in only about 50% of cases,199 whereas dilatation of the bile duct to a diameter greater than 6 mm is seen in about 75% of cases. Ultrasonography can confirm, or at least suggest, the presence of bile duct stones but cannot exclude choledocho­ lithiasis definitively. EUS, although clearly more invasive than standard ultrasonography, has the advantage of visualizing the bile duct more accurately. In preliminary studies, EUS has excluded or confirmed choledocholithiasis with sensitivity and specificity rates of approximately 98% as compared with ERCP.203 ERCP is the standard method for the diagnosis and therapy of bile duct stones,251 with sensitivity and speci­ ficity rates of approximately 95%. When the clinical probabi­lity of choledocholithiasis is low, however, less invasive studies, such as EUS and MRCP, should be performed first.216 Percutaneous transhepatic cholangiography (percutaneous THC) is also an accurate test for confirming the presence of choledocholithiasis. The procedure is most readily accomplished when the intrahepatic bile ducts are dilated and now is performed primarily when ERCP is unavailable or has been technically unsuccessful. Laparoscopic ultrasonography may be used in the surgical suite immediately before mobilization of the gallbladder during cholecystectomy. Laparoscopic ultrasono­ graphy may be as accurate as surgical cholangiography in detecting bile duct stones and may thereby obviate the need for the latter.252

Differential Diagnosis

Symptoms caused by obstruction of the bile duct cannot be distinguished from those caused by obstruction of the cystic duct. Therefore, biliary pain is always in the differential diagnosis in patients with an intact gallbladder. The presence of jaundice or abnormal liver biochemical test results strongly points to the bile duct rather than the gallbladder as the source of the pain. In patients who present with jaundice, malignant obstruction of the bile duct or obstruction from a choledochal cyst may be indistinguishable clinically from choledocholithiasis (see Chapters 62 and 69). Acute passive congestion of the liver, associated with cardiac decompensation, may cause intense RUQ pain, tenderness, and even jaundice with serum bilirubin levels higher than 10 mg/dL (see Chapter 83); however, fever is usually absent, and the WBC count is normal or only slightly elevated. The patient typically has other obvious signs of cardiac decompensation. Constrictive pericarditis and cor pulmonale also may cause acute congestion of the liver with only subtle cardiac findings. Acute viral hepatitis rarely may cause severe RUQ pain with tenderness and fever. The WBC count, however, usually is not elevated, whereas serum alanine aminotransferase and aspartate aminotransferase levels are markedly elevated. Acquired immunodeficiency syndrome (AIDS)-associated cholangiopathy253 and papillary stenosis must be considered in human immunodeficiency virus–positive patients with RUQ pain and abnormal liver biochemical test results (see Chapter 33).

Treatment

Because of its propensity to result in serious complications such as cholangitis and acute pancreatitis, choledocho­ lithiasis warrants treatment in nearly all cases.254 The optimal therapy for a given patient depends on the severity

of symptoms, presence of coexisting medical problems, availability of local expertise, and presence or absence of the gallbladder. Bile duct stones discovered at the time of a laparoscopic cholecystectomy present a dilemma to the surgeon. Some surgeons may attempt laparoscopic exploration of the bile duct. In other cases, the operation can be converted to an open cholecystectomy with bile duct exploration, but this approach results in greater morbidity and a more prolonged hospital stay. Alternatively, the laparoscopic cholecystectomy can be carried out as planned, and the patient can return for ERCP with removal of the bile duct stones. Such an approach, if successful, cures the disease but runs the risk of necessitating a third procedure, namely a bile duct exploration, if the stones cannot be removed at ERCP. In general, the greater the expertise of the therapeutic endoscopist, the more inclined the surgeon should be to complete the laparoscopic cholecystectomy and have the bile duct stones removed endoscopically.254 In especially high-risk patients, endoscopic removal of bile duct stones may be performed without cholecystectomy. This approach is particularly appropriate for elderly patients with other severe illnesses.255 Cholecystectomy is required subsequently for recurrent symptoms in only 10% of patients. The surgical management and endoscopic treatment of gallstones are discussed in detail in Chapters 66 and 70, respectively.

CHOLANGITIS

Of all the common complications of gallstones, the most serious and the most lethal is acute bacterial cholangitis. Pus under pressure in the bile ducts leads to rapid spread of bacteria via the liver into the blood, with resulting septicemia. Moreover, the diagnosis of cholangitis is often problematic (especially in the critical early phase of the disease) because clinical features that point to the biliary tract as the source of sepsis are often absent.26 Table 65-2 delineates the symptoms, signs, and laboratory findings that can aid in an early diagnosis of cholangitis.

Etiology and Pathophysiology

In approximately 85% of cases, cholangitis is caused by a stone embedded in the bile duct, with resulting bile stasis.257 Other causes of bile duct obstruction that may result in cholangitis are neoplasms (see Chapters 60 and 69), biliary strictures (see Chapters 68 and 70), parasitic infections (see Chapters 68 and 82), and congenital abnormalities of the bile ducts (see Chapter 62). This discussion deals specifically with cholangitis caused by gallstones in the bile duct. Bile duct obstruction is necessary, but not sufficient, to cause cholangitis. Cholangitis is relatively common in patients with choledocholithiasis and nearly universal in patients with a post-traumatic bile duct stricture but is seen in only 15% of patients with neoplastic obstruction of the bile duct. It is most likely to result when a bile duct that already contains bacteria becomes obstructed, as is the case in most patients with choledocholithiasis and stricture but in few patients with neoplastic obstruction. Malignant obstruction is more often complete than obstruction by a stricture or a bile duct stone and less commonly permits the reflux of bacteria from duodenal contents into the bile ducts.258 The bacterial species most commonly cultured from the bile are E. coli, Klebsiella, Pseudomonas, Proteus, and enterococci. Anaerobic species such as Bacteroides fragilis and Clostridium perfringens are found in about 15% of appropriately cultured bile specimens. Anaerobes usually accompany aerobes, especially E. coli. The shaking chills

1117

1118

Section VIII  Biliary Tract and fever of cholangitis are caused by bacteremia from bile duct organisms. The degree of regurgitation of bacteria from bile into hepatic venous blood is directly proportional to the biliary pressure and, hence, the degree of obstruction.258 For this reason, decompression alone often effectively treats the illness.

Clinical Features

The hallmark of cholangitis is Charcot’s triad, consisting of RUQ pain, jaundice, and fever (see Table 65-2). The full triad is present in only 70% of patients.258 The pain of cholangitis may be surprisingly mild and transient but is often accompanied by chills and rigors. Elderly patients in particular may present solely with mental confusion, lethargy, and delirium. Altered mental status and hypo­ tension in combination with Charcot’s triad, known commonly as Reynolds’ pentad, occur in severe suppurative cholangitis. On physical examination, fever is almost universal, occurring in 95% of patients. RUQ tenderness is elicited in approximately 90% of patients, but jaundice is clinically detectable in only 80%. Notably, peritoneal signs are found in only 15% of patients. The combination of hypo­ tension and mental confusion indicates gram-negative septicemia. In overlooked cases of severe cholangitis, intrahepatic abscess may manifest as a late complication (see Chapter 82). Laboratory study results are often helpful in pointing to the biliary tract as the source of sepsis. In particular, the serum bilirubin level exceeds 2 mg/dL in 80% of patients. When the bilirubin level is normal initially, the diagnosis of cholangitis may not be suspected.249 The WBC count is elevated in 80% of patients. In many patients who have a normal WBC count, examination of the peripheral blood smear reveals a dramatic shift to immature neutrophil forms. The serum alkaline phosphatase level is usually elevated, and the serum amylase level may also be elevated if pancreatitis is also present. In the majority of cases, blood culture results are positive for enteric organisms, especially if culture specimens are obtained during chills and fever spikes. The organism found in the blood is invariably the same as that found in the bile.

Diagnosis

The principles of radiologic diagnosis of cholangitis are the same as those for choledocholithiasis. As shown in Table 65-3, stones in the bile duct are seen ultrasonographically in only about 50% of cases155 but can be inferred by detection of a dilated bile duct in about 75% of cases. Normal ultrasonography findings do not exclude the possibility of choledocholithiasis in a patient in whom the clinical presentation suggests cholangitis.241 An abdominal CT is an excellent test for excluding complications of gallstones such as acute pancreatitis and abscess, but a standard abdominal CT scan is not capable of excluding bile duct stones. EUS and MRC, as noted earlier, have a much higher accuracy rate than CT for detecting and excluding stones in the bile duct. ERCP is the standard test for the diagnosis of bile duct stones and cholangitis. Moreover, the ability of ERCP to establish drainage of infected bile under pressure can be life-saving. If ERCP is unsuccessful, percutaneous THC can be performed (see Chapter 70).

Treatment

In cases of suspected bacterial cholangitis, blood culture specimens should be obtained immediately and therapy

started with antibiotics effective against the likely causative organisms.259 In mild cases, initial therapy with a single drug, such as cefoxitin, 2.0 g intravenously every six to eight hours, is usually sufficient. In severe cases, more intensive therapy (e.g., gentamicin, ampicillin, and metronidazole or a broad-spectrum agent such as piperacillintazobactam 3.375 g intravenously every six hours or, if resistant organisms are suspected, meropenem 1 g intravenously every eight hours) is indicated. The patient’s condition should improve within 6 to 12 hours, and in most cases, the infection comes under control within 2 to 3 days, with defervescence, relief of discomfort, and a decline in the WBC count. In these cases, definitive therapy can be planned on an elective basis. If, however, after 6 to 12 hours of careful observation, the patient’s clinical status declines with worsening fever, pain, mental confusion, or hypotension, the bile duct must be decompressed immediately.259 If available, ERCP with stone extraction, or at least decompression of the bile duct with an intrabiliary stent, is the treatment of choice. Controlled studies in which ERCP and decompression of the bile duct were compared with emergency surgery and bile duct exploration have shown dramatically lower morbidity and mortality rates in patients treated endoscopically.254 The surgical treatment and endoscopic management of cholangitis are discussed in detail in Chapters 66 and 70, respectively.

UNCOMMON COMPLICATIONS Table 65-4 describes the clinical manifestations, diagnosis, and treatment of several uncommon complications of gallstone disease.

EMPHYSEMATOUS CHOLECYSTITIS

Patients who have emphysematous cholecystitis present with the same clinical manifestations as patients with uncomplicated acute cholecystitis, but in the former, gasforming organisms have secondarily infected the gallbladder wall. Pockets of gas are evident in the area of the gallbladder fossa on plain abdominal films, ultrasonog­ raphy, and abdominal CT (see Fig. 65-13).260 Emergency antibiotic therapy with anaerobic coverage and early cholecystectomy are warranted because the risk of gallbladder perforation is high. Emphysematous cholecystitis often occurs in diabetic persons or older men who do not have gallstones, in whom atherosclerosis of the cystic artery with resulting ischemia may be the initiating event (see Chapter 67).

CHOLECYSTOENTERIC FISTULA

A cholecystoenteric fistula occurs when a stone erodes through the gallbladder wall (usually the neck) and into a hollow viscus. The most common entry point into the bowel is the duodenum, followed in frequency by the hepatic flexure of the colon, the stomach, and the jejunum. Symptoms are initially similar to those of acute cholecystitis, although at times the stone may pass into the bowel and may be excreted without causing any symptoms.261 Because the biliary tract is decompressed, cholangitis is not common despite gross seeding of the gallbladder and bile ducts with bacteria. The diagnosis of a cholecystoenteric fistula is suspected from radiographic evidence of pneumobilia and may be confirmed by barium contrast studies of the upper or lower gastrointestinal tract; often the precise anatomic location of the fistula is not identified until surgery.

Chapter 65  Gallstone Disease Table 65-4  Uncommon Complications of Gallstone Disease COMPLICATION

PATHOGENESIS

CLINICAL FEATURES

DIAGNOSIS/TREATMENT

Emphysematous cholecystitis

Secondary infection of the gallbladder wall with gasforming organisms (Clostridium welchii, Escherichia coli, and anaerobic streptococci) More common in elderly, diabetic men; can occur without stones (see Chapter 67) Erosion of a (usually large) stone through the gallbladder wall into adjacent bowel, most often the duodenum, followed in frequency by the hepatic flexure, stomach, and jejunum

Symptoms and signs similar to those of severe acute cholecystitis

Mirizzi’s syndrome

Impacted stone in the gallbladder neck or cystic duct with extrinsic compression of the common hepatic duct from accompanying inflammation or fistula

Jaundice and right upper quadrant pain

Porcelain gallbladder

Intramural calcification of the gallbladder wall, usually in association with stones

No symptoms attributable to the calcified wall per se, but carcinoma of the gallbladder is a late complication in ≈20% (see Chapter 69)

Plain abdominal films may show gas in the gallbladder fossa Ultrasonography and CT are sensitive for confirming gas Treatment with intravenous antibiotics, including anaerobic coverage, and early cholecystectomy High morbidity and mortality rates Plain abdominal films may show gas in the biliary tree and/or a small bowel obstruction in gallstone ileus as well as a stone in the right lower quadrant, if the stone is calcified Contrast gastrointestinal series may demonstrate the fistula A fistula from a solitary stone that passes may close spontaneously Cholecystectomy and bowel closure are curative Gallstone ileus requires emergency laparotomy; the diagnosis is often delayed, with a resulting mortality rate of ≈20% ERCP demonstrates dilated intrahepatic ducts and extrinsic compression of the common hepatic duct and possible fistula Preoperative diagnosis is important to guide surgery and minimize the risk of BD injury Plain abdominal films or CT shows intramural calcification of the gallbladder wall Prophylactic cholecystectomy is indicated to prevent carcinoma

Cholecystoenteric fistula

Symptoms and signs similar to those of acute cholecystitis; although sometimes the fistula may be clinically silent Stones >25 mm, especially in elderly women, may produce a bowel obstruction, or “gallstone ileus”; the terminal ileum is the most common site of obstruction Gastric outlet obstruction (Bouveret’s syndrome) may occur rarely

BD, bile duct; CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography.

If the gallstone exceeds 25 mm in diameter, it may manifest, especially in elderly women, as a small intestinal obstruction (gallstone ileus); the ileocecal area is the most common site of obstruction.262 In such cases, a plain abdominal film may show the pathognomonic features of pneumobilia, a dilated small bowel, and a large gallstone in the right lower quadrant. Unfortunately, the diagnosis of a gallstone ileus is often delayed, with a resulting mortality rate of approximately 20%. Bouveret’s syndrome is characterized by gastric outlet obstruction resulting from duodenal impaction of a large gallstone that has migrated through a cholecystoduodenal fistula.263

tion of gallstones but is mentioned here because of the remarkable tendency of carcinoma to develop as a late complication of gallbladder calcification (specifically, a gallbladder with focal rather than diffuse wall calcification).267 The diagnosis of a porcelain gallbladder can be made with a plain abdominal film or abdominal CT, which shows intramural calcification of the gallbladder wall. Prophy­ lactic cholecystectomy, preferably through a laparoscopic approach, is indicated to prevent the subsequent development of carcinoma, which may otherwise occur in up to 20% of cases (see Chapter 69).268

MIRIZZI’S SYNDROME

The authors wish to acknowledge the significant contributions of colleagues in the gallstone field. This work was supported in part by research grants DK54012 and DK73917 (D.Q.-H.W.) from the National Institutes of Health (U.S. Public Health Service).

Mirizzi’s syndrome is a rare complication in which a stone embedded in the neck of the gallbladder or cystic duct extrinsically compresses the common hepatic duct with resulting jaundice, bile duct obstruction, and, in some cases, a fistula.264,265 Typically the gallbladder contracted and contains stones. ERCP usually demonstrates the characteristic extrinsic compression of the common hepatic duct. Treatment is traditionally by an open cholecystectomy, although endoscopic stenting and laparoscopic cholecystectomy have been performed successfully. Preoperative diagnosis of Mirizzi’s syndrome is important so that bile duct injury can be avoided.266

PORCELAIN GALLBLADDER

Strictly speaking, porcelain gallbladder, defined as intramural calcification of the gallbladder wall, is not a complica-

ACKNOWLEDGMENTS

KEY REFERENCES

Buch S, Schafmayer C, Volzke H, et al. A genome-wide association scan identifies the hepatic cholesterol transporter ABCG8 as a suscepti­ bility factor for human gallstone disease. Nat Genet 2007; 39:995-9. (Ref 159.) Buhman KK, Accad M, Novak S, et al. Resistance to diet-induced hypercholesterolemia and gallstone formation in ACAT2-deficient mice. Nat Med 2000; 6:1341-7. (Ref 164.) Collins C, Maguire D, Ireland A, et al. A prospective study of common bile duct calculi in patients undergoing laparoscopic cholecystectomy: Natural history of choledocholithiasis revisited. Ann Surg 2004; 239:28-33. (Ref 250.)

1119

1120

Section VIII  Biliary Tract Konikoff FM, Chung DS, Donovan JM, et al. Filamentous, helical, and tubular microstructures during cholesterol crystallization from bile. Evidence that cholesterol does not nucleate classic monohydrate plates. J Clin Invest 1992; 90:1155-60. (Ref 74.) Maurer KJ, Ihrig MM, Rogers AB, et al. Identification of cholelithogenic enterohepatic helicobacter species and their role in murine cho­ lesterol gallstone formation. Gastroenterology 2005; 128:1023-33. (Ref 122.) Paigen B, Carey MC. Gallstones. New York: Oxford University Press; 2002. p 298. (Ref 3.) Portincasa P, Di Ciaula A, Wang HH, et al. Coordinate regulation of gallbladder motor function in the gut-liver axis. Hepatology 2008; 47:2112-26. (Ref 97.) Wang DQ, Carey MC. Complete mapping of crystallization pathways during cholesterol precipitation from model bile: Influence of physical-chemical variables of pathophysiologic relevance and identi­ fication of a stable liquid crystalline state in cold, dilute and hydrophilic bile salt-containing systems. J Lipid Res 1996; 37:606-30. (Ref 52.) Wang DQ, Paigen B, Carey MC. Phenotypic characterization of Lith genes that determine susceptibility to cholesterol cholelithiasis in inbred mice: Physical-chemistry of gallbladder bile. J Lipid Res 1997; 38:1395-411. (Ref 73.)

Wang DQ, Schmitz F, Kopin AS, et al. Targeted disruption of the murine cholecystokinin-1 receptor promotes intestinal cholesterol absorption and susceptibility to cholesterol cholelithiasis. J Clin Invest 2004; 114:521-8. (Ref 115.) Wang HH, Portincasa P, Wang DQ. Molecular pathophysiology and physical chemistry of cholesterol gallstones. Front Biosci 2008; 13:401-23. (Ref 4.) Wang HH, Afdhal NH, Gendler SJ, et al. Evidence that gallbladder epithelial mucin enhances cholesterol cholelithogenesis in MUC1 transgenic mice. Gastroenterology 2006; 131:210-22. (Ref 89.) Wang HH, Afdhal NH, Wang DQ. Estrogen receptor alpha, but not beta, plays a major role in 17beta-estradiol-induced murine cholesterol gallstones. Gastroenterology 2004; 127:239-49. (Ref 21.) Wang HH, Afdhal NH, Wang DQ. Overexpression of estrogen receptor alpha increases hepatic cholesterogenesis, leading to biliary hypersecretion in mice. J Lipid Res 2006; 47:778-86. (Ref 22.) Yu L, Hammer RE, Li-Hawkins J, et al. Disruption of Abcg5 and Abcg8 in mice reveals their crucial role in biliary cholesterol secretion. Proc Natl Acad Sci U S A 2002; 99:16237-42. (Ref 60.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

66 Treatment of Gallstone Disease Robert E. Glasgow and Sean J. Mulvihill

CHAPTER OUTLINE Medical Treatment  1121 Dissolution Therapy  1121 Extracorporeal Shock-Wave Lithotripsy  1122 Surgical Treatment  1124 Open Cholecystectomy  1125 Laparoscopic Cholecystectomy  1126 Choice of Treatment  1129 Indications for Treatment  1129 Asymptomatic Gallstones  1129 Biliary Pain  1130 Acute Cholecystitis  1130 Gallstone Pancreatitis  1132 Special Problems  1132

Many options are available for the treatment of patients with symptomatic gallstone disease. Improvements in endoscopic, radiologic, and chemical therapies for gallstones have enhanced the overall management of these patients. Nevertheless, surgery remains the most important therapeutic option. Laparoscopic cholecystectomy has become the standard method for the management of patients with biliary pain and complications of gallstone disease, such as acute cholecystitis, gallstone pancreatitis, and choledocholithiasis. In this chapter, treatment of patients with symptomatic and asymptomatic gallstones is discussed, with particular emphasis on indications for medical and surgical treatment, patient selection, and outcomes.

MEDICAL TREATMENT Medical treatment of gallstone disease was first proposed by Schiff in Italy in 1873.1 Dabney of Virginia first reported the effective treatment of gallstones with bile acids in 1876, an observation later confirmed by Rewbridge of Minnesota in 1937.2,3 Despite these initial reports, the use of medical dissolution treatment did not gain acceptance until large clinical series were reported in the 1970s. Contact dissolution of gallstones with solvents and percutaneous cholecystolithotomy techniques also have been reported, but these modalities have not proved superior to oral dissolution, shock-wave lithotripsy, or laparoscopic cholecystectomy and have been abandoned. The mainstay of current nonsurgical treatment of gallstone disease is oral dissolution with ursodeoxycholic acid, with or without extracorporeal shockwave lithotripsy.

Choledocholithiasis  1133 Choledocholithiasis Known Preoperatively  1134 Choledocholithiasis Identified during Cholecystectomy  1134 Choledocholithiasis Identified after Cholecystectomy  1134 Bile Duct Stricture  1134 Postcholecystectomy Syndrome  1135 Choledocholithiasis  1135 Cystic Duct Remnant  1137 Sphincter of Oddi Dysfunction  1137 Gallstones, Cholecystectomy, and Cancer  1137 Biliary Tract Cancer  1137 Colorectal Cancer  1137

Although nonsurgical treatment of gallstones has proved effective in carefully selected patients, only a limited number of patients are candidates for this treatment option. Nonsurgical treatments are effective only in patients with small, cholesterol gallstones. Significant admixtures of pigment or calcium salts make stones indissoluble. In addition, long-term success with medical treatment of gallstones occurs only in patients in whom the lithogenic disturbance that led to gallstone formation is transient. For most patients, gallstone formation represents an imbalance in biliary lipid excretion, gallbladder stasis, or infection of the bile (see Chapter 65). In these patients, successful dissolution is usually followed by recurrence of gallstones in 30% to 50% of patients within five years.4-7 Therefore, the proper choice of treatment must take into account the type and severity of symptoms, the physical characteristics of the stones, gallbladder function, and the characteristics and preference of the patient.

DISSOLUTION THERAPY

The rationale for oral dissolution therapy is the reversal of the condition that led to formation of cholesterol gallstones, namely, the supersaturation of bile with cholesterol (see Chapter 65). Cholesterol stones dissolve if the surrounding medium is capable of solubilizing the cholesterol in the stones. Both chenodeoxycholic acid and ursodeoxycholic acid dissolve gallstones by decreasing biliary cholesterol secretion and desaturating bile. These agents encourage the removal of cholesterol from stones via micellar solubilization, formation of a liquid crystalline phase, or both. Chenodeoxycholic acid was the first bile acid used for gallstone dissolution but has been abandoned because of side effects, including diarrhea and increased serum aminotransferase

1121

1122

Section VIII  Biliary Tract and cholesterol levels. Ursodeoxycholic acid is well tolerated and is currently used in oral dissolution regimens. In randomized comparisons, ursodeoxycholic acid was just as effective as chenodeoxycholic acid alone or in combination with ursodeoxycholic acid.8-10 The rate of stone dissolution is a function of (1) thermodynamic forces, including the degree of bile desaturation and the concentration of ursodeoxycholic acid in bile; (2) kinetic forces, including stirring of bile; and (3) the surfaceto-volume ratio of the stones. Oral dissolution targets the thermodynamic forces.11 Because small stones have a smaller surface-to-volume ratio, they respond more quickly and reliably to oral dissolution therapy. The use of oral dissolution therapy does not address the problem of gallbladder stasis.12 Although prokinetic agents, including alpha-adrenergic antagonists and clarithromycin, have been shown to increase gallbladder motility, their use in preventing and treating gallstones has not been studied.13,14

Patient Selection

Selection of patients for oral dissolution therapy is a function of the stage of gallstone disease, gallbladder function, and the characteristics of the stones. Selection criteria are summarized in Table 66-1. Oral dissolution therapy should be considered for patients with uncomplicated gallstone disease, including those with mild, infrequent biliary pain. Patients with severe or frequent biliary pain and patients with complications of gallstones, including cholecystitis, pancreatitis, and cholangitis, should not be treated with oral dissolution therapy; these patients should be referred for surgery. In addition, the gallbladder must function and the cystic duct must be patent to allow unsaturated bile and stones to clear from the gallbladder. The patency of the cystic duct is usually evaluated by oral cholecystography. More recently, stimulated cholescintigraphy and functional ultrasonography have been used. These latter modalities assess cystic duct patency as well as gallbladder function. The characteristics of the stones play an important role in determining the efficacy of dissolution treatment. Oral dissolution therapy works only on cholesterol stones. Although verifying the composition of gallstones can be difficult, the appearance of stones on plain films or computed tomography (CT) images can be useful. Cholesterol

Table 66-1  Selection Criteria for Oral Bile Acid Dissolution Therapy Stage of gallstone disease Gallbladder function*

Stone characteristics

Symptomatic (biliary pain) without complications Opacification of gallbladder on oral cholecystography (patent cystic duct) Normal result of stimulated cholescintigraphy (normal gallbladder emptying) Normal result of functional ultrasonography (normal gallbladder emptying after a test meal) Radiolucent on radiography Isodense or hypodense to bile and absence of calcification on computed tomography Diameter <6 mm (optimal) or 6-10 mm (acceptable)

*Either stimulated cholescintigraphy or functional ultrasonography may be done.

stones are radiolucent on conventional radiographs, and they are hypodense or isodense to bile and lack stone calcification on CT images.15 During oral cholecystography, the specific gravity of cholesterol stones is less than or equal to that of contrast-enriched bile, thereby resulting in stone buoyancy.16 The number of stones does not influence the success of oral dissolution therapy; however, only patients with stones that occupy less than one half of the gallbladder volume should be considered for treatment. Although oral dissolution therapy has been effective in stones up to 10 mm in diameter, results are best in stones less than 5 mm in size.17,18 The ideal stones for oral dissolution treatment are shown in Figure 66-1.

Therapeutic Regimens

Ursodeoxycholic acid (ursodiol) is the preferred drug for oral dissolution treatment. It is taken in a dose of 10 to 15 mg/kg of body weight per day. Nighttime dosing is more effective and is associated with better patient compliance than mealtime dosing.19 Unlike chenodeoxycholic acid (chenodiol), ursodeoxycholic acid is well tolerated and has no important side effects. Treatment should continue until stone dissolution is documented by two consecutive negative ultrasonograms at least one month apart. Treatment should be stopped if the patient does not tolerate the drug or experiences a complication of gallstones during therapy or if the stones fail to dissolve after six months or dissolve only partially after six months with lack of progression to complete dissolution by two years.

Efficacy

With ursodeoxycholic acid, complete dissolution is achieved in 20% to 70% of patients. The variability in the reported response rates is a function of differences in patient selection, doses of bile acid, treatment times, and diagnostic techniques used to document stone dissolution.9,10,20,21 A meta-analysis of all randomized trials of dissolution treatment showed stone dissolution in 37% of patients.9 The frequency of stone dissolution was 29% for stones larger than 10 mm, 49% for stones smaller than 10 mm, and 70% for stones less than 5 mm. The time to resolution varies among patients, with a median rate of 0.7 mm per month.11 Improvement in symptoms occurs before stones have dissolved completely. In addition, long-term treatment has been reported to decrease the risk of biliary pain and acute cholecystitis, independent of gallstone dissolution.18 Despite initial dissolution of stones in properly selected patients, the rate of gallstone recurrence after oral dissolution therapy is 50% after five years, with most recurrences within the first two years.5,7 The risk is lower in patients with a solitary stone than in those with multiple stones.

EXTRACORPOREAL SHOCK-WAVE LITHOTRIPSY

The application of extracorporeal shock-wave lithotripsy to the treatment of gallstones was first applied to patients by Sauerbruch in 1985.22 The rationale for shock-wave lithotripsy is to diminish the surface-to-volume ratio of a stone, thereby increasing the efficacy of oral dissolution and decreasing stone size to allow small stones and debris to pass directly from the gallbladder into the intestine without causing symptoms. The technique involves the delivery of focused high-pressure sound waves to gallstones. Three types of lithotripters have been developed: underwater spark-gap, piezoelectric crystals, and electromagnetic membrane lithotriptors.23 Regardless of the energy source, the shock waves from the lithotriptor are delivered from an underwater source to the soft tissue. Passage of the shock wave through the soft tissue does not diminish the energy

Chapter 66  Treatment of Gallstone Disease

A

C

B

D

Figure 66-1.  A and B, Gallstones for which oral dissolution therapy is appropriate: A, ultrasonogram showing small gallstones; B, multiple small cholesterol stones. C and D, Gallstones for which oral dissolution therapy is inappropriate: C, radiopaque gallstones on a plain film; D, large pigmented gallstones.

wave significantly. Passage of the shock wave through the anterior and posterior walls of the stone liberates compressive and tensile forces and causes cavitation at the anterior surface of the stone, thereby leading to stone fragmentation. Factors that influence fragmentation include the size, microcrystalline structure, and architecture of the stone. Although the composition of the stone does not influence successful lithotripsy, only cholesterol stone fragments are dissolved effectively by bile acid therapy.

Table 66-2  Selection Criteria for Extracorporeal Shock-Wave Lithotripsy Stage of gallstone disease Gallbladder function*

Patient Selection

Because shock-wave lithotripsy is usually combined with oral dissolution therapy, patient selection criteria for shockwave lithotripsy are similar to those for oral dissolution treatment and are summarized in Table 66-2. Gallbladder function and cystic duct patency are required and are demonstrated by oral cholecystography, functional ultrasonography, or stimulated cholescintigraphy (see Chapter 65). Lithotripsy should be considered only for patients with mild, uncomplicated biliary pain. Pregnant patients and

Stone characteristics

Symptomatic (biliary pain) without complications Opacification of gallbladder on oral cholecystography (patent cystic duct) Normal result of stimulated cholescintigraphy (normal gallbladder emptying) Normal result of functional ultrasonography (normal gallbladder emptying after a test) Radiolucent on radiography Isodense or hypodense to bile and absence of calcification on computed tomography Single Diameter <20 mm

*Either stimulated cholescintigraphy or functional ultrasonography may be done.

1123

1124

Section VIII  Biliary Tract

A

B

C

Figure 66-2.  Ultrasonograms of a gallbladder with a single stone before (A), one day after (B), and six weeks after (C) extracorporeal shock-wave lithotripsy. Multiple small stone fragments seen one day after lithotripsy have disappeared six weeks after lithotripsy.

patients on anticoagulants should not undergo lithotripsy. Shock-wave lithotripsy is reserved for patients with a solitary stone, measuring less than 2 cm in size.6,24 Because only cholesterol stones are reliably cleared by the addition of oral dissolution therapy, stones should have radiographic features, such as radiolucency, suggestive of cholesterol stones.

Therapeutic Approach

Patients usually are given sedatives and analgesics or are anesthetized and placed in the prone position to minimize the distance between the energy source and the stones and to eliminate interference from intestinal gas and the costal margin. Targeting and monitoring for fragmentation are accomplished with ultrasonography. Multiple treatment sessions are often required to achieve sufficient pulverization. Factors that predict the success of lithotripsy include the degree of fragmentation and gallbladder emptying.25-27 Fragmentation depends on stone characteristics and the dose of the shock wave. Important stone characteristics include the size and number of stones as well as their structure and the presence of calcification.25,28,29 The energy of shock waves, number of shock waves per session, and number of sessions also influence the success rate.30,31 Ursodeoxycholic acid, 10 to 15 mg/kg of body weight per day, is administered orally to dissolve stone fragments, especially when residual stone fragments are larger than 2 mm in size, gallbladder function is poor, or the gallbladder has not cleared small fragments within three to six months of lithotripsy.32 An example of successful combined treatment is shown in Figure 66-2.

Efficacy

The percentage of patients who are free of stones after 6 and 12 months is 47% to 77% and 68% to 84%, respectively.24,25,33-37 Follow-up data reveal 27%, 41%, and 54% cumulative recurrence rates at 3, 5, and 10 years, respectively.38 Recurrence is most often related to the presence of lithogenic bile and gallbladder dysmotility, rather than patient variables such as gender, age, and weight. Factors that predict higher rates of treatment failure include stones larger than 16 mm in size, multiple stones, and stones with a CT density of greater than 84 Hounsfield units.39 Recurrent stones are usually small and multiple and cause recurrent biliary pain. Maintenance therapy with ursodeoxycholic acid after lithotripsy is not cost effective and has not been shown to be effective.7,40 Side effects of lithotripsy include petechiae of the skin at the site of shock-wave delivery (8%), hematuria (4%), and liver hematomas (<1%). No long-term liver biochemical abnormalities have been noted. Biliary pain develops in

approximately one third of patients; cystic duct obstruction develops in 5%; and complications of stone passage, such as biliary pancreatitis, develop in less than 2%.25 Lithotripsy is more cost-effective in the elderly than in the young and less cost-effective in patients with multiple stones than in those with a single stone. When combined with ursodeoxycholic acid, lithotripsy is at least as costeffective as open cholecystectomy for patients with small stones and less cost-effective for those with large stones.41,42 When lithotripsy is compared with laparoscopic cholecystectomy, patients who undergo laparoscopic cholecystectomy experience a greater incremental improvement in quality of life at six months, whereas those who undergo lithotripsy have higher rates of recurrent stones and biliary symptoms.43

Bile Duct Stones

Extracorporeal shock-wave lithotripsy has also been used in the management of choledocholithiasis. Intracorporeal electrohydraulic lithotripsy has been shown to be effective in this setting as well. These treatment options are reserved for patients who fail conventional endoscopic measures (see Chapter 70), mechanical lithotripsy, or surgical treatment of choledocholithiasis. Appropriate indications for shock-wave lithotripsy are large stones impacted in the bile duct that are not amenable to endoscopic extraction, intrahepatic stones, stones above a biliary stricture, cystic duct remnant stones, and bile duct stones associated with Mirizzi’s syndrome (compression of the common hepatic duct) (see Chapter 65). Selection of patients for shock-wave treatment of bile duct stones is similar to that for treatment of uncomplicated gallbladder gallstones. The success rate for treatment of bile duct stones is 70% to 90%.44-49 Most patients require endoscopic extraction of large stone fragments following treatment. Mild, transient hemobilia occurs in 10% of patients, and biliary sepsis develops in 4% following the procedure. Other complications are similar to those seen after lithotripsy for gallbladder stones. Because of the potential for septic complications, preprocedure endoscopic, nasobiliary, or percutaneous biliary drainage is performed. Antibiotics are given to minimize the risk of biliary sepsis.

SURGICAL TREATMENT Approximately 700,000 cholecystectomies are performed for gallstone disease in the United States each year. Currently, the vast majority of these operations are performed using laparoscopic techniques. For example, 7888 cholecys-

Chapter 66  Treatment of Gallstone Disease Table 66-3  Mortality Rates for Open Cholecystectomy as a Function of Clinical Setting Clinical Setting of Cholecystectomy (% Mortality) REFERENCE

YEAR(S)

NO. OF PATIENTS

BILIARY PAIN

ACUTE CHOLECYSTITIS

BILE DUCT EXPLORATION

55 56 57 58

1962-1966 1977-1981 1932-1978 1989

28,621 13,854 11,808 42,474

1.5 0.4 0.5 0.02

3.5 1.6 2.9 0.26

NA 2.3 3.5 NA

NA, not available.

6.0

5.0

Mortality rate (%)

tectomies were performed in Utah in 2005; 96% of these operations were laparoscopic cholecystectomies, and 4% were open procedures. A review of the National Hospital Discharge Database from 1998 to 2001 showed that 25% of cholecystectomies were performed by an open approach. Of the remaining 75% done laparoscopically, the rate of conversion to open cholecystectomy was 5% to 10%.50 Patients with complicated gallstone disease, including acute cholecystitis, gallstone pancreatitis, and choledocholithiasis, are more likely than those with uncomplicated disease to require an open cholecystectomy or conversion from a laparoscopic to an open approach.50 Despite the increased reliance on minimally invasive techniques in the care of these patients, open cholecystectomy continues to play an important role in the management of complications of gallstones.

4.0

3.0

2.0

1.0

OPEN CHOLECYSTECTOMY

Karl Langenbuch, a surgeon in Berlin, is credited with performing the first cholecystectomy in 1882. Since then, cholecystectomy has remained the main therapeutic option for the management of patients with gallstones, largely because of its remarkable success in relieving symptoms and its low morbidity. In prospective studies, 90% to 95% of patients who undergo cholecystectomy experience substantial or complete relief of their symptoms.51,52 Cholecystectomy is more effective in relieving biliary pain than in relieving dyspepsia and flatulence, which are less strongly correlated with gallstone disease.

0 <39

40–49 50–59 60–69 70–79

>80

Age (years) Figure 66-3.  Relationship between age and risk of death from open cholecystectomy. Data include all patients operated on in Denmark between 1977 and 1981. (Data from Bredesen J, Jorgensen T, Andersen TF, et al. Early postoperative mortality following cholecystectomy in the entire female population of Denmark, 1977-1981. World J Surg 1992; 16:530-5.)

Technique

The technique of open cholecystectomy has not changed substantially since its first description. With the surgeon standing on the patient’s right side, a right subcostal (Kochar) incision is made two fingerbreadths below the right costal margin. Alternatively, a midline incision may be used. After exploring the abdomen and taking down any adhesions to the gallbladder, the gallbladder is dissected from the gallbladder fossa in a retrograde fashion, from the fundus down to the infundibulum. When the gallbladder has been mobilized, the cystic artery and duct are readily identified. A cholangiogram may be performed to look for bile duct stones or to confirm the anatomy. The cystic duct and artery are ligated and divided. An alternative approach is to perform a dissection of the Triangle of Calot structures, as is done during laparoscopic cholecystectomy (see later), prior to removing the gallbladder from the liver. The Triangle of Calot is the space bordered by the cystic duct, cystic artery, and inferior edge of the gallbladder. Dissection and identification of these structures permits safe division of the cystic duct and minimizes the chance of bile duct injury. The abdominal incision is then closed. Closed suction drains are rarely indicated after cholecystectomy.

Results

The risk of open cholecystectomy has declined over the years. The overall mortality rate of cholecystectomy in 35,373 patients operated on before 1932 was 6.6%.53 This rate decreased to 1.8% by 1952.54 Since then, the overall mortality rate for cholecystectomy has averaged about 1.5%. The mortality rate is considerably lower in patients operated on electively for biliary pain, with an average of less than 0.5% (Table 66-3).55-57 The risk of death is several-fold higher when cholecystectomy is performed as an emergency for acute cholecystitis and when bile duct exploration is required (see Table 66-3). Additionally, the mortality rate is directly proportional to the patient’s age (Fig. 66-3). In a report of the entire Danish experience with cholecystectomy from 1977 to 1981, patients under the age of 50 years had a risk of death of 0.028% from elective cholecystectomy56; the rate rose to 5.56% in patients older than 80 years of age. The experience in the United States has been similar. Of 11,808 patients who underwent cholecystectomy at the New York Hospital–Cornell Medical Center between 1932 and 1978, the risk of death from elective cholecystectomy for chronic cholecystitis was 0.1% in patients less than age

1125

1126

Section VIII  Biliary Tract 50 years and 0.8% in those 50 years of age or older.57 In a later series, the overall mortality rate of 42,474 patients who underwent cholecystectomy in 1989 in California and Maryland was 0.17%.58 In this series, the mortality rate in patients less than 65 years of age was 0.03%, compared with more than 0.5% in patients 65 years of age and older. Likewise, the morbidity rate, mean length of hospital stay, and average hospital charges were significantly higher in the older patients than in the younger group. Most mortality following cholecystectomy is related to cardiac disease, particularly myocardial infarction. Major complications after open cholecystectomy are rare. In a large survey of 28,621 patients who underwent cholecystectomy in the 1960s, complications occurred in 4.0%.55 Subsequent studies have confirmed a 4.0% to 5.0% rate of perioperative morbidity.59-61 Most complications are relatively minor, such as wound infections or seromas, urinary retention or infection, and atelectasis. Complications related specifically to cholecystectomy include bile leaks, bile duct injury, and acute pancreatitis. Of these complications, bile duct injury is the most serious and often requires endoscopic therapy and, in some cases, complicated and technically difficult surgical repair (see Chapter 70). Alternatively, bile duct injury can lead to benign biliary stricture formation and bile duct obstruction with secondary biliary cirrhosis and liver failure. The rate of bile duct injury during open cholecystectomy is not known precisely but has been estimated to be 1 in 200 to 600 cases.62,63 Two ductal injuries occurred in 1200 cholecystectomies at the Cedars-Sinai Medical Center in Los Angeles from 1982 to 1988,64 and no bile duct injuries occurred in 1252 patients who underwent elective cholecystectomy at two large North American and European centers.65 In general, bile duct injuries are preventable complications and are commonly the result of inadequate surgical training, unrecognized variations in bile duct anatomy, or misidentification of normal anatomy. Unusual amounts of bleeding, severe inflammation, and emergency operations do not play as great a role in these injuries as might be supposed.

LAPAROSCOPIC CHOLECYSTECTOMY

Since the first reports in the late 1980s, laparoscopic cholecystectomy has rapidly gained acceptance as the technique of choice for the management of the typical patient with biliary pain. The driving force behind the rapid development of laparoscopic cholecystectomy initially was patient demand. The benefits of this minimally invasive approach compared with open cholecystectomy, including minimal scarring, less pain, and quicker return to normal activities were recognized by the lay public. Laparoscopic cholecystectomy was introduced and has gained acceptance not through organized and carefully conceived clinical trials, as was the case in the United States for hip replacement surgery, but by acclamation. Historically, laparoscopic cholecystectomy was an outgrowth of diagnostic laparoscopy and the early efforts of gynecologists at operative laparoscopy. The development of laparoscopic cholecystectomy was predicated on technical advances in miniaturized video cameras and other specialized equipment. Few other surgical operations are as dependent on the safe and reliable function of instruments. Advances in instrument and equipment design and manufacture have led to significant improvements in the safety and utility of minimally invasive surgery in the treatment of most gastrointestinal diseases. Laparoscopic cholecystectomy is now the preferred approach to the treatment of gallstone disease and its complications.

Technique

The operative team for laparoscopic cholecystectomy generally consists of a surgeon, assistant surgeon, scrub nurse, circulating nurse, and anesthesiologist. Patients are asked to void immediately prior to entering the operating room. Prophylactic antibiotics are not routinely administered to patients with uncomplicated gallstone disease, including biliary pain.66 Patients with potential infectious com­ plications of gallstones, including acute cholecystitis and cholangitis, and patients with long-standing symptoms or advanced age should receive antibiotics if these agents were not already started before surgery.67 Sequential compression stockings are used to reduce the risk of lower extremity thromboembolism. In order to view the abdominal contents and provide room for instruments, a space is developed by inducing a pneumoperitoneum with a nonflammable, physiologically benign gas such as carbon dioxide. Pneumoperitoneum is achieved by either a closed technique in which a Veress needle is inserted into the peritoneum through a small incision, followed by placement of an operating trocar, or by a direct, open technique in which the operating trocar is placed directly into the abdomen under direct visualization through a small incision. After the pneumoperitoneum has been established, a trocar is placed at the umbilicus and a laparoscope is introduced. Three additional trocars are placed in the upper abdomen under direct vision for insertion of operating instruments and retractors. The current technique of laparoscopic cholecystectomy is best described as “the critical view” approach,68 as summarized in Figure 66-4. In this approach, the entire hepatocystic triangle is dissected, exposing the cystic duct and artery, infundibulum of the gallbladder, and junction of the gallbladder and cystic duct before a cholangiogram is performed or the cystic duct and artery are divided. The assistant retracts the gallbladder fundus cephalad, anterior to the liver, and the infundibulum laterally. The surgeon, operating through the epigastric port, identifies and dissects the cystic duct and artery circumferentially. Special care must be taken to identify the junction of the cystic duct and gallbladder, to ensure that the bile duct has not been isolated inadvertently. Cholangiography is performed via cannulation of the cystic duct. If the cholangiogram shows normal anatomy and no evidence of choledocholithiasis, the cholangiocatheter is removed and the cystic duct and artery are divided between small metal clips. The gallbladder is then dissected out of the liver bed and delivered through the umbilical incision. Care is taken to avoid perforation of the gallbladder during its dissection from the liver because the spillage of gallstones and bile has been shown to increase the risk of postoperative fever and intraabdominal abscess formation.69 The operation concludes with evacuation of the pneumoperitoneum and closure of the incisions.

Rationale for Cholangiography

Cholangiography during laparoscopic cholecystectomy has two main purposes. First, the cholangiogram may detect unsuspected bile duct stones. Second, the cholangiogram confirms the surgeon’s impression of the anatomy of the bile ducts. In the era before laparoscopic cholecystectomy, the value of routine cholangiography during cholecystectomy was debated, with some surgeons arguing in favor of its selective use.70 Routine cholangiography has been criticized because of its relatively low yield, failure to identify all retained stones, occasional false-positive results, cost, and risk. Nevertheless, 8% to 16% of all patients with chole­

Chapter 66  Treatment of Gallstone Disease

A

B

C

D

E

F

G Figure 66-4.  Laparoscopic cholecystectomy. A, Gallbladder in situ. B, Cephalad retraction of the fundus toward the right shoulder exposes the infundibulum of the gallbladder. C, Retraction of the infundibulum toward the right lower quadrant opens the hepatocystic triangle. The hepatocystic triangle is the area bordered by the cystic duct, gallbladder edge, and liver edge. D, Division of the peritoneum overlying the anterior and posterior aspects of the hepatocystic triangle exposes “the critical view.” E, Cholangiogram catheter in the cystic duct. F, Normal cholangiogram. G, Gallbladder removed from the gallbladder fossa with electrocautery.

1127

1128

Section VIII  Biliary Tract Table 66-4  Experience with Laparoscopic Cholecystectomy reference 80 81 82 84 85 a b c d

NO. OF PATIENTS

MORBIDITY RATE (%)

MORTALITY RATE (%)

BILE DUCT INJURY (%)

CONVERSION RATE (%)

3319 6076 13,833 2201 114,005 3285 56,591 33,563 22,953

6.7 4.3 4.3 4.3 5.4 10.1 NA 8.5 14.6

0.15 0.12 0.14 0 0.06 0.2 NA 0.09 0.3

0.33 0.86 0.59 0.14 0.5 0.25 0.42 0.2 0.3

5.2 6.8 5.3 4.3 2.2 3.6 NA 3.5 5.3

a

Data from Velanovich V, Morton JM, McDonald M, et al. Analysis of the SAGES outcomes initiative cholecystectomy registry. Surg Endosc 2006; 20:43-50. Data from Nuzzo G, Giuliante F, Giovannini I, et al. Bile duct injury during laparoscopic cholecystectomy: Results of an Italian national survey on 56,591 cholecystectomies. Arch Surg 2005; 140:986-92. c Data from Savassi-Rocha PR, Ferreira RH, Diniz MT, Sanches SR. Laparoscopic cholecystectomy in Brazil: Analysis of 33,563 cases. Int Surg 1997; 82:208-13. d Data from Giger UF, Michel JM, Opitz I, et al. Risk factors for perioperative complications in patients undergoing laparoscopic cholecystectomy: Analysis of 22,953 consecutive cases from the Swiss Association of Laparoscopic and Thoracoscopic Surgery database. J Am Coll Surg 2006; 203:723-8. NA, not available. b

lithiasis harbor bile duct stones. Routine use of operative cholangiography detects unsuspected bile duct stones in about 5% of patients who undergo cholecystectomy and detects anatomic ductal abnormalities in 12%.71 During laparoscopic cholecystectomy, the two-dimensional video image and inability to palpate structures of the porta hepatitis make identification of the cystic duct–bile duct junction problematic. Cholangiography plays an especially important role in delineating bile duct anatomy prior to division of any important structures. Large population studies from Australia and Sweden have demonstrated the importance of routine intraoperative cholangiography in decreasing the frequency of major bile duct injuries.72,73 The rate of bile duct injury during laparoscopic cholecystectomy when routine cholangiography is performed is 0.2% to 0.4%, compared with 0.4% to 0.6% when cholangiography is not performed routinely.72-74 Routine cholangiography is cost effective when the cost associated with bile duct injuries is considered.75 In addition, routine cholangi­ ography permits earlier identification of intraoperative bile duct injuries, if they occur,76 and thereby improves the rate of repair.

Results

Several large series have now described the experience with laparoscopic cholecystectomy (Table 66-4).76-87 A review of the experience with laparoscopic cholecystectomy in the United States showed an operative mortality rate of 0.06%. Internationally, operative mortality rates have ranged from 0% to 0.15%. Conversion to an open procedure was required in 2.2% in the United States and 3.6% to 8.2% of patients internationally, generally because of inflammation that precludes safe dissection of the porta hepatis. Major morbidity occurred in approximately 5% of patients, and bile duct injuries occurred in 0.14% to 0.5% of patients. Operating time ranged from one to two hours. Initially, most patients were observed overnight after laparoscopic cholecystectomy, but same-day discharge has become standard for elective cases. Most patients return to full activities, including work, within one week. No randomized, prospective trials have compared the results of laparoscopic cholecystectomy with those of open cholecystectomy in the United States, nor are any likely. Patient enthusiasm for the laparoscopic approach and the rapid acceptance of the procedure by surgeons have made direct, controlled comparison of the two procedures difficult. Nonrandomized data from the United States and small,

Table 66-5  Cholecystectomy-Related Mortality in Maryland before (1989) and after (1992) the Introduction of Laparoscopic Cholecystectomy VARIABLE

1989

1992

% CHANGE

Number of cholecystectomies Crude rate of cholecystectomies per 1000 population Operative mortality rate (%) Number of deaths

7416

9993

+35

1.57

2.04

+30

0.84

0.56

−33

62

56

−10

Data from Steiner CA, Bass EB, Talamini MA, et al. Surgical rates and operative mortality for open and laparoscopic cholecystectomy in Maryland. N Engl J Med 1994; 330:403-8.

randomized trials from other countries support the contention that the laparoscopic approach is superior to the open approach.59,88-92 In these analyses, the main benefits of the laparoscopic approach have included a shortened hospital stay, decreased pain, reduced disability, and reduced costs. Population studies have shown a substantial decline in cholecystectomy-related mortality rates following the introduction of the laparoscopic technique (Table 66-5).93 Against the perceived benefits of laparoscopic cholecystectomy over the open approach is concern about unacceptably high complication rates, especially for bile duct injury. Although the exact frequency of bile duct injury around the world is not known, two lines of evidence suggest that the rate has declined. First, regional studies have demonstrated a decrease in the rate of bile duct injury as overall experience with laparoscopic cholecystectomy has increased (Fig. 66-5).94 Curiously, however, the frequency of bile duct injury does not continue to fall with increasing experience of the individual surgeon, but rather plateaus.95-97 Although duct injuries are more common early in an individual surgeon’s experience, they still occur in the hands of seasoned surgeons, albeit at a lower rate. As overall experience has increased, the rate of bile duct injury for laparoscopic cholecystectomy has approximated that seen with open cholecystectomy. Second, the number of patients with bile duct injury treated at tertiary referral medical centers has declined since the early days of laparoscopic cholecystectomy.98 Introduction of laparoscopic cholecystectomy in

Chapter 66  Treatment of Gallstone Disease of stones. Although oral dissolution therapy is effective in treating selected patients, the low morbidity rate associated with laparoscopic cholecystectomy may negate any potential advantages of nonsurgical treatment. In addition, nonsurgical treatment is less desirable because of the duration of treatment and high likelihood of recurrent stones. Patient selection also has a significant bearing on the choice of treatment. Some patients are not candidates for cholecystectomy or general anesthesia or choose not to have surgery. Only in these patients should medical treatment be offered. All other patients should be encouraged to undergo laparoscopic cholecystectomy for symptomatic gallbladder pain. For patients with complications of gallstones, laparoscopic cholecystectomy with cholangiography is the preferred treatment; open cholecystectomy is reserved for patients who are not candidates for a successful and safe open cholecystectomy. The emergence of natural orifice translumenal endoscopic surgery, or NOTES, may offer additional options for cholecystectomy in the future.

1.8 1.6

Bile duct injury (%)

1.4 1.2 1 0.8 0.6 0.4 0.2 0 0

5

10

15

20

30

50

Number of cases Figure 66-5.  Effect of a surgeon’s experience on the risk of bile duct injury during laparoscopic cholecystectomy. The dramatic decline in risk as experience is gained has been attributed to a “learning curve.” (Data adapted from Moore MJ, Bennett CL. The learning curve for laparoscopic cholecystectomy. The Southern Surgeons Club. Am J Surg 1995; 170:55-9.)

the United States was rapid and may have exceeded the capability of the medical educational system to train all practitioners adequately. The initial relatively high rates of bile duct injury have been ascribed to a “learning curve.” This experience is a cautionary example for other new technologies that may be introduced into medical practice. Concern has been raised about the increased use of laparoscopic cholecystectomy for gallstone disease when compared with historical rates for open cholecystectomy. In a defined health maintenance organization population in Pennsylvania, the rate of cholecystectomy increased from 1.35 per 1000 enrollees in 1988, just before the introduction of the laparoscopic approach, to 2.15 per 1000 enrollees in 1992, just after its introduction.99 No significant changes in the rates of herniorrhaphy or appendectomy were observed. Similarly, statewide data from Maryland have shown that the rate of cholecystectomy rose from 1.69 per 1000 residents in 1987 to 1989 to 2.17 per 1000 residents in 1992,93 and Scottish nationwide data have shown a 20% increase in the age-standardized rate of cholecystectomy.100 The reasons for the increases in use are not yet clear. The consensus of experts in the field is that selection of patients for cholecystectomy should not be altered by the availability of the laparoscopic approach.

CHOICE OF TREATMENT Several factors influence the choice of treatment for symptomatic gallstone disease. These factors include the stage of gallstone disease, characteristics of the stone and gallbladder function, and preference of the patient. For patients with uncomplicated biliary pain, treatment options include surgery and oral dissolution with or without lithotripsy. Surgery has the advantage of dealing with the underlying cause of gallstones regardless of the number, size, and type

INDICATIONS FOR TREATMENT ASYMPTOMATIC GALLSTONES

Decisions regarding the management of the patient with asymptomatic gallstones must be predicated on knowledge of the natural history of the condition, as discussed in Chapter 65. In general, patients with asymptomatic gallstones should be reassured that life-threatening complications are uncommon and that symptoms related to the stones develop in only a minority of patients.101-105 In the event that an asymptomatic patient becomes symptomatic, the initial presentation is most often with uncomplicated biliary pain. In fact, most patients in whom complications of gallstones develop have antecedent biliary pain.106 Decision analysis calculations suggest that the risks of cholecystectomy approximate the potential benefit in preventing future serious sequelae of gallstones.107 These calculations were based on historical data regarding the outcome of open cholecystectomy; the rate of serious sequelae of gallstones was determined from long-term follow-up of a group of male faculty members at a major mid-western university. Whether these data are applicable to the more common female patient considering laparoscopic cholecystectomy today is not known. Nevertheless, the strategy of prophylactic cholecystectomy in all asymptomatic patients probably has no major advantage over the recommendation that cholecystectomy be limited to symptomatic patients.108,109 In addition, studies that have analyzed health-related quality of life do not support cholecystectomy for asymptomatic patients.110 In certain subgroups, the benefits of prophylactic cholecystectomy for asymptomatic gallstones may outweigh the risks. American Indians, for example, appear to have a rate of gallstone-associated gallbladder cancer high enough to justify prophylactic cholecystectomy.110 In morbidly obese persons as well as recipients of heart and lung transplants, complications of gallstone disease carry high morbidity rate, and prophylactic cholecystectomy may be indicated.111-113 Curiously, renal transplant patients with asymptomatic gallstones have a low risk of complications related to gallstone disease and, therefore, should not be considered for prophylactic cholecystectomy (see later).114,115 The risks of complications of gallstone disease in children may outweigh the risk of cholecystectomy (see later). Diabetic persons have been thought to be particularly prone to gallstone formation and to complications from the stones. Morbidity and mortality rates for diabetic patients

1129

1130

Section VIII  Biliary Tract who undergo emergency operations for complications of gallstone disease have also been thought to be excessive. These perceptions have not been borne out when confounding variables, such as hyperlipidemia, obesity, cardiovascular disease, and chronic kidney disease, are taken into account.116 Therefore, prophylactic cholecystectomy in an asymptomatic diabetic patient with gallstones does not appear to be warranted.117 Data do support, however, early intervention in diabetic patients in whom symptoms develop because these patients are at an increased risk of developing gangrenous cholecystitis.118 Therefore, the severity of complications may be higher when complications of gallstones arise in these patients.

BILIARY PAIN Patient Selection

The majority of operations for biliary tract disorders are performed to relieve symptoms related to intermittent obstruction of the cystic duct by gallstones. This constellation of symptoms, including intermittent epigastric or right upper quadrant pain, nausea, and vomiting has been termed biliary pain (“biliary colic” in the past) (see Chapter 65). Histologically, gallbladders from patients suffering from repeated attacks of biliary pain usually, but not always, show fibrosis and mononuclear cell infiltration that are characteristic of chronic cholecystitis. Furthermore, patients with biliary pain are more likely than patients with asymptomatic stones to experience complications of gallstones. Cholecystectomy is indicated in these symptomatic patients. As with any operation, the potential benefits in terms of relief of symptoms and prevention of future complications must be weighed against the risk of surgery. Fortunately, the physiologic stress of cholecystectomy is minimal, and the operation may be undertaken safely even in the elderly and infirm. In the poorly compensated cirrhotic patient, the risk of cholecystectomy is substantially higher.118-120 Operation in this setting is justified only if the symptoms are severe, complications arise, or the cirrhosis is well com­ pensated.121,122 When cholecystectomy is performed for biliary pain, routine perioperative antibiotics are seldom indicated.123

Diagnostic Evaluation

The diagnosis of biliary pain is generally suspected from the clinical history (see Chapter 65). Few important findings specific to gallstone disease are elicited on physical examination. Symptoms and signs of heart disease, especially congestive heart failure, should be detected during the preoperative evaluation because heart disease adversely affects the risk of surgery. Few preoperative laboratory tests are routinely necessary; liver biochemical tests should be

obtained to screen for unsuspected choledocholithiasis. Imaging evaluation may be limited to ultrasonography in most patients with biliary pain. Ultrasonography has a high sensitivity (95%) and specificity (98%) in this setting and is also useful for detecting gallbladder inflammation— which is suggested by thickening of the gallbladder wall and pericholecystic fluid—and dilatation of the bile ducts. Ancillary tests, including oral cholecystography, magnetic resonance cholangiopancreatography (MRCP), endoscopic retrograde cholangiography (ERCP), or cholecystokinin scintigraphy, are useful for confirming the diagnosis in the unusual patient in whom gallstones are suspected but ultrasonography is negative. In patients with atypical symptoms, endoscopy, upper gastrointestinal barium radiography, or both, may be performed to exclude other disorders such as esophagitis or peptic ulcer disease.

ACUTE CHOLECYSTITIS

Management of the patient with acute cholecystitis begins with intravenous hydration and restoration of tissue perfusion and electrolyte balance. Vomiting is prominent in most patients, and nasogastric suction may be required. Intravenous antibiotics are indicated because bile or gallbladder wall cultures are positive for bacteria in more than 40% of patients.124 A cephalosporin such as cefoxitin is satisfactory for mildly ill patients, but in more severe cases, broadspectrum antibiotics, such as piperacillin-tazobactam or a third-generation cephalosporin with metronidazole, should be given. If gangrenous or emphysematous cholecystitis is suspected, an agent effective against anaerobic organisms should be included. Subsequent management depends on the certainty of the diagnosis, severity of the attack, and general condition of the patient. If cholecystitis is severe and complications such as perforation appear imminent, cholecystectomy should be undertaken urgently. If the nature of the symptoms is uncertain, surgery may be indicated to establish the diagnosis. Conversely, the elderly patient with concurrent illnesses such as congestive heart failure may benefit from an initial nonoperative approach. In the past, the timing of cholecystectomy for the typical patient with acute cholecystitis was controversial. Multiple prospective, randomized, controlled clinical trials have compared the strategies of early (within days of presentation) and delayed (after six to eight weeks) surgery for acute cholecystitis (Table 66-6).89,125-131 A meta-analysis of these trials has shown that, for the average patient, early operation is preferable because the total length of hospitalization and costs are reduced, morbidity is less, and deaths related to progressive acute cholecystitis are prevented.130 Early operation does not appear to increase the major risks of cholecystectomy, such as bile duct injury, substantially.

Table 66-6  Early versus Delayed Cholecystectomy for Acute Cholecystitis: Combined Results from Seven Randomized Trials TIMING OF CHOLECYSTECTOMY Early† Delayed‡

NO. OF PATIENTS

MORTALITY RATE (%)

BILE DUCT INJURIES

TOTAL MEAN HOSPITAL STAY (DAYS)

FAILURE OF REGIMEN*

378 364

0 2.0

0 0

9.6 17.8

NA 26%

*Failure is defined as worsening acute symptoms requiring early surgery. † Within days of presentation. ‡ After 6-8 weeks. NA, not applicable. Data from references 89, 125-129, and Lai PB, Kwong KH, Leung KL, et al. Randomized trial of early versus delayed laparoscopic cholecystectomy for acute cholecystitis. Br J Surg 1998; 85:764-7.

Chapter 66  Treatment of Gallstone Disease Although acute cholecystitis was initially considered to be a contraindication to laparoscopic cholecystectomy, it now is clear that laparoscopic cholecystectomy is feasible in the majority of cases. Technical problems are encountered occasionally in patients with a gangrenous gallbladder, coagulopathy, or severe inflammation that obscures the structures of the hepatocystic triangle. In these settings, an open approach may be necessary. Cholangiography is particularly valuable in patients with acute cholecystitis to confirm the ductal anatomy. The benefits of laparoscopic cholecystectomy in patients with biliary pain, including decreased incisional pain, shortened hospital stay, and more rapid return to work, also apply to patients with acute cholecystitis. For the high-risk patient with severe concurrent illnesses, such as liver, pulmonary, or heart failure, cholecystostomy (gallbladder drainage) is preferable to cholecystectomy. Operative cholecystostomy has been superseded by a percutaneous approach in the majority of patients. After the patient has recovered from the attack of acute cholecystitis, laparoscopic cholecystectomy should be performed if the patient’s overall condition permits. Alternatively, residual stones can be removed via the cholecystostomy tube, and the patient may be managed expectantly. Recurrent biliary symptoms develop in approximately one half of all patients treated with a cholecystostomy. An example of a patient best managed by percutaneous cholecystostomy is shown in Figure 66-6. The indication for a percutaneous cholecystostomy is the patient’s high surgical risk rather than the severity of the acute cholecystitis or appearance of the gallbladder on an imaging study. Prompt laparoscopic cholecystectomy is the preferred treatment for acute cholecystitis. Acute cholecystitis in diabetic patients is associated with a significantly higher frequency of infectious complications such as sepsis, compared with nondiabetic patients.132,133 Cholecystectomy should be performed expeditiously in this group of patients. Similarly, acute cholecystitis in the elderly patients population may have a deceptively benign clinical presentation but is associated with a high rate of empyema, gangrene, and perforation.134-136 Factors asso­ ciated with gangrenous or emphysematous cholecystitis include male gender, diabetes mellitus, cardiovascular disease, and an initial white blood cell count in excess of

A

15,000/mm3. As with diabetic patients, early cholecystectomy is warranted in elderly patients to ensure prompt control of infection. The routine use of surgical drainage catheters after laparoscopic cholecystectomy for acute cholecystitis is not warranted and may be deleterious.137

Acalculous Cholecystitis

Acute cholecystitis that occurs in the absence of gallstones is termed acalculous cholecystitis (see Chapter 67). Most commonly, acalculous cholecystitis occurs in a patient hospitalized for other serious illnesses, such as trauma, burns, or major surgery. It may develop in outpatients, among whom elderly male patients with peripheral vascular disease appear to be at highest risk.138 Acalculous cholecystitis may also complicate the treatment of patients with the acquired immunodeficiency syndrome.139 The pathophysiology of acalculous cholecystitis is unclear, but biliary stasis caused by fasting, alterations in gallbladder blood flow, activation of factor XII, prostaglandins, and endotoxin all may play roles (see Chapter 67).140-144 Sludge is generally present in the gallbladder and may obstruct the cystic duct. Gangrene, empyema, and per­ foration of the gallbladder complicate the course of acal­ culous cholecystitis more commonly than they complicate the course of acute cholecystitis caused by gallstones. In some series, the frequency of these complications approaches 75%.145 Cholecystectomy has been the mainstay of therapy for acalculous cholecystitis. Prompt removal of the gallbladder is particularly important when gangrene or empyema is suspected and when perforation is imminent. In some patients, however, the risk of surgery is high because of the severity of their underlying illness. These patients may be managed initially with placement of a percutaneous tube cholecystostomy under ultrasound guidance. Most patients treated with tube cholecystostomy recover. Those in whom evidence of intra-abdominal sepsis develops or persistent obstruction of the cystic duct is seen on cholangiography require cholecystectomy.

Emphysematous Cholecystitis

Emphysematous cholecystitis is an uncommon condition characterized by infection of the gallbladder wall by gasforming bacteria, particularly anaerobes (see Chapter 65).

B

Figure 66-6.  Imaging studies in a 47-year-old woman with severe acute cholecystitis complicating a prolonged intensive care unit stay for multisystem organ failure after surgery for a perforated viscus. A, Computed tomography scan showing acute cholecystitis with gallbladder wall thickening, pericholecystic fluid, and stones (arrow). B, Percutaneous cholecystostomy (small arrow) showing a gallstone impacted at the neck of the gallbladder (large arrow). A cholecystostomy tube was left in place and the patient improved clinically.

1131

1132

Section VIII  Biliary Tract Diabetes mellitus has been cited as a risk factor.118,146 Gangrene and perforation commonly complicate the course of emphysematous cholecystitis. The treatment of emphysematous cholecystitis is prompt laparoscopic cholecystectomy after restoration of fluid and electrolyte balance. Antibiotics are indicated, with coverage directed against gram-negative rods and anaerobic bacteria.

GALLSTONE PANCREATITIS

The pathophysiology and clinical presentation of patients with gallstone pancreatitis are discussed in Chapters 58 and 65. Initial management of patients with gallstone pancreatitis includes fluid resuscitation, bowel rest, and monitoring for complications. The majority of patients have a relatively mild illness that resolves clinically within one week with conservative management. The presence of cholelithiasis should be determined by ultrasonography early in the course of the treatment of a patient with acute pancreatitis. If cholelithiasis is present, laparoscopic cholecystectomy generally should be performed prior to the patient’s discharge. In the past, cholecystectomy early in the course of gallstone pancreatitis carried significant risk. For that reason, the timing of cholecystectomy was delayed for one to two months to allow resolution of the inflammatory process. A major disadvantage of this delayed approach was that up to one half of patients had further attacks of pancreatitis during the observation period. It is now recognized that cholecystectomy may be performed safely during the same hospitalization when the clinical signs of pancreatitis have resolved.147-149 This approach shortens the total duration of illness and hospitalization.150 Additionally, it prevents recurrent pancreatitis. Cholangiography should be performed during the cholecystectomy to exclude residual bile duct stones, as recommended by the International Association of Pancreatology.151 In patients with severe or necrotizing pancreatitis, surgery is delayed for several weeks to allow (1) patients to recover from the sequelae of pancreatitis; (2) inflammation of the hepatoduodenal ligament to decrease, thereby permitting safe dissection; and (3) identification of the small subset of patients in whom pancreatic pseudocysts develop and may require additional surgical treatment. A small subset of patients with necrotizing gallstone pancreatitis appear to benefit from early endoscopic sphincterotomy and clearance of the bile duct (see Chapter 58).152-154 Bile duct stones are found in a substantial proportion of patients with severe gallstone pancreatitis when sphincterotomy is performed within the first 24 to 48 hours of hospitalization. The morbidity of this approach is less than that for early surgery with bile duct exploration. Most patients who undergo endoscopic sphincterotomy for gallstone pancreatitis should have elective cholecystectomy when the pancreatitis has subsided. In elderly patients or those at high risk for surgery, cholecystectomy may be deferred, but further symptoms of gallstone disease may be expected in up to 31% on shortterm follow-up.150 The risk of symptomatic gallstone disease is even higher in patients with cystic duct obstruction on cholangiography.155

SPECIAL PROBLEMS Gallstone Disease during Pregnancy

Occasionally, gallbladder disease is first noted or becomes more troublesome during pregnancy. The most common clinical presentations in this setting are worsening biliary pain and acute cholecystitis. Jaundice and acute pancreatitis caused by choledocholithiasis are rare. Imaging evalua-

tion of symptoms suggestive of biliary tract disease can nearly always be limited to ultrasonography. The potential teratogenic effects of conventional radiography and radionuclide scanning make these techniques unjustified in the pregnant patient. Therapeutic options for gallstone disease are limited in pregnancy. In the past, cholecystectomy during pregnancy was discouraged because of the fear of fetal loss. Complications such as spontaneous abortion and preterm labor were common in operated women in the first and third trimesters of gestation, respectively. In addition, pregnancy was formerly considered an absolute contraindication to laparoscopic surgery because of concern about potential trocar injury to the uterus and the unknown effects of pneumoperitoneum on the fetal circulation. Improvements in anesthesia and tocolytic agents appear to have made abdominal surgery safer during pregnancy. Several large case series have suggested that cholecystectomy may be undertaken during pregnancy with minimal fetal and maternal morbidity.156,157 Laparoscopic cholecystectomy is performed during pregnancy only when necessary. Indications include complicated gallstone disease, including acute cholecystitis and pancreatitis, when the underlying disease poses a threat to the pregnancy, and severe biliary pain, when the mother is unable to maintain adequate nutrition. Surgery is probably safest during the second trimester, when the risk of fetal loss and teratogenicity that may occur in the first trimester and the risk of preterm labor that may occur in the third trimester are both low.156,157

Gallstone Disease during Childhood

Gallstone disease in the pediatric population appears to be increasing in frequency. Chronic hemolysis leading to pigment gallstones is the cause in about 20% of patients.158 A history of prolonged fasting with total parenteral nutritional support is an increasingly important risk factor. Ileal disorders or previous bowel resection increase the risk of gallstone development. Management of childhood cholelithiasis must take into account the type of stone (pigment or cholesterol), presence or absence of symptoms, and underlying factors such as total parenteral nutrition. Cholecystectomy is indicated in all symptomatic patients with gallstones. The management of asymptomatic gallstones is less clear. Gallstones in infants who are receiving total parenteral nutrition occasionally resolve following reinstitution of oral feedings. Therefore, observing the asymptomatic infant in this setting for up to 12 months seems reasonable. Persistent gallstones and asymptomatic pigment stones (which do not resolve spontaneously) are best treated with laparoscopic cholecystectomy.

Mirizzi’s Syndrome

Mirizzi’s syndrome refers to common hepatic duct obstruction resulting from compression by a gallstone impacted in the cystic duct. Two types of Mirizzi’s syndrome have been described.159 In type I, the hepatic duct is compressed by a large stone impacted in the cystic duct or Hartman’s pouch. Associated inflammation may contribute to the obstruction and formation of a stricture in the central section of the extrahepatic bile duct. In type II, the calculus has eroded into the hepatic duct to produce a cholecystocholedochal fistula. Mirizzi’s syndrome is rare, occurring in about 1% of all patients who undergo cholecystectomy. Most patients present with repeated bouts of pain, fever, and jaundice. Ultrasonography generally reveals gallstones with a contracted gallbladder and moderate intrahepatic ductal dilatation with normal extrahepatic biliary anatomy.

Chapter 66  Treatment of Gallstone Disease ERCP is useful in delineating the hepatic duct anatomy. The appearance of the obstruction and surrounding inflammation may be confused with a Klatskin tumor (see Chapter 69). The possibility of Mirizzi’s syndrome should be considered during a difficult cholecystectomy to reduce the likelihood of hepatic duct injury. Management of type I Mirizzi’s syndrome includes cholecystectomy with or without bile duct exploration. In the presence of severe inflammation, in which identification of the anatomy is difficult, partial cholecystectomy with postoperative endoscopic sphincterotomy to ensure clearance of bile duct stones is preferable. Management of type II Mirizzi’s syndrome is best accomplished by partial cholecystectomy and cholecystocho­ ledochoduodenostomy. Frozen section examination of the gallbladder wall may be necessary to exclude a carcinoma. Rarely, a Roux-en-Y hepaticojejunostomy is required to repair a large defect in the common hepatic duct.

Gallstone Ileus

Gallstone ileus is an uncommon form of bowel obstruction caused by impaction of a large gallstone in the intestinal lumen. Bouveret’s syndrome refers to impaction of a gallstone in the distal duodenum or at the pylorus with resulting symptoms of gastric outlet obstruction. Gallstone ileus represents a true mechanical obstruction rather than a defect in motility as the name “ileus” would suggest. The median age of affected patients is more than 70 years. Most are women. Gallstone ileus is the cause of intestinal obstruction in less than 1% of patients younger than 70 years in age but nearly 5% of those 70 years of age or older.159 Symptoms are typical of mechanical intestinal obstruction and include cramping abdominal pain, vomiting, and abdominal distention. Only a minority of patients have symptoms suggestive of acute cholecystitis, but one half are known to have a history of gallstones.160 Liver biochemical test levels are elevated in 40% of patients, but overt jaundice is rare. Plain abdominal films reveal an intestinal gas pattern compatible with intestinal obstruction in most patients. Pneumobilia is present in about one half of all patients, and the aberrant gallstone is visible in a minority. Upper or lower gastrointestinal barium studies may occasionally identify the site of obstruction or the fistula, but these tests are unnecessary in most cases. Ultrasonography is useful for confirming the presence of cholelithiasis and may allow visualization of the fistula. The pathophysiology of gallstone ileus involves erosion of a gallstone, generally over 2.5 cm in diameter, into the intestinal lumen via a cholecystoenteric fistula. Most commonly, the fistula occurs in the duodenum and, less often, the colon. As the gallstone is passed down the length of the intestine, it obstructs the lumen intermittently. Characteristically, complete obstruction occurs in the ileum, where the lumen is narrowest. The obstruction has been described as “tumbling,” because the symptoms wax and wane during the passage of the stone. Management should be directed initially at restoration of fluid and electrolyte balance, followed by exploratory laparotomy. A laparoscopic approach is technically feasible and effective. Removing the stone via a small enterotomy relieves the intestinal obstruction. A search should be made for additional stones. Bowel resection is necessary only when perforation or intestinal ischemia has occurred. Treatment of the cholecystoenteric fistula is not necessary at the initial operation because many fistulas close spontaneously.161 Elective cholecystectomy and closure of the fistula are indicated if symptoms of chronic cholecystitis persist. Mortality rates in this high-risk patient population are

high, averaging 15% to 18%. Gallstone ileus recurs in about 5% of patients.

Incidental Cholecystectomy

Occasionally, gallstones are identified unexpectedly before or during another operation. When this happens, incidental cholecystectomy should be considered at the time of the original, planned, procedure. The rationale for incidental cholecystectomy is to prevent later development of symptomatic gallstone disease, including early postoperative acute cholecystitis. As expected, addition of a cholecystectomy increases the risk of postoperative complications. The decision to proceed with an incidental cholecystectomy is based on an assessment of the expected benefits and risks. Some data are available to assist in quantifying these factors. The typical patient with gallstones tends to remain asymptomatic. On long-term follow-up, symptoms develop in 18% to 35% of these initially asymptomatic per­ sons.101,107,108 Certain groups, however, are at higher risk. Gallstones in morbidly obese patients, for example, tend to have a more aggressive natural history, making incidental cholecystectomy at the time of gastric bypass surgery appealing.162 Patients with large (>2.5 cm) gallstones and those with calcification of the gallbladder wall (porcelain gallbladder) have an increased risk for the development of acute cholecystitis and gallbladder cancer, and in these patients, incidental cholecystectomy is warranted.163-165 In patients with sickle cell disease, who are at risk for the development of pigment gallstones because of chronic hemolysis, distinguishing the clinical presentation of a sickle cell crisis from acute cholecystitis may be difficult, and incidental cholecystectomy is indicated in these patients.166,167 Similarly, patients with other hemolytic anemias, such as b-thalassemia, are at high risk for the development of gallstones, and a high percentage of them become symptomatic.168 Cholecystectomy appears warranted for asymptomatic patients with stones if splenectomy is undertaken for the hemolytic anemia. Finally, laparotomy for reasons other than cholecystectomy is associated with a high frequency of postoperative biliary symptoms if the gallbladder containing stones is left in situ. Of 68 asymptomatic patients with stones who underwent laparotomy in one study, 54% became symptomatic postoperatively, and 22% required cholecystectomy within 30 days.169,170 The risk of adding incidental cholecystectomy to another abdominal procedure appears to be low.171 If the patient is in otherwise reasonable health, the primary operation has proceeded smoothly, and operative exposure is adequate, incidental cholecystectomy can be done safely at the time of another operation, including colectomy. The risk does not appear to be increased in the elderly. The risk of postoperative wound infections, however, may be increased in some cases by the addition of an incidental cholecystectomy.172

CHOLEDOCHOLITHIASIS Choledocholithiasis may be detected at the same time that gallbladder stones are discovered during an evaluation for biliary tract symptoms, during cholecystectomy, or after a cholecystectomy. Several management options are available, including oral dissolution therapy, interventional radiologic and endoscopic techniques, and surgery (see Chapter 70). Which management strategy is most appropri-

1133

1134

Section VIII  Biliary Tract ate for a given patient depends on the clinical situation in which the stones have been identified (jaundice, cholangitis, pancreatitis, or absence of symptoms), status of the gallbladder, and age and general condition of the patient. Additional factors to consider are the expertise of the available surgical, endoscopic, and radiologic specialists.

CHOLEDOCHOLITHIASIS KNOWN PREOPERATIVELY

When choledocholithiasis is known to be present preoperatively, an acceptable approach is to clear the bile duct by endoscopic sphincterotomy and then proceed with laparoscopic cholecystectomy. An alternative approach is an open or laparoscopic cholecystectomy with bile duct exploration. In the era of laparoscopic cholecystectomy and endoscopic stone retrieval, laparoscopic bile duct exploration has been performed less commonly than open bile duct exploration was performed during the open cholecystectomy era.173 Laparoscopic bile duct exploration via either a transcystic duct approach or direct incision of the bile duct is a technically demanding procedure. When performed by an experienced surgeon, however, the success rate for clearing the duct of stones ranges from 83% to 97%.174-176 Decision analyses and randomized trials have shown that laparoscopic cholecystectomy with bile duct exploration via either a transcystic or transcholedochal approach results in lower rates of morbidity and mortality as well as a shorter hospital stay, quicker return to health, and lower number of procedures than preoperative endoscopic retrieval followed by laparoscopic cholecystectomy.174,175 The decision to proceed with the surgery-only approach depends greatly on the experience and technical skill of the surgeon and his or her team. If a surgeon lacks adequate training or adequately trained staff and equipment, a two-stage approach is preferable.

CHOLEDOCHOLITHIASIS IDENTIFIED DURING CHOLECYSTECTOMY

If unsuspected choledocholithiasis is identified by cholangiography during laparoscopic cholecystectomy, the following three options are available: (1) conversion to an open operation with bile duct exploration; (2) laparoscopic bile duct exploration; and (3) completion of the laparoscopic cholecystectomy with postoperative endoscopic sphincterotomy and stone extraction. An algorithm of these options is shown in Figure 66-7. Factors that influence this decision include the number and location of bile duct stones, the presence of associated ductal pathology, and the skill and experience of the surgeon and endoscopist. Completion of the laparoscopic cholecystectomy with postoperative endoscopic sphincterotomy is satisfactory for most patients and has the advantage of preserving the minimally invasive approach. Endoscopic sphincterotomy may be technically unsuccessful, however, in 5% to 10% of patients—even in the hands of a skilled endoscopist—and complete clearance of stones from the bile duct is possible in only 70% to 80% of patients.177 In such patients, a second attempt may be required. Increasing experience has shown that laparoscopic bile duct exploration is safe and effective. Stone clearance rates average 95%, with an operative mortality rate of 0.5%.174,177-179 Laparoscopic bile duct exploration compares favorably with endoscopic sphincterotomy in terms of efficacy, cost, and safety.180 As with open bile duct exploration, laparoscopic bile duct exploration is done preferably through a trancystic approach, in which the bile duct is explored and stones are removed

Unsuspected BD stone

Laparoscopic choledochoscopy

Unsuccessful

Successful

Complete the laparoscopic cholecystectomy

Convert to open chole and BDE

Lap chole + postop ERCP/ES

Successful

Unsuccessful

Reoperate for BDE Figure 66-7.  A suggested algorithm for the management of bile duct (BD) stones found unexpectedly during laparoscopic cholecystectomy (lap chole). BDE, bile duct exploration; ERCP/ES, endoscopic retrograde cholangiopancreatography/endoscopic sphincterotomy.

via the cystic duct. On rare occasions in which the stones are too large or located above the insertion of the cystic duct into the bile duct, a transcholedochal approach is needed. This approach involves exploration of the bile duct and removal of stones via an incision directly into the bile duct. The incision in the bile duct must be closed with sutures, often over a T-tube to prevent stricturing of the bile duct, especially when the duct is small and suture closure in the absence of a T-tube may lead to a stricture that is difficult to dilate. Therefore, a small bile duct is a contraindication to direct transcholedochal exploration. T-tubes are also left in the bile duct when evacuation of stones is incomplete to allow biliary decompression and percutaneous stone extraction by an interventional radiologist.

CHOLEDOCHOLITHIASIS IDENTIFIED AFTER CHOLECYSTECTOMY

Choledocholithiasis identified in patients who previously have undergone cholecystectomy is best managed with endoscopic sphincterotomy and stone extraction. If a T-tube is still present from a recent bile duct exploration, radiologic extraction of the stone via the T-tube tract is usually possible. Surgery is rarely required in this situation, but if required, an open bile duct exploration is usually performed because access to the bile duct via the cystic duct is lost after cholecystectomy.

BILE DUCT STRICTURE The majority of benign bile duct strictures are the result of iatrogenic injury during cholecystectomy. A minority are the sequelae of chronic pancreatitis, primary sclerosing cholangitis, trauma, liver transplantation, and choledocho-

Chapter 66  Treatment of Gallstone Disease lithiasis. Operative injury to the bile duct during cholecystectomy may occur because of misinterpretation of the biliary ductal anatomy; inaccurate placement of clips, sutures, or cautery to control hemorrhage; tenting of the bile duct during control of the cystic duct; and ineffective retraction and exposure.181-183 These injuries commonly occur during an otherwise uneventful cholecystectomy and may be unnoticed by the surgeon. Bile duct injury presents in one of three patterns. If the bile duct has been completely occluded, jaundice develops rapidly in the early postoperative period after cholecystectomy. In the second pattern, the injury is manifested by the development of bile ascites that results from transection of an extrahepatic bile duct, ineffective placement or dislodgement of cystic duct ligatures, or a bile leak from the gallbladder fossa as a result of a divided cystohepatic duct or duct of Luschka. A bile leak is often associated with an infected bile collection in the subhepatic space. In the third pattern, partial bile duct obstruction leads to intermittent episodes of pain, jaundice, or cholangitis, usually within two years of the cholecystectomy. The three patterns of injury are illustrated in Figure 66-8. In the early postoperative period following laparoscopic cholecystectomy, the clinician should suspect the possibility of bile duct injury in any patient with persistent abdominal pain. The differential diagnosis of cholangitis in a patient with a history of cholecystectomy consists mainly of bile duct stricture and choledocholithiasis. Stricture and choledocholithiasis may be difficult to differentiate on clinical grounds because the symptoms, signs, and liver biochemical test levels may be identical. The imaging evaluation of a patient with suspected bile duct stricture usually begins with ultrasonography to identify dilated ducts or a subhepatic fluid collection. In the early postoperative period, a technetium-labeled radionuclide scan may expeditiously and noninvasively demonstrate patency of the biliary tree and exclude a bile leak. If these studies suggest bile duct injury, ERCP is indicated to define the lesion. The initial goals of management include control of subhepatic infection, usually via percutaneous drainage of any fluid collection, and biliary drainage, either via an endoscopic or transhepatic route (see Chapter 70). Following control of infection and biliary drainage, complete cholangiography, either endoscopic or radiologic, is necessary to define the anatomy and plan reconstruction. Most patients with a benign postoperative biliary stricture are best managed with surgical repair. Although numerous operations have been described, the best results are obtained with resection of the stricture and an end-to-side Roux-en-Y choledochojejunostomy or hepaticojejunostomy. The principles of a successful repair include complete dissection of the strictured segment, creation of a tension-free anastomosis, accurate mucosa-to-mucosa approximation of the anastomosis with fine absorbable suture material and unscarred proximal ductal tissue, and preservation of the ductal blood supply. The mortality rate of operations to correct benign biliary strictures averages 0% to 2% in modern series.184-186 The risk of surgery is related directly to the presence of risk factors such as cirrhosis, renal failure, uncontrolled cholangitis, age, and malnutrition. The long-term results of biliary reconstruction for a benign bile duct stricture are good, with cure achieved in 85% to 98% of patients.184-187 Results are worse in patients with high strictures or cirrhosis. In those with high strictures, special techniques may be necessary to obtain healthy ductal tissue that is uninvolved in the inflammatory process for anastomosis; liver resection may be required. Recurrent strictures pose technical difficulties, but

satisfactory results are still achieved in about 75% of patients.184-186,188 Postoperative strictures may be treated with endoscopic or percutaneous balloon dilation with or without stent placement as long as the remaining duct has not been disrupted. Benign postoperative strictures often can be managed endoscopically with placement of plastic, removable stents. Although several endoscopic procedures are often required, good results can be achieved in appropriately selected patients (see Chapter 70). No randomized, prospective trials have compared surgical, endoscopic, and radiologic approaches. In one nonrandomized trial, long-term bile duct patency was achieved in 88% of patients treated with hepaticojejunostomy compared with 55% of those who underwent balloon dilation.189 No procedure-related mor­ tality was observed. In view of the excellent long-term results and low mortality rate of hepaticojejunstomy in experienced hands, surgery should be offered as the initial treatment to all fit patients with bile duct stricture. Non­ operative management is best reserved for patients with biliary cirrhosis, significant comorbid illness, or high recurrent strictures.

POSTCHOLECYSTECTOMY SYNDROME Postcholecystectomy syndrome refers to the occurrence of abdominal symptoms that often resemble biliary pain following cholecystectomy (see also Chapters 63 and 65). The term is misleading in that it encompasses a wide spectrum of biliary and non-biliary disorders that are rarely related to the operation itself. The frequency of such symptoms following cholecystectomy ranges from 5% to 40%.51,190-193 The most common postoperative symptoms are dyspepsia, flatulence, and bloating, which usually antedate the cholecystectomy. Other patients have persistence of right upper quadrant or epigastric abdominal pain. A small percentage of patients with postcholecystectomy symptoms present with severe abdominal pain, jaundice, or emesis; investigation is much more likely to reveal a distinct, treatable cause in this group of patients than in those with mild or nonspecific symptoms. If the symptoms arise early in the post­ operative period, bile peritonitis secondary to iatrogenic biliary injury must be suspected. The differential diagnosis of symptoms after cholecystectomy includes extraintestinal disorders such as cardiac ischemia, nonbiliary gastrointestinal conditions such as peptic ulcer disease, biliary disorders such as choledocholithiasis, functional illnesses such as irritable bowel syndrome, and psychiatric diseases. The spectrum of possibilities is listed in Table 66-7. The clinician must carefully consider the possibility of nonbiliary causes of pain and direct the evaluation appropriately. The most common biliary causes are described in the sections that follow.

CHOLEDOCHOLITHIASIS

Bile duct stones are the most common cause of postcholecystectomy symptoms. They may be residual stones overlooked at the time of cholecystectomy or, less frequently, stones that have formed primarily in the duct. The natural history of choledocholithiasis is not known, but in some patients, these stones clearly can cause biliary-type pain, jaundice, pancreatitis, or cholangitis. The diagnosis of choledocholithiasis is suggested by the clinical picture. Liver biochemical test values, particularly the alkaline phosphatase level, may be elevated. Ultrasonography may show a dilated bile duct, but visualization of the stone is

1135

1136

Section VIII  Biliary Tract

A

B

C

D

E

F

Figure 66-8.  Common patterns of bile duct injury during laparoscopic cholecystectomy. A and B, With the first pattern, the patient may present with a biloma or bile ascites from a cystic duct stump leak (arrow in A) or a bile leak from a duct of Luschka (arrow in B). C and D, With the second pattern, the patient presents with jaundice, with or without a bile leak, as a result of excision of the bile duct secondary to misinterpretation of the bile duct for the cystic duct. These problems usually involve injury to the confluence of the hepatic ducts and to the right hepatic artery as well. C, An endoscopic retrograde cholangiopancreatogram (ERCP) showing surgical clips occluding the bile duct (arrow); D, the corresponding transhepatic cholangiogram demonstrating excision of the hepatic duct confluence (arrow). E and F, With the third pattern, the patient presents with jaundice caused by a stricture resulting either from a surgical clip placed on the bile duct instead of the cystic duct or from a thermal injury. E, An ERCP showing a stricture from a surgical clip (arrow); F, the corresponding transhepatic cholangiogram showing the bile duct stricture (arrow).

Chapter 66  Treatment of Gallstone Disease Table 66-7  Causes of Abdominal Pain after Cholecystectomy Biliary causes

Pancreatic causes Other gastrointestinal disorders

Extraintestinal disorders

Biliary stricture Biliary tract malignancy Choledocholithiasis Choledochocele Cystic duct remnant Sphincter of Oddi dysfunction Pancreatitis Pancreatic pseudocyst Pancreatic malignancy Esophageal motor disorders Gastroesophageal reflux disease Intestinal malignancy Intra-abdominal adhesions Irritable bowel syndrome Mesenteric ischemia Peptic ulcer disease Coronary artery disease Intercostal neuritis Neurologic disorders Psychiatric disorders Wound neuroma

uncommon. MRCP and ERCP may confirm the presence of ductal stones and exclude other possibilities such as a bile duct stricture or tumor. Endoscopic sphincterotomy with stone extraction is curative in most patients.194

CYSTIC DUCT REMNANT

In some patients, the cause of postcholecystectomy symptoms has been attributed to pathology in the cystic duct remnant195,196 Abnormalities that have been described include cystic duct stones, fistulas, granulomas, and neuromas. Associated bile duct stones are common. Although the existence of such a syndrome has been controversial, in one randomized trial, complete excision of the cystic duct during cholecystectomy was associated with fewer postoperative sequelae than a standard operative technique in which a portion of the cystic duct was left in situ.197 In the era of laparoscopic cholecystectomy, the cystic duct is divided closer to its origin from the gallbladder to minimize the risk of bile duct and right hepatic artery injury that may arise from dissection at the insertion of the cystic duct into the common hepatic duct; as a result, the frequency of cystic duct stump syndrome may be higher.198 MRCP and ERCP are useful for delineating the biliary anatomy in patients with suspected cystic duct remnant pathology. Treatment is with surgical excision of the cystic duct remnant.

SPHINCTER OF ODDI DYSFUNCTION

Up to 10% of patients with postcholecystectomy pain are found to have a structural or functional abnormality of the sphincter of Oddi (see Chapter 63).199,200 Structural problems have been referred to as papillary stenosis, which is characterized by a fixed narrowing of the sphincter in association with an elevated basal sphincter pressure. The stenosis may occur as a result of trauma such as passage of gallstones, instrumentation, pancreatitis, or infection. Functional or motility disorders have been referred to as biliary or sphincter of Oddi dyskinesia and ampullary spasm. Biliary manometry in these patients reveals elevated sphincter pressure resulting from abnormal tonic or phasic smooth muscle contractions. Because the cause of this disorder is unknown, and in many cases a distinction

between a structural or functional process cannot be made, the generic term sphincter of Oddi dysfunction is now preferred. Clinical manifestations of sphincter of Oddi dysfunction include biliary-type pain, jaundice, and pancreatitis. ERCP findings of a dilated bile duct and delayed (>45 minutes) drainage of contrast medium from the bile duct are typical. The combination of biliary-type pain, abnormal liver biochemical test levels, and a dilated bile duct is highly pre­ dictive of a response to endoscopic sphincterotomy.201 In patients in whom the diagnosis is not as clear, biliary manometry is indicated. Treatment is with endoscopic sphincterotomy. Selected patients may require transduo­ denal sphincteroplasty and septoplasty.202 The pathophys­ iology, classification, and treatment of this disorder are discussed more fully in Chapter 63.

GALLSTONES, CHOLECYSTECTOMY, AND CANCER A number of reports have demonstrated associations between either gallstones or cholecystectomy and the development of cancers in organs as diverse as the gallbladder, bile ducts, stomach, colon, breast, and uterus. Whether a causal relationship exists between gallbladder disease or its treatment and the development of these malignancies is unclear. Common environmental factors, perhaps dietary, may influence the rates of all these diseases. On the other hand, alterations in the composition of bile in patients with gallstones could influence the development of carcinoma. Moreover, cholecystectomy increases the enterohepatic circulation of bile acids, which increases mucosal exposure to potentially carcinogenic secondary bile acids such as deoxycholate (see Chapter 64).203,204

BILIARY TRACT CANCER

The strongest association between gallstones and cancer is with cancers of the biliary tree itself, particularly gallbladder carcinoma (see Chapter 69). Most patients with gallbladder cancer have gallstones, and epidemiologic data show a strong relationship between the two diseases. The risk of gallbladder cancer is greater in patients with large gallstones than in those with small gallstones, in those with multiple gallstones than a single gallstone, and in Native Americans.205-207 A weaker statistical association exists between gallstones and cholangiocarcinoma, and a causal relationship is suggested by the finding that the risk is lower in patients who undergo cholecystectomy than in those whose gallstones are untreated.208,209

COLORECTAL CANCER

Studies from the early 1980s identified a statistical association between cholecystectomy and the subsequent development of colorectal cancer, particularly in the right colon.204,210-212 The magnitude of the risk of colorectal cancer, although statistically significant, was low (relative risk 1.5 to 2.0). Subsequent studies have disputed the association or attributed it to the gallstones rather than the cholecystectomy (see Chapter 123). These findings should not represent a deterrent to cholecystectomy in a patient with a clear indication for the procedure.

KEY REFERENCES

Cameron DR, Goodman AJ. Delayed cholecystectomy for gallstone pancreatitis: Re-admissions and outcomes. Ann R Coll Surg Engl 2004; 86:358-62. (Ref 147.)

1137

1138

Section VIII  Biliary Tract Choudhary A, Bechtold ML, Puli SR, et al. Role of prophylactic antibiotics in laparoscopic cholecystectomy: A meta-analysis. J Gastrointest Surg 2008; 12:1847-53. (Ref 123.) Flum DR, Dellinger EP, Cheadle A, et al. Intraoperative cholangiography and risk of common bile duct injury during cholecystectomy. JAMA 2003; 289:1639-44. (Ref 74.) Flum DR, Flowers C, Veenstra DL. A cost-effectiveness analysis of intraoperative cholangiography in the prevention of bile duct injury during laparoscopic cholecystectomy. J Am Coll Surg 2003; 196:38593. (Ref 75.) Ito K, Ito H, Whang EE. Timing of cholecystectomy for biliary pancreatitis: Do the data support current guidelines? J Gastrointest Surg 2008; 12:2164-70. (Ref 150.) Keus F, de Jong JA, Gooszen HG, van Laarhoven CJ. Laparoscopic versus open cholecystectomy for patients with symptomatic cholecystolithiasis. Cochrane Database Syst Rev 2006:CD006231. (Ref 59.) Kharbutli B, Velanovich V. Management of preoperatively suspected choledocholithiasis: A decision analysis. J Gastrointest Surg 2008; 12:1973-80. (Ref 175.) Papi C, Catarci M, D’Ambrosio L, et al. Timing of cholecystectomy for acute calculous cholecystitis: A meta-analysis. Am J Gastroenterol 2004; 99:147-55. (Ref 130.) Petroni ML, Jazrawi RP, Pazzi P, et al. Ursodeoxycholic acid alone or with chenodeoxycholic acid for dissolution of cholesterol gallstones:

A randomized multicentre trial. The British-Italian Gallstone Study group. Aliment Pharmacol Ther 2001; 15:123-8. (Ref 10.) Shao T, Yang YX. Cholecystectomy and the risk of colorectal cancer. Am J Gastroenterol 2005; 100:1813-20. (Ref 204.) Siddiqui T, MacDonald A, Chong PS, Jenkins JT. Early versus delayed laparoscopic cholecystectomy for acute cholecystitis: A metaanalysis of randomized clinical trials. Am J Surg 2008; 195:40-7. (Ref 131.) Strasberg SM, Eagon CJ, Drebin JA. The “hidden cystic duct” syndrome and the infundibular technique of laparoscopic cholecystectomy— the danger of the false infundibulum. J Am Coll Surg 2000; 191:661-7. (Ref 68.) Uhl W, Warshaw A, Imrie C, et al. IAP guidelines for the surgical management of acute pancreatitis. Pancreatology 2002; 2:565-73. (Ref 151.) Williams EJ, Green J, Beckingham I, et al. Guidelines on the management of common bile duct stones (CBDS). Gut 2008; 57:1004-21. (Ref 194.) Zaliekas J, Munson JL. Complications of gallstones: The Mirizzi syndrome, gallstone ileus, gallstone pancreatitis, complications of “lost” gallstones. Surg Clin North Am 2008; 88:1345-68. (Ref 159.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

67 Acalculous Biliary Pain, Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis, and Polyps of the Gallbladder Karin L. Andersson and Lawrence S. Friedman

CHAPTER OUTLINE Acalculous Biliary Pain  1139 Definition and Clinical Features  1139 Epidemiology and Pathophysiology  1140 Diagnosis and Treatment  1140 Acute Acalculous Cholecystitis  1141 Definition  1141 Epidemiology  1141 Pathogenesis  1141 Clinical Features  1142 Diagnosis  1142 Treatment  1143 Prevention  1144 Cholesterolosis  1144 Definition  1144 Epidemiology  1144 Pathology  1145 Pathogenesis  1145 Clinical Features  1145

Although gallstones and their complications account for most cholecystectomies,1 a consistent 15% of these operations are performed in patients without gallstones.2 In these patients, the majority of cholecystectomies are performed as treatment for one of two distinct clinical syndromes: acalculous biliary pain and acalculous cholecystitis. As shown in Table 67-1, acalculous biliary pain is generally a disorder of young, predominantly female, ambulatory patients and mimics calculous biliary pain. Acute acalculous cholecystitis is typically a disease of immobilized and critically ill older men with coexisting vascular disease. Because the clinical features and prognosis of these two entities are quite different, they are considered separately in this chapter. Three typically asymptomatic conditions of the gallbladder—cholesterolosis, adenomyomatosis, and gallbladder polyps—are also reviewed. This chapter is dedicated to the memory of Lyman E. Bilhartz, MD.

Diagnosis  1146 Treatment  1146 Adenomyomatosis  1146 Definition  1146 Epidemiology  1146 Pathology  1146 Pathogenesis  1146 Clinical Features  1147 Diagnosis  1148 Treatment  1149 Polyps of the Gallbladder  1149 Definition  1149 Epidemiology  1149 Pathology  1149 Clinical Features and Diagnosis  1150 Natural History  1151 Treatment  1151

ACALCULOUS BILIARY PAIN DEFINITION AND CLINICAL FEATURES

Intense epigastric or right upper quadrant pain that starts suddenly, rises in intensity over a 15-minute period, and continues at a steady plateau for 30 minutes or more before slowly subsiding is characteristic of biliary pain. The localization of pain to the right hypochondrium or radiation to the right shoulder is the most specific finding for a biliary tract origin.3 The attacks of pain are frequently precipitated by ingestion of a meal and may be accompanied by restlessness, nausea, or vomiting. Between attacks, the physical findings are usually normal, with the possible exception of residual upper abdominal tenderness. When a patient presents with such a history and ultrasonography confirms the presence of gallstones, the management is straightforward—namely, elective cholecystectomy (or perhaps an attempt at medical dissolution of the stones)

1139

1140

Section VIII  Biliary Tract Table 67-1  Comparison of Acalculous Biliary Pain and Acute Acalculous Cholecystitis ACALCULOUS BILIARY PAIN

ACUTE ACALCULOUS CHOLECYSTITIS

Epidemiology

Female preponderance (80%) Young to middle-aged ambulatory patient Risk factors are similar to those for cholelithiasis (i.e., obesity and multiparity)

Male preponderance (80%) Critically ill elderly patient in intensive care unit Risk factors are preexisting atherosclerosis, recent surgery, and hemodynamic instability

Clinical features

Episodic right upper quadrant or epigastric pain identical to calculous biliary pain Physical findings are usually normal Laboratory findings are usually normal

Unexplained sepsis with few localizing signs; rapid progression to gangrene and perforation Physical examination may show fever; right upper quadrant tenderness is present in only 25% Leukocytosis and hyperamylasemia may be present

Diagnostic tests

Ultrasonography shows no stones and usually a normal gallbladder Biliary drainage (Meltzer-Lyon test) typically demonstrates cholesterol crystals Stimulated cholescintigraphy using cholecystokinin to measure the gallbladder ejection fraction may identify patients who are likely to improve after cholecystectomy

See Table 67-2

Treatment

Elective cholecystectomy for patients with classic biliary pain and either biliary cholesterol crystals or a gallbladder ejection fraction <35%

Urgent cholecystostomy or emergency cholecystectomy for gangrene or perforation

Prognosis

Good; attacks continue unless cholecystectomy is performed

Poor, with a mortality rate of 10%-50%

(see Chapters 65 and 66). In comparison, the management of acalculous biliary pain represents a significant challenge. Patients with acalculous biliary pain have clinical features and biliary-type pain similar to those of patients with cholelithiasis, but a normal gallbladder on ultrasonography and normal serum levels of liver and pancreatic enzymes.4,5 Acalculous biliary pain may stem from a spectrum of overlapping disorders, including chronic acalculous cholecystitis, acalculous biliary disease, gallbladder dysmotility, and biliary dyskinesia, which share symptomatology but differ in the pathologic findings of the resected gallbladder. In patients with acalculous biliary pain, symptomatic improvement following cholecystectomy is more variable.

EPIDEMIOLOGY AND PATHOPHYSIOLOGY

Ultrasound-negative biliary pain is common in population studies, with reported prevalences of approximately 7% in men and 20% among women.6 Acalculous biliary pain is predominantly a disorder of young women. In one series of more than 100 patients, 83% were female, and the mean age was approximately 30 years.5 The cause of the acalculous biliary pain syndrome is not known, but indirect evidence suggests that several different etiologies may culminate in the same clinical presentation. Stimulated duodenal bile from patients with acalculous biliary pain is more dilute with respect to both bile acids and phospholipids than bile from patients with gallstones or from control women without biliary symptoms.7 The low bile acid concentration may be related to the sluggish or incomplete gallbladder contraction that has been observed in patients with acalculous biliary pain. The lower molar percentage of phospholipids is consistent with the hypothesis that biliary phospholipids are hydrolyzed to free fatty acids, which incite inflammation. The striking preponderance of young, fertile women among patients with acalculous biliary pain closely parallels the epidemiology of cholelithiasis, suggesting that the two conditions have similar risk factors. Some studies have shown that up to half of patients with acalculous biliary pain actually have microscopic cholelithiasis in resected gallbladder specimens,8 indicating that the original ultrasonogram was falsely negative. Examination of a bile speci-

men for microlithiasis (Meltzer-Lyon test, discussed later) can be helpful in identifying these patients. Several studies have shown that a subset of patients with acalculous biliary pain have histologic evidence of cholesterolosis in their resected gallbladders (see later).9-11 Although usually an incidental pathologic finding, cholesterolosis of the gallbladder may, in some patients, disrupt normal gallbladder contraction and result in biliary pain. In other patients, the resected gallbladder demonstrates significant inflammation, characteristic of chronic acalculous cholecystitis.12 Finally, acalculous biliary pain is listed as a functional gastrointestinal disorder by a multinational working committee of gastrointestinal investigators (Rome III classification [see Chapter 118]), with the implication that a pathologic lesion is not required for the diagnosis.4 In patients with a histologically normal gallbladder, a lack of coordination between gallbladder contraction and sphincter of Oddi relaxation, or gallbladder dyskinesia, may cause biliary pain (see Chapter 63). Alternatively, the strong link between acalculous biliary pain and other functional bowel disorders suggests that visceral hypersensitivity may also contribute to biliary pain in patients with a normal gallbladder.6

DIAGNOSIS AND TREATMENT

As described earlier, the symptoms of acalculous biliary pain may be indistinguishable from those of cholelithiasis. A careful review of the patient’s complaints should confirm that the symptoms are genuinely suggestive of biliary pain rather than dyspepsia, heartburn, cramping abdominal pain, or flatulence.3 If the symptoms are consistent with biliary pain, a detailed review of the ultrasonogram with a radio­ logist is warranted. Although gallstones larger than 2 mm are unlikely to be missed (the sensitivity of ultrasound for detecting stones exceeds 95%), other ultrasonographic evidence of gallbladder disease may be overlooked if the primary focus is to exclude stones. Patients with adenomyomatosis of the gallbladder or small cholesterol polyps may have biliary pain that is relieved by cholecystectomy (see later). Determining in whom and when to pursue cholecystectomy for patients with biliary pain and a normal ultrasound result presents a challenge.

Chapter 67  Acalculous Biliary Pain, Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis Examination of the Bile for Cholesterol Crystals (Meltzer-Lyon Test)

If the ultrasonogram is normal, the bile may be examined for evidence of cholesterol crystals. Long before the advent of ultrasonography, biliary drainage was used to identify patients who were likely to have gallstones. The test has been modified so that the bile is now aspirated during an upper endoscopy after stimulation of gallbladder contraction with intravenous cholecystokinin (CCK).13 The bile should be kept at room temperature and examined imme­ diately (after completion of the endoscopy) under a microscope for the presence of characteristic birefringent, notched rhomboid cholesterol crystals or calcium biliru­ binate granules. Limited clinical studies in patients with acalculous biliary pain have shown that approximately one third have crystals in their bile.9,10 At the time of surgery, most of these patients have documented microlithiasis and pathologically confirmed cholecystitis, and their symptoms resolve following cholecystectomy. The remaining two thirds of patients who do not have crystals in their bile generally have a benign course and rarely return with evidence of significant biliary tract disease.

Stimulated Cholescintigraphy

A second approach to determining which patients with acalculous biliary pain are likely to benefit from cholecystectomy involves calculation of a gallbladder ejection fraction (GBEF) using cholescintigraphy (see Chapter 63). An intravenously administered radiolabeled hepatobiliary agent (e.g., 99mTc-diisopropyl iminodiacetic acid) is concentrated in the gallbladder, and a computer-assisted gamma camera measures activity before and after stimulation of gallbladder contraction with a slow intravenous infusion of CCK over 30 minutes. The GBEF is defined as the change in activity divided by the baseline activity. Studies in healthy volunteers have shown that normal GBEF averages 75% and virtually always exceeds 35%.5 Fatty meal chole­ scintigraphy is a less costly alternative to the CCKstimulated test and uses oral fat intake (typically half-and-half milk) to stimulate gallbladder contraction physiologically; normal values for GBEF tend to be lower than those for CCK-stimulated cholescintigraphy.14 Ironically, as stimulated cholescintigraphy gains clinical acceptance, its positive predictive value is expected to fall. When the test was first developed, most patients referred for testing had been experiencing biliary pain for years, thereby allowing ample time for other causes of pain to become evident; therefore, the pretest probability of having a primary gallbladder motility derangement was high, and the specificity of the test was excellent. Now, the test is employed earlier in the evaluation of patients with biliary pain (sometimes immediately after ultrasonography fails to demonstrate gallstones), and patients with nonbiliary or self-limiting diseases have not been weeded out. The earlier that cholescintigraphy is employed, the lower the pretest probability of acalculous biliary pain and, unfortunately, the lower the predictive value of a positive result.15 Fewer than half of patients with acalculous biliary pain have a depressed GBEF, but most of those who do have a depressed GBEF continue to have symptoms when followed for as long as 3 years. If cholecystectomy is performed in these patients, histologic evidence of chronic cholecystitis is found in approximately 90%, cystic duct narrowing in 80%, and cholesterolosis in 30%.11 Long-term symptom relief following cholecystectomy typically occurs in 65% to 80% of patients with an abnormal GBEF5,16,17; however, up

to 50% of patients managed without surgery also experience symptom relief.18 Patients with acalculous biliary pain and a normal GBEF have a variable, although generally benign, course. Some are found to have a nonbiliary cause of the symptoms, and in others the pain resolves with time. Cholecystectomy has not typically been recommended for patients with acalculous pain and a normal GBEF, although symptom relief following cholecystectomy in this population may be similar to that for patients with a depressed GBEF who undergo cholecystectomy. As a result, although the GBEF is used commonly to evaluate patients with acalculous biliary-type pain, it is not a reliable predictor of the response to cholecystectomy.

ACUTE ACALCULOUS CHOLECYSTITIS DEFINITION

Acute acalculous cholecystitis is acute inflammation of the gallbladder in the absence of stones. Acute cholecystitis resulting from calculi is discussed in Chapter 65. The designation acalculous cholecystitis has been questioned as incorrectly suggesting that the disease is simply cholecystitis without stones. Instead, the term necrotizing cholecys­ titis has been proposed to reflect the distinct etiology, pathology, and prognosis of the disease.19

EPIDEMIOLOGY

Acute acalculous cholecystitis accounts for 5% to 10% of cholecystectomies performed in the United States. In fact, of the cholecystectomies performed in postoperative or hospitalized patients recovering from trauma or burns, more than half are for acalculous disease.20 Less commonly, acute acalculous cholecystitis may occur in the absence of antecedent trauma or stress, especially in children,21 elderly patients with coexisting vascular disease,22 bone marrow transplant recipients, patients who receive cytotoxic drugs via the hepatic artery,23 and patients with the acquired immunodeficiency syndrome.24 In some cases, specific infectious causes can be identified, such as Salmonella,25 Staphylococcus aureus,26 cytomegalovirus in immunocompromised patients,27 and Epstein-Barr virus in children.21 Systemic vasculitides such as polyarteritis nodosa and systemic lupus erythematosus may manifest as acute acalculous cholecystitis caused by ischemic injury to the gallbladder.28 Finally, acute acalculous cholecystitis is being recognized increasingly in otherwise healthy people without any risk factors.29,30 As a group, patients with acute acalculous cholecystitis are more likely to be men and old than are patients with cholecystitis caused by calculi, cases of which cluster in younger women.31

PATHOGENESIS

Most cases of acute acalculous cholecystitis occur in the setting of prolonged fasting, immobility, and hemodynamic instability. The gallbladder epithelium, although normally a robust tissue, is exposed continuously to one of the most noxious agents in the body: a concentrated solution of bile acid detergents. In the course of a normal day, the gallbladder empties the concentrated bile several times and is replenished with dilute (and presumably less noxious) hepatic bile. With prolonged fasting, the gallbladder is not stimulated by CCK to empty, and concentrated bile stagnates in the gallbladder lumen.32 In addition, the gallbladder epithelium has relatively high metabolic energy requirements in order to absorb electrolytes and water from the

1141

1142

Section VIII  Biliary Tract bile. Therefore, in an immobile, fasting patient with splanchnic vasoconstriction (often resulting from septic shock in a patient in the intensive care unit), ischemic and chemical injury to the gallbladder epithelium may occur.33 A study that compared the microcirculation of gallbladders removed for gallstone disease or acute acalculous cholecystitis showed that the capillaries barely filled in acalculous cholecystitis, indicating that disturbed microcirculation may play an important role in its pathogenesis.34 Inappropriate activation of factor XII (demonstrated to initiate gallbladder inflammation in animals)35 and local release of prostaglandins in the gallbladder wall36,37 have also been implicated in the tissue injury associated with acalculous cholecystitis. In animal models, tissue destruction can be attenuated by inhibiting prostaglandin synthesis with indomethacin. Expression of tight junction proteins in the gallbladder epithelium of patients with acute acalculous cholecystitis differs from calculous cholecystitis, perhaps reflecting the role of increased gallbladder wall permeability in the systemic inflammatory response.38 Infection of the gallbladder mucosa with bacteria, usually gram-negative enteric organisms and anaerobes,39 is thought to be a secondary event in acute acalculous cholecystitis, following rather than causing the initial injury. One postulated explanation for the rising incidence of acute acalculous cholecystitis, particularly in younger patients, is obesity and the accompanying increase in gallbladder wall fat, which has been demonstrated to interfere with gallbladder emptying in animal models. In one study, sixteen patients with acute acalculous cholecystitis had significantly more gallbladder wall fat than normal subjects without cholecystitis.40

Table 67-2  Diagnostic Criteria for Acute Acalculous Cholecystitis TECHNIQUE

FINDINGS

Clinical examination

Right upper quadrant tenderness is helpful, if present, but is lacking in three quarters of cases Unexplained fever, leukocytosis, or hyperamylasemia is frequently the only finding Thickened gallbladder wall (defined as >4 mm) in the absence of ascites and hypoalbuminemia (defined as serum albumin <3.2 g/dL) Presence of sonographic Murphy’s sign (maximum tenderness over the ultrasonographically localized gallbladder) Pericholecystic fluid collection Bedside availability is a major advantage Thickened gallbladder wall (defined as >4 mm) in the absence of ascites and hypoalbuminemia Pericholecystic fluid, subserosal edema (in the absence of ascites), intramural gas, or sloughed mucosa Best test for excluding other intra-abdominal diseases but requires moving the patient to a scanner Nonvisualization of the gallbladder with normal excretion of radionuclide into the bile duct and duodenum indicates a positive result for acute cholecystitis Results in critically ill, immobilized patients may be falsely positive because of viscous bile Morphine augmentation may reduce the number of false-positive results (see text) Better at excluding than confirming acute cholecystitis

Ultrasonography

Computed tomography

Hepatobiliary scintigraphy

CLINICAL FEATURES

The clinical features of acute acalculous cholecystitis often differ from those of acute cholecystitis caused by stone disease. Although right upper quadrant pain, fever, localized tenderness overlying the gallbladder, and leukocytosis may be evident in classic presentations, such as those of younger outpatients, some or all of these features are commonly lacking in elderly postoperative patients.41 Symptoms or signs referable to the right upper quadrant are initially absent in 75% of cases. Unexplained fever or hyperamylasemia may be the only clues that something is amiss. Compared with the clinical course of typical calculous cholecystitis, that of acute acalculous cholecystitis is more fulminant. By the time the diagnosis has been made, at least half of the patients have experienced a complication of cholecystitis, such as gangrene or a confined perforation of the gallbladder.42 Empyema of the gallbladder and ascending cholangitis may further complicate cases in which bacterial superinfection of the gallbladder has occurred. Because the disease often occurs in debilitated patients and complications occur rapidly, the mortality rate of acute acalculous cholecystitis is high, ranging from 10% to 50%, as compared with a 1% mortality rate in patients with calculous cholecystitis. Such high mortality rates have led some investigators to propose that empirical cholecystostomy be considered in gravely ill patients in the intensive care unit in whom no source of sepsis can be found.43

DIAGNOSIS

The rapid development of complications in acute acalculous cholecystitis makes early diagnosis critical for avoiding excessive mortality. Unfortunately, the lack of specific clinical findings pointing to the gallbladder, combined with a confusing clinical picture related to antecedent surgery or

trauma, makes early diagnosis difficult. For elderly patients at risk, a high index of suspicion for biliary tract sepsis is the best hope for early recognition and treatment. Table 67-2 delineates several diagnostic criteria for acute acalculous cholecystitis.

Ultrasonography

In the evaluation of patients with suspected acute acalculous cholecystitis, ultrasonography offers the distinct advantages of being widely available and easily transportable to the bedside.44 Three ultrasonographic findings indicative of gallbladder disease are a (1) thickened gallbladder wall (defined as >4 mm) in the absence of ascites or hypoalbuminemia, (2) sonographic Murphy’s sign (defined as maximum tenderness over the ultrasonographically localized gallbladder), and (3) pericholecystic fluid collection. A thickened gallbladder wall (Fig. 67-1) is not specific for cholecystitis but in the proper clinical setting is suggestive of gallbladder involvement and should prompt further evaluation. A sonographic Murphy’s sign is operator dependent and requires a cooperative patient but, when present, is a reliable indicator of gallbladder inflammation.45 A pericholecystic fluid collection indicates advanced disease. Sensitivity rates of ultrasonography for detecting acute acalculous cholecystitis have been reported to range from 67% to 92%, with specificity rates of more than 90%.44 Investigators have proposed an ultrasonographic scoring system to improve the diagnostic accuracy of ultrasonography in critically

Chapter 67  Acalculous Biliary Pain, Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis

ill patients.46 Two points are given for distention of the gallbladder or thickening of the gallbladder wall, and 1 point each is given for “striated” thickening (alternating hypoechoic and hyperechoic layers) of the gallbladder wall, sludge, or pericholecystic fluid. Scores of 6 or higher accurately predict acalculous cholecystitis.

pathogenesis of which involves obstruction of the cystic duct by a stone, filling of the gallbladder on scintigraphy virtually excludes cholecystitis as the cause of the patient’s symptoms.48 Hepatobiliary scintigraphy is less precise in acute acalculous cholecystitis. Gallbladder and cystic wall edema can cause an obstructive picture similar to calculous cholecystitis on scintigraphy. Patients with acute acalculous cholecystitis have often fasted for prolonged periods, a state that can result in concentrated, viscous bile that flows poorly through the cystic duct and causes a false-positive hepatobiliary scan result. Most patients with acute acalculous cholecystitis (in contrast to those with calculi) do not have an obstructed cystic duct; hence, hepatobiliary scans can be falsely negative as well.49 The sensitivity of the test may exceed 90%, but the lack of specificity in fasted, critically ill patients limits the usefulness of the test primarily to excluding acute acalculous cholecystitis rather than confirming the diagnosis. A study in which ultrasonography and cholescintigraphy were performed in critically ill patients found cholescintigraphy to be useful for the early diagnosis of acute acalculous cholecystitis, whereas ultrasonography alone did not permit an early decision regarding the need for surgery.50 In an effort to improve the accuracy of biliary scintigraphy, investigators have proposed the use of morphineaugmented cholescintigraphy, in which morphine sulfate is administered intravenously (0.05 to 0.1 mg/kg) to patients in whom the gallbladder has not been visualized on standard cholescintigraphy.51 The rationale for this procedure is that morphine increases resistance to the flow of bile through the sphincter of Oddi and thus forces filling of the gall­ bladder if the cystic duct is patent, thereby reducing the likelihood of a false-positive result. In approximately 60% of critically ill patients with possible biliary tract sepsis and a nonvisualized gallbladder on standard cholescintigraphy, the gallbladder is visualized after morphine augmentation, and, therefore, acute cholecystitis can be excluded as the source of sepsis.

Computed Tomography

TREATMENT

Figure 67-1.  Ultrasonogram demonstrating thickening of the gallbladder wall to 17 mm (denoted by asterisks) characteristic of acute acalculous cholecystitis. Point tenderness was noted when the transducer was pressed onto the abdomen over the gallbladder (sonographic Murphy’s sign). The diagnosis was confirmed at laparotomy. (Courtesy of David Hurst, MD, Dallas, Tex.)

Computed tomography (CT) findings suggestive of cholecystitis include gallbladder wall thickening (>4 mm), pericholecystic fluid, subserosal edema (in the absence of ascites), intramural gas, and sloughed gallbladder mucosa. Sensitivity and specificity rates of these findings for predicting acute acalculous cholecystitis at surgery have been reported to exceed 95%. CT is also superior to ultrasonography in detecting disease elsewhere in the abdomen that could be the cause of a patient’s fever or abdominal pain.47 An obvious disadvantage of CT is that it cannot be performed at the bedside, which is necessary in many critically ill patients. Several investigators have emphasized that CT is complementary to ultrasonography and often detects gallbladder disease in high-risk patients with normal ultrasonographic findings.

Hepatobiliary Scintigraphy

Hepatobiliary scintigraphy may be useful for excluding cystic duct obstruction in patients with clinical features suggestive of acute cholecystitis. Under normal conditions, intravenously administered radionuclide is taken up by the liver, secreted into bile, concentrated in the gallbladder (where it produces a “hot spot” on a scan), and emptied into the duodenum. A positive scan result for cystic duct obstruction is defined as failure of filling of the gallbladder despite the normal passage of radionuclide into the duodenum. In suspected calculous cholecystitis, the

In light of the rapid progression of acute acalculous cholecystitis to gangrene and perforation, early recognition and intervention are required. Supportive medical care should include restoration of hemodynamic stability as well as antibiotic coverage for gram-negative enteric organisms and anaerobes if biliary tract infection is suspected.

Surgical Cholecystectomy and Cholecystostomy

Traditionally, the definitive therapeutic approach for acute acalculous cholecystitis has been urgent laparotomy and cholecystectomy (see Chapter 66). Nowadays, laparoscopic cholecystectomy is the standard approach.52 In patients too unstable to tolerate anesthesia, radiographically guided percutaneous cholecystostomy can be performed53; definitive cholecystectomy can be undertaken when the patient is stable, if necessary.

Percutaneous Cholecystostomy

Several investigators have reported favorable results with the ultrasonographically guided percutaneous transhepatic placement of a cholecystostomy drainage tube, coupled with intravenous administration of antibiotics, as definitive therapy in patients in whom surgery poses a high risk.43,54-55 Studies suggest that most patients with acute acalculous cholecystitis can be treated with percutaneous drainage; if the postdrainage cholangiogram is normal, the catheter can be removed, and cholecystectomy is not necessary.54,56

1143

1144

Section VIII  Biliary Tract Transpapillary Endoscopic Cholecystostomy

Some critically ill patients with suspected acute acalculous cholecystitis are poor candidates for ultrasonographically guided percutaneous cholecystostomy, let alone surgery, because of massive ascites or uncorrectable coagulopathy. Such patients may benefit from an endoscopic approach in which the cystic duct is selectively cannulated during endoscopic retrograde cholangiopancreatography with an obliquely angled guidewire that tracks along the lateral wall of the bile duct and facilitates cannulation of the cystic duct.57 If the wire can negotiate the spiral valves within the cystic duct successfully, a nasobiliary catheter is introduced over the guidewire into the gallbladder, the contents are aspirated, and the gallbladder is lavaged with 1% Nacetylcysteine in saline to dissolve mucus and sludge. The nasocholecystostomy catheter is allowed to drain by gravity for several days and can be easily removed when the patient has recovered and is stable. Studies have shown that successful intubation of the gallbladder can be achieved in 90% of attempts and that drainage and lavage of the viscous black bile and sludge from the gallbladder result in clinical resolution in most of these critically ill patients. The technique is more cumbersome and expensive than ultrasonographic placement of a cholecystostomy tube and should be reserved for patients who would not tolerate a percutaneous approach.58

PREVENTION

Daily stimulation of gallbladder contraction with intravenously administered CCK, 50 ng/kg over 10 minutes, has been shown to prevent the formation of gallbladder sludge in fasting patients receiving total parenteral nutrition (see Chapter 62).59 The efficacy and cost-effectiveness of such prophylaxis remain to be established.

CHOLESTEROLOSIS DEFINITION

Cholesterolosis is an acquired histologic abnormality of the gallbladder epithelium characterized by excessive accumulation of cholesterol esters and triglyceride within epithelial macrophages (Fig. 67-2).60 Clinicians generally encounter the lesion only as an incidental pathologic finding after surgical resection of the gallbladder, although the diagnosis may be suspected in certain patients before surgery. Cholesterolosis, as well as adenomyomatosis of the gallbladder (see later), has been classified as one of the hyperplastic cholecystoses, a term introduced in 1960 to describe several diseases of the gallbladder thought to share the common features of mucosal hyperplasia, hyperconcentration and hyperexcretion of dye on cholecystography, and absence of inflammation.61 The proponents of this concept believed that biliary pain, in the absence of gallstones, could be explained by the presence of one of the hyperplastic cholecystoses. Other investigators, citing the lack of a common etiology and the nonspecificity of the clinical features, have recommended that the term hyperplastic cholecystoses be abandoned.

EPIDEMIOLOGY

Although cholesterolosis has been recognized as a distinct pathologic entity for more than a century, its actual prevalence remains a matter of some dispute. Depending on whether gross or microscopic criteria are used for diagnosis, the frequency of cholesterolosis in autopsy specimens has ranged from 5% to 40%. A large autopsy series involving more than 1300 cases in which each gallbladder was examined microscopically found the prevalence of cholesterol­

Lipid-laden foamy macrophages

Epithelium

Lamina propria

Muscle layer

Adventitia

Normal gallbladder

Diffuse cholesterolosis

Cholesterol polyp

Figure 67-2.  Schematic representation of a normal gallbladder, diffuse cholesterolosis, and a cholesterol polyp. Note the distribution of lipid-laden foamy macrophages in cholesterolosis and the cholesterol polyp. The diffuse form of cholesterolosis (center; see also Fig. 67-3) accounts for 80% of cases and generally causes no symptoms. Cholesterol polyps (right), present in 20% of cases, are typically small, fragile excrescences that have a tendency to ulcerate or detach spontaneously from the mucosa. Although usually asymptomatic, these polyps have been associated with biliary pain and even acute pancreatitis.

Chapter 67  Acalculous Biliary Pain, Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis osis to be 12%.62 When surgically resected gallbladders were examined, the frequency was, not surprisingly, about 50% higher (18%) than that found in autopsy material.63 The incidence of cholesterolosis has not been calculated because its onset is rarely known. The epidemiology of cholesterolosis is analogous to that of cholesterol gallstone disease,64 in that similar groups of persons are predisposed; however, the two lesions occur independently and do not usually coexist in the same person. Like gallstone disease, cholesterolosis is uncommon in children (the youngest reported patient was a 13-year-old girl) and shows a marked predilection for women up to 60 years of age. After that, the gender differences are less pronounced. No racial, ethnic, or geographic differences in prevalence have been described, although if the analogy with cholesterol gallstone disease is extended, the prevalence would be expected to be higher in Western than nonWestern societies. Obesity also appears to be a risk factor for cholesterolosis; a frequency of 38% has been observed in gallbladders resected during weight loss surgery.65

PATHOLOGY

Cholesterolosis is defined pathologically by the accumulation of lipid (cholesteryl esters and triglyceride) within the gallbladder mucosa. The four patterns of lipid deposition are as follows60: Diffuse: The lipid is distributed throughout the epithelial lining of the gallbladder and ends abruptly at the cystic duct. This pattern accounts for 80% of all cases. Cholesterol polyps: The excess lipid is confined to one or more areas of the epithelium that eventually form excrescences into the lumen of the gallbladder. Isolated cholesterol polyps in the absence of diffuse cholesterolosis account for about 10% of the total cases. Combined diffuse cholesterolosis and cholesterol polyps: Cholesterol polyps occur on a background of diffuse cholesterolosis. This pattern accounts for about 10% of cases. Focal cholesterolosis: Excess lipid deposition is limited to a small area of the mucosa.

Figure 67-3.  Photomicrograph of diffuse cholesterolosis. Note the hyperplastic, elongated villi and the foamy macrophages (arrows). (Hematoxylin and eosin.) (Courtesy of Pamela Jensen, MD, Dallas, Tex.)

cosa. Although extracellular deposits of lipid are rare, small yellow particles (lipoidic corpuscles) representing detached masses of foam cells are occasionally seen floating in the bile.

PATHOGENESIS

When the gallbladder is inspected visually at the time of laparotomy or laparoscopy, a diagnosis of cholesterolosis can be made in 20% of the cases on the basis of the gross appearance of the gallbladder mucosa as seen through the translucent serosal surface. When the gallbladder is opened, the mucosa characteristically has pale, yellow linear streaks running longitudinally, giving rise to the term strawberry gallbladder (although the mucosa is usually bile stained rather than red). When cholesterolosis is diagnosed at the time of surgical resection of the gallbladder, gallstones are also present in 50% of cases. If the diagnosis of cholester­ olosis is made at autopsy, stones are present in only 10%,62 demonstrating that the two disease processes are independent of each other.

The cause of the accumulation of cholesteryl esters and triglyceride in cholesterolosis remains obscure.66 Postulated mechanisms are that the cholesterol is derived from the blood67 or that mechanical factors that impede emptying of the gallbladder lead to local deposition of lipid.68 Data have shown unequivocally that the gallbladder epithelium is capable of absorbing cholesterol from the bile, as might be expected in epithelium that is embryologically and histologically similar to intestinal absorptive cells.69,70 Moreover, the cholesterol in gallbladder bile is already in the ideal physical state for absorption (i.e., a mixed micelle). The question remains as to why, in some patients, resorbed biliary cholesterol is esterified and then stored in foamy macrophages as cholesterolosis.71 Like cholesterol stones, cholesterolosis is frequently, but not always, found in gallbladders exposed to bile that is supersaturated with cholesterol.72 The two disorders (cholesterolosis and stone disease), both of which lead to the ectopic accumulation of cholesterol, probably share common pathogenic mechanisms (such as the secretion of abnormal bile) but progress independently in a given patient, depending on other factors such as the presence of nucleating proteins in bile and the rate of mucosal esterification of cholesterol.73 Cholesterolosis is not associated with high serum cholesterol levels.64

Microscopic Appearance

CLINICAL FEATURES

Gross Appearance

Hyperplasia of the mucosa is invariably present and is described as marked in 50% of cases. Usually, the hyperplasia is of the villous type. The most prominent feature is an abundance of macrophages within the elongated villi. Each macrophage is stuffed with lipid droplets and has a characteristic appearance of a foam cell (Fig. 67-3). In milder cases, the foam cells are limited to the tips of the villi (accounting for the linear streaks seen on gross examination); with more severe involvement, the foam cells may fill the entire villi and spill over into the underlying submu-

Cholesterolosis usually does not cause symptoms, as is evident by how frequently autopsy specimens show the lesion in patients who never had biliary symptoms. On occasion, individual patients have dull, vague, right upper quadrant or epigastric pain that resembles biliary pain and are found subsequently to have cholesterolosis without stones or gallbladder inflammation after cholecystectomy. Of the patients who undergo cholecystectomy for the syndrome of acalculous biliary pain, pain is more likely to resolve in those in whom incidental cholesterolosis is found

1145

1146

Section VIII  Biliary Tract on pathologic examination of the gallbladder than in those in whom cholesterolosis is not found.74 In retrospective surgical series of nearly 4000 gallbladders removed by cholecystectomy, 55 patients with acalculous cholesterolosis were identified.75 The investigators found that nearly half of these patients had presented with recurrent pancreatitis of unknown etiology and speculated that small cholesterol polyps had detached from the gallbladder wall and transiently obstructed the sphincter of Oddi, thereby provoking the acute pancreatitis. In 5 years of postoperative follow-up, pancreatitis did not recur. These investigators and others76,77 have suggested that cholester­ olosis (or more specifically, cholesterol polyps) should be considered in the differential diagnosis of idiopathic pancreatitis.

DIAGNOSIS

Diffuse cholesterolosis (which, as noted earlier, constitutes 80% of cases) is only rarely detectable by either ultrasonography or oral cholecystography. In the polypoid form, however, polyps of sufficient size have a characteristic appearance on ultrasonography as single or multiple, nonshadowing, fixed echoes that project into the lumen of the gallbladder.78 Most of the polyps are small (2 to 10 mm). The polyps can be identified accurately as cholesterolosis polyps by endoscopic ultrasonography, which demonstrates a characteristic aggregation of hyperechoic spots.79 On oral cholecystography, the polyps appear as small, round radiolucencies in the lumen of the opacified gallbladder and are best demonstrated after the gallbladder has emptied partially and abdominal compression has been applied.

TREATMENT

Because cholesterolosis is only rarely diagnosed before resection of the gallbladder, the issue of treatment is usually irrelevant. In the rare case of polypoid cholesterolosis diagnosed on ultrasonography or cholecystography, the absence of biliary tract symptoms argues against any intervention. If the patient has symptoms consistent with biliary pain or pancreatitis, a cholecystectomy is indicated.75 There is no medical therapy for cholesterolosis.

ADENOMYOMATOSIS DEFINITION

Adenomyomatosis (an unwieldy term that obscures its meaning) of the gallbladder is an acquired, hyperplastic lesion characterized by excessive proliferation of surface epithelium with invaginations into the thickened muscularis or even more deeply.80 Despite the prefix adeno-, the lesion is generally benign and unrelated to adenomatous epithelia elsewhere in the gastrointestinal tract. Simple adenomyomatosis is not thought to have the potential for malignant transformation. The literature on this obscure condition is complicated by the use of a number of different terms to describe the same lesion. One researcher noted that adenomyomatosis has been described by at least 18 distinct names, the more common of which are adenomyoma (used when the lesion is localized to the gallbladder fundus), diverticulosis of the gallbladder (ignores the hyperplasia), cholecystitis glandularis proliferans (overemphasizes the role of inflammation), Rokitansky-Aschoff sinuses (familiar but anatomically incorrect), adenomyosis, and adenomyomatous hyperplasia.81 Some terms are used in the radiologic literature,

whereas others are used exclusively by pathologists. None is familiar to most gastroenterologists.

EPIDEMIOLOGY

The prevalence of adenomyomatosis of the gallbladder varies greatly according to the criteria used for diagnosis and whether resected gallbladders or autopsy specimens are examined. In a large series of more than 10,000 cholecystectomy specimens, Shepard and associates82 found only 103 cases of adenomyomatosis, for a prevalence of about 1%. The lesion is more common in women than men by a 3 : 1 ratio, and the prevalence rises with age. Neither ethnic nor geographic differences in prevalence have been described.

PATHOLOGY

A review of the normal histologic architecture of the gallbladder and Rokitansky-Aschoff sinuses is useful for understanding the pathology of adenomyomatosis (Fig. 67-4). Unlike the small intestine, the gallbladder has no muscularis mucosa, and the lamina propria abuts directly on the muscular layer. In childhood, the epithelial layer is cast up into folds and supported by the lamina propria. As the gallbladder ages, the valleys of the epithelial layer may deepen so that they penetrate into the muscular layer and form Rokitansky-Aschoff sinuses. These sinuses are acquired lesions present in about 90% of resected gallbladders. If the Rokitansky-Aschoff sinuses are deep and branching and are accompanied by thickening (hyperplasia) of the muscular layer, a diagnosis of adenomyomatosis can be made.80 Rupture of Rokitansky-Aschoff sinuses is thought to underlie the rare entity xanthogranulomatous cholecystitis, in which the gallbladder is involved in an inflammatory process with lipid-laden macrophages.

Gross Appearance

Adenomyomatosis may involve the entire gallbladder (diffuse or generalized adenomyomatosis) or, more commonly, may be localized to the gallbladder fundus, in which case the lesion is often termed adenomyoma. On rare occasions, the process may be limited to an annular segment of the gallbladder wall (segmental adenomyomatosis) and may give rise to luminal narrowing and a “dumbbell-shaped” gallbladder (Fig. 67-5). In any case, the involved portion of the gallbladder wall is thickened to 10 mm or more, and the muscle layer is three to five times its normal thickness. On cut sections, cystic dilatations of the Rokitansky-Aschoff sinuses are evident and may be filled with pigmented debris or calculi.

Microscopic Appearance

Hyperplasia of the muscle layer is invariably present, and the epithelial lining occasionally undergoes intestinal metaplasia. Mild chronic inflammation is often present.

PATHOGENESIS

The pathogenesis of adenomyomatosis is unknown. Increased intraluminal pressure in the gallbladder from mechanical obstruction (e.g., from an obstructing calculus, kink in the cystic duct, or congenital septum) has been postulated to result in cystic dilatation of the RokitanskyAschoff sinuses, subsequent hyperplasia of the muscle layer, and adenomyomatosis.80 Like pressure-related colonic diverticula, Rokitansky-Aschoff sinuses are most likely to be found where the muscle layer is weakest (at the site of a penetrating blood vessel). Nevertheless, evidence of outflow obstruction of the gallbladder is not always found; for example, calculi are present in only about 60% of cases of

Chapter 67  Acalculous Biliary Pain, Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis Lumen

Epithelium

Lumen

Lumen

Lamina propria

Adventitia

Muscle layer

Normal gallbladder

Rokitansky-Aschoff sinus

Adenomyomatosis

Figure 67-4.  Schematic representation of a normal gallbladder, Rokitansky-Aschoff sinus, and adenomyomatosis. Rokitansky-Aschoff sinuses, which are present in about 90% of resected gallbladders, consist of invaginations of the epithelium into the muscle layer to produce tiny intramural diverticula. By themselves, they have no clinical significance. A histologic diagnosis of adenomyomatosis requires that the Rokitansky-Aschoff sinuses be deep, branching, and accompanied by hyperplasia of the muscle layer.

Lamina propria

Muscle layer

Epithelium

Normal gallbladder

Fundic adenomyomatosis (adenomyoma)

adenomyomatosis.82 Some investigators have proposed that adenomyomatosis is a consequence of chronic inflammation, but inflammation is not always present, particularly when the lesion is localized to the fundus.83 Finally, several investigators have noted an association between adenomyomatosis and anomalous pancreaticobiliary ductal union (see Chapter 55). In one study, half of the patients with adenomyomatosis had anomalous pancreaticobiliary ductal union,84 and in another study, one third of patients with anomalous pancreaticobiliary ductal union had adenomyo-

Generalized adenomyomatosis

Segmental adenomyomatosis

Figure 67-5.  Schematic representation showing the different patterns of adenomyomatosis. Most of the cases are localized to the fundus of the gallbladder (in which case the lesion is termed an adenomyoma); generalized and segmental patterns are much less common. An adenomyoma is usually 10 to 20 mm in diameter and may be largely confined to the wall or may project into the lumen to produce a polypoid lesion.

matosis.85 The pathogenic link between these two peculiar entities is unclear.

CLINICAL FEATURES

Adenomyomatosis, like cholesterolosis, usually causes no symptoms and is typically an incidental finding at autopsy or surgical resection. As noted earlier, gallstones are present in more than half of the resected gallbladders that are found to have adenomyomatosis; in these cases the symptoms can be ascribed to the stones.82 Uncommonly,

1147

1148

Section VIII  Biliary Tract A

B

Figure 67-6.  A, Gross pathologic appearance of a gallbladder adenomyoma involved by adenocarcinoma. B, Histologic examination shows a moderately differentiated adenocarcinoma of the gallbladder undermining the mucosa of the adenomyoma. (Hematoxylin and eosin.) (Courtesy of Aviva Hopkowitz, MD, Dallas, Tex.)

acalculous adenomyomatosis appears to cause symptoms indistinguishable from the biliary pain of cholelithiasis. On rare occasions, adenocarcinoma of the gallbladder has been found in association with adenomyomatosis (Fig. 67-6)86; however, the malignancy is often far removed from the localized area of adenomyomatosis, and the association has been thought to be coincidental rather than causal. Nevertheless, several reports of adenocarcinoma occurring in an area of gallbladder wall involved with adenomyomatosis have created diagnostic uncertainty on ultrasonography or cholecystography.87 A retrospective review of more than 3000 resected gallbladders revealed a significantly higher frequency (6.4%) of gallbladder cancer in gallbladders with the segmental form of adenomyomatosis than would have been expected by chance alone. The investigators proposed that segmental adenomyomatosis should be considered a potentially premalignant lesion.88 A second review of gallbladder cancers associated with segmental adenomyomatosis revealed a spectrum of cytologic atypia in the specimens ranging from hyperplastic to malignant epithelium, suggestive of neoplastic progression.89 When simple adenomyomatosis of the gallbladder is discovered incidentally, the lesion is likely to be benign. If there is any suspicion of an associated mass lesion, particularly one larger than 10 mm, or if segmental adenomyomatosis is found, however, a thorough radiologic evaluation of the gallbladder is warranted, and cholecystectomy should be considered.

Figure 67-7.  Oral cholecystogram showing segmental adenomyomatosis in a 28-year-old man with postprandial epigastric pain radiating through to the back. The film demonstrates an annular segment of the gallbladder wall (arrowhead) involved with adenomyomatosis, which has produced a constriction of the lumen. Although no gallstones were present, a cholecystectomy was performed, and the patient’s symptoms were relieved. (Courtesy of W. J. Kilman, MD, Dallas, Tex.)

DIAGNOSIS

As noted previously, adenomyomatosis is frequently diagnosed only after resection and direct examination of the gallbladder; however, several specific radiologic and ultrasonographic findings may, if present, allow the diagnosis to be made preoperatively. On oral cholecystography (see Chapter 65), the mural diverticula that constitute Rokitansky-Aschoff sinuses may fill with contrast material and produce characteristic radiopaque dots that parallel the margin of the gallbladder lumen.90 Any portion of the gallbladder wall may be involved (Fig. 67-7). Localized, fundal adenomyomatosis (adenomyoma) may manifest as a filling defect in the fundus, whereas segmental adenomyomatosis may appear as a circumferential narrowing of the gallbladder lumen. As is the

case with cholesterolosis, the radiologic findings in adenomyomatosis are best appreciated when the gallbladder has partially emptied of contrast material and external pressure has been applied during the examination.90 Although ultrasonography has largely replaced oral cholecystography in the evaluation of the gallbladder, the ultrasonographic findings in adenomyomatosis are less specific. A thickened gallbladder wall (>4 mm) is not specific for adenomyomatosis and can also be seen in many other conditions such as liver disease with ascites.91 Carefully performed studies in which radiologic and ultrasonographic findings of adenomyomatosis were correlated with pathologic findings have shown that diffuse or segmental thicken-

Chapter 67  Acalculous Biliary Pain, Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis ing of the gallbladder wall in association with intramural diverticula (seen as round anechoic foci) accurately predicts adenomyomatosis.92 If the intramural diverticula (dilated Rokitansky-Aschoff sinuses) are filled with sludge or small calculi, the lesions may appear echogenic with acoustic shadowing or a reverberation artifact.93 Endoscopic ultrasonography may demonstrate the characteristic finding of multiple microcysts, corresponding to the proliferated Rokitansky-Aschoff sinuses.79 CT and magnetic resonance imaging findings in adenomyomatosis include94 differential enhancement of gallbladder wall layers, detection of Rokitansky-Aschoff sinuses within a thickened gallbladder wall,95 and subserosal fatty proliferation.96 In a study of 20 patients with surgically proved adenomyomatosis who had preoperative ultrasound, helical CT, and MRI evaluation, the diagnostic accuracies of the three modalities were 66%, 75%, and 93%, respectively.97 In one case report, an adenomyoma without histologic evidence of cancer was the cause of a false-positive finding on 18-fluorodeoxyglucose positron emission tomography (PET), likely because of the associated inflammatory activity.98

variants of cholesterolosis that result from infiltration of the lamina propria with lipid-laden foamy macrophages. The pathogenesis of cholesterol polyps is discussed in the section on cholesterolosis (see earlier). Cholesterol polyps are typically small (>10 mm in diameter), pedunculated polyps that are attached to the mucosa by a thin, fragile stalk.102 Frequently, detached tiny cholesterol polyps are found floating in the bile when the gallbladder is opened in the operating room.103 Although they may be solitary in 20% of cases, the mean number of cholesterol polyps present in one series was eight.104

TREATMENT

Inflammatory Polyps

In the absence of biliary tract symptoms, adenomyomatosis requires no treatment. If the patient has biliary pain and radiographic or ultrasonographic evidence of adenomyomatosis with calculi, a cholecystectomy is indicated. A more difficult clinical problem arises when a patient is symptomatic and has suspected adenomyomatosis but no stones.87 In such cases, the more extensive or severe the adenomyomatosis appears to be, the more likely that the symptoms are related to the lesion and that the patient will benefit from cholecystectomy. Fear of malignant transformation is not a reason to operate, unless an ultrasonographic or radiologic image suggests a mass or perhaps shows the segmental form of adenomyomatosis.99

POLYPS OF THE GALLBLADDER DEFINITION

The term polyp of the gallbladder is used to describe any mucosal projection into the lumen of the gallbladder.100 The vast majority of gallbladder polyps are the result of lipid deposits or inflammation, rather than neoplasms. Because the nature of a polyp cannot be defined without histologic evaluation, however, clinicians must decide whether the concern of malignancy is sufficient to perform cholecys­ tectomy based on indirect information such as the radiographic appearance of the polyp, patient demographics, and symptoms.

EPIDEMIOLOGY

The frequency of gallbladder polyps, defined either pathologically or radiologically,101 ranges from 1% to 4%. Often, gallbladder polyps are an incidental finding at the time of cholecystectomy.

PATHOLOGY

Polyps of the gallbladder may be classified as shown in Table 67-3 as either non-neoplastic (95% of all gallbladder polyps) or neoplastic.102

Cholesterol Polyps

Cholesterol polyps (also known as papillomas of the gall­ bladder, although the term should be discarded) are the most common type of gallbladder polyp. They are benign

Adenomyomas

Adenomyomatosis of the gallbladder localized to the fundus may produce a hemispheric projection into the lumen that resembles a polyp. Such a lesion has come to be known as an adenomyoma, although it is not neoplastic in origin. The pathogenesis of an adenomyoma is discussed in the section on adenomyomatosis (see earlier). The lesion is usually approximately 15 mm in size, and its bulk is confined to the muscular wall of the gallbladder.102 Inflammatory polyps are small sessile lesions that consist of granulation and fibrous tissue infiltrated with lymphocytes and plasma cells. The average size is 5 to 10 mm. A solitary polyp is found in 50% of cases, and two to five polyps are found in the remainder.102 When discovered at the time of cholecystectomy, an inflammatory polyp is almost always an incidental finding.

Adenomas

In light of the high frequency of adenomatous polyps in the gastrointestinal tract, gallbladder adenomas are surprisingly uncommon. Their frequency in resected gallbladder specimens is only about 0.15%.105 Adenomas are typically solitary, pedunculated masses from 5 to 20 mm in diameter. They may occur anywhere in the gallbladder. When multiple, as they are in approximately one third of cases, two to five polyps are usually present. Histologically, they are classified as either papillary or nonpapillary. The former type consists of a branching, tree-like skeleton of connective tissue covered with tall columnar cells, whereas the latter consists of a proliferation of glands encased by a fibrous stroma. On rare occasions, the entire gallbladder mucosa may undergo adenomatous transformation that results in innumerable tiny mucosal polyps termed multicentric papillomatosis. Notably, gallstones are present in half of cases of adenomatous polyps.102 Unlike the colon, in which adenomas are much more common than adenocarcinomas, the gallbladder is affected less commonly by adenomas than by carcinomas (by a 1 : 4 ratio). The frequency of progression from adenoma to adenocarcinoma is not well defined. In a series of more than 1600 consecutive cholecystectomies from Japan, 18 of the operated patients were found to have gallbladder adenomas.106 Seven of the adenomas contained foci of carcinoma. In the same series, 79 cases of invasive carcinoma were found; 15 (19%) of the lesions were thought to have residual adenomatous tissue within the cancer, suggesting that the initial lesion may have been an adenoma. Notably, all the adenomas that contained foci of carcinoma were larger than 12 mm, a finding that suggests that large adenomas may represent premalignant lesions.

Miscellaneous Polyps

Although a wide variety of benign lesions may manifest as polyps in the gallbladder, these lesions are rare. Fibromas,

1149

1150

Section VIII  Biliary Tract Table 67-3  Types of Gallbladder Polyps HISTOLOGIC TYPE

RELATIVE FREQUENCY (%)

NEOPLASTIC

SIZE RANGE (mm)

NUMBER OF POLYPS

Cholesterol polyp (a polypoid form of cholesterolosis)

60

No

2-10

Multiple (average of 8)

Adenomyoma (a localized form of adenomyomatosis)

25

No

10-20

One

Inflammatory polyp

10

No

5-10

One in half of cases (2-5 in remainder)

Adenoma

4

Yes

5-20

One in two thirds of cases (2-5 in remainder)

Miscellaneous neoplasms

<1

Yes

5-20

One

leiomyomas, and lipomas of the gallbladder are extraordinarily rare, particularly considering how commonly they are found elsewhere in the gastrointestinal tract. Neuro­ fibromas, carcinoids,107 and heterotropic gastric glands occur even less frequently.108 Taken together, the combined frequency of nonadenomatous neoplastic polyps of the gallbladder is considerably less than 1 per 1000 resected specimens.102

CLINICAL FEATURES AND DIAGNOSIS

Polyps of the gallbladder typically do not cause symptoms. They are often noted as an incidental finding during cholecystectomy for gallstones or by imaging studies performed for other indications. In the exceptional case in which a polyp (without gallstones) is identified ultrasonographically or radiographically because of symptoms, the clinical symptomatology may resemble that of biliary pain, although classic features (e.g., intense epigastric or right upper quadrant pain starting suddenly, rising in intensity over a 15-minute period, and continuing at a steady plateau for several hours before slowly subsiding) may be

COMMENTS May detach and behave clinically as a stone; may cause biliary pain, bile duct obstruction, or pancreatitis Surgery is not required unless the patient is symptomatic Always localized to the gallbladder fundus Forms a hemispheric projection into the lumen with the bulk confined to the muscular wall Surgery is not required unless the patient is symptomatic or a neoplasm cannot be excluded Consists of granulation tissue and fibrous tissue with lymphocytes and plasma cells that infiltrate the lamina propria Surgery is not required Rare lesion, found in only 0.15% of resected gallbladders Usually is pedunculated and coexists with stones in half of cases Only polyp in the gallbladder with a premalignant potential; the frequency of progression from adenoma to carcinoma is much lower than that for colon polyps Virtually all adenomas with a focus of carcinomas are >12 mm in diameter; lesions <10 mm can be monitored ultrasonographically For lesions >10-18 mm in size, laparoscopic cholecystectomy should be considered in good surgical candidates For lesions >18 mm in size, open rather than laparoscopic cholecystectomy should be considered, because invasive cancer is more likely and extended resection may be required Extremely rare lesions (see text), with frequencies of <0.10% of polyps

absent. Rare instances of acute acalculous cholecystitis and even hemobilia have been ascribed to benign gallbladder polyps.109 The histologic types of gallbladder polyps cannot be distinguished on clinical grounds alone.99 Nor do ultrasonographic and cholecystographic findings predict histology reliably (Fig. 67-8).91 The sensitivity of conventional ultrasonography may be as high as 80% for detecting polyps greater than 10 mm in diameter, but its accuracy in characterizing the type of polyp may be as low as 20%.110 Endoscopic ultrasonography is a more sensitive and specific method for diagnosing gallbladder polyps. One study comparing conventional and endoscopic ultrasonography found that the diagnostic accuracy of endoscopic ultrasonography for differentiating polyp types exceeded 90%.111 Several studies have shown that an endoscopic ultrasonographic scoring system that incorporates the size, number, shape, and echogenicity of polyps and polyp margins may predict the neoplastic potential of gallbladder polyps.111-113 Several cases in which preoperative 18-fluorodeoxyglucose PET accurately predicted the presence of malignant tumor of the

Chapter 67  Acalculous Biliary Pain, Acalculous Cholecystitis, Cholesterolosis, Adenomyomatosis TREATMENT

Figure 67-8.  Ultrasonogram (right longitudinal view) showing a gall­ bladder polyp in a 55-year-old woman with mild biliary pain. A 10-cm luminal filling defect is demonstrated (arrow). It does not cast an acoustic shadow and is fixed to the gallbladder wall. The findings are consistent with a gallbladder polyp, although the histology cannot be predicted from the ultrasonogram. A cholecystectomy demonstrated multiple cholesterol polyps, one of which was unu­sually large. (Courtesy of R. S. Harrell, MD, Dallas, Tex.)

gallbladder in patients with gallbladder polyps have been reported.114 Other studies have evaluated clinical predictors of malignancy. Aside from polyp size (>10 mm), patient age greater than 60 years is the strongest predictor of neoplastic disease. The presence of concurrent gallstones is also associated with a higher risk of malignancy.115 Single polyps and symptomatic polyps may be more likely to be malignant than multiple polyps and asymptomatic polyps, respectively.

NATURAL HISTORY

The few studies that have attempted to define the natural history of untreated gallbladder polyps highlight the benign nature of most polyps and support a “watch and wait” approach in most cases.116 On the basis of records at the Mayo Clinic, one study identified approximately 200 patients in whom cholecystograms demonstrated gallbladder polyps and immediate cholecystectomy was not performed.117 After 15 years of follow-up, symptoms sufficient to warrant surgery developed in fewer than 10% of the patients, and none of the patients available for follow-up had evidence of gallbladder cancer. One group of investigators performed annual or semiannual ultrasound for a 5-year period on 109 patients with polyps smaller than 10 mm. During this time, gallbladder cancer developed in no patient, and the polyp exhibited no growth in more than 88% of patients.118 Another study identified 224 patients with gallbladder polyps, 95% of which were predicted to be cholesterol polyps on the basis of the ultrasonographic appearance and the remainder of which were classified as “polypoid lesions of uncertain benignity.”119 After an average follow-up of 9 months, all the polyps thought initially to be benign remained the same size or were proved to be benign at resection. Two thirds of the polypoid lesions in which a benign nature was uncertain were found to be adenomas or carcinomas when resected. These findings suggest that although most gallbladder polyps are benign, high-risk polyps often have an identifiable characteristic such as larger size.

Patients who are symptomatic with biliary pain and have ultrasonographic evidence of both polyps and stones in the gallbladder should undergo elective cholecystectomy. The decision is more complicated for patients in whom gallbladder polyps without concurrent gallstones are discovered. For these patients, the decision to operate depends on the severity of symptoms, the confidence of the clinician that the symptoms are biliary in origin, and ultrasonographic features (particularly the size) of the polyp. The likelihood of malignancy is correlated with polyp size. Polyps less than 10 mm in diameter are unlikely to be cancerous and generally do not require intervention in the absence of symptoms. Because polyps larger than 10 mm have a greater likelihood of being cancerous, elective laparoscopic cholecystectomy should be considered in acceptable surgical candidates.120-122 In a patient who is a poor surgical risk with a polyp larger than 10 mm, periodic monitoring for polyp growth (perhaps every 6 to 12 months) with ultrasound or endoscopic ultrasound may be reasonable.120,122 Polyps larger than 18 mm in diameter pose a significant risk of malignancy and should be resected if possible. One study found that lesions of this size often contain advanced, invasive cancer that involves the serosal surface of the gallbladder and requires a more extensive dissection than can be accomplished by laparoscopy.123 As a result, the investigators advocate open cholecystectomy for these large polypoid lesions of the gallbladder. Because of the uncertainty regarding the neoplastic potential of polyps, common practice is to offer cholecystectomy to any good surgical candidate with a polyp greater than 10 mm in diameter. Patients should be warned, however, that false-positive results for polypoid lesions on ultrasonography are common and that polyps may not be found in all resected gallbladders. The 10-mm cut-off rule for following gallbladder polyps expectantly may not apply to patients with primary sclerosing cholangitis, in whom the risk of malignancy in polypoid lesions of the gallbladder may be as high as 60%.124 In this high-risk population, cholecystectomy for polyps smaller than 10 mm may be justifiable. In low-risk populations, such as asymptomatic patients with small, presumably benign polyps, periodic surveillance for polyp growth may be prudent. One group of investigators recommends conventional ultrasonographic evaluation every 3 to 6 months in the immediate post-diagnostic period to exclude a rapidly growing tumor, but less frequent or no investigation after 1 to 2 years of stability in polyp size.125

KEY REFERENCES

Akatsu T, Aiura K, Shimazu M, et al. Can endoscopic ultrasonography differentiate nonneoplastic from neoplastic gallbladder polyps? Dig Dis Sci 2006; 51:416-21. (Ref 79.) Albores-Saavedra J, Shukla D, Carrick K, Henson D. In situ and invasive adenocarcinomas of the gallbladder extending into or arising from Rokitansky-Aschoff sinuses: A clinicopathologic study of 49 cases. Am J Surg Pathol 2004; 28:621-8. (Ref 89.) Boulton R, Adams D. Gallbladder polyps: When to wait and when to act. Lancet 1997; 349:817-18. (Ref 116.) Hansel S, DiBaise J. Gallbladder dyskinesia. Curr Treat Options Gastroenterol 2008; 11:78-84. (Ref 6.) Kalliafas S, Ziegler DW, Flancbaum L, Choban PS. Acute acalculous cholecystitis: Incidence, risk factors, diagnosis, and outcome. Am Surg 1998; 64:471-5. (Ref 31.) Kmiot WA, Perry EP, Donovan IA, et al. Cholesterolosis in patients with chronic acalculous biliary pain. Br J Surg 1994; 81:112-15. (Ref 74.) Laurila J, Syrajala H, Laurila P. Acute acalculous cholecystitis in critically ill patients. Acta Anaesth Scand 2004; 48:986-91. (Ref 52.) Lee K, Wong J, Li J, Lai P. Polypoid lesions of the gallbladder. Am J Surg 2004; 188:186-90. (Ref 125.)

1151

1152

Section VIII  Biliary Tract Mirvis SE, Vainright JR, Nelson AW, et al. The diagnosis of acute acalculous cholecystitis: A comparison of sonography, scintigraphy, and CT. AJR Am J Roentgenol 1986; 147:1171-5. (Ref 44.) Rastogi A, Slivka A, Moser A, Wald A. Controversies concerning pathophysiology and management of acalculous biliary-type abdominal pain. Dig Dis Sci 2005; 50:1391-401. (Ref 16.) Terzi C, Sökmen, Seckin S, et al. Polypoid lesions of the gallbladder: report of 100 cases with special reference to operative indications. Surgery 2000; 127:622-7. (Ref 115.)

Weedon D. Adenomyomatosis. In: Sternberg SS (ed): Pathology of the Gallbladder. New York: Masson; 1984. pp 185-94. (Ref 80.) Weedon D. Cholesterolosis. In: Sternberg SS (ed): Pathology of the Gallbladder. New York: Masson; 1984. pp 161-5. (Ref 60.) Yap L, Wycherley AG, Morphett AD, Toouli J. Acalculous biliary pain: Cholecystectomy alleviates symptoms in patients with abnormal cholescintigraphy. Gastroenterology 1991; 3:786-93. (Ref 5.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

68 Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis Andrew S. Ross and Kris V. Kowdley

CHAPTER OUTLINE Primary Sclerosing Cholangitis  1153 Diagnosis  1153 Epidemiology  1155 Primary Sclerosing Cholangitis and Inflammatory Bowel Disease  1156 Etiology and Pathogenesis  1156 Natural History and Prognostic Models  1157 Clinical Features  1159 Pathology  1160 Complications  1161 Treatment  1164

Sclerosing cholangitis encompasses a spectrum of cholestatic conditions that are characterized by patchy inflammation, fibrosis, and destruction of the intrahepatic and extrahepatic bile ducts. These conditions are typically chronic, progressive disorders in which persistent biliary damage may lead to biliary obstruction, biliary cirrhosis, and hepatic failure, with associated complications. The first description of sclerosing cholangitis is credited to Delbet in 1924.1 Although considered for many years to be an extremely rare disorder, the advent of endoscopic retrograde cholangiopancreatography (ERCP) in the 1970s has allowed an improved understanding of the true prevalence of this disorder and facilitated careful study of its natural history. Nevertheless, many aspects of sclerosing cholangitis remain poorly understood; most notably lacking are a detailed knowledge of its etiology and proven effective medical therapy. A cholangiographic appearance of diffuse stricturing and segmental dilatation of the biliary system, designated sclerosing cholangitis, may be observed in many distinct conditions. The most frequent is primary sclerosing cholangitis (PSC), an idiopathic disorder that usually occurs in association with inflammatory bowel disease (IBD) but may develop independently. PSC may also be associated with a wide variety of fibrotic, autoimmune, and infiltrative disorders, although whether such associations imply a common pathogenesis or epiphenomena is unclear (Table 68-1). PSC is also associated with various immunodeficiency states; in such cases biliary abnormalities may be caused by infection with Dr. Bruce Y. Tung contributed to this chapter in the previous edition of this book.

Recurrent Pyogenic Cholangitis  1167 Epidemiology  1167 Etiology and Pathogenesis  1167 Clinical Features  1168 Pathology  1169 Treatment  1169 Prognosis and Complications  1169

an opportunistic pathogen. The term secondary sclerosing cholangitis refers to a clinical and radiologic syndrome that is similar to PSC but develops as a consequence of a known pathogenesis or injury. Obstructive, toxic, ischemic, and neoplastic causes of secondary sclerosing cholangitis have been described (see Table 68-1). This chapter focuses on PSC and recurrent pyogenic cholangitis.

PRIMARY SCLEROSING CHOLANGITIS DIAGNOSIS

No standardized criteria for the diagnosis of PSC have been universally adopted. Early diagnostic criteria included diffuse intra- and extrahepatic bile duct strictures occurring in the absence of prior biliary surgery or cholelithiasis and after exclusion of cholangiocarcinoma.2 These criteria were later modified because of the recognition that the clinical spectrum of PSC is broader than initially appreciated, and strict adherence to the original criteria underestimates the prevalence of the disease. It is now apparent that a form of PSC, termed small-duct PSC, involves only the intra­ hepatic biliary tree, without obvious extrahepatic duct abnormalities.3 In addition, both cholelithiasis and chole­ docholithiasis may develop as a consequence of PSC, and their presence does not exclude a diagnosis of underlying PSC.4,5 Furthermore, cholangiocarcinoma is a relatively common complication of PSC, and both conditions frequently coexist.6 The diagnosis of PSC is based on typical cholangiographic findings in the setting of consistent clinical, biochemical, serologic, and histologic findings as well as exclusion of

1153

1154

Section VIII  Biliary Tract Table 68-1  Classification and Diseases Associated with Sclerosing Cholangitis Primary Sclerosing Cholangitis (PSC) Principal Disease Associations Inflammatory bowel disease Crohn’s colitis or ileocolitis Ulcerative colitis Other Disease Associations Systemic disease with fibrosis Inflammatory pseudotumor Mediastinal fibrosis Peyronie’s disease Pseudotumor of the orbit Retroperitoneal fibrosis Riedel’s thyroiditis Autoimmune or Collagen Vascular Disorders Autoimmune hemolytic anemia Celiac disease Chronic sclerosing sialadenitis Membranous nephropathy Rapidly progressive glomerulonephritis Rheumatoid arthritis Sjögren’s syndrome Systemic lupus erythematosus Systemic sclerosis Type I diabetes mellitus Alloimmune Diseases Hepatic allograft rejection Hepatic graft-versus-host disease after bone marrow transplantation Infiltrative Disease Hypereosinophilic syndrome Histiocytosis X Sarcoidosis Systemic mastocytosis Immunodeficiency Congenital immunodeficiency Combined immunodeficiency Dysgammaglobulinemia X-lined agammaglobulinemia Acquired immunodeficiency Acquired immunodeficiency syndrome Angioimmunoblastic lymphadenopathy Selective IgA deficiency Secondary Sclerosing Cholangitis Obstructive Autoimmune pancreatitis Biliary parasites Caroli’s disease Choledocholithiasis Chronic pancreatitis Congenital abnormalities Cystic fibrosis Choledochal cyst Fungal infection Recurrent pyogenic cholangitis Surgical stricture Toxic Intra-arterial floxuridine (FUDR) Intraductal formaldehyde or hypertonic saline (echinococcal cyst removal) Ischemic Hepatic allograft arterial occlusion Paroxysmal nocturnal hemoglobinuria Toxic vasculitis (FUDR) Vascular trauma Neoplastic Cholangiocarcinoma Hepatocellular carcinoma Lymphoma Metastatic cancer

secondary causes of sclerosing cholangitis. The characteri­ stic cholangiographic findings are multifocal stricturing and ectasia of the biliary tree. Areas of narrowing are interspersed with areas of normal or near-normal caliber and of post-stenotic dilatation. Although the majority of patients with PSC have coexisting abnormalities of the intra- and extrahepatic bile ducts, a small percentage have an isolated lesion. Patients with small-duct PSC may have a normal cholangiogram. Gallbladder abnormalities, including tumors, may exist in up to 41% of patients with PSC.7 ERCP is considered the standard for establishing a diagnosis of PSC but carries a risk for complications of up to 10% in patients with PSC.8,9 Magnetic resonance cholangiopancreatography (MRCP) has largely replaced ERCP for diagnostic cholangiography as a result of improvements in image quality and the noninvasive nature of MRCP (Fig. 68-1). In the few studies that have compared MRCP and ERCP in patients with PSC, MRCP has demonstrated comparable sensitivity for the detection of biliary structuring,10-13 although performance and interpretation of magnetic resonance cholangiograms vary with the technique and institution. ERCP has the advantage of combining highresolution cholangiography with the potential for advanced diagnostic and therapeutic interventions, including brush cytology or intraductal biopsy for the diagnosis of cholangiocarcinoma, balloon or catheter dilation of strictures, biliary stent placement, sphincterotomy, and stone removal. Percutaneous transhepatic cholangiography (THC) may also yield diagnostic images and allow therapeutic intervention but requires percutaneous puncture and may be technically difficult if the intrahepatic bile ducts are not sufficiently dilated (see Chapter 70). Patients with IBD and a cholestatic pattern of liver biochemical test elevations should undergo imaging of the hepatobiliary system because of the relatively high pretest probability of PSC. Ultrasonography or computed tomography (CT) may be useful for planning further diagnostic and therapeutic strategies in selected patients, but they are usually insufficient for a diagnosis of PSC because normal findings do not exclude the diagnosis. The decision as to which method of cholangiography to perform must be individualized. In most cases, ERCP is the initial test of choice for patients in whom a therapeutic intervention or the need for brush cytology is anticipated. In an asymptomatic patient with mild liver biochemical abnormalities who is unlikely to require therapeutic intervention, MRCP is the preferred initial test if the images are reliable.13 When MRCP is nondiagnostic and clinical suspicion for PSC remains, diagnostic ERCP is indicated.

Differential Diagnosis

In a patient with a cholangiographic appearance characteristic of sclerosing cholangitis, secondary causes of sclerosing cholangitis must be excluded (see Table 68-1). Patients with the acquired immunodeficiency syndrome (AIDS) and a CD4+ T-lymphocyte count below 100/mm3 can exhibit a cholangiographic appearance identical to that of PSC; this entity is termed AIDS cholangiopathy. Cryptosporidium, Microsporidium, cytomegalovirus, and other organisms have been isolated from the bile of affected patients.14,15 Exposure of the bile ducts to toxins such as intra-arterial floxuridine (FUDR)16 and formaldehyde administered to treat a hydatid cyst, when the cyst communicates with the biliary tract,17 can produce a similar cholangiographic appearance. After these secondary causes of sclerosing cholangitis are excluded, distinguishing PSC from other disorders of

Chapter 68  Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis

A

B

Figure 68-1.  A, Reconstructed images from a magnetic resonance cholangiopancreatogram (MRCP) performed in a patient with primary sclerosing cholangitis. The image suggests atrophy of the right-sided intrahepatic bile ducts with slight dilatation of the ducts on the left side. B, A film from an endoscopic retrograde cholangiogram from the same patient is shown, revealing findings similar to those on MRCP.

the bile ducts may still be challenging. Choledocholithiasis and cholangiocarcinoma may both develop in conjunction with, or independent of, PSC. In the presence of extensive choledocholithiasis or diffuse cholangiocarcinoma, identifying underlying PSC may be difficult. Cholangiographic findings in patients with cirrhosis from causes other than PSC may at times be mistaken for PSC; however, cholangiography in cirrhotic patients without PSC typically shows diffuse intrahepatic attenuation of bile ducts without the ductal irregularity or stricturing seen in patients with PSC. Primary biliary cirrhosis (PBC) is another chronic cholestatic condition that shares some clinical features with PSC (see Chapter 89); however, PBC predominantly affects middle-aged women, has no association with IBD, and is associated strongly with high titers of antimitochondrial antibodies. Whereas liver histologic findings in the two disorders overlap substantially,18 the distinction between the two is readily apparent on cholangiography. Patients with advanced PBC may demonstrate smooth tapering and narrowing of the intrahepatic bile ducts, but ductal irregularity or strictures are not seen and extrahepatic lesions do not occur. Antimitochondrial antibody-negative PBC (autoimmune cholangitis) may be difficult to distinguish from small-duct PSC because serologic profiles and cholangiographic findings may overlap, but the demographic and histologic features of the two disorders are distinct (see Chapter 89). Autoimmune hepatitis may also be difficult to distinguish from PSC (see Chapter 88). In the pediatric population, PSC typically manifests with features of autoimmune hepatitis, and cholangiography is necessary to distinguish the two disorders (see Chapter 62).19 With use of a standardized scoring system for the diagnosis of autoimmune hepatitis, 7.5% of patients with PSC are characterized as “definite” or “probable” for the diagnosis of autoimmune hepatitis, thereby underscoring the need for cholangiography when PSC is suspected.20 Features suggestive of autoimmune hepatitis include female predominance, a hepatocellular rather than cholestatic pattern of liver biochemical test abnormalities, hypergammaglobulinemia, high titers of antinuclear and anti-smooth muscle antibodies, histologic evidence of

periportal necroinflammation, and clinical response to glucocorticoid therapy. An overlap syndrome between PSC and autoimmune hepatitis has been described; it consists of a mixed cholestatic and hepatocellular pattern of liver biochemical test abnormalities, the presence in serum of autoantibodies including antineutrophil cytoplasmic antibodies (ANCA), cholangiography consistent with PSC, and histologic evidence of periductular fibrosis as well as periportal necroinflammation.21,22 A disorder of the pancreaticobiliary tree termed autoimmune pancreatitis, sclerosing pancreatocholangitis, or immunoglobulin (Ig) G4–associated cholangitis has been described.23 This disorder shares cholangiographic and clinical features with PSC but differs in its responsiveness to glucocorticoid therapy. Serum levels of IgG4 are often elevated in this disorder, and high numbers of IgG4 positive lymphocytes (>20 per high-powered field) are identified in pinch biopsies obtained from the major papilla or bile duct and may be diagnostic. Although specific diagnostic criteria for this disorder are still emerging, persons without IBD who present with symptoms and cholangiographic findings consistent with PSC should undergo measurement of serum IgG4 levels as well as endoscopy and biopsy of the major papilla to exclude IgG4-associated cholangitis (see Chapter 59).23-25

EPIDEMIOLOGY

Determination of the true incidence and prevalence of PSC is complicated by the variable presentation of the disease, inconsistent diagnostic criteria, and referral bias inherent in many published studies. Two population-based studies have provided the most accurate epidemiologic estimates of PSC in Western populations. On the basis of these studies performed in the United States and Norway, the incidence of PSC is estimated to be 0.9 to 1.3 per 100,000, and the point prevalence is estimated to be 8.5 to 13.6 per 100,000.26,27 Although PSC has been diagnosed in neonates and as late as the eighth decade of life, most patients present between the ages of 25 and 45 years, with a mean age of approximately 39 years.19,28-33 Approximately 70% of patients with PSC are men,27-31 but in the subset of patients without IBD, the male-to-female ratio is lower (0.72:1).34 Women with

1155

1156

Section VIII  Biliary Tract PSC are generally older at diagnosis.27,35 PSC is also associated with nonsmoking, an effect that cannot be explained entirely by the association between ulcerative colitis (UC) and nonsmoking.36,37

PRIMARY SCLEROSING CHOLANGITIS AND INFLAMMATORY BOWEL DISEASE

The relationship between PSC and IBD is striking and incompletely understood. Approximately 80% of all patients with PSC have concomitant IBD.27-29,31,38,39 Conversely, PSC is present in 2.4% to 4.0% of all patients with chronic UC and 1.4% to 3.4% of patients with Crohn’s disease.35,38,40,41 Of patients with both PSC and IBD, approximately 85% to 90% have UC and the remainder have Crohn’s colitis or ileocolitis. The association with IBD is stronger with more extensive colonic involvement; the prevalence of PSC is approximately 5.5% in those with pancolitis, in contrast to 0.5% in those with only distal colitis.35 PSC is not thought to occur in association with Crohn’s disease isolated to the small intestine. Racial differences in the association between PSC and IBD may exist; concomitant IBD is seen in only 21% of Japanese patients with PSC.42 Despite the strong association between PSC and UC, the two diseases often progress independently of each other.43 Although IBD is typically diagnosed before PSC, UC may be newly diagnosed years after liver transplantation for endstage liver disease caused by PSC. Conversely, PSC may be diagnosed years after total proctocolectomy for UC.44,45 Whether PSC differs clinically in patients with and without concomitant IBD is unclear. Older reports demonstrated no histologic46 or cholangiographic47 differences between patients with or without IBD. One study,34 however, suggested that patients without IBD are more likely to be female, have disease isolated to the extrahepatic ducts, and be symptomatic at the time of diagnosis. Of the multiple multivariate analyses performed to identify risk factors for progression of PSC (see later), only one found that the presence of IBD has a significant independent effect on progression of PSC.29 Some patients without overt IBD may have subclinical histologic changes detected in the colon or may develop overt colitis at a later date.43 Therefore, a high index of suspicion for the emergence of IBD is warranted, and colonoscopy with random biopsies of the colonic mucosa is recommended for all patients with a new diagnosis of PSC.

ETIOLOGY AND PATHOGENESIS

The etiology and pathogenesis of PSC remain poorly understood. Genetic and immunologic factors appear to play key roles in disease susceptibility and progression. The importance of nonimmunogenetic (infectious, vascular, toxic) factors remains controversial. Currently, the most attractive model of disease pathogenesis postulates that PSC represents an immunologic reaction that develops in immuno­ genetically susceptible persons who are exposed to an environmental or toxic trigger, such as bacterial cell wall products. Any theory of the pathogenesis of PSC must explain the strong association with IBD.

Genetic Factors

The importance of genetic factors in the pathogenesis of PSC is demonstrated by familial occurrence of the disease and its associations with specific human leukocyte antigen (HLA) haplotypes. Although uncommon, familial clustering of cases of PSC have been reported.48,49 Furthermore, PSC is strongly associated with specific HLA haplotypes. Early

studies described an overrepresentation of HLA B8 and DR3 in patients with PSC; these haplotypes are also associated with other autoimmune disorders such as myasthenia gravis and autoimmune hepatitis.50,51 These findings are not explained simply by the association between PSC and IBD because HLA B8 and DR3 are not overrepresented in patients with IBD but without PSC. The subsequent development of molecular genotyping demonstrated that the most common allele in patients with PSC is DRB3*0101, which encodes the DRw52a antigen. One study found this allele in 100% of 29 patients with PSC who underwent liver transplantation,52 but subsequent studies have demonstrated this allele in only 50% to 55% of patients with PSC.53-55 Currently, the extended HLA haplotypes that are most strongly associated with PSC are as follows:53,55,56 B8-TNF*2-DRB3*0101-DRB1*0301-DQA1*0501DQB1*0201; DRB3*0101-DRB1*1301-DQA1*0103-DQB1*0603; and DRB5*0101-DRB1*1501-DQA1*0102-DQB1*0602. Haplotypes associated with protection from PSC include the following: DRB4*0103-DRB1*0401-DQA1*03-DQB1*0302 and MICA*002. The strongest association maps to the HLA class I/III boundary on chromosome 6p21. Strong disease associations have been identified with the MICA*008 allele57 and the tumor necrosis factor a-2 allele.58,59 Despite the multiple HLA associations described, however, a single HLA-encoded gene that determines susceptibility to PSC appears unlikely. More likely are multiple HLA susceptibility loci, which may explain in part why PSC is a relatively rare disease even though the HLA haplotypes associated with PSC are relatively common in populations of Northern European descent. Also controversial is whether specific haplotypes are associated with disease outcomes. One study suggested a poor prognosis in patients with PSC and HLA DR4,60 but this finding was not confirmed.55 A more recent multicenter study involving 256 patients with PSC showed that the heterozygous haplotype DR3,DQ2 was associated with a greater risk of liver transplantation or death and the DQ6 haplotype was associated with a decreased risk of disease progression.61 The relationship between several non-major histocompatibility complex (MHC) genes and susceptibility to PSC has also been investigated. An initial study reported an association with polymorphisms in the gene encoding matrix metalloproteinase 3 (MMP-3) and postulated a role for MMP-3 in progression of PSC because of its ability to regulate fibrosis and immune activation.62 A subsequent report, however, did not confirm an association between either MMP-1 or MMP-3 polymorphisms and PSC.63 Similarly, no associations between PSC and polymorphisms in the interleukin (IL)-1 or IL-10 genes have been noted.64

Immunologic Factors

Evidence suggests that the immune system plays a key role in the etiology and pathogenesis of PSC, including the multiple associations between PSC and other autoimmune disorders. The most frequently associated autoimmune disorders include type I diabetes mellitus and Graves’ disease, which are more common in patients with PSC and IBD than in patients with IBD alone.65 In addition, as described earlier, an overlap syndrome that includes features of both PSC and autoimmune hepatitis has been

Chapter 68  Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis described.21,22,66 In rare cases, well-characterized autoimmune hepatitis may evolve into sclerosing cholangitis, suggesting that both diseases may be part of the same clinical spectrum.67 Unlike most other autoimmune disorders, however, PSC has an approximately 2:1 male predominance, is not associated with disease-specific autoanti­ bodies, and does not exhibit a consistent clinical response to immunosuppressive therapy. A wide range of serum autoantibodies are found in patients with PSC, although none is specific for the disease. Whether any of these associated antibodies plays a key role in the pathogenesis of the disease process or whether they represent simple epiphenomena is unclear. Antinuclear antibodies may be present in 24% to 53%, anti-smooth muscle antibodies in 13% to 20%, and an atypical perinuclear ANCA (pANCA) in 65% to 88% of patients with PSC.68-75 Antibodies directed against cardiolipin, bactericidal/permeability-increasing protein, cathepsin G, and lactoferrin have also been detected.74-75 Antibodies directed against an epitope shared by colonic and biliary epithelial cells have been demonstrated and may suggest a mechanism for the association between IBD and PSC.76 Autoantibodies that bind to human biliary epithelial cells (anti-BEC) have been shown to induce expression of IL-6 and the cell adhesion molecule CD44; this finding could represent a potential mechanism for the inflammatory bile duct destruction seen in patients with PSC.77 Abnormalities of both humoral and cellular immunity have been described in patients with PSC. They include an increase in circulating immune complexes, deficient clearance of immune complexes, and activation of the classical pathway of the complement system.78-80 Serum elevations of IL-8 and IL-10 also suggest exaggerated humoral immunity.81 Some of the abnormalities in cellular-mediated immunity that have been described include a decrease in circulating CD8+ cytotoxic T cells,82 increased numbers of γδ T cells in peripheral blood as well as portal areas of the liver,83 and overrepresentation of Vβ3 T-cell receptor gene segments in hepatic (but not peripheral) T-cell populations.84

Biliary Epithelial Cells

The role of biliary epithelial cells in the pathogenesis of PSC remains unclear. Biliary epithelial cells could serve as a trigger and a target for immune-mediated injury. Biliary epithelial cells have been shown to express MHC class II antigens85 and adhesion molecules such as intracellular adhesion molecule-1 (ICAM-1)86 and could play a role as antigen-presenting cells to T lymphocytes. The expression of these molecules can be regulated on biliary epithelial cells by various cytokines, including IL-2 and interferon-γ.87 Biliary epithelial cells, however, may not express the costimulatory ligands necessary for activation of T lymphocytes.88 In addition, many of the same findings are seen in patients with PBC and extrahepatic bile duct obstruction as well, making it less likely that they play a primary pathogenic role in PSC.85

Infectious and Toxic Factors

The strong association between PSC and colitis has provoked the theory that penetration of infectious or toxic agents through an inflamed colon into the portal system may play an important role in the pathogenesis of PSC. Bile culture results have been positive in explanted livers in a majority of patients with PSC, although the number of bacterial strains has correlated inversely with the time since the last endoscopic intervention.89 In addition, bacterial endo-

toxin has been shown to accumulate in biliary epithelial cells in patients with PSC and PBC.90 In patients with AIDS cholangiopathy, a variety of organisms, including Cryptosporidium, Microsporidium, and cytomegalovirus, have been isolated from the bile.14,15 A study that evaluated serologic profiles in 41 patients with PSC found a higher percentage with Chlamydia lipopolysaccharide antibodies than in a large control population. No association was seen with any other microorganisms, including Mycoplasma and 22 viruses tested.91 Further study is necessary before a direct link between PSC and Chlamydia, or any other infectious agent, can be established. A loss of normal colonic mucosal barrier because of inflammation could allow portal inflow of noninfectious toxins. Toxic damage leading to sclerosing cholangitis has been demonstrated in humans as well as animal models. Biliary exposure to caustic agents17 or hepatic artery infusion of chemotherapeutic agents such as FUDR16 can produce a cholangiographic appearance identical to that of PSC. In a rat model, administration of the biliary toxin α-naphthylisothiocyanate led to the development of a chronic cholangitis similar to sclerosing cholangitis in humans.92 The toxic injury hypothesis, however, does not explain why PSC is not associated with the severity of colonic inflammation in patients with IBD and why PSC may develop years after a patient has undergone total proctocolectomy.

Vascular Factors

Ischemia has been postulated to play a role in the pathogenesis of PSC because a similar cholangiographic appearance may be found after surgical trauma to the biliary vascular supply93 and after hepatic artery thrombosis or arterial fibrointimal hyperplasia after liver transplantation.94,95 In addition, PSC is associated with the presence of autoantibodies such as pANCA and anti-cardiolipin antibodies. These autoantibodies, in turn, are strongly associated with vasculitides such as Wegener’s granulomatosis, polyarteritis nodosa, and thrombotic syndromes. These associations suggest that immune-mediated vascular injury plays a role in the pathogenesis of PSC.

NATURAL HISTORY AND PROGNOSTIC MODELS

PSC is typically a progressive disease, although the natural history is incompletely understood.29-31,96-98 The disease may be considered to progress through the following four clinical phases, although some phases may not develop or be apparent in an individual patient: 1. Asymptomatic phase: Patients may have cholangiographic evidence of PSC but normal serum liver biochemical values and no symptoms. These patients typically are identified as a result of incidental findings on imaging studies. 2. Biochemical phase: Patients remain asymptomatic but have biochemical abnormalities, typically elevations of serum alkaline phosphatase levels with variable elevations of serum bilirubin and aminotransferase levels. 3. Symptomatic phase: Symptoms of cholestasis or liver injury, or both, develop. Pruritus, fatigue, symptoms of cholangitis, and jaundice may occur, often in combination. 4. Decompensated cirrhosis: The final phase is characterized by worsening symptoms and complications of endstage liver disease, such as ascites, encephalopathy, and variceal bleeding.

1157

1158

Section VIII  Biliary Tract This model of disease progression provides a useful framework for understanding the variability of the natural history in studies of patients with PSC.

Asymptomatic Primary Sclerosing Cholangitis

Asymptomatic patients with PSC make up 15% to 44% of cohorts examined in published studies29,31,32,98,99 Some reports have suggested that asymptomatic patients typically have a benign course of disease. Helzberg and colleagues98 reported on 11 asymptomatic patients with PSC who were followed for a mean of 37 months, and all 11 remained asymptomatic without evidence of progressive disease. By contrast, Porayko and colleagues99 followed 45 asymptomatic patients with PSC for a median of 6.25 years, and during the surveillance period, liver failure, resulting in liver transplantation or death, developed in 13 (31%). Overall, symptoms developed in 24 (53%), and progressive liver disease, demonstrated by new symptoms or signs, worsening cholangiographic findings, or progressive liver histologic abnormalities, developed in 34 (76%) patients. The Kaplan-Meier estimate of median survival free of liver failure in this study was 71% at seven years for the asymptomatic patients, significantly lower than the 96% expected on the basis of an age-, sex-, and race-matched U.S. control population. Differences in the rates of progression between these studies98,99 may be the result of differences in patient populations, the definition of “asymptomatic,” and the duration of clinical follow-up.

Symptomatic Primary Sclerosing Cholangitis

Patients with symptoms at the time of diagnosis generally have a worse prognosis than asymptomatic patients.29,30 The clinical stage is likely more advanced at the time of diagnosis in symptomatic patients, who have more severe biochemical derangements, more abnormalities on cholangiography, and a higher histologic stage on liver biopsy specimens than asymptomatic patients. Wiesner and colleagues29 compared the natural history of PSC in 37 asymptomatic patients with that in 137 patients who were symptomatic at the time of diagnosis. After a mean followup of six years, 55 (40%) of the symptomatic patients had died, compared with 4 (11%) in the asymptomatic group. The Kaplan-Meier estimate of median survival for the entire cohort was 11.9 years; for the symptomatic cohort, the estimated median survival was between 8 and 9 years. Farrant and colleagues30 described the natural history of PSC in 126 patients, of whom 84% were symptomatic. After a median follow-up of 5.8 years, the estimated median survival was 12 years. Similar findings were reported in a large study by Broome and colleagues.31 In 305 patients with PSC followed for a median of 5.25 years, of whom 44% were asymptomatic, the estimated median survival was 12 years. Patients who were symptomatic at the time of entry into the study had a significantly worse expected survival (9.3 years) than asymptomatic patients. A study of 174 patients with PSC by Ponsioen and colleagues97 suggested a better overall prognosis, with a median expected survival of 18 years. The reason for improved survival in this most recent study is not known, but patient data were predominantly from the 1990s, compared with data from the 1970s and 1980s in the other studies described. Although therapeutic advances were not dramatic in the interim, earlier diagnosis in the 1990s may have led to differences in patient selection that appeared to affect outcomes.

Small-Duct Primary Sclerosing Cholangitis

Patients who have histologic, biochemical, and clinical features of PSC but a normal cholangiogram are considered

to have small-duct PSC, which accounts for 5% to 20% of all patients with PSC.3,100 Three studies have performed extended clinical follow-up in patients with small-duct PSC.100-102 In these studies, 12% to 17% of patients progressed to classic large-duct PSC over long-term follow-up, although the true rate may be higher because cholangiograms were not obtained routinely in all patients. Cholangiocarcinoma did not develop in any patient over a median follow-up of 63 to 126 months, and survival in the smallduct PSC group was better than that of matched control groups with classic PSC.100-102 Therefore, small-duct appears to represent an early stage of PSC, may progress to largeduct PSC in a small percentage of patients, and is associated with a better prognosis than classic PSC.

Prognostic Models

Natural history studies have provided insight into specific clinical, biochemical, and histologic features of PSC that may influence prognosis. Whereas early studies described individual factors that were associated with poor survival in PSC, subsequent studies utilized multivariable regression analysis techniques, such as Cox proportional hazards analysis, to develop more sophisticated mathematical models to predict survival. Such models predict expected survival for a specific patient at a specific time. This information is essential for counseling patients about their prognosis and planning future care, such as liver transplantation. Multivariable prognostic models that have been developed to predict survival in patients with PSC are shown in Table 68-2. In an early multivariable analysis, hepatomegaly and a serum bilirubin level > 1.5 mg/dL were found to be independently associated with a poor prognosis in PSC. The patient’s age, histologic findings, presence of concomitant IBD, and pattern of cholangiographic involvement did not correlate independently with survival in this study.98 Wiesner and colleagues29 developed a prognostic model based on age, serum bilirubin level, hemoglobin value, presence or absence of IBD, and histologic stage. With this model, three risk groups (low, intermediate, high) were formed, and predicted survival curves were shown to be similar to observed survival curves. Farrant and colleagues30 developed a multivariable prognostic model in which hepatomegaly, splenomegaly, serum alkaline phosphatase level, histologic stage, and age at presentation were found to be important independent factors. Dickson and colleagues103 then presented a model developed from a multicenter collaboration in which data from 426 patients with PSC were pooled. In this analysis, the patient’s age, serum bilirubin level, histologic stage, and presence of splenomegaly were found to correlate independently with survival, and the model was validated against the observed survival data in a subgroup of the entire cohort. Broome and colleagues31 found the patient’s age, histologic stage, and serum bilirubin level to be independent predictors of survival in 305 patients with PSC, but this prognostic model was not validated independently. Most recently, Kim and colleagues,104 using easily obtainable clinical and biochemical factors, revised an earlier predictive model that did not require liver biopsy and did not rely on subjective physical findings such as splenomegaly or hepatomegaly. This revised natural history model (revised Mayo risk model) found the patient’s age, serum bilirubin level, serum aspartate aminotransferase (AST) level, and serum albumin level and a history of variceal bleeding to be independent predictors of survival. The model was generated from data on 529 patients from 5 centers and was validated using data from another center that had not been used in the development of the model.

Chapter 68  Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis

Table 68-2  Independent Predictors of Survival and Prognostic Index Formulas Used in Natural History Models of Primary Sclerosing Cholangitis* MAYO CLINIC MODEL29 Predictors of Survival† Age Bilirubin Histologic stage Hemoglobin IBD Prognostic Index Formula‡ R = 0.06 × age + 0.85 × loge(min[bilirubin or 10]) 4.39 × loge(min[hemoglobin or 12]) + 0.51 × histologic stage + 1.59 × indicator for IBD

KING’S COLLEGE MODEL30

MULTICENTER MODEL103

REVISED MAYO MODEL104

Age Hepatomegaly Histologic stage Splenomegaly Alkaline phosphatase

Age Bilirubin Histologic stage Splenomegaly

Age Bilirubin Albumin AST Variceal bleeding

R = 1.81 × hepatomegaly + 0.88 × splenomegaly + 2.66 × log(alk phos) + 0.58 × histologic stage + 0.04 × age

R = 0.535 × loge(bilirubin) + 0.486 × histologic stage + 0.041 × age + 0.705 × splenomegaly

R = 0.03 × age + 0.54 × loge(bilirubin) + 0.54 × loge(AST) + 1.24 × variceal bleeding − 0.84 × albumin

*Superscript numbers indicate references. † Age expressed in years, bilirubin in mg/dL; hemoglobin in gm/dL; alkaline phosphatase in U/L; AST in U/L; albumin in g/dL. ‡ Values for IBD, hepatomegaly, splenomegaly, and variceal bleeding are 1 if present, 0 if absent. Alk phos, alkaline phosphatase; AST, aspartate aminotransferase; IBD, inflammatory bowel disease; min, minimum of; R, risk score.

The Child (Child-Pugh) classification may also be used to predict survival in patients with PSC (see Chapter 90). Shetty and colleagues105 found that Kaplan-Meier sevenyear survival rates for patients with Child class A, B, and C cirrhosis caused by PSC were 89.8%, 68%, and 24.9%, respectively. Subsequent evaluation, however, suggested that the Child classification is less accurate than the revised Mayo risk model, especially for patients with early-stage PSC.106 Despite the cumbersome nature of the mathematical formulas used in the various natural history models, these models can be useful in the clinical care of patients with PSC. They may facilitate selection for and timing of liver transplantation by comparing predicted survival with readily available post-liver transplantation survival rates. The availability of multiple models with differing prognostic variables, however, may be confusing in clinical practice. The models also may not account for other clinical events, such as the development of cholangiocarcinoma or variceal bleeding, that may affect prognosis in patients with PSC. Further refinement of these prognostic models, including consensus on the use of specific prognostic variables, may ultimately clarify their role in clinical practice.

CLINICAL FEATURES Symptoms

The initial clinical presentation of PSC can be quite varied and may run the gamut from asymptomatic elevations of serum alkaline phosphatase levels to decompensated cirrhosis with jaundice, ascites, hepatic ence­ phalopathy, or variceal bleeding. The most common symptoms at the time of presentation include jaundice, fatigue, pruritus, and abdominal pain.19,28-33,107,108 Other associated symptoms may include fever, chills, night sweats, and weight loss (Table 68-3). The onset of these symptoms is typically insidious, although an acute hepatitis-like presentation has been described.40 Increasingly, PSC is diagnosed in an asymptomatic or minimally symptomatic stage. Large series have shown that 15% to 44% of patients with PSC are asymptomatic at the time of diagnosis,29,31,32,98,99 probably because of the routine liver biochemical screening in patients with IBD, as well as the

Table 68-3  Most Common Symptoms and Signs at the Time of Diagnosis of Primary Sclerosing Cholangitis Symptoms Fatigue Abdominal pain Pruritus Fever/night sweats Asymptomatic Weight loss Signs Jaundice Hepatomegaly Splenomegaly Hyperpigmentation Ascites

Rate (%) 65-75 24-72 15-69 13-45 15-44 10-34 30-73 34-62 32-34 14-25 4-7

Data from references 19, 29-33, 98, 99, 107, 108.

widespread availability of MRCP and ERCP for evaluating elevated serum alkaline phosphatase levels. Symptoms of PSC are often intermittent. Episodes of pruritus, jaundice, abdominal pain, and fever are typically interspersed with asymptomatic periods of varying duration.39,107,108 The intermittency of the symptoms is thought to reflect intermittent biliary obstruction caused by microlithiasis and sludge.5,109 This obstruction may predispose to cholestasis and induce an acute inflammatory reaction. Secondary bacterial infection may result in low-grade cholangitis and predispose to pigment stone formation.5

Physical Examination

Physical findings may be normal in patients with PSC, particularly those who are asymptomatic. When physical abnormalities are present, the most common include hepatomegaly, jaundice, and splenomegaly (see Table 68-3). Skin findings are common and include cutaneous hyperpigmentation, excoriations resulting from pruritus, and xanthomata. As liver disease progresses, spider angiomas, muscular atrophy, peripheral edema, ascites, and other signs of advanced liver disease may appear.28-30

1159

1160

Section VIII  Biliary Tract Laboratory Findings

Chronic elevation of serum alkaline phosphatase levels, typically three to five times normal, is the biochemical hallmark of PSC. A normal alkaline phosphatase level, however, may be found in up to 6% of patients with cholangiographically proved PSC.110,111 In some cases, an advanced histologic stage has been demonstrated on a liver biopsy specimen despite normal serum alkaline phosphatase levels.110 Serum aminotransferase levels are typically elevated, although rarely above four to five times normal except in the pediatric population.112 The serum bilirubin level may be normal or elevated and often fluctuates. When the serum bilirubin level is elevated, the bilirubin is predominantly conjugated. Reductions in the serum albumin level and prolongation of the prothrombin time may reflect hepatic synthetic dysfunction with advanced liver disease. In addition, malnutrition and underlying IBD may lower serum albumin levels. Vitamin K malabsorption related to cholestasis may play a role in prolonging the prothrombin time. Other nonspecific consequences of cholestasis are elevations in serum copper, serum ceruloplasmin, and hepatic copper levels, increased urinary copper excretion, and elevated serum cholesterol levels. Several immunologic markers and serum autoantibodies are found in the majority of patients with PSC, although none is specific for the disease. Hyperglobulinemia is frequent; serum IgM levels are elevated in up to 50% of patients, and IgG and IgA levels also may be elevated.28,29,28,107 Antinuclear antibodies, often in low titer, may be detected in 24% to 53% of patients. Anti-smooth muscle antibodies are found in 13% to 20% of patients, but antimitochondrial antibodies are found in less than 10%.26,28,70,73,75 Most commonly found in patients with PSC are pANCA,72 which are detected in 65% to 88% of patients and appear to react to a heterogeneous group of antigens.70,71,74,75 These antigens have been found to represent neutrophil nuclear envelope proteins predominantly, and the corresponding antibodies have been referred to as “antineutrophil nuclear antibodies” (ANNA).113 In contrast to Wegener’s granulomatosis, titers of pANCA do not appear to correlate with disease activity, severity, or response to medical therapy in patients with PSC.72 Furthermore, the presence of autoantibodies does not appear to differ in patients with and without IBD. Anticardiolipin antibodies are detected in 66% of patients with PSC, and the titer has been reported to correlate with disease severity.75 In general, despite the high frequency of autoantibodies in patients with PSC, a clear association between the presence of these antibodies, pathogenesis of the disease, and prognosis or response to treatment remains unproved. Measurement of autoantibodies is therefore of limited clinical value in patients with PSC.

Imaging Findings

Cholangiography by ERCP, MRCP, or percutaneous THC establishes a diagnosis of PSC and provides information regarding the distribution and extent of disease. The characteristic cholangiographic findings include multifocal stricturing and ectasia of the biliary tree. Areas of narrowing are interspersed with areas of normal or near-normal caliber and areas of post-stenotic dilatation. The result is a classic “beaded” appearance to the biliary tree. The strictured segments are usually short, annular, or band-like in appearance (Fig. 68-2), although longer confluent strictures may be seen in more advanced disease. Localized segments of dilated ducts may have a saccular or diverticular appearance. Major areas of focal, tight narrowing known as dominant strictures may be seen and often involve the bifurcation of the hepatic duct.114 At times, diffuse involvement of the intrahepatic

A

B Figure 68-2.  Endoscopic retrograde cholangiopancreatography (ERCP) in two patients with primary sclerosing cholangitis (PSC). A, ERCP with contrast injected through a balloon catheter (seen in distal bile duct). The intrahepatic ducts are mainly affected and show diminished arborization (pruning), with diffuse segmental strictures alternating with normal-caliber or mildly dilated duct segments (cholangiectasias), resulting in a beaded appearance. B, ERCP in a patient with PSC. Radiologic features include diffuse irregularity of the intrahepatic ducts, multiple short strictures and cholangiectasias, small diverticula in the wall of the common hepatic duct (arrow), and clips from a prior cholecystectomy.

biliary tree may give a pruned appearance that is difficult to distinguish from the diffuse intrahepatic duct attenuation seen in patients with cirrhosis of any cause; irregularity of the duct wall or concomitant involvement of the extrahepatic bile duct supports a diagnosis of PSC. Both the extrahepatic and intrahepatic bile ducts are abnormal in approximately 75% of cases. The intrahepatic ducts alone may be involved in 15% to 20% of cases.28,31,35,107 Abnormalities of the extrahepatic biliary tree in the absence of intrahepatic involvement are less common.98,99 The cystic duct and gallbladder may be involved in up to 15% of patients but may not be well visualized on routine cholangiography.115 Pancreatic duct irregularities similar to those seen in chronic pancreatitis may rarely be noted.116

PATHOLOGY

Gross and histologic specimens from the extrahepatic bile ducts demonstrate a diffusely thickened, fibrotic duct wall.

Chapter 68  Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis

A

B Figure 68-3.  Liver histopathology in primary sclerosing cholangitis. A, A segmental bile duct is completely obliterated by fibrosis (arrow), demonstrating “fibro-obliterative cholangitis.” (Hematoxylin and eosin, ×200.) B, A medium-sized bile duct is surrounded by concentric fibrosis with an onionskin appearance. (Hematoxylin and eosin, ×400.) (A and B, Courtesy of Matthew Yeh, MD, PhD, Seattle, Wash.)

The fibrosis is accompanied by a mixed inflammatory infiltrate that may involve the epithelium and biliary glands.117,118 Florid hyperplasia of the biliary glands with accompanying neural proliferation has been described.119 Examination of PSC explants removed at the time of liver transplantation has demonstrated areas of thin-walled saccular dilatation, termed “cholangiectasias,” that correspond to the beaded appearance on cholangiography.120 A wide range of liver biopsy findings may be seen in patients with PSC. For this reason, histologic findings are not typically diagnostic for PSC. The characteristic bile duct lesion is a fibro-obliterative process that may lead to an “onionskin” appearance of concentric fibrosis surrounding medium-sized bile ducts (Fig. 68-3); however, this finding is seen in less than one half of biopsy specimens.117,121,122 The smaller interlobular and septal bile duct branches may be entirely obliterated by this process, resulting in fibroobliterative cholangitis. This finding is present in only 5% to 10% of biopsy specimens but is thought to be virtually pathognomonic of PSC.121 In this process, the biliary epithelium may degenerate and atrophy and be replaced entirely by fibrous cords. Other characteristic histopathologic find-

ings may include bile duct proliferation, periductal inflammation, and ductopenia. The degree of inflammation can be quite variable but is typically a portal-based mixture of lymphocytes, plasma cells, and neutrophils with a periductal focus. Lymphoid follicles or aggregates may also be seen.121,122 The histologic progression of PSC can be divided into four stages, analogous to a similar staging system in PBC46 (See Chapter 89). In stage I (portal stage) changes are confined to the portal tracts and consist of portal inflammation, connective tissue expansion, and cholangitis. Stage II (periportal stage) is characterized by expansion of inflammatory and fibrotic processes beyond the confines of the limiting plate, resulting in “piecemeal necrosis” (interface hepatitis) and periportal fibrosis. Depending on the degree of biliary obstruction, ductular proliferation and cholangitis may be of varying severity. Stage III (septal stage) is characterized by fibrous septa that bridge from one portal tract to the next. Bridging necrosis may occasionally be seen but is uncommon. Stage IV (cirrhotic stage) implies progression to biliary cirrhosis. The degree of inflammatory activity may subside as the stage of disease progresses, and focal bile ductular proliferation may be striking. A study that examined the time course of progression through the histologic stages revealed that for patients with stage II disease, 42%, 66%, and 93% progressed to a higher histologic stage at one, two, and five years, respectively.123 For patients initially with stage II disease, progression to biliary cirrhosis (stage IV) occurred in 14%, 25%, and 52%, respectively. Regression of stage was observed in 15% of patients and probably reflected sampling variability in the histologic assessment. Many of the histologic findings of PSC are nonspecific and may be seen in other disorders. In particular, the histologic distinction between PSC and PBC may be difficult to discern. In one study, histologic examination could classify only 28% of patients who had one of the two diseases.18 When lymphocytic interface hepatitis is prominent, the distinction from autoimmune hepatitis may be challenging, especially because hypergammaglobulinemia and autoantibodies may be present in both conditions. In addition, an overlap syndrome with features of both PSC and autoimmune hepatitis has been described.21,22,66 When severe cholestasis develops, hepatic copper accumulation can be dramatic and may mimic that seen in Wilson disease.124

COMPLICATIONS Cholestasis

The complications associated with all causes of chronic cholestasis may develop in patients with PSC (see also Chapters 20 and 89). Pruritus is one of the most common symptoms of PSC and may adversely affect a patient’s quality of life. Severe excoriations and debilitating symptoms may develop. The pathogenesis of pruritus in chronic cholestasis is poorly understood, and response to therapy is inconsistent (see Chapter 89). The accumulation of bile acids in the plasma and tissue of cholestatic patients has been cited as a potential cause of pruritus, and the pruritus of cholestasis is typically treated with oral administration of bile-acid binding resins such as cholestyramine. Not all patients with elevated serum bile acid levels itch, however. In addition, pruritus is frequently intermittent, despite the relative stability of serum bile acid levels. Several lines of evidence suggest that cholestasis is asso­ ciated with an increased level of endogenous opioids. In animal models, cholestasis is associated with an increase in plasma levels of endogenous opioids.125 In humans, cholestatic patients may experience opiate withdrawal-like

1161

1162

Section VIII  Biliary Tract symptoms after the administration of an opioid antagonist. In addition, administration of naloxone and naltrexone, which have opioid antagonist properties, has been reported to relieve pruritus in cholestatic patients in small clinical trials.126,127 Nutritional deficiencies may complicate chronic chole­ stasis in patients with PSC. Intestinal absorption of the fatsoluble vitamins A, D, E, and K is particularly affected and is thought to be related to decreased intestinal concentrations of conjugated bile acids.128 Concomitant disease such as IBD, chronic pancreatitis, and celiac disease may also contribute to intestinal malabsorption. Clinical consequences include night blindness (vitamin A deficiency) and coagulopathy (vitamin K deficiency). The importance of metabolic bone disease, also referred to as hepatic osteodystrophy, is often underrecognized in patients with PSC. Two forms of metabolic bone disease may develop: osteomalacia and osteoporosis. With improvements in nutritional management, osteomalacia (decreased bone mineralization) is now relatively rare, and most bone disease in cholestatic patients is osteoporosis. Bone mineral density is significantly lower in patients with PSC than in age- and sex-matched controls.129 The pathogenesis of bone density loss in PSC and other chronic cholestatic liver diseases is unknown. Intestinal malabsorption of vitamin D is probably not the primary abnormality because serum vitamin D levels are often normal and vitamin D repletion does not usually have a major impact on the severity of osteoporosis. In patients with concomitant IBD, the use of glucocorticoids may play a role in exacerbating bone loss in patients with PSC. Overall, severe osteoporosis is less common in patients with PSC than in those with PBC because a majority of patients with PSC are young men who have a higher baseline bone mineral density and a slower rate of bone loss than middle-aged women, who account for most cases of PBC.

Biliary Stones

Cholelithiasis and choledocholithiasis are more common in patients with PSC than in the general population. Gallstones are found in approximately 25% of patients with PSC and are often pigmented calcium bilirubinate stones.5 Biliary strictures may predispose to bile stasis and intraductal sludge and stone formation. Ultrasonography has only an intermediate sensitivity for detecting intraductal stones. Therefore, patients with PSC and worsening cholestasis or jaundice should undergo ERCP to distinguish biliary stone disease from the development of a dominant stricture or cholangiocarcinoma.

Cholangiocarcinoma

Cholangiocarcinoma is a feared complication of PSC and can arise from bile duct epithelium anywhere in the biliary tract (see Chapter 69). PSC should be considered a premalignant condition of the biliary tree, analogous to the relationship between UC and carcinoma of the colon. The reported frequency of cholangiocarcinoma in patients with PSC has ranged from 6% to 11% in natural history studies and from 7% to 36% in patients with PSC who undergo liver transplantation.130,131 Tumors are most commonly found in the common hepatic duct and perihilar region but may involve only the bile duct, intrahepatic ducts, cystic duct, or gallbladder. Cholangiocarcinoma remains the leading cause of death in patients with PSC. The pathogenesis of cholangiocarcinoma in PSC is poorly understood. Although cholangiocarcinoma may complicate any stage of the disease, chronic inflammation is thought to predispose to epithelial dysplasia and an increased risk of

malignant transformation. The role of chronic inflammation is supported by the observation that patients with chronic Clonorchis sinensis and other liver fluke infections are also at increased risk of cholangiocarcinoma (see Chapter 82).132 A role for proinflammatory cytokines in stimulating oxidative DNA damage and inactivation of DNA repair processes has been postulated. Biliary malignancy should be suspected when a patient with PSC exhibits rapid clinical deterioration with worsening jaundice, weight loss, and abdominal pain. Advanced PSC without cholangiocarcinoma may present with the identical clinical presentation. The diagnosis of cholan­ giocarcinoma presents a particular challenge in patients with PSC. A malignant biliary stricture may be indistinguishable cholangiographically from the underlying PSC (Fig. 68-4). Because of the tendency of cholangiocarcinoma to grow in sheets as opposed to a discrete mass, crosssectional imaging with CT or magnetic resonance imaging (MRI) may be insensitive for detection of cholangiocarcinoma. Several serum tumor markers of cholangiocarcinoma have been evaluated. Serum CA 19-9 has been the most commonly utilized tumor marker, with one small study reporting a sensitivity of 89% and specificity of 86% for a serum CA 19-9 level greater than 100 U/mL in diagnosing cholangiocarcinoma.133 A later study from the same group found a lower sensitivity for serum CA 19-9 but a correlation between the CA 19-9 level and tumor stage; no tumor was resectable in patients with a CA 19-9 level greater than 1000 U/mL.134 Another study described a biochemical index using CA 19-9 and carcinoembryonic antigen (CEA) levels (CA 19-9 + (CEA × 40) > 400), with a reported sensitivity of 86% for cholangiocarcinoma.135 More recent studies, however, suggested a poor sensitivity (33%) for this combined biochemical index despite a relatively high specificity (85%).136,137 Obtaining an adequate tissue sample presents a particular challenge in the diagnosis of cholangiocarcinoma. Dominant strictures resulting from PSC may be indistinguishable cholangiographically from those harboring cholangiocarcinoma. Brush cytology can be obtained at the time of ERCP, but the sensitivity of this approach is only 50% to 60% at best.138,139 False-positive results are also possible because chronically inflamed cells may take on a malignant cytologic appearance. The addition of a sampling technique to brush cytology, such as endobiliary biopsy or fine-needle aspiration (FNA), improves sensitivity,138-140 and when clinical or cholangiographic suspicion for cholangiocarcinoma is high, two tissue sampling techniques should be used. The sensitivity of cytologic examination is increased by use of specialized techniques such as fluorescent in situ hybridization and digital image analysis. Endoscopic ultrasound (EUS) with FNA has an emerging role in the evaluation of suspected cholangiocarcinoma when brush cytology and other methods have failed to yield a diagnosis.141,142 The sensitivity of EUS-FNA for diagnosing cholangiocarcinoma in patients with PSC has been reported to be as high as 89%.141,142 High-frequency intraductal ultrasound (IDUS) is under study as a means of discriminating benign from malignant dominant bile duct strictures, and several studies143-145 have demonstrated that the addition of IDUS to ERCP improves the ability to distinguish benign from malignant dominant bile duct strictures in patients with PSC. Direct endoscopic visualization of bile duct strictures is also possible with the availability of choledoscopy (cholangioscopy) in clinical practice. A single small study has shown that the addition of choledoscopy to ERCP in patients with PSC and a dominant stricture

Chapter 68  Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis The development of cholangiocarcinoma is an ominous sign, with a median survival of five months after diagnosis.6 In addition, recurrence of cholangiocarcinoma after liver transplantation is nearly universal, and cholangiocarcinoma is generally considered a contraindication to liver transplantation. Early identification of patients with PSC who are at high risk of cholangiocarcinoma is crucial so that liver transplantation may be undertaken before the bile duct cancer develops. One study reported a strong association between current or former smoking and the development of cholangiocarcinoma in patients with PSC.147 A second study, however, could not confirm an association with smoking.130 In addition, duration of PSC, distribution of biliary strictures, and past medical or surgical intervention do not appear to be associated with an increased risk of cholangiocarcinoma. Longer duration of IBD has been shown to be a risk factor for cholangiocarcinoma in some,148 but not all,130 studies. Overall, therefore, studies to date have not defined clear risk factors for cholangiocarcinoma that are clinically useful for identifying patients with PSC at particularly high risk. Nor has the optimal surveillance strategy been delineated.

A

B Figure 68-4.  Cholangiographic progression of cholangiocarcinoma complicating primary sclerosing cholangitis (PSC). This 43-year-old man with a history of a mild ulcerative colitis for 3 years presented with pruritus and an elevated serum alkaline phosphatase level. A, Initial endoscopic retrograde cholangiopancreatography (ERCP) showed mild, diffuse changes compatible with uncomplicated PSC. Several short, annular strictures and cholangiectasias are present in the intrahepatic ducts (arrows), with a single, short, annular stricture of the bile duct. The cystic duct is not filled and may already be obstructed by tumor. B, Second ERCP performed seven months later after progressive jaundice and weight loss developed rapidly. Now a 2-cm mass is projecting into, and obstructing, the common hepatic duct (arrows). A catheter has been passed beyond the obstructing mass. There is marked dilatation of the left main duct proximal to the obstruction; the right main duct is completely occluded.

enhanced detection of malignancy, with an overall sensitivity of 92%.146 Therefore, in patients with PSC, suspected cholangiocarcinoma, and negative brush cytology or endobiliary biopsy results, repeat ERCP plus EUS-FNA, IDUS, or choledoscopy improves the sensitivity of diagnosing cholangiocarcinoma.

Colonic Neoplasia

Most patients with PSC also have IBD; the majority have ulcerative pancolitis. UC is known to be associated with an increased risk of colonic dysplasia and carcinoma (see Chapter 112); the risk of colon cancer increases with the duration, extent, and severity of colitis. For patients with pancolitis, the cumulative risk of colon cancer is approximately 5% to 10% after 20 years and 12% to 20% after 30 years.149-151 A growing body of evidence suggests that patients with concomitant PSC and UC are at significantly higher risk for developing colonic neoplasia (dysplasia or carcinoma) than patients with UC alone.152-157 In one report, 132 patients with UC and PSC were compared with a randomly selected historical cohort of patients with UC but without PSC.155 Colonic carcinoma or dysplasia developed in significantly more patients with both UC and PSC than those with UC alone (25% versus 6%), and significantly more deaths from colorectal cancer were observed in the patients with PSC (4.5% versus 0%). A similarly designed study included 20 patients with both UC and PSC and 25 matched controls with UC alone.153 Colonic dysplasia was observed significantly more often in patients with both UC and PSC (45% versus 16%), although the time to the development of dysplasia was similar in the two groups. Patients with PSC and UC were also more likely than patients with UC alone to have synchronous sites of dysplasia in the colon. Another study examined 40 patients with PSC and UC matched with 80 control subjects with UC alone.152 The cumulative risk of colorectal dysplasia or carcinoma in patients with both UC and PSC was 9%, 31%, and 50% after 10, 20, and 25 years, respectively. These rates were significantly higher than those for the control group (2%, 5%, and 10%, respectively). The mechanisms by which PSC confers an added risk of colonic neoplasia are not well understood. A high colonic concentration of secondary bile acids may play a role because patients with UC and colonic dysplasia or carcinoma have higher fecal bile acid concentrations than patients with UC who do not have dysplasia or carcinoma.158 This theory is supported by the higher incidence of rightsided colon cancer in patients with UC and PSC than in those with UC alone.155,156 In the study by Shetty and colleagues,155 76% of the colon cancers in patients with UC and PSC occurred proximal to the splenic flexure, compared with only 20% in patients with UC alone.155 Increased

1163

1164

Section VIII  Biliary Tract colonic secondary bile acid concentrations may also explain the apparent chemoprotective effect of ursodeoxycholic acid (UDCA) against the development of colonic neoplasia. Two studies have reported that UDCA use is associated with a lower risk of colonic dysplasia or cancer in patients with UC and PSC.159,160 UDCA may confer protection against colonic neoplasia by reducing colonic concentrations of secondary bile acids, as well as by affecting expression of protein kinase C isoforms, metabolism of arachidonic acid, and expression of cyclooxygenase-2.159-162 These data need to be confirmed in larger-scale, prospective trials before UDCA use can be recommended routinely for preventing colon cancer in patients with PSC. Whether UDCA has a chemoprotective effect in patients with UC but without PSC is unknown. Patients with PSC who have UC should undergo annual colonoscopic surveillance for the detection of colonic dysplasia or cancer. As for colonoscopic surveillance in patients with UC alone, multiple mucosal biopsy specimens should be obtained (see Chapter 112).163 Most experts agree that colonoscopic surveillance should start immediately after the diagnosis of PSC. Surveillance should continue even after liver transplantation because these patients remain at increased risk for colonic neoplasia.164-166

Peristomal Varices

Varices at the stoma may develop in patients with PSC and portal hypertension who have previously undergone proctocolectomy with ileostomy for IBD.167,168 These varices may bleed spontaneously, and the bleeding may be dramatic. Treatment modalities that may initially be effective in achieving hemostasis include injection sclerotherapy,169 percutaneous transhepatic coil embolization,170 surgical stomal revision,171 and transjugular intrahepatic portosystemic shunt placement (see Chapter 90).172 Nevertheless, recurrent bleeding is common, and liver transplantation should be considered to relieve portal hypertension and treat the underlying liver disease.

TREATMENT

Except for liver transplantation, no specific therapy has proved effective for treating PSC. The objectives of management should be to treat the complications of disease, such as flares of bacterial cholangitis, jaundice, and pruritus, and prevent complications, such as osteoporosis and nutritional deficiencies. Other complications such as cholangiocarcinoma and liver failure should be diagnosed as early as possible to allow the possibility of treatment.

Medical Treatment of Underlying Disease

A wide variety of medications have been studied in patients with PSC (Table 68-4). Many of the published studies have been small and uncontrolled with limited follow-up. Because of the varied course of PSC, with spontaneous remissions and unpredictable flares, adequate clinical trials in patients with PSC require long-term follow-up. In addition, the defined study endpoints, whether clinical, biochemical, histologic, or a mathematical risk score, have varied greatly among published studies. To date, no medical treatment has been shown clearly to alter the course of PSC. UDCA has been the most extensively studied drug in patients with PSC. At least five controlled clinical trials have been reported, with varying doses of UDCA.173-177 The mechanisms by which UDCA is thought to exert a beneficial effect in cholestatic conditions include protection of cholangiocytes against cytotoxic hydrophobic bile acids, stimulation of hepatobiliary secretion, protection of hepatocytes against bile-acid induced apoptosis, and induction of anti-

Table 68-4  Medical Therapy for Primary Sclerosing Cholangitis* NO PROVEN BENEFIT

POSSIBLE BENEFIT

Antibiotics Cholestyramine Glucocorticoids183,184 Azathioprine Methotrexate187 Cyclosporine Tacrolimus188 Pentoxifylline189 Colchicine191 d-penicillamine192 Nicotine193,194 Perfenidone195

Ursodeoxycholic acid (20-30 mg/kg/day)173-177,180

*Superscript numbers indicate references.

oxidants (see also Chapter 89).178,179 In a randomized, controlled trial by Beuers and colleagues,173 patients with PSC treated with UDCA, 13 to 15 mg/kg/day for one year, had significant improvements in biochemical and histologic endpoints in comparison with those given placebo, but no effect on symptoms was noted. In the largest controlled trial of UDCA for PSC to date, Lindor and colleagues176 randomized 105 patients to UDCA, 13 to 15 mg/kg/day, or placebo for a median of 2.2 years. Significant biochemical improvements were seen with UDCA therapy, but no difference was seen in the primary endpoint, which was a composite of death, liver transplantation, histologic progression of at least two stages, hepatic decompensation, or quadrupling of the serum bilirubin level. Because of the disappointing results with standard-dose UDCA, several groups have published results on the use of high-dose UDCA in the treatment of PSC. Mitchell and colleagues177 performed a two-year controlled trial of high-dose UDCA, 20 mg/kg/day, versus placebo in 26 patients with PSC. Compared with placebo, treatment with high-dose UDCA led to improvement in biochemical markers and a reduction in the progression of fibrosis and cholangiographic changes. In addition, high-dose UDCA was well tolerated, with no significant adverse events. Harnois and colleagues180 performed an open-label pilot study of UDCA, 25 to 30 mg/kg/day, for one year in 30 patients with PSC and compared changes in the revised Mayo risk score with those in patients from a previous randomized trial of standard-dose UDCA. After one year of treatment, improvement in the revised Mayo risk score in the high-dose UDCA group was significantly greater than that seen in placebotreated patients from the prior study but no better than that in the patients treated with standard-dose UDCA. With the encouraging results of these two small studies, a larger study was undertaken in which 219 patients with PSC were randomized to high-dose UDCA or placebo. After five years, no differences in symptoms, liver biochemical abnormalities, quality of life, or transplant-free survival were found between the two groups.181 A subsequent study in 31 patients randomized to low-dose, standard-dose, and high-dose UDCA found improvement in the revised Mayo risk score in all groups, with statistical significance only in the group receiving high-dose UDCA.182 Therefore, the precise benefit of UDCA in general, and high-dose UDCA in particular, remains uncertain. Given the immunologic alterations in patients with PSC, immunosuppressive therapy would appear to be a reasonable consideration. Glucocorticoids, administered both

Chapter 68  Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis orally and via nasobiliary lavage, have not shown a clear benefit in uncontrolled studies,183,184 although subgroups of patients may have a clinical response.185 Lack of long-term data demonstrating clear response and concerns about longterm adverse effects, including exacerbation of metabolic bone disease, have limited the use of glucocorticoids. Oral budesonide, a newer glucocorticoid with limited systemic toxicity, has been evaluated in an uncontrolled pilot study in 21 patients with PSC.186 After one year of therapy, treated subjects had minimal biochemical or histologic improvement and no change in revised Mayo risk score. In addition, significant loss of bone mass was seen with use of budesonide. Other immunomodulators also have been evaluated. In a small prospective, controlled trial of methotrexate, no biochemical, histologic, or cholangiographic differences from therapy with placebo were seen after two years of treatment.187 In a small pilot study of tacrolimus therapy, significant biochemical improvement was observed after one year, but no change in cholangiographic or histologic severity was documented.188 Pentoxifylline, which inhibits tumor necrosis factor-α (TNF-α), led to no biochemical or symptomatic improvements in a one-year pilot study in 20 patients with PSC.189 Etanercept, a recombinant inhibitor of TNF-α, also showed no benefit in a small number of patients with PSC.190 Colchicine has been evaluated as a potential therapy for PSC because of its antifibrogenic potential. In a randomized, controlled clinical trial comparing colchicine therapy with placebo for three years, no differences were seen between the two groups in rates of mortality or liver transplantation or in symptoms or biochemical and histologic findings.191 D-penicillamine has also been studied in a randomized, controlled clinical trial192 because of the increased hepatic copper concentrations seen in patients with PSC and other chronic cholestatic conditions. In addition to its cupruretic effects, penicillamine may have antifibrogenic and immunosuppressive properties. Therapy with penicillamine therapy for three years, however, led to no difference in mortality or in biochemical or histologic progression as compared with therapy with placebo. In addition, penicillamine was associated with substantial toxicity. Other studies have failed to demonstrate a significant response to nicotine193,194 or the antifibrotic drug pirfenidone.195 Finally, combinations of various agents such as azathioprine, glucocorticoids, UDCA, and antibiotics have been studied in a limited fashion.196-198 The results of these studies have been mixed, with some showing no benefit and others demonstrating histologic improvement in small numbers of patients. A problem with combining agents is an increased risk of adverse drug reactions.

Medical Treatment of Complications

An important component in the medical care of patients with PSC is the management of complications of the disease, such as pruritus, nutritional deficiencies, and bacterial cholangitis. As discussed earlier, therapy with UDCA does not have a consistent effect on pruritus, although some patients may notice symptomatic improvement. Although treatment with antihistamines may improve pruritus, anion-exchange resins such as cholestyramine, colestipol hydrochloride, or colesevelam are typically more effective, although compliance is a problem with the use of bile acid resins. These drugs are relatively unpalatable, frequently produce con­ stipation, and may interfere with the absorption of other medications. Rifampin appears to be an effective and safe alternative for patients who do not respond to the preceding measures.199,200 Opiate antagonists such as naloxone and

naltrexone have also been shown to be effective for cholestatic pruritus, although self-limited episodes of opioid withdrawal-like symptoms may occur.126,127,201 Patients who are unresponsive to these measures and who do not obtain relief from endoscopic therapy of a dominant stricture (see later) may need to be considered for plasmapheresis (which has shown benefit in anecdotal reports) or even liver transplantation (see also Chapter 89). Patients with PSC should be screened for nutritional deficiencies by measurement of fat-soluble vitamin levels and the prothrombin time. In most patients, vitamin supplements are given orally, but a parenteral route may be necessary in patients with severe intestinal fat malabsorption. Administration of vitamin A is usually effective for correcting subclinical vitamin A deficiency. Correction of vitamin D deficiency, with or without calcium supplements, is of unproven benefit in cholestatic liver disease but is generally recommended because of its safety.202 The use of bisphosphonates and other agents to promote bone formation requires further study in patients with PSC and osteoporosis. In patients with PSC, prolongation of the prothrombin time is more likely to be the result of advanced liver disease than of vitamin K deficiency, although a trial of oral vitamin K is warranted in patients with coagulopathy (see Chapters 89 and 92). Bacterial cholangitis may develop spontaneously or after manipulation of the biliary tree. In some patients, recurring bouts of bacterial cholangitis can be debilitating. Antibiotic prophylaxis is indicated in any patient with known or suspected PSC who undergoes manipulation of the biliary tree via ERCP, percutaneous THC, or surgery. Typically, a fiveto seven-day course of a broad-spectrum antibiotic such as a fluoroquinolone, cephalosporin, or beta-lactamase inhibitor is prescribed following biliary manipulation. For patients with recurring cholangitis, long-term antibiotic prophylaxis may be helpful. The standard approach involves rotating antibiotics (e.g., amoxicillin-clavulanate, ciprofloxacin, and trimethoprim-sulfamethoxazole), given in three- to fourweek cycles, in an attempt to reduce the risk of resistance. In patients with severe, recurrent bacterial cholangitis that does not respond to this approach, liver transplantation may be the only beneficial option.

Endoscopic Management

Endoscopic therapy for PSC carries the potential to relieve jaundice, pruritus, and abdominal pain; improve bio­ chemical cholestasis; decrease the frequency of episodes of bacterial cholangitis; and improve biliary flow. In theory, improved long-term biliary patency could slow the progression of the disease and prevent or delay biliary cirrhosis. Studies of endoscopic intervention in patients with PSC, however, have typically been small, retrospective series and uncontrolled trials, and firm conclusions are not available. Patients most likely to benefit from endoscopic intervention are those with one or more dominant stricture. These patients are more likely to present with specific symptoms such as worsening jaundice or pruritus, cholangitis, or abdominal pain. Multiple studies have reported significant improvements in clinical, biochemical, and cholangiographic endpoints in patients with a dominant stricture treated with endoscopic therapy,203-207 usually dilation with a balloon or graduated dilators, with or without temporary placement of a biliary stent. Sphincterotomy is often performed for improved access and to treat choledocholithiasis, if present. One retrospective study suggested that balloon dilation followed by stent placement offered no improvement and increased the risk of complications

1165

1166

Section VIII  Biliary Tract compared with balloon dilation alone.208 Because the study was not randomized, however, these findings may have been attributable to differences between the treatment groups. Three studies have suggested that progression of the underlying disease process may be slowed by endoscopic therapy of a dominant stricture. Baluyut and colleagues209 performed graduated and balloon dilation, with or without stent placement, in 63 patients with PSC with a median follow-up of 34 months, and demonstrated an observed 5-year survival that was significantly better than survival predicted from the revised Mayo model score (see Table 68-3). Stiehl and colleagues210 performed endoscopic balloon dilation and occasional stent placement in 52 patients with PSC in whom a dominant stricture developed while the patients were on therapy with UDCA. Actuarial survival free of liver transplantation at three, five, and seven years was significantly better than that predicted from the multicenter model score (see Table 68-3). Finally, a retrospective chart review by Gluck and colleagues8 reported a single-center 20-year experience with endoscopic therapy for PSC. Endoscopic therapy for a dominant stricture was performed in 84 of 106 patients who underwent a total of 317 procedures during the 20-year observation period. The patients in whom endoscopic therapy was performed had a significantly higher survival rate than that predicted by the revised Mayo model score at years three and four. These studies were not randomized trials, and in some cases were retrospective, but they provide some supporting evidence to suggest that endoscopic management of a dominant stricture may alter the course of PSC. Endoscopic therapy in PSC also has important limitations. Complications of ERCP, such as pancreatitis, cholangitis, worsening cholestasis, and perforation, occur at an overall rate of 7.3% to 10%.8,210,211 Patients with diffuse intrahepatic biliary stricturing and no dominant stricture are less likely to derive benefit from endoscopic intervention and may be at higher risk for post-ERCP cholangitis.208 If ERCP is performed in expert hands and only for specific indications such as worsening of jaundice, pruritus, or cholangitis—that is, in the subgroup of patients who are most likely to benefit from therapy—the risks in patients unlikely to benefit will be minimized (see Chapter 70).

Percutaneous Management

Percutaneous THC with balloon dilation, stenting, or both, can also be undertaken to treat biliary strictures in patients with PSC. This approach is typically recommended only when endoscopic intervention is contraindicated or unsuccessful because of the added risks of bleeding and bile peritonitis, as well as increased patient discomfort, associated with percutaneous intervention (see Chapter 70).

Surgical Management

Biliary Surgery The role of biliary surgery in PSC has diminished considerably with improvements in endoscopic therapy and liver transplantation. Resections of a dominant stricture of the bile duct or near the hepatic bifurcation followed by hepaticojejunostomy or choledochojejunostomy have been the most commonly performed operations.212,213 Postoperative mortality is increased significantly in patients with PSC and cirrhosis.214 In addition, biliary surgery may complicate future liver transplantation. Currently, biliary surgery in patients with PSC is rarely indicated and should be reserved for the small subset of patients who have early-stage PSC and biliary strictures that are not amenable to endoscopic or percutaneous intervention.

Liver Transplantation Liver transplantation (see also Chapter 95) is the only therapy that has been shown conclusively to improve the natural history of PSC. In addition, quality of life improves after liver transplantation.215,216 The procedure is recommended for patients with PSC in whom decompensated cirrhosis and complications of portal hypertension develop. Recurrent cholangitis or pruritus that is refractory to medical and endoscopic management rarely may also be indications for liver transplantation. Determination of the appropriate timing for liver transplantation in patients with PSC may be challenging, although use of available natural history models can be helpful. When a patient’s expected survival after liver transplantation exceeds survival predicted from the natural history models, liver transplantation should be undertaken, in the absence of contraindications. Intraoperatively, patients who undergo liver transplantation for PSC should have a Roux-en-Y choledochojejunostomy anastomosis, instead of a standard choledochocholedochostomy. This approach is recommended to allow removal of as much of the native biliary tree as possible, to reduce the risk of recurrent strictures and cholangiocarcinoma.217 Patient and graft survival after liver transplantation for PSC is excellent. A large single-center experience demonstrated 1-, 5-, and 10-year actuarial patient survival rates of 93.7%, 86.4%, and 69.8%, respectively. Corresponding graft survival rates were 83.4%, 79.0%, and 60.5%.218 Similar results have been reported in other series.219,220 Overall, survival rates after liver transplantation for PSC are significantly better than those for any other disease except PBC.221 Recipient factors that have been associated with a worse prognosis after liver transplantation for PSC are older age, decreased serum albumin level, renal failure, Child’s class C cirrhosis, and advanced United Network for Organ Sharing status.221,222 The presence of cholangiocarcinoma has a major impact on the outcome after liver transplantation for PSC. Early studies demonstrated that even in cases in which cholangiocarcinoma was discovered incidentally in the explant, recipient survival was poor, with a one-year survival rate of 30% in one series.223 On the basis of such studies, cholangiocarcinoma has generally been considered a contraindication to liver transplantation. Another report confirmed the poor post-liver transplantation outcome in patients with known cholangiocarcinoma but suggested a good survival rate for those who had a small cholangiocarcinoma found incidentally at the time of transplantation.220 Subsequent studies have demonstrated improved results of liver transplantation in patients with cholangiocarcinoma, with oneand five-year survival rates of 65% to 82% and 35% to 42%, respectively.224,225 Preoperative chemoradiation in highly selected patients with cholangiocarcinoma has shown promise in reducing the rate of tumor recurrence after liver transplantation (see Chapter 69).226,227 Biliary strictures commonly recur after liver transplantation for PSC and may represent recurrent PSC. In addition to recurrent PSC, potential causes of biliary strictures after liver transplantation include ABO blood group incompatibility, hepatic artery occlusion, chronic ductopenic graft rejection, Roux-en-Y–related cholangitis, and preservationrelated ischemia. The diagnosis of recurrent PSC has been proposed to be confined to those patients who have a consistent cholangiographic pattern and compatible liver histology showing fibrous cholangitis, fibro-obliterative lesions, biliary fibrosis, or biliary cirrhosis, and who lack other risk factors for biliary strictures, such as hepatic artery occlusion, ABO incompatibility, or ductopenic graft rejection, or who develop non-anastomotic strictures within 90 days of

Chapter 68  Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis transplantation.228 With these stringent criteria, the risk of recurrent PSC after liver transplantation ranges from 5.7% to 21.1%,219,220,229-230 and patient and graft survival do not appear to be adversely affected for the first five years of follow-up.220,230 No specific risk factors for the development of recurrent PSC have been clearly identified. In addition, no specific therapy has been shown to treat or prevent recurrent PSC effectively after liver transplantation. The effect of liver transplantation on the course of underlying IBD and risk for subsequent colonic neoplasia remains controversial. The clinical course of IBD after liver transplantation has varied greatly among studies.45,219,232,233 Immunosuppression was thought possibly to ameliorate the clinical course of IBD but does not in most instances. A large single-center study showed that liver transplantation was associated with an increased risk for colectomy.234 The risk of colorectal neoplasia is generally agreed to be increased after liver transplantation in patients with UC,164-166,219 although this finding has not been confirmed in all studies.234 Annual surveillance colonoscopy is recommended after liver transplantation for PSC in patients with UC.

RECURRENT PYOGENIC CHOLANGITIS Recurrent pyogenic cholangitis (RPC) was originally defined by Cook and colleagues as a syndrome characterized by recurrent bacterial cholangitis, intrahepatic pigment stones, and biliary strictures, possibly leading to chronic liver disease and cho­langiocarcinoma.235 RPC has also been called oriental cholangiohepatitis, “Hong Kong disease,” “biliary obstruction syndrome of the Chinese,” and hepatolithiasis.236

EPIDEMIOLOGY

Digby first described RPC in 1930 in Chinese patients in Hong Kong and was the first to recognize that RPC was clinically and pathologically distinct from biliary disease caused by gallstones in Western populations.237 Subsequently, most cases have occurred in patients from East Asia; men and women are affected equally. Patients in rural areas and those of lower socioeconomic status appear to be at increased risk, and the incidence is highest among persons between ages 30 and 40.238,239 In certain parts of Southeast Asia such as Taiwan, over one half of the cases of gallstone disease are estimated to be caused by RPC.240 In Singapore, Japan, and Hong Kong, 2% to 5% of biliary calculous disease is attributed to RPC.241 The incidence of RPC in East Asia appears to be decreasing. A study from the Queen Mary Hospital, a major referral center for RPC, reported that only 6.7% of a total of 490 patients who underwent surgery for hepatobiliary disease were classified as having RPC.241 By contrast, typical gallstone disease was the reason for surgery in 44%. These numbers likely reflect referral bias; a nationwide survey found that RPC or intrahepatic stones represented only about 20% of the total cases of biliary tract disease. The incidence of intrahepatic stones appears to be decreasing in Taiwan as well; one study reported that the incidence of hepatolithiasis decreased from 21% to 18% between 1981 and 1989.240 By contrast, the incidence and prevalence of RPC are increasing in Western countries, reflecting immigration patterns. One study from San Francisco found that the prevalence of RPC doubled between 1983 and 1995.242 Similarly, a review from a county hospital in Los Angeles reported that 57% of 18 patients with RPC seen over a seven-year period were of Asian descent, and 36% of the patients were of Hispanic descent.243

The reasons for the changing epidemiology of intrahepatic stones in East Asia are unclear. Several possible explanations have been proposed, including adoption of a Western-style diet with a higher protein content; improved hygiene, which may lead to reduced gastrointestinal infections and consequently decreased entry of bacteria into the portal circulation (an important cause of cholangitis); and, less likely, reduction in disease burden related to Clonorchis sinensis and Ascaris lumbricoides. A low-protein diet may lead to reduced biliary levels of glucaro-1:4-lactone, an inhibitor or bacterial b-glucuronidase, which helps promote the formation of calcium bilirubinate stones by decon­ jugating bilirubin into unconjugated bilirubin.244 Furthermore, a diet low in fat may be associated with reduced gallbladder contractility and thus promote stasis, which also is a factor in stone formation.245 The rising incidence of cholesterol gallstones in Asia suggests that environmental factors, such as adoption of a more Western-style diet, rather than genetic factors, are major factors in the pathogenesis of RPC.241

ETIOLOGY AND PATHOGENESIS

The etiology of RPC remains unknown. The most attractive hypothesis links biliary tract infection with the parasites Clonorchis sinensis, Opisthorchis species, and Ascaris lumbricoides (see Chapters 82 and 110). Infection with these parasites is endemic in the same geographic region where RPC is prevalent. C. sinensis, a trematode (liver fluke), is endemic in China, Japan, Taiwan, Korea, and Vietnam.246,247 O. felineus and O. viverrini are the two species most commonly implicated in opisthorchiasis. O. felineus has been described in Southeast Asia and in parts of the former Soviet Union, and A. lumbricoides, a roundworm, is a ubiquitous parasite and may infect over one billion people throughout the world.248,249 Both organisms colonize the biliary tree and lead to infection, biliary tract obstruction, and secondary bacterial cholangitis. The Clonorchis worm can survive for decades in the biliary tree and may lead to inflammatory changes in the bile ducts, as well as direct bile duct obstruction by the flukes, and shows a specific predilection for the left hepatic ducts.248 Analysis of biliary stones from patients with RPC has demonstrated evidence of ova.250 Patients with RPC, however, have not been shown to have an increased prevalence of infection with these parasites when compared with the general population, and approximately one half of patients with RPC demonstrate no evidence of infection.251-253 Furthermore, some parts of Asia with a high prevalence of RPC have low or undetectable rates of infection with C. sinensis.254 Nevertheless, it remains possible that parasitic infection may account for a significant proportion of cases of RPC. Moreover, currently available serologic tests lack precision to detect prior infection, and structural or functional changes in the bile ducts, gallbladder, or sphincter of Oddi caused by infection in the remote past may alter the biliary epithelia and thereby promote the later formation of intrahepatic stones. Bacterial infections have also been proposed as a cause of RPC because of the high incidence of bacterial cholangitis in patients with RPC. Portal bacteremia, possibly related to gastrointestinal infection and bacterial translocation, has been associated with low socioeconomic status and malnutrition.239 The hypothesis that bacterial infections lead to RPC, however, does not explain the unique epidemiologic features of the disease. Dietary factors and hygiene have also been implicated in the pathogenesis of RPC. Because cholecystokinin, a potent mediator of gallbladder contractility, is secreted in response

1167

1168

Section VIII  Biliary Tract to dietary fat, diets high in carbohydrate and low in saturated fat may be associated with reduced gallbladder contractility and promote stone formation. Furthermore, deconjugation of bilirubin by bacteria or endogenous enzymes, as noted earlier, may be related to dietary and environmental factors and may facilitate formation of pigment stones. Several investigators have suggested that patients with RPC may lack an inhibitor of bacterial b-glucuronidase.255,256 Another factor that may contribute to the pathogenesis of RPC is abnormal motor activity of the sphincter of Oddi. A study of biliary sphincter manometry in 15 patients with RPC and 15 normal control subjects257 demonstrated various abnormalities, including retrograde phasic waves, abnormal propagation of phasic contractions, elevated basal sphincter pressure, and increased frequency of phasic contractions in the sphincter among the patients with RPC (see Chapter 63). Almost one half of the patients had evidence of papillitis at endoscopy. The investigators suggested that papillitis may lead to altered function in the sphincter of Oddi, resulting in turn in delayed biliary drainage and recurrent cholangitis. Whether papillitis and sphincter of Oddi dysfunction are a cause or effect of RPC is difficult to ascertain from the study; recurrent bouts of cholangitis may, in fact, have led to changes in the biliary tract, and recurrent passage of small pigment stones and debris may have resulted in endoscopic and manometric changes.

Figure 68-5.  Computed tomographic scan with intravenous contrast in a patient with recurrent pyogenic cholangitis. The scan demonstrates marked dilatation of the biliary tree, most prominent in the left lobe of the liver. High-density foci in the dilated segments represent calcified stones.

CLINICAL FEATURES

Patients with RPC often present with symptoms of ascending cholangitis (45% of patients).258 Patients characteri­ stically present with fever, right upper quadrant pain, and jaundice, also referred to as Charcot’s triad. A prior history of such attacks is elicited in the majority of patients, whereas up to 30% of patients present with an initial episode.242,250 Patients may also present with abdominal pain or pancreatitis. On physical examination, abdominal tenderness is a common finding. Hepatomegaly is present in approximately 20% of patients. A palpable gallbladder may be present in approximately 10% of patients and may point to emphysematous cholecystitis.250,251,253 Laboratory findings are compatible with biliary obstruction, with elevation of serum total and direct bilirubin, aminotransferase, and alkaline phosphatase levels; leukocytosis may be present. Imaging findings in patients with RPC are characteristic (Figs. 68-5 and 68-6). The majority of patients (75% to 80%) have intrahepatic stones, with predominant involvement of the left hepatic duct. Up to 70% may also have associated stones in the gallbladder, and cholecystitis (calculous or acalculous) is a frequent associated finding.259-261 Dilatation of the bile ducts is found almost universally. The central bile ducts are dilated disproportionately, with abrupt tapering and attenuation of more peripheral bile ducts within the liver. The presence of bile duct calculi is usually associated with intrahepatic bile duct dilatation and downstream strictures. The left hepatic ducts are involved more often than are the right.250 Cross-sectional imaging studies of the liver, such as ultrasonography and CT, as well as direct cholangiography, are useful for evaluating a patient with known or suspected RPC. Ultrasonography is a reasonable screening test early in the evaluation of the patient and may identify intra- and extrahepatic bile duct dilatation in the majority of cases. CT is helpful in delineating the complete segmental anatomy of the liver and is invaluable in planning surgical therapy (see Figure 68-5). CT may also demonstrate associated atrophic segments of the liver, abscesses, and cholangiocarci-

Figure 68-6.  Endoscopic retrograde cholangiopancreatography in a patient with recurrent pyogenic cholangitis. Multiple filling defects are seen in the bile duct, consistent with stones and/or air bubbles. The poorly opacified left hepatic duct is strictured at its takeoff, markedly dilated, and filled with intrahepatic stones. The right intrahepatic ducts demonstrate characteristic reduction in arborization, widening of the branching angles, and abrupt peripheral tapering.

noma.250 A study of more than 1300 patients with RPC from Korea described characteristic CT findings in 82 patients in whom cholangiocarcinoma was associated with RPC; cholangiocarcinoma tended to be located in atrophic segments associated with biliary calculi and was often accompanied by portal vein occlusion or narrowing.262

Chapter 68  Sclerosing Cholangitis and Recurrent Pyogenic Cholangitis Direct cholangiography, whether performed by the percutaneous or endoscopic route, allows localization of intrahepatic stones and strictures and placement of drains or extraction of stones (see Figure 68-6). ERCP is the first choice for therapeutic intervention because it is less invasive than percutaneous THC.263 Removal of intrahepatic stones by this technique may be exceedingly difficult, however, because of the presence of a tight stricture and the sharp angulation of the intrahepatic bile ducts. MRCP permits selective visualization of the biliary tree and has become the preferred test for diagnostic evaluation of the bile ducts. In a study from Korea,264 18 of 24 patients who had MRCP before undergoing surgical treatment for RPC also underwent direct cholangiography. Two examiners evaluated MRCP images and direct cholangiograms and compared them with surgical findings. All dilated bile ducts and 98% of focal duct strictures and intraductal stones were identified by MRCP. By contrast, direct cholangiography identified only 44% to 47% of segmental bile duct abnormalities. The authors concluded that MRCP was more sensitive than direct cholangiography for complete evaluation of the biliary tree. MRCP is an important complementary technique for examination of the biliary tree, especially in patients being considered for surgery because of its ability to visualize the bile ducts upstream from obstructed areas completely and to correlate the observed abnormalities with the segmental anatomy of the liver.265 Patients who present with an initial episode of cholangitis associated with intrahepatic stones and strictures should undergo evaluation for infection with Clonorchis and Opisthorchis species, particularly if the patient comes from or has traveled to an endemic area. The diagnosis of a parasitic infection is made by identification of eggs in fecal specimens; concentrated stool may be required. Eggs are present in stool after four weeks of infection.250 Duo­ denal or biliary fluid also may demonstrate eggs or intact worms. Peripheral eosinophilia may be present in cases of parasitic infection and may be associated with elevated serum IgE levels.246

PATHOLOGY

The characteristic findings in RPC include strictures and dilatation of the intra- and extrahepatic bile ducts. In classic cases, the left hepatic duct is more commonly and more severely affected than the right.250,251 With chronic disease, atrophy of the left lobe or the left lateral segment of the liver may occur and may be the site at which cholangiocarcinoma develops.250,251 The bile ducts are often obstructed with pigment stones, in addition to sludge and inspissated bile, and are frequently contaminated by bacteria and purulent material. The stones are composed predominantly of calcium bilirubinate, although cholesterol has been reported to be found in increasing proportions.239 Ova of parasites with a predilection for the biliary tree, such as C. senensis, have been described within the stones.250,252 An inflammatory infiltrate is frequently present in the walls of involved bile ducts and may be associated with periductal fibrosis and abscesses.250,251

TREATMENT

Antibiotic therapy should be initiated promptly, after cultures of blood and bile (if accessible) have been obtained. Abdominal ultrasonography is a reasonable initial imaging modality.259 In patients with evidence of cholangitis and dilated intra- and extrahepatic bile ducts, ERCP is the preferred interventional procedure. ERCP with sphincterotomy, with or without placement of a nasobiliary drain or an endobiliary drain, may adequately remove bile

duct stones and traverse strictures. After adequate biliary drainage has been achieved, additional cross-sectional imaging with CT or MRI, with or without MRCP, may be considered. Several studies have reviewed the success of various nonoperative interventions for initial management of patients with RPC presenting acutely. Sperling and colleagues258 compared outcomes in 41 patients with RPC based on whether they received immediate therapeutic ERCP, hepatobiliary surgery, or no intervention. Symptoms recurred in 62% of patients who underwent only diagnostic ERCP but one half as often in those treated with therapeutic ERCP or surgery. Therapeutic ERCP was particularly effective in patients with disease involving the extrahepatic bile ducts and was comparable in efficacy to surgery. Patients with disease involving both the right and left hepatic bile ducts tend to undergo more imaging studies, percutaneous cholangiograms, and endoscopic or surgical procedures.242 Hepaticojejunostomy has been a commonly used surgical procedure for the treatment of intrahepatic stones in patients with RPC. Kusano and colleagues,266 from Japan, reviewed the long-term outcomes of hepaticojejunostomy for intrahepatic hepatolithiasis in 159 patients over a 23-year period. Surgical approaches included hepatectomy (n = 94), biliary lithotripsy (n = 65), or hepaticojejunostomy (n = 72) (or combinations). Residual or recurrent stones were identified in approximately one third of patients after presumably complete removal. The rate of cholangitis was higher among patients who underwent a biliary-enteric anastomosis (22 of 72; 31%) than in those who did not (3 of 87; 3.4%). This study suggests that surgical therapy involving a biliaryenteric anastomosis may be associated with a higher rate of cholangitis than nonsurgical therapy. Laparoscopic biliary bypass surgery has been proposed as a technically feasible and effective option for patients with RPC.267 A large retrospective case series from Canada of 42 patients who underwent surgical intervention for RPC demonstrated excellent postoperative outcomes for both liver resection and bile duct exploration with biliary bypass; the overall operative mortality rate was 0%, and the complication rate was 35% for resection and 30% for bile duct exploration.268 Treatment with an antihelminthic agent is indicated in patients with evidence of active parasitic infection. Prazi­ quantel, 75 mg/kg in three divided doses for one day, is the treatment of choice (see also Chapter 82). The treatment is almost universally effective in both Clonorchis and Opisthorchis infections.269 Side effects of the medication include headache and gastrointestinal symptoms. Even patients without gastrointestinal symptoms but with Clonorchis or Opisthorchis in the stool should be treated to reduce the risk of cholangiocarcinoma.132

PROGNOSIS AND COMPLICATIONS

The natural history of RPC has been evaluated in a large series from Korea. The records of 193 patients with newly diagnosed RPC were reviewed to examine the rates and risk factors for recurrence.270 The mean follow-up period was 56 months (range, 1 to 242 months). Cumulative recurrence rates of cholangitis were 25% at three years and 37% at five years, with an overall rate of 45% during the follow-up period. Factors associated with cholangitis included recurrent stones (hazard ratio = 4.02; 95% CI: 1.3 to 12.4), residual stones (hazard ratio = 1.77; 95% CI: 1.1 to 3.0); prior hepatic resection was associated with a lower rate of recurrent cholangitis (hazard ratio = 0.28; 95% CI: 0.1 to 0.7). Strictures in the extrahepatic bile ducts were associated

1169

1170

Section VIII  Biliary Tract with a higher rate of cholangitis, whereas strictures in the intrahepatic ducts and disruption of the sphincter of Oddi (e.g., sphincterotomy) were not. Cholangiocarcinoma is another long-term complication in patients with RPC (see Chapter 69). The frequency of cholangiocarcinoma has not been clearly delineated. One large study found that the frequency of cholangiocarcinoma among patients with RPC was approximately 3%.271 Other studies have reported the rate to be as high as 9%.272,273 Secondary biliary cirrhosis may develop and require liver transplantation.

KEY REFERENCES

Angulo P, Pearce DH, Johnson CD, et al. Magnetic resonance cholangiography in patients with biliary disease: Its role in primary sclerosing cholangitis. J Hepatol 2000; 33:520-7. (Ref 10.) Bambha K, Kim WR, Talwalkar J, et al. Incidence, clinical spectrum, and outcomes of primary sclerosing cholangitis in a United States community. Gastroenterology 2003; 125:1364-9. (Ref 27.) Bjornsson E. Immunoglobulin G4-associated cholangitis. Curr Opin Gastroenterol 2008; 24:389-94. (Ref 25.) Bjornsson E, Chari ST, Smyrk TC, Lindor K. Immunoglobulin G4 associated cholangitis: Description of an emerging clinical entity based on review of the literature. Hepatology 2007; 45:1547-54. (Ref 23.) Feldstein AE, Perrault J, El-Youssif M, et al. Primary sclerosing cholangitis in children: A long-term follow-up study. Hepatology 2003; 38:210-17. (Ref 33.) Gluck M, Cantone N, Brandabur J, et al. A twenty-year experience with endoscopic therapy for symptomatic primary sclerosing cholangitis. J Clin Gastroenterol 2008; 42:1032-9. (Ref 8.) Gregorio GV, Portmann B, Karani J, et al. Autoimmune hepatitis/ sclerosing cholangitis overlap syndrome in childhood: A 16-year prospective study. Hepatology 2001; 33:544-53. (Ref 66.)

Kamisawa T, Okamoto A. IgG4-related sclerosing disease. World J Gastroenterol 2008; 14:3948-55. (Ref 24.) Kaya M, Angulo P, Lindor KD. Overlap of autoimmune hepatitis and primary sclerosing cholangitis: An evaluation of a modified scoring system. J Hepatol 2000; 33:537-42. (Ref 20.) Mitchell SA, Grove J, Spurkland A, et al. Association of the tumour necrosis factor alpha -308 but not the interleukin 10 -627 promoter polymorphism with genetic susceptibility to primary sclerosing cholangitis. Gut 2001; 49:288-94. (Ref 59.) Mitchell SA, Bansi DS, Hunt N, et al. A preliminary trial of high-dose ursodeoxycholic acid in primary sclerosing cholangitis. Gastroenterology 2001; 121:900-7. (Ref 177.) Moff SL, Kamel IR, Eustace J, et al. Diagnosis of primary sclerosing cholangitis: A blinded comparative study using magnetic resonance cholangiography and endoscopic retrograde cholangiography. Gastrointest Endosc 2006; 64:219-23. (Ref 13.) Olsson R, Boberg KM, de Muckadell OS, et al. High dose ursodeoxycholic acid for the treatment of primary sclerosing cholangitis: A 5-year multicenter, randomized, controlled study. Gastroenterology 2005; 129:1464-72. (Ref 181.) Said K, Glaumann H, Bergquist A. Gallbladder disease in patients with primary sclerosing cholangitis. J Hepatol 2008; 48:598-605. (Ref 7.) Stiehl A, Rudolph G, Kloters-Plachky P, et al. Development of dominant bile duct stenoses in patients with primary sclerosing cholangitis treated with ursodeoxycholic acid: Outcome after endoscopic treatment. J Hepatol 2002; 36: 151-6. (Ref 9.) Textor HJ, Flacke S, Pauleit D, et al. Three-dimensional magnetic resonance cholangiopancreatography with respiratory triggering in the diagnosis of primary sclerosing cholangitis: Comparison with endoscopic retrograde cholangiography. Endoscopy 2002; 34:984-90. (Ref 11.) Tung BY, Emond MJ, Haggitt RC, et al. Ursodiol use is associated with lower prevalence of colonic neoplasia in patients with ulcerative colitis and primary sclerosing cholangitis. Ann Intern Med 2001; 134:89-95. (Ref 159.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

69 Tumors of the Bile Ducts, Gallbladder, and Ampulla Boris Blechacz and Gregory J. Gores

CHAPTER OUTLINE Cholangiocarcinoma  1171 Epidemiology  1171 Etiology  1172 Pathology  1172 Pathogenesis  1172 Clinical Features and Diagnosis  1173 Staging  1175 Treatment  1176 Gallbladder Carcinoma  1177 Epidemiology  1177 Etiology  1177 Pathology  1178 Pathogenesis  1178

Biliary malignancies comprise the vast majority of biliary neoplasms and are divided into the following three categories: (1) carcinomas of the intra- and extrahepatic bile ducts; (2) carcinoma of the gallbladder; and (3) carcinoma of the ampulla of Vater.1 In the United States and other Western nations, biliary malignancies are rare. In certain parts of the world, however, their prevalence rates are high, making them leading causes of cancer death. Biliary cancers are highly aggressive with dismal prognoses. Often these cancers are diagnosed at an advanced stage. In general, they are resistant to chemotherapy. Advances in the understanding of the molecular pathogenesis of these tumors has allowed the development of new experimental models and targeted therapies, and advances in diagnosis and surgical treatment have resulted in improved outcomes for subsets of patients.

CHOLANGIOCARCINOMA Cholangiocarcinoma is an epithelial carcinoma with differentiated features of biliary epithelium that arises from the intra- and extrahepatic biliary tree.2 It is the most common bile duct tumor and second most common primary hepatic malignancy (after hepatocellular carcinoma, see Chapter 94). Since the 1970s, the incidence has increased substantially in Western societies.3 The classification of cholangiocarcinomas into intra- and extrahepatic cancers is based on differences in anatomy, etiology, pathogenesis, molecular signature, and treatment. The second-order bile ducts are the anatomic margin for the distinction between these two subsets (see Chapter 62). Extrahepatic cholangiocarcinomas account for 80% to 90%

Clinical Features and Diagnosis  1179 Staging  1180 Treatment  1180 Ampullary Carcinoma  1181 Epidemiology  1181 Etiology  1181 Pathology  1181 Pathogenesis  1182 Clinical Features and Diagnosis  1182 Staging  1182 Treatment  1182 Other Tumors of the Bile Ducts and Gallbladder  1183

of these cancers and can be further subclassified into hilar or distal bile duct cancers. Hilar cholangiocarcinoma, also referred to as Klatskin tumors, are described clinically according to the Bismuth-Corlette classification as types I to IV (Fig. 69-1). Type I cholangiocarcinomas involve the common hepatic duct distal to the union of the right and left hepatic ducts; type II tumors involve the union of the right and left hepatic ducts; type IIIa tumors involve the union of the right and left hepatic ducts and extend up the right hepatic duct; type IIIb tumors involve the union of the right and left hepatic ducts and extend up the left hepatic duct; and type IV tumors are multifocal or involve the biliary confluence and extend up the right and left hepatic ducts. The natural course of cholangiocarcinoma is aggressive, with a median survival of less than 24 months following diagnosis.4 The only potentially curative treatment is surgical. Unfortunately, the majority of patients are diagnosed at an advanced stage that precludes curative surgery.

EPIDEMIOLOGY

Cholangiocarcinoma is an uncommon cancer that accounts for less than 2% of all malignancies. It is the ninth most common gastroenterologic malignancy and the second most common primary hepatic and biliary malignancy. Hepatobiliary malignancies account for 13% and 3% of overall cancer-related mortality in the world and in the United States, respectively; 10% to 20% of these deaths are caused by cholangiocarcinoma. Global incidence rates for cholangiocarcinoma are heterogeneous. The highest incidence is observed in Southeast Asia, with rates up to 96 per 100,000 population, and the lowest incidence is observed in Australia, with rates as low as 0.1 per 100,000 population.5 In the United States,

1171

Section VIII  Biliary Tract Type I

Type II

Type IIIa

Type IIIb

Intrahepatic

Liver Common hepatic duct Gallbladder

Bile duct

Hilar Distal

Ampulla of Vater

Extrahepatic

1172

Type IV

Duodenum

A

B

Figure 69-1.  Classification of cholangiocarcinoma. A, Anatomic classification of intra- and extrahepatic cholangiocarcinoma. Extrahepatic cholangiocarcinoma is further subclassified as hilar and distal. B, Bismuth-Corlette classification of hilar cholangiocarcinoma as types I to IV. Tumor is depicted in yellow and normal bile ducts in green. (From Blechacz BR, Gores GJ. Cholangiocarcinoma. Clin Liver Dis 2008; 12:131-150.)

incidence rates are 0.95 per 100,000 population for intra­ hepatic cholangiocarcinoma and 0.82 per 100,000 for extrahepatic cholangiocarcinoma. Ethnic differences in the incidence of cholangiocarcinoma have been observed in the United States, with the highest rates observed in patients of Hispanic descent and the lowest in African Americans. Since the 1980s, age-adjusted incidence rates for intrahepatic cholangiocarcinoma have increased, whereas those for extrahepatic cholangiocarcinoma have remained stable; the cause of the overall increase in incidence is unknown.3 Fifty-two percent to 54% of patients with cholangiocarcinoma are male. Globally, the average age at diagnosis is older than 50 years. In Western industrialized nations, most cases are diagnosed around age 65, and cholangiocarcinoma is uncommon before age 40 except in patients with primary sclerosing cholangitis (see Chapter 68).

ETIOLOGY

In the majority of cases, the etiology of cholangiocarcinoma is unknown. Several risk factors have been identified (Table 69-1). These risk factors are characterized by their association with inflammation and cholestasis. Primary sclerosing cholangitis (PSC) is one of the most common risk factors (see Chapter 68). In patients with PSC, the annual frequency rate of cholangiocarcinoma is 0.6% to 1.5%, and the overall prevalence rate of cholangiocarcinoma is 5% to 15%. In the majority of patients with PSC in whom cholangiocarcinoma develops, the diagnosis of cholangiocarcinoma is made within two-and-a-half years of the diagnosis of PSC.6 Other risk factors for cholangiocarcinoma include biliary infections with Opisthorchis viverrini and Clonorchis sinensis, which are endemic in East Asia (see Chapter 82).7 Biliary malformations such as Caroli’s disease and choledochal cysts are associated with a 10% to 15% risk for the development of cholangiocarcinoma (see Chapter 62),8 whereas hepatolithiasis (“recurrent pyogenic cholangitis”) carries a 10% risk for the development of cholangiocarcinoma (see Chapter 68).9 Recurrent bacterial cholangitis in the setting of biliary-enteric drainage procedures has also been associated with the development of cholangiocarcinoma.10 Carcinogens such as thorotrast (used as a radiologic contrast agent in the past) and dioxins have been associated with an

Table 69-1  Risk Factors for Cholangiocarcinoma Biliary-enteric drainage procedures Caroli’s disease Choledochal cyst Cirrhosis Clonorchis sinensis infection Hepatitis C Hepatolithiasis Opisthorchis viverrini infection Primary sclerosing cholangitis Thorotrast Toxins (dioxins, polyvinyl chloride)

increased risk of cholangiocarcinoma.11 Hepatitis C and cirrhosis are considered possible risk factors for cholangiocarcinoma (see Chapter 79).5

PATHOLOGY

Cholangiocarcinoma is a paucicellular, highly desmoplastic tumor. Macroscopically, it can be described according to its growth characteristics as mass forming, periductalinfiltrating, or intraductal-papillary. Intrahepatic cholangiocarcinomas are typically mass forming, whereas ductal carcinomas can present in any of the three growth forms. Histologically, 90% of cholangiocarcinomas are adenocarcinomas. Other histologic types include intestinal-type adenocarcinoma, clear cell adenocarcinoma, signet-ring cell carcinoma, adenosquamous carcinoma, squamous cell carcinoma, and small cell carcinoma.12 Intraductal-papillary adenocarcinomas spread superficially along the biliary mucosa without deep invasion of the fibromuscular wall layers and have a better prognosis than nonpapillary cancers.13 Metastases to regional and peripancreatic lymph nodes are frequently observed with this growth type.

PATHOGENESIS

As noted earlier, etiologic associations as well as experimental data provide evidence for inflammation and bile acids as key factors in the molecular pathogenesis of cholangiocarcinoma. Inflammation provides a microenvi-

Chapter 69  Tumors of the Bile Ducts, Gallbladder, and Ampulla Table 69-2  Molecular Changes in Cholangiocarcinoma

Table 69-3  Diagnostic Criteria for Cholangiocarcinoma

MALIGNANT PHENOTYPE

DYSREGULATED GENE products AND PATHWAYS

Proliferation

Interleukin (IL)-6, IL-6 receptor (gp130) Hepatocyte growth factor (HGF)/c-Met ErbB2 K-ras BRAF Cyclooxygenase-2 (COX-2) Caspase-9 Mcl-1 bcl-2 Bcl-XL COX-2 Telomerase Cyclin D1 p21waf1/cip1 p27kip1 pRb p53 mdm-2 DCP4/Smad4 p16INK4a E-cadherin α/β-catenin Aspartyl (asparaginyl) β-hydroxylase WISP1v Matrix metalloproteinases Vascular endothelial growth factor (VEGF) Transforming growth factor-β (TGF-β)

Malignant appearing stricture AND persistent serum CA 19-9 level >129 U/mL in the absence of bacterial cholangitis* Mass lesion on cross-sectional imaging Positive conventional cytology result Positive (transluminal) biopsy specimen Stricture plus polysomy on fluorescence in situ hybridization (FISH)

Apoptosis evasion

Senescence evasion Cell cycle dysregulation

Invasion/Metastases

Angiogenesis

ronment that promotes malignant transformation of bile duct–associated cells such as cholangiocytes, biliary stem cells, or epithelial cells within peribiliary glands (see Chapter 62). In subsets of cholangiocarcinoma and mixed hepatocellular-cholangiocarcinoma, hepatic stem/progenitor cells have been suggested to be the cells of origin.14 On a molecular level, increased tissue concentrations of cytokines, bile acids, and growth factors can lead to functional inactivation of mismatch DNA repair genes and tumor suppressor genes and promote the expression of protooncogenes. Cytokines stimulate expression of inducible nitric oxide synthase (iNOS) in epithelial cells, thereby resulting in increased intracellular levels of nitric oxide (NO) and reactive nitrogen oxide species (RNOS). NO and RNOS interact with cellular DNA and proteins and thus result in DNA mutations and strand breaks, with inactivation of DNA repair proteins.15 A variety of different oncogenic mutations have been described in cholangiocarcinoma (Table 69-2). Their frequencies differ depending on the anatomic location, stage, type, and etiology of the tumor and the ethnicity of the patient. Growth factors, dysregulated signaling pathways, and tyrosine kinases promote further tumor cell proliferation, as well as inhibition of senescence and apoptosis. Key signaling pathways in cholangiocarcinoma involve interleukin (IL)-6, cyclooxygenase-2 (COX-2), epidermal growth factor receptor (EGFR), ErbB2 (also known as human epidermal growth factor receptor 2 [HER 2]), hepatocyte growth factor (HGF), and vascular endothelial cell growth factor receptor (VEGFR). The IL-6 signaling axis in particular has been shown to be critical to the pathogenesis of cholangiocarcinoma.16,17 Increased IL-6 signaling is sustained in cholangiocarcinoma cells by the following mechanisms: (1) autocrine secretion of high levels of IL-6; (2) COX-2-mediated up-

*Serum CA 19-9 >100 U/mL in patients with primary sclerosing cholangitis.

regulation of the IL-6 receptor subunit gp130; and (3) inactivation of the negative feedback loop through epigenetic silencing of suppressor of cytokine signaling 3 (SOCS3). Increased IL-6 signaling constitutively activates the PI3K pathway, Janus Kinase (JAK) signal transducers and activators of transcription (STAT) pathway, and p38- and p42/44mitogen-activated protein kinase (MAPK) pathways, which in turn mediate cell survival, cell proliferation, and evasion of cell senescence. The EGFR pathway is also constitutively activated, resulting in further activation of p38- and p42/44MAPK.18 The EGFR homolog ErbB2 is overexpressed in many cholangiocarcinomas, resulting in downstream induction of Raf/MAPK pathways. In addition to its oncogenic effects, MAPK induces overexpression of COX-2, which is further stimulated by increased concentrations of bile acids, oxysterols, and iNOS.19,20 COX-2 also mediates cell survival and proliferation. Sustained cell proliferation is further mediated by increased HGF secretion and overexpression of its receptor, c-Met.21 The multiple pathways form a complex interacting network.

CLINICAL FEATURES AND DIAGNOSIS

The diagnosis of cholangiocarcinoma is challenging because of the paucicellular character of the malignancy and requires a multidisciplinary approach that includes clinical evaluation and laboratory, endoscopic, and radiologic studies (Fig. 69-2 and Table 69-3). Intrahepatic cholangiocarcinoma manifests predominantly with abdominal pain and systemic symptoms such as cachexia, malaise, and fatigue. Extrahepatic cholangiocarcinoma manifests in most cases with painless jaundice secondary to malignant biliary obstruction. In 10% of patients, bacterial cholangitis is the initial presenting symptom. An “atrophy-hypertrophy” complex can be documented by physical examination as palpable hypertrophy of the contralateral, unaffected lobe of the liver, with atrophy of the affected lobe as a result of vascular encasement and bile ductal obstruction. Laboratory analysis may reveal evidence of obstructive cholestasis (e.g., elevation of serum alkaline phosphatase and bilirubin levels). Serum levels of several serum tumor markers (CA 19-9, carcinoembryonic antigen [CEA], and Ca-125) may be elevated in patients with cholangiocarcinoma; however, none of these serum markers is specific, and they each can be elevated in other gastroenterologic or gynecologic malignancies and in the setting of biliary inflammation or infection.22 The most commonly used marker is CA 19-9. In patients with PSC, the sensitivity and specificity of CA 19-9 for the diagnosis of cholangiocarcinoma is 79% and 98%, respectively, when the cutoff value is 129 U/mL. In patients without PSC, the sensitivity is 53% with a cutoff value of 100 U/mL.23,24 Importantly, patients with a negative Lewis antigen status do not express CA 19-9, and significant CA 19-9 elevations can occasionally be observed in patients with bacterial cholangitis and choledocholithiasis.

1173

1174

Section VIII  Biliary Tract Clinical suspicion of CCA

CA 19-9 level Endoscopic cholangiography (brushing, cytology, DIA, FISH)

Figure 69-2.  Diagnostic algorithm for hilar cholangiocarcinoma. In cases of clinically suspected hilar cholangiocarcinoma, a serum CA 19-9 level, endoscopic retrograde cholangiopancreatography, and conventional as well as molecular cytologic analysis of endoscopically obtained biliary brushings of malignant-appearing areas should be performed. If results of these tests are normal or negative, observation is recommended. Management of cholangiocarcinoma should be prompted by identification of a dominant stricture, serum CA 19-9 level >129 U/mL, or a biopsy or cytology result that is positive for carcinoma or polysomy. In indeterminate cases, gadolinium-enhanced magnetic resonance imaging (MRI) of the liver with ferumoxide contrast is recommended. If a mass lesion or vascular encasement is identified, management of cho­ langiocarcinoma should be initiated. If the MRI study is negative but clinical concern about cholangiocarcinoma persists, positron emission tomography (PET) can be performed. If “hot spots” are identified on PET (positive result), further management should be directed toward cholangiocarcinoma. If the result of the PET scan is negative, close follow-up is recommended. If MRI is negative and cholangiocarcinoma is considered unlikely, the patient can be observed expectantly. CCA, cholangiocarcinoma; DIA, digital image analysis; FISH, fluorescence in situ hybridization; pos., positive. (Published with permission from Wiley InterScience.)

Dominant stricture CA 19-9 >129 U/mL Pos. biopsy, cytology, or FISH polysomy

Indeterminate

MRI

Mass Vascular encasement

No dominant stricture Neg. biopsy/cytology/ advanced cytology CA 19-9 <129 U/mL

Negative

Minimal Significant clinical concern clinical concern PET scan

“Hot spot” Management of CCA

An essential diagnostic tool for cholangiocarcinoma is cholangiography, which provides both anatomic information and material for a tissue diagnosis. Several approaches can be used, including endoscopic retrograde cholangiopancreatography (ERCP), percutaneous transhepatic cholangiography (THC), and magnetic resonance cholangiography (MRCP) (Figs. 69-3 and 69-4). The choice of ERCP versus percutaneous THC depends on the location of a suspicious biliary stricture, local expertise, and accessibility of the stricture to the technique. Either technique provides information on intrabiliary tumor extension and allows cytologic sampling and therapeutic intervention for malignant biliary obstruction. Although such interventions are not possible with MRCP, this technique is noninvasive and provides additional information on the extent of extrabiliary tumor, vascular encasement, the relation of the primary tumor to surrounding structures, and intra- as well as extrahepatic metastases.2 Computed tomography (CT) is used primarily for preoperative planning; it provides information on vascular and other anatomic structures that affect surgical decision making. The sensitivity of CT for detection of nodal N2 metastases (see later), however, is only 50%, and its accuracy in the evaluation of resectability is only 60% to 75%. Magnetic resonance imaging (MRI) is currently the best imaging technique for cholangiocarcinoma; its sensitivity and tumor tissue depiction are improved by the use of ferumoxide as a contrast agent.25 Cross-sectional imaging should be performed prior to cholangiography to help guide decisions regarding biliary drainage (see later). The use of positron emission tomography (PET)/CT should be limited to cases that are indeterminate with regard to extent and other features (including malignancy) after the other studies are

Negative Observation

performed. Its sensitivity for regional lymph node metastases is only 12%. Also, the sensitivity of CT and PET for the detection of cholangiocarcinoma do not differ. PET/CT achieves sensitivity and specificity rates of 93% and 80% for the detection of intrahepatic cholangiocarcinoma but only 55% and 33% for the detection of extrahepatic cholangiocarcinoma. False-positive PET results can be observed in the setting of inflammation.26 Endoscopic ultrasound (EUS) can contribute additional information about tumor dimension and anatomic location in relation to surrounding structures. In addition, it provides information about regional lymphadenopathy and allows sampling of lymph nodes by fine-needle aspiration. Biopsy of the primary lesion via EUS is highly discouraged, however, because of the potential for tumor cell spread.2 As for most malignancies, a tissue diagnosis of cholangiocarcinoma is desirable; however, a pathologic diagnosis is challenging because of the tumor’s paucicellular character. Tissue is difficult to obtain, and cellular reactive changes resulting from inflammation often complicate the diagnosis. The most common technique for collecting tumor tissue is brush cytology during ERCP, but conventional cytology is limited because of low specificity. Fluorescence in situ hybridization (FISH) has been demonstrated to increase the sensitivity and specificity of cytology for diagnosing cholangiocarcinoma in patients with and without PSC.27 Typical findings include (1) trisomy 7; (2) tetrasomy or duplication of all labeled chromosomes; or (3) polysomy or amplification of at least three chromosomes (Fig. 69-5). Polysomy is diagnostic of cholangiocarcinoma in the appropriate clinical setting, whereas trisomy 7 is not diagnostic but increases a patient’s risk for the development of cholangiocarcinoma.

Chapter 69  Tumors of the Bile Ducts, Gallbladder, and Ampulla

A A

B B Figure 69-4.  Imaging of intrahepatic cholangiocarcinoma. A gadoliniumenhanced magnetic resonance image with ferumoxide in a patient with an intrahepatic cholangiocarcinoma is depicted. The arrows point to the tumor in the T2-weighted images. A, Axial view. B, Coronal view.

that remain unconfirmed after use of the aforementioned methods when the level of suspicion for cholangiocarcinoma is high. The value of these techniques remains to be determined.

C Figure 69-3.  Radiologic presentation of extrahepatic cholangiocarcinoma. A, Film from endoscopic retrograde cholangiography in a patient with hilar cholangiocarcinoma demonstrating dominant strictures of the biliary tree consistent with Bismuth-Corlette type IV. B, Gadolinium-enhanced magnetic resonance imaging with ferumoxide in the same patient. The arrow points to the biliary tumor seen on a T2-weighted image. C, Positron emission tomography/computed tomography scan of the same patient. The biliary tumor is seen as an enhancing region (arrow).

Tetrasomy has to be interpreted with caution because it can be observed during the M phase of the cell cycle during mitosis (see Chapter 3).2 Newer techniques for diagnosing cholangiocarcinoma include choledoscopy (cholangioscopy), intraductal ultrasound, and choledochoenteroscopy. They are not yet established in routine practice and should be reserved for cases

STAGING

Cholangiocarcinoma has been classified according to the American Joint Committee on Cancer (AJCC)/ International Union Against Cancer (UICC) TNM (tumor node metastasis) system.28 This system is based on hepatocellular carcinoma and is therefore only of limited use in cholangiocarcinoma. It is especially suboptimal for hilar cholangiocarcinoma. An improved, preoperative staging system for extrahepatic cholangiocarcinoma is needed. In oncologic surgery, negative surgical margins (R0) correlate with lack of postoperative recurrence and with prognosis. A staging system for hilar cholangiocarcinoma that incorporates resectability requires that the extent of biliary disease, vascular encasement, and hepatic lobar atrophy be assessed, in addition to the information provided by a clinical TNM system. Unilobar atrophy is clinically important, because it indicates unilateral biliary obstruction with ipsilateral vascular encasement (atrophy-hypertrophy complex). This finding does not indicate unresectability unless it occurs in combination with contralateral vascular or biliary tumor involve-

1175

1176

Section VIII  Biliary Tract

A

Normal

B

Trisomy 7

C

Polysomy

Figure 69-5.  Fluorescence in situ hybridization (FISH) analysis of biliary brushings. Fluorescent photographic micrographs of FISH analysis of biliary brushings are shown. Each colored signal represents one chromosome. Two signals of the same color indicate a normal, diploid status (A). Three signals of the same color are representative of trisomy 7 (B), and more than three signals of the same color indicate polysomy (C). (From Jarnagin WR, Fong Y, DeMatteo RP, et al. Staging, resectability, and outcome in 225 patients with hilar cholangiocarcinoma. Ann Surg 2001; 234:507-17, with permission from Wiley InterScience.)

Table 69-4  Sloan-Kettering Cancer Staging System for Cholangiocarcinoma

Table 69-5  Exclusion Criteria for Resection of Hilar Cholangiocarcinoma

STAGE

CRITERIA

T1

Tumor involving biliary confluence ± unilateral extension to secondary radicles Tumor involving biliary confluence ± unilateral extension to secondary radicles AND Ipsilateral portal vein involvement ± ipsilateral hepatic lobe atrophy Tumor involving biliary confluence + bilateral extension to secondary radicles OR Unilateral extension to secondary radicles + contralateral portal vein involvement OR Main or bilateral portal vein involvement

Atrophy of one liver lobe with encasement of the contralateral portal vein branch Atrophy of one liver lobe with contralateral secondary biliary radicle involvement Bilateral portal vein branch encasement Bilateral hepatic artery encasement Distant lymph node metastases Hilar cholangiocarcinoma, Bismuth-Corlette type IV Intrahepatic or distant metastases Primary sclerosing cholangitis Significant comorbid conditions

T2

T3

ment.29 Such a staging system has been proposed at Memorial Sloan-Kettering Cancer Center (MSKCC) (Table 69-4). Resectability, likelihood of curative (R0) resection, metastatic spread to N2-level lymph nodes (where N1 signifies metastasis to lymph nodes within the hepatoduodenal ligament and N2 signifies metastasis to peripancreatic, periduodenal, celiac, superior mesenteric, or posterior pancreaticoduodenal lymph nodes), and survival correlate with the tumor stage in this classification system.29

TREATMENT

The only curative treatment option for cholangiocarcinoma is surgical extirpation. Surgical outcomes have improved substantially in the 2000s because of careful patient selection with lower surgical mortality rates and higher rates of R0 (negative surgical margin) resection.30 Surgical treatment options include resection and liver transplantation (see Chapter 95). Solitary intrahepatic cholangiocarcinoma can be resected by hepatic segmentectomy or lobectomy. Five-year survival rates after resection of intrahepatic cholangiocarcinoma range from 22% to 42%. Survival is positively correlated with early tumor stage (localized tumor), younger age, and

better performance status.30 Liver transplantation is not an option for intrahepatic cholangiocarcinoma because of disappointing five-year survival rates of 0% to 18%.31 Resection is also the treatment of choice for extrahepatic cholangiocarcinoma in the absence of PSC (Table 69-5). Five-year survival rates for extrahepatic CCA after R0 resection are 11% to 41% for hilar and 27% to 37% for distal cholangiocarcinoma. Unfortunately, R0 resectability rates are less than 50%.32 Occasionally, increased resectability can be achieved by preoperative portal vein embolization, resulting in compensatory hyperplasia of the contralateral hepatic lobe. This technique allows extended partial hepatectomy because of the increased volume of the remnant liver.33 Neither adjuvant nor neoadjuvant chemotherapy or radiation therapy has been shown to be effective and is therefore not recommended in patients with cholangiocarcinoma who undergo resection. In patients with PSC, resection may be complicated by liver failure or a second cholangiocarcinoma, which has a tendency to develop in the area of a biliary-enteric anastomosis. Therefore, these patients should be considered for liver transplantation instead of resection. In the past, liver transplantation was not an option for patients with extrahepatic cholangiocarcinoma because five-year survival rates were only 23% to 26%.34 Subsequently, new liver transplantation protocols were developed and led to 5-year survival rates of 76% in carefully selected patients with hilar cholangiocarcinoma.35 The

Chapter 69  Tumors of the Bile Ducts, Gallbladder, and Ampulla selection criteria are strict (see Table 69-3 and Chapter 95). The treatment protocol consists of neoadjuvant external beam radiation therapy with concurrent systemic 5fluorouracil chemotherapy, followed by brachytherapy and oral maintenance chemotherapy with capecitabine until transplantation. At the current time, no curative medical therapies for cholangiocarcinoma are available. A variety of chemotherapeutic agents such as gemcitabine, other antimetabolites, taxanes, platinum analogs, anthracyclines, and mitomycin have been evaluated as single or combination therapies. The only chemotherapeutic drug approved by the U.S. Food and Drug Administration (FDA) for cholangiocarcinoma is gemcitabine; however, no randomized, controlled phase III trials have shown a significant survival benefit with chemotherapy in patients with cholangiocarcinoma. Available studies have been statistically flawed or underpowered or have shown poor response rates. Therefore, chemotherapy is not a treatment option for these malignancies unless used in the palliative (see below) or neoadjuvant setting. Similarly, the use of radiation therapy remains controversial.2 The growing understanding of the molecular pathogenesis of a variety of human malignancies has led to the development of targeted therapies such as tyrosine kinase inhibitors. Targeted agents have achieved significant prolongation of survival when used to treat other cancers, such as hepatocellular carcinoma and renal cell carcinoma. Randomized, controlled trials of these agents are needed in patients with cholangiocarcinoma, with survival, rather than tumor response, as the primary endpoint.

Palliative Treatments

Cholangiocarcinoma is associated with substantial morbidity. Commonly, patients experience cholestasis, abdominal pain, and cachexia, which limit the quality of life. Therefore, palliative treatments are essential in the management of patients with cholangiocarcinoma. Options for restoration of biliary drainage include endoscopic, percutaneous, and surgical techniques. Endoscopic and percutaneous methods are based on placement of biliary stents (see Chapter 70), whereas surgical approaches create a bypass via a choledocho- or hepaticojejunostomy. The efficacies of similar endoscopic and surgical approaches are similar, but the mortality rate, frequency of procedure-related complications, and duration of hospital stay are higher for surgical palliation.36 The decision to pursue endoscopic or percutaneous biliary stent deployment is based on the anatomic location of the malignant stricture (see Chapter 70). In general, unilateral restoration of bile flow is sufficient. Cross-sectional imaging is critical before a stent is placed to avoid attempts at endoscopic drainage of an atrophic lobe or a lobe in which adequate drainage is not feasible. Retrograde injection of dye without drainage carries a high risk of iatrogenic bacterial cholangitis, which can be severe. Early intervention in a patient with malignant biliary obstruction is recommended because the time to normalization of the serum bilirubin level doubles from three to six weeks when the total bilirubin serum level is greater than 10 mg/dL.37 Most recently, photodynamic therapy (PDT) has been shown to be a feasible palliative treatment option that reduces cholestasis, improves quality of life, and possibly provides a survival benefit.38,39 PDT involves the systemic application of a photosensitizing agent (e.g., hematoporphyrin), followed by localized illumination of the tumor at a specific wavelength. Tumor cell cytotoxicity is achieved

through reactive oxygen species-mediated cell death, tumorvessel thrombosis, and tumor-specific immune reactions. The procedure is well tolerated and characterized by a low complication rate.

GALLBLADDER CARCINOMA Gallbladder carcinoma is the second most common primary biliary malignancy and the fifth most common malignancy of the gastrointestinal tract. Like other biliary malignancies, gallbladder carcinoma is diagnosed at an advanced stage in the majority of cases. In only one third of the cases is the specific diagnosis of gallbladder carcinoma made prior to surgical exploration.40 The growth kinetics of gallbladder carcinoma are faster than those for cholangiocarcinoma, and, in general, gallbladder carcinoma is diagnosed at a later stage than ampullary carcinoma (see later). Gallbladder carcinoma is not amenable to medical or radiation therapy, and surgical resection is the only potentially curative treatment. Unfortunately, only a minority of patients are surgical candidates at the time of diagnosis. The prognosis of gallbladder carcinoma is dismal, with five-year survival rates of 0% to 10% and median survival of less than six months. In the 2000s, more aggressive surgical approaches have been advocated to improve outcomes.

EPIDEMIOLOGY

The distribution of gallbladder carcinoma is geographically heterogeneous, with the highest incidence rates (up to 21.5 per 100,000 population) observed in India. Incidence rates are also high in South America, Asia, and certain Eastern European countries such as Poland.41 Gallbladder carcinoma is rare in Western European countries and the United States, where the National Cancer Institute reported an age-adjusted incidence rate of 1.2 per 100,000 in 2007. Global incidence rates of gallbladder carcinoma parallel the incidence rates of cholelithiasis. With the exception of Japan, where the incidence has increased, age-adjusted incidence rates of gallbladder carcinoma have remained relatively stable in most countries since the 1970s.41 Ethnic differences in the incidence of gallbladder carcinoma have been described within the United States, with higher rates in whites than in African Americans. Mortality rates also vary globally. The age-adjusted mortality rate in the United States between 2000 and 2005 was 0.7 per 100,000, with an overall decrease since 1990. The highest mortality rate (35 per 100,000) was reported in southern Chile.42 The average age at diagnosis is 65 years, and the peak incidence is observed in the seventh and eighth decades of life. Globally, there is a female predisposition to gallbladder carcinoma.41

ETIOLOGY

The cause of gallbladder carcinoma is not well understood but is thought to be multifactorial. Several risk factors for gallbladder carcinoma have been described (Table 69-6). The primary risk factor for gallbladder carcinoma is cholelithiasis (see Chapter 65). Gallstones are found in 65% to 90% of patients with gallbladder carcinoma. Populations with high rates of cholelithiasis also have high rates of gallbladder carcinoma. Autopsy-based studies from Chile have suggested a seven-fold increased risk of gallbladder carcinoma in patients with cholelithiasis, whereas epidemio-

1177

1178

Section VIII  Biliary Tract Table 69-6  Risk Factors for Gallbladder Carcinoma Adenomyomatosis Anomolous union of the pancreaticobiliary ductal system Carcinogens* Cholangiocarcinoma Cholelithiasis (stone size >1 cm) Chronic Salmonella typhi or paratyphi carrier status First-degree relative with gallbladder cancer Inflammatory bowel disease Intrahepatic biliary dysplasia Porcelain gallbladder Primary sclerosing cholangitis *Methylcholanthrene, O-aminoazotoluene, nitrosamines, possibly others.

logic studies in the United States have observed only a marginally significant three-fold increased risk of gallbladder carcinoma in men with cholelithiasis.43,44 Gallbladder carcinoma actually develops in only 1% to 3% of patients with cholelithiasis, and 20% of patients with gallbladder carcinoma do not have evidence of cholelithiasis. Therefore, a prophylactic cholecystectomy in an asymptomatic patient with gallstones to prevent gallbladder carcinoma cannot be recommended. A positive correlation between the risk of gallbladder carcinoma and the size and number of gallstones has been reported but reflects the duration of cholelithiasis.45 No differences in the risk of gallbladder carcinoma have been observed with different types of gallstones. Porcelain gallbladder (extensive calcification of the gallbladder wall) is a classic, although controversial, risk factor for gallbladder carcinoma.46 Although an increased risk of gallbladder carcinoma has been reported in patients with a porcelain gallbladder, the risk may be limited to patients with selective mucosal calcification (types II and III porcelain gallbladder) rather than those with diffuse mucosal calcification (type I).47 Adenomatous polyps of the gallbladder constitute another risk factor for gallbladder carcinoma (see Chapter 67). The risk correlates positively with the size, type, and growth rate of the polyps. Patients with polyps that are greater than 1 cm in size, sessile, and associated with gallstones, exhibit a rapid increase in size, demonstrate arterial flow on Doppler ultrasonography, or are symptomatic are at increased risk of malignant transformation, and patients with such polyps warrant prophylactic cholecystectomy.46,48 Anomalous union of the pancreaticobiliary ductal system (AUPBD) has been associated with the development of gallbladder carcinoma. In this congenital defect, the pancreatic and bile ducts unite outside the duodenal wall in a long common channel. The anomaly is found incidentally in 1.5% to 2% of patients who undergo ERCP and leads to cholestasis and reflux of pancreatic secretions into the gallbladder with resulting chronic inflammation of the mucosa. Approximately 10% of patients with gallbladder carcinoma have coexisting AUPBD, and gallbladder carcinoma develops in 15% to 40% of those with AUPBD. Patients with an associated choledochal cyst have a lower frequency of gallbladder carcinoma than those without a choledochal cyst.49 Patients with AUPBD are usually 10 years younger at the time of diagnosis of gallbladder carcinoma and have a lower frequency of cholelithiasis than those without AUPBD. On the basis of a significantly increased risk of gallbladder carcinoma, several Japanese hepatobiliary oncology associations have recommended an aggressive, prophylactic surgical approach to patients with AUPBD.46

PSC has been associated with gallbladder carcinoma, and studies have reported that adenocarcinoma of the gallbladder develops in up to 20% of patients with PSC and that 40% to 60% of gallbladder masses in patients with PSC are malignant.50,51 Therefore, patients with PSC and a gallbladder mass of any size should undergo prophylactic cholecystectomy or be monitored closely for gallbladder carcinoma. Adenomyomatosis of the gallbladder, which is characterized by microscopic invaginations (Rokitansky-Aschoff sinuses) of the mucosa with cyst formation in the muscularis propria, has been linked to gallbladder carcinoma (see Chapter 67). The magnitude of the malignant potential of adenomyomatosis has not been clearly defined and appears to depend on morphologic features and the patient’s age.52 Adenomyomatosis generally is viewed as a benign condition. Other conditions associated with gallbladder carcinoma include inflammatory bowel disease, intrahepatic biliary dysplasia, and cholangiocarcinoma.51 Chronic carriers of Salmonella typhi or paratyphi have been shown to be at increased risk for the development of gallbladder carcinoma.53 Other bacteria such as Escherichia coli and Helicobacter pylori also have been associated with gallbladder carcinoma, but the data are not conclusive. First-degree relatives of patients with gallbladder carcinoma have a relative risk of 13.9 for developing this malignancy.54 Carcinogens, including methylcholanthrene, O-aminoazotoluene, and nitrosamines, have been identified in animal models of gallbladder carcinoma. Other potential carcinogens include mustard oil, products of free radical oxidation, and secondary bile acids.55

PATHOLOGY

From 80% to 95% of gallbladder carcinomas are adenocarcinomas; the majority of these are moderately-to-well differentiated.42 Less common types, in order of frequency, include undifferentiated or anaplastic carcinoma, squamous cell carcinoma, and adenosquamous carcinoma. Rare types include carcinoids, small cell carcinomas, malignant melanomas, lymphomas, and sarcomas.55 Sixty percent of gallbladder carcinomas are located in the gallbladder fundus, 30% in the body, and 10% in the gallbladder neck.56 Analogous to cholangiocarcinoma, the papillary form of gall­bladder carcinoma has a lower potential for invasion and metastatic spread to lymph nodes.57 Gallbladder carcinoma spreads via direct invasion, lymphogenic or hematogenic metastasis, perineural invasion, and intraperitoneal or intraductal invasion. Lymphatic tumor cell spread is determined by the physiologic gallbladder lymphatic plexus, including the first-level lymph nodes along the biliary tract (cystic duct, bile duct, and hepatic duct) followed by pancreaticoduodenal lymph nodes as well as lymph nodes along the common hepatic artery and celiac axis. Lymph node metastases are described in 54% to 64% of patients and correlate with the depth of invasion. Gallbladder carcinoma has a predisposition to involve the liver bed because of venous drainage, predominantly into hepatic segments IVb and V (see Chapter 71), and the anatomic approximation that allows direct hepatic invasion. Perineural spread is observed in 24% and intraductal spread in 19% of cases.

PATHOGENESIS

Gallbladder carcinoma can develop from foci of mucosal dysplasia or carcinoma in situ (CIS) that progress to adenocarcinoma or from an adenoma-carcinoma sequence similar

Chapter 69  Tumors of the Bile Ducts, Gallbladder, and Ampulla to that seen with colon cancer (see Chapter 123).58 Foci of dysplasia and CIS are frequently found adjacent to gallbladder carcinoma in surgically resected gallbladder specimens and are thought to be precursors of invasive adenocarcinoma.42 The time of progression of dysplasia to carcinoma is estimated to be 10 to 15 years.59 Like cholangiocarcinoma, the major pathogenic factor is inflammation. Increased iNOS and COX-2 expression has been demonstrated immunohistochemically in gallbladder carcinoma samples as well as in hyperplastic gallbladder mucosa from patients with AUPBD and also has been associated with TP53 tumor suppressor gene mutations in patients with gallbladder carcinoma. High expression and mutation rates of the TP53 gene have been demonstrated in 35% to 92% of gallbladder carcinomas, 86% of carcinomas in situ, and 28% of dysplastic foci, supporting an early role for TP53 mutation in the dysplasia-carcinoma progression sequence.60,61 In up to 60% of patients with gallbladder carcinoma, mutations of the K-ras oncogene have been detected; the frequency is highest in patients with AUPBD.62 Single studies have reported upregulation of ErbB2 and the nm23 metastasis suppressor protein.63,64 Mutations and increased expression of the nuclear oncogene that encodes p16INK4 have been shown in several studies.65 Other studies have demonstrated loss of heterozygosity or microsatellite instabilities in chromosomal regions that harbor known or putative tumor suppressor genes.66 Therefore, available data indicate pathogenetic mechanisms similar to those of other biliary malignancies, but the molecular pathogenesis appears to depend on the specific etiology and mode of progression.

A

CLINICAL FEATURES AND DIAGNOSIS

In two thirds of cases, gallbladder carcinoma is diagnosed incidentally during or after cholecystectomy for presumed benign disease, reflecting the initial clinically silent nature of this malignancy. Common clinical presentations include biliary or abdominal pain and jaundice secondary to direct invasion of the biliary ducts or metastases to the hepatoduodenal ligament. Weight loss, abdominal distention, or other symptoms resulting from compression or invasion of adjacent organs indicate more advanced disease. CEA and CA 19-9 are the most commonly used tumor markers in gallbladder carcinoma. As indicated earlier, these tests aid in diagnosis but should not be relied on, because levels can be elevated in inflammatory conditions and gastroenterologic and gynecologic malignancies; moreover, a subset of the population does not produce CEA. Abdominal ultrasound is often one of the first imaging studies performed in a patient who presents with the aforementioned symptoms. The sensitivity and accuracy of ultrasound for gallbladder carcinoma are 85% and 80%, respectively; early cancers, especially sessile polyps, can be missed. Typical imaging presentations of gallbladder carcinoma include focal or diffuse mural thickening of the gallbladder, an intraluminal mass greater than 2 cm in size that originates in the gallbladder wall, and a subhepatic mass that replaces or obscures the gallbladder and often invades adjacent organs (Fig. 69-6). Findings indicative of the malignant nature of a gallbladder lesion include irregular, asymmetrical mural thickening greater than 1 cm in depth and a nodular or smooth intraluminal mass greater than 1 cm in size, with fixation to the gallbladder wall, that is not displaced by the patient’s movements and has no acoustic shadow. In indeterminate cases, Doppler ultrasound can be attempted to differentiate a malignant from a benign gallbladder lesion on the basis of the pattern of the color signal, blood flow velocity, and resistive index (a measure of

B

C Figure 69-6.  Imaging of gallbladder carcinoma by computed tomography (CT) and ultrasonography. A, Axial CT view of the abdomen. Cholelithiasis is seen inferior to the gallbladder mass (arrow). B, Coronal view of the same patient. C, Ultrasonogram in the same patient showing a large mass (arrow) originating from the gallbladder wall and protruding into the lumen.

resistance to arterial blood flow).67 MRI and CT can be helpful in the diagnosis if the ultrasound findings are indeterminate. The role of PET in gallbladder carcinoma is evolving, and currently PET is not part of the routine evaluation. The sensitivity of PET for detecting gallbladder carcinoma is only 75% to 78%.68,69 Its main impact is in the

1179

1180

Section VIII  Biliary Tract Table 69-7  TNM and American Joint Committee on Cancer (AJCC)/International Union Against Cancer (UICC) Staging Systems for Gallbladder Carcinoma TNM STAGE

CRITERIA

Tx T0 Tis T1a T1b T2 T3

Primary tumor cannot be assessed No evidence of primary tumor Carcinoma in situ Tumor invades lamina propria Tumor invades muscularis propria Tumor invades perimuscular connective tissue without extension beyond serosa or into liver Tumor invades serosa OR Tumor invades one adjacent organ OR Both (extension ≤2 cm into liver) Tumor extends >2 cm into liver AND/OR Tumor extends into ≥2 adjacent organs (stomach, duodenum, colon, pancreas, omentum, extrahepatic bile ducts, liver)

T4

Nx N0 N1 N2

Regional lymph nodes cannot be assessed No regional lymph node metastases Metastases in cystic duct, pericholedochal, and/or hilar lymph nodes Metastases into peripancreatic (head only), periduodenal, periportal, celiac, and/or superior mesenteric lymph nodes

Mx M0 M1

Distant metastases cannot be assessed No distant metastases Distant metastases

AJCC/UICC Stage

Tumor

Node

Metastasis

0 IA IB II A II B III IV

Tis T1 T2 T3 T1-3 T4 Any T

N0 N0 N0 N0 N1 Any N Any N

M0 M0 M0 M0 M0 M0 M1

TNM, tumor, node, metastasis.

detection of distant metastases that result in a change in management.26

STAGING

The most commonly used staging system is the TNM system described by the AJCC and UICC. The TNM-based staging system correlates with survival. Reported five-year survival rates for patients with stages 0, I, II, III, and IV gallbladder carcinoma are 60%, 39%, 15%, 5%, and 1%, respectively (Table 69-7).

TREATMENT

Surgery is the only potential curative therapeutic option for gallbladder carcinoma. Only 15% to 47% of patients are candidates for surgery at the time of diagnosis because of the advanced stage of their disease. Contraindications to resection include multiple hepatic or peritoneal metastases, malignant ascites, distant metastases, extensive involvement of the hepatoduodenal ligament, encasement or occlusion of major vessels, and poor performance status. Direct involvement of the colon, duodenum, or liver is not considered an absolute contraindication to surgical resection, however. The goal of surgical treatment is an R0 resection, defined as negative margins and nodal dissection one level past microscopically involved lymph nodes. R0 resection in gallbladder carcinoma has been shown to correlate with survival and with significantly increased five-year survival rates.70 Surgical procedures with curative extent include (1) simple cholecystectomy; (2) extended or radical cholecystectomy with additional resection of greater than 2 cm

of the gallbladder bed plus lymphadenectomy of the hepatoduodenal ligament behind the second part of the duodenum, head of the pancreas, and celiac axis; (3) extended cholecystectomy with hepatic, segmental, or lobar resection; (4) extended cholecystectomy with extensive para-aortic lymph node resection; and (5) extended cholecystectomy with bile duct resection or pancreaticoduodenectomy. The surgical approach is dictated by the extent of tumor. Stage Tis and T1a (see Table 69-7) gallbladder cancer can be treated with simple cholecystectomy, with five-year survival rates of 85% to 100%. A few reports favor simple cholecystectomy also for stage 1b gallbladder carcinoma and report similar survival rates after either simple or radical cholecystectomy.71,72 Up to 15% of patients with stage 1b gallbladder carcinoma, however, are positive for lymph node metastases, compared with 2.5% of patients with stage 1a gallbladder carcinoma. Also, higher recurrence rates have been observed after simple (versus radical) cholecystectomy; therefore, radical cholecystectomy is preferred for stage 1b gallbladder carcinoma.73,74 Invasion of the muscularis propria, as in stage T2 or higher tumors, requires radical cholecystectomy. The extent and benefit of partial hepatectomy in stage T2 tumors are controversial, and partial hepatectomy has not been shown to prolong survival. Also, the surgical approach to stage T3 and T4 tumors is controversial. Some studies show no five-year survival benefit after radical cholecystectomy for stage T3 and T4 tumors, but other studies report five-year survival rates of 15% to 63% and 7% to 25%, respectively. Because of the poor prognosis of gallbladder carcinoma and the possibility of a survival benefit, as well as prolongation of

Chapter 69  Tumors of the Bile Ducts, Gallbladder, and Ampulla Postoperative diagnosis of gallbladder carcinoma

Staging

T2, T3, T4

T1

T1a

Re-exploration

T1b

Resectable

No further treatment if margins are negative

M1

Radical cholecystectomy

Unresectable

Palliative treatment

survival until recurrence, a radical surgical approach to these advanced-stage gallbladder carcinomas is recommended by many centers. When gallbladder carcinoma is diagnosed during laparoscopy, the procedure should be converted to an open procedure, and the laparoscopic port sites should be resected, because tumor may recur at these sites secondary to iatrogenic dissemination.75 Further surgical management then depends on the tumor stage, as outlined above and in Figure 69-7. When gallbladder carcinoma is diagnosed postoperatively, further management depends on the tumor stage and the presence or absence of tumor at the margins of the surgical specimen. Currently, chemotherapy and radiation therapy are not considered standard treatment for gallbladder carcinoma. A variety of different chemotherapeutic drugs, either alone or in combination, have been evaluated, including gemcitabine, fluoropyrimidines, mitomycin, and platinumbased compounds. Partial response rates of up to 36% and stabilization of disease in up to 48% of patients have been reported. The number of trials is limited, the number of patients in these trials has been small, and none of the studies has been randomized and controlled. Similarly, the role of radiation therapy is not defined in gallbladder carcinoma. In general, gallbladder carcinoma is considered radioresistant. Use of radiation therapy is further limited by the radiosensitivity of the surrounding tissue.

AMPULLARY CARCINOMA Carcinomas of the ampulla of Vater belong to the family of periampullary carcinomas. This family includes carcinomas of the duodenum, ampulla of Vater, distal bile duct, and pancreas. Ampullary carcinomas are the second most common form of periampullary carcinoma (after pancreatic head cancer). The distinction between the different forms is important because ampullary carcinomas are often diagnosed earlier than the others and therefore at a resectable stage, thus resulting in a better prognosis.

Figure 69-7.  Diagnostic algorithm for gallbladder carcinoma diagnosed intra- or postoperatively at laparoscopic cholecystectomy. In cases in which pathologic examination of the surgical cholecystectomy specimen identifies a stage T1a tumor with negative surgical margins, no further treatment is indicated. If the tumor is found to be a stage T1b tumor or the margins of resection are positive for malignant tissue, re-exploration for further resection is indicated. Similarly, patients with gallbladder carcinoma found to be stage T2, T3, or T4 should undergo surgical re-exploration. If re-exploration reveals resectable gallbladder carcinoma, radical cholecystectomy should be performed. If the tumor is deemed unresectable, palliative management is indicated. When postoperative staging reveals metastatic spread, palliative management is indicated. M, metastasis stage; T, tumor stage. (Modified from Misra S, Chaturvedi A, Misra NC, Sharma ID. Carcinoma of the gallbladder. Lancet Oncol 2003; 4:167-76.)

EPIDEMIOLOGY

Ampullary carcinomas are rare, accounting for fewer than 1% of all gastrointestinal cancers and 4% to 8% of periampullary carcinomas. The annual incidence has been estimated to be 0.6 per 100,000 population.76,77 Peak incidence is in the seventh decade of life. There is a slight male predominance, with a male-to-female ratio of 1.48 : 1.77 Racial heterogeneity has been observed; the vast majority of patients are white, followed by patients of Hispanic and Asian descent. African Americans have the lowest incidence rates in the United States.76

ETIOLOGY

Although the etiology of ampullary carcinomas is unknown in the majority of cases, several conditions have been associated with this malignancy, mostly in case reports or small series. Familial adenomatous polyposis (FAP) is an important risk factor for the development of ampullary carcinomas, with a relative risk of 124 (see Chapter 122).78 Periampullary carcinoma is the second most common cause of death (after colon cancer) in patients with FAP. Usually, periampullary carcinoma arises later than colorectal carcinoma in this patient group but earlier in comparison with sporadic ampullary carcinomas.77 Screening for upper gastrointestinal lesions (polyps or carcinoma) at regular intervals of six months to four years, depending on the degree of duodenal polyposis, is therefore recommended in patients with FAP. Similarly, increased rates of ampullary carcinoma have been described in patients with Gardner’s syndrome, a variant of FAP (see Chapter 122).79 Hereditary nonpolyposis colorectal cancer does not appear to pre­ dispose to ampullary carcinoma.80 Other genetic diseases reported to predispose to the development of ampullary carcinoma include neurofibromatosis type I and Muir-Torre syndrome.81,82 As for cholangiocarcinoma, chronic liver fluke infection has been reported to be a risk factor for ampullary carcinoma (see Chapter 82).77

PATHOLOGY

The ampulla of Vater is an anatomically complex area that consists of the papilla, common pancreaticobiliary channel,

1181

1182

Section VIII  Biliary Tract type typically develop from adenomas, whereas pancreaticobiliary and ulcerating carcinomas often lack a precursor lesion.77 On the molecular level, an increased frequency of K-ras mutations has been observed in 24% to 47% of tumors and is more common in the intestinal than pancreaticobiliary type.88 Also, p53 overexpression has been observed in 46% of tumors and is thought to be associated with ulcerating ampullary carcinomas. In an immunohistochemical study, aberrant expression of cell cycle regulators (i.e., p21WAF1/CIP1, p27Kip1, p16INK4, cyclin D1, type 8, and cyclin E, and the retinoblastoma protein [pRb]), was observed.89 These changes are similar to those seen in colorectal and pancreatic carcinomas, yet are distinctive. More research is necessary to understand the development of these tumors on a molecular level.

CLINICAL FEATURES AND DIAGNOSIS

Figure 69-8.  Endoscopic appearance of ampullary carcinoma. A catheter has been placed in the ampulla of Vater for biliary drainage after a sphincterotomy was performed.

distal bile duct, and distal main pancreatic duct. Macroscopically, ampullary carcinomas are classified into the following three subtypes: (1) intramural protruding (intraampullary), (2) extramural protruding (periampullary), and (3) ulcerating ampullary77 (Fig. 69-8). The ulcerating type is usually diagnosed at an advanced stage and has the highest rates of lymph node metastases. Consistent with its anatomic heterogeneity, the ampulla includes several different histologic cell types, such as epithelia of the common pancreaticobiliary channel, bile duct, pancreatic duct, or duodenal mucosa, Brunner’s glands, and aberrant pancreatic acini in the wall of the bile duct. The most common site of cellular atypia is found in the area of the common pancreaticobiliary channel, followed by the pancreatic duct, duodenal epithelium, and Brunner’s glands.83 Seventy-five percent of ampullary neoplasias are adenocarcinomas, 20% are benign adenomas, and 5% are neuroendocrine tumors.84 Adenocarcinomas account for 90% of ampullary malignancies; the rest include unusual types, such as mucinous, signet-ring cell, and undifferentiated carcinomas.84 Histopathologically, 90% of ampullary adenocarcinomas can be classified into pancreaticobiliary or intestinal types.85,86 Immunohistochemically, the two types can be distinguished by high cytokeratin 7 expression and a lack of intestinal apomucin (MUC2) in the pancreaticobiliary type and cytokeratin 20 as well as MUC2 expression in the intestinal type.87 The frequencies of the two different histologic types, as well as their correlation with invasion, lymph node metastases, and prognosis, differ in various studies, with some studies indicating that the frequency is higher and the prognosis is worse for the pancreaticobiliary type and other studies finding the opposite results.85-87

PATHOGENESIS

The majority of ampullary carcinomas follow an adenomacarcinoma sequence. In 30% to 91% of ampullary carcinomas, residual adenomatous tissue is found.77 Although precursor lesions can develop from intestinal as well as pancreaticobiliary-type tissue, carcinomas of the intestinal

Like the other periampullary and biliary malignancies, ampullary carcinomas present initially with obstructive jaundice in 70% to 82% of cases. Pancreaticobiliary ampullary carcinomas in particular have been reported to present initially with obstructive jaundice.85 Because of their anatomic location, cholestasis develops at an earlier stage than in other periampullary and biliary malignancies, and the resectability rate is therefore higher at the time of diagnosis. Nonicteric patients may present with bacterial cholangitis. Rare patients have “silver stools” as a result of the combination of acholic stools that result from bile duct obstruction and bleeding of the tumor. When obstructive cholangitis is suspected, further diagnostic evaluation is similar to that for other biliary malignancies. Immunohistochemical analysis has shown high expression of CEA and CA 19-9 in the tumor,90 but no studies have evaluated the serum concentrations of these markers in patients with ampullary carcinoma. Usually, ampullary carcinomas are diagnosed by endoscopy on the basis of their macroscopic appearance and findings on biopsy specimens (see Fig. 69-8). Subsequent diagnostic tests are directed toward an assessment of resectability and detection of metastases. As for other biliary and periampullary carcinomas, radiologic techniques such as CT and MRI are used commonly in this setting. On MRI/ MRCP, ampullary carcinoma usually is seen as a discrete, hypodense mass on T2-weighted images. Occasionally, the tumor can present as irregular thickening around the bile duct or bulging into the duodenum. Frequently, dilatation of both the bile and pancreatic ducts (“double-duct sign”) or only the bile duct is seen; dilatation of the pancreatic duct alone is seen rarely.91 Often, EUS is used in the preoperative evaluation. Its accuracy for detecting invasion of adjacent organs is 80% to 90%, and its sensitivity and specificity for detecting vascular invasion are 73% and 90%, respectively.92,93

STAGING

Ampullary carcinomas are classified according to the AJCC/ UICC TNM classification.28 The T stage was shown to be predictive of survival in a univariate analysis but not in a multivariate analysis.85,87 Independent predictors of survival by multivariate analysis are lymphovascular invasion, perineural invasion, stage greater than or equal to III, and pancreaticobiliary subtype.85 Several studies have confirmed the significance of nodal involvement in predicting survival.87,94

TREATMENT

As for other biliary malignancies, surgical resection is the only curative treatment for ampullary carcinomas. In

Chapter 69  Tumors of the Bile Ducts, Gallbladder, and Ampulla contrast to the other biliary malignancies, however, 77% to 88% of ampullary carcinomas are resectable at the time of diagnosis.95 The standard surgical approach is pancreaticoduodenectomy. Outcomes are excellent in the absence of lymph node metastases, with five-year survival rates of 68% to 78%.96,97 In the presence of lymph node–positive disease, the prognosis worsens significantly, with five-year survival rates of 16% to 25%.96,97 Extracapsular lymph node involvement results in further worsening of the prognosis, with a five-year survival rate of only 9%.90 Nodal microinvolvement has been reported to be an adverse prognostic factor, and immunohistochemical analysis of resected nodes has been recommended.98 Limited surgical or endoscopic papillectomy has been reported but is not recommended, because recurrence rates are higher than with pancreaticoduodenectomy.1 Chemotherapy and radiation therapy have not been evaluated in randomized controlled trials. A few studies have used 5-fluorouracil and radiation as adjuvant therapy in patients with node-positive disease and reported improved survival.99,100 In the absence of large, randomized controlled trials, however, such treatment is not standard. Palliative treatment should be directed at alleviating tumor-associated complications with the goal of optimizing the patient’s quality of life. Obstructive cholestasis is a major cause of morbidity and can usually be treated palliatively either by endoscopic or percutaneous placement of a biliary stent or by a surgical bypass similar to that carried out for other biliary or periampullary malignancies.

OTHER TUMORS OF THE BILE DUCTS AND GALLBLADDER Other neoplastic diseases may involve the biliary tract (Table 69-8). Their inclusion in the differential diagnosis of biliary tumors is essential, because management differs depending on the tumor type. Tumors of neuroectodermal origin such as carcinoids (see Chapter 31) and paragangliomas are rare and typically nonfunctioning.101 They are located most commonly in the ampulla of Vater. Occasionally, carcinoids develop in the extrahepatic biliary tree, predominantly in the bile duct. Patients are usually female and young. Primary carcinoids of the biliary tract constitute less than 1% of all gastrointestinal carcinoids and usually are not associated with the carcinoid syndrome.102,103 Approximately one third of patients have metastases at diagnosis. The treatment of choice is surgical resection, and the prognosis is generally good.104-106 Patients with paragangliomas often present with gastrointestinal bleeding; only 25% present with jaundice. Their malignant potential has been estimated to be 33%, and some investigators recommend pancreaticoduodenectomy as the treatment of choice.107 Granular cell tumors, which are of neuronal derivation, are extremely rare; only a few cases have been described. Usually, they are located in the extrahepatic biliary tree, particularly at the junction of the cystic duct and the bile duct.108 Occasionally, they can cause biliary obstruction, as occurs when they are located in the hepatic hilum.108 Because of their benign character, resection is usually curative.109 Rarely, neuromas of the extrahepatic biliary tree develop after cholecystectomy.110 Mesenchymal tumors, such as lipomas, leiomyomas, hemangiomas, and lymphangiomas, have been described in the gallbladder. In general, mesenchymal tumors are extremely rare and restricted to case reports. Lymphangiomas are often asymptomatic and only incidentally detected;

Table 69-8  Other Tumors of the Gallbladder and Bile Ducts GALLBLADDER Benign

Malignant

Tumor-like lesions

Adenoma Granular cell tumor Mesenchymal tumor (lipoma, leiomyoma, hemangioma, lymphangioma) Paraganglioma Adenosquamous carcinoma Carcinoid Small cell carcinoma Spindle cell sarcomatoid carcinoma Others (angiosarcoma, carcinosarcoma, Kaposi’s sarcoma, leiomyosarcoma, malignant fibrous histiocytoma, melanoma, metastatic tumors, non-Hodgkin’s lymphoma, rhabdomyosarcoma) Adenomyoma/adenomyomatosis Cholesterol polyp Heterotopia (gastric, pancreatic, liver, adrenal, thyroid) Inflammatory polyp

BILE DUCTS Benign

Malignant

Precursor lesions

Adenomyoma Bile duct adenoma Biliary adenofibroma Biliary cystadenoma and cystadenocarcinoma Biliary hamartoma Ciliated hepatic foregut cyst Granular cell tumor Neuroma Serous cystadenoma Solitary or multiple cysts Carcinoid Embryonal (botryoid) rhabdosarcoma Leukemia Lymphoma Melanoma Metastatic tumor Paraganglioma Biliary dysplasia (intraepithelial neoplasia/ atypical hyperplasia) Intraductal papillary neoplasia

however, they can increase in size and result in abdominal pain or jaundice. Ultrasound, CT, and MRI/MRCP aid in the preoperative diagnosis. Usually, they manifest as a multilocular, fluid-filled, cystic mass with thin walls and septa and show enhanced signal density with contrast administration.111 Most of the reported cases have been treated successfully with surgical resection, including cholecystectomy if the tumor is located within the gallbladder, or endoscopic resection if the tumor is in the area of the ampulla of Vater.112-115 Hamartomas also have been reported in the area of the ampulla of Vater and have been resected successfully by endoscopy.116 Heterotopia of the gallbladder may be caused by gastric, pancreatic, hepatic, adrenal, or thyroid tissue. Clinical complications such as hemorrhage are extremely rare.117,118 Benign bile duct lesions include adenomas, cystadenomas, adenofibromas, cysts, and granular cell tumors. Adenomyomas are found more commonly in the ampulla of Vater. Cystadenomas are more common in women and manifest primarily with abdominal pain. They are found predominantly in the intrahepatic biliary tree and are characterized on ultrasound by papillary extrusions of the wall

1183

1184

Section VIII  Biliary Tract and septa. They are considered premalignant because of their potential to transform into cystadenocarcinomas; hence, the treatment of choice is complete resection (see Chapter 94).119-121 Malignant tumors of the biliary tree other than cholangiocarcinoma include cystadenocarcinomas, lymphomas, and malignant melanomas. These malignancies arise primarily in the extrahepatic bile ducts. Cystadenocarcinomas can be distinguished morphologically from cholangiocarcinomas by their cystic character.122 They are rarely located in the gallbladder.123 Symptoms are nonspecific, and CT and MRI can be helpful in making the diagnosis. The treatment of choice is surgical resection.123 Malignant melanoma of the biliary tree is uncommon and should prompt investigation for a cutaneous melanoma, because cases of metastatic spread to the bile ducts have been described.124,125 Lymphomas can occasionally involve the extrahepatic biliary tree and often are mistaken for cholangiocarcinoma.126,127 In general, biliary lymphomas are very rare and account for less than 1% of lymphomas.127,128 Embryonal rhabdomyosarcoma of the biliary tree is extremely rare in adults but is the most common malignant tumor at this anatomic location in children.129 Frequently, it is misdiagnosed preoperatively as a choledochal cyst.130 Complete surgical resection is rarely possible, and a multidisciplinary approach to treatment is recommended. The prognosis of biliary rhabdosarcomas is good, with reported five-year survival rates of up to 78%.131 Few reports exist of follicular lymphomas originating in the extrahepatic biliary tree and gallbladder. Often, these tumors are diagnosed after resection.

ACKNOWLEDGMENTS

This work was supported by a grant from the NIH DK59427, the Mayo Clinic Clinical Investigator Program, and the Mayo Foundation.

KEY REFERENCES

Blechacz B, Gores GJ. Cholangiocarcinoma: Advances in pathogenesis, diagnosis, and treatment. Hepatology 2008; 48:308-21. (Ref 2.)

Buckles DC, Lindor KD, Larusso NF, et al. In primary sclerosing cholangitis, gallbladder polyps are frequently malignant. Am J Gastroenterol 2002; 97:1138-42. (Ref 50.) Csendes A, Becerra M, Rojas J, Medina E. Number and size of stones in patients with asymptomatic and symptomatic gallstones and gallbladder carcinoma: A prospective study of 592 cases. J Gastrointest Surg 2000; 4:481-5. (Ref 45.) Gores GJ, Nagorney DM, Rosen CB. Cholangiocarcinoma: Is transplantation an option? For whom? J Hepatol 2007; 47:455-9. (Ref 35.) Isomoto H, Kobayashi S, Werneburg NW, et al. Interleukin 6 upregulates myeloid cell leukemia-1 expression through a STAT3 pathway in cholangiocarcinoma cells. Hepatology 2005; 42:1329-38. (Ref 16.) Ito H, Matros E, Brooks DC, et al. Treatment outcomes associated with surgery for gallbladder cancer: A 20-year experience. J Gastrointest Surg 2004; 8:183-90. (Ref 70.) Jaiswal M, LaRusso NF, Burgart LJ, Gores GJ. Inflammatory cytokines induce DNA damage and inhibit DNA repair in cholangiocarcinoma cells by a nitric oxide-dependent mechanism. Cancer Res 2000; 60:184-90. (Ref 15.) Kimura W, Futakawa N, Yamagata S, et al. Different clinicopathologic findings in two histologic types of carcinoma of papilla of Vater. Jpn J Cancer Res 1994; 85:161-6. (Ref 86.) Levy C, Lymp J, Angulo P, et al. The value of serum CA 19-9 in predicting cholangiocarcinomas in patients with primary sclerosing cholangitis. Dig Dis Sci 2005; 50:1734-40. (Ref 23.) Moreno Luna LE, Kipp B, Halling KC, et al. Advanced cytologic techniques for the detection of malignant pancreatobiliary strictures. Gastroenterology 2006; 131:1064-72. (Ref 27.) O’Connell JB, Maggard MA, Manunga J, et al. Survival after resection of ampullary carcinoma: A national population-based study. Ann Surg Oncol 2008; 15:1820-7. (Ref 76.) Petrowsky H, Wildbrett P, Husarik DB, et al. Impact of integrated positron emission tomography and computed tomography on staging and management of gallbladder cancer and cholangiocarcinoma. J Hepatol 2006; 45:43-50. (Ref 26.) van der Gaag NA, ten Kate FJ, Lagarde SM, et al. Prognostic significance of extracapsular lymph node involvement in patients with adeno­ carcinoma of the ampulla of Vater. Br J Surg 2008; 95:735-43. (Ref 94.) Yoon JH, Canbay AE, Werneburg NW, et al. Oxysterols induce cyclooxygenase-2 expression in cholangiocytes: Implications for biliary tract carcinogenesis. Hepatology 2004; 39:732-8. (Ref 19.) Yoon JH, Gwak GY, Lee HS, et al. Enhanced epidermal growth factor receptor activation in human cholangiocarcinoma cells. J Hepatol 2004; 41:808-14. (Ref 18.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

70 Endoscopic and Radiologic Treatment of Biliary Disease Andrew H. Stockland and Todd H. Baron

CHAPTER OUTLINE Imaging of the Biliary Tract  1185 Ultrasonography  1185 Magnetic Resonance Cholangiopancreatography and Multidetector Computed Tomography Cholangiography  1185 Percutaneous Transhepatic Cholangiography  1186 Technique  1186 Postoperative Biliary Strictures  1188 Primary Sclerosing Cholangitis  1188 Bile Leaks  1188 Bile Duct Injury  1188 Bile Duct Stones  1189 Malignant Biliary Obstruction  1189 Percutaneous Cholecystostomy Tube Placement  1191

Endoscopic therapy and radiologic treatment of biliary disease have evolved in separate but parallel manners. Endoscopic therapy is performed using endoscopic retrograde cholangiopancreatography (ERCP) and, more recently, using endoscopic ultrasound (EUS)-guided techniques. ERCP is performed primarily by endoscopists trained in a gastroenterology training program, but in some centers it is performed by surgeons. ERCP is one of the most technically demanding endoscopic procedures, and for the successful management of complex cases, the learning curve is steep. Radiologic therapy of the biliary tree is performed via a percutaneous approach by interventional radiologists. The two approaches should be seen as complementary rather than competitive. The decision to proceed with an endoscopic or radiologic approach is often based on local expertise; other considerations include physician referral patterns, location of a lesion within the biliary tree, failure of one method, and altered anatomy as a result of prior surgery.

IMAGING OF THE BILIARY TRACT Imaging of the biliary tree is of utmost importance in planning the management approach to patients with biliary disorders and is discussed briefly in this context.

ULTRASONOGRAPHY

Noninvasive imaging of the biliary tree frequently begins with transabdominal ultrasound, which provides a global

Endoscopic Retrograde Cholangiopancreatography  1191 Bile Duct Stones  1191 Bile Leaks  1191 Primary Sclerosing Cholangitis  1192 Benign Biliary Strictures  1193 Indeterminate Biliary Strictures  1193 Malignant Biliary Strictures  1193 Sphincter of Oddi Dysfunction  1196 Complications  1197 Combined Percutaneous and Endoscopic Approaches  1197 Endoscopic Ultrasonography  1197 Diagnostic Role  1197 Therapeutic Role  1197

picture of the liver and is nearly universally available. There is no radiation exposure, and contrast agents are not required. Intrahepatic ductal dilatation can be visualized easily and the size of the bile duct can be documented. Ultrasound also provides imaging of the gallbladder and detects gallstones. For detection of choledocholithiasis, ultrasound has a high specificity, but the sensitivity does not exceed 68% and is often lower than 50%.1,2 The sensitivity decreases if the stones are small and the bile ducts are not dilated. Ultrasound is highly accurate (78% to 98%) for detecting extrahepatic biliary obstruction.2 When used in conjunction with the clinical evaluation, ultrasound allows differentiation between liver parenchymal disease and extrahepatic biliary obstruction with a reasonable sensi­ tivity and high specificity.2 Ultrasound is less accurate, however, at defining the level and cause of obstruction, with accuracy rates ranging from 27% to 95% and 23% to 88%, respectively.2 In addition, ultrasound is limited in the ability to distinguish malignant from benign causes of obstruction.2

MAGNETIC RESONANCE CHOLANGIOPANCREATOGRAPHY AND MULTIDETECTOR COMPUTED TOMOGRAPHY CHOLANGIOGRAPHY

Magnetic resonance cholangiopancreatography (MRCP) is a magnetic resonance imaging (MRI) study (and thus noninvasive) that is dependent on the high T2-signal charac­ teristics of bile. It does not require administration of oral or

1185

1186

Section VIII  Biliary Tract

Figure 70-1.  Choledocholithiasis on magnetic reso­ nance cholangiopancreatography (MRCP) and endo­ scopic retrograde cholangiopancreatography (ERCP). A, MRCP showing filling defect in distal bile duct (arrow). B, Corresponding ERCP with same filling defect.

A

intravenous contrast material. For the detection of choledocholithiasis, MRCP has a sensitivity ranging from 81% to 100%, a specificity ranging from 96% to 100%, and high overall diagnostic accuracy (Fig. 70-1).3 In addition, MRCP is highly accurate in demonstrating the presence of benign and malignant strictures4 and allows a thorough evaluation of the intrahepatic bile ducts. In patients suspected of having post-liver transplant biliary complications, intravenous administration of mangafodipir trisodium (Teslascan, Amersham Health, Princeton, NJ) may be used. This agent is excreted primarily in the bile and may improve imaging sensitivity for post-liver transplant biliary leaks and strictures.5 An MRI can be performed as well with an intravenous contrast agent, such as gadodiamide (Omniscan, GE Healthcare, United Kingdom) or gadopentetate dimeglumine (Magnevist, Bayer Healthcare, Leverkusen, Germany or Multihance, Bracco, Princeton, NJ), to detect and characterize mass lesions in the liver, porta hepatis, or pancreas. Contraindications to MRI include a cardiac pacemaker, automatic implantable cardioverter defibrillator, and some types of cerebral aneurysm clips. A particular concern about gadolinium-based intravenous contrast agents is that they may precipitate nephrogenic systemic fibrosis, a rare scleroderma-like disease manifested by hardening of the skin and fibrotic changes that affect multiple organs. The cause remains unclear, but reports suggest that patients with preexisting kidney disease (renal failure) are at greatest risk.6,7 Multidetector computed tomography cholangiography (MDCT) with multiplanar reformation is a computed tomography (CT)-based imaging study. MDCT is a combination of rapid volume acquisition and thin-slice imaging. Water is used as an oral contrast agent for the biliary tree, and intravenous iodinated contrast is also administered. Images acquired in the axial plane can be reconstructed sagittally or coronally and reformatted three dimensionally. The intravenous contrast dye is not excreted in bile but enhances adjacent surrounding visceral structures such as the liver,

B

pancreas, and other soft tissues. Bile ducts thus appear as low attenuation structures that are best visualized if dilated. The sensitivity and specificity of MDCT for bile duct strictures have been reported to be 85.7% and 100%, respectively.8 MDCT also has a high sensitivity and specificity for the detection of bile duct stones.9

PERCUTANEOUS TRANSHEPATIC CHOLANGIOGRAPHY Percutaneous transhepatic cholangiography (THC) is an invasive diagnostic test and can be therapeutic if necessary. In light of the array of noninvasive imaging studies, percutaneous THC is rarely performed purely for diagnostic purposes. Subsequent decompression of biliary obstruction, removal of a stone, balloon dilation of a stricture, and placement of a stent for a stricture can be performed. The procedure is generally reserved for patients for whom ERCP is precluded because of difficult endoscopic access across a biliary-enteric (Roux-en-Y) anastomosis, gastric bypass, or extrahepatic biliary stricture that cannot be traversed endoscopically. Serious procedure-related complications such as bleeding, sepsis, or bile leakage occur in approximately 2% to 4% of cases.10 The procedure generally can be performed with monitored moderate (“conscious”) sedation.11 Broadspectrum intravenous antibiotics are usually administered prophylactically.

TECHNIQUE

Review of CT and MR imaging of the liver prior to percutaneous THC can help determine the best approach (i.e., from the right or left side) and the location of the dominant dilated ducts and help avoid traversing adjacent structures, such as the colon, unintentionally. Dilated bile ducts on the left side may be easily accessible with a minimal number of needle passes with use of ultrasound guidance from a

Chapter 70  Endoscopic and Radiologic Treatment of Biliary Disease subxiphoid approach.12 A standard right-sided approach is used most frequently, however, and is performed from an intracostal approach, usually via the mid-axillary region below the 10th intercostal space. Higher punctures increase the risk of pneumothorax or biliary pleural effusion. From either side, the procedure is initiated by advancing a 22-gauge needle under fluoroscopic guidance centrally toward the liver hilum and gently injecting contrast as the needle is withdrawn slowly. The initial use of such a small needle reduces hepatic trauma as well as the likelihood of bleeding despite the potential need for multiple needle passes to cannulate a bile duct, particularly when the bile ducts are not dilated. Ultrasound guidance and CT guidance can be used to access nondilated bile ducts.13,14 When a bile duct is cannulated, a diagnostic cholangiogram can be performed. Isolated ducts, because of strictures or stones that do not communicate with the rest of the biliary tree, may need to be opacified via additional needle passes. If the procedure is only diagnostic, and biliary obstruction is not evident, the needle is simply withdrawn. If the biliary system is obstructed, however, serious consideration should be given to traversing the obstruction and leaving a decompressive “external-internal” tube in place; abandoning an obstructed biliary system may lead to bile leakage from the puncture site. The risk of hepatic arterial injury is reduced by using a peripheral intrahepatic bile duct for final access. If the duct cannulated initially is too central (the larger branches of the hepatic artery tend to be more central), a more peripheral duct should be chosen for access into the biliary tree. Frequently, use of a second needle to puncture a more peripheral duct is required, and the initial needle is used to opacify and visualize this new and safer access duct. A 0.018-inch “micro” wire is then advanced via the needle into the biliary tree and the access system “upsized” by the passage of catheters of increasing diameter over the wire. When access is gained, the obstruction can be traversed and an external-internal biliary tube can be placed (Fig. 70-2). These tubes provide drainage

A

C

holes positioned above the level of obstruction; the distal pigtail is configured within the small intestine. The size of the tube usually ranges from 8 to 12 French. A larger tube may yield better decompression, but care must be taken not to place a tube in which the size may actually obstruct drainage of smaller ducts, particularly in the setting of primary sclerosing cholangitis (PSC), in which many of the obstructed ducts are not dilated. If the obstruction cannot be traversed during the initial attempt, a drainage catheter can be left proximal to the obstruction in the biliary tree (external drainage), and subsequent attempts can be made via this access after several days of drainage. This delay often allows inflammation to decrease and increases the likelihood of subsequent internalization of a catheter. Generally, the external-internal drainage tube is left to external drainage until fever or blood in the biliary tree resolves. Capping of the external end of the tube to permit internal drainage only decreases biliary fluid losses, which can be more than 1 liter per day, and prevents associated dehydration or electrolyte abnormalities. Bile samples obtained during the initial procedure can be sent for culture or cytology. Contraindications to percutaneous THC include coagulopathy. Generally the procedure is thought to be safe with an international normalized ratio (INR) of less than 1.8 and platelet count greater than 50,000/mm3. Any abnormalities should be corrected immediately before the procedure. Marked ascites between the liver and puncture site increases the risk of bile leakage, whereas a tortuous biliary catheter course may lead to malposition of the catheter or difficulty with future manipulations. In the presence of a substantial amount of perihepatic ascites, a pre-procedure paracentesis can be performed or ultrasound guidance can be used to place a small temporary peritoneal drainage catheter adjacent to the liver for the duration of the procedure. Biliary sepsis can be minimized by avoiding overdistention of the bile ducts and limiting the number of manipulations during the procedure. As soon as a tube is placed, it

B

D

Figure 70-2.  Schematic showing percutaneous trans­ hepatic cholangiography. A, A peripheral bile duct is identified and entered with a needle. B, A guidewire is passed through the needle across the obstructing lesion into the duodenum. C, The needle has been withdrawn. D, An internal-external catheter is inserted over the guidewire.

1187

1188

Section VIII  Biliary Tract can be used as an access for further manipulations or interventions. Following initial biliary decompression, further intervention should be avoided until fever and sepsis have resolved. Patients need to be monitored closely for the first 24 to 48 hours following the procedure. Brisk bleeding around the catheter site, through the catheter itself, or from the gastrointestinal tract suggests the possibility of hepatic arterial injury.15 Presentation of a hepatic artery pseudoaneurysm can be delayed, sometimes for a week or two after the initial procedure. If bleeding persists, the hemoglobin level drops substantially or the patient becomes hemodynamically unstable, hepatic angiography should be considered, and, if an injured arterial branch is demonstrated, embolization should be performed. A small amount of blood in the biliary tube or bile ducts following the original procedure, or during subsequent manipulations, is frequently self-limited and clears within one or two days.

POSTOPERATIVE BILIARY STRICTURES

Postoperative strictures may occur following laparoscopic cholecystectomy, major hepatic resection, and liver transplantation at a choledochocholedochal anastomosis or within an intrahepatic duct as a result of ischemia or recurrent PSC (Table 70-1) (see Chapters 66, 68, and 95). Dilation of a postoperative or other benign biliary stricture can be performed via percutaneous THC or through a mature, surgically-placed T-tube tract. Maturation of the T-tube tract usually requires six weeks. Percutaneous THC and biliary balloon dilation may be performed at the same session in the absence of clinical signs of cholangitis or sepsis. An 8-French or 10-French transhepatic tube is left in place, and the patient returns for repeat cholangiography six weeks later, at which time further stricture dilation is performed if bile duct narrowing of 30% or greater persists. The tube is then repeatedly upsized to a 12-French tube to facilitate healing of the stricture at a larger diameter. If the stricture resolves on follow-up, the biliary tube can be removed; otherwise, a similar procedure should be performed after six to eight weeks. In one of the largest series published with long-term follow-up, percutaneous biliary balloon dilation was performed in 85 patients with a benign biliary stricture.16 In the 75 patients with follow-up, 205 percutaneous procedures were performed during 112 treatments of 84 biliary strictures. Stricture balloon dilation from 8 to 12 mm was performed. Procedures were repeated at 2- to 14-day intervals until cholangiography demonstrated free drainage of contrast material to the small intestine and no residual stenosis. An internal-external biliary drain was left in place for a mean of 14 to 22 days and removed if the patient did well when the catheter was clamped and had a normal cholangiogram. All procedures were technically successful. A total of 52, 11, 10, and 2 patients underwent a total of one, two, three, and four dilations, respectively. Major complications occurred in 2% of procedures: two subphrenic abscesses,

Table 70-1  Principal Causes of Benign Biliary Strictures Chronic pancreatitis Postoperative Bile duct resection with choledochocholedochal anastomosis Laparoscopic cholecystectomy Liver transplantation Primary sclerosing cholangitis

one hepatic arterial pseudoaneurysm, and one case of hemobilia. The probability that clinically significant restenosis did not develop at 5, 10, 15, 20, and 25 years was 0.52, 0.49, 0.49, 0.41, and 0.41, respectively, after the first treatment, and 0.43, 0.30, 0.20, 0.20, and 0.20, respectively, after the second treatment. No significant difference was found in the rate of restenosis for strictures at anastomotic and nonanastomotic sites. Overall, 56 of 75 patients (75%) had successful management with percutaneous therapy. Following liver transplantation, percutaneous THC is used for treating complications in patients with a ductto-duct anastomosis and especially in patients in whom hepaticojejunostomy has been performed; these latter anastomoses frequently cannot be accessed via an endoscopic approach. Hepaticojejunal anastomosis is performed at the time of liver transplantation in children, persons with PSC, persons who undergo reoperation for a complication of a duct-to-duct (choledococholedochal) anastomosis, and living-related donors. For treatment of both duct-to-duct and hepaticojejunal anastomotic strictures, percutaneous therapy provides a high nonoperative success rate.17-19 In addition, in those patients in whom the bile duct is approachable via ERCP but who fail an endoscopic approach, a percutaneous approach is often successful.20

PRIMARY SCLEROSING CHOLANGITIS

Most nonsurgical therapeutic interventions for PSC are now performed via ERCP (see Chapter 68). In the past, percutaneous therapy for a dominant stricture using balloon dilation followed by biliary drain placement for two to three months was found to be highly effective for treating obstructive biliary symptoms in patients with PSC21 but less effective in patients with jaundice for more than six months because of liver parenchymal dysfunction. More recently, only case reports of percutaneous therapy for PSC have appeared in the literature.22 In our experience, percutaneous therapy is useful for patients with a dominant stricture that cannot be accessed endoscopically (Fig. 70-3). In these cases, a guidewire or catheter passed percutaneously can be left in the duodenum to facilitate future endoscopic access (see later).

BILE LEAKS

Bile leaks are almost always postsurgical in etiology and arise from anastomotic (e.g., post-liver transplant) and nonanastomotic sites. The latter include cut surfaces of the liver and bile ducts following hepatectomy and laparoscopic injury. Percutaneous management may include drainage of free bile from the peritoneal cavity and of localized bile collections (bilomas) as well as placement of a biliary catheter above or across the leaking site to allow successful closure in the majority of cases.23-25

BILE DUCT INJURY

Widespread performance of laparoscopic cholecystectomy has led to an increased frequency of major bile duct injuries (see Chapter 66). Other causes of bile duct injury include bile duct exploration or biliary injury resulting from abdominal surgery or trauma. Percutaneous transhepatic biliary drain placement can be used as primary treatment of the injury or to augment surgical repair. Misra and colleagues26 retrospectively evaluated 51 patients who underwent percutaneous biliary management following laparoscopic cholecystectomy-related bile duct injuries over a 10-year period; 45 had operative repair prior to referral. Overall, 46 of the 51 were initially managed percutaneously, and 5 were managed percutaneously following failed hepaticojejunostomy. Nonoperative percutaneous management with balloon

Chapter 70  Endoscopic and Radiologic Treatment of Biliary Disease

A

B

C Figure 70-3.  Percutaneous therapy for recurrent primary sclerosing cholangitis after liver transplantation. The patient has a hepaticojejunostomy. A, Initial puncture and passage of a guidewire into the jejunum. B, Balloon dilation of right and left hepatic ducts. C, Follow-up cholangiogram with indwelling internal-external biliary catheter.

dilation resulted in an overall success rate of 58.8% at a mean follow-up of 76 months.

BILE DUCT STONES

Bile duct stones can be managed percutaneously via cholecystostomy tubes, percutaneous placed drains, or surgical T-tubes. Gallbladder tube or T-tube tracts require approximately six weeks to mature prior to use. In many cases, bile duct stones can be cleared percutaneously by dilating the papilla from an antegrade approach.27,28 The stones, which may require mechanical fragmentation, are flushed into the duodenum. The percutaneous catheter is replaced for several days and then removed. With this approach, in one study27 stones were removed in 95 of 100 patients. In some cases, particularly in the setting of complex intrahepatic stones, a small-caliber choledochoscope (cholangioscope) can be passed through a mature percutaneous tract. Stones are then fragmented using a variety of techniques, with a high rate of success (see Chapter 65).29,30

MALIGNANT BILIARY OBSTRUCTION

Stents can be placed percutaneously for relief of malignant biliary obstruction, either preoperatively or for palliation. Stents are composed of either rigid plastic or self-expanding metal. Self-expanding metal stents (SEMS) were designed to avoid occlusion from bacterial biofilm, which invariably occurs in plastic stents and results in the need for re-intervention. Distal bile duct strictures (e.g., caused by pancreatic head cancer) are preferably managed via ERCP (see later) because endoscopic stent placement is less painful than percutaneous stent placement and is associated with fewer compli­ cations. This conclusion is based on a randomized trial of endoscopic and percutaneous approaches using plastic stents.31 Percutaneous stent placement can easily be achieved in these patients, however. Multiple interventions are often needed to place plastic stents prior to final internalization of the stent because large-bore (≥10-French) stents require dilation of a tract through the liver, which

1189

1190

Section VIII  Biliary Tract

A

B

C Figure 70-4.  Percutaneous management of calculous cholecystitis. A, Percutaneous catheter is placed into the gallbladder. Note multiple gallstones. B, A large-caliber tube allows extraction of stones. C, Complete clearance of gallstones is demonstrated.

often cannot be accomplished in one stage. Bleeding, which occurs with such aggressive dilation, often requires main­ tenance of an external catheter to drain blood within the biliary tree. More recently, SEMS have been used. In a classic study, percutaneous placement of SEMS was asso­ ciated with significantly longer stent patency, reduction in the need for re-intervention, and shorter hospital stays.32 In addition, advantages of SEMS when placed percutaneously are the availability of small-diameter pre-deployment delivery systems, so that the percutaneous tract does not require dilation, and the capability for stent insertion in one step.32,33 In a randomized trial34 of endoscopic versus percutaneous palliation of malignant bile duct obstruction in which metal biliary stents were placed percutaneously in one step and plastic stents were placed endoscopically,

percutaneous placement of SEMS was associated with a 34% lower rate of recurrent biliary obstruction. Relief of hilar biliary obstruction (e.g., caused by hilar cholangiocarcinomas, or Klatskin tumors) is more difficult to achieve endoscopically than relief of distal bile duct obstruction. Several studies have suggested that the percutaneous approach to these tumors is superior to the endoscopic approach, with a lower rate of post-procedure cholangitis.34,35 Covered SEMS were designed to improve stent patency by reducing the frequency of occlusion resulting from tumor ingrowth and tissue hyperplasia. Studies have shown promising results,36,37 although no randomized trials of covered versus uncovered stents placed via the percutaneous approach have been published.

Chapter 70  Endoscopic and Radiologic Treatment of Biliary Disease PERCUTANEOUS CHOLECYSTOSTOMY TUBE PLACEMENT The standard treatment of acute calculous cholecystitis is cholecystectomy (see Chapters 65 and 66). Even with the advent of laparoscopic cholecystectomy, some patients are still not surgical candidates. Percutaneous cholecystostomy tube placement is a minimally invasive way to treat these patients and can be performed with a local anesthetic or with moderate sedation. Tube placement enables immediate decompression of the gallbladder. Bile samples obtained during tube placement can be used to guide antimicrobial therapy, and the tube can be used for cholangiography to confirm cystic duct obstruction or, if the cystic duct becomes patent, bile duct obstruction. Percutaneous gallbladder therapy is useful for the management of severe acute calculous cholecystitis as a nonoperative approach in elderly patients or persons who are poor candidates for surgery and as a way to avoid emergency surgery.38 In the last situation, an elective cholecystectomy can be performed subsequently, often laparoscopically.39,40 If the patient remains a poor surgical candidate, the cholecystostomy tube can remain in place long term. Alternatively, stones in the gallbladder, cystic duct, or bile duct can be managed percutaneously. The mature percutaneous tract can be dilated, after which the stones can be extracted (Fig. 70-4). Patients with intrahepatic gallbladders and small, shrunken, thick-walled gallbladders are not candidates for this approach. Despite the high success rate of percutaneous stone removal, stones can recur.41 An additional challenge in patients in an intensive care unit is the management of suspected acute acalculous cholecystitis (see Chapter 67). A Murphy sign can be difficult, if not impossible, to demonstrate, particularly in intubated or unresponsive patients. Delayed diagnosis and treatment can lead to gallbladder gangrene and perforation and to mortality. In patients in whom clinical suspicion for acute acalculous cholecystitis is high, a gallbladder tube should be placed percutaneously. If the gallbladder is not the source of the patient’s clinical problem, the cholecystostomy tube remains in place for six weeks. In one study of 55 critically ill patients with suspected acute acalculous cholecystitis who underwent percutaneous gallbladder tube placement, clinical improvement was seen in 58.7% within 24 hours and 95.7% within 72 hours.42

ENDOSCOPIC RETROGRADE CHOLANGIOPANCREATOGRAPHY Endoscopic retrograde cholangiopancreatography has evolved from a purely diagnostic to an almost exclusively therapeutic procedure. ERCP is commonly performed using moderate sedation,43 although in severely ill patients and in cases anticipated to be complex, an anesthesiologist is often needed. ERCP is performed with a side-viewing duodenoscope that allows identification of the major papilla. The bile duct is cannulated under endoscopic and fluoroscopic guidance. A variety of catheters, guidewires, and stents are available to allow therapeutic interventions to be performed. Diagnostic ERCP is still used for facilitating manometry in patients with suspected sphincter of Oddi dysfunction (see Chapter 63) and for establishing the diagnosis of PSC when other imaging techniques have been nondiagnostic (see Chapter 68).44 A variety of biliary indications for ERCP45 will each be discussed.

BILE DUCT STONES

ERCP is usually performed in patients with known choledocholithiasis or in those patients with at least a moderate clinical suspicion of choledocholithiasis (see Chapter 65). In patients with gallbladder stones and a low clinical suspicion of choledocholithiasis, noninvasive imaging studies (MRCP, MDCT) or EUS are preferred to minimize the potential for complications of ERCP.46 In patients with a low clinical suspicion of choledocholithiasis in whom cholecystectomy is planned, intraoperative cholangiography can be performed, and, if stones are identified, laparoscopic exploration and stone removal can be undertaken. ERCP can then be reserved for patients in whom the stones are not extracted.46 The standard method for stone removal is endoscopic biliary sphincterotomy to allow enlargement of the papilla and subsequent extraction of stones with a balloon or basket (Fig. 70-5). With this approach, more than 80% of all stones can be removed successfully.47 Larger stones may require additional removal techniques (discussed later). An alternative to biliary sphincterotomy is balloon dilation of the papilla (balloon sphincteroplasty), which can be performed using small-diameter balloons (4 to 8 mm). The technique was introduced as a way to preserve sphincter of Oddi function, especially in young patients. The stones are removed using balloon or basket techniques. Most of the literature on balloon sphincteroplasty comes from outside the United States. Two meta-analyses of randomized trials of balloon sphincteroplasty versus sphincterotomy have shown that the rates of pancreatitis and need for mechanical lithotripsy are significantly higher, but the risk of bleeding is significantly lower, with balloon sphincteroplasty than with sphincterotomy.48,49 In the United States, the only randomized trial that compared balloon sphincteroplasty and sphincterotomy was closed prematurely because of two deaths in young patients from post-ERCP pancreatitis after sphincteroplasty.50 Sphincteroplasty still remains, however, an alternative approach in patients with coagulopathy,48 persons with underlying cirrhosis (particularly Child’s class C48 [see Chapter 90]), and those with altered anatomy (e.g., Billroth II gastrojejunostomy [see Chapter 53]), in which sphincterotomy is technically difficult.48 Removal of large bile duct stones (defined arbitrarily as ≥1.5 cm in diameter) may require additional techniques than those described earlier to be removed successfully. One such technique is lithotripsy. One form of lithotripsy is mechanical lithotripsy, in which the stone is captured in a specialized large basket and crushed51 (Fig. 70-6). The fragments are removed using standard extraction techniques. Another form of lithotripsy is intraductal lithotripsy, which is performed by passing laser or electrohydraulic catheters into the bile duct. The stones are fragmented under direct endoscopic visualization using a transpapillary choledochoscope.52 Direct visualization is necessary to ensure that the lithotripsy device is directed at the stone and not the bile duct wall. More recently, the combination of biliary sphincterotomy and large-diameter (>12 mm) balloon dilation has been used to remove large stones and decrease the need for mechanical lithotripsy.53 This large-diameter dilation method appears to be safe and not associated with an increased risk of post-ERCP pancreatitis.54 If large stones cannot be removed, a biliary stent is placed to relieve the obstruction.55 Additional procedures can then be undertaken electively to remove residual stones.

BILE LEAKS

As discussed previously, bile leaks arise as a result of postsurgical complications and trauma. Most commonly, post-

1191

1192

Section VIII  Biliary Tract

A

B

Figure 70-5.  Endoscopic images during removal of a bile duct stone. A, Bulging papilla consistent with an impacted stone is seen. B, After endoscopic sphincterotomy, the stone is extracted.

A

B

C

Figure 70-6.  Mechanical lithotripsy of a bile duct stone. A, Endoscopic retrograde cholangiogram showing a large stone (arrow). B, Lithotripsy basket crushing the stone. C, Follow-up cholangiogram shows clearance of the stone from the bile duct.

cholecystectomy leaks arise from either the cystic duct or duct of Luschka (see Chapter 62). These smaller leaks can usually be managed with biliary sphincterotomy alone or placement of a small-caliber (7-French) plastic biliary stent (or both).56 This approach diverts bile away from the leak into the duodenum and negates the effect of the otherwise high-pressure biliary sphincter. More complex leaks usually require placement of one or more large-caliber plastic

biliary stents in combination with biliary sphincterotomy (Fig. 70-7).57 The use of removable, covered SEMS for treatment of refractory leaks has also been described.58

PRIMARY SCLEROSING CHOLANGITIS

Patients with PSC may benefit from endoscopic intervention to treat a dominant stricture or biliary lithiasis.59 Patients with a dominant stricture usually present with progressive

Chapter 70  Endoscopic and Radiologic Treatment of Biliary Disease large diameter (10 mm) of the SEMS results in dilation of the stricture over time. The stent is removed after an interval of three to six months. Results using this approach have been encouraging, although these devices are not yet approved for use in benign diseases.

INDETERMINATE BILIARY STRICTURES

A

Some biliary strictures cannot be readily classified as benign or malignant on the basis of imaging studies and tissue sampling. Tissue sampling techniques at ERCP consist of wire-guided biliary brush cytology and intraductal forceps biopsy.67 Additional techniques that can be used to assess indeterminate strictures include intraductal ultrasonography68 and direct choledochoscopy with or without directed biopsy.69 In a small percentage of patients, the diagnosis still remains unclear. In some patients, the final diagnosis can only be established during long-term follow-up or at surgical exploration and resection.

MALIGNANT BILIARY STRICTURES

Endoscopic relief of malignant biliary obstruction is achieved by placement of large-bore plastic stents or SEMS across the malignant stricture. The approach to the patient depends on whether the stricture is distal to the bifurcation of the common hepatic duct or involves the bifurcation (hilar obstruction).

Distal Bile Duct Strictures

B Figure 70-7.  Intrahepatic bile leak treated endoscopically after a hepatec­ tomy. A, Contrast extravasation (long arrow) is seen near the percutaneous drain; an internal biliary stent (short arrows) is placed. B, Follow-up chol­ angiogram shows resolution of the leak.

biliary obstruction. Cholangiocarcinoma must be considered in these patients. Routine brush cytology has a low sensitivity in these patients, but fluorescence in situ hybri­ dization (FISH) has been shown to have a high sensitivity for the detection of cholangiocarcinoma (see Chapter 69).60 Choledochoscopy also may improve detection of malignancy in these patients.61 Endoscopic treatment of a dominant stricture involves balloon dilation, often in combination with short-term (<8 weeks) placement of a large-bore (10-French) stent (Fig. 70-8). Endpoints following endoscopic therapy have included clinical, biochemical, and radiologic improvement, with success rates ranging from 65% to 100%.62

BENIGN BILIARY STRICTURES

Benign biliary strictures are caused by a variety of disorders (see Table 70-1), and the response to therapy varies with the cause. Endoscopic therapy consists of balloon dilation followed by placement of plastic biliary stents. Data in the 2000s suggest that, for most causes of benign strictures, placement of multiple side-by-side large-bore plastic stents over the course of several endoscopic sessions (Fig. 70-9), with stents remaining in place for up to one year, allows a higher rate of successful stricture resolution than placement of only one or two stents.63,64 Chronic pancreatitis produces distal bile duct strictures that are usually refractory to endoscopic therapy with a single plastic stent, particularly in patients with calcific chronic pancreatitis.65 Multiple plastic stents can be placed. Covered SEMS also have been used for the treatment of chronic pancreatitis–induced bile duct strictures.66 The

Pancreatic head cancer is the most common cause of distal bile duct obstruction (see Chapters 60 and 61). In patients with known pancreatic cancer in whom surgical resection (pancreaticoduodenectomy) is planned, routine preoperative ERCP for biliary decompression is discouraged. Several studies have shown that this approach does not improve surgical outcome and may cause postoperative morbidity and that complications from ERCP may delay or prevent surgical resection.70 The indications for preoperative ERCP include acute cholangitis and severe pruritus.71 In addition, stent placement is indicated when neoadjuvant chemoradiation is administered because the time to surgical resection is usually prolonged. In this situation, the use of a short-length SEMS (covered or uncovered) appears to be the best option; one study showed a high rate of stent occlusion in patients with pancreatic cancer who underwent preoperative chemoradiation and in whom plastic stents had been placed compared with those in whom a SEMS was placed.72 Uncovered SEMS are removed along with the tumor at the time of surgical resection. Indeed, in one study, a covered SEMS was placed in all patients with pancreatic cancer regardless of resectability, and this approach was found to be cost-effective when compared with placement of plastic stents.73 ERCP with biliary stent placement has been shown in randomized trials to be an acceptable alternative to palliative surgical bypass.74 Biliary stents can be placed safely in an outpatient setting.75 The comparative studies of endoscopy and surgery for palliation of distal biliary obstruction were performed using plastic stents, prior to the advent of SEMS. The main limitation to plastic stent placement is stent occlusion as a result of bacterial biofilm or reflux of vegetable matter.76 Therefore, in the comparative trials of surgery and endoscopy, the shorter length of initial hospital stay in the endoscopy group was offset by the need for subsequent hospitalizations and need for repeat ERCP to manage stent occlusion. The median time until stent occlusion for a standard large-bore stent is approximately three months. Stent occlusion results in recurrent jaundice, usually with cholangitis. SEMS have overcome the problem

1193

1194

Section VIII  Biliary Tract

A

D

B

C

E

Figure 70-8.  Endoscopic therapy of primary sclerosing cholangitis. A, Cholangiogram demonstrates irregular strictures of the right and left hepatic ducts with intrahepatic stones. B, Balloon dilation of the left hepatic duct. C, Balloon dilation of the right hepatic duct. D, Bilateral plastic biliary stents are placed. E, Follow-up cholangiogram with balloon occlusion shows marked improvement.

of bacterial biofilm, and randomized, controlled trials have shown superior patency rates for SEMS when compared with plastic stents (Fig. 70-10).77 Because the cost of SEMS is much greater than that of plastic stents, placement of a SEMS is cost-effective only if the patient survives longer than three to six months. Therefore, projected life expectancy should be taken into consideration when choosing between plastic and metal stents.78 Other factors to be considered include the patient’s adherence and ability to return for care.77 Uncovered SEMS occlusion is generally managed easily with placement of a plastic stent or a new SEMS within the existing one.79 More recently, covered metal stents have been developed in an attempt to overcome

occlusion caused by tumor overgrowth and tissue hyperplasia. Early comparative studies have demonstrated prolonged patency with covered SEMS compared with uncovered SEMS80, although this advantage has not been firmly established.81 Covered SEMS are associated with higher migration rates82 and possibly an increased risk of acute cholecystitis.83

Hilar Biliary Obstruction

Hilar strictures may be caused by cholangiocarcinoma or metastatic disease. The clinical success rates for achieving adequate palliation for hilar tumors is less than that for distal bile duct tumors.84 Furthermore, technical success

Chapter 70  Endoscopic and Radiologic Treatment of Biliary Disease

A

C

B

D

rates for bilateral endoscopic stent placement (right and left hepatic ducts) are also lower. Most patients with hilar obstruction will be adequately palliated when only one side of the liver is drained (unilateral drainage)—in other words, when only one side has been accessed and therefore contaminated.85 Patients who have had contrast instilled in both the left and right biliary systems require stenting of both to prevent progressive cholangitis.86 Not as firmly established is that metal stents offer superior prolongation of palliation compared with plastic stents for hilar tumors, as is the case for distal strictures. In a prospective, single-

Figure 70-9.  Endoscopic treatment of choledochochole­ dochal anastomotic stricture with multiple stents. A, Chol­ angiogram of stricture (arrow) at the anastomosis. B, Balloon dilation of the stricture. C, Multiple plastic stents are placed. D, Follow-up cholangiogram showing improvement.

arm, pilot study of metal stent placement in 17 patients with Bismuth type II and III hilar cholangiocarcinoma (see Chapter 69), median stent patency was 12 months.87 A noncomparative, single-arm study showed that insertion of a Wallstent (a type of metal stent) is safe and feasible and achieves successful palliation without the need for further biliary re-intervention in the majority (69%) of patients with nonresectable hilar cholangiocarcinoma.88 In a more recent prospective, observational, cohort study of patients with hilar tumors treated with plastic or metal stents, patients in whom a metal stent was placed had significantly lower rates

1195

1196

Section VIII  Biliary Tract

Figure 70-10.  Endoscopic palliation of malignant obstructive jaundice caused by unresectable pancreatic cancer. A, Chol­ angiogram showing obstructive stricture in distal bile duct. B, Cholangiogram immediately after deployment of a selfexpandable metal stent.

A

of post-procedural complications and need for percutaneous drainage.89 This finding suggests that metal stents may offer the same benefits for hilar biliary obstruction as they do for distal biliary obstruction. Few comparative studies of endoscopic and percutaneous approaches to hilar tumors have been published. One study of the outcome following endoscopic and percutaneous treatment of hilar cholangiocarcinoma with use of only SEMS found that the rate of successful biliary decompression was significantly higher in the group in whom a SEMS was placed percutaneously than in the group in whom it was placed endoscopically.90 The median survival of patients in whom biliary drainage was successful initially, regardless of which procedure was performed, was much longer than that of patients in whom biliary drainage was not achieved. In addition, after successful biliary decompression had been achieved, the durations of median survival and stent patency were similar in the two groups. In summary, in patients with unresectable hilar cholangiocarcinoma who undergo ERCP for palliation, unilateral stent placement should be performed with contrast injection confined to one side. A pre-ERCP abdominal CT may reveal atrophy of one lobe of the liver, and this lobe should specifically be avoided at ERCP because contamination will require drainage to prevent cholangitis but will not likely aid palliation. Alternatively, MRI can define the biliary anatomy prior to ERCP. One stent—plastic or metal—is usually adequate to achieve palliation.91 SEMS appear to offer superior palliation. Finally, from an endoscopic perspective, achieving successful drainage is more difficult technically for hilar tumors than for nonhilar tumors. Photodynamic therapy (PDT) has been used for palliation of patients with unresectable hilar cholangiocarcinoma in whom jaundice does not resolve after endoscopic placement of plastic stents.92 Significant improvements in cholestasis,

B

quality of life, and survival (as compared with historical controls) has been demonstrated with PDT93 and can be maintained for an extended period.94 The relevant studies were performed outside the United States, in countries in which smaller, more flexible laser fibers are available. At the present time, passage of laser fibers used for the treatment of esophageal cancer into the biliary system is difficult, but feasible. Nonetheless, PDT trials in the United States have shown promising results.95-97 PDT for cholangiocarcinoma is generally confined to selected centers with expertise in this procedure.

SPHINCTER OF ODDI DYSFUNCTION

Biliary sphincter of Oddi dysfunction (SOD) is classified as types I, II, and III (see Chapter 63).98,99 All types are characterized by intermittent biliary-type abdominal pain. Sphincter of Oddi manometry (SOM) is traditionally performed during ERCP by passing a water-perfused catheter into the bile duct or pancreatic duct to measure the biliary or pancreatic sphincter pressure, respectively. More recently, solid-state catheters have been used. SOD type I is associated with objective abnormalities in laboratory test results, often during attacks, and an abnormally dilated extrahepatic bile duct on an imaging study. SOD type II is associated with either abnormal laboratory test results or a dilated bile duct. SOD type III is characterized by pain alone. Patients with SOD type I invariably respond to endoscopic biliary sphincterotomy and do not require SOM for diagnosis.98 Patients with SOD type II who have a high basal sphincter pressure on SOM generally respond to biliary sphincterotomy.99 Patients with SOD type III do not appear to respond to biliary sphincterotomy at a greater rate than they do to placebo, and the role of ERCP in these patients is unclear. Indeed, one study has suggested that postcholecystectomy pain may be explained by persistent hyperexcitability of the nociceptive neurons in the central nervous system and may

Chapter 70  Endoscopic and Radiologic Treatment of Biliary Disease be unrelated to objective motility disorders of the sphincter of Oddi.100

COMPLICATIONS

Five major types of complications of ERCP may occur: sedation-related, pancreatitis, bleeding, perforation, and infection. Rates of post-ERCP pancreatitis vary because of differences in patient selection and operator technique and experience. Patients at highest risk are young, otherwise healthy women, especially those with known or suspected SOD.101 Elderly patients and those with chronic pancreatitis or pancreatic cancer have lower rates of pancreatitis. Prophylactic placement of a stent into the main pancreatic duct reduces the risk of pancreatitis in high-risk patients and nearly eliminates the risk of severe pancreatitis.102 Bleeding may occur after biliary sphincterotomy. Risk factors for post-sphincterotomy bleeding include coagulopathy and institution of anticoagulation within 72 hours of the sphincterotomy.103 Perforation of the duodenum occurs in less than 1% of patients and may require surgical management.104 Infection occurs primarily in patients in whom drainage of the biliary tree after ERCP is inadequate. Such patients include those with extensive intrahepatic PSC or advanced hilar tumors and those who have undergone failed stent placement for biliary obstruction. Data are accruing that a lower ERCP volume by an endoscopist is associated with a lower success rate and higher complication rate (see also Chapter 40).105

COMBINED PERCUTANEOUS AND ENDOSCOPIC APPROACHES In some situations in which ERCP is unsuccessful but the bile duct still needs to be accessed endoscopically, a combined percutaneous-endoscopic approach can be undertaken, a so-called rendezvous procedure106 (Fig. 70-11). An example is a patient with a large duodenal diverticulum and a bile duct stone. If the diverticulum prevents endoscopic biliary access, a guidewire is passed percutaneously into the duodenum; the patient is then brought to the ERCP suite, and an ERCP is repeated. The wire is grasped by a forceps, and accessories are passed over the wire, thereby allowing

Figure 70-11.  Schematic of combined percutaneous and endoscopic approach to the biliary tract. The guidewire is passed into the duodenum and identified endoscopically.

sphincterotomy and stone extraction. The technique is not needed for most malignant strictures, which can be managed entirely with a percutaneous approach if the endoscopic approach fails.

ENDOSCOPIC ULTRASONOGRAPHY DIAGNOSTIC ROLE

EUS has expanded the diagnostic role of endoscopy. EUS combines endoscopy and ultrasound to provide highresolution images of the biliary system and is highly accurate for determining the cause of extrahepatic biliary obstruction, with a sensitivity of 97% and a specificity of 88%.2 The sensitivity, specificity, and accuracy rates of EUS for the diagnosis of bile duct stones are 95%, 98%, and 96%, respectively.2 EUS has the ability to distinguish different causes of malignant obstruction. In particular, EUS is more sensitive (93% to 100%) than CT (53% to 77%), transabdominal ultrasound (50% to 67%), MRI (50% to 67%), and ERCP (90%) for the detection of pancreatic tumors (see Chapter 61).2 EUS is less invasive than ERCP and has no associated radiation or contrast exposure. EUS combined with fine-needle aspiration provides tissue diagnosis of masses and lymph nodes.

THERAPEUTIC ROLE

Like ERCP, EUS is evolving into a therapeutic procedure, but the current therapeutic applications of biliary EUS are limited.107 EUS-guided biliary drainage has been described in which a needle followed by catheters or guidewires is passed into the biliary system, either through the left lobe of the liver108-111 into the intrahepatic ducts or through the duodenal wall directly into the bile duct.112 Another use of EUS is the drainage of bilomas.113

KEY REFERENCES

Adler DG, Baron TH, Davila RE, et al. Standards of Practice Committee of American Society for Gastrointestinal Endoscopy. ASGE guideline: The role of ERCP in diseases of the biliary tract and the pancreas. Gastrointest Endosc 2005; 62:1-8. (Ref 45.) Aytekin C, Boyvat F, Harman A, et al. Percutaneous management of anastomotic bile leaks following liver transplantation. Diagn Interv Radiol 2007; 13:101-4. (Ref 24.) Cantwell CP, Pena CS, Gervais DA, et al. Thirty years’ experience with balloon dilation of benign postoperative biliary strictures: Long-term outcomes. Radiology 2008; 249:1050-7. (Ref 16.) Costamagna G, Bulajic M, Tringali A, et al. Multiple stenting of refractory pancreatic duct strictures in severe chronic pancreatitis: Longterm results. Endoscopy 2006; 38:254-9. (Ref 65.) De Palma GD, Galloro G, Siciliano S, et al. Unilateral versus bilateral endoscopic hepatic duct drainage in patients with malignant hilar biliary obstruction: Results of a prospective, randomized, and controlled study. Gastrointest Endosc 2001; 53:547-53. (Ref 85.) Freeman ML. Understanding risk factors and avoiding complications with endoscopic retrograde cholangiopancreatography. Curr Gastroenterol Rep 2003; 5:145-53. (Ref 101.) Gluck M, Cantone NR, Brandabur JJ, et al. A twenty-year experience with endoscopic therapy for symptomatic primary sclerosing cholangitis. J Clin Gastroenterol 2008; 42:1032-9. (Ref 59.) Lammer J, Hausegger KA, Flückiger F, et al. Common bile duct obstruction due to malignancy: Treatment with plastic versus metal stents. Radiology 1996; 201:167-72. (Ref 32.) Leveau P, Andersson E, Carlgren I, et al. Percutaneous cholecystostomy: A bridge to surgery or definite management of acute cholecystitis in high-risk patients? Scand J Gastroenterol 2008; 43:593-6. (Ref 38.) Matlock J, Freeman ML. Endoscopic therapy of benign biliary strictures. Rev Gastroenterol Disord 2005; 5:206-14. (Ref 63.) Misra S, Melton GB, Geschwind JF, et al. Percutaneous management of bile duct strictures and injuries associated with laparoscopic

1197

1198

Section VIII  Biliary Tract cholecystectomy: A decade of experience. J Am Coll Surg 2004; 198:218-26. (Ref 26.) Shah RJ, Langer DA, Antillon MR, Chen YK. Cholangioscopy and cholangioscopic forceps biopsy in patients with indeterminate pancreaticobiliary pathology. Clin Gastroenterol Hepatol 2006; 4:219-25. (Ref 69.) Weinberg BM, Shindy W, Lo S. Endoscopic balloon sphincter dilation (sphincteroplasty) versus sphincterotomy for common bile duct stones. Cochrane Database Syst Rev 2006; 4:CD004890. (Ref 49.)

Williams EJ, Green J, Beckingham I, et al. British Society of Gastroenterology. Guidelines on the management of common bile duct stones (CBDS). Gut 2008; 57:1004-21. (Ref 46.) Yoon WJ, Lee JK, Lee KH, et al. A comparison of covered and uncovered Wallstents for the management of distal malignant biliary obstruction. Gastrointest Endosc 2006; 63:996-1000. (Ref 81.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

71 Embryology, Anatomy, Histology, and Developmental Anomalies of the Liver Joseph Misdraji

CHAPTER OUTLINE Embryology  1201 Vascular Development  1202 Anatomy  1202 Nerves  1203 Lymphatics  1203 Histology  1203 Organization of Liver Parenchyma  1204

Knowledge of embryology, anatomy, and histology of the liver is basic to understanding pathologic processes. This chapter includes brief descriptions of the development, normal anatomy, and histology of the liver, followed by a short introduction to major developmental anomalies.

EMBRYOLOGY The liver develops at three to four weeks’ gestation as an outgrowing bud of proliferating endodermal cells from the ventral wall of the foregut in response to signals from the adjacent developing heart (Fig. 71-1).1-3 At this stage, the liver bud is separated from the mesenchyme of the septum transversum by basement membrane.2 Shortly thereafter, the basement membrane surrounding the liver bud is lost, and cells delaminate from the bud and invade the septum transversum as cords of hepatoblasts—bipotential cells that will differentiate into hepatocytes and cholangiocytes.1,4 As they invade the septum transversum mesenchyme, the hepatoblasts intermingle with endothelial cells; this interaction is necessary to support hepatic morphogenesis.2 Hepatocytic differentiation involves the development of abundant rough endoplasmic reticulum and the Golgi apparatus needed for synthesis of secreted proteins. The establishment of hepatocyte cell apical-basal polarity begins as early as seven weeks.2,3 Hepatic differentiation is highly dependent on signals from the cardiogenic mesoderm and septum transversum mesenchyme, which produce fibroblast growth factor (FGF) and bone morphogenetic protein (BMP), respectively; these factors induce the expression of

Developmental Anomalies of the Liver  1205 Riedel’s Lobe  1205 Abernethy Malformation  1205 Simple Cysts  1205 Choledochal Cysts  1205 Ductal Plate Malformation  1205 Biliary Atresia  1206 Alagille Syndrome  1206

hepatic messenger ribonucleic acids (mRNAs).1,4 In the absence of FGF signaling from the cardiac mesoderm, the ventral endoderm develops into pancreas.2 Biliary development begins at six weeks when a subset of hepatoblasts close to the portal mesenchyme strongly express biliary-specific antigens (see Chapter 62).3,5 These biliary precursor cells form a continuous single layered ring around the portal mesenchyme, called the ductal plate. This plate becomes partly bilayered in the next step. A period of remodeling follows in which focal dilations appear between the two cell layers and eventually form lumens. The parts of the ductal plate not involved in the formation of ducts regress by apoptosis, and, around the time of birth, the remaining ducts are incorporated into the portal mesenchyme.5 This process begins in the hilum and extends along a gradient to the periphery of the liver. The mechanism by which hepatoblasts differentiate into biliary cells is not fully understood. Hepatocyte nuclear factor (HNF)-6 is a transcription factor that regulates the number of cells that enter the biliary differentiation pathway and their localization; Hnf-6-deficient mouse embryos show numerous hepatoblasts that express biliary antigens extending into the liver parenchyma, whereas wild-type mouse embryos have far fewer cells that express biliary antigens and that are restricted to the vicinity of the portal mesenchyme.5 The Notch pathway is also involved in bile duct development; the Jagged1/Notch2 interaction may be critical for induction of biliary differentiation and repression of hepatocytic differentiation.6 Mutations in the gene that codes for the Notch ligand Jagged-1 are associated with Alagille syndrome (see Chapter 62). Longitudinal studies of patients with Alagille syndrome have shown that duct paucity develops progressively after birth, suggesting that

1201

1202

Section IX  Liver

Larynx Stomach Heart Vitelline duct

Liver bud Duodenum Midgut

Esophagus Septum transversum Liver

Stomach

Allantois

A

Cloacal membrane

Hindgut

B

Cloaca

Figure 71-1.  Embryology of the liver. A, At the 3-mm embryo stage, the liver bud forms in response to signals from the developing heart. B, At the 5-mm stage, the hepatoblasts penetrate the septum transversum.

the Notch pathway is required to maintain a differentiated phenotype rather than to initiate differentiation and morphogenesis.5 Mutations of several genes, the products of which are involved with ciliary function, lead to disrupted morphogenesis of the bile ducts and cyst formation.7 The extrahepatic bile ducts and gallbladder develop separately and become anastomosed to the intrahepatic bile ducts by an unknown mechanism. In the initial stages of sinusoidal development, sinusoidal endothelium is continuous and non-fenestrated. This continuous fetal endothelial lining is eventually replaced by a discontinuous fenestrated endothelial lining. In the process, the sinusoidal endothelial cells lose cell markers of continuous endothelial cells, such as platelet-endothelial adhesion molecule-1 and CD34, and acquire markers of adult sinusoidal endothelial cells, such as CD4 and intracellular adhesion molecule-1.8 Thereafter, portal veins develop, followed by centrilobular veins and arteries. Mesenchymal cells from the septum transversum transform to hepatic stellate cells and fibroblasts.8 Kupffer cells presumably originate from the yolk sac because their presence in fetal liver precedes bone marrow development.1

VASCULAR DEVELOPMENT

The developing liver incorporates the omphalomesenteric (vitelline) and umbilical (placental) veins. The vitelline veins run from the yolk sac to the heart, and as the liver invades them, the midsection of the veins becomes capillarized.9 Portions of the inferior segments of the vitelline veins regress and form a single portal vein, although the right and left vitelline circulations persist as the right and left intrahepatic portal circulations in the adult liver.9 The superior segment of the left vitelline vein regresses and the superior segment of the right vitelline vein becomes the common hepatic vein, which is incorporated into the inferior vena cava.9 The umbilical veins run from the placenta to the heart. During the 6- to 7-mm stage of human development, part of the left umbilical vein becomes the ductus venosus, which shunts placenta-derived arterial blood from the umbilical vein to the inferior vena cava, bypassing the liver; the remainder of the left umbilical vein and the right umbilical vein disappear. After birth, the obliterated prehepatic segment of the left umbilical vein becomes the round ligament of the liver (ligamentum teres hepatis) in the free edge of the falciform ligament, and the ductus venosus becomes the ligamentum venosum.3

ANATOMY Parietal peritoneum covers the liver except for the bare area, where the liver comes in direct contact with the diaphragm and is suspended by fibrous tissue and the hepatic veins.10 The peritoneal reflections that surround the bare area comprise the superior and inferior coronary ligaments and the right and left triangular ligaments, which attach the liver to the diaphragm. Traditionally, four lobes are distinguished in the liver based on its external appearance: right, left, caudate, and quadrate. On the anterior surface, the falciform ligament divides the liver into the right and left anatomic lobes. On the inferior surface, the quadrate lobe is defined by the gallbladder fossa, porta hepatis, and ligamentum teres hepatis. The caudate lobe is delineated by the inferior vena cava groove, porta hepatis, and ligamentum venosum fissure.11 Although these lobes are convenient and well known, these structures are not true functional lobes. The true right and left lobes of the liver are of roughly equal size and are divided not by the falciform ligament, but by a plane passing through the bed of the gallbladder and the notch of the inferior vena cava. This plane, which has no external indications, is called the Cantlie line.10,11 Based on arterial blood supply, portal venous blood supply, biliary drainage, and hepatic venous drainage, the liver is divided into right and left functional lobes, each of which is divided into two segments, and these are further subdivided into two subsegments.10 Several systems of subdivision have been proposed, but the most widely used systems are those of Couinaud, which follows the distribution of the portal and hepatic veins,12 and Healey and Schroy, which is based on the distribution of bile ducts.13 In these systems, the subsegments are assigned numbers from 1 to 8, with the caudate lobe being subsegment 1 and the others following in a clockwise pattern (Fig. 71-2).11 The liver receives approximately 70% of its blood supply and 40% of its oxygen from the portal vein and 30% of its blood supply and 60% of its oxygen from the hepatic artery.14 The portal vein is formed from the confluence of the superior mesenteric vein and the splenic vein. At the hilum, the portal vein divides into right and left branches, upon which the right and left lobes of the liver are based.15 The hepatic artery commonly arises from the celiac trunk, although occasionally it arises from the superior mesenteric artery.15 A common variant is a left hepatic artery that branches from the left gastric artery and a right hepatic

Chapter 71  Embryology, Anatomy, Histology, and Developmental Anomalies of the Liver IVC

II VII

VIII

III IV

VI

Falciform ligament

V

Ligamentum teres hepatis

A

Gall bladder

V

Ligamentum teres hepatis

IV

VI

III

I

VII

II

IVC

Ligamentum venosum fissure Figure 71-2.  Segmental anatomy of the liver based on the Couinaud terminology. Eight segments are identified. A, Anterior view. B, Inferior view. IVC, inferior vena cava.

B

artery branch that arises from the superior mesenteric artery.15 Within the hilum, the hepatic artery lies anterior to the portal vein and to the left of the bile duct. In the liver, the arteries, portal veins, and bile ducts are surrounded by a fibrous sheath, the Glissonian sheath, whereas the hepatic veins lack this structure.10 Three major hepatic veins drain into the inferior vena cava, although in 60% to 85% of persons, the left and middle veins unite to enter the inferior vena cava as a single vein.10,15 The extrahepatic biliary tree is composed of the common hepatic duct, cystic duct, gallbladder, and right and left hepatic ducts. The right and left hepatic ducts drain the right and left lobes of the liver, respectively. The fusion of the right and left hepatic ducts gives rise to the common hepatic duct. The caudate lobe usually drains to the origin of the left hepatic duct or to the right hepatic duct. The cystic duct usually drains into the lateral aspect of the common hepatic duct below its origin.16

NERVES

Figure 71-3.  Normal portal tract with a portal vein (upper left) and paired arteriole and bile duct (center); the bile duct is to the left of the arteriole in the portal tract. (Masson trichrome, ×200.)

Sympathetic or adrenergic nerve fibers form a rich plexus around blood vessels. Branches from the plexus supply the lobules where terminal branches surround perisinusoidal cells and hepatocytes.17 Parasympathetic (cholinergic) nerve fibers innervate extrahepatic and intrahepatic branches of the hepatic artery, portal vein, and hepatic vein.17 Intrinsic

nerves regulate hepatic blood flow, biliary physiology, and metabolism. With the advent of liver transplantation, however, the importance of the hepatic nervous system has been questioned, given the adequate functioning of the denervated allograft.17,18

LYMPHATICS

Superficial lymphatics from the convex surface of the liver run through the right or left triangular ligament and the falciform ligament. They cross the diaphragm to enter precardiac, superior phrenic, and juxtaesophageal lymph nodes or travel alongside the right or left inferior phrenic artery to the celiac nodes.19 Superficial lymphatics from the visceral surface of the liver mostly run to the hepatic lymph nodes. From the caudate lobe, lymph vessels drain into precaval nodes. Deep lymphatic vessels leave the liver at the porta hepatis to drain into the foraminal node at the epiploic foramen and the superior pancreatic nodes. Lymphatic vessels that leave the liver with the hepatic veins continue in the wall of the inferior vena cava.19

HISTOLOGY The majority of the liver is composed of hepatocytes arranged in plates or “muralium” one or two cells thick, separated by sinusoids (see also Chapter 72). Hepatocytes appear as polygonal cells with round nuclei of varying sizes with frequent binucleate cells. Portal tracts within the parenchyma contain a branch of the hepatic arteriole, portal vein, and bile duct running together as a triad and accompanied by nerve fibers and lymphatic vessels (Fig. 71-3). Terminal hepatic arterioles and terminal portal venules originate from portal tracts and supply blood to the sinusoids. The sinusoids lead mixed portal and arterial blood from the portal tract to the terminal hepatic venules (also known as central veins). These terminal hepatic venules drain into sublobular veins, then into hepatic veins, and eventually to the vena cava. The terminal portal venules do not possess a muscle layer and so have no inlet sphincters at their junction with sinusoids; however, large endothelial cells at that junction bulge

1203

1204

Section IX  Liver their nuclei into the lumen and, by means of contraction, control the flow of blood into the sinusoids.18 A similar outlet sphincter-like activity occurs at the site where the sinusoid connects with the terminal hepatic vein.18 In contrast to terminal portal venules, terminal hepatic arterioles are invested by smooth muscle and are capable of forming presinusoidal sphincters.18 Sinusoidal endothelial cells are the primary barrier between blood and hepatocytes (see Fig. 72-1). They are a unique type of endothelial cell in that they have fenestrations, have a limited capacity for endocytosis, and lack a basal lamina.14 The size of the fenestrae is different in the lobular periphery and in the center of the lobule.20 Contraction of actin filaments within endothelial cells controls the pore size, which in turn is controlled by calcium and calmodulin.20 Kupffer cells are macrophages that also line sinusoids and have the capacity to phagocytose large particles.20 They are more numerous, larger, and more phagocytically active in the periportal region.14 Their major role is to clear blood of senescent red blood cells and toxic endogenous and exogenous substances.20 Apart from phagocytosis, Kupffer cells handle low-density lipoproteins and produce lymphokine mediators that direct hepatocyte protein synthesis, inflammatory mediators, and hepatocyteprotective prostaglandins.18 Hepatic stellate cells were formerly known as Ito cells, or fat-storing cells; they are perisinusoidal cells that are sites of fat metabolism and vitamin A storage.18 They encircle the sinusoidal wall and may regulate the width of the lumen. When hepatic stellate cells are activated, they transform into myofibroblasts that express desmin and smooth muscle actin.14 A perisinusoidal space, the space of Disse, remains between the sinusoidal lining and the vascular pole of hepatocytes and communicates with the sinusoidal space through multiple fenestrations.18 This space contains plasma and collagen types I, III, IV, and V, which act as the scaffolding of the organ.18 The space of Mall is a space between the periportal hepatocytes and portal connective tissue. Lymphatic fluid accumulates in the space of Disse and then passes into the space of Mall before draining into lymphatic vessels.14 Lymphatic vessels form a network in the portal spaces in association with branches of the hepatic artery.18 One aspect of the hepatocyte borders the sinusoid and another borders the bile canaliculus. The canaliculi direct bile to the terminal canals of Hering, which are lined partly by hepatocytes and partly by cholangiocytes.21 The canals of Hering do not stop at the limiting plate of the portal tract but extend into the periportal region of the lobule. The canals of Hering pass into bile ductules, which are lined entirely by cholangiocytes.21 The ductules in turn connect to the smallest interlobular bile ducts. Interlobular bile ducts connect to septal bile ducts and then into hepatic bile ducts. Histologically, the smaller ducts are lined by cuboidal cells, whereas the larger ducts are lined by columnar epithelial cells.

ORGANIZATION OF LIVER PARENCHYMA

The classic lobule of the liver was described in 1833 by Kiernan as a hexagon with a central vein at its center and portal tracts at three corners. Because many glands have as the center of their functional unit a duct, Mall envisioned the basic unit of the liver to be the portal unit, defined at its center by a portal tract and at its periphery by central veins.18 The liver acinus was defined in 1954 by Rappaport as the parenchyma around terminal afferent portal and arterial vessels that supply this group of hepatocytes with

TVH

CV

12 3

TVH

P

N

P

N

N

N

Figure 71-4.  Schematic drawing of liver architecture. At the left is the classic hepatic lobule, with the central vein as its center and portal tracts at three corners. In the middle is the portal unit, with the portal tract at its center, and central veins and nodal points at its periphery. At the right is the liver acinus, the center of which is the terminal afferent vessel (in the portal tract) and the periphery of which is drained by the terminal hepatic venule, or central vein. Zones 1, 2, and 3 extending from the portal tract to the terminal hepatic venule are shown. CV, central vein; N, nodal point; TVH, terminal hepatic venule; P, portal tract.

blood. At the periphery of the acinus lies the terminal hepatic venule (the “central vein”) which drains several acini.18 In this model, the following three zones exist: (1) the periportal zone, or zone 1, which is supplied by blood with high oxygen content; (2) the intermediate zone (zone 2); and (3) the perivenular zone (zone 3), which receives blood that is relatively low in oxygen content.18 The acinus represents a functional and a structural unit that facilitates the description of lesions such as bridging necrosis and fibrosis (Fig. 71-4).14 In 1982, Matsumoto and Kawakami presented a view of liver architecture based on its angioarchitecture.22 In this concept, the portal and hepatic venous systems are divided into a conducting portion, which delivers and drains blood from the parenchyma, and a parenchymal portion, which is the basis for the primary lobule. The parenchymal portion of the portal and hepatic venous systems consists of minute side branches that originate as orderly rows along the terminal branches of the conducting portion. The portal venous branches divide several times more often than the hepatic venous branches, thereby creating a larger number of portal venous channels for each hepatic venous channel. The final ramifications of the portal venous system are known as septal branches. The “central vein,” meanwhile, is actually six to eight draining venules that individually face a corresponding inflow unit. The conical cluster of hepatocytes fed by a septal branch and drained by a hepatic vein branch forms a “primary lobule.” Several primary lobules together form a classic lobule. Matsumoto and Kawakami also noted that the sinusoids that arise from the septal branches have a transverse course near the portal tract before turning radially to the central vein, and this bed of transverse sinusoids forms a sickle-

Chapter 71  Embryology, Anatomy, Histology, and Developmental Anomalies of the Liver PT

PT

PT

THV 3

2

1

1

THV 2

3

PT Septal branches

PT

Figure 71-5.  Drawing that compares liver blood flow in the three zones of the acinus model with Matsumoto’s concept of liver architecture. According to the model by Matsumoto, sinusoids that abut portal tracts and terminal afferent vessels (septal branches) form a hemodynamically equipotential sickle-shaped perfusion front (dotted lines). This model conforms to the concept of the classic lobule rather than to the acinus. Zones 1, 2, and 3 of the hepatic acinus are labeled. PT, portal tract; THV, terminal hepatic venule.

shaped “inflow front” for perfusion of the lobule that differs from the linear supply proposed by the acinus model (Fig. 71-5).22 The convex aspect of the sickle abuts a portal tract, its arms extend along septal branches, and the concave aspect faces the central vein. This arrangement defines two zones: the peripheral part of the classic lobule composed of adjoining sickle-shaped areas and the centrilobular portion bound by these sickle-shaped areas. Immunohistochemical studies of hepatic enzymes and many pathologic processes, such as centrilobular necrosis in chronic passive congestion and steatosis of the liver, highlight the presence of a continuous periportal network around portal tracts and terminal afferent vessels and a distinct concentric perivenous area around the central vein, supporting the idea that the liver architecture resembles the classic lobule more than the acinus.23

DEVELOPMENTAL ANOMALIES OF THE LIVER RIEDEL’S LOBE

Riedel’s lobe denotes a prominent right liver lobe that extends below the level of the umbilicus. Riedel’s lobe is an anatomic variation that occurs more often in women than in men. It may be mistaken for an abdominal mass. Liver biochemical test levels are normal, and the diagnosis is established by ultrasound.3

ABERNETHY MALFORMATION

The Abernethy malformation is a congenital extrahepatic portocaval shunt. Two types of shunts are known to occur. In a type 1 shunt, portal blood is diverted completely into the inferior vena cava, with absence of the portal vein (Fig. 71-6). This type of shunt occurs more often in girls than

Figure 71-6.  Portal tract in a patient with Abernethy malformation type 1. The portal tract contains a bile duct (upper right) and two hepatic arterial branches, but no portal vein branch. (Hematoxylin and eosin, ×200.)

in boys; is associated with other congenital abnormalities such as cardiac defects, biliary atresia, and polysplenia; may manifest with hypergalactosemia, hyperbilirubinemia, hyperammonemia, or variceal bleeding; and may be complicated by the formation of hepatic tumors such as focal nodular hyperplasia.24 Type 1 Abernethy malformation can be further divided into subtype 1a, in which the superior mesenteric vein and the splenic vein do not join and thus there is no anatomic portal vein, and subtype b, in which the superior mesenteric vein and splenic vein do join to form a portal vein, which then drains into a systemic vein.24,25 In type 2 Abernethy malformation, the portal vein is intact, but a side-to-side anastomosis with the inferior vena cava leads to shunting; technically, therefore, type 2 is not a true absence of the portal vein. A type 2 shunt occurs in both girls and boys and is not associated with other malformations.25,26

SIMPLE CYSTS

Simple cysts lined by biliary-type epithelium may arise from the biliary tree. These cysts may lose connection with the biliary tree, in which case their contents will not be bile stained (see Chapter 62).27

CHOLEDOCHAL CYSTS

Choledochal cysts are congenital dilatations of the biliary tree. They occur more often in females than in males. The location (extrahepatic, intrahepatic, or both) and character (saccular, diverticular, or fusiform) of these cysts is the basis for their classification (see Chapter 62).

DUCTAL PLATE MALFORMATION

Defects in duct morphogenesis result from a variety of mutations that affect ductal plate remodeling; they manifest as a discontinuous ring of bile duct structures encircling the portal tracts. Many of these disorders are associated with hepatic and renal cyst formation. Included in this group of disorders is congenital hepatic fibrosis, autosomal dominant and recessive polycystic kidney disease, polycystic liver disease, Caroli’s disease, and von Meyenburg complexes.3 Caroli’s disease is characterized by dilatation of the intrahepatic biliary tree. Caroli’s syndrome is defined as Caroli’s disease in conjunction with congenital hepatic fibrosis. Spo-

1205

1206

Section IX  Liver disease (most often peripheral pulmonary stenosis), characteristic facies (broad forehead, deep set eyes, straight nose, pointed chin) (see Figure 62-10), ocular abnormalities (typically posterior embryotoxon), and skeletal abnormalities (commonly butterfly vertebrae).28 The syndrome is caused by mutations or deletions of the jagged1 (JAG1) gene, which encodes the ligand for the Notch transmembrane receptor. The Notch signaling pathway functions in many cell types to regulate cell fate decisions during development (see Chapter 62).29

KEY REFERENCES

Figure 71-7.  Portal tract in a patient with Alagille syndrome. The portal tract contains a large portal vein branch and several hepatic arterial branches, but no bile duct. (Hematoxylin and eosin, ×100.)

radic von Meyenburg complexes occur frequently in the periphery of the adult liver, suggesting that they represent somatic mutations that occur postnatally in a small portion of the biliary tree (see Chapter 62).27

BILIARY ATRESIA

Extrahepatic biliary atresia is obliteration of the bile duct caused by inflammation and fibrosis. Affected infants are well until several weeks of age, when they develop conjugated hyperbilirubinemia.27 Biliary atresia occurs in syndromic and nonsyndromic settings. Syndromic biliary atresia is probably of antenatal onset and is associated with cardiac and other extrahepatic malformations, often involving laterality.27 Liver biopsy specimens from syndromic cases have histologic evidence of ductal plate malformation. The predisposition to inflammation and fibrosis may be related to abnormally toxic bile or abnormal cholangiocytes that cannot withstand the detergent action of bile acids (see Chapter 62).

ALAGILLE SYNDROME

Alagille syndrome is characterized by bile duct paucity and chronic cholestasis (Fig. 71-7; see also Fig. 62-11), cardiac

Bezerra JA. Liver development: A paradigm for hepatobiliary disease in later life. Semin Liver Dis 1998; 18:203-16. (Ref 3.) Couinaud C. Le Foie. Etudes Anatomiques et Chirurgicales. Paris: Masson & Cie; 1957. (Ref 12.) David H, Reinke P. The concept of the “perisinusoidal functional unit” of the liver—importance to pathological processes. Exp Pathol 1987; 32:193-224. (Ref 20.) Deshpande RR, Heaton ND, Rela M. Surgical anatomy of segmental liver transplantation. Br J Surg 2002; 89:1078-88. (Ref 15.) Duncan SA. Mechanisms controlling early development of the liver. Mech Dev 2003; 120:19-33. (Ref 2.) Enzan H, Himeno H, Hiroi M, et al. Development of hepatic sinusoidal structure with special reference to the Ito cells. Microsc Res Tech 1997; 39:336-49. (Ref 8.) Healey JJ, Schroy P. Anatomy of the biliary ducts within the human liver: Analysis of the prevailing pattern of branching and the major variations of the biliary ducts. Arch Surg 1953; 66:599-616. (Ref 13.) Howard ER, Davenport M. Congenital extrahepatic portocaval shunts—the Abernethy malformation. J Pediatr Surg 1997; 32:494-7. (Ref 25.) Knisely AS. Biliary tract malformations. Am J Med Genet 2003; 122A: 343-50. (Ref 27.) Malarkey DE, Johnson K, Ryan L, et al. New insights into functional aspects of liver morphology. Toxicol Pathol 2005; 33:27-34. (Ref 14.) Matsumoto T, Kawakami M. The unit-concept of hepatic parenchyma— a re-examination based on angioarchitectural studies. Acta Pathol Jpn 1982; 32 Suppl 2:285-314. (Ref 22.) Rutkauskas S, Gedrimas V, Pundzius J, et al. Clinical and anatomical basis for the classification of the structural parts of liver. Medicina (Kaunas, Lithuania) 2006; 42:98-106. (Ref 11.) Sasse D, Spornitz UM, Maly IP. Liver architecture. Enzyme 1992; 46:832. (Ref 18.) Saxena R, Theise ND, Crawford JM. Microanatomy of the human liver-exploring the hidden interfaces. Hepatology 1999; 30:1339-46. (Ref 17.) Skandalakis JE, Skandalakis LJ, Skandalakis PN, et al. Hepatic surgical anatomy. Surg Clin North Am 2004; 84:413-35. (Ref 10.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

72 Liver Physiology and Energy Metabolism Namita Roy-Chowdhury and Jayanta Roy-Chowdhury

CHAPTER OUTLINE Liver Cell Types and Organization  1207 Parenchymal Cells  1207 Sinusoidal Cells  1210 Perisinusoidal Cells  1211 Integration of the Functions of the Different Cell Types  1211 Cell-Matrix Interactions  1211 Components of the Extracellular Matrix  1212 Regeneration and Apoptosis of Liver Cells  1212 Regeneration  1212 Programmed Cell Death  1214

LIVER CELL TYPES AND ORGANIZATION Liver cells can be classified into three groups: parenchymal cells include hepatocytes and bile duct epithelia; sinusoidal cells are composed of hepatic sinusoidal endothelial and Kupffer cells (hepatic macrophages); and perisinusoidal cells consist of hepatic stellate cells and pit cells. Hepato­ cytes comprise 60% of the adult liver cell population, representing ~78% of the tissue volume (see Chapter 71).1

PARENCHYMAL CELLS Hepatocytes

Hepatocytes are large polyhedral cells approximately 20 to 30 µm in diameter.2 Consistent with their high synthetic and metabolic activity, hepatocytes are enriched in organ­ elles. About 30% of human hepatocytes are binucleate. Hepatocytes are polarized epithelial cells. Their plasma membranes have three distinct domains—(1) the sinusoidal surface (~37% of the cell surface) that comes in direct contact with plasma through the fenestrae of the specialized hepatic sinusoidal endothelial cells; (2) the canalicular surface (~13% of the cell surface) that encloses the bile canaliculus; and (3) contiguous surfaces. By analogy with glandular epithelia, the sinusoidal, canalicular, and con­ tiguous plasma membrane domains are also termed baso­ lateral, apical, and lateral surfaces, respectively.3 The sinusoidal and canalicular surfaces contain microvilli, which greatly extend the surface area of these domains. The space between the endothelia and the sinusoidal villi is termed the space of Disse. There is a bidirectional exchange of liquids and solutes between the plasma and hepatocytes at the sinusoidal surface. In many cases, the molecular transfer is augmented by proteins that promote facilitated diffusion or energy-consuming active transport. The canalicular domains of two adjacent hepatocytes are sealed at the periphery by tight junctions (desmosomes),

Protein Synthesis and Degradation in the Liver  1214 Hepatic Gene Expression  1214 Protein Folding  1216 Protein Catabolism  1216 Hepatic Nutrient Metabolism  1216 Carbohydrate Metabolism  1217 Lipid Metabolism  1220

thereby delimiting the bile canaliculus, which is the begin­ ning of the biliary drainage system (see Chapter 62). In contrast to the bidirectional flow at the sinusoidal surface, flow from hepatocytes into the bile canaliculi is predomi­ nantly unidirectional. Plasma Membranes The plasma membranes consist of lipid bilayers composed of glycerophospholipids, cholesterol, and sphingolipids that provide barrier to water and most polar substances.3,4 The inner and outer leaflets of the plasma membrane differ in lipid, protein, and carbohydrate composition, reflecting their functional differences. Protein molecules within the leaflets mediate transport of specific molecules and serve as a link with cytoskeletal structures and the extracellular matrix. Hepatocyte plasma membranes consist of 36% lipid, 54% protein, and 10% carbohydrate by dry weight. Outer leaflets of hepatocyte plasma membranes are enriched in carbohydrates. Lipid rafts are microdomains (∼50 nm diameter) of the outer leaflets of the plasma membrane that are highly enriched in cholesterol and sphingolipids.5 These are coupled to cholesterol-rich microdomains in the inner leaflet by an unknown mechanism. Raft lipids and associ­ ated proteins diffuse together laterally on the membrane surface. Some surface receptors become associated with the rafts on ligand binding, or they can lead to “clustering” of smaller rafts into larger ones. Lipid rafts are important in signal transduction, apoptosis, cell adhesion and migration, cytoskeletal organization, and protein sorting during both exocytosis and endocytosis (see later). Certain viruses enter cells via the lipid rafts. Membrane proteins perform receptor, enzyme, and trans­ port functions.6 Integral membrane proteins traverse the lipid bilayer once or multiple times or are buried in the lipid. Additional “extrinsic” protein molecules are associ­ ated with plasma membrane. Membrane proteins can rotate

1207

1208

Section IX  Liver Gap junction Hepatocyte

Tight junction

Bile canaliculus

Hepatocyte sinusoidal (basolateral) membrane Space of Disse

Kupffer cell

Stellate cell Sinusoid

Sinusoidal endothelial cell (with fenestrae)

Figure 72-1.  The spatial relationship among the different cell types of the liver. Sinusoidal plasma comes in direct contact with hepatocytes in the space of Disse. The endothelial cells are fenestrated and lack a basement membrane. Kupffer cells are located in the lumen of the sinusoid, where they are in direct contact with the sinusoidal endothelial cells and portal blood. Stellate cells are situated between the endothelial cells and hepatocytes, and come in direct contact with both cell types. The hepatocytes are joined with each other by tight junctions and the communicating gap junctions. The canalicular domain of the plasma membrane of two adjacent hepatocytes encloses the bile canaliculus.

or diffuse laterally but usually do not flip-flop from one leaflet to another. Concentration of specific membrane pro­ teins is maintained by a balance between their synthesis and degradation by shedding of membrane vesicles, proteolytic digestion within the membrane, or internalization into the cell. Receptor proteins internalized into the cell may be degraded or recycled to the cell surface. Cell Junctions.  Hepatocytes are organized into sheets (seen as chords in two-dimensional sections) by occluding (“tight”), communicating (“gap”), and anchoring junctions (Fig. 72-1). Tight junctions or desmosomes form gasket-like seals around the bile canaliculi, thereby permitting a con­ centration difference of solutes between the cytoplasm and bile canaliculus. Desmosomes are specialized membrane structures that anchor intermediate filaments to the plasma membrane and link cells together. Gap junctions are subdo­ mains of contiguous membranes of hepatocytes that com­ prise ~3% of the total surface membrane. They consist of hexagonal particles with hollow cores, termed connexons, made up of six connexin molecules.7 Connexons of one cell are joined to those of an adjacent cell to form a radially symmetrical cylinder that can open or close the central channel. Gap junctions are involved in nutrient exchange, synchronization of cellular activities, and conduction of electrical impulses. Cytoskeleton The hepatocyte cytoskeleton supports the organization of subcellular organelles, cell polarity, intracellular movement of vesicles, and molecular transport.8,9 It comprises micro­ filaments, microtubules, and intermediate filaments, as well as the cytoskeleton-associated proteins.10 Intermediate fila­ ments are polymers of fibrous polypeptides (cytokeratins

and lamins) that provide structural support to the cells. In addition, vimentin is expressed by hepatocytes in tissue culture, and neurofilaments appear in injured hepatocytes and form Mallory bodies (also termed Mallory-Denk bodies or Mallory’s hyaline). Hepatocytes express two cytokeratins, CK8 and CK18. Bile duct epithelial cells express these pro­ teins and CK19. Plectin is a giant protein that cross-links intermediate filaments to each other and to the plasma membrane, microtubules, and actin filaments. Microtubules are hollow tubular structures (with an outer diameter of 24 nm) that consist of polymerized dimers of α and β tubulin, that are involved in intracellular transport and cellular organization.11,12 Microtubules serve as tracks to the movement of cytoplasmic vesicles, mediated by ade­ nosine triphosphatase (ATPase)-powered motor proteins, kinesin, dynein, and dynamin. Depolymerization of the microtubules, for example, by colchicine treatment, inhibits plasma protein secretion without affecting protein synthe­ sis. Microtubules participate in cellular organization by interacting with the Golgi apparatus, intermediate filaments, and F-actin.13 They also maintain the integrity of the surface membrane during canalicular contraction.14 Microfilaments are composed of double-helical F-actin strands, which are polymers of G-actin. A large number of actin-associated proteins control the polymerization, depolymerization, and splicing of F-actin. Together with myosins, actins maintain the integrity of the cell matrix, facilitate bile canalicular contraction, and control tight junction permeability. Microfilaments are also important in receptor-mediated endocytosis and various transport processes. Collapse of the cellular structure of hepatocytes during apoptosis and formation of apoptotic bodies may be related to remodeling of the actin cytoskeleton of hepatocytes.15 Nucleus Nuclei of hepatocytes are relatively large and have promi­ nent nucleoli. The two concentric nuclear membranes are stabilized by networks of intermediate filaments, one inside the inner membrane and one outside the outer membrane.16 The outer nuclear membrane is in direct continuity with the ER membranes. The perinuclear space between the two nuclear membranes surrounds the nucleus and is con­ tinuous with the endoplasmic reticulum (ER) lumen. The nuclear membrane contains pores through which molecules are selectively transported to and from the cytoplasm. The ribonuclear protein (RNP) network and the perinucleolar chromatin radiate from the nucleolus. The nuclear chromatin contains the chromosomes and associated proteins. The chromosomes comprise a series of genes, interspersed with intragenic deoxyribonucleic acid (DNA). The DNA is transcribed into ribonucleic acid (RNA), which undergoes multiple processing steps, giving rise to messenger RNA (mRNA) molecules that are translocated across the nuclear pores into the cytoplasm, where they become associated with ribosomes. Nuclear DNA also encodes additional RNA types that have accessory roles in protein synthesis and other functions. Ribosomal RNAs (rRNAs) are encoded by DNA within the nucleolus. Transfer RNA (tRNA) binds to amino acids and provides a necessary link between the nucleic acid code and sequential amino acid incorporation in the growing protein chain during translation. Other RNAs are involved in the processing of mRNA, rRNA, and tRNA molecules. Just before cell divi­ sion, both the DNA and protein components of chromatin are duplicated. The two copies of each duplicated chromo­ some are separated and distributed precisely so that the two daughter cells each receive a complete set of genes.

Chapter 72  Liver Physiology and Energy Metabolism Transport between the Nucleus and the Cytoplasm.  Pores of the nuclear envelope are associated with a large number of proteins, organized in an octagonal symmetry.17 The nuclear pore complex (NPC) is a large macromolecular assembly that protrudes into both the cytoplasm and the nucleoplasm. Bidirectional nucleocytoplasmic transport occurs through the central aqueous channel in NPCs.18 His­ tones, DNA- and RNA-polymerases, transcription factors, and RNA-processing proteins are selectively transported into the nucleus from the cytoplasm, where they are synthesized, whereas tRNAs and mRNAs are synthesized in the nucleus and exported to the cytoplasm through the NPCs. Often, the export and import processes are interrelated. For example, ribosomal proteins are imported into the nucleus from the cytoplasm and, after assembly with ribo­ somal RNA, are exported to the cytoplasm as a ribosomal subunit. Proteins containing nuclear localization motifs that consist of specific cationic amino acid sequences are recog­ nized by pore complex receptors, termed importins or karyopherins, and are rapidly transported into the nucleus via an energy-consuming process powered by specific ade­ nosine triphosphatase (ATPase)/guanosine triphosphase (GTPase) enzymes. In other cases, large molecules diffuse slowly through the nuclear pores and are retained in the nucleus by binding to specific intranuclear sites. Molecules that are smaller than 5 kd diffuse freely across the nuclear pores. Endoplasmic Reticulum The ER is the largest intracellular membrane compartment, consisting of membranous tubules or flattened sacs (cyster­ nae) that enclose a continuous lumen or space and extend throughout the cytoplasm.19 The domain of ER in which active protein synthesis occurs has attached ribosomes and is termed the rough ER. The other domain, termed smooth ER, is devoid of ribosomes and is the site of lipid biosyn­ thesis, detoxification, and calcium regulation. The nuclear envelope is a specialized domain of the ER.20 Golgi Complex The Golgi complex consists of a stack of flat sac-like mem­ branes (cysternae) that are dilated at the margins.21 Many proteins synthesized in the rough ER are transported to the Golgi apparatus in protein-filled transition vesicles. The aspect of the Golgi complex facing the ER is the cis face; the opposite side is termed the trans face. Glycoproteins are thought to be transported between the Golgi sacs via shuttle vesicles. The highly mannosylated glycosyl moiety of pro­ teins that are N-glycosylated in the ER are processed in the Golgi sacs into mature forms. Some other proteins are O-glycosylated in the Golgi complex. These proteins are then sorted for transport to appropriate cellular organelles (see later discussion of exocytosis and endocytosis).22 Lysosomes Lysosomes consist of a system of membrane-bound sacs and tubules that contain hydrolytic enzymes that are active at pH 4.5 to 5.23,24 The ATPase-powered proton pump main­ tains the acid pH by importing hydrogen ions into the lyso­ somal lumen.23 Lysosomal enzymes are glycoproteins with N-linked oligosaccharides. Following synthesis in the ER, the carbohydrate moieties are modified in the Golgi appara­ tus, where their mannose residues are phosphorylated. Rec­ ognition of these mannose 6-phosphate (M6P) groups by the M6P receptor in trans-Golgi stacks25 results in their segrega­ tion and translocation into late endosomes, which transform into lysosomes.26,27

Mitochondria Mitochondria constitute about 20% of the cytoplasmic volume of hepatocytes and are responsible for cellular res­ piration.28-30 They contain the enzymes of the tricarboxylic acid cycle, fatty acid oxidation, and oxidative phosphoryla­ tion.30,31 Mitochondria conserve the energy generated by oxidation of substrates as high-energy phosphate bonds of ATP. In addition, parts of the urea cycle, gluconeogenesis, fatty acid synthesis, regulation of intracellular calcium con­ centration, and heme synthesis take place in the mitochon­ dria. Mitochondria play a key role in programmed cell death, or apoptosis (see later).32 The outer smooth surface membrane of the mitochon­ drion is functionally different from the inner membrane, which is highly folded to form cristae. Mitochondria are positioned at major sites of ATP utilization by translocation along microtubules. In addition to soluble enzymes, the mitochondrial matrix includes large intramitochondrial granules that store calcium and other ions and smaller gran­ ules that contain mitochondrial ribosomes. Mitochondrial DNA, embedded within the matrix, encodes a number of mitochondrial proteins. The remaining mitochondrial pro­ teins are encoded by nuclear genes. Glycolysis and fatty acid oxidation in the mitochondria generate chemical intermediates that feed into the citric acid cycle of energy-yielding reactions.33,34 The citric acid cycle breaks down acetyl coenzyme A (acetyl CoA) into three molecules of nicotinamide adenine dinucleotide (NADH), one molecule of flavin adenine dinucleotide (FADH2), and two molecules of carbon dioxide. Electrons derived from NADH and FADH2 drive an electron transport pathway in the inner mitochondrial membrane, leading to ATP production. Passage of electrons across the inner mito­ chondrial membrane to the space between the inner and outer membrane generates a proton gradient that drives ATP synthesis.35 Peroxisomes Peroxisomes are spherical-appearing structures that enclose a matrix that contains a lattice or crystalline core.36 Peroxi­ somes are abundant in hepatocytes and are thought to be essential for life. Several oxidative catabolic reactions, as well as anabolic reactions, take place in peroxisomes, which provide important links between the metabolism of carbo­ hydrates, lipids, proteins, fats, and nucleic acids. Exocytosis and Endocytosis Exocytosis and endocytosis are pathways involved in exporting, importing, and intracellular trafficking of mole­ cules. Addition of new proteins and lipids to the plasma membrane by exocytosis and removal of membrane com­ ponents into cytoplasmic compartments by endocytosis keep the cell surface in a state of dynamic polarization. During exocytosis, secreted proteins, synthesized in the ER, pass sequentially through the cis-, medial-, and trans-Golgi stacks and the trans-Golgi network and finally appear at the cell surface.37,38 This vectorial transport through the Golgi stacks occurs via vesicles that are coated by proteins termed coatamers or COP (COPI and COPII), which are distinct from clathrin (see later).39,40 Guanosine triphosphate-guanosine diphosphate (GTP-GDP) exchange factors and GTPactivating proteins that are specific for each type of vesicle stimulate membrane binding and catalytic activation of small guanosine triphosphatases (GTPases). Once bound to the membrane, GTPases induce recruit­ ment of COP proteins. In the ER, the first coat protein to be recruited is COPII, and vesicular/tubular clusters are formed. These clusters are thought to coalesce to form a complex

1209

1210

Section IX  Liver tubular network, termed the ER/Golgi intermediate com­ partment. Acquisition of COPI proteins by the membranes of this tubular network results in the formation of vesicles that carry out bidirectional protein transport to and from the Golgi stacks. Some vesicles that emerge from the exit side of the Golgi apparatus, termed the trans-Golgi network (TGN), can transport multiple protein molecules simul­ taneously and release them together into the extracellular medium. Other types of vesicles that carry membrane pro­ teins and enzymes destined for specific intracellular organ­ elles also pass through this secretory pathway. These vesicles are sorted at the TGN, and vesicles carrying specific cargo are delivered to appropriate target organelles.41 Endocytosis is the import of extracellular macromole­ cules by processes that include pinocytosis, phagocytosis, receptor-mediated endocytosis (RME), and caveolar inter­ nalization.42 Pinocytosis refers to nonselective bulk-phase uptake of extracellular fluid via engulfment by plasma membrane invaginations. Phagocytosis is the ingestion of particles as well as regions of the cell surface. In contrast to these nonspecific modes of uptake, RME is a mechanism of uptake of specific molecules (ligands). After the ligands bind to their specific cell surface receptors, the ligandreceptor complexes concentrate in “pits” that are coated on the cytoplasmic surface by three-pronged structures (triske­ lions) composed of three heavy chains and three light chains of clathrin. The assembled coats consist of a geometric array of 12 pentagons and a variable number of hexagons, depending on the size of the coat. The coated pits pinch off into the underlying cytoplasm as coated vesicles.43 In the next step, the vesicles lose their clathrin coat and are termed endosomes. Endosomal vesicles travel along micro­ tubules and can take three distinct pathways. Some endo­ somes return to the cell surface, and the contained ligand-receptor complexes are secreted out of the cells by a process termed diacytosis. Transferrin is a prototype ligand for diacytosis. Some other ligands, such as immuno­ globulin A (IgA) oligomers, may traverse the cells to be secreted into bile along with the receptor. This process is termed transcytosis.44 The best studied type of RME is the classical endocytotic pathway, in which the interior of the endosome is acidified by the action of a proton pump, thereby leading to ligandreceptor uncoupling.45 By mechanisms that have not been elucidated fully, the dissociated ligands and receptors are sorted into different vesicles. The ligand-containing vesicles proceed to lysosomes, where the ligand is degraded by lyso­ somal hydrolases. A majority of the ligand-free receptors translocate to the cell surface and replenish the receptor pool. Some receptors, such as the insulin receptor, do not undergo recycling and are rapidly degraded in lysosomes. In addition to the recruitment of clathrin, initiation of the formation of endocytotic vesicles requires adaptor proteins, particularly AP-2, which localizes between the lipid bilayer and clathrin. Non-scaffold proteins, such as the GTPases and dynamin, are also important in the conversion of a coated pit to a coated vesicle. This function of dynamin requires association with a protein termed amphiphysin. Genetic, cell biological, and biochemical studies are iden­ tifying additional proteins that are required for clathrin coat and vesicle formation (reviewed by Stockert45). In addition to physiological ligands, many viruses use receptormediated endocytosis to enter cells. Internalization via caveolae is another pathway by which macromolecules can enter cells. Binding of caveolin to the cytoplasmic aspect of cholesterol-rich lipid rafts on the plasma membrane generates 50- to 60-nm flask-shaped invaginations of the plasma membrane. These invaginations bud off into the cytoplasm to form vesicles, termed caveolae

or plasmalemmal vesicles. Caveolae perform various func­ tions, including signal transduction, calcium regulation, non-clathrin-dependent internalization, and transcytosis. Glucosyl phosphatidylinositol (GPI)-anchored proteins, the β-adrenergic receptor, and tyrosine kinase are concentrated in caveolae.46

Bile Duct Epithelial Cells

Bile duct epithelial cells, or cholangiocytes, comprise large and small subpopulations of cells, the cell volumes of which correlate roughly with the diameter of the intrahe­ patic bile ducts (see Chapter 62). The large cholangiocytes have a relatively more developed ER and a lower nucleusto-cytoplasmic ratio than do the small cholangiocytes.47 The paucity of expression of cytochrome P450–dependent monooxygenase activity imparts a survival advantage to the small cholangiocytes against injury by chemicals. For example, cytochrome P450 2E1–mediated formation of toxic intermediates of carbon tetrachloride leads to the loss of large cholangiocyte function after administration of the pro-toxin, whereas the small cholangiocytes are resistant to the toxic injury. Bile ducts are not mere passive conduits for biliary drain­ age but play an active role in the secretion and absorption of biliary components and regulation of the extracellular matrix composition. Cholangiocytes are highly polarized. A sodium-dependent bile salt transporter (ABAT), located at the apical (luminal) surface of cholangiocytes, mediates the uptake of conjugated bile acids by cholangiocytes, whereas an alternatively spliced truncated form of the protein (ASBT), located at the basolateral surface, mediates the efflux of the bile acids in a sodium-independent manner. The sodium-dependent glucose transporter (SGLT1), located at the apical domain, and GLUT1, a facilitative glucose transporter on the basolateral domain, are responsible for glucose reabsorption from bile. Aquaporin-1 at the apical and basolateral surfaces constitute water channels that may mediate hormone-regulated transport of water into bile by cholangiocytes. The purinergic receptor (P2u) stimulates chloride ion efflux. Activation of apical P2u by ATP, which is secreted into the bile by hepatocytes, mobilizes Ca2+ stores, thereby stimulating Cl− efflux from cholangiocytes. The large, but not the small, cholangiocytes express secretin and somatostatin receptors, the chloride/bicarbonate exchanger, and the cystic fibrosis transmembrane regulator, which may enable this population of cholangiocytes to modulate water and electrolyte secretion in response to secretin and somatostatin (see also Chapter 64).48

SINUSOIDAL CELLS Hepatic Sinusoidal Endothelial Cells

Hepatic sinusoidal endothelial cells (HSECs) account for 20% of total liver cells. These cells are distinguished by the fenestrae (pores) in their flat, thin extensions that form sieve plates. Unlike capillary endothelial cells, HSECs do not form intracellular junctions and simply overlap each other (see Fig. 72-1B). The presence of fenestrae and the absence of a basement membrane permit plasma to enter the space of Disse and come in direct contact with the sinusoidal surface of hepatocytes.49 Diameters of the fenestrae are actively controlled by the actin-containing components of the cytoskeleton in response to changes in the chemical milieu.50 Thus, the specialized endothelial lining of hepatic sinusoids serve as a selective barrier between the blood and the hepatocytes. Hepatic endothelial cells can secrete pros­ taglandins and a wide variety of proteins, including inter­ leukin (IL)-1 and IL-6, interferon, tumor necrosis factor-α (TNF-α) and endothelin.

Chapter 72  Liver Physiology and Energy Metabolism Kupffer Cells

Kupffer cells are specialized tissue macrophages that account for 80% to 90% of the total population of fixed macrophages in the body. These cells are derived from bone marrow stem cells or monocytes and are highly active in removing particulate matter and toxic or foreign substances that appear in the portal blood from the intestine.51 Kupffer cells are located in the sinusoidal lumen and are in direct contact with endothelial cells (see Fig. 72-1). They possess bristle-coated micropinocytic vesicles, fuzzy-coated vacu­ oles, and worm-like structures that are special features of cells that are active in pinocytosis and phagocytosis. An abundance of lysosomes reflects their prominent role in degrading substances taken up from the bloodstream. Kupffer cells secrete a variety of vasoactive toxic mediators, which may be involved in host defense mechanisms and in pathophysiologic processes in some liver diseases. Kupffer cells increase in number and activity in chemical, infec­ tious, or immunologic injury to the liver.52

PERISINUSOIDAL CELLS Hepatic Stellate Cells

Hepatic stellate cells (HSCs) are also known as Ito cells, vitamin A–storing cells, fat-storing cells, or lipocytes. These cells are a part of the stellate cell system, which includes similar cells in the pancreas, lung, kidney, and intestine. Hepatic stellate cells are located between the endothelial lining and hepatocytes (see Fig. 72-1B). These mesenchymal cells represent 5% to 8% of all liver cells and are important sources of paracrine, autocrine, juxtacrine, and chemoat­ tractant factors that maintain homeostasis in the micro­ environment of the hepatic sinusoid. Microfilament and microtubule-enriched flat cytoplasmic extensions of quies­ cent stellate cells store vitamin A–enriched lipid droplets and spread out parallel to the endothelial lining, contacting several cells.53 HSCs express receptors for retinol-binding protein (RBP), which mediates the endocytosis of RBPretinol complexes.54 After chronic liver injury, the slender star-shaped HSCs become activated to elongated myofibroblasts. They lose retinoids and up-regulate the synthesis of extracellular matrix components, such as collagen, proteoglycan, and adhesive glycoproteins. Stellate cell activation is the central event in hepatic fibrosis.55 Activation of HSCs is initiated by paracrine stimulation by neighboring sinusoidal endo­ thelial cells, Kupffer cells, endothelial cells, and hepato­ cytes, as well as platelets and leukocytes. Endothelial cells participate in activation by producing cellular fibronectin and by converting the latent form of TGF-β to its active, profibrogenic form. Binding of TGF-β to its receptor on HSCs plays a critical role in stellate cell activation. Binding of bacterial lipopolysaccharides (LPS) arriving to the liver from the intestine to Toll-like receptor 4 (TLR4) enhances the effect of TGF-β on HSCs by two different mechanisms.56 First, increased chemokine expression by stellate cells results in chemotaxis of Kupffer cells, which secrete TGF-β. Second, LPS binding to TLR4, activates nuclear factor kappa B (NF-κB) via the adapter protein MyD88 (myeloid differentiation response protein), thereby down-regulating the TGF-β pseudoreceptor Bambi (bone morphogenic protein and the activin membrane-bound inhibitor) and thereby sensitizing the HSCs to TGF-β signaling. The threedimensional structure of the extracellular matrix modulates the shape, proliferation, and function of HSCs, probably by signal transduction via binding to cell surface integrins, followed by changes in cytoskeleton assembly. Activation of HSCs is perpetuated by continued effect of these stimuli, leading to several discrete changes in cell

behavior, such as proliferation, contractility, overexpres­ sion of extracellular matrix proteins (e.g., collagens I, III, IV, V, and VI; laminin; tenascin; undulin; hyaluronic acid; and proteoglycans), matrix degradation by releasing metallopro­ teinases, and the release of leukocyte chemoattractants and cytokines. The overall number of HSCs increases during fibrosis because of a change in the balance between pro­ liferation and apoptosis, which is influenced by soluble growth factors and the matrix.

Pit Cells

Pit cells, the natural killer (NK) cells of the liver, are located mainly within the sinusoidal lumen, close to Kupffer cells. They have the appearance of large lymphocytes and are adherent to the sinusoidal wall, often anchored with villous extensions (pseudopods).57 In the human liver, pit cells have pronounced polarity, abundant cytoplasm containing dense granules, a conspicuous cytocenter, and a locomotory shape characterized by hyaloplasmic pseudopods and a uropod (a tail-like structure that forms on the trailing end of a moving cell). The cytoplasmic granules appear as pits by microscopy, hence the name pit cells. Pit cells are shortlived and are replenished from extrahepatic sources. In common with circulating NK cells, the pit cells express OX-8 antigen, and some express asialoganglioside ganglio­ tetrasylceramide (asialo-GMr1). Pit cells do not express the pan–T-cell marker, OX-19, which is expressed by circulat­ ing NK cells. Although the source of pit cells remains debated, they are antigenically related to NK cells of other viscera. Pit cells have tumor cell-killing activity in the liver and are also thought to remove virus-infected liver cells. Their per-cell cytolytic activity is greater than that of circu­ lating NK cells. Pit cells may also have a role in controlling the growth and differentiation of liver cells and possibly in liver graft rejection.58

INTEGRATION OF THE FUNCTIONS OF THE DIFFERENT CELL TYPES Functional integration of the various groups of liver cells occurs through direct cell-to-cell communication (e.g., via gap junctions), paracrine secretion that affects neighboring cells, cell signaling, interaction with the extracellular matrix, and generalized response to endocrine and meta­ bolic fluxes.59 Hepatocytes and hepatic sinusoidal endothe­ lial cells lack a continuous basement membrane, and the spatial relationship of the cells is maintained through inter­ action with the extracellular matrix. Anchoring to the extra­ cellular matrix is important for the survival of hepatocytes. Anchoring also provides traction for movement and permits liver cells to receive signals from matrix components and matrix-bound growth factors. Hepatic extracellular matrix components are produced during development along the migration path of the hepatocytes and exhibit unique patterns of distribution and organization. Stellate cells, hepatocytes, and, to some extent, endothelial cells are major producers of the extracellular matrix in the liver. Excess deposition of connective tissue causes changes in hemodynamic properties and eventually impairs liver function.55

CELL-MATRIX INTERACTIONS

Cell-matrix interactions in the liver are important in main­ taining hepatocyte morphology and proliferation. For example, when plated on a flat layer of collagen, hepato­ cytes synthesize DNA at a level that is four-fold higher than when they are grown on gels composed of basement mem­

1211

1212

Section IX  Liver brane proteins. The type of matrix determines the level of expression of albumin and other hepatocyte-specific gene products in cultured hepatocytes.59,60 On the other hand, cell-cell and cell-matrix interactions determine the level of synthesis and deposition of hepatic extracellular matrix proteins by the various types of liver cells. Such interaction also modulates the production of specific enzymes and their inhibitors that mediate remodeling of the extracellular matrix. Integrin and non-integrin receptors mediate the interac­ tion of liver cells with extracellular matrix. Integrins bind to extracellular matrix proteins at specialized cell attach­ ment sites that often contain the arginine-glycine-aspartate motif, thereby resulting in attachment of the extracellular matrix to the intracellular cytoskeleton network. This attachment results in changes in cell shape, spreading, and migration. Integrins also influence cell proliferation, differentiation, survival, apoptosis, and gene expression via signal transduction.61,62 Non-integrin surface receptors mediate cell attachment by different mechanisms.

COMPONENTS OF THE EXTRACELLULAR MATRIX

Components of the extracellular matrix include collagens, noncollagenous glycoproteins, and proteoglycans. The liver contains five types of collagen (I, III, IV, V, and VI) and seven classes of noncollagenous glycoproteins (fibronectin, laminin, entactin/nidogen, tenascin, thrombospondin, SPARC [secreted protein, acidic, and rich in cysteine], and undulin). Hepatic extracellular matrix also includes a large number of proteoglycans and glycosaminoglycans, such as membrane-associated syndecan, thrombomodulin, and betaglycan, and extracellular matrix-associated versican, biglycan, decorin, fibromodulin, and perlecan.59,63

REGENERATION AND APOPTOSIS OF LIVER CELLS REGENERATION

Normal adult hepatocytes divide infrequently, with fewer than 1 in 10,000 hepatocytes undergoing mitosis at any given time—yet the liver possesses a unique capacity to replace tissue mass after liver injury or loss of liver mass. The capacity of the liver to regulate its own growth is evident in liver transplantation, where the size of the trans­ planted organ increases or decreases as appropriate to the size of the recipient. Such finely regulated hyperplasia of the liver is also seen after successful single-lobe liver trans­ plantation in children.64 Hepatic regeneration has been studied extensively in rodents. Following resection of two thirds of the liver in rats, the residual liver cells proliferate and restore the liver mass within days to weeks. Although generally termed “regeneration,” this process is, in fact, restorative hyperpla­ sia because the total liver mass, rather than the lobulated anatomic configuration, is reconstituted. In the rat, DNA synthesis peaks at 24 hours after partial hepatectomy, when approximately 35% of hepatocytes are in cell cycle. Cell division occurs six to eight hours after DNA synthesis. The time frame of DNA synthesis varies from species to species. For example, in mice, maximum DNA synthesis occurs 36 to 40 hours after hepatic resection. Because 80% to 95% of hepatocytes undergo mitosis, liver mass is restored after one or two cell divisions. All classes of hepatocytes, including diploid, tetraploid, and octaploid cells participate in this

quasi-synchronized proliferation, either by mitosis of mono­ nucleated cells or by cytokinesis of binucleated or tetranu­ cleated hepatocytes, after DNA synthesis in all nuclei. Interestingly, adult hepatocytes, rather than liver progenitor cells, contribute to liver regeneration after partial hepatec­ tomy. Only when the proliferation of adult hepatocytes is inhibited because of certain toxic or physical injuries do progenitor cells, often termed oval cells, proliferate. The oval cells are thought to give rise to both hepatocytes and bile duct epithelial cells.65 After liver injury, early signals for hepatocyte replication come from nonparenchymal cells (see Fig. 72-2B).66,67 LPS and intestine-derived cytokines stimulate Kupffer cells and hepatic sinusoidal endothelial cells to produce TNF-α and IL-6. Growth factors, such as hepatic growth factor (HGF), are released from stores in the hepatic matrix and are secreted by HSCs, whereas epidermal growth factor is secreted into portal blood by epithelial cells of the proximal small intestine and salivary glands.66 Hormones, such as triiodothyronine (T3), insulin, and nonrepinephrine, are important cooperative factors in liver regeneration.68 Repli­ cation of nonparenchymal cells lag behind that of hepato­ cytes by 24 to 72 hours. Initially, the newly proliferated hepatocytes form clusters, first in zone 1 and later in other zones of the liver (see Chapter 71). Regenerating endothelial cells invade these clusters and restore the single-cell thick liver plates. Early as well as late changes occur in the expression of extracellular matrix components and the enzymes that mod­ ulate them. The mitotic phase is mostly completed in three days, and the liver mass is restituted in about seven days. Liver cells return to their quiescent state when the liver mass is restored to the original size, give or take ~10%. A balance between mitosis and apoptosis fine-tunes the restoration of hepatic mass. The strictly self-limited nature of hepatocyte replication suggests that strong regulatory pressures are present that favor replicative repression. The ability of the liver to regulate its size is dependent on signals generated outside the liver, such as hormonal or metabolic signals, as well as internal signals generated within the liver.65 Signals for cessation of growth of the regenerating liver are understood less well than those governing hepato­ cellular replication.

Gene Expression During Hepatic Regeneration

The regenerative process is a cascade of events that move cells from their resting G0 phase through the G1 phase, S (DNA synthesis) phase, G2 phase, and then to M (mitotic cell division) phase (Fig. 72-2A) (see Chapter 3). Expression of a large number of genes is induced or down-regulated after partial hepatectomy at transcriptional or posttranscriptional levels.66,68 The sequence of activation of various genes during liver regeneration has been elucidated by studies using partial hepatectomy and gene knockout mice that lack specific cytokines. These genes include cell cycle genes, metabolic genes, genes coding for extracellular matrix proteins, growth factors, cytokines, and transcription factors. Chronologically, these genes can be grouped into immediate early genes, delayed early genes, and cell cycle– associated genes. Expression of these genes is modulated by signal transduction pathways that receive and transduce stimuli for cell replication and tissue remodeling. Immediate Early Genes Immediate early genes are activated almost immediately after partial hepatectomy without the need for protein syn­ thesis. More than 70 immediate early genes have been iden­ tified, and more are expected to be discovered by microarray

Chapter 72  Liver Physiology and Energy Metabolism Stellate cells and matrix degradation

Liver injury or loss of liver mass Loss of liver mass

Delayed early genes

G1

G

0

Quiescence

HGF

G0/G1

Cyclin D

Kupffer cells

TNF-α IL-6

Endothelial cells

sis

) (M

(S)

s Mitosi

Cyclin E and A

2

EGF TGF-β

Immediate early genes

G

A

Portal blood flow LPS, cytokines

Proximal small intestine

DN A

e th syn

B

NF-κB/PHF STAT3

G1/S

G1 IE genes TFs (AP-1, Myc)

Cyclin D1 cdks

Hepatocyte

T3

Insulin

Norepinephrine

Figure 72-2.  A, The cell cycle of hepatocytes in response to liver injury or loss of liver mass. Quiescent hepatocytes (G0) rapidly enter G1 after loss of liver mass (e.g., partial hepatectomy), along with expression of immediate early genes. This phase is followed sequentially by the expression of delayed early genes and cyclins. DNA synthesis (S phase) reaches a peak in 24 hours in rats and 36 to 40 hours in mice. Shortly thereafter, the cell enters G2 and undergoes mitosis (M). B, The sequence of signals that leads to liver regeneration following liver damage or partial hepatectomy. Intestine-derived lipopolysaccharides (LPS) and cytokines in the portal venous blood activate Kupffer cells and endothelial cells, which release tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6). These signals lead to the activation of nuclear factor kappa B (NF-κB), also known as post hepatectomy factor (PHF), and STAT3 (signal transducer and activator of transcription-3), without the need for new protein synthesis. Hepatic growth factor (HGF) is released by hepatic stellate cells and also may be derived from storage sites following matrix degradation. Epidermal growth factor (EGF), secreted by proximal small intestinal and salivary gland epithelial cells, as well as insulin, triiodothyronine (T3), and norepinephrine serve as cooperative factors for transition of hepatocytes through G1 to the S phase. Immediate early (IE) genes and transcription factors (TFs), including AP-1 and Myc, are expressed as the hepatocyte enters the initial phase of G1. Delayed early genes and cyclins are expressed later in G1. Transforming growth factor-β (TGF-β), which inhibits hepatocyte DNA synthesis, is blocked during the proliferative phase. Removal of the block at the end of the cell cycle may be one of the factors that permit the hepatocyte to return to the quiescent state. AP-1, activator protein-1; cdks, cyclin-dependent kinases. (Data from Taub R. Liver regeneration: From myth to mechanism. Nat Rev Mol Cell Biol 2004; 5:836-47.)

analysis of gene expression following partial hepatectomy. Many of these immediate early genes are involved in meta­ bolic processes not directly linked to DNA synthesis. In addition to the proto-oncogenes, c-fos, c-jun, c-myc, and c-ets, the immediate early genes include transcription factors, such as NF-κB, STAT3 (signal transducer and activator of transcription), activator protein-1 (AP-1), C/ EBPβ (CCAAT enhancer binding protein β), insulin-like growth factor-binding protein-1, phosphatases, cyclic AMP responsive promoter element modulator gene (CREM), X-box–binding protein 1 (XBP-1), and metabolic genes such as phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase. In the quiescent liver, NF-κB remains in the cytosol and is inactivated by binding to its inhibitor (IκB). Binding of TNF to its cell surface receptor initiates a signaling cascade that culminates in phosphorylation of IκB, causing the release of NF-κB and its translocation to the nucleus and resulting in transcriptional activation of more than a dozen genes likely to be involved in the immediate early response. IL-6 is one of the target genes of NF-κB. IL-6 is a strong inducer of STAT3 activation and is thought to play an important role in hepatic regeneration. C/EBPα expression is down-regulated during liver regeneration, whereas C/ EBPβ expression is induced. C/EBPα may repress hepato­ cyte replication by inhibiting the proteolytic degradation of the cell cycle inhibitor p21 and by reducing E2F complexes containing the retinoblastoma protein p107. On the other hand, C/EBPβ activates the expression of mitogen-activated protein kinase phosphatase (MKP-1), Egr-1 transcription factor, and the cell cycle proteins cyclin B and E. CREM and XBP-1 participate in the regulation of liver regeneration through their effect on cAMP-responsive genes.

Delayed Early Genes Delayed early genes are transcribed after the immediate early gene response but before the cell cycle genes reach maximum levels of expression. Expression of these genes occurs during the G0→G1 phase transition and is dependent on protein synthesis. This group of genes includes those that encode HRS/SRp40 (a splicing factor and modulator of alternative splicing of RNA transcripts) and the antiapoptotic gene, bcl-x. In contrast, the pro-apoptotic genes, BAK, BAD, and BAX are initially down-regulated after partial hepatectomy and are induced at a later time. Cell Cycle Genes Cyclins and cyclin-dependent kinases (cdks) are expressed during cell cycle progression from the G1 through S to M phase. During the G1 phase, cdks catalyze the phosphoryla­ tion of retinoblastoma gene protein (pRb), causing its dis­ sociation from the E2F family of proteins. This dissociation eliminates the repression of gene expression by pRb. In regenerating mouse liver cyclin D1, mRNA is expressed before DNA synthesis, whereas the expression of cyclin E mRNA coincides with DNA synthesis. Cyclin D1 forms a complex with cdk4, which causes phosphorylation of pRb, resulting in E2F activation. Cyclin D1 may also sequester the cell cycle inhibitor p27.

Integration of Cytokine and Growth Factors in Liver Regeneration

The early, reversible phase of liver regeneration, during which hepatocytes can enter the cell cycle by moving from the quiescent G0 state to early G1 phase, is termed priming.66 This phase is initiated by the effect of cytokines, the best

1213

1214

Section IX  Liver studied of which include TNF-α and IL-6. Generation of reactive oxygen species as a consequence of the acute meta­ bolic changes and release of LPS that occur in response to the loss of hepatic functional mass may have a role in trig­ gering the initial cytokine response. During priming, NF-κB and STAT3 are activated, and AP-1 and C/EBP are expressed. Together, these factors lead to the immediate early gene expression response after partial hepatectomy (see earlier). The priming events sensitize hepatocytes to growth factors. In the absence of growth factors, the cells cannot move past a certain “restriction point” in G1. The second phase of liver regeneration, termed progression, requires HGF and TGF-α as well as cyclins D1 and E. During the progression phase, the cells move past the restriction point in G1 to S and beyond. When the peak level of cyclin D1 expression is reached, cells progress autonomously through the cell cycle, without further need for growth factors. Expression of HGF, TGF-α, and probably epidermal growth factor (EGF) increase after partial hepatectomy. These factors are the direct mitogens for liver regeneration. EGF binds to both the EGF receptor and the TGF-α receptor, and c-met is the receptor for HGF. Growth hormone, thyroid hormones, and parathormone are permissive for liver regeneration, whereas insulin and norepinephrine are considered adjuvant factors.66 HGF and c-met Major sources of HGF in the liver are Kupffer cells and HSCs. HGF is produced as a single 87- to 90-kd pro-protein by nonparenchymal cells and is cleaved into ~64-kd and ~32-kd peptides that form heterodimers.65,66 HGF mRNA levels are increased 12 to 24 hours after partial hepatectomy in rats. Elevated levels of HGF have been observed in the serum of patients with fulminant hepatic failure, thus sug­ gesting an important role for HGF in regeneration of human liver. C-met, the HGF receptor, is a heterodimer consisting of a 145-kd β-chain and a 45-kd α-chain, linked by disulfide bonds. The two polypeptide chains of c-met are also derived from proteolytic cleavage of a single precursor protein. The β-chain contains the transmembrane region and the intracel­ lular tyrosine kinase domain. HGF binding to the extracel­ lular domain of c-met activates tyrosine kinase, thereby initiating a signal transduction pathway.

PROGRAMMED CELL DEATH

Programmed cell death, or apoptosis, is an integral part of hepatic regeneration. Apoptosis is involved in a fine tuning and remodeling process that results in reconstruction of the hepatic architecture. Apoptosis results in the removal of damaged, senescent, or supernumerary cells, without altering the cellular microenvironment. Loss of function of pro-apoptotic proteins, overexpression of anti-apoptotic proteins, or loss of apoptotic signaling in cells can lead to the survival of DNA-damaged cells, leading in turn to several forms of cancer.69 Apoptotic signals can originate within the cells through mechanisms that sense DNA damage and inappropriate pro­ liferative signals. In other cases, the apoptotic signals come from other cells in at least three ways.70 First, cells recog­ nized as foreign or as pathogens may receive apoptotic signals from immune mediator cells. Second, the nurturing signals of neighboring cells or extracellular matrix may be lost, thus resulting in apoptosis of anchor-dependent cells. Third, some cells undergo apoptosis in response to certain growth factors such as TGF-β1. In contrast to necrosis, apoptosis is an active process that culminates in cell death. During the latent phase of apop­ tosis, the cell undergoes molecular and biochemical change but remains morphologically intact. In the execution phase,

a series of dramatic structural changes take place that culminate in the fragmentation and condensation of the cell into membrane-enclosed apoptotic bodies. Initially, a variety of stimuli, including DNA damage, growth factor withdrawal, toxins, or radiation, trigger the apoptotic pathway. The signal is transduced by a series of defined protein-protein interactions Finally, cell death is executed by the activation of specific proteases called caspases that cleave multiple substrates, leading to DNA fragmentation, chromatin condensation, cell shrinkage, and membrane blebbing. The apoptotic cell may be phagocytosed or simply lose contact with neighboring cells. Apoptosis does not cause an acute inflammatory reaction. All these morpho­ logic features of apoptosis contrast with those of necrosis, in which the cell swells and releases proinflammatory mate­ rial into the neighboring space.69 The two major apoptotic pathways include activation of cell surface death receptors70 and mitochondrial permeabi­ lity transition.71 At least six different cell surface molecules can function as death receptors. One of the best character­ ized death receptors is Fas (also known as Apo1 or CD95). Fas belongs to the family of TNF receptors. Binding of Fas to Fas ligand leads to an interaction between the cytoplasmic domain of the Fas receptor and the death domain of the adaptor protein, FADD (Fas-associated protein with death domain), which in turn recruits and activates procaspase-8. Once activated, caspase-8 activates downstream caspases such as caspase-3. The second major pathway involves mito­ chondria and is triggered by various toxic insults. Either Bax or Bak opens channels and thereby releases the electron transport protein cytochrome c and other proteins from the intermembranous space into the cytoplasm. Cytochrome c binds the scaffolding protein Apaf-1. The C-terminal portion of Apaf1 is a negative regulator of apoptosis. The N-terminal region contains a caspase recruitment domain and an ATPase domain. Binding of cytochrome c and deoxyadenosine triphosphate (dATP) removes the negative regulatory influ­ ence of the C-terminus of Apaf-1, thereby permitting binding and autoactivation of caspase 9. Activated caspase 9, in turn, activates caspases-3 and -7, thus initiating cell death. In addition, permeabilization of the mitochondrial outer membrane results in the loss of function of the electron transport chain, which is essential for most mitochondrial functions, including ATP generation.

Expression of Genes Involved in Apoptosis During Liver Regeneration

Liver regeneration is a complex process that involves a balance of cell replication, apoptosis, and remodeling and is orchestrated by a number of molecular mediators. Genes involved in apoptosis are actively expressed in the regener­ ating liver. These genes include the inducing genes c-fos, c-jun, c-myc, TP53, Bax, Bad, Bak, and TGF-β; the apoptosis inhibitory genes, Bcl-2, Bcl-XL, TRPM-2/clusterin; and the Rb gene. Some of these genes are also involved in cell pro­ liferation through regulation of the cell cycle.

PROTEIN SYNTHESIS AND DEGRADATION IN THE LIVER HEPATIC GENE EXPRESSION

Compared with most organs, the liver expresses a large number of genes. Over 90% of plasma proteins and about 15% of the total protein mass of the body are produced in the liver.72 As in all mammalian cells, gene expression is initiated by transcription of the gene into an RNA transcript,

Chapter 72  Liver Physiology and Energy Metabolism mediated by RNA polymerase II. The nascent RNA is modi­ fied by capping of the 5′-terminus with 7-methylguanosine, excision of the noncoding intervening sequences (introns), splicing together of the coding sequences (exons), and, in most cases, addition of polyadenylate at the 3′-end. The processed mRNA is actively transported out of the nucleus. In the cytoplasm, association of the mRNA with the 40s ribosomal subunit and methionine RNA requires several initiation factors, a cap binding protein, and ATP hydrolysis. Once this initiation complex is formed, the 60s ribosomal subunit is recruited and polypeptide chain elongation pro­ ceeds as specific tRNAs recognize corresponding codons and sequentially attach appropriate amino acids. Chain elongation requires elongation factors and energy provided by GTP hydrolysis. Cessation of translation at the stop codons requires recognition by a termination factor. In most cases, the nascent protein is processed by cleavage of an amino terminal signal peptide. Many proteins undergo further proteolytic cleavage, cotranslational glycosylation, and modification of the carbohydrate moieties in the Golgi apparatus, before being secreted or transported to other intracellular organelles (see earlier). Gene expression is regulated at multiple levels. Gene trans­cription is regulated by the state of the chromatin, which determines the accessibility of specific genes to the transcription machinery, and binding of specific transcrip­ tion factors that promote or repress gene transcription. Posttranscriptional regulation can involve differential splicing, modulation of mRNA stability and efficiency of translation, protein folding, association with self or other proteins, or phosphorylation or other forms of protein modification. Modulation of protein degradation is another important mechanism that regulates net protein content. All of these modes of regulation are active in liver cells and are areas of intensive investigation. Some genes expressed in hepatocytes, loosely termed “housekeeping genes,” are expressed in many other organs

as well. In addition, the expression of many other genes occurs preferentially or uniquely in the liver. Expression of these liver-specific genes permits the liver to perform essen­ tial functions of the body, including secretion of plasma proteins, gluconeogenesis, glycogen storage, glucose meta­ bolism, cholesterol homeostasis, bile salt production, and detoxification of endogenous metabolites and exogenous substances. A series of cis-acting elements in specific genes mediate their hepatocyte-preferred expression.73 These cisacting DNA elements bind different families of HNFs. Although none of these factors is entirely liver-specific, high levels of liver-preferred gene expression occur only in the presence of combinatorial interaction of these transcription factors. Maintenance of hepatocyte-enriched expression of specific transcription factors involves cross-regulation by other unrelated liver-enriched transcription factors. Some of the transcription factors involved in hepatocyte specific­ ity are also important in hepatic tissue specification during embryogenesis. Many of the transcription factors are nor­ mally located in the cytoplasm. Binding of hormones or cytokines to their respective cell surface receptors causes conformational changes in the cytoplasmic domain of these receptors, often through phosphorylation. Such conforma­ tional changes lead to a series of events that eventually lead to the translocation of specific transcription factors to the nucleus and their binding to the respective cis-acting elements in the regulatory regions of genes. Thus, extracel­ lular signals are transduced to a series of intracellular events, culminating in the induction or repression of gene expression. Regulation of gene transcription is the most important, but not the only, mechanism of modulation of gene expres­ sion. Stability of the RNA, translational regulation, and post-translational modifications can all affect the steadystate concentration, intracellular or extracellular location, and activity of a given gene product. The major plasma proteins synthesized and secreted by the liver are shown in Table 72-1.

Table 72-1  Some Serum Proteins Produced by the Liver PROTEIN

MOLECULAR WEIGHT (DALTONS)

α1-Acid glycoprotein (orosomucoid)

40,000

Albumin

66,500

Alpha-fetoprotein α1-Antichymotrypsin

66,300 68,000

α1-Antitrypsin (α1-AT) Ceruloplasmin Complement C3 Complement C4 C-reactive protein

54,000 132,000 185,000 200,000 118,000

Ferritin Fibrinogen

450,000 340,000

Haptoglobin Serum amyloid A Transferrin

≈100,000 9,000 79,500

FUNCTION Inhibits proliferating response of peripheral lymphocytes to mitogens Binding protein, osmotic regulator Binding protein Inhibits chymotrypsin-like serine proteinase Inhibitor of elastin Ferroxidase Complement pathway Complement pathway Binds pathogens and damaged cells to initiate their elimination Intracellular iron storage Precursor to fibrin in hemostasis, wound healing Binds hemoglobin released by hemolysis Unknown Iron-binding protein

ASSOCIATION WITH LIVER DISEASE

ACUTE-PHASE RESPONSE

—

Increased

Decreased in chronic liver disease

Decreased

Increased in hepatocellular carcinoma —

Decreased Increased

Missense mutations associated with liver disease Decreased in Wilson disease — — —

Increased

Increased in hemochromatosis Decreased in chronic liver disease

Increased Increased

—

Increased

— Increased in iron deficiency

Increased Decreased

Increased Increased Increased Increased

Data from Katz N, Jungermam K. Metabolic heterogeneity of the liver. In: Tavoloni N, Berk PD, editors. Hepatic Transport and Bile Secretion: Physiology and Pathophysiology. New York: Raven Press; 1993. p 55 and Putnan FW. Progress in plasma proteins. In: The Plasma Proteins: Structure, Function, and Genetic Control. Orlando: Academic Press; 1984. p 45.

1215

1216

Section IX  Liver Nuclear Receptors

Modulation of metabolic pathways and detoxicating mech­ anisms according to the needs of the body often requires coordinated up-regulation or repression of the expression of a set of genes. In many cases, such coordination is medi­ ated by nuclear receptors, such as retinoid X receptor (RXR), liver X-receptor (LXR), farnesoid X-receptor (FXR), constitutive androstane receptor (CAR), peroxisome pro­ liferator activator receptor (PPAR), and thyroid hormone receptor (TR).74 For example, expression of proteins that mediate bilirubin uptake by hepatocytes, intracellular storage of bilirubin, glucuronidation of bilirubin, and bile canalicular excretion of bilirubin glucuronides may all be regulated by CAR. Nuclear receptors mediate induction or repression of genes by small nonprotein molecules. For example, phenobarbital binds to CAR in the cytoplasm, leading to the translocation of CAR to the nucleus and thereby resulting in simultaneous induction of multiple genes that have CAR-binding elements in their cisregulatory regions. Similarly, bile acids bind to FXR, fibrates bind to PPAR, and thyroid hormones bind to TR. In most cases, nuclear receptors function by forming heterodimers with RXR, although some nuclear receptors can function as homodimers.

PROTEIN FOLDING

Proteins that are destined for export to intracellular mem­ branes or secretion into the plasma are translocated into the ER where folding takes place prior to secretion through the Golgi apparatus.75 The ER contains a number of molecular chaperones and folding catalysts that promote efficient folding. All chaperones enable and promote protein folding and assembly, but their specific functions differ. Many chaperones work in tandem with one other. Some molecular chaperones bind to nascent chains as they emerge from the ribosome and protect aggregationprone hydrophobic regions. Other chaperones are involved in later stages of folding, particularly for complex proteins that include oligomeric species and multimolecular assemblies. In addition to promoting proper folding, chaperones play an important role in the “quality control” of proteins, through a complex series of glycosylation and deglycosyl­ ation processes and prevention of misfolded proteins from being secreted from the cell.76 The unfolded or misfolded proteins are targeted for degradation through the ubiquitinproteasome pathway.77 Up to one half of all polypeptide chains fail to satisfy the quality control mechanism in the ER, and for some proteins, such as the cystic fibrosis transmembrane conductance regulator (CFTR), the success rate is even lower. The proportion of molecules that misfold is increased greatly in mutant proteins with amino acid substitutions. Some molecular chaperones are able to rescue misfolded proteins to enable them to have a second chance to fold correctly. Under some circumstances, chaperones can solubilize proteins that have aggregated because of mis­ folding. In some cases, energy for such active intervention may be derived from ATP hydrolysis. Many molecular chaperones, such as the heat shock protein, are up-regulated in stressful situations, when protein misfolding is more prone to occur. In addition to molecular chaperones, several classes of folding catalysts accelerate steps in the folding process. For example, peptidylprolyl isomerases increase the rate of cis/trans isomerization of peptide bonds involving proline residues, and protein disulfide isomerases enhance formation and reorganization of disulfide bonds within proteins.

PROTEIN CATABOLISM

Like protein synthesis, proteolysis, is a major process that contributes to the body protein turnover. The autophagiclysosomal pathway and the ubiquitin/proteasome pathway are the two major mechanisms of protein degradation. The autophagic-lysosomal mechanism is responsible for bulk degradation of endogenous proteins, as well as degradation of other cellular components such as RNA, carbohydrates, and lipids. This pathway may be seen as a cell restructuring mechanism. The autophagy system is regulated physiolo­ gically by plasma levels of the amino acids leucine, glutamine, tyrosine, phenylalanine, proline, methionine, tryptophan, and histidine, probably through binding to cell surface receptors and subsequent intracellular signaling. Protein kinase cascades such as mTOR, Erk, eIF2a, and others may be involved in the regulation of autophagy. Amino acids may exert their effects through these pathways in combination with insulin.78 Chaperone-mediated autoph­ agy (CMA) is a selective mechanism for the degradation of altered cytosolic proteins in lysosomes. Synthesis of a lyso­ somal receptor that is critical for CMA declines in aged animals, thereby leading to the accumulation of altered pro­ teins and eventually leading to the characteristic functional alterations in the aging liver and other organs. In transgenic mice in which the abundance of the lysosomal receptor for CMA is retained until advanced age, less damaged proteins accumulate intracellularly, and liver functions are main­ tained at youthful levels.79 The ubiquitin/proteasome pathway is the principal mech­ anism for turnover of normally short-lived proteins in mam­ malian cells.80 Ubiquitin is a small protein that can link covalently to itself or to other proteins, either as monomers or as chains of polyubiquitin. Ubiquitin is added to a target protein by ubiquitin-activating, ubiquitin-conjugating, and ubiquitin-ligating enzymes. The first function attributed to ubiquitin was the covalent binding to misfolded proteins, thereby directing proteasome-dependent proteolysis. Now, ubiquitin and ubiquitin-related proteins are also known to direct specific proteins through the endocytotic pathway by modifying cargo proteins, as well as by regulating compo­ nents of the cytoplasmic protein trafficking machinery. By regulating the turnover of mitotic cyclins, ubiquitination plays an important role in cell cycle regulation.81,82 Although the ubiquitin/proteasome pathway is generally considered to be separate from the lysosomal proteolysis mechanism, ubiquitination is sometimes required for lyso­ somal proteolysis. A subset of endocytosed proteins must be conjugated to ubiquitin as a trigger for internalization from the plasma membrane.83,84 Thus, ubiquitin conjugation appears to be important in several protein trafficking steps, including endocytosis.

HEPATIC NUTRIENT METABOLISM The liver is at the hub of numerous metabolic pathways. In this chapter, we will present a brief description of the role of the liver in energy metabolism. The liver provides energy continuously to the entire body through its ability to store and modulate the availability of systemic nutrients.85 In turn, the metabolic function of the liver is regulated by hormones secreted by the pancreas, adrenals, thyroid, and adipose tissue, as well as neuronal inputs. A liver-adipose tissue-brain-pancreas axis,86 as well as a gut-brain-liver axis,87 orchestrates the management of the energy supply to body tissues. In addition to serving as a store for excess energy as lipids, the adipose tissue, particularly visceral

Chapter 72  Liver Physiology and Energy Metabolism fatty tissue that drains into the portal circulation, plays an active role in hepatic energy metabolism by releasing free fatty acids (FFAs) into plasma and releasing a series of adi­ pokines that either increase or decrease insulin sensitivity in the liver and other tissues.86 During nutrient absorption (fed state), the liver regulates nutrient flux as the absorbed nutrients are metabolized, modified for storage in the liver and fatty tissue, or made available to other organs as an energy source. During fasting, the energy supply is main­ tained from the stored fuel and by synthesis. Starvation induces breakdown of triglycerides in adipose tissues into FFAs and glycerol. FFAs reduce insulin sensitivity, thereby affecting glucose metabolism in muscles and the liver. FFAs bind and activate PPARs in liver and other tissues, thereby affecting their gene expression.88 In the hepatocyte, most of the acetyl CoA produced by oxidation of FFA is used to synthesize ketone bodies (e.g., acetoacetate and β-hydroxybutyrate) that are released into circulation and used as an energy source by many peripheral tissues. The glycerol released by triglyceride hydrolysis is used by the liver for the synthesis of glucose, which is the only source of energy for neurons and red blood cells, or of triglyceride. The triglyceride is packaged into very-lowdensity lipoproteins (see later) and returns to the adipose tissue.89 The role of a gut-brain-liver axis in glucose homeostasis has also been established.87 In rats, lipids arriving in the intestine give rise to long-chain fatty acyl-coenzyme A by the action of acyl-CoA synthase, which sends an afferent signal to the nucleus of the solitary tract in the hindbrain through the vagus nerve. This signal leads to N-methyl-daspartate ion channel-dependent glutamatergic neurotrans­ mission through the efferent vagal fibers that supply the liver, thereby resulting in a reduction in glucose production by the liver that precedes the actual post-absorptive glucose influx from the intestine. Thus, the rapid gut-brain-liver communication helps prevent excessive fluctuation of the blood glucose level. Unfortunately, this mechanism becomes inoperative with continued intake of excessive calories for several days. Detailed reviews of hepatic nutrient metabo­ lism are available elsewhere.90

CARBOHYDRATE METABOLISM

Glucose is the primary energy source for the brain, erythro­ cytes, muscle, and renal cortex. Maintaining adequate cir­ culating levels of glucose is essential for the central nervous system, which normally uses glucose as its major metabolic fuel. After a person fasts for 24 to 48 hours, the brain can use ketones as a metabolic fuel, thereby reducing its glucose requirement by 50% to 70%.90 The liver is the principal organ that maintains total carbohydrate stores by synthesiz­ ing glycogen and generating glucose from precursors.91 Glucose is synthesized from nonoxidative metabolic prod­ ucts of glucose (pyruvate and lactate) that are generated predominantly by red blood cells (RBCs) and from amino acid precursors that are derived predominantly from muscle during prolonged starvation or exercise.

Regulation of Glucose Uptake and Efflux from the Hepatocyte

Glucose is a critical molecule in the metabolic pathway because it can be converted to amino acids, fatty acids, or glycogen, the major storage form of glucose. Glucose enters hepatocytes via the glucose transporter-2, which facilitates the diffusion of glucose across the sinusoidal membrane.92 Glucose transporter-2 differs from other members of the glucose transporter family in that it is independent of meta­

bolic conditions or insulin levels. Because of the low-affinity, high-capacity characteristics of glucose transporter-2, intra­ hepatic glucose concentration is determined by the plasma glucose level, which, in turn, is regulated by glucokinase activity (see “Formation of Glucose-6-Phosphate” later).93 Glucose transporter-1, which is present in the brain, RBCs, and hepatocytes, particularly in zone 3, is a low-capacity, high-affinity glucose transporter that permits glucose uptake by hepatocytes when the circulating glucose concentration is low. Increased expression of glucose transporter-1 during fasting enhances glucose uptake by hepatocytes. Hepatocel­ lular glucose homeostasis is maintained by interlinking path­ ways that are regulated by multiple signals, which prevent competing pathways from operating at the same time.94 Figure 72-3 illustrates these pathways and the modulating influences that control the metabolic flux of glucose and other sugars, such as fructose. Formation of Glucose-6-Phosphate Rapid conversion of glucose to glucose-6-phosphate (glucose-6-P) modulates the glucose concentration within the hepatocyte, thereby regulating influx or efflux of glucose from the hepatocyte.91 Glucose-6-P is a nodal branch point compound that can enter three independent metabolic path­ ways: (1) synthesis of glycogen, which can be mobilized rapidly during fasting; (2) anaerobic glycolysis via the Embden-Meyerhof pathway, which generates pyruvate or lactate as a substrate for the tricarboxylic acid (Krebs) cycle in mitochondria; or (3) the pentose-phosphate shunt, which generates reducing equivalents necessary for anaerobic gly­ colysis and fatty acid synthesis. The pentose-phosphate shunt is regulated by the activity of mitochondrial glucose6-P dehydrogenase.92 Conversion of glucose to glucose-6-P is catalyzed by hexokinase, which accepts several different hexose substrates, and glucokinase (GK, also termed hexo­ kinase type 4 or D), which is expressed predominantly in the liver and pancreas and is specific for glucose.95 A low-affinity, high-capacity system, GK is not inhibited by the reaction product, glucose-6-P. Therefore, the level of GK activity regulates hepatocellular glucose concentration, which determines the net uptake of glucose by hepatocytes from hepatic sinusoidal plasma. GK is activated by insulin and inhibited by glucagon.90 Mutations in the GK gene are associated with some rare cases of maturity-onset diabetes of young adults (MODY).95 Fructose-1-phosphate modulates GK activity by regulating the inhibitory activity of a GK regulatory protein.96 The regulation of GK by fruc­ tose is thought to prevent futile cycling between glucose and glucose-6-P that consumes ATP. Starvation decreases GK activity, thereby promoting glucose efflux from the hepatocyte. Conversion of Glucose-6-Phosphate to Glucose Conversion of glucose-6-P to glucose is catalyzed by glucose-6-phosphatase (glu-6-Pase), a multisubunit enzyme with its active site located within the ER lumen.97 Thus, glucose-6-P needs to traverse the ER membrane to be dephosphorylated. Inherited deficiency of glu-6-Pase causes glycogen storage disease type Ia (see Chapter 76).97 Glucose6-P transport is mediated by a microsomal transport protein, which, when defective, causes type Ib glycogen storage disease. As expected, glu-6-Pase activity is increased by starvation, resulting in an increase in hepatocellular glucose concentration and consequent efflux of glucose into the sinusoidal space by the bidirectional glucose transporter-2. Glucose-6-P can enter the pentose monophosphate shunt that generates the reduced form of nicotinamide dinucleo­

1217

1218

Section IX  Liver Blood Glucose Fructose Hepatocyte

Glucose

Glu-6-Pase

+

–

–

+

Pentose phosphate pathway

Fructose 6-P 6-Fru kinase/Pase

Glucose 1-P Fruc-1,6-P2ase

–

Fructose-2,6-P2 Fructose-1,6-P2

6 PF-1-K

+

Triglycerides D-glyceraldehyde 3-phosphate α-glycerol

Fatty Acyl-CoA

Glycerol

H 2 O2

Peroxisomal β oxidation

PEP Fatty Acyl carnitine

Fatty acid Carnitine +

Fatty Acyl-CoA

Fatty acid synthase

+

O2 Fatty Acyl-CoA

Acyl-CoA synthase

Fru-1-K Fructose-1-P

Dihydroxyacetone phosphate Acetyl CoA

Regulatory protein

GK

Glucose-6-P

se

UDPG

yla or

Glycogen cycle

Synthase

Glut 2

P ho sp h

Glycogen

CPT

I

T

Carnitine T II + Fatty Acyl-CoA

CP

FAD FADH2

PEPCK OAA

PYR

Fatty acids

Malonyl-CoA

PK

Acetyl CoA carboxylase

PYRDH

Mitochondrion Fatty Acyl PYR carnitine Acetyl CoA

β oxidation

Acetyl CoA ATP citrate lyase

Acetyl CoA

OAA Tricarboxylic acid cycle

Citrate

Citrate

CO2 + energy

Figure 72-3.  Hepatic carbohydrate and lipid metabolism. Gluconeogenic pathways are identified by dashed lines. 6-Fru Kinase/Pase, 6-phosphofructo-2kinase/fructose-2,6-biphosphatase; 6 PF-1-K, 6-phosphofructo-1-kinase; ATP, adenosine triphosphate; CoA, coenzyme A; CPT, carnitine palmitoyl­ transferase; FAD, flavine adenine dinucleotide; FADH2, reduced flavine adenine dinucleotide; Fru-1-K, hepatic fructokinase; Fruc-1,6-P2ase, fructose-1,6-biphosphatase; Fructose-1-P, fructose-1-phosphate; Fructose-1,6-P2, fructose-1,6-diphosphate; Frucose-2,6-P2, fructose-2,6-diphosphate; Fructose 6-P, fructose-6-phosphate; GK, glucokinase; Glucose-6-P, glucose-6-phosphate; Glu-6-Pase, glucose-6-phosphatase; Glut 2, glucose transporter 2; OAA, oxaloacetate; PEP, phosphoenol pyruvate; PEPCK, phosphoenol pyruvate carboxykinase; PK, pyruvate kinase; PYR, pyruvate; PYRDH, pyruvate dehydrogenase; T, carnitine:acylcarnitine transferase; UDPG, uridine diphosphate glucose. (Data from Piklis SJ, Granner DK. Molecular physiology of the regulation of hepatic gluconeogenesis and glycolysis. Annu Rev Physiol 1992; 54:885-909).

tide phosphate (NADPH). The other possible metabolic fate of glucose-6-P is conversion to fructose 6-P, which can enter the fructose 6-P-fructose 1,6-diphosphate (fructose-1,6-P2) pathway. Fructose-1,6-P2 modulates the activity of pyruvate kinase (PK), which can affect substrate cycling in the subsequent pyruvate (PYR)-phosphoenol pyruvate (PEP) pathway. These opposing enzyme reactions regulate the formation of gluconeogenesis precursors and glycolysis. The relative production of fructose 6-P and fructose-1,6-P2 is regulated by the opposing action of 6-phosphofructo-1phosphokinase (6-PK-1-K) and fructose-1,6-bisphosphatase (fruc-1,62Pase).91 Within this cycle is a unique enzyme: 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase (6-fru kinase/Pase). This enzyme, which combines the pro­ perties of both a 6-PF-2K and its corresponding phosphory­ lase enzyme activity, produces the regulatory product fructose-2,6-P2. Fructose-2,6-P2 is a potent activator of 6-PF1-K and inhibitor of fruc-1,62Pase. Moreover, it favors the formation of the fructose-1,6-P2 product. The enzyme is regulated by both hormonal and nutrient regulations and serves as another modulator of glucose metabolism. During starvation, when fructose-2,6-P2 levels are low, gluconeo­

genesis is enhanced. On the other hand, high levels of 6-fru kinase/Pase found during refeeding and insulin administration promote glycolysis and fatty acid synthesis. The phosphorylation status of the 6-fru kinase/Pase is regu­ lated by the cAMP-dependent kinase site and phosphatase 2A activity. From fructose-1,6-P2, a sequence of four biochemical reac­ tions leads to the formation of PEP with generation of eight molecules of ATP.85 PEP can then be metabolized into PYR as part of the third regulatory cycle in glucose metabolism. Pyruvate kinase, which transforms PEP to PYR, generates two ATP molecules. PYR is another nodal branch point in the metabolic pathway, from which it can undergo further metabolism in mitochondria to form acetyl-CoA. Thereafter, it can undergo aerobic metabolism by the tricarboxylic acid cycle. In this pathway, PYR may be metabolized ultimately to water and carbon dioxide, with the production of 15 molecules of ATP per molecule of PYR. Other products of the tricarboxylic acid cycle are also precursors for fatty acid (citrate) or amino acids by means of oxaloacetate formation. Fructose-1,6-P2 is also an inducer of PK.98 In the reverse reaction, PYR is metabolized to oxaloacetate, which is a

Chapter 72  Liver Physiology and Energy Metabolism precursor to the amino acid L-aspartate. Oxaloacetate is converted by the energy-dependent activity of phospho­ enolpyruvate carboxykinase (PEPCK), an important regula­ tor of gluconeogenesis. PEPCK expression is inhibited by insulin at the transcriptional level90,98 and is up-regulated during fasting and in diabetes mellitus.

Hepatic Metabolism of Galactose and Fructose

Lactose, a major disaccharide present in human and cow milk, is split into glucose and galactose. Galactose can be converted to glucose-6-P, after which it can be used for glycogen synthesis; or it can be oxidized further to form PYR or acetyl-CoA for additional energy generation or fatty acid synthesis.85 Galactose is initially phosphorylated by galactokinase to form galactose-1-phosphate (galactose-1-P). In the presence of uridine diphosphoglucose (UDPG), further metabolism by uridyltransferase forms glucose-1phosphate (glucose-1-P) and UDP-galactose. UDP-galactose can be epimerized by UDP-glucose-4-epimerase to form UDP-glucose, which is a precursor to glucose-1-P. Glucose1-P can be converted to glucose-6-P. Thus, like glucose, galactose can participate in the glycolytic pathway. Fructose, an abundant sugar in the diet, is absorbed by the intestinal epithelium by a sodium-independent carrier distinct from the intestinal glucose transporter. It is con­ verted to fructose-1-phosphate (fructose-1-P) by hepatic fructokinase (Fru-1-K), using either ATP or GTP as a cofac­ tor. Fructose-1-P activates GK activity by removing the inhibitory regulatory protein. Fructose-1-P does not enter the glucogenic pathway but is further metabolized by fructose-1-phosphate aldolase to form two trioses: dihydroxylacetone phosphate and glyceraldehyde-3phosphate. Dihydroxylacetone phosphate may be isomer­ ized to glyceraldehyde phosphate and enter the glycolytic pathway or may be reduced to glyceraldehyde-3-phosphate and provide the glycerol backbone for triacylglycerol and phospholipids. Glyceraldehyde-3-phosphate may be com­ bined with dihydroxylacetone phosphate by aldolase B ultimately to form fructose-1,6-P2. Depending on the meta­ bolic requirements of the liver, fructose-1,6-P2 can be used for gluconeogenesis and glycogen synthesis or may be sub­ jected to glycolysis, ultimately resulting in the formation of lactate. Because fructose enters the carbohydrate cycle at the second regulatory step, fructose is a better substrate for lipogenesis in the liver than is glucose. Aldolase B defi­ ciency results in hereditary fructose intolerance as a result of excess fructose-1-P build-up. Treatment consists of avoid­ ance of sucrose and fructose in the diet.

Glycogen Formation

Glycogen stored in the liver is the main source of rapidly available glucose for the glucose-dependent tissues, such as RBCs, retina, renal medulla, and brain.99 Hepatic glycogen stores contain up to a two-day supply of glucose before gluconeogenesis occurs, mainly from lactate, a three-carbon end-product of anaerobic glucose metabolism.85,100 Hepatic gluconeogenesis produces up to 240 mg of glucose a day, which is approximately twice the metabolic need of the RBCs, retina, and brain. The three-carbon precursors gener­ ated by anaerobic metabolism from muscle, intestine, liver, or RBCs may account for up to 50% of the glycogen pool formed during nonabsorptive states. Alanine, another major glucose precursor, is generated by the catabolism of muscle proteins, which is a major cause of muscle wasting during prolonged fasting. Glycogen stored in muscle is used locally and cannot be exported out of the cell because muscles lack glu-6-Pase. The relative contribution of each of the precur­ sors to glycogen synthesis depends on the nutritional status,

amount, and route of glucose administration (oral vs. intra­ venous) and on hormonal regulation. Rapid switching between glycogen synthesis and break­ down is mediated by a cascade of enzymes that are regulated by local nutrients and hormones.85 Glycogen phosphorylase, which is activated by phosphorylation, cata­ lyzes the breakdown of glycogen subunits, and glycogen synthase, which is activated by dephosphorylation, cata­ lyzes the addition of UDP-glucose to the expanding glyco­ gen chain. In addition, glucose and glucose-6-P are allosteric activators of the synthase enzyme, whereas glucose binding inactivates the phosphorylase. Glycogen exists as two distinct populations consisting of proglycogen, with a molecular weight of approximately 4 × 105, and macroglycogen, with a molecular weight of 1 × 107, the concentrations of which depend on the relative activi­ ties of enzymes favoring proglycogen formation (phos­ phorylase and debranching enzymes) and those favoring glycogenin formation (branching enzymes). The ability of glycogenin to initiate the formation of glycogen is important in hepatic carbohydrate metabolism. The existence of these two distinct pools of glycogen permits subtle control of glucose levels, and their relative contributions could have a physiologic role in disease states such as diabetes mellitus.

Regulation of Glycolytic-Gluconeogenic Pathways

The glycolytic-gluconeogenic pathways are regulated by hormonal signals and the relative availability of nutrients. Insulin up-regulates the expression of genes that encode the glycolytic enzymes and represses the expression of meta­ bolic enzymes responsible for gluconeogenesis. Glucagon, catecholamines, corticosteroids, and growth hormone increase cellular cAMP levels, thereby augmenting the glu­ coneogenic pathway. In many cases, post-transcriptional mRNA stabilization or degradation, post-translational phos­ phorylation or end-product inhibition, or allosteric modula­ tion contributes to the relative abundance or activity of specific enzymes.91,98 Glucose and fructose modulate the enzyme activities by direct inhibition or by allosteric modu­ lation of the enzymes. In the fed state, high activity of GK, 6-PF-1-K, and PK induced by insulin favors formation of PYR, with low activity of PEPCK and other gluconeogenic enzymes. During fasting, the fall in plasma insulin levels removes the inhibition of the gluconeogenic enzymes PEPCK and fruc-1,6-P2ase. Simultaneously, an increase in glucagon and β-adrenergic agonists raises intracellular cAMP levels, leading to inhibition of 6-PK-2 kinase activity and stimulation of fruc-2,6-Pase, thereby reducing fructose2,6-P2 concentration and activation of fruc-1,62Pase, with a net increase in gluconeogenesis. After a prolonged fast, glu­ coneogenesis is further stimulated by an increase in the supply of substrate and alterations in the concentration of various enzymes.

Carbohydrate Metabolism in Cirrhosis

Patients with cirrhosis have an increased frequency of hyperglycemia and relative hyperinsulinemia.101 The hyper­ glycemia may be explained by decreased glucose uptake by muscle and reduced glycogen storage in liver and muscle. These changes lead to insulin resistance, which causes an increase in plasma insulin levels. Other causes of relative insulin resistance include increased serum FFA levels that can inhibit glucose uptake by muscle, altered second mes­ senger activity after insulin binds to its receptor, and increased serum concentrations of cytokines that result from elevated serum levels of LPS. Increased levels of glu­ cagon and catecholamines may be contributing factors. The

1219

1220

Section IX  Liver net result is impaired nonoxidative use of glucose with decreased storage of glycogen and impaired uptake of glucose by muscle, thereby causing a relative insulinresistant state similar to that found in patients with diabetes mellitus and obesity.

LIPID METABOLISM

Fatty acids are an important energy source for the liver and serve as an efficient fuel store within and outside the liver because oxidation of fatty acids yields the highest ATP production of any metabolic fuel.85 In addition, most organs are capable of using fatty acids as a fuel.102 The liver plays a central role in regulating the body’s total fatty acid needs. Excess glucose can be converted to fatty acid for future use and stored at distal sites such as adipose tissue and deliv­ ered by lipoproteins (see “Lipid Transport” later). Triglyc­ erides are stored in the cytoplasm of hepatocytes, where they are enclosed in a monolayer of phospholipids to form lipid droplets, which are important in the energy balance of the cell and the whole organism. Under conditions of excess lipid accumulation in the hepatocyte, for example, in over­ nutrition, the risk of acquiring insulin resistance increases. Lipid droplet formation may require a family of PPARαinduced ER proteins, termed fat-inducing transcripts 1 and 2 (FIT-1 and FIT-2).103 Beta oxidation of fatty acids in mitochondria and peroxisomes has different physiologic consequences.104 Furthermore, fatty acids are structural components of cell membranes and are important in cellular function and cell anchoring. The regulation of fatty acid synthesis and transport of fatty acids to other organs in association with lipoproteins constitutes another critical role of the liver in managing the metabolic needs of the entire body.

Fatty Acid Synthesis

Fatty acid synthesis occurs in the cytosol and is regulated closely by the availability of acetyl-CoA, which forms the basic subunit of the developing fatty acid carbon chain.85 Acetyl-CoA is synthesized predominately in mitochondria and is derived mainly from carbohydrate metabolism, with a small fraction coming from amino acids.4,12,13 Acetyl-CoA is condensed with oxaloacetate to form citrate, which is exported from the mitochondria and is then cleaved by the cytosolic ATP citrate lyase to produce oxaloacetate and acetyl-CoA. Conversion of acetyl-CoA to malonyl-CoA by the action of acetyl-CoA carboxylase is the first step in fatty acid synthesis. Acetyl-CoA carboxylase is the key enzyme in regulating fatty acid synthesis because it provides the necessary building blocks for elongation of the fatty acid carbon chain.105 Malonyl-CoA is used by a set of enzymatic activities con­ tained within a single peptide chain that comprises the remarkable fatty acid synthase system.85 Malonyl-CoA binds to acyl carrier protein (ACP). Catalytic activity is contained within two distinct domains that catalyze sequential condensation, reduction, dehydrogenation, and reduction, which constitute the fatty acid synthetic cycle. Two NADPH molecules are required for each two-carbon unit that is added to the growing fatty acid chain. After completion of the first cycle, the 4-carbon butyl group is transferred from ACP to a peripheral thiol, thereby allowing it to accept the next malonyl-CoA group to restart the entire cycle. The cycle continues for an additional six or seven rounds until a carbon-16 (palmitate) or carbon-18 (stearate) fatty acid is synthesized. Fatty acid-CoA is then released and used for other metabolic pathways. Further elongation of the fatty acid chain can occur either in the mitochondrion or within the microsomal membrane.85

In the mitochondrion, the first step is mediated by enoylCoA reductase. Microsomal elongation uses malonyl-CoA to increase the size of fatty acyl-CoA in a process that involves four separate enzymatic reactions. The elongation ability of microsomes is tissue dependent and serves the needs of specific organs. The fatty acid chain elongates until an appropriate length has been achieved, and the fatty acid is then esterified with glycerol to form triglycerides. These newly formed triglycerides can be transported by lipopro­ teins to distal sites for storage and use. In situations of excess carbohydrates, PYR can be converted to acetyl-CoA by the mitochondrial pyruvate dehydrogenase complex to serve as fatty acid precursors, although lipogenesis from carbohydrates consumes about 25% of the energy contained in the carbohydrates.

Beta Oxidation of Fatty Acids

Fatty acid beta oxidation is an important source of energy for many organs, including the liver. Beta oxidation occurs in mitochondria and peroxisomes, and the process requires transport of substrates across the membranes delimiting these organelles. Mitochondrial Beta Oxidation Fatty acids are translocated across the mitochondrial mem­ branes by first undergoing fatty acyl-CoA formation by the activity of distinct fatty acyl-CoA synthetases that are spe­ cific for short-, medium-, or long-chain fatty acids in the mitochondrial outer membrane.85,106 In the inner mitochon­ drial membrane, conjugation of fatty acyl-CoA with carni­ tine is catalyzed by carnitine palmitoyltransferase I, with formation of fatty acylcarnitine, which is translocated into the mitochondrion, in exchange for free carnitine, by an integral inner membrane protein, fatty acylcarnitine:carnitine translocase.107 Inside the mitochondrion, a reverse reaction mediated by carnitine palmitoyltransferase II releases fatty acyl-CoA, which is now a substrate for beta oxidation. The first step that is unique to beta oxidation is formation of trans-enol fatty acid, which is generated by acyl-CoA dehy­ drogenase. Acyl-CoA dehydrogenase transfers two electrons to flavin adenine dinucleotide (FAD), which then transfers them to the electron transport chain in the mitochondrion. 3-Keto fatty acyl-Co then undergoes a series of sequential reactions to acetyl-CoA and fatty acyl-CoA, which undergo another round of beta oxidation. Acetyl-CoA can enter the tricarboxylic acid cycle, thereby generating 12 ATP, or it can enter the 3-hydroxyl methyl glutaryl-CoA cycle to form ketone bodies. Only mitochondria in the liver are capable of forming ketone bodies. Regulation of mitochondrial beta oxidation lies with fatty acylcarnitine formation, which is catalyzed by carnitine palmitoyltransferase I.107 MalonylCoA, the basic subunit of fatty acid synthesis, is a potent inhibitor of carnitine palmitoyltransferase I, and thus pre­ vents beta oxidation and fatty acid synthesis from occurring concurrently. Peroxisomal Beta Oxidation Peroxisomes have lesser capacity than mitochondria for beta oxidation of fatty acid. The relative contribution of peroxisomes to beta oxidation depends on the fatty acid chain length and administration of peroxisome prolifera­ tors. In contrast to fatty acid oxidation in the mitochon­ drion, initial fatty acyl-CoA formation within the peroxisome does not require fatty acyl carnitine formation for entry into peroxisomes. During the next metabolic step, in which trans-enoyl fatty acyl-CoA is formed, another significant difference occurs in the peroxisomes as compared with mitochondria: Two electrons produced are transferred to

Chapter 72  Liver Physiology and Energy Metabolism FAD to form FADH2, which is then transferred directly to oxygen to form hydrogen peroxide. Hydrogen peroxide is detoxified by catalase to form water and oxygen. (In the mitochondrion, electrons are delivered to the mitochondrial electron transport system that ultimately generates water and ATP.) The significance of this difference lies in both lack of ATP production and generation of hydrogen perox­ ide in the peroxisomes, which in the presence of transi­ tional metals can yield toxic hydroxyl radicals and can promote lipid peroxidation and oxidant injury. NADH generated in subsequent reactions needs to be removed from the peroxisomes, whereas, in mitochondria, NADH can enter the electron transport cycle and generate additional ATP molecules. Peroxisomal enzymes can metabolize only long-chain fatty acids with a minimal chain length of 10 carbons and a maximal length of 24 carbons. As in mitochondria, beta oxidation in peroxisomes pro­ ceeds similarly by 2-carbon acetyl-CoA cleavage until octanoyl-CoA is formed. Octanoyl-CoA is then combined with carnitine to form fatty acyl carnitine, which can be transported by the mitochondrial inner membrane trans­ porter and undergo completion of beta oxidation. Acyl-CoA formed in peroxisomes by beta oxidation of fatty acids can diffuse out of the peroxisomes after formation of acetyl carnitine.107 The regulation of peroxisomal metabolism of fatty acids appears to be solely at the level of substrate availability, which may be regulated by a family of soluble fatty acid binding proteins present in the cytosol of all cells. The peroxisomal pathway provides a supply of acetyl-CoA that does not require citrate formation and that can be used in fatty acid synthesis. Because the initial electron transfer is not coupled to the mitochondrial electron transport system, peroxisomal fatty acid oxidation is less efficient than mito­ chondrial beta oxidation and may provide a means of elimi­ nating fatty acids with energy loss. Peroxisomes proliferate on administration of a large number of hypolipidemic agents, such as clofibrate, with a resulting 5- to 10-fold increase in the relative contribution of peroxisomal fatty acid beta oxidation. Because peroxisomal beta oxidation produces less ATP than does beta oxidation in mitochon­ dria, a relative increase in peroxisomal fatty acid beta oxida­ tion can lead to a reduction in lipid mass and to weight loss. This pathway also provides a means of generating hydrogen peroxide, which can be used by catalase for the oxidation of substrates such as ethanol. Increased triglyceride synthesis, reduced synthesis of lipid transport proteins (see later), and a decreased level of beta oxidation can result in the accumulation of fat within hepatocytes (steatosis). A classic example of this process is alcoholic steatosis, which occurs when a large percentage of the total caloric intake is derived from ethanol. Alteration in the redox potential with excess NADH produced by ethanol metabolism results in an increased NADH/NAD ratio, which favors the formation of α-glycerol phosphate, which in turn promotes triglyceride formation. In addition, a decrease in NAD content in mitochondria may reduce fatty acid beta oxidation, thereby contributing to fatty acid accumulation.108

Lipoproteins

Apolipoproteins (apo), which are synthesized by the liver, in combination with triglycerides, phospholipids, choles­ terol, and cholesterol esters, constitute circulating lipopro­ teins, which mediate the transport of lipids from the liver into the plasma and from the plasma into the liver and other tissues. The liver also expresses cell surface receptors for circulating lipoproteins and modulates plasma levels of

these important macromolecules. Lipoprotein trafficking has been reviewed elsewhere109 and is summarized in the following section. Types of Lipoproteins Lipoproteins were originally classified according to their relative density, which is inversely related to their particle size. Listed in increasing order of density, they are: chylomicrons, very-low-density lipoproteins (VLDL), intermediate-density lipoproteins (IDL), low-density lipo­ proteins (LDL), and high-density lipoproteins (HDL). Density differences in these particles reflect the type and amount of specific lipids and the proportion of protein present within these lipoprotein fractions.109 Specific apoli­ poproteins bind lipids to form lipoproteins, which are mod­ ified by enzymes in plasma or endothelial cells and act as ligands for specific lipoprotein receptors that mediate their uptake by target tissues. The lipid components are in constant dynamic flux because of delivery of lipids and cholesterol to cells, trans­ fer to other lipoproteins (mediated by lipid transfer pro­ teins), and catalysis by lipolytic enzymes. Triglycerides are the major lipids contained in chylomicrons that are gener­ ated in the intestinal epithelial cells and VLDL produced in the liver. They are the energy source for peripheral tissues and components of cellular membrane structures. Choles­ terol is the major lipid in LDL and HDL. Cholesterol, unlike triglycerides, is not used as a fuel source but as a structural component of membranes and as precursors for steroid hor­ mones. Trafficking of cholesterol is usually in the form of cholesteryl ester, which is generated in the plasma by the activity of lecithin-cholesterol acyltransferase (LCAT) (see later). Tangier’s disease, a rare autosomal recessive disorder characterized by the accumulation of cholesteryl esters in reticuloendothelial cells, including the tonsils, thymus, and lymph nodes, as well as liver, spleen, and gallbladder, in combination with the near absence of serum HDL choles­ terol, is now recognized to be caused by mutations in the ATP-binding cassette transporter A1 (ABCA1), a member of the ABC supergene family.110 Affected patients classically present with enlarged, orange-colored tonsils and have a four- to six-fold increased risk of atherosclerotic heart disease. Although the function of the transporter is not completely known, its location at the plasma membrane suggests that it mediates the active transport (“flipping”) of cholesteryl ester from the inner to the outer leaflet of the plasma membrane, from which it can be transferred to apo­ lipoproteins and secreted.110 Apolipoproteins The major apolipoproteins associated with triglyceride transport are apoB-100, which is synthesized in the liver, and apoB-48, which is synthesized in the intestine.111 Both proteins are translated from the same mRNA. In human intestinal epithelium, the apoB mRNA undergoes posttranscriptional RNA editing, which generates a stop codon by cytidine deamination that results in the translation of a form of apoB that is approximately 48% of the size of the full-length apoB-100 generated in the liver. The carboxyterminal domain that is absent in apoB-48 is essential for binding to the LDL receptor. Unlike the apoB-100containing VLDL, chylomicron remnants, which contain apoB-48, are rapidly cleared from plasma and do not give rise to LDL.111 ApoC is synthesized predominately in the liver, with minor expression in the intestine and other organs, and is composed of three different gene products that may inhibit

1221

1222

Section IX  Liver the uptake of chylomicron remnants by the liver. ApoC-1 is a minor component of VLDL, HDL, and IDL and is of unknown function. ApoC-II is present in VLDL, IDL, HDL, and chylomicrons and is an essential activator of lipopro­ tein lipase (LPL) (see “Intestinal Lipoprotein Metabolism,” later). Inherited deficiency of apoC-II causes hypertriglyc­ eridemia. ApoC-III is present in IDL, HDL, and chylomi­ crons and may be an inhibitor of LPL activity.112 ApoE is synthesized in the liver and is found on all lipoproteins. ApoE is important for removal of lipoprotein remnants in the serum, can bind to the LDL receptor and other membrane proteins, and is important in targeting lipoproteins to specific receptors on peripheral cells. Three major alleles of the apoE gene exist (ε2, ε3, and ε4), with the ε3 allele being the most abundant and ε2/ε3 genotype being the most frequent. Each allele possesses a different ability to bind to the LDL receptor. Absence of apoE leads to reduced clearance of chylomicron and VLDL remnants, resulting in elevated plasma levels and a consequent increase in the risk of atherosclerosis.113 ApoE is also impor­ tant in lipid transport in the central nervous system espe­ cially after neuronal injury. Inheritance of a single apoε4 allele is associated with a six- to eight-year earlier onset of Alzheimer’s disease than that associated with the ε3/ε3 genotype.114 ApoA-I and -II are synthesized in the liver and intestine. ApoA-I is the major component of HDL lipoproteins. In a lipid-poor state, apoA-I accepts cholesterol from the cell membrane. ApoA-I is a key activator of LCAT, which enhances cholesterol esterification in the plasma, and the absence of a specific, conserved region in apoA-I causes loss of its LCAT activating property. ApoA-II is another compo­ nent of HDL. ApoA-IV is a minor constituent synthesized in the intestine.115 Lipolytic Enzymes LPL is synthesized in fat and muscle cells and is located in the luminal surface of the capillary bed of adipose, lung, and muscle tissues.116 LPL catalyzes lipolysis of tri­ glycerides present in VLDL, chylomicrons, or HDL. LPL is stimulated by fasting, fatty acids, hormones, and catechol­ amines. Patients who are homozygous for LPL deficiency present with severe hypertriglyceridemia in childhood and pancreatitis. Hepatic triglyceride lipase (HTGL) is another member of the lipase family. It is synthesized in the liver and binds to the luminal surface of hepatic endothelial cells. It is involved in lipolysis of VLDL or IDL and thus plays a major role in LDL formation. HDL may be another substrate for HTGL activity. Inherited deficiency of LPL leads to accumulation of large particles containing both apoB-100 and apoB-48, with almost complete absence of smaller apoB-containing lipoprotein. In animal studies, inhibition of HTGL results in accumulation of VLDL and IDL, with the enrichment of HDL in triglycerides. Lipid Transport Proteins In plasma, lipid exchange between particles is facilitated by the activity of LCAT and cholesteryl ester transfer protein (CETP).116 LCAT is synthesized in the liver, and apoA-I is a cofactor for LCAT activity. CETP is synthesized predomi­ nantly in the liver and circulates in association with HDL. CETP mediates the exchange of cholesteryl esters from HDL with triglycerides from chylomicrons or VLDL. The activity of LCAT in combination with the lipid transfer proteins, CETP and phospholipid transfer protein (PLTP), is essential for the transfer of cholesterol from nonhepatic tissue to the liver.116

Intestinal and Hepatic Lipid Transport

The liver functions as the hub for receiving fatty acids and cholesterol from the diet and peripheral tissues, packages them into lipoprotein complexes, and releases the com­ plexes into the circulation (Fig. 72-4). Following absorption by intestinal epithelial cells, fatty acids are formed into triglycerides, and cholesterol is esterified. Both lipids are packaged into nascent chylomicrons composed predomi­ nantly of triglycerides (85% to 92%), phospholipids (6% to 12%), cholesteryl ester (1% to 3%), fat-soluble vitamins, and the following apolipoproteins (1% to 3%): apoB-48, apoA-I, apoA-II, and apoA-IV.117 Nascent chylomicrons enter the interstitial space and are carried into the systemic venous circulation via the thoracic duct. In the interstitial space, chylomicrons acquire apoC-II, which activates LPL, thereby promoting triglyceride release. Triglyceride release may be reduced by acquisition of apoC-III, which may inhibit LPL activity. The addition of apoE is critical for targeting the chylomicron remnant, which can then be taken up by hepatocytes through the chylomicron remnant receptor. Release of triglycerides by LPL and extraction by periph­ eral tissues increases the relative cholesteryl ester concen­ tration in chylomicron remnants, which are taken up by hepatocytes via a hepatocyte membrane transporter that recognizes a binding domain on apoE. The endocytosed chylomicron remnants are targeted to lysosomes, where they are degraded. Inherited mutations of the binding domain of apoE reduce chylomicron remnant clearance. When chylomicron excretion is delayed, as occurs with mutations of the binding domain of apoE or reduced LPL activity or apoC-II levels, chylomicron remnants that accu­ mulate in the serum may be taken up by endothelial cells or macrophages, which transform into foamy cells. The foamy cells are precursors of fatty streaks and atheromas. Increased VLDL secretion resulting from excess fatty acid absorption can also compete with the chylomicron remnant uptake system. Fatty acids released from adipocytes by the action of intracellular hormone-sensitive lipase are bound to serum albumin and transported to other tissues including the liver, where they are used for synthesis of phospholipids and triglycerides.118 The liver synthesizes cholesterol from lowmolecular-weight precursors. Hepatic cholesterol synthesis is regulated by the rate-limiting enzyme 3-hydroxyl-3methylglutaryl coenzyme A reductase (HMG-CoA reduc­ tase). Lipids are exported from the liver as VLDL particles, which are the major carriers of plasma triglycerides during nonabsorptive states.85 Lipids may be stored temporarily in the liver as fat droplets and cholesteryl esters, excreted directly into bile, or metabolized into bile acids. The liver is the major site of sterol excretion from the body and is the site of bile acid synthesis (see Chapter 64). The coordinated input, synthesis, and excretion of sterols require complex regulation of multiple enzymatic path­ ways. Bile acids returning to the liver via the enterohepatic circulation modulate these enzyme activities. Bile acids recycle 20 to 30 times per day via the enterohepatic circula­ tion and use specific transmembrane transporters at apical and basolateral domains of hepatocyte plasma membrane, as well as intracellular binding proteins.119 In the terminal ileum, a great majority of the bile acid molecules are reab­ sorbed via a sodium-dependent bile acid transporter. Bile acids are also important in micellization of fats for intestinal absorption and as coactivators of bile acid–dependent lipase activity. FXR, a member of the sterol nuclear receptor family, binds to and is activated by bile salts. Heterodimers of activated FXR and RXR modulate the coordinated

Chapter 72  Liver Physiology and Energy Metabolism Dietary cholesterol and triglyceride

Chylomicrons

Intestine

FA energy

FA storage

Feces

Nascent HDL

Remnants

FFA

Cholesterol

Triglyceride

VLDL Liver FFA

VLDL Hormone-sensitive lipase

Scavenger receptor

Bile acids Hepatic lipase

Triglyceride

LCAT

ACAT Cholesteryl ester

LDL receptor

Macrophage HDL3

IDL Triglyceride CETP

Lipoprotein lipase

Cholesterol

LDL

HDL binding site

LCAT HDL2

Cholesteryl ester

Adipose tissue Figure 72-4.  Lipoprotein metabolism. ACAT, acylcholesterol acyltransferase; CETP, cholesteryl ester transfer protein; FA, fatty acid; FFA, free fatty acid; HDL, high-density lipoproteins; LCAT, lecithin-cholesterol acyltransferase; LDL, low-density lipoproteins; VLDL, very-low-density lipoproteins. (Modified from Shepherd J. Lipoprotein metabolism. An overview. Drugs 1994; 47[Suppl 2]:1-10).

regulation of multiple genes that encode key bile salt transporters, such as the sodium-dependent taurocholate pump (NTCP) at the sinusoidal domain of hepatocytes, bile salt export pump (BSEP) at the canalicular domain, intestinal bile acid transporter (IBAT) in the terminal ileum, and cholesterol-7α-hydroxylase in hepatocytes (see Chapter 64).120 Transport of ApoB-Containing Lipoproteins In the fasting state, VLDL, which is synthesized in the liver, replaces chylomicrons as the major transporter of triglycer­ ides and cholesterol. In addition to the full-length apoB-100, VLDL contains triglycerides (taken up from plasma or syn­ thesized in the liver), cholesteryl esters (exogenous or endogenous), and phospholipids.121 During fasting, fatty acids in VLDL are derived predominantly from the activity of hormone-sensitive lipase in adipocytes, whereas after a meal, dietary fatty acids are the major source. Fatty acids may be taken up by hepatocytes by passive diffusion or via fatty acid transport proteins in the sinusoidal domain of the cell membrane. In hepatocyte cytosol, fatty acids are stored bound to the abundant 12-kd fatty acid binding protein (FABP) family, which may direct fatty acids to specific sub­ cellular targets, such as the smooth ER for VLDL synthesis or peroxisomes for beta oxidation. FABPs are transcription­ ally regulated by peroxisome proliferating agents (e.g.,

fibrates), suggesting that their role is physiologic in global lipid metabolism. ApoB-100 is the predominant transport carrier in VLDL; apoC-I, C-II, C-III, and apoE arise from other lipoproteins within the serum. ApoB-100 synthesis and VLDL secretion are regulated by the availability of cotransported lipids and sterols in the smooth ER. ApoB-100 synthesis may change dramatically without alteration in apoB-100 mRNA levels.122 Following synthesis in the smooth ER, apoB-100 interacts with newly synthesized triglycerides and cholesteryl esters that enter the ER via specific membrane transporters. The apoB-lipid complex is translocated into the lumen, trans­ ported through the Golgi apparatus, and secreted into the sinusoidal space as VLDL. When the lipid components are not available, apoB-100 undergoes degradation in the ER. During periods of low plasma triglyceride levels, the liver secretes smaller IDL-like particles or even LDL-type particles. In the plasma, the activity of LPL and HTGL removes triglycerides from VLDL, generating progressively smaller and denser IDL and LDL particles. Conversion of IDL to LDL requires the activity of apoE. LDL particles become enriched in cholesteryl esters, both by removal of triglycerides and acquisition of cholesteryl esters from other lipopro­ teins, predominantly HDL, with release of apoC to HDL. LDL is subsequently removed from the circulation by LDL

1223

1224

Section IX  Liver receptors in the liver and peripheral tissues. Subpopula­ tions of VLDL that begin as large VLDL undergo lipolysis and are converted to IDL, which is taken up via the LDL receptor. Transport of ApoA-Containing High-Density Lipoprotein HDL, another major class of lipoproteins secreted by the liver, appears to have a protective role against atherosclero­ sis. HDL is a heterogeneous population of lipoproteins that can be separated by sophisticated analytic centrifugation techniques. Nascent HDL is formed in the liver and intestine by lipolysis of VLDL and chylomicrons, respectively, with modification by peripheral tissue. The major protein con­ stituents of HDL are apoA-I and apoA-II, with minor amounts of apoA-IV, apoC, apoE, and others.123 In humans, apoA-I is synthesized in the liver and intestine. Nascent apoAcontaining lipoprotein complexes that appear as discoid particles can be transformed into HDL particles in the serum by the action of LCAT and the lipid transfer proteins CETP and PLTP. The HDL3 subclass is particularly important because these cholesterol-poor particles are able to deliver cholesterol extracted from peripheral membranes and provide a sub­ strate for plasma LCAT activity. Cholesteryl esters formed by LCAT are extremely hydrophobic and move into the core of the lipoprotein complex, thereby providing space on the surface of the lipoprotein for extraction of additional cho­ lesterol from cell membranes. This complex enlarges with increasing amounts of cholesteryl esters, which are able to accommodate apoC-II and C-III, thereby resulting in HDL2 formation. CETP removes esterified cholesterol from HDL in exchange for triglycerides, which are eventually hydro­ lyzed by HTGL, thereby regenerating small HDL. Acquisi­ tion of apoC-II also promotes LPL activity, thereby increasing lipolysis.116 The movement of apolipoproteins between HDL and chylomicrons allows the recycling of lipids and proteins between these two pools. Cholesterol and phospholipids are also transferred to the chylomicrons as triglycerides are released by LPL activity to local tissues. As the remnant is further processed, apoC-II and apoC-III, phospholipids, and cholesterol are transferred back to HDL. Triglycerides that are transferred from VLDL and chylomicrons to HDL are more accessible to lipolysis by endothelial-based lipases because of their smaller size. With the removal of triglycerides, these particles revert to HDL3 and apoC-II, after which apoC-II and apoE recycle to chylomicrons and VLDL.

Lipoprotein Receptors

The major lipoprotein receptors for LDL, chylomicron rem­ nants, HDL, and the scavenger receptor are members of the larger LDL receptor supergene family.124 These receptors share four major structural features: (1) cysteine-rich complement-type repeats; (2) epidermal growth factor precursor-like repeats; (3) a transmembrane domain; and (4) a cytoplasmic domain.125 LDL Receptor The LDL receptor exists as an oligomeric surface glycopro­ tein that plays a pivotal role in LDL clearance and choles­ terol homeostasis. It binds ligands at the cell surface, after which the ligand-receptor complex is internalized via the classic endocytotic pathway. The ligand dissociates from the receptor in acidic endosomal vesicles. Subsequently, the ligand is delivered to lysosomes for degradation and the receptor returns to the surface. The LDL receptor is present

on all cell types; however, the liver contains approximately 70% of the total body pool of LDL receptors. The LDL recep­ tor recognizes apoE and apoB-100, but not apoB-48. ApoEcontaining chylomicron remnants, VLDL, LDL, IDL, and HDL can all be taken up via the LDL receptor. Approxi­ mately two thirds of LDL is cleared by this receptor. Homo­ zygous deficiency of the functional LDL receptor occurs in approximately 1 in 1 million persons and is associated with accelerated atherosclerosis manifesting in childhood (familial hypercholesterolemia). LDL receptors are highly conserved among species.124 Very-Low-Density Lipoprotein Receptor The VLDL receptor has a high-sequence homology with the LDL receptor but is expressed predominantly in extrahe­ patic tissues such as heart, muscle, and fat. Unlike the LDL receptor, the VLDL receptor does not bind to apoB and may serve specifically to take up triglyceride-rich apoEcontaining lipoproteins, such as VLDL or IDL.124,125 Chylomicron Remnant Receptor The chylomicron remnant receptor accepts apoE as a ligand. Chylomicron remnants are removed from the circulation exclusively by the liver, probably because these large complexes can penetrate the unique sinusoidal vascular space. The multifunctional, α2-macroglobulin/LDL receptorrelated protein (LRP) is the chylomicron remnant recep­ tor.126 LRP is present in liver, brain, and muscle. In cultured cells, LRP is able to mediate the endocytosis of apoE-con­ taining chylomicron remnants. Mice that lack LRP in the liver do not have hepatic chylomicron remnant uptake, con­ firming that LRP is the major chylomicron remnant receptor. Unlike the LDL receptor, LRP is able to bind a number of unrelated ligands, such as lipoprotein, proteinase-inhibitor complex, and protein-lipid complex. Low-Density Lipoprotein Scavenger Receptor Ligands for the scavenger receptor A (SR-A) include lipo­ polysaccharides, polyanionic lipids, and LDL in which some of the free lysine residues have been chemically modi­ fied.127 These receptors exist in two forms as trimeric integral membrane glycoproteins in endothelial cells, mac­ rophages, and Kupffer cells. Oxidized LDL is internalized via the scavenger receptors but is metabolized poorly in macrophages, leading to the accumulation of cholesteryl esters within the cell. Monocytes, which migrate into lipidenriched atherosclerotic lesions, can also be induced to express the scavenger receptor. High-Density Lipoprotein Receptor A high-affinity HDL binding protein has been identified in the plasma membrane of hepatocytes, macrophages, adrenal cells, and adipocytes.127 These receptors appear to recog­ nize specifically apoA present in HDL particles. The HDL receptor does not mediate endocytosis but allows only selective delivery of lipids to and from the HDL lipopro­ teins. By mediating the transfer of cholesterol from the plasma membrane to the HDL lipoprotein, the HDL receptor facilitates reverse cholesterol transport. The HDL receptor is a class B scavenger receptor, referred to as SR-B1.123 This receptor is most abundant in the liver, ovary, and adrenal glands—organs previously shown to be the principal sites of cholesterol uptake from HDL in vivo. HDL is a major source of cholesterol secreted in bile. Overexpression of SR-B1 in mouse liver increases biliary cholesterol secre­ tion and reduces plasma HDL.128 Conversely, deficiency of this receptor results in decreased biliary cholesterol secretion.129

Chapter 72  Liver Physiology and Energy Metabolism Derangement of Lipid Metabolism in Liver Disease

The most common lipid abnormality in patients with chronic liver disease is hypertriglyceridemia (plasma levels of 250 to 500 mg/dL), which is found in patients with alco­ holic or viral liver disease and tends to resolve when the liver disease improves. Excess ethanol ingestion causes pre­ dominantly hypertriglyceridemia, due to increased fatty acid synthesis and decreased beta oxidation of fatty acids, resulting from increased NADH production by alcohol metabolism. Moderate alcohol ingestion is associated with increased HDL3 levels, which may reduce the risk of athero­ sclerosis. LDL, HDL, and total serum cholesterol levels decrease progressively with cirrhosis advancing from Child class A to class C (see Chapters 73 and 90). Serum choles­ terol level may be a useful prognostic marker in patients with noncholestatic liver diseases.130 Cholestatic disorders manifest with a distinct pattern of dyslipoproteinemia because of the retention of cholesterol, phospholipids, and bile salts that are normally secreted in bile.131 A prolonged increase in total serum cholesterol and lipid levels, as seen in primary biliary cirrhosis, for example, can be associated with formation of xanthoma. Within the LDL fraction of the serum of cholestatic patients, three dis­ tinct lipoproteins can be identified, namely, β2-lipoprotein (triglyceride rich), also known as lipoprotein Y (LP-Y), lipo­ protein X (LP-X), and normal LDL. LP-Y appears to be a remnant of a triglyceride-rich lipoprotein that is distinct from IDL. Cholestatic patients with elevated triglyceride levels often have clear (not lipemic) serum because most of the triglycerides are contained in LP-Y and LDL. LP-X is a complex composed of equimolar amounts of excess phos­ pholipid and cholesterol in combination with albumin and certain members of the apoC family. The phospholipid flip­ pase activity of multidrug resistance protein-3 (MDR3), also termed ATP binding cassette protein B4 (ABCB4) (see Chapter 64), is essential for LP-X formation. Mice lacking mdr2 (the murine homolog of MDR3) are unable to form LP-X during cholestasis caused by complete bile duct obstruction.132 In patients with chronic parenchymal liver disease, plasma cholesteryl ester levels are often reduced, a finding that suggests that LCAT activity is diminished because of

impaired hepatic synthesis. Alternatively, decreased LCAT activity may result from reduced apoC-II levels or release of cholesteryl ester hydrolase from damaged hepatocytes, with conversion of cholesteryl esters to cholesterol. Chronic dyslipoproteinemia in these patients can also lead to altera­ tions in cellular membrane lipids, resulting in formation of abnormal RBCs, such as echinocytes, and altered membrane function with potential pathophysiologic consequences.

KEY REFERENCES

Anghel SI, Wahli W. Fat poetry: A kingdom for PPARgamma. Cell Research 2007; 17:486-511. (Ref 86.) Arias IM. The biology of hepatic endothelial fenestrae. In: Schaffner F, Popper H, editors. Progress in Liver Diseases, vol IX. Philadelphia: WB Saunders; 1990. pp 11-26. (Ref 50.) Bedossa P, Paradis V. Liver extracellular matrix in health and disease. J Pathol 2003; 200:504-15. (Ref 59.) Conner SD, Schmid SL. Regulated portals of entry into the cell. Nature 2003; 422:37-44. (Ref 42.) Hui AY, Friedman SL. Molecular basis of hepatic fibrosis. Expert Rev Mol Med 2003; 14:1-23. (Ref 55.) Karpen SJ. Nuclear receptor regulation of hepatic function. J Hepatol 2002; 36:832-50. (Ref 74.) Klionsky DJ, Emr SD. Autophagy as a regulated pathway of cellular degradation. Science 2000; 290:1717-21. (Ref 78.) Kumar NM, Gilula NB. The gap junction communication channel. Cell 1996; 84:381-8. (Ref 7.) Nakielny S, Dreyfuss G. Transport of proteins and RNAs in and out of the nucleus. Cell 1999; 99:677-90. (Ref 18.) Newmeyer DD, Ferguson-Miller S. Mitochondria: Releasing power for life and unleashing the machineries of death. Cell 2003; 112:481-90. (Ref 32.) Nordlie RC, Foster JD, Lange AJ. Regulation of glucose production by the liver. Annu Rev Nutr 1999; 19:379-406. (Ref 93.) Miller JP. Liver disease, alcohol and lipoprotein metabolism. In: Zakim D, Boyer TD, editors. Hepatology: A Textbook of Liver Diseases, 4th ed. Philadelphia: WB Saunders; 2003. p 127. (Ref 109.) Taub R. Liver regeneration: From myth to mechanism. Nat Rev Mol Cell Biol 2004; 5:836-47. (Ref 66.) Yin XM, Ding WX. Death receptor activation-induced hepatocyte apop­ tosis and liver injury. Curr Mol Med 2003; 3:491-508. (Ref 70.) Zegers MMP, Hoekstra D. Mechanisms and functional features of polar­ ized membrane traffic in epithelial and hepatic cells. Biochem J 1998; 336:257-69. (Ref 3.) Full references for this chapter can be found on www.expertconsult.com.

1225

CHAPTE R

73 Liver Chemistry and Function Tests Daniel S. Pratt

CHAPTER OUTLINE Bilirubin  1227 Bilirubin Metabolism  1227 Measurement of Serum Bilirubin  1228 Approach to the Patient with an Elevated Bilirubin Level  1228 Aminotransferases  1229 Approach to the Patient with an Elevated Aminotransferase Level  1230 Alkaline Phosphatase  1231 Approach to the Patient with an Elevated Alkaline Phosphatase Level  1232 Tests of Hepatic Synthetic Function  1234 Albumin  1234 Prothrombin Time  1234

When appropriately ordered and interpreted, serum biochemical tests, the so-called “liver function tests” or “liver chemistries,” can be useful in the evaluation and management of patients with liver disorders. The term liver biochemical tests is preferable to liver function tests because the most commonly used tests—the aminotransferases and alkaline phosphatase—do not measure a known function of the liver. These tests have the potential to identify liver disease, distinguish among types of liver disorders, gauge the severity and progression of liver dysfunction, and monitor response to therapy. Understanding the shortcomings of these tests, however, is important. No test can accurately assess the liver’s total functional capacity; biochemical tests measure only a few of the thousands of biochemical functions performed by the liver. Further, considered individually, these tests lack sensitivity and specificity for liver injury; a battery of tests must be used to evaluate the liver. The standard battery of tests that is most helpful in assessing liver disease includes total and direct bilirubin, albumin, prothrombin time, and the serum enzymes: alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), and occasionally gamma glutamyl transpeptidase (GGTP) and 5′ nucleotidase (5′NT). Inter­ pretation of these results in concert with careful history taking and a physical examination may suggest a specific type of liver injury, thereby allowing a directed evaluation, risk assessment for surgical procedures, and estimation of prognosis. In addition to the most commonly used tests, other more specialized tests include quantitative tests of liver function and a growing number of options to assess the degree of hepatic fibrosis. This chapter reviews commonly used and some emerging tests individually and provides a framework for interpreting them.

Tests to Detect Hepatic Fibrosis  1234 Quantitative Liver Function Tests  1235 Indocyanine Green Clearance  1235 Galactose Elimination Capacity  1235 Caffeine Clearance  1235 Lidocaine Metabolite Formation  1235 Aminopyrine Breath Test  1235 Bile Acids  1236 Specific Applications of Liver Biochemical Testing  1236 Drug-Induced Liver Injury  1236 Surgical Candidacy and Organ Allocation  1236

BILIRUBIN (see Chapter 20) BILIRUBIN METABOLISM

Bilirubin is a breakdown product of heme (ferroprotoporphyrin IX). About 4 mg/kg body weight of bilirubin is produced each day, nearly 80% from the breakdown of hemoglobin in senescent red blood cells and prematurely destroyed erythroid cells in the bone marrow and the remainder from the turnover of hemoproteins such as myoglobin and cytochromes distributed throughout the body.1 The initial steps of bilirubin metabolism occur in reticuloendothelial cells, predominantly in the spleen. Heme is converted to biliverdin by the microsomal enzyme heme oxygenase. Biliverdin is then converted to bilirubin by the cytosolic enzyme biliverdin reductase. Bilirubin formed in the reticuloendothelium is lipid soluble and virtually insoluble in water. In order to be transported in blood, this unconjugated bilirubin must be solubilized. The process is initiated by reversible, noncovalent binding to albumin, which has both high-affinity and lower-affinity binding sites for unconjugated bilirubin. The unconjugated bilirubin-albumin complex readily passes through the fenestrations in the endothelium lining the hepatic sinusoids into the space of Disse, where the bilirubin dissociates from albumin and is taken up by the hepatocytes via a protein-mediated, facilitated process, possibly mediated by a liver-specific organic anion transport protein. After entering the hepatocyte, unconjugated bilirubin is bound in the cytosol to a number of proteins, including proteins in the glutathione S-transferase superfamily.2 These proteins serve to reduce efflux of bilirubin back into the serum and present the bilirubin for conjugation. The enzyme uridine-5′-diphosphate (UDP) glucuronyl transfer-

1227

1228

Section IX  Liver ase found in the endoplasmic reticulum solubilizes bilirubin by conjugating it to glucuronic acid to produce bilirubin monoglucuronide and diglucuronide.3 The now hydrophilic bilirubin diffuses to the canalicular membrane for excretion into the bile canaliculi. Conjugated bilirubin is transported across the canalicular membrane by the multiple drug resistance-associated protein 2 (MRP2) via an adenosine triphosphate (ATP)-dependent process.4 This is the only energy-dependent step in bilirubin metabolism and explains why even patients with fulminant hepatic failure have a predominantly conjugated hyperbilirubinemia. Once in the bile, conjugated bilirubin passes undisturbed until it reaches the distal ileum and colon, where bacteria containing β-glucuronidases hydrolyze conjugated bilirubin to unconjugated bilirubin, which is further reduced by bacteria to colorless urobilinogen.5 The urobilinogen is either excreted unchanged, oxidized and excreted as urobilin, which has an orange color, or absorbed passively by the intestine into the portal system as urobilinogen. The majority of the absorbed urobilinogen is re-excreted by the liver. A small percentage filters across the renal glomerulus and is excreted in urine. Unconjugated bilirubin is never found in urine because in the serum it is bound to albumin and not filtered by the glomerulus. The presence of bilirubin in urine indicates a conjugated hyperbilirubinemia and hepatobiliary disease.

MEASUREMENT OF SERUM BILIRUBIN

The terms direct and indirect bilirubin, which correspond roughly to conjugated and unconjugated bilirubin, respectively, derive from the original van den Bergh reaction.6 Serum bilirubin is still measured in clinical laboratories by some modification of this diazo reaction.7 In this assay, bilirubin is exposed to diazotized sulfanilic acid. The conjugated fraction of bilirubin reacts promptly, or “directly,” with the diazo reagent without the need for an accelerant and thereby allows measurement of the con­ jugated bilirubin fraction by photometric analysis within 30 to 60 seconds. The total bilirubin is measured 30 to 60 minutes after the addition of an accelerant such as alcohol or caffeine. The unconjugated, or indirect, fraction is then determined by subtracting the direct component from the total bilirubin. Newer and more accurate methods of measuring bilirubin, such as high-performance liquid chromatography, have been developed but are not generally available because they are more difficult to perform and do not add additional information beyond that provided by the diazo method in most clinical situations. These new methods allow the identification of delta bilirubin, a conjugated bilirubin tightly linked to albumin through covalent binding. Delta bilirubin is found in cases of prolonged and severe elevation of serum conjugated bilirubin levels, and because of the strength of the covalent binding, delta bilirubin has the half-life of albumin, approximately 14 to 21 days, which far exceeds the usual half-life of bilirubin in serum of 4 hours. The identification of delta bilirubin explains why the decline in serum bilirubin in some patients with prolonged jaundice seems to lag behind clinical recovery and why some patients with conjugated hyperbilirubinemia do not have bilirubinuria. Using the diazo method, normal values of total serum bilirubin are between 1.0 and 1.5  mg/dL, with 95% of a normal population falling between 0.2 and 0.9  mg/dL.8 Normal values for the indirect component are between 0.8 and 1.2  mg/dL. The diazo method, however, tends to overestimate the amount of conjugated bilirubin, particularly within the normal range. As a result, “normal” ranges

for conjugated bilirubin have crept upward over time. In general, if the direct acting fraction is less than 15% of the total, the bilirubin can be considered to be entirely indirect. The most frequently reported upper limit of normal for conjugated bilirubin is 0.3  mg/dL. The presence of even a mild increase in conjugated bilirubin in the serum should raise the possibility of liver injury. The measurement and fractionation of serum bilirubin in patients with jaundice does not allow differentiation between parenchymal (hepatocellular) and obstructive (cholestatic) jaundice. The magnitude and duration of hyperbilirubinemia have not been critically assessed as prognostic markers. In general, the higher the serum bilirubin level in patients with viral hepatitis, the greater the hepatocellular damage and the longer the course of disease. Patients may die, however, of acute liver failure with only a modest elevation of serum bilirubin. Total serum bilirubin correlates with poor outcomes in alcoholic hepatitis and is a critical component of the model for end-stage liver disease (MELD) score, which is used to estimate survival of patients with end-stage liver disease (see later and Chapter 95).

APPROACH TO THE PATIENT WITH AN ELEVATED BILIRUBIN LEVEL

Hyperbilirubinemia may be the result of overproduction of bilirubin through excessive breakdown of hemoglobin; impaired hepatocellular uptake, conjugation, or excretion of bilirubin; or regurgitation of unconjugated and conjugated bilirubin from damaged hepatocytes or bile ducts. The presence of conjunctival icterus suggests a total serum bilirubin level of at least 3.0  mg/dL but does not allow differentiation between conjugated and unconjugated hyperbilirubinemia. Tea- or cola-colored urine may indicate the presence of bilirubinuria and thus conjugated hyperbilirubinemia. The evaluation of the patient with an isolated elevation of the serum bilirubin level is quite different from that of the patient with an elevated bilirubin associated with elevated liver enzyme levels; the latter suggests either a hepatocellular or cholestatic process, as discussed later. The first step in the evaluation of a patient with an isolated elevation of the serum bilirubin level is to fractionate the bilirubin to determine if it is conjugated or unconjugated bilirubin (Fig. 73-1). If less than 15% of the total is conjugated, one can be assured that virtually all the serum bilirubin is unconjugated. An overproduction of bilirubin as a result of excessive breakdown of hemoglobin can occur with any of a number of inherited or acquired disorders (Table 73-1). The patient’s medication history should be reviewed for drugs that can cause impaired hepatocellular uptake of bilirubin. If no cause is identified, a genetic enzyme deficiency that results in impaired conjugation of bilirubin, the most common of which is Gilbert’s syndrome, is likely. As discussed in Chapter 20, Gilbert’s syndrome is common, with a reported incidence of 6% to 12% (see Table 20-2). A mutation in the TATAA element in the 5′ promoter region of the UDP glucuronyl transferase gene results in a reduction in enzyme activity to approximately one third of normal. The mildly elevated indirect serum hyperbilirubinemia seen in Gilbert’s syndrome is of no clinical consequence. This benign clinical course contrasts with those of much rarer conditions, Crigler-Najjar syndrome, types I and II (see Table 20-2). The mutations in these conditions result in significantly reduced UDP glucuronyl transferase activity: <10% in Crigler-Najjar type II and complete absence of enzyme activity in Crigler-Najjar type I, resulting in much

Chapter 73  Liver Chemistry and Function Tests Fractionate bilirubin

Table 73-1  Causes of Isolated Hyperbilirubinemia in Adults

≥15% direct (often >50% direct)

CAUSE

<15% direct

Dubin-Johnson or Rotor’s syndrome

Evaluate for hemolysis: haptoglobin, LDH, peripheral blood smear Negative

Positive

Review drugs: (e.g., rifampin, probenecid)

Hemolytic disorders (See Table 73-1)

No culprit medications Inherited disorders: Gilbert’s or Crigler-Najjar syndrome, types 1 or 2 Figure 73-1.  Evaluation of an isolated elevation of the serum bilirubin level. LDH, lactate dehydrogenase.

greater elevations of unconjugated serum bilirubin to levels that carry an increased risk of kernicterus. When an isolated hyperbilirubinemia is associated with a conjugated fraction of >15%, and typically >50%, the diagnosis is either the uncommon Dubin-Johnson syndrome or the even rarer Rotor’s syndrome (see Fig. 73-1, Table 20-2, and Table 64-4). The defect in Dubin-Johnson syndrome is in the MRP2 gene. The defect in Rotor’s syndrome has yet to be defined, but in both syndromes excretion of conjugated bilirubin across the bile canalicular membrane is reduced, resulting in an increased conjugated serum bilirubin level. Neither syndrome is associated with adverse clinical outcomes. Additional genetic disorders of bile acid transport that may be associated with hyperbilirubinemia are discussed in Chapters 64 and 76.

AMINOTRANSFERASES The serum aminotransferases (also called transaminases), the most sensitive markers of acute hepatocellular injury, have been used to identify liver disease since the 1950s.9 ALT (formerly serum glutamic pyruvic transaminase, or SGPT) and AST (formerly serum glutamic oxaloacetic transaminase, or SGOT) catalyze the transfer of the α-amino groups of alanine and l-aspartic acid, respectively, to the α-keto group of ketoglutaric acid. AST, found in cytosol and mitochondria, is widely distributed throughout the body; it is found, in order of decreasing concentration, in liver, cardiac muscle, skeletal muscle, kidney, brain, pancreas, lung, leukocytes, and erythrocytes. ALT, a cytosolic enzyme also found in many organs, is present in greatest concentration by far in the liver and is, therefore, a

Indirect Hyperbilirubinemia Hemolytic Disorders   Inherited    Red cell enzyme defects (e.g., glucose-6-phosphate dehydrogenase deficiency)    Sickle cell disease    Spherocytosis and elliptocytosis   Acquired    Drugs and toxins    Hypersplenism    Immune mediated    Paroxysmal nocturnal hemoglobinuria    Traumatic: macro- or microvascular injury Ineffective Erythropoiesis   Cobalamin deficiency   Folate deficiency   Profound iron deficiency   Thalassemia Drugs: Rifampin, Probenecid Inherited Conditions   Crigler-Najjar syndrome types I and II   Gilbert’s syndrome Other   Hematoma Direct Hyperbilirubinemia Inherited Conditions   Dubin-Johnson syndrome   Rotor’s syndrome

MECHANISM Overproduction of bilirubin

Overproduction of bilirubin

Impaired hepatocellular uptake Impaired conjugation of bilirubin

Overproduction of bilirubin Impaired excretion of conjugated bilirubin

more specific indicator of liver injury. Increases in serum values of the aminotransferases reflect either damage to tissues rich in these enzymes or changes in cell membrane permeability that allow ALT and AST to leak into serum; hepatocyte necrosis is not required for the release of aminotransferases, and the degree of elevation of the aminotransferases does not correlate with the extent of liver injury.10 Aminotransferases have no function in serum and act like other serum proteins. They are distributed in plasma and interstitial fluid and have half-lives measured in days. The activity of ALT and AST at any moment reflects the relative rate at which they enter and leave the circulation. They are probably cleared by cells in the reticuloendothelial system, with AST cleared more rapidly than ALT. Normal values for aminotransferases in serum vary widely among laboratories, but values gaining general acceptance are <30  U/L for men and <19 U/L for women. The inter-laboratory variation in the normal range is the result of technical issues; no reference standards exist to establish the upper limits of normal for serum ALT and AST levels. Therefore, each reference laboratory is responsible for identifying a locally defined reference population or for using a normal range first established in the 1950s.9 The normal range is defined as the mean of the reference population plus 2 standard deviations; approximately 95% of a uniformly distributed population will fall within this “normal” range. Some investigators have recommended revisions of normal values for the aminotransferases with adjustments for sex and body

1229

1230

Section IX  Liver mass index, but others have raised concern about the potential costs and unclear benefits of implementing such a change.11-15 A serum aminotransferase level below normal is of no clinical importance. It has been reported in patients with chronic kidney disease on hemodialysis and is believed to be caused in part by vitamin B6 deficiency.

APPROACH TO THE PATIENT WITH AN ELEVATED AMINOTRANSFERASE LEVEL

Serum aminotransferase levels are typically elevated in all forms of liver injury; levels up to 300 U/L are nonspecific. In certain circumstances the degree and pattern of elevation of the aminotransferases, evaluated in the context of a patient’s characteristics, symptoms, and physical examination findings, can suggest particular diagnoses and direct the subsequent evaluation (Table 73-2). The differential diagnosis of marked elevations of aminotransferase levels (>1000 U/L) includes viral hepatitis (A to E), toxin or druginduced liver injury, ischemic hepatitis, and less commonly, autoimmune hepatitis, acute Budd-Chiari syndrome, fulminant Wilson disease, and acute obstruction of the biliary tract.

Table 73-2  Causes of Elevated Serum Aminotransferase Levels* Chronic, Mild Elevations, ALT > AST (<150 U/L or 5 × normal) Hepatic Causes a1-antitrypsin deficiency Autoimmune hepatitis Chronic viral hepatitis (B, C, and D) Hemochromatosis Medications and toxins Steatosis and steatohepatitis Wilson disease Nonhepatic Causes Celiac disease Hyperthyroidism Severe, Acute Elevations, ALT > AST (>1000 U/L or >20-25 × normal) Hepatic Causes Acute bile duct obstruction Acute Budd-Chiari syndrome Acute viral hepatitis Autoimmune hepatitis Hepatic artery ligation Ischemic hepatitis Medications/toxins Wilson disease Severe, Acute Elevations, AST > ALT (>1000 U/L or >20-25 × normal) Hepatic Cause Medications or toxins in a patient with underlying alcoholic liver injury Nonhepatic Cause Acute rhabdomyolysis Chronic, Mild Elevations, AST > ALT (<150 U/L, <5 × normal) Hepatic Causes Alcohol-related liver injury (AST : ALT > 2 : 1, AST nearly always <300 U/L) Cirrhosis Nonhepatic Causes Hypothyroidism Macro-AST Myopathy Strenuous exercise ALT, alanine aminotransferase; AST, aspartate aminotransferase. *Virtually any liver disease can cause moderate aminotransferase elevations (5-15 × normal)

The ratio of AST to ALT in serum is helpful in a few specific circumstances—perhaps most importantly in the recognition of alcoholic liver disease. If the AST level is less than 300 U/L, a ratio of AST to ALT of more than 2 suggests alcoholic liver disease, and a ratio of more than 3 is highly suggestive of alcoholic liver disease.16 The ratio results from a deficiency of pyridoxal 5′-phosphate in patients with alcoholic liver disease; ALT synthesis in the liver requires pyridoxal phosphate more than does AST synthesis.17 When a patient with chronic alcoholic liver disease sustains a superimposed liver injury, particularly acetaminophen hepatotoxicity, the aminotransferase level can be strikingly elevated, yet the AST/ALT ratio is maintained. An increased ratio of AST to ALT may also be seen in muscle disorders. The degree of elevation is typically less than 300  U/L, but in rare cases, such as rhabdomyolysis, levels observed in patients with acute hepatocellular disease can be reached. In cases of acute muscle injury, the AST/ ALT ratio may initially be greater than 3 : 1, but the ratio quickly declines toward 1 : 1 because of the shorter serum half-life of AST.18 The ratio typically is close to 1 : 1 in patients with chronic muscle diseases. Although the AST to ALT ratio is typically less than 1 in patients with chronic viral hepatitis and nonalcoholic fatty liver disease (NAFLD), a number of investigators have observed that, as cirrhosis develops, the ratio rises and may become greater than 1. Studies have shown that an AST/ ALT ratio of greater than 1 as an indicator of cirrhosis in patients with chronic hepatitis C has a high specificity (94% to 100%) but a relatively low sensitivity (44% to 75%).19 The increase in AST/ALT ratio with the development of cirrhosis is believed to result from impaired functional hepatic blood flow, with a consequent decrease in hepatic sinusoidal uptake of AST.20 The majority of patients evaluated for elevated serum aminotransferase levels are asymptomatic and have mild elevations (≤5-fold) identified during routine screening. The first step in the evaluation of mildly elevated serum aminotransferase levels is to repeat the test to confirm persistence of the elevated value. If the aminotransferase level remains elevated, the recommended evaluation is illustrated in Figure 73-2. The initial step is to take a careful history focused on identifying all of the patient’s medications, including over-the-counter (OTC) medications, complementary and alternative medications (CAM), and substances of abuse. Correlating the use of medications temporally with the laboratory abnormalities will sometimes reveal a specific culprit. Almost any medication, including OTC medications, CAM, and substances of abuse, has the potential to elevate serum aminotransferase levels. Relatively common offending agents include nonsteroidal anti-inflammatory drugs, antibiotics, hydroxymethylglutaryl-coenzyme A reductase inhibitors, antiepileptics, and antituberculous medications (see Chapter 86). The association between use of a medication and liver enzyme elevations is readily established by stopping the medication and observing return of the enzyme levels to normal. Re-challenge with the suspect medication followed by a rise in serum aminotransferase levels is confirmatory but often not undertaken. Muscle disease should also be excluded by obtaining serum creatine kinase and aldolase levels. The next step in the evaluation is to assess the patient for the more common and treatable causes of liver disease, including chronic hepatitis B and C, hemochromatosis, autoimmune hepatitis, Wilson disease, and NAFLD. Although autoimmune hepatitis is commonly considered a disease of young to middle-aged women, it also is seen in men and has been reported in all ethnic groups (see Chapter

Chapter 73  Liver Chemistry and Function Tests History: review medications, including OTC and CAM drugs, alcohol, and illicit drug use

CK and aldolase to exclude muscle disease

HBeAg, anti-HBe HBV DNA

Liver biopsy Ophthalmoscopic examination, 24-hour urine for copper Liver biopsy

History unrevealing Anti-HCV

HBsAg+

ANA+ and/or SMA+ ↑ Globulins

Stop potential toxins Follow response of aminotransferase levels

HBsAg, Anti-HBs, Anti-HBc, Anti-HCV, Iron Iron, TIBC saturation >45% ANA, SMA, SPEP Ceruloplasmin (age <40) RUQ US Steatosis on US

↓ Ceruloplasmin

Evaluation unrevealing HCV RNA α1-AT phenotype TFTs TTG antibody

Evaluation unrevealing

HCV RNA HCV genotype Ferritin HFE gene testing Liver biopsy Recommend weight loss

Liver biopsy if aminotransferase levels >2 times ULN

Figure 73-2.  Evaluation of asymptomatic elevation of serum aminotransferase levels. a1-AT, a1-antitrypsin; ANA, antinuclear antibodies; Anti-HBc, antibody to hepatitis B core antigen; Anti-HBe, antibody to hepatitis B e antigen; Anti-HBs, antibody to hepatitis B surface antigen; Anti-HCV, antibody to hepatitis C virus; CAM, complementary and alternative medicines; CK, creatine kinase; HBeAg, hepatitis B e antigen; HBV DNA, hepatitis B virus deoxyribonucleic acid; HBsAg, hepatitis B surface antigen; HFE, hemochromatosis; OTC, over-the-counter; RNA, ribonucleic acid; RUQ US, right upper quadrant ultrasound; SMA, smooth muscle antibodies; SPEP, serum protein electrophoresis; TIBC, total iron binding capacity, TFTs, thyroid function tests; TTG, tissue transglutaminase; ULN, upper limit of normal.

88). The clinical onset of Wilson disease is usually between the ages of 5 and 25 years; the diagnosis should be considered initially in all patients age 40 or younger and those older than age 40 with aminotransferase elevations that remain unexplained after other causes are excluded (see Chapter 75). NAFLD is the most common cause of elevated serum aminotransferase levels in the United States (see Chapter 85). If testing for the disorders listed earlier fails to provide a diagnosis, the less common causes of liver disease, such as alpha-1 antitrypsin deficiency, and extrahepatic causes of persistently elevated liver enzyme levels, such as thyroid disease and celiac disease, should be sought. If testing for these disorders is negative, the decision to perform a liver biopsy is determined by the degree of aminotransferase elevation, with the recognition that the results of the biopsy are unlikely to alter management.

ALKALINE PHOSPHATASE The term alkaline phosphatase applies generally to a group of isoenzymes distributed widely throughout the body.21 The isoenzymes of greatest clinical importance in adults are in the liver and bone because these organs are the major sources of serum alkaline phosphatase. Other isoenzymes originate from the placenta, small intestine, and kidneys. In the liver, alkaline phosphatase is found on the canalicular membrane of hepatocytes; its precise function is undefined. Alkaline phosphatase has a serum half-life of approximately seven days, and although the sites of degradation are unknown, clearance of alkaline phosphatase from serum is

independent of either patency of the biliary tract or functional capacity of the liver. Hepatobiliary disease leads to increased serum alkaline phosphatase levels through induced synthesis of the enzyme and leakage into the serum, a process mediated by bile acids.22 A number of individual physiologic variations in serum alkaline phosphatase levels have been identified. Patients with blood groups O and B have elevations in serum alkaline phosphatase levels caused by release of intestinal alkaline phosphatase after a fatty meal.23 This observation is the basis for the recommendation by some authorities that the serum alkaline phosphatase level be checked in the fasting state. An increased serum alkaline phosphatase level of intestinal origin is seen in a benign familial elevation of serum alkaline phosphatase. Serum alkaline phosphatase values vary with age. Male and female adolescents have serum alkaline phosphatase levels twice the level seen in adults; the level correlates with bone growth, and the increase in serum is in bone alkaline phosphatase. Although the level of serum alkaline phosphatase increases after age 30 years in both men and women, the increase is more pronounced in women than in men; a healthy 65-year-old woman has a serum alkaline phosphatase level 50% higher than that of a healthy 30-year-old woman.24 The reason for this difference is not known. In a person with isolated elevation of the serum alkaline phosphatase level, the serum GGTP or 5′NT are used to distinguish a liver origin from bone origin of the alkaline phosphatase elevation. A low serum alkaline phosphatase level may occur in patients with Wilson disease, especially those presenting with fulminant hepatitis and hemolysis, possibly because of reduced activity of the enzyme owing to displacement of the co-factor zinc by copper (see Chapter 75).

1231

1232

Section IX  Liver Gamma Glutamyl Transpeptidase

Gamma glutamyl transpeptidase (GGTP) is found in the cell membranes of a wide distribution of tissues including liver (both hepatocytes and cholangiocytes), kidney, pancreas, spleen, heart, brain, and seminal vesicles. It is present in the serum of healthy persons. Serum levels are not different between men and women and do not rise in pregnancy. Although an elevated serum GGTP level has high sensitivity for hepatobiliary disease, its lack of specificity limits its clinical utility. The primary use of serum GGTP levels is to identify the source of an isolated elevation in the serum alkaline phosphatase level; GGTP is not elevated in bone disease (Fig. 73-3).25 GGTP is elevated in patients taking phenytoin, barbiturates, and some drugs used in highly active antiretroviral therapy, including non-nucleoside reverse transcriptase inhibitors and the protease inhibitor abacavir.26,27 Serum GGTP levels are also elevated in patients who drink alcohol, and some experts have advocated use of the GGTP level for identifying unreported alcohol use (see Chapter 84). The sensitivity of an elevated serum GGTP level for alcohol use ranges from 52% to 94%, but a low specificity limits its usefulness for this purpose.28 One study has suggested an association between high serum GGTP levels and the risk of hepatocellular carcinoma.29 The GGTP level had a negative predictive value of 97.9%, higher than that for alkaline phosphatase, total bilirubin, ALT, and AST, for detecting bile duct stones in patients undergoing laparoscopic cholecystectomy.30

5′-Nucleotidase

5′-Nucleotidase (5′NT) is associated with the canalicular and sinusoidal plasma membranes. Its function is undefined. 5′NT is also found in the intestine, brain, heart, blood vessels, and endocrine pancreas. Serum levels of 5′NT are unaffected by sex or race, but age affects the level; values are lowest in children and increase gradually, reaching a plateau at approximately age 50 years. As with GGTP, the primary role of the serum 5′NT level is to identify the organ source of an isolated serum alkaline phosphatase elevation (see Fig. 73-3). The 5′NT level is not increased in bone disease and is primarily increased in hepatobiliary disease.

Fractionate ALP Bone origin or measure GGTP or 5´NT level

APPROACH TO THE PATIENT WITH AN ELEVATED ALKALINE PHOSPHATASE LEVEL

The first step in the evaluation of a patient with an isolated and asymptomatic elevation of the serum alkaline phosphatase is to identify the tissue source (see Fig. 73-3). The most precise way of doing this is via fractionation through electrophoresis; each isoenzyme of alkaline phosphatase has different electrophoretic mobilities.31 Tests used in the past involving heat and urea denaturation of alkaline phosphatase are neither sensitive nor specific. An acceptable alternative method is to check either the serum GGTP or 5′NT level; elevation of either verifies that the elevated alkaline phosphatase is the result of hepatobiliary disease. A normal 5′NT level, however, does not rule out the possibility of hepatobiliary disease because the 5′NT and alkaline phosphatase do not necessarily increase in parallel in early or mild hepatic injury, thus making GGTP the preferred test. The primary value of an elevated serum level of alkaline phosphatase of liver origin is to allow the recognition of cholestatic disorders (i.e., disorders associated with impaired bile flow, often with jaundice). In general, a serum alkaline phosphatase elevation out of proportion to the level of the aminotransferases suggests a cholestatic disorder. A four-fold elevation of the serum alkaline phosphatase is seen in approximately 75% of patients with chronic cholestasis, both intrahepatic or extrahepatic, whereas lesser elevations are nonspecific and can occur in a wide range of conditions. Figures 73-3 and 73-4 illustrate the recommended evaluation of cholestatic liver enzymes— either an isolated alkaline phosphatase elevation (see Fig. 73-3) or a disproportionate elevation of the alkaline phosphatase compared with the aminotransferases (see Fig. 73-4). Central to the evaluation of an elevated alkaline phosphatase level is imaging of the biliary tract. An absence of dilated intrahepatic bile ducts focuses the search on intrahepatic causes of cholestasis (Table 73-3), whereas dilated ducts should lead to an evaluation of extrahepatic causes of cholestasis (Table 73-4). As with elevated aminotransferase levels, the evaluation of intrahepatic causes of cholestatic

Review drug list RUQ US

Bone evaluation

Ducts not dilated: intrahepatic cholestasis

Ducts dilated: extrahepatic cholestasis

AMA, ACE levels Viral serologies: HAV, HBV, HCV, EBV, CMV

CT ERCP

Liver origin Review drug list AMA ACE level RUQ US AMA+

ACE level elevated

Consider liver biopsy

Evaluation unrevealing MRCP Liver biopsy

Bile ducts dilated ERCP CT

Figure 73-3.  Evaluation of an isolated elevation of the serum alkaline phosphatase level. ACE, angiotensin-converting enzyme; ALP, alkaline phosphatase; AMA, antimitochondrial antibodies; CT, computed tomog­ raphy; ERCP, endoscopic retrograde cholangiopancreatography; GGTP, gamma glutamyl transpeptidase; MRCP, magnetic resonance cholan­ giopancreatography; 5′NT, 5′ nucleotidase; RUQ US, right upper quadrant ultrasound.

AMA+

Evaluation negative

Liver biopsy

MRCP Liver biopsy

Figure 73-4.  Evaluation of cholestatic liver enzyme elevations. ACE, angiotensin-converting enzyme; AMA, antimitochondrial antibodies; CMV, cytomegalovirus; CT, computed tomography; EBV, Epstein-Barr virus; ERCP, endoscopic retrograde cholangiopancreatography; HAV, hepatitis A virus; HBV, hepatitis B virus; HCV, hepatitis C virus; MRCP, magnetic resonance cholangiopancreatography; RUQ US, right upper quadrant ultrasound.

Chapter 73  Liver Chemistry and Function Tests Table 73-3  Intrahepatic Causes of Cholestatic Liver Enzyme Elevations in Adults

Table 73-4  Extrahepatic Causes of Cholestatic Liver Enzymes in Adults

Drugs* Bland cholestasis Anabolic steroids Estrogens Cholestatic hepatitis Angiotensin-converting enzyme inhibitors: captopril, enalapril Antimicrobials: amoxicillin-clavulanic acid, ketoconazole Azathioprine Chlorpromazine Nonsteroidal anti-inflammatory drugs: sulindac, piroxicam Granulomatous hepatitis Allopurinol Antibiotics: sulfonamides Antiepileptics: carbamazepine, phenytoin Cardiovascular agents: hydralazine, procainamide, quinidine Phenylbutazone Vanishing bile duct syndrome Amoxicillin-clavulanic acid Chlorpromazine Dicloxacillin Erythromycins Flucloxacillin Primary Biliary Cirrhosis Primary Sclerosing Cholangitis Granulomatous Liver Disease Infections Brucellosis Fungal: histoplasmosis, coccidioidomycosis Leprosy Q fever Schistosomiasis Tuberculosis, Mycobacterium avium complex, bacillus Calmette-Guérin Sarcoidosis Idiopathic granulomatous hepatitis Other Crohn’s disease Heavy metal exposure: beryllium, copper Hodgkin’s disease Viral Hepatitis Hepatitis A Hepatitis B and C, including fibrosing cholestatic hepatitis Epstein-Barr virus Cytomegalovirus Idiopathic Adult Ductopenia Genetic Conditions Progressive familial intrahepatic cholestasis Type 1 (Byler’s disease) Type 2 Type 3 Benign recurrent intrahepatic cholestasis Type 1 Type 2 Cystic fibrosis Malignancy Hepatocellular carcinoma Metastatic disease Paraneoplastic syndrome Non-Hodgkin’s lymphoma Prostate cancer Renal cell cancer Infiltrative Liver Disease Amyloidosis Lymphoma Intrahepatic Cholestasis of Pregnancy Total Parenteral Nutrition Graft-versus-Host Disease Sepsis

Intrinsic Choledocholithiasis Immune-Mediated Duct Injury Autoimmune pancreatitis Primary sclerosing cholangitis Malignancy Ampullary cancer Cholangiocarcinoma Infections AIDS cholangiopathy Cytomegalovirus Cryptosporidiosis Microsporidiosis Parasitic infections Ascariasis Extrinsic Malignancy Gallbladder cancer Metastases, including portal adenopathy from metastases Pancreatic cancer Mirizzi’s syndrome* Pancreatitis Pancreatic pseudocyst

*Categorized by histologic pattern. Drug lists are not meant to be comprehensive.

*Compression of the common hepatic duct by a stone in the neck of the gallbladder. AIDS, acquired immunodeficiency syndrome.

liver enzymes begins with a careful history of medication use, including OTC medications, CAM, and drugs of abuse, and temporal correlation of their use with elevation of the liver enzymes. A liver biopsy generally is not required for the diagnosis; withdrawal of the offending agent and resolution of the liver enzymes is sufficient to confirm the diagnosis. The rate of improvement can be slow, and if bile duct destruction has developed (“vanishing bile duct syndrome”), the changes may be irreversible. Primary biliary cirrhosis (PBC) is a classic autoimmune disease. The immunologic injury is characterized by a T cell–mediated destruction of the intrahepatic bile ducts. Although predominantly a disease of middle-aged women with a median age at diagnosis of approximately 50 years, 5% to 10% of affected patients are men. The reported age range is 22 to 93 years. Antimitochondrial antibodies (AMA) are found in serum in 95% of patients and are diagnostic; a liver biopsy that demonstrates characteristic histologic findings is confirmatory (see Chapter 89). Primary sclerosing cholangitis (PSC) is a disease of altered immunity marked by inflammation and fibrosis of the intraor extrahepatic bile ducts, or both. The disorder is strongly associated with inflammatory bowel disease and is found most commonly in younger men. The diagnosis is confirmed by cholangiography, either endoscopic retrograde cholangiopancreatography (ERCP) or magnetic resonance cholangiopancreatography (MRCP) (see Chapter 68). Granulomatous liver disease can be caused by a number of disorders (see Table 73-3). Infectious etiologies must be excluded because treatment for many of the other causes of granulomatous liver disease is immunosuppressive therapy. Sarcoidosis is the most common etiology. The diagnosis is based on an elevated angiotensin-converting enzyme (ACE) level and typical extrahepatic manifestations. Hepatic involvement, however, is uncommonly the impetus for initiating therapy for sarcoidosis (see Chapter 35). Viral hepatitis, particularly cases cased by Epstein-Barr virus (EBV) and cytomegalovirus (CMV), can manifest with

1233

1234

Section IX  Liver a prominent cholestatic liver enzyme pattern (see Chapter 81). A number of familial conditions produce intrahepatic cholestasis (see Table 64-5). Progressive forms of these disorders manifest in childhood, whereas the benign forms— benign recurrent intrahepatic cholestasis types 1 and 2—can manifest for the first time in adulthood. Other intrahepatic causes of cholestatic liver enzymes are listed in Table 73-3. If imaging shows intrahepatic ductal dilatation, the evaluation focuses on the extrahepatic biliary tract to identify an intrinsic or extrinsic process causing biliary obstruction (see Table 73-4). The evaluation often includes an ERCP for tissue acquisition and placement of a biliary stent if obstruction is present (see Chapter 70). Computed tomography (CT) provides assessment for an extrinsic process, and tissue acquisition can be performed with CT or endoscopic ultrasound guidance.

TESTS OF HEPATIC SYNTHETIC FUNCTION ALBUMIN

Quantitatively, the most important plasma protein, albumin, accounts for 75% of the plasma colloid oncotic pressure and is synthesized exclusively by hepatocytes. The average adult produces approximately 15 g/day and has 300 to 500 g of albumin distributed in body fluids. The liver has the ability to double the rate of synthesis in the setting of rapid albumin loss or a dilutional decrease in the serum albumin concentration.32 The half-life of albumin is 14 to 21 days; the site of degradation is not known. Albumin synthesis is regulated by changes in nutritional status, osmotic pressure, systemic inflammation, and hormone levels.33 Thus, the differential diagnosis of serum hypoalbuminemia, in addition to hepatocellular dysfunction, includes malnutrition, excessive loss from protein-losing enteropathy or nephrotic syndrome, chronic systemic inflammatory conditions, and hormonal imbalances. The long serum half-life of albumin in serum accounts for its unreliability as a marker of hepatic synthetic function in acute liver injury. Serum albumin levels less than 3 g/dL in a patient with newly diagnosed hepatitis should raise suspicion of a chronic process. Serum albumin is an excellent marker of hepatic synthetic function in patients with chronic liver disease and cirrhosis, with the exception of patients with cirrhosis and ascites, who may have normal or increased albumin production but an increased volume of distribution that results in a low serum albumin level. Albumin has no utility as a screening test in patients for whom there is low suspicion of liver disease; a study in which the serum albumin level was measured in 449 consecutive patients yielded 56 abnormal results, of which only 2 (0.4%) were of clinical importance.34

PROTHROMBIN TIME

Clotting is the end result of a complex series of enzymatic reactions involving clotting factors, all of which are produced in the liver except factor VIII, which is produced by vascular endothelial cells. The prothrombin time is a measure of the rate at which prothrombin is converted to thrombin, reflecting the extrinsic pathway of coagulation. Factors involved in the synthesis of prothrombin include II, V, VII, and X. The international normalized ratio (INR) is used to express the degree of anticoagulation on warfarin therapy. The INR standardizes prothrombin time measurement according to the characteristics of the thromboplastin reagent used in a particular laboratory; the initial measurement is expressed as an international sensitivity index (ISI),

which is then used in calculating the INR. Because the ISI is validated only for patients taking a vitamin K antagonist, concern has been raised about the validity of using the ISI (and INR) in patients with chronic liver disease.35 Two studies have demonstrated, in fact, that the ISI, as currently determined, is not accurate for calculating the INR in patients with cirrhosis, and the investigators have proposed that specific ISI and INR determinations using control patients with liver disease be used to eliminate interlaboratory variability in calculating the INR in patients with cirrhosis.36,37 A prolonged prothrombin time can be caused by a number of conditions besides reduced hepatic synthetic function: congenital deficiency of clotting factors, vitamin K deficiency (vitamin K is required for normal functioning of factors II, VII, IX, and X), and disseminated intravascular coagulation (DIC). DIC can be identified by measuring a factor VIII level in serum; the level is decreased in DIC and normal or increased in liver disease. Vitamin K deficiency is identified by demonstrating that subcutaneous administration of vitamin K (e.g., 10 mg) leads to improvement in the prothrombin time; a 30% or more improvement in the prothrombin time is consistent with hypovitaminosis K. Oral vitamin K is not absorbed by the intestine in patients with jaundice. Measurement of the prothrombin time in patients with liver disease is most useful in cases of acute liver disease. Unlike the serum albumin, the prothrombin time allows an assessment of current hepatic synthetic function; factor VII has the shortest serum half-life (six hours) of all the clotting factors. The prothrombin time has prognostic value in patients with acute acetaminophen- and nonacetaminophen-related liver failure (see Chapter 93), as well as alcoholic hepatitis (see Chapter 84). The INR, along with total serum bilirubin and creatinine levels, are components of the MELD score, which is used to allocate donor organs for liver transplantation. The MELD score accurately predicts survival in patients with decompensated cirrhosis (see later). The prothrombin time is not an accurate measure of bleeding risk in patients with cirrhosis because it assesses only the activity of procoagulant clotting factors, not anticoagulants such as protein C and antithrombin, the production of which is also reduced in cirrhosis. The partial throm­ boplastin time (PTT) assesses the intrinsic pathway of the coagulation cascade. The PTT can be prolonged in patients with advanced cirrhosis, but prolongation of the PTT is less sensitive than the PT for detecting coagulopathy.

TESTS TO DETECT HEPATIC FIBROSIS Although liver biopsy is the standard for the assessment of hepatic fibrosis, noninvasive measures of hepatic fibrosis have been developed and show promise. These measures include single serum biochemical markers that potentially reflect the activity level of hepatic fibrogenesis (hyaluronan is the best to date) and multiparameter tests aimed at detecting and staging the degree of hepatic fibrosis (more than 20 such tests are described in the literature). Hyaluronan is a glucosaminoglycan produced in mesenchymal cells and widely distributed in the extracellular space. Typically degraded by hepatic sinusoidal cells, serum levels of hyaluronan are elevated in patients with cirrhosis as a result of sinusoidal capillarization (see Chapter 90). A fasting hyaluronan level greater than 100 mg/L had a sensitivity of 83% and specificity of 78% for the detection

Chapter 73  Liver Chemistry and Function Tests of cirrhosis in patients with a variety of chronic liver diseases.38 Hyaluronan has been shown to be useful for identifying advanced fibrosis in patients with chronic hepatitis C, chronic hepatitis B, alcoholic liver disease, and nonalcoholic steatohepatitis.39 Preoperative serum hyaluronan levels also have been shown to correlate with the development of hepatic dysfunction after hepatectomy.40 FibroTest (marketed as FibroSure in the United States) is the best evaluated of the multiparameter blood tests. The test incorporates haptoglobin, bilirubin, GGTP, apolipoprotein A-I, and α2-macroglobulin and has been found to have high positive and negative predictive values for diagnosing advanced fibrosis in patients with chronic hepatitis C. One study showed that use of a more sensitive index cut-off had a sensitivity of 90%, specificity of 36%, positive predictive value of 88%, and negative predictive value of 40% for the diagnosis of bridging fibrosis in patients with chronic hepatitis C.41 The test has similar performance characteristics in patients with chronic hepatitis B and alcoholic liver disease and has been shown to predict advanced fibrosis in patients taking methotrexate for psoriasis.42 The newer FIBROSpect II assay incorporates hyaluronate, tissue inhibitor of metalloproteinase 1, and α2-macroglobulin. In a group of patients with chronic hepatitis C, FIBROSpect II had a sensitivity of 72% and a specificity of 74% for identifying advanced fibrosis.43 Transient elastography (TE), marketed as FibroScan, uses ultrasound waves to measure hepatic stiffness noninvasively. Central to this technique’s development was the principle that fibrosis leads to increased stiffness of the hepatic tissue and that a shear wave would propagate faster through stiff material than through elastic material.44 The ultrasound transducer emits a low-frequency (50 Hz) shear wave, and the amount of time required for the wave to go through a set “window” of tissue is measured.45 The window of tissue is 1 cm by 4 cm, 100 times the area of an average liver biopsy. A meta-analysis showed that TE performed best at differentiating cirrhosis from absence of cirrhosis but was less accurate for the estimation of lesser degrees of fibrosis.46 TE has been shown to be accurate for identifying advanced fibrosis in patients with chronic hepatitis C, PBC, hemochromatosis, NAFLD, and recurrent chronic hepatitis after liver transplantation.47-50 Magnetic resonance elastography (MRE) is another noninvasive technique under study. The shear elasticity of the liver is measured after low-frequency (65  Hz) waves are transmitted into the right lobe of the liver. In one study,51 MRE was found to be superior to TE for staging liver fibrosis in patients with a variety of chronic liver diseases.

densitometry and fingertip optical sensors) generate data that appear to correlate well with levels determined by blood sampling. Possible uses of ICG include the assessment of hepatic dysfunction, measurement of hepatic blood flow, and prediction of clinical outcomes in patients with liver disease. Unfortunately, measurement of ICG has proved to be insensitive for detecting hepatic dysfunction and is inaccurate for measuring blood flow in patients with cirrhosis because of decreased ICG extraction by the diseased liver. Although ICG measurement has shown some promise for predicting outcomes in certain clinical situations, it has not been employed widely outside of research protocols.

GALACTOSE ELIMINATION CAPACITY

The galactose elimination capacity has been studied as a measure of functional hepatic mass. Galactose is given as a single intravenous bolus (0.5 g/kg), and blood samples are collected. Patients with cirrhosis and chronic hepatitis have reduced galactose clearance from serum as compared with healthy controls, but the test has proved no better than standard serum testing in differentiating healthy persons from those with liver disease.

CAFFEINE CLEARANCE

Caffeine clearance tests quantify functional hepatic capacity by assessing the activity of cytochrome P450 1A2, Nacetyltransferase, and xanthine oxidase. Caffeine is given orally (200 to 366  mg), and levels are measured in blood, urine, saliva, breath, or scalp hair. The alternative methods correlate well with the plasma clearance method. Tobacco use increases caffeine clearance, and drug interactions can affect results. Increasing age correlates with decreased caffeine clearance. Overnight salivary caffeine clearance has been shown to correlate with ICG measurements and galactose clearance as well as with results of the aminopyrine breath test (see later).52

LIDOCAINE METABOLITE FORMATION

Lidocaine is metabolized to its major metabolite monoethylglycinexylidide (MEGX) by the hepatic cytochrome P450 system.53 Serum samples are taken 15, 30, and 60 minutes after an intravenous administration of lidocaine (1 mg/kg). Neither MEGX nor galactose elimination were found to be superior to the Child-Turcotte-Pugh (CTP) (see Chapter 90) or MELD score in predicting prognosis in patients with cirrhosis secondary to viral hepatitis.54 Other studies have suggested that a decline in MEGX concentration correlates well with histologic worsening in patients with chronic liver disease.55

AMINOPYRINE BREATH TEST

QUANTITATIVE LIVER FUNCTION TESTS Quantitative function tests have been developed in the hope of evaluating the excretory or detoxification capacity of the liver more specifically than the serum bilirubin level. Unfortunately, although these tests lead to improved sensitivity, their lack of specificity and often cumbersome methodology have limited their widespread acceptance, except in research settings.

INDOCYANINE GREEN CLEARANCE

Indocyanine green (ICG) is a nontoxic dye that is cleared exclusively by the liver; 97% of an administered dose (0.646.4  mol/kg given as an intravenous bolus) is excreted unchanged into bile. ICG can be measured directly by spectrophotometry. Noninvasive methods (dichromatic earlobe

The 14C and 13C aminopyrine breath tests measure hepatic mixed-function oxidase mass. The radioactive methyl groups of aminopyrine undergo demethylation and eventual conversion to labeled CO2, which is then exhaled and can be measured. After an overnight fast, a known dose of 14C aminopyrine (1 to 2 µCi) is administered orally, and breath samples are taken every 30 minutes for 4 hours; some investigators check a single sample at either 1 or 2 hours. Healthy subjects excrete 6.6% ± 1.3% of the administered dose in the breath in two hours; patients with hepatocellular injury excrete considerably less. The degree of decrease in excretion of aminopyrine overlaps considerably in patients with all types of severe liver disease, including cirrhosis, chronic hepatitis, alcoholic liver disease, and hepatocel­ lular carcinoma.56 Data are conflicting regarding the ability of this test to predict survival in patients with chronic liver disease.

1235

1236

Section IX  Liver BILE ACIDS Bile acids are synthesized from cholesterol in hepatocytes, conjugated to glycine or taurine, and secreted into bile (see Chapter 64). After passage into the small intestine, most bile acids are actively reabsorbed. The liver efficiently extracts bile acids from the portal blood. In healthy persons, all bile acids in serum emanate from the reabsorption of bile acids in the small intestine. Maintenance of normal serum bile acid concentrations depends on hepatic blood flow, hepatic uptake, secretion of bile acids, and intestinal transit. Serum bile acids are sensitive but nonspecific indicators of hepatic dysfunction and allow some quantification of functional hepatic reserve. Serum bile acid levels correlate moderately well with the results of aminopyrine breath tests in patients with chronic hepatitis.57 Unfortunately, the correlation between serum bile acid levels and the histologic severity of chronic hepatitis and alcoholic liver disease is poor.58 Serum bile acids are elevated in patients with cholestatic liver diseases but normal in patients with Gilbert’s syndrome and Dubin-Johnson syndrome and can be used to make the distinction. Although decreased serum bile acid levels are highly specific indicators of liver dysfunction, they are not as sensitive as initially hoped.

SPECIFIC APPLICATIONS OF LIVER BIOCHEMICAL TESTING Liver biochemical tests have been used to monitor for and assess the severity of drug-induced liver injury, assess operative risk, identify candidates for liver transplantation, and direct donor organ allocation.

DRUG-INDUCED LIVER INJURY

Most drugs that are hepatotoxic cause idiosyncratic liver injury, defined as injury that is unpredictable, occurs at therapeutic drug levels, and is infrequent. The estimated frequency of idiosyncratic drug-induced liver injury for any particular medication ranges from 1 in 1000 to 1 in 100,000. These reactions are marked by a variable latency period ranging from 5 to 90 days, or even longer.59 Other drugs produce dose-dependent toxicity. These injuries are predictable, have a high incidence, and generally have a well understood mechanism. Acetaminophen is the classic example of a drug that causes dose-dependent liver injury. The dose of acetaminophen exceeds 15 g, almost four times the recommended daily dose, in 80% of cases. Acetaminophen doses within the therapeutic range (<4 g/day) can be sufficient to cause liver injury in susceptible persons, such as those who use ethanol chronically. The King’s College criteria identify patients with a poor prognosis from acetaminophen-induced liver injury: those with an arterial pH <7.3 or those with an INR >6.5, serum creatinine level >3.4  g/dL, and stage 3 to 4 hepatic encephalopathy (see Chapter 86).60 Most occurrences of drug-induced liver injury are mild and respond promptly to drug withdrawal with complete resolution. Isolated elevation of the serum aminotransferase levels, even to values greater than three times the upper limit of normal, is associated with a positive outcome. When aminotransferase elevations are associated with clinical jaundice (so-called Hy’s Law, after the late Dr. Hyman Zimmerman), the risk of mortality is increased to as high as 10%.61

SURGICAL CANDIDACY AND ORGAN ALLOCATION

Patients with acute and chronic liver disease are potentially at increased risk of morbidity and mortality if they undergo surgery. The risk depends on the etiology of the liver disease, severity of the liver disease, and planned operation. Although routine preoperative liver biochemical testing is not recommended in otherwise healthy people, the identification of unexpected elevated liver enzyme levels should prompt a postponement of surgery until the cause of the abnormalities has been identified. A retrospective analysis found that patients with acute viral hepatitis who undergo laparotomy had an operative mortality rate of approximately 9.5%.62 Elective surgery should be postponed in patients with acute hepatitis. The surgical risk in patients with chronic hepatitis correlates with the severity of histologic inflammation in the liver. Those with only portal inflam­ mation and interface hepatitis have low operative risk, whereas those with panlobular hepatitis have an increased risk. The etiology of chronic hepatitis does not influence outcome. Examination of histology also is critical in assessing the surgical risk in patients with alcoholic liver disease. Hepatic steatosis alone is associated with a low operative risk, whereas alcoholic hepatitis is associated with a mortality rate as high as 55% in patients undergoing portosystemic shunt surgery, for example. A period of abstinence of three to six months before elective surgery is recommended in these patients. Few data exist for surgical risk in patients with NAFLD, but the mortality rate appears to correlate with the severity of steatosis in patients undergoing liver resection. Whether steatohepatitis carries a higher risk than that for steatosis is not clear. An estimated 10% of patients with advanced liver disease undergo surgery in the last two years of their lives. Cirrhosis is associated with increased operative risk, particularly with certain types of surgery, including cardiothoracic surgery, hepatic resection, and other abdominal operations. The data evaluating the surgical risk in these patients were derived retrospectively but point consistently toward the usefulness of the CTP scoring system for predicting perioperative mortality. Two studies performed more than 10 years apart examined mortality after abdominal surgery in cirrhotic patients and reported nearly identical rates of mortality for patients with Child’s class A, B, and C cirrhosis: 10%, 30% to 31%, and 76% to 82%, respectively.63,64 In general, surgery may be undertaken in patients with Child’s class A cirrhosis, whereas the medical condition of patients with Child’s class B cirrhosis should be optimized prior to planned surgery. The mortality rate in patients with Child’s class C cirrhosis is prohibitive, and surgery should be avoided. The MELD score was created originally to predict survival in patients with cirrhosis and portal hypertension undergoing placement of a transjugular intrahepatic portosystemic shunt (TIPS).65 The score has subsequently been validated as an accurate predictor of survival in patients with advanced liver disease. The MELD score incorporates three objective variables into a mathematical formula: 9.57 × loge(creatinine) + 3.78 × loge(total bilirubin) + 11.2 × loge(INR) + 6.43. The working range is 6 to 40, and the score has been shown to correlate with mortality in patients undergoing surgery other than liver transplantation, including hepatic resection, other abdominal procedures, and cardiac surgery.66-68 MELD is used most often for prioritizing the allocation of donor organs for liver transplantation. Before 2002, organs were allocated by the United Network for Organ Sharing

Chapter 73  Liver Chemistry and Function Tests (UNOS) on the basis of the CTP score and time on the wait list. Questions of fairness with this system were raised in light of subjective variables in the CTP score (specifically, the amount of ascites and grade of encephalopathy). Wait time was also believed to be an unfair component that placed patients referred at a late stage of disease at a disadvantage. The MELD score addressed these concerns by using only three objective variables and eliminating wait time as a criterion for organ allocation.69 Since implementation of the MELD score for prioritizing organ allocation, the number of deaths among patients on the wait list has decreased, suggesting that use of the MELD score is achieving its primary goal—allocation of organs to the sickest patients first (see Chapter 95).

KEY REFERENCES

Addario L, Scaglione G, Tritto G, et al. Prognostic value of quantitative liver function tests in viral cirrhosis: A prospective study. Eur J Gastroenterol Hepatol 2006; 18:713-20. (Ref 54.) Bellest L, Eschwege V, Poupon R, et al. A modified international norma­ lized ratio as an effective way of prothrombin time standardization in hepatology. Hepatology 2007; 46:528-34. (Ref 36.) Friedrich-Rust M, Martens S, Sarrazin C, et al. Performance of transient elastography for the staging of liver fibrosis: A meta-analysis. Gastroenterology 2008; 134:960-74. (Ref 46.) Kamath PS, Kim WR. The Model for End-Stage Liver Disease (MELD). Hepatology 2007; 45:797-805. (Ref 65.) Kaneda H, Hashimoto E, Yatsuji S, et al. Hyaluronic acid levels can predict severe fibrosis and platelet counts can predict cirrhosis in

patients with nonalcoholic fatty liver disease. J Gastroenterol Hepatol 2006; 21:1459-65. (Ref 39.) Kaplan M. Alkaline phosphatase. Gastroenterology 1972; 62:452-68. (Ref 21.) Kim HC, Nam CM, Jee SH, et al. Normal serum aminotransferase concentration and risk of mortality from liver diseases: Prospective cohort study. BMJ 2004; 328:983. (Ref 13.) Kunde SS, Lazenby AJ, Clements RH, et al. Spectrum of NAFLD and diagnostic implications of the proposed new normal range for serum ALT in obese women. Hepatology 2005; 42:650-6. (Ref 15.) Lee WM. Drug-induced hepatotoxicity. N Engl J Med 2003; 349:474-85. (Ref 59.) Mansour A, Watson W, Shayani V, et al. Abdominal operation in patients with cirrhosis: Still a major surgical challenge. Surgery 1997; 122: 730-6. (Ref 64.) Nathwani RA, Pais S, Reynolds TB, et al. Serum alanine aminotransferase in skeletal muscle diseases. Hepatology 2005; 41:380-2. (Ref 18.) Poynard T, McHutchison J, Manns M, et al. Biochemical surrogate markers of liver fibrosis and activity in a randomized trial of peginterferon alpha-2b and ribavirin. Hepatology 2003; 38:481-92. (Ref 41.) Prati D, Taioli E, Zanella A, et al. Updated definitions of healthy ranges for serum alanine aminotranferase levels. Ann Intern Med 2002; 137:1-10. (Ref 12.) Tripodi A, Chantarangkul V, Prirnignani M, et al. The international normalized ratio calibrated for cirrhosis normalizes prothrombin time results for model for end-stage liver disease calculation. Hepatology 2007; 46:520-7. (Ref 37.) Wiesner R, Edwards E, Freeman R, et al. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology 2003; 124:91-6. (Ref 69.) Full references for this chapter can be found on www.expertconsult.com.

1237

CHAPTE R

74 Hemochromatosis Bruce R. Bacon and Robert S. Britton

CHAPTER OUTLINE Causes of Iron Overload  1239 Pathophysiology  1240 Intestinal Iron Absorption  1240 Hepcidin  1240 HFE Protein  1241 Iron-Induced Tissue Injury and Fibrosis  1242

Trousseau was the first to describe a case of hemochromatosis in the French pathology literature in 1865.1 Almost 25 years later, in 1889, von Recklinghausen, thinking that the disease was a blood disorder that caused increased skin pigmentation, introduced the term hemochromatosis.1 In 1935, Sheldon published a description of all 311 cases of the disease that had been reported in the world’s literature to that point, including several from his own records. He recognized that hemochromatosis was an inborn error of iron metabolism and that all the pathologic manifestations of the disease were caused by increased iron deposition in the affected organs.1 In 1976, Simon and coworkers2 demonstrated that the gene for hereditary hemochromatosis (HH) was linked to the HLA region on the short arm of chromosome 6. The benefit of early diagnosis on survival was shown in a classic paper by Niederau and colleagues,3 who demonstrated that if HH was identified and treated before the development of cirrhosis or diabetes mellitus, survival of affected patients was equivalent to that of an age- and gender-matched population. In 1996, the HFE gene was identified on chromosome 6, thereby permitting genetic testing for the two major mutations (C282Y, H63D) that are responsible for HFE-related HH.4 Several prospective population studies have shown that the frequency of the C282Y homozygous state is approximately 1 in 250 in white populations of northern European descent.5 It is now recognized that C282Y homozygosity has incomplete clinical penetrance, with a strong male predominance for symptomatic disease.6 HH is characterized by increased intestinal iron absorption that results from low expression of the iron-regulatory protein hepcidin.5-7 In addition to the discovery of HFE and hepcidin, several additional genes and proteins involved in the regulation of iron homeostasis have been identified, contributing to a better understanding of cellular iron uptake and release. Also, numerous clinical and pathophysiologic studies have been performed and have led to improved diagnosis, better family screening, and new insights into normal and abnormal iron homeostasis. HFE-related HH is a common autosomal recessive disorder of iron metabolism; if it is diagnosed early and treated appropriately, every patient with the disorder can have a normal lifespan.

Clinical Features  1242 Diagnosis  1243 Treatment and Prognosis  1246 Family Screening  1247

CAUSES OF IRON OVERLOAD Hereditary hemochromatosis comprises several inherited disorders of iron homeostasis characterized by increased intestinal iron absorption that results in tissue iron deposition (Table 74-1). The older terms primary hemochromatosis and idiopathic hemochromatosis should no longer be used. The liver is always the principal recipient of most of the absorbed iron and is always involved in symptomatic HH. The most common form of HH by far is HFE-related HH.5-9 It is an autosomal recessive disorder usually identified in adults of northern European ancestry. Most patients who present with HH are homozygous for the C282Y mutation of HFE, although some persons who are compound heterozygotes (C282Y/H63D) also have iron overload. Other inherited forms of iron overload, classified as nonHFE-related HH, are juvenile hemochromatosis and iron overload resulting from mutations in the genes for hepcidin,10 transferrin receptor 2 (TFR2),11 or ferroportin.12 Juvenile HH is characterized by rapid iron accumulation. Mutations in two different genes have been shown to cause forms of juvenile HH. The more common mutation occurs in the HJV gene on chromosome 1q; this gene encodes a protein called hemojuvelin.13 Mutations in the hepcidin gene (HAMP) also produce a form of juvenile HH10; hepcidin is a hepatic peptide that acts to down-regulate iron absorption (see later). Mutations of the gene TFR2 produce an autosomal recessive form of HH that is clinically similar to HFE-related HH.11 How these TFR2 mutations result in iron overload is not yet known; they possibly cause abnormal iron sensing by hepatocytes, the predominant site of TFR2 expression.14 A rare autosomal dominant form of HH results from two categories of mutations in the gene for the iron transporter ferroportin.15 “Loss-of-function” mutations decrease the cell surface localization of ferroportin, thereby reducing its ability to export iron. The result is iron depo­ sition primarily in macrophages, and this disorder is sometimes termed ferroportin disease. The second category includes “gain-of-function” ferroportin mutations that abolish hepcidin-induced ferroportin internalization and degradation; the distribution of excess iron is similar to that in HFE-related HH, primarily parenchymal.

1239

1240

Section IX  Liver Table 74-1  Iron Overload Conditions Hereditary hemochromatosis

Secondary iron overload

Miscellaneous

HFE-related hereditary hemochromatosis (type 1)   C282Y homozygosity   C282Y/H63D compound heterozygosity   Other HFE mutations Non–HFE-related hereditary hemochromatosis:   Hemojuvelin (HJV) mutations (type 2A)   Hepcidin (HAMP) mutations (type 2B)   Transferrin receptor 2 (TFR2) mutations (type 3)   Ferroportin (SLC40A1) mutations (type 4)    Loss-of-function mutations    Gain-of-function mutations   African iron overload Iron-loading anemias   Aplastic anemia   Chronic hemolytic anemia   Pyridoxine-responsive anemia   Pyruvate kinase deficiency   Sideroblastic anemia   Thalassemia major Parenteral iron overload   Iron-dextran injections   Long-term hemodialysis   Red blood cell transfusions Chronic liver disease   Alcoholic liver disease   Hepatitis B   Hepatitis C   Nonalcoholic steatohepatitis   Porphyria cutanea tarda   Portacaval shunt Insulin resistance syndrome with iron overload Dietary iron overload Aceruloplasminemia Congenital alloimmune hepatitis (neonatal hemochromatosis) Congenital atransferrinemia

African iron overload occurs primarily in sub-Saharan Africa and is now considered to be the result of a non-HFE– related genetic trait that can be exacerbated by dietary iron loading.16 Some persons who manifest African iron overload consume an iron-rich fermented maize beverage, but iron overload also can occur in people who do not drink this beverage. In most cases, iron-loaded Kupffer cells are prominent in African iron overload; by contrast, Kupffer cells are relatively spared in HFE-related HH. A similar form of iron overload has been suggested to occur in African Americans,17 and further investigations are needed to determine the genetic basis, prevalence, and clinical consequences of this condition. Persons who absorb excessive amounts of iron as a result of an underlying cause other than any of the previously mentioned inherited defects have secondary iron overload18 (see Table 74-1). Examples are persons with disorders of ineffective erythropoiesis, liver disease (in some cases), increased oral intake of iron, or the rare condition congenital atransferrinemia. Both HH and secondary iron overload should be distinguished from parenteral iron overload, which is always iatrogenic and which leads to iron deposition that is found initially in the reticuloendothelial system. In patients with ineffective erythropoiesis who require red blood cell transfusions, parenchymal and reticuloendothe-

lial iron overload coexists because these people have a stimulus to increased iron absorption and receive iron in the form of red blood cell transfusions. Congenital alloimmune hepatitis is responsible for most cases of neonatal hemochromatosis.19 In these cases, immune-mediated liver injury in the fetus is associated with the development of iron overload. Treatment with intravenous immunoglobulin during pregnancy markedly slows or prevents the development of this condition.

PATHOPHYSIOLOGY The pathophysiologic mechanisms of HH fall into the following four main categories: (1) increased intestinal absorption of dietary iron, (2) decreased expression of the iron-regulatory hormone hepcidin, (3) altered function of HFE protein, and (4) iron-induced tissue injury and fibrogenesis.

INTESTINAL IRON ABSORPTION

An increase in intestinal iron absorption is a pathogenic characteristic of HFE-related HH.5-9 Understanding the pathogenesis of HH, therefore, requires a review of the determinants of duodenal iron absorption. Because there are no important physiologic mechanisms to regulate iron loss, iron homeostasis depends on a tight linkage between body iron requirements and intestinal iron absorption. Nearly all absorption of dietary iron occurs in the duodenum, where iron may be taken up as either ionic iron or heme.14,20 The absorption of both forms of iron is increased in patients with HFE-related HH. Absorption of ionic iron across the enterocyte occurs in two stages: uptake across the apical membrane and transfer across the basolateral membrane (Fig. 74-1A). Before uptake, ionic iron must be reduced from the ferric to the ferrous state; this step is accomplished by ferric reductases that are expressed on the luminal surface of duodenal enterocytes. The ferrous iron crosses the apical membrane via divalent metal transporter 1 (DMT-1). Iron taken up by the enterocyte may be stored as ferritin (and excreted in the feces when the senescent enterocyte is sloughed) or transferred across the basolateral membrane to the plasma. This latter process occurs via the transporter, ferroportin. The basolateral transfer of iron requires oxidation of iron to the ferric state by the ferroxidase hephaestin. Uptake of heme occurs by a transporter whose identity remains uncertain. Once internalized, the heme is degraded and the liberated iron is handled by the enterocyte in the same manner as absorbed ionic iron. Patients with HFE-related HH demonstrate increased basolateral transfer of iron from the enterocytes to the plasma. This increased transfer may be the driving force behind the increased intestinal iron absorption that is characteristic of HH. Some studies of patients with HFErelated HH have demonstrated higher duodenal expression of ferroportin and DMT-1.14 The major regulator of intestinal iron absorption is the peptide hormone hepcidin.

HEPCIDIN

Hepcidin is an iron-regulatory hormone that plays a central role in iron homeostasis by coordinating iron absorption, mobilization, and storage to meet the iron requirements of erythropoiesis and other iron-dependent processes20-23 (see Fig. 74-1B). Hepcidin is expressed predominantly in hepatocytes and is secreted into the circulation. It binds to ferroportin, which is highly expressed on macrophages and the basolateral surface of enterocytes, thereby

Chapter 74  Hemochromatosis Fe3+

Ferric reductase

Fe2+ DMT-1

Liver (HFE, HJV, TFR2)

Enterocyte Ferritin

Ferroportin

Low MW Fe pool

Fe2+

Hephaestin

Fe3+

Iron release by macrophages

Hepcidin

Iron absorption by enterocytes

A B Transferrin Figure 74-1.  Iron absorption pathway in duodenal enterocytes and the role of hepcidin. A, Duodenal enterocytes are the major site of iron absorption. Before uptake, dietary ionic iron requires reduction from the ferric (Fe3+) to the ferrous (Fe2+) state. This is accomplished by ferric reductases that are expressed on the luminal surfaces of enterocytes. Ferrous iron is taken up by the apical transporter, DMT-1. Iron may be stored within the cell as ferritin, and then lost with the sloughed senescent enterocyte, or transferred across the basolateral membrane to the plasma. This latter process occurs via the transporter ferroportin and requires oxidation of iron back to the ferric state by the ferroxidase hephaestin. B, Hepcidin is produced by the liver and secreted into the blood. HFE protein, hemojuvelin (HJV), and transferrin receptor 2 (TFR2) may participate in the hepatic iron-sensing mechanism that regulates hepcidin expression. Hepcidin reduces iron release by macrophages (and thereby increases macrophage iron stores) and also reduces iron absorption by duodenal enterocytes to reduce the amount of dietary iron in the circulation. In HFE-related hereditary hemochromatosis, loss of functional HFE protein leads to aberrant hepatocellular sensing of plasma iron, inappropriately low levels of hepcidin, diminished macrophage iron stores, and greater duodenal iron absorption. MW, molecular weight.

causing ferroportin to be internalized and degraded, thus inhibiting iron export. Hepcidin expression is regulated by total body iron, erythropoiesis, hypoxia, and inflammation. Excess iron and inflammation induce hepcidin expression, which, in turn, results in decreased intestinal iron absorption and diminished iron release from macrophages. By contrast, hepcidin expression is decreased by iron deficiency, erythropoiesis, and hypoxia, with resulting increases in iron absorption from the intestine and release of iron from macrophages. In all types of HH, iron overload results from impairment in the hepcidin regulatory pathway. In humans and mice, mutations or knockout of the genes for HFE, hemojuvelin, hepcidin, or TFR2 decrease hepcidin expression, with a resulting increase in intestinal iron absorption via upregulation of ferroportin levels.20-25 Studies have revealed that iron-induced regulation of hepcidin expression involves a bone morphogenetic protein (BMP)-dependent signaling pathway.20,23 BMPs bind to specific receptors on hepatocytes, thereby triggering SMAD protein-dependent activation of hepcidin expression. Selective inhibition of BMP signaling abrogates ironinduced upregulation of hepcidin. Hemojuvelin is a BMP co-receptor and facilitates the binding of BMP to its receptor; knockout of the hemojuvelin gene markedly decreases BMP signaling and hepcidin expression and causes iron overload. The inflammatory cytokine interleukin-6 upregulates hepcidin via STAT3 (signal transducer and activator of transcription-3) signaling, causing iron retention in macrophages and decreased intestinal iron absorption. The resulting hypoferremia plays a major causal role in the anemia of chronic disease.20-23 Reactive oxygen species inhibit hepcidin expression via a C/EBPα (CCAAT/ enhancer binding protein alpha)-mediated mechanism,

which may account for the hepatic iron loading associated with alcoholic liver disease and chronic hepatitis C.26

HFE PROTEIN

Studies of HFE protein structure and function were a direct consequence of the cloning of the HFE gene. The HFE gene encodes a 343-amino acid protein consisting of a 22-amino acid signal peptide, large extracellular domain, single transmembrane domain, and short cytoplasmic tail (Fig. 74-2).4 The extracellular domain of HFE protein consists of three loops (α1, α2, and α3), with intramolecular disulfide bonds within the second and third loops. The structure of the HFE protein is similar to that of other major histocompatibility complex (MHC) class I proteins, but evidence indicates that HFE protein does not participate in antigen presentation. Like MHC class I molecules, however, HFE protein is physically associated with β2-microglobulin (see Fig. 74-2). The major mutation responsible for HH results in the substitution of tyrosine for cysteine at amino acid 282 in the α3 loop (C282Y) and abolishes the disulfide bond in this domain.4 Loss of this disulfide bond interferes with the interaction of HFE protein with β2-microglobulin, and the C282Y mutant protein demonstrates decreased presentation at the cell surface, increased retention in the endoplasmic reticulum, and accelerated degradation.27 A second mutation associated with HH results in the change of a histidine to an aspartate at position 63 in the α1 chain (H63D), but this mutation has less biological impact than the C282Y mutation. Like HH patients, HFE-knockout mice manifest higher hepatic iron levels, elevated transferrin saturation (TS), increased intestinal iron absorption, and relative sparing of iron loading in reticuloendothelial cells.14 The molecular mechanisms by which HFE influences iron-dependent regulation of hepcidin remain unclear. HFE can bind to both TFR2 and the classic transferrin receptor,

1241

1242

Section IX  Liver His63 → Asp, H63D

HFE protein

α1

S S

α2

Studies of iron-induced tissue damage in organs other than the liver, such as the heart, pancreas, and endocrine glands, have been limited. Studies in myocardial cells have shown functional abnormalities resulting from iron-induced peroxidation.31

CLINICAL FEATURES NH2 β2-microglobulin

Extracellular

Cytosol

S S

S S

α3 Cys282 → Tyr, C282Y

Plasma membrane

COOH Figure 74-2.  Schematic model of HFE protein in association with β2microglobulin at the cell surface. The three extracellular domains of HFE protein are designated α1, α2, and α3. β2-Microglobulin is physically associated with the α3 domain. HFE protein also contains a transmembrane domain and a short intracellular domain. Positions of the two common HFE mutations, C282Y and H63D, are shown.

TFR1.20 In addition, both HFE and TFR2 may interact with hemojuvelin, suggesting that a complex of HFE and TFR2 may play a regulatory role in BMP signaling. One proposed model suggests that the complex of TFR1 and HFE acts as an iron sensor at the cell membrane of the hepatocyte; as the TS increases, diferric transferrin displaces HFE from TFR1, thereby making HFE available to bind to TFR2. The complex of HFE and TFR2 is then postulated to influence hepcidin expression.20,23

IRON-INDUCED TISSUE INJURY AND FIBROSIS

Another major pathophysiologic mechanism in HH relates to the liver damage that results from iron overload. In patients with advanced HH, hepatic fibrosis and cirrhosis are the principal pathologic findings. A number of studies of experimental hepatic iron overload have identified irondependent lipid peroxidation and associated impairment of membrane-dependent functions of mitochondria, microsomes, and lysosomes.28 A relationship between ironinduced lipid peroxidation and fibrosis has been shown in several studies.29,30 One hypothesis is that iron-induced lipid peroxidation occurs in hepatocytes and causes hepatocellular injury or death. Kupffer cells may become activated by products released from injured iron-loaded hepatocytes and produce profibrogenic cytokines, which can, in turn, stimulate hepatic stellate cells to synthesize larger amounts of collagen, thereby leading to pathologic fibrosis.29,30

Many patients with HFE-related HH come to medical attention without any symptoms or physical findings. They are identified as homozygous relatives of probands during family screening studies or by the results of serum iron studies in routine screening blood chemistry panels (Table 74-2).32,33 Nonetheless, the clinician should know the typical clinical manifestations in patients who do present with symptomatic disease. Most patients with symptomatic HFE-related HH are 40 to 50 years of age at the time of detection. Although C282Y homozygosity is distributed equally between men and women, the clinical penetrance is much lower in women, as a result of iron loss from normal menses and childbirth and possible gender-related disease modifier genes (Table 74-3).34-39 When patients present with symptoms, the most common are weakness and lethargy, arthralgias, abdominal pain, and loss of libido or potency in men.3,40 Patients with HFErelated HH may have nonspecific right upper quadrant abdominal pain that is most likely caused by hepatic capsular distention. Hepatomegaly is found on physical examination in a majority of patients; splenomegaly and other complications of chronic liver disease, including ascites, edema, and jaundice, may be present. Diabetes mellitus has decreased in frequency with earlier diagnosis of hemochromatosis and is typically not seen in the absence of cirrhosis (see later). Detection of the often subtle bronzed or slate gray skin pigmentation of HFE-related HH requires astuteness on the part of the clinician (see later). Organ damage and symptoms are usually related to the extent of iron loading. When patients are identified prospectively by either family or population screening, the frequency of patients who are asymptomatic increases dramatically.34-39 All patients with HFE-related HH who have an elevated serum ferritin value should also have increased hepatic iron stores, but the extent of hepatic iron loading is often not high enough to cause liver damage. In the late 1960s, cirrhosis was found in more than 50% of the patients identified with HH3; in studies from the 1970s through the 1990s, cirrhosis was found in only 5% to 10% of patients.32,33 Later population screening studies have reported an even lower frequency of cirrhosis in C282Y homozygotes.35,38,39 Serum aminotransferase elevations are usually mild. With regular phlebotomy and depletion of excess iron stores, elevated liver enzyme values typically revert to normal. When HFErelated HH is diagnosed and treated before the development of hepatic fibrosis or cirrhosis, long-term hepatic abnor­ malities do not develop. When HFE-related HH is detected after cirrhosis has developed, however, hepatocellular carcinoma can occur even after successful phlebotomy,41 thus emphasizing the importance of early diagnosis and treatment. Other clinical manifestations that can occur relate to the level of iron loading in nonhepatic organs. In older series, diabetes mellitus was a common complication of pancreatic iron loading,3 but in later series in which the diagnosis of HFE-related HH was made earlier in its course, diabetes mellitus has rarely been present.32,33 Other endocrinologic abnormalities are loss of libido and impotence in men,

Chapter 74  Hemochromatosis Table 74-2  Clinical Features of Hereditary Hemochromatosis in Three Studies from Different Time Periods Reference

Variable Dates of study Case selection method Number of patients Men Women Mean age (yrs) Age range (yrs) Symptoms (%) Weakness or lethargy Abdominal pain Arthralgias Loss of libido, impotence (% of men) None Findings (%) Hepatomegaly Skin pigmentation Diabetes mellitus Elevated liver enzymes Cirrhosis

3

32

33

1959-1983 Symptomatic index cases, family screening 163 145 18 46

Before 1990 Family screening by HLA typing 37 19 18 Men: 49 Women: 53 11-79

1990-1995 Screening chemistry panels, family screening 40 26 14 Men: 46 Women: 47 23-73

83 58 43 38 9

19 3 40 32 46

25 3 13 12 73

83 75 55 62 69

3 9 11 27 3

13 5 5 33 13*

18-77

*Five of 40 patients had cirrhosis; 1 had concomitant chronic hepatitis C, and 1 had alcoholic liver disease. HLA, human leukocyte antigen.

Table 74-3  Clinical Penetrance of C282Y Homozygosity in Women and Men FINDING 39

Iron-overload–related disease* Liver fibrosis34,37-39 Cirrhosis34,35,37-39 Abnormal metacarpophalangeal joints34,37,39

WOMEN (%)

MEN (%)

1 0-5 0-2 2-12

28 11-18 1-12 4-26

*Defined as iron overload accompanied by at least one of the following conditions: hepatocellular carcinoma, liver fibrosis or cirrhosis, characteristic arthropathy, raised serum aminotransferase levels, or diagnosis due to symptoms of hereditary hemochromatosis. Superscript numbers are references.

owing to both primary testicular failure and gonadotropin insufficiency resulting from the effects of iron on pituitary function,42 and occasionally hypothyroidism; adrenal function is typically normal. Other endocrinologic effects can occur as a result of complications of cirrhosis (see Chapter 92). Cardiac manifestations occur rarely because patients are now diagnosed earlier in the course of HFE-related HH than in the past. Cardiomyopathy, atrial and ventricular dysrhythmias, and congestive heart failure can occur.43 Characteristic of the arthropathy of HFE-related HH are changes in the second and third metacarpophalangeal joints. Joint space narrowing, chondrocalcinosis, subchondral cyst formation, osteopenia, and swelling of the joints may be seen.8,9,40 Unfortunately, the arthritic symptoms of HFErelated HH typically do not improve with phlebotomy. The skin pigmentation of HFE-related HH, which can be subtle, is characterized by either a bronze discoloration due to predominant melanin pigmentation or a gray pigmentation resulting from iron deposition in the basal layers of the

epidermis.40 The frequency of certain infections, including those caused by Vibrio vulnificus, Listeria monocytogenes, Yersinia enterocolitica, and Yersinia pseudotuberculosis, is more common in iron-loaded patients, although still rare.

DIAGNOSIS The requirements for diagnosis of HH have changed since the availability of HFE mutation analysis.8,9 As in the past, the disorder must be considered in any patient with typical symptoms or abnormal screening iron test results. If the physical examination or family history raises suspicions, then the appropriate serum iron tests along with HFE mutation analysis should be obtained. With the advent of genetic testing, the need for liver biopsy has diminished. In symptomatic patients, as discussed earlier, the most common symptoms are fatigue, malaise, right upper quadrant abdominal pain, and arthralgias. Less commonly, symptoms of chronic liver disease, diabetes mellitus, and congestive heart failure are identified. Because many of these symptoms are nonspecific or are related to other common diseases, HH is often overlooked by clinicians. In the early 1990s, patients with HH commonly presented after the discovery of abnormal results of screening blood chemistry tests obtained as part of routine health maintenance or for another reason.33 Many commercial laboratories added iron and total iron-binding capacity (TIBC), with TS calculated as iron ÷ TIBC × 100%, to their panel of screening serum chemistry tests, and in many patients, a TS was obtained inadvertently even though the test had not been specifically ordered. In one series, 62% of patients newly identified between 1990 and 1995 came to medical attention in this way.33 Another 14% of cases were identified through screening of family members of a known proband. Therefore, as many as 75% of patients came to

1243

1244

Section IX  Liver Table 74-4  Representative Iron Measurements in Serum and Liver and HFE Mutation Analysis Findings in Patients with Phenotypic HFE-Related Hereditary Hemochromatosis TEST Serum Iron   (µg/dL)   (µmol/L) Transferrin saturation (%) Ferritin Men (ng/mL; µg/L) Women (ng/mL; µg/L) Liver Iron staining Iron concentration   (µg/g dry weight)   (µmol/g dry weight) Hepatic iron index ([µmol/g dry weight] ÷ age in years) HFE Mutation Analysis

NORMAL VALUE/RESULT

VALUE/RESULT IN HFE-RELATED HEMOCHROMATOSIS

60-180 11-32 20-45*

180-300 32-54 45-100

20-200 15-150

300-3000 250-3000

0, 1+

3+, 4+

300-1500 5-27 <1.1

3000-30,000 53-536 >1.9

wt/wt C282Y/wt H63D/wt

C282Y/C282Y C282Y/H63D

*A value of 45% is considered elevated. wt, wild type.

medical attention by way of screening laboratory tests. The majority of these patients were asymptomatic and had no physical findings of HH, and the frequency of end-stage complications of HFE-related HH, such as cirrhosis and diabetes mellitus, was much lower than that reported in earlier series of patients who presented with symptoms of the disease.33 In 1998, the Health Care Finance Administration (now the Center for Medicare and Medicaid Services) stopped providing reimbursement for screening tests of any kind, and since then, fewer American patients with HH have been identified through routine screening. When HFE-related HH has been considered, the diagnosis is relatively straightforward. Measurements of serum iron and TIBC or transferrin, with calculation of TS, and serum ferritin should be obtained (Table 74-4). Studies have shown that it is not necessary for blood samples to be drawn in the fasting state. A TS value greater than 45% is the earliest phenotypic manifestation of HFE-related HH. As a result, TS is a more sensitive and specific test for HFE-related HH than serum ferritin, which can be normal in young persons with HFE-related HH or elevated in unaffected persons for a variety of reasons, including various types of necroinflammatory liver disease (e.g., chronic viral hepatitis, alcoholic liver disease, nonalcoholic steatohepatitis), certain malignancies, and other inflammatory conditions. An elevated serum ferritin level with a normal TS value in a person who has an inflammatory disorder generally suggests that the person does not have HFE-related HH. On the other hand, an elevated TS value with a normal ferritin value in a young person does not exclude HFE-related HH. A large North American population screening study demonstrated that 1 in 227 white persons was homozygous for the C282Y mutation, but only 57% and 88% of the female and male homozygotes, respectively, had an elevated ferritin value.36 This finding indicates that a higher proportion of C282Y homozygotes do not express iron overload than had previously been thought. The proportion of the non-expressing cohort that will subsequently show evidence of iron loading is uncertain. In a longitudinal follow-up study of patients identified by genetic screening, progressive iron loading, as indicated by rising serum ferritin levels, developed in 40% of C282Y homozygotes.44

As soon as serum iron parameters have been determined to be abnormal, HFE mutation analysis should be performed. If the patient is a C282Y homozygote or a compound heterozygote (C282Y/H63D) and has a serum ferritin level lower than 1000 ng/mL and normal liver enzyme values, a liver biopsy is not needed.45-47 If, however, the serum ferritin value is higher than 1000 ng/mL or liver enzymes are elevated, liver biopsy is indicated. If liver biopsy is determined to be appropriate, sufficient tissue for histopathologic evaluation and biochemical measurement of the hepatic iron concentration (HIC) should be obtained. With the advent of genetic testing, liver biopsy is performed solely to assess the damage (if any) to the liver. A proposed algorithm for evaluating people for possible HFE-related HH is shown in Figure 74-3. When a liver biopsy specimen is obtained, Perls’ Prussian blue stain is used for the determination and loca­ lization of storage iron. Iron stores in HFE-related HH are typically found in periportal hepatocytes, with little or no iron found in Kupffer cells (Fig. 74-4).48 In patients with a higher HIC, iron distribution becomes panlobular, and storage iron can be seen in Kupffer cells and bile duct cells. Grade 1 or 2 Perls’ Prussian blue staining can be seen in specimens from normal livers or specimens from patients with very early HH confirmed by HFE mutation analysis. Grade 3 stainable iron occasionally can be seen in specimens from patients with alcoholic cirrhosis or end-stage liver disease, in which iron staining correlates poorly with HIC. In the absence of other disorders, grade 3 to 4 stainable iron in an HFE pattern is consistent with HFE-related HH. In addition to histochemical staining, biochemical iron measurement in the liver is important (see Table 74-4). Typically, patients with HFE-related HH who present with symptoms have a HIC greater than 10,000 µg/g (dry weight) (normal < 1500 µg/g); HIC values may be more than 30,000 µg/g. Fibrosis and cirrhosis are usually not seen until the HIC exceeds 20,000 µg/g.49 In patients with both HFE-related HH and other forms of chronic liver disease, such as alcoholic liver disease or chronic viral hepatitis, increased fibrosis or cirrhosis can occur at a much lower HIC and at a younger age.8,9,50,51 In asymptomatic or younger

Chapter 74  Hemochromatosis Transferrin saturation Normal (<45%)

Elevated (≥45%)

No further evaluation Genetic testing C282Y/H63D

C282Y/wt H63D/wt H63D/H63D wt/wt

C282Y/C282Y

Serum ferritin <1000 ng/mL

Normal ALT and AST Evaluate for other causes of iron overload; liver biopsy may be considered

Proceed with therapeutic phlebotomy

Serum ferritin ≥1000 ng/mL

Elevated ALT or AST

Perform liver biopsy for further evaluation

A

Consider liver biopsy

Figure 74-3.  Algorithm for the evaluation of possible HFE-related hereditary hemochromatosis in a person with a negative family history. ALT, serum alanine aminotransferase level; AST, serum aspartate aminotransferase level; wt, wild type.

B

Figure 74-4.  Histology of HFE-related hereditary hemochromatosis. A, This liver biopsy specimen was obtained from a 47-year-old C282Y homozygous woman who presented with a transferrin saturation of 63% and a serum ferritin level of 1190 ng/mL. The hepatic iron concentration was 9840 µg/g with a hepatic iron index of 3.7. At low power, iron deposition is seen to be much greater in the periportal zone (acinar zone 1) (arrows) than in the centrilobular zone (acinar zone 3). (Perls’ Prussian blue; ×100.) B, At a higher magnification of a specimen from another patient with HFE-related hereditary hemochromatosis, iron deposition is seen to be primarily in hepatocytes arranged in cords, with less iron accumulation in reticuloendothelial (Kupffer) cells that line the intervening sinusoids. In patients with higher hepatic iron concentrations, iron deposition becomes panlobular, and storage iron can be seen in Kupffer cells and bile duct cells. (Perls’ Prussian blue.) (A, Courtesy of Elizabeth M. Brunt, MD, St. Louis; B, Courtesy of Edward Lee, MD, Washington, DC.)

patients with early HFE-related HH, HIC is increased to a lesser degree and often is much less than 10,000 µg/g. A common diagnostic dilemma occurs when it is not clear whether a patient has liver disease with abnormal iron parameters or HFE-related HH with elevations in liver enzymes. In this setting, HFE mutation analysis is extremely useful. If the patient is a C282Y homozygote or a compound heterozygote (C282Y/H63D), the iron loading is most likely caused predominantly by the genetic abnormality. On the other hand, if the patient has underlying liver disease, is a C282Y heterozygote, or is an H63D heterozygote or H63D

homozygote or has neither mutation, it is likely that the iron loading is caused by the underlying liver disease, perhaps with a minor contribution from the HFE genotype. In the past, the hepatic iron index (HIC in µmol/g ÷ the patient’s age in years) was used to distinguish HH (>1.9) from secondary iron overload (≤1.9) .49 With HFE mutation analysis, the value of the hepatic iron index has diminished. Computed tomography (CT), magnetic resonance imaging (MRI), and magnetic susceptibility testing have all been proposed as techniques to quantify the HIC without the need for a liver biopsy (Fig. 74-5). Magnetic susceptibility

1245

1246

Section IX  Liver Table 74-5  Treatment of HFE-Related Hereditary Hemochromatosis Perform phlebotomy of 500 mL (1 unit) of whole blood weekly unless hematocrit value drops below 37%. Check transferrin saturation and ferritin levels at two- to three-month intervals to monitor response (optional). When iron stores are depleted (ferritin < 50 ng/mL and transferrin saturation < 50%), proceed to maintenance phlebotomy of 1 unit of whole blood every two to three months. Aim to keep transferrin saturation < 50%; if successful, ferritin should remain < 50 ng/mL.

Figure 74-5.  Computed tomography (CT) of a patient with hemochromatosis. Hepatic density measured by CT scanning is increased in heavily iron-loaded patients. The iron-containing liver is much brighter than the spleen.

testing is available in only a few centers in the United States and Europe as a research tool. In early studies, CT and MRI were generally not reliable for detecting mild iron overload, but newer MRI techniques have shown improved sensitivity.52,53

TREATMENT AND PROGNOSIS The treatment of HFE-related HH is relatively straightforward; most patients can be treated with routine therapeutic phlebotomy8,9 (Table 74-5). Ideally, diagnosis and initiation of treatment should begin before the development of hepatic fibrosis or cirrhosis; if they are, patients will have a normal lifespan. Each unit of whole blood (500 mL) contains approximately 200 to 250 mg of iron, depending on the hemoglobin concentration; therefore, C282Y homozygotes who have 10 to 20 g of excess storage iron require extended phlebotomy regimens (40 to 80 units of blood removed). Most patients can tolerate weekly phlebotomy of 1 unit of whole blood, and occasional younger patients can tolerate the removal of 2 or 3 units per week. Some older patients and occasional patients with a coexisting underlying hematologic disorder resulting in anemia can tolerate phlebo­tomy of only 0.5 unit per week or every other week. The iron-chelating drug deferoxamine is used in patients with HFE-related HH and cardiac manifestations or in patients who cannot tolerate phlebotomy. Deferoxamine, 20 to 50 mg/kg/day, is administered 5 days per week as a con­ tinuous subcutaneous infusion (over a 12-hour period each day) via a portable pump. Although not absolutely necessary, obtaining a TS value and serum ferritin level every two to three months is useful for predicting the eventual return of iron stores to normal. Typically, the serum ferritin level falls progressively as hepatic iron stores decrease, whereas TS usually remains elevated until just before iron stores return to normal. In uncomplicated cases, the ferritin level drops about 30 ng/mL with each unit of blood removed. When the iron stores have reached a level in the low-normal range, the serum ferritin level should be less than 50 ng/mL and the TS less than 50%. At this point, maintenance phlebotomies every two to three months are required in most patients.

The rate of reaccumulation of iron varies among indivi­ duals, and patients may require regular maintenance phlebotomy at more or less frequent intervals. Occasional patients do not reaccumulate iron, for reasons that are unknown. The prognosis for patients with HFE-related HH is improved significantly by therapeutic phlebotomy.3,54,55 Life expectancy is reduced in patients who present with cirrhosis or diabetes mellitus, and the risk of death from hepatocellular carcinoma is higher in patients with HFE-related HH. Hepatocellular cancer is usually seen only in patients who already have cirrhosis. Established cirrhosis typically does not reverse with phlebotomy, but many patients will have a decrease in hepatic fibrosis with aggressive treatment.37 Unfortunately, neither arthritis nor hypogonadism improves; however, management of diabetes mellitus may become easier after iron removal. The value of screening for hepatocellular carcinoma in cirrhotic patients with HH is controversial because the cost-effectiveness of the screening approaches has not been validated. Most authorities suggest ultrasonography or CT scanning and a serum alpha fetoprotein measurement every six months in cirrhotic patients with HFE-related HH. With improved methods of detection and treatment of small, early hepatocellular carcinomas (e.g., radiofrequency ablation, chemoembolization, resection, and liver transplantation), screening seems to be reasonable (see Chapter 94). When diagnosis and treatment are delayed and complications of end-stage liver disease develop, liver transplantation (LT) may be undertaken (see also Chapter 95). In the era before HFE genotyping was available, studies addressing the outcome of iron-loaded patients after LT reported that the survival rate was substantially lower than that for other patients. Significant hepatic iron loading is now known to occur in 35% to 78% of patients with end-stage liver disease, regardless of the cause of cirrhosis.56 Only about 10% of patients with iron overload and end-stage liver disease are C282Y homozygotes.57 Two studies have analyzed post-LT outcome in patients with confirmed HFE-related HH and found five-year survival rates of 34% and 55%, which were lower than the overall post-LT survival rate of 72% to 75%.58,59 The most common causes of death were infections, cardiovascular complications, and recurrence of hepatocellular carcinoma. The five-year survival rate for patients with non-HH iron overload was higher (63%) than that for patients with HFE-related HH (34%) but still somewhat reduced compared with the overall survival rate.58 If iron overload of any cause is diagnosed early, it should be treated to reduce the chance of death following LT. Unfortunately, HFE-related HH or secondary iron overload resulting from liver disease is often not diagnosed before LT. One factor that may contribute to the increased post-LT mortality in patients with undiagnosed or untreated HFE-related HH is the extent of iron deposition in extrahe-

Chapter 74  Hemochromatosis patic sites, such as the heart. A high index of suspicion for iron overload in patients with end-stage liver disease should lead to improved diagnosis and prompt institution of phlebotomy or iron-chelation therapy before LT. These changes in management should reduce the frequency of postoperative complications and improve long-term post-LT survival.

FAMILY SCREENING When a proband with HFE-related HH has been identified and therapy has been initiated, the clinician still has a responsibility to the patient’s family.8,9,60 For asymptomatic C282Y homozygotes and compound heterozygotes (C282Y/ H63D) identified by HFE mutation analysis within a sibship, a liver biopsy is unnecessary. In family members in whom serum ferritin levels are increased, proceeding to thera­ peutic phlebotomy is reasonable. Liver biopsy should be reserved for patients in whom another underlying liver disease is possible. Persons who are C282Y heterozygotes are not at risk for progressive iron overload. If the spouse of a C282Y homozygote is a C282Y heterozygote, their offspring have a 50% chance of being homozygous for C282Y. HFE mutation analysis in the children can eliminate the need for subsequent serum iron testing if a genotype of C282Y/C282Y or C282Y/H63D is not found in a proband’s spouse, although the issues of genetic discrimination and stigmatization must be acknowledged and considered. In children who are C282Y homozygotes or compound heterozygotes, serum ferritin measurements should be obtained yearly and phlebotomy instituted when ferritin values become elevated.

KEY REFERENCES

Adams PC, Barton JC. Haemochromatosis. Lancet 2007; 370:1855-60. (Ref 5.) Adams PC, Reboussin DM, Barton JC, et al. Hemochromatosis and iron-overload screening in a racially diverse population. N Engl J Med 2005; 352:1769-78. (Ref 36.) Allen KJ, Gurrin LC, Constantine CC, et al. Iron-overload-related disease in HFE hereditary hemochromatosis. N Engl J Med 2008; 358:221-30. (Ref 39.) Andrews NC. Forging a field: The golden age of iron biology. Blood 2008; 112:219-30. (Ref 20.) Bacon BR. Hemochromatosis: Diagnosis and management. Gastroenterology 2001; 120:718-25. (Ref 8.) Beutler E, Felitti VJ, Koziol JA, et al. Penetrance of 845G→A (C282Y) HFE hereditary haemochromatosis mutation in the USA. Lancet 2002; 359:211-18. (Ref 35.) Feder JN, Gnirke A, Thomas W, et al. A novel MHC class 1-like gene is mutated in patients with hereditary haemochromatosis. Nat Genet 1996; 13:399-408. (Ref 4.) Ganz T. Iron homeostasis: Fitting the puzzle pieces together. Cell Metab 2008; 7:288-90. (Ref 23.) Gordeuk VR. African iron overload. Semin Hematol 2002; 39:263-9. (Ref 16.) Nemeth E, Ganz T. Regulation of iron metabolism by hepcidin. Annu Rev Nutr 2006; 26:323-42. (Ref 21.) Olynyk JK, Trinder D, Ramm, GA, et al. Hereditary hemochromatosis in the post-HFE era. Hepatology 2008; 48:991-1001. (Ref 6.) Philippe MA, Ruddell RG, Ramm GA. Role of iron in hepatic fibrosis: One piece in the puzzle. World J Gastroenterol 2007; 13:4746-54. (Ref 30.) Pietrangelo A. Hereditary hemochromatosis: A new look at an old disease. N Engl J Med 2004; 350:2383-97. (Ref 7.) Pietrangelo A. Non-HFE hemochromatosis. Hepatology 2004; 39:21-9. (Ref 12.) Powell LW. Hereditary hemochromatosis and iron overload diseases. J Gastroenterol Hepatol 2002; 17(Suppl):S191-5. (Ref 9.) Full references for this chapter can be found on www.expertconsult.com.

1247

CHAPTE R

75 Wilson Disease Diane W. Cox and Eve A. Roberts

CHAPTER OUTLINE The Copper Pathway  1249 The Basic Molecular Defect  1249 Clinical Features  1251 Hepatic Presentation  1251 Neurologic Presentation  1252 Psychiatric Presentation  1252 Ocular Signs  1252 Involvement of Other Systems  1252

Copper, a component of several essential enzymes, is toxic to tissues when present in excess. Dietary intake of copper generally exceeds the trace amount required, and mechanisms to control influx and efflux from cells must maintain an appropriate balance. Two disorders of copper transport are known: Menkes disease, an X-linked defect in transport of copper from the intestine, and Wilson disease, an autosomal recessive disorder of copper overload. Wilson disease was first described, in 1912 by Kinnear Wilson, as a familial disease characterized by progressive, lethal neurologic dysfunction with chronic liver disease and a corneal abnormality, the Kayser-Fleischer (KF) ring.1 Wilson also observed that some younger siblings of typical patients died of severe liver disease without experiencing neurologic abnormalities. In this disease, inadequate hepatic copper excretion leads to copper accumulation in the liver, brain, kidney, and cornea. The incidence in most populations is on the order of 1 in 30,000.

THE COPPER PATHWAY Dietary copper is absorbed in the upper intestine and binds loosely to albumin in serum. Albumin and copper-histidine transport copper to a variety of tissues; most copper is transported to the liver. Trace amounts of copper are required for essential enzymes that affect connective tissue and elastin cross-linking (lysyl oxidase), free radical scavenging (superoxide dismutase), electron transfer (cytochrome oxidase), pigment production (tyrosinase), and neurotransmission (dopamine b-monooxygenase). Copper in hepatocytes and in other cells is bound to metallochaperones, low-molecular-weight proteins that each deliver copper to a specific target molecule. In the liver, copper is incorporated into the protein apoceruloplasmin to produce ceruloplasmin (also called holoceruloplasmin). More than 90% of the copper in plasma is an integral part of ceruloplasmin, an α2-glycoprotein that contains six molecules of copper and has a molecular weight of 132 kd. The normal serum concentration of ceruloplasmin in adults, as measured by immunochemical or enzymatic techniques, is 200 to 400 mg/L, rising from a very low

Pathology  1252 Diagnosis  1253 Mutation Analysis  1254 Treatment  1255 Prognosis  1257

level at birth to 300 to 500 mg/L in the first years of life. Ceruloplasmin is an acute-phase reactant that is elevated by inflammation (including inflammatory hepatic disease), pregnancy, and the use of exogenous estrogen. The majority of ingested copper is excreted via the bile; a small fraction is excreted in urine. When intestinal or liver cells are overloaded with copper, the metallothioneins, a class of lowmolecular-weight cysteine-rich proteins, are induced and sequester copper in a nontoxic form. The normal pathways of copper transport in the body and in the hepatocyte are shown in Figures 75-1 and 75-2.

THE BASIC MOLECULAR DEFECT Our understanding of the basic defect in Wilson disease increased dramatically with the cloning of the genes, first for X-linked Menkes disease and then for Wilson disease. The gene for Menkes disease (ATP7A), which was cloned by using a chromosomal breakpoint in an affected female patient, was found to be related to bacterial copper resistance genes. Cloning of the Wilson disease gene (ATP7B) was accomplished by a combination of linkage analysis, physical mapping of the relevant region of chromosome 13q14, and recognition of its extensive homology with the Menkes disease gene.2,3 The coding region of the gene is 4.1 kilobases in length, with messenger RNA (mRNA) of about 8 kilobases. The product, ATP7B (or the Wilson ATPase), is a membrane P-type adenosine triphosphatase (ATPase) that consists of 1443 amino acid residues and has a molecular mass of 160 kd. The predicted structure,2 as confirmed by structural studies,4 has six copper binding domains, a phosphorylation domain, an ATP-binding region, and eight transmembrane domains. (Fig. 75-3). All functionally important regions of the gene are conserved between bacteria and yeast. Mutations in the ATP7B gene result in retention of copper in the liver as well as impaired incorporation of copper into ceruloplasmin. The Long-Evans cinnamon (LEC) rat and the toxic milk (tx) mouse have mutations in their homologous ATP7B genes and are thus suitable models for the study of Wilson disease mechanisms and therapy.5,6

1249

1250

Section IX  Liver Oral intake of copper (1.5–4 mg daily) Intestinal absorption Plasma albumin binding (rapid clearance) Menkes disease (X chromosome) Other tissues, proteins: Brain Eye KIDNEY Enzymes Urine

Wilson disease (chromosome 13)

LIVER Apoceruloplasmin Cu2+

Aceruloplasminemia (chromosome 3)

MT Ceruloplasmin (ferroxidase)

Ferritin Fe2+

Iron mobilization

Plasma transferrin Fe3+

Biliary excretion (1–4 mg daily)

Figure 75-1.  Simplified overview of the pathways for copper ion transport and steps affected in genetic disorders of copper metabolism. MT, metallothioneins. (Modified from Cox DW. Genes of the copper pathway. Am J Hum Genet 1995; 56:828-34.)

Although the gene for Menkes disease is expressed in many tissues, including muscle, kidney, heart, and intestine, the gene for Wilson disease is expressed predominantly in the liver and kidney, with minor expression in brain, lungs, and placenta.2 Studies in cultured cells show localization of the Wilson ATPase in the trans-Golgi network, with trafficking from the trans-Golgi network to cytoplasmic vesicles in the presence of increased copper.7 When intracellular copper concentrations are elevated, the Wilson ATPase is found near the apical (bile canalicular) membrane in hepatocytes, consistent with its proposed function of facilitating excretion of copper via bile.8 Additional proteins are involved in the intracellular transport of copper. Molecular copper is not free in the cell but is transported to specific proteins by copper chaperones.9 The chaperone ATOX1 transports copper to the Wilson ATPase. Other chaperones deliver copper to superoxide dismutase in the cytoplasm and various coppercontaining proteins in mitochondria. The study of inherited hepatic copper toxicosis in Bedlington terriers has identified a new component of the copper transport system. Affected dogs show clinical variability that ranges from death or hepatic disease at two to three years of age to less severe chronic disease to a high level of hepatic copper. The proposed defective canine gene was identified by positional cloning through the use of markers to identify a region containing the COMMD1 (initially called the MURR1) gene. The gene has a deletion of one exon in some,10 but not all,11 affected dogs. Although

Ceruloplasmin Apoceruloplasmin Metallothionein (copper storage)

Nucleus

Bile canaliculus

ATP7B TGN

COMMD1

ATOX1 ATP7B COX17

Cu CTR1

SCO1

Cytochrome oxidase Mitochondrion

CCS

Cu/Zn superoxide dismutase

Figure 75-2.  Model of a hepatocyte showing major proteins in the copper transport pathway. Low-molecular-weight copper chaperones (ATOX1, COX17, and CCS) deliver copper to specific target proteins (ATP7B, cytochrome oxidase, and superoxide dismutase, respectively). SCO1 transports copper across the mitochondrial membrane. ATP7B (shown as a channel) traffics from the trans-Golgi network (TGN) to cytoplasmic vesicles that deliver copper to the bile canaliculus. COMMD1 may be involved in excretion of copper into bile.

Chapter 75  Wilson Disease 2 Metal binding 1 NH2

3

ATP binding domain

4 Transduction H1069Q 5

ATP hinge region Phosphorylated D residue

6

COOH

R778L

Channel

Conserved cysteine-proline-cysteine motif

the gene responsible for the excess copper remains to be identified, the disease in Bedlington terriers has highlighted other genes that may be involved in response to excess copper; COMMD1 interacts not only with the Wilson ATPase,12 but also with the X-linked inhibitor of apoptosis (XIAP), which can also bind copper.13 COMMD1 could be involved in response to, and facilitate elimination of, excess copper.

CLINICAL FEATURES The clinical presentation of Wilson disease is extremely variable. The age at onset of symptoms is much broader than previously thought, ranging from 3 to 55 years. Wilson disease with hepatic involvement has been identified in one- to two-year-old patients and in patients older than age 60.14 Patients may present with chronic or fulminant liver disease, a progressive neurologic disorder without clinically prominent hepatic dysfunction, isolated acute hemolysis, or psychiatric illness. The clinical variability often makes confirmation of the diagnosis difficult. Further description of the clinical features and treatment has been published.14,15

HEPATIC PRESENTATION

The hepatic presentation of Wilson disease is more common in younger patients than in older patients. Wilson disease should be considered as a possible diagnosis in any child, symptomatic or not, with hepatomegaly, persistently elevated serum aminotransferase levels, or evidence of fatty liver. Symptoms may be vague and nonspecific, such as fatigue, anorexia, or abdominal pain. Occasionally patients present with a self-limited clinical illness that resembles acute hepatitis, with malaise, anorexia, nausea, jaundice, elevated serum aminotransferase levels, and abnormal coagulation test results. Some patients have a history of self-limited jaundice, apparently caused by unexplained hemolysis. Patients may present with severe, established

Figure 75-3.  A model of the predicted product of the Wilson disease gene, ATP7B. Functional domains conserved in the Menkes disease gene are indicated as blocks. Numerous mutations occur in functionally important regions. Positions of common missense mutations, H1069Q and R778L, are shown. ATP, adenosine triphosphate. (Model modified from Bull PC, Cox DW. Wilson disease and Menkes disease: New handles on heavy-metal transport. Trends Genet 1994; 10:246-52.)

chronic liver disease with hepatosplenomegaly, ascites, congestive splenomegaly, a low serum albumin level, and persistently abnormal coagulation test results. Some patients have isolated splenomegaly without hepatomegaly. Many of these findings relate more to portal hypertension as a consequence of Wilson disease than to the metabolic disorder itself. Wilson disease may present in children and young adults with clinical liver disease indistinguishable from autoimmune hepatitis.16 As in autoimmune hepatitis, the onset may be acute. Fatigue, malaise, arthropathy, and rashes may occur; laboratory findings include elevated aminotransferase levels, a greatly increased serum immunoglobulin (IgG) concentration, and detectable nonspecific autoantibodies such as antinuclear antibodies and anti-smooth muscle (anti-actin) antibodies. Wilson disease must be specifically ruled out because the treatment of the two diseases is entirely different. With appropriate treatment, the long-term outlook for patients with Wilson disease that manifests as autoimmune hepatitis appears to be favorable, even if cirrhosis is present. Wilson disease may present as fulminant hepatic failure, with severe coagulopathy and encephalopathy. Acute intravascular hemolysis is usually present, and renal failure may develop. Because the patient typically has not been suspected of having underlying liver disease, fulminant viral hepatitis is usually the working diagnosis. Unlike fulminant viral hepatitis, fulminant Wilson disease is typically characterized by disproportionately low aminotransferase levels (usually much less than 1500 U/L) at the onset of clinically apparent disease. The serum alkaline phosphatase level is in the normal range or even low for the patient’s age, and the bilirubin level is often disproportionately elevated as a result of hemolysis. In adults who present with Wilsonian fulminant liver failure, the combination of a ratio of the alkaline phosphatase level to the total bilirubin level of less than 4 and a ratio of the aspartate aminotransferase (AST) to alanine aminotransferase (ALT) level of greater than 2.2 can be extremely helpful for making the diagnosis, whereas

1251

1252

Section IX  Liver the serum ceruloplasmin level is not (see later).17 Slit-lamp examination may reveal Kayser-Fleischer rings (see later). Urinary copper excretion is greatly elevated. These patients require urgent liver transplantation because they do not respond well to chelation treatment; albumin dialysis and related techniques may serve as temporary procedures until liver transplantation can be performed (see later).18 This presentation of Wilson disease is not rare, and affected patients account for a substantial proportion of persons transplanted for fulminant hepatic failure (see Chapters 93 and 95). Recurrent bouts of hemolysis may predispose to the development of gallstones. Cirrhosis, if present, may be a further predisposing factor. Children with unexplained cholelithiasis, particularly with small bilirubinate stones, should be tested for Wilson disease. Compared with other chronic liver diseases, Wilson disease is rarely complicated by hepatocellular carcinoma; however, patients may have an increased propensity to abdominal malignancies,19 and hepatocellular carcinoma has been reported in a child with Wilson disease. In patients who have predominantly hepatic disease, evidence of subtle neurologic involvement can often be found. Mood disturbance (mainly depression, but sometimes impulsive or neurotic behavior), deterioration in school performance or handwriting, and clumsiness may be identified by careful direct questioning. A soft whispery voice (hypophonia) is another early feature of neurologic involvement.

NEUROLOGIC PRESENTATION

The neurologic presentation of Wilson disease tends to occur in the second and third decades or later but has been reported in children as young as 6 to 10 years old. Most patients with a neurologic presentation have hepatic involvement, albeit often asymptomatic. Neurologic involvement follows two main patterns: movement disorder or rigid dystonia. Movement disorders tend to occur earlier and include tremors, poor coordination, and loss of fine motor control. Spastic dystonic disorders generally develop later, with mask-like facies, rigidity, gait disturbance, and pseudobulbar involvement such as dysarthria, drooling, and swallowing difficulty. Rarely, patients present with peripheral neuropathy or dysautonomia. Seizures are uncommon, and intellect is not impaired. Imaging of the brain is important for assessing neurologic Wilson disease, and results may be abnormal in the absence of overt neurologic symptoms. Magnetic resonance imaging may be the most sensitive modality.

PSYCHIATRIC PRESENTATION

As many as 20% of patients may present with purely psychiatric symptoms. These symptoms are highly variable, although depression is common. Phobias and compulsive behaviors have been reported; aggressive or antisocial behavior may also be seen.

OCULAR SIGNS

The classic Kayser-Fleischer ring is caused by copper deposition in Descemet’s membrane of the cornea. Copper is actually distributed throughout the cornea, but fluid streaming favors accumulation near the limbus, especially at the superior and inferior poles, and eventually circumferentially around the iris. Kayser-Fleischer rings are visible on direct inspection only when iris pigmentation is light and copper deposition is heavy. A careful slit-lamp examination is mandatory. Copper deposition in the lens (sunflower cataract), which does not interfere with vision, may be seen

on slit-lamp examination and, like Kayser-Fleischer rings, disappear with chelation therapy. Kayser-Fleischer rings may be absent in 40% to 60% of patients with exclusively hepatic involvement and in presymptomatic patients. Most patients with a neurologic or psychiatric presentation of Wilson disease have Kayser-Fleischer rings; only 5% do not. Kayser-Fleischer rings are not specific for Wilson disease; they may be found occasionally in patients with other types of chronic liver disease, usually with a prominent cholestatic component, such as primary biliary cirrhosis, primary sclerosing cholangitis, or familial cholestatic syndromes, and rarely in patients with nonhepatic diseases.

INVOLVEMENT OF OTHER SYSTEMS

Wilson disease can be accompanied by various extrahepatic disorders apart from neurologic disease. Episodes of hemolytic anemia can result from the sudden release of copper into the blood. Renal disease, mainly Fanconi’s syndrome, may be prominent. Findings include microscopic hematuria, aminoaciduria, phosphaturia, and defective acidification of the urine. Nephrolithiasis has also been reported. Arthritis, affecting mainly the large joints, may occur as a result of synovial copper accumulation. Other musculo­ skeletal problems include osteoporosis and osteochondritis dissecans. Vitamin D–resistant rickets may develop as a result of the renal damage. Copper deposition in the heart can lead to cardiomyopathy or cardiac arrhythmias. Sudden death in Wilson disease has been attributed to cardiac involvement but is rare. Copper deposition in skeletal muscle can cause rhabdomyolysis. Endocrine disorders can occur. Hypoparathyroidism has been attributed to copper deposition. Amenorrhea and testicular problems appear to result from Wilson disease itself, not from cirrhosis. Infertility or repeated spontaneous abortion may be a sign of Wilson disease. Pancreatitis, possibly resulting from copper deposition in the pancreas, may also occur.

PATHOLOGY In the earliest stages before cirrhosis develops, histologic findings in the liver include steatosis, focal necrosis, glycogenated nuclei in hepatocytes, and sometimes apo­ ptotic bodies. As parenchymal damage progresses, possibly through repeated episodes of lobular necrosis, periportal fibrosis develops. Cirrhosis is usually macronodular but may be micronodular. Early in the course of Wilson disease, hepatocellular copper is bound mainly to metallothionein and is distributed diffusely in the cytoplasm of hepatocytes; therefore, histochemical stains for copper are negative. As the disease progresses, the copper content exceeds the storage capacity of metallothionein, and copper is deposited in lysosomes. Lyosomal aggregates of copper can be detected by special staining techniques for copper or copper-binding protein (such as rubeanic acid or orcein, respectively). In the cirrhotic liver, some areas may have no stainable copper at all. If the clinical presentation mimics autoimmune hepatitis, a liver biopsy specimen may reveal classic histologic features such as interface hepatitis. Inflammation may be severe. Mallory (Mallory-Denk) bodies may be found. In patients who present with fulminant hepatic failure, histologic findings confirm preexisting liver disease; cirrhosis may be present; and parenchymal copper is located mainly in Kupffer cells rather than hepatocytes. Changes in hepatocellular mitochondria, identified with electron microscopy, are an important feature in Wilson

Chapter 75  Wilson Disease disease.20 The mitochondria vary in size, and the numbers of dense bodies in mitochondria may be increased. The most striking change is dilatation of the tips of the mitochondrial cristae as a result of separation of the inner and outer membranes of the cristae, with widening of the intercristal space until the appearance is irregularly cystic. The crista resembles a tennis racquet if only the tip is dilated. This finding, although not entirely specific for Wilson disease, can be helpful diagnostically, even in young and minimally affected patients. Involvement of hepatocytes may be not uniform, and abnormalities may be found in some hepatocytes in some lobules and not in others. The mitochondrial changes are probably a consequence of oxidative damage from excessive copper in hepatocytes.21

DIAGNOSIS The patient with the classic combination of chronic liver disease, tremor or dystonia, and Kayser-Fleischer rings is readily diagnosed on clinical grounds, but such patients are uncommon. A diagnostic scoring system has been proposed22 but has been evaluated only in children. Suggestive clinical symptoms are often the main prerequisite for diagnosing Wilson disease, and laboratory investigations may provide confirmation. Kayser-Fleischer rings should be sought through a careful slit-lamp examination and repeated if necessary. Lack of Kayser-Fleischer rings does not exclude the diagnosis of Wilson disease. Routine liver biochemical test levels are usually abnormal, with mild-to-moderate elevations of the serum aminotransferase levels. Serum levels of the ALT may be much lower than those of AST, possibly reflecting damage to hepatocellular mitochondria. Two major disturbances of copper disposition in Wilson disease are defective incorporation of copper into ceruloplasmin and decreased biliary excretion of copper. A summary of biochemical features in Wilson disease in comparison with normal persons is shown in Table 75-1. The classic feature of a low ceruloplasmin concentration has proved less typical than previously thought because hepatic inflammation may be sufficient to elevate serum ceruloplasmin levels. Also, the normal range for serum ceruloplasmin is increased in very young children. The method of measuring ceruloplasmin may account in part for finding normal ceruloplasmin levels in patients with Wilson disease. Immunologic methods, which are commonly used in most laboratories, measure both apoceruloplasmin and holoceruloplasmin and typically overestimate the true amount of functional ceruloplasmin in plasma. The oxidase assay, although technically less convenient for laboratories that perform automated testing, provides a more reliable measure of ceruloplasmin for diagnosis because the assay

measures enzymatically active, copper-containing ceru­ loplasmin. This method permits an accurate estimate of non–ceruloplasmin-bound copper23 and can also indicate possible early copper deficiency in treated patients.24 Serum ceruloplasmin measurement by itself is not a sufficient diagnostic test for Wilson disease. A low serum level of ceruloplasmin is not unique to Wilson disease; synthesis of ceruloplasmin may be reduced in other types of chronic liver disease, intestinal malabsorption, nephrosis, and malnutrition. Furthermore, a low ceruloplasmin concentration is found in at least 10% of heterozygotes for Wilson disease. Almost complete absence of ceruloplasmin is found in hereditary aceruloplasminemia, a rare autosomal recessive condition that is associated with neurologic, retinal, and pancreatic degeneration caused by iron accumulation in the brain, retina, and pancreas.25,26 Anemia and an increase in the plasma ferritin level are observed. Aceruloplasminemia has confirmed the important function of ceruloplasmin as a ferroxidase that oxidizes iron for transport from ferritin to transferrin, a function proposed in the late 1960s. Targeted disruption of the ceruloplasmin gene in a mouse model has confirmed the critical role of ceruloplasmin in transporting iron out of cells.27 Rarely, patients with Wilson disease who undergo rigorous, prolonged chelation therapy may resemble those with clinical aceruloplasminemia if ceruloplasmin oxidase activity is reduced to undetectable levels.28 The concurrence of Wilson disease and mutation-confirmed hereditary hemochromatosis has been reported.29 In patients with Wilson disease, the serum copper concentration is low, in parallel with the low serum cerulo­plasmin level. The non–ceruloplasmin-bound copper concentration, which can be estimated by subtracting the amount of copper associated with ceruloplasmin from the total serum copper, is elevated. The amount of ceruloplasmin-bound copper (in µg/dL) is estimated by multiplying serum ceruloplasmin (in mg/dL) by 3.15 (amount of copper [in µg] per mg of ceruloplasmin). In normal persons the non–ceruloplasmin-bound copper concentration is <15 µg/ dL. For results reported under the conventions of the Système Internationale, the conversion factor is the same, but reference values are converted to µg/L (by multiplying by 10); total serum copper reported in µmol/L must be converted to µg/L by multiplying that value by 63.5, the molecular weight of copper. In Wilson disease, the concentration of non–cerulopasmin-bound copper is more than 20 µg/dL (200 µg/L), and even ten times higher in patients with fulminant liver failure with intravascular hemolysis. The usefulness of this calculation, which is highly dependent on the accuracy of the copper and ceruloplasmin measurements, has not been validated as a diagnostic criterion. Serum uric acid and phosphate concentrations may be low in patients with untreated Wilson disease, reflecting renal tubular dysfunction. Urinalysis may show micro-

Table 75-1  Biochemical Parameters in Patients with Wilson Disease and in Normal Adults

Serum ceruloplasmin (mg/L) Serum copper (µg/L)   (µmol/L) Urinary copper (µg/d)   (µmol/d) Liver copper (µg/g dry weight)

NORMAL ADULTS

WILSON DISEASE*

200-350 700-1520 11-24 <40 <0.6 20-50

0-200 190-640 3-10 100-1000 >1.6 >200

*For all the assays, results in homozygotes and heterozygotes may overlap in some cases.

1253

1254

Section IX  Liver scopic hematuria; if possible, aminoaciduria, phosphaturia, and proteinuria should be quantified. Studies of urinary copper excretion, preferably with three separate 24-hour collections, have proved useful for diagnosis. Urinary copper excretion reflects the non– ceruloplasmin-bound copper concentration in plasma. The collection must be complete, and the volume and total creatinine excretion should be measured; precautions against contamination with copper in the collection process are essential. The basal 24-hour urinary copper excretion is elevated at least two to three times normal in the vast majority of patients. The conventional cutoff criterion of 100 µg/day, although typical, is not sufficiently sensitive. A patient with a basal 24-hour urinary copper excretion of greater than 40 µg/day merits further investigation for Wilson disease. Heterozygotes usually have a normal 24-hour urinary copper excretion, although the value may be borderline abnormal in some cases. Although a normal person may excrete as much as 20 times the baseline level of copper after administration of penicillamine, a patient with symptomatic Wilson disease will excrete considerably more. Urinary copper excretion of 2.5 µmol (1600 µg) or more per 24 hours is diagnostic of Wilson disease. This provocative test lacks sensitivity for identifying presymptomatic affected siblings.30 Hepatic tissue copper concentration, which usually is measured by neutron activation analysis or atomic absorption spectrometry, may provide important diagnostic information. A hepatic copper content greater than 250 µg per g dry weight of liver is considered diagnostic of Wilson disease. On the basis of a large series of geneticallydiagnosed patients, a value of greater than 70 µg per g dry weight has been proposed as a better diagnostic threshold, although some specificity is lost.31 Hepatic parenchymal concentrations of less than 40 µg per g dry weight are regarded as strong evidence against a diagnosis of Wilson disease. Liver biopsy samples must be collected without extraneous copper contamination, but in general ordinary disposable liver biopsy needles can be used. The sample submitted must be adequate in size. In the early stages of Wilson disease, when copper is distributed diffusely in the liver cell cytoplasm, this measurement may clearly indicate hepatic copper overload. In later stages of hepatic Wilson disease, the measurement of hepatic copper is less reliable because copper is distributed unequally in the liver. Moreover, liver biopsy may not be safe in such patients because of coagulopathy or ascites—therefore this diagnostic para­ meter may not be available. Some heterozygotes have minor elevations of liver tissue copper. An elevated hepatic copper concentration is not specific for Wilson disease; patients with chronic cholestasis or diseases such as Indian childhood cirrhosis may have elevated hepatic copper levels. In view of the numerous available diagnostic tests, a methodical approach is required. The classic patient with Wilson disease, whether displaying hepatic or neurologic findings, may be considered to be someone between 6 and 40 years of age with serum ceruloplasmin level less than 5 mg/dL (<50 mg/L) and definite Kayser-Fleischer rings. Otherwise, in the presence of chronic liver disease (indicated by hepatomegaly or biochemical abnormalities) or typical neurologic symptoms, the combination of a low serum ceruloplasmin level (less than 140 mg/dL32) and elevated basal 24-hour urinary copper excretion (>40 µg/day) is highly suggestive of Wilson disease. The measurement of 24-hour urinary copper excretion after administration of penicillamine may be so definitive that the diagnosis is no longer in doubt. Typical ocular findings complete the clinical diagnosis but may be lacking. A percutaneous

liver biopsy is useful for assessing the severity of liver damage and measuring parenchymal copper concentration, which is regarded by some to be the sine qua non for the diagnosis of Wilson disease. This procedure, however, may have to be delayed in patients with severe liver dysfunction. Other clinical entities in the differential diagnosis must be appropriately excluded. In the patient who does not have classic manifestations, extensive studies must be pursued meticulously; ultimately a gene mutation analysis may be the only convincing and reliable diagnostic procedure.

MUTATION ANALYSIS

More than 500 reported mutations in the ATP7B gene have been detected in many different populations since the original mutations were described. These mutations are recorded in the Wilson Disease Mutation Database (http://www. wilsondisease.med.ualberta.ca/database.asp), as described by Kenney and Cox.33 This database includes references and population sources. The usual approach to detection of a mutation is high-throughput sequencing of either selected or all exons of the gene. Although identification of a mutation is technically straightforward, care must be taken that the change detected causes disease and is not a rare normal variant, particularly for missense mutations, in which a single amino acid is substituted. Because of the similarity between yeast and mammalian copper transport systems, yeast and cell assay systems have been developed for the functional assessment of variants.34-36 The identification of one mutation may be adequate to confirm the diagnosis, if characteristic clinical symptoms and at least some of the biochemical features are present and if the one mutation detected is clearly established as a diseasecausing mutation. The majority of mutations in ATP7B identified to date are missense mutations. Small deletions, insertions, nonsense, and splice-site mutations occur throughout the gene (Cox, Mutation Database). Large gene deletions, which are found in about 20% of patients with Menkes disease, are rare, and the mutation spectrum for the Menkes gene, ATP7A, is different from that for ATP7B.37 Various ethnic groups have different specific mutations. The common histidine1069glutamine (His1069Gln) mutation3 is present, at least in the heterozygous state, in 35% to 75% of Europeans with Wilson disease; the higher number is relevant only for eastern Europe.38 Exon 8 of the gene is particularly rich in mutations in European populations. The mutation arginine778leucine is common in Chinese populations.39 Because no mutation is present in high frequency in Japanese and Mediterranean populations, mutation detection is more challenging in such populations. In populations with ethnic homogeneity or in which the spectrum of mutations is established, testing strategies can identify the mutations in more than 90% of patients, as in Sardinians, in whom the disease frequency is 1 in 7000 live births.40 In populations with a limited number of mutations, use of a “Wilson disease chip”41 may be cost-effective. If the patient is clinically normal, has only mild signs of the disease, or has a late age of onset, the possibility exists that the patient could actually be a heterozygote. Up to now, however, heterozygotes have not been known to become clinically affected or to require treatment. Gene deletions, duplications, nonsense mutations, and splice-site mutations would be predicted to prevent almost completely the formation of the gene product and thus to produce a severe defect. This prediction is true to some extent,39,42 and the onset of liver disease at three years of age has been reported in a patient with a severe gene defect that

Chapter 75  Wilson Disease was predicted to abolish the gene product.43 The common His1069Gln mutation tends to be associated with neurologic disease and later onset44; however, this mutation has been reported in homozygotes as young as nine years of age with hepatic disease. The positions of this and other missense mutations (amino acid substitutions) are shown in Figure 75-4. Most patients are compound heterozygotes, carrying two different mutations for the gene. Because homozygotes for a single mutation are relatively infrequent, correlation of clinical features with specific mutations is difficult. With the opportunity of confirming a diagnosis of Wilson disease by direct identification of mutations, the spectrum of manifestations of Wilson disease has been found to be even wider than previously recognized. No individual biochemical test is reliable for the identification of patients. In some cases, even all combinations of tests prove inadequate for a diagnosis. For example, Kayser-Fleischer rings, once thought to occur inevitably in the presence of neurologic symptoms, are not necessarily present even in conjunction with well-established neurologic symptoms. The serum ceruloplasmin concentration, once thought to be reduced in a great majority of patients, may be normal in a major proportion of patients with hepatic manifestations of Wilson disease. The use of molecular tests in patients with any clinical symptoms of the disease may become routine in the near future and is already feasible in some populations. Because of the difficulty in distinguishing asymptomatic patients from heterozygotes, mutation analysis should be carried out if possible.

D13S314 D13S301 WND D13S133 D13S316 D13S137

D13S314 9 D13S301 7 WND W D13S133 17 D13S316 3 D13S137 2 Affected (proband)

3 6 W 17 6 9

11 4 N 17 7 9

9 7 W 17 3 2

9 6 N 15 2 2

9 6 N 15 2 2 Normal

11 9 4 7 N W 17 17 7 3 9 2

3 6 W 17 6 9

3 6 W 17 6 9

9 7 W 17 3 2

Presymptomatic Diagnosis of Siblings

If mutations have been identified in a patient, mutational analysis is easily carried out in siblings (sibs) by direct testing for the mutations found in the patient. If mutations have not been identified, accurate diagnosis can be achieved using markers flanking the gene. The most useful genetic markers are stretches of dinucleotides or trinucleotides, which show such extensive variability in the normal population so that parents within any one family will carry different alleles of these markers. This variability allows the tracking of the disease gene as it segregates within families, as shown in Figure 75-4. It is important that informative markers flank the gene because an erroneous diagnosis could rarely result if markers on only one side of the gene are informative and a recombinant event has occurred close to the gene. The combination of markers, or haplotype, reliably indicates the genetic status within the family. According to marker studies, an occasional person considered, as a result of biochemical testing, to have a high probability of being a presymptomatic patient has been shown to be a heterozygote. Therefore, confirmation of the genotype is highly recommended before treatment is initiated. If a heterozygote is found to accumulate copper stores in later decades of life, he or she may need to be followed for signs of copper overload. In the absence of marker analysis, screening should include physical examination, liver biochemical tests, serum copper and ceruloplasmin measurements, a 24-hour urinary copper determination, and a careful slit-lamp examination. Children 6 years old or under who appear to be unaffected should be rechecked at yearly intervals over the next 5 to 10 years. Genetic screening with the use of flanking markers, or by direct mutation analysis, however, is the most reliable way to identify an affected sibling when the patient’s DNA is available for mutational analysis. For deceased patients, tissue from autopsy or biopsy material can be used.

TREATMENT

11 4 N 17 7 9

Affected Heterozygote

Figure 75-4.  Diagnostic use of polymorphic DNA microsatellite markers for siblings of a confirmed patient with Wilson disease in a pedigree. DNA markers are listed in centromeric to telomeric order. Three markers are usually sufficient for an unambiguous result: D13S314, D13S301, and D13S316. Numbers represent alleles of each marker listed. The proband (arrow) and presymptomatic sibling confirmed as affected are shown as filled circles.

Three treatments for Wilson disease are generally recognized: d-penicillamine, trientine, and zinc (Table 75-2). Chelation with tetrathiomolybdate is a relatively new and still experimental option. With effective lifelong chelation treatment, most patients live normal, healthy lives. Starting treatment early is critical, and the outcome is best for patients in whom the disease is diagnosed and treatment begun when the patient is presymptomatic. Whether routine institution of chelation therapy in infancy (<2 years old) is advantageous or deleterious remains unknown. Likewise, the potential role of gene transfer therapy remains uncertain.45 Dietary treatment alone is ineffective, but most patients should eliminate copper-rich foods from the diet. These foods include organ meats, shellfish, nuts, chocolate, and mushrooms. Vegetarians require specific dietary counseling. If the concentration of copper in the patient’s drinking water is believed to be high, the water should be analyzed and a copper-removing device should be installed in the plumbing system. Penicillamine, introduced in 1956 by J. M. Walshe, is effective in most patients with Wilson disease. Penicillamine, which is the sulfhydryl-containing amino acid cysteine substituted with two methyl groups, greatly increases urinary excretion of copper. Only the d-penicillamine form is used clinically. Studies in the LEC rat model indicate that penicillamine inhibits the accumulation of copper

1255

1256

Section IX  Liver Table 75-2  Approach to the Treatment of Wilson Disease DRUG

DOSE*

Penicillamine (+ pyridoxine 25 mg daily)

Initial: 1-1.5 g/day (adults) or 20 mg/kg/day (children) divided into two or three doses

Trientine† Zinc

Maintenance: 0.75-1 g/day as needed to maintain cupruresis Initial: 1-1.2 g/day divided into two or three doses‡ Maintenance: Same Initial: 50 mg elemental zinc three times daily (adults)§ Maintenance: Titrate dose based on efficacy monitoring data¶

TESTS FOR MONITORING EFFICACY

TESTS FOR MONITORING SIDE EFFECTS

24-hour urinary copper: 200-500 µg (3-8 µmol)/day as target; estimated non–ceruloplasminbound copper <200 µg/L Estimated non–ceruloplasminbound copper <100 µg/L Same as for penicillamine

Complete blood count; urinalysis; skin examination

Same as for penicillamine 24-hour urinary copper: <75 µg (1.2 µmol)/day as target Estimated non–ceruloplasminbound copper <100 µ/L

Complete blood count; iron studies Serum zinc level

*All medications should be taken before or after mealtime if possible, but the timing of the dose may need to be adjusted to enhance compliance. † Requires refrigeration. ‡ The dose of trientine in children is not established (approximately 20 mg/kg/day). § The dose of zinc in children is not established. ¶ The 24-hour urinary copper excretion reflects total body copper load and thus can be used to monitor zinc treatment even though zinc does not cause cupruresis; some groups prefer to use the estimated non–ceruloplasmin-bound copper determination as a guide.

in hepatocellular lysosomes and solubilizes copper for mobilization from these particles, but not from cytoplasmic metallothioneins.46 In addition to its chelating action, penicillamine inhibits collagen cross-linking and has some immunosuppressive properties. The neurologic status of patients with mainly neurologic symptoms may worsen initially after treatment with penicillamine is started47; most, but not all, recover with continued use of penicillamine. A febrile reaction with rash and proteinuria develops in some patients within 7 to 10 days of beginning treatment. Penicillamine can be restarted slowly, along with gluco­ corticoids, but changing to an alternative chelator is preferred. Penicillamine, although effective, can cause serious side effects. Adverse reactions involving the skin include various types of rashes, pemphigus, and elastosis perforans serpiginosa. Hypothyroidism has been reported. Other side effects vary from minor (loss of taste, gastrointestinal upset, and arthralgias) to severe (proteinuria, leukopenia, or thrombocytopenia). Aplastic anemia occurs rarely and does not always reverse when penicillamine is stopped. Nephrotic syndrome, Goodpasture’s syndrome, a myasthenia syndrome, and a systemic disease that resembles systemic lupus erythematosus have all been reported. These severe side effects require immediate discontinuation of penicillamine and use of a different chelator. Whether lifelong treatment with penicillamine is free of adverse consequences is not yet clear. Patients who have taken penicillamine for 30 to 40 years may have chronic skin changes with loss of elastic tissue. Whether the antifibrotic effect weakens other connective tissues is not known. Theoretically, chronic depletion of other trace metals may occur. In view of these side effects, penicillamine should be used in the lowest effective dose. Trientine, or triethylene tetramine dihydrochloride (2,2,2-tetramine), the official short name of which is “trien,” is the usual second-line treatment for patients who are intolerant of penicillamine. Trientine differs chemically from penicillamine in lacking sulfhydryl groups. Copper is chelated by forming a stable complex with the four constituent nitrogens in a planar ring. Trientine increases urinary copper excretion and may interfere with intestinal absorp-

tion of copper. Trientine is a less potent chelator than penicillamine, but the difference is not clinically important. Trientine produces little significant toxicity in patients with Wilson disease—apart from causing occasional gastritis and inducing iron deficiency, apparently by chelating dietary iron. Bone marrow suppression is extremely rare. Adverse effects of penicillamine resolve and do not recur during treatment with trientine.48 Neurologic worsening after treatment with trientine is begun has rarely been reported. Trientine is highly effective, even in patients with advanced liver fibrosis or as initial treatment in children.49 Oral zinc, used in Europe since the 1970s,50 has been investigated extensively as a treatment modality in North America. Its mechanism of action is entirely different from that of the chelators. In pharmacologic doses, zinc interferes with the absorption of copper from the gastrointestinal tract and increases the excretion of copper in the stool. The postulated mechanism of action is through the induction of metallothionein in enterocytes. The metallothionein has a greater affinity for copper than for zinc and preferentially binds copper from the intestinal contents. Once bound, the copper is not absorbed but is lost in the feces as enterocytes are shed during normal turnover.51 Additionally, zinc may interfere with lipid peroxidation and enhance the availability of glutathione.51 Problems with zinc therapy include gastritis, which is a common side effect, and uncertainty about dosing. Using a zinc salt other than zinc sulfate may minimize gastritis; any zinc salt is equally acceptable for the treatment of Wilson disease. Food interferes with the effectiveness of zinc, and some investigators recommend that no food be eaten for one hour before or after a dose of zinc is taken. This dosing regimen tends to increase the severity of gastritis and may be sufficiently inconvenient to compromise compliance, as in adolescents. An alternative approach is to be less rigorous about avoiding zinc at mealtimes and to titrate the dose according to the urinary copper excretion or serum non–ceruloplasmin-bound copper concentration. Treatment with zinc appears to have few side effects52 but may not be effective in all patients. Rare patients experience a deterioration in hepatic Wilson disease when started on zinc. Zinc may have immunosuppressant effects and reduce

Chapter 75  Wilson Disease leukocyte chemotaxis. Studies in rats suggest a possible interference with bone formation. The long-term effectiveness of zinc requires further investigation, but current data indicate that zinc is effective as maintenance therapy and has low toxicity. For patients who present with decompensated chronic liver disease, combining zinc with a conventional chelator (preferably trientine) has become a popular treatment strategy despite a lack of extensive validation. The two types of treatments must be temporally dispersed through the day, with at least four to five hours between administration of the two drugs, or else they may neutralize each other. This intensive induction regimen is best suited to patients with severe hepatic or neurologic disease53 and remains investigational. Some patients with severe hepatic Wilson disease may fail on this regimen and require urgent liver transplantation. Ammonium tetrathiomolybdate may be especially suitable for treatment of severe neurologic Wilson disease because, unlike penicillamine, it is not associated with early neurologic deterioration.54 Tetrathiomolybdate interferes with the absorption of copper from the intestine and binds to plasma copper with high affinity. Unlike penicillamine, tetrathiomolybdate has been found in LEC rats to remove copper from metallothionein at low doses; at higher doses, an insoluble copper complex is deposited in the liver.55 Although tetrathiomolybdate is regarded as non-toxic, bone marrow suppression and hepatotoxicity are noteworthy adverse side effects. Little is known about where the mobilized copper and molybdate may be deposited. The optimal dose and length of treatment, as well as long-term side effects, require further study. Such a potent copper-binding drug could produce copper deficiency. Antioxidants may be a useful adjunct for preventing tissue damage. Studies in copper-loaded animals and in patients with Wilson disease indicate that copper enhances free radical production in tissues and that this effect may be an important cause of liver damage.56 Oxidative damage may be reflected in mutations of the TP53 tumor suppressor gene, increased activity of nitric oxide synthase in patients with hepatic Wilson disease,55 and DNA strand breaks in the brains of LEC rats.57 Oxidative effects of excess copper may contribute to development and progression of liver damage.58 Alpha-tocopherol may be beneficial adjunctive treatment for patients with severe hepatic decompensation. For pregnant patients with Wilson disease, treatment must be continued throughout pregnancy. Postpartum hepatic decompensation is a risk if treatment is stopped completely during pregnancy. Although many successful pregnancies have occurred during treatment with penicillamine, the drug is officially classified as a teratogen. Occasional reports of severe collagen defects in the offspring of a patient treated with penicillamine may be caused in part by copper deficiency as a result of prolonged aggressive treatment as well as the teratogenic effects of penicillamine.59 Treatment with zinc may be less likely to produce adverse effects on developing collagen in the fetus. The safety of trientine during pregnancy is unknown, apart from favorable anecdotal reports. Judicious reduction of the dose of penicillamine or trientine by approximately 25% of the pre-pregnancy dose is advisable, especially if delivery by cesarean section is anticipated. Family screening, as discussed earlier, is an important and essential preventive measure. Early initiation of continuous treatment can prevent clinical manifestations from developing in asymptomatic patients with Wilson disease.

PROGNOSIS Patients with Wilson disease are generally regarded as having a good prognosis if the disease is diagnosed promptly and treated consistently. An asymptomatic sibling who is diagnosed on biochemical or genetic grounds before any sign of clinical impairment appears to have the best longterm outlook. Patients with early hepatic disease have a generally favorable prognosis as long as treatment is consistent and well tolerated. Severe neurologic disease may not resolve entirely on treatment. The role of liver transplantation in Wilson disease is limited (see Chapters 93 and 95). Fulminant liver failure in a patient with Wilson disease necessitates liver transplantation (see earlier). Some patients with severe liver disease that is unresponsive to drug therapy may also proceed to early transplantation. The outcome is favorable, with one-year survival rates of 80% to 90% and excellent survival beyond one year.60,61 Severe neurologic disease may improve after liver transplantation, but experience is conflicting14 and liver transplantation cannot be recommended for neurologic Wilson disease. Patients with neurologic or psychiatric manifestations of Wilson disease appear to have poor outcomes after liver transplantation and adhere poorly to medical regimens.62 Therefore, liver transplantation should be reserved for patients who present with severe, decompensated liver disease that is unresponsive to therapy or with fulminant liver failure. Live donorrelated liver transplantations, even when the graft is from a family member who may be a heterozygote, have been found to yield adequately functioning grafts.63,64 If the donor is the brother or sister of the patient, genetic analysis of ATP7B should be included in the assessment of the donor. Patients who stop taking chelating treatment have a poor prognosis. New neurologic abnormalities, such as dysarthria, may develop. Rapidly progressive hepatic decompensation has been observed and occurs on average within three years and as early as eight months after treatment is stopped. The liver damage is usually refractory to reinstitution of chelation therapy. Affected patients require liver transplantation. The quality of life of patients with Wilson disease may be compromised by drug toxicity. Anecdotal observations suggest that damage to collagen may accrue over decades in patients who are maintained indefinitely on penicillamine, but the risk has not been assessed adequately. Deficiencies in trace metals may develop with the use of any chelator, although whether these deficiencies are clinically important is not yet clear. Abnormal iron metabolism, leading to hepatic iron overload and anemia, can be predicted if ceruloplasmin oxidase activity is reduced to zero. Patients should be encouraged to maintain a healthy lifestyle, including avoidance of alcohol and obesity.

KEY REFERENCES

Brewer GJ, Dick RD, Johnson VD, et al. Treatment of Wilson’s disease with zinc: XV long-term follow-up studies. J Lab Clin Med 1998; 132:264-78. (Ref 52.) Bull PC, Thomas GR, Rommens JM, et al. The Wilson disease gene is a putative copper transporting P-type ATPase similar to the Menkes gene. Nat Genet 1993; 5:327-37. (Ref 2.) Emre S, Atillasoy EO, Ozdemir S, et al. Orthotopic liver transplantation for Wilson’s disease: A single-center experience. Transplantation 2001; 72:1232-6. (Ref 61.) Ferenci P, Steindl-Munda P, Vogel W, et al. Diagnostic value of quantitative hepatic copper determination in patients with Wilson’s disease. Clin Gastroenterol Hepatol 2005; 3:811-18. (Ref 31.)

1257

1258

Section IX  Liver Harris ZL, Takahashi Y, Miyajima H, et al. Aceruloplasminemia: Molecular characterization of this disorder of iron metabolism. Proc Natl Acad Sci USA 1995; 92:2539-43. (Ref 25.) Hsi G, Cullen LM, Macintyre G, et al. Sequence variation in the ATPbinding domain of the Wilson disease transporter, ATP7B, affects copper transport in a yeast model system. Hum Mutat 2008; 29:491501. (Ref 36.) Kenney SM, Cox DW. Sequence variation database for the Wilson disease copper transporter, ATP7B. Hum Mutat 2007; 28:1171-7. (Ref 33.) Korman JD, Volenberg I, Balko J, et al. Screening for Wilson disease in acute liver failure: A comparison of currently available diagnostic tests. Hepatology 2008; 48:1167-74. (Ref 17.)

Merle U, Schaefer M, Ferenci P, Stremmel W. Clinical presentation, diagnosis and long-term outcome of Wilson’s disease: A cohort study. Gut 2007; 56:115-20. (Ref 47.) Roberts EA, Schilsky ML. Diagnosis and treatment of Wilson disease: An update. Hepatology 2008; 47:2089-111. (Ref 14.) Wilson DC, Phillips MJ, Cox DW, Roberts EA. Severe hepatic Wilson’s disease in preschool-aged children. J Pediatr 2000; 137:719-22. (Ref 43.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

76 Other Inherited Metabolic Disorders of the Liver Mike A. Leonis and William F. Balistreri

CHAPTER OUTLINE Clinical Features of Metabolic Liver Disease  1259 α1-Antitrypsin Deficiency  1260 Glycogen Storage Diseases  1262 Type I  1262 Type III  1264 Type IV  1264 Congenital Disorders of Glycosylation  1265 Porphyrias  1266

Inborn errors of metabolism encompass a vast variety of disorders with myriad presentations including liver disease and are characterized by complex pathophysiology. Metabolic liver diseases may manifest as acute, lifethreatening illnesses in the neonatal period or as chronic liver disease in adolescence or adulthood, with progression to liver failure, cirrhosis, or hepatocellular carcinoma. In the 2007 report of the Scientific Registry of Transplant Recipients in the United States, 3% of all liver transplantations in the United States were performed because of complications resulting from metabolic liver disease.1 When the pediatric population alone is analyzed, this percentage is substantially higher. Nationwide, more than 20% of the liver transplantations performed over the five-year period from 1995 to 2000 were for complications of metabolic liver disease.2 Liver transplantation has become a life-saving measure for many patients with metabolic liver diseases. New nontransplant treatment options have become available that may, in certain cases, obviate the need for liver transplantation and thereby help alleviate the shortage of donor organs.3,4

CLINICAL FEATURES OF METABOLIC LIVER DISEASE The diverse presenting symptoms of metabolic liver disease are listed in Table 76-1. Certain metabolic liver diseases in young patients may mimic other illnesses, such as acute infections and intoxications. By contrast, the older patient with metabolic liver disease may present with symptoms of chronic disease. Because metabolic diseases can resemble multiple other disorders, a high index of suspicion is required for correct diagnosis. An infant presenting with cholestasis should undergo an evaluation for metabolic liver disease. (The approach to the

Tyrosinemia  1269 Urea Cycle Defects  1271 Arginase Deficiency  1274 Bile Acid Synthesis and Transport Defects  1274 Bile Acid Synthesis Defects  1274 Bile Acid Transport Defects  1276 Cystic Fibrosis  1277 Mitochondrial Liver Diseases  1277

patient with jaundice is discussed in detail in Chapter 20.) Any patient with progressive neuromuscular disease, developmental delay, or regression of developmental milestones also requires evaluation. Metabolic liver disease should be an immediate consideration in patients with elevated serum aminotransferase levels, hepatomegaly, acidosis, hypoglycemia, ascites, bleeding diathesis, hyperammonemia, coma, recurrent vomiting, or poor weight gain. A detailed history can often raise the possibility of metabolic liver disease and help direct the investigation. A family history of consanguinity, multiple miscarriages, or early infant deaths may suggest a metabolic derangement. Close relatives with undiagnosed liver disease, progressive neurologic or muscle disease, or undiagnosed developmental delays should also raise suspicion. A carefully obtained dietary history is also important in dissecting the nature of the illness; introduction of certain foods may correlate with the onset of symptoms, as in patients with urea cycle defects, galactosemia, or fructosemia. A history of specific dietary aversions may also be revealing.5 Recommended initial screening tests, which should direct further diagnostic evaluation, are listed in Table 76-2. Because patients with metabolic liver disease often present with acute and recurrent symptoms, it is of utmost importance that the physician obtain the diagnostic studies as soon as possible and at the time the patient is experiencing symptoms. The laboratory values for many of these illnesses may normalize between acute episodes. In enigmatic cases, serum and urine samples should be obtained during the acute illness and saved (frozen) for definitive studies, if available. A liver biopsy can also be a valuable diagnostic tool. If a metabolic liver disease is suspected, in addition to obtaining specimens for standard histology, a frozen specimen should be saved and a sample prepared for later electron microscopic study to look at the subcellular organelles, which may exhibit characteristic changes in some disorders.

1259

1260

Section IX  Liver Table 76-1  Presenting Features of Metabolic Liver Disease Symptoms

Signs

Other Findings

Coma Developmental delay Growth failure Hyperammonemic symptoms Hypoglycemic symptoms Neurologic or motor skill deterioration Recurrent vomiting Seizures Ascites Abdominal distention Cardiac dysfunction Cataracts Dysmorphic features Hepatomegaly Hypotonia Jaundice Short stature Splenomegaly Unusual odors Acidosis Cholestasis Coagulopathy Fulminant hepatic failure Ketosis Rickets

Table 76-2  Screening Laboratory Studies for Metabolic Liver Disease* Serum

Urine

Albumin Alkaline phosphatase Amino acids Aminotransferases Ammonia Anion gap calculation Coagulation profile Electrolytes Ferritin Fractionated bilirubin Gamma glutamyl transpeptidase Glucose Lactate† Peripheral blood smear Pyruvate† Uric acid† Organic acids Orotic acid Reducing substances

*Specimens of serum and urine obtained during acute episodes should be saved for later studies. † Obtain if patient is acidotic or has neurologic symptoms.

α 1 -ANTITRYPSIN DEFICIENCY Deficiency of α1-antitrypsin (α1-AT) is transmitted in an autosomal recessive fashion and leads to an increased risk of lung and liver disease. This deficiency is one of the most common genetic diseases in the world and the second most common metabolic disease affecting the liver, the most common being hereditary hemochromatosis (see Chapter 74).6,7 The following discussion focuses on the effects of α1-AT deficiency on the liver.

Pathophysiology

The prototypical member of the serpin family of protease inhibitors, α1-AT binds with and promotes the degradation of serine proteases in the serum and tissues. The most

important of these serine proteases is neutrophil elastase, which is inhibited by α1-AT through formation of a tight 1 : 1 α1-AT-to-elastase complex. Loss of serum α1-AT activity, as occurs in the most common form of α1-AT deficiency, leads to uninhibited neutrophil elastase activity and is one of the primary mechanism for the premature development of pulmonary emphysema in affected patients.8 Another proposed mechanism derives from the observation that α1ATZ polymers are proinflammatory toward neutrophils and are found in lung lavage fluid of patients with the protease inhibitor (Pi) ZZ phenotype of α1-AT deficiency (see later) and emphysema; moreover, instillation of α1-ATZ polymers in murine lung leads to an influx of neutrophils to the lung.8,9 Allelic α1-AT mutant variants produce Pi gene products that can be distinguished from the normal product by electrophoretic methods; the normal allelic representation is designated PiM. The PiZ variant produces a mutant α1-ATZ protein. Homozygosity at the PiZ allele is the most common and classic pathologic form of α1-AT deficiency and is capable of leading to liver and lung disease. The α1-ATZ molecule represents a single amino acid replacement of glutamine with a lysine residue as a result of a mutation at position 342 of the α1-AT (SERPINA1) gene. More than 100 naturally occurring variants of α1-AT have been described. Although most of these variants are either of no clinical significance or are extremely rare, two additional variants, PiSiiyama and PiMmalton, have been reported to be associated with liver injury and even cirrhosis.8,10-12 α1-AT is produced almost exclusively in the rough endoplasmic reticulum (ER) of hepatocytes and is subsequently targeted to the secretory pathway via the Golgi apparatus. Structural misfolding and polymerization of the mutant α1ATZ protein causes its aberrant retention in the ER, failure of progression through the secretory pathway, and diminished intracellular degradation. In persons with the phenotype PiZZ, serum α1-AT activity levels are reduced to less than 15% of normal; this loss of function accounts for the development of pulmonary disease. Studies suggest that the rate of intracellular degradation may itself be genetically determined and may influence the expression of disease; α1-ATZ appears to be degraded more slowly in the ER of PiZZ patients who are susceptible to liver disease than in PiZZ patients who are not susceptible to liver disease.13 The exact mechanism for α1-ATZ-induced liver injury is not known. Studies of transgenic mice that express the human ATZ gene suggest a gain-of-function mechanism by which retention in the ER and accumulation in hepatocytes of mutant α1-ATZ is responsible for hepatotoxicity.14 Multiple intracellular signaling pathways, including caspase activation, ER stress responses, and the autophagic response, are activated by the retention of malformed proteins in the ER.15 Autophagy is an intracellular degradative pathway that targets proteins and organelles for destruction during development as well as at times of stress or nutrient deprivation. One hypothesis under investigation is that increased stimulation of autophagy by accumulation of α1-ATZ protein leads to ongoing mitochondrial injury, cellular apoptosis, and a chronic cycle of hepatocellular death and regene­ ration that may, over time, lead to liver injury and cirr­ hosis.7,16 Other genetic and environmental modifiers of the ER “quality control” process for handling mutant α1-ATZ protein are undoubtedly involved and account for the wide variation in clinical phenotype observed in the hepatic presentation of PiZZ patients.7

Clinical Features

Although the prevalence of the classic α1-AT deficiency allele, PiZ, is highest in populations derived from northern

Chapter 76  Other Inherited Metabolic Disorders of the Liver European ancestry, many racial subgroups are affected worldwide, and millions of persons have combinations of deficiency alleles (i.e., PiSS, PiSZ, or PiZZ).8,17 In the United States, the overall prevalence of deficiency allele combinations is approximately 1 in 490 (i.e., 1 in 1058 for PiSS, 1 in 1124 for PiSZ, and 1 in 4775 for PiZZ).18 Mounting evidence suggests that heterozygous α1-AT deficiency states can contribute to the development of cirrhosis and chronic liver failure in adults through mechanisms similar to those encountered with the PiZZ phenotype.10-12,19 In addition, the heterozygous α1-AT deficiency state may contribute to worsening of chronic liver disease caused by hepatitis C viral infection or nonalcoholic fatty liver disease in adults, as well as cholestatic liver diseases in children.7,19-22 In the most unbiased study to date, reported by Sveger,23 on the epidemiology of liver disease in patients with α1-AT deficiency, 200,000 Swedish infants were screened for α1-AT deficiency, 184 infants were found to have abnormal allelic forms of α1-AT (127 PiZZ, 2 PiZnull, 54 PiSZ, and 1 PiSnull), and 6 (5 PiZZ and 1 PiSZ) died in early childhood, although only 2 from cirrhosis.23 Although about 10% of newborns with α1-AT deficiency (PiZZ) present with cholestasis and as many as 50% continue to have elevated serum aminotransferase levels at age three months, most are clinically asymptomatic.23,24 Liver disease does not develop in patients with null α1-AT phenotypes, whereas earlyonset emphysema develops in all of them.25 Therefore, the prognosis for patients with liver disease manifesting in infancy as a result of α1-AT deficiency (PiZZ) is highly variable. Even those children in whom cirrhosis develops can have a highly variable progression to end-stage liver disease (ESLD), which infrequently leads to liver transplantation.26 Moreover, siblings with PiZZ have a variable degree of liver involvement; in a study reported by Hinds and colleagues, five of seven children with PiZZ requiring liver transplantation had siblings with PiZZ who had no persistent liver involvement.27 This finding suggests that environmental or additional genetic factors must be involved in determining the severity of liver disease in α1-AT deficiency; this area is under active scientific investigation.28 Of 150 patients with α1-AT deficiency from Sveger’s original study23 who subsequently underwent evaluation at age 16 and 18 years, none had clinical signs of liver disease. Elevated serum aminotransferase or gamma glutamyl transpeptidase (GGTP) levels were found in fewer than 20% of patients with a PiZZ phenotype and in fewer than 15% of those with a PiSZ phenotype.24 By the third decade of life, analysis of this same cohort of affected persons showed that 6% of PiZZ and 9% of PiSZ patients had a marginal increase in serum alanine aminotransferase levels.29 Even though liver disease is often (but not always) mild during infancy and childhood, patients with α1-AT deficiency have an increased risk for development of cirrhosis during adulthood; 42% of all PiZZ patients have histologic evidence of cirrhosis at autopsy.30,31 Moreover, homozygous α1-AT deficiency raises the risk for development of hepatocellular carcinoma, especially in men older than 50 years.31 The diagnosis of α1-AT deficiency should be considered in any patient presenting with noninfectious chronic hepatitis, hepatosplenomegaly, cirrhosis, portal hypertension, or hepatocellular carcinoma.

Histopathology

Histopathologic features of α1-AT deficiency change as the affected patient ages. In infancy, liver biopsy specimens may show bile duct paucity, intracellular cholestasis with or without giant cell transformation, mild inflammatory changes, or steatosis, with few of the characteristic periodic acid–Schiff (PAS)-positive, diastase-resistant glob-

ules. These inclusions, which result from polymerized α1ATZ protein, are most prominent in periportal hepatocytes, and may also be seen in Kupffer cells. Immunohistochemistry with monoclonal antibody to α1-ATZ can also be performed to verify the diagnosis. As the patient ages, these changes may resolve completely or progress to chronic hepatitis or cirrhosis.

Diagnosis

α1-AT is considered a hepatic acute-phase reactant, and its release may be stimulated by stress, injury, pregnancy, or neoplasia. Because these factors can influence α1-AT production, even in patients with the PiZZ phenotype, the diagnosis of α1-AT deficiency should be based on phenotype analysis and not solely on the serum α1-AT level.32 A liver biopsy specimen, although not universally recommended, should confirm the diagnosis. Commercial tests are available to detect the most common mutant alleles by polymerase chain reaction analysis of genomic DNA. In addition, a re-sequencing molecular array chip is available for rapid sequencing of the entire SERPINA1 gene to allow identification of rare mutant alleles.33 Screening guidelines to diagnose α1-AT deficiency in asymptomatic persons have been proposed in an effort to reduce the risk of emphysema by counseling patients to avoid smoking.34 Adults with chronic lung disease and siblings of affected patients with lung or liver disease should be targeted for screening, and appropriate education and genetic counseling should be offered to patients with α1-AT deficiency identified by screening.

Treatment

The initial treatment of α1-AT deficiency is with symptomatic care. Breast-feeding until the end of the first year of life has been suggested to decrease the manifestations of cholestatic disease, as may the use of ursodeoxycholic acid.35 The importance of providing fat-soluble vitamins, when indicated, adequate nutrition, and counseling to avoid smoking and second-hand smoke cannot be overemphasized. The role of neonatal screening for α1-AT deficiency is still not settled, although the effect on smoking practices in patients diagnosed at an early age appears to be positive. If effective therapy for liver disease caused by α1-AT deficiency becomes available, neonatal screening may become more useful for preventing the need for liver transplantation. Although progression to ESLD is uncommon, α1-AT deficiency is the most common metabolic liver disease for which liver transplantation is performed. Besides replacing the injured organ, transplantation corrects the metabolic defect, thereby preventing further progression of systemic disease. Between 1995 and 2004, 161 children and 406 adults underwent liver transplantation for ESLD associated with α1-AT deficiency in the United States; of these, 4.4% of the children and fewer than 1% of the adults were African Americans. Five-year patient survival rates following liver transplantation for the pediatric and adult patients with α1-AT deficiency were 83% and 90% respectively.36 Liver transplantation with grafts from donors with unrecognized α1-AT deficiency appears to have a comparable outcome to transplants using grafts without unrecognized liver disease.37 Replacement therapy with purified α1-AT is the only treatment option approved by the U.S. Food and Drug Administration (FDA) for lung disease associated with α1-AT deficiency. Patients who received replacement therapy in several studies, including a small randomized placebocontrolled trial, have had a slower rate of decline in lung tissue and function compared with control groups, although clinical efficacy has yet to be conclusively demonstrated.38

1261

1262

Section IX  Liver This therapy would not be expected to benefit α1-AT deficiency–associated liver disease, which results from malprocessed α1-ATZ, not loss of α1-AT protein function. Other mechanistically based treatment options aimed at influencing the stability or secretion rates of α1-ATZ within the hepatocyte ER are being investigated. Chemical chaperones such as phenylbutyric acid (PBA) markedly increase secretion of α1-ATZ in experimental in vitro and in vivo models of α1-AT deficiency.39 A small pilot study investigated the potential benefits of PBA in the treatment of children with α1-AT–deficient liver disease. Unfortunately, no statistically significant increase in serum α1-AT levels occurred in PBA-treated patients, many of whom experienced unacceptable side effects.40 Another pharmacologic approach in the early stages of investigation involves the design of small molecules to inhibit α1-ATZ protein polymerization and increase clearance of intracellular aggregates.41 α1-AT deficiency is one of many diseases for which reconstitution of the normal genotype through gene therapy is being studied. Long-term expression of human α1-AT in murine liver, at therapeutic levels, has been achieved after hydrodynamics-based intravenous injection of nonviral deoxyribonucleic acid (DNA) constructs.42 Also, delivery of a vector carrying a ribozyme designed to target human α1ATZ messenger ribonucleic acid (mRNA) in the portal vein of transgenic mice that carry the human PiZ allele led to a 50% reduction in α1-ATZ mRNA levels. In a separate experiment, administration of a vector carrying the normal PiM allele, which is resistant to ribozyme cleavage, led to longterm expression of human α1-AT protein in both liver and serum up to one year following injection.43 These results are promising first steps toward the goal of achieving successful gene replacement therapy for α1-AT deficiency.

GLYCOGEN STORAGE DISEASES More than 10 distinct inborn disorders of glycogen metabolism have been described in the literature, but only 3 are associated with significant liver disease–glycogen storage disease (GSD) types I, III, and IV.44,45 Other GSDs may cause hepatomegaly or liver histologic changes but generally do not cause clinically important liver disease. The overall incidence of GSD types I, III, and IV is estimated to be between 1 in 50,000 and 1 in 100,000 population. Glycogen metabolism occurs in many tissues, but the areas of clinical importance are the muscle and liver. The body uses glycogen as a storage system for glucose and as a ready reserve for times when systemic glucose is required (see Chapter 72). Glycogen is composed of long-chain glucose molecules arranged in a linear 1,4 linkage. From 8% to 10% of the glucose molecules are attached in a 1,6 linkage to form branching chains, which permit efficient storage of glucose while minimizing the impact on intracellular osmolality. The substrates for glycogen synthesis, glucose-6phosphate (Glu-6-P) and glucose-1-phosphate (Glu-1-P), are derived from several pathways, including fructose and galactose metabolic cycles as well as gluconeogenesis and glycogenolysis (Fig. 76-1). Through the action of uridine diphosphate glucose (UDPG) pyrophosphorylase and glycogen synthase, Glu-1-P is metabolized to UDPG and glycogen, sequentially. The 1,4 linkages can be converted to 1,6 linkages by the actions of branching enzymes. Amylo-1,6-glucosidase is a debranching enzyme that can release 8% to 10% of the glucose stored in glycogen. The remaining glucose is released as Glu-1-P

1, 4 Branched linkages Branching enzyme

Debranching enzyme Glycogen

Glycogen synthase

Amylo-1,6-glucosidase Glucose

UDP Glucose Phosphorylase a UDP glucose phosphorylase Glucose-1-Phosphate Phosphoglucomutase Glucose-6-Phosphate Glucose-6-Phosphatase Glucose Figure 76-1.  Pathway of glycogen synthesis and glycogenolysis. Enzymes are shown in italics. UDP, uridine diphosphate.

through the action of phosphorylase a and is converted to Glu-6-P by phosphoglucomutase. Phosphorylase exists in an active (a) and an inactive form (b); protein kinase is responsible for the conversion of phosphorylase b to a. Protein kinase is stimulated by epinephrine, glucagon, and fasting, and thereby increases glycogenolysis. High levels of glucose influence the conversion of phosphorylase a back to phosphorylase b, thereby decreasing glycogenolysis. Glycogen synthase also exists in active (a) and inactive (b) forms. Phosphorylase a inhibits the conversion to glycogen synthase a, thereby reducing glycogen synthesis. High levels of glycogen favor the formation of glycogen synthase b.45

TYPE I

GSD type I, the most common inborn error of glycogen metabolism, results from deficiency of a two-component enzyme system involved in the transport of Glu-6-P from the cytosol into the ER by Glu-6-P translocase and subsequent cleavage of Glu-6-P by glucose-6-phosphatase (Glu-6Pase), located on the luminal side of the ER. Clinical and molecular genetic observations have disclosed two subtypes of GSD type I, Ia and Ib, that account for virtually all cases.46 The clinical phenotype with respect to liver disease is similar in the two forms; however, patients with GSD type Ib often have intermittent severe neutropenia and neutrophil dysfunction, making them prone to recurrent episodes of severe bacterial infections and Crohn’s-like intestinal disease.47,48 Disruption of the function of Glu-6-Pase (type Ia) or Glu-6-P translocase (type Ib) inhibits the utilization of glucose by gluconeogenesis, glycogenolysis, and the metabolism of fructose or galactose. This inability to release stored glucose leads to hypoglycemia within 90 to 180 minutes of the last ingested glucose. Lactate and fatty acid metabolism and glycolytic pathways are then used as sources of energy.

Chapter 76  Other Inherited Metabolic Disorders of the Liver Clinical Features

Most patients with GSD type I present in infancy with symptoms of metabolic derangement, such as lethargy, seizures, or coma as a result of profound hypoglycemia or metabolic acidosis, a protruding abdomen caused by hepatomegaly, muscular hypotonia, and delayed psychomotor development.49 Physical signs invariably include hepatomegaly, usually with a normal-sized spleen. Patients in whom the disease is poorly controlled for a long time exhibit short stature and growth failure and may be prone to adiposity. Delayed bone age and reduced postpubertal bone mineral density are common.50 Xanthomas can appear after puberty and localize to the elbows, knees, buttocks, or nasal septum, the last leading to epistaxis. Patients with GSD type I are susceptible to a wide spectrum of brain damage that may result in epilepsy, hearing loss, and abnormal neuroradiologic findings, most likely as a result of recurrent episodes of hypoglycemia.51 Other metabolic derangements can be seen. Lactic acid levels can reach four to eight times normal; the accompanying metabolic acidosis may manifest as muscle weakness, hyperventilation, malaise, headache, or recurrent fever. Serum adenosine triphosphate (ATP) and phosphorus levels are low, secondary to an increase in purine synthesis and the inability to release phosphorus from Glu-6-P. Hyperuricemia is common and may lead to gout, arthritis, or progressive nephropathy. Nephromegaly secondary to increased glycogen deposition is common, and with advancing age, progressive renal disease, hypertension, and renal failure requiring dialysis and transplantation may develop.49 Because of hypoglycemia, patients have chronically high serum levels of glucagon with depressed levels of insulin. Hypertriglyceridemia and hypercholesterolemia are present in both GSD Ia and GSD Ib (but more prominently in GSD Ia) and may account for the greater frequency of xanthoma formation.52 In addition to the features already noted, patients with GSD type Ib often have severe intermittent neutropenia and neutrophil dysfunction as well as high platelet counts. Crohn’s-like inflammatory bowel disease often occurs in patients with GSD type Ib at the time of severe neutropenia, and patients are prone to the development of severe bacterial infections, with abscess formation at numerous locations throughout the body.47 No correlation has been found between the severity of bacterial infections or degree of neutropenia and the molecular defect in Glu-6-P transport activity in humans with GSD type Ib, suggesting that other as yet unknown factors, such as modifying genes, define the ultimate disease phenotype.53

Hepatic Involvement

Hepatomegaly in GSD type I results from increased glycogen storage in the liver as well as a large degree of fatty infiltration; the latter likely develops because of a wide array of perturbations in lipid metabolism, including increased free fatty acid flux into the liver.52 Patients demonstrate mild elevations in serum aminotransferase levels but generally do not develop cirrhosis or liver failure. Hepatic adenomas develop in 22% to 75% of patients, as early as three years of age but most commonly in the second decade of life, and tend to increase in both size and number as the patient ages (see Chapter 94).54 Adenomas in rare instances can transform to hepatocellular carcinoma; unfortunately, serum α-fetoprotein and carcinoembryonic antigen levels and features of the lesions on hepatic imaging are not predictive of malignant transformation.54-56 In the past, reports suggested that the risk of hepatic

adenoma was increased in patients who were nonadherent with treatment of GSD type I, but more recent reports have found no differences in the risk of adenoma formation between adherent and nonadherent patients.57 In some patients, hepatic adenomas have been demonstrated to regress and disappear after adequate nutritional therapy, but in general, the course is unpredictable, especially in nonadherent patients.54,55 Because of the unreliability of radiographic imaging and serum marker levels in predicting malignant transformation in this patient population, the better approach—resection of an adenoma or liver transplantation—is uncertain.58

Diagnosis

The hepatic glycogen content is elevated in patients with GSD type I, and the most accurate diagnostic measure is direct analysis of enzyme activity performed on fresh, rather than frozen, liver tissue. Analysis of fresh liver tissue is important to avoid disrupting microsomal Glu-6-Pase activity.59 Fasting serum glucose and lactate levels, a positive glucagon response test result, and the response to fructose or galactose administration (patients with GSD type I do not show the expected rise in serum glucose concentration after administration of glucagon or either sugar) often provide supportive evidence but may not yield a definitive diagnosis. Intermittent severe neutropenia is noted in most patients with GSD type Ib.47,53 DNA analysis-based approaches that integrate biochemical features and the presence or absence of persistent neutropenia have been proposed and may provide a diagnostic alternative to liver biopsy.60

Treatment

Patients with undiagnosed or undertreated GSD type I are at increased risk of death, usually from hypoglycemic comas, seizures, metabolic acidosis, or, in those with GSD type Ib, sepsis from neutropenia.49 Rarely, hepatocellular carcinoma is a cause of death. Management centers on preventing the acute metabolic derangements and potential long-term complications and enabling the patient to attain normal psychological development and a good quality of life.59,61 Consensus guidelines for the management of GSD type I have been proposed.59,61 Biomedical targets for good metabolic control include a preprandial blood glucose level higher than 63 to 77 mg/dL (3.5 to 4.3 mmol/L), urine lactate-to-creatinine ratio higher than 0.06, high-normal serum uric acid level, venous blood base excess higher than −5 mmol/L, bicarbonate level higher than 20 mmol/L, serum triglyceride level lower than 6 mmol/L, and body mass index (BMI) between 0 and 2 standard deviations from normal. In addition, for GSD type Ib, demonstrating a normal fecal α1-AT level is desirable (see Chapter 28).59,62 Because optimal glycemic control is not always possible and the risk of severe hypoglycemia is high if delivery of glucose is interrupted inadvertently, serum lactate levels should be kept at the high end of normal because lactate is an alternative fuel for the brain. Nutritional supplementation has become the mainstay of therapy for GSD type I. Frequent, high-carbohydrate, daytime feedings, such as uncooked cornstarch, or con­ tinuous nighttime drip feedings, or both, allow the steady release of glucose and lead to improved metabolic control and normalized growth and development.52,62 A bioche­ mical target is to maintain the serum glucose level above 70 mg/dL (3.9 mmol/L). Uncooked cornstarch in a dose of 2 g/kg every six hours (6 to 8 mg of glucose/kg/minute) has been suggested; however, alternative regimens have been implemented successfully.

1263

1264

Section IX  Liver For infants, when the diagnosis of GSD type 1 is confirmed, a formula that does not contain fructose or galactose should be prescribed. Frequent daytime feedings and continuous nocturnal administration are required, with the rate of delivery needed to maintain euglycemia being ≈8 mg/kg/ minute. Morning feedings should be given quickly after discontinuation of the nighttime drip to avoid hypogly­ cemia. As solids are introduced, high-carbohydrate foods should be emphasized. These patients require special attention during acute illnesses that may affect oral intake or metabolism because they can become hypoglycemic quickly. Prophylaxis with antibiotics (e.g., trimethoprimsulfamethoxazole) is recommended for patients with GSD type 1b and severe neutropenia or recurrent bacterial infections.61 Granulocyte colony stimulating factor (GCSF) has been used with success in patients with GSD type 1b to improve hematologic parameters and neutrophil function and reduce the morbidity associated with severe bacterial infections.63 Splenomegaly may worsen with GCSF therapy, and bone marrow aspiration before and during GCSF therapy may be prudent, given rare occurrences of acute myelogenous leukemia (AML) in patients with GSD type 1b.61 Both GCSF and inflammatory bowel disease raise the risk of osteopenia, and monitoring of bone density is advised. Adenoviral-mediated gene replacement therapy of recom­ binant Glu-6-Pase in a canine model of GSD type Ia deficiency, which has all of the major features of GSD type 1a in humans, has led to encouraging results and may be an option in humans in the future.64 An alternative approach, human hepatocyte transplantation, has been performed in a single patient with GSD type 1a with nearresolution of hypoglycemic episodes; however, hypoglyce­ mia subsequently recurred, likely because of the lack of ongoing immunosuppression to reduce the likelihood of recipient rejection of transplanted donor hepatocytes.65 Hepatocyte transplantation, with the use of standard posttransplantation immunosuppression, has been performed successfully in an 18-year-old male adolescent with GSD type 1b, with euglycemia maintained up to 2 years post-transplantation.66 Liver transplantation has corrected the metabolic error in patients with GSD type I and permitted catch-up growth, even in patients in the third decade of life.67,68 Neutrophil counts and function, however, are only variably improved after liver transplantation in patients with GSD type 1b.67

TYPE III

GSD type III results from deficiency of glycogendebranching enzyme (GDE) and leads to the accumulation of limit dextrin units, which restrict subsequent glucose release by phosphorylase. Because deficiency of GDE does not interfere with metabolism of Glu-6-P, patients with GSD type III still have effective mechanisms for gluconeogenesis. Therefore, affected patients have a milder clinical course than those with GSD type I and are able to fast for longer periods. In infancy, however, GSD type III may be indistinguishable from GSD type I. GDE is encoded by a single gene and possesses two independent catalytic activities, an amylo-1,6-glucosidase and oligo-1,4→1,4 glucan transferase. Both of these activities are deficient in the two main clinical subtypes of GSD type III, types IIIa and IIIb. Differential expression of four major GDE mRNA isoforms in liver and muscle tissue distinguishes the two types: type IIIa affects liver and muscle and accounts for 80% of patients, and type IIIb affects the liver only and accounts for 15% of patients. Although subtype-specific

mutations in the GDE gene are increasingly identified, the molecular basis for this differential tissue-specific expression of GDE is unknown, and no clear genotype-phenotype correlations have been identified.44,69 Rare isolated loss of one of the two GDE activities has been observed (i.e., glucosidase activity in type IIIc and transferase activity in type IIId).

Clinical Features

Persons with GSD type III typically exhibit hypoglycemia, hepatomegaly, and growth failure. Liver enlargement results from increased glycogen deposition and not fatty infiltration. The liver may show fibrotic septa that rarely lead to frank cirrhosis and liver failure. Serum lactate and uric acid levels are normal, and aminotransferase levels are increased only moderately until advanced liver disease occurs. Hyperlipidemia may be present but is not as pronounced as in GSD type I. Patients have normal responses to fructose and galactose loading. Patients with GSD type III may also display progressive muscle weakness, which worsens with activity, and muscle wasting. Nephromegaly is not seen, but cardiomegaly may be present. The diagnosis can be made by direct enzyme analysis of muscle or liver tissue or peripheral leukocytes; mutation analysis of the GDE gene will be increasingly important for diagnosis in the future.45,70 Hepatic adenomas develop in approximately 25% of patients, and isolated reports of cirrhosis leading to hepatocellular carcinoma have been reported.55,71

Treatment

A high-protein, low-carbohydrate diet has been suggested to normalize metabolic activity, ensure normal growth, normalize muscle function, and minimize hepatomegaly. This diet provides adequate substrates for gluconeogenesis while reducing the need for glycogen storage. Patients with refractory hypoglycemia or persistent hepatomegaly may require a nighttime continuous infusion or cornstarch therapy, as used for GSD type I.

TYPE IV

Deficiency of the branching enzyme is seen in GSD type IV, a rare syndrome also known as amylopectinosis. Glycogen and amylopectin accumulate in hepatocytes, leading to hepatomegaly, abdominal distention, and failure to thrive, most commonly during infancy. Signs of liver disease, when present, predominate later in the course of the disease. Several variable forms of GSD type IV have been observed—a severe congenital form that manifests as fetal hydrops, neonatal hypotonia, or fetal death72; a childhood subtype that manifests as cardiomyopathy and abnormal neuromuscular development; and other milder, nonprogressive presentations of hepatic disease that do not lead to cirrhosis and are not associated with skeletal muscle or neurologic involvement.73 Genotype-phenotype analyses of the branching enzyme gene have revealed a high degree of molecular heterogeneity without clear clinical associations.73 Hypoglycemia is relatively uncommon, and responses to fructose and galactose challenges are normal. Serum lactate and pyruvate levels are normal, and aminotransferase levels are only moderately elevated until more severe liver involvement becomes apparent. Progressive macronodular cirrhosis is present with an abundance of PAS-positive deposits (amylopectin) in hepatocytes (Fig. 76-2). Cirrhosis may progress to liver failure, and adenomas and hepatocellular carcinoma may develop rarely.74 The diagnosis of GSD type IV can be made by direct enzyme analysis of liver tissue or fibroblasts. Most patients

Chapter 76  Other Inherited Metabolic Disorders of the Liver

B

A Figure 76-2.  Photomicrograph of a liver biopsy specimen from a patient with glycogen storage disease type IV. A, At low power, loss of the normal trabecular hepatic architecture is seen. B, Higher power shows accumulation of “ground-glass” cytoplasmic inclusions within the hepatocytes; this finding is a consequence of abnormal intracellular processing of amylopectin-like material. (A, Hematoxylin and eosin, ×20; B, ×100.)

die within the first three years of life if the disease is untreated. Diets high in protein and low in carbohydrate have been associated with improved growth but have had little effect on liver involvement. Liver transplantation has been used successfully and results in correction of the metabolic error and normal growth; however, persistence of amylopectin deposits in the heart (with progressive cardiomyopathy) and leukocytes of affected patients has been described.75

CONGENITAL DISORDERS OF GLYCOSYLATION Congenital disorders of glycosylation (CDGs) comprise a group of inherited defects in the enzymes that synthesize the glycan moiety of glycoproteins or the macromolecules that affect intracellular trafficking and functioning of glycoproteins.76-78 More than 20 CDGs involving both asparagine (N)- and serine/threonine (O)-linked protein glycosylation have been reported to date, and many of these disorders lead to dysfunction of the liver, intestine, or both.76-78 Protein glycosylation is complex and involves multiple enzymatic steps and subcellular compartments.77,78 Secretory glycoproteins with altered carbohydrate moieties in CDGs include coagulation factors, albumin and other binding proteins, growth hormone, apolipoproteins, insulin, and thyroxinebinding globulin.79 Because protein glycosylation occurs in all cells, it is not surprising that patients with a CDG exhibit multisystemic abnormalities, often dominated by central nervous system manifestations. Two main groups of protein N-glycosylation disorders, groups I and II, have been delineated on the basis of characteristic isoelectric focusing patterns of serum transferrin, a marker protein for this group of disorders.77 Group I disorders (of which there are 12 types) involve aberrant processing of lipid-linked oligosaccharides before transfer to protein targeted for glycosylation and include the three most common and best-characterized types of CDG, types Ia, Ib, and Ic. Clinical features in common among these three disorders are protein-losing enteropathy, coagulopathy (both procoagulant and anticoagulant states), feeding difficulties, and hepatomegaly.77,80

Glucose

Glucose-6-Phosphate

Fructose-6-Phosphate Phosphomannose Isomerase Mannose-6-Phosphate Mannose Hexokinase Phosphomannomutase Mannose-1-Phosphate GDP-Mannose Asparagine-linked glycosylation Figure 76-3.  Pathway of mannose metabolism. Enzymes are shown in italics. GDP, guanosine diphosphate.

CDG type Ia is caused by defects in phosphomannomutase (PMM), an enzyme that converts mannose-6-phosphate to mannose-1-phosphate (Fig. 76-3). Almost 60 mutations have been found in the encoding PMM2 gene; most patients are compound heterozygotes for mutations that likely preserve some residual PMM enzymatic activity, suggesting that complete loss of PMM activity is incompatible with life; indeed, targeted disruption of the PMM2 gene in the mouse leads to early embryonic lethality.81,82 Patients typically have severe neurologic abnormalities, dysmorphisms (inverted nipples, abnormal fat distribution, and esotropia), and congenital hepatic fibrosis and steatosis in addition to the features described earlier.80,83 Patients with CDG type Ib have a defect in phosphomannose isomerase (PMI), which converts fructose-6-phosphate

1265

1266

Section IX  Liver to mannose-6-phosphate. In addition to intractable diarrhea, protein-losing enteropathy, and congenital hepatic fibrosis, recurrent episodes of hyperinsulinemic hypoglycemia and cyclic vomiting can occur in patients with CDG type Ib. Neurologic symptoms are usually absent and dysmorphisms are less common than in CDG type 1a.76,80,84 Many patients with CDG type Ib have been treated effectively with dietary mannose, making CDG type Ib the only treatable form of CDG to date, although liver fibrosis can still develop despite improvement in clinical symptoms.78,85,86 Transient hepatomegaly, without congenital hepatic fibrosis, has been noted in a patient with CDG type Ic; otherwise, the clinical features of CDG type Ic are similar to but milder than those of CDG type Ia.80 Patients with CDG type Ih also have protein-losing enteropathy with chronic diarrhea, hepatic fibrosis, and variable degrees of hepatic dysfunction.87 Children with untyped cases of CDG, labeled CDG-X until the genes and protein products can be identified, have been found and have isolated cryptogenic chronic liver disease, mild coagulopathy, and mild portal fibrosis and focal steatosis on liver biopsy specimens.88 In some cases, numerous myelinosomes have been identified in hepatocytes by electron microscopy. Group II CDGs (six types) involve defects that affect the processing of N-linked glycoproteins.77,78 Most types result from defects in enzymes involved in the trimming of protein-bound oligosaccharides and the subsequent addition of terminal sugars. Patients have marked dysmorphic features and severe developmental retardation. Two infants with hepatosplenomegaly, progressive jaundice, severe epilepsy, recurrent infections, and heart failure have been shown to have mutations in a subunit of the conserved oligomeric Golgi (COG) complex that result in disruption of glycosylated protein intracellular trafficking.89 Another CDG type II disorder with liver involvement has been reported in an infant with recurrent infections, chronic diarrhea, progressive cirrhosis, and normal hepatic excretory and synthetic function.89 The biochemical and clinical features of the remaining types of CDG are less well characterized. Isoelectrofocusing of apolipoprotein C-III, which carries a single O-linked glycan moiety, has been proposed as a screening method for the O-glycan biosynthesis defects (six types) encountered thus far.76,90 Hepatic dysfunction is usually mild in CDG and usually does not lead to symptoms. Mild hepatic steatosis and fibrosis typically are seen on light microscopy; on electron microscopy, lysosomal vacuoles, termed myelosomes, with concentric electrondense membranes and variable electron-lucent and electron-dense material, are noted.91 Patients uncommonly can progress to liver failure, with micronodular cirrhosis noted at autopsy. Any patient with unexplained congenital hepatic fibrosis, protein-losing enteropathy, or a procoagulant tendency should be evaluated for the possibility of CDG (especially type Ib). An initial screening of serum transferrin with isoelectric focusing should be performed, followed by confirmatory enzymatic analysis in fibroblasts, leukocytes, or liver tissue. If the diagnosis of CDG type Ib is confirmed, oral mannose therapy should be initiated.86

PORPHYRIAS The porphyrias are a diverse group of metabolic diseases that result from deficiencies in enzymes involved in the heme synthetic pathway. In this section, this synthetic pathway and those porphyrias for which the primary site of

expression is the liver or in which direct hepatotoxicity occurs are reviewed.

Pathophysiology

The metabolic pathways of heme synthesis are essentially the same in the two tissues in which heme synthesis primarily occurs, the liver (15% to 20%) and the bone marrow (75% to 80%), although synthetic control may be different in the two tissues. The rate-limiting step in hepatic heme synthesis begins with the conversion of glycine and succinyl coenzyme A (CoA) to 5-aminolevulinic acid (ALA) by the action of ALA synthase (Fig. 76-4). ALA synthase activity is decreased by the end-product of the pathway, heme, and is increased by substances that induce the hepatic cytochrome P450 pathway. Six additional enzymatic steps convert ALA to protoporphyrin IX (see Fig. 76-4). In the final step of the pathway, protoporphyrin IX is coupled to ferrous iron by ferrochelatase to create heme. Enzyme deficiencies arising from any of these eight steps of the heme synthetic pathway lead to the clinically apparent diseases known as the porphyrias. Porphyrias are commonly classified according to clinical features into two main groups, acute porphyrias, which are characterized by dramatic and potentially life-threatening Glycine + Succinyl CoA ALA synthase 5-aminolevulinic acid ALA dehydratase

ADD

Porphobilinogen PBG deaminase

AIP

Hydroxymethylbilane Uroporphyrinogen III cosynthase

CEP

Uroporphyrinogen III Uroporphyrinogen III decarboxylase

PCT; HEP

Coproporphyrinogen III Coproporphyrinogen oxidase

HCP

Protoporphyrinogen IX Protoporphyrinogen oxidase

VP

Protoporphyrin IX Ferrochelatase

EPP

Heme Figure 76-4.  Pathway of heme synthesis. The location of enzymatic deficiency in the various forms of porphyria is noted. On the left, enzymes are shown in italics. On the right, abbreviations for cutaneous porphyrias are shown in italics, and those for acute porphyrias are shown in roman typeface. ADD, 5-aminolevulinic acid (ALA) dehydratase deficiency; AIP, acute intermittent porphyria; CEP, congenital erythropoietic porphyria; EPP, erythropoietic protoporphyria; HCP, hereditary coproporphyria; HEP, hepatoerythropoietic porphyria; PBG, porphobilinogen; PCT, porphyria cutanea tarda; VP, variegate porphyria.

Chapter 76  Other Inherited Metabolic Disorders of the Liver Table 76-3  The Porphyrias ENZYMATIC DEFECT

MODE OF INHERITANCE

CLINICAL FINDINGS

SITE OF EXPRESSION

MAJOR BIOCHEMICAL FINDINGS

Acute intermittent porphyria ALA dehydratase deficiency Hereditary coproporphyria

PBG deaminase ALA dehydratase Coproporphyrinogen oxidase

Autosomal dominant Autosomal recessive Autosomal dominant

Neurologic Neurologic Neurologic, cutaneous

Liver Liver Liver

Variegate porphyria

Protoporphyrinogen oxidase

Autosomal dominant

Neurologic, cutaneous

Liver

Urine: ALA < PBG Urine: ALA Urine: ALA > PBG, coproporphyrin Stool: Coproporphyrin Urine: ALA > PBG, coproporphyrin Stool: Coproporphyrin, protoporphyrinogen

Congenital erythropoietic porphyria Erythropoietic protoporphyria

Uroporphyrinogen III cosynthase Ferrochelatase

Autosomal recessive

Cutaneous

Bone marrow

Autosomal dominant

Cutaneous, rarely neurologic

Liver, bone marrow

Hepatoerythropoietic porphyria

Uroporphyrinogen III decarboxylase

Autosomal recessive

Cutaneous

Liver, bone marrow

Porphyria cutanea tarda

Uroporphyrinogen III decarboxylase

Autosomal dominant or acquired

Cutaneous

Liver

ACUTE PORPHYRIAS

Cutaneous Porphyrias Urine and stool: Coproporphyrin I Urine: None Stool: Protoporphyrin, coproporphyrin Urine: Uroporphyrin, 7-carboxylate porphyrin Stool: Isocoproporphyrin Urine: Uroporphyrin, 7-carboxylate porphyrin Stool: Isocoproporphyrin

ALA, 5-aminolevulinic acid; PBG, porphobilinogen.

neurologic symptoms, and cutaneous porphyrias, which typically cause few or no neurologic symptoms but instead give rise to a variety of severe skin lesions (Table 76-3). In five of the porphyrias, the liver is the major site of expression; in two others, both the liver and bone marrow are involved; and in one, only the bone marrow contributes.

Acute Porphyrias

The symptoms and signs of the acute neurovisceral attacks that occur in the four acute porphyrias vary considerably. Abdominal pain is present in more than 90% of patients, followed in frequency by tachycardia and dark urine in about 80% of patients. Neuropsychiatric features include hysteria, depression, psychosis, confusion, hallucinations, seizures, and coma, although little evidence suggests that chronic psychiatric illness occurs. Other features are constipation, extremity pain, paresthesias, nausea, vomiting, urinary retention, hypertension, peripheral sensory deficits (often in a “bathing trunk” distribution), and weakness leading to ascending paralysis or quadriplegia. These neurologic attacks appear to be related to the overproduction of ALA and porphobilinogen (PBG), which leads to higher serum and tissue levels of these neurotoxic products.92 Acute episodes are about five times as common in women as in men and may be precipitated by many factors, most commonly drugs, alcohol ingestion, and smoking. Steroids, sex hormones, and medications that stimulate the hepatic cytochrome P450 system, perhaps by increasing the requirements for heme production, are often identified as precipitants.93 Other inciting factors are fasting, infections, and pregnancy; some women report greater problems during the luteal phase of their menstrual cycles.92,94 The disease is clinically latent in 65% to 80% of patients. ALA dehydratase deficiency is a rare syndrome with autosomal recessive transmission in which the enzyme activity is less than 3%. The enzyme activity is 50% of normal in carriers, who are asymptomatic. Affected patients have severe, recurrent neurologic attacks that may be life threatening. They excrete large amounts of ALA in their urine.

Liver transplantation was reported to result in complete resolution of symptoms in one patient with ALA dehydratase deficiency.95 The three remaining acute porphyrias—acute intermittent porphyria (AIP), hereditary coproporphyria (HCP), and variegate porphyria (VP)—result from partial deficiency of the enzymes PBG deaminase, coproporphyrinogen oxidase, and protoporphyrinogen oxidase, respectively. All three disorders are inherited in an autosomal dominant fashion with variable expression. AIP is the most common of the three conditions, occurring in 5 to 10 per 100,000 people, and manifests primarily as derangements in the autonomic nervous system or as a psychiatric disorder.96 VP is more common in South Africa than elsewhere. Although HCP and VP give rise to neurologic symptoms similar to those of AIP, cutaneous lesions also occur in HCP and predominate in VP.97

Cutaneous Porphyrias

The cutaneous porphyrias differ from the acute porphyrias in that affected patients exhibit few or no neurologic symptoms. In these illnesses, excess porphyrins or porphyrinogens are deposited in the upper dermal capillary walls, where these photoreactive compounds cause tissue damage that manifests as cutaneous vesicles and bullae in areas exposed to light or excessive mechanical manipulation. Scarring, infection, pigment changes, and hypertrichosis can follow and even lead to severe mutilation.98 Porphyria cutanea tarda (PCT), the most common of the porphyrias, typically involves a 50% reduction in activity of the enzyme uroporphyrinogen decarboxylase. Patients usually present after 20 years of age. Two types of PCT are recognized. Type I PCT affects 80% of patients and is a sporadic (acquired) form with the enzyme deficiency restricted to the liver. Type II, which affects the other 20% of patients, is familial and inherited in an autosomal dominant fashion with incomplete penetrance; the enzyme deficiency occurs in all tissues.99 Symptoms develop in fewer than 10% of patients with type II PCT. Type I PCT is associ-

1267

1268

Section IX  Liver ated strongly with high alcohol intake, estrogen therapy, and systemic illnesses, including systemic lupus erythematosus, diabetes mellitus, chronic kidney disease, and the acquired immunodeficiency syndrome. For unclear reasons, concomitant hepatitis C infection is strongly associated with expression of PCT (see Chapter 79).100,101 The frequency of mutations of the HFE gene, which causes hereditary hemochromatosis, is increased in patients with types I and II PCT, and these mutations are thus susceptibility factors for clinical expression of the PCT phenotype (see Chapter 74).100,102,103 This association is consistent with pathologic findings in liver biopsy specimens from patients with PCT, of whom 80% have siderosis, 15% have cirrhosis, and most have evidence of iron overload. A patient in whom hepatocellular carcinoma developed has been reported to have unsuspected underlying PCT and hereditary hemochromatosis with bridging fibrosis.104,105 Patients usually do not show signs of overt clinical liver disease, apart from elevated serum aminotransferase levels. Hepatoerythropoietic porphyria (HEP) is a rare form of porphyria with a pathogenesis similar to that of PCT. HEP results from homozygous uroporphyrinogen decarboxylase deficiency, yielding less than 10% of normal enzyme activity. The cutaneous lesions, which resemble those of PCT, are typically severe and mutilating. The disease usually manifests in the first year of life, and as the patient ages, the dermatologic manifestations may subside, but liver disease, characterized by a nonspecific hepatitis, worsens. Mutation analysis of the uroporphyrinogen decarboxylase gene has revealed numerous mutations that are usually unique to an individual family; no clear genotype-phenotype correlations have been observed.99 Congenital erythropoietic porphyria (CEP) is a rare form of porphyria with autosomal recessive transmission that is caused by deficiency of uroporphyrinogen III cosynthase, which mainly affects erythropoietic tissue. Affected patients typically present in the first year of life with blisters and disfiguring skin lesions in exposed areas. Infants may present with pink urine and photosensitivity. As patients age, erythrodontia, a pathognomonic red or brownish discoloration of the teeth, is commonly seen. CEP can be distinguished clinically from HEP by the presence in some cases of a Coombs-negative hemolytic anemia, which can be quite severe. Splenomegaly is common. Erythropoietic protoporphyria (EPP) is caused by partial deficiency of the enzyme ferrochelatase (FECH), the final step in the heme synthetic pathway. EPP is the second most common type of porphyria and is inherited in an autosomal dominant manner with variable penetrance. Patients with EPP and asymptomatic carriers exhibit an FECH enzyme activity of 30% to 40% and 50%, respectively, even though both groups inherit a defective FECH allele. The mechanism for the variable clinical expression is explained by coinheritance of a “low-expressing” normal FECH allele in symptomatic patients with EPP and a normal FECH allele in asymptomatic patients.106,107 Although the bone marrow is the predominant source of excess protoporphyrin, with a variable contribution from the liver and other tissues, the skin is the primary site of deposition of this phototoxic compound in patients with EPP. Therefore, the principal clinical manifestation is exquisite photosensitivity, which may present during infancy and can lead to a wide spectrum of symptoms (e.g., itching, burning, or pain) and to scars and lichenification of the skin. Vesicles are rare. Patients with EPP may show a mild hypochromic, microcytic anemia. Clinical liver disease, which develops in 5% to 10% of patients with EPP, results from progressive hepatic accumu-

lation of protoporphyrin.108 Liver disease typically occurs after age 30 but has been described in children. The liver appears black and nodular, with hepatocellular necrosis, portal inflammation, cholestasis, and extensive deposits of dark brown pigment in hepatocytes, Kupffer cells, and biliary structures; birefringence of pigment deposits is seen on polarization microscopy.109 Of 57 patients with EPP followed for more than 20 years in one study,110 50% had normal serum aminotransferase levels and liver histologic findings. Of the remaining patients, cirrhosis occurred in seven and liver failure developed in two. Liver disease is not believed to be secondary to alcohol consumption, viral infections, or external toxins, although these insults can worsen liver function.111 Genetic heterogeneity in the FECH gene has been noted in multiple studies and is seen in patients who need liver transplantation.112

Diagnosis

The approach to the diagnosis of the porphyrias is summarized in Table 76-3. Clinical features alone are usually not specific enough to confirm a diagnosis or distinguish among the various forms of porphyrias, and biochemical test results must be interpreted correctly for accurate diagnosis and management. The diagnosis of porphyria should be considered in patients with recurrent bouts of severe abdominal pain, dark urine, constipation, and neuropsychiatric disturbances or in patients with typical dermatologic findings. To differentiate among the different porphyrias, urine and stool samples should be obtained for porphyrin studies and a urine specimen collected for quantitative ALA and PBG determinations. In AIP, excretion of PGB and ALA (PGB more than ALA) in dark urine is common during porphyric attacks, but the levels may be normal during asymptomatic periods and in prepubertal patients.113 Patients with HCP and VP excrete high levels of ALA and PBG in the urine; in contrast to those with AIP, these patients excrete more ALA than PGB. A rapid “spot” urine test to detect urinary PGB is recommended to diagnose the acute porphyrias, except for the rare patient with ALA dehydratase deficiency.92 Fecal coproporphyrins are increased in both HCP and VP, whereas only in VP is the amount of fecal protoporphyrin also increased. Important advances have been made in the identification of a large number of gene mutations in several of the acute porphyrias, including AIP, VP, and HCP. Given the high degree of genetic heterogeneity, the lack of clear genotype-phenotype correlations, and the failure to find mutations in 5% to 10% of families with current techniques, however, genetic testing is not recommended as a general screening tool. If a mutation that causes porphyria has been identified in a particular subject, however, screening of asymptomatic family members has a sensitivity and specificity of nearly 100% for that family and may be helpful, together with appropriate genetic counseling. Enzyme activity measurement can likewise help to confirm the diagnosis, although to a lesser degree than genetic testing given the variable tissue expression of many of the heme synthetic enzymes.

Hepatic Involvement

Hepatic involvement in porphyria is variable; in general, patients with acute porphyria may have elevations of serum aminotransferase and bile acid levels, with further increases during acute episodes. Liver biopsy specimens may show steatosis and iron deposition. Although these changes are minor, patients with acute porphyria are at increased risk for development of hepatocellular carcinoma.114

Chapter 76  Other Inherited Metabolic Disorders of the Liver PCT and HEP are more commonly associated with hepatic complications, including liver enlargement with fatty infiltration, inflammation, and granulomatous changes. Siderosis and fibrosis may lead to cirrhosis and liver failure. The risk of hepatocellular carcinoma is increased only slightly in patients with these disorders.105 The patterns of liver injury in CEP are similar to those in PCT and HEP.115

Treatment

The overall survival of patients with acute porphyria is good. Treatment is based on avoidance of drugs and other precipitating factors.93 Generous fluid and glucose administration is recommended during acute attacks and can elicit the “glucose effect” that diminishes ALA synthase activity. Intravenous administration of hematin, a congener of heme, has been shown to decrease the drive for heme synthesis and its excessive by-products. Hematin can also have a dramatic effect on neurologic symptoms, especially if given early in an attack, with clinical improvement often occurring within one to two days.92 Women in whom symptoms are affected by the phases of their menstrual cycle can show improvement while taking oral contraceptives. Liver transplantation has been attempted for several of the porphyrias, with mixed results.92,116,117 Because of the increased frequency of hepatocellular carcinoma, patients with AIP should undergo standard surveillance for this tumor (see Chapter 94).92 Because of the wavelengths of light absorbed by the porphyrins, patients affected by porphyria are at risk from exposure to sunlight as well as to household and fluorescent lights. Patients must use special sunscreen lotions that block rays in the 400- to 410-nm range. Skin trauma should be minimized as much as possible; early treatment of skin infections can decrease scarring. Special screens may be especially useful for protection against indoor lighting. Some patients have incurred severe or lethal internal burns during surgery, including liver transplantation; thus, appropriate precautions must be taken.98 Treatment of PCT initially consists of removal of any offending agent. Historically, treatment has included phlebotomy to decrease iron overload and hepatic siderosis. This approach may provide relief of cutaneous symptoms in four to six months. Chloroquine complexes with uroporphyrin and facilitates its excretion, but caution must be used during therapy with chloroquine because the drug is potentially hepatotoxic.98 The efficacy of chloroquine has been variable in patients with PCT who are homozygous for mutations in the HFE gene; for these patients, phlebotomy should be first-line therapy.118,119 Treatment strategies for HEP are similar to, but have not been as successful as, those for PCT. Blood transfusions and administration of hematin, charcoal, and cholestyramine all have led to clinical improvement in patients with EPP, but long-term resolution has not been demonstrated. One report of a 63-year-old man with severe cholestatic liver disease caused by EPP showed complete biochemical and histologic resolution of cholestasis with the use of blood transfusions, hematin, cholestyramine, and ursodeoxycholic acid. Subsequently, the patient underwent bone marrow transplantation, with apparently complete resolution of EPP.120 Liver transplantation has been accomplished in patients with EPP and ESLD, with mixed results; the erythropoietic defect persists.117,121 A retrospective review of all 20 patients with EPP who have undergone liver transplantation in the United States revealed unique perioperative complications, including light-induced tissue damage in four patients and neuropathy in six, as well as recurrent EPP-associated liver disease

in 65% of patients who survived more than two months. Overall patient and graft survival rates were statistically similar to those for all other patients transplanted in the United States during the same period.117 Therefore, liver transplantation must be considered symptomatic therapy, except in patients with fulminant hepatic failure, given the high risk of recurrent disease in the graft and the added risk of intraoperative photodynamic injury to internal organs.117,122 Bone marrow transplantation after liver transplantation has been performed successfully in a child with EPP-induced ESLD.123 Splenectomy, which lengthens the lifespan of circulating red blood cells and decreases the erythropoietic drive, has been shown to be effective in many patients with CEP. Frequent blood transfusions or hematin infusions inhibit the stimulus for heme production, thereby diminishing or eliminating the cutaneous manifestations of the disease.124 Bone marrow transplantation has proved curative in severely affected patients.124,125

TYROSINEMIA Four known human diseases are caused by enzymatic deficiencies in the catabolic pathway for the amino acid tyrosine: alkaptonuria and hereditary tyrosinemia types I, II, and III. Although all of the enzymes involved in this pathway are found in the liver, only hereditary tyrosinemia type I (HT1) leads to progressive liver dysfunction. Formerly known as hepatorenal tyrosinemia, HT1 also affects other organ systems, in particular the kidneys and periph­ eral nerves. Advances in our understanding of the pathophysiology of the disease process and new treatment options, such as an inhibitor of an early step in the degradation pathway, have improved the clinical course dramatically for affected persons. A disease with autosomal recessive transmission, HT1 has a worldwide incidence of about 1 in 100,000. The incidence is much greater in northern Europe (1 per 8000) and in the Saguenay-Lac-St. Jean region of Quebec, Canada (1 per 1846), where a founder effect has been documented.126

Pathophysiology

The pathway for tyrosine metabolism is shown in Figure 76-5. The enzymatic defect in patients with tyrosinemia has been identified in fumarylacetate hydrolase (FAH), the final step in the degradation process. More than 30 mutations in FAH have been found in patients with HT1, but no clear correlation between FAH genotype and HT1 phenotype has been found.127 FAH deficiency leads to accumulation of the upstream metabolites fumarylacetoacetate (FAA) and ma­leylacetoacetate, which are then converted to the toxic inter­mediates succinylacetoacetone (SAA) and succinylacetone (SA). FAA has been shown to deplete blood and liver of glutathione, the consequence of which may be augmentation of the mutagenic potential of FAA.127,128 SA inhibits renal glucose and amino acid transport and the degradation of ALA to PBG, probably via direct modification of amino acids in enzyme active sites. It also inhibits DNA ligase activity in fibroblasts isolated from patients with HT1.127,129 Over time, the combined effects of high levels of FAA and SA on the integrity of DNA and cellular repair mechanisms may account for increased chromosomal breakage in fibroblasts isolated from patients with HT1, as well as an increased risk of hepatocellular carcinoma.127,130

1269

1270

Section IX  Liver Phenylalanine Phenylalanine hydroxylase Tyrosine Tyrosine aminotransferase p-Hydroxyphenylpyruvate p-Hydroxyphenylpyruvate dioxygenase

NTBC

Homogentisate Homogentisate oxidase Maleylacetoacetate Maleylacetoacetate isomerase Fumarylacetoacetate Fumarylacetoacetate hydrolase

Succinylacetoacetate + Succinylacetone

HT1

Fumarate + Acetoacetate Figure 76-5.  Pathway of tyrosine metabolism. The location of the enzymatic defect in hereditary tyrosinemia type 1 (HT1) and the site of action of 2-(2-nitro-4-trifluoro-methylbenzoyl)-1,3-cyclohexanedione (NTBC) are shown. Enzymes are shown in italics.

Clinical and Pathologic Features

Patients with HTI present either acutely with liver failure or with chronic liver disease, with or without hepatocellular carcinoma. In the acute form of HTI, patients present with liver disease in the first six months of life; symptoms include those associated with hepatic synthetic dysfunction, such as hypoglycemia, ascites, jaundice, and bleeding diathesis, as well as anorexia, vomiting, and irritability.131 Laboratory studies show elevations of serum aminotransferase, GGTP, and bilirubin levels and decreased levels of coagulation factors. Serum tyrosine, methionine, and α-fetoprotein levels are elevated substantially. Analysis of the urine reveals phosphaturia, glucosuria, hyperaminoaciduria, renal acidosis, and elevated excretion of ALA and phenolic acids.127 The acute form is usually fatal within the first two years of life. In a multicenter study, van Spronsen and associates131 showed that 77% of patients with tyrosinemia presented before the age of six months. The one- and twoyear survival rates were 38% and 29%, respectively, if patients presented between zero and two months, and 74% and 74%, respectively, if they presented between two and six months. Survival for both time intervals rose to 96% if the first symptoms appeared after age six months. The cause of death was usually recurrent bleeding and liver failure (35 of 47 deaths); however, hepatocellular carcinoma (7 of 47) and neurologic crisis (3 of 47) accounted for some deaths. Patients with the chronic form of HTI classically have symptoms that are similar to but milder than those of the acute presentation; sometimes, serum aminotransferase levels as well as plasma tyrosine and methionine levels can be within the normal range.127 These patients usually present after one year of age with hepatomegaly, rickets, nephromegaly, hypertension, and growth retardation. They also are likely to have neurologic problems and hepatocellular carcinoma.

The pathologic changes differ in the acute and chronic forms of the disease. In the acute form, the liver may appear enlarged with a pale nodular pattern or may be shrunken, firm, and brown. Micronodular cirrhosis, fibrotic septa, bile duct proliferation and plugging, steatosis, pseudoacinar and nodular formations, and giant cell transformation may be found on histologic examination. Varying amounts of FAH enzyme activity have been found in liver tissue from patients with HTI as a result of spontaneous reversion of FAH gene mutations. Patients with the chronic form of the disease have a higher level of reversion and a lower frequency of liver dysplasia.132 In an analysis of mutations in the FAH gene in members of 13 unrelated families with HT1, no mutation type predominated in the affected families, and no correlation between genotype and phenotype was observed.133 In the chronic form of tyrosinemia, the liver appears enlarged, coarse, and nodular. In histologic specimens, micronodular and macronodular cirrhosis may be present, as may steatosis, fibrotic septa, and a mild lymphoplasmacytic infiltrate. Cholestasis is less pronounced than in the acute form of HT1. Large- or small-cell dysplasia may be present, reflecting premalignant changes. Because of the nodular changes, identification of progression to hepato­ cellular carcinoma can be difficult. Because the serum α-fetoprotein value is elevated before hepatocellular carcinoma develops, measurement of α-fetoprotein is not helpful in the diagnosis of hepatocellular carcinoma in this setting. Visualization of both low- and high-attenuation hepatic nodules on computed tomography (CT) is thought to be highly suggestive of hepatocellular carcinoma.134 Renal involvement is nearly universal in patients with tyrosinemia. Findings include a decreased glomerular filtration rate, proximal renal tubular dysfunction, nephromegaly, phosphaturia (which is responsible for the development of rickets), glucosuria, and aminoaciduria. The toxic meta­ bolites SA and SAA are thought to have a direct effect on kidney function. Some patients progress to renal failure and need renal transplantation.135,136 One third of patients develop a cardiomyopathy, most commonly interventricular septal hypertrophy, that is reversible with either medical or surgical management of the disease.137 The neurologic manifestations may be the most concerning feature in older patients with tyrosinemia. More than 40% of patients experience porphyria-like symptoms.138 In a study of 20 children with HTI and 104 neurologic crises, the most common symptoms were pain (96%), hypertonia (76%), vomiting and ileus (69%), weakness (29%), and diarrhea (12%). Eight of the 104 patients required mechanical ventilation.138 A neurologic crisis has been considered to be a frequent cause of death, but the onset of a neurologic crisis may not be associated with worsening liver disease. Blockage of the degradation of ALA by SA is thought to be responsible for the neurotoxicity.

Diagnosis

The diagnosis of tyrosinemia should be suspected in any child with neonatal liver disease or a bleeding diathesis or in any child older than one year with undiagnosed liver disease or rickets. The diagnosis is suggested by increased tyrosine, methionine, phenylalanine, and α-fetoprotein levels. Elevated serum and urine SA and urine ALA levels are regarded as pathognomonic for tyrosinemia. The diagnosis can be confirmed with an assay for FAH in lymphocytes, erythrocytes, skin fibroblasts, or liver tissue.127 Prenatal diagnosis can be performed by determining SA levels in amniotic fluid or measuring FAH activity in chorionic villus biopsy specimens. If the specific gene mutation

Chapter 76  Other Inherited Metabolic Disorders of the Liver in a family is known, early genetic diagnosis can be made from chorionic villus biopsy specimens as well.127 Current newborn screening programs measure tyrosine levels on dried blood specimens; unfortunately, neonatal hypertyrosinemia is not specific for a diagnosis of HT1, and the tyrosine levels in newborns with HT1 may overlap substantially with tyrosine levels found in unaffected newborns. Improved newborn screening methodologies have been developed that measure SA in addition to amino acid levels in dried blood specimens; however, these methods are not yet widely employed.139,140

Treatment

Historically, the treatment of tyrosinemia has been dietary management, with restriction of tyrosine and phenylalanine. Dietary restriction has been shown to reverse renal damage and improve metabolic bone disease, although the liver disease progresses. An adequate intake of these amino acids is needed to ensure normal growth and development.127 Few studies of the long-term outcome in tyrosinemia treated with strict dietary management alone are available. Liver transplantation has become a mainstay of therapy for patients with tyrosinemia. The transplant corrects the phenotype and normalizes FAH activity and liver function. Additionally, the biochemical profiles normalize and kidney disease abates, with improvement in glomerular filtration rate, tubular acidosis, and hypercalcemia in most patients.141,142 Abnormal renal size and architecture persist after liver transplantation,143 and many patients continue to excrete SA despite normal serum values.142 One report has suggested that patients with tyrosinemia who have undergone liver transplantation and receive tacrolimus-based immunosuppression have a high rate of post-transplant neurologic complications.144 In 1992, Lindstedt and associates145 published data on the treatment of tyrosinemia with the herbicide 2-(2-nitro4-trifluoro-methylbenzoyl)-1,3-cyclohexanedione (NTBC). Later, Holme and Lindstedt146 published the results of a large long-term study of 220 patients with HT1 who were treated with this agent for up to seven years. NTBC is a potent inhibitor of 4-hydroxyphenylpyruvate dioxygenase, one of the initial steps in tyrosine metabolism. Blocking the degradation of tyrosine to its downstream toxic metabolites (i.e., FAA, SA, and SAA) was postulated to lead to improved hepatic function. Treated patients exhibit improved liver synthetic function, as reflected by a shortening of the prothrombin time, as well as decreased serum aminotransferase levels and a reduction in liver parenchymal heterogeneity and nodules on CT. In addition, serum α-fetoprotein and ALA levels are diminished and renal tubular dysfunction reverses.147 Therefore, elevated α-fetoprotein levels in a patient receiving NTBC therapy should raise concern about the patient’s nonadherence to therapy or the development of hepato­cellular cancer.127,148 Long-term results show continued improvement in all parameters noted in the earlier report as well as a lower risk for the development of hepatocellular carcinoma in patients who start therapy and were free of hepatocellular carcinoma before the age of two years.146 No patient withdrew from the study because of adverse side effects of the drug. Transient thrombocytopenia and neutropenia as well as ocular symptoms suggestive of corneal irritation have been noted rarely. Cognitive impairment resulting in learning problems may be a long-term complication of chronic use of NTBC in this patient population, possibly from the effects of chronic hypertyrosinemia.149 In another study from Quebec,150 only 4 of 35 patients treated with NTBC underwent liver transplantation; 1

patient received a transplant because of concern about the heterogeneous texture of the liver (suggestive of cirrhosis) shown by ultrasonography coupled with a persistent moderate elevation in the serum α-fetoprotein level. At resection, this child was found to have a small nodule with hepatocellular dysplasia. The 31 remaining patients were monitored while receiving NTBC therapy for up to three years, and none experienced neurologic crises or deterioration of liver disease.150 Therefore, therapy with NTBC significantly improves the clinical course of patients treated at an early age. For those in whom therapy is initiated at a later age, NTBC offers a palliative benefit, although the risk of hepatocellular carcinoma is still high in this group of patients. Because the strategy of treating patients with HT1 with NTBC is relatively new, greater experience is required to enable assessment of the relative costs and long-term outcome and recognition of any possible adverse effects of long-term NTBC therapy.149 Early diagnosis, achieved by inclusion of HT1 in neonatal screening programs, may allow the prompt initiation of effective therapy with NTBC and avoidance of liver transplantation.

UREA CYCLE DEFECTS Although the syndromes related to the urea cycle defects (UCDs) are not associated with serious liver injury, the basic genetic defect is located within the liver, and the manifestations can mimic those of other metabolic liver diseases. The urea cycle consists of five enzymes that, through several steps, process ammonia derived from amino acid metabolism to urea. Genetic defects in each of these enzymes have been reported, and their overall incidence has been estimated to be 1 in 20,000 to 1 in 30,000.151

Pathophysiology

The steps of the urea cycle are illustrated in Figure 76-6. Carbamyl phosphate synthetase (CPS) I forms carbamyl phosphate from ammonium and bicarbonate. This step requires N-acetyl glutamate as a cofactor, which is synthesized from N-acetyl CoA and glutamic acid by N-acetyl glutamate synthetase. Ornithine transcarbamylase (OTC) combines carbamyl phosphate with ornithine to form citrulline. A second nitrogen enters the cycle as aspartate, which combines with citrulline by the action of argininosuccinate synthetase (AS) to form argininosuccinate. Argininosuc­ cinate is converted to arginine and fumarate by arginino­ succinase, or argininosuccinate lyase (AL). Arginase then catalyzes the breakdown of arginine to urea and ornithine in the final step of the pathway. Several amino acid transporters, such as citrin, an aspartate/glutamate carrier protein that supplies aspartate to the urea cycle, are involved in shuttling metabolites into the urea cycle.152 CPS II, through the pyrimidine synthetic pathway, leads to the formation of orotic acid. Excess carbamyl phosphate can be used by this pathway if a block occurs distal to OTC in the metabolic pathway. Excess nitrogen in the form of amino acids can be shunted to alternative pathways of waste-nitrogen excretion by the medicinal use of sodium benzoate and sodium phenylacetate, leading to the generation of hippurate and phenylacetylglutamine, respectively. Enzymatic defects have been identified in all five steps involved in the urea cycle. Deficiency of four of these enzymes is transmitted through autosomal recessive inheritance, whereas OTC deficiency is transmitted as an X-linked trait. More than 340 different mutations in the OTC gene

1271

1272

Section IX  Liver +

NH 4 Glycine

Glutamine

+

+

Benzoate

Phenylbutyrate

HCO 3

–

Carbamyl phosphate synthetase Carbamyl phosphate Orotate

Hippurate

Ornithine transcarbamylase

Phenylacetylglutamine Ornithine

Citrulline Aspartate Argininosuccinate synthetase

Urea

Urine Figure 76-6.  The urea cycle. Alternative pathways that are used therapeutically for waste nitrogen disposal are also illustrated. Enzymes are shown in italics.

give rise to OTC deficiency, the most common urea cycle defect.153 Numerous defects in the other enzymes or amino acid transporters of the cycle (e.g., N-acetylglutamate synthetase or citrin) have been characterized as well.154-156 Moreover, numerous mRNA instability mutations have been found in patients with CPS I deficiency.157 A UCD has two main biochemical consequences: Arginine becomes an essential amino acid (except in arginase deficiency), and nitrogen accumulates in a variety of mole­ cules, some of which can have deleterious toxic effects.

Clinical Features

The spectra of clinical presentations in patients with any of the UCDs are virtually identical; in the neonatal period, these disorders usually manifest as acute lifethreatening events. Contrary to the commonly held belief that UCDs are primarily disorders of the newborn period, a large report of the clinical presentation and outcome data of 260 patients with UCDs revealed that two thirds of patients present at greater than 30 days of age, even when female OTC carriers are excluded from the analysis.158 Another large cross-sectional study of 183 patients with UCDs had similar findings.159 Late-onset adult presentations also have been described. Affected infants appear normal for the first 24 to 72 hours until they are exposed to their first feeding, which provides the initial protein load that fosters ammonia production. Symptoms include irritability, poor feeding, vomiting, lethargy, hypotonia, seizures, coma, and hyperventilation, all secondary to hyperammonemia.158 Initially, neonates are often mistakenly thought to have sepsis, despite the absence of perinatal risk factors, and

Arginase Arginine

Argininosuccinate

Argininosuccinase Fumarate

diagnostic laboratory testing is delayed.160 Plasma ammonia levels should be obtained whenever an evaluation for sepsis is initiated in a neonate; levels may exceed 2000 µmol/L (3400 mg/dL), normal levels being 50 µmol/L (85 mg/dL) or less. For all age groups, overall survival decreases as the peak plasma ammonia level rises for a given episode of hyperammonemia with survival rates of 98% and 47% for peak ammonia levels of less than 200 µmol/L and greater than 1000 µmol/L, respectively.161 Newborns have a survival rate of 73% after their presenting episode of hyperammonemia, whereas patients older than 30 days of age have a survival rate of 98%. Male patients with OTC deficiency have a survival rate of 91% following an episode of hyperammonemia, a rate significantly less than those (93% to 98%) for all other forms of UCDs. Blood gas analysis shows respiratory alkalosis secondary to the hyperventilation caused by the effects of ammonia on the central nervous system. Blood urea nitrogen levels are typically low but can be elevated during times of dehydration or hypoperfusion. Serum levels of liver enzymes are usually normal or minimally elevated. Severe hepatomegaly can occur in early-onset forms of AL deficiency.151 As stated earlier, one third of patients with UCDs present in the neonatal period; the remainder are diagnosed at variable times from infancy to adulthood.158,159 OTC deficiency is the most common UCD (55%), followed by argininosuccinic aciduria (AL deficiency, 16%) and citrullinemia (AS deficiency, 14%).159 Male patients with OTC deficiency have been diagnosed as late as 40 years of age with varied phenotypic presentations.162 As many as 10% of female car-

Chapter 76  Other Inherited Metabolic Disorders of the Liver Table 76-4  Laboratory Values in Urea Cycle Defects ENZYME DEFICIENCY Carbamyl phosphate synthetase Ornithine transcarbamylase Argininosuccinate synthetase Argininosuccinase Arginase

AMMONIA (PLASMA)

CITRULLINE (SERUM)

ARGININOSUCCINATE (URINE OR SERUM)

OROTIC ACID (URINE)

ARGININE/ORNITHINE (SERUM)

↑–↑↑↑

↓

↓

↓

↓

↑–↑↑↑ ↑–↑↑↑ ↑–↑↑↑ ↑

↓ ↑↑↑ ↑↑↑ ↑↑

↓ ↓ ↑↑↑ ↑↑

riers of OTC deficiency can have symptoms, which may be severe and fatal, although most female carriers have no symptoms or report only nausea after high-protein meals.163 Late-onset CPS deficiency has also been described,164 and the adult form of AS deficiency is relatively common in Japan.155 Symptoms and signs of late-onset UCDs, especially OTC and CPS deficiencies, include episodic irritability, lethargy, or vomiting; self-induced avoidance of protein such as milk, eggs, and meats; and short stature or growth delays. Neurologic symptoms, which can also be episodic, include ataxia, developmental delays, behavioral abnormalities, combativeness, biting, confusion, hallucinations, headaches, dizziness, visual impairment, diplopia, anorexia, and seizures.158 Acute hyperammonemic episodes can resemble Reye’s syndrome (see Chapter 86). Such episodes can be precipitated by high-protein meals, viral or bacterial infections, medications, trauma, or surgery. Infants may present after being weaned from breast milk to infant formulas, which have a higher protein content. Patients with OTC and CPS deficiencies have been reported to present in the postpartum period with acute decompensation and death.165,166 Citrin deficiency is caused by mutations in the SLC25A13 gene and is associated with both adult-onset type 2 citrullinemia and neonatal intrahepatic cholestasis resulting from citrin deficiency (NICCD), a syndrome that primarily affects newborns in East Asia.155,167,168 A case of citrin deficiency in a white infant of European descent who presented with poor weight gain and a bleeding diathesis, but without cholestasis, has been reported, expanding the clinical spectrum of presentations of citrin deficiency.169 NICCD is associated with hyperaminoacidemia (e.g., elevated citrulline, methionine, and tyrosine levels) and cholestasis.155 Hypergalac­ tosemia and elevated acylcarnitine levels also may be observed.170 Hepatic steatosis may be seen in the liver histologically.171 In most patients with NICCD, all biochemical abnormalities resolve spontaneously or with minimal dietary restrictions (e.g., the use of lactose-free formulas); however, several affected infants have required liver transplantation before age one year.167 Therefore, jaundiced infants with multiple abnormal newborn metabolic screen results (e.g., elevation of blood phenylalanine, methionine, or galactose levels) must be observed closely because of the risk for development of ESLD caused by NICCD; a chubby face outside of the realm of normal may be a diagnostic clue.172 The diagnosis can be made by sequence analysis of the SLC25A13 gene or by failure to detect citrin protein in peripheral lymphocytes.168,173

Diagnosis

A high index of suspicion is required for prompt diagnosis of UCDs. Symptoms can mimic those of other acute neonatal problems, such as infections, seizures, and pulmonary or cardiac disease.158,174 Later presentations can mimic other behavioral, psychiatric, or developmental disorders. The

↑↑ Normal–↑ Normal–↑ Normal–↑

↓ ↓ ↓ ↑↑

first clue may be an elevated serum ammonia level with normal serum aminotransferase levels and without metabolic acidosis. Therefore, if a UCD is considered, the following laboratory measurements should be obtained: serum ammonia, arterial blood gases, urine organic acids, serum amino acids, and urinary orotic acid; Table 76-4 reviews the expected laboratory results. Urinary organic acid profiles are typically normal in UCDs. The plasma amino acid profiles are distinctive, with abnormal levels of arginine, ornithine, and citrulline. Citrulline levels are barely detectable in OTC or CPS deficiencies but markedly raised in AS and AL deficiencies. AS deficiency can be distinguished from AL deficiency by the finding of argininosuccinic acid in the plasma and urine of patients with AL deficiency. OTC deficiency is differentiated from CPS deficiency by the excessive urinary excretion of orotic acid in OTC deficiency. Direct enzyme analysis can be performed and can be useful in patients who have a partial deficiency or who present in adulthood. Prenatal enzyme and genetic linkage analysis can be carried out in family members of known carriers.175 Early neonatal diagnosis leads to improved survival. An N-carbamoyl-glutamic acid test has been advocated for all patients presenting with a suspected UCD while further testing is carried out; patients with N-acetylglutamate synthase deficiency, one of the less common UCDs, have a significant drop in plasma ammonia levels within eight hours of receiving an oral dose of N-carbamoyl glutamic acid, a structural analog of N-acetyl glutamic acid.176,177 An allopurinol loading test, which leads to excretion of orotic acid in amounts that are 10- to 20-fold greater than normal in heterozygote female carriers of OTC deficiency, is nonspecific, and its results must be interpreted with caution; the result can be positive in some patients with mitochondrial disease or defects in pyrimidine metabolism.178 Liver biopsy specimens typically show minimal fatty infiltration; fibrosis or cirrhosis are uncommon findings but have been reported.151,179

Treatment

All external protein intake should be discontinued in infants with UCDs presenting acutely. Serum ammonia levels should be restored to normal. The use of oral lactulose to lower the nitrogen load has not been studied in this patient population. Given the extremely high ammonia levels often encountered, continuous arteriovenous hemodialysis or hemofiltration is often required, but exchange transfusions and peritoneal dialysis are ineffective. Alternative pathways for waste nitrogen disposal should be employed, specifically intravenous administration of sodium benzoate and sodium phenylacetate; however, sodium benzoate should be used with caution in patients with cirrhosis because a paradoxical rise in blood ammonia levels has been observed.180 Arginine, carnitine, and long-chain fatty acids are usually present in low levels in these patients and should be supplemented.181,182 Once the patient stabilizes,

1273

1274

Section IX  Liver low levels of dietary protein, 0.5 to 1.0 gm/kg, may be introduced, with progressive increases as tolerated to provide sufficient protein for growth and tissue repair while minimizing urea production. Long-term therapy and protein restriction are then tailored to the patient; patients with a severe disorder may need essential amino acids to supplement their protein intake.182 Oral phenyl­butyrate can be substituted for phenylacetate to improve palatability. The outcome for patients who present with hyperam­ monemic coma and a delayed diagnosis is poor.161 The level of ammonia at the time of the first hyperammonemic episode is a rough guide to the eventual neurodevelopmental outcome.183 The sooner the hyperammonemia is treated and the correct diagnosis is made, the better the long-term survival, although for patients who survive the neonatal period, the median survival without liver transplantation is less than four years and is associated with severe developmental delay and neurologic morbidity.160 With optimal dietary and medical management, patients may still have repeated hyperammonemic crises, often during intercurrent viral infections. Symptomatic female OTC heterozygotes also benefit from therapy, which leads to fewer hyperammonemic episodes and a reduced risk of further cognitive decline.184 Patients with a UCD and a deterioration or lack of im­provement despite therapy have undergone either total or auxiliary liver transplantation successfully, with normalization of enzyme activity and ammonia levels, the ability to tolerate a normal diet, and five-year survival rates of 90%.185,186 Liver transplantation, if considered, should be done before neurologic damage is permanent because the patient’s neurologic status does not improve after transplantation. The metabolic condition of the patients normalizes completely, and neurologic status does not worsen after transplantation.164 In addition, hepatocyte transplantation has been successful in improving metabolic function in patients with UCDs and has been applied successfully as a bridge to auxiliary liver transplantation.186,187 For patients without severe neurologic compromise before liver transplantation, this therapeutic approach is worthwhile. In addition, the annual cost of care for this group of patients is likely to be reduced dramatically after liver transplantation. The importance of identifying the deleterious mutation in a patient with a UCD will likely become increasingly important not only as a means of allowing carrier testing and prenatal diagnosis, but also as an aid to treatment decisions. For example, patients with the mutations of OTC deficiency that result in the most severe disease (e.g., abolished enzyme activity) may benefit preferentially from immediate liver transplantation to prevent severe mental retardation or death, whereas those with mutations that lead to milder disease may be better managed medically with dietary restrictions and ammonia scavengers to facilitate growth before possible liver transplantation.

ARGINASE DEFICIENCY

At least two forms of arginase activity occur in humans: Arginase I (AI) predominates in the liver and red blood cells, and arginase II (AII) is found predominantly in kidney and prostate. Arginase deficiency involving AI is the least common of the UCDs. Hyperammonemia is unusual in affected persons, but hyperammonemic coma and death have been reported.188 The clinical disorder is distinct from the other UCDs. It is characterized by indolent deterioration of the cerebral cortex and pyramidal tracts, leading to progressive dementia and psychomotor retardation, spastic diplegia progressing to quadriplegia, seizures, and

growth failure. The syndrome is often confused with cerebral palsy.189 Laboratory studies may reveal elevated blood arginine values, mild hyperammonemia, and a mild increase in urine orotic acid excretion. Varying amounts of urea are still produced in these patients secondary to the compensatory elevated expression of AII in the kidneys that ameliorates the clinical disorder. The diagnosis is confirmed by enzymatic analysis, which can be performed prenatally on cord blood samples. Treatment consists of protein restriction and, when needed, medical therapy with sodium phenylbutyrate.189

BILE ACID SYNTHESIS AND TRANSPORT DEFECTS The pathways for bile acid synthesis and transport within the hepatobiliary system are complex, involving several enzymes and transport processes located in multiple sub­ cellular fractions of the hepatocyte (see Chapter 64). With technologic advances in molecular biology and mass spectrometry, several different inborn errors in bile acid synthesis and transport have been identified as causes of clinical disease. The definition and classification of these disorders have improved, particularly in the clinically heterogeneous subset of cases that comprise progressive familial intrahepatic cholestasis (PFIC) syndromes. For some of these disorders, this progress has led to dramatic advances in often life-saving therapy. The diagnosis of PFIC is imprecise; the accepted criteria are (1) the presence of chronic, unremitting intrahepatic cholestasis, (2) exclusion of identifiable metabolic or anatomic disorders, and (3) characteristic clinical, biochemical, and histologic features.190 Other symptoms and signs are severe pruritus, hepatomegaly, wheezing and cough, short stature, delayed sexual development, fat-soluble vitamin deficiency, and cholelithiasis. Affected persons exhibit severe and progressive intrahepatic cholestasis, usually manifesting within the first few months of life and often proceeding to cirrhosis and ESLD by the second decade of life. Patients with PFIC syndromes have been found to have defects in bile acid synthetic and transport processes. Some patients believed previously to have idiopathic neonatal hepatitis or an undiagnosed familial hepatitis syndrome may now be diagnosed accurately as having a form of PFIC. An estimated 1% to 2.5% of patients with idiopathic cholestasis may have defects in bile acid metabolism and transport.191 Table 76-5 lists the known errors of primary and secondary bile acid synthesis and transport.

BILE ACID SYNTHESIS DEFECTS

Bile acid synthetic pathways are discussed in detail in Chapter 64. Only the most common enzyme deficiencies are described here; all can be diagnosed through mass spectrometry of the urine or serum. General principles of therapy rely on the hypothesis that inborn errors of bile acid biosynthesis lead to underproduction of normal trophic and choleretic primary bile acids and overproduction of hepatotoxic primitive bile acid metabolites.192 Bile acids found in patients with inborn errors also act as cholestatic agents by inhibiting canalicular ATP-dependent bile acid transport (the rate-limiting step in the overall process of bile acid transport across the canalicular membrane).193 Cerebrotendinous xanthomatosis (CTX), C27-steroid-27hydroxylase deficiency, is characterized by bilateral juvenile cataracts and chronic diarrhea, followed by progressive

Chapter 76  Other Inherited Metabolic Disorders of the Liver Table 76-5  Inborn Errors of Bile Acid Synthesis and Transport* Defective Bile Acid Synthesis Primary defects Cerebrotendinous xanthomatosis (C27-steroid27-hydroxylase deficiency) 3β-hydroxy-ΔC27-steroid dehydrogenase/ isomerase (HSD3B7) deficiency Δ4-3-oxosteroid 5β-reductase (AKR1D1) deficiency C24-steroid-7α-hydroxylase (CYP7B1) deficiency α-Methylacyl-CoA racemase (AMACR) deficiency Secondary defects Peroxisomal biogenesis disorders (PBDs)   Rhizomelic chondrodysplasia punctata (due to organelle   Zellweger spectrum damage)    Zellweger’s syndrome    Neonatal adrenoleukodystrophy    Infantile Refsum’s disease   Other    Hyperpipecolic acidemia    Leber’s congenital amaurosis Disorders with loss of single peroxisomal function   Acatalasemia   Acyl-CoA oxidase deficiency   Adult Refsum’s disease   D-bifunctional protein deficiency   Hyperoxaluria type I   Sterol carrier protein X deficiency   Thiolase deficiency (pseudo-Zellweger’s syndrome)   X-linked adrenoleukodystrophy Contiguous gene syndrome Generalized hepatic synthetic dysfunction   Fulminant hepatic failure (multiple causes)   Neonatal iron storage disease   Tyrosinemia Defective Bile Acid or Phospholipid Transport FIC1 (FIC1) deficiency: progressive familial intrahepatic cholestasis (PFIC) type 1:   Byler’s disease   Benign recurrent intrahepatic cholestasis (BRIC) syndrome   Greenland familial cholestasis BSEP (ABCB11) deficiency: PFIC type 2 MDR3 (ABCB4) deficiency: PFIC type 3 Others   Alagille (Jagged) syndrome (see Chapter 62)   Cholestasis-lymphedema syndrome (Aagenaes syndrome)   North American Indian childhood cirrhosis (Cirhin) *Corresponding genes are shown in italics. Modified from Balistreri WF. Inborn errors of bile acid biosynthesis: Clinical and therapeutic aspects. In: Hofmann AF, Paumgartner G, Stiehl A, et al, editors. Bile Acids in Gastroenterology, Basic and Clinical Advances. London: Kluwer Academic Publishers; 1995. p 333; Jonas MM, Perez-Atayde AR, editors. Liver disease in infancy and childhood. In: Schiff ER, Sorrell MF, Maddrey WC, editors. Schiff’s Diseases of the Liver, 9th ed. Philadelphia: Lippincott Williams & Wilkins; 2003. p 1459.

neurologic dysfunction, hypercholesterolemia, atherosclerosis, and deposition of cholesterol and cholestanol in tissues.192,194,195 CTX should be treated with chenodeoxycholic acid. In one series,196 five children with CTX showed dramatic improvement in biochemical and electroencephalographic abnormalities and resolution of diarrhea with this therapy. In addition, no further delay in motor development was noted, and three patients showed an improved intelligence quotient.196 Patients with 3β-hydroxy-C27-steroid dehydrogenase/ isomerase (3β-HSD) deficiency may present with pruritus,

jaundice, hepatomegaly, steatorrhea, and fat-soluble vitamin deficiencies. Deficiency of Δ4-3-oxosteroid 5β-reductase also leads to neonatal cholestasis but rapidly progresses to synthetic dysfunction and liver failure. Both conditions have been treated with a primary bile acid (i.e., cholic acid) and ursodeoxycholic acid supplementation.191,197 Cholic acid bypasses the enzymatic block and provides negative feedback to earlier steps in the synthetic pathways, and ursodeoxycholic acid displaces toxic bile acid metabolites and serves as a hepatobiliary cytoprotectant. Other known disorders result from C24-steroid-7α-hydroxylase, α-methylacylCoA racemase, and amidation deficiencies.191,192 Many peroxisomopathies have been described; these disorders are associated with multiple clinical abnormalities and up to 50 wide-ranging biochemical abnormalities.198 They are diagnosed through a combination of specialized tests, such as that for very-long-chain fatty acids (VLCFAs) and ultrastructural analysis of tissue biopsy specimens.199 Peroxisomes are responsible for beta-oxidation in the final steps of bile acid synthesis to yield the primary bile acids cholic acid and chenodeoxycholic acid. The disorders can be divided into three groups: peroxisome biogenesis disorders (PBDs), which cause multiple abnormalities; disorders of single proteins, which result in limited dysfunction; and a contiguous gene syndrome (see Table 76-5).198 PBDs, which are grouped because they share similar clinical and biochemical features, include rhizomelic chondrodysplasia punctata (RCDP), which is characterized by severe rhizomelic shortening of the limbs, severe skeletal abnormalities, cataracts, and facial abnormalities, and the three Zellweger spectrum disorders, which include Zellweger’s syndrome (ZS), with the most severe clinical abnormalities, neonatal adrenoleukodystrophy, and infantile Refsum’s disease, with the mildest features.198 ZS is a primary disorder of peroxisome biogenesis. The multiple features of ZS include distinctive dysmorphic features (hypertelorism, large anterior fontanelle, deformed earlobes), neonatal hypotonia, impaired hearing, retinopathy, cataracts, seizures, and skeletal changes. Hepatomegaly is common, and the progressive liver disease that develops in patients with ZS is similar to that identified in other errors of bile acid synthesis.200 Peroxisome biogenesis involves more than 13 PEX genes and requires the targeting and importation of cytosolic proteins into the peroxisomal membrane and matrix. Importation of proteins fated for the peroxisomal matrix requires guidance from one of two peroxisome-targeting signals, PTS1 and PTS2. Patients with Zellweger spectrum disorders display defects in the importation of proteins that use PTS1 and PTS2, whereas patients with RCDP have a defect in the importation of proteins using PTS2.198 The most common disorder of peroxisomes, X-linked adrenoleukodystrophy (X-ALD), is included in the second grouping of peroxisomopathies and results from a defect in the peroxisomal adrenoleukodystrophy protein (ALDP), which is a member of the ATP-binding cassette (ABC) superfamily of membrane transporters. Several other single peroxisomal enzyme deficiencies exist, as noted in Table 76-5. The last group, contiguous gene syndrome, involves a large deletion spanning chromosome Xq28 and affecting multiple genes involved in peroxisome function.198,201 Historically, the treatment of ZS has been supportive, with most patients not surviving the first year of life. The goals of medical therapy have been to improve nutrition and growth, control central nervous system symptoms, and limit progression of liver disease. Treatment with primary bile acids has been shown to improve biochemical studies and histology, increase growth, and improve neurologic

1275

1276

Section IX  Liver symptoms.202 Adrenal hormone replacement therapy is necessary for patients with X-ALD; the only known effective treatment for the cerebral form of this condition is hematopoietic stem cell transplantation, which if performed early, leads to good neurologic outcomes.203,204 In an uncontrolled study, the use of docosahexaenoic acid (DHA) led to improved myelination in patients with peroxisomal disorders; currently, however, the use of DHA should be limited to controlled clinical trials.205

BILE ACID TRANSPORT DEFECTS

The study of intrahepatic cholestasis syndromes has greatly enhanced our understanding of hepatic excretory function and bile acid metabolism, and vice versa. The spectrum of diseases associated with mutations in genes involved in bile acid transport physiology is large and growing with continued scientific advances; the precise terminology used to describe these disorders continues to evolve as well. Three specific disorders of bile acid transport defects will be reviewed here: familial intrahepatic cholestasis 1 (FIC1) disease, bile salt export pump (BSEP) disease, and multidrug resistance protein 3 (MDR3) disease. FIC1 disease encompasses a spectrum of at least two disease states (1) Byler’s disease and Byler’s syndrome, which generally present in infancy and lead to progressive cholestasis often associated with severe pruritus, and (2) benign recurrent intrahepatic cholestasis (BRIC) type I, which gives rise to recurrent episodes of intrahepatic cholestasis beginning in childhood or adulthood that can last days to months and resolve spontaneously without causing detectable lasting liver damage. Byler’s disease was first described in 1965 in an Amish kindred descended from Jacob Byler. Patients with a clinical disorder similar to Byler’s disease, but who are not of Amish descent, are said to have Byler’s syndrome. FIC1 disease is caused by mutations in the ATP8B1 gene (initially named the FIC1 gene) that encodes the FIC1 protein, a P-type adenosine triphosphatase involved in ATP-dependent aminophospholipid transport. FIC1 protein is expressed in many organs in addition to the liver, including the intestine and pancreas. Severe mutations in FIC1 lead to progressive early-onset FIC1 disease (PFIC type 1), whereas mutations in FIC1 that are predicted to affect protein structure minimally are more commonly seen in patients with BRIC type 1.206,207 Patients with Greenland familial cholestasis also have distinct defects in the FIC1 gene, as do other kindreds of Dominican and Saudi descent.207,208 In PFIC type 1, serum GGTP and cholesterol levels are normal or mildly elevated, and levels of bile acids are elevated in the serum and low in the bile. Serum aminotransferase and bilirubin levels are mildly elevated as well. Impaired bile acid transport in the intestine may account for the striking malabsorption and diarrhea manifested by some patients with PFIC type 1; these intestinal clinical features do not resolve after liver transplantation. Histology of liver tissue from patients with PFIC type 1 disease typically shows a bland canalicular cholestasis, with varying degrees of hepatocellular ballooning and giant cell transformation; portal fibrosis and eventually cirrhosis may be seen later in the course of the disease. The liver histology of patients with BRIC is normal between episodes. On electron microscopic evaluation of liver tissue from patients with PFIC type 1, characteristic coarse, granular bile deposits are seen in the canaliculus (“Byler’s bile”).208 BSEP disease is caused by a wide spectrum of mutations in the ABCB11 gene, which encodes an ABC protein that serves as the canalicular BSEP.206 Patients with the progres-

sive form of BSEP disease (PFIC type 2) present with high serum bile acid levels but low or low-normal serum GGTP levels, and they usually have intense pruritus, jaundice, poor weight gain, and hepatosplenomegaly.208,209 In addition, a genetically distinct form of BRIC (type 2) is associated with mutations in ABCB11; interestingly, patients with BRIC type 2 commonly have cholelithiasis and the absence of extrahepatic manifestations.206,210 Early in the course of PFIC type 2, a nonspecific giant cell hepatitis is found on routine histology of the liver, and on electron microscopy, amorphous bile deposits are seen in the canaliculi. For unclear reasons, patients with PFIC type 2 appear to have an increased risk of developing malignancies of the hepatobiliary system, such as hepatoblastoma, hepato­ cellular carcinoma, and cholangiocarcinoma, in contrast to patients with other forms of progressive intrahepatic cholestasis.208,211-213 MDR3 disease is caused by mutations in the ABCB4 gene that encodes the MDR3 glycoprotein, an ABC phosphatidylcholine translocase located on the canalicular membrane.206,214 Patients with MDR3 disease present with several disease phenotypes as well, including PFIC type 3, intrahepatic and gallbladder lithiasis, intrahepatic cholestasis of pregnancy, and adult-onset ductopenic cholestatic liver disease.215-217 Moreover, mutations in MDR3 may serve as disease modifiers for primary sclerosing cholangitis, primary biliary cirrhosis, and drug-induced cholestasis.218 MDR3 deficiency is thought to lead to decreased excretion of cytoprotective biliary phospholipids, leaving an increased pool of cytotoxic biliary bile salts that gives rise to subsequent bile duct damage and proliferation and release of GGTP into the serum. Patients with PFIC type 3 present with high serum levels of GGTP and bile acids as well as bile ductular proliferation on routine microscopy. Some female patients with familial intrahepatic cholestasis of pregnancy have been shown to be heterozygous carriers of a mutation in MDR3; the mutations likely lead to a genetic predisposition that requires the coexistence of other nongenetic factors for full expression of the disease.215,219 Other chronic intrahepatic cholestatic diseases such as North American Indian childhood cirrhosis (NAICC), cholestasis-lymphedema syndrome (Aagenaes syndrome),220 neonatal icthyosis and sclerosing cholangitis syndrome, and arthrogryposis-renal dysfunction—cholestasis (ARC) syndrome—have disease causing loci that are genetically distinct from those that cause FIC1, BSEP, or MDR3 diseases.221 Further molecular genetic characterization of these disorders has identified a single-point mutation in the cirhin gene, which encodes a nucleolar protein of unknown function, in patients with NAICC.222 Moreover, mutations in the claudin-1 and VPS33B (vacuolar protein sorting 33B) genes have been identified in patients with neonatal icthyosis and sclerosing cholangitis and in ARC syndromes, respectively; both of these genes encode proteins that are important in membrane fusion events.223,224 Medical treatment of patients with disorders of bile acid transport as a group, with phenobarbital, cholestyramine, opioid antagonists, rifampin, and phototherapy, has been largely ineffective or highly variably effective and of short duration of benefit at best. Therapy with ursodeoxycholic acid may be effective in reducing pruritus and improving liver biochemical test levels in up to 50% of patients with PFIC, regardless of serum GGTP levels.186 Surgical approaches such as ileal exclusion and partial external biliary diversion have provided satisfactory symptomatic relief to some patients by decreasing the bile acid pool and pruritus, especially in patients with PFIC type 1 and 2.225-227 Long-term marked improvement in pruritus and growth has

Chapter 76  Other Inherited Metabolic Disorders of the Liver been reported in six pediatric patients treated with biliary diversion.225 If all else fails, liver transplantation leads to good overall outcomes, with normalization of bile acid synthesis and growth, even in patients who receive a living donor organ from a potentially heterozygous parent.228,229

CYSTIC FIBROSIS Cystic fibrosis (CF) is discussed in detail in Chapter 57; here, a brief discussion of the hepatic complications seen in this multisystemic disorder is presented. Liver disease can be the presenting symptom of CF in the newborn, and CF-associated liver disease has been seen with meconium ileus syndrome. Other risk factors for CF-associated liver disease include male sex, pancreatic insufficiency, and a CF transmembrane regulator genotype associated with severe disease.230-232 Although CF has been identified in fewer than 2% of patients with neonatal cholestasis, the diagnosis should be considered in any infant who presents with neonatal cholestasis. CF-associated liver disease may become more common as the mean age of survival for patients with CF rises; however, liver involvement is not universal and seems to peak during the adolescent years. Up to 30% of patients may have clinical or symptomatic liver disease after the neonatal period.230 Hepatobiliary diseases noted in patients with CF can be grouped into three categories (Table 76-6). The pathognomonic lesion of CF, focal biliary cirrhosis (FBC), presumably results from defective function of the CF transmembrane regulator protein, which is expressed in bile duct cells. Obstruction of small bile ducts leads to chronic inflammatory changes, bile duct proliferation, and portal fibrosis. At

Table 76-6  Spectrum of Hepatobiliary Disease in Patients with Cystic Fibrosis Lesions specific to cystic fibrosis

Lesions secondary to extrahepatic disease

Lesions that occur with a higher frequency in patients with cystic fibrosis

Hepatic   Focal biliary cirrhosis with inspissation   Multilobular biliary cirrhosis with inspissation Biliary   Microgallbladder   Mucocele   Mucous hyperplasia of the gallbladder Hepatic lesions associated with cardiopulmonary disease   Centrilobular necrosis   Cirrhosis Pancreatic lesions   Fibrosis (leading to bile duct compression/stricture) Hepatic   Drug hepatotoxicity   Fatty liver   Neonatal cholestasis   Viral hepatitis Biliary   Biliary sludge   Cholangiocarcinoma   Cholelithiasis   Sclerosing cholangitis

Modified from Balistreri WF. Liver disease in infancy and childhood. In: Schiff ER, Sorrell MF, Maddrey WC, editors. Schiff’s Diseases of the Liver, 9th ed. Philadelphia: Lippincott-Raven; 1999. p 1379.

autopsy, FBC has been identified in 25% to 30% of patients older than one year of age.233 Progression to multilobular biliary cirrhosis occurs in approximately 5% to 10% of patients with CF and leads to symptoms associated with portal hypertension, such as splenomegaly and variceal bleeding.230,231 Hepatic steatosis also develops in roughly one half of patients but does not appear to correlate with outcome. Biliary abnormalities range from microgall­bladder, which is largely asymptomatic and is found in up to 20% of patients, to cholelithiasis and cholangiocarcinoma.231,234 Interestingly, the presence of liver disease does not correlate with the severity of pulmonary disease.230,232 The variable occurrence and clinical course of liver disease in patients with CF may be related to other genetic factors. For example, elevated concentrations of endogenous ursodeoxycholic acid have been documented in patients with CF without liver disease, thus raising the possibility that ursodeoxycholic acid may protect against liver injury in these patients.235 The diagnosis of liver disease in patients with CF can be difficult because the presenting signs are subtle. Hepatomegaly, which is present in approximately 30% of patients, has been shown to correlate well with the presence of cirrhosis and is often the first finding of liver disease. Liver biochemical test levels may remain relatively normal despite histologic evidence of cirrhosis. Needle biopsy of the liver can be helpful; however, because of the focal distribution of histologic abnormalities, sampling error may occur. Ultrasonography can detect the presence of biliary tree abnormalities as well as heterogeneous or nodular liver parenchyma; however, a normal study does not preclude significant hepatic fibrosis.236,237 Treatment with ursodeoxycholic acid improves liver biochemical test levels in patients with CF; however, conclusive evidence that the drug halts the progression to cirrhosis is not yet available.238 Because patients with CF rarely have true hepatocellular dysfunction, the role of liver transplantation in this patient population remains controversial, particularly given the scarcity of donor livers. Nevertheless, liver transplantation has been performed successfully in patients with CF who have portal hypertension and stable pulmonary function, with post-transplant outcomes and improvement in nutritional status comparable to those for patients without CF.239,240 A portosystemic shunt can be an effective treatment for a patient with CF and portal hypertension, with long-term outcomes comparable to that for patients with CF who undergo liver transplantation.241

MITOCHONDRIAL LIVER DISEASES A growing number of liver diseases have been attributed to defects in mitochondrial function. In addition to defects in mitochondrial enzymes involved in the urea cycle or energy metabolism, several mitochondrial hepatopathies involve respiratory chain/oxidative phosphorylation/ electron transport defects or alterations in mitochondrial DNA (mtDNA) levels. The mitochondrial genome is especially vulnerable to oxidative injury not only because of its spatial relationship to the respiratory chain, but also because of its lack of protective histones and an adequate excision and recombination repair system. Mitochondrial DNA is inherited almost entirely from the maternal ovum; therefore, many primary mitochondrial deficiencies are inherited in a dominant fashion. Many nuclear genes, however, such as DNA polymerase-γ (POLG), thymidine kinase 2 (TK2), deoxyguanosine kinase (DGUOK), SCO1,

1277

1278

Section IX  Liver BCS1L, and MPV17, encode proteins critical to maintaining proper amounts of mitochondrial DNA and to allowing normal mitochondrial respiratory function. Most mito­ chondrial diseases with primary involvement of the liver are caused by nuclear rather than mitochondrial gene mutations.242 Mitochondrial respiratory chain disorders can affect 1 in 20,000 children, with liver involvement occurring in 10% to 20% of patients.242,243 Striking heterogeneity of clinical features, ranging from single-organ involvement to multisystem disease, can lead to a delayed or missed diagnosis and can confound therapeutic decision making, for example, with respect to the advisability of liver transplantation.242 The diagnosis of a mitochondrial respiratory chain defect should be considered in a patient with liver disease who has unexplained neuromuscular symptoms, including a seizure disorder; involvement of seemingly unrelated organ systems; a rapidly progressive course; or a chronic course that proves to be a diagnostic dilemma.242 In about 80% of patients, symptoms appear early in life, before age two. The plasma lactate level and the ratio of lactate to pyruvate are often, but not universally, elevated, especially with insidious presentations.242 Neonatal liver failure has been reported in association with cytochrome c oxidase deficiency caused by mutations in the SCO1 or BCS1L genes. The key features of these disorders are lactic acidemia and an elevated ratio of plasma lactate to pyruvate levels. Infants with AlpersHuttenlocher syndrome (progressive neuronal degeneration in childhood with liver disease ascribed to mitochondrial dysfunction) experience vomiting, hypotonia, seizures, and liver failure, often beginning by age six months. Frequently, the liver disease is unsuspected clinically and becomes evident late in the course of the disease; AlpersHuttenlocher syndrome has been shown to be caused by mutations in POLG.244 Alternatively, in mtDNA depletion syndrome (caused by mutations in the POLG, DGUOK, or MPV17 genes), hypoglycemia, acidosis, and liver failure develop early in infancy, and neurologic abnormalities are less prominent.242,245 Navaho neurohepatopathy has been shown to be caused by mtDNA depletion and a defect in the MPV17 gene product, which is involved in mtDNA maintenance and regulation of oxidative phosphorylation.242,246 Other multisystemic mitochondrial diseases with liver involvement are Pearson’s marrow-pancreas syndrome (caused by large deletions of mtDNA segments) and chronic diarrhea and intestinal pseudo-obstruction with liver involvement.242 Liver biopsy specimens in mitochondrial disorders typically show macrovesicular and microvesicular steatosis, with increased density and occasional swelling of mitochondria on electron microscopy. Immunohistochemical techniques are being used more frequently (e.g., to diagnose cytochrome c oxidase deficiency). Cholestasis may be present, and conditions associated with chronic liver disease can show micronodular cirrhosis. Lactic acidemia may be constant, intermittent, or absent in mitochondrial disorders.247 Direct measurement of the enzymatic activity of the respiratory chain electron transport protein complexes can be performed on frozen tissue from the organ

that expresses the clinical disease, although skin fibroblasts and lymphocytes may also be used. Few academic centers around the world can perform the assays for mitochondrial respiration (polarographic studies) or mtDNA analysis at this time. No known effective therapy has been developed for mitochondrial respiratory chain disorders that alters the course of disease. Several strategies have been proposed to delay the progression of such disorders, including the use of antioxidants such as vitamin E or ascorbic acid; electron acceptors and cofactors, such as coenzyme Q10, thiamine, or riboflavin; and agents proposed to work by other mechanisms, such as carnitine, creatine, or succinate supple­ mentation. The use of these agents, however, is clearly experimental at this time.242 Liver transplantation is generally contraindicated in these patients, given their dire outcome and high frequency of severe extrahepatic organ involvement.242,248

KEY REFERENCES

Alissa FT, Jaffe R, Shneider BL. Update on progressive familial intra­ hepatic cholestasis. J Pediatr Gastroenterol Nutr 2008; 46:241-52. (Ref 208.) Anderson KE, Bloomer JR, Bonkovsky HL, et al. Recommendations for the diagnosis and treatment of the acute porphyrias. Ann Intern Med 2005; 142:439-50. (Ref 92.) Campbell KM, Arya G, Ryckman FC, et al. High prevalence of alpha-1antitrypsin heterozygosity in children with chronic liver disease. J Pediatr Gastroenterol Nutr 2007; 44:99-103. (Ref 20.) Desmond CP, Wilson J, Bailey M, et al. The benign course of liver disease in adults with cystic fibrosis and the effect of ursodeoxycholic acid. Liver Int 2007; 27:1402-8. (Ref 238.) Eklund EA, Sun L, Westphal V, et al. Congenital disorder of glycosylation (CDG)-Ih patient with a severe hepato-intestinal phenotype and evolving central nervous system pathology. J Pediatr 2005; 147:84750. (Ref 87.) Jaeken J, Matthijs G. Congenital disorders of glycosylation: A rapidly expanding disease family. Annu Rev Genomics Hum Genet 2007; 8:261-78. (Ref 76.) Lee WS, Sokol RJ. Liver disease in mitochondrial disorders. Semin Liver Dis 2007; 27:259-73. (Ref 242.) Lindstedt S, Holme E, Lock EA, et al. Treatment of hereditary tyrosinaemia type I by inhibition of 4-hydroxyphenylpyruvate dioxygenase. Lancet 1992; 340:813-17. (Ref 145.) Liu C, Aronow BJ, Jegga AG, et al. Novel resequencing chip customized to diagnose mutations in patients with inherited syndromes of intrahepatic cholestasis. Gastroenterology 2007; 132:119-26. (Ref 33.) Shimozawa N. Molecular and clinical aspects of peroxisomal diseases. J Inherit Metab Dis 2007; 30:193-7. (Ref 198.) Summar ML, Dobbelaere D, Brusilow S, Lee B. Diagnosis, symptoms, frequency and mortality of 260 patients with urea cycle disorders from a 21-year, multicentre study of acute hyperammonaemic episodes. Acta Paediatr 2008; 97:1420-5. (Ref 158.) Sundaram SS, Bove KE, Lovell MA, Sokol RJ. Mechanisms of disease: Inborn errors of bile acid synthesis. Nat Clin Pract Gastroenterol Hepatol 2008; 5:456-68. (Ref 192.) Sveger T. Liver disease in alpha1-antitrypsin deficiency detected by screening of 200,000 infants. N Engl J Med 1976; 294:1316-21. (Ref 23.) Trauner M, Fickert P, Wagner M. MDR3 (ABCB4) defects: A paradigm for the genetics of adult cholestatic syndromes. Semin Liver Dis 2007; 27:77-98. (Ref 215.) Tuchman M, Lee B, Lichter-Konecki U, et al. Cross-sectional multicenter study of patients with urea cycle disorders in the United States. Mol Genet Metab 2008; 94:397-402. (Ref 159.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

77 Hepatitis A Maria H. Sjogren and Joseph G. Cheatham

CHAPTER OUTLINE Virology  1279 Epidemiology  1280 Pathogenesis  1281 Clinical Features  1281 Fulminant Hepatitis A  1282 Extrahepatic Manifestations  1282 Autoimmune Hepatitis after Acute Hepatitis A  1282

The development of liver biopsy techniques in the 1930s allowed the recognition of hepatic necroinflammation that characterizes all forms of viral hepatitis. Subsequent experi­ mental work in humans led to the clinical recognition that viruses are etiologic agents of hepatitis A (“infectious hepa­ titis”) and hepatitis B (“serum hepatitis”).1,2 Later, the exis­ tence of two hepatitis viruses was demonstrated—hepatitis A virus (HAV) and hepatitis B virus (HBV) (see Chapter 78).3 Additional viral causes of acute and chronic hepatitis were identified subsequently (see Chapters 79 to 81). HAV was first characterized in 1973, when scientists detected the virus in stools from human volunteers who were infected with HAV.4 The ensuing development of sensitive and specific serologic assays for the diagnosis of HAV infection and the isolation of HAV in cell culture5 were important advances that permitted understanding of the epidemiology of HAV infection and, ultimately, control of the disease.

VIROLOGY In 1982, HAV was classified as an enterovirus type 72 belonging to the Picornaviridae family. Subsequent deter­ mination of the sequence of HAV nucleotides and amino acids led to questioning of this classification, and a new genus, Hepatovirus, was created for HAV.6 HAV has an icosahedral shape and lacks an envelope. It measures 27 to 28 nm in diameter, has a buoyant density of 1.33 to 1.34 g/cm3 in cesium chloride, and has a sedimenta­ tion coefficient of 156 to 160S on ultracentrifugation. HAV survives exposure to ether and an acid environment at pH 3. It also survives heat exposure of 60°C for 60 minutes but is inactivated at 85°C for 1 minute. HAV is capable of sur­ viving in sea water (4% survival rate), in dried feces at room temperature for 4 weeks (17% survival), and in live oysters for 5 days (12% survival).7 HAV has only one known serotype, and no antigenic cross-reactivity with the hepatitis B, C, D, E, or G agents. The HAV genome consists of a positive-sense RNA that is 7.48 kb long, single-stranded, and linear (Fig. 77-1). HAV RNA has a sedimentation coefficient of 32 to 33S and a molecular weight of 2.8 × 10.4 The HAV RNA has a long open reading frame, consisting of 6681 nucleotides, and is

Diagnosis  1282 Prevention and Treatment  1283 Immunization against Hepatitis A Virus in Patients with Chronic Illnesses  1284

covalently linked to a 5′ terminal protein and a 3′ terminal polyadenosine tract. The onset of HAV replication in cell culture systems takes from weeks to months. Primate cells, including African green monkey kidney cells, primary human fibroblasts, human diploid cells (MRC-5), and fetal rhesus kidney cells, are favored for cultivation of HAV in vitro. The virus is not cytopathic, and persistent infection in the cell cultures is the rule. Two conditions control the outcome of HAV rep­ lication in cell culture.8 The first condition is the genetic makeup of the virus; HAV strains mutate in distinct regions of the viral genome as they become adapted to cell culture. The second condition is the metabolic activity of the host cell at the time of infection. Cells in culture, although infected simultaneously, initiate HAV replication in an asynchronous manner. This asynchronicity may be caused by differences in the metabolic activity of individual cells, but definitive evidence of cell-cycle dependence of HAV replication is lacking.9 An initial step in the life cycle of a virus is its attach­ ment to a cell surface receptor. The location and function of these receptors determine tissue tropism. Little is known about the mechanism of entry of HAV into cells. Some work has suggested that HAV could infect cells by a surrogate-receptor binding mechanism (involving a non­ specified serum protein). HAV infectivity in tissue culture has been shown to require calcium and to be inhibited by the treatment of the cells with trypsin, phospholipases, and b-galactosidase.10 A surface glycoprotein, named HAVcr-1, on African green monkey kidney cells has been identified as a receptor for HAV. Blocking of HAVcr-1 with specific monoclonal antibodies prevents infection of otherwise susceptible cells. Experimental data suggest that HAVcr-1 not only serves as an attachment receptor but also may facilitate uncoating of HAV and its entry into hepatocytes.11 Whatever the entry mechanism, once HAV enters a cell, the viral RNA is uncoated, cell host ribosomes bind to viral RNA, and polysomes are formed. HAV is translated into a large polyprotein of 2227 amino acids. This polyprotein is organized into three regions: P1, P2, and P3. The P1 region encodes structural proteins VP1, VP2, VP3, and a putative VP4. The P2 and P3 regions encode nonstructural proteins associated with viral replication (see Fig. 77-1).

1279

1280

Section IX  Liver HAV RNA Open reading frame

5'

3' AAA

VPg Translation

Noncoding

Noncoding

HAV polyprotein NH2

Nonstructural

Structural

P2

P1

P3

COOH

Proteins VP4?

3A (pre-VPg?)

2A

VP2 VP3

VP1

3B (VPg)

2B 2C

patients with HAV infection, including 12 with FHF.17 The investigators observed fewer nucleotide substitutions in the HAV genome from patients with FHF than in those from patients without FHF (P < .001). The differences were most prominent between nucleotides 200 and 500, suggesting that nucleotide variations in the central portion of the 5′ untranslated region influence the clinical severity of HAV infection.

3C (protease) 3D (RNA polymerase)

Figure 77-1.  Genomic organization of hepatitis A virus (HAV). VP, viral protein; VPg, 5′ terminal protein. (From Levine JE, Bull FG, Millward-Sadler GH, et al. Acute viral hepatitis. In: Millward-Sadler GH, Wright R, Arther MJP, editors. Wright’s Liver and Biliary Disease, 3rd ed. London: WB Saunders, 1992. p 679.)

The HAV RNA polymerase copies the plus RNA strand. The RNA transcript in turn is used for translation into pro­ teins, which are used for assembly into mature virions. Down-regulation of HAV RNA synthesis appears to occur as defective HAV particles appear.12 In addition, a group of specific RNA-binding proteins has been observed during persistent infection.13 The origin and nature of these pro­ teins is unknown, but they exert activity on the RNA tem­ plate and are believed to play a regulatory role in the replication of HAV.14 Numerous strains of HAV exist, with considerable nucle­ otide sequence variability (15% to 25% difference within the P1 region of the genome). Human HAV strains can be grouped into four different genotypes (I, II, III, and VII), whereas simian strains of HAV belong to genotypes IV, V, and VI.15 Despite the nucleotide sequence heterogeneity, the antigenic structure of human HAV is highly conserved among strains. The HAV VP1/2A and 2C genes are thought to be respon­ sible for viral virulence. This conclusion is based on experi­ ments in which the genotypes and phenotypes of viruses were compared after animals were infected with one of 14 chimeric virus genomes derived from two infectious cDNA clones that encoded a virulent HAV isolate and an attenu­ ated HAV isolate (HM175 strain), respectively.16 Among the many strains of HAV, the HM175 and CR326 human HAV strains are important because they were used for production of commercially available vaccines. Strain HM175 was isolated in 1978, from human feces from Aus­ tralian patients in a small outbreak of hepatitis A. CR326 was isolated from Costa Rican patients infected with HAV. The nucleotide and amino acid sequences showed 95% identity between the two strains. Vaccines prepared from these strains are thought to provide protection against all relevant human strains of HAV. Variations in the HAV genome are thought to play a role in the development of fulminant hepatic failure (FHF) during acute HAV infection. The 5′ untranslated region of the HAV genome was sequenced in serum samples from 84

EPIDEMIOLOGY Acute hepatitis A is a reportable infectious disease in the United States. The incidence has declined by 90% since 1995. In 2006, 3579 cases of HAV infection were reported, corresponding to a rate of infection of 1.2 per 100,000, com­ pared with 4 per 100,000 in 2001. With the underreporting of cases and the occurrence of asymptomatic infections taken into consideration, the true number of HAV infections in 2006 was calculated to be 32,000, compared with 93,000 in 2001. The greatest rate of decline has been among chil­ dren from states where routine vaccination of children was recommended in 1999. The highest rate of reported disease historically has been among children ages 5 to 14 years. Because of the rapid rate of decline of disease in children, however, rates are now similar among age groups, with adults ages 20 to 44 having the highest rate of disease in 2006.18 The epidemiologic risk factors for HAV infection reported for the U.S. population in 2006 were as follows: unknown, 65%; international travel, 15%; contact with a patient who has hepatitis, 12%; sexual or household contact with a patient who has hepatitis A, 10%; men having sex with men, 9%; food or waterborne outbreak, 7%; child or employee in a daycare center, 4%; contact with a daycare child or employee, 4%; and injection drug use, 2%.18 HAV infection generally follows one of three epidemio­ logic patterns.19 In countries where sanitary conditions are poor, most children are infected at an early age. Although earlier seroepidemiologic studies routinely showed that 100% of preschool children in these countries had detect­ able antibody to HAV (anti-HAV) in serum, presumably reflecting previous subclinical infection, subsequent studies have shown that the average age of infection has risen rapidly to 5 years and older, when symptomatic infection is more likely. For example, 82% of 1393 Bolivian school children were shown to have detectable anti-HAV, but when they were stratified into two groups according to family income, a significant difference was found between the groups: 95% of children from low-income families, but only 56% of children from high-income families, had detectable anti-HAV.20 The second epidemiologic pattern is seen in industrial­ ized countries, where the prevalence of HAV infection is low among children and young adults. In the United States, prior to universal HAV vaccination, the prevalence of antiHAV was approximately 10% in children but 37% in adults.21 The third epidemiologic pattern is observed in closed or semiclosed communities, such as some isolated com­ munities in the South Pacific, in which HAV is capable (through epidemics) of infecting the entire population, which then becomes immune. Thereafter, newborns remain susceptible until the virus is reintroduced into the community. Whatever the epidemiologic pattern, the primary route of transmission of HAV is the fecal-oral route, by either

Chapter 77  Hepatitis A Table 77-1  Detection of Hepatitis A Virus (HAV) and Infectivity of Human Secretions or Excretions SECRETION/EXCRETION

COMMENT

REFERENCEs

Stool

Main source of infection. HAV is detectable during the incubation period and for several weeks after the onset of disease. After the onset of symptoms, HAV is detectable in 45% and 11% of fecal specimens collected during the first and second weeks of illness, respectively, whereas HAV RNA (by polymerase chain reaction assay) is detectable for 4 to 5 months. Viremia is present during the incubation period. Blood collected 3 and 11 days before the onset of symptoms has caused post-transfusion infection in recipients. Chronic viremia does not occur. HAV has been detected in the bile of chimpanzees infected with HAV. HAV is detected in low titer during the viremic phase. A urine sample was reported to be infectious after oral inoculation. Urine contaminated with blood was also infectious. Unknown in humans. HAV has been identified in the oropharynx of experimentally infected chimpanzees. Uncertain. HAV may be detectable during the viremic phase.

26,27

Blood

Bile Urine

Nasopharyngeal secretions Semen, vaginal fluid

person-to-person contact or ingestion of contaminated food or water. Although rare, transmission of HAV by a paren­ teral route has been documented after transfusion of blood22,23 or blood products.24 Cyclical outbreaks among users of injection and noninjection illicit drugs and among men who have sex with men (up to 10% may become infected in outbreak years) have been reported.25 Table 77-1 provides information about the detection of HAV and its infectivity in human body fluids.26-33 From 11% to 22% of patients with acute hepatitis A require hospitalization, with an average length of stay of 4.6 days, costing on average $7926 per patient in 2004. In one outbreak involving 43 persons, the total cost was approxi­ mately $800,000. On average, 27 workdays are lost per adult case of hepatitis A. In adolescents and adults, the combined direct and indirect costs associated with HAV infection in the United States totaled approximately $488.8 million in 1997, compared with $93 million in 2006.25,34,35 The decline in costs is a direct result of the dramatic reduction in the number of infections seen since the introduction of the HAV vaccine combined with changes in vaccination policies in the United States (see later).18

PATHOGENESIS After HAV is ingested and survives gastric acid, it traverses the small intestinal mucosa and reaches the liver via the portal vein. The precise mechanism of hepatic uptake in humans is unknown (see earlier). In an experimental model using African green monkey kidney cells,11 the putative cel­ lular receptor for HAV has been identified as a surface glycoprotein. Once the virus enters the hepatocyte, it starts replicating in the cytoplasm, where it is seen on electron microscopy as a fine granular pattern, but it is not present

28,29

30 31,32

33 —

in the nucleus. HAV is distributed throughout the liver. Although HAV antigen has been detected in other organs (lymph nodes, spleen, kidney), the virus appears to repli­ cate exclusively in hepatocytes. When the virus is mature, it reaches the systemic circulation via the hepatic sinusoids and is released into the biliary tree through bile canaliculi, passed into the small intestine, and eventually excreted in the feces. The pathogenesis of HAV-associated hepatocyte injury is not completely defined. The lack of injury to cells in cell culture systems suggests that HAV is not cytopathic. Immunologically mediated cell damage is more likely. The emergence of anti-HAV could result in hepatic necrosis during immunologically mediated elimination of HAV.

CLINICAL FEATURES Infection with HAV does not result in chronic infection, only in an acute self-limited episode of hepatitis. Rarely, acute hepatitis A can have a prolonged or a relapsing course, and occasionally profound cholestasis can occur.36 The incubation period is commonly 2 to 4 weeks, rarely up to 6 weeks. The mortality rate is low in previously healthy persons. Morbidity can be significant in adults and older children. The clinical characteristics of cases of hepatitis A reported in 2002 were similar to those in previous years, with a pre­ ponderance of cases in men in all age groups. Overall, 72% of patients manifested jaundice, 25% required hospitaliza­ tion, and 0.5 % died.37 The need for hospitalization rose with age, from 5% among children younger than 5 years to 34% among persons 60 years or older.37 Patients with HAV infection usually present with one of the following five clinical patterns: (1) asymptomatic

1281

1282

Section IX  Liver without jaundice, (2) symptomatic with jaundice and selflimited after approximately 8 weeks, (3) cholestatic, with jaundice lasting 10 weeks or more,36 (4) relapsing, with two or more bouts of acute HAV infection occurring over a 6- to 10-week period, and (5) FHF. Children younger than 2 years are usually asymptomatic; jaundice develops in only 20% of them, whereas symptoms develop in most children (80%) 5 years or older. The rate of symptoms is high in adolescents and adults. HAV infection with prolonged cholestasis is a rare variant but occasionally leads to invasive diagnostic procedures (inappropriately) because the diagnosis of acute hepatitis may not be readily accepted in patients who have jaundice for several months, even in the presence of detectable anti-HAV of the immunoglobulin M (IgM) class (see later).36 A relapsing course is observed in approximately 10% of patients with acute hepatitis A. Shedding of HAV in stool has been documented during the relapse phase.38 This variant is benign, and the infection ultimately resolves.38 Neither the cholestatic variant nor relapsing hepatitis A is associated with an increase in mortality. In all cases, treat­ ment is symptomatic. Acute hepatitis A, unlike hepatitis E, is not associated with a higher mortality rate in pregnant women. Prodromal symptoms in patients with acute hepatitis A include fatigue, weakness, anorexia, nausea, vomiting, and abdominal pain. Less common symptoms are fever, head­ ache, arthralgias, myalgias, and diarrhea. Dark urine pre­ cedes other symptoms in approximately 90% of infected persons; this symptom occurs within 1 to 2 weeks of the onset of prodromal symptoms. Symptoms of hepatitis may last from a few days to 2 weeks and usually decrease with the onset of clinical jaundice. Right upper quadrant tender­ ness and mild liver enlargement are found on physical examination in 85% of patients; splenomegaly and cervical lymphadenopathy are each present in 15%. Complete clinical recovery is achieved in 60% of affected persons within 2 months and in almost everyone by 6 months. The overall prognosis of acute hepatitis A in otherwise healthy adults is excellent. Potentially fatal complications (e.g., FHF) develop in a few patients (see later).39 Acute HAV infection must be differentiated by appro­ priate serologic testing from other causes of acute viral hepatitis, autoimmune hepatitis, and other causes of acute hepatitis (see Chapters 20 and 73). In some cases the diag­ nosis may be difficult to make because the patient may harbor another viral infection, such as chronic hepatitis B or chronic hepatitis C, with superimposed acute HAV infection.

FULMINANT HEPATITIS A

FHF due to HAV is rarely seen in children, adolescents, or young adults. The case-fatality rate in people older than 49 years with acute hepatitis A is reported to be 1.8%, com­ pared with an overall rate of 0.3% in persons of all ages.37 Hepatic failure caused by hepatitis A becomes manifest in the first week of illness in about 55% of affected patients and during the first 4 weeks in 90%; the onset of FHF rarely occurs after 4 weeks of illness.39 The contribution of HAV to acute liver failure has been reported to be greater in populations classified as hyperen­ demic for HAV. In a report from India, where 276 patients with FHF were seen between 1994 and 1997, 10.6% of the cases among adults were caused by HAV. HAV had been responsible for only 3.5% of cases among 206 patients with FHF seen in the same community from 1978 to 1981.40 Certain populations have increased morbidity and a high risk of acute liver failure from HAV infection. Among these

groups are the elderly41 and persons with chronic liver disease and HIV infection. A 1998 report described the clinical outcome of 256 persons hospitalized for acute hepa­ titis A in Tennessee from January 1994 through December 1995.42 On admission, 89% of the patients had experienced sustained nausea or vomiting, and 26% had a prolonged prothrombin time (>3 seconds); 39 had serious complica­ tions (19 hepatobiliary and 20 extrahepatic complications), and 5 (2%) died. Morbidity and mortality correlated with age. Twenty-five percent of patients 40 years and older had at least one complication, compared with 11% of patients younger than 40 years (P = .014). Although two reports since the late 1990s have described a decline in the number of cases of acute viral hepatitis among patients with FHF in the United States,43,44 this decline is attributable principally to the control of hepatitis B. The contribution of HAV infection to FHF has remained unchanged since the 1970s despite the availability of highly efficacious vaccines (see also Chapter 90).

EXTRAHEPATIC MANIFESTATIONS

Extrahepatic manifestations are less frequent in acute HAV infection than in acute HBV infection and consist most com­ monly of an evanescent rash (14%) and arthralgias (11%) and uncommonly of leukocytoclastic vasculitis, glomerulo­ nephritis, and arthritis, in which immune-complex disease is believed to play a pathogenic role. Cutaneous vasculitis is typically seen on the legs and buttocks; skin biopsies reveal the presence of IgM anti-HAV and complement in the blood vessel walls. The arthritis also appears to have a predilection for the lower extremities. Both vasculitis and arthritis have been associated with cryoglobulinemia, although cryoglobulinemia, in general, is more frequently associated with hepatitis C virus (HCV) infection. The cryoglobulin has been shown to contain IgM anti-HAV. Other rare extrahepatic manifestations that may be immunecomplex related include toxic epidermal necrolysis, fatal myocar­ditis, renal failure in the absence of liver failure, optic neuritis, transverse myelitis, polyneuritis, and chole­ cystitis. Hematologic complications include thrombocyto­ penia, aplastic anemia, and red-cell aplasia. Patients with more protracted illness appear to have a higher frequency of extrahepatic manifestations.

AUTOIMMUNE HEPATITIS AFTER ACUTE HEPATITIS A

Several viruses have been reported to trigger the onset of autoimmune hepatitis (AIH). In rare cases, hepatitis A has been followed by the development of type 1 AIH. Genetic predisposition is thought to play a role (see Chapter 88).45,46

DIAGNOSIS Acute hepatitis A is clinically indistinguishable from other forms of viral hepatitis. The diagnosis of infection is based on the detection of specific antibodies against HAV (antiHAV) in serum (Fig. 77-2). A diagnosis of acute hepatitis A requires demonstration of IgM anti-HAV in serum. The test result is positive from the onset of symptoms47 and usually remains positive for approximately 4 months.48 Some patients may have low levels of detectable IgM anti-HAV for more than a year after the initial infection.48 IgG antiHAV is also detectable at the onset of the disease, remains present usually for life, and, after clinical recovery, is interpreted as a marker of previous HAV infection (as

Chapter 77  Hepatitis A Table 77-2  Groups at High Risk of Hepatitis A Virus Infection

Jaundice symptoms

Healthy persons who travel to endemic areas, work in occupations for which the likelihood of exposure is high, are family members of infected patients, or adopt infants or children from endemic areas Men who have sex with men Persons who have tested positive for human immunodeficiency virus Persons with chronic liver disease Persons with clotting factor disorders Users of injection and noninjection illicit drugs

Anti-HAV

ALT HAV in feces

0

1

IgM Anti-HAV

2

3

4

5

6

12

14

Months after exposure Figure 77-2.  Typical course of a case of acute hepatitis A. ALT, serum alanine aminotransferase level; Anti-HAV, antibody to hepatitis A virus; HAV, hepatitis A virus; IgM, immunoglobulin M. (From Hoofnagle JH, DiBisceglie AM. Serologic diagnosis of acute and chronic viral hepatitis. Semin Liver Dis 1991; 11:73-83.)

demonstrated by a positive result on a commercial assay for total anti-HAV and negative result for IgM anti-HAV). Testing for HAV RNA is limited to research laboratories. HAV RNA has been detected in serum, stool, and liver tissue. Viral RNA can be amplified by polymerase chain reaction (PCR) methodology.49 With a PCR assay, HAV RNA has been documented in human sera for up to 21 days after the onset of illness.50 The use of HCV RNA testing has been described in a report of 76 French patients with acute HAV infection seen between January 1987 and April 2000; 19 had FHF,51 10 of whom required liver transplantation and 1 of whom died while awaiting liver transplantation. The HAV RNA status was determined in 39 of the 50 patients in whom sera and clinical data were available, including the 19 with FHF. HAV RNA was detected in 36 of these 50 patients (72%). The likelihood that HAV RNA was unde­ tectable was greater in patients with FHF than in those with nonfulminant hepatitis (P < .02). Of those in whom HAV RNA was detectable, titers were lower in patients with encephalopathy than in patients with nonfulminant hepati­ tis (3.6 log vs 4.4 log; P = .02). These data suggest that detection of IgM anti-HAV coupled with nondetection or finding of low-titer HAV RNA in patients with severe acute hepatitis may signal an ominous prognosis and the need for early referral for liver transplantation. As in other studies, HAV genotype did not seem to play a role in the severity of clinical manifestations.52

PREVENTION AND TREATMENT Recommendations concerning immunoprophylaxis against HAV were published in December 1999 by the Advisory Committee on Immunization Practices (ACIP).25 The overall strategy was to protect persons from disease and to lower the incidence of HAV infection in the United States. The available monovalent vaccines were initially licensed for use in children greater than 2 years of age but are now licensed for use after 12 months of age.25,53 In accordance

with the 1999 ACIP guidelines, high-risk populations were targeted for immunization. After achieving a dramatic reduction in incidence rates, a universal childhood vaccina­ tion policy was adopted in 2006, with the hope of eventu­ ally eliminating indigenous HAV transmission in the United States. The decline in incidence rates not surprisingly has been greater in children than in adults, effectively removing children as a high-risk population and potentially removing the primary reservoir for the virus in the United States.18,25 Table 77-2 lists the populations now considered to be at highest risk of HAV infection. No specific medications are available to treat acute hepa­ titis A; symptomatic treatment is the rule. Attention to sani­ tation and administration of serum immune globulin (IG) have been the mainstays of preventing HAV infection. The availability of excellent HAV vaccines has rendered use of IG for pre-exposure prophylaxis unnecessary. Furthermore, in June 2007, the HAV vaccine was approved for use in postexposure prophylaxis of immunocompetent persons, ages 12 months to 40 years, without chronic liver disease.54 This new indication for the HAV vaccine was based on the results of a study that compared the efficacy of the HAV vaccine with that of IG for postexposure prophylaxis against HAV infection. The rates of clinical infection were low in both groups, with clinical hepatitis A developing in 4.4% of subjects in the vaccine group compared with 3.3% of those in the IG group.55 This study, however, likely excluded persons with asymptomatic infection. In the vaccine group, 162 persons with IgM HAV in serum were excluded, com­ pared with 50 persons in the IG group because of either a lack of symptoms or absence of an elevated serum alanine aminotransferase level of at least two times the upper limit of normal. The possibility exists that a number of persons with asymptomatic hepatitis A still posed an infectious risk to others. Taking into account data from Canada and the United Kingdom, where the HAV vaccine has been used for post­ exposure prophylaxis since the early 2000s, the ACIP con­ cluded the HAV vaccine is safe and comparable to IG in protecting recipients against clinical hepatitis A. The ACIP guidelines allow persons who have recently been exposed to HAV and who have not been vaccinated previously to be given a single dose of single-antigen HAV vaccine or IG (0.02 mL/kg) as soon as possible, within 2 weeks of expo­ sure. Some of the benefits of the vaccine include long-term immunity if a second dose of the vaccine is administered in accordance with the standard vaccine schedule (see Table 77-3), as well as cost savings and wide availability compared with IG.54 Although IG is considered safe, the perception is widespread that it poses a risk because it is a blood-derived product. IG can cause fever and myalgias just as the vaccine can, but pain at the injection site is usually more pronounced with IG than with the vaccine. Postexposure prophylaxis with IG can be administered at

1283

1284

Section IX  Liver Table 77-3  Recommended Regimens for Hepatitis A Vaccination* VACCINE

SCHEDULE

AGE (YR)

DOSE

VOLUME (mL)

DOSING SCHEDULE

HAVRIX

Standard Standard Accelerated Postexposure prophylaxis Standard Standard Accelerated Postexposure prophylaxis†† Standard Accelerated

1-18 >18 ≥1 ≥1 1-18 >18 ≥1 ≥1 ≥18 ≥18

720 ELU 1440 ELU Single age-appropriate dose Single age-appropriate dose 25 U 50 U Single age-appropriate dose Single age-appropriate dose 720 ELU HAV, 20 mg HBV 720 ELU HAV, 20 mg HBV

0.5 1.0 Age appropriate Age appropriate 0.5 1.0 Age appropriate Age appropriate 1.0 1.0

0, 6-12 months 0, 6-12 months ≥2 weeks prior to travel† <2 weeks after exposure† 0, 6-18 months 0, 6-18 months ≥2 weeks prior to travel† <2 weeks after exposure† 0, 1, 6 months 0, 7, 21-30 days†

VAQTA

TWINRIX

ELU, enzyme-linked immunoassay (ELISA) units; HBV, hepatitis B virus; U, units. *Vaccines are injected intramuscularly into the deltoid muscle. † Timing of booster dose (necessary for long-term protection): VAQTA, 6 mos; HAVRIX, 6-12 mos; TWINRIX, 12 mos. †† Not approved by the U.S. Food and Drug Administration.

the same time as initiation of active immunization with the vaccine.56 The HAV vaccine was first licensed in the United States in 1995; two inactivated HAV vaccines are commercially available. Extensive use of the vaccines in clinical trials and postmarketing surveillance support the safety and efficacy of these products. HAVRIX is manufactured by SmithKline Biologicals, Rixensart, Belgium, and VAQTA by Merck Sharp & Dohme, West Point, Pennsylvania. Both vaccines are derived from HAV grown in cell culture. The final prod­ ucts are purified and formalin-inactivated; they contain alum as an adjuvant. The basic difference between the two commercially available vaccines is the HAV strain used for preparation. HAVRIX was prepared with the HM175 strain, whereas VAQTA was prepared with the CR326 strain.57,58 The difference is of little practical importance because both vaccines are safe and immunogenic. The doses and schedule of immunization are shown in Table 77-3. After vaccination with HAVRIX, anti-HAV is estimated to remain detectable in serum for approximately 20 years; immunity may last longer.59 From the time the HAV vaccine was licensed in the United States through 2005, more than 50 million doses were administered. Worldwide, more than 188 million doses of HAV vaccine were administered through 2005. Among adults, the most common local side effects have been soreness at the injection site (56%), head­ ache (14%), and malaise (7%). In children, the most common side effects have been soreness at the injection site (15%), feeding pro­blems (8%), headache (4%), and induration at the injection site (4%).25 In the United States, through 2005, the Vaccine Adverse Event Reporting System received 6136 reports of unex­ plained adverse events after immunization with the HAV vaccine alone or in combination with other vaccines. Of the 6136 reports, 871 were considered serious and included Guillian-Barré syndrome, immune thrombocyto­ penic purpura, elevated serum aminotransferase levels, and seizures in children.25 No reported serious event, however, could be attributed definitively to the HAV vaccine, and the reported rates did not exceed the expected background rates. For example, the general population incidence of Guillain-Barré syndrome ranges from 0.5 to 2.4 cases per 100,000 person-years, and among adult HAV vaccine recipi­ ents the incidence of Guillain-Barré was 0.2 cases per 100,000 person-years.25 A combined formulation of hepatitis A and B vaccines (TWINRIX) is available and has an excellent record of effi­ cacy and safety.60 Although some long-term studies have

shown persistence of anti-HAV in children and adolescents, seroconversion rates for TWINRIX are apparently lower in children ages 1 to 6 years than those for standard monova­ lent vaccines.61 Currently, therefore, TWINRIX is approved only for persons 18 years of age and older. As a result of the reduction in endemic cases of hepatitis A in the United States, the largest proportion of patients who now become infected with HAV are non-immune adults traveling to endemic areas. Even if medical advice is sought before travel, the time is usually insufficient for completing the standard immunization schedule. HAVRIX and VAQTA are approved by the U.S. Food and Drug Administration (FDA) for use in an accelerated vaccination schedule before planned travel. If given at least 2 weeks before travel, a single dose of either monovalent vaccine results in protective anti-HAV titers.54 In 2008, the FDA also approved an accelerated vaccination schedule for TWINRIX that can be completed within 30 days, with a booster at 12 months, after studies showed equivalent protection when TWINRIX was compared with standard and alternative schedules of the individual monovalent vaccines. After 1 year, HAV seroconversion rates were 100%, and HBV sero­ conversion rates were 96.4% to 100% with TWINRIX.62,63 The TWINRIX accelerated schedule is being considered for use in new inmates at U.S. correctional facilities, where high-risk activities place the inmates at risk for both HAV and HBV infections.64 The dosing schedules are shown in Table 77-3.

IMMUNIZATION AGAINST HEPATITIS A VIRUS IN PATIENTS WITH CHRONIC ILLNESSES

Persons with chronic liver disease are at increased risk of HAV-related morbidity and mortality if they acquire the infection. Therefore, pre-exposure prophylaxis with the HAV vaccine has been recommended for patients with chronic liver disease who are susceptible to HAV.65 This recommendation should be extended to patients awaiting liver transplantation as well as those who have already undergone liver transplantation, although the immunoge­ nicity of the HAV vaccine is reduced in such persons.66 An episode of acute hepatitis in a patient with underlying chronic liver disease poses the risk of considerable mor­ bidity and mortality. Although the current guidelines recommend immunization against HAV for all patients with chronic liver disease,25 the results of several costeffectiveness analyses have been conflicting. A report pub­ lished in 2000 found that saving the life of one patient with

Chapter 77  Hepatitis A HCV infection through HAV vaccination would cost 23 million Canadian dollars,67 although some of the assump­ tions in this report have been challenged.68 Two other studies of patients with chronic hepatitis C showed a decided benefit to immunization against HAV.69,70 The methods used in these studies were dissimilar, and some analyses may have been insensitive to the incidence of HAV or may have underestimated the economic and societal costs of a case of FHF. Universal immunization against HAV during childhood, before the possible occurrence of chronic liver disease, offers the greatest promise of prevent­ ing HAV infection.71 Patients infected with the human immunodeficiency virus (HIV) should be vaccinated against HAV. The response to vaccination, however, may be reduced because of a blunted immune system. Earlier studies suggested sero­ positivity rates above 97% in HIV-infected children on antiretroviral therapy72; however, a more recent study found that a CD4+ count of less than 25/mm3 and an HIV viral load of more than 400 copies/mL predicted a reduced serocon­ version rate.73 Although the discrepancies among studies can be explained in part by different sensitivities of assays for anti-HAV, it appears that the more immunosuppressed a person is, the less likely the person is to respond to vac­ cination. In this population, consideration should be given to checking post-vaccination IgG anti-HAV titers to assess immunity. Small studies have shown limited additional benefit to a third dose of HAV vaccine in persons who fail to respond to the standard vaccine schedule.73

KEY REFERENCES

American Academy of Pediatrics, Committee on Infectious Disease. Hepatitis A vaccine recommendations. Pediatrics 2007;120:189-99. (Ref 53.)

Centers for Disease Control and Prevention. Hepatitis Surveillance Report No. 59, Atlanta, CDC, 2004. (Ref 37.) Centers for Disease Control and Prevention. Prevention of hepatitis A thorough active or passive immunization. MMWR 2006;55(No. RR07):1-23. (Ref 25.) Centers of Disease Control and Prevention: Prevention of hepatitis A after exposure to hepatitis A virus and in international travelers. MMWR 2007;56:1080-4. (Ref 54.) Centers for Disease Control and Prevention. Surveillance for Acute Viral Hepatitis, United States-2006. MMWR 2008;57(SS02);1-24. (Ref 18.) Connor BA, Blatter MM, Beran J, et al. Rapid and sustained immune response against hepatitis A and B achieved with combined vaccine using an accelerated administration schedule. J Travel Med 2007;14:915. (Ref 63.) FDA approval for a combined hepatitis A and B vaccine. MMWR 2001;50:806. (Ref 60.) Feinstone SM, Kapikian AZ, Purcell RH. Hepatitis A: Detection by immune electron microscopy of a viruslike antigen associated with acute illness. Science 1973;182:1026-8. (Ref 4.) Krugman S, Ward R, Giles JP, et al. Infectious hepatitis: Detection of virus during the incubation period and in clinically inapparent infec­ tions. N Engl J Med 1959;261:729-34. (Ref 3.) Mathiesen LR, Feinstone SM, Purcell RH, Wagner JA. Detection of hepatitis A antigen by immunofluorescence. Infect Immun 1977;18:524-30. (Ref 15.) Nothdurft HD, Dietrich M, Zuckerman JN, et al. A new accelerated vac­ cination schedule for rapid protection against hepatitis A and B. Vaccine 2002;20:1157-62. (Ref 62.) Rezende G, Roque-Alsonso M, Samuel D, et al. Viral and clinical factors associated with fulminant course of hepatitis A infection. Hepatology 2003;38:613-18. (Ref 51.) Rosenthal P. Cost-effectiveness of hepatitis A vaccination in children, adolescents and adults. Hepatology 2003;37:44-51. (Ref 71.) Victor JC, Monto AS, Surdina TY, et al. Hepatitis A vaccine versus immune globulin for postexposure prophylaxis. N Engl J Med 2007;357:1685-94. (Ref 55.) Zhou F, Shefer A, Weinbaum C, et al. Impact of hepatitis A vaccination on health care utilization in the United States, 1996-2004. Vaccine 2007;25:3581-7. (Ref 34.) Full references for this chapter can be found on www.expertconsult.com.

1285

CHAPTE R

78 Hepatitis B and D Robert Perrillo

CHAPTER OUTLINE Hepatitis B  1287 Epidemiology  1287 Clinical Outcomes  1288 Molecular Biology  1289 Pathogenesis  1292 Natural History  1293 Clinical Features  1294 Diagnosis  1298 Treatment  1300 Prevention  1307

HEPATITIS B An estimated 400 million persons are carriers of hepatitis B virus (HBV) in the world today; of these, 75% reside in Asia and the Western Pacific. Effective vaccines against HBV have been available since the early 1980s, but perinatal and early life exposures continue to be major sources of infection in high-prevalence areas. High-risk behaviors such as promiscuous heterosexual contact and injection drug use account for many new cases in young adults. Fulminant acute hepatitis B accounts for several hundred deaths per year in the United States, and chronic HBV infection accounts for one million deaths worldwide each year from complications of end-stage liver disease, including hepatocellular carcinoma (HCC). Hepatitis B is the chief cause of cirrhosis and HCC in the world today, and nationwide vaccination has been shown to diminish greatly the number of new cases of infection and HCC in Taiwanese children.1 Universal hepatitis B vaccination is likely to have the greatest impact on liver disease–related mortality in future generations.

EPIDEMIOLOGY Geographic Distribution and Sources of Exposure

The prevalence of hepatitis B varies markedly around the world. In highly endemic regions, such as Southeast Asia (excluding Japan), China, and much of Africa, 8% or more of the population are chronic HBV carriers, and the lifetime risk of infection ranges from 60% to 80%.2 In these areas, perinatal transmission and horizontal spread among children are the major sources of infection. Approximately 60% of the world’s population reside in these areas where HBV is highly endemic.3 Regions of intermediate risk include parts of southern and eastern Europe, the Middle East, Japan, the Indian subcontinent, much of the former Soviet Union, and northern Africa. In intermediate-risk areas, the lifetime risk of infection is between 20% and 60%. Persons of all age groups are infected, but as in high-risk areas, most infections occur during infancy or early childhood. Regions

Hepatitis D  1309 Epidemiology  1309 Virology  1310 Pathogenesis  1310 Diagnosis  1310 Natural History  1310 Clinical Features  1311 Treatment  1311 Prevention  1312

of low prevalence are North America, western Europe, certain parts of South America, and Australia. In these areas, the lifetime risk of HBV infection is less than 20%, and transmission is primarily horizontal (i.e., between young adults). Sexual transmission is the main mode of transmission in Europe and North America, and injection drug use is a major contributor to new cases as well.4 Transmission of infection from an HBV carrier mother to her neonate accounts for the majority of new infections in the world today. Sixty to ninety percent of hepatitis B surface antigen (HBsAg)-positive mothers who are hepatitis B e antigen (HBeAg)-positive transmit the disease to their offspring, whereas mothers who are positive for antibody to HBeAg (anti-HBe) transmit the disease less frequently (15% to 20%) (see later discussion of diagnosis). Other less common sources of infection are household contact with an HBV carrier, hemodialysis, exposure to infected health care workers, tattooing, body piercing, artificial insemination, and receipt of blood products or organs. Since routine screening of the blood supply was implemented in the early 1970s, transfusion-associated hepatitis B has become rare in the United States. Hepatitis B can be transmitted by blood that tests negative for HBsAg but positive for antibody to hepatitis B core antigen (anti-HBc) because of low levels of circulating HBV DNA in such blood.5 HBsAg-negative blood that is positive for anti-HBc is excluded from the donor pool in the United States and many countries around the world. In 0% to 30% of persons who are seropositive for anti-HBc alone, HBV DNA is detectable in serum by polymerase chain reaction (PCR) testing.6 HBV is transmitted efficiently by percutaneous and mucous membrane exposure to infectious body fluids. The virus is 100 times as infectious as human immunodeficiency virus (HIV) and 10 times as infectious as hepatitis C virus (HCV). HBeAg seropositivity indicates a higher risk of transmission from mother to child, after needlestick exposure, and in the setting of household contact. HBV DNA has been detected by sensitive techniques such as PCR testing in most body fluids, except for stool that has not been contaminated with blood. Although HBV replicates primarily in hepato-

1287

1288

Section IX  Liver cytes, the presence of replicative intermediates and virally encoded proteins in other sites, such as the adrenal gland, testis, colon, nerve ganglia, and skin, suggests that a vast extrahepatic reservoir for infectious virus exists.7 Small amounts of HBV DNA have been demonstrated in peripheral mononuclear cells and liver tissue years after apparent resolution of chronic infection.8,9 Extrahepatic localization of low levels of replicating virus explains the relatively high rate of transmission of infection from anti-HBc–positive organ donors.10

liver disease and do not experience complications. Some of these patients, however, ultimately have one or more episodes of reactivated hepatitis in which the levels of viremia and serum aminotransferase activity increase. Also, some patients with the inactive carrier state may demonstrate HCC. Active carriers, on the other hand, have evidence of HBV replication on non–PCR-based assays for HBV DNA, intermittently or persistently elevated serum aminotransferase levels, and evidence of chronic hepatitis in a liver biopsy specimen.

Rates of Infection in the United States

Clinical Sequelae of Acute Hepatitis B Virus Infection

The incidence of hepatitis B has been declining in the United States over the past decade because of universal vaccination of newborns, adult vaccination programs for high risk persons, changes in sexual lifestyle, refinements in blood screening procedures, and the availability of virusinactivated blood components.11 Most striking has been the decrease among children and health care workers, groups with the highest rates of vaccination. Data from the Centers for Disease Control and Prevention (CDC) indicate that more than 95% of pregnant women in the United States are tested for HBsAg, and infant vaccine coverage levels are now equivalent to those of other vaccines in the childhood schedule. Nonetheless, an estimated 78,000 new HBV infections occurred in 2001, with the highest incidence rates among sexually active young adults (20 to 29 years old) and higher rates occurring among black and Hispanic persons than in white persons.12 Since 1995, approximately 40% of cases of acute hepatitis B reported to the CDC were caused by intimate contact among heterosexuals, 15% to 20% were related to intravenous drug use, and 12% occurred in men who have sex with men. No identifiable source of exposure was demonstrated in approximately 15% of cases. Nearly one third of prison inmates have been infected with hepatitis B, and 2% are chronically infected. According to the third National Health and Nutrition Examination Survey (1988 to 1994), one or more serologic markers of HBV infection were demonstrated in 4.9% of the U.S. population, and the prevalence of chronic infection was 0.2%.13 Traditional estimates based on the results of blood donation screening in the late 1970s also indicated a prevalence rate for chronic infection of 0.2% to 0.4% in the United States. Although traditional estimates are that the number of HBV carriers in the United States is between 1.25 and 1.5 million, this figure is likely to be a serious underestimate because of changing immigration patterns and underrepresentation of certain minority groups in field surveys. For example, 15 million Asians live in the United States, and even a conservative estimate that the prevalence in this group is 5% would raise the overall number of HBV carriers in the United States by more than 750,000.

CLINICAL OUTCOMES Definitions

In common usage, the term HBV carrier has often been used to refer to persons persistently infected with HBV who have normal serum aminotransferase levels. (They are sometimes inappropriately referred to as “healthy” HBV carriers.) Because the nomenclature is potentially confusing, the proposal has been made that the carrier state be categorized as inactive or active. Inactive carriers are patients who have evidence of HBV replication on a PCR-based assay only (but not with a less sensitive non–PCR-based assay) and normal or only mildly elevated serum aminotransferase values (see later).14 Long-term follow-up of inactive carriers suggests that the majority of these patients do not have progressive

The age at which a person becomes infected with HBV is a principal determinant of the clinical outcome. HBV infection in adults with an intact immune system is likely to cause clinically apparent acute hepatitis B; only 1% to 5% of these persons become chronically infected.4 By contrast, as many as 95% of infected neonates become chronic HBV carriers because of immunologic tolerance to the virus. In adults, fulminant liver failure caused by acute hepatitis B occurs in less than 1% of cases, but this group still accounts for 5% of all cases of acute liver failure and approximately 400 deaths annually in the United States.15 Rapid viral elimination may result in clearance of HBsAg from serum by the time of initial presentation. In these cases, the accurate diagnosis of fulminant hepatitis B may require testing with immunoglobulin (Ig) M antibody to hepatitis B core antigen (HBcAg) (IgM anti-HBc) (see later discussion of serologic markers of infection).16 The rate of spontaneous survival in acute liver failure caused by hepatitis B is only approximately 20%. Liver transplantation has resulted in survival rates of 50% to 60%. Recurrent disease in the allograft is now uncommon because of administration of hepatitis B immune globulin (HBIG) and potent orally administered antiviral agents (see later and Chapters 93 and 95).

Clinical Sequelae of Chronic Hepatitis B Virus Infection

Chronic hepatitis B develops in 2% to 5% of persons who acquire HBV infection in adulthood. Progressive liver disease (including cirrhosis and HCC) can be expected to develop in one quarter to one third of people who acquire infection in the first few years of life. An estimated 15% to 25% of patients ultimately die of liver-related causes, with the greatest risk in male HBV carriers. The presence of active viral replication and long-standing necroinflammatory liver disease caused by HBV strongly influences the rate of progression to cirrhosis. The major determinant of survival is the severity of the liver disease when the patient first comes to medical attention.17 Cirrhosis is associated with decreased survival and an increased frequency of HCC. Five- and 20-year survival rates of 55% and 25%, respectively, have been reported in patients with cirrhosis at presentation, whereas rates of 97% and 63%, respectively, have been reported for those with mild (noncirrhotic) disease.18 Survival rates differ most dramatically between patients with compensated and decompensated cirrhosis. In one study, an 84% five-year survival rate was reported for patients with compensated HBV-related cirrhosis, compared with 14% for patients with cirrhosis complicated by ascites, jaundice, encephalopathy, or a history of variceal bleeding.19 Multivariate analyses in several large cohort studies have identified age, ascites, hyperbilirubinemia, and other features of advanced liver disease as correlating independently with survival in patients with HBV-related cirrhosis. Interferon-induced clearance of HBeAg (see later) has been associated with prolongation of survival without complications or the need for liver transplantation.20

Chapter 78  Hepatitis B and D functions that are critical for packaging and DNA replication (including priming, RNA- and DNA-dependent DNA polymerase, and RNase H activities). Although HBV is a DNA virus, replication occurs through an RNA intermediate and requires an active viral reverse transcriptase/polymerase enzyme. The mutation rate is higher for HBV than for other DNA viruses (an estimated 1010 to 1011 point mutations per day).22 Complete HBV genomic sequencing has identified a large number of mutations within the HBV genome, many of which are silent or do not alter the amino acid sequence of encoded proteins. Because of genomic overlap, however, some of the silent mutations in one ORF (for example, the polymerase gene) may result in an amino acid substitution in an overlapping ORF (surface gene), although with currently uncertain clinical implications. Figure 78-2 illustrates the life cycle of HBV. The initial phase of hepadnaviral infection involves the attachment of mature virions to host cell membranes. The human receptor for HBV remains unknown. Entry of the virus results from fusion of the viral and host membranes as the nucleocapsid is released into the cytoplasm. Mechanisms of intracellular transport of viral genome into the nucleus are poorly understood, but the first step in genomic replication involves conversion of the relaxed circular form of HBV DNA into a double-stranded, covalently closed circular form (cccDNA). The cccDNA, which serves as the template for viral transcription, is the major form of viral DNA in the nucleus of infected hepatocytes. Subgenomic (0.7 to 2.4 kb) and prege-

Clearance of HBsAg in patients with HBV-related cirrhosis has been associated with an excellent prognosis, including improvement in liver histology and function, a decreased chance of viral reactivation, and prolonged survival.17 Even HBsAg clearance, however, is not an absolute safeguard against the future development of HCC in persons who already have cirrhosis.21

MOLECULAR BIOLOGY

HBV is a small DNA virus that belongs to the Hepadnaviridae family. Other members of this virus family are human HBV-like agents that infect the woodchuck, ground and tree squirrels, woolly monkey, crane, heron, Ross goose, and duck. HBV is a small (3.2-kilobase [kb]) virus with a DNA genome that has a relaxed, circular, partially doublestranded configuration (Fig. 78-1). The genome is composed of four open reading frames (ORFs) and has a compact design in which several genes overlap and use the same DNA to encode different viral proteins. The four viral genes are the core, surface, X, and polymerase genes. The core gene encodes the core nucleocapsid protein, which is important in viral packaging and production of HBeAg. The surface gene encodes the pre-S1, pre-S2, and S proteins (comprising the large [L], middle [M], and small [S] surface proteins). The X gene encodes the X protein, which has transactivating properties and may be important in hepatic carcinogenesis. The polymerase gene has a large ORF (approximately 800 amino acids) and overlaps the entire length of the surface ORF. It encodes a large protein with

Pre-

S2

+s tr

– -C

F-S

d an

F-P OR

56

ORF

3.5 kb RNA

NA

OR

d an

DR1

Pr eC

AAA AAA AAA

kb R

1 -S re st r

P

4 2.

2. 1

A RN kb

DR2

56

ORF-X

AAA

0.7 kb RNA

Figure 78-1.  Overlapping open reading frames (ORFs) of the hepatitis B virus (HBV) genome and major transcripts (wavy lines). The genome is partially double stranded with four overlapping ORFs, or genes. The S gene encodes the viral surface envelope proteins (HBsAg) and is composed of the pre-S1, pre-S2, and S regions. The core gene consists of the precore and core regions, which give rise to the hepatitis B e antigen (HBeAg) and core protein, respectively. The polymerase (P) gene overlaps the entire S gene, and mutations in this region may, in theory, give rise to changes in the HBsAg protein that affect neutralization by antibody to HBsAg. The fourth gene codes for an incompletely understood protein, HBX. Two 11-basepair direct repeats (DR1 and DR2) are required for strandspecific HBV DNA synthesis during viral replication.

1289

1290

Section IX  Liver Excess HBsAg

HBV Budding

Assembly Endoplasmic reticulum

Uncoating Figure 78-2.  Life cycle of the hepatitis B virus (HBV). The receptor for viral entry has not been identified. Once inside the hepatocyte, the virus undergoes uncoating, and the HBV genome enters the nucleus, followed by repair of the singlestranded DNA strand and formation of the covalently closed circular (ccc) DNA template. Viral transcripts are formed for the hepatitis B surface antigen (HBsAg), DNA polymerase, X protein, and RNA pregenome; the pregenome and polym­ erase are incorporated into the maturing nucleocapsid and removed after translation. The surface protein enveloping process occurs in the endoplasmic reticulum. Some of the nonenveloped nucleocapsid re-circulates back to the nucleus, and the cycle begins again. Excess tubular and spherical forms of HBsAg are secreted in great abundance.

nomic (3.5 kb) RNA molecules are transcribed from this template. The L protein is translated from the 2.4 kb RNA, the M and S proteins from the 2.1 kb RNA, and the X protein from the 0.7 kb transcript. The pregenomic RNA serves as the template for reverse transcription as well as the messenger RNA (mRNA) for translation of the core and polymerase proteins; the precore RNA codes for the precore gene product. HBV replication begins with encapsidation of the pregenomic RNA through complex interactions between host and viral proteins. HBV DNA polymerase reverse transcribes the pregenomic RNA into a negative-strand HBV DNA, which in turn serves as the template for positive-strand synthesis to form a partially double-stranded genome. Concurrent with HBV DNA synthesis, the nucleocapsid undergoes maturation and, through a yet incompletely understood mechanism, interacts with the S protein to initiate viral assembly in the endoplasmic reticulum. S protein is synthesized in the endoplasmic reticulum, where monomer aggregates that exclude host membrane proteins subsequently bud into the lumen as subviral particles. When formed, HBsAg undergoes glycosylation in the endoplasmic reticulum and the Golgi apparatus. Noninfectious subviral particles (spherical and filamentous forms of HBsAg) are secreted in great abundance when compared with mature virions.

Hepatitis B Virus Genotypes

A genetic classification based on comparisons of complete genomes has demonstrated eight genotypes of HBV, designated A through H (Table 78-1).23 Several methods have been used for HBV genotyping, including a commercially available line probe assay. Genotypic differences are based on an intergroup divergence of 8% or more in the complete nucleotide sequence. Genotype A is the predominant genotype in northern Europe and the United States. Genotypes B and C are confined to populations in eastern Asia and the Far East, but changes in immigration patterns have resulted in an influx of Asian HBV carriers with these genotypes into the United States.24 Genotype D is found worldwide but

HBsAg cccDNA

Translation

Positive strand synthesis

Removal of pregenome

Repair Transcription 2.4 / 2.1 kb RNA 3.5 kb RNA

Negative strand synthesis

Translation

Table 78-1  Hepatitis B Genotypes (A-H) and Their Possible Clinical Associations Geographic Distributions A: Northwestern Europe, North America, Central Africa B: Southeast Asia, including China, Japan, and Taiwan (prevalence is increasing in North America) C: Southeast Asia (prevalence is increasing in North America) D: Southern Europe, Middle East, India E: West Africa F: Central and South America, United States (Native Americans), Polynesia G: United States, France H: Central and South America Proposed Clinical Associations Time to HBeAg seroconversion and probability of HBsAg loss: B < C Response to treatment with interferon-α: A > B ≥ C > D Precore/core promoter mutant frequency: precore mutation not selected with A and F Liver disease activity and risk of progression: B < C Evolution to chronic liver disease: A < D Hepatocellular carcinoma risk: B > C in younger age group in Taiwan but B < C in older age group in Japan HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus.

is especially prevalent in the Mediterranean area, Middle East, and south Asia. Genotype E is indigenous to western sub-Saharan areas, and genotype F prevails in Central America. Cases of genotype G have been reported in the United States and France. Genotype H has been described in Mexico. Clinical associations appear to exist with the various genotypes (see Table 78-1). Currently, the strongest clinical associations appear to be that (1) HBeAg seroconversion occurs earlier in patients with genotype B than in those with genotype C and (2) response to therapy with interferon varies with genotype (see later).25 The viral genotype also has implications for the frequency of precore and core mutations (see later) and may have an effect on the frequency of HCC.

Chapter 78  Hepatitis B and D Mutations of the Hepatitis B Virus Genome

The vast majority of mutations in the HBV genome identified by comparing nucleotide sequences with those of wildtype HBV are silent or do not alter the amino acid sequence in a particular ORF. Some mutations have potentially important disease associations, however, and are described next. Hepatitis B Surface Antigen Mutants Mutations in the surface gene can result in changes in the antibody-binding domain. Accordingly, both virus neutralization by polyclonal antibody to HBsAg and testing for HBsAg by methods that depend on antibody binding can be affected. Large-scale vaccination programs in regions endemic for HBV have revealed a 2% to 3% frequency of vaccine escape mutants that result from alterations in the “a” determinant of the HBsAg protein, which is the major neutralizable epitope. The typical mutation results in the substitution of glycine for arginine at amino acid position 145; this substitution prevents binding of neutralizing antibodies (i.e., antibody to HBsAg [anti-HBs]). The clinical significance of these mutants for neonatal vaccination programs is highly controversial because the frequency of these variants among HBV-infected mothers whose infants respond to vaccination has been found to be similar to that of mothers whose infants do not respond. The “a” determinant mutants also are proposed to have clinical relevance after liver transplantation for hepatitis B. As many as 50% of patients in whom recurrent HBV infection develops despite the use of HBIG have been shown to have these escape mutants, and the rate at which the mutations are detected appears to correlate with the length of time over which HBIG is repeatedly administered.26 Mutations in the Precore, Basal Core Promoter, and Core Genes Mutations in the precore and basal core promoter regions of the HBV genome can influence the production of HBeAg. A precore mutation results in a stop codon at nucleotide 1896 that abolishes the synthesis of HBeAg,27 whereas mutations in the basal core promoter at nucleotides 1762 and 1764 decrease HBeAg synthesis by approximately 70% while maintaining pregenomic RNA levels.28 Both types of mutations have been observed in cases of severe or fulminant hepatitis, which has been attributed to the loss of the immune-tolerizing effects of HBeAg antigen (see later). The presence of core promoter mutations has been linked to a significantly increased risk of HCC.29 Precore and basal core promoter mutants have been described in the same patients and are particularly common in Asian and European patients with chronic hepatitis B.30 A large serosurvey of HBV carriers residing in the United States has found that precore and core promoter mutations are common (fre­ quencies of 27% and 44%, respectively), depending on the ethnicity and places of birth of the patients. Both mutant forms of HBV were observed to occur far more commonly in HBeAg-negative patients (precore mutation in 38% of HBeAg-negative versus 9% of HBeAg-positive patients; core promoter mutation in 51% versus 36%).31 In addition to these mutations, upstream mutations in the core gene can influence immunologic responses to HBV. Core gene mutations have been shown to block recognition of HBV by cytotoxic T lymphocytes (CTLs), a key mode of viral clearance. Therefore, the mutations contribute to HBV immune escape and possibly influence the response to interferon.32 Core gene mutations within the immunodominant epitopes of the HBV nucleocapsid also can affect CD4+ T-cell reactivity.33

In patients with perinatally acquired chronic hepatitis B, a prolonged immune tolerant phase with minimal to absent hepatic necroinflammatory activity is typically seen for the first 20 to 30 years of HBV infection. Sequencing studies have shown stable core gene sequences during this phase. Precore mutations are also uncommon during this phase. Core gene mutations become more common as patients pass from the immune tolerant phase, at which time a growing number of mutations are observed in the region of the core gene that includes many B- and T-cell epitopes. Both precore stop codon mutants and core gene mutants have been associated with a poor response to interferon therapy. Hepatitis B Virus DNA Polymerase Mutants The polymerase gene encodes a DNA polymerase enzyme needed for encapsidation of viral RNA into core particles, conversion of the pregenomic viral RNA into a negative strand of viral DNA (reverse transcription), and conversion of this first HBV DNA strand into a second DNA strand of positive polarity. In general, the HBV reverse transcriptase function of the polymerase gene is highly conserved because major mutations that impair the efficiency of viral replication lead to selection pressure against such variant forms. As indicated earlier, HBV has a high rate of replication (1011 virions per day) and low replication fidelity, meaning that it has a propensity to mispair nucleotide bases when it reverse transcribes viral RNA to DNA. HBV DNA polym­ erase also lacks any proofreading activity, so it cannot repair its mistakes. Therefore, when a nucleotide base is misplaced, it remains in the growing viral DNA strand as a base mutation, and the new HBV DNA molecule has a different sequence from the original (wild-type) genome. The overall error rate of HBV DNA polymerase is estimated to be 1 per 10,000 nucleotides copied, which translates to the potential for 10 million base-pair errors per day in an infected person. All possible single-base mutations can be produced in a 24-hour period, although many such mutations will yield nonviable viruses.34 Mutations in the sequence of HBV polymerase can lead to drug resistance to nucleoside analogs used to treat HBV infection because some HBV polymerase mutants have decreased susceptibility to these drugs and are selected during treatment. The mutations in the sequence of HBV DNA polymerase that confer drug resistance result in amino acid substitutions in the reverse transcriptase domain of the enzyme. The changes in the structure of the enzyme, in turn, are thought to sterically inhibit binding of the drugs to their active sites. The amino acids in the reverse transcriptase are numbered from 1 to 344, and an amino acid identity is given by the single letter amino acid code. By convention, substitutions in reverse transcriptase are designated by the wildtype amino acid, followed by the number of the amino acid, followed by the substituted amino acid. For example, for the nucleoside analog lamivudine (see later), two types of mutations occur at nucleotide position 204 of domain C (the catalytic site of the polymerase) that result in substitution of the amino acid methionine (M) for either isoleucine (I) or valine (V). These mutations are designated M204I and M204V, respectively, and are referred to collectively as YMDD mutants; the letters stand for the amino acids (Y = tyrosine, M = methionine, D = aspartate) in the C domain. The M204V mutation tends to occur in conjunction with a mutation in domain B that results in substitution of leucine (L) with methionine (L180M). The M204I mutation or the combined M204V-L180M mutations result in marked resistance to the effect of lamivudine (>10,000-fold reduction in susceptibility). After prolonged exposure to adefovir, drug resistant mutations in domains B and D are selected

1291

1292

Section IX  Liver L180M A181V/T T184G/S S202I M204V/I N236T M250V LAM LdT ADV

Table 78-2  Hepatitis Flares in Patients with Chronic Hepatitis B CAUSE OF FLARES

COMMENT

Spontaneous

Factors that precipitate antecedent viral replication are unclear Flares are often observed during withdrawal; requires preemptive antiviral therapy

TDF

Immunosuppressive therapy

ETV

Antiviral therapy for HBV   Interferon

Figure 78-3.  Common amino acid changes (top row) that result from nucleotide substitutions in the hepatitis B virus polymerase gene and confer drug resistance to oral antiviral agents (first column) used to treat hepatitis B (see text for description of nomenclature). Note that cross resistance exists among lamivudine, telbivudine, and entecavir, but entecavir requires the presence of one or more secondary substitutions at codons 184, 202, or 250 for clinically apparent drug resistance to appear. Note also that the potential for cross-resistance exists between adefovir and tenofovir, but the greater antiviral potency of tenofovir reduces the chances of resistance greatly. A, alanine; ADV, adefovir; ETV, entecavir; G, glycine; I, isoleucine; L, leucine; LAM, lamivudine; LdT, telbivudine; M, methionine; TDF, tenofovir; N, asparagine; S, serine; T, threonine; V, valine.

(A181V/T and N236T). Other nucleotide substitutions have been described that are instrumental for telbivudine and entecavir resistance (Fig. 78-3). The inherent mutability of HBV indicates that single and even double polymerase mutants preexist as minor “quasispecies” even before treatment of HBV infection is begun. Because of the limitations in the sensitivities of current genotype assays, these mutants would not be detectable until they are selected and expanded under the pressure of drug treatment. Resistance to lamivudine is found in approximately 20% of patients after one year of treatment but in nearly 70% after five years (see Fig. 78-3).35 Rates of resistance to adefovir, a nucleotide analog, are 0% at one year and 29% after five years.36 The efficacy of both of these drugs against HBV is impaired by a single nucleotide substitution. The more mutations necessary for drug resistance (indicating a higher genetic barrier to resistance), the slower the emergence of and lower overall rate of resistance. For example, resistance to entecavir, another nucleoside analog, occurs in less than 1% of patients at five years because the preexistence of lamivudine-resistant mutations and one or more additional mutations in the viral polymerase gene are required for resistance.37 Persistent infection with drugresistant HBV ultimately is associated with progression of disease and blunting of hepatic histologic improvement with antiviral therapy.38 Severe flares of hepatitis have also been reported after the emergence of drug-resistant mutants,39 and acquisition of these mutants may lead to rapidly progressive liver disease after liver transplantation (Table 78-2).40 Horizontal transmission of these mutants, which can complicate drug therapy in secondarily infected persons, is also possible.

PATHOGENESIS

HBV is generally not a cytopathic virus, and the severity of HBV-associated liver disease is considered to be related to the intensity of the host immunologic response to the virus. Whereas both humoral and cellular immune responses are needed for effective clearance of the virus, the cellular immune response appears to be the arm principally involved in the pathogenesis of disease. The immunologic response to HBV encompasses both an innate, or nonspecific, response

Lamivudine   During treatment   YMDD mutant   On withdrawal*

HIV treatment

Genotypic variation   Precore and core promoter mutants Superinfection with other hepatitis viruses

Flares are often observed during the second to third month; may herald virologic response Flares are no more common than with placebo Can have severe consequences in patients with advanced liver disease Flares are caused by rapid re-emergence of wild-type HBV; can have severe consequences in patients with advanced liver disease Flares can occur with HAART or with immune reconstitution; in addition, HBV increases the risk of antiretroviral drug hepatotoxicity Fluctuations in serum ALT levels are common with precore mutants May be associated with suppression of HBV replication

*Has also been reported with adefovir and entecavir. ALT, alanine aminotransferase; HIV, human immunodeficiency virus; HAART, highly active antiretroviral therapy; HBV, hepatitis B virus; YMDD, tyrosine-methionine-aspartate-aspartate.

(for example, natural killer cells and interferons) and an adaptive immune response, including antibodies to viral antigens, human leukocyte antigen (HLA) class II–restricted CD4+ T cells, and HLA class I–restricted CD8+ CTLs.41 Induction of the antigen-specific T-cell response is thought to occur in lymphoid organs, where the host T cells encounter viral peptide antigens (or epitopes) that are presented by antigen-presenting cells such as dendritic cells, B cells, and macrophages. This process results in the maturation and expansion of T cells that are specific for these viral epitopes and is followed by their migration to the liver, where they perform their effector function. During acute HBV infection, most HBV DNA molecules are cleared rapidly from the liver via noncytopathic mechanisms mediated by cytokines that are released initially by cells of the innate immune system42 and later by liverinfiltrating HBV-specific CD8+ cells. Cell-mediated immune responses are efficient in self-limited infection because the responses are vigorous, multispecific, and oriented toward type 1 helper T (Th1) cells. Persons with chronic HBV infection, by contrast, exhibit infrequent, narrowly focused, and weak HBV-specific T-cell responses.43 In chronic hepatitis B, the majority of mononuclear cells in liver infiltrates of patients with chronic hepatitis B at any given time are non–antigen-specific.44 CD8+ CTLs are thought to contribute to the disease process in the liver and result in apoptosis of infected hepatocytes. To be recognized by the CD8+ CTLs, targeted hepatocytes must present viral epitopes as short peptides that have been endogenously processed and fit within the peptide-binding groove of the class I major histocompatibility complex (MHC) molecules.45 The binding of the CTL

Chapter 78  Hepatitis B and D HBeAg

Anti-HBe

Normal range

HBV DNA

ALT

Phase

Immune tolerant

Immune clearance

Inactive carrier state

Optimal treatment time Liver

Minimal, if any, inflammation

Chronic inflammation/ fibrosis

Reactivation Optimal treatment time

Mild inflammation and minimal fibrosis

Active inflammation

Figure 78-4.  Natural evolution and phases of chronic hepatitis B in a person with perinatal or early-life acquisition of the infection. The four phases are designated the immune tolerant phase, immune clearance phase, inactive carrier state, and reactivation phase, as determined by biochemical, virologic, and histologic activity of the disease. In some patients, hepatitis B e antigen (HBeAg) seroconversion to antibody to HBeAg (anti-HBe) is followed by the selection of precore and/or core promoter mutant forms of HBV and continuing hepatitis. This evolution generally occurs after several decades of infection and accounts for the frequent occurrence of active HBeAg-negative hepatitis in middle age. At this point, some patients may have high-normal or slightly elevated serum ALT levels despite having significant underlying fibrosis and inflammation. These patients are not in the immune tolerant phase but instead should be considered to be in the immune clearance phase with mild disease. Ultimately, prolonged histologic activity (necroinflammation) leads to cirrhosis in at least 20% of patients; cirrhosis, in turn, as well as ongoing necroinflammatory activity in the liver, can result in hepatocellular carcinoma. The optimal times for antiviral therapy are during the immune clearance and reactivation phases.

T-cell receptor (TCR) to the peptide-MHC complex on the hepatocyte surface can then result in the direct killing of the infected cell and release of potent antiviral cytokines by the activated CTL. Recognition by MHC class II– restricted CD4+ helper T cells requires the appropriate presentation of viral peptides in the context of class II MHC molecules. The CD4+ cells produce antiviral cytokines and provide help in neutralizing antibody production. Antibody neutralization limits intrahepatic spread of virus during primary infection and serves an important role in preventing reinfection.

NATURAL HISTORY

Four phases of HBV infection have been described: immune tolerance, immune clearance, the inactive carrier state, and reactivation (Fig. 78-4). Patients who acquire the infection in the perinatal period often have high serum levels of HBV DNA without biochemical evidence of active hepatitis and are considered to be immune tolerant to HBV. When followed longitudinally, many of these patients ultimately exhibit elevated serum aminotransferase levels in association with histologic evidence of chronic hepatitis. The trigger mechanisms for this apparent change in tolerance are poorly understood but likely reflect changes in the immune reactivity of the host. Experiments in transgenic mice suggest that HBeAg induces a state of immunologic tolerance to HBV in neonates.46 Perinatal transmission of HBeAg has been considered to be a potential mechanism for the immune-tolerant state. As persons enter the immune clearance phase, HBV DNA concentrations diminish, serum alanine aminotransferase (ALT) levels rise, and hepatic histologic activity, reflecting

immune-mediated lysis of infected hepatocytes, increases. The duration of this second phase varies, often lasting many years. The third phase (inactive HBV carrier state) occurs after seroconversion from HBeAg to anti-HBe and is usually preceded by a marked reduction in serum HBV DNA to levels that are detectable only by PCR methodology, followed by normalization of serum ALT levels and resolution of liver necroinflammation. This phase may last a lifetime, but a proportion of patients ultimately undergo spontaneous or immunosuppression-mediated reactivation of HBV replication with reappearance of high levels of HBV DNA in serum, with or without HBeAg seroreversion and a rise in serum ALT levels. For unclear reasons, precore or core promoter mutants that prevent or down-regulate HBeAg production may be selected during or after HBeAg seroconversion (see earlier).47 A key event in the natural history of HBeAg-positive chronic hepatitis is seroconversion of HBeAg to anti-HBe, which is associated with marked reduction in HBV replication and biochemical and histologic remission in the majority of patients. Regression of liver fibrosis occurs gradually months to years after HBeAg seroconversion.48 Most studies have found that the mean annual rate of spontaneous HBeAg seroconversion ranges from 8% to 15% in HBV-infected children or adults with serum ALT elevations. Longitudinal studies of untreated patients with predominantly HBeAg-positive chronic hepatitis B have shown that the frequency of development of cirrhosis ranges from 2 to 5 per 100 person-years and the five-year cumulative frequency of progression to cirrhosis from 8% to 20%.49 The rate of cirrhosis has been suggested to be higher in HBeAg-

1293

1294

Section IX  Liver negative patients than in HBeAg-positive patients. Risk factors for the development of cirrhosis have been identified; of these, older age, the stage of fibrosis at presentation, and ongoing HBV replication with persistent or intermittent detection of HBV DNA by a non–PCR-based assay are perhaps the most important clinically. Combined infection with hepatitis D virus (HDV [see later]), HCV, or HIV and concomitant alcohol abuse have also been linked to a higher rate of development of cirrhosis. When cirrhosis develops, two major complications may occur: hepatic decompensation and HCC. In a large European cohort with HBV-related compensated cirrhosis, the five-year cumulative frequency of hepatic decompensation was 16%, and the incidence per 100 person-years was 3.3.50 Similar rates have been reported in Asians. The cumulative five-year frequency of HCC can be as high as 14%.50 Factors associated with an increased risk of HCC include male gender, age more than 45 years, having a first-degree relative with HCC, the presence of cirrhosis, HBeAg positivity, and reversion from anti-HBe to HBeAg positivity.51 HCC can still develop in HBsAg-positive persons with none of the identified risk factors, but less frequently. In addition, HCC has been described in persons who lose HBsAg. Recommendations about ultrasonography and alpha fetoprotein screening for HCC are controversial, but in general, screening is recommended in all patients with cirrhosis and in male HBV carriers older than age 40 years in whom the likely route of transmission has been perinatal or early childhood exposure; some authorities recommend screening after age 30 years in highly viremic patients when perinatal acquisition is suspected (see Chapter 94). ALT as a Surrogate Marker for Disease Activity The serum ALT level has been used conventionally as a measure of disease activity in patients with chronic hepatitis B. Use of the standard reference range (0 to 40  U/L), however, can be misleading for evaluating HBV-related disease activity. A serum ALT level within the normal range is an imperfect surrogate marker for the absence of disease activity because determination of standard reference ranges has not traditionally taken into account increased body mass index (BMI), diabetes mellitus, and other features, such as alcohol intake, that tend to inflate values in a “normal” reference population (see Chapter 73). An insurance record-based study in Korea that included more than 140,000 persons who were followed for eight years demonstrated that all-cause liver-related mortality is increased when the serum ALT level exceeds 20  U/L in women and 30  U/L in men.52 This finding is particularly relevant to the management of Asians with hepatitis B and viremia. Because these patients tend to have a small body mass, a normal serum ALT value according to the standard laboratory reference range may be misleading. Such patients often have been excluded from treatment and assumed to be immune tolerant (that is, without liver disease). Studies in Asia and the United States have shown that as many as 20% to 30% of HBV carriers with persistently normal serum ALT levels and serum HBV DNA levels >104 copies/ mL have stage 2 or greater (of 4) inflammation and stage 2 or greater (of 4) fibrosis on a liver biopsy specimen.53 Moreover, Asian HBV carriers with high-normal serum ALT levels (>0.5 × upper limit of normal [ULN]) have been shown to have more fibrosis on liver biopsy specimens than do those with low-normal serum ALT levels (<0.5 × ULN) and often demonstrate higher serum ALT elevations on prolonged follow-up.53-55 These data are consistent with the hypothesis that viremic HBV carriers who acquired HBV infection early in life and who have high-normal serum ALT levels may represent a subgroup of patients who have

already entered the immune clearance phase and are not in the immune-tolerant phase. This occurrence should be suspected in particular in persons older than age 35 to 40 because immune tolerance often subsides after two to three decades of infection.54 Liver biopsy can be a useful tool to distinguish persons in the immune clearance phase despite normal or near-normal ALT levels but with active liver disease from those in the immune tolerant phase and absence of active liver disease (see Fig. 78-4). HBV DNA Level and Long-Term Complications Population-based Asian cohort studies have established that the serum HBV DNA level is the single best predictor of future progression to cirrhosis and HCC in HBV-infected persons.56,57 In a prospective cohort study, more than 3600 HBV carriers from Taiwan, of whom 60% were male, 70% were older than age 40, 85% were HBeAg negative, and 95% had normal serum ALT levels, were followed for a mean of 11 years. The calculated relative risks for cirrhosis and HCC were shown to correlate with the level of HBV DNA on entry into the study when compared with a reference population of HBV carriers in whom HBV DNA was undetectable in serum by a PCR assay.57 Even serum HBV DNA levels as low as 10,000 copies/mL (equivalent to 2000 IU/mL) were associated with a higher relative risk of cirrhosis and HCC. The relative risk of HCC was highest in persons with a serum HBV DNA level of more than 1 million copies/mL and intermediate in those in whom follow-up serum samples indicated spontaneous reduction of the serum HBV DNA level from greater than 100,000 copies/mL to less than 10,000 copies/mL. These data can be interpreted to mean that both the duration and level of viremia are important risk factors for the development of HCC. The data also suggest that suppression of serum HBV DNA levels, whether spontaneously or as a result of antiviral therapy, lowers the risk of HCC. The serum HBV DNA level remained a significant predictor of cirrhosis and HCC even after adjustment for patient age, gender, serum ALT level, and HBeAg status. Other studies in Chinese HBV carriers have demonstrated that persons in whom liver decompensation and HCC develop often have modest serum ALT elevations that are frequently below the recommended threshold (2 × ULN) for antiviral treatment in current practice guidelines (see later).58 On the basis of these data, some authorities have suggested that all persons who acquire HBV infection early in life and who are age 45 to 50 years or older and demonstrate serum HBV DNA levels of 100,000 or more copies/mL should receive long-term therapy with a nucleoside analog to prevent cirrhosis and HCC.59,60 In a landmark study,61 more than 600 Asian patients with advanced fibrosis and a serum HBV DNA level greater than 100,000 copies/mL were randomized in a ratio of 2 : 1 to active treatment with lamivudine or placebo. Treatment was planned for 5 years, but the study was discontinued after a mean duration of only 32 months because disease progression and HCC occurred significantly more frequently in the group of patients randomized to placebo.61 Therapeutic benefit was independent of the serum ALT level at baseline. These data strongly suggest that continuing suppression of serum HBV DNA by antiviral therapy can alter the long-term outcome of active chronic hepatitis B.

CLINICAL FEATURES Acute Hepatitis B

The incubation period of acute hepatitis B varies from a few weeks to 6 months (average, 60 to 90 days), depending on the amount of replicating virus in the inoculum. The disease

Chapter 78  Hepatitis B and D may be more severe in patients coinfected with other hepatitis viruses and in those with established underlying liver disease.62 Abstention from alcohol is usually recommended, but the chance of an uneventful recovery does not appear to be affected by the consumption of moderate amounts of alcohol (20 to 30 g daily) during the convalescent phase.63 Acute infections are heralded by a serum sickness–like prodrome of fever, arthralgia or arthritis, and rash, which is most commonly maculopapular or urticarial, in 10% to 20% of patients. This prodrome results from circulating HBsAg–anti-HBs complexes that activate complement and are deposited in the synovium and walls of cutaneous blood vessels.64 These features generally abate before the manifestations of liver disease and peak serum aminotransferase elevations are observed. Jaundice develops in only about 30% of patients. Clinical symptoms and jaundice generally disappear after one to three months, but some patients have prolonged fatigue even after serum ALT levels return to normal. In general, elevated serum ALT levels and serum HBsAg titers decline and disappear together, and in approximately 80% of cases, HBsAg disappears by 12 weeks after the onset of illness.65 In 5% to 10% of cases, HBsAg is cleared early and is no longer detectable by the time the patient first presents to a health care provider. Persistence of HBsAg after six months implies development of a carrier state, with only a small likelihood of recovery during the next 6 to 12 months. Delayed clearance of HBsAg has been reported to be preceded by a decline in HBsAg titers. Serum aminotransferase levels of 1000 to 2000 U/L are typical, with ALT being higher than aspartate aminotransferase (AST) levels. In patients with icteric hepatitis, the rise in serum bilirubin levels often lags behind that in ALT levels. The peak ALT level does not correlate with prognosis, and the prothrombin time (international normalized ratio [INR]) is the best indicator of prognosis. After clinical recovery from acute hepatitis B and HBsAg seroconversion, HBV DNA often remains detectable in serum as determined by a PCR assay (see later discussion of diagnosis). After resolution of acute hepatitis, the numbers of HBV-specific CD4+ and CD8+ cells in blood and liver decrease rapidly. Nonetheless, T-cell responsiveness remains high on re-encounter with HBV antigens, indicating that traces of virus can maintain the CTL response indefinitely after clinical recovery, thereby exerting control over the virus and preventing reactivated infection.41,66 Fulminant hepatitis occurs in less than 1% of cases (see Chapter 93). Fulminant hepatitis B generally occurs within four weeks of the onset of symptoms and is associated with encephalopathy, multiorgan failure, and a high mortality rate (>80%) if not treated by liver transplantation. Patients older than age 40 years appear to be more susceptible than younger persons to “late-onset liver failure,” in which encephalopathy, renal dysfunction, and other extrahepatic complications of severe liver insufficiency become manifest over the course of several months. The pathogenic mechanisms of fulminant hepatitis are poorly understood but are presumed to involve massive immune-mediated lysis of infected hepatocytes. This proposed mechanism may explain why many patients with fulminant hepatitis B have no evidence of HBV replication in serum at presentation.

Chronic Hepatitis B

A history of acute or symptomatic hepatitis is often lacking in patients with chronic HBV infection. When symptoms are present, fatigue tends to predominate over other constitutional symptoms, such as poor appetite and malaise. Right upper quadrant pain also may occur but is generally low grade. Patients may remain asymptomatic even during

periods of reactivated hepatitis. In other instances, particularly when superimposed on cirrhosis, reactivation of HBV infection may be associated with frank jaundice and signs of liver failure (see later discussion of acute flares in chronic hepatitis B). Physical findings may be normal, or hepatosplenomegaly may be found. In decompensated cirrhosis, spider angiomata, jaundice, ascites, and peripheral edema are common. Liver biochemical test results are usually completely normal during the inactive HBV carrier state. In contrast with patients in the immune-tolerant phase of HBV infection, most patients in the immune-clearance phase of chronic HBV infection have mild to moderate elevations in serum AST and ALT levels. During exacerbations of disease, serum ALT levels may be as high as 1000 U/L or more, and the clinical and laboratory picture is indistinguishable from that of acute hepatitis B, including the presence in serum of IgM anti-HBc. Progression to cirrhosis should be suspected whenever hypersplenism, hypoalbuminemia (in the absence of nephropathy), or prolongation of the prothrombin time is found. The serum AST level is typically higher than the serum ALT level in patients with advanced cirrhosis (see Chapter 73). Extrahepatic Manifestations Extrahepatic syndromes seen in association with acute or chronic hepatitis B are important to recognize because they may occur without clinically apparent liver disease and can be mistaken for independent disease processes in other organ systems. The pathogenesis of these extra­ hepatic disorders has not been fully elucidated but likely involves an aberrant immunologic response to extrahepatic viral proteins.67 Many of the extrahepatic manifestations (e.g., arthritis, dermatitis, glomerulonephritis, polyarteritis nodosa, cryoglobulinemia, papular acrodermatitis, and polymyalgia rheumatica) are observed in association with circulating immune complexes that activate serum com­ plement. Antiviral therapy may be indicated for persistent symptoms. Arthritis-Dermatitis.  The constellation of fever, arthralgias, rash, angioneurotic edema, and, less commonly, hematuria and proteinuria is seen as a prodromal manifestation of acute hepatitis B and rarely in patients with chronic hepatitis B. The pro­ximal interphalangeal joints, knees, ankles, shoulders, and wrists are the joints most commonly affected. During the period of acute joint symptoms, HBsAg titers in the blood are high and complement levels are low. Correspondingly, HBsAg has been detected in synovial membranes, and complement levels in synovial fluid are low. Evidence of activation of the complement system by HBsAg–anti-HBs complexes exists. After the joint symptoms subside, complement levels return to normal, and HBsAg titers in serum begin to decline. This syndrome must be distinguished from inflammatory forms of arthritis because glucocorticoid therapy mistakenly given to patients with such HBV manifestations enhances HBV replication, and abrupt withdrawal of these agents may be associated with a flare in disease activity. Polyarteritis Nodosa.  One of the most serious extrahepatic syndromes associated with chronic HBV infection is polyarteritis nodosa. As many as 30% of patients with polyarteritis nodosa are infected with HBV, but the disorder develops in less than 1% of patients with HBV infection, either after acute or recent hepatitis B or, more commonly, in association with chronic HBV infection. Typical features include arthralgias, mononeuritis, fever, abdominal pain, renal disease, hypertension, central nervous system abnormalities, and rash. Medium to small arteries and arterioles are involved by fibrinoid necrosis and perivascular infiltra-

1295

1296

Section IX  Liver tion. The disease is thought to result from deposition of circulating immune complexes that contain HBsAg; for this reason, therapy with plasmapheresis may be indicated. Good therapeutic responses also have been observed with antiviral agents, given alone or in combination with plasmapheresis. No apparent relationship exists between the severity of the vasculitis and the severity of the hepatic disease, and the hepatic disease often is relatively mild despite high levels of viral replication. The course of polyarteritis nodosa is variable, but the prognosis is gravest for patients with substantial proteinuria (>1 g/day), renal insufficiency (serum creatinine > 1.6 mg/dL), gastrointe­ stinal involvement, cardiomyopathy, and involvement of the central nervous system. Glomerulonephritis.  Several types of glomerular lesions have been described in patients with chronic HBV infection; membranous glomerulonephritis and membranoproliferative glomerulonephritis are the most common.68 Renal biopsy specimens have demonstrated immune complex deposition and cytoplasmic inclusions in the glomerular basement membrane. The immune complexes activate complement and production of cytokines with a subsequent inflammatory response. Nephrotic syndrome is the most common presentation of HBV-associated glomerulonephritis. In affected children, renal failure at presentation is almost always mild, and a history of clinical liver disease is uncommon. Nevertheless, liver biopsy specimens almost always demonstrate varying degrees of chronic hepatitis. The diagnosis of HBV-associated glomerulonephropathy is usually established by serologic evidence of HBV antigens or antibodies, the presence of immune-complex glomerulonephritis in a renal biopsy specimen, and the demonstration of glomerular deposits of one or more HBV antigens, such as HBsAg, HBcAg, or HBeAg, by immunohistochemistry. Most patients have detectable HBeAg in serum and, in addition, demonstrate low serum C3 and occasionally low C4 levels. The renal disease typically resolves in months to several years, especially in children. Often, resolution occurs in conjunction with HBeAg seroconversion. Rarely, however, renal failure may ensue. The natural history of HBV-related glomerulonephritis in adults has not been well defined, but several reports suggest that glomerular disease is often slowly and relentlessly progressive.69 Successful treatment has been accomplished with interferon alpha and

has been linked to long-term control of HBV replication.70 Therapy with nucleoside analogs has resulted in improved renal function and diminished proteinuria. Cryoglobulinemia.  Type II cryoglobulins consist of a polyclonal IgG and monoclonal IgM, whereas type III cryoglobulins contain polyclonal IgG and rheumatoid factor. Type II and type III cryoglobulinemia have been associated with hepatitis B, but the association is uncommon. In a large patient cohort, the frequency of cryoglobulinemia was significantly higher in patients with chronic HCV infection (54%) than in patients with chronic HBV infection (15%) (see Chapter 79). Cryoglobulinemia may be associated with systemic vasculitis (purpura, arthralgias, peripheral neuropathy, and glomerulonephritis) but is often pauci­ symptomatic or asymptomatic. Interferon has been used successfully to treat symptomatic cryoglobulinemia in association with chronic hepatitis B. Experience with nucleoside analog therapy has not been reported. Histopathologic Features Chronic HBV infection is characterized by mononuclear cell infiltration in the portal tracts. Periportal inflammation often leads to the disruption of the limiting plate of hepatocytes (interface hepatitis), and inflammatory cells often can be seen at the interface between collagenous extensions from the portal tracts and liver parenchyma (referred to as active septa). During reactivated hepatitis B, lobular inflammation is more intense and reminiscent of that seen in acute viral hepatitis. Steatosis is not a feature of chronic hepatitis B, as it is in chronic hepatitis C. The only histologic feature noted on routine light microscopy that is specific for chronic hepatitis B is the presence of ground-glass hepatocytes (Fig. 78-5). This morphologic finding results from accumulation of HBsAg particles (20 to 30 nm in diameter) in the dilated endoplasmic reticulum. Because of high levels of cysteine in HBsAg, ground-glass cells have a high affinity for certain dyes, such as orcein, Victoria blue, and aldehyde fuchsin. Ground-glass hepatocytes also may be seen in HBV carriers, in whom they may be detected in up to 5% of cells. When present in abundance, ground-glass hepatocytes often indicate active viral replication.71 Immunofluorescence and electron microscopic studies have shown HBcAg inside the hepatocyte nuclei of affected cells. During periods of intense hepatitis

Hepatitis B surface antigen

Ground-glass inclusions

A

B

Figure 78-5.  A, Liver biopsy specimen showing ground-glass inclusions in hepatocytes. These inclusions represent large amounts of hepatitis B surface antigen (HBsAg) in the endoplasmic reticulum of infected hepatocytes (Hematoxylin and eosin, ×630.) B, Immunohistochemical stain for HBsAg. Note that the brownish inclusions correspond to the ground-glass inclusions seen in A (×630). (Courtesy of Dr. Gist Farr, New Orleans, La.)

Chapter 78  Hepatitis B and D activity, cytoplasmic core antigen staining is generally observed. After successful treatment of HBV infection with a nucleoside analog, the cytoplasmic core antigen staining often disappears, but nuclear core antigen staining may remain, indicating persistence of the HBV cccDNA template.

deterioration occurs, often without the subsequent loss of HBeAg. Multiple episodes of reactivation and remission have been shown to accelerate the progression of chronic hepatitis B and are particularly likely to occur in patients infected with the precore mutant form of chronic hepatitis B (see earlier).47

Acute Flares in Chronic Hepatitis B

Immunosuppressive Therapy-Induced Flares Reactivation of HBV replication is a well-recognized complication in patients with chronic HBV infection who receive cytotoxic or immunosuppressive therapy.76 Suppression of the normal immunologic responses to HBV leads to enhanced viral replication and is thought to result in widespread infection of hepatocytes by HBV. On dis­ continuation of immunosuppressive medications, such as cancer chemotherapy, antirejection drugs, and glucocorticoid therapy, immune competence is restored and infected hepatocytes are rapidly destroyed. The more potent the immunosuppression, the higher the level of viral replication and, thus, the greater the potential for serious clinical consequences of sudden withdrawal of the therapy and restoration of immunologic competence. Postmortem studies of liver tissue from patients with severe liver injury have documented sparse staining of viral antigens, suggesting that the patients were in an active state of immune clearance.77 The vast majority of patients who experience immunosuppressive therapy–induced flares have been positive for HBsAg in serum before treatment, but some studies have described the reappearance of HBsAg in patients who were initially positive for anti-HBs, anti-HBc, or both.78 Reactivated hepatitis in patients who are negative for HBsAg and positive for either anti-HBc or anti-HBs is explainable by the possible latency of HBV in liver and mononuclear cells and the large extrahepatic reservoir of HBV. Chemotherapy given to patients with cancer who are HBV carriers is associated with an increased risk of liver-related morbidity and mortality.79 Reactivated hepatitis B also occurs in patients who are given immunosuppressive medications to prevent organ transplant rejection. The frequency of reactivated hepatitis appears to be particularly high in patients who undergo bone marrow transplantation because of extensive immunologic conditioning before transplantation and treatment of graft-versus-host disease.80 Rarely, fibrosing cholestatic hepatitis, a rapidly progressive form of liver injury associated with inordinately high levels of HBsAg and HBcAg in liver tissue, may develop in such patients.81 Acute flares of hepatitis B resulting from cancer chemotherapy and other immunosuppressive drugs are often detected after substantial increases in serum aminotransferase levels have been noted. Initiation of antiviral treatment after detection of such biochemical abnormalities has little effect on reducing liver injury because much of the immunologic response to HBV and viral elimination has already occurred. Instead, the key to management lies in anticipating the occurrence of a flare, initiating antiviral treatment preemptively (e.g., 4 to 6 weeks before the start of chemotherapy), and continuing the treatment for 6 to 12 months after completion of chemotherapy.82

Chronic hepatitis B is often punctuated by sudden flares of disease activity that are reflected by a rise in serum aminotransferase levels. Although a uniform definition is lacking, a flare has frequently been described as a rise in serum ALT levels to at least two times the baseline value. Spontaneous flares are an important part of the natural history of hepatitis B because when they occur repeatedly, they lead to histologic progression. Acute flares in chronic hepatitis B occur in association with a number of circumstances and clinical situations (see Table 78-2). Most flares result from a change in the balance between immunologic responses to HBV and the level of viral proliferation. Acute flares in chronic hepatitis B that are not explainable by infection with other hepatotropic viruses often occur as a secondary response to increased levels of replicating wild-type or mutant HBV or as a result of therapeutic intervention with immunologic modifiers such as interferon, glucocorticoids, and cancer chemotherapy. In some instances, the event that initiates an acute exacerbation of chronic hepatitis B may not be readily identifiable, and the flare is considered spontaneous. Spontaneous Flares Spontaneous exacerbations of chronic hepatitis B often result from reactivated infection, and an increase in serum HBV DNA levels often precedes an increase in serum aminotransferase levels. Histologic evidence of acute lobular hepatitis superimposed on the changes of chronic viral hepatitis is frequently observed during these flares.72 IgM anti-HBc, a marker that is often diagnostic of acute viral hepatitis, may also appear in serum at this time.73 The reasons for reactivated infection are unknown but likely relate to subtle changes in the immunologic control of viral replication. Reactivation seems to occur more commonly in persons who are infected with HIV.74 In persons who acquire HBV infection early in life, flares become more common during adulthood, presumably because of a breakdown in immune tolerance to HBV.75 Fatigue may be reported during flares of chronic hepatitis B, but in many instances, patients remain asymptomatic. Occasionally symptoms and signs of frank liver failure become apparent, particularly when the flare is superimposed on advanced chronic hepatitis B. Most clinically recognizable flares occur in patients who are in the nonreplicative phase of HBV infection (i.e., initially testing positive for anti-HBe and negative for serum HBV DNA on a molecular hybridization assay). During such flares, serum HBV DNA levels increase, and HBeAg often reappears in serum (seroreversion). HBV DNA and HBeAg are often detectable in serum when the patient is first seen, but if the flare has been ongoing for several weeks or longer, the accompanying enhancement of the immune response may make it difficult to detect a rise in serum HBV DNA levels. Frequently, subsidence of these flares of hepatitis is accompanied by loss of HBV DNA and HBeAg in serum. Flares also can occur in patients who are in the replicative phase of infection (i.e., already positive for HBV DNA and HBeAg in serum). In these instances, HBV replication intensifies, serum HBV DNA levels rise, and liver biochemical

Antiviral Therapy–Induced Flares Antiviral treatment of chronic hepatitis B can be associated with flares of hepatitis in several circumstances. Flares may occur during interferon therapy, after withdrawal of nucleoside analogs or glucocorticoid therapy, and in association with lamivudine-resistant mutants (see also later discussion of treatment).

1297

1298

Section IX  Liver Interferon.  Interferon-induced flares of chronic hepatitis B occur in approximately one third of treated patients and result from the immunostimulatory properties of the drug. Flares generally occur during the second or third month of treatment with conventional preparations of interferon. Flares also occur in patients treated with pegylated interferon and have been reported to occur more frequently in patients infected with HBV genotype A than with other genotypes. This finding may explain the higher rate of sustained virologic response and HBsAg clearance in patients infected with HBV genotype A.83 Serum ALT flares have been shown to be a predictor of sustained virologic response, particularly in patients with high levels of viremia.84 Flares tend to be particularly common in patients who have decompensated liver disease, with rates as high as 50% reported in one series.85 Such flares are frequently associated with clinical deterioration in the patient. Nucleoside Analogs.  Serum ALT flares occur in approximately 20% to 25% of patients after withdrawal of nucleoside analogs such as lamivudine and adefovir. These flares probably are caused by rapid resurgence of wild-type HBV, and although generally well tolerated, they have been associated with serious clinical exacerbations in patients with advanced liver disease.86 Reinstitution of therapy is often associated with a decline in serum HBV DNA and aminotransferase levels. Flares have been seen to follow the emergence of YMDD mutant HBV during therapy with lamivudine (see earlier).87 Initial reports emphasized the temporal occurrence of these flares at the time of or shortly after detection of lamivudine resistance. Further follow-up of patients with lamivudine-resistant HBV mutants, however, has indicated that the frequency of moderate or severe serum ALT flares (defined as >5 and >10 × UNL, respectively) increases with time after detection of lamivudine resistance. In one long-term study, the cumulative frequencies of such ALT flares were as follows: 24% at less than one year, 29% at one to two years, 30% at two to three years, 37% at three to four years, and 61% at more than four years after detection of lamivudine resistance.88 Glucocorticoid Withdrawal.  Serum ALT levels increase, often with an inverse decline in HBsAg concentration and HBV DNA level, after withdrawal of glucocorticoids in patients with chronic HBV infection.89 In clinical trials of HBV therapy, a short course of glucocorticoid therapy given before conventional antiviral therapy was reported to enhance virologic response rates.90,91 The immune rebound following withdrawal of glucocorticoid therapy after a fourto eight-week course may result from increased activation of lymphocytes that promote Th1 cytokine responses at a time when viral antigen expression is increased. Serious hepatic decompensation has been reported in patients with advanced disease, however, and this therapeutic approach is no longer used. Antiretroviral Therapy.  Serum ALT flares occur in patients coinfected with HIV and HBV who receive highly active antiretroviral therapy (HAART).92 A number of potential causes have been identified. Lamivudine resistance and withdrawal may be associated with ALT flares. HBV infection raises the risk of toxicity from antiretroviral therapy, usually within six months after the initiation of treatment. Immune reconstitution resulting from HAART may also be associated with ALT flares. Affected patients may also be particularly susceptible to flares because of infection with other hepatitis viruses. Flares Associated with Genotypic Variation Chronic infection with precore mutant HBV (referred to as HBeAg-negative chronic hepatitis B) often is associated

with multiple flares of liver cell necrosis interspersed with periods of normal serum ALT and low serum HBV DNA levels.47 Approximately 45% of patients have episodic serum ALT flares with normal levels between episodes, and 20% have flares superimposed on persistent ALT elevations.93 These flares have been attributed to rises in the concentration of precore mutants in the liver and changes in the ratio of concentrations of precore to wild-type HBV. Mutations at the basal core promoter region of the HBV genome are associated with greater HBeAg synthesis, histologic evidence of liver inflammation, and increased viral replication.94 Multiple exacerbations of hepatitis resulting from reactivated HBV infection have been described in patients with basal core promoter mutations, either alone or in association with precore mutations. HBeAg-negative patients who have both precore and core promoter mutants may be particularly predisposed to episodes of severe reactivation after cancer chemotherapy.95 Flares Caused by Infection with Other Viruses Patients with chronic HBV infection may exhibit severe flares in serum aminotransferase levels and even frank liver failure when superinfected with other hepatotropic viruses, such as hepatitis A virus (HAV), HCV, and HDV. Increased mortality has been reported when HDV superinfection is superimposed on chronic hepatitis B, and chronic HDV infection is often associated with multiple fluctuations in serum aminotransferase levels (see later discussion of HDV).96 Acute hepatitis C superimposed on chronic hepatitis B has been reported to be as severe as HDV superinfection and has been associated with a high rate of liver failure (34%) and death (10%).97 A cumulative frequency of cirrhosis and HCC that is higher than that attributable to chronic HDV infection or chronic HBV infection alone has been demonstrated. Acute hepatitis C often leads to chronic HCV infection, and the subsequent course also may be characterized by frequent fluctuations in serum aminotransferase levels. Patients with chronic hepatitis B who become infected with other hepatotropic viruses (HAV, HCV, HDV, or both HCV and HDV) may test negative for both HBeAg and HBV DNA on non–PCR-based assays because of viral interference.

DIAGNOSIS

HBsAg appears in serum 2 to 10 weeks after exposure to HBV and before the onset of symptoms or elevation of serum aminotransferase levels. In self-limited acute hepa­ titis, HBsAg usually becomes undetectable after four to six months. Persistence of HBsAg for more than six months implies progression to chronic HBV infection. The disappearance of HBsAg is followed several weeks later by the appearance of anti-HBs. In most patients, antiHBs persists for life and provides long-term immunity. In some patients, anti-HBs may not become detectable after disappearance of HBsAg, but these patients do not appear to be susceptible to recurrent infection.98 Anti-HBs may not be detectable during a window period of several weeks to months after the disappearance of HBsAg. During this period, the diagnosis of acute HBV infection is made by the detection of IgM anti-HBc in serum.99 Coexistence of HBsAg and anti-HBs in serum has been reported in approximately 25% of HBsAg-positive persons and occurs more commonly in persons with chronic hepatitis B than in those with acute hepatitis B.100 In most instances, the anti-HBs is present in a low level, nonneutralizing, and heterotypic—that is, directed against a

Chapter 78  Hepatitis B and D subtype of HBsAg different from the subtype present in the infected patient. The mechanisms behind this finding are not clear but relate to antibody formed against minor variants of the HBsAg protein. The presence of these heterotypic antibodies is not associated with specific risk factors or changes in clinical course and may occur in patients with or without active liver disease and viral replication. Anti-HBc is detectable in acute and chronic HBV infection. During acute infection, anti-HBc is predominantly of the IgM class and is usually detectable for four to six months after an acute episode of hepatitis and rarely for up to two years. As noted earlier, IgM anti-HBc may become detectable during exacerbations of chronic hepatitis B. Anti-HBc persists in persons who recover from acute hepatitis B and also in association with HBsAg in those who progress to chronic infection. In areas where HBV is not endemic, isolated anti-HBc in serum has been detected in 1% to 4% of the general population. Isolated reactivity for anti-HBc may occur in the following situations: (1) during the window period of acute hepatitis B, when anti-HBc is predominantly of the IgM class; (2) many years after recovery from acute hepatitis B, when anti-HBs has fallen to undetectable levels; (3) as a false-positive serologic test result; (4) after many years of chronic HBV infection, when the HBsAg titer has fallen below the level of detection; (5) in HBV-infected persons who are coinfected with HCV; and (6) rarely, as a result of varying sensitivity of HBsAg assays.101 Evidence for coinfection with HCV has been demonstrated in as many as 60% of persons in whom anti-HBc is the only marker of HBV.102 Results of PCR testing of sera have shown that 0% to 30% of patients with isolated anti-HBc have HBV DNA in serum. Usually, the HBV DNA is detectable at a low level and not by standard hybridization assays, which are less sensitive than PCR assays.103 The presence of low-level viremia in these HBsAg-negative subjects has clinical implications with regard to potential infectivity. For example, in the past, anti-HBc testing of blood donors prevented some cases of post-transfusion hepatitis B.104 Also, the risk of transmission of HBV infection from a liver donor with isolated anti-HBc has been found to be as high as 50% to 70% in some series; lower rates of transmission have been observed in other forms of solid organ transplantation.105 Low-level viral replication also has implications with regard to the possibility of underlying liver disease. HBV DNA in serum and liver tissue has been confirmed by PCR methodology in some HBsAg-negative patients with cirrhosis and HCC and in some patients with fulminant non-A, non-B, non-C hepatitis as defined by conventional serologic testing.106 HBeAg is a soluble viral protein that is found in serum early during acute HBV infection. HBeAg reactivity usually disappears at or soon after the peak in serum aminotransferase levels, and persistence of HBeAg three or more months after the onset of illness indicates a high likelihood of transition to chronic HBV infection. The finding of HBeAg in the serum of an HBV carrier indicates greater infectivity, a high level of viral replication, and the need for antiviral therapy. With a commercially available PCR assay, nearly 90% of patients with HBeAg-positive chronic hepatitis B were found to have serum HBV DNA levels persistently above 105 copies/mL, with a mean value of 8.37 log10 (>108) copies/mL.107 By contrast, anti-HBe–positive patients had much lower serum HBV DNA levels; higher values were found in those with persistently or intermittently elevated serum ALT levels (mean of 5.1 log10 copies/mL) than in those with persistently normal ALT levels (3.10 log10 copies/mL).

Most HBeAg-positive patients have active liver disease; the exceptions are HBeAg-positive children and young adults with perinatally acquired HBV infection, who usually have normal serum ALT levels and minimal inflammation of the liver.14 In general, seroconversion from HBeAg to anti-HBe is associated with a reduction in serum HBV DNA levels of 3 log10 copies/mL or greater and remission of liver disease. Some patients, however, continue to have active liver disease and detectable HBV DNA in serum because of low levels of wild-type virus or the selection of precore or core promoter mutations that impair HBeAg secretion. HBV DNA can be measured in serum with qualitative or quantitative assays. The clinical utility of testing for serum HBV DNA has been hampered by the absence of a licensed test in the United States as well as an accepted international reference standard. A number of non–PCR-based assays are available with levels of sensitivity ranging from 103 to 105 genomic copies/mL of serum. Although these assays are less sensitive than those that are PCR-based, their results correlate with clinical response to antiviral therapy, and several of the currently available antiviral therapies were licensed on the basis of clinical trials in which these assays were used (see later). The use of these less sensitive non–PCRbased assays has several shortcomings, however, so most clinical laboratories use one of several commercially available PCR assays with enhanced sensitivity (102 genomic copies/mL or less). The measurement of serum HBV DNA is commonly used to evaluate a patient’s candidacy for antiviral therapy and to monitor response during treatment. Patients with a high serum HBV DNA level at baseline less commonly respond to therapy with conventional interferon than patients with a low level.108 With the use of solution hybridization testing, a baseline HBV DNA level of 200 pg/mL (roughly equivalent to 56 million copies/mL on a PCR assay) or greater has been found to be associated with a low rate of response to standard interferon. By contrast, baseline serum HBV DNA levels have not been shown to correlate with response to nucleoside analog therapy because of the more potent inhibition of viral replication by these agents. Monitoring of HBV DNA levels at key intervals during therapy allows one to predict the likelihood of HBeAg clearance. Several studies have found that the level of serum HBV DNA at 12 weeks of nucleoside analog treatment may help predict the likelihood of HBeAg seroconversion.109,110 Other studies have suggested that measuring the HBV DNA level at baseline or during treatment can be used to evaluate the likelihood of relapse after treatment is discontinued and development of resistance to lamivudine.111,112 Reappearance of HBV DNA in serum during treatment suggests that drug resistance has occurred,113 and high pretreatment levels of serum HBV DNA have been shown to correlate with a higher rate of recurrent HBV infection in liver transplant recipients who are treated with lamivudine.114 Qualitative PCR is an even more sensitive method of detecting HBV DNA than quantitative PCR. Use of qualitative PCR has altered traditional concepts about the clearance of HBV DNA from serum in acute and chronic HBV infection. Small amounts of HBV DNA can be detected in serum and peripheral mononuclear cells years after recovery from acute hepatitis B.66 Even after disappearance of HBsAg and apparent loss of HBV DNA from serum in patients with chronic hepatitis B, small amounts of HBV DNA persist in liver tissue and peripheral mononuclear cells years later.9 Detection of HBV DNA in serum by a qualitative PCR assay before liver transplantation may identify patients who are at increased risk of apparent de novo

1299

1300

Section IX  Liver hepatitis after transplantation and may pinpoint HBV as the cause of liver disease in HBsAg-negative patients.115,116 Finally, detection of minute amounts of HBV DNA may be particularly important in diagnosing patients with ful­ minant hepatitis B—who frequently have cleared HBsAg by the time they seek medical attention.117

TREATMENT Virologic Endpoints and Definitions of Response

The primary goal of treatment for chronic hepatitis B is durable suppression of serum HBV DNA to levels below those associated with liver disease. This goal can be accomplished with either interferon alpha or nucleoside analogs. The level at which serum HBV DNA is suppressed adequately is generally considered to be less than 105 copies/ mL (<20,000 IU/mL) for patients with HBeAg-positive chronic hepatitis B and often lower for those with HBeAgnegative hepatitis.14 Definitions of response vary, but the most important clinically is a lasting or durable suppression of serum HBV DNA when the patient is no longer receiving treatment (Table 78-3). HBeAg seroconversion (loss of HBeAg and appearance of anti-HBe in serum) is an addi-

Table 78-3  Treatment of Chronic Hepatitis B: Definitions of Response to Antiviral Therapy Virologic response

Biochemical response On-treatment response   Initial response

  Maintained response Off-treatment response   Sustained response   Durable response

Decrease in serum HBV DNA level to <105 copies/mL or <20,000 IU/mL in HBeAg-positive cases and <104 copies/mL or <2000 IU/mL in HBeAg-negative cases Loss of HBeAg with or without seroconversion to anti-HBe* Normalization of serum ALT levels Suppression of HBV DNA levels to <104-5 copies/mL with or without loss of HBeAg, in addition to normalization of serum ALT levels Requiring continuation of therapy Virologic and biochemical response observed for 6-12 months after treatment is discontinued Indefinite virologic and biochemical response after treatment is discontinued

*Pertains to HBeAg-positive patients only. ALT, alanine aminotransferase; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus.

tional endpoint that can be used to determine the appropriate length of treatment with a nucleoside analog. Although most experts agree that achieving HBsAg seroconversion is more desirable than HBV suppression, the former occurs so infrequently with current antiviral therapies that it is considered an impractical endpoint. Even so, a systematic review of clinical trials with interferon alpha indicated that early HBsAg seroconversion occurs significantly more frequently in treated than in nontreated patients. In a meta-analysis of 16 randomized, controlled trials, loss of HBsAg from serum occurred six times as frequently in interferon-treated patients as in nontreated patients.118 Long-term follow-up (mean, 6.2 years; range, 1 to 11 years) of HBeAg-positive patients treated with standard interferon alpha demonstrated that 71% of sustained responders became HBsAg negative.119 By contrast, a one-year course of lamivudine or adefovir does not result in a higher rate of HBsAg seroconversion than does placebo, and the frequency of HBsAg seroconversion with prolonged therapy has yet to be evaluated extensively. The observation that early HBsAg seroconversion occurs more often with interferon than with nucleoside analogs emphasizes that the mechanisms of action of the two treatments differ fundamentally and provides a rationale for the use of combination therapy using both types of drugs (see sections on the individual agents and combination therapy).

Factors Involved in the Choice of Agents

In deciding on the appropriate type of therapy for patients with chronic hepatitis B, the physician should consider the serum ALT level, serum HBV DNA level, and liver histology at baseline as well as the expense of treatment, potential for and ability of the patient to withstand adverse effects, age and other comorbid conditions of the patient, and realistic expectations about the need for monitoring. Interferon and nucleoside analogs each have advantages and disadvantages, and no one therapy is suitable for all patients (Table 78-4). One major advantage of therapy with interferon is that it tends to be time limited, in that durable responses do not require maintenance therapy. By contrast, prolonged treatment with a nucleoside analog is often necessary to maintain viral suppression.

Guidelines for the Management of Hepatitis B

Consensus guidelines for the treatment of hepatitis B have been published by the American Association for the Study of Liver Diseases (AASLD), Asian-Pacific Association for the Study of the Liver, and European Association for the Study of the Liver.120-122 In general, the three sets of guidelines are quite similar. As innovations in medical therapy are made, the guidelines are updated, and differences among the

Table 78-4  Advantages and Disadvantages of Currently Available Antiviral Agents AGENT(S)

ADVANTAGES

DISADVANTAGES

Peginterferon

Finite duration of treatment Durable off-treatment response Disappearance of HBsAg (5%-8%)

Given by injection Frequent side effects Expensive Low response rate patients with a high level of viremia

Nucleos(t)ide analogs

Negligible side effects Potent inhibition of virus replication Less expensive than interferon

Oral delivery Drug resistance Long or indefinite treatment duration Low rate of HBsAg disappearance Moderately expensive when given long term*

*Average retail price is approximately $200-$700 (USD) per month, depending on drug. HBsAg, hepatitis B surface antigen.

Chapter 78  Hepatitis B and D guidelines represent, in part, differences in the availability of the various therapeutic agents around the world as well as unavoidable delays in publication that prevent the incorporation of new data. Because the practice guidelines of the liver societies are evidence based, recommendations have not been made about combination therapy with more than one nucleoside analog or interferon plus a nucleoside analog. In fact, combination therapy has not been shown in the available clinical trials to have additional therapeutic benefit (see later). The three guidelines have changed as new information has become available.

The recommendations made in the three sets of published guidelines have many similarities. In general, the published guidelines recommend treatment of persons who have both biochemical evidence of liver injury and serum HBV DNA levels in excess of 20,000 IU/mL, or roughly 100,000 copies/ mL.123 Nucleoside analog therapy is recommended specifically in patients with decompensated cirrhosis. Emphasis also is given to the treatment of patients with serum ALT levels that are at least double the upper limit of normal (Table 78-5), although some experts disagree with setting arbitrary serum ALT and HBV DNA thresholds.59,123 This

Table 78-5  Recommendations for Treatment of Chronic Hepatitis B* HBeAg STATUS

SERUM HBV DNA LEVEL

SERUM ALT LEVEL

TREATMENT STRATEGY

+

>20,000 IU/mL

≤2 × ULN

+

>20,000 IU/mL

>2 × ULN

−

>20,000 IU/mL

>2 × ULN

−

>2000 IU/mL

1-2 × ULN

− + or − (cirrhosis)

≤2,000 IU/mL Detectable

≤ULN —

+ or − (cirrhosis)

Undetectable

—

Observe; consider treatment if serum ALT level rises Low efficacy of currently available treatments Pegylated interferon alpha, lamivudine, adefovir, entecavir, telbivudine, or tenofovir may be used as initial therapy but lamivudine and telbivudine are not preferred because of high rates of resistance Duration of therapy: Pegylated interferon alpha: 48 weeks Lamivudine: minimum of 1 year§ Adefovir: minimum of 1 year§ Entecavir: minimum of 1 year§ Telbivudine: minimum of 1 year§ Tenofovir: minimum of 1 year§ For interferon nonresponders or those with contraindications, use tenofovir or entecavir For lamivudine resistance, add adefovir or tenofovir; or switch to emtricitabine/tenofovir combination; or switch to high-dose entecavir For adefovir resistance, add lamivudine; or switch to emtricitabine/ tenofovir combination; or switch to or add entecavir For entecavir resistance, switch to or add adefovir or tenofovir For telbivudine resistance, add adefovir or tenofovir; or switch to emtricitabine/tenofovir combination; or switch to high-dose entecavir Pegylated interferon alpha, lamivudine, adefovir, entecavir, telbivudine or tenofovir may be used as initial therapy (see above) Endpoints of treatment: Sustained normalization of serum ALT level and undetectable HBV DNA on PCR assay Duration of therapy: Interferon alpha: 1 year Lamivudine: >1 year Adefovir: >1 year Entecavir: >1 year Telbivudine: >1 year Tenofovir: >1 year For pegylated interferon alpha nonresponders or those with contraindications to interferon alpha, use tenofovir or entecavir For drug resistance, considerations are same as above Consider liver biopsy and treat if moderate or severe inflammation or significant fibrosis Observe, treat if serum HBV DNA or ALT level increases Compensated cirrhosis: Treat if HBV DNA >2000 IU/mL (lamivudine and telbivudine not preferred) If HBV DNA <2000 IU/mL, consider treatment if ALT elevated Decompensated cirrhosis: Refer for liver transplantation and coordinate treatment with transplant center; adefovir plus either lamivudine or telbivudine; or entecavir; or tenofovir preferred. Compensated cirrhosis: Observe Decompensated cirrhosis: Refer for liver transplantation

†

‡

*Revised from reference 120. † Approximately equal to serum HBV DNA level >105 copies/mL. ‡ Some authorities treat these patients indefinitely to try to reduce the rate of future complications if the patient is older than age 40 and infection was acquired early in life. In this paradigm, patients with high-normal serum ALT levels are considered appropriate for treatment and liver biopsy is optional. Other authorities prefer to treat patients with a serum HBV DNA level >20,000 IU/mL and any degree of serum ALT elevation. Liver biopsy is preferred for patients older than 35-40 years of age (see text). § Treatment should be continued for 6-12 months after HBeAg seroconversion and may vary in duration with the severity of liver disease before treatment. ALT, alanine aminotransferase; HBeAg, hepatitis B e antigen; IU, international units; PCR, polymerase chain reaction; ULN, upper limit of normal.

1301

1302

Section IX  Liver recommendation is based on the observation that rates of sustained virologic response in patients with minimal pretreatment serum ALT elevations are low with either interferon or nucleoside analogs.108,124 All the guidelines indicate that treatment decisions ideally should be made in the context of liver histologic findings and that treatment should be directed preferentially to patients with histologically moderate to severe hepatitis, although currently treatment decisions are not based on specific grading or staging of liver histology.

Antiviral Agents

Seven antiviral agents have been licensed for the treatment of hepatitis B as of 2009. One of these drugs, standard or conventional interferon alpha-2b, largely has been replaced in clinical usage by pegylated interferon alpha-2a. The choice of drug therapy should be individualized based on patient and viral specific features whenever possible. Interferons Interferon alpha.  Interferon alpha was licensed for the treatment of chronic hepatitis B in 1992. Interferon is effective after a relatively short course of treatment (four months to one year) and, unlike the nucleoside analogs, has not been associated with drug resistance. Also, unlike nucleoside analogs, interferon has direct immunomodulatory properties. Interferon enhances HLA class I antigen expression on the surface of infected hepatocytes and augments CD8+ CTL activity. These effects could be important mechanistically in reducing the amount of the HBV cccDNA (the genomic template for viral transcription) and may thereby explain the loss of HBsAg that occurs in approximately 5% to 8% of interferon-treated patients. The major disadvantages of interferon relate to its poorer acceptance because of side effects (see Chapter 79) and lower level of HBV DNA suppression. Flares of serum ALT have been described during therapy with interferon alpha, and although these flares are potentially important in achieving a virologic response, they occur unpredictably and are inconsistently associated with antiviral efficacy. The magnitude of an ALT flare has been shown to predict the likelihood of a sustained virologic response in patients with high levels of viremia, suggesting that vigorous cell-mediated immune responses often are required to overcome high levels of viral replication.84 Pegylated Interferon Alpha.  Pegylated interferon has been found to be more potent than conventional or standard interferon and is currently licensed in more than 75 countries.125 Doses of 1.0 µg per kg of body weight of peginterferon alfa-2b and 180 µg of peginterferon alfa-2a given once weekly have been studied in clinical trials.126-128 No data are available for judging whether the increased effectiveness of pegylated interferon is primarily a function of a more pronounced effect on viral replication or of greater immunomodulatory action. Viral genotype appears to affect the response to interferon. In a report from Taiwan, patients with genotype-B, HBeAg-positive chronic hepatitis B were found to have a response to conventional interferon more frequently than patients with genotype-C chronic hepatitis B.129 A relationship between virologic response and genotype was reaffirmed in a large multicenter study of peginterferon alfa-2b. In this study, HBeAg-positive patients infected with HBV genotype A demonstrated HBeAg loss more frequently than those with genotypes B, C, and D (47% vs. 44%, 28%, and 25%, respectively).126 A subsequent follow-up study of this cohort after a mean interval of three years demonstrated that patients infected with HBV genotypes A or B had the highest

rates of durable virologic response (96% and 86%, respectively) and the highest rate of HBsAg clearance (58% and 14%, respectively).130 By contrast, rates for patients infected with HBV genotype D were 76% and 6%, respectively. These results confirm and extend those of earlier studies in HBeAg-positive patients that suggested that patients infected with HBV genotype A respond more frequently than those infected with genotype D. The relationship between sustained virologic response and HBV genotype is less clear in patients with HBeAg-negative hepatitis B, but a systematic analysis of more than 500 patients treated with either conventional or pegylated interferon alpha found that patients infected with HBV genotype C had the highest rates of sustained virologic and biochemical response and those infected with HBV genotype D had the lowest rates (50% and 21%, respectively).131 The effect that genotype exerts on the response to interferon could be particularly relevant to the treatment of North American patients with chronic HBV infection, in light of the influx of Asian HBV carriers infected with HBV genotypes B and C beginning in the last decades of the 20th century. Nucleoside and Nucleotide Analogs Nucleoside analogs have excellent oral bioavailability, a good safety record, and antiviral efficacy comparable to that observed with interferon alfa-2b. These drugs have proved to be particularly useful in the management of patients with decompensated cirrhosis, in whom even small doses of interferon can lead to worsening liver failure and severe infections. Nucleoside and nucleotide analogs replace natural nucleosides during the synthesis of the first or second strand (or both) of HBV DNA. They thus serve as competitive inhibitors of the viral reverse transcriptase and DNA polymerase. Because nucleos(t)ide analogs partially suppress viral replication, they have to be given for more than one year in most cases to achieve maximal efficacy. Unfortunately, drug resistance occurs with prolonged monotherapy. Figure 78-3 illustrates the common HBV nucleotide substitutions associated with drug resistance and the potential for cross resistance. Nucleos(t)ide analogs have several other limitations as well. With these agents, demonstrating the clearance of HBV cccDNA has been difficult, and in contrast to treatment with interferon, HBsAg clearance rarely occurs after one year of treatment with nucleoside analogs. These problems may result, in part, from the fact that these agents, unlike interferon, do not have a direct, enhancing effect on the immunologic response to HBV.78 Also, as indicated earlier, postwithdrawal serum ALT flares have been seen in approximately 25% of cases after discontinuation of nucleoside analog therapy. Lamivudine.  The approval of the nucleoside analog lamivudine in 1998 was a major breakthrough in the treatment of hepatitis B. The drug has been shown to be a relatively potent inhibitor of viral replication, convenient to administer, and free of severe adverse effects. Clinical trials demonstrated that a one-year course of lamivudine resulted in suppression of viral replication and improvement in histologic findings in the liver.132,133 In one study, HBeAg loss and HBeAg seroconversion occurred in 32% and 17% of patients, respectively.133 A two-year course of lamivudine proved to be more effective, resulting in an increase in the rate of HBeAg seroconversion from 17% at one year to 27% at two years.134 Prolongation of treatment beyond one year, however, has been associated with incremental changes in viral resistance (38% at two years), and the longer treatment is continued, the more frequently resistance is seen (65% at year five).88 Resistance is even more commonly encountered

Chapter 78  Hepatitis B and D (90% at four years) in patients coinfected with HIV because of the early use of lamivudine in HAART regimens.92 Lamivudine resistance for more than two years has been associated with a blunted histologic response, and patients in whom lamivudine resistance has developed experience more hepatitis flares.38,88 For these reasons, the drug is no longer recommended as first-line therapy. Fortunately, a number of alternative nucleos(t)ide analogs have considerably lower resistance profiles. Adefovir Dipivoxil.  Adefovir dipivoxil is the acyclic phosphonate nucleotide analog of adenosine monophosphate. The drug was approved in 2002 for the treatment of HBeAg-positive and HBeAg-negative chronic hepatitis B on the basis of the findings of randomized, controlled trials in the United States, Europe, and Asia.135,136 In these pivotal studies, treatment with adefovir for 48 weeks resulted in median serum HBV DNA reductions of 3.52 log10 and 3.91 log10 copies/mL in HBeAg-positive and HBeAg-negative patients, respectively. The rates of HBeAg loss and HBeAg seroconversion were slightly lower than those achieved with lamivudine for 52 weeks. A rise in the frequency of HBeAg seroconversion and nondetectability of HBV DNA by PCR methodology has been observed during the second year of adefovir treatment.137 The level of HBV DNA suppression has been the same irrespective of viral genotype.138 Although the extent of viral suppression is only 0.5 to 1.0 log value less with adefovir than with lamivudine, the two drugs differ greatly in their resistance profiles. Point mutations (A181V, N236T) in the B and D domains, respectively, of the HBV polymerase gene that affect HBV susceptibility to adefovir occur in only 3% of patients after two years of treatment but increase thereafter, with rates of 6% to 8% at three years, 15% to 18% at four years, and 20% to 29% at five years.139-143 HBV isolates with the N236T mutation have remained susceptible to lamivudine and appear to be sensitive to entecavir and telbivudine in vitro (see later).144 Adefovir has been shown to be clinically and virologically effective in patients with lamivudine-resistant HBV, whether they have clinically stable disease, decompensated cirrhosis, or recurrent hepatitis B after liver transplantation.145,146 Adefovir-resistant mutants, on the other hand, remain susceptible to lamivudine, and adefovir resistance has been shown to occur more frequently when lamivudine is discontinued in lamivudine-resistant patients than when lamivudine is continued.147,148 Therefore, most experts choose to continue lamivudine when starting adefovir therapy in patients who have become resistant to lamivudine. The sooner resistance to lamivudine is appreciated, the more rapid and complete the virologic response to the addition of adefovir.149 Adefovir has the disadvantage of potential nephrotoxicity, and dose reductions may be necessary in patients with or at risk of compromised renal function.146 Adefovir is still used frequently, but incomplete viral suppression occurs in 30% of patients, particularly those with high viral levels at the initiation of treatment. Therefore, adefovir is being replaced as first-line therapy with more potent drugs such as entecavir and tenofovir (see later). Entecavir.  Entecavir is a deoxyguanine nucleoside analog that inhibits HBV replication selectively. The drug blocks HBV replication by inhibiting the priming of HBV DNA polymerase and the synthesis of the first and second strands of HBV DNA. Entecavir was approved by the U.S. Food and Drug Administration (FDA) in 2005 on the basis of registration trials that demonstrated no resistance after one year of treatment and improved virologic efficacy when compared with lamivudine.150-151 Entecavir is effective against both

wild-type and lamivudine-resistant HBV. In phase 3 clinical trials, an entecavir dose of 0.5 mg was used to treat HBeAgpositive and HBeAg-negative patients who were previously untreated with a nucleoside analog, whereas 1 mg was used in patients who were resistant to lamivudine. Entecavir is more potent than either lamivudine or adefovir and results in nondetectable HBV DNA in 67% and 90% of previously untreated (“treatment-naïve”) HBeAg-positive and HBeAgnegative patients, respectively, after one year. In HBeAgpositive patients treated for an additional year, the cumulative rate of HBV DNA negativity by a PCR assay was 80% for entecavir-treated patients compared with 37% for lamivudine-treated patients.152 Entecavir resistance is rare (approximately 1% at five years) in treatment-naïve patients but is common in patients with prior lamivudine resistance. Virologic “rebound” as a result of resistance to entecavir has been demonstrated to occur in 1% of lamivudine-refractory patients after one year of treatment and in an additional 9% after two years, with further increases as treatment is continued.153 Telbivudine.  Telbivudine is an l-nucleoside analog of thymidine that has been shown to be more potent than lamivudine in randomized, controlled trials in patients with HBeAg-positive and HBeAg-negative hepatitis B.154,155 In the registration trial, 1367 patients with HBeAg-positive or HBeAg-negative hepatitis B were randomized to receive telbivudine 600 mg or lamivudine 100 mg, each once daily, for 104 weeks.155 Virologic and biochemical responses associated with telbivudine were superior to those with lamivudine. Several randomized studies also have reported rapid and marked reductions in serum HBV DNA levels in patients who had been switched from adefovir to telbivudine. Unfortunately, after one and two years of treatment, genotypic resistance was found in 5% and 25% of HBeAgpositive patients, respectively, and 2% and 11% of HBeAgnegative patients, respectively.155,156 The highest rates of virologic response and, conversely, the lowest rates of resistance, were found in patients who had less than 3 log10 copies of HBV DNA at week 24 of treatment, with the best responses found in those with a negative HBV DNA result by a PCR assay. Resistance to telbivudine is conferred by the M204I mutation either alone or in conjunction with the L180M mutation. These mutations also confer resistance to lamivudine, and for this reason switching to telbivudine is not preferred in cases of lamivudine resistance. Despite the greater antiviral potency and lower resistance profile of telbivudine compared with lamivudine, telbivudine has a higher rate of resistance than entecavir and tenofovir (see later) and has fallen out of favor as a treatment for HBV infection. In addition, a multicenter study of telbivudine in combination with pegylated interferon was stopped prematurely because of the development of myopathy and elevations in serum creatinine kinase levels in a higher percentage of patients than reported with other nucleoside analogs. Tenofovir Disoproxil Fumarate.  Tenofovir, an acyclic nucleotide inhibitor of HBV polymerase and HIV reverse transcriptase, is similar chemically to adefovir dipivoxil but is produced in a 300-mg rather than a 10-mg formulation. Tenofovir was originally licensed for the treatment of HIV infection and approved in 2008 for the treatment of HBV infection. Its antiviral activity against HBV has been reported to be significantly greater than that of the 10-mg dose of adefovir in lamivudine-resistant and treatment-naïve patients.157,158 Resistance to tenofovir has not occurred after two years of treatment in either HBeAg-positive or HBeAgnegative patients.158 Preliminary studies have suggested that adefovir-resistant HBV may be less sensitive than wild-type

1303

1304

Section IX  Liver HBV to tenofovir.159 Whether unique nucleotide substitutions in the HBV polymerase will ultimately be found that confer resistance to tenofovir in either previously untreated or adefovir-resistant cases is unknown. Prolonged tenofovir therapy has been associated with bone loss and decreased bone density in HIV-infected patients.160 As with adefovir, renal tubular damage and Fanconi syndrome have been associated with prolonged use of tenofovir in a small number of cases and may be a particular risk in the elderly or persons with existing mild renal disease.161 Serum creatinine levels should be monitored regularly during prolonged use of tenofovir, and if the serum creatinine level rises, the dose may need to be modified or the drug discontinued to reduce the risk of further nephrotoxicity. Concomitant use of ritonavir or didanosine in HIV-infected patients has been found to be associated with a greater risk of Fanconi syndrome.162 Other Nucleoside Analogs Emtricitabine.  Emtricitabine is a fluorinated cytosine analog that inhibits HBV DNA polymerase and HIV reverse transcriptase. This drug is currently licensed in the United States and other countries for the treatment of HIV type 1 (HIV-1) infection. Preliminary results from a placebocontrolled, phase 3 study in previously untreated patients with chronic HBV infection have demonstrated that emtri­ citabine, 200 mg daily for 48 weeks, reduces serum HBV DNA levels by a median of 3 log10 copies/mL and improves liver histology significantly.163 Emtricitabine treatment of patients coinfected with HBV and HIV has led to levels of suppression of HBV DNA similar to those that occur in treated patients infected with HBV alone.163 The drug is structurally related to lamivudine, however, and has similar mutational sites and rates of resistance. In HBV monoinfected patients, the frequencies of YMDD mutations in persons who receive emtricitabine, 200 mg daily, have been shown to be between 9% and 13% at week 48 and 19% at week 96.164 On the basis of these findings, this agent is unlikely to play an important role in the management of HBV monoinfection. Clevudine.  Clevudine, a pyrimidine analog, is a potent inhibitor of HBV replication both in vitro and in vivo. This drug has been studied in woodchucks and in humans and is licensed for treatment of hepatitis B in Korea on the basis of 24-week clinical trials. In one study, therapy with clevudine given daily for 12 weeks resulted in a reduction in serum HBV DNA levels of greater than 4 log10 copies/mL.165 After therapy was discontinued, rebound to pretreatment HBV DNA levels occurred slowly in this and other studies.166 In one study, a 3-log10 reduction in HBV DNA levels was still evident 6 months after discontinuation of clevudine after a 24-week course of treatment.167 This pattern is different from that reported with other nucleoside analogs and might be explainable by a suppressive effect of clevudine on HBV cccDNA, the genomic template that has been shown to be relatively resistant to treatment with other oral agents. To date, little is known about the safety and efficacy of longterm treatment with clevudine. In 2009, clinical trials were halted because of the occurrence of myopathy in patients treated with clevudine for more than 24 weeks. Combination of a Nucleoside and Nucleotide Agent.  The FDA has approved a combination formulation of tenofovir (300 mg) and emtricitabine (200 mg) for use in patients with HIV infection. This formulation is not specifically licensed for treatment of hepatitis B, but the drug could be useful in cases of resistance to lamivudine, telbivudine, entecavir, adefovir, or tenofovir. HBV mutants resistant to the first four drugs remain susceptible to tenofovir, and tenofovirresistant HBV remains susceptible to emtricitabine. The

drug is occasionally used as first-line therapy of HBV monoinfection in patients with advanced cirrhosis because drug resistance has greater consequences in this clinical setting. As a dual formulation tablet, the drug may improve patient adherence and reduce the cost of care when compared with combination therapy with two nucleoside analogs. Viral Resistance to Nucleoside and Nucleotide Analogs.  With the exception of tenofovir, all currently available nucleos(t)ide analogs have been associated with drug resistance when used as monotherapy. Genotypic resistance is the term that describes the finding of a nucleotide sub­stitution in the HBV DNA polymerase gene that has been associated with clinical evidence of drug resistance. Such a mutation can be detected by a commercially available reverse hybridization assay (InnoLipa, Innogenetics, Belgium). The detection method can detect only known mutants and is limited by the ability to detect only the most common mutations. Furthermore, to be detectable, the drugresistant mutant HBV has to constitute at least 10% of the viral population in an infected patient. Testing for drugresistant HBV should be done prior to discontinuation of the drug because of the rapid re-emergence of wild-type HBV on discontinuation of the drug, at which point the mutant HBV can no longer be detected. Phenotypic resistance refers to the findings of an in vitro assay that uses a human liver cell line transfected with the relevant HBV DNA polymerase mutant. Various concentrations of test drug are added to the culture to assay for loss of susceptibi­ lity of the virus to the drug. This test is a research tool and is helpful in determining whether new genotypic changes cause clinically meaningful drug resistance. Registration trials for nucleos(t)ide analogs have incor­ porated serial genotypic testing for drug-resistant HBV mutants. These studies have shown quite clearly that genotypic resistance to a nucleos(t)ide analog often occurs weeks to months before a virologic breakthrough (defined as greater than a 1-log10 [10-fold] increase in serum HBV DNA levels above the previous nadir) (Fig. 78-6). Ultimately, if the patient is continued on the drug, virologic rebound (defined as an increase in serum HBV DNA levels to greater than 100,000 copies or 20,000 IU/mL) will occur and will be followed by elevation of the serum ALT level (see Fig. 78-6). Rescue therapy can modify this sequence of events if a second agent that lacks cross resistance to the original drug is used alone to replace the first drug or added to the first drug. many experts prefer to add another nucleos(t)ide analog that does not share the same resistance pattern rather than switch to an alternative monotherapy in this clinical situation. A patient treated with a nucleos(t)ide analog should be tested at periodic intervals to assess for virologic breakthrough.168 Such testing enables assessment of suppression of viremia and permits detection of viral breakthrough as early during the course of treatment as possible. To do this, the serum HBV DNA level should be monitored at threemonth intervals. Patients who do not have at least a 1-log10 decline in the serum HBV DNA level after three months of treatment are considered to have primary treatment failure and should be given an alternative agent. The next key interval is at 24 weeks of treatment. In the GLOBE trial, the relative efficacies of lamivudine and telbivudine were compared, and persons who remained positive in serum for HBV DNA at 24 weeks of treatment with either drug were more likely to have a failed response and eventually developed resistance; the risk of a negative outcome to treatment was proportional to the level of detectable HBV DNA in serum at the 24-week point.155 In clinical practice, patients

Chapter 78  Hepatitis B and D Virologic rebound

HBV DNA

ALT Virologic breakthrough

Resistance mutation

Biochemical breakthrough

Time Figure 78-6.  Time course of events leading to clinically apparent drug resistance in patients treated for hepatitis B virus (HBV) infection. In this case, HBV DNA levels decline in serum during antiviral therapy. The levels may or may not become undetectable, and with continued selection pressure by the drug, compensatory mutations develop and improve the ability of the mutant to replicate. The drug-resistant HBV mutant expands within the viral quasispecies. Detection of genotypic resistance occurs weeks to months before virologic breakthrough (a rise in HBV DNA levels in serum), which in turn may precede biochemical breakthrough elevation of serum alanine aminotransferase (ALT) levels by weeks to months. Ultimately, if the patient continues to take the drug, virologic rebound (a marked rise in HBV DNA levels) and a further rise in serum ALT levels occur.

often have a serum HBV DNA level of greater than 10,000 copies/mL at week 24 if their baseline level of viremia is high, even with very potent drugs. If such a patient is taking a nucleos(t)ide analog with a high genetic barrier to resistance (e.g., entecavir or tenofovir), continuing the drug is reasonable, with the expectation of an eventual response. By contrast, if such a patient is taking a nucleos(t)ide analog that is associated with either a high rate of resistance (e.g., lamivudine) or limited antiviral potency (e.g., adefovir), switching to an alternative drug is probably best.169 Unless the choice of a particular nucleos(t)ide analog is restricted or the baseline serum HBV DNA level is relatively low (less than 106 copies/mL), first-line therapy with a highly potent nucleos(t)ide analog such as entecavir or tenofovir is preferred because the rapidity of HBV DNA suppression associated with these agents makes it far less likely that drug-resistant mutants will emerge. Whenever a virologic breakthrough occurs, the patient should be questioned about adherence to therapy. If poor adherence is not a factor, genotypic testing for resistance should be done, as described earlier. If drug resistance is confirmed, the treating clinician may prescribe an alternative nucleos(t)ide analog that lacks cross resistance to the first drug or add the new drug while continuing the first. Clinical experience has indicated, however, that adding a second drug may be preferred to switching to another single agent because sequential monotherapy can result in multidrug-resistant HBV.170 When multidrug resistance occurs, combination therapy is not likely to be effective. Combination Nucleoside or Nucleotide Analog Treatment The combination of two or more nucleos(t)ide analogs may be more effective in the treatment of HBV infection than a single agent. In vitro data and studies in the woodchuck

model of hepatitis B support a role for combination therapy of hepatitis B. Combination treatment is hoped to prevent or delay the emergence of drug resistance and lead to more rapid clinical stabilization. This outcome could be particularly important for patients with decompensated cirrhosis or those in urgent need of liver transplantation. Disadvantages of combination therapy are the added cost and the potential for greater toxicity. In addition, certain combinations could theoretically lead to multidrug resistance. Somewhat surprisingly, the results of early clinical trials of combination therapy in previously untreated patients with HBeAg-positive chronic hepatitis B have shown that the combination of two nucleoside analogs (telbivudine and lamivudine) or the combination of a nucleoside analog and a nucleotide analog (lamivudine and adefovir) does not lead to greater viral inhibition during the first year of treatment than that seen with monotherapy.154,171 The reasons for the lack of apparent additive effect in these studies remain unexplained. Possibly, nucleoside analogs such as telbivudine and lamivudine compete sterically for binding to the HBV DNA polymerase or compete for phosphorylation enzymes (kinases) required for drug activation. Another possible explanation is that a measurable increase in viral suppression may be difficult to demonstrate whenever a drug with substantially less antiviral activity is added to a more potent drug (for example, when lamivudine is added to telbivudine or when adefovir is added to lamivudine). A study in which the combination of lamivudine and adefovir was compared with lamivudine alone, each given for two years, clearly demonstrated a lower rate of lamivudine resistance in the combination therapy arm (15% vs. 43%, respectively).171 On the basis of these and similar observations and the clinical experience with combination reverse transcriptase inhibitors in HIV infection, some authorities have recommended combination therapy as a first-line approach to prevent drug resistance. With the newer nucleos(t)ide agents, however, resistance occurs so infrequently that combination therapy might be better reserved for patients with clinical and laboratory features that have been associated with drug resistance, such as a high level of viremia and high body mass index.168 Combination Interferon and Nucleoside Analog Treatment From a conceptual standpoint, treatment with the combination of interferon and a nucleoside analog might prove to be more effective than either drug alone because these agents have different mechanisms of action and might also permit a shorter course of nucleoside analog therapy, thereby reducing the risk of viral resistance. Three large multicenter studies evaluated these effects in patients given a combination of pegylated interferon and lamivudine. In one study, HBeAg-positive patients received peginterferon alfa-2b with either lamivudine or placebo for one year.126 At the end of treatment, 44% of the patients who received combination therapy had lost HBeAg, whereas only 29% who received peginterferon alone had done so; however, response rates in the two groups were no longer significantly different 26 weeks after the end of treatment (35% and 36%, respectively). It is possible that the low dose of peginterferon used in this study (100 µg weekly for eight months followed by 50 µg weekly until the end of treatment) may have contributed to the relatively high relapse rates after remission. In the second study,127 HBeAg-positive patients were treated with peginterferon alfa-2a, 180 µg weekly for 48 weeks, or combined therapy with lamivudine or with lamivudine monotherapy. Although the greatest degree of HBV DNA suppression was observed in the group

1305

1306

Section IX  Liver that received combination therapy, the proportion of patients who underwent HBeAg seroconversion did not differ significantly between the two interferon-containing arms six months after completion of therapy. In the third study,128 patients with HBeAg-negative chronic hepatitis B were treated with peginterferon alfa-2a in an identical design strategy to that of the HBeAg-positive trial. Again, viral suppression was greater in the combination therapy arm, but this advantage did not translate into a higher rate of sustained virologic response. An interesting finding from the three-arm studies that evaluated peginterferon alfa-2a was that the rate of lamivudine resistance was significantly lower with combination therapy than with lamivudine monotherapy. Taken together, these three studies provide proof of concept that pegylated interferon alpha and lamivudine have additive antiviral effects during treatment. Trials of combinations of pegylated interferon and entecavir or tenofovir given for more than one year are in progress.

Antiviral Therapy in Special Populations

Pregnant Women Nucleos(t)ide analog therapy may be considered during pregnancy for two reasons: to protect the health of the mother and to prevent breakthrough HBV infection in HBV-vaccinated newborns. Several studies have shown that treatment of mothers who have high serum levels of HBV DNA during the last 4 to 12 weeks of pregnancy decreases the rate of HBV infection in newborns vaccinated at birth against HBV.172,173 These studies have been small and generally have had problematic study designs or incomplete follow-up. Furthermore, breakthrough HBV infections develop in only 5% of newborns when the newborns are given a three-dose regimen of HBV vaccine and single dose of HBIG (see later). Therefore, antiviral treatment of the mother to prevent newborn HBV infection remains highly controversial and cannot be recommended at this time. None of the current antiviral agents is licensed for use in pregnancy, and prominent warnings exist for the potential risk of harmful effects on the fetus. Telbivudine and tenofovir are considered category B drugs by the FDA (defined as a lack of animal embryologic toxicity without studies in humans). Extensive experience with tenofovir exists in HIVHBV–coinfected mothers. Lamivudine is a category B drug in HIV-infected pregnant women but a category C drug in HBV-infected women (defined as embryotoxic or teratogenic in animals without study in humans). A large amount of safety data on lamivudine in HIV-infected mothers also exists. Women who are of child-bearing age should be warned not to become pregnant while undergoing treatment with a nucleos(t)ide analog. If pregnancy occurs, the risk of drug withdrawal, including ALT flares, must be balanced against the uncertainity of harmful effects to the fetus. Because lamivudine has a long record of safety and has had the most extensive use during pregnancy in HIV-infected women, many hepatologists prefer to prescribe this agent whenever they feel compelled to treat the hepatitis B in a pregnant woman. Because defects in bone mineral density, including osteomalacia, have been described with tenofovir in HIVinfected patients, this drug seems to be a poor choice during pregnancy because of uncertainty about the effects on fetal bone maturation. The degree of risk to the fetus from the use of nucleo­ s(t)ide analogs that are licensed only for hepatitis B (adefovir, telbivudine, entecavir) is likely to be small. The Antiretroviral Pregnancy Registry in the United States has been tracking spontaneously reported maternal and fetal outcomes in women receiving oral nucleos(t)ide

drugs since 1989. Of the nearly 10,000 reported pregnancies during which the mother had received an oral nucleos(t)ide analog, 95% had HIV infection, and less than 1% had HBV infection alone. Nonetheless, the overall frequency of birth defects in the infants of these women has not been shown to be significantly different from that reported in the general U.S. population. Interferon is contraindicated during pregnancy largely because of its antiproliferative effects. In the event of pregnancy, interferon should be discontinued. Breast feeding is not recommended during the first year of the infant’s life for mothers who are undergoing antiviral therapy. Persons with Acute Hepatitis Because of the high rate (>95%) of complete immunologic recovery from acute hepatitis B, definitive recommendations about the treatment of acute hepatitis B cannot be made. Some experts recommend nucleos(t)ide analog therapy when HBeAg remains detectable in serum for more than 10 to 12 weeks because of the high likelihood of evolution to the chronic HBV carrier state without treatment. A National Institutes of Health–sponsored clinical workshop on hepatitis B proposed that persons with acute viral hepatitis complicated by an increase in INR above 1.5 and deep jaundice persisting for more than four weeks should receive antiviral therapy.174 Antiviral treatment of patients with fulminant hepatitis B is recommended in the AASLD guidelines because of the safety of nucleos(t)ide analog therapy and the need for liver transplantation in many of these patients.120 Persons with Cirrhosis Nucleos(t)ide analog therapy has been shown to be safe in patients with cirrhosis and has made a major difference in the care of patients with advanced liver disease. Interferon is contraindicated in patients with even mildly decompensated cirrhosis because immune-mediated flares of serum ALT levels may occur and may be associated with further clinical deterioration. Also, serious infections have been reported in treated patients.175 Practice guidelines of the AASLD suggest that nucleos(t)ide analog therapy is preferred in all cases of HBV-related cirrhosis because of a greater chance of dose-limiting side effects and clinical worsening in the event of an ALT flare with interferon. Nevertheless, some studies have shown that patients with stable, well compensated cirrhosis can be treated safely and actually may have a higher rate of virologic response when compared with patients without cirrhosis.176,177 Treatment of patients with HBV cirrhosis needs to be individualized, and all of those treated with interferon should be carefully monitored. Persons with Human Immunodeficiency Virus–Hepatitis B Virus Coinfection With improved control of HIV disease with HAART, liver disease has emerged as one of the leading causes of death in patients with HIV.178 Antiviral therapy for hepatitis B should be considered for all HIV-HBV–coinfected patients with evidence of liver disease, irrespective of the CD4 count. In coinfected patients not requiring HAART, therapy for HBV should be based on agents with no HIV activity such as adefovir or pegylated interferon.179 Entecavir treatment is associated with a decline in HIV RNA levels; thus, it also is not recommended for use in patients who are not receiving concomitant HIV treatment.180 In patients with CD4 counts less than 350/mm3, the use of agents with dual anti-HIV and anti-HBV activity should be considered. Combination therapy with either emtricitabine and tenofovir or

Chapter 78  Hepatitis B and D lamivudine and tenofovir should ideally be used to avoid or delay the development of antiviral resistance. Persons with Hepatitis B Virus–Hepatitis C Virus Coinfection When compared with monoinfected patients, HBV-HCV– coinfected patients tend to have more severe liver injury and a higher probability of cirrhosis.181 Limited data are available, however, to define the best approach to treatment in this group of patients. In most instances, one virus, often HCV, is dominant through the process of viral interference. The typical patient is positive for HCV RNA but negative for HBV DNA in serum. Close monitoring has been recommended before treatment is initiated, however, because some patients exhibit alternating viremia. In a prospective clinical trial of 19 patients with combined HBV-HCV infection, all were positive for HCV RNA, and only 5 were positive for HBV DNA prior to treatment. A high rate of sustained virologic response (74%) was observed after a 48-week course of pegylated interferon alpha and ribavirin (see Chapter 79), and two of the five HBV DNA-positive patients also had a virologic response. Unfortunately, HBV DNA became detectable again in four patients who were initially negative, suggesting that a risk exists that HBV may reactivate with eradication of HCV.182 The optimal therapy for patients who are positive in serum for both HBV DNA and HCV RNA is also unclear. In such instances, the author has had success in treating both infections simultaneously with a combination of a nucleos(t)ide analog, pegylated interferon alpha, and ribavirin.

Unresolved Issues

Major advances in antiviral therapy have occurred since 2000, but many unresolved issues remain. Although clearance of HBsAg from serum occurs in some patients after a relatively short course of interferon therapy, interferon is used infrequently because of its adverse effects. Nevertheless, data based on serial monitoring of HBsAg and HBeAg concentrations during therapy with pegylated interferon strongly suggest that failure to achieve a decline in serum concentrations of these viral parameters at key treatment intervals (12 and 24 weeks) has a high negative predictive value for a sustained response and could provide stopping rules similar to those used for the treatment of HCV infection (see Chapter 79). Clinical trials also have suggested that serial monitoring of the HBsAg concentration allows determination of the appropriateness of extending the duration of interferon therapy in patients with HBeAg-negative hepatitis, a condition in which HBsAg clearance is the most reliable endpoint. Multinational trials to evaluate the clinical utility of these markers to monitor virologic response during therapy with interferon are in progress. Commercial assays for HBsAg or HBeAg concentrations are not yet available in the United States but are widely available in Europe and elsewhere in the world and are likely to become available in the United States. Whether 24 weeks of pegylated interferon therapy is preferable to 48 weeks of therapy in patients with a favorable HBV genotype (A or B) or in those who demonstrate an early loss of HBeAg in serum is unknown and is under study. Because combination therapy with pegylated interferon alpha and lamivudine has been associated with enhanced viral suppression, studies of pegylated interferon and newer, more potent nucleos(t)ide analogs are also is progress. The excellent tolerability of oral antiviral therapy makes long-term treatment of hepatitis B feasible. Long-term treatment could be especially important for middle-aged or older patients with high levels of viremia who acquired infection

early in life. Many such patients have normal or only slightly elevated serum ALT levels. For such patients, the intent of treatment would be to keep the serum HBV DNA at the lowest possible level to prevent disease complications and to prolong life, rather than to achieve potentially shortterm endpoints such as HBeAg seroconversion and normalization of serum ALT levels. Additional studies on survival and complication rates after prolonged treatment will determine whether these goals can be achieved, but a placebocontrolled design in patients with active disease will be difficult because data already exist to suggest that the progression of disease is slowed over a few years when HBVinfected patients with advanced fibrosis are treated. The lessons of HIV treatment with regard to the need for combination therapy to prevent drug resistance should not be ignored, but the availability of potent nucleos(t)ide analogs with a high genetic barrier to resistance makes the need for combination therapy a less pressing issue at this time. Because of the low resistance rates associated with these agents, study of the relative benefits of combination nucleos(t)ide analog therapy and monotherapy will require large numbers of patients followed for a long period of time. Retrospective analysis of the data from drug registration trials and prospective evaluation in large clinical trials are needed to define further the common predictors of drug resistance for each drug. Once this is done, combination therapy could be selected for those patients who are more likely to develop resistance with a given nucleos(t)ide analog, thereby limiting the use of combination therapy to selected patients. Finally, the availability of increasingly potent nucleo­ s(t)ide analogs has inadvertently led to diminished emphasis on the importance of the host immune response in achieving therapeutic endpoints. Pegylated interferon is currently the only approved therapeutic agent with immunomo­dulatory as well as viral inhibitory properties. The development and validation of practical tools to measure the immune response before, during, and after interferon therapy would provide a major step forward in understanding the relationship between the host immune response and the viral response, including HBsAg clearance.

PREVENTION

Immunoprophylaxis against HBV is of two types: passive immunization using HBIG and active immunization using inactive HBsAg. Active immunization gives long-term immunity, whereas passive immunization confers only immediate and short-lived protection.

Hepatitis B Immunoglobulin

HBIG is prepared from plasma that is known to contain high titers of anti-HBs. Numerous clinical trials have established the efficacy of HBIG in preventing HBV infection in highrisk persons, such as hemodialysis patients, sexual partners of patients with hepatitis B, and newborn infants of HBsAgpositive mothers. HBIG licensed in the United States has an anti-HBs titer of 1 : 100,000. In Europe, several preparations of HBIG with different concentrations and pharmacokinetic properties are available. HBIG is safe, although rare anaphylactic reactions can occur. Myalgias, rash, and arthralgias have also been reported and are believed to result from formation of antigen-antibody complexes.

Hepatitis B Vaccine

Currently marketed HBV vaccines make use of DNA recombinant technology by introducing the gene for HBsAg (S

1307

1308

Section IX  Liver gene) into the genome of the yeast Saccharomyces cerevisiae. The two vaccines available in the United States are Recombivax HB (Merck, licensed in 1986) and Engerix-B (GlaxoSmithKline, licensed by SmithKline Beecham in 1989). Aluminum hydroxide is used as an adjuvant in both vaccines. Because thimerosal, a preservative used in the vaccines, contains mercury, thimerosal-free vaccines have become available, especially for use in infants. The HBV vaccine is administered intramuscularly in the deltoid muscle of adults and the anterolateral thigh of infants and neonates. The vaccines induce HBsAg-specific helper T cells and T cell–dependent B cells to produce neutralizing antibody against the “a” epitope (amino acid sequence 124 to 148) of HBsAg as early as two weeks after the first injection.183 HBV vaccines are highly efficacious in preventing HBV infection.184 Because the vaccines contain HBsAg only, anti-HBs is the sole antibody produced. Consequently, a vaccinee who tests positive for anti-HBc after vaccination should be considered to have had a subclinical HBV infection. The vaccines typically achieve an anti-HBs titer greater than 100 mIU/mL. Antibody titers greater than 100 mIU/mL confer 100% protection against HBV infection, and a lower antibody titer (up to 10 mIU/mL) is seroprotective in most instances. Peak antibody titers and persistence of antibody levels vary among different persons. The titers drop steadily over the first two years after vaccination, sometimes to levels less than 10 mIU/mL. Two studies in different populations have demonstrated that anti-HBs titers decrease to nonprotective levels in at least 25% to 50% of recipients over a period of 5 to 10 years.185 Although protective anti-HBs response rates after HBV vaccination typically exceed 90%, a number of factors can impede an adequate antibody response. Smoking, obesity, injection into the buttock, chronic liver disease, presence of HLA-DR3, DR7, and DQ2 alleles, absence of the HLA-A2 allele, and extremes of age may be associated with reduced immunogenicity. Such “hyporesponders” may benefit from a higher dose of vaccine. Response rates also are lower in immunocompromised patients, such as transplant recipients, patients receiving chemotherapy, and those with endstage liver disease. Only 50% to 60% of patients on hemodialysis respond adequately to vaccination. Therefore, patients with chronic kidney disease should be vaccinated early in the course of their disease, before renal disease progresses, to ensure optimal response to vaccination.186 Five percent to 8% of HBV vaccine recipients do not achieve detectable anti-HBs levels (“nonresponders”). Studies conducted mostly in animals indicate that intradermal injection of the vaccine may produce a stronger humoral and cellular immune response than conventional intramuscular administration.187,188 Intradermal injection, by recruiting “professional” dendritic cells, stimulates primary MHC class I- and class II–restricted T cell responses. In one study, intradermal vaccination resulted in protective anti-HBs responses in nonresponders to intramuscular adminis­ tration.188 Repeated dosing with intradermal vaccination (5 µg every two weeks to provide an anti-HBs titer of 1000 mIU/mL or greater or a total of 52 doses) has resulted in a protective antibody response rate of nearly 100% in patients undergoing long-term hemodialysis. At present, intradermal HBV vaccination has not been recommended officially—in part because of concerns about standardization of the technique for intradermal delivery. Because HBV vaccination results in strong immunologic memory capable of preventing infection even in patients with low or undetectable antibody titers, no role exists for

Table 78-6  High-Risk Groups for Whom Hepatitis B Virus (HBV) Vaccination Should Be Considered Heath care workers Hemodialysis patients Household contacts and sexual partners of HBV carriers or patients with acute hepatitis B Injection drug users Inmates of correctional facilities International travelers to areas endemic for HBV who may have intimate contact with the local population or take part in medical activities Men who have sex with men Patients who are likely to require multiple transfusions with blood or blood products Patients with chronic liver disease (other than chronic hepatitis B) Potential organ transplant recipients Public safety workers with likelihood of exposure to blood Sexually active heterosexual men and women, if they have more than one partner Staff and clients of institutions for developmentally disabled

a booster vaccine dose in immunocompetent adults and children. Current recommendations include booster doses only for patients undergoing hemodialysis, in whom antiHBs titers should be tested annually and a booster dose given if the titer is lower than 10 mIU/mL.189 No serious side effects of the HBV vaccine have been reported. The frequency of neurologic diseases such as aseptic meningitis and Guillain-Barré syndrome is not increased in vaccine recipients. Targeted High-Risk Groups Table 78-6 lists the high-risk groups for whom HBV vaccination is recommended. Targeted vaccination has not achieved its objective in certain high-risk groups, such as injection drug users, but has achieved great success among health care workers and newborns. The CDC has extended its original recommendations for routine HBsAg screening to include persons born in countries in which the prevalence of hepatitis B exceeds 2%.190 This new recommendation will facilitate identification of susceptible persons who are in need of vaccination and those in need of antiviral therapy. Vaccination Schedule The doses of currently available HBV vaccines and recommendations for the schedules of administration are shown in Table 78-7. The typical vaccination schedule is zero, one, and six months. The first two doses have no effect on the final anti-HBs titer. The third dose acts as a booster to achieve a high anti-HBs titer. In immunocompromised patients and patients undergoing hemodialysis, four vaccine doses are recommended, with the fourth dose given to ensure the highest possible anti-HBs titer. If vaccination is interrupted, the second dose should be administered as soon as possible after the first.190 If the third dose is not given on schedule, it should be given at least two months after the second dose. The HBV vaccine is currently administered to all children and infants as a part of the universal immunization program. Combination HBV vaccines with diphtheria-pertussistetanus (DPT) and Haemophilus influenzae type B (Hib) (DTPw-HB/Hib), the vaccines in current use for immunization of infants, do not reduce the immunogenicity of any of the components of HBV infection.191 Adolescents who have not been vaccinated in infancy or childhood should also be vaccinated.

Chapter 78  Hepatitis B and D Postexposure and Perinatal Prophylaxis Table 78-8 summarizes recommendations for prevention of perinatal transmission of HBV. Table 78-9 lists recommendations for prophylaxis after exposure to a known HBsAg-positive source. Postexposure vaccination should be considered for any percutaneous, ocular, or mucous membrane exposure. The type of immunoprophylaxis is determined by the HBsAg status of the source and the vaccination-response status of the exposed person. Bivalent Vaccine A combined HAV and HBV vaccine has been licensed commercially (TWINRIX, GlaxoSmithKline, Research Triangle Park, NC) and has been shown to be highly immunogenic and protective against both infections (see Chapter 77). This vaccine offers ease of administration for persons at increased risk of both HAV and HBV infection (e.g., world travelers or men who have sex with men) and in patients with underlying chronic liver disease.192 HBV Escape Mutants and Implications for Immunization As described earlier, mutations in the HBV genome that encodes HBsAg can result in mutant HBV virus strains that can escape neutralization by anti-HBs. The mutation involves the “a” determinant and has shown decreased binding to monoclonal anti-a antibodies. Such mutants have been reported worldwide and are particularly common in areas endemic for HBV. These mutant viruses account for some instances in which the HBV vaccine has failed to prevent perinatal transmission. The frequency of this mutation is currently low, but in the future this HBV mutant could emerge as an important threat, in which case HBV vaccines may have to incorporate the mutant antigen to remain effective.

†

Infants and children age <11 yr Children age 11-19 yr Adults (≥20 yrs) Hemodialysis patients Immunocompromised patients

2.5 µg

Hepatitis D (delta) virus (HDV) was discovered by Rizzetto and associates in 1977, as a unique nuclear antigen in the hepatocytes of patients infected with HBV.193 The antigen was identified subsequently as a novel pathogen and was linked to severe chronic hepatitis B and fulminant HBV infection.

EPIDEMIOLOGY

HDV is distributed globally with wide variations in prevalence. At least 5% of HBV carriers worldwide are estimated to be infected with HDV, and therefore, the overall burden of HDV infection is between 15 and 20 million cases. The highest prevalence is seen in South America and the Mediterranean basin. The prevalence is low in Northern Europe and North America, where HDV infection is confined to injection drug users. The rate of infection among HBsAgpositive blood donors in the United States has been found to be 3.8%.194 The incidence of transfusion-associated HDV infection has been declining steadily because of HBV vaccination and screening of donor sera for HBsAg. In fact, many epidemiologists believe that the epidemic of HDV, which started in the 1970s, is coming to an end. For example, epidemiologic data from Italy show that the prevalence of HDV infection in HBsAg carriers by 2000 was 8.3%, compared with 25% in the early 1970s.195-197 HDV infection remains an important problem among injection drug users, however.198

Table 78-9  Postexposure Prophylaxis of Hepatitis B if the Source Is HBsAg Positive VACCINATION STATUS OF EXPOSED PERSON

Table 78-7  Recommended Dosing for the Currently Available Hepatitis B Vaccines* RECOMBIVAX HB (10 µg/mL)

HEPATITIS D

Unvaccinated

ENGERIX-B (20 µg/mL)

Previously vaccinated:   Known responder*   Known nonresponder

10 µg

5 µg 10 µg 40 µg (1.0 mL)‡ 40 µg (1.0 mL)‡

20 µg 20 µg 40 µg (2.0 mL)§ 40 µg (2.0 mL)§

*The standard schedule is 0, 1, and 6 months. † Infants born to hepatitis B surface antigen-negative mothers. ‡ Special formulation (40 µg/mL). § Two 1.0-mL doses administered at one site in four-dose schedule (0, 1, 2, 6 months).

  Antibody response unknown

IMMUNE PROPHYLAXIS HBIG (0.06 mL/kg) and initiate hepatitis B vaccine series No treatment HBIG × 2 doses (one month apart) OR HBIG × 1 dose and initiate revaccination Test for anti-HBs If adequate*: No treatment If inadequate†: HBIG × 1 dose and give vaccine booster dose

*Anti-HBs titer ≥10 mIU/mL. † Anti-HBs titer <10 mIU/mL. Anti-HBs, antibody to hepatitis B surface antigen; HBIG, hepatitis B immune globulin; HBsAg, hepatitis B surface antigen.

Table 78-8  Hepatitis B Prophylaxis of Infants Born to Hepatitis B Surface Antigen-Positive Mothers Vaccine: Give One of the Following: AGE OF THE INFANT

HBIG

VACCINATION

RECOMBIVAX HB

ENGERIX-B

Within 12 hours of birth 1 month 6 months†

0.5 mL IM* None None

First dose Second dose Third dose†

5 µg (0.5 mL) 5 µg (0.5 mL) 5 µg (0.5 mL)

10 µg (0.5 mL) 10 µg (0.5 mL) 10 µg (0.5 mL)

HBIG, hepatitis B immune globulin; IM, intramuscular. *HBIG should be administered at a site different from that used for the vaccine. † If four doses of vaccine are administered, the third dose is given at 2 months, and the fourth dose is given at 12-18 months.

1309

1310

Section IX  Liver Among the three genotypes of HDV (I to III), genotype I is the most prevalent and is the most common genotype in Mediterranean countries, Africa, Europe, and North America.199,200 Different subtypes within this genotype may exist in certain parts of Africa. Genotype II is reported mostly in Japan and Taiwan and is associated with milder liver disease than that seen with genotype I.201 Genotype III has been isolated from epidemics in South America.202 Different genotypes of HDV may interact variably with different HBV genotypes. Whether the interaction between HDV and HBV genotypes specifically increases the severity of HDV infection is unclear, although infection with HDV genotype III and HBV genotype F is reported to cause severe hepatitis. The mode of HDV transmission is linked closely to that of HBV transmission, with the parenteral route being the most efficient. Sexual transmission of HDV has been reported, and familial clustering of cases has been seen in endemic areas.

HDV or an enhanced immune response against the two viruses. Direct cytotoxicity of HDV has been questioned on the basis of studies in transgenic mice. Mice expressing either HDAg-L or HDAg-S show no evidence of hepatocyte injury. The lack of a direct cytotoxic effect also is supported by the observation that liver transplant recipients who express HDAg in their allografts do not manifest evidence of cellular damage.207 Instead, the pathogenic mechanism of HDV-induced liver damage is most likely related to the immunologic response to HDV. The occurrence of classic necroinflammatory changes in the liver and several auto­ antibodies, such as antibodies to liver-kidney microsome (anti-LKM), thymocytes, and nuclear lamin C, also suggest a role for immune-mediated liver injury. One fact is certain: The ability of HDV to cause hepatic necrosis is determined by expression of HBV, as illustrated after liver transplantation, when HDV infection becomes pathogenic only if HBV infection also recurs.207

VIROLOGY

DIAGNOSIS

HDV is a unique agent that bears no similarity to other transmissible agents that infect animals. In fact, the 1.7-kb single-strand HDV RNA genome shares several features with plant viroids, such as intramolecular base pairing and autocatalytic RNA segments.203 Unlike plant viroids, however, HDV RNA encodes a protein, hepatitis delta antigen (HDAg). The virion consists of the HDV genome complexed with approximately 70 copies of HDAg in an envelope protein composed of lipids and HBsAg. The protein envelope that is contributed by HBV protects the HDV RNA-HDAg complex. The protein envelope is not required for replication of HDV and is the only helper function provided by HBV. Once HDV with its HBV envelope protein enters the host, the HDV RNA-HDAg complex migrates to the nucleus. Viral replication then proceeds in the nucleus according to a double-rolling model, aided possibly by host DNAdependent RNA polymerase.204 During translation, two forms of HDAg (encoded by the same regions of RNA) are produced, a short form (HDAg-S) and a long form (HDAg-L). HDAg-L has 19 to 20 more amino acids than HDAg-S. The additional amino acids in HDAg-L are incorporated by a process of RNA editing, another unique aspect of the HDV genome.205 Interestingly, HDAg-S and HDAg-L have opposite effects on viral replication; HDAg-S acts as a facilitator, and HDAg-L as an inhibitor.206 The extent of RNA editing determines the amount of HDAg-L formed and, consequently, influences the rate of replication. In states of high replication, only HDAg-S is produced. Ultimately, the intracellular ratio of HDAg-S to HDAg-L determines the rate of replication, assembly, and transport from infected hepatocytes. Because it has a unique genome, HDV is classified in a separate genus of the Deltaviridae family. No other virus has been identified in this genus. The current consensus is that HDV is a satellite virus. Satellite viruses are subviral particles that carry a distinct nucleic acid, usually RNA, that requires a helper virus for transmission and multiplication. In addition, the nucleic acid of satellite viruses is distinct from the nucleic acid of helper viruses. No other animal virus has been identified as a satellite virus. HDV is not a viroid, as previously believed because HDV RNA codes for a structural protein, HDAg.

PATHOGENESIS

The pathogenic mechanisms of HDV hepatitis remain poorly understood. Because HBV is not known to be directly cytotoxic, the severity of combined infection with HBV and HDV may be attributed either to a direct cytotoxic effect of

The most useful markers of HDV infection include HDAg, antibody to HDAg (anti-HDV), HDV RNA, and immunohistochemical staining of HDAg in liver tissue. Detection of HDV RNA by reverse transcriptase PCR amplification (RTPCR), with a detection limit of 50 to 100 copies/mL, is the most reliable diagnostic technique, with nearly 100% sensitivity. HDV RNA is the earliest marker to appear during the course of HDV infection and may be seen in the absence of other markers.208 Higher levels of HDV RNA in serum may be associated with more severe disease. The level of HDV RNA in serum is a reliable marker for monitoring the efficacy of treatment and documenting viral eradication. HDV RNA also can be detected in liver cells by hybridization techniques, which are generally less sensitive than RT-PCR. The HDV genomic product, HDAg, is another marker of HDV infection. HDAg can be demonstrated in hepatocytes by immunohistochemical staining, but the reliability of this method decreases as the disease becomes chronic. Measurement of HDAg in serum is also problematic because of the presence of high titers of neutralizing antibodies, which interfere with detection of HDAg. The most readily available marker of HDV infection has been anti-HDV. Anti-HDV does not confer protection against HDV. Either IgM anti-HDV or total anti-HDV, which is composed of both IgM and IgG anti-HDV, can be detected. IgM anti-HDV appears in serum at the time of acute infection, and IgG anti-HDV develops later in the course. IgM antiHDV often persists as the disease progresses to chronicity and is detectable in high titers in patients with chronic HDV infection. It is frequently regarded as a marker of serious liver damage.209 As the infection evolves from the acute to the chronic phase, the type of IgM antibody also changes from a monomeric (S) form to a multimeric (19S) form. IgG anti-HDV persists for a long time in immunocompetent persons and may indicate chronic or previous HDV infection. Some patients with IgG anti-HDV may not have active infection and test negative for HDV RNA.210

NATURAL HISTORY

Because of the obligate relationship of HDV to HBV infection, the natural history of HDV infection depends on the clinical course of HBV. Two distinct types of HDV infections are possible: One is a coprimary infection (or coinfection), in which infection of HBV and HDV occurs simultaneously, and the other is superinfection, in which HDV infection is superimposed on established chronic HBV infection.

Chapter 78  Hepatitis B and D Serologic results

Superinfection (previously HBsAg+)

Clinical features

Prognosis

HBsAg+ HBV DNA+ IgM anti-HBc –* IgM anti-HDV+

Abrupt increase in ALT from baseline Decompensation of previously compensated liver disease

Progression to cirrhosis

HBsAg+ HBV DNA+ IgM anti-HBc+* IgM anti-HDV+

Acute hepatitis Double peak in ALT (as HDV is established after HBV)

Resolution in most cases as HBV resolves

HDV

Coinfection (previously HBsAg –)

* Distinguishing feature between superinfection and coinfection. Figure 78-7.  Serologic results, clinical features, and prognosis of hepatitis D superinfection and coinfection with HDV and HBV. ALT, serum alanine aminotransferase level; anti-HBc, antibody to hepatitis B core antigen; anti-HDV, antibody to hepatitis D virus; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HDV, hepatitis D virus; IgM, immunoglobulin M.

Early epidemiologic studies suggested that HDV infection aggravates the severity of HBV infection in coprimary infection, but subsequent reports have disputed this claim.211 In a European multicenter study on prognostic factors in 366 patients with chronic hepatitis B and compensated cirrhosis, HDV infection did not influence the prognosis.212 In a long-term follow-up study of 302 patients with chronic hepatitis B (76 with HDV infection), HDV infection was not an independent predictor of mortality.213 Therefore, HDV infection appears to have a varying influence on the course of hepatitis B and is not necessarily associated with severe hepatitis. The severity of HDV infection may vary with the frequency of HDV in a population, with the level of HBV viremia, and with interactions between specific HBV and HDV genotypes. Coprimary infection is seen most often in injection drug users (Fig. 78-7). Because HBV infection resolves in a majority of patients, HDV also disappears in most patients, and the risk of chronicity after coprimary infection is less than 5%. Some data suggest, however, that coin­ fection with HDV enhances the risk of fulminant hepatitis B.214 Superinfection of HDV in an HBV carrier can lead to severe hepatitis and acute decompensation of preexisting liver disease. Affected patients often express a high level of HDV viremia because high serum levels of HBsAg readily protect the replicating HDV genome. Superinfection may also coincide with a decline in serum HBV DNA levels because HDV replication inhibits HBV replication.215 In a study involving 185 patients with HDV superinfection, HDV RNA was detectable in 63 of 64 patients with acute HDV infection, but HBV DNA was detectable in only 40% of patients.215 Rarely, HDV superinfection may lead to disappearance of HBsAg and appearance of anti-HBs.215 In contrast to coinfection, chronic HDV infection develops frequently after HDV superinfection. HDV superinfection evolves to chronic HDV infection in 70% of patients and is characterized by persistent HDV viremia and detectable HDV RNA in serum. Although the clinical course of chronic HDV infection varies, persistent replication of HDV and HBV often leads to progressive hepatitis and cirrhosis within a few years. More rapid clinical progression leading to end-stage liver disease within two years may be seen in some injection drug users; HCC also may develop.216 The clinical course of a triple infection with HBV, HDV, and HCV is usually dominated by the HCV infection.

Affected patients often have a severe episode of acute hepatitis at the time of HDV or HCV superinfection, but the chronic stage is slowly progressive and does not differ much from that of chronic HDV and HBV infection.

CLINICAL FEATURES

HDV and HBV coinfection typically manifests as self-limited acute hepatitis. Some patients may demonstrate a double peak in serum aminotransferase levels because of a delay in HDV replication after HBV replication. Markers detectable in serum at the time of acute hepatitis include IgM anti-HBc, IgM anti-HDV, HDV RNA, HBV DNA, and HBsAg. Acute hepatitis usually resolves in a few weeks, with a gradual return of liver biochemical test levels to normal. As the infection resolves, HDV RNA and HBV DNA titers steadily decrease, and anti-HBs appears after the disappearance of HBsAg. Occasionally, IgM anti-HDV may persist after anti-HB appears and serum aminotransferase levels return to normal. HDV superinfection manifests clinically as acute hepatitis in an otherwise stable chronic HBV carrier. Clinically, HDV superinfection can mimic a spontaneous flare of chronic HBV infection. These two diagnostic possibilities can be differentiated easily because patients with HDV superinfection have markers of HDV infection, such as HDV RNA and IgM anti-HDV, in serum. IgM anti-HDV is seen in both coinfection and superinfection and is not useful for differentiating the two. Determining whether an HDV infection is a superinfection or coinfection depends on whether the HBV infection is chronic or acute, and the distinguishing serologic feature of coinfection is the presence of IgM anti-HBc in serum.

TREATMENT

Despite developments in the treatment of HBV monoinfection, results of therapy for HDV-HBV infection have been disappointing. Hepatitis D is the least common cause of chronic viral hepatitis worldwide, but it is the most severe form. The only therapeutic option currently available is interferon alpha, the efficacy of which is related to the dose and duration of treatment. Nucleos(t)ide analogs, currently the mainstay of treatment for HBV infection, are not effective in HDV infection. The lack of efficacy of nucleos(t)ide analogs may be explained in part by the observation that nucleos(t)ide analogs seldom lead to disappearance of HBsAg, the only HBV protein that is required by HDV. The

1311

1312

Section IX  Liver rate of sustained HDV clearance after a one-year course with high doses of conventional interferon alpha (9 million units three times a week) is low (20%).217 Better results have been reported with peginterferon alpha, both in interferon-naïve patients and previous nonresponders to conventional interferon-α. Doses of 1.5 ug/kg of peginterferon alfa-2b and 180 ug of peginterferon alfa-2a have been used successfully.217 Current recommendations from an Italian workshop include the use of pegylated interferon alpha for 48 to 72 weeks.218 Because the ultimate goal of treatment is the eradication of both HBV and HDV, some authorities have advocated that therapy be continued as long as possible in responders and preferably until the loss of HBsAg occurs.217 Nucleos(t)ide analog therapy used in conjunction with interferon does not improve response rates compared with interferon alone.219 Reliable predictors of a long-term response have not been identified, although patients with disease of short duration or without cirrhosis are more likely to respond, thereby underscoring the importance of early diagnosis and treatment.

PREVENTION

Because the ability of HDV to infect a host depends on the preexistence of HBsAg, vaccination against HBV confers protection against HDV. Groups of persons who exhibit a high rate of HDV infection, such as injection drug users, should be targeted for vaccination.

KEY REFERENCES

Buster EH, Flink HJ, Cakaloglu Y, et al. Sustained HBeAg and HBsAg loss after long-term follow-up of HBeAg-positive patients treated with peginterferon a-2b. Gastroenterology 2008; 135:459-67. (Ref 130.) Chen CJ, Yang HI, Su J, et al. Risk of hepatocellular carcinoma across a biological gradient of serum hepatitis B virus DNA level. JAMA 2006; 295:65-73. (Ref 57.) Doo E, Liang TJ. Molecular anatomy and pathophysiologic implications of drug resistance in hepatitis B virus infection. Gastroenterology 2001; 120:1000-8. (Ref 34.)

Farci P, Chessa L, Balistrieri C, et al. Treatment of chronic hepatitis D. J Viral Hepatitis 2007; 14(Suppl 1):58-63. (Ref 217.) Fattovich G. Bortolotti F, Donato F. Natural history of chronic hepatitis B: Special emphasis on disease progression and prognostic factors. J Hepatol 2008; 48:335-52. (Ref 17.) Hoofnagle JH, Doo E, Liang TJ, et al. Management of hepatitis B: Summary of a clinical research workshop. Hepatology 2007; 45:105675. (Ref 174.) Iloeje UH, Yang HI, Su J, et al. Predicting liver cirrhosis risk based on the level of circulating hepatitis B viral load. Gastroenterology 2006; 130:678-86. (Ref 56.) Keeffe EB, Dieterich DT, Han SH, et al. A treatment algorithm for the management of chronic hepatitis B virus infection in the United States: 2008 Update. Clin Gastroenterol Hepatol 2008; 6:1315-41. (Ref 123.) Keeffe EB, Zeuzem S, Koff RS, et al. Report of an international workshop: Roadmap for management of patients receiving oral therapy for chronic hepatitis B. Clin Gastroenterol Hepatol 2007; 5:890-7. (Ref 169.) Lai CL, Yuen MF. The natural history and treatment of chronic hepatitis B: A critical evaluation of standard treatment criteria and end points. Ann Intern Med 2007; 147:58-61. (Ref 60.) Lau GK, Piratvisuth T, Luo KX, et al. Peginterferon alfa-2a, lamivudine, and the combination for HBeAg-positive chronic hepatitis B. N Engl J Med 2005; 352:2682-95. (Ref 127.) Liaw YF, Sung JJ, Chow WC, et al. Lamivudine for patients with chronic hepatitis B and advanced liver disease. N Engl J Med 2004; 351:152131. (Ref 61.) Lok AS, Lai CL, Leung N, et al. Long-term safety of lamivudine treatment in patients with chronic hepatitis B. Gastroenterology 2003; 125:1714-22. (Ref 35.) Lok AS, McMahon BJ. AASLD Practice Guidelines. Chronic hepatitis B. Hepatology 2007; 45:507-39. (Ref 120.) Marcellin P, Lau GK, Bonino F, et al. Peginterferon alfa-2a alone, lamivudine alone, and the two in combination in patients with HBeAgnegative chronic hepatitis B. N Engl J Med 2004; 351:1206-17. (Ref 128.) Perrillo RP. Acute flares in chronic hepatitis B. The natural and unnatural history of an immunologically mediated liver disease. Gastroenterology 2001; 120:1009-22. (Ref 78.) Zoulim F, Perrillo R. Hepatitis B: Reflections on the current approach to antiviral therapy. J Hepatol 2008; 48:S2-19. (Ref 168.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

79 Hepatitis C Jacqueline G. O’Leary and Gary L. Davis

CHAPTER OUTLINE Virology  1313 Structure  1313 Genomic Organization  1314 Viral Replication and Life Cycle  1314 Genotypes and Quasispecies  1316 Epidemiology  1317 Incidence and Prevalence  1317 Transmission  1317 Pathogenesis  1318 Viral Mechanisms  1319 Immune-Mediated Mechanisms  1319 Clinical Features  1319 Acute and Chronic Hepatitis C  1319 Extrahepatic Manifestations  1320 Diagnosis  1321 Indirect Assays  1321 Direct Assays  1321 Selection of Serologic and Virologic Tests  1322 Liver Biopsy and Noninvasive Assessment of Fibrosis  1322 Natural History  1324 Factors Associated with Progression of Chronic Hepatitis C  1325

Hepatitis C virus (HCV) infects 170 million people worldwide and 1.6% of the population of the United States.1-4 Unfortunately, HCV successfully evades the host immune response in 55% to 85% of acutely infected persons, thus leading to chronic infection. The natural history of hepatitis C varies greatly; reasons for this heterogeneity remain incompletely understood but are related to both viral, host, and environmental factors. Chronic HCV infection can lead to cirrhosis and hepatocellular carcinoma. The incidence of these complications has risen dramatically in the 2000s but is expected to decline over the next 20 years.3 In fact, whereas HCV-related mortality increased dramatically after 1995, it has reached a plateau since 2002.5 Complications of HCV-related cirrhosis are currently the leading indication for liver transplantation in the United States and Europe. Chronic hepatitis C is the only chronic viral infection that can be cured by antiviral therapy. Currently, 40% to 50% of patients infected with HCV genotype 1 who tolerate fulldose treatment with pegylated interferon and ribavirin achieve a sustained virologic response (SVR) to treatment, defined as absence of HCV RNA in serum six months after discontinuation of treatment; an SVR is almost always associated with a durable eradication of the virus.6,7 From 70% to 80% of genotype 2- and 3-infected patients achieve an SVR. Substantial progress in understanding the mechanisms of virus entry into the hepatocyte, replication, and the host immune response has led to the development of new therapeutic agents that target the steps in the viral life

Patients with Persistently Normal Serum Aminotransferase Levels  1326 Immunocompromised Patients  1326 Prevention  1327 General Measures and Treatment of Acute Infection  1327 Immunoprophylaxis  1328 Treatment  1328 Goals  1328 End Points  1328 Drugs  1329 Efficacy  1329 Factors That Predict a Sustained Virologic Response  1331 Indications and Contraindications  1331 Monitoring and Safety  1332 Retreatment  1333 HIV-HCV Coinfected Patients  1333 Liver Transplant Recipients  1334 Future Therapies  1334

cycle. New agents under investigation promise to improve the SVR further when combined with interferons and ribavirin.

VIROLOGY STRUCTURE

The HCV virion has been visualized by electron microscopy and is an enveloped virus 50 nm in diameter.8 The two envelope proteins, E1 and E2, heterodimerize and assemble into tetramers, which create a smooth outer layer. This layer has a “fishbone” configuration with icosahedral symmetry. The envelope proteins are anchored to a host cell–derived lipid bilayer envelope membrane that surrounds the nucleocapsid. The nucleocapsid is believed to be composed of multiple copies of the core protein and forms an internal icosahedral viral coat that encapsulates the genomic ribonucleic acid (RNA). HCV circulates in various forms in the serum of an infected host, including (1) virions that are bound to very-low-density and lowdensity lipoproteins and appear to represent the infectious fraction; (2) virions bound to immunoglobulins; and (3) free virions. In addition, viral particles that exhibit phys­ icochemical, morphologic, and antigenic properties of nonenveloped HCV nucleocapsids have been detected in plasma.9

1313

1314

Section IX  Liver GENOMIC ORGANIZATION

tion.14,15 Understanding of the mechanisms of viral replication comes from studies in chimpanzees, extrapolation of the mechanisms used by other flaviviruses, and infection of cells in vitro with subgenomic replicons (segments of HCV RNA that are capable of amplification and synthesis of viral proteins but that do not produce mature viruses). Early events in viral binding to the hepatocyte surface are still not completely understood (Fig. 79-1). HCV entry involves the attachment of envelope proteins E1 and E2 to cell surface molecules. The expression and function of CD81, a member of the tetraspan superfamily, are essential for HCV entry into hepatocytes.16 In addition, human scavenger receptor class B type 1 (SR-B1), a selective importer of cholesteryl esters from high-density lipoproteins (HDL) into cells, has been shown to interact with E2 and is essential for HCV entry.17 Whereas CD81 and SR-B1 are required early in the process of viral entry, claudin-1 (CLDN1), a tight junction component that is highly expressed on hepatocytes, is required later in the cell entry process.18 Heparin sulfated proteoglycans have also been shown to be essential

HCV is a single-stranded positive-sense RNA virus that belongs to the Flaviviridae family and has been classified as the sole member of the genus Hepacivirus.10 The genome of HCV contains approximately 9600 nucleotides with an open reading frame (ORF) that encodes one large viral polypeptide precursor of 3008 to 3033 amino acids. The HCV ORF is flanked upstream by a 5′ untranslated region (UTR) that functions as an internal ribosome entry site (IRES) to direct cap-independent translation (i.e., without the addition of an extra ribonucleotide to the 5′ end of the viral messenger RNA) and downstream by a 3′ UTR that is critical for initiation of new RNA strand synthesis.11-13 The 5′ and portions of the 3′ UTR are the most conserved parts of the HCV genome.

VIRAL REPLICATION AND LIFE CYCLE

Although peripheral blood mononuclear cells, B cells, T cells, and dendritic cells have been reported to support HCV replication, hepatocytes are the major site of viral replica-

12 Vesicle fusion/ release of infectious mature virion

1 Receptor-virus binding 11 2

Virion transport and glycoprotein maturation

Endocytosis

3 Fusion and uncoating

9

10 Assembly/ packaging

(+)

Disproportionate synthesis of (+) strand viral genomic RNA from (−) template

(−)

(+)

NS5B RdRp

8

4

P7 Translation into polyprotein precursor

Generation of (−) strand RNA template

5 Polyprotein processing NS2 C E1 E2 NS5A NS5B Cleavage of downstream NS proteins by NS3/4A cofactors

(−)

NS3 unwinds 7 double-strand RNA

(+) NS5B RdRp 6

Formation of a replication complex

NS3 NS4B NS4A Figure 79-1.  Putative life cycle of hepatitis C virus. (Reproduced with permission from Pawlotsky JM, Chevaliez S, McHutchison JG. The hepatitis C virus life cycle as a target for new antiviral therapies. Gastroenterology 2007; 132:1979-98.) (See text for details and Fig. 79-2 for functions of the hepatitis C virus proteins.) NS, nonstructural; RdRp, RNA-dependent RNA polymerase.

Chapter 79  Hepatitis C for HCV cell entry.19 Other receptors are also likely required for viral entry.18 There is some evidence that the low-density lipoprotein (LDL) receptor is involved during endocytosis of HCV.20 Not only do viral receptors on hepatocytes play a role in entry of HCV into the cell, but also other cofactors affect the efficiency of viral infectivity. C-type lectins DC-SIGN (dendritic cell–specific intercellular adhesion molecule-3grabbing nonintegrin; CD209) and L-SIGN (DC-SIGNr, liver and lymph node specific; CD209L) are expressed on dendritic cells and liver sinusoidal endothelial cells, respectively.21 Both receptors bind E2 on the HCV virus and facilitate infection of adjacent hepatocytes by acting as a capture and delivery mechanism. In addition, HDL increases infectivity of hepatocytes ten-fold, although the mechanism is not understood.22 Once the HCV virus attaches to the cell, endocytosis of the bound virion is presumed to occur, as in other flavi­ viruses. A pH drop in the vesicle causes conformational changes in the glycoproteins that lead to fusion of the viral and cellular membranes23 and release of viral RNA into the cytoplasm. In the cytosol, the 5′ UTR contains several highly conserved and structured domains that function as an IRES, which directs the RNA to its docking site on the endoplasmic reticulum and mediates cap-independent internal initiation of HCV polyprotein translation by recruiting both cellular proteins, including eukaryotic initiation factors (eIF) 2 and 3, and viral proteins.24,25 The large polyprotein generated by translation of the HCV genome is co- and post-translationally processed proteolytically into at least 11 viral proteins, including both structural (nucleocapsid [C], or p21; envelope 1 [E1], or gp31; and envelope 2 [E2], or gp70) and nonstructural (NS2, NS3, NS4A, NS4B, NS5A, and NS5B) proteins (Fig. 79-2).11,13 The functions of these specific proteins are described later in the chapter. After polyprotein processing, NS4B expression causes the membrane alterations that are seen on electron microscopy as a membranous web.26,27 The replication complex associates viral proteins, cellular components, and nascent RNA strands and is essential for HCV replication, as demonstrated in replicon cell culture systems.28 HCV replication is catalyzed by the NS5B RNA-dependent RNA polymerase (RdRp). The positive-strand genomic RNA serves as a template for the synthesis of a negative-strand intermediate. Negative-strand RNA serves as a template for production of numerous strands of RNA of positive polarity that are used for polyprotein translation and synthesis of new interme­ diates of replication and are packaged into new virus particles.29 Finally, viral particle formation is initiated by the interaction of the core protein with genomic RNA in the endoplas-

C

E1

E2

p7

NS2

NS3

mic reticulum, although the details of this process and subsequent export of mature virions from the hepatocyte are poorly understood.30,31 By analogy with pestiviruses, HCV packaging and release are likely to be inefficient because much of the virus remains in the cell. Following release, viral particles may infect adjacent hepatocytes or enter the circulation, where they are available for infection of another cell or host.

Virus Protein Function

The large polyprotein generated by translation of the HCV genome is cleaved by cellular and viral proteases to form structural and nonstructural proteins.11,13 The structural proteins are separated from the nonstructural proteins by the short membrane peptide p7, which is believed to be a viroporin, a protein that plays a role in viral particle maturation and release.32,33 At least one, and possibly three, alternative reading frame proteins (ARFP, or F for “frameshift”) exists.34-36 The exact number of alternative reading frames, the number of proteins that result, and the functions of the proteins are not known.36 One such protein is 17 kd in size; it can be expressed in vitro, and antibodies to it have been found in infected patients. The crystal structures of most of the ORF proteins have been elucidated and have been helpful in understanding protein interactions and functions. Although these proteins are most important for viral replication, some also interact with host proteins and may facilitate persistence of the virus by impairing the immune response. The core protein is first cleaved from the large polypeptide and then further processed by a host signal peptidase.11 In infectious HCV virions, core protein forms the viral nucleocapsid and binds RNA; it has many other functions as well. Core protein has been found attached to lipid rafts and the endoplasmic reticulum, and it translocates into the nucleus. When core protein attaches to lipid rafts, it recruits nonstructural proteins, resulting in the assembly of infectious virions. Core protein can also interact with the host immune system by inactivating the RNA silencing activity of Dicer, a cellular endoribonuclease that produces small interfering RNA to bind and target HCV RNA for destruction by the cell.37 Core protein can also bind to Janus kinase-1 (JAK1) and JAK2 and alter the activation of signal transducer activator of transcription (STAT) proteins, leading to impairment of interferon production.38 Extracellularly, core protein inhibits T-cell activation and proliferation, possibly by down-regulating co-stimulatory molecules on dendritic cells.39 Specific polymorphisms in core protein have also been associated with intracellular lipid accumulation40; this may be the result of facilitation of phosphorylation of insulin receptor substrate-1 (IRS-1), thereby leading to insulin resistance.41 Mutations in core protein have also

NS4A

NS4B

NS5A

NS5B

Serine NS3 ? RNA RNA Envelope Viroporin RNA-dependent protease helicase protease binding site RNA polymerase glycoproteins Cystine cofactor protease Figure 79-2.  Schematic representation of the hepatitis C virus polyprotein. The structural proteins C (core), E1, and E2 (envelope proteins) are cleaved from the polyprotein by the host signal peptidase. p7, a viroporin protein, is cleaved by the endoplasmic reticulum signal peptidase and forms an ion channel that is essential for assembly and release of infectious virions. The NS2 cysteine protease autocatalytically cleaves itself from the polyprotein. The NS3 protease cleaves the remainder of the nonstructural proteins: NS3 (serine protease and RNA helicase), NS4A (NS3 protease cofactor), NS4B, NS5A (RNA binding site), and NS5B (RNA-dependent RNA polymerase). Core

1315

1316

Section IX  Liver been associated with an increased risk of hepatocellular carcinoma (HCC) in patients; core protein alone can cause HCC in transgenic mice.42 E1 and E2 proteins are cleaved from the polypeptide by host signal peptidase.11 The two proteins form highly glycosylated heterodimers and then tetramers that are essential for viral assembly (see earlier). They also mediate cell entry by binding to surface receptors.43 Subsequently, they are responsible for fusion between the host cell membrane and the viral envelope. Because E1 and E2 are expressed on the surface of the virion, they are targets of host antibodies. The first 27 amino acids of E2 form hypervariable region 1 (HVR1); alterations in HVR1 are believed to be an attempt by the virus at antibody-mediated immune evasion. P7 is cleaved by the endoplasmic reticulum signal peptidase and forms an ion channel. This viroporin protein is essential for efficient assembly and release of infectious virions but not for cell entry.11,33 Because p7 is needed later in the viral life cycle, cleavage of the polypeptide is delayed.33 NS2 complexes with NS3 and zinc to form a cysteine protease, with two composite active sites, that autocatalytically cleaves NS2 from NS3.44 No other function of NS2 has been discovered to date. NS3 has several functions in addition to complexing with NS2 for autocatalytic cleavage of the NS2-NS3 site.44 Its function as a serine protease is markedly enhanced by its association with NS4A. The enzyme results in cleavage of the polyprotein at the NS3-NS4A, NS4A-NS4B, NS4B-NS5A, and NS5A-NS5B sites.11,45 The NS3 protease also cleaves and thereby destroys the function of Cardif and TRIF (Toll/interleukin receptor domaincontaining adapter-inducing interferon-β), which are intermediates in two separate pathways of host-cell interferon secretion in response to viral infection.46-48 This property may have a significant effect in impairing the host response to HCV infection. Finally, a portion of the NS3 protein functions as a helicase that unwinds viral RNA as well as host DNA. The helicase function is dependent on adenosine triphosphate (ATP), may require dimerization of NS3, and progresses in discrete steps like an inchworm.12,49 NS4A complexes with NS3 and functions to stabilize the protease and helicase activities and anchor the complex to the endoplasmic reticulum membrane.11,12 It also regulates hyperphosphorylation of NS5A.50 The only known function of NS4B is to induce the formation of the “membranous web” on which HCV transcription occurs.26 NS5A binds zinc and forms homodimers that are bound to the endoplasmic reticulum membrane.12 NS5A is essential for viral replication and is believed to provide an RNA-binding site within the replication complex.51 In addition, NS5A inhibits apoptosis in infected cells,52,53 and some mutations confer improved sensitivity to interferon therapy.54 NS5B is the viral RNAdependent RNA polymerase.11 The crystal structure elucidates the tunnel of the enzyme that directs single-stranded RNA into the active site.55 It can synthesize both negativestrand HCV RNA templates and positive-strand HCV RNA genomes.

GENOTYPES AND QUASISPECIES

HCV has an inherently high mutational rate that results in considerable heterogeneity throughout the genome.56 This high mutational rate is in part a consequence of the RNAdependent RNA polymerase of HCV, which lacks 3′-to-5′exonuclease proofreading ability that ordinarily would remove mismatched nucleotides incorporated during replication. An average of one error occurs for every 104 to 105 nucleotides copied. This phenomenon is favored by a high

viral turnover rate; 1010 to 1012 virions are produced per day.57 A substantial proportion of newly synthesized viral genomes have alterations. Because of the functional differences in HCV proteins, genetic variation in some parts of the genome confers advantages by evading or inhibiting the host immune system, whereas other mutations may be lethal to the virus if they lead to defective replication machinery. Therefore, genetic variation is distributed irregularly along the genome. Each new genetic variant is produced in a single cell and may or may not spread through the liver and into the serum. The result is not only genetic diversity in the serum, but also compartmentalization of variant virions in different parts of the liver and perhaps in extrahepatic sites. Because of the vast genetic variation, a classification scheme was devised whereby viral sequences are given a genotype and subtype. The first division used to describe the genetic heterogeneity of HCV is the viral genotype, which refers to genetically distinct groups of HCV isolates that have arisen during the evolution of the virus. Nucleotide sequencing has shown variation of up to 34% between genotypes.56 The most conserved region (5′ UTR) has a maximum nucleotide sequence divergence of 9% between genotypes, whereas the highly variable regions that encode the envelope proteins (E1 and E2) exhibit a nucleotide sequence divergence of 35% to 44% between genotypes. The sequences cluster into 6 major genotypes (designated by numbers), with sequence similarities of 60% to 70%, and more than 70 subtypes (designated by a lower case letter) within these major genotypes, with sequence similarities of 77% to 80%.56 In this scheme, the first variant, which was cloned by Choo and colleagues, is designated type 1a.58 The HCV genotype is an intrinsic characteristic of the infecting HCV strain and does not change over time; therefore, the genotype only needs to be determined once in an infected person. Mixed-genotype infections may be seen and reflect either coinfection with more than one HCV virus or methodologic problems in genotype testing. In addition, inter-genotypic HCV recombinants have been described59; these are thought to arise because of recombination between different genotypes in patients with repeated exposure. The recombination events have been reported to occur in or between NS2 and NS3.60 Global geographic differences exist in the distribution of HCV genotypes, as well as in the mode of acquisition. In the United States, genotype 1a is the most prevalent, accounting for approximately 57% of HCV infections, followed by genotype 1b in 17%, genotype 2 in 14%, genotype 3 in 7%, and genotype 4, 5, or 6 in less than 5%.61 Racial differences are seen in the prevalence of genotypes; approximately 90% of African Americans are infected with HCV genotype 1, whereas only 70% of whites and 71% of Hispanics are infected with genotype 1.62 In Europe, the most prevalent genotype is 1b (47%), followed by 1a (17%), 3 (16%), and 2 (13%).63 Genotype 4 is found mainly in Egypt, the Middle East, and Central Africa.64 In Egypt, approximately 15% of the population is infected with HCV, and more than 90% have HCV genotype 4. Because of the high prevalence rate in Egypt, genotype 4 represents 20% of the world’s HCV-infected population. Genotype 5, although originally isolated in South Africa, is also seen in specific regions of France, Belgium, and Spain.65 Genotype 6 is found predominantly in Asia. The distribution of genotypes is ever changing with immigration and alterations in the primary modes of viral transmission. Therefore, the frequencies of viral genotypes change over time.66 An important clinical correlation with HCV genotype is response to treatment (see later). HCV genotype does

Chapter 79  Hepatitis C not, however, appear to influence the severity of liver disease, as defined by the stage of fibrosis, or the likelihood of progression of acute to chronic HCV infection. The second component of genetic heterogeneity is quasispecies generation.56 Quasispecies are closely related, yet heterogeneous, sequences of HCV RNA within a single infected person that result from mutations that occur during viral replication. The rate of nucleotide changes varies signi­ficantly among the different regions of the viral genome. The highest proportion of mutations is found in the E1 and E2 regions, particularly in HVR1. Even though this region represents only a minor part of the E2 region, it accounts for approximately 50% of the nucleotide changes and 60% of the amino acid substitutions within the envelope region. The development of quasispecies may be one mechanism by which the virus escapes the host’s immune response and establishes persistent infection.67 During acute infection or during treatment, the lack of quasispecies is associated with viral clearance, and the development of numerous quasispecies is associated with viral persistence.68 In acute disease, patients in whom genetic variation in the HVR1 region develops after antibody seroconversion progress to chronic disease, whereas those in whom such genetic variation does not develop are more likely to achieve viral clearance.67 Genetic variation before seroconversion does not correlate with outcome, indicating that quasispecies formation results from antibody-mediated immune pressure. Interestingly, no intrinsically interferon-resistant variants of HCV have been defined, indicating that both viral and host factors play important roles in determining whether the virus persists or is cleared. An increased number of quasispecies has also been associated with more rapid progression to cirrhosis and the development of HCC.69

EPIDEMIOLOGY INCIDENCE AND PREVALENCE

The worldwide seroprevalence of HCV infection, based on detection of antibody to HCV (anti-HCV), is estimated to be 3%, with more than 170 million people infected chronically. Marked geographic variation exists, with infection rates ranging from 1.3% to 1.6% in the United States to 15% in Egypt.1,64 Currently, between 3.2 and 5 million persons are infected with HCV in the United States.1 The prevalence rate is higher in persons 40 to 49 years old than in older or younger persons, in males (2.1%) than in females (1.1%), and in African Americans (3%) than in whites (1.5%).1 Other risk factors for HCV infection and the associated frequencies of infection are injection drug use (57.5%), blood transfusion before 1992 (5.8%), greater than 50 lifetime sexual partners (12%), and family income below the poverty level (3.2%). The current prevalence of HCV infection in the United States may be underestimated because the National Health and Nutrition Examination Survey (NHANES) data did not evaluate persons who are homeless, incarcerated, or in the military. Among incarcerated persons, 12% to 35% are positive for HCV RNA serum,70 whereas those in military service have a seroprevalence rate for antiHCV of 0.5%.71 Worldwide, three different epidemiologic patterns of HCV infection have emerged: (1) previous exposure through health care with a peak prevalence in older persons; (2) exposure through intravenous drug use, the major risk factor since data first became available in about 1960, with a peak prevalence among middle-aged persons; and (3) ongoing

high levels of infection in areas where high rates of infection occur in all age groups.63 Given the factors that influence viral diversity (see earlier), estimating the site of origin and age of HCV by phylogenetic analysis is difficult. The best estimate is that HCV originated in western and sub-Saharan Africa.72 Subsequent global spread probably occurred coincident with trade and human migration. Evolution of the virus led to a geographic distribution of genotypes, so that genotypes 1, 2, and 3 are most common in North America and Europe, genotype 4 is most common in the Middle East, and genotypes 5 and 6 are most common in Southeast Asia. In Japan, HCV transmission transitioned from constant to exponential growth in the 1920s, and the prevalence of HCV infection is highest in older persons.73 In Japan, and later in southern and Eastern Europe, health care–related procedures—particularly reuse of contaminated syringes—played a major role in viral spread. In the United States, Australia, and other developed countries, peak prevalence is in persons ages 40 to 49 years, and analysis of risk factors suggests that most HCV transmission occurred between the mid-1980s and the mid-1990s, through intravenous drug use. In Egypt, the spread of HCV increased exponentially from the 1930s to the 1980s because of mass vaccination campaigns with reuse of medical equipment.64 In Egypt and other developing countries, high rates of infection are observed in all age groups, suggesting that an ongoing risk of HCV acquisition exists. In the United States, the incidence of acute hepatitis C is falling. The peak incidence was estimated to be 180,000 cases per year in the mid-1980s, but the rate declined to approximately 19,000 new cases by 2006.74 Many factors have contributed to the falling incidence of acute hepatitis C. In the 1980s, when blood was purchased from donors, 2% to 10% of blood units were infected with HCV, leading to a high rate of transfusion-acquired HCV infection.75 The institution of volunteer blood donation, creation of recombinant clotting factors, and implementation of HCV blood testing (between 1990 and 1992) dramatically decreased transfusion-acquired HCV infection.66 An important mechanism of transmission worldwide has been the lack of sterilization of medical instruments such as syringes. Although the incidence of HCV transmission by medical instruments has also decreased markedly, the risk has not been eliminated, even in the United States. Currently, new HCV infections in the United States and other developed countries occur primarily as a result of injection drug use.

TRANSMISSION

Modes of transmission of HCV can be divided into percutaneous (blood transfusion and needlestick inoculation) and nonpercutaneous (sexual contact and perinatal exposure). Patients are often unwilling to disclose percutaneous risk factors, and therefore “nonpercutaneous” transmission may represent occult percutaneous exposure.

Percutaneous Transmission

Blood transfusion, before the introduction of screening, and injection drug use are the most clearly documented risk factors for HCV infection. Following the introduction of anti-HCV screening of blood donors between 1990 and 1992, the number of transfusion-related cases of HCV infection declined sharply, and currently less than 1 case occurs per 2,000,000 units transfused.76,77 Injection drug use has always been the major route of HCV acquisition in the United States and accounts for an increasingly large portion of cases, at least 68% of new cases of HCV infection, since the virus was essentially eliminated

1317

1318

Section IX  Liver from the blood supply.74 The prevalence of HCV infection in injection drug users ranges from 57% to 90%.1,78 Although risk factors for hepatitis B virus (HBV) and human immunodeficiency virus (HIV) infection overlap with those for HCV infection, the prevalence of HCV infection in this population is the highest among the three viruses. The majority of injection drug users become anti-HCV positive within six months of initiating injection drug use with shared paraphernalia. Chronic hemodialysis is also associated with increased rates of HCV infection. The frequency of anti-HCV in patients on hemodialysis ranges from 11.6% in the United States to 55% to 85% in Jordan, Saudi Arabia, and Iran.79 Serologic assays for anti-HCV may underestimate the frequency of HCV infection in this relatively immunocompromised population, and virologic assays may be necessary for accurate diagnosis.80 Transmission may occur from infected patients to health care workers. A serologic survey of emergency department patients found that 18% were infected with HCV.81 The proportion with HCV infection was even higher in patients with a history of intravenous drug use (83%), past blood transfusion (21%), or a male homosexual lifestyle (21%). Although all potential routes of transmission of HCV infection to hospital workers are not obvious, needlestick injuries probably account for a large proportion of cases. Anti-HCV seroconversion rates are approximately 0.3% to 4% in longitudinal studies of health care workers after percutaneous inoculation from anti–HCV-positive sources, although the risk is dependent on the type of needle (hollow versus solid, infusion versus withdrawal), volume of inoculum, depth of injury, time the body fluid has spent ex vivo, level of viremia (viral load), and HIV status of the inoculating body fluid.82,83 In one study, 99% of surgeons in training experienced at least one needlestick by their final year of residency. Fifty-three percent of these injuries involved a high-risk patient, and only 49% were reported to the employee health service.84 As a result, as many as 16,000 new cases of HCV are estimated to have been transmitted to health care workers worldwide in 2000.85 Although less common, transmission of HCV also may occur from health care workers to patients. Because acute HCV infection often is subclinical, nosocomial transmission may occur with greater frequency than has been recognized previously. Strict adherence to universal precautions to protect health care workers and patients is critically important. At this time, no treatment is effective for post-exposure prophylaxis, and no data support such treatment even if it were available.

Nonpercutaneous Transmission

Nonpercutaneous modes of HCV transmission include sexual practices and childbirth. Available evidence indicates that transmission by nonpercutaneous routes occurs but is inefficient. From 10% to 20% of patients with HCV infection report that their only risk factor is sexual exposure to a partner with HCV infection. Most seroepidemiologic studies, however, have demonstrated anti-HCV in only a small proportion of sexual contacts of infected persons. In a large prospective study of monogamous seronegative partners of HCV-infected patients who denied anal intercourse and intercourse during menstruation, no instances of HCV transmission of the same sequenced virus occurred over a 10-year period of time.86 Therefore, many of the cases presumed to be the result of sexual transmission are likely the result of other, perhaps unreported or unrecognized, exposures. If the index sexual partner is infected with HIV or the partners engage in high-risk sexual practices, such as

anal intercourse, however, the transmissibility of HCV is likely increased.87 Furthermore, epidemiologic studies have shown that persons with multiple sex partners have a higher prevalence of HCV infection.1 Whether sexually transmitted diseases promote transmission of HCV through breakdown of mucosal or immune barriers is unclear. HCV-infected persons commonly are counseled to notify sexual partners of their HCV status. The risk of sexual transmission is negligible in monogamous couples that do not engage in high-risk sexual practices.86 Barrier methods should be recommended, however, to persons in nonmonogamous relationships or those engaging in high-risk sexual practices. Compared with the high efficiency of perinatal transmission of HBV infection (see Chapter 78), the risk of perinatal transmission of HCV infection is low, averaging 5.1% to 6.7% for HCV-monoinfected patients and two to three times higher for HIV-HCV-coinfected patients.88,89 Mothers with a high viral load are more likely to transmit HCV to their infants, a finding that may explain why infants born to mothers with HIV-HCV coinfection are at higher risk of HCV infection. Interestingly, the use of highly active antiretroviral therapy (HAART) in HIV-HCV-coinfected mothers may decrease the risk of perinatal transmission of both HIV and HCV.89 Data regarding the risk associated with vaginal delivery as opposed to cesarean delivery are uncontrolled, but evidence for a higher risk of HCV transmission with vaginal delivery is unconvincing. This issue remains controversial, and some authorities recommend elective cesarean section before membrane rupture.88 Although little data exist, the risk of HCV transmission from breastfeeding is negligible to small. The Centers for Disease Control and Prevention have concluded that breastfeeding by HCV-infected mothers is generally safe. Some authorities have suggested, however, that mothers with a high viral load (greater than 108 copies/mL, see later) may pose a risk.90 Because anti-HCV can be acquired passively by the infant, molecular testing for HCV RNA is required if the diagnosis of HCV infection is suspected. Infants of infected mothers should not undergo serologic testing for anti-HCV before the age of 18 months because maternal antibodies may persist in the infant’s serum and lead to diagnostic confusion.

Sporadic HCV Infection

The source of transmission is unknown in 9% to 27% of cases of HCV infection.78 Such sporadic HCV infection pro­ bably results from an undisclosed or unrecognized percutaneous route of infection. This presumption is supported by the observation that intranasal cocaine use is not considered a risk factor for HCV transmission (although it was con­ sidered a risk factor in the past).91 HCV infection can be acquired from non-commercial tattooing and body piercing when equipment is reused, shared, or improperly sterilized. Commercial tattooing is now well controlled and probably conveys little risk of HCV infection. Iatrogenic transmission of HCV is well documented in a variety of circumstances, most notably via contaminated multi-use vials and inadequately sterilized multi-use instruments and syringes, as seen with schistosomal treatment campaigns in Egypt.92

PATHOGENESIS Determinants of persistence of HCV include (1) the evasion of immune responses through several viral mechanisms, (2) inadequate induction of the innate immune response, (3)

Chapter 79  Hepatitis C insufficient induction or maintenance of an adaptive immune response, (4) the production of viral quasispecies, and (5) the induction of immunologic tolerance.67,93,94 In 55% to 85% of acute HCV infections, the net result of the host-virus interplay is the inability to clear virus despite the development of antibodies against several viral proteins. In the minority of patients in whom acute HCV resolves, an early and multispecific CD4+ T-cell proliferative response occurs, with predominance of type 1 CD4+ helper T (Th1) cells in the peripheral blood,95 most of which produce interferon-α. This “protective” response is still detected 18 to 20 years after infection in a majority of asymptomatic recovered patients but in only a minority of patients in whom chronic HCV infection develops.96 Although the immune response is essential in preventing viral persistence after acute HCV infection in 15% to 45% of cases, in those without viral clearance, the immune response mediates hepatic cell destruction and fibrosis.

VIRAL MECHANISMS

In chronically infected patients, the pathogenesis of liver damage is largely immune mediated. In a small subset of immunocompromised HCV-infected patients among both HIV-infected patients and organ transplant recipients, however, a syndrome termed fibrosing cholestatic hepatitis develops.97,98 Such cases are thought to result from direct viral hepatotoxicity of infected cells because viral levels are typically greater than 30 million copies/mL and hepatocytes contain enormous concentrations of virus and viral proteins.99 Survival in such patients is quite poor. The majority of patients with HCV infection have a variable immune response that, although inadequate to eradicate acute infection, appears to regulate the vigor of persistent infection and avoid the development of fibrosing cholestatic hepatitis. The immune response to HCV is incompletely understood because animal models are not readily available and most studies in man rely on observations in peripheral blood rather than the hepatic immune environment.

IMMUNE-MEDIATED MECHANISMS

HCV infection elicits an immune response in the host that involves both an initial innate response as well as a subsequent adaptive response. The innate response is the first line of defense against the virus and includes several arms such as natural killer (NK) cell activation and cellular antiviral mechanisms triggered by pathogen-associated mole­ cular patterns (PAMPs) recognized by the cell.100-102 These processes can lead to apoptosis of infected cells within the first few hours of infection. NK cells, as the effector cells of the innate immune system, also produce tumor necrosis factor (TNF)-b and interferon-α, cytokines that are critical for dendritic cell maturation and subsequent induction of adaptive immunity.94 After this, however, the virus initiates a number of mechanisms that undermine the ability of the host to control the infection. Virus-related disruption of the innate, and later adaptive, immune response occurs at several levels.93 NK cell function is slowed possibly because NK cell–mediated cytotoxicity and production of cytokines are interrupted when the HCV E2 protein binds its cellular receptor CD81.103,104 PAMPs activate several cellular processes including the JAK-STAT pathway and Toll-like receptor-3 (TLR-3), activation of both of which ultimately results in production of cellular interferons and interferon-regulated factors that convey antiviral properties to the cell.93 NS3/4 protease degrades TRIF, an essential intermediate in this pathway, and cleaves interferon promoter stimulator-1 (IPS-1), an

intermediate in the signaling cascade, to activate interferon when retinoic inducible gene-1 (RIG-1) binds viral intermediates.47,105,106 In addition, HCV core protein promotes STAT-1 degradation, inhibits STAT-1 phosphorylation, promotes suppressor of cytokine signaling (SOCS) induction (an inhibitor of JAK-STAT signaling), and impairs interferon-stimulated gene factor-3 (ISGF3), a heterotrimer of STAT-1, STAT-2, and interferon-b promoter stimulator (IRF-9) from binding to the promoter regions of interferonstimulated response elements (ISRE), thereby inhibiting transcription of interferon-response genes. Even when interferon-response genes are activated, NS5A and E2 both can disrupt protein kinase R (PKR) function to suppress translation, thereby allowing viral replication to continue.93 In addition, NS5A inhibits 2′-5′-oligoadenylate synthetase (OAS). 2′-5′-OAS is expressed in response to HCV infection and leads to HCV RNA degradation.93 Taken together, HCV is able to disorient the innate immune response at several levels, and these strategies appear to be pivotal in establishing the chronicity of infection. The ability of HCV to impair the innate immune response prevents development of a vigorous adaptive immune response to the infection. NK cells do not adequately activate dendritic cells, and as a result, the priming of CD8+ and CD4+ T cells in HCV-infected patients is inadequate.107-109 Dendritic cells from HCV-infected patients are more likely to produce IL-10, a cytokine that inhibits antigen-specific T-cell responses. In addition, CD4+ T cells primed from dendritic cells of HCV-infected patients are more likely to produce IL-10. Even if an adequate T-cell response is created, HCV-infected patients have a large number of regulatory T cells in their portal tracts110; intrahepatic immune regulation by these cells has not been demonstrated but is presumed. HCV-specific T cells are enriched at the site of viral replication, with an increased number in the liver when compared with the peripheral blood.95,111 CD8+ lymphocytes predominate, suggesting that cytotoxic T lymphocytes are the main perpetrators of hepatocellular injury. The T-cell immune response in the liver may result in direct lysis of infected cells and inhibition of viral replication by secreted antiviral cytokines.95,111 Whereas the cellular immune response plays a pivotal role in the pathogenesis of HCV infection, the importance of the humoral immune response is less clear. Antibodies to viral proteins are produced in low levels and do not appear to correlate with the stage of infection or immune reactivity. Furthermore, administration of high-titer HCVenriched or HCV-specific immunoglobulin has little effect on viral levels or persistence in humans.112 In summary, viral products play an integral role in the immune regulation that leads to chronic infection instead of viral clearance. Both the virus and the immune response probably play a role in the development of hepatocellular injury.

CLINICAL FEATURES ACUTE AND CHRONIC HEPATITIS C

HCV accounted for an estimated 20% of cases of acute hepatitis in 2006.74 Acute hepatitis C is rarely seen in clinical practice because nearly all cases are asymptomatic.87,113 Jaundice probably occurs in about 10% of patients with acute HCV infection, whereas 20% to 30% of patients present with nonspecific symptoms such as fatigue, nausea, and vomiting. HCV RNA is detectable within 2 to 3 weeks

1319

Section IX  Liver Symptoms +/−

Table 79-1  Extrahepatic Manifestations of Hepatitis C Virus Infection

Anti-HCV

HCV RNA Titer

1320

ALT Normal 0

1

2

3

4

Months

5

6

1

2

3

4

Years

Time after exposure Figure 79-3.  Serologic pattern of acute hepatitis C virus (HCV) infection followed by recovery. Symptoms may or may not be present during acute infection. ALT, serum alanine aminotransferase level; anti-HCV, antibody to hepatitis C virus. (Modified from the Centers for Disease Control and Prevention, www.cdc.gov/hepatitis/Resources/Professionals/ Training/Serology/training.htm#one.)

of exposure, and anti-HCV seroconversion occurs between day 15 and month 3. Serum aminotransferase levels peak at about the first month after exposure (Fig. 79-3), exceed 1000 IU/L in 20% of cases, and generally follow a fluctuating pattern for the first few months. In patients in whom jaundice develops, peak serum bilirubin levels usually are less than 12 mg/dL, and jaundice typically resolves within one month. Severe impairment of liver function and liver failure are rare. The presentation may be more apparent and the clinical course more severe when acute HCV infection occurs in patients who drink large amounts of alcohol or have coinfection with HBV or HIV. The rate of viral persistence after acute infection varies, ranging from 45% to more than 90%. Age and gender clearly influence the risk of chronicity, with younger and female patients having the lowest rates of chronicity. Other factors that may play a role include the source of infection and size of inoculum (chronicity may be less common in injection drug users than in those who acquire HCV infection by blood transfusion), immune status of the host (chronicity rates are higher in persons with immunodeficiency states such as agammaglobulinemia and HIV infection), and the patient’s race (rates of viral persistence are higher in African Americans than in whites and Hispanic Americans in the United States). Finally, the rate of spontaneous clearance is higher in symptomatic patients in whom jaundice develops during acute infection than in those who remain asymptomatic.87,113,114 In patients with community-acquired hepatitis C in whom the infection resolves spontaneously, loss of HCV RNA from serum usually occurs within three to four months of the onset of clinical disease.114 Serum alanine aminotransferase (ALT) levels are usually elevated in patients with chronic HCV infection. Because levels commonly fluctuate, however, as many as one half of patients may have a normal ALT level at any given time.1 The ALT level may remain normal for prolonged periods of time in about 20% of cases,115 although transient elevations occur even in these cases.116 Persistently normal ALT levels are more common in women, and such cases typically are associated with lower serum HCV RNA levels and less inflammation and fibrosis on liver biopsy specimens.115 Most patients with chronic hepatitis C are asymptomatic before the onset of advanced hepatic fibrosis. Patients who have been diagnosed with chronic infection, however, often complain of fatigue or depression, and they consistently

Proven Associations Autoimmune thyroiditis B-cell non-Hodgkin’s lymphoma Diabetes mellitus Lichen planus Mixed cryoglobulinemia Monoclonal gammopathies Porphyria cutanea tarda Possible Associations Chronic polyarthritis Idiopathic pulmonary fibrosis Non-cryoglobulinemic nephropathies Sicca syndrome Thyroid cancer Renal cell carcinoma Vitiligo

score lower than HCV-negative persons in all aspects of health-related quality of life (HRQL).117,118 Whether the decrease in HRQL is related to viral factors, social factors (e.g., intravenous drug use), social stigmatization, or worry related to the diagnosis itself is unclear. Nonetheless, HRQL scores improve if the patient achieves a sustained response to antiviral therapy (see later). Less common symptoms may include arthralgias, paresthesias, myalgias, sicca syndrome, nausea, anorexia, and difficulty with concentration. The severity of these symptoms may be, but is not necessarily, related to the severity of the underlying liver disease.

EXTRAHEPATIC MANIFESTATIONS

Patients with HCV infection may present with extrahepatic conditions, or these manifestations may occur in patients known to have chronic hepatitis C. Classification of the extrahepatic manifestations of HCV is shown in Table 79-1 and is based on the strength of available data to prove a correlation.119 Types 2 and 3 cryoglobulinemia, characterized by polyclonal immunoglobulin G (IgG) plus monoclonal IgM and polyclonal IgG plus polyclonal IgM, respectively, can both be caused by HCV infection. Among HCV-infected patients 19% to 50% have cryoglobulins in serum, but clinical manifestations of cryoglobulinemia are reported in only 5% to 10% of these patients and are more common in patients with cirrhosis. Symptoms and signs include fatigue, arthralgias, arthritis, purpura, Raynaud’s phenomenon, vasculitis, peripheral neuropathy, and nephropathy. The diagnosis is clear when a rheumatoid factor is detected, cryoglobulins are present, and complement levels are low in serum; however, the reliability of cryoglobulin measurements is dependent on proper handling and processing of the sample. Glomerular disease generally manifests as cryoglobu­ linemic nephropathy, membranoproliferative glomerulonephritis (MPGN), and membranous nephropathy.119,120 Cryoglobulinemic nephropathy manifests as hematuria, proteinuria, edema, and renal insufficiency of varying degrees, and on renal biopsy specimens it has features of MPGN. At diagnosis, 20% of patients with type 2 cryoglobulinemia have renal involvement, and renal involvement develops in another 35% to 60% over time. In about 15% of patients, cryoglobulinemic nephropathy progresses to end-stage kidney disease requiring dialysis. Therapy should be considered in patients with symp­ tomatic cryoglobulinemia. Cryoglobulinemia resolves in patients who achieve an SVR with pegylated interferon and ribavirin therapy (see later).119 Unfortunately, patients with

Chapter 79  Hepatitis C significant renal involvement are at a disadvantage with respect to antiviral therapy because administration of ribavirin is generally contraindicated if the creatinine clearance is less than 50 mL/min. In these patients, treatment with rituximab should be considered.121,122 A durable clinical response to four doses of rituximab has been reported, although no prospective clinical trials have been completed to date. Prednisone, cyclophosphamide, other chemotherapeutic agents as well as plasmapheresis have been used with variable success; however, these approaches do not treat the underlying HCV infection.119 If cryoglobulinemia and renal disease improve with such treatment, then subsequent treatment of the HCV infection with pegylated interferon-α and ribavirin should be reconsidered. HCV infection is associated with the development of B-cell non-Hodgkin’s lymphoma and monoclonal gammopathy of uncertain significance.119,123,124 The relative risk of lymphoma is small (1.28) in the United States.123 The most prevalent forms of lymphoma found in patients infected with HCV are follicular lymphoma, chronic lymphocytic lymphoma, lymphoplasmacytic lymphoma, and marginal zone lymphoma. In addition, marginal splenic lymphoma has been reported to regress after therapy for HCV infection alone. An estimated 8% to 10% of patients with type 2 cryoglobulinemia evolve into lymphoma over time. Despite the known association of HCV infection with lymphoma, HCV RNA does not integrate into the host genome and cannot be considered a typical oncogenic virus. Rather, HCV shows lymphotropism and may facilitate the development and selection of abnormal B-cell clones by chronic stimulation of the immune system. In addition, genetic rearrangements in B cells, specifically the Bcl2/JH rearrangement and the t(14;18) translocation, have been found in HCV-infected patients.125,126 Patients with B-cell genetic alterations are less likely to respond to antiviral therapy.127 Other extrahepatic manifestations of HCV infection include porphyria cutanea tarda, lichen planus, and sicca syndrome.119 In addition, insulin resistance and diabetes mellitus are associated with HCV infection. The SVR to antiviral therapy for HCV infection is reduced in insulinresistant patients; however, if HCV can be eradicated, insulin resistance often improves, an observation that further supports the relationship between HCV infection and insulin resistance.128 Although associations between HCV infection and both thyroid cancer and idiopathic pulmonary fibrosis have been described, data about the effect of HCV eradication on disease progression are lacking (see Table 79-1). A myriad of other conditions have been observed in association with HCV infection, but a true link has not been firmly established for these disorders.119 Although not associated with disease, seropositivity for autoantibodies is found in many HCV-infected persons (e.g., antinuclear antibodies with a titer greater than 1 : 40 in 9%, smooth muscle antibodies with a titer greater than 1 : 40 in 20%, anti-liver-kidney microsomal antibodies in 6%).129 In addition, anti-thyroid peroxidase is found in serum in 5% to 12% of HCV-infected patients, although associated thyroid disease is found in only 2% to 5% of patients.119 Therefore, the diagnosis of an autoimmune condition in a patient with HCV infection can never be based on serology alone.

DIAGNOSIS Several immunologic and molecular assays are used to detect and monitor HCV infection.130 The presence of antiHCV in high titer in serum (generally an enzyme immunoas-

say [EIA] ratio greater than 9) indicates exposure to the virus but does not differentiate among acute, chronic, and resolved infection. Anti-HCV usually persists for many years, and perhaps for life, in patients after spontaneous resolution of infection or an SVR following antiviral therapy. Serologic assays are used initially for diagnosis, whereas virologic assays are required for confirming infection, monitoring response to treatment, and evaluating immunocompromised patients.130,131

INDIRECT ASSAYS

EIAs detect antibodies against different HCV antigens. The time course of the development of symptoms, detection of anti-HCV, and appearance of HCV RNA after acute infection is shown in Figure 79-3. Three generations of EIAs have been developed. The latest, third-generation, EIAs detect antibodies against HCV core, NS3, NS4, and NS5 antigens as early as 7 to 8 weeks after infection, with sensitivity and specificity rates of 99%.131 Despite ongoing viral replication, serologic test results can be negative in patients who are on hemodialysis or are immunocompromised.80 Because the performance characteristics of third-generation EIAs are so good, confirmation with a recombinant immunoblot assay (RIBA) is no longer required. Instead, patients who are antiHCV positive should undergo HCV RNA testing to determine if they have active viremia or have cleared the infection.

DIRECT ASSAYS

Two general types of direct assays exist: qualitative and quantitative HCV RNA tests.130 Qualitative HCV RNA nucleic acid tests (NAT) only report whether HCV RNA is found in serum or not and do not quantitate the amount of HCV RNA. These tests should be used only for screening purposes now (e.g., screening of blood donated to a blood bank) and should not be used in clinical practice. As a result of NAT technology, transfusion-related HCV infection has decreased to less than one case per two million units of blood transfused.76,77 Unlike NAT testing, quantitative HCV RNA tests are essential for monitoring the response to antiviral therapy (see later). Currently “real-time” tests, such as Taqman, are performed using polymerase chain reaction (PCR) methodology, with a lower limit of detection of 10 to 15 international units (IU)/mL.132 These assays have a linear dynamic range of 1 to 7 log10 IU/mL and are the preferred testing method in practice. Transcription-mediated amplification (TMA) is also extremely sensitive, but available assays are not quantitative in the lower dynamic range of the test. The advantages of these very sensitive tests include positivity within one to three weeks after acute infection and detection of low-level residual infection during anti­ viral therapy. A disadvantage of all quantitative tests is the lack of comparability among different assays. Although conversion to a standard IU/mL concentration attempted to resolve such discrepancies, results are still variable. Thus, conversion factors vary from 0.9 copies/mL to 5.2 copies/mL per IU/mL reported. For this reason, the same laboratory and assay should be used during antiviral treatment when comparisons of levels at different points in time are critical to decision making about treatment.

HCV Genotyping

Identifying the genotype of HCV can be accomplished by several methods. The most accurate method is PCR and direct sequencing of the NS5B or E1 region; however, this approach is not practical in clinical practice. HCV genotyping can be done by evaluating type-specific antibodies

1321

1322

Section IX  Liver and has a 90% concordance in immunocompetent patients when results are compared with sequence analysis of the HCV genome.131 Testing can also be accomplished with reverse hybridization to genotype-specific probes, restriction fragment length polymorphism analysis, or PCR amplification of the 5′ noncoding region of the HCV genome. These tests have 92% to 96% concordance with the correct genotype; genotype 1 is identified with the highest accuracy. Because of mutations in the regions studied, regardless of the technique used, errors in subtype identification occur in 10% to 25% of cases. A line-probe assay (Inno-LIPA) using genotype-specific probes for reverse transcription of the 5′ portion of the HCV genome is currently the most popular commercial assay for HCV genotyping.133

SELECTION OF SEROLOGIC AND VIROLOGIC TESTS

For patients at low risk for HCV infection, a negative result on an EIA for anti-HCV is sufficient to exclude HCV infection. A positive anti-HCV result is sufficient to confirm the diagnosis if the serum ALT level is elevated. If the serum ALT level is not elevated, HCV RNA should be measured. In high-risk patients, such as those with an elevated ALT level who have a known risk factor for HCV, have experienced recent exposure, or are either immunocompromised or on dialysis, a positive anti-HCV result is sufficient to confirm HCV infection. If the anti-HCV result is negative, then HCV RNA testing should be done.

Screening in Blood Banks

The risk of acquiring HCV infection from blood products has declined dramatically since blood donors have been screened routinely. Many blood banks have now switched from third-generation EIA testing to NAT testing, and this change has decreased the risk of transfusion-acquired HCV infection to one infection per two million units of blood transfused.76,77

Diagnosis Following Known Exposure

Following an occupational or recreational exposure or in the context of mother-to-infant transmission, the diagnosis of HCV infection is now based on HCV RNA testing by the most sensitive molecular method, usually a real-time PCR assay. If transmission has occurred, HCV RNA is detectable in serum 1 to 3 weeks following exposure, whereas antiHCV may not be seen for 7 to 8 weeks following exposure. Treatment should be considered in adults with acute exposure because the risk of chronicity is high without treatment and clearance of HCV RNA with antiviral therapy is likely (see later). Therefore, early diagnosis of HCV infection is important in these settings.130 In patients with a continued risk of infection, as through injection drug use, periodic testing for HCV RNA should be done. Because spontaneous clearance of HCV RNA from serum is more likely in infants than in adults, an infant should not be tested until 18 months after birth if the mother is HCV RNA positive.

is usually done by percutaneous liver biopsy (Table 79-2), but indirect and noninvasive methods to assess liver injury and fibrosis are under study and becoming commercially available. Examination of liver biopsy specimens is used to quantify hepatic injury into discrete grades of inflammation and stages of fibrosis.134,135 Several scoring systems have been used and differ in range of possible scores and definitions (Fig. 79-4). The first system used was the Histology Activity Index (HAI) described by Knodell and colleagues. The components of this system include periportal inflammation and necrosis (graded as 0 to 10), lobular inflammation and necrosis (0 to 4), portal inflammation (0 to 4), and fibrosis (0 to 4). This scoring system combines inflammation and fibrosis into one score. Scheuer created a simplified scoring system that separates grade from stage: Portal inflammation and interface hepatitis (0 to 4), lobular activity (0 to 4), and fibrosis stage (0 to 4). The Ishak system is a modification of Knodell’s system but separates histologic grade from stage. Ishak’s fibrosis scores range from 0 to 6 (1 or 2, portal fibrotic expansion; 3 or 4, bridging fibrosis; 5 or 6, cirrhosis) (Fig. 79-5). The higher number of gradations of fibrosis has made the Ishak system popular for scoring progression of fibrosis in clinical trials. Currently, the METAVIR scoring system is the most popular; it is simpler than all the aforementioned systems. Inflammation is graded from 0 to 4 (none, mild, moderate, and severe), and fibrosis is staged from 0 to 4 (1, portal fibrotic expansion; 2, portal fibrosis with septa formation; 3, bridging fibrosis; 4, cirrhosis). Although examination of liver biopsy specimens is still considered the standard for establishing the grade of inflammation and stage of fibrosis, limitations of liver biopsy include (1) associated morbidity (pain occurs in as many as

Table 79-2  Reasons to Perform a Liver Biopsy in a Patient with Hepatitis C Assessment of the need for surveillance for hepatocellular carcinoma Evaluation for concomitant liver diseases Guidance for decisions regarding treatment of hepatitis C Staging of fibrosis

Fibrosis

Knodell

None

0

0

0

Portal

1

1

1

Periportal

2

2

2

3

3

4 5

- marked Cirrhosis

The risk of progressive hepatic injury from HCV infection varies considerably, with some patients showing little or no progression after decades of infection and others progressing rapidly to cirrhosis.134 Therefore, an assessment of the degree of liver injury is usually advisable. This assessment

3

Bridging - focal - diffuse

LIVER BIOPSY AND NONINVASIVE ASSESSMENT OF FIBROSIS

Ishak

METAVIR

4

4

6

Figure 79-4.  Comparison of the Knodell, METAVIR, and Ishak hepatic fibrosis staging systems. The METAVIR staging system is similar to the Scheuer system. Portal, periportal, bridging, and cirrhosis describe the degree (stage) of fibrosis (see also Fig. 79-5).

Chapter 79  Hepatitis C Stage 0 No fibrosis Hepatic artery

Stage 1 Portal fibrosis

Stage 2 Portal fibrosis with septa

Portal vein Bile duct

Stage 3 Bridging fibrosis Focal

Diffuse

Marked

Stage 4 Cirrhosis

Figure 79-5.  Visual depictions of fibrosis staging (METAVIR system) in patients with chronic hepatitis C. Stage 0 represents the absence of abnormal fibrosis, stage 1 shows portal fibrosis, and stage 2 shows portal fibrosis with septa. Bridging fibrosis (stage 3) is a continuous process ranging from focal periportal fibrosis with occasional bridging between portal structures to marked bridging without nodule formations, sometimes called early cirrhosis. Depending on the staging system (see Figure 79-4), bridging fibrosis may be stage 3, 4, or 5, emphasizing the importance of the reporting of the staging system used in pathology reports. A diagnosis of cirrhosis requires the presence of nodule formation.

30% in some series and hemorrhage or bile leak occurs in 0.3% of patients) and mortality (0.03%); (2) cost; (3) poor patient acceptance; (4) intraobserver and interobserver variability in the interpretation of findings (with current scoring systems, intraobserver and interobserver concordance for staging fibrosis among hepatopathologists is approximately 90% and 85%, respectively); (5) inaccuracy in interpretation of findings, particularly for the diagnosis of cirrhosis (with a false-negative rate of 15%); and (6) sampling error (a 33% difference in one stage of fibrosis and 2.4% difference in two stages of fibrosis is seen in simultaneously obtained biopsy specimens from the right and left hepatic lobes).136,137 Interobserver and intraobserver variability is increased when inexperienced pathologists use a complicated scoring system to evaluate liver tissue. Sampling error is especially common when small biopsy specimens are obtained. A biopsy should be done with at least a 16-gauge needle, be 15- to 20-mm or more in length, and contain at least 6 portal triads, although 11 or greater is considered optimal.138,139 Because of the limitations of liver biopsy, several noninvasive tests to estimate fibrosis have been developed (Table 79-3). Although many such tests have been evaluated, few have entered clinical practice. FibroSure (also known as FibroTest) is a noninvasive measure of fibrosis that creates a composite score, adjusted for gender and age, derived from the serum levels of a2-macroglobulin, haptoglobin, apolipoprotein A-1, gamma glutamyl transpeptidase, and

total bilirubin.137 The test accurately categorizes patients with stage 0 and 1 fibrosis and those with cirrhosis; however, it is less useful in patients with intermediate scores. The AST-to-platelet ratio index (APRI) is used primarily to diagnose or exclude cirrhosis.140 In an initial evaluation, 81% of cirrhotic patients were accurately categorized with a score less than or equal to 0.5; however, the index does not discriminate among lower levels of fibrosis, and its reproducibility has been called into question.138,141 A new technol­ ogy, transient elastography (Fibroscan), uses ultrasound to assess the elasticity of the liver, which correlates with the amount of hepatic fibrosis. In a meta-analysis, the area under the receiver operating curve (an estimate of accuracy) of Fibroscan for predicting cirrhosis was 0.94.142 When FibroSure and Fibroscan were evaluated, the areas under the curve (an estimate of accuracy) for predicting cirrhosis were 0.9 for FibroSure and 0.95 for Fibroscan.143 Although noninvasive testing has improved dramatically, all currently available tests have limitations. First, these tests generally have not been applied to the evaluation of other potentially contributing disease processes, such as nonalcoholic steatohepatitis. Second, the degree of hepatic inflammation is not assessed by these tests. Third, the utility of these tests has not been evaluated in select populations such as dialysis patients, HIV-positive persons, liver transplant recipients, and persons with active autoimmune conditions. Fourth, although cirrhosis is accurately predicted by several noninvasive tests, the finer discrimination of

1323

1324

Section IX  Liver every four to five years.144-146 In patients who initially had stage 0 or 1 fibrosis, a noninvasive measurement may be adequate for following progression of fibrosis. In those with stage 2 or 3 fibrosis, however, liver biopsy is typically recommended because development of cirrhosis, if demonstrated, signifies the need for surveillance for hepatocellular carcinoma and varices. In the future, noninvasive testing for fibrosis will likely play a larger role in the management of patients with HCV infection.

fibrosis score is not as reliable as that with examination of liver biopsy specimens. Despite the limitations, in patients with a relative contraindication to liver biopsy (such as hemophilia or anticoagulation therapy in patients who are at high risk of developing thrombotic events if treatment is interrupted) or in patients who refuse a liver biopsy, noninvasive testing can be helpful in disease management and in assessing prognosis. The following recommendations can be made regarding the role of liver biopsy in the management of patients with chronic hepatitis C (see Table 79-2). Regardless of the degree of serum aminotransferase elevations, an examination of a baseline liver biopsy specimen is recommended, but not mandatory, for initial assessment of patients with chronic hepatitis C, genotype 1 or 4, when examination of liver histology will influence treatment plans, and in patients infected with HCV genotype 2 or 3 who wish to defer treatment.144 Liver biopsy is not required when cirrhosis is already suggested by clinical findings (e.g., ascites, splenomegaly, spider angioma, low platelet count, or prolonged prothrombin time) or imaging (e.g., nodularity of the liver, evidence of portal hypertension). It is also not indicated following successful antiviral therapy, although histology generally improves significantly over time following eradication of HCV (see later). In persons who do not achieve an SVR through treatment, a repeat assessment of liver fibrosis should be considered

NATURAL HISTORY Progression of chronic hepatitis C is largely silent, as graphically depicted in Figure 79-6. The reported rate of progression of fibrosis depends on the cohort studied and the strategy used to collect data. Between 2% and 24% of HCVinfected patients progress to cirrhosis after 20 years of infection.147 Iatrogenic outbreaks of HCV infection have provided insight into the natural history of HCV infection. In 1977, young Irish women were infected with HCV genotype 1b when a contaminated lot of anti-D immunoglobulin was administered to them. After more than 17 years of followup, cirrhosis developed in only 2%, and complications of liver disease developed in none.148 These women continue to be followed, and 184 untreated patients with HCV infec-

Table 79-3  Accuracy of Noninvasive Tests for Predicting Hepatic Fibrosis in Patients with Hepatitis C

TEST

NUMBER OF FIBROSIS PATIENTS STAGING STUDIED SYSTEM

APRI

270

Ishak

Transient elastography (Fibroscan) FibroSure

327

METAVIR

339

METAVIR

HISTOLOGIC FIBROSIS (F) STAGES COMPARED

SENSITIVITY (%)*

PPV FOR FIBROSISTEST SPECIFICITY (%)* CIRRHOSIS (%) ACCURACY (%)†

F0-2 vs. F3-6 F0-4 vs. F5-6 F0-1 vs. F2-4 F0-3 vs. F4

41 89 56 86

95 75 91 96

88 57 88 78

70 77 68 94

F0-1 vs. F2-4 F0-2 vs. F3-4

100 70

22 95

50 91

57 84

*Sensitivity and specificity for distinguishing higher stages of fibrosis from lower stages of fibrosis. † Accuracy = (sensitivity)(prevalence) + (specificity)(1 − prevalence) APRI, AST-to-platelets ratio index; PPV, positive predictive value. From Wai CT, Greinson JT, Fontana RJ, et al. A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. Hepatology 2003; 38:518-26; Ziol M, Handra-Luca A, Kettaneh A, et al. Noninvasive assessment of liver fibrosis by measurement of stiffness in patients with chronic hepatitis C. Hepatology 2005; 41:48-54; Imbert-Bismut F, Ratziu V, Pieroni L, et al. Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: A prospective study. Lancet 2001; 357:1069-75.

Hepatic decompensation 3% per year

Acute HCV infection

Figure 79-6.  Natural history of hepatitis C virus (HCV) infection. Hepatic decompensation includes ascites, hepatic encephalopathy, variceal hemorrhage, hepatorenal syndrome, or hepatic synthetic dysfunction.

55–89%

Chronic HCV infection

2–24% over 20 years

Cirrhosis

1–4% per year Hepatocellular carcinoma

Chapter 79  Hepatitis C tion had paired biopsy data over the subsequent decade, with cirrhosis developing in only an additional 2.1%.149 A similar study from Germany involved 1018 women infected through contaminated batches of anti-D immunoglobulin.150 Twenty years after exposure, 85% of patients tested positive for anti-HCV, and 55% tested positive for HCV RNA. An additional 3% had achieved an SVR with interferon therapy. Of the 220 who underwent liver biopsy, 3% had bridging fibrosis, and none had cirrhosis. These two studies demonstrate that the prognosis is excellent in women infected at a young age with a low viral inoculum. Longer-term studies are still needed to ensure that fibrosis progression does not accelerate later in life. Most estimates of the rate of progression of fibrosis in HCV-infected patients are substantially higher than those reported in the aforementioned studies. The cohort followed appears to influence the observed rate of progression dramatically. The 20-year expected rate of progression to cirrhosis is 25% in patients referred to liver centers, 23.8% in patients followed after transfusion-associated hepatitis, 3.8% in healthy blood donors, and 4.7% in communitybased studies.147 The factor that most distinguishes these different groups is the age at the time of infection. Fibrosis appears to progress more rapidly in older than in younger patients, although other factors such as alcohol intake may influence the rate of progression as well. Among patients with HCV-induced cirrhosis, manifestations of liver failure (e.g., ascites, gastrointestinal bleeding, encephalopathy, hepatorenal syndrome, or synthetic dysfunction) develop in 3% per year, and HCC develops in 1% to 4% per year.151-154 The Hepatitis C Long-Term Treatment against Cirrhosis (HALT-C) trial, a prospective study of patients with bridging fibrosis or cirrhosis (see later), showed that over a period of three and one half years, the risk of death was 4.6% to 6.6%, the risk of decompensation was 13.2% to 14.3%, and the risk of HCC was 2.8% to 3.2%.155 The five-year survival rate is only about 50% after hepatic failure develops.151,152 An estimated 8000 to 10,000 deaths from HCV-related liver complications occur annually in the United States, although this number, which is based on death certificates, is likely an underestimate.156 The age-adjusted mortality rate from HCV infection has decreased slightly in the 2000s.5,157 Cirrhosis and HCC resulting from HCV infection are the leading indications for liver transplantation in the United States and Europe and account for approximately 40% of transplantations since 2000 in the United States. In patients who achieve an SVR with therapy and who have compensated cirrhosis from HCV, the risk of decompensation and HCC is reduced dra-

matically.155,158 Nevertheless, HCC can still occur, and surveillance for HCC is still needed in these patients.

FACTORS ASSOCIATED WITH PROGRESSION OF CHRONIC HEPATITIS C

Numerous host and environmental factors contribute to the presence or absence of histologic progression of HCV infection (Table 79-4). Identification of these factors is important because modifiable factors can be altered and high-risk patients can be treated promptly. Age at infection is a strong predictor of outcome, as discussed earlier. Natural history studies also suggest that progression of fibrosis is not linear over time and that the progression is slower at younger ages and increases over time, with the greatest acceleration after the age of 50 years.159 This phenomenon likely explains some of the discrepancies seen in natural history studies. Accelerated fibrosis seen in older persons is independent of the age at acquisition. The mechanisms behind the accelerated rate of progression of fibrosis with aging are poorly understood. Male gender is associated with more rapid progression to cirrhosis and HCC.157,159 The reason for this association is not clear, and hormonal effects on fibrogenesis have been suggested. Estrogen inhibits proliferation and activation of hepatic stellate cells in vitro.160 In addition, fibrosis appears to accelerate in postmenopausal women, and this acceleration may be attenuated with use of estrogen.157,161,162 Race also has been implicated in disease progression. African Americans have a higher prevalence of HCV infection than whites,1 but their risk of progression of the disease to cirrhosis is lower.163,164 Because African Americans have a higher prevalence of disease, decreased response to therapy, and possibly decreased access to health care and liver transplantation, they have higher rates of HCV-related inpatient mortality.165 Unlike African Americans, HCVinfected Hispanics have a higher rate of progression of fibrosis than do whites.166 Whether the higher risk is explained by a higher prevalence of alcohol abuse, fatty liver disease, or other factors is unclear. Alcohol use of greater than 20 g/day clearly increases the risk of progression of fibrosis and development of HCC in patients who are chronically infected with HCV.167-171 Whether occasional or lesser amounts of alcohol also contribute to disease progression is less clear; however, in the absence of definitive evidence, most hepatologists recommend that HCV-infected patients avoid alcohol altogether. In addition to alcohol, smoking cigarettes or cannabis leads to an increased rate of progression of fibrosis in HCV-

Table 79-4  Factors Associated with Progression of Hepatic Fibrosis in Patients with Chronic HCV Infection ESTABLISHED

POSSIBLE

NOT ASSOCIATED

Age >40 years Alcohol consumption Hepatitis B virus coinfection HIV coinfection Immunosuppressed state Insulin resistance Marijuana use Obesity Schistosomiasis Severe hepatic necroinflammation Smoking White race

Increased hepatic iron concentration Male gender Serum ALT level

Viral genotype Viral load

ALT, alanine aminotransferase; HIV, human immunodeficiency virus.

1325

1326

Section IX  Liver infected patients.172,173 Daily users of cannabis also have a higher risk of liver steatosis. In experimental models, cannabinoid receptor antagonists have been shown to decrease hepatic steatosis.174 Several metabolic abnormalities and comorbid conditions, including insulin resistance or type 2 diabetes mellitus, obesity, and liver steatosis have been associated with accelerated progression of fibrosis.157 The frequencies of insulin resistance and of type 2 diabetes mellitus are markedly increased in patients infected with HCV genotypes 1 or 4.175-178 HCV infection results in down-regulation of IRS-1 and 2 through up-regulation of SOCS-3, an effect that is independent of other factors such as obesity that can also cause hepatic steatosis.179,180 Either insulin resistance or hepatic steatosis alone increases the risk of progression of fibrosis.179,180 Weight loss is associated with reductions in hepatic steatosis and the rate of fibrosis.181 Steatosis is also common in patients infected with HCV genotype 3, but this effect is virally induced and not associated with progression of fibrosis.182,183 Patients with insulin resistance have a lower SVR to HCV therapy.184 Mild-to-moderately increased hepatic iron stores are associated with more advanced fibrosis. A consistent relationship between C282Y or H63D heterozygosity and increased progression of fibrosis has not been established, however (see Chapter 74).185,186 Reduction of hepatic iron concentrations does not reduce the risk of progression of fibrosis or improve the response to antiviral treatment.187 Polymorphisms in genes implicated in the immune response and fibrogenesis have been studied extensively and may account for some of the variability in progression of fibrosis in patients who are otherwise similar, as reviewed elsewhere.157 The candidate genes are numerous, and most studies are limited by small sample sizes and lack of reproducibility; larger scale, genome-wide investigations are in progress.188

PATIENTS WITH PERSISTENTLY NORMAL SERUM AMINOTRANSFERASE LEVELS

Up to one half of anti-HCV-positive persons have a normal serum ALT level at any given time,1 and the ALT level will remain normal for at least six months in 8% to 20% of them.145,189 Serum HCV RNA levels are lower in these persons than in persons with elevated ALT levels, and they tend to have lesser degrees of hepatic inflammation and fibrosis. In one series, 86.5% of HCV-infected persons with a normal serum ALT level had METAVIR stage 0 or 1 fibrosis, 12% had stage 2 fibrosis, and just 1.5% had cirrhosis.189 In a study of 159 patients with persistently normal serum ALT levels, the mean Ishak fibrosis score was 0.87.116 Therefore, these patients tend to have less fibrosis, with the implication that fibrosis progresses more slowly in these patients than in those with an elevated ALT level.116,157,189 Progression to cirrhosis, although far less common in those with a normal ALT level, remains possible, and antiviral therapy may still be indicated.

IMMUNOCOMPROMISED PATIENTS

Studies of HCV-infected patients with humoral or cellular immune impairment have shown rates of progression to cirrhosis that are significantly higher than those observed in immunocompetent patients.98,190,191 Given the apparent predominance of cellular immunity in the pathogenesis of chronic hepatitis C (see earlier), it is fascinating that HCVinfected patients with common variable immune deficiency (CVID), which is characterized by a humoral immune defect,

have an accelerated rate of progression of fibrosis. In patients with CVID who received pooled plasma products before the initiation of HCV screening, the frequency of HCV infection is high, and the rate of progression of liver disease is accelerated in these patients.190 During a follow-up period of approximately 10 years, 40% of 71 such patients progressed to decompensated cirrhosis.190 Cellular immune impairment is far more common than humoral immune impairment. The most common cause is iatrogenic as a result of exogenous immunosuppression following transplantation. Unfortunately, many patients with cellular immune impairment already have HCV infection before becoming immunosuppressed. Chronic HCV infection is the most common indication for liver transplantation and is present, either alone or in association with alcoholic liver disease, in 40% of patients who undergo liver transplantation in the United States. Recurrence of HCV infection is almost universal in patients who undergo transplantation and who are positive for HCV RNA in serum. Typically, levels of viremia increase at least one log following transplantation. The clinical course in patients with recurrent HCV infection after liver transplantation is variable; overall, however, the disease progresses more quickly to cirrhosis than in nontransplanted persons. Rapid recurrence of HCV infection and graft loss resulting from fibrosing cholestatic hepatitis occur in 1% to 10% of patients, and most of these persons die within one year.98 Of those without fibrosing cholestatic hepatitis, cirrhosis develops in 20% to 40% in five years.191 High serum HCV RNA levels before or early after liver transplantation, older age of the organ donor, severe and early recurrence of HCV infection after transplantation, and the use of high doses of prednisone or lymphocyte-depleting drugs to treat acute graft rejection are factors that most consistently are associated with a poor outcome.191 In patients in whom the disease progresses to cirrhosis after liver transplantation, decompensation occurs in 40% within one year, of whom 50% die in the following year.191,192 As a result, long-term survival of patients with HCV infection who undergo transplantation is inferior to long-term survival of patients who undergo transplantation for other indications.193 Fibrosing cholestatic hepatitis is generally accepted as a contraindication to retransplantation because of the rapid destruction of the new graft. For the same reason, most liver transplant centers are reluctant to consider retransplantation in patients with severe recurrent HCV infection within the first year. Because progression to cirrhosis is quicker after retransplantation, no consensus has been established as to whether retransplantation should be offered to a patient with recurrent HCV infection, even to a patient with slowly progressive disease (see also Chapter 95).194 Transmission of HCV to other solid-organ transplant recipients has been described. Before testing for HCV infection became widely available, such transmission was thought to occur infrequently in patients who underwent transplantation. In one large study of cardiac transplant recipients,195 survival of the 261 recipients of hearts from HCV-positive donors was significantly lower at 1, 5, and 10 years (83%, 53%, and 25%), respectively, than survival in recipients of hearts from HCV-negative donors (92%, 77%, and 53%, respectively). The additional risk of death was not affected by the recipients’ pretransplant HCV status or age. Excess mortality was attributable to liver disease and coronary vasculopathy. Patients on hemodialysis and renal transplant recipients have a higher prevalence of HCV infection (up to 11% to 14% in some series in the United States) than that in the general population (1.6% in the United States).1,79,196 The

Chapter 79  Hepatitis C higher prevalence is attributed to a higher frequency of risk factors, particularly prior blood transfusion, and possibly nosocomial spread within dialysis units. Survival is lower in HCV-infected patients on hemodialysis than in uninfected hemodialysis patients, with a relative risk of death of 1.57.197,198 Most mortality is related to HCC and complications of cirrhosis. Antiviral treatment is difficult in these patients, but the chance of response to treatment occurs in about 30% in genotype 1–infected patients, a surprisingly high rate considering that ribavirin is typically avoided in this group (see later).199 The impact of HCV infection on renal transplant recipients is controversial. HCV is the leading cause of liver disease in the postrenal transplant setting, and the rate of histologic progression is more rapid than in immunocompetent persons.200 Although the effect of HCV infection on short-term survival appears to be minimal, early studies found that the rate of late survival is reduced because of infectious complications rather than complications of liver disease mellitus. Moreover, HCV increases the risk of postrenal transplant glomerulopathy and diabetes mellitus. Long-term patient and graft survival rates are lower in renal transplant recipients for reasons that require further study. Despite the risks, renal transplantation still confers a survival benefit over hemodialysis in both HCV-positive and HCV-negative recipients. Heart transplant recipients with preexisting chronic HCV infection appear to do well, although the experience is limited.201 On the other hand, as noted earlier, uninfected heart transplant recipients who receive a graft from an HCVinfected donor have poorer outcomes, with a reduction in the one-year survival rate from 92% to 83% and a reduction in the five-year survival rate from 77% to 55%.195,201 Fibrosing cholestatic hepatitis also has been reported.99 Few data are available on outcomes in lung and small bowel transplant recipients who are acutely or chronically infected with HCV.196 Interferon is generally contraindicated in recipients of solid organ transplants other than liver transplants because of the risk of graft rejection and loss.196 Therefore, treatment of HCV infection should be considered before solid organ transplantation is undertaken in transplant candidates with chronic hepatitis C (see later). Liver disease is common in the first several weeks after allogeneic bone marrow transplantation (BMT), but most cases are the result of sinusoidal obstruction syndrome or graft-versus-host disease (see Chapter 34). Chronic hepatitis C is the most common cause of chronic liver disease after BMT. The cumulative frequency of cirrhosis has been estimated to be 11% after 15 years and 24% after 20 years.202 HCV is a major cause of cirrhosis and of liver-related mortality in this population. Unlike the case of solid-organ transplant recipients, however, antiviral treatment appears to be safe in BMT recipients.203 Graft-versus-host disease is a contraindication to interferon therapy, and cytopenias may limit the doses of interferon-α or ribavirin that can be used safely. HCV infection is common in HIV-infected persons because the two infections share similar routes of transmission. Approximately 25% of HIV-infected persons are coinfected with HCV, and up to 8% of HCV-infected patients are coinfected with HIV.204 Differences in the efficiency of transmission of HCV and HIV by parenteral or sexual routes explain the wide variation in HCV seropositivity among HIVinfected persons. Higher rates of coinfection are seen among injection drug users and recipients of blood transfusions than among sexual contacts of HIV-infected persons. HIV infection decreases the spontaneous rate of HCV clearance during acute HCV infection and leads to a correspondingly

higher rate of chronic HCV infection. Before the introduction of HAART, HCV had little impact on morbidity and mortality in HIV-infected persons. The reduction in the rate of mortality from the acquired immunodeficiency syndrome (AIDS) since the introduction of HAART, however, has allowed the sequelae of chronic HCV infection to become apparent. Indeed, complications of liver disease, mostly related to hepatitis C, are now almost as common as AIDSrelated complications, and a major cause of mortality, in patients with HIV infection.204,205 Progression of hepatic fibrosis is accelerated in HIV-HCV-coinfected patients when compared with HCV-monoinfected patients. In a large study,206 10.6% of HIV-HCV coinfected patients with hemophilia progressed to decompensated cirrhosis over a median of 12.1 years as compared with 1.6% of HCV-monoinfected patients with hemophilia.206 Potential mechanisms of increased progression of fibrosis have been investigated. HIV increases the production of TGF-b1 in the liver and leads to increased HCV replication.207 HIV-positive patients with a low CD4+ count have greater gastrointestinal bacterial translocation, leading to increased levels of lipopolysaccharide in the portal circulation and greater progression of fibrosis.208 HIV proteins may also cause activation of collagen production by hepatic stellate cells.209 In addition, steatosis and steatohepatitis are common in HIV-HCV-coinfected patients, and both are associated with an increased risk of advanced fibrosis on liver biopsy specimens.210 All forms of immunosuppression, whether endogenous or exogenous, appear to result in more rapid progression of fibrosis in HCV-infected persons. More research is needed to elucidate the mechanisms and determine whether different immunosuppressive agents or regimens might make transplantation safer for HCV-infected persons.

PREVENTION GENERAL MEASURES AND TREATMENT OF ACUTE INFECTION

Because neither an effective vaccine nor post-exposure prophylaxis against HCV infection is available, major efforts should be placed on preventing HCV infection. Needle exchange and methadone programs have decreased the rate of HCV transmission among injection drug users.211 In addition, universal precautions, disposable equipment, and rigorous sterilization of reusable medical and surgical equipment have reduced nosocomial HCV infections. Nonetheless, HCV infection is common among patients, and exposure to these patients is common in hospital workers. In one study, the frequency of hepatitis C was 15-fold higher in hospitalized patients than in the general population.212 Needlesticks are common in the health care setting and are under-reported. Almost all surgical house staff will sustain at least one needlestick, and in nearly one half the needlestick will be from a high-risk patient.84 Fortunately, the risk of transmission of HCV is about 0.3% when exposure occurs from hollow bore needles used to draw patients’ blood, although deep injuries increase the risk of transmission.83,213 Postexposure treatment with interferon-α has been used after occupational exposure to HCV, but the experience to date is uncontrolled and no benefit has been shown (not surprisingly given the low risk of transmission in this setting).214 Although postexposure prophylaxis is not effective, early treatment of acute HCV infection is effective.215-217 Treatment should be strongly considered in patients with acute

1327

1328

Section IX  Liver HCV infection. Therapy with pegylated interferon alone should be started between weeks 8 and 12 after presentation if HCV RNA has not cleared spontaneously from serum by then.216,217 The addition of ribavirin to pegylated interferon does not increase the SVR in patients with acute HCV infection.215 Patients with acute HCV genotype 1 infection who are treated for 24 weeks have a greater than 80% chance of an SVR. A shorter duration of therapy may be acceptable for patients with acute HCV genotype 2 or 3 infection, but data are limited. Because most patients with HCV infection are asymptomatic or minimally symptomatic, the vast majority of HCVinfected persons are not detected during acute infection. Therefore, screening of groups at high risk of HCV infection is recommended (see earlier). Persons at high risk include recipients of blood and blood products before 1992, past or present injection drug users (including those with a single exposure), persons with multiple sexual partners, patients on hemodialysis, infants of HCV-infected mothers, persons with occupational exposure to HCV-positive blood, and patients with persistently elevated serum ALT levels.1 HCVinfected patients should be instructed to avoid sharing razors. In addition, “safe sex” practices, such as the use of latex condoms, should be encouraged in persons with multiple sexual partners.87 Monogamous sexual partners who do not engage in high-risk sexual activity have a negligible risk of transmission.86 In view of the low rate of vertical transmission, pregnancy and breastfeeding are not contraindicated in HCV-infected women.88-90 HCV-infected patients are advised to undergo vaccination against hepatitis A and B because of the high risk of severe liver disease if superinfection with either of these viruses occurs.

IMMUNOPROPHYLAXIS

In contrast to hepatitis B, for which immunoprophylaxis or vaccination that results in a high level of envelope antibodies leads to protective immunity, the presence of anti-HCV in serum as a result of either administration of HCV-specific immunoglobulin or experimental vaccination does not prevent HCV infection effectively. HCV has an inherently high mutational rate that results in considerable heterogeneity in the envelope proteins and facilitates rapid escape from antibody recognition (see earlier). Therefore, traditional viral envelope or multiprotein vaccines have had limited success in preventing reinfection with homologous strains of HCV and no effect in preventing reinfection with heterologous strains of HCV. Monoclonal and polyclonal immunoglobulins have been used intra- and postoperatively in liver transplant recipients in the hope of avoiding reinfection of the graft, but this approach has been repeatedly unsuccessful.218,219 Given these difficulties, work on developing an HCV vaccine has focused instead on stimulating both humoral and cellular immune responses against the virus.220 If a vaccine were able to produce both neutralizing antibodies and a cellular immune response, it might have a role as both a prophylactic and therapeutic vaccine. Several strategies have been used to create a vaccine. Combining HCV proteins with an adjuvant such as the Toll-like receptor agonist dipalmitoyl-S-glyceryl cystine lipid moiety elicit CD8+ T-cell responses.221 Similarly, DNA vaccines lead to antibody production and cross-strain protective immunity by cytotoxic T lymphocytes in vivo. Dendritic cell vaccines have been created by transducing protein or transfecting viral constructs into dendritic cells. The source of the dendritic cells used and the mechanisms of maturation are crucial to their function. All of these approaches require further investigation to determine whether protective immu-

nity is induced. Combinations of techniques are likely to be studied.

TREATMENT Interferon-α monotherapy was approved for the treatment of chronic hepatitis C, then known as non-A, non-B hepatitis, before HCV was even identified. Substantial advances have been made since then with the introduction of prolonged treatment periods, longer-acting pegylated formulations of interferon, and the oral guanosine analog ribavirin. Treatment with the current standard of care—pegylated interferon and ribavirin—achieves an SVR (i.e., eradication of HCV) in about 42% to 52% of genotype 1–infected patients and 70% to 80% of genotype 2– or 3–infected patients.6,7,222 More recently, elucidation of the mechanisms of HCV replication has led to the development of the firstgeneration protease and polymerase inhibitors, some of which are in the final phases of clinical development and appear to improve response rates even further.

GOALS

The primary goal of therapy for HCV infection is eradication of the virus. A consequence of achieving this goal is prevention of liver-related deaths associated with the development of HCC and decompensated cirrhosis. SVR—the absence of detectable virus in blood 24 weeks after completion of therapy—is an excellent surrogate marker for the resolution of HCV infection. Seven retrospective follow-up studies that collectively included more than 500 patients, many of whom were immunosuppressed, followed for a mean of three and one half years found the risk of recurrent or new infection following SVR to be just 0.1% per year.223-229 SVR is also associated with a reduction in hepatic inflammation, regression of fibrosis, and improvement in HRQL.117,118 The risk of liver failure is essentially eliminated in patients with cirrhosis who achieve an SVR.158,230,231 Although the risk of HCC after an SVR in patients with cirrhosis is reduced by more than one half, the risk is not eliminated, and therefore, screening for HCC must continue.158

END POINTS

Although SVR is considered the endpoint of therapy, other indicators of response are often monitored during therapy because they may help guide and refine treatment (Figs. 79-7 and 79-8). A rapid virologic response (RVR), defined as undetectable HCV RNA in serum after the first 4 weeks of antiviral therapy, identifies those patients who are most sensitive to treatment and is associated with an 80% to 90% SVR rate in genotype 1–infected patients who complete 48 weeks of therapy or in genotype 2– or 3–infected patients who complete 24 weeks of therapy.232 An early virologic response (EVR), defined as a greater than 2-log drop in viral load at 12 weeks of therapy, is needed to continue therapy beyond 12 weeks because the absence of an EVR predicts failure of treatment in more than 98% of cases.233 An EVR can be subdivided further into a complete EVR (cEVR), defined as undetectable HCV RNA in serum at 12 weeks of therapy, and a partial EVR (pEVR), defined as a greater than 2-log decrease in the level of HCV RNA in serum despite residual detectable HCV RNA after 12 weeks of therapy. An end-of-treatment response (ETR) is defined as undetectable HCV RNA in serum at the end of treatment; a small and variable proportion of patients with an ETR will relapse when treatment is stopped. The absence of an ETR is considered nonresponse to treatment.

Chapter 79  Hepatitis C

Serum HCV RNA level (log IU/mL)

8 7 6 Null response 5 2-log drop

4 3 Partial viral response

2

Partial early viral response Complete early viral response Rapid viral response (RVR)

1 0 0

6

12

18

24

30

36

42

48

54

}

60

EVR

66

Limit of detection 72

Serum HCV RNA level (log IU/mL)

Weeks

78

Figure 79-7.  Viral responses to therapy of hepatitis C virus (HCV) infection with peginterferon and ribavirin (beginning at week 0). Rapid virologic response (RVR) is defined as undetectable HCV RNA in serum at 4 weeks. Early viral response (EVR) is defined as a 2-log decline in HCV RNA levels from baseline; it can be further defined as a complete EVR (undetectable HCV RNA at week 12) or partial EVR (HCV RNA still detectable at week 12 but undetectable by week 24) or a partial viral response (a 2-log decline in HCV RNA levels by week 12 but with residual virus detectable at week 24). A null response (or nonresponder) signifies failure to achieve an EVR.

8 7 6

Null response 2-log drop

5 4

Relapse

3 Limit of detection

2 1

SVR

0 0

6 12 18 24 30 36 42 48 54 60 66 72 78 Weeks

DRUGS

Interferon-based regimens are, and for some time will continue to be, the cornerstone of antiviral therapy for HCV infection. Interferons are naturally occurring glycoproteins that exert a wide array of antiviral, antiproliferative, and immunomodulatory effects. Pegylated interferons consist of interferon bound to a molecule of polyethylene glycol (PEG) of varying length. The large size of the molecule increases the half-life of the interferon, thereby allowing once-weekly dosage. Two pegylated interferons are licensed for use in the United States and elsewhere. The first is 40-kd peginterferon alfa-2a, which is used in a fixed dose of 180 µg per week. The second is 12-kd peginterferon alfa-2b, which is prescribed according to the patient’s body weight in a dose of 1.5 µg/kg per week. Pegylated interferons have replaced standard interferon, used in the past, and have resulted in a significant increase in the SVR. Ribavirin is an oral guanosine analog with activity against DNA and RNA viruses. When ribavirin is used in combination with interferon, the ETR improves and the relapse rate decreases. Several mechanisms to explain the synergistic effect of ribavirin when administered in combination with interferon have been proposed, including (1) alterations of the cytokine milieu leading to a change from a type 2 T-helper cell (Th2) to a Th1 immune response; (2) depletion of intracellular guanosine triphosphate through inhibition of the host enzyme inosine monophosphate dehydrogenase (IMPDH); (3) inhibition of the action of the HCV RNA-

Figure 79-8.  Viral responses to therapy of hepatitis C virus (HCV) infection with peginterferon and ribavirin (beginning at week 0). Sustained virologic response is undetectable HCV RNA in serum six months after discontinuation of therapy (at 48 weeks) for HCV infection. Relapse is a return of HCV RNA in serum after discontinuation of therapy. A null response (or nonresponder) signifies failure to achieve an EVR.

dependent RNA polymerase; and (4) induction of lethal mutagenesis during HCV RNA replication.234 Ribavirin generally is well tolerated, although it results in a dosedependent hemolytic anemia. The dose administered is based on the patient’s weight, and the patient’s hemoglobin level must be monitored during treatment. Furthermore, in patients with a history of cardiopulmonary disease who cannot tolerate a sudden fall in the hemoglobin level, ribavirin must be used with caution, if at all. In addition, ribavirin is teratogenic; patients taking ribavirin and their partners are required to avoid pregnancy during therapy and for six months after cessation of the drug. Ribavirin has a long cumulative half-life in serum and is excreted by the kidneys; as a result it can lead to severe side effects, particularly hemolysis, in patients with kidney disease. The dose of ribavirin must be adjusted for renal function, and the drug should be administered with extreme caution to patients with a creatinine clearance less than 50 mL/min. Ribavirin is not removed by hemodialysis.

EFFICACY

The current standard of care for the treatment of HCV infection is the combination of a pegylated interferon administered subcutaneously once per week and ribavirin taken orally every day. A treatment algorithm is depicted in Figure 79-9. The dose of interferon is the same for all HCV genotypes, but the dose of ribavirin is based on the patient’s weight for genotypes 1 and 4 (13.3 mg/kg/day divided twice

1329

1330

Section IX  Liver Determine genotype and viral load

Genotype 2 or 3 Viral load < 400,000 IU/mL

Genotype 2 or 3 Viral load > 400,000 IU/mL

Genotype 1 Liver biopsy

Consider

PEG-IFN and weightbased RBV 12–16 weeks

PEG-IFN and 800 mg/d RBV 24 weeks

Fibrosis Stage 3–4

Fibrosis Stage 0–2

No treatment required at this time PEG-IFN and weight-based RBV Check HCV RNA at 4 and 12 weeks

RVR

cEVR

pEVR

< 2-log drop

Treat 48 weeks 80–90% SVR

Treat 48 weeks 70% SVR

Treat 72 weeks 38% SVR

Nonresponder Discontinue drugs

Figure 79-9.  Algorithm for the treatment of hepatitis C virus (HCV) infection. Patients infected with HCV genotype 2 or 3 and with a low viral load (level of viremia <400,000 IU/mL) may be treated with pegylated interferon (PEG-IFN) and weight-based ribavirin (RBV) for 12 to 16 weeks, with sustained virologic response (SVR) rates similar to those for patients treated for 24 weeks with PEG-IFN and 800 mg/day of RBV; either regimen is acceptable for this group. Patients at higher risk of relapse or nonresponse (those with genotype 1 or with cirrhosis or a high viral load) should not be considered for a shortened course of therapy. Patients with HCV genotype 1 infection and advanced fibrosis (stage 3-4 according to the METAVIR system) should be treated, but patients infected with genotype 1 who have stage 0-2 fibrosis may not require immediate therapy. (Dotted arrow indicates that treatment is an option in these patients.) The viral load should be checked at weeks 4 and 12 of treatment because the likelihood of response is determined by the viral load at these time points (see Figure 79-7). cEVR, complete early virologic response; pEVR, partial early virologic response; RVR, rapid virologic response; IU, international units. (Based on Mangia A, Minerva N, Bacca D, et al. Individualized treatment duration for hepatitis C genotype 1 patients: A randomized controlled trial. Hepatology 2008; 47:43-50; Pearlman B, Ehleben G, Saifee S, et al. Treatment extension to 72 weeks of peginterferon and ribavirin in hepatitis C genotype 1–infected slow responders. Hepatology 2007; 46:1671-4; Shiffman ML, Suter F, Bacon BR, et al. Peginterferon alfa-2a and ribavirin for 16 or 24 weeks in HCV genotype 2 or 3. N Engl J Med 2007; 357:124-34.)

daily) and is fixed at 800 mg per day (in two divided doses) for genotypes 2 and 3. Treatment is administered for 48 weeks in patients with genotype 1 or 4 infection and 24 weeks in those with genotype 2 or 3 infection. In genotype 1–infected patients, treatment should be stopped if an EVR is not achieved. The SVR in genotype 1–infected patients is 42% to 52%.6,7,222 In the Individualized Dosing Efficacy versus Fixed Dosing to Assess Optimal Pegylated Interferon Therapy (IDEAL) trial, 3070 genotype 1–infected patients were randomized to one of the two pegylated interferons, and no difference in SVR was noted between the two formulations.235 Higher doses of pegylated interferon or ribavirin may be helpful in some patients. In one study,236 patients with HCV genotype 1 infection, high viral load, and a weight greater than 85 kg—all factors that predict a poor response to treatment—had a higher SVR when the dose of peginterferon alfa-2a or the dose of ribavirin was increased. In this difficult-to-treat patient population, peginterferon alfa-2a in a dose of 270 µg/week improved the SVR rate when compared with a standard dose of 180 µg/week. When higher doses of ribavirin are used (15.2 mg/kg/day) in genotype 1–infected patients, the relapse rate decreases (from 38% for standard-dose to 8% for high-dose ribavirin);

however, hemolysis can limit treatment with high doses of ribavirin.236,237 Some patients may not require the standard duration of treatment to achieve an SVR. Patients with HCV genotypes 2 or 3 infection may achieve an SVR in some cases with just 12 to 16 weeks of treatment.238-240 If a truncated course of treatment is anticipated, the patient should be given a ribavirin dose based on weight rather than a fixed dose of 800 mg/day and should only be treated with a shortened course if an RVR is achieved.241 In general, patients with HCV genotype 1 infection should not have the treatment duration shortened, even if they achieve an RVR because SVR rates are about 10% less than those that follow a full year of therapy. Occasionally, however, drug intolerance may make a shorter course of treatment preferable.242,243 A longer than standard course of therapy appears to be helpful in genotype 1–infected patients who have a slow response to antiviral therapy. In patients with a pEVR, extending the duration of therapy from 48 to 72 weeks improves the SVR rate from 30% to 45.5% (with use of weight-based ribavirin).244 Furthermore, if genotype 1–infected patients are unable to tolerate the full dose of ribavirin, extending treatment to 72 weeks increases the

Chapter 79  Hepatitis C SVR rate from 32% to 45%, a rate similar to that for 48 weeks of treatment with a weight-based dose of ribavirin and pegylated interferon.245 Mangia and colleagues examined the effect of tailoring the duration of treatment–either longer or shorter than usual— to the time when HCV RNA first became undetectable in serum. They randomized patients to 48 weeks of therapy or a variable length of therapy of 24, 48, or 72 weeks based on a first negative HCV RNA result at weeks 4, 8, and 12, respectively.246 If HCV RNA was negative in serum at week 4, the SVR rate was 87% with 48 weeks and 77% with 24 weeks of therapy. When HCV RNA became negative after eight weeks, all patients received 48 weeks of treatment, and the SVR rate was 70% to 72%. When the HCV RNA first became undetectable at week 12, the SVR rate was 38% with 48 weeks and 64% with 72 weeks of therapy.

FACTORS THAT PREDICT A SUSTAINED VIROLOGIC RESPONSE

Sustained viral response rate

The best predictor of response to pegylated interferon and ribavirin is the rate of the initial fall in serum HCV RNA levels during treatment. The highest SVRs occur in patients with an RVR, followed respectively by those with a cEVR, those with a pEVR, and those without an EVR. Patients are usually interested in an estimate of their chances of a response before making a decision to proceed with therapy. Pretreatment factors associated with a greater chance of an SVR include infection with non-genotype 1 HCV, a low baseline serum HCV RNA level, absence of bridging fibrosis or cirrhosis on a liver biopsy specimen, age younger than 40 years, absence of obesity, lack of hepatic steatosis or insulin resistance, absence of HIV infection, and white race (Fig. 79-10). Although the likelihood of an SVR is marginally lower in patients without these favorable factors, patients should not be discouraged and treatment should not be withheld because of the presence of any or all of these factors. Of known pretreatment variables, the most powerful predictor of a response to treatment is the viral genotype. Genotype 2 is the most responsive to therapy. Response rates for genotype 3 infections are close to those of genotype 2, although genotype 3–infected patients with a high viral load have been shown to have lower response rates (86% for those with a serum HCV RNA level less than 600,000 IU/mL and 70% for those with a HCV RNA level of 600,000 IU/mL or greater).247 Genotype 4 infection is associated with an SVR similar to that for genotype 3 infection but requires drug doses and a duration of therapy

100%

60% 40%

41– 42

INDICATIONS AND CONTRAINDICATIONS

Antiviral therapy should be considered in all patients with chronic hepatitis C. In most cases, the decision as to whether or not to proceed with treatment is based on the patient’s desire and need for therapy. The degree of need is a subjective risk-benefit or cost-benefit decision that is made after considering the histologic stage of disease, presence or absence of favorable factors for a response to treatment, and comorbid conditions that might preclude any potential benefit of eradication of the virus. As treatments have become more effective, these decisions have become easier, and more patients have been considered to be appropriate candidates. Physicians who are not comfortable with admin-

80– 90

66– 75

80%

similar to those for genotype 1 infection. Of the common HCV genotypes, HCV genotype 1 is the least responsive to therapy. Pretreatment viral load also predicts the SVR. Studies have used different cutoff values to define a “low” viral load, with a range of <400,000 IU/mL to <800,000 IU/mL. A low viral load has consistently been associated with a higher SVR, independent of HCV genotype.241-243,246 Obesity, insulin resistance, and hepatic steatosis all decrease the chance of SVR.184 These factors appear to affect the SVR independently. Weight reduction leads to an improved SVR to HCV therapy. African Americans respond poorly to standard therapy for HCV infection, with SVR rates ranging from 19% to 28%.248-250 African Americans have a higher proportion of genotype 1 infections than do whites, but this difference does not explain the difference in SVR rates. Some evidence suggests that African Americans have improved response to higher doses of interferon.251 Advanced hepatic fibrosis is a negative predictor of SVR to therapy.252 Patients with advanced fibrosis, however, are most in need of antiviral therapy and should be treated. Patients who have advanced fibrosis should expect a 10% reduction in SVR compared with patients who do not have advanced fibrosis.158 Thrombocytopenia and neutropenia occur more frequently in these patients and may necessitate a reduction in the dose of pegylated interferon or discontinuation of therapy, further limiting the success of therapy in cirrhotic patients (see later). Finally, strict adherence by the patient to the prescribed regimen improves the likelihood of an SVR, and compliance must be emphasized at each visit. Adherence to therapy appears to be particularly relevant in the first months of therapy; the impact of a reduction in dose is less if it occurs after an RVR has been achieved.253,254

56– 68 30– 39

54– 41– 55 44

70 61

39– 52 19– 28

15– 30

40 20

20% 0% 1 non-1 High Low 3–4 0–2 Genotype Viral load

Fibrosis stage*

AA White RVR cEVR pEVR <2 Race

Response

Genotype 1

2–4 >4 HOMA

Figure 79-10.  Sustained virologic response to pegylated interferon and weight-based ribavirin in relation to patient and virus characteristics. AA, African American; cEVR, complete early virologic response; high viral load ≥400,000 IU/mL; HOMA, homeostatic model assessment, a measure of insulin resistance (higher value indicates greater insulin resistance); IU, international units; low viral load, <400,000 IU/mL; pEVR, partial early virologic response; RVR, rapid virologic response. *METAVIR system.

1331

1332

Section IX  Liver istering and monitoring the therapy should refer patients to a gastroenterologist or hepatologist experienced in treating patients with hepatitis C. The best time to treat a patient with chronic hepatitis C is before the development of cirrhosis because response rates are better and the risk of later complications of cirrhosis can be eliminated if treatment is successful. Treatment of patients with decompensated cirrhosis can still be successful but should only be undertaken by an experienced clinician. Managing patients with decompensated cirrhosis who are treated for HCV infection is labor intensive and requires reductions in dose because of cytopenias in 50% to 75% of cases as well as administration of growth factors in some cases; however, the treatment is generally poorly tolerated. Many prescribers prefer to begin with low doses of both drugs and increase the dose as tolerated by the patient.224 The chance of an SVR is low in patients with decompensated cirrhosis (13% in those with genotype 1 and 50% in those with genotype 2 and 3 infections).224 Nonetheless, the risk-benefit ratio clearly favors treatment in most of these cases. If these patients subsequently require liver transplantation, those who have achieved an SVR will avoid recurrence of HCV infection.224 Relative and absolute contraindications to interferon and ribavirin therapy are listed in Table 79-5. In general, more patients with relative contraindications to treatment have been treated successfully as practitioners have gained experience and familiarity with the treatment. Many relative contraindications to therapy may resolve over time or as a result of a specific intervention. A few absolute contraindications remain. Because ribavirin is a teratogen, unwillingness of the patient and his or her partner to practice adequate contraception and avoid pregnancy during treatment or for six months after the discontinuation of therapy is an absolute contraindication to starting or continuing treatment. Any severe or uncontrolled psychiatric condition is considered an absolute contraindication to therapy. If a patient’s psychiatric disorder is treated appropriately and is stable, however, the patient may become a candidate for HCV therapy. Severe cardiac or pulmonary disease is an absolute contraindication to HCV therapy because of the risk of worsening tissue hypoxia if severe hemolytic anemia

Table 79-5  Contraindications to Therapy with Pegylated Interferon and Ribavirin Absolute

Relative

Acute pancreatitis Autoimmune hepatitis Comorbid conditions that markedly limit life expectancy History of hypersensitivity to one of the drugs Pregnancy or unwillingness to use birth control during and for six months after treatment Severe cardiac disease Severe pulmonary disease Uncontrolled psychiatric condition Uncontrolled seizure disorder Active alcohol or drug abuse Active infection Baseline hemoglobin level <10 g/dL (for ribavirin) Baseline neutrophil count <1500/mm3 Baseline platelet count <90,000/mm3 Creatinine clearance <50 mL/min (use ribavirin with extreme caution) Decompensated cirrhosis Hemoglobinopathy Ophthalmologic disorders (may worsen during therapy) Other autoimmune conditions Uncontrolled hyperthyroidism or hypothyroidism

occurs because of ribavirin. Moreover, in most patients with severe comorbid disease, successful antiviral therapy would offer no survival benefit. Patients with an underlying autoimmune condition are at risk of experiencing an exacerbation of the condition on pegylated interferon. Although previously considered an absolute contraindication, an autoimmune disorder should probably be considered a strong relative contraindication, and the risk of an exacer­ bation of the underlying autoimmune disease must be weighed against the potential benefit of clearing the HCV infection.

MONITORING AND SAFETY

Before antiviral therapy is started, baseline liver bioche­ mical test levels, a complete blood count (CBC), and a thyroid-stimulating hormone (TSH) level should be obtained. A pregnancy test is required in women before ribavirin is initiated. The HCV genotype and serum HCV RNA level are necessary to determine the dose of ribavirin and duration of therapy (see earlier). Most hematologic side effects, particularly hemolytic anemia, occur within the first month of therapy, and initially a CBC should be performed weekly. Approximately 10% of patients will have a fall in hemoglobin levels to less than 10 g/dL, with a mean decrease of approximately 3 g/ dL. After the first month, a serum ALT level and a CBC should be obtained monthly, and a TSH level should be obtained every three months. Drug doses should be reduced according to the severity of side effects. A serum HCV RNA level should be obtained at baseline in all patients. Assessment of the decline in HCV RNA level during treatment predicts the likelihood of an SVR, can be used to determine the duration of therapy (see earlier), and is essential for determining treatment failure in genotype 1–infected patients. Therefore, quantitative HCV RNA levels should be drawn at baseline and at weeks 4 and 12. If HCV RNA is still detectable in serum at week 12, a level should also be drawn at week 24. If the HCV RNA level has not fallen by at least 2 logs (99%) after 12 weeks of treatment or is still detectable at week 24, treatment should be stopped. Use of the serum HCV RNA level at week 12 to indicate an EVR does not apply to patients with genotype 2 or 3 infection. The HCV RNA level should be measured at the end of treatment in all patients. If HCV RNA is undetectable, a level should be repeated six months after completion of therapy to determine whether an SVR or relapse has occurred. If an SVR is achieved, HCV RNA testing should be performed annually for at least two years. In nonresponders or relapsers in whom no additional treatment is considered, follow-up testing should be similar to that in patients who receive no treatment, with at least yearly check-ups and laboratory testing and a repeat liver biopsy every four to five years to assess disease progression, particularly if retreatment is considered. Drugs are discontinued consequent to an adverse event or laboratory test abnormality in approximately 5% and 16% of patients who receive a combination of pegylated interferon and ribavirin for 24 and 48 weeks, respectively.6,7,222,255 A reduction in the dose of pegylated interferon is required in 26% to 36% of patients and of ribavirin in 19% to 38%, depending on the initial dose and anticipated duration of therapy. Neutropenia, anemia, and thrombocytopenia are the most frequent reasons for reductions in dose. To ensure maximal response rates, dose reductions should be avoided, and the patient’s adherence to therapy should be encouraged. The most frequent side effects of interferon include flulike symptoms (in greater than 90% of patients) and alope-

Chapter 79  Hepatitis C cia (in 10% to 30%). Flu-like side effects are common but usually do not necessitate dose adjustments. Other adverse events include depression and hypothyroidism or hyperthyroidism. Because adherence to therapy is an important factor in optimizing efficacy, the various drug-related toxicities—in particular, psychiatric symptoms and thyroid dysfunction—should be managed medically as soon as they are detected. With ribavirin, anemia, cough, pharyngitis, insomnia, dyspnea, pruritus, rash, nausea, and anorexia are the most common side effects. The most serious side effects are anemia and teratogenic effects. Hemolytic anemia is reversible and usually resolves within the first month after therapy is stopped. If anemia is severe or slow to recover, the patient’s iron stores should be assessed by laboratory testing. Administration of hematopoietic growth factors (e.g., erythropoietin, filgrastim) may enable a patient to continue full-dose pegylated interferon and ribavirin. Use of growth factors may improve the patient’s subjective well-being but has not been shown to have an effect on response to antiviral treatment.256

RETREATMENT

The decision to retreat a patient who failed to respond to an earler course of antiviral therapy should take into consideration the previous regimen used, the appropriateness of the doses given throughout the course, the patient’s ability to tolerate therapy, and the response to treatment, especially whether and when HCV RNA became undetectable. Generally, retreatment with the same regimen, even with a different brand of interferon, is not justified. In nonresponders to an initial course of therapy consisting of either standard interferon monotherapy or standard interferon and ribavirin, the response rate to retreatment with pegylated interferon and ribavirin is 21% to 28% and 12% to 29%, respectively (Fig. 79-11).257-259 Persons who relapsed after a course of standard interferon alone or a combination of interferon with ribavirin achieve an SVR rate with

Sustained virologic response rate

60%

HIV-HCV COINFECTED PATIENTS

55

50% 43 40% 33 30%

28

20%

18

10%

6

0% Prior therapy

IFN

IFN/ RBV

PEG/ RBV

Nonresponders

IFN

pegylated interferon and ribavirin of 40% to 58% and 42%, respectively.257-259 Response rates are higher in patients who are infected with genotypes 2 and 3, have lower baseline levels of serum HCV RNA, and have lesser degrees of hepatic fibrosis. Other strategies for retreating nonresponders include use of higher doses of drugs or maintenance therapy. In the prospective Daily-Dose Consensus Interferon and Ribavirin: Efficacy of Combined Therapy (DIRECT) trial, 343 patients, who were previous nonresponders to pegylated interferon and ribavirin, were randomized to interferon alfacon-1, a non-pegylated recombinant interferon, 9 mg or 15 mg daily, and ribavirin for 48 weeks.260 The SVR rates were 5% and 9%, respectively.260 In the Retreatment with Pegasys Plus Ribavirin in Patients Not Responding to Prior Peg-Intron/ Ribavirin Combination Therapy (REPEAT) trial, 942 nonresponders to pegylated interferon and ribavirin were randomized to pegylated interferon and ribavirin for 48 or 72 weeks. A subset of each group also received a higher (induction) dose of interferon for the first 12 weeks. The SVR rate was 7% to 9% in the patients in the 48-week arms and 14% to 16% for those in the 72-week arms.261 The use of an induction dose made no difference. Maintenance therapy for nonresponders to pegylated interferon and ribavirin with low-dose (e.g., half of the usual dose) pegylated interferon has been studied in two large randomized prospective trials.155,262 In the Colchicine versus Peg-Intron Long-Term (COPILOT) trial, a decrease in the frequency of variceal hemorrhage was observed in treated patients in the subset with portal hypertension, but no differences in the frequency of liver failure, liver transplantation, HCC, or liver-related death after four years were observed.262 In the HALT-C trial (see earlier), no differences in these end points were observed after three and one half years.155 Although neither study was adequately powered to detect a benefit in the subset of patients in whom HCV RNA was suppressed throughout therapy, the conclusion can be drawn that maintenance therapy with low-dose peginterferon confers no benefit in nonresponders to full-dose therapy.

IFN/ RBV

PEG/ RBV

Relapsers

Figure 79-11.  Sustained virologic response rate to retreatment with pegylated interferon and weight-based ribavirin in HCV genotype 1–infected patients who had been treated previously and either did not respond (nonresponders) or relapsed (relapsers). IFN, standard interferon; PEG, pegylated interferon; RBV, ribavirin.

Patients with HIV-HCV coinfection probably progress more rapidly to cirrhosis than do HCV-monoinfected patients, and therefore treatment of HCV infection should always be considered in this group.263 Patients are at even higher risk of progression of fibrosis if they are female, older than age 33 years, have an increase in the CD4+ count of less than 100/mm3 with HAART, continue to have a detectable HIV viral load during antiretroviral therapy, or have untreated HCV infection. SVR rates were lower in HIV-HCV coinfected patients in clinical trials than in HCV-monoinfected historical controls.264-266 SVR rates with pegylated interferon and ribavirin (800 mg/day) for 48 weeks are 44% to 73% in genotype 2– or 3–infected patients and 14% to 29% in genotype 1– and 4–infected patients.264-266 One reason for a lower SVR in coinfected patients is the use of lower doses of ribavirin in most of the studies. In one study,267 however, standard weight-based dosing of ribavirin was used, and the SVR rate in genotype 1–infected patients was 35%. No doubt other factors played a role in the lower response rates, including high dose-reduction and drug-discontinuation rates and the presence among study subjects of factors that predict a poor response to treatment, including a high viral load, preponderance of genotype 1, a high rate of advanced hepatic fibrosis, concurrent alcohol use, and a high proportion of African Americans. Despite a lower SVR among HIV-HCV

1333

1334

Section IX  Liver coinfected patients, however, an RVR is still associated with a high SVR, and failure to achieve an EVR signifies nonresponse.268 The safety and efficacy of pegylated interferon and ribavirin have not been established in patients with a CD4+ count less than 200/mm3.269 Ideally, therapy for HCV infection should be started before antiretroviral therapy for HIV is initiated. When HCV therapy cannot be started first, zidovudine, stavudine, and didanosine should not be used in combination with ribavirin because of the additive risk of mitochondrial toxicity. Contraindications to antiviral therapy for HCV infection in HIV-HCV-coinfected patients do not differ from those for monoinfected patients.

LIVER TRANSPLANT RECIPIENTS

Complications of chronic hepatitis C are the most common indication for liver transplantation (see Chapter 95). Patients who have detectable HCV RNA in serum at the time of liver transplantation almost universally experience reinfection of the allograft. Therefore, therapy with interferon and ribavirin should be considered in cirrhotic patients, preferably before liver failure or HCC occurs. Treatment is difficult to tolerate in this setting, and response rates are poor when cirrhosis is decompensated (see earlier). Another option in these patients is to initiate antiviral treatment after transplantation. Preemptive therapy of all HCV-infected patients within the first few weeks after transplantation is associated with substantial toxicity and a low chance of an SVR; it is not recommended.270 Still another option is to follow the patient expectantly post-transplantation and begin antiviral therapy when surveillance liver biopsy specimens demonstrate progression of disease, usually to grade 3 inflammation or stage 2 fibrosis. Often, this degree of liver injury occurs at least one year post-transplantation, but initiating treatment before then is not contraindicated. Treatment is better tolerated more than one year post-transplantation, but reductions in drug doses are required in most patients. Few patients tolerate full doses of ribavirin because of the renal dysfunction associated with calcineurin inhibitors. Nonetheless, 23% to 26% of genotype 1–infected patients achieve an SVR with 48 weeks of pegylated interferon and ribavirin. The SVR rate may be higher than 80% in patients infected with HCV genotype 2 or 3.270-272 Although patients tolerate therapy poorly and SVR rates are low, a survival advantage has been demonstrated for patients who achieve an SVR compared with those who remain viremic.273 Although triple therapy with pegylated interferon, ribavirin, and a protease inhibitor will likely become available to nontransplant patients with HCV infection in the near future (see later), protease inhibitors such as telaprevir alter the metabolism of immunosuppressants. Therefore, transplant recipients are unlikely to benefit from these newer agents, at least initially.

FUTURE THERAPIES

A greater understanding of the HCV replication machinery has led to the identification of numerous potential drug targets. Several HCV-specific protease and polymerase inhibitors are in clinical trials, and two of the former class of agents are in phase III trials. These drugs promise to improve the chances of eradication of HCV in coming years. Nonetheless, the same issues that have plagued these classes of drugs in the treatment of HIV infection, including intolerance and viral resistance, will be important considerations in the treatment of HCV infection as well. Therefore, initial trials have used these novel agents in combination with pegylated interferons and ribavirin to reduce the chances of drug resistance and viral breakthrough during treatment.

Several inhibitors of the NS3/4A serine protease are currently in clinical trials. Telaprevir (VX-950) and boceprevir (SCH503034) are the two that are furthest along in development. Initial monotherapy studies demonstrated early selection of preexisting drug-resistant strains of the HCV virus, particularly if the dose of the drug was insufficient or the dosing interval was greater than eight hours.274 The combination of peginterferon alfa-2a and ribavirin with telaprevir, given orally three times a day for 12 weeks, followed by 12 weeks of peginterferon alfa-2a and ribavirin, in previously untreated genotype 1–infected patients led to a rapid reduction in serum HCV RNA levels that prevented the emergence of resistance and resulted in an SVR rate of 61%, compared with 41% in those treated with pegylated interferon and ribavirin alone for 48 weeks (standard treatment).275 A second trial in previously untreated genotype 1-infected patients showed an SVR rate of 68% with the 24-week triple-drug regimen and an SVR rate of 48% with standard treatment.276 Ribavirin remains a critical component of the regimen because telaprevir and pegylated interferon dual therapy produce an inferior SVR compared with triple-drug therapy. Side effects of telaprevir include rash, gastrointestinal complaints, and anemia, although most are mild to moderate and do not require a dose reduction or discontinuation. Boceprevir, a second protease inhibitor, is also given orally three times a day, and, in combination with peginterferon alfa-2b and ribavirin, led to an SVR rate of 55% in previously untreated genotype 1–infected patients after 28 weeks of therapy.277 Because of concerns about resistance, a second group of patients was treated with a 4-week lead-in phase of pegylated interferon and ribavirin to reduce HCV RNA levels before the introduction of boceprevir. The SVR rate was 57% after completion of an additional 24 weeks of triple-drug therapy. A second target of HCV replication, the RNA-dependent RNA polymerase, was inhibited by R1626, a nucleoside analog, given orally twice daily, in combination with peginterferon alfa-2a and ribavirin.278 Triple-drug therapy in previously untreated genotype 1–infected patients resulted in undetectable HCV RNA levels in serum at four weeks in 74% of patients, compared with 5% of patients treated with standard therapy. Another orally bioavailable nucleoside analog inhibitor of the RNA-dependent RNA polymerase, R7128, reduced serum HCV RNA levels in patients by 5.1 log when used in combination with pegylated interferon and ribavirin, compared with a 2.5-log reduction with standard treatment, after 28 days.279 Other nonnucleoside inhibitors target allosteric sites on the RNAdependent RNA polymerase and lead to modest reductions in serum HCV RNA levels when given as monotherapy; studies of combination therapy with pegylated interferon and ribavirin have not yet been undertaken. Another strategy for impairing viral polymerase function is to inhibit cyclophylin B, which facilitates attachment of HCV RNA to the replicase complex. The cyclophilin inhibitor DEBIO-025, given daily in combination with pegylated interferon, caused a 4.75-log drop in serum HCV RNA levels in 29 days, compared with a 2.49-log decline in HCV RNA levels with pegylated interferon alone and a 2.2-log decline in HCV RNA levels with DEBIO-025 alone.280 Other approaches to antiviral therapy include targeting the HCV helicase and viral assembly; these approaches to antiviral therapy are in the early stages of development. Ultimately, combination therapy with several virusspecific agents, similar to the strategy used to treat HIV infection, will likely become the favored approach to treating HCV infection and hopefully will achieve eradication of

Chapter 79  Hepatitis C the virus in most cases. Drug interactions, viral resistance, and drug side effects will all be major hurdles to overcome.

KEY REFERENCES

Armstrong GL, Wasley A, Simard EP, et al. The prevalence of hepatitis C virus infection in the United States, 1999 through 2002. Ann Intern Med 2006; 144:705-14. (Ref 1.) Bruno S, Stroffolini T, Colombo M, et al. Sustained virological response to interferon-alpha is associated with improved outcome in HCVrelated cirrhosis: A retrospective study. Hepatology 2007; 45:579-87. (Ref 158.) Everson GT, Trotter J, Forman L, et al. Treatment of advanced hepatitis C with a low accelerating dosage regimen of antiviral therapy. Hepatology 2005; 42:255-62. (Ref 224.) Fried MW, Shiffman ML, Reddy KR, et al. Peginterferon alfa-2a plus ribavirin for chronic hepatitis C virus infection. N Engl J Med 2002; 347:975-82. (Ref 6.) Mangia A, Minerva N, Bacca D, et al. Individualized treatment duration for hepatitis C genotype 1 patients: A randomized controlled trial. Hepatology 2008; 47:43-50. (Ref 246.) Manning DS, Afdhal NH. Diagnosis and quantitation of fibrosis. Gastroenterology 2008; 134:1670-81. (Ref 138.) Manns MP, McHutchison JG, Gordon SC, et al. Peginterferon alfa-2b plus ribavirin compared with interferon alfa-2b plus ribavirin for

initial treatment of chronic hepatitis C: A randomised trial. Lancet 2001; 358:958-65. (Ref 7.) Missiha SB, Ostrowski M, Heathcote EJ. Disease progression in chronic hepatitis C: Modifiable and nonmodifiable factors. Gastroenterology 2008; 134:1699-714. (Ref 157.) Poordad F, Reddy KR, Martin P. Rapid virologic response: A new milestone in the management of chronic hepatitis C. Clin Infect Dis 2008; 46:78-84. (Ref 232.) Powell EE, Jonsson JR, Clouston AD. Steatosis: Co-factor in other liver diseases. Hepatology 2005; 42:5-13. (Ref 178.) Sangiovanni A, Prati GM, Fasani P, et al. The natural history of compensated cirrhosis due to hepatitis C virus: A 17-year cohort study of 214 patients. Hepatology 2006; 43:1303-10. (Ref 154.) Shiffman ML, Salvatore J, Hubbard S, et al. Treatment of chronic hepatitis C virus genotype 1 with peginterferon, ribavirin, and epoetin alpha. Hepatology 2007; 46:371-9. (Ref 237.) Shiffman ML, Suter F, Bacon BR, et al. Peginterferon alfa-2a and ribavirin for 16 or 24 weeks in HCV genotype 2 or 3. N Engl J Med 2007; 357:124-34. (Ref 241.) Strader DB, Wright T, Thomas DL, Seeff LB. Diagnosis, management, and treatment of hepatitis C. Hepatology 2004; 39:1147-71. (Ref 144.) Wiesner RH, Sorrell M, Villamil F. Report of the first International Liver Transplantation Society expert panel consensus conference on liver transplantation and hepatitis C. Liver Transpl 2003; 9:S1-9. (Ref 191.) Full references for this chapter can be found on www.expertconsult.com.

1335

CHAPTE R

80 Hepatitis E Rakesh Aggarwal and Krzysztof Krawczynski

CHAPTER OUTLINE Virology  1337 Epidemiology  1337 Routes of Transmission  1338 Seroepidemiology  1339

Hepatitis E is a form of acute, icteric, self-limited viral hepatitis caused by the hepatitis E virus (HEV). The disease was first recognized in the 1980s, when sera collected during the first recorded epidemic in Delhi, India, in 1955,1 and during another epidemic in Kashmir, India, in 1978,2 were found to lack serologic markers of hepatitis A and B.3 In retrospect, several epidemics of enterically transmitted hepatitis with epidemiologic features resembling those of hepatitis E outbreaks occurred in Europe in the 18th and 19th centuries. HEV was identified in 1983 by immune electron microscopy4; its genome was cloned in 1990 and fully sequenced shortly thereafter.5,6

VIROLOGY HEV is a small RNA virus, 32 to 34 nm in diameter, nonenveloped, and icosahedral. HEV RNA is approximately 7.2 kilobases in length, single- and positive-stranded, 5′-capped, and polyadenylated. The crystal structure of HEV-like particles that consist of capsid protein has been found to exhibit the protruding domain that is involved in the binding of HEV to susceptible cells and that contains some neutralization epitopes.7 Other investigators have found that each capsid protein contains 3 linear domains that form distinct structural elements: the continuous capsid (S), 3-fold protrusions (P1), and 2-fold spikes (P2) and have speculated that binding of the susceptible cell receptor to P1 may result in conformational changes that eventually lead to cell membrane penetration and genome release into the infected cell.8 The virus is currently classified in a separate genus named Hepevirus in the family Hepeviridae.9 The HEV genome contains three open reading frames (ORFs) (Fig. 80-1).6 ORF1 encodes nonstructural proteins, ORF2 encodes the viral capsid protein, and ORF3 encodes a protein of unknown function. Details of HEV replication in hepatocytes and its release from infected cells remain unknown; the application of the replicon system to HEV research may generate a better understanding of the intracellular events during virus replication.10 Phylogenetic analysis of HEV isolates shows the existence of four geographically distinct genotypes, termed genotypes 1 to 4 (Table 80-1).11 Genotype 1 includes various isolates from Asia that have a nucleotide sequence homology of 92% to 99% (amino acid sequence homology 95% to 99%) with each other. These isolates have nucleotide homology of 75% (amino acid homology 86%)

Pathogenesis  1340 Clinical Features  1340 Diagnosis  1341 Treatment and Prevention  1341

with genotype 2 isolates, which include a strain from Mexico and some isolates from western Africa. Isolates from the United States, classified as genotype 3, are 92% identical to each other but only 73.5% to 74.5% identical to genotypes 1 and 2.12 Genotype 3 isolates have also been reported from human cases in several European countries (including the United Kingdom, France, the Netherlands, and Spain), Japan, and South America. Genotype 4 consists of isolates from some parts of China, Taiwan, Japan, and Vietnam. All HEV genotypes belong to a single serotype because they share at least one major serologically cross-reactive epitope. Swine HEV was identified in pigs in the midwestern United States13; the virus naturally infects pigs and induces transient viremia and antibodies that react with human HEV strains. A comparison of swine HEV with the Burmese and Mexican isolates of human HEV showed approximately 90% to 92% and 79% to 83% identity at the amino acid level in the ORF2 and ORF3 regions, respectively. HEV-like genomic sequences have been isolated from pigs in several parts of the world and occasionally from other animals, including deer and wild boar. In geographical regions where genotype 3 and 4 HEV has been reported among humans, swine isolates of HEV have also belonged to these genotypes. Also, swine HEV strains in Taiwan and Spain have greater genetic similarity to human HEV strains from these regions than to swine and human HEV isolates from other parts of the world, suggesting the possibility of zoonotic transmission. By contrast, swine HEV identified in India is genetically different from the HEV in patients from the same geographic region.14,15 An HEV-like virus has been recovered from chickens and its genome partially sequenced; it appears to be noninfectious to mammals.16

EPIDEMIOLOGY Several epidemics of hepatitis E affecting several hundred to several thousand persons have occurred on the Indian subcontinent and in southeast and central Asia, where this infection is endemic.1,2,17-19 Outbreaks of hepatitis E also have been reported from the northern and western parts of Africa and the Middle East. Two small outbreaks were reported in Mexico in 1986 and 1987 (Fig. 80-2). Overall attack rates range from 1% to 15% and are higher for adults (3% to 30%) than for children (0.2% to 10%). The male-tofemale ratio among cases has ranged from 1:1 to 4:1, but the

1337

1338

Section IX  Liver Polycysteine region

Transmembrane sequence ORF3

Hypervariable region

Hydrophobic region

ORF1 ORF2 Signal peptide

Figure 80-1.  The genome of hepatitis E virus. The three open reading frames—ORF1, ORF2, ORF3—are shown.

Methyl transferase

Domain X

Domain Y

Helicase

Papain-like cysteine protease

RNA replicase

Proline “hinge”

Glycosylation site

Table 80-1  Hepatitis E Virus Genotypes and Their Geographic Distribution GENOTYPE

HUMAN CASES

1

1A: India, Myanmar (formerly Burma), Nepal 1B: China, Pakistan, countries of the former Soviet Union 1C: Northern Africa 1D: An isolate from a patient with fulminant hepatic failure and a few other Indian isolates Mexico, Western Africa USA, South America, Europe (France, Spain, UK, the Netherlands), Japan China, Taiwan, Japan, Vietnam

2 3

4

ANIMALS —

— USA, China, Japan, Southeast Asia, Australia, New Zealand, South America India, China

Adapted from Schlauder GG, Frider B, Sookoian S, et al. Identification of 2 novel isolates of hepatitis E virus in Argentina. J Infect Dis 2000; 182:294-7.

outbreaks have been characterized by particularly high attack and mortality rates among pregnant women. The epidemics vary in nature, ranging from single-peaked, shortlived outbreaks to prolonged, multipeaked epidemics lasting more than a year (Table 80-2). In endemic areas, hepatitis E accounts for up to 50% to 70% of cases of sporadic acute hepatitis; these cases are demographically and clinically similar to those observed during disease outbreaks. In nonendemic regions, hepatitis E is related mostly to travel to HEV-endemic regions and accounts for fewer than 1% of cases of acute viral hepatitis. Solitary cases or small series of cases with indigenously-acquired cases of acute hepatitis E have been reported from several countries in Europe, North America, and Australia. A report from the United Kingdom described 40 patients with hepatitis E and no history of travel to an area endemic for this disease. HEV

infection was thought to be acquired locally, and the patients were described as a group of “autochthonous” hepatitis E cases.20 Although most patients found to have HEV infection in this series had jaundice, a few had anicteric illness with nonspecific symptoms or asymptomatic aminotransferase elevations. The number of cases peaked in spring and summer, and disease appeared to be more common in residents of coastal and estuarine areas. The disease showed seasonal variations with peaks in spring and summer and no cases occurring during November and December. Case series with similar characteristics have been described from southwest France and the Netherlands.21,22 Table 80-3 highlights differences in the epidemiologic and clinical features of HEV genotypes 1 and 3.

ROUTES OF TRANSMISSION

HEV infection is transmitted predominantly through the fecal-oral route. Most reported outbreaks have been related to consumption of fecally contaminated drinking water (see Table 80-2). The outbreaks frequently follow heavy rains and floods, but some epidemics occur during the hot summer months, when decreased flow in rivers may increase the risk of water contamination. Person-to-person transmission of hepatitis E seems to be uncommon during both epidemic and sporadic settings,23,24 and secondary attack rates among household contacts are only 0.7% to 2.2%. Therefore, recurrent epidemics in endemic regions are probably related to nearly continuous fecal contamination of water. This conclusion is supported by the demonstration of HEV RNA in waste water, sewage, and drinking water in endemic regions25 and of viable HEV (infectious to primates) in occasional sewage specimens in nonendemic regions.26 Mechanisms postulated to contribute to contamination of water sources by HEV and maintenance of the virus in endemic areas include the occurrence of subclinical HEV infection, animal reservoirs that harbor HEV, and prolonged fecal shedding of the virus by humans. Although data on subclinical HEV infection in humans in endemic regions are scant, viral excretion has been demonstrated in an experimental macaque model of subclinical HEV infection.27 Proven incidents of animal-to-human transmission of HEV have been shown to occur in Japan, where hepatitis E developed after ingestion of uncooked deer meat; the viral genomic sequences obtained from these cases were similar to those retrieved from left-over meat.28 Genomic sequences from human cases of hepatitis E in the United Kingdom

Chapter 80  Hepatitis E have resembled those from swine isolates of HEV, both belonging to genotype 3. HEV genomic sequences were isolated from pig livers sold in grocery stores in Japan and in the United States and were closely related to those from human cases of hepatitis E in those countries.29,30 Zoonotic transmission of HEV infection as a critical factor in HEV transmission, however, is not compatible with the rarity of clinically overt human HEV infection in non-

Table 80-2  Epidemiologic Features of Hepatitis E Large outbreaks involving several thousand persons in developing countries Sporadic cases Frequent in endemic areas Uncommon in nonendemic areas (occur mainly among travelers to endemic areas) Fecal-oral transmission, usually through contaminated water Highest attack rates among young adults aged 15 to 40 years, with relative sparing of children Infrequent person-to-person transmission No evidence of parenteral or sexual transmission High attack rates in pregnant women, particularly in second and third trimesters High mortality rates (15% to 25%) in pregnant women, especially in the third trimester Mother-to-newborn (transplacental) transmission is likely

endemic areas despite the high prevalence of antibody to HEV (anti-HEV) among animals (domestic swine, wild rats and mice, cattle, dogs). Moreover, genetic differences between human and animal HEV isolates in India14,15 and failure of epidemic strains of HEV to induce experimental infection in pigs cast doubts on the epidemiologic signi­ ficance of the zoonotic route of HEV transmission in endemic regions. Vertical transmission from pregnant mothers to newborn babies is well documented.31 One study showed evidence of transmission of HEV by blood transfusion,32 but further data are needed to determine the frequency of transmission by this mode.

SEROEPIDEMIOLOGY

Prevalence rates of anti-HEV, which is detectable in all geographic areas, are higher among populations in endemic areas (10% to 40%) than among those in nonendemic areas (1% to 5%). Studies from nonendemic areas, however, have reported seroprevalence rates among healthy people of approximately 20% in the United States33 and 16% in the United Kingdom.20 Whether anti-HEV seroreactivity in nonendemic areas reflects subclinical HEV infection, serologic cross-reactivity with other agents, exposure to animal reservoirs of HEV-like viruses, or false-positive serologic results is unclear. In endemic areas such as India, where antibody to hepatitis A virus is nearly universally detectable by adolescence,

Table 80-3  Comparison of Epidemiologic and Clinical Features Associated with HEV Genotypes 1 and 3 HEV GENOTYPE 1

HEV GENOTYPE 3

Epidemiologic patterns Animal-to-human transmission Water-borne transmission Animal reservoir Age group Chronic infection

Large epidemics, small outbreaks, and sporadic cases Not reported Well known to occur; most common route No Young men most commonly affected Not known to occur

Severity

Variable severity, including fulminant hepatic failure

Only sporadic cases Demonstrated; a likely mode of transmission Unknown Yes Usually elderly Reported in transplant recipients receiving immunosuppressive drugs Severity and poor outcome are related to comorbid conditions

Figure 80-2.  Geographic distribution of areas endemic for hepatitis E virus (red areas).

1339

1340

Section IX  Liver anti-HEV is uncommon in children, and the seroprevalence rate increases during young adulthood to reach about 40% in adults.34 These rates of HEV seropositivity appear to be relatively lower than those expected from the frequency of clinical disease and outbreaks. By contrast, in Egypt, antiHEV is common in children and the seroprevalence rate of anti-HEV reaches 70% or more in young adults,35 although epidemics of disease are not known. These distinctive seroprevalence patterns may be related to infection with an attenuated or animal strain of HEV or differences in environmental conditions.

PATHOGENESIS Understanding of virologic, serologic, and pathologic events in the pathogenesis of HEV infection (Fig. 80-3) is based on data from human patients and experimentally infected primates (cynomolgus macaques, rhesus monkeys, and chimpanzees). The virus enters the host primarily through the oral route, although the mechanism(s) by which the virus reaches the liver is unknown. In human volunteers, the incubation period after oral exposure is four to five weeks, and HEV is detected in the stool approximately one week before the onset of illness and for up to two weeks there­ after. In clinical cases, the fecal shedding of the virus is observed until about four weeks after the onset of illness. HEV RNA can be detected in serum for two weeks after the onset of illness in virtually all patients. In experimentally infected primates, HEV RNA has been found in serum, bile,

Acute

Convalescent phase HEV RNA in stool

and feces a few days before elevation of the serum alanine aminotransferase (ALT) level. HEV antigen (HEVAg) is expressed in hepatocytes as early as seven days after infection and may be identified in more than 50% of the cells, but the number of HEVAg-positive cells decreases sharply when serum ALT levels increase markedly.36 The onset of ALT elevation in the serum and of histopathologic changes in the liver generally corresponds with the appearance of anti-HEV in serum. Patients with acute hepatitis E show cellular immune responses to HEV proteins37-39 but the number of cytokineproducing HEV-reactive CD8+ cells is not increased.40 Helper T cell type 1/type 2 (Th1/Th2) cytokine balance is biased toward Th2 in pregnant women,41 in whom hepatitis E is particularly severe. Although the temporal concordance of the HEV-specific immune response with the onset of liver pathology suggests that liver injury is immune mediated, further studies are needed to define the role of the immune response in the induction of liver injury; no evidence for cytopathic properties of HEV itself exists. Histopathologic features of hepatitis E are similar to those of other forms of acute hepatitis and include ballooned hepatocytes, acidophilic bodies, focal parenchymal necrosis, and inflammatory infiltrates in the lobules and enlarged portal tracts. Nearly one half of patients with hepatitis E have cholestatic hepatitis, characterized by canalicular bile stasis and gland-like transformation of parenchymal cells, with less marked hepatocytic degeneration and necrosis.42 Submassive or massive necrosis and collapse of liver parenchyma are seen in patients with severe liver injury. Chronic HEV viremia with genotype 3 virus has been reported in some renal and liver transplant recipients in Europe, with elevated aminotransferase levels and no evidence of infection with other hepatotropic viruses.43 Viremia and liver enzyme elevation appeared nearly at the same time, and liver biopsies showed evidence of liver injury. A case report has suggested that such infection can lead to chronic hepatitis and cirrhosis.44

HEVAg in the liver HEV RNA in serum

CLINICAL FEATURES

ALT

The most common recognizable form of HEV infection is acute icteric hepatitis. The clinical manifestations (Table 80-4) are similar to those observed in patients with acute hepatitis A or B.45 The illness usually is insidious in onset. An initial prodromal phase lasts one to four days and is characterized by a varying combination of flu-like symptoms, fever, mild chills, abdominal pain, anorexia, nausea,

IgG anti-HEV

IgM anti-HEV

0

1

2

3

Table 80-4  Clinical Features of Hepatitis E

4

5

Months post-infection Figure 80-3.  The time course of events during hepatitis E virus (HEV) infection (based on studies in human subjects and in experimentally infected primates). ALT, serum alanine aminotransferase level; anti-HEV, antibody to HEV; HEVAg, HEV antigen; IgG, immunoglobulin G; IgM, immunoglobulin M.

Incubation period of 2 to 10 weeks Varying clinical manifestations Anicteric hepatitis Icteric hepatitis Severe hepatitis leading to fulminant hepatic failure Inapparent, asymptomatic infection Clinical illness is similar to that with other types of acute viral hepatitis (except among pregnant women) Milder illness in children Low mortality rate (0.07%-0.6%) (except in pregnant women, see Table 80-2) No relation to chronic hepatitis (except in immunosuppressed persons), cirrhosis, or hepatocellular carcinoma

Chapter 80  Hepatitis E aversion to smoking, vomiting, clay-colored stools, dark urine, diarrhea, arthralgias, asthenia, and a transient macular skin rash. Most prodromal symptoms tend to diminish with the onset of jaundice (icteric phase), but dark urine, light stool color, and itching persist for a varying amount of time. Physical examination reveals jaundice and a mildly enlarged, soft, and slightly tender liver. Some patients have splenomegaly. Laboratory test abnormalities include bilirubinuria, conjugated hyperbilirubinemia, and marked elevations in serum aminotransferase and gamma glutamyl transpeptidase levels. Elevation of the serum ALT level may precede symptoms, and the magnitude of elevation does not correlate with the severity of liver injury. Mild leukopenia and relative lymphocytosis may occur. As the illness subsides, serum aminotransferase levels return to normal, followed by a decline in the serum bilirubin, which usually returns to normal levels by six weeks. Ultrasonography may demonstrate normal findings or show a mildly enlarged liver, increase in hepatic parenchymal echogenicity, gallbladder wall edema, prominence of portal venules, and a slightly enlarged spleen. The main use of ultrasonography is to exclude biliary obstruction as the cause of jaundice. Acute hepatitis E usually is self-limited. A few patients have a prolonged course with marked cholestasis (cholestatic hepatitis), including persistent jaundice lasting 2 to 6 months, prominent itching, and marked elevation of the serum alkaline phosphatase level, ultimately with spontaneous resolution. Reported case-fatality rates have ranged from 0.5% to 4%; however, these rates seem to be overestimated because they are based on data from hospitals. Population surveys during outbreaks have reported lower mortality rates of 0.07% to 0.6%.17 Some HEV-infected persons may exhibit only nonspecific symptoms resembling those of an acute viral febrile illness with serum aminotransferase elevations but without jaundice (anicteric hepatitis), and some may remain entirely asymptomatic. Anicteric and asymptomatic infections may occur more frequently than icteric disease, because a large proportion of HEV-seropositive persons in endemic areas do not recall ever having had jaundice. The lower attack rates among children than adults during HEV outbreaks may represent a higher frequency of anicteric or subclinical HEV infections in the younger age group. In a small proportion of patients, the disease is severe and associated with subacute or fulminant hepatic failure. Pregnant women, particularly those in the second or third trimester, are affected more frequently during hepatitis E outbreaks than are others in the population and have a worse outcome, with mortality rates of 5% to 25%. In an epidemic in Kashmir, India, clinical hepatitis E developed in 17.3% of pregnant women, compared with 2.1% of nonpregnant women and 2.8% of men, all ages 15 to 45 years.46 Among the pregnant women, attack rates during the first, second, and third trimesters were 8.8%, 19.4%, and 18.6%, respectively. Fulminant hepatic failure developed in approximately 22% of the affected pregnant women, with an increased frequency of abortions, stillbirths, and neonatal deaths. The pathogenic elements that lead to severe liver damage during pregnancy, especially during the last trimester, remain unknown.

DIAGNOSIS The diagnosis of human HEV infection is based on detection of either HEV RNA in stool and serum specimens with use of a reverse transcription–polymerase chain reaction assay

or demonstration of a virus-specific host immune response (see Fig. 80-3). Tests to detect HEV RNA are available only in research laboratories. Enzyme immunoassays (EIAs) for the detection of immunoglobulin M (IgM) and IgG antibodies to HEV have been developed using recombinant HEV antigens expressed in Escherichia coli or insect cells, synthetic peptides corresponding to immunogenic epitopes of HEV, and protein expressed from a synthetic gene encoding multiple linear antigenic epitopes from the ORF2 and ORF3 regions.47-49 The presence in serum of IgM anti-HEV indicates acute infection, whereas detection of IgG anti-HEV may indicate the convalescent phase or past infection. IgM anti-HEV appears in the early phase of clinical illness, lasts 4 to 5 months, and can be detected in 80% to 100% of cases during outbreaks of acute hepatitis E. IgG anti-HEV appears in serum a few days after IgM anti-HEV, and the titer of IgG anti-HEV increases during the convalescent phase and remains high for at least one to four and one half years; the exact duration of persistence of IgG anti-HEV in serum is not known. Although several commercial kits for the detection of IgM and IgG anti-HEV are available in various countries, none is currently licensed for clinical use in the United States. Use of different target antigens from different HEV strains and of different expression systems to produce recombinant proteins makes comparison of different tests problematic; the sensitivity of various assays has ranged from 17% to 100% when such assays are applied to a panel of coded specimens from nonendemic areas.50

TREATMENT AND PREVENTION Acute hepatitis E usually is self-limited and requires only supportive care and no specific intervention. Isolation of infected persons is not indicated, because person-to-person transmission is uncommon. Patients with fulminant hepatitis need measures to control cerebral edema and consideration of liver transplantation (see Chapter 93). In pregnant women, no proven benefit to terminating the pregnancy has been documented; postpartum hemorrhage resulting from deranged coagulation requires treatment with fresh-frozen plasma. Prevention of hepatitis E in endemic areas depends primarily on the supply of clean drinking water and strict attention to sewage disposal. In an epidemic setting, measures to improve the quality of water—as simple as boiling water—have been shown to lead to rapid abatement in the number of new cases. The occurrence of large epidemics of hepatitis E among adults in endemic areas suggests that anti-HEV is not fully protective or that antibody levels in serum decline with time, gradually reaching a nonprotective level. In the few studies that have evaluated the role of immune globulin manufactured in endemic areas for pre- or postexposure prophylaxis, no significant reduction in disease rates with use of this agent was found. Cloning of the HEV genome and subsequent availability of recombinant proteins have led to the development of candidate vaccines. Immunization with recombinant HEV capsid protein in HEV-susceptible primates has shown protection against hepatitis and viremia, although viral excretion is not prevented.51 An experimental HEV DNA vaccine has also been shown to induce serum anti-HEV that is protective against challenge with a heterologous HEV strain in cynomolgus macaques.52 A vaccine using a recombinant truncated form of HEV capsid protein, produced in Spodoptera frugiperda cells infected with a recombinant baculovirus and adsorbed on

1341

1342

Section IX  Liver aluminium hydroxide, has undergone a phase 2, doubleblind, randomized, placebo-controlled safety and efficacy trial among human volunteers in Nepal.53 In this trial, nearly 2000 young adults, almost all (>99%) male, were randomly assigned to receive three doses of either the recombinant HEV protein or a matched placebo (at 0, 1, and 6 months). The study subjects underwent active surveillance for clinical acute hepatitis E for a period extending over two years. The vaccine showed 95.5% protective efficacy against hepatitis E disease among those who received three doses and 88.5% efficacy in an intention-to-treat analysis among those who had received only the first dose. A high IgG anti-HEV level was observed in all the vaccinated volunteers at one month after the third vaccine dose, but only 56% of them had this high an antibody titer by the end of the study. More studies are needed to determine the efficacy of the vaccine among pregnant women, the duration of its protective efficacy, its effect on subclinical HEV infection and fecal shedding of the virus, and thus its impact on transmission of the virus. The vaccine may be useful for travelers to HEVendemic regions and for pregnant women and persons with chronic liver disease who reside in such regions. Finally, cost considerations will determine whether this new and promising vaccine can be used in populations living in developing countries where hepatitis E is endemic and vaccination is likely to be the most beneficial.

KEY REFERENCES

Aggarwal R, Krawczynski K. Hepatitis E: An overview and recent advances in clinical and laboratory research. J Gastroenterol Hepatol 2000; 15:9-20. (Ref 17.) Berke T, Matson DO. Reclassification of the Caliciviridae into distinct genera and exclusion of hepatitis E virus from the family on the basis of comparative phylogenetic analysis. Arch Virol 2000; 145:1421-36. (Ref 9.)

Dalton HR, Stableforth W, Thurairajah P, et al. Autochthonous hepatitis E in southwest England: natural history, complications and seasonal variation, and hepatitis E virus IgG seroprevalence in blood donors, the elderly and patients with chronic liver disease. Eur J Gastroenterol Hepatol 2008; 20:784-90. (Ref 20.) Fix AD, Abdel-Hamid M, Purcell RH, et al. Prevalence of antibodies to hepatitis E in two rural Egyptian communities. Am J Trop Med Hyg 2000; 62:519-23. (Ref 35.) Kamar N, Selves J, Mansuy JM, et al. Hepatitis E virus and chronic hepatitis in organ-transplant recipients. N Engl J Med 2008; 358:81117. (Ref 43.) Khuroo MS, Kamili S. Aetiology, clinical course and outcome of sporadic acute viral hepatitis in pregnancy. J Viral Hepat 2003; 10:61-9. (Ref 32.) Meng XJ, Purcell RH, Halbur PG, et al. A novel virus in swine is closely related to the human hepatitis E virus. Proc Natl Acad Sci U S A 1997; 94:9860-5. (Ref 13.) Purcell RH, Nguyen H, Shapiro M, et al. Pre-clinical immunogenicity and efficacy trial of a recombinant hepatitis E vaccine. Vaccine 2003; 21:2607-15. (Ref 51.) Reyes GR, Purdy MA, Kim JP, et al. Isolation of a cDNA from the virus responsible for enterically transmitted non-A, non-B hepatitis. Science 1990; 247:1335-9. (Ref 5.) Shata MT, Barrett A, Shire NJ, et al. Characterization of hepatitis Especific cell-mediated immune response using IFN-gamma ELISPOT assay. J Immunol Methods 2007; 328:152-61. (Ref 38.) Shrestha MP, Scott RM, Joshi DM, et al. Safety and efficacy of a recombinant hepatitis E vaccine. N Engl J Med 2007; 356:895-903. (Ref 53.) Srivastava R, Aggarwal R, Jameel S, et al. Cellular immune responses in acute hepatitis E virus infection to the viral open reading frame 2 protein. Viral Immunol 2007; 20:56-65. (Ref 40.) Tsarev SA, Tsareva TS, Emerson SU, et al. ELISA for antibody to hepatitis E virus (HEV) based on complete open-reading frame-2 protein expressed in insect cells: Identification of HEV infection in primates. J Infect Dis 1993; 168:369-78. (Ref 49.) Vishwanathan R. Infectious hepatitis in Delhi (1955-56): A critical study: Epidemiology. Indian J Med Res 1957; 45(Suppl 1):1-29. (Ref 1.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

81 Hepatitis Caused by Other Viruses Jordan J. Feld and E. Jenny Heathcote

CHAPTER OUTLINE Hepatitis G and GB Agent Infection  1343 Virology  1343 Epidemiology  1344 Clinical Features  1344 Diagnosis  1344 GB Virus Type C and Human Immunodeficiency Virus  1344 Treatment  1346 TT Virus Infection  1346 Virology  1346 Epidemiology  1346

A number of viruses may be hepatotropic in that viremia is occasionally associated with elevations in serum aminotransferase levels and viral replication may occur in hepatocytes, but little, if any, liver disease ensues. Such viruses include hepatitis G virus (HGV) and the GB agents, TT virus (TTV), Sanban virus, Yonban virus, SEN virus, and TTV-like mini-virus. Other novel agents such as the NV-F virus-like agent, which may exacerbate the severity of chronic hepatitis C, have been reported, but little is known about them. Other viral diseases may sometimes involve the liver as part of a systemic infection. The agents of such infections include human immunodeficiency virus (HIV) (see Chapter 33), Epstein-Barr virus (EBV), cytomegalovirus (CMV), herpes simplex virus (HSV), varicella-zoster virus (VZV), and the virus that causes severe acute respiratory syndrome (SARS). Infection with any of these viruses may rarely lead to severe, sometimes fatal, liver disease.

HEPATITIS G AND GB AGENT INFECTION During the long search for the cause of transfusionassociated non-A, non-B hepatitis (see Chapter 79), the GB agent (GBV) and hepatitis G virus (HGV) were discovered and later shown to be two isolates of the same virus. A 35-year-old surgeon with the initials GB developed an acute icteric hepatitis. When his serum was serially inoculated into healthy marmosets, they too developed hepatitis. Analysis of the marmosets infected with derivations of the GB serum led to the identification of two distinct viruses, labeled GBV-type A (GBV-A) and GBV-type B (GBV-B).1 A third virus closely related to the GB agents was subsequently identified by the same investigators from a human sample and was classified as GBV-C.2 At approximately the same time, another group independently identified a virus from

Clinical Features  1346 Treatment  1346 Sanban, Yonban, and SEN Viruses and TTV-Like Mini-virus Infections  1346 Systemic Viral Infections That May Involve the Liver  1347 Epstein-Barr Virus  1347 Cytomegalovirus  1347 Herpes Simplex Virus  1348 Varicella-Zoster Virus  1349 Coronavirus Causing Severe Acute Respiratory Syndrome  1349

the serum of a patient with cryptogenic non-A-to-E hepatitis, which they named HGV.3 Subsequent studies revealed 96% homology between the genomes of HGV and GBV-C, indicating that they were actually two strains of the same virus.4 Because large epidemiologic studies have not demonstrated any association between infection with GBV-C/ HGV and acute or chronic hepatitis, the use of the term “hepatitis G virus” has been questioned. In fact, even the index patient (“GB”) was subsequently shown to be infected with hepatitis C virus (HCV) as the cause of his liver disease. For clarity, GBV-C/HGV will be referred to as GBV-C.

VIROLOGY

GBV-C is a positive-strand RNA virus with a genome of 9400 nucleotides encoding approximately 2900 amino acids and is classified as a member of the Flaviviridae family. GBV-C shares 44% and 28% nucleotide homology with GBV-A and GBV-B, respectively, and has five known genotypes with distinct global geographic distributions.4 Although similar in many respects to HCV, GBV-C shares only 27% nucleotide homology with HCV, and the two viruses are clearly distinct. The GBV-C genome is organized like that of HCV. One long open reading frame encodes a single large polyprotein with structural proteins encoded at the 5′ aminoterminus and nonstructural proteins encoded at the 3′ carboxyterminus. A nontranslated region at the 5′ end serves as an internal ribosomal entry site (IRES), allowing translation of the uncapped messenger RNA.5 The structural proteins differ between HCV and GBV-C. Two glycoproteins—E1 and E2—predicted to compose the GBV-C viral envelope are cleaved from the polyprotein, likely by a host cell signal peptidase. Whereas HCV E1 and E2 have 5 and 11 N-linked glycosylation sites, GBV-C E1 and E2 have only 1 and 3 such sites, respectively.6 Perhaps of greater importance, amino acid polymorphisms do not cluster in the hypervariable region of GBC-C E2, as they do

1343

1344

Section IX  Liver in HCV E2. The hypervariable region of HCV E2 is thought to account in part for the ability of HCV to evade immune attack and cause persistent infection. This difference in E2 polymorphisms may account for the 60% to 75% rate of chronic HCV infection, compared with only 25% in GBV-C infection. The GBV-C genome does not encode a core protein, but biophysical and electron microscopic studies suggest that the virus does have a nucleocapsid structure, presumably with a core protein.7 Some GBV-C infected patients have antibodies that react with a synthetic peptide that corresponds to the region immediately proximal to the E1 coding region, suggesting that GBV-C may have a truncated core protein at the amino terminus of the genome. In contrast with the structural differences, the GBV-C nonstructural proteins—designated NS1, NS2, and so on—are similar to those of HCV, including a zinc-dependent thiol protease that cleaves NS2 from NS3, a serine protease at the amino terminus of NS3 that probably cleaves all downstream proteolytic sites, and an RNA helicase (as in all positive-strand RNA viruses) that is also found in NS3 downstream from the protease region. NS4A likely serves as a cofactor for the NS3 protease. The NS5A protein is believed to interact with double-strand RNA protein kinase, and NS5B serves as the RNA-dependent RNA polymerase.8 Another important difference between GBV-C and HCV may be tissue tropism. Negative-strand RNA (indicating the presence of active viral replication) has been demonstrated in liver tissue during HCV infection, implying hepatotropism, but has not been clearly demonstrated during GBV-C infection. Negative-strand RNA, however, has been demonstrated in peripheral blood mononuclear cells (PBMCs), bone marrow, and spleens of patients with GBV-C infection, suggesting that GBV-C is a lymphotropic virus.9 The demonstration of replication of a GBV-C clone in CD4+ T cells confirms that GBV-C is able to replicate in lymphocytes and may help explain the interaction between GBV-C and HIV infection (see later). GBV-C can be grown in cell culture and has been proposed as a model for studying HCV. Persistent GBV-C infection has been achieved in marmosets, which provide a small animal model of HCV infection in which a new serine protease inhibitor targeted to HCV has produced a marked reduction in GBV-C viremia.10 GBV-B has been proposed as a better model for HCV infection than GBV-C because GBV-B causes acute and rarely chronic hepatitis in tamarins and marmosets.11 Therefore, although the GB viruses do not cause overt human liver disease, they may be useful tools to study HCV infection.

EPIDEMIOLOGY

GBV-C is found worldwide. At least five genotypes have been identified, each with a specific geographic distribution: genotype 1 predominates in West Africa, genotype 2 in Europe and the United States, genotype 3 in parts of Asia, genotype 4 in Southeast Asia, and genotype 5 in South Africa.12 The development of GBV-C E2 antibodies correlates with loss of GBV-C viremia and suggests past exposure and clearance of GBV-C infection.13 Evidence of current and past GBV-C infection is found frequently in persons with parenteral risk factors and also is common among volunteer blood donors. Between 14% and 38% of persons with frequent exposures to blood are viremic with GBV-C, whereas 50% to 70% of such persons are seropositive for E2 anti­ bodies.14 Up to 16% of healthy blood donors are positive for E2 antibodies, with much lower rates of active viremia.15

Past or current GBV-C viremia is found as often in blood donors with a normal serum alanine aminotransferase (ALT) level as in donors rejected because of an elevated serum ALT level.16 Consequently, GBV-C transmission is not prevented by exclusion of donors with a normal ALT value. GBV-C also has been shown to be transmitted sexually and vertically much more frequently than HCV, and the risk may vary with the GBV-C genotype.17 No evidence exists for hepatitis or other clinical sequelae in infected babies.18 Because both GBV-C and HCV are transmitted parenterally, GBV-C-HCV coinfection is common. GBV-C viremia is present in about 20% of HCV-infected persons, and 80% of the remaining subjects are seropositive for antibodies to E2.8 These findings suggest that the rate of natural clearance of GBV-C is higher (~75%) than that for HCV (~25%).

CLINICAL FEATURES

Although GBV-C is detected in many patients with non-Ato-E acute and chronic hepatitis and may persist for years, it does not appear to cause liver (or any other) disease, even in immunocompromised persons.16,19-21 Nor does GBV-C appear to modulate the course or response to treatment of chronic HCV or hepatitis B virus (HBV) infection.21 GBV-C infection does not affect the outcome of liver transplantation, even though liver transplant recipients have high rates of GBV-C viremia, likely because of their high transfusion requirements.22 The duration of GBV-C infection may depend on the immune status and age of the host. As with HBV infection, childhood acquisition of GBV-C appears to lead to chronic infection, whereas sexual transmission in immunocompetent adults typically leads to rapid clearance of viremia.23 HIV-infected persons are also more likely than non–HIVinfected persons to develop chronic GBV-C infection.24 In contrast to HCV infection, the development of antibodies to GBV-C E2 seems to protect against reinfection.25 No association between GBV-C and hepatocellular carcinoma,26 non-Hodgkin’s lymphoma,27 aplastic anemia, or lichen planus28 has been documented.

DIAGNOSIS

Because GBV-C rarely causes disease in humans, diagnostic tests are not widely available and generally are reserved for research purposes. GBV-C RNA can be detected by using polymerase chain reaction (PCR) methodology with commercially available primers. A test for GBV-C antibody, to document past infection, is also available.

GB VIRUS TYPE C AND HUMAN IMMUNODEFICIENCY VIRUS

Once GBV-C was shown not to cause liver disease, interest in this virus diminished. In 1998, however, two independent groups of investigators observed that among a small number of HIV-infected persons, lower HIV viral loads as well as slower progression to acquired immunodeficiency syndrome (AIDS) and death correlated with the presence of GBV-C viremia (Fig. 81-1).29,30 Subsequently, larger studies have confirmed this finding in varied populations and have suggested that HIV–GBV-C–coinfected persons have a better prognosis than HIV-monoinfected persons.30-34 In addition, coinfected persons respond better to antiretroviral therapy, with a more rapid increase in CD4+ counts and suppression of HIV viral load.34,35 Children infected with GBV-C at birth also have a reduced risk of contracting HIV via vertical transmission, independent of the mother’s GBV-C status.36 Some studies, however, have not confirmed a protective effect of GBV-C infection in HIV-infected persons.37-41 The

Chapter 81  Hepatitis Caused by Other Viruses 1.0

A Proportion of patients surviving

GBV-C viremic .8

.6

Anti-E2 positive

.4

.2

No GBV-C exposure

0 0

1000

2000

3000

4000

5000

6000

7000

Days after first diagnosis of HIV infection Number at risk GBV-C–positive 33 Anti-E2–positive 112 Unexposed 52

31 101 37

25 72 27

19 35 12

11 16 3

6 6 0

1 0 0

0 0 0

1.0

B Proportion of patients surviving

GBV-C viremic .8

.6 Anti-E2 positive .4

.2

No GBV-C exposure

0 0

1000

3000

2000

Days after determination of GBV-C status Number at risk GBV-C–positive 33 Anti-E2–positive 112 Unexposed 52

28 65 21

25 44 11

21 38 9

12 24 4

discrepant results may relate to differences in the HIVinhibitory effect of different GBV-C genotypes and the duration of GBV-C infection42; the longer duration of GBV-C infection, the greater the benefit on HIV progression.33 The viral interaction may work in both directions. In a longitudinal study of coinfected persons, GBV-C viral titers increased during effective highly active antiretroviral therapy (HAART) and fell with interruptions in HAART, suggesting that HIV may also inhibit GBV-C.43 Although GBV-C does not affect the outcome of HCV monoinfection, a study of the effect of GBV-C infection on the progression of liver disease in HIV-HCV coinfected persons found that patients with current or past GBV-C infection had a

1 0 0

Figure 81-1.  Survival according to GBV-C status in patients with human immunodeficiency virus (HIV) infection. Survival from the time of diagnosis of HIV infection (A) and survival from the time the blood sample was drawn to determine the GBV-C status (B) are shown in relation to the GBV-C status. For both measures, the patients who were seropositive for GBV-C RNA had significantly better survival. Anti-E2, antibody to E2 glycoprotein. (From Tillmann HL, Heiken H, Knapik-Botor A, et al. Infection with GB virus C and reduced mortality among HIV-infected patients. N Engl J Med 2001; 345:715-24.)

lower risk of cirrhosis and hepatic complications but no difference in survival compared with patients without GBV-C infection.44 Why HIV-GBV-C coinfected persons have an improved response to HAART is unknown. HIV replication is diminished in PBMCs that are infected with both HIV and GBV-C, and a higher inoculum of HIV is required to infect cells already infected with GBV-C. The NS5A protein of GBV-C has been shown to down-regulate the expression of the HIV co-receptor CXCR4, and the E2 protein of GBV-C interferes with early steps in the HIV life cycle.45 Virologic inhibition may explain the interaction, but immune mechanisms, including GBV-C-induced stimulation of HIV-inhibitory

1345

1346

Section IX  Liver cytokines, endogenous interferon production, and promotion of a T helper cell type 1 (Th1) immune profile, have also been proposed.46,47

TREATMENT

Because GBV-C infection is not associated with clinical liver disease, no treatments have targeted GBV-C specifically. In HIV–HCV–GBV-C coinfected persons, peginterferon and ribavirin treatment led to sustained GBV-C clearance in 31% of patients, with no observable subsequent effect on the course of HCV or HIV infection.48 In patients coinfected with GBV-C and HCV who were treated with interferon and ribavirin, GBV-C RNA disappeared from serum during therapy but reappeared in all patients following discontinuation of therapy.49 Importantly, no effect of GBV-C infection was observed on the response to treatment of HCV infection.50

TT VIRUS INFECTION TTV was first identified in 1977 by the use of representational difference analysis in a patient (with the initials TT) in Japan who had acute post-transfusion non-A-to-G hepatitis.51 TTV is also referred to as the transfusion-transmitted and Torque-Teno virus.

VIROLOGY

TTV is a non-enveloped, single-stranded, negative-polarity, circular DNA virus. It is closely related to a family of animal viruses known as Circoviridae, which have not been associated with human disease. TTV is the first human singlestranded circular DNA virus to be identified and does not fit precisely into any known virus family. Other TTV-like viruses were subsequently discovered and together make up the human Anellovirus family. The TTV genome is 3965 nucleotides long and contains at least three overlapping open reading frames. Three messenger RNAs (mRNAs) are expressed by TTV; each mRNA is translated from two start codons, leading to the production of six viral proteins. One protein product of the largest mRNA (3.0 kb) functions as the capsid protein and also serves as a replicase enzyme. The product of the second open reading frame interferes with host nuclear factor kappa B signaling, but its role in viral replication is unknown.52 TTV displays remarkable genomic sequence diversity, even among isolates found in an individual patient.53 At least 23 genotypes have been identified, with greater than 30% sequence divergence between isolates. Some genotypes differ in sequence by greater than 50%, and genotype prevalence rates vary geographically.54 TTV is believed to be hepatotropic on the basis of the observation that viral levels are higher in the liver than in the serum of infected patients. TTV has also been identified within hepatocytes, and shown to replicate, by in situ hybridization and PCR; however, no or only minor morphologic changes have been seen in cells with positive hybridization signals.55 TTV has also been shown to replicate in stimulated PBMCs and bone marrow cells.56

EPIDEMIOLOGY

TTV is found worldwide and is common. Initial studies documented infection in 1% to 40% of healthy blood donors.57 As more inclusive primers have been used to detect differing genotypes, the reported prevalence among blood donors has increased dramatically, approaching 100% in some studies.58 The prevalence of TTV infection

increases with age but appears to reach a plateau by early childhood.59 TTV is also found in a variety of nonhuman primate species. TTV is transmitted effectively by all parenteral routes; high prevalence rates have been documented in hemophiliacs, injection drug users, patients on hemodialysis, and organ transplant recipients.60 TTV has also been shown to be transmitted enterically, and high TTV DNA titers are present in the feces of viremic patients. Fecal-oral spread may account for the high prevalence in low-risk, healthy blood donor populations.58

CLINICAL FEATURES

Although TTV was associated with acute hepatitis in the patient in whom it was first identified, other studies have not supported a causal association between TTV and liver disease.61,62 In the original study,51 viremia was detected six weeks after exposure and two weeks before a rise in serum ALT levels. The viral DNA was cleared from serum, as documented by PCR testing, and serum ALT levels subsequently returned to normal. Viremia may persist for years in both immunocompetent and immunosuppressed persons. Most infected persons have no biochemical or histologic evidence of liver disease.61,63 Like GBV-C, TTV does not appear to alter the natural history or response to treatment of either chronic HCV or HBV infection.64 A report that TTV infection may increase the risk of hepatocellular carcinoma in patients with chronic HCV infection remains to be confirmed.65

TREATMENT

Formal studies of treatment of TTV infection have not been performed. A small study of HCV-TTV-coinfected patients showed that TTV infection had no effect on sustained virologic response of HCV to therapy with peginterferon and ribavirin. Although TTV viremia cleared after therapy in 6 of 10 patients, 4 of the 6 relapsed within six months.66

SANBAN, YONBAN, AND SEN VIRUSES AND TTV-LIKE MINI-VIRUS INFECTIONS Since the discovery of TTV in 1997,51 several similar viruses with a small DNA genome have been isolated in Japan and named Sanban, Yonban, and TTV-like mini-virus (TLMV). These viruses have been divided into 29 genotypes, with sequence divergence of greater than 30%.67 Like TTV, they are readily transmitted parenterally and can also be passed by the fecal-oral route. None has been clearly associated with human liver disease to date. In 1999, a novel virus was identified in an HIV-positive person (with the initials SEN) who had post-transfusion hepatitis of unknown etiology. This virus was found with the use of degenerative primers from the prototype TTV. The SEN virus is a small, nonenveloped, single-stranded DNA virus, but unlike that of TTV, the SEN genome is linear. Nucleotide sequencing has shown 50% homology with the prototype TTV, but only 30% of amino acids are homologous. Sequencing of multiple isolates has demonstrated sequence divergence of 15% to 50%.68 Like TTV, SEN virus is transmitted both parenterally and by the fecal-oral route.69 Vertical transmission occurs but in most cases does not lead to chronic infection. Natural clearance of both perinatally and parenterally acquired SEN virus does not appear to protect against reinfection.70 The prevalence varies markedly and is highest among

Chapter 81  Hepatitis Caused by Other Viruses patients with parenteral risk factors, particularly those coinfected with HCV.71 The prevalence among healthy blood donors is approximately 2% in the United States and 10% in Japan.72 The clinical significance of SEN infection remains controversial. One study of patients with post-transfusional nonA-to-E hepatitis suggested that SEN was the cause in a majority (11 of 12) of cases. SEN viremia persisted for more than one year in 45% of those infected; however, clinical hepatitis did not develop in a majority (86%) of transfused patients who acquired SEN infection. None of the patients with hepatitis had a fulminant course, nor did chronic liver disease or cirrhosis develop during follow up.73 Other reports and case series have identified SEN or TTV viremia in patients with both fulminant and chronic hepatitis, but causation has been difficult to establish. Most studies have shown neither an association between SEN or any of the other viruses in this group and human disease, nor an effect of these viruses on the course or response to treatment for chronic viral hepatitis.69,74

SYSTEMIC VIRAL INFECTIONS THAT MAY INVOLVE THE LIVER EPSTEIN-BARR VIRUS

EBV infection is common and covers a wide spectrum of clinical presentations. Most infants and children are either asymptomatic or have mild, nonspecific complaints, whereas adolescents and adults typically present with the triad of pharyngitis, fever, and lymphadenopathy.75 Although usually subclinical, liver involvement is nearly universal in patients with EBV mononucleosis and ranges from serum aminotransferase elevations to rare cases of fulminant and even fatal hepatitis.76 Up to 90% of patients with acute mononucleosis have serum aminotransferase and lactate dehydrogenase elevations two to three times the upper limit of normal. The enzyme levels typically rise over a one- to two-week period, and peak levels are usually much lower than those normally seen in patients with acute hepatitis A, B, D, or E. Elevated levels of alkaline phosphatase are common, and mild hyper­ bilirubinemia is observed in as many as 45% of cases.77 In most patients, liver test levels normalize within one month, often before complete resolution of clinical symptoms.78 As with infectious mononucleosis, EBV hepatitis tends to be more severe in adults older than age 30 than in younger adults and children.79 Elderly patients occasionally present with jaundice, fever, and right upper quadrant pain that may suggest a diagnosis of extrahepatic biliary obstruction.80 Although jaundice may be caused by viral-induced cholestasis, autoimmune hemolytic anemia should be excluded in all hyperbilirubinemic patients. Cholestatic jaundice with pruritus may be observed in young women with EBV infection who continue taking oral contraceptive pills. Rarely, EBV may cause persistent chronic infection with involvement of the liver, lungs, and other organs. In the largest series of fatal mononucleosis, hepatic involvement was universal and was the cause of death in 13 of 30 patients.81 Granulomatous hepatitis also has been attributed to chronic EBV infection on occasion.82 EBV infection detected on a liver biopsy specimen may account for some cases of otherwise unexplained hepatitis. Fatal, fulminant EBV hepatitis has been described in both immunocompetent and immunocompromised persons and appears to be associated with a greater than usual EBV viral burden, particularly in T cells as opposed to B cells.83 A hemophago-

cytic syndrome characterized by fever, hepatosplenomegaly, hepatic synthetic dysfunction, cytopenias, and marked hyperferritinemia may develop in patients with EBV infection. The syndrome is caused by natural killer T cell dysregulation, leading to lymphoctye proliferation and activation with uncontrolled hemophagocytosis and cytokine production. Although rare, the syndrome is usually severe and may be fatal.84 The diagnosis of EBV hepatitis is based on clinical features of mononucleosis and laboratory data suggestive of acute EBV infection. Most patients (70%) have a leukocytosis with a predominance of lymphocytes and monocytes, and up to 50% have mild thrombocytopenia.85 The Monospot test is sensitive for the detection of heterophile antibodies but is not a specific test for EBV infection. Levels of EBVspecific immunoglobulin M (IgM) antibodies peak early in serum and may persist for many months, at which point IgG antibodies appear. Findings on abdominal ultrasonography are usually nonspecific and may include hepatosplenomegaly, lymphadenopathy, and possibly gallbladder thickening, which has been reported to portend more severe liver disease.86 Liver biopsy is rarely necessary for diagnosis but, if done, shows portal and sinusoidal mononuclear cell infiltration with no disruption of hepatic architecture; multinucleated giant cells are not a feature. In more severe cases, focal hepatic necrosis may be evident. In situ hybridization or PCR testing of biopsy samples may be used to confirm the diagnosis, but immunohistochemistry for EBV proteins is rarely positive.87 No specific treatment for EBV hepatitis exists. Acyclovir inhibits EBV replication and reduces viral shedding from the nasopharynx but has no effect on clinical symptoms or outcome.88 Improvement in acute and chronic EBV hepatitis has been reported with gancyclovir treatment, but this approach has not been well studied.89 Liver transplantation has been performed for fulminant EBV hepatitis. EBV rarely causes hepatitis after liver transplantation but has been associated with post-transplantation lymphoproliferative disease (see Chapter 95). In a patient with chronic EBV hepatitis after kidney transplantation, treatment with rituximab was reported to lead to improvement in serum liver enzyme levels and hepatic pathology.90

CYTOMEGALOVIRUS

CMV is the largest member of the Herpesviridae family and, like other herpesviruses, persists lifelong in a latent, nonreplicative state after resolution of primary infection. Consequently, clinical disease caused by CMV may occur as a primary infection or, more commonly, as reactivation of latent infection,91 particularly in immunocompromised persons. In immunocompetent children and adults, primary CMV infection is usually subclinical but may cause a mononucleosis-like illness. Liver involvement is common and is characterized by mild to moderate aminotransferase (88%) and alkaline phosphatase (64%) elevations with or without hepatosplenomegaly.92 Although the clinical course is mild in most patients, rare instances of granulomatous, cholestatic CMV hepatitis, with or without jaundice, and even massive, fatal hepatic necrosis have been described.93 In addition to the congenital CMV syndrome (jaundice, hepatosplenomegaly, thrombocytopenic purpura, and severe neurologic impairment), CMV is a common cause of neonatal hepatitis.94 A number of case reports also suggest a possible association between CMV infection and acute portal vein thrombosis; however, the mechanism is unclear.95 Disseminated, life-threatening CMV infection with multiorgan involvement may develop in patients with impaired

1347

1348

Section IX  Liver

BD

A

B

BD

C

D

Figure 81-2.  Ultrasonographic findings in AIDS cholangiopathy and cytomegalovirus infection. A, A thick rind of echogenic tissue (arrows) surrounds the central portal triads and causes irregular narrowing of the intrahepatic bile ducts. B, The bile duct (BD) is dilated and its wall is minimally irregular. C, The dilated BD tapers abruptly at an echogenic, inflamed ampulla (arrow), indicative of papillary stenosis. D, The ampulla (arrow) is enlarged and echogenic, as viewed transversely in the caudal aspect of the pancreatic head. AIDS, acquired immunodeficiency syndrome.

cell-mediated immunity (see Chapters 33 and 34). Hepatobiliary involvement by CMV is common in patients with AIDS and may manifest as hepatitis, pancreatitis, or acalculous gangrenous cholecystitits.96 CMV also may cause AIDSassociated cholangiopathy, which manifests with chronic cholestasis and mimics primary sclerosing cholangitis (PSC) clinically and radiographically.97 Patients may have papillary stenosis alone or in combination with intra- or extrahepatic (or both) biliary stricturing and dilatation (Fig. 81-2; see also Fig. 33-7). Antiviral therapy has no effect on this syndrome, but papillotomy, with or without placement of a biliary stent, may lead to symptomatic improvement.96 Organ transplant recipients are also at risk for aggressive CMV hepatitis, including fibrosing cholestatic hepatitis (see Chapter 95), but for unclear reasons cholangiopathy does not develop in these patients.98 CMV hepatitis can be difficult to distinguish from graft rejection in liver transplant recipients.99 The diagnosis of CMV infection is based on results of serologic studies, liver biopsy, or both. In acute primary CMV infection, IgM antibodies to CMV are present. For patients with reactivation of latent CMV, “shell vial” assays, which use monoclonal antibodies to detect CMV antigens, or direct CMV antigenemia testing are necessary. Because CMV viremia precedes organ involvement, testing for CMV antigenemia or CMV PCR testing of blood are useful screen-

ing tools in immunocompromised patients.100 Multinucleated giant cells with mononuclear portal and parenchymal inflammatory infiltrates and cholestasis are commonly seen on liver biopsy specimens. Large nuclear inclusions, sometimes referred to as “owl’s eye” inclusions may be seen in hepatocytes or biliary epithelial cells (Fig. 81-3). With mild CMV disease in an immunocompetent adult, treatment is unnecessary. In immunocompromised patients, antiviral therapy is indicated. Ganciclovir, a guanosine nucleoside analog with a much longer intracellular half-life than that of acyclovir, has proved to be most effective. The major toxicity is bone marrow suppression, particularly granulocytopenia. Because viremia correlates with disease outcome, ganciclovir should be continued until CMV antigenemia is undetectable.101 For patients resistant to or intolerant of ganciclovir, alternative agents include foscarnet and cidofovir.

HERPES SIMPLEX VIRUS

HSV typically causes mucocutaneous vesicular oral or genital lesions; visceral involvement may occur in certain clinical settings. HSV hepatitis is seen in neonates, pregnant women, and immunocompromised persons and can be aggressive and life-threatening.102 Cases of severe HSV hepatitis in immunocompetent persons have been reported as well.103 Severe hepatitis with multiorgan involvement

Chapter 81  Hepatitis Caused by Other Viruses

Figure 81-3.  Histopathologic findings in cytomegalovirus hepatitis. In the center (arrow) is a large hepatocyte with a large nucleus that contains an “owl’s eye” inclusion. (Hematoxylin and eosin.) (Courtesy of Maha Guindi, MD, Toronto, Canada.)

and, often, adrenal insufficiency, may develop in neonates exposed to infected maternal genital secretions at the time of delivery.104 In pregnant women, HSV hepatitis usually has a fulminant course. The disease is most common in late gestation, typically (in 65% of patients) in the third trimester. Mucocutaneous lesions are present in only 50% of cases, and a high index of clinical suspicion is important to ensure timely diagnosis.105 Maternal and perinatal mortality rates approach 40%, and in the largest series, 25% of patients were diagnosed only at autopsy. Early diagnosis and initiation of antiviral therapy are critical.106 Mild, asymptomatic liver enzyme elevations may be seen in 14% of immunocompetent patients with acute genital HSV infection. By contrast, immunocompromised patients may present with fulminant hepatitis.107 Hepatitis is more common with acute infection than reactivation and presents with fever, leukopenia, and markedly elevated serum aminotransferase levels. Coagulopathy, including disseminated intravascular coagulation, and jaundice may be seen. Of reported cases, only 50% had a rash at presentation, and 58% were diagnosed at autopsy.108 Liver biopsy is essential for diagnosis, particularly in pregnancy. The transjugular route may be required because liver failure may develop rapidly, precluding percutaneous biopsy. Focal or extensive hemorrhagic or coagulative necrosis, with few inflammatory infiltrates, is seen. Intranuclear (Cowdry A type) inclusions may be identified in hepatocytes at the margins of the necrosis. In addition, some periportal multinucleated hepatocytes show a ground-glass appearance suggestive of viral inclusions (Fig. 81-4).109 Electron microscopy, immunohistochemical staining, and PCR techniques can be used to confirm the diagnosis.110 HSV hepatitis constitutes an emergency, and empirical therapy should be instituted pending diagnostic confirmation. High-dose intravenous acyclovir (at least 10 mg/kg every eight hours) is effective and appears to be safe in pregnancy.111 Prolonged therapy may be required because severe relapse has been reported.103 Successful liver transplantation has also been reported.112

VARICELLA-ZOSTER VIRUS

Like HSV infection, VZV infection occasionally can be complicated by hepatitis. Serum liver enzyme levels may be elevated in up to 3.4% of children with chickenpox;

Figure 81-4.  Histopathologic findings in herpes simplex hepatitis. At the edge of a necrotic zone, some hepatocytes are multinucleated, and many nuclei contain eosinophilic viral (Cowdry type A) inclusions. (Hematoxylin and eosin.) (From Lucas SB. Other viral and infectious diseases and HIVrelated liver disease. In Burt AD, Portmann BC, Ferrell LD, editors. Pathology of the Liver, 5th ed. London: Churchill Livingstone; 2007. p 446.)

however, clinically significant hepatitis has been reported only rarely.113 Although VZV reactivation in adults usually is limited to the skin, dissemination with liver, lung, and pancreatic involvement may occur.114 Rarely, visceral involvement has been reported to develop before cutaneous manifestations in bone marrow or solid organ transplant recipients. If visceral involvement is suspected, treatment with high-dose intravenous acyclovir should be instituted.

CORONAVIRUS CAUSING SEVERE ACUTE RESPIRATORY SYNDROME

During the 2003 outbreak of SARS, a rapidly progressive respiratory illness, elevated serum aminotransferase levels were commonly observed during the acute illness. Subsequently, cases of SARS hepatitis were reported in which the coronavirus that causes SARS was demonstrated in the liver by reverse transcriptase–PCR techniques in three patients; no viral particles were seen on electron microscopy. All three cases fulfilled the World Health Organization criteria for SARS. Examination of liver tissue revealed marked apoptosis, ballooning of hepatocytes, and moderate lobular lymphocytic infiltration.115

KEY REFERENCES

Alter HJ, Nakatsuji Y, Melpolder J, et al. The incidence of transfusionassociated hepatitis G virus infection and its relation to liver disease. N Engl J Med 1997; 336:747-54. (Ref 16.) Alter MJ, Gallagher M, Morris TT, et al. Acute non-A-E hepatitis in the United States and the role of hepatitis G virus infection. Sentinel Counties Viral Hepatitis Study Team. N Engl J Med 1997; 336:741-6. (Ref 20.) Berzsenyi MD, Bowden DS, Kelly HA, et al. Reduction in hepatitis C-related liver disease associated with GB virus C in human immunodeficiency virus coinfection. Gastroenterology 2007; 133:1821-30. (Ref 44.) Charlton M, Adjei P, Poterucha J, et al. TT-virus infection in North American blood donors, patients with fulminant hepatic failure, and cryptogenic cirrhosis. Hepatology 1998; 28:839-42. (Ref 63.) Goodgame RW. Gastrointestinal cytomegalovirus disease. Ann Intern Med 1993; 119:924-35. (Ref 91.) Linnen J, Wages J Jr, Zhang-Keck ZY, et al. Molecular cloning and disease association of hepatitis G virus: A transfusion-transmissible agent. Science 1996; 271:505-8. (Ref 3.) Markin RS. Manifestations of Epstein-Barr virus-associated disorders in liver. Liver 1994; 14:1-13. (Ref 81.)

1349

1350

Section IX  Liver Pellise M, Miquel R. Liver failure due to herpes simplex virus. J Hepatol 2000; 32:170. (Ref 110.) Rouphael NG, Talati NJ, Vaughan C, et al. Infections associated with haemophagocytic syndrome. Lancet Infect Dis 2007; 7:814-22. (Ref 84.) Sheng WH, Hung CC, Wu RJ, et al. Clinical impact of GB virus C viremia on patients with HIV type 1 infection in the era of highly active antiretroviral therapy. Clin Infect Dis 2007; 44:584-90. (Ref 37.) Stapleton JT. GB virus type C/Hepatitis G virus. Semin Liver Dis 2003; 23:137-48. (Ref 8.) Tanaka E, Alter HJ, Nakatsuji Y, et al. Effect of hepatitis G virus infection on chronic hepatitis C. Ann Intern Med 1996; 125:740-3. (Ref 21.)

Tillmann HL, Heiken H, Knapik-Botor A, et al. Infection with GB virus C and reduced mortality among HIV-infected patients. N Engl J Med 2001; 345:715-24. (Ref 32.) Williams CF, Klinzman D, Yamashita TE, et al. Persistent GB virus C infection and survival in HIV-infected men. N Engl J Med 2004; 350:981-90. (Ref 33.) Xiang J, McLinden JH, Chang Q, et al. Characterization of a peptide domain within the GB virus C NS5A phosphoprotein that inhibits HIV replication. PLoS ONE 2008; 3:e2580. (Ref 45.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

82 Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess Arthur Y. Kim and Raymond T. Chung

CHAPTER OUTLINE Bacterial Infections Involving or Affecting the Liver  1351 Gram-Positive and Gram-Negative Bacteria  1351 Chlamydia  1353 Rickettsia  1353 Spirochetes  1354 Tuberculosis and Other Mycobacteria  1355 Parasites  1355 Protozoa  1355 Helminths  1359

The liver serves as the initial site of filtration of absorbed intestinal luminal contents and is particularly susceptible to contact with microbial antigens of all varieties. In addition to infection by viruses (see Chapters 77 to 81), the liver can be affected by (1) spread of bacterial or parasitic infection from outside the liver; (2) primary infection by spirochetal, protozoal, helminthic, or fungal organisms; or (3) systemic effects of bacterial or granulomatous infections.

BACTERIAL INFECTIONS INVOLVING OR AFFECTING THE LIVER GRAM-POSITIVE AND GRAM-NEGATIVE BACTERIA

A number of extrahepatic infections can lead to derangements of hepatic function, ranging from mild abnormalities of liver biochemical tests to frank jaundice and, rarely, hepatic failure.

Toxic Shock Syndrome: Staphylococcus aureus or Group A Streptococci

Toxic shock syndrome is a multisystem disease caused by toxic shock syndrome toxins, which are superantigens that cause T cell activation and massive cytokine release. Origi-

Lawrence S. Friedman served as an author of this chapter on previous editions of this textbook.

Fungi  1365 Candidiasis  1365 Histoplasmosis  1365 Liver Abscess  1366 Pyogenic  1366 Amebic  1368

nally described in association with serious infections caused by Staphylococcus aureus, this syndrome is now more frequently a complication of group A streptococcal infections, particularly necrotizing fasciitis.1 Risk factors for S. aureus toxic shock syndrome include tampon use and surgical wound infection. Typical findings include a scarlatiniform rash, mucosal hyperemia, hypotension, vomiting, and diarrhea.2 Hepatic involvement is almost always present and can range from elevations of serum aminotransferase levels to jaundice and extensive necrosis. Histologic findings in the liver include microabscesses and granulomas. The diagnosis is confirmed by culture of toxigenic Streptococcus pyogenes or S. aureus from the wound, blood, or other body sites. For wound infections or necrotizing fasciitis, surgical intervention is critical. Clindamycin, in conjunction with another active agent, is recommended to interfere with bacterial toxin production. Antibiotics effective against S. aureus include nafcillin for methicillin-sensitive isolates and vancomycin or linezolid for methicillin-resistant isolates, whereas penicillin remains active against S. pyogenes. Intravenous immunoglobulin may have a benefit in the setting of toxic shock associated with S. pyogenes.3

Clostridia

Clostridial myonecrosis involving Clostridium perfringens usually is a mixed anaerobic infection that results in the rapid development of local wound pain, abdominal pain, and diarrhea. The skin lesions become discolored and even bullous, and gas gangrene spreads rapidly, leading to a high mortality rate. Jaundice may develop in up to 20% of patients with gas gangrene and is predominantly a conse-

1351

1352

Section IX  Liver quence of massive intravascular hemolysis caused by an exotoxin elaborated by the bacterium.4 Evidence of liver involvement may include abscess formation and gas in the portal vein. Hepatic involvement does not appear to affect mortality. The presence of clostridial bacteria portends a poor prognosis in persons with cirrhosis.5 Surgical débridement with wide excision is essential; penicillin and clindamycin are effective antibiotics.

Actinomyces

Actinomycosis is caused most commonly by Actinomyces israelii, a gram-positive anaerobic bacterium. Although cervicofacial infection is the most frequent manifestation of actinomycotic infection, gastrointestinal involvement occurs in 13% to 60% of patients.6,7 Hepatic involvement is present in 15% of cases of abdominal actinomycosis and is believed to result from metastatic spread from other abdominal sites. Common presenting manifestations of actinomycotic liver abscess include fever, abdominal pain, and anorexia with weight loss.8,9 The course is more indolent than that seen with the usual causes of pyogenic hepatic abscess and thus may be mistaken for a tumor.8 Fistula formation and invasion of other surrounding tissues such as the pleural space can occur. Anemia, leukocytosis, an elevated erythrocyte sedimentation rate, and an elevated serum alkaline phosphatase level are nearly universal. Radiographic findings are nonspecific; multiple abscesses may be seen in both lobes of the liver. The diagnosis is based on aspiration of an abscess cavity and either visualization of characteristic sulfur granules or positive results on an anaerobic culture. Most abscesses resolve with prolonged courses of intravenous penicillin or oral tetracycline. Large abscesses can be drained percutaneously or resected surgically.10

Listeria

Hepatic invasion in adult human Listeria monocytogenes infection is uncommon. One report described thirty-four cases of listeriosis involving the liver, ranging from solitary to multiple abscesses and acute and granulomatous hepa­ titis.11 Hepatic histologic features include multiple abscesses and granulomas. Predisposing conditions include immunosuppression, diabetes mellitus, and underlying liver disease, including cirrhosis, hemochromatosis, or chronic hepatitis. The diagnosis of disseminated listerial infection is based on a positive culture result from blood or isolation from an aspirate in the case of a liver abscess. Treatment is with ampicillin or penicillin, often with gentamicin for synergy.12

Shigella and Salmonella

Several case reports have described cholestatic hepatitis attributable to enteric infection with Shigella.13,14 Histologic findings in the liver have included portal and periportal infiltration with polymorphonuclear neutrophils, hepatocyte necrosis, and cholestasis. Typhoid fever, caused by Salmonella typhi, is a systemic infection that frequently involves the liver. Elevation of serum aminotransferase levels is common, whereas the serum bilirubin level may rise in a minority of cases.15 Some patients may present with an acute hepatitis-like picture, characterized by fever and tender hepatomegaly.16 Cholecystitis and liver abscess may complicate hepatic involvement with S. typhi infection.17 Hepatic damage by S. typhi appears to be mediated by bacterial endotoxin, although organisms can be visualized within the liver tissue. Endotoxin may produce focal necro-

sis, a periportal mononuclear infiltrate, and Kupffer cell hyperplasia in the liver. These changes resemble those seen in gram-negative sepsis. Characteristic typhoid nodules scattered throughout the liver are the result of profound hypertrophy and proliferation of Kupffer cells. The clinical course can be severe, with a mortality rate approaching 20%, particularly with delayed treatment or in patients with other complications of Salmonella infection. The suggestion has been made that severe typhoid fever with jaundice and encephalopathy can be differentiated from acute liver failure by the presence of an elevated serum alkaline phosphatase level, mild hypoprothrombinemia, thrombocytopenia, hepatomegaly, and an aspartate aminotransferase (AST) level greater than the alanine aminotransferase (ALT) level.18 Ciprofloxacin and ceftriaxone are first-line agents for the treatment of typhoid fever. S. paratyphi A and B (Salmonella enterica serotypes paratyphi A and B) are the predominant causes of paratyphoid fever. As in typhoid fever, abnormalities in liver bioche­ mical tests, particularly serum aminotransferase levels, with or without hepatomegaly, are common.19 Liver abscess is a rare complication.20 Treatment is with a third-generation cephalosporin or a fluoroquinolone.

Yersinia

Infection with Yersinia enterocolitica manifests as ileocolitis in children and as terminal ileitis or mesenteric adenitis in adults. Arthritis, cellulitis, erythema nodosum, and septicemia may complicate Yersinia infection. Most patients with complicated disease have an underlying comorbid condition, such as diabetes mellitus, cirrhosis, or hemochromatosis. Excess tissue iron, in particular, may be a predisposing factor because growth of the Yersinia bacterium is enhanced by iron. The subacute septicemic form of the disease resembles typhoid fever or malaria. Multiple abscesses are distributed diffusely in the liver and spleen. In some cases, the occurrence of Y. enterocolitica liver abscesses may lead to the detection of underlying hemochromatosis.21,22 The mortality rate is approximately 50%. Fluoroquinolones are the preferred antibiotics.

Gonococci

In approximately 50% of patients with disseminated gonococcal infection, serum alkaline phosphatase levels are elevated, and in 30% to 40% of patients, AST levels are elevated.23 Jaundice is uncommon. The most common hepatic complication of gonococcal infection is the Fitz-Hugh–Curtis syndrome, a perihepatitis that is believed to result from direct spread of the infection from the pelvis (see later).23 Clinically, patients describe a sudden, sharp pain in the right upper quadrant. The pain may be confused with that of acute cholecystitis or pleurisy. Most patients have a history of pelvic inflammatory disease. The syndrome is distinguished from gonococcal bacteremia by a characteristic friction rub over the liver and negative blood culture results. The diagnosis is made by vaginal culture for gonococci. The overall prognosis of gonococcal infection appears to be unaffected by the presence of perihepatitis.24 Ceftriaxone is the antibiotic of choice.

Legionella

Legionella pneumophila, a fastidious gram-negative bacterium, is the cause of Legionnaires’ disease. Although pneumonia is the predominant clinical manifestation, abnormal liver biochemical test results are frequent, with elevations in serum aminotransferase levels in 50%, alkaline phospha-

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess tase levels in 45%, and bilirubin levels in 20% of cases (but usually without jaundice). Involvement of the liver does not influence clinical outcome. Liver histologic changes include microvesicular steatosis and focal necrosis; organisms can be seen occasionally. The diagnosis is confirmed by direct fluorescence of antibody in the serum or sputum or of antigen in the urine.25 The antibiotic of choice is azithromycin or a fluoroquinolone.

Burkholderia pseudomallei (Melioidosis)

Burkholderia pseudomallei is a soil-borne and water-borne gram-negative bacterium that is found predominantly in Southeast Asia. The clinical spectrum of melioidosis ranges from asymptomatic infection to fulminant septicemia with involvement of the lungs, gastrointestinal tract, and liver. Histologic changes in the liver include inflammatory infiltrates, multiple microabscesses, and focal necrosis. Organisms can be visualized with a Giemsa stain of a liver biopsy specimen.26 With chronic disease, granulomas may be seen. Some liver abscesses may demonstrate a “honeycombing” appearance on computed tomography.27 Abscesses may need to be drained or débrided, and ceftazidime or meropenem is the initial drug of choice followed by a prolonged course of trimethoprim-sulfamethoxazole, with or without doxycycline.28

Brucella

Brucellosis may be acquired from infected pigs, cattle, goats, and sheep (Brucella suis, Brucella abortus, Brucella melitensis, and Brucella ovis, respectively) and typically manifests as an acute febrile illness. Hepatic abnormalities are seen in a majority of infected persons, and jaundice may be present in severe cases. Typically, multiple noncaseating hepatic granulomas are found in liver biopsy specimens; less often, focal mononuclear infiltration of the portal tracts or lobules is seen.29 Rarely, brucellosis also may produce hepatosplenic abscesses.30,31 The diagnosis can be made by isolation of the organism from a cultured specimen of liver tissue and is confirmed by serologic testing in combination with a history of exposure to animals. Surgical drainage may be required for management of Brucella abscesses. The combination of streptomycin and doxycycline is the most effective antimicrobial therapy.

Coxiella burnetii (Q Fever)

Infection by Coxiella burnetii, typically acquired by inhalation of animal dusts, causes the clinical syndrome of Q fever, which is characterized by relapsing fevers, headache, myalgias, malaise, pneumonitis, and culture-negative endocarditis. Liver involvement is common.32 The predominant abnormality is an elevated serum alkaline phosphatase level, with minimal elevations of AST or bilirubin levels. The histologic hallmark in the liver is the presence of characteristic fibrin ring granulomas. The diagnosis is confirmed by serologic testing for complement-fixing antibodies.33 Treatment with doxycycline usually is effective.

Bartonella (Oroya Fever)

Endemic to Colombia, Ecuador, and Peru, Bartonella bacilliformis is a gram-negative coccobacillus that causes an acute febrile illness accompanied by jaundice, hemolysis, hepatosplenomegaly, and lymphadenopathy. Centrilobular necrosis of the liver and splenic infarction may occur. As many as 40% of patients die of sepsis or hemolysis. Prompt treatment with chloramphenicol in combination with penicillin, clindamycin, or trimethoprim-sulfamethoxazole prevents fatal complications.34

Bacillary Angiomatosis and the Acquired Immunodeficiency Syndrome

Bacillary angiomatosis is an infectious disorder that primarily affects persons with the acquired immunodeficiency syndrome (AIDS) or other immunodeficiency states. The causative agents have been identified as the gram-negative bacilli Bartonella henselae and, in some cases, Bartonella quintana.35 Infection frequently is associated with exposure to cats. Bacillary angiomatosis is characterized most commonly by multiple blood-red papular skin lesions, but disseminated infection with or without skin involvement also has been described.36 The causative bacilli can infect liver, lymph nodes, pleura, bronchi, bones, brain, bone marrow, and spleen. Additional manifestations include persistent fever, bacteremia, and sepsis. Hepatic infection should be suspected when serum aminotransferase levels are elevated in the absence of other explanations. Hepatic infection in persons with bacillary angiomatosis may manifest as peliosis hepatis, or blood-filled cysts (see Chapter 83). Histologically, peliosis in patients with AIDS is characterized by an inflammatory myxoid stroma containing clumps of bacilli and dilated capillaries surrounding the blood-filled peliotic cysts. Increasingly, diagnosis of Bartonella infection is by polymerase chain reaction (PCR)-based methods.37 Bacillary angiomatosis responds uniformly to therapy with erythromycin. For visceral infection, prolonged treatment with erythromycin or doxycycline should be administered.38

Bacterial Sepsis and Jaundice

Jaundice may complicate systemic sepsis caused by gramnegative or gram-positive organisms. Exotoxins and endotoxin liberated in overwhelming infection can directly or indirectly, through cytokines such as tumor necrosis factor-α (TNF-α), inhibit the transport of bile acids and other organic anions across the hepatic sinusoidal and bile canalicular membranes, thereby leading to intrahepatic cholestasis (see Chapter 20).39 Serum bilirubin levels can reach 15 mg/dL or higher. The magnitude of the jaundice does not correlate with mortality. Results of cultures of liver biopsy specimens usually are negative.

CHLAMYDIA Fitz-Hugh–Curtis Syndrome

Although perihepatitis was first associated with gonococcal salpingo-oophoritis (see earlier), it is now most frequently associated with Chlamydia trachomatis infection.40 The presentation is similar to perihepatitis caused by gonococcal infection, with right upper quadrant pain accompanying a urogenital infection such as pelvic inflammatory disease. The diagnosis can be made by direct visualization at laparoscopy or laparotomy and supported by pathologic demonstration of endometritis, salpingitis, and microbiologic detection of C. trachomatis in the genital tract. Liver biochemical test results are generally normal. The treatment of choice is a single dose of azithromycin or seven days of doxycycline.

RICKETTSIA Rocky Mountain Spotted Fever

Mortality from Rocky Mountain spotted fever, a systemic tick-borne rickettsial illness, has decreased considerably as a result of prompt recognition of the classic maculopapular rash in association with fever and an exposure history. A

1353

1354

Section IX  Liver small subset of patients, however, present with multiorgan manifestations and have a high mortality rate.41 A characteristic severe vasculitis develops in these patients and is believed to be the result of a microbe-induced coagulo­ pathy. Hepatic involvement is frequent in multiorgan disease. In one postmortem study, rickettsiae were iden­ tified in the portal triads of eight of nine fatal cases. Portal tract inflammation, portal vasculitis, and sinusoidal erythrophago­cytosis were consistent findings, but hepatic necrosis was negligible. The predominant clinical mani­ festation was jaundice; elevations of serum aminotrans­ ferase and alkaline phosphatase levels varied. Jaundice probably results from a combination of inflammatory bile ductular obstruction and hemolysis and is associated with increased mortality.32,42

caused by immune complexes.45 Spirochetes are seen in renal tubules in a majority of autopsy specimens but rarely are found in the liver. Hepatic histologic findings generally are nonspecific and do not include necrosis. Altered mitochondria and disrupted membranes in hepatocytes on electron microscopy suggest the possibility of a toxin-mediated injury. The diagnosis of leptospirosis is made on clinical grounds in conjunction with a positive result of a blood or urine culture specimen in the first and second phase, respectively. Serologic testing confirms the diagnosis when culture results are unrevealing. Doxycycline is effective if given within the first several days of illness. Most patients recover without residual organ impairment.

Ehrlichia

Secondary Syphilis Liver involvement is characteristic of secondary syphilis.47 The frequency of hepatitis in secondary syphilis ranges from 1% to 50%.47,48 Symptoms and signs usually are nonspecific, including anorexia, weight loss, fever, malaise, and sore throat. A characteristic pruritic maculopapular rash involves the palms and soles. Jaundice, hepatomegaly, and tenderness in the right upper quadrant are less common. Almost all patients exhibit generalized lymphadenopathy. Biochemical testing generally reveals low-grade elevations of serum aminotransferase and bilirubin levels, with a disproportionate elevation of the serum alkaline phosphatase level; isolated elevation of the alkaline phosphatase is common.49 Proteinuria may be present. Histologic examination of the liver in syphilitic hepatitis generally discloses focal necrosis in the periportal and centrilobular regions. The inflammatory infiltrate typically includes polymorphonuclear neutrophils, plasma cells, lymphocytes, eosinophils, and mast cells.47,48 Kupffer cell hyperplasia may be seen, but bile ductule injury is rare. Granulomas may be seen. Spirochetes may be demonstrated by silver staining in as many as 50% of patients. Resolution of these findings without sequelae follows treatment with penicillin.

Ehrlichiae are rickettsiae that parasitize leukocytes. In the United States, human monocytic ehrlichiosis is caused principally by Ehrlichia chaffeensis and, less often, by Ehrlichia canis. Human granulocytic anaplasmosis (formerly known as human granulocytic ehrlichiosis) is caused by Anaplasma phagocytophilum.32,43 In contrast with Rocky Mountain spotted fever, a rash is often absent. Hepatic involvement is seen in greater than 80% of cases, usually in the form of mild, transient serum aminotransferase elevations. More marked aminotransferase elevations may occur occasionally, in association with cholestasis, hepatosplenomegaly, and liver failure. Liver injury is attributable to proliferation of organisms within hepatocytes and provocation of an immune response. Focal necrosis, fibrin ring granulomas, and cholestatic hepatitis can be observed. A mixed portal tract infiltrate and lymphoid sinusoidal infiltrate usually are seen. The disease generally resolves with appropriate antibiotic therapy with doxycycline.44

SPIROCHETES Leptospirosis

Leptospirosis is one of the most common zoonoses in the world and it has a wide range of domestic and wild animal reservoirs. Humans acquire the spirochete by contact with infected urine or contaminated soil or water. In humans, disease can occur as anicteric leptospirosis or as Weil’s syndrome (described subsequently). Anicteric leptospirosis accounts for more than 90% of cases and is characterized by a biphasic illness. The first phase begins, often abruptly, with viral illness-like symptoms associated with fever, leptospiremia, and conjunctival suffusion, which serves as an important diagnostic clue. Following a brief period of improvement, the second phase in 95% of cases is characterized by myalgias, nausea, vomiting, abdominal tenderness, and, in some cases, aseptic meningitis.45 During this phase, a few patients have elevated serum aminotransferase and bilirubin levels with hepatomegaly. Weil’s syndrome is a severe icteric form of leptospirosis and constitutes 5% to 10% of all cases. The first phase of this illness often is marked by jaundice, which may last for weeks. During the second phase, fever may be high, and hepatic and renal manifestations predominate. Jaundice may be marked, with serum bilirubin levels approaching 30 mg/dL (predominantly conjugated). Serum aminotransferase levels usually do not exceed five times the upper limit of normal.46 Acute tubular necrosis often develops and can lead to renal failure, which may be fatal. Hemorrhagic complications are frequent and are the result of capillary injury

Syphilis

Tertiary (Late) Syphilis Tertiary syphilis is now rare. Although hepatic lesions are common in late syphilis, most patients are asymptomatic. Some patients describe anorexia, weight loss, fatigue, fever, or abdominal pain. The characteristic hepatic lesion in tertiary syphilis is the gumma, which can be single or multiple. It is necrotic centrally, with surrounding granulation tissue consisting of a lymphoplasmacytic infiltrate and endarteritis; exuberant deposition of scar tissue may occur, giving the liver a lobulated appearance (hepar lobatum). If hepatic involvement is unrecognized, hepatocellular dysfunction and portal hypertension with jaundice, ascites, and gastroesophageal varices can ensue. Hepatic gummas may resolve after therapy with penicillin.50

Lyme Disease

Lyme disease is a multisystem disease caused by the tick-borne spirochete Borrelia burgdorferi. Predominant manifestations are dermatologic, cardiac, neurologic, and musculoskeletal. Hepatic involvement has been described. Among 314 patients, abnormal liver biochemical test results with generally increased serum aminotransferase and lactate dehydrogenase levels were seen in 19%.51 Clinical findings included anorexia, nausea and vomiting, weight loss, right upper quadrant pain, and hepatomegaly, usually within days to weeks of the onset of illness and often accompanied by the sentinel rash, erythema migrans.52

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess In early stages of the illness, the spirochetes are believed to disseminate hematogenously from the skin to other organs, including the liver.53 One report has suggested that the Lyme spirochete also can cause acute hepatitis as a manifestation of reactivation,54 although the possibility of reinfection cannot be fully excluded. Histologic examination of the liver in Lyme hepatitis reveals hepatocyte ballooning, marked mitotic activity, microvesicular fat, Kupffer cell hyperplasia, a mixed sinusoidal infiltrate, and intraparenchymal and sinusoidal spirochetes.53 The diagnosis of Lyme disease is confirmed with serologic studies in patients with a typical clinical history. Hepatic involvement tends to be more frequent in disseminated disease but does not appear to affect overall outcome, which is excellent in primary disease after institution of treatment with oral doxycycline, amoxicillin, clarithromycin, or azithromycin.55 Ceftriaxone is the drug of choice for late disease.44,53

TUBERCULOSIS AND OTHER MYCOBACTERIA

Granulomas are found in liver biopsy specimens in approximately 25% of persons with pulmonary tuberculosis and 80% of those with extrapulmonary tuberculosis. Tuberculous granulomas can be distinguished from sarcoid granulomas by central caseation, acid-fast bacilli, and the presence of fewer granulomas, with a tendency to coalesce.56 Multiple granulomas in the liver also may be seen following vaccination with bacille Calmette-Guérin, especially in persons with an impaired immune response. Patients with multiple granulomas caused by tuberculosis rarely have clinically significant liver disease. Occasionally, tender hepatomegaly is found. Jaundice with elevated serum alkaline phosphatase levels may occur in miliary infection. The treatment of tuberculous granulomatous disease of the liver is the same as that for active pulmonary tuberculosis— namely, four-drug therapy.56 Hepatic involvement in Mycobacterium avium complex infection is discussed in Chapter 33.

Hepatocyte Sporozoite

Mosquito

Hypnozoite (P. vivax, P. ovale)

Erythrocytic schizonts

Pathobiology of the Plasmodium Life Cycle Malarial sporozoites injected by an infected mosquito circulate to the liver and enter hepatocytes. Maturation to schizonts ensues. When the schizont ruptures, merozoites are released into the bloodstream, where they enter erythrocytes. The major species of Plasmodium responsible for malaria differ with respect to the number of merozoites released and the maturation times. Infection by Plasmodium falciparum and Plasmodium malariae is not associated with a residual liver stage after the release of merozoites, whereas infection by Plasmodium vivax and Plasmodium ovale is associated with a persistent exoerythrocytic stage, the hypnozoite, which persists in the liver and, when activated, can divide and mature into schizont forms. Plasmo-

Erythrocytes

Figure 82-1.  The life cycle of Plasmodium species.

Table 82-1  Parasitic Infections of the Liver and Biliary Tract: Classification by Pathologic Process PATHOLOGIC PROCESS Liver Disease Granulomatous hepatitis

Portal fibrosis Hepatic abscess or necrosis Cystic liver disease Peliosis hepatis Reticuloendothelial Disease Kupffer cell infection or hyperplasia Biliary Tract Disease Cholangitis

PARASITES (Tables 82-1 and 82-2)

An estimated 300 to 500 million persons in more than 100 countries are infected with malaria each year. The liver is affected during two stages of the malarial life cycle: first in the pre-erythrocytic phase, and then in the erythrocytic phase, which coincides with clinical illness. The life cycle of the prototypical malarial parasite is illustrated in Figure 82-1.

Merozoites

Gametocytes

Biliary hyperplasia

PROTOZOA (see also Chapter 109) Malaria

Preerythrocytic schizont (all species)

Cholangiocarcinoma

DISEASES Capillariasis Fascioliasis Schistosomiasis Strongyloidiasis Toxocariasis Schistosomiasis Amebic abscess Toxoplasmosis Echinococcosis Bacillary angiomatosis Babesiosis Malaria Toxoplasmosis Visceral leishmaniasis Clonorchiasis/opisthorchiasis Fascioliasis Ascariasis Clonorchiasis Cryptosporidiosis Fascioliasis Clonorchiasis/opisthorchiasis

dium knowlesi has been identified as a fifth species capable of infecting humans and occasionally results in severe manifestations including jaundice, hepatic dysfunction, and acute kidney injury.57 The extent of hepatic injury varies with the malarial species (most severe with P. falciparum) and the severity of infection. Unconjugated hyperbilirubinemia most commonly is seen as a result of hemolysis, but hepatocellular dysfunction is also possible, leading to conjugated hyperbilirubinemia. Moderate elevations of serum aminotransferase and 5′-nucleotidase levels may be observed.58 Synthetic dysfunction (e.g., prolongation of the prothrombin time, hypoalbuminemia) may be seen as well. In severe falciparum malaria, hypoglycemia and lactic acidosis are late and life-threatening complications.59 Reversible reductions in portal venous blood flow have been described during the acute phase of falciparum malaria, presumably as a consequence of micro-occlusion of portal venous branches by parasitized erythrocytes.59

1355

United States

Eurasia, Central America, South America

Worldwide

Babesiosis (Babesia spp.)

Visceral leishmaniasis (Leishmania donovani)

Toxoplasmosis (Toxoplasma gondii)

Worldwide

Exposure to deer tick

Africa, Asia, South America

Malaria (Plasmodium falciparum, P. malariae, P. vivax, P. ovale, P. knowlesi)

Nematodes Toxocariasis (Toxocara canis, T. cati)

Blood transfusion, intravenous drug use

Worldwide, especially Africa, Asia, Mexico, South America

Protozoans Amebiasis (Entamoeba histolytica)

Exposure to dogs or cats, especially with children younger than five years of age

Congenital infection, immunosuppression (AIDS, organ transplantion)

Immunosuppression (AIDS, organ transplantation)

Poor sanitation, sexual transmission

ENDEMIC AREAS

DISEASE (CAUSE)

PREDISPOSIng factors

Table 82-2  Parasitic Diseases of the Liver and Biliary Tract

Migration of larvae to the liver (visceral larva migrans)

Replication in the liver leading to inflammation, necrosis

Infection of RE cells

Hemolysis with multiorgan involvement

Sporozoite clearance by hepatocytes; exoerythrocytic replication in the liver

Hematogenous spread and tissue invasion, abscess formation (see Fig. 82-10)

PATHOPHYSIOLOGY

Granuloma formation with eosinophilia

Fever, lymphadenopathy, occasionally hepatosplenomegaly, atypical lymphocytosis

Fever, weight loss, hepatosplenomegaly, secondary bacterial infection, skin hyperpigmentation (kala-azar)

Fever, anemia, hepatosplenomegaly, abnormal liver test results, hemoglobinuria

Tender hepatomegaly, splenomegaly, rarely hepatic failure (P. falciparum)

Fever, RUQ pain, peritonitis, elevated right hemidiaphragm, rupture

MANIFESTATIONS

Larvae in tissue, serology (ELISA)

Serology (IF, ELISA), isolation of the organism in the tissue

Amastigotes seen in the spleen, liver, or bone marrow

Identification of the parasite on a blood smear, PCR

Identification of the parasite on a blood smear

Cysts in the stool, serology (e.g., ELISA, CIE, IHA), hepatic imaging

DIAGNOSIS

Albendazole 10 mg/kg/d × 5 d or mebendazole, 100-200 mg bid × 5 d

Metronidazole 750 mg [PO or IV] tid × 7-10 d or tinidazole 2 g × 3 d, followed by iodoquinol 650 mg tid × 20 d or diloxanide furoate 500 mg tid × 10 d or aminosidine (paromomycin) 25-35 mg/kg/d in 3 divided doses × 7-10 d P. falciparum: Chloroquine (chloroquine-sensitive); mefloquine; or quinine and either doxycycline or clindamycin; or pyrimethamine-sulfadoxine (Fansidar); or atovaquone/ proguanil (chloroquine-resistant); or artesiminins. P. malariae: Chloroquine P. vivax, P. ovale, P. knowlesi. Chloroquine and primaquine (chloroquine-sensitive) or mefloquine and primaquine (chloroquine-resistant) (eliminate exoerythrocytic forms)† Azithromycin 500 mg on day 1, then 250 mg qd and atovaquone 750 mg bid × 7-10 d or clindamycin 300-600 mg IV q6h or 600 mg PO q8h and quinine 650 mg q8h × 7-10 d Pentavalent antimonial (stibogluconate sodium and meglumine antimoniate) 20 mg/kg/d × 28 d; or liposomal amphotericin B [IV] 3 mg/kg/d on days 1-5, 14, and 21; or aminosidine (paromomycin) 16-20 mg/kg/d × 21 d; or pentamidine isethionate, 2-4 mg/kg/d for up to 15 d; or miltefosine, 2.5 mg/kg/d × 28 d Pyrimethamine, 100 mg loading dose followed by 25-50 mg/d; plus sulfadiazine, 2-4 g/d in 4 divided doses; or clindamycin, 300 mg four times daily, plus folinic acid, 10-25 mg daily for 2-4 wk

TREATMENT*

1356 Section IX  Liver

Worldwide

Tropical climates

Asia, Africa, South America, Southern Europe, United States

Temperate climates

Hepatic capillariasis (Capillaria hepatica)

Ascariasis (Ascaris lumbricoides)

Strongyloidiasis (Strongyloides stercoralis)

Trichinosis (Trichinella spiralis)

Southeast Asia, China, Japan, Korea, Eastern Europe

Clonorchiasis and opisthorchiasis (Clonorchis sinensis, Opisthorchis viverrini, O. felineus) Cestodes Echinococcosis (Echinococcus granulosus, E. multilocularis)

Migration of larvae to the liver; encystment (hydatid cyst)

Migration of larvae through the liver; penetration of the bile ducts or surgery Migration through the ampulla; egg deposition in the bile ducts

Fibrogenic host immune response to eggs in the portal vein

Hematogenous dissemination to the liver

Larval penetration from the intestine to the liver

Migration of larvae to the liver; invasion of the bile ducts by adult worms

Migration of larvae to the liver; inflammatory reaction to eggs

PATHOPHYSIOLOGY

Tender hepatomegaly, fever, eosinophilia, cyst rupture, biliary obstruction

Acute: eosinophilic infiltrate; chronic: hepatosplenomegaly, presinusoidal portal hypertension, perioval granuloma formation Acute: fever, abdominal pain, jaundice, hemobilia; chronic: hepatomegaly Biliary hyperplasia, obstruction, sclerosing cholangitis, stone formation, cholangiocarcinoma

Occasionally jaundice, biliary obstruction, larvae in hepatic sinusoids

Hepatomegaly, occasionally jaundice, larvae in the portal tract or lobule

Acute, subacute hepatitis, tender hepatomegaly, occasionally splenomegaly, eosinophilia Abdominal pain, fever, jaundice, biliary obstruction, perioval granulomas

MANIFESTATIONS

Serology (ELISA, IHA), hepatic imaging

Ova in the stool, flukes in the bile ducts at ERC or surgery

Surgical resection or percutaneous drainage. Perioperative albendazole 400 mg bid continuing × 8 wk

Praziquantel 75 mg/kg in 3 divided doses × 1 d

Praziquantel 40-60 mg/kg in 2-3 divided doses × 1 d; or oxamniquine (not available in United States). Acute toxemic schistosomiasis: Praziquantel 40-60 mg/kg in 2-3 divided doses × 1 d + glucocorticoids Triclabendazole 10 mg/kg × 1 dose

Ova in the stool, rectal or liver biopsy

Ova in the stool, flukes in the bile ducts at ERC

Glucocorticoids for allergic symptoms; albendazole 400 mg bid × 10-15 d; or mebendazole 200 mg/d × 10-15 d

Albendazole 400 mg × 1 dose; or mebendazole 100 mg bid × 3 d; or pyrantel pamoate 11 mg/kg up to 1g; or ivermectin 200 µg/kg × 1 dose Ivermectin 200 µg/kg/d × 2 d; or albendazole 400 mg/d × 3 d

Supportive; possibly dithiazine iodide, sodium stibogluconate, albendazole, or thiabendazole

TREATMENT*

History, eosinophilia, fever, muscle biopsy

Larvae in the stool or duodenal aspirate

Ova or adult in stool or contrast study

Adult worms or eggs in a liver biopsy specimen (see Fig. 82-2)

DIAGNOSIS

*All drugs are given orally unless otherwise specified. † For dosing guidelines for malaria, please refer to http://www.cdc.gov/malaria/pdf/treatmenttable.pdf. AIDS, acquired immunodeficiency syndrome; CDC, Centers for Disease Control and Prevention; CIE, counterimmunoelectrophoresis; d, day; ELISA, enzyme-linked immunosorbent assay; ERC, endoscopic retrograde cholangiography; IF, immunofluorescence; IHA, indirect hemagglutination assay; IV, intravenously; PO, by mouth; PCR, polymerase chain reaction assay; RE, reticuloendothelial; RUQ, right upper quadrant.

Cattle and sheep raising (E. granulosus)

Cattle or sheep raising; ingestion of contaminated watercress Ingestion of raw fresh-water fish

Worldwide

Fascioliasis (Fasciola hepatica)

Worldwide

Travelers exposed to bodies of fresh water

Immunosuppression (AIDS, chemotherapy, organ transplantation) predisposes to hyperinfection Ingestion of undercooked pork

Ingestion of raw vegetables

Exposure to rodents

PREDISPOSIng factors

Asia, Africa, South America, Caribbean

Trematodes Schistosomiasis (Schistosoma mansoni, S. japonicum)

ENDEMIC AREAS

DISEASE (CAUSE)

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess 1357

1358

Section IX  Liver Histopathologic Features In acute falciparum malaria in a previously unexposed person, hepatic macrophages hypertrophy, and large quantities of malarial pigment (the result of hemoglobin degradation by the parasite) accumulate in Kupffer cells, which phagocytose parasitized and unparasitized erythrocytes.60 Histopathologic features include Kupffer cell hyperplasia with pigment deposition and a mononuclear infiltrate. Hepatocyte swelling and centrizonal necrosis may be seen. All abnormalities are reversible with treatment. Clinical Features Only the erythrocytic stage of malaria is associated with clinical illness. Symptoms and signs of acute infection typically develop 30 to 60 days following exposure and include fever, which often is hectic, malaise, anorexia, nausea, vomiting, diarrhea, and myalgias. Jaundice caused by hemolysis is common in adults, especially in heavy infection with P. falciparum. In general, hepatic failure is seen only in association with concomitant viral hepatitis or with severe P. falciparum infection.61,62 One series iden­ tified evidence of hepatic encephalopathy in 15 of 86 patients with falciparum malaria and jaundice; four cases were fatal.61 Tender hepatomegaly with splenomegaly is common. Cytopenias are common in acute infection. The differential diagnosis includes viral hepatitis, gastroen­ teritis, amebic liver abscess, yellow fever, typhoid, tuberculosis, and brucellosis. Diagnosis The diagnosis of acute malaria rests on the clinical history, physical examination, and identification of parasites on peripheral thin and thick blood smears. Because the number of parasites in the blood may be small, repeated smear examinations should be performed by an experienced examiner when the index of suspicion is high. P. knowlesi may resemble P. malariae in morphology, and PCR-based tests may help distinguish these two species.57 Rapid antigen detection assays are available but are less reliable than other diagnostic approaches.63 Treatment The treatment of acute malaria depends on the species of parasite and, for falciparum infection, the pattern of chloroquine resistance. Chloroquine generally is effective in areas endemic for chloroquine-sensitive species. Resistant falciparum infections can be treated with mefloquine alone; quinine and either doxycycline or clindamycin; pyrimethamine-sulfadoxine (Fansidar); a combination of atovaquone and proguanil; or artemisinin derivatives including artemisinin, artemether, and artesunate.64 For P. vivax and P. ovale infections, the addition of primaquine (in persons without glucose-6-phosphate dehydrogenase deficiency) to chloroquine or mefloquine is indicated to eliminate the exoerythrocytic hypnozoites in the liver.65 Hyperreactive Malarial Splenomegaly (Tropical Splenomegaly Syndrome) In endemic areas, repeated exposure to malaria may lead to an aberrant immunologic response characterized by overproduction of B lymphocytes, circulating malarial antibody, and increased levels of circulating immune complexes, resulting in dense hepatic sinusoidal lymphocytosis and stimulation of the reticuloendothelial cell system. The clinical picture includes massive splenomegaly, markedly elevated antimalarial antibody levels, and high serum immunoglobulin M (IgM) levels. Severe debilitating anemia caused by hypersplenism, especially in women of child-

bearing age, can result.66 Variceal bleeding is uncommon but may result from portal hypertension consequent to markedly increased splenic and portal venous blood flow. Treatment consists of lifelong antimalarial therapy and blood transfusions.

Babesiosis

Babesiosis, caused by Babesia species, is a malaria-like illness transmitted by the deer tick Ixodes scapularis.67 The disease is endemic to coastal areas of the Northeast and areas of the Midwest in the United States. Clinical features include fever, anemia, mild hepatosplenomegaly, abnormalities on liver biochemical tests, hemoglobinuria, and hemophagocytosis on bone marrow biopsy specimen. The disease is especially severe in asplenic and immunocompromised patients. In rare cases, marked pancytopenia occurs. Hepatic involvement reflects the severity of the systemic illness but generally is not severe. Uncomplicated cases are treated with a combination of the following active agents: (1) oral azithromycin, 500 mg single dose followed by 250 mg once daily, plus atovaquone, 750 mg twice daily, for 7 to 10 days; or (2) oral clindamycin, 600 mg three times daily, in combination with quinine, 650 mg three times daily, for 7 to 10 days. In severe cases, the clindamycin may be given intravenously and partial or complete exchange transfusion should be considered.44

Leishmaniasis

Visceral leishmaniasis is caused by Leishmania donovani and is endemic in the Mediterranean, central Asia, the former Soviet Union, the Middle East, China, India, Pakistan, Bangladesh, Africa, Central America, and South America.68 This entity should be considered in returning travelers and military personnel from these areas. Amastigotes are ingested by the sand fly (Lutzomyia in the New World, Phlebotomus in the Old World) and become flagellated promastigotes. Following injection into the human host, the promastigotes are phagocytosed by macrophages in the reticuloendothelial system, where they multiply. Histopathologic Features In visceral leishmaniasis, organisms usually can be found in mononuclear phagocytes of the liver, spleen, bone marrow, and lymph nodes. Proliferation of Kupffer cells often is seen, and amastigotes (Leishman-Donovan bodies) can be detected within these cells.69 Occasionally, parasitebearing cells aggregate within noncaseating granulomas.70 Hepatocyte necrosis can range in degree from mild to severe. Healing is accompanied by fibrous deposition, and occasionally the liver takes on a cirrhotic appearance. Nevertheless, complications of chronic liver disease are rare. Clinical Features Visceral infection caused by L. donovani begins with a papular or ulcerative skin lesion at the site of the sand fly bite. Following an incubation period of two to six months (sometimes years), intermittent fevers, weight loss, diarrhea (of bacillary, amebic, or leishmanial origin), and progressive painful hepatosplenomegaly develop, often accompanied by pancytopenia and a polyclonal hypergammaglobulinemia. Secondary bacterial infections resulting from suppression of reticuloendothelial cell function are important causes of mortality and include pneumonia, pneumococcal infection, and tuberculosis. Physical findings include hepatomegaly, massive splenomegaly, jaundice or ascites in severe disease, generalized lymphadenopathy, and muscle wasting.71 Cutaneous gray hyperpigmentation, which prompted the name kala-azar

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess (black fever), is characteristically seen in patients in India. Oral and nasopharyngeal nodules resulting from granuloma formation also may be seen.

ing techniques and tissue culture systems can provide adjunctive diagnostic support. PCR analysis of serum and liver also can be helpful in ambiguous cases.84

Diagnosis The diagnosis is based on the history, physical examination, and microscopic demonstration of amastigotes by a Wright or Giemsa stain of affected tissue samples. The highest yield (90%) comes from aspiration of the spleen. Liver biopsy is less risky and associated with a yield nearly as great as that of splenic aspiration. The yield of bone marrow aspirates is 80%, and that of lymph node aspirates is 60%.54 Culture requires specialized media and may take several weeks. Serologic testing (enzyme-linked immunosorbent assay [ELISA], immunofluorescence, direct agglutination) can be used to support a presumptive diagnosis of visceral leishmaniasis but is insensitive, particularly in immunocompromised hosts.72 The leishmanin skin test (Montenegro test) is not helpful in acute visceral disease. PCR-based testing of blood or other tissue may also be useful for diagnosis as well as monitoring.73

Treatment Antibiotic therapy should be administered to all persons with severe symptomatic infection and to immunocompromised or pregnant patients with acute uncomplicated infection. Treatment consists of a combination of pyrimethamine and sulfadiazine, plus folinic acid to minimize hematologic toxicity, for two to four weeks.78

Treatment Pentavalent antimonial compounds are the drugs of choice for all forms of leishmaniasis. Parenteral sodium stibogluconate and meglumine antimoniate are available through the Centers for Disease Control and Prevention for treatment of infections in the United States. Treatment with antimonials should be administered for at least four weeks. Alternative parenteral agents include liposomal amphotericin B and aminosidine (paromomycin).74 Patients with AIDS and leishmaniasis often fail to respond to or relapse following treatment with conventional regimens.72 Miltefosine, a phosphocholine analog administered orally, has shown promise in visceral leishmaniasis, with a reported cure rate of 82% to 97%.75,76

Toxoplasmosis

Toxoplasmosis, caused by Toxoplasma gondii, is found worldwide. In the United States, serologic surveys suggest that exposure to T. gondii has decreased from 14% to 9% among persons ages 12 to 49.77 The infection may be transmitted congenitally or occur as an opportunistic infection that causes cerebral mass lesions in patients with AIDS. Oocysts of T. gondii in soil, water, or contaminated meat are ingested and mature in the intestinal tract of humans to become sporozoites, which penetrate the intestinal mucosa, become tachyzoites, and circulate systemically, invading a wide array of cell types.78 Hepatic involvement has been observed in severe, disseminated infection. Clinical Features Although most primary infections are asymptomatic, acquired toxoplasmosis can manifest as a mononucleosislike illness with fever, chills, headache, and regional lymphadenopathy.79 Hepatomegaly, splenomegaly, and minimal elevations of serum aminotransferase levels are uncommon findings.80,81 Infections of immunocompromised hosts can result in pneumonia, myocarditis, encephalitis, and, rarely, hepatitis.78,82 Toxoplasmosis can produce atypical lym­ phocytosis, an otherwise unusual feature of parasitic disease. Diagnosis The diagnosis is best made by detecting specific IgM or IgG antibody using highly specific indirect immunofluorescence or an enzyme immunoassay.83 Specialized histologic stain-

HELMINTHS (see also Chapter 110) Nematodes (Roundworms)

Nematodes are nonsegmented roundworms that have a thick cuticle covering the body. Toxocariasis and capillariasis manifest with major hepatobiliary features, whereas ascariasis, strongyloidiasis, and trichinosis affect the liver less frequently or less severely. Toxocariasis Toxocara canis and Toxocara cati infect dogs and cats, respectively. Infection occurs worldwide, especially in children, and is acquired when embryonated eggs are ingested in soil or contaminated food. The eggs hatch in the small intestine and release larvae that penetrate the intestinal wall, enter the portal venous circulation, and reach the liver and systemic circulation. Blocked by narrowing vascular channels, the immature worms bore through vessel walls and migrate through the tissues, where they cause hemorrhagic, necrotic, and secondary inflammatory responses. When larvae become trapped in tissue, they provoke granuloma formation with a predominance of eosinophils. Tissue larvae may remain in inflammatory capsules or granulomas for months to years. The liver, brain, and eye are affected most frequently.85 Clinical Features.  Most infected persons are asymptomatic. Two clinical syndromes are recognized (1) visceral larva migrans and (2) “occult” infections associated with nonspecific symptoms, including abdominal pain, anorexia, fever, and wheezing.85 Visceral larva migrans is seen most commonly in children with a history of pica. Findings include fever, hepatomegaly, urticaria, leukocytosis with persistent eosinophilia, hypergammaglobulinemia, and elevated blood group isohemagglutinins.85 Toxocariasis has been implicated in the development of chronic cholestatic hepatitis86 as well as pyogenic liver abscess.87 Pulmonary manifestations include asthma and pneumonitis. Neurologic involvement can result in focal or generalized seizures, encephalopathy, and abnormal behavior.85 Ocular larva migrans often is associated with visual loss and strabismus and can manifest as a unilateral raised retinal lesion that resembles an ocular tumor. Diagnosis.  The possibility of toxocariasis should be considered in persons with a history of pica, exposure to dogs or cats, and persistent eosinophilia.88 Stool studies are not useful for toxocariasis because these organisms do not produce eggs in humans, nor do they remain in the gastrointestinal tract. A definitive diagnosis is made by identification of the larvae in affected tissues, although blind biopsies are not routinely recommended.85 The finding of an eosinophilic granuloma may be specific for visceral larva migrans.89 A liver biopsy may be necessary to differentiate visceral larva migrans from hepatic capillariasis (see later). A strongly positive result on an ELISA using larval antigens provides support for the diagnosis.

1359

1360

Section IX  Liver Treatment.  Treatment is primarily supportive, as visceral larva migrans is generally self-limited. If required, antihelminthic therapy with albendazole, 10 mg/kg/day in two divided doses for five days, or mebendazole, 100 to 200 mg twice daily for five days, may be used. Severe pulmonary, cardiac, ophthalmologic, or neurologic manifestations may warrant use of systemic glucocorticoids.85 Hepatic Capillariasis Human infection with Capillaria hepatica is rare. Infection with C. hepatica is acquired by ingesting soil, food, or water contaminated with embryonated eggs. Larvae released in the cecum penetrate the intestinal mucosa, enter the portal venous circulation, and lodge in the liver. Four weeks after infection, adult worms disintegrate, releasing eggs into the hepatic parenchyma and producing an intense inflammatory reaction with macrophages, eosinophils, and giant cells. Resolution is accompanied by marked peri-egg fibrosis. Clinical Features.  Hepatic capillariasis typically manifests as acute or subacute hepatitis. Findings include fever, nausea, vomiting, diarrhea or constipation, anorexia, myalgias, arthralgias, tender hepatomegaly, and occasionally splenomegaly. Laboratory investigation may reveal leukocytosis with eosinophilia; mild elevations of serum AST, alkaline phosphatase, and bilirubin levels; anemia; and an increased erythrocyte sedimentation rate. A chest radiograph may show pneumonitis.90 Diagnosis.  The diagnosis is established by detection of adult worms or eggs in the liver (Fig. 82-2). Histologic findings in the liver include necrosis, fibrosis, and granulomas.90 A finding of C. hepatica eggs in stools is not indicative of acute infection and probably reflects passage of undercooked liver from an infected animal. Treatment.  Treatment of hepatic capillariasis has, in general, been unsuccessful. Anecdotal benefit has been reported in end-stage cases with therapy with dithiazanine iodide, sodium stibogluconate, albendazole, or thiabendazole.91 Ascariasis Ascaris lumbricoides infects at least 1 billion persons, particularly in areas of lower socioeconomic standing.92 Humans are infected by ingesting embryonated eggs, usually adherent to raw vegetables. The eggs hatch in the small

Figure 82-2.  Histopathology of hepatic capillariasis. Intrahepatic granulomas may be seen surrounding numerous eggs. (Hematoxylin and eosin.) (From Burt AD, Portmann BC, Ferrell LD, editors. MacSween’s Pathology of the Liver, 5th ed. London: Churchill Livingstone; 2007. p 465.)

intestine, and the larvae penetrate the mucosa, enter the portal circulation, and reach the liver, pulmonary artery, and lungs; they grow in the alveolar spaces, are regurgitated and swallowed, and become mature adults in the intestine two to three months after ingestion. Then the cycle repeats itself. Clinical Features.  Symptoms generally occur in persons with a large worm burden; most infected persons are asymptomatic. Cough, fever, dyspnea, wheezing, substernal chest discomfort, and hepatomegaly may occur in the first two weeks. Chronic infection more frequently is characterized by episodic epigastric or periumbilical pain. If the worm burden is particularly heavy, small bowel complications such as obstruction, intussusception, volvulus, perforation, or appendicitis may occur.93 Fragments of disintegrating worms within the biliary tree can serve as nidi for the development of biliary calculi.94 Preexisting disease of the biliary tree or pancreatic duct can predispose the patient to migration of the worm into the bile ducts, with development of obstructive jaundice, cholangitis, or intrahepatic abscesses.92,95 Diagnosis.  A history of regurgitating a worm or passing a large worm (15 to 40 cm long) in the stool suggests ascariasis. In the absence of such a history, the diagnosis is made by identification of characteristic eggs in stool specimens. Larvae also may be identified in sputum and gastric washings and in liver and lung biopsy specimens. In patients with biliary or pancreatic symptoms, ultrasonography, magnetic resonance cholangiopancreatography, or endoscopic retrograde cholangiopancreatography (ERCP) is performed. ERCP also allows extraction of the worm.96 Chest radio­ graphy may show an infiltrate, and eosinophilia may be present. Treatment.  One of the following regimens may be used: (1) a single dose of albendazole, 400 mg; (2) mebendazole, 100 mg twice daily for three days; (3) a single dose of ivermectin, 200 µg/kg; or (4) pyrantel pamoate, 11 mg/kg to a maximum of 1 g.97 Intestinal or biliary obstruction may require endoscopic or surgical intervention. Strongyloidiasis Strongyloides stercoralis is prevalent in the tropics and subtropics, southern and eastern Europe, and the United States. Infection usually is asymptomatic. Humans are infected by the filariform larvae, which penetrate intact skin, are carried to the lungs, migrate through the alveoli, and are swallowed to reach the intestine, where maturation ensues. Autoinfection can occur if the rhabditiform larvae transform into infective filariform larvae in the intestine; reinfection occurs by penetration of the bowel wall or perianal skin. Symptomatic infection results from a heavy infectious burden or infection in an immunocompromised patient. In the latter case, a hyperinfection syndrome may result from dissemination of filariform larvae into tissues that usually are not infected.98 Clinical Features.  Acute infection can lead to a pruritic eruption, followed by fever, cough, wheezing, abdominal pain, diarrhea, and eosinophilia. In immunocompromised patients, the hyperinfection syndrome may be characterized by invasion of any organ, including the liver, lung, and brain. Hyperinfection should be considered particularly in the setting of sepsis caused by multiple organisms found in intestinal flora, a consequence of burrowing of larvae through the intestinal mucosa.99 When the liver is affected, features include jaundice and cholestatic liver biochemical test abnormalities. A liver biopsy specimen may show periportal inflammation, eosinophilic granulomatous hepatitis, or both. Larvae may be observed in intrahepatic bile

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess canaliculi, lymphatic vessels, and small branches of the portal vein.98 Diagnosis.  Serologic tests include counterimmunoelectrophoresis and ELISA and can be used for post-treatment evaluation.100 The diagnosis of active infection is firm when larvae are identified in the stool or intestinal biopsy specimens. An obstructive hepatobiliary picture in a person with known strongyloidiasis should alert the clinician to the possibility of dissemination. Treatment.  For treatment of acute infection, the drug of choice is ivermectin, 200 mg/kg/day for two days. Clearance rates are high. An alternative agent is albendazole, 400 mg/day for three days for adults and children older than two years of age, but retreatment may be necessary and this drug is less effective for disseminated disease. The hyperinfection syndrome requires longer courses of treatment than those used for the primary acute infection.101 Trichinosis Humans may be infected with Trichinella spiralis by eating raw or undercooked pork-bearing larvae, which are released in the small intestine, penetrate the mucosa, and disseminate through the systemic circulation. Larvae can be found in the myocardium, cerebrospinal fluid, brain, and, less commonly, liver and gallbladder. The larvae then re-enter the circulation and reach striated muscle, where they become encapsulated. Clinical Features.  Clinical manifestations occur when the worm burden is high and include diarrhea, fever, myalgias, periorbital edema, and leukocytosis with marked eosinophilia. Rarely, larvae can be seen invading hepatic sinusoids on examination of a liver biopsy specimen. Jaundice may result from biliary obstruction. Diagnosis.  The diagnosis is suggested by a characteristic history in a patient with fever and eosinophilia. Serologic

Miracidia penetrate snail

Cercariae leave snail

assays for antibody to Trichinella may not be helpful in the acute phase of infection but can be useful after two weeks.102 Muscle biopsy may help to confirm the diagnosis. DNAbased tests are investigational. Treatment.  Treatment consists of glucocorticoids to relieve allergic symptoms, followed by antihelminthic treatment with albendazole, 400 mg twice daily for 10 to 15 days, or mebendazole, 200 mg/day for 10 to 15 days.102

Trematodes (Flukes)

Schistosomiasis (Bilharziasis) About 200 million persons worldwide are infected with trematodes of the genus Schistosoma. Schistosoma mansoni is found in the western hemisphere, Africa, and the Middle East; S. haematobium is found in Africa and the Middle East; S. japonicum and S. mekongi are found in the Far East; and S. intercalatum is found in parts of central Africa. The last two species are much less common than the other three and cause liver disease and colonic disease, respectively.103 The infectious cycle is initiated by penetration of the skin by free cercariae in fresh water (Fig. 82-3). The cercariae reach the pulmonary vessels within 24 hours, pass through the lungs, and reach the liver, where they lodge, develop into adults, and mate. Adult worms then migrate to their ultimate destinations in the inferior mesenteric venules (S. mansoni), superior mesenteric venules (S. japonicum), or veins around the bladder (S. haematobium). These locations correlate with the clinical complications associated with each species. Each female fluke can lay 300 to 3000 eggs daily. The eggs are deposited in the terminal venules and eventually migrate into the lumen of the involved organ, after which they are expelled in the stool or urine. Eggs remaining in the organ provoke a robust granulomatous response. Excreted eggs hatch immediately in fresh water

Cercariae enter skin

Schistosomulae migrate to liver Eggs deposited into water

Fertilization takes place

Eggs pass into feces

Adult worms migrate to the mesenteric vessels

Females deposit fertilized eggs and these pass into intestine Figure 82-3.  The life cycle of Schistosoma species. (From Gitlin N, Strauss R. Atlas of Clinical Hepatology. Philadelphia: WB Saunders; 1995. p 72.)

1361

1362

Section IX  Liver

A

B

Figure 82-4.  Pipestem fibrosis. A, Liver resection specimen demonstrates characteristic pipestem fibrosis due to long-term infection with Schistosoma mansoni. (Courtesy of Dr. Fiona Graeme-Cook, Boston, Mass.) B, Ultrasound image of the liver from a patient with schistosomiasis demonstrates the pipestem fibrosis, seen as echodense circles surrounding vessels (arrow). (Courtesy of Dr. Mark Feldman, Dallas, Tex.)

and liberate early intermediate miracidia, which infect their snail hosts. The miracidia transform into cercariae within the snails and then are released into the water, from which they may again infect humans.103 Clinical Features.  Acute toxemic schistosomiasis (Katayama syndrome or Katayama fever), presumably a consequence of the host immunologic response to mature worms and eggs, occurs approximately four to six weeks after exposure. Manifestations include headache, fever, chills, cough, diarrhea, myalgias, arthralgias, tender hepatomegaly, and eosinophilia. Untreated acute schistosomiasis invariably progresses to chronic disease. Mesenteric infection leads to hepatic complications, including periportal fibrosis, presinusoidal occlusion, and, ultimately, portal hypertension, as a result of the inflammatory reaction to eggs deposited in the liver. The development of periportal fibrosis appears to be related to production of TNF-a.104 The lungs and central nervous system may be affected when eggs or adult worms pass through the liver into the systemic circulation, especially in S. japonicum infection; pulmonary hypertension and cor pulmonale may result.105 With severe schistosomal infection, portal hypertension becomes progressive, leading to gastroesophageal varices, splenomegaly, and rarely ascites. Chronic schistosomal infection may be complicated by increased susceptibility to Salmonella infections.106 Hepatitis B or hepatitis C viral coinfection also is common in persons living in endemic areas and may accelerate the progression of liver disease and development of hepatocellular carcinoma.107 In African intestinal schistosomiasis, pseudopolyps of the colon may develop, leading in some cases to protein-losing colopathy and formation of an inflammatory mass in the descending colon. Laboratory findings in chronic schistosomiasis include anemia from recurrent luminal gastrointestinal bleeding or hypersplenism, leukocytosis with eosinophilia, an elevated erythrocyte sedimentation rate, and increased serum IgE levels. Results of liver biochemical tests generally are normal until the disease is at an advanced stage. Diagnosis.  The possibility of acute schistosomiasis should be considered in a patient with a history of exposure, abdominal pain, diarrhea, and fever. Multiple stool examinations for ova may be required to confirm the diagnosis because results frequently are negative in the early phase of disease. Serologic testing using counterimmunoelectrophoresis or ELISA cannot distinguish between past infection

and active disease but may be useful in a returned traveler. Sigmoidoscopy or colonoscopy may reveal rectosigmoid or transverse colon involvement and may be useful in chronic disease, when few eggs pass in the feces. Ultrasonography and liver biopsy are useful for demonstrating periportal (“pipestem” or “clay pipestem”) fibrosis (Fig. 82-4) but not for diagnosing acute infection because of their insensitivity for detecting schistosomal eggs.108 CT may show lowattenuation rings around main portal vein branches with marked enhancement.109 Treatment.  Praziquantel, 40 mg/kg for S. mansoni or S. haematobium and 60 mg/kg for S. japonicum or S. mekongi given in one day in two to three divided doses four hours apart, is the therapeutic agent of choice. Oxamniquine is an effective alternative agent in patients who cannot tolerate praziquantel, but it is no longer available in the United States. Treatment of acute toxemic schistosomiasis often requires prednisone to suppress immune-mediated helminthicidal or drug reactions, in conjunction with praziquantel at the dose appropriate for the particular species for three to six days.103 Retreatment after two to three months is often necessary after Katayama fever.110 Band ligation and injection sclerotherapy of varices are effective in controlling variceal bleeding (see Chapter 90). Management of advanced chronic schistosomal liver disease may require placement of a distal splenorenal shunt or esophagogastric devascularization with splenectomy. Fortunately, since the advent of praziquantel, complicated schistosomal liver disease has become uncommon. Fascioliasis Fascioliasis is endemic in parts of Europe and Latin America, North Africa, Asia, the Western Pacific, and some parts of the United States. Fascioliasis is caused by the sheep liver fluke Fasciola hepatica. Eggs passed in the feces of infected mammals into fresh water give rise to miracidia that penetrate snails and eventually emerge as mobile cercariae, which attach to aquatic plants such as watercress. Hosts become infected when they consume plants containing encysted metacercariae, which then bore into the intestinal wall, enter the abdominal cavity, penetrate the hepatic capsule, and eventually settle in the bile ducts, where they attain maturity. Mature flukes release eggs that are passed in the host’s feces to complete the life cycle.111 Clinical Features.  Three syndromes are recognized: acute or invasive, chronic latent, and chronic obstructive.112 The

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess acute phase corresponds to the migration of young flukes through the liver and is marked by fever, pain in the right upper quadrant, and eosinophilia. Urticaria with dermatographia and nonspecific gastrointestinal symptoms are common. Physical examination often reveals fever and a tender, enlarged liver. Splenomegaly is seen in as many as 25% of cases, but jaundice is rare and liver biochemical test abnormalities are mild.113 Eosinophilia can be profound, with eosinophils sometimes exceeding 80% of the differential leukocyte count.111 The latent phase corresponds with the settling of the flukes into the bile ducts and can last for months to years. Affected patients may experience vague gastrointestinal symptoms. Eosinophilia persists, and fever can occur.113 The chronic obstructive phase is a consequence of intrahepatic and extrahepatic bile ductal inflammation and hyperplasia evoked by adult flukes. Recurrent biliary pain, cholangitis, cholelithiasis, and biliary obstruction may result. Blood loss from epithelial injury occurs, but overt hemobilia is rare. Liver biochemical testing commonly demonstrates a pattern suggestive of biliary obstruction.114 Long-term infection may lead to biliary cirrhosis and secondary sclerosing cholangitis, but no convincing association with biliary tract or hepatic malignancy has been demonstrated.115 Diagnosis.  The diagnosis should be considered in a patient with prolonged fever, abdominal pain, diarrhea, tender hepatomegaly, and eosinophilia. Because eggs are not passed during the acute phase, diagnosis depends on the detection of antibody by counterimmunoelectrophoresis or ELISA. In the latent and chronic phases, a definitive diagnosis is based on the detection of eggs in stool, duodenal aspirate specimens, or bile.116 On occasion, ultrasonography or ERCP will demonstrate flukes in the gallbladder and bile duct.117 If one member of a family is diagnosed with fascioliasis, all household members should be evaluated. Hepatic histologic findings include necrosis and granulomas with eosinophilic infiltrates and Charcot-Leyden crystals. Eosinophilic abscesses, epithelial hyperplasia of the bile ducts, and periportal fibrosis may be seen.118 Treatment.  The drug of choice is triclabendazole, 10 mg/kg given once orally. Praziquantel, mebendazole, and albendazole are not effective for fascioliasis. Other medications that are potentially efficacious are bithionol and nitazoxanide.119

Clonorchiasis and Opisthorchiasis

Clonorchis sinensis, Opisthorchis viverrini, and Opisthorchis felineus are trematodes of the family Opisthorchiidae. Infection by C. sinensis and O. viverrini is widespread in East and Southeast Asia and is linked to lower socioeconomic status. O. felineus infects humans and domestic animals in eastern Europe. All three have similar lifecycles and result in similar clinical manifestations. Eggs are passed in the feces into fresh water, consumed by snails, and hatch as free-swimming cercariae, which seek and penetrate fish or crayfish and encyst in skin or muscle as metacercariae. The mammalian host is infected when it consumes raw or undercooked fish. The metacercariae excyst in the small intestine and migrate into the ampulla of Vater and bile ducts, where they mature into adult flukes. Infection can be maintained for 2 decades or longer.116 Clinical Features.  In general, acute infection is clinically silent. Occasional symptoms include fever, abdominal pain, and diarrhea. Chronic manifestations correlate with the fluke burden and are dominated by hepatobiliary features: fever, pain in the right upper quadrant, tender hepatomegaly, and eosinophilia. If the worm burden in the bile ducts

Figure 82-5.  Clonorchis sinensis. (Courtesy of Dr. Fiona Graeme-Cook, Boston, Mass.)

is heavy, chronic or intermittent biliary obstruction can ensue, with frequent cholelithiasis, cholecystitis, jaundice, and, ultimately, recurrent pyogenic cholangitis (see Chapter 68). Liver biochemical test results, especially serum alkaline phosphatase and bilirubin levels, are elevated. Longstanding infection leads to exuberant inflammation, resulting in periportal fibrosis, marked biliary epithelial hyperplasia and dysplasia, and, ultimately, a substantially increased risk of cholangiocarcinoma.116,120 Cholangiocarcinoma resulting from clonorchiasis or opisthorchiasis tends to be multicentric and arises in the secondary biliary radicles of the hilum of the liver. Cholangiocarcinoma should be suspected in infected persons with weight loss, jaundice, epigastric pain, or an abdominal mass (see Chapter 69). Diagnosis.  The diagnosis of clonorchiasis or opisthorchiasis is made by detection of characteristic fluke eggs in the stool, except late in the disease when biliary obstruction has supervened. In these cases, the diagnosis is made by identifying flukes in the bile ducts or gallbladder at surgery or in bile obtained by postoperative drainage or percutaneous aspiration (Fig. 82-5). Endoscopic or intraoperative cholangiography reveals slender, uniform filling defects within intrahepatic ducts that are alternately dilated and strictured, mimicking sclerosing cholangitis. Serologic methods of diagnosis cannot distinguish between past or current infection.49,121 Treatment.  All patients with clonorchiasis or opisthorchiasis should receive praziquantel, which is uniformly effective in a dose of 75 mg/kg in three divided doses over one day. Side effects are uncommon and include headache, dizziness, and nausea. After treatment, dead flukes may be seen in the stool or biliary drainage. When the burden of infecting organisms is high, the dead flukes and surrounding debris or stones may cause biliary obstruction, necessitating endoscopic or surgical drainage.115

Cestodes (Tapeworms)

Echinococcosis Infections with Echinococcus granulosus can be found worldwide in areas where dogs are used to help raise livestock. Echinococcus multilocularis is distributed in northern North America and Eurasia, whereas Echinococcus vogeli is found in scattered areas of Central and Latin America. Infection occurs when humans eat vegetables contaminated by dog feces containing embryonated eggs. The eggs hatch in the small intestine and liberate oncospheres that penetrate the mucosa and migrate via vessels or lym-

1363

1364

Section IX  Liver

Figure 82-6.  Liver resection specimen of a hydatid cyst caused by Echinococcus granulosus. Multiple daughter cysts are seen. (Courtesy of Dr. Fiona Graeme-Cook, Boston, Mass.) Figure 82-7.  Computed tomographic scan showing the typical appearance of a hydatid cyst in the liver. (Courtesy of Dr. Mukesh Harisinghani, Boston, Mass.)

phatics to distant sites. The liver is the most common destination (70%), followed by the lungs (20%), kidney, spleen, brain, and bone. In these organs, a hydatid cyst develops by vesiculation and produces thousands of protoscolices. The cyst wall contains three layers: an outer adventitial layer, which is host-derived and can calcify, and an intermediate acellular layer and inner germinal layer, which are wormderived. A protoscolex is produced asexually within small secondary cysts that develop from the inner layer. Rupture of the hydatid cyst releases the viable protoscolices, which set up daughter cysts in secondary sites. The adult Echinococcus tapeworm consists of a scolex, which contains a rostellum with 20 to 50 hooklets and 4 suckers, a neck, and an immature, mature, and gravid proglottid. Dogs acquire the infection by consuming organs of sheep, cattle, or other livestock bearing the hydatid cyst. Clinical Features.  Most patients with a hydatid cyst in the liver have no symptoms. As the cysts of E. granulosus grow within the liver (Fig. 82-6), they begin to cause lowgrade fever, pain, tender hepatomegaly (usually affecting the right hepatic lobe), and eosinophilia. If the cysts grow large enough, they may rupture spontaneously or after trauma into the lungs, leading to dyspnea and hemoptysis. More extensive rupture into the peritoneum or lungs may lead to a life-threatening anaphylactic reaction to the cyst contents. Rupture into the biliary tract can cause cholangitis and obstruction; marked eosinophilia may be present. Superinfection of the hepatic cysts can lead to pyogenic liver abscesses in up to 20% of patients with hepatic disease. Rare complications of hydatid cysts or cyst rupture include pancreatitis, portal hypertension, Budd-Chiari syndrome, and rupture into the pericardial sac. E. multilocularis is highly invasive; infection leads to formation of solid masses in the liver that are easily confused with cirrhosis or carcinoma. Alveolar hydatid disease is the term applied to hepatic nodules that appear on microscopy as alveoli-like microvesicles.122 Daughter cysts bud from the germinal membrane in an uncontrolled manner, with “invasion” of the surrounding liver parenchyma by the scolices. Infection of bile ducts and vessels and necrosis of parenchyma may result in cholangitis, liver abscess, sepsis, portal hypertension, hepatic vein occlusion, and biliary cirrhosis. Unfortunately, infection generally is

not diagnosed until the lesions are inoperable because of extensive invasion or distant metastatic disease, and mortality rates are high, approaching 90%.122 Infection with E. vogeli has clinical features intermediate between those of infections caused by the other two species and is characterized by multiple fluid-filled cysts containing daughter cysts and protoscolices. Although not as aggressive as E. multilocularis infection, E. vogeli infection can spread to contiguous sites. Diagnosis.  A history of exposure in a patient with hepatomegaly and an abdominal mass is highly suggestive of hepatic echinococcosis, but the most important diagnostic tools are radiology and serology. Ring-like calcifications in up to one fourth of hepatic cysts are visible on plain abdominal radiographs in patients infected with E. granulosus. The sensitivity and specificity of both ultrasonography and CT in confirming the diagnosis are high (Fig. 82-7).123 Both modalities can demonstrate intracystic septations and daughter cyst formation in about one half of the cysts. Contrast-enhanced CT may display avascular cysts with ring enhancement. Percutaneous aspiration of the cyst had traditionally been discouraged because of concern about anaphylactic reactions. Encouraging reports, however, suggest that under carefully controlled conditions, with use of thin needles and concomitant antihelminthic therapy, percutaneous aspiration for diagnosis and therapy may be safe.124,125 The detection of protoscolices or acid-fast hooklets in the cyst fluid confirms the diagnosis.126 An ELISA is the best serologic assay for diagnosis, with a sensitivity rate of 84% to 90%.127 Assays for detecting circulating antigen are likely to provide additional diagnostic benefit in the future. The Casoni skin test, used in the past, is nonspecific and no longer recommended. E. multilocularis infection can be diagnosed with a combination of ELISA and CT, which often shows scattered areas of calcified necrotic tissue. In E. vogeli infection, CT demonstrates polycystic lesions in the liver or peritoneal space. Treatment.  In the past, accessible cysts in younger persons were always treated surgically, and surgery is still

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess considered the preferred treatment in many cases. The goal has been removal of the cestode without disruption of cyst contents. Care must be taken to isolate the cyst and to inject cidal agents before the cyst is aspirated. Successful approaches have included cystectomy, endocystectomy, omentoplasty, and marsupialization. A laparoscopic app­ roach is feasible in some cases. In complicated cases, hepatic lobectomy or hemihepatectomy may be necessary. Calcified cysts need not be removed. Promising data indicate that careful percutaneous drainage is a safe and effective alternative to surgery for the treatment of complicated cysts.128 In addition to surgery or drainage, administration of an antihelminthic, such as albendazole, 10 mg/kg/day for eight weeks, is recommended.129 Puncture, aspiration, injection (of a scolicidal agent), and re-aspiration (PAIR) can be performed safely with long-term control of echinococcal cysts.125 Injection of hydatid liver cysts with albendazole has also been described.124 Therefore, nonsurgical approaches are now available for management of hydatid cysts. The decision between surgical and nonsurgical techniques depends on the extent and type of lesions.130 Cysts that cannot be treated surgically or percutaneously should be treated with albendazole, preferably, or mebendazole. Large doses and prolonged treatment are required (e.g., albendazole 10 mg/kg daily in two divided doses for 28 days, repeated three or four times, with 2-week breaks between courses). Surgical resection is curative in up to one third of cases of E. multilocularis infection. In most cases the disease is advanced when the diagnosis is made. In such cases, palliative drainage procedures or long-term treatment with albendazole or other benzimidazole carbamates may prolong survival.122,131 Surgery appears to be the most effective approach for management of E. vogeli infection.

FUNGI CANDIDIASIS

Candida species may cause invasive systemic infection with hepatic involvement in severely immunocompromised persons (see Chapters 33 and 34). The liver can become infected by C. albicans and related species in the setting of disseminated, multiorgan disease. Most disseminated infections occur in leukemic patients undergoing high-dose chemotherapy and become clinically evident during the period of recovery from severe neutropenia. In several series, hepatic candidiasis was present in 51% to 91% of predominantly leukemic patients with disseminated candidiasis.132,133 Disease often is overwhelming, with high mortality rates.133 Other, less frequent presentations in the compromised host include isolated or focal hepatic or hepatosplenic candidiasis.134 Focal candidiasis is believed to result from colonization of the gastrointestinal tract by Candida, which disseminates locally following the onset of neutropenia and mucosal injury caused by high-dose chemotherapy.134 Resulting fungemia of the portal vein seeds the liver and leads to formation of hepatic microabscesses and macroabscesses. In either focal or disseminated candidiasis involving the liver, clinical features include fever, abdominal pain and distention, nausea, vomiting, diarrhea, and tender hepatomegaly. The serum alkaline phosphatase level is almost invariably elevated, with varying elevations in serum aminotransferase and bilirubin levels. CT of the abdomen is the most sensitive test to detect hepatic or splenic abscesses,

Figure 82-8.  T2-weighted magnetic resonance image showing charac­ teristic small high-intensity foci (arrows) of hepatosplenic candidiasis. (Courtesy of Dr. Mukesh Harisinghani, Boston, Mass.)

which often are multicentric (Fig. 82-8).135 In cases diagnosed antemortem, liver biopsy or laparoscopy reveals macroscopic nodules, necrosis with microabscesses, and characteristic yeast or hyphal forms of Candida.136,137 The results of cultures of biopsy material are negative in most cases. PCR methodology has been used to diagnose hepatic candidiasis.138 Response rates to therapy with intravenous amphotericin B are better (almost 60%) for focal hepatic candidiasis than for disseminated disease. The success of treatment is currently far from optimal, however. Alternatives to amphotericin B are fluconazole; liposomal amphotericin; and intravenous echinocandins such as caspofungin, micafungin, or anidulafungin.139 The widespread use of prophylactic fluconazole in high-risk patients has resulted in lower rates of fatal visceral fungal infection while promoting a shift toward infections caused by other molds resistant to this agent.140

HISTOPLASMOSIS

Infection with Histoplasma capsulatum is acquired through the respiratory tract and in most cases is confined to the lungs. Severely immunocompromised persons (e.g., those with AIDS), however, are predisposed to disseminated histoplasmosis (see Chapter 33). The liver can be invaded in both acute and chronic progressive disseminated histoplasmosis. Fever, oropharyngeal ulcers, hepatomegaly, and splenomegaly may be present in patients with chronic disease.141 In children with acute hepatic disease, which appears to be an extension of primary pulmonary infection, marked hepatosplenomegaly is universal and is associated with high fever and lymphadenopathy. In one series of 111 cases of disseminated histoplasmosis, serum ALT levels were elevated in 39%, AST levels were elevated in 27%, and alkaline phosphatase levels were >200 U/L in 55%.142 Hepatosplenomegaly is present in approximately 30% of adults with acute disease (often the AIDS-defining illness). Yeast forms can be identified in liver biopsy specimens with standard hematoxylin and eosin staining. The silver methenamine method is superior for detecting yeast forms in areas of caseating necrosis or in granulomas. The organism is difficult to culture and almost never grows from biopsy specimens. Serologic testing for complement-fixing

1365

1366

Section IX  Liver antibodies is therefore helpful in confirming the diagnosis. In immunocompromised persons who may not be capable of mounting an antibody response, detection of H. capsulatum antigens in urine and serum can be useful.143 Treatment options include therapy with amphotericin B, fluconazole, or itraconazole.

LIVER ABSCESS PYOGENIC

In the past, most cases of pyogenic liver abscess were a consequence of appendicitis complicated by pylephlebitis (portal vein inflammation) in a young patient. This presentation is less common today as a result of earlier diagnosis and effective antibiotic therapy. Most cases now are cryptogenic or occur in older men with underlying biliary tract disease.144 Predisposing conditions include malignancy, immunosuppression, diabetes mellitus, and previous biliary surgery or interventional endoscopy.

Pathogenesis

Infections of the biliary tract (e.g., cholangitis, cholecystitis) are the most common identifiable source of liver abscess. Infection may spread to the liver from the bile duct, along a penetrating vessel, or from an adjacent septic focus (including pylephlebitis). Pyogenic liver abscess may arise as a late complication of endoscopic sphincterotomy for bile duct stones or within three to six weeks of a surgical biliaryintestinal anastomosis.144 Pyogenic liver abscesses may complicate recurrent pyogenic cholangitis, which is found predominantly in East and Southeast Asia and is characterized by recurring episodes of cholangitis, intrahepatic stone formation, and, in many cases, biliary parasitic infections (see Chapter 68). Less commonly, liver abscess is a complication of bacteremia arising from underlying abdominal disease, such as diverticulitis, appendicitis, perforated or penetrating peptic ulcer, gastrointestinal malignancy, inflammatory bowel disease, or peritonitis, or rarely from bacterial endocarditis or penetration of a foreign body through the wall of the colon. The risk of liver abscess may be increased in patients with underlying diabetes mellitus or cirrhosis.145,146 Occasionally, a pyogenic liver abscess may be the presentation of a hepatocellular or gallbladder carcinoma or a complication of chemoembolization or percutaneous ablation of a hepatic neoplasm.147

In approximately 40% of cases of pyogenic liver abscess, no obvious source of infection can be identified. Oral flora have been proposed to be a potential source in such cases, particularly in patients (often alcoholics) with severe periodontal disease.

Microbiology

Most pyogenic liver abscesses are polymicrobial. The bacterial organisms that have been cultured from liver abscesses are listed in Table 82-3. The most frequently isolated organisms are Escherichia coli and Klebsiella, Proteus, Pseudomonas, and Streptococcus species, particularly the Streptococcus milleri group. Certain virulent strains of Klebsiella pneumoniae can cause liver abscess in the absence of underlying hepatobiliary disease, often with metastatic infection.148 With improved cultivation methods and earlier diagnosis, the number of cases caused by anaerobic organisms has increased. The most commonly identified anaerobic species are Bacteroides fragilis and Fusobacterium necrophorum; anaerobic streptococci also have been identified. Pyogenic abscess associated with recurrent pyogenic cholangitis may be caused by Salmonella typhi. Clostridium and Actinomyces species are uncommon causes of liver abscess, and rare cases are caused by Yersinia enteroco­ litica, Pasteurella multocida, Haemophilus parainfluenzae, and Listeria species. Septic melioidosis also has been described. Liver abscesses caused by Staphylococcus aureus infection are most common in children and patients with septicemia or other conditions associated with impaired host resistance, including chronic granulomatous disease.149 Fungal abscesses of the liver may occur in immunocompromised hosts, particularly those with a hematologic malignancy (see earlier).

Clinical Features and Diagnosis

In the preantibiotic era, patients with a pyogenic liver abscess typically presented with acutely spiking fevers, pain in the right upper quadrant, and, in many cases, shock. After the introduction of antibiotics, the presentation of pyogenic liver abscess became less acute. Today’s presentation often is insidious, particularly in elderly patients, and is characterized by malaise, low-grade fever, anorexia, weight loss, and dull abdominal pain that may increase with movement. Symptoms may be present for one month or more before a diagnosis is made. Multiple abscesses are typical when biliary disease is the source and are associated with a more acute systemic presentation, often with sepsis

Table 82-3  Organisms That May Be Isolated from the Abscess and the Blood in Patients with a Pyogenic Liver Abscess GRAM-NEGATIVE AEROBES

GRAM-POSITIVE AEROBES

ANAEROBES

OTHERS

Escherichia coli Klebsiella pneumoniae Enterobacter spp. Pseudomonas spp. Citrobacter spp. Morganella spp. Proteus spp. Salmonella spp. Serratia marcescens* Yersinia spp.* Burkholderia pseudomallei* Capnocytophaga canimorsus* Pasteurella multocida* Achromobacter xylosoxidans*

Enterococcus spp. Streptococcus pyogenes Staphylococcus aureus Streptococcus milleri group Listeria monocytogenes* Bacillus cereus*

Bacteroides spp. Fusobacterium spp. Streptococcus spp. Peptostreptococcus spp. Peptococcus spp. Prevotella spp.* Clostridium spp.* Actinomyces spp.*

Candida spp. Mycobacterium tuberculosis

*Rare cause.

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess and shock, than is the case with solitary abscesses. When an abscess is situated near the dome of the liver, pain may be referred to the right shoulder, or a cough resulting from diaphragmatic irritation or atelectasis may be present. Physical examination usually discloses fever, hepatomegaly, and liver tenderness, which is accentuated by movement or percussion. Splenomegaly is unusual, except with a chronic abscess. Ascites is rare, and in the absence of cholangitis, jaundice is present only late in the course of the illness. Portal hypertension may follow recovery if the portal vein has been thrombosed. Laboratory findings include anemia, leukocytosis, an elevated erythrocyte sedimentation rate, and abnormal liver biochemical test results, especially an elevated serum alkaline phosphatase level. Blood culture specimens will identify the causative organism in at least 50% of cases.150 Direct cultures of aspirated fluid are useful for identification of the organism and determination of antibiotic susceptibility and should be sent for both aerobic and anaerobic culture.151 Chest x-rays may show elevation of the right hemidiaphragm and atelectasis. Ultrasonography and CT are the initial imaging modalities of choice. Abscesses as small as 1 cm in diameter can be detected. Ultrasonography is inexpensive and accurate and can guide needle aspiration of the abscess. Culture specimens of aspirated material yield positive results in 90% of cases (although the yield probably is lower if the patient has been receiving antibiotics). CT also is accurate, with a sensitivity rate approaching 100%, but is more expensive than ultrasonography. Hepatic abscesses are usually hypodense on a CT scan and may display a rim of contrast enhancement in less than 20% of cases (Fig. 82-9). CT permits precise localization of an abscess, assessment of its relationship to adjacent structures, and detection of gas in the abscess, which is associated with increased mortality. An abscess must be distinguished from other mass lesions in the liver, including cystic lesions, benign and malignant neoplasms, soft tissue tumors (neurofibroma, leiomyoma, and malignant fibrous histiocytoma), focal nodular hyperplasia, and hemangiomas (see Chapter 94), as well as inflam-

Figure 82-9.  Computed tomographic scan showing multiple pyogenic abscesses in the liver. (Courtesy of Dr. Mukesh Harisinghani, Boston, Mass.)

matory pseudotumors. Magnetic resonance imaging is more sensitive than CT for detecting small abscesses, which have low signal intensity on T1-weighted images and high signal intensity on T2-weighted images and enhance with gadolinium. ERCP is indicated in patients with imaging evidence of biliary stones or prominent cholestasis.152 Rarely, arteriography may be of value in distinguishing an abscess from a tumor. Inflammatory pseudotumor of the liver (also called plasma cell granuloma) is a rare, benign lesion characterized by proliferating fibrous tissue infiltrated by inflammatory cells. The cause is unknown. Affected persons (typically young men) often have a history of recent infection, but a causative infectious agent is rarely isolated from the lesion. Additional associated disorders include chronic inflammatory and autoimmune disorders, particularly ascending cholangitis and primary sclerosing cholangitis, as well as diabetes mellitus, Sjögren’s syndrome, gout, ulcerative colitis, Crohn’s disease, HIV infection, Epstein-Barr virus infection, and acute myeloblastic leukemia. Patients typically present with intermittent fever, abdominal discomfort, vomiting, diarrhea, weight loss, and malaise and have hepatomegaly, right upper quadrant tenderness, and jaundice on physical examination. Portal hypertension may develop. Laboratory findings also are similar to those associated with liver abscess, including polyclonal hyperglobulinemia in 50% of cases, and imaging studies generally are interpreted as showing a tumor or an abscess. Treatment generally has been by surgical resection of the lesion, although some patients may recover spontaneously or after treatment with antibiotics or glucocorticoids, once the diagnosis is made on the basis of needle biopsy findings.153,154

Prevention and Treatment

Pyogenic liver abscesses are best prevented by prompt treatment of acute biliary and abdominal infections and by adequate drainage of infected intra-abdominal collections under appropriate antibiotic coverage. Treatment of a hepatic abscess requires antibiotic therapy directed at the causative organism(s) and, in most cases, drainage of the abscess, usually percutaneously with radiologic guidance. An indwelling drainage catheter may be placed in the abscess until the cavity has resolved, particularly for lesions greater than 5 cm in size, although intermittent needle aspiration may be as effective as continuous catheter drainage for smaller lesions.155,156 With multiple abscesses, only the largest abscess may need to be aspirated; smaller lesions often resolve with antibiotic treatment alone, but rarely, each lesion may need drainage. For a small abscess, anti­ biotic therapy without drainage may suffice. Biliary decompression is essential when a hepatic abscess is associated with biliary tract obstruction or communication and may be accomplished through the endoscopic or transhepatic route (see Chapter 70). Surgical drainage of a hepatic abscess may be necessary in patients with incomplete percutaneous drainage, unresolved jaundice, renal impairment, a multiloculated abscess, or a ruptured abscess.157 A laparoscopic approach may be feasible in select cases. Initial antibiotic coverage, pending culture results, should be broad in spectrum and include ampicillin and an aminoglycoside (when a biliary source is suspected) or a third-generation cephalosporin (when a colonic source is suspected), plus, in either case, metronidazole, to cover anaerobic organisms. If amebiasis is suspected, metronidazole should be started before aspiration is performed. Alternative regimens include combinations of a beta-lactam and beta-lactamase inhibitor active against enteric organisms, including anaerobes. After culture results and sensi-

1367

1368

Section IX  Liver tivity profiles have been obtained, antibiotic therapy directed at the specific organism(s) should be administered intravenously for at least two weeks and then orally for up to six weeks.158 For streptococcal infections, the use of high-dose oral antibiotics for six months may be preferable. The mortality rate for patients with hepatic abscesses treated with antibiotics and percutaneous drainage has improved since the 1980s.157,159 A worse prognosis is associated with a delay in diagnosis, multiple abscesses, multiple organisms cultured from blood, a fungal cause, shock, jaundice, hypoalbuminemia, a pleural effusion, an underlying biliary malignancy, multiorgan dysfunction, sepsis, or other associated medical diseases.157,160-164 Complications of pyogenic liver abscess include empyema, pleural or pericardial effusion, portal or splenic vein thrombosis, rupture into the pericardium, thoracic and abdominal fistula formation, and sepsis. Metastatic septic endophthalmitis occurs in as many as 10% of diabetic patients with a liver abscess caused by Klebsiella pneumoniae.148

AMEBIC

Amebiasis occurs in 10% of the world’s population and is most common in tropical and subtropical regions (see also Chapter 109).165,166 In the United States, it is a disease of young, often Hispanic adults. Endemic areas include Africa, Southeast Asia, Mexico, Venezuela, and Colombia. Amebic liver abscess is the most common extraintestinal manifestation of amebiasis. Compared with affected persons who reside in an endemic area, persons in whom an amebic liver abscess develops after travel to an endemic area are older and more likely to be male, have marked hepatomegaly, and have a large abscess or multiple abscesses. The occurrence of an amebic liver abscess in a person who has not traveled to or resided in an endemic area should raise the suspicion of underlying immuno­ suppression, particularly AIDS.167,168 Other persons at increased risk include inpatients in residential institutions

and men who have sex with men. Host factors that contribute to the severity of disease include younger age, pregnancy, malnutrition, alcoholism, glucocorticoid use, and malignancy.

Pathogenesis

During its life cycle, Entamoeba histolytica exists as trophozoite or cyst forms (Fig. 82-10). After infection, amebic cysts pass through the gastrointestinal tract and become trophozoites in the colon, where they invade the mucosa and produce typical “flask-shaped” ulcers. The organism is carried by the portal vein circulation to the liver, where an abscess may develop. Occasionally, organisms travel beyond the liver and can establish abscesses in the lung or brain. Rupture of an amebic liver abscess into the pleural, pericardial, and peritoneal spaces can also occur.

Clinical Features

Amebic liver abscess is 10 times as common in men as in women and is rare in children.165 An amebic liver abscess is more likely than a pyogenic liver abscess to be associated with an acute presentation. Symptoms are present on average for two weeks by the time a diagnosis is made. A latency period between intestinal and subsequent liver infection of up to many years is possible, and less than 10% of patients report an antecedent history of bloody diarrhea with amebic dysentery. Abdominal pain is typically well localized to the right upper quadrant. Fever is nearly universal but may be intermittent. Malaise, myalgias, and arthralgias are common. Jaundice is uncommon and signifies a poor prognosis. Pulmonary symptoms and signs may be present, but a pericardial rub and peritonitis are rare. Occasionally a friction rub is heard over the liver. Laboratory features resemble those found in pyogenic abscess. Coinfection with bacterial pathogens is uncommon. Rare complications of amebic abscesses can include intraperitoneal, intrathoracic, and pericardial rupture and multiorgan failure.

In the small intestine, the cyst wall disintegrates and trophozoites are released

Cysts are consumed

Trophozoites invade the colon

Cysts

Trophozoite Hepatic abscess Ameboma Rupture of abscess may cause: empyema bronchohepatic fistula pericarditis peritonitis Figure 82-10.  The life cycle of Entamoeba histolytica in amebiasis. (From Gitlin N, Strauss R. Atlas of Clinical Hepatology. Philadelphia: WB Saunders; 1995. p 64.)

Trophozoites are carried to Trophozoites the portal circulation invade the mesenteric vessels Trophozoites spread hematogenously to: lungs spleen brain kidneys

Cyst

Chapter 82  Bacterial, Parasitic, and Fungal Infections of the Liver, Including Liver Abscess Table 82-4  Pyogenic and Amebic Liver Abscess: Clinical Comparisons PARAMETER

PYOGENIC LIVER ABSCESS

AMEBIC LIVER ABSCESS

Number Location Presentation Jaundice Diagnosis Treatment

Often multiple Either lobe of liver Subacute Mild US or CT ± aspiration Drainage (if technically feasible) + antibiotics IV (see text)

Usually single Usually right hepatic lobe, near the diaphragm Acute Moderate US or CT and serology Metronidazole, 750 mg tid for 7-10 d orally or IV, or tinidazole, 2 g orally for 3 d, followed by iodoquinol, 650 mg orally tid for 20 d; diloxanide furoate, 500 mg orally tid for 10 d; or aminosidine (paromomycin) 25-35 mg/kg/d orally in three divided doses for 7-10 d

CT, computed tomography; IV, intravenously; US, ultrasonography.

Treatment

Standard therapy consists of metronidazole, 750 mg three times daily by mouth or, if necessary, intravenously for 7 to 10 days. Tinidazole or chloroquine may be substituted for metronidazole. The response to treatment usually occurs within 96 hours. Following a course of metronidazole, most authorities recommend the addition of an oral luminal amebicide—such as iodoquinol, 650 mg three times daily for 20 days; diloxanide furoate, 500 mg three times daily for 10 days; or aminosidine (paromomycin) 25 to 35 mg/kg daily in three divided doses for 7 to 10 days—to eradicate residual amebae in the gut.174 The development of a vaccine against E. histolytica has been hampered in part because natural infection does not result in long-term immunity.

KEY REFERENCES Figure 82-11.  Computed tomographic scan showing a large amebic abscess in the left lobe of the liver. (Courtesy of Dr. Mark Feldman, Dallas, Tex.)

Diagnosis The diagnosis of amebic liver abscess is based on clinical suspicion, hepatic imaging, and serologic testing. The organism is isolated from the stool in only 50% of patients. Hepatic imaging studies cannot distinguish a pyogenic from an amebic liver abscess (Fig. 82-11). An amebic abscess is commonly localized to the right hepatic lobe, close to the diaphragm, and usually is single (Table 82-4).169 Available serologic tests include an ELISA and indirect hemagglutination, cellulose acetate precipitin, counterimmunoelec­trophoresis, immunofluorescent antibody, and rapid latex agglutination tests. Serologic test results must be interpreted in the clinical context because serum antibody levels may remain elevated for years after recovery or cure. The sensitivity of these tests is approximately 95%, and the specificity is more than 95%. Falsenegative results may occur within the first 10 days of infection.166 PCR-based tests to detect amebic DNA and an ELISA to detect amebic antigens in serum are available in the research setting.170,171 Aspiration of an amebic abscess should be performed if the diagnosis remains uncertain. The presence of a reddish-brown pasty aspirate (“anchovy paste” or “chocolate sauce”) is typical; trophozoites rarely are identified. Aspiration also may be considered when no response to antibiotic therapy has occurred after five to seven days or when an abscess in the left lobe of the liver is close to the pericardium.172,173

Akritidis N, Tzivras M, Delladetsima I, et al. The liver in brucellosis. Clin Gastroenterol Hepatol 2007; 5:1109-12. (Ref 29.) Bhattacharya SK, Sinha PK, Sundar S, et al. Phase 4 trial of miltefosine for the treatment of indian visceral leishmaniasis. J Infect Dis 2007; 196:591-8. (Ref 76.) Craig P. Echinococcus multilocularis. Curr Opin Infect Dis 2003; 16:43744. (Ref 131.) Fang CT, Lai SY, Yi WC, et al. Klebsiella pneumoniae genotype k1: An emerging pathogen that causes septic ocular or central nervous system complications from pyogenic liver abscess. Clin Infect Dis 2007; 45:284-93. (Ref 148.) Karp CL, Auwaerter PG. Coinfection with Hiv and tropical infectious diseases. I. Protozoal pathogens. Clin Infect Dis 2007; 45:1208-13. (Ref 167.) Keiser J, Utzinger J. Efficacy of current drugs against soil-transmitted helminth infections: Systematic review and meta-analysis. Jama 2008; 299:1937-48. (Ref 97.) Keiser PB, Nutman TB. Strongyloides stercoralis in the immunocompromised population. Clin Microbiol Rev 2004; 17:208-17. (Ref 98.) Khuroo MS, Wani NA, Javid G, et al. Percutaneous drainage compared with surgery for hepatic hydatid cysts. N Engl J Med 1997; 337:881-7. (Ref 128.) Kontoyiannis DP, Luna MA, Samuels BI, Bodey GP. Hepatosplenic candidiasis. A manifestation of chronic disseminated candidiasis. Infect Dis Clin North Am 2000; 14:721-39. (Ref 140.) Lodhi S, Sarwari AR, Muzammil M, et al. Features distinguishing amoebic from pyogenic liver abscess: A review of 577 adult cases. Trop Med Int Health 2004; 9:718-23. (Ref 169.) Manzella A, Ohtomo K, Monzawa S, Lim JH. Schistosomiasis of the liver. Abdom Imaging 2008; 33:144-50. (Ref 109.) Marcos LA, Terashima A, Gotuzzo E. Update on hepatobiliary flukes: Fascioliasis, opisthorchiasis and clonorchiasis. Curr Opin Infect Dis 2008; 21:523-30. (Ref 116.) Salles JM, Salles MJ, Moraes LA, Silva MC. Invasive amebiasis: An update on diagnosis and management. Expert Rev Anti Infect Ther 2007; 5:893-901. (Ref 174.) Scholing M, Schneeberger PM, van den Dries P, Drenth JP. Clinical features of liver involvement in adult patients with listeriosis. Review of the literature. Infection 2007; 35:212-18. (Ref 11.) Full references for this chapter can be found on www.expertconsult.com.

1369

CHAPTE R

83 Vascular Diseases of the Liver William E. Stevens and Abhitabh Patil

CHAPTER OUTLINE Budd-Chiari Syndrome  1371 Etiology  1371 Clinical Features  1372 Pathology  1373 Diagnosis  1373 Treatment  1374 Sinusoidal Obstruction Syndrome (Veno-occlusive Disease)  1375 Etiology  1376 Clinical Features and Diagnosis  1376 Pathology  1376 Treatment  1377

Vascular disorders of the liver are relatively uncommon but frequently result in serious liver disease and portal hypertension.1,2 The continuously high metabolic activity of the liver makes it particularly susceptible to vascular compromise; however, its complex dual blood supply offers unique protection against ischemic injury. Hypercoagulable states play an important role in the pathogenesis of many of these disorders, and knowledge of them is essential for understanding and treating the associated hepatic vascular con­ditions. This chapter reviews a heterogeneous group of disorders resulting from hepatic vascular and cardiovascular diseases. Vasculitis involving the liver is discussed in Chapter 35.

BUDD-CHIARI SYNDROME Hepatic venous outflow obstruction is the hallmark of the Budd-Chiari syndrome. Reductions in hepatic venous outflow can occur anywhere from the right atrium to the small hepatic venules and result in dramatic anatomic and physiologic changes. Classic Budd-Chiari syndrome results from thrombosis of one or more hepatic veins at their openings into the inferior vena cava. The deleterious physiologic changes of hepatic venous obstruction are transmitted directly to the hepatic sinusoids, resulting in sinusoidal congestion, portal vein hypertension, and reduced portal vein blood flow. The result is hepatomegaly, pain, ascites, and impaired hepatic function. The ascitic fluid typically has a high serum-ascites albumin gradient and a high protein content as a result of the increased filtration of serum proteins through the highly permeable sinusoidal spaces (see Chapter 91). The progression of disease is rarely fulminant; in most patients, the clinical course is subacute

Portal Vein Thrombosis  1377 Etiology  1377 Clinical Features and Diagnosis  1378 Treatment  1379 Ischemic Hepatitis  1379 Etiology  1379 Clinical Features and Diagnosis  1379 Prognosis and Treatment  1380 Congestive Hepatopathy  1380 Peliosis Hepatis  1381 Hepatic Artery Aneurysm  1381 Hepatic Artery Atherosclerosis  1382

and less than six months in duration.1 In older series, the mortality rate in untreated cases was as high as 90% at three and one half years.3 With advances in diagnostic imaging and improved medical, surgical, and radiologic treatments for Budd-Chiari syndrome, however, survival has improved significantly since 1980. The literature on Budd-Chiari syndrome is extensive. Two major reviews of the world literature collected data on cases reported before 1980.4,5 Reviews of more than 100 cases have appeared from Japan, India, China, and South Africa.6-9 The review from India is especially helpful in demonstrating the geographic diversity of this syndrome.7

ETIOLOGY

Anatomically, Budd-Chiari syndrome results from hepatic vein obstruction, inferior vena cava obstruction (above or at the level of the hepatic veins), or both. The main causes of the syndrome are listed in Table 83-1. In Western countries, thrombosis of the hepatic veins is the most common presentation, whereas in Asia and Africa, membranous obstruction of the inferior vena cava (MOVC) accounts for more than 40% of cases.10 Thrombogenic states can be identified in at least 75% of cases not caused by MOVC. Increasingly sophisticated testing for hypercoagulable states has reduced the frequency of idiopathic cases to less than 10%.1,2,11 Hematologic disorders are the most common causes of Budd-Chiari syndrome. Primary myeloproliferative diseases, particularly polycythemia vera, may account for 50% of cases.1,2 In addition, latent myeloproliferative disorders may be detected using cell culture techniques12 or by testing for mutations (specifically the V617F mutation) in the gene coding for the tyrosine kinase Janus kinase 2 (JAK2).13 Testing for JAK2 mutations is likely to play an increasing role in the evaluation of patients with Budd-Chiari syndrome and a suspected myeloproliferative disorder. Tumors,

1371

1372

Section IX  Liver Table 83-1  Causes of Budd-Chiari Syndrome Hypercoagulable States Antiphospholipid syndrome Antithrombin deficiency Factor V Leiden mutation Lupus anticoagulant Methylenetetrahydrofolate reductase mutation TT677 Myeloproliferative disorders (including polycythemia vera and essential thrombocytosis) Oral contraceptives Paroxysmal nocturnal hemoglobinuria Postpartum thrombocytopenic purpura Pregnancy Protein C deficiency Protein S deficiency Prothrombin mutation G20210A Sickle cell disease Infections Aspergillosis Filariasis Hydatid cysts Liver abscess (amebic or pyogenic) Pelvic cellulitis Schistosomiasis Syphilis Tuberculosis Malignancies Adrenal carcinoma Bronchogenic carcinoma Hepatocellular carcinoma Leiomyosarcoma Leukemia Renal carcinoma Rhabdomyosarcoma Miscellaneous Behçet’s syndrome Celiac disease Dacarbazine therapy Inflammatory bowel disease Laparoscopic cholecystectomy Membranous obstruction of the vena cava (see Fig. 83-1) Polycystic liver disease Sarcoidosis Trauma to hepatic veins

infections, and pregnancy each account for about 10% of cases. Other hypercoagulable states associated with BuddChiari syndrome include paroxysmal nocturnal hemoglobinuria, antiphospholipid syndrome,14 and deficiencies of antithrombin, protein C, and protein S.15 More recently recognized causes include factor V Leiden mutation16 and mutations of the prothrombin gene and the methylene­ tetrahydrofolate reductase gene.17 Oral contraceptive use increases the risk of Budd-Chiari syndrome by more than two-fold, especially in the presence of other hypercoagu­ lable states. More than 25% of patients are found to have multiple thrombogenic risk factors.17 The pathophysiologic characteristics of MOVC are poorly understood. The disorder is much more common in developing countries, especially Asia and Africa, than in Western countries. More than 70% of cases of Budd-Chiari syndrome in China are due to MOVC. In India the incidence of MOVC, in contrast to classic Budd-Chiari syndrome, seems to be decreasing.18 The clinical presentation usually is subacute or chronic. The membranous webs may be thick or thin and typically occur in the intrahepatic inferior vena cava, often with occlusion of the ostia of the hepatic veins (Fig. 83-1). A congenital origin for the lesion has been proposed on the

Figure 83-1.  Venography demonstrating membranous obstruction of the vena cava. Injection of contrast medium into the inferior vena cava demonstrates retrograde flow toward the legs (downward arrow), rather than antegrade flow toward the right atrium. The membranous obstruction is nearly complete; only a tiny opening is visible as a small protrusion (curved arrow). (Courtesy of Drs. Jeanne LaBerge, Roy Gordon, Robert Kerlin, and Ernest Ring, San Francisco, Calif.)

basis of the complex embryologic development of the inferior vena cava and reported presentations in childhood. An acquired origin, however, is supported by the peak occurrence in the fourth decade of life and histologic studies suggesting that the membrane develops from an organizing thrombus. Hypercoagulability is relatively less common in MOVC than in hepatic vein thrombosis. Therefore, other explanations for thrombosis in MOVC have been proposed, such as chronic infection, endothelial trauma caused by movement of the diaphragm with respiration and coughing, and venous turbulence resulting from the right-angle flow of blood from the hepatic veins into the inferior vena cava. Another feature of MOVC is the propensity of affected persons to develop hepatocellular carcinoma, which is less common in classic Budd-Chiari syndrome.19,20 The distinctive features of MOVC have led some investigators to consider MOVC a separate clinical entity termed obliterative hepatocavopathy.21

CLINICAL FEATURES

The epidemiologic characteristics of Budd-Chiari syndrome (except those cases associated with MOVC) parallel those of its underlying conditions. The syndrome is rare in infants and young children; the largest pediatric series describes South African children with MOVC. More than one half of the cases of classic Budd-Chiari syndrome occur between the ages of 20 and 39 years.5 Budd-Chiari syndrome occasionally may be identified in asymptomatic persons undergoing evaluation for mildly elevated liver biochemical test levels.22 In these patients, the lack of symptoms probably is the result of thrombosis of only one hepatic vein or the development of large venous collaterals.

Chapter 83  Vascular Diseases of the Liver Fulminant Budd-Chiari syndrome is uncommon and occurs most often in women with a hypercoagulable state due to pregnancy, usually in the setting of an additional predisposing factor (e.g., factor V Leiden mutation). Patients present with abrupt and severe abdominal pain and vomiting, marked hepatomegaly, jaundice, ascites, high serum aminotransferase levels (greater than 1000 U/L), and rapid deterioration of hepatic function, with resulting encephalopathy and renal failure. Few of these patients survive without prompt intervention, usually including liver transplantation. Acute Budd-Chiari syndrome is more common, accounting for 20% to 30% of cases. The clinical findings depend on the location of the thrombus, stage and rapidity of evolution, and percentage of liver tissue deprived of venous drainage. Symptoms and signs develop over one to two months and include abdominal pain, tender hepatomegaly, and ascites, but not hepatic encephalopathy. Typically the serum bilirubin level is less than 5 mg/dL, and the serum aminotransferase levels are elevated two- to three-fold. Some patients may have a remitting course. Most patients with Budd-Chiari syndrome present with subacute or chronic symptoms and signs evolving over three to six months. Many of these patients already have cirrhosis and exhibit complications of chronic hepatic decompensation. Hepatosplenomegaly and ascites are usually present. If the inferior vena cava is also occluded, then signs of collateral circulation, such as dilated venous collaterals over the flanks and back, will be present and most patients will have lower extremity edema. Liver biochemical test levels may be nonspecifically, mildly elevated. Some patients will experience variceal bleeding requiring endoscopic therapy. Because Budd-Chiari syndrome is uncommon and symptoms and signs are nonspecific, misdiagnosis or delays in diagnosis are frequent. The manifestations of Budd-Chiari syndrome are similar to those of other causes of fulminant hepatitis, but massive hepatomegaly and rapid development of ascites are suggestive features. Acute alcoholic hepatitis with prominent hepatomegaly, ascites, and mildly abnormal liver biochemical test levels may be confused with Budd-Chiari syndrome. Right-sided congestive heart failure, severe tricuspid regurgitation, and constrictive pericarditis all are associated with impaired venous return from the liver and may mimic subacute or chronic Budd-Chiari syndrome (see later). These cardiac conditions are suggested by the presence of increased jugular venous pressure and hepatojugular reflux; careful cardiovascular examination and echocardiography will assist in the diagnosis. Patients with Budd-Chiari syndrome presenting with right upper quadrant pain and gallbladder wall thickening on ultrasound examination may be misdiagnosed as having acute cholecystitis. Budd-Chiari syndrome should be considered in patients presenting with decompensated cirrhosis or refractory ascites out of proportion to the magnitude of liver biochemical test abnormalities. In addition, anyone with a hypercoagulable disorder who presents with hepatomegaly and ascites should be evaluated for Budd-Chiari syndrome. Although the course of Budd-Chiari syndrome may be indolent, few cases ever regress. Some form of intervention is needed in most patients to prevent ongoing hepatic ischemic necrosis and to control ascites and other complications.

PATHOLOGY

Acutely, the hepatic histologic features of centrilobular congestion, hemorrhage, sinusoidal dilatation, and noninflam­

Figure 83-2.  Histopathologic features of Budd-Chiari syndrome. This lowpower photomicrograph shows the centrizonal congestion, hemorrhage, and hepatocyte necrosis typical of the acute type of Budd-Chiari syndrome. Masson trichrome stain. (Courtesy of Dr. Edward Lee, Washington, DC.)

matory cell necrosis predominate (Fig. 83-2). Within weeks, fibrosis develops in the centrilobular areas, more so than in periportal areas. Over time, these lesions evolve into cirrhosis. Large regenerative nodules are common, especially in areas of decreased portal venous perfusion.23 Indeed, portal vein thrombosis can be seen in 10% of cases and in 50% of hepatic explants at the time of liver transplantation for Budd-Chiari syndrome.24 Because hepatic vein occlusion is asymmetrical, the pathologic effects may vary in different regions of the liver. Massive caudate lobe hypertrophy is a common feature, probably because of preservation of venous drainage directly into the inferior vena cava, and may contribute to compression of the inferior vena cava. Pathologic changes (except for cirrhosis) can be reversed with adequate decompression of the hepatic sinusoids (see later).

DIAGNOSIS

Liver biochemical test levels may be normal or demonstrate only mild elevations or nonspecific changes. Occasionally, serum aminotransferase levels may be more than five times the upper limit of normal in fulminant and acute cases. Often, the severity of the hepatic illness seems out of proportion to the mild degree of abnormality on liver biochemical tests. Doppler ultrasonography, with sensitivity and specificity rates greater than 80%, is the diagnostic procedure of first choice.25 It is relatively inexpensive, safe, and available in most hospitals. Typical Doppler ultrasonographic features of Budd-Chiari syndrome include lack of visualization of normal hepatic venous connections to the inferior vena cava, comma-shaped intrahepatic or subcapsular collateral vessels, and absence of flow signal in the hepatic veins. The diagnostic accuracy of ultrasonography is decreased by a large body habitus and is operator dependent. Magnetic resonance imaging (MRI) and computed tomography (CT)26 may also demonstrate characteristic features of Budd-Chiari syndrome but do not add much to the findings of an adequate ultrasonographic examination.27 MRI may be a better second-line test than CT because of the ability to provide accurate angiographic detail of the hepatic vein and inferior vena cava anatomy with minimal risk of nephro­ toxicity. The combination of Doppler ultrasonography and either MRI or CT imaging should be sufficient to diagnose most cases of Budd-Chiari syndrome.

1373

1374

Section IX  Liver For years, venography was the standard for diagnosis of Budd-Chiari syndrome; however, with improvements in less invasive radiologic imaging, venography is often unnecessary for diagnostic purposes alone. Venography should be performed in cases of suspected Budd-Chiari syndrome when first- and second-line imaging tests are nondiagnostic and when surgery and other therapeutic interventions are planned. Measurement of the hepatic venous pressure gradient is required when vena cava stenosis is present, to plan for portosystemic shunt surgery (see Chapter 90). Venography also allows access for transjugular biopsy of both the right and left hepatic lobes, to confirm the diagnosis of Budd-Chiari syndrome and guide therapy. Liver biopsy is not essential for making the diagnosis of Budd-Chiari syndrome, however, and clinical staging of hepatic synthetic function (e.g., Child-Turcotte-Pugh score) may be better for planning treatment (see later).28

TREATMENT

The therapy of Budd-Chiari syndrome depends on the etiology, anatomic characteristics, and pace of the disease (Fig. 83-3). The precipitating causes of Budd-Chiari syndrome must be evaluated and treated. Most patients need some form of intervention, and collaboration among a hepatologist, interventional radiologist, and hepatobiliary surgeon is optimal. Treatment options consist primarily of combinations of medical therapy with diuretics and anticoagulants, interventional therapy to decompress the hepatic sinusoids and prevent further hepatic necrosis, and liver transplantation for hepatic failure. Medical management alone may be appropriate for the few patients with milder forms of Budd-Chiari syndrome. Treatment typically consists of diuretic therapy with spironolactone and furosemide and dietary sodium restriction to achieve a negative sodium balance (see Chapter 91). Large-volume paracentesis may be needed to relieve tense ascites. Anticoagulation is recommended and consists of

intravenous heparin followed by warfarin to achieve an international normalized ratio (INR) for prothrombin of 2.0 to 2.5. Medical therapy is considered successful if ascites is controlled, liver biochemical test results improve or normalize, and symptoms resolve. Most patients with Budd-Chiari syndrome do not respond adequately to medical therapy alone, however, and require some form of intervention to decompress the hepatic sinusoids. Many case reports and small series have described the use of thrombolytic therapy for acute Budd-Chiari syndrome. Thrombolytic agents are most effective if administered within three weeks of the onset of symptoms, if flow is demonstrated in the thrombosed vein, and if the agent is infused directly into the occluded hepatic vein. Systemic and hepatic arterial infusions are less effective.29 Angioplasty is often performed in conjunction with thrombolysis to improve vein patency and to reopen an acutely thrombosed stent or transjugular intrahepatic portosystemic shunt (TIPS) (see later). The role of interventional radiology in the treatment of Budd-Chiari syndrome has expanded greatly. Angioplasty, with or without placement of a stent through short and localized stenoses of the hepatic veins or the inferior vena cava, can relieve symptoms in more than 80% of patients with MOVC and is the primary treatment for this disorder in many parts of the world.30-33 The rate of restenosis is high, and regular follow-up with Doppler ultrasonography is required.34 Angioplasty can also be combined successfully with surgical creation of a portacaval shunt in patients with both inferior vena cava and hepatic vein obstruction (see later).35 Use of TIPS has gained popularity as a treatment for refractory complications of portal hypertension (see Chapter 90). Procedural mortality is low, and shunt placement is usually successful,36 even in complete hepatic vein occlusion, in which a shunt can be placed between the intrahepatic vena cava and portal vein.37 TIPS is also useful for

Diagnosis of Budd-Chiari syndrome confirmed by ultrasonography or venography

Chronic

Acute

Fulminant hepatic failure

Liver transplantation

Stable hepatic function

Preserved hepatic function

Angioplasty ± stent placement Fails or thrombolysis

Portosystemic shunt or TIPS

Decompensated cirrhosis

Liver transplantation ±pre-transplant TIPS

Figure 83-3.  Approach to the management of patients with Budd-Chiari syndrome. Although each case must be individualized, this algorithm shows typical treatment options for the various forms of Budd-Chiari syndrome. Patients with fulminant hepatic failure or decompensated cirrhosis should be considered for liver transplantation. Patients with clinically stable, acute disease have the most options available to them. Special effort should be made to determine local (nonsystemic) or transient causes of the syndrome that might benefit from less invasive therapies. TIPS, transjugular intrahepatic portosystemic shunt.

Chapter 83  Vascular Diseases of the Liver treating combined hepatic vein and inferior vena cava obstruction and can be effective in patients with fulminant Budd-Chiari syndrome awaiting liver transplantation. Short-term mortality rates in the latter setting may be as high as 50%38; however, long-term clinical stabilization after TIPS placement without liver transplantation also has been reported.39 TIPS has been used successfully in patients with acute Budd-Chiari syndrome in whom thrombolytic therapy and angioplasty have failed. Finally, TIPS is an effective “bridge” to liver transplantation in patients with chronic Budd-Chiari syndrome who have liver failure and refractory ascites or variceal bleeding. Unfortunately, TIPS dysfunction requiring revision occurs in as many as 70% of patients40 and seems to be more common in patients with Budd-Chiari syndrome than in those with other indications for TIPS placement. Polytetrafluoroethylenecovered stents significantly reduce rates of TIPS dysfunction and the need for reintervention and improve primary patency rates at one year from 19% to 67%.41 TIPS placement that extends too far into the portal vein or into the suprahepatic vena cava makes liver transplantation more complicated. In Western countries, surgical therapy for Budd-Chiari syndrome consists mainly of portosystemic shunting and liver transplantation. Surgical therapy of MOVC that is unsuitable for or refractory to angioplasty and stent placement may entail resection or transatrial “finger fracture” of vena cava webs42 or dorsocranial liver resection with a hepaticoatrial anastomosis (the Senning procedure).43 For classic Budd-Chiari syndrome, portosystemic shunting relieves portal hypertension effectively, thereby alleviating hepatic ischemic necrosis, refractory ascites, and variceal bleeding. When shunt surgery is successful, the portal vein becomes the hepatic outflow tract, hepatomegaly resolves, hepatic histologic findings improve and may even normalize, and survival is prolonged in more than 90% of cases.44 The choice of a portosystemic shunt depends on the degree of hepatomegaly and caudate lobe hypertrophy, presence or absence of inferior vena cava stenosis, and surgical expertise. Portacaval and mesocaval shunts are associated with the best results. With mesocaval shunts, the rate of shunt thrombosis is higher—33% at five years45—but placement is technically simpler when portal dissection is impeded because of caudate lobe hypertrophy. Thrombosis of portacaval shunts is uncommon, with a rate of only 3% over 13 years,44 but portacaval shunts increase the technical difficulty of subsequent liver transplantation more than do mesocaval shunts. Dysfunction of a surgical portosystemic shunt significantly reduces long-term survival and is more likely with prosthetic grafts or with high portal venous pressures.46 Placement of a portacaval or mesocaval shunt is contraindicated if inferior vena cava stenosis is present and the vena cava pressure is greater than 20 mm Hg or if the portacaval pressure gradient is less than 10 mm Hg. Surgical options in these patients include placement of a mesoatrial shunt,47,48 a combined portacaval and cavoatrial shunt,44 and combinations of surgical shunt creation with vena cava angioplasty and stent placement.35 Because of high rates of shunt thrombosis, however, TIPS may be the best option in these patients. Liver transplantation is appropriate in patients with liver failure resulting from fulminant or chronic Budd-Chiari syndrome and in patients with a failed surgical porto­ systemic shunt (see Chapter 95).49 In patients with protein C, protein S, or antithrombin deficiency, liver transplan­ tation also cures the underlying hypercoagulable state, although most patients will require lifelong anticoagulation.

Underlying myeloproliferative disorders can be managed with hydroxyurea and aspirin after liver transplantation,50 but there is a risk of progression and leukemic transformation.51 Recurrent Budd-Chiari syndrome after liver transplantation occurs in 4% to 10% of patients, and the risk of thrombosis of the hepatic artery and portal vein is increased as well.52,53 In addition, bleeding complications are more common because of anticoagulant therapy.54 Despite these drawbacks, five-year survival rates for patients with Budd-Chiari syndrome who undergo liver transplantation exceed 85%.53 The choices of therapy in patients with Budd-Chiari syndrome are complicated, in great part because of the lack of large, controlled studies comparing various treatments and the lack of standardization in classifying Budd-Chiari syndrome into acute, subacute, and chronic forms.55 One multivariate analysis failed to show a survival advantage in surgically shunted patients compared with those receiving medical treatment alone.56 Although medium-term studies of TIPS as definitive therapy in Budd-Chiari syndrome have been reported, there are no long-term controlled studies comparing TIPS with surgical therapy. TIPS is clearly an effective and widely available therapy; however, shunt dysfunction is common and leads to frequent reinterventions. On the other hand, surgical shunts are associated with significant perioperative morbidity and mortality, but when performed successfully, they offer good long-term outcomes. Surgical expertise in complex shunt surgery, however, is declining with the increasing use of TIPS. Surgical shunt surgery seems to be most efficacious in patients with acute or chronic Budd-Chiari syndrome with refractory symptoms and preserved hepatic synthetic function (Child classes A and B [see Chapter 90]).57 Patients with fulminant or chronic Budd-Chiari syndrome with impaired hepatic synthetic function (Child classes B and C) should be considered for liver transplantation, with or without prior TIPS placement. (As compared with the Child class, the Model for End-stage Liver Disease (MELD) score does not incorporate an assessment of refractory ascites and is less useful for estimating prognosis in patients with Budd-Chiari syndrome than in those with other causes of advanced chronic liver disease [see Chapter 95]). In patients with acute Budd-Chiari syndrome, attempts should be made to decompress the hepatic sinusoids with combinations of angioplasty, stent placement, and thrombolytic therapy, followed by TIPS placement or creation of a surgical portosystemic shunt in those in whom these treatment approaches fail. A proposed stepwise (“minimally invasive”) approach beginning with anticoagulation therapy, then attempted hepatic vein recanalization, followed by TIPS, and then liver transplantation in treatment failures has been reported to achieve an overall five-year survival rate of nearly 90%.58

SINUSOIDAL OBSTRUCTION SYNDROME (VENO-OCCLUSIVE DISEASE) Occlusion of the terminal hepatic venules and hepatic sinusoids resembles the Budd-Chiari syndrome clinically; however, the causes, epidemiologic and pathophysiologic characteristics, and prognosis of this entity are sufficiently distinct to justify a separate designation. In the past, the entity was known as veno-occlusive disease. Because the occlusion consistently involves the hepatic sinusoids, the term sinusoidal obstruction syndrome has been proposed as a more appropriate name for this disorder.59

1375

1376

Section IX  Liver ETIOLOGY

Liver disease caused by Senecio poisoning was originally reported from South Africa in 1920 (see Chapter 87),60 but the term veno-occlusive disease was not used until 1954, when it was related to the ingestion of pyrrolizidine alkaloids contained in plants of the genera Senecio, Crotalaria, and Heliotropium in Jamaica.61 Ingestion of alkaloids in inadequately winnowed wheat or in “bush tea,” especially in malnourished persons, is the main cause of venoocclusive disease (sinusoidal obstruction syndrome) worldwide. Epidemics have been reported in India, Afghanistan, South Africa, the Middle East, and the United States. More recently, the herbal remedy comfrey (genus Symphytum) has been associated with sinusoidal obstruction syndrome. Rare familial clusters have been reported in association with immunodeficiency states.62 Since the advent of cancer chemotherapy in the 1950s, sinusoidal obstruction syndrome in Western countries has occurred most commonly after bone marrow transplantation (see Chapter 34).63 A variety of antineoplastic drugs have been implicated as causes of sinusoidal obstruction syndrome including gemtuzumab ozogamicin, actinomycin D, dacarbazine, cytosine arabinoside, mithramycin, and 6thioguanine. Hepatic irradiation and therapy with busulfan plus cyclophosphamide also are established risk factors (see Chapters 34 and 39). In addition, long-term immunosuppression with azathioprine and 6-thioguanine used in patients with inflammatory bowel disease and in kidney and liver transplant recipients has been reported to cause sinusoidal obstruction syndrome.63 The frequency of sinusoidal obstruction syndrome after bone marrow transplantation varies, ranging from 0% to 70%, depending on patient and treatment-related variables.

CLINICAL FEATURES AND DIAGNOSIS

In the early stages of the disorder, features of portal hypertension predominate. Classically, sinusoidal obstruction syndrome manifests with mild hyperbilirubinemia (bilirubin levels greater than 2 mg/dL), painful hepatomegaly, weight gain of more than 2%, and development of ascites. Weight gain and painful hepatomegaly usually precede the onset of jaundice, which can be followed by the development of ascites, encephalopathy, and multiorgan failure. Sinusoidal obstruction syndrome occurs most commonly within 10 to 20 days after bone marrow transplantation (see Chapter 34). Serum bilirubin levels typically peak at day 17 after transplantation. Predisposing factors include pretreatment with norethisterone to suppress menses, preexisting elevated serum aminotransferase levels and chronic liver diseases such as hepatitis C, the HFE C282Y mutation for genetic hemochromatosis,64 advanced age, recent systemic bacterial or viral infections, previous bone marrow transplantation, and allogeneic (as opposed to autologous) bone marrow transplantation.63 The diagnosis of sinusoidal obstruction syndrome is often based on characteristic clinical features and exclusion of other conditions. Distinguishing this disease from the many other disorders that may complicate bone marrow transplantation is challenging. The most frequent conditions that can mimic sinusoidal obstruction syndrome include graftversus-host disease, hepatic dysfunction caused by sepsis and drug toxicity, and cholestasis resulting from hemolysis and congestive heart failure. Graft-versus-host disease is rare before day 15 after transplantation, and sepsis and drug toxicity rarely cause painful hepatomegaly and ascites. Routine liver biochemical test results are not specific for sinusoidal obstruction syndrome. Serum alkaline phos­

phatase and aminotransferase elevations can accompany the hyperbilirubinemia and probably indicate coexisting hepatic ischemic necrosis. Thrombocytopenia is common and may be aggravated by portal hypertension and splenic sequestration. Findings on ultrasound examination, CT, and MRI are nonspecific in early sinusoidal obstruction syndrome but are helpful in excluding biliary obstruction, infiltrative liver lesions, and hepatic vein occlusion. Common findings in sinusoidal obstruction syndrome include gallbladder wall thickening, hepatosplenomegaly, ascites, portal vein enlargement with sluggish or reversed flow, and umbilical vein recanalization. Ultrasound and Doppler findings can predict disease severity noninvasively.65 If the diagnosis is uncertain, a transjugular liver biopsy and measurement of the hepatic venous pressure gradient can be performed (see Chapter 90). In this setting, a gradient of greater than 10 mm Hg is highly suggestive of sinusoidal obstruction syndrome and predictive of increased disease severity.66 Serum levels of a variety of mediators of coagulation, inflammation, and fibrosis have been found to be predictive of sinusoidal obstruction syndrome before it becomes clinically apparent. Levels of protein C, factor VII, and antithrombin are decreased in plasma before the onset of sinusoidal obstruction syndrome, whereas levels of plasminogen activator inhibitor type 1, tumor necrosis factor-α, and procollagen type III are increased.67 In patients who undergo hematopoietic stem cell transplantation, an elevated procollagen type III level—even before chemotherapy—is a risk factor for the later development of the disorder.68 It is unclear if these abnormalities cause the occlusion or result from the occlusion or hepatic injury. In most patients with sinusoidal obstruction syndrome, the disorder resolves gradually over two to three weeks. The overall mortality rate is 20% to 50%, with most deaths resulting from multiorgan failure rather than hepatic failure.63 Severe sinusoidal obstruction syndrome, which carries a mortality rate of almost 100%, can be predicted by more rapid increases in the serum bilirubin level and weight. Other predictors of severe disease include the presence of ascites, a hepatic venous pressure gradient greater than 20 mm Hg, and the onset of multiorgan failure.69

PATHOLOGY

The principal histologic features of sinusoidal obstruction syndrome result from toxic injury to the centrilobular (zone 3) endothelial cells in the hepatic sinusoids and terminal hepatic venules (see Fig. 34-5A and B).70 The resulting cellular debris, exfoliated hepatocytes, activated coagulation factors, and extravasated red blood cells produce pro­ gressive occlusion of the sinusoids and venules, causing sinusoidal dilatation and severe hepatic congestion. Inflammation is notably absent. Progressive venular sclerosis ensues, with deposition of collagen in the sinusoids and venules eventually leading to venular obliteration, hepatocellular ischemic necrosis, and widespread fibrosis. Animal studies have shown that toxin-mediated depolymerization of F-actin in sinusoidal epithelial cells leads to rounding and swelling of the cells with dissection of the sinusoidal lining off the space of Disse. This leads to embolization of cellular debris and obstruction of the hepatic sinusoids. F-actin depolymerization seems to be mediated by metalloproteinases, which in turn are affected by nitric oxide production; sinusoidal obstruction syndrome can be prevented in animal models by inhibition of metalloprotenase-971 and stimulation of nitric oxide production.72 In addition, sinusoidal obstruction syndrome

Chapter 83  Vascular Diseases of the Liver is less likely to occur in humans who have enhanced urea cycle activity because of the carbamyl-phosphate synthetase (CPS) A allele, which increases hepatic nitric oxide production.64

TREATMENT

The lack of effective and safe treatment for sinusoidal obstruction syndrome has increased interest in preventive strategies. Recognizing risk factors for the development of the disorder and altering chemotherapeutic regimens and doses can decrease associated morbidity and mortality. Studies of ursodeoxycholic acid, heparin, and lowmolecular-weight heparin (LMWH) for the prevention of sinusoidal obstruction syndrome have yielded inconclusive results. Ursodeoxycholic acid has minimal toxicity, and in two studies it reduced the frequency of the disease by more than 50% compared with placebo or no treatment but had no effect on overall survival.73,74 One large study failed to show any preventive benefit of ursodeoxycholic acid. Similarly, heparin reduced the frequency of sinusoidal obstruction syndrome in a large, prospective, randomized trial from 13.7% in control subjects who received no treatment to 2.5% in patients in the group treated with heparin,75 but two other studies showed no preventive benefit from heparin and a 7% risk of hemorrhagic complications.74,76 The combination of ursodeoxycholic acid and heparin offers no advantage over heparin alone.77 Use of LMWH is associated with a lower rate of bleeding complications than that observed with heparin and may be more effective. In a retrospective study in which LMWH was compared with heparin and no treatment, the frequency of sinusoidal obstruction syndrome was 4% with LMWH, 11% with heparin, and 22% with no treatment.78 Studies of prostaglandin E1 to prevent sinusoidal obstruction syndrome have also yielded inconclusive results, and use of this agent is limited by its significant toxicity. Treatment of sinusoidal obstruction syndrome primarily entails supportive care, including diuresis, analgesia for pain, paracentesis for tense ascites, and avoidance of nephro­ toxins and other hepatotoxins. Survival when multiorgan failure has occurred is dismal. Treatment with tissue plasminogen activator, as described in many case reports and small series, has shown response rates of 30% and a rate of life-threatening hemorrhage of 20% to 30%.74 Treatment is ineffective in patients in whom multiorgan failure has developed. Other treatments reported in small studies or in trials with inconclusive results include the use of highdose glucocorticoids, N-acetylcysteine, antithrombin,74 and protein C. Defibrotide is a novel oligodeoxyribonucleotide with antiischemic, antithrombotic, and thrombolytic activity but minimal systemic anticoagulant effect. Several studies have shown efficacy of defibrotide in the prevention and treatment of sinusoidal obstruction syndrome, with no major toxicity. Prophylactic defibrotide plus heparin prevented the development of sinusoidal obstruction syndrome in 52 successive bone marrow or stem cell transplant patients, compared with a 19% disease incidence in a historical control group.79 A large multi-institutional study of defibrotide in the treatment of severe sinusoidal obstruction syndrome showed a 36% rate of response (improvement in serum bilirubin levels and multiorgan failure) in a group of patients with an otherwise dismal prognosis.80 Defibrotide has gained orphan drug status in the United States, and a large prospective phase III study is in progress. TIPS has been performed successfully in patients with sinusoidal obstruction syndrome, but with generally poor

results. Clinical improvement after TIPS may be seen in 50% of patients, but only 10% survive long term.81 Until more studies are available, TIPS should be reserved for patients with severe sinusoidal obstruction syndrome who have refractory ascites or who are candidates for liver transplantation. Liver transplantation has been performed successfully for liver failure caused by sinusoidal obstruction syndrome. Most patients with sinusoidal obstruction syndrome, however, are unsuitable for liver transplantation because of underlying malignancy and the severity of the associated multiorgan failure. Because a majority of patients with mild to moderate sinusoidal obstruction syndrome do well with conservative therapy, the focus on future investigation should be directed at prevention and treatment of severe sinusoidal obstruction syndrome. As the ability to predict severe sinusoidal obstruction syndrome improves and as newer agents, such as defibrotide, become available, morbidity and mortality from sinusoidal obstruction syndrome will likely decrease.

PORTAL VEIN THROMBOSIS Portal vein obstruction results from thrombosis, constriction, or invasion of the portal vein. The resulting portal hypertension leads to splenomegaly and formation of portosystemic collaterals and esophageal, gastric, duodenal, and jejunal varices. Varices proliferate in the porta hepatis and involve the gallbladder and bile duct. As portal vein thrombosis evolves, fibroblasts transform the clot into a firm, collagenous plug in which tortuous venous channels develop. This cavernous transformation (resulting in a portal cavernoma) begins within days after acute thrombosis and continues to evolve over weeks to months. Upstream from the obstruction, the small intestine and colon become congested, and the stomach exhibits changes of portal hypertensive gastropathy (see Chapter 90). Mesenteric is­ chemia can occur if the thrombus extends into the mesenteric veins (see Chapter 114). Downstream from the clot, the liver usually maintains normal function and appears unaffected. Ascites, abdominal pain, and fever may develop with acute thrombosis, but usually recede subsequently. As more venous collaterals form, a state of equilibrium is reached in which some portal perfusion is maintained and some portal hypertension persists. Clinically, chronic portal vein thrombosis usually is asymptomatic until variceal bleeding occurs.

ETIOLOGY

Most cases of portal vein thrombosis have an identifiable cause related to hypercoagulability or to local factors such as inflammation, trauma, or malignancy (Table 83-2). Fewer than 20% of cases are considered idiopathic. Better understanding of the multiple causes of hypercoagulability has led to the recognition that multiple coexisting risk factors are present in as many as 40% of affected patients.82 Infection, most often from umbilical vein sepsis, is the main cause of portal vein thrombosis in children. Portal vein thrombosis is well documented after neonatal umbilical vein catheterization but resolves in more than 50% of cases.83 In adults, cirrhosis or abdominal malignancies are responsible for more than 50% of the cases of portal vein thrombosis.84 The disorder occurs in at least 10% of patients with cirrhosis, presumably as a result of sluggish portal vein blood flow, but acquired and inherited hypercoagulable states can be identified in many patients with cirrhosis and portal vein thrombosis.85 Prior variceal sclerotherapy,

1377

1378

Section IX  Liver Table 83-2  Causes of Portal Vein Thrombosis Hypercoagulable States Antiphospholipid syndrome Antithrombin deficiency Factor V Leiden mutation Methylenetetrahydrofolate reductase mutation TT677 Myeloproliferative disorder (including polycythemia vera) Nephrotic syndrome Oral contraceptives Paroxysmal nocturnal hemoglobinuria Pregnancy Prothrombin mutation G20210A Protein C deficiency Protein S deficiency Sickle cell disease Inflammatory Diseases Behçet’s syndrome Inflammatory bowel disease Pancreatitis Infections Appendicitis Cholangitis Cholecystitis Diverticulitis Liver abscess (amebic or pyogenic) Schistosomiasis Umbilical vein infection

Complications of Therapeutic Interventions Alcohol injection Colectomy Endoscopic sclerotherapy Fundoplication Gastric banding Hepatic chemoembolization Hepatobiliary surgery Islet cell injection Liver transplantation Peritoneal dialysis Radiofrequency ablation of hepatic tumor(s) Splenectomy TIPS procedure Umbilical vein catheterization Impaired Portal Vein Flow Budd-Chiari syndrome Cirrhosis Cholangiocarcinoma Hepatocellular carcinoma Nodular regenerative hyperplasia Pancreatic carcinoma Sinusoidal obstruction syndrome Miscellaneous Bladder cancer Choledochal cyst Living at high altitude

Figure 83-4.  Doppler ultrasound image of portal vein thrombosis. The arrow points to the portal vein thrombus. The red coloring indicates disordered blood flow around the thrombus toward the liver.

TIPS, transjugular intrahepatic portosystemic shunt.

abdominal surgeries,86 and hepatocellular carcinoma are major risk factors for portal vein thrombosis in patients with cirrhosis. Hepatocellular and pancreatic carcinomas are the most common malignant causes of portal vein thrombosis,84 usually because of a combination of hypercoagulability and invasion or constriction of the portal vein. Local inflammatory reactions resulting from acute or chronic pancreatitis are a common cause of portal vein thrombosis. Pylephlebitis, or septic portal vein thrombosis, can complicate intraabdominal infections such as appendicitis, diverticulitis, and cholangitis. Splenic vein trauma during splenectomy results in portal vein thrombosis in 8% of cases; the risk increases to 40% if a myeloproliferative disorder is present.87 In addition, portal vein thrombosis acutely complicates 5% of pancreaticoduodenectomies with portal vein reconstruction and is chronic in up to 17% of such cases.88 In this setting, portal vein thrombosis seems to have a low morbidity but is commonly associated with recurrent malignancy.

CLINICAL FEATURES AND DIAGNOSIS

Portal vein thrombosis is found with equal frequency in adults (mean age, 40 years) and children (mean age, 6 years). The initial manifestation is almost always hematemesis from variceal bleeding. Abdominal pain is unusual unless the thrombosis is acute or involves the mesenteric veins and causes intestinal ischemia. Splenomegaly is usually present, but ascites is uncommon, except in acute portal vein thrombosis or when the thrombosis complicates cirrhosis. Liver biochemical test results are usually normal. Occasionally, bile duct varices can cause biliary obstruction89 and even mimic cholangiocarcinoma on cholangiography.90 Other unusual locations for ectopic varices in portal vein thrombosis include the gallbladder, duodenum, and rectum.

Figure 83-5.  Magnetic resonance angiographic appearance of cavernous transformation of the portal vein (portal cavernoma). The arrow points to numerous small tortuous collateral vessels that replace the normal portal vein anatomy.

Doppler ultrasonography is highly sensitive for detection of portal vein thrombosis and reveals an echogenic thrombus in the portal vein (Fig. 83-4), extensive collateral vessels in the porta hepatis, an enlarged spleen, and occasionally nonvisualization of the portal vein. When the diagnosis of portal vein thrombosis is still uncertain, magnetic resonance angiography is better than CT in demonstrating the typical changes of portal vein thrombosis (portal cavernoma) (Fig. 83-5). Portal venography usually is unnecessary unless a surgical portosystemic shunt is being

Chapter 83  Vascular Diseases of the Liver considered. Evaluating the patient for precipitating hypercoagulable risk factors may require a consultation with a hematologist. The natural history of portal vein thrombosis is related primarily to the underlying disorder. In the absence of cirrhosis, cancer, and mesenteric vein thrombosis, the 10-year survival rate for patients with portal vein thrombosis is greater than 80%; only 2% experience fatal variceal hemorrhage.84 Variceal bleeding caused by portal vein thrombosis has a much better outcome than that observed with variceal bleeding caused by cirrhosis because of preserved hepatic function and lack of coagulopathy in patients with thrombosis alone. In addition, development of spontaneous portosystemic collaterals can lead to a reduced frequency of recurrent variceal bleeding in patients with portal vein thrombosis.

Because hepatitis refers to inflammation of the liver, the term ischemic hepatitis is somewhat of a misnomer because inflammation is typically not present. A more physiologic term would be hypoxic hepatitis because the primary cause of this syndrome is tissue hypoxia, which may be the result of hypoperfusion from cardiac failure, systemic hypoxemia from respiratory failure, or increased oxygen requirements from sepsis.100 The name ischemic hepatitis is used, however, because of clinical similarities to other forms of acute hepatitis and the characteristic pathologic feature of acute centrilobular necrosis. Ischemic hepatitis is probably the most commonly encountered form of vascular liver disease.

TREATMENT

ETIOLOGY

Endoscopic band ligation or sclerotherapy is first-line therapy for variceal bleeding in patients with portal vein thrombosis (see Chapter 90). Sessions should be repeated until the varices are obliterated. Therapy with a beta blocker is beneficial in preventing initial91 and, in combination with endoscopic therapy, recurrent variceal bleeding. Recurrent or refractory variceal bleeding or bleeding from varices distal to the esophagus is an indication for placement of a portosystemic shunt. TIPS is an option if the technical challenge of gaining access to the portal vein can be overcome. Focal malignant portal vein obstruction can be stented via a percutaneous approach to control refractory variceal bleeding and ascites.92 Elective mesocaval and splenorenal shunts93 and the extended Sugiura procedure (esophagogastric devascularization and transection)94 have also been performed successfully in patients with portal vein thrombosis, with low mortality and long survival. In children with refractory complications of portal vein thrombosis, a Rex shunt (mesenteric-to-left-portal-vein bypass) is also an effective option.95 Anticoagulation is recommended in patients with acute portal vein thrombosis, to prevent cavernous transformation and complications of portal hypertension. Spontaneous recanalization with acute thrombosis is rare, but therapeutic recanalization can be achieved in more than 80% of the cases with anticoagulants (intravenous heparin or subcutaneous LMWH, followed by warfarin to achieve an INR of 2.0 to 2.5 for at least six months).96 Prompt use of broadspectrum antibiotics in cases of septic pylephlebitis also promotes thrombolysis. Selective transcatheter infusions of thrombolytic agents have been used successfully in acute portal vein thrombosis97 when associated with mesenteric vein thrombosis and intestinal ischemia but are associated with a high complication rate. Long-term anticoagulation should be considered in patients with portal vein thrombosis and a recognized hypercoagulable state, surgical portosystemic shunt, or concomitant mesenteric vein thrombosis. The safety of chronic anticoagulation in patients with portal vein thrombosis and varices has been a major concern, but one study has shown that long-term anticoagulation does not increase the risk or severity of variceal bleeding and prevents further portal and mesenteric venous thrombotic complications.98 In general, however, anticoagulants are not recommended for chronic portal vein thrombosis, especially when associated with cavernous transformation. Liver transplantation for liver failure complicated by portal vein thrombosis is associated with increased technical complexity and a high rethrombosis rate, but overall survival rates are not reduced when compared with rates for other indications.99

ISCHEMIC HEPATITIS

Of all cases of extreme serum aspartate aminotransferase (AST) elevations (to more than 3000 U/L), ischemic hepatitis accounts for about half.101 The most common cause of ischemic hepatitis is cardiovascular disease, which accounts for more than 70% of cases, followed in frequency by respiratory failure and sepsis, each of which accounts for less than 15% of cases.100 Hypotension is documented as a precipitating factor in more than 50% of patients with ischemic hepatitis but does not need to be evident for ischemic hepatitis to occur. Hypotension often is clinically apparent as a result of acute myocardial infarction, severe congestive heart failure, or sepsis but may be less obvious following a transient arrhythmia or silent coronary ischemic event. The presence of congestive heart failure significantly increases the likelihood that a drop in the cardiac output from any cause will result in ischemic hepatitis. Indeed, more than 80% of cases of ischemic hepatitis occur in the setting of congestive heart failure.102 Acute trauma, hemorrhage, burns, and heat stroke can also cause ischemic hepatitis, but the likelihood is substantially less in the absence of underlying heart disease.

CLINICAL FEATURES AND DIAGNOSIS

Ischemic hepatitis often is first considered when extreme serum aminotransferase elevations are detected in a patient hospitalized for problems not primarily associated with the liver. Findings on physical examination are usually dominated by the underlying precipitating medical condition. The patient’s mental status is often altered because of diminished cerebral perfusion. Laboratory studies show extreme elevations of the aminotransferase levels (more than 3000 U/L). The serum lactate dehydrogenase (LDH) level is profoundly elevated, often more so than the alanine aminotransferase (ALT), and an ALT-to-LDH ratio of less than 1.5 is more typical of ischemic hepatitis than of viral hepatitis.103 The prothrombin time may be prolonged by two or three seconds, and the serum bilirubin level is often mildly increased, with peak levels seen after the aminotransferase levels peak. Serum creatinine and blood urea nitrogen levels are often elevated because of acute tubular necrosis. Characteristically, serum aminotransferase levels peak 1 to 3 days after the hemodynamic insult and return to normal within 7 to 10 days. The differential diagnosis for this type of severe acute injury includes acute hepatitis caused by viral infections, autoimmunity, toxins, and medications (see also Chapter 73). Liver biopsy specimens, although usually unnecessary, reveal bland, centrilobular necrosis with preservation of the hepatic architecture (Fig. 83-6). Occasionally a definitive diagnosis of ischemic hepatitis can be difficult to make, but

1379

Section IX  Liver likely to occur when chronic congestive heart failure or cirrhosis is also present. The overall prognosis depends primarily on the severity of the underlying predisposing condition, not the severity of the liver disease. No specific therapy exists for ischemic hepatitis, and treatment is directed at improving cardiac output and systemic oxygenation.

CONGESTIVE HEPATOPATHY

Figure 83-6.  Histopathologic features of ischemic hepatitis. This lowpower photomicrograph demonstrates centrilobular necrosis, loss of hepatocytes, and sinusoidal congestion with red blood cells, but only a scant inflammatory infiltrate. Perivenular fibrosis is evident. (Hematoxylin and eosin.) (Courtesy of Dr. Pamela Jensen, Dallas, Tex.)

10,000

8000 LDH 6000 U/L

1380

4000

2000

AST

0 0

1

2

3

4

5

6

7

8

9

10

Days Figure 83-7.  A frequently observed course of serum aspartate aminotransferase (AST) and lactate dehydrogenase (LDH) levels in ischemic hepatitis. (Adapted from Gitlin NG, Serio KM. Ischemic hepatitis: Widening horizons. Am J Gastroenterol 1992; 7:831.)

the typical prompt rise in serum aminotransferase and LDH levels followed by a rapid fall within a few days is more characteristic of ischemic hepatitis than of other causes of severe acute liver injury (Fig. 83-7).104

PROGNOSIS AND TREATMENT

Most cases of ischemic hepatitis are transient and selflimited. In the most severely affected patients, ischemic hepatitis is just one manifestation of multiorgan failure and signals a poor prognosis. Fulminant hepatic failure resulting from ischemic hepatitis is uncommon but is more

The effects of congestive heart failure on the liver predictably include decreased hepatic blood flow, increased hepatic vein pressure, and decreased arterial oxygen saturation.105 Right-sided heart failure results in transmission of increased central venous pressure from the heart directly to the hepatic sinusoids. The result is centrilobular congestion and sinusoidal edema that further decrease oxygen delivery. The injurious effects of superimposed ischemic hepatitis are common in these patients. The acute and chronic damage results in progressive centrilobular fibrosis. Sinusoidal hypertension and congestion can lead to the development of ascites, with a characteristically high serum-ascites albumin gradient and a high protein concentration (see Chapter 91). Clinically, the symptoms and signs of congestive heart failure are the predominant features. Dull right upper quadrant pain in association with hepatomegaly is common. The liver may be pulsatile if tricuspid regurgitation is present, and hepatojugular reflux is often apparent on compression over the liver. Spider angiomata and varices are usually not present, and variceal bleeding caused by congestive hepatopathy alone does not occur. Mild elevation of the serum bilirubin level (to less than 3 mg/dL) is common, and jaundice is seen in fewer than 10% of patients, occurring in those with severe or acute congestive heart failure.106 The prothrombin time is prolonged in more than 75% of cases and usually is resistant to therapy with vitamin K. Other liver biochemical test levels are often normal or only mildly elevated. Liver test results improve slowly or normalize with effective therapy of the underlying congestive heart failure. Ultrasound is useful in excluding other hepato­ biliary diseases. Typical CT findings in congestive hepatopathy include hepatomegaly, ascites, dilatation of the inferior vena cava and hepatic veins, and inhomogeneous hepatic enhancement during the portal phase of contrast administration. The histologic features of congestive hepatopathy include atrophy of hepatocytes, sinusoidal distention, and centrilobular fibrosis (Fig. 83-8). Centrilobular necrosis, consistent with ischemic hepatitis, is frequent in liver biopsy specimens that show congestive hepatopathy and usually correlates with recent hypotension.107 Bridging fibrosis typically extends between central veins (rather than between portal tracts) to produce a pattern of “reverse lobulation” characteristic of cardiac cirrhosis. The distribution of fibrosis throughout the liver is highly variable and correlates with focal sinusoidal thrombosis, with obliteration of central and portal veins that leads in turn to localized ische­ mia, parenchymal extinction, and fibrosis.108 The presence of congestive hepatopathy does not affect the prognosis in patients with heart failure; the mortality rate is determined primarily by the severity of the underlying cardiac disease. Occasionally, paracentesis may be needed to alleviate tense ascites, but therapy is generally directed at improving cardiac disease.

Chapter 83  Vascular Diseases of the Liver

Figure 83-8.  Histopathologic features of cardiac cirrhosis. This low-power view shows a portal tract in the center of a regenerative nodule and fibrotic bands bridging central veins. The size of the scar and the presence of the nodule attest to the long-term course of the fibrotic process. Even at low power, the bland nature of the cirrhosis is apparent. No inflammatory cells are evident. The sinusoids are dilated and congested. Masson trichrome stain. (Courtesy of Dr. Edward Lee, Washington, DC.)

Figure 83-9.  Histopathologic features of peliosis hepatis. Note the presence of three blood-filled cysts without lining cells and an adjacent portal tract. Sinusoidal dilatation is also present. (Hematoxylin and eosin.) (Courtesy of Dr. Edward Lee, Washington, DC.)

PELIOSIS HEPATIS Peliosis hepatis is characterized by the presence of multiple blood-filled cavities distributed randomly throughout the liver. The cavities range in size from a few millimeters to 3 cm across and usually are seen in association with dilated hepatic sinusoids (Fig. 83-9). Two histologic types of peliosis hepatis occur. In the parenchymal type, blood-filled cavities are lined by hepatocytes, and hemorrhagic parenchymal necrosis and congestion usually are present. In the phlebectatic type, the cavities are lined by endothelial cells associated with aneurysmal dilation of the central vein.109 Fibrosis, cirrhosis, regenerative nodules, and tumors also may be seen with peliosis. In the past, peliosis hepatis was largely a histologic curiosity, but its association with renal transplantation and the acquired immunodeficiency syndrome (AIDS) have increased clinical awareness of this syndrome.

Although jaundice, painful hepatomegaly, liver failure, and fatal hemorrhage may be manifestations of peliosis, more often the disorder is detected during evaluation of abnormal liver biochemical test results in an asymptomatic patient. If the hemorrhagic cavities are large enough, they can be detected by ultrasonography, CT, and magnetic resonance imaging.110 The pathogenesis of peliosis hepatis is unknown, but the leading theories include damage to sinusoidal endothelial cells, outflow obstruction of blood flow at the sinusoidal level, and hepatocellular necrosis. Before the 1970s, peliosis was most often identified in patients dying of wasting diseases, particularly tuberculosis and carcinomatosis.109 Since then, peliosis has been associated with the use of anabolic steroids, oral contraceptives, tamoxifen, danazol, vitamin A, glucocorticoids, 6-thioguanine and azathioprine, and exposure to urethane, vinyl chloride, and thorium dioxide (see Chapter 87). The syndrome can regress with discontinuation of the causative agent. Myeloproliferative diseases such as agnogenic myeloid metaplasia111 and malignant histiocytosis,112 infections such as E. coli pyelonephritis,113 and Castleman disease (giant lymph node hyperplasia)114 have also been associated with peliosis. Bacillary peliosis is caused by the bacterium responsible for cat-scratch disease, in the genus Bartonella. Traumatic exposure to cats is a recognized risk factor. The syndrome has been reported in immunocompetent persons but is associated primarily with AIDS (see Chapters 33 and 82).115 In affected patients, vascular proliferative lesions are found commonly in the skin (where they are termed bacillary angiomatosis) or in the liver and spleen (where they are termed bacillary peliosis and where bacilli can be identified histologically adjacent to the peliotic lesions). Symptoms and signs include anorexia, abdominal pain, fever, lymphadenopathy, hepatosplenomegaly, and cutaneous vascular lesions or nodules.116 Anemia, an elevated serum alkaline phosphatase level, and a CD4+ lymphocyte count of less than 200/mm3 are typical laboratory findings. Bacillary peliosis responds to antibiotic therapy (e.g., erythromycin, doxycycline) (see Chapter 82). Peliosis hepatis occurs in 20% of patients after kidney transplantation and is associated mainly with the prolonged use of azathioprine and possibly cyclosporine.117 The lesions often are asymptomatic or associated with abnormal liver biochemical test levels, but progressive fibrosis, cirrhosis, and portal hypertension may be additional findings. Hepatic lesions may regress on withdrawal of azathioprine, but the overall course is not clearly modified and the risk of transplant rejection is increased substantially.118

HEPATIC ARTERY ANEURYSM Hepatic artery aneurysms (HAAs) are uncommon, but they are the second leading category of visceral artery aneurysms (after splenic artery aneurysms) and account for more than 20% of cases. A majority of true HAAs are isolated, saccular, extrahepatic lesions involving the full arterial wall. In the past, HAAs were mainly mycotic (infectious) in etiology, but today they typically result from atherosclerosis, mediointimal degeneration, trauma, and, less commonly, infection. Other rare causes of true HAA are vasculitides, such as polyarteritis nodosa, systemic lupus erythematosus, Takayasu arteritis, and Kawasaki disease, and connective tissue disorders, such as Marfan syndrome, Ehlers-Danlos syndrome, and Osler-Weber-Rendu disease.119 Approxi-

1381

1382

Section IX  Liver mately one half of HAAs are pseudoaneurysms (aneurysms resulting from injury). Procedures commonly associated with hepatic artery pseudoaneurysms include liver biopsy, transhepatic biliary drainage, cholecystectomy, hepatectomy, and liver transplantation.120 Symptoms of HAA include epigastric or right subchondral pain, but most affected persons are asymptomatic until the aneurysm ruptures. Rarely, a pulsatile right upper quadrant mass or thrill may be detected. Patients may present with rupture into the biliary tree, with hemobilia, epigastric pain, and icterus; rupture into the portal vein, with portal hypertension and variceal bleeding; or rupture into the peritoneal cavity, with abdominal pain and shock. The mortality rate from rupture of a HAA is more than 30%. Nonatherosclerotic aneurysms and multiple HAAs carry an increased risk of rupture and should be treated. Although the risk of rupture of an aneurysm is independent of its size, atherosclerotic aneurysms greater than 2 cm in diameter also should be treated.119 Doppler ultrasound studies and CT readily demonstrate HAAs, but angiography is especially useful for defining these lesions, accessing the collateral circulation, and planning treatment. Hepatic artery pseudoaneurysms are treated effectively by angiographic embolization.120 True extrahepatic aneurysms may be treated with embolization, provided that presence of collateral circulation, distance from the gastroduodenal artery, absence of cirrhosis, and patency of the portal vein can be confirmed, but surgical resection of the aneurysm may be preferable, to minimize the risk of hepatic infarction.121

HEPATIC ARTERY ATHEROSCLEROSIS Despite its frequency in the general population, atherosclerosis is rarely a cause of liver disease. Intimal thickening and atherosclerosis in hepatic arteries are less common and occur later in life than is typical for coronary arteries.122 Hepatic infarction resulting from atherosclerosis alone is rare. The dual blood supply to the liver undoubtedly confers protection from ischemia. Nevertheless, atherosclerosis is the primary cause of approximately one third of HAAs (see earlier).121 In addition, because the common bile duct derives all of its blood supply from the hepatic artery, atherosclerosis can result in ischemic cholangiopathy with biliary strictures and obstruction.123 The presence of athero-

sclerosis occasionally prevents the use of some donor livers for transplantation. Atherosclerosis makes arterial anastomoses technically more difficult to secure and may predispose the liver to ischemic injury during transport and reperfusion.

KEY REFERENCES

Abbas MA, Fowl RJ, Stone WM, et al. Hepatic artery aneurysm: Factors that predict complications. J Vasc Surg 2003; 38:41-5. (Ref 119.) Bearman SI. Avoiding hepatic veno-occlusive disease: What do we know and where are we going? Bone Marrow Transplant 2001; 27:1113-20. (Ref 74.) Condat B, Pessione F, Hillaire S, et al. Current outcome of portal vein thrombosis in adults: Risk and benefit of anticoagulant therapy. Gastroenterology 2001; 120:490-7. (Ref 98.) DeLeve LD, Valla DC, Garcia Tsao G; American Association for the Study of Liver Disease. Vascular disorders of the liver. Hepatology 2009; 49:1729-64. (Ref 2.) Henrion J, Schapira M, Luwaert R, et al. Hypoxic hepatitis: Clinical and hemodynamic study in 142 consecutive cases. Medicine 2003; 82:392-406. (Ref 100.) Janssen, HL, Wijnhoud A, Haagsma EB, et al. Extrahepatic portal vein thrombosis: Aetiology and determinants of survival. Gut 2001; 49:720-4. (Ref 84.) Mitchell MC, Boitnott JK, Kaufman S, et al. Budd-Chiari syndrome: Etiology, diagnosis and management. Medicine 1982; 61:199-218. (Ref 4.) Mohle-Boetani JC, Koehler JE, Berger TG, et al. Bacillary angiomatosis and bacillary peliosis in patients infected with human immunodeficiency virus: Clinical characteristics in a case-control study. Clin Infect Dis 1996; 22:794-800. (Ref 116.) Okuda K. Inferior vena cava thrombosis at its hepatic portion (obliterative hepatocavopathy). Semin Liver Dis 2002; 22:15-26. (Ref 21.) Orloff MJ, Daily PO, Orloff SL, et al. A 27 year experience with surgical treatment of Budd-Chiari syndrome. Ann Surg 2000; 232:340-52. (Ref 44.) Plessier A, Sibert A, Consigny Y, et al. Aiming at minimal invasiveness as a therapeutic strategy for Budd-Chiari syndrome. Hepatology 2006; 44:1308-16. (Ref 58.) Richman SM, Delman AJ, Grob D. Alterations in indices of liver function in congestive heart failure with particular reference to serum enzymes. Am J Med 1961; 30:211-25. (Ref 106.) Rossle M, Olschewski M, Siegerstetter V, et al. The Budd-Chiari syndrome: Outcome after treatment with the transjugular intrahepatic portosystemic shunt. Surgery 2004; 135:394-403. (Ref 36.) Valla DC. Primary Budd-Chiari syndrome. J Hepatol 2009; 50:195-203. (Ref 1.) Yoshimoto K, Ono N, Okamura T, et al. Recent progress in the diagnosis and therapy for veno-occlusive disease of the liver. Leukemia Lymphoma 2003; 44:229-34. (Ref 67.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

84 Alcoholic Liver Disease Robert L. Carithers, Jr. and Craig J. Mcclain

CHAPTER OUTLINE Epidemiology  1383 Spectrum of Disease  1383 Pathogenesis  1384 Ethanol Metabolism and Toxic Metabolites  1384 Other Metabolic Mechanisms  1385 Immune and Inflammatory Mechanisms  1387 Gender and Genetic Factors  1388 Emerging Mechanisms  1389 Fibrosis  1389 Diagnosis of Alcohol Abuse  1389 Diagnosis of Alcoholic Liver Disease  1389 History  1389 Physical Examination  1390 Laboratory Features  1390 Histopathology  1390 Conditions That May Resemble Alcoholic Liver Disease  1391 Differential Diagnosis of Clinical Deterioration  1392

Alcoholic liver disease remains a challenging enigma for basic scientists and clinicians. Despite considerable research since the 1950s, many important facets of this disease have yet to be resolved. Paramount among these important questions are the following: (1) Why does cirrhosis develop in only a small fraction of heavy alcohol abusers? (2) What is the pathogenesis of severe alcoholic liver disease? (3) What are the most effective treatments for patients with advanced alcoholic liver disease?

EPIDEMIOLOGY Although two thirds of American adults drink alcohol, only a minority are problem drinkers. Nevertheless, the number of alcoholics in the United States is estimated to be 14 million.1 The total costs of alcohol abuse amount to $185 billion annually, most of which are related to lost productivity and motor vehicle collisions. Alcoholic liver disease also is a major health care problem, accounting for 40% of deaths from cirrhosis and more than 30% of cases of hepatocellular carcinoma in the United States.1,2 In both Europe and the United States, alcoholic liver disease and its complications account for 50,000 deaths annually.3 Numerous studies have shown that alcoholic liver disease develops in women after a shorter duration of drinking and with a lower daily alcohol intake than in men.4,5 Populationbased surveys have documented that men usually must drink 40 to 80 g of alcohol daily and women 20 to 40 g daily for 10 to 12 years to achieve a significant risk of liver disease.4-6 Table 84-1 illustrates the alcohol content of various beverages, their typical serving sizes, and the daily

Cofactors That May Influence Progression of Alcoholic Liver Disease  1392 Chronic Hepatitis C  1392 Obesity and Smoking  1392 Prognosis  1393 Alcoholic Hepatitis  1393 Alcoholic Cirrhosis  1394 Treatment  1394 Abstinence and Lifestyle Modification  1394 Nutritional Support  1395 Anti-inflammatory and Anticytokine Drugs  1396 Antioxidants  1397 Drugs of Unlikely Benefit  1398 Liver Transplantation  1398 Optimal Management  1399

alcohol intake, for at least 10 years, that puts both men and women at risk for the development of alcoholic liver disease.

SPECTRUM OF DISEASE Chronic alcohol abuse can result in a spectrum of liver injury that ranges from mild fatty infiltration to cirrhosis and hepatocellular carcinoma (Fig. 84-1).7,8 Fat accumulation in liver cells, which is the earliest and most predictable response to alcohol ingestion, is seen in 90% of heavy drinkers. Although fatty liver generally is a benign condition that usually reverses quickly with abstinence, cirrhosis can develop within five years in 10% of patients who continue to drink heavily.9 Cirrhosis is particularly likely to develop if the steatosis manifests as a mixed micro- and macrovesicular pattern rather than the macrovesicular pattern seen in most alcoholics.8 Much more important than steatosis alone is the development of necroinflammation, with or without fat infiltration, and fibrosis (alcoholic hepatitis) that occurs in approximately 10% to 35% of heavy drinkers. Alcoholic hepatitis is an important clinical entity for the following two reasons: (1) patients with severe alcoholic hepatitis have extremely high short-term mortality rates, and (2) alcoholic hepatitis is a well-documented precursor of cirrhosis, with an associated long-term risk of cirrhosis that is nine times higher than that for patients with fatty liver alone.9 With continued alcohol abuse, a fine mesh-like pattern of cirrhosis with prominent involvement of the central vein (micronodular, or Laennec’s, cirrhosis) develops in 8% to 20% of heavy drinkers. Over time, and especially with continued abstinence, this lesion can evolve

1383

1384

Section IX  Liver Table 84-1  Alcohol Content of Various Beverages Daily Intake Needed to Exceed Threshold for Alcoholic Liver Disease* BEVERAGE

ALCOHOL CONTENT

SERVING SIZE

AMOUNT OF ALCOHOL

Beer Wine Fortified wine Hard liquor

5% 12% 20% 40%

12 oz 4 oz 4 oz 1.5 oz

13.85 g 10.7 g 17.8 g 13.4 g

MEN

WOMEN

3-6 cans 4-8 glasses 2-4 glasses 3-6 drinks

1.5-3 cans 2-4 glasses 1-2 glasses 1.5-3 drinks

*Alcohol intake of 40-80 g/day for men and 20-40 g/day for women for 10 years.

Pericentral fibrosis in fatty liver

Micronodular cirrhosis

Normal liver

Fatty liver

Alcoholic hepatitis

Macronodular cirrhosis with hepatocellular carcinoma Figure 84-1.  Histologic spectrum of alcoholic liver disease.

to include broad bands of fibrosis that separate large nodules of liver tissue (macronodular cirrhosis). Hepatocellular carcinoma typically develops in this setting.

PATHOGENESIS ETHANOL METABOLISM AND TOXIC METABOLITES

The liver is the main organ responsible for ethanol metabolism; other organs such as the stomach contribute to much lesser degrees. Ethanol is metabolized by three major systems in the liver: alcohol dehydrogenase (ADH), cytochrome P450 2E1 (CYP2E1), and, of least importance, catalase.10 ADH is actually a set of cytoplasmic enzymes with multiple isoforms. ADH is the primary enzyme system responsible for metabolism of ethanol at low concentrations, whereas CYP2E1 contributes at higher concentrations of ethanol (greater than 10 mM) and is induced by exposure

to ethanol. Both ADH and CYP2E1 convert ethanol to acetaldehyde, which is then converted to acetate by aldehyde dehydrogenase (ALDH). Acetaldehyde is a highly reactive and potentially toxic compound that is responsible for many of the systemic toxic effects of alcohol, such as nausea, headaches, and flushing. The “Oriental flush syndrome” results from impaired metabolism of acetaldehyde caused by inheritance of the ALDH22 allele, which encodes an inactive form of ALDH2. Persons from East Asia who are homozygous for this mutation rarely drink ethanol because they invariably experience toxic systemic effects, such as flushing and tachycardia, when they do drink. Acetaldehyde also is postulated to play an etiologic role in alcoholic liver disease. Acetaldehyde can form adducts with reactive residues on proteins or small molecules (e.g., cysteines). These chemical modifications can alter or interfere with normal biologic processes and can be directly toxic to the cell. Modified molecules also may stimulate the host’s immune response and cause autoimmune-like manifestations. Antibodies against such oxidatively modified

Chapter 84  Alcoholic Liver Disease proteins have been reported in both human and animal models of alcoholic liver disease.11 An example is the hybrid adduct of malondialdehyde and acetaldehyde (MAA), unique to alcohol exposure, which induces an immune reaction in human alcoholics and in animal models of alcoholic liver disease.11 Acetaldehyde also has been shown to impair mitochondrial glutathione transport and to sen­ sitize hepatocytes to tumor necrosis factor (TNF)-mediated killing.12 In addition to forming cytotoxic metabolites such as acetaldehyde, ethanol metabolism can alter the cellular oxidation-reduction (redox) state, thereby modulating liver injury. Specifically, the oxidation of ethanol uses nicotinamide-adenine dinucleotide (NAD+) as an electron acceptor and thereby causes a shift in the ratio of reduced NAD (NADH) to NAD+ to a more reduced state.10 This change in the redox state can impair normal carbohydrate and lipid metabolism; multiple effects ensue, including a decrease in the supply of adenosine triphosphate (ATP) to the cell and an increase in hepatic steatosis.

OTHER METABOLIC MECHANISMS Oxidative Stress

Oxidative stress is an imbalance between pro-oxidants and antioxidants. Reactive oxygen species (ROS) and reactive nitrogen species (RNS) are products of normal metabolism and can be beneficial to the host (e.g., by contributing to bacterial killing).13 Overproduction of ROS and RNS or inadequate antioxidant defenses (e.g., low levels of vitamins, selenium, mitochondrial glutathione), or both, can lead to liver injury. Oxidative stress is well documented in alcoholic liver disease.14 Studies in normal volunteers have shown that acute alcohol consumption causes a dose-related increase in urinary isoprostane levels (a marker of lipid peroxidation, which is an indirect marker of oxidative stress), and patients with alcoholic hepatitis exhibit high isoprostane levels.15 Major ROS and RNS include superoxide anion, hydrogen peroxide, and hydroxyl radical (ROS) and nitric oxide, peroxynitrite (RNS), and hypohalous acid. Oxidative stress usually is documented by detection of one of several indirect markers (1) protein oxidation (e.g., protein thiol or carbonyl products); (2) lipid oxidation (e.g., isoprostanes, malondialdehyde); (3) DNA oxidation (e.g., oxodeoxyguanosine); or (4) depletion or induction of antioxidant defenses (e.g., vitamin E, glutathione, thioredoxin). The stimulus for oxidative stress in the liver comes from multiple sources. In hepatocytes, CYP2E1 activity increases after alcohol consumption—in part because of stabilization of messenger RNA (mRNA). The CYP2E1 system leaks electrons to initiate oxidative stress.13 CYP2E1 is localized in the hepatic lobule in areas of alcohol-induced liver injury. Moreover, overexpression of CYP2E1 in mice and in HepG2 cells (a hepatocyte cell line) in vitro leads to enhanced alcohol hepatotoxicity.16,17 On the other hand, alcoholinduced liver injury still develops in CYP2E1-knockout mice. These findings suggest that increased CYP2E1 probably plays a role in alcoholic liver injury but is not the sole or dominant factor. In the CYP2E1-knockout mice, other compensatory mechanisms may be operational, such as induction of other microsomal enzymes. Nonparenchymal cells and infiltrating inflammatory cells (e.g., polymorphonuclear neutrophils) are another major source of prooxidants that are used for normal cellular processes, such as killing invading organisms. Major enzyme systems for pro-oxidant production in Kupffer cells include nicotinamide-adenine dinucleotide phosphate (NADPH) oxidase

and inducible nitric oxide synthase (iNOS). Mice deficient in NADPH oxidase or mice treated with the drug diphenyleneiodonium sulfate, which blocks NADPH oxidase, are resistant to ethanol-induced liver injury.18,19 Infiltrating neutrophils use enzyme systems such as myeloperoxidase to generate hypochlorus acid (HOCl−, a halide species that causes oxidative stress) and RNS. Oxidative stress can mediate liver injury through at least two major pathways: direct cell injury and cell signaling. Direct cell injury is indicated by markers such as lipid peroxidation and DNA damage. An even greater role is played by signaling pathways, for example, activation of tran­ scription factors such as nuclear factor kappa B (NF-κB), which plays a critical role in the production of cytokines such as TNF. The critical role of oxidative stress in the development of alcoholic liver disease has been validated in multiple studies in rats and mice fed alcohol and treated with various antioxidants, ranging from ebselen to green tea polyphenols, that overexpress both superoxide dismutase I and II. Various antioxidants have been shown to block or attenuate the development of alcoholic liver disease in rodents.13

Mitochondrial Dysfunction

Mitochondria are the major consumers of molecular oxygen and major generators of ROS in the liver. Mitochondrial dysfunction is well documented in alcoholic liver disease and contributes to oxidative stress.20,21 Mitochondrial abnormalities in alcoholic liver disease include megamitochondria observed on light and electron microscopy and functional mitochondrial abnormalities as documented by an abnormal 13C ketoacid breath test result (ketoacids are metabolized by mitochondria).22 Short-term administration of alcohol causes increased hepatic superoxide generation in liver mitochondria, with an increased flow of electrons along the respiratory electron transport chain. The increased NADH/NAD+ ratio caused by ethanol intake favors superoxide generation.13 Because hepatic mitochondria lack catalase, glutathione plays a critical role in protecting mitochondria against oxidative stress. Mitochondria do not make glutathione but instead import it from the cytosol. In alcoholic liver disease, the transport of glutathione into mitochondria is impaired, and selective mitochondrial glutathione depletion is observed.23 Glutathione depletion also sensitizes the liver to the toxic effects of TNF, and TNF also impairs mitochondrial function. Finally, an increase in mitochondrial membrane depolarization and permeabi­lity leads to hepatocyte death, especially apoptotic (programmed), rather than necrotic, cell death.

Hypoxia

The centrilobular area of the hepatic lobule (the functional unit of the liver) has the lowest oxygen tension and greatest susceptibility to hypoxia. Chronic alcohol intake increases oxygen uptake by the liver and increases the lobular oxygen gradient. A chronic intragastric feeding model in rats has been used to define the mechanisms underlying hepatic hypoxia and the association of these mechanisms with cycling of urinary alcohol levels (UALs).24 At high UALs, hepatic hypoxia is observed, along with reduced ATP levels; the NADH/NAD+ ratio is shifted to the reduced state; and the hypoxia-inducible factor (HIF) 1 and 2 genes are upregulated. When UALs fall, reperfusion injury occurs, with free radical formation and peak liver enzyme release from hepatocytes. Stimuli for this cycle of events include catecholamines (levels of which coincide with peak UALs) and modulators of vascular tone (e.g., nitric oxide,

1385

1386

Section IX  Liver endothelin-1); modulation of these stimuli offers future therapeutic possibilities.

Impaired Proteasome Function

The 26S ubiquitin-proteasome pathway is the primary proteolytic pathway of eukaryotic cells (see Chapter 72). It controls the levels of numerous proteins involved in gene regulation, cell division, and surface receptor expression, as well as stress response and inflammation. The proteasome system is now considered a cellular defense mechanism because it also removes irregular and damaged proteins generated by mutations, translational errors, or oxidative stress.25 This pathway involves two major steps: (1) covalent attachment of multiple ubiquitin molecules to the protein substrate and (2) degradation of the targeted protein by the 26S proteasome complex. The 26S ubiquitin-proteasome pathway may play a pathogenic role in the development of alcoholic liver disease.26 Early clinical studies showed that hepatomegaly caused by chronic alcohol consumption resulted in part from accumulation of protein in the liver.27 Animal studies have demonstrated that chronic ethanol feeding results in a significant decrease in proteolytic activity of the proteasome; this decreased activity can lead to abnormal protein accumulation, including accumulation of oxidized proteins.28,29 The decrease in proteasome function correlates significantly with the level of hepatic oxidative stress. Patients with alcoholic cirrhosis have increased serum ubiquitin levels, suggesting damaged proteasome function.30 Also, hepatocytes from alcoholics contain large amounts of ubiquitin in the form of cellular inclusions, or Mallory (or Mallory-Denk) bodies, which accumulate because they are not degraded efficiently by the proteasome.31 The formation of Mallory bodies, which occurs when the ubiquitin-proteasome system is inhibited or overwhelmed, is a pathway of liver injury caused by diverse toxins, including alcohol.31 As hepatocytes die as a result of proteasome inhibition, they inappropriately release cytokines such as interleukin (IL)-8 and IL-18. IL-8 recruits neutrophils and probably plays a role in neutrophil infiltration in alcoholic hepatitis, and IL-18 sustains inflammation in the liver.32

Abnormal Metabolism of Methionine, S-adenosylmethionine, and Folate

In mammals, the liver plays a central role in methionine metabolism; nearly one half of the daily intake of methionine is metabolized in the liver (Fig. 84-2). The first step in methionine metabolism is the formation of S-adenosyl­ methionine (SAM) in a reaction catalyzed by methionine adenosyltransferase (MAT). Activity of this enzyme is depressed in alcoholic liver disease.33 SAM is the principal biological methyl donor through the transmethylation pathway; the precursor of aminopropyl groups used in polyamine biosynthesis; and a precursor of glutathione through its conversion to cysteine along the transsulfuration pathway. Under normal conditions, most of the SAM generated daily is used in transmethylation reactions, in which methyl groups are added to a large number of molecules by means of specific methyltransferases.34 These compounds include DNA, RNA, biogenic amines, phospholipids, histones, and other proteins; methylation of these compounds may modulate cellular functions and integrity. In this process, SAM is converted to S-adenosylhomocysteine (SAH), which is a potent competitive inhibitor of most methyltransferases. Both an increase in SAH and a decrease in the SAM/SAH ratio are known to inhibit transmethylation reactions.34-36

Methionine a

b e

Betaine

THF

↓ SAM

Dimethyl glycine

Methyl acceptors

f

↑ SAH c

Adenosine

5-methyl-THF

↑ Homocysteine d Cysteine

g, h

GSH

Figure 84-2.  Hepatic methionine metabolism. Chronic alcohol consumption causes S-adenosylmethionine (SAM) deficiency and an increase in homocysteine and S-adenosylhomocysteine (SAH) levels. a, methionine adenosyltransferase; b, enzymes involved in transmethylation reactions, including phosphatidylethanolamine N-methyltransferase; c, SAH hydrolase; d, cystathionine B-synthase; e, betaine-homocysteine methyltrans­ ferase; f, methionine synthetase; g, glutamate-cysteine synthetase; h, glutathione (GSH) synthetase. ↑↓, effects of alcohol; THF, tetrahydrofolate.

Deficiency of SAM in patients with alcoholic liver disease was first noted in the early 1980s, when it was observed that alcoholic subjects had delayed clearance of an oral bolus of methionine (presumably because of blocked conversion of methionine to SAM).37 Functional MAT activity was subsequently shown to be subnormal in liver biopsy specimens from alcoholic subjects.38 Subnormal hepatic SAM levels also are noted in various experimental models of liver injury. Exogenous administration of SAM corrects the deficiency and attenuates the severity of these experimental forms of liver injury.33,39 Because SAM is a precursor of glutathione, SAM deficiency results in glutathione deficiency, which is observed in many forms of liver disease.40 In animal studies, exogenous SAM corrects hepatic deficiencies of both SAM and glutathione.41 Because glutathione is required for optimal expression of MAT activity in liver, hepatic deficiency of MAT may be caused in part by glutathione deficiency. Also, hepatic MAT is sensitive to oxidative stress, and the subnormal hepatic MAT activity in patients with alcoholic liver disease could result from oxidation of MAT.42 In models of alcohol-induced hepatotoxicity, SAM has been shown to maintain mitochondrial glutathione levels. Depletion of mitochondrial glutathione is thought to be one pathogenic factor in the development of alcoholic liver disease, and SAM, but not other glutathione prodrugs, prevents mitochondrial glutathione depletion in experimental alcoholic liver disease (possibly by protecting mitochondrial glutathione transport systems).43 SAM also decreases lipopolysaccharide (LPS)-stimulated TNF release and increases IL-10 release in a monocyte cell line.33 Similarly, in rats fed a diet to induce SAM deficiency, serum TNF levels increase, and sensitivity to endotoxin-induced hepatotoxicity, which can be blocked by injection of SAM, increases markedly.44 These data support the concept that SAM may have direct hepatoprotective functions and may modify LPS-stimulated cytokine production. Although serum SAM levels are decreased in patients with alcoholic liver disease, levels of the downstream pro­ ducts SAH and homocysteine are elevated. Homocysteine has been postulated to play a role in the pathogenesis of

Chapter 84  Alcoholic Liver Disease fatty liver seen with alcoholic liver disease, and in experimental animals reduction of homocysteine levels with administration of betaine (to convert homocysteine to methionine) attenuates the severity of alcoholic liver disease.45 Elevated SAH levels have been shown to sensitize hepatocytes to TNF-mediated destruction, and SAH may be a critical physiologic sensitizer of TNF-mediated killing in liver injury.36 Homocysteine and SAH can be removed by giving betaine, which facilitates regeneration of methionine from homocysteine. Folic acid also can play a critical role in the regeneration of homocysteine to methionine by means of 5-methyltetrahydrofolate (5-MTHF).46 Fatty liver develops in mice that lack the gene for MTHF reductase (MTHFR), and steatohepatitis develops in MAT1A– knockout mice; these findings further support a role for this critical pathway in the development of steatosis and steatohepatitis. Halsted and colleagues have shown that folic acid deficiency enhances the development of alcohol-induced liver injury in micropigs and that alcohol feeding interferes with normal folic acid metabolism in multiple different pathways—from impaired intestinal uptake to increased renal excretion.46 Collectively, the data support a role for altered methionine-transmethylation-transsulfuration metabolism in alcoholic liver disease and link these pathways to TNF hepatotoxicity.35

IMMUNE AND INFLAMMATORY MECHANISMS Kupffer Cell Activation and Dysregulated Cytokine Production

Cytokines are low-molecular-weight mediators of cellular communication (see Chapter 2).47 Multiple cell types in the liver are potential sources of the increased release of proinflammatory cytokines observed in alcoholic liver disease. Kupffer cells are prominent producers of pro­ inflammatory cytokines such as TNF-α, as well as certain anti-inflammatory cytokines such as IL-10. Sinusoidal endothelial cells express adhesion molecules, which modulate white blood cell adhesion and transmigration. Activated stellate cells produce collagen in response to signals such as the profibrotic cytokine transforming growth factor-b (TGF-b). Hepatocytes are a relatively newly recognized source of cytokine production, including IL-8, a major neutrophil chemotactic peptide and angiogenesis factor. Major stimuli for the observed increase in proinflammatory cytokine production in alcoholic liver disease are believed to be ROS and intestine-derived LPS.47 In alcoholic liver disease, intestinal permeability is increased and the frequency of endotoxemia is high. LPS activates the redoxsensitive transcription factor NF-κB in Kupffer cells, thereby resulting in the production of certain cytokines such as TNF (Fig. 84-3). TNF can increase intestinal permeability, induce oxidative stress, and perpetuate this cycle. Generation of ROS through the metabolism of alcohol also can activate NF-κB and stimulate proinflammatory cytokine production.48 Although necrosis has traditionally been thought to be the major mechanism of hepatocyte cell death in alcoholic liver disease, increased apoptosis also has been documented. As hepatocytes die of apoptosis, they can be taken up by Kupffer cells and stimulate TNF production.49 Moreover, when alcohol-fed rodents are treated with caspase inhibitors to block apoptosis, liver inflammation and injury are markedly attenuated.50 When hepatocytes die of proteasome inhibition-mediated apoptosis, the dying hepatocytes release IL-8 and IL-18, which cause sustained inflammation.32 Therefore, in alcoholic liver disease, hepatocyte apoptosis may sustain proinflammatory cytokine production and cell injury or death.

TNF metabolism is dysregulated in alcoholic hepatitis, as suggested by the observation that cultured monocytes (which produce substantial amounts of TNF) from patients with alcoholic hepatitis produce significantly increased amounts of TNF in response to LPS stimulation.51 Increased serum TNF concentrations in patients with alcoholic hepatitis show a strong correlation with disease severity and risk of mortality.47 Serum concentrations of TNF-inducible cytokines and chemokines, such as IL-6, IL-8, IL-18, monocyte chemoattractant protein 1 (MCP-1), and others, are elevated in patients with alcoholic hepatitis or cirrhosis, and the levels often correlate with markers of the acutephase response, reduced liver function, and poor clinical outcomes.47 This enhanced cytokine response to a physiologic stimulus such as LPS is termed priming. Increased serum or urinary levels of neopterin and other markers indicate that monocytes and Kupffer cells are primed in alcoholic liver disease. This priming for LPS-stimulated TNF production has been reproduced in vitro by culturing monocyte cell lines with relevant concentrations of alcohol. This response appears to be mediated, at least in part, by induction of CYP2E1 and oxidative stress.52 Studies in rats, mice, and tissue culture have validated a pathogenic role for cytokines (especially TNF) in the development of alcoholic liver disease.47 Rats chronically fed alcohol are more sensitive to the hepatotoxic effects of injected LPS and have much higher LPS-stimulated plasma levels of TNF in comparison with control rats. Liver injury can be attenuated by giving a prostaglandin analog to downregulate TNF production.53 Because rats have a natural aversion to alcohol, an intragastric route is often used to deliver high amounts of alcohol. Studies using the intragastric alcohol-feeding model have demonstrated that the development of liver injury coincides with an increase in TNF mRNA expression in the liver and in isolated Kupffer cells.54,55 Rats fed ethanol intragastrically also have high blood LPS levels and increased expression of CYP2E1, as well as markers of oxidative stress and lipid peroxidation. Not only are levels of proinflammatory cytotoxic cytokines increased in alcoholic liver disease, but also monocyte and Kupffer cell production of protective anti-inflammatory cytokines such as IL-10 is decreased.56 The importance of this observation for humans has been confirmed using IL10–knockout mice, in which more severe ethanol-induced hepatotoxicity develops and increased levels of proinflammatory cytokines such as TNF are seen.57 Several strategies have been devised to decrease cytokine production or activity in an attempt to block or attenuate liver injury. Examples include antibiotics to modulate intestinal flora and LPS, gadolinium chloride to destroy Kupffer cells, and antioxidants such as glutathione prodrugs to inhibit cytokine production. Each of these strategies has been successful in attenuating alcohol-induced liver injury in rats.47 Prebiotics such as oat bran also have been shown to decrease endotoxemia in experimentally induced alcoholic liver injury. Perhaps the most compelling data that relate TNF to alcohol-induced liver injury are from studies in which anti-TNF antibody has been used to prevent liver injury in alcohol-fed rats.58 Similarly, alcoholic liver injury does not develop in mice that lack the TNF type I receptor.59 Hepatocytes normally are resistant to TNF killing. Hepatocytes from rats fed alcohol or hepatocytes incubated in alcohol are sensitized to TNF killing, however.16,60 Some potentially relevant mechanisms for this sensitization include mitochondrial depletion of glutathione, accumulation of SAH, and inhibition of proteasomes, among others.

1387

1388

Section IX  Liver Lipopolysaccharide

ROS

Other mediators

Kupffer cell NF-κB Apoptotic hepatocytes

“Priming”

TNF IL-8

Tumor necrosis factor Hepatocyte

Neutrophil ROS Proteases

“Sensitization”

Cell death

Oxidative stress Lipid peroxidation products

NF-κB Other mediators

Proteasome Mitochondrial function S-adenosylmethionine

Acetaldehyde

S-adenosylhomocysteine Interleukin-8 (IL-8) Interleukin-18

Figure 84-3.  Cytokine production in alcoholic liver disease. Kupffer cells are primed to overproduce cytotoxic cytokines in response to a lipopolysaccharide stimulus, and hepatocytes are sensitized to the hepatotoxic effects of these cytokines through mechanisms such as decreased production of anti-apoptotic survival proteins. Hepatocyte death causes increased production of interleukin-8 (neutrophil chemoattractant) and interleukin-18 (which primes tumor necrosis factor release and a type 1 helper T-cell response) (see text for details). NF-κB, nuclear factor kappa B; ROS, reactive oxygen species.

Therefore, in alcoholic liver disease, monocytes and Kupffer cells are primed to increase production of TNF and to sensitize hepatocytes to TNF killing. These processes are closely intertwined with previously described mechanisms such as oxidative stress, mitochondrial dysfunction, abnormal metabolism of methionine, and dysfunction of proteasomes.

Immune Responses to Altered Hepatocellular Proteins

Alcoholic hepatitis may persist histologically for many months after exposure to ethanol has ceased, suggesting an ongoing immune or autoimmune response. Autoimmune reactions are now well documented in patients with alcoholic liver disease, with autoantibodies directed against phospholipids, alcohol dehydrogenase, heat shock protein, and other potential antigens.61 Patients with alcoholic liver disease are at increased risk for the development of immune responses directed at neoantigens generated from the interactions of metabolites of alcohol (e.g., acetaldehyde or hydroxyethyl radicals) with liver proteins.62 Studies have linked genetic susceptibility and autoimmunity in alcoholic liver disease. Some humans have a genetic mutation in the cytotoxic T lymphocyte–associated antigen 4G (CTLA-4G) allele that leads to inappropriately activated T cell function. One of the breakdown products of alcohol metabolism, the hydroxyethyl radical, can modify CYP2E1 and, in the presence of the CTLA-4G mutation, increase the risk that anti-CYP2E1 autoantibodies will develop.63 This is one

pathway whereby alcohol abuse may break “self-tolerance” in the liver.

GENDER AND GENETIC FACTORS

Because the development of liver injury varies among people who drink the same amount of alcohol, great interest exists in possible gender-associated and genetic predisposition to alcoholic liver disease. Female gender is now a well-accepted risk factor for the development and rapid progression of alcoholic liver disease.4,5 Although rates of metabolism and elimination of alcohol have been reported to be more rapid in women than in men, when adjusted for liver volume, elimination rates are similar between genders.64 Studies in rats or mice fed alcohol chronically have demonstrated that females are more susceptible than males to liver injury. Risk factors for the development of liver disease in females appear to include increased endotoxemia, lipid peroxidation, and chemokine (e.g., monocyte chemotactic protein-1) mRNA levels and activation of the critical transcription factor NF-κB.65 These risk factors are critical for determining “safe” levels of alcohol consumption in women. Many authorities consider any amount of alcohol above 20 g a day to be a risk factor for the development of liver disease in women; however, differences in levels of alcohol dehydrog­ enase in gastric mucosa between men and women are not thought to play a major role in the greater susceptibility of women to alcoholic liver injury.

Chapter 84  Alcoholic Liver Disease Genetic polymorphisms in alcohol-metabolizing systems such as CYP2E1 and ADH have been postulated to play a role in the development of alcoholic liver disease.66 None of these polymorphisms, however, adequately explains the diverse pathologic responses seen among patients with alcoholic liver disease. Polymorphisms in the promoter regions of cytokines TNF and IL-10 also have been reported to predispose affected persons to the development of alcoholic liver disease and are under active study.67,68

EMERGING MECHANISMS

Three areas deserve recognition as emerging mechanisms of liver injury. The first is the area of epigenetics, which refers to changes in gene expression that do not involve DNA coding sequence modifications. Epigenetic mechanisms typically involve DNA methylation and a variety of modifications of histones. Epigenetic modifications can be transient or long term and can be tissue or organ specific. The critical steps and enzymes (e.g., histone acetyltrans­ ferase, DNA methyltransferase) responsible for epigenetic modifications are under active investigation by multiple laboratories.69 A second area is the endoplasmic reticulum (ER) stress response, which is induced by the accumulation of unfolded or misfolded proteins. To deal with the ER stress response, cells activate a series of signaling pathways termed the unfolded protein response (UPR), which can be either protective (usually in the short term) or detrimental (usually in the long term). One of the effects of a prolonged UPR can be increased production of triglycerides and cholesterol, leading to fatty liver. Some potential inducers of the ER stress in alcoholic liver disease include elevated homocysteine levels, acetaldehyde adducts, and oxidative stress.70 A third area is the endogenous cannabinoids, which are ubiquitous lipid signaling molecules that mediate their effects by specific cannabinoid receptors, CB1 and CB2. Studies have demonstrated that inhibition of CB1 receptors can cause weight loss and attenuate fatty liver and hyperlipidemia in animal models of obesity and steatohepatitis. Moreover, CB1 blockade reduces hepatic fibrosis in a variety of animal models of cirrhosis.71

FIBROSIS

The development of hepatic fibrosis, leading to cirrhosis, indicates major progression of alcoholic liver disease and represents a maladaptive wound healing response. The development of fibrosis is a dynamic state, with constant remodeling of scar tissue; fibrosis may regress with discontinuation of exposure to alcohol. The stellate cell is the major source of collagen production in the liver. It normally exists in a quiescent state and serves as a major storehouse for vitamin A. Following activation, the stellate cell assumes a myofibroblast-like contractile phenotype and produces collagen. The cytokine TGF-b is a major stimulus for stellate cell activation and collagen production. Other cytokines implicated in activation of stellate cells include plateletderived growth factor and connective tissue growth factor (see Chapter 90). Oxidative stress plays a major role in stellate cell activation, and a variety of antioxidants can block both stellate cell activation and collagen production in vitro. Serum levels of 4-hydroxy-nonenal, a specific product of lipid peroxidation, are elevated in patients with alcoholic liver disease and up-regulate both procollagen type I and tissue inhibitor of metalloproteinase-1 (TIMP-1) gene expression. Matrix metalloproteinase-1 plays a major role in degrading type I collagen. TIMP-1 levels also are elevated in alcoholic liver disease. The result appears to be an increase in stellate cell activation and collagen production

on the one hand and a decrease in matrix degradation on the other hand.72-74

DIAGNOSIS OF ALCOHOL ABUSE The diagnosis of alcohol abuse is based on a history of heavy alcohol intake and the presence of other organ system damage or an excessive frequency of falls, lacerations, and fractures. Physicians typically identify only 50% of patients with drinking problems because of both inadequate questioning by physicians and denial by patients.75 Underdiagnosis is particularly common in older patients.76 The uniform application of screening tools such as the four-item AUDIT-C (Alcohol Use Disorders Identification Test) consumption questions and the four-item CAGE (need to cut down, annoyed by criticism, guilty about drinking, need for an eye-opener in the morning) questionnaire dramatically improves the recognition of patients with problem drinking in primary care clinics.77 Even the use of a single question: “When was the last time you had more than x drinks in one day?,” where x equals four for women and five for men, dramatically improves the diagnosis of alcoholism in primary care settings.78 Because of the reluctance of many patients to share their drinking histories candidly, continued interest has been directed toward laboratory measures that can reliably identify patients with problem drinking. Blood or breath alcohol measurements are the most sensitive and specific indicators of recent alcohol abuse, particularly among binge drinkers.79 The major limitation of these tests is the short half-life of ethanol in blood, urine, and breath. As a result, continued efforts have focused on developing biomarkers of alcohol abuse that are detectable over longer periods of time. The most specific of these biomarkers is carbohydrate-deficient transferrin (CDT).80 CDT levels increase in serum with ingestion of 50 to 80  g/day of ethanol for two to three weeks and decline gradually during abstinence, with a half-life of approximately 15 days; however, the sensitivity of this test is only 35% to 40% among alcoholics who consume 100  g of ethanol daily.80 Conditions that can influence the CDT level include gender, age, body mass index, smoking, anorexia, pregnancy, and acute trauma with blood loss.80 These issues, as well as a lack of methodologic standardization, has limited the use of CDT measurement in clinical practice. Mean corpuscular erythrocyte volume (MCV), serum gamma glutamyl transpeptidase (GGTP) levels, and the ratio of mitochondrial aspartate aminotransferase (AST) to total AST (mitochondrial AST/total AST) and combinations of these markers have been touted as more accurate measures of alcohol abuse. A mathematically formulated equation incorporating serum GGTP and CDT levels appears to have the best diagnostic accuracy of any currently available assays for chronic alcohol abuse (Fig. 84-4).81 Measurement of two alcohol metabolites, phosphatidylethanol and ethyl glucuronide, also shows promise as an innovative way to detect recent alcohol use.80

DIAGNOSIS OF ALCOHOLIC LIVER DISEASE HISTORY

Most patients with fatty liver are asymptomatic. Although patients with alcoholic hepatitis and cirrhosis may be asymptomatic, many present with a variety of complaints

1389

Section IX  Liver Table 84-2  Symptoms and Signs in Hospitalized Patients with Alcoholic Liver Disease* Patients Affected (%) SYMPTOM OR SIGN

MILD DISEASE (n = 89)

Hepatomegaly Jaundice Ascites Hepatic encephalopathy Splenomegaly Fever

MODERATE DISEASE (n = 58)

SEVERE DISEASE (n = 37)

OVERALL

94.7 100 79.3 55.2 30.9 31.0

79.4 100 86.5 70.3 39.4 21.6

86.7 60.1 57.1 44.6 26.0 22.8

84.3 17.4 30.3 27.3 18.0 18.0

*Moderate disease was defined by a serum bilirubin level >5 mg/dL, severe disease by a bilirubin level >5 mg/dL and a prolonged prothrombin time >4 seconds. Data from Mendenhall CL. Alcoholic hepatitis. Clin Gastroenterol 1981; 10:417-41.

100% Sensitivity (%)

1390

80% 60%

90% 63%

58% 45%

47%

50%

LABORATORY FEATURES

40% 20% 0% CDT

GGTP

MCV

GGTPCDT Figure 84-4.  The sensitivities of serum levels of carbohydrate-deficient transferrin (CDT), gamma glutamyl transpeptidase (GGTP), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and GGTP plus CDT in combination and the mean corpuscular volume (MCV) for detecting heavy drinkers for a population of 165 heavy drinkers consuming a mean of approximately 130 g of ethanol per day. The corresponding specificities for the above markers in this group were 98% for CDT, 99% for GGTP, 94% for MCV, 95% for AST, 87% for ALT, and 98% for GGTP plus CDT. (From Niemelä O. Biomarkers in alcoholism. Clin Chimica Acta 2007; 377:39-49, with permission.)

AST

ALT

including anorexia, nausea and vomiting, weakness, jaundice, weight loss, abdominal pain, fever, and diarrhea.82

PHYSICAL EXAMINATION

tractures. Enlarged parotid and lacrimal glands often are seen, and severely ill patients may have Muercke’s lines or white nails. Patients with hepatopulmonary syndrome often have digital clubbing (see Chapter 92).

The most extensive demographic information on alcoholic liver disease in the United States comes from studies of hospitalized patients who were assigned the diagnosis on the basis of clinical and histologic parameters.82,83 Although these studies included few asymptomatic patients, they provide a useful guide to diagnosis. The most common physical finding in patients with fatty liver and alcoholic hepatitis is hepatomegaly, which is detectable in more than 75% of patients, regardless of disease severity. Patients with alcoholic hepatitis and cirrhosis also may have hepatic tenderness, an audible bruit over the liver, spider angiomata, splenomegaly, and peripheral edema. Jaundice and ascites, which are found in approximately 60% of patients, are more frequent in patients with severe disease (Table 84-2). Various degrees of hepatic encephalopathy can be seen, usually in the most severely ill patients. Some patients with alcoholic hepatitis have a fever, with temperatures as high as 104°F, that can persist for weeks. In patients with well-compensated cirrhosis, findings on the physical examination can be normal; however, most patients have obvious hepatomegaly and splenomegaly. As the disease progresses, the liver decreases in size and has a hard and nodular consistency. Patients with decompensated cirrhosis typically have muscle wasting, ascites, spider angiomata, palmar erythema, and Dupuytren’s con-

Only one third of hospitalized patients with fatty liver have laboratory abnormalities, which usually consist of mild increases in serum AST and alanine aminotransferase (ALT) levels. As illustrated in Table 84-3, surprisingly modest elevations of serum aminotransferase levels are seen in patients with alcoholic hepatitis and cirrhosis, even when the disease is severe.82,83 Serum AST levels are almost always less than 300 to 500 U/L and typically are associated with trivial elevation of serum ALT levels, resulting in an AST/ALT ratio greater than 2, which is characteristic of alcoholic liver disease, in part because of deficiency of pyridoxal 5′ phosphate (a cofactor of aminotransferases) in alcoholic patients (see Chapter 73). Serum alkaline phosphatase levels can range from normal to values greater than 1000 U/L. Serum bilirubin levels range from normal to 20 to 40 mg/dL, and serum albumin levels may be normal or depressed to a value as low as 1.0 to 1.5 g/dL. Most patients with alcoholic liver disease are anemic and have some degree of thrombocytopenia. By contrast, the white blood cell count usually is normal or elevated, occasionally to levels consistent with a leukemoid state. Severely ill patients usually have marked prolongation of the prothrombin time—often expressed as the international normalized ratio (INR)—and often have elevated serum creatinine values.82-85

HISTOPATHOLOGY

The clinical diagnosis of alcoholic liver disease is quite sensitive and specific; therefore, liver biopsy is rarely needed to establish the diagnosis.86 A liver biopsy is essential for determining precisely the severity of hepatic injury, however, and for clarifying the diagnosis in atypical cases (see Fig. 84-1). Centrilobular and perivenular fatty infiltration in the liver is seen in most persons who drink more than 60 g of alcohol daily. Classic histologic features of alcoholic hepatitis include ballooning degeneration of hepatocytes, alcoholic hyaline (Mallory, or Mallory-Denk, bodies) within damaged hepatocytes, and a surrounding infiltrate composed of polymorphonuclear leukocytes. Most patients have moderate to severe fatty infiltration. Varying degrees of fibrosis may be present, and many patients exhibit an unusual perisinusoidal distribution of fibrosis, at times with partial or complete obliteration of the terminal hepatic venules (sclerosing hyaline necrosis). Cirrhosis can be identified by the presence of nodules of hepatic tissue that are completely surrounded by fibrous tissue.7,8

Chapter 84  Alcoholic Liver Disease Table 84-3  Laboratory Values in Hospitalized Patients with Alcoholic Liver Disease* Mean Value LABORATORY TEST

MILD DISEASE (n = 89)

Hematocrit value (%) MCV (µm3) WBC count (per mm3) Serum AST level (U/L) Serum ALT level (U/L) Serum alkaline phosphatase level (U/L) Serum bilirubin level (mg/dL) Prolongation of prothrombin time (seconds) Serum albumin level

38 100 8000 84 56 166 1.6 0.9 3.7

MODERATE DISEASE (n = 58) 36 102 11,000 124 56 276 13.5 2.4 2.7

SEVERE DISEASE (n = 37) 33 105 12,000 99 57 225 8.7 6.4 2.4

*Moderate disease was defined by a serum bilirubin level >5 mg/dL, severe disease by a bilirubin level >5 mg/dL and a prolonged prothrombin time >4 seconds. ALT, alanine aminotransferase; AST, aspartate aminotransferase; MCV, mean corpuscular volume; WBC, white blood cell. Data from Mendenhall CL. Alcoholic hepatitis. Clin Gastroenterol 1981; 10:417-41.

Alcoholic cirrhosis typically is micronodular or mixed micro- and macronodular. In patients with coexisting alcoholic hepatitis, alcoholic hyaline is almost universal, and sclerosing hyaline necrosis and moderate-to-severe fatty infiltration are common.7,8 In patients with alcoholic cirrhosis who abstain from alcohol for long periods, a frequent finding is a gradual transformation to macronodular cirrhosis, which is indistinguishable from cirrhosis caused by other forms of liver disease.

CONDITIONS THAT MAY RESEMBLE ALCOHOLIC LIVER DISEASE

Although the clinical diagnosis of alcoholic liver disease usually is quite straightforward, the similarity of clinical and histologic features of other disorders to those of alcoholic liver disease sometimes causes diagnostic confusion. The most commonly encountered conditions that have clinical or histologic features in common with alcoholic liver disease are nonalcoholic fatty liver disease (NAFLD), hereditary hemochromatosis, amiodarone hepatotoxicity, and Budd-Chiari syndrome.

Nonalcoholic Fatty Liver Disease

The condition that is most challenging to differentiate from alcoholic liver disease is NAFLD. The two conditions are histologically indistinguishable. As a consequence, the differentiation between alcoholic liver disease and NAFLD has to be made on clinical grounds. The strongest evidence in support of a diagnosis of NAFLD rather than alcoholic liver disease is a history of daily alcohol intake less than 20 g/ day. When a patient’s alcohol intake is questionable, differentiating the two conditions can be difficult, if not impossible. Patients with NAFLD are more likely than patients with alcoholic liver disease to be asymptomatic and often have peripheral insulin resistance, obesity, hypertension, and dyslipidemia.87,88 A model that incorporates the MCV, AST/ALT ratio, body mass index, and gender shows promise in more clearly differentiating patients with alcoholic liver disease from those with NAFLD.89 The serum CDT level can be useful for distinguishing heavy drinkers from abstinent patients with NAFLD; however, the accuracy of this test for detecting moderate but clinically significant levels of alcohol intake is less clear (see earlier and Chapter 85).80

Hereditary Hemochromatosis

On occasion, distinguishing patients with alcoholic liver disease and secondary iron overload from those with liver

disease caused by hereditary hemochromatosis, particularly those with decompensated cirrhosis, can be difficult. Patients with end-stage liver disease from alcoholic cirrhosis may have elevated serum iron and ferritin levels and increased hepatic iron levels suggestive of hereditary hemochromatosis.90 In fact, more than 20% of patients with endstage alcoholic cirrhosis have clinically important hepatic siderosis.91 To complicate matters further, 15% to 40% of patients with hereditary hemochromatosis consume more than 80 g of alcohol daily.92 The overlapping clinical features of hereditary hemochromatosis and alcoholic liver disease include hepatomegaly, testicular atrophy, cardiomyopathy, and glucose intolerance. Testing for mutations in the gene for hereditary hemochromatosis, HFE, is the best method for differentiating the two conditions among whites. Few patients with alcoholic cirrhosis and iron overload are homozygous for C282Y or heterozygous for the C282Y and H63D HFE genes, whereas some have a hepatic iron index value greater than 1.9 that might otherwise suggest hereditary hemochromatosis (see Chapter 74).90,91

Amiodarone Hepatotoxicity

Much less common and less difficult than NAFLD to distinguish from alcoholic liver disease is amiodarone hepatotoxicity. Although the hepatic histologic features of this condition may be similar to those of alcoholic hepatitis with or without cirrhosis, the clinical setting usually distinguishes amiodarone hepatotoxicity from alcoholic liver disease (see Chapter 86).8,93

Budd-Chiari Syndrome

Occasional patients with severe alcoholic liver disease can be misdiagnosed as having acute Budd-Chiari syndrome (hepatic vein thrombosis) on the basis of rapid clinical deterioration, marked hepatomegaly, caudate lobe hypertrophy, and failure to visualize the hepatic veins by Doppler ultrasonography.94 Careful evaluation of these patients usually reveals clinical and biochemical features typical of severe alcoholic hepatitis. Patent hepatic veins usually can be demonstrated by venography. Liver biopsy is particularly useful in distinguishing the characteristic histologic features of alcoholic liver disease from those of Budd-Chiari syndrome. Failure to recognize alcoholic hepatitis as the underlying cause of the liver disease before initiating anticoagulation or performing portacaval shunt surgery can result in high mortality rates (see Chapter 83).94

1391

Section IX  Liver DIFFERENTIAL DIAGNOSIS OF CLINICAL DETERIORATION

Occasionally, determining why a patient with stable, wellcompensated alcoholic liver disease has suddenly and inexplicably deteriorated may be difficult.

Acetaminophen Hepatotoxicity

The most common cause of severe drug-induced liver injury encountered in the United States is acetaminophen hepatotoxicity (see Chapter 86). Two clinical patterns of liver injury have been identified: (1) suicidal or accidental ingestion of large quantities of acetaminophen sufficient to cause hepatic injury and (2) ingestion of lesser quantities of acetaminophen by patients predisposed to injury because of up-regulation of the hepatic enzymes that convert acetaminophen to a hepatotoxic metabolite. The latter type of toxicity is seen most commonly in chronic alcoholics who take excessive acetaminophen over a period of days to weeks for relief of a headache, toothache, or other minor pain.95,96 The clinical features in these patients are indistinguishable from those of alcoholic liver disease, with one obvious exception: AST values are typically more than 1000 U/L, much higher than those in patients with alcoholic liver disease. Because liver injury typically has occurred by the time of hospita­ lization, acetaminophen levels are not helpful for diagnosis or management. Recognition of the cause of the unusually elevated serum aminotransferase levels comes from careful questioning of the patient and family about acetaminophen ingestion in the days to weeks before hospitalization. The morbidity and mortality associated with this condition are considerable.95,96 Because many of these patients have a history of recent heavy alcohol use, few are candidates for liver transplantation.

CHRONIC HEPATITIS C

The cofactor that influences progression of alcoholic liver disease most profoundly is HCV infection. Between one fourth and one third of patients with alcoholic liver disease have serologic or virologic evidence (or both) of HCV infection.100 The prevalence of HCV infection is highest in patients who have used injection drugs; however, the risk is high even among those who deny drug use. Histologic features of focal lymphoid aggregates, portal inflammation, and periportal or bridging fibrosis are common in liver biopsy specimens from alcoholics with HCV infection.101 Of greater importance, liver disease is more severe, advanced disease develops at a younger age, and survival is shorter in patients with both alcoholic liver disease and HCV infection than in patients with alcoholic liver disease and no evidence of HCV infection.100 In one of the more striking examples of the interaction between alcohol abuse and hepatitis C, Corrao and colleagues found that the .relative risk of cirrhosis was 10-fold higher among heavy drinkers with chronic hepatitis C than among those who had no evidence of HCV infection (Fig. 84-5).102 In addition, alcohol and HCV act synergistically in the development of hepatocellular carcinoma (see Chapter 79).103-105

OBESITY AND SMOKING

The risk of liver disease is two to three times higher in drinkers who are obese than in drinkers who have a normal body mass index.106 Although an increased risk of fatty liver is not surprising in obese persons (see Chapter 85), obesity also appears to be an independent risk factor for both alcoholic hepatitis and cirrhosis.106,107 Cigarette smoking also has been shown to accelerate the progression of fibrosis in patients with alcoholic liver disease,108,109 and smoking

Acute Viral Illness

Patients with alcoholic cirrhosis are vulnerable to decompensation from a variety of viral illnesses. Acute viral hepatitis can result in the sudden onset of liver failure, with extremely high mortality rates (see Chapters 77 through 81).97,98 Sudden decompensation and liver failure also have been reported during infection with influenza A virus.99 These dramatic cases illustrate the importance of routine immunization against hepatitis A and B and influenza in patients with alcoholic cirrhosis.

Hepatocellular Carcinoma

Occasional patients with alcoholic cirrhosis who have been abstinent for many years decompensate suddenly, with the abrupt onset of hepatic encephalopathy, variceal bleeding, or ascites. Not infrequently, the underlying cause is hepatocellular carcinoma. Unfortunately, the sudden onset of symptoms frequently results from tumor invasion of the portal or hepatic veins; as a result, the prognosis for these patients is dismal.3 The risk of hepatocellular carcinoma in patients with alcoholic cirrhosis underscores the need for surveillance for this neoplasm in these patients, especially those who abstain from alcohol and in whom the long-term prognosis is otherwise good (see Chapter 94).

160 140

HCV negative HCV positive

120 Odds ratio for cirrhosis

1392

100 80 60 40 20 0 None

COFACTORS THAT MAY INFLUENCE PROGRESSION OF ALCOHOLIC LIVER DISEASE A number of factors have been reported to have an adverse effect on the progression of liver disease in chronic alcoholics. The most important of these factors are chronic hepatitis C virus (HCV) infection, obesity, and smoking.

25 or 50

75 or 100 125 or 150

>175

Average lifetime daily alcohol intake (g) Figure 84-5.  Odds ratio for developing cirrhosis in patients who chronically drink varying amounts of alcohol based on the presence or absence of hepatitis C virus (HCV) infection. (Data from Corrao G, Lepore AR, Torchio P, et al. The effect of drinking coffee and smoking cigarettes on the risk of cirrhosis associated with alcohol consumption. A case-control study. Provincial Group for the Study of Chronic Liver Disease. Eur J Epidemiol 1994; 10:657-64.)

Chapter 84  Alcoholic Liver Disease appears to accelerate disease progression in patients with HCV infection who drink heavily.110

PROGNOSIS The prognosis for an individual patient with alcoholic liver disease depends on the degree of pathologic injury, patient’s nutritional status, presence of complications of advanced liver disease, presence of other comorbid conditions such as HCV infection, and patient’s ability to discontinue destructive patterns of drinking. In studies that have examined the natural history of alcoholic liver disease on the basis of histologic characteristics at diagnosis, patients with fatty liver have had the best outcome (70% to 80% survival rate at four to five years); those with alcoholic hepatitis or cirrhosis have had an intermediate outcome (50% to 75% survival rate at four to five years); and those with cirrhosis combined with alcoholic hepatitis have had the worst outcome (30% to 50% survival rate at four to five years) (Fig. 84-6).111 Among all patients with alcoholic liver disease, the average one-year and five-year survival rates are approximately 80% and 50%, respectively.80 Alcoholic cirrhosis also appears to be an independent risk factor for hepatocellular carcinoma.2,3,100-102 Among alcoholics, men older than 50 years of age appear to be most vulnerable to the development of hepatocellular carcinoma (see Chapter 94).112 Estimating the prognosis of patients with alcoholic liver disease is particularly important for determining the need for (1) specific therapy in a patient with severe alcoholic hepatitis and (2) liver transplantation in a patient with alcoholic cirrhosis.

ALCOHOLIC HEPATITIS

The prognosis among patients with alcoholic hepatitis can vary dramatically. In patients with severe disease, the mortality rate is high, approaching that for patients with fulminant hepatic failure. Clinical features associated with severe disease include hepatic encephalopathy, marked prolongation of the prothrombin time, elevation of the serum bilirubin level above 25 mg/dL, depression of the serum albumin level, an elevated serum creatinine level, and older age. Other important prognostic variables in patients with severe alcoholic hepatitis are spontaneous hepatic encephalopathy and hepatorenal syndrome (Fig. 84-7).79,81,82 The one-month mortality rate in patients with spontaneous hepatic encephalopathy is approximately 50%, and the rate in those with hepatorenal syndrome is 75%.79,81,82,113-115 Three models have been shown to predict short-term prognosis in these often critically ill patients. Maddrey and Boitnott discovered a simple formula they called the discriminant function (DF), calculated as [4.6 × prothrombin time − control value (seconds)] + serum bilirubin (mg/dL). The DF has proved useful for identifying patients with poor short-term survival rates.113 Three prospective studies have demonstrated that patients with a DF value of 32 or more have a poor prognosis, with one-month mortality rates of 35% to 45%.82,114,115 By contrast, patients with a DF value less than 32 have short-term survival rates of 90% to 100%.113,116 The Model for End-stage Liver Disease (MELD) score (which includes the serum bilirubin level, INR, and serum creatinine level) and the Glasgow alcoholic hepatitis score (which includes age, WBC count, blood urea nitrogen level, prothrombin time ratio [ratio of the patient’s prothrombin time to the control value], and serum bilirubin level) also have been shown to predict survival in patients with severe alcoholic hepatitis.117,118 Because both the MELD and Glasgow scores include measures of renal function, they appear to be more accurate than the DF in determining the prognosis of patients with concomitant kidney injury.

1.0

80

Mortality rate at day 28 (%)

Probability of survival

0.9

0.8

0.7

0.6

0.5 40

80

120

160

200

240

60

40

20

280

Weeks No cirrhosis, no hepatitis n = 58 Cirrhosis, no hepatitis n = 42 No cirrhosis, hepatitis n = 19 Cirrhosis, hepatitis n = 98 Figure 84-6.  Survival of patients with alcoholic liver disease stratified by histologic severity of disease. (From Orrego H, Black JE, Blendis LM, Medline A. Prognosis of alcoholic cirrhosis in the presence or absence of alcoholic hepatitis. Gastroenterology 1987; 92:208-14, with permission.)

0 DF > 32

Encephalopathy

Hepatorenal syndrome

Clinical features of severe alcoholic hepatitis Figure 84-7.  Mortality rate at day 28 in patients with severe alcoholic hepatitis who received no treatment and who had a discriminant function (DF) score greater than 32, spontaneous hepatic encephalopathy, or hepatorenal syndrome. See text for calculation of DF. (Data from references 84, 85, and 114.)

1393

Section IX  Liver 1.0

70 60 Probability of survival

0.8 5-year survival rate (%)

1394

50 40 30 20

Abstinence 0.6 Relapsing 0.4 Excessive 0.2

10

Unknown

0 0

0 5–7

8–10

11–15

Child-Turcotte-Pugh score Figure 84-8.  Five-year survival rates in patients with alcoholic cirrhosis according to their Child-Turcotte-Pugh scores. (Data from Poynard T, Naveau S, Doffoel M, et al. Evaluation of efficacy of liver transplantation in alcoholic cirrhosis using matched and simulated controls: Five-year survival. Multi-centre group. J Hepatol 1999; 30:1130-7.)

120 240 360 480 600 720 840 960 1080 Survival duration in days

Figure 84-9.  Survival curves during the three years following hospital discharge according to alcohol consumption: abstinence: patients who were abstinent; relapsing: patients with one or more periods of abstinence alternating with one or more periods of excessive consumption; excessive: patients with excessive consumption of alcohol at the first follow-up point. Survival differed significantly between abstinent and excessively drinking patients (P < .001). (Modified with permission of Veldt BJ, Laine F, Guillygomarc’h A, et al. Indication of liver transplantation in severe alcoholic liver cirrhosis: Qualitative evaluation and optimal timing. J Hepatol 2002; 36:93-8.)

ALCOHOLIC CIRRHOSIS

The clinical tool used most widely to determine prognosis in patients with alcoholic cirrhosis is the Child-TurcottePugh (CTP) score and Child (or Child-Pugh) classification (see Chapter 90). This simple classification system, which was designed specifically to assess the risk of mortality following portacaval shunt surgery in cirrhotic patients with variceal bleeding, has gained favor as a rapid method for determining the prognosis of patients with various chronic liver diseases. The Child classification is as effective as quantitative liver function tests (see Chapter 73) and disease-specific prognostic models for determining shortterm prognosis in groups of patients awaiting liver transplantation (see Chapter 95).119 Despite its limitations, the Child classification has been adopted widely for riskstratifying patients with cirrhosis because of its simplicity and ease of use. Five-year survival rates for patients with alcoholic cirrhosis decrease dramatically as the CTP score and Child’s class become higher at the time of clinical presentation (Fig. 84-8).117 The development of ascites, variceal bleeding, hepatic encephalopathy, spontaneous bacterial peritonitis, or hepatorenal syndrome also has a significant impact on the prognosis of patients with alcoholic cirrhosis. The five-year survival rate for persons in whom any of these complications develop is only 20% to 50% of that for patients with compensated cirrhosis.120 The most ominous complications are spontaneous bacterial peritonitis and rapid-onset hepatorenal syndrome (see Chapters 91 and 92). Fewer than half of the patients in whom spontaneous bacterial peritonitis develops can be expected to survive one year; the median survival of patients with hepatorenal syndrome is less than two weeks.121,122 Other models that have been used to predict prognosis in patients with alcoholic cirrhosis are the Beclere model, a proportional hazards model developed by Poynard and colleagues, and the MELD score.119,123 The Beclere model,

which was developed from a database of 818 patients with alcoholic cirrhosis who were followed prospectively for four years, includes the serum bilirubin level, serum albumin level, patient’s age, and presence or absence of hepatic encephalopathy.119 The MELD model was developed at the Mayo Clinic to assess short-term prognosis in patients undergoing transjugular intrahepatic portosystemic shunt placement.123 This model has been shown to be useful for predicting short-term survival in groups of patients with various liver diseases (see Chapter 90). Abstinence from continued excessive drinking is the most important predictor of survival in patients who survive an initial hospitalization for alcoholic cirrhosis.124 The rate of survival over the ensuing two years is 70% to 80% among patients who abstain or dramatically reduce their excessive drinking, compared with only 20% to 30% in those who continue to drink heavily (Fig. 84-9).124

TREATMENT ABSTINENCE AND LIFESTYLE MODIFICATION

Virtually every study of abstinence in alcoholic liver disease shows beneficial effects on patient survival, even in patients with decompensated cirrhosis (see Fig. 84-9). Reducing but not completely stopping alcohol consumption also has been shown to improve projected survival in patients with alcoholic liver disease.125,126 Heavy drinkers who receive socalled brief interventions, which are less than 1 hour in duration and incorporate simple motivational counseling techniques, are twice as likely as control patients to moderate or stop their drinking 6 to 12 months later.126 Regular meetings with a nurse or other health care professional to

Chapter 84  Alcoholic Liver Disease

NUTRITIONAL SUPPORT

Nutritional abnormalities are pervasive in alcoholics. Two divergent patterns are common: obesity and malnutrition. Persons who combine alcohol abuse with a high calorie diet frequently develop truncal obesity, which can accelerate the progression of the underlying alcoholic liver disease.41,42,125 The high frequency of obesity is not surprising given the high caloric content (7.1 kcal/g) of alcohol. By contrast, malnutrition is a widespread clinical problem among patients with alcoholic liver disease when a substantial proportion of nutrient-rich dietary calories are replaced with alcohol.131 Every patient with moderate to severe alcoholic hepatitis or cirrhosis shows some signs of malnutrition, with up to 50% of total energy intake derived from alcohol.132 The frequency of malnutrition increases dramatically with the severity of liver disease.128 For example, the frequency of profound malnutrition increases from 20% in patients with Child’s class A cirrhosis to 60% in those with Child’s C cirrhosis.127 Furthermore, a strong association exists between protein-calorie malnutrition and complications of alcoholic disease such as infections, encephalopathy, ascites, and variceal bleeding.131,132 Levels of folate, vitamin B6, thiamine, and vitamin A, as well as those of trace elements such as selenium, zinc, copper, and magnesium, are often severely reduced in patients with alcoholic liver disease.125 The mechanisms underlying the profound malnutrition frequently observed in patients with moderate to severe alcoholic liver disease are complex and multifactorial and include (1) anorexia induced by increased proinflammatory cytokines such as TNF and leptin; (2) intestinal fat and protein malabsorption; and (3) a catabolic state that promotes gluconeogenesis from skeletal and visceral proteins.131-133 Evaluating malnutrition in patients with liver disease can be difficult because the tests most commonly used to assess nutritional status (e.g., serum albumin concentration, anthropometry, immune status) often are affected by the liver disease. The creatinineheight index appears to be the most reliable indicator of loss of muscle mass in patients with moderate-to-severe liver disease.133,134

100 90 80 Percent surviving

emphasize abstinence and adherence to medication appear to be more effective than intense counseling by alcohol treatment specialists.124,127 Abstinence invariably causes resolution of hepatic steatosis. An additional goal of abstinence is to prevent ongoing injury, fibrosis, and the possible development of hepatocellular carcinoma, but few studies have addressed the effects of abstinence on disease progression. Short-term treatment with naltrexone, an opioid antagonist, decreases the chance of relapse in one third of heavy drinkers; however, this agent is contraindicated in patients with liver disease because of its extensive hepatic metabolism and potential hepatotoxicity.127,129 Baclofen, a gamma aminobutyric acid (GABA) B-receptor agonist, shows promise as the first safe and effective agent to improve abstinence and decrease the likelihood of relapse in patients with alcoholic cirrhosis.130 As discussed earlier, obesity, which is increasing in frequency among alcoholics as well as the general population, is associated with the development of fatty liver, steatohepatitis, and cirrhosis and appears to be a major risk factor for progression of alcoholic liver disease (see Chapter 85).41,42 The majority of alcoholics smoke cigarettes, another risk factor for more severe alcoholic liver disease. Therefore, lifestyle modifications including significant reduction or cessation of alcohol consumption, weight control, and elimination of cigarette smoking, are important initial approaches to the treatment of alcoholic liver disease.

70

TEN: 62%

60 50 40 P = 0.028

30

GC: 39%

20 10 0 0

30 60 90 120 150 180 210 240 270 300 330 360 Days

Figure 84-10.  Probability of survival for one year after randomization of 72 patients to total enteral nutrition (TEN) or to glucocorticoid therapy (GC). (From Cabre E, Rodriguez-Iglesias P, Caballeria J, et al. Short- and longterm outcome of severe alcohol-induced hepatitis treated with steroids or enteral nutrition: A multicenter randomized trial. Hepatology 2000; 32:3642, with permission.)

Treating nutritional abnormalities in alcoholics can be quite challenging. The management of obesity combines abstinence with dietary restriction and regular exercise— goals that are difficult to achieve in any group of patients, especially alcoholics. Malnutrition can be equally difficult to overcome. Two thirds of hospitalized patients provided with and encouraged to consume a balanced 2500-kcal diet with careful monitoring by a dietitian are unable to meet this objective. These difficulties have stimulated research into various forms of nutritional support for patients with severe alcoholic hepatitis and cirrhosis.135-138 In the first large study of intensive nutritional support in patients with moderate to severe alcoholic hepatitis, Mendenhall and colleagues demonstrated increased survival in patients provided aggressive nutritional support.132 In a pivotal multicenter study by Cabre and coworkers, patients were randomized to receive prednisone 40 mg daily (see later) or a liverspecific formula containing 2000 calories per day through a feeding tube.136 The one-month mortality rates were similar in both groups, but the one-year mortality rate was significantly lower in the patients who received the enteral nutrition, in great part because of reduced infectious complications, in comparison with patients who received glucocorticoids (Fig. 84-10). This study clearly demonstrates the important role of enteral nutrition in hospitalized patients with severe alcoholic liver disease. By contrast, no study has, to date, demonstrated a survival benefit for treatment with parenteral nutrition.131 Tube feeding in patients with alcoholic liver disease probably is underused in most hospitals because of concerns about precipitating hepatic encephalopathy or stimulating bleeding from esophageal varices, neither of which has been documented. Most patients probably can tolerate standard enteral products, and only selected patients with overt hepatic encephalopathy require liver-specific products rich in branched-chain amino acids. Studies of nutritional support in outpatients are limited. Hirsch and colleagues demonstrated that patients attending

1395

Section IX  Liver an outpatient liver clinic who took an enteral nutritional support product that contained 1000 kcal and 34 g of protein had significantly improved protein intake and fewer hospitalizations in comparison with those not receiving the supplement.137 Cirrhotics have decreased hepatic glycogen stores and an accelerated metabolic reaction to short-term fasting. Studies have shown that nighttime snacks (710 kcal = 2 cans of Ensure Plus) attenuate this catabolic response and significantly improve lean tissue over a one-year period of supplementation.138 Therefore, nutritional support can improve nutritional status and, in some patients, may enhance liver function and decrease the risk of death. Assessment of nutritional status and nutritional supplementation should be pursued aggressively in both inpatients and outpatients with alcoholic liver disease, especially those with moderate-to-severe alcoholic hepatitis and cirrhosis.

ANTI-INFLAMMATORY AND ANTICYTOKINE DRUGS Glucocorticoids

Glucocorticoid therapy has been the most extensively studied and the most controversial treatment for patients with alcoholic hepatitis. A total of 10 small, single-center, placebo-controlled randomized trials of glucocorticoid therapy were published from 1971 to 1984, and only two showed a benefit.114 Helman and colleagues demonstrated improved survival only in patients who had hepatic encephalopathy within the first 10 days after hospital admission.17 Maddrey and colleagues confirmed the prognostic importance of encephalopathy and found that a DF value greater than 32 (see earlier) was as effective as detecting encephalopathy in selecting patients at high risk for early mortality and that these patients appeared to benefit from gluco­ corticoid therapy.113 These two prognostic tools—hepatic encephalopathy and an elevated DF—were used to select patients for entry into a subsequent multicenter study that demonstrated a dramatic improvement in short-term survival with glucocorticoid therapy.114 The cumulative 28-day mortality rate for this severely ill group of patients was 35% in the placebo recipients, compared with only 6% in patients who received methylprednisolone. Using the same selection criteria for study entry, Ramond and colleagues confirmed the improvement in short-term survival and also demonstrated a continued survival benefit for up to six months after treatment with glucocorticoids (Fig. 84-11).115 Additional follow-up of these patients revealed that the survival benefit of glucocorticoid therapy persisted for one but not two years after treatment.133 No major complications were associated with glucocorticoid therapy in these studies.114,115 These two clinical trials included only patients with severe disease; patients with gastrointestinal bleeding requiring transfusions and active infection were excluded. Furthermore, none of the patients had evidence of hepatorenal syndrome before entry into the studies. Glucocorticoids should not be used in patients with mild alcoholic hepatitis; however, a short course of glucocorticoids (e.g., prednisone, 40 mg daily for 28 days, followed by 20 mg daily for 7 days and 10 mg daily for 7 days) may be beneficial in patients with severe disease. The DF, MELD score, and Glasgow index can each be used to select patients for treatment. Glucocorticoids should not be used in patients with gastrointestinal bleeding requiring blood transfusions or with evidence of active infection and probably are not effective in patients with hepatorenal syndrome.19,139 The response to glucocorticoids can be determined within seven

100

Glucocorticoid

Percent surviving

1396

50

Placebo

0 90

180

Day Figure 84-11.  Survival in 61 patients with alcoholic hepatitis randomly assigned to receive glucocorticoid therapy or placebo. Survival rates at six months were 84% in the glucocorticoid treatment group and 45% in the placebo group (P = .002). (From Raymond MJ, Poynard T, Rueff B, et al. A randomized trial of prednisone in patients with severe alcoholic hepatitis. N Engl J Med 1992; 326:507, with permission.)

days (primarily by a reduction in serum bilirubin levels), and treatment can be discontinued at that time if a response has not been achieved.140 Figure 84-12 illustrates the factors that should be taken into account when glucocorticoid therapy is considered in patients with severe alcoholic hepatitis.

Pentoxifylline

Pentoxifylline is a nonselective phosphodiesterase inhibitor that increases intracellular concentrations of adenosine 3′,5′-cyclic monophosphate (cAMP) and guanosine 3′,5′cyclic monophosphate (cGMP) and may thereby inhibit TNF production. Pentoxifylline also has been shown to decrease gene transcription and to affect multiple steps in the cytokine/chemokine inflammatory pathway, either directly or indirectly by inhibiting TNF.141 Selected effects of pentoxifylline include inhibition of cytokine/chemokine synthesis (e.g., MCP-1, IL-8, macrophage inflammatory protein [MIP]-1α and MIP-1b), decreased expression of adhesion molecules on endothelial cells, decreased activation of neutrophils, decreased proliferation of lymphocytes and monocytes, and decreased binding and transmigration of leukocytes. Pentoxifylline also reduces fibroblast proliferation and secretion of collagen and other interstitial matrix proteins. Akriviadis and colleagues performed a prospective, randomized, double-blind clinical trial of pentoxifylline in patients with severe alcoholic hepatitis (DF greater than 32).20 Forty-nine patients received pentoxifylline, 400 mg orally three times daily, and 52 received placebo (vitamin B12) for four weeks. Only 12 patients (24.5%) who received pentoxifylline died, compared with 24 (46%) who received placebo (Fig. 84-13). Pentoxifylline therapy was associated with a significant decrease in the frequency of hepatorenal syndrome as a cause of death and was well tolerated with

Chapter 84  Alcoholic Liver Disease no major side effects. On the basis of this single trial, pen­ toxifylline appears to be a viable alternative to glucocorticoids, particularly in patients with clinically important renal dysfunction.

Alcoholic hepatitis

Hepatic encephalopathy or at least one of the following: DF* ≥ 32, MELD† score ≥ 18, or Glasgow score‡ ≥ 9 Yes No Poor prognosis: consider specific therapy

Glucocorticoids Followed by Pentoxifylline

Good prognosis: nutritional support and conservative management

Louvet and his colleagues in France explored the possibility of switching patients to pentoxifylline if they failed to demonstrate a response to the first seven days of corticosteroid therapy.142 Unfortunately, these patients did not obtain any benefit from the early switch. Therefore, the optimal approach to the management of nonresponders to glucocorticoid therapy remains unresolved.

Specific Anti-Tumor Necrosis Factor Therapy

Active gastrointestinal bleeding, systemic infection, or renal insufficiency Yes

No

Pentoxifylline 400 mg three times daily for 28 days

Prednisone 40 mg for 7 days. If serum bilirubin level decreases, continue prednisone 40 mg daily for an additional 21 days, followed by a two-week taper. If bilirubin level does not decrease, stop treatment after 7 days.

Figure 84-12.  Algorithm for the management of patients with alcoholic hepatitis. *The DF is calculated as follows: 4.6 (prothrombin time of patient − prothrombin time of control) + serum bilirubin level (in mg/dL). † The Model for End-stage Liver Disease (MELD) score is based on the serum bilirubin level, international normalized ratio, and serum creatinine level (see Chapter 90). ‡ The Glasgow alcoholic hepatitis score is based on the patient’s age, white blood cell count, blood urea nitrogen level, ratio of prothrombin time to a control value, and serum bilirubin level. DF, discriminant function.

1.0 0.9

Probability of survival

0.8 0.7

Dysregulated cytokine metabolism was described in alcoholic hepatitis long before it was recognized in inflammatory bowel disease and rheumatoid arthritis. An initial concern in alcoholic liver disease arose from early observations that low (“basal”) amounts of TNF were important for liver regeneration.143 Therefore, many investigators suggested that down-regulating, without totally blocking, TNF activity would be a preferred therapeutic intervention. Indeed, many therapies used in alcoholic liver disease (e.g., glucocorticoids, pentoxifylline, S-adenosylmethionine) decrease but do not abolish TNF activity. Because therapy with anti-TNF antibodies has been shown to block development of alcohol-induced liver injury in rats, it was initially studied in small clinical trials in patients with alcoholic hepatitis, with apparent success.144 A large, double-blind, randomized controlled trial in France in which patients with acute alcoholic hepatitis were treated with prednisolone or prednisolone plus high-dose infliximab was terminated, however, because of an increased rate of infectious complications in the patients who received combined therapy.145 Etanercept was postulated to be more appropriate than infliximab in patients with alcoholic liver disease because of its shorter duration of action, but a National Institutes of Health (NIH)-sponsored multicenter trial reported similar one-month mortality rates and significantly worse six-month mortality rates in patients with moderate-to-severe alcoholic hepatitis treated with etanercept compared with those treated with placebo.146 Therefore, at the present time, agents that inhibit, but do not totally block, inflammatory mediators such as TNF appear to be preferable to those that totally block these mediators in patients with alcoholic hepatitis.

0.6

ANTIOXIDANTS S-adenosylmethionine

0.5 0.4 0.3 0.2 0.1 0 0

20

40

60

80

100

120

140

160

Time (days) Figure 84-13.  Probability of survival in 101 patients with alcoholic hepatitis treated with pentoxifylline (red line) or placebo (blue line) (P = .037). (From Akriviadis E, Botla R, Briggs W, et al. Pentoxifylline improves short-term survival in severe acute alcoholic hepatitis: A double blind, placebocontrolled trial. Gastroenterology 2000; 119:1637-48, with permission.)

As discussed earlier, alcoholic liver disease is characterized by elevated plasma methionine concentrations and decreased clearance of intravenously or orally administered methionine. Decreased MAT activity results in decreased plasma (and presumably intrahepatic) SAM levels.147 Administration of SAM has been reported to protect against experimental liver injury caused by alcohol, acetaminophen, carbon tetrachloride, and galactosamine.69,75 Theoretical benefits of SAM in alcoholic liver disease include its roles as an antioxidant and a critical methyl donor and its actions in maintaining mitochondrial function, decreasing TNF levels, and producing glutathione. A multicenter clinical study reported that SAM in a daily dose of 1200 mg significantly reduced the mortality rate and decreased the need for liver transplantation in patients with Child’s A and B alcoholic cirrhosis.148

1397

1398

Section IX  Liver Silymarin

Silymarin, the active ingredient extracted from Silybum marianum (also known as milk thistle), has been shown in experimental animals to protect against various hepatotoxins, including carbon tetrachloride, acetaminophen, iron (in iron overload), and poisonous mushrooms (see Chapter 87).149 It has antioxidant properties, protects against lipid peroxidation, and exerts anti-inflammatory and antifibrotic effects. Despite these properties, insufficient data are available from well-conducted clinical trials to demonstrate improvement in mortality, complications of cirrhosis, or histology in patients with alcoholic liver disease.150,151 Nevertheless, silymarin has become the most popular form of complementary and alternative medicine therapy for patients with liver disease because of its good safety profile, and ongoing NIH studies should answer questions concerning its efficacy (see Chapter 127).

Vitamin E

Vitamin E deficiency has been well documented in patients with alcoholic liver disease.152 Vitamin E has hepatoprotective effects in experimental liver injury, with potentially beneficial effects that include membrane stabilization, reduced NF-κB activation and TNF production, and inhibition of hepatic stellate cell activation and collagen production.152-154 Unfortunately, the largest randomized study of vitamin E supplementation in patients with alcoholic liver disease did not show a significant benefit, possibly because a relatively low dose was used.154

Combination Antioxidant Therapy

Oxidative stress plays an etiologic role in the development of alcoholic liver disease (see earlier), and antioxidants block the development of alcoholic liver disease in experimental animals. An initial study of a combination of antioxidants in patients with alcoholic hepatitis reported beneficial effects,155 whereas a subsequent trial in which glucocorticoids were compared with an antioxidant cocktail was stopped after an interim analysis found a significant benefit in the glucocorticoid-treated group.156 These inconsistent results led to a third trial in which patients with acute alcoholic hepatitis were randomized to antioxidant therapy alone or with glucocorticoids; neither treatment improved six-month survival.157 None of these studies evaluated whether or not antioxidant therapy actually decreases oxidative stress. At this time, antioxidants should not be used as sole therapy for alcoholic hepatitis; whether they are of benefit as adjunctive therapy in patients with alcoholic hepatitis or alcoholic cirrhosis and which antioxidant should be used remain unclear.

DRUGS OF UNLIKELY BENEFIT Colchicine

Colchicine has many potential therapeutic mechanisms of action in alcoholic liver disease, including inhibition of collagen production, enhancement of collagenase activity, and anti-inflammatory activity. Initial positive studies158 led to a large Veterans Administration (VA) Cooperative Study of colchicine therapy in patients with alcoholic cirrhosis that showed no beneficial effects on overall or liver-related mortality.159 A smaller study from Europe also showed no beneficial effects of colchicine therapy in patients with alcoholic liver disease.160

Propylthiouracil

Chronic alcohol feeding in animal models can induce a hypermetabolic state with increased oxygen consumption

similar to the hypermetabolic state associated with hyperthyroidism. This hypermetabolic state may lead to relative hypoxia in the centrilobular area of hepatic lobules. Propylthiouracil has been postulated to attenuate the hypermetabolic state, function as an antioxidant, and improve portal blood flow. Nevertheless, a Cochrane review of six randomized trials involving more than 700 patients found no beneficial effect of propylthiouracil therapy in patients with alcoholic liver disease.161

Anabolic Steroids

Anabolic steroids have been shown to decrease fatty infiltration in the liver and are hepatoprotective. As noted earlier, patients with end-stage liver disease frequently are malnourished and often have low circulating levels of the anabolic hormone insulin-like growth factor-1. These observations provide a rationale for using anabolic steroids to treat alcoholic liver disease.132 Nevertheless, a Cochrane review was not able to demonstrate efficacy for anabolic steroids (specifically oxandrolone) in patients with alcoholic liver disease, although such therapy did appear to be safe.162

Ursodeoxycholic Acid

Urosodeoxycholic acid, an agent used for a variety of cholestatic liver disorders (see Chapter 89), was evaluated in one large multicenter controlled clinical trial in patients with severe alcoholic liver disease and found to have no beneficial effect on six-month survival.163

Polyenylphosphatidylcholine

Polyenylphosphatidylcholine, or lecithin, a lipid extract obtained from soybeans, has been shown to prevent septal fibrosis and cirrhosis in alcohol-fed baboons and to stimulate the release of collagenase activity by cultured hepatic stellate cells. It also has antioxidant effects and decreases TNF production.164 Multiple positive studies of poly­ enylphosphatidylcholine in animal models of liver disease led to a VA Cooperative Study that evaluated the effects of this drug in humans with early alcoholic liver disease.165 Results of this study were negative; however, patients decreased their alcohol use markedly during the trial, thus decreasing the likelihood that a beneficial effect of poly­ enylphosphatidylcholine could be demonstrated.

LIVER TRANSPLANTATION

Liver transplantation for alcoholic liver disease remains one of the most contentious and controversial areas in transplantation medicine. Short-term outcomes following liver transplantation in patients with alcoholic liver disease are comparable to those for patients who receive transplants for most other conditions, with seven-year survival rates of 60% (see also Chapter 95).166 Profound confusion in the early postoperative period is more likely to develop in patients with alcoholic cirrhosis, however, than in those undergoing transplantation for other liver diseases.167 The result can be a prolonged hospitalization and an increase in the cost of transplantation. A number of transplant centers have reported distinctly lower long-term survival rates among patients with alcoholic liver disease, particularly among those who return to heavy drinking.168 In addition, patients with alcoholic liver disease have an increased risk of pharyngeal, esophageal, and gastric malignancies after transplantation.169 This risk is particularly high among persons who begin smoking heavily again after the operation.170 Many patients with apparently advanced alcoholic liver disease can recover to the degree that transplantation is not

Chapter 84  Alcoholic Liver Disease 1.00

0.8

0.75 Probability of survival

Probability of improving liver function

1.0

0.6

0.4

Transplanted: 58% 5-year survival

0.50

Simulated: 35% 5-year survival

Matched controls: 31% 5-year survival

0.25

0.2

0

0.00 0

60

120

180

240

300

360

0

500

required if they reduce their alcohol intake significantly or abstain completely (Fig. 84-14).124 Because the benefits of abstinence can be so dramatic, requiring a period of abstinence before proceeding with transplantation is reasonable for patients with alcoholic liver disease. Patients who have a CTP score of 11 or greater despite at least six months of abstinence have improved survival with liver transplantation compared with predicted survival based on the Beclere model (Fig. 84-15).119 Similar, although less impressive, results have been shown using other prognostic models.171 The optimal length of pretransplant abstinence remains controversial.40 Some experts have argued that patients with severe alcoholic hepatitis should be abstinent for one year before being considered for transplantation, whereas others have argued that patients should be considered for transplantation if they continue to have CTP scores of 11 or greater after only three months of abstinence.124,139 Evidence of a survival benefit following transplantation is less clear for patients with milder alcoholic liver disease, unless they have hepatocellular carcinoma. Patients with CTP scores of 5 to 7 do not benefit from liver transplantation.119 The survival benefit from transplantation for patients with a CTP score of 8 to 10 after 6 months of abstinence is minimal compared with predicted survival using the Beclere and MELD models.119,171 Furthermore, a trial in which patients with a CTP score of 8 to 10 were randomized to receive immediate transplantation or to be observed expectantly showed a lower two-year survival rate among the patients randomized to undergo immediate transplantation (73% versus 80%), primarily because of a high risk of postoperative malignancy.172 Efforts to refine risk scores in patients with severe alcoholic liver disease and incorporate them into standard transplantation selection criteria as well as efforts to select

1500

2000

Days

Follow-up duration in days Figure 84-14.  Probability that liver function will improve with time in patients with cirrhosis caused by alcoholic liver disease who reduce alcohol intake or abstain from alcohol. Improvement was defined as a change from Child-Turcotte-Pugh class C to class B or A. Red line, patients who reduced intake; blue line, abstinent patients; circles, deaths; triangle, patient undergoing transplantation. (From Veldt BJ, Laine F, Guillygomarc’h A, et al. Indication of liver transplantation in severe alcoholic liver cirrhosis: Qualitative evaluation and optimal timing. J Hepatol 2002; 36:93-8, with permission.)

1000

Figure 84-15.  Probability of survival over five years in patients with ChildTurcotte-Pugh scores of 11 to 15 after six months of abstinence from alcohol who underwent liver transplantation (top line), compared with matched control subjects (P = .008) and simulated control subjects (i.e., predicted from a model) (P = .001). (Modified from Poynard T, Naveau S, Doffoel M, et al. Evaluation of efficacy of liver transplantation in alcoholic cirrhosis using matched and simulated controls: 5 year survival. Multicentre group. J Hepatol 1999; 30:1130-7.)

patients who can maintain long-term abstinence for alcohol and tobacco abuse are ongoing.

OPTIMAL MANAGEMENT

Alcoholic liver disease accounts for 50,000 deaths annually in the United States and Europe. The optimal management of patients with alcoholic liver disease begins with a dramatic reduction in or elimination of alcohol intake, which often can be accomplished successfully using “brief interventions” by a nurse, primary care physician, or gastroenterologist. Abstinence can have a profound impact on survival even in patients with decompensated cirrhosis. The next important step is to eliminate other factors, such as cigarette smoking and obesity, which can enhance disease progression. Treatment of concomitant HCV infection may be an important aspect of management in some patients (see Chapter 79). Patients with severe alcoholic hepatitis should receive enteral feedings to ensure adequate calorie and protein intake. In patients with severe alcoholic hepatitis who do not have a systemic infection or gastrointestinal bleeding, a short course of glucocorticoid therapy should be considered. An alternative strategy is the use of pentoxifylline, especially in patients with marginal renal function or hepatorenal syndrome. For patients with alcoholic cirrhosis, no drugs available in the United States have documented efficacy. Many patients already are taking agents such as milk thistle or SAM that may be of benefit and appear to be safe. These agents are not covered by insurance, however, and access to such agents often depends on the socioeconomic status of the patient. Because of the risk of decompensation with superimposed infections, all patients with alcoholic cirrhosis should receive vaccinations for hepatitis A and B and annual vaccinations for influenza. In addition, they should

1399

1400

Section IX  Liver undergo regular screening for hepatocellular carcinoma (see Chapter 94). Screening is particularly important in older patients with cirrhosis who have been abstinent for sustained periods. Finally, liver transplantation has been shown to be effective in carefully selected patients who have discontinued drinking (see Chapter 95).

KEY REFERENCES

Akriviadis E, Botla R, Briggs W, et al. Pentoxifylline improves shortterm survival in severe acute alcoholic hepatitis: A double-blind, placebo-controlled trial. Gastroenterology 2000; 119:1637-48. (Ref 85.) Arteel GE. Oxidants and antioxidants in alcohol-induced liver disease. Gastroenterology 2003; 124:778-90. (Ref 13.) Boetticher NC, Peine CJ, Kwo P, et al. A randomized, double-blinded, placebo-controlled multicenter trial Etanercept in the treatment of alcoholic hepatitis. Gastroenterology 2008; 135;1953-60. (Ref 146.) Cabre E, Rodriguez-Iglesias P, Caballeria J, et al. Short- and long-term outcome of severe alcohol-induced hepatitis treated with steroids or enteral nutrition: A multicenter randomized trial. Hepatology 2000; 32:36-42. (Ref 136.) Ji C. Dissection of endoplasmic reticulum stress signaling in alcoholic and non-alcoholic liver injury. J Gastroenterol Hepatol 2008; 23(Suppl 1):S16-24. (Ref 70.) Kunos G, Osei-Hyiaman D, Batkai S, Gao B. Cannabanoids hurt, heal in cirrhosis. Nat Med 2006; 12:608-10. (Ref 71.) Mathurin P, Mendenhall CL, Carithers RL, et al. Glucocorticoids improve short-term survival in patients with severe alcoholic hepa-

titis (AH): Individual data analysis of the last three randomized placebo controlled double blind trials of glucocorticoids in severe AH. J Hepatol 2002; 36:480-7. (Ref 116.) Niemelä O. Biomarkers in alcoholism. Clin Chimica Acta 2007; 377:3949. (Ref 80.) Pfitzmann R, Schwenzer J, Rayes N, et al. Long-term survival and predictors of relapse after orthotopic liver transplantation for alcoholic liver disease. Liver Transplant 2007; 13:197-205. (Ref 168.) Plank LD, Gane EJ, Peng S, et al. Nocturnal nutritional supplementation improves total body protein status of patients with liver cirrhosis: A randomized 12-month trial. Hepatology 2008; 48:557-66. (Ref 138.) Purohit V, Abdelmalek MF, Barve S, et al. Role of S-adenosylmethionine, folate, and betaine in the treatment of alcoholic liver disease: Summary of a symposium. Am J Clin Nutr 2007; 86:14-24. (Ref 35.) Shukla SD, Velazquez J, French SW, et al. Emerging role of epigenetics in the actions of alcohol. Alcohol Clin Exp Res 2008; 32:1525-34. (Ref 69.) Stewart S, Prince M, Bassendine M, et al. A randomized trial of antioxidant therapy alone or with corticosteroids in acute alcoholic hepatitis. J Hepatol 2007; 47:277-83. (Ref 157.) Veldt BJ, Laine F, Guillygomarc’h A, et al. Indication of liver transplantation in severe alcoholic liver cirrhosis: Quantitative evaluation and optimal timing. J Hepatol 2002; 36:93-8. (Ref 124.) Yip WW, Burt AD. Alcoholic liver disease. Semin Diagn Pathol 2006; 23:149-60. (Ref 8.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

85 Nonalcoholic Fatty Liver Disease Andrea E. Reid

CHAPTER OUTLINE Nonalcoholic Fatty Liver and Steatohepatitis  1401 Epidemiology  1401 Etiology  1401 Pathogenesis  1402 Clinical, Laboratory, and Imaging Features  1405

In 1980, Ludwig and colleagues from the Mayo Clinic coined the term nonalcoholic steatohepatitis (NASH) to describe a form of liver disease observed in middle-aged patients with abnormal liver biochemical test results and histologic evidence of alcoholic hepatitis but no history of alcohol abuse.1 Much has been learned about NASH since this initial description. NASH is part of the spectrum of nonalcoholic fatty liver disease (NAFLD), which encompasses simple fatty liver, NASH, and NAFLD-associated cirrhosis. NAFLD has emerged as a burgeoning clinical entity, now recognized as an important component of the metabolic syndrome, as well as an exciting area of basic and clinical investigation in the field of hepatology.

NONALCOHOLIC FATTY LIVER AND STEATOHEPATITIS EPIDEMIOLOGY

The prevalence of NAFLD in the general population is undefined. Several studies have estimated the scope of this disorder in the United States. The Dallas Heart Study of more than 2200 adults documented hepatic triglyceride content by proton magnetic resonance spectroscopy (MRS) and found fatty liver in 31% of participants; the highest prevalence (45%) was among Hispanics.2 Most of the patients with fatty liver by MRS had normal liver biochemical test levels, although the normal range for serum aminotransferase levels in this study was wider than generally accepted. The National Health and Nutrition Examination Survey (NHANES) III, which included more than 15,700 adults, documented unexplained elevations of serum aminotransferase levels, presumably caused by NAFLD, in 2.8% to 5.5% of participants.3,4 Population-based estimates of NAFLD have been reported for other countries as well. These studies have documented NAFLD in 10% to 24% of the population, with the highest prevalence (up to 76%) among obese nondrinkers.5-7 Prevalence estimates vary widely depending on the information available in a given population and the diagnostic criteria that are used to estab-

Histopathologic Features  1405 Diagnosis  1407 Natural History  1408 Treatment  1409 Focal Fatty Liver  1410

lish the diagnosis (i.e., liver biochemical test levels, radiologic study results, or liver biopsy findings). Most cases of NAFLD are discovered in the fourth to sixth decades of life, although NAFLD is also described, with increasing frequency, in obese children and adolescents, as well as in older adults. NAFLD may be present long before a diagnosis is established. In early clinical studies, the majority of patients with NAFLD were female; however, subsequent data have suggested that men may be affected as often as women and may be at greater risk for advanced forms of NAFLD, including NASH. The prevalence of NAFLD appears to vary by ethnicity. In the Dallas Heart Study, Hispanics demonstrated the highest prevalence (45%) of NAFLD, compared with 33% for whites and 24% for African Americans. The reasons for racial and ethnic disparities in the prevalence of NAFLD is not known but may be related, at least in part, to racial differences in body fat distribution8 and the prevalence of the metabolic syndrome, which is greatest in people of Hispanic descent.9 Other studies have also shown that African Americans and Mexican Americans have higher frequencies of unexplained serum aminotransferase elevations than do whites.3,4,9,10 Familial clustering of NAFLD may occur,11,12 which likely reflects both genetic and environmental predisposition to the metabolic conditions associated with NAFLD (see later).13

ETIOLOGY

Many agents and conditions have been associated with NAFLD. The causes may be divided into two broad categories (1) drugs and toxins and (2) metabolic abnormalities, either acquired or congenital. Potential causes of NAFLD are listed in Table 85-1. Obesity is the condition most often reported in association with NAFLD. Since 1980, the proportion of Americans who are overweight (defined as a body mass index [BMI] > 25 kg/m2) or obese (BMI > 30 kg/m2) has increased markedly. In 2004, 66.2% of Americans adults were classified as overweight or obese, as were 17.4% of children ages 12 to 19 years.14 The health implications of the unremitting obesity epidemic are staggering, and NAFLD is

1401

1402

Section IX  Liver Table 85-1  Causes of Nonalcoholic Fatty Liver Disease Acquired Metabolic Disorders Diabetes mellitus Dyslipidemia Kwashiorkor and marasmus Obesity Starvation Cytotoxic and Cytostatic Drugs l-Asparaginase Azacitidine Azaserine Bleomycin Methotrexate Puromycin Tetracycline* Other Drugs and Toxins Amiodarone 4,4´-diethylaminoethoxyhexestrol Dichlorethylene Ethionine Ethyl bromide Estrogens Glucocorticoids Highly active antiretroviral therapy Hydrazine Hypoglycin Orotate Perhexilene maleate Safrole Tamoxifen Metals Antimony Barium salts Chromates Phosphorus Rare earths of low atomic number Thallium compounds Uranium compounds Inborn Errors of Metabolism Abetalipoproteinemia Familial hepatosteatosis Galactosemia Glycogen storage disease Hereditary fructose intolerance Homocystinuria Systemic carnitine deficiency Tyrosinemia Weber-Christian syndrome Wilson disease Surgical Procedures Biliopancreatic diversion Extensive small bowel resection Gastric bypass Jejunoileal bypass Miscellaneous Conditions Industrial exposure to petrochemicals Inflammatory bowel disease Partial lipodystrophy Jejunal diverticulosis with bacterial overgrowth Severe anemia Total parenteral nutrition *Tetracycline is cytotoxic by virtue of inhibiting mitochondrial b-oxidation.

a common byproduct in both adults and children. As noted earlier, most patients with NAFLD are obese. In morbidly obese patients (BMI > 35 kg/m2), including those referred for bariatric surgery, the frequency of NAFLD is as high as 90%, with advanced disease (i.e., NASH) seen in 9% to 40%.15-19 A correlation among BMI, degree of steatosis, and severity of liver injury has been demonstrated in several studies20-22; however, the distribution of body fat may be

more important than the total adipose mass for the development of hepatic steatosis. Studies have shown a significant correlation between the risk of the metabolic syndrome, degree of hepatic steatosis, and waist-to-hip ratio, thus highlighting the importance of intra-abdominal or visceral fat as a pre­dictor of NAFLD.23-25 NAFLD also is strongly associated with type 2 diabetes mellitus and glucose intolerance, with or without superimposed obesity.26 Type 2 diabetes mellitus, hyperglycemia, or glucose intolerance has been described in 20% to 75% of adult patients with NASH and may increase the risk of NASH more than twofold compared with that for nondiabetic persons. The presence of NAFLD in diabetic patients may also increase the risk of cardiovascular disease significantly.27 The association between type 2 diabetes mellitus and NAFLD appears strongest in morbidly obese patients.15 NAFLD has been associated with insulin resistance and hyperinsulinemia even in lean subjects with normal glucose tolerance.28 Diabetes mellitus may be an independent predictor of advanced NAFLD, including cirrhosis and hepatocellular carcinoma.29-32 Hyperlipidemia is found in a substantial proportion of patients with NAFLD. Data from the Dallas Heart Study revealed NAFLD in 60% of patients with mixed hyperlipidemia,33 and a study from Korea of potential living liver donors showed that hyperlipidemia was associated with a greater than twofold risk of significant (>30%) steatosis.34 Most patients with NAFLD have multiple risk factors, including central obesity, type 2 diabetes mellitus, and hyperlipidemia, although some affected persons lack all recognized risk factors. NAFLD has been associated with many drugs and toxins and metabolic, surgical, and genetic conditions (see Table 85-1) that have abnormal fat metabolism and mitochondrial injury or dysfunction in common. NAFLD is now recognized as the hepatic component of the metabolic syndrome, which includes hyperlipidemia, glucose intolerance, obesity, and systemic hypertension. The risk and severity of NAFLD increase with the number of components of the metabolic syndrome.28,35

PATHOGENESIS

The pathogenesis of NAFLD is poorly understood, in part because of a lack of suitable animal models that mimic human NAFLD. In light of the variety of conditions that have been associated with NAFLD, it is not surprising that no single pathogenic mechanism has been identified. The prevailing theory is the “two-hit hypothesis,” first proposed by Day and James in 1998.36 This hypothesis states that dysregulation of fatty acid metabolism leads to steatosis, which is the first hepatic insult in NAFLD. Steatosis is associated with several cellular adaptations and altered signaling pathways, which render hepatocytes vulnerable to a “second hit.” The second insult may be one or more environmental or genetic perturbations, which cause hepatocyte necrosis and inflammation and activate the fibrogenic cascade, thereby leading to fibrosis and cirrhosis in a minority of patients with NAFLD. Hepatic steatosis is the hallmark histologic feature of NAFLD. Normally, free fatty acid (FFA) is supplied to the liver through intestinal absorption (in the form of chylomicron remnants) or from lipolysis of adipose tissue, where FFA is stored as triglycerides. In the liver, FFA is oxidized by mitochondria, esterified into triglycerides, synthesized into phospholipids and cholesteryl esters, and secreted from the liver as very-low-density lipoprotein (VLDL). Under normal circumstances, fatty acid metabolism is under tight regulatory control by catecholamines, glucagon, growth hormone, and insulin. Hepatic triglyceride accumulation

Chapter 85  Nonalcoholic Fatty Liver Disease occurs when fatty acid metabolism shifts to favor net lipogenesis, rather than lipolysis. This shift occurs when the amount of FFA supplied to the liver from the intestine or adipose tissue exceeds the amount needed for mitochondrial oxidation, phospholipid synthesis, and synthesis of cholesteryl esters. Triglycerides also accumulate in the liver when synthesis of lipoprotein decreases or export of lipids from the liver is impeded (see also Chapter 72). Current evidence points to insulin resistance and hyperinsulinemia as the primary pathogenic factors in steatosis in most patients with NAFLD. Strong laboratory and clinical evidence supports the association of peripheral insulin resistance and hyperinsulinemia with NAFLD, even in lean patients without obvious glucose intolerance.37-39 The molecular mechanism leading to insulin resistance is complex and not understood completely. In the setting of obesity and hyperinsulinemia, alterations in several molecules, including FFA, tumor necrosis factor-α (TNF-α), membrane glycoprotein PC-1, and leptin, interfere with the insulin signaling pathway. Diabetes mellitus and obesity are associated with increased amounts of FFA in plasma, caused in part by abnormal release of FFA by insulin-resistant adipocytes. Excess FFA contributes to hepatic insulin resistance by down-regulating insulin receptor substrate-1 (IRS-1) signaling.40,41 Insulin resistance and hyperinsulinemia lead to steatosis by means of a number of aberrant mechanisms of FFA disposal. In the liver, insulin stimulates fatty acid synthesis, down-regulates mitochondrial boxidation of FFA, blocks the secretion of triglycerides from hepatocytes by increasing intracellular degradation of VLDL and apolipoprotein B-100 (apoB-100), and blocks exocytosis of VLDL-containing vesicles.40,42,43 Also, patients with NASH have impaired hepatic synthesis of apoB-100, which also may contribute to hepatic triglyceride accumulation.44 Insulin resistance in NAFLD may be potentiated by aberrant levels or function of several important peptide mediators secreted by adipocytes, including TNF-α, leptin, and adiponectin. In noninflammatory states, TNF-α is derived from adipose tissue (including adipose tissue macrophages), and plasma levels of TNF-α correlate with body fat mass.45 TNF-α interferes with insulin signaling by down-regulating IRS-1 signaling via serine phosphorylation, likely through activation of stress-related protein kinases including Jun N-terminal kinase (JNK), which plays a key role in obesityrelated insulin resistance. Activation of the inhibitor kappab kinase (IKK-b)/nuclear factor kappa b (NF-κb) pathway by FFA may also play a role in reduced hepatic insulin sensitivity,46,47 and may increase production of additional inflammatory cytokines such as TNF-α and interleukin (IL)-6.48 Elevated TNF-α levels have been demonstrated in several studies of NAFLD49-52; however, the independent contribution of TNF-α to the pathogenesis and risk of progression of NAFLD is still unclear. Adipocytokines are peptides produced by visceral adipose tissue. Adiponectin is secreted by adipocytes in inverse proportion to BMI and is a potent inhibitor of TNF-α. Serum adiponectin levels are reduced in obesity, insulin resistance, diabetes mellitus, and the metabolic syndrome.50 Delivery of recombinant adiponectin to mice fed a high-fat, alcohol-containing diet and to genetically obese (ob/ob) mice dramatically alleviates hepatomegaly, steatosis, inflammation, and elevated liver biochemical test levels in both murine populations.53 These therapeutic effects result in part from the ability of adiponectin to enhance hepatic fatty acid b-oxidation, decrease hepatic triglyceride content, and decrease hepatic insulin resistance. Furthermore, adiponectin suppresses hepatic and plasma concentrations of TNF-α. Studies have reported an inverse relationship

between serum adiponectin levels and the degree of steatosis and hepatocyte injury in humans with NAFLD, and this inverse association may be independent of insulin resistance.51,54 Further studies are needed to determine whether an increase in the TNF-α/adiponectin ratio has a primary pathogenic role in the development of steatosis or is more directly correlated with progression from steatosis to steatohepatitis. Leptin is a satiety hormone, derived from adipocytes, that controls food intake and energy regulation (see Chapter 1). Leptin is intimately involved with insulin signaling and regulation of glucose metabolism in peripheral tissues and may play an important role in regulating the partitioning of fat between mitochondrial b-oxidation and triglyceride synthesis in the liver.55 Severe steatosis and steatohepatitis develop in leptin-deficient (ob/ob) mice. Obesity in humans is associated with relative leptin resistance and high leptin levels, which may contribute to the genesis of steatosis by a negative impact on insulin signaling or may be a consequence of the chronic hyperinsulinemia associated with obesity. Several studies have examined the relationship between serum leptin levels and NAFLD, with conflicting results.56-59 One study has suggested that serum leptin levels in patients with NASH correlate with the severity of hepatic steatosis, independent of BMI, but not with the degree of hepatic inflammation or fibrosis.58 At present, the contri­bution of leptin to the pathogenesis of NAFLD is unclear. Although insulin resistance and hyperinsulinemia are clearly pivotal to the development of steatosis, consensus is lacking on the subsequent insults that cause steatosis to progress to steatohepatitis and fibrosis in some patients. Similarities in the histologic features and natural history of alcoholic liver disease and NAFLD suggest that common mechanisms may be involved in the pathogenesis of these disorders. Chronic oxidative stress is believed to be central to the pathogenesis of alcohol-related liver damage. Processes that increase the production of oxidants in the liver during chronic alcohol exposure include the metabolism of ethanol to its reactive intermediate acetaldehyde; induction of microsomal ethanol-oxidizing enzymes, such as cytochrome P450 2E1 (CYP2E1), which generates reactive oxygen species (ROS) that can peroxidize cellular membranes, thereby causing cellular injury60; inhibition of mitochondrial electron transport chain activity; and depletion of mitochondrial glutathione.61 Activation of microsomal enzymes, including CYP2E1, in patients with NAFLD62-63 and mitochondrial production of ROS in murine models of NAFLD64,65 suggest that chronic oxidative stress and lipid peroxidation may also be central to the pathogenesis of NAFLD. Increased levels of FFA can be directly toxic to hepatocytes through a number of mechanisms. An increased FFA concentration leads to lysosomal destabilization and stimulation of TNF-α.66 FFA also up-regulates cytochrome P450 isoenzymes, leading to enhanced generation of ROS and lipid peroxidation.67 An increased intracellular FFA concentration can lead to sustained up-regulation of peroxisomal proliferator-activated receptor-α (PPAR-α), which promotes fatty acid oxidation and disposal but also may increase oxidative stress through the production of dicarboxylic acid derivatives; PPAR-α also may predispose affected persons to carcinogenesis.44 FFA can be directly toxic to cellular membranes, lead to the formation of toxic fatty acid ethyl ethers, and cause overall disruption of mitochondrial function, thereby overwhelming the over­ lapping protective mechanisms designed to combat FFA hepatotoxicity.45

1403

1404

Section IX  Liver Endotoxin and endotoxin-mediated cytokine release are suspected in the pathogenesis of alcoholic steatohepatitis, in which increased serum levels of bacterial endotoxin and lipopolysaccharide (LPS) stimulate hepatic production of TNF-α, IL-6, and IL-8 and activate an inflammatory response that leads to hepatic necrosis (see Chapter 84).61 Bacterial endotoxin also may contribute to the development of NAFLD in some circumstances. Portal endotoxemia was believed to contribute to NASH and hepatic failure associated with surgical jejunoileal bypass (performed in the past to treat obesity), the risk of which was reduced with antibiotics. Yang and colleagues have demonstrated that ob/ob mice with steatosis are highly vulnerable to endotoxininduced hepatocyte damage, and NASH rapidly develops in these animals after exposure to low doses of bacterial LPS.68 In addition, Zucker diabetic (fa/fa) rats and ob/ob mice demonstrate decreased Kupffer cell function, which may increase the vulnerability of steatotic hepatocytes to TNFα–mediated liver damage.69 Small studies suggest a possible pathogenic role of bacterial endotoxins in human NAFLD as well,70,71 but these studies are far from conclusive. A growing body of evidence suggests that mitochondrial changes and altered hepatic energy homeostasis may play roles in the pathogenesis of NAFLD. Studies have shown a decrease in the activity of mitochondrial respiratory chain complexes in steatotic livers, with a concomitant increase in mitochondrial ROS formation; these changes correlate with serum TNF-α levels, insulin resistance, and BMI.72,73 Ob/ob mice have increased levels of uncoupling protein, UCP-2, an inner mitochondrial membrane protein that mediates proton leak, uncouples adenosine triphosphate (ATP) synthesis, regulates ROS production, and may render fatty hepatocytes vulnerable to metabolic stressors65; however, the role of UCP-2 in NAFLD in humans is unknown. Studies have shown that mice and humans with NAFLD have diminished capacity for replenishing ATP stores after ATP depletion. Mitochondrial structural defects may be one cause of reduced ATP stores. Megamitochondria and crystalline mitochondrial inclusions have been iden­ tified in patients with NAFLD and may represent an adaptive process to oxidative stress or secondary injury.39,74 Limited data in patients with NASH suggest differential expression of several genes important for proper mitochondrial functioning, including genes involved in ROS scavenging, glucose metabolism, and fatty acid metabolism.75 In addition, mitochondrial DNA damage similar to that found in alcoholic liver disease and Wilson disease also may contribute to the development of NASH. Further animal and human studies are needed to determine whether mitochondrial dysfunction and ATP depletion are causes or consequences of NAFLD. Fibrosis is a frequent histologic finding in advanced NAFLD but has not been well studied in this disease. Hepatic fibrosis results from activation and proliferation of hepatic stellate cells in the subendothelial space of Disse, with subsequent secretion of extracellular matrix components, including collagen types I and III (see Chapter 90). Factors proposed to initiate and perpetuate the fibrogenic process in stellate cells include inflammatory cytokines, angiotensin, alterations in the extracellular matrix, growth factors, and oxidative stress. In NAFLD, lipid peroxidation products may enhance hepatic production of transforming growth factor-b (TGF-b), which activates stellate cells. Endothelial cells, leukocytes, and Kupffer cells may stimulate the stellate cells to proliferate, possibly through the release of platelet-derived growth factor (PDGF), TGF-b, and other cytokines.76 In addition, hyperinsulinemia and hyperglycemia associated with NAFLD may stimulate release of

connective tissue growth factor, an intermediate molecule involved in fibrogenesis.77 Finally, animal data suggest that leptin may perpetuate fibrogenesis in NAFLD by stimulating Kupffer cells and sinusoidal endothelial cells to produce TGF-b.78 Research into the pathogenesis of NAFLD is proliferating at a rapid pace, but the picture is far from clear (Fig. 85-1). Any one of the putative mechanisms discussed here is unlikely to explain the pathogenesis of NAFLD in all affected patients. More likely, NAFLD develops as a con­ sequence of a “multi-hit” process. The first “hit” is steatosis

Obesity Insulin resistance Hyperinsulinemia Increased serum leptin levels Altered cytokine levels: Increased TNF-α Decreased adiponectin

Excessive dietary carbohydrates, dyslipidemia, genetic mutations, drugs, toxins, nutritional deficiencies, other factors

Fat-containing hepatocytes

Direct cytotoxic effects of increased FFA

Oxidative stress and lipid peroxidation

Impaired β-oxidation of FFA

Mitochondrial damage: Increased ROS Altered ATP homeostasis Structural abnormalities

Steatohepatitis ? Environmental, genetic, dietary factors

Increased TGF-β and other cytokines

Fibrosis/cirrhosis Figure 85-1.  Proposed pathogenesis of nonalcoholic fatty liver disease (NAFLD). NAFLD is believed to occur as a result of a multi-hit process. Insulin resistance and hyperinsulinemia are present in many patients with NAFLD and, possibly in association with other metabolic or genetic abnormalities and altered cytokine levels, may lead to hepatic steatosis. Hepatocellular injury leading to steatohepatitis occurs in a minority of patients with fatty liver, probably as a result of multiple overlapping insults. Increased levels of free fatty acid (FFA), uncompensated oxidative stress, lipid peroxidation, cytokine dysregulation, mitochondrial dysfunction, and other environmental and genetic factors may contribute to the development of hepatocellular injury and fibrosis in susceptible persons. ATP, adenosine triphosphate; ROS, reactive oxygen species; TNF-α, tumor necrosis factor-α; TGF-b, transforming growth factor-b.

Chapter 85  Nonalcoholic Fatty Liver Disease Table 85-2  Clinical and Laboratory Features of Nonalcoholic Fatty Liver Disease SYMPTOMS

SIGNS

LABORATORY FEATURES

Common None (48%-100% of patients)

Hepatomegaly

Two- to fourfold elevation of serum ALT and AST levels AST/ALT ratio less than 1 in most patients Serum alkaline phosphatase level is slightly elevated in one third of patients Normal serum bilirubin and serum albumin levels and prothrombin time Elevated serum ferritin level

Splenomegaly Spider angiomata Palmar erythema Ascites

Low-titer (less than 1 : 320) ANA Elevated transferrin saturation HFE gene mutation (C282Y)

Uncommon Vague right upper quadrant pain Fatigue Malaise

ANA, antinuclear antibodies; ALT, alanine aminotransferase; AST, aspartate aminotransferase.

induced primarily by insulin resistance and hyperin­ sulinemia. After steatosis develops, a number of factors, including lipid peroxidation, oxidative stress, cytokine alterations, mitochondrial dysfunction, and Kupffer cell activation, may incite an inflammatory response and fibrosis in some patients with genetic or environmental susceptibilities. The exact interplay among these and other factors remains to be elucidated, but understanding of the pathogenesis should be enhanced by the development and refinement of appropriate animal models, including the genetically obese, leptin-deficient ob/ob mouse and Zucker diabetic rats, in which steatosis develops; S-adenosylmethionine (SAM)-deficient mice, in which severe steatohepatitis deve­ lops79; the Otsuka-Long-Evans-Tokushima fatty (OLETF) rat model, in which a cholecystokinin-A receptor defect leads to hyperglycemia, obesity, insulin resistance, and hepatic steatosis80; and a high fat-fed rat model in which insulin resistance, elevated serum TNF-α levels, increased oxidative stress, mitochondrial lesions, and early fibrosis develop.81

CLINICAL, LABORATORY, AND IMAGING FEATURES

The clinical and laboratory features of NAFLD are summarized in Table 85-2. NAFLD usually is discovered incidentally because of elevated liver biochemical test levels or hepatomegaly noted during an evaluation for an unrelated medical condition. Most patients with NAFLD are asymptomatic, but some may describe vague right upper quadrant pain, fatigue, and malaise. Hepatomegaly has been described in up to 75% of patients with NAFLD but often is difficult to appreciate on physical examination because of obesity. Stigmata of chronic liver disease, such as splenomegaly, spider angiomata, and ascites, are rare, except in patients with NAFLD-associated cirrhosis. Elevated liver biochemical test levels may be found in up to 50% of patients with simple steatosis and are present in approximately 80% of patients with advanced NAFLD. A mild-to-moderate (1.5- to 4-fold) elevation of the serum aspartate aminotransferase (AST) or alanine aminotransferase (ALT) level, or both, is usual, and levels seldom exceed 10 times the upper limit of normal. The serum ALT level usually is greater than the AST level, in contrast with the pattern of alcoholic hepatitis, in which the AST level is at least twofold higher than the ALT level (see Chapters 73 and 84). The alkaline phosphatase and gamma glutamyl transpeptidase (GGTP) levels may be elevated, but the

serum bilirubin level, prothrombin time, and serum albumin level typically are normal, except in patients with NAFLDassociated cirrhosis. Up to one fourth of patients with NAFLD may have antinuclear antibodies (ANA) in low titers (less than 1 : 320).82 Antimitochondrial antibodies (AMA) and hepatitis B surface antigen are not detected. Antibody to hepatitis C virus (anti-HCV) must be absent to implicate NAFLD as the sole cause of abnormal liver biochemical test levels; however, steatosis, often in association with visceral obesity, frequently accompanies HCV infection and may be associated with a more aggressive course (see Chapter 79).83 Serum ceruloplasmin and α1-antitrypsin levels are within normal limits. Serum and hepatic iron levels may be elevated in patients with NAFLD. In particular, the serum ferritin level may be elevated in 20% to 50% of patients with NAFLD and may be a marker of more advanced disease.29,84 Nevertheless, the frequency of genetic hemochromatosis among patients with NAFLD is not increased. Clinical and laboratory findings do not correlate with the histologic severity of NAFLD. The entire histologic spectrum of NAFLD, including cirrhosis, can be seen in patients with normal or near-normal serum aminotransferase levels.31 Imaging studies often are obtained during the evaluation of unexplained liver biochemical abnormalities or suspected NAFLD. Hepatic ultrasonography, the imaging modality employed most commonly, may reveal a “bright” liver of increased echogenicity, consistent with hepatic steatosis (Fig. 85-2). Fatty liver also can be documented by abdominal computed tomography (CT) scan (a fatty liver is lower in density than the spleen), and by magnetic resonance imaging (MRI), with which fat appears bright on T1-weighted imaging. A study that assessed the sensitivities of MRI, abdominal CT, and ultrasonography for distinguishing advanced NASH from simple steatosis showed that ultrasonography and CT had sensitivity rates of 100% and 93% for detecting hepatic fat involving greater than 33% of the liver, with positive predictive values of 62% and 76%, respectively.85 No radiologic modality, however, was able to distinguish simple steatosis from more advanced forms of NAFLD. Imaging studies such as ultrasonography may support the diagnosis of NAFLD but cannot predict the severity of disease and cannot replace liver biopsy for establishing the diagnosis with certainty.

HISTOPATHOLOGIC FEATURES

The major histologic features of NAFLD resemble those of alcohol-induced liver disease and include steatosis (fatty

1405

1406

Section IX  Liver

A

B

Figure 85-2.  Imaging studies of fatty liver. A, Ultrasound demonstrating increased echogenicity. B, T1-weighted magnetic resonance image demonstrating a bright liver. (Courtesy of Dr. Mukesh Harisinghani, Boston, Mass.)

Figure 85-3.  Histologic features of simple steatosis (fatty liver). The characteristic feature is diffuse macrovesicular steatosis. Glycogenated nuclei are common. (Hematoxylin and eosin.) (Courtesy of Dr. Gregory Y. Lauwers, Boston, Mass.)

liver), steatohepatitis (fatty liver plus parenchymal inflammation with or without accompanying focal necrosis), and varying degrees of fibrosis, including cirrhosis. Steatosis is predominantly macrovesicular and usually is distributed diffusely throughout the liver lobule, although prominent microvesicular steatosis and zone 3 (perivenular) steatosis have been reported occasionally (Fig. 85-3). Mild lymphocytic, neutrophilic, or mixed inflammatory infiltrates also may be observed, and glycogenated nuclei are common. NASH, which is an advanced form of NAFLD, is indistinguishable histologically from alcoholic hepatitis (Fig. 85-4 and Table 85-3). Steatosis is present in all cases and can affect the hepatic lobules either diffusely or primarily in the central zones. The degree of steatosis may correlate with the patient’s BMI and generally is more severe in NASH than in alcoholic hepatitis.86 Lobular inflammation is a hallmark feature of NASH and is characterized by infiltration of lymphocytes, other mononuclear cells, and polymorphonuclear neutrophils. The intensity of the inflammation varies with

Figure 85-4.  Histologic features of nonalcoholic steatohepatitis. Diffuse or perivenular macrovesicular steatosis is present. Lobular inflammation consists of neutrophils, lymphocytes, and other mononuclear cells. Hepatocyte ballooning and necrosis of varied degrees are hallmark features. Glycogenated nuclei are present. Mallory bodies, which may be small, sparse, and inconspicuous, are seen. (Hematoxylin and eosin.) (Courtesy of Dr. Gregory Y. Lauwers, Boston, Mass.)

the severity of steatohepatitis and may be milder in NASH than in alcoholic hepatitis.87 Glycogenated nuclei may be present. Hepatocyte ballooning and hepatocyte necrosis of varying degrees often are present and may portend a worse prognosis.88,89 Mallory (or Mallory-Denk) bodies, which may be small, sparse, and inconspicuous, are seen frequently. Mild stainable iron may be present in up to 50% of the patients. Pericellular, perisinusoidal, and periportal fibrosis has been described in 37% to 84% of patients with NASH. The extent of fibrosis varies considerably, ranging from delicate strands surrounding small veins or groups of cells to densely fibrotic septa with distortion of the hepatic architecture. Perisinusoidal fibrosis is most common, especially in adults, is initially mild, and predominates in zone 3 around the terminal hepatic veins.41 Cirrhosis is found on initial biopsy in 7% to 16% of patients with NAFLD and

Chapter 85  Nonalcoholic Fatty Liver Disease Table 85-3  Histologic Features of Nonalcoholic Fatty Liver Disease Present in All or Most Cases Macrovesicular steatosis Diffuse or centrilobular steatosis; degree may correlate with BMI Parenchymal inflammation Polymorphonuclear neutrophils, lymphocytes, other mononuclear cells Hepatocyte necrosis Ballooning hepatocyte degeneration Observed with Varied Frequencies Perivenular, perisinusoidal, or periportal fibrosis (37%-84%), moderate to severe in 15%-50%; most prevalent in zone 3 (perivenular) Cirrhosis (7%-16% on index biopsy specimen) Mallory bodies Glycogenated nuclei Lipogranulomas Stainable hepatic iron

Elevated serum aminotransferase levels and/or hepatomegaly

Exclude excessive alcohol use and other forms of liver disease by history and laboratory tests

Image liver with US, CT, or MRI

Normal

Fatty liver present

Liver biopsy

Consider liver biopsy to stage disease and define risk of progression

BMI, body mass index.

abnormal liver biochemical test levels.15,29 The risk of cirrhosis in the setting of NAFLD may be greatest in morbidly obese patients. In NAFLD-associated cirrhosis, the typical histologic features of NAFLD may be minimal or absent, potentially leading to the misdiagnosis of cryptogenic cirrhosis. Strict histologic criteria for NASH have not yet been defined. At least two scoring systems have been proposed.90,91 To reach consensus on the pathologic classification of NASH, the Pathology Committee of the National Institutes of Health NASH Clinical Research Network has proposed a scoring system incorporating 14 histologic features.90 The unweighted sum of scores for steatosis, lobular inflammation, and hepatocellular ballooning is used to construct an activity score, which can be used to classify cases into categories designated “not NASH,” “borderline NASH,” or “NASH” with reasonable inter-rater agreement. Thus far, however, no scoring system for NAFLD or NASH has been widely adopted.

DIAGNOSIS

Establishing a definitive diagnosis of NAFLD requires both clinical and histologic data (Fig. 85-5). Most patients with NAFLD are evaluated because of chronically elevated liver biochemical test levels, with or without hepatomegaly. The combination of the patient’s history, physical examination, blood test results, and radiologic findings is useful for excluding other causes of liver disease. Laboratory testing should include liver biochemical tests, complete blood count, prothrombin time, anti-HCV, hepatitis B surface antigen, iron indices, ceruloplasmin in persons younger than 40 years of age, α1-antitrypsin, and AMA. Imaging studies may support the diagnosis (see earlier), but the absence of characteristic findings does not preclude a diagnosis of NAFLD. To establish a diagnosis of NAFLD, alcoholic liver disease must be excluded. Clinical and histologic data unreliably differentiate NAFLD from alcoholic liver disease in ambulatory patients. Therefore, the diagnosis of NAFLD should be entertained only in the absence of significant alcohol use (consumption of less than 20 to 40 g of alcohol per day in most clinical studies).

The Role of Liver Biopsy

The role of liver biopsy in establishing the diagnosis of NAFLD has been debated. Many practitioners consider NAFLD a diagnosis of exclusion, when clinical and labora-

Figure 85-5.  Diagnostic approach to patients with suspected nonalcoholic fatty liver disease (NAFLD). The diagnosis of NAFLD is based on clinical and histologic criteria. Most patients are evaluated because of elevated serum aminotransferase levels and/or hepatomegaly. The diagnosis of NAFLD should be considered when excessive alcohol use is absent and laboratory test results exclude other causes of liver disease. Radiologic studies may demonstrate fatty liver. Liver biopsy is the standard means of diagnosis and the only test that can reliably differentiate simple steatosis from advanced NAFLD (i.e., nonalcoholic steatohepatitis), although noninvasive methods for assessing fibrosis are under study. CT, computed tomography; MRI, magnetic resonance imaging; US, ultrasonography.

tory examinations fail to reveal another cause of chronic liver disease. The diagnosis of NAFLD is also suggested when an imaging study provides evidence of fatty liver. Most patients with NAFLD do not undergo a liver biopsy, largely because the results usually will not affect management (therapeutic options are limited, see later). The argument for liver biopsy in all patients with presumed NAFLD is centered on several lines of evidence: The correlation among clinical, laboratory, and histologic findings in NAFLD is poor, and patients with normal liver biochemical test results can have significant liver injury on biopsy specimens31; liver biopsy is the only diagnostic test that can reliably quantify hepatic steatosis, necrosis, and fibrosis; and the histologic stage of NAFLD is the best prognostic indicator.92 In addition, understanding of the natural history and treatment of NAFLD have been hampered by the lack of histologic data in most clinical studies. The most prudent approach may be to select patients for whom liver biopsy might influence management. Several clinical and biochemical risk factors for progressive disease have been identified and may facilitate selection of patients for liver biopsy. Patients with obesity, longstanding or persistent liver biochemical test abnormalities (even with good glycemic control and after weight loss), older age, multiple com­ ponents of the metabolic syndrome, an AST/ALT ratio greater than 1, markedly elevated liver biochemical test levels, symptoms and signs of portal hypertension, or evidence of fibrosis on an imaging study are more likely to have advanced disease (Table 85-4).29,89 The results of a liver biopsy in these patients might lead to a more aggressive treatment strategy, participation in clinical trials, or

1407

1408

Section IX  Liver Table 85-4  Risk Factors for Advanced* Nonalcoholic Fatty Liver Disease Clinical Older age (>50 years) Obesity Diabetes mellitus/insulin resistance Hypertension Laboratory AST/ALT ratio > 1 Serum ALT level > twice the upper limit of normal Serum triglyceride levels > 155 mg/dL Histologic Severe steatosis Necroinflammatory activity (hepatocyte ballooning, necrosis) Stainable iron ALT, alanine aminotransferase; AST, aspartate aminotransferase. *Nonalcoholic steatohepatitis and advanced fibrosis.

screening for hepatocellular carcinoma in the setting of cirrhosis.

Noninvasive Markers of Fibrosis in NAFLD

Although percutaneous liver biopsy remains the standard for the diagnosis of NAFLD, it is costly, invasive, and associated with a small risk of complications. Sampling variability is common, and the large number of persons with NAFLD far outstrips the manpower available to perform liver biopsies. Significant progress has been made in developing simple, noninvasive, and quantitative tests to estimate the degree of hepatic fibrosis in a number of liver diseases, including NAFLD. The FibroTest (called FibroSure in the United States) is the best studied of these noninvasive tests (see Chapters 73 and 79). The panel of blood tests used to estimate hepatic fibrosis includes serum α2-macroglobulin, apolipoprotein A-1, haptoglobin, total bilirubin, and GGTP levels, and the necroinflammatory activity index combines the same five markers plus the serum ALT level. In a study of 167 patients with NAFLD, FibroTest was highly sensitive for detecting bridging fibrosis and cirrhosis.93 FibroTest cutoff value of .70 had a positive predictive value of 73% and a specificity of 98% for advanced fibrosis. A cutoff value of 0.30 had a negative predictive value of 90% for advanced fibrosis. Unfortunately, 33% of patients had a FibroTest score between 0.30 and 0.70, and in this range, the test is inaccurate for assessing the stage of fibrosis. Therefore, patients with a score in this range would need a liver biopsy for accurate staging. Angulo and colleagues developed and validated another noninvasive fibrosis scoring system called the NAFLD Fibrosis Score, which is derived from clinical and laboratory information that is obtained easily in the context of any clinical encounter.94 Using this scoring algorithm, which incorporates age, BMI, hyperglycemia, AST/ALT ratio, platelet count, and serum albumin level, the authors defined a low cutoff value with a negative predictive value of 88% to 93% and a high cutoff value with a positive predictive value of 82% to 90%. Only 25% to 28% of cases were indeterminate and would therefore require liver biopsy for accurate staging. Additional noninvasive tests for fibrosis have been evaluated with variable success in small studies of NAFLD, including transient elastography (Fibroscan), which uses ultrasound to quantify liver stiffness and estimate fibrosis,95 serum dehydroepiandrosterone levels,96 and serum hyaluronic acid levels.97 One or more noninvasive indices of fibrosis is likely to be validated in the future and may supplant the need for liver biopsy in many, but not all, patients with NAFLD.

NATURAL HISTORY

The natural history of NAFLD is largely unknown because no prospective, longitudinal studies with histologic followup have been conducted in patients with NAFLD. In addition, the long-term complications of NAFLD are probably under-recognized and under-reported, in part because the characteristic feature of macrovesicular steatosis may diminish or disappear in late-stage NAFLD. The available data from retrospective studies suggest that NAFLD is a benign disease in most patients. The prognosis in patients with steatosis in the absence of hepatocyte necrosis and fibrosis clearly is favorable, with little potential for histologic or clinical progression.98 In some patients, however, NAFLD can lead to cirrhosis, liver failure, or hepatocellular carcinoma. Between 2003 and 2008, five retrospective studies were published that included a total of 215 patients with NAFLD who underwent paired liver biopsies.6,99-101 Follow-up periods ranged from 1 to 14 years, and histologic progression of fibrosis was documented in 32% to 41% of the patients. NAFLD was stable histologically in 34% to 50% and appeared to improve in 16% to 29%. Substantial variability in the rate of progression of fibrosis was observed among patients; histologic progression did not signal clinical deterioration in most cases, and no clinical or laboratory data reliably predicted the course of liver disease. These data must be interpreted with caution because variability in histologic staging may be the result of sampling error.102 A large population-based study from Olmsted County, Minnesota, included 420 patients with definite or presumed NAFLD based on imaging studies or, less frequently, liver biopsy findings.103 Cirrhosis developed in 3% of the patients over a mean follow-up of 7.6 years. Liver disease was the third leading cause of death in this cohort, behind malignancy and ischemic heart disease. These limited data suggest that NAFLD is an indolent condition, with few clinical sequelae in most patients, but can progress to irreversible, clinically important liver disease over a relatively short period of time in a minority of affected persons. This conclusion is supported by data from a classic study of 132 patients with NAFLD evaluated over an 18-year period. The study compared clinical outcomes based on the degree of injury on an index liver biopsy specimen.104 Each biopsy specimen was assigned to a histologic subgroup—types 1 to 4—that represented progressively severe disease as determined by the cumulative occurrence of steatosis, inflammation, hepatocyte necrosis, and fibrosis. Cirrhosis and liver-related deaths were more common (25% and 11%, respectively) in patients with NAFLD types 3 and 4 (steatosis, inflammation, and necrosis, with or without fibrosis) than in patients with NAFLD types 1 and 2 (steatosis without necrosis), although the mortality difference failed to reach statistical significance. In addition, the liver-related mortality rate of 11% among patients with NAFLD types 3 and 4 was higher than the age-adjusted death rate from chronic liver disease/cirrhosis in the general U.S. population (9.5 per 100,000 per year). These data suggest that the risk of liver-related complications in NAFLD correlates, at least to some extent, with the degree of hepatocellular injury and fibrosis found on an index liver biopsy specimen. NAFLD and alcoholic hepatitis are similar histologically, but differ substantially in clinical outcomes (see Chapter 84). The five-year survival rate of patients with alcoholic hepatitis is only 50% to 75% because of the large proportion of patients (greater than 50%) in whom cirrhosis and its complications develop. A study has shown that the longterm survival of patients with NASH is significantly better than the long-term survival of patients with alcoholic hepatitis.86 In the minority of patients in whom NAFLD-

Chapter 85  Nonalcoholic Fatty Liver Disease associated cirrhosis develops, however, the outcome may be similar to that for other causes of cirrhosis. Strong circumstantial evidence suggests that NAFLD is the likely cause of many cases of cryptogenic cirrhosis105-107 and may be associated with the development of hepatocellular carcinoma.30,106,108 One study showed that the 5- to 10-year outcome of NAFLD-associated cirrhosis was similar to that for HCV-associated cirrhosis, although hepatocellular carcinoma was significantly less common in the patients with NAFLD.109 Large prospective studies are needed to define the natural history of NAFLD, but emerging evidence confirms that NAFLD can be progressive and associated with significant morbidity and mortality in some patients. The risks of liverrelated morbidity and mortality are greatest in persons with evidence of advanced NAFLD (steatohepatitis with necrosis and fibrosis) on the initial liver biopsy specimen.104 If clinical and biochemical risk factors for progressive disease can be established, a subset of patients can be identified in whom a liver biopsy will have the greatest prognostic and therapeutic value (see earlier). Several potential risk factors have been identified in different populations (see Table 85-4). Women are over-represented in studies of NAFLD, but whether gender is an independent risk factor for advanced disease is unclear. Older age, obesity, diabetes mellitus, and an AST/ALT ratio greater than 1 were demonstrated in one study to be significant predictors of severe fibrosis (bridging/cirrhosis) in patients with NAFLD.29 In another study of overweight patients with abnormalities on liver biochemical tests, liver fibrosis was independently associated with hepatic necroinflammatory activity, BMI greater than 28 kg/m2, age older than 50 years, serum triglyceride level higher than 1.7 mmol/L, and serum ALT more than twice normal.43 In another study of morbidly obese patients referred for bariatric surgery, systemic hypertension, an elevated serum ALT level, and a high insulin resistance index were highly predictive of advanced NAFLD.15 As noted previously, however, the full spectrum of liver damage, including cirrhosis, has been documented in nonobese patients with near-normal liver biochemical test results.

TREATMENT

The optimal therapy for NAFLD has not been established. To date, no large, randomized treatment trials demonstrating resolution of steatosis, inflammation, and fibrosis have been conducted in patients with NAFLD. Small numbers of patients, varying inclusion criteria, and varying end points have limited the clinical impact of published studies. Historically, the treatment of NAFLD has consisted of weight loss, removal of offending drugs and toxins, and control of associated metabolic disorders, including diabetes mellitus and hyperlipidemia. Several case reports and small studies of diet and exercise have shown improvements in biochemical, ultrasonographic, and in some cases, histologic abnormalities in children and adults with NASH.52,110-112 Intensive nutritional counseling may lead to sustained weight loss and significant histologic improvement in some patients.113 Several small, largely uncontrolled studies also showed improvements in liver biochemical test results, steatosis, and fibrosis in a few patients who achieved modest weight loss with orlistat, a reversible inhibitor of gastric and pan­ creatic lipases.114,115 Bariatric surgery leads to massive weight loss and improves insulin sensitivity in most patients, normalizes some of the metabolic abnormalities involved in the pathogenesis of NAFLD, decreases the hepatic expression of mediators of liver inflammation and fibrosis, and improves hepatic histology in patients with NAFLD.116-120

Table 85-5  Potential Therapies for Nonalcoholic Fatty Liver Disease Avoidance of toxins Discontinue potentially offending medications/toxins Minimize alcohol intake Exercise and diet Moderate, sustained exercise and weight loss in overweight patients Effects of specific diets are not known Antidiabetic/insulin-sensitizing agents Metformin Thiazolidinediones Lipid-lowering agents Gemfibrozil Statins Antioxidants Betaine N-acetylcysteine Superoxide dismutase Vitamin E Iron reduction by phlebotomy Inflammatory mediators by: Agents that affect increasing mitochondrial ATP stores and/or activity Agents that affect modulating leptin activity Agents that affect modulating TNF-α activity Agents that affect raising adiponectin levels Bariatric surgery for morbid obesity ATP, adenosine triphosphate; TNF-α, tumor necrosis factor-α.

No scientific data exist to support a particular commercial or medicinal diet plan for NAFLD, and no one correct dietary approach is likely to be suitable for all patients with NAFLD.121 Unfortunately, the therapeutic benefit of weight loss achieved with diet, medicinal aids, exercise, or surgery has not been examined in randomized, prospective studies with firm histologic end points. Until such studies are performed, a recommendation for moderate weight loss is reasonable in overweight patients with NAFLD, although sustained weight loss is seldom achieved. Rapid weight loss can exacerbate steatohepatitis in morbidly obese patients, especially after bariatric surgery122; therefore, the rate of weight loss and serial liver biochemical test results should be monitored carefully in patients on a weight reduction regimen. New therapeutic methods should capitalize on today’s improved understanding of the pathogenesis of NAFLD (Table 85-5).

Antioxidants

Medications that minimize oxidative stress may prove useful. Vitamin E, an inexpensive yet potent antioxidant, has been examined as an agent for treatment of NAFLD in several small pediatric and adult studies, with varying results.52,123,124 In all studies, vitamin E was well tolerated, and most studies showed modest improvements in serum aminotransferase levels, ultrasonographic appearance of the liver, and, infrequently, histologic findings. Randomized controlled studies with histologic inclusion criteria and end points are needed, however, to determine if vitamin E, either alone or in combination with other medications, leads to histologic improvement in NAFLD. In light of the potentially negative effects of vitamin E on cardiovascular health,125 caution should be exercised in treating NAFLD with vitamin E until better studies are available. Betaine, a metabolite of choline that raises SAM levels and decreases cellular oxidative damage, has shown promise in a small pilot study as a therapeutic agent for NASH.126 N-acetylcysteine, superoxide dismutase, and PPAR-α

1409

1410

Section IX  Liver agonists such as ragaglitazar also may hold therapeutic promise,64,127,128 although clinical studies in humans are lacking.

Insulin-Sensitizing Agents

The association between hyperinsulinemic insulin resistance and NAFLD provides a logical target for treatment. Metformin, a biguanide that reduces hyperinsulinemia and improves hepatic insulin sensitivity, reduces hepatomegaly and hepatic steatosis in ob/ob mice,129 but results in human studies have been less impressive.130,131 Thiazolidinediones (TZDs), potent PPAR-α agonists, also are being investigated as possible agents for the treatment of NAFLD. PPAR-α is a nuclear receptor expressed in adipose tissue, muscle, and liver. In adipocytes, PPAR-α promotes cell differentiation and decreases lipolysis and FFA release. TZDs improve insulin sensitivity and hyperinsulinemia by increasing glucose disposal in muscle and decreasing hepatic glucose output. Treatment with troglitazone, a first-generation TZD, was associated with biochemical and histologic improvements in patients with NASH, but troglitazone subsequently was withdrawn from the market because of rare but serious hepatotoxicity. Rosiglitazone and pioglitazone, TZDs with low rates of hepatotoxicity, have been investigated in separate 48-week, single-arm treatment trials in patients with histologically proven NASH.132,133 In both studies, treatment was well tolerated and was associated with improved insulin sensitivity, normalization of liver biochemistries, and histologic improvement in most patients. A drawback of both TZDs, however, was substantial weight gain (4.0% to 7.3%) and increased total body adiposity. In addition, the durability of biochemical and histologic improvements after completion of therapy was not examined in either study. A follow-up study has suggested, however, that the beneficial effects of pioglitazone diminish when the drug is discontinued.134 A placebo-controlled European trial of rosiglitazone in 63 patients with NAFLD showed improvement in steatosis and serum aminotransferase levels but not in necroinflammation or fibrosis.135 Rosiglitazone was also associated with significant weight gain in the TZD-treated patients in this study. TZDs must be assessed in large, placebocontrolled trials before they can be recommended for routine use in patients with NAFLD.

untreated control subjects, but histologic features were not assessed.139 Treatment of NASH with atorvastatin, a 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitor, showed promise in a small pilot study,140 but statins have not been assessed in a large clinical trial. Ursodeoxycholic acid, a cytoprotective agent, showed promise in a pilot study of NASH but was not effective in a randomized placebo-controlled trial.141 The combination of ursodeoxycholic acid and vitamin E has shown some efficacy. Future therapies for NAFLD might include agents that increase adiponectin levels, neutralize TNF-α, improve mitochondrial ATP homeostasis, or alter leptin levels. When the pathogenesis of NAFLD is elucidated further, new therapeutic agents will likely be developed.

Liver Transplantation

Patients with NAFLD in whom end-stage liver disease develops should be evaluated for liver transplantation. The outcome of liver transplantation in these patients is good, although NAFLD can recur after liver transplantation.142,143 The risk factors for recurrent or de novo NAFLD after liver transplantation probably are multifactorial and include hypertriglyceridemia, obesity, diabetes mellitus, and glucocorticoid therapy.

FOCAL FATTY LIVER In contrast with NAFLD, which is a diffuse parenchymal process, focal fatty liver is a localized or patchy process that simulates a space-occupying lesion in the liver on imaging studies. This condition has been recognized increasingly in adults and children as a result of the improved sensitivity of abdominal imaging. Focal fatty liver has characteristic patterns on CT: usually a nonspherical shape, absence of mass effect, and CT attenuation values consistent with those of soft tissue.144 The density of focal fatty liver is close to that of water, unlike that of liver metastases, which have a density that is closer to that of hepatocytes. Ultrasonography and MRI can help confirm a diagnosis of focal fatty liver (Fig. 85-6). A presumptive diagnosis of focal fatty liver

Iron Reduction

High serum iron and ferritin levels have been identified in some patients with NAFLD, most of whom do not have genetic hemochromatosis or hepatic iron overload. Most investigators believe that increased serum iron indices are a by-product of hepatic inflammation, rather than a contributor to the pathogenesis of NAFLD, but a few small studies have suggested that iron depletion may have a therapeutic role in NAFLD by decreasing plasma insulin, glucose, and serum aminotransferase levels.136,137 The relationship between iron and insulin is complex, but the insulinsparing effect of iron depletion may be the result of enhanced skeletal muscle glucose transport and metabolism and increased hepatic extraction and metabolism of insulin.138 The primary limitation of these studies is the lack of histologic inclusion criteria and end points, but the results are intriguing and merit further investigation.

Lipid-Lowering and Cytoprotective Agents

The usefulness of lipid-lowering and cytoprotective drugs for the treatment of NAFLD has been assessed in a few small trials, with varying results. Treatment with gemfibrozil was associated with biochemical improvement in 74% of patients in the treatment group, compared with 30% of

Figure 85-6.  Focal fatty liver. Focal fatty liver (arrow) on computed tomography (CT) scan. The characteristic features are the nonspherical shape, absence of a mass effect, and CT attenuation values consistent with those of soft tissue. Ultrasound studies and magnetic resonance imaging also may confirm the diagnosis of focal fatty liver. (Courtesy of Dr. Mukesh Harisinghani, Boston, Mass.)

Chapter 85  Nonalcoholic Fatty Liver Disease should not be made when a mass effect, areas of mixed hypo- and hyperechogenicity, an irregular shape, or a history of malignancy is present. In such cases, ultrasonographically guided fine-needle biopsy is recommended. No evidence exists to suggest that the pathogenesis of focal fatty liver is similar to that of NAFLD. In fact, the pathogenesis of focal fatty liver is uncertain and may involve altered venous blood flow to the liver, tissue hypoxia, and intestinal malabsorption of lipoproteins. Furthermore, in the absence of accompanying or background liver disease, the lesion often regresses. Therefore, no specific treatment is justified.

KEY REFERENCES

Adams LA, Lymp JF, St. Sauver J, et al. The natural history of nonalcoholic fatty liver disease: A population-based cohort study. Gastroenterology 2005; 129:113-21. (Ref 103.) Adams LA, Sanderson S, Lindor KD, Angulo P. The histological course of nonalcoholic fatty liver disease: A longitudinal study of 103 patients with sequential liver biopsies. J Hepatol 2005; 42:132-8. (Ref 99.) Angulo P, Hui JM, Marchesini G, et al. The NAFLD fibrosis score: A noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology 2007; 45:846-54. (Ref 94.) Day C, James O. Steatohepatitis: A tale of two “hits”? [editorial]. Gastroenterology 1998; 114:842-5. (Ref 36.)

Dixon J, Bhathal P, O’Brien P. Nonalcoholic fatty liver disease: Predictors of nonalcoholic steatohepatitis and liver fibrosis in the severely obese. Gastroenterology 2001; 121:91-100. (Ref 15.) El-Serag H, Tran T, Everhart J. Diabetes increases the risk of chronic liver disease and hepatocellular carcinoma. Gastroenterology 2004; 126:460-8. (Ref 30.) Farrell GC, Larter CZ. Nonalcoholic fatty liver disease: From steatosis to cirrhosis. Hepatology 2006; 43:S99-S112. (Ref 47.) Hui J, Hodge A, Farrell G, et al. Beyond insulin resistance in NASH: TNF-alpha or adiponectin? Hepatology 2004; 40:46-54. (Ref 51.) Lin H, Yang S, Chuckaree C, et al. Metformin reverses fatty liver disease in obese, leptin-deficient mice. Nature Med 2000; 6:998-1003. (Ref 129.) Marchesini G, Bugianesi E, Forlani G, et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 2003; 37:917-23. (Ref 28.) Matteoni C, Younossi Z, Gramlich T, et al. Nonalcoholic fatty liver disease: A spectrum of clinical and pathological severity. Gastroenterology 1999; 116:1413-19. (Ref 104.) Promrat K, Lutchman G, Uwaifo G, et al. A pilot study of pioglitazone treatment for nonalcoholic steatohepatitis. Hepatology 2004; 39:18896. (Ref 133.) Ratziu V, Giral P, Charlotte F, et al. Liver fibrosis in overweight patients. Gastroenterology 2000; 118:1117-23. (Ref 89.) Ratziu V, Giral P, Jacqueminet S, et al. Rosiglitazone for nonalcoholic steatohepatitis: One-year results of the randomized placebocontrolled Fatty Liver Improvement with Rosiglitazone Therapy (FLIRT) Trial. Gastroenterology 2008; 135:100-10. (Ref 135.) Full references for this chapter can be found on www.expertconsult.com.

1411

CHAPTE R

86 Liver Disease Caused by Drugs Narci C. Teoh, Shivakumar Chitturi, and Geoffrey C. Farrell

CHAPTER OUTLINE Definitions and Importance  1413 Epidemiology  1414 Case Definition: Which Agent?  1414 Frequencies of Hepatic Drug Reactions  1414 Importance of Drugs as a Cause of Liver Disease  1415 Risk Factors  1415 Pathophysiology  1417 Role of the Liver in Drug Elimination  1417 Pathways of Drug Metabolism  1417 Toxic Mechanisms of Liver Injury  1419 Immunologic Mechanisms  1422 Clinicopathologic Features of Drug-Induced Liver Disease  1422 Classification  1422 Histopathologic Features  1424 Clinical Features  1424 A Practical Approach to Diagnosis  1424 Physician Awareness  1424 Exclusion of Other Disorders  1425 Extrahepatic Features  1425 Chronologic Relationships  1425 Which Drug?  1425 Indications for Liver Biopsy  1425 Considerations in Patients with Viral Hepatitis  1425 Prevention and Management  1426 Dose-Dependent Hepatotoxicity  1427

DEFINITIONS AND IMPORTANCE Drugs are a relatively common cause of liver injury, which usually is defined by abnormalities of liver biochemical test levels, particularly an increase in the serum alanine aminotransferase (ALT), alkaline phosphatase, or bilirubin level, to more than twice the upper limit of normal. Drug-induced liver injury can be difficult to define in clinical practice because the biochemical tests used to detect liver injury may also be elevated as part of an adaptive response to drugs. Further, the severity of drug-induced liver injury varies from minor nonspecific changes in hepatic structure and function to acute liver failure, cirrhosis, and liver cancer. The term drug-induced liver disease should be confined to cases in which the nature of liver injury has been characterized histologically. With the exception of acetaminophen, anticancer drugs, and some botanical or industrial hepatotoxins (see Chapter 87), most cases of drug-induced

Acetaminophen  1427 Other Types of Cytopathic Liver Injury  1429 Drug-Induced Acute Hepatitis  1431 Immunoallergic Reactions  1431 Metabolic Idiosyncrasy  1434 Drug-Induced Granulomatous Hepatitis  1437 Drug-Induced Chronic Hepatitis  1438 Diclofenac  1439 Minocycline  1439 Drug-Induced Acute Cholestasis  1439 Importance, Types of Reactions, and Diagnosis  1439 Cholestasis without Hepatitis  1439 Cholestasis with Hepatitis  1440 Cholestatic Hepatitis with Bile Duct Injury  1441 Drug-Induced Chronic Cholestasis  1441 Flucloxacillin  1441 Fibrotic Bile Duct Strictures  1441 Drug-Induced Steatohepatitis and Hepatic Fibrosis  1441 Amiodarone  1442 Tamoxifen and Other Causes of Drug-Induced Steatohepatitis  1442 Methotrexate  1443 Drug-Induced Vascular Toxicity  1444 Azathioprine  1445 Liver Tumors  1445

liver disease represent adverse drug reactions or hepatic drug reactions. These effects are noxious and unintentional and occur at doses recommended for prophylaxis or therapy. The latent period is longer (typically one week to three or six months) than that for direct hepatotoxins (hours to a few days), and extrahepatic features of drug hypersensitivity may be present. Although drug-induced liver disease is a relatively uncommon cause of jaundice or acute hepatitis in the community, it is an important cause of more severe types of acute liver disease, particularly among older people (see Epidemiology). The overall mortality rate among patients hospitalized for drug-induced liver injury is approximately 10%1 but varies greatly for individual drugs.2,3 The reported frequencies of individual hepatic drug reactions are often underestimated because of the inadequacy of spontaneous reporting by physicians and pharmacists.2,3 With more reliable prospective and epidemiologic techniques, the frequency (or risk) of most types of drug-induced liver disease is between one per 10,000 and one per 100,000 persons

1413

1414

Section IX  Liver exposed.4 Because these responses to drug exposure are clearly rare and unpredictable, they are often termed idiosyncratic drug reactions. Their rarity blunts diagnostic acumen because most clinicians will see few, if any, cases and therefore do not have an appropriate level of clinical suspicion. This concern applies especially to comple­ mentary and alternative medicine (CAM), as discussed in Chapter 87. Failure to withdraw the causative agent after the onset of symptoms of drug hepatitis or inadvertent reexposure to such a drug is a common and avoidable factor in acute liver failure attributable to drug-induced liver injury.5-8 Another challenge is that hepatic drug reactions produce an array of clinical syndromes and pathologic findings that mimic known hepatobiliary diseases. Furthermore, although individual agents (and some drug classes) typically produce a characteristic “signature syndrome,” they can also be associated with other and sometimes multiple clinicopathological syndromes. Drug-induced liver injury is the most common reason for withdrawal of an approved drug from the market. The subject therefore has medicoeconomic, legal, and regulatory ramifications. Because of the low frequency of most types of idiosyncratic drug reactions that involve the liver, serious hepatotoxicity is not usually detected until post-marketing surveillance is conducted. Historically, drugs that have developed a reputation for potential hepatotoxicity usually have been replaced by more acceptable alternatives. Examples include troglitazone, the prototypic thiazolidine­ dione and bromfenac, a nonsteroidal anti-inflammatory drug (NSAID), both of which were withdrawn from the market because of several cases of fatal acute liver failure.5-9 The burgeoning number of available conventional medications and CAM preparations now includes many hundreds that can be cited as rare causes of drug-induced liver disease. This increasing number of potentially causative agents poses several challenges to clinicians,5-10 including concern about what constitutes an adequate level of patient information at the time a drug is prescribed and the reliability of evidence linking an individual agent to a particular type of liver injury.5,11-13 Another development is the appreciation that drug toxicity, in the context of complex medical situations, can interact with other causes of liver injury. Noteworthy examples of such situations are bone marrow transplantation; cancer chemotherapy; highly active antiretroviral therapy (HAART) for human immunodeficiency virus (HIV) infection and the acquired immunodeficiency syndrome (AIDS); use of antituberculosis drugs in patients with chronic viral hepatitis; rifampin hepatitis in patients with primary biliary cirrhosis (Chapter 89); nonalcoholic fatty liver disease (NAFLD)— particularly nonalcoholic steatohepatitis (NASH)—preci­ pitated by tamoxifen; and possibly other drugs in overweight persons with type 2 diabetes mellitus and the metabolic syndrome.

EPIDEMIOLOGY Frequency or risk—the number of adverse reactions for a given number of persons exposed—is the best term for expressing how common a drug reaction is. Timedependent terms such as incidence and prevalence are not appropriate for drug reactions because the frequency is not linearly related to the duration of exposure. For most reactions, the onset occurs within a relatively short exposure time, or latent period, although some rarer types of

chronic liver disease occur after many months or years. The frequency of drug-induced liver disease is usually based on the reported rate of drug reactions; such reports are usually a voluntary part of post-marketing surveillance and are submitted to pharmaceutical companies or adverse drug reaction monitoring bodies. In the United States, following approval by the U.S. Food and Drug Administration (FDA), pharmaceutical companies are required to report serious adverse events (any incident resulting in death, a threat to life, hospitalization, or permanent disability [Code of Federal Regulations]). Surveillance becomes a more passive process, however, when a drug is approved for marketing, and physicians and pharmacists are encouraged to file voluntary written reports through the MediWatch program. Similar systems operate in most industrialized countries. Nevertheless, MediWatch receives reports for fewer than 10% of adverse drug reactions,2 and in France fewer than 6% of hepatic adverse drug reactions are reported.3 The situation may be somewhat better in Sweden, but the annual reported incidence of adverse drug reactions of 2.2 per 100,000 in the population over the age of 15 is still much lower than the predicted incidence of 14 per 100,000.3 A prospective surveillance study in Spain measured the annual incidence of drug-related acute serious liver disease as 7.4 per million inhabitants.4

CASE DEFINITION: WHICH AGENT?

At least 300 agents have been implicated in drug-induced liver injury.10 The evidence for most drugs, however, is confined to individual or small numbers of case reports, especially in letters to scientific journals or to regulatory authorities, or small observational series. Therefore, for most agents, the evidence that they could cause liver injury is circumstantial and incomplete. Reports often lack pathologic definition, full exclusion of other disorders (for older reports), and logistic imputation of causality, especially with respect to temporal associations (see Diagnosis).5,9,10 Overall, probably fewer than 50 agents have been implicated reliably as causes of drug-induced liver disease. In general, agents used most commonly in clinical practice and in the community, including antimicrobials, antineoplastic agents, and NSAIDs, are those that have been implicated most often in drug-induced liver injury in larger series. The challenge of identifying the culprit drug among multiple candidates is discussed later.1,4,5,6,11

FREQUENCIES OF HEPATIC DRUG REACTIONS

Because of incomplete reporting, frequencies of hepatic drug reactions may often be underestimated. These estimated frequencies are also crude indicators of risk because of the inherent inaccuracies of case definitions (see Diagnosis)5,9,10 and because case recognition and reporting depend on the skill and motivation of observers. The increased interest of prescribers when initial cases of drug-induced liver disease have been described, together with inappropriate prescribing (e.g., prolonged use of bromfenac, which was approved only for seven days of use, and overprescribing of flucloxacillin and amoxicillin-clavulanic acid in some countries) can give rise to apparent “mini-epidemics.” More appropriate epidemiologic methods applied to hepatotoxicity have included prescription event monitoring, record linkage, and case-control studies. Prescription event monitoring and record linkage have been used to estimate the frequency of liver injury with some antimicrobials (erythromycins, sulfonamides, tetracyclines, flucloxacillin, amoxicillin-clavulanate) and NSAIDs.11 Epidemiologic studies confirm the rarity of drug-induced liver disease with currently used agents. For NSAIDs, the

Chapter 86  Liver Disease Caused by Drugs risk of liver injury is between 1 and 10 per 100,000 individuals exposed1,4,5,6; amoxicillin-clavulanic acid has been associated with cholestatic hepatitis in 1 to 2 per 100,000 exposed persons1,4,5,6,8; and low-dose tetracyclines have caused hepatotoxicity in less than one case per million persons exposed.1,4-6 The frequency of liver injury may be higher for agents that exert a metabolic type of hepatotoxicity. For example, isoniazid causes liver injury in up to 2% of persons exposed; the risk depends on the patient’s age and gender, concomitant exposure to other agents, and presence of hepatitis B virus (HBV) and possibly hepatitis C virus (HCV) infections.12 For some drugs in which other host factors play an etiopathogenic role, casecontrol studies have been used to define attributable risk. Examples include the implication of aspirin in Reye’s syndrome and oral contraceptives in liver tumors and hepatic vein thrombosis. A relationship may exist between the frequency and severity of serum ALT elevations that indicate liver injury and the risk of severe hepatotoxicity. This relationship was proposed in the 1970s by the late Hyman Zimmerman.6 According to “Hy’s rule,” elevations of serum ALT levels to eight-fold or more above the upper limit of normal or associated increases in the serum bilirubin concentration indicate a potential for the drug to cause acute liver failure at a rate of about 10% of the number of cases of jaundice. Therefore, if two cases of jaundice associated with druginduced liver injury are observed in a total phase 3 clinical trial experience of 2500 patients, approximately one case of acute liver failure would be expected for every 12,500 subjects who were prescribed the drug during the marketing phase.

IMPORTANCE OF DRUGS AS A CAUSE OF LIVER DISEASE

Hepatotoxicity accounts for less than 5% of cases of jaundice or acute hepatitis in the community and for even fewer cases of chronic liver disease5,6; however, drugs are an important cause of more severe types of liver disease and for liver disease in older people. They account for 10% of cases of severe hepatitis admitted to the hospital in France6 and for 43% of cases of hepatitis among patients 50 years of age or older.7 Drugs account for more than half of the cases of acute liver failure referred to special units in the United States7 and between 20% and 75% of cases of acute liver failure in other industrialized countries.4,7 The pattern of agents incriminated varies among countries; for example, herbal medicines are a relatively more common cause in Asian countries than in other countries (Chapter 87). In most cases of drug-related liver injury, drugs are the sole cause of hepatic damage. In other cases, drugs increase the relative risk for types of liver disease that may occur in the absence of drug exposure. Examples include salicylates in Reye’s syndrome, oral contraceptive steroids in hepatic venous thrombosis, methotrexate in hepatic fibrosis associated with alcoholic liver disease and NAFLD, and tamoxifen in NASH. Predisposition of patients with preexisting liver disease to drug-induced injury is minimal, but some interesting potential interactions between chronic HCV infection and several groups of drugs and between chronic HBV infection and antituberculosis chemotherapy are now reasonably established, as discussed later. On the other hand, liver failure may be more likely to develop if the patient with a hepatic drug reaction (e.g., to amoxicillin-clavulanic acid), which usually is associated with a good outcome, has underlying chronic liver disease.

RISK FACTORS

For dose-dependent hepatotoxins such as acetaminophen and methotrexate and for some idiosyncratic reactions that are partly dependent on dose (e.g., bromfenac, tetracyclines, dantrolene, tacrine, oxypenicillins), the factors that influence the risk of drug-induced liver disease include the dose of the drug, blood level of the drug, and duration of intake. For idiosyncratic reactions, however, host determinants appear to be central to liver injury. The most critical determinant is likely to be genetic predisposition, but other “constitutional” and environmental factors can influence the risk of liver injury, as summarized in Table 86-1. The most important factors are age,4 gender, exposure to other substances, a history or family history of previous drug reactions, other risk factors for liver disease, and concomitant medical disorders.

Genetic Factors

Genetic determinants predispose to drug-induced liver disease,13 as they do for other types of drug reaction, such as penicillin allergy. Atopic patients have been thought to have an increased risk of some types of drug hepatitis, but this increase in risk has not been proved. Genetic factors determine the activity of drug-activating and antioxidant pathways, encode pathways of canalicular bile secretion, and modulate the immune response, tissue stress responses, and cell death pathways (see Chapter 72). Documented examples of drugs associated with a familial predisposition to adverse hepatic drug reactions are few and include valproic acid and phenytoin.5,8,13 Inherited mitochondrial diseases are a risk factor for valproic acid–induced hepatotoxicity.14 Some forms of drug-induced liver disease, particularly drug-induced hepatitis and granulomatous reactions, can be associated with the reactive metabolite syndrome (see later). Weak associations have been reported between specific human leukocyte antigen (HLA) haplotypes and some types of drug-induced liver disease. Andrade and colleagues13 found positive associations between the class II HLA haplotype and cholestatic or mixed liver damage for some drugs. They suggested that no specific HLA allele predisposed to the overall risk of drug-induced liver disease but that the pattern of liver injury could be influenced by these genetic determinants. Other investigators have found stronger associations between the HLA haplotype and cholestatic reactions to amoxicillin-clavulanic acid and ticlopidine (see later).13,15

Age

Most hepatic drug reactions are more common in adults than in children. Exceptions include valproic acid hepatotoxicity, which is most common in infants younger than three years of age and rare in adults, and Reye’s syndrome, in which salicylates play a key role.16,17 As discussed later, both may be examples of mitochondrial toxicity.14 In adults, the risk of isoniazid-associated hepatotoxicity is greater in persons older than 40 years of age. Similar observations have been made for nitrofurantoin, halothane, etretinate, diclofenac, and troglitazone.1,4,5,10 The increased frequency of adverse drug reactions in older subjects is largely the result of increased exposure, the use of multiple agents, and altered drug disposition.1,4,5,10 In addition, clinical severity of hepatotoxicity increases strikingly with age, as exemplified by fatal reactions to isoniazid and halothane.

Gender

Women are particularly predisposed to drug-induced hepatitis, a difference that cannot be attributed simply to

1415

1416

Section IX  Liver Table 86-1  Factors Influencing the Risk of Liver Diseases Caused by Drugs FACTOR

EXAMPLES OF DRUGS AFFECTED

INFLUENCE

Age

Isoniazid, nitrofurantoin, halothane, troglitazone Valproic acid, salicylates

Age >60 years: increased frequency, increased severity More common in children

Gender

Halothane, minocycline, nitrofurantoin

More common in women, especially with chronic hepatitis More common in men

Amoxicillin-clavulanic acid, azathioprine Dose

Acetaminophen, aspirin; some herbal medicines Tetracycline, tacrine, oxypenicillins Methotrexate, vitamin A

Blood levels are directly related to the risk of hepatotoxicity Idiosyncratic reactions, but partial relationship to dose Total dose, dosing frequency, duration of exposure is related to the risk of hepatic fibrosis

Genetic factors

Halothane, phenytoin, sulfonamides Amoxicillin-clavulanic acid Valproic acid

Multiple cases in families Strong HLA association Familial cases, association with mitochondrial enzyme deficiencies

History of other drug reactions

Isoflurane, halothane, enflurane Erythromycins Diclofenac, ibuprofen, tiaprofenic acid Sulfonamides, COX-2 inhibitors

Instances of cross-sensitivity have been reported among members of each drug class but are rare

Other drugs

Acetaminophen Valproic acid Anticancer drugs

Isoniazid, zidovudine, and phenytoin lower dose threshold and increase severity of hepatotoxicity Other antiepileptics increase risk of hepatotoxicity Interactive vascular toxicity

Acetaminophen hepatotoxicity Isoniazid, methotrexate

Lowered dose threshold, poorer outcome Increased risk of liver injury, hepatic fibrosis

Halothane, troglitazone, tamoxifen, methotrexate Acetaminophen

Increased risk of liver injury; hepatic fibrosis Increased risk of hepatotoxicity

Hycanthone, pemoline Antituberculosis drugs, ibuprofen

Increased risk of liver injury Increased risk of liver injury with chronic hepatitis B and C

Methotrexate Sulfonamides Tetracycline, methotrexate Azathioprine, thioguanine, busulfan

Increased Increased Increased Increased

Excessive alcohol use Nutritional status:   Obesity   Fasting Preexisting liver disease

Other diseases/conditions:   Diabetes mellitus   HIV infection/AIDS   Renal failure   Organ transplantation

risk of hepatic fibrosis risk of hypersensitivity risk of liver injury, hepatic fibrosis risk of vascular toxicity

AIDS, acquired immunodeficiency syndrome; COX-2, cyclooxygenase-2; HIV, human immunodeficiency virus; HLA, human leukocyte antigen.

increased exposure. Examples include toxicity caused by halothane, nitrofurantoin, sulfonamides, flucloxacillin, minocycline, and troglitazone.5,6 Drug-induced chronic hepatitis caused by nitrofurantoin, diclofenac, or minocycline has an even more pronounced female preponderance.6 Conversely, equal sex frequency or even male preponderance is common for some drug reactions characterized by cholestasis, as for amoxicillin-clavulanic acid. Azathioprine-induced liver disease is more likely to develop in male renal transplant recipients than in female recipients.18

Concomitant Exposure to Other Agents

Patients who are taking multiple drugs are more likely to experience an adverse reaction than those who are taking one agent.5,9,10 The mechanisms include enhanced cytochrome P450 (CYP)-mediated metabolism of the second drug to a toxic intermediate (see later). Examples discussed later include toxicity caused by acetaminophen, isoniazid, valproic acid, other anticonvulsants, and anticancer drugs. Alternatively, drugs may alter the disposition of other agents by reducing bile flow or competing with canalicular pathways for biliary excretion (phase 3 drug elimination) (see later). This mechanism may account for apparent interactions between oral contraceptive steroids and other drugs to produce cholestasis. Drugs or their metabolites may also interact in mechanisms of cellular toxicity and cell death that involve mitochondrial injury, intracellular signaling

pathways, activation of transcription factors, and regulation of hepatic genes involved in controlling the response to stress and injury that triggers pro-inflammatory and cell death processes.19,20

Previous Drug Reactions

A history of an adverse drug reaction generally increases the risk of reactions to the same drug and also to some other agents (see later). Nevertheless, instances of cross-sensitivity to related agents in cases of drug-induced liver disease are surprisingly uncommon. Examples of cross-sensitivity between drugs (or drug classes) include the haloalkane anesthetics (see Chapter 87), erythromycins, phenothiazines and tricyclic antidepressants, isoniazid and pyrazinamide, sulfonamides and other sulfur-containing compounds (e.g., some clyclooxygenase-2 [COX-2] inhibitors), and some NSAIDs. A crucial point is that a previous reaction to the same drug is a major risk factor for an increase in the severity of drug-induced liver injury.6

Alcohol

Chronic excessive alcohol ingestion decreases the dose threshold for, and enhances the severity of, acetaminophen-induced hepatotoxicity and increases the risk and severity of isoniazid hepatitis, niacin (nicotinic acid, nicotinamide) hepatotoxicity, and methotrexate-induced hepatic fibrosis.

Chapter 86  Liver Disease Caused by Drugs Nutritional Status

Obesity is strongly associated with the risk of halothane hepatitis (see Chapter 87) and appears to be an independent risk factor for NASH and hepatic fibrosis in persons taking methotrexate or tamoxifen. Fasting also predisposes to acetaminophen hepatotoxicity,21 and a role for undernutrition has been proposed in isoniazid hepatotoxicity.22

Preexisting Liver Disease

In general, liver diseases such as alcoholic cirrhosis and cholestasis do not predispose to adverse hepatic reactions. Exceptions include toxicity to some anticancer drugs, niacin, pemoline, and hycanthone. Preexisting liver disease is a critical determinant of methotrexate-induced hepatic fibrosis (discussed later). Patients with chronic HBV infection12 and possibly those with chronic HCV infection or HIV/AIDS appear to be at heightened risk of liver injury during anti-tuberculosis or HAART therapy,23 after exposure to ibuprofen and possibly other NSAIDs, after myeloablative therapy in preparation for bone marrow transplantation (resulting in sinusoidal obstruction syndrome [see later]),24 and possibly after taking antiandrogens, such as flutamide and cyproterone acetate.25 A particularly strong association has been observed between HCV infection (present in 33% of patients with HIV/AIDS) and the risk of liver injury during HAART; the risk may be increased 2- to 10-fold.26-30

Other Diseases

Rheumatoid arthritis appears to increase the risk of salicylate hepatotoxicity, and a curious, unexplained observation is that hepatitis associated with sulfasalazine appears to be more common in patients with rheumatoid arthritis than in those with inflammatory bowel disease.8-10,31,32 Diabetes mellitus, obesity, and chronic kidney disease predispose to methotrexate-induced hepatic fibrosis, whereas HIV/AIDS confers a heightened risk of sulfonamide hypersensitivity.31-33 A retrospective cohort study of five health maintenance organizations found that the age- and sex-standardized incidence of drug-induced acute liver failure in patients with diabetes mellitus was 0.08 to 0.15 per 1000 personyears, irrespective of the therapeutic agent used (the number using troglitazone was small); the incidence was highest (approximately 0.3 per 1000) during the first six months of exposure.32 Renal transplantation is a risk factor for azathioprine-associated vascular injury, whereas kidney disease predisposes to tetracycline-induced fatty liver.6 Finally, sinusoidal obstruction syndrome induced by anticancer drugs is more common after bone marrow transplantation24 and in persons with HCV infection.5,6,8-10,26

PATHOPHYSIOLOGY ROLE OF THE LIVER IN DRUG ELIMINATION

By virtue of the portal circulation, the liver is highly exposed to drugs and other toxins absorbed from the gastrointestinal tract. Most drugs tend to be lipophilic compounds that are readily taken up by the liver but that cannot be easily excreted unchanged in bile or urine. The liver is well equipped to handle these agents by an adaptable (inducible) series of metabolic pathways. These pathways include those that alter the parent molecule (phase 1); synthesize conjugates of the drug or its metabolite with a more water-soluble moiety, such as a sugar, amino acid, or sulfate molecule (phase 2); and excrete in an energy-dependent manner the parent molecule, its metabolites, or conjugates into bile

(phase 3). For any given compound, one, two, or all three of these steps may participate in drug elimination. Expression and subcellular location of the proteins (enzymes, membrane transporters) that mediate these steps are controlled by a set of nuclear receptors that function as transcriptional regulators and coregulators, thereby accounting for coordinated regulation among the three phases of hepatic drug elimination.

PATHWAYS OF DRUG METABOLISM

As reviewed elsewhere,5,34 phase 1 pathways of drug metabolism include oxidation, reduction, and hydrolytic reactions. The products can be readily conjugated or excreted without further modification.

Cytochrome P450

Most type 1 reactions are catalyzed by microsomal drug oxidases, the key component of which is a hemoprotein of the CYP gene superfamily. The apparent promiscuity of drug oxidases toward drugs, environmental toxins, steroid hormones, lipids, and bile acids results from the existence of multiple closely related CYP proteins. More than 20 CYP enzymes are present in the human liver.34 The reaction cycle involves binding of molecular oxygen to the iron in the heme prosthetic group, with subsequent reduction of oxygen by acceptance of an electron from nicotinamide-adenine dinucleotide phosphate (NADPH) cytochrome P450 reductase, a flavoprotein reductase. The resulting “activated oxygen” is incorporated into the drug or another lipophilic compound. Reduction of oxygen and insertion into a drug substrate (“mixed function oxidation”) can result in formation of chemically reactive intermediates, including free radicals, electrophilic “oxy-intermediates” (e.g., unstable epoxides, quinone imines), and reduced (and therefore reactive) oxygen species (ROS). The quintessential example is the CYP2E1-catalyzed metabolite of acetaminophen, N-acetyl-p-benzoquinone imine (NAPQI), an oxidizing and arylating metabolite that is responsible for liver injury associated with acetaminophen hepatotoxicity. Other quinone compounds are potential reactive metabolites of troglitazone, quinine, and methyldopa. Epoxide metabolites of diterpenoids may be hepatotoxic products of the hepatic metabolism of some plant toxins (see Chapter 87).35 ROS have broad significance in the production of tissue injury, particularly by contributing to the production of oxidative stress and triggering tissue stress responses and cell death pathways, as discussed later. The hepatic content of CYP proteins is higher in acinar zone 3 (see Chapter 71). Localization of CYP2E1 is usually confined to a narrow rim of hepatocytes 1 to 2 cells thick around the terminal hepatic venule. This finding explains in part the zonality of hepatic lesions produced by drugs and toxins, such as acetaminophen and carbon tetrachloride, which are converted to reactive metabolites.

Genetic and Environmental Determinants of Cytochrome P450 Enzymes

Pharmacogenetics and Polymorphisms of Cytochrome P450 Expression The hepatic expression of each CYP enzyme is genetically determined. This finding largely explains the four-fold or greater differences in rates of drug metabolism among healthy subjects. Some CYPs, particularly minor forms, are also subject to polymorphic inheritance; therefore, occasional persons completely lack the encoded protein.34 One example is CYP2D6, the enzyme responsible for the metabolism of debrisoquine and perhexiline. Poor metabolizers

1417

1418

Section IX  Liver lack CYP2D6 and accumulate perhexiline when given usual doses; lack of CYP2D6 is the critical determinant in serious adverse effects of perhexiline, including chronic hepatitis and cirrhosis.36 Other examples include CYPs 2C9 and 2C19, which affect the metabolism of S-warfarin, omeprazole, tolbutamide, and phenytoin and of S-mephenytoin, respectively34; 3% of white populations and 15% of Asians are poor metabolizers of S-mephenytoin. Developmental Regulation and Constitutive Expression Expression of several CYPs is developmentally regulated. During adult life, the expression of some CYPs declines slightly (by up to 10%) with advancing age, but this change is minor compared with the effects of genetic variation, environmental influences, and liver disease. Differences in the expression of CYPs 3A4 and 2E1 between men and women may explain the slighly enhanced metabolism of certain drugs (erythromycin, chlordiazepoxide, midazolam) in women, but whether this difference contributes to the increased risk of hepatic drug reactions in women remains unclear. Nutrition and Disease-Related Changes A person’s nutritional status influences the expression of certain CYPs, both in health and with liver disease.5,10,20,34 Expression of CYP2E1 is increased by obesity, high fat intake, and fasting.20,34 Diseases that alter the expression of hepatic CYPs include diabetes mellitus (increased CYP2E1), hypothyroidism (decreased CYP1A), and hypopituitarism (decreased CYP3A4).34 Cirrhosis is associated with decreased levels of total cytochrome P450 and also with reduced hepatic perfusion; the result is a decrease in the clearance of drugs such as propranolol that are metabolized rapidly by the liver.34 The effects of cirrhosis vary, however, among individual CYP families (e.g., CYP1A levels are lowered, but CYP2C and CYP2D6 levels often are preserved) and with the type of liver disease (e.g., CYP3A4 levels are preserved with cholestatic liver disease but lowered with hepatocellular liver disease). Adaptive Response and Enzyme Induction Exposure to lipophilic substances results in an adaptive response that usually involves synthesis of new enzyme protein, a process termed enzyme induction. The molecular basis for genetic regulation of constitutive and inducible expression of the major human hepatic cytochrome P450, CYP3A4, has been determined.37 Agents such as rifampin interact with the pregnane X-receptor (PXR), a member of the orphan nuclear receptor family of transcriptional regulators.37 Activated PXR and the analogous constitutive androstane receptor (CAR) in turn bind to cognate nucleotide sequences upstream to the CYP3A4 structural gene within a “xenobiotic-regulatory enhancer module” (XREM). This interaction regulates the CYP3A4 promoter downstream and ultimately the transcription of CYP3A4 protein. Similar control mechanisms apply to several other CYP pathways,37,38 particularly those involved with bile acid synthesis in which the nuclear receptors implicated include the farnesoid X-receptor (FXR), which down-regulates bile acid synthesis and up-regulates bile salt excretory pathways, and liver X receptor, a positive regulator of bile acid synthesis via CYP7A (see also Chapter 64).37 Common examples of the induction of microsomal enzymes by environmental compounds include the effect of smoking cigarettes and cannabis on CYP1A238 and of alcohol on CYP2E1 and possibly CYP3A4.39 Several drugs are potent inducers of CYP enzymes. Isoniazid induces CYP2E1, whereas phenobarbital and phenytoin increase the expression of multiple CYPs.34 Rifampin is a potent inducer of

CYP3A4, as is hypericum,40 the active ingredient of St John’s wort, a commonly used herbal medicine, thereby causing interactions between conventional medicines and a CAM preparation. Further descriptions of the regulation of hepatic drug metabolizing enzymes have been published elsewhere.34,38 The implications for drug-induced liver disease are twofold. First, enzyme induction often involves more than the CYP system, possibly because of activation of PXR and CAR; this observation could account for increases in serum levels of alkaline phosphatase and gamma glutamyl transpeptidase (GGTP), which reflect “hepatic adaptation” to chronic drug ingestion. Second, the influence of one drug on expression and activity of drug metabolizing enzymes and drug elimination (phase 3) pathways can alter the metabolism or disposition of other agents. Such drug-drug interactions are important pharmacologically and may be relevant to mechanisms of drug-induced liver injury. Inhibition of Drug Metabolism Some chemicals inhibit drug metabolism. In persons taking more than one medication, for example, competition for phase 2 pathways such as glucuronidation and sulfation facilitates the presentation of unconjugated drug to the CYP system. This mechanism appears to explain in part why agents such as zidovudine and phenytoin lower the dose threshold for acetaminophen-induced hepatotoxicity.

Other Pathways of Drug Oxidation

In addition to CYP enzymes, electron transport systems of mitochondria may lead to the generation of tissue-damaging reactive intermediates during the metabolism of some drugs. Examples of such reactive intermediates include nitroradicals from nitrofuran derivatives (nitrofurantoin, cocaine). Subsequent electron transfer by flavoprotein reductases into molecular oxygen generates superoxide and other ROS. Some anticancer drugs such as doxorubicin and the imidazole antimicrobial agents can participate in other oxidation-reduction (redox) cycling reactions that generate ROS.

Phase 2 (Conjugation) Reactions

Phase 2 reactions involve formation of ester links to the parent compound or a drug metabolite. The responsible enzymes include glucuronosyl (or glucuronyl) transferases, sulfotrans­ferases, glutathione-S transferases, and acetyl and amino acid N-transferases. The resulting conjugates are highly water soluble and can be excreted readily in bile or urine. Conjugation reactions can be retarded by depletion of their rate-limiting cofactors, such as glucuronic acid and inorganic sulfate, and the relatively low capacity of these enzyme systems restricts the efficacy of drug elimination when substrate concentrations exceed enzyme saturation. In general, drug conjugates are nontoxic, and phase 2 reactions are considered to be detoxification reactions, with exceptions. For example, some glutathione conjugates can undergo cysteine S-conjugate β-lyase-mediated activation to highly reactive intermediates. Little is known about the regulation of such enzymes or their potential significance for drug-induced liver disease or hepatocarcinogenesis.

Phase 3 Pathways

Several transporters secrete drugs, drug metabolites, or their conjugates into bile, and this mechanism is often referred to as phase 3 of hepatic drug elimination. These pathways involve ATP-binding cassette (ABC) proteins, which derive the energy for their transport functions from hydrolysis of ATP. ABC transport proteins are widely distributed in nature and include the cystic fibrosis transmembrane con-

Chapter 86  Liver Disease Caused by Drugs ductance regulator (CFTR) (see Chapter 76) and the canalicular and intestinal copper transporters (see Chapter 75). The role of ABC transport proteins in secretion of bile has been reviewed (see Chapter 64).37,38,41,42 Multidrug resistance protein 1 (MDR1) is highly expressed on the apical (canalicular) plasma membrane of hepatocytes, where it transports cationic drugs, particularly anticancer agents, into bile. Another family of ABC transporters, the multidrug resistance-associated proteins (MRPs), is also expressed in the liver. At least two members of this family excrete drug (and other) conjugates from hepatocytes: MRP-3 on the basolateral surface facilitates passage of drug conjugate into the sinusoidal circulation, and MRP-2, expressed on the canalicular membrane, pumps endogenous compounds (e.g., bilirubin diglucuronide, leukotrieneglutathionyl conjugates, glutathione) and drug conjugates into bile. The bile salt export pump (BSEP) and MDR3 (in humans) and Mdr2 (in mice) are other canalicular transporters concerned, respectively, with bile acid and phospholipid secretion into bile. Genetic polymorphisms of these genes are associated with human cholestatic liver diseases (see Chapters 64 and 76). BSEP interacts with several drugs.42 Regulation of the membrane expression and activity of these drug elimination pathways is complex. The possibility that their altered expression or impaired activity (by competition between agents, changes in membrane lipid composition, or damage from reactive metabolites or covalent binding) could lead to drug accumulation, impairment of bile flow, or cholestatic liver injury has been demonstrated for estrogens,43,44 troglitazone,45 terbinafine,46 and flucloxacillin47 and may have wider mechanistic importance for drug-induced cholestasis and other forms of liver injury.42

TOXIC MECHANISMS OF LIVER INJURY Direct Hepatotoxins and Reactive Metabolites

Highly hepatotoxic chemicals injure key subcellular structures, particularly mitochondria and the plasma membrane. The injury arrests energy generation, dissipates ionic gradients, and disrupts the physical integrity of the cell. This type of overwhelming cellular injury does not apply to currently relevant hepatotoxins, most of which require metabolic activation to mediate damage to liver cells. The resulting reactive metabolites can interact with critical cellular target molecules, particularly those with nucleophilic substituents such as thiol-rich proteins and nucleic acids. Together with ROS, they act as oxidizing species within the hepatocyte to establish oxidative stress, a state of imbalance between pro-oxidants and antioxidants. ROS are also key signaling molecules that mediate biological responses to stress, as discussed later. Alternatively, reactive metabolites bind irreversibly to macromolecules, particularly proteins and lipids. Such covalent binding may produce injury by inactivating key enzymes or by forming protein-drug adducts that could be targets for immunodestructive processes that cause liver injury.

Oxidative Stress and the Glutathione System

The liver is exposed to oxidative stress by the propensity of hepatocytes to reduce oxygen, particularly in mitochondria and also in microsomal electron transport systems (such as CYP2E1), and by NADPH-oxidase-catalyzed formation of ROS and nitroradicals in Kupffer cells, endothelial cells, and stimulated polymorphs and macrophages. To combat oxidative stress, the liver is well-endowed with antioxidant mechanisms, including micronutrients, such as vitamin E and vitamin C, thiol-rich proteins (e.g., metallothionein,

ubiquinone), metal-sequestering proteins (e.g., ferritin), and enzymes that metabolize reactive metabolites (e.g., epoxide hydrolases), ROS (e.g., catalase, superoxide dismutase), and lipid peroxides (e.g., glutathione peroxidases). Glutathione (l-gamma-glutamyl-l-cyteine-glycine) is the most important antioxidant in the mammalian liver.19 Hepatocytes are the exclusive site of glutathione synthesis. Hepatic levels of glutathione are high (5 to 10 mmol/L) and can be increased by enhancing the supply of cysteine for glutathione synthesis; this mechanism is the cornerstone of thiol antidote therapy for acetaminophen poisoning. Hepatocyte glutathione synthesis increases in response to pro-oxidants, as occurs when CYP2E1 is overexpressed as a result of signaling via the redox-sensitive transcription factor Nrf.19,20,48,49 Glutathione is a critical cofactor for several antioxidant pathways, including thiol-disulfide exchange reactions and glutathione peroxidase. Glutathione peroxidase has a higher affinity for hydrogen peroxide than does catalase, and it disposes of lipid peroxides, free radicals, and electrophilic drug metabolites. Reduced glutathione is a cofactor for conjugation reactions catalyzed by the glutathione S-transferases. Other reactions proceed nonenzymatically. In turn, the products include glutathione-protein mixed disulfides and oxidized glutathione. The latter can be converted back to glutathione by proton donation catalyzed by glutathione reductase. Normally, most glutathione within the hepatocyte is in the reduced state, indicating the importance of this pathway for maintenance of the redox capacity of the cell. The reduced form of NADPH is an essential cofactor for glutathione reductase; NADPH formation requires ATP, thereby illustrating a critical link between the energy-generating capacity of the liver and its ability to withstand oxidative stress. Glutathione is also compartmentalized within the hepatocyte, with the highest concentrations found in the cytosol. Adequate levels of glutathione are essential in mitochondria, where ROS are constantly being formed as a minor by-product of oxidative respiration and in response to some drugs or metabolites that interfere with the mitochondrial respiratory chain. Mitochondrial glutathione is maintained by active uptake from the cytosol, a transport system that is altered by chronic ethanol exposure, and is, therefore, another potential target of drug toxicity.19

Biochemical Mechanisms of Cellular Injury

Mechanisms once thought to be central to hepatotoxicity, such as covalent binding to cellular enzymes and peroxidation of membrane lipids, are no longer regarded as exclusive pathways of cellular damage. Rather, oxidation of proteins, phospholipid fatty acyl side chains (lipid peroxidation), and nucleosides appear to be components of the biochemical stress that characterizes toxic liver injury. Secondary reactions also may play a role; these reactions include post-translational modification of proteins via adenosine diphosphate (ADP) ribosylation or protease activation, cleavage of DNA by activation of endogenous endonucleases, and disruption of lipid membranes by activated phospholipases.20 Some of these catabolic reactions could be initiated by a rise in the cytosolic ionic calcium concentration [Ca2+]i, as a result of increased Ca2+ entry or release from internal stores in the endoplasmic reticulum and mitochondria.19,20 The concept that hepatotoxic chemicals cause hepatocyte cell death by a biochemical final common pathway (e.g., activation of catalytic enzymes by a rise in [Ca2+]i) has proved inadequate to explain the diverse processes that can result in lethal hepatocellular injury. Rather, a variety of processes can damage key organelles, thereby causing intracellular stress that activates signaling pathways and tran-

1419

1420

Section IX  Liver scription factors. In turn, the balance between these factors can trigger the onset of cell death or facilitate protection of the cell, as discussed in the following sections.

Types of Cell Death

Apoptosis Apoptosis is an energy-dependent, genetically programmed form of cell death that typically results in controlled deletion of individual cells. In addition to its major roles in developmental biology, tissue regulation, and carcinogenesis, apoptosis is important in toxic, viral, and immunemediated liver injury.50-53 The ultrastructural features of apoptosis are cell and nuclear shrinkage, condensation and margination of nuclear chromatin, plasma membrane blebbing, and ultimately fragmentation of the cell into membrane-bound bodies that contain intact mitochondria and other organelles. Engulfment of these apoptotic bodies by surrounding epithelial and mesenchymal cells conserves cell fragments that contain nucleic acid and intact mitochondria. These fragments are then digested by lysosomes and recycled without release of bioactive substances. As a consequence, apoptosis in it purest form (usually found only in vitro) does not incite an inflammatory tissue reaction. The cellular processes that occur in apoptosis are often mediated by caspases, a family of proteolytic enzymes that contain a cysteine at their active site and cleave poly­ peptides at aspartate residues; non–caspase-mediated programmed cell death has also been described in experimental hepatotoxicity (see also Chapter 72). Apoptosis rarely, if ever, is the sole form of cell death in common forms of liver injury, such as ischemia-reperfusion injury, cholestasis, and toxic liver injury, all of which are typically associated with a hepatic inflammatory response. Whether or not activation of pro-death signals causes cell death depends on several factors, including pro-survival signals, the rapidity of the process, the availability of glutathione and ATP, and the role of other cell types. Some of these issues are discussed briefly here and are reviewed in more detail elsewhere.20,50-53 Hepatocytes undergo apoptosis when pro-apoptotic intracellular signaling pathways are activated, either because of toxic biochemical processes within the cell (intrinsic pathway) or because cell surface receptors are activated to transduce cell death signals (external pathway). Proapoptotic receptors are members of the tumor necrosis factor-α (TNF-α) receptor superfamily, which possess a so-called death domain. These receptors include Fas, for which the cognate ligand is Fas-ligand (Fas-L), TNF-R1 receptor (cognate ligand is TNF), and TNF-related apoptosis-inducing ligand (TRAIL) receptors (cognate ligand is TRAIL). In addition to model hepatotoxins such as the quinone, menadione, and hydrogen peroxide, some drugs (e.g., acetaminophen, plant diterpenoids) have been shown to be converted into pro-oxidant reactive metabolites, thereby initiating the following sequence: CYP-mediated metabolism to form reactive metabolites → glutathione depletion → mitochondrial injury with release of cytochrome c and operation of the mitochondrial membrane permeability transition (MPT) → caspase activation → apoptosis. Mitochondria play a pivotal role in pathways that provoke or oppose apoptosis.50,51,53 In the external pathway, activation of the death domain of pro-apoptotic receptors recruits adapter molecules—Fas-associated death domain (FADD) and TNF receptor-associated death domain (TRADD)— which bind and activate procaspase 8 to form the deathinducing signaling complex (DISC). In turn, caspase 8

cleaves Bid, a pro-apoptotic member of the B cell lymphoma/leukemia (Bcl-2) family, to tBid. tBid causes translocation of Bax to the mitochondria, where it aggregates with Bak to promote permeability of the mito­ chondria.50 Release of cytochrome c and other pro-death molecules, including Smac (which binds caspase inhibitor proteins, such as inhibitor of apoptosis proteins [IAPs]) and apoptosis-inducing factor (AIF, also known as Apaf)51 allows formation of the “aptosome,” which activates caspase 9 and eventually caspase 3 to execute cell death (Fig. 86-1). Intracellular stresses in various sites release other mitochondrial permeabilizing proteins (e.g., Bmf from the cytoskeleton and Bim from the endoplasmic reticulum), whereas members of the Bcl-2 family, Bcl-2 and Bcl-XL, antagonize apoptosis and serve as survival factors by regulating the integrity of mitochondria; the protective mechanism is poorly understood. Stress-activated protein kinases, particularly c-jun N-terminal kinase (JNK) may also be proapoptotic52 by phosphorylating and inactivating the mitochondrial protective protein Bcl-XL. Execution of cell death by apoptosis usually occurs via activation of caspase 3, but more than one caspaseindependent pathway of programmed cell death has been described.53 Stresses to the endoplasmic reticulum can bypass mitochondrial events by activation of caspase 12, which in turn activates caspase 9 independently of the apoptosome. The final steps of programmed cell death are energy dependent. Therefore, depletion of ATP abrogates the controlled attempt at “cell suicide,” resulting instead in necrosis (see later) or an overlapping pattern that has been designated as “apoptotic necrosis” or “necraptosis.”54,55 Furthermore, when apoptosis is massive, the capacity for rapid phagocytosis can be exceeded, and “secondary” necrosis can occur.55 Intracellular processes and activation of pro-apoptotic death receptors are not mutually exclusive pathways of cell death in toxic liver injury. In fact, drug toxicity could predispose the injured hepatocyte to apoptosis mediated by TNF-R or Fas-operated pathways by several mechanisms, including blockade of nuclear factor-kappa B (NF-κB), which usually is a hepatoprotective transcription factor in hepatocytes, and inhibition of purine and protein synthesis. Furthermore, activation of Kupffer cells (e.g., by endotoxin) and recruitment of activated inflammatory cells can increase production of TNF-α. Apoptosis-initiating pathways have broad relevance in several areas of hepatology. Inhibition of caspases is an important protective mechanism against cell death. Such anti-apoptotic pathways include chemical blockade of the cysteine thiol group by nitric oxide (NO) or ROS and cellular depletion of glutathione.20 Protein inhibitors include IAP family members, heat shock proteins (HSPs), and FLICE (caspase-8)inhibitory proteins (FLIP).50-52 FLIP inhibit caspase-8 activation as a decoy for FADD binding. Bcl-2 and Bcl-XL inhibit mitochondrial permeability, whereas phosphatidylinositol 3-kinase/Akt phosphorylates caspase 9 and activates NF-κB. Necrosis In contrast to apoptosis, necrosis has been conceptualized as a relatively uncontrolled process that can result from extensive damage to the plasma membrane with disturbance of ion transport, dissolution of membrane potential, cell swelling, and eventually rupture of the cell. Drug-induced injury to the mitochondrion can impair energy generation, whereas MPT can release stored Ca2+ into the cytosol and perturb other ionic gradients. Mitochondrial enzymes appear to be a particular target of NAPQI, the reactive

Chapter 86  Liver Disease Caused by Drugs Death ligand e.g., TNF, Fas, TRAIL

Death receptor e.g., TNF-R1

Plasma membrane Cytoplasm FLIP

FADD

Caspase 8 (FLICE)

TRADD

RIP DISC IAP

IAP Bid TRAF2 Bax

Bcl-xL Lipid mediators (ceramide)

Effector caspases

Apoptosis

Apoptosome

Mitochondrion MPT

Necrosis Figure 86-1.  Apoptosis and necrosis pathways in mammalian cells. See text for details. Bcl, B-cell lymphoma/leukemia family (Bax, Bid, and Bcl-xL are members); DISC, death-inducing signaling complex; FADD, Fas-associated death domain; FLIP, FLICE-inhibitory proteins; IAP, inhibitor of apoptosis proteins; MPT, mitochondrial permeability transition; RIP, receptor-interacting protein; TNF, tumor necrosis factor; TNF-R1, TNF receptor-1; TRADD, TNF receptor-associated death domain; TRAF2, TNF receptor-associated factor-2; TRAIL, TNF-related apoptosis ligand.

metabolite of acetaminophen. Reye’s syndrome–like disorders (e.g., toxicity caused by valproic acid; some nucleoside analogs, such as fialuridine, didanosine, zidovudine, zalcitabine; and possibly “ecstasy” [see Chapter 87]) may also result from mitochondrial injury. Mitochondrial injury can result in cell death by either apoptosis or necrosis54,55; the type of cell death pathway may depend primarily on the energy state of the cell, as well as the rapidity and severity of the injury process. In the presence of ATP, cell death can proceed by apoptosis, but when mitochondria are deenergized, the mechanism of cell death is necrosis. This apparent dichotomy between cell death processes is probably artificial, and apoptosis and necrosis more likely represent the morphologic and mechanistic ends of a spectrum of overlapping cell death processes.19,55 One important way in which necrosis differs from apoptosis is that uncontrolled dissolution of the cell liberates macromolecular breakdown products, including lipid peroxides, aldehydes, and eicosanoids. These products act as chemoattractants for circulating leukocytes, which enjoin an inflammatory response in the hepatic parenchyma. Even before cell death occurs, oxidative stress produced during drug toxicity can up-regulate adhesion molecules and chemokines that are expressed or secreted by endothelial cells.

These processes contribute to recruitment of the hepatic inflammatory response, which is prominent in some types of drug-induced liver disease. Lymphocytes, polymorphonuclear leukocytes (neutrophils and eosinophils), and macrophages also may be attracted to the liver as part of a cell-mediated immune reaction. Role of Oxidative Stress Although severe oxidative stress in hepatocytes, particularly when focused on mitochondria, is likely to induce necrosis, lesser (or more gradual) exposure can trigger apoptosis because ROS and oxidative stress can activate Fas signaling, JNK and other kinases, p53, and microtubular assembly and impair protein folding, thereby resulting in an unfolded protein response by the endoplasmic reticulum. Oxidative stress also may amplify cell death processes by uncoupling of the mitochondrial respiratory chain, release of cytochrome c, or massive oxidation and export of glutathione (intact glutathione is required for Fas signaling). Conversely, oxidative stress may protect against apoptosis in some circumstances through inhibition of caspase or activation of NF-κB. As a result of these opposing effects, predicting the consequences of hepatic oxidative stress in terms of liver injury is not easy.

1421

1422

Section IX  Liver Role of Hepatic Nonparenchymal Cells and the Innate Immune Response

In addition to migratory cells, activation of nonparenchymal liver cell types is likely to play an important role in drug and toxin-induced liver injury. Kupffer cells function as resident macrophages and antigen-presenting cells. Some of the toxic effects of activated Kupffer cells, as well as of recruited leukocytes, may be mediated by release of cytokines, such as TNF and Fas-L, which under some circumstances can induce cell death in hepatocytes by apoptosis or necrosis.55 In addition, activated Kupffer cells release ROS, nitroradicals, leukotrienes, and proteases. Endothelial cells of the hepatic sinusoids or terminal hepatic veins are vulnerable to injury by some hepatotoxins because of their low glutathione content. Such hepatotoxins include the pyrrolizidine alkaloids, which are an important cause of the sinusoidal obstruction syndrome (hepatic venoocclusive disease).56 Other types of drug-induced vascular injury may be caused primarily by involvement of the sinusoidal endothelial cells (see Chapter 83). Hepatic stellate cells (formerly fat-storing or Ito cells) are the principal liver cell type involved in matrix deposition in hepatic fibrosis. Stellate cells are activated in methotrexate-induced hepatic fibrosis. The possibility that vitamin A, ROS, or drug metabolites can transform stellate cells into collagen-synthesizing myofibroblasts is of considerable interest.

IMMUNOLOGIC MECHANISMS

In addition to the activation of innate inflammatory processes in the liver by toxic mechanisms, (extrinsic) immunologic mechanisms could account for certain aspects of idiosyncratic drug-induced liver disease. Immune attack involves liganding of death receptors, as discussed earlier, or porin-mediated introduction of granzyme.19 The most convincing evidence for drug allergy includes (1) delayed onset after initial exposure and accelerated onset after rechallenge, (2) hepatic inflammatory infiltrates with neutrophils and eosinophils, and (3) fever, rash, lymphadenopathy, peripheral eosinophilia, and involvement of other organs. In some types of drug hepatitis, the liver is clearly implicated as part of a systemic hypersensitivity reaction, as described later for the reactive metabolite syndrome (RMS); sulfonamides, phenytoin, nitrofurantoin, minocycline, nevirapine, and some Chinese herbal medicines are causative agents. Why the liver is the predominant site of injury in some persons whereas other organs are involved in other persons is unclear; genetic factors relevant to tissuespecific gene expression could be involved. One possible immunopathogenic mechanism for druginduced liver disease is the altered antigen concept, in which an initial interaction between drug metabolites and cellular proteins results in the formation of neoantigens (haptens) or drug-protein adducts. An example is the formation of trifluoroacetylated (TFA) adducts after exposure to halothane or other haloalkane anesthetics (see Chapter 87). For these adducts to initiate tissue-damaging immune responses (1) processing should be presented in an immunogenic form (e.g., by Kupffer cells, in association with major histocompatibility complex [MHC] molecules); (2) appropriately responsive CD4+ T cells must be present to provide help to induce an immune response; and (3) the drug-derived antigen, together with a class II MHC molecule, must be expressed on the target cells in order to attract CD8+ (cytotoxic) T cells. That bile duct epithelial cells are more likely to express class II MHC antigens may explain why they are possible targets in drug-induced cholestatic hepatitis.

Although antibodies directed against TFA-protein adducts circulate in the majority of patients following recovery from halothane-induced liver injury,57 the specificity and pathogenicity of these antibodies remain in doubt. Another way in which circulating drug-induced antibodies could result in immune-mediated lysis of hepatocytes is through molecular mimicry of host enzymes.58 Experimental evidence suggests that for diclofenac antibody-dependent cell-mediated immunity could operate as a mechanism for drug-induced liver disease.59 A second type of immunopathogenic mechanism is dysregulation of the immune system, termed drug-induced autoimmunity. This mechanism can lead to the formation of drug-induced autoantibodies (e.g., anti–liver-kidney microsome [LKM] antibodies) directed against microsomal enzymes. For tienilic acid, CYP2C9 is the target of antiLKM, whereas after halothane hepatitis anti-LKM are directed against CYP2E1. Non-tissue specific autoantibodies, such as antinuclear and smooth muscle antibodies, may be detected in patients with nitrofurantoin, methyldopa, or minocycline hepatitis. Like spontaneous autoimmunity, drug-induced autoimmunity may involve genetic predisposition through anomalies of immune tolerance.

CLINICOPATHOLOGIC FEATURES OF DRUG-INDUCED LIVER DISEASE CLASSIFICATION

Hepatic drug reactions mimic all known liver diseases, but classification is often difficult because of overlap among categories. Although a classic (“signature”) syndrome is associated with many individual agents, a given drug may be associated with more than one clinicopathologic syndrome. Furthermore, the clinical and laboratory features of liver disease and the hepatic histologic findings may be discordant. Therefore, although recognition of specific patterns or syndromes is a vital clinical clue to the diagnosis of drug-induced liver disease, the chronologic relationship between administration of the drug and liver injury is a more important clue. Drugs are often divided into dose-dependent, or predictable, hepatotoxins and dose-independent, or unpredictable (idiosyncratic), hepatotoxins. Dose-dependent hepatotoxins generally require metabolic activation to toxic metabolites or interfere with subcellular organelles and biochemical processes at key sites, such as mitochondria or canalicular bile secretion.43 Liver injury produced by dose-dependent hepatotoxins usually occurs after a short latent period (hours), is characterized by zonal necrosis or microvesicular steatosis, and can be reproduced in other species. By contrast, idiosyncratic hepatotoxins cause a wide range of histologic changes and do not reliably cause injury in other species; in addition, the latent period before the onset of injury is variable in duration. The distinction between dosedependent and idiosyncratic hepatotoxins is blurred with agents such as dantrolene, tacrine, perhexiline, flucloxacillin, cyclophosphamide, nucleoside analogs, bromfenac, anticancer drugs, and cyclosporine. Liver injury caused by each of these drugs is partly dose dependent, but reactions occur in only a small proportion of exposed persons. Two general types of mechanisms may account for idiosyncratic hepatotoxicity: metabolic idiosyncrasy and immunoallergy. Metabolic idiosyncrasy refers to the susceptibility of rare persons to hepatotoxicity from a drug that, in conventional doses, is usually safe. Such susceptibility may result from genetic or acquired differences in drug metabo-

Chapter 86  Liver Disease Caused by Drugs lism or canalicular secretion, mitochondrial defects, or cell death receptor signaling. Immunoallergy indicates operation of the immune system in mediating the response to a drug. These two mechanisms may be interrelated (see later). Other pathogenic mechanisms may include indirect mediation of liver injury, as in vascular and possibly hyperthermic changes produced by cocaine, ecstasy, intraarterial fluroxuridine, and possibly anesthetics (see Chapter 87). The most practical classification of drug hepatotoxicity is based on clinical and laboratory features and liver histology, as summarized in Table 86-2. This classification provides a framework for discussing drug-induced hepatic disease in comparison with other hepatobiliary disorders but is imperfect because the clinical and pathologic features are not always congruent. Moreover, much overlap between categories exists, particularly in the spectrum from severe necrosis (which may result from dose-dependent or idiosyncratic hepatotoxicity) to focal necrosis with lobular inflammation (hepatitis) to cholestasis. Many drugs produce a spectrum of syndromes from hepatitis to cholestasis, and some authorities include a further category of mixed cholestatichepatocellular reactions. Granulomatous hepatitis is asso­ ciated with liver biochemical test abnormalities that are usually indistinguishable from those typical of hepatitis, cholestasis, or mixed reactions.

Drugs can alter liver test results without causing significant liver injury. Such adaptive responses include hyperbilirubinemia associated with rifampin, cyclosporine, and indinavir and raised serum GGTP and alkaline phosphatase levels associated with phenytoin and warfarin.5,6 The latter effect is probably attributable to microsomal enzyme induction. For agents such as isoniazid, however, the distinction between adaptation and minor injury is blurred; adaptation in such cases may be a response to oxidative injury. Conversely, liver tumors or hepatic fibrosis may develop insidiously without significant abnormalities of liver biochemical tests—the former in association with sex steroids or vinyl chloride monomer and the latter with methotrexate, arsenic, or hypervitaminosis A. The duration of the disorder is another consideration in classifying drug-induced liver diseases. In general, chronic liver disease is much less commonly attributable to drugs and toxins than are acute reactions,8 but not to consider drugs as a possible eitology of chronic liver disease can lead to a missed diagnosis, with serious clinical consequences.8,9 In contrast to most other types of hepatic pathobiology, drugs and toxins constitute the most important cause of vascular disorders of the liver (see later). Drugs also have been associated with chronic cholestasis, chronic hepatitis, steatohepatitis, hepatic fibrosis, cirrhosis, and benign and malignant liver tumors.

Table 86-2  Clinicopathologic Classification of Drug-Induced Liver Disease CATEGORY

DESCRIPTION

IMPLICATED DRUGS: EXAMPLES

Hepatic adaptation

No symptoms; raised serum GGTP and AP levels (occasionally raised ALT) Hyperbilirubinemia Symptoms of hepatitis; zonal, bridging, and massive necrosis; serum ALT level >5-fold increased, often >2000 U/L Microvesicular steatosis, diffuse or zonal; partially dose dependent, severe liver injury, features of mitochondrial toxicity (lactic acidosis) Symptoms of hepatitis; focal, bridging, and massive necrosis; serum ALT level >5-fold increased; extrahepatic features of drug allergy in some cases Duration >3 months; interface hepatitis, bridging necrosis, fibrosis, cirrhosis; clinical and laboratory features of chronic liver disease; autoantibodies with some types of reaction (see Table 86-6) Hepatic granulomas with varying hepatitis and cholestasis; raised serum ALT, AP, GGTP levels Cholestasis, no inflammation; serum AP levels >twice-normal Cholestasis with inflammation; symptoms of hepatitis; raised serum ALT and AP levels Bile duct lesions and cholestatic hepatitis; clinical features of cholangitis Cholestasis present >3 months Paucity of small bile ducts; resembles primary biliary cirrhosis, but AMA negative Strictures of large bile ducts Steatosis, focal necrosis, Mallory’s hyaline, pericellular fibrosis, cirrhosis; chronic liver disease, portal hypertension Sinusoidal obstruction syndrome, nodular regenerative hyperplasia, others Hepatocellular carcinoma, adenoma, angiosarcoma, others

Phenytoin, warfarin

Dose-dependent hepatotoxicity Other cytopathic toxicity, acute steatosis Acute hepatitis

Chronic hepatitis

Granulomatous hepatitis Cholestasis without hepatitis Cholestatic hepatitis Cholestasis with bile duct injury Chronic cholestasis:   Vanishing bile duct syndrome   Sclerosing cholangitis Steatohepatitis Vascular disorders Tumors

Rifampin, flavaspidic acid Acetaminophen, nicotinic acid, amodiaquine, hycanthone Valproic acid, didanosine, HAART agents, fialuridine, l-asparaginase, some herbal medicines Isoniazid, dantrolene, nitrofurantoin, halothane, sulfonamides, phenytoin, disulfiram, acebutolol, etretinate, ketoconazole, terbinafine, troglitazone Nitrofurantoin, etretinate, diclofenac, minocycline, nefazodone (see also Table 86-6) Allopurinol, carbamazepine, hydralazine, quinidine, quinine (see also Table 86-5) Oral contraceptives, androgens Chlorpromazine, tricyclic antidepressants, erythromycins, amoxicillin-clavulanic acid Chlorpromazine, flucloxacillin, dextropropoxyphene Chlorpromazine, flucloxacillin, trimethoprim-sulfamethoxazole Intra-arterial floxuridine, intralesional scolicidals Perhexiline, amiodarone, others (see Chapter 85) Many (see Table 86-8) Many (see Chapter 94)

ALT, alanine aminotransferase; AMA, antimitochondrial antibodies; AP, alkaline phosphatase; AST, aspartate aminotransferase; GGTP, gamma glutamyl transpeptidase; HAART, highly active antiretroviral therapy.

1423

1424

Section IX  Liver HISTOPATHOLOGIC FEATURES

Although no pathognomonic hallmarks of drug-induced liver disease have been identified, certain histologic patterns suggest drug-induced liver injury. These patterns include zonal necrosis or microvesicular steatosis (which accompanies mitochondrial injury) and mixed histologic features of hepatocellular necrosis and cholestasis. Necrotic lesions that are disproportionately severe compared with the clinical picture also indicate a possible drug cause, whereas destructive bile duct lesions, prominent neutrophils, and eosinophils (at least 25% of inflammatory cells) are suggestive of drug-induced cholestatic hepatitis. Hepatic granuloma formation is another common type of hepatic drug reaction. In cases of steatohepatitis, hepatic fibrosis, or liver tumors, no specific clues to a drug cause have been recognized, although sex steroids increase the vascularity of hepatic tumors and are frequently associated with sinusoidal dilatation or peliosis hepatis. Drug-induced steatohepatitis caused by amiodarone and perhexilene tends to be associated with severe lesions that more closely resemble alcoholic hepatitis than NASH.60 Other drugs (e.g., tamoxifen, methotrexate) cause lesions that are indistinguishable from NASH associated with diabetes mellitus and the metabolic syndrome.6,8,61

CLINICAL FEATURES

The history and physical examination can provide important clues to the diagnosis of hepatic drug reactions. Most important is the temporal pattern of disease evolution in relation to exposure to drugs or toxins. The identification of specific risk factors for hepatotoxicity (e.g., chronic excessive alcohol intake in a person taking acetaminophen) and the presence of systemic features of drug hypersensitivity may indicate the correct diagnosis. Systemic features include fever, rash, mucositis, eosinophilia, lymphadenopathy, a mononucleosis-like syndrome, bone marrow suppression, vasculitis, renal failure, pneumonitis, and pancreatitis. These features may be part of a characteristic syndrome thought to have a genetic basis and likely mediated by formation of drug metabolites that act as haptens to initiate an immunodestructive tissue reaction termed the reactive metabolite syndrome (RMS).62

Reactive Metabolite Syndrome

Drugs implicated as a cause of RMS include sulfonamides, aminopenicillins, fluoroquinolones, clozapine, anticonvulsants (phenytoin, lamotrigine, phenobarbital, carbama­ zepine), minocycline, protease inhibitors (nevirapine, abacavir), some NSAIDs, and Chinese herbal medicines.62 Risk factors for RMS include a family history of an affected first-degree relative (increases the risk to 1 in 4). Use of other drugs, such as glucocorticoids or valproic acid, at the time the new agent is started increases the risk 4- to 10-fold. The presence of a disorder associated with immune dysregulation (e.g., systemic lupus erythematosus) increases the risk 10-fold, whereas HIV/AIDS increases the risk 100-fold. The illness characteristically begins between 1 and 12 weeks (typically 2 to 4 weeks) after the drug is started; “sentinel symptoms” include fever, pharyngitis, malaise, periorbital edema, headache or otalgia, rhinorrhea, and mouth ulcers. A severe drug rash is an essential feature. Erythematous reactions are usual and may evolve to toxic epidermal necrolysis or erythema multiforme, often with mucositis (Stevens-Johnson syndrome). Early abnormalities on blood testing include neutrophilia and elevated levels of acute-phase reactants; atypical lymphocytosis and eosinophilia may be noted later. Hepatic reactions are found in

about 13% of cases. Findings include cholestasis, acute hepatitis, and granulomas. Other features include lymphadenopathy (16%), nephritis (6%), pneumonitis (6%), and more severe hematologic abnormalities (5%).

Latent Period to Onset

For idiosyncratic reactions, a latent period occurs between the commencement of drug intake and the onset of clinical and laboratory abnormalities. The duration of the latent period is commonly 2 to 8 weeks for immunoallergic types of drug hepatitis (such as the RMS) and often 6 to 20 weeks or longer for agents such as isoniazid, dantrolene, and troglitazone. Occasionally, liver injury may become evident after discontinuation of the causative agent; for oxypenicillins and amoxicillin-clavulanate, the onset of hepatotoxicitiy may occur as long as two weeks after the end of therapy. In other cases, hepatotoxicity is rare after the first exposure to a drug but more frequent and more severe after subsequent courses. Examples include halothane, nitrofurantoin, and dacarbazine. A history of a previous reaction to the drug in question (inadvertent rechallenge) may, therefore, be the key to the diagnosis of drug-induced liver disease.

Dechallenge and Rechallenge

Another aspect of the temporal relationship between ingestion of a drug and hepatotoxicity is the response to discontinuation of the drug, or dechallenge. Dechallenge should be accompanied by discernible and progressive improvement within days or weeks of stopping the incriminated agent. Exceptions occur with ketoconazole, troglitazone, coumarol, etretinate, and amiodarone; with these agents, reactions may be severe, and clinical recovery may be delayed for months. Although some types of drug-induced cholestasis also can be prolonged, failure of jaundice to resolve in a suspected drug reaction most often is indicative of an alternative diagnosis. Rarely, deliberate rechallenge may be used to confirm the diagnosis of drug-induced liver disease or to prove the involvement of one particular agent when the patient has been exposed to several drugs; however, this approach is potentially hazardous and should be undertaken only with a fully informed and consenting (in writing) patient and preferably with the approval of an institutional ethics committee.

A PRACTICAL APPROACH TO DIAGNOSIS In the absence of specific diagnostic tests, diagnosis of druginduced liver disease requires clinical suspicion, a thorough drug history, careful consideration of the temporal relationships between drug ingestion and liver disease, and exclusion of other disorders. The objective weighing of evidence for and against an individual agent—causality assessment— is a probabilistic form of diagnosis.63,64 Several clinical scales that incorporate and weigh various features of hepatic adverse drug reactions have been described.9,65-67 A liver biopsy may be indicated in some cases to exclude other diseases and to provide further clues to a drug etiology.68 In the future, in vitro tests may provide confirmatory evidence for particular drugs,57-65 but rechallenge is currently the standard test for drug-induced liver disease.66

PHYSICIAN AWARENESS

Physician awareness is critical for the diagnosis of druginduced liver disease. The sources of potential hepatotoxins include not only prescribed medications, but also over-the-

Chapter 86  Liver Disease Caused by Drugs counter drugs (e.g., ibuprofen), CAM preparations (see Chapters 87 and 127), substances taken for recreational use (e.g., cocaine, ecstasy) or self-poisoning, and environmental contaminants in food and water supplies, the home, the workplace, and the community. Unfortunately, patients and physicians do not always heed early nonspecific symptoms associated with reactions to hepatotoxic drugs. For example, preventable deaths from liver failure still occur more than 40 years after the recognition that isoniziad can cause drug hepatitis.69 Although continuing education and availability of information about potentially hepatotoxic drugs are important issues, physicians have a professional and legal obligation to inform patients about possible adverse drug reactions. A study from Switzerland found that the frequency of new cases of drug-induced liver injury among over 4000 hospital admissions was 1.4% (57 cases). Nevertheless, the drug reaction was not mentioned as a diagnosis in the physicians’ discharge note in 52% to 67% of cases.66 Drug toxicity should be considered a possibility in cases of obscure or poorly explained liver disease, particularly in cases in which mixed or atypical patterns of cholestasis and hepatitis, cholestasis in which common causes have been excluded, especially in the elderly, and histologic features suggestive of a drug etiology are observed. In such cases, the drug history must be addressed as a special investigation, with attention paid to additional sources of information (household members, primary care providers), household drugs, non-prescribed medications, and environmental toxins (see Chapter 87).

EXCLUSION OF OTHER DISORDERS

Other diseases must be excluded before hepatobiliary disease can be ascribed to a drug. For acute and chronic hepatocellular reactions, viral and autoimmune causes of hepatitis and vascular and metabolic disorders must be considered. Some types of drug-induced chronic hepatitis are associated with autoantibodies and superficially resemble autoimmune hepatitis. An approach to the correct diagnosis is described later (see Nitrofurantoin). Drug-induced cholestasis should be considered only when dilatation of the bile duct has been excluded by imaging. In older patients, and particularly when drug exposure does not include agents known to cause cholestasis, cholangiography (e.g., magnetic resonance imaging [MRCP], endoscopic retrograde cholangiography [ERCP]) is obligatory, as is liver biopsy. Drugs and metabolic factors may interact to cause steatohepatitis, as discussed later.

EXTRAHEPATIC FEATURES

The constellation of rash, eosinophilia, and other organ involvement is relatively specific for an adverse drug reaction as a cause of liver disease (see earlier). These findings, however, are present in only a minority of cases, so their absence is not helpful. In particular, drugs that cause idiosyncratic liver injury by nonimmunologic mechanisms are not usually associated with extrahepatic features. Specific diagnostic tests for individual drug-induced liver diseases have been described57 but are not generally accepted or available. In the case of dose-dependent hepatotoxins, blood levels may be helpful (see later).

CHRONOLOGIC RELATIONSHIPS

For most drugs, the chronologic relationship among drug ingestion, onset, and resolution of liver injury remains the most important consideration in diagnosis. The criteria for temporal eligibility include the relationship of drug ingestion to onset, course of the reaction after discontinuation of

the drug, and response to readministration of the drug.5,6-9 Deliberate rechallenge can be hazardous and is rarely indicated for logistic and ethical reasons, but inadvertent rechallenge may have occurred already. The rechallenge is regarded as positive if the serum ALT or alkaline phosphatase level increases at least two-fold.1,6,9 Deliberate rechallenge may be considered to ascertain whether a drug that is important for an individual patient is responsible for hepatotoxicity (e.g., amiodarone needed for refractory ventricular tachycardia). In other cases, documenting the propensity of newer agents, hitherto unrecognized as hepatotoxins, to cause liver injury may be desirable. Written informed consent is required for a deliberate rechallenge.

WHICH DRUG?

New and nonproprietary compounds should arouse particular suspicion. For patients who are taking multiple drugs, the agent started most recently before the onset of liver injury is most likely to be responsible. If that agent is unlikely to be the culprit and another well-known hepatotoxin is being taken, the latter is the more likely culprit. When possible, the most likely hepatotoxin or all therapeutic agents should be discontinued. If the patient improves, the drugs that are unlikely to be responsible can be carefully reintroduced.

INDICATIONS FOR LIVER BIOPSY

Liver biopsy may be helpful in difficult cases, especially when temporal relationship between the ingestion of a known hepatotoxic agent and the onset of liver injury is unclear. In practice, for example, the onset of jaundice within two to six weeks of starting an agent such as amoxicillin-clavulanic acid or of acute hepatitis in the presence of other features of RMS in a person taking nevirapine as part of HAART would be contexts in which the suspicion of a drug etiology is so strong that liver biopsy is unnecessary. Conversely, substantially abnormal liver biochemical test levels (e.g., a serum ALT level elevated more than five-fold) in a person who has one or more autoantibodies suggestive of autoimmune hepatitis and has been taking a statin or other cardiovascular drug for three to six months may constitute a clinical challenge that often can be resolved only by liver biopsy. The community may benefit when new instances or patterns of drug-induced liver disease are adequately defined; this benefit may persuade the clinician (but not always the informed patient) to proceed with a liver biopsy in equivocal cases.

CONSIDERATIONS IN PATIENTS WITH VIRAL HEPATITIS

Patients with chronic hepatitis B or C may be at higher risk of liver injury from antituberculosis chemotherapy, ibuprofen and possibly other NSAIDs, anti-cancer drugs, and HAART compared with persons without viral hepatitis.26-30 A more common clinical problem is the finding of a serum ALT level greater than 300 U/L at a routine office visit in a patient with previous levels less than 150 U/L. In patients with hepatitis C, the rise in serum ALT is more likely the result of drug toxicity than a spontaneous change in the activity of the hepatitis C, particularly when the ALT level is greater than 1000 U/L.68 The most commonly implicated agents are acetaminophen taken in moderate doses under conditions of increased risk (e.g., fasting, alcohol excess, use of other medication) and CAM preparations, typically Chinese herbal remedies (see Chapter 87). Clinical suspicion is essential for recognizing the drug cause of liver

1425

1426

Section IX  Liver injury so that appropriate advice can be given. Determination of blood levels of acetaminophen also may be useful in difficult cases, but levels (particularly undetectable levels) can be difficult to interpret in the context of regular ingestion, as opposed to a single episode of self-poisoning.

PREVENTION AND MANAGEMENT With the exception of acetaminophen hepatotoxicity (discussed later), little effective treatment for drug-induced liver disease is available. Special emphasis, therefore, must be placed on prevention and early detection of liver injury as well as on prompt withdrawal of the offending agent. Safe use of self-medication of agents such as acetaminophen, NSAIDs, and CAM preparations is important. Clear and open communication between the physician and patient and appropriate recommendations about dose limitations could prevent most instances of liver injury from these agents. The majority of drugs associated with drug-induced liver disease are idiosyncratic hepatotoxins, for which liver injury occurs rarely. The overall frequency of adverse hepatic reactions can be minimized by avoiding overuse of these drugs; antibiotics such as amoxicillin-clavulanic acid and flucloxacillin are pertinent examples. Similarly, polypharmacy should be avoided when possible. The rarity of adverse drug reactions also means that the hepatotoxic potential of new agents may not be recognized until after their introduction. Therefore, all physicians share the responsibility to report suspected adverse effects to monitoring agencies during postmarketing surveillance of new drugs. For dose-dependent hepatotoxins, prevention depends on adherence to dosage guidelines or use of blood levels. This approach has virtually abolished some forms of druginduced liver injury, such as tetracycline-induced fatty liver, aspirin hepatitis, and methotrexate-induced hepatic fibrosis. In cases with specific risk factors, strategies to prevent toxicity are essential (e.g., avoid use of valproic acid with other drugs in the very young; do not prescribe methotrexate to persons who consume alcohol in excess). Moderate doses of acetaminophen are contraindicated in heavy drinkers and after fasting,21 and administration of halothane should not be repeated within 28 days or in persons suspected of previous sensitivity to a haloalkane anesthetic. Early detection is also critical. Patients should be warned to report any untoward symptoms, particularly unexplained nausea, malaise, right upper quadrant abdominal pain, lethargy, or fever. These nonspecific features may represent the prodrome of drug-induced hepatitis. They are an indication for liver biochemical testing and, if the results suggest liver injury, for cessation of treatment. A more difficult issue is whether regular (protocol) screening with liver biochemical tests should be performed when a drug is prescribed. Although such screening often is recommended by authors and drug manufacturers, the efficiency and cost-effectiveness of this approach are unknown. The onset of liver injury is often rapid, rendering once-a-month or every-second-week screening futile. Furthermore, 7.5% of persons who receive placebo in clinical trials have persistently raised serum ALT levels.70 If liver biochemical test levels are monitored, the level of abnormality at which a drug should be discontinued is uncertain, as illustrated by isoniazid, which causes some liver biochemical test abnormality in 30% of exposed sub-

jects. Generally, the recommendation is that isoniazid be stopped if serum ALT levels exceed 250  U/L or more than five times the upper limit of normal, but elevation of the serum bilirubin or albumin concentration or prolongation of the prothrombin time provides a clearer indication to stop the drug. Conversely, a rise in the serum GGTP level or a minor elevation of serum alkaline phosphastase level usually indicates hepatic adaptation rather than liver injury. We do not routinely recommend protocol screening, but this approach could be useful for agents such as valproic acid, isoniazid, pyrazinamide, ketoconazole, dantrolene, tacrine, thiazolidinediones, and synthetic retinoids, either because the onset of liver injury may be delayed and gradual in some cases or because such screening can emphasize the hepatotoxic potential of these drugs to patients and physicians. Liver biopsy has a role in the assessment of hepatic fibrosis in patients who take methotrexate (see later). Highly toxic solvents should be avoided in the workplace, and such agents have been abandoned. Adequate ventilation and use of masks and protective clothing are vital to prevent occupational exposure to hepatotoxic chemicals. In some cases, liver biochemical tests are performed routinely in exposed persons, but abnormalities are more likely to reflect diseases such as chronic hepatitis C, alcoholism, and NAFLD than toxic liver injury. In the case of vinyl chloride exposure, periodic physical examination (for hepatomegaly) and hepatic imaging with ultrasonography may be useful (see Chapter 87). Active management of drug-induced liver injury includes removal of the drug and administration of antidotes and anti-inflammatory and cytoprotective agents. In practice, treatment usually is confined to discontinuation of the hepatotoxic drug. Failure to discontinue a drug that is the cause of liver injury is the single most important factor leading to poor outcomes, such as acute liver failure and chronic liver disease.8,9 For ingested toxins such as metals, poisonous mushrooms, and acetaminophen, removal of the unabsorbed drug through the aspiration of stomach contents may be appropriate. Methods to remove absorbed toxins, such as hemodialysis through a charcoal column and forced diuresis, are not effective for hepatotoxins. For chlordecone, an organochlorine insecticide that is lipid-soluble and excreted in bile, oral administration of cholestyramine enhances removal of the agent from the body by interrupting the enterohepatic cycle.71 Thiol replacement therapy, usually with N-acetylcysteine (NAC), is indicated as an antidote for acetaminophen poisoning. Whether NAC or other antioxidants have a role in other types of acute hepatotoxicity is unclear, but the flavonoid, silybin (silymarin), is traditionally used for Amanita phalloides toxicity72 and tocopherol analogs show promise in experimental hepatotoxicity (see Chapter 87). Beyond discontinuation of the offending agent, the management of drug hepatitis and cholestasis is symptomatic and supportive. In cases of acute liver failure, hepatic transplantation should be considered (see Chapters 93 and 95).7 Ursodeoxycholic acid has shown some promise in the management of chronic cholestasis and pruritus caused by drug hepatotoxicity. Glucocorticoids have little role in the management of drug-induced cholestasis or hepatitis and are ineffective in chlorpromazine-, methyldopa-, and isoniazidinduced hepatitis and in drug-induced fulminant hepatic failure. Case reports attest to the occasional effectiveness of glucocorticoids in protracted cases of hepatitis caused by etretinate, allopurinol, diclofenac, or ketoconazole.5 Glucocorticoids should be reserved for atypical and refractory cases, particularly those associated with vasculitis. Clinical

Chapter 86  Liver Disease Caused by Drugs evidence of the effectiveness of putative hepatoprotective agents, such as prostaglandin analogs, is lacking.

DOSE-DEPENDENT HEPATOTOXICITY Few dose-dependent hepatotoxins are clinically important today. Examples include acetaminophen, some herbal medicines (CAM preparations), plant and fungal toxins, amodiaquine, hycanthone, vitamin A, methotrexate, cyclophosphamide, anti-cancer drugs, carbon tetrachloride, phosphorus, and metals (especially iron, copper, and mercury). Acetaminophen is by far the most important of these; hepatotoxicity caused by CAM preparations is discussed in Chapter 87.

Table 86-3  Risk Factors for Acetaminophen-Induced Hepatotoxicity FACTOR

RELEVANCE

Age

Children may be more resistant than adults Minimal hepatotoxic dose: 7.5g (≈100 mg/kg) in adults, 150 mg/kg in children Severe toxicity possible with dose >15 g Influenced by dose, time after ingestion, gastric emptying Best indicator of risk of hepatotoxicity (see text and Fig. 86-2) Toxic dose threshold lowered; worsens prognosis (also related to late presentation); nephrotoxicity common Toxic dose threshold lowered— therapeutic misadventure (see text) Toxic dose threshold lowered— therapeutic misadventure; worsens prognosis (e.g., isoniazid, phenytoin, zidovudine) Late presentation or delayed treatment (>16 hr) predicts worse outcome

Dose

Blood level

Chronic excessive alcohol ingestion Fasting

ACETAMINOPHEN General Nature, Frequency, and Predisposing Factors

Acetaminophen (paracetamol) is a widely used analgesic available without prescription. It is safe when taken in the recommended therapeutic dose of 1 to 4 g daily, but hepatotoxicity produced by self-poisoning with acetaminophen has been recognized since the 1960s. Despite the effectiveness of thiol-based antidotes, acetaminophen remains the most common cause of drug-induced liver injury in most countries and an important cause of acute liver failure.7,73 Parasuicide and suicide are the usual reasons for overdose.73,74 Although controversial,75,76 hepatologists and pediatricians see cases of acetaminophen poisoning that have arisen through what Zimmerman and Maddrey termed therapeutic misadventure.77 This occurrence is especially common in persons who habitually drink alcohol to excess and has also been recognized after daily ingestion of moderate therapeutic doses (10 to 20 g over three days) of acetaminophen in adults and children who are fasting or malnourished21 or who are taking drugs that interact with the metabolism of acetaminophen.77 Single doses of acetaminophen that exceed 7 to 10 g (140 mg/kg body weight in children) may cause liver injury, but this outcome is not inevitable. Severe liver injury (serum ALT level greater than 1000 U/L) or fatal cases usually involve doses of at least 15 to 25 g, but because of interindividual variability, survival is possible even after ingestion of a massive single dose of acetaminophen (greater than 50 g).78 Among persons with an untreated acetaminophen overdose, severe liver injury occurred in only 20%, and among those with severe liver injury, the mortality rate was 20%.78 Conversely, among heavy drinkers, daily acetaminophen doses of 2 to 6 g have been associated with fatal hepatotoxicity.75-78 Risk factors for acetaminophen-induced hepatotoxicity are summarized in Table 86-3. Children are relatively resistant to acetaminophen-induced hepatotoxicity,79 possibly because of their tendency to ingest smaller doses, greater likelihood of vomiting, or biological resistance. Therapeutic misadventure after multiple doses, especially during fasting and when weight-based recommendations have been exceeded, has a high mortality rate. By contrast, the presence of underlying liver disease does not predispose to acetaminophen hepatotoxicity. Self-poisoning with acetaminophen is most common in young women, but fatalities are most frequent in men, possibly because of alcoholism and late presentation.73-75 The time of presentation is critical because thiol therapy given within 12 hours of acetaminophen poisoning virtually abolishes significant liver injury (see later). Therapeutic

Concomitant medication

Time of presentation

misadventure is also associated with a worse outcome.76 Concomitant use of agents such as phenobarbital, pheny­ toin, isoniazid, and zidovudine is another risk factor for acetaminophen hepatotoxicity. These drugs may promote the oxidative metabolism of acetaminophen to NAPQI by inducing CYP2E1 (for isoniazid) or CYP3A4 (for phenytoin) or by competing with glucuronidation pathways (for zidovudine). Alcohol and fasting have dual effects by enhancing expression of CYP2E1 and by depleting hepatic glutathione. Fasting also may impair acetaminophen conjugation by depleting cofactors for the glucuronidation and sulfation pathways.21 Acetaminophen hepatotoxicity produces zone 3 hepatic necrosis, with extension into submassive (bridging) or panacinar (massive) necrosis in severe cases. Inflammation is minimal, and recovery is associated with complete resolution without fibrosis. The zonal pattern of acetaminophen-induced necrosis is related to the mechanism of hepatotoxicity, particularly the role of CYP2E1, which is expressed in zone 3, and to lower levels of glutathione in zone 3 hepatocytes than in hepatocytes in the other zones.

Clinical Course, Outcomes, and Prognostic Indicators

In the first two days after acetaminophen self-poisoning, features of liver injury are not present. Nausea, vomiting, and drowsiness are often caused by concomitant ingestion of alcohol and other drugs. After 48 to 72 hours, serum ALT levels may be elevated, and symptoms such as anorexia, nausea and vomiting, fatigue, and malaise may occur. Hepatic pain may be pronounced. In severe cases, the course is characterized by repeated vomiting, jaundice, hypoglycemia, and other features of acute liver failure, particularly coagulopathy and hepatic encephalopathy. The liver may shrink as a result of severe necrosis. Serum levels of ALT are often between 2000 and 10,000 U/L. These high levels, together with high levels of other intracellular proteins (ferritin, glutathione S-transferases), may provide a clue to the diagnosis in complex settings, as may occur with alcoholic patients and those with viral hepatitis.77 Indicators of a poor outcome73-76 include grade IV hepatic coma, acidosis, severe and sustained impairment of coagu-

1427

Section IX  Liver lation factor synthesis, renal failure, and a pattern of falling serum ALT levels in conjunction with a worsening prothrombin time (see also Chapter 93). Renal failure reflects acute tubular necrosis or the hepatorenal syndrome. Myocardial injury also has been attributable to acetaminophen toxicity.78 Death occurs between 4 and 18 days after the overdose and generally results from cerebral edema and sepsis complicating hepatic and multiorgan failure. A majority of patients recover completely. Cases of apparent chronic hepatotoxicity rarely have been attributed to continued ingestion of acetaminophen (2 to 6 g/day), usually in a susceptible host, such as a heavy drinker or a person with preexisting, unrecognized liver disease.5,6 Rare cases of acetaminophen hypersensitivity, typically involving skin or lung, have been reported in association with liver injury.80,81

Management

In patients who present within four hours of ingesting an excessive amount of acetaminophen, the stomach should be emptied with a wide-bore gastric tube. Osmotic cathartics or binding agents have little if any role in management. Charcoal hemoperfusion has no established role. The focus of management is on identifying patients who should receive thiol-based antidote therapy and, in those with established severe liver injury, assessing the patient’s candidacy for liver transplantation. Blood levels of acetaminophen should be measured at the time of presentation. Because of delayed gastric emptying, however, blood levels within four hours of ingestion may underestimate the extent of exposure. After four hours, acetaminophen blood levels give a reliable indicator of the risk of liver injury in patients with an acute overdose (not in those with a therapeutic misadventure). The risk of liver injury is then estimated by reference to the Prescott nomogram (Fig. 86-2).78 Indications for antidote therapy include a reliable history of major poisoning (more than 10 g) or blood acetaminophen levels in the moderate or high-risk bands on the monogram, or both.74,78 At-risk patients should be hospitalized for monitoring. Hepatic necrosis occurs only when glutathione concentrations fall below a critical level, thereby allowing NAPQI to produce liver injury. Administration of cysteine donors stimulates hepatic synthesis of glutathione. Many cysteine precursors or thiol donors could be used, but NAC has become the agent of choice. Oral administration is preferred in the United States,73,78 with a loading dose of 140 mg/kg followed by administration of 70 mg/kg every 4 hours for 72 hours. This regimen is highly effective, despite the theoretical disadvantage that delayed gastric emptying and vomiting may reduce intestinal absorption of NAC. In Europe and Australia, NAC is administered by slow bolus intravenous injection followed by infusion (150 mg/kg over 15 minutes in 200 mL of 5% dextrose, with a second dose of 50 mg/kg 4 hours later, if the blood acetaminophen levels indicate a high risk of hepatoxicity, and a total dose over 24 hours of 300 mg/kg).78 The intravenous route may be associated with a higher rate of hypersensitivity reactions because of the higher systemic blood levels achieved.5 Adverse reactions to NAC may be severe, with rash, angioedema, and shock, which occasionally is fatal.5 Therefore, NAC must be administered under close supervision and only for appropriate indications. In patients known to be sensitized to NAC, methionine is probably just as effective but is not available in a commercial preparation; it must be made up fresh and often causes vomiting.78 Cases of acetaminophen-induced severe liver injury are virtually abolished if NAC is administered within 12 hours and possibly within 16 hours of acetaminophen inges-

µg/mL µmol/L

Acetaminophen plasma concentration

1428

300

2000

200 150

1300 1000 900 800 700 600 500 400

100 90 80 70 60 50 40 30

300 250 200

Potential for toxicity

20

10 9 8 7 6 5 4 3

Toxicity unlikely

100 90 80 70 60 50 40 30

Treatment recommended if level is above broken line

20

2 10

4 Measure level at least 4 hours post-ingestion

8

12

16

20

24

28

32

36

Hours post-ingestion

Figure 86-2.  Acetaminophen toxicity nomogram. The risk of hepatotoxicity correlates with the plasma acetaminophen level and the time after ingestion. (From Smilkstein MJ, Knapp GL, Kulig KW, et al. Efficacy of oral N-acetylcysteine in the treatment of acetaminophen overdose. Analysis of the National Multicenter Study [1976-1985]. N Engl J Med 1988; 319:1557-62.)

tion.73,74,78 After 16 hours, thiol donation is unlikely to affect the development of liver injury because oxidation of acetaminophen to NAPQI with consequent oxidation of thiol groups is complete and mitochondrial injury and activation of cell death pathways are likely to be established. Nevertheless, NAC has been reported to decrease the mortality associated with acetaminophen-induced hepatotoxicity when administered 16 to 36 hours after self-poisoning,73,74,78 possibly because NAC stabilizes vascular reactivity in patients with liver failure. Therefore, administration of NAC is recommended for patients with a late presentation after acetaminophen overdose. Other strategies to protect the liver against acetaminophen poisoning, such as inhibition of CYP-dependent metabolism through the use of cimetidine or administration of prostaglandin analogs, which are efficacious in rats, have not been established as clinically useful. The constitutive androstane receptor (CAR) has been identified as a regulator of acetaminophen metabolism and hepatotoxicity in mice.82 Inhibition of CAR activity by administration of androstanol one hour after acetaminophen dosing blocks liver injury.83 Liver transplantation has been advocated as a therapeutic option for select patients in whom liver failure develops after acetaminophen poisoning.73,74 The selection of cases is based on the prognostic indicators discussed earlier and is strongly influenced by the prospects for successful psychological rehabilitation (see Chapter 95).74 In several series,

Chapter 86  Liver Disease Caused by Drugs about 60% of listed patients have been transplanted, and survival rates have exceeded 70%.74

Prevention

Safe use of acetaminophen involves adherence to the recommended maximum dose for healthy adults and children and education about the risk factors that lower the toxic dose threshold. Acetaminophen doses of more than 2 g a day are contraindicated in heavy drinkers, in those taking other medications (particularly phenytoin, zidovudine, and isoniazid), and during fasting. Prolonged use of acetaminophen requires caution in patients with severe cardiorespiratory disease or advanced cirrhosis. Use of acetaminophen for self-poisoning continues despite attempts at public education about the risks involved. The chances of harm from a suicidal gesture may be reduced by the sale of acetaminophen in smaller package sizes and in blister packs, which hamper ready access to the tablets or capsules.84,85

OTHER TYPES OF CYTOPATHIC LIVER INJURY

Some hepatotoxins are not as clearly dose dependent as acetaminophen but cause cytopathic or cytotoxic changes, such as extensive hydropic change, diffuse or zonal microvesicular steatosis, and zonal necrosis.5,6 Injury likely represents metabolic idiosyncrasy, in which the drug or one of its metabolites accumulates and interferes with protein synthesis or intermediary metabolism, or both. The mitochondrion often appears to be the main subcellular target, and other metabolically active tissues can be involved. Pancreatitis and renal tubular injury may accompany severe liver injury caused by valproic acid, tetracycline, and HAART, and metabolic acidosis with a shock-like state is common. The first agent recognized to cause this clinicopathologic syndrome was tetracycline administered in high doses (greater than 2 g/day for more than four days, usually intravenously) to pregnant women, to men taking estrogens, or to patients with renal failure.6 With appropriate dose limitations, this reaction is entirely preventable.

Niacin (Nicotinic Acid)

Hepatotoxicity associated with use of niacin, or nicotinic acid (3-pyridinecarboxylic acid), has been noted since the 1960s. When used to treat hypercholesterolemia, niacin has been an important cause of liver injury.86 It is a dosedependent hepatotoxin; liver injury usually occurs at doses that exceed 2 g/day, but in rare instances, low-dose (500 mg/ day) sustained-release niacin has been implicated in fulminant hepatic failure.87 Patients who are taking sulfonylurea drugs and those with preexisting liver disease, particularly alcoholic hepatitis, are at increased risk. No association with age, diet, or insulin-managed diabetes mellitus has been recognized. The symptoms of niacin hepatotoxicity begin as early as one week to as long as four years after the drug is started. The clinicopathologic spectrum encompasses mild and transient increases in serum ALT levels, jaundice, acute hepatitis, and cholestasis. Liver injury resolves completely when the drug is stopped. Liver biopsy specimens show hepatic necrosis and centrilobular cholestasis. Well-documented cases of fulminant hepatitis, some necessitating liver transplantation, also have been attributed to niacin. Substitution of one niacin preparation for another without a dose adjustment should be avoided; switching from immediate- to sustained-release preparations requires a 50% to 70% reduction in the dose of niacin.

Valproic Acid (Sodium Valproate)

Valproic acid-associated hepatic injury occurs almost exclusively in children, particularly those younger than three

years of age. Also at risk are persons with a family history of a mitochondrial enzyme deficiency (particularly involving the urea cycle or long-chain fatty acid metabolism), Friedreich’s ataxia, or Reye’s syndrome or with a sibling affected by valproic acid hepatotoxicity. Another risk factor is multiple drug therapy. Cases in adults have been described rarely. The overall risk of liver injury among persons taking valproic acid varies from 1 per 500 persons exposed among high-risk groups (children under age 3, polypharmacy, genetic defects of mitochondrial enzymes) to less than 1 in 37,000 in low-risk groups.88 No relationship exists between valproic acid toxicity and dose, but blood levels of valproic acid tend to be high in one half of affected persons. The metabolite 4-en-valproic acid, produced by CYP-catalyzed metabolism of valproic acid, is a dose-dependent hepatotoxin in animals and in vitro. The concept has emerged that valproic acid is an occult dose-dependent toxin in which accumulation of a hepatotoxic metabolite (favored by coexposure to CYPinducing antiepileptic agents) produces mitochondrial injury in a susceptible host (e.g., young children, especially those with partial deficiencies of mitochondrial enzymes).89 Symptoms begin 4 to 12 weeks after treatment with valproic acid is started and are often nonspecific, including lethargy, malaise, poor feeding, somnolence, worsening seizures, muscle weakness, and facial swelling. In typical cases, features of hepatotoxicity follow, including anorexia, nausea, vomiting, abdominal discomfort over the liver, and weight loss.88,89 When jaundice ensues, hypoglycemia, ascites, coagulation disorders, and encephalopathy indicate liver failure with imminent coma and death. In some cases, a neurologic syndrome characterized by ataxia, mental confusion, and coma predominates, with little evidence of hepatic involvement. In other cases, fever and tender hepatomegaly suggestive of Reye’s syndrome may be present (see later); such cases tend to have a better prognosis. Additional extrahepatic features may include alopecia, hypofibrinogenemia, thrombocytopenia, and pancreatitis. The terminal phase is often indicated by renal failure, hypoglycemia, metabolic acidosis, and severe bacterial infection. Laboratory features include modest elevations of serum bilirubin and aminotransferase levels; the aspartate aminotransferase (AST) level is usually higher than the ALT level. A profound decrease in clotting factor levels, hypoalbuminemia, and raised serum ammonia levels are common. A small liver with increased echogenicity suggestive of steatosis or extensive necrosis is seen on hepatic imaging. Histologic examination of the liver shows submassive or massive hepatic necrosis in two thirds of cases with either zonal or generalized microvesicular steatosis.89 Ultrastructural studies indicate conspicuous abnormalities of the mitochondria. Treatment is supportive. Prevention depends on careful adherence to prescribing guidelines, including the avoidance of valproic acid in combination with other drugs in the first three years of life and in children who may have mitochondrial enzyme abnormalities. Elevations in liver biochemical test levels develop in at least 40% of patients who take valproic acid and therefore are an unreliable predictor of severe hepatotoxicity. Warning patients and parents about the need to report any adverse symptoms during the first six months of valproic acid therapy is most important.

Antiretroviral Agents

Elevated liver biochemical test levels and clinical evidence of liver disease are common in patients with HIV/AIDS. Reasons for this finding include HBV or HCV infection,

1429

1430

Section IX  Liver other hepatobiliary infections, lymphoma and other tumors, and possibly effects of HIV infection itself. The frequency of hepatic injury with HAART (which often includes three or four agents) is at least 10%.23,27,28,90 The agents used can be broadly categorized as nucleoside (or nucleotide) reverse transcriptase inhibitors, non-nucleoside reverse transcriptase inhibitors, and protease inhibitors. Because HIV coinfection with HBV or HCV increases the risk of toxicity, all patients should be screened for viral hepatitis before starting HAART.28 Nucleoside (or Nucleotide) Reverse Transcriptase Inhibitors Nucleosides and nucleotides that block HIV reverse transcriptase are also weak inhibitors of mitochondrial DNA polymerase gamma in vitro; the order of their potency is: zalcitabine > didanosine > stavudine > lamivudine > zi­ dovudine > abacavir.91 The mechanisms of hepatotoxicity may also involve oxidative stress, resulting in further deletion of mitochondrial DNA, and the consequences of impaired oxidative phosphorylation, fatty acyl β-oxidation, and insulin resistance. In clinical studies, zidovudine, didanosine, and stavudine are the agents implicated most often in liver injury.90-93 Risk factors for mitochondrial drug toxicity among persons with HIV infection include obesity, female gender, and absence of an AIDS-defining illness.90-94 Hallmarks of mitochondrial hepatotoxicity include extensive microvesicular or macrovesicular steatosis (or both), lactic acidosis, and liver biochemical test abnormalities with progression to acute liver failure. Asymptomatic hyperlactatemia is common (especially with stavudine) among persons treated with HAART,94 but life-threatening lactic acidosis with hepatic steatosis is rare, with an estimated risk of 1.3 per 1000 person-years of antiretroviral use. The onset is a median of 6 months (with a range of 3 to 17 months) after treatment is started. Patients present with symptoms that are nonspecific and include nausea, vomiting, diarrhea, dyspnea, lethargy, and abdominal pain. Extrahepatic manifestations, such as myopathy or peripheral neuropathy, and in severe cases pancreatitis and renal failure, may follow the onset of the lactic acidosis and liver injury. Discontinuation of the drug is mandatory but does not prevent fatalities. Nevertheless, the overall mortality rate is low. One suggested approach to prevention is to monitor therapy with nucleos(t)ide reverse transcriptase inhibitors by coupling serum ALT and AST testing with serial measurements of the HIV load and CD4 count. Any new aminotransferase elevation should be followed immediately by measurement of serum lactate, muscle, and pancreatic enzyme levels.92 Non-nucleoside Reverse Transcriptase Inhibitors Non-nucleoside reverse transcriptase inhibitors occasionally may cause hepatitis as part of a hypersensitivity reaction within the first six weeks of use.90,95,96 Reactions are usually accompanied by peripheral and tissue eosinophilia, skin rash, and lymphadenopathy. Resolution occurs within four weeks of discontinuing the drug.95 Nevirapine also has been implicated in several instances of severe hepatotoxicity,97,98 including cases among healthcare workers in whom nevirapine was used for post-exposure prophylaxis against HIV infection.97 The FDA received 12 reports of such hepatotoxic reactions between 1997 and 2000; liver failure requiring hepatic transplantation developed in one person, seven had clinical features of hepatitis (jaundice, fever, nausea, vomiting, abdominal pain, and hepatomegaly), and four others had elevated serum aminotransferase levels without symptomatic illness. The recommended two-week

dose escalation regimen was not adhered to in some of the cases.99 Sequential toxicity with nevirapine followed by efavirenz has been reported in an HIV-HCV coinfected person.100 Protease Inhibitors Elevation of liver enzymes occurs commonly with protease inhibitors, but clinical hepatitis is infrequent. The agents most often implicated in liver injury are ritonavir and indinavir. The latter also may be associated with unconjugated hyperbilirubinemia in 7% of treated persons, a finding that is of no clinical consequence.9 Severe acute hepatitis may occur rarely. The association with peripheral or tissue (in liver biopsy specimens) eosinophilia in some cases suggests an immunoallergic basis for liver injury.101,102 Acute hepatitis also has been reported in 2.9% to 30% of persons who take ritonavir.103 The course of the illness is generally mild, and the liver injury responds favorably to drug withdrawal. Rarely, acute liver failure may develop; in these cases, liver histologic examination has shown severe microvesicular steatosis, cholestasis, and extensive fibrosis. Several studies have addressed the potential influence of underlying chronic viral hepatitis on the toxicity of protease inhibitors. Although hepatotoxicity appeared to be more common, liver injury was rapidly reversible in most cases; this observation suggests that the overall effect of protease inhibitors in coinfected persons is not detrimental.104 Many protease induce or inhibit CYP3A4, thereby causing important drug-drug interactions.105 Furthermore, the immune reconstitution that can follow successful HAART may cause a flare-up of previously quiescent chronic hepatitis B (see Chapter 33).

Aspirin

Aspirin occasionally has been associated with major increases in serum ALT levels suggestive of drug hepatitis, but hepatotoxicity occurs only when blood salicylate concentrations exceed 25 mg/100 mL.106 In addition, individual susceptibility factors include hypoalbuminemia, active juvenile rheumatoid arthritis, and systemic lupus erythematosus. Most cases of aspirin-induced hepatotoxicity have been identified by biochemical testing, rather than clinical features. If present, symptoms usually begin within the first few days or weeks of high-dose aspirin therapy. Acute liver failure and fatalities have been rare. Resolution occurs rapidly after drug withdrawal, and salicylates can be reintroduced at a lower dose. All salicylates appear to carry hepatotoxic potential so there is no advantage to replacing aspirin with another salicylate. Liver biopsy specimens reveal a nonspecific focal hepatitis with hepatocellular degeneration and hydropic changes. Steatosis is not usually present, and the absence of steatosis distinguishes aspirin hepatotoxicity from Reye’s syndrome. Reye’s syndrome has been linked with use of aspirin in febrile children. Although Reye’s syndrome is not simply a form of drug-induced liver disease, aspirin plays an important role in its multifactorial pathogenesis. Reye’s syndrome usually occurs between three and four days after an apparently minor viral infection. It is characterized by acute encephalopathy and hepatic injury, the latter documented by a three-fold or greater rise of serum aminotransferase or ammonia levels and by characteristic histologic findings. Because of effective public health campaigns against the use of aspirin in young febrile children, the incidence of Reye’s syndrome has declined markedly.107 Patients with juvenile rheumatoid arthritis (Still’s disease) or systemic lupus erythematosus appear to be at particular risk of Reye’s syndrome. Clinical and laboratory features of

Chapter 86  Liver Disease Caused by Drugs Table 86-4  Types of Drug-Induced Acute Hepatitis: Immunoallergic Reaction versus Metabolic Idiosyncrasy CHARACTERISTIC

IMMUNOALLERGIC reaction

METABOLIC IDIOSYNCRASY

Frequency Gender predilection Latent period to onset Relationship to dose Interactions with other agents

<1 case per 10,000 persons exposed Women, often ≥2:1 Fairly constant, 2-10 weeks None None

Course after stopping drug

Prompt improvement (rare exceptions, e.g., minocycline) Always; often fever within 3 days Usual; often initial symptom, part of prodrome Common

1-50 cases per 10,000 persons exposed Variable, slightly more common in women More variable, 2-24 weeks, occasionally longer than 1 year Usually none (occasional exceptions) Alcohol; occasionally other drugs (e.g., isoniazid with rifampin) Variable; occasionally slow improvement or deterioration (e.g., troglitazone) Usual (in two thirds of cases), abnormal liver biochemical test levels in 2-21 days Infrequent, less prominent Rare

33%-67% of cases Usual, pronounced Often present Nitrofurantoin, phenytoin, methyldopa, sulfonamides, etretinate, minocycline

<10% of cases Common but mild Rarely present Isoniazid, pyrazinamide, ketoconazole, dantrolene, troglitazone

Positive rechallenge Fever Extrahepatic features (rash, lymphadenopathy) Eosinophilia   Blood   Tissue Autoantibodies Examples

chronic liver disease do not develop in affected persons, and features of drug allergy are not present. Management requires suspecting the correct diagnosis and reduction in the dose (or discontinuation) of aspirin. Recovery is usually rapid. Aspirin can be used again in lower doses, but other NSAIDs have displaced the use of high-dose aspirin for most conditions. The incidence of Reye’s syndrome in the United States, Great Britain, and elsewhere has fallen dramatically since aspirin use has been avoided in children with a viral illness,16,17 although misdiagnosis in the past of cases that subsequently were shown to be caused by inborn errors of metabolism that mimic Reye’s syndrome may account in part for the apparent decline in the incidence of Reye’s syndrome.107

Other Drugs

is an antileukemic drug that often causes hepatotoxicity, which usually is reversible but can result in liver failure associated with diffuse microvesicular steatosis.108 Amodiaquine, a 4-aminoquinolone antimalarial agent, has been associated with fatal hepatotoxicity, as well as with agranulocytosis.109 Toxicity may be related to the total dose of the drug. Amodiaquine should be reserved for active treatment of chloroquine-resistant falciparum malaria, and dose recommendations should be strictly observed. Hycanthone is an antischistosomal agent that causes dose-dependent hepatoxicity. Risk factors for hepatotoxicity include concomitant administration of phenothiazines or estrogens, preexisting liver injury, and bacterial infection, but the most important risk factor is dose.110 The hepatotoxic effects of environmental toxins, illicit drug abuse substances, and metals are discussed in Chapter 87. l-asparaginase

DRUG-INDUCED ACUTE HEPATITIS The term acute hepatitis is used to describe lesions characterized by the presence of hepatic inflammation with conspicuous hepatocyte cell death or degeneration. More severe lesions include zonal and bridging necrosis or massive (pan-

lobular) hepatic necrosis; these lesions may be associated with acute (fulminant or subfulminant) hepatic failure.5,6 Acute hepatitis accounts for nearly 50% of reported adverse drug reactions involving the liver,1-4 and potential causative agents are numerous.5,6,8,110,111 Two broad types of drug hepatitis are those cases with clinical and laboratory features consistent with drug allergy (immunoallergic reactions) and those without such features. The latter type could be the result of metabolic idiosyncrasy; partial dose dependence, a relationship between hepatitis and metabolism of the drug, and histologic or ultrastructural features consistent with chemical toxicity are often found. The clinical and laboratory features that suggest one or the other type of drug hepatitis are summarized in Table 86-4. Nitrofurantoin is discussed as an example of immunoallergy, and isoniazid is used to illustrate metabolic idiosyncrasy. Other relatively frequent examples of drug hepatitis are described briefly, including those associated with granulomatous reactions and chronic hepatitis.

IMMUNOALLERGIC REACTIONS Nitrofurantoin

Nitrofurantoin, a synthetic furan-based compound, is a urinary antiseptic agent that continues to lead to cases of hepatic injury.112 The frequency of nitrofurantoin hepatic injury ranges from 0.3 to 3 cases per 100,000 exposed persons.113,114 The risk increases with age, particularly after the age of 64. Two thirds of acute cases occur in women, and the female-to-male ratio is 8 : 1 for chronic hepatitis.113,114 The range of liver diseases associated with nitrofurantoin includes acute hepatitis, occasionally with features of cholestasis, hepatic granulomas, chronic hepatitis with autoimmune phenomena, acute liver failure, and cirrhosis.113,114 Causality has been proved by rechallenge, and no relationship to dose has been observed; cases have even been described after ingestion of milk from a nitrofurantointreated cow.115 The relative frequencies of hepatocellular and cholestatic or mixed reactions and of acute and chronic hepatitis caused by nitrofurantoin have been the subject of debate. The nature of the adverse reactions covers a spectrum of biochemical and histologic features that have no apparent

1431

1432

Section IX  Liver relevance to the patient’s clinical outcome. Chronicity depends mostly on the duration of drug ingestion, which has been less than six weeks in acute cases but more than six months in 90% of chronic cases.113,114 Patients with chronic hepatitis often have continued to take nitrofurantoin despite symptoms attributable to an adverse drug effect or have been exposed to another course of the drug after previous toxicity. The mortality rate for chronic nitrofurantoin hepatitis is 20%, compared with 5% to 10% for acute hepatitis.113 The latent period between initial exposure to the drug and the onset of liver disease ranges from a few days to six weeks. Early symptoms may be nonspecific (e.g., fever, myalgia, arthralgia, fatigue, malaise, anorexia, and weight loss) and are followed by more specific features of hepatitis, such as nausea and vomiting, hepatic pain or discomfort, dark urine, jaundice, and, occasionally, pruritus. Rash occurs in 20% of affected persons, and lymphadenopathy may be present. Pneumonitis, which may be complicated by pulmonary fibrosis, can develop concurrently with hepatitis in 20% of cases and is suggested by cough and dyspnea. Rarely, liver failure develops, with ascites, coagulopathy, and encephalopathy. In patients with chronic hepatitis, clinical findings (such as spider angiomata, hepatosplenomegaly, muscle wasting, and ascites) may suggest chronic liver disease. Liver biochemical testing may show pronounced elevation of serum ALT levels, but more often the pattern is mixed, with an increase in the serum alkaline phosphatase level as well. In other cases, the results suggest cholestasis. Serum bilirubin levels tend to be increased in proportion to the severity of the reaction. In contrast to most types of acute drug hepatitis, serum albumin concentrations often are low. Hypergammaglobulinemia is more likely in patients with chronic hepatitis than in those with acute hepatitis.113 Eosinophilia occurs in 33% of cases. Autoantibodies (antinuclear antibodies and smooth muscle antibodies) are present in some patients with acute hepatitis and in 80% of those with chronic hepatitis. The presence of autoantibodies can make differentiation of nitrofurantoin-induced fulminant hepatitis from autoimmune hepatitis challenging.116 In contrast to idiopathic autoimmune hepatitis, the frequency of the human leukocyte antigens HLA-B8 and -DRw3 is not increased.113,114 No specific treatment for nitrofurantoin hepatitis exists. Glucocorticoids have no role—even in patients with chronic hepatitis with autoimmune features. Recovery is rapid after nitrofurantoin is discontinued. Monitoring liver biochemical test levels in users of nitrofurantoin is unlikely to be useful or cost-effective.

Other Drugs

Methyldopa was one of the first drugs reported to cause immunoallergic drug hepatitis. Cases are now rare because better antihypertensive agents are available, except for pregnancy-related cases.117 Hepatic reactions to methyldopa vary from abnormal liver biochemical test levels, severe acute hepatitis, granuloma formation, and cholestasis to chronic hepatitis with bridging necrosis and cirrhosis. The female predilection, clinical and laboratory changes, course, and extrahepatic features of drug allergy are similar to those for nitrofurantoin. Phenytoin causes severe acute drug hepatitis in less than one per 10,000 persons exposed.118 Incidence rates are equal in men and women, and cases can occur in childhood. Blacks may be more often affected than whites. Rash, fever, eosinophilia, lymphadenopathy, a pseudomononucleosis syndrome, and other allergic features are common. The

clinical features are suggestive of immunoallergy as part of RMS. An individual or familial enzymatic defect that leads to reduced disposal of phenytoin arene oxide has been detected among patients with phenytoin reactions.118 This finding implicates a possible metabolic factor in the pathogenesis of phenytoin toxicity. The mortality rate is 10% to 40%. Some deaths are caused by liver failure, whereas others are the result of severe systemic hypersensitivity, bone marrow suppression, exfoliative dermatitis, or vasculitis involving the skin and kidney. Rarer hepatic associations with phenytoin reactions include cholestatic hepatitis and bile duct injury. The most common association with phenytoin therapy is an adaptive response of the liver with induction of microsomal enzymes; at least two thirds of patients have raised serum GGTP levels, and one third exhibit raised serum alkaline phosphatase levels. On histopathologic examination, ground-glass cytoplasm, which represents hypertrophied smooth endoplasmic reticulum, is usually present in hepatocytes. Barbiturates, including phenobarbital, rarely are associated with acute hepatitis. Described cases have been similar to phenytoin reactions; fever and rash are usual, and the rate of mortality as a result of liver failure is high.119 Among newer antiepileptic drugs, felbamate120 and topiramate121 have been associated with acute liver failure. Sulfonamides are a cause of drug hepatitis that is relatively common with combination drugs such as cotrimoxazole (sulfamethoxazole and trimethoprim).122 Tri­ methoprim alone has been associated with some cases of cholestatic hepatitis; the estimated risk is 1.4 cases per 100,000 exposed persons.122 Reactions to co-trimoxazole resemble those associated with trimethoprim more closely than they resemble sulfonamide reactions; cholestasis or cholestatic hepatitis is more common than is hepatitis. Patients with HIV/AIDS are predisposed to sulfonamide hypersensitivity. Some other drugs have a sulfa moiety that differs from that of sulfonamides but that may increase the risk of cross-sensitivity reactions; for example, celecoxib, a COX-2 inhibitor, has been observed to cause severe hepatitis in two women with a history of sulfonamide sensitivity.123 Likewise, sulfonylureas, such as gliclazide, rarely have been associated with drug hepatitis with features of immunoallergy.124 The latent period between exposure to the drug and the onset of sulfonamide hepatitis is 5 to 14 days, and clinical features often include fever, rash, mucositis (StevensJohnson syndrome), lymphadenopathy, and vasculitis (that is, features of RMS). Reactions may be severe, and deaths have occurred. The serum ALT level is usually increased to a greater degree than the serum alkaline phosphatase level, but mixed or cholestatic reactions occur. A few cases of hepatic granuloma formation and of chronic hepatitis also have been associated with sulfonamides. Sulfasalazine (salicylazosulfapyrine, salazopyrine) has been associated with rare cases of often severe acute hepatitis.125 Although the sulfonamide moiety has been associated to be responsible, this assumption has been challenged by the report of one patient in whom hepatitis recurred after exposure to mesalamine (mesalazine, 5-aminosalicylate).126 This finding implicates the salicylate moiety, and like salicylate hepatitis (discussed earlier) sulfasalazine hepatotoxicity appears to be more common in patients with rheumatoid arthritis than in those with inflammatory bowel disease. A case of mesalamine-induced chronic hepatitis with autoimmune features was diagnosed after 21 months of treatment with the drug.127 Granulomatous hepatitis has been reported with mesalamine.128

Chapter 86  Liver Disease Caused by Drugs Minocycline and other tetracyclines used in conventional low doses are a rare but important cause of drug hepatitis,129,130 including cases that have resulted in acute liver failure requiring liver transplantation.131 Minocycline is one of the few agents in current use that can lead to druginduced autoimmune hepatitis, as discussed later. Disulfiram (Antabuse) rarely has been associated with acute hepatitis, occasionally leading to liver failure.132 Disulfiram hepatitis usually is easy to distinguish from alcoholic hepatitis by the ten-fold or greater elevation of serum ALT activity. b-adrenergic blocking agents rarely have been incriminated in hepatotoxicity. Acebutolol,133 carvedilol,134 labetalol,135 and metoprolol136 each have been associated with cases of acute hepatitis; some cases were proved by rechallenge. Reactions were hepatocellular and severe. Data are insufficient to determine whether or not immunoallergy is likely. The calcium channel blockers, nifedipine,137 verapamil,138 diltiazem,139 and amlodipine140 have good safety records, but rare cases of acute hepatitis with a short incubation period (five days to six weeks) and other features of immunoallergy have been reported. Of the angiotensin II receptor blockers, irbesartan has been linked to two reports of cholestasis.141,142 In both patients, jaundice developed within one month of the start of therapy. Liver biochemical testing showed predominant cholestasis, and findings on ultrasonography were normal. Histologic examination revealed marked cholestatic features in both patients, with an inflammatory infiltrate and eosinophils in one patient. Clinical resolution occurred when the medication was stopped; however, liver biochemical abnormalities persisted for more than one year in one patient. Biliary ductopenia is a rare complication. Other angiotensin II receptor blockers implicated in cases of acute hepatitis or cholestatic hepatitis include losartan, valsartan, and candesartan.143-145 Angiotensin-converting enzyme inhibitor-induced liver disease is a rare but important complication of this widely prescribed class of drugs. The incidence has been estimated to be 9 per 100,000 patients treated.146 Reactions to captopril (the oldest and possibly most hepatotoxic representative) and enalapril usually manifest as cholestatic hepatitis, but hepatocellular or mixed hepatocellular reactions can occur.147-149 Features of hypersensitivity, such as fever, skin rashes, and eosinophilia have been observed in patients with captopril hepatotoxicity.147 Histologic examination reveals marked centrilobular cholestasis with eosinophilic portal infiltrates.147 Liver biochemical abnormalities usually resolve after withdrawal of the drug, but resolution may take up to six months in some cases. Fulminant hepatic failure has been attributed to lisinopril,150 whereas fosinopril has been associated with bland cholestasis.151 Ramipril has been implicated in three cases of cholestatic liver injury, one of which progressed to biliary cirrhosis.152 Biliary ductopenia also has been observed with enalapril.153 Hydroxymethylglutaryl-coenzyme A reductase inhibitors (“statins”), as a class of drugs, are not strongly associated with important hepatic injury, although literature reports and data contributed to drug safety surveillance authorities appear to be discordant. Use of statins has increased greatly as new guidelines have lowered the target level for lowdensity lipoprotein (LDL) cholesterol, thereby resulting in use of higher doses of statins. A dose-related rise in serum aminotransferase levels develops in 1% to 3% of people who take statins.154 A minor (less than two-fold normal) rise in the serum ALT and AST levels (without symptoms) is the most common manifestation of liver injury with these

compounds. These elevations usually reverse rapidly with discontinuation of the statin and also reverse if therapy is not interrupted. Lovastatin,155 pravastatin,156 atorvastatin,157 simvastatin,158 and rosuvastatin159 have been implicated in a few reports of acute hepatitis or cholestatic hepatitis. Use of some members of this group of lipid-lowering agents in combination with gemfibrozil is associated with an increased rate of myositis but apparently not with an increased rate of drug-induced liver injury. Prescribing guidelines for statins invariably warn about the risk of liver injury when these agents are prescribed to persons with preexisting liver abnormalities. Nevertheless, evidence that fatty liver (or steatohepatitis), hepatitis C, or other common liver disorders predispose to drug-induced liver disease is lacking. Moreover, a controlled trial of highdose pravastatin in patients with preexisting liver disease has demonstrated the safety of statins in this setting.160 Similar conclusions were reached in the Dallas Heart Study, in which statin users were no more likely to exhibit serum ALT elevations than non-users.161 Likewise, statin hepatotoxicity has been unrelated to the serum ALT level at baseline in other retrospective studies162,163 and has even been shown to reduce overall progression of hepatic fibrosis.163 Monitoring serum amionotransferase levels is recommended, but this approach is unlikely to predict toxicity164 or to be cost-effective.165 In the Air Force Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TEXCAPS) increases of more than three times the upper limit of normal were observed in only 18 of 100,000 aminotransferase determinations in users of lovastatin, and in no case did hepatitis develop.165 Etretinate, a synthetic retinoid, is useful for treating several skin diseases. Unlike vitamin A (see Chapter 87), synthetic retinoids are not predictable hepatotoxins, but etretinate has been associated with elevated liver biochemical test levels in 10% to 25% of treated patients.166 Levels may normalize with a reduction in drug dose, thereby suggesting partial dose dependency. Approximately 10 cases of severe hepatitis have been attributed to etretinate, and some have been proved by rechallenge.166 Most patients were women older than age 50; two cases were associated with chronicity, and one patient appeared to respond to glucocorticoids. Because etretinate has a half-life of 100 days, monitoring serum ALT levels is recommended in users of the drug. A progressive increase in the serum ALT level to values above twice the upper limit of normal are an indication to stop the etretinate or to perform a liver biopsy.166 Acitretin is another synthetic retinoid that has been associated with a few instances of acute hepatitis, including cases associated with bile duct injury and progressive hepatic fibrosis.167,168 Gastric acid suppression drugs have an excellent safety record, although rare adverse hepatic reactions have been reported. The histamine H2 receptor antagonist oxmetidine was removed from clinical trials because of hepatotoxicity, and subsequently, ebrotidine was withdrawn because of many cases of liver injury.169 Cimetidine,170 ranitidine,171 and famotidine172 have been associated with cases of acute hepatitis, mostly mild and often with cholestatic features. Some cases have been proved by rechallenge. Features of immunoallergy were present in some of the cimetidine reactions. Cases of hepatotoxicity have been attributed to the proton pump inhibitors omeprazole, lansoprazole, and pantoprazole; these reports have been isolated and causality was not established unequivocally in some of the cases.173,174 Zafirlukast, a leukotriene receptor antagonist effective against asthma, has been reported to cause severe liver

1433

1434

Section IX  Liver injury, with several instances of acute liver failure.175,176 Montelukast has been implicated in three cases of acute hepatitis or cholestatic hepatitis.177 Ticlopidine, an antiplatelet agent, has been associated with more than 30 reports of hepatotoxicity. Examination of liver histology has shown bland cholestasis in most cases and occasionally microvesicular steatosis; cholestatic hepatitis with bile duct injury also has been reported.178 An association with a specific HLA haplotype (A*3303) was observed in Japanese patients with ticlopidine hepatotoxicity.179 Currently, clopidogrel is preferred to ticlopidine, but clopidogrel also can cause hepatocellular or mixed hepatocellular-cholestatic liver injury.180

METABOLIC IDIOSYNCRASY Isoniazid

Isoniazid-induced liver injury has been characterized since the 1970s, but deaths still occur.181-184 Hepatitis develops in about 21 per 1000 persons exposed to isoniazid; 5% to 10% of cases are fatal. The risk and severity of isoniazid hepatitis increase with age; the risk is 0.3% in the third decade of life and increases to 2% or higher after age 50.181,182 Overall frequency rates are the same in men and women, but 70% of fatal cases are in women; black and Hispanic women may be at particular risk.181,182 The risk of toxicity is not related to the dose or blood level of isoniazid. The role of genetic factors has been controversial. Associations have been described with specific genes that code for enzymes involved in aspects of drug metabolism or detoxification (CYP2E1, N-acetyltransferase, glutathione-S-transferase), but data are conflicting.185-187 Chronic excessive alcohol intake increases the frequency and severity of isoniazid hepatotoxicity,181,182 as may rifampin and pyrazinamide.188 Concomitant use of pyrazinamide or acetaminophen have been associated with several cases that were fatal or led to liver transplantation.189 Some studies have found that the risk of liver injury from isoniazid and other anti-tuberculosis drugs is increased among persons with chronic HBV infection, but reports are conflicting.189 Malnutrition may play a role in isoniazid hepatotoxicity in some countries. Likewise, in patients with HCV or HIV infection (or both), the risk of significant serum ALT elevations during anti-tuberculosis treatment has been reported to be increased several-fold; successful antiviral treatment of hepatitis C allowed the safe reintroduction of anti-tuberculosis drugs in four patients. Serum ALT levels increase in 10% to 36% of persons taking isoniazid in the first 10 weeks. The elevations typically are minor and resolve spontaneously. In persons in whom hepatitis develops, the latent period from exposure to disease ranges from 1 week to more than 6 months, with a median of approximately 8 weeks and, in severe cases, 12 weeks.181,182 Re-exposure to isoniazid may be associated with an accelerated onset, although the experience in India is that gradual reintroduction of isoniazid and rifampin therapy can be achieved in a majority of cases after drug hepatitis has resolved. Prodromal symptoms occur in one third of patients and include malaise, fatigue, and early symptoms of hepatitis, such as anorexia, nausea, and vomiting. Jaundice appears several days later and is the only feature in approximately 10% of cases. Fever, rash, arthralgias, and eosinophilia are uncommon. Liver biochemical testing indicates hepatocellular injury; serum AST levels exceed serum ALT levels in one half of patients. Serum bilirubin levels usually are elevated; values that are increased more than ten-fold indicate a poor prognosis. In one study,181 one third of patients had a prolonged prothrombin time, and 60% of these cases were fatal. Liver

biopsy samples generally show hepatocellular injury, which is focal in approximately 50% of cases, often with marked hydropic change in residual hepatocytes. In the remaining cases, hepatocellular necrosis is zonal, submassive, or massive, with inflammation confined to the portal tracts. Cholestasis and lobular regeneration suggestive of early cirrhosis is a rare feature. Cases with a fatal outcome have been associated with a longer duration of treatment with isoniazid or continued ingestion of isoniazid after the onset of symptoms.181,182 Therefore, most deaths from isoniazid hepatitis could be prevented if patients report symptoms early in the course and isoniazid is discontinued. In the United States, isoniazid hepatotoxicity is second only to acetaminophen as an indication for liver transplantation for drug-induced liver injury.189 Children are less susceptible than adults, but serious hepatotoxicity can occur in children; over a 10-year period (1987-1997), eight children required liver transplantation for isoniazid hepatotoxicity in the United States (0.2% of pediatric liver transplants).190 Recovery is rapid if isoniazid is discontinued before severe liver injury is established. Management of liver failure is supportive (see Chapter 93); liver transplantation is indicated in the most severe cases. Prevention is the most appropriate way to prevent isoniazid hepatotoxicity, and determining whether the risks of isoniazid preventive therapy outweigh those of latent tuberculosis is critical. The optimal approach to monitoring a patient for isoniazid toxicity is uncertain; every-other-week or monthly monitoring of serum ALT levels will not always prevent the rapid onset of severe hepatotoxicity. Effective prevention depends on awareness of early symptoms, no matter how nonspecific.

Other Drugs

Other Antituberculosis Drugs Most cases in which rifampin has been implicated in liver injury have occurred in patients who also are taking isoniazid,191 but a few cases have been observed when rifampin was given alone to patients with underlying liver disease.192 Pyrazinamide (and the related ethionamide) has long been known to be a dose-dependent hepatotoxin. The drug is now used in lower doses (1.5 to 2 g/day) because of the emergence of resistant strains of mycobacteria. Hepatotoxicity in patients who are taking combinations that include isoniazid and pyrazinamide may be particularly severe.183 Monitoring of serum ALT levels during therapy is recommended. Cross-sensitivity among isoniazid, pyrazinamide, and ethionamide may occur. Treatment of latent tuberculosis with the combination of pyrazinamide and rifampin, levofloxacin, or ethambutol has been associated with an increased risk of hepatic injury.193-195 Antifungal Agents Ketoconazole is associated with raised serum levels of aminotransferases in 5% to 17% of treated patients.196,197 Symptomatic hepatitis occurs in 7 to 20 of 100,000 exposed persons. Women (with a female-to-male ratio of 2 : 1) and persons older than 40 years of age are particularly susceptible to ketoconazole-induced liver injury.196-198 Concurrent use of drugs (e.g., lovastatin) that share similar metabolic pathways of elimination (CYP3A4) with ketoconazole can lead to hepatotoxicity.199 Reactions are usually mild but can be severe, with rare cases of acute liver failure.200 The mortality rate is 3% to 7%.196,197 The onset of toxicity is 6 to 12 weeks after ketoconazole is started, and rarely after the drug is stopped. Toxicity is unrelated to the dose of the drug.

Chapter 86  Liver Disease Caused by Drugs Continued ingestion of ketoconazole after the onset of symptoms leads to an adverse outcome. Jaundice occurs in 50% of patients in whom acute hepatitis develops, and up to one third may present with nonspecific symptoms, such as nausea, anorexia, and vomiting. Fever, rash, eosinophilia, and other immunoallergic characteristics are rare. Liver biochemical test levels are primarily hepatocellular or mixed, but cholestatic hepatitis or bland cholestasis may occur.196 Jaundice usually resolves within 12 weeks, but resolution may take months.196,197 Cirrhosis is a rare complication following acute hepatic injury.201 The role of glucocorticoid treatment in cases that are slow to resolve is unclear. Fulminant hepatic failure requiring liver transplantation has been reported.202 Fatal liver failure also has followed the use of ketoconazole to reduce hypercortisolism in patients with Cushing’s syndrome.203 Terbinafine is an allylamine antifungal agent that is effective in the treatment of onychomycosis. Several cases of cholestatic hepatitis attributed to terbinafine have been reported.204 The frequency of hepatotoxicity associated with this drug has been estimated to be two to three cases per 100,000 persons exposed.204 The onset is usually within four to six weeks after the drug is started. Liver biopsy specimens show hepatocyte degeneration and canalicular cholestasis with variable portal tract inflammation. Recovery is usual with discontinuation of the drug, although prolonged cholestasis with ductopenia has been reported.205 Ursodeoxycholic acid has been prescribed to affected patients to hasten recovery when cholestasis is protracted.205,206 Fulminant hepatic failure also has been described,207 and a case of sinusoidal obstruction syndrome has been associated with use of terbinafine in a liver transplant recipient.208 The FDA has received at least 16 reports of fulminant liver failure possibly linked to terbinafine209; the frequency of this outcome has been estimated to be 1 per million persons exposed.210 Fluconazole and itraconazole appear to be less hepatotoxic than ketoconazole and terbinafine211; elevations of liver biochemical test levels occur in fewer than 5% of patients. Rare cases of severe hepatic necrosis have been ascribed to fluconazole, but other causes were not excluded. Instances of acute liver failure associated with itraconazole have been reported.212-214 Among more than 69,000 patients who received an oral antifungal agent, ketoconazole and itraconazole were most often associated with liver injury; the relative risks were 228 and 17.7, respectively, in comparison with non-users.215 Antidiabetic Drugs Thiazolidinediones.  Troglitazone was the first peroxisome proliferator-activated receptor-γ (PPARγ) agonist used in patients with type 2 diabetes mellitus for improving glycemic control and lowering serum lipid levels by reducing insulin resistance. Elevations of serum aminotransferase levels were noted in 0.5% to 1.9% of recipients in early trials, which failed to reveal serious hepatotoxicity.216 Reports of acute liver failure emerged in the post-marketing phase,217 in which troglitazone was associated with more than 75 instances of fatal hepatotoxicity or liver failure requiring hepatic transplantation.218 Reported cases of troglitazone hepatotoxicity generally were in older women and obese persons—the common phenotype of persons with type 2 diabetes mellitus. Detailed epidemiologic studies have not been performed to define the risk factors clearly. Evidence that preexisting liver disease or other drugs predispose to troglitazone hepatotoxicity is lacking, although a progressive course in one patient was attributed to concurrent use of simvas-

tatin and troglitazone.219 Mitochondrial injury is favored as the mechanism of hepatic injury, but other mechanisms (e.g., reactive metabolites, inhibition of BSEP) have been proposed.220-222 The onset of troglitazone hepatotoxicity was often as late as 9 to 12 months after treatment was started223,224; rare cases had a much earlier onset (8 days).225 Presenting symptoms included nausea, fatigue, jaundice, vomiting, and symptoms of liver failure. Progression to acute liver failure was often rapid, and in some cases, deterioration continued despite discontinuation of troglitazone.226 Histologic examination of liver biopsy specimens, explanted livers, or autopsy material showed submassive or massive hepatic necrosis, with post-collapse scarring, bile duct proliferation, and some eosinophils.227 Severe cholestasis also was reported,228 as is sometimes observed with other causes of fulminant hepatic failure (e.g., valproic acid) and does not necessarily imply a pathogenic mechanism different from that in cases not associated with cholestasis. Troglitazone was withdrawn from the market in 1999. Serious liver injury appears to be rare with the secondgeneration thiazolidinediones rosiglitazone and pioglitazone. In clinical trials, a raised serum ALT level (to greater than three times the upper limit of normal) was reported in 0.25% of patients treated with rosiglitazone and 0.26% treated with pioglitazone.218 Six reports of hepatotoxicity associated with rosiglitazone have been published.229 In some cases, the onset was earlier (e.g., 8 days, 21 days) than in most reported cases of troglitazone hepatotoxicity. On the other hand, the presentation of hepatic injury can be delayed beyond one year.230 An alternative diagnosis of ischemic hepatitis (or ischemia-reperfusion injury) was proposed for one of the cases, and confounding factors were present in two other cases.230,231 Examination of liver histologic findings showed cholestatic hepatitis or granulomatous hepatitis.232 One death has been reported, but all other patients have recovered. Pioglitazone has been implicated in five reports of acute hepatocellular injury.229 Cholestatic hepatitis with bile duct injury was observed in two cases.233 Most patients have recovered after discontinuing pioglitazone. Acute liver failure is a rare complication.234 An isolated reversible increase in the serum alkaline phosphatase level has been described in a patient taking pioglitazone.235 Before treatment with drugs of this class is begun, the FDA recommends that liver biochemical test levels be measured; the pretreatment serum ALT level should be less than 2.5 times the upper limit of normal. Monitoring the serum ALT level every two months during the first year of therapy and periodically thereafter is advised. If ALT levels remain persistently elevated (greater than three times the upper limit of normal), the thiazolidinedione should be discontinued. Symptoms suggestive of hepatitis should be assessed immediately. Persons in whom jaundice developed with troglitazone should not take other thiazolidinediones.236 Other Oral Hypoglycemic Drugs.  Hepatocellular liver injury was common with older sulfonylureas, such as carbutamide, metahexamide, and chlorpropamide.237 Tolbutamide, tolazamide, glimepiride, and glibenclamide, which are currently used agents, rarely have been associated with cholestasis or cholestatic hepatitis.238-242 Hypersensitivity phenomena (fever, skin rash, eosinophilia) were present in some cases, as would be expected in view of the structural relationship between sulfonylureas and sulfonamides. Most cases resolved after withdrawal of the drug; however, chronic cholestasis progressing to vanishing bile duct syndrome (see Chapter 20) has been described with tolbutamide and tolazamide.241,242 Fatal liver failure has been

1435

1436

Section IX  Liver reported in at least two patients, one of whom had underlying cirrhosis.243 Gliclazide15,244 and glibenclamide also have been associated with hepatocellular liver injury and, with the latter drug, hepatic granulomas.239 Metformin, acarbose, repaglinide, and human insulin rarely have been associated with liver injury.245-248

Drugs Used for Neurologic Disorders

Several neuroleptic agents have been associated with drug hepatitis. Some reactions appear to be immunoallergic, whereas others conform to the pattern of apparent metabolic idiosyncrasy, depending on the structure of the drug. Such reactions have been reported for commonly used antidepressants, such as fluoxetine,249,250 aroxetine,251 venlafaxine,252 trazodone,253 tolcapone,254 and nefazodone255; the last two drugs also have been implicated in several cases of acute liver failure. Antidepressants Monoamine Oxidase Inhibitors.  Iproniazid was one of the first drugs associated with acute hepatitis. Reactions occurred in 1% of recipients and were often severe, with instances of fatal fulminant liver failure. The hydrazine substituent (which iproniazid shares in part with isoniazid, ethionamide, pyrazinamide, and niacin) was determined to be hepatotoxic moiety.256 Phenelzine and isocarboxazid also have been associated with occasional instances of hepatocellular injury, but monoamine oxidase inhibitors are now prescribed infrequently.257 Tricyclic Antidepressants.  Tricyclic antidepressants bear a structural resemblance to the phenothiazines and are occasional causes of cholestatic or, less commonly, hepatocellular injury. Recovery following cessation of the drug is usual, but prolonged cholestasis has been observed with amitriptyline258 and imipramine.259 Selective Serotonin Reuptake Inhibitors and Other Modern Antidepressants.  Selective serotonin reuptake inhibitors (SSRIs) have a better overall safety profile than tricyclic antidepressants. Liver enzyme elevations have been observed in asymptomatic persons taking fluoxetine and paroxetine.249 A few reports of acute and chronic hepatitis have been attributed to the use of SSRIs.249,250 Nefazodone has been associated with cases of subacute liver failure.255,260 Centrilobular, massive, or submassive hepatic necrosis was observed on histologic examination of the livers of affected persons. Nefazodone has been withdrawn from the market. Trazodone has been implicated in cases of acute and chronic hepatocellular injury.253,261 The onset can be delayed as long as 18 months or can occur within 5 days of the start of the drug.262 Occasional reports have noted the occurrence of severe hepatotoxicity with combinations of antidepressants or with antidepressants used in combination with other neuroleptic agents.263 Drug regulatory authorities have been alerted about hepatic adverse events with atomoxetine, a norepinephrine reuptake inhibitor used in treating attention deficit hyperactivity disorder. Only three of the reported cases have been linked conclusively with the drug; all showed a pattern of acute hepatocellular injury.264 Other Neurologic Drugs Tolcapone, a catechol-o-methyl transferase inhibitor used in the treatment of Parkinson’s disease, has been associated with at least four cases of acute liver failure.265,266 All reported cases occurred in women older than 70 years of age, who presented with jaundice and high serum ALT

levels. Centrilobular necrosis was observed on liver histologic examination at autopsy in one case.265 Serious liver injury has not been reported in users of the drug who adhered to monitoring guidelines.266 Postmarketing surveillance has identified three additional patients with acute hepatocellular injury caused by tolcapone. The general consensus, however, is that tolcapone is safe if users are monitored appropriately.267 Current FDA guidelines recommend serum ALT testing every two to four weeks for the first six months. Thereafter, the frequency of testing is left to the discretion of the treating doctor. Patients in whom signs of hepatic impairment or a rise in the serum ALT level (at least one to two times the upper limit of normal) develops should be monitored closely; persistent serum ALT elevations (more than two times the upper limit of normal) are an indication to discontinue the drug.268 Alpidem,269 zolpidem,270 and bentazepam271 are sedative hypnotics that have been implicated in hepatotoxicity. In three reported cases of bentazepam hepatotoxicity, the clinicopathologic pattern resembled chronic hepatitis, but without autoantibodies or other immunologic features.271 Tacrine is a reversible choline esterase inhibitor that improves cognition in patients with Alzheimer’s disease. In a survey of tacrine-related adverse effects in 2446 patients with Alzheimer’s disease, serum ALT levels more than three times the upper limit of normal occurred in 25% of patients and more often in women than in men; levels were elevated more than 20-fold in 2% of patients.272 No dose effect was observed. Serum ALT levels rose abruptly, rather than gradually, and the elevations resolved after discontinuation of tacrine. Symptoms were rare; only nausea and vomiting correlated with major serum ALT elevations. In liver biopsy specimens from three patients, steatosis and mild lobular hepatitis were observed. According to this study,272 minor degrees of hepatocellular injury occur in one half of users of tacrine, but tolerance to this minor form of liver injury eventually develops. Isolated reports of patients in whom jaundice developed indicate a rare potential for tacrine to cause more severe hepatotoxicity. Weekly monitoring of serum ALT levels during the first three months of therapy with tacrine and discontinuation of the drug if values reach three times the upper limit of normal should prevent important hepatotoxicity.272 Although the mechanism of tacrine-induced hepatotoxicity is unclear, a metabolic idiosyncrasy seems likely. Genetic factors that may underlie individual susceptibility to tacrine toxicity have been suggested (e.g., glutathione S-transferase polymorphisms),273 but examination of a panel of 19 candidate genes failed to reveal relevant mutations.274 Mitochondrial injury has been implicated in an animal model of tacrine hepatotoxicity.275 Dantrolene, a skeletal muscle relaxant used to treat spasticity, causes hepatitis in about 1% of exposed persons, with a case-fatality rate of approximately 28%.276 Most affected patients have been older than 30 years of age. One third of patients are asymptomatic, and the remainder present with jaundice and symptoms of hepatitis. Hepatocellular necrosis, often submassive or massive, has been noted on liver biopsy specimens.276 When therapy with dantrolene is initiated, liver biochemical tests should be monitored every two weeks. Liver enzyme elevations are an indication to stop dantrolene. Other neurotropic drugs and muscle relaxants implicated as idiosyncratic hepatotoxins include tizanidine (a centrally acting muscle relaxant),277 alverine (a smooth muscle relaxant),278 and riluzole.279 Patients with cirrhosis who take tizanidine are at risk of developing hypotension; levels of this CYP1A2-metabolized drug are increased in persons with

Chapter 86  Liver Disease Caused by Drugs cirrhosis as a consequence of diminished cytochrome activity.280 Riluzole is a glutamate antagonist approved for the treatment of amyotrophic lateral sclerosis. Increased serum ALT levels were reported in 1.3% to 10% of users of the drug in clinical trials. Two cases of acute hepatitis with microvesicular steatosis have since been reported, with onset at four and eight weeks, respectively, after the start of treatment.279,281 Rarely, hepatocellular injury may be delayed for as long as six months. Liver biochemical test abnormalities resolve rapidly after riluzole is discontinued. Nonsteroidal Anti-Inflammatory Drugs NSAIDs rarely cause drug-induced liver disease, with or without immunoallergic features and with varying degrees of hepatocellular injury and cholestasis. Bromfenac, a phenylacetic acid derivative, was withdrawn from the U.S. market in 1998 because of several cases of severe hepatotoxic reactions that resulted in acute liver failure leading to liver transplantation or death.282 Most affected patients had received therapeutic doses of brom­ fenac for more than 90 days before experiencing a prodrome of malaise and fatigue, followed by symptoms of severe hepatitis and progressive liver failure for 5 to 37 days; use of the drug had been recommended for only up to 7 days. No features suggestive of immunoallergy were evident. Histologic examination of the liver showed zonal or confluent necrosis with a predominantly lymphocytic infiltrate. Most COX-2 inhibitors appear to be relatively free of hepatoxicity, although a small number of cases have been reported in association with nimesulide and celecoxib.283 An exception is lumiracoxib, which was associated with severe hepatotoxicity and was withdrawn from use. Celecoxib has a low potential for liver injury. In a review of 14 controlled trials, the frequency of hepatic dysfunction (0.8%) was not significantly different from that in placebotreated patients (0.9%) and appeared to be lower than that observed with other NSAIDs.284 Elevations in serum aminotransferase levels often occurred in persons taking diclofenac concurrently. When serious hepatocellular injury was attributed to celecoxib, female gender was a predisposing factor.285 The onset of symptoms was between four days and four weeks after the drug was started. Liver biochemical

abnormalities were consistent with a pattern of hepatocellular or mixed liver injury. Eosinophilia and skin rash suggestive of RMS occurred in some patients. All patients recovered within one to four months of discontinuation of the drug. Acute liver failure has been a rare complication of celecoxib toxicity. The manufacturer currently recommends that celecoxib not be administered to persons with a documented sulfonamide allergy because of published reports of cross-reactivity and toxicity. Nimesulide is an NSAID that has COX-2 selectivity. It has been linked to several cases of acute hepatitis and fatal hepatic failure, especially in women,286 although the risk of liver injury is very small.287 The time to the onset of symptoms has ranged from 1 to 15 weeks, although a delay of up to 8 months is possible.288,289 Hypersensitivity features with peripheral eosinophilia may occur. Centrilobular or bridging necrosis and occasionally bland cholestasis have been described on liver histologic examination. Resolution usually occurs 2 to 17 months after nimesulide is discontinued.289

DRUG-INDUCED GRANULOMATOUS HEPATITIS Granulomatous reactions are a common type of druginduced liver disease,290-294 and drugs account for 2% to 29% of cases of granulomatous hepatitis (see Chapter 35). The number of drugs and foreign compounds associated with hepatic granulomas exceeds 40; some of these compounds are summarized in Table 86-5. Not all the agents reported to cause granulomatous hepatitis are associated with systemic inflammation or with persuasive evidence of causality. Many (e.g., halothane, methyldopa, nitrofurantoin, troglitazone, amiodarone, amoxicillin-clavulanic acid) are more commonly associated with other patterns of liver injury. Some of these associations may be fortuitous. The clinical picture is heralded by fever and systemic symptoms (e.g., malaise, headache, myalgias) from 10 days to 4 months after the start of treatment. Hepatomegaly and hepatic tenderness are common; splenomegaly is present in 25% of patients. Extrahepatic features of drug hypersensi-

Table 86-5  Drug-Induced Granulomatous Hepatitis: Major Causative Agents, Frequency, Risk Factors, Clinicopathologic Characteristics, and Outcomes CAUSATIVE AGENT*

FREQUENCY

RISK FACTORS

Allopurinol

Rare (<40 cases)

Carbamazepine Phenylbutazone

16:100,000 treatment years 1:5000 exposed

Older men, black race, renal failure, use of thiazides Age >40 yr, no gender predilection

Hydralazine

Rare

Older patients, possibly slow acetylators

Quinine

Rare

No recognized risk factors

No age or gender predilection

CLINICOPATHOLOGIC CHARACTERISTICS Acute hepatitis, cholestatic hepatitis, bile duct injury also frequent; rash (exfoliative dermatitis), nephritis, vasculitis usual Two thirds of cases show granulomatous hepatitis; remainder show acute hepatitis, cholangitis; no features of drug allergy Severe acute hepatitis, cholestasis and bile duct injury also reported; features of drug allergy common; occasionally vasculitis Other types of reaction also common: acute hepatitis, cholestatic hepatitis, cholangitis; features of drug allergy uncommon; vasculitis not described Acute hepatitis in two thirds of cases; rash, interstitial pneumonitis, positive Coombs test, thrombocytopenia; vasculitis not described

OUTCOMEs Mortality rate 15%, especially with vasculitis No reported fatality, rapid recovery Mortality rate 25%, particularly in cases with hepatocellular necrosis Reactions severe but no mortality reported Good prognosis

*Other drugs that have been reliably reported to cause granulomatous hepatitis include quinidine, phenytoin (usually with vasculitis), sulfonamides (usually with vasculitis), nitrofurantoin, aspirin, papaverine, procainamide, sulfasalazine, mesalamine, and glyburide. Single case reports have implicated many other agents, as referred to briefly in the text.

1437

1438

Section IX  Liver tivity are common, as is eosinophilia (30%). Liver biochemical test levels are typically mixed because of the infiltrative nature of hepatic granulomas and the frequent presence of some hepatocellular necrosis or cholestasis. For several drugs that cause granulomatous hepatitis, continued exposure leads to more severe types of liver disease, such as cholestatic hepatitis with or without bile duct injury and hepatic necrosis (see Table 86-5). Small-vessel vasculitis is another potential complication (see Table 86-5) and may involve the kidneys, bone marrow, skin, and lungs; the mortality rate is high.

DRUG-INDUCED CHRONIC HEPATITIS Chronic hepatitis is defined as hepatitis that continues for more than six months. For drug reactions, however, the definition often has been made inappropriately on hepatic histologic features alone. The histologic features include periportal inflammation (piecemeal necrosis), bridging necrosis, and fibrosis. Because these features may be present as early as six weeks after the onset of severe reactions, they do not confirm chronicity. The diagnosis of chronic hepatitis is more convincing when clinical or biochemical evidence of hepatitis has been present for more than three months and when clinical and laboratory features of chronic liver disease or histologic evidence of established hepatic fibrosis are present. Drugs are an uncommon cause of chronic hepatitis (Table 86-6). Implicated agents such as oxyphenisatin5 and methyldopa (see earlier) are now used rarely, and with the increasing importance of chronic viral hepatitis (see Chapters 78 and 79), only a small proportion of cases of chronic hepatitis (2% to 6%, depending on geographic reporting region) are caused by drugs. In fact, several cases of pre-

sumed drug-induced chronic hepatitis reported in the past were subsequently found to be cases of hepatitis C.68 Nevertheless, recognition of a drug cause remains important because a poor outcome is usually attributable to continued ingestion of the drug after the clinical onset of the disorder. Drug-induced chronic hepatitis is approximately four times as likely to develop in women as in men. Older patients appear to be at greater risk (as in the case of nitrofurantoin), and the reaction is virtually unknown in children. Drugs associated with chronic hepatitis more commonly cause acute hepatitis, and the latent period to recognition tends to be longer in cases of chronic hepatitis; therefore, the duration of drug ingestion may be a risk factor for chronic hepatitis. Two syndromes of drug-induced chronic hepatitis occur. In the first, cases appear to be identical to acute hepatitis but more severe, more prolonged, or later in onset, perhaps as a result of failure of recognition. These cases may appropriately be termed chronic toxicity. Clinical and laboratory features of chronic liver disease are rare, and hallmarks of autoimmunity are absent. Management consists of withdrawal of the drug and treatment of liver failure (see Table 86-6). The second syndrome more closely resembles autoimmune hepatitis based on the presence of spider angiomata, firm liver edge, splenomegaly, and potential for liver failure. Ascites, bruising, bleeding esophageal varices, and hepatic encephalopathy are common. In addition to raised serum ALT and bilirubin levels, hypoalbuminemia and hyperglobulinemia are usual. The prothrombin time is prolonged in severe cases. Autoantibodies, particularly antinuclear and smooth muscle antibodies, are frequent. In contrast to idiopathic autoimmune hepatitis, other hallmarks of autoimmunity, such as a history of other autoimmune diseases and genetic predisposition indicated by HLA-B8 and -DRw3

Table 86-6  Drug-Induced Chronic Hepatitis: Causative Agents, Risk Factors, Clinicopathologic Characteristics, and Outcomes CAUSATIVE AGENT*

RISK FACTORS

CLINICOPATHOLOGIC CHARACTERISTICS

OUTCOMEs

Nitrofurantoin

Age >40 yr; 90% of cases in women; continued ingestion after onset

Mortality rate 10%

Methyldopa

Age >50 yr; 80% of cases in women; repeated courses, continued ingestion in sensitized patient Age >65 yr; most cases in women

Clinical features of chronic hepatitis, liver failure; some cases with features of cholestasis; 20% with pneumonitis; hyperglobulinemia usual, ANA, SMA Jaundice, diarrhea, liver failure; hyperglobulinemia, ANA, SMA positive; protracted course Clinical features of chronic hepatitis, liver failure; ANA, SMA, hyperglobulinemia Often part of drug-induced systemic lupus erythematosus syndrome (arthritis, rash, rarely nephritis); ANA, hyperglobulinemia

Diclofenac Minocycline

Young women; prolonged use of drug

Isoniazid

Age >50 yr; continued drug ingestion after onset; duration of therapy Age >30 yr; dose, duration of therapy Age >50 yr; two thirds in women

Dantrolene Etretinate Acetaminophen

Regular intake at moderate doses (2 to 6 g/day); alcohol, fasting, other drugs

High mortality rate

Severe and fatal cases with cirrhosis; no immune phenomena

Response to glucocorticoids in a few cases Cases may be severe, with fatal outcome or need for liver transplantation; glucocorticoid treatment may be indicated High mortality rate or need for liver transplantation

Jaundice, liver failure; no immune phenomena

High mortality rate

Jaundice, weight loss, liver failure; deterioration after stopping drug No features of chronic liver disease, no autoimmune phenomena; there are cases of chronic toxicity

Response to glucocorticoids in two reported cases Rapid normalization of liver biochemical test levels after drug is stopped

*Other drugs include oxyphenisatin and tienilic acid, which are now of historical interest,5,6 and clometacin, for which many presumably affected patients have now been shown to have had hepatitis C.100 Several other agents, including sulfonamides, aspirin, halothane, cimetidine, methotrexate, trazadone, fluoxetine, fenfibrate, and germander, have been mentioned as associated with chronic hepatitis, but details are not always convincing.5 ANA, antinuclear antibodies; SMA, smooth muscle antibodies.

Chapter 86  Liver Disease Caused by Drugs alleles, are not found. Treatment with immunosuppressants usually is not indicated; the clinical condition improves spontaneously after withdrawal of the causative drug. In individual cases, however, glucocor­ticoids occasionally appear to hasten recovery (see later).

DICLOFENAC

Diclofenac is one of the world’s most prescribed NSAIDs and appears to be at least as safe as comparable agents. Significant hepatotoxicity occurs in about 1 to 5 per 100,000 persons exposed, or 0.4 per 1 million defined daily doses; the latter rate is minimally greater than that for phenylbutazone (0.2 per 1 million) and piroxicam (0.3 per 1 million) but less than that for benoxaprofen (12.6 per 1 million) and bromfenac, which was withdrawn from the market (see earlier). More than 200 cases of diclofenac hepatitis have been reported,295 including several proved by indavertent rechallenge. Only four cases have been fatalities, and five cases can reasonably be regarded as chronic hepatitis. Genetic susceptibility to diclofenac hepatotoxicity has been documented.296,297 In these cases, polymorphisms have been observed within genes that affect metabolic pathways that lead to formation of reactive metabolites of the drug; biliary excretion and the immune response to drug metaboliteprotein adducts have been identified.296,297 The risk of diclofenac hepatitis is increased in women and with aging. A prodromal illness characterized by anorexia, nausea, vomiting, and malaise heralds the onset of liver injury, which usually occurs within 3 months (range, 1 to 11 months) of the start of treatment. Fever and rash occur in 25% of patients.295 Liver biochemical test results reflect acute hepatitis with or without cholestasis. Reactions tend to be severe, with jaundice occurring in 50% of cases. Liver biopsy specimens reveal acute lobular hepatitis, and in severe cases, bridging or confluent necrosis, interface hepatitis, and fibrous expansion of the portal tracts had been noted.298 The prognosis is usually good; resolution occurs after discontinuation of the drug. Cases of druginduced chronic hepatitis have been described in which the clinical and laboratory features (ascites, hypoalbuminemia, hyperglobulinemia, jaundice) suggested autoimmune hepatitis, although the frequency of autoantibodies is unclear. These cases usually improve spontaneously after discontinuation of the drug, but glucocorticoids have been used successfully in a few protracted cases.299 Cross-sensitivity with other NSAIDs seems to be rare, but one patient with diclofenac hepatitis also had an adverse reaction to ibuprofen and another had an adverse reaction to tiaprofenic acid.299 The rarity of severe diclofenac-induced hepatotoxicity makes liver biochemical monitoring unrealistic. Patients need to be advised to report adverse effects, and clinicians must be aware that diclofenac can cause both acute and chronic hepatitis.

MINOCYCLINE

Minocycline has been associated with rare cases of druginduced systemic lupus erythematosus syndrome (rash, polyarthritis, hyperglobulinemia, and antinuclear antibodies), chronic hepatitis with autoimmune features, and both syndromes in the same patient.300,301 The onset often occurs after treatment with minocycline for more than six months, and young women appear to be particularly affected. The reactions are severe; some patients have died or required liver transplantation.302 Progression to cirrhosis has been reported.303 The course may be prolonged after the drug is discontinued; several patients have been treated with glucocorticoids.301

DRUG-INDUCED ACUTE CHOLESTASIS IMPORTANCE, TYPES OF REACTIONS, AND DIAGNOSIS

Drugs are an important cause of acute cholestasis, with or without hepatitis.304 The full spectrum of drug-related disorders includes cholestatic hepatitis with cholangitis and chronic cholestasis, either with a vanishing bile duct syndrome resembling primary biliary cirrhosis or with biliary strictures reminiscent of sclerosing cholangitis.305,306 The clinical and biochemical features of drug-induced cholestasis resemble those of several other hepatobiliary disorders, and clinicians must take a thorough drug history from all patients with cholestasis. The prompt discontinuation of exposure to a causative agent prevents an adverse outcome and avoids unnecessary invasive investigations or surgery. The clinical syndrome of cholestasis is characterized by pruritus, dark urine, pale stools, and, in more serious cases, jaundice. Liver biochemical test results show a predominant elevation of the serum alkaline phosphastase level, with a lesser increase in the serum ALT level, elevation of GGTP and 5′-nucleotidase levels, raised serum bile acid levels, and conjugated hyperbilirubinemia. The serum ALT level may be elevated up to eight-fold, as a result of either the toxic effects of acute bile retention on hepatocellular integrity or concomitant “hepatitis.” In such cases, the ratio of the relative increases in serum ALT and alkaline phosphatase levels (based on multiples of the upper limits of normal) is typically less than 2 : 1 in patients with cholestasis.305 Cases of mixed cholestasis and hepatitis are highly suggestive of a drug reaction. Hepatobiliary imaging is essential for excluding dilatation of bile ducts produced by biliary obstruction and for detecting a hepatic or pancreatic mass lesion. In the absence of such findings, drug-induced cholestasis is more likely, and a liver biopsy is often advisable. Certain histologic features suggest a hepatic drug reaction, whereas others (e.g., edema of the portal tracts) suggest biliary obstruction. When the temporal relationship to drug ingestion indicates a high probability of a drug reaction, particularly when the agent is known to be potentially hepatotoxic, the incriminated drug should be discontinued and the patient observed for improvement. Management should focus on symptom relief; with particular attention to pruritus (see Chapters 20 and 89).305-307 Pruritus is often ameliorated with cholestyramine. In intractable cases, ursodeoxycholic acid has shown promise.307,308 Rifampin can be tried, and phototherapy, plasmapheresis, and morphine receptor antagonists (e.g., naloxone, naltrexone, nalmefene) have been used as third-line therapies.307 Glucocorticoids have no role. Phenobar­bital and antihistamines are usually ineffective or cause oversedation.

CHOLESTASIS WITHOUT HEPATITIS

Cholestatic reactions are characterized by the retention of bile in canaliculi, Kupffer cells, and hepatocytes, with minimal inflammation or hepatocellular necrosis. Synonyms are pure, canalicular, or bland cholestasis. Cholestasis without hepatitis reflects a primary disturbance in bile flow. Sex steroids are the typical causative agents. Some drugs generally associated with cholestatic hepatitis occasionally produce bland cholestasis, (e.g., amoxicillinclavulanic acid, sulfonamides, griseofulvin, ketoconazole, tamoxifen, warfarin, ibuprofen).305,306 Cyclosporine is associated with liver biochemical test abnormalities; the fea-

1439

1440

Section IX  Liver tures resemble those of cholestasis, but hyperbilirubinemia usually is predominant.5 The reaction is mild and reverses rapidly with a reduction in dose. Cholestasis also has been documented with tacrolimus,309 whereas sirolimus has been implicated in cases of mild acute hepatitis.310

oxypenicillins,319 ketoconazole [see earlier],211 sulfonylureas, sulindac,320 ibuprofen, phenylbutazone, piroxicam,321 captopril,322 flutamide,323 enalapril,153 pravastatin,324 ator­ vastatin,325 ticlopidine,326 ciprofloxacin,327 norfloxacin,328 and metformin.329)

Steroids

Chlorpromazine

Oral Contraceptive Steroids The frequency of cholestasis with oral contraceptive steroids (OCS) is 2.5 per 10,000 women exposed. The occurrence of cholestasis with OCS is partly dose dependent and less likely with low-dose than high-dose estrogen preparations.311 Genetic factors influence the frequency of this complication, with a particularly high rate observed among women in Chile and Scandinavia.306 Persons with a history of cholestasis of pregnancy are also at risk (50%). The estrogenic component is most likely responsible. Symptoms develop two to three months, rarely as late as nine months, after OCS are started. A mild transient prodrome of nausea and malaise may occur and is followed by pruritus and jaundice. The serum alkaline phosphatase level is moderately elevated, and serum aminotransferase levels are increased transiently, occasionally to levels exceeding 10 times the upper limit of normal. The serum GGTP level is often normal. Recovery is usually prompt, within days to weeks after cessation of the drug. Chronic cholestasis is rare.306 Hormonal replacement therapy (HRT) is safe in patients with liver disease. Jaundiced patients, however, may experience an increase in serum bilirubin levels, and liver biochemical test levels should be monitored in HRT users with liver disease.306 Anabolic Steroids At high doses, anabolic steroids often produce reversible bland cholestasis, usually within one to six months of the start of treatment. Recovery usually follows drug withdrawal, but protracted cholestasis with biliary ductopenia can occur.312,313 Rarely, anabolic steroid use can also cause a predominantly acute hepatocellular pattern of injury.314 Both OCS and the 17-alkylated anabolic steroids are associated with cholestasis, vascular lesions, and hepatic neoplasms (see later). The strength of these associations with individual lesions varies. Benign hepatic neoplasms, except hemangiomas,315 are clearly associated with use of OCS, whereas their association with hepatocellular carcinoma is controversial.316 By contrast, hepatocellular carcinoma is well documented in users of anabolic steroids. Likewise, hepatic and portal vein thrombosis is an established adverse effect of OCS but not of anabolic steroids. Other vascular lesions, such as peliosis hepatis (see Chapter 83), are observed more often with anabolic steroids than with OCS.

CHOLESTASIS WITH HEPATITIS

Cholestasis with hepatitis is a common type of hepatic drug reaction and is characterized by conspicuous cholestasis and hepatocellular necrosis. Histologic lesions in the liver include lobular and portal tract inflammation, often with neutrophils and eosinophils, as well as mononuclear cells. This type of reaction overlaps with drug-induced acute hepatitis (occasionally resulting in acute liver failure), cholestasis without hepatitis, and cholestasis with bile duct injury. Causative agents include chlorpromazine (see later), antidepressants and other psychotropic agents, erythromycins and other macrolides, and related ketolide anti­biotics (telithromycin,317 clindamycin,318 sulfonamides,

Chlorpromazine hepatitis, the prototypical drug-induced cholestatic hepatitis,330 has been recognized since the 1950s, and cases still occur. The full spectrum of hepatic reactions to chlorpromazine includes asymptomatic liver biochemical test abnormalities in 20% to 50% of persons exposed to the drug and rare cases of fulminant hepatic necrosis. The frequency of cholestatic hepatitis varies from 0.2% to 2.0%, depending on the type of study; the lower value probably is representative of the risk in the general population. No relationship to dose or to underlying liver disease has been recognized. Female predominance is evident. Reactions do not appear to be more common with increasing age but are rare in children. The onset of cholestatic hepatitis is generally 1 to 6 weeks after the start of chlorpromazine and occasionally 5 to 14 days after its discontinuation. Accelerated onset occurs with rechallenge. A prodromal illness of fever and nonspecific symptoms is usual and is followed by gastrointestinal symptoms and jaundice. Pruritus is common and occurs later with chlorpromazine hepatitis than with druginduced cholestasis without hepatitis. In a small proportion of affected patients, right upper quadrant abdominal pain is severe. Rash is infrequent. Liver biochemical test results show elevation of both serum ALT and alkaline phosphatase levels and hyperbilirubinemia. Eosinophilia is present in 10% to 40% of patients. Most patients with chlorpromazine hepatitis recover completely—one third within four weeks, another third between four and eight weeks, and the remainder after eight weeks.308,330 In about 7% of cases, full recovery has not occurred by six months (see later).

Amoxicillin-Clavulanic Acid

At least 150 cases of cholestatic hepatitis have been attributed to the use of amoxicillin-clavulanic acid (e.g., Augmentin), a commonly prescribed antibiotic. The overall frequency is 1.7 cases per 10,000 prescriptions19; male gender, increasing age (older than 55 years), and possibly use of the antibiotic for a prolonged period of time are risk factors.331 The clavulanic acid component has been implicated because similar lesions have been noted with ticarcillin-clavulanic acid,332 whereas amoxicillin rarely causes liver disease.333 The onset of symptoms is within 6 weeks (mean 18 days) of the start of drug therapy, although rarely the onset is delayed up to 6 weeks after discontinuation of the drug. Features of hypersensitivity such as fever, skin rash, and eosinophilia are seen in 30% to 60% of patients. Liver biopsy specimens show cholestasis with mild portal inflammation. Bile duct injury (usually mild) and perivenular bilirubinostasis with lipofuscin deposits are often present. Other histologic features include hepatic granulomas, biliary ductopenia, and cirrhosis.334 Most patients recover completely in 4 to 16 weeks; fatal outcomes are rare. A strong association with the HLA-DRB1*1501-DRB5*0101DQB1*062 haplotype, supports the view that an immunologic idiosyncrasy mediated through HLA class II antigens could play a pathogenic role in this form of drug-induced cholestatic hepatitis. The presence of this haplotype, however, has no influence on the clinical characteristics, severity, and outcome of the disease.

Chapter 86  Liver Disease Caused by Drugs CHOLESTATIC HEPATITIS WITH BILE DUCT INJURY

Bile duct (cholangiolytic) injury is observed with several drugs that cause cholestatic hepatitis, such as chlorpromazine308 and flucloxacillin.319 The severity of bile duct injury may be a determinant of the vanishing bile duct syndrome (see later).335,336 The clinical features may resemble those of bacterial cholangitis, with upper abdominal pain, fever, rigors, tender hepatomegaly, jaundice, and cholestasis. Liver biochemical test levels are typical of cholestasis. Compounds associated with this syndrome include arsphenamine, carbamazepine,337 dextropropoxyphene,338 and methylenediamine, an industrial toxin responsible for Epping jaundice, an outbreak of jaundice associated with the ingestion of bread made from contaminated flour (see Chapter 87).339

Dextropropoxyphene

Dextropropoxyphene, an opioid analgesic used alone or in compound analgesics, has caused cholestasis with bile duct injury in at least 25 reported cases,338 some proved by inadvertent rechallenge. A female predominance has been recognized. The onset of symptoms is usually within two weeks of the start of dextropropoxyphene. Illness is often heralded by abdominal pain, which may be severe and simulate other causes of cholangitis. Jaundice is usual. ERCP shows normal bile ducts. Liver biopsy specimens demonstrate cholestasis with expansion of the portal tracts by inflammation and mild fibrosis; portal tract edema also may be present. Other features include irregularity and necrosis of the biliary epithelium, together with an infiltrate of neutrophils and eosinophils on the outer surface of bile ducts. Bile ductule proliferation is universal. Recovery has occurred in all reported cases.338 Liver biochemical test levels normalize between one and three months after discontinuation of the drug. This type of hepatic drug reaction must be distinguished from bile duct obstruction and bacterial cholangitis.

DRUG-INDUCED CHRONIC CHOLESTASIS Drug-induced liver disease is considered to be chronic when typical liver biochemical changes last longer than 3 months306; earlier definitions required the presence of jaundice for more than 6 months or anicteric cholestasis (raised alkaline phosphatase and GGTP levels) for more than 12 months after the implicated agent was stopped.305 Drug-induced chronic cholestasis is uncommon but has been ascribed to more than 45 compounds.305-307,330,340-342 Chronicity complicates approximately 7% of cases of chlorpromazine hepatitis308 and is a feature in 10% to 30% of cases of flucloxacillin hepatitis.306 Chronicity has been reported in less than 5% of cases of erythromycin hepatitis341 and in only isolated instances for other agents, such as tetracycline,342 amoxicillin-clavulanic acid,343 ibuprofen,344,345 trimethoprim-sulfamethoxazole,346 and ciprofloxacin.347 Chronic cholestasis always is preceded by an episode of acute cholestatic hepatitis. The episode of acute cholestatic hepatitis tends to be severe and occasionally is associated with the Stevens-Johnson syndrome.345 One study indicated that the severity of bile duct lesions at the time of the initial hepatic reaction is a critical determinant of a chronic course.336 Other possible mechanisms include continuing toxic or immunologic destruction of the biliary epithelium.340 The hepatic histologic lesion is characterized by a

paucity of smaller (septal, interlobular) bile ducts and ductules, often with residual cholestasis, and portal tract inflammation directed against injured bile ducts. This process may lead to an irreversible loss of biliary patency and the vanishing bile duct syndrome.348 The clinical features are those of chronic cholestasis. Pruritus is the dominant symptom and is often severe. Continuing jaundice, dark urine, and pale stools are possible but not invariable findings and may resolve despite persistence of liver biochemical abnormalities. In severe cases, intestinal malabsorption, weight loss, and bruising caused by vitamin K deficiency may occur; xanthelasma, tuberous xanthomata, and other complications of severe hypercholesterolemia also have been noted. Firm hepatomegaly may be found on physical examination, but splenomegaly is unusual unless portal hypertension develops. Antimitochondrial antibodies are not a feature of drug-induced chronic cholestasis. Cases usually have a favorable outcome, with resolution of jaundice in most instances. Progression to biliary cirrhosis is rare305,306 and is associated with a severe reduction in the number of bile ducts.

FLUCLOXACILLIN

Flucloxacillin is one of the most important causes of druginduced hepatitis in Europe, Scandinavia, and Australia.319,349 Flucloxacillin-induced hepatotoxiciy is usually severe, and several fatalities have resulted from the systemic features and associated cholestatic hepatitis. The course is prolonged, and a high proportion of cases result in chronic cholestasis and the vanishing bile duct syndrome.349 Other oxypenicillins appear to be less prone to cause this complication, but cholestasis has been reported with cloxacillin and dicloxacillin.319,350 Acute hepatocellular injury has been reported with oxacillin.351

FIBROTIC BILE DUCT STRICTURES

Chronic cholestasis caused by some drugs may result from the development of fibrotic strictures of the larger bile ducts. This complication has been associated with intralesional therapy of hepatic hydatids with formalin352 and intraarterial infusion of floxuridine for metastatic colorectal carcinoma.353 After several months of floxuridine infusion, the frequency of toxic hepatitis or bile duct injury, or both, is 25% to 55%. Acalculous cholecystitis also may occur. ERCP shows strictures, typically in the common, left, and right hepatic ducts. Unlike primary sclerosing cholangitis, the bile duct and the smaller intrahepatic bile ducts are spared. Ischemia has been suspected, and toxicity to biliary epithelial cells is another possibility. Recovery may occur after floxuridine is discontinued. Some patients require dilation or stenting of biliary strictures.

DRUG-INDUCED STEATOHEPATITIS AND HEPATIC FIBROSIS Drug-induced liver disease can produce cirrhosis by a variety of processes; chronic hepatitis and chronic cholestasis with the vanishing bile duct syndrome have already been discussed. Steatohepatitis is a form of chronic liver disease in which fatty change is associated with focal liver cell injury, Mallory’s hyaline, focal inflammation of mixed cellularity, including neutrophils, and progressive hepatic fibrosis in a pericentral (zone 3) and pericellular distribution.354 Alcohol is a common etiologic factor. NASH is associated with insulin resistance, diabetes mellitus, obesity, and several drugs (e.g., perhexiline maleate and amioda-

1441

1442

Section IX  Liver rone).355 In addition to causing steatohepatitis or chronic injury to liver cells or bile ducts, some exogenous compounds appear to promote hepatic fibrogenesis directly, most likely through effects on hepatic nonparenchymal cells; stellate cells are central to this process (see Chapter 71). Compounds that stimulate hepatic fibrosis include arsenic, vitamin A, and methotrexate.

AMIODARONE

Amiodarone is an iodinated benzofuran derivative used for therapy-resistant ventricular tachyarrhythmias. Adverse effects lead to discontinuation of therapy in 25% of patients. These adverse effects include pulmonary infiltrates, worsening cardiac failure, hypothyroidism, peripheral neuropathy, nephrotoxicity, and corneal deposits, but liver disease is one of the most serious. The spectrum of abnormalities includes abnormal liver biochemical test levels in 15% to 80% of patients and clinically significant liver disease, including rare cases of acute liver failure, in 0.6%. Liver disease also has been reported in patients who receive an intravenous loading dose of amiodarone; the toxic ingredient is likely to be the vehicle (polysorbate 80) rather than amiodarone because oral amiodarone has been used subsequently in these cases without a problem.356-361 The most typical hepatic lesion is steatohepatitis; cirrhosis develops in 15% to 50% of patients with hepatoxicity.357,358 A notable feature of amiodarone-induced liver disease is that progression of the disease may occur despite discontinuation of amiodarone.358,361 Amiodarone is highly concentrated in the liver, and after a few weeks of treatment, the drug accounts for as much as 1% of the wet weight of the liver. The iodine content absorbs radiation, so that the liver appears opaque on computed tomography scans.361 Although odd, this appearance is not clinically significant. Hepatic storage of amiodarone also produces phospholipidosis, a storage disorder characterized by enlarged lysosomes stuffed with whorled membranous material (myeloid bodies). In animals fed amiodarone, development of phospholipidosis is time and dose dependent.359 Phospholipi­ dosis may result from the direct inhibition of phospholipase or from the formation of nondegradable drug-phospholipid complexes and appears to have no relationship to the development of NASH and hepatocyte injury. Other occasional hepatic abnormalities include granuloma formation and acute liver failure, apparently caused by severe acute hepatitis or a Reye’s syndrome-like illness.362 Amiodarone, by virtue of its physicochemical properties, is concentrated in mitochondria and may interrupt mitochondrial electron transport.363 In rats and mice, treatment with amiodarone produces microvesicular steatosis, augments mitochondrial production of ROS, and causes lipid peroxidation.364,365 Chronic liver disease is detected only 1 year or more (median, 21 months) after amiodarone is started. The duration of amiodarone therapy and possibly the total dose,362,366 but not the incremental dose, are related to the development of chronic liver disease. Cases of cirrhosis with low-dose amiodarone have been documented.367 The frequency of other toxic effects of amiodarone (most of which are thought to be dose dependent) is increased in patients with liver disease.366 Patient complaints include fatigue, nausea and vomiting, malaise, weight loss, and abdominal swelling as a result of ascites. Hepatomegaly, jaundice, bruising, and other features of chronic liver disease may be present. Liver biochemical test abnormalities include increased aminotransferase levels, often to at least five times the upper limit

of normal, and minor increases in the serum alkaline phosphatase level. The ratio of serum AST to ALT levels is close to unity and thus differs from the ratio seen in patients with alcoholic hepatitis. In severe cases, hyperbilirubinemia, a low serum albumin level, and prolongation of the prothrombin time are evident. Diagnosing the cause of abnormal liver biochemical test results and hepatomegaly is often difficult in patients taking amiodarone, and a liver biopsy may be indicated. The histologic changes in the liver include phospholipidosis, steatosis, focal necrosis with Mallory’s hyaline, infiltration with neutrophils, and pericellular fibrosis.358 Cirrhosis is often present. Prevention and management of amiodarone-induced liver disease are problematic because liver biochemical test abnormalities are common in patients who take amiodarone, particularly in those with congestive heart failure. Furthermore, the frequency of amiodarone hepatotoxicity does not appear to differ between patients with and without baseline serum ALT elevations, and amiodarone should not be withheld in patients with an elevated serum ALT level.368 In asymptomatic or less severe cases of amiodarone hepatotoxicity, abnormalities resolve in two weeks to four months after amiodarone is discontinued. In cases of severe liver disease, the mortality rate is high.358,366 Cessation of amiodarone therapy does not always result in clinical improvement, presumably because of prolonged hepatic storage of amiodarone, and in one study, outcome was worse (usually from fatal arrhythmias) in patients who discontinued amiodarone than in those who did not.358 Although serial liver biochemical test measurements are recommended in patients who take amiodarone,366 whether such testing is adequate to prevent serious hepatotoxicity and reduce the overall mortality rate is unknown.

TAMOXIFEN AND OTHER CAUSES OF DRUG-INDUCED STEATOHEPATITIS

For agents reported to be associated with steatohepatitis during the 1990s, causality has been difficult to prove,369 particularly because NASH is a common disorder among patients with obesity, insulin resistance, or metabolic syndrome (see Chapter 85). Calcium channel blockers, used to treat hypertension and cardiac arrhythmias, have rarely been associated with steatohepatitis,370,371 and methyldopa has been reported to be associated with cirrhosis in obese middle-aged women372; however, these associations may have been fortuitous. Other drugs, including estrogens373 and glucocorticoids,374 may precipitate NASH in predisposed persons because of their effects on the risk factors for NASH—insulin resistance, type 2 diabetes mellitus, obesity, and hypertriglyceridemia. On the other hand, the association between NASH and tamoxifen appears to be much stronger. Tamoxifen is an estrogen-receptor ligand with both agonist and antagonist actions. It is widely used for the prevention and treatment of breast cancer. Several forms of liver injury have been attributed to tamoxifen375: cholestasis,376 hepatocellular carcinoma,377 peliosis hepatis,378 acute hepatitis, massive hepatic necrosis,375 steatosis, and steatohepatitis—occasionally with cirrhosis.379-385 In one series of 66 women with breast cancer who had received tamoxifen for three to five years, 24 showed radiologic evidence of hepatic steatosis.382 Seven other patients have been diagnosed with NASH (proved by liver biopsy) after taking tamoxifen for 7 to 33 months.380,381,383 The metabolic profile of women with radiologic evidence of hepatic steatosis (or histologic proof of steatohepatitis) during tamoxifen therapy appears similar to that of most

Chapter 86  Liver Disease Caused by Drugs patients with NASH; one half have been obese, and the increase in body mass index has correlated with hepatic steatosis.386 Tamoxifen can induce hypertriglyceridemia, another risk factor for NASH. Treatment with bezafibrate, a PPAR-α stimulator, has been reported to decrease the severity of hepatic steatosis, as assessed radiologically, markedly.387 Therefore, tamoxifen may play a synergistic role with other factors such as insulin resistance, hyperli­ pidemia, and obesity in causing steatohepatitis. This hypothesis is supported by results of an Italian tamoxifen chemoprevention trial in which development of fatty liver or steatohepatitis was confined mainly to overweight or obese women with the metabolic syndrome.388 Physicians need to be aware of the high frequency (approximately 30%) of hepatic steatosis, as determined by hepatic imaging, or steatohepatitis in women who receive tamoxifen. Patients who take tamoxifen should be monitored for this adverse effect by physical examination (to detect hepatomegaly) and liver biochemical testing; some authors also advocate annual hepatic imaging (by ultrasonography or computed tomography).389 Liver biopsy may be indicated to establish the severity of the disorder, particularly if liver biochemical test abnormalities do not resolve after tamoxifen is discontinued, or, in some cases, to exclude metastatic breast cancer. Many patients appear to improve after tamoxifen is discontinued, but whether treatment should always be withdrawn permanently is not clear, particularly because the effect of tamoxifen on survival from breast cancer is impressive. An alternative option to bezafibrate (see earlier) is to consider the use of exemstane (or other aromatose inhibitor), which can lower serum triglyceride levels.390 Toremifene, an analog of tamoxifen, also has been reported to cause steatosis or steatohepatitis, but with a lower frequency (less than 10%) than that reported for tamoxifen.389 Raloxifene, a selective estrogen receptor modulator, has been implicated in a case of steatohepatitis391 and a case of acute hepatocellular injury accompanied by eosinophilia.392 Although evidence of acute hepatocellular injury was provided in the first case, causality could not be established in

the second case because preexisting liver disease (NASH) could not be excluded. Perhexiline maleate and coralgil (4,4′-diethylaminoethoxyhexestrol) are definite toxic causes of steatohepatitis and have been withdrawn from the U.S. market.369

METHOTREXATE

Methotrexate is a dose-dependent toxin. In high doses, methotrexate can result in bone marrow suppression, mucocutaneous reactions, pneumonitis, and hepatotoxicity. In the 1950s, previous methotrexate treatment of acute childhood leukemia was shown to be complicated by severe hepatic fibrosis and cirrhosis; a few cases were complicated by hepatocellular carcinoma.393 In the 1960s, the use of methotrexate for psoriasis was found to be associated with the development of hepatic fibrosis and cirrhosis in as many as 25% of cases.394 Since then, a clearer picture of methotrexate as a dose-dependent promoter of hepatic fibrosis has emerged, particularly in persons who drink alcohol excessively or have preexisting liver disease. Guidelines have been instituted for scheduled pretreatment and interval liver biopsies to monitor the safety of methotrexate therapy. Methotrexate is now used more often as a low-dose weekly regimen in the management of rheumatoid arthritis, psoriasis, and other immunologic conditions, including inflammatory bowel disease. The problem of hepatotoxicity has largely been overcome by the avoidance of daily dosing with methotrexate and a reduction in the weekly dose to 5 to 15 mg (see also Chapter 22).395-397

Risk Factors

Risk factors for methotrexate-induced hepatic fibrosis are listed in Table 86-7; dose, alcohol intake, and preexisting liver disease are the most important.396,397 Total dose, incremental dose, dose interval, and duration of methotrexate therapy each influence the risk of hepatic fibrosis. After the cumulative ingestion of 3 g of methotrexate, the chance of histologic progression is 20%, but only 3% of patients are found to have advanced hepatic fibrosis.398 Obesity and diabetes mellitus may be important risk factors for hepatic

Table 86-7  Risk Factors for Methotrexate-Induced Hepatic Fibrosis* RISK FACTOR

IMPORTANCE

IMPLICATIONS FOR PREVENTION

Age

Increased risk >60 yr; possibly related to reduced renal clearance and/or biologic effect on fibrogenesis Incremental dose Dose frequency Duration of therapy Cumulative (total) dose Increased risk with daily levels >15 g (1-2 drinks)

Care in use of methotrexate in older people

Dose

Alcohol consumption Obesity Diabetes mellitus Preexisting liver disease

Increased risk Increased risk in obese persons (type 2 diabetes mellitus) Greatly increased risk, particularly related to alcohol, obesity, and diabetes mellitus (NASH)

Systemic disease

Possibly risk greater with psoriasis than rheumatoid arthritis (may depend on preexisting liver disease, alcohol intake) Increased risk because of reduced clearance NSAIDs, vitamin A, and arsenic may increase risk

Impaired renal function Other drugs

NASH, nonalcoholic steatohepatitis; NSAIDs, nonsteroidal anti-inflammatory drugs. *See Chapter 22.

5-15 mg/wk is safe Weekly bolus (pulse) safer than daily schedules Consider liver biopsy every 2 years Consider liver biopsy after each 2 g of methotrexate Avoid methotrexate use if intake not curbed Consider pretreatment liver biopsy with relevant history Consider pretreatment and interval liver biopsies Consider pretreatment and interval liver biopsies Pretreatment liver biopsy mandatory Avoid methotrexate, or schedule interval biopsies according to severity of hepatic fibrosis, total dose, and duration of methotrexate therapy Monitor liver biochemical tests during therapy None Reduce dose; greater caution with use of methotrexate Greater caution with use; monitor liver biochemical tests

1443

1444

Section IX  Liver fibrosis because they predispose to NASH and are associated with induction of CYP2E1; the strong association between NASH and methotrexate in causing liver injury during long-term, low-dose methotrexate treatment has been highlighted,399 as has the possibility that methotrexate itself can cause a pattern of injury resembling steatohepatitis. Increasing age, impaired renal function, and concomitant use of certain drugs decrease the elimination of methotrexate or facilitate tissue uptake by displacing metho­ trexate from plasma-protein binding sites. Psoriasis and rheumatoid arthritis are associated with hepatic abnormalities that range from abnormal liver biochemical test levels (25% to 50% of cases) and minor histologic changes (50% to 70%) to fibrosis (11%) and cirrhosis (1%). In patients with psoriasis, alcoholism often is a complicating factor, and in a meta-analysis,397 alcohol consumption was the most important determinant of advanced hepatic fibrosis in patients treated with methotrexate; the risk of progressive hepatic fibrosis was 73% in persons who drank more than 15 g of alcohol daily, compared with 26% in those who did not. The possibility that low-dose (5 to 15 mg) methotrexate given as a single weekly dose can cause hepatic fibrosis has been debated.395-397 The available data are limited by a lack of controlled studies in which liver histologic findings were evaluated; the lack of pretreatment evaluation of liver histology is a particularly serious deficiency in view of the high frequency of liver abnormalities among patients with rheumatoid arthritis and psoriasis. The conclusion has been reached that current regimens that are in use can promote hepatic fibrosis, at least at the ultrastructural level, but that cases of clinically significant liver disease are now virtually unknown. Indeed, repeat liver biopsies in some series have shown a reduction in fibrosis despite continuation of methotrexate in lower doses.398 Therefore, although methotrexate remains a potential cause of liver disease, advanced hepatic fibrosis is in large part preventable.

Clinicopathologic Features

Liver biochemical test abnormalities are common among patients who take methotrexate, but advanced hepatic fibrosis occasionally can develop in the absence of such abnormalities. Likewise, nausea, fatigue, and abdominal pain are common adverse effects of methothrexate, but patients with hepatic fibrosis are typically asymptomatic unless complications of liver failure or portal hypertension, such as bleeding esophageal varices, develop. A firm liver edge, hepatomegaly, splenomegaly, and ascites may be noted. Liver biochemical test levels are either normal or show nonspecific changes, including minor elevations of serum ALT and GGTP levels. In more advanced cases, hypoalbuminemia is present, but elevation of the serum bilirubin level and coagulation disturbances are rare. Thrombocytopenia may be present in patients with cirrhosis. Liver histologic findings often are graded according to the system of Roenigk, which has been useful in analyzing the published literature.397 In this system, grades I and II indicate a variable amount of steatosis, nuclear pleomorphism, and necroinflammatory activity but no fibrosis. Higher grades reflect increasing degrees of fibrosis, as follows: grade IIIa, few septa; grade IIIb, bridging fibrosis; and grade IV, cirrhosis. The pattern of hepatic fibrosis includes pericellular fibrosis, a feature of both alcoholic steatohepatitis and NASH; the possibility that methotrexate itself causes steatohepatitis or accentuates fibrogenesis among persons with underlying “primary NASH” has been suggested.399 Cases of hepatic fibrosis in livers with a relative paucity (or complete

absence) of portal and lobular inflammation have been reported.

Outcome and Prevention

Serious clinical sequelae (portal hypertension, liver failure, hepatocellular carcinoma) resulting from methotrexateinduced liver disease are now rarely seen. In a study of 32 patients with inflammatory bowel disease who received long-term methotrexate (mean dose of 2.6 g; follow-up period of 131 weeks), minor histologic changes in the liver were common, but advanced hepatic fibrosis was rare. Cases that have come to liver transplantation generally have been associated with suboptimal supervision of methotrexate therapy.400 Cases of severe hepatic fibrosis (Roenigk grades IIIb and IV) are often associated with lack of progression and even improvement after discontinuation of methotrexate or a reduction in the dose.398 In less severe cases, a balanced judgment must be made about the appropriateness of continuing or discontinuing methotrexate. An interval liver biopsy after an additional two years or 2 g of methotrexate may be judicious in a patient who is found to have minor hepatic fibrosis on a liver biopsy specimen. Recommendations for preventing methotrexate-induced hepatic fibrosis have been made (see Chapter 35).396,401 If possible, methotrexate should be avoided when the risk of liver injury is high. Persons treated with methotrexate should abstain from alcohol, and those who drink more than 100 g of ethanol per week should not be given methotrexate.396,397,401 A pretreatment liver biopsy is indicated only if the liver biochemical test levels are abnormal or if the history (e.g., alcoholism) and clinical features (e.g., hepatomegaly, risk factors for NASH) indicate possible underlying liver disease.402 The use of liver biochemical testing to monitor the patient’s progress during treatment with methotrexate is recommended but is problematic because of the lack of specificity and sensitivity of the tests; four to six sets of liver biochemical tests are often performed each year in patients undergoing treatment with methotrexate. Persistent or recurrent elevations of the serum AST or ALT levels, any decrease in the serum albumin level, or the development of hepatomegaly warrants investigation by liver biopsy. Scheduled liver biopsies are recommended after a cumulative dose of 4 g of methotrexate or duration of therapy of two years.397 Other authorities have suggested that a liver biopsy be performed after 5 g of methotrexate.403 Whether a liver biopsy is necessary in patients with normal liver biochemical test levels and without major risk factors for hepatic fibrosis remains unclear.398,401 Some authors have suggested a lower threshold for liver biopsy (cumulative dose of 1.5 g and after every 1 g of methotrexate thereafter) for patients with no risk factors for hepatoxicity.404 Serum biochemical tests that indicate progressive hepatic fibrosis, such as procollagen peptide-3 levels, have not proved sufficiently accurate for monitoring patients taking methotrexate.

DRUG-INDUCED VASCULAR TOXICITY Vascular injury to the liver may give rise to several unusual types of liver disease, including sinusoidal obstruction syndrome (formerly veno-occlusive disease, a form of hepatic venous outflow obstruction), peliosis hepatis (dilatation and destruction of hepatic sinusoids), noncirrhotic portal hypertension, and nodular regenerative hyperplasia. Drugs and chemical toxins are the most common causes of hepatic

Chapter 86  Liver Disease Caused by Drugs Table 86-8  Types of Drug-Induced Hepatic Vascular Disorders: Clinicopathologic Features, Outcome, and Implicated Etiologic Agents DISORDER

CLINICOPATHOLOGIC FEATURES

OUTCOMEs

IMPLICATED ETIOLOGIC AGENTS

Sinusoidal obstruction syndrome (veno-occlusive disease) Nodular regenerative hyperplasia

Abdominal pain, tender hepatomegaly, ascites, liver failure; occasionally chronic liver disease, other signs of portal hypertension Portal hypertension, encephalopathy— especially after variceal bleeding; diagnosed by histology Splenomegaly, hypersplenism, varices; ascites if associated hepatocellular disease Incidental finding, hepatomegaly, hepatic rupture, liver failure; diagnosed from appearances at surgery, vascular imaging Hepatomegaly, abdominal pain

High mortality rate; some cases may evolve into nodular regenerative hyperplasia Relatively good prognosis

Especially in bone marrow transplantation: 6-thioguanine, busulfan; dactinomycin, azathioprine, mitomycin; pyrrolizidine (e.g., in comfrey) Anticancer drugs: busulfan, dactinomycin; azathioprine

Prognosis depends on cause and associated liver injury

Vitamin A, methotrexate, azathioprine, arsenic, vinyl chloride, anticancer drugs

Prognosis depends on cause and complications

Anabolic steroids, azathioprine, 6-thioguanine

May regress after stopping oral contraceptives

Oral contraceptive steroids

Noncirrhotic portal hypertension Peliosis hepatis

Sinusoidal dilatation

vascular injury.405 The mechanism of injury is primarily dose-dependent toxicity to sinusoidal and other vascular endothelial cells, particularly when drugs are used in combination or concurrently with radiotherapy. Activation of inflammatory cells also may be important. Individual drugs (e.g., azathioprine) have been associated with more than one vascular syndrome, and the individual disorders overlap and may evolve from one type to another. Vascular toxicity may give rise to a continuum of disorders, each resulting from injury to different components of the hepatic vasculature. The essential features of these disorders are summarized in Table 86-8, and the more important conditions are discussed in Chapter 83. Hepatic imaging and measurement of portal pressure play a role in the diagnosis of these conditions; some, particularly nodular regenerative hyperplasia, are difficult to confirm in needle biopsy specimens.

AZATHIOPRINE

Hepatic complications of azathioprine, although rare, may be severe, diverse, and often late in onset. The overall frequency of azathioprine hepatotoxicity is less than 0.1%. Many cases occur in complex medical situations, particularly organ transplantation, in which activation of the immune system, viral infections, and other agents may increase the risk of hepatotoxicity (see Chapter 34). The central role of azathioprine has been confirmed in some cases that resolved after discontinuation of the drug and others in which a positive rechallenge was documented.406,407 Disturbances associated with azathioprine include bland cholestasis, cholestatic hepatitis with bile duct injury,407,408 zonal necrosis, and vascular toxicity.406 Vascular toxicity may give rise to the diverse syndromes of sinusoidal obstruction syndrome, peliosis hepatis, nodular regenerative hyperplasia, and noncirrhotic portal hypertension.406-409 Hepatocellular carcinoma with focal glycogenosis also has been reported in a long-term recipient of azathioprine.410 Azathioprine is associated with an extraordinary range of hepatic disorders, including liver biochemical test abnormalities in asymptomatic patients, bland cholestasis, cholestatic hepatitis, bile duct injury, and vascular injury. Cholestatic hepatitis is probably the most common; several cases have been associated with zone 3 necrosis and congestion, suggesting acute vascular injury, and azathioprine shares the vascular toxicity of other thiopurines. All hepatic

syndromes that result from vascular injury have been associated with azathioprine, particularly after organ transplantation. Cases of azathioprine-induced nodular regenerative hyperplasia and sinusoidal obstruction syndrome also have been reported in patients treated with azathioprine for other medical conditions, including inflammatory bowel disease.405,411 No relation between toxicity and the dose or duration of azathioprine therapy has been observed, but men are almost exclusively involved in cases of hepatic vascular injury following renal transplantation. The onset of cholestatic reactions is 2 weeks to 22 months after the start of azathioprine, but vascular toxicity is recognized later, typically 3 months to 3 years, and occasionally more than 9 years, after transplantation.409,410 The presentation and clinical features depend on the type of reaction. Cases of later onset are the result of delayed recognition and tend to be associated with complications of portal hypertension, including ascites, and liver failure. Recovery can occur in such cases,407 but the overall mortality rate is high. Azathioprine is used as an immunosuppressive agent following solid organ transplantation and as a glucocorticoid-sparing agent in some autoimmune diseases, including autoimmune hepatitis (see Chapter 88). In contrast to azathioprine, 6-mercaptopurine is a dose-dependent cause of hepatocellular necrosis and has been fatal in a few cases and, even more rarely, associated with cholestasis.5

LIVER TUMORS Several associations between pharmacologic and environmental agents and benign and malignant liver tumors have been described, but causality has been difficult to prove because of the rarity of these associations. For some sex steroid-related tumors, as well as for vinyl chloride-induced angiosarcoma, the relative risk attributable to the causative agent has been determined. Prevention and early detection are critical for improving outcomes. The major tumors of interest include cavernous hemangioma, focal nodular hyperplasia, hepatic adenoma, hepatocellular carcinoma, angiosarcoma, hepatoblastoma, cholangiocarcinoma, mixed carcinosarcoma, and epithelioid hemangioendothelioma (see Chapters 69 and 94).

1445

1446

Section IX  Liver KEY REFERENCES

Biour M, Poupon R, Grange JD, et al. Drug-induced hepatotxicity. The 13th updated edition of the bibliographic database of drug-related liver injuries and responsible drugs. Gastroenterol Clin Biol 2000; 24:1052-91. (Ref 10.) Björnsson E, Olsson R. Outcome and prognostic markers in severe druginduced liver disease. Hepatology 2005; 42:481-9. (Ref 8.) Bruno S, Maisonneuve P, Castellana P, et al. Incidence and risk factors for non-alcoholic steatohepatitis: Prospective study of 5408 women enrolled in Italian tamoxifen chemoprevention trial. BMJ 2005; 330:932-5. (Ref 388.) Chalasani N, Aljadhey H, Kesterson J, et al. Patients with elevated liver enzymes are not at higher risk for statin hepatotoxicity. Gastroenterology 2004; 128:1287-92. (Ref 162.) Daly AK, Aithal GP, Leathart JB, et al. Genetic susceptibility to diclofenac-induced hepatotoxicity: Contribution of UGT2B7, CYP2C8, and ABCC2 genotypes. Gastroenterology 2007; 132:272-81. (Ref 297.) Fontana RJ, McCashland TM, Benner KG, et al. Acute liver failure associated with prolonged use of bromfenac leading to liver transplantation. The Acute Liver Failure Study Group. Liver Transpl Surg 1999; 5:480-4. (Ref 282.) Ibanez L, Perez E, Vidal X, et al. Prospective surveillance of acute serious liver disease unrelated to infectious, obstructive, or metabolic

diseases: Epidemiological and clinical features, and exposure to drugs. J Hepatol 2002; 37:592-600. (Ref 4.) Kosters A, Karpen SJ. Bile acid transporters in health and disease. Xenobiotica 2008; 38:1043-71. (Ref 38.) Lewis JH, Mortensen ME, Zweig S, et al. Efficacy and safety of high-dose pravastatin in hypercholesterolemic patients with well-compensated chronic liver disease. Hepatology 2007; 46:1453-63. (Ref 160.) Lucena MI, Andrade RJ, Fernández MC, et al. Determinants of the clinical expression of amoxicillin-clavulanate hepatotoxicity: A prospective series from Spain. Hepatology 2006; 44:850-6. (Ref 334.) Meier Y, Cavaliaro M, Roos M, et al. Incidence of drug-induced liver injury in medical patients. Eur J Clin Pharmcol 2005; 61:135-43. (Ref 66.) Rostom A, Goldkind L, Laine L. Nonsteroidal anti-inflammatory drugs and hepatic toxicity: A systematic review of randomized controlled trials in arthritis patients. Clin Gastroenterol Hepatol 2005; 3:489-98. (Ref 285.) Zimmerman HJ. Hepatotoxicity. The Adverse Effects of Drugs and Other Chemicals on the Liver, 2nd ed. Philadelphia: Lippincott, Williams and Wilkins; 1999. (Ref 6.) Zollner G, Marschall HU, Wagner M, et al. Role of nuclear receptors in the adaptive response to bile acids and cholestasis: Pathogenetic and therapeutic considerations. Mol Pharm 2006; 3:231-51. (Ref 37.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

87 Liver Disease Caused by Anesthetics, Toxins, and Herbal Preparations James H. Lewis

CHAPTER OUTLINE Anesthetic Agents  1447 Halothane  1447 Other Anesthetic Agents  1449 Jaundice in the Postoperative Period  1449 Chemicals  1450 Commercial and Industrial Chemicals  1450 Pesticides  1452 Metals  1453 Iron  1453 Phosphorus  1453 Copper Salts  1453 Thorium Dioxide  1453 Other Metals  1453

Although halothane hepatitis is now largely of historical interest, it holds an important place in the annals of causality assessment in drug-induced liver disease.1 In contrast with the largely unpredictable hepatotoxicity seen with more modern anesthetics and most other medicinal agents (as discussed in Chapter 86), liver damage caused by occupationally and environmentally encountered chemical compounds and other toxins often is more predictable, dose related, and predominantly cytotoxic in nature.1-4 Industrial exposure to hepatotoxic chemicals is a less frequent occupational hazard today than in the past, but reports of toxicity from chemical agents, as well as metals, adulterated cooking oils, and botanical toxins, have not disappeared.3,4 Additionally, the use of complementary and alternative medicine (CAM) preparations continues to increase, and reports of liver injury from potentially hepatotoxic herbal and weight loss products continue to appear (see Chapter 127).5,6 Anesthetics, herbal products, mushrooms, and other toxins continue to account for a substantial percentage of emergency liver transplants for acute liver failure.7

ANESTHETIC AGENTS The volatile inhalational anesthetics in current use are derivatives of some of the most potent chemical hepato­ toxins developed for medicinal purposes. Chloroform, the original haloalkane anesthetic, has long been abandoned but remains an important experimental hepatotoxin, as does carbon tetrachloride (another chlorinated aliphatic hydro-

Adulterated Cooking Oils and Contaminated Foods  1453 Drugs of Abuse  1454 Cocaine  1454 Others  1454 Botanical and Environmental Hepatotoxins  1454 Mushrooms  1454 Other Foodstuffs  1455 Vitamins and Herbal Preparations  1455 Hypervitaminosis A  1455 Herbal Remedies and Nutritional Supplements  1456

carbon), which found use as an early vermifuge and is still employed as a household reagent in some parts of the world.1,8 Halothane (fluothane), introduced in the 1950s as a safer, nonexplosive alternative to ether, is a haloalkane compound that produced a well-described but rare syndrome of acute hepatotoxicity, usually after repeat exposure.9 The anesthetics that followed-methoxyflurane, enflurane, isoflurane—all have been implicated as a cause of similar injury, albeit less commonly for enflurane and isoflurane than for halothane; even fewer instances have been reported for the newest agents, sevoflurane and desflurane,10,11 because of their proportionally lower degree of metabolism.12 Halothane is no longer being produced in the United States but continues to be employed in other countries13 and is a case study in the elucidation of immunologic-mediated liver injury.14

HALOTHANE

The retrospective National Halothane Study, cited in the past as the basis for exonerating halothane as a cause of hepatotoxicity,15 is now considered flawed.1 Nearly 1000 cases of halothane hepatotoxicity were reported worldwide during the 1960s and 1970s.1,9,16 A fairly uniform clinical picture of postoperative fever, eosinophilia, jaundice, and hepatic necrosis occurred a few days to weeks after administration of anesthesia, usually after repeat exposure to halothane, and the case-fatality rate was high (Table 87-1). Rare cases of halothane-induced liver injury were reported after workplace exposure among anesthesiologists, surgeons, nurses, and laboratory staff and after halothane sniffing for “recreational” use; in affected persons, antibodies to trifluo-

1447

1448

Section IX  Liver Table 87-1  Clinicopathologic Features of Halothane Hepatitis Estimated incidence After first exposure: 0.3 to 1.5 per 10,000 After multiple exposures: 10 to 15 per 10,000 Female-to-male ratio 2 : 1 Latent period to first symptom After first exposure: 6 days (11 days to jaundice) After multiple exposures: 3 days (6 days to jaundice) Jaundice as presenting symptom in 25% (range of serum bilirubin: 3-50 mg/dL) Fever in 75% (precedes jaundice in 75%); chills in 30% Rash in 10% Myalgias in 20% Ascites, renal failure, and/or gastrointestinal hemorrhage in 20%-30% Eosinophilia in 20%-60% Serum ALT and AST levels: 25-250 × ULN Serum alkaline phosphatase level: 1-3 × ULN Histologic features Zone 3 massive hepatic necrosis in 30%; submassive necrosis in 70% (autopsy series) Inflammation usually less marked than in viral hepatitis Eosinophilic infiltrate in 20% Granulomatous hepatitis occasionally Course and outcome Mortality rate (pretransplantation era): 10%-80% Symptoms can resolve within 5-14 days Full recovery can take 12 weeks or longer Chronic hepatitis not well documented Adverse prognostic findings Age >40 years Obesity Short duration to the onset of jaundice Serum bilirubin level >20 mg/dL Coagulopathy ALT, alanine aminotransferase; AST, aspartate aminotransferase; ULN, upper limit of normal.

roacetylated (TFA) proteins were demonstrated, indicating previous exposure.17 Two types of postoperative liver injury have been asso­ ciated with halothane. A minor form is seen in 10% to 30% of patients, in whom mild asymptomatic elevations in serum alanine aminotransferase (ALT) levels develop between the first and tenth postoperative days; the risk of hepatotoxicity is higher after two or more exposures to halothane than with subsequent use of alternative agents such as enflurane, isoflurane, and desflurane. Evidence of immune activation is lacking in these patients,18 in whom the ALT elevations generally are rapidly reversible. The major form of halothane-induced hepatotoxicity is a rare, doseindependent, severe hepatic drug reaction with elements of immunoallergy and metabolic idiosyncrasy (see Table 87-1). After an initial exposure to halothane, the frequency of this form of toxicity is only approximately 1 per 10,000,19 but the rate increases to approximately 1 per 1000 after two or more exposures, especially when the anesthetic agent is readministered within a few weeks.1 Typically, zone 3 (centrilobular) hepatic necrosis is seen histologically.20 The case-fatality rate ranged from 14% to 71% in the pre-liver transplantation era.1

Risk Factors for Halothane Hepatitis

Host-related risk factors for halothane hepatitis are listed in Table 87-2. The reaction is rare in childhood11; patients younger than 10 years of age represent only about 3% of the total, and cases in persons younger than 30 years account for less than 10%.11,16 The disease tended to be more severe in persons older than 40 years of age. Two thirds of cases

Table 87-2  Risk Factors for Halothane Hepatitis Older age (>40 years) Female gender Two or more exposures (documented in 80%-90% of cases) Obesity Familial predisposition Induction of CYP2E1 by phenobarbital, alcohol, or isoniazid CYP2E1, cytochrome P450 2E1.

have been in women, and repeat exposure to halothane (especially within a few weeks or months) is documented in as many as 90% of cases.1 The time between exposures can be as long as 28 years.21 After repeat exposure, hepatitis is earlier in onset and more severe. Obesity is another risk factor, possibly because of storage of halothane in body fat. The induction of cytochrome P450 (CYP) enzymes (especially CYP2E1) that metabolize halothane to its toxic intermediate has been produced experimentally with phenobarbital, alcohol, and isoniazid; valproate inhibits and phenytoin has no specific effect on halothane hepatotoxicity.1 Inhibition of CYP2E1 by administration of a single dose of disulfiram has been suggested as a means of preventing halothane hepatitis—by inhibiting the production of the metabolite responsible for neoantigen formation.22 Familial predisposition to halothane-induced liver injury has been reported in closely related family members.23 Serum antibodies to volatile anesthetics have been found in pediatric anesthesiologists,17 who, like patients with halothane hepatitis, had higher levels of serum autoantibodies to CYP2E1 and to endoplasmic reticulum protein (ERp58) than those found in general anesthesiologists and control subjects who had never been exposed to inhalational anesthetics. The autoantibodies are not thought to have a role in pathogenesis.1

Pathology

In a study of 77 cases of halothane hepatitis reviewed by the Armed Forces Institute of Pathology,20 various degrees of liver injury were seen, depending on the severity of the reaction. Massive or submassive necrosis involving zone 3 was present in all autopsy specimens, whereas biopsy material revealed a broader range of injury—from spotty necrosis in about one third of cases to zone 3 necrosis in two thirds. The zone 3 injury is sharply demarcated, and the inflammatory response is less severe than in acute viral hepatitis.

Pathogenesis

Halothane injury occurs by one or more of three potential mechanisms: hypersensitivity, production of hepatotoxic metabolites, and hypoxia, in decreasing order of importance.1 Evidence for the role of hypersensitivity is found in the increased susceptibility and shortened latency after repeat exposure, the hallmark symptoms and signs of drug allergy (fever, rash, eosinophilia, and granuloma formation), and the detection of neoantigens and antibodies. Halothane oxidation yields trifluoroacetylchloride, which acts on hepatocyte proteins to produce neoantigens that are responsible for the major form of injury. By contrast, reductive pathways produce free radicals that can act as reactive metabolites that may have a role in causing minor injury. A unifying hypothesis set forth by Zimmerman1 suggests that halothane injury most likely is the result of immunologic enhancement of zone 3 necrosis produced by the reductive metabolite(s). Accordingly, the hepatotoxic potential of halothane depends on the susceptibility of the patient and on factors that promote production of hepatotoxic or immu-

Chapter 87  Liver Disease Caused by Anesthetics, Toxins, and Herbal Preparations Table 87-3  Hepatotoxic Anesthetics Other Than Halothane anesthetic

PERCENT METABOLIZED

INCIDENCE OF hepatotoxicity

CROSS-REACTIVITY WITH OTHER HALOALKANES

OTHER clinical FEATURES

Methoxyflurane Enflurane Isoflurane Desflurane

>30% 2% 0.2% <0.2%

Low 1 in 800,000 Rare Few reports

Yes Yes Yes Yes

Sevoflurane

Minimal

Rare

Uncertain

Nephrotoxicity Similar to halothane Similar to halothane Cardiac toxicity, malignant hyperthermia None reported

nogenic metabolites.1 A protective role for zinc pretreatment has been proposed based on studies in rats.24

Course and Outcome

Mortality rates for halothane hepatitis were high in early series; since then, successful treatment has been achieved with liver transplantation in many patients.13 When spontaneous recovery occurs, symptoms usually resolve within 5 to 14 days, and recovery is complete within several weeks.1 Immunosuppressive agents have only rarely been reported to improve the outcome.11 It is doubtful that halothane causes chronic hepatitis.1 Adverse prognostic factors include age older than 40 years, obesity, severe coagulopathy, serum bilirubin level greater than 20 mg/dL, and a shorter interval to onset of jaundice.1,16,19

OTHER ANESTHETIC AGENTS

The likelihood that individual haloalkane anesthetics will cause liver injury appears to be related to the extent to which they are metabolized by hepatic CYP enzymes: 20% to 30% for halothane, greater than 30% for methoxyflurane, 2% for enflurane, 1% for sevoflurane, and 0.2% or less for isoflurane and desflurane.12 Accordingly, the estimated frequency of hepatitis from the newer agents is much less than that for halothane (Table 87-3). Methoxyflurane caused hepatotoxicity and a high frequency of nephrotoxicity that led to its withdrawal.25 Enflurane caused a clinical syndrome similar to that for halothane, with the onset of fever within 3 days and jaundice in 3 to 19 days after anesthesia26,27; the estimated incidence of enflurane-induced liver injury was about 1 in 800,000 exposed patients.10 Despite its low rate of metabolism,12 numerous instances of isoflurane-associated liver injury have been reported.28-30 In one case, cross-sensitivity was suspected 22 years after an initial exposure to enflurane.29 TFA liver proteins have been detected in patients with suspected isoflurane liver toxicity.30 The newest haloalkane anesthetics, desflurane and sevoflurane, appear to be nearly free of adverse hepatic effects. Desflurane undergoes minimal biotransformation and was not associated with the development of TFA antibodies in exposed rats.12 Only isolated reports of liver injury in patients receiving desflurane anesthesia have been published.31 The biotransformation of sevoflurane also is minimal, and only rare reports have implicated this agent in postoperative hepatic dysfunction.32 Ether, nitrous oxide, and cyclopropane apparently are devoid of signi­ ficant hepatotoxic potential because of their lack of halogen moieties.1

JAUNDICE IN THE POSTOPERATIVE PERIOD

From 25% to 75% of patients undergoing surgery experience postoperative hepatic dysfunction, ranging from mild

Table 87-4  Causes of Postoperative Hepatic Dysfunction Hepatocellular Injury (predominant serum ALT elevation, with or without hyperbilirubinemia) Acute transfusion-associated hepatitis Hepatic allograft rejection Hepatic artery thrombosis Inhalational anesthetics—halothane, others Ischemic hepatitis (shock liver) Other drugs—antihypertensives (e.g., labetalol), heparin Unrecognized chronic liver disease—NASH, hepatitis C, other disorders Cholestatic Jaundice (elevated serum alkaline phosphatase ± ALT; direct hyperbilirubinemia) Acalculous cholecystitis Benign postoperative cholestasis Bile duct injury—post-cholecystectomy, post-liver transplantation Bile duct obstruction—gallstones, pancreatitis Cardiac bypass of prolonged duration Cholangitis Drugs—amoxicillin-clavulanate, chlorpromazine, erythromycin, telithromycin, trimethoprim-sulfamethoxazole, warfarin, others Hemobilia Microlithiasis (biliary sludge) Prolonged total parenteral nutrition Sepsis Indirect Hyperbilirubinemia (serum alkaline phosphatase and ALT often normal) Gilbert’s syndrome Hemolytic anemia (G6PD deficiency, other causes) Multiple transfusions Resorbing hematoma ALT, alanine aminotransferase; G6PD, glucose-6-phosphate dehydrogenase; NASH, nonalcoholic steatohepatitis.

elevations in liver biochemical tests to hepatic failure; jaundice has been reported in nearly 50% of patients with underlying cirrhosis in the postoperative period (see Chapter 20).33 Patients undergoing upper abdominal surgical procedures are at highest risk of postoperative liver dysfunction, as well as pancreatitis, cholecystitis, and bile duct injury because of impaired blood flow to the liver.33 Table 87-4 lists many causes of postoperative jaundice and hepatic dysfunction, broadly divided into hepatocellular injury, cholestasis, and indirect hyperbilirubinemia. Drugs that may cause hepatoxicity in this setting include antibiotics (e.g., erythromycin, telithromycin, amoxicillin-clavulanate, trimethoprim-sulfamethoxazole) and the halogenated anesthetics discussed earlier; most produce injury by hyper­ sensitivity mechanisms within one to two weeks of administration.1,2 Table 87-5 contrasts the features of halogenated anesthetic-induced hepatitis, ischemic hepatitis (shock liver) (see Chapter 83), and cholestatic injury (see Chapter 86) in the early postoperative period.

1449

1450

Section IX  Liver Table 87-5  Comparative Features of Causes of Acute Postoperative Liver Injury FEATURE Incidence Latency Fever, rash, eosinophilia Serum ALT (× ULN) Jaundice Histology Mortality Recovery time Risk factors   Age   Gender   Body weight   Hypotension

HALOALKANE ANESTHETIC TOXICITY

ISCHEMIC HEPATITIS

POSTOPERATIVE CHOLESTASIS

Rare 2-15 days Present 25-200× Common Zone 3 necrosis High Up to 12 weeks

Not uncommon Within 24 hours Absent Can exceed 200× Rare Coagulative necrosis, sinusoidal congestion Varies with diagnosis 10-12 days with supportive care

Common Few days Absent Minimal or normal Common (direct hyperbilirubinemia) Bile plugs, cholestasis Not from liver disease Variable, may be prolonged

Adults, age >40 years F > M 2 : 1 Obese May or may not be present

Any F=M Any Documented in 50%

Any F=M Any Absent

ALT, alanine aminotransferase; F, female; M, male; ULN, upper limit of normal.

Table 87-6  Phases of Illness after Ingestion of Various Hepatotoxins Toxin phase Phase I (1-24 hours) Onset of toxicity A,N,V,D Shock Neurologic symptoms Phase II (24-72 hours) Asymptomatic latent period Phase III (>72 hours) Jaundice Hepatic failure Renal failure Maximum AST and ALT (× ULN) Zonal necrosis Steatosis Case-fatality rate

ACETAMINOPHEN

PHOSPHORUS

AMANITA PHALLOIDES

CARBON TETRACHLORIDE

Immediate + − −

Immediate ++++ + +

Delayed 6-20 hr ++++ ± ±

Immediate + − −

+

±

+

+

+ + + 1000 3 − 5%-15%

+ + + <10-100 1 ++++ 25%-50%

+ + + 500 3 + 20%-25%

+ + + 500 3 + 20%-25%

ALT, serum alanine aminotransferase level; AST, serum aspartate aminotransferase level; A,N,V,D, anorexia, nausea, vomiting, diarrhea; ULN, upper limit of normal. Adapted from Zimmerman HJ: Hepatotoxicity. The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 1999.

CHEMICALS COMMERCIAL AND INDUSTRIAL CHEMICALS

Among tens of thousands of chemical compounds in commercial and industrial use, several hundred are listed as causing liver injury by the National Institute for Occupational Safety and Health (NIOSH), as published in their Pocket Guide to Chemical Hazards.34 The National Library of Medicine also maintains a database of chemical toxins in its Toxicology and Environmental Health Information Program (TEHIP).35 Toxic exposure to chemical agents occurs most often from inhalation or absorption by the skin and less often from absorption by the gastrointestinal tract after oral ingestion or through a parenteral route. Because most chemical toxins are lipid soluble, when absorbed they can easily cross biological membranes to reach their target organ(s), including the liver.3,4 Hepatotoxic chemical exposure (as with carbon tetrachloride and phosphorus) usually results in an acute cytotoxic injury that typically consists of three distinct phases, similar to those observed after an acetaminophen overdose or ingestion of toxic mushrooms (Table 87-6).1,3

Less commonly, acute cholestatic injury may occur.36 Many chemicals (e.g., vinyl chloride) also are carcinogenic, and hepatic malignancies have been part of the clinicopathologic spectrum of chemical injury (Table 87-7).37 Although liver injury is the dominant toxicity for some agents, hepatic damage may be only one facet of more generalized toxicity for other agents.3

Carbon Tetrachloride and Other Chlorinated Aliphatic Hydrocarbons

Carbon tetrachloride (CCl4) is a classic example of a zone 3 hepatotoxin that causes necrosis leading to hepatic failure (see Table 87-6). Injury is mediated by its metabolism to a toxic trichloromethyl radical catalyzed by CYP2E1.7,38 Alcohol potentiates the injury through induction of this cytochrome.1 Most cases have been the result of industrial or domestic accidents, such as inhalation of CCl4containing dry cleaning fluids that are used as household reagents or ingestion of these compounds by alcoholics who mistake them for potable beverages.1 At the cellular level, direct damage to cellular membranes results in leakage of intracellular enzymes and electrolytes, leading in turn to

Chapter 87  Liver Disease Caused by Anesthetics, Toxins, and Herbal Preparations Table 87-7  Clinicopathologic Spectrum of Chemical Hepatotoxins

Table 87-8  Relative Hepatotoxicity of Haloalkane Compounds

Acute Injury Necrosis Carbon tetrachloride and other haloalkanes Cocaine, “ecstasy,” phencyclidine Haloaromatics, nitroaromatics, nitroaliphatics Hydrochlorofluorocarbons Phosphorus, iron, copper salts, inorganic arsenic Microvesicular Steatosis Boric acid Chlordecone Cocaine Dimethylformamide Hydrazine Hypoglycin Thallium Toluene, xylene Cholestasis Alpha-naphthylisocyanate Aniline—rapeseed oil Dinitrophenol Methylene dianiline Paraquat Subacute Injury Necrosis Trinitrotoluene Sinusoidal Obstruction Syndrome Pyrrolizidine alkaloids, arsenic, thorium dioxide Toxic Cirrhosis Hexachlorobenzene, polychlorinated biphenyls Tetrachlorethane Peliosis Hepatis Dioxin Chronic Injury Cirrhosis Chloroaliphatics, trinitrotoluene, arsenic, pyrrolizidine alkaloids Hepatoportal sclerosis Arsenic, vinyl chloride Neoplasia Hepatocellular carcinoma Arsenic, aflatoxins, thorium dioxide Angiosarcoma Vinyl chloride, thorium dioxide, arsenic Hemangioendothelioma Arsenic

COMPOUND

Data from references 1, 3, 4, 36, and 37.

calcium shifts and lipid peroxidation.8 Hepatic steatosis develops as a result of triglyceride accumulation caused by haloalkylation-dependent inhibition of lipoprotein micelle transport out of the hepatocyte.38 CCl4 is more toxic than other haloalkanes and haloalkenes because toxicity correlates inversely with the level of bond dissociation energy, number of halogen atoms, and chain length (Table 87-8).1,38 In older series, complete clinical and histologic recovery from CCl4-induced liver damage was the rule with modest exposures, but supervening acute tubular necrosis and gastrointestinal hemorrhage were associated with a casefatality rate of 10% to 25%.1,3 Activation of endonucleases, causing chromosomal damage and mutations, may result in carcinogenesis.38 Chloroform remains an important experimental hepatotoxin, although its use as an anesthetic has long been abandoned (see later).1,3 Hepatic injury, including chronic hepatitis, has been reported with 1,1,1-trichloroethane.39 Hydrochlorofluorocarbons (HCFCs) have been associated with liver injury in several industrial workers exposed to dichlorotrifluoroethane (HCFC 123) and 1-chlorotetraflu-

RELATIVE TOXICITY

Carbon tetrachloride Tetrachlorethane Chloroform Trichloroethylene 1,1,2-Trichloroethane Tetrachloroethylene 1,1,1-Trichloroethane Dichloromethane Dibromomethane Methylchloride

++++ ++++ ++ + to ++ + to ++ + + ± ± −

Scale from ++++, maximal injury to −, trivial or no injury. Data from references 1 and 3.

oroethane (HCFC 124), both of which are metabolized to reactive trifluoroacetyl halide intermediates similar to those implicated in halothane toxicity.40 Zone 3 necrosis is present on liver biopsy specimens, and autoantibodies against CYP2E1 or P58 are detected in the serum of many affected persons. As with halothane, liver toxicity may be potentiated by ethanol.41

Vinyl Chloride and Other Chlorinated Ethylenes

In the past, exposure to vinyl chloride monomer (VCM), or monochloroethylene, occurred in polymerization plants where vinyl chloride was heated to form polyvinyl chloride (PVC) in the manufacture of plastics; the toxic gas containing VCM was inhaled in this process.1 Vinyl chloride is ubiquitous in the environment and has been estimated by the Environmental Protection Agency to exist in at least 10% of toxic waste sites.4 Although PVC appears to be nontoxic, long-term exposure to VCM has led to chronic liver injury, including nodular subcapsular fibrosis, sinusoidal dilatation, peliosis hepatis, and periportal fibrosis associated with portal hypertension.1,3 Nonalcoholic fatty liver disease, including lipogranulomas, has been described in more than 50% of nonobese chemical workers with high exposure levels to VCM; some of these workers continued to have nonalcoholic steatohepatitis up to six years later.42 Vinyl chloride is carcinogenic. Angiosarcoma develops after a mean latency of 25 years after exposure; the risk is related to the duration and extent of contact.43 Alcohol appears to enhance the hepatocarcinogenicity of vinyl chloride, in rodents and possibly in humans, by inducing CYP2E1, which converts vinyl chloride to a toxic or carcinogenic metabolite (e.g., 2-chloroethylene oxide).1 A history of vinyl chloride exposure was found in 15% to 25% of all cases of hepatic angiosarcoma reported in the late 1970s,3 and strict hygienic measures instituted in 1974 have resulted in a marked decrease in the frequency of angiosarcoma since then; however, persons with the highest exposure still have a four-fold increased risk of developing periportal hepatic fibrosis, which may be a precursor of angiosarcoma.44 Persons previously exposed to vinyl chloride should undergo regular clinical examination for early detection of liver tumors, and those with known chronic liver disease or high levels of exposure should undergo regular hepatic imaging. Persons who work in PVC plants should undergo regular monitoring of liver biochemical test levels, and those with persistent abnormalities should be removed from workplace exposure.44 High serum levels of hyaluronic acid were correlated with the development of angiocarcinoma

1451

1452

Section IX  Liver in 26 of 82 workers occupationally exposed to PVC in Kentucky.45

Nonhalogenated Organic Compounds

Benzene has been associated with minor hepatic injury in animals. Toluene led to steatosis and necrosis in a “glue sniffer”46 and has been associated with acute fatty liver of pregnancy; it caused elevations in serum gamma glutamyl transpeptidase levels after industrial exposure. Xylene can cause mild hepatic steatosis, and styrene (vinyl benzene) has led to elevated serum aminotransferase levels after prolonged exposure.

Trinitrotoluene and Other Nitroaromatic Compounds

Trinitrotoluene (TNT), or nitroglycerin, was first observed to be hepatotoxic during World War I, when severe acute and subacute hepatic necrosis developed in munitions workers in England, Germany, and the United States; the case-fatality rate was more than 25%.1,3 The frequency of hepatotoxicity during World War II was lower, with approximately 1 in 500 workers affected, but the estimated frequencies of methemoglobinemia and aplastic anemia were 50 times higher.3 Subacute hepatic necrosis followed two to four months of regular exposure to TNT. Percutaneous absorption was the major source of exposure. In some patients, rapidly progressive liver failure and death occurred within days to months, with massive hepatic necrosis at autopsy. In others, the subacute injury progressed over several months to micronodular cirrhosis and portal hypertension. The relatively low incidence of injury suggests that formation of a toxic metabolite was involved.1 Nitrobenzene and dinitrobenzene also were observed to be hepatotoxic during World War I. As with TNT, excessive exposure led to methemoglobinemia.

Nitroaliphatic Compounds

Nitromethane, nitroethane, and nitropropane cause variable degrees of hepatic injury. 2-Nitropropane (2-NP) has caused fatal massive hepatic necrosis after occupational exposure as a solvent, fuel additive, varnish remover, and rocket propellant. Toxic hepatitis associated with the chronic inhalation of propane and butane also has been reported.47

Polychlorinated Biphenyls and Other Halogenated Aromatic Compounds

Polychlorinated biphenyls (PCBs) are mixtures of trichloro-, tetrachloro-, pentachloro-, and hexachloro-derivatives of biphenyls, naphthalenes, and triphenyls that are used in the manufacture of electrical transformers, condensers, capacitors, insulating materials for electrical cables, and industrial fluids. Acute and chronic hepatotoxicity from PCB exposure seen during World War II resembled that caused by TNT.3,4 Inhalation of toxic fumes released by the melting of PCBs and chloronaphthalene mixtures during soldering of electrical materials was the most common means of exposure.1 The severity of liver injury correlated with the number of chlorine molecules.3 Liver damage appeared as early as seven weeks after ongoing exposure and was accompanied by anorexia, nausea, and edema of the face and hands. Acne-like skin lesions (chloracne) usually preceded hepatic injury. Once jaundice appeared, death occurred within two weeks in fulminant cases, which were characterized by massive necrosis (so-called acute yellow atrophy), or after one to three months in the subacute form. Cirrhosis developed in some persons who survived the acute injury.1 Polybrominated biphenyls (PBBs) appear to be even more toxic than PCBs. Consumption of milk and meat from live-

stock given feed mistakenly contaminated by a PBB has led to hepatomegaly and minor elevations in liver enzyme levels in exposed persons.3

Miscellaneous Chemical Compounds

Dimethylformamide is a solvent used in the synthetic resin and leather industries that causes dose-related massive hepatic necrosis in animals48 and is capable of producing focal hepatic necrosis and microvesicular steatosis in humans.3 Most persons exposed for more than one year have symptomatic disease that slowly resolves when they are removed from the workplace. Disulfiram-like symptoms can occur.49 Alcohol use, hepatitis B virus infection, and a high body mass index are risk factors.50 Hydrazine and its derivatives used in jet and rocket fuel cells are also experimental hepatotoxins and carcinogens and have been reported to cause hepatic steatosis in animals1 and reversible injury in humans after inhalation.51 Bromoalkanes and iodoalkanes, used in insecticides and aircraft fuels, have rarely caused hepatic injury.3 Ethylene dibromide (dibromoethane) has led to zone 3 hepatic necrosis after ingestion in attempted suicide or as fatal hepatotoxicity associated with nephrotoxicity and cardiotoxicity following occupational exposure or inadvertent poisoning.52

PESTICIDES

Although exposure to insecticides, herbicides, and other pesticides is common, acute liver injury resulting from these compounds, many of which are chlorinated hydrocarbons, is rare.1,3 Evidence that dichlorodiphenyltrichloroethane (DDT) and other organochlorines (aldrin, amitrole, chlordane, dieldrin, lindane, mirex) lead to liver damage or carcinogenicity is limited.1 Agent Orange (2.4-dichlorophenoxyacetic acid), the defoliant widely used in Vietnam, has been reported to cause acute hepatitis after chronic exposure; however, contaminating dioxins have been suggested to be responsible for the toxic effects.53,54 Moreover, chronic liver injury among Vietnam veterans is more likely to have been related to viral infections or alcohol than to Agent Orange,55 and hepatocarcinogenesis is more likely to have been related to chronic hepatitis B infection.56 Ingestion of or dermal exposure to dichloride dimethyl­ dipyridilium (paraquat) has been implicated in several instances of hepatotoxicity as a result of attempted suicide and homicide.57 Patients may present with severe vomiting and profuse diarrhea leading to hypokalemia and often have evidence of oral, pharyngeal, and esophageal caustic injury after ingestion. Death results from a combination of renal, respiratory, cardiac, and hepatic failure; mortality rates are as high as 70%, and death often occurs within the first 48 hours. Treatment with charcoal hemoperfusion in conjunction with cyclophosphamide, dexamethasone, furosemide, and vitamins B and C—the so-called Caribbean scheme—has been attempted, but persons who ingest more than 45 mL are likely to die with or without this treatment.57 Histo­ pathologic changes include zone 3 necrosis followed by injury to small- and medium-sized interlobular bile ducts.58 Chlordecone (Kepone) has been shown to impair biliary excretion and lipid transport and storage,59 but neurologic toxicity appears to dominate the clinical injury. Occupational exposure has led to hepatic steatosis and elevated serum aminotransferase levels. Trivial hepatic enzyme abnormalities have been seen in persons heavily exposed to chloretone.3 Hexachlorobenzene in contaminated grain has been associated with an epidemic of porphyria cutanea tarda and liver injury.3

Chapter 87  Liver Disease Caused by Anesthetics, Toxins, and Herbal Preparations Inorganic arsenic has long been used as a homicidal or suicidal agent, and toxic exposure in the past also followed ingestion of Fowler’s solution (arsenic trioxide) used as a treatment for psoriasis and asthma.1,3 Other sources of exposure are contaminated ground and well water and homemade alcohol. Doses greater than 3 g can cause death in one to three days; hepatic injury generally is overshadowed by gastrointestinal, neurologic, and vascular effects,1 leading ultimately to central nervous system depression and vascular collapse. A lesion resembling hepatic sinusoidal obstruction syndrome (hepatic veno-occlusive disease) can develop,3 and noncirrhotic portal hypertension developed in more than 90% of 248 patients who consumed contaminated drinking water for up to 15 years.60 Occupational exposure to arsenic is still observed among vineyard workers, farmers, and gold miners,61 although its use as an insecticide has been curtailed since the 1940s. Lumber treated with chromated copper arsenate as a pre­ servative may be an additional source of exposure.62 The clinical syndrome associated with arsenicosis includes skin lesions (blackfoot disease), anemia, diabetes mellitus, hearing loss, neurobehavioral disorders, and cardiovascular diseases, in addition to benign and malignant liver disease.63 Chronic hepatic injury, including cirrhosis and noncirrhotic portal hypertension, may be a precursor to hepatic neoplasms, such as angiosarcomas, hemangioendotheliomas, and hepatocellular carcinomas, after exposure of more than 10 years.64 Treatment with thiol chelators has had variable success in cases of prolonged exposure, and coadministration of antioxidants, such as vitamins C and E, may be of added benefit.65

METALS IRON

Most of the 5000 cases of accidental iron poisoning in the United States each year occur in young children who mistake iron supplements for candy.1 The severity of injury correlates with the dose ingested3; ingestion of less than 20 mg/kg of elemental iron is unlikely to produce serious toxicity, whereas doses of more than 200 mg/kg can be fatal.1 Severe injury has been seen only with serum iron concentrations above 700  mg/dl measured within the first 12 hours after ingestion.66 Iron, per se, is not hepatotoxic, but ferric and ferrous ions can act through free radicals and lipid peroxidation to cause membrane disruption and necrosis.67 Clinically evident liver injury is uncommon, but zone 1 necrosis occurs in the most severe cases.1 Clinical illness is characterized by sequential phases of gastrointestinal injury, subsidence of symptoms, and overt hepatotoxicity accompanied by renal failure. Deferiprone, an oral iron chelator, was implicated in causing worsening hepatic fibrosis in a long-term study of patients with thalassemia68; however, these findings were not confirmed by subsequent histopathologic analysis,69 and injury from this agent appears unlikely.

PHOSPHORUS

Poisoning by white phosphorus has been rare since its use in firecrackers and matches was outlawed in the midtwentieth century.3 Cases reported since then usually have been the result of ingestion of rat or roach poison.1 Shortly after ingestion, vomiting, gastrointestinal bleeding, convulsions, shock, and death occur within 24 hours. Phosphorescence of the vomitus and stools and a typical garlic-like odor on the breath are characteristic, when present. The

pre­dominant hepatic lesion is steatosis and necrosis, most prominent in the periportal region. Serum aminotransferase levels generally are no higher than 10 times the upper limit of normal.1

COPPER SALTS

Acute poisoning by copper leads to a syndrome resembling iron toxicity. Ingestion of toxic amounts (1 to 10 mg) usually is seen with suicidal intent, especially on the Indian subcontinent.3,67 Vomiting, diarrhea, and abdominal pain accompanied by a metallic taste are seen during the first few hours after ingestion. Gastrointestinal tract erosions, renal tubular necrosis, and rhabdomyolysis often accompany zone 3 hepatic necrosis by the second or third day. Jaundice results from both hepatic injury and acute hemolysis caused by high blood copper levels.3 The mortality rate is 15%, with early deaths resulting from shock and circulatory collapse and late deaths resulting from hepatic and renal failure.1

THORIUM DIOXIDE

Thorium dioxide (Thorotrast) was used as an intravenous contrast medium for radiographic procedures in the first half of the twentieth century; more than 50,000 persons may have been exposed.1 Thorotrast was subsequently found to cause hepatic angiosarcomas after latency periods of 20 to 40 years. As with arsenic, reports of hepatic sinusoidal obstruction syndrome and a Budd-Chiari-like syndrome of portal hypertension also have appeared. Given the extraordinarily long half-life (hundreds of years) of the compound, which is a radioactive alpha emitter, exposed persons remain at risk for the development of leukemia, in addition to hepatocellular cancer.70 Histologically, thorium dioxide is found in Kupffer cells and macrophages as dark brown refractile granules, the identity of which can be confirmed by spectrographic analysis.38

OTHER METALS

Although cadmium produces hepatic necrosis and cirrhosis in laboratory animals,71 evidence is lacking that exposure to cadmium causes important human injury.1 Several metals are associated with apoptosis, which might explain their potential for hepatotoxicity.72 Beryllium has led to midzonal liver necrosis as a result of phagocytosis of insoluble beryllium phosphate by Kupffer cells.1 Chronic industrial exposure (usually by inhalation of high concentrations of oxide or phosphorus mixtures) is associated with the formation of hepatic (and pulmonary) granulomas.3 Therapy with chelating agents and antioxidants has been used in animal models of beryllium toxicity.73 Lead hepatotoxicity may be seen as part of the larger symptom complex of abdominal pain, constipation, and encephalo­pathy that occurs with chronic ingestion or environmental exposure.74

ADULTERATED COOKING OILS AND CONTAMINATED FOODS A number of contaminated foodstuffs and cooking oils have been associated with epidemics of hepatotoxicity, now largely of historical interest only. The Spanish toxic oil syndrome occurred in 1981, after exposure of up to 100,000 Spaniards to rapeseed cooking oil that was contaminated by anilines and acetanilides. Nearly 20,000 persons became ill, many with hepatic injury and jaundice. Approximately 2500 died.75 Among 332 patients followed for up to eight years, hepatic injury devel-

1453

1454

Section IX  Liver oped in 43%, usually at the onset of a multisystem disease. A mixed cholestatic-hepatocellular injury pattern was seen, with jaundice or hepatomegaly in fewer than 20%. After an eight-year follow-up, liver disease persisted in only four patients.75 Epping jaundice refers to an epidemic of toxic liver injury that occurred in Epping, England, in 1965.3,36,76 The outbreak involved 84 persons who had eaten bread contaminated with methylenedianiline that had spilled onto the floor of a van carrying flour. The clinical syndrome consisted of abdominal pain, fever, and chills, followed by cholestatic jaundice resembling that seen with biliary obstruction; eosinophilia was seen in about one half of patients. Liver biopsy specimens revealed Kupffer cell hyperplasia with portal inflammation but little or no necrosis.3 Most persons recovered in four to six weeks, with jaundice lasting up to four months in a few. The mechanism of injury was thought to be a chemically induced cholangitis, possibly as a result of a hypersensi­ tivity reaction. Cholangiocarcinoma later developed in one patient.76 Yusho oil disease in western Japan, and a related epidemic referred to as yu-cheng in Taiwan, involved nearly 2000 persons who had eaten rice prepared in oil contaminated by PCBs, dioxins, and polychlorinated dibenzofurans in 1968. The disease was characterized by chloracne, skin hyperpigmentation, eyelid edema, and neuropathy, with jaundice reported in approximately 10% of patients.77 Exposed persons still harbored high levels of these agents nearly three decades after the outbreak.77 Hexachlorobenzene contamination of wheat in the 1950s led to an epidemic of toxic porphyria cutanea tarda and severe liver disease involving more than 3000 Turkish Kurds, with a mortality rate that exceeded 10%. This fungicide had been added to seed grain that was used for food during a famine.3,36

DRUGS OF ABUSE COCAINE

Cocaine is a dose-dependent hepatotoxin.3 Acute cocaine intoxication affects the liver in 60% of patients,73 and many affected persons have markedly elevated serum ALT levels (greater than 1000 U/L). Associated features include rhabdomyolysis, hypotension, hyperpyrexia, disseminated intravascular coagulation, and renal failure. Hepatic injury probably is the result of toxic metabolites (e.g., norcocaine nitroxide) formed by the CYP system, specifically CYP2E1 and CYP2A,79 and enhanced hepatotoxicity is seen in persons who regularly consume alcohol.3 In animals, pretreatment with N-acetylcysteine decreases the risk of cocaine hepatotoxicity,80 although the usefulness of Nacetylcysteine for treating human cocaine-induced hepatic injury has not been determined.

OTHERS

“Ecstasy” (3,4-methylenedioxymethamphetamine) is a euphorigenic and psychedelic amphetamine derivative that can lead to hepatic necrosis as part of a heat stroke–like syndrome resulting from exhaustive dancing in hot nightclubs (“raves”).81 The injury can be fatal and has necessitated liver transplantation in some instances.82,83 The role of CYP enzymes in the toxicity of this and other so-called “designer drugs” may relate to specific genetic polymorphisms of CYP2D6 or other cytochromes.84 Phencyclidine (“angel dust”) is another stimulant that can lead to hepatic injury as part of a syndrome of malignant hyperthermia that produces zone 3 hepatic necrosis, congestion, and collapse, with high serum aspartate aminotransferase (AST) and ALT levels reminiscent of ischemic hepatitis.85

BOTANICAL AND ENVIRONMENTAL HEPATOTOXINS Examples of hepatotoxic mushrooms, fruits, and other foodstuffs, including grains and nuts contaminated by fungal mycotoxins or other potentially injurious compounds, are listed in Table 87-9.

MUSHROOMS

Poisonous varieties of mushrooms number approximately 100 among the more than 5000 species, and more than 8000 mushroom poisonings were reported in the United States in 2001.86 Greater than 90% of cases of fatal poisoning are caused by Amanita phylloides (death cap) or Amanita verna (destroying angel), found in the Pacific Northwest and eastern United States.87 A fatal outcome can follow ingestion of a single 50-g (2-oz) mushroom; the toxin is one of the most potent and lethal in nature.88 Alpha-amatoxin is thermostable, can resist drying for years, and is not inactivated by cooking. Rapidly absorbed through the gastrointestinal tract, the amatoxin reaches hepatocytes through the enterohepatic circulation and inhibits production of messenger RNA and protein synthesis, leading in turn to cell necrosis. A second toxin, phalloidin, is responsible for the severe gastroenteritis that precedes hepatic and central nervous system injury.89 Phalloidin disrupts cell membranes by interfering with polymerization of actin. A latent period of 6 to 20 hours after ingestion of a mushroom precedes the first symptoms of intense abdominal pain, vomiting, and diarrhea. Hepatocellular jaundice and renal failure occur over the next 24 to 48 hours and are followed by confusion, delirium, convulsions, and eventually coma by 72 hours.1,89 The characteristic hepatic lesion is steatosis and zone 3 hepatic necrosis, with nucleolar inclusions seen on electron microscopy.3 In a case series of eight patients,90 the mean serum AST level was 5488 U/L (range, 1486 to 12,340), ALT 7618

Table 87-9  Botanical and Environmental Hepatotoxins AGENT

TOXIC COMPONENT

TYPE OF INJURY

COMMENT

Ackee fruit Aspergillus flavus Aspergillus tamari Cycasin Toxic mushrooms

Hypoglycin Aflatoxin B1 Cyclopiazonic acid Methylazoxymethanol Fulminant hepatic failure

Microvesicular steatosis Acute hepatitis, portal hypertension Acute hepatitis Acute hepatitis Fulminant hepatic failure

Jamaican vomiting sickness Hepatocarcinogenic — — Resembles APAP injury

APAP, N-acetyl-p-aminophenol (acetaminophen).

Chapter 87  Liver Disease Caused by Anesthetics, Toxins, and Herbal Preparations (range, 3065 to 15,210), and bilirubin 10.5 mg/dL (range, 1.8 to 52), with peak levels on days four and five. Acute kidney injury requiring dialysis developed in one patient, and three exhibited encephalopathy. Mortality rates traditionally have been high, especially when the serum ALT level exceeds 1000 U/L, and emergency liver transplantation often is required90; however, some patients survive with conservative management, which includes nasogastric lavage with activated charcoal, intravenous penicillin G, N-acetylcysteine (using a standard acetaminophen [Nacetyl-p-aminophenol, or APAP] protocol of a loading dose of 140 mg/kg orally followed by 15 additional oral doses of 70 mg/kg or one of two intravenous protocols: for presentation within 10 hours of exposure, a loading dose of 150 mg/ kg infused over 1 hour, followed by 50 mg/kg over 4 hours, followed by 100 mg/kg over 16 hours; for presentation more than 10 hours after exposure, a loading dose of 140 mg/kg over 1 hour, followed by 70 mg/kg every 4 hours for at least 12 doses), and milk thistle (Silybum marianum).89 The use of these therapeutic modalities is not always effective, and in a large review of 2108 cases over a 20-year period in the United States and Europe,91 penicillin G, either alone or in combination with other therapy, demonstrated limited benefit. Similarly, no role for glucocorticoids was found. The utility of plasmapheresis or hemoperfusion is unproved.

OTHER FOODSTUFFS

The unripe fruit of the ackee tree (Blighia sapida), native to Jamaica, contains a hepatotoxin, hypoglycin A, that produces a clinical syndrome of gastrointestinal distress and microvesicular steatosis known as Jamaican vomiting sickness, which resembles Reye’s syndrome (see Chapter 86).3,36,92 Cholestatic jaundice has been described after chronic ingestion.93 Cycasin is a potent hepatotoxin and hepatocarcinogen found in the fruit of the cycad tree (Cycas circinalis, Cycas revoluta). A small epidemic of acute hepatic injury attributable to the ingestion of cycad nuts was reported from Japan. The purported toxin is methylazoxymethanol, which normally is eliminated or rendered inactive in preparing the nuts before ingestion.3 Aflatoxins are a family of mycotoxins found in Asper­ gillus flavus and related fungi that are ubiquitous in tropical and subtropical regions. They contaminate peanuts, cashews, soybeans, and grains stored under warm, moist conditions and are well-known hepatotoxins and hepatocarcinogens.1,3 Aflatoxin B1, a potent inhibitor of RNA synthesis, is the most hepatotoxic member of the family. Reactive metabolites are formed by the CYP system, and malnutrition is a possible potentiating factor (perhaps because of the depletion of glutathione). When consumed in large quantities, aflatoxin B1 is responsible for a clinical syndrome characterized by fever, malaise, anorexia, and vomiting followed by jaundice. Portal hypertension with splenomegaly and ascites may develop over the next few weeks. In large epidemics, mortality rates have approached 25% and correlate with the dose ingested.3 Zone 3 hepatic necrosis without inflammation is the characteristic lesion. Other histologic findings include cholestasis, microvesicular steatosis, and bile duct proliferation.3 The risk of hepatocellular carcinoma (HCC) correlates with the amount of aflatoxin consumed, especially in subSaharan Africa and eastern China, where wheat often exceeds rice as a staple in the diet (see Chapter 94).3 Alcohol and possibly exposure to DDT may play an enhancing role in hepatocarcinogenesis.94 An even more important cofactor may be the hepatitis B virus.95 The frequency of a mutation

in the TP53 tumor suppressor gene correlates with the development of HCC in these regions, but this mutation is rare in HCC from Western countries (see Chapter 94).95

VITAMINS AND HERBAL PREPARATIONS The use of vitamins, dietary supplements, and herbal and nonproprietary remedies is an important aspect of CAM. This field continues to grow in the United States and around the world (see Chapter 127).96 In the United States, alternative medicines were used by 34% of the population in 1990 and 42% in 1997; nearly 20% of the population took complementary medicines at the same time as conventional prescriptions.97 The use of herbal products is even more popular among patients with chronic liver disease,98,99 despite the absence of controlled clinical trials to assess safety and efficacy in this setting.100 Many so-called health foods, dietary and weight loss supplements, and herbal products are potent hepatotoxins that have led to acute liver failure and the need for emergency liver transplan­ tation.5,6,8,101,102 Dietary supplements containing anabolic androgenic steroids may cause severe cholestatic liver injury.103,104

HYPERVITAMINOSIS A

Vitamin A (retinol) is a dose- and duration-dependent hepatotoxin capable of causing injury ranging from asymptomatic elevations in serum aminotransferase levels with minor hepatic histologic changes to perisinusoidal fibrosis leading to noncirrhotic portal hypertension and, in some cases, cirrhosis.105 Approximately one third of the U.S. population is estimated to take vitamin supplements containing vitamin A, with as many as 3% of products providing a daily dose of at least 25,000 IU. Hypervitaminosis A usually is the result of self-ingestion, rather than intentional overdose, and all age groups are represented.106 The average daily dose of vitamin A in reported cases of liver disease has been nearly 100,000 IU over an average duration of 7.2 years, for a mean cumulative dose of 229 million IU, but liver injury has been described with daily doses of 10,000 to 45,000 IU,107 and cirrhosis has occurred after a daily intake of 25,000 IU for at least 6 years.105,107 Long-term use of low-dose vitamin A supplements (250 to 5000 retinol equivalents per day) does not appear to be toxic.108 Because of the long half-life of vitamin A in the liver (50 days to 1 year),107,109 the fibrotic process may continue due to the slow release of hepatic vitamin A stores despite discontinuation of oral intake of the vitamin. Genetic factors may play a role, and apparent familial hypervitaminosis A occurred in four siblings who ingested large doses as treatment for congenital ichthyosis.110 Vitamin A toxicity has been reported in native Alaskans who ingest large amounts of fresh polar bear liver.105 Water-soluble, emulsified, and solid formulations of vitamin A are up to 10 times as toxic as oil-based preparations.111 Hepatotoxicity from vitamin A has been attributed to activation of hepatic stellate cells, the body’s principal storage site of the vitamin. Resulting hyperplasia and hypertrophy produce sinusoidal obstruction and increased collagen synthesis, leading in turn to portal hypertension.112 Rare cases of peliosis hepatis also have been attributed to hypervitaminosis A. Beta carotene, a precursor of vitamin A, is involved in the neoplastic transformation of squamous cell lung and tracheal tissues, especially in smokers who consume alcohol.113 Liver biopsy specimens show increased storage of vitamin A, seen as characteristic greenish autofluorescence on irra-

1455

1456

Section IX  Liver diation with ultraviolet light.105 The excess vitamin A is stored initially in stellate cells that lie in the space of Disse and become hyperplastic and hypertrophic. The enlarged clear stellate cells compress the hepatic sinusoids, giving rise to a “Swiss cheese,” or honeycombed, appearance.105 Hepatocellular injury usually is minor, with microvesicular steatosis and focal degeneration and without significant necrosis or inflammation. Hepatic fibrosis in a perisinusoidal distribution can arise from activated stellate cells that transform into myofibroblasts. In one series,105 cirrhosis was present in 59%, chronic hepatitis in 34%, microvesicular steatosis in 21%, perisinusoidal fibrosis in 14%, and peliosis in 3% of cases. Hypervitaminosis A also can involve the skin and central nervous system.1 Hepatomegaly is common, and in severe cases, splenomegaly, ascites, and esophageal variceal bleeding may be features.1,5 Liver biochemical test abnormalities, present in two thirds of cases, are nonspecific, with only modest elevations in serum aminotransferase and alkaline phosphatase levels. The diagnosis of vitamin A toxicity rests on a dietary and medication history and clinical suspicion. Plasma vitamin A levels may be normal, and the diagnosis is supported by the demonstration of increased hepatic stores of vitamin A and characteristic histologic findings.114 The diagnosis may be delayed for several years if hepatotoxicity is not recognized or is misdiagnosed.105,107 Symptoms resolve and liver enzymes normalize gradually after discontinuation of vitamin A ingestion in less severe cases, but deterioration may continue in cases of severe intoxication, particularly when cirrhosis is already present.107 Features of liver failure and cirrhosis at the time of diagnosis indicate a poor prognosis, and liver trans­ plantation may be required.1 Alcohol can potentiate hepatotoxicity and should be avoided. Vitamin A supplements generally should be avoided in other types of liver disease because of possible accentuation of hepatic injury and fibrosis.113 Severe liver injury rarely has been reported with the use of acitretin, a vitamin A metabolite.115

HERBAL REMEDIES AND NUTRITIONAL SUPPLEMENTS

The increasing use of CAM is well described in patients with liver disease.96-99 Silymarin (Silybum marianum, milk thistle) is the most commonly used herbal preparation among these patients, and while it appears to be quite safe, there is an increasing number of reports of hepatotoxicity from several classes of herbal and weight reduction agents that has paralleled the rise in CAM therapies.5,6,37,102 Warnings have been issued for several agents, and in a few instances, the U.S. Food and Drug Administration (FDA) and other health authorities have requested their removal from the marketplace (e.g., kava kava, ephedra [ma huang], Lipokinetix, and Hydroxycut in the United States [see FDA .gov/Food/ResourcesforYou/Consumers/ucm08542.htm] and germander in France). Any patient with liver disease should be questioned about the ingestion of herbal remedies. Estes and colleagues101 documented the use of several commonly promoted herbal agents (including Lipokinetix, skullcap, ma huang, chaparral, and kava kava) in 50% of their patients with acute liver failure over a two-year period. The agents were used by an equal number of men and women, for several months to 15 years. Six of 10 patients underwent emergency liver transplantation (with 2 deaths), 3 died before transplantation, and only 1 patient recovered spontaneously. Table 87-10 lists various herbal remedies according to their toxic constituent and the nature of the associated liver disease.

Pyrrolizidine Alkaloids

Pyrrolizidine alkaloids are found in approximately 3% of all flowering plant species throughout the world, and ingestion of such plants, often as medicinal teas or in other formulations, can produce acute and chronic liver disease, including sinusoidal obstruction syndrome (SOS), in humans and livestock.116 SOS was first reported in the 1950s as a disease of Jamaican children, manifesting with acute abdominal distention, marked hepatomegaly, and ascites, a triad that resembled Budd-Chiari syndrome (see Chapter 83).3 The disease was linked to consumption of “bush tea,” made largely from plants of Senecio, Heliotropium, and Crotalaria species and taken as a folk remedy for acute childhood illnesses that are characterized histologically by centrilobular hepatic congestion with occlusion of the hepatic venules leading to congestive cirrhosis. Comfrey (Symphytum officinale) remains commercially available even though it is a dose-dependent hepatotoxin.6,116 In Afghanistan, ingestion of pyrrolizidine alkaloidcontaminated grains and bread led to a large epidemic of SOS, affecting 8000 persons and innumerable sheep.3 Hepatotoxic pyrrolizidine alkaloids are cyclic diesters, and some forms (e.g., fulvine, monocrotaline) cause both liver and lung injury.116 The mechanism of injury is postulated to be impairment of nucleic acid synthesis by reactive metabolites of pyrrolizidine alkaloids generated by hepatic microsomes, leading, in turn, to progressive loss of sinusoid cells and sinusoidal hemorrhage, as well as injury to the endothelium of the terminal hepatic venule, with deposition of fibrin.116,117 SOS causes acute, subacute, and chronic injury. The acute form is characterized by zone 3 necrosis and sinusoidal dilatation, leading to a Budd-Chiari-like syndrome with abdominal pain and the rapid onset of ascites within three to six weeks of ingestion. In Jamaica, the course was rapidly fatal in 15% to 20% of affected persons. Approximately one half of the patients with the acute form recovered spontaneously; transition to a more chronic form of injury occurred in the remainder.1,3 In the subacute and chronic forms, central fibrosis and bridging between central veins led to a form of cirrhosis similar to that seen with chronic passive hepatic congestion (so-called cardiac cirrhosis). At one time, this form of injury accounted for one third of the cases of cirrhosis seen in Jamaica, with death often resulting from complications of portal hypertension in as few as one to three years.3 Certain pyrrolizidine alkaloids, such as comfrey extracts, are hepatocarcinogenic and, like aflatoxins, induce mutations of the TP53 gene.116

Germander

The blossoms of plants from the Labiatae family (Teucrium chamaedrys) were used for years in herbal teas and in the mid-1980s as capsules for weight reduction in France, until several dozen cases of liver injury, including fatal hepatic failure,118,119 forced its withdrawal from the French marketplace in 1992.119 Most patients were middle-aged women who had ingested germander for 3 to 18 weeks, with consequent development of acute hepatocellular injury, often with jaundice.119 The injury usually resolved within 1.5 to 6 months after the germander was discontinued, with prompt recurrence after rechallenge in many persons. The cause of germander hepatotoxicity is an interplay between toxic metabolites and immunoallergic mechanisms. Germander is composed of several compounds, including glycosides, flavonoids, and furan-containing diterpenoids, all of which are converted by the CYP system (especially CYP3A) to reactive metabolites. Covalent binding to cellular proteins, depletion of hepatic glutathione, apoptosis, and cytoskeleton membrane injury (bleb formation) cause

Chapter 87  Liver Disease Caused by Anesthetics, Toxins, and Herbal Preparations Table 87-10  Hepatotoxic Herbal Remedies HEPATOTOXIC COMPONENT

TYPE OF LIVER INJURY

Uncertain Uncertain Pyrrolizidine alkaloids

Reversible hepatitis or cholestasis FHF SOS

Anthracene glycoside N-nitro-fenfluramine Nordihydroguaiaretic acid

Cholestatic hepatitis Acute hepatitis, FHF Acute and chronic hepatitis, FHF

Ephedra spp. Lycopodium serratum

Ephedrine ?Levo-tetrahydropalmitine

Multiple uses

Scutellaria root

Diterpenoids

  Shou-wu-pian Comfrey Germander

Traditional medicine Herbal tea Weight loss, fever

Uncertain Pyrrolizidine alkaloid Diterpenoids, epoxides

Greater celandine Herbalife Hydroxycut Impila Kava kava

Gallstones, IBS Nutritional supplement Weight loss Multiple uses Anxiolytic

Polygonum multiflorum Symphytum spp. Teucrium chamaedry, T. capitatum, T. polium Chelidonium majus — Camellia sineusis Callilepsis laureola Piper methysticum

Uncertain Various Green tea Potassium atractylate Kava lactone, pyrone

Kombucha Lipokinetix Mistletoe

Weight loss Weight loss Asthma, infertility

Lichen alkaloid Lichen alkaloid Viscus album

Usnic acid Usnic acid Uncertain

Oil of cloves Pennyroyal (squawmint oil) Prostata Sassafras Senna Skullcap Valerian

Dental pain Abortifacient

Various foods, oils Hedeoma pulegoides, Mentha pulegium Multiple sources Sassafras albidum Cassia angustifolia Scutellaria Valeriana officinalis

Eugenol Pulegone, monoterpenes

Severe hepatitis, FHF Acute or chronic hepatitis or cholestasis, steatosis Hepatocellular necrosis, cholestasis, steatosis, granulomas Acute hepatitis or cholestasis Acute SOS, cirrhosis Acute and chronic hepatitis, ?autoimmune injury, FHF Cholestatic hepatitis, fibrosis Severe hepatitis, FHF Acute hepatitis, FHF(?) Hepatic necrosis Acute hepatitis, cholestasis, FHF(?) Acute hepatitis Acute hepatitis, jaundice, FHF Hepatitis (in combination with skullcap) Zonal necrosis Severe hepatocellular necrosis

Uncertain Safrole Sennoside alkaloids; anthrone Diterpenoids Uncertain

Chronic cholestasis HCC (in animals) Acute hepatitis Hepatitis Elevated liver enzymes

REMEDY

popular USE

SOURCE

Barakol Black cohosh “Bush tea”

Anxiolytic Menopausal symptoms Fever

Cascara Chaso/onshido Chaparral leaf (greasewood, creosote bush) Chinese medicines   Ma huang   Jin bu huan

Laxative Weight loss “Liver tonic,” burn salve, weight loss

Cassia siamea Cimicifuga racemosa Senecio, Heliotropium, Crotalaria spp. Cascara sagrada — Larrea tridenta

Weight loss Sleep aid, analgesic

  Syo-saiko-to

Prostatism Herbal tea Laxative Anxiolytic Sedative

FHF, fulminant hepatic failure; HCC, hepatocellular carcinoma; IBS, irritable bowel syndrome; SOS, sinusoidal obstruction syndrome. Individual references can be found in Stedman C. Herbal hepatotoxicity. Semin Liver Dis 2002; 22:195-206, and references 37, 102, 139-146, 150, 151, 154, 155.

cell disruption in animal models.6,8 Epoxide hydrolase on plasma membranes is a target of germander antibodies, which have been found in the sera of patients who have consumed germander teas over long periods of time.120 Reports of liver injury also have appeared with other species of Teucrium, including Teucrium capitatum121 and Teucrium polium.122

Chaparral

The dried leaf of the desert shrub chaparral (Larrea tridentata), also known as greasewood or creosote bush, is ground into a tea or used in capsules or tablets for various ailments. Multiple reports of hepatitis have appeared; most cases have occurred within 1 to 12 months of use and resolved within a few weeks to months of discontinuation.123 Among 13 cases reported to the FDA,123 acute hepatocellular or cholestatic injury was observed, with 2 cases of fulminant hepatitis requiring liver transplantation and 4 cases of progression to cirrhosis. Renal toxicity and skin rash can accompany liver injury. The active ingredient, nordihydroguaiaretic acid, an inhibitor of cyclooxygenase and lipoxygenase pathways, is the likely cause of hepatic injury, although the mechanism also may involve phytoestrogen-induced effects on the liver.101 A case of recurrence on rechallenge suggests a possible role for immunoallergy.6

Pennyroyal

The leaves of pennyroyal (the common name for two related plant species, Hedeoma pulegoides and Mentha pulegium) are used to make oils (squawmint oil), tablets, and home-brewed mint teas. The plant contains pulegone and smaller amounts of other monoterpene ketones. Oxidative metabolites of pulegone (e.g., menthofuran) bind to cellular proteins and deplete hepatic glutathione, thereby leading to liver injury.124 Cases of hepatocellular injury, including fatal necrosis, were associated with gastrointestinal and central nervous system toxicity within a few hours of ingestion. In animals, inhibition of pulegone metabolism by the CYP system with disulfiram and cimetidine has limited pennyroyal hepatotoxicity.125 The use of N-acetylcysteine may protect against pennyroyal toxicity in human cases.124

Chinese Herbal Medications

Jin bu huan (Lycopodium serratum) is a traditional herbal remedy that has been used as a sedative and analgesic for more than 1000 years.6 Numerous cases of hepatic injury have appeared,126,127 with a mean latency of 20 weeks (range, 7 to 52 weeks) after the start of jin bu huan in recommended doses. Associated symptoms and signs included fever, fatigue, nausea, pruritus, abdominal pain, hepatomegaly,

1457

1458

Section IX  Liver and jaundice. Liver biopsy specimens from a small number of patients showed a range of histopathologic changes, including lobular hepatitis with prominent eosinophils, mild hepatitis with microvesicular steatosis, and fibrotic expansion of the portal tracts. The injury resolved within a mean of 8 weeks (range, 2 to 30 weeks) but could recur on rechallenge.126 The only predisposing factor was female gender. Serum ALT levels were increased 20- to 50-fold, with minor increases in the alkaline phosphatase level, except in one patient with cholestasis. Hyperbilirubinemia was prominent in the more severe cases. A case of chronic hepatitis has been described.6 The mechanism of injury may involve levo-tetrahydropalmatine, a neuroactive metabolite with structural similarity to pyrrolizidine alkaloids. At present, the FDA has banned the importation of jin bu huan anodyne tablets into the United States.6 Syo-saiko-to (xiao-chai-hu-tang, dai-saiko-to) contains Scutellaria root (skullcap), which is a postulated hepatotoxin.128 The spectrum of liver injury has included hepatocellular necrosis, microvesicular steatosis, cholestasis, granuloma formation, and a flare of autoimmune hepatitis.129 Reversible acute hepatitis or cholestasis has followed the consumption of shou-wu-pian, a product derived from Polygonum multiflorum.130 Ma huang, derived from plants of Ephedra species, has been reported to cause acute, sometimes severe, hepatitis, including acute liver failure.101,131,132 The active constituent, ephedrine, also has been linked to severe adverse cardiovascular and central nervous system effects, including fatalities, when used as a stimulant and weight loss aid.133 The FDA has issued a ruling that ephedra-containing products present an unreasonable risk and should be avoided.134

Weight Loss Products

Chaso and onshido are Chinese herbal dietary weight loss supplements that were reported to cause severe liver injury, with a mean serum ALT level of 1978 U/L (range, 283 to 4074 U/L), in 12 patients.135 Fulminant hepatic failure developed in two persons: one died, and the other survived after receiving a liver transplant. The suspected hepatotoxic ingredient was N-nitroso-fenfluramine, a derivative of the appetite suppressant fenfluramine, which was withdrawn from the U.S. market in 1997.136 Another dietary supplement used for weight loss, Lipokinetix (composed of norephedrine, sodium usniate [usnic acid], diiodothyronine, yohimbine, and caffeine), has been associated with acute hepatitis, including fulminant hepatic failure requiring liver transplantation.101,137 In a case series of seven previously healthy patients (four women, three men; mean age, 27 years), acute hepatitis developed after a latent period of less than 4 weeks in five patients and 8 to 12 weeks in the other two. Mean serum ALT levels were 4501  U/L (range, 438 to 14,150  U/L), and mean serum bilirubin levels were 6.5  mg/dL (range, 2.2 to 14.6  mg/dL). No evidence of immunoallergy was evident. All of the patients recovered spontaneously, with normalization of serum ALT and bilirubin levels within four months. Fulminant hepatic failure necessitating emergency liver transplantation was reported in a previously healthy 28-year-old nonobese woman who had taken an over-thecounter preparation of usnic acid for weight loss,138 suggesting that this agent may be the hepatotoxic component of Lipokinetix. Usnic acid also is a component of Kombucha tea, which has been associated with hepatic injury.6 Usnic acid is a potent inhibitor of CYP2C19 and CYP2C9 and may interact with other medications or supplements to produce

hepatotoxic drug-drug interactions.139 Herbalife is the latest nutritional supplement to be reported to cause severe liver injury, including the need for liver transplantation.140,141 A mixed hepatic infiltrate with eosinophils and lymphocytes was seen on liver biopsy specimens, along with other changes, including necrosis, cholestasis, and sinusoidal obstruction. A positive rechallenge reaction was observed in a number of patients after their liver biochemical abnormalities had normalized.140 Hydroxycut was withdrawn from the U.S. market in 2009 after an FDA review of reports found that the slimming aid was associated with possible hepatotoxicity.142 One of its active ingredients, green tea (Camellia sinensis), despite being used widely for millennia,143 has been implicated in liver injury.144 A safety review of green tea by the U.S. Pharmacopeia145 found that in each reported case a relation with liver injury was possible, although the study used the Naranjo causality scale, which is not specific for assessing hepatotoxicity.146

Kava Kava

Kava kava is a natural sedative and antianxiety agent derived from the root of the pepper plant (Piper methysticum). This herbal product has been the subject of an FDA consumer alert6 after it was banned in the European Union and Canada147 because of severe hepatotoxicity, including fatal liver failure.101,148,149 A review of 78 cases of hepatic injury reported to the FDA included 11 cases of liver failure requiring liver transplantation and 4 deaths.149 Other investigators, however, have questioned the validity of the causality assessment, and only rare instances of hepatotoxicity are found when a more accurate causality scale is used.150,151 Although kavalactone has been shown to inhibit cytochrome P450 enzymes, deplete hepatic glutathione, and possibly inhibit cyclooxygenase,149 the hepatotoxic component may be the major kava alkaloid pipermethystine.152 Induction of apoptosis and mitochondrial toxicity are the suspected hepatotoxic mechanisms.153

Black Cohosh

Black cohosh (Actaea racemosa and Cimicifuga racemosa), used for menopausal symptoms, has been implicated in reports of possible hepatic injury.154 As for kava and green tea (see earlier), the causality assessment methodology is considered inaccurate.155

KEY REFERENCES

Chitturi S, Farrell GC. Hepatotoxic slimming agents and other herbal hepatotoxins. J Gastroenterol Hepatol 2008; 23:366-73. (Ref 5.) Chojkier M. Hepatic sinusoidal-obstruction syndrome: Toxicity of pyrrolizidine alkaloids. J Hepatol 2003; 39:437-46. (Ref 116.) Elinav E, Pinsker G, Safadi R, et al. Association between consumption of Herbalife nutritional supplements and acute hepatotoxicity. J Hepatol 2007; 47:514-20. (Ref 140.) Faust TW, Reddy KR. Postoperative jaundice. Clin Liver Dis 2004; 8:151-66. (Ref 33.) Favreau JT, Ryu ML, Braunstein G, et al. Severe hepatotoxicity associated with the dietary supplement LipoKinetix. Ann Intern Med 2002; 136:590-5. (Ref 137.) Foti RS, Dickmann LJ, Davis JA, et al. Metabolism and related human risk factors for hepatic damage by usnic acid containing nutritional supplements. Xenobiotica 2008; 38: 264-80. (Ref 139.) Kenna JG. Mechanism, pathology, and clinical presentation of hepatotoxicity of anesthetic agents. In Kaplowitz N, DeLeve L, editors. Drug-Induced Liver Disease. New York: Marcel Dekker; 2004. pp 405-24. (Ref 11.) Lewis JH, Kleiner D. Hepatic injury due to drugs, chemicals and toxins. In: Burt AD, Portmann BC, Ferrell LD, editors. MacSween’s Pathology of the Liver, 5th ed. Philadelphia: Churchill Livingstone Elsevier; 2007. pp 649-759. (Ref 37.)

Chapter 87  Liver Disease Caused by Anesthetics, Toxins, and Herbal Preparations Mandibusan MK, Odin M, Eatmond DA. Postulated carbon tetrachloride mode of action: A review. J Environ Sci Health C Environ Carcinog Ecotoxicol Rev 2007; 25:185-209. (Ref 38.) Mohi-ud-din R, Lewis JH. Drug- and chemical-induced cholestasis. Clin Liver Dis 2004; 8:95-132. (Ref 93.) Rengstorff DS, Osorio RW, Bonacini M. Recovery from severe hepatitis caused by mushroom poisoning without liver transplantation. Clin Gastroenterol Hepatol 2003; 1:392-6. (Ref 89.) Schoepfer AM, Engel A, Fattinger K, et al. Herbal does not mean innocuous: Ten cases of severe hepatotoxicity associated with dietary supplements from Herbalife products. J Hepatol 2007; 47:521-6. (Ref 141.) Seeff LB, Lindsay KL, Bacon BR, et al. Complementary and alternative medicine in chronic liver disease. Hepatology 2001; 34:595-603. (Ref 98.)

Verma S, Thuluvath PJ. Complementary and alternative medicine in hepatology: Review of evidence of efficacy. Clin Gastroenterol Hepatol 2007; 5:408-16. (Ref 99.) Zimmerman HJ. Hepatotoxicity. The Adverse Effects of Drugs and Other Chemicals on the Liver, 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 1999. (Ref 1.) Zimmerman HJ, Lewis JH. Chemical- and toxin-induced hepatotoxicity. Gastroenterol Clin North Am 1995; 24:1027-45. (Ref 3.) Full references for this chapter can be found on www.expertconsult.com.

1459

CHAPTE R

88 Autoimmune Hepatitis Albert J. Czaja

CHAPTER OUTLINE Diagnostic Criteria  1461 Clinical Criteria  1462 Scoring Criteria  1462 Pathogenesis  1463 Classification  1465 Type 1 Autoimmune Hepatitis  1465 Type 2 Autoimmune Hepatitis  1465 Variant Forms  1466 Variant with Primary Biliary Cirrhosis  1466 Variant with Primary Sclerosing Cholangitis  1466 Variant with Cholestatic Features  1467 Autoantibody-Negative Autoimmune Hepatitis  1467 Autoimmune Hepatitis and Chronic Hepatitis C  1467 Epidemiology  1468 Prognostic Indices  1468 Laboratory Indices  1468 Histologic Findings  1469 Human Leukocyte Antigen Status  1469 Ethnicity  1469 Serologic Markers  1470 Model for End-Stage Liver Disease Score  1470

Autoimmune hepatitis (AIH) is a disorder of unknown cause characterized by unresolving inflammation of the liver and by the presence of interface hepatitis on histologic examination (Fig. 88-1), hypergammaglobulinemia, and autoantibodies.1 Diagnosis requires the exclusion of other chronic liver diseases that have similar features, including Wilson disease, chronic viral hepatitis, drug-induced liver disease, nonalcoholic fatty liver disease, and the immune cholangiopathies of primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC). Panacinar or lobular hepatitis is within the histologic spectrum of AIH (Fig. 88-2), and centrilobular (Rappaport zone 3) necrosis has been described in AIH and may indicate an early stage of the disease before the development of interface hepatitis.2

DIAGNOSTIC CRITERIA An international panel codified the diagnostic criteria of AIH in 1992, and an expanded panel updated them in 1999.3 The propensity for an acute, rarely fulminant, presentation has been recognized,2 and the requirement for six months of disease activity to establish chronicity has been waived.3 Cholestatic histologic changes, including bile duct injury and ductopenia, are incompatible findings, but trivial biliary changes within the background of classic histologic features do not preclude the diagnosis.4,5 The serologic tests essential for diagnosis are assays for antinuclear antibodies (ANA), smooth muscle antibodies (SMA), and antibodies to liver-kidney microsome type 1

Clinical Features  1470 Severe or Fulminant Presentation  1471 Advanced Age  1471 Asymptomatic Patients  1471 Associated Diseases  1471 Hereditary Forms  1471 Treatment  1471 Indications  1471 Regimens  1471 Drug Actions  1471 Drug-Related Side Effects  1472 End Points  1473 Results  1474 Relapse  1475 Treatment Failure  1475 Incomplete Response  1475 Drug Toxicity  1475 Liver Transplantation  1475

(anti-LKM1).1,6,7 These assays are based on the indirect immunofluorescence of rodent tissues or Hep-2 cell lines or on enzyme immunoassays using microtiter plates with adsorbed recombinant or highly purified antigens. Atypical perinuclear anti-neutrophil cytoplasmic antibodies (atypical pANCA) are common in type 1 AIH, PSC, and chronic ulcerative colitis.6,7 They are directed against antigens within the nucleus, rather than the cytoplasm, of granulocytes, and the reactivities localize to the proteins within the lamina of the nucleus. Because these antibodies are directed against nuclear envelope antigens rather than cytoplasmic antigens, a better term for them is anti-neutrophil nuclear antigens or “ANNA.”8 Atypical perinuclear anti-neutrophil cytoplasmic antibodies have been useful in evaluating patients who lack the conventional autoantibodies.6,7 Celiac disease can be associated with liver disease that resembles AIH and should be excluded in patients with cryptogenic chronic hepatitis by screening for immunoglobulin (Ig) A antibodies to tissue transglutaminase or endomysium.6,7,9 IgA endomysial antibodies are more predictive of celiac disease in patients with AIH than are IgA antibodies to tissue transglutaminase, which can be stimulated by hepatic inflammation and fibrogenesis.6,7 New autoantibodies continue to be characterized in the hope of improving diagnostic specificity and prognostic value, but none has been incorporated into conventional diagnostic algorithms.6,7,10 Antibodies to soluble liver antigen (anti-SLA), actin (anti-actin), chromatin (antichromatin), asialoglycoprotein receptor (ASGPR), and liver cytosol type 1 (anti-LC1) have been associated with severe AIH, poor treatment response, and relapse after drug

1461

1462

Section IX  Liver Elevated serum AST and gamma globulin levels AST: alkaline phosphatase >3

AMA negative Ceruloplasmin normal Normal α1-antitrypsin phenotype Normal or near-normal serum iron level HBsAg, anti-HCV, IgM anti-HAV negative Liver biopsy Interface hepatitis ± lobular hepatitis (see Figs. 88-1 and 88-2)

Definite Figure 88-1.  Histopathology of interface hepatitis. The limiting plate of the portal tract is disrupted by a lymphoplasmacytic infiltrate. This histologic pattern is the hallmark of autoimmune hepatitis, but it is not disease specific. (Hematoxylin and eosin, ×200.)

Gamma globulin level ≥1.5 normal ANA, SMA, or anti-LKM1 ≥1:80 No exposure to drugs or blood products Alcohol intake <25 g/day

Type 1 ANA and/or SMA +

Figure 88-2.  Histopathology of panacinar (lobular) hepatitis. Mononuclear inflammatory cells line the sinusoidal spaces. Typically, panacinar (lobular) hepatitis coexists with interface hepatitis, but it may be pronounced during an acute onset of autoimmune hepatitis or during a relapse after treatment withdrawal. (Hematoxylin and eosin, ×200.)

withdrawal.10-13 Their major clinical limitation has been their infrequent individual occurrence in AIH.

CLINICAL CRITERIA

The definite diagnosis of AIH requires exclusion of other similar diseases; laboratory findings that indicate substantial immunoreactivity; and histologic features of interface hepatitis (Fig. 88-3).3 A probable diagnosis is justified when findings are compatible with AIH but insufficient for a definite diagnosis (see Fig. 88-3).3 Patients who lack con­ ventional autoantibodies but who are seropositive for investigational markers, such as antibodies to ASGPR, SLA, actin, or LC1, are classified as having probable disease.3

SCORING CRITERIA

The original scoring system proposed by the International Autoimmune Hepatitis Group accommodates the diverse

Probable Gamma globulin level <1.5 normal ANA, SMA, or anti-LKM1 ≤1:40 Previous drugs or blood products Alcohol use Other liver-related autoantibodies (see Table 88-1)

Type 2 Anti-LKM1 +

Figure 88-3.  Diagnostic algorithm for autoimmune hepatitis. Diagnosis requires predominant elevation of the serum aminotransferase levels, exclusion of other similar disorders (especially Wilson disease, druginduced hepatitis, and viral hepatitis), interface hepatitis on histologic examination, and manifestations of immunoreactivity, including serum gamma globulin elevation and seropositivity for antinuclear antibodies (ANA), smooth muscle antibodies (SMA), or antibodies to liver-kidney microsome type 1 (anti-LKM1). The degree of immunoreactivity and the presence of confounding etiologic factors, such as alcohol or drug exposure, distinguish definite from probable autoimmune hepatitis. Classification into one of the descriptive categories of type 1 and type 2 autoimmune hepatitis is based on the nature of the autoantibodies. AMA, antimitochondrial antibodies; AST, aspartate aminotransferase; anti-HAV, antibody to hepatitis A virus; anti-HCV, antibody to hepatitis C virus; HBsAg, hepatitis B surface antigen; IgM, immunoglobulin M.

manifestations of AIH and renders an aggregate score that reflects the net strength of the diagnosis before and after glucocorticoid treatment (Table 88-1).3 By weighing each component of the syndrome, discrepant features can be accommodated and biases associated with isolated inconsistencies can be avoided. The original scoring system ensures the comparability of study populations in clinical trials, provides a comprehensive template for systematically assessing all features of the disease, and can render a diagnosis of AIH in patients with few or atypical features.3 The original scoring system is not a discriminative diagnostic index, and it should not be used to distinguish AIH from other liver diseases. A simplified scoring system has been developed to ease clinical application and is based on four clinical components that include the presence and level of autoantibody expression by indirect immunofluorescence, serum IgG concentration, histologic features, and viral markers (Table 88-2).14 The original scoring system has greater sensitivity for the diagnosis than the simplified system (100% vs.

Chapter 88  Autoimmune Hepatitis Table 88-1  Revised Original Scoring System for the Diagnosis of Autoimmune Hepatitis CATEGORY

VARIABLE

Gender AP/AST

Female >3 <1.5 >2.0* 1.5-2.0* 1.0-1.5* <1.0 >1 : 80 1 : 80 1 : 40 <1 : 40 Positive Positive Negative Yes No <25 g/day >60 g/day DR3 or DR4 Thyroiditis, ulcerative colitis, synovitis, others Anti-SLA, anti-actin, anti-LC1, pANCA Interface hepatitis Plasmacytic infiltrate Rosettes None of above Biliary changes Other features Complete Relapse

Gamma globulin or IgG level above normal ANA, SMA, or anti-LKM1 titer AMA Viral markers Drug history Alcohol HLA Immune disease Other liver-defined autoantibodies Histologic features

Treatment response Pretreatment Score Definite diagnosis Probable diagnosis Post-treatment Score Definite diagnosis Probable diagnosis

Table 88-2  Simplified Scoring System for Diagnosis of Autoimmune Hepatitis* SCORE +2 −2 +2 +3 +2 +1 0 +3 +2 +1 0 −4 −3 +3 −4 +1 +2 −2 +1 +2 +2 +3 +1 +1 −5 −3 −3 +2 +3

>15 10-15 >17 12-17

*Times upper limit of normal. AMA, antimitochondrial antibodies; ANA, antinuclear antibodies; anti-LC1, antibodies to liver cytosol type 1; anti-LKM1, antibodies to liver-kidney microsome type 1; anti-SLA, antibodies to soluble liver antigen; AP/AST (or AP/ALT), ratio of serum alkaline phosphatase level to serum aspartate aminotransferase (or serum alanine aminotransferase) level; HLA, human leukocyte antigen; IgG, immunoglobulin G; pANCA, perinuclear anti-neutrophil cytoplasmic antibodies; SMA, smooth muscle antibodies. Adapted from Alvarez F, Berg PA, Bianchi FB, et al. International Autoimmune Hepatitis Group report: Review of criteria for diagnosis of autoimmune hepatitis. J Hepatol 1999; 31:929-38.

95%), but the simplified system has greater specificity (90% vs. 73%) and predictability (92% vs. 82%).15 Whereas the original scoring system is useful for evaluating patients in whom every component must be assessed because of few or atypical findings, the simplified scoring system is useful for excluding AIH in patients with other conditions and concurrent immune features.

PATHOGENESIS The pathogenic mechanisms of AIH are unknown. The most popular hypotheses invoke a constellation of interactive factors that include a triggering agent, genetic predisposition, and various determinants of autoantigen display, immunocyte activation, and effector cell expansion (Fig.

CATEGORY

VARIABLE

Autoantibodies† Antinuclear antibodies or smooth muscle antibodies

1 : 40

+1

Antibodies to liver-kidney microsome type 1 Antibodies to soluble liver antigen Immunoglobulin Level Immunoglobulin G

≥1 : 80 ≥1 : 40

+2 +2

Positive

+2

>Upper limit of normal >1.1 times upper limit of normal

+1

Compatible with autoimmune hepatitis Typical of autoimmune hepatitis

+1

No viral markers

+2

Histologic Findings Morphologic features

Viral Disease Absence of viral hepatitis Pretreatment Aggregate Score Definite diagnosis Probable diagnosis

SCORE

+2

+2

≥7 6

*Adapted from Hennes EM, Zeniya M, Czaja AJ, et al. Simplified diagnostic criteria for autoimmune hepatitis. Hepatology 2008; 48:169-76. † Autoantibody titers as determined by indirect immunofluorescence.

88-4).16-18 Proposed triggering factors include infectious agents, drugs, and toxins. The lag time between exposure to the trigger and onset of the disease can be long, and the triggering factor may not be needed for perpetuation of the disorder. The CD4+ helper T cell is the principal effector cell, and its activation is the initial step in the pathogenic pathway. Molecular mimicry of a foreign antigen and a self-antigen is the most common explanation for the loss of selftolerance, but this mechanism has not been established for any autoimmune disease.16-18 Genetic factors influence autoantigen presentation and CD4+ helper T cell recognition. The antigen-binding groove of the class II molecule of the major histocompatibility complex (MHC) is encoded by alleles that determine the groove’s configuration and ability to activate immunocytes. The susceptibility alleles of AIH in white North Americans and northern Europeans reside on the DRB1 gene and are DRB1*0301 and DRB1*0401 (see Fig. 88-4).19-21 Different ethnic groups have different susceptibility alleles, a finding that supports a “shared motif hypothesis” of pathogenesis.19-21 According to this hypothesis, the risk of disease relates to amino acid sequences in the antigenbinding groove of the class II MHC molecule, and multiple alleles encode the same or similar sequence (“shared motif”). The critical shared motif in white North Americans and northern Europeans with AIH is a six-amino-acid sequence represented by the code LLEQKR.19-21 This sequence is located between positions 67 and 72 of the DRβ polypeptide chain of the class II MHC molecule, and lysine (K) in position 71 is the critical determinant of susceptibility. DRB1*0301 and DRB1*0401 encode identical amino acid sequences in the DRβ67-72 region and affect susceptibility similarly. DRB1*0404 and DRB1*0405 are the susceptibility alleles in Mexican, Japanese, mainland Chinese, and Argentine

1463

1464

Section IX  Liver 1st signal

B7 2nd signal CD28

APC P4

P4 Aspartic acid Glutamic acid

Lysine Arginine

DRB1*0401

Antigen-binding groove

DRB1*1301 Diverse autoantibodies DRB1*03 DRB1*07

T cell

DRβ71 DRB1*0301

Antigenic peptide

CD4+

CTLA4 Type 1 cytokine response TNFA*2

Autoimmune promoters Cytokine Polymorphisms Increased bcl-2 expression Cytotoxic FAS Granzyme T cells Polymorphism perforin

Activated CD4+ T-helper cell

Deficient T-reg cells

FAS ligand

Type 2 cytokine response

Plasma cells (from B cells)

FAS Host DNA cytosolic fragments

FAS Polymorphism Hepatocyte apoptosis

Cell-mediated cytotoxity

Fc receptor Hepatocyte NK Antibody-mediated cellular cytoxicity

Figure 88-4.  Interactive mechanisms that contribute to the development of autoimmune hepatitis in white North American and northern European adults. The initial stimulus for immune activity is an antigenic peptide (upper left corner) that has a negatively charged aspartic acid or glutamic acid at a position within its structure (P4) that can form a salt bridge with a positively charged lysine or arginine residue at position 71 within the DR beta polypeptide chain (DRb71) of the antigen binding groove of the class II DR molecule of the major histocompatibility complex (top center). The DR molecule-antigen complex of the antigen-presenting cell (APC) then interacts with the antigen receptor of a CD4+ T-helper cell (interaction not shown), and the first co-stimulatory signal is completed (1st signal). The CD28 molecule on the surface of the CD4+ T-helper cell ligates with the B7 ligand on the surface of the APC, and the second co-stimulatory signal (2nd signal) is completed (upper right corner). The activated CD4+ T-helper cell can then differentiate and proliferate along type 1 and type 2 cytokine pathways (middle right). Deficiencies in the function or amount of cytotoxic T lymphocyte antigen 4 (CTLA4) can enhance the strength of the 2nd signal. Differentiation along the type 1 cytokine pathway can be promoted by polymorphisms of the tumor necrosis factor gene (TNFA*2) and tumor necrosis factor receptor superfamily gene (FAS), resulting in cell-mediated cytotoxicity by sensitized liver-infiltrating cytotoxic T cells and increased hepatocyte apoptosis (middle bottom). The apoptosis of hepatocytes can, in turn, release DNA cytosolic fragments that contribute to the production of diverse collateral autoantibodies (middle left). Autoantibody expression is, in part, host-dependent and influenced by the susceptibility alleles DRB1*03 and DRB1*07. Host genetic predispositions are also important in encoding the antigen-binding groove of the class II DR molecule through the actions of DRB1*0301, DRB1*0401, and DRB1*1301 alleles and in generating autoimmune promoters (cytokine and FAS polymorphisms) that enhance cell-mediated cytotoxicity and hepatocyte apoptosis. The enhanced expression of the anti-apoptotic protein (bcl-2) on the surface of cytotoxic T cells can protect them from programmed cell death and perpetuate their autoreactivity (middle). The cytokine pathways can be enhanced by deficiencies in the actions of T-regulatory cells (T-reg cells), which have suppressive effects that can be reversed by glucocorticoid treatment. Differentiation of B cells into plasma cells, via the type 2 cytokine response of activated CD4+ T-helper cells, can result in immunoglobulin production that, in turn, generates an antibodymediated cellular toxicity. Natural killer (NK) cells with Fc receptors are directed against complexes of immunoglobulin with normal hepatocyte membrane constituents. (Adapted from Czaja AJ. Autoimmune hepatitis—Part A: Pathogenesis. Expert Rev Gastroenterol Hepatol 2007; 1:113-128.)

adults and encode a similar sequence, except for arginine (R) instead of lysine (K) at the DRβ71 position.20,21 Arginine is a positively charged amino acid that is structurally similar to lysine, and its substitution for lysine should not greatly alter the antigen-binding properties of the class II MHC molecule. By contrast, DRB1*1501 protects against AIH in white North Americans and northern Europeans, and this allele encodes isoleucine (I) instead of leucine (L) at position DRβ67 and alanine (A) instead of lysine (K) at position DRβ71. Alanine is a neutral, nonpolar amino acid that, when substituted for lysine, should greatly affect antigen presentation and immunocyte activation. Antigenic peptides are selected for display by the nature of the amino acids that interact with residues within the antigen-binding groove (see Fig. 88-4).21,22 The critical sixamino-acid motif in AIH restricts the range of peptides that can be accommodated. Multiple self-antigens or foreign

antigens may satisfy the minimal structural requirements and serve as immunogenic peptides. The ideal triggering epitope must have a negatively charged amino acid residue (aspartic acid or glutamic acid) at peptide position P4 to form a salt bridge with the positively charged lysine or arginine at DRβ71.21 Molecular modeling indicates that a negatively charged P4 residue in the antigenic peptide and the positively charged lysine or arginine at DRβ71 can form a P4-DRβ71 immunoreactive unit that is independent of the other residues within the antigen and antigen-binding groove. This minimal immunoreactive unit can be created by multiple antigenic peptides and class II MHC molecules, and the number of these units may affect susceptibility by a “dose effect.” DRB1*1301 is associated with AIH in Argentine children22 and Brazilian patients23,24 and encodes ILEDER at positions DRβ67-72. Glutamic acid (E), aspartic acid (D),

Chapter 88  Autoimmune Hepatitis and glutamic acid (E) are at positions DRβ69, DRβ70, and DRβ71, respectively, in the class II MHC molecule, and the presence of these critically located but negatively charged amino acid residues argues against the “shared motif” hypothesis of pathogenesis. The “molecular footprint” hypothesis of pathogenesis holds that susceptibility to AIH in different regions and ethnic groups relates to indigenous factors or agents favored by certain genetic phenotypes.20,21 In South America, DRB1*1301 is associated with protracted hepatitis A virus infection, and persons with this allele may be “selected” from their environment to have prolonged exposure to viral and hepatic antigens that favor the development of AIH.25 An understanding of the individual susceptibility allele in different geographic regions may allow use of this “footprint” to track the cause of the disease. The “autoimmune promoter hypothesis” of pathogenesis complements the “shared motif” and “molecular footprint” hypotheses by proposing that genetic promoters inside and outside the MHC can affect disease occurrence, either in synergy (epistasis) with the principal susceptibility factors or in lieu of them.18,20,21 Polymorphisms of the tumor necrosis factor (TNF)-α gene (TNFA*2),26 the cytotoxic T lymphocyte antigen 4 gene (CTLA4),27 and Fas gene promoter at position -670 (TNFRSF6)28 have been associated with increased immunoreactivity, disease severity, and early progression to cirrhosis in white North American and northern European patients. Constellations of autoimmune promoters, as yet undefined, may individualize the disease by affecting its occurrence, clinical phenotype, and outcome. Liver cell destruction is accomplished by either cellmediated cytotoxicity, antibody-dependent cell-mediated cytotoxicity, or a combination of both mechanisms (see Fig. 88-4).16-18 Cell-mediated cytotoxicity depends on the clonal expansion of CD8+ cytotoxic T cells that accomplish liver cell injury through the release of lymphokines. This mechanism is regulated by type 1 cytokines, and the −308 polymorphism of TNFA*2 may facilitate this pathway.26 Antibody-dependent cell-mediated cytotoxicity is regulated by type 2 cytokines, and the natural killer cell accomplishes liver cell destruction by binding of its Fc receptor with an antigen-antibody complex on the hepatocyte surface.16-18 The predominant mechanism depends on the phenotypic differentiation of the CD4+ helper T cell, which in turn reflects the cytokine milieu. The cytokine milieu may reflect polymorphisms of the cytokine genes that favor excessive production of some modulators, such as TNF-α, or deficient production of others. Defects in the counter-regulatory cytokine milieu may also reflect reduced numbers of intrahepatic natural killer T (NKT) cells and the failure of T-regulatory (T-reg) cells (CD4+CD25+ cells) to modulate CD8+ T cell proliferation and cytokine production.18,29 Increased numbers of gd T cells may also contribute to the cytodestructive process by recognizing antigens presented by the non-classic MHC molecules, and the recruitment and intrahepatic trafficking of cytotoxic T lymphocytes may be enhanced by the upregulation of chemokines, such as CXCL16.18 Fibrogenesis, in turn, is fueled by the resulting inflammatory activity as perivascular hepatic stellate cells transform into myofibroblasts. The matrix proteins accumulate as tissue inhibitors retard the counteractive degradative actions of matrix metalloproteinases, and stellate cells continue to be activated in an autocrine fashion by transforming growth factor-β (TGF-β).30 Glucocorticoid therapy can favorably alter the cytokine milieu, improve the number and function of the T-reg cells, impair activation of TGF-β, promote disappear-

ance of the metalloproteinase inhibitors, enhance degradation of the fibrotic liver matrix, and foster apoptosis of the hepatic stellate cells (see also Chapter 2).31

CLASSIFICATION Two types of AIH have distinctive serologic profiles. Neither has been ascribed a unique cause, specific treatment strategy, or special type of behavior (Table 88-3). The terms are useful as clinical descriptors and as research designations to ensure homogeneous study populations.

TYPE 1 AUTOIMMUNE HEPATITIS

Type 1 AIH is characterized by SMA, ANA, or both (see Table 88-3).1,7 Antibodies to actin have greater specificity, but less sensitivity, for the diagnosis of AIH than SMA.6 Atypical pANCA are found in as many as 90% of patients with type 1 AIH and typically are absent in type 2 AIH.6,7 Type 1 AIH can occur at any age and in either gender (see Table 88-3).7 Initial studies that suggested a bimodal age distribution probably reflected referral biases to tertiary medical centers. The disease has been described in infants and probably is underdiagnosed in the elderly.32 Seventyeight percent of patients are women (female-to-male ratio 3.6 : 1), and 38% have concurrent extrahepatic immunologic diseases.7 Autoimmune thyroiditis (occurring in 12% of the cases), Graves’ disease (6%), and ulcerative colitis (6%) are the most common associated immune disorders. Rheumatoid arthritis, pernicious anemia, scleroderma, Coombs-positive hemolytic anemia, autoimmune throm­ bocytopenic purpura, symptomatic cryoglobu­linemia, leukocytoclastic vasculitis, nephritis, erythema nodosum, systemic lupus erythematosus, and fibrosing alveolitis also may occur (less than 1% each). Cholangiography is warranted to exclude PSC in all patients who have concurrent ulcerative colitis (see Chapter 68).33 Type 1 AIH is associated with an abrupt onset of symptoms in 40% of cases and may manifest in a fulminant fashion.2 The acute presentation frequently reflects preexisting subclinical disease that is unmasked by progression or represents a spontaneous exacerbation of inflammatory activity. Features of chronicity are lacking in 8% of patients, in whom the presentation of the disorder is indistinguishable from that of acute viral or toxic hepatitis. The target autoantigen of type 1 AIH is unknown. Human leukocyte antigen (HLA)-DR3 (DRB1*0301) and HLA-DR4 (DRB1*0401) are independent risk factors for the disease in white North Americans and northern Europeans.19-21 More than 80% of white patients in Great Britain and the United States possess either DRB1*0301, DRB1*0401, or both, compared with 42% of the unaffected white population. In South America, especially in children, DRB1*1301 is the principal susceptibility allele. These findings indicate that type 1 AIH is a complex polygenic disorder.

TYPE 2 AUTOIMMUNE HEPATITIS

Type 2 AIH is characterized by the expression of anti-LKM1 (see Table 88-3).1,7,34 Most affected persons are children (ages 2 to 14 years), but in Europe, especially in Germany and France, 20% of patients are adults.34 In the United States, type 2 AIH is rare, and only 4% of patients older than 18 years have anti-LKM1.35 The regional differences in prevalence may relate to ethnic differences in the genetic predisposition to the disease.36 Type 2 patients are younger than type 1 patients and may have different clinical and laboratory features (see Table

1465

1466

Section IX  Liver Table 88-3  Classification of Autoimmune Hepatitis Based on Autoantibodies CLINICAL FEATURE

TYPE 1

TYPE 2

Signature autoantibodies

Smooth muscle Nuclear Atypical pANCA Actin* Asialoglycoprotein receptor* Chromatin* Soluble liver antigen* Unknown Infants to elderly 78% 38% Autoimmune thyroiditis Graves’ disease Ulcerative colitis

Liver-kidney microsome type 1

Associated autoantibodies

Putative autoantigen Age Female Concurrent immune diseases Typical concurrent autoimmune diseases

Organ-specific antibodies Serum gamma globulin elevation HLA associations Allelic risk factors

Glucocorticoid responsive

4% +++ B8, DR3, DR4, DR13 DRB1*0301 and *0401   (white North Americans and northern Europeans) DRB1*1301   (South Americans, especially children) +++

Liver cytosol type 1* Soluble liver antigen*

CYP2D6 Children (2-14) 89% 34% Autoimmune thyroiditis Vitiligo Type 1 diabetes mellitus APECED 30% + B14, DR3, C4A-Q0, DR7 DQB1*0201, DRB1*07, DRB1*03   (Germans and Brazilians) ++

*Autoantibodies with an asterisk are investigational only and not available for routine clinical use. APECED, autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy; CYP2D6, cytochrome P450 2D6; pANCA, perinuclear anti-neutrophil cytoplasmic antibodies.

88-3).1,7,34 As with type 1 AIH, an acute or fulminant presentation is possible and important to recognize and treat early. Susceptibility to type 2 AIH has been associated with DQB1*0201, DRB1*07, and DRB1*03.37 DQB1*0201 is in strong linkage disequilibrium with DRB1*07 and DRB1*03 and has been proposed as the principal genetic determinant of the disease. The expression of anti-LKM1 has been associated with DRB1*07, and various aspects of type 2 AIH may have different genetic determinants.38 The target antigen of type 2 AIH is the cytochrome P450 2D6 mono-oxygenase (CYP2D6).39 This protein is a 50-kd microsomal drug-metabolizing enzyme, and its expression on the hepatocyte surface can be modulated by interleukins and TNF-α. Antibodies to LKM1 inhibit the activity of CYP2D6 in vitro but not in vivo, and lymphocytes extracted from the liver tissue of patients who have the disease exhibit immunoreactivity specific to the antigen. CYP2D6 has been sequenced, cloned, and mapped, and five epitopes are recognized by anti-LKM1.40 The amino acid sequence spanning 193-212 of the CYP2D6 molecule is the target of anti-LKM1 in 93% of patients. Homologies exist between CYP2D6 and the genomes of the hepatitis C virus, cytomegalovirus, and herpes simplex type 1 virus. These molecular mimicries may result in cross-reacting antibodies and support the hypothesis that repeated viral infections may break self-tolerance and cause the disease.17

VARIANT FORMS Patients who have atypical features of AIH currently lack an official designation and confident treatment strategy.41 They may have manifestations of AIH and another type of chronic liver disease or findings that are incompatible with AIH by current diagnostic criteria (Table 88-4).42,43

VARIANT WITH PRIMARY BILIARY CIRRHOSIS

AIH in patients who also have antimitochondrial antibodies (AMA) and histologic features of cholangitis constitutes an

overlap syndrome with PBC (see Table 88-4).41-44 Typically, affected patients have low titers of AMA and concurrent features of bile duct injury or loss.45 Antibodies against the PBC-specific M2 mitochondrial antigens may be present46; histologic features of cholangitis, including destructive cholangitis, may be seen45; and copper staining of hepatic tissue indicative of cholestasis may be observed.45 Occurrence rates range from 5% of patients initially diagnosed as having AIH to 19% of patients initially diagnosed as having PBC (see also Chapter 89).42 The clinical course of the disease and response to treatment depend mainly on the predominant component of the disease. Patients who have high serum aspartate aminotransferase (AST) levels, serum alkaline phosphatase levels less than twice the upper limit of normal, moderate to severe interface hepatitis on histologic examination, and high diagnostic scores for AIH commonly respond to glucocorticoid therapy.42,45 By contrast, patients who have serum alkaline phosphatase levels greater than twice the upper limit of normal, serum gamma glutamyl transpeptidase levels at least five times the upper limit of normal, and florid bile duct lesions on histologic examination mainly have PBC and commonly respond to ursodeoxycholic acid in combination with glucocorticoids (prednisone, prednisolone, or budesonide).44 AMA can be detected in 18% of patients with AIH in the absence of cholestatic features and can appear and disappear during the course of the disease without evolution to a different syndrome or the need for different therapy.47,48 Histologic changes of bile duct injury are required in addition to AMA and other classic features of AIH to designate a variant form of AIH.

VARIANT WITH PRIMARY SCLEROSING CHOLANGITIS

Histologic changes of lymphocytic, pleomorphic, or fibrous cholangitis; cholestatic laboratory findings; concurrent inflammatory bowel disease; or failure to respond to glucocorticoids constitute indications for cholangiography in

Chapter 88  Autoimmune Hepatitis Table 88-4  Variant Forms of Autoimmune Hepatitis AIH + PBC

AIH + PSC

Clinical and laboratory features

AIH features AMA +

Histology

Cholangitis Cholestasis Prednisone if AP ≤ 2 × ULN Prednisone and UDCA if AP >2 × ULN and/or florid duct lesions

AIH features Ulcerative colitis AMA − Abnormal cholangiogram Normal cholangiogram (small duct disease) Cholangitis Cholestasis Prednisone and UDCA

Treatment

AIH + CHOLESTATIC FEATURES

autoantibodynegative AIH

ANA and/or SMA + AMA − No ulcerative colitis Normal cholangiogram

AIH features No autoantibodies HLA-DR3 or DR4

Cholangitis Cholestasis Prednisone and/or UDCA depending on AP level and histologic features

Interface hepatitis Conventional regimens for AIH

AIH, autoimmune hepatitis; AMA, antimitochondrial antibodies; ANA, antinuclear antibodies; AP, serum alkaline phosphatase level; HLA, human leukocyte antigen; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; SMA, smooth muscle antibodies; UDCA, ursodeoxycholic acid; ULN, upper limit of normal.

patients who have AIH.41-43 As many as 41% of persons with one of these features have cholangiographic changes of PSC and are classifiable as having a variant syndrome of AIH (see Table 88-4).33 Furthermore, 54% of patients who have PSC have features that support a probable or definite diagnosis of AIH.42 The absence of characteristic cholangiographic changes does not preclude the diagnosis of PSC because small-duct disease may be present (see Table 88-4) (see Chapter 68).33,49 Magnetic resonance cholangiography has demonstrated unsuspected PSC of the large bile ducts in 8% of adults with AIH, and the possibility of PSC must be evaluated in all individuals with AIH and disease refractory to glucocorticoid therapy regardless of other features.50 Children with AIH may also have unsuspected bile duct changes. Autoimmune sclerosing cholangitis is a disorder described in children who have the clinical phenotype of AIH but abnormal findings on cholangiographic studies.51 Inflammatory bowel disease is frequently absent, and these children can respond to glucocorticoid therapy; in this respect, they may differ from adults with AIH and PSC. Treatment is empirical and typically ineffective in adults.33,42,43 Glucocorticoids and ursodeoxycholic acid (13 to 15 mg/kg orally daily), alone or in combination, can be considered, depending on whether hepatitic or cholestatic features predominate. Some studies have suggested that high-dose ursodeoxycholic acid (20 to 25 mg/kg daily) has some value in typical PSC, and this treatment can be considered in conjunction with glucocorticoids.

sent collateral damage associated with severe inflammatory activity; they do not constitute a variant syndrome, nor do they change the diagnosis or affect the treatment strategy.

AUTOANTIBODY-NEGATIVE AUTOIMMUNE HEPATITIS

Thirteen percent of adults with chronic hepatitis of undetermined cause satisfy international criteria for the diagnosis of AIH but lack the characteristic autoantibodies (see Table 88-4).53 These patients commonly are designated as having cryptogenic chronic hepatitis and may be excluded inappropriately from therapies of potential benefit. Autoantibody-negative patients are similar in age, female predominance, frequency of concurrent immunologic diseases, histologic features, and laboratory findings to patients with classic AIH.53 Furthermore, they have similar frequencies of HLA-B8, HLA-DR3, and HLA-A1-B8-DR3,53 and they respond as well to glucocorticoid treatment as do their autoantibody-positive counterparts.53,54 These persons probably have a form of AIH that has escaped detection by conventional serologic assays, and they are candidates for a closely monitored treatment trial of glucocorticoids. Assays for atypical pANCA, anti-SLA, and IgA endomysial antibodies or antibodies to tissue transglutaminase occasionally yield positive results in these patients,6,7 and successive testing for conventional autoantibodies may demonstrate the late appearance of typical autoimmune markers in some cases (see Table 88-4).55

VARIANT WITH CHOLESTATIC FEATURES

Eight percent of patients with AIH have histologic features of bile-duct injury and laboratory changes of cholestasis in the absence of AMA and cholangiographic changes of largeduct PSC.42,52 These individuals may have AMA-negative PBC, small-duct PSC, or a separate syndrome (“autoimmune cholangitis”) (see Table 88-4 and Chapter 89).42,52 This variant is probably a heterogeneous category that encompasses patients with atypical, early, or transitional features of classic disease.52 Persons who have this variant are inconsistently responsive to glucocorticoids and ursodeoxycholic acid.42,52 Preliminary experience suggests that these therapies can help to improve the clinical and laboratory abnormalities but not the histologic changes. Histologic findings of bile duct injury, including destructive cholangitis, may be seen in classic AIH in the absence of other cholestatic features.4,5 These changes are transient and probably repre-

AUTOIMMUNE HEPATITIS AND CHRONIC HEPATITIS C Eight percent of white North American adults with classic AIH have concurrent infection with hepatitis C virus (HCV), and 52% of patients with chronic hepatitis C have autoantibodies, concurrent immune diseases, or both.56 Identification of these patients as belonging to one or the other group is important because interferon therapy can enhance the immune manifestations of persons with AIH and concurrent HCV infection, and immunosuppressive treatment can increase serum viral levels in persons with chronic hepatitis C and background autoimmune features. The nature and degree of the associated immune mani­ festations distinguish AIH with coincidental HCV infection

1467

1468

Section IX  Liver

Autoimmune predominant

Autoab titers ≥1:320 Multiple autoab species Autoag-driven disease Interface hepatitis Plasma cells

Viral predominant

Autoab titers <1:320 Single autoab species Immune complex disease Lymphoid aggregates Steatosis

Figure 88-5.  Key features in determining the predominant autoimmune or viral nature of autoimmune hepatitis in patients with with concurrent hepatitis C infection. Autoimmune-predominant disease is characterized by high-titer (≥1 : 320) autoantibodies (Autoab), multiple autoantibody species, concurrent autoantigen (Autoag)-driven immune disease (autoimmune thyroiditis, Graves’ disease, or ulcerative colitis), and histologic features of interface hepatitis and portal plasma cell infiltration. Viruspredominant disease is charac­terized by low-titer autoantibodies, single autoantibody expression, concurrent immune complex diseases, and histologic features of portal lymphoid aggregates and hepatic steatosis. Immune complex diseases may include vasculitis, glomerulonephritis, and symptomatic cryoglobulinemia.

from chronic hepatitis C with autoimmune features (Fig. 88-5).56,57 Concurrent immune diseases that reflect a cellmediated response against autoantigens (autoimmune thyroiditis, Graves’ disease, inflammatory bowel disease) typify an autoimmune process and occur more commonly in AIH than in chronic hepatitis C.56,57 They also are more likely to have clinical consequences if the immune diseases are exacerbated during antiviral therapy. Serum autoantibody titers can increase during antiviral treatment, but this increase is neither pathogenic nor associated with clinical deterioration. Patients with chronic hepatitis C and concurrent autoimmune features have immune complex diseases (vasculitis, glomerulonephritis, and symptomatic cryo­ globulinemia) more frequently than do patients with classic AIH, and these findings support a predominantly viral disease. The immune manifestations are driven mainly by viral antigen.56 Patients with classic AIH typically express multiple antibodies in titers that exceed 1 : 320.56 By contrast, patients with chronic hepatitis C typically express one type of autoantibody (ANA or SMA) in titers that are usually less than 1 : 320. Furthermore, patients with AIH have higher serum levels of AST, gamma globulin, and IgG than do patients with chronic hepatitis C. High serum titers (greater than 1 : 320) of autoantibodies, marked hypergammaglobulinemia, multiple autoantibody types, or the presence of autoantigen-driven concurrent immune diseases indicates an autoimmune-predominant syndrome. Further differentiation requires liver biopsy evaluation. Patients with classic AIH more commonly have severe interface hepatitis, moderate to severe portal plasma cell

infiltration, and panacinar hepatitis in liver biopsy specimens than do patients with typical chronic hepatitis C.57 By contrast, patients with typical chronic hepatitis C have a higher frequency of portal lymphoid aggregates and steatosis than is seen in patients with classic AIH (see also Chapter 79). Histologic diagnoses based on these predominant patterns have high specificity (81% and 91%) and predictability (62% and 82%) for AIH and chronic hepatitis C, respectively.58 Their sensitivity for each clinical diagnosis, however, is low (40% and 57%, respectively). Treatment of patients with mixed features of AIH and chronic hepatitis C should be appropriate for the predominant disease and must be based on the nature of concurrent immune diseases, number and titers of associated autoantibodies, and histologic pattern. Combined therapies with glucocorticoid and antiviral drugs should be avoided.

EPIDEMIOLOGY The incidence of AIH among white northern Europeans is 1.9 cases per 100,000 persons per year, and its point prevalence is 16.9 cases per 100,000 persons per year.59 In the United States, AIH affects 100,000 to 200,000 persons and accounts for 2.6% of the transplantations in the European Liver Transplant Registry and 5.9% in the National Institutes of Health Liver Transplantation Database. The frequency of AIH among patients with chronic liver disease in North America is between 11% and 23%. The impact of genetic risk factors must be considered in assessing the occurrence of disease in different regions. The prevalence of AIH is greatest among northern European white persons who have a high frequency of HLA-DR3 and HLA-DR4, and AIH is found with similar frequency in the derivative populations of North America and Australia. The Japanese have a low frequency of HLA-DR3, and AIH in Japan is associated with HLA-DR4.20,21 All populations are susceptible to AIH, which has been described in African Americans, South Americans, Arabs, Japanese, mainland Chinese, and subcontinental Indians. The prevalence of AIH among Alaskan natives (43 per 100,000 population) is higher than that reported in a white Norwegian population (16.9 per 100,000).60

PROGNOSIS INDICES The prognosis for AIH relates mainly to the severity of liver inflammation at the initial medical consultation, as reflected in the laboratory indices and the histologic findings. HLA status and ethnicity influence disease occurrence, clinical phenotype, and treatment outcome (Table 88-5). Certain serologic markers have shown promise as pro­gnostic instruments, and the Model for End-stage Liver Disease (MELD) score may be useful for identifying problematic patients early.

LABORATORY INDICES

Serum AST and gamma globulin levels reflect the severity of disease and immediate prognosis. Sustained severe elevations indicate a poor outcome unless therapy is started. Less severe laboratory abnormalities are associated with a better prognosis (see Table 88-5).1 Mild AIH may still progress, and the laboratory indices reflective of mild inflammatory activity are unable to predict disease behavior in individual patients.61

Chapter 88  Autoimmune Hepatitis Table 88-5  Prognostic Factors and Associated Reported Outcomes in Autoimmune Hepatitis Outcome Prognostic Factor

CIRRHOSIS

POOR TREATMENT RESULT

MORTALITY

Common

—

49% at 15 yr 17% at 5 yr 82% at 5 yr — — — — —

— — — — 97% treatment failure 100% relapse Relapse common Relapse common Histologic activity 27% treatment failure 38% poor result 5% treatment failure

50% at 3 yr 90% at 10 yr 10% at 10 yr Normal at 5 yr 45% at 5 yr 58% at 5 yr — — — —

Before Treatment AST ≥ 10-fold ULN or AST ≥ 5-fold ULN and serum gamma globulin ≥ 2-fold ULN AST <10-fold ULN and serum gamma globulin <2-fold ULN Interface hepatitis Bridging necrosis or multilobular necrosis Cirrhosis Model for End-stage Liver Disease score ≥12 points Anti-soluble liver antigen* Anti-asialoglycoprotein receptor* Anti-chromatin* HLA-DR3 or DRB1*0301

—

HLA-DR4 or DRB1*0401 After Treatment Multilobular necrosis and failure of pretreatment hyperbilirubinemia to improve after 2 wk Failure to enter remission and first sign of decompensation within 4 yr (e.g., ascites) Male gender, cirrhosis ≥10 yr, portal hypertension, continuous treatment >3 yr, or previous treatment failure

— — —

— — —

—

High mortality

—

—

High mortality

—

Hepatocellular cancer

—

*Autoantibodies with an asterisk are investigational only and not available for routine clinical use. AST, serum aspartate aminotransferase level; HLA, human leukocyte antigen; ULN, upper limit of normal.

Spontaneous resolution is possible in 13% to 20% of patients regardless of disease activity. No features predict this outcome, and patients should not be managed with this expectation.61 Of persons who survive the early, most active stage of the disease, inactive cirrhosis develops in 41%. Patients who receive no treatment and who have initially severe disease and survive the first two years of illness typically survive long term.

HISTOLOGIC FINDINGS

The histologic findings at presentation also are indices of disease severity, and each pattern of liver cell injury has its own prognostic implication (see Table 88-5).1 Esophageal varices develop in 54% of patients with cirrhosis, and death from variceal hemorrhage occurs in 20% of those with varices if treatment of AIH is not instituted. Hepatocellular carcinoma (HCC) also can occur in patients with cirrhosis. The overall Kaplan-Meier 10-year probability of HCC is 2.9%, and the overall incidence of HCC is four cases per 1000 patient-years. The risk of HCC is small and mainly affects males and patients with longstanding cirrhosis (>10 years).62

HUMAN LEUKOCYTE ANTIGEN STATUS

White North American and northern European patients with HLA-DR3 are typically younger than patients with other HLA alleles and have more active disease (see Table 88-5).63,64 Whites with HLA-DR3 (DRB1*0301) respond less well to glucocorticoids than do patients with HLA-DR4 (DRB1*0401), whereas those with HLA-DR4 (DRB1*0401) have concurrent immune diseases more commonly and better outcomes than those with HLA-DR3 (DRB1*0301).63,64 By contrast, DRB1*1501 protects against AIH in white North Americans and northern Europeans. Other ethnic groups and individuals in other geographic regions may have different genetic predispositions that reflect diverse etiologic agents or antigenic exposures. At present, the clinical appli-

cations of HLA testing in AIH are uncertain, and determinations are not made routinely in practice.

ETHNICITY

Ethnicity may affect disease severity as well as presentation.65 Cirrhosis is present at accession more commonly, and hepatic synthetic function is decreased more frequently,66 in black North American patients with AIH than in white North American patients (85% vs. 38%). Both groups respond similarly to glucocorticoids, but black North American patients are younger at presentation than their white counterparts. These findings suggest that black North Americans have more aggressive disease than is seen in white North Americans and that their higher frequency of advanced disease reflects intrinsic disease behavior. Studies in Alaskan natives, Turks, Japanese, South Americans, non-European, non-white patients, and Somalians also have emphasized clinical and prognostic differences among the racial groups. Alaskan natives have a higher frequency of acute icteric disease, asymptomatic illness, and advanced fibrosis at presentation than is characteristic of their white counterparts.60 Japanese patients lack the DRB1*0301 allele, which has been associated with disease occurrence, early age at onset, and poor treatment response in white populations.64 The clinical phenotype and treatment requirements in Japanese patients with AIH differ from those of their white counterparts. They typically have mild, late-onset disease that may respond to nonsteroidal medications such as ursodeoxycholic acid. South American patients in Brazil and Argentina are younger than their North American white counterparts.23,24 They have more severe laboratory abnormalities, and DRB1*1301 is their principal susceptibility allele. African, Asian, and Arab patients also have an earlier age at onset than their white northern European counterparts and a higher frequency of cholestatic laboratory findings, greater occurrence of biliary changes on histologic examination,

1469

1470

Section IX  Liver and poorer initial response to standard therapy than are found for white patients or other ethnic groups.67 Interwoven into the natural history of AIH in each ethnic population and geographic region are cultural and socioeconomic factors that may affect the time to diagnosis and access to treatment.

SEROLOGIC MARKERS

Emerging serologic tests for AIH reflect the goal of identifying markers that have prognostic as well as diagnostic value (see Table 88-5).6,7 Antibodies to SLA satisfy this role with a specificity of 99% for AIH and a strong association with severe disease and relapse after glucocorticoid withdrawal.11,12 The target autoantigen of anti-SLA is a 50-kd protein that may be a transfer ribonucleoprotein complex (tRNP(Ser)Sec) involved in the incorporation of selenocysteine into peptide chains.68 The expression of anti-SLA is closely associated with HLA-DR3, and antibodies to SLA may be surrogate markers of a genetic propensity for severe disease and relapse after drug withdrawal.11,12 Testing for anti-SLA may also be useful in reclassifying patients with cryptogenic chronic hepatitis.6,7 Antibodies to ASGPR are closely associated with histologic activity and also identify individuals who relapse after drug withdrawal possibly because of residual hepatic inflammation.6,7,13 Antibodies to chromatin and antibodies to actin are each being investigated as prognostic markers and can also identify patients who relapse after drug withdrawal (anti-chromatin) or die of liver failure or require liver transplantation (anti-actin).6,7,10 Performance parameters of each marker and standardi­ zation of individual assays are still required. The major limitations of most prognostic markers are the low frequency of occurrence in patients with poor outcomes and their inability to exclude a poor result. Antibodies to SLA, which are the strongest contenders as prognostic markers, are found in only 16% of white North American patients with AIH.

MODEL FOR END-STAGE LIVER DISEASE SCORE

The MELD score is an established method by which to predict early mortality associated with severe liver disease; it is based on the serum creatinine level, total serum bilirubin concentration, and international normalized ratio (INR).69 Survival predicted by the MELD score relates mainly to the degree of impaired liver function and not the presence of cirrhosis or the cause of liver disease. In AIH, a MELD score of ≥12 points at presentation identifies 97% of patients who fail to respond to glucocorticoid treatment. The specificity of this score for treatment failure is 68%, and its positive predictive value is 19% (see Table 88-5).69 Sensitivity for treatment failure is preferred over specificity in order to capture at presentation all patients at risk who can then be monitored more closely or considered for salvage therapies at the earliest sign of deterioration.

CLINICAL FEATURES The clinical features of AIH frequently reflect the inflammatory activity of the liver disease or the complications of cirrhosis (Table 88-6).1,7 Cholestatic features may be present but do not dominate the clinical picture. Similarly, manifestations of liver decompensation, such as ascites, hepatic encephalopathy, and variceal bleeding, are uncommon findings at the initial medical consultation. Ready fatigability is the most common symptom (seen in 86% of adult patients) (see Table 88-6). Weight loss is

Table 88-6  Clinical Features of Autoimmune Hepatitis OCCURRENCE (%) Symptoms Fatigue Jaundice Upper abdominal discomfort Pruritus (mild) None (at presentation) Anorexia Myalgias Diarrhea Cushingoid features Fever (≤40°C) Physical Findings Hepatomegaly Jaundice Spider angiomata Concurrent immune disease Splenomegaly None Ascites Encephalopathy Laboratory Features Elevated aspartate aminotransferase level Hypergammaglobulinemia Increased immunoglobulin G level Hyperbilirubinemia Alkaline phosphatase level ≥ 2-fold normal Immunoserologic Markers* SMA, ANA, or anti-LKM1 Atypical pANCA Anti-asialoglycoprotein receptor* Anti-actin* Anti-chromatin* Anti-liver cytosol 1* Anti-soluble liver antigen*

86 77 48 36 25-34 30 30 28 19 18 78 69 58 ≤38 ≥32 <25 20 14 100 92 91 83 33 100 92 (type 1 AIH only) 82 74 42 (ANA+ only) 32 (type 2 AIH only) 11-17

*Autoantibodies with an asterisk are investigational or are not available for routine clinical use. AIH, autoimmune hepatitis; ANA, antinuclear antibodies; anti-LKM1, antibodies to liver-kidney microsome type 1; pANCA, perinuclear anti-neutrophil cytoplasmic antibodies; SMA, smooth muscle antibodies.

uncommon, and intense pruritus argues against the diagnosis. Hepatomegaly is the most common physical finding (78%), and jaundice is found in 69% of patients. Splenomegaly can be present in patients with and without cirrhosis (56% and 32%, respectively), as can spider angiomata. As many as 34% of patients are asymptomatic at initial consultation, and 25% of adults have a normal physical examination.7 The discordance between the severity of inflammatory activity and the presence of symptoms is most common in children, whose clinical status frequently does not accurately reflect the severity of the underlying liver disease. Hyperbilirubinemia is present in 83% of patients, but the serum bilirubin level is greater than three times the upper limit of normal in only 46%.45 Similarly, the serum alkaline phosphatase level commonly is increased (81%), but elevations are two times the upper limit of normal in 33% and four times the upper limit of normal in only 10% of patients (see Table 88-6).45 The hypergammaglobulinemia of AIH is polyclonal; the IgG fraction predominates. Paraproteins are common, and patients may have diverse, nonspecific serologic findings, including antibodies to bacteria (Escherichia coli, Bacteroides, and Salmonella species) and viruses (measles virus, rubella virus, and cytomegalovirus). Cryoglobulinemia may be present, but symptomatic cryoglobulinemia is rare.

Chapter 88  Autoimmune Hepatitis Concurrent immunologic diseases are common and involve diverse organ systems, most frequently the thyroid.7 SMA, ANA, and anti-LKM1 are required for the diagnosis, and other autoantibodies may be present, as shown in Table 88-6,7 although these other autoantibodies do not have routine clinical applications.

SEVERE OR FULMINANT PRESENTATION

AIH can have an acute severe or fulminant presentation that can be mistaken for acute viral or toxic hepatitis, and misdiagnosis can delay or defer the institution of potentially life-saving therapy.2 Glucocorticoid therapy can be effective in suppressing the inflammatory activity in 36%-100% of patients, whereas delay in treatment can have a strong negative impact on outcome. Furthermore, unrecognized chronic disease can have a spontaneous exacerbation and appear acute. Affected patients invariably die if they are untreated or there is no response to glucocorticoid therapy within two weeks.70

ADVANCED AGE

Twenty percent of adults with AIH develop the disease after the age of 60 years, and they may be under-recognized because the liver disease is ascribed to the toxicity of medication or to concurrent illnesses that may also affect liver biochemical test levels.32,71 These elderly patients have a greater degree of hepatic fibrosis at presentation than young adults, and their higher frequencies of ascites and cirrhosis indicate that they have an aggressive disease that is commonly indolent and unsuspected.71 Enthusiasm for therapy is justifiably tempered by concerns regarding pre-existing osteopenia or other comorbid conditions. Elderly patients, however, typically respond well to glucocorticoid treatment, and preemptive efforts to maintain bone density with bisphosphonates, calcium and vitamin D supplements, and a regular exercise program can improve tolerance to the treatment.32,71 Elderly patients do have higher cumulative frequencies of drug-related complications and vertebral compression fractures than do young patients, but these consequences are associated with protracted durations of treatment or repeated relapses and re-treatment. Therapy in the elderly must be individualized to limit the duration of initial treatment, total cumulative dose of glucocorticoids, and re-treatment after relapse.32

ASYMPTOMATIC PATIENTS

AIH can be asymptomatic in 34% to 45% of patients.72,73 Typically, affected patients are men, and they have significantly lower serum alanine aminotransferase levels at presentation than do symptomatic patients. Histologic findings, including the frequency of cirrhosis, are similar in asymptomatic and symptomatic patients, and as many as 70% of asymptomatic patients become symptomatic during the course of their disease. The asymptomatic state does not preclude the need for glucocorticoid therapy if other objective manifestations of disease activity are present or emerge, and close surveillance of these patients for worsening inflammation is justified.

ASSOCIATED DISEASES

As many as 38% of patients with AIH have concurrent immune diseases, which may mask the underlying liver disease.1,7 Autoimmune thyroiditis, Graves’ disease, synovitis, and ulcerative colitis are the most common immunemediated disorders associated with AIH in North American adults, whereas insulin-dependent diabetes mellitus, vitiligo, and autoimmune thyroiditis are the most common concurrent disorders in European children with AIH and

anti-LKM1.34 Autoimmune sclerosing cholangitis can be present with or without inflammatory bowel disease in children with AIH,51 and PSC can be detected in 8% of adults with classic AIH50 and 44% of those with AIH and inflammatory bowel disease.33 Concurrent immune diseases typically do not affect the prognosis of AIH unless the underlying liver disease is unrecognized or associated with bile duct changes.33,50 Patients presenting with immune-mediated rheumatic diseases, celiac disease, or inflammatory bowel disease should be assessed for underlying AIH.7

HEREDITARY FORMS

Autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED) is a syndrome caused by a single-gene mutation localized to chromosome 21q22.3 that affects the generation of an autoimmune regulator (AIRE) involved in the negative selection of autoreactive immunocytes in the thymus.74 The syndrome is characterized by multiple endocrine organ failure (parathyroids, adrenals, or ovaries), mucocutaneous candidiasis, and ectodermal dysplasia in various combinations that may include AIH. Unlike other autoimmune diseases, APECED has a recessive mendelian pattern of inheritance, complete penetrance of the gene, no HLA-DR associations, and no female predominance. Patients who have APECED and AIH may have particularly aggressive liver disease that does not respond well to standard immunosuppressive regimens.74

TREATMENT INDICATIONS

The indications for treatment of AIH are shown in Table 88-7 and are based on manifestations of hepatic inflammation.1 The prognosis of untreated mild AIH is unclear, and the decision to treat patients with less than severe AIH is a highly individualized clinical judgment that reflects the perceived risk of the untreated disease, possible side effects of therapy, and patient’s physical and mental tolerance of the disease and its management. Untreated asymptomatic mild AIH may still progress, and there are no clinical indices that predict the course of disease. Most patients with AIH are treated regardless of disease activity at presentation, because of physicians’ uncertainty about the benign nature of mild or asymptomatic disease and the likelihood that the disease will remain mild long-term.61

REGIMENS

Prednisone, alone or at a lower dose in combination with azathioprine, is effective therapy (Table 88-8).1,75 No findings at presentation preclude a satisfactory response to therapy. The presence of ascites or hepatic encephalopathy identifies patients with a poor prognosis, but these findings do not preclude a full response to glucocorticoid therapy.70 Decompensated patients with multilobular necrosis on histologic examination in whom at least one laboratory parameter fails to normalize, or in whom hyperbilirubinemia does not improve, after two weeks of treatment have a high immediate mortality rate. These patients should be evaluated for liver transplantation (see Table 88-5).70,76 Patients in whom these parameters improve during the first two weeks of therapy have excellent immediate survival rates, and their drug treatment should be continued.70

DRUG ACTIONS

Glucocorticoids limit T cell activation by inhibiting cytokine production and the expression of adhesion mole-

1471

1472

Section IX  Liver Table 88-7  Indications for Treatment in Autoimmune Hepatitis Indications FINDINGS

ABSOLUTE

RELATIVE

NONE

Clinical

Incapacitating symptoms Relentless clinical progression

Mild or no symptoms

Laboratory

AST ≥10-fold ULN AST ≥5-fold ULN and gamma globulin ≥ 2-fold ULN Bridging necrosis Multilobular necrosis

AST 3- to 9.9-fold ULN AST ≥ 5-fold ULN and gamma globulin <2-fold ULN Interface hepatitis

Asymptomatic with minimal laboratory changes Previous intolerance of prednisone and/or azathioprine AST <3-fold ULN Severe cytopenia

Histologic

Inactive cirrhosis Portal hepatitis Decompensated cirrhosis with variceal bleeding

AST, serum aspartate aminotransferase level, ULN, upper limit of normal.

Table 88-8  Preferred Treatment Regimens in Autoimmune Hepatitis Single-Drug Therapy

Combination Therapy Prednisone (mg/day)

Azathioprine (mg/day)

Prednisone (mg/day)

30 mg × 1 wk 20 mg × 1 wk 15 mg × 2 wk 10 mg until end point

50 mg until end point

60 mg 40 mg 30 mg 20 mg

cules.31,77 Glucocorticoids are lipophilic and can diffuse into the cytosol of cells to bind the glucocorticoid receptor. The complex of drug and receptor then translocates to the nucleus where it inhibits cytokine gene expression and reduces the activity of nuclear factor-κB. Type 1 and type 2 cytokine pathways are affected, and both cellular and humoral immune responses are blunted. Repeated administration of glucocorticoids is required to achieve results in AIH because of a short biological half-life. Azathioprine is a purine antagonist that blocks the proliferation of lymphocytes.77 It is converted to 6-mercaptopurine in blood by a nonenzymatic, glutathione-based pathway, and this metabolite is, in turn, converted to 6-thioguanine by hypoxanthine guanine phosphoribosyl transferase. The 6-thioguanines are the active metabolites that interfere with purine nucleotide synthesis within the cell cycle, and they thereby impair proliferation of rapidly dividing T and B lymphocytes. Competing enzymatic pathways can convert 6-mercaptopurine to either 6-thiouric acid by xanthine oxidase or 6-methyl mercaptopurine by thiopurine methyltransferase (TPMT). Each end product is inactive, and the integrity of the enzymatic routes responsible for their production influences the erythrocyte concentrations of the active 6-thioguanine nucleotides. Drugs that inhibit xanthine oxidase activity, such as allopurinol, or deficiencies in TPMT activity can increase the therapeutic efficacy or the toxicity of the 6-thioguanine metabolites, respectively. Blood TPMT activity is also inducible by azathioprine, and this effect may further modulate the erythrocyte concentrations of the 6-thioguanine nucleosides. The immunosuppressive action of azathioprine is achieved slowly (three months or longer) because of the drug transformations and nuclear incorporations that are required to limit immunocyte proliferation. Azathioprine has weak or no immunosuppressive effect in AIH unless combined with prednisone.

× 1 wk × 1 wk × 2 wk until end point

DRUG-RELATED SIDE EFFECTS

Cosmetic changes, such as facial rounding, dorsal hump formation, obesity, acne, striae, alopecia, and facial hirsutism, occur in 80% of patients after two years of treatment regardless of the regimen used.77 Severe side effects include osteopenia with vertebral compression, diabetes mellitus, cataracts, emotional instability, opportunistic infections, pancreatitis, and hypertension. Severe complications are uncommon and develop only after protracted therapy (more than 18 months) and on the regimen with the higher dose of prednisone (20 mg per day). Azathioprine with prednisone is preferred to a higher dose of prednisone alone because the combination produces fewer glucocorticoidrelated side effects during comparable periods of treatment (10% vs. 44%). Treatment must be discontinued prematurely in 13% of patients, mainly because of intolerable obesity, cosmetic changes, brittle diabetes mellitus, or osteoporosis with vertebral compression (Table 88-9). Treatment with azathioprine can be complicated by cholestatic liver disease, nausea, emesis, rash, pancreatitis, arthralgias, opportunistic infections, and cytopenias.77 Five percent of patients treated with azathioprine develop early adverse reactions (nausea, vomiting, arthralgias, fever, skin rash, or pancreatitis), which warrant its discontinuation. The overall frequency of azathioprine-related side effects in patients treated with 50 mg daily is 10%, and the side effects typically improve after the dose is reduced or the therapy is discontinued.77 Cytopenia is the most common consequence of treatment; bone marrow failure is rare. Cytopenia occurs in 46% of patients, and severe hematologic abnormalities occur in 6%.77,78 These toxicities are not predictable by either genotyping or phenotyping for TPMT activity,78-80 and the most common association with cytopenia in these patients is cirrhosis and presumed hypersplenism associated with portal hypertension (see Table 88-9).78,80

Chapter 88  Autoimmune Hepatitis Table 88-9  Side Effects Associated with Therapy for Autoimmune Hepatitis* Prednisone

Azathioprine

TYPE

FREQUENCY

TYPE

FREQUENCY

Cosmetic (usually mild) Alopecia Dorsal hump Facial rounding Hirsutism Striae Weight gain Somatic (severe) Cataracts Diabetes mellitus Emotional instability Hypertension Osteopenia Vertebral compression Inflammatory/neoplastic Malignancy Opportunistic infection Pancreatitis

80% (after 2 yr)

Hematologic (mild) Cytopenia Hematologic (severe) Leucopenia Thrombocytopenia Somatic (usually mild) Arthralgias Fever Nausea and vomiting Rash Neoplastic Nonhepatic cell types Hematologic/enteric Bone marrow failure Villus atrophy and malabsorption Teratogenic during pregnancy

46% (especially with cirrhosis)

13%

Rare

6% 5%

3% (after 10 yr) Rare Rare (theoretical)

*Adapted from Czaja AJ. Safety issues in the management of autoimmune hepatitis. Expert Opin Drug Saf 2008; 7:319-33.

Teratogenicity is a theoretical complication of therapy with azathioprine.77 Azathioprine has been administered successfully in pregnant women with AIH, pregnant mothers with inflammatory bowel disease, and women who have conceived while taking azathioprine after liver transplantation.77 Nevertheless, azathioprine has been associated with congenital malformations in pregnant mice, and these defects have included cleft palate, skeletal anomalies, hydrops fetalis, reduced thymic size, anemia, and hematopoietic depression. Furthermore, the placenta is only a partial barrier to the metabolites of azathioprine, and low levels of 6-thioguanine are detectable in the newborns of mothers treated for Crohn’s disease.77 The odds ratio of having a child with congenital malformations while taking azathioprine for inflammatory bowel disease is 3.4. Azathioprine is not an essential medication in the treatment of AIH and can be discontinued during pregnancy, in which case the liver disease can be managed successfully by adjustments in the dose of prednisone. Oncogenicity is another possible complication of therapy with azathioprine.77 The incidence of extrahepatic neoplasms is 1 per 194 patient-years; the probability of tumor occurrence is 3% after 10 years; and the risk of malignancy is 1.4-fold greater than normal.81 The low but increased risk of malignancy does not contraindicate azathioprine therapy in AIH but emphasizes the importance of maintaining strict indications for treatment. Blood TPMT activity is significantly lower in patients with AIH and intolerance to azathioprine than in patients with uncomplicated courses of treatment.79 Similar findings have been described in patients with inflammatory bowel disease and rheumatic conditions (see Chapter 111). These observations have suggested that routine screening of the blood TPMT level may identify persons with AIH at risk for azathioprine-related complications. The blood TPMT activity level, however, has not been predictive of toxicity in individual patients.79,80 Genotypic and phenotypic screening for blood TPMT activity has not reduced the frequency of azathioprine-induced side effects in patients with AIH compared with unscreened patients,78-80 nor has the occurrence of azathioprine-related side effects been associated

with below-normal levels of TPMT activity.78 Near-zero enzyme activity occurs rarely in otherwise normal indivi­ duals (0.3% to 0.5%), and the value of screening to detect this unusually low enzyme deficiency remains uncertain, especially if some patients with low levels do not exhibit azathioprine toxicity.77,78 Testing for blood TPMT activity seems most appropriate in patients with pre-existing or progressive cytopenias and in those subjected to doses of azathioprine higher than the conventional schedule of 50 mg daily. Avoidance of azathioprine in patients with pre-existing or progressive cytopenias (leukocyte count below 2.5 × 109/L or platelet count below 50 × 109/L) and close monitoring (at three-month intervals) of the blood leukocyte and platelet counts in all patients taking the drug may be the best preventive strategy.77,78

END POINTS

Glucocorticoid therapy is continued until remission, treatment failure, incomplete response, or drug toxicity occurs (Fig. 88-6).1,75 Remission connotes absence of symptoms, resolution of all laboratory indices of active inflammation, and histologic improvement to normal liver tissue or inactive cirrhosis.82 Evaluation of liver tissue before drug withdrawal is essential to establish remission because histologic activity may be present in 55% of patients who satisfy other requirements for remission. Typically, histologic improvement lags behind clinical and laboratory resolution by three to eight months, and treatment should be extended for at least this period to compensate for the lag. Patients with normal serum AST and gamma globulin levels and normal liver biopsy findings immediately before drug withdrawal have a significantly lower frequency of relapse after drug withdrawal than patients with near-normal test levels and histology (60% vs. 90%, P < 0.001).82 Only 40% of treated patients, however, are able to improve fully, and complete resolution of the clinical and histologic manifestations of the disease does not preclude relapse after drug withdrawal.82 Treatment failure connotes deterioration during therapy (see Fig. 88-6).1,69,75 It is characterized by worsening of the serum AST or bilirubin levels by at least 67% of previous

1473

1474

Section IX  Liver Prednisone + azathioprine or Prednisone alone (higher dose)

Remission

Treatment failure

No symptoms AST ≤2x normal Gamma globulin normal Normal or minimal histologic activity

Rise in AST and/or bilirubin level Worsening histologic activity

Drug toxicity

Improvement in all aspects Failure to satisfy remission criteria

Intolerable symptoms Vertebral compression Cytopenia

Prolonged therapy (≥3 years)

Dose reduction or withdrawal

High-dose therapy

No therapy

Sustained remission

Incomplete response

Relapse

Stable liver function

Decompensation

Chronic therapy

Liver transplantation

Indefinite low-dose prednisone or azathioprine

Maintenance therapy

Indefinite low-dose prednisone

or

Indefinite azathioprine

Figure 88-6.  Treatment algorithm for autoimmune hepatitis. Patients who satisfy indications for glucocorticoid therapy are given prednisone in combination with azathioprine or a higher dose of prednisone alone (see Table 88-8). Treatment is continued until the criteria for a treatment end point are met. Possible end points are remission, treatment failure, incomplete response, and drug toxicity. Therapy can then be discontinued, increased in dose, or reduced in dose. Responses to the dose adjustments determine the need for other actions. The principal indices of inflammatory activity are serum aspartate aminotransferase (AST) and gamma globulin levels.

values, progressive histologic activity, or onset of ascites or encephalopathy. Conventional glucocorticoid therapy should be stopped, and a high-dose regimen should be instituted. Incomplete response connotes improvement that is insufficient to satisfy remission criteria (see Fig. 88-6).1,75 Failure to achieve remission within three years indicates that remission is unlikely and warrants discontinuation of conventional treatment. Drug toxicity justifies premature withdrawal of medication or a reduction in dose (see Fig. 88-6).1,75,77 Most side effects are reversible, and some consequences such as cataracts and osteopenia with vertebral compression have effective therapies. Weight gain, acne, edema, and diabetes may be consequences of the disease rather than the drugs.

RESULTS

Prednisone alone or in combination with azathioprine induces clinical, biochemical, and histologic remission in 65% of patients within three years.1,75 The average treatment interval until remission is 22 months. Therapy improves survival. The 10-year life expectancies following treatment for patients with and without cirrhosis at the time of the initial medical consultation are 89% and 90%, respectively.

The overall 10-year survival rate is 93% and is comparable to that of an age- and sex-matched cohort from the population at large (94%). Patients who have histologic cirrhosis respond as well as do noncirrhotic patients and should receive similar treatment, with the same expectation of success. Twenty-one percent of patients managed with conventional regimens of prednisone alone or in combination with azathioprine sustain their remission for a median of 76 months after drug withdrawal, and an effort should be made to withdraw all patients from initial therapy after criteria for remission have been satisfied.83 Treatment with glucocorticoids also may reduce hepatic fibrosis.84 Fibrosis scores have improved in 56% of patients followed for 55 ± 9 months, and fibrosis did not progress in 33% of patients followed for 62 ± 14 months. Histologic activity indices decreased concurrently, and patients in whom the histologic activity indices improved had a higher frequency of improvement in the fibrosis scores (80% vs. 25%, P = .002). These findings suggest that improvement in hepatic fibrosis occurs in conjunction with reduction in liver inflammation and that glucocorticoids may facilitate the disappearance of fibrosis by suppressing inflammatory activity.30,84 Small case studies have also suggested that cirrhosis can disappear during treatment, but this possibility must await confirmation by assessments more reliably

Chapter 88  Autoimmune Hepatitis reflective of cirrhosis than conventional needle biopsy of the liver.

least one year. This effort can be repeated if earlier withdrawal attempts have failed.83

RELAPSE

TREATMENT FAILURE

Patients who enter remission commonly experience an exacerbation after drug withdrawal.1,75,83 Relapse is defined as the reappearance of histologic disease after discontinuation of drug therapy. An increase in the serum AST level to more than three-fold the upper limit of normal after discontinuation of drug therapy is invariably associated with interface hepatitis on histologic examination. This biochemical change is sufficient to diagnose relapse without requiring liver tissue assessment. Relapse occurs in 50% of patients within six months of discontinuing therapy, and most patients (70% to 86%) experience an exacerbation within three years.82,83 Reinstitution of the original treatment induces another remission, but relapse commonly recurs after termination of therapy. Repeated relapse and re-treatment is associated with a cumulative morbidity and mortality. Cirrhosis develops more commonly (38% vs. 4%, P = 0.004); death from hepatic failure or need for liver transplantation occurs more often (20% vs. 0%, P = 0.008)85; and drug-induced side effects are more frequent (70% vs. 30%, P = 0.01) in persons who relapse than in those who sustain remission after drug withdrawal.77 The frequencies of each complication increase with each subsequent relapse and re-treatment. The optimal time to interrupt this sequence is after the first course of treatment and relapse, and the preferred treatment strategy is to implement maintenance therapy with azathioprine.86 After the initial relapse, therapy with prednisone and azathioprine is re-started and continued until clinical and laboratory resolution is achieved again (see Fig. 88-6).86 The dose of azathioprine is then increased to 2 mg/kg daily as the dose of prednisone is reduced. Azathioprine is then continued indefinitely as a chronic maintenance regimen. Eighty percent of patients are able to sustain remission in this fashion over a 10-year period of observation. Azathioprine-induced cytopenias compel dose reduction in 9% of patients; glucocorticoid-related side effects improve; and arthralgias associated with glucocorticoid withdrawal eventually resolve (but may be protracted). Malignancies of various types develop in 7% of patients, but an association with azathioprine is disputed.86 An alternative management strategy after the first relapse is to maintain suppression of inflammatory activity using daily, low-dose prednisone (see Fig. 88-6).87 Eighty-seven percent of patients can be managed long term on prednisone at less than 10 mg per day (median dose, 7.5 mg per day). The dose is titrated to the lowest level needed to prevent symptoms and to maintain serum AST levels below three times the upper limit of normal. Side effects attributable to previous glucocorticoid therapy resolve in 85% of patients; the immediate survival rate is comparable to that for persons who have relapsed and been treated with a conventional regimen of prednisone alone or a lower dose of prednisone and azathioprine (91% vs. 90%), and new complications do not occur. The azathioprine and low-dose prednisone maintenance regimens have not been compared directly, but the azathioprine schedule has intuitive appeal because it eliminates the need for prednisone. Treatment after relapse does not need to be indefinite.83 The probability of ultimately achieving inactive disease that does not require continuous drug therapy is 28%. This possibility justifies an effort to withdraw treatment in all patients who have relapsed previously and who have stably maintained inactive disease on maintenance therapy for at

Nine percent of patients deteriorate during glucocorticoid therapy (treatment failure) (see Fig. 88-6).1,69,75 High doses of prednisone alone (60 mg per day) or prednisone (30 mg per day) in conjunction with azathioprine (150 mg per day) constitute standard treatment in this group.1,75,88 Each schedule induces clinical and biochemical improvement in 70% of patients within two years. Histologic resolution, however, occurs in only 20%, and long-term therapy is frequently necessary. These patients are at risk of liver failure and serious drug toxicity. Liver transplantation must be considered at the first sign of hepatic decompensation. The development of ascites typically heralds the need for a transplant evaluation (see Table 88-5). Alternative management strategies for treatment failure have included the administration of cyclosporine, tacrolimus, mycophenolate mofetil, ursodeoxycholic acid, budesonide, 6-mercaptopurine, methotrexate, and cyclophosphamide.75,89 In each instance, experience has been limited, and in most reports the preliminary results have been encouraging but uncorroborated. Among the new drugs used in treatment failure, only ursodeoxycholic acid has been evaluated by a randomized controlled clinical trial, and the findings did not support its use.75 Site-specific interventions will be possible as soon as the pathogenic mechanisms are clarified.75 These therapies may include peptides to block autoantigen display within the class II MHC molecules, agents such as CTLA4 to temper immunocyte response, T-cell vaccination, oral tolerance regimens, cytokine manipulations, mesenchymal stem cell transplantation, and gene therapies that can offset the over-expression of certain cytokines, limit fibrosis, and promote regeneration.

INCOMPLETE RESPONSE

In 13% of patients, clinical improvement is observed during therapy without achieving remission criteria (see Fig. 88-6).1,75 The diminishing benefit-to-risk ratio of protracted therapy justifies an alternative strategy. The administration of azathioprine (2 mg/kg daily) as the sole drug86 or a lowdose prednisone regimen87 is a reasonable approach. The goal of treatment is to reduce and stabilize disease activity on a drug schedule that is well tolerated.

DRUG TOXICITY

For patients with serious side effects from therapy, treatment usually can be continued with the single tolerated drug (prednisone or azathioprine) in an adjusted dose (see Fig. 88-6).1,75,77 Cyclosporine, 6-mercaptopurine, and cyclophosphamide also have been used successfully after drug toxicity in isolated cases.1,75

LIVER TRANSPLANTATION Liver transplantation is effective in the treatment of decompensated AIH (see Chapter 95).1,75 Five-year survival rates for patients and grafts range from 83% to 92%, and the actuarial 10-year survival rate after transplantation is 75%. AIH recurs in at least 17% of patients, and AIH develops de novo in 3% to 5% of patients who undergo transplantation for nonautoimmune liver disease. Acute rejection, glucocorticoid-resistant rejection, and chronic rejection occur more commonly in patients undergoing transplantation for AIH than for other conditions, and

1475

1476

Section IX  Liver patients with AIH are more difficult to withdraw from glucocorticoids.90 Recurrent AIH typically is mild and develops in patients who are inadequately immunosuppressed.89 Dose adjustments usually are sufficient to suppress the disease, but progression to cirrhosis and graft failure have been reported. De novo AIH is a clinical syndrome that affects both children and adults who undergo transplantation for nonautoimmune liver disease.90 Patients with de novo AIH in whom glucocorticoid therapy fails, experience worsening hepatic fibrosis with possible graft loss, and those who do not receive glucocorticoids progress to cirrhosis, require retransplantation, or die of liver failure. De novo AIH in some adults has been associated with severe centrilobular necrosis, and adult patients have been reported to express an atypical anti-liver-kidney cytosolic antibody of uncertain pathogenic significance.90 AIH should be included in the differential diagnosis of allograft dysfunction in all transplant recipients.

KEY REFERENCES

Abdalian R, Dhar P, Jhaveri K, et al. Prevalence of sclerosing cholangitis in adults with autoimmune hepatitis: Evaluating the role of routine magnetic resonance imaging. Hepatology 2008; 47:949-57. (Ref 50.) Alvarez F, Berg P, Bianchi F, et al. International Autoimmune Hepatitis Group report: Review of criteria for diagnosis of autoimmune hepatitis. J Hepatol 1999; 31:929-38. (Ref 3.) Czaja AJ. Autoantibodies in autoimmune liver disease. Adv Clin Chem 2005; 40:127-64. (Ref 6.)

Czaja AJ. Autoimmune hepatitis—Part A: Pathogenesis. Expert Rev Gastroenterol Hepatol 2007; 1:113-28. (Ref 18.) Czaja AJ. Autoimmune hepatitis—Part B: Diagnosis. Expert Rev Gastroenterol Hepatol 2007; 1:129-43. (Ref 7.) Czaja AJ. Performance parameters of the diagnostic scoring systems for autoimmune hepatitis. Hepatology 2008; 48:1540-8. (Ref 15.) Czaja AJ. Genetic factors associated with the occurrence, clinical phenotype and outcome of autoimmune hepatitis. Clin Gastroenterol Hepatol 2008; 6:379-88. (Ref 64.) Czaja AJ. Safety issues in the management of autoimmune hepatitis. Expert Opin Drug Saf 2008; 7:319-33. (Ref 77.) Czaja AJ, Carpenter HA. Empiric therapy of autoimmune hepatitis with mycophenolate mofetil: Comparison with conventional treatment for refractory disease. J Clin Gastroenterol 2005; 39:819-25. (Ref 88.) Czaja AJ, Freese DK. Diagnosis and treatment of autoimmune hepatitis. Hepatology 2002; 36:479-97. (Ref 1.) Hennes EM, Zeniya M, Czaja AJ, et al. Simplified diagnostic criteria for autoimmune hepatitis. Hepatology 2008; 48:169-76. (Ref 14.) Montano-Loza AJ, Carpenter HA, Czaja AJ. Features associated with treatment failure in type 1 autoimmune hepatitis and predictive value of the model for end-stage liver disease. Hepatology 2007; 46:1138-45. (Ref 69.) Montano-Loza A, Carpenter HA, Czaja AJ. Improving the end point of corticosteroid therapy in type 1 autoimmune hepatitis to reduce the frequency of relapse. Am J Gastroenterol 2007; 102:1005-12. (Ref 82.) Montano-Loza A, Carpenter HA, Czaja AJ. Consequences of treatment withdrawal in type 1 autoimmune hepatitis. Liver Int 2007; 27:50715. (Ref 85.) Montano-Loza A, Czaja AJ. Current therapy for autoimmune hepatitis. Nature Clin Pract Gastroenterol Hepatol 2007; 4:202-14. (Ref 75.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

89 Primary Biliary Cirrhosis Paul Angulo and Keith D. Lindor

CHAPTER OUTLINE Epidemiology  1477 Pathogenesis  1478 Autoantibodies  1478 Genetic Factors  1478 Molecular Mimicry  1479 Xenobiotics and Other Implicated Agents  1479 Clinical Features  1479 Asymptomatic Disease  1479 Symptomatic Disease  1479 Associated Diseases  1480 Diagnosis  1480 Biochemical Features  1480 Serologic Diagnosis  1480 Histopathologic Features  1480 Imaging Studies  1481

Primary biliary cirrhosis (PBC) is an autoimmune liver disease that generally affects middle-aged women from a variety of racial groups and represents the most common chronic cholestatic liver disease in adults in the United States. PBC is characterized by ongoing inflammatory destruction of the intralobular bile ducts, which leads to chronic cholestasis and biliary cirrhosis, with consequent complications such as portal hypertension and liver failure. Although the designation primary biliary cirrhosis has been used for several decades, the term is potentially misleading because most patients do not have “cirrhosis” on liver biopsy specimens when the disease is diagnosed. Evidence for an immunologic cause of PBC includes the presence of activated T cells in areas of bile duct destruction, presence of highly specific autoantibodies that react with antigens localized on biliary epithelial cells, and association of PBC with other disorders thought to be autoimmune in nature. Fatigue and pruritus are the most common presenting symptoms of PBC; however, as many as one half of patients are asymptomatic when the disease is diagnosed on the basis of incidentally discovered elevated serum liver enzyme levels in a cholestatic pattern. Rarely, patients present with advanced disease manifested by esophageal variceal hemorrhage, ascites, or hepatic encephalopathy. PBC should be considered in a patient with an elevated serum alkaline phosphatase level, hypercholesterolemia, and an elevated serum immunoglobulin M (IgM) level. The presence of antimitochondrial antibodies in serum is highly characteristic of the disease. Ursodeoxycholic acid (UDCA) is the only medication of proven benefit for patients with PBC, and liver transplantation offers a life-extending alternative for patients with end-stage PBC. Complications of chronic cholestasis such as osteopenic bone disease, fat-soluble vitamin deficiency, hypercholesterolemia, and steatorrhea should be

Natural History  1482 Asymptomatic Primary Biliary Cirrhosis  1482 Symptomatic Primary Biliary Cirrhosis  1482 Predicting Survival  1482 Treatment  1482 Ursodeoxycholic Acid  1483 Other Drugs  1484 Complications of Chronic Cholestasis  1485 Bone Disease  1485 Fat-Soluble Vitamin Deficiency  1486 Hyperlipidemia  1486 Pruritus  1486 Steatorrhea  1487 Liver Transplantation  1487 AMA-Negative Primary Biliary Cirrhosis  1487

recognized and treated. Survival models are used to help predict the prognosis of patients with the disease and the timing of liver transplantation.

EPIDEMIOLOGY PBC occurs worldwide and predominantly in women, with a female-to-male ratio of 9 : 1. The diagnosis of PBC usually is made between the ages of 30 and 60 years, with a range of 21 to 93 years. The disease has been documented in even younger patients—two teenagers 15 and 16 years of age, respectively.1 Until the early 1970s, PBC was considered a rare condition that manifested with persistent jaundice and almost inevitably progressed to end-stage liver disease. A better understanding of its pathogenesis, along with subsequent clinical and epidemiologic studies, has modified current concepts regarding this condition. PBC seems to be more common than was formerly believed because of increasing awareness of the disease and because asymptomatic patients are identified through the widespread use of screening tests such as determination of serum cholesterol levels and liver biochemical test levels in otherwise healthy persons. The reported prevalence of PBC varies among countries, with a range of 19 cases per 1 million population in Israel to 402 cases per 1 million population in Olmsted County, Minnesota.2 Whether the difference in prevalence is real or a result of different methodologies used to detect the disease is unknown. Inconsistency in case definition and case finding methods, as well as imprecision in defining the study area, the populations evaluated, and the dates of diagnosis, particularly in earlier reports, makes comparisons among studies difficult. Estimates of the annual incidence

1477

1478

Section IX  Liver of PBC range from 0.7 to 49 per 1 million population. Both the prevalence and incidence of PBC seem to have increased over time.2 The increase in prevalence possibly reflects an increase in survival time in patients with PBC. In the United States,3 the age-adjusted reported incidence of PBC per 1 million person-years is 45 for women and 7 for men (27 overall). The reported prevalence per 1 million population is 654 for women and 121 for men (402 overall); these figures represent the highest prevalence rates for PBC ever reported.

PATHOGENESIS Although the cause of PBC remains unknown, several lines of evidence suggest an autoimmune pathogenesis. The evidence includes the intense humoral and cellular response to an intracytoplasmic antigen, presence of highly specific antimitochondrial antibodies (AMA), involvement of T lymphocytes in the destruction of bile ducts, and numerous defects in immunologic regulation. Like other autoimmune diseases, PBC has a clear female predominance. Both PBC and the presence of AMA occur more frequently in close relatives of patients who have PBC than in controls. PBC is associated with an increased incidence of autoimmune disease of other types in both patients with PBC and their first-degree relatives.4,5 PBC seems to be triggered by an immune-mediated response to one or more allo- or autoantigens, which leads to progressive destruction of bile ducts, chronic cholestasis, and eventual biliary cirrhosis. Immunohistochemical phenotyping of inflammatory cells surrounding the bile ducts shows a combination of CD4+ and CD8+ T lymphocytes, accompanied by B lymphocytes and natural killer cells. Bile duct destruction is induced directly by the cytotoxicity of CD4+ and CD8+ T cells in contact with biliary epithelium. B lymphocytes are relatively uncommon in the inflammatory reaction but sometimes can be seen in clusters. Intracellular adhesion molecules (e.g., intracellular adhesion molecule-1 [ICAM-1]) are strongly expressed on many epithelial cells, particularly in areas of lymphocyte damage; these molecules may facilitate the interaction between destructive lymphocytes and their targets. In the early biliary lesions of PBC, eosinophilic infiltration and granulomas often are seen. PBC is principally a disease of the small intrahepatic bile ducts, with loss of biliary epithelial cells that line these ducts and resulting cholestatic damage. PBC is not restricted to the liver; abnormalities of salivary and lacrimal glands with an associated cellular phenotypic change similar to that seen in the biliary epithelial cells also occur. Three spontaneous autoimmune biliary disease mouse models6-8 and two induced models of PBC9,10 have been reported.

AUTOANTIBODIES

AMA were described for the first time in patients with PBC in the 1960s and continue to be regarded as the most sensitive and specific immunologic hallmark of the disease. AMA are directed to the E2 component of the pyruvate dehydrogenase complex (PDC-E2), the E2 unit of the branched-chain 2-oxo-acid dehydrogenase complex (BCOADC-E2), and the E2 subunit of the 2-oxo-glutarate dehydrogenase complex (OGDC-E2). Other AMA recognize the E1a subunit of PDC (PDC E1a) and the dihydrolipoamide dehydrogenase-binding protein (E3BP) of PDC. These molecules all are located on the inner mitochondrial membranes. At least one of these components usually reacts with AMA in a patient with PBC. The most frequent

antigen against which AMA are directed is PDC-E2; PDCE2–reacting antibodies are present in 90% to 95% of PBC sera. The mechanisms by which AMA directed to proteins located on the inner surface of mitochondrial membranes develop are unknown; PDC-E2 or a cross-reactive molecule is overexpressed on biliary epithelial cells in PBC, predominantly at the luminal domain, and PDC-E2–specific CD4+ T cells are present in portal inflammatory infiltrates of affected persons. Although AMA are predominantly of the IgG1 and IgG3 classes, most patients who have PBC exhibit polyclonal elevation of serum IgM levels; the IgM is not directed at mitochondrial or nuclear antigens. This phenomenon is suggestive of polyclonal activation of the B-cell compartment with an associated failure of isotype switching, representing aberrant B-cell activation.11 AMA do not appear to be cytotoxic: (1) They persist after liver transplantation without evidence of disease recurrence; (2) disease severity is unrelated to antibody titer; (3) they are not always present in PBC; and (4) they develop in animal models after the injection of recombinant PDC-E2 protein, but without resulting bile duct destruction or inflammation. Further, the different types and numbers of mitochondrial antigens recognized by Western immunoblot analysis at the time of the patient’s presentation is independent of the stage of the liver disease and not associated with specific clinical, biochemical, histologic, and immunologic features or with the Mayo risk score (see later).12 Antinuclear antibodies (ANA) are present in nearly one half of patients with PBC and in up to 85% of patients with AMA-negative PBC (see later). The most relevant immunofluorescent reactivities of ANA in patients with PBC are anti-multiple nuclear dots antibodies ([anti-MND], with the molecular target being a 100-kd soluble protein called Sp100), anticentromere antibodies, and antinuclear envelope antibodies. The immunofluorescence pattern of the antinuclear envelope antibodies is characterized as being rim-like and membranous; its molecular targets are structural components of the nuclear pore complex, such as gp210 and nucleoprotein p62, and of the nuclear membrane, such as lamin B receptors. Antibodies against the nuclear pore protein gp210 (anti-gp210) are found in 25% of patients with AMA-positive PBC and in up to 50% of those with AMA-negative PBC. ANA with the MND and rim-like and membranous patterns, which are relatively rare or absent in normal and pathologic controls, are strongly associated with PBC and can be considered to be surrogate markers of PBC in AMA-negative patients.13-15 The specificity of anti-gp210 for PBC when detected by immunoblotting is greater than 99%, whereas antibodies to p62 (anti-p62) are found in approximately 25% of patients and are highly specific for PBC. Anti-p62 antibodies seem to be mutually exclusive with anti-gp210 antibodies. Further, anti-gp210 and possibly anti-p62 also offer prognostic information in that they seem to be associated with aggressive disease with a poor prognosis.13,16,17 In Japanese patients with PBC, anticentromere antibodies are associated with the development of portal hypertension.18

GENETIC FACTORS

The occurrence of PBC in relatives of affected persons plus abnormalities of cell-mediated immunity in first-degree relatives of patients with PBC suggests a genetic association. This association is further supported by the finding that PBC exhibits a higher concordance rate in monozygotic than dizygotic twin pairs, suggesting a genetic component to disease susceptibility and expression of the susceptibility through genes that regulate the immune response.19 Many

Chapter 89  Primary Biliary Cirrhosis of the familial risk and genetic studies of PBC have failed to distinguish, however, between true genetic risk and shared environmental exposure. Although no link between PBC and a specific human leukocyte antigen (HLA) class I phenotype has been found, HLA class II molecules may contribute to the development of this condition. The HLA associations with PBC detected most commonly have been with the DRB1*0801 allele in European and North American white populations and DRB1*0803 in Japanese populations.20,21 Although the DRB1*08 allele seems to impart a significant risk for PBC, with odds ratios of 3 or higher in white and Japanese populations, a study from China reported that the frequencies of the DRB1*0701 and DRB1*03 alleles were increased significantly in patients with PBC compared with controls, with no difference in the frequencies of the DRB1*08 allele.22 A few class II HLA alleles demonstrate protective associations against PBC. These alleles include DQA1*0102 in United States and Japanese studies and DQB1*0602 in a United States study. More recently, DRB*13 was found to be protective against PBC in patients from the United Kingdom and Italy, and DRB1*11 was protective in patients from Italy but not those from the United Kingdom.20,23 An association with HLA class III genes, which code for complement components C2 and C4, cytokines, and complement factor B, has been studied less extensively. Earlier studies reported an increased frequency of haplotype C4B2 and haplotype C4A*Q0 in patients with PBC. Smaller, older studies of the association between PBC and alleles affecting the expression of tumor necrosis factor-α have been reported, but an association remains unresolved.

MOLECULAR MIMICRY

Molecular mimicry between host autoantigens and unrelated exogenous proteins is one of the hypotheses to explain how autoantibodies to self-proteins arise, break tolerance, and lead to autoimmune disease. Molecular mimicry of an extrinsic protein produced by an infectious agent has long been suggested as a possible initiating event in PBC. Infectious agents incriminated in the immune response in PBC include various bacteria and viruses, most recently Chlamydia pneumoniae,24 Novospingobium aromaticivorans,25 and human betaretrovirus.26 Microorganisms produce a multitude of foreign antigens that collectively constitute the major determinants recognized by the immune system. These antigens potentially include a variety of carbohydrates, lipids, and proteins that can be recognized by specific receptors on inflammatory cells. In PBC, PDC-E2 appears to be an ideal candidate for foreign antigens to mimic. PDC-E2, particularly its inner lipoyl domain, is highly conserved among bacteria, yeasts, and mammals. Autoimmune phenomena in PBC could result from peptides that mimic T-cell epitopes of microbial proteins and that are derived from, and presented by, abnormally expressed HLA class II molecules. Molecular mimicry has been invoked to explain the breaking of tolerance against mitochondrial antigens. Definitive evidence for this theory is still lacking, however.

XENOBIOTICS AND OTHER IMPLICATED AGENTS

Xenobiotics are foreign compounds that may alter selfproteins by inducing a change in the molecular structure of the native protein sufficient to induce an immune response. The immune response may then result in the recognition of both the modified and the native proteins.27 The continued presence of the self-protein may perpetuate the immune response initiated by the xenobiotic-induced adduct, thereby leading to chronic autoimmunity. Because many

xenobiotics are metabolized in the liver, the potential for liver-specific alteration of proteins is substantial. To address the hypothesis that PBC is induced by xenobiotic exposure, Long and colleagues28 synthesized the inner lipoylated domain of PDC-E2, replaced the lipoic acid moiety with synthetic structures, and quantified the reactivity of these structures with sera from patients with PBC. AMA from all patients reacted more strongly to 3 of the 18 modified organic autoepitopes than to the native domain. Defective sulfoxidation of certain compounds, such as bile acids, estrogen, or drugs, and selenium deficiency have been proposed as underlying mechanisms that may lead to this process. These hypotheses remain unproved.

CLINICAL FEATURES ASYMPTOMATIC DISEASE

Widespread use of screening laboratory tests has led to diagnosis of PBC at an asymptomatic stage in up to 60% of patients with this condition. Such patients are found incidentally to have an elevated serum alkaline phosphatase level and AMA during routine health evaluations or during investigation of an unrelated complaint, such as a clinical manifestation of an autoimmune disease known to be associated with PBC. Some asymptomatic persons with AMA and normal results of liver biochemical tests are found to have features on liver biopsy specimens that are diagnostic of or consistent with PBC; symptoms, signs, and laboratory evidence of chronic cholestasis eventually develop in these persons.

SYMPTOMATIC DISEASE

The typical patient with symptomatic disease (Table 89-1) is a middle-aged woman with a complaint of fatigue or pruritus. Other symptoms include right upper quadrant abdominal pain, anorexia, and jaundice. Fatigue, although relatively nonspecific, is considered to be the most disabling symptom by many patients, and it worsens in some patients as the disease progresses.29,30 Fatigue in patients with PBC does not correlate with several markers of disease severity, or with the patient’s age or thyroid status, but does correlate with sleep disturbance and depression.29 In a four-year follow-up study, the severity of fatigue was quite stable overall; only liver transplant recipients experienced a significant improvement in fatigue.31 In that study, fatigue was also found on multivariate analysis to be an independent predictor of mortality, particularly cardiac death.31 Pruritus may occur at any point, early or late, in the course of the disease, or intermittently throughout the

Table 89-1  Symptoms and Signs of Primary Biliary Cirrhosis at Presentation symptom or sign Fatigue Pruritus Hyperpigmentation Hepatomegaly Splenomegaly Xanthelasma Jaundice Right upper quadrant pain None

FREQUENCY (%) 21-85 19-55 25 25 15 10 3-10 8 25-61

1479

1480

Section IX  Liver Table 89-2  Diseases Associated with Primary Biliary Cirrhosis DISEASE Keratoconjunctivitis sicca (Sjögren’s syndrome) Renal tubular acidosis Arthritis/arthropathy Gallstones Autoimmune thyroiditis Scleroderma and its variants   Raynaud’s disease   CREST or any of its components   Scleroderma Cutaneous disorders—lichen planus, discoid lupus, pemphigoid Hepatocellular carcinoma Pulmonary fibrosis Celiac disease

FREQUENCY (%) 72-100 50-60 4-42 33 15-20 15-19 8 7 3-4 11 1-2 Rare Rare

CREST, calcinosis, Raynaud’s phenomenon, esophageal dysmotility, sclerodactyly, telangiectasia.

course. Pruritus generally is intermittent during the day and is most troublesome in the evening and at night. Pruritus often resolves as the disease progresses, but in some patients, severe, intractable pruritus can develop in earlier stages of the disease and may require liver transplantation for effective management. In a population-based study of 770 patients with PBC from England, the cumulative risk of developing pruritus was 19%, 45%, and 57% at 1, 5, and 10 years, respectively.30 Most patients with PBC do not have jaundice at the time of diagnosis. Jaundice occurs later in the course of the disease and usually is persistent and associated with a worse prognosis. Symptoms also may relate to fat-soluble vitamin deficiency, bone pain with or without spontaneous fractures, or an associated autoimmune disease that may occur in patients with PBC (Table 89-2). Symptoms and signs of advanced liver disease, such as ascites, bleeding from gastroesophageal varices, and encephalopathy, usually occur late in the course of PBC. On physical examination, the most common signs are skin hyperpigmentation (caused by deposition of melanin), hepatosplenomegaly, xanthelasma, and, in more advanced disease, jaundice. Symptoms appear to be less frequent in men than in women, and autoimmune manifestations, especially Sjögren’s syndrome, also are less frequent in men. Otherwise, PBC is identical clinically in men and in women.

ASSOCIATED DISEASES

Many of the diseases found frequently in patients with PBC (see Table 89-2) are thought to be related to disturbances in immune mechanisms. These associated disorders include Sjögren’s syndrome (characterized by dry eyes [keratoconjunctivitis sicca] and dry mouth), scleroderma and its variants, rheumatoid arthritis, some cutaneous disorders, renal tubular acidosis, and thyroiditis. The frequency of malignancy is increased in patients with PBC. An increased risk of breast cancer in women with PBC was found in earlier studies but was not confirmed by subsequent, larger studies. Although hepatocellular carcinoma occurs in only 1% to 2% of patients with PBC, the risk is much higher than that in the general population. Gallstones can be found in up to one third of patients with PBC, but inflammatory bowel disease and interstitial pulmonary fibrosis are rare.

DIAGNOSIS The diagnosis of PBC is established by liver biochemical test results consistent with chronic cholestasis plus the presence in serum of AMA. Liver biopsy helps to confirm the diagnosis of PBC but may not be necessary for establishing the diagnosis in patients with characteristic chronic cholestasis and AMA.31

BIOCHEMICAL FEATURES

Liver biochemical test results show a cholestatic picture. Almost all patients have increased serum levels of alkaline phosphatase (three to four times the upper limit of normal) and gamma glutamyl transpeptidase. Serum aminotrans­ ferase (aspartate aminotransferase [AST], alanine aminotransferase [ALT]) levels are mildly elevated (usually less than three times normal); marked elevations (more than five times normal) are distinctly unusual and may suggest PBCautoimmune hepatitis overlap syndrome (see Chapter 88) or coexisting viral hepatitis. Serum bilirubin levels usually are normal in early stages and increase slowly over the course of the disease; levels ultimately may exceed 20 mg/dL. A high serum bilirubin level, low serum albumin, and prolonged prothrombin time indicate a poor prognosis and advanced disease. Serum immunoglobulin levels, especially IgM, are increased, as are serum levels of bile acids, in particular cholic and chenodeoxycholic acids, and cholesterol.

SEROLOGIC DIAGNOSIS

Indirect immunofluorescence, immunoblotting, and enzyme-linked immunosorbent assay (ELISA) can detect AMA. Indirect immunofluorescence is by far the most commonly used serologic test and detects AMA in 90% to 95% of patients with PBC; however, indirect immunofluorescence testing requires interpretation by a skilled observer, and the result may be interpreted erroneously as negative for AMA in some patients with PBC. Immunoblotting and ELISA have sensitivity and specificity rates higher than 95% for the detection of AMA in PBC and can detect AMA in patients with PBC who are AMA negative by direct immunofluorescence testing. Other autoantibodies found in patients with PBC include rheumatoid factor (70%), smooth muscle antibodies (66%), antinuclear antibodies (50%), and antithyroid (antimicrosomal, antithyroglobulin) antibodies (41%).

HISTOPATHOLOGIC FEATURES

The initial lesion on a liver biopsy specimen in PBC (Figs. 89-1 and 89-2A and B) is damage to epithelial cells of the small bile ducts. The most important and only diagnostic clue in many cases is ductopenia, defined as the absence of interlobular bile ducts in greater than 50% of portal tracts. The florid duct lesion, in which the epithelium of the interlobular and segmental bile ducts degenerates segmentally, with formation of poorly defined, noncaseating epithelioid granulomas, is nearly diagnostic of PBC but is found in a relatively small number of cases, mainly in early stages. The two most popular histologic staging systems are those proposed by Ludwig and colleagues and Scheuer, which classify the disease in four stages. Both systems describe progressive pathologic changes, beginning initially in the portal areas surrounding the bile ducts and culminating in cirrhosis. Ludwig stage 1 disease is characterized by inflammatory destruction of the intrahepatic septal and interlobular bile ducts that range up to 100 µm in diameter. These

Chapter 89  Primary Biliary Cirrhosis lesions often are focal and described as florid duct lesions, characterized by marked inflammation and necrosis around a bile duct. The portal tracts usually are expanded by lymphocytes, with only sparse neutrophils or eosinophils seen. In stage 2 disease (see Fig. 89-2A), the inflammation extends from the portal tract into the hepatic parenchyma, a lesion called interface hepatitis, or formerly, piecemeal necrosis.

Stage 1

Stage 2

Stage 3

Destruction of bile ducts with proliferation of bile ductules can be seen. Stage 3 disease is characterized by scarring and fibrosis. Lymphocytic involvement of the portal and periportal areas, as well as the hepatic parenchyma, can be seen, but the hallmark of this stage is the presence of fibrosis without regenerative (or regenerating) nodules. Stage 4 disease is characterized by cirrhosis with fibrous septa and regenerative nodules (see Fig. 89-2B). Most patients with PBC demonstrate progression of the liver disease; a few patients have a prolonged course of histologic stability, and only rare patients have sustained regression. A time course Markov model has been used to describe the rate of histologic progression over time (Table 89-3). Liver biopsy has been considered necessary for confirming the diagnosis of PBC and excluding other liver diseases. This routine indication for diagnostic purposes has been questioned, however.32 In a patient with AMA in serum, the combination of a serum alkaline phosphatase level greater than 1.5 times the upper limit of normal plus a serum AST level less than 5 times the upper limit of normal yields a 98.2% positive predictive value for a diagnosis of PBC. Therefore, a liver biopsy is not necessary to confirm the diagnosis in most patients with PBC and should be performed in only the minority of AMA-positive patients with a serum alkaline phosphatase level less than 1.5 times normal or a serum AST level greater than 5 times normal.32

IMAGING STUDIES

Cross-sectional imaging with ultrasonography, computed tomography, or magnetic resonance imaging is useful for excluding biliary obstruction. Other than increased liver echogenicity and signs of parenchymal liver disease or portal hypertension, findings on cross-sectional imaging

Stage 4

Table 89-3  Time Course of Histologic Progression to a Higher Stage in Patients with Primary Biliary Cirrhosis

Figure 89-1.  Schematic representation of the staging system of primary biliary cirrhosis (Ludwig’s classification). The left side of the schematic shows five portal tracts surrounding a central vein at each stage. The right side shows a larger single portal tract at each stage (the bile ductule is blue). In stage 1 the inflammation is confined to the portal space, focused on the bile duct. In stage 2 the inflammation extends into the hepatic parenchyma (interface hepatitis or piecemeal necrosis). In stage 3, fibrosis is present; in stage 4, cirrhosis is present.

A

Initial Histologic Stage (%)

RATE OF PROGRESSION

1

2

3

1 year 2 years

41 62

43 62

35 50

B

Figure 89-2.  A, Photomicrograph of stage 2 primary biliary cirrhosis (PBC). Mononuclear inflammatory cells expand the portal tracts with some disruption of the limiting plates (interface hepatitis). The bile ducts are surrounded by inflammatory cells, and no fibrosis is evident. (Hematoxylin and eosin, ×100.) B, Stage 4 PBC. Cirrhosis, with areas of fibrosis surrounding the hepatic parenchyma, is present. A dense mononuclear inflammatory infiltrate is still seen in the portal tract, with interface hepatitis. (Hematoxylin and eosin, ×100.)

1481

1482

Section IX  Liver usually are unremarkable. Adenopathy in the portal area is found in approximately 24% of patients with PBC, is not progressive, and is important to recognize so as to avoid confusion about the diagnosis or undue concern about the presence of an underlying malignancy. Large, bulky adenopathy, however, should raise the question of asso­ ciated malignancy.

NATURAL HISTORY The natural history of PBC has been described in patients with symptoms attributable to PBC, as well as in asympto­ matic patients who have normal or abnormal liver bio­ chemical test levels. Prognostic models useful in predicting survival in an individual patient have been developed.

ASYMPTOMATIC PRIMARY BILIARY CIRRHOSIS

An early series reported on 29 patients with AMA (in a titer of 1 : 40 or greater), normal liver biochemical test levels, and no symptoms of liver disease. Liver histology was compatible with or diagnostic of PBC in 24 patients (83%) and normal in only 2 patients. The entire cohort of patients was followed for a median of 17.8 years (range, 11 to 24 years). Liver biochemical test values became persistently abnormal in 24 patients (83%), and persistent symptoms attributable to PBC, including fatigue, pruritus, and right upper abdo­ minal discomfort, developed in 22 (76%). Five patients died, none because of liver disease, after a median period of 11.7 years (range, 6.4 to 16.8 years) from the first positive AMA titer. The median time from the first positive AMA result to persistent liver biochemical abnormalities was 5.6 years (range, 0.9 to 19 years). Four of 10 patients who underwent a second liver biopsy during a median follow-up of 11.4 years (range, 1.3 to 14.3 years) showed progression of disease stage, but cirrhosis or portal hypertension did not develop in any of the patients during the follow-up period. This study shows clearly that asymptomatic patients who have AMA and normal liver biochemical test levels have early PBC; with time, clinically obvious PBC may develop. These patients may represent a subgroup of patients with PBC whose natural history is different from that in the general PBC patient population. Several reports have described the natural history of asymptomatic patients who have AMA, abnormal liver biochemical test levels consistent with cholestasis, and liver histologic features diagnostic of or compatible with PBC. Asymptomatic patients have less advanced disease than that typically seen in symptomatic patients.33 Patients who are asymptomatic at presentation may survive longer, but a majority will eventually have progressive disease. A median survival of approximately 10 years has been reported for this group of patients. Patients who remain asymptomatic for several years may have a significantly longer survival than that in symptomatic patients, but their life expectancy is still less than that of an age- and gender-matched population. Symptoms of PBC will develop in approximately 40% of the initially asymptomatic patients within five to seven years of follow-up, and most asymptomatic patients ultimately will become symptomatic if the follow-up period is long enough (95% after 20 years).33 When symptoms develop, life expectancy falls significantly and is the same as that for other symptomatic patients. The mortality rate for liver-related causes is significantly higher in initially symptomatic patients than in initially asympto­ matic patients; however, an excess rate of non–liver-related mortality in initially asymptomatic patients has been reported to decrease the median survival in these patients to that in initially symptomatic patients.33

Table 89-4  Independent Predictors of Survival in Patients with Primary Biliary Cirrhosis in Various Clinical Studies Clinical Age Ascites Edema Hepatomegaly Variceal bleeding Laboratory Serum albumin level Serum alkaline phosphatase level Serum bilirubin level Prothrombin time Liver Histology Cholestasis Cirrhosis Fibrosis Mallory’s hyaline

SYMPTOMATIC PRIMARY BILIARY CIRRHOSIS

When compared with asymptomatic patients, patients with PBC who have symptoms of chronic cholestasis show a more rapid progression to end-stage liver disease and have a worse prognosis. Several independent predictors of a poor prognosis have been identified in this group of patients (Table 89-4). The manifestations of portal hypertension and its complications in patients with PBC are similar to those in other forms of cirrhosis. Most patients with PBC and portal hypertension have cirrhosis; however, portal hypertension can be found in some patients with PBC and moderate-to-severe hepatic inflammation without cirrhosis on a liver biopsy specimen. Development of gastroesophageal varices is an ominous sign that is observed in approximately one third of patients with PBC during extended follow-up. Approximately 40% of these patients will experience one or more episodes of variceal bleeding within three years of developing varices and, as a group, have a decreased survival rate.

PREDICTING SURVIVAL

When untreated, PBC may follow a course that extends over a 15- to 20-year period. In patients with a serum bilirubin level greater than 10 mg/dL, however, the average life expectancy is reduced to two years. In order to predict survival in patients with PBC, prognostic models, some of which rely on Cox’s proportional hazard analysis, have been developed (Table 89-5). Among these models, the Mayo risk score has been cross-validated and is widely used in predicting survival and in guiding referral of patients for liver transplantation. These prognostic models also can be used for monitoring the efficacy of experimental drugs in clinical trials. Although all the prognostic models are of help in clinical decision-making, they should not replace clinical judgment in determination of the optimal timing of liver transplantation in an individual patient (see Chapter 95).

TREATMENT A large number of published controlled and uncontrolled trials have evaluated various drugs in PBC. These drugs can be classified according to their mechanisms of action as bile acids, immunosuppressive, anti-inflammatory, cupruretic, or antifibrotic.

Chapter 89  Primary Biliary Cirrhosis Table 89-5  Prognostic Models in Primary Biliary Cirrhosis REFERENCE

PREDICTIVE VARIABLES

FORMULA (IF USED)

VALIDATED BY OTHER GROUPS

A

Age Total bilirubin Serum albumin Prothrombin time Edema score Bleeding varices Bilirubin Bilirubin Ascites Albumin Age Gastrointestinal bleeding Central cholestasis Cirrhosis Immunoglobulin M Age Bilirubin Albumin Alkaline phosphatase

R = 0.871 loge (bilirubin in mg/dL) −2.53 loge (albumin in g/dL) +0.039 (age in years) +2.38 loge (prothrombin time in seconds) +0.859 edema Loge R = 1.68 (bleeding − 0.25) +2.03 log (bilirubin − 30.3) Calculated from pocket chart and tables in published article

Yes

R = exp [(0.0742 × age in years) + (0.195 × loge bilirubin ratio*) − (2.7878 × albumin ratio*) + (0.2610 × loge alkaline phosphatase ratio*)]

No

B C

D

No Yes

*Ratio equals times lower limit of normal for albumin and times upper limit of normal for bilirubin and alkaline phosphatase. R, risk score. A, Dickson ER, Grambsch PM, Flemming TR, et al. Prognosis in primary biliary cirrhosis: Model for decision making. Hepatology 1989; 10:1-7; B, Rydning A, Schrumpf E, Abdelnoor M, et al. Factor of prognostic importance in primary biliary cirrhosis. Scand J Gastroenterol 1990; 25:119-26; C, Christensen E, Neuberger J, Crowe J, et al. Beneficial effect of azathioprine and prediction of progression in primary biliary cirrhosis. Final results of an international trial. Gastroenterology 1985; 89:1084-95; D, Prince M, Chetwynd A, Newman W, et al. Survival and symptom progression in a geographically based cohort of patients with primary biliary cirrhosis: Follow-up for up to 28 years. Gastroenterology 2002; 123:1044-51.

URSODEOXYCHOLIC ACID

UDCA, the 7-b epimer of chenodeoxycholic acid, occurs naturally in small quantities in human bile (less than 4% of total bile acids). It was first introduced for the dissolution of radiolucent gallstones in the 1970s and is the only medication approved by the U.S. Food and Drug Administration for treatment of PBC. Several mechanisms for the protective actions of UDCA have been proposed, including inhibiting absorption of toxic, hydrophobic, endogenous bile salts; stabilizing hepatocyte membranes against toxic bile salts; replacing endogenous bile acids, some of which may be hepatotoxic, with the nonhepatotoxic UDCA; and reducing expression of major histocompatibility complex (MHC) class I and class II antigens. During UDCA therapy, a variable increase in the concentration of total bile acids in serum is observed. The proportion of UDCA in serum and bile increases to approximately 30% to 60% of total bile acids, and the proportion of endogenous bile acids, such as cholic, chenodeoxycholic, deoxycholic, and lithocholic acids, declines consequently. The degree of enrichment of the bile acid pool with UDCA is similar in all histologic stages of PBC and correlates with improvement in liver biochemical test levels and the Mayo risk score. Because of its safety and patient adherence to treatment with the drug, UDCA has received the most attention of any drug used to treat PBC. Treatment with UDCA leads to rapid improvement in liver biochemical test levels and a decrease in the histologic severity of interface hepatitis, inflammation, cholestasis, bile duct paucity, and bile duct proliferation.34,35 UDCA significantly decreases the risk of development of gastroesophageal varices and ascites and delays progression to cirrhosis.36,37 The predicted probability that cirrhosis will develop after 5 years of therapy with UDCA for patients with stage 1, 2, or 3 disease at diagnosis is 4%, 12%, and 59%, respectively; at 10 years of therapy with UDCA, the probability of cirrhosis is 17%, 27%, and 76%, respectively.37 These figures confirm the beneficial

effect of UDCA on delaying progression to cirrhosis, as compared with disease progression in the absence of treatment. Moreover, UDCA reduces proliferation of colonic epithelial cells, and its long-term use in patients with PBC significantly reduces the probability that colorectal adenomas will recur following removal.38 The beneficial effect of UDCA on long-term survival in patients with PBC was questioned in the past on the basis of results of two meta-analyses.39 These meta-analyses had serious methodologic flaws derived primarily from mixing different patient populations, including some in which the duration of treatment—with suboptimal doses of UDCA— was too short to demonstrate an effect.39 When an effective dose of UDCA (13 to 15 mg/kg/day) has been used and an appropriate number of patients received treatment for an appropriate period of time, UDCA has been shown clearly to improve survival free of liver transplantation (Fig. 89-3).40 On the contrary, a more recent meta-analysis that included only randomized, controlled trials in which the doses of UDCA administered and the duration of follow-up were adequate concluded that the frequency of liver transplantation was reduced significantly, with a marginally significant reduction in the rate of death or liver transplantation, but that the rate of death alone was not reduced in the group of patients who received UDCA.41 In addition, in the sensitivity analyses that included studies that administered placebo as control, long-term studies (>48 months), and large-sized studies (>100 patients), long-term treatment with UDCA reduced the frequencies of liver transplantation and death or liver transplantation significantly.41 Four reports have provided further data on the long-term effects of UDCA in PBC.42-45 A study of 262 patients who received the standard dose of UDCA daily for a mean of eight years found that survival without liver transplantation was only slightly lower than survival of an age- and gendermatched healthy control population and that treatment with UDCA could normalize the survival rate of patients with

1483

Section IX  Liver OTHER DRUGS Prednisolone and Prednisone

1.0

Prednisolone and prednisone may improve serum alkaline phosphatase and aminotransferase levels and liver histologic features in patients with PBC, at least in the short term. Unfortunately, the development of side effects with glucocorticoids, particularly a reduction in bone mass, precludes their use in the treatment of PBC.

UDCA

Probability of survival (%)

1484

Budesonide

Placebo and then UDCA

0.5 0

24

48

Months Figure 89-3.  Survival in 548 patients with primary biliary cirrhosis. The probability of survival was significantly greater in patients who received treatment with ursodeoxycholic acid (UDCA) for four years than in those who first received placebo and then received UDCA (P < .001; relative risk, 1.92; 95% CI: 1.30 to 2.82). (Adapted from Poupon R, Lindor KD, CauchDudek K, et al. Combined analysis of French, American and Canadian randomized controlled trials of urso­deoxycholic acid therapy in primary biliary cirrhosis. Gastroenterology 1997; 113:884-90.)

PBC when given at early stages.42 Another study of 192 patients with PBC treated with UDCA for a mean of 6.7 years found that patients with PBC who have a biochemical response (decrease of serum alkaline phosphatase levels to less that 40% of the baseline or to normal) after 1 year of treatment had a survival rate better than that predicted by the Mayo risk score and similar to that of a historical control population.43 A study of 297 patients treated with UDCA for a median of 5.7 years showed that survival without liver transplantation in low-risk patients was significantly better than survival predicted by the Mayo risk score and slightly lower than that for an age- and gender-matched control group of the general population.44 Another report evaluated the effect of UDCA on hepatocellular carcinoma and found that the increased risk in UDCA-treated patients with PBC was three-fold, in contrast to an eight-fold increase in patients with PBC not treated with UDCA.45 A few studies have failed to demonstrate a survival benefit in UDCAtreated patients with PBC,8,46,47 but several flaws in study design, most notably small numbers of patients enrolled46 and use of suboptimal doses of UDCA,8,47 preclude meaningful conclusions.48 The experience in the United Kingdom8 is different from that of other European countries in that lower doses (6 to 8 mg/kg/day) of UDCA are used and therapy is started later in the course of the disease. The Mayo risk score, a cross-validated index of survival in PBC, decreases significantly during UDCA therapy. With UDCA therapy, the Mayo risk score (recalculated at six months of UDCA therapy) retains its validity in predicting survival, just as it does in the absence of effective therapy. The most cost-effective dose of UDCA in patients with PBC is 13 to 15 mg/kg/day, which can be given in one or two divided doses taken with meals. In patients also taking cholestyramine (or other bile-salt binding agents [see later]), UDCA should be taken at least two hours before or after cholestyramine to ensure intestinal absorption.

Budesonide is a newer glucocorticoid structurally related to 16α-hydroxyprednisolone, with extensive first-pass hepatic metabolism and minimal systemic availability. In a randomized, multicenter trial,49 79 patients with noncirrhotic (stage 1 to 3) PBC were enrolled, and 41 were randomized to treatment with oral budesonide (6 mg daily) in com­ bination with UDCA (15 mg/kg/day), whereas 36 received UDCA alone. At three years of treatment, the combination of budesonide and UDCA led to greater histologic improvement compared with UDCA alone. Side effects of gluco­ corticoids led to discontinuation of treatment in only one patient, and seven other patients reported mild glucocorticoid-related side effects. In an earlier, open-label study,50 22 patients with PBC who had experienced a suboptimal response to UDCA for a number of years were treated with oral budesonide (3 mg three times daily) for one year. The addition of budesonide to treatment with UDCA was associated with improvement in liver enzyme levels, without affecting other important prognostic markers such as the bilirubin level and Mayo risk score. In that study,50 the addition of budesonide was associated with a significant worsening of osteoporosis and cosmetic effects, particularly in those patients with more advanced (stage 3 to 4) PBC. A pharmacokinetic study showed that the hepatic metabolism of budesonide is markedly reduced in patients with stage 4 PBC compared with those who have stage 1 to 2 disease.51 In that study,51 portal vein thrombosis, possibly related to budesonide, developed in two of the seven patients with stage 4 PBC. Collectively, the data suggest that budesonide may be of potential benefit for patients with early-stage PBC but is associated with important systemic glucocorticoidrelated adverse events in patients with more advanced-stage disease. Therefore, before budesonide can be recommended for the treatment of PBC, appropriately designed controlled trials of long-term duration are necessary.

Colchicine

Three placebo-controlled trials of colchicine involving a total of 181 patients have been conducted. Colchicine was associated with some improvement in liver biochemical test values, but no effect on symptoms related to cholestasis, histologic progression to cirrhosis, or overall survival was seen. Analysis of these studies suggests that colchicine is not of benefit in patients with PBC.

Methotrexate

Patients with PBC who demonstrated clinical, biochemical, and histologic improvement with methotrexate therapy have been described in anecdotal reports. In a placebocontrolled trial of methotrexate for PBC, methotrexate in a dose of 7.5 mg per week for up to six years was not only of no benefit, but also was associated with more unfavorable outcomes than was observed with placebo. A large randomized trial evaluating UDCA (15 mg/kg/day) plus metho­ trexate (15 mg/m2 of body surface area weekly, maximal dose of 20 mg/week) versus UDCA plus placebo has been reported.52 In that study, 265 patients with PBC and a serum bilirubin level below 3 mg/dL were assigned to one of the

Chapter 89  Primary Biliary Cirrhosis two treatment groups; the mean period of study was 7.5 years. The hazard ratio for death with or without liver transplantation was no better in the methotrexate-UDCA combination group than in the UDCA-placebo group. Therefore, methotrexate should not be recommended routinely as monotherapy or as an adjuvant to UDCA.

Other Medications and Combination Therapy

Other medication such as d-penicillamine, azathioprine, chlorambucil, cyclosporine, malotilate, tetracycline, tacrolimus, thalidomide, and silymarin have been evaluated for the treatment of PBC, but no convincing evidence of efficacy was reported for any of these agents, and some were associated with serious adverse events. None of these medications can be recommended for the treatment of PBC at this time. Small pilot, open-label studies of short duration have been reported for mycophenolate mofetil, rituximab, bezafibrate, rifampin, lamivudine plus zidovudine (Combivir), and sulindac.53-56 Although some improvement in liver biochemical test levels was observed with these medications, none of these agents can be recommended outside of clinical trials. The use of combination therapy with drugs that have different properties has been evaluated in open and controlled trials. Combinations studied include UDCA and methotrexate, UDCA and colchicine, cyclosporine and prednisone, chlorambucil and prednisolone, UDCA and prednisone or prednisolone, UDCA and sulindac, and UDCA, prednisone, and azathioprine. Although some liver biochemical improvement in the short term has been reported with some of these combinations, the small numbers of patients enrolled, short follow-up period, and risk of drug-related side effects do not allow recommendation of any of these combinations for the treatment of PBC. Furthermore, none of these combinations seems to be more effective than UDCA alone. The encouraging results of the combination of UDCA with budesonide in patients with early PBC49 require confirmation in larger, controlled trials.

COMPLICATIONS OF CHRONIC CHOLESTASIS BONE DISEASE

Osteopenic bone disease with a predisposition to spontaneous fracturing is a common complication of chronic cholestatic liver disease. In North America, most patients with osteopenia from cholestasis have osteoporosis rather than osteomalacia. Osteoporosis is defined as defective bone formation, whereas osteomalacia is defective bone mineralization resulting from vitamin D deficiency. Women with PBC lose bone mass at a rate approximately twice that seen in age-matched controls, and this accelerated bone loss is the result of decreased formation rather than increased resorption of bone. The cause of osteoporosis associated with PBC is poorly understood, but the pathogenesis seems to be related to cholestasis itself. In one study,57 patients with PBC and a rate of bone loss higher than 2% per year were identified as those with more severe cholestasis, whereas in another study,58 bone mass was significantly lower in patients with stage 3 to 4 PBC than in the general population matched for age and gender, but such a difference was not seen for patients with stage 1 to 2 PBC.58 Genetic susceptibility for the development of osteoporosis in PBC has been suggested, including vitamin D receptor gene polymorphisms,57,59-61 collagen type Ia1 gene polymorphisms,57 and insulin-like growth factor-1 gene polymorphisms,62 but con-

tradictory results have been reported and no definitive genetic susceptibility and bone disease in PBC has been confirmed. The severity and progression of bone disease can be assessed by measurement of bone mineral density in different sites, in particular the lumbar spine and femur. Dualenergy x-ray absorptiometry and dual-photon absorptiometry are noninvasive techniques that quantify bone mass accurately. At the time of referral for or diagnosis of PBC, approximately 20% of patients have osteoporosis, as defined by a T-score below −2.5 in either the lumbar spine or the femoral neck, and approximately 10% have severe bone disease, as defined by a Z-score below −2.58,59 (The T-score is the number of standard deviations below the mean peak value in young gender-matched normal subjects, whereas the Z-score is the number of standard deviations below mean normal values corrected for age and gender.) The risk of osteoporosis (T-score below −2.5) is eight times higher in patients with PBC than in a gender-matched population, whereas the risk of severe bone disease (Z-score below −2) is four times higher in patients with PBC than in a healthy gender- and age-matched population.58 In patients with PBC, as in the general population, older age, postmenopausal status, and lower body-mass index are independent risk factors for the development of osteoporosis. In patients with PBC, however, the severity of osteoporosis increases as liver disease advances; bone mass in patients with stage 1 or 2 PBC is similar to that in a normal age- and gender-matched population, but bone mass is significantly lower in patients with stage 3 or 4 disease.58 Higher serum bilirubin levels, possibly as an indication of more advanced PBC, correlate significantly with a higher rate of bone loss.58 The reported cumulative frequency of fragility fractures in patients with PBC ranges from 10% to 26%, with a cumulative frequency of vertebral fractures of 10% to 20%. One half of patients with PBC who undergo liver transplantation have severe bone disease and one half of patients with PBC (almost exclusively those with preexisting osteopenia) experience a pathologic fracture during the first months after liver transplantation. Treatment of the bone disease includes adequate exercise and supplemental calcium (1200 to 1500 mg daily orally) and vitamin D (600 to 800 IU daily orally, or if deficiency is present, 25,000 to 50,000 IU orally once or twice per week). Treatment with estrogens significantly prevents loss of bone mass in postmenopausal patients with PBC63 and was not associated with worsening cholestasis in a series of 46 patients with PBC who received estrogen treatment for a mean period of almost five years.63 Because of the carcinogenic properties of estrogens, their lack of protective cardiovascular effects, a possible increased risk of dementia, and resumption of menses, however, postmenopausal women are not enthusiastic about taking estrogens. Raloxifene, a selective estrogen receptor modulator, looks promising as an alternative to estrogen replacement therapy for postmenopausal osteoporosis. Raloxifene was evaluated in a pilot study of nine postmenopausal women with PBC who showed a significant increase in bone mass after one year of treatment with raloxifene; this improvement in bone mass was not seen in an age- and menopausal status-matched control group.64 Raloxifene deserves evaluation in a large controlled trial in patients with PBC. Bisphosphonates also hold promise in the treatment of osteoporosis in patients with PBC. Although etidronate was no better than placebo in one randomized controlled study,65 alendronate (70 mg/ wk) was found to improve bone mass significantly after two years of treatment in 13 patients with PBC when compared with etidronate66 and in 15 patients with PBC when com-

1485

1486

Section IX  Liver pared with 13 patients treated with placebo.67 Other bisphosphonates, including risedronate and pamidronate, deserve evaluation, as do parathormone derivatives.68

FAT-SOLUBLE VITAMIN DEFICIENCY

Most patients with PBC and fat-soluble vitamin deficiency have advanced liver disease with jaundice. Fat-soluble vitamin deficiency is almost always caused by malabsorption resulting from decreased amounts of bile salts in the intestinal lumen. Vitamin D deficiency should be excluded in patients with PBC by measurement of 25-hydroxyvitamin D, a major metabolite of vitamin D. When vitamin D deficiency is encountered, vitamin D in a dose of 25,000 to 50,000 IU, given once or twice per week, usually is sufficient to achieve a normal serum vitamin D level. Because 25-hydroxylation of vitamin D is normal in patients with PBC, vitamin D (rather than the more expensive 25-hydroxyvitamin D or 1,25-hydroxyvitamin D) can be prescribed. Vitamin A deficiency, which can cause reduced night vision, can occur in patients with PBC. When blood levels of vitamin A are low and the patient is symptomatic, replacement therapy with oral vitamin A, 100,000 IU daily for 3 days, and then 50,000 IU daily for 14 days, should be instituted. If patients are deficient but asymptomatic, a dose of 25,000 to 50,000 IU two or three times per week is adequate. The adequacy of replacement therapy is assessed by repeating serum vitamin A assays and evaluating the patient for darkness adaptation, if indicated. Vitamin K deficiency occurs with severe cholestasis and is manifested by a prolonged prothrombin time. A trial of oral vitamin K, 5 to 10 mg daily, should be given to determine if the prothrombin time improves. If it does, the patient should be maintained on a water-soluble vitamin K, 5 mg per day. Deficiency of vitamin E has been reported in a few patients with PBC. Typically, vitamin E deficiency causes a neurologic abnormality that primarily affects the posterior columns and is characterized by areflexia, loss of proprioception, and ataxia. In patients with chronic cholestasis and low serum levels of vitamin E, oral replacement therapy with high-dose vitamin B may halt progression of neuropathy.

HYPERLIPIDEMIA

Lipid abnormalities are found in up to 85% of patients with PBC. High-density lipoprotein (HDL) levels usually are most

prominently elevated in the early stages of PBC; as the disease progresses, HDL levels decrease and low-density lipoprotein (LDL) levels increase (see Chapter 72). The risk of atherosclerosis in these patients with hyperlipidemia does not appear to be increased.69 Xanthelasmas (deposits of cholesterol in the skin) may develop in some patients with hyperlipidemia and can be troublesome. Therapy with UDCA has been shown to lower the LDL levels in patients with PBC and has been useful in some patients with xanthelasmas. Surgical removal of xanthelasmas is seldom successful, and such attempts should be avoided. The use of simvastatin has been evaluated in six patients with PBC and hypercholesterolemia.70 In all patients, serum LDL levels decreased, whereas HDL levels remained largely unchanged after two months of therapy; of interest, the serum alkaline phosphatase and gamma glutamyl transpeptidase levels also improved. Further evaluation of simvastatin for PBC is warranted.

PRURITUS

The cause of pruritus in patients with PBC remains an enigma. Various agents may provide symptomatic relief (Table 89-6). The bile acid-binding resin cholestyramine was the first medication described to alleviate this symptom. Therapy with cholestyramine is successful in a majority of patients who can tolerate the unpleasant side effects of bad taste, bloating, and occasional constipation. The recommended total dose is 3 to 12 g/day orally, and the drug is most effective when one half of the dose is given 30 minutes before and one half is given 30 minutes after breakfast, to permit maximal bile acid binding as the gallbladder empties. All drugs that can potentially bind to cholestyramine should be taken several hours before or after the cholestyramine. Other bile acid sequestrants such as colestipol and colesevelam have not been studied. Not all patients with pruritus are helped by cholestyramine. The antibiotic rifampin also is effective in relieving the pruritus of PBC. A majority of patients respond to rifampin, and benefit occurs within one week of the start of therapy. The starting dose is 150 mg twice daily orally; occasionally, higher doses are needed. Rifampin induces drug-metabolizing enzymes, so caution is needed when concurrent drugs are administered. Rifampin has been associated with liver injury in up to 15% of patients. Occasionally, treatment with UDCA alleviates pruritus, although on occasion pruritus may worsen with initiation of UDCA. In warm countries, exposure to ultraviolet light

Table 89-6  Medical Therapy of Cholestasis-Associated Pruritus DRUG

oral REGIMEN

EFFICACY

ADVERSE EFFECTS

Cholestyramine*

3-4 g 30 min before meals and 2 hours apart from ursodeoxycholic acid

Beneficial in most patients

Rifampin

150-300 mg two times a day

Beneficial in some, but not all, controlled trials to date

Ursodeoxycholic acid

13-15 mg/kg/day

Naltrexone

50 mg every day

Beneficial in intrahepatic cholestasis of pregnancy Beneficial in small controlled trials

Fat malabsorption, decreased intestinal absorption of other medications, constipation Inducer of hepatic enzymes involved in drug metabolism, potential hepatotoxicity, red-orange discoloration of urine and secretions No major toxicity reported

Antihistamines:   Diphenhydramine   Hydroxyzine

25-50 mg four times a day 25 mg three times a day

Rarely provide significant relief apart from sedation

Opiate withdrawal symptoms, rare hepatotoxicity Drowsiness

*Other bile salt sequestrants such as colestipol and colesevelam have not been studied as well as cholestyramine for the treatment of cholestasis-associated pruritus.

Chapter 89  Primary Biliary Cirrhosis without sun block can alleviate pruritus, and not surprisingly, the pruritus of PBC subsides during the summer months. The hypothesis has been proposed that pruritus may be related to the release of endogenous opioids. Intravenous infusion of the opiate receptor antagonist naloxone has shown a clear benefit in a double-blind trial. Oral opiate receptor antagonists such as nalmefene and naltrexone have led to amelioration of pruritus in patients with PBC, although further trials are needed to evaluate their safety. The serotonin antagonist ondansetron decreased the pruritus associated with cholestasis in studies using subjective methodology but not in a study of patients with PBC that applied behavior methodology (i.e., measurement of scratching activity).71 The serotonin reuptake inhibitor sertraline (75 to 100 mg orally) was associated with relief of pruritus as assessed by a visual analog scale and healing of excoriations as assessed by physical examination in a randomized, placebo-controlled trial.72 Because of their sedative effects, antihistamines such as diphenhydramine and hydroxyzine are helpful for treating the insomnia associated with pruritus, which is always more troublesome at night. Pheno­ barbital may have the same effect. The pruritus of PBC is almost always cured by liver transplantation, which is a viable option for patients with severe intractable pruritus (see Table 89-6).

STEATORRHEA

Steatorrhea can occur in patients with advanced PBC. Several causes have been described. The most important cause is decreased bile acid delivery with insufficient micellar concentration of bile acids in the small intestine (see Chapters 64 and 101). Occasionally, exocrine pancreatic insufficiency can be found as part of a widespread glandular dysfunction seen in some patients with PBC. Coexisting celiac disease has been reported in a small number of patients with PBC, and small intestinal bacterial overgrowth may be the cause of steatorrhea in some patients with PBC and scleroderma. Because each of these causes has specific and different treatments, determining the exact cause of steatorrhea is important. Patients with decreased intestinal bile acid concentrations usually benefit from substitution of medium-chain triglycerides for long-chain triglycerides in their diets and a decrease in total fat intake. Patients with exocrine pancreatic insufficiency will benefit from pancreatic replacement therapy; patients with celiac disease require gluten withdrawal from the diet; and patients with small intestinal bacterial overgrowth should receive intermittent broad-spectrum oral antibiotic therapy.

LIVER TRANSPLANTATION The best therapeutic alternative for patients with end-stage PBC is liver transplantation (see Chapter 95). The major manifestations of chronic liver disease that should prompt an evaluation for liver transplantation in patients with other causes of chronic liver disease apply to patients with PBC. These indications include complications related to portal hypertension, including bleeding from gastroeso­ phageal varices, diuretic-resistant ascites, hepatorenal syndrome, and hepatic encephalopathy. In patients with PBC, the development of complications associated with chronic cholestasis, such as a poor quality of life secondary to disabling fatigue, intractable pruritus, and severe muscle wasting, as well as persistent increases in the serum bilirubin level in the absence of hepatic malignancy, should prompt clinicians to consider referral for liver transplanta-

tion, even in patients without cirrhosis on a liver biopsy specimen. Data from the United Network for Organ Sharing show a clear trend toward decreased rates of liver transplantation for PBC.73 From 1995 to 2006, the absolute number of liver transplants in the United States increased an average of 249 cases per year, but the absolute number of transplants performed for PBC decreased by an average of 5.4 cases per year,73 despite the steady increase in the incidence and prevalence of PBC.2 A similar trend has been observed in Europe, where the proportion of liver transplants performed in patients with PBC decreased from 55% in the early years of transplantation (late 1980s) to 11% in 2006.73 Because UDCA is now prescribed nearly universally to patients with PBC, the decline in the number of liver transplants for PBC is consistent with improved survival resulting from treatment with UDCA. Liver transplantation clearly improves survival, as well as quality of life, for patients with PBC. One-year survival rates after liver transplantation are currently higher than 90%, with five-year survival rates higher than 80% in most transplant centers. The advent of prognostic models has helped identify factors that predict survival (see Table 89-5). The Mayo risk score (http://www.mayoclinic.org/gi-rst/ mayomodel2.html), which is based on the patient’s age, serum bilirubin level, serum albumin level, prothrombin time, presence of edema, and need for treatment with diuretics is superior to the Child-Turcotte-Pugh score in predicting survival. The Model for End-stage Liver Disease (MELD) score (http://www.mayoclinic.org/meld/mayomodel6.html) is also a reliable measure of mortality risk in patients with end-stage liver disease, including PBC, and is used as a disease severity index to determine organ allocation priorities (see Chapter 95).75 PBC-specific autoantibodies against mitochondria (AMA) and gp210 protein generally persist in a patient’s serum after liver transplantation. PBC may recur in the allograft, with a frequency of recurrence ranging from 11% to 34% and a median time to diagnosis of recurrence of 36 to 61 months after transplantation.76-81 The risk of recurrence increases with time, so that by 10 years, histologic recurrence may be found in 30% to 50% of patients. Tacrolimusbased immunosuppression has been the most consistently identified risk factor for disease recurrence, although older recipient age and male gender of the recipient also have been identified as risk factors for recurrence. One study found that the use of UDCA in patients with recurrent PBC improves serum alkaline phosphatase and aminotransferase levels at three years post-transplantation but has no effect on progression of histologic stage.81 Recurrent PBC following liver transplantation does not seem to decrease survival significantly, although in some studies a small proportion of patients had graft failure.

AMA-NEGATIVE PRIMARY BILIARY CIRRHOSIS As discussed earlier, AMA are detected in the serum of 90% of patients with PBC by the indirect immunofluorescence technique and in 95% of patients by the immunoblotting technique. AMA-negative PBC is the designation for those patients who clinically, biochemically, and histologically appear to have the classic features of PBC but are found not to have AMA in serum by indirect immunofluorescence or immunoblotting techniques. Of patients who have PBC by all other criteria, 5% are confirmed AMA negative.82 Terms such as “autoimmune cholangitis” and “autoimmune chol-

1487

1488

Section IX  Liver angiopathy” have been used in the past to describe this entity, but the current consensus holds that there is no practical difference between AMA-positive and AMAnegative PBC and, therefore, no need for an alternative label to describe AMA-negative PBC.83 Most patients with AMA-negative PBC have antinuclear (perinuclear/rim-like or multiple nuclear dot pattern) or smooth muscle antibodies (or both). Although these patients may be distinguished by the lack of AMA in serum, the specific AMA antigen PDC-E2 is expressed on the apical region of their biliary epithelium, as occurs in AMApositive patients—an observation suggesting that the pathogenesis of both conditions may be identical. Whether a different genetic susceptibility exists for the development of AMA-positive PBC and AMA-negative PBC is still uncertain. In one study, the frequencies of HLA DRb1*08 and DQb1*04 were reduced significantly in patients with AMAnegative PBC as compared with AMA-positive patients.84 When regulatory T cells and the subgroup of T cells suggested to have a role in the genesis of autoimmune disease were examined in patients with PBC, however, no difference was found between those who had and those who did not have AMA.85 Patents with AMA-negative PBC tend to follow a clinical course, and to demonstrate a therapeutic response to UDCA, similar to those in AMA-positive patients.83,86 When a patient with AMA-negative PBC is evaluated, other diseases that may manifest in a similar manner should be excluded. The absence of AMA makes liver biopsy mandatory to look for features of PBC and rule out other liver diseases. Also, imaging by magnetic resonance cholangiopancreatography (MRCP) is essential to identify other cholangiopathies such as primary sclerosing cholangitis. Liver biopsy and MRCP, along with select laboratory tests, will allow the exclusion of conditions that should be considered in the differential diagnosis, such as celiac disease, hepatitis C, sarcoidosis, small-duct primary sclerosing cholangitis, and IgG4-associated autoimmune cholangitis (see Chapter 68).

Patients with AMA-negative PBC should be treated with UDCA in a dose of 13 to 15 mg/kg/day. When histologic features of superimposed autoimmune hepatitis are detected, the combination of glucocorticoids and UDCA should be considered (see Chapter 88).86

KEY REFERENCES

Invernizzi P, Selmi C, Mackay IR, et al. From bases to basis: Linking genetics to causation in primary biliary cirrhosis. Clin Gastroenterol Hepatol 2005; 3:401-10. (Ref 19.) Jackson H, Solaymani-Dodaran M, Card TR, et al. Influence of ursodeoxycholic acid on the mortality and malignancy associated with primary biliary cirrhosis: A population-based cohort study. Hepatology 2007; 46:1131-7. (Ref 45.) Jacob DA, Neumann UP, Bahra M, et al. Long-term follow-up after recurrence of primary biliary cirrhosis after liver transplantation in 100 patients. Clin Transplant 2006; 20:211-20. (Ref 80.) Jones DE, Bhala N, Burt J, et al. Four year follow-up of fatigue in a geographically defined primary biliary cirrhosis cohort. Gut 2006; 55:536-41. (Ref 31.) Lazaridis KN, Juran BD, Boe GM, et al. Increased prevalence of antimitochondrial antibodies in first-degree relatives of patients with primary biliary cirrhosis. Hepatology 2007; 46:785-92. (Ref 5.) Mayo MJ, Handem I, Saldana S, et al. Sertraline as a first-line treatment for cholestatic pruritus. Hepatology 2007; 45:666-74. (Ref 72.) Menon KV, Angulo P, Weston S, et al. Bone disease in patients with primary biliary cirrhosis: Independent predictors and rate of progression. J Hepatol 2001; 35:316-23. (Ref 58.) Pares A, Caballeria J, Rodes J. Excellent long-term survival in patients with primary biliary cirrhosis and biochemical response to ursodeoxycholic acid. Gastroenterology 2006; 130:715-20. (Ref 43.) Poupon RE, Lindor KD, Pares A, et al. Combined analysis of the effect of treatment with ursodeoxycholic acid on histologic progression in primary biliary cirrhosis. J Hepatol 2003; 39:12-16. (Ref 35.) Prince M, James OFW. The epidemiology of primary biliary cirrhosis. Clin Liver Dis 2003; 7:795-819. (Ref 2.) Zein CO, Angulo P, Lindor KD. When is liver biopsy needed in the diagnosis of primary biliary cirrhosis? Clin Gastroenterol Hepatol 2003; 1:89-95. (Ref 32.) Zein C, Jorgensen RA, Clarke B, et al. Alendronate improves bone mineral density in patients with PBC: A randomized placebocontrolled trial. Hepatology 2005; 42:762-71. (Ref 67.) Full references for this chapter can be found on the www.expertconsult.com.

CHAPTE R

90 Portal Hypertension and Gastrointestinal Bleeding Vijay H. Shah and Patrick S. Kamath

CHAPTER OUTLINE Normal Portal Circulation  1489 Hemodynamic Principles of Portal Hypertension  1491 Increased Intrahepatic Resistance  1492 Hyperdynamic Circulation  1493 Collateral Circulation and Varices  1494 Measurement of Portal Pressure  1495 Hepatic Vein Pressure Gradient  1495 Splenic Pulp Pressure  1496 Portal Vein Pressure  1496 Endoscopic Variceal Pressure  1496 Detection of Varices  1496 Upper Gastrointestinal Endoscopy  1496 Ultrasonography  1497 Computed Tomography  1497 Magnetic Resonance Imaging  1498 Endoscopic Ultrasonography  1498

Variceal hemorrhage, hepatic encephalopathy, and ascites—the major complications of cirrhosis of the liver—result from portal hypertension, defined as an increase in hepatic sinusoidal pressure to 6 mm Hg or greater. Portosystemic collaterals decompress the hypertensive hepatic sinusoids and give rise to varices at the gastroesophageal junction and elsewhere. These portosystemic collaterals also may allow ammonia derived from the intestine to reach the brain, thereby resulting in hepatic encephalopathy through a pathologic process of several intermediary steps involving the peripheral benzodiazepine-type receptors, neuro­ steroids, and γ-aminobutyric acid (GABA) receptors (see Chapter 92). Additionally, portal hypertension is associated with renal retention of sodium and water and the formation of ascites (see Chapter 91). Indeed, portal hypertension and its complications remain important clinical problems despite advances in treatment and improved understanding of both the molecular basis and pathophysiology of portal hypertension.

NORMAL PORTAL CIRCULATION The portal venous system carries capillary blood from the esophagus, stomach, small and large intestine, pancreas, gallbladder, and spleen to the liver. The portal vein is formed by the confluence of the splenic vein and the superior mesenteric vein behind the neck of the pancreas.1 The inferior mesenteric vein usually drains into the splenic

Causes of Portal Hypertension  1498 Common Causes  1498 Less Common Causes  1500 Clinical Assessment of Patients with Portal Hypertension-Related Bleeding  1501 Treatment of Portal Hypertension-Related Bleeding  1501 Pharmacologic Therapy  1501 Endoscopic Therapy  1503 Transjugular Intrahepatic Portosystemic Shunt  1504 Surgical Therapy  1506 Management of Specific Lesions  1508 Esophageal Varices  1508 Gastric Varices  1510 Ectopic Varices  1513 Portal Hypertensive Gastropathy and Gastric Vascular Ectasia  1514

vein. The left gastric vein, also called the left coronary vein, usually drains into the portal vein at the confluence of the splenic vein and superior mesenteric vein (Fig. 90-1). The portal vein is approximately 7.5 cm in length and runs dorsal to the hepatic artery and bile duct into the hilum of the liver. The uppermost 5 cm of the portal vein does not receive any tributaries.2 In the hilum of the liver, the portal vein divides into the left and right portal vein branches, which supply the left and right sides of the liver, respectively. The umbilical vein drains into the left portal vein. The cystic vein from the gallbladder drains into the right portal vein, whereas the portal venules drain into hepatic sinusoids that, in turn, are drained by the hepatic veins into the inferior vena cava. The left and middle hepatic veins usually join and drain into the inferior vena cava separately but adjacent to the confluence of the right hepatic vein with the inferior vena cava. The caudate lobe drains separately into the inferior vena cava (see Chapter 71). The circulatory system of the normal liver is a highcompliance, low-resistance system that is able to accommodate a large blood volume, as occurs after a meal, without substantially increasing portal pressure. The liver receives a dual blood supply from the portal vein and the hepatic artery that constitutes nearly 30% of total cardiac output. Portal venous blood derived from the mesenteric venous circulation constitutes approximately 75% of total hepatic blood flow, whereas the remainder of blood to the liver is derived from the hepatic artery, which provides highly oxygenated blood directly from the celiac trunk of the aorta. Portal vein–derived and hepatic artery–derived blood flow

1489

1490

Section IX  Liver converge in high-compliance, specialized vascular channels termed hepatic sinusoids. A dynamic and compensatory interplay occurs between hepatic blood flow derived from the portal vein and that from the hepatic artery. Specifically, when portal venous blood flow to the liver is diminished, as occurs in portal vein thrombosis, arterial inflow increases in an attempt to maintain total hepatic blood flow at a constant level. Similarly, after hepatic artery occlusion, portal venous inflow increases in a compensatory manner. This autoregulatory mechanism, aimed at maintaining total hepatic blood flow at a constant level, is termed the hepatic arterial buffer response. The sinusoids are highly permeable and thus facilitate the transport of macromolecules to the parenchymal hepatocytes that reside on the extraluminal side of the endothelial cells. The hepatic sinusoids are highly permeable because they lack a proper basement membrane and because the

Inferior vena cava Left, right, and middle hepatic veins

Portal vein

Liver

Splenic vein

Spleen

Superior mesenteric vein

Hepatic sinusoids

Left colon Inferior mesenteric vein

Right colon/small intestine Figure 90-1.  Anatomy of the portal circulation. Blood vessels that constitute the portal circulation and hepatic outflow tracts are depicted.

endothelial cells that line the sinusoids contain fenestrae. Other unique aspects of the hepatic sinusoids are the space of Disse, a virtual space located extraluminal to the endothelial cell and adjacent to the hepatocyte, and its cellular constituents, the hepatic stellate cell (HSC) and the Kupffer cell (Fig. 90-2; see also Chapters 71 and 72). These two cell types probably play an important role, in concert with the endothelial cell, in regulating sinusoidal hemodynamics and homeostasis and may contribute to the sinusoidal derangements that occur in portal hypertension. Under basal conditions, HSCs maintain a quiescent phenotype and accumulate vitamin A. On activation, however, as occurs in cirrhosis and portal hypertension, these cells are postulated to develop contractile abilities that permit them to function as sinusoidal pericytes. Kupffer cells contribute to vascular homeostasis by generating cytokines with potent cellular and vasoregulatory actions, including tumor necrosis factor. Endothelial cells and smooth muscle cells in nonsinusoidal hepatic vessels such as the portal venule and the terminal hepatic venule are important in hepatic vasoregulation, particularly in the normal liver, where HSCs are quiescent, unactivated, and presumably less contractile. Many studies have established the important role of nitric oxide (NO), derived from endothelial NO synthase (eNOS), in hepatic vasodilatation. Shear stress, caused by the frictional force of blood within the sinusoids, is one of the most potent physiologic stimuli of eNOS-derived NO production in hepatic sinusoids. By contrast, endothelin-1 (ET-1), also released by endothelial cells, promotes hepatic vasoconstriction by binding to ET-A receptors located on HSCs. ET-1 also appears to be generated within HSCs themselves and promotes HSC contraction through an autocrine loop. Of interest, ET-1 may alternatively bind to ET-B receptors on endothelial cells. This signaling pathway paradoxically promotes vasodilatation by activating eNOS. Other vascular mediators implicated in hepatic vasoregulation include carbon monoxide generated by the heme oxygenase system, the sympathetic adrenergic agonist norepinephrine, the renin-dependent vasoconstrictor angiotensin, prostaglandins, thromboxane, leukotrienes, and hydrogen sulfide. Of these mediators, angiotensin is of particular interest because it is a potent constrictor of HSCs and is released in increased amounts in cirrhosis owing to systemic sympathetic hyperactivity. A number of angiotensin receptor blockers

Normal

Cirrhosis Hepatocytes Deposition of collagen in space of Disse

Hepatic stellate cell Space of Disse

Activation of hepatic stellate cells

Sinusoidal endothelial cell Fenestrae

A

Sinusoidal lumen

Constriction of sinusoids

B

Defenestration of sinusoids

Figure 90-2.  Anatomy of the hepatic microvasculature. A, Normal sinusoidal microanatomy is depicted. The sinusoidal lumen is lined by fenestrated sinusoidal endothelial cells that allow the transport of macromolecules to the abluminal space of Disse. Quiescent hepatic stellate cells reside within this space, adjacent to hepatocytes and endothelial cells. B, In cirrhosis, a number of changes occur in the hepatic microcirculation, including loss of fenestrae in endothelial cells (defenestration), constriction of sinusoids, and activation of hepatic stellate cells with ensuing deposition of collagen and increased contractility.

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding are undergoing evaluation for the treatment of portal hypertension.

HEMODYNAMIC PRINCIPLES OF PORTAL HYPERTENSION In cirrhosis, as well as in most noncirrhotic causes of portal hypertension, portal hypertension results from changes in portal resistance in combination with changes in portal inflow. The influence of flow and resistance on pressure can be represented by the formula for Ohm’s law: ∆P = F × R in which the pressure gradient in the portal circulation (ΔP) is a function of portal flow (F) and resistance to flow (R).

Increases in portal resistance or portal flow can contribute to increased pressure. Portal hypertension almost always results from increases in both portal resistance and portal flow (Fig. 90-3). One exception is that of an arteriovenous fistula, which in the initial stages causes portal hypertension largely through an increase in portal flow in the absence of an increase in resistance. The mechanism of the increase in portal resistance depends on the site and cause of portal hypertension; in the Western world, the most common cause is liver cirrhosis (see later). Because of the increase in hepatic resistance and the decrease in hepatic compliance, small changes in flow that do not increase pressure in the normal liver can have a prominent stimulatory effect on portal pressure in the cirrhotic liver. The increase in portal venous inflow is part of a generalized systemic derangement termed the hyperdynamic circulatory state. Collateral vessels that dilate and new vascular sprouts that form connect the high-pressure portal venous system with

Beta adrenergic blockers

Surgical and transjugular intrahepatic shunts Sclerotherapy Variceal ligation

Hepatic circulation EC ↓ NO Nitrates Angiotensin receptor blockers Prazosin

Collateral circulation

↑ ET-1

EC

CC ↑ NO

Constriction

↑ NO

Dilatation Angiogenesis

Beta adrenergic blockers Angiogenesis inhibitors

Splanchnic circulation EC ↑ NO Somatostatin analogs Vasopressin analogs Beta adrenergic blockers Nitrates

CC Dilatation

Figure 90-3.  Vascular disturbances in portal hypertension and sites of action of portal pressure-reducing therapies. Portal hypertension typically results from increased resistance, usually from within the liver, in combination with increased portal venous flow. The increase in hepatic resistance results from mechanical factors in combination with dynamic vasoconstriction mediated by decreased nitric oxide (NO) production and increased endothelin-1 (ET-1) production. The increase in portal venous flow occurs as a result of vasodilatation in the splanchnic circulation that is mediated by increased NO production. A collateral circulation, including esophageal varices, develops between the hypertensive portal vasculature and systemic venous system; however, these collaterals are inadequate to decompress the hypertensive portal circulation fully. Collateral development is mediated by dilatation of existing collateral vessels, as well as development of new blood vessels and sprouts (angiogenesis). Therapies aimed at the different sites of hemodynamic disturbances are shown. CC, contractile cell (e.g., hepatic stellate cell, vascular smooth muscle cell); EC, endothelial cell.

1491

1492

Section IX  Liver

Vasodilators Adenosine Carbon monoxide Endothelium-derived hyperpolarizing factor Hydrogen sulfide Nitric oxide Prostacyclin Serotonin

Vasoconstrictors Angiotensin Endothelin-1 Leukotrienes Norepinephrine Serotonin Thromboxane

Figure 90-4.  Representative vasodilator and vasoconstrictor molecules implicated in the vascular abnormalities in portal hypertension.

lower-pressure systemic veins. Unfortunately, this process of angiogenesis and collateralization is insufficient for normalizing portal pressure and actually causes complications of portal hypertension, such as esophageal varices.3 Approaches to block this angiogenic process are a compelling target for drug development. The changes in portal flow and resistance also can be viewed as originating from mechanical and vascular factors. Mechanical factors include the fibrosis and nodularity of the cirrhotic liver with distortion of the vascular architecture and the remodeling that is recognized to occur in the systemic and splanchnic vasculature in response to the chronic increases in flow and shear stress that characterize the hyperdynamic circulatory state. Vascular factors include intrahepatic vasoconstriction, which contributes to increased intrahepatic resistance, and the splanchnic and systemic vasodilatation that accompanies the hyperdynamic circulatory state. The vascular factors that contribute to portal hypertension are particularly important because they are reversible and dynamic and therefore compelling targets for experimental therapies (Fig. 90-4). Conversely, effective therapies for the fixed, mechanical component of portal hypertension caused by scar, regenerative nodules, and vascular remodeling are currently lacking. Indeed, most available therapies for portal hypertension focus on correction of hemodynamic alterations in the portal circulation. Approaches include use of nonselective b-adrenergic blocking agents, octreotide, and vasopressin to reduce the hyperdynamic circulation, portal venous inflow, and splanchnic vasodilatation.4,5 Alternative agents reduce the increased intrahepatic resistance and include angiotensin receptor blockers and mononitrates.

INCREASED INTRAHEPATIC RESISTANCE

In cirrhosis, increased portal resistance occurs in great part as a result of mechanical factors that reduce vessel diameter. In addition to regenerative nodules and fibrotic bands, these mechanical factors include capillarization of the sinusoids and swelling of cells, including hepatocytes and Kupffer cells. As discussed earlier, however, reduced hepatic vessel diameter resulting in increased portal resistance, even when caused by cirrhosis, is not a purely mechanical phenomenon.6 Hemodynamic changes in the hepatic circulation also contribute to increased intrahepatic resistance.7,8 These changes are characterized by hepatic vasoconstriction and impaired responses to vasodilatory stimuli. The increase in intrahepatic resistance is determined largely by changes in vessel radius, with small reductions in vessel radius causing prominent increases in resistance. Blood viscosity and vessel length also can influ-

ence resistance, albeit to a much smaller extent. The factors that regulate resistance can be viewed in the context of the law of Poiseuille: R = 8ηL πr 4 in which R is resistance, ηL is the product of blood viscosity and vessel length, and r is vessel radius. Although vasoactive changes were estimated initially to account for 10% to 30% of the increase in portal resistance in cirrhosis, subsequent studies have suggested that these figures actually may underestimate the contribution of hepatic vasoconstriction to the increased resistance observed in the cirrhotic liver. In noncirrhotic causes of portal hypertension, the increase in resistance may occur at sites upstream (prehepatic) or downstream (posthepatic) of the liver, as in portal vein thrombosis and hepatic vein thrombosis, respectively (Fig. 90-5). Furthermore, the site of increased intrahepatic resistance can be further delineated as the sinusoids (sinusoidal), upstream from the sinusoids within the portal venules (presinusoidal), or downstream from the sinusoids in the hepatic venules (postsinusoidal), as in alcoholic cirrhosis, schistosomiasis, and sinusoidal obstruction syndrome, respectively. Pressure is increased only in the portal circulation behind the site of increased resistance, and in isolated portal vein thrombosis, hepatic function frequently remains largely preserved despite prominent portal hypertension. Most evidence suggests that a decrease in the production of the vasodilator NO and an increase in the production of the vasoconstrictor ET-1 jointly contribute to the increase in hepatic vascular resistance. In experimental models of cirrhosis, the bioavailability of hepatic NO is diminished because of a reduction in the production of NO by endothelial cells.7,9,10 A similar paradigm is observed in the human cirrhotic liver.11 Most studies indicate that the reduction in NO production occurs not through a reduction in hepatic eNOS protein levels9,10 but through defects in the steps necessary to activate existing eNOS protein. For example, increases in the production of the eNOS-inhibiting protein caveolin-1 have been observed in experimental models of cirrhosis10 and in human cirrhosis. Another pathway that contributes to deficient generation of NO by eNOS is a reduction in the level of AKT (protein kinase B) phosphorylation of eNOS and upregulation of the eNOS inhibiting protein, GRK (G protein-coupled receptor kinase), in the cirrhotic liver.12 Irrespective of the mechanism of deficiency, the lack of availability of NO is thought to allow HSCs, which are activated and highly contractile in liver cirrhosis, to constrict the sinusoids that they envelop, thereby increasing portal pressure. The role of the HSCs in this process remains controversial, however, because evidence is mixed regarding whether the site of the increase in intrahepatic resistance in cirrhosis is the sinusoids, where stellate cells reside, or the pre- or postsinusoidal venules (or both), which are devoid of stellate cells and in which endothelial cells signal smooth muscle cells. Furthermore, increasing evidence points toward diverse origins of these myofibroblastic cells within the cirrhotic sinusoids, with portal myofibroblasts as well as HSC postulated to play important roles.13 In this regard, therapies that target myofibroblast migration may be a compelling therapeutic target by limiting the density of these contractile cells within the hepatic sinusoids.14 Finally, the contribution of HSCs to hepatic angiogenesis may also be an important target for treating fibrosis and portal hypertension.15 In clinical practice, NO can be delivered by NO donor agents such as mononitrates. NO donor agents exert their

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding

Posthepatic Intrahepatic Prehepatic Budd-Chiari syndrome Presinusoidal Portal vein thrombosis Constrictive pericarditis Idiopathic portal hypertension Splenic vein thrombosis Inferior vena caval obstruction Primary biliary cirrhosis Right-sided heart failure Sarcoidosis Severe tricuspid regurgitation Schistosomiasis Sinusoidal Alcoholic cirrhosis Alcoholic hepatitis Cryptogenic cirrhosis Postnecrotic cirrhosis Postsinusoidal Sinusoidal obstruction syndrome

beneficial effects in part by relaxing the actively contractile stellate cells.16,17 The systemic actions of these agents, however, tend to cause side effects and exacerbate the hyperdynamic circulatory state. In studies utilizing a liverspecific NO donor compound, the increased intrahepatic vascular resistance could be corrected by the generation of additional NO and consequent relaxation of HSCs.18 In cirrhosis, however, deficient endothelial cell NO generation may be accompanied by impaired stellate cell relaxation in response to NO,19 perhaps because of diminished response of the NO second messenger cyclic guanosine monophosphate (cGMP) in activated cells.16 In this situation, a prominent beneficial effect of NO donors is less predictable. Excessive ET-1 also contributes to increased intrahepatic vasoconstriction in portal hypertension through vasoconstrictive effects in the liver, presumably by enhancing HSC contractility.20,21 In experimental models, ET-1 protein and receptor expression are increased, most notably in HSCs and endothelial cells.20,22,23 In humans with portal hypertension, plasma and liver ET-1 levels also are increased.24 The reason for activation of the ET-1 system in portal hypertension is not known, but this effect may be secondary to transforming growth factor-β (TGF-β), a key fibrogenic growth factor.23 Clinical trials of ET antagonists in patients with portal hypertension are in progress; however, the variable effects of ET modulation in experimental models of portal hypertension, as well as the possible hepatotoxicity of these compounds, have limited enthusiasm for studies in humans.25 Other therapies for portal hypertension may provide benefit through the ET pathway. For example, somatostatin, which reduces portal pressure by constricting the splanchnic circulation, also may act by inhibiting ET1–dependent HSC contraction.26

Figure 90-5.  Classification of portal hypertension. The different sites of increased resistance to portal flow (posthepatic, intrahepatic, and prehepatic) and associated diseases are shown. Many diseases cause a mixed pattern. Portal hypertension rarely can occur exclusively as a result of increased portal flow, as occurs with arteriovenous shunts (not shown).

Other vasoactive mediators, including cysteinyl leukotrienes, thromboxane, angiotensin, and hydrogen sulfide, also have been implicated in the development of increased intrahepatic resistance in cirrhosis.27,28 Some of these mediators, particularly angiotensin, which causes contraction of HSCs, have been studied in humans. Attempts to reduce portal pressure using pharmacologic agents that inhibit angiotensin activation of HSC contraction have met with mixed results thus far.29

HYPERDYNAMIC CIRCULATION

In addition to the increases in portal resistance discussed earlier, a major factor in the development and perpetuation of portal hypertension is an increase in portal venous flow, or the hyperdynamic circulation. The term portal venous inflow indicates the total blood that drains into the portal circulation, not the blood flow in the portal vein itself, which may actually be diminished in portal hypertension because of portosystemic collateral shunts. The hyperdynamic circulation is characterized by peripheral and splanchnic vasodilatation, reduced mean arterial pressure, and increased cardiac output. Vasodilatation, particularly in the splanchnic bed, permits an increase in inflow of systemic blood into the portal circulation.30 Splanchnic vasodilatation is caused in large part by relaxation of splanchnic arterioles and ensuing splanchnic hyperemia. Studies of experimental portal hypertension have demonstrated that splanchnic vascular endothelial cells are primarily responsible for mediating splanchnic vasodilatation and enhanced portal venous inflow through excess generation of NO.31-39 This excess generation of NO and ensuing vasodilatation, hyperdynamic circulation, and hyperemia in the splanchnic and systemic circulation con-

1493

1494

Section IX  Liver trasts with the hepatic circulation, in which NO deficiency contributes to increased intrahepatic resistance. The mechanism of excess NO production from the endothelial cells of the systemic and splanchnic arterial circulation is an area of active investigation. Some of the increase in NO production probably occurs from shear stress– dependent and shear stress–independent increases in the expression of eNOS, which can be corrected in part by beta blockers.37,40-45 Activation of existing eNOS by cytokines or mechanical factors also seems to contribute to excess systemic and splanchnic NO generation through pathways that include eNOS phosphorylation and protein interactions.42-46 The physiologic stimuli that mediate this process are not well understood but may include ET-1, which is increased in the serum of patients with portal hypertension, and the cytokine tumor necrosis factor-α (TNF-α) because inhibitors of TNF improve portal pressure and the splanchnic circulatory disturbances in both human and experimental portal hypertension. TNF-α may be derived from intestinal endotoxin, and intestinal decontamination appears to correct the hyperdynamic circulation in humans, suggesting a link with intestinal inflammation.47 Vascular endothelial growth factor (VEGF) has also been implicated in this process by excessively activating eNOS.48 In humans with portal hypertension, therapeutic inhibition of NOS has met with mixed clinical results. In one study, inhibition of NOS corrected altered systemic hemodynamics,49 but other studies have not demonstrated significant portal pressure–reducing effects of systemic NOS inhibition.50 Other mediators that may contribute to systemic and splanchnic vasodilatation include anandamide, an endogenous vasodilatory cannabinoid,51-53 heme oxygenase,17,54-56 and cyclooxygenase.57 Compelling evidence also supports a primary defect in smooth muscle cells in portal hypertension, perhaps because of defects in potassium channels.58-62 In fact, many pharmacologic therapies for portal hypertension target the splanchnic arteriolar smooth muscle cells, rather than endothelial cells, to reduce splanchnic vasodilatation. For example, octreotide, a synthetic analog of somatostatin, causes marked but transient reductions in portal pressure by contracting splanchnic smooth muscle cells, thereby limiting portal venous inflow, especially after meals. Nonselective beta blockers and vasopressin also reduce portal pressure by constricting splanchnic arterioles and thereby reducing portal venous inflow. Because intrahepatic resistance persists, therapies targeted toward the increase in portal venous inflow usually do not normalize portal pressure entirely but often blunt the prominent increases in portal venous inflow that occur in response to a meal. Combination therapy with an agent that reduces increased intrahepatic resistance, such as a nitrate, and an agent that reduces portal venous inflow, such as a beta blocker, are more effective in reducing portal pressure than is either agent alone.

COLLATERAL CIRCULATION AND VARICES

The portal vein–systemic collateral circulation develops and expands in response to elevation of the portal pressure.63 Blood flow in the low volumes that normally perfuse these collaterals and flow toward the portal circulation is reversed in portal hypertension because the increased portal pressure exceeds systemic venous pressure. Therefore, flow is reversed in these collateral vessels, and blood flows out of the portal circulation toward the systemic venous circulation. The sites of collateral formation are the rectum, where the inferior mesenteric vein connects with the pudendal vein and rectal varices develop; the umbilicus, where the vesti-

gial umbilical vein communicates with the left portal vein and gives rise to prominent collaterals around the umbilicus (caput medusae); the retroperitoneum, where collaterals, especially in women, communicate between the ovarian vessels and iliac veins; and the distal esophagus and proximal stomach, where gastroesophageal varices are the major collaterals formed between the portal venous system and systemic venous system. Four distinct zones of venous drainage at the gastroesophageal junction are particularly relevant to the formation of esophageal varices.64 The gastric zone, which extends for 2 to 3 cm below the gastroesophageal junction, comprises veins that are longitudinal and located in the submucosa and lamina propria. They come together at the upper end of the cardia of the stomach and drain into short gastric and left gastric veins. The palisade zone extends 2 to 3 cm proximal to the gastric zone into the lower esophagus. Veins in this zone run longitudinally and in parallel in four groups corresponding to the esophageal mucosal folds. These veins anastomose with veins in the lamina propria. The perforating veins in the palisade zone do not communicate with extrinsic (periesophageal) veins in the distal esophagus. The palisade zone is the dominant watershed area between the portal and systemic circulations. More proximal to the palisade zone in the esophagus is the perforating zone, where there is a network of veins. These veins are less likely to be longitudinal and are termed perforating veins because they connect the veins in the esophageal submucosa and the external veins. The truncal zone, the longest zone, is approximately 10 cm in length, located proximally to the perforating zone in the esophagus, and usually characterized by four longitudinal veins in the lamina propria. Veins in the palisade zone in the esophagus are most prone to bleeding because no perforating veins at this level connect the veins in the submucosa with the periesophageal veins. Varices in the truncal zone are unlikely to bleed because the perforating vessels communicate with the periesophageal veins, allowing the varices in the truncal zone to decompress. The periesophageal veins drain into the azygous system, and as a result, an increase in azygous blood flow is a hallmark of portal hypertension. The venous drainage of the lower end of the esophagus is through the coronary vein, which also drains the cardia of the stomach, into the portal vein. The fundus of the stomach drains through short gastric veins into the splenic vein. In the presence of portal hypertension, varices may therefore form in the fundus of the stomach. Splenic vein thrombosis usually results in isolated gastric fundal varices. Because of the proximity of the splenic vein to the renal vein, spontaneous splenorenal shunts may develop and are more common in patients with gastric varices than in those with esophageal varices.65,66 The predominant collateral flow pattern in intrahepatic portal hypertension is through the right and left coronary veins, with only a small portion of flow through the short gastric veins. Therefore, most patients with intrahepatic causes of portal hypertension have esophageal varices or gastric varices in continuity with esophageal varices. Unfortunately, portal hypertension caused by cirrhosis generally persists and progresses despite the development of even an extensive collateral circulation. Progression of portal hypertension results from (1) the prominent obstructive resistance in the liver; (2) resistance within the collaterals themselves; and (3) continued increase in portal vein inflow. The collateral circulatory bed develops through a combination of angiogenesis, the development of new blood vessels, and dilatation and increased flow through preexisting collaterals.3,67 Experimental evidence suggests that VEGF, a key NO

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding stimulatory growth factor, may contribute to both the angiogenic and collateral vessel responses.55,68 Inhibition of VEGF or NO may attenuate the collateral vessel propagation by inhibiting angiogenic responses in experimental models of portal hypertension and collateralization.67-72 Some pharmacologic agents used in the management of portal hypertension, such as beta blockers and octreotide, may act in part by constricting collateral vessels.73-76 Approaches to inhibiting VEGF and angiogenesis are worth studying therapeutically.48 The development of gastroesophageal varices requires a portal pressure gradient of at least 10 mm Hg. Furthermore, a portal pressure gradient of at least 12 mm Hg is thought to be required for varices to bleed; other local factors that increase variceal wall tension also are needed77 because all patients with a portal pressure gradient of greater than 12 mm Hg do not necessarily bleed. Factors that influence variceal wall tension can be viewed in the context of the law of Laplace: T = Pr w where T is variceal wall tension, P is the transmural pressure gradient between the variceal lumen and esophageal lumen, r is the variceal radius, and w is the variceal wall thickness. When the variceal wall thins and the varix increases in diameter and pressure, the tolerated wall tension is exceeded and the varix will rupture. These physiologic observations are manifested clinically by the observation that patients with larger varices (r) in sites of limited soft tissue support (w), with elevated portal pressure (P), tend to be at greatest risk for variceal rupture from variceal wall tension (T) that becomes excessive. One notable site in which soft tissue support is limited is at the gastroesophageal junction. The lack of tissue support and high vessel density may contribute to the greater frequency of bleeding from varices at the gastroesophageal junction. The law of Laplace also has implications for the relevance of pharmacologic therapies aimed at reducing portal pressure. Reductions in portal pressure will reduce the variceal transmural pressure gradient, thereby reducing the risk that variceal wall tension will become excessive and varices will rupture. Clinically, a reduction in the hepatic venous pressure gradient to less than 12 mm Hg almost negates the risk of variceal hemorrhage. The changes in portal pressure and local variceal factors, however, are dynamic and influenced by a number of physiologic (an increase in intra-abdominal pressure, meal-induced increases in portal pressure), diurnal (circadian changes in portal pressure), and pathophysiologic (acute alcohol use) factors, and portal pressure and esophageal variceal pressure may vary at different times.

MEASUREMENT OF PORTAL PRESSURE Portal pressure may be measured indirectly or directly. The most commonly used method of measuring portal pressure is determination of the hepatic vein pressure gradient (HVPG), which is an indirect method. Measurement of splenic pulp pressure and direct measurement of the portal vein pressure are invasive, cumbersome, and infrequently used approaches. Variceal pressure also can be measured but is not routinely performed in clinical practice. Measurement of liver stiffness using ultrasound fibroelastography or magnetic resonance elastography (MRE) may indicate the presence of portal hypertension but cannot yet be used to measure portal pressure.

HEPATIC VEIN PRESSURE GRADIENT

The HVPG is the difference between the wedged hepatic venous pressure (WHVP) and free hepatic vein pressure (FHVP). The HVPG has been used to assess portal hyper­ tension since its first description in 1951,78 and has been validated as the best predictor for the development of complications of portal hypertension. Measurement of the HVPG requires passage of a catheter into the hepatic vein under radiologic guidance until the catheter can be passed no further, that is, until the catheter has been “wedged” in the hepatic vein. The catheter can be passed into the hepatic vein through the femoral vein or using a transjugular venous approach. The purpose of wedging the catheter is to form a column of fluid that is continuous between the hepatic sinusoids and the catheter. Therefore, the measured pressure of fluid within the catheter reflects hepatic sinusoidal pressure. One of the drawbacks of using a catheter that is wedged in the hepatic vein is that the WHVP measured in a more fibrotic area of liver may be higher than the pressure measured in a less fibrotic area because of regional variation in the degree of fibrosis. Using a balloon-occluding catheter in the right hepatic vein to create a stagnant column of fluid in continuity with the hepatic sinusoids eliminates this variation in measurement of WHVP because the balloon catheter measures the WHVP averaged over a wide segment of the liver.79 HVPG is not effective for detecting presinusoidal causes of portal hypertension. For example, in portal hypertension secondary to portal vein thrombosis, the HVPG is normal. Moreover, the HVPG may underestimate sinusoidal pressure in primary biliary cirrhosis and other presinusoidal causes of portal hypertension.80 Therefore, HVPG is accurate for detecting only sinusoidal and postsinusoidal causes of portal hypertension. The HVPG represents the gradient between the pressure in the portal vein and the intra-abdominal inferior vena caval pressure. An elevation in intra-abdominal pressure increases both WHVP and FHVP equally, so that the HVPG is unchanged. The advantage of the HVPG is that variations in the “zero” reference point have no impact on the HVPG.81 The HVPG is measured at least three times to demonstrate that the values are reproducible. Total occlusion of the hepatic vein by the inflated balloon to confirm that the balloon is in a wedged position is demonstrated by injecting contrast into the hepatic vein. A sinusoidal pattern should be seen, with no collateral circulation to other hepatic veins. The contrast washes out promptly with deflation of the balloon. Correct positioning of the balloon also is demonstrated by a sharp increase in the recorded pressure on inflation of the balloon. The pressure then becomes steady until the balloon is deflated, when the pressure drops sharply. In experienced hands, measurement of the HVPG is highly reproducible, accurate, and safe. Measurement of the HVPG has been proposed for the following indications: (1) to monitor portal pressure in patients taking drugs used to prevent variceal bleeding; (2) as a prognostic marker82; (3) as an end-point in trials using pharmacologic agents for the treatment of portal hypertension83; (4) to assess the risk of hepatic resection in patients with cirrhosis; and (5) to delineate the cause of portal hyper­ tension (i.e., presinusoidal, sinusoidal, or postsinusoidal (Table 90-1), usually in combination with venography, right-sided heart pressure measurements, and transjugular liver biopsy. Although the indication for HVPG measurement with the most potential for widespread use is monitoring the efficacy of therapies to reduce portal pressure, HVPG monitoring is not done routinely in clinical practice because no controlled trials have yet demonstrated its usefulness.84

1495

1496

Section IX  Liver Table 90-1  The Use of Hepatic Venous Pressure Gradient in the Differential Diagnosis of Portal Hypertension TYPE OF PORTAL HYPERTENSION

WHVP

FHVP

HVPG

Prehepatic Presinusoidal Sinusoidal Postsinusoidal Posthepatic   Heart failure   Budd-Chiari syndrome

Normal Normal Increased Increased

Normal Normal Normal Normal

Normal Normal Increased Increased

Increased —

Increased Hepatic vein cannot be cannulated

Normal —

FHVP, free hepatic vein pressure; HVPG, hepatic venous pressure gradient; WHVP, wedged hepatic venous pressure.

SPLENIC PULP PRESSURE

Determination of splenic pulp pressure is an indirect method of measuring portal pressure and involves puncture of the splenic pulp with a needle catheter. Splenic pulp pressure is elevated in presinusoidal portal hypertension, when the HVPG is normal. Because of the potential risk of complications, especially bleeding, associated with splenic puncture, however, the procedure is rarely used.

PORTAL VEIN PRESSURE

Direct measurement of the pressure in the portal vein is a rarely used method that can be carried out through a percutaneous transhepatic route, transvenous approach, or, rarely, intraoperatively (although anesthesia can affect portal pressure). The transhepatic route requires portal vein puncture performed under ultrasound guidance. A catheter is then threaded over a guidewire into the main portal vein. With increasing use of the transjugular intrahepatic portosystemic shunt (TIPS) (see later), radiologists have gained expertise in puncturing the portal vein and measuring portal vein pressure by a transjugular route. Direct portal pressure measurements are carried out when HVPG cannot be measured, as in patients with occluded hepatic veins caused by the Budd-Chiari syndrome, in whom a surgical portosystemic shunt is being contemplated,85 or in patients with intrahepatic, presinusoidal causes of portal hypertension, such as idiopathic portal hypertension, in which the HVPG may be normal.

ENDOSCOPIC VARICEAL PRESSURE

Varices rupture and bleed when the expanding force of intravariceal pressure exceeds variceal wall tension. Measurement of the difference between intravariceal pressure and pressure within the esophageal lumen (the transmural pressure gradient across the varices) is potentially a more important indicator of bleeding risk than measurement of HVPG,86,87 especially in patients with portal vein thrombosis and other causes of portal hypertension associated with a normal HVPG. Variceal pressure can be measured by inserting a needle connected to a pressure transducer, through a fluid-filled catheter, into a varix; this approach is currently not justified except when measurement of variceal pressure is followed by variceal injection sclerotherapy. Because variceal sclerotherapy has fallen out of favor (see later), measurement of variceal pressure by variceal puncture is seldom carried out. A miniature pneumatic pressure sensitivity gauge attached to the tip of an endoscope (Varipres Solid Components, Barcelona, Spain) allows noninvasive measurement of variceal pressure. Patients with previous variceal bleeding have been demonstrated to have higher variceal pressures than those in patients without previous bleeding.88 A variceal pressure greater than 18 mm Hg during a bleeding episode is associated with failure to control bleeding and

predicts early rebleeding.89 Moreover, patients on pharmacologic therapy who show a decrease in variceal pressure of greater than 20% from baseline have a low probability of bleeding, as compared with patients who do not demonstrate a greater than 20% decrease in variceal pressure, in whom the risk of variceal bleeding is 46%.88 Variceal pressure measurements determined with use of Varipres are considered satisfactory when they meet the following criteria: (1) a stable intraesophageal pressure; (2) absence of artifacts caused by esophageal peristalsis; and (3) correct placement of the gauge over the varix, as shown by fine fluctuations in the pressure tracing that correspond to the cardiac cycle and respirations. Therefore, measurement of variceal pressure requires both a skilled endoscopist and a cooperative patient, and, even in expert hands, accurate variceal pressure measurements cannot be obtained in 25% of patients. Manometry using an endoscopic balloon to measure variceal pressure is subject to observer bias because it relies on visual appearance to determine whether the varices have collapsed.90-92 With this technique, a balloon is inserted into the esophagus and inflated until the varices are noted on endoscopy to collapse. The pressure in the balloon required to collapse the varices represents the variceal pressure. In general, techniques of measuring variceal pressure are still considered experimental and not suitable for routine clinical use.

DETECTION OF VARICES UPPER GASTROINTESTINAL ENDOSCOPY

Upper gastrointestinal endoscopy is the most commonly used method to detect varices. The current consensus is that all patients with cirrhosis of the liver should be screened for esophageal varices by endoscopy. In patients in whom no varices are detected on initial endoscopy, endoscopy to look for varices should be repeated in 2 to 3 years. If small varices are detected on the initial endoscopy, endoscopy should be repeated in 1 to 2 years.93,94 None of the various noninvasive methods of determining which patients benefit most from endoscopic screening are accurate enough to recommend for routine use in clinical practice.95 The role of noninvasive markers in predicting the risk of large esophageal varices requires study in large multicenter trials. Preliminary data suggest that wireless video capsule endoscopy (see Chapter 19)96 and computed tomography (CT) imaging are alternative screening modalities in patients who are not candidates for upper endoscopy. Moreover, CT screening may be more cost-effective than endoscopy.97 Endoscopic grading of esophageal varices is subjective. Various criteria have been used to try to standardize the reporting of esophageal varices. The best known of these

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding

A

B

Figure 90-6.  Endoscopic appearances of esophageal varices. A, Upper gastrointestinal endoscopy demonstrates dilated and straight veins (small esophageal varices) in the distal esophagus (arrows). B, Upper gastrointestinal endoscopy demonstrates large esophageal varices, greater than 5 mm in diameter, with a fibrin plug (arrow) representing the site of a recent bleed.

criteria are those compiled by the Japanese Research Society for Portal Hypertension. The descriptors include red color signs, color of the varix, form (size) of the varix, and location of the varix.98 Red color signs include red “wale” markings, which are longitudinal whip-like marks on the varix; cherry-red spots, which usually are 2 to 3 mm or less in diameter; hematocystic spots, which are blood-filled blisters 4 mm or greater in diameter; and diffuse redness. The color of the varix can be white or blue. The form of the varix at endoscopy is described most commonly. Esophageal varices may be small and straight (grade I); tortuous and occupying less than one third of the esophageal lumen (grade II); or large and occupying more than one third of the esophageal lumen (grade III). Varices can be in the lower third, middle third, or upper third of the esophagus. Of all of the aforementioned descriptors, the size of the varices in the lower third of the esophagus is the most important. The size of the varices in the lower third of the esophagus is determined during withdrawal of the endoscope (Fig. 90-6). As much air as possible should be aspirated from the stomach while the esophageal lumen is fully inflated. Small varices, that is, those occupying less than one third of the lumen, are less than 5 mm in diameter, whereas large varices are greater than 5 mm in diameter.98,99 As a point of reference, any varix larger in diameter than an open pinch biopsy forceps is likely to be greater than 5 mm in diameter. Patients with large esophageal varices, Child (or Child-Pugh) class C cirrhosis (see later), and red color signs on varices have the highest risk of variceal bleeding within 1 year.100 The increase in bleeding risk attributable to the presence of red color signs, however, is not independent of the risk associated with large variceal size. Therefore, prophylactic treatment to prevent variceal bleeding is recommended in all patients with large esophageal varices irrespective of the presence or absence of red color signs (see later).

ULTRASONOGRAPHY

Ultrasound examination of the liver with Doppler study of the vessels has been used widely to assess patients with portal hypertension. Features suggestive of portal hyperten-

sion on ultrasonography include splenomegaly, portosystemic collateral vessels, and reversal of the direction of flow in the portal vein (hepatofugal flow). Some studies have demonstrated that a portal vein diameter greater than 13 mm and the absence of respiratory variations in the splenic and mesenteric veins are sensitive but nonspecific markers of portal hypertension.101,102 These criteria are not used routinely in clinical practice in most centers. Ultrasound examination can detect thrombosis of the portal vein, which appears as nonvisualization or cavernous transformation (a cavernoma) of the portal vein; the latter finding indicates an extensive collateral network in place of the portal vein.103 Splenic vein thrombosis also can be demonstrated. Portal blood flow can be measured by Doppler ultrasonography, which is the easiest research method for detecting post­ prandial increases in splanchnic blood flow.104 Although Doppler ultrasonography is clinically useful in the initial evaluation of portal hypertension, the technique is not widely used to provide quantitative assessments of the degree of portal hypertension. Transient elastography may be useful in detecting portal hypertension but is not sufficiently sensitive to recommend as a modality to monitor decreases in portal pressure in patients on pharmacotherapy (see Chapter 73).105

COMPUTED TOMOGRAPHY

Computed tomography (CT) is useful for demonstrating many features of portal hypertension, including abnormal configuration of the liver, ascites, splenomegaly, and collateral vessels (Fig. 90-7). Detection of varices may be an emerging indication for CT. Diagnosis of fundal varices by multidetector row CT (MDCT) is at least as accurate as endoscopic ultrasonography (see later). CT is especially helpful in distinguishing submucosal from perigastric fundal varices106 and is considered a less invasive alternative to conventional angiographic portography in assessing portosystemic collaterals. At present, however, CT is not a recommended screening method for detecting large esophageal varices, but it may be a cost-effective method of screening for varices and preferred to endoscopy by patients.97

1497

1498

Section IX  Liver

A

B

C

D

Figure 90-7.  Abdominal computed tomography (CT) in patients with portal hypertension. A, CT scan shows an irregular contour of the liver typical of cirrhosis (arrowheads). A small right pleural effusion is evident (straight arrow). The liver is hypointense relative to the spleen (curved arrow), typical of fatty infiltration of the liver in alcoholic cirrhosis. B, Coronal section of a CT scan showing contrast-enhanced esophageal varices (cursor). C, CT scan demonstrating two large esophageal varices (arrows) 5 mm and 6 mm in diameter. Varices are almost opposed to each other. D, CT scan demonstrates a tuft of gastroesophageal collaterals (straight arrow). The enlarged spleen also is seen (curved arrow).

MAGNETIC RESONANCE IMAGING

Gadolinium-enhanced magnetic resonance imaging (MRI) is becoming recognized as a potentially useful method of detecting esophageal varices.107 In addition, MRI can be used to measure portal and azygous blood flow, which is increased in patients with portal hypertension.108 MRI provides excellent detail of the vascular structures of the liver and can detect portal venous thrombosis and spleen stiffness in patients with portal hypertension, but the role of MRI in the assessment of portal hypertension requires further study. Unlike transient elastography using ultrasound, MRI can accurately assess the stiffness of even fatty livers.109

ENDOSCOPIC ULTRASONOGRAPHY

Endoscopic ultrasound examination (endosonography) using radial or linear array echo-endoscopes or endoscopic ultrasound mini-probes passed through the working channel of a diagnostic endoscope has been applied as an investigational tool in the evaluation of patients with varices. Endoscopic ultrasonography has been used to study several aspects of esophageal varices, including the cross-sectional area of varices to identify patients at increased risk of bleeding77; size of and flow in the left gastric vein, azygous vein, and paraesophageal collaterals; changes after endoscopic therapy; and recurrence of esophageal varices following variceal ligation (see later).110 Endosonography can be combined with endoscopic measurement of transmural variceal pressure to allow estimation of variceal wall tension, which is a predictor of variceal bleeding (see earlier).111-113

CAUSES OF PORTAL HYPERTENSION The usual classification of causes of portal hypertension is based on the site of increased resistance to portal blood flow—namely, prehepatic, intrahepatic, and posthepatic— and is outlined in Figure 90-5. Intrahepatic sites of increased resistance can be presinusoidal, sinusoidal, or postsinusoidal. Many causes of portal hypertension are associated with an increase in resistance at more than one site. For example, alcoholic cirrhosis may be associated with increased resistance at the presinusoidal, sinusoidal, and postsinusoidal levels. Therefore, classification based on the site of resistance may not be possible for all diseases that cause portal hypertension. A more useful classification is clinically based and considers common and less common causes of portal hypertension (Table 90-2).

COMMON CAUSES Cirrhosis

Complications related to portal hypertension are the usual clinical manifestations of cirrhosis of the liver. Although all causes of cirrhosis are associated with portal hypertension, some features are disease specific. In alcoholic liver disease, elevation of the portal pressure is accurately reflected by the HVPG; moreover, portal hypertension may occur in the absence of cirrhosis but is more marked when cirrhosis is present. Perivenular lesions implicated in the pathogenesis in noncirrhotic alcoholic liver injury account for the presinusoidal component of portal hypertension in these

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding Table 90-2  Causes of Portal Hypertension Common Cirrhosis Extrahepatic portal vein thrombosis Idiopathic portal hypertension Schistosomiasis Less Common Fibropolycystic liver disease Hereditary hemorrhagic telangiectasia Malignancy Myeloproliferative disorders Nodular regenerative hyperplasia Partial nodular transformation of the liver Sarcoidosis Splanchnic arteriovenous fistula

patients.114 Autoimmune hepatitis also may be associated with portal hypertension in the absence of cirrhosis115; however, the risk of variceal bleeding is low in patients with autoimmune hepatitis. In patients with hemochromatosis, portal hypertension may be seen even before cirrhosis; the severity of portal hypertension increases with increasing fibrosis. Patients with hemochromatosis may bleed from varices despite an HVPG less than 12 mm Hg, indicating a presinusoidal component of portal hypertension. Phlebotomy therapy in patients with hemochromatosis may result in a decrease in portal hypertension.116 In patients with primary biliary cirrhosis, portal hypertension also may occur before cirrhosis has developed. The risk of variceal bleeding increases with an increase in the histologic stage of the disease.117 In earlier stages of primary biliary cirrhosis, portal hypertension is predominantly presinusoidal, but as the disease progresses, a sinusoidal component develops. Therefore, the HVPG may underestimate portal pressure in patients with primary biliary cirrhosis.80 Portal hypertension occurs in patients with primary sclerosing cholangitis and in those with biliary strictures. A long duration of biliary obstruction usually is required, although portal hypertension has been known to develop in a few months in patients with chronic bile duct obstruction caused by chronic alcoholic pancreatitis.118 Portal hypertension in patients with biliary obstruction regresses following relief of the biliary obstruction.

Schistosomiasis

Schistosomiasis may be the most common cause of portal hypertension worldwide (see Chapter 82). Bleeding from esophageal varices is a major cause of death in patients with hepatosplenic schistosomiasis. Portal hypertension results from presinusoidal obstruction caused by deposition of eggs of Schistosoma mansoni and Schistosoma japonicum in the presinusoidal portal venules. The host reaction results in granulomatous inflammation, which causes presinusoidal and periportal fibrosis.119 The fibrosis that results is sometimes called “clay pipestem” or simply “pipestem” fibrosis and usually is associated with sustained heavy infection. The periportal collagen deposition leads to progressive obstruction of portal blood flow, portal hypertension, and variceal bleeding, along with splenomegaly and hypersplenism. Lobular architecture usually is preserved. Coinfection with hepatitis B or C virus in patients with hepatic schistosomiasis can result in more rapid progression of fibrosis, hepatic failure, and an increased risk of hepatocellular carcinoma.120 In the initial stages of schistosomiasis, the HVPG is normal owing to the presinusoidal nature of the obstruction.

Figure 90-8.  Computed tomography image of choledochal varices (arrows).

Some patients with schistosomiasis and portal hypertension also may have portal vein thrombosis. Patients with schistosomiasis often undergo portosystemic shunt surgery to treat variceal bleeding, with excellent long-term outcomes.

Extrahepatic Portal Vein Thrombosis

Extrahepatic portal vein thrombosis is a prehepatic, presinusoidal cause of portal hypertension and a common cause of portal hypertension in children (see Chapter 83). The most common causes of portal vein thrombosis include hematologic disorders such as polycythemia vera or other myeloproliferative disorders. Other causes include a prothrombotic state, such as antithrombin, protein C, or protein S deficiency; antiphospholipid syndrome (or antiphos­ pholipid antibody syndrome); paro­xysmal nocturnal hemoglobinuria; oral contraceptive use; a neoplasm, usually intra-abdominal; an inflammatory disease, such as pan­ creatitis, inflammatory bowel disease, or diverticulitis; abdominal trauma; and postoperative states, especially postsplenectomy. Cirrhosis is a cause of portal vein thrombosis.121 Older studies suggested that portal vein thrombosis occurs in approximately 6% of patients with cirrhosis and in up to 25% of those with cirrhosis and hepatocellular carcinoma.122 With improved imaging, portal vein thrombosis is now known to be a more common complication of cirrhosis, and the association with hepatocellular carcinoma may not be as strong as previously thought. Isolated splenic vein thrombosis caused by a pancreatic neoplasm or pancreatitis usually is not associated with a thrombophilia. Umbilical vein sepsis may be an etiologic factor in children with portal vein thrombosis, but even in these cases, an associated prothrombotic state may predispose the patient to portal vein thrombosis. Acute and subacute portal vein thrombosis usually does not manifest with variceal bleeding.1 Chronic portal vein thrombosis is suggested by nonvisualization of the portal or splenic vein and an extensive collateral circulation. Patients may present with nonspecific symptoms or with variceal bleeding and hypersplenism. Bleeding usually is from gastroesophageal varices but may be from duodenal varices and, rarely, other ectopic sites. Gallbladder varices (Fig. 90-8) also have been described in patients with portal vein thrombosis.123 The treatment of portal vein thrombosis is symptomatic, with the aim of controlling variceal bleeding or preventing

1499

1500

Section IX  Liver recurrent variceal bleeding. Patients in whom esophageal varices are not large, and a thrombophilia is detected, are best managed with anticoagulation because in these patients, the benefits of anticoagulation outweigh the risks.124 Local or systemic thrombolytic therapy is seldom required and is generally reserved for patients in whom a portal vein thrombus extends into the superior mesenteric vein, with danger of impending intestinal ischemia. Endoscopic therapy is used to control acute variceal bleeding and to prevent recurrent bleeding. Use of pharmacologic agents such as beta blockers to prevent variceal bleeding is probably also effective in patients with portal vein thrombosis, but this approach has not been well studied. Patients with portal vein thrombosis have lower mortality and morbidity rates from variceal bleeding than those reported in patients with cirrhosis and variceal bleeding, owing to the lack of coagulopathy and synthetic liver dysfunction. Surgical portosystemic shunt procedures are carried out in patients in whom bleeding cannot be controlled by conservative measures. If a suitable vein is not available for anastomosis, a large collateral vein may be anastomosed to a systemic vein.125 Placement of a TIPS is possible in some patients with chronic portal vein thrombosis.

Idiopathic Portal Hypertension

Idiopathic portal hypertension is uncommon in Western countries but is common in parts of Asia such as India and Japan. This disorder is diagnosed when the portal pressure is elevated in the absence of significant histologic changes in the liver or extrahepatic portal vein obstruction.126 A liver biopsy specimen from affected patients may be entirely normal,122 although increased concentrations of ET-1 have been noted in the periportal hepatocytes, portal venules, and hepatic sinusoids of patients with idiopathic portal hypertension.127 Various terms used (rather loosely) to describe idiopathic portal hypertension include hepatoportal sclerosis, noncirrhotic portal fibrosis, and Banti’s syndrome.128,129 Use of the term idiopathic portal hypertension probably is best restricted to portal hypertension in patients in whom no hepatic lesion is found on light microscopy. The term hepatoportal sclerosis suggests obliterative portal venopathy with subendothelial thickening of the intrahepatic portal veins; thrombosis and recanalization of these veins may follow. Fibrosis of the portal tracts is prominent later in the course. The cause of idiopathic portal hypertension is unclear in a majority of patients, although chronic arsenic intoxication, exposure to vinyl chloride, and hypervitaminosis A have been implicated (see Chapter 87). These etiologic factors are present in only a minority of patients. The dominant clinical features of the condition are variceal bleeding and hypersplenism related to a markedly enlarged spleen. Liver biochemical test levels are usually normal, although the serum alkaline phosphatase level may be mildly elevated. Ascites is uncommon. The HVPG in this disorder usually is normal because the site of increased resistance is presinusoidal.130 Surgical portosystemic shunts are well tolerated in these patients, although hepatic encephalopathy may occur on long-term follow-up evaluation.122 Liver transplantation is rarely required in these patients. Idiopathic portal hypertension may be confused with incomplete septal cirrhosis, which probably is an unrelated condition characterized by incomplete septa and liver nodularity.131 Patients with incomplete septal cirrhosis are clinically similar to patients with cirrhosis and may progress to end-stage liver disease and require liver transplantation.

LESS COMMON CAUSES Nodular Regenerative Hyperplasia

Nodular regenerative hyperplasia is a histopathologic diagnosis characterized by atrophy of zone 3 hepatocytes and hypertrophy of zone 1 hepatocytes without significant fibrosis (see Chapters 35 and 94).132 This disorder has been recognized increasingly as a cause of portal hypertension and may even occur after liver transplantation.133 Similar histologic changes may be seen in well-established Budd-Chiari syndrome.134 The nodular hyperplasia may not be apparent on histologic examination unless a reticulin stain is carried out to demonstrate the micronodules. These regenerative nodules are believed to result from an imbalance between hyperperfused areas of the liver, with resulting regenerative nodules, and poorly perfused areas, with resulting atrophy. Nodular regenerative hyperplasia is associated with a variety of conditions, predominantly hematologic and rheumatologic in nature. Liver biochemical abnormalities include mild elevation of the serum aminotransferase levels. Portal hypertension manifesting as variceal bleeding is the predominant clinical presentation. Ascites also may develop in these patients, suggesting that an increase in sinusoidal pressure occurs.135 Hepatocellular carcinoma does not occur, but liver transplantation may be required in some patients.

Partial Nodular Transformation of the Liver

Partial nodular transformation of the liver is an uncommon lesion that is characterized by large nodules in the perihilar region.136 These nodules may be visible on imaging studies of the liver. The rest of the liver may be normal or may show changes of nodular regenerative hyperplasia. Liver biochemical test levels usually are normal. Like nodular regenerative hyperplasia, partial nodular transformation of the liver is believed to be related to an imbalance in portal perfusion of the liver, but the abnormality is restricted to the hilar branches, whereas in nodular regenerative hyperplasia the abnormality is more diffuse. Variceal bleeding is the predominant presentation in partial nodular transformation of the liver, although patients with large nodules may experience abdominal pain. Hepatocellular carcinoma may rarely develop in these regenerating nodules. Treatment with a surgical portosystemic shunt is associated with good long-term results.

Fibropolycystic Liver Disease

Fibropolycystic liver disease is a term which encompasses Caroli’s disease, Caroli’s complex (Caroli’s disease with congenital hepatic fibrosis), congenital hepatic fibrosis, and polycystic liver disease. Congenital hepatic fibrosis usually occurs in association with Caroli’s disease of the liver, polycystic disease of the kidney, and medullary sponge kidney (see Chapter 62). The major manifestation of congenital hepatic fibrosis is variceal bleeding.137 A portosystemic shunt may be placed in these patients to treat refractory variceal bleeding, with a low long-term risk of hepatic encephalopathy. Patients with polycystic liver disease, whether associated with polycystic kidney disease or not, rarely present with portal hypertension (see Chapter 94).138 Portal hypertension may decrease after treatment of the cysts.

Sarcoidosis

Portal hypertension is an uncommon manifestation of hepatic sarcoidosis (see Chapter 35).139 The site of increased intrahepatic resistance in patients with sarcoidosis seems to be postsinusoidal, in view of the elevated HVPG. In early disease, however, the resistance is predominantly at a

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding presinusoidal level. Treatment with glucocorticoids may decrease portal hypertension in some patients with hepatic sarcoidosis.

Malignancy

Portal hypertension has been associated with leukemias, lymphomas, and systemic mastocytosis (see Chapters 34 and 35).140 Portal hypertension also may occur in patients with hepatocellular carcinoma independent of the presence of cirrhosis (see Chapter 94). The pathogenesis of portal hypertension in patients with hepatocellular carcinoma is thought to be multifactorial; contributing factors include portal vein thrombosis, pressure by the tumor on the portal vein, and, in some cases, a hepatic artery–portal vein fistula. Esophageal varices may be seen in patients with hepatic metastases, although variceal bleeding is unusual.141

Splanchnic Arteriovenous Fistula

Splanchnic arteriovenous fistula should be suspected when the onset of ascites and variceal bleeding is acute, especially in the presence of an abdominal bruit. When a splanchnic artery ruptures into a mesenteric vein, the portal pressure increases acutely, reaching levels of systemic arterial pressure.142 The result is acute portal hypertension with development of ascites and variceal bleeding. A bruit may be heard in the left upper quadrant of the abdomen with a splenic arteriovenous fistula and in the right upper quadrant with a hepatic artery–portal vein fistula. With a longstanding fistula, secondary hepatic changes of perisinusoidal fibrosis related to an increase in portal venous inflow may be present. In the early stages, embolization or ligation of the fistula will ameliorate the portal hypertension. In late stages, however, portal fibrosis may be advanced, and the portal hypertension may not correct completely with embolization of the fistula.

Hereditary Hemorrhagic Telangiectasia

Hereditary hemorrhagic telangiectasia (HHT), or OslerWeber-Rendu disease, is an unusual cause of portal hypertension (see also Chapters 19 and 36). Diagnostic criteria include mucocutaneous telangiectasias, epistaxis, arteriovenous fistulas of the viscera (usually lung or liver), and a family history of the disorder. Manifestations of HHT depend on the site of fistula formation. A fistula between the hepatic artery and hepatic vein manifests predominantly as biliary disease, mainly biliary strictures and cho­ langitis, and high-output cardiac failure. A fistula between the hepatic artery and portal vein results in portal hypertension and biliary strictures, whereas a fistula between the portal vein and hepatic vein, which is rare, results in hepatic encephalopathy.143 Nodular regenerative hyperplasia, which develops in some patients with HHT, may worsen portal hypertension.144 Although symptomatic liver disease in HHT is rare, involvement of the liver is found in a majority of patients.145

CLINICAL ASSESSMENT OF PATIENTS WITH PORTAL HYPERTENSION-RELATED BLEEDING Patients with esophageal or gastric variceal bleeding present with hematemesis or melena (or both). Chronic blood loss is a more common presentation of portal hypertensive gastropathy or gastrointestinal vascular ectasia. The classic presentation of patients with variceal bleeding is with effortless and recurrent hematemesis; the vomitus is described as dark red in color.

Portal hypertension should be suspected in all patients with gastrointestinal bleeding and peripheral stigmata of liver disease—namely, jaundice, spider angiomata, palmar erythema, Dupuytren’s contractures, parotid enlargement, testicular atrophy, loss of secondary sexual characteristics, ascites, and encephalopathy. Splenomegaly is an important clue to the presence of portal hypertension, and the presence of ascites makes the presence of esophageal varices even more likely. Caput medusae, suggestive of an intrahepatic cause of portal hypertension, is present around the umbilicus; the flow of blood is away from the umbilicus. In Budd-Chiari syndrome, by contrast, veins are dilated in the flanks and back, and blood flows in a cephalic direction.85 A bruit may be heard in the left or the right upper quadrant in a patient with a splanchnic arteriovenous fistula. A venous hum may be heard in the epigastrium of a patient with portal hypertension and represents collateral flow in the falciform ligament. Laboratory studies frequently reveal evidence of hepatic synthetic dysfunction, including prolongation of the prothrombin time, hypoalbuminemia, and hyperbilirubinemia, as well as anemia. Thrombocytopenia and leukopenia, reflecting hypersplenism and, in alcoholics, bone marrow suppression, may be noted. Patients with severe bleeding may present with hypovolemic shock and renal insufficiency. Abdominal imaging studies frequently reveal splenomegaly, collateral vessels, abnormal liver echotexture and contour, and ascites.

TREATMENT OF PORTAL HYPERTENSION-RELATED BLEEDING The treatment of portal hypertension is aimed either at reducing portal blood flow with pharmacologic agents— such as beta blockers or vasopressin and its analogs—or at decreasing intrahepatic resistance, with pharmacologic agents—such as nitrates or by radiologic or surgical creation of a portosystemic shunt. Treatment also may be directed at the varices with use of endoscopic or radiologic techniques.

PHARMACOLOGIC THERAPY

The pharmacologic agents used in the treatment of portal hypertension are divided into two groups: those that decrease splanchnic blood flow and those that decrease intrahepatic vascular resistance (Table 90-3). The agents that decrease splanchnic blood flow acutely are vasopressin and its analogs and somatostatin and its analogs. badrenergic blocking agents (beta blockers) also decrease portal blood flow but are used only to prevent variceal bleeding and rebleeding. Agents that target intrahepatic vascular resistance include α-adrenergic blocking agents,

Table 90-3  Drugs Used in the Treatment of Portal Hypertension Drugs That Decrease Portal Blood Flow Nonselective β-adrenergic blocking agents Somatostatin and its analogs Vasopressin Drugs That Decrease Intrahepatic Resistance α1-Adrenergic blocking agents (e.g., prazosin) Angiotensin receptor blocking agents Nitrates

1501

1502

Section IX  Liver angiotensin receptor blocking agents, and nitrates, but only nitrates are now considered for clinical use. Diuretics, by decreasing plasma volume, may reduce portal pressure but are not recommended as sole agents for the treatment of portal hypertension. Metoclopramide and other gastric prokinetic agents may decrease intravariceal pressure by contracting the lower esophageal sphincter, but these agents have not been evaluated in clinical trials and are not recommended.

Vasopressin and Its Analogs

Vasopressin is an endogenous peptide hormone that causes splanchnic vasoconstriction, reduces portal venous inflow, and reduces portal pressure. This drug is associated with serious systemic side effects, however. By causing constriction of systemic vessels, vasopressin may result in necrosis of the bowel. Additionally, vasopressin has direct negative inotropic and chronotropic effects on the myocardium that lead to reduced cardiac output and bradycardia, respectively. An increase in cardiac afterload can result in myocardial infarction, and antidiuresis, resulting from the action of vasopressin on the kidney, can result in hyponatremia. Terlipressin, or triglycyl-lysine-vasopressin, is a semisynthetic analog of vasopressin that is cleaved by endothelial peptidases to release lysine vasopressin. Compared with vasopressin, terlipressin results in lower circulatory levels of the vasopressin analog and a lower rate of systemic side effects. Vasopressin and terlipressin have been used in combination with nitrates to decrease the risk of systemic side effects. Terlipressin is preferred over vasopressin because of its superior safety profile. In addition, an increase in survival has been demonstrated in patients with variceal bleeding treated with terlipressin. Terlipressin is not currently available in the United States, although it is undergoing evaluation by the U.S. Food and Drug Administration (FDA).

Somatostatin and Its Analogs

Somatostatin is a 14-amino-acid peptide. Five somatostatin receptors—SRTR 1 to SRTR 5—are recognized, but the actual distribution of the receptors in humans is not clear. Following intravenous injection, somatostatin has a half-life in the circulation of one to three minutes; therefore, longeracting analogs of somatostatin have been synthesized. The best known of these analogs are octreotide, lanreotide, and vapreotide.146 Somatostatin decreases portal pressure and collateral blood flow by inhibiting release of glucagon.147 The optimal dose and duration of use of somatostatin have not been adequately studied. Following a single 250-µg bolus injection of somatostatin, portal and azygous blood flow decrease, but the effect lasts only a few minutes.148 Use of higher doses is associated with a more impressive decrease in HVPG. Somatostatin also decreases portal hypertension by decreasing postprandial splanchnic blood flow.149 Following a variceal bleed, blood in the gastrointestinal tract acts like a meal, leading to an increase in portal flow and elevation in the portal pressure; this elevation in pressure is ameliorated by the use of somatostatin. Following intravenous administration, octreotide has a half-life in the circulation of 80 to 120 minutes. Its effect on reducing portal pressure is not prolonged, however. Moreover, continuous infusion of octreotide does not decrease portal pressure despite decreasing the postprandial increase in portal pressure.62,150 Available evidence is insufficient to prove the superiority of somatostatin and its analogs to placebo in the control of acute variceal bleeding.134 Some randomized controlled

trials, however, support the view that somatostatin or octreotide may be equivalent in efficacy to sclerotherapy or terlipressin for controlling acute variceal bleeding. Also, early administration of vapreotide may be associated with improved control of bleeding but without a significant reduction in mortality rate.151 In clinical practice, treatment with somatostatin or octreotide is combined with endoscopic management of variceal bleeding (see later).

β-Adrenergic Blocking Agents

Nonselective β-adrenergic blocking agents have been used extensively since the landmark study of Lebrec and colleagues demonstrated the efficacy of these agents in preventing variceal rebleeding.152 Nonselective beta blockers such as propranolol or nadolol are preferred. Blockade of β1adrenergic receptors in the heart decreases cardiac output. Blockade of β2-adrenergic receptors, which cause vasodilatation in the mesenteric circulation, allows unopposed action of α1-adrenergic receptors and results in decreased portal flow. The combination of decreased cardiac output and decreased portal flow leads to a decrease in portal pressure. Nadolol has advantages over propranolol in that it is excreted predominantly by the kidney, has low lipid solubility, and is associated with a lower risk of central nervous system side effects such as depression. The effectiveness of beta blockers is assessed most accurately by monitoring the HVPG; this approach is not widely used in clinical practice. The usual method of monitoring the efficacy of beta blockers is to observe a decrease in the heart rate, which is a measure of β1-adrenergic receptor blockade. Despite adequate β1-adrenergic receptor blockade, however, some patients might benefit from a further increase in the dose of beta blocker, to increase the degree of β2-adrenergic blockade. Raising the dose, however, results in more side effects and the likelihood that treatment will need to be withdrawn.153

Nitrates

Short-acting (nitroglycerin) or long-acting (isosorbide mononitrate) nitrates result in vasodilatation. The vasodilatation results from a decrease in intracellular calcium in vascular smooth muscle cells. Nitrates cause venodilatation, rather than arterial dilatation, and decrease portal pressure predominantly by decreasing portal venous blood flow. The effect on intrahepatic resistance is less impressive than generally has been believed. Nitroglycerin has been used in combination with vasopressin to control acute variceal bleeding. The rate of infusion of nitroglycerin is 50 to 400 µg per minute, provided that the systolic blood pressure is greater than 90 mm Hg; however, the combination of vasopressin and nitroglycerin is seldom used nowadays. Nitrates are no longer recommended, either alone or in combination with a beta blocker, for primary prophylaxis to prevent first variceal bleeds. For secondary prophylaxis (to prevent variceal rebleeding), isosorbide mononitrate may be added to a beta blocker if the beta blocker alone has not resulted in an appropriate decrease in HVPG. In clinical practice, it is unusual for patients to tolerate nitrates for any length of time because of side effects, especially hypotension and headaches.

Drugs That Decrease Intrahepatic Vascular Resistance

The ideal agent for treatment of portal hypertension would be a drug that decreases intrahepatic vascular resistance. Unfortunately, such a drug is not currently available. A desirable drug would be one that selectively decreases intrahepatic vascular resistance without worsening systemic vasodilatation. Agents that may decrease intrahepatic resis-

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding tance include α1-adrenergic blockers such as prazosin,154 but long-term administration of prazosin causes worsening of the systemic hyperdynamic circulation associated with portal hypertension and consequent sodium retention and ascites.154 The addition of propranolol to prazosin may ameliorate the adverse affects of prazosin on the systemic circulation. Losartan, an angiotensin II receptor type I antagonist, causes a reduction in portal pressure without significant effects on the systemic circulation.155 In rando­ mized, controlled trials of losartan or another angiotensin II receptor antagonist, irbesartan, however, portal pressure was not reduced significantly. In fact, renal function has worsened in patients given losartan or irbesartan.156,157 ET-receptor blockers and liver-selective NO donors are promising investigational agents for therapies that target intrahepatic vascular resistance.16 Studies suggest that simvastatin may decrease intrahepatic resistance and maintain hepatic blood flow while decreasing portal pressure. This effect probably results from simvastatin-mediated NO release.158

ENDOSCOPIC THERAPY

Endoscopic therapy is the only treatment modality that is widely accepted for the prevention of variceal bleeding, control of acute variceal bleeding, and prevention of variceal rebleeding. Endoscopic variceal therapy includes variceal sclerotherapy and band ligation.

Sclerotherapy

Endoscopic sclerotherapy has largely been supplanted by endoscopic band ligation, except when poor visualization precludes effective band ligation of bleeding varices. Available evidence does not support emergency sclerotherapy as first-line treatment of variceal bleeding (Table 90-4).159 The technique involves injection of a sclerosant into (intravariceal) or adjacent to (paravariceal) a varix. Some paravariceal injection usually takes place during attempted intravariceal therapy. The sclerosants used include sodium tetradecyl sulfate, sodium morrhuate, ethanolamine oleate, and absolute alcohol; the choice of a sclerosant is based on availability, rather than on superior efficacy of one agent over another. Complications of endoscopic sclerotherapy may arise during or after the procedure. During injection, the patient may experience some degree of retrosternal discomfort, which may persist postoperatively. More serious complications include sclerosant-induced esophageal ulcer-related bleeding, strictures, and perforation. The risk of ulcers caused by sclerotherapy may be reduced by use of oral sucralfate or a proton pump inhibitor after sclerotherapy.

A

B

Variceal Ligation

Endoscopic variceal ligation is the preferred endoscopic modality for control of acute esophageal variceal bleeding and prevention of rebleeding; however, the utility of band ligation in the treatment of gastric varices is limited. Variceal ligation is simpler to perform than injection sclerotherapy. The procedure involves suctioning of the varix into the channel of an endoscope and deploying a band around the varix. The band strangulates the varix, thereby causing thrombosis. Multi-band devices can be used to apply several bands without requiring withdrawal and reinsertion of the endoscope. Varices at the gastroesophageal junction are banded initially, and then more proximal varices are banded in a spiral manner at intervals of approximately 2 cm; the endoscope is then withdrawn. Varices in the mid- or proximal esophagus do not need to be banded. Endoscopic variceal ligation is associated with fewer complications than sclerotherapy and requires fewer sessions to achieve variceal obliteration. Moreover, esophageal variceal ligation during an acute bleed is not associated with a sustained elevation in HVPG, as occurs with sclerotherapy.160 Endoscopic variceal ligation can cause local complications, including esophageal ulcers (Fig. 90-9), strictures, and dysmotility, albeit less frequently than does sclerotherapy. Banding-induced ulcers can be large and potentially serious if gastric fundal varices are banded. Now that overtubes are no longer used to facilitate repeated insertion of the endoscope during a banding session, the mechanical complications seen in the past (mucosal tears and esophageal

Table 90-4  Complications of Endoscopic Variceal Therapy* During Procedure Aspiration pneumonia Retrosternal chest pain Following Procedure Bleeding Esophageal dysmotility Esophageal stricture Local ulcers Mediastinitis Perforation Systemic (Usually with Sclerotherapy) Mesenteric venous thrombosis Pulmonary embolism Sepsis *Sclerotherapy and band ligation.

Figure 90-9.  Endoscopic views of gastric varices and esophageal variceal ligation-related ulcers. A, The gastroesophageal junction is seen on a retroflexed view following ligation of multiple gastric varices (arrowheads), which resemble polyps. B, Upper endoscopy in the same patient 4 weeks later demonstrates ulcers at the sites of earlier ligation (arrowheads).

1503

1504

Section IX  Liver perforations) are uncommon. A proton pump inhibitor is usually recommended after variceal ligation even though data to support the use of a proton pump inhibitor are limited.

TRANSJUGULAR INTRAHEPATIC PORTOSYSTEMIC SHUNT

A transjugular intrahepatic portosystemic shunt (TIPS)— also referred to as a transjugular intrahepatic portosystemic stent shunt (TIPSS)—reduces elevated portal pressure by creating a communication between the hepatic vein and an intrahepatic branch of the portal vein. A percutaneous transjugular approach is used to insert the shunt. A TIPS functions as a side-to-side portacaval shunt and has been used to treat complications of portal hypertension, mainly variceal bleeding and refractory ascites, as well as BuddChiari syndrome, hepatic hydrothorax, and hepatorenal syndrome (see Chapters 83, 91, and 92). A TIPS can be placed by an interventional radiologist, with a mortality rate of less than 1% to 2%. TIPS placement usually is carried out with the patient under sedation. A platelet count greater than 60,000/mm3 and an acceptable prothrombin time as reflected by an international normalized ratio (INR) less than 1.4 usually are recommended but are not essential in an emergency. Broad-spectrum antibiotic coverage is recommended when TIPS placement is carried out in a patient with primary sclerosing cholangitis and as an emergency procedure.

For this procedure, the hepatic vein is cannulated through a transjugular approach, and using a Rosch needle, the portal vein is cannulated. A guidewire is then passed to connect the hepatic vein and a branch of the portal vein. Following dilation of the tract, a stent is placed and dilated as required to reduce the portacaval pressure gradient (the pressure difference between the portal vein and the inferior vena cava at the confluence of the hepatic vein) to below 12 mm Hg (Fig. 90-10). Whether a lesser reduction of portacaval pressure gradient, to only 15 mm Hg or so (instead of 12 mm Hg), could be associated with reduced bleeding or a lower frequency of hepatic encephalopathy requires further study. In the past, the stents most commonly used for a TIPS were the Wallstent and the Palmaz stent. Nowadays, a coated stent is used (Viatorr, Gore, Flagstaff, Arizona). This stent has an uncoated portion that anchors the stent to the portal vein and a polytetrafluoroethylene-coated portion that lines the tract in the liver parenchyma and the draining hepatic vein. The frequency of shunt stenosis is reduced when coated stents are used instead of uncoated stents.161 A TIPS can be placed successfully by an experienced operator in greater than 95% of cases. Complications following the procedure are classified as procedure related, early (occurring before 30 days), or late (after 30 days) (Table 90-5). The prevention and treatment of pro­cedure-related, early, and late post-TIPS complications are outlined in Table 90-6.

A

B

C

D

Figure 90-10.  Creation of a transjugular intrahepatic portosystemic shunt (TIPS). A, Portogram with a catheter in the portal venous system (arrowheads). The portal venous system is clearly outlined (straight arrows). Gastroesophageal collaterals are also demonstrated (curved arrows). B, A stent (arrow) has been placed to bridge the hepatic vein and the portal vein. A balloon (arrowheads) is being used to dilate the parenchymal tract within the liver. C, Following expansion of the stent (arrow), injection into the portal vein demonstrates persistence of the gastroesophageal varices (arrowheads). D, Following embolization of the varices with steel coils (arrowheads), the intrahepatic portal vasculature is no longer demonstrated, indicative of hepatofugal flow of portal blood through the shunt.

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding Portal Hypertension-Related Bleeding

The most common indication for placement of a TIPS is refractory variceal bleeding. TIPS has been used to control acute variceal bleeding and to prevent variceal rebleeding when pharmacologic and endoscopic therapies have failed,

Table 90-5  Complications of Transjugular Intrahepatic Portosystemic Shunt Placement TIMING OF COMPLICATION

COMPLICATION

Procedure-related (life-threatening)

Cardiopulmonary failure Carotid artery puncture Intraperitoneal hemorrhage Sepsis Cardiac arrhythmias Fever Hematoma at puncture site Hemolytic anemia Hepatic encephalopathy Pain Progressive hepatic failure Pulmonary artery hypertension Shunt thrombosis Stent migration Reactions to contrast media Hepatic encephalopathy Liver failure Portal vein thrombosis Progressive hepatic failure Shunt stenosis

Early post-procedure (1-30 days)

Late post-procedure (>30 days)

Modified from Kamath PS, McKusick M. Transjugular portosystemic shunt (TIPS). Bailliere Clin Gastroenterol 1997; 11:327-49.

especially in patients with Child class B or C cirrhosis, in whom bleeding is more likely to be refractory to therapy than in patients with Child class A cirrhosis. Refractory ascites and prevention of variceal rebleeding are the only indications for TIPS that have been subjected to controlled trials. When bleeding from varices cannot be controlled after two sessions of endoscopic therapy within a 24-hour period, TIPS placement is the usual salvage treatment. TIPS also is used to treat bleeding from isolated gastric fundal varices, for both control of bleeding and prevention of rebleeding. A surgical portosystemic shunt may be preferred over a TIPS in patients with preserved synthetic liver function (Child class A) in centers that have the surgical expertise. TIPS has been effective in the management of uncontrolled esophageal variceal bleeding in patients with decompensated cirrhosis of the liver.162 Hemorrhage is controlled in more than 90% of patients, but the mortality rate in such patients is high—greater than 60% within 90 days. A similar outcome is observed in patients who undergo TIPS placement for refractory gastric variceal bleeding.163 In a meta-analysis of twelve randomized controlled trials that compared TIPS with endoscopic therapy, the rate of rebleeding was lower with TIPS, but the frequency of encephalopathy was higher, and no effect on survival was observed.164 Therefore, TIPS cannot be recommended as a first choice of treatment for preventing variceal rebleeding; rather, it is reserved for patients who have failed endoscopic or pharmacologic therapy.

Follow-up Evaluation

The frequency of stenosis of a TIPS is high, ranging from 20% to 78% depending on the surveillance technique used

Table 90-6  Prevention and Treatment of Transjugular Intrahepatic Portosystemic Shunt-Related Complications COMPLICATION

PREVENTION

TREATMENT

Inadvertent injury to carotid artery during jugular vein access

Perform with ultrasound guidance to facilitate venous access

Manual compression of carotid puncture site to prevent hematoma

Hepatic capsular laceration during portal vein access

Avoid atrophic lobes and limit needle passes to 3-4 cm of excursion

Usually requires no treatment For severe hemorrhage, transfuse with blood products until stable; obtain abdominal CT and surgical consultation

Extrahepatic puncture of portal venous system

Delineate bifurcation of portal vein on preprocedure CT

Leave catheter in place for portogram; use as a guide for intrahepatic portal vein puncture Work quickly to establish a functioning shunt, then remove the errant catheter

Intrahepatic arterial or biliary puncture

Work centrally within the liver

Usually no treatment is required; remove the catheter and continue If a fistula develops, embolize the arterial feeder with steel coils

Sepsis after shunt placement

Give prophylactic antibiotics Adhere to strict sterile technique

Broad-spectrum antibiotic coverage

Early shunt thrombosis

Avoid sharp angles when placing the stent Ends should not abut against the intima of the vein

Shunt venogram and clot lysis using tPA delivered by pulse-spray technique Extend the shunt to ensure stent coverage of the intrahepatic tract and to ensure adequate length in hepatic and portal veins

Uncontrollable encephalopathy after shunt placement

Use narrow shunts in high-risk patients

Reduce the diameter of the shunt with additional concentrically placed stents Embolize the shunt with steel coils

Shunt stenosis

Use wider or covered stents Avoid bile duct injury

Dilation or atherectomy of the shunt Place an additional stent if necessary

Post-shunt liver failure

Avoid procedure in patients with a MELD score ≥24

Consider early liver transplantation

CT, computed tomography; MELD, Model for End-stage Liver Disease; tPA, tissue plasminogen activator. Modified from Kamath PS, McKusick M. Transjugular portosystemic shunt (TIPS). Bailliere Clin Gastroenterol 1997; 11:327-49.

1505

1506

Section IX  Liver and the definition of stenosis. Neither the optimal interval nor the most cost-effective method of surveillance for TIPS stenosis has been determined. Doppler ultrasound evaluation generally is used to identify TIPS stenosis, but the negative predictive value of this approach is low and the positive predictive value is only acceptable. The best indicator that the shunt has stenosed is recurrence of the problem that necessitated the TIPS. The only certain method of demonstrating shunt patency is by means of a TIPS venogram and measurement of the portacaval pressure gradient. An increase in the gradient to greater than 12 mm Hg warrants dilation of the stent or placement of an additional stent.

may be carried out routinely in such patients when indicated. Patients with a MELD score higher than 24 have reduced survival following TIPS placement, with a mor­ tality rate approaching 30% at three months. Because these patients are at high priority for liver transplantation, TIPS should be avoided unless needed to control acute variceal bleeding. In the intermediate group with MELD scores ranging from 15 to 24, TIPS placement can be carried out depending on the patient’s preference and the physician’s judgment and taking into consideration the likelihood of liver transplantation. This approach has been validated independently.168

Selection of Patients

Surgical treatment of portal hypertension falls into three groups: non-shunt procedures, portosystemic shunt procedures, and liver transplantation. Surgical procedures are used as salvage therapy when standard management with pharmacologic and endoscopic therapy fails in patients with Child class A cirrhosis. Surgical treatment also may be considered early in the course of portal hypertension in patients who live at a great distance from centers that can manage variceal bleeding adequately or in whom crossmatching blood products (in case of bleeding) is difficult. How failure of standard therapy is defined depends on the specific circumstances of the patient’s presentation, availability of surgical expertise, and outcome of conservative management. Liver transplantation should be considered in all patients with variceal bleeding who meet minimal listing criteria for liver transplantation (currently, a CTP score of 7 or greater). Selection and prioritization of patients for liver transplantation are discussed in Chapter 95.

The presence of a TIPS may worsen liver function by depriving the liver of portal venous blood, thereby increasing the risk of hepatic encephalopathy, with decreased survival in some patients. Therefore, the procedure should be used selectively. Emergency TIPS is clearly associated with a high mortality rate.165,166 In patients in whom TIPS placement has been carried out to prevent variceal rebleeding, 30-day mortality rates may be as high as 44%. Factors associated with a poor prognosis include a serum alanine aminotransferase (ALT) level greater than 100 U/L, serum bilirubin level greater than 3 mg/dL, and pre-TIPS hepatic encephalopathy unrelated to bleeding.166 Patients with a high CTP score (Table 90-7) also have reduced survival. The Child classification has some limitations, however; for example, it does not discriminate survival well among patients within each Child class. Furthermore, some parameters that make up the CTP score, such as ascites and encephalopathy, are assessed by subjective interpretation. The need for a more accurate method to assess survival in patients undergoing TIPS has led to creation of a new tool to predict survival, the Model for End-stage Liver Disease (MELD) (see http://www.mayoclinic.org/gi-rst/ mayomodel6.html and Chapter 73).165 With data from four centers within the United States, this mathematical model originally was composed of the serum creatinine level, INR as a measure of the prothrombin time, serum bilirubin level, and etiology of liver disease. Subsequently, the MELD formula was modified to include only the first three parameters (creatinine, INR, and bilirubin).167 The MELD has been widely validated for predicting survival in patients with cirrhosis, including patients who have undergone TIPS placement, and is more accurate for this purpose than the Child classification. Patients with a MELD score of 14 or less have an excellent survival rate after TIPS placement; therefore, TIPS

SURGICAL THERAPY

Non-shunt Procedures

Non-shunt procedures include esophageal transection and gastroesophageal devascularization. They are performed infrequently but may be required in selected cases. Esophageal Transection Esophageal transection, in which the esophagus is stapled and transected, is highly effective in controlling variceal bleeding and is associated with a lower risk of encephalopathy than that for portosystemic shunts. Esophageal transection was considered in the past when two sessions of endoscopic therapy had failed to control variceal bleeding within a 24-hour period.79 Mortality rates are not improved over those observed with endoscopic sclerotherapy, however. With the advent of TIPS, esophageal staple transection is now seldom used.

Table 90-7  Child-Turcotte-Pugh Scoring System and Classification Numerical Score PARAMETER

1

2

3

Ascites Encephalopathy Bilirubin (mg/dL) Albumin (g/dL) Prothrombin time (seconds increased)

None None <2 >3.5 1-3

Slight Slight/moderate 2-3 2.8-3.5 4-6

Moderate/severe Moderate/severe >3 <2.8 >6

Total Numerical Score 5-6 7-9 10-15 *Or Child-Pugh class.

Child Class* A B C

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding Devascularization Procedures Devascularization procedures typically have been used to prevent recurrent variceal bleeding in patients with extensive splenic and portal vein thrombosis when a suitable vein is not available for creation of a portosystemic shunt.169 In the original operation described by Sugiura and Futagawa, both a thoracotomy and a laparotomy were carried out.170 Subsequently, the operation has been carried out through an abdominal approach and combined with a splenectomy. The procedure consists of total devascularization of the greater curvature of the stomach combined with devascularization of the upper two thirds of the lesser curvature of the stomach and circumferential devascularization of the lower 7.5 cm of the esophagus. The rate of recurrent bleeding following this procedure is variable but may be as high as 40%, depending on the population being treated and duration of follow-up.

Portosystemic Shunts

With the increasing availability of TIPS, the use of surgical shunts for refractory variceal bleeding has declined markedly. Surgical portosystemic shunts are categorized as selective shunts such as distal splenorenal shunts, partial shunts such as the side-to-side calibrated portacaval shunt, and total portosystemic shunts such as the side-to-side portacaval shunt or end-to-side portacaval shunt. Selective Shunts The most widely used selective shunt is the distal splenorenal shunt, originally described by Warren and colleagues.171 With this shunt, only varices at the gastroesophageal junction and spleen are decompressed, and portal hypertension is maintained in the superior mesenteric vein and portal vein; therefore, variceal bleeding is controlled, but ascites persists. The shunt procedure involves a portalazygous disconnection and subsequent anastomosis between

the splenic vein and left renal vein in an end-to-side fashion (Fig. 90-11). The entire length of the pancreas must be mobilized, and the left adrenal vein must be ligated. The distal splenorenal shunt has been associated with control of variceal bleeding in approximately 90% of patients and a lower rate of hepatic encephalopathy than that reported for total shunts.172 Partial Portosystemic Shunts A partial portosystemic shunt is carried out using a synthetic interposition graft between the portal vein and the inferior vena cava. When the shunt diameter is 8 mm, portal pressure is reduced below 12 mm Hg, and antegrade flow to the liver is maintained in most patients.173 Rates of preventing variceal rebleeding and encephalopathy following the shunt are similar to those seen with a distal splenorenal shunt. As in patients who have had a distal splenorenal shunt, ascites may occur in approximately 20% of patients who have had a partial portosystemic shunt because hepatic sinusoidal pressure is not reduced.174,175 Portacaval Shunts End-to-side and side-to-side portacaval shunts have been described, but nowadays only the side-to-side portacaval shunt is in use.176 Any portacaval shunt that is greater than 12 mm in diameter is likely to result in a total shunting of portal blood. A shunt with a diameter <12 mm is created with an interposition graft, or alternatively a direct vein-tovein anastomosis may be constructed. Variceal bleeding, as well as ascites, is well controlled because the hepatic sinusoids are decompressed. Variceal rebleeding following a total shunt is seen in less than 10% of patients, but hepatic encephalopathy occurs in 30% to 40% of patients.99 Liver transplantation in patients who have had a portacaval shunt is associated with increased operative morbidity and intraoperative transfusion requirements. The outcome of liver

Short gastric v.

Coronary v. Portal v. R. gastric v.

L. gastroepiploic v.

Superior mesenteric v. L. renal v.

Splenic v.

Figure 90-11.  Distal splenorenal shunt. Surgical anatomy following completion of a distal splenorenal shunt is depicted. For this procedure, the splenic vein is disconnected from the superior mesenteric vein and is separated from the pancreas; all its collaterals are ligated. The portal system is thus disconnected from the azygous system so that all flow from the gastroesophageal junction is through the short gastric veins into the splenic vein. The splenic vein is then anastomosed to the left renal vein in an end-to-side fashion.

1507

1508

Section IX  Liver transplantation is not otherwise significantly different, however, from that for patients who have not had a portacaval shunt. Nonetheless, surgical portacaval shunts should be avoided in patients who are potential candidates for liver transplantation.

MANAGEMENT OF SPECIFIC LESIONS ESOPHAGEAL VARICES Natural History

Esophageal varices are present in approximately 40% of patients with cirrhosis and in as many as 60% of patients with cirrhosis and ascites.99 In cirrhotic patients who do not have esophageal varices at initial endoscopy, new varices will develop at a rate of approximately 5% per year. In patients with small varices at initial endoscopy, progression to large varices occurs at a rate of 10% to 15% per year and is related predominantly to the degree of liver dysfunction.177 On the other hand, improvement in liver function in patients with alcoholic liver disease who abstain from alcohol is associated with a decreased risk, and sometimes even disappearance, of varices.178 Up to 25% of patients with newly diagnosed varices will bleed within two years.177 The best clinical predictor of bleeding appears to be variceal size. The risk of bleeding in patients with varices less than 5 mm in diameter is 7% by two years, and the risk in patients with varices greater than 5 mm in diameter is 30% by two years.177 Even more important, however, is the HVPG because the risk of bleeding is virtually absent when the HVPG is below 12 mm Hg.87 Nevertheless, measurement of HVPG is not routinely performed in clinical practice to assess bleeding risk. The prognosis for variceal bleeding in patients with cirrhosis has improved since the 1980s. Initial treatment is associated with cessation of bleeding in approximately 90% of patients.177,179 Approximately one half of patients with a variceal bleed stop bleeding spontaneously because hypovolemia leads to splanchnic vasoconstriction, which results in a decrease in portal pressure. Excessive transfusions may, in fact, increase the chance of rebleeding.180 Active bleeding at endoscopy, a lower initial hematocrit value, higher serum aminotransferase levels, higher Child class, bacterial infection, an HVPG greater than 20 mm Hg, and portal vein thrombosis are associated with failure to control bleeding at five days.179,181-183 Of patients who have stopped bleeding, approximately one third will rebleed within the next six weeks. Of all rebleeding episodes, approximately 40% will take place within five days of the initial bleed.184 Predictors of rebleeding include active bleeding at emergency endoscopy, bleeding from gastric varices, hypoalbuminemia, renal insufficiency, and an HVPG greater than 20 mm Hg.177 The risk of death with acute variceal bleeding is 5% to 8% at one week and about 20% at six weeks.177 Patients who rebleed early, have a MELD score >18, require >4 units of packed red blood cell transfusions,185 and in whom renal failure develops have the highest risk of death. Alcohol as the cause of cirrhosis, a higher serum bilirubin level, a lower serum albumin level, hepatic encephalopathy, and hepatocellular carcinoma are additional factors associated with an increased six-week mortality rate. Treatment of esophageal variceal bleeding is classified as either primary prophylaxis, that is, prevention of variceal hemorrhage in patients who have never bled; control of acute variceal bleeding; or secondary prevention of rebleeding in patients who have survived an initial bleeding episode. Effective treatments to prevent the development of

varices and ascites in patients with cirrhosis are not yet available, although beta blockers may slow enlargement of small varices into large varices.

Prevention of Bleeding

The utility of pre-primary prophylaxis, that is, the efficacy of beta blockers to prevent the formation of varices, has not been demonstrated.79,186 Patients with Class C cirrhosis who have small varices may be considered for treatment with a beta blocker. All patients with large varices (diameter greater than 5 mm) should be considered for prophylactic therapy (“primary prophylaxis”) to prevent variceal bleeding. The presence of additional endoscopic signs such as red wales does not influence the decision regarding prophylactic therapy. Twelve trials have addressed the use of a nonselective beta blocker for primary prophylaxis of variceal bleeding and have demonstrated a decrease in the risk of variceal bleeding from 25% in patients in the control group to 15% in patients taking a beta blocker. The absolute risk reduction is thus approximately 10%, and the number needed to treat to prevent one variceal bleed is approximately 10 patients. The mortality rate is reduced from 28.4% in control patients to 23.9% in patients taking a beta blocker; the absolute risk reduction is 4.5%. The number of patients needed to be treated to prevent one death is approximately 22. In patients who do not bleed during therapy and who do not experience side effects, treatment should be continued indefinitely because withdrawal of a beta blocker can result in an increased risk of bleeding.187,188 The side effects of beta-blocker treatment are probably overemphasized because only approximately 15% of patients need to discontinue the drug.189 A baseline heart rate and blood pressure recording will help determine whether a patient is a candidate for pharmacologic treatment with a beta blocker. A resting heart rate of less than 55 to 60 beats per minute or a systolic blood pressure less than 90 mm Hg indicates that the patient is likely to be intolerant of beta blockers. In other patients, the HVPG ideally should be measured at baseline (Fig. 90-12). A longacting preparation of propranolol or nadolol may be started; the usual starting dose of long-acting propranolol is 60 mg once daily and that of nadolol is 20 mg once daily. Because the risk of bleeding is greatest at night, the beta blocker should probably be administered in the evening.108 The dose of propranolol or nadolol can be increased gradually every three to five days until the target heart rate of 25% below baseline or 55 to 60 beats per minute or the maximum tolerated dose is reached, provided that the systolic blood pressure remains above 90 mm Hg. The daily dose of long-acting propranolol or nadolol required to reach the target heart rate ranges from 40 to 160 mg. Patients with a decrease in systolic blood pressure below 90 mm Hg are most likely to experience side effects. In patients on pharmacologic therapy, follow-up endoscopy is unnecessary unless gastrointestinal bleeding occurs. When the target heart rate is reached, however, a repeat HVPG measurement may be carried out as close to one month as possible. In patients in whom the HVPG has decreased to less than 12 mm Hg, the risk of bleeding is virtually eliminated. Patients in whom the HVPG decreases by at least 20% have a risk of variceal bleeding of less than 10%. Unfortunately, only 30% to 40% of patients respond to a beta blocker; those with better liver function show the best response.189 In patients who are intolerant of or who have contraindications to beta blockers, isosorbide mononitrate has been tried but is no better than placebo in preventing variceal bleeding.190 In these patients, endoscopic prophylaxis should be pursued. Unfortunately,

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding ligation to prevent primary prophylaxis is not currently recommended.

Upper endoscopy

No varices

Large esophageal varices

Small esophageal varices

Repeat endoscopy in 2–3 years

Consider HVPG measurement

Repeat endoscopy in 1–2 years

Nonselective β-adrenergic blockers

Consider HVPG remeasurement

Contraindication to β-adrenergic blockers Intolerance to β-adrenergic blockers Target HVPG not reached Patient preference

Perform EVL Figure 90-12.  Algorithm for primary prophylaxis of esophageal variceal hemorrhage. The hepatic venous pressure gradient (HVPG) may be measured before a nonselective b-adrenergic blocking agent is started and one month after the maximum tolerated dose of the beta blocker is reached. The goal of treatment is to reduce the HVPG to <12 mm Hg or by ≥20%. EVL, endoscopic variceal ligation; HVPG, hepatic venous pressure gradient.

patients who do not achieve a decrease in the HVPG to less than 12 mm Hg, or of greater than 20%, on a beta blocker may not respond well to endoscopic variceal ligation either.191 Endoscopic Prevention Prophylactic sclerotherapy for the prevention of variceal bleeding has been studied extensively but cannot be recommended.192 The preferred method of endoscopic treatment is variceal band ligation. Variceal band ligation has been compared with no treatment in five trials. A meta-analysis of these studies demonstrated that, compared with no treatment, endoscopic variceal ligation decreases the risk of first bleeding and the mortality rate.193 Meta-analysis of the nine trials that compared endoscopic variceal ligation with a beta blocker demonstrated a lower bleeding risk with endoscopic variceal ligation, with no difference in mortality rates.194 A subsequent study has suggested that nonbleedingrelated mortality may actually be reduced by beta blockers.195 Side effects with beta blockers are more frequent than with variceal ligation, but complications of variceal ligation can be potentially life threatening. Therefore, current evidence does not support endoscopic variceal ligation as the preferred method of primary prophylaxis against variceal bleeding. The risks and benefits of the alternative therapies should be discussed and treatment individualized. Beta blockers are cheaper and more convenient to use and may potentially reduce the risk of bleeding from gastric varices and portal hypertensive gastropathy. Band ligation is the only option for patients with high-risk varices who have contraindications to beta blockers or who have not responded to or are intolerant of beta blockers. Combined use of propranolol and endoscopic variceal

Control of Acute Bleeding

Acute esophageal variceal bleeding constitutes a lifethreatening emergency and requires management by a welltrained team of hepatologists, endoscopists, intensive care personnel, radiologists, and surgeons. Treatment is aimed at resuscitating the patient, controlling the bleeding, and preventing complications (see Chapter 19). Two large-bore intravenous access lines should be inserted immediately. Red blood cells should be transfused with the goal of maintaining the hematocrit value around 25%. Normal saline may be infused intravenously until packed red blood cells are available for transfusion. In patients with active bleeding, the airway needs to be protected, and endotracheal intubation is advised. Antibiotics should be administered to all patients to prevent bacteremia and spontaneous bacterial peritonitis. Norfloxacin, 400 mg orally twice daily for seven days, is the preferred choice.196 When oral intake is not possible, intravenous ciprofloxacin, 400 mg every 12 hours; levofloxacin, 500 mg every 24 hours; or ceftriaxone, 1 g every 24 hours for 7 days is recommended. The addition of recombinant factor VIIa to standard therapy has not been shown to improve control of bleeding.197 A combination of endoscopic therapy and pharmacologic therapy of variceal bleeding may be superior to pharmacologic treatment alone. Pharmacologic agents should be started as early as possible; in some centers, they are started while the patient is being transferred by ambulance to the hospital. Somatostatin, octreotide, terlipressin, and vasopressin plus nitroglycerin are the options for pharmacologic therapy. The specific agent chosen depends on availability and physician preference. In the United States, octreotide is the agent used most commonly. Terlipressin is the first choice in many other countries because it is the only drug that has been associated with improved survival.198 Pharmacologic treatment should be continued for up to five days to prevent early rebleeding. Endoscopic therapy is carried out as soon as the patient is hemodynamically stabilized. At upper endoscopy, bleeding from esophageal varices is diagnosed if active bleeding from the varices is seen; signs of recent hemorrhage, such as a white fibrin plug or a red blood clot over a varix, are present; varices with risk signs for bleeding, such as a cherry-red spot, hematocystic spot, or red wale sign, are seen; or esophageal varices are seen in the absence of any other lesion that could give rise to gastrointestinal bleeding. Endoscopic treatment is recommended at the time of initial endoscopy, and endoscopic variceal ligation is the preferred method. At upper endoscopy, the actively bleeding varix is ligated (Fig. 90-13). Ligation initially should be at or immediately below the bleeding site. Other large varices also should be banded during the same session. If active bleeding is not seen, ligation should be carried out beginning with varices at the gastroesophageal junction and proceeding proximally at intervals of 2 cm in a spiral fashion. If bleeding obscures the varices, then multiple bands are placed at the gastroesophageal junction circumferentially until bleeding can be controlled, but the long-term risks of esophageal stricture are increased in such patients. Bleeding can be controlled in up to 90% of patients with a combination of pharmacologic and endoscopic treatment. Bleeding cannot be controlled in approximately 10% of patients, as defined by any of the following three factors: (1) transfusion of four units of red blood cells or more to maintain the hematocrit value between 25% and 30%; (2) inabil-

1509

1510

Section IX  Liver

A

Figure 90-13.  Band ligation for control of esophageal variceal bleeding. A, On upper endoscopy, an actively bleeding varix can be seen in the distal esophagus (arrow). B, With the variceal banding device in position, the varix is suctioned into the device at the site of active bleeding (arrow). C, After the band is in place (arrow) and the varix has been ligated, the bleeding has stopped. D, Visualization of the varix with the band in place with complete control of bleeding. (Images courtesy of Dr. Louis M. Wong Kee Song, Rochester, Minn.)

C

ity to increase the systolic blood pressure by 20 mm Hg or to greater than 70 mm Hg; or (3) persistence of a heart rate greater than 100 beats per minute.199 Rebleeding is defined as recurrence of bleeding after initial control for 24 hours during which the vital signs and hemoglobin level are stable. When two sessions of endoscopic treatment within a 24-hour period have failed to control variceal bleeding, salvage therapies such as TIPS should be carried out (Fig. 90-14), although the mortality rate in this group of patients is high. Surgical portosystemic shunts, although extremely effective in controlling variceal bleeding, have largely been abandoned because of high mortality rates. Balloon tamponade sometimes is used to stabilize the patient until definitive treatment can be carried out.

Prevention of Rebleeding

B

All patients who have had a variceal bleed should receive prophylactic therapy (“secondary prophylaxis”) to reduce the risk of rebleeding, which otherwise occurs in up to 80% of patients at two years. Patients with a CTP score of 7 or greater also should be evaluated for liver transplantation. Options for preventing variceal rebleeding are pharmacologic therapy, endoscopic therapy, and portosystemic shunts (surgical or radiologic), or combinations of these therapies. Combined therapy with endoscopic variceal ligation and a nonselective beta blocker is the preferred treatment, and long-acting propranolol or nadolol may be used. Ideally, the hemodynamic response to a beta blocker should be monitored with the goal of reducing the HVPG by greater than 20% or to less than 12 mm Hg. If these goals are not achieved, isosorbide mononitrate may be added. The extended-release form of isosorbide mononitrate is preferred, with an initial starting dose of 30 mg/day. Unfortu-

D

nately, it is unusual for patients in our practice to tolerate nitrates after they have achieved beta blockade. Hypotension and headaches are the usual reasons for discontinuing isosorbide mononitrate. Endoscopic variceal ligation alone may be performed to prevent variceal rebleeding in patients who have poor liver function (Fig. 90-15). In practice, the first endoscopic session is carried out 7 to 14 days after the initial variceal ligation to control bleeding. Endoscopic therapy is then repeated at three- to four-week intervals; this approach has been suggested because bands might still be in place if endoscopy is repeated earlier. If the HVPG is monitored, a reduction in HVPG to less than 12 mm Hg or by greater than 20% should obviate the need for variceal ligation. For patients who bleed during pharmacologic treatment, variceal ligation should be carried out. Conversely, for patients who have undergone variceal ligation alone and experience recurrent bleeding, a beta blocker should be started. Patients who have a variceal rebleed despite receiving optimal pharmacologic and endoscopic treatment require a portosystemic shunt. Even in patients with Child class A cirrhosis, a TIPS may be as effective as a distal splenorenal shunt, and the choice of therapy should depend on local expertise.200

GASTRIC VARICES

The most widely used classification of gastric varices is the Sarin classification.201 According to this classification, type 1 gastroesophageal varices (GOV1) extend 2 to 5 cm below the gastroesophageal junction and are in continuity with esophageal varices; type 2 gastroesophageal varices (GOV2) are in the cardia and fundus of the stomach and in continuity with esophageal varices; varices that occur in the fundus of the stomach in the absence of esophageal varices are

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding called isolated gastric varices type 1 (IGV1), whereas varices that occur in the gastric body, antrum, or pylorus are called isolated gastric varices type 2 (IGV2). Approximately 25% of patients with portal hypertension have gastric varices, most commonly GOV1, which comprise approximately 70% of all gastric varices. Intrahepatic causes of portal hypertension may be associated with both GOV1 and GOV2. Splenic vein thrombosis usually results in IGV1, but the most common cause of fundal gastric varices may be cirrhosis.

Acute bleeding from esophageal varices Resuscitation Vasoactive agent (e.g., octreotide) Antibiotic (e.g., norfloxacin) Upper endoscopy Variceal ligation (or sclerotherapy)

Natural History

Bleeding controlled No

Yes

Repeat endoscopic therapy

Secondary prophylaxis (see Fig. 90-15)

Bleeding controlled

No Balloon tamponade

Yes

Secondary prophylaxis (see Fig. 90-15)

Child class A

Child class B/C

Surgical portosystemic shunt or TIPS

TIPS

Figure 90-14.  Algorithm for the management of bleeding esophageal varices. TIPS, transjugular intrahepatic portosystemic shunt.

Gastric varices typically occur in association with more advanced portal hypertension. Bleeding is thought to be more common in patients with GOV2 and IGV1 than in those with other types of gastric varices; in other words, bleeding is more common from fundal varices than from varices at the gastroesophageal junction. Whereas intra­ esophageal pressure is negative, intra-abdominal pressure is positive, and the transmural pressure gradient across gastric varices is smaller than that across esophageal varices. Gastric varices, however, tend to be larger in diameter than esophageal varices. Gastric varices are supported by gastric mucosa, whereas esophageal varices tend to be unsupported in the lower third of the esophagus. Therefore, gastric varices are likely to bleed only when they are large, as demonstrated in a study in which larger gastric varices (greater than 5 to 10 mm in diameter) were more likely to bleed than smaller ones.202 Although gastric varices have been thought to bleed less frequently than esophageal varices, the bleeding rates probably are comparable if patients are matched for the severity of cirrhosis (CTP score).201 In contrast with esophageal varices, bleeding from gastric varices has been described with an HVPG less than 12 mm Hg.203,204 Gastric varices in continuity with esophageal varices may regress following treatment of the esophageal varices. When gastric varices persist despite obliteration of esophageal varices, the prognosis is poorer, probably because of the severity of liver disease.

Prevention of Bleeding Child class A, B

Child class B, C

Consider baseline HVPG measurement

EVL

β-adrenergic blocker

Recurrent variceal bleeding

Unfortunately, no studies have evaluated pharmacologic or endoscopic treatment for primary prophylaxis of gastric variceal hemorrhage, and recommendations are based primarily on the guidelines for managing esophageal varices. Large gastric varices with red color signs, especially in patients with Child class C liver disease, are most likely to bleed. Because these gastric varices usually are associated with esophageal varices, pharmacologic treatment with a nonselective beta blocker may be initiated to prevent variceal hemorrhage. Endoscopic treatment or TIPS is currently not recommended for the primary prevention of gastric variceal bleeding. Balloon-occluded retrograde transhepatic occlusion of varices has been used in uncontrolled studies to prevent bleeding from gastric varices, with some success.

Consider HVPG in 1 month HVPG ↓ to <12 mm Hg or HVPG ↓ ≥20% Yes Continue therapy

No Perform EVL

Control of Acute Bleeding Yes Add βadrenergic blocker

No Repeat EVL until varices are obliterated

Figure 90-15.  Algorithm for the prevention of recurrent variceal bleeding (secondary prophylaxis). EVL, endoscopic variceal ligation; HVPG, hepatic venous pressure gradient.

The approach to treating esophageal variceal hemorrhage also applies to acute gastric variceal hemorrhage and includes volume resuscitation, avoidance of overtransfusion, and antibiotic prophylaxis with norfloxacin, 400 mg twice daily for seven days. Upper endoscopy is carried out after patients have been volume resuscitated and stabilized and often following endotracheal intubation to protect the airway. The endoscopic diagnosis of gastric variceal bleeding may be difficult because of pooling of blood in the fundus. A diagnosis of gastric variceal hemorrhage should

1511

1512

Section IX  Liver Acute bleeding from gastric varices

Resuscitation Vasoactive agent (e.g., octreotide) Antibiotic (e.g., norfloxacin)

Endoscopic therapy possible?

Yes

A

B

Bleeding controlled

Figure 90-16.  Gastric variceal bleeding. A, Active bleeding from a gastric varix (arrowhead) can be seen. B, Bleeding from the varix (straight arrow) is controlled following injection of sodium tetradecyl sulfate. Pooling of blood in the stomach is indicated by the curved arrow.

No

No TIPS

Yes Cyanoacrylate available? No

be considered if bleeding is noted from a gastric varix (Fig. 90-16); blood is found to appear at the gastroesophageal junction or the gastric fundus; blood is found in the stomach and gastric varices with a “white nipple sign” (indicating a fibrin-platelet plug) are seen in the absence of other causes of bleeding; and gastric varices are noted in the absence of other lesions in the esophagus and stomach.205 Because controlled studies evaluating pharmacologic therapy for gastric variceal bleeding are lacking, the agents used are based on extension of the data relating to esophageal varices. Medical management with vasoactive agents should be started as early as possible, preferably at least 30 minutes before endoscopic therapy is carried out. The preferred endoscopic therapy for fundal gastric variceal bleeding is injection of polymers of cyanoacrylate, usually N-butyl-2-cyanoacrylate,206,207 but these tissue adhesives are not currently available in the United States. Obliteration of the varices occurs when the injected cyanoacrylate adhesive hardens on contact with blood. The mucosa overlying the varix eventually sloughs, and the hardened polymer is extruded. Fortunately, the resulting ulcers occur late, and the risk of bleeding is lower than that associated with sclerotherapy-related ulcers. Cyanoacrylate injection has been found to be superior to both variceal band ligation and sclerotherapy using alcohol.207 Complications of cyanoacrylate injection include bacteremia and variceal ulceration. Pulmonary and cerebral emboli have been reported on occasion, usually in patients with spontaneous large portosystemic or intrapulmonary shunts. The endoscope may be damaged by the glue, but the risk is minimized if silicone gel is used and suction is avoided for 15 to 20 seconds following injection.208 For injection of GOV2 or IGV1, a retroflexed endoscopic approach is recommended. Sclerosants such as sodium tetradecyl sulfate, ethanolamine oleate, and sodium morrhuate are not particularly effective for control of gastric variceal bleeding.209 When sclerotherapy is carried out for gastric varices, the volume of sclerosant required is larger than that used for esophageal varices, and fever and retrosternal pain are more common. It is much easier to obliterate GOV1 than GOV2 or IGV1. IGV1 are the most difficult gastric varices to obliterate and, when present, should prompt early consideration of definitive treatment such as portosystemic shunting if cyanoacrylate is not available. Although some investigators recommend ligation of gastric varices up to 20 mm in diameter,210 this recommendation is not supported by our experience. Band ligation of varices greater than 10 mm in diameter usually is unsafe.

Yes Consider obliteration of varices

Child class A

Child class B or C

Surgical portosystemic shunt

TIPS

Figure 90-17.  Algorithm for the management of bleeding gastric varices in patients with portal hypertension. TIPS, transjugular intrahepatic portosystemic shunt.

Ligation is safest if the varices are in the cardia of the stomach. Because gastric fundal varices are covered by mucosa, drawing the entire varix into the ligation device is often not possible. Application of bands results in creation of a large ulcer on the varix, sometimes with disastrous results (see Fig. 90-9). If endoscopic and pharmacologic therapies fail to control gastric variceal bleeding, then a Linton-Nachlas tube, which has a 600-mL balloon, may be passed as a temporizing measure. The commonly used Minnesota tube and Sengstaken-Blakemore tube, with only 250-mL gastric balloons, are not as effective as the Linton-Nachlas tube for controlling bleeding from gastric fundal varices.211,212 Nevertheless, most patients in whom endoscopic and pharmacologic treatment fails to control gastric variceal bleeding will require a TIPS, which can control bleeding in greater than 90% of patients—a rate of efficacy equivalent to that for TIPS in controlling esophageal variceal bleeding (Fig. 90-17).163,213

Prevention of Rebleeding

Endoscopic Therapy In the absence of large, well-conducted trials, recommendations regarding pharmacologic or endoscopic therapy to prevent gastric variceal rebleeding are derived from retrospective studies. Cyanoacrylate glue has been used to prevent gastric variceal rebleeding, with favorable results.214 In a small study, the 2-octyl-cyanoacrylate polymer (Dermabond) has been used to prevent gastric variceal rebleeding, with excellent results.215 Patients require an average of two or three sessions for obturation of gastric varices with cyanoacrylate polymers. Detachable snares have been used to ligate gastric varices that may be too large for band ligation, but the experience with this procedure is limited.216

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding Transvenous Obliteration of Gastric Varices Balloon-occluded retrograde transvenous obliteration of gastric fundal varices is carried out in patients with demonstrable splenorenal shunts. Such shunts can be demonstrated on multidetector CT in a large number of patients with bleeding gastric varices. In these patients, a catheter is passed into the left renal vein, usually through a transfemoral approach, and into the varices, which drain into the renal vein. Following balloon occlusion of the varices, a sclerosant is injected retrograde into the varix under fluoroscopic guidance. An extension of this technique is to occlude the varix through a transfemoral approach, while, at the same time, the other end of the varix is approached through a percutaneous transhepatic portal venous route. This technique allows occlusion of the varices at both ends, after which sclerosant can be injected.217 The use of these techniques, which require considerable radiologic skill, is currently limited to some centers in the Far East.

Portosystemic Shunts

Limited data are available regarding use of surgical portosystemic shunts for the treatment of gastric varices in patients with cirrhosis. Two studies performed in patients with good liver function, most of whom had extrahepatic portal vein thrombosis, demonstrated excellent results, with a low long-term risk of bleeding and encephalopathy, after creation of a surgical shunt.218,219 TIPS also is effective in preventing gastric variceal rebleeding. Because TIPS for this indication does not always result in a decrease in the size of gastric varices,220 the target HVPG is uncertain in these patients.204 Patients with an HVPG less than 12 mm Hg after TIPS are protected from esophageal variceal bleeding but have been known to bleed from gastric varices. Therefore, if the HVPG is reduced to a level below 12 mm Hg but gastric fundal varices are still prominent when contrast is injected into the portal vein (especially if the patient has bled from gastric fundal varices), the gastric varices should be embolized.

ECTOPIC VARICES

Varices that occur at a site other than the gastroesophageal junction are termed ectopic varices and account for less than 5% of all varix-related bleeding episodes. Ectopic varices most commonly manifest with melena or hemate­ mesis. They also may manifest with hemobilia, hematuria, hemoperitoneum, or retroperitoneal bleeding. Ectopic varices occur with both extrahepatic portal vein obstruction and cirrhosis. The duodenum is a common site of ectopic varices, and varices typically are associated with portal vein obstruction, but in the West, the usual cause of duodenal varices is cirrhosis. The common occurrence of duodenal varices in patients with portal vein obstruction probably relates to the formation of collateral vessels around the thrombosed portal vein, which connect pancreaticoduodenal veins to retroduodenal veins, which drain into the inferior vena cava.221 In some of those patients with extrahepatic portal vein obstruction, varices form around the gallbladder and bile duct. The other common site of ectopic varices is peristomal, in patients with inflammatory bowel disease and primary sclerosing cholangitis who have undergone a proctocolectomy with creation of an ileostomy.222 Varices develop at the level of the mucocutaneous border of the stoma and are termed stomal varices. They are recognized by a bluish halo surrounding the stoma and by a dusky appearance and friable consistency of the stomal tissue; no obvious variceal lesions are seen.

Figure 90-18.  Endoscopic image of a colonic varix (arrow).

Anorectal varices are reported in 10% to 40% of cirrhotic patients who undergo colonoscopy and must be distinguished from hemorrhoids (Fig. 90-18). Rectal varices are dilated superior and middle hemorrhoidal veins, whereas hemorrhoids are dilated vascular channels above the dentate line. Rectal varices collapse with digital pressure, but hemorrhoids do not. Bleeding from stomal varices is readily apparent on presentation. Ectopic variceal bleeding should be considered in all patients with portal hypertension and overt gastrointestinal bleeding without an obvious bleeding source on endoscopy or a drop in the hemoglobin level associated with abdominal pain and increasing abdominal girth. CT of the abdomen demonstrates layering of free fluid in the peritoneal cavity in patients who have intra-abdominal hemorrhage, typical of fresh blood mixed with ascitic fluid. The diagnosis of intra-abdominal hemorrhage secondary to ectopic variceal bleeding is confirmed by a paracentesis that yields bloody ascitic fluid with clots.

Management

In patients suspected of having ectopic variceal bleeding, vasoactive drugs may be administered initially to control the bleeding. If the bleeding ectopic varix is visualized at endoscopy, as typically is the case with duodenal or colonic varices, then endoscopic therapy can be carried out.223 Endoscopic glue injection or band ligation is the preferred approach for bleeding duodenal varices. Colonic varices tend to be larger in diameter and may require application of hemostatic clips. Patients with bleeding stomal varices can be trained to compress the site locally if bleeding is obvious. Because bleeding from stomal varices is visible and detected early, the mortality rate for bleeding stomal varices is low.224 At present, no recommendations support primary prophylaxis to prevent bleeding from ectopic varices. To prevent rebleeding from ectopic varices, pharmacologic treatment with a beta blocker is usually tried, although no studies are available to support this approach. If the portal vein is patent, then transhepatic embolization of stomal varices can be carried out (Fig. 90-19). Embolization of varices using a transhepatic approach can control bleeding in most patients with stomal varices. The rate of rebleeding is high, however, because the portal hypertension persists. In patients in whom embolization fails to prevent rebleed-

1513

1514

Section IX  Liver ing, TIPS placement may be considered. A surgical portosystemic shunt is recommended in patients with Child class A cirrhosis and in patients with portal hypertension from extrahepatic portal vein thrombosis in whom a vein suitable for a shunt is available. Placement of a nonselective portosystemic shunt, such as a portacaval shunt, mesocaval Acute bleeding from ectopic varices

PORTAL HYPERTENSIVE GASTROPATHY AND GASTRIC VASCULAR ECTASIA

Resuscitation + vasoactive agent (e.g., octreotide) + antibiotic (e.g., norfloxacin)

Endoscopic therapy possible?

Yes

No

Bleeding successfully controlled Yes

No

Start β-adrenergic blocker

Patent portal vein ?

Yes

Transhepatic embolization of varices

Yes

Bleeding controlled

Start βadrenergic blocker

Child class A

No

Surgical ligation of varices

Start β-adrenergic blocker

No

Child class B or C

Surgical nonselective portosystemic shunt

TIPS

Figure 90-19.  Algorithm for the management of bleeding from ectopic varices in patients with portal hypertension. TIPS, transjugular intrahepatic portosystemic shunt.

A

shunt, or central splenorenal shunt, should be carried out in a patient with stomal varices. Patients with ectopic varices who present with intraperitoneal hemorrhage have a poor outcome because the diagnosis usually is not considered and often is made at laparotomy. Acute bleeding may be controlled by trans­ hepatic obliteration or surgical ligation of the varices. In patients who are critically ill, a TIPS should be placed, followed by embolization of the bleeding varix.

Mucosal changes in the stomach in patients with portal hypertension include portal hypertensive gastropathy (PHG) and gastric vascular ectasia. In all likelihood, these lesions are distinct, as demonstrated by histologic features and differences in the response to a TIPS. An appearance analogous to PHG in the colon is termed portal hypertensive colopathy (see Chapter 36). The diagnosis of PHG is based on the presence of a characteristic mosaic-like pattern of the gastric mucosa on endoscopic examination. This pattern is characterized by small polygonal areas with a depressed border. Superimposed on this mosaic-like pattern may be red point lesions that usually are greater than 2 mm in diameter. PHG is considered mild when only a mosaic-like pattern is present and severe when superimposed discrete red spots are also seen (Fig. 90-20).225 The cause and pathogenesis of PHG are poorly understood. Development of PHG correlates with the duration of cirrhosis but not necessarily the degree of liver dysfunction. The frequency of PHG following endoscopic treatment of esophageal varices is increased—possibly a result of longer duration of portal hypertension in these patients. In gastric vascular ectasia, aggregates of ectatic vessels can be seen on endoscopic examination as red spots without a mosaic background.226 When the aggregates are confined to the antrum of the stomach, the term gastric antral vascular ectasia (GAVE) is used (see Chapters 19 and 36). If aggregates in the antrum are linear, the term watermelon stomach is used to describe the lesion (Fig. 90-21). When the red spots are distributed diffusely, in both the distal and the proximal stomach, the term diffuse gastric vascular ectasia is preferred.227

B

Figure 90-20.  Endoscopic views of portal hypertensive gastropathy. A, Mild portal hypertensive gastropathy is characterized by a mosaic appearance without red color signs. B, Severe portal hypertensive gastropathy is characterized by a cobblestone appearance with superimposed red spots.

Chapter 90  Portal Hypertension and Gastrointestinal Bleeding Chronic bleeding

Iron replacement

β-adrenergic blocker

Bleeding controlled?

No

No

Figure 90-21.  Endoscopic images of watermelon stomach.

Table 90-8  Comparison of Portal Hypertensive Gastropathy (PHG) and Gastric Antral Vascular Ectasia (GAVE) FEATURE

PHG

GAVE

Distribution Mosaic pattern Red color signs Findings on gastric mucosal biopsy   Thrombi   Spindle cell proliferation   Fibrohyalinosis Treatment

Proximal stomach Present Present

Distal stomach Absent Present

− +

+++ ++

− β-adrenergic blockers TIPS

+++ ?Antrectomy

PHG accounts for approximately one fourth of all cases of bleeding (acute and chronic) in patients with portal hypertension, but for less than 10% of all acute bleeding episodes. The more common presentation is one of chronic, slow bleeding and anemia. Pharmacologic therapy to prevent bleeding (primary prophylaxis) in patients with severe portal hypertensive gastropathy is not currently recommended. Small studies have suggested that octreotide may be useful for controlling acute bleeding.229 Beta blockers are recommended for preventing chronic blood loss in patients who have bled from severe PHG.230,231 When patients are transfusion-dependent despite beta blockade and iron supplementation, a TIPS may be inserted (Fig. 90-22). A TIPS decreases transfusion requirements and results in reversal of the mucosal lesions on endoscopic examination.227

TIPS

Figure 90-22.  Algorithm for the management of chronic bleeding from portal hypertensive gastropathy. TIPS, transjugular intrahepatic portosystemic shunt.

Chronic bleeding Iron replacement Hemoglobin stable

No

?Liver transplantation

Distinguishing PHG from GAVE is sometimes difficult. A background mosaic pattern and proximal distribution favor PHG (Table 90-8). GAVE is less common, occurs in the absence of a background mosaic pattern, and typically is antral in location, although lesions may be present in the proximal stomach. Mucosal biopsies are recommended when the endoscopic diagnosis is uncertain. GAVE appears histologically as dilated mucosal capillaries with focal areas of fibrin thrombi or ectasia in combination with proliferation of spindle cells.228 Similar ectatic lesions may be seen in the small bowel and may cause acute or chronic gastrointestinal blood loss.

Continue β-adrenergic blocker

Yes

Continue iron, β-adrenergic blocker; transfusions as required

TIPS, transjugular intrahepatic portosystemic shunt.

Management

Patient transfusion dependent?

Yes

GAVE localized Platelets >45,000/mm3 INR <1.4

Continue iron replacement

Yes

No

Yes

Thermoablative therapy

Estrogen/progesterone No

No

No

Bleeding controlled

Is patient a liver transplant candidate?

? Antrectomy

Yes Continue therapy

Yes

Bleeding controlled Yes Continue iron, red cell transfusions; thermoablation as required

Consider liver transplantation

Figure 90-23.  Algorithm for the management of chronic bleeding from gastric vascular antral ectasia (GAVE). INR, international normalized ratio.

Management of GAVE is more problematic. Initial treatment involves repletion of iron and red blood cell transfusions to treat symptomatic anemia. If lesions are localized, the platelet count is greater than approximately 45,000/ mm3, and the INR is less than 1.4, thermoablative therapy, as with argon plasma coagulation, may be helpful (Fig. 90-23). The usual settings for argon plasma coagulation

1515

1516

Section IX  Liver are an energy level of 60 to 90 watts and a gas flow rate of 1 to 2 L per minute. If the coagulation parameters are suboptimal, thermal coagulation is associated with an increase in mucosal bleeding in many patients. When the vascular ectasias are diffuse and extensive in the stomach, cryotherapy using liquid nitrogen or CO2 may be used.232 If endoscopic therapy fails, therapy with an oral estrogenprogesterone combination may be useful in reducing transfusion requirements.233 The usual dose is estradiol, 35 µg, plus norethindrone, 1 mg, daily. Because the medication is taken daily, no risk of breakthrough vaginal bleeding exists. Rarely, painful gynecomastia may limit use of this combination in men. In the patient with preserved hepatic synthetic function who continues to bleed despite thermoablative therapy and an estrogen-progesterone combination, a surgical antral resection may be carried out. TIPS does not reduce the bleeding risk in patients with GAVE and is associated with a substantial risk of hepatic encephalopathy.227 Therefore, TIPS placement is not recommended as therapy for GAVE. Nevertheless, GAVE is reversed with liver transplantation, even in the presence of portal hypertension, suggesting that GAVE is related to liver failure, rather than to portal hypertension.50,234

KEY REFERENCES

Bernard B, Grange JD, Khac EN, et al. Antibiotic prophylaxis for the prevention of bacterial infections in cirrhotic patients with gastrointestinal bleeding: A meta-analysis. Hepatology 1999; 29:1655-61. (Ref 196.) Bosch J, Thabut D, Albillos A, et al. Recombinant factor VIIa for variceal bleeding in patients with advanced cirrhosis: A randomized, controlled trial. Hepatology 2008; 47:1604-14. (Ref 197.) Cales P, Masliah C, Bernard B, et al. French Club for the study of portal hypertension. Early administration of vapreotide for variceal bleeding in patients with cirrhosis. N Engl J Med 2001; 344:23-8. (Ref 151.) de Franchis R, Eisen GM, Laine L, et al. Esophageal capsule endo­ scopy for screening and surveillance of esophageal varices in

patients with portal hypertension. Hepatology 2008; 47:1595-603. (Ref 96.) Failli P, DeFranco R, Caligiuri A, et al. Nitrovasodilators inhibit plateletderived growth factor-induced proliferation and migration of activated human hepatic stellate cells. Gastroenterology 2000; 119:479-92. (Ref 16.) Fernandez M, Vizzutti F, Garcia-Pagan J, et al. Anti-VEGF receptor-2 monoclonal antibody prevents portal-systemic collateral vessel formation in portal hypertensive mice. Gastroenterology 2004; 126:88694. (Ref 55.) Garcia-Tsao G, Bosch J, Groszmann RJ. Portal hypertension and variceal bleeding—unresolved issues. Summary of an American Association for the Study of Liver Diseases and European Association for the Study of the Liver single-topic conference. Hepatology 2008; 47:176472. (Ref 83.) Henderson JM, Boyer TD, Kutner MH, et al. Distal splenorenal shunt versus transjugular intrahepatic portal systematic shunt for variceal bleeding: A randomized trial. Gastroenterology 2006; 130:1643-51. (Ref 200.) Iwakiri Y, Grisham M, Shah V. Vascular biology and pathobiology of the liver: Report of a single-topic symposium. Hepatology 2008; 47:1754-63. (Ref 48.) Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients with end-stage liver disease. Hepatology 2001; 33:464-70. (Ref 167.) Lebrec D, Poynard T, Hillon P, et al. Propranolol for prevention of recurrent gastrointestinal bleeding in patients with cirrhosis: A controlled study. N Engl J Med 1981; 305:1371-4. (Ref 152.) Perri RE, Chiorean MV, Fidler JL, et al. A prospective evaluation of computerized tomographic (CT) scanning as a screening modality for esophageal varices. Hepatology 2008; 47:1587-94. (Ref 97.) Pinzani M, Milani S, De Franco R, et al. Endothelin 1 is overexpressed in human cirrhotic liver and exerts multiple effects on activated hepatic stellate cells. Gastroenterology 1996; 110:534-48. (Ref 21.) Semela D, Das A, Langer D, et al. Platelet-derived growth factor signaling through ephrin-b2 regulates hepatic vascular structure and function. Gastroenterology 2008; 135:671-9. (Ref 15.) Talwalkar JA, Yin M, Fidler JL, et al. Magnetic resonance imaging of hepatic fibrosis: Emerging clinical applications. Hepatology 2008; 47:332-42. (Ref 109.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

91 Ascites and Spontaneous Bacterial Peritonitis Bruce A. Runyon

CHAPTER OUTLINE Pathogenesis of Ascites  1517 Cirrhotic Ascites  1517 Noncirrhotic Ascites  1517 Clinical Features  1518 History  1518 Physical Examination  1518 Diagnosis  1519 Abdominal Paracentesis  1519 Ascitic Fluid Analysis  1521 Differential Diagnosis of Ascites  1527 Complications  1528 Ascitic Fluid Infection, Including Spontaneous Bacterial Peritonitis  1528

Ascites is of Greek derivation (“askos”) and refers to a bag or sack. The word is a noun and describes pathologic fluid accumulation within the peritoneal cavity. The adjective ascitic is used in conjunction with the word fluid to describe the liquid per se. Therefore, “ascitic fluid” is preferred to “ascites fluid.”

PATHOGENESIS OF ASCITES CIRRHOTIC ASCITES

Ascites occurs in the setting of cirrhosis as a result of the sequence of events detailed in Figure 91-1. The most recent theory of ascitic fluid formation, the “peripheral arterial vasodilation hypothesis,” proposes that both older hypotheses, the underfill and overflow theories, are correct, but that each is operative at a different stage.1 The first abnormality that develops appears to be portal hypertension. Portal pressure increases above a critical threshold, and circulating nitric oxide levels increase. Nitric oxide leads to vasodilatation. As the state of vasodilatation worsens, plasma levels of vasoconstrictor, sodium-retentive hormones increase, renal function deteriorates, and ascitic fluid forms—that is, decompensation occurs. In the setting of volume overload in a patient with cirrhosis and ascites, the explanation for the neurohumoral excitation, which is characteristic of volume depletion, may relate to volume sensors. Animals have sophisticated systems for detecting and preserving vascular perfusion pressures and intravascular osmolality. An organism’s ability to detect changes in intravascular volume (especially volume overload) is limited, however, and is linked to pressure receptors. This observation may explain, in part, the paradox of dramatic volume overload in the face of sympa-

Cellulitis  1534 Tense Ascites  1535 Pleural Effusions  1535 Abdominal Wall Hernias  1535 Treatment of Ascites  1536 Low-Albumin-Gradient Ascites  1536 High-Albumin-Gradient Ascites  1536 Refractory Ascites  1538 Prognosis  1540

thetic nervous traffic and hormone levels that are indicative of intravascular volume depletion.

NONCIRRHOTIC ASCITES

The mechanism of fluid retention in patients with malignancy-related ascites depends on the location of the tumor. Peritoneal carcinomatosis appears to cause ascites through the production of proteinaceous fluid by tumor cells lining the peritoneum. Extracellular fluid enters the peritoneal cavity to reestablish oncotic balance. Fluid accumulates in patients with massive liver metastases because of portal hypertension caused by stenosis or occlusion of portal veins by tumor nodules or tumor emboli.2 In patients with hepatocellular carcinoma, ascites arises because of the underlying cirrhosis-related portal hypertension, tumor-induced portal vein thrombosis, or both. Chylous ascites in patients with malignant lymphoma appears to be caused by lymph node obstruction by tumor and rupture of chyle-containing lymphatics. Ascites can complicate high-output or low-output heart failure or nephrotic syndrome. As in cirrhosis, effective arterial blood volume appears to be decreased, and the vasopressin, renin-aldosterone, and sympathetic nervous systems are activated.3 These changes lead to renal vasoconstriction and sodium and water retention. Fluid then “weeps” from the congested hepatic sinusoids as lymph, as in cirrhotic ascites. Tuberculosis, Chlamydia infection, and coccidioidomycosis probably cause ascites through the production of proteinaceous fluid, as in peritoneal carcinomatosis. Spontaneous bacterial peritonitis does not appear to cause fluid to accumulate; infection develops only in preexisting ascites. In patients with pancreatic or biliary ascites, fluid accumulates by leakage of pancreatic juice or bile into the peritoneal cavity or forms secondary to a “chemical burn” of

1517

1518

Section IX  Liver PHT ↑ Nitric oxide Vasodilatation Renal sodium retention

↑ Sympathetic nervous activity, renin, aldosterone

Overfill of intravascular volume

Ascites formation Figure 91-1.  Pathogenesis of ascites in the setting of cirrhosis. PHT, portal hypertension.

Table 91-1  Causes of Ascites CAUSE

%

Cirrhosis (with or without infection) Miscellaneous portal hypertension-related disorder (including 5% with two causes) Cardiac disease Peritoneal carcinomatosis Miscellaneous nonportal hypertension-related disorders

85 8 3 2 2

Data from Runyon BA, Montano AA, Akriviadis EA, et al. The serum-ascites albumin gradient is superior to the exudate-transudate concept in the differential diagnosis of ascites. Ann Intern Med 1992; 117:215-20.

the peritoneum. After abdominal surgery, especially extensive retroperitoneal dissection, lymphatics may be transected and may leak lymph for varying amounts of time. The mechanism of development of ascites in this condition is similar to that for malignant chylous ascites, namely, lymphatic leak.

CLINICAL FEATURES HISTORY

Most patients (approximately 85%) with ascites in the United States have cirrhosis. The three most common causes of cirrhosis are excess alcohol use, chronic hepatitis C, and nonalcoholic steatohepatitis (NASH) related in many cases to obesity. As the obesity epidemic evolves, NASH could become the most common cause of cirrhosis. Many patients have two of these conditions, and some have all three.4 In approximately 15% of patients with ascites, a nonhepatic cause of fluid retention is identified (Table 91-1). Ascites frequently develops during a patient’s first episode of decompensation of alcoholic liver disease. Ascites can develop early in alcoholic liver disease in the precirrhotic, alcoholic hepatitis stage. At this stage, portal hypertension and the resulting predisposition to sodium retention are reversible with abstinence from alcohol. Patients with precirrhotic alcoholic liver disease may lose their predisposi-

tion to fluid retention when they reduce or cease consumption of alcohol. Evidence is accumulating that cirrhosis unrelated to alcohol use can also be reversible with effective therapy.5 Whether a decompensated cirrhotic liver can revert to a normal liver, however, remains to be seen. Many patients with cirrhosis and ascites will ultimately require liver transplantation. Patients with ascites should be questioned about risk factors for liver disease other than alcohol, such as injection drug use, blood transfusions, sex with a same-gender partner, acupuncture, tattoos, ear piercing, and country of origin. Commonly, the cause of ascites in a middle-aged or elderly woman is viral hepatitis–induced cirrhosis resulting from a remote, often forgotten blood transfusion. Another cause of “cryptogenic” cirrhosis and ascites is NASH from long-standing obesity.6 Many patients who have been obese will spontaneously lose 50 or even 100 pounds after their liver disease decompensates. Unless the patient is questioned about lifetime maximum body weight and usual adult body weight, the possibility of NASH-related cirrhosis may not be considered. With careful history-taking and appropriate laboratory testing, the percentage of patients with cirrhosis who are now labeled cryptogenic is approaching zero.6 Patients with a long history of stable cirrhosis and the sudden development of ascites should be suspected of harboring a hepatocellular carcinoma that has precipitated the decompensation. Patients with ascites who have a history of cancer should be suspected of having malignancy-related ascites. Cancer in the past, however, does not guarantee a malignant cause of ascites. For example, patients with tobacco-related lung cancer and a history of alcohol abuse may have ascites due to cirrhosis. Breast, lung, colon, and pancreatic cancers are regularly complicated by ascites.2 Abdominal pain is a helpful distinguishing feature. Malignancy-related ascites frequently is painful, whereas cirrhotic ascites usually is not, unless bacterial peritonitis or alcoholic hepatitis is superimposed. A history of heart failure may raise the possibility of cardiac ascites. Alcoholic patients in whom ascites develops may have alcoholic cardiomyopathy or alcoholic liver disease, but usually not both. Tuberculous peritonitis usually manifests as fever and abdominal pain. Many affected patients are recent immigrants from an endemic area. In the United States, more than one half of the patients with tuberculous peritonitis have underlying alcoholic cirrhosis, which may contribute to the formation of ascitic fluid. Ascites may occur in patients with acute pancreatitis with necrosis or a ruptured pancreatic duct from chronic pancreatitis or trauma. Often troublesome ascites also may develop in a small percentage of patients undergoing hemodialysis. Fitz-Hugh–Curtis syndrome caused by Chlamydia or gonorrhea may cause inflammatory ascites in a sexually active woman. Patients in whom ascites and anasarca develop in the setting of diabetes mellitus should be suspected of having nephrotic ascites. Ascites in a patient with symptoms and signs of myxedema should prompt assessment of thyroid function. Serositis in a patient with a connective tissue disease may be complicated by ascites.7

PHYSICAL EXAMINATION

On the basis of the history and the appearance of the abdomen, the diagnosis of ascites is readily suspected and usually confirmed easily on physical examination. The presence of a full, bulging abdomen should lead to percussion of the flanks. If the degree of flank dullness is greater

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis than usual (i.e., if the percussed air-fluid level is higher than that normally found on the lateral aspect of the abdomen with the patient supine), the examiner should check for “shifting.” If flank dullness is absent, checking for shifting is unnecessary. Approximately 1500 mL of fluid must be present before dullness is detected.8 If flank dullness is not present, the chance that the patient has ascites is less than 10%.8 A fluid wave is not worth testing for.8 Gaseous distention of the bowel, a thick panniculus, and an ovarian mass can mimic ascites. Gaseous distention should be readily apparent on percussion. Ovarian masses usually cause tympanitic flanks with central dullness. Also, the speed of increase in abdominal girth can be helpful; ascites develops in days to weeks, whereas thickening of omentum and panniculus takes months to years. An obese abdomen may be diffusely dull to percussion, and abdominal ultrasonography may be required to determine if fluid is present. Ultrasonography can detect as little as 100 mL of fluid in the abdomen.9 The presence of palmar erythema, large pulsatile spider angiomata, large abdominal wall collateral veins, or fetor hepaticus is suggestive of parenchymal liver disease and portal hypertension. The presence of large veins on the patient’s back suggests inferior vena cava blockage. An immobile mass in the umbilicus, the Sister Mary Joseph nodule, is suggestive of peritoneal carcinomatosis. The neck veins of patients with ascites should always be examined. Alcoholic cardiomyopathy with cardiac ascites can mimic cirrhosis with ascites; an elevated jugular venous pressure helps with this aspect of the differential diagnosis. Constrictive pericarditis is one of the few curable causes of ascites. Most patients with cardiac ascites have impressive jugular venous distention. Some have no visible jugular venous distention but such high central venous pressures that their bulging forehead veins rise to the top of their skulls. When present, peripheral edema in patients with liver disease is usually found in the lower extremities and occasionally may involve the abdominal wall. Patients with nephrotic syndrome or cardiac failure may have total body edema (anasarca).

DIAGNOSIS Although the diagnosis of ascites may be suspected on the basis of the history and physical examination, final confirmation is based on successful abdominal paracentesis or detection of ascites on imaging. Determination of the cause of ascites is based on the results of the history, physical examination, and ascitic fluid analysis. In general, few other tests are required.

ABDOMINAL PARACENTESIS Indications

Abdominal paracentesis with appropriate ascitic fluid analysis is probably the most rapid and cost-effective method of diagnosing the cause of ascites. Also, because of the possibility of ascitic fluid infection in a cirrhotic patient admitted to the hospital, a surveillance paracentesis performed on admission may detect unexpected infection.9 Not all patients with ascitic fluid infection are symptomatic; many have subtle symptoms, such as mild confusion noticed only by the family. Detection of infection at an early asymptomatic stage may reduce mortality. Therefore, ascitic fluid should be sampled in all inpatients and outpatients with new-onset ascites and in all patients with ascites who are

admitted to the hospital. Paracentesis should be repeated in patients (whether hospitalized or not) in whom symptoms, signs, or laboratory abnormalities suggestive of infection develop (e.g., abdominal pain or tenderness, fever, encephalopathy, hypotension, renal failure, acidosis, peripheral leukocytosis).

Contraindications

Few contraindications to paracentesis have been recognized. Coagulopathy is a potential contraindication; however, most patients with cirrhotic ascites have coagulopathy, and if mild to moderate coagulopathy were viewed as a contraindication to paracentesis, few patients with cirrhosis would undergo this procedure.10 Coagulopathy should preclude paracentesis only when clinically evident fibrinolysis or disseminated intravascular coagulation is present.10 These conditions occur in fewer than 1 per 1000 paracenteses. No data are available to support cutoff values for coagulation parameters beyond which paracentesis should be avoided. Global coagulation is usually normal in the setting of cirrhosis despite abnormal tests of coagulation because there is a balanced deficiency of procoagulants and anticoagulants.11 Even after multiple paracenteses, bloody ascites usually does not develop in patients with severe prolongation of the prothrombin time. Patients with cirrhosis and without clinically obvious coagulopathy simply do not bleed excessively from needlesticks unless a blood vessel is entered.10 Studies regarding complications of paracentesis in patients with ascites have documented no deaths or infections caused by paracentesis.9,10 No episodes of hemoperitoneum or entry of the paracentesis needle into the bowel have been reported in these studies. Complications have included only abdominal wall hematomas in approximately 2% of paracenteses, even though 71% of the patients had an abnormal prothrombin time and 21% had a prothrombin time prolonged by more than five seconds.10 Complication rates may be higher when paracentesis is performed by an inexperienced operator. Transfusion of blood products (fresh frozen plasma or platelets) routinely before paracentesis in cirrhotic patients with coagulopathy, presumably to prevent hemorrhagic complications, is not supported by data. Because a hematoma that necessitates blood transfusion develops in only approximately 1% of patients who undergo paracentesis without prophylactic transfusion of plasma or platelets, approximately 100 to 200 units of fresh frozen plasma or platelets would have to be given to prevent the transfusion of approximately 2 units of red blood cells. In a prospective study of 1100 therapeutic paracenteses, no blood products were given prior to the procedure nor were they needed after the procedure despite a platelet count as low as 19,000 cells/mm3 [0.25 × 109/L]) and international normalized ratio (INR) as high as 8.7.12

Patient Position and Choice of Needle and Entry Site

The volume of fluid in the abdomen and the thickness of the abdominal wall determine, in part, how the patient should be positioned in preparation for paracentesis. Patients with a large volume of ascites and thin abdominal wall can be “tapped” successfully in the supine position, with the head of the bed or examining table elevated slightly. Patients with less fluid can be placed in the lateral decubitus position and tapped in the midline or in the right or left lower quadrant while supine (see later). Patients with small amounts of fluid may be tapped successfully only in the face-down position or with ultrasound guidance.13

1519

1520

Section IX  Liver The choice of the site for inserting the needle has changed over the years because of the increasing prevalence of obesity and frequency of therapeutic paracentesis. Paracentesis in obese patients poses special challenges. In obese patients, the abdominal wall usually is substantially thicker in the midline than in the lower quadrants on ultrasound examination.13 The abdominal wall may be even thicker than the length of a 3.5-inch paracentesis needle. Also, on physical examination, determining whether ascites is present or absent in the obese patient is frequently difficult. Ultrasound examination is helpful in confirming the presence of fluid and in guiding the paracentesis needle. Preferably, the needle is inserted into the left lower quadrant, rather than the right lower quadrant because the cecum may be distended with gas from lactulose therapy. Also, the right lower quadrant is more likely than the left to have a surgical scar (e.g., from an appendectomy). When therapeutic paracentesis is performed, more fluid can be obtained using a lower quadrant needle insertion site than a midline site. The needle must be placed several centimeters from a surgical scar. The bowel may be adherent to the peritoneal surface of the abdomen near a scar, and a needle inserted there may enter the bowel.9 A long midline scar precludes midline paracentesis. An appendectomy scar precludes a right lower quadrant site, in general. I usually choose a site in the left lower quadrant two fingerbreadths (3 cm) cephalad and two fingerbreadths medial to the anterior superior iliac spine.13 In a patient with multiple abdominal scars, ultrasound guidance may be required. In a patient who is not overweight, I prefer to use a standard metal 1.5-inch, 22-gauge needle. Paracentesis in obese patients requires the use of a longer needle, for example, one that is 3.5 inches and 22 gauge. Steel needles are preferable to plastic-sheathed cannulas because plastic sheaths may shear off into the peritoneal cavity, with the potential to kink and obstruct the flow of fluid after the cannula is removed. Metal needles do not puncture the bowel unless the bowel is adherent to a scar or severe gaseous distention is present.

Technique

Diagnostic Paracentesis Drapes, gown, hat, and mask are optional, but sterile gloves should be used when paracentesis is performed. The skin is disinfected with an iodine solution. The skin and subcutaneous tissue should be infiltrated with a local anesthetic. The sterile package insert enclosing the gloves can be used as a sterile field on which to place syringes, needles, gauze, and other supplies. When sterile gloves are not used, ascitic fluid cultures frequently grow skin contaminants; a single viable organism will grow to detectable levels in blood culture bottles. To prevent leakage of fluid after the needle is withdrawn, a special technique is required. The previously used term “Z tract” led to confusion about the precise technique: It does not involve manipulating the needle up and down, as this could lead to tissue injury. This technique of needle insertion is accomplished by displacing (with one gloved hand) the skin approximately 2 cm downward and then slowly inserting the paracentesis needle mounted on the syringe held in the other hand. The hand holding the syringe stabilizes the syringe and retracts its plunger simultaneously. A steady hand and experience are needed. The skin is released only after the needle has penetrated the peritoneum and fluid flows. When the needle is ultimately removed, the skin resumes its original position and

seals the needle pathway. (If the needle were inserted straight into the peritoneum from the skin surface, the fluid would leak out easily because the pathway would be straight.) The needle should be advanced slowly through the abdominal wall in approximately 5-mm increments. Slow insertion allows the operator to see blood if a vessel is entered, so that the needle can be withdrawn immediately before further damage is done. Slow insertion also allows the bowel to move away from the needle, thereby avoiding bowel puncture. The syringe that is attached to the needle should be aspirated intermittently during insertion. If continuous suction is applied, bowel or omentum may be drawn to the end of the needle as soon as the needle enters the peritoneal cavity, thereby occluding flow and resulting in an apparently unsuccessful tap. Slow insertion also allows time for the elastic peritoneum to “tent” over the end of the needle and be pierced by it. The most common causes of an unsuccessful paracentesis are continuous aspiration during insertion of the needle and rapid insertion and withdrawal of the needle before the peritoneum is pierced. If the operator is certain that the needle tip is inserted far enough but no fluid is apparent, the syringe and needle can be twisted 90 degrees to pierce the peritoneum, thereby permitting flow of fluid. Approximately 30 mL of fluid is obtained using one or more syringes. I prefer to use a 5- or 10-mL syringe for the initial portion of a diagnostic tap and then twist this syringe off the needle and replace it with a 20- or 30-mL syringe to obtain the remainder of the sample. The initial use of a small syringe allows the operator to have better control and to see fluid more easily as it enters the hub of the syringe. The syringe and attached needle are then pulled out of the abdomen, and the needle is removed and discarded. A sterile needle is then placed on the larger syringe, and an appropriate amount of fluid is inoculated into each of a pair of prepared blood culture bottles (see later). Usually, 5 to 10 mL is inoculated into 50-mL bottles, and 10 to 20 mL into 100-mL bottles. The next aliquot is placed into a “purple-top” ethylenediaminetetraacetic acid tube for a cell count, and the final aliquot is placed into a “red-top” tube for chemistries. Inoculating the culture bottles first with a sterile needle minimizes contamination. The fluid must be placed promptly into the anticoagulant-containing tube to avoid clotting; clotted fluid cannot be analyzed for cell count. Therapeutic Paracentesis Therapeutic paracentesis is similar to diagnostic paracentesis except that a larger-bore needle is used and additional equipment is required. In the patient who is not overweight, I prefer to use a standard metal 1.5-inch, 16- to 18-gauge needle. Obese patients may require a longer needle, for example, one that is 3.5 inches and 18 gauge. A set of 15-gauge five-hole needles has been produced specifically for therapeutic abdominal paracentesis; these needles may replace the spinal needles used currently for paracentesis in obese patients. The 15-gauge needles have a removable sharp inner component and a blunt outer cannula; they range in length from 3.25 to 5.9 inches. A tiny scalpel nick is required to permit the large needle to enter the skin. An old method of using a 60-mL syringe, stopcock, and collection bag is tedious; use of vacuum bottles (1 or 2 L) connected to the needle with noncollapsible tubing is much faster. Use of a pump is even faster than vacuum bottles. Unless the needle is allowed to drift subcutaneously, the needle (or blunt steel cannula) can be left in the abdomen during a therapeutic paracentesis without injury. Larger-

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis bore needles or cannulas permit more rapid removal of fluid but leave larger defects if they enter vessels or the bowel inadvertently. Once fluid is flowing, the needle should be stabilized to ensure steady flow. Not unusually, flow ceases intermittently. With respiratory movement, the needle may gradually work its way out of the peritoneal cavity and into the soft tissue, and some serosanguineous fluid may appear in the needle hub or tubing. When this happens, the pump should be turned off or a clamp placed on the tubing connected to the vacuum bottle. The tubing is removed from the needle, and the needle is twisted a few degrees. If flow does not resume, the needle is twisted a bit more. If flow still does not resume, the needle is inserted in 1- to 2-mm increments until brisk dripping of fluid from the needle hub is seen. The tubing is then reattached, and more fluid is removed. Occasionally, fluid cannot be aspirated but drips from the needle hub. In this situation, fluid is allowed to drip into a sterile container for collection, as in a lumbar puncture. As the fluid is removed, the bowel and omentum draw closer to the needle and eventually block the flow of ascitic fluid. The patient must then be repositioned so that gravity causes the fluid to pool near the needle. It is useful to reposition the patient a few times during a total paracentesis to maximize the amount of fluid removed. Excessive manipulation of the needle is avoided, to minimize the risk of trauma to the bowel or blood vessels. After samples of fluid are obtained for testing, 2 to 4 L of fluid is removed to relieve the pressure of tense ascites in patients with new or diuretic-sensitive ascites. A sodiumrestricted diet and diuretics are prescribed to reduce the fluid further (see later). If a patient is known to be diuretic-resistant, a “total tap” is performed—that is, all of the fluid that is accessible is removed. If less is removed, the tap will need to be repeated soon (see later—“Refractory Ascites”).

ASCITIC FLUID ANALYSIS Gross Appearance

Non-neutrocytic (i.e., ascitic fluid polymorphonuclear neutrophil [PMN] count less than 250/mm3 [0.25 × 109/L]) ascitic fluid is transparent and usually slightly yellow (Fig. 91-2). Ascitic fluid with a very low protein concentration may have no pigment and look like water. The opacity of many cloudy ascitic fluid specimens is caused by neutrophils. The presence of neutrophils leads to a shimmering effect when a glass tube containing the fluid is rocked back and forth in front of a light. Fluid with an absolute neutrophil count less than 1000/mm3 (1.0 × 109/L) may be nearly clear. Fluid with a count greater than 5000/mm3 (5.0 × 109/L) is quite cloudy, and fluid with a count greater than 50,000/mm3 (50.0 × 109/L) resembles mayonnaise. Ascitic fluid specimens frequently are blood-tinged or frankly bloody. A red blood cell count of 10,000/mm3 (10.0 × 109/L) is the threshold for a pink appearance; lower concentrations result in clear or turbid fluid. Ascitic fluid with a red blood cell count greater than 20,000/mm3 (20.0 × 109/L) is distinctly red. Many ascitic fluid specimens are bloody because of a traumatic tap; these specimens are blood-streaked and frequently clot unless the fluid is transferred immediately to the anticoagulant-containing tube for the cell count. By contrast, nontraumatic or remotely traumatic blood-tinged ascitic fluid is homogeneous and does not clot because it has already clotted and the clot has lysed. Some patients with portal hypertension have bloody hepatic lymph, resulting in bloody ascitic fluid—perhaps because

of rupture of lymphatics that are under high pressure. Samples from patients with hepatocellular carcinoma are regularly bloody, but only about 10% of samples from patients with peritoneal carcinomatosis are red.2 Although many physicians have the impression that tuberculosis results in bloody ascitic fluid, less than 5% of tuberculous samples are hemorrhagic in my experience. Ascitic fluid frequently is lipid-laden. Lipid opacifies the fluid. The degree of opalescence of ascitic fluid ranges from slightly cloudy to completely opaque and chylous. Most opaque, milky fluid samples have a triglyceride concentration greater than 200 mg/dL (2.26 mmol/L) and usually greater than 1000 mg/dL (11.30 mmol/L). Fluid that has the appearance of dilute skim milk has a triglyceride concentration between 100 mg/dL (1.13 mmol/L) and 200 mg/dL (2.26 mmol/L). A substantial minority of cirrhotic ascitic fluid samples are neither transparent nor frankly milky. These opalescent samples have slightly elevated triglyceride concentrations ranging from 50 mg/dL (0.56 mmol/L) to 200 mg/dL (2.26 mmol/L).14 The opacity of these fluids does not have the shimmering characteristics of ascitic fluid with an elevated white blood cell count. The lipid usually layers out when a tube of ascitic fluid is placed in the refrigerator for 48 to 72 hours. In contrast with findings in older published reports, most patients with chylous or opalescent ascites have cirrhosis.14,15 Dark-brown fluid with a bilirubin concentration greater than that of serum usually indicates biliary perforation.16 Deeply jaundiced patients have bile-stained ascitic fluid, but the bilirubin level and the degree of pigmentation are visually less than those of the corresponding serum. Pancreatic ascites may be pigmented because of the effect of pancreatic enzymes on red blood cells. The red blood cells may have to be centrifuged before the discolored supernatant is revealed. The degree of pigmentation ranges from tea-colored to jet black, as in pancreatic necrosis (formerly hemorrhagic pancreatitis). Black ascitic fluid also may be found in patients with malignant melanoma.

Tests

The practice of ordering every available body fluid test on every ascitic fluid specimen is expensive and can be more confusing than helpful, especially when unexpectedly abnormal results are encountered. An algorithm for the analysis of ascitic fluid is shown in Figure 91-2. The basic concept is that screening tests are performed on the initial specimen; additional testing is performed only when necessary as indicated by the results of the screening tests. Further testing may require another paracentesis, but because most specimens consist of ascitic fluid resulting from uncomplicated cirrhosis, no further testing is needed in a majority of cases. Also, because laboratories frequently store the fluid for a few days, additional testing can often be ordered on the stored fluid. On the basis of cost analysis, tests can be classified as routine, optional, unusual, and unhelpful (Table 91-2).9 The cell count is the single most helpful ascitic fluid test. Only approximately 10 µL of fluid is required for a standard manual hemocytometer count. Therefore, if only one drop of fluid can be obtained, it should be sent for cell count. More fluid is almost always obtainable, however. The fluid should be submitted in an anticoagulant-containing tube (i.e., ethylenediaminetetraacetic acid) to prevent clotting. Because the decision to begin empirical antibiotic treatment of suspected ascitic fluid infection is based largely on the absolute neutrophil count (which should have a turnaround time of a few minutes), rather than the culture (which takes

1521

Subtract 1 WBC/ 750 RBCs Subtract 1 PMN/ 250 RBCs

Triglyceride concentration

Bilirubin concentration

Bloody

Milky

Dark brown

Special testing or Cell count correction

≥500

<500

<50% PMNs

≥250

<250

<1.1

≥1.1

<1.1

≥1.1

<1.1

≥1.1

Serum-ascites albumin gradient (SAAG) (g/dL)

≥50% PMNs

White blood Polymorphocell (WBC) nuclear count neutrophil (cells/mm3) (PMN) count (cells/mm3)

Tuberculous peritonitis and underlying cirrhosis Peritoneal carcinomatosis

Send fluid for tuberculosis testing Positive cytology

Tuberculous peritonitis

Search for primary tumor

Peritoneal carcinomatosis and portal hypertension Positive cytology

Send fluid for tuberculosis testing

Abdominal computed tomography Pancreatic ascites

Ascitic fluid amylase > 100 U/L

Mycobacterial growth on culture of laparoscopic biopsy specimen of peritoneum

Upright abdominal film, water soluble contrast studies of the GI tract

Secondary bacterial peritonitis

Polymicrobial infection, TP > 1 g/dL, Glucose < 50 mg/dL, LDH ≥ 225 U/L

Clinical response to antibiotic SBP

24-hour urine protein excretion

Chest roentgenogram and echocardiogram

Ultrasound and/or liver biopsy

Confirmatory testing

Single organism in culture, TP < 1 g/dL, Glucose > 50 mg/dL, LDH < 225 U/L

Nephrotic ascites

Cardiac ascites

Total protein ≥2.5 g/dL Total protein <2.5 g/dL

Uncomplicated cirrhotic ascites

Working diagnosis

Total protein <2.5 g/dL

Other testing

Figure 91-2.  Algorithm for the approach to the differential diagnosis of ascites. GI, gastrointestinal; LDH, lactate dehydrogenase; PMN, polymorphonuclear neutrophil; RBC, red blood cell; SBP, spontaneous bacterial peritonitis; TP, total protein.

Abdominal paracentesis

Transparent yellow or Crystal clear or Cloudy yellow

Gross appearance of fluid

1522 Section IX  Liver

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis Table 91-2  Ascitic Fluid Laboratory Tests ROUTINE

OPTIONAL

UNUSUAL

UNHELPFUL

Cell count Albumin Total protein

Amylase Culture in blood culture bottles Glucose Gram stain LDH

Bilirubin Cytology TB smear, culture, and PCR test Triglycerides

Cholesterol Fibronectin Lactate pH

LDH, lactate dehydrogenase; PCR, polymerase chain reaction; TB, tuberculosis.

12 to 48 hours to demonstrate growth), the cell count is more important than the culture in the early detection and treatment of ascitic fluid infection. Even samples from asymptomatic outpatients undergoing therapeutic paracentesis should be sent for a cell and differential count; the information obtained can lead to early, life-saving treatment of bacterial infection. Cell Count Surprisingly, ascitic fluid cell counts have not been standardized. Some laboratories count mesothelial cells in addition to white blood cells (WBCs) and label the sum as “nucleated cells.” The usefulness of mesothelial cell counts is not clear. The WBC count in uncomplicated cirrhotic ascites is usually less than 500 cells/mm3 (0.5 × 109/L) (see Fig. 91-2).9,17 During diuresis in patients with cirrhotic ascites, the WBC count can concentrate to more than 1000 cells/mm3 (1.0 × 109/L).17 A diagnosis of diuresis-related elevation of the ascitic fluid WBC count, however, requires that a prediuresis count be available, that normal lymphocytes predominate in the fluid, and that unexplained clinical symptoms or signs (e.g., fever or abdominal pain) be absent. The upper limit of normal for the absolute PMN count in uncomplicated cirrhotic ascitic fluid is usually stated to be lower than 250/mm3 (0.25 × 109/L).9,17 The short survival of PMNs results in relative stability of the absolute PMN count during diuresis.17 Therefore, the 250 cells/mm3 (0.25 × 109/L) cutoff value remains reliable even after diuresis. New methods have been developed to estimate the number of ascitic fluid cells.18 Dipsticks can detect an ascitic fluid PMN count greater than 250/mm3 (0.25 × 109/L) in 90 to 120 seconds. Urine-specific dipsticks have been used to date and are not very sensitive.19 What is now needed is an ascitic fluid–specific dipstick. Any inflammatory process can result in an elevated ascitic fluid WBC count. Spontaneous bacterial peritonitis is the most common cause of inflammation of ascitic fluid and the most common cause of an elevated ascitic WBC count (see later). The total WBC count, as well as the absolute PMN count, is elevated in spontaneous bacterial peritonitis, and PMNs usually account for more than 70% of the total WBC count. Also, in tuberculous peritonitis and peritoneal carcinomatosis, the total ascitic WBC count is frequently elevated, but usually with a predominance of lymphocytes.2 In most instances, bloody ascitic fluid is the result of a slightly traumatic tap. Leakage of blood into the peritoneal cavity leads to an elevated ascitic fluid WBC count. Because neutrophils predominate in blood, the ascitic fluid differential count may be altered by contamination of ascitic fluid with blood. To correct for this, 1 PMN is subtracted from the absolute ascitic fluid PMN count for every 250 red blood cells17 (see Fig. 91-2). If the leakage of blood occurred at a remote time, the PMNs will have lysed, and the corrected

Table 91-3  Classification of Ascites by Serum-Ascites Albumin Gradient HIGH GRADIENT ≥1.1 g/dL (11 g/L)

LOW GRADIENT <1.1 g/dL (11 g/L)

Alcoholic hepatitis Budd-Chiari syndrome Cardiac ascites Cirrhosis Fatty liver of pregnancy Fulminant hepatic failure Massive liver metastases “Mixed” ascites Myxedema Portal vein thrombosis Sinusoidal obstruction syndrome

Biliary ascites Bowel obstruction or infarction Nephrotic syndrome Pancreatic ascites Peritoneal carcinomatosis Postoperative lymphatic leak Serositis in connective tissue diseases Tuberculous peritonitis

PMN count will be a negative number. If the corrected PMN count in a bloody specimen is greater than or equal to 250 cells/mm3 (0.25 × 109/L), the patient must be assumed to be infected. Exudate/Transudate Classification Before the 1980s, the ascitic fluid total protein concentration was used to classify ascites as either exudative (greater than 2.5 g/dL [25 g/L]) or transudative (less than 2.5 g/dL [25 g/L]). Unfortunately, this classification does not work well in ascitic fluid, and these terms as applied to ascitic fluid were never carefully defined or validated. Attempts at using combinations of lactate dehydrogenase (LDH) and serum-to–ascitic fluid ratios of LDH and protein also have not been shown to classify ascitic fluid accurately into exudates and transudates.20 Serum-Ascites Albumin Gradient The serum-ascites albumin gradient (SAAG) has been proved to categorize ascites better than the total protein concentration or other parameters21 (Table 91-3). The SAAG is based on oncotic-hydrostatic balance. Portal hypertension results in an abnormally high hydrostatic pressure gradient between the portal bed and ascitic fluid. A similarly large difference must exist between ascitic fluid and intravascular oncotic forces. Albumin exerts greater oncotic force per gram than that exerted by other proteins. Therefore, the difference between the serum and ascitic fluid albumin concentrations correlates directly with portal pressure. Calculating the SAAG involves measuring the albumin concentration of serum and ascitic fluid specimens and simply subtracting the ascitic fluid value from the serum value. Unless a laboratory error has been made, the serum albumin concentration is always the larger value. The gradi-

1523

1524

Section IX  Liver ent is calculated by subtraction and is not a ratio. If the SAAG is 1.1 g/dL (11 g/L) or greater, the patient can be considered to have portal hypertension with an accuracy of approximately 97%.21 Also, if the serum albumin minus ascitic fluid total protein gradient is 1.1 g/dL (11 g/L) or greater, the patient has portal hypertension because the ascitic fluid albumin concentration cannot be greater than the ascitic fluid total protein concentration. Conversely, if the SAAG is less than 1.1 g/dL (11 g/L), the patient is unlikely to have portal hypertension. The SAAG does not explain the pathogenesis of ascites formation, nor does it explain where the albumin came from—that is, liver or bowel. It simply gives the physician an indirect but accurate index of portal pressure. The accuracy of the test is excellent, even with ascitic fluid infection, diuresis, therapeutic paracentesis, intravenous infusions of albumin, and various causes of liver disease.21 Measurement of the ascitic fluid albumin concentration has been routine in some laboratories since the 1980s. Nevertheless, before sending ascitic fluid for the first time to a laboratory to measure the albumin concentration, a physician should discuss the test with the laboratory chemist. The accuracy of the albumin assay at low albumin concentrations (e.g., less than 1 g/dL [10 g/L]) should be confirmed because many patients with ascites have a serum albumin concentration in the range of 2.0 g/dL (20 g/L) and an ascitic fluid albumin concentration in the range of 0 to 1.0 g/dL (0 to 10 g/L). If a patient with cirrhosis has a serum albumin level of less than 1.1 g/dL (11 g/L), as occurs in less than 1% of patients with cirrhotic ascites, the SAAG will be falsely low. The accuracy of the SAAG is also reduced when specimens of serum and ascites are not obtained nearly simultaneously. The specimens should be obtained on the same day, preferably within the same hour. Both serum and ascitic fluid albumin concentrations change over time; however, these values change in parallel, so the difference is stable. Arterial hypotension may result in a decrease in the portal pressure and a narrowing of the SAAG. Lipid interferes with the assay for albumin, and chylous ascites may result in a falsely high SAAG. Serum hyperglobulinemia (serum globulin level greater than 5 g/dL [50 g/L]) leads to a high ascitic fluid globulin concentration and can narrow the albumin gradient by contributing to the oncotic forces. A narrowed gradient caused by high serum globulin levels occurs in only approximately 1% of ascitic fluid specimens. To correct the SAAG in the setting of a high serum globulin level, the following formula is used22: Corrected SAAG = uncorrected SAAG × 0.16 × (serum globulin [ g dL] + 2.5) Approximately 5% of patients with ascites have “mixed” ascites (that is, two causes of ascites) (see Table 91-1). Most of these patients have portal hypertension from cirrhosis as well as another cause of ascites, such as tuberculous peritonitis or peritoneal carcinomatosis.21 The albumin gradient is high (1.1 g/dL [11 g/L] or greater) in mixed ascites, as a reflection of the underlying portal hypertension.21 The presence of a high SAAG does not confirm a diagnosis of cirrhosis; it simply indicates the presence of portal hypertension. Many causes of portal hypertension other than cirrhosis are recognized (see Tables 91-1 and 91-3 and Chapter 90). A low SAAG does not confirm a diagnosis of peritoneal carcinomatosis. Although peritoneal carcinomatosis is the most common cause of a low SAAG, other causes

exist (see Table 91-3). The SAAG needs to be determined only on the first paracentesis specimen in a given patient; it does not need to be repeated on subsequent specimens, if the first value is definitive. If the first result is borderline (e.g., 1.0 or 1.1 g/dL [10 or 11 g/L]), repeating the paracentesis and analysis usually provides a definitive result. Highalbumin-gradient and low-albumin-gradient should replace the modifiers “transudative” and “exudative” in the classification of ascites.21 Culture In the past, culture methodology for ascitic fluid was based on the notion that most episodes of ascitic fluid infection were polymicrobial with high colony counts, as in surgical peritonitis. The most common bacterial infection of ascitic fluid, spontaneous bacterial peritonitis, is monomicrobial, however, with a low bacterial concentration (median colony count of only 1 organism/mL).23 The older method of culture consisted of inoculation (in the microbiology laboratory) of each of three agar plates and some broth with a few drops of fluid. This method of culturing ascitic fluid, as is used for urine or stool, is predictably insensitive for detecting monomicrobial infections with a low colony count. Spontaneous bacterial peritonitis is more like bacteremia in terms of the number of bacteria present; culturing ascitic fluid as if it were blood has a high yield.23 In fact, the sensitivity of culture in detecting bacterial growth in neutrocytic ascites (i.e., ascitic fluid with a PMN count of 250 cells/mm3 [0.25 × 109/L] or greater) depends on the method of culture used. The older method of culture has been found to detect bacterial growth in approximately 50% of neutrocytic samples, whereas bedside inoculation of blood culture bottles with ascitic fluid detects growth in approximately 80%.9 Multiple prospective studies have demonstrated the superiority of the blood culture bottle method.9 Also, bedside inoculation is superior to delayed laboratory inoculation of blood culture bottles in the laboratory.24 Gene probes are now commercially available for the detection of bacteremia; hopefully, they will also lead to rapid (30-minute) and accurate detection of organisms in ascitic fluid. Culture will continue to be required, however, for assessment of the susceptibility of the organism to antibiotics. Total Protein As noted earlier, the antiquated exudate/transudate system of ascitic fluid classification, which is based on ascitic fluid total protein concentration, is problematic. The protein concentration in ascitic fluid in the setting of cirrhosis is determined almost entirely by the serum protein concentration and portal pressure. A patient with cirrhosis and a relatively high serum protein concentration will have a relatively high ascitic fluid protein concentration. Because of this relationship, almost 20% of ascitic samples in patients with cirrhosis will have a protein concentration greater than 2.5 g/dL (25 g/L). The ascitic fluid total protein concentration does not increase during spontaneous bacterial peritonitis; it remains stable before, during, and after infection.25 In fact, patients with the lowest ascitic protein concentrations are the most susceptible to spontaneous peritonitis.26 During a 10-kg diuresis, the ascitic fluid total protein concentration doubles, and 67% of such patients with cirrhotic ascites have a protein concentration greater than 2.5 g/dL (25 g/L) by the end of diuresis.17 In almost one third of patients with malignant ascites, the ascites is caused by massive liver metastases or hepatocellular carcinoma, and the ascitic fluid in these patients has a low protein concen-

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis tration.2 In cardiac ascites, the ascitic fluid protein concentration is greater than 2.5 g/dL (25 g/L).27 Therefore, the exudate/transudate method of classification of ascites places many patients with cirrhosis and ascites and all patients with cardiac ascites in the exudate category and many patients with malignant ascites and essentially all patients with spontaneously infected ascites in the transudate category. Clearly, this method of classification is not useful. By contrast, the SAAG classifies fluid by the presence or absence of portal hypertension and is much more physiologic and intuitive in nature.21 The albumin gradient classifies cardiac ascites in the high-SAAG category, similar to cirrhotic ascites. The high SAAG of cardiac ascites is presumably the result of high right-sided cardiac pressures. In patients with cardiac ascites, the SAAG may narrow with diuresis; such narrowing does not happen in patients with cirrhosis. The combination of ascitic fluid total protein, glucose, and LDH is of value in distinguishing spontaneous bacterial peritonitis from intestinal perforation with leakage of gut contents into ascites28 (Fig. 91-3). Patients who have neutrocytic ascitic fluid, in whom the clinical picture suggests bacterial peritonitis (rather than peritoneal carcinomatosis or tuberculous peritonitis) and who meet two of the following three criteria, are likely to have surgical peritonitis and merit immediate radiologic evaluation to determine if intestinal perforation with leakage of intestinal contents into ascites has occurred: total protein greater than 1 g/dL (10 g/L), glucose less than 50 mg/dL (2.8 mmol/L), and LDH greater than the upper limit of normal for serum.28 Glucose The glucose molecule is small enough to diffuse readily into body fluid cavities. Therefore, the concentration of glucose in ascitic fluid is similar to that in serum, unless glucose is being consumed by ascitic fluid WBCs or bacteria.28 In early spontaneous bacterial peritonitis, the ascitic fluid glucose concentration is similar to that of sterile fluid.25 By contrast, in spontaneous bacterial peritonitis detected late in its course (as well as in the setting of intestinal perforation into

Free air or extravasation of contrast Yes medium on abdominal imaging study

ascitic fluid), the ascitic fluid glucose concentration usually drops to 0 mg/dL (0 mmol/L) because of large numbers of stimulated neutrophils and bacteria.28 Lactate Dehydrogenase The LDH molecule is too large to enter ascitic fluid readily from blood,28 and the ascitic fluid concentration of LDH usually is less than one half of the serum level in uncomplicated cirrhotic ascites. In spontaneous bacterial peritonitis, the ascitic fluid LDH level rises because of the release of LDH from neutrophils, and the ascitic fluid concentration is greater than that of serum. In secondary peritonitis, the LDH level is even higher than that seen in spontaneous bacterial peritonitis and may be several-fold higher than the serum LDH level.28 Amylase In uncomplicated ascites in the setting of cirrhosis, the ascitic fluid amylase concentration usually is one half that of the serum value, approximately 50 U/L.29 In patients with acute pancreatitis or intestinal perforation (with release of luminal amylase into the ascitic fluid), the fluid amylase concentration is elevated markedly, usually greater than 2000 U/L and approximately five-fold greater than simultaneous serum values.28-30 Gram Stain Gram stains of body fluids demonstrate bacteria only when more than 10,000 bacteria/mL are present. The median ascitic concentration of bacteria in spontaneous bacterial peritonitis is only 1 organism/mL, similar to the colony count in bacteremia.23 Requesting an ascitic fluid Gram stain to detect bacteria in spontaneous bacterial peritonitis is analogous to requesting a Gram stain of blood to detect bacteremia. Bacteria are detected on Gram stain only with overwhelming infection, as in advanced spontaneous bacterial peritonitis or asplenic pneumococcal sepsis. Gram stain of ascitic fluid is most helpful in the diagnosis of free perforation of the intestine into ascitic fluid. In this setting, sheets of multiple different bacteria are found. Gram stain

Perforation peritonitis

Laparotomy

No

If necessary

Yes Fulfillment of at least 2 of the following: Total protein >1 g/dL Glucose <50 mg/dL LDH > upper limit of normal

Non-perforation secondary bacterial peritonitis No

Ascitic fluid PMN count ≥250 cells/mm3

Percutaneous drainage

Evidence for loculated infection (US, CT)

No

Ascites PMN count < Yes baseline after 48 hours of therapy with antibiotic

Spontaneous bacterial peritonitis

Continue antibiotic

No Ascitic fluid bile-stained

Ascitic fluid bilirubin >6 mg/dL and ascitic fluid/serum bilirubin >1.0

Yes

Biliary perforation

Figure 91-3.  Algorithm for differentiating spontaneous from secondary bacterial peritonitis in patients with neutrocytic ascites (i.e., neutrophil count of 250 cells/ mm3 [0.25 × 109/L] or greater) in the absence of hemorrhage into ascitic fluid, tuberculosis, peritoneal carcinomatosis, or pancreatitis. Antibiotic therapy should be started at the time peritonitis (ascitic fluid PMN count ≥250 cells/mm3) is detected. CT, computed tomography; LDH, lactate dehydrogenase; PMN, polymorphonuclear neutrophil; US, ultrasound. (Reproduced with permission from Akriviadis EA, Runyon BA. The value of an algorithm in differentiating spontaneous from secondary bacterial peritonitis. Gastroenterology 1990; 98:12733. Copyright 1990 by the American Gastroenterological Association.)

1525

1526

Section IX  Liver of the centrifuged sediment of 50 mL of ascites has a sensitivity rate of only 10% for visualizing bacteria in spontaneous bacterial peritonitis.23 Smear and Culture for Tuberculosis A direct smear of ascitic fluid to detect mycobacteria is almost never positive because of the rarity of tuberculous peritonitis and the low concentration of mycobacteria in ascitic fluid in tuberculous peritonitis.31 The older literature suggests that 1 L of fluid should be cultured. The largest centrifuge tube found in most laboratories, however, has a capacity of 50 mL. In general, only one 50-mL aliquot of fluid is centrifuged, and the pellet is cultured. In contrast to a sensitivity rate of approximately 50% for ascitic fluid mycobacterial culture with optimal processing, laparoscopy with histology and culture of peritoneal biopsies has a sensitivity approaching 100% for detecting tuberculous peritonitis.31 Tuberculous peritonitis can easily be confused with spontaneous bacterial peritonitis because both conditions are associated with abdominal pain and fever, and one half of the patients with tuberculous peritonitis have cirrhosis. A negative bacterial culture and predominance of mononuclear cells in the differential count, however, provide clues to the diagnosis of tuberculous peritonitis. DNA probes are now available to detect mycobacteria and probably will replace older methods of detection.32 Nevertheless, cultures still will be required to determine susceptibility to antimicrobial agents. Cytologic Examination In the past, ascites related to malignancy was assumed to be caused only by peritoneal carcinomatosis; massive liver metastases and hepatocellular carcinoma superimposed on cirrhosis were not recognized as causes of malignant ascites. These studies did not compare cytologic examination with a standard diagnostic test, such as autopsy, laparotomy, or laparoscopy, and cytologic study was reported to have a sensitivity of only about 60% in detecting malignant ascites.33 Cytologic studies, however, can be expected to detect malignancy only when tumor cells line the peritoneal cavity and exfoliate into the ascitic fluid—that is, in peritoneal carcinomatosis. Such studies should not be expected to detect tumor when the peritoneum is uninvolved, as in ascites resulting from portal hypertension in patients with hepatocellular carcinoma or massive liver metastases or from lymph node obstruction in patients with malignant lymphoma.2 In one study in which the location and type of tumor that caused ascites were confirmed by a standard test, only approximately two thirds of patients with malignancy-related ascites were found to have peritoneal carcinomatosis, but nearly 100% of patients with peritoneal carcinomatosis were reported to have positive findings on cytologic examination of ascitic fluid; the remaining one third of patients with massive liver metastases, chylous ascites caused by lymphoma, or hepatocellular carcinoma had negative cytologic findings.2 Therefore, the sensitivity of cytology is approximately 100% for detecting peritoneal carcinomatosis but much lower for detecting malignancyrelated ascites caused by conditions other than peritoneal carcinomatosis. Cytologic studies should not be falsely positive if performed carefully; I have never encountered a falsepositive result. Because hepatocellular carcinoma rarely metastasizes to the peritoneum, a positive ascitic fluid cytology in a patient with hepatocellular carcinoma is unusual enough to be the subject of a case report.34 Measurement of the serum alpha fetoprotein concentration (which is always higher in serum than in ascitic fluid) may be of value in detecting hepato­

cellular carcinoma; serum alpha fetoprotein is much more sensitive than ascitic cytology for this purpose.2 In malignancy-related ascites, the fluid may have an elevated PMN count, presumably because dying tumor cells attract neutrophils.2 The elevated PMN count may cause confusion with spontaneous bacterial peritonitis; however, a predo­ minance of lymphocytes in malignancy-related ascites is usual. Flow cytometry and magnetic enrichment of ascitic fluid as an adjunct to cytology may further increase diagnostic accuracy.35 Triglyceride A triglyceride level should be measured in opalescent or frankly milky ascitic fluid (see Fig. 91-2). By definition, chylous ascites has a triglyceride concentration greater than 200 mg/dL (2.26 mmol/L) and greater than the serum level; usually, the level is greater than 1000 mg/dL (11.30 mmol/L).36 In sterile ascitic fluid specimens in the setting of cirrhosis that are slightly cloudy, without an elevated cell count (i.e., opalescent), the triglyceride concentration is elevated—64 ± 40 mg/dL (0.72 ± 0.45 mmol/L), compared with 18 ± 9 mg/dL (0.20 ± 0.10 mmol/L) for clear ascites in the setting of cirrhosis.14 Bilirubin The bilirubin concentration should be measured in ascitic fluid that is dark brown. An ascitic fluid bilirubin level greater than 6 mg/dL (102 µmol/L) and greater than the serum level of bilirubin suggests biliary or proximal small intestinal perforation into ascitic fluid.16,28 Tests That Are Seldom Helpful Tests that have been proposed to be helpful in the analysis of ascitic fluid but shown subsequently to be of no benefit include determination of pH, lactate, fibronectin, and cholesterol. The studies that attempted to validate the value of pH and lactate included small numbers of patients and used suboptimal culture techniques. In the two largest and most recent studies, which did not have some of the deficiencies of the earlier studies, the ascitic fluid pH and lactate were found not to be helpful.37,38 The pH was found to have no impact on decision-making regarding the use of empirical antibiotic therapy.37 Fibronectin and cholesterol have been proposed to be useful in detecting malignant ascites. The basic premise in studies of these markers was that ascitic fluid cytologic examination is insensitive. Unfortunately, the design of the studies was problematic, several subgroups of malignancy-related ascites (e.g., massive liver metastases, hepatocellular carcinoma with cirrhosis) were not considered, and appropriate control groups (e.g., patients with ascites caused by conditions other than cirrhosis or peritoneal carcinomatosis) were not included. Other studies have demonstrated that in patients with massive liver metastases, ascitic fluid fibronectin and cholesterol concentrations are not abnormally elevated.39,40 Therefore, in patients with malignancy-related ascites and negative cytologic findings, these “humoral tests of malignancy” are usually negative. Additionally, patients with high-protein non­ cirrhotic ascites nearly always have ascitic fibronectin and cholesterol elevations despite the absence of malignancy.2,39,40 Carcinoembryonic antigen (CEA) in ascitic fluid has been proposed as a helpful marker for detecting malignant ascites.41 The study that attempted to validate this proposal, however, was flawed, and more studies, with various subgroups of patients, are required before testing for ascitic fluid CEA can be considered validated.

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis Measurement of adenosine deaminase has been proposed as a useful test for detecting peritoneal tuberculosis. In the United States, however, where greater than 50% of patients with tuberculous peritonitis have underlying cirrhosis, the adenosine deaminase level has been found to be too insensitive to be helpful.31

Table 91-4  Classification of Malignancy-Related Ascites Hepatocellular carcinoma Malignant Budd-Chiari syndrome (tumor emboli in hepatic veins) Malignant lymph node obstruction Massive liver metastases Peritoneal carcinomatosis Peritoneal carcinomatosis with massive liver metastases

DIFFERENTIAL DIAGNOSIS OF ASCITES Although cirrhosis is the cause of ascites in most patients with ascites evaluated by an internist, a cause other than liver disease is found in approximately 15% of patients (see Table 91-1). Approximately 5% of patients have two causes of ascites, that is, “mixed” ascites.21 Usually, these patients have cirrhosis plus one other cause, such as peritoneal carcinomatosis or tuberculous peritonitis (see Table 91-1). Because tuberculosis is potentially fatal but curable and frequently occurs in cirrhotic patients with preexisting ascites, the physician must not assume that liver disease is the only cause of ascites in a febrile alcoholic patient if the ascitic fluid analysis is atypical. For example, if the ascitic fluid lymphocyte count is unusually high, tuberculous peritonitis may be present. Interpretation of the results of ascitic fluid analysis is difficult in patients with mixed ascites but crucial to accurate diagnosis and treatment. Additionally, liver diseases other than cirrhosis (e.g., alcoholic hepatitis or fulminant hepatic failure) may cause ascites (see Table 91-1). An algorithm for the differential diagnosis of ascites is shown in Figure 91-2. This proposed strategy is applicable to a majority of patients with ascites, including many with the causes listed in Table 91-1. Not every patient (including patients with rare causes of ascites) can be categorized readily with such an algorithm, however. Many patients with enigmatic ascites eventually are found to have two or even three causes of ascites (e.g., heart failure, cirrhosis caused by NASH, diabetic nephropathy). In these cases, the sum of predisposing factors leads to sodium and water retention, even though each factor alone may not be severe enough to cause fluid overload. In most patients with ascites, cirrhosis is the cause. This form of ascites, especially when low in protein, is complicated frequently by spontaneous bacterial peritonitis (see later).26 Other forms of ascites are complicated by spontaneous peritonitis so rarely that they are the subjects of case reports or small series. The intestine can perforate with spillage of contents in patients with ascites of any cause, cirrhosis or otherwise. The ascitic fluid analysis in intestinal perforation is dramatically different from that in spontaneous bacterial peritonitis (see Fig. 91-3).28 Distinguishing spontaneous bacterial peritonitis from surgical peritonitis in a patient with cirrhosis is critical to the patient’s survival; spontaneous bacterial peritonitis is treated with antibiotics alone, whereas surgical peritonitis is treated with antibiotics and emergency surgical intervention (see Chapter 37). Cancer accounts for fewer than 10% of cases of ascites (see Table 91-1). Not all cases of malignancy-related ascites are caused by peritoneal carcinomatosis; the characteristics of the ascitic fluid and the treatments vary, depending on the pathophysiology of the ascites—for example, peritoneal carcinomatosis versus massive liver metastases2 (Table 91-4; see also “Ascitic Fluid Analysis”). Congestive heart failure accounts for less than 5% of cases of ascites (see Chapter 83). Cardiac ascites is characterized by a high-albumin gradient, high ascitic fluid protein con-

centration, and normal blood hematocrit value.27 The gradient may narrow with diuresis, in contrast to cirrhosis. Patients with cardiac ascites often have alcoholic cardiomyopathy, with cardiomegaly on a chest radiograph and fourchamber enlargement of the heart on an echocardiogram. Clinically, heart failure may mimic cirrhosis, including the presence of small nonbleeding esophageal varices and hepatic encephalopathy.42 Ascites in the setting of cirrhosis is characterized by a high albumin gradient, as in cardiac ascites, but a low protein concentration, and patients with cirrhosis and ascites have a lower mean blood hematocrit value of 32%.27 Serum pro-brain-type natriuretic peptide also can be useful in distinguishing cardiac ascites from ascites due to cirrhosis. The median value is 6100 pg/mL in the former but only 166 pg/mL in the latter.43 In the United States, tuberculous peritonitis generally is a disease of Asian and Latin American immigrants to the West Coast, poor African Americans, and the elderly. Tuberculous peritonitis was a rare disease between 1955 and 1985, but it has increased in prevalence since then because of the acquired immunodeficiency syndrome (AIDS).44 Fifty percent of patients with tuberculous peritonitis have underlying cirrhosis (and thus, “mixed” ascites). Although most patients with liver disease are not unusually predisposed to the hepatotoxicity of antituberculosis drugs, they tolerate drug toxicity less well than do patients with a normal liver.45 Underdiagnosis can lead to unnecessary deaths from untreated tuberculosis, whereas overdiagnosis and overtreatment of suspected but unproven tuberculous peritonitis may lead to unnecessary deaths from the hepatotoxicity of isoniazid. If the clinical circumstances (e.g., fever in an immigrant from an area endemic for tuberculosis) and results of the initial ascitic fluid analysis (high lymphocyte count) suggest tuberculosis, strong consideration should be given to an urgent laparoscopy with histologic examination and culture of peritoneal biopsy specimens. If at laparoscopy the peritoneum demonstrates the typical “millet-seed” and “violin-string” appearance, antituberculosis therapy can be started immediately. Blind peritoneal biopsy may be performed in the patient without cirrhosis; however, in a patient with cirrhosis, the predictable presence of peritoneal collateral veins makes blind biopsy potentially hazardous, and laparoscopically guided biopsy is prefer­ able. Suspected tuberculous peritonitis is one of the few remaining indications for diagnostic laparoscopy. Peritoneal coccidioidomycosis can mimic tuberculous peritonitis, including its appearance at laparoscopy, and can occur in patients without AIDS.46 The high sensitivities of cytology for peritoneal carcinomatosis and ultrasound-guided biopsy for focal liver lesions have obviated the need for laparoscopy in detecting tumor, for all practical purposes.2 Pancreatic ascites, an uncommon condition, occurs in patients with clinically obvious severe acute pancreatitis or a history of chronic pancreatitis or pancreatic trauma (see Chapters 58 and 59).29 Ordering an ascitic fluid amylase level on all ascitic fluid samples is unnecessary; the test is indicated only in patients in whom pancreatitis is suspected

1527

1528

Section IX  Liver or the initial ascitic fluid is nondiagnostic (see Table 91-2). Patients with alcohol-related pancreatic ascites may also have underlying alcoholic cirrhosis. Pancreatic ascites frequently is neutrocytic and may also be complicated by bacterial infection. Patients with an ascitic fluid neutrophil count of 250 cells/mm3 (0.25 × 109/L) or greater merit empirical antibiotic coverage, at least until the cause of the elevated neutrophil count is explained. Nephrogenous ascites is a poorly understood form of ascites that develops in patients undergoing hemodialysis.47 On careful evaluation, most patients with ascites in the setting of hemodialysis are found to have another cause of ascites, usually cirrhosis from alcohol abuse or from hepatitis C. The presence of a second cause of fluid overload explains why these patients have ascites, whereas a majority of patients on dialysis do not. Although the nephrotic syndrome used to be a common cause of ascites in children, it is rare in adults.48 When it occurs in adults, a second cause of ascites usually is present, just as in nephrogenous ascites.48 The ascitic fluid is usually characterized by a low protein concentration and low SAAG and can be complicated by spontaneous bacterial peritonitis. Chlamydia (or rarely gonococcal) peritonitis should be suspected in sexually active young women with fever and neutrocytic, high-protein, low-gradient ascites and no evidence of liver disease. This infection responds rapidly to oral doxycycline and is one of the few curable causes of ascites. In some patients, pathologic accumulation of fluid develops in the peritoneal cavity as a result of leakage from a ruptured viscus (e.g., “bile ascites” from a ruptured gallbladder).16,28 The ascitic fluid analysis is critical to the preoperative diagnosis of this condition (see earlier “Ascitic Fluid Analysis,” and Fig. 91-3). Chylous ascites develops when intra-abdominal lymphatics containing chyle rupture. The older literature suggests that this form of ascites is caused by a malignancy in nearly 90% of cases.36 By contrast, cirrhosis is the cause of chylous ascites in more than 90% of the patients whom I have encountered (see Table 91-1).15,21 The high lymphatic flow and pressure are presumed to be the cause of lymphatic rupture in patients with cirrhosis. In addition, retroperitoneal surgery and radical pelvic surgery in patients with cancer can transect lymphatics and thereby lead to chylous ascites. Additional causes of ascites include ambulatory peritoneal dialysis, Budd-Chiari syndrome, myxedema, connective tissue disease, postoperative ascites, and rare entities. With the iatrogenic form of ascites associated with peritoneal dialysis, the patient is usually not under the care of a gastroenterologist. Although Budd-Chiari syndrome is regularly complicated by ascites, hepatic vein thrombosis is rare and accounts for less than 0.1% of cases of ascites (see Chapter 83). Ascites in patients with myxedema appears to be related to heart failure49; treatment of the hypothyroidism cures the fluid retention. Serositis with development of ascites may complicate systemic lupus erythematosus (see Chapter 35).7 Ascites after abdominal surgery (often after cholecystectomy in the setting of asymptomatic gallstones and abnormal liver biochemical test results) is a common mode of presentation of previously undiagnosed cirrhosis.50 Resection of hepatocellular carcinoma in the setting of cirrhosis regularly leads to hepatic decompensation, which all too often starts a downward spiral ending in death.51 Aggressive hormone administration to induce ovulation can lead to ascites from “ovarian hyperstimulation syn-

drome.”52 Other rare causes of ascites include the POEMS syndrome (polyneuropathy, organomegaly, endocrinopathy, M component, and skin changes) and hemophagocytic syndrome.53,54 The latter is a rare syndrome that usually occurs in patients with leukemia or lymphoma and can masquerade as decompensated cirrhosis.54 Ascites that recurs or does not resolve after liver transplantation appears to be due to relative hepatic venous outflow obstruction or hepatitis C but frequently is enigmatic.55,56

COMPLICATIONS ASCITIC FLUID INFECTION, INCLUDING SPONTANEOUS BACTERIAL PERITONITIS

Ascitic fluid infection can be classified into five categories based on ascitic culture results, PMN count, and presence or absence of a surgical source of infection (Table 91-5). An abdominal paracentesis must be performed and ascitic fluid must be analyzed before a confident diagnosis of ascitic fluid infection can be made. A “clinical diagnosis” of infected ascitic fluid without a paracentesis is inadequate.

Classification

Of the three subtypes of spontaneous ascitic fluid infection, the prototype is spontaneous bacterial peritonitis. The diagnosis of spontaneous bacterial peritonitis is made when there is a positive ascitic fluid culture and an elevated ascitic fluid absolute PMN count (i.e., at least 250 cells/mm3 [0.25 × 109/L]) without evidence of an intra-abdominal surgically treatable source of infection.9 When Correia and Conn coined the term “spontaneous bacterial peritonitis” in 1975, their goal was to distinguish this form of infection from surgical peritonitis,57 an important distinction. Therefore, although many patients with spontaneous bacterial peritonitis have a focus of infection (e.g., urinary tract infection or pneumonia), the diagnosis of spontaneous bacterial peritonitis is still appropriate unless the focus requires surgical intervention (e.g., a ruptured viscus). I have not encountered a convincing case of polymicrobial spontaneous bacterial peritonitis; all of the patients presumed to have spontaneous bacterial peritonitis in whom ascitic fluid cultures initially grew more than one organism eventually were found to have surgical peritonitis or an erroneous culture result (e.g., a pathogen plus a contaminant or two colony morphologies of one species of bacteria). The criteria for a diagnosis of monomicrobial nonneutrocytic bacterascites (MNB) include (1) a positive ascitic fluid culture for a single organism, (2) an ascitic fluid PMN count lower than 250 cells/mm3 (0.25 × 109/L), and (3) no evidence of an intra-abdominal surgically treatable source of infection.58 In the older literature, MNB was either grouped with spontaneous bacterial peritonitis or labeled “asymptomatic bacterascites.” Because many patients with bacterascites have symptoms, the modifier “asymptomatic” seems inappropriate. Culture-negative neutrocytic ascites (CNNA) is diagnosed when (1) the ascitic fluid culture grows no bacteria, (2) the Table 91-5  Classification of Ascitic Fluid Infection Culture-negative neutrocytic ascites Monomicrobial non-neutrocytic bacterascites Polymicrobial bacterascites (needle perforation of the bowel) Secondary bacterial peritonitis Spontaneous bacterial peritonitis

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis ascitic fluid PMN count is 250 cells/mm3 (0.25 × 109/L) or greater, (3) no antibiotics have been given (not even a single dose), and (4) no other explanation for an elevated ascitic PMN count (e.g., hemorrhage into ascites, peritoneal carcinomatosis, tuberculosis, or pancreatitis) can be identified.59 This variant of ascitic fluid infection seldom is diagnosed when sensitive culture methods are used.23 Secondary bacterial peritonitis is diagnosed when (1) the ascitic fluid culture is positive (usually for multiple organisms), (2) the PMN count is 250 cells/mm3 (0.25 × 109/L) or greater, and (3) an intra-abdominal surgically treatable primary source of infection (e.g., perforated intestine, perinephric abscess) has been identified.28 The importance of distinguishing this variant from spontaneous bacterial peritonitis is that secondary peritonitis usually requires emergency surgical intervention (see also Chapter 37). Polymicrobial bacterascites is diagnosed when (1) multiple organisms are seen on Gram stain or cultured from the ascitic fluid and (2) the PMN count is lower than 250 cells/ mm3 (0.25 × 109/L).60 This diagnosis should be suspected when the paracentesis is traumatic or unusually difficult because of ileus or when stool or air is aspirated into the paracentesis syringe. Polymicrobial bacterascites is essentially diagnostic of intestinal perforation by the paracentesis needle.

Bowel flora ? Altered permeability Bacteria in mesenteric lymph nodes

Bacteria in abdominal lymphatics

Bacteria in thoracic duct lymph

Pathogenesis

Since the 1990s, the elusive cause of spontaneous bacterial peritonitis has become clearer, and the pathogenesis of spontaneous forms of ascitic fluid infection has been partially elucidated (Fig. 91-4). The body of currently available evidence suggests that the spontaneous forms of ascitic fluid infection are the result of overgrowth of a specific organism in the intestine, “translocation” of that microbe from the intestine to mesenteric lymph nodes, and resulting spontaneous bacteremia and subsequent colonization of susceptible ascitic fluid61-62 (see Fig. 91-4). When bacteria enter the fluid in the abdomen, by whatever route, a battle ensues between the virulence factors of the organism and the immune defenses of the host.63 The ascitic fluid protein concentration does not change with development of spontaneous infection.25 Low-protein ascitic fluid (e.g., protein content less than 1 g/dL [10 g/L]) is particularly susceptible to spontaneous bacterial peritonitis.26 The endogenous antimicrobial (opsonic) activity of human

Urinary tract infection

Respiratory tract infection Complement deficiency

Reticuloendothelial system dysfunction Bacteremia

? Lymphatic rupture

Bacteria in hepatic lymph

Clinical Setting

The spontaneous variants of ascitic fluid infection—spontaneous bacterial peritonitis, CNNA, and MNB—occur almost exclusively in the setting of severe liver disease. The liver disease usually is chronic (cirrhosis), but may be acute (fulminant hepatic failure) or subacute (alcoholic hepatitis). Cirrhosis of all causes can be complicated by spontaneous ascitic fluid infection. Spontaneous infection of noncirrhotic ascites is rare enough to be the subject of case reports. Essentially all patients with spontaneous bacterial peritonitis have an elevated serum bilirubin level and abnormal prothrombin time due to advanced cirrhosis.9 Ascites appears to be a prerequisite for the development of spontaneous bacterial peritonitis. The peritonitis is unlikely to precede the development of ascites. Usually, the infection develops when the volume of ascites is at its maximum. Secondary bacterial peritonitis and polymicrobial bacterascites can develop with ascites of any type. The only prerequisite, in addition to the presence of ascites, is an intra-abdominal surgical source of infection.28 Such an infection can result from penetration of a needle into the bowel during attempted paracentesis.60

Altered flora

Bacterascites

Poor opsonic activity

Moderate opsonic activity

Good opsonic activity

Sterile non-neutrocytic ascites Figure 91-4.  Proposed pathogenesis of spontaneous ascitic fluid infection. CNNA, culture-negative neutrocytic ascites; SBP, spontaneous bacterial peritonitis. SBP

CNNA

ascitic fluid correlates directly with the protein concentration of the fluid.62 Patients with deficient ascitic fluid opsonic activity are predisposed to spontaneous bacterial peritonitis.64 Patients with detectable ascitic fluid opsonic activity appear to be protected from spontaneous bacterial peritonitis unless they are exposed to a particularly virulent organism (e.g., Salmonella).63,64 Studies in both patients and animals with cirrhosis demonstrate that MNB is common.58,65 Pieces of bacterial DNA are commonly present in serum and ascitic fluid of patients with cirrhosis.66 In both humans and rats, most episodes of bacterascites resolve without antibiotic treatment.58,65 The fluid frequently becomes sterile without an increase in ascitic PMNs. Apparently, the host’s defense mechanisms are able to eradicate the invading bacteria on most occasions. Uncontrolled infection probably develops only when the defenses are weak or the organism is virulent (see Fig. 91-4). Bacterascites probably is more common than spontaneous bacterial peritonitis. Conceivably, ascitic fluid in the setting of cirrhosis is colonized regularly by bacteria, and almost just as regularly, the colonization resolves. The entry of PMNs into the fluid probably signals failure of the peri-

1529

1530

Section IX  Liver toneal macrophages to control the infection.67 A majority of episodes of MNB appear to resolve in cirrhotic rats and humans, whereas untreated spontaneous bacterial peritonitis is frequently fatal. In summary, MNB probably represents an early stage of ascitic fluid infection, which can resolve or progress to CNNA or to spontaneous bacterial peritonitis. Most episodes of CNNA are diagnosed by insensitive culture methods for which numbers of bacteria are insufficient to reach the threshold of detectability.23 Inoculation of ascitic fluid into blood culture bottles can lead to detection of a single organism in the cultured aliquot of fluid, whereas the older method of culture by inoculation of agar plates and broth probably requires at least 100 organisms/mL.23 Even when optimal culture methods are used, however, a small percentage of specimens of neutrocytic ascitic fluid grow no bacteria. A study of rapid sequential paracenteses (before the initiation of antibiotic treatment) in patients with CNNA demonstrated that, in most cases, the PMN count dropped spontaneously and the culture results remained negative in the second specimen.68 When sensitive culture techniques are used, CNNA probably results from (1) previous antibiotic treatment (even one dose), (2) an inadequate volume of fluid inoculated, or (3) spontaneously resolving spontaneous bacterial peritonitis in which the paracentesis is performed after all bacteria have been killed by host defenses but before the PMN count has normalized. The pathogenesis of secondary bacterial peritonitis is more straightforward than that of spontaneous bacterial peritonitis. When the intestine perforates, billions of bacteria flood into the ascitic fluid. In the absence of a frank perforation, bacteria may cross inflamed tissue planes and enter the fluid. The pathogenesis of polymicrobial bacterascites is also obvious.60 A paracentesis needle enters the bowel, and the bowel contents are released into the ascites.

approximately 10% of patients with ascites were infected at the time of hospital admission; of the subgroup of patients with cirrhosis, about 27% were infected.9 At present, because of measures to prevent spontaneous bacterial peritonitis, the frequency has dropped significantly (see later). Of patients with culture-positive ascitic fluid, about two thirds have neutrocytic ascitic fluid (spontaneous bacterial peritonitis), and one third have MNB.58 The frequency of CNNA depends largely on the culture technique (see earlier). Polymicrobial bacterascites occurs in only 1 in 1000 paracenteses. Secondary bacterial peritonitis is found in only 0% to 2% of patients with ascites at the time of hospital admission.9,28

Symptoms and Signs

Risk Factors

Bacteriology

Escherichia coli, streptococci (mostly pneumococci), and Klebsiella cause most episodes of spontaneous bacterial peritonitis and MNB in patients who are not receiving selective intestinal decontamination (Table 91-7; see later); CNNA is, by definition, culture-negative and polymicrobial bacterascites is, by definition, polymicrobial. The most apparent difference between the spontaneous forms of ascitic fluid infection and the secondary forms (secondary peritonitis and polymicrobial bacterascites) is that the former always are monomicrobial and the latter usually are polymicrobial. Although older papers reported that anaerobic bacteria were present in approximately 6% of cases of spontaneous bacterial peritonitis, the detection of anaerobes probably reflected unrecognized cases of secondary bacterial peritonitis. In more recent series, anaerobes have been found in approximately 1% of cases of spontaneous bacterial peritonitis and MNB.23,58 Selective intestinal decontamination causes a change in the bacteria isolated from patients in whom an ascitic infection develops. Gram-positive organisms are frequently cultured from the ascitic fluid of these patients (see Table 91-7).69

Although 87% of patients with spontaneous bacterial peritonitis are symptomatic at the time the infection is diagnosed, the symptoms and signs of infection are often subtle, such as a slight change in mental status.58 Without prompt paracentesis, the diagnosis and treatment of infected ascites may be delayed, often resulting in the death of the patient. The symptoms and signs manifested in all five variants of ascitic fluid infection are listed in Table 91-6.

Patients with cirrhosis are unusually predisposed to bacterial infection because of multiple defects in immune defense. The concept that cirrhosis is a form of acquired immunodeficiency (in the generic sense) is rather new. In a prospective study, a bacterial infection occurred in 34% of 405 patients with cirrhosis at the time of admission to the hospital or during the hospitalization.70 Low ascitic fluid total protein concentrations, as well as the phagocytic (both motile and stationary) dysfunction associated with cirrhosis, are risk factors for bacterial infection. Paracentesis itself has been proposed as a risk factor for ascitic fluid infection. This theoretical risk has not been substantiated in prospective studies of paracentesis-related

Frequency

Since the 1980s, routine paracenteses at the time of hospitalization in patients with ascites have provided data regarding the frequency of ascitic fluid infection. In the 1980s,

Table 91-6  Symptoms and Signs of Ascitic Fluid Infection Frequency (%)* SYMPTOM OR SIGN Fever Abdominal pain Abdominal tenderness Rebound tenderness Altered mental status

SBP

Bacterascites

CNNA

Secondary Peritonitis

Polymicrobial Bacterascites

68 49 39 10 54

57 32 32 5 50

50 72 44 0 61

33 67 50 17 33

10 10 10 0 0

*Data presented as % of the total number of patients in that group. CNNA, culture-negative neutrocytic ascites; SBP, spontaneous bacterial peritonitis. Data from references 28, 58-60.

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis Table 91-7  Pathogens in Ascitic Fluid Infection Frequency (%)* ORGANISM Monomicrobial Escherichia coli Klebsiella pneumoniae Streptococcus pneumoniae Streptococcus viridans Staphylococcus aureus Miscellaneous gram-negative Miscellaneous gram-positive Polymicrobial

SBP 37 17 12 9 0 10 14 1

Monomicrobial Non-Neutrocytic Bacterascites 27 11 9 2 7 14 30 0

Secondary Bacterial Peritonitis

SBP with SID

20 7 0 0 13 7 0 53

0 7 29 0 0 7 50 7

*Data reported as % of total patients in that group. SBP, spontaneous bacterial peritonitis; SID, selective intestinal decontamination. Data from references 23, 58.

complications.10 Spontaneous bacterial peritonitis is statistically more likely to be diagnosed on the first paracentesis than on subsequent taps.10 Needle-induced ascitic fluid infections do not occur unless the bowel is penetrated by the paracentesis needle10,60; fortunately, this occurs in only 1 in 1000 taps. One would expect bacteria of the skin flora such as Staphylococcus aureus to be isolated more frequently if poor paracentesis technique were the cause of many cases of spontaneous bacterial peritonitis; yet skin flora microorganisms are seldom isolated from ascitic fluid when sterile technique is used.23 Iatrogenic peritonitis is most likely to occur when the paracentesis needle enters the bowel during a difficult paracentesis. Gastrointestinal hemorrhage is an under-recognized risk factor for the development of spontaneous bacteremia and spontaneous bacterial peritonitis. The cumulative pro­ bability of infection during a single hospitalization for bleeding is approximately 40%.71 The risk appears to peak 48 hours after the onset of hemorrhage. The high risk of infection probably is mediated by a shock-induced increase in the translocation of bacteria from the intestine to extraintestinal sites. Urinary tract infections also constitute an under-recognized risk factor for spontaneous bacterial peritonitis.72

Diagnosis

Timely diagnosis of ascitic fluid infection requires a high index of suspicion and a low threshold for performing a paracentesis. Clinical deterioration, especially fever or abdominal pain, in a patient with ascites should raise the suspicion of infection and prompt a paracentesis. If the ascitic fluid PMN count is elevated, the working diagnosis is ascitic fluid infection until proved otherwise. Although peritoneal carcinomatosis, pancreatitis, hemorrhage into ascites, and tuberculosis can lead to an elevated ascitic fluid PMN count, most cases of neutrocytic ascites are caused by bacterial infection. A predominance of PMNs in the WBC differential count lends further support for the diagnosis of infection. In patients with peritoneal carcinomatosis, pancreatitis, and tuberculosis, a predominance of PMNs in the ascites would be an uncommon finding. An elevated absolute ascitic fluid PMN count with a predominance of neutrophils in a clinical setting compatible with infection should prompt empirical antibiotic therapy (Table 91-8; see later). Although spontaneous bacterial peritonitis is approximately six times as common as surgical peritonitis in a patient with ascites, secondary peritonitis should be con-

Table 91-8  Indications for Empirical Antibiotic Therapy of Suspected Spontaneous Ascitic Fluid Infection Ascitic fluid neutrophil count ≥ 250/mm3 (0.25 × 109/L) or positive “dipstick” test Convincing symptoms or signs of infection

sidered in any patient with neutrocytic ascites (see also Chapter 37). Clinical symptoms and signs do not distinguish patients with secondary peritonitis from those with spontaneous bacterial peritonitis (see Fig. 91-3).28 Even with free perforation of the colon into ascitic fluid, a classic surgical abdomen does not develop. Peritoneal signs require contact of inflamed visceral and parietal peritoneal surfaces, and such contact does not occur when there is a large volume of fluid separating these surfaces. Intestinal perforation can be suspected and pursued if a specimen of ascites is neutrocytic and meets two of the following three criteria (see Fig. 91-3): (1) total protein greater than 1 g/dL (10 g/L), (2) glucose less than 50 mg/dL (2.8 mmol/L), and (3) LDH greater than the upper limit of normal for serum.28 In the setting of a perforated viscus, cultures of ascitic fluid nearly always disclose multiple organisms, except in gallbladder rupture, which is usually monomicrobial.16 Brown ascitic fluid with a bilirubin concentration that is greater than 6 mg/dL (102 µmol/L) and greater than the serum level is indicative of biliary or proximal small intestinal perforation into ascites.16 An ascitic fluid amylase level that is greater than five-fold that of the serum level also may be indicative of intestinal rupture (but not gallbladder rupture) with the release of luminal amylase.28,29 The initial ascitic fluid analysis is helpful in delineating which patients are likely to have a ruptured viscus (see Fig. 91-3). Within minutes of the detection of neutrocytic ascitic fluid, these patients should undergo imaging studies to confirm and localize the site of rupture. Plain and upright abdominal films and water-soluble contrast studies of the upper and lower intestines or abdominal computed tomog­ raphy should be obtained. If perforation is documented, emergency surgical intervention is the next step. Timing is crucial; after septic shock occurs, death is nearly certain. Antibiotic therapy without surgical intervention in the treatment of a ruptured viscus is predictably unsuccessful. In contrast to patients with peritonitis resulting from perforation of a viscus, patients with secondary peritonitis

1531

1532

Section IX  Liver unrelated to perforation tend not to have a diagnostic initial ascitic fluid analysis.28 The need to make the diagnosis of secondary peritonitis in patients without free perforation is less urgent, and there may be time to evaluate the response of the ascitic PMN count and fluid culture to treatment with antibiotics. It is best to repeat the paracentesis to assess the response to treatment after 48 hours of therapy; by 48 hours, the ascitic PMN count will be lower than the pretreatment value and the ascitic culture will be negative in essentially every patient with spontaneous bacterial peritonitis who has been treated with an appropriate antibiotic.28 Before 48 hours of treatment, the ascitic PMN count may rise to a value higher than baseline in either spontaneous bacterial peritonitis or secondary peritonitis.28 The culture remains positive in secondary peritonitis and becomes rapidly negative in spontaneous bacterial peritonitis (see Fig. 91-3).28 Whereas antibiotics alone cannot control secondary peri­ tonitis, medical therapy cures spontaneous bacterial peritonitis rapidly.28

Treatment

Patients with an ascitic fluid PMN count of 250 cells/mm3 (0.25 × 109/L) or greater and a clinical scenario compatible with ascitic fluid infection should receive empirical antibiotic treatment (Table 91-9; see also Table 91-8).9,73 Patients with hemorrhage into the ascitic fluid, peritoneal carcinomatosis, pancreatic ascites, or tuberculous peritonitis may have an elevated ascitic PMN count that is unrelated to spontaneous bacterial peritonitis and usually do not require empirical antibiotic treatment. If they do receive antibiotics, the ascitic PMN count usually fluctuates randomly, in contrast to the dramatic reduction in PMN count typical of spontaneous bacterial peritonitis. If the clinical picture is unclear initially, the physician should err on the side of

Table 91-9  Treatment of Subtypes of Ascitic Fluid Infection DIAGNOSIS

TREATMENT

Spontaneous bacterial peritonitis

Five days of intravenous antibiotic to which the organism is highly susceptible (e.g., cefotaxime 2 g every 8 hours empirically followed by more specific therapy after susceptibility results are available) Five days of intravenous antibiotic to which the organism is highly susceptible, if the patient is symptomatic or persistently culture-positive; not all patients with bacterascites require treatment Five days of intravenous third-generation cephalosporin (e.g., cefotaxime 2 g every 8 hours) Surgical intervention plus approximately 2 weeks of intravenous cephalosporin (e.g., cefotaxime 2 g every 8 hours) plus an antianaerobic drug such as metronidazole* Intravenous third-generation cephalosporin (e.g., cefotaxime 2 g every 8 hours) plus an antianaerobic drug such as metronidazole* Duration is determined by clinical response and serial ascitic fluid PMN counts and cultures

Monomicrobial non-neutrocytic bacterascites Culture-negative neutrocytic ascites Secondary bacterial peritonitis

Polymicrobial bacterascites

*Dose of intravenous metronidazole is 15 mg/kg × 1, then 7.5 mg/kg every 6 hours. PMN, polymorphonuclear neutrophil.

antibiotic treatment (with a non-nephrotoxic antibiotic). If ascitic fluid cultures are negative, the antibiotic can be stopped after 48 hours. In patients with uninfected neutrocytic ascitic fluid (except those with hemorrhage), lymphocytes usually predominate in the ascitic fluid differential count, in contrast to those with spontaneous bacterial peritonitis, in whom PMNs predominate. In patients with bloody ascitic fluid, a “corrected” PMN count should be calculated (as discussed earlier). Antibiotic therapy is not necessary for patients with bloody fluid unless the corrected ascitic fluid PMN count is 250 cells/mm3 (0.25 × 109/L) or greater. The decision to begin empirical antibiotic treatment in patients with bacterascites must be individualized. Many episodes resolve without treatment58; however, the hospital mortality rate of 32% in patients with MNB is attributable, at least, in part, to infection.58 Therefore, treatment appears to be warranted in many patients. By definition, the ascitic PMN count is lower than 250 cells/mm3 (0.25 × 109/L) in this variant of ascitic fluid infection, and the PMN count cannot be the only parameter on which to base the decision about empirical therapy. Most patients with MNB in whom the colonization does not resolve progress to spontaneous bacterial peritonitis and have symptoms or signs of infection at the time of the paracentesis that documents bacterascites.58 Therefore, patients with cirrhosis and ascites who have convincing symptoms or signs of infection should receive treatment regardless of the ascitic fluid PMN count. Empirical treatment can be discontinued after only two to three days if the culture demonstrates no growth. Asymptomatic patients may not need treatment.58 The paracentesis should be repeated for cell count and culture in patients without clinical evidence of infection, as soon as it is known that the initial culture result is positive. If the PMN count has risen to at least 250/mm3 (0.25 × 109/L) or if symptoms or signs of infection have developed, treatment should be started. Culture results usually are negative in patients without a rise in the ascitic fluid PMN count on repeat paracentesis and without clinical evidence of infection, and these persons do not require treatment58 because colonization has been eradicated by host immune defenses. The physician will not know initially that the ascitic fluid culture is destined to be negative in a patient with CNNA; therefore, empirical antibiotic treatment should be started. When the preliminary culture demonstrates no growth, it is helpful to repeat the paracentesis after 48 hours of therapy to assess the response of the PMN count to antibiotics. A dramatic decline in PMN count (always below the baseline pretreatment value and frequently a reduction of more than 80%) confirms a response to treatment.28 In such cases, a few more days of therapy is probably warranted. A stable ascitic fluid PMN count, especially with a predominance of lymphocytes and monocytes, suggests a nonbacterial (or mycobacterial) cause of ascitic fluid neutrocytosis, and the fluid should be sent for cytologic examination and mycobacterial culture. Because a negative culture result may be due to insensitive culture techniques, the prevalence of CNNA in a hospital that still uses conventional methods of culture can be reduced by convincing the microbiology laboratory to accept and process ascitic fluid submitted in blood culture bottles.23 Gram stain of the ascitic fluid is most helpful in detecting secondary peritonitis, in which multiple different bacterial forms are seen, but is of little value in guiding the choice of empirical antibiotic treatment for spontaneous ascitic infections. I have found that use of the Gram stain did not help narrow the antibiotic coverage in even 1 patient of approximately 500 with spontaneous bacterial peritonitis. Only

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis approximately 10% of Gram stains demonstrate organisms in spontaneous bacterial peritonitis.23 If a Gram stain indicates secondary peritonitis, coverage of anaerobic flora, in addition to coverage of aerobic and facultative anaerobic flora, is required, as is an emergency search for the source of the infection (see Fig. 91-3; Table 91-9).28 Therefore, a positive Gram stain may lead to broader antibiotic coverage, rather than narrower coverage. Choosing narrow coverage (e.g., penicillin alone) based on a misinterpretation of the significance of the results of the Gram stain may lead to the patient’s death from uncontrolled infection before it becomes apparent that the isolated organism is resistant to the chosen antibiotic. Until the results of susceptibility testing are available, relatively broad-spectrum antibiotic therapy is warranted in patients with suspected ascitic fluid infection. After sensitivities are known, the spectrum of coverage can usually be narrowed. The antibiotics that have been recommended for empirical treatment have changed over the years. In the late 1970s, the combination of ampicillin and gentamicin was promoted, but this recommendation was not based on susceptibility testing or efficacy data. Subsequently, gentamicin was shown to have an unpredictable volume of distribution in patients with ascites, and the serum creatinine level (and even the creatinine clearance) was found to be a poor index of the glomerular filtration rate in patients with ascites.74 Therefore, determining the appropriate loading and maintenance doses of gentamicin for this patient population is difficult, and no evidence-based guidelines are available for the prescribing physician to follow. In my experience, even if high serum levels are avoided, nephrotoxicity still develops in most cirrhotic patients with ascites who receive an aminoglycoside. One study has documented an adjusted odds ratio of 4.0 for aminoglycosides as a risk factor for renal dysfunction in patients with cirrhosis.75 Evidence that newer aminoglycosides are less nephrotoxic than gentamicin is lacking. Several antibiotics are now available for the treatment of ascitic fluid infection. Cefotaxime, a third-generation cephalosporin, has been shown in a controlled trial to be superior to ampicillin plus tobramycin for the treatment of spontaneous bacterial peritonitis.76 Fully 98% of causative organisms were susceptible to cefotaxime, which did not result in superinfection or nephrotoxicity.76 Cefotaxime or a similar third-generation cephalosporin appears to be the treatment of choice for suspected spontaneous bacterial peritonitis.9 Anaerobic coverage is not needed, nor is coverage for Pseudomonas or Staphylococcus.23 Cefotaxime, 2 g intravenously every eight hours, has been shown to result in excellent ascitic fluid levels (20-fold killing power after one dose).77 In patients with a serum creatinine level greater than 3 mg/dL, the dosing interval may be extended to 12 hours.77 Neither a loading dose nor an intraperitoneal dose appears to be necessary or appropriate. The clinician should, however, write “first dose STAT” when ordering treatment, to avoid a delay in administration of the life-saving agent. Other Intravenous Antibiotics Amoxicillin-clavulanic acid has been shown to be as effective as cefotaxime in a randomized trial but is not available in a parenteral formulation in the United States.78 Other antibiotics have been recommended as well but have been less well studied than has cefotaxime. Some newer drugs have been used to treat spontaneous bacterial peritonitis (without any data on antibiotic penetration into the ascitic fluid) on the basis of their spectrum of coverage and formulary constraints. Infection with organisms that are resistant

to the empirical antibiotic or use of drugs that do not enter the ascitic fluid in high enough concentrations to kill the bacteria may lead to the patient’s death. Intravenous Albumin Renal impairment occurs in 33% of episodes of spontaneous bacterial peritonitis.79 Spontaneous bacterial peritonitis leads to increased intraperitoneal nitric oxide production, which in turn further increases systemic vasodilatation and promotes renal failure (see Chapter 92).80 Intravenous albumin (1.5 g/kg of body weight at the time the infection is detected and 1.0 g/kg on day three) can increase intravascular volume and, in combination with cefotaxime, has been shown in a large randomized trial to reduce the risk of renal failure and improve survival compared with cefotaxime without albumin.81 Albumin appears to be effective by decreasing vasodilatation.82 A confirmatory randomized trial is needed. Because of the survival advantage, however, the use of intravenous albumin as an adjunct to antibiotic treatment has been recommended.83 Oral Antibiotic Treatment Oral ofloxacin has been reported in a controlled trial to be as effective as parenteral cefotaxime in the treatment of spontaneous bacterial peritonitis in patients who do not have vomiting, shock, bleeding, or renal failure.84 The dose studied was 400 mg twice daily.84 Another study has demonstrated the efficacy of intravenous ciprofloxacin, 200 mg every 12 hours for 2 days, followed by oral ciprofloxacin, 500 mg every 12 hours for 5 days.85 Because of the possibility of fluoroquinolone resistance in patients receiving fluoroquinolones to prevent spontaneous bacterial peritonitis (see later), however, the empirical use of a fluoroquinolone to treat suspected spontaneous bacterial peritonitis should be avoided.86 Fortunately, bacterial isolates from patients with spontaneous bacterial peritonitis who were receiving fluoroquinolones for prophylaxis of this disorder remain susceptible to cefotaxime.73 Narrowing the Spectrum of Coverage After the results of susceptibility testing are available, an antibiotic with a narrower spectrum of activity usually can be substituted for the broad-spectrum drug (e.g., pneumococci will usually be sensitive to penicillin, and most E. coli species will usually be sensitive to ampicillin). Duration of Treatment Infectious disease subspecialists generally recommend 10 to 14 days of antibiotic therapy for life-threatening infections; however, no data are available to support this duration of treatment in spontaneous ascitic fluid infections. The ascitic fluid culture becomes sterile after one dose of cefotaxime in 86% of patients.28 After 48 hours of therapy, the ascitic fluid PMN count is always less than the pretreatment value in patients with a spontaneous ascitic fluid infection treated with appropriate antibiotics; frequently, an 80% reduction is observed at 48 hours.28 A randomized, controlled trial involving 100 patients has demonstrated that 5 days of treatment is as efficacious as 10 days in patients with spontaneous bacterial peritonitis or CNNA.87 I have been treating spontaneous bacterial peritonitis and CNNA for five days since the late 1980s, with excellent results. The average duration of oral ofloxacin treatment was eight days in the only published trial.84 Follow-up Paracentesis On the basis of a large database of repeat paracenteses during and after the treatment of spontaneous bacterial peri-

1533

1534

Section IX  Liver tonitis,28 a follow-up paracentesis does not appear to be needed if the setting (advanced cirrhosis with symptoms and signs of infection), bacterial isolate (monomicrobial with a typical organism), and response to treatment (dramatic reduction in symptoms and signs of infection) are typical.28 Paracentesis should be repeated after 48 hours of treatment if the course is atypical.28 Treatment of Ascitic Fluid Infection Other than Spontaneous Bacterial Peritonitis Because of the predictable presence of anaerobes, patients with suspected secondary peritonitis require empirical antibiotic coverage that is broader in spectrum than that used for spontaneous bacterial peritonitis. They also require an emergency evaluation to assess the need for surgical intervention (see earlier discussion, and Table 91-8 and Fig. 91-3). Cefotaxime plus metronidazole appears to provide excellent initial empirical therapy of suspected secondary peritonitis.28 Polymicrobial bacterascites (from needle perforation of the bowel) is tolerated relatively well. Peritonitis developed in only 1 in 10 patients with a needle perforation of the intestine with spillage of intestinal contents into ascitic fluid in the one relevant study.60 The single episode of paracentesis-related peritonitis was not fatal. Patients with low-protein ascitic fluid appear to be at most risk for development of a PMN response and clinical peritonitis related to needle perforation of the intestine.60 Most of the patients with a higher ascitic protein concentration (e.g., greater than 1 g/dL [10 g/L]) did not receive antibiotics, yet did well. Many physicians, however, probably would feel uncomfortable in withholding antibiotic treatment if needle perforation is suspected. If a decision to treat is made, anaerobic coverage should be included (e.g., cefotaxime plus metronidazole; see Table 91-9). Whether or not treatment is begun, a follow-up paracentesis is helpful (if it can be performed safely) to monitor the ascitic fluid PMN count and culture results. If a decision was made to defer antibiotic treatment initially and the number of organisms in the ascitic fluid does not decrease or the PMN count rises in the second specimen, antibiotic treatment should be initiated (see Table 91-9).

Prognosis

In the past, 48% to 95% of patients with a spontaneous ascitic fluid infection died during the hospitalization in which the diagnosis was made, despite antibiotic treatment.9,20 The most recent series report the lowest mortality rates (less than 5% if antibiotics are administered in a timely fashion), probably because of earlier detection and treatment of infection, as well as the avoidance of nephrotoxic antibiotics.87 The trial in which cefotaxime plus albumin was studied reported the lowest hospitalization mortality rate yet—10%.81 Even now, however, some patients are cured of their infection but die of liver failure or gastrointestinal bleeding because of the severity of the underlying liver disease. In fact, spontaneous ascitic fluid infection is a good marker of end-stage liver disease and has been proposed as an indication for liver transplantation in a patient who is otherwise a candidate. To maximize survival, it is important that paracentesis is performed in all patients with ascites at the time of hospitalization, so that infection can be detected and treated promptly. The ascitic fluid cell count should be reviewed as soon as the results are available (approximately 60 minutes), and appropriate treatment should be instituted if indicated. The first dose of antibiotic should be given immediately. Because the “dipstick” test results are available in

90 to 120 seconds, this new tool may speed treatment of spontaneous bacterial peritonitis and improve survival.18 Paracentesis should be repeated during the hospitalization if any manifestation of clinical deterioration develops, including abdominal pain, fever, change in mental status, renal failure, acidosis, peripheral leukocytosis, or gastrointestinal bleeding. If the physician waits to perform a paracentesis until convincing symptoms and signs of infection have developed, the infection is likely to be advanced by the time the diagnosis is made. No survivors of spontaneous bacterial peritonitis have been reported when the diagnosis was made after the serum creatinine level had risen above 4 mg/dL (350 µmol/L) or after shock had developed. Without surgical intervention, the mortality rate for secondary peritonitis in hospitalized patients with ascites approaches 100%. When secondary peritonitis is diagnosed early and treated with emergency laparotomy, the mortality rate is approximately 50%.28

Prevention

The identification of risk factors for spontaneous bacterial peritonitis (including an ascitic fluid protein concentration less than 1.0 g/dL, variceal hemorrhage, and previous episode of spontaneous bacterial peritonitis) has led to controlled trials of prophylactic antibiotics.26,88-90 Norfloxacin, 400 mg per day orally, has been reported to reduce the risk of spontaneous bacterial peritonitis in inpatients with lowprotein ascites and those with previous spontaneous bacterial peritonitis.88,89 Norfloxacin, 400 mg orally twice daily for seven days, helps prevent infection in patients with variceal hemorrhage90 and is cost-effective in preventing recurrent spontaneous bacterial peritonitis.91 More recently, intravenous ceftriaxone 1 g daily for seven days was found to be even more effective than norfloxacin in the setting of gastrointestinal bleeding; this regimen allows administration of antibiotics to patients who are vomiting blood.92 Oral antibiotics select for resistant organisms in the intestinal flora in patients, and in animals these organisms can then cause spontaneous ascitic fluid infection.86,93 Despite this concern, two randomized trials of primary prevention of ascitic fluid infection with prophylactic norfloxacin or ciprofloxacin have demonstrated a survival advantage for the antibiotic-treated patients (Table 91-10).94-96 Trimethoprim-sulfamethoxazole has also been shown to prevent spontaneous bacterial peritonitis in an animal model and in patients; in animals survival was increased.97,98 The recommended dose for patients is one double-strength tablet daily.98 Use of parenteral antibiotics to prevent endoscopic sclerotherapy–related or band ligation–related infections in nonbleeding patients does not appear to be warranted, as indicated by a controlled trial.99 Active bleeding, not endoscopic treatment, appears to be the risk factor for ascitic fluid infection. On the other hand, bacterial infection is associated with failure to control variceal hemorrhage.100 This observation provides additional incentive to try to prevent, detect, and treat infections aggressively in this setting to minimize mortality related not only to infection, but also to hemorrhage.

CELLULITIS

Cellulitis of the lower extremities or abdominal wall is a common cause of soft tissue infection in obese patients with edema. One study has documented a 19% cumulative probability of cellulitis during hospitalization of patients with cirrhosis and ascites, compared with only a 4% likelihood of spontaneous bacterial peritonitis.101 Risk factors for cellulitis included obesity (which is increasing in frequency in

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis Table 91-10  Prevention of Spontaneous Bacterial Peritonitis (SBP) INDICATION

DRUG

RESULTS

Prior SBP

Norfloxacin 400 mg orally once daily until death or liver transplantation Norfloxacin 400 mg orally twice daily × 7 days Ceftriaxone 1 g intravenously/day × 7 days

66% Reduction in recurrence

Cirrhosis with gastrointestinal hemorrhage Cirrhosis with ascitic fluid total protein <1.5 g/dL and either Child-Turcotte-Pugh score ≥9 and total bilirubin ≥3 mg/dL, or creatinine ≥1.2 mg/dL, or blood urea nitrogen ≥25 mg/dL, or serum sodium ≤130 mEq/L Cirrhosis with ascitic fluid total protein <1.5 g/dL

Norfloxacin 400 mg/day orally × 1 year

Ciprofloxacin 500 mg orally daily × 1 year

73% Reduction in infection 67% Reduction in infection compared with norfloxacin 89% Reduction in SBP 32% Reduction in hepatorenal syndrome 52% Increase in 3-month survival 25% Increase in 1-year survival 31% Reduction in infection 30% Improvement in survival

Data from references 89, 90, 92, 94, 95.

patients with cirrhosis), homelessness, and greater degree of edema.101 A high index of suspicion and low threshold for treatment with a first-generation cephalosporin or other antibiotic may help decrease morbidity and mortality from uncontrolled cellulitis. Some patients with ascites do not seek medical attention until they can no longer breathe or eat comfortably because of the pressure of the intra-abdominal fluid on the diaphragm. Tense ascites requires urgent therapeutic paracentesis (see later). Contrary to folklore, tense ascites can be drained without untoward hemodynamic effects.102 “Total paracentesis,” even more than 22 L, has been demonstrated to be safe.102 In the setting of tense ascites, therapeutic paracentesis improves venous return and hemodynamics; the myth of paracentesis-related hemodynamic catastrophes was based on anecdotal observations in small numbers of patients.

authors have recommended chest tube insertion and sclerosing of the pleurae with tetracycline; however, chest tubes inserted to treat hepatic hydrothorax are usually difficult to remove105; moreover, shortness of breath may recur when the tube is clamped, and fluid may leak around the insertion site of the tube. A peritoneovenous shunt (see later) can be considered when the patient with hepatic hydrothorax has large-volume ascites, but the shunt usually clots after a short time. Direct surgical repair of the diaphragmatic defect can be considered, but the patients typically are poor operative candidates. Video thoracoscopic suture of the hole in the diaphragm followed by pleurodesis has been reported to be successful in one patient.106 Sodium restriction plus use of diuretics with intermittent thoracentesis is the safest and most effective first-line therapy of hepatic hydrothorax. TIPS placement has been reported to be successful and constitutes reasonable second-line treatment.103 If the patient is a candidate for liver transplantation, proceeding with a transplantation evaluation may be the best approach.

PLEURAL EFFUSIONS

ABDOMINAL WALL HERNIAS

TENSE ASCITES

“Sympathetic” pleural effusions are common in patients with cirrhotic ascites. They usually are unilateral and rightsided but occasionally may be bilateral and larger on the right side than on the left. A unilateral left-sided effusion suggests tuberculosis. A large effusion in a patient with cirrhotic ascites is designated hepatic hydrothorax.103 Most carefully studied patients with hepatic hydrothorax have been shown to have a small defect in the right hemidiaphragm. Occasionally, the effusion develops acutely, with sudden onset of shortness of breath as the abdomen decompresses. With large diaphragmatic defects, ascites may be undetectable on clinical examination despite a large pleural effusion. The most common symptom associated with hepatic hydrothorax is shortness of breath. Infection of the fluid can occur, usually as a result of spontaneous bacterial peritonitis and transmission of bacteria across the diaphragm.104 The analysis of uncomplicated hepatic hydrothorax fluid is similar, but not identical, to that of ascitic fluid because the pleural fluid is subject to hydrostatic pressures different from those that affect the portal bed. The total protein concentration is higher (by approximately 1.0 g/dL [10 g/L]) in the pleural fluid than in ascitic fluid.103 The treatment of hepatic hydrothorax was difficult until the transjugular intrahepatic portosystemic shunt (TIPS) became available (see later).103 The effusions tend to occur in patients who are the least adherent to treatment regimens or in whom ascites is most refractory to therapy. Some

Abdominal wall hernias are common in patients with ascites. They usually are umbilical or incisional and occasionally inguinal. Up to 20% of patients with cirrhosis and ascites have umbilical hernias at the time of hospitalization.107 Some of these hernias incarcerate or perforate. Because of these potential complications, elective surgical treatment should be considered in a patient with a hernia and ascites. Insertion of mesh should be avoided because of the potential for the mesh to become infected. The ascitic fluid should be medically removed preoperatively because the hernia recurs in 73% of patients who have ascites at the time of hernia repair but in only 14% of those who have no ascitic fluid at the time of repair.108 Nevertheless, hernia repair is not without hazard. Successful laparoscopic repair of a recurrent strangulated umbilical hernia has been described.109 TIPS has also been reported to lead to good control of symptoms and may obviate the need for surgical repair.110 Many transplant surgeons prefer to avoid repair of the hernia or postpone it until the time of liver transplantation. An elastic abdominal binder can be used as a temporizing measure to reduce pain and hernia enlargement. Surgical repair of a hernia or TIPS should be performed urgently in patients with skin ulceration, crusting, or black discoloration and emergently for refractory incarceration or rupture. Rupture is the most feared complication of an umbilical hernia. If TIPS is used, it must be performed prior to bacteremia. Infection of the TIPS may be difficult to eradicate.

1535

1536

Section IX  Liver TREATMENT OF ASCITES Appropriate treatment of ascites depends on the cause of fluid retention. Accurate determination of the etiology of ascites is crucial. The SAAG is helpful diagnostically and for therapeutic decision-making. Patients with a low SAAG usually do not have portal hypertension and do not respond to salt restriction and diuretics (except for those with nephrotic syndrome). Conversely, patients with a high SAAG have portal hypertension and are usually responsive to these measures.9

LOW-ALBUMIN-GRADIENT ASCITES

Peritoneal carcinomatosis is the most common cause of lowalbumin-gradient ascites.2 Peripheral edema in affected patients can be managed with diuretics. By contrast, patients without peripheral edema who receive diuretics lose only intravascular volume, without loss of ascitic fluid. The mainstay of treatment of nonovarian peritoneal carcinomatosis is outpatient therapeutic paracentesis. Patients with peritoneal carcinomatosis usually live only a few months. Patients with ovarian malignancy are an exception to this rule and may exhibit a good response to surgical debulking and chemotherapy. Ascites caused by tuberculous peritonitis (without cirrhosis) is cured by antituberculosis therapy. Diuretics do not speed weight loss unless the patient has underlying portal hypertension from cirrhosis. Pancreatic ascites may resolve spontaneously, require endoscopic placement of a stent in the pancreatic duct or operative intervention, or respond to treatment with somatostatin.111 A postoperative lymphatic leak from a distal splenorenal shunt or radical lymphadenectomy also may resolve spontaneously but on occasion may require surgical intervention or placement of a peritoneovenous shunt. Chlamydia peritonitis is cured by tetracycline. Ascites caused by lupus serositis may respond to glucocorticoids.7 Dialysis-related ascites may respond to aggressive dialysis.47

HIGH-ALBUMIN-GRADIENT ASCITES

Cirrhosis is the most common cause of liver disease that leads to high-albumin-gradient ascites (see Table 91-1). Many patients with cirrhosis experience multiple insults to the liver, including excessive alcohol use, NASH, and chronic hepatitis C.4 One of the most important steps in treating high-albumin-gradient ascites in a patient with alcoholic liver disease, with or without other causes of liver injury, is to convince the patient to stop drinking alcohol. In a period of months, abstinence from alcohol can result in healing of the reversible component of alcoholic liver disease, and the ascites may resolve or become more responsive to medical therapy. Similarly, patients with other forms of treatable liver disease (e.g., autoimmune hepatitis, hemochromatosis, Wilson disease) should receive specific therapy for these diseases. Occasionally, cirrhosis due to causes other than alcohol is reversible5; however, these diseases are usually less reversible than alcoholic liver disease, and by the time ascites is present, these patients may be better candidates for liver transplantation than for protracted medical therapy.

Hospitalization

Outpatient treatment of patients with small-volume ascites can be attempted initially. However, patients with largevolume ascites and those who are resistant to outpatient treatment usually require hospitalization for definitive diagnosis and management.9 Many of these patients also have

gastrointestinal hemorrhage, encephalopathy, infection, or hepatocellular carcinoma. An intensive period of inpatient education and treatment may be required to convince the patient that the prescribed diet and diuretics are actually effective and worth the effort to prevent future hospitalizations.

Precipitating Cause

Determining the immediate precipitant of ascites (e.g., dietary indiscretion or nonadherence to therapy with diuretics) may be of value. Ascites may be precipitated by saline infusions given perioperatively or to treat variceal hemorrhage or by sodium bicarbonate tablets; in such cases the ascites may resolve without the need for long-term treatment.

Diet Education

Fluid loss and weight change are related directly to sodium balance in patients with portal hypertension–related ascites. In the presence of avid renal retention of sodium, dietary sodium restriction is essential. The patient and the food preparer should be educated by a dietitian about a sodiumrestricted diet. Severely sodium-restricted diets (e.g., 500 mg, or 22 mmol, of sodium per day) are feasible (but not palatable) in an inpatient setting but unrealistic for outpatients. The dietary sodium restriction that I recommend for both inpatients and outpatients is 2000 mg (88 mmol) per day. Protein is not restricted unless the patient has hepatic encephalopathy refractory to two drugs on a vegetable protein diet.

Fluid Restriction

Indiscriminate restriction of fluid in the treatment of cirrhotic ascites is inappropriate and serves only to alienate patients, nurses, and dietitians; moreover, hypernatremia may result. Sodium restriction, not fluid restriction, results in weight loss; fluid follows sodium passively. The chronic hyponatremia usually seen in patients with cirrhotic ascites is seldom morbid. Attempts to correct hyponatremia rapidly in this setting can lead to more complications than those related to the hyponatremia. Severe hyponatremia (e.g., serum sodium concentration less than 120 mmol/L) does warrant fluid restriction in the patient with cirrhosis and ascites but fortunately occurs in only 1.2% of patients.112 Unless the decline in sodium concentration is rapid, symptoms of hyponatremia usually do not develop in cirrhotic patients until the serum sodium concentration is below 110 mmol/L.

Role of Bed Rest

Although bed rest has traditionally been prescribed, no controlled trials support this practice; bed rest was part of the treatment of heart failure in the past and was extrapolated to the treatment of cirrhosis with ascites without data.113 An upright posture may aggravate the plasma renin elevation found in most cirrhotic patients with ascites and, theoretically, increase renal sodium retention. In all likelihood, however, strict bed rest is unnecessary and may lead to decubitus ulcer formation in emaciated patients.

Urine Sodium Excretion

The 24-hour urinary sodium excretion is a helpful parameter to follow in patients with portal hypertension–related ascites. The completeness of the urine collection can be assessed by measuring the urinary creatinine excretion: Men with cirrhosis should excrete 15 to 20 mg/kg per day of creatinine, and women should excrete 10 to 15 mg/kg per day9; excretion of less creatinine indicates an incomplete

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis collection. Only the 10% to 15% of patients who have significant spontaneous natriuresis can be considered for dietary sodium restriction as sole therapy of ascites (i.e., without diuretics).9 When given a choice, however, most patients would prefer to take some diuretics with more liberal intake of sodium than to take no pills with severe restriction of sodium intake. Contrary to popular belief, most patients, including outpatients, can comply with instructions to collect complete 24-hour urine specimens. Because urine is the most important route of excretion of sodium in the absence of diarrhea or hyperthermia, and because dietary intake is the only source of nonparenteral sodium, dietary intake and urinary excretion of sodium should be roughly equivalent, if the patient’s weight is stable. Nonurinary sodium losses are less than 10 mmol per day in these patients.114 A suboptimal decline in body weight may be the result of inadequate natriuresis, failure to restrict sodium intake, or both. Monitoring 24-hour urinary sodium excretion and daily weight will clarify the issue. Patients who are adherent to an 88 mmol per day sodium diet and who excrete more than 78 mmol per day of sodium in the urine should lose weight. If the weight is increasing despite urinary losses in excess of 78 mmol per day, one can assume that the patient is consuming more sodium than is prescribed in the diet.

Urine Sodium-to-Potassium Ratio

Although 24-hour urine specimens constitute the diagnostic standard, one study has demonstrated that when a random urine specimen has a sodium concentration greater than the potassium concentration, a 24-hour specimen will reveal sodium excretion greater than 78 mmol per day in approximately 90% of cases.115 Therefore, a random urine sodiumto-potassium concentration ratio greater than 1 predicts that the patient should lose weight if a sodium-restricted diet is followed. Patients who do not lose weight despite a random urine sodium-to-potassium ratio greater than 1 probably are not adherent to the diet.

Avoidance of Urinary Bladder Catheters

Many physicians promptly insert a bladder catheter in hospitalized patients with cirrhosis to monitor urine output accurately. Unfortunately, many of these immunocompromised patients have urinary tract infections on hospital admission,72 and urethral trauma from insertion of the catheter in the setting of cystitis can lead to bacteremia. Prolonged catheterization predictably leads to cystitis and possibly sepsis in these patients. I insert urinary catheters only briefly and only in the intensive care unit setting; these portals of entry for bacteria should be removed as soon as possible. Twenty-four-hour urine specimens can be collected completely without catheters.

Diuretics

Spironolactone is the mainstay of treatment for patients with cirrhosis and ascites but increases natriuresis slowly. Single-agent diuretic therapy with spironolactone requires several days to induce weight loss. Although spironolactone alone has been shown to be superior to furosemide alone in the treatment of cirrhotic ascites,116 I prefer to start spironolactone and furosemide together on the first hospital day in initial doses of 100 mg and 40 mg, respectively, each taken once in the morning.9 Amiloride, 10 mg per day, can be substituted for spironolactone; amiloride is less widely available and more expensive than spironolactone but more rapidly effective, and it does not cause gynecomastia. A new potassium-sparing diuretic, eplerenon, has been used in the treatment of heart failure and does not cause gynecomastia,

but studies of its use in cirrhosis are lacking. The half-life of spironolactone is approximately 24 hours in normal control subjects but is markedly prolonged in patients with cirrhosis; almost one month is required to reach a steady state.117 In view of its long half-life, dosing the drug multiple times per day is unnecessary. A loading dose may be appropriate but has not been studied. Single daily doses maximize adherence; 25-, 50-, and 100-mg spironolactone tablets are available generically. Furosemide also should be given once a day.118 If the combination of spironolactone, 100 mg per day (or amiloride, 10 mg per day) and furosemide, 40 mg per day orally, is ineffective in increasing urinary sodium or decreasing body weight, the doses of both drugs should be increased simultaneously, as needed (e.g., spironolactone, 200 mg plus furosemide, 80 mg, then 300 mg plus 120 mg, and finally 400 mg plus 160 mg). In my experience, as well as in a randomized trial, starting both drugs at once speeds the onset of diuresis, whereas slowly increasing the daily dose of spironolactone to 400 mg or even higher before adding furosemide delays diuresis and results in hyperkalemia.119 The 100 : 40 ratio of the daily doses of spironolactone and furosemide usually maintains normokalemia. The ratio of doses can be adjusted to correct abnormal serum potassium levels. Occasionally, an alcoholic patient who has had no recent food intake will have hypokalemia at the time of admission and for a variable interval thereafter. Such a patient should receive spironolactone alone until the serum potassium normalizes; furosemide can then be added. When combined with a sodium-restricted diet in a study of almost 4000 patients, the regimen of spironolactone and furosemide has been demonstrated to achieve successful diuresis in more than 90% of cirrhotic patients.120 Intravenous diuretics cause acute decreases in the glomerular filtration rate in patients with cirrhosis and ascites and generally should be avoided.121 Many patients are given intravenous furosemide when they are hospitalized because of failure of outpatient treatment of ascites in the setting of cirrhosis. The approach of switching from oral to intravenous administration is effective for heart failure, but in patients with cirrhosis, repeated doses of intravenous furosemide regularly lead to azotemia and then to an erroneous diagnosis of hepatorenal syndrome. (The correct diagnosis is diuretic-induced azotemia that resolves when the diuretics are withheld and fluid is administered intravenously.) Some physicians give intravenous albumin with intravenous furosemide, but a randomized crossover study has shown no benefit to albumin in this setting.122 Repeated intravenous dosing of furosemide appears to be too “harsh” for the patient with cirrhosis; oral diuretics are better tolerated. If rapid weight loss is desired, therapeutic paracentesis should be performed (see later). No limit has been identified for acceptable daily weight loss in patients who have massive edema. As soon as the edema has resolved, a reasonable maximum weight loss is probably 0.5 kg per day.123 Encephalopathy, a serum sodium concentration less than 120 mmol/L despite fluid restriction, and a serum creatinine level greater than 2.0 mg/dL (180 mmol/L) are indications to discontinue diuretics and reassess the patient. Abnormalities in potassium levels almost never prohibit diuretic use because the ratio of the two diuretics can be readjusted. Patients with parenchymal renal disease (e.g., diabetic nephropathy) usually require relatively higher doses of furosemide and lower doses of spironolactone; otherwise, they develop hyperkalemia. Patients in whom

1537

1538

Section IX  Liver complications develop despite a careful attempt at diuretic treatment usually require second-line therapy. Prostaglandin inhibitors (e.g., nonsteroidal anti-inflammatory drugs) should be avoided in patients with cirrhosis and ascites because they inhibit diuresis, may promote renal failure, and may cause gastrointestinal bleeding.124 Reducing the quantity of fluid in the abdomen can improve the patient’s comfort and prevent hepatic hydrothorax and hernias. Also, by concentrating the ascitic fluid, diuresis increases the opsonic activity of ascitic fluid 10-fold and theoretically may be of value in preventing spontaneous ascitic fluid infection.125 An issue that nurses regularly raise is whether diuretics should be withheld when a patient’s blood pressure is low. No data exist to support this practice in the setting of cirrhosis. Baseline blood pressure, mental status, and creatinine must be factored into the decision to continue, hold, or discontinue diuretics. The baseline blood pressure is usually low (e.g., 70 to 100 systolic, in a patient with cirrhotic ascites). Unless it has dropped significantly or the patient has confusion or azotemia, diuretics should be given. In the past, patients with ascites frequently occupied hospital beds for prolonged durations because of uncertainty regarding the diagnosis and optimal treatment and because of iatrogenic complications. Although a “dry” abdomen is a reasonable ultimate goal, complete resolution of ascites should not be a prerequisite for discharge from the hospital. Patients who are stable, with ascites as their major problem, can be discharged after they are demonstrated to be responding to the medical regimen and are normokalemic, are not azotemic, and have a normal or slightly to moderately reduced serum sodium level. Following discharge from the hospital, a patient should be seen in the outpatient setting within 7 to 14 days.

responding to treatment and are not experiencing problems. Intensive outpatient follow-up helps prevent subsequent hospitalizations. Diuretic doses and dietary sodium intake are adjusted to achieve weight loss and negative sodium balance. Patients who are gaining fluid weight despite diuretic therapy should not be considered to have diuretic-resistant ascites until they are demonstrated to be adherent to the prescribed diet. Monitoring the urine sodium concentration provides insight into adherence. Patients who excrete more than 78 mmol per day of sodium in the urine or have a random urine sodium-to-potassium ratio greater than 1 should be losing weight if they are consuming less than 88 mmol of sodium per day. In my experience, most patients who initially are thought to be diuretic-resistant eventually are found to be nonadherent to the diet; they demonstrate weight gain and urinary sodium excretion as high as 500 mmol per day or more. Diet education is crucial to the successful management of such patients. Patients with truly diuretic-resistant ascites excrete nearly sodium-free urine despite maximal doses of diuretics. During long-term follow-up, abstinent alcoholic patients may become more sensitive to diuretics. In these cases, the dose of diuretics may be tapered and the drugs even discontinued.

REFRACTORY ASCITES

Mild renal tubular acidosis develops in a substantial minority of patients with cirrhosis and ascites. Although oral sodium bicarbonate administration has been recommended in this setting, such treatment increases sodium intake dramatically and cannot be advocated in the absence of evidence to support its use.

Refractory ascites is defined as ascites unresponsive to a sodium-restricted diet and high-dose diuretic treatment. Refractoriness may manifest as minimal or no weight loss despite diuretics or the development of complications of diuretics.127 Several studies have shown that ascites in the setting of cirrhosis is refractory to standard medical therapy in fewer than 10% of patients.116,120 In the 1960s, portacaval shunts were used to treat refractory ascites, but operative hemorrhagic complications and portosystemic encephalopathy led to abandonment of this approach.113 In Europe in the 1970s, the Paris pump was used to ultrafilter ascitic fluid and reinfuse it intravenously.113 Unfortunately, this approach was complicated by disseminated intravascular coagulation and was abandoned. Viable options for patients refractory to routine medical therapy include liver transplantation, serial therapeutic paracenteses, TIPS, and peritoneovenous shunts (Fig. 91-5).9

Aquaretics

Liver Transplantation

Role of Sodium Bicarbonate

The aquaretics are a new class of drugs that have been used in animals and preliminarily in patients with cirrhosis to increase urinary water excretion and to increase the serum sodium concentration. Patients with mild hyponatremia (serum sodium less than 130 mmol/L) can respond with an increase in the serum sodium level, although dose reductions were common in a randomized trial.126 Whether these drugs will improve severe hyponatremia without causing hypotension awaits further investigation.

Outpatient Management

After discharge from the hospital, the patient’s body weight, orthostatic symptoms, and serum electrolyte, urea, and creatinine levels should be monitored. Twenty-four-hour or random urine specimens for a sodium-to-potassium ratio can be collected to assist with treatment decisions. It is my experience that adherent outpatients can collect complete specimens successfully, when adequate written and oral instructions are provided. The subsequent frequency of follow-up evaluations is determined by the response to treatment and stability of the patient. I usually evaluate these patients every one to four weeks until they clearly are

Liver transplantation should be considered among the treatment options for patients with cirrhosis and ascites— whether the fluid is diuretic-sensitive or diuretic-refractory (see also Chapter 95). In many areas of the United States, patients with ascites are not offered transplantation until hepatorenal syndrome has developed (see Chapter 92). The 12-month survival rate for patients with ascites refractory to medical therapy is only 32%.128 The survival rate for liver transplantation is much higher. In patients who are candidates for liver transplantation, procedures that could make transplantation difficult should be avoided. Surgery in the right upper quadrant causes adhesions that become vascularized and difficult to remove during transplant surgery. Even peritoneovenous shunting can lead to the formation of a “cocoon” in the right upper quadrant that can involve the bowel and liver.129

Serial Paracenteses

Therapeutic abdominal paracentesis is one of the oldest medical procedures. In the 1980s, after 2000 years of use, scientific data regarding large-volume paracentesis were reported, and patients were documented to tolerate large-

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis First-line treatment

Second-line treatment

Third-line treatment

Liver transplantation if appropriate candidate

Peritoneovenous shunt

Non-tense ascites

Diagnostic paracentesis

2 gm/day Na diet and diuretics (spironolactone and furosemide)

Good response 90% 10% Poor response

Therapeutic paracentesis (4-5 L) ± intravenous albumin (see text) Tense ascites

Outpatient therapeutic paracentesis

Transjugular intrahepatic portosystemic shunt

volume paracentesis well, just as patients had in the 1940s and earlier.130 In one large randomized, controlled trial, therapeutic paracentesis plus intravenous infusion of colloid led to fewer minor (asymptomatic) changes in serum electrolyte and creatinine levels than those reported with diuretic therapy; however, no differences in morbidity or mortality rates could be demonstrated.130 Therapeutic paracentesis now appears to be first-line therapy for patients in whom ascites is tense and second-line therapy for cirrhotic patients in whom ascites is refractory to diuretics (see Fig. 91-5).9 The world record for volume of fluid removed at one time appears to be 41 L.131

Colloid Replacement

A controversial issue regarding therapeutic paracentesis is the role of colloid replacement. In one study, patients with tense ascites were randomized to receive intravenous albumin (10 g/L of fluid removed) or no albumin after therapeutic paracentesis.132 More statistically significant (asymptomatic) changes in serum electrolyte, plasma renin, and serum creatinine levels developed in the patients who did not receive albumin than in those who received albumin, but no greater frequency of clinical morbidity or mortality was seen. Although another study has documented that the patients who have a postparacentesis rise in plasma renin levels have a decreased life expectancy compared with those who have stable renin levels, no study has demonstrated a decreased survival rate in patients not given a plasma expander compared with patients given albumin after paracentesis.133 A new phrase, “paracentesis-induced circulatory dysfunction,” has been coined to describe the rise in plasma renin levels after paracentesis.134 Despite the lack of benefit of albumin infusion on survival, the authors of the two studies cited previously recommend routine infusion of albumin after therapeutic paracentesis.132,134 Albumin infusions markedly increase the degradation of albumin, however, and albumin is expensive.135,136 In a study performed in the 1960s, 58% of infused albumin was offset by

Figure 91-5.  Algorithm for the treatment of patients with cirrhosis and ascites.

increased degradation, and a 15% increase in the serum albumin level led to a 39% increase in degradation.135 Increasing the concentration of albumin in cell culture media has been shown to decrease albumin synthesis.137 In view of the cost ($7 to $25/g or $350 to $1250/tap), it is difficult to justify the expense of routine infusions of albumin based on the available data. The confusion regarding albumin infusion relates, in part, to the design of the relevant studies. In the studies from Barcelona, patients with “tense” ascites could be entered into the trial of albumin versus no albumin, and 31% of these patients were not even receiving diuretics.131 It seems more appropriate to study the population in which chronic paracenteses are really needed, specifically the diureticresistant group, rather than all patients with tense ascites.138 Another group of investigators has shown that patients with cirrhosis and diuretic-resistant ascites tolerate a 5-L paracentesis without a change in plasma renin levels.139 My approach to patients with tense ascites is to take off enough fluid (4 to 5 L) to relieve intra-abdominal pressure and then to rely on diuretics to eliminate the remainder. To remove all of the fluid by paracentesis when most of it can be removed with diuretics seems inappropriate, partly because paracentesis removes opsonins, whereas diuresis concentrates opsonins.125 Patients with early cirrhosis and diureticsensitive ascites should be managed with diuretics, not large-volume paracentesis; these patients may be more sensitive to paracentesis-related volume depletion than are patients with advanced cirrhosis.140 Chronic therapeutic paracenteses should be reserved for the 10% of patients in whom diuretic treatment fails to relieve the ascites. Other studies have compared less expensive plasma expanders with albumin. No differences in electrolyte imbalance or clinically relevant complications between the groups have been found.141 In addition, some authors advocate giving one half of the plasma expander immediately after the paracentesis and the other half six hours later.133,141 This approach converts an otherwise simple outpatient pro-

1539

1540

Section IX  Liver cedure into an all-day clinic visit or even a brief hospitalization and seems unwarranted. A colloid that specifically should be avoided is hetastarch, which can accumulate in Kupffer cells and cause portal hypertension in patients without preexisting liver disease.142 Consensus statements, a randomized trial of albumin versus saline in 6997 critically ill patients, and systematic reviews have pointed out some of the hazards of albumin infusion and have recommended against its liberal use.143-145 Until more convincing data involving appropriate groups of patients are available, it seems reasonable to (1) avoid serial large-volume paracenteses in patients with diureticsensitive ascites; (2) withhold albumin after taps of 5 L or less; and (3) consider albumin infusion optional after taps of larger volume in patients with diuretic-resistant ascites.9 A small, randomized trial has shown that terlipressin may be equivalent to albumin after therapeutic paracentesis in preventing paracentesis-induced circulatory dysfunction; if this drug receives approval for use in the United States and further studies support its efficacy, terlipressin would be an alternative to albumin.146

Transjugular Intrahepatic Portosystemic Shunt

TIPS is a side-to-side portacaval shunt that is placed by an interventional radiologist (or hepatologist), usually with the use of local anesthesia. TIPS placement was first used for the treatment of refractory variceal bleeding, but it also has been advocated for diuretic-resistant ascites147 (see Chapter 90). TIPS was received with great enthusiasm in the 1990s, similar to the enthusiasm for the peritoneovenous shunt in the 1970s. Just as with peritoneovenous shunting, TIPS was overused until serious complications and suboptimal efficacy were reported. Four large-scale randomized trials in diuretic-resistant patients have demonstrated consistent superiority of TIPS over repeated paracentesis but no survival advantage.148-151 Multiple meta-analyses have been published confirming efficacy but with more hepatic encephalopathy in the TIPS group.152-156 One meta-analysis has demonstrated a trend toward improved survival in the TIPS group.153 Another metaanalysis, which analyzed individual patient data, did show improved transplant-free survival with TIPS.156 Although TIPS dysfunction was common when an uncoated (or uncovered) shunt was used, polytetrafluoroethylene-coated stents have been reported to improve patency and survival when compared with uncoated stents in a nonrandomized study and to improve patency, with no survival advantage, when compared with uncoated stents in a randomized trial.157,158 Also, the four older TIPS trials preceded deve­ lopment and implementation of the Model for End-stage Liver Disease (MELD) score, which predicts 90-day mor­ tality after TIPS placement (see Chapter 90); new trials using the coated stent and selecting patients according to their MELD scores may demonstrate a survival advantage for TIPS compared with repeated taps. TIPS also is useful in the treatment of hepatic hydrothorax and umbilical hernia.111,118 A direct intrahepatic portosystemic shunt connects the portal vein directly to the inferior vena cava and has applicability in patients with Budd-Chiari syndrome (see Chapter 83).159

Peritoneovenous Shunt

In the mid-1970s, the peritoneovenous shunt was promoted as a new “physiologic” treatment for the management of ascites. Reports of shunt failure, fatal complications following shunt insertion, and randomized trials demonstrating no survival advantage have led to the relegation of this procedure to third-line therapy in patients with cirrhosis

and ascites9,120 (see Fig. 91-5). Patients who are not candidates for liver transplantation and who have a scarred abdomen that is not amenable to repeated paracenteses, who are not candidates for a TIPS, or in whom an attempt at TIPS placement has failed make up the small subset of candidates for a peritoneovenous shunt. A randomized trial has shown that even an uncoated TIPS stent has better “assisted patency” than the peritoneovenous shunt.160

Novel Treatments

Novel treatment options for patients with refractory ascites include weekly infusions of intravenous albumin, ascites reinfusion, ultrafiltration, terlipressin infusion (not available in the United States), partial splenic artery embolization, peritoneal-urinary drainage of the fluid using a surgically implanted pump, and percutaneous placement of a peritoneovenous shunt by an interventional radiologist.161-168 More data are needed before these treatments can be advocated.

PROGNOSIS Cirrhosis complicated by ascites is associated with significant morbidity and mortality, related, in part, to the severe underlying liver disease and, in part, to the ascites per se. In one half of the patients in whom cirrhosis is detected before decompensation (i.e., development of ascites, jaundice, or encephalopathy or gastrointestinal hemorrhage), ascites occurs within 10 years.169 When ascites appears, the expected mortality rate is approximately 50% in just two years.170 With liver transplantation, survival is improved dramatically.

KEY REFERENCES

Akriviadis EA, Runyon BA. The value of an algorithm in differentiating spontaneous from secondary bacterial peritonitis. Gastroenterology 1990; 98:127-33. (Ref 28.) Fernandez J, Navasa M, Garcia-Pagan JC, et al. Effect of intravenous albumin on systemic and hepatic hemodynamics and vasoactive neurohumoral systems in patients with cirrhosis and spontaneous bacterial peritonitis. J Hepatol 2004; 41:384-90. (Ref 82.) Fernandez J, Navasa M, Planas R, et al. Primary prophylaxis of spontaneous bacterial peritonitis delays hepatorenal syndrome and improves survival in cirrhosis. Gastroenterology 2007; 133:818-24. (Ref 94.) Fernandez J, Ruiz del Arbol L, Gomez C, et al. Norfloxacin vs. ceftriaxone in the prophylaxis of infections in patients with advanced cirrhosis and hemorrhage. Gastroenterology 2006; 131:1049-56. (Ref 92.) Follo A, Llovet JM, Navasa M, et al. Renal impairment after spontaneous bacterial peritonitis in cirrhosis: Incidence, clinical course, predictive factors and prognosis. Hepatology 1994; 20:1495-501. (Ref 79.) Perez-Ayuso RM, Arroyo V, Planas R, et al. Randomized comparative study of efficacy of furosemide vs. spironolactone in nonazotemic cirrhosis with ascites. Gastroenterology 1983; 84:961-8. (Ref 116.) Runyon BA. Management of adult patients with ascites caused by cirrhosis. Hepatology 2004; 39:841-56. (Ref 9.) Runyon BA, Canawati HN, Akriviadis EA. Optimization of ascitic fluid culture technique. Gastroenterology 1988; 95:1351-5. (Ref 23.) Runyon BA, Hoefs JC, Morgan TR. Ascitic fluid analysis in malignancyrelated ascites. Hepatology 1988; 8:1104-9. (Ref 2.) Runyon BA, McHutchison JG, Antillon MR, et al. Short-course vs. longcourse antibiotic treatment of spontaneous bacterial peritonitis: A randomized controlled trial of 100 patients. Gastroenterology 1991; 100:1737-42. (Ref 87.) Runyon BA, Montano AA, Akriviadis EA, et al. The serum-ascites albumin gradient is superior to the exudate-transudate concept in the differential diagnosis of ascites. Ann Intern Med 1992; 117:215-20. (Ref 21.) Runyon BA, Squier SU, Borzio M. Translocation of gut bacteria in rats with cirrhosis to mesenteric lymph nodes partially explains the

Chapter 91  Ascites and Spontaneous Bacterial Peritonitis pathogenesis of spontaneous bacterial peritonitis. J Hepatol 1994; 21:792-6. (Ref 61.) Salerno F, Camma C, Enea M, et al. Transjugular intrahepatic porto­ systemic shunt for refractory ascites: A meta-analysis of individual patient data. Gastroenterology 2007; 133:825-34. (Ref 156.) Salerno F, Merli M, Riggio O, et al. Randomized controlled study of TIPS vs. paracentesis with albumin in cirrhosis with refractory ascites. Hepatology 2004; 40:629-35. (Ref 151.)

Wong C, Holroyd-Leduc J, Thorpe KE, Straus SE. Does this patient have bacterial peritonitis or portal hypertension? How do I perform a paracentesis and analyze the results? JAMA 2008; 299:1166-78. (Ref 145.) Full references for this chapter can be found on www.expertconsult.com.

1541

CHAPTE R

92 Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome, and Systemic Complications of Liver Disease Moises Ilan Nevah and Michael B. Fallon

CHAPTER OUTLINE Hepatic Encephalopathy  1543 Pathophysiology  1544 Clinical Features and Diagnosis  1544 Treatment  1545 Hepatorenal Syndrome  1546 Pathophysiology  1546 Clinical Features and Diagnosis  1547 Classification  1547 Prevention and Treatment  1548 Hepatopulmonary Syndrome and Portopulmonary Hypertension  1549 Pathophysiology  1549

The liver is a key regulator of the normal function of other organ systems by virtue of its central role in nutrition, metabolism, and secretion. Chronic liver disease and acute liver failure disrupt this normal homeostasis and cause sys­ temic manifestations that may dominate the clinical fea­ tures of liver disease. Despite the diverse organ systems affected, the extrahepatic manifestations of liver disease share common mechanistic pathways that reflect secondary effects of cirrhosis rather than primary abnormalities in the target organ. Therefore, most of the extrahepatic syndromes are reversible with liver transplantation. This chapter pre­ sents an overview of the pathophysiology, diagnosis, and treatment of common extrahepatic manifestations of liver disease.

Clinical Features and Diagnosis  1550 Treatment  1551 Cirrhotic Cardiomyopathy  1553 Pathophysiology  1553 Clinical Features and Diagnosis  1553 Treatment  1553 Endocrine Dysfunction  1554 Adrenal Insufficiency  1554 Gonadal Dysfunction  1554 Thyroid Dysfunction  1554 Bone Disease  1554 Coagulation Disorders  1554

HEPATIC ENCEPHALOPATHY The term hepatic encephalopathy (HE) encompasses a wide array of transient and reversible neurologic and psychiatric manifestations usually found in patients with chronic liver disease and portal hypertension, but also seen in patients with acute liver failure. HE develops in 50% to 70% of patients with cirrhosis, and its occurrence is a poor prog­ nostic indicator, with projected one- and three-year survival rates of 42% and 23%, respectively, without liver transplan­ tation.1 Symptoms may range from mild neurologic distur­ bances to overt coma.2,3 HE is often triggered by an inciting event that results in a rise in the serum ammonia level. The precise underlying pathophysiologic mechanisms are not

1543

1544

Section IX  Liver Cirrhosis

Inflammation, bacterial translocation, and vasoactive mediators

Splanchnic and peripheral vasodilatation

Hepatocyte dysfunction

Portosystemic shunting

↑ Blood NH3 levels

Brain edema Astrocyte swelling (↑ Glu and Gln) Neurotransmitter and receptor alterations (↑ GABA) Altered brain glucose metabolism

Increased permeability of blood-brain barrier

Hepatic encephalopathy Figure 92-1.  Pathophysiology of hepatic encephalopathy. GABA, γ-aminobutyric acid; Gln, glutamine; Glu, glutamate; NH3, ammonia.

well understood, and the mainstay of therapy is the elimina­ tion of the precipitating event and excess ammonia.4 Liver transplantation generally reverses HE.

PATHOPHYSIOLOGY

A number of factors, occurring alone or in combination, have been implicated in the development of HE. These factors may differ in acute and chronic liver disease and include the production of neurotoxins, altered permeability of the blood-brain barrier, and abnormal neurotransmission (Fig. 92-1). The best-described neurotoxin involved in HE is ammonia, which is produced primarily in the colon, where bacteria metabolize proteins and other nitrogenbased products into ammonia. Enterocytes synthesize ammonia from glutamine.4-6 Once produced, ammonia enters the portal circulation and, under normal conditions, is metabolized and cleared by hepatocytes. In cirrhosis and portal hypertension, reduced hepatocyte function and por­ tosystemic shunting contribute to increased circulating ammonia levels. Arterial hyperammonemia is observed in up to 90% of patients with HE, although serum levels are neither sensitive nor specific indicators of its presence. Increased permeability of the blood-brain barrier increases the uptake and extraction of ammonia by the cerebellum and basal ganglia.7-9 Acute hyperammonemia appears to have a direct effect on brain edema, astrocyte swelling, and the transport of neurally active compounds such as myoino­ sitol, and thereby contributes to HE.10-12 Other alterations in HE affect neuronal membrane fluidity, central nervous system (CNS) neurotransmitter expression, and neurotransmitter receptor expression and activation.13,14 The γ-aminobutyric acid (GABA)–benzodiaz­ epine system has been the most well studied. Although CNS benzodiazepine levels and GABA receptor concentra­ tions are unchanged in animal models of HE, increased sensitivity of the astrocyte (peripheral-type) benzodiazepine receptor enhances activation of the GABA-benzodiazepine system.15,16 This activation occurs in part through a feed-forward system in which production of neurosteroids (allopregnanolone and tetrahydrodeoxycorticosterone) by

astrocytes further activates the GABAA-benzodiazepine receptor system.17,18 Other factors that influence CNS neu­ rotransmission, including serotonin (5-hydroxytryptamine, 5-HT),19-21 nitric oxide (NO), circulating opioid peptides, manganese, and increased oxygen free radical production, have also been postulated to contribute to HE.4 Finally, hyperammonemia, particularly in acute liver failure, also increases astrocyte glutamine production via glutamine synthetase. The rise in astrocyte glutamine and glutamate concentrations contributes to factors associated with CNS dysfunction.5,22,23

CLINICAL FEATURES AND DIAGNOSIS

HE may present as a spectrum of reversible neuropsychiat­ ric symptoms and signs, ranging from mild changes in cog­ nition to profound coma, in patients with acute or chronic liver disease. It is often precipitated by an inciting event (e.g., gastrointestinal bleeding, electrolyte abnormalities, infections, medications, dehydration). The diagnosis of HE, therefore, requires careful consideration in the appropriate clinical situation. Occasionally, HE may be the initial pre­ sentation of chronic liver disease. Subtle findings in HE may include forgetfulness, alterations in handwriting, dif­ ficulty with driving, and reversal of the sleep-wake cycle.24,25 Overt findings may include asterixis, agitation, disinhibited behavior, seizures, and coma. Other causes of altered mental status, particularly hypoglycemia, hyponatremia, medica­ tion ingestion, and structural intracranial abnormalities resulting from coagulopathy or trauma, should be consid­ ered and rapidly excluded in patients suspected of having HE. No specific laboratory findings indicate the presence of HE definitively. The most commonly used test to assess a patient with possible HE is the blood ammonia level. An elevation in the blood ammonia level in a patient with cir­ rhosis and altered mental status supports a diagnosis of HE. Blood ammonia levels may be elevated in the absence of HE, however, because of gastrointestinal bleeding or the ingestion of certain medications (e.g., diuretics, alcohol, narco­tics, valproic acid).15,26,27 In addition, blood ammonia levels may be elevated in the presence of HE, even in the absence of cirrhosis and portal hypertension, in patients with metabolic disorders that influence ammonia genera­ tion or meta­bolism, such as urea cycle disorders (see Chapter 76) and dis­orders of proline metabolism (Table 92-1).28,29 Use of a tourniquet when blood is drawn and delayed pro­ cessing and cooling of a blood sample may raise the blood ammonia level.15 Measurement of arterial ammonia offers no advantage over measurement of venous ammonia levels in patients with chronic liver disease.30 In patients with acute liver failure, however, elevated arterial ammonia levels (150 to 200 mg/dL or higher) may be predictive of the presence of brain edema and herniation (see Chapter 93).12,31,32 Of the scoring systems used to grade the severity of HE, the West Haven system, based on a scale of 0 to 4, is the most widely used in clinical practice (Table 92-2).25 Although clinically useful, the West Haven criteria are insensitive and have led to the development of standardized psychometric tests and rapid bedside mental status assess­ ments to aid in the diagnosis of HE and facilitate research.33-37 One simple paper and pencil test, the portosystemic enceph­ alopathy syndrome test (PSET), evaluates the patient’s attention, concentration, fine motor skills, and orientation and has been shown to be highly specific for the diagnosis of HE.33,38 The development of these tests has led to recogni­ tion of the syndrome of minimal HE, in which abnormalities are observed on testing but clinically recognizable altera­ tions of HE are minimal or not detected. The presence of

Chapter 92  Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome minimal HE is common in patients with cirrhosis, appears to influence the patient’s quality of life and driving ability, and confers an increased risk that overt HE will develop in the patient. Whether treatment of minimal HE confers any benefit is an area of active investigation.24,39-41 A number of novel imaging and functional tests have been studied in the diagnosis of HE. Magnetic resonance spec­ troscopy (MRS) has been used to measure brain concentra­ tions of choline and glutamine noninvasively.42 Magnetic resonance (MR) T1 mapping with partial inversion recovery (TAPIR) has been investigated as a means to measure changes in the brain quantitatively over clinically relevant measurement times.33 Whether MR-based techniques can be standardized and become practical diagnostic tests is uncer­ tain. The critical flicker frequency test, a simple light-based test that has been used to assess cerebral cortex function in a number of disorders, has been shown to be a reliable marker of minimal HE and may become a clinically useful screening test.34-36

TREATMENT

Current treatments for HE are directed primarily toward the elimination or correction of precipitating factors (e.g.,

Table 92-1  Differential Diagnosis of Hyperammonemia Acute liver failure Chronic kidney disease Cigarette smoking Cirrhosis Gastrointestinal bleeding Inborn errors of metabolism Proline metabolism disorders Urea cycle disorders (e.g., carbamyl phosphate synthetase I deficiency, ornithine transcarbamylase deficiency, argininosuccinate lyase deficiency, N-acetyl glutamate synthetase deficiency) Medications Alcohol Diuretics (e.g., acetazolamide) Narcotics Valproic acid Muscle exertion and ischemia Portosystemic shunts Technique and conditions of blood sampling High body temperature High-protein diet Tourniquet use

bleeding, infection, hypokalemia, medications, dehydra­ tion), reduction in elevated blood ammonia levels, and avoidance of the toxic effects of ammonia in the CNS. In the past, dietary protein restriction was considered an impor­ tant component of the treatment of HE. Subsequent work, however, has suggested that limiting protein-calorie intake is not beneficial in patients with HE.43-45 Vegetable and dairy proteins are preferred to animal proteins because of a more favorable calorie-to-nitrogen ratio. Although branchedchain amino acid supplementation may improve symptoms modestly, the benefits of such supplementation are not suf­ ficient to justify its routine use.4 Nonabsorbable disaccharides have been the cornerstone of the treatment of HE. Oral lactulose or lactitol (the latter is not available in the United States) are metabolized by colonic bacteria to byproducts that appear to have beneficial effects by causing catharsis and reducing intestinal pH, thereby inhibiting ammonia absorption.46 These agents improve symptoms in patients with acute and chronic HE when compared with placebo but do not improve psycho­ metric test performance or mortality. The most common side effects experienced by patients who take lactulose are abdominal cramping, flatulence, diarrhea, and electrolyte imbalance. Lactulose may also be administered per rectum (as an enema) to patients who are at increased risk of aspira­ tion, although the efficacy of enema administration has not been evaluated. Oral antibiotics also have been used to treat HE, with the aim of modifying the intestinal flora and lowering stool pH to enhance the excretion of ammonia. Antibiotics are gener­ ally used as second-line agents after lactulose or in patients who are intolerant of nonabsorbable disaccharides. Neomy­ cin has been approved by the U.S. Food and Drug Admin­ istration (FDA) for use in acute HE in a dose of 1 to 3 g orally every six hours for up to six days but has been used more commonly off-label to treat chronic HE in doses of 0.5 to 1 g every 12 hours, in addition to lactulose. The efficacy of neomycin in acute or chronic HE, however, is not clearly established,47 and ototoxicity and nephrotoxicity caused by neomycin have been reported, particularly in patients with preexisting renal dysfunction.4 Rifaximin has been studied and approved by the FDA for the treatment of chronic HE on the basis of the results of a multicentered, randomized, controlled trial in which the overall clinical efficacy and rate of side effects were similar in patients treated with lactitol and those treated with rifaximin.48 The usual dose is 400 mg orally three times daily. Two systematic reviews of randomized controlled trials that compared rifaximin

Table 92-2  Clinical Stages of Hepatic Encephalopathy Impairment GRADE

INTELLECTUAL FUNCTION

NEUROMUSCULAR FUNCTION

0 Minimal, subclinical

Normal Normal examination findings. Subtle changes in work or driving Personality changes, attention deficits, irritability, depressed state Changes in sleep-wake cycle, lethargy, mood and behavioral changes, cognitive dysfunction Altered level of consciousness (somnolence), confusion, disorientation, and amnesia Stupor and coma

Normal Minor abnormalities of visual perception or on psychometric or number tests Tremor and incoordination

1 2 3 4

Asterixis, ataxic gait, speech abnormalities (slow and slurred) Muscular rigidity, nystagmus, clonus, Babinski sign, hyporeflexia Oculocephalic reflex, unresponsiveness to noxious stimuli

Modified from the West Haven Criteria; in Ferenci P, Lockwood A, Mullen K, et al. Hepatic encephalopathy—definition, nomenclature, diagnosis, and quantification: Final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, 1998. Hepatology 2002; 35:716-21.

1545

1546

Section IX  Liver with other therapies (nonabsorbable disaccharides and other antibiotics) for the treatment of acute or chronic HE have confirmed that the efficacy and side effect profiles are comparable.49,50 Other antibiotics, including metronidazole and vancomycin, have been reported to be effective in small trials and case series, but the data to support their use are insufficient. In addition to antibiotics, several other agents that may modify intestinal flora and modulate ammonia generation or absorption have been evaluated as potential treatments for HE. Acarbose, an intestinal α-glucosidase inhibitor used to treat type 2 diabetes mellitus, inhibits the intestinal absorption of carbohydrates and glucose and results in their enhanced delivery to the colon. As a result, the ratio of saccharolytic to proteolytic bacterial flora is increased, and blood ammonia levels are decreased. A randomized, controlled, double-blind, crossover trial has demonstrated that acarbose improves mild HE in patients with cirrhosis and adult-onset diabetes mellitus.51 Similarly, probiotic regimens have been used to modify intestinal flora and diminish ammonia generation. Four small studies have suggested that these agents may be beneficial in humans with mild HE.40,52-55 These agents merit further evaluation and may be alternatives for patients who do not tolerate lactulose. Strategies to enhance ammonia clearance may also be useful in the treatment of HE. Sodium benzoate, sodium phenylbutyrate, and sodium phenylacetate, all of which increase ammonia excretion in urine, are approved by the FDA for the treatment of hyperammonemia resulting from urea cycle enzyme defects and may improve HE in cirrhosis (see Chapter 76). Administration of sodium benzo­ ate, however, results in a high sodium load, and the efficacy of this agent is not clearly established.4,56 The combination of intravenous sodium phenylacetate and sodium benzoate (Ammonul, Ucyclyd Pharma, Scottsdale, Ariz) in HE is being studied. Administration of zinc, which has been used because zinc deficiency is common in patients with cirrho­ sis57-59 and because zinc increases the activity of ornithine transcarbamylase, an enzyme in the urea cycle, may also improve HE; however, clear efficacy has not been estab­ lished. Extracorporeal albumin dialysis using the molecular adsorbent recirculating system (MARS) has resulted in a reduction in blood ammonia levels and improvement in severe HE in patients with acute-on-chronic liver failure (see Chapter 93).60 Further studies are needed to clarify whether albumin dialysis has a role in treatment of HE. Finally, l-ornithine–l-aspartate (LOLA), a salt of the amino acids ornithine and aspartic acid that activates the urea cycle and enhances ammonia clearance, has been shown in several randomized controlled studies to improve HE com­ pared with lactulose61-63; however, this agent is not available in the United States. Flumazenil is a specific benzodiazepine (GABAA recep­ tor) antagonist that has been used in patients with HE. It improves the degree of encephalopathy and electrophysio­ logic findings in approximately one fourth of patients with grade 3 or 4 HE. It has a short half-life and a number of potential side effects, including seizures, arrhythmias, and withdrawal symptoms, that limit its clinical usefulness.4

HEPATORENAL SYNDROME The term hepatorenal syndrome (HRS) was first used in 1932 to describe acute kidney injury, mainly acute tubular necrosis (ATN) or interstitial nephritis, in a group of patients

who had undergone biliary tract surgery.64 As pathophy­ siologic mechanisms were better elucidated, HRS was found to be part of a cascade of events associated with intense dilatation of the splanchnic arterial vasculature in the setting of cirrhosis or acute liver injury and resulting in profound renal arterial vasoconstriction and progressive renal failure.65 Histologically, the kidneys are normal in HRS. Function may be restored by correction of portal hypertension, liver transplantation, removal of the kidneys and transplantation of them into a noncirrhotic recipient, and, in some cases, medical therapy (see later).66-68 Acute renal dysfunction occurs in 15% to 25% of hospi­ talized patients with cirrhosis. Among the multiple causes of acute kidney injury, prerenal azotemia (resulting from intravascular volume depletion) is most common, account­ ing for 60% to 80% of cases. ATN is the second most common cause of acute kidney injury in this setting and accounts for 20% to 40% of cases.69,70 HRS appears to be an extension of the pathophysiology of prerenal azotemia and is therefore potentially reversible. The annual frequency of HRS in cirrhotic patients with ascites is roughly 8%71 and, in some reports, as high as 40%.72 HRS develops in approxi­ mately 30% of cirrhotic patients who are admitted with spontaneous bacterial peritonitis (SBP) or other infection, 25% who are hospitalized with severe alcoholic hepatitis, and 10% who require serial large-volume paracentesis.73 The observation that morbidity and mortality remain high once the syndrome is established has led to a focus on the prevention and early therapy of renal dysfunction in patients with cirrhosis.71

PATHOPHYSIOLOGY

The pathophysiology of HRS is complex and incompletely characterized. Three important components are recognized to contribute to the initiation and perpetuation of altered renal perfusion (Fig. 92-2): (1) arterial vasodilatation in the splanchnic and systemic circulation; (2) renal vasoconstric­ tion; and (3) cardiac dysfunction.74 These components influ­ ence renal function in concert and form the basis for current therapies and preventive strategies.

Splanchnic Arterial Vasodilatation

Splanchnic and systemic arterial vasodilation are hallmarks of the progression of portal hypertension in cirrhosis and lead to decreased effective circulating blood volume and ultimately to a decrease in blood pressure. This process is mediated by a number of endogenous substances, includ­ ing NO, carbon monoxide (CO), glucagon, prostacyclin, adrenomedullin, and endogenous opiates that are released or act locally in the vasculature in response to mechanical and inflammatory signals.65,71,75,76 In the early stages of portal hypertension, compensation for the decrease in effective circulatory volume includes an increase in heart rate and cardiac output, thereby creating a hyperdynamic circulation.77 As liver disease and splanchnic vasodila­ tation progress, additional compensatory mechanisms are activated.

Renal Arterial Vasoconstriction

Splanchnic and systemic vasodilatation also lead to com­ pensatory renal vasoconstriction and renal sodium and water retention. These responses are mediated by stimula­ tion of the sympathetic nervous system, activation of the renin-angiotensin-aldosterone system (RAAS), and nono­ smotic release and activity of arginine vasopressin (as a result of increased secretion and decreased clearance of arginine vasopressin and apparent increased expression of vasopressin-regulated water channels), as well as intra­

Chapter 92  Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome Inflammation, bacterial translocation, and vasoactive mediators

Cirrhosis

Table 92-3  Diagnostic Criteria for Hepatorenal Syndrome* Cirrhosis with ascites Serum creatinine level ≥ 1.5 mg/dL (133 µmol/L) No or insufficient improvement in serum creatinine level (remains ≥1.5 mg/dL) 48 hr after diuretic withdrawal and adequate volume expansion with intravenous albumin Absence of shock No evidence of recent use of nephrotoxic agents Absence of intrinsic renal disease

Splanchnic and peripheral vasodilatation

CCM

?

↓ Effective circulatory volume ↓ Blood pressure ↑ Cardiac output

↑ RAAS and SNS activity, ↑ plasma ADH Na+ /H2O retention

Ascites/ hyponatremia ACEIs ARBs NSAIDs

Diarrhea Diuretics Gastrointestinal bleeding Large-volume paracentesis SBP/sepsis Hepatic dysfunction

Reduction in GFR or renal vasoconstriction

↓ Renal perfusion

Hepatorenal syndrome Figure 92-2.  Pathophysiology and triggers of hepatorenal syndrome. ACEIs, angiotensin-converting enzyme inhibitors; ADH, antidiuretic hormone; ARBs, angiotensin receptor blockers; CCM, cirrhotic cardio­ myopathy; GFR, glomerular filtration rate; NSAIDs, nonsteroidal antiinflammatory drugs; RAAS, renin-angiotensin-aldosterone system; SBP, spontaneous bacterial peritonitis; SNS, sympathetic nervous system.

renal events. Although the precise intrarenal mechanisms are speculative, altered production or action of endothe­ lins, prostaglandins, kallikreins, and F2-isoprostanes may contribute to renal vasoconstriction.72,78,79 Ultimately, the balance between vasoconstrictive responses in the kidney and systemic and splanchnic vasodilatation is lost, thus leading to a prominent increase in renal vascular resis­ tance, decrease in renal perfusion, and reduction in the glo­merular filtration rate.71,77

Cardiac Dysfunction

Impaired cardiac function also may contribute to renal hypoperfusion in HRS. In one prospective study, HRS developed in cirrhotic patients with more severe arterial vasodilatation and lower cardiac output.80 In another study of a cohort of patients who were treated for SBP, renal dys­ function (including HRS in some cases) developed in those with lower cardiac output and lower arterial pressure measurements associated with higher circulating levels of norepinephrine and renin plasma activity, despite effective treatment of the infection.81 These data demonstrate that cardiac output is impaired in patients with cirrhosis in whom HRS develops as compared with those in whom HRS does not develop, and suggests that cardiac dysfunction may be an important additional factor in the pathogenesis of HRS. The relationship between cardiac dysfunction in patients with HRS and cirrhotic cardiomyopathy (see later) has not been studied.82,83

*As defined by the International Ascites Club Consensus Workshop in 2007 (Salerno F, Gerbes A, Ginès P, et al. Diagnosis, prevention and treatment of hepatorenal syndrome in cirrhosis. Gut 2007; 56:1310-8).

CLINICAL FEATURES AND DIAGNOSIS

HRS is a functional disorder, and therefore diagnostic labo­ ratory and imaging studies are not available. The diagnosis of HRS requires a high index of clinical suspicion and exclusion of other potential causes of kidney injury. The diagnostic criteria for HRS as defined by the International Ascites Club Consensus Workshop in 2007 include the fol­ lowing: (1) cirrhosis with ascites; (2) serum creatinine level higher than 1.5 mg/dL (133 µmol/L); (3) lack of improve­ ment in the serum creatinine level to 1.5 mg/dL (133 µmol/L) or less after at least two days of diuretic withdrawal and volume expansion with albumin (1 g/kg of body weight/ day, up to a maximum of 100 g/day); (4) absence of shock, (5) lack of current or recent treatment with nephrotoxic drugs; and (6) absence of parenchymal kidney disease as indicated by proteinuria of more than 500 mg/day, micro­ hematuria (>50 red blood cells/high power field), or abnor­ mal renal ultrasonographic findings (Table 92-3).74 Several aspects of the diagnosis of HRS deserve emphasis. First, in patients with no prior evidence or history of renal impairment, the diagnostic criteria for HRS include an increase in the serum creatinine level by 50% above baseline to a level higher than 1.5 mg/dL (133 µmol/L).71 Although this definition is standardized, a subset of patients with cirrhosis and end-stage liver disease have a profound decrease in muscle mass and urea synthesis that may in turn result in reduced serum creatinine and blood urea nitrogen levels, thereby potentially delaying recognition of HRS.73,84 Second, many medications, most notably diuretics, lactulose, angiotensin-converting enzyme inhibi­ tors, angiotensin receptor blockers, and nonsteroidal antiinflammatory drugs, may influence intravascular volume status and renal perfusion and should be identified expedi­ tiously and discontinued in the setting of acute renal dys­ function. Third, even though SBP may not be accompanied by obvious symptoms and signs, HRS may develop in as many as 20% of affected patients.85 Therefore, a low thresh­ old for evaluating cirrhotic patients with ascites for the presence of SBP is required.

CLASSIFICATION

HRS has traditionally been classified into two types on the basis of clinical characteristics and prognosis (types 1 and 2).74 Some authors have advocated expanding the classification to include patients who are not encompassed within the current framework (i.e., type 3 HRS).72 Type 1 HRS presents as a rapidly progressive form of renal dysfunction. Typically, the serum creatinine level doubles to a value higher than 2.5 mg/dL in a period of two weeks or less. Type 1 HRS is often triggered by an inciting event that causes a rapid decline in the hemodynamic parameters that maintain renal homeostasis in cirrhotic patients.70 The most common triggers include severe bacte­

1547

1548

Section IX  Liver rial infections,86-88 gastrointestinal bleeding, surgical proce­ dures, and acute liver injury.74,89,90 SBP is the main bacterial infection that predisposes cirrhotic patients to develop HRS. Other bacterial infections have less of an impact on the development of the syndrome, unless sepsis is present or the response to antibiotic therapy is poor.71,87,91 Patients who exhibit high levels of inflammatory response markers and patients with severe circulatory depression prior to the onset of infection are most susceptible to the development of HRS. Some degree of adrenal insufficiency has been found in 80% of patients with septic shock in whom HRS develops, and administration of glucocorticoids may improve survival.92 Type 2 HRS is a more slowly progressing entity compared with type 1 HRS but still carries a median survival of only approximately six months. Type 2 HRS is observed in patients with severe ascites (diuretic resistant) and is char­ acterized by serum creatinine levels lower than 2.5 mg/dL. The degree of arterial hypotension and circulatory dysfunc­ tion is less than that seen with type 1 HRS. Type 1 HRS may develop in patients with type 2 HRS following a triggering event.71,72 Many patients with cirrhosis and portal hypertension also have underlying chronic kidney disease, which complicates recognition of HRS, even in the presence of underlying pathophysiologic mechanisms that favor the development of HRS. Whether these patients should be considered to have a unique form of HRS (type 3) and whether they should be treated in a fashion similar to that for other patients with HRS have not been clearly defined.72

PREVENTION AND TREATMENT

The high mortality rate of HRS underscores the importance of prevention. In particular, intravascular volume depletion (resulting from overdiuresis, diarrhea caused by lactulose, gastrointestinal bleeding from gastroesophageal varices, or large-volume paracentesis without colloid administration), nephrotoxins (e.g., nonsteroidal anti-inflammatory drugs, nephrotoxic antibiotics), and infection (SBP, bacteremia) should be avoided or addressed if present. Specific guide­ lines address the primary and secondary prophylaxis of variceal bleeding (see Chapter 90), administration of colloid (albumin) to patients with a rising serum creatinine level after a large-volume paracentesis and in the presence of SBP (see Chapter 91), and prophylactic administration of anti­ biotics to patients at high risk of SBP or other infections because of hospitalization for gastrointestinal bleeding (see Chapters 19 and 90).85,86,89,91 Routine invasive hemody­ namic monitoring of cirrhotic patients with a rising serum creatinine level does not have a clear benefit and is not recommended. The concept that treatment of HRS is possible and may improve survival has emerged since 2000. Current options include medical therapies, transjugular intrahepatic por­ tosystemic shunt (TIPS) placement, and liver transplan­ tation. Medical therapies for HRS are directed toward reversing the underlying splanchnic and systemic vaso­ dilatation with vasoconstrictors and increasing effective circulatory volume with the use of colloid. Such treatment is used increasingly as a temporizing measure until defini­ tive treatment for liver disease (liver transplantation) or portal hypertension (TIPS) is undertaken, or until an acute process (SBP, gastrointestinal bleeding) has been reversed (Table 92-4).93,94

Medical Therapy

The use of vasoconstrictors with or without administration of colloid in HRS was initially reported in the 1960s. Since then, several regimens, including terlipressin and albumin,

Table 92-4  Management of Hepatorenal Syndrome Measures to prevent variceal bleeding (e.g., beta blockers, band ligation) Pentoxifylline for severe alcoholic hepatitis Prevention Avoid intravascular volume depletion (diuretics, lactulose, gastrointestinal bleeding, large-volume paracentesis without adequate volume repletion) Judicious management of nephrotoxins (ACEIs, ARBs, NSAIDs, antibiotics) Prompt diagnosis and treatment of infections (SBP, sepsis) SBP prophylaxis Medical management Stop all nephrotoxic agents (ACEIs, ARBs, NSAIDs, diuretics) Antibiotics for infections Intravenous albumin—bolus of 1 g/kg/day on presentation (maximum dose, 100 g daily). Continue at dose of 20-60 g daily as needed to maintain central venous pressure between 10 and 15 cm H2O Vasopressor therapy (in addition to albumin): Terlipressin*—start at 1 mg IV every 4 hr and increase up to 2 mg IV every 4 hr if baseline serum creatinine level does not improve by 25% at day 3 of therapy or Midodrine and octreotide—begin midodrine at 2.5-5 mg orally three times daily and increase to a maximum dose of 15 mg three times daily. Titrate to a MAP increase of at least 15 mm Hg; begin octreotide at 100 µg subcutaneously three times daily and increase to a maximum dose of 200 µg subcutaneously three times daily or begin octreotide at a 25-µg IV bolus and continue a rate of 25 µg/hr or Norepinephrine—0.1-0.7 µg/kg/min as an IV infusion. Increase by 0.05 µg/kg/min every 4 hr and titrate to a MAP increase of at least 10 mm Hg Duration of vasopressor treatment is generally a maximum of two weeks until reversal of hepatorenal syndrome or liver transplantation Evaluate patient for liver transplantation *Not available in the United States. Data from references 69, 74, 85, 86, 89, 91, 93-94, 97-124. ACEIs, angiotensin-converting enzyme inhibitors; ARBs, angiotensin receptor blockers; IV, intravenous; MAP, mean arterial pressure; NSAIDs, nonsteroidal anti-inflammatory drugs; SBP, spontaneous bacterial peritonitis.

midodrine, octreotide, and albumin, and norepinephrine and albumin, have been developed and studied. Terlipressin is an intravenously administered selective vasopressin 1 receptor agonist vasoconstrictor that has been used in Europe and is under review by the FDA for the treatment of type 1 HRS.95,96 It has been evaluated in approx­ imately 330 patients included in four randomized, con­ trolled trials and two meta-analyses.69,97-109 In two multicenter studies of patients with type 1 HRS, terlipressin (given in two different doses in the two studies) in combination with albumin improved serum creatinine levels relative to albumin alone (30% to 43% vs. 8% to 13%), although sur­ vival was not significantly different in the two groups.106,107 In addition, in one study,106 terlipressin was associated with a significantly increased rate of cardiovascular complica­ tions compared with albumin alone; therefore, this agent should only be used with close monitoring. Finally, the response to terlipressin in patients with type 1 HRS appeared to be better in patients with less severe renal dysfunction at baseline, thus supporting the early initiation of therapy. These studies indicate that administration of terlipressin in combination with albumin can improve renal function in HRS, although the optimal dose and duration of therapy are not defined. Midodrine, an orally administered α1-adrenergic agonist, and octreotide, a somatostatin analog that inhibits endoge­

Chapter 92  Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome nous vasodilators,110,111 have been used in combination with albumin for type 1 HRS in three small nonrandomized studies. In two studies, treatment with midodrine, titrated to cause a rise in mean arterial blood pressure, was associ­ ated with improved serum creatinine levels and improved survival compared with no treatment and was associated with few major side effects.112-114 This regimen has the advantage of ease of administration and appears to have a favorable safety profile; however, its efficacy has not been established in randomized controlled trials. Norepinephrine, a widely available intravenously admin­ istered α1-adrenergic agonist, in combination with albumin, has been proposed as an alternative to terlipressin on the basis of two small pilot studies.115,116 In one study of 22 patients with type 1 or 2 HRS, norepinephrine appeared to be as effective and safe as terlipressin. Significant cardio­ vascular side effects, however, have been reported with the use of norepinephrine to treat HRS, and whether the efficacy and safety of norepinephrine are similar to those of terli­ pressin have not been fully defined.

Radiologic and Surgical Therapy

Transjugular Intrahepatic Portosystemic Shunt Insertion of a transjugular intrahepatic portosystemic shunt (TIPS) creates a portosystemic shunt and lowers portal venous pressure, thereby decreasing venous pooling in the splanchnic circulation and increasing venous return. TIPS is effective for the treatment of diuretic-resistant ascites, a precursor to type 2 HRS.117 Four pilot studies have evalu­ ated the use of TIPS in nontransplant candidates with type 1 or 2 HRS.118-121 In these studies, serum creatinine levels declined, sodium excretion increased, and neurohumoral responses improved after TIPS, although survival was not affected. The major benefit was seen in patients with type 2 HRS. In one study, the use of midodrine, octreotide, and albumin followed by TIPS appeared to be effective in a small cohort of patients with type 1 HRS. An important limitation of the use of TIPS in HRS is the potential to worsen hepatic function. Therefore, important questions regarding the safety and benefit of TIPS in HRS remain. Liver Transplantation Liver transplantation is the only therapeutic modality that has the potential to reverse both liver dysfunction and HRS and should be considered in any patient found to have HRS.65,66,71,122 Rates of postoperative complications and inhospital mortality are higher in patients transplanted with HRS than in those without HRS,75,122 and up to 35% of patients with HRS require long-term renal replacement therapy.66 Still, the three-year survival rate of patients trans­ planted with HRS is approximately 60% compared with 70% to 80% for patients transplanted without HRS. The duration and degree of renal dysfunction preoperatively may be independent predictors of survival, and patients who require hemodialysis carry a mortality risk that is 1.77 times higher than that of patients who do not need dialy­ sis.122-124 In one study, patients with HRS who responded to treatment with a vasopressin analog prior to liver trans­ plantation had outcomes similar to those of patients who underwent liver transplantation without HRS,94 a finding that supports the use of such therapy as a bridge to liver transplantation. Larger trials are needed to confirm this observation.

Other Therapies

Extracorporeal albumin dialysis with MARS is an experi­ mental therapeutic modality that enhances the removal of water-soluble and albumin-binding toxins from the circu­ lation.125 The results of one small randomized trial have

supported the use of MARS to improve serum creatinine levels and survival rates in patients with HRS, although larger studies are needed to confirm these findings.126,127 Several other vasoconstrictive medical therapies, includ­ ing dopamine and octreotide in combination with albumin, have not improved the outcome of HRS, and in one study use of a nonselective endothelin receptor antagonist to inhibit intrarenal vasoconstriction in HRS proved deleterious.95

HEPATOPULMONARY SYNDROME AND PORTOPULMONARY HYPERTENSION Cirrhosis and portal hypertension are accompanied by alter­ ations in the vascular beds of multiple organ systems. In the pulmonary circulation, two distinct clinical entities, termed the hepatopulmonary syndrome (HPS) and portopulmonary hypertension (PPH), have been recognized. HPS occurs when pulmonary microvascular alterations impair gas exchange and is found in up to 30% of patients evaluated for liver transplantation.128-130 PPH occurs when vasocon­ striction and remodeling in resistance vessels increase pul­ monary arterial pressures and is found in as many as 5% of patients with cirrhosis. The mechanisms whereby these two entities develop are incompletely characterized, although they occur in similar clinical settings and may share patho­ genic pathways. The presence of HPS or PPH increases mortality in affected patients. No effective medical thera­ pies are available for HPS, although liver transplantation reverses the syndrome in most patients. Medical therapies that improve pulmonary hemodynamics in patients with PPH have become available, but the specific role of liver transplantation in PPH is not clearly defined.131-138

PATHOPHYSIOLOGY Hepatopulmonary Syndrome

HPS is characterized by microvascular alterations and dila­ tation in the precapillary and capillary pulmonary arterial circulation. In human HPS, the production of vasodilatory substances within the pulmonary vasculature, most notably NO, is increased. Although increased circulating and pul­ monary NO levels appear to be features of human HPS, improvement in oxygenation in response to acute inhibition of NO is variable,139-142 and HPS may take more than one year to resolve after liver transplantation in advanced cases.135 These findings suggest that other vasoactive media­ tors or angiogenesis in the pulmonary microvasculature may contribute to vascular alterations and hypoxemia in human HPS. In experimental HPS induced by bile duct ligation in the rat, pulmonary NO overproduction has also been observed and is triggered by a series of pathophysiologic events. During the onset of pulmonary vascular alterations, increased biliary production and release of endothelin-1143 in conjunction with shear stress induces pulmonary micro­ vascular endothelin-B receptor overexpression, which leads in turn to endothelin-1–mediated endothelial NO synthase (eNOS)-derived NO production.144,145 As HPS progresses, tumor necrosis factor-α (TNF-α) levels rise as a result of bacterial translocation, which leads to adherence of macro­ phages in the pulmonary microvasculature and inducible NO synthase (iNOS)–derived NO production and heme oxygenase-1–derived carbon monoxide production.146-150 Endothelin receptor antagonists and inhibition of NOS, bac­ terial translocation, TNF-α, and heme oxygenase all improve experimental HPS.148-151

1549

1550

Section IX  Liver Cirrhosis

Inflammation, bacterial translocation, and vasoactive mediators

Hepatocyte/cholangiocyte injury ↑ TGF-β → ↑ ET1

↑ TNF-α Endotoxemia

Monocyte adhesion/activation ↑ iNOS ↑ HO-1

↑ eNOS ↑ VEGF

Vasodilatation Angiogenesis

Hypoxemia

Hepatopulmonary syndrome Figure 92-3.  Pathophysiology of hepatopulmonary syndrome. eNOS, endothelial nitric oxide synthase; ET1, endothelin-1; HO-1, heme oxygenase; iNOS, inducible nitric oxide synthase; TGF-β, transforming growth factor-β; TNF-α, tumor necrosis factor-α; VEGF, vascular endothelial growth factor.

Studies have found that experimental HPS is accompa­ nied by enhanced pulmonary vascular endothelial growth factor (VEGF) expression and angiogenesis.151 Both antian­ giogenic peptides and pentoxifylline, a phosphodiesterase inhibitor with TNF-α antagonist properties, have been shown to inhibit angiogenesis and decrease the severity of HPS. Two small studies of pentoxifylline in human HPS have reported conflicting results, and the role of angiogen­ esis and angiogenesis inhibition in human HPS has yet to be defined (Fig. 92-3).152-155

Portopulmonary Hypertension

The mechanisms whereby PPH develops are poorly under­ stood. Histologically, PPH shares the characteristic features of other forms of pulmonary arterial hypertension (PAH): medial proliferation and hypertrophy, plexiform arterio­ pathy, and in situ vascular thrombosis.156-158 The precise role of portal hypertension in this process is not clear, however, and whether PPH shares pathophysiologic mecha­ nisms with PAH is unknown. Studies have found PPH, like PAH, to be more common in women than men.159 In addition, endothelin-1 levels are increased and thought to contribute to vascular alterations in PAH, a finding that has also been observed in cirrhotic patients with PPH compared with cirrhotic patients without PPH.160 By con­ trast, genetic polymorphisms in serotonin metabolism, which appear to increase vascular tone in a subset of patients with PAH, have not been found in patients with PPH.161 Nevertheless, the observation that therapies used for PAH also appear to be effective for PPH supports the notion that downstream effector mechanisms are similar in the two disorders.

CLINICAL FEATURES AND DIAGNOSIS Hepatopulmonary Syndrome

HPS is defined as a widened age-corrected alveolar-arterial oxygen gradient (AaPo2) on room air in the presence or absence of hypoxemia (AaPo2 = 15 mm Hg, or 20 mm Hg in patients older than 64 years) as a result of intrapulmonary vasodilation. HPS can be graded, according to a task force, on the basis of the degree of hypoxemia, as follows: mild (Pao2 ≥ 80 mm Hg); moderate (Pao2 = 61 to 80 mm Hg), severe (Pao2 = 50 to 60 mm Hg), and very severe (Pao2 < 50 mm Hg).162 The frequency of HPS ranges from 10% to 35% in patients with cirrhosis who undergo evaluation for liver transplantation.128,131,136 In addition, HPS has been found to increase mortality significantly in patients with cirrhosis. Moreover, survival of patients with HPS after liver transplantation is diminished compared with those without HPS, particularly when hypoxemia is severe.163 Patients with HPS present most commonly with respira­ tory complaints in the setting of chronic liver disease. Other specific pulmonary abnormalities may accompany certain liver diseases and cause respiratory symptoms. These abnormalities include panacinar emphysema in patients with α1-antitrypsin deficiency, pulmonary granulomas and interstitial fibrosis in those with primary biliary cirrhosis and sarcoidosis, and characteristic lung disease in patients with cystic fibrosis. Occasionally, HPS may be the initial manifestation of cirrhosis, and it also may be found in the setting of noncirrhotic and posthepatitic portal hyper­ tension, ischemic hepatitis, and chronic hepatitis in the absence of confirmed cirrhosis. A syndrome similar to HPS has been described in children with congenital abnor­ malities that divert hepatic blood from the pulmonary circulation.164-166 Classic clinical manifestations of HPS include platypnea (dyspnea worsened by an erect position and improved by a supine position), orthodeoxia (exacerbation of hypoxia and hypoxemia in an upright position), an insidious onset and slow progression of dyspnea, clubbing, and distal cyano­ sis.128,131,167 Although clubbing and hypoxemia (partial pres­ sure of oxygen < 60 mm Hg) in patients with liver disease in the absence of intrinsic cardiopulmonary disease are highly suggestive of HPS,131 other clinical features are not reliable for detecting HPS, and many patients, particularly those with early HPS, are asymptomatic or have symptoms only on exertion. Cough has also been described as a pre­ senting symptom of HPS.168 In most studies, the presence or severity of HPS does not appear to correlate with the degree of hepatic dysfunction.135 The diagnosis of HPS requires a high degree of clinical suspicion, measurement of arterial blood gases, detection of intrapulmonary shunting, and exclusion of intrinsic cardiopulmonary disease as the cause of hypoxemia. The most sensitive test for the diagnosis of intrapulmonary shunting is contrast echocardiography,169 which is per­ formed by peripheral intravenous injection of agitated saline to produce microbubbles that are visualized during transthoracic echocardiography. Normally, microbubbles travel from the right ventricle to the lungs, where they are absorbed and do not reach the left ventricle. In patients with intracardiac shunting, microbubbles reach the left ventricle early (within one to three cardiac cycles after injection). In patients with intrapulmonary shunting, microbubbles reach the left ventricle in a delayed fashion (three to six cardiac cycles after injection). Up to 60% of patients with cirrhosis have intrapulmonary vasodilatation on contrast echocardiography, but only a subset of patients have sufficient vasodilatation to cause abnormal arterial blood gas results and HPS.

Chapter 92  Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome O2 saturation ≥96%

<96%

TTE

Contrast TTE

Positive

ABGs Exclude lung disease

Negative

ABGs Exclude lung disease

Hypoxemia HPS Figure 92-4.  Approach to screening for hepatopulmonary syndrome in potential candidates for liver transplantation. ABGs, arterial blood gases; TTE, transthoracic echocardiography.

In patients with pulmonary symptoms and hypoxemia who are found to have intrapulmonary shunting, intrinsic cardiopulmonary disease should be excluded. Chest radio­ graphy or computed tomography (CT) and pulmonary function tests are generally performed in patients being con­ sidered for liver transplantation. If potentially reversible cardiopulmonary disorders are detected, treatment is under­ taken, and the assessment of oxygenation is repeated. In addition to its use in the evaluation of patients for pulmo­ nary disease, high-resolution chest CT has been reported to detect peripheral pulmonary vascular dilatations in a small cohort of patients with HPS.139,170 The diagnostic usefulness of chest CT for the diagnosis of HPS, however, is not defined. In the small subset of patients found to have advanced hypoxemia (Pao2 < 60 mm Hg) and both intrapulmonary shunting and significant cardiopulmonary disease, a techne­ tium-labeled macroaggregated albumin scan may confirm that HPS is contributing to the gas exchange abnormalities. In this test, the shunting of intravenously administered 99m Tc-labeled macroaggregated albumin (20 µm in diameter) through the pulmonary vasculature is quan­tified by measur­ ing activity and uptake in both the lungs and brain.169,171,172 In one study, the scan result was positive (>6% shunt fraction) in patients with HPS and a Pao2 of <60 mm Hg, but not in patients with chronic obstructive pulmonary disease and a similar degree of hypoxemia.171 Screening algorithms for HPS have been evaluated, particularly in patients who are candidates for liver trans­ plantation. One validated approach uses pulse oximetry screening (Po2 < 96%) to detect patients with a Pao2 < 70 mm Hg (sensitivity 100%, specificity 88%; Fig. 92-4). This subgroup of patients (~30% of all liver transplantation candidates) is then targeted for arterial blood gas analysis, contrast echocardiography, and evaluation for intrinsic car­ diopulmonary disease, thereby limiting full evaluation for HPS to those patients likely to have at least moderate HPS.173,174 Currently, however, no practice guidelines for screening for HPS in liver transplantation candidates have been developed.

Portopulmonary Hypertension

PPH is defined as the development of PAH in the setting of portal hypertension. The diagnostic criteria for PPH include the presence of PAH as defined by the World Health Orga­ nization: mean arterial pulmonary pressure (mPAP) > 25 mm Hg at rest or 30 mm Hg with exercise; pulmonary

capillary wedge pressure < 15 mm Hg; and pulmonary vas­ cular resistance > 240 dynes • s • cm−5) occurring in the setting of portal hypertension (splenomegaly, thrombocyto­ penia, portosystemic shunts, or portal vein hemodynamic abnormalities).129,175,176 PPH is generally graded according to the degree of elevation in mPAP, which correlates with the mortality risk associated with liver transplantation and influences decisions regarding therapy.172 Mild PPH (mPAP = 25 to 35 mm Hg) is not associated with an increased operative risk for liver transplantation and may not require medical therapy. Moderate PPH (mPAP = 35 to 50 mm Hg) is associated with an increased operative risk for liver trans­ plantation and is an indication for medical therapy. Severe PPH (mPAP > 50 mm Hg) has a prohibitive operative mor­ tality and is managed with medical therapy. PPH has been found in as many as 6% of cirrhotic patients who are evalu­ ated for liver transplantation, and outcomes are worse than in cirrhotic patients without PPH. The most common symptom associated with PPH is exertional dyspnea; other nonspecific symptoms such as orthopnea, fatigue, chest pressure, syncope, edema, and lightheadedness also may occur.176 Characteristic physical examination features of PAH, including an elevated jugular venous pressure, loud second pulmonic heart sound, murmur of tricuspid regurgitation, and lower extremity edema, have been reported but are not sufficiently sensitive nor specific to be diagnostically useful. In cirrhotic patients, peripheral edema out of proportion to the degree of ascites should prompt consideration of right ventricular dysfunc­ tion secondary to pulmonary hypertension. In a number of screening studies, most cirrhotic patients with significant PPH were asymptomatic.133,177 The diagnosis of PPH warrants a high degree of clinical suspicion, and all patients considered for liver transplanta­ tion as well as patients with suggestive symptoms or physi­ cal findings should be evaluated for PPH. Transthoracic echocardiography is the recommended screening test because it evaluates right-sided cardiac function and allows an estimation of right ventricular systolic pressure by evalu­ ating the tricuspid regurgitant jet.178 In addition, other causes of elevated right-sided cardiac pressures, including secondary pulmonary hypertension, volume overload, and a hyperdynamic circulation, should be considered and assessed. Methods for estimating right ventricular systolic pressure vary among centers but, in general, in the absence of significant pulmonary artery stenosis, an estimated right ventricular systolic pressure higher than 40 mm Hg or the presence of right ventricular abnormalities support further evaluation for PPH. The absence of both these echo­ cardiographic findings essentially excludes PPH from the differential diagnosis.133 In all patients found to have echo­ cardiographic features suggestive of PPH, pulmonary artery catheterization should be performed to establish the diagnosis and assess the severity of PPH. Findings on pulmonary artery catheterization are useful for distin­ guishing volume overload and a hyperdynamic circulation from PPH.

TREATMENT Hepatopulmonary Syndrome

Treatment options for HPS are limited. Currently, no estab­ lished medical therapies exist, although case reports and small case series have suggested that some treatments may improve oxygenation. Therefore, patients with wellpreserved hepatic synthetic function who have hypoxemia are generally treated symptomatically until oxygenation worsens sufficiently to permit listing for liver transplanta­

1551

1552

Section IX  Liver tion based on an exception to the Model for End-stage Liver Disease (MELD) score (see Chapter 95). Liver transplanta­ tion reverses HPS in most affected patients, although mor­ tality after liver transplantation appears to be higher in patients with HPS than in those without HPS.163,179-181

patients with HPS, MELD exception points may be given to patients with HPS and a resting Pao2 of <60 mm Hg, thereby increasing their priority status for transplantation (see Chapter 95).179-181

Medical Therapy A number of agents have been tried in patients with HPS empirically or on the basis of data from experimental models in uncontrolled trials, case series, and case reports. Among the compounds studied are garlic preparations, pentoxifylline, aspirin, N-acetylcysteine, glucocorticoids, antimetabolites (cyclophosphamide), antibiotics (norfloxa­ cin), somatostatin, and almitrine (Duxil, a respiratory stimulant).139,172,182-184 Of these agents, only two have been evaluated in small clinical studies. An open-label trial of a garlic preparation reported an increase in Pao2 of approximately 10 mm Hg in 40% of patients when used for periods of at least six months.182 Two small studies have evaluated pentoxifylline as a treatment for HPS; one found improvement in oxygenation, and the other observed dose-limiting gastrointestinal toxicity and no clinical improvement.185,186 The usefulness of oxygen supplementation in patients with HPS has not been studied, although oxygen supple­ mentation is a common long-term intervention used for patients with resting hypoxemia (Pao2 < 60 mm Hg) or exercise-induced oxygen desaturation. In general, respira­ tory symptoms improve with oxygen, and tissue oxygen­ ation within the liver may be improved.

Treatment for PPH has changed significantly during the 2000s because of the availability of oral vasodilators devel­ oped for use in patients with PAH. In general, liver trans­ plantation is contraindicated in patients with moderate to severe PPH because of increased perioperative mortality resulting from poor right-sided cardiac function. The ability to lower pulmonary arterial pressures and decrease pulmo­ nary vascular resistance with medications may decrease perioperative complications and has raised the possibility that subsequent liver transplantation may result in the reso­ lution of PPH.

Interventional Radiologic Therapy Two radiologic techniques—TIPS,187 to lower portal pres­ sure, and pulmonary angiography with embolization, to occlude areas of intrapulmonary shunting,172 have been reported to be useful in HPS. These invasive approaches have been described only in case reports or small retro­ spective analyses, and whether either treatment reliably improves oxygenation is not clear. Therefore, placement of a TIPS specifically to treat HPS in the absence of other indications is not recommended.187,188 Pulmonary angi­ ography has been considered in patients with severe hypox­ emia to identify focal arteriovenous shunting that might be diminished with embolization. Little evidence to support such an approach is available, however, and arteriovenous shunting sufficient to cause hypoxemia and amenable to embolization may be detected by high-resolution CT.172 Liver Transplantation Liver transplantation reverses HPS in as many as 80% of affected patients.139 Hypoxemia may persist after transplan­ tation, however, particularly when severe before transplan­ tation, and may require more than one year to resolve. Mortality also appears to be higher in patients with HPS who undergo liver transplantation than in patients trans­ planted without HPS, although this observation has not been completely uniform.181 The increase in postoperative mortality is related in part to the severity of HPS, with one prospective study demonstrating that patients with pro­ found hypoxemia (Pao2 < 50 mm Hg) and marked intrapul­ monary shunting (shunt fraction > 20%) had a marked increase in mortality.135,136,163,179 In addition, unique compli­ cations such as transient worsening of hypoxemia following liver transplantation, the development of pulmonary hyper­ tension, and embolic cerebrovascular events have been observed and may contribute to adverse postoperative out­ comes. On the basis of the relationship between the severity of hypoxemia and poor outcomes in nontransplanted

Portopulmonary Hypertension

Medical Therapy The therapy of PPH is largely empirical. In general, diuretics are used for volume overload, supplemental oxygen is provided if hypoxemia is present, and anticoagulation may be appropriate in the subset of patients with PPH who do not have hepatic decompensation, marked coagulopathy, or gastroesophageal varices. In one study, withdrawal of β-adrenergic blocker therapy improved right-sided cardiac function in patients with PPH, and some clinicians advocate treating varices with band ligation if feasible and withdrawing medical therapy for varices in patients with PPH.189 Despite data from trials in patients with PAH, no ran­ domized controlled studies have evaluated vasodilator therapy in patients with PPH. Nevertheless, case reports and series have evaluated prostacyclin analogs, several of which require complex intravenous administr­ation (e.g., epoprostenol, iloprost, treprostinil),138,177,190,191 oral endothelin receptor antagonists including bosentan (a mixed endothelin-A and -B receptor antagonist),192-194 ambrisentan (an endothelin-A blocker),65 and oral phosphodiesterase-5 inhibitors (e.g., sildenafil).195-197 Bosen­ tan hepatotoxicity has been described uncommonly and is potentially mediated by inhibitory effects on the cana­ licular bile salt export protein. Close monitoring of liver biochemical test levels and use only in patients with well-compensated liver disease is appropriate. Use of ambrisentan in patients with PPH is increasing, although few data about its efficacy are available. Sildenafil generally has been well tolerated in patients with PPH. The magni­ tude of improvement in pulmonary arterial pressures has been modest with oral agents, and some clinicians use endothelin receptor antagonists and phosphodiesterase-5 inhibitors in combination, although the beneficial effects are limited. Liver Transplantation Traditionally, moderate to severe PPH has been a contrain­ dication to liver transplantation, particularly when the mPAP is higher than 50 mm Hg.198 With the availability of agents to lower pulmonary pressures, consideration has been given to improving pulmonary hemodynamics with medical therapy followed by liver transplantation to attempt to resolve PPH.199 Three case series have described such an approach.190,200,201 Although results have varied among centers, some patients with PPH appear to have resolved after liver transplantation. Whether liver transplantation is an effective treatment for PPH and whether subgroups of patients may derive the specific benefit from such an approach have not been clarified.

Chapter 92  Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome

Although systemic hemodynamic changes in cirrhotic patients have been observed for more than 100 years, not until the 1950s were patients with alcoholic cirrhosis found to have a so-called hyperdynamic circulation—decreased arterial blood pressure, decreased peripheral resistance, and increased cardiac output.202 Later studies described vascular hyporesponsiveness to vasoconstrictors in cirrhotic patients that was attributed to effects of alcohol on the heart and termed alcoholic cardiomyopathy.203 Subsequently, specific vascular and cardiac abnormalities have been found in human and animal models of cirrhosis, independent of alcohol ingestion, and have led to the concept that cirrhosis itself triggers cardiac dysfunction; this has been termed cirrhotic cardiomyopathy (CCM).204

myocytes have been observed.214 Whether the severity of hepatic synthetic dysfunction correlates with the degree of cardiac dysfunction is not clear.215 The impaired ventricular response to stress and exercise observed in human and animal models of cirrhosis83,204,208,210 has been attributed to impaired β-adrenergic signaling pathways that lead to sub­ normal chronotropic and contractile responses,216-218 as well as to cardiomyocyte dysfunction resulting from overproduc­ tion of NO, carbon monoxide, and endocannabinoids.219-224 Electrophysiologic abnormalities, most notably prolonga­ tion of the QT interval (QTc), have also been observed in patients with CCM.225,226 The degree of prolongation of the QTc appears to correlate with the severity of liver disease and may contribute to the dissociation between electrical and mechanical events in the heart and to cardiac dysfunc­ tion.227 These abnormalities appear to improve after liver transplantation.

PATHOPHYSIOLOGY

CLINICAL FEATURES AND DIAGNOSIS

CIRRHOTIC CARDIOMYOPATHY

Although diagnostic criteria for CCM have not been established, three major pathophysiologic features have been observed: (1) structural and functional ventricular abnormalities; (2) an abnormal ventricular response in the presence of pharmacologic, physiologic, or surgical stress; and (3) cardiac electrophysiologic abnormalities (Fig. 92-5).83,204-207 The major structural and functional ventricular abnor­ malities, found in histologic and echocardiographic studies, are left ventricular hypertrophy and diastolic dysfunc­ tion.83,208-211 The structural changes have been attributed to a hypertrophic response to the hyperdynamic circulation. Diastolic alterations may be precipitated or worsened by the presence of a significant amount of ascites.212,213 In addition, impaired systolic function and histologic injury to cardio­

Cirrhosis

Inflammation, bacterial translocation, and vasoactive mediators Splanchnic and peripheral vasodilatation ↓ Effective circulatory volume ↓ Blood pressure

↑ Cardiac output

Cardiac structural, functional, and electrophysiologic abnormalities Ventricular hypertrophy β Adrenergic receptor dysfunction ↑ QTc

Diastolic or systolic dysfunction

Cirrhotic cardiomyopathy Figure 92-5.  Pathophysiology of cirrhotic cardiomyopathy. QTc, rate-­ corrected QT interval.

Because CCM typically becomes clinically detectable under circumstances of stress, the diagnosis is difficult to make and the process may not be recognized. Overt cardiac dys­ function may occur after common clinical interventions in cirrhosis. For example, a prospective multicenter analysis has shown that more than 10% of patients who undergo TIPS placement for refractory ascites exhibit signs of heart failure when compared with patients treated by repeated large-volume paracentesis.228 Subsequent studies have revealed that both the severity of liver disease (as measured by the MELD score) and diastolic dysfunction 28 days after the procedure are independent predictors of death in patients with cirrhosis who are treated with a TIPS.229 Simi­ larly, in a case series of patients who underwent liver trans­ plantation, 47% were found to have radiologic evidence of pulmonary edema within the first 24 hours after transplan­ tation. Volume replacement was considered unlikely to be the sole culprit, suggesting that cardiac dysfunction may have played a role.230,231 Finally, in prospective studies, a lower cardiac output has been found in patients in whom HRS developed than in those in whom HRS did not develop, thereby supporting the role of cardiac dysfunction in HRS.80,81 No precise diagnostic criteria for CCM have been estab­ lished, because no baseline clinical, imaging, or biochemi­ cal findings have been found to predict the development of overt cardiac dysfunction under stress definitively.232-236 The presence of QTc prolongation and echocardiographic evi­ dence of diastolic dysfunction are readily detected but do not appear to correlate with the risk of development of cardiac dysfunction under stress. Serum levels of markers of cardiac dysfunction—brain natriuretic peptide, atrial natriuretic peptide, and troponin I—are elevated in patients with cirrhosis and appear to correlate with QTc prolonga­ tion, diastolic dysfunction, and the severity of liver disease. Evaluation of cardiac function under stress conditions with echocardiography or ventriculography has been reported, and cardiac function may be impaired but does not correlate clearly with subsequent clinical cardiac dysfunction. Further work is needed to define the role of serum markers and stress testing in the diagnosis of CCM.235,236 In cirrhotic patients found to have evidence of congestive heart failure, alternative causes of cardiac dysfunction, including coro­ nary artery disease, valvular abnormalities, and other causes of cardiomyopathy, should be excluded.

TREATMENT

Therapy of volume overload in patients suspected of having CCM includes standard supportive measures and diure­

1553

1554

Section IX  Liver sis.237 Use of preload and afterload reducing agents should be considered with caution, because these agents may worsen hypotension in the setting of underlying systemic vasodilatation.37,204 Inotropic and chronotropic agents do not appear to be of benefit.205 In patients without overt heart failure but with electrocardiographic or echocardiographic features consistent with CCM, chronic use of aldosterone antagonists has been shown to improve echocardiographic features marginally,237,238 and short-term administration of a noncar­dioselective b-adrenergic antagonist has been shown to improve QTc prolongation.239 Whether these agents prevent or ameliorate cardiac dysfunction in response to stress is unknown. The effects of liver transplantation on cardiac abnor­ malities in patients with cirrhosis have not been fully char­ acterized. A small study has demonstrated that resting echocardiographic and stress radionuclide ventriculo­ graphic abnormalities found in patients with cirrhosis were improved or reversed after liver transplantation.212,240 Further data are needed to determine whether CCM reverses after liver transplantation.

ENDOCRINE DYSFUNCTION Cirrhosis has been linked to abnormalities in the endocrine system, including abnormal sex hormone metabolism, thyroid disease, osteoporosis, and, most recently, adrenal insufficiency.

ADRENAL INSUFFICIENCY

The recognition that adrenal insufficiency worsens out­ comes in sepsis, a syndrome that shares physiologic abnor­ malities seen in liver failure, has led to evaluation of adrenal dysfunction in patients with liver disease. Five studies have demonstrated the presence of relative adrenal insufficiency in up to 69% of critically ill patients with liver disease.241-245 The presence of relative adrenal insufficiency was associ­ ated with greater hemodynamic instability and increased mortality. In one study,245 glucocorticoid therapy improved survival, but, in two others,242,243 such therapy was associ­ ated with increased mortality secondary to nosocomial and opportunistic infections. Inadequate hepatic cholesterol production as a result of liver disease may predispose patients to impaired adrenal cortisol synthesis from choles­ terol during stress. Currently, no standardized diagnostic criteria have been developed for relative adrenal insuffi­ ciency, and whether a subset of critically ill patients with liver disease will benefit from glucocorticoid treatment remains an area of active investigation.

GONADAL DYSFUNCTION

Historical data have suggested a high frequency (70% to 80%) of central and peripheral hypogonadism in cirrhotic patients. Hypogonadism in this setting is associated with a decreased concentration of free or bioavailable testosterone in direct proportion to the degree of liver dysfunction.246,247 Sex hormone-binding globulin (SHBG), which binds testos­ terone and 17b-estradiol in serum, has a lower affinity for estrogens than testosterone. In patients with cirrhosis, ele­ vated concentrations of SHBG (with a resulting shift in the balance of hormones in favor of estrogens) and a decreased production of dehydroepiandrosterone sulfate (a precursor of androgenic hormones) may account for the “feminization syndrome” seen in male patients with cirrhosis.247-249 Acute and chronic alcohol consumption is responsible for direct toxic effects on Leydig cells and for alterations of the hypothalamic-pituitary-gonadal axis. These alterations

include a decrease in the serum concentration of luteinizing hormone and decreased responsiveness to gonadotropinreleasing hormone.250,251 Spironolactone, which frequently is used to treat fluid overload, causes painful gynecomastia by displacing androgen from its receptor and binding protein and by increasing estradiol production and testos­ terone clearance.252 Topical testosterone appears to improve muscle strength and survival in patients with chronic allograft failure after liver transplantation, but whether tes­ tosterone supplementation improves the symptoms and signs related to hypogonadism in pretransplantation cir­ rhotic patients is unknown.253

THYROID DYSFUNCTION

A number of thyroid abnormalities, including increased thyroid volume and decreased serum levels of free triiodo­ thyronine, have been found in patients with cirrhosis. These alterations appear to correlate with the severity of liver disease, and the presence of thyroid disease may be a pre­ dictor of decreased survival.254-257 In patients with hepatitis C and autoimmune liver disease, the incidence of hypothy­ roidism and autoimmune thyroid disease is increased.258-260 The detection of hepatitis C virus in thyroid cells raises the possibility of direct viral cytotoxicity.261 Interferon-based treatments for viral hepatitis may cause thyroiditis and have been implicated in both hyper- and hypothyroidism in 10% to 15% of treated patients (see Chapter 79).262 Whether routine screening for thyroid disease influences survival or quality of life is unknown.

BONE DISEASE

The frequency of osteoporosis among patients with all causes of chronic liver disease ranges from 12% to 55%.263 Potential risk factors include cholestasis, a maternal history of hip fracture, progression of liver disease, alcohol con­ sumption, lower body mass index, oral glucocorticoid use for more than three months, and older patient age. Women with primary biliary cirrhosis have a four-fold higher risk of developing osteoporosis and a two-fold higher risk of bone fractures than age-matched controls.264 Ursodeoxycho­ lic acid treatment does not influence bone density in patients with primary biliary cirrhosis,265 but lower glucocorticoid use and better nutrition appear to decrease the frequency of osteoporosis.266 Bone mineral density is an appropriate screening tool for osteoporosis in patients with cirrhosis (especially primary biliary cirrhosis and primary sclerosing cholangitis) and those who require more than three months of glucocorticoid therapy.266-268 Treatment of osteoporosis in chronic liver disease is based on studies of postmenopausal women and is an area of ongoing investigation. The use of calcium and vitamin D and of therapeutic bisphosponates improves bone mineral density and appears not to have significant side effects (see Chapter 89).268,269

COAGULATION DISORDERS Cirrhosis is well recognized to be associated with a bleeding diathesis because of the presence of a prolonged prothrom­ bin time and thrombocytopenia. Studies in the 2000s have suggested, however, that the interplay among abnormalities in both pro- and anticoagulant factors may result in increased bleeding and hypercoagulability. The precise mechanisms for these clinical events are not fully characterized.270 In cirrhosis, the progressive loss of hepatocytes leads to decreased synthesis of procoagulant factors, including vitamin K–dependent factors (II, VII, IX, X), factor V, and factor XI. The severity of clotting abnormalities, measured

Chapter 92  Hepatic Encephalopathy, Hepatorenal Syndrome, Hepatopulmonary Syndrome by the prothrombin time (PT), activated partial thrombo­ plastin time (aPTT), and international normalized ratio (INR) increase as liver disease progresses,271 and the increase in INR predicts survival.272 A prolonged PT alone is not generally considered to be a major risk factor for spontane­ ous bleeding in patients with cirrhosis but does increase the severity of bleeding when it occurs.273 Common practice is to administer fresh frozen plasma (FFP), vitamin K, and occasionally recombinant factor VIIa to correct coagulopa­ thy in patients with chronic liver disease, particularly in the setting of bleeding or prior to invasive procedures.274 Despite current practice, however, clinical evidence that vitamin K, FFP, or recombinant factor VIIa administration reduces the severity of variceal bleeding, for example, is not strong.275-277 In addition, the volume of FFP (>6 units) required to achieve a clinically significant reduction in the PT in patients with cirrhosis has been associated with a significant increase in the risk of acute lung injury and volume overload.278,279 In one trial of patients who underwent laparoscopic liver biopsy, recombinant factor VIIa therapy normalized the PT, and no patients required blood transfusions at the time of biopsy.280 Nevertheless, no placebo-controlled trial of recombinant factor VIIa in cirrhotic patients undergoing invasive procedures has been performed. One concern about use of the INR in patients with cirrhosis is the sub­ stantial interlaboratory variability with use of the standard international sensitivity index (ISI) to normalize varying sensitivities of thromboplastin reagents.281 This variability may result in as much as a 25% difference in mean INR for a single patient sample assayed with different reagents. Adjusting the ISI calibration by using samples from cir­ rhotic patients appears to eliminate this variation but is time-consuming and not widely used.282 Therefore, use of the INR to predict bleeding risk and to calculate the MELD score for organ allocation priority has limitations. Thrombocytopenia is also a common feature in cirrhotic patients with portal hypertension. It is associated with the presence of hypersplenism, but decreased hepatic thrombo­ poietin synthesis and, in some cases, direct bone marrow toxicity (e.g., from alcohol or hepatitis C virus) also play a role.283 In addition to quantitative abnormalities, platelet thrombin generation is impaired, particularly at platelet counts below 50,000/mL, and may contribute to impaired clot formation.284 The presence of thrombocytopenia, however, does not appear to increase the risk of bleeding in patients with cirrhosis. The data on whether the administra­ tion of platelets influences the bleeding risk in patients who undergo invasive procedures or decreases blood loss or transfusion requirements in patients with variceal hemor­ rhage are limited. Nevertheless, common clinical practice is to administer platelets to achieve a platelet count of approx­ imately 50,000/mL prior to invasive procedures and in the setting of active bleeding.278,279 Although impaired platelet function, as assessed by measuring the bleeding time, is commonly observed in patients with cirrhosis, neither prolongation of the bleeding time nor its correction with desmopressin administration influence the risk of bleeding.285,286 Dysfibrinogenemia, reflected by elevated circulating levels of D-dimer and fibrinogen degradation products, and by prolongation of the clot lysis time, is seen in up to 46% of cirrhotic patients.287,288 These abnormalities result from altered production of activators and inhibitors of fibrinoly­ sis, activation of the coagulation cascade by endotoxemia, and decreased clearance of fibrinolytic proteins in the setting of hepatic synthetic dysfunction. Dysfibrinogenemia becomes more severe as liver disease progresses and may progress to overt disseminated intravascular coagulation, further increasing the risk of bleeding. Results of tests for

D-dimer and fibrinogen degradation products are variable in cirrhosis, but no uniformly accepted standard for assess­ ing dysfibrinogenemia is available. Antihyperfibrinolytic therapy with compounds such as ε-aminocaproic acid has been used to prevent blood loss in stable cirrhotic patients and during liver transplantation, but evidence to support such an approach is limited.289,290 In addition to effects that predispose to bleeding, hepatic synthetic dysfunction in patients with cirrhosis also impairs the production of endogenous anticoagulant proteins, including protein C, protein S, antithrombin, tissue plasminogen acti­ vator, and thrombomodulin.291 These abnormalities may result in hypercoagulability and a risk for thrombosis. For example, portal vein thrombosis is a well-recognized com­ plication of cirrhosis, and case-control studies have found up to a two-fold increased risk of thromboembolic events (deep venous thrombosis and pulmonary embolism) in patients with cirrhosis and noncirrhotic liver disease com­ pared with controls.292-294 Whether a subset of these patients has underlying inherited causes of hypercoagulability that contribute to thrombosis has not been fully defined.295 In some studies, anticoagulation has appeared to be safe and of benefit for patients with noncirrhotic, nonmalignant portal vein thrombosis associated with gastroesophageal varices, but the role of anticoagulation therapy in cirrhotic portal vein thrombosis remains unclear.296 The potential role of hypercoagulability in the progression of liver disease and in some complications of cirrhosis (e.g., HRS, PPH) are areas for future investigation.297

KEY REFERENCES

Alqahtani SA, Fouad TR, Lee SS. Cirrhotic cardiomyopathy. Semin Liver Dis 2008; 28:59-69. (Ref 205.) Bass NM. Review article: The current pharmacological therapies for hepatic encephalopathy. Aliment Pharmacol Ther 2007; 25(Suppl 1):23-31. (Ref 43.) Esrailian E, Pantangco ER, Kyulo NL, et al. Octreotide/Midodrine therapy significantly improves renal function and 30-day survival in patients with type 1 hepatorenal syndrome. Dig Dis Sci 2007; 52:7428. (Ref 114.) Gaskari SA, Honar H, Lee SS. Therapy insight: Cirrhotic cardio­ myopathy. Nat Clin Pract Gastroenterol Hepatol 2006; 3:329-37. (Ref 210.) Martin-Llahi M, Pépin M-N, Guevara M, et al. Terlipressin and albumin vs. albumin in patients with cirrhosis and hepatorenal syndrome: A randomized study. Gastroenterology 2008; 134:1352-9. (Ref 106.) Mas A, Rodes J, Sunyer L, et al. Comparison of rifaximin and lactitol in the treatment of acute hepatic encephalopathy: Results of a random­ ized, double-blind, double-dummy, controlled clinical trial. J Hepatol 2003; 38:51-8. (Ref 48.) Møller S, Henriksen JH. Cardiovascular complications of cirrhosis. Gut 2008; 57:268-78. (Ref 206.) Palma DT, Fallon MB. The hepatopulmonary syndrome. J Hepatol 2006; 45:617-25. (Ref 139.) Rodríguez-Roisin R, Krowka MJ. Hepatopulmonary syndrome—a liver-induced vascular disorder. N Engl J Med 2008; 358:2378-87. (Ref 172.) Salerno F, Gerbes A, Ginès P, et al. Diagnosis, prevention and treatment of hepatorenal syndrome in cirrhosis. Gut 2007; 56:1310-18. (Ref 74.) Sanyal AJ, Boyer T, Garcia-Tsao G, et al. A randomized, prospective, double-blind, placebo-controlled trial of terlipressin for type 1 hepatorenal syndrome. Gastroenterology 2008; 134:1360-8. (Ref 107.) Swanson K, Wiesner R, Krowka M. Natural history of hepatopulmonary syndrome: Impact of liver transplantation. Hepatology 2005; 41:11229. (Ref 135.) Torregrosa M, Genesca J, Gonzalez A, et al. Role of Doppler echocar­ diography in the assessment of portopulmonary hypertension in liver transplantation candidates. Transplantation 2001; 71:572-4. (Ref 134.) Wong F, Pantea L, Sniderman K. Midodrine, octreotide, albumin, and TIPS in selected patients with cirrhosis and type 1 hepatorenal syndrome. Hepatology 2004; 40:55-64. (Ref 118.) Full references for this chapter can be found on www.expertconsult.com.

1555

CHAPTE R

93 Acute Liver Failure Robert J. Fontana

CHAPTER OUTLINE Definition  1557 Causes  1558 Acetaminophen Toxicity  1558 Idiosyncratic Drug Toxicity  1559 Viral Infections  1560 Miscellaneous Causes  1560 Indeterminate Acute Liver Failure  1560 Clinical Features  1561 Hepatic Encephalopathy and Cerebral Edema  1561 Coagulopathy and Bleeding  1562 Infection  1562 Multiple Organ Failure Syndrome  1562

Acute liver failure is a rare clinical syndrome that results from the sudden loss of hepatic parenchymal and metabolic functions and manifests as coagulopathy and encephalo­ pathy. The cause of acute liver failure varies widely through­ out the world, with drug, infectious, and immunologic diseases reported most commonly. Acute liver failure is a medical emergency associated with a high mortality rate because of the development of cerebral edema, infectious complications, and multiorgan failure. Despite advances in medical management, mortality rates in patients with acute liver failure remain high in the absence of emergency liver transplantation.

DEFINITION Acute liver failure is defined as the rapid development of hepatocellular dysfunction, specifically coagulopathy and mental status changes (encephalopathy) in a patient without known prior liver disease.1 Acute liver failure is a clinical syndrome that represents the final common pathway of severe liver injury resulting from numerous infectious, immunologic, metabolic, vascular, and infiltrative disor­ ders. The mechanism of liver injury in acute liver failure is most often severe hepatocellular necrosis, as occurs with acetaminophen toxicity or viral hepatitis. Acute liver failure can also result from severe cellular or mitochondrial dys­ function, as occurs with some forms of drug toxicity (e.g., antiretroviral agents), Wilson disease, and acute fatty liver of pregnancy.2 Acute liver failure (or fulminant hepatic failure) origi­ nally was defined by an interval of eight weeks or less between the onset of illness and appearance of encepha­ lopathy.3 In an attempt to improve the prediction of the

Prognosis  1563 Treatment  1563 Initial Evaluation and Management  1564 Encephalopathy and Cerebral Edema  1565 Coagulopathy and Bleeding  1566 Infection  1566 Multiple Organ Failure Syndrome  1566 Liver Transplantation  1567 Investigational Approaches  1567 Auxiliary Liver Transplantation  1567 Extracorporeal Liver Support  1568 Hepatocyte Transplantation  1568

prognosis and outcome, O’Grady and colleagues divided patients into three groups based on the time interval between the onset of jaundice and encephalopathy: those with hyperacute liver failure (seven days or less), those with acute liver failure (eight to 28 days), and those with sub­ acute liver failure (four to 24 weeks).4 In general, patients with hyperacute liver failure are more likely to develop cerebral edema and to recover without liver transplantation. By contrast, patients with subacute or late-onset hepatic failure are more likely to present with evidence of portal hypertension such as ascites and to have a low rate of sur­ vival without transplantation.5-7 Although the duration of illness may help predict prognosis, the overlap among patients with acute liver failure with varying presentations is great, and the duration of symptoms is largely related to the cause of liver failure. The original definition of acute liver failure (encephalopathy and coagulopathy within eight weeks of the illness onset) is used in this chapter, because this definition is the most widely used in clinical studies and in criteria for liver transplantation in the United States. The diagnosis of acute liver failure is made clinically on the basis of the physical examination (altered mental status) and supportive laboratory findings (hyperbilirubinemia, prolonged prothrombin time). Infrequently, acute liver failure may be confused with other clinical entities that are associated with jaundice, coagulopathy, and encephalopa­ thy, such as sepsis, systemic disorders with liver and brain involvement (e.g., systemic lupus erythematosus, throm­ botic thrombocytopenic purpura), and an acute decompen­ sation of chronic liver disease. In particular, sepsis and acute liver failure have similar clinical profiles, and severe sepsis is frequently accompanied by a change in mental status; in this situation, jaundice and coagulopathy may result from intrahepatic cholestasis arising from high levels

1557

1558

Section IX  Liver of proinflammatory cytokines and disseminated intravascu­ lar coagulation (DIC), respectively. Measurement of plasma factor VIII levels may help differentiate sepsis (low factor VIII level) from acute liver failure (factor VIII level generally not reduced). Alcoholic hepatitis and flares of chronic hepa­ titis B virus (HBV) infection may occasionally be mistaken for acute liver failure. In these cases, a careful review of the medical history, laboratory and imaging studies, and, in selected cases, liver biopsy findings is helpful.

CAUSES The underlying cause of acute liver failure in an individual patient is established by the patient’s history, serologic and molecular diagnostic test results, and characteristic radiologic or histologic features. The predominant cause of acute liver failure differs markedly throughout the world. In the United States and other Western countries, medications, including acetaminophen, and idiosyncratic drug toxicity are the most commonly identified causes of acute liver failure (Table 93-1).8,9 In France, Japan, and India, severe acute HBV infection is a leading cause of acute liver failure.10-12 In addition to these causes, numerous other, often rare, conditions can lead to acute liver failure (Table 93-2).

ACETAMINOPHEN TOXICITY

Acetaminophen is a dose-dependent hepatotoxin that, when ingested in excessive doses, can lead to life-threatening liver injury characterized by hypoprothrombinemia, towering aminotransferase elevations, and a normal or minimally elevated serum bilirubin level (Table 93-3). Measurement of serum acetaminophen levels is helpful in assessing the risk of hepatotoxicity following an acute overdose, but false-positive detection of acetaminophen in serum has been reported in some patients with deep jaundice if a colori­metric assay for acetaminophen is used.13 Because of the widespread availability of acetaminophen, intentional acetaminophen overdose (i.e., >10 g) is a common mode of attempted suicide, with over 60,000 cases reported each year in the United States.14 Although most patients who take an intentional overdose of acetaminophen recover, acute liver failure develops in a minority, and acetaminophen toxicity has become the leading cause of acute liver failure in both the United States and United Kingdom (Fig. 93-1).8,9 An increasing frequency of cases of unintentional acet­ aminophen overdose leading to acute liver failure has also

been reported since the 1990s.15-17 In many such “therapeu­ tic misadventures,” patients have ingested over-the-counter products containing acetaminophen as well as narcoticacetaminophen congeners prescribed for an acute medical condition. Almost 10% of U.S. patients discharged from urban emergency rooms are given a prescription for an acetaminophen-narcotic congener, but most do not receive proper instruction regarding the need to reduce the dose or discontinue the use of other acetaminophen-based analge­ sics.18 Chronic heavy alcohol consumption may lower the threshold for acetaminophen toxicity in some patients by inducing cytochrome P450 enzyme activity (see Chapter 86).19 In addition, preexisting hepatic dysfunction and resulting glutathione depletion may predispose some patients to acetaminophen toxicity.20 Most patients with an unintentional overdose of acetaminophen have ingested large doses of acetaminophen-containing products over several days, but one study has suggested that ingestion of only 4 g of acetaminophen daily can lead to mild, tran­ sient serum aminotransferase elevations in 40% of healthy volunteers.21 In the United Kingdom, restrictions on the quantity of acetaminophen dispensed, as well as blister packaging of products, were introduced in 1998 to reduce the incidence of acetaminophen toxicity. Since then, rates of hospitaliza­ tion, acute liver failure, and the need for liver transplanta­ tion for acetaminophen toxicity have declined.22,23 In the United States, black box warnings on acetaminophen prod­

Table 93-2  Uncommon Causes of Acute Liver Failure Autoimmune hepatitis (initial presentation) Eclampsia, preeclampsia Fatty liver of pregnancy Hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome Viral infections (e.g., Epstein-Barr virus, herpes simplex virus, cytomegalovirus) Hepatic ischemia (e.g., cardiogenic shock, volume depletion) Malignant infiltration (e.g., lymphoma, hematologic malignancy, metastatic lung or breast cancer) Primary graft nonfunction following liver transplantation Toxins (e.g., Amanita phalloides ingestion, sea anenome sting) Vascular abnormalities (e.g., Budd-Chiari syndrome, sinusoidal obstruction syndrome, iatrogenic occlusion of portal vein during surgery) Wilson disease (initial presentation)

Table 93-1  Various Causes of Acute Liver Failure Worldwide COUNTRY, YEAR(S) OF STUDY*

Cause (% of Cases)

NO. OF PATIENTS IN STUDY

HEPATITIS A

HEPATITIS B

ACETAMINOPHEN

DRUG OR TOXIN*

OTHER

INDETERMINATE

308

4

7

39

13

24

13

941

9

9

53

7

5

17

330 93 458

4 4 2

47 44 16

2 0 0

15 0 5

10 11 30

22 41 47

2130

6

18

29

8

14

25

United States, 1998-20018 United Kingdom, 1987-19939 France, 198610 Japan, 199811 India, 1992-199812 Total *Other than acetaminophen.

Chapter 93  Acute Liver Failure Table 93-3  Presenting Clinical Features of Acute Liver Failure*

PARAMETER

ACETAMINOPHEN (n = 605)

IDIOSYNCRATIC DRUG REACTION (n = 156)

INDETERMINATE (n = 180)

36 75 —†

46 68 9

38 57 9

51

37

4016

626

Age (yr) Female (%) Time to jaundice (days) Grade 3 or 4 encephalopathy (%) Serum ALT level (U/mL) Serum bilirubin level (mg/dL) Transplanted (%) Spontaneous survival (%) Overall survival (%)

4.5

HEPATITIS A (n = 34)

OTHER causes (n = 244)

42 43 6.5

49 44 3

42 73 7

49

53

53

41

846

1702

2275

668

HEPATITIS B (n = 102)

20.4

22.4

18.5

12.3

15.7

9 65

44 27

44 25

44 26

29 56

31 37

73

68

66

64

82

64

700

70

600

60

500

50

400

40

300

30

200

20

100

10

Other

Indeterminate

Shock Wilson disease Budd-Chiari syndrome Pregnancy

Hepatitis A Autoimmune hepatitis

Hepatitis B

Drugs

0

Acetaminophen

0

Spontaneous survival (%)

Number of patients

*1321 consecutive adults enrolled in the U.S. Acute Liver Failure Study Group (1998 to July 2008). All data reported as median or percentage; survival determined at three weeks. † Most patients do not develop jaundice. ALT, alanine aminotransferase. Data provided courtesy of Dr. William M. Lee and the U.S. Acute Liver Failure Study Group, September 2008.

Cause Spontaneous survival Transplanted

Died before transplantation Spontaneous survival rate (%)

Figure 93-1.  Outcomes of acute liver failure in the United States by cause. The U.S. Acute Liver Failure Study Group identified acetaminophen (in 49% of cases) and idiosyncratic drug reactions (in 12%) as the leading causes of acute liver failure among 1,321 consecutive adults between January 1, 1998 and July 1, 2008. In patients with hepatitis B, Wilson disease, or acute liver failure of indeterminate cause, the three-week spontaneous survival rate was less than 25% in the absence of emergency liver transplantation. By contrast, in patients with acute liver failure related to acetaminophen toxicity, hepatitis A, or pregnancy, the three-week spontaneous survival rate exceeded 50%. Rapid identification of the cause of acute liver failure is recommended to facilitate transfer of patients with a poor prognosis to a liver transplantation center. (Data provided with permission from Dr. William M. Lee, Dallas, Tex.)

ucts and package labeling have been instituted, but the impact of these measures on the incidence and severity of acute liver failure caused by acetaminophen has not been evaluated. Additional measures that may prove useful include imposing limits on the number of tablets sold at one

time, unbundling or reducing the dose of acetaminophen in prescription-narcotic congeners, and placing stronger warnings on package labeling.24,25

IDIOSYNCRATIC DRUG TOXICITY

Numerous prescription drugs, including various antibiotics, nonsteroidal anti-inflammatory drugs, and antiseizure medications, have been implicated in acute liver failure (see Chapter 86).26,27 In most cases, drug-induced acute liver failure is a rare and unpredictable event resulting from a metabolic idiosyncrasy that occurs in 1 in 10,000 to 1,000,000 patient–exposure years. Patients with druginduced acute liver failure are frequently female (70%) and develop jaundice within six months of starting the sus­ pected agent.8,26 Among 141 U.S. liver transplant recipients with drug-induced acute liver failure, isoniazid (16%), pro­ pylthiouracil (9%), phenytoin (7%), and valproic acid (7%) were the most commonly identified causative medications.28 Various over-the-counter herbal products and dietary supplements also have been associated with acute liver failure, including kava kava, weight loss supplements, and ephedra.29-31 In addition, severe hepatotoxicity has been associated with extracts of green tea, black cohosh, and adulterated traditional Chinese medicines (see Chapter 87).32-34 Unfortunately, herbal products and dietary supple­ ments are not closely regulated during development, manufacturing, or marketing, and, in many mixtures, the hepatotoxic ingredient(s) cannot be identified. Establishing a diagnosis of drug-induced acute liver failure is usually difficult because of the lack of specific laboratory markers, inability to rechallenge the patient, and limitations of available causality assessment instruments.35 In addition, less than 10% of patients have evidence of a hypersensitivity reaction associated with a rash or eosino­ philia at presentation. The Drug-Induced Liver Injury Network (DILIN) was established to improve our under­ standing of the risk factors, mechanisms, and outcomes of drug-induced liver injury in the United States.36 The development of an evidence-based causality assessment instrument to assist with the early recognition and diagno­ sis of idiosyncratic drug-induced liver injury is a priority (see http://dilin.dcri.duke.edu for additional information).

1559

1560

Section IX  Liver The primary treatment of drug-induced acute liver failure is to withdraw the culprit drug immediately once druginduced liver injury is suspected. In selected patients with severe hepatotoxicity caused by α-methyldopa, nitrofurantoin, or minocycline and autoimmune features (i.e., hypergammaglobulinemia, autoantibodies, plasma cell infiltration on a liver biopsy specimen), treatment with glucocorticoids may be of benefit, but controlled studies are lacking.37,38 Overall, the outcome of drug-induced acute liver failure is poor, with a spontaneous survival rate of only 2% to 40% unless liver transplantation is performed emergently.

VIRAL INFECTIONS Hepatotropic Viruses

Hepatitis A virus (HAV) and HBV infections are major causes of acute liver failure in many parts of the world, including India and other developing countries (see Table 93-1 and Chapters 77 and 78). Acute infection with HAV rarely leads to acute liver failure (<0.01% of cases), and, when it does, the prognosis is relatively good (see Fig. 93-1). Data from the U.S. Acute Liver Failure Study Group have indicated that the incidence of fulminant HAV infection is declining in parallel with the declining incidence of acute HAV in the general population because of increased immu­ nization against HAV.39 Although HBV is the most common viral cause of acute liver failure, acute liver failure is an uncommon manifestation of acute HBV infection. Infection with hepatitis D virus (HDV), which requires coinfection with HBV, accounts for 4% of cases of acute liver failure in areas endemic for HBV.40 Older studies have suggested that virologic factors, including infection with precore or core promoter variants of HBV, may account for the devel­ opment of acute liver failure in some patients.41 Subsequent studies have failed to demonstrate unique HBV mutants or variants associated with acute liver failure; the role of HBV genotypes in acute liver failure requires further investiga­ tion.42,43 Acute hepatitis E virus (HEV) infection is a leading cause of acute liver failure in India and other tropical coun­ tries but is rarely seen in Western countries (see Chapter 80).12,44 Pregnant women may be particularly prone to develop acute liver failure caused by HEV. European series have suggested that organ transplant recipients may be at increased risk of acquiring acute HEV infection and may develop chronic HEV infection.45,46 Hepatitis C virus (HCV) rarely causes acute liver failure.

Other Viruses

Nonhepatotropic viruses, including Epstein-Barr virus (EBV), cytomegalovirus (CMV), varicella-zoster virus, herpes simplex virus (HSV), and parvovirus B-19 infection account for less than 1% of cases of acute liver failure in adults.47,48 Whether these rare causes of acute liver failure are the result of viral variants or an aberrant host immune response to the virus is unclear. Making a diagnosis of acute liver failure caused by a nonhepatotropic virus is often difficult and frequently requires histologic confirma­ tion, as well as polymerase chain reaction testing for viral deoxyribonucleic acid (DNA) in the serum. Almost 50% of patients with HSV-related acute liver failure have no char­ acteristic skin lesions at presentation; the mortality rate is high because the diagnosis is usually delayed.49 Severe acute EBV, CMV, and HSV infection should be considered as possible causes of acute liver failure, particularly in immunosuppressed patients, because they can be treated successfully with antiviral therapy (see Chapter 81).

MISCELLANEOUS CAUSES

Acute liver failure occasionally develops in pregnant women, particularly during the third trimester (see Table 93-2).50 Acute fatty liver of pregnancy occurs in 0.0008% of all pregnancies and is associated with preeclampsia in over 50% of cases (see Chapter 38).51 Some affected women have an inherited deficiency in a fatty acid oxidation enzyme that can be identified by genetic testing. Wilson disease, a rare autosomal recessive disorder characterized by impaired biliary excretion of copper, can present as acute liver failure in up to 25% of cases (see Chapter 75). Most of these patients present in the second or third decade of life and have prom­ inent hemolysis, a low serum alkaline phosphatase level, an elevated serum aspartate aminotransferase (AST)–to– alanine aminotransferase (ALT) ratio, increased urinary copper excretion, and Kayser-Fleischer rings.52 Prompt rec­ ognition and listing for liver transplantation are essential, because the outcome is otherwise fatal. Infrequent causes of acute liver failure include mushroom (Amanita phalloides) poisoning (see Chapter 87), Budd-Chiari syndrome (see Chapter 83), autoimmune hepatitis (see Chapter 88), and malignant infiltration of the liver (see Chapter 35). All the miscellaneous causes of acute liver failure combined account for 5% to 30% of cases of acute liver failure (see Table 93-1).

INDETERMINATE ACUTE LIVER FAILURE

Acute liver failure of unknown cause, defined by negative serologic testing for hepatitis A, B, C, D, and E and the absence of other known causes, constitutes 15% to 44% of cases of acute liver failure worldwide (see Table 93-1). Because many of these patients present with a viral pro­ drome, the hope has been that new, more sensitive molecu­ lar laboratory methods would identify a viral cause of acute liver failure of unknown cause. Occult HBV infection has been identified in the sera or livers of some patients with acute liver failure of unknown cause by some41,53 but not other42,43 investigators. Although HCV has been impli­ cated as a cause of acute liver failure in a few patients, it is an exceedingly rare cause of acute liver failure in Western countries.53-55 Togavirus-like particles have been identified by electron microscopy in 7 of 18 liver explants from patients who underwent transplantation for indeterminate acute liver failure but are unlikely to be responsible for a substantial portion of cryptogenic cases of acute liver failure.56 The transfusion-transmitted virus (TTV) was found in the sera of patients with acute liver failure in initial studies, but TTV infection is not thought to be pathogenic.57,58 Studies have failed to demonstrate a link between hepatitis G virus (GB agent, or GBV-C), parvovirus B19, or SEN virus and indeterminate acute liver failure (see Chapter 81).59-61 Using a highly sensitive and specific assay for serum acetaminophen-cysteine adducts, Davern and colleagues identified adducts in 7 of 36 (19%) patients presumed to have indeterminate acute liver failure; the adduct levels were similar to those in patients with known acet­ aminophen hepatotoxicity.62 These patients tended to have high serum aminotransferase levels and low serum bilirubin levels at presentation, features similar to those seen in patients with a known acetaminophen overdose (see Table 93-3). Whether acetaminophen was the primary cause of liver injury or a cofactor requires further study.63 Regardless, patients with indeterminate acute liver failure should be evaluated rapidly for liver transplantation because of the low likelihood of spontaneous recovery (see later).

Chapter 93  Acute Liver Failure HEPATIC ENCEPHALOPATHY AND CEREBRAL EDEMA

CLINICAL FEATURES The clinical features of acute liver failure may result from the loss of critical hepatocellular functions (e.g., protein synthesis, intermediary metabolism, detoxification) and from effects on organs other than the liver. The major com­ plications of acute liver failure, as well as their pathogenesis and medical management, are outlined in Table 93-4. The initial presentation usually includes nonspecific complaints such as nausea, vomiting, and malaise, and jaundice usually develops soon after. Hepatocellular injury leads to impaired elimination of bilirubin, depressed synthesis of coagulation factors I, II, V, VII, IX, and X, and diminished synthesis of glucose. In addition, decreased uptake and increased gen­ eration of intracellular lactate occur as a result of anaerobic glycolysis. These derangements manifest clinically as jaun­ dice, coagulopathy, hypoglycemia, and metabolic acidosis. In addition to portending liver failure, coagulopathy increases the risk of gastrointestinal and intracranial hemor­ rhage, hypoglycemia can contribute to brain injury, and acidosis can contribute to hypotension.

Hepatic encephalopathy is a defining criterion for acute liver failure. Encephalopathy in acute liver failure is thought to arise primarily from the development of cerebral edema and resulting intracranial hypertension, rather than from portosystemic shunting of toxins. In addition to cerebral edema, many of the other complications of acute liver failure, including hypoglycemia, sepsis, fever, hypoxemia, and hypotension, may contribute to neurologic abnor­ malities. The staging of encephalopathy in acute liver failure is similar to that used for patients with cirrhosis (see Chapter 92). In stage 1, patients have subtle changes in affect, altered sleep patterns, or difficulties with concentra­ tion. Stage 2 is characterized by drowsiness, disorientation, and confusion. Stage 3 is marked by somnolence and inco­ herence. In stage 4, frank coma with minimal (4A) or no (4B) response to noxious stimuli is detected. On physical examination, many patients have asterixis or a tremor in stage 1 or 2, whereas hyperreflexia, clonus, and muscular rigidity are common in stages 3 and 4. Although worrisome, these upper motor neuron signs do not portend a poor prog­

Table 93-4  Pathogenesis and Management of Major Complications of Acute Liver Failure COMPLICATION

PATHOGENESIS

MANAGEMENT

Hypoglycemia

Diminished hepatic glucose synthesis

Blood glucose monitoring Intravenous glucose supplementation (10% or 20% dextrose)

Encephalopathy

Cerebral edema ICP monitoring (if stage 3 or 4 encephalopathy)

CT scan (if advanced encephalopathy) Elevate head of the bed > 30 degrees Consider osmotherapy (mannitol) or barbiturates Treat other contributing factors (e.g., hypoglycemia, hypoxemia, fever) Reduce fever (cooling blankets, antibiotics) Avoid benzodiazepines and other sedative medications ? Moderate hypothermia (see text)

Infections

Reduced immune function Invasive procedures

Aseptic medical, nursing care Daily surveillance cultures of blood, urine, and sputum High index of suspicion for bacterial and fungal infection Preemptive antibiotics for gram-negative organisms, anaerobes, and skin flora Consider antifungal therapy if patient worsens despite antibacterial coverage

Gastrointestinal hemorrhage

Stress ulceration

Nasogastric tube placement Intravenous H2 receptor antagonist or proton pump inhibitor

Coagulopathy

Reduced clotting factor synthesis Thrombocytopenia Fibrinolysis

Parenteral vitamin K Platelet infusions for bleeding and before procedures Plasma infusions for bleeding and before procedures Cryoprecipitate for bleeding with hypofibrinogenemia Recombinant factor VIIa (?) (see text)

Hypotension

Hypovolemia Decreased vascular resistance

Hemodynamic monitoring of central venous pressures Volume repletion with blood or colloid α-Adrenergic agents

Respiratory failure

ARDS (DAD)

Hemodynamic monitoring of central venous pressures Mechanical ventilation

Pancreatitis

?Hypoxia

Supportive care, including supplemental oxygen if needed Abdominal CT to exclude necrotizing pancreatitis

Renal failure

Hypovolemia Hepatorenal syndrome Acute tubular necrosis

Hemodynamic monitoring of central venous pressures Volume repletion with blood or colloid Avoidance of nephrotoxic agents (e.g., aminoglycosides, NSAIDs, contrast dye) Oral N-acetylcysteine prior to intravenous contrast agent Hemofiltration, dialysis

ARDS, acute respiratory distress syndrome; CT, computed tomography; DAD, diffuse alveolar damage; ICP, intracranial pressure; NSAIDs, nonsteroidal anti-inflammatory drugs.

1561

1562

Section IX  Liver nosis and can reverse with recovery or replacement of the failing liver. Cerebral edema is found in up to 80% of patients who die of acute liver failure and is almost universal in patients with coma.64 The pathogenesis of cerebral edema in acute liver failure is poorly understood. It has been proposed to result, in part, from the actions of intestine-derived neurotoxins that escape hepatic clearance and are released into the sys­ temic circulation.65 The demonstration of swollen endothe­ lial and astroglial cells in the brains of patients with acute liver failure suggests a potential role for cytotoxic edema, possibly resulting from increased brain glutamine levels. On the other hand, vacuolization in the basement membranes of capillaries, consistent with disruption of the blood-brain barrier, suggests a vasogenic mechanism of cerebral edema in acute liver failure. In any event, an increased production of glutamine in the central nervous system as a result of high circulating levels of ammonia and intracerebral lactate is believed to be critical to the pathogenesis of cerebral edema. In one study, arterial ammonia levels were associ­ ated with the risk of uncal herniation and death in patients with acute liver failure.66 Another study suggested that a lack of reduction in arterial ammonia levels over time was associated with an increased risk of progression to cerebral edema.67 Progressive cerebral edema caused by the development of intracranial hypertension can also result in cerebral hypo­ perfusion and resulting hypoxia, which can lead to irrevers­ ible neurologic damage, uncal herniation, and brain death. Loss of intracranial vascular tone can also lead to surges in intracranial pressure (ICP) with changes in systemic hemo­ dynamics based on the following equation: Cerebral perfusion pressure = mean arterial pressure − intracranial pressure. Because the management of cerebral edema is complex and challenging, systemic hemodynamic monitoring and direct ICP monitoring are frequently recommended for patients with stage 3 or 4 encephalopathy and those under­ going liver transplantation.

COAGULOPATHY AND BLEEDING

The liver is the major site of synthesis of coagulation factor and related inhibitory proteins. The reticuloendothelial cell system of the liver is involved in the clearance of activated clotting factors and their degradation products. Therefore, patients with acute liver failure frequently have a multi­ factorial coagulopathy and a resulting increased risk of bleeding and clotting. Laboratory features of fibrinolysis, hypofibrinogenemia, dysfibrinogenemia, and DIC are fre­ quent in patients with acute liver failure.68 Thrombocytope­ nia also develops in most patients with acute liver failure and may be the result of increased destruction of platelets from a consumptive coagulopathy or reduced thrombopoi­ etin production, bone marrow dysfunction, or the effects of medications.69 Clinically significant bleeding has been reported to occur in 10% to 20% of patients with acute liver failure; the most common sources are the upper gas­ trointestinal tract, nasopharynx, and skin puncture sites. Critically ill patients with acute liver failure have a particu­ lar propensity for gastrointestinal bleeding caused by acute portal hypertension, increased intracranial pressure, and coagulopathy.70 To monitor the coagulopathy of acute liver failure, experts advise obtaining serial assessments of the platelet count, hemoglobin level, international normalized ratio (INR), fibrinogen level, and factor V level. Although measurement

of the INR has limitations, the INR can be readily obtained in most hospitals. Because factor V has the shortest serum half-life of the clotting factors synthesized by the liver, serial measurement of factor V levels has been proposed as a sensitive way to monitor recovery of liver function over time.

INFECTION

Bacterial infections may develop in as many as 80% of patients with acute liver failure, and bacteremia is present in up to 25%.65,71 Uncontrolled infection accounts for exclu­ sion of approximately 25% of patients with acute liver failure from liver transplantation and approximately 40% of post-transplantation deaths. At least three factors place patients with acute liver failure at increased risk for infec­ tion. First, intestine-derived microorganisms may enter the systemic circulation from portal venous blood as a result of damage to hepatic macrophages (Kupffer cells). Second, impaired neutrophil function may result from reduced hepatocellular synthesis of acute-phase reactants, including components of the complement cascade. Third, patients with acute liver failure are often subjected to invasive pro­ cedures (e.g., intravascular and urethral catheterization, endotracheal intubation), and physical barriers to infection, including skin and airway, are thus breached. Indeed, the major sites of infection are the respiratory and urinary tracts. Not surprisingly, the most commonly isolated bacte­ ria are staphylococcal and streptococcal species and gramnegative aerobes.72 Fungal infections develop in up to one third of patients with acute liver failure.73 Most of these infections are caused by Candida albicans and typically develop after the second or third week of hospitalization. Although Aspergillus infections were thought to be uncommon in the setting of acute liver failure, they may be more prevalent than previ­ ously appreciated, and aspergillosis may account for up to half of fatal infections in the post-transplantation period.74 Risk factors for fungal infections are renal failure, prolonged antibiotic therapy for bacterial infections, and use of inva­ sive monitoring devices. Characteristically, fungal infection is associated with fever or leukocytosis refractory to broadspectrum antibiotics. Patients in whom symptoms and signs of the systemic inflammatory response syndrome develop in association with bacterial or fungal infection are more likely to experience worsening encephalopathy and to die; this observation highlights the importance of infections in the outcome of acute liver failure.75,76

MULTIPLE ORGAN FAILURE SYNDROME

The multiple organ failure syndrome manifests clinically as peripheral vasodilatation with hypotension, pulmonary edema, renal failure, and DIC. Liver failure may trigger the microcirculatory derangements that underlie this syndrome by two mechanisms. First, polymerization of actin (released from dying hepatocytes) within the capillary lumen and platelet activation may cause endothelial injury.77 Second, impaired hepatic clearance may lead to the accumulation of vasoactive substances in the systemic circulation.78 Multi­ ple organ failure syndrome is an important contributor to mortality and constitutes a major contraindication to liver transplantation. Hypotension is observed frequently in patients with acute liver failure and can result from reduced vascular resistance or intravascular volume depletion. Subclinical cardiac injury, defined by elevated serum troponin levels, is common in patients with acute liver failure, and the level of troponin correlates with the risk of an adverse outcome.79 Acute pancreatitis may also develop in patients with acute

Chapter 93  Acute Liver Failure liver failure, particularly those with an acetaminophen overdose, as a result of tissue hypoxia and hypoperfusion. In one series, 44% of patients who died of acute liver failure had acute pancreatitis.80 Acute pancreatitis is not a contra­ indication to liver transplantation unless evidence of extensive pancreatic necrosis is seen by computed tomog­ raphy (CT). Respiratory failure is commonly associated with acute liver failure. In one series, 37% of patients with acute liver failure had pulmonary edema,81 whereas acute respiratory distress syndrome (ARDS) was present in 33% of patients with acetaminophen-associated acute liver failure.82 Fur­ thermore, ARDS was associated with intracranial hyperten­ sion, the requirement for vasopressor agents, and a higher rate of mortality. The cause of renal failure, which develops in 30% to 50% of patients with acute liver failure, is usually multifacto­ rial.78 Hepatorenal syndrome is often difficult to differenti­ ate from intravascular volume depletion because both entities are associated with oliguria, azotemia, and a low fractional excretion of sodium. Acute tubular necrosis is associated with a 50% decrease in survival in patients with acetaminophen-induced acute liver failure,83 and the mor­ tality rate is more than doubled in patients with multiple organ failure syndrome.84

PROGNOSIS Patients with acute liver failure fall into two broad catego­ ries: (1) those in whom intensive medical care enables recovery of hepatic function through liver regeneration and (2) those who require liver transplantation to survive. Rapid identification of patients with an unfavorable prognosis is critical. The cause of acute liver failure and clinical presen­ tation are important correlates of prognosis. For example, patients with acute liver failure caused by acetaminophen have a better prognosis than those with indeterminate acute liver failure (see Fig. 93-1).8 Similarly, patients who reach stage 3 or stage 4 encephalopathy tend to do worse than those who reach only stage 1 or stage 2 encephalopathy.4 In an individual patient, however, these indicators do not allow accurate prediction of the need for liver transplantation. Investigators at King’s College in London have performed a multivariate analysis of clinical and biochemical variables and their relationship to mortality in 588 patients with acute liver failure.83 In this analysis, a distinction was made between patients with acetaminophen toxicity and those with other causes of acute liver failure (Table 93-5). For patients with acetaminophen-induced acute liver failure, the presence of any single adverse characteristic was associated with a mortality rate of at least 55%, and severe acidosis was associated with a mortality rate of 95%. Sub­ sequently, the addition of arterial lactate levels to these predictors has been shown to improve the positive and negative predictive values in patients with acetaminophenassociated acute liver failure.85 Among patients with non­ acetaminophen acute liver failure, the presence of any single adverse prognostic factor was associated with a mor­ tality rate of 80%, and the presence of three adverse char­ acteristics was associated with a mortality rate of more than 95%. These mortality rates vastly exceed those associated with liver transplantation. Therefore, the presence of any single indicator of a poor prognosis should prompt early referral to a liver transplantation center. These selection criteria are simple, and acquisition of the necessary data

Table 93-5  King’s College Criteria for Liver Transplantation in Acute Liver Failure83,85 ACETAMINOPHEN CASES

NON-ACETAMINOPHEN CASES

Arterial pH < 7.3* or Arterial lactate level > 3.5 mmol/L at 4 hr or Arterial lactate level > 3.0 mmol/L at 12 hr* or INR > 6.5 (or PT > 100 sec); and Serum creatinine > 3.4 mg/dL; and Stage 3 or 4 encephalopathy

INR > 6.5 (PT > 100 sec) or Any three of the following five: Age <10 or >40 yr Duration of jaundice > 7 days Cause: non-A, non-B hepatitis, halothane hepatitis, idiosyncratic drug reaction, indeterminate INR > 3.5 (or PT > 50 sec) Serum bilirubin > 17.5 mg/dL

*Measured after fluid resuscitation. INR, international normalized ratio; PT, prothrombin time.

requires only brief history taking, routine laboratory studies, and serologic testing for HAV and HBV.86 Another study from the U.S. Acute Liver Failure Study Group has demonstrated a lower sensitivity rate and negative predic­ tive value for these prognostic factors in 108 patients with acetaminophen-induced acute liver failure.15 Other groups have found that markers of organ failure, such as the Acute Physiology and Chronic Health Evaluation (APACHE) II score and Sequential Organ Failure Assessment (SOFA) index, are better predictors of outcome in patients with acetaminophen-related acute liver failure than the King’s College criteria and Model for End-stage Liver Disease (MELD) score (Table 93-6).16,87,88 Liver histologic evaluation in acute liver failure is associated with substantial sampling error and potential complications and does not reliably predict outcome.89 Therefore, percutaneous or transjugular liver biopsy is not recommended for prognosis or staging purposes but can be helpful in confirming a diagnosis of malignant infiltration, autoimmune hepatitis, and nonhepatotropic viral infection (Fig. 93-2). The predictive value of serum Gc-globulin levels is comparable to the King’s College criteria, but the assay is technically difficult and not generally available.90 Other investigators have examined the prognostic useful­ ness of measuring plasma factor V levels (see Table 95-5)91 and hepatic volumetry,92 but these parameters do not appear to add significantly to the assessment of outcome. The U.S. Acute Liver Failure Study Group and others have reported on the potential use of elevated serum phosphate levels (>3.7  mg/dL) as a marker of impaired liver regenera­ tion and poor prognosis in patients with acetaminopheninduced acute liver failure.93,94 Other groups, however, have failed to confirm the clinical usefulness of serum phosphate measurements in acute liver failure.95 Serial assessment of serum alpha fetoprotein levels, which cor­ relate with hepatic regeneration, have been reported to be of value in predicting prognosis in patients with acute liver failure.96

TREATMENT A variety of therapies have been proposed and studied in patients with acute liver failure, including glucocorticoid therapy, prostaglandin infusions, and exchange transfu­

1563

1564

Section IX  Liver Table 93-6  Criteria Used to Predict the Prognosis of Patients with Acetaminophen-Induced Acute Liver Failure SOURCE

CRITERIA

U.S. ALFSG, N = 275

Royal Free Hospital, N = 100

SENSITIVITY (%)

SPECIFICITY (%)

PPV

NPV

APACHE II score ≥ 20 King’s College criteria ≥ 1 (see Table 93-5) MELD score ≥ 35

68 26

87 92

0.77 0.63

0.81 0.69

61

71

0.54

0.76

APACHE II score ≥ 12 Arterial lactate level ≥ 3.3 mmol/L King’s College criteria ≥ 1 (see Table 93-5) SOFA score ≥ 12

67 68 47

76 73 83

0.69 0.68 0.70

0.75 0.73 0.65

67

80

0.74

0.74

ALFSG, Acute Liver Failure Study Group; APACHE, Acute Physiology and Chronic Health Evaluation; MELD, Model for End-stage Liver Disease; PPV, positive predictive value; NPV, negative predictive value; SOFA, sequential organ failure assessment. Adapted from Larson AM, Polson J, Fontana RJ, et al. Acetaminophen–induced acute liver failure: Results of a United States multicenter prospective study. Hepatology 2005; 43:1364-72; and Cholongitas EB, Betrossian A, Leandro G, et al. King’s criteria, APACHE II, and SOFA scores in acute liver failure. Hepatology 2006; 43:881.

Patient with acute liver failure

Intensive medical management Cause-specific treatment (if appropriate) Immediate consultation with liver transplantation center

Contraindications to transplantation

Yes

Continued intensive support

No Transfer to liver transplantation center Place on waiting list for emergency transplantation

Donor organ available Figure 93-2.  Liver histopathology in acute liver failure. The specimen shown is from a 59-year-old man who underwent a transjugular liver biopsy for presumed amoxicillin-clavulanate liver toxicity. Findings include severe acute hepatitis with confluent areas of necrosis predominating in the central zones, consistent with, but not diagnostic of, an idiosyncratic drug reaction. Despite supportive care, multiorgan failure developed, culminating in death. (Hematoxylin and eosin, ×40.)

Yes Assess prognosis

Recovery likely

Yes

No

sions. Only liver transplantation, however, has permitted salvage of patients with irreversible liver failure. Unfortu­ nately, many patients with irreversible acute liver failure do not undergo liver transplantation because of late referral, contraindications, or the lack of a donor liver. Therefore, patients with acute liver failure should be evaluated for liver transplantation as soon as possible and, if no contra­ indications are identified, placed on a liver transplant waiting list. If and when a donor organ becomes available, a patient listed for transplantation should be reassessed for the continued need for transplantation. An algorithm depicting the management of acute liver failure is shown in Figure 93-3.

INITIAL EVALUATION AND MANAGEMENT

The initial management of a patient with acute liver failure should include rapid identification of the cause of acute liver failure, with an emphasis on treatable conditions. For example, acetaminophen toxicity is treated initially with

No

Contraindications to transplantation

Ongoing intensive support Reassess for recovery and presence of contraindications to transplantation Yes

No Liver transplantation Figure 93-3.  Algorithm for the management of acute liver failure. The initial approach to management includes rapid identification and treatment of reversible causes of acute liver failure. Supportive care with careful monitoring for complications of the disorder in an intensive care unit is recommended. Contact with a liver transplantation center should be established promptly for potential transfer and transplantation evaluation as early as possible.

gastric lavage, oral charcoal, and prompt administration of oral N-acetylcysteine (see Chapter 86).97 For patients with severe nausea and vomiting, an approved intravenous for­ mulation of N-acetylcysteine can be administered safely in a monitored setting.98 Similarly, patients with Amanita

Chapter 93  Acute Liver Failure mushroom poisoning should be treated with immediate gastric lavage and instillation of charcoal in an attempt to reduce the toxin load.99 In addition, hemodialysis can remove toxin from the serum, and intravenous penicillin, milk thistle (silymarin), and cytochrome c may further lower the enterohepatic toxin load (see Chapter 87); however, the clinical benefit of these measures is uncer­ tain.100 HSV-induced acute liver failure has been reported to respond to intravenous acyclovir.101 Rapid delivery and supportive care constitute the treatment of choice for preg­ nant women with acute fatty liver of pregnancy, the he­ molysis, elevated liver enzyme levels, and low platelet (HELLP) syndrome, and preeclampsia (see Chapter 38). The benefit of oral antiviral agents in patients with fulminant hepatitis B remains unproven; many experts advise pre­ scribing a nucleoside analog, such as entecavir or lamivu­ dine, in light of their favorable safety profiles, in the hope of salvaging liver function and reducing the level of viremia prior to liver transplantation (see Chapter 78).102-104 N-acetylcysteine has been proposed as a potential treat­ ment for non–acetaminophen-related acute liver failure on the basis of studies demonstrating improvements in tissue oxygenation and systemic hemodynamics and the drug’s antioxidant properties.105 The U.S. Acute Liver Failure Study Group reported on a multicenter, randomized, con­ trolled trial that compared a 72-hour infusion of intravenous N-acetylcysteine and placebo in 173 adult patients with acute liver failure.106 The overall patient survival at three weeks was similar in the two groups (70% N-acetylcysteine vs. 66% placebo, P = 0.28), but transplant-free survival was significantly better in patients who received N-acetylcyste­ ine (40% vs. 27%, P = 0.04). The benefit of N-acetylcysteine appeared to be limited to the subgroup of patients with grade 1 or 2 encephalopathy at entry (52% vs. 31%, P = 0.021). In addition, a single-center retrospective study of 170 children with nonacetaminophen acute liver failure demonstrated that treatment with N-acetylcysteine was associated with a shorter length of hospital stay and a higher rate of spontaneous recovery.107 In this study, however, the untreated controls were not contemporaneous and had more severe illness at presentation. In both studies,106,107 N-acetylcysteine was generally well tolerated, with a low rate of side effects (e.g., rash, bronchospasm, arrhyth­ mia). Additional studies of N-acetylcysteine for non– acetaminophen-related acute liver failure are in progress to identify which patients may benefit from this treatment. All patients with acute liver failure should be cared for in an intensive care unit because they can deteriorate rapidly.104 Serial laboratory studies, including acid-base status, arterial ammonia levels, and INR, should be carried out to monitor the patient’s condition. Urgent transfer to a liver transplantation center is advisable early in the course, prior to the development of advanced encephalopathy or complications.108

ENCEPHALOPATHY AND CEREBRAL EDEMA

Encephalopathy associated with acute liver failure tends to be progressive, unless liver failure is reversed. Sedativehypnotic drugs, which may exacerbate encephalopathy, should be avoided unless patients require mechanical ven­ tilation. Lactulose is of uncertain benefit and may be asso­ciated with bowel ischemia. Reversible conditions that may contribute to altered mental status (e.g., hypogly­ cemia, hypoxemia) should be treated immediately. Hypoglycemia generally responds to parenteral administra­ tion of glucose. Similarly, underlying infection and sepsis should be treated aggressively with fluids and anti­

biotics, because systemic cytokines may alter brain function (see Table 93-4). Patients with stage 3 or 4 encephalopathy should undergo elective endotracheal intubation and mechanical ventila­ tion for protection of the airway, particularly before being transported to a liver transplantation center. Many mechani­ cally ventilated patients are also deeply sedated or para­ lyzed, and evidence of generalized seizure activity that can worsen encephalopathy may be concealed. Therefore, continuous electroencephalographic monitoring of deeply sedated or paralyzed patients with acute liver failure has been proposed. Treatment of subclinical seizures with phe­ nytoin or other antiepileptic medications is appropriate, but the efficacy of prophylactic therapy to prevent seizure activity has not been established.109 Intracranial hypertension can be suspected on the basis of noninvasive assessment or direct measurement, but non­ invasive assessment by physical examination and radiologic imaging has important limitations. Impaired pupillary responses, posturing, or seizures, which may suggest the presence of intracranial hypertension, are not sensitive signs for intracranial hypertension, particularly when seda­ tives or neuromuscular blocking agents are used in mechan­ ically ventilated patients. CT of the head is useful for identifying mass lesions, intracranial hemorrhage, and evi­ dence of brainstem herniation because these conditions may affect clinical decision making. CT scans of the head should be obtained in all patients with advanced encepha­ lopathy. Nevertheless, the correlation between CT evidence of cerebral edema and measured ICP is imperfect, with a sensitivity varying from 60% to 75%.110,111 Transcranial Doppler measurements of middle cerebral artery blood flow continue to improve, but further refinements are needed to make this noninvasive modality reliable enough for clinical decision-making.112 Monitoring of ICP represents the most accurate way to detect intracranial hypertension but has several potential limitations. First, placement of an ICP transducer requires correction of underlying coagulopathy. Second, the ICP transducer represents a potential portal of entry for infec­ tious organisms. Third, placement of the transducer can precipitate intracranial hemorrhage, which can be fatal. The frequency of serious complications ranges from 4% to 20%; parenchymal catheters are associated with a higher rate of complications than subdural or epidural transducers.113 Nevertheless, ICP transducers can provide invaluable physi­ ologic data that influence management and decisions regard­ ing liver transplantation. In one study, 92 patients with advanced encephalopathy had ICP monitors placed and were compared with 239 unmonitored patients. The moni­ tored patients tended to have more severe liver failure and multiorgan failure.114 In subjects listed for liver trans­ plantation, ICP monitoring was associated with a greater number of medical treatments and procedural interven­ tions. Although overall and post-transplantation survival rates were similar in the two groups, the monitored patients had a low rate of intracranial bleeding (i.e., 10%). The data suggest that future studies of therapies for acute liver failure should include ICP monitoring. For patients with refractory intracranial hypertension, technetium perfusion scans are useful to detect irreversible brain death, which may other­ wise be unrecognized in a sedated patient. Elevation of the head of the bed to at least 30 degrees from horizontal (and avoidance of the head-down position) is a simple measure to reduce ICP. If this maneuver fails, spe­ cific treatment is required. Osmotherapy and barbiturates are two options for treating intracranial hypertension. Osmotherapy with intravenous mannitol (0.5 to 1 g/kg)

1565

1566

Section IX  Liver requires preserved renal function (or concomitant hemofil­ tration, if necessary) and effectively controls intracranial hypertension in approximately 60% of cases.115 Hypertonic saline may also be of value in patients with cerebral edema, but controlled trials are needed to confirm a benefit.116 Uncontrolled data have supported the use of intravenous thiopental, a barbiturate; its efficacy is similar to that of mannitol.117 Thiopental has two relative advantages—its onset of action is rapid, and its use does not require pre­ served renal function. Potential drawbacks of thiopental are a risk of hypotension and, of greater importance, the poten­ tial to mask clinical indicators of neurologic recovery or deterioration. In general, it is reasonable to use mannitol as first-line therapy and to reserve a barbiturate for patients with renal insufficiency or refractory intracranial hyperten­ sion. Glucocorticoids are of no benefit.113 Pilot studies have suggested that moderate hypothermia (e.g., 32° C to 33° C), achieved by the use of an external cooling blanket, may be of benefit in patients with acute liver failure and refractory cerebral edema.118,119 All patients treated with hypothermia require the placement of an ICP monitor because paralytic agents are required to prevent shivering. Moreover, the effect of moderate hypothermia on the risk of bleeding and infection require further study. A prophylactic role for hypothermia also requires further study.

COAGULOPATHY AND BLEEDING

Placement of a nasogastric tube to monitor gastrointestinal bleeding and gastric pH is recommended for intubated patients with acute liver failure. The risk of upper gastroin­ testinal hemorrhage can be reduced by intravenous admin­ istration of a histamine H2 receptor antagonist,120 and proton pump inhibitors probably have a similar benefit. Adminis­ tration of subcutaneous vitamin K to attempt to reverse hypoprothrombinemia is also reasonable. Coagulation parameters, including the INR, plasma factor V level, plate­ let count, and plasma fibrinogen level, should be assessed serially in all patients with acute liver failure. The decision to administer clotting factors prophylacti­ cally in nonbleeding patients should be tempered by the recognition that infusion of plasma will mitigate the value of the INR as a prognostic factor. Furthermore, infusion of plasma can lead to volume overload and respiratory failure, particularly in patients with renal failure, and prophylactic administration of plasma has not been shown to improve the clinical outcomes of patients with acute liver failure. Therefore, unless a patient is actively bleeding or an inva­ sive procedure is planned, prophylactic infusions of plasma are not recommended. For patients with acute liver failure who undergo an inva­ sive procedure and whose INR fails to improve with plasma, recombinant factor VIIa has been shown to be efficacious, but the optimal dose has not been established and use of this agent carries a risk of thrombosis.121 Cryoprecipitate, plasma, and platelets should be given to patients with hypo­ fibrinogenemia, DIC, and active bleeding.

INFECTION

Clinical recognition of infection may be difficult because signs such as hypotension, leukocytosis, and acidosis may reflect the underlying liver failure. Therefore, daily surveil­ lance cultures of blood, urine, and ascitic fluid are recom­ mended in patients with acute liver failure. The advisability of prophylactic antibiotics in the setting of acute liver failure is debatable. On one hand, prophylactic antibiotics may delay the development of infections that limit the

applicability of liver transplantation. On the other hand, antibiotics may increase the risk of superinfection with resistant bacteria or fungi. This issue has been addressed in a small randomized trial.122 Patients treated with prophylac­ tic intravenous cefuroxime had a significant reduction in the rate of documented infections (from 61% to 32%) com­ pared with those treated conservatively and a modest (but statistically insignificant) increase in the rate of survival (from 45% to 67%). Enteral decontamination with orally administered antibiotics (as well as systemic antibiotics) does not appear to alter the clinical outcome of patients with acute liver failure, compared with that observed with systemic antibiotics alone.123 The usefulness of systemic prophylactic antibiotics warrants further investigation. At the least, a high level of suspicion for infection and a low threshold for administering antibiotics are required in man­ aging patients with acute liver failure. If infection is sus­ pected, the choice of antibiotics should be based on the spectrum of likely bacterial pathogens (e.g., Staphylococcus, gram-negative aerobes) and local hospital patterns of microbial sensitivity. A reasonable empirical regimen is intravenous vancomycin and a third-generation cephalo­ sporin or fluoroquinolone.

MULTIPLE ORGAN FAILURE SYNDROME

The fundamental goal of management of multiple organ failure syndrome in patients with acute liver failure is similar to that in patients with other causes of multiple organ failure—to optimize arterial pressure and tissue oxy­ genation. Ideally, the mean arterial pressure (MAP) should be kept above 60 mm Hg to maintain cerebral perfusion.64 A central venous or right heart catheter may be useful for monitoring the patient’s intravascular volume status. Hypo­ tension resulting from intravascular volume depletion should be corrected with blood or colloids. If hypotension is caused by reduced vascular resistance, administration of an α-adrenergic agonist may be useful. Although pressors can be used to maintain MAP within a physiologic range, they have the potential to impair tissue oxygenation further; terlipressin (a long-acting vasopressin analog not available in the United States) may worsen cerebral edema (see Chapters 90 and 92).124,125 Most experts recommend using norepinephrine or dopamine rather than vasopressin because of adverse effects of the latter on intracranial hyper­ tension.125 In small short-term studies, N-acetylcysteine has been shown to improve tissue oxygenation without adverse effects on hemodynamics126; however, the impact of this agent on overall patient outcome has not yet been determined. Endotracheal intubation and mechanical ventilation are frequently necessary for patients with acute liver failure. Hypoxemia can result from respiratory depression caused by coma or impaired gas exchange caused by ARDS or superimposed pneumonia. Vigorous suctioning and Val­ salva maneuvers should be avoided to prevent surges in ICP. Patients with acute liver failure tolerate volume overload poorly in light of their propensity to develop ARDS and cerebral edema. Early measurement of the central venous or pulmonary capillary wedge pressure is preferable in oligu­ ric patients to empirical administration of fluid boluses. If oliguria persists in the face of adequate central filling pres­ sures, continuous renal replacement therapy should be ini­ tiated. Continuous venovenous hemofiltration has been shown to be superior to intermittent hemodialysis, with less hemodynamic instability and improved tissue oxygen deliv­ ery, in oliguric patients with acute liver failure.127 Nephro­ toxic drugs such as aminoglycosides and nonsteroidal

Chapter 93  Acute Liver Failure Table 93-7  Results of Liver Transplantation for Acute Liver Failure in the United States LIVER TRANSPLANTATION CENTER LOCATION* Chicago132 Michigan131 Nebraska134 Pittsburgh133 Philadelphia129 San Francisco132 United States (multicenter)130 United States (multicenter)8 Total

NO. OF PATIENTS

STUDY PERIOD

EARLY PATIENT SURVIVAL (%)†

ONE-YEAR PATIENT SURVIVAL (%)

19 19 30 42 18 35 121 89 373

1984-1988 1985-1990 1986-1991 1980-1987 1985-1990 1988-1992 1994-1996 1998-2001 —

74 NR 75 74 65 94 NR 84 80 (187/233)

58 68 42 59 65 92 76 NR 70 (198/284)

*Superscript numbers indicate references. † Early patient survival reflects discharge from the hospital following transplantation. NR, not reported.

anti-inflammatory drugs should be avoided in all patients with acute liver failure, and appropriate precautions should be taken if intravenous contrast dye is required.

LIVER TRANSPLANTATION

Liver transplantation has transformed the management of patients with acute liver failure and is discussed in greater detail in Chapter 95. Before the advent of liver transplanta­ tion, less than 30% of patients with acute liver failure sur­ vived. By contrast, survival rates for patients with acute liver failure who undergo liver transplantation have been substantially higher, with a short-term survival rate of 80% and a one-year survival rate of 70% when the results of several major transplantation centers are combined (Table 93-7).8,128-134 The decision to perform transplantation in a patient with acute liver failure must balance the likelihood of spontaneous recovery with the risks of surgery and longterm immunosuppression. Furthermore, contraindications to transplantation, particularly irreversible brain damage, active extrahepatic infection, or multiple organ failure syn­ drome, must be considered. Although changes in the rules governing allocation of donor livers have shortened waiting times for patients with acute liver failure in the United States, the decision to place a patient on the waiting list for transplantation must still be made promptly.135,136 A patient’s clinical status needs to be assessed frequently to determine whether the patient is likely to recover or has developed a contraindication to transplantation. In one series, a contraindication that pre­ cluded transplantation developed in 22% of liver transplan­ tation candidates, whereas 12% improved and were removed from the waiting list.8 The shorter median waiting times among liver transplant recipients compared with the times to exclusion for nontransplanted patients (three versus five days) highlights the critical, ongoing shortage of donor organs. Because of the shortage of donor organs, patients with acute liver failure are more likely to receive an ABOincompatible rather than an ABO-compatible or ABOidentical graft.137 In addition, marginal donor grafts that are older or steatotic are used more frequently in transplant recipients with acute liver failure than in recipients with other indications for transplantation. These factors may explain in part the higher rate of primary nonfunction and rejection among transplant recipients with acute liver failure compared with other recipients.138 Posttransplantation seronegative chronic hepatitis is also more common in transplant recipients with acute liver failure than in recipients with cirrhosis (41% versus 14% at one

year).139 The long-term functional and cognitive outcomes of transplant recipients with acute liver failure have not been well studied but may be inferior to those of transplant recipients with cirrhosis.140

Live Donor Liver Transplantation

In countries in which cadaveric livers are not readily avail­ able, live donor liver transplantation has been performed successfully in highly selected patients with acute liver failure (see Chapter 95).141,142 In the United States, adult live donor liver transplantation has rarely been undertaken for acute liver failure, and a moratorium has been placed on using this procedure for patients with acute liver failure in New York. One study, however, has demonstrated that seven of ten live donor organ recipients with acute liver failure survived, compared with two of three cadaveric transplant recipients with acute liver failure.143 In addition, the frequency of complications in live donors to a patient with acute liver failure (50%) was similar to that reported in donors who were evaluated electively for donation to a transplant recipient with cirrhosis (33%). Nevertheless, because of concerns regarding the ability to evaluate donors safely in an accelerated time frame, the potential for coer­ cion, and the potential for poorer outcomes, most experts recommend a whole-sized cadaveric transplant whenever possible in a patient with acute liver failure.

INVESTIGATIONAL APPROACHES Treatment strategies, such as charcoal hemoperfusion and administration of prostaglandin E1, which showed promise in uncontrolled trials, have not been shown to be superior to standard care when studied in randomized studies.144,145 Plasmapheresis and hepatectomy have been suggested as a possible bridge to liver transplantation, but prospective trials have yet to be performed.146,147 Three additional forms of therapy may provide a bridge to liver transplantation or to regeneration of the native liver with spontaneous recovery—auxiliary liver transplantation, extracorporeal liver support devices, and hepatocyte transplantation.

AUXILIARY LIVER TRANSPLANTATION

Auxiliary liver transplantation, in which the donor graft is implanted orthotopically beside the surgically reduced native liver or heterotopically inferior to the native liver, has been investigated by a number of centers.148,149 The

1567

1568

Section IX  Liver advantage of this procedure is that by providing a temporary auxiliary liver, the severely diseased native liver may be allowed to regenerate. Ideally, immunosuppression may then be gradually withdrawn, thereby allowing the trans­ planted liver to involute or be surgically removed. The usefulness of this operation is limited by technical compli­ cations as well as by the difficulty in predicting which patients with acute liver failure are likely to experience hepatic regeneration. This approach should only be under­ taken in centers with specialized expertise.

EXTRACORPOREAL LIVER SUPPORT

Extracorporeal liver support devices fall into two broad categories, hemodiadsorption systems and bioartificial livers.150 Hemodiadsorption systems use hemodialysis in combination with perfusion of the patient’s plasma or blood through three hollow fiber filters impregnated with char­ coal, resins, and albumin, respectively. Although these devices may remove circulating toxins, they do not replace other liver functions. Albumin dialysis (e.g., the molecular adsorbent recirculating system [MARS]) uses hemodialysis of whole blood in series with an albumin dialyzer and char­ coal filter. This technology is simpler to use than plasma­ pheresis and has shown some promise in trials of cirrhotic patients with hepatic encephalopathy,151,152 but information regarding the safety and efficacy of these systems in patients with acute liver failure is limited.153,154 Bioartificial liver devices contain liver cells grown within specialized hollow fiber cartridges through which the patient’s plasma is perfused. The success of such devices depends largely on the mass of cells they contain, the extent to which these cells maintain liver-specific functions, and the duration for which these functions are maintained. Because the devices under clinical investigation contain only hepatocytes, derangements attributable to nonparen­ chymal cells, such as Kupffer cells and biliary epithelia, are not replaced. The results of the HepatAssist bioartificial liver device trial were reported in 2004.155 This device uses a dialysis cartridge loaded with approximately 100 g of cryopreserved porcine hepatocytes, or 7 billion cells, and also has a charcoal filter. The 85 patients with acute liver failure who were treated with the HepatAssist device did not experience an improvement in 30-day survival com­ pared with the 86 patients with acute liver failure who received standard care (71% versus 62%, P = 0.26). Treat­ ment with the device was well tolerated, and the rate of thrombocytopenia, hypotension, and other adverse events was not significantly greater in treated patients. Further­ more, there were no reports of inadvertent transmission of porcine retroviruses or development of xenogenic antibod­ ies in treated patients.156 Although this pioneering trial failed to demonstrate a significant benefit in outcome with the HepatAssist device, the proof of concept that an extra­ corporeal device with porcine hepatocytes can be used and can lead to a trend toward improvement in metabolic, hemodynamic, and clinical parameters was realized. The development of devices with a larger hepatocyte mass, sim­ plified circuitry, and differentiated hepatocyte function is eagerly awaited.

HEPATOCYTE TRANSPLANTATION

The potential role of hepatocyte transplantation in patients with acute liver failure is likely to be as a bridge to liver transplantation or regeneration.157 Human hepatocyte trans­ plantation has demonstrated efficacy in preliminary studies of patients with metabolic disorders and decompensated

cirrhosis.158,159 Stable expression of transplanted hepato­ cytes has been difficult to achieve. In one trial, three of six patients with acute liver failure survived 14, 20, and 52 days after transplantation of 109 to 1010 hepatocytes, representing 1% to 10% of normal liver cell mass.121 Although metabolic parameters improved within 72 hours of transplantation, transient respiratory insufficiency was observed in several patients. This report demonstrated that transplanted human hepatocytes can engraft into a regenerating liver, but further work on enhancing graft function, native liver regeneration, and cell delivery is needed. Pluripotent hepatocyte stem cells derived from bone marrow may prove useful for hepatocyte transplantation. In one study, liver biopsy specimens from human recipients of gender-discordant bone marrow or liver transplants were analyzed for marrow-derived hepatocytes and cholangio­ cytes.160 Differences in the extent of engraftment were asso­ ciated with the degree of allograft injury, thereby suggesting a possible role for bone marrow–derived stem cells in the treatment of severe acute hepatitis, acute liver failure, and metabolic defects. Further studies are needed to identify and isolate pluripotent liver stem cells from human bone marrow and clarify the factors that govern cellular differen­ tiation and liver regeneration.161

KEY REFERENCES

Bernal W, Donaldson N, Wyncoll D, et al. Blood lactate as an early predictor of outcome in paracetamol-induced acute liver failure: A cohort study. Lancet 2002; 359:558-63. (Ref 85.) Bernal W, Hall C, Karvellas CJ, et al. Arterial ammonia and clinical risk factors for encephalopathy and intracranial hypertension in acute liver failure. Hepatology 2007; 46:1844-52. (Ref 67.) Davern TJ, James LP, Hinson JA, et al. Measurement of serum acetaminophen-protein adducts in patients with acute liver failure. Gastroenterology 2006; 130:687-94. (Ref 62.) Demetriou AA, Brown RS, Busuttil RW, et al. Prospective, randomized, multicenter controlled trial of a bioartificial liver in treating acute liver failure. Ann Surg 2004; 239:660-70. (Ref 155.) Fontana RJ. Acute liver failure due to drugs. Sem Liv Dis 2008; 28:17588. (Ref 27.) Houlihan DD, Newsome PN. Critical review of clinical trials of bone marrow stem cells in liver Disease. Gastroenterology 2008; 135:43850. (Ref 161.) Kumar M, Satapathy S, Monga R, et al. A randomized controlled trial of lamivudine to treat acute hepatitis B. Hepatology 2007; 45:97-101. (Ref 102.) Larson AM, Polson J, Fontana RJ, et al. Acetaminophen-induced acute liver failure: Results of a United States Multicenter prospective study. Hepatology 2005; 43:1364-72. (Ref 16.) McDiarmid SV, Goodrich NP, Harper AM, et al. Liver transplantation for status 1: The consequences of good intentions. Liver Transpl 2007; 13:699-707. (Ref 136.) O’Grady JG, Alexander GJ, Hayllar KM, et al. Early indicators of prog­ nosis in fulminant hepatic failure. Gastroenterology 1989; 97:439-45. (Ref 83.) Ostapowicz G, Fontana RJ, Schiodt FV, et al. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med 2002; 137:947-54. (Ref 8.) Stravitz RT, Kramer AH, Davern TJ, et al. Intensive care of patients with acute liver failure: Recommendations of the U.S. Acute Liver Failure Study Group. Crit Care Med 2007; 35:2498-508. (Ref 104.) Vaquero J, Fontana RJ, Larson AM, et al. Complications and use of intracranial pressure monitoring in patients with acute liver failure and severe encephalopathy. Liver Transpl 2005; 11:1581-9. (Ref 114.) Wade J, Rolando N, Philpott-Howard J, et al. Timing and a cause of bacterial infections in a liver intensive care unit. J Hosp Infection 2003; 53:144-6. (Ref 72.) Watkins PB, Seeff LB. Drug-induced liver injury: Summary of a singletopic clinical research conference. Hepatology 2006; 43:618-31. (Ref 36.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

94 Tumors and Cysts of the Liver Adrian M. Di Bisceglie and Alex S. Befeler

CHAPTER OUTLINE Primary Malignant Tumors  1569 Hepatocellular Carcinoma  1569 Intrahepatic Cholangiocarcinoma  1580 Hepatoblastoma  1582 Hemangiosarcoma  1582 Epithelioid Hemangioendothelioma  1583 Other Primary Malignant Tumors of the Liver  1583 Hepatic Metastases  1583 Benign Tumors  1584 Hepatocellular Adenoma  1584

Mass lesions of the liver occur sufficiently often that clinicians interested in liver diseases should have a thorough understanding of their presentations, diagnosis, and treatment. Hepatic mass lesions include tumors, tumor-like lesions, abscesses, cysts, hematomas, and confluent granulomas. The frequency with which each is seen varies in different geographic regions and different populations. The more common hepatic tumors and cysts and those important for other reasons are reviewed in this chapter. Hepatic tumors may originate in the liver—from hepatocytes, bile duct epithelium, or mesenchymal tissue—or spread to the liver from primary tumors in remote or adjacent organs. In adults in most parts of the world, hepatic metastases are more common than primary malignant tumors of the liver, whereas in children, primary malignant tumors outnumber both metastases and benign tumors of the liver. Except for cavernous hemangiomas, benign hepatic tumors are rare in all geographic regions and in all age groups.

PRIMARY MALIGNANT TUMORS Among primary malignant tumors of the liver, hepatocellular carcinoma is by far the most common.

HEPATOCELLULAR CARCINOMA Epidemiology

Hepatocellular carcinoma is the commonest primary malignant tumor of the liver. It is the fifth most common cancer in men and the eighth most common in women, and it ranks fourth in annual cancer mortality rates.1,2 Information on incidence is derived from an increasing but still limited number of cancer registries, and it is possible to classify countries into broad risk categories only. Moreover, in

Cavernous Hemangioma  1586 Infantile Hemangioendothelioma  1587 Other Benign Tumors of the Liver  1587 Tumor-like Hepatic Lesions  1587 Focal Nodular Hyperplasia  1587 Other Nodular Disorders  1588 Hepatic Cysts  1589 Fibrocystic Diseases of the Liver  1589 Approach to the Patient with a Hepatic Mass Lesion  1590

low-income (developing) countries, especially in subSaharan Africa, hepatocellular carcinoma is underdiagnosed and underreported, in some cases by as much as 50%. Despite these sources of inaccuracy, hepatocellular carcinoma clearly has an unusual geographic distribution (Fig. 94-1). Moreover, the tumor is not necessarily uniformly common throughout countries with a high incidence, such as China3 and Mozambique.4 The incidence of hepatocellular carcinoma has increased considerably in Japan since the 1980s, and lesser increases have been recorded in developed Western countries, including North America and Western Europe.5 Interestingly, a study from Japan has shown that the rate of hepatocellular carcinoma began to decline in 2000, presumably because of the aging of the cohort of persons infected with hepatitis C virus (HCV).6 A similar downward trend has been noted in some European countries, including France and Italy.7 By contrast, in the United States, hepatocellular carcinoma is the cancer that has been increasing in incidence most rapidly since 2000, at a time when other major cancers such as cancers of the lung, breast, prostate, and colon are decreasing.8 Considerable racial and ethnic variation exits in the incidence of hepatocellular carcinoma in the United States. The incidence among Asians is highest, almost double that of white Hispanics and more than four times higher than that of whites.9 Migrants from countries with a low incidence to areas with a high incidence of hepatocellular carcinoma usually retain the low risk of their country of origin, even after several generations in the new environment. The consequences for migrants from countries with a high incidence to those with a low incidence differ, depending on the major risk factors for the tumor in their country of origin and whether chronic hepatitis B virus (HBV) infection, if this is the major risk factor, is acquired pre­ dominantly by the perinatal or horizontal route (see later and Chapter 78).2,10,11

1569

1570

Section IX  Liver

High Intermediate Low Figure 94-1.  Incidence of hepatocellular carcinoma in different parts of the world. High, age-adjusted rate of more than 15 cases/100,000 population/ year; intermediate, age-adjusted rate of 5 to 15 cases/100,000/year; low, age-adjusted rate of fewer than 5 cases/100,000/year.

Men are generally more susceptible than women to hepatocellular carcinoma. Male predominance is, however, more obvious in populations at high risk for the tumor (mean male-to-female ratio, 3.7 : 1.0) than in those at low or intermediate risk (2.4 : 1.0).1,2 In industrialized countries, the number of men and number of women with hepatocellular carcinoma in the absence of cirrhosis are almost equal. The incidence of hepatocellular carcinoma increases progressively with advancing age in all populations, although it tends to level off in the oldest age groups.1,2 In Chinese and particularly in black African populations, however, the mean age of patients with the tumor is appreciably younger than in other populations. This finding is in sharp contrast to the age distribution in Japan, where the incidence of hepatocellular carcinoma is highest in the cohort of men ages 70 to 79 years.6 Hepatocellular carcinoma is rare in children.12,13

Clinical Features

Although the typical clinical features of hepatocellular carcinoma are well recognized (including abdominal pain and weight loss in patients with cirrhosis), more patients are now being diagnosed at an early stage, when they have no specific symptoms or signs. This trend toward earlier diagnosis is probably the result of surveillance programs in patients with chronic liver disease (see later). In faradvanced disease, patients with hepatocellular carcinoma usually present with typical symptoms and signs, and diagnosis is easy. In addition, hepatocellular carcinoma often coexists with cirrhosis,14 and the onset of hepatocellular carcinoma is marked by a sudden unexplained change in the patient’s condition. Patients with hepatocellular carcinoma often are unaware of its presence until the tumor has reached an advanced stage. The most common, and frequently first, symptom is right hypochondrial or epigastric pain. Other symptoms are listed in Table 94-1. Physical findings vary with the stage of disease. Early in the course, evidence of cirrhosis alone may be present, or abnormal findings may be absent (see Table 94-1). When the

Table 94-1  Symptoms and Signs of Hepatocellular Carcinoma SYMPTOM Abdominal pain Weight loss Weakness Abdominal swelling Nonspecific gastrointestinal symptoms Jaundice

FREQUENCY (%) 59-95 34-71 22-53 28-43 25-28 5-26

SIGN Hepatomegaly Ascites Fever Splenomegaly Wasting Jaundice Hepatic bruit

54-98 35-61 11-54 27-42 25-41 4-35 6-25

tumor is advanced at the time of the patient’s first medical visit, the liver is almost always enlarged, sometimes massively. Hepatic tenderness is common and may be profound, especially in the later stages. The surface of the enlarged liver is smooth, irregular, or frankly nodular. An arterial bruit may be heard over the tumor15; the bruit is heard in systole, rough in character, and not affected by changing the position of the patient. Although not pathognomonic, a bruit is a useful clue to the diagnosis of hepatocellular carcinoma. Less often, a friction rub may be heard over the tumor, but this sign is more characteristic of hepatic metastases or abscesses. Ascites may be present when the patient is first seen or may appear with progression of the tumor. In most patients, ascites is the result of long-standing cirrhosis and portal hypertension (see Chapter 91), but in some cases it is caused by invasion of the peritoneum by the primary tumor or

Chapter 94  Tumors and Cysts of the Liver Table 94-2  Paraneoplastic Syndromes Associated with Hepatocellular Carcinoma Carcinoid syndrome Hypercalcemia Hypertrophic osteoarthropathy Hypoglycemia Neuropathy Osteoporosis Polycythemia (erythrocytosis) Polymyositis Porphyria Sexual changes—isosexual precocity, gynecomastia, feminization Systemic arterial hypertension Thyrotoxicosis Thrombophlebitis migrans Watery diarrhea syndrome

metastases. The ascitic fluid may be blood-stained. In a small proportion of patients, hepatocellular carcinoma invades the hepatic veins, thereby causing Budd-Chiari syndrome, and tense ascites results (see Chapter 83).16 Splenomegaly, if present, reflects coexisting cirrhosis and portal hypertension. Physical evidence of cirrhosis may also be noted. Severe pitting edema of the lower extremities extending up to the groins occurs when hepatocellular carcinoma has invaded the hepatic veins and propagates into and obstructs the inferior vena cava.16 A Virchow-Trosier (supraclavic­ ular) node, Sister Mary Joseph’s (periumbilical) nodule, or enlarged axillary lymph node is rarely present. Paraneoplastic Manifestations Some of the deleterious effects of hepatocellular carcinoma are not caused by local effects of the tumor or metastases (Table 94-2). Each of the paraneoplastic syndromes in hepatocellular carcinoma is rare or uncommon. One of the more important is type B hypoglycemia, which occurs in less than 5% of patients, manifests as severe hypoglycemia early in the course of the disease,16 and is believed to result from the defective processing by malignant hepatocytes of the precursor to insulin-like growth factor II (pre-IGF II).17 By contrast, type A hypoglycemia is a milder form of glycopenia that occurs in the terminal stages of hepatocellular carcinoma (and other malignant tumors of the liver). It results from the inability of a liver extensively infiltrated by tumor, and often cirrhotic, to satisfy the demands for glucose by a large, often rapidly growing tumor and by the other tissues of the body. Another important paraneoplastic syndrome is polycythemia (erythrocytosis), which occurs in less than 10% of patients with hepatocellular carcinoma.18 This syndrome appears to be caused by the synthesis of erythropoietin or an erythropoietin-like substance by malignant hepatocytes. Patients with hepatocellular carcinoma, especially the sclerosing variety, may present with hypercalcemia in the absence of osteolytic metastases. When hypercalcemia is severe, it may result in the typical complications of hypercalcemia, including drowsiness and lethargy. The probable cause is secretion of parathyroid hormone–related protein (PTHrP) by the tumor.19 Cutaneous paraneoplastic manifestations of hepatocellular carcinoma are rare except for pityriasis rotunda (circumscripta), which may be a useful marker of the tumor in black Africans. The rash consists of single or multiple, round or oval, hyperpigmented, scaly lesions on the trunk and thighs that range in diameter from 0.5 to 25 cm.20

Diagnosis

The gold standard for the diagnosis of hepatocellular carcinoma is pathology. For practical purposes (i.e., to apply treatment), hepatocellular carcinoma can only be diagnosed in the presence of an abnormality on imaging of the liver. The development of hepatocellular carcinoma is thought to occur as a result of a multistep sequential process from a dysplastic focus of hepatocytes to a low-grade dysplastic nodule to a high-grade dysplastic nodule to early welldifferentiated hepatocellular carcinoma, and then to less differentiated states.21,22 In early hepatocellular carcinoma, particularly when a needle biopsy specimen is examined, controversy may exist among pathologists as to whether a particular specimen is consistent with dysplasia or carcinoma. Dysplastic nodules and even regenerative cirrhotic nodules can be seen on imaging studies and are potentially confused with hepatocellular carcinoma. Although there are specific imaging features based on the enhancement patterns with dynamic imaging of dysplastic nodules and hepatocellular carcinoma (see later), some overlap occurs.23,24 Nevertheless, there is a growing consensus, based on guidelines from the major European and American liver societies and now backed up by published experience, that the diagnosis of hepatocellular carcinoma can be made in the appropriate clinical setting based on specific imaging characteristics, with or without an elevated serum alpha fetoprotein (AFP) level.24-27 Serum Tumor Markers Serum tumor markers generally are not diagnostic for hepatocellular carcinoma by themselves but can be used in conjunction with imaging findings to diagnose hepatocellular carcinoma. Additionally, they may raise the suspicion for hepatocellular carcinoma and lead to more sensitive and serial imaging of the liver. Conventional liver biochemical tests do not distinguish hepatocellular carcinoma from other hepatic mass lesions or cirrhosis. Many of the substances synthesized and secreted by hepatocellular carcinoma are not biologically active. Nevertheless, a few are produced by a sufficiently large proportion of tumors to warrant their use as serum markers for hepatocellular carcinoma. The most helpful of these markers is AFP (Table 94-3). Alpha Fetoprotein.  AFP is an α1-globulin normally present in high concentrations in fetal serum but in only minute amounts thereafter. Reappearance of high serum levels of AFP strongly suggests the presence of hepatocellular carcinoma (or hepatoblastoma [see later]),28 especially in populations in which hepatocellular carcinoma is most prevalent: The great majority of Chinese and black African patients have a raised serum concentration of AFP (>10 ng/mL), and approximately 75% have a diagnostic level (>500 ng/mL). These percentages are lower in populations at low or intermediate risk for the tumor, in which the sensitivity ranges from 25% to 65%, with a specificity of 79% to 95% and cutoff values for an elevated and diagnostic level of 16 and 200 ng/mL, respectively.29-35 With higher levels of AFP, the confidence in the diagnosis of hepatocellular carcinoma is greater. Although levels higher than 500 ng/mL usually indicate hepatocellular carcinoma, they sometimes can be seen in patients with active viral hepatitis. In the setting of a cirrhotic patient with a hepatic mass lesion larger than 2 cm in diameter and suggestive features of hepatocellular carcinoma, an AFP level higher than 200 ng/mL is con­ sidered diagnostic for hepatocellular carcinoma.25,26,33,36,37 The mean serum value of AFP in affected patients in regions with a high incidence of hepatocellular carcinoma is

1571

1572

Section IX  Liver Table 94-3  Tumor Markers of Hepatocellular Carcinoma* MARKER

SENSITIVITY (%)

SPECIFICITY (%)

CUTOFF†

Alpha fetoprotein (AFP) High-incidence populations Low-incidence populations Des-γ-carboxyprothrombin AFP-L3

80-90 25-65 28-89 71

90 79-95 87-95 63

>10 ng/mL 16-200 ng/mL 10-125 mAU/mL >10%

COMMENTS Most widely available; most extensively studied Precise cutoff value not determined Adds specificity to AFP if AFP level is 10-200 ng/mL

*Note that sensitivity and specificity rates vary both with the population under study and with the absolute level of the marker. Thus, the specificity of a markedly elevated serum AFP level in patients with cirrhosis greatly exceeds the sensitivity of mildly elevated levels in persons without cirrhosis (and similarly for the other markers). † Cutoff is the value above which the level is considered abnormal. AFP-L3, Lens culinaris agglutinin reaction fraction of alpha fetoprotein; mAU, milliarbitrary units.

60,000 to 80,000 ng/mL, compared with approximately 8,000 ng/mL in regions with a low or intermediate incidence of the tumor. Raised serum values range over six orders of magnitude, although an AFP concentration higher than 1 million ng/mL is rare. False-positive results also may occur in patients with tumors of endodermal origin, nonseminomatous germ cell tumors, and pregnancy. A progressively rising serum AFP concentration is highly suggestive of hepatocellular carcinoma. AFP is not essential to hepatocarcinogenesis, and thus not all hepatocellular carcinomas produce AFP. The levels of AFP appear to be affected by ethnicity, underlying cause of liver disease, and tumor stage.30,33 Synthesis of AFP by a tumor is permanent and age-related; the younger the patient, the more likely the serum value is to be raised and the higher the level attained. According to the American Association for the Study of Liver Diseases (AASLD) guidelines, hepatocellular carcinoma can be diagnosed with confidence in patients with a serum AFP level higher than 200 ng/mL and a mass in the liver.25 An AFP level higher than about 500 ng/mL predicts worse outcomes with liver transplantation compared with lower levels.38 Attempts to correlate the degree of differentiation of hepatocellular carcinoma with production of AFP have produced conflicting results. Because both false-positive and false-negative results are obtained when AFP is used as a serum marker for hepatocellular carcinoma, the search for an ideal marker continues. Alternative markers have not proved to be more useful than AFP. Fucosylated Alpha Fetoprotein.  AFP is heterogeneous in structure. Its microheterogeneity results from differences in the oligosaccharide side chain and accounts for the differential affinity of the glycoprotein for lectins. AFP secreted by malignant hepatocytes contains unusual and complex sugar chains that are not found in AFP produced by nontransformed hepatocytes. One variant, Lens culinaris agglutinin reactive fraction (AFP-L3), appears to improve the specificity of AFP, particularly AFP serum levels from 10 to 200 ng/mL.39,40 The recommended cutoff value for AFP-L3 to diagnose hepatocellular carcinoma is higher than 10%, although the specificity varies depending on the absolute level of AFP. A Western series has suggested that a cutoff value of 35% is necessary to achieve 100% specificity.40 Therefore, AFP-L3 is not sufficiently validated to confirm the diagnosis of hepatocellular carcinoma without other supporting findings, such as suggestive imaging. Des-γ-Carboxy Prothrombin.  Serum concentrations of desγ-carboxy prothrombin (DCP) (also known as prothrombin

produced by vitamin K absence or antagonist II [PIVKA II]) are raised in most patients with hepatocellular carcinoma.41 DCP is an abnormal prothrombin that is thought to result from a defect in the post-translational carboxylation of the prothrombin precursor in malignant cells.42 In Western populations, DCP may be a better marker than, or at least a complementary marker to, AFP.43-45 In black Africans, however, DCP is less sensitive and less specific than AFP.46 The appropriate cutoffs are not well established, and thus the precise role of DCP in the diagnosis of hepatocellular carcinoma requires validation. Other Markers.  Multiple other potential serum markers for hepatocellular carcinoma are in the exploratory phase of evaluation, including glypican 3, Golgi protein 73, hepatocyte growth factor, insulin growth factor 1, transforming growth factor-β1, and proteomic profiling using surfaceenhanced laser desorption/ionization time-of-flight (SELDITOF) mass spectrometry.47-51 All these novel markers have been shown to be elevated in patients with hepatocellular carcinoma compared with those with only chronic liver disease, but clear cutoff values and comparisons with other markers have not been established. Some of these markers may be complementary to established markers, although none of them has an established high throughput method of measurement, as required for a clinical test. The roles of these markers in the diagnosis of hepatocellular carcinoma await further study.

Imaging

The diagnosis of hepatocellular carcinoma generally requires imaging evidence of a focal lesion in the liver, although large infiltrating lesions can also be diagnostic. Arterial hyperenhancement, particularly seen on dynamic contrast imaging of the liver, is observed because the blood supply of hepatocellular carcinoma comes from newly formed abnormal arteries (neoangiogenesis).23,52,53 As a nodule transforms from low- to high-grade dysplasia and then to hepatocellular carcinoma, the primary blood supply shifts from portal to arterial—especially new abnormal arterial branches that produce characteristic findings on dynamic contrast imaging of the liver.27 Ultrasonography Ultrasonography detects most hepatocellular carcinomas but may not distinguish this tumor from other solid lesions in the liver. As with all imaging methods, the sensitivity increases with increasing size of the lesion. A systematic review of eight studies using histologic reviews of liver explants has shown that ultrasound has fair sensitivity

Chapter 94  Tumors and Cysts of the Liver (pooled estimate, 48%; 95% confidence interval [CI], 34% to 62%) with good specificity, estimated at 97% (95% CI, 95% to 98%).24 Advantages of ultrasonography include safety, availability, and cost-effectiveness, although it has the drawbacks of being nonstandardized and examinerdependent. Body habitus, particularly obesity, may limit the sensitivity of this test. Approximately two thirds of symptomatic hepatocellular carcinomas are uniformly hyper­ echoic, whereas the remainder are partly hyperechoic and partly hypoechoic.54 Small tumors are uniformly hypoechoic. The ultrasonographic appearance is influenced by the presence of fat, calcium, and necrosis. Tumors located immediately under the right hemidiaphragm may be difficult to detect. In Japanese patients in particular, hepatocellular carcinoma may have a well-defined, even thick capsule, which can be seen on ultrasonography. Ultrasonography with Doppler technology is useful for assessing the patency of the inferior vena cava, portal vein and its larger branches, hepatic veins, and biliary tree. Dynamic contrast-enhanced Doppler ultrasonography with intra-arterial infusion of CO2 microbubbles and intravenous enhanced color Doppler ultrasonography are refinements that, by characterizing hepatic arterial and portal venous flow in tumorous nodules, facilitate the diagnosis of malignant and benign hepatic nodules.55 These techniques are not often performed in the United States. Computed Tomography Multiphase, also called dynamic, helical computed tomography (CT) is the imaging technique of choice for the diagnosis of hepatocellular carcinoma.24,54,55 CT during arterial portography is also helpful but rarely done because it is invasive. Phases in dynamic contrast-enhanced CT can include noncontrast, arterial, portal venous, and delayed phases. The classic and most diagnostic pattern for hepatocellular carcinoma is a combination of enhancement in the arterial phase (with the uninvolved liver lacking enhancement), loss of central nodule enhancement compared with the uninvolved liver (washout), and capsular enhancement in the portal-venous and delayed phases (Fig. 94-2).25,56

Noncontrast

Arterial

Portal venous

Delayed

When the lesion is larger than 2 cm in diameter, this pattern has almost 100% specificity for hepatocellular carcinoma.36,37,56 When the nodule is 1 to 2 cm, guidelines recommend a second type of dynamic imaging (magnetic resonance imaging [MRI] or contrast ultrasonography) to confirm the diagnosis of hepatocellular carcinoma, although the specificity of one dynamic study is higher than 90%.57 CT often finds so-called hypervascular-only lesions, which enhance in the arterial phase and become isodense to the surrounding liver in the portal-venous and delayed phases. These lesions may be dysplastic nodules, arterial portal shunts, atypical hemangiomas, hepatocellular carcinoma, confluent fibrosis, or aberrant venous drainage. When less than 2 cm in diameter, only about 30% are hepatocellular carcinomas, which grow over time. Other causes disappear or remain stable on follow-up studies. Current guidelines recommend biopsy of lesions larger than 1 cm if the serum AFP level is less than 200 ng/mL and serial imaging for lesions smaller than 1 cm.58 Hepatocellular carcinoma may also have other patterns on CT, such as washout only on delayed imaging, a hypovascular nodule, or a fat-containing nodule.27,58 Overall, the pooled estimate of sensitivity and specificity for detecting hepatocellular carcinoma by CT is 67.5% (95% CI, 55% to 80%) and 92.5% (95% CI, 89% to 96%), respectively. Dynamic CT is also useful for detecting invasion into the portal or hepatic veins and identifying the location and number of tumors; these findings are critical for planning treatment. Magnetic Resonance Imaging Dynamic MRI using gadolinium contrast agents provides another way of distinguishing hepatocellular carcinoma from normal liver tissue. The performance of MRI and the findings on multiphase contrast enhancement are similar to those described for CT (Fig. 94-3). Typically, the signal intensity on T1-weighted images is low.27,54 The pooled estimate of sensitivity and specificity for detecting hepatocellular carcinoma by MRI is 80.6% (95% CI, 70% to 91%) and 84.8% (95% CI, 77% to 93%), respectively.24 MRI may be slightly superior overall to CT, although local expertise should dictate the choice of imaging technique. Hepatic Angiography Since the advent of CT and MRI, the diagnostic role of hepatic angiography has decreased. Digital subtraction angiography is helpful for recognizing small hyper­vascular hepatocellular carcinomas but may miss early, welldifferentiated hypovascular tumors. Hepatocellular carcinomas often are densely vascular, although multinodular tumors may be relatively avascular.59 The arteries in the tumor are irregular in caliber and do not taper in the usual way, and the smaller branches may show a bizarre pattern. The hepatic veins fill early, and retrograde filling of the portal veins results from the presence of arteriovenous anastomoses within the tumor. An additional finding is a delay in capillary emptying, which is seen as a blush. The center of some large tumors may be avascular as a result of necrosis or, less often, hemorrhage. Angiography is essential for delineating the hepatic arterial anatomy in planning embolization or chemoembolization of the tumor or infusion of cytotoxic drugs directly into the hepatic artery or its branches (see later).

Laparoscopy Figure 94-2.  Dynamic computed tomography scan of a patient with hepatocellular carcinoma showing no lesion in the noncontrast phase, an enhancing lesion in the arterial phase of contrast administration, and a faint lesion in the portal venous phase seen better in the delayed phase.

Laparoscopy can be used to detect peritoneal and other extrahepatic spread, ascertain whether the nontumorous part of the liver is cirrhotic, and obtain biopsies under direct vision.

1573

1574

Section IX  Liver T2

T1

Delayed venous

Arterial Venous Figure 94-3.  Multiphasic magnetic resonance imaging of the liver showing hepatocellular carcinoma with characteristic features, including hyperintensity (arrow) on T2-weighted image (top left panel) but not on T1-weighted image (top right panel), enhancement during the arterial phase of contrast administration (bottom left panel), with central washout of contrast and capsular enhancement during the venous and delayed phases (bottom middle and right panels).

Pathology

Definitive diagnosis of hepatocellular carcinoma depends on demonstrating the typical histologic features. Suitable samples generally can be obtained by percutaneous biopsy or fine-needle aspiration. The yield and safety of the procedure can be increased by directing the needle under ultrasonographic or CT guidance. Laparoscopically directed biopsy is an alternative approach. Needle biopsy of the tumor carries a small but definite risk of spread along the needle track. Gross Appearance Hepatocellular carcinoma may take one of three forms— nodular, massive, or diffusely infiltrating. The nodular variety of hepatocellular carcinomas is most common and usually coexists with cirrhosis. It is characterized by numerous round or irregular nodules of various sizes scattered throughout the liver; some of the nodules are confluent. The massive type is characterized by a large circumscribed mass, often with small satellite nodules. This type of tumor is most prone to rupture and is more common in younger patients with a noncirrhotic liver. In the rare diffusely infiltrating variety, a large part of the liver is infiltrated homogeneously by indistinct minute tumor nodules, which may be difficult to distinguish from the regenerating nodules of cirrhosis that are almost invariably present. The portal vein and its branches are infiltrated by tumor in up to 70% of cases seen at autopsy; the hepatic veins and bile ducts are invaded less often. Microscopic Appearance Hepatocellular carcinoma is classified histologically into well-differentiated, moderately differentiated, and undifferentiated (pleomorphic) forms.60

Well-Differentiated Appearance.  Despite the aggressive nature and poor prognosis of hepatocellular carcinoma, most tumors are well differentiated. Trabecular and acinar (pseudoglandular) varieties occur, sometimes in a single tumor. In the trabecular variety, the malignant hepatocytes grow in irregular anastomosing plates separated by often inconspicuous sinusoids lined by flat cells resembling Kupffer cells. The trabeculae resemble those of normal adult liver but often are thicker and may be composed of several layers of cells. Scanty collagen fibers may be seen adjacent to the sinusoid walls. The malignant hepatocytes are polygonal, with abundant, slightly granular cytoplasm that is less eosinophilic than that of normal hepatocytes. The nuclei are large and hyperchromatic, with prominent nucleoli. Bile production is the hallmark of hepatocellular carcinoma, regardless of the pattern. Gland-like structures are present in the acinar variety. The structures are composed of layers of malignant hepatocytes surrounding the lumen of a bile canaliculus, which may contain inspissated bile. A tubular or pseudopapillary appearance may be produced by degeneration and loss of cells, or cystic spaces may form in otherwise solid trabeculae. The individual cells may be more elongated and cylindrical than in the trabecular variety. Moderately Differentiated Appearance.  Solid, scirrhous, and clear cell varieties of hepatocellular carcinoma are described. In the solid variety, the cells usually are small, although they vary considerably in shape. Pleomorphic multinucleated giant cells occasionally are present. The tumor grows in solid masses or cell nests. Evidence of bile secretion is rare, and connective tissue is inconspicuous. Central ischemic necrosis is common in larger tumors. In the scirrhous variety, the malignant hepatocytes grow in

Chapter 94  Tumors and Cysts of the Liver narrow bundles separated by abundant fibrous stroma. Duct-like structures occasionally are present. In most tumors, the cells resemble hepatocytes. In an occasional tumor, the malignant hepatocytes are predominantly or exclusively clear cells. More often, tumors contain areas of clear cells. The appearance of these cells results from a high glycogen or, in some cases, fat content. Undifferentiated Appearance.  The cells are pleomorphic, varying greatly in size and shape. The nuclei also are extremely variable. Large numbers of bizarre-looking giant cells are present. The cells may be spindle-shaped, resembling those of sarcomas. Globular hyaline structures may be seen in all types of hepatocellular carcinoma. These structures reflect the presence of AFP, α1-antitrypsin, or other proteins. Mallory’s hyaline occasionally is present. Progenitor Cell Hepatocellular Carcinoma.  A class of primary liver cancer appears to have its origins in progenitor cells, the stem cells of the liver, located in association with the canals of Hering. Progenitor cell activation is seen in association with chronic viral hepatitis and cirrhosis, presumably relegated to senescence of hepatocytes. These tumors may appear morphologically like typical hepatocellular carcinoma or mixed cholangiohepatocellular carcinoma. Tumor cells stain positively for cytokeratin 19, and the tumor appears to have a more aggressive course than typical hepatocellular carcinoma.61

Metastases

Extrahepatic metastases are present at autopsy in 40% to 57% of patients with hepatocellular carcinomas.62 The most common sites are the lungs (up to 50% in some reports) and regional lymph nodes (approximately 20%). The adrenal glands are frequently involved.

Fibrolamellar Hepatocellular Carcinoma

The fibrolamellar variant of hepatocellular carcinoma typically occurs in young patients, has an approximately equal gender distribution, does not secrete AFP, is not caused by chronic hepatitis B or C, and almost always arises in a noncirrhotic liver.63-65 Fibrolamellar hepatocellular car­ cinoma is more often amenable to surgical treatment and therefore generally carries a better prognosis than that for conventional hepatocellular carcinoma. It does not, however, respond to chemotherapy any better than other forms of hepatocellular carcinoma. The hepatocytes are characteristically plump, deeply eosinophilic, and encompassed by abundant fibrous stroma composed of thin, parallel fibrous bands that separate the cells into trabeculae or nodules. The cytoplasm is packed with swollen mitochondria and, in approximately half of the tumors, contains pale or hyaline bodies. Nuclei are prominent, and mitoses are rare.

Staging

Accurate staging of hepatocellular carcinoma is necessary for prognostication and also to assist with selection of therapy. Determining the optimal staging system for hepatocellular carcinoma has been controversial, in part because it needs to take into account both the severity of the underlying liver disease and the size and degree of spread of the tumor. As with all cancers, the TNM (tumornode-metastasis) system can be used to stage hepatocellular carcinoma, but this system does not factor in the underlying liver disease. A study66 comparing the usefulness of seven staging systems, including the Okuda, TNM, Cancer of the Liver Italian Program (CLIP), Barcelona Clinic Liver Cancer

(BCLC), Chinese University Prognostic Index (CUPI), Japanese Integrated Staging (JIS), and Group d’Etude et Traitement du Carcinome Hépatocellulaire (GETCH) systems in a cohort of patients from the United States, has found the BCLC staging system to have the best independent predictive power for survival. The BCLC system has been adopted by the AASLD for use in its practice guidelines on management of hepatocellular carcinoma.25 This staging classification also includes a treatment schedule based on stage (Fig. 94-4).67

Causes and Pathogenesis

In contrast to many other malignancies, for which risk factors can only sometimes be identified, the immediate cause of hepatocellular carcinoma can usually be identified and is most commonly chronic viral hepatitis or cirrhosis. Hepatocellular carcinoma is multifactorial in cause and complex in pathogenesis. Four major causative factors have been identified (Table 94-4). The differing blend of risk factors in various parts of the world may explain, in part, the diverse biologic characteristics of hepatocellular carcinoma in various populations.68 Hepatitis B Virus Some 387 million carriers of HBV exist in the world today, and hepatocellular carcinoma will develop in as many as 25% of them (see Chapter 78). HBV accounts for up to 80% of hepatocellular carcinomas, which occur with high frequency in East Asian and African populations.68,69 Persistent HBV infection antedates the development of hepatocellular carcinoma by several to many years, an interval commensurate with a cause and effect relationship between the virus and the tumor. Indeed, in at-risk populations, the HBV carrier state is largely established in early childhood by perinatal or horizontal infection.70,71 Approximately 90% of children infected at this stage of life become chronic carriers of the virus, and these early-onset carriers face a lifetime relative risk for developing hepatocellular carcinoma of more than 100, compared with uninfected controls.72 An effective vaccine against HBV has been available since the early 1980s and, in countries in which this vaccine has been included in the expanded program of immunization for a sufficient length of time, the HBV carrier rate among children has decreased by 10-fold or more. Studies in Taiwan, where universal immunization was started in 1984

Table 94-4  Risk Factors for Hepatocellular Carcinoma Major Risk Factors Chronic hepatitis B virus infection Chronic hepatitis C virus infection Cirrhosis Dietary exposure to aflatoxin B1 Other Liver Conditions α1-Antitrypsin deficiency Hemochromatosis Membranous obstruction of the inferior vena cava Type 1 and type 2 glycogen storage disease Type 1 hereditary tyrosinemia Wilson disease Inherited Conditions Not Associated with Liver Disease Ataxia-telangiectasia Hypercitrullinemia Other Factors Cigarette smoking Diabetes mellitus Oral contraceptive steroids

1575

1576

Section IX  Liver HCC

Stage 0 PST 0 and Child-Pugh A

Stage D PST >2 or Child-Pugh C

Stage A–C PST 0-2 and Child-Pugh A-B

Single nodule <2 cm or carcinoma in situ

Early stage (A) Single or 3 nodules <3 cm each, PST 0

Intermediate stage (B) Multinodular, PST 0

Advanced stage (C) Portal invasion, N1, M1, PST 1-2

End stage (D)

3 nodules ≤3 cm

Single

Portal pressure/serum bilirubin level Increased Normal

Comorbid disease

No

Resection

Liver transplantation (CLT/LDLT)

Yes PEI/RFA

Curative treatments (30% of cases) 5-yr survival: 40%-70%

TACE

Sorafenib

Consider entry into randomized controlled trial (50%) 3-yr survival: 10%-40%

Symptomatic treatment (20%) Survival: <3 mo

Figure 94-4.  Barcelona Clinic Liver Cancer (BCLC) staging classification and treatment schedule. Staging is based on tumor size and spread, the patient’s performance status (PST) on a scale of 0 (good) to >2 (poor), and liver function as assessed by the Child-Pugh class (see Chapter 90). Patients with very early (stage 0) hepatocellular carcinoma (HCC) are optimal candidates for surgical resection. Patients with early (stage A) HCC are candidates for radical therapy (resection, cadaveric liver transplantation [CLT] or live-donor liver transplantation [LDLT], or local ablation via percutaneous ethanol injection [PEI] or radiofrequency ablation [RFA]). Patients with intermediate (stage B) HCC benefit from transarterial chemoembolization (TACE). Patients with advanced HCC, defined as the presence of macroscopic vascular invasion, extrahepatic spread, or cancer-related symptoms (PST 1 or 2) (stage C), benefit from sorafenib. Patients with end-stage disease (stage D) should receive symptomatic treatment. The treatment strategy will transition from one stage to another when treatment fails or is contraindicated. M, metastasis stage; N, nodal stage. (Adapted from Llovet J, Di Bisceglie A, Bruix J, et al. Design and endpoints of clinical trials in hepatocellular carcinoma. J Natl Cancer Inst 2008; 100:698-711.)

and where the rate of HBV carriage among children has decreased by more than 10-fold, have already shown a 70% reduction in the mortality rate from hepatocellular carcinoma in children in the vaccinated age groups.73 This finding gives promise for the ultimate eradication of HBVinduced hepatocellular carcinoma and provides further evidence of the causal role of the virus in the development of this tumor. HBV DNA is integrated into cellular DNA in approximately 90% of HBV-related hepatocellular carcinomas.74 The sites of chromosomal insertion appear to be random, and whether viral integration is essential for hepatocarcinogenesis is still uncertain. The virus appears to be directly and indirectly carcinogenic.75 Possible direct carcinogenic effects include cis-activation of cellular genes as a result of viral integration, changes in the DNA sequences flanking the integrated viral DNA, transcriptional activation of remote cellular genes by HBV-encoded proteins, particularly the X protein, and effects resulting from viral mutations. The transcriptional activity of the HBV X protein may be mediated by interaction with specific transcription factors, activation of the mitogen-activated protein (MAP) kinase and Janus

kinase–signal transducer and activator of transcription (JAK/STAT) pathways, an effect on apoptosis, and modulation of DNA repair. Studies have shown a clear link between the amount of HBV replication (measured as serum level of HBV DNA [viral load]) and subsequent risk of hepatocellular carcinoma. The long-term risk of hepatocellular carcinoma increases markedly in patients with serum HBV DNA levels higher than 104 copies/mL.76 A randomized controlled trial of antiviral therapy has also shown a reduction in the incidence of hepatocellular carcinoma in association with reductions in serum levels of HBV DNA on therapy.77 Indirect carcinogenic effects are the result of the chronic necroinflammatory hepatic disease, in particular cirrhosis, induced by the virus. The increased hepatocyte turnover rate resulting from continuous or recurring cycles of cell necrosis and regeneration acts as a potent tumor promoter. In addition, the distorted architecture characteristic of cirrhosis contributes to the loss of control of hepatocyte growth, and hepatic inflammation generates mutagenic reactive oxygen species. The transgenic mouse model of Chisari and coworkers has provided indirect

Chapter 94  Tumors and Cysts of the Liver support for the role of prolonged hepatocyte injury in hepatocarcinogenesis.78 Hepatitis C Virus Approximately 170 million people in the world today are chronically infected with HCV and are at greatly increased risk for the development of hepatocellular carcinoma. In Japan, Italy, and Spain, HCV is the cause of about 75% of hepatocellular carcinomas, and, in other industrialized countries, HCV infection, often in combination with alcohol abuse, is emerging as a major cause of the tumor.68,79 Patients with HCV-induced hepatocellular carcinoma generally are older than those with HBV-related tumors, and it is likely that the HCV infection is acquired mainly in adult life. Almost all HCV-induced hepatocellular carcinomas arise in cirrhotic livers, and most of the exceptions are in livers with chronic hepatitis and fibrosis. This observation strongly suggests that chronic hepatic parenchymal disease plays a key role in the genesis of HCV-related tumors. Because the HCV genome does not integrate into host DNA, the virus would have to exert a direct carcinogenic effect by some other means. Cirrhosis In all parts of the world, hepatocellular carcinoma frequently coexists with cirrhosis.80 All causative forms of cirrhosis may be complicated by tumor formation. A long-term follow-up study of 2126 U.S. military veterans with cirrhosis found that hepatocellular carcinoma developed in 100 (4.7%) over an average period of 3.6 years.80 The calculated rate was 1.3/100 patient-years. Risk factors for hepatocellular carcinoma included obesity, low platelet count, and the presence of antibody to hepatitis B core antigen. A similar study from Italy found an incidence of hepatocellular carcinoma of 3.7/100 patient-years among persons with HCV infection and 2.0/100 patient-years among persons with HBV infection. Older age and male gender were confirmed as risk factors among patients with cirrhosis.81 Aflatoxin B1 Dietary exposure to aflatoxin B1, derived from the fungi Aspergillus flavus and Aspergillus parasiticus, is an important risk factor for hepatocellular carcinoma in parts of Africa and Asia. These molds are ubiquitous in nature and contaminate a number of staple foodstuffs in tropical and subtropical regions (see Chapter 87). Epidemiologic studies have shown a strong correlation between the dietary intake of aflatoxin B1 and incidence of hepatocellular carcinoma.82 Moreover, aflatoxin B1 and HBV interact synergistically in the pathogenesis of hepatocellular carcinoma. Heavy dietary exposure to aflatoxin B1 may contribute to hepatocarcinogenesis through an inactivating mutation of the third base of codon 249 of the TP53 tumor suppressor gene.83,84 Other Liver Conditions Hepatocellular carcinoma develops in as many as 45% of patients with hemochromatosis (see Chapter 74).85 Malignant transformation was thought previously to occur only in the presence of cirrhosis (and is certainly more likely to do so), but this complication also has been reported in patients without cirrhosis.86 Excessive free iron in tissues may be carcinogenic, perhaps by generating mutagenic reactive oxygen species.87 Further support for this theory comes from the observations that black Africans with dietary iron overload are at increased risk of hepatocellular carcinoma88 and that rats fed a diet high in iron develop iron-free dysplastic foci and hepatocellular carcinoma in the absence of

cirrhosis.89 Hepatocellular carcinoma develops occasionally in patients with Wilson disease, but only in the presence of cirrhosis (see Chapter 75).90 Malignant transformation has been attributed to the cirrhosis but also may result from oxidant stress secondary to the accumulation of copper in the liver.91 Hepatocellular carcinoma also may develop in patients with other inherited metabolic disorders that are complicated by cirrhosis, such as α1-antitrypsin deficiency and type 1 hereditary tyrosinemia, and in patients with certain inherited diseases in the absence of cirrhosis—for example, type 1 glycogen storage disease (see Chapter 76). Hepatocellular carcinoma develops in approximately 40% of patients with membranous obstruction of the inferior vena cava, a rare congenital or acquired anomaly (see Chapter 83). Continuous cycles of hepatocyte necrosis followed by regeneration resulting from the severe and unremitting hepatic venous congestion render the cells susceptible to environmental mutagens, as well as to spontaneous mutations.92 More recently, the role of obesity, diabetes mellitus, and fatty liver disease have come to be recognized in the causation of hepatocellular carcinoma,93-95 although the mechanisms whereby these overlapping conditions contribute to causing hepatocellular carcinoma is unknown. Certainly, cirrhosis caused by nonalcoholic steatohepatitis may give rise to hepatocellular carcinoma, but it appears that these risk factors may also be additive to chronic viral hepatitis and cirrhosis. A statistically significant correlation between the use of oral contraceptive steroids and the occurrence of hepatocellular carcinoma has been demonstrated in countries in which the incidence of hepatocellular carcinoma is low and no overriding risk factor for development of the tumor is present.96 Epidemiologic evidence of a link between cigarette smoking and the occurrence of hepatocellular carcinoma is conflicting, although most of the evidence suggests that smoking is a minor risk factor.97 Heavy smokers have an approximately 50% higher risk than nonsmokers. The incidence of hepatocellular carcinoma is increased in patients with human immunodeficiency virus (HIV) infection compared with controls in the general population, presumably because of the increased rate of chronic viral hepatitis in the HIV-positive population.98 Although the aforementioned risk factors have been identified, the precise mechanisms whereby they lead to hepatocellular carcinoma still need to be elucidated. Multiple cellular pathways are involved in causing unconstrained proliferation of hepatocytes and increased angiogenesis against a background of chronic liver disease. These pathways have become the targets for new molecular therapies against hepatocellular carcinoma (Table 94-5).99

Table 94-5  Key Molecular Pathways Involved in Hepatocarcinogenesis Angiogenic signaling Epigenetic promoter methylation and histone acetylation Growth factor-stimulated receptor tyrosine kinase JAK/STAT signaling PI3-kinase/AKT/mTOR p53 and cell cycle regulation Ubiquitin-proteasome Wnt/β-catenin JAK/STAT, janus kinase/signal transducers and activators of transcription; mTOR, mammalian target of rapamycin. Adapted from Roberts L. Emerging experimental therapies for hepatocellular carcinoma: What if you can’t cure? In: McCullough A, ed. AASLD Postgraduate Course, 2007. Boston: AASLD, 2007; p 185.

1577

1578

Section IX  Liver Natural History and Prognosis

Symptomatic hepatocellular carcinoma carries a grave prognosis; in fact, the annual incidence and mortality rates for the tumor are almost identical. The main reasons for the poor outcome are the extent of tumor burden when the patient is first seen and the frequent presence of coexisting cirrhosis and hepatic dysfunction. The natural history of hepatocellular carcinoma in its florid form is one of rapid progression, with increasing hepatomegaly, abdominal pain, wasting, and deepening jaundice, and with death ensuing in two to four months. In industrialized countries, however, the tumor appears to run a more indolent course with longer survival times.100 Rare cases of spontaneous tumor regression have been reported.

Treatment

Important advances in the treatment of hepatocellular carcinoma have occurred since the 1980s; these advances include randomized controlled trials that support the benefits of certain treatments such as chemoembolization and the multikinase inhibitor sorafenib. Overwhelming evidence supports the superiority of liver transplantation over other therapies for patients with portal hypertension and cirrhosis. Because hepatocellular carcinoma is usually a combination of two diseases—the underlying liver disease (usually cirrhosis with varying degrees of decompensation) and the cancer itself—both factors must be taken into account when selecting treatment. When presented with a patient with hepatocellular carcinoma, the clinician should decide which is the best initial therapy: surgical resection or liver transplantation, if the patient is a candidate for either; ethanol or radiofrequency ablation, if possible based on the size of the tumor; chemoembolization; and, if the tumor is too advanced, sorafenib or a clinical trial. Table 94-6 describes the treatment options for hepatocellular carcinoma. The BCLC staging classification and treatment schedule can help guide the clinician in choosing the most appropriate treatment (see Fig. 94-4). Surgical Resection Surgical therapy, whether by tumor resection or liver transplantation, offers the best chance of cure for hepatocellular

Table 94-6  Treatment Options for Hepatocellular Carcinoma Modality

Comments

Surgical resection

Curative but limited to noncirrhotic patients and cirrhotic patients without portal hypertension May be technically difficult High recurrence rate Successful in selected patients (Milan criteria; see text and Chapter 95) Requires lifelong immunosuppression Expensive and not available worldwide Potentially curative for small tumors, including multiple tumors High recurrence rate Prolongs survival in unresectable tumors if hepatic function is preserved Palliative only; can be used as an adjunct to surgical resection or transplantation Drug toxicity common Sorafenib is the first such agent shown to improve patient survival

Liver transplantation

Alcohol injection and radiofrequency ablation Chemoembolization Chemotherapy Targeted molecular therapies

carcinoma. For resection to be considered, the tumor should be confined to one lobe of the liver, favorably located, and, ideally, the nontumorous liver tissue should not be cirrhotic. Expert surgical centers can achieve five- and ten-year survival rates of 40% and 26%, respectively, with a mean tumor diameter of 8.8 cm in noncirrhotic patients.101 Unfortunately, these patients represent less than 5% of Western cases.102,103 Resection is also effective if the tumor is limited to the left lobe or a portion of the right lobe, thereby permitting a segmental resection if the patient has Child (ChildPugh) class A cirrhosis, the serum bilirubin level is normal, and portal hypertension is not present (based on imaging, a normal platelet count, and lack of varices on endoscopy or on direct measurement of the hepatic venous pressure gradient). Using these criteria, five-year survival rates of 50% can be achieved. In parts of the world where liver transplantation is not available, surgical resection is a viable option, particularly for Child class A patients without portal hypertension and with a Model for End-stage Liver Disease (MELD) score of 9 (see Chapter 90). All the tumor nodules need to be removed with negative margins, and the patient needs to be left with enough functional liver volume (usually defined as ≅40%) to survive the postoperative period.104-106 Overall, resection is feasible in only approximately 15% of patients. Resection performed at expert surgical centers carries an operative mortality rate of less than 5%, but at low volume centers the mortality rate is almost three times greater.107 Unfortunately, recurrence after resection occurs in more than 50% in the long term, and salvage liver transplantation is rarely possible.108 Liver Transplantation Liver transplantation is performed in patients in whom the tumor is not resectable but is confined to the liver or in whom advanced cirrhosis and poor liver function preclude resection (see Chapter 95).25 Liver transplantation is the ideal therapy for hepatocellular carcinoma because it provides the largest possible resection margin, removes the remaining liver, which is at high risk for de novo tumors, and replaces the dysfunctioning liver. Liver transplantation can fail in patients with extrahepatic tumor, which tends to grow rapidly under the influence of post-transplantation immunosuppression. Because the availability of donor livers is limited, the consensus is that the outcomes of liver transplantation for hepatocellular carcinoma should be similar to those for other indications for liver transplantation and superior to those for other treatments for hepatocellular carcinoma. Several large series have demonstrated that if one selects candidates based on the Milan criteria—a single lesion up to 5 cm in size or two to three lesions, each up to 3 cm, with no large-vessel vascular invasion or metastasis—the five-year survival rate is 70% to 75%, and the tumor recurrence rate is 10% to 15%.102,109-111 These criteria led to the hepatocellular carcinoma MELD exception pathway, which was adopted in the United States in 2002. As a result of the change, the frequency of hepatocellular carcinoma as an indication for liver transplantation rose from 4.6% to 26% of the total adult liver transplant population. Additionally, progression of the tumor beyond the Milan criteria before a patient undergoes transplantation has largely been eliminated.38,112 In other parts of the world, waiting times before transplantation remain critical, and when the waiting time increases to one year, as many as one half of patients will not receive a transplant.102 An analysis of four-year survival rates for all patients transplanted in the United States has confirmed that overall outcomes for those transplanted with hepatocellular carcinoma are only minimally worse than for those transplanted for other indications.38 Certain subgroups of patients do worse, including

Chapter 94  Tumors and Cysts of the Liver those with nodules 3 to 5 cm in diameter, a MELD score of 20, and a serum AFP level of ≥455 ng/mL. Some authorities have advocated a modest expansion of the Milan criteria, with use of the so-called University of California, San Francisco (UCSF) criteria (a single lesion up to 6.5 cm in diameter or two or three lesions up to 4.5 cm each, with a total tumor diameter of 8 cm), based on excellent prospective outcomes from a small, single-center series, but these patients generally need a special exception from the regional review board in the United States.113 Other groups who use similar criteria have shown similarly good results.114 A larger multicenter study is needed before these criteria can be widely adopted. Local Ablation Local ablative therapies are potentially curative treatments for patients with small tumors, usually smaller than 3 to 5  cm in diameter, that are not amenable to resection or liver transplantation because of patient pre­ference, the number and location of lesions, or significant hepatic dysfunction (Child class B or C; see Fig. 94-4).25,115 The first of these techniques available was percutaneous ethanol injection (PEI), a relatively effective and safe method that is still used and is most effective for lesions smaller than 2 to 3  cm in diameter.116 PEI requires multiple sessions and, in patients with small tumors and intact hepatic function, can lead to survival rates similar to those for surgical resection, although no randomized studies have been performed to demonstrate equivalent outcomes.117 Complications are rare but include tumor seeding of the needle track. More recently, radiofrequency ablation (RFA) has supplanted PEI, because it is more effective, particularly with larger tumors (up to 3 to 5  cm), requires fewer sessions, and has similar complication rates.118 RFA can be performed percutaneously or by a laparoscopic or open surgical approach. Survival rates are similar to those for surgical resection, although recurrence rates are higher and complications are uncommon.117,119 PEI is generally favored over RFA for lesions adjacent to a major vessel or large bile ducts. A randomized study has suggested that a combination of RFA and chemoembolization for tumors larger than 3  cm in diameter produces a survival benefit when compared with either treatment alone.120 PEI and RFA have been used to stabilize tumor growth in patients awaiting liver transplantation, but their use for this purpose is controversial and probably unnecessary, unless the waiting time for transplantation is more than six months or the tumor burden is near the limits of acceptability for transplantation.102,121-123 Chemoembolization Transarterial chemoembolization (TACE) is a palliative treatment reserved for patients with relatively intact hepatic function (Child class A or B) and tumors not amenable to local ablative treatments because of size, number, or location (see Fig. 94-4).25,121 Six randomized trials and a metaanalysis have compared embolization or chemoembolization with supportive care and have shown overall improved survival with treatment.124-130 The effectiveness of TACE before liver transplantation has not been fully elucidated, but TACE can be considered if the waiting time for transplantation is more than six months or the tumor size is near the acceptable limit.131,132 Theore­tically, TACE can be used to reduce the size of the tumor to make resection or transplantation possible (downstaging) or to allow a more conservative resection, although study results are mixed as to whether this approach is effective.133,134

Chemotherapy A large number of anticancer drugs, including alkylating agents, antitumor antibiotics, antimetabolites, plant alkaloids, platinum derivatives, procarbazine, estrogen receptor modulators, and somatostatin, have been tried alone and in various combinations and by different routes of administration for the treatment of hepatocellular carcinoma, but response rates have invariably been less than 20%.126,135 Several small-molecule, targeted anticancer agents have been developed and studied for the treatment of hepatocellular carcinoma. Sorafenib, an inhibitor of Raf kinase and the tyrosine kinase activity of vascular endothelial growth factor receptors (VEGFRs) and platelet-derived growth factor receptor (PDGFR), is the first of these new agents to show modest improvement in survival compared with supportive care.136 The drug should be considered for patients with intact hepatic function (Child class A) and portal vein thrombosis, extrahepatic tumor, or failures of other therapies (see Fig. 94-4). Other targeted agents, alone and in combination with each other and with traditional chemotherapy, are being studied. Patients with advanced hepatic dysfunction (Child class C) or advanced tumor symptoms (Eastern Cooperative Oncology Group [ECOG] performance status > 2) have such a poor prognosis that only supportive care should be offered (see Fig. 94-4).25 New Alternative Techniques New local ablative techniques include cryoablation, microwave ablation, and laser ablation and are being studied in hepatocellular carcinoma, but these techniques have not been adequately compared with PEI and RFA; thus, their use should be limited to clinical trials. Use of other local regional therapies include stereotactic radiotherapy and radioembolization with 90Y microspheres, but should be limited to clinical trials or as a last resort until they are compared with chemoembolization.

Screening

Because symptomatic hepatocellular carcinoma seldom is amenable to surgical cure and responds poorly to conservative treatments, a pressing need exists to prevent the tumor or detect it at a presymptomatic stage when surgical intervention is still possible. Programs for detecting subclinical hepatocellular carcinomas are of two types: (1) screening whole populations with a high incidence of the tumor; and (2) case finding and long-term surveillance of persons at high risk for the development of hepatocellular carcinoma. Mass population screening has rarely been attempted, whereas case finding and surveillance of high-risk persons are more feasible25 and have been shown to be cost-effective in countries with a high incidence of the tumor. An AASLD practice guideline published in 2005 provides recommendations for screening (Table 94-7).25 Briefly, patients at high risk of developing hepatocellular carcinoma should be entered into a surveillance program. Surveillance for hepatocellular carcinoma should be performed using ultrasonography at 6- to 12-month intervals, but AFP alone should not be used for screening unless ultrasound is not available. Although CT and MRI are effective imaging modalities for the diagnosis of hepatocellular carcinoma, they are not recommended for routine use in surveillance but may be considered if adequate ultrasound images cannot be obtained because of body habitus. Growing evidence suggests that surveillance for hepato­cellular carcinoma in patients with cirrhosis improves outcome.137

Prevention

Although great progress has been achieved in the primary prevention of HBV-induced hepatocellular carcinoma with

1579

1580

Section IX  Liver Table 94-7  High-Risk Groups for Whom Surveillance for Hepatocellular Carcinoma Is Recommended High-Risk Group

FACTORS THAT ADD TO RISK

Hepatitis B viral carriers

Africans, age > 20 yr Asian men, age > 40 yr Asian women, age > 50 yr Family history of hepatocellular carcinoma Patients with cirrhosis Patients with high serum HBV DNA level and ongoing hepatic injury Alcoholic cirrhosis α1-Antitrypsin deficiency* Autoimmune hepatitis* Hemochromatosis Hepatitis B Hepatitis C Nonalcoholic steatohepatitis* Primary biliary cirrhosis

Patients with cirrhosis

*No data on efficacy of surveillance available. HBV DNA, hepatitis B viral deoxyribonucleic acid. Adapted from Bruix J, Sherman M. Management of hepatocellular carcinoma. Hepatology 2005; 42:1208-36.

universal infant vaccination against HBV in many countries, the full impact of universal HBV vaccination on the occurrence of the tumor will not be realized for many years. A significant reduction has already been noted in childhood hepatocellular carcinoma in Taiwan, where universal infant vaccination was adopted in the mid-1980s.138 In the meantime, the huge numbers of existing HBV carriers worldwide remain at risk for hepatocellular carcinoma, and little progress has been made in preventing malignant transformation in persons with chronic viral hepatitis, nor has much progress has been made on other fronts. A vaccine against HCV will not be available in the near future, and prevention of aflatoxin-induced tumors is far from a becoming a reality, despite ongoing trials of chemopreventive agents. Considerable interest has been expressed in the impact of antiviral therapy against HBV and HCV in reducing the incidence of hepatocellular carcinoma. One randomized controlled trial of long-term therapy of lamivudine versus placebo in patients with chronic hepatitis B has shown a significant decrease in the frequency of clinical events in the treated group, including a decrease in the frequency of hepatocellular carcinoma.77 Several large retrospective studies have shown a decrease in frequency of hepatocellular carcinoma in patients successfully treated for chronic hepatitis C with interferon-based regimens.139

INTRAHEPATIC CHOLANGIOCARCINOMA

Cholangiocarcinoma is a malignant neoplasm arising from the biliary duct epithelium. It often carries different names based on the particular portion of the biliary tree involved—small intrahepatic bile ducts (peripheral cholangiocarcinoma), hepatic duct bifurcation (perihilar cholangiocarcinoma, or Klatskin tumor), and extrahepatic bile ducts (bile duct carcinoma). The location of the tumor has a major impact on the presenting symptoms and treatment approach. Perihilar cholangiocarcinoma is classified with the intrahepatic group based on International Classification of Diseases, 9th revision (ICD-9) codes even though it is extrahepatic in origin and is the most common form.140,141 This section will be limited to a discussion of intrahepatic cholangiocarcinoma; extrahepatic cholangiocarcinoma is discussed in Chapter 69.

Epidemiology

Intrahepatic cholangiocarcinoma represents approximately 10% to 20% of all primary liver cancers and 20% to 25% of cholangiocarcinomas. The geographic variation in prevalence rates is marked, ranging from 0.2 to 96/100,000 in men and from 0.1 to 38/100,000 in women, because of differences in the frequencies of known risk factors in various populations.142 The highest prevalence rates are found in parts of Asia, most notably certain regions of Thailand, Hong Kong, China, Japan, and Korea. Chronic infestation of the biliary tree with one of the liver flukes is thought to be the cause of these high rates.143 The overall prevalence rate in the United States is 0.85/100,000, with a 1.5-fold higher rate in men than women. The rate in whites is about equal to that in African Americans and about half that in Asians. Although the underlying predisposing factor for most cases of cholangiocarcinoma is unknown, a number of risk factors have been recognized. The strongest association is with Opisthorchis viverrini, a liver fluke endemic in parts of Southeast Asia and acquired by ingestion of raw or uncooked fish.142,144,145 The association with Clonorchis sinensis, a related liver fluke, is weaker.146 An association with the radiographic contrast agent thorium dioxide (Thorotrast), which was banned in the 1950s, has been well established.147 Primary sclerosing cholangitis is linked to a diagnosis of cholangiocarcinoma at a young age, with a lifetime risk of 8% to 20% (see Chapter 68).148-150 Congenital and acquired abnormalities of the biliary tract that may result in bile stasis, chronic inflammation, and infection, as in biliary atresia,151 von Meyenburg complexes,152 Caroli’s disease,153 choledochal cyst,153 and intrahepatic cholelithiasis, have been associated with the development of cholangiocarcinoma. Cirrhosis, particularly caused by HCV, also has been associated with cholangiocarcinoma.154 Intrahepatic cholangiocarcinoma is rare before the age of 40 years, and historically the worldwide approximate average age at presentation is 50 years. Epidemiologic data indicate that the age at presentation has shifted to more than 65 years. Additionally, the incidence and mortality rates are increasing worldwide.140 Surveillance, Epidemiology and End Results (SEER) registry data from the United States have shown a 165% increase between the late 1970s and the late 1990s.142 This increase may be a result, in part, of the increased prevalence of cirrhosis, particularly HCVassociated cirrhosis.154

Molecular Pathogenesis

Malignant transformation of the bile duct cells generally occurs in an environment of inflammation or cholestasis (or both), usually as a result of one of the known risk factors. The proposal has been made that a combination of these environmental factors and genetic predisposition—for example, defects in oncogenes or bile salt transporters— leads to an accumulation of genetic defects that results in carcinoma.140,155 A polymorphism in the gene for the natural killer cell receptor G2D (NKG2D) has been associated with an increased risk of cholangiocarcinoma in patients with primary sclerosing cholangitis.156 At the molecular level, numerous changes have been described, including mutations of the K-ras gene, the gene for interleukin-6, and allelic loss or mutations of TP53 and p16, as well as many others (see Chapter 69).

Clinical Features

Peripheral cholangiocarcinoma seldom produces symptoms until the tumor is advanced. The clinical features are then similar to those of hepatocellular carcinoma, including

Chapter 94  Tumors and Cysts of the Liver malaise, weight loss, abdominal pain, and jaundice, which may be more frequent and prominent than with hepatocellular carcinoma.153,157 The clinical presentation of perihilar and extrahepatic cholangiocarcinoma is with progressive painless jaundice, acholic stools, pruritus with or without weight loss, and, rarely, cholangitis.158 Patients with primary sclerosing cholangitis may present with worsening jaundice caused by a dominant bile duct stricture, weakness, and weight loss.

Diagnosis

In patients with perihilar cholangiocarcinoma and extrahepatic cholangiocarcinoma, obstructive jaundice is evident, with elevated serum levels of bilirubin, alkaline phosphatase, gamma glutamyl transpeptidase (GGTP), and often aminotransferases. In patients with peripheral cholangiocarcinoma, often only the alkaline phosphatase level is elevated. CA 19-9 is the most frequently used serum tumor marker for cholangiocarcinoma but has significant limitations because CA 19-9 levels are also elevated in pancreatic, colorectal, gastric, and gynecologic cancers and in acute bacterial cholangitis.159 CA 19-9 is always undetectable in the 7% of the population that is Lewis blood group– negative. In patients with unexplained biliary obstruction without primary sclerosing cholangitis, the sensitivity of CA 19-9 is 53%, and the negative predictive value is 72% to 92%, for a cutoff value of 100 U/mL. In patients with primary sclerosing cholangitis, the sensitivity ranges from 38% to 89% and specificity from 50% to 98%. The addition of carcinoembryonic antigen (CEA) probably does not improve the performance of CA 19-9 in the setting of primary sclerosing cholangitis. Initial imaging with ultrasound helps identify biliary obstruction. Dynamic contrast-enhanced CT or MRI further aids in localizing the lesion and determining the possibility of resection.140,160 MRI with magnetic resonance cholangiography (MRCP) is a superior modality because of a higher sensitivity than CT for detecting lesions and localizing biliary obstruction. The tumor is hypodense on T1-weighted images and moderately intense on T2-weighted images. Endoscopic retrograde cholangiopancreatography (ERCP) or transhepatic cholangiography allows for localization of the tumor, sampling of tissue and bile, and relief of biliary obstruction if the tumor is unresectable. A perihilar tumor may demonstrate a classic appearance on ERCP, but for other biliary strictures, particularly in patients with primary sclerosing cholangitis, determining whether cholangiocarcinoma is present may be difficult. Cytology has a 30% sensitivity, which improves to 40% to 70% with the addition of brushings and biopsies. Newer cytologic techniques that assess the cells for aneuploidy and chromosomal aberrations may improve the diagnostic yield but are not widely available (see Chapter 69). Endoscopic ultrasound (EUS) with fine-needle aspiration (FNA) in patients without primary sclerosing cholangitis has the advantage of improving sensitivity and specificity for diagnosis of the primary lesion and nodal metastasis but the disadvantage of causing peritoneal seeding, and thus should be avoided if surgical resection is contemplated. Percutaneous biopsies also carry the risk of peritoneal seeding and are generally avoided if the tumor is potentially resectable.

Pathology

Peripheral cholangiocarcinoma usually is a large and solitary tumor, but it may be multinodular.161 It is grayishwhite, firm, and occasionally umbilicated and can produce a focal hepatic mass, a tumor growing along and infiltrating the bile ducts, or an intraductal papillary lesion.160 The

tumor is poorly vascularized and rarely bleeds internally or ruptures. Perihilar cholangiocarcinoma may take the form of a firm intramural tumor that encircles the bile duct, a bulky mass centered on the duct or hilar region that radiates into the hepatic tissue, or a spongy friable mass within the lumen of the duct. Metastatic nodules may be distributed irregularly throughout the liver. The bile ducts peripheral to the tumor may be dilated, resulting in some cases in biliary cirrhosis. Metastases in regional lymph nodes occur in about 50% of cases. Microscopically, cholangiocarcinoma exhibits acinar or tubular structures that resemble those of other adenocarcinomas.161 Most tumors are well differentiated. Secretion of mucus may be demonstrable, but bile production is not seen. The tumor cells provoke a variable desmoplastic reaction, and, in many tumors, the collagenized stroma may be the most prominent feature. Distinguishing the tumor from metastatic adenocarcinoma may be difficult, and some experts have advocated assuming that an adenocarcinoma in the liver is cholangiocarcinoma if no primary tumor can be found elsewhere.162

Treatment and Prognosis

Early diagnosis of intrahepatic cholangiocarcinoma is unusual, and the annual mortality rate is almost identical to the annual incidence of the tumor.153,157 Long-term survival after diagnosis in the United States based on the SEER database is dismal, with a one-year survival rate of 28% and a five-year survival rate less than 5%. The five-year survival rate has not improved since the late 1980s.142 In a person with suspected or proven intrahepatic cho­ langiocarcinoma, staging is recommended to determine surgical resectability, which is the only opportunity for cure. The staging evaluation usually includes dynamic MRI and MRCP of the abdomen (or dynamic helical CT, if MRI is unavailable) and a chest x-ray or CT.140 Positron emission tomography (PET) has been assessed in small series and does not clearly add to other modalities. EUS with FNA of suspicious lymph nodes may detect otherwise unrecognized metastasis in up to 20% of cases.160 Surgical resectability of intrahepatic cholangiocarcinoma should be determined in conjunction with an experienced hepato­ biliary surgeon and requires the ability to achieve clear surgical margins, which usually necessitates a major hepatectomy. Criteria for resection include absence of all the following: evidence of extrahepatic metastasis; main portal vein or hepatic artery invasion or encasement; bilateral segmental bile duct involvement; and contralateral hepatic lobar atrophy. Additionally, the patient must be medically fit to undergo surgery and have sufficient hepatic reserve. Without clear margins of resection, surgery provides benefits similar to those of endoscopic or biliary drainage. Patients well selected for surgical resection achieve a one- to two-year median survival and a 29% to 36% five-year survival rate. If resection is not possible and major biliary obstruction is present, biliary drainage, either endoscopic or percutaneous, should be performed, because drainage appears to improve symptoms and survival (see Chapter 70).162 Placement of an expandable metal stent is generally preferred to plastic stents if the expected survival of the patient is more that three to six months.140 The rates of response and survival following radiation therapy and chemotherapy are modest, as suggested by predominantly small uncontrolled series, so these modalities generally should be restricted to clinical trials. Photodynamic therapy in addition to biliary stent placement may provide benefit. Liver transplantation alone results in unacceptably high recurrence rates and

1581

1582

Section IX  Liver limited survival.160 In a single center report, aggressive preoperative therapy of unresectable hilar cholangiocarcinoma with external beam radiation, brachytherapy, and chemosensitization, followed by liver transplantation in patients who survived the treatment and had contained disease, produced a five-year survival rate of 82%.163

HEPATOBLASTOMA Epidemiology

In children, hepatoblastoma is the third most common malignant tumor and the most common malignant hepatic tumor. It occurs almost exclusively in the first three years of life; boys are affected twice as often as girls.164,165

Clinical Features

Most children with hepatoblastoma come to medical attention because of abdominal swelling.166 Other reasons include failure to thrive, weight loss, poor appetite, abdominal pain, irritability, and intermittent vomiting and diarrhea. The tumorous liver almost always is enlarged and firm and may be tender. Its surface is smooth or nodular. Hepatoblastomas rarely rupture. Distant metastases are evident, usually in the lung, in 20% of patients at the initial visit.167 The tumor occasionally causes isosexual precocity in boys as a result of the ectopic production of human chorionic gonadotropin.168

Diagnosis

AFP is present in high concentrations in the serum of 80% to 90% of patients with hepatoblastoma and is a useful clue to diagnosis.169 The few patients with a low serum AFP level appear to have a worse prognosis.170 Anemia is common, as is thrombocytosis, which is attributed to raised serum thrombopoietin levels. Pulmonary metastases and, rarely, mottled calcification in the tumor may be seen on plain radiography. Ultrasonography is the most widely used initial imaging technique, although the findings are not specific. CT and MRI are used to define the extent of the tumor and plan definitive surgery. The tumor is seen as an avascular mass on hepatic arteriography.171

Pathology

Hepatoblastomas are the malignant derivatives of incompletely differentiated hepatocyte precursors. Their constituents are diverse, reflecting both the multipotentiality of their mesodermal origin and the progressive stages of embryonic and fetal development. Hepatoblastomas are classified morphologically into an epithelial type, composed predominantly of epithelial cells of varying maturity, and a mixed epithelial and mesenchymal type, which contains, in addition, tissues of mesenchymal derivation.166,169 The tumors usually are solitary, ranging in diameter from 5 to 25 cm, and always well circumscribed (about half are encapsulated). They vary in color, ranging from tan to grayish-white, and contain foci of hemorrhage, necrosis, and calcification. Vascular channels may be prominent on the capsular surface. Epithelial hepatoblastomas are solid, whereas tumors of the mixed variety often are separated into lobules by white bands of collagen tissue. Two types of epithelial cells are present in the tumor.172 Cells of the first type resemble fetal hepatocytes and are arranged in irregular plates, usually two cells thick, with bile canaliculi between individual cells and sinusoids between plates. Cells of the second type are embryonal and are less differentiated than the fetal type. Mixed hepatoblastomas contain mesenchymal tissue consisting of areas of a highly cellular primitive type of mesenchyme intimately admixed with epithelial elements. Cartilage and striated

muscle may be present. Hepatoblastomas may show foci of squamous cells, with or without keratinization, and foreign body–type giant cells. Vascular invasion may be evident. Metastases most commonly involve lung, abdominal lymph nodes, and brain.

Cause and Pathogenesis

Hepatoblastoma may occur sporadically or in association with hereditary syndromes such as familial adenomatous polyposis (FAP) and Beckwith-Wiedemann syndrome, suggesting a possible role for chromosomes 5 and 11 in the genesis of the tumor. The FAP tumor suppressor gene downregulates β-catenin. Sporadic hepatoblastoma is not associated with any known environmental risk factor, and its pathogenesis is unclear. Most patients with hepatoblastoma have mutations of the FAP gene, and a similar number have activating mutations of the β-catenin gene, raising the possibility that the wnt signaling pathway plays a role in the development of the tumor.173

Treatment and Prognosis

Hepatoblastomas are rapidly progressive. If the lesion is solitary and sufficiently localized to be resectable, surgery often is curative, with five-year survival rates as high as 75%.166 The current practice is to pretreat the patient with cisplatin and doxorubicin. When the tumor is judged to be inoperable, neoadjuvant chemotherapy may reduce the size of the tumor sufficiently to permit resection. Encouraging results also have been obtained with liver transplantation in patients with bilobar multifocal tumors without extrahepatic extension.174 If surgery is not possible or the tumor recurs after surgery, the prognosis generally is poor.

HEMANGIOSARCOMA Epidemiology

Although rare, angiosarcoma is the most common malignant mesenchymal tumor of the liver.175,176 It occurs almost exclusively in adults and is most prevalent in the sixth and seventh decades of life.177,178 Men are affected four times as often as women.

Pathogenesis

Despite its rarity, hepatic angiosarcoma is of special interest because specific risk factors have been identified, although no cause is discerned in most cases. In early reports, the tumor became evident approximately 20 years after the patient had been exposed to thorium dioxide (see Chapter 87).179 Angiosarcoma also has occurred in German vintners who used arsenic-containing insecticides and drank wine adulterated with arsenic.180 A few patients with angiosarcoma had taken potassium arsenite (Fowler’s solution) for many years to treat psoriasis.181 Hepatic angiosarcoma in workers exposed to vinyl chloride monomer (VCM) was first reported in 1974.177,182,183 The monomer is converted by enzymes of the endoplasmic reticulum to reactive metabolites that form DNA adducts and guanosine-to-adenine transitions in the K-ras and TP53 genes. Angiosarcomas have occurred after exposures of 11 to 37 years (or after shorter periods with a heavy initial exposure). The mean age of patients at diagnosis is 48 years. In addition to angiosarcoma, persons exposed to VCM may be at increased risk of hepatocellular carcinoma and soft tissue sarcoma.

Clinical Features

The most common presenting symptom is upper abdominal pain. Other frequent complaints are abdominal swelling, rapidly progressing liver failure, malaise, weight loss, poor appetite, and nausea.176,177 Vomiting occurs

Chapter 94  Tumors and Cysts of the Liver occasionally. The duration of symptoms generally ranges from one week to six months, but a few patients have had symptoms for as long as two years before seeking medical attention. The liver almost always is enlarged and usually is tender. Its surface may be irregular, or a definite mass may be felt. An arterial bruit occasionally is heard over the enlarged liver. Splenomegaly may be present and is attributed to the hepatic fibrosis and consequent portal hypertension that also may complicate exposure to VCM. Ascites is frequent, and the fluid may be blood-stained. The patient often has jaundice. Fever and dependent edema are less common. Approximately 15% of patients present with acute hemoperitoneum following tumor rupture. Rarely, pulmonary or skeletal metastases are present.

Diagnosis

A rising serum bilirubin level and other evidence of pro­ gressive hepatic dysfunction may be present, especially in the later stages of the tumor. Plain radiography may show pulmonary metastases, a raised right hemidiaphragm, or, rarely, skeletal metastases. In patients who received thorium dioxide, radiopaque deposits of the material may be evident in the liver and spleen.179 One or more mass lesions may be demonstrated on ultrasonography, CT, or MRI, but diffusely infiltrating tumor may not be visualized. Hepatic arteriography reveals a characteristic appearance.184 The hepatic arteries are displaced by the tumor, which shows a blush and “puddling” during the middle of the arterial phase that persist for many seconds, except in the central area, which may be hypovascular.

Complications and Prognosis

Hepatic angiosarcomas grow rapidly, and the prognosis is poor; death ensues within six months. Patients may have thrombocytopenia resulting from entrapment of platelets within the tumor (Kasabach-Merritt syndrome), disseminated intravascular coagulation with secondary fibrinolysis,185 or microangiopathic hemolytic anemia as a result of fragmentation of erythrocytes within the tumor circulation.186

Pathology

Angiosarcomas usually are multicentric.187 Their hallmark is the presence of blood-filled cysts, although solid growth also is seen. The lesions are fairly well circumscribed but not encapsulated. Larger masses are spongy and bulge beneath Glisson’s capsule. The earliest microscopic change is the presence of hypertrophic sinusoidal lining cells with hyperchromatic nuclei in ill-defined loci throughout the liver. With progression of the lesion, sinusoidal dilatation and disruption of hepatic plates occur, and the malignant cells become supported by collagen tissue. Enlarging vascular spaces lined by malignant cells cause the tumor to become cavernous. The malignant endothelial cells usually are multilayered and may project into the cavity in intricate fronds and tufts supported by fibrous tissue. The fronds commonly are elongated, with ill-defined borders. The cytoplasm is clear and faintly eosinophilic. Nuclei are hyperchromatic and vary greatly in size and shape; some cells are multinucleated. Evidence of phagocytosis may be seen. Foci of extramedullary hematopoiesis are common, and invasion of the portal and central veins occurs in most cases. Distant metastases are present in 50% of tumors.

Treatment

Operative treatment usually is precluded by the advanced stage of the tumor. Even when surgery is undertaken, the

patient commonly survives only one to three years, although long-term survival may be achieved in the few patients with a solitary tumor.176 The results of irradiation and chemotherapy are poor.

EPITHELIOID HEMANGIOENDOTHELIOMA Epidemiology

Epithelioid hemangioendothelioma is a rare tumor whose incidence is not known. A case series of 137 cases has been collected at a specialized referral center.188 Two thirds of patients were female, and the tumor occurred at all ages in adulthood.

Clinical Features

Patients typically present with nonspecific symptoms, such as abdominal pain and weight loss.

Diagnosis

Imaging studies show a characteristically highly vascular mass, which may infiltrate throughout the liver. Case reports indicate that the tumor can be visualized on PET. Correct diagnosis requires histologic examination of tissue obtained by biopsy.

Complications and Prognosis

The tumor has low-grade malignant potential and must be distinguished from hemangiosarcoma, because it has a much better prognosis if treated appropriately and aggressively. Epithelioid hemangioendothelioma may metastasize, both within and beyond the liver.

Pathology

Tumors are often multiple and may be diffuse throughout the liver. Histologically, they are characterized by the presence of dendritic and epithelioid cells that contain vacuoles, representing intracellular lumina. These cells stain positively for endothelial markers, such as factor VIII– related antigen, CD34, or CD31.

Treatment

The primary treatment modality for epithelioid hemangioendothelioma is surgical, including resection or liver transplantation. Transplantation appears to be effective for this tumor, even in the presence of advanced or even metastatic disease. The tumor does not appear to be sensitive to radiation or chemotherapy.

OTHER PRIMARY MALIGNANT TUMORS OF THE LIVER

Undifferentiated (embryonal) sarcoma is a rare primary malignancy of the liver that occurs in both children and adults.189,190 The tumor tends to be aggressive, but long-term survival can be achieved with radical surgery and chemotherapy. Other rare sarcomas arising in the liver include liposarcoma,191 lymphoma,176,192 and rhabdomyosarcoma.190

HEPATIC METASTASES The liver is the most frequent target for metastatic spread of tumors. Hepatic metastases occur in 40% to 50% of adult patients with extrahepatic primary malignancies.193 Foremost among the reasons for the high frequency of hepatic metastases are the double blood supply of the liver and the presence of fenestrations in the sinusoidal endo­thelium that facilitate penetration of malignant cells

1583

1584

Section IX  Liver into the hepatic parenchyma.194 Hepatic metastases commonly originate from primary sites in the distribution of the portal venous system, including the pancreas, stomach, and colon. Outside this distribution, tumors of the lung and breast are the most common origins of hepatic metastases.

Clinical Features

Symptoms resulting from hepatic metastases often are absent or overshadowed by those of the primary tumor. Occasionally, the symptoms and signs attributable to metastases are the presenting manifestations of an asymptomatic primary tumor. In such cases, the likely symptoms are malaise, weight loss, and upper abdominal pain. Jaundice, when present, is seldom attributable to replacement of hepatic tissue by metastases. Depending on the extent of the metastatic disease, the liver may be enlarged, sometimes markedly. Its surface may be irregular, and umbilicated nodules may be felt by the examiner. A friction rub may be heard over hepatic metastases.

Diagnosis

CT is the most useful imaging technique.195 Multiphase helical CT and CT during arterial portography are more sensitive than conventional CT. Dynamic contrast-enhanced Doppler ultrasonography with intra-arterial infusion of CO2 microbubbles also is useful for the diagnosis of hepatic metastases.54 T1-weighted MRI also may be helpful, and iron oxide–enhanced MRI is even better.

Pathology

Macroscopic Appearance Hepatic metastases almost always are multiple.193 Their pathologic features vary, depending on the site of origin. Metastases are expansive, when they are discrete, or infiltrative. Individual metastases may reach a large size, and, with multiple metastases, the liver may be greatly enlarged. Metastases commonly are gray-white and may show scattered hemorrhages or central necrosis. Individual metas­ tases may be surrounded by a zone of venous stasis. Subcapsular lesions often are umbilicated. The dictum that cirrhotic livers are less likely than noncirrhotic livers to harbor metastatic deposits remains to be verified. Microscopic Appearance The microscopic features, including the degree of stromal growth, of most hepatic metastases duplicate those of the tumor of origin. Metastatic deposits usually are easily delineated from the surrounding liver tissue. Invasion of portal or hepatic veins may be seen, although less often than with hepatocellular carcinoma.193 It may be difficult to distinguish metastatic adenocarcinoma from primary cholangiocarcinoma.162

Treatment and Prognosis

The extent of replacement of liver tissue by metastases generally determines the patient’s prognosis. The greater the tumor burden, the worse the outlook, with only approximately 50% of patients surviving three months after the onset of symptoms and less than 10% surviving more than one year.196 Improved imaging modalities, advances in sur­ gical techniques for resection and transplantation, and new chemotherapeutic agents and regional therapies have made it possible to achieve long-term survival in individual patients. Long-term survival has been accomplished most often by resection of hepatic metastases in patients with colorectal cancer, a substantial number of whom have been cured or have obtained up to 20 years of disease-free sur-

vival.196-198 Survival for five years can be achieved in up to 60% who undergo resection of a solitary colon cancer metastasis to the liver.199 If the primary tumor has been removed completely and metastases are confined to the liver, resection of hepatic metastases should be considered. Liver transplantation, with or without chemotherapy, has been undertaken in a few patients but is generally not considered. RFA is a valid therapy for colorectal metastases in patients who are unable to tolerate or refuse surgical resection. Other invasive methods of destroying metastases, such as ethanol injection, freezing with cryoprobes, and laser vaporization, warrant further study. Radiation therapy and intra-arterial infusion of cytotoxic drugs have limited roles.

BENIGN TUMORS HEPATOCELLULAR ADENOMA Epidemiology and Pathogenesis Hepatocellular adenomas (or hepatic adenomas) were extremely rare before the use of oral contraceptive steroids became widespread. They are still rare in men, and the development of this tumor in women who are taking or have taken contraceptive steroids strongly implies a cause and effect relationship.200,201 Nevertheless, in light of the large number of women who use this form of contraception, the risk of hepatocellular adenoma is small, and its occurrence implies some form of genetic predisposition (see later). The association is particularly strong with prolonged use of oral contraceptive steroids; the estimated risk for women who use oral contraceptive agents continuously for five to seven years is five times the normal rate, and this risk increases to 25 times with use for longer than nine years. Using preparations with high hormone potency further increases the risk. Contraceptive pill–associated hepatocellular adenomas are more likely to develop in older than younger women. Hepatocellular adenomas have been linked to both types of synthetic estrogen and all forms of progestogen contained in oral contraceptive preparations.202 Current evidence favors estrogens as the culprit, although progestogens may contribute through their enzyme-inducing properties. The growth of hepatocellular adenomas appears to be hormone-dependent, as evidenced by an increase in size during pregnancy and occasional cases of regression (and even disappearance) of the tumor after cessation of oral contraceptive use. Although hormonal replacement therapy has not proved to be associated with the development of hepatic adenoma, avoidance of this form of therapy in women with a history of oral contraceptive–related hepatic adenoma is prudent, unless compelling reasons to the contrary exist. Hepatocellular adenomas and adenomatosis also may occur in those receiving long-term anabolic androgenic steroids63 and in those with certain inherited metabolic disturbances, especially type 1 glycogen storage disease, in which one or more hepatocellular adenomas occur in approximately 60% of patients (see Chapter 76).201 Genetic alterations have been identified in hepatocellular adenomas. Bilallelic mutations of the TCF1 gene that codes for hepatocyte nuclear factor 1α (HNF-1α) have been identified in up to 60% of patients with adenoma.203 A second pathway, the wnt pathway, which also is activated in about 25% of patients with hepatocellular carcinoma, is activated in at least some cases of adenoma.204 β-Catenin activation via this pathway appears to confer a higher risk of malignant transformation.205 The third identified pathway for formation of hepatocellular adenomas includes acute inflamma-

Chapter 94  Tumors and Cysts of the Liver Genetic Altered molecular mutations and pathway and risk factors its frequency TCF1 mut. MODY3

Obesity Alcohol

b-catenin mut. Glycogenolysis

Oral contraceptive

CYP1B1 mut.

Main clinical and pathologic characteristics

HNF1α inactivation 35%-45%

Familial adenomatosis (MODY3) Marked steatosis

β-catenin activation* 15%-19%

Males Cytologic abnormalities Pseudoglandular formation Higher risk of HCC

Acute phase inflammation* 30%-35%

Inflammatory infiltrates Dystrophic vessels Telangiectasia

Unidentified 10%-20%

A

B

Figure 94-5.  Schematic representation of the different molecular pathways altered in hepatocellular adenoma. Left, Main risk factors and known genetic predispositions. Center, Altered molecular pathways and their frequencies. Right, Principal clinical and pathologic features of the subtypes of adenomas. Arrows indicate the significant relationships. *Some tumors may be simultaneously inflammatory and β-catenin activated. CYP1B1, cytochrome P450 1B1; HCC, hepatocellular carcinoma; HNF1α, hepatocyte nuclear factor 1α (gene symbol TCF1); MODY3, maturity-onset diabetes of the young type 3; mut, mutation. (Adapted from Rebouissou S, Bioulac-Sage P, Zucman-Rossi J. Molecular pathogenesis of focal nodular hyperplasia and hepatocellular adenoma. J Hepatol 2008; 48:163-70.)

tory responses demonstrable by histologic examination of the tumor205 and associated with obesity and alcohol (Fig. 94-5).

Clinical Features

Hepatocellular adenomas manifest in a number of ways. They may produce no symptoms and be discovered during routine physical examination, if large, or during imaging of the upper abdomen for other reasons, if small. Approximately 25% of patients experience pain in the right hypochondrium or epigastrium. The pain usually is mild and ill-defined but may be severe as a result of bleeding into or infarction of the tumor. If the liver is enlarged, the surface usually is smooth, and the liver may be slightly tender. The most alarming presentation is with an acute hemoperitoneum following rupture of the adenoma. This complication is not uncommon, especially with tumors linked to oral contraceptive use, and carries an appreciable mortality rate.206,207 Tumors that rupture generally are large and solitary, although the most important determinant of rupture is a superficial location. Often, the affected woman is menstruating at the time; rupture also may occur during pregnancy.208

Diagnosis

Serum AFP concentrations are normal. Ultrasonography is used for initial imaging. Multiphase helical CT or MRI may also be used.209,210 The tumor has a clearly defined margin and, often, has almost parallel vessels entering it from the periphery (spoke wheel appearance). Alternatively, the lesion may contain tortuous vessels coursing irregularly through it. The adenoma may have focal avascular areas as a result of hemorrhage or necrosis. Because adenomas mimic normal liver tissue microscopically, needle biopsy and FNA may be of limited diagnostic value because the

Figure 94-6.  A, Surgical specimen of a large hepatocellular adenoma. The tumor is yellowish and slightly lobular, with a pseudocapsule and areas of necrosis and hemorrhage. B, Photomicrograph of a hepatocellular adenoma showing the resemblance to normal liver tissue, with cords of normal-looking, although generally slightly larger, hepatocytes, as well as Kupffer cells (but fewer in number than normal) lining the sinusoids. Bile ducts and central veins are not seen, but the presence of abnormal vascular structures is evident. (Hematoxylin and eosin.) (A courtesy of Elizabeth Brunt, MD, St. Louis, MO; B courtesy of Professor A.C. Paterson, Johannesburg, South Africa.)

cytologic appearance of the cells is similar to those of normal hepatocytes, although the tumor does not have normal portal structures.

Pathology

Hepatocellular adenoma generally occurs as a solitary, relatively soft, light brown to yellow tumor. It is sharply circumscribed but does not have a true capsule, although a pseudocapsule is formed by compression of the surrounding liver tissue (Fig. 94-6A).211 Hepatocellular adenomas arise in an otherwise normal liver. Occasionally, two or more tumors are present. Adenomas range in diameter from 1 to 30 cm and are commonly 8 to 15 cm in diameter. They are larger on average in women taking contraceptive steroids than in those not taking contraceptive steroids; they usually occupy a subcapsular position and project slightly from the surface of the liver. A pedunculated variety is seen occasionally. The cut surface of the tumor may show illdefined lobulation but is never nodular or fibrotic. Foci of

1585

1586

Section IX  Liver hemorrhage or necrosis are frequent, and bile staining may be evident. Microscopically, hepatocellular adenoma may mimic normal liver tissue to an astonishing degree (see Fig. 94-6B).211 The tumor is composed of sheets or cords of normal-looking or slightly atypical hepatocytes that show no features of malignancy. Few or no portal tracts or central veins are present, and bile ducts are conspicuously absent. Only an infrequent fibrous or vascular septum traverses the lesion. An essentially normal reticulin pattern is demonstrable throughout the adenoma. The walls of arteries and veins are thickened. Some areas with vascular abnormalities are infarcted, and thrombi may be seen. Peliosis hepatis may be found in relation to the tumor.

Treatment and Prognosis

Because of the danger that a hepatocellular adenoma may rupture, surgical treatment is recommended.211,212 Resection usually is feasible in an uncomplicated case. When rupture has occurred, emergency resection should be performed if possible. If resection cannot be accomplished, the hepatic artery should be ligated. Arterial embolization has also been used successfully to control hemorrhage from a ruptured adenoma, either in preparation for surgery or when surgery is not possible.206 Whether or not the tumor is removed, the patient must refrain from taking oral contraceptive steroids. If the adenoma is not resected, pregnancy should be avoided. Hepatocellular carcinoma occurs in a small number of women taking oral contraceptive steroids, and speculation has been raised that hepatocellular adenomas might undergo malignant transformation. Indeed, this sequence has been documented in a few instances. Therefore, managing hepatocellular adenomas merely by discontinuing the use of contraceptive steroids carries the risk that malignant transformation may still occur. The management of hepatic adenomatosis is problematic.213 Often, in these cases, the number of tumors is large, and they cannot be resected entirely. The role of liver transplantation for adenomatosis is not clear at present.

CAVERNOUS HEMANGIOMA Epidemiology

Cavernous hemangioma is the most common benign tumor of the liver and is found in as many as 7% of autopsies.63,211 The lesion is thought to be a congenital malformation or hamartoma that increases in size, initially with growth of the liver and thereafter by ectasia. Cavernous hemangiomas affect persons of all ages, although they manifest most often in the third, fourth, and fifth decades of life. Women are predominantly affected (4 : 1 to 6 : 1) and often present at a younger age and with larger tumors in comparison with men. Cavernous hemangiomas may increase in size with pregnancy or the administration of estrogens and are more common in multiparous than in nulliparous women.

Occasionally, an arterial bruit is heard over the tumor. Arteriovenous shunting has been described with cavernous hemangiomas.

Diagnosis

The ultrasonographic appearance is variable and nonspecific, although the lesion usually is echogenic.215,216 Provided that the cavernous hemangioma is larger than 3 cm in diameter, single photon emission computed tomography (SPECT) with colloid 99mTc-labeled red blood cells shows the tumor to be highly vascular and has a sensitivity and accuracy similar to those of MRI.217 Almost all cavernous hemangiomas can be diagnosed by bolus-enhanced CT with sequential scans.218 The center of the lesion remains hypodense, whereas the peripheral zone, which varies in thickness and may have a corrugated inner margin, is enhanced. MRI has a high degree of specificity and a central role in the diagnosis of small hemangiomas (Fig. 94-7).219,220 With small hemangiomas, the contrast material may assume a ring-shaped or C-shaped configuration, with an avascular center resulting from fibrous obliteration; this appearance is pathognomonic. Thrombocytopenia resulting from sequestration and destruction of platelets in large hemangiomas (KasabachMerritt syndrome) is seen occasionally in infants but rarely in adults.214,221 Malignant transformation has not been reported.

A

Clinical Features

The great majority of cavernous hemangiomas are small and asymptomatic and are discovered incidentally during imaging of the liver for another reason, at autopsy, or at laparotomy. Larger or multiple lesions produce symptoms.214 Those larger than 4 cm in diameter are called giant cavernous hemangiomas, which may be as large as 27 cm. Upper abdominal pain is the most common complaint and results from partial infarction of the lesion or pressure on adjacent tissues. Early satiety, nausea, and vomiting also may occur. Cavernous hemangiomas occasionally rupture. The only physical finding may be an enlarged liver.

B Figure 94-7.  Magnetic resonance imaging of a small cavernous hemangioma in the liver (arrow). A, T1-weighted image showing a rounded mass with a uniform increase in T1 signal intensity (low signal). B, Heavily T2-weighted image showing a mass with a uniform increase in signal intensity (bright signal relative to the water signal of cerebrospinal fluid). (Courtesy of Dr. P. Sneider, Johannesburg, South Africa.)

Chapter 94  Tumors and Cysts of the Liver Because of the risk of severe bleeding, percutaneous needle biopsy should not be performed if a cavernous hemangioma is suspected. Moreover, a needle biopsy is of limited diagnostic value.

Pathology

Cavernous hemangiomas usually are solitary lesions, although multiple tumors occur in 10% of patients.63,211 Reddish-purple or bluish masses are seen under Glisson’s capsule or deep in the substance of the liver. The larger lesions may be pedunculated. Cavernous hemangiomas are well circumscribed but seldom encapsulated. They may show central necrosis and, in some cases, the whole tumor is firm in consistency and grayish-white in appearance. Microscopically, hemangiomas are composed of multiple vascular channels of varying sizes lined by a single layer of flat epithelium and supported by fibrous septa. The vascular spaces may contain thrombi. The demonstration of mast cells within hemangiomas suggests that they may have a role in pathogenesis.222 Sclerosing cavernous hemangiomas may sometimes be seen and probably represent natural involution of these lesions. Occasionally, cavernous hemangiomas are associated with hemangiomas in other organs. They also may coexist with cysts in the liver or pancreas,223 von Meyenburg complexes (see later),224 or focal nodular hyperplasia (see later).225

Treatment

The great majority of cavernous hemangiomas can safely be left untreated. A cavernous hemangioma that is large but localized, and the cause of incapacitating symptoms, should be resected.214 If resection is not feasible, reduction in the size of the tumor with relief of symptoms is rarely achieved with irradiation, arterial ligation, arteriographic embolization, or systemic glucocorticoids.226,227 RFA has been used with some success. If a cavernous hemangioma has ruptured, it may be necessary to embolize or clamp the hepatic artery to stop bleeding before proceeding with resection, although this complication is exceedingly rare.

INFANTILE HEMANGIOENDOTHELIOMA Epidemiology and Clinical Features

Although rare, infantile hemangioendothelioma is the most common tumor of the liver in infants. Its importance stems from the high incidence of congestive heart failure in infants with this tumor and the resulting high mortality rate. The tumor almost invariably manifests in the first six months of life and is twice as common in girls as in boys.164,228 Hepatic hemangioendothelioma often coexists with hemangiomas in other organs, especially the skin (in approximately 50% of patients). Small hemangioendotheliomas are usually asymptomatic. The presence of a large lesion is recognized clinically by the diagnostic triad of an enlarged liver, high-output cardiac failure, and multiple cutaneous hemangiomas.201,228 The liver is larger than expected on the basis of the severity of the cardiac failure, and hepatomegaly persists after the heart failure has been treated successfully. When hemangioendotheliomas occur diffusely throughout the liver, as they usually do, their combined effect is to act as a large peripheral arteriovenous shunt. Shunts of this size are responsible for the cardiac failure. Approximately one third of patients have jaundice. Patients may be anemic, partly because of the dilutional effect of the increased circulating plasma volume that develops with large peripheral arteriovenous fistulas. A microangiopathic hemolytic anemia may con­tribute. In addition, thrombocytopenia may be present (KasabachMerritt syndrome). Malignant change is a rare complication.

Diagnosis

Ultrasonography may show one or more echogenic masses in the liver. Hepatic angiography is particularly helpful in diagnosis and shows stretching, but not displacement, of the intrahepatic arteries.229 Abnormal vessels arise from the hepatic arteries and promptly opacify the liver, thereby giving rise to the characteristic blush of an arteriovenous shunt. The circulation time through the liver is short. Focal avascular areas may be evident when hemorrhage into or necrosis of the tumor has occurred. CT with enhancement and MRI are as specific as hepatic arterio­graphy for the diagnosis of hemangioendotheliomas.230 Percutaneous biopsy is contraindicated because of the danger of bleeding.

Pathology

Two types of infantile hemangioendothelioma are recognized. Type I lesions are often calcified and have a fibrous stromal separation (with bile ductules) between channels. Type II lesions have a more malignant and disorganizedappearing endothelial cell lining and no stromal bile ductules.164,231 Infantile hemangioendotheliomas typically are multifocal and produce a nodular deformity of the entire liver. The nodules range in size from a few millimeters to many centimeters and are well demarcated but not encapsulated. At laparotomy, the nodules can be seen to pulsate. They are reddish purple, although large tumors are gray to tan. They may show hemorrhages, fibrosis, or calcification. Microscopically, infantile hemangioendothelioma is composed of layers of plump endothelial cells. A single layer characterizes the type I pattern, whereas several layers characterize the type II pattern. In some areas of the tumor, solid masses of mesoblastic primordial cells that differentiate early into vascular structures are observed. Fibrous septa may be prominent, and extramedullary hematopoiesis occurs frequently. Thrombosis may be followed by scarring and calcification.

Treatment and Prognosis

The course of infantile hemangioendothelioma is characterized by tumor growth during the early months of life, followed by gradual involution.228 If the child survives, the tumor involutes completely. Life-threatening aspects of the disorder are intractable congestive heart failure and, to a lesser extent, consumptive coagulopathy or rupture of the tumor. Cardiac failure should be treated by conventional means initially, but if these measures fail, more aggressive treatment of the tumor, such as embolization, ligation of the hepatic artery, surgical resection, or liver transplantation, should be considered.232,233 Use of glucocorticoids has been successful in many (but not all) patients,234 whereas irradiation has seldom been beneficial. When the tumor is confined to one lobe, surgical resection is curative, even in the presence of cardiac failure.228

OTHER BENIGN TUMORS OF THE LIVER

Other rare benign tumors of the liver include angiomyolipoma,235 bile duct adenoma, biliary cystadenoma, and biliary adenofibroma.236,237

TUMOR-LIKE HEPATIC LESIONS FOCAL NODULAR HYPERPLASIA

Focal nodular hyperplasia is a circumscribed, usually solitary lesion composed of nodules of benign hyperplastic hepatocytes surrounding a central stellate scar.238

1587

1588

Section IX  Liver Epidemiology and Pathogenesis

Focal nodular hyperplasia is more common than hepatocellular adenoma. The lesion is seen more often in women than in men, although the gender difference is much less striking than that for hepatocellular adenoma. Focal nodular hyperplasia occurs at all ages, but most patients present in the third and fourth decades of life201; the age distribution is similar to that of hepatocellular adenomas and the two lesions may coexist. The cause of focal nodular hyperplasia is unknown. Abnormalities in arteries in small and medium-sized portal tracts have been described, suggesting a role of vascular malformation in its pathogenesis.239,205 It has been described to occur with other vascular lesions, such as cavernous hemangioma, epithelioid hemangioendothelioma, and hereditary hemorrhagic telangiectasia.240,241 A role for oral contraceptive steroids in the development of the lesion was suggested but has been disputed.201 Nevertheless, some evidence suggests that focal nodular hyperplasia may be hormone-dependent.242,243 Contraceptive steroids may accentuate the vascular abnormalities in focal nodular hyperplasia and cause the lesion to enlarge, become more symptomatic, and, rarely, rupture.

Clinical Features

Most of these lesions do not produce symptoms and are often discovered during upper abdominal imaging for other reasons or because an enlarged liver is felt on routine examination or found during abdominal surgery or at autopsy.201,244,245 Patients may experience mild pain, particularly with bleeding into or necrosis of the lesion.

Diagnosis

Serum AFP levels are normal. The mass lesion seen on ultrasonography and CT is not specific for focal nodular hyperplasia,246,247 unless the central scar and feeding artery are seen (Fig. 94-8). MRI may be useful for the diagnosis of focal nodular hyperplasia.248

Pathology

Focal nodular hyperplasia manifests as a firm, coarsely nodular, light brown or yellowish-gray mass of variable size with a dense, central stellate scar and radiating fibrous septa that divide the lesion into lobules.249 The nodule may be small, resembling a cirrhotic nodule, or extremely large.

Figure 94-8.  Contrast-enhanced computed tomography scan of the liver during arterial phase showing a typical focal nodular hyperplasia (arrow) with enhancement of the mass lesion and the central stellate scar apparent by its lack of enhancement.

The lesion of focal nodular hyperplasia usually occupies a subcapsular position and may be pedunculated. It generally is solitary. Larger lesions may show foci of hemorrhage or necrosis, although these features are seen less frequently than in hepatocellular adenomas. The fibrous septa sometimes are poorly developed, and the central scar may be absent. The lesion is sharply demarcated from the surrounding liver tissue, which is normal, but a true capsule is absent. Focal nodular hyperplasia is associated with hepatic hemangiomas in as many as 20% of cases. Microscopically, focal nodular hyperplasia closely resembles a focal form of inactive cirrhosis. Individual hepatocytes are indistinguishable from those of normal liver but lack the normal cord arrangement in relation to sinusoids, central veins, and portal tracts. Kupffer cells are present. Characteristically, the fibrous septa contain numerous bile ductules and vessels. Other features include heavy infiltrations of lymphocytes and, to a lesser extent, plasma cells and histiocytes. Bile duct proliferation in portal tracts also may be evident. Branches of the hepatic artery and portal vein show various combinations of intimal and smooth muscle hyperplasia, subintimal fibrosis, thickening of the wall, occlusive luminal lesions, and, at times, occluding thrombosis. Whether these vascular changes are primary or secondary is not known. Peliosis hepatis may be an associated lesion. The histologic features almost always make it possible to distinguish focal nodular hyperplasia from hepatocellular adenoma, although the distinction may be extremely difficult to make, particularly in small biopsy specimens.

Treatment

Large symptomatic or complicated lesions should be resected, usually by segmental resection or enucleation. Recurrence after resection is rare. These lesions may also be treated with RFA. Otherwise, focal nodular hyperplasia should be left alone. If the lesion is not resected, discontinuation of contraceptive steroids is recommended and may result in regression of the lesion. Periodic ultrasonography should be performed if a firm diagnosis of focal nodular hyperplasia has not been made, and a lesion seen to increase substantially in size should be resected. The available evidence argues against the notion that focal nodular hyperplasia is a premalignant condition.

OTHER NODULAR DISORDERS

Nodular regenerative hyperplasia is characterized by nodularity of the liver without fibrosis (see Chapter 35).250 This rare condition may be associated with a number of diseases, such as rheumatoid arthritis and Felty’s syndrome. Although generally diffuse, the nodularity occasionally is focal, in which case the lesion may be mistaken for a tumor. Patients with nodular regenerative hyperplasia typically present clinically with portal hypertension. Partial nodular transformation is characterized by nodules that are limited to the perihilar region of the liver. These patients also present with portal hypertension. Macroregenerative nodules may occur in advanced cirrhosis or after massive hepatic necrosis. In the presence of cirrhosis, they are believed to be premalignant and may, in addition, be mistaken for hepatic tumors during hepatic imaging.60 Inflammatory pseudotumor is a rare entity, resulting from focal infection, that may be mistaken for a hepatic tumor (see Chapter 82).251 It occurs particularly in young men, who present with intermittent fever, abdominal pain, jaundice, vomiting, and diarrhea. Leukocytosis, an elevated erythrocyte sedimentation rate, and polyclonal hyperglobulinemia are present in approximately 50% of patients. The lesion

Chapter 94  Tumors and Cysts of the Liver may be solitary or multiple and shows a mixture of chronic inflammatory cells, with plasma cells predominating. Focal fatty infiltration, or focal fatty sparing in the presence of diffuse fatty infiltration, may also be mistaken for a hepatic tumor (see Chapter 85).252

HEPATIC CYSTS Hepatic cysts are abnormal fluid-filled spaces in the hepatic parenchyma and biliary tree. They are categorized into three main types—fibrocystic diseases of the liver, cystadenomas and cystadenocarcinomas, and hydatid cysts. Cystadenomas and cystadenocarcinomas are discussed in Chapter 69. Hydatid cysts are discussed in Chapter 82.

FIBROCYSTIC DISEASES OF THE LIVER

Fibrocystic diseases of the liver originate from abnormal persistence or defects in the progressive remodeling of the ductal plate during development, resulting in dilated fluid-filled spaces, including hepatic and choledochal cysts, portal fibrosis, and ductal plate malformations (see Chapter 62).253,254 Fibrocystic disorders of the liver described here include simple hepatic cysts, polycystic liver disease (PCLD), von Meyenburg complexes, and Caroli’s disease (type V choledochal cyst). (The other diseases are congenital hepatic fibrosis and type IV choledochal cysts; see Chapter 62.)

Simple Hepatic Cysts

Simple hepatic cysts are thought to be congenital in origin and have a frequency of about 2.5% of the population.255 They generally are smaller than 5 cm in diameter and can number up to three before being considered part of PCLD.256 The cysts usually are asymptomatic and discovered incidentally during upper abdominal imaging. They occur more often in women than in men, and their prevalence increases with age. When symptomatic, they can produce complications similar to those of PCLD, including intracystic bleeding, infection, rupture, or compression of adjacent organs. Typically, initial imaging with ultrasound, CT, or MRI provides an accurate diagnosis and distinguishes a simple cyst from a hydatid cyst and cystadenoma. Septations, papillary projects, or calcification should raise the suspicious for an alternative diagnosis.257 Asymptomatic solitary hepatic cysts should be left alone. If intervention is required because of symptoms, percutaneous aspiration and sclerosis with alcohol or doxycycline will almost always ablate the cyst, but recurrence is frequent.256 An alternative approach is laparoscopic (or, rarely, open surgical) fenestration, which is seldom followed by recurrence but has greater morbidity.

Polycystic Liver Disease

PCLD is a rare condition in which multiple cysts form in the hepatic parenchyma; usually it comes to clinical attention in adulthood (Fig. 94-9). PCLD usually presents in association with autosomal dominant polycystic kidney disease (ADPKD)258,259 but can appear as isolated polycystic liver disease.260,261 The cysts range in diameter from a few millimeters to 10 cm or more. They contain clear, colorless, or strawcolored fluid and are lined by a single layer of cuboidal or columnar epithelium, resembling that of bile ducts.258-262 Rarely, the cysts may be lined by squamous epithelium; these cysts may be complicated by the development of squamous cell carcinoma. In addition to the nature of the lining

Figure 94-9.  Magnetic resonance imaging of the abdomen in a patient with severe polycystic liver disease. This coronal T2-weighted image shows a massively enlarged liver with numerous bright fluid-filled cysts. (Courtesy of Dr. N. Cem Balci, St. Louis, Mo.)

epithelium, evidence for a biliary origin of these cysts is suggested by the composition of the cystic fluid, which has a low glucose content and contains secretory immunoglobulin A and gamma glutamyl transpeptidase. The cysts are thought to arise as a result of ductal plate malformation. This process gives rise to von Meyenburg complexes (see later), which become disconnected from the biliary tree during development and growth and dilate progressively to form cysts. PCLD is fairly common in patients with ADPKD. It occurs in approximately 24% of patients in the third decade of life to 80% in the sixth decade of life, but the kidney disease usually dominates the clinical course.263 Cysts also may be present in the pancreas, spleen and, less often, other organs. Symptomatic liver disease correlates with advancing age, severity of renal cysts, and renal dysfunction.264 Women tend to have larger and more numerous cysts, and a correlation with the number of pregnancies has been found. The use of exogenous female sex hormones may accelerate the rate of growth and size of the cysts. PCLD may coexist with other fibrocystic liver diseases, such as congenital hepatic fibrosis (in which the patient is likely to present with portal hypertension), Caroli’s disease, or von Meyenburg complexes.258-262 PCLD also is associated with other conditions such as berry aneurysms, mitral valve prolapse, diverticular disease, and inguinal hernias. Isolated PCLD not associated with ADPKD is rare, representing 7% of all PCLD in autopsy series.265 It usually is asymptomatic.266 Like ADPKD-associated PCLD, isolated PCLD is associated with pregnancy and appears to be more symptomatic in women than men. ADPKD is a common genetic disease with a frequency of 1:1,000 in whites.267 Two genes are responsible. The gene affected in ADPKD1 is PKD-1, which is located on chromosome 16q13-q23 and expresses a ubiquitous protein, polycystin-1.268,269 The gene responsible for ADPKD2 is PKD-2, which is located on chromosome 4 and expresses polycystin-2. The two polycystins are transmembrane

1589

1590

Section IX  Liver glycoproteins that complex and localize in the primary cilium, a microtubule-based structure found on renal and biliary tubule epithelium and thought to act as a flow sensor and regulator of Ca2+ influx.270 Although the mutation is inherited as an autosomal dominant trait, a second somatic mutation is thought to be necessary to produce the monoclonally derived cysts.271 Isolated PCLD has been shown in North American and Finnish families to be linked to the gene PRKCSH (also known as protein kinase C substrate 80K-H) on chromosome 19 locus p13.2-13.1 and to SEC63 on chromosome 6q21, although other associated genes undoubtedly exist.268,272,273 The gene products hepatocystin and SEC63p are thought to be involved in the folding and quality control of glycoproteins and protein translocation in the endoplasmic reticulum, respectively.274 The genes appear to be autosomal dominant, and a second somatic mutation is thought to be needed to cause disease. The hepatic cysts in polycystic liver disease, whether or not they occur in association with renal cysts, rarely cause morbidity, and many affected persons are asymptomatic.258-262 The livers of these patients contain only a few cysts or cysts smaller than 2  cm in diameter. Symptoms occur in patients with more numerous and larger cysts (10% to 15% of patients, usually women), generally with markedly enlarged livers. Abdominal discomfort or pain, postprandial fullness, awareness of an upper abdominal mass, a protuberant abdomen, and shortness of breath may be present. Severe pain may be experienced with rupture or infection of a cyst, bleeding into a cyst, or torsion of a pedunculated cyst. Jaundice is evident in approximately 5% of patients and is caused by compression of the major intrahepatic or extrahepatic bile ducts. Ascites, if present, is the result of portal hypertension, which generally is caused by associated congenital hepatic fibrosis but occasionally by compression of the hepatic veins by the cysts. Gastroesophageal variceal bleeding has rarely been reported.275 Liver biochemical test results generally are normal, with intact hepatic function, although serum alkaline phosphatase and GGTP levels may be increased. A raised right hemidiaphragm may be evident on a plain x-ray of the chest in patients with severe PCLD. The diagnosis of PCLD is confirmed by ultrasound, CT, or MRI (see Fig. 94-9). On the rare occasions when cysts require treatment, fenestration (unroofing) should be performed.258,276 Cyst fenestration originally was done at laparotomy but is now performed laparoscopically. A high recurrence rate is observed for cysts treated in this way. Cysts also have been treated by percutaneous injection of a sclerosing substance such as alcohol or doxycycline, but most patients have too many small cysts to warrant this approach. Patients who fail to respond to cyst fenestration may be considered for partial hepatic resection. Liver transplantation (sometimes combined with renal transplantation) is generally reserved for patients with hepatic failure or severe symptoms that interfere with activities of daily living.277

Fibrocystic Disease Associated with Autosomal Recessive Polycystic Kidney Disease

Fibrocystic liver disease may manifest in childhood as an autosomal recessive disorder that is usually rapidly fatal as a consequence of associated autosomal recessive polycystic kidney disease (ARPKD).258,259 A proportion of patients maintain renal function into adulthood, however, and complications of the associated liver disease then predominate. The liver cysts are microscopic, rather than macroscopic,

and present a clinical picture of congenital hepatic fibrosis. Complications of portal hypertension are the usual hepatic manifestations of the disease. The gene responsible for this disease, PKHD1, has been identified at chromosomal locus 6p21-cen, and the ARPKD protein, fibrocystin, is predicted to be an integral receptor-like protein.278 Many different mutations throughout the gene have been identified in patients with ARPKD.

Von Meyenburg Complexes

Von Meyenburg complexes (also known as biliary microhamartomas) are common and do not produce symptoms; they are small and usually multiple. Each complex is composed of cystically dilated intra- and interlobular bile ducts embedded in a fibrous stroma.175,279 The cysts are lined by cuboidal or flat epithelium. They occur in almost all patients with congenital hepatic fibrosis and may coexist with Caroli’s disease or ADPKD. Von Meyenburg complexes are found in or adjacent to portal tracts and are believed to arise as a result of malformation of the ductal plate (see Chapter 62), and they may be complicated by the development of peripheral cholangiocarcinoma.280

Caroli’s Disease

Caroli’s disease is a rare disorder characterized by congenital nonobstructive gross dilatation of the segmental intrahepatic bile ducts.256 The disease has been included in the classification of choledochal cysts (as type V)256,258 and may occur in association with medullary sponge kidney (in 60% to 80% of patients) or congenital hepatic fibrosis (see Chapter 62). Caroli’s disease is believed to be caused by an intrauterine event that arrests ductal plate remodeling at the level of the larger intrahepatic bile ducts.253 The resulting bile duct ectasia may be diffuse or localized. Both autosomal recessive and autosomal dominant modes of inheritance have been proposed. Caroli’s disease affects men and women equally and usually becomes symptomatic in early adulthood; more than 80% of patients present with symptoms before the age of 30 years. Patients typically present with recurrent episodes of fever and abdominal pain caused by cholangitis. The liver often is enlarged. Ductal ectasia predisposes to bile stagnation, which in turn may lead to cholangitis, abscess formation, and septicemia.281 Gallstones form in the ectatic ducts in one third of patients. The result of these complications may be cholangiocarcinoma, which develops in less than 10% of patients. Caroli’s disease usually is discovered when the liver is imaged during investigation of suspected cholangitis. Irregular dilatations of the larger intrahepatic bile ducts are seen. Attacks of cholangitis require treatment with antibiotics. Endoscopic retrograde cannulation of the biliary system may be used to facilitate removal of sludge or stones from the accessible part of the biliary system, and the cysts may be drained by an endoscopic or percutaneous route. Liver resection for unilobar Caroli’s disease and liver transplantation for diffuse Caroli’s disease are associated with excellent long-term patient survival and a low rate of complications.282

APPROACH TO THE PATIENT WITH A HEPATIC MASS LESION The approach to the diagnosis of a mass lesion in the liver will be influenced by the age and gender of the patient and the presence or absence of symptoms (Fig. 94-10). Making

Chapter 94  Tumors and Cysts of the Liver A

Ultrasound

Possible abscess (see Chapter 82)

Solid

Cystic

Simple benign cyst (single) Polycystic disease (multiple) Echinococcal cyst (daughter cysts) Biliary cystadenoma

Suspicious for hemangioma

Not hemangioma

MRI or

Characterize with further imaging

99mTc-labeled

Evaluate further and treat if symptomatic or if echinococcosis or malignancy is suspected

red blood cell study

Consider needle biopsy

B

Ultrasound, CT, or MRI Not suspicious for HCC

Suspicious for HCC

<1 cm

1-2 cm

>2 cm

Repeat imaging in 3-4 mo

Two dynamic imaging studies

Dynamic imaging Serum AFP

Typical of HCC

Not typical of HCC

Biopsy

Typical of HCC or Serum AFP >200 ng/mL

Cyst Hemangioma Metastases

Not typical of HCC

Biopsy

a definitive diagnosis of a mass lesion in the liver solely on clinical grounds is seldom possible. Nevertheless, detailed history taking will provide important clues about the prob­ able benign or malignant nature of the lesion. The approach to a mass in the liver differs, depending on whether or not cirrhosis is present. In a noncirrhotic liver, masses or tumors may be found in the liver incidentally or because of symptoms; the main concern is cancer metastatic from elsewhere. Initial imaging studies such as ultrasonography, CT, or MRI will indicate if the lesion is cystic. Cystic lesions should be investigated further and treated only if echinococcal cysts are suspected (see Chapter 82). Possible solid lesions in a noncirrhotic liver include a hemangioma, which can be confirmed by MRI or a 99mTc-labeled red blood cell scan. Alternatively, a biopsy should be considered to exclude malignancy, unless the lesion has the characteristic radiologic appearance of focal nodular hyperplasia. In a patient known to have cirrhosis, the presence of a nodular or mass lesion should be presumed to be hepatocellular carcinoma until proven otherwise. The AASLD practice

Figure 94-10.  A, Algorithm depicting the approach to management of a patient, not known to have cirrhosis, with a hepatic mass (often incidental, possibly symptomatic). B, Algorithm depicting the approach to management of a patient with known or suspected cirrhosis and a hepatic mass (found on routine surveillance because of symptoms or increasing AFP level). AFP, alpha fetoprotein; CT, computed tomography; HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging.

guidelines provide criteria for the noninvasive diagnosis of hepatocellular carcinoma based on the vascularity of the tumor (see earlier). Contrast enhancement during the arterial phase of a multiphase CT or MRI study with subsequent washout during the venous phase is considered diagnostic of hepatocellular carcinoma if the lesion is larger than 2 cm in diameter. For lesions between 1 and 2 cm in diameter, this characteristic vascularity should be demonstrated on two imaging modalities; biopsy of the lesion may be required. For a lesion smaller than 1 cm, biopsy may be difficult to accomplish technically, and the lesion should be observed; if it becomes larger than 1 cm in diameter, the diagnosis of hepatocellular carcinoma is essentially confirmed. If ascertaining whether a patient has underlying cirrhosis is difficult on clinical and imaging grounds, a biopsy of the nontumorous liver may be done. Ultrasonography generally is the first imaging study performed to evaluate a hepatic mass to determine whether the lesion is cystic or solid and whether the intrahepatic bile ducts are dilated. Ultrasonography is the cheapest and

1591

1592

Section IX  Liver most widely available hepatic imaging modality but is operator-dependent. Ultrasonography also is an effective way to demonstrate the presence of daughter cysts within hydatid cysts. Some tumors, such as biliary cystadenoma and cystadenocarcinoma, may be partly cystic. In addition, solid tumors may show hypoechoic areas interspersed with the expected hyperechoic picture. The hypoechoic areas result from the histologic characteristics of the tumor or from necrosis or hemorrhage in the lesion. Focal fatty change produces focal hypoechoic areas that also are hypodense on CT. Hemangiomas typically are hyperechoic. Ultrasonography does not permit diagnosis of a specific hepatic tumor to be made, and additional information needs to be obtained from dynamic contrast-enhanced ultrasonography, CT, or MRI. These dynamic studies demonstrate the vascular nature of certain tumors and can distinguish the arterial vascularity of hepatocellular carcinomas from the general increase in vascularity of hepatic adenomas. Metastases are generally much less vascular, with the exception of neuroendocrine tumors that have spread to the liver. Hemangiomas, although very vascular, enhance slowly with these dynamic studies because the vascularity is of capillary origin. Rapid washout of contrast dye from lesions that enhance during the arterial phase is a feature of hepatocellular carcinoma. Each technique has limitations, however, and, at present, an ideal method for imaging the liver is not available. Cavernous hemangiomas larger than 3 cm in diameter may be recognized on scintigraphy following the injection of radiolabeled red blood cells. Hepatic arteriography may show the typical features of vascular malignant tumors but is not always helpful and is required less frequently than in the past. If hepatocellular carcinoma or hepatoblastoma is suspected, the serum AFP concentration should be measured. Laparoscopy may demonstrate unsuspected tumor seeding or the presence of cirrhosis. Definitive diagnosis of a solid hepatic lesion depends on demonstrating the typical histologic features of the tumor. Histologic confirmation usually can be achieved by percutaneous needle biopsy or FNA, often with CT or ultrasound guidance.

KEY REFERENCES

Bismuth H, Chiche L, Castaing D. Surgical treatment of hepatocellular carcinomas in noncirrhotic liver: Experience with 68 liver resections. World J Surg 1995; 19:35-41. (Ref 101.) Bruix J, Sherman M. Management of hepatocellular carcinoma. Hepatology 2005; 42:1208-36. (Ref 25.) Colli A, Fraquelli M, Casazza G, et al. Accuracy of ultrasonography, spiral CT, magnetic resonance, and alpha-fetoprotein in diagnosing hepatocellular carcinoma: A systematic review. Am J Gastroenterol 2006; 101:513-23. (Ref 24.) Desmet VJ. Congenital diseases of intrahepatic bile ducts: Variations on the theme “ductal plate malformation.” Hepatology 1992; 16:1069-83. (Ref 253.) El-Serag H, Mason A. Rising incidence of hepatocellular carcinoma in the United States. N Engl J Med 1999; 340:745-50. (Ref 5.) Everson GT. Hepatic cysts in autosomal dominant polycystic kidney disease. Am J Kidney Dis 1993; 22:520-5. (Ref 262.) Khan SA, Thomas HC, Davidson BR, Taylor-Robinson SD. Cholangiocarcinoma. Lancet 2005; 366:1303-14. (Ref 162.) Lazaridis KN, Gores GJ. Cholangiocarcinoma. Gastroenterology 2005; 128:1655-67. (Ref 160.) Lencioni R, Cioni D, Crocetti L, et al. Early-stage hepatocellular carcinoma in patients with cirrhosis: Long-term results of percutaneous image-guided radiofrequency ablation. Radiology 2005; 234:961-7. (Ref 117.) Liaw YF, Sung JJ, Chow WC, et al. Lamivudine for patients with chronic hepatitis B and advanced liver disease. N Engl J Med 2004; 351:152131. (Ref 77.) Llovet J, Bruix J. Systematic review of randomized trials for unresectable hepatocellular carcinoma: Chemoembolization improves survival. Hepatology 2003; 37:429-42. (Ref 126.) Llovet JM, Ricci S, Mazzaferro V, et al. Sorafenib in advanced hepatocellular carcinoma. N Engl J Med 2008; 359:378-90. (Ref 136.) Marrero J, Fontana R, Fu S, Conjeevaram H, Su G, Lok A. Alcohol, tobacco and obesity are synergistic risk factors for hepatocellular carcinoma. J Hepatol 2005; 42:218-24. (Ref 93.) Nguyen MH, Garcia RT, Simpson PW, et al. Racial differences in effectiveness of alpha-fetoprotein for diagnosis of hepatocellular carcinoma in hepatitis C virus cirrhosis. Hepatology 2002; 36:410-17. (Ref 33.) Ribero D, Curley SA, Imamura H, et al. Selection for resection of hepatocellular carcinoma and surgical strategy: Indications for resection, evaluation of liver function, portal vein embolization, and resection. Ann Surg Oncol 2008; 15:986-92. (Ref 105.) Zucman-Rossi J. Genetic alterations in hepatocellular adenomas: Recent findings and new challenges. J Hepatol 2004; 40:1036-9. (Ref 204.) Full references for this chapter can be found on www.expertconsult.com.

ACKNOWLEDGMENT

The authors gratefully acknowledge the contributions of Professor Michael C. Kew, who authored this chapter in previous editions of the textbook.

CHAPTE R

95 Liver Transplantation Paul Martin and Hugo R. Rosen

CHAPTER OUTLINE Indications  1594 Listing Criteria and Policies of the United Network for Organ Sharing  1595 Absolute and Relative Contraindications  1595 Transplantation Evaluation and Listing  1598 Disease-Specific Indications  1598 Alcoholic Liver Disease  1598 Hepatitis B  1599 Hepatitis C  1600 Acute Liver Failure  1601 Cholestatic Liver Disease  1601 Hepatic Malignancy  1602 Metabolic Disorders  1603 Nonalcoholic Fatty Liver Disease  1603 Vascular Disorders  1603

Despite continued advances in the treatment of chronic liver disease, most notably antiviral therapy, and management of complications of chronic liver disease, such as transjugular intrahepatic portosystemic shunt (TIPS) placement, liver transplantation remains the only prospect for long-term survival in patients with advanced liver disease who have reached the limits of purely medical interventions. The major indications for liver transplantation in adults are decompensated cirrhosis, unresectable primary hepatic malignancies, and acute liver failure in which spontaneous recovery is not anticipated.1 Liver transplantation has continued to evolve in response to the shortage of deceased-donor organs and the frequency of recurrent disease. The availability of a wider array of immunosuppressive agents has made graft rejection a lesser threat than disease recurrence.2 Recurrence of hepatitis C virus (HCV) infection, because of its frequency as an indication for liver transplantation, high rate of graft reinfection and failure, and lack of effective prophylaxis, has become a major challenge in the care of liver transplant recipients.3 By contrast, modern regimens of prophylaxis to prevent hepatitis B virus (HBV) reinfection now allow liver transplantation to be undertaken with a low likelihood of recurrence.4 Recurrence of nonviral liver disease is now recognized as a threat to the graft, albeit of an order of magnitude less frequent than that for HCV reinfection.5 Because immunosuppression has a proviral effect, a number of liver transplantation centers have attempted earlier reduction or withdrawal of glucocorticoids in recipients with chronic HBV or HCV infection.6 By contrast, more intensive immunosuppression may be necessary in liver transplant recipients with autoimmune liver diseases.7 The apparently intractable shortage of deceased donors is reflected in continuing deaths of potential recipients listed for liver transplantation. A major change in the allocation of available organs has occurred with the introduction of

Autoimmune Hepatitis  1603 Other Indications  1603 Surgical Aspects of Liver Transplantation  1604 Native Hepatectomy  1604 Live-Donor Liver Transplantation  1605 Immunosuppression  1605 Postoperative Course  1606 Initial Phase to Discharge from Hospital  1606 Following Discharge from Hospital  1608 Long-Term Management  1609 General Preventive Medicine  1609 Immunizations and Antibiotic Prophylaxis  1610 When to Call the Transplantation Center  1611 Hepatic Retransplantation  1611

the Model for End-stage Liver Disease (MELD) score, which has achieved its stated aim of reducing the number of deaths on the liver transplantation waiting list.8 This method of organ allocation assigns organs to recipients on the basis of an objective, continuous measure of severity of liver disease, thereby removing time spent on the waiting list as a determining factor. Extension of live-donor liver transplantation (LDLT) to adult recipients has increased the organ pool despite the tempering of enthusiasm following an increased appreciation of potential risks to the donor.9 Other innovations such as splitting deceased-donor grafts to benefit two recipients and use of so-called marginal or extended criteria grafts, including those from older and non–heart-beating donors, also have expanded the organ supply modestly.10 An increased frequency of complications, most notably of the biliary tract, has been a consequence of expanding the donor criteria, however.11 Efforts to expand the deceaseddonor supply by public education programs have succeeded in enhancing organ donation, although the major dis­ crepancy between the number of potential recipients and the number of available organs persists, resulting in continuing attrition of listed patients who succumb to com­ plications of decompensated liver disease while awaiting liver transplantation.12 Although the organ allocation system will undoubtedly continue to evolve and recurrence of disease will remain a threat, the prospects for long-term survival are very good to excellent for most liver transplant recipients who otherwise would succumb to their liver disease. For instance, the likely one-year survival rate for patients with decompensated cirrhosis is less than 10% without liver transplantation but approximately 85% to 90% at one year and 75% at five years after transplantation for most indications.13 Access to liver transplantation has transformed the management of advanced liver disease but has resulted in an expanding cohort of decompensated potential recipients

1593

1594

Section IX  Liver who require frequent medical attention.14 Because the best outcomes following liver transplantation are obtained in patients who have not already experienced multiple complications of liver disease,15 referral for transplantation evaluation is appropriate when a cirrhotic patient has had an index complication, such as the onset of ascites. For at least some potential recipients, access to LDLT may avoid a lengthy waiting period with the risk of further and potentially life-threatening complications of liver disease. In parallel with the evolution of liver transplantation, the care of transplantation candidates with advanced disease has become a major clinical challenge. The transplantation hepatologist must combine the skills necessary to practice gastroenterology, multidisciplinary internal medicine, and intensive care. This skill set has been formally recognized by the development of a secondary subspecialty in transplantation hepatology by the American Board of Internal Medicine.16

INDICATIONS The major indications for liver transplantation in adults reflect the most frequent causes of adult liver disease, notably chronic hepatitis C, alcoholic liver disease, and, to a lesser extent, chronic hepatitis B, primary biliary cirrhosis (PBC), primary sclerosing cholangitis (PSC), autoimmune hepatitis, and hemochromatosis (Fig. 95-1). Many liver transplantation candidates previously diagnosed as having “cryptogenic” cirrhosis are now considered to have underlying nonalcoholic fatty liver disease (NAFLD). An uncommon but important indication for liver transplantation is acute liver failure, which has a high mortality rate in the absence of liver transplantation. The role of liver transplantation in primary hepatic malignancy has become better defined; a subset of patients with primary hepatocellular carcinoma (HCC) have a high likelihood of cure with transplantation, although the roles of adjunctive therapies such as transarterial chemoembolization (TACE), radiofrequency ablation, and the oral chemotherapy agent sorafenib need to be defined (see Chapter 94).17 The other major primary adult hepatic malignancy, cholangiocarcinoma, had been regarded as a contraindication to liver transplantation because of its rapid and almost invariable recurrence post–liver transplantation, although acceptable outcomes have been reported in a subset of patients with hilar tumors who receive adjuvant external beam radiation and chemo-

10%

33%

7% 4% 1% 6%

4%

14% 9%

12%

Hepatitis C Hepatitis B Alcohol Cryptogenic/nonalcoholic steatohepatitis Cholestatic disorders Hepatocellular carcinoma Other malignancies Metabolic disorders Pediatric diseases Miscellaneous causes

Figure 95-1.  Proportion of liver transplants performed for specific indications, 1992 to 2007. (From O’Leary JG, Lepe R, Davis GL. Indications for liver transplantation. Gastroenterology 2008; 134:1764-76, with permission.)

sensitization (see Chapter 69).18 The major indication for pediatric liver transplantation is biliary atresia following a failed Kasai procedure (portoenterostomy) or delayed recognition of the diagnosis (see Chapter 62). Other major pediatric indications include α1-antitrypsin deficiency and other metabolic disorders (see Chapter 76). Recognition of cirrhosis per se is not an indication for liver transplantation, although a key issue in managing a cirrhotic patient is deciding whether transplantation will be needed in the future and when referral for evaluation is appropriate (Table 95-1). Another important aspect of the management of a compensated cirrhotic patient is the anticipation of complications. Endoscopic surveillance of patients with cirrhosis for gastroesophageal varices, for instance, to offer prophylaxis, either pharmacologic or endoscopic, to prevent initial or recurrent gastrointestinal bleeding is now routine practice (see Chapter 90).14 Surveillance for HCC is also regarded as the standard of care in a cirrhotic patient (see Chapter 94). Discovery of a hepatic mass suggestive of HCC in a cirrhotic patient should prompt evaluation to determine whether hepatic resection or liver transplantation is the most appropriate curative approach. Otherwise, liver transplantation should normally be a consideration only when the limits of medical therapy for complications of cirrhosis have been reached, particularly if LDLT is an option. The risk of surgery must always be weighed against a realistic assessment of the potential recipient’s prognosis in the absence of liver transplantation. For example, in a patient with decompensated cirrhosis caused by HBV infection, effective suppression of viral replication by an effective antiviral agent may result in impressive clinical improvement, thereby delaying or even obviating the need for liver transplantation (see Chapter 78). Similarly, abstinence from alcohol can result in resolution of signs of hepatic decompensation in a patient with alcoholic liver disease (see Chapter 84). The course of chronic liver disease remains unpredictable, however, and observing an apparently well-compensated patient deteriorate dramatically because of an intercurrent complication such as variceal bleeding is sobering. Although recognition of cirrhosis implies a risk of major complications and diminished life expectancy, a cirrhotic patient can remain stable for a protracted period of time. For example, Fattovich and colleagues19 observed that in patients with well-compensated

Table 95-1  Indications for Liver Transplantation Acute liver failure Complications of cirrhosis Ascites Chronic gastrointestinal blood loss due to portal hypertensive gastropathy Encephalopathy Liver cancer Refractory variceal hemorrhage Synthetic dysfunction Liver-based metabolic conditions with systemic manifestations α1-Antitrypsin deficiency Familial amyloidosis Glycogen storage disease Primary oxaluria Tyrosemia Urea cycle enzyme deficiencies Wilson disease Systemic complications of chronic liver disease Hepatopulmonary syndrome Portopulmonary hypertension

Chapter 95  Liver Transplantation Probability of decompensation (%)

50

28% 10 yrs

25

0 0

12

24

36

48

60

72

84

96

108 120

355 336 304 253 205 133 104

65

41

29

Months 18

Patients at risk Figure 95-2.  Probability of development of major complications (decompensation) in patients with well-compensated cirrhosis due to hepatitis C. (From Fattovich G, Giustina G, Degos F, et al. Morbidity and mortality in compensated cirrhosis type C: A retrospective follow-up study of 384 patients. Gastroenterology 1999; 112:463-72, with permission.)

cirrhosis caused by HCV infection, major complications of portal hypertension such as ascites and variceal hemorrhage occurred in less than 30% at 10 years; in the absence of an index complication, survival was excellent (Fig. 95-2). Once a complication supervenes, however, survival diminishes rapidly. For example, after the development of ascites refractory to diuretics, only 25% of patients survive beyond 1 year.20 A prospective study of more than 200 Italian patients with HCV-related compensated cirrhosis followed for up to 17 years found that HCC was the most common complication of cirrhosis detected, occurring in 32%, followed by ascites. These findings imply that HCC, rather than complications of portal hypertension, is the most frequent reason for a worsening prognosis in a patient with cirrhosis.21 The development of predictive models based on the natural history of PBC (see Chapter 89) and PSC (see Chapter 68)22 has helped clinical decision making for patients with these cholestatic disorders, which tend to progress in a fairly stereotypical fashion. Before the introduction of the MELD score, analogous models had not been available for the noncholestatic forms of cirrhosis, and the decision to refer a patient for liver transplantation generally was based on an estimate of disease severity using objective parameters such as the serum albumin level as well as more subjective factors such as the presence of hepatic encephalopathy, as in the Child-Turcotte-Pugh score (see Chapter 90). On clinical rather than biochemical grounds, important indications for liver transplantation remain disease severity reflective of hepatocellular failure, reflected by coagu­ lopathy and jaundice; complications of portal hypertension, such as refractory ascites and recurrent variceal bleeding; or the combination of portosystemic shunting and diminished hepatocellular function, as in hepatic encephalopathy (see Table 95-1). Following validation of predictive models for the natural history of PBC and PSC, prediction of an individual patient’s course has been possible on the basis of simple clinical and laboratory parameters, the most ominous of which is a rising serum bilirubin level. Although

deterioration in a patient’s quality of life typically may not be reflected adequately in predictive models, including MELD, the presence of potentially disabling symptoms such as pruritus and osteopenia in patients with cholestatic and other forms of cirrhosis, as well as recurrent bacterial cholangitis in those with PSC, are important considerations in deciding when to refer a patient for liver transplantation. MELD exceptions, such as adding points to the so-called biological MELD score in order to increase the likelihood of liver transplantation, require approval by a local regional review board, which includes representatives from local transplantation programs. The awarding of extra MELD points recognizes that although this system is a major advance in organ allocation, at least some patients may be disadvantaged by its use of purely objective parameters and exclusion of factors such as intractable ascites or encephalopathy that were incorporated into older allocation schemes. Ideally, liver transplantation should occur before a protracted period of disability reduces the likelihood that post-transplantation rehabilitation will lead to full employment and normal social functioning.

LISTING CRITERIA AND POLICIES OF THE UNITED NETWORK FOR ORGAN SHARING Organ allocation within the United States is administered by the United Network for Organ Sharing (UNOS), which now considers only disease severity (and not waiting time, as in the past) to determine a patient’s priority for liver transplantation. A variety of organ allocation systems have been used since the 1980s to allocate the limited number of deceased-donor organs in an equitable manner. Older systems were based on a combination of clinical and biochemical parameters such as the Child-Turcotte-Pugh score. The MELD score is a mathematical formula (available at www.unos.org) that incorporates the serum bilirubin level, creatinine level, and international normalized ratio (INR) and provides a more objective and accurate way to stratify liver transplantation candidates for organ allocation and to eliminate time waiting for a donor organ as a determining factor.23 The MELD score overcomes some of the inherent limitations of the Child-Turcotte-Pugh score, including limited discriminatory ability, subjective interpretation of parameters such as presence or absence of ascites on the basis of the physical examination, and the “ceiling effect” of the Child-Turcotte-Pugh score (e.g., no greater weight is given to a serum bilirubin level of 35 mg/dL than to a level of 3.5 mg/dL, even though a patient with the markedly higher bilirubin level clearly has more advanced liver disease). Inclusion of the serum creatinine level reflects the major prognostic importance of renal dysfunction in patients with advanced liver disease. Adoption of the MELD score has been a major step in achieving an equitable organ allocation system in the United States, although it will undoubtedly continue to undergo refinement.

ABSOLUTE AND RELATIVE CONTRAINDICATIONS Contraindications to liver transplantation have also evolved, reflecting innovations in surgical technique and postoperative care. As an example of the latter, effective prophylaxis against HBV recurrence now allows excellent graft and patient survival rates in patients with chronic hepatitis B.24 On the other hand, greater experience has highlighted the

1595

1596

Section IX  Liver Table 95-2  Absolute Contraindications to Liver Transplantation Acquired immunodeficiency syndrome Active alcoholism or substance abuse Advanced cardiac or pulmonary disease Anatomic abnormality that precludes liver transplantation Child-Turcotte-Pugh score <7 Cholangiocarcinoma Extrahepatic malignancy Fulminant hepatic failure with sustained ICP >50 mm Hg or CPP <40 mm Hg* Hemangiosarcoma Persistent noncompliance Uncontrolled sepsis *ICP, intracranial pressure; CPP, cerebral perfusion pressure; CPP, equals the mean arterial pressure minus ICP.

futility of retransplantation in many debilitated recipients with a failing graft caused by recurrent HCV infection.25 By contrast, the introduction of highly effective antiretroviral therapy has allowed consideration of liver transplantation in human immunodeficiency virus (HIV)–infected patients with decompensated liver disease, typically caused by either HCV or HBV infection.26 Still, absolute and relative contraindications remain. An absolute contraindication (Table 95-2) to liver transplantation is a clinical circumstance in which the likelihood of a successful outcome is so remote that liver transplantation should not be offered. A relative contraindication implies a suboptimal likelihood of a good outcome, although liver transplantation may still be considered in some patients. The role of liver transplantation in the management of HCC has become better defined with the recognition that a large tumor burden is associated with a high probability of metastatic spread within a short time of liver trans­ plantation.27 Some tumor characteristics predictive of a poor outcome, most notably vascular invasion, may only be apparent once the explant is examined, despite the sophistication of current imaging techniques. Although results of liver transplantation for cholangiocarcinoma have been poor because of a high rate of tumor recurrence, a subset of patients with hilar tumors may benefit from multimodal therapy and liver transplantation. The results of liver transplantation remain poor for angiosarcoma, and its recognition pre–liver transplantation remains an absolute contraindication. By contrast, at least some patients with epithelioid hemangioendothelioma have been transplanted successfully, despite an extensive tumor burden, with documented regression of extrahepatic metastases. For a transplantation candidate with a prior extrahepatic malignancy, therapy needs to have been curative, with the resection specimen indicating a low likelihood of metastatic spread. A two-year recurrence-free interval is adequate for most nonhepatic malignancies prior to liver transplantation, but for breast cancer, colon cancer, and malignant melanoma, a longer period following resection is desirable.28 Myeloproliferative disorders frequently underlie BuddChiari syndrome (see Chapter 83), but fortunately their evolution to acute leukemia is not accelerated following liver transplantation.29 Ongoing alcohol or recreational drug use remains an absolute contraindication to liver transplantation. If con­ tinued abuse is a concern, random toxicology screens are appropriate. Although medicinal marijuana may be used legitimately for palliation, most transplantation programs discourage the use of marijuana because of concerns about

the overall compliance of users and possible pulmonary side effects, as well as some evidence that its use may adversely affect the course of liver disease (a point that may be moot in a patient already listed for liver transplantation). A history of prescription narcotic abuse is also a cause for concern because it may contribute to difficulties with pain management post–liver transplantation. Non-narcotic alternatives should be attempted for the management of chronic pain. Other analgesics such as nonsteroidal antiinflammatory drugs (NSAIDs) also are contraindicated in persons with end-stage liver disease because of potential renal and gastrointestinal complications. Cigarette smoking is prohibited in liver transplantation candidates as well because of its multiple adverse effects, including an asso­ ciated risk of hepatic artery thrombosis and malignancy post–liver transplantation.30 With the increasing use of herbal compounds and other so-called health products, a discussion of their unproven efficacy and unknown toxicities and caution against their use in the post–liver transplantation setting because of the potential for drug interactions are appropriate (see Chapter 87).31 The careful medical evaluation necessary prior to liver transplantation frequently uncovers important comorbid­ ities, typically cardiac and pulmonary. Although patients with decompensated cirrhosis were previously thought to be somewhat protected against coronary artery disease (CAD) because of low afterload, reflecting peripheral vasodilatation, decreased hepatic synthesis of cholesterol, and increased circulating estrogen levels, cirrhotic patients have a prevalence of CAD at least as great as that of an agematched control population.32 Risk factors for CAD in cirrhotic patients include a high prevalence of diabetes mellitus. Additional risk factors in the post–liver trans­ plantation period include immunosuppressive drugs that contribute to systemic hypertension, hyperlipidemia, and obesity. Assessment of cardiac risk in cirrhotic patients may be inadequate because of poor physical stamina during routine stress testing. Administration of dobutamine mimics the physiologic effects of exercise and is used with stress echocardiography to exclude clinically significant CAD in liver transplant recipients with cirrhosis. Patients who reach 85% of the maximal predicted heart rate without an abnormality on stress echocardiogram have a low likelihood of peri- and postoperative ischemic cardiac events.33 Discrete coronary artery stenoses can be managed by angioplasty and stenting pre–liver transplantation. Although surgical bypass grafting may be contraindicated because of a significant perioperative risk of excessive bleeding in a patient with decompensated cirrhosis, such surgery, if successful, may render a patient an acceptable candidate for liver transplantation. Pre–liver transplantation cardiac evaluation may overestimate cardiac performance. impaired cardiac function may become apparent only after the protective effect of the decreased systemic vascular resistance that characterizes cirrhosis is lost after liver transplantation, when afterload increases because of the hypertensive effects of the primary immunosuppressive agents and when overvigorous volume repletion may occur.34 Specific forms of cirrhosis may be associated with extrahepatic manifestations that diminish long-term survival. For example, lethal cardiac arrythmias may result in poorer survival in patients who undergo liver transplantation for decompensated cirrhosis caused by hemochromatosis.35 Pulmonary evaluation in the liver transplantation candidate may reveal abnormal arterial oxygenation (see Chapter 92). Although severe chronic obstructive pulmonary disease or pulmonary fibrosis precludes liver transplantation, respiratory restriction as a result of ascites or diminished respira-

Chapter 95  Liver Transplantation tory muscle strength caused by chronic illness is reversible and does not preclude liver transplantation. Even patients who undergo liver transplantation for α1-antitrypsin deficiency may show improvement in pulmonary function tests postoperatively.36 The hepatopulmonary syndrome (HPS) is characterized by the triad of chronic liver disease, pulmonary vascular dilatations (with right-to-left shunting), and hypoxemia.37 The diagnosis is suggested by the finding of a Pao2 value of less than 70 mm Hg on an arterial blood gas determination obtained from the patient in the supine position. Definitive diagnosis is made by the demonstration of intrapulmonary vascular dilatations by contrast-enhanced echocardiography (which is the most sensitive technique), perfusion lung scanning with 99mTc–labeled macroaggregated albumin, or pulmonary arteriography. Detection of contrast in the left side of the heart within several beats after its appearance in the right atrium indicates intrapulmonary shunting. Predictors of potential reversibility of HPS after liver transplantation include younger age, a lesser degree of preoperative hypoxemia, and adequate correction of hypoxemia with inspiration of 100% oxygen (Pao2 >200 mm Hg).38 In the majority of patients with HPS, hypoxemia resolves within several months of liver transplantation, although a protracted period of ventilatory support in the immediate postoperative period may be required. Because of the potential for improvement with liver transplantation, extra MELD points may be allocated to a patient with HPS. HPS must be distinguished from portopulmonary hypertension, because the latter is associated with high perioperative mortality and frequently unchanged pulmonary hemodynamics despite liver transplantation. Specifically, documentation of a mean pulmonary arterial pressure greater than 35 mm Hg, pulmonary vascular resistance greater than 300 dynes • s • cm−5, and cardiac output less than 8 L/minute are indicative of a high perioperative risk because the patient will be unable to increase the cardiac output appropriately in response to altered intra- and postoperative hemodynamics. Vasodilator therapy may reduce pulmonary arterial pressure and permit liver transplantation.39 Hepatic hydrothorax is a frequent manifestation of portal hypertension characterized by collection of a transudate in the pleural cavity, usually on the right side and often with relatively little ascites remaining in the abdominal cavity (see Chapter 91). Hepatic hydrothorax can be particularly difficult to manage and often requires repeated tho­ racentesis or placement of a TIPS pre–liver transplantation.40 The temptation to insert a permanent chest tube needs to be resisted because a chest tube can lead to infection in the pleural cavity with the risk of fistula formation. Similarly, interventions such as pleurodesis or pleural decortication should be avoided because hepatic hydrothorax can be controlled only by a reduction in portal pressure. Active uncontrolled extrahepatic infection is an absolute contraindication to liver transplantation, which should be deferred if sepsis is suspected. In a patient with decompensated cirrhosis, an unexplained clinical deterioration such as altered mental status or systemic hypotension in the absence of gastrointestinal hemorrhage must be presumed to reflect sepsis and is an indication to start broad antibiotic coverage while culture results are awaited. Liver transplantation, however, may be the only option for a patient with recurrent bacterial cholangitis in the setting of PSC (see Chapter 68). Repeated bouts of spontaneous bacterial peritonitis need to be controlled by antibiotic therapy prior to attempting liver transplantation (see Chapter 91). A particularly ominous finding is fungemia, which is typically impossible to eradicate in a debilitated cirrhotic patient and precludes liver transplantation.

An important technical consideration in the liver transplantation candidate is the presence of vascular abnor­ malities that may increase the difficulty of surgery. With increased surgical experience, however, such abnormalities, most notably portal vein thrombosis, are less likely to be obstacles to liver transplantation. More extensive vascular thrombosis with involvement of the superior mesenteric vein may require extensive vascular reconstruction.41 The presence of a prior portosystemic shunt, particularly a nonselective (side-to-side or end-to-side) portocaval shunt, increases the technical complexity of liver transplantation but is no longer regarded as a contraindication. With the widespread use of TIPS to control manifestations of portal hypertension, including variceal hemorrhage, intractable ascites, and hydrothorax, without disrupting the vascular anatomy, TIPS is now the most frequent shunt encountered in liver transplant recipients and does not usually present an additional operative challenge during liver transplantation. Age restrictions have been relaxed in liver transplantation candidates, although close attention must be paid to comorbid conditions in older patients that not only increase the risk of perioperative mortality, but also may decrease the likelihood that the liver transplant recipient will be able to return to an active lifestyle, particularly because severe liver disease may cause more debility in older than in younger patients.42 Because a subset of robust older recipients have good outcomes after liver transplantation, candidates in their late 60s or even older who are otherwise in good health should not be precluded a priori from liver transplantation. The differential diagnosis of renal insufficiency in patients with advanced liver disease includes hepatorenal syndrome, which is potentially reversible. Renal failure remains an important predictor of a poor outcome post–liver transplantation (see Chapter 92).43 Typically, renal dysfunction in patients with decompensated cirrhosis reflects a variety of insults, including sepsis, hypotension, and use of nephrotoxic medications. In liver transplant recipients with decompensated cirrhosis, renal insufficiency severe enough to require dialysis has been associated consistently with poorer patient outcomes. Assessment of the potential for improvement following liver transplantation is critical. A rule of thumb is that return of adequate renal function is unlikely after liver transplantation if dialysis has been required for more than one month prior to liver transplantation. Inclusion of the serum creatinine level in the MELD score reflects the major prognostic importance of renal insufficiency in patients with advanced liver disease. A consequence of this recognition has been an increased rate of combined liver-kidney transplantation in the MELD era of liver transplantation, with a consequent depletion in the supply of kidneys for patients awaiting isolated deceaseddonor renal transplantation.44 An important reflection of impaired free-water handling in patients with decompensated cirrhosis is dilutional hyponatremia. Consequences of hyponatremia include altered mental status with more profound degrees of hyponatremia and an increased risk of calcineurin-induced neurotoxicity after liver transplantation (see later). Incorporation of the serum sodium level into the MELD formula (MELDNa) may increase the prognostic accuracy of the MELD score, particularly in patients with relatively low MELD scores.45 A major systemic manifestation of decompensated cirrhosis is malnutrition. Loss of muscle mass increases the likelihood of perioperative morbidity, with a need for more protracted ventilator support and poorer patient survival. Peripheral edema and ascites make changes in body weight

1597

1598

Section IX  Liver or anthropometric measurements unreliable for assessing nutritional status in a patient with advanced cirrhosis. More profound nutritional deficiencies may reflect the specific cause of cirrhosis, as in a malnourished alcoholic person with multiple vitamin and electrolyte deficiencies or a patient with cholestatic liver disease and depletion of fatsoluble vitamins as a result of small intestinal malabsorption. Evaluation by a dietitian is an integral part of the assessment of the liver transplantation candidate. Attempts to improve the nutritional status of liver transplantation candidates have included enteral and parenteral feeding, which may result in improvement, albeit modest, in some patients.46 An increasingly obese liver transplantation candidate pool is raising concerns about the role of obesity in the pathogenesis of NAFLD and in postoperative mortality resulting from cardiovascular events, as well as postoperative complications such as wound infections.47

TRANSPLANTATION EVALUATION AND LISTING The details of the formal liver transplantation evaluation vary from center to center, but the essentials are to establish that liver transplantation is indicated in the management of the potential recipient’s liver disease, the patient has no comorbidities severe enough to preclude transplantation, and the patient has adequate emotional and social resources to undergo a major surgical procedure and continue on long-term immunosuppression afterward (Table 95-3). Generally in the United States, clearance is needed from the patient’s insurance carrier before the extensive testing necessary for liver transplantation evaluation is undertaken. The patient is seen by a transplantation surgeon, hepatologist, psychiatrist, dietitian, and social worker, with additional consultations as clinically indicated. As increasingly frailer and older candidates are evaluated, identifying potential causes of perioperative morbidity, such as carotid artery stenosis, is imperative. Detailed abdominal imaging is performed not only to screen for HCC, but also to uncover vascular abnormalities such as portal vein thrombosis that may make surgery technically challenging. Disease-specific issues need to be addressed, such as the likelihood of recidivism in the alcoholic patient or management of a large

tumor burden in the patient with HCC. The appropriateness of liver transplantation for each candidate is then discussed formally at a meeting of the patient selection committee with input from the members of the transplantation team. If the patient’s candidacy is deemed to be appropriate, formal listing is undertaken with UNOS by matching of the recipient by blood type and weight with potential deceased donors. Once listed, the patient’s priority for organ allocation is determined by the MELD score, based either on the pure score consisting of the serum bilirubin level, serum creatinine level, and INR or that score plus additional points awarded for specific indications, such as HCC. Waiting time on the list is no longer a determining factor. With the critical and seemingly intractable shortage of deceaseddonor organs, a major challenge for UNOS and organ retrieval agencies elsewhere in the world has been to develop an equitable system of allocation in an effort to ensure that hepatic allografts are not used for patients whose prognosis without liver transplantation remains good. Patients with a MELD score of less than 15 appear have a better survival without transplantation than with transplantation. As shown in Figure 95-3, the MELD score has been shown to correlate with the three-month survival rate. Patients with a MELD score of less than 10 are ineligible for active listing with UNOS unless they are eligible to receive extra points because of an additional complication of liver disease, such as HCC or HPS. Once the evaluation process is complete and the patient is accepted for liver transplantation, financial clearance is sought from the patient’s private, state, or federal insurer to fund the procedure. Unfortunately, criteria for coverage for liver transplantation vary among insurers; however, in the United States, if Medicare, the major federal payor, funds a particular indication for liver transplantation, the other insurance carriers generally follow suit.

DISEASE-SPECIFIC INDICATIONS ALCOHOLIC LIVER DISEASE

Despite the high frequency of chronic HCV infection as an indication for liver transplantation (see later), alcoholic liver disease (ALD) remains the most frequent cause of

Table 95-3  Transplantation Evaluation Process Financial screening Medical evaluation Hepatology evaluation Laboratory testing Cardiac evaluation Hepatic imaging General health assessment Transplantation surgery evaluation Anesthesia evaluation Psychiatry or psychology consultation Social work evaluation Financial and insurance counseling Nutritional evaluation

Secure approval for evaluation As discussed in text Confirm diagnosis and optimize management Assess hepatic synthetic function, serum electrolytes, renal function, viral serologies, markers of other causes of liver disease, tumor markers, ABO-Rh blood typing; 24-hour urine for creatinine clearance; urinalysis and urine drug screen Electrocardiography and two-dimensional echocardiography, stress testing and cardiology consult if risk factors are present and/or patient is age 40 years or older Ultrasonography with Doppler to document portal vein patency, triple-phase computed tomography or gadolinium magnetic resonance imaging for tumor screening Chest film, prostate specific antigen level (men), Pap smear and mammogram (women), colonoscopy if patient is age 50 years or older or has primary sclerosing cholangitis Assess technical issues and discuss risks of procedure Required if unusually high operative risk, e.g., patient has portopulmonary hypertension, hypertrophic obstructive cardiomyopathy, previous anesthesia complications If history of substance abuse, psychiatric illness, or adjustment difficulties Address potential psychosocial issues and possible effect of transplantation on patient’s personal and social system Itemize costs of transplantation and post-transplantation care; help develop financial management plans Assess nutritional status and patient education

From O’Leary JG, Lepe R, Davis GL. Indications for liver transplantation. Gastroenterology 2008; 134:1764-76, with permission.

Chapter 95  Liver Transplantation pathologic drinking after liver transplantation appear to have a higher rate of medical problems, including pneumonia, cellulitis, and pancreatitis, that require hospital admission and occasionally lead to graft loss and death. In addition, alcoholic liver transplant recipients are prone to develop de novo oropharyngeal and lung tumors, likely reflecting other aspects of an alcoholic lifestyle, most notably cigarette smoking.48

100

3-month survival (%)

80

60

HEPATITIS B

40

20

0 0

20

40 50 30 MELD score Figure 95-3.  Relationship between the three-month survival rate and Model for End-stage Liver Disease (MELD) score in patients with cirrhosis. 10

decompensated chronic liver disease (see Chapter 84). Decompensated alcoholic cirrhosis is now firmly established as a legitimate indication for liver transplantation despite some lingering controversy.48 Concerns in the past included the risk of recidivism following liver transplantation as well as potentially poor compliance by transplant recipients with a history of alcoholism. In addition, the large number of patients with ALD was thought to have the potential to outstrip the donor supply. These fears have not been confirmed, and even patients with evidence of acute alcoholic hepatitis in the explant do not appear to have inferior post–liver transplantation survival rates despite some earlier reports to the contrary. Excellent graft and patient survival rates are the norm following liver transplantation for ALD. Key factors in determining an alcoholic patient’s suitability for liver transplantation include recognition by the patient of the key role alcohol has played in the genesis of the liver disease, participation in some form of alcohol rehabilitation, such as attendance at Alcoholics Anonymous, stable social support, and a defined period of abstinence from alcohol before transplantation. Conventionally this period of abstinence has been six months, although rigorous studies have failed to confirm that this duration of abstinence per se confers a high likelihood of continued sobriety but have emphasized the importance of factors such as a lack of either isolation or depression. Despite these strategies, however, a substantial proportion of alcoholic recipients resume drinking after liver transplantation, although surprisingly, graft loss or early death attributable to alcohol abuse has been uncommon. A higher rate of return to alcohol use is elicited by use of anonymous questionnaires or toxicology screening than by direct questioning of patients. With longer-term follow-up, as many as 40% of alcoholic recipients resume alcohol use.49 A particularly difficult dilemma arises in the alcoholic patient with severely decompensated liver disease and alcohol use until the time of hepatic decompensation in whom the likelihood of surviving without prompt liver transplantation is low. Clearly enunciated criteria, including a contractual commitment by the patient to sobriety and active involvement in some form of alcohol rehabilitation such as participation in Alcoholics Anonymous, ensure that the selection process is equitable under these circumstances. Patients who return to

Effective prevention of graft reinfection in the HBV-infected candidate has been a major triumph of liver transplantation. HBV recurrence was frequent and resulted in reduced patient and graft survival rates during the 1980s; as a result, Medicare refused to fund liver transplantation in HBVinfected persons. Key factors in improving outcomes included recognition of the key role of pre–liver transplantation active viral replication, as demonstrated by detection in serum of hepatitis B e antigen or HBV deoxyribonucleic acid (DNA) by molecular hybridization techniques as a predictor of recurrent HBV infection in the graft and the protective effect of long-term use of high-dose hepatitis B immunoglobulin (HBIG). In a seminal study50 Samuel and colleagues observed that patients with a fulminant presentation of acute HBV infection or hepatitis D virus coinfection had a reduced risk of recurrent hepatitis B post–liver transplantation in the absence of immunoprophylaxis, reflecting the lower level of HBV replication in patients with fulminant hepatitis B than in those with chronic hepatitis B. Long-term administration of high-dose HBIG, initially administered intravenously, resulted in markedly reduced rates of recurrence of hepatitis B post-transplantation; HBIG administered in combination initially with the nucleoside analog lamivudine further decreased the rate of HBV recurrence. Lamivudine had been used as monotherapy to prevent recurrent HBV infection post-transplantation but was limited by frequent mutations in HBV polymerase gene, leading in turn to graft reinfection.51 The optimal dosing regimen for HBIG has been difficult to establish in the absence of controlled clinical data. Some groups have titrated HBIG doses according to trough serum levels of antibody to hepatitis B surface antigen (anti-HBs). Sub­ sequently, intramuscular administration of HBIG has been confirmed as an efficacious, and less expensive, alternative to intravenous regimens, when used in combination with lamivudine.52 HBV infection that recurs despite administration of HBIG may reflect inadequate doses of HBIG or a genomic mutation in the “a” moiety of HBsAg that results in less avid binding of the virus to HBIG.53 Lamivudine resistance, acquired before liver transplantation, has also been implicated in HBV recurrence despite apparently adequate immunoprophylaxis after liver transplantation. The availability of additional oral antiviral agents with efficacy against HBV has expanded options for preventing graft reinfection (see Chapter 78). A large, multicenter study, for example, has demonstrated that adefovir dipivoxil taken for 48 weeks results in significant virologic, biochemical, and clinical improvement in patients with chronic HBV infection both pre– and post–liver transplantation and may obviate the need of liver transplantation because of improved hepatocellular function.54 Despite the complexities of managing HBV infection in the liver transplantation patient, excellent graft and patient survival rates are now routine, in contrast to the gloomy picture for HBV-infected liver transplantation candidates prior to the use of HBIG. The ever-expanding list of oral antiviral agents promises to improve outcomes further, even in patients in whom antiviral resistance develops. With the licensing of several

1599

1600

Section IX  Liver effective oral agents to treat HBV, HBIG use will likely be superseded by the use of combinations of oral antiviral agents. The liver transplantation candidate with chronic HBV infection is now readily accepted for liver transplantation, albeit with the requirement for long-term antiviral immunoprophylaxis.

HEPATITIS C

A major concern in the liver transplantation community, apart from the shortage of donor organs, is the burden of recurrent HCV infection after liver transplantation. After HCV was initially identified, early reports suggested that recurrent HCV infection with detection of viremia post– liver transplantation was frequent and did not have an adverse effect on overall patient or graft survival rates in the first several years after liver transplantation. Extensive experience now indicates, however, that HCV-infected liver transplant recipients have inferior long-term outcomes, compared with patients transplanted for other causes of cirrhosis, because of graft injury from recurrent HCV infection. Analysis of serial liver biopsy specimens from liver transplant recipients with recurrent HCV infection has identified accelerated fibrosis and progression to cirrhosis compared with immunocompetent patients with HCV infection. A particular challenge is to identify liver transplant recipients with recurrent HCV infection at increased risk of rapidly progressive graft injury. Less than 10% of patients with mild recurrent hepatitis C at one year post–liver transplantation appear to progress to allograft cirrhosis by five years. By contrast, two thirds of patients with at least moderately severe hepatitis C at one year post-transplantation progress to cirrhosis by 5 years.55 Concern has been raised, however, that with prolonged follow-up, some patients with initially mild recurrent hepatitis C after liver transplantation may experience a more aggressive course. A study by Berenguer and colleagues56 evaluated serial protocol liver biopsy specimens to assess the histologic outcome of 57 HCV genotype 1b-infected liver transplant recipients with an initially mild recurrence, defined as no or minimal fibrosis (fibrosis stage F0 or F1) during the first three years posttransplantation (see Chapter 79).56 Progression to bridging fibrosis or cirrhosis (stage F3 or F4) occurred in 35% (n = 20) of such patients, with 12 recipients progressing to stage F3 and 8 recipients progressing to stage F4. Variables associated with progressive fibrosis on univariate analysis were the baseline fibrosis stage and activity grade (P < 0.0001), recipient female gender (P = 0.04), the serum alanine aminotransferase (ALT) level at one year post-transplantation (P = 0.02), and the aspartate aminotransferase (AST) and ALT levels at baseline (P = 0.008 and P = 0.005, respectively). By multivariate analysis, the only variable that was significant was fibrosis stage at baseline (relative risk, 11; 95% CI, 3 to 41; P = 0.0007). Therefore, delayed HCVrelated severe liver damage is frequent, even in transplant recipients with an initially mild recurrence, and is seen in approximately one third of patients. Some degree of fibrosis at baseline appears to predict accelerated recurrent HCV infection. A particularly ominous finding is prominent biochemical and histologic cholestasis that frequently is a precursor to rapid allograft failure. Antiviral therapy for patients with a cholestatic HCV recurrence may need to be continued indefinitely because discontinuation leads to a rapid relapse of the cholestatic syndrome and death.3 Reliable predictors of severe recurrent HCV infection have been difficult to identify, although a number of viral and nonviral factors have been implicated (Table 95-4). Infection with viral genotype 1b had been suggested to be a

Table 95-4  Factors Associated with Severe Hepatitis C Virus Recurrence Following Liver Transplantation Viral Factors Absence of pretransplantation hepatitis B viral coinfection Cytomegalovirus coinfection High serum HCV RNA levels before transplantation and within two weeks after transplantation Viral genotype 1b Immunosuppression Multiple episodes of rejection (indicating a high cumulative prednisone dose) Use of OKT3 to treat rejection Other Factors High tumor necrosis factor-α production in the graft Impaired HCV-specific CD4+ T-cell responses Ischemic-preservation injury Nonwhite recipient HCV, hepatitis C virus; RNA, ribonucleic acid.

key predictor, but this observation has not been universal. Higher serum levels of HCV RNA before liver transplantation and immediately after liver transplantation and possibly more rapid evolution of HCV quasispecies have been described in patients with more aggressive recurrent HCV56 (see also Chapter 79). Older deceased-donor age has been implicated repeatedly. Episodes of acute cellular rejection, particularly if multiple, lead to a greater likelihood of severe recurrent hepatitis C. A major challenge is to distinguish recurrent HCV infection from graft rejection, particularly because many of the histologic hallmarks of acute rejection, including bile duct injury, are also consistent with recurrent HCV infection. Examination of serial liver biopsy specimens may help clarify this issue and allow avoidance of inappropriate additional immunosuppression in the recipient with recurrent HCV infection rather than graft rejection. Once recurrent HCV infection in the graft progresses to cirrhosis, overt hepatic decompensation is frequent. In contrast to recurrent HBV infection, effective prophylaxis against recurrent HCV infection has not been possible (Fig. 95-4).3 Current treatment strategies generally fall into three categories: (1) pretransplantation antiviral therapy; (2) preemptive therapy started in the early post–liver transplantation period before the development of clinically apparent acute hepatitis C; and (3) post-transplantation therapy at the time of diagnosis of acute hepatitis C or for established or severe chronic hepatitis C.3 The approach followed by most transplantation centers is to initiate antiviral therapy when clinically significant evidence of recurrent HCV infection is identified (as defined by either grade 3 or 4 (of 4) hepatic inflammation or a fibrosis stage of F2 or more). Accurate assessment of the efficacy of treating recurrent HCV infection is difficult, however, because most reported studies have been uncontrolled, single-centered, and small in size. Interferon alpha monotherapy generally has been ineffective in treating established recurrent hepatitis C (see Chapter 79). When interferon is combined with ribavirin, the rate of virologic response is increased, but at the cost of frequent and potentially severe side effects.3 Leukopenia is a par­ ticularly vexing problem in patients undergoing such treatment, and adjunctive administration of granulocyte colony– stimulating factor (G-CSF) may permit continuation of interferon therapy. The long-acting pegylated interferons have been evaluated in a number of clinical trials. Generally, prophylactic interferon-based regimens administered

Chapter 95  Liver Transplantation .8

Table 95-5  Criteria for Liver Transplantation in Acute Liver Failure +++

Probability (%)

.6 +++

++

+ +

.4 +++

.2

+

++

++ 0.0 0

90 180 270 360 450 540 630 720 810 900 9901080 Follow-up in days

Figure 95-4.  Probability of hepatic decompensation in patients with cirrhosis resulting from recurrent hepatitis C following liver transplantation. (From Berenguer M, Prieto M, Rayon JM, et al. Natural history of clinically compensated hepatitis C virus-related graft cirrhosis after liver transplantation. Hepatology 2000; 32:852-8, with permission.)

shortly post–liver transplantation in an effort to prevent graft reinfection have resulted in low rates of sustained virologic response.57 On the basis of reports of rejection and graft loss in renal transplant recipients treated with interferon, as well as preliminary experience in liver transplant recipients, concern has been raised that therapy with interferon increases the risk of graft rejection (although as noted earlier the distinction between recurrent HCV infection and graft rejection is difficult even on histologic grounds).58 Increasingly, recurrent HCV infection is recognized as the cause of graft failure in transplant recipients, and the dilemma arises as to whether repeat liver transplantation in affected patients is justified.25 A subset of patients retransplanted for graft loss caused by recurrent hepatitis C have reasonable survival rates, if they do not have deep jaundice or renal failure at the time of retransplantation (see later).

ACUTE LIVER FAILURE

Acute liver failure (ALF) is an uncommon but important indication for liver transplantation because of its high mortality rate but excellent outcomes with prompt liver transplantation (unless major neurologic complications have occurred). ALF is defined as the onset of hepatic encephalopathy within 26 weeks of the initial recognition of acute liver disease and reflects a variety of causes (see Chapter 93). Despite an abrupt onset, antecedent chronic liver disease is absent, and hepatic recovery is possible. In the past, liver transplantation for ALF resulted in poorer patient survival rates than those for benchmark indications, such as PBC. Subsequent experience, however, has shown that excellent patient survival rates are possible if ALF is identified early in its course and the patient is listed for liver transplantation before irreversible complications, especially neurologic, have supervened.59 The absence of papilledema on funduscopy and of typical features of cerebral edema on computed tomography (CT) of the head do not preclude the presence of cerebral edema complicating worsening encephalopathy; therefore, direct

Criteria of King’s College, London Acetaminophen cases Arterial pH <7.3, or INR >6.5 and serum creatinine >3.4 mg/dL Non-acetaminophen cases INR >6.5, or Any three of the following: Age <10 years or >40 years Duration of jaundice before encephalopathy >7 days Etiology: non-A, non-B hepatitis; halothane hepatitis; idiosyncratic drug reaction; indeterminate Serum bilirubin >17.6 mg/dL INR >3.5 (PT >50 seconds) Criteria of Hôpital Paul-Brousse, Villejuif Hepatic encephalopathy, and Factor V level <20% in patients
intracranial pressure monitoring may be required to detect and manage this frequently lethal complication of ALF. Direct intracranial pressure monitoring can only be recommended, however, if local neurosurgical expertise and interest are available, because a high rate of complications has tempered enthusiasm for use of this technique in many centers. Specific criteria to identify patients with ALF who are unlikely to recover spontaneously without liver transplantation are shown in Table 95-5. The challenge in managing patients with ALF is to avoid unnecessary liver transplantation in patients who will recover spontaneously, while not delaying liver transplantation in patients in whom the only option for survival is liver transplantation. The role of liver-assist devices in managing ALF, either as definitive therapy or as a “bridge to transplantation,” remains an area of active investigation (see Chapter 93).

CHOLESTATIC LIVER DISEASE

PBC and PSC continue to be relatively common indications for liver transplantation in many transplantation centers and have had a key role in the development of disease models for liver disease, with PBC serving as a benchmark for patient and graft survival. Development of the Mayo disease models to predict the course of cholestatic disorders has aided in decision making regarding timely referral for liver transplantation (see Chapters 68 and 89). Patients with PBC and PSC should be referred for liver transplantation evaluation if their Mayo risk scores predict a one-year survival rate of less than 95%. The Mayo model to predict survival in patients with PBC incorporates the serum bilirubin level, albumin level, patient’s age, prothrombin time, and presence of edema, whereas the model to predict survival in patients with PSC includes the serum bilirubin level, patient’s age, serum AST level, variceal bleeding, and serum albumin level. These models, however, do not take into account prominent and frequently disabling complications of cholestatic liver disease, such as pruritus, osteopenia, or, in PSC, recurrent bouts of bacterial cholangitis and have now been effectively superseded by the MELD score, which is used to determine organ allocation. Despite the generally excellent results of liver transplantation for the cholestatic disorders, concern has increased that they may recur in the graft.

1601

1602

Section IX  Liver Biliary stricturing similar to that in the native diseased liver can be identified in a minority of patients following liver transplantation for PSC and may represent recurrent disease.5 Differentiation of recurrent disease from other important causes of graft injury such as chronic rejection or ischemia may be difficult. Recurrent PSC results in nonanastomotic stricturing of the intrahepatic biliary tree. Although some improvement in symptoms can be obtained by balloon dilation and stent placement, long-term graft viability is reduced. Graft loss caused by recurrent PBC appears to be less frequent than that for PSC. Management of recurrent PBC includes excluding other causes of hepatic dysfunction. Primary immunosuppression with tacrolimus has been implicated in recurrence of PBC by some, but not all, investigators. A controversial issue is whether colectomy in transplant recipients with PSC and associated inflammatory bowel disease reduces the risk of recurrent PSC.60

HEPATIC MALIGNANCY

HCC is the most common primary hepatic malignancy and results in more than 600,000 deaths annually worldwide, usually in patients with underlying cirrhosis. A notable exception is chronic HBV infection, in which HCC can arise in the absence of cirrhosis (see Chapter 94). Tumors with diameters of less than 2 cm discovered incidentally in the explanted liver typically do not have an adverse effect on patient survival. The likelihood of tumor recurrence increases markedly, however, with greater tumor burden, vascular invasion, and the presence of multiple lesions. Liver transplantation is the most definitive treatment of HCC; indeed, 26% of patients who received a liver allograft between 2002 and 2007 had HCC, an observation that reflects the frequency of HCC in cirrhotic patients and the awarding of extra priority to patients with this indication for liver transplantation.61 Improvements in outcome of liver transplantation for HCC in the 2000s are almost entirely attributable to better patient selection rather than improved surgery or adjuvant therapy.62 The preoperative metastatic workup should include a bone scan and chest CT in addition to abdominal imaging. Portal vein occlusion in a patient with HCC is regarded as evidence of metastatic spread and precludes liver transplantation. On the basis of the seminal experience reported by Mazzaferro and colleagues from Milan, generally accepted criteria for liver transplantation in patients with HCC have included a tumor diameter of less than 5 cm, if the tumor is solitary, or no more than three lesions, with the diameter of the largest lesion measuring no greater than 3 cm.63 Survival rates comparable to those for transplantation for decompensated cirrhosis in the absence of complicating HCC (75% at four years) have been reported. With the growing success of liver transplantation for HCC, the Milan criteria have been criticized as being excessively restrictive by excluding many patients who otherwise would have done well with a low risk of recurrence after liver transplantation.62 Various expanded criteria have been proposed to extend the limits of tumor size and number while preserving patient survival rates.64,65 After initial adoption of the MELD score for organ allocation, increased priority was given to patients with HCC who met the Milan criteria, in recognition of the potential for cure of HCC by liver transplantation and the concern that a protracted wait for liver transplantation could result in an increase in tumor burden and metastatic spread. Adoption of the MELD score has resulted in proportionally more patients with HCC undergoing liver transplantation.61 In addition, waiting times for patients with HCC to receive a

deceased donor organ have decreased significantly, and the number of patients dropping out from the waiting list because of advanced-stage disease has also decreased. In the most recent modification of allocation policies, fewer additional MELD score points have been added for the diagnosis of HCC because of the relatively good short-term prognosis of patients with a small HCC.61 Concern has been expressed that cirrhotic recipients without HCC are now disadvantaged because of the preference given to patients with HCC. Several adjuvant interventions have been reported in patients transplanted for HCC (see Chapter 94).17 Recurrent tumor occurs frequently in the graft, and the rationale for adjuvant therapy has been to eliminate micrometastatic disease that typically is disseminated via the vascular system. Conventional systemic chemotherapy administered perioperatively as well as for varying durations before and after liver transplantation, and usually incorporating doxorubicin, is of uncertain benefit. A potentially important strategy for expanding the criteria for liver transplantation in patients with HCC is to downstage the tumor with the use of locoregional therapy so that the Milan criteria are met; whether this approach ultimately improves patient survival remains to be determined.65 For example, TACE administered directly into the tumor arterial supply is often employed to reduce tumor burden during the often protracted wait for liver transplantation. This intervention can be hazardous in patients with decompensated cirrhosis, and its benefit in patients with favorable tumor characteristics remains to be determined. Radiofrequency ablation is used increasingly to manage HCC. Confounding the management of the liver transplantation candidate with HCC, however, is the frequent observation that the tumor burden in the explant is significantly underestimated by preoperative imaging studies. Despite these caveats, a subset of patients with HCC can be cured by liver transplantation and would not have tolerated surgical resection of the tumor because of associated cirrhosis. Oral therapy of HCC with sorafenib will undoubtedly be expanded into liver transplantation populations. The fibrolamellar variant of HCC presents in younger adults without underlying cirrhosis and, as a result, often manifests only when the tumor burden is already large. Extensive resection can be tolerated because cirrhosis is absent. Liver transplantation may be performed in patients who have recurrent tumor after resection. Tumor recurrence after liver transplantation may be relatively indolent and, although not as infrequent as was once thought, survival rates are acceptable.66 Hepatoblastoma is a rare pediatric tumor that also occurs in the absence of underlying parenchymal liver disease. Initial management consists of sur­ gical resection; adjuvant chemotherapy is indicated for metastatic disease. Liver transplantation is an option when the tumor cannot be resected (see Chapter 94). Cholangiocarcinoma remains the only major primary hepatic tumor for which a definitive role for liver transplantation has been difficult to establish. The results of liver transplantation for cholangiocarcinoma diagnosed preoperatively have been so poor that its presence has been regarded as a contraindication to liver transplantation, and even tumors discovered only incidentally in the explant have a high recurrence rate. A subset of patients with a hilar location of the tumor and absence of nodal involvement have been reported to have good five-year survival rates. The extent of cholangiocarcinoma frequently is more extensive than suspected on pre–liver transplantation imaging; often there is local, lymphatic, and perineural spread. The addition of en bloc pancreaticoduodenectomy has not

Chapter 95  Liver Transplantation resulted in improved survival after liver transplantation. Newer approaches to treatment have included preoperative irradiation and chemotherapy, and careful intraoperative tumor staging followed by liver transplantation, with encouraging preliminary results (see Chapter 69).

METABOLIC DISORDERS

Patients with congenital hepatic enzyme deficiencies and other inborn errors of metabolism may be cured by liver transplantation67 (see Chapters 74 to 76). Metabolic disorders potentially cured by liver transplantation fall into two broad categories: diseases dominated clinically by obvious hepatocellular disease (e.g., Wilson disease, hemochro­ matosis) and those without any clinical evidence of liver disease (e.g., primary hyperoxaluria, familial hypercholesterolemia). Although metabolic disorders are most prominent as indications for liver transplantation in the pediatric population, important adult diseases managed by liver transplantation include Wilson disease and hemochroma­ tosis. Substantial neurologic improvement can occur following liver transplantation for Wilson disease in patients who present with decompensated cirrhosis with neurologic involvement. A Wilsonian crisis with severe hemolysis is an indication for urgent liver transplantation; chelation therapy is ineffective in such cases. Hemochromatosis has been associated with poorer outcomes following liver transplantation than have other forms of cirrhosis because of increased rates of adverse cardiac and infectious outcomes. Ongoing studies will clarify whether iron depletion before liver transplantation improves survival rates after liver transplantation. Iron reaccumulation in the grafts of patients transplanted with hemochromatosis is a theoretical concern, and continued iron depletion is not typically required.68 Liver transplantation also has been performed in patients with a variety of systemic disorders, including adult polycystic disease, and as a curative procedure in combination with renal transplantation for primary hyperoxaluria, in which end-organ damage is confined to the kidney but the metabolic defect is hepatic. Liver transplantation has been successful in arresting manifestations of familial amyloid polyneuropathy, with the explant, which is the source of the abnormal protein, available for use in a “domino” fashion in an older recipient who will not live long enough for neurologic injury to develop.69 The biliary cirrhosis associated with cystic fibrosis also has been managed successfully with liver transplantation, although patients remain at risk of infectious and other complications of this systemic disorder.

NONALCOHOLIC FATTY LIVER DISEASE

NAFLD is now recognized as a major cause of chronic liver disease, including cirrhosis and HCC, and is implicated in many cases of cirrhosis (formerly termed cryptogenic) (see Chapter 85). Many of the key precipitants of NAFLD— obesity, hyperlipidemia, and diabetes mellitus—are exacerbated by the post-transplantation immunosuppressive regimen.70 Recurrence of NAFLD post–liver trans­ plantation causes graft injury, although graft loss does not typically result. De novo NAFLD after liver transplantation has also been described. In the absence of any specific therapy for NAFLD, therapeutic efforts after liver transplantation should center on weight control, optimal diabetic management, and use of a lipid-lowering regimen.

VASCULAR DISORDERS

Budd-Chiari syndrome is characterized by hepatic venous outflow obstruction with a presentation that often mimics

decompensated cirrhosis (see Chapter 83).71 Important associations are myeloproliferative disorders, hypercoagulable states, and vena caval webs. Medical approaches to management often are disappointing and fail to retard progression to liver failure and death. Examination of liver biopsy specimens may be helpful in determining whether the therapeutic approach should be decompression with a portosystemic shunt or liver transplantation. Good longterm results have been described in patients who undergo prompt TIPS or portosystemic shunt surgery, but patients with advanced fibrosis on a liver biopsy specimen should undergo liver trans­plantation. Although many patients have an underlying myeloproliferative disorder, an accelerated progression to leukemia or bone marrow failure is not typically observed after liver transplantation. Long-term anticoagulation is continued in liver transplant recipients transplanted for Budd-Chiari syndrome. Sinusoidal obstruction syndrome (SOS) is a similar disorder manifested by necrosis of zone 3 hepatocytes and fibrous obliteration of the central venule lumen. Most commonly seen after bone marrow transplantation (BMT), SOS may lead to hepatic failure and death in up to 25% of patients despite an otherwise successful BMT. Although experience with liver transplantation for hepatic complications of BMT is limited,72 it appears to be the only inter­ vention that consistently alters the course of advanced SOS. Similarly, liver transplantation has been shown to be effective in the management of severe post-BMT graft-versus-host disease with predominantly hepatic involvement (see also Chapter 34). Hypocoagulable (e.g., hemophilia A and B) as well as hypercoagulable (e.g., protein C and S deficiencies) hematologic disorders have been cured with liver transplantation.

AUTOIMMUNE HEPATITIS

Failure of immunosuppressive therapy to arrest progression of severe autoimmune hepatitis with the development of hepatic decompensation is an indication for liver transplantation (see Chapter 88).73 The presence of human leukocyte antigen (HLA)-DR3 is associated with a lower likelihood of a therapeutic response to immunosuppressive therapy in patients with autoimmune hepatitis. Excellent long-term survival is usual after liver transplantation, although the autoimmune dia­thesis may result in higher rates of acute cellular rejection. In addition, recurrent autoimmune hepatitis has been recognized increasingly and may require higher maintenance doses of immunosuppression. Graft survival is generally not diminished by recurrent autoimmune hepatitis.5 Recurrent disease mimics the features of disease in the native liver, is associated with hypergammaglobulinemia and autoanti­bodies, and is generally responsive to glucocorticoids.

OTHER INDICATIONS

A variety of other diagnoses have been reported as indications for liver transplantation (see Table 95-1). Adult polycystic disease with marked abdominal distention resulting from multiple hepatic cysts that are not amenable to resection has been treated successfully by liver transplantation. If chronic kidney disease is present, a combined liver-renal transplantation is indicated. Cerebral imaging is indicated to exclude intracranial aneurysms, which are a feature of this syndrome (see Chapter 94). Liver transplantation also is indicated in cases of multiple adenomas associated with glycogen storage disease and not only eliminates the risk of progression to HCC but also corrects the underlying metabolic disease (see Chapter 76). Diseases with multiorgan involvement for which liver transplantation has been per-

1603

1604

Section IX  Liver formed include Alagille syndrome, amyloidosis, and sarcoidosis (see Chapters 35 and 62).

Superior view II

VIII

VII

III

SURGICAL ASPECTS OF LIVER TRANSPLANTATION Once a potential deceased organ donor is identified, the local organ procurement organization coordinates harvesting and supplies pertinent donor medical information to centers with suitable potential recipients listed with UNOS. In contrast with other types of organ transplants, including kidney and bone marrow transplants, absence of HLA compatibility does not appear to affect liver graft survival, and donor-recipient matching is based only on ABO blood compatibility and physical characteristics such as the recipient’s weight. In critically ill recipients, an ABO-incompatible organ may be implanted, with the recognition that graft survival may be diminished.74 In addition to screening serologic studies and routine liver biochemical testing, particular attention is paid to the donor’s medical history, including cardiovascular instability and need for pressor support before determination of brain death. With the critical shortage of deceased organ donors, expansion of the donor pool has included acceptance of donors ages 70 years and older. As noted earlier, however, use of older donors in recipients with HCV infection may lead to more severe HCV recurrence, potentially limiting the use of older donors in at least some recipients. The typical donor has had a catastrophic head injury or an intracerebral bleed with brain death but without multisystem organ failure. Electrolyte imbalance and hepatic steatosis in the donor are particular concerns because they are predictors of subsequent graft nonfunction. Analogous to the MELD score, a “donor risk index” has been derived to assess the likelihood of good graft function.75 Key adverse factors include older donor age (especially >60 years of age), use of a split or partial graft, and a non–heart-beating donor, from which the organs are harvested after the donor’s cardiac output ceases, in contrast with the more typical deceased donation in which the organs are harvested prior to car­ diovascular collapse. Use of non–heart-beating donors is associated with reduced rates of long-term graft survival and an increased risk of biliary complications and correlates with the duration of “warm ischemia” after cardiovascular collapse and before retrieval of the organ. The harvesting team makes a visual and, if necessary, histologic assessment of the donor organ. Particular note is made during the harvesting procedure of anatomic variants in the hepatic arterial supply that need to be preserved to ensure graft viability. Once the circulation is interrupted, the organ is rapidly infused with cold preservation solution. Donor iliac arteries and veins are also retrieved in case vascular grafting is required. After its arrival at the recipient institution, further vascular dissection, with arterial reconstruction if necessary, is performed before implantation. Splitting cadaveric donor livers either in situ during harvesting or ex vivo on return to the transplantation center allows two recipients to receive portions of the same hepatic allograft, if graft volume and quality are sufficient. An adult cadaveric liver is divided into two functioning grafts. The left lateral segment (segments 2 and 3) is used for a pediatric recipient, and the right trisegment (segments 4 to 8) is used for an adult recipient. Acceptable graft and patient survival rates can be obtained with split grafts, although high-risk unstable recipients may have poorer outcomes with this

V

VI

IV

Inferior view V

IV III

VI

VII

I

II

Figure 95-5.  Segmental anatomy of the liver in the superior and inferior views. Segment VIII is visible only on the superior view, and segment I (caudate lobe) is visible only on the inferior view. (From Keeffe EB. Liver transplantation: Current status and novel approaches to liver replacement. Gastroenterology 2001; 120:749-62, with permission.)

technique. Figure 95-5 shows the segmental anatomy of the liver, which forms the basis of dissection for both split and living-donor liver transplantation.

NATIVE HEPATECTOMY

Removal of the native liver is the most technically challenging part of deceased-donor liver transplantation. Previous abdominal surgery, especially a portosystemic shunt, and severe portal hypertension add to the complexity of hepatectomy. Hepatectomy is technically easier, however, after placement of a TIPS than after a surgical portosystemic shunt. Hilar dissection is performed to access the major hepatic vessels and devascularize the liver. Clamping of the portal vein during hepatectomy and liver implantation results in increased bleeding during dissection, mesenteric congestion, and production of lactate, whereas clamping of the inferior vena cava aggravates venous stasis and causes renal hypertension, with diminished venous return to the heart. To circumvent these problems, venovenous bypass is achieved by cannulation of the portal vein and inferior vena cava via the femoral vein and return of blood via the axillary vein to the right side of the heart. This technique is commonly performed in adults and older pediatric recipients. In some recipients, only a suprahepatic anastomosis to the vena cava is performed, and the inferior vena cava is ligated below the graft. This is called the piggyback technique, in contrast to the more usual circumstance in which anastomosis to the vena cava is performed above and below the graft. The piggyback technique may be applicable if uninterrupted caval flow during liver transplantation is particularly beneficial, as in a recipient with cardiac instability, if a prior portosystemic shunt obviates the need for portal bypass, or if the recipient is a pediatric patient in whom venovenous bypass may not be possible. The portal venous anastomosis

Chapter 95  Liver Transplantation is performed after portal bypass is terminated and is followed by the hepatic arterial anastomosis. The bile duct anastomosis is then fashioned directly, “duct-to-duct,” with or without a T-tube. Hepaticojejunostomy is the preferred anastomosis if there is intrinsic bile duct disease such as PSC or a major discrepancy in donor and recipient bile duct diameters. Microscopic surgical techniques facilitate the donor-recipient biliary and vascular anastomoses. Vascular anatomic anomalies increase the complexity of surgery further. In the past, a direct duct-to-duct anastomosis was typically stented by placement of a T-tube, with the added advantage of easy assessment of bile flow and quality and potential access for cholangiography postoperatively. The risk of a bile leak during sub­sequent removal of the T-tube, however, has led many transplantation programs to abandon routine placement of a T-tube. The use of a live donor involves implantation of only a portion of the liver and is even more technically challenging than use of a whole cadaveric organ (see later). Auxiliary cadaveric liver transplantation is the placement of a graft without removal of the native liver. This technique usually has been performed in critically ill patients such as those with acute liver failure who are too unstable to tolerate native hepactectomy. Irrespective of the type of graft used, after the anastomoses are complete, the newly implanted graft is reperfused with restoration of normal blood flow. The resulting release of vasoactive agents from pooled blood in the lower half of the body can lead to lethal cardiovascular instability and tachyarrhythmias. Prompt bile production should occur if graft function is adequate. Hyperacute rejection is rare but devastating after liver transplantation and leads to rapid graft necrosis within hours and the need for urgent retransplantation.

LIVE-DONOR LIVER TRANSPLANTATION

A major surgical advance has been the extension of LDLT from pediatric recipients to adult recipients, although debate continues about the role of LDLT in adult recipients because of the magnitude of the risk to the donor in light of the large volume of donor liver required.76 The potential donor must be a healthy adult, typically a family member or close friend of the recipient, who volunteers to be evaluated. A series of checks and balances is necessary to ensure that the potential donor undergoes an adequate medical assessment and is not proceeding only as a result of pressure from the patient or family. It is crucial that the potential recipient not be privy to details of the potential donor’s evaluation. In most centers, a hepatologist not involved in the care of the recipient performs an assessment of the donor. Often an independent advocate is also appointed to safeguard the donor’s interests. At each stage of the process, the potential donor is given the opportunity to withdraw from consideration.77 Preoperative evaluation of the donor is best performed in four stages over a period of one to three months; the more invasive testing such as liver biopsy is undertaken later in the evaluation (Table 95-6). After undergoing complete evaluation, only a relatively small proportion of potential donors are deemed satisfactory candidates. One consequence of the evaluation of many potential donors has been the recognition that anatomic abnormalities of the biliary and vascular system and unsuspected abnormalities on liver biopsy specimens are common in apparently healthy persons. Right lobes (segments 5 to 8), extended right grafts (segments 4 to 8), or left hepatic grafts (segments 2 to 4) have been used successfully in adult-to-adult LDLT. Adult LDLT

Table 95-6  Protocol for Evaluation of Potential Living-Related Donors Stage 1

Stage 2 Stage 3 Stage 4

Complete history and physical examination Liver biochemical tests, blood chemistries, hematology, coagulation profile, urinalysis, alpha fetoprotein, carcinoembryonic antigen, and serologic tests for hepatitis A, B, and C, cytomegalovirus, Epstein-Barr virus, and human immunodeficiency virus Abdominal ultrasound examination, chest film Complete psychiatric and social evaluation Computed tomography (CT) of the abdomen Pulmonary function tests, echocardiography Liver biopsy Celiac and superior mesenteric angiography with portal phase* Magnetic resonance cholangiogram Informed consent

*Increasingly, CT angiography is done, instead of standard angiography in stage 2. Adapted from Ghobrial RM, Amersi F, Busuttil RW. Surgical advances in liver transplantation. Living related and split donors. Clin Liver Dis 2000; 4:553-65, with permission.

provides obvious advantages to the recipient, including reduction in mortality rates for patients awaiting a donor organ.78 An expected reduction in the risk of graft rejection because of receipt of a graft from a relative has been not realized, and concern has been raised that recurrence of HCV infection may be accelerated. The overriding concern about LDLT is the short- and long-term consequences to the donor, including the risk of immediate perioperative morbidity and mortality, and financial losses resulting from time lost from work, possible uninsurability in the future, and a lack of long-term follow-up data to ensure that hepatic resection and subsequent regeneration do not result in biliary or other abnormalities.

IMMUNOSUPPRESSION Administration of immunosuppressive agents following liver transplantation is divided into induction (initial) and maintenance (long-term) phases. In addition, episodes of acute cellular and chronic ductopenic rejection require therapy. A wide array of immunosuppressive agents are currently used.79 In practice, new immunosuppressive agents are introduced for use in renal transplantation before they are applied to liver transplantation. Therefore, the need for effective antirejection regimens in other areas of solid organ transplantation is great (see also Chapter 34). The primary goal of immunosuppression is to prevent graft rejection and loss; a secondary goal is to avoid the adverse consequences of the antirejection regimen.80 A list of the commonly used immununosuppressive agents, routes of administration, methods of monitoring, and common adverse effects is shown in Table 95-7. Common drug-drug interactions are shown in Table 95-8. The calcineurin inhibitors cyclosporine and tacrolimus are the basis for a majority of induction and maintenance immunosuppressive regimens. Both agents have substantial toxicity. Tacrolimus is now favored over cyclosporine for primary immunosuppression. In addition, patients may be converted from cyclosporine to tacrolimus following glucocorticoidor OKT3-refractory rejection, late rejection (>6 months post–liver transplantation), histologically diagnosed chronic

1605

1606

Section IX  Liver Table 95-7  Immunosuppressive Agents Used in Liver Transplantation AGENT

MODE OF ACTION

MONITORING

TOXIC EFFECTS

Cyclosporine

Calcineurin inhibitor: suppresses IL-2–dependent T-cell proliferation Same as cyclosporine Cytokine inhibitor (IL-1, IL-2, IL-6, TNF, and IFN-γ)

Blood level

Renal, neurologic, hyperlipidemia, hypertension, hirsutism Renal, neurologic, diabetes mellitus Hypertension, diabetes mellitus, obesity, osteoporosis, infection, depression, psychosis Bone marrow suppression, hepatotoxicity Diarrhea, bone marrow suppression

Tacrolimus Prednisone Azathioprine Mycophenolate mofetil Sirolimus OKT3 IL-2 receptor blocker

Blood level None

Inhibition of T- and B-cell proliferation by interfering White blood cell count with purine synthesis Selective inhibition of T- and B-cell proliferation by White blood cell count interfering with purine synthesis Inhibition of late T-cell functions Blood level Blocking of T cell CD3 receptor, preventing stimulation by antigen Competitive inhibition of IL-2 receptor on activated lymphocytes

CD3+ count None

Neutropenia, thrombocytopenia, hyperlipidemia Cytokine release syndrome, pulmonary edema, increased risk of infections Hypersensitivity reactions with basiliximab

IFN, interferon; IL, interleukin; TNF, tumor necrosis factor. Adapted from Everson GT, Karn I. Immediate postoperative care. In: Maddrey WC, Schiff ER, Sorrell MF, editors. Transplantation of the liver. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2001. p 131.

Table 95-8  Clinically Relevant Drug Interactions with Immunosuppressive Drugs Drugs that increase blood levels of cyclosporine and tacrolimus: Antifungals: fluconazole, ketoconazole, itraconazole Antibiotics: clarithromycin, erythromycin Calcium channel blockers: diltiazem, verapamil Others: allopurinol, bromocriptine, metoclopramide Drugs that decrease blood levels of cyclosporine and tacrolimus: Anticonvulsants: phenobarbitone, phenytoin Antibiotics: nafcillin, rifampin Drugs that increase nephrotoxicity of cyclosporine and tacrolimus: Gentamicin, ketoconazole, nonsteroidal anti-inflammatory drugs Drugs that interact with mycophenolate mofetil: Acyclovir, ganciclovir (increase levels) Antacids (inhibit absorption) Cholestyramine (inhibits absorption) Drugs that interact with azathioprine: Allopurinol, angiotensin-converting enzyme (ACE) inhibitors (increase hematologic side effects) Warfarin (decreased anticoagulant effect)

rejection, severe cholestasis, intestinal malabsorption of cyclosporine, or cyclosporine toxicity (e.g., hirsutism, gin­ givitis, severe hypertension). When used as rescue therapy for chronic rejection, tacrolimus is less effective once the serum bilirubin levels rise above 10 mg/dL, underscoring the importance of early recognition of chronic rejection. Although implicated in hepatic artery thrombosis as well as delayed wound healing and infections, sirolimus, an inhibitor of the protein mTOR (mammalian target of rapamycin), has been used increasingly in liver transplantation as a calcineurin-sparing strategy and also in patients transplanted for HCC to reduce tumor recurrence.81 Another newer agent is basiliximab, a monoclonal antibody directed against CD25, which may be an alternative to glucocorticoids as an induction agent in liver transplantation.82 Considerable differences exist among transplantation centers in the rate at which the level of induction immunosuppression is reduced to avoid toxicity and lessen the risk of recurrent disease. Generally, the strategy includes tapering and in some cases discontinuing maintenance glucocorticoids.83

POSTOPERATIVE COURSE INITIAL PHASE TO DISCHARGE FROM HOSPITAL

Because of the complexity of liver transplantation and the often markedly decompensated state of liver transplant recipients, invasive monitoring with arterial and pulmonary venous lines is necessary in the first few postoperative days. If a T-tube has been placed, dark copious bile provides evidence of satisfactory graft function. The patient’s overall status, including neurologic recovery from anesthesia, urinary output, and cardiovascular stability, also reflect graft function. Routine antimicrobial prophylaxis includes bowel decontamination with oral nonabsorbable antibiotics, perioperative systemic broad-spectrum antibiotics, antifungal agents, and ganciclovir to prevent cytomegalovirus (CMV) infection. Markedly abnormal liver biochemical test levels are typical during the initial 48 to 72 postoperative hours and reflect a number of insults to the graft, including ischemia following harvesting, during preservation, and on subsequent reperfusion. The overall trend in serum aminotransferase levels should be downward, with a corresponding improvement in coagulopathy and a falling serum bilirubin level. Thrombocytopenia in the immediate postoperative period reflects a variety of processes, including residual splenomegaly, medications, and, importantly, reduced graft function. Worrisome clinical features include scanty, pale bile if a T-tube has been used, metabolic acidosis, depressed mentation, and the continued need for pressor support with worsening liver biochemical test levels. Hepatic artery thrombosis needs to be excluded promptly by Doppler ultrasound and is an indication for urgent retransplantation. Hepatic artery thrombosis is more common in pediatric recipients because of the smaller size of the vessels. Antiplatelet therapy is now administered routinely to reduce the risk of hepatic artery thrombosis.84 Primary nonfunction of the graft is also an indication for urgent retransplantation and is suggested by the absence of bile production in the first several hours after transplantation, as well as an unstable overall clinical status. Donor characteristics that are associated with an increased likelihood of primary nonfunction include marked hepatic steatosis and profound hyponatremia. If graft function is adequate, however, pressor support can be

Chapter 95  Liver Transplantation gested by a return of liver biochemical test levels toward normal values. For the occasional patient with presumed acute cellular rejection who fails to have a biochemical response to glucocorticoids, immunosuppression may be enhanced with the monoclonal antibody OKT3. Liver biopsy should be repeated first to confirm a lack of histologic response before more intensive therapy is initiated and to exclude other important causes of graft dysfunction such as ischemia. The ability of recurrent HCV infection to mimic virtually all the features of acute cellular rejection histologically has led to a reevaluation of the need to treat apparent acute cellular rejection aggressively under all circumstances. Routine (protocol) liver biopsies also have fallen out of favor because histologic evidence of acute cellular rejection can be noted in the absence of worsening graft function with no apparent clinical significance. A cholangiogram should be obtained routinely prior to clamping the T-tube, if one is present; a nonsustained rise in liver biochemical test levels can be anticipated as a result. The timing of various infectious complications following liver transplantation is shown in Figure 95-7.

tapered and ventilator weaning parameters monitored to facilitate extubation, although the recipient who is markedly debilitated from advanced cirrhosis may require several days of ventilatory support. Poor graft function and renal insufficiency also can impede weaning from the ventilator. Within the first week after liver transplantation, liver biochemical test levels should steadily improve as ischemia and reperfusion injury resolve. Acute cellular rejection becomes an important and frequent cause of graft dysfunction at one week and beyond and is suggested by a rise in serum aminotransferase, alkaline phosphatase, and bilirubin levels. Because the biochemical features are nonspecific, the threshold for performing a liver biopsy to evaluate other diagnostic possibilities, which include slowly resolving reperfusion injury, biliary tract obstruction, and cholestasis related to sepsis, should be low. Histologic findings characteristic of acute cellular rejection are bile duct injury, portal inflammation with eosinophils, and, with more severe injury, endotheliitis (Fig. 95-6). High doses of glucocorticoids (1000 mg of methylprednisolone or the equivalent), followed by a taper (200 to 20 mg/day) extending over several days, constitute first-line therapy. A response is sug-

A

B

Figure 95-6.  Histopathology of acute cellular rejection of a liver graft. A, The portal tract shows a lymphocytic and plasma cell infiltrate that spills over into the periportal hepatocytes and bile duct. B, The central vein shows attachment of lymphocytes to the endothelium (endotheliitis). (From Cotran RS, Kumar V, Collins T, editors. Robbins’ pathologic basis of disease. 6th ed. CD-ROM. Philadelphia: WB Saunders; 1999, with permission.)

Bacterial infections Herpetic stomatitis Mucocutaneous candidiasis CMV hepatitis or syndrome EBV, VZV, adenovirus infections Pneumocystis jiroveci infection Aspergillosis or mucormycosis Nocardiosis, listeriosis, tuberculosis Recurrent hepatitis B or C 0

20

40

60

80

Days post transplantation

100

Figure 95-7.  Time course of various infectious complications in liver transplant recipients. CMV, cytomegalovirus; EBV, Epstein-Barr virus; VZV, varicella-zoster virus. (From Everson GT, Kam I. Immediate postoperative care. In: Maddrey WC, Schiff ER, Sorrell MF, editors. Transplantation of the liver. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2001. p 131.)

1607

1608

Section IX  Liver Table 95-9  Medical Complications in the Immediate Post-Transplantation Period Infections Bacterial Viral Cytomegalovirus Epstein-Barr virus Fungal Aspergillosis, mucormycosis Candidiasis, torulopsosis Pneumocystis jiroveci pneumonia Respiratory complications Acute respiratory distress syndrome Hepatopulmonary syndrome Pneumonia Portopulmonary hypertension Pulmonary edema Acute kidney injury Cardiovascular disease Cardiomyopathy Hemochromatosis Idiopathic hypertrophic subaortic stenosis Hypertension Myocardial ischemia Valvular heart disease Neurologic complications Central nervous system hemorrhage Central pontine myelinolysis Ischemic events Seizures Coagulopathy Disseminated intravascular coagulation Thrombocytopenia Diabetes mellitus From Everson GT, Karn I. Immediate postoperative care. In: Maddrey WC, Schiff ER, Sorrell MF, editors. Transplantation of the Liver. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2001, with permission.

A number of other important medical issues are common in the first weeks following liver transplantation (Table 95-9). Neurologic dysfunction can present as an acute confusional state or seizures, and the differential diagnosis includes the lingering effects of hepatic encephalopathy, electrolyte imbalance, poor graft function, sepsis, uremia, and side effects of medications. A particular concern is the development of neurologic toxicity caused by the major immunosuppressive agents. Overly rapid correction of hyponatremia perioperatively has been implicated in the genesis of central pontine myelinolysis, with evidence of demyelination demonstrable on magnetic resonance imaging. Management includes correcting the electrolyte imbalance if present and reducing the dose of calcineurin inhibitor, which can be facilitated by the use of myco­ phenolate mofetil.85 Diabetes mellitus, which is common in persons with cirrhosis, can occur for the first time in the postoperative period and usually requires insulin to control. HCV infection further increases the risk of post–liver transplantation diabetes mellitus.86 Renal impairment post–liver transplantation may reflect a number of factors, including slowly resolving pre–liver transplantation hepatorenal syndrome or renal failure of other causes, intraoperative hypotension resulting in acute tubular necrosis, and, importantly, the nephrotoxic effects of cyclosporine and tacrolimus, which cause renal afferent arteriolar vasoconstriction and a corresponding reduction in glomerular filtration rate. Adjunctive therapy with mycophenolate mofetil allows a reduction in the doses of cyclosporine and tacrolimus while providing adequate immunosuppression. Short-term hemo-

dialysis may be necessary until renal function improves. In the first three to four weeks post–liver transplantation, infections are typically bacterial and related to surgical complications such as intra-abdominal bleeding, bile leak, or wound infection.

FOLLOWING DISCHARGE FROM HOSPITAL

If the initial postoperative course has been smooth, planning for discharge is possible by the second week post–liver transplantation. Recovery often is more protracted, particularly in debilitated recipients. Once discharged, patients are seen at frequent intervals during the first postoperative month. The liver biochemical test levels should fall to normal values within a few weeks following liver transplantation. Further graft dysfunction is an indication for prompt liver biopsy because acute cellular rejection remains a concern during this time; in addition, CMV becomes an important consideration three or more weeks posttransplantation.87 Histologic features suggestive of CMV hepatitis include “owl’s eye” inclusion bodies in the hepatocytes as well as neutrophilic abscesses with focal necrosis of the parenchyma (see Chapter 81). Recipients who have had no prior exposure to CMV are at high risk of CMV infection, particularly if they receive a graft from a CMVseropositive donor, and are candidates for more intensive prophylaxis. A distinction is made between CMV viremia and CMV disease with systemic manifestations such as diarrhea, because the viremia does not invariably imply disease. Early recurrence of HCV infection also may become apparent, and, as noted earlier, it is crucial to recognize that many of the histologic features of acute cellular rejection, such as bile duct inflammation and endotheliitis, are mimicked by HCV infection (Table 95-10). In addition to graft hepatitis, other manifestations of de novo CMV infection include pneumonitis and diarrhea. Reactivation of CMV in a previously infected recipient tends to be less clinically severe than de novo infection. The diagnosis of CMV infection is confirmed by culture of tissue or blood, but not by isolation from urine. Many transplantation centers now have access to rapid tissue culture techniques that use indirect immunofluorescence and that allow prompt diagnosis. High-dose intravenous ganciclovir is highly effective in the treatment of CMV infection, but viral resistance has been described. Other therapies include a CMV hyperimmune globulin and foscarnet. Not only is CMV infection an important cause of morbidity and mor­ tality in liver transplant recipients, but it also has been implicated in other complications, notably chronic graft rejection and severe recurrent HCV infection. If a liver biopsy specimen shows features suggestive of biliary obstruction or if graft dysfunction is associated with clinical features of cholangitis such as fever and abdominal pain, a cholangiogram should be obtained. Increasingly, magnetic resonance cholangiopancreatography (MRCP) is the initial method of choice because of its noninvasive nature and high degree of accuracy irrespective of the type of biliary anastomosis.88 Before the advent of MRCP, a cho­ langiogram was generally obtained through the T-tube if present, by endoscopic retrograde cholangiopancreatography (ERCP) if a T-tube was not present, or by percutaneous transhepatic cholangiography if a choledochojejunostomy had been performed. An anastomotic stricture in a chole­ dochocholedochostomy is usually easily managed by endoscopic balloon dilation and temporary internal stenting (see Chapter 70). Surgical intervention is reserved for patients who do not respond to this approach, in which case the anatomy is converted to a Roux-en-Y anastomosis. Anastomotic stricturing also can occur at the site of a cho-

Chapter 95  Liver Transplantation Table 95-10  Histologic Features of Recurrent Hepatitis C Virus Infection versus Acute Cellular Rejection FEATURE

HCV RECURRENCE

REJECTION

Time of onset after liver transplantation Portal inflammation   Lymphocytes   Lymphoid aggregates   Lymphoid follicles   Eosinophils Steatosis Acidophilic bodies Bile ductule damage Atypical features

Any time; onset usually within the first year Most cases Bland, uniform Usually 50% of cases Inconspicuous Often Common About 50% of cases Cholestasis, ballooning degeneration without significant inflammation, marked ductular proliferation mimicking obstruction, granulomas

Usually within the first two months Always Activated Occasionally Rarely Almost always Never Uncommon Common Prominent periportal and lobular necroinflammatory activity without subendothelial venular inflammation

HCV, hepatitis C virus. From Rosen HR, Martin P. Liver transplantation. In: Schiff ER, Sorrell MF, Maddrey WC, editors. Schiff’s Diseases of the liver. 8th ed. Philadelphia: Lippincott-Raven; 1999. p. 1589.

ledochojejunostomy and requires access, usually by a percutaneous approach, to dilate the stenotic area. A critical issue in management is distinguishing anastomotic stricturing from nonanastomotic stricturing caused by ischemia or other insult to the graft. The bile duct in the liver transplant recipient is prone to ischemia because of its relatively tenuous arterial blood supply and the development of a biliary stricture (unless it is obviously anastomotic) may reflect hepatic artery thrombosis. Ischemic stricturing is generally diffuse but can be predominantly hilar. Although temporizing measures such as balloon dilation may have some utility, such efforts are generally futile if hepatic artery thrombosis is present or stricturing is widespread, and retransplantation will be required. Other causes of nonanastomotic stricturing include the use of an ABO-incompatible graft and protracted cold ischemia after harvesting. Biliary stricturing also can be a feature of recurrent PSC. A T-tube, if present, is removed by the sixth postoperative month, and removal is best performed at the transplantation center, because bile leaks are common. If a bile leak occurs, prompt ERCP with nasobiliary drainage or stenting usually allows the tear in the bile duct to heal uneventfully (see Chapter 70). In addition to prophylaxis against CMV infection in the early postoperative months, antibiotics, most frequently trimethoprim-sulfamethoxazole, are prescribed long term to prevent infection with Pneumocystis jiroveci. In patients intolerant of sulfa drugs, alternative options are dapsone tablets or inhaled pentamidine. The duration of prophylaxis varies by program but generally needs to be at least one year following transplantation. Fungal infections pose a major threat to the liver transplant recipient, particularly in the setting of marked debilitation, intensive immunosuppression for rejection, or retransplantation. Major sites of infection are mucocutaneous (oral and esophageal), pulmonary, and intracerebral. Despite prolonged therapy with amphotericin or itraconazole, a fatal outcome is usual. A diagnosis of a brain abscess in a patient with invasive Aspergillus infection implies a dismal prognosis. Superficial skin infections and simple colonization must be distinguished from invasive fungal infections, because topical antifungal agents such as nystatin or clotrimazole can eradicate the former. Similarly, bladder irrigation with amphotericin can cure candidal cystitis without the need for systemic antifungal therapy.

Opportunistic infections remain a concern in the liver transplant recipient during long-term follow-up. Patients need prompt assessment for symptoms of infection. Standard antibiotic therapy is appropriate for communityacquired respiratory infections, but a more extensive workup is indicated when symptoms are unusually severe or an infection fails to resolve rapidly with treatment. Enteric bacteremia may be an initial clue to hepatic artery thrombosis in an otherwise stable recipient. Reactivation of tuberculosis may present in an atypical fashion after liver transplantation. Bronchoscopy or lumbar puncture with cultures may be necessary, as clinically indicated.

LONG-TERM MANAGEMENT GENERAL PREVENTIVE MEDICINE

Satisfactory long-term management of the liver transplant recipient requires cooperation and communication between the primary care physician and the transplantation center. Many of the disorders that impact long-term survival after liver transplantation are common diseases, including systemic hypertension, hyperlipidemia, and diabetes mellitus.89 Regular determination of a complete blood count, electrolytes, liver biochemical test levels, and immunosuppressive drug levels are indicated, and the results should be forwarded to the transplantation center. Generally, the frequency of blood work can be reduced beyond the first postoperative year in recipients with stable graft function. Systemic hypertension is a frequent complication of liver transplantation and is related to calcineurin inhibitor– induced renal vasoconstriction, as well as to the effects of other drugs such as glucocorticoids. Reduction in the level of immunosuppression unfortunately is generally ineffective in ameliorating hypertension. Other contributing factors include mild renal insufficiency, which is frequent after liver transplantation even when absent preoperatively. Initial antihypertensive therapy usually consists of a calcium channel blocker. Angiotensin-converting enzyme inhibitors and potassium-sparing diuretics are relatively contraindicated because of their propensity to accentuate hyperkalemia, which is frequent in liver transplant recipients who often have renal tubular acidosis caused by the calcineurin inhibitor. Because cyclosporine and tacrolimus levels are

1609

1610

Section IX  Liver increased by verapamil and diltiazem, nifedipine is the agent of choice. β-Adrenergic–blocking agents are secondline antihypertensive agents; diuretics are generally avoided because of concern about exacerbating renal insufficiency and electrolyte imbalance in the liver transplant recipient. Furosemide is the diuretic of choice if fluid overload is present. In the minority of patients in whom hypertension is not controlled, a centrally acting agent such as clonidine is introduced. For the occasional patient with intractable hypertension on cyclosporine-based immunosuppression, substitution of tacrolimus for cyclosporine may aid blood pressure control. Both cyclosporine and tacrolimus, however, are nephrotoxic and accentuate renal impairment that may have existed perioperatively. Although acute nephrotoxicity may respond to interruption of or a reduction in the dose of these drugs, chronic renal impairment is usually irreversible, and drastic dose reductions should be avoided for fear of precipitating graft rejection. Cofactors implicated in the progression to advanced chronic kidney disease after liver transplantation include recurrent HCV infection with associated glomerulonephritis as well as diabetes mellitus and systemic hypertension.90 Renal transplantation may be considered in liver transplant recipients who become dialysis dependent after otherwise successful liver transplantation. Osteopenia is a frequent cause of morbidity in liver transplant recipients.91 Although hepatic osteodystrophy is typically associated with the cholestatic liver diseases, it is also common in patients with other forms of cirrhosis. Factors implicated in the pathogenesis of hepatic osteodystrophy include poor nutritional status, immobility, the calciuric effect of many diuretics, hypogonadism, and glucocorticoid use in patients with autoimmune hepatitis. In the initial several months after liver transplantation, osteopenia is accelerated further by high-dose glucocorticoid therapy as well as the other major immunosuppressive agents. Atraumatic fractures may occur in trabecular bone such as vertebrae or ribs. Patients begin to rebuild bone mass after the doses of immunosuppressive drugs are reduced and the patient’s mobility increases. Supplemental calcium and vitamin D are frequently prescribed to patients with symptomatic osteopenia, as is a bisphosphonate. De novo malignancies are increased in frequency following liver transplantation.92 Recipients need ongoing ageappropriate surveillance for common tumors such as breast and colon cancer. Screening for prostatic carcinoma should be performed by yearly digital rectal examination in male liver transplant recipients older than age 40, in conjunction with serum prostate specific antigen (PSA) testing. Screening for colorectal cancer by colonoscopy should also be performed every 3 to 5 years after age 50 in asymptomatic recipients; in patients with a history of PSC and ulcerative colitis, yearly colonoscopy with surveillance mucosal biopsies is recommended.93 In the setting of chronic immunosuppression, screening female transplant recipients older than age 40 for breast cancer by yearly mammography seems appropriate, although the cost effectiveness of this approach is undefined. Other malignancies that are increased in frequency in organ transplant recipients include skin, female genital tract, and perineal cancers; alcoholic patients may be particularly prone to malignancies of the oropharynx (see Chapter 84). Patients should be encouraged to wear sunscreen and have regular examinations by a dermatologist. Post-transplantation lymphoproliferative disorder (PTLD) varies from a low-grade indolent process to an aggressive neoplasm.94 Uncontrolled proliferation of B cells after liver transplantation, typically in response to primary Epstein-

Barr virus infection, can be polyclonal or monoclonal. Pediatric recipients are at particular risk because of the absence of prior Epstein-Barr viral infection. Intensive immunosuppression with OKT3 for severe rejection increases the risk of PTLD, which can present as a mononucleosis-like syndrome, lymphoproliferation, or malignant lymphoma. Clinical features suggestive of PTLD include lymphadenopathy, unexplained fever, and systemic symptoms such as weight loss. After the diagnosis is made histologically by biopsy of involved areas (which can include the liver graft and gastrointestinal tract as well as lymph nodes), therapy includes a reduction in the level of immunosuppression and antiviral therapy with ganciclovir directed against Epstein-Barr virus. Systemic chemotherapy may be required in patients who present with a malignant lymphoma. The higher frequency of PTLD in pediatric recipients has led to surveillance by polymerase chain reaction methodology for Epstein-Barr viremia and reduction in the level of immunosuppression in patients with a positive result before clinical features of PTLD occur. In addition, antiviral prophylaxis is prescribed to high-risk recipients, including those who are seronegative for Epstein-Barr virus and receive a liver from a seropositive donor. Chronic graft rejection is increased in frequency in survivors of PTLD because of the deliberate reduction in the level of immunosuppression, which may be increased cautiously after PTLD is contained. Hyperlipidemia is observed in up to one half of liver transplant recipients and reflects a number of pathogenic factors, including diabetes mellitus, obesity, renal dysfunction, and the use of immunosuppressive agents, especially cyclosporine.95 Pharmacologic therapy is indicated if hypercholesterolemia fails to respond to weight reduction and tight diabetic control. Pravastatin, a 3-hydroxy-3methylglutaryl coenzyme A reductase inhibitor, is well tolerated and efficacious in liver transplant recipients.96 Diabetes mellitus is common in liver transplant recipients and occurs in approximately one third of patients for the first time after transplantation. The pathogenesis is multifactorial; immunosuppressive therapy is a major factor because of the hyperglycemic effects of prednisone, cyclosporine, tacrolimus, azathioprine, and mycophenolate mofetil. HCV infection is also implicated. In most diabetic recipients, therapy with insulin is required. The high frequency of diabetes mellitus following liver transplantation has led to the development of glucocorticoid-sparing immunosuppressive regimens. A related problem is obesity, which is frequent even in liver transplant recipients who were profoundly malnourished preoperatively. Factors responsible for weight gain after liver transplantation include glucocorticoid use, increased caloric intake, and decreased physical activity during recuperation from surgery. Immunosuppression with tacrolimus has been reported to result in less weight gain than occurs with cyclosporine; to a large extent, this difference may reflect the lower glucocorticoid doses used with tacrolimus. Management of obesity in this population includes a reduction in glucocorticoid doses and even complete withdrawal if possible. Use of mycophenolate mofetil may permit maintenance immunosuppression without glucocorticoids.

IMMUNIZATIONS AND ANTIBIOTIC PROPHYLAXIS

Immunization against hepatitis A and B, influenza, pneumococcus, tetanus, and diphtheria is part of the standard pretransplantation evaluation. A substantial proportion of patients may be unable to mount adequate antibody

Chapter 95  Liver Transplantation

WHEN TO CALL THE TRANSPLANTATION CENTER

A number of common symptoms, signs, and laboratory abnormalities warrant a call by the local physician to the transplantation center. These problems include fever, abdominal pain, neurologic symptoms, anticipated surgery, and a possible change in the patient’s immunosuppressive regimen. When an unexplained abnormality of liver biochemical test levels occurs, a complete workup for possible causes is imperative. Although a liver biopsy can be obtained by the patient’s local physician, a local pathologist may be inexperienced in allograft interpretation, and it is critical that the specimen be reviewed at the transplantation center so that appropriate decisions regarding management can be obtained. Additionally, many transplantation programs prefer to perform interventional biliary tract studies because therapeutic intervention is often required, and immediate access to the transplantation team permits more rapid decision making. Any evidence of graft failure needs to be attended to immediately by referral to the transplantation center.

HEPATIC RETRANSPLANTATION

Although improved immunosuppressive regimens have led to a lower rate of graft loss because of chronic rejection, recurrence of the underlying liver disease has been recognized increasingly as a cause of graft failure, as illustrated most strikingly in HCV-infected recipients.97 The rates and severity of recurrent diseases are highly variable and probably related to a complex interplay among host factors (including the underlying liver disease), therapeutic regimens (e.g., immunosuppression, antiviral treatment), and possibly genetic variability of the allograft (perhaps through effects on the nature and magnitude of the inflammatory response within the graft). Understanding the full effect of recurrent disease, especially nonviral disease, on patient and graft survival will require long-term follow-up studies. For example, although the rate of histologic recurrence of

viral hepatitis is greatest in the first year following liver transplantation, recurrent PBC or PSC develops in less than 5% of patients by the first year, whereas more than 20% demonstrate histologic recurrence 10 years after liver transplantation.98 As patients enter their second and third decades following liver transplantation, the number of patients who require retransplantation may deplete the donor pool further. This issue is compounded by the observation that patients who undergo retransplantation experience an approximate 20% overall reduction in the rate of survival but consume an increased amount of resources when compared with primary transplant recip­ ients. On the basis of these considerations, a number of investigators have developed models to predict survival following retransplantation, especially for HCV infection (Table 95-11 and Fig. 95-8). For example, the liver transplant recipient with jaundice, graft failure caused by recurrent HCV infection, and renal failure has a markedly diminished likelihood of surviving retransplantation. Although these models only estimate the probability of

1.0 0.9 0.8 Proportion surviving

responses because of the immunosuppression associated with end-stage liver disease. Vaccines based on live or attenuated microorganisms (e.g., measles, mumps, rubella, oral polio, bacille Calmette-Guérin [BCG], vaccinia) are contraindicated because of the risk of reactivation. Prophylactic antibiotics are recommended for any dental procedure, even basic cleaning.

Low risk

0.7 0.6

Medium risk

0.5 0.4

High risk

0.3 0.2 0.1 0 0

500

1000

1500

2000

2500

Days following retransplantation Figure 95-8.  Survival of 1356 patients undergoing hepatic retransplantation stratified into low-risk, medium-risk, and high-risk groups using prognostic predictive factors listed in Table 95-11 (P < 0.00001 by Wilcoxon rank sum). (From Rosen HR, Madden JP, Martin P. A model to predict survival following hepatic retransplantation. Hepatology 1999; 29: 365-70.)

Table 95-11  Multivariate Models Developed to Predict Survival Following Liver Retransplantation NO. OF PATIENTS 1356 70 418 150 447 207

PROGNOSTIC PREDICTIVE FACTORS

COMMENTS

Recipient’s age, bilirubin and creatinine levels, UNOS status, and cause of graft failure Recipient’s age, UNOS status, inpatient status, serum bilirubin and creatinine levels Recipient’s age, mechanical ventilatory status, serum bilirubin and creatinine levels Recipient’s age group (pediatric vs. adult), mechanical ventilatory status, organ ischemia time >12 hr, serum bilirubin and creatinine levels Interval to retransplantation (better prognosis if within 30 days) Bilirubin and serum creatinine levels

UNOS database used; HCV positivity and donor age are significant predictors by univariate analysis83 King’s College84 University of Pittsburgh85 UCLA86 Mayo Clinic; limited to patients with PBC or PSC87 UNOS database used; limited to HCV-positive patients retransplanted for causes other than primary nonfunction88

HCV, hepatitis C virus; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; UNOS, United Network for Organ Sharing. From Rosen HR. Disease recurrence following liver transplantation. Clin Liver Dis 2000; 4:675-89, with permission.

1611

1612

Section IX  Liver survival for an individual patient and do not take into account the patient’s quality of life, they can be used as adjuncts to clinical judgment. Application of retransplantation to low-risk patients is associated with survival com­ parable to that for primary liver transplantation; determining whether retransplantation is justified in patients with highrisk scores will require prospective studies. In the current era of extreme organ shortage, the utility of primary and repeat transplantation (i.e., duty to promote the best outcome in the aggregate) needs to be considered. An analysis by Burton and colleagues, using the Cobbs-Douglas equation, has demonstrated that for primary liver transplantation, maximal utility is achieved by allocating organs at the highest MELD score (i.e., “sickest first”). For retransplan­ tation, however, maximal utility for HCV and non-HCV diagnoses is achieved at MELD scores of 21 and 24, respectively. Utility starts to decline at MELD scores greater than 28.25 A multicenter U.S. study of 272 patients (73 retransplanted for causes not related to HCV) indicated that oneand three-year survival rates following retransplantation were equivalent in HCV-positive and HCV-negative patients. These data were likely shaped by selection bias, considering that nationally more than one third of patients with HCVrelated allograft failure are not even considered for retransplantation and only half of those reevaluated are even relisted. This study also showed that MELD scores greater than 30 are associated with particularly poor survival following retransplantation.99 Major challenges remain in liver transplantation, including the shortage of donor organs, threat of recurrent disease, and morbidity associated with lifelong therapeutic immunosuppression. Nevertheless, the availability of liver transplantation has transformed the lives of patients with advancing liver disease and their health care providers from an ultimately futile effort to manage the complications of cirrhosis into a life-prolonging and life-enhancing intervention.

KEY REFERENCES

Ahmed A, Keeffe EB. Current indications and contraindications for liver transplantation. Clin Liver Dis 2007; 11:227-47. (Ref 13.) Arjal RR, Burton JR Jr, Villamil F, et al. Review article: The treatment of hepatitis C virus recurrence after liver transplantation. Aliment Pharmacol Ther 2007; 26:127-40. (Ref 3.) Benten D, Sterneck M, Panse J, et al. Low recurrence of preexisting extrahepatic malignancies after liver transplantation. Liver Transpl 2008; 14:789-98. (Ref 29.) Cardenas A, Kelleher T, Chopra S. Review article: Hepatic hydrothorax. Aliment Pharmacol Ther 2004; 20:271-9. (Ref 40.) Fallon MB, Krowka MJ, Brown RS, et al. Impact of hepatopulmonary syndrome on quality of life and survival in liver transplant candidates. Gastroenterology 2008; 135:1168-75. (Ref 37.) Grewal P, Martin P. Pretransplant management of the cirrhotic patient. Clin Liver Dis 2007; 11:431-49. (Ref 14.) Ioannou GN, Perkins JD, Carithers RL Jr. Liver transplantation for hepatocellular carcinoma: Impact of the MELD allocation system and predictors of survival. Gastroenterology 2008; 134:1342-51. (Ref 61.) Kim WR, Biggins SW, Kremers WK, et al. Hyponatremia and mortality among patients on the liver-transplant waiting list. N Engl J Med 2008; 359:1018-26. (Ref 45.) Lopez PM, Villanueva A, Roayaie S, et al. Neoadjuvant therapies for hepatocellular carcinoma before liver transplantation: A critical appraisal. Liver Transpl 2006; 12:1747-54. (Ref 17.) Lucey MR. Liver transplantation for alcoholic liver disease. Clin Liver Dis 2007; 11:283-9. (Ref 48.) McAvoy NC, Kochar N, McKillop G, et al. Prevalence of coronary artery calcification in patients undergoing assessment for orthotopic liver transplantation. Liver Transpl 2008; 14:1725-31. (Ref 32.) O’Leary JG, Lepe R, Davis GL. Indications for liver transplantation. Gastroenterology 2008; 134:1764-76. (Ref 1.) Oo YH, Neuberger J. Recurrence of nonviral diseases. Clin Liver Dis 2007; 11:377-95. (Ref 5.) Pham PT, Pham PC, Rastogi A, et al. Review article: Current management of renal dysfunction in the cirrhotic patient. Aliment Pharmacol Ther 2005; 21:949-61. (Ref 43.) Rosen HR. Transplantation immunology: What the clinician needs to know for immunotherapy. Gastroenterology 2008; 134:1789-801. (Ref 80.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

96 Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine Ellen Kahn and Fredric Daum

CHAPTER OUTLINE Anatomy  1615 Macroscopic Features  1615 Microscopic Features  1617 Embryology  1623 Molecular Regulation of Intestinal Morphogenesis  1623 Specific Structures and Systems  1624 Clinical Implications  1626 Abnormalities in Normal Embryologic Development  1626 Abdominal Wall  1626

ANATOMY MACROSCOPIC FEATURES Small Intestine

The small intestine is a specialized tubular structure within the abdominal cavity in continuity with the stomach proximally and the colon distally. The small bowel increases in length from about 250 cm in the full-term newborn to 600 to 800 cm in the adult. The duodenum, the most proximal portion of the small intestine, begins at the duodenal bulb, travels in the retroperitoneal space around the head of the pancreas, and ends on its return to the peritoneal cavity at the ligament of Treitz. The remainder of the small intestine is suspended within the peritoneal cavity by a thin broad-based mesentery that is attached to the posterior abdominal wall and allows relatively free movement of the small intestine within the abdominal cavity. The proximal 40% of the mobile small intestine is the jejunum, and the remaining 60% is the ileum. The jejunum occupies the left upper portion of the abdomen, and the ileum is positioned in the right abdomen and upper part of the pelvis. No distinct anatomic demarcation exists between jejunum and ileum. Visual examination of the luminal surface of the small intestine reveals mucosal folds, the plicae circulares. More numerous in the proximal jejunum, the plicae circulares

Meckel’s Diverticulum and Other Vitelline Duct Abnormalities  1628 Malrotations  1630 Proliferation  1631 Intestinal Atresia and Stenosis  1632 Anorectum  1633 Enteric Nervous System  1636 Miscellaneous and Genetic Defects  1640

decrease in number in the distal small bowel and are absent in the terminal ileum. Aggregates of lymphoid follicles are scattered throughout the small intestine but are found in highest concentration in the ileum, where they are designated Peyer’s patches. Peyer’s patches normally are more prominent during infancy and childhood than they are in adulthood. The small bowel transitions to the colon at the ileocecal valve, which consists of two semilunar lips that protrude into the cecum. The ileocecal valve provides a barrier to retrograde flow of colonic contents into the small intestine. This barrier appears to be a function of the angulation between the ileum and cecum and is maintained by the superior and inferior ileoceal ligaments; a true tonic sphincter-type pressure does not appear to be present in this region.

Colon and Rectum

The colon is a tubular structure approximately 30 to 40 cm in length at birth in the full-term infant. In the adult, the colon measures approximately 150 cm, about one quarter of the length of the small intestine. The diameter of the colon is greatest in the cecum (7.5 cm) and narrowest in the sigmoid (2.5 cm). The colon is continuous with the small intestine proximally at the ileocecal valve and ends distally at the anal verge (Fig. 96-1). The external appearance of the colon differs from that of the small intestine because the

1615

1616

Section X  Small and Large Intestine Hepatic flexure

Greater omentum (cut away) Epiploic taenia

Free taenia (taenia libera)

Transverse mesocolon Haustra

Hook exposing epiploic Appendices taenia epiploicae Semilunar folds (plicae semilunares)

Ileocecal valve Figure 96-1.  Macroscopic characteristics of the colon. Note the taeniae, haustra between the taeniae and appendices epiploicae on the outer surface, and the semilunar folds on the luminal side. (From Netter FH. The Netter Collection of Medical Illustration. vol 3. Teterboro, NJ: Icon Learning System; 2002.)

Greater omentum (cut away)

Free taenia (taenia libera) Sigmoid mesocolon

Splenic flexure

Hook exposing mesocolic taenia Haustra

Cecum Appendix

Rectosigmoid junction

longitudinal muscle fibers of the colon coalesce into three discrete bands called taeniae, located at 120-degree intervals about the colonic circumference: taenia liberis, taenia omentalis, and taenia mesocolica. The taeniae start at the base of the appendix and extend continuously to the proximal rectum. Outpouchings of the colon, the haustra, are found between the taeniae. Semilunar folds characterize the mucosa between haustra. Small sacs of peritoneum filled with adipose tissue, the appendices epiploicae, are found on the external surface of the colon. The most proximal portion of the colon, the cecum, lies in the right iliac fossa and projects downward as a blind pouch below the entrance of the ileum. The cecum is a sacculated structure 6 to 8 cm in length and breadth. Because of its large diameter, it is the part of the colon most apt to rupture with distal obstruction, and cecal tumors can grow to be quite large without producing symptoms of obstruction. The mobility of the cecum normally is fixed by a small mesocecum; an anomaly in fixation exists in 10% to 20% of people, especially women, predisposing them to cecal volvulus. The vermiform appendix is a blind outpouching of the ceum that begins inferior to the ileocecal valve. Appendiceal anatomy is discussed further in Chapter 116. The ascending colon extends from the cecum distally for 12 to 20 cm along the right side of the peritoneal cavity to the hepatic flexure. The ascending colon is covered with peritoneum and thus constitutes a retroperitoneal organ. At the hepatic flexure, the colon turns medially and anteriorly to emerge into the peritoneal cavity as the transverse colon. This is the longest portion (40 to 50 cm) and the most mobile segment of the colon and drapes itself across the anterior abdomen between the hepatic and splenic flexures. When a person assumes the upright posi-

tion, the transverse colon may actually dip down into the pelvis. The transverse colon may become fixed in this festooned position by adhesions, most commonly resulting from hysterectomy, potentially leading to a technically difficult colonoscopy. The descending colon, about 30 cm in length, travels posteriorly and then inferiorly in the retroperitoneal compartment to the pelvic brim. There, it emerges into the peritoneal cavity as the sigmoid colon. This is an S-shaped redundant segment of variable length, tortuosity, and mobility, which challenges the endoscopist and radiologist, and is susceptible to volvulus. Because the sigmoid is the narrowest part of the colon, tumors and strictures of this region typically cause obstructive symptoms early in the course of disease. The rectum, 10 to 12 cm in length in the adult, begins at the peritoneal reflexion and follows the curve of the sacrum, ending at the anal canal.

Anal Canal

The anal canal is approximately 5 cm in length in the adult and has discrete upper and lower demarcations. The anorectal ring is located proximally and is composed of the upper portion of the internal sphincter, the longitudinal muscle of the rectum, the deep portion of the external sphincter, and the puborectalis portion of the levator ani muscle; distally, the anal verge represents the transition of anoderm to true skin. The mucosa of the distal 3 cm of the rectum and the anal canal contains 6 to 12 redundant longitudinal folds called the columns of Morgagni, which terminate in the anal papillae. These columns are joined together by mucosal folds called the anal valves, which are situated at the dentate line. The muscularis mucosae disap-

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine pears in the anorectal canal, and the inner circular coat of muscularis propria thickens to form the internal anal sphincter. The external anal sphincter surrounds the anal canal, and its fibers blend with those of the levator ani muscle to attach posteriorly to the coccyx and anteriorly to the perineal body. The anatomy and function of these muscles are described in more detail in Chapter 125.

mp m

s

Vasculature

The superior mesenteric artery delivers oxygenated blood to the distal duodenum, the jejunum and ileum, the ascending colon, and the proximal two thirds of the transverse colon. The remainder of the colon is supplied by branches of the inferior mesenteric artery. The arterial supply of the anal area is from the superior, middle, and inferior hemorrhoidal arteries, which are branches of the inferior mesenteric, hypogastric, and internal pudendal arteries respectively. Venous drainage of the anus is by both the systemic and portal systems. The internal hemorrhoidal plexus drains into the superior rectal veins and then into the inferior mesenteric vein, which, with the superior mesenteric vein, joins the splenic vein to form the portal vein. The distal anus drains by the external hemorrhoidal plexus through the middle rectal and pudendal veins into the internal iliac vein. (See Chapter 114 for additional discussion of the intestinal blood supply.)

mm

sm

Figure 96-2.  Photomicrograph of the small intestine showing its general microscopic architecture. m, mucosa; mm, muscularis mucosae; mp, muscularis propria; s, serosa; sm, submucosa. (Hematoxylin and eosin, ×25.)

Lymphatic Drainage

The lymphatic drainage of both the small intestine and colon follows their respective blood supplies to lymph nodes in the celiac, superior preaortic, and inferior preaortic regions. Lymphatic drainage proceeds to the cisterna chyli and then via the thoracic duct into the left subclavian vein. Proximal to the dentate line, lymphatic drainage is to the inferior mesenteric and periaortic nodes, whereas distal to the dentate line, lymphatic drainage is to the inguinal lymph nodes. Therefore, inguinal lymphadenopathy can be seen with inflammatory and malignant disease of the lower anal canal.

ge

lp

Extrinsic Innervation

The autonomic nervous system—sympathetic, parasympathetic, and enteric—innervates the gastrointestinal tract. The sympathetic and parasympathetic nerves constitute the extrinsic nerve supply and connect with the intrinsic nerve supply, which is composed of ganglion cells and nerve fibers within the intestinal wall. Innervation of the small intestine and colon is discussed in detail in Chapters 97 and 98, respectively.

A

mv

MICROSCOPIC FEATURES General Considerations

The small and large intestine share certain histologic characteristics. The wall of the small intestine and colon is composed of four layers: mucosa (or mucous membrane), submucosa, muscularis (or muscularis propria), and adventitia (or serosa) (Fig. 96-2). Mucosa The mucosa is the innermost layer formed by glandular epithelium, lamina propria, and muscularis mucosae (Fig. 96-3A and B). The glandular epithelium forms cylindrical structures called crypts. The lamina propria, which supports the epithelium, is a layer of reticular connective tissue with elastin, reticulin, and collagen fibers, lymphocytes, plasma cells, and eosinophilic granulocytes, as well as lymphatics and capillaries. The muscularis mucosae consists of

B

mm

C

Figure 96-3.  Histologic and electron microscopic photographs of the small intestine. A, Components of the mucosa: ge, glandular epithelium; lp, lamina propria. Note the absorptive cells that appear as high columnar cells with eosinophilic cytoplasm (arrow). (Hematoxylin and eosin, ×250.) B, Goblet cells (arrow) and brush border are stained red. mm, muscularis mucosae. (Periodic acid–Schiff stain, ×150.) C, Microvilli (mv) are seen as delicate finger-like projections on electron microscopic examination, ×9000. (C, Courtesy of S. Teichberg, PhD.)

1617

1618

Section X  Small and Large Intestine a thin layer of smooth muscle at the boundary of the mucosa and submucosa. The glandular epithelium is composed of various cell types: stem cells, undifferentiated crypt cells, absorptive cells (also called columnar cells), secretory cells (goblet cells, Paneth cells, enteroendocrine cells), and M cells. Signaling pathways such as Wnt, bone marrow protein (Bmp), PTEN/PI3K, Notch, hedgehog, platelet-derived growth factor, and SOX9 play important roles in the development of the intestinal epithelium.1-4 Wnt signaling plays a role in promoting cell proliferation; maintains stem cells in an undifferentiated state; defines compartmentalization into Paneth cells, proliferative, and differentiation zones along the crypt-villus axis; and directs early secretory lineage development as well as terminal differentiation of Paneth cells through the transcription factor SOX9.1 Bmps belong to the transforming growth factor-β family. Bmp signaling is important in intestinal development and homeostasis. It antagonizes crypt formation and stem cell self-renewal and has a role in directing maturation of all three secretory cell types (goblet, enteroendocrine, and Paneth). Bmp signaling in the mesenchyme plays a signi­ ficant role in crypt morphogenesis; loss of Bmp leads to multiplication and elongation of crypts.2 PTEN/PI3K pathway plays a role in cell survival, proliferation, and growth.1 Notch proteins mediate cell fate decisions and pattern by regulating the helix-loop-helix factor that controls terminal differentiation. Notch directs development of absorptive cells and depletion of secretory lineage cells, and increases proliferation.1 The hedgehog (Hh) signaling pathway is important in crypt and villus morphogenesis and maintenance of stem cells.3 Both Sonic (Shh) and Indian (Ihh) play a role. Ihh is critical for the maintenance of intestinal stem cells, whereas Shh inhibits the growth of the villi. The contractile subepithelial pericryptal myofibroblasts represent a major target for Hh signaling. Hh signals sent to the epithelium-associated subepithelial myofibroblasts localize the precrypt structure and maintain the organization of the crypt-villus axis. Hh signaling also inhibits the proliferation or differen­tiation of smooth muscle and the proliferation compartment of the intestinal epithelium.3 Platelet-derived growth factor A stimulates mesenchymal condensation, proliferation, and evagination of overlying epithelium to form villi.3 Studies in animals also have contributed to the understanding of the molecular mechanism of the different pathways.1,2 Stem cells are pluripotential cells located at the base of the intestinal crypts. Stem cells give rise to all types of mature intestinal epithelial cells and at the same time replenish themselves through self-renewal. Undifferen­ tiated cells have fewer intracellular organelles and microvilli than do absorptive cells. The absorptive cells (see Fig. 96-3A) are high columnar cells with oval basal nuclei, eosinophilic cytoplasm, and a periodic acid–Schiff (PAS)– positive free surface, the brush border (see Fig. 96-3B). On electron microscopic examination, the brush border is seen to be composed of microvilli (see Fig. 96-3C), which are more numerous in the small intestinal than in the colonic epithelium. Small bowel enterocyte microvilli are estimated to increase the luminal surface area of the cell 14- to 40-fold. Goblet cells are oval or round, with flattened basal nuclei (Fig. 96-4A); their cytoplasm is basophilic, metachromatic (see Fig. 96-4B), and PAS positive (see Fig. 96-4C). Paneth cells are flask shaped and have an eosinophilic granular

A

B

C Figure 96-4.  Photomicrographs of the large and small intestine demonstrating goblet cells. A, Clear, empty-looking cytoplasm (arrow) and basal nuclei are seen with use of hematoxylin and eosin, ×250. B, Metachromatic staining of the cytoplasm results with use of the alcian blue stain, ×150. C, The cells demonstrate red staining with use of periodic acid–Schiff stain, ×150.

cytoplasm and a broad base positioned against the basement membrane (Fig. 96-5). Paneth cells contain zinc, antimicrobial peptides, and growth factors and secrete lysoenzymes. Enteric antimicrobial peptides produced by Paneth cells protect against intestinal infection and maintain enteric homeostasis.5 A cathelin-related antimicrobial peptide (CRAMP) identified in neonatal epithelium during the first weeks after birth, confers protection from Listeria monocytogenes.5 The mucosa also contains specialized cells that because of their specific endocrine function are called enteroendocrine or neuroendocrine cells. These neuroendocrine cells historically have been divided into argentaffin cells (granules able to reduce silver nitrate) and argyrophilic cells (granules that reduce silver nitrate only in the presence of a chemical reducer). Argentaffin cells stain positive with bichromate salts and also are called enterochromaffin cells. These cells are oval or triangular

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine

lc

Figure 96-5.  Photomicrograph of the small intestinal mucosa demonstrating the crypts of Lieberkühn (lc) and Paneth cells (arrow), which are characterized by granular eosinophilic cytoplasm. (Hematoxylin and eosin, ×250.)

(also called “halo cells”) and have a basal position in relation to the remaining epithelial cells (Fig. 96-6A) and a pale cytoplasm filled with dark-stained granules. Variation in shapes and cell types has been detected with immunohistochemical staining. The unifying APUD concept— amine precursor, uptake, and decarboxylation—ascribes common characteristics to these neuroendocrine cells. APUD cells are a group of cells with a common embryonic neural crest origin and with similar cytochemical and electron microscopic features; however, embryologic and morphologic data support an endodermal origin of these cells. Ultrastructurally, enteroendocrine cells contain membrane-bound granules with variably sized electrodense cores (see Fig. 96-6B), averaging 100 to 250 nm in diameter, and consisting of large dense-core vesicles and smaller, synaptic-type microvesicles. Neurosecretory granules can be demonstrated with the Grimelius stain by light microscopy as dark granules (see Fig. 96-6C), or, more specifically, by immunofluorescence, and with immunohistochemical stains such as neuron-specific enolase, chromogranin, and synaptophysin. Chromogranin enables visualization of the large dense-core vesicles, and synaptophysin targets the small synaptic-like microvesicles (see Fig. 96-6D).6 Vesicular monoamine transporter 1 (VMAT1) and 2 (VMAT2) are two isoforms of the adenosine triphosphate (ATP)– dependent vesicular monoamine transporters. These antigens, derived from both the large and small dense-core vesicles, are expressed differentially in small dense-core vesicles. Both are expressed in neuroendocrine cells, but VMAT1 is restricted to serotonin-producing enterochromaffin cells, and VMAT2 is expressed in histamine-producing cells, enterochromaffin-like cells, and pancreatic islet cells.7 Specific immunohistochemical stains allow for identification of individual protein products of the neuroendocrine cells. Besides releasing hormones in the blood, neuroendocrine cells also regulate secretion, absorption, motility, mucosal cell proliferation, and possibly immunobarrier control.6 Electron microscopy and immunohistochemistry have led to the identification of a variety of cell types (Table 96-1). Designation according to the nature of the stored peptide is preferable to characterization of neuroendocrine cells by letters. Serotonin-producing enterochro-

maffin cells, vasoactive intestinal polypeptide (VIP), and somatostatin D cells are distributed throughout the small and large intestine. Gastrin-, ghrelin-, gastric inhibitory peptide (GIP)-, secretin-, and cholecystokinin-producing cells are found predominantly in the stomach and proximal small intestine; peptide YY-, glucagon-like peptide (GLP)-1-, GLP-2-, and neurotensin-secreting cells are found in the ileum.8 Neuroendocrine cells originate from a common precursor cell in the intestinal crypt. The earliest cell fate is regulated by the Notch signaling pathway (see earlier). Math1 is the first factor involved in endocrine specification, followed by neurogenin3.8 Pax4 and Pax6, paired ox homeodomain transcription factors, and Nkx2.2 also are required for neuroendocrine differentiation.8,9 As mature neuroendocrine cells migrate to the tip of the villi, they undergo apoptosis and are extruded into the lumen. M cells are specialized epithelial cells overlying lymphoid follicles in the small intestine and colon. M cells selectively bind, process, and deliver pathogens directly to lymphocytes, macrophages, or other components of the mucosal lymphoid system. Interstitial cells of Cajal (ICC) are present in both the small intestine and the colon and are mesenchymal cells, located in the myenteric plexus, the muscularis propria, and the submucosa (Fig. 96-7). The distribution of the ICC is similar in children and in adults although a difference in their distribution is seen in fetuses of different gestational ages.10 Recognized as the pacemaker cells of the intestine, the ICC regulate intestinal motility by generating slow waves and determining frequency of smooth muscle contraction; they also amplify the neuronal signals, mediate neurotransmission from enteric motor neurons to smooth muscle cells, and set the smooth muscle membrane potential gradient. The ICC are spindle shaped or stellate, with long ramified processes, and have large, oval light-staining nuclei with sparse perinuclear cytoplasm. The ICC express the receptor for tyrosine kinase (c-Kit) or CD117 which is necessary for their maintenance. Serotonin regulates the number of the ICC by increasing their proliferation.11 Immunohis­ tochemical stains that use antibodies against c-Kit allow the ICC to be labeled. The distribution and onset of appearance of these cells in the gastrointestinal tract have been described.10 Submucosa The submucosa, between the muscularis mucosae and the muscularis propria, is a fibrous connective tissue layer that contains fibroblasts, mast cells, blood and lymphatic vessels, and a nerve fiber plexus—Meissner’s plexus—comprised of nonmyelinated, postganglionic sympathetic fibers, and parasympathetic ganglion cells. Muscularis or Muscularis Propria The muscularis propria, mainly responsible for contractility, consists of two layers of smooth muscle: an inner circular coat and an outer longitudinal coat arranged in a helicoidal pattern. A prominent nerve fiber plexus called the myenteric plexus, or Auerbach’s plexus, is found between these two muscle layers (Fig. 96-8). Parasympathetic and postganglionic sympathetic fibers terminate in parasympathetic ganglion cells, and postganglionic parasympathetic fibers terminate in smooth muscle. Adventitia or Serosa The adventitia is the outermost layer of connective tissue. When covered by a single layer of mesothelial cells, it is called the serosa.

1619

1620

Section X  Small and Large Intestine

A

B

C

D

Figure 96-6.  Microscopic characteristics of neuroendocrine cells of the small intestine. A, Features include clear cytoplasm and a round nucleus (arrow). (Hematoxylin and eosin, ×250.) B, Neurosecretory granules are seen as electron-dense, round black bodies (arrow) on electron microscopic examination, ×20,000. C, Granules in neuroendocrine cells are stained black with the Grimelius stain (arrow), ×150. D, Cells stained with synaptophysin have brown cytoplasm (arrow), ×250. (B, Courtesy of S. Teichberg, PhD.)

Table 96-1  Enteroendocrine Cells of the Intestinal Tract: Cell Types and Products, Vesicle Markers, and Distribution Vesicle Markers CELL TYPE

CELL product

LDCV

P/D1 EC D L PP G CCK S GIP M N

Ghrelin 5-HT Somatostatin GLI/PYY PP Gastrin Cholecystokinin Secretin, 5-HT GIP/Xenin Motilin Neurotensin

CgA, VMAT2 CgA, VMAT1 CgA SgII > CgA CgA, SgII, VMAT2 CgA CgA CgA CgA

SLMV Syn Syn Syn Syn Syn

DUOD

JEJ

ILEUM

APP

Colon

Rec

f + + f e + + + + + f

f + + +

f + f +

+ f +

+ f +

+ f +

+ + + + +

f f f +

App, appendix; CgA, chromogranin A; DUOD, duodenum; e, presence of cells in fetus and newborn; EC, enterochromaffin cell, 5-HT, 5-hydroxytryptamine (serotonin); f, presence of few cells; GIP, gastric inhibitory polypeptide; GLI, glucagon-like immunoreactants (glicentin, glucagon-37, glucagon-29, GLP-1, GLP-2); JEJ, jejunum; LDCV, large dense-core vesicles; NESP55, neuroendocrine secretory protein 55; PP, pancreatic polypeptide; PYY, PP-like peptide with N-terminal tyrosine amide; REC, rectum; SgII, secretogranin II (also known as chromogranin C); SLMV, synaptic-like microvesicles; Syn, synaptophysin; VMAT1, VMAT2, vesicular monoamine transporter 1, 2; +, presence of cells; >, heavier staining than. Adapted from Solcia E, Capela C, Fiocca R, et al. Disorders of the endocrine system. In: Ming SC, Goldman H, editors. Pathology of the Gastrointestinal Tract. Philadelphia: Williams & Wilkins; 1998. p 295.

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine v

A

Figure 96-7.  Photomicrograph showing the interstitial cells of Cajal in the small intestine. Brown-staining, elongated cells are evident around the myenteric plexus (arrow). (CD117 immunostain, ×250.)

il

bg

B Figure 96-9.  Photomicrograph of the duodenal mucosa. A, Villi are seen as finger-like projections. B, Brunner glands (bg) are found below the mucosa. (Hematoxylin and eosin. A, ×250; B, ×150.)

mp

ol

Figure 96-8.  Photomicrograph of the muscularis propria of the small intestine. The myenteric plexus (mp) is seen as a pale area with ganglion cells between the inner and outer layers (il, ol) of the muscularis propria (arrow). (Hematoxylin and eosin, ×250.)

Small Intestine

The mucosa of the small intestine is characterized by mucosal folds (plicae circulares, or valves of Kerckring) and villi. The mucosal folds are composed of mucosa and submucosa. Villi are mucosal folds that decrease in size from the proximal to distal small intestine and are of different shapes in the various segments of the small intestine: they may be broad, short, or leaf-like in the duodenum; tonguelike in the jejunum; and finger-like more distally (Fig. 96-9A). The villous pattern also may vary in different ethnic groups. Thus, for example, biopsy specimens from Africans, Indians, South Vietnamese, and Haitians have shorter and thicker villi, an increased number of leaf-shaped villi, and more mononuclear cells in comparison with specimens from North Americans. Various methods have been suggested to determine normal villus height. The height of the normal villus is 0.5 to 1.5 mm; villus height should be more than one half of the total thickness of the mucosa, and three to five times the length of the crypts. Villi are lined by enterocytes, goblet cells, and enteroendocrine cells.

Intestinal villus morphogenesis begins when mesen­ chymal aggregates impinge on the basal aspect of the epithelium to produce primitive folds. By nine to 10 weeks of gestation, the pseudostratified squamous epithelium converts to a single layer of columnar cells that lines mesenchymal stalks or the lamina propria.12 During mid- to late gestation, the basic tissue architecture of the intestine is established through epithelial-mesenchymal interaction. Induced by signals from mesoderm-derived mesenchyme, the endoderm-derived epithelium evaginates to form villi and intervillus regions. The intervillus region eventually invaginates into the mucosa to form crypts.1 Contractile subepithelial pericryptal myofibroblasts contribute mechanically to crypt formation and are the major source of instructive signals to the epithelium.3 Two types of glands are present in the small intestine: Brunner’s glands and crypts of Lieberkühn (intestinal crypts). Brunner’s glands are submucosal glands (see Fig. 96-9B) found primarily in the first portion of the duodenum and in decreased numbers in the distal duodenum; their function is to secrete a bicarbonate-rich alkaline secretion that helps neutralize gastric chyme. In children these glands also may be present in the proximal jejunum. Brunner’s glands open into the intestinal crypts and morphologically resemble pyloric glands. Crypts of Lieberkühn are tubular glands that extend to the muscularis mucosae (see Fig. 96-5). The crypts are occupied mainly by undifferentiated cells and Paneth cells. Cells are generated at the crypt base and proceed to migrate toward the villus. During this migration, these cells mature and differentiate into a secretory lineage (goblet cells, enteroen-

1621

1622

Section X  Small and Large Intestine docrine cells, Paneth cells) and enterocytes. The commitment of the stem cells to differentiate is acquired in the upper third of the crypt, where cells lose their ability to divide. The constant renewal of enterocytes is regulated by human acyl-coenzyme A synthetase.13 Paneth and columnar cells predominate in the base of the crypt. Above the base are absorptive cells and oligomucin cells; the latter originate from undifferentiated cells and differentiate into goblet cells. Goblet cells predominate in the upper half of the crypt. Enteroendocrine cells are admixed with goblet cells. A certain number of CD3+ intraepithelial T lymphocytes (up to 30 per 100 epithelial cells) normally are present in the villi. Smooth muscle is found in the lamina propria of the small intestinal villus, extending vertically up from the muscularis mucosae. Plasma cells containing primarily immunoglobulin A (IgA), and mast cells also are present. Lymphoid tissue is prominent in the lamina propria as solitary nodules and as confluent masses—Peyer’s patches—and is seen in the submucosa. Peyer’s patches are distributed along the antimesenteric border and are most numerous in the terminal ileum; their numbers decrease with age. Most types of enteroendocrine cells are present in the duodenum. Cells that produce ghrelin, gastrin, cholecystokinin, motilin, neurotensin, GIP, and secretin are restricted to the small intestine.6 The proportions of these cells differ in the villi and crypts, as well as in different segments of the intestine. Ninety percent of the villus epithelial cells are absorptive cells intermingled with goblet and enteroendocrine cells. The proportion of goblet to absorptive cells is increased in the ileum. The ICC are more abundant in the myenteric plexus of the small bowel than in the colon.10

Colon

Colonic epithelial cells are generated from stem cells at the base of the crypts and migrate toward the intestinal lumen after three to five days, on initiation of apoptosis. Most epithelial cells undergo apoptosis when they lose contact with the extracellular matrix and are shed into the lumen through caspase activation. Caspase activation is responsible for the cleavage of essential intracellular proteins leading to apoptosis and therefore loss of anchorage.14 The mucosa of the large intestine is characterized by the presence of crypts of Lieberkühn, associated predominantly with goblet cells intermixed with a few absorptive and enteroendocrine cells. Glucagon-like immunoreactant (GLI), pancreatic polypeptide-like peptide (PYY) with N-terminal tyrosine amide–producing L cells predominate in the large intestine. Enterochromaffin, enterochromaffin-like, and pancreatic polypeptide–producing cells also are found. Paneth cells are scarce and normally are noted only in the proximal colon. The lamina propria of the large intestine contains solitary lymphoid follicles extending into the submucosa. Lymphoid follicles are more developed in the rectum and decrease in number with age. Confluent lymphoid tissue is present in the appendix. Macrophages (muciphages) predominate in the subepithelial portion of the lamina propria. These cells are weakly PAS positive and are associated with stainable lipids.

Anal Canal

Microscopically the anal canal is divided into three zones: proximal, intermediate or pectinate, and distal or anal skin. The proximal zone is lined by stratified cuboidal epithelium, and the transition with the rectal mucosa, which is lined by high columnar mucus-producing cells, is called the

rg

ep

A ep as

B Figure 96-10.  Photomicrograph of the anal canal. A, Anorectal histologic junction. Transition from rectal glandular mucosa (rg) to proximal anal mucosa lined by stratified squamous epithelium (ep) is evident. B, The pectinate line is characterized by anal mucosa with stratified squamous epithelium (ep) and anal skin (as) containing adnexae (arrow). (A and B, Hematoxylin and eosin, ×150.)

anorectal histologic junction (Fig. 96-10A). The intermediate or pectinate zone is lined by stratified squamous epithelium but without adnexae (e.g., hair, sebaceous glands) and also is referred to as anoderm. Its proximal margin, in contact with the proximal zone, is called the dentate line; its distal margin, in contact with the anal skin, constitutes the pectinate line, also referred to as the mucocutaneous junction (see Fig. 96-10B). The anal skin is lined by squamous stratified epithelium and contains hair and sebaceous glands.

Vasculature

Large arterial branches enter the muscularis propria and pass to the submucosa, where they branch to form large plexuses. In the small intestine, two types of branches arise from the submucosal plexuses: some arteries branch on the inner surface of the muscularis mucosae and break into a capillary meshwork that surrounds the crypts of Lieberkühn. Other arteries are destined for villi, each receiving one or two arteries, and set up the anatomic arrangement that allows a

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine countercurrent mechanism during absorption. These vessels enter at the base of the villus and form a dense capillary network immediately underneath the epithelium of the entire villus structure. One or several veins originate at the tip of each villus from the superficial capillary plexus, anastomose with the glandular venous plexus, and then enter the submucosa joining the submucosal venous plexus. In the colon, branches from the submucosal plexus extend to the surface, giving rise to capillaries supplying the submucosa, and there branch to form a capillary meshwork around the crypts of Lieberkühn. From the periglandular capillary meshwork, veins form a venous plexus between the base of the crypts and the muscularis mucosae. From this plexus, branches extend into the submucosa and form another venous plexus, from which large veins follow the distribution of the arteries and pass through the muscularis propria into the serosa.

Lymph Vessels

The lymphatics of the small intestine are called lacteals and become filled with milky-white lymph called chyle after eating. Each villus contains one central lacteal, except in the duodenum, where two or more lacteals per villus may be present. The wall of the lacteal consists of endothelial cells, reticulum fibers, and smooth muscle cells. The central lacteals anastomose at the base of the villus with the lymphatic capillaries between the crypts of Lieberkühn. They also form a plexus on the inner surface of the muscularis mucosae. Branches of this plexus extend through the muscularis mucosae to form a submucosal plexus. Branches from the submucosal plexus penetrate the muscularis propria, where they receive branches from plexuses between the inner and outer layers. Lymphatic vessels are absent in the colonic mucosa, but the distribution of lymphatics in the remaining colonic layers is similar to that in the small intestine.

Nerves

The intrinsic nervous system (enteric nervous system) consists of subserosal, muscular, and submucosal plexuses. The subserosal plexus contains a network of thin nerve fibers, without ganglia, that connects the extrinsic nerves with the intrinsic plexus. The myenteric plexus, or Auerbach’s plexus, is situated between the outer and inner layers of the muscularis propria (see Fig. 96-8); it consists of ganglia and bundles of unmyelinated axons that connect with the ganglia forming a meshwork. These axons originate from processes of the ganglion cells and extrinsic vagus and sympathetic ganglia. The deep muscular plexus is situated on the mucosal aspect of the circular muscular layer of the muscularis propria. It does not contain ganglia; it innervates the muscularis propria and connects with the myenteric plexus. The submucosal plexus, or Meissner’s plexus, consists of ganglia and nerve bundles. The nerve fibers of this plexus innervate the muscularis mucosae and smooth muscle in the core of the villi. Fibers from this plexus also form a mucosal plexus that is situated in the lamina propria and provides branches to the intestinal crypts and villi. The ganglion cells of the submucosal plexus are distributed in two layers: one is adjacent to the circular muscular layer of the muscularis propria; the other is contiguous to the muscularis mucosae. Ganglion cells are large cells, isolated or grouped in small clusters called ganglia (Fig. 96-11). Ganglion cells have an abundant basophilic cytoplasm, a large vesicular round nucleus, and a prominent nucleolus. Ganglion cells are scarce in the physiologically hypogan­ glionic segment 1 cm above the anal verge.

g Figure 96-11.  Photomicrograph showing a normal submucosal plexus of the colon. The ganglia (g) are identified by their oval structure and the nerve trunks are thin (arrow). (Hematoxylin and eosin, ×150.)

EMBRYOLOGY The embryo begins the third week of development as a bilaminar germ disk. During week three, in a process called gastrulation, this disk becomes a trilaminar disk. The surface facing the yolk sac becomes the definitive endoderm; the surface facing the amniotic sac becomes the ectoderm. The middle layer is called mesoderm. The long axis and leftright axis of the embryo also are established at this time. The oral opening is marked by the buccopharyngeal membrane; the future openings of the urogenital and the digestive tracts become identifiable as the cloacal membrane. At four weeks of gestation, the alimentary tract is divided into three parts: foregut, midgut, and hindgut. The endoderm forms the intestinal tube, which communicates only with the yolk sac. Narrowing of the communication of the yolk sac with the endoderm forms the vitelline duct. With folding of the embryo during the fourth week of development, the mesodermal layer splits. The portion that adheres to endoderm forms the visceral peritoneum, whereas the part that adheres to ectoderm forms the parietal peritoneum. The space between the two layers becomes the peritoneal cavity.

MOLECULAR REGULATION OF INTESTINAL MORPHOGENESIS

The induction of endoderm appears to be governed by nodal or transforming growth factor-β signaling.15 Specification is initiated by transcription factors expressed in the different regions of the intestinal tube. Thus, PDX1 specifies the duodenum, CDXC the small intestine, and CDXA the large intestine and rectum.16 Differentiation of the gastrointestinal tract depends on the interaction between the endoderm and mesoderm through the Hox code. Signaling from the mesoderm to endoderm is regulated by the Hox genes that encode homeodomain-containing transcription factors. Induction of the Hox code in the mesoderm results from expression of Shh through the endoderm of the midgut and hindgut. Shh is a signaling molecule that acts as a morphogen or form-producing substance in a variety of organ systems. When prompted by this code, the mesoderm instructs the endoderm to form the various components of the midgut

1623

1624

Section X  Small and Large Intestine A Endoderm Ectoderm Angiogenic cell cluster Buccopharyngeal membrane

C Buccopharyngeal membrane

B

Amniotic cavity

Foregut Connecting stalk

Heart tube

Allantois

Pericardial cavity

Cloacal membrane

Cloacal membrane

Hindgut

D Lung bud

Liver bud Midgut

Heart tube

Remnant of the buccopharyngeal membrane Vitelline duct

Allantois

Yolk sac Figure 96-12.  Formation of the foregut, midgut, and hindgut (see text for details). (From Sun B, editor. Langman’s Medical Embryology. 9th ed. Philadelphia: Lippincott Williams & Wilkins; 2004.)

and hindgut regions, for example, the small bowel, cecum, colon, and cloaca.16 As indicated by animal studies, Hox genes contribute to the subdivision of the intestine, and formation of the ileocecal valve that separates the small and the large intestine. Shh also plays a crucial role in the development of the hindgut.17 The primitive gut results from incorporation of the endoderm-lined yolk sac cavity into the embryo, following embryonal cephalocaudal and lateral folding. The primitive gut is composed of a blind-ended tube in the cephalic and caudal portions of the embryo, which is the progenitor of the foregut and hindgut; the midgut (Fig. 96-12) is connected to the yolk sac by the vitelline duct. The endoderm gives rise to the epithelial lining of the gastrointestinal tract; muscle, connective tissue, and peritoneum originate from the splanchnic mesoderm. During the ninth week of development, the epithelium begins to differentiate from the endoderm with villus formation and differentiation of epithelial cell types. Organogenesis is complete by 12 weeks of gestation. Initially the foregut, midgut, and hindgut are in broad contact with the mesenchyma of the posterior abdominal wall. The intraembryonic cavity is in open communication with the extraembryonic cavity. Subsequently the intraembryonic cavity loses its wide connection with the extraembryonic cavity. By week five of embryonic development, splanchnic mesoderm layers are fused in the midline and form a double-layered membrane, the dorsal mesentery, between the right and left halves of the body cavity. The mesoderm surrounds the intestinal tube and suspends it from the posterior body wall, allowing it to hang into the body cavity. The caudal portions of the foregut, the midgut, and most of the hindgut thus are suspended from the abdominal wall by the dorsal mesentery extending from the duodenum to the cloaca. The dorsal mesentery forms the mesoduodenum in the duodenum, the dorsal mesocolon

in the region of the colon, and the mesentery proper in the region of the jejunum and ileum.16

SPECIFIC STRUCTURES AND SYSTEMS Duodenum

The duodenum originates from the terminal portion of the foregut and cephalic part of the midgut. With rotation of the stomach, the duodenum becomes C-shaped and rotates to the right; the fourth portion becomes fixed in the left upper abdominal cavity. The mesoduodenum fuses with the adjacent peritoneum; both layers disappear, and the duodenum becomes fixed in its retroperitoneal location. The lumen of the duodenum is obliterated during the second month of development by proliferation of its cells; this phenomenon is shortly followed by recanalization. Because the foregut is supplied by the celiac artery and the midgut by the superior mesenteric artery, the duodenum is supplied by both arteries and therefore is relatively protected from ischemic injury.16

Midgut

In a 5-week embryo, the midgut is suspended from the dorsal abdominal wall by a short mesentery and communicates with the yolk sac by way of the vitelline duct. The midgut gives rise to the duodenum distal to the ampulla, to the entire small bowel, and to the cecum, appendix, ascending colon, and the proximal two thirds of the transverse colon. The midgut rapidly elongates with formation of the primary intestinal loop. The cephalic portion of this loop, which communicates with the yolk sac by the narrow vitelline duct, gives rise to the distal portion of the duodenum, the jejunum, and a portion of the ileum; the distal ileum, cecum, appendix, ascending colon, and proximal two thirds of the transverse colon originate from the caudal limb. During week 6 of embryonic development, the primary

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine Diaphragm

Esophagus

Lesser omentum

Liver Falciform ligament Stomach

Gallbladder Duodenum Cecum Vitelline duct

Allantois

Descending colon

Jejunoileal loops

Cloacal membrane

Rectum

intestinal loop enters the umbilical cord (physiologic umbilical herniation) (Fig. 96-13), and by week 10 it re-enters the abdominal cavity. The proximal portion of the jejunum is the first portion of the intestine to re-enter the abdominal cavity and becomes located on the left side; the subsequent loop that re-enters the abdominal cavity locates to the right. The cecal bud is the last segment to re-enter the abdominal cavity. The cecum originates as a small dilatation of the caudal limb of the primary intestinal loop by approximately 6 weeks of development. Initially it lies in the right upper quadrant; then it descends to the right iliac fossa, placing the ascending colon and hepatic flexure in the right side of the abdominal cavity. The appendix originates from the distal end of the cecal bud. Because the appendix develops during descent of the colon, its final position frequently is retrocecal or retrocolonic. The primary intestinal loop rotates counterclockwise for approximately 270 degrees around an axis formed by the superior mesenteric artery. This rotation occurs in three stages (Fig. 96-14): the first stage occurs between six and eight weeks (90 degrees), the second stage is at nine weeks (180 degrees), and the third stage is at 12 weeks of gestation (270 degrees). Elongation of the bowel continues, and the jejunum and ileum form a number of coiled loops within the peritoneal cavity.16

Mesentery

When the caudal limb of the primitive intestine moves to the right side of the abdominal cavity, the dorsal mesentery twists around the origin of the superior mesenteric artery. After the ascending and the descending portions of the colon reach their final destinations, their mesenteries fuse with the peritoneum of the posterior abdominal wall, and they become retroperitoneal organs. The appendix, cecum, and descending colon retain their free mesentery. The transverse mesocolon fuses with the posterior wall of the greater omentum. The mesentery of the jejunum and ileum at first

Figure 96-13.  Physiologic umbilical herniation of the intestinal loop during normal development. Coiling of the small intestinal loops and formation of the cecum occur during the herniation. The first 90 degrees of rotation occur during herniation; the remaining 180 degrees occur during the return of the intestine to the abdominal cavity. (From Sun B, editor. Langman’s Medical Embryology. 9th ed. Philadelphia: Lippincott Williams & Wilkins; 2004.)

is in continuity with the ascending mesocolon; after the ascending colon becomes retroperitoneal, the mesentery only extends from the duodenum to the ileocecal junction.16

Hindgut

The distal third of the transverse colon, the descending colon and sigmoid, the rectum, and the upper part of the anal canal originate from the hindgut. Initially the urinary, genital, and rectal tracts empty into a common channel, the cloaca. They become separated by the caudal descent of the urorectal septum into an anterior urogenital sinus and a posterior intestinal canal. The lateral fold of the cloaca moves to the midline, and the caudal extension of the urorectal septum develops into the perineal body. In a man, the lateral genital ridges coalesce to form the urethra and scrotum; in a woman, no fusion occurs, and the labia minora and majora evolve. The cloaca is lined by endoderm and covered anteriorly by ectoderm. The most distal portion of the hindgut enters into the posterior region of the cloaca, the primitive anorectal canal. The boundary between the endoderm and the ectoderm forms the cloacal membrane. This membrane ruptures by the seventh week of embryonic development, creating the anal opening for the hindgut. This portion is obliterated by the ectoderm but recanalizes by week nine. Thus, the distal portion of the anal canal originates from the ectoderm and is supplied by the inferior rectal artery; the proximal portion of the anal canal originates from the endoderm and is supplied by the superior rectal artery. The pectinate line is situated at the junction of the endoderm and the ectoderm.

Arterial System

Vascular endothelial growth factor (VEGF)-A and its receptors, VEGFR-1 and VEGFR-2, are important for endothelial cell proliferation, migration, and sprouting. Angiopoietins

1625

1626

Section X  Small and Large Intestine 1st stage

3rd stage

2nd stage

Figure 96-14.  The three stages of normal intestinal rotation (see text for details). (From Gosche JR, Touloukian RJ. Congenital anomalies of the midgut. In: Wyllie R, Hyams JS, editors. Pediatric Gastrointestinal Disease. Pathophysiology, Diagnosis, Management. 2nd ed. Philadelphia: WB Saunders; 1999.)

and their receptors, Tie1 and Tie2, play a role in remodeling and maturation of the developing vasculature. Mutation in Tie2 has been reported in vascular dysmorphogenesis. Vascular malformation is briefly discussed in Chapter 36. Arteries of the dorsal mesentery, originating from fusion of the vitelline arteries, give rise to the celiac, superior mesenteric, and inferior mesenteric arteries. Their branches supply the foregut, midgut, and hindgut, respectively.

Venous System

Vitelline veins give rise to a periduodenal plexus that develops into a single vessel, the portal vein. The superior mesenteric vein originates from the right vitelline vein that receives blood from the primitive intestinal loop. The left vitelline vein disappears. The umbilical veins become connected to the hepatic sinusoids after which the right umbilical vein disappears and the left umbilical vein joins the inferior vena cava; ultimately the umbilical vein is obliterated and forms the ligamentum teres. The cardinal veins are involved with forming the inferior vena cava as is the proximal portion of the right vitelline vein.

Lymphatic System

Lymphatics originate from endothelial budding of veins, after which the peripheral lymphatic system spreads by endothelial sprouting into the surrounding tissues and organs. Flt4 (also known as VEGFR-3), a receptor for VEGF, plays a role in development of the vascular as well as the lymphatic systems. Overexpression of VEGF-C, a ligand of Flt4, results in hyperplasia of lymphatic vessels in transgenic mice. The homeobox gene Prox1 is essential for normal development of the lymphatic system based on animal studies. Homeobox genes contain a conserved sequence of 183 nucleotides. The proteins encoded by homeobox-containing genes act as regulatory molecules that control the expression of other genes. Several families of homeobox-containing genes are known, including the murine Hox family, which has been implicated in pattern formation during embryogenesis. Disruption of this gene in mice causes chyle-filled intestine. Abnormalities in the development of the lymphatic system can result in lymphangiectasia (see Chapter 28).

Enteric Nervous System

The enteric nervous system originates from vagal, truncal, and sacral neural crest cells. Most of the enteric nervous

system cells derive from the vagal and truncal neural crest, enter the foregut mesenchyma, and colonize the developing intestine in a cephalocaudal direction. The truncal neural crest gives rise to ganglia of the proximal stomach, whereas the vagal neural crest supplies ganglia to the entire intestine including the rectum; this colonization is complete by 13 weeks of embryonic development. A small component of the enteric nervous system originates from sacral neural crest cells. These cells form extraintestinal pelvic ganglia that colonize the hindgut mesenchyma before the arrival of the vagal-derived neural crest cells.18 The normal development of the enteric nervous system depends on the survival of cells developed from the neural crest, and their proliferation, movement, and differentiation into neurons and glial cells. Microenvironmental, genetic, or molecular mechanisms may intervene in these processes (see “Disturbance in the Enteric Nervous System”).

CLINICAL IMPLICATIONS

Table 96-2 summarizes the different congenital clinical entities that result from disturbances in embryologic development. Gastrointestinal malformations can be associated with extraintestinal defects when genes such as those that determine left-right asymmetry are involved. The CFC1 gene plays a role in establishing left-right axis. Mutations of this gene have been reported in extrahepatic biliary atresia, in the polysplenia syndrome (inferior vena cava abnormalities, preduodenal portal vein, intestinal malrotation, and situs inversus), and in right-sided stomach and congenital heart disease.19,20

ABNORMALITIES IN NORMAL EMBRYOLOGIC DEVELOPMENT ABDOMINAL WALL Omphalocele

Current theories suggest that a teratogenic event during the first three weeks of gestation prevents return of the bowel to the abdomen and causes failure of lateral embryonic fold development, which results in an omphalocele. Omphalocele occurs with a frequency of 2.5 in 10,000 births. Associated anomalies (e.g., sternal defects) result from failure of closure of the cephalic folds; failure of caudal fold develop-

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine Table 96-2  Causes of Abnormalities in Normal Embryologic Development Body wall   Omphalocele   Gastroschisis Mesentery   Mobile cecum   Volvulus Vitelline duct   Meckel’s diverticulum   Omphalomesenteric cyst   Patent omphalomesenteric duct Rotation   Malrotation   Nonrotation   Reverse rotation Proliferation   Duplication Intestinal atresia and stenosis   “Apple-peel” atresia   Duodenum   Small and large intestine Anorectum Enteric nervous system   Hirschsprung’s disease   Intestinal neuronal dysplasia   Pseudo-obstruction Miscellaneous   Intestinal epithelial dysplasia   Microvillus inclusion disease Other genetic defects   Congenital chloride diarrhea   Congenital glucose or galactose malabsorption   Congenital lactase deficiency   Congenital sodium diarrhea   Congenital sucrase/isomaltase deficiency   Cystic fibrosis

Failure of the intestine to return to the abdominal cavity after its physiologic herniation Weakening of the abdominal wall Persistence of mesocolon Failure of fusion of mesocolon with posterior abdominal wall Persistence of the vitelline duct (see Fig. 96-17) Focal failure of vitelline duct obliteration Total failure of vitelline duct obliteration Failure of rotation of the proximal midgut; distal midgut rotates 90 degrees clockwise Failure of stage 2 rotation (see Fig. 96-18) Rotation of 90 degrees instead of 270 degrees Abnormal proliferation of the intestinal parenchyma Coiling of proximal jejunum distal to the atresia around the mesenteric remnant Lack of recanalization Vascular “accident” Disturbance in hindgut development Failure of migration of ganglion cells; microenvironment changes Controversial Multifactorial (see Chapter 120) Abnormalities of the basement membrane Defective protein trafficking and abnormal cytoskeletal and microfilament function Abnormal Cl−-HCO3− exchange in the ileum and colon Absence of Na+-glucose cotransporter for glucose and galactose Decrease in lactase-phlorizin hydrolase Defective sodium-proton exchange Abnormal intracellular transport, aberrant processing, and defective function of sucrase or isomaltase Defective cystic fibrosis transmembrane conductance regulator

ment results in exstrophy of the bladder and, in extreme cases, exstrophy of the cloaca. Omphalocele is a congenital hernia involving the umbilicus. It is covered by an avascular sac composed of fused layers of amnion and peritoneum (Fig. 96-15). The umbilical cord usually is inserted into the apex of the sac, and the blood vessels radiate within the sac wall. Although a central defect is present in the skin and the linea alba, the remainder of the abdominal wall is intact, including the surrounding musculature. Because a small occult omphalocele of the umbilical cord may not be observed at birth, it is recommended that the umbilical cord be tied at least 5 cm from the abdominal wall at the time of delivery. Close inspection of the umbilical cord before clamping will avoid clamping an occult omphalocele. With a large omphalocele, the liver and spleen frequently are outside the abdominal cavity. Associated anomalies occur in about 75% of children with omphalocele and include chromosomal abnormalities such as trisomy 13 or 18, nonchromosomal syndromes such as BeckwithWiedemann syndrome (mental retardation, hepatomegaly, large body stature, hypoglycemia), fetal valproate syndrome, exstrophy of the bladder or cloaca, and OEIS (omphalocele, exstrophy of the bladder, imperforate anus, spinal defect). Malformations of the musculoskeletal, cardiovascular, and central nervous systems, also can occur.21,22 Prenatally, increased levels of maternal serum alpha fetoprotein suggest the possible presence of an omphalocele. Ultrasound examination during pregnancy allows the diagnosis of this abdominal wall defect in most infants.

Figure 96-15.  A newborn with an omphalocele. Note the translucent saclike structure with its attached umbilical cord.

1627

1628

Section X  Small and Large Intestine Increased maternal levels of alpha fetoprotein are suggestive of gastroschisis, as well as omphalocele. Most children with gastroschisis can undergo primary closure safely; however, for the child with significant intestinal atresia as a complication of gastroschisis, bowel exteriorization and secondary closure often are preferred treatment. It is crucial to conserve intestinal length in these children. Adhesive small bowel obstruction is a frequent and a serious complication, especially in the first year of life.23

MECKEL’S DIVERTICULUM AND OTHER VITELLINE DUCT ABNORMALITIES

Figure 96-16.  Gastroschisis. In this newborn, there are full-thickness disruption of the abdominal wall and protruding viscera without accom­ panying peritoneum. (From Feldman’s Online Gastro Atlas, Current Medicine.)

Fetal management, including possible termination of pregnancy, is determined by the physician in consultation with the family. If pregnancy is continued, mode of delivery and provision for care of a child with possibly coexisting anomalies should be considered before labor and delivery. Operative treatment is required in all patients with omphalocele. The size of the omphalocele determines whether a primary repair or delayed primary closure is selected. Escharification of the intact omphalocele sac has been used. Reoperation is necessary in up to 25% of cases of omphalocele, either for reclosure of stomas or for subsequent bowel obstruction.

Gastroschisis

Gastroschisis is an abdominal wall defect most commonly located to the right of an intact umbilical cord (Fig. 96-16). The incidence of gastroschisis is approximately 1 in 10,000 births overall, but approaches 7 in 10,000 among mothers younger than 20 years of age. Gastroschisis occurs more frequently in whites and in Hispanic infants than in other races or ethnicities. In gastroschisis, a sac is absent, and the extruded bowel is “padded” and thickened along its length from its extended exposure to the amniotic fluid. Histologically, the bowel usually is normal. Atresia occurs in 10% to 15% of children with gastroschisis. Almost all infants with gastroschisis also exhibit malrotation. Whereas prematurity is more common in children born with gastroschisis than it is in children with omphalocele, extraintestinal anomalies are much more common with omphalocele than they are with gastroschisis. The morbidity and mortality in patients with gastroschisis are largely related to intestinal atresia; other congenital anomalies also have been reported in a small number of patients.21,22 Gastroschisis may be complicated by necrotizing enterocolitis, with all its attendant short-term and long-term complications.

Persistence of the ductal communication between the intestine and the yolk sac beyond the embryonic stage may result in several anomalies of the omphalomesenteric (vitelline) duct (Fig. 96-17) including (1) a blind omphalomesenteric duct, or Meckel’s diverticulum; (2) a central cystic dilatation in which the duct is closed at both ends but patent in its center, an omphalomesenteric or vitelline cyst; (3) an umbilical-intestinal fistula (see Fig. 96-17A), resulting from the duct remaining patent throughout its length; and (4) complete obliteration of the duct, resulting in a fibrous cord or ligament extending from the ileum to the umbilicus, as an omphalomesenteric band.24 In approximately 1% to 4% of all infants, a remnant of the embryonic yolk sac is retained, making the omphalomesenteric or vitelline duct the most common site of congenital gastrointestinal anomaly. Between the fifth and seventh weeks of gestation, the omphalomesenteric duct, which has connected the embryo to the yolk sac, attenuates, involutes, and separates from the intestine. Before this separation, the epithelium of the yolk sac develops an appearance similar to that of the gastric mucosa. Under normal circumstances the omphalomesenteric duct becomes a thin fibrous band that fragments and is absorbed spontaneously during the fifth to tenth week of gestation. Partial or complete failure of involution of the duct results in the variety of retained structures described above. A Meckel’s diverticulum is an antimesenteric outpouching of the ileum that usually is found approximately 2 feet from the ileocecal junction (see Fig. 96-17B). It occurs in 1.2% to 2% of the population and has a male-to-female ratio of 3 : 1.25 Meckel’s diverticula account for 67% of all omphalomesenteric duct remnants.24 Length of the diverticulum varies, ranging from 1 to 10 cm. Ectopic gastrointestinal mucosa—duodenal, gastric, biliary or colonic, or aberrant pancreatic tissue—is present in about 50% of Meckel’s diverticula; most common is ectopic gastric mucosa, accounting for 80% to 85% of all Meckel’s diverticula– associated ectopic tissue (see Fig. 96-17C). Painless bleeding per rectum is the most common manifestation of a Meckel’s diverticulum. Blood in the stool usually is maroon, even in patients with massive bleeding and hypovolemic shock. Bright red blood per rectum, as might be seen with bleeding from the left colon, is almost never encountered, but melena may be seen in patients with intermittent or continual, less severe bleeding. The cause of bleeding is peptic ulceration secondary to acid production by the ectopic gastric mucosa within the Meckel’s diverticulum. A “marginal” ulcer often develops at the junction of the gastric and ileal mucosae. Although Helicobacter pylori has been observed in the gastric mucosa within a Meckel’s diverticulum, a relationship between bleeding from a Meckel’s diverticulum and presence of this organism is unlikely. Despite massive bleeding, death seldom, if ever, occurs in children from complications of a Meckel’s diver-

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine Meckel’s diverticulum

Vitelline cyst

Umbilicus

Vitelline ligament

Vitelline ligaments

Vitelline fistula

A

C

B 5 mins

10 mins

15 mins

20 mins

25 mins

30 mins

D

Figure 96-17.  Vitelline duct abnormalities and features of Meckel’s diverticulum. A, Schematic representations of a Meckel’s diverticulum, vitelline cyst, and vitelline fistula. B, Surgical specimen revealing an outpouching of the ileum (Meckel’s diverticulum). C, Photomicrograph showing replacement of the small intestinal mucosa by ectopic oxyntic mucosa lining a Meckel’s diverticulum. (Hematoxylin and eosin, ×150.) D, Meckel’s diverticulum scan demonstrating uptake of 99mtechnetium-pertechnetate (arrows) by the diverticulum by 10 minutes. (A from Sun B, editor. Langman’s Medical Embryology. 9th ed. Philadelphia: Lippincott Williams & Wilkins; 2004.) (See text for details.) (D, Courtesy of Dr. I. Zanzi.)

ticulum. Once hypovolemia occurs from blood loss, the splanchnic blood vessels contract, and bleeding tends to diminish or cease. Intestinal obstruction is the next most common manifestation of a Meckel’s diverticulum. This obstruction is caused either by intussusception with the diverticulum as a lead point or by herniation through or volvulus around a persistent fibrous cord remnant of the vestigial vitelline duct. In children older than 4 years of age, intussusception almost always is secondary to a Meckel’s diverticulum. However, diverticulum-related intestinal obstruction may occur at any age. Volvulus around a vitelline cord has been described in the neonatal period. Bilious vomiting and abdominal distention usually are the initial signs of obstruction. Intestinal obstruction in these patients, as with other causes of obstruction, can lead to intestinal ischemia and death. Diverticulitis of a Meckel’s diverticulum occurs as a result of acute inflammation. Most commonly, affected patients are diagnosed as having acute appendicitis, and the diagnosis of Meckel’s diverticulitis is made at exploratory laparotomy. Perforation occurs in approximately one third of patients with Meckel’s diverticulitis and may result from peptic ulceration.26 A chronic form of Meckel’s diverticuli-

tis (Meckel’s ileitis) may mimic the presentation of Crohn’s disease of the ileum. Meckel’s diverticulum may be an incidental finding.25 The presence of a Meckel’s diverticulum always should be considered in an infant or child with significant painless rectal bleeding. Standard abdominal plain films, barium contrast studies, and ultrasonographic imaging rarely are helpful in making the diagnosis. Because bleeding almost always is from ectopic gastric mucosa within the diverticulum, the Meckel’s scan, which allows imaging of the gastric mucosa, should be the initial diagnostic study (see Fig. 96-17D). Uptake of the 99mTc-pertechnetate is by the mucussecreting cells of the gastric mucosa, not the parietal cells. Unfortunately, this study has only 85% sensitivity and 95% specificity. When the diagnosis of a bleeding Meckel’s diverticulum is entertained and the Meckel’s scan is negative, splanchnic angiography and 99mTc-labeled red blood cell studies may be used; however, diagnosis is usually made at surgery. It is reasonable to perform esophagogastroduodenoscopy and colonoscopy to rule out other possible etiologic disorders. Although symptomatic Meckel’s diverticulum is far more common in pediatric patients, it may occur in adults.

1629

1630

Section X  Small and Large Intestine Omphalomesenteric (or Vitelline) Cyst

Omphalomesenteric (or vitelline) cyst is more common in men and is characterized by a mucosa-lined intestinal cystic mass within the center of a fibrous cord.24 Infection of the cyst or intestinal obstruction may result.

Patent Omphalomesenteric (Vitelline) Duct

Patent omphalomesenteric (vitelline) duct represents a persistent connection between the distal ileum and the umbilicus. This fistula has a male-to-female ratio of 5 : 1, and accounts for 6% to 15% of omphalomesenteric duct remnants. The diagnosis usually is made in the first few weeks of life after separation of the umbilical cord from the newborn umbilicus. Foul-smelling discharge from the umbilicus occurs.27 Examination of the umbilicus reveals either an opening or a polypoid mass resulting from limited prolapse of the patent omphalomesenteric duct. Definitive diagnosis can be made by fistulography. Complications of this type of fistula include prolapse of the patent duct, or of the duct and the attached ileum, through the umbilicus, which may lead to partial intestinal obstruction. Prolapse should not be mistaken for an umbilical polyp, because excision of involved tissue might result in perforation. Resection is warranted.27

Omphalomesenteric Band

In omphalomesenteric band, the solid cord connecting the ileum to the umbilicus remains intact. This cord may result in intestinal obstruction from an internal hernia or volvulus.

Vitelline Blood Vessel Remnants

Failure of involution of vitelline blood vessel remnants results in complications similar to those seen with a retained fibrous cord within the peritoneal cavity. Intestinal obstruction occurs when a portion of the small intestine wraps itself around the band. Treatment of all vitelline duct abnormalities is surgical.

MALROTATIONS

Rotation defects result from errors in the normal embryonic development of the midgut, which gives rise to the distal duodenum, jejunum, ileum, cecum, and appendix, as well as the ascending colon and proximal two thirds of the transverse colon. Aberrations in midgut development may result in a variety of anatomic anomalies, including (1) rotation and fixation, (2) atresias and stenoses, (3) duplications, and (4) persistence of embryonic structures. These congenital anomalies may cause symptoms not only in the newborn or neonatal period, but also later in childhood and adulthood. Therefore, congenital anomalies of the midgut are considerations in the differential diagnosis of intestinal obstruction and ischemia in patients of all ages. Because anomalies of intestinal rotation may remain asymptomatic throughout life, their true incidence is unknown; a prevalence of 1 in 500 live births has been reported.28 Symptoms usually manifest within the first month of life, with bilious emesis and abdominal distention, but presentation may be delayed in mild cases to the fourth decade of life. Patients may have cramping abdominal pain, vomiting, diarrhea, abdominal tenderness, and blood or even mucosal tissue in the stool from ischemia. If ischemia is allowed to progress, peritonitis and hypovolemic shock may develop, potentially culminating in death. Delay in surgery in patients with ischemic injury may result in a short bowel, necessitating chronic total parenteral nutrition therapy and eventually small bowel transplantation, with or without liver transplantation. Most adult

patients with anomalies of intestinal rotation have chronic symptoms for several months or years before diagnosis.

Classification

Anomalies of rotation usually are characterized by the stage in the rotational process at which normal embryonic development of the midgut has been interrupted. Most anomalies of midgut rotation occur during the second stage of rotation and have been characterized as nonrotation, reverse rotation, and malrotation (Fig. 96-18). Of these, nonrotation is most common and reflects a complete failure of the second stage of rotation. With this anomaly the intestinal tract occupies the same position in the abdomen as it does in an eight-week-old embryo; the small intestine is located to the right of the midline and the colon is positioned to the left. Defects in the first and third stages of rotation are uncommon. Abnormalities in the first stage are associated with extroversion of the cloaca; abnormalities of the third stage cause failure of cecal elongation, and the cecum remains in the right upper quadrant. In adults, reverse rotation of the midgut loop is the most commonly diagnosed defect of the midgut. Reverse rotation of the midgut loop is rare, however, and accounts for only 4% of all rotational anomalies. In reverse rotation, the midgut rotates 180 degrees clockwise during the second stage of rotation, resulting in a net 90 degrees of clockwise rotation. This may produce either the retroarterial colon type (the colon is located behind the superior mesenteric artery) or the liver and entire colon ipsilateral type of reverse rotation. Malrotation of the midgut loop, a developmental anomaly of intestinal fixation and rotation, occurs when the proximal midgut fails to rotate around the mesenteric vessels during the second stage of rotation. The distal midgut does rotate 90 degrees in a counterclockwise direction, however, with the result that the jejunum and ileum remain to the right of the superior mesenteric artery and the cecum is situated in the subpyloric region. With the potential for the small intestine and cecum to twist around the superior mesenteric artery and each other, this is the rotation anomaly in adults most frequently associated with ischemic damage, therefore mandating surgical correction.

Associated Abnormalities

Associated anomalies are seen in 30% to 60% of patients with defects in intestinal rotation. Nonrotation of the midgut is a significant finding in patients with omphalocele, gastroschisis, and diaphragmatic hernia. Rotation defects are seen in approximately 30% to 50% of infants with duodenal or jejunal atresia and in 10% to 15% of children with intestinal pseudo-obstruction; they also are associated with a variety of other conditions including Hirschsprung’s disease, esophageal atresia, biliary atresia, annular pancreas, meconium ileus, intestinal duplications, mesenteric cysts, Meckel’s diverticulum, urologic anomalies, and imperforate anus.29 Anomalies of rotation can cause acute or chronic intermittent obstruction from volvulus (see Fig. 96-18D,E). Venous and lymphatic obstruction, also from volvulus, can lead to malabsorption and abnormalities in intestinal motility. Patients may fail to thrive and present with chylous ascites and other symptoms and signs of lymphangiectasia resulting from chronic lymphatic obstruction. Duodenal obstruction can occur as the result of midgut volvulus, and as the result of peritoneal bands between a malpositioned cecum in the subpyloric region and the peritoneum. These bands, called Ladd’s bands, cross the second

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine

Duodenum

Duodenum

Ladd’s bands

Transverse colon

Ascending colon

Jejunoileal loops

Transverse colon Descending colon

A

B

Cecum

D

Jejunoileal loops

Descending colon

C

E

Figure 96-18.  Rotation defects. A and B, Two examples of nonrotation. A, Ladd’s bands are seen crossing the duodenum; some authors would refer to this as a “mixed rotation.” B, In nonrotation, the small intestine is located to the right of the midline, and the colon is to the left of the midline. C, Reverse rotation. The transverse colon passes behind the duodenum. D, Malrotation with volvulus characterized by a clockwise twist of the mesentery and strangulation. E, Radiologic appearance of malrotation depicting the duodenum to the right of the spine, with a volvulus. A and B, from Gosche JR, Touloukian J. Congenital anomalies of the midgut. In: Wyllie R, Hyams JS, editors. Pediatric Gastrointestinal Disease: Pathophysiology, Diagnosis, Management. 2nd ed. Philadelphia: WB Saunders; 1999. C, From Sun B, editor. Langman’s Medical Embryology. 9th ed. Philadelphia: Lippincott Williams & Wilkins; 2004. D, From Netter FH. The Netter Collection of Medical Illustration. vol 3. Teterboro, NJ: Icon Learning System; 2002. (D, Courtesy of Dr. J. Levenbrown.)

or third portion of the duodenum and cause obstruction by compression or kinking. Ladd’s bands are an anomaly of peritoneal embryogenesis and persist throughout life.

Diagnosis and Management

If time allows, diagnosis can be made by upper gastrointestinal contrast examination and delineation of the site of the duodenojejunal junction. Findings on ultrasonography may suggest malrotation if the superior mesenteric vein is seen located to the left of the superior mesenteric artery, in contradistinction to the normal anatomy. In the child with acute onset of bilious vomiting and peritoneal signs, no diagnostic studies should be performed if they delay surgical intervention. In the full-term infant with bilious emesis, anomalies of rotation should be considered first and foremost, to avoid the morbidity and mortality associated with these lesions. Ladd’s procedure, including dividing

Ladd’s bands, if present, widening of the mesentery, appendectomy, and fixation of the small intestine on the right and the colon on the left side of the abdomen, is the operation of choice.30

PROLIFERATION Enteric Duplication

Enteric duplications are rare with an incidence of 1 in 4500 births. Enteric duplications are either tubular or spherical; the tubular type communicates with the normal intestinal tract, whereas the spherical type does not. Tubular duplications may join the intestine at one or at both of its ends. Except for duodenal duplications, duplications occur on the mesenteric side of the bowel, and a common blood supply and muscular coat are shared by the duplicated segment and the adjacent bowel. Duplication cysts may be completely

1631

1632

Section X  Small and Large Intestine isolated and have their own exclusive blood supply. Small bowel duplications often contain pancreatic tissue or gastric mucosa; the latter can be diagnosed by 99mTc radioisotopic imaging.31 The etiology of duplications is unclear, but may involve a defect in intestinal recanalization. Enteric duplications occur throughout the gastrointestinal tract but are most common in the ileum.31 Gastric duplications occur least commonly. Depending on the site of the duplication, and whether ectopic gastric mucosa is present (seen in approximately 50% of the cases), complications include intestinal hemorrhage, ulceration, perforation, intestinal obstruction, volvulus, intussusception, infection, pancreatitis, jaundice, hematobilia, and cutaneous enteric fistulas. Duplication of the rectum is the most common of the large bowel duplications and may be associated with constipation or obstipation. Colonic duplications frequently involve the entire colon. Occasionally, large bowel duplications affect several segments of the colon, leaving “skip areas” of normal colon. A high percentage of children with duplications have associated malfor­mations. Adenocarcinoma, neuroendocrine carcinoma, and squamous carcinoma have been documented with gastric, small bowel, and colonic duplications,31,32 and carcinoid has been described in duplications of the rectum. Neuroenteric cysts attach posteriorly to the spinal cord, are associated with asymptomatic hemivertebrae, and may occur at any level of the gastrointestinal tract. An intra-abdominal mass may be appreciated in a child with intestinal duplication, either by abdominal palpation or on rectal examination. Stool may contain occult blood from ulcerated ectopic gastric mucosa or ischemic damage. Other symptoms and signs include abdominal distention, constipation, vomiting, and respiratory distress.33 Generalized peritonitis can be the first manifestation of a perforated duplication cyst. In adults, acute abdomen, intra-abdominal mass, symptoms of colonic diverticulitis and chronic abdominal pain have been observed.34 Preoperative diagnosis by radiographic evaluation is problematic, but radioisotope studies may prove diagnostic if ectopic mucosa is present in sufficient quantities.

INTESTINAL ATRESIA AND STENOSIS

Of all of the congenital anomalies of the midgut, atresias and stenoses occur most frequently. Intestinal atresia refers to a congenital complete obstruction of the intestinal lumen, whereas stenosis indicates a partial or incomplete obstruction. Atresias occur more commonly than do stenoses, and small bowel atresias have a reported incidence rate of 1 in 1500 live births.35 Small bowel atresias are more common in black infants, low birth weight infants, and twins. Jejunoileal atresias are distributed equally throughout the jejunum and ileum, and multiple atresias are found in up to 20% of children. Colonic atresia occurs infrequently and accounts for less than 10% of all atresias. In the duodenum, atresia results from failure of recanalization of the solid stage of duodenal development, whereas in the remaining small intestine and colon, atresia is the result of intestinal ischemia. Evidence of a vascular “accident” is noted in 30% to 40% of infants with atresia; proposed mechanisms include volvulus, constriction of the mesentery in a tight abdominal wall defect such as gastroschisis, internal hernia, intussusception, and obstruction with perforation. Jejunoileal atresia may follow maternal use of ergotamine (in Cafergot for headaches) or cocaine taken during pregnancy and also is associated with congenital rubella. Atresias also may result from low-flow states and placental insufficiency35; in such cases, evidence of a vas-

Figure 96-19.  Plain film of the abdomen showing a “double bubble,” typical of duodenal atresia. The larger bubble is the gastric bubble; the smaller bubble is the duodenal bubble. (Courtesy of Dr. J. Levenbrown.)

cular accident will be absent. Absence of fibroblastic growth factor 10 may result in intestinal atresia.36,37 In familial cases of jejunoileal atresia there is probably a disruption of a normal embryonic pathway, making this type of atresia a true embryologic malformation rather than an acquired lesion.38 Duodenal obstruction may result from atresia (40% to 60%), stenosis (35% to 40%), or an intestinal web (5% to 15%). Eighty percent of these atresias are contiguous with or distal to the ampulla of Vater, and virtually all webs are within a few millimeters of the ampulla. Atresias may be multiple. The incidence of duodenal obstruction varies, ranging from 1 in 10,000 to 20,000 live births. About 25% of patients with duodenal atresia are born preterm. Stenosis most often is due to extrinsic duodenal obstruction from an annular pancreas. Other anomalies that may cause duodenal obstruction in children with malrotation are Ladd’s bands, an anterior or preduodenal portal vein, or aberrant intramural pancreatic tissue. Clinically, the presentation is that of a proximal intestinal obstruction with bilious vomiting on the first day of life, usually without abdominal distention. With gastric dilatation, the epigastrium may appear to be full by inspection and palpation. Excessive retention of gastric bile–stained fluid is typical. Duodenal obstruction is diagnosed easily by abdominal films revealing a typical “double bubble” sign with a paucity of small intestinal air (Fig. 96-19). Mothers of infants with duodenal obstruction often have polyhydramnios, and uterine ultrasonography may even demonstrate a double bubble in the unborn fetus. Vomiting, abdominal distention, delayed meconium passage, and jaundice are more frequent with jejunoileal than duodenal atresia.39 The classification system of Grosfeld and colleagues comprises five different types of jejunoileal and colonic atresias

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine I

IIIa

II

IIIb

IV

Figure 96-20.  Classification of jejunoileal atresias. Type I: The mucosa and submucosa form a web or intraluminal diaphragm, resulting in obstruction. A defect in the mesentery is not present, and the intestine is not shortened. Type II: The dilated proximal intestine has a bulbous blind end connected by a short fibrous cord to the blind end of the distal intestine. The mesentery, however, is intact, and the overall length of the small bowel is not usually shortened. Type IIIa: The defect in type IIIa is similar to that in type II in that both types have blind proximal and distal ends. In type IIIa, however, complete disconnection exists. In addition, a V-shaped mesenteric defect is present. The proximal blind end is usually markedly dilated and not peristaltic. The compromised intestine undergoes intrauterine absorption, and, as a result, the intestine is shortened. Type IIIb: In addition to a large defect of the mesentery, the intestine is significantly shortened. This lesion is also known as Christmas tree deformity because the bowel wraps around a single perfusing vessel, like the tinsel coil wrapped around a Christmas tree; it also has been called an apple-peel deformity. The distal ileum receives its blood supply from a single ileocolic or right colic artery because most of the superior mesenteric artery is absent. Type IV: Multiple small intestinal atresias are present in any combination of types I to III. This defect often takes on the appearance of a string of sausages because of the multiple lesions. (From Grosfeld JL, Ballantine TVN, Shoemaker R. Operative management of intestinal atresia and stenosis based on pathologic findings. J Pediatr Surg 1979; 14:368.)

(Fig. 96-20).39a In the “apple-peel” atresia or “Christmas tree” deformity (type IIIb), proximal atresia with wide separation of the bowel loops is associated with absence of the distal superior mesenteric artery. The distal ileum receives its blood supply by retrograde perfusion through the ileocolic artery. Type IIIb atresias account for less than 5% of all atresias. Atresias are far more common than stenoses, with a frequency ratio of 15 : 1. With the exception of multiple atresias and perhaps the apple-peel atresia, heredity appears to be of little significance in most cases. Approximately 50% of children with duodenal atresia have associated malformations. Of this group, 30% have Down syndrome.39 Major anomalies occur less frequently with jejunoileal atresias and colonic atresias than with duodenal atresia. The most common anomalies are malrotation, volvulus, and gastroschisis, all of which can cause intestinal ischemia in utero.40 Extragastrointestinal anomalies associated with atresias include cardiovascular, pulmonary, and renal malformations, and skeletal deformities. Prematurity is common, ranging in incidence from 25% in ileal atresias to 40% in jejunal lesions; 50% percent of babies with multiple atresias are born prematurely. If the obstruction occurs beyond the ampulla of Vater, bilious or feculent vomiting with abdominal distention is seen. The presence of meconium in the colon is uncommon at surgery, but variable amounts may be noted. With distal obstruction, abdominal films may demonstrate multiple dilated air-filled bowel loops. If perforation has occurred in utero, extraluminal air and intraperitoneal calcifications or calcifications within the scrotal sac may be present, suggesting meconium peritonitis. A “soap bubble” appearance of the ileum may suggest meconium ileus (cystic fibrosis). Air-fluid levels

rarely are seen in meconium ileus. Prenatal ultrasonographic findings in jejunoileal atresia include dilated bowel and polyhydramnios.41 Considerations in the differential diagnosis of distal bowel obstruction include small intestinal and colonic atresias, meconium ileus, Hirschsprung’s disease, and meconium plug with or without small left colon syndrome. In the small left colon syndrome, the descending and sigmoid colon are narrowed, usually with a caliber transition at or near the splenic flexure. Typically, neonates are born to mothers with gestational diabetes and may experience resolution of obstruction without operation. Contrast studies of the colon are helpful in making a proper diagnosis. An upper gastrointestinal contrast study may provide additional important information. Surgery is required to relieve the intestinal obstruction in the atretic or narrowed segment. Postoperative complications include fluid and electrolyte disorders, nutritional and feeding problems from diarrhea due to short bowel and small bowel failure, and failure to thrive.

ANORECTUM

Anorectal malformations comprise a wide spectrum of diseases that can involve the male and female anus and rectum as well as the urinary and genital tracts.42 Anorectal malformations occur in 1 in 4000 to 5000 newborns and are more common among boys and in children with Down syndrome.43 During normal development, after appearance of the urorectal septum, migration of the primitive anus down the posterior wall of the cloaca may occur. Some experts postulate that a craniocaudal fusion of the lateral urorectal

1633

Section X  Small and Large Intestine

Type 2. Pouch ≤1.5 cm from the anal dimple

Type 3. A blind pouch >1.5 cm from the anal dimple

Type 4. Atresia of the rectum with a normal anus

In females

A

Type 1. A thin membrane over the anus

Rectovaginal

Rectofourchet

Rectoperineal

In males

1634

B

Rectovesical

Rectourethral

Rectoperineal

Figure 96-21.  Anorectal malformations. A, Types of imperforate anus. B, Types of associated fistulas. (From Netter FH. The Netter Collection of Medical Illustration. vol 3. Teterboro, NJ: Icon Learning System; 2002.)

ridges occurs from the walls of the cloaca. Migration of the anus is completed when the urorectal septum reaches the perineum. Anorectal malformations during the fourth to twelfth weeks of gestation are believed to result from failure of migration of the anus and excessive fusion. Vascular accidents, maternal diabetes, and maternal ingestion of thalidomide, phenytoin, and trimethadione all have been proposed causes. Defective development of the dorsal cloaca also has been implicated44 and distal 6q deletions have been reported in sacral or anorectal malformations.45 Alteration in Shh signaling also may play a role in producing abnormal notochord development and sacral or anorectal malformations.46,47 Anorectal malformations may occur with higher frequency in infants born after in vitro fertilization.48

Different types of anorectal malformations are illustrated in Figure 96-21. Anorectal malformations are divided into low (infra- or translevator), high (supralevator), and intermediate categories. A functional and practical classification of these malformations, the Wingspread classification, is summarized in Table 96-3A. The classification in Table 96-3B is designed, according to Pena,49 to increase the physician’s awareness of the possibility of the presence of these lesions, as well as to establish therapeutic priorities.

Anocutaneous Fistula

In anocutaneous fistula, the rectum traverses normally through most of the anal sphincter, but its lower portion

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine Table 96-3  Classifications of Anorectal Malformations Wingspread Classification MALE

FEMALE

Low* Anocutaneous fistula Anal stenosis

Anovestibular fistula Anal stenosis Anocutaneous fistula

†

Intermediate Anal agenesis without fistula Rectobulbar urethral fistula High‡ Anorectal agenesis   With rectoprostatic urethral fistula   Without fistula Rectal agenesis

Anal agenesis without fistula Rectovaginal fistula Rectovestibular fistula Anorectal agenesis With rectovaginal fistula Without fistula Cloaca

Classification Based on the Need for Colostomy49 MALE

FEMALE

Colostomy Not Required Perineal (cutaneous) fistula Colostomy Required Rectourethral fistula   Bulbar   Prostatic Rectovesical fistula Imperforate anus without fistula

Colostomy Not Required Perineal (cutaneous) fistula Colostomy Required Vestibular fistula

Rectal atresia

Persistent cloaca Imperforate anus without fistula Rectal atresia

*Low: infra-, or translevator. † Intermediate: between high and low. ‡ High: supralevator.

deviates anteriorly and ends as a perineal, cutaneous fistula anterior to the center of the external anal sphincter (anocutaneous or perineal fistula). These anomalies are similar in the male and the female child. Perineal fistula is the most benign of anorectal defects, and associated urologic defects are uncommon (10%). All patients achieve bowel control after proper surgical treatment. Examination of the perineum may demonstrate features indicative of a perineal fistula, including a prominent midline skin ridge (“bucket-handle” malformation) and subepithelial midline raphe fistula having the appearance of a black ribbon because of its meconium content. Surgery consists of a simple anoplasty, usually done without a protective colostomy.

Rectourethral Fistula

In rectourethral fistula, by far the most frequent anorectal malformation in male children, the rectum descends through a portion of the pelvic floor musculature but focally deviates anteriorly and communicates with the posterior urethra. This fistula may end in the lower posterior (bulbar) or in the upper posterior (prostatic) urethra.49 Prenatal echogenic calcifications within the bowel, due to a mixture of meconium and urine, should suggest an anorectal malformation with rectourinary fistula and bladder outlet obstruction.50 Children with prostatic urethral fistulas more commonly have sacral and urologic defects (60%) than do children with bulbar prostatic fistula (30%). Eighty-five percent of children with rectourethral bulbar fistula achieve fecal continence after repair, compared with 60% of children with rectoprostatic fistula.

Rectovesical Fistula

In rectovesical fistula, the most proximal anorectal defect in male children, the rectum opens into the bladder neck. Ninety percent of these malformations are associated with significant urologic defects, and only 15% of children achieve bowel control after surgical repair.

Vestibular Fistula

In vestibular fistula, the most common anorectal defect of female children, the rectum opens into the vestibular bulb of the clitoris. The vesticular bulbs are erectile structures situated on either side of the vulvovaginal orifice. The rectum and the vagina share a thin common wall. Thirty percent of affected children have associated urologic defects, and approximately 90% of these children achieve bowel control after surgery. In the vaginal fistula, the rectum opens in the lower or, less frequently, the upper half of the vagina.

Anorectal Agenesis (Imperforate Anus) Without Fistula

In anorectal agenesis the rectum ends blindly without a fistula approximately 1 to 2 cm above the perineum. Sphincter function usually is preserved, with 80% of these patients achieving bowel control after surgery. Approximately 50% of children with imperforate anus have Down syndrome. Conversely, 95% of children with Down syndrome who have anorectal malformations will have this specific type of defect.

Rectal Agenesis (Atresia)

Rectal agenesis occurs more frequently in female than in male children, and consists of complete (atresia) or partial (stenosis) interruption of the rectal lumen between the anal canal and the rectum. On visual inspection of the perineum, the anus appears normal; however, an obstruction can be found 1 to 2 cm above the mucocutaneous junction of the anus. Sphincter function is normal in these patients, and associated urologic defects are rare. In these children, prognosis is excellent, with 100% achieving full bowel control after anorectoplasty.

Anal Stenosis

Anal stenosis, a fibrous ring located at the anal verge, causes constipation and gives the stool a ribbon-like appearance. Response to dilation or surgical disruption is excellent.

Persistent Cloaca

In the complex defect of persistent cloaca, the rectum, vagina, and urethra are fused into a single common channel that opens into one perineal orifice situated at the site of what should be the opening of the normal urethra. Prognosis depends on the intactness of the sacrum and the length of the common channel. Prognosis is better in those children with a shorter common channel (less than 3 cm) than in those with a common channel longer than 3 cm; the latter have a higher incidence of urologic anomalies.51 Associated urologic problems are an important consideration with persistent cloaca. For example, urologic emergencies from obstructive uropathy are common, and hydrocolpos may compress the opening of the ureters, resulting in bilateral megaureters and massive vesicoureteral reflux.

Associated Abnormalities

Other associated abnormalities have been reported in 70% of children with anorectal malformation (Table 96-4).42,43 Anorectal malformations occur in malformation syndromes and with chromosomal anomalies.43,52 The higher and more complex the anorectal defect, the greater the chance of severe urologic anomalies (72%);

1635

1636

Section X  Small and Large Intestine Table 96-4  Common Abnormalities Associated with Anorectal Malformations Cardiovascular Atrial septal defect Dextrocardia Pulmonary stenosis Tetralogy of Fallot Ventricular septal defect Central nervous system Aqueductal stenosis Cerebral atrophy Microcephaly Myelomeningocele Teratoma Chromosomal abnormalities Trisomy 13 Trisomy 18 Trisomy 21 Craniofacial Cleft palate Epicanthal folds Low-set ears Potter facies Simian creases Gastrointestinal Duodenal atresia Esophageal atresia Malrotation Tracheoesophageal fistula Genitourinary Ambiguous genitalia Cryptorchidism Multicystic dysplastic kidney Renal agenesis Malformation sequences Caudal regression syndrome Malformation syndrome Cat’s-eye syndrome Opitz syndrome Potter syndrome type 1 Malformation associations VATER complex (vertebral defects, anal atresia, tracheoesophageal fistula with esophageal atresia, radial and renal anomalies) VATERL complex (vertebral, anal, cardiac, tracheal, esophageal, renal, and limb anomalies) Musculoskeletal Abnormal rib number Deformed or reduced number of sacral vertebrae Dislocated hip Hemisacrum Hemivertebra Micrognathia Omphalocele Polydactyly Respiratory Choanal atresia Diaphragmatic hernia Hypoplastic lungs Subglottic stenosis Data adapted from Cho S, Moore SP, Fangman T. One hundred three consecutive patients with anorectal malformations and their associated anomalies. Arch Pediatr Adolesc Med 2001; 155:587-91.

sacral abnormalities also are frequent. Children with a persistent cloaca or rectovesical fistula have a 99% chance of having an associated genitourinary anomaly, whereas less than 10% of children with low fistula have such abnormalities. Overall, patients with additional anomalies are more likely to have high lesions than are patients with isolated anorectal malformations.43 Boys with low and high anorectal malformation have a high incidence of genital and

gastrointestinal anomalies, whereas urologic anomalies are more frequent in girls with high anorectal malformations.53 Long-term bowel dysfunction occurs in one third of boys with perineal fistula. In the first 24 hours of life, a decision should be made whether a child needs a colostomy or simple anoplasty. The presence of an associated defect, either urologic or cardiac, that might be life threatening requires immediate evaluation. A cloaca with a common channel shorter than 3 cm can be repaired by posterior sagittal intervention, whereas a common channel longer than 3 cm requires a laparotomy.51

ENTERIC NERVOUS SYSTEM Hirschsprung’s Disease

Hirschsprung’s disease (HD) is due to a congenital absence of ganglion cells in both the submucosal (Meissner’s) and myenteric (Auerbach’s) plexuses. Aganglionosis extends continuously for a variable distance proximal to the internal sphincter. Short-segment HD is most common with a transition zone from aganglionic colon to ganglionic colon at the level of the sigmoid. In long-segment HD the entire colon and even the small intestine may lack ganglia. With an incidence of 1 in 5000 live births, approximately 700 new cases of HD occur each year in the United States. The incidence is lowest in Hispanic and highest in Asian individuals. Approximately 10% of babies with Down syndrome have HD. Deletion of 17q21 and other chromosomal anomalies also have been reported.54 Familial occurrence has been reported in about 7% of cases. Familial cases have a male predominance with an increased incidence of long-segment aganglionosis. Affected families carry a high risk of familial recurrence of long-segment HD.55 HD is seen most commonly in full-term infants but on occasion does occur in premature births. In the short-segment type, a 4 : 1 male preponderance is observed, and in the long-segment type, the ratio is reduced to about 2 : 1. Short-segment HD accounts for nearly 90% of cases in childhood, and long-segment HD accounts for the remainder. It is rare that ultrashort-segment HD manifests in the pediatric population, but it does explain certain cases of chronic constipation that come to attention in adulthood. Pathogenesis Two pathogenetic mechanisms have been proposed for HD: (1) failure of migration of neural cells and (2) alteration of the colonic microenvironment. Genetic, vascular, and infectious factors are invoked to explain these alterations. Failure of Migration.  Between the fifth and twelfth weeks of gestation, premature arrest of the craniocaudal migration of vagal neural cells will result in HD. Colonic Microenvironment Changes.  A basic defect in the microenvironment necessary for the migration, development, and survival of ganglion cells has been postulated. Levels of various substances such as laminin, nicotinamide adenine dinucleotide phosphate (NADPH)-diaphorase, and neural cell adhesion molecules, as well as other polypeptides, have been shown to be reduced in the aganglionic segment. Some investigators have postulated that an alteration in the extracellular matrix with decreased concentrations of laminin and collagen IV constitutes a barrier to neutrophin 3, thereby perhaps impairing the neuroblastic migration and colonization. Neutrophin 3 promotes survival of sympathetic and sensory neurons in vitro and supports the growth and survival of differing subsets of neurons.

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine Table 96-5  Genes Involved in Hirschsprung’s Disease GENE

chromosome LOCATION

INHERITANCE

PHENOTYPE

PENETRANCE OF HD TRAIT

RET GDNF NTN SOX10 EDNRB EDN3 ECE1 ZFHX1B (SIP1) PHOX2B TCF4

10q11.2 5p13 19p13 22q19 13q22 20q13 1p36 2q22 4p12 18q21

AD AD AD AD AR/AD AD AD AD AD AD

HD HD HD WS4 WS4/HD WS4/HD HD, CFD, CD MCA-MR CCHS Epileptic encephalopathy

70% (male), 50% (female) Low Low 80% Low 5% Low 60% 20% Low

AD, autosomal dominant; AR, autosomal recessive; CCHS, congenital central hypoventilation syndrome; CD, cardiac defect; CFD, craniofacial defect; HD, Hirschsprung’s disease; MCA-MR, multiple congenital anomalies–mental retardation syndrome; SIP1, Smad-interacting protein; WS4/HD, combination of Shah-Waardenburg syndrome with Hirschsprung’s disease (see Table 96-6). Data from Amiel J, Sproat-Emison E, Garcia-Barcelo M, et al. Hirschsprung disease, associated syndromes and genetics: A review. J Med Genet 2008; 45:1-14.

Nitric oxide synthase is reduced in the aganglionic segment in HD, explaining the failure of relaxation of the affected colonic segment. Isolated case reports have linked the destruction of ganglion cells in segmental HD to cytomegalovirus infection and muscular hyperplasia of pericolonic vessels. The genetics of HD have now been characterized.18 Inheritance of the disease can be autosomal dominant, autosomal recessive, or polygenic. Penetration of mutations generally is low and depends on the extent of aganglionosis in affected family members. RET (rearranged during transfection) mutation penetrance is incomplete and sex dependent. It appears that the mutation, although increasing a child’s odds of having HD, is not predictive of the specific abnormality. Alterations of several genes have been implicated (Table 96-5).56-59 RET, a proto-oncogene that codes for a receptor tyrosine kinase protein (c-Kit), is the major susceptibility gene in HD, and maps to chromosome 10q11.2. More than 100 mutations of this gene have been identified and reduced c-Kit levels in the colon of patients with HD have been observed.54 Identified gene mutations currently account for only approximately half of all cases of HD, but it is recommended that RET exon 10 mutation analysis be done in all children with HD18; germline RET mutations also can cause multiple endocrine neoplasia type IIA (MEN-IIA). Although the test results will be negative in the vast majority of cases, the significance of identifying MEN-IIA mutation carrier status for that individual and family appear to justify such testing.54 Mutation of the RET has been noted in familial and sporadic HD. Congenital birth defects are found 5% to 33% of patients with HD.54 Although HD usually occurs as an isolated event, in 30% of the patients it may be part of a syndrome (Table 96-6). Clinical Features Most children with HD should be diagnosed in the newborn nursery. Any full-term infant who does not pass meconium within the first 48 hours of life should be suspected of having this disorder. Frequently, such infants will have abdominal distention and feeding difficulties. They also may have bilious emesis from partial bowel obstruction. Dilation of the empty rectum by the first examiner usually results in the explosive expulsion of retained fecal material and decompression of the proximal normal bowel. HDassociated enterocolitis occurs more frequently in the first three months of life, in patients with delayed diagnosis, in

Table 96-6  Some Congenital Anomalies and Syndromes Associated with Hirschsprung’s Disease Congenital Anomalies Cardiac (5% of cases) Septal defects Central nervous system (4% of cases) Distal limbs Gastrointestinal (4% of cases) Meckel’s diverticulum Pyloric stenosis Small bowel atresia Genital (2%-3% of cases) Hypospadias Renal (4% of cases) Dysplasia Agenesis Sensorineural Skin Syndromes Congenital central hypoventilation MEN-II (medullary thyroid cancer, pheochromocytoma, parathyroid hyperplasia) Movat-Wilson (characteristic facies, microcephaly, mental retardation) Piebaldism (hypopigmentation of skin and hair) Shah-Waardenburg (regional hyperpigmentation, white forelock, bicolored irides, sensorineural deafness) Smith-Lemli-Opitz (anteverted nostrils, ptosis of eyelids, syndactyly of second and third toes, hypospadias and cryptorchidism in males) Syndromes with limb abnormalities (metaphyseal dysplasia, McKusick-type—mild bowing of legs, irregular metaphyses, fine sparse hair) MEN, multiple endocrine neoplasia. Data from Amiel J, Sproat-Emison E, Garcia-Barcelo M, et al. Hirschsprung disease, associated syndromes and genetics: A review. J Med Genet 2008; 45:1-14.

children with trisomy 21, and with long-segment involvement; girls and children with a positive familial history also are more frequently affected. Enterocolitis may develop due to ischemia from colonic distention proximal to the aganglionic segment, with secondary infection from colonic bacteria; cases also have been reported of HD-associated enterocolitis in the aganglionic segment; C. difficile has been isolated in children with this enterocolitis. Mortality rates of up to 30% have been reported for enterocolitis, which remains the major cause of death in HD. Colonic

1637

1638

Section X  Small and Large Intestine perforation, most frequently involving the cecum and rarely the appendix, may occur, even in utero. Most commonly, infants younger than six months of age with HD will continue to have variable but significant constipation, punctuated by recurrent obstructive crises or bouts of fecal impaction, often with failure to thrive. The abdomen may be distended with fecal masses, and peristaltic waves may be visible. Anemia and hypoalbuminemia are common. Blood-flecked diarrhea should suggest the presence of enterocolitis, and immediate evaluation should be undertaken. As the child with HD grows older, problems continue, and fecal soiling occasionally may occur. An infant with HD who is breast-fed may have fewer difficulties with defecation because the high concentration of lactose in breast milk causes watery stools that are passed more easily. Once breast milk is discontinued, symptoms of HD may worsen. Diagnosis The child with symptomatic HD usually demonstrates signs and symptoms of bowel obstruction. The diagnosis may be made by one or a combination of the following tests: contrast enema, rectal biopsy, and anal manometry. Flexible sigmoidoscopy plays a complementary role in diagnosis. A contrast enema performed on an unprepared colon will show the distal narrowed hypertonic segment of bowel (usually seen best in a lateral projection). The transition zone between the narrowed distal and dilated proximal

intestine will be seen in the most common form of HD—the rectosigmoid form (Fig. 96-22A)—but may not be seen with long- or ultrashort-segment intestinal involvement. In ultrashort-segment HD, a radiologic picture indistinguishable from that in functional constipation with dilated bowel extending to the anus usually is seen. The transition zone may not be evident in rectosigmoid HD if the patient has undergone cleansing enemas or colonic irrigation before the study. Although it has been suggested that the transition zone may not be evident in the first six weeks of life, it almost always is noted in the neonate with partial bowel obstruction. Flexible sigmoidoscopy typically reveals a normal but empty rectum. The dilated proximal bowel, if within reach of the scope, is traversed easily, unless there is abundant feces in the lumen; occasionally stercoral ulcers may be seen. Anal manometry is the most reliable method by which the gastroenterologist can make the diagnosis of ultrashortsegment HD. A normal physiologic response to distention of the rectum is relaxation of (smooth muscle) internal sphincter pressure. In HD, not only does rectal distention fail to induce internal sphincter relaxation, but a paradoxical rise in external sphincter pressure often is seen (see Fig. 96-22B). Sufficient volumes of air must be used to stimulate rectal distention for a reliable study. A false-positive result most commonly is caused by a capacious rectum in a constipated child or with megacolon, in which case balloon

External sphincter

External sphincter

Internal sphincter

Internal sphincter

dg

10 mm Hg 5 sec

na

A

B

Rectal balloon 50 mL air

Rectal balloon 50 mL air

nt

C

D

Figure 96-22.  Hirschsprung’s disease. A, Film from a barium enema examination showing the transition zone between the narrowed distal aganglionic segment (na) and the proximal dilated ganglionic segment (dg). B, Anal manometry. Left tracing illustrates normal function. In the right tracing note the lack of relaxation of the internal sphincter on rectal distention in a patient with Hirschsprung’s disease. C, Photomicrograph of a rectal suction biopsy specimen showing the absence of ganglion cells and the presence of thickened nerve trunks (nt) characteristic of Hirschsprung’s disease. (Hematoxylin and eosin, ×125.) D, Acetylcholinesterase-positive fibers stained brown (arrows) in the muscularis mucosae and lamina propria, ×250. (A, Courtesy of Dr. J. Levenbrown; B, from Markowitz J. Gastrointestinal motility. In: Silverberg M, Daum F, editors. Textbook of Pediatric Gastroenterology. 2nd ed. Chicago: Year Book Medical Publishers; 1988.)

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine distention may not stimulate the reflex. Up to 20% of normal children have a falsely absent reflex, especially if they are premature or of low birth weight. Nonetheless, a positive response such as internal sphincter relaxation is strong evidence against HD. Suction biopsy of the rectal mucosa is the most reliable method of diagnosis, except in patients with ultrashortsegment HD. The biopsy capsule should be placed at least 2 cm above the mucocutaneous junction in infants and 3 cm above the junction in older children to avoid the physiologic hypoganglionic zone. To be certain of the absence of ganglion cells in the submucosal plexus, an experienced pathologist may need to review many serial sections. Hyperplastic sympathetic nerve fibers and pro­ liferating Schwann cells are associated findings (see Fig. 96-22C), but can be absent in total aganglionosis. Controversy exists regarding the type of stains necessary to make a diagnosis of HD. Because acetylcholinesterase is increased in the muscularis mucosae and lamina propria in the aganglionic segment (see Fig. 96-22D), staining for this enzyme has been used for many years. This technique requires fresh, non–formalin-fixed tissue and technical expertise; at best, this stain is confirmatory. False-positive and false-negative reports have been documented in total colonic aganglionosis.60 A variety of histochemical staining methods have been proposed for the identification of ganglion cells, but all are expensive, time-consuming, and unnecessary. In the neonate, considerations in the differential diagnosis of HD include other causes of intestinal obstruction, such as meconium ileus, ileal atresia, meconium plug syndrome, and the microcolon seen in infants of diabetic mothers. When symptoms and signs of enterocolitis are present, diagnostic possibilities in the neonate also include primary necrotizing enterocolitis, HD-associated enterocolitis, milk protein–induced colitis (see Chapter 9), and sepsis with possible disseminated intravascular coagulation. In the toddler or older child, HD must be differentiated from functional constipation (stool withholding, fecal retention). In the latter condition, history indicates that the child did pass meconium in the newborn nursery and that clinical problems did not arise until the child usually was at least 18 months of age. Fecal impaction almost always is present in fecal retention, and fecal soiling is characteristic. Children with anterior displacement of the anus may be more prone to fecal retention. Idiopathic pseudo-obstruction and intestinal neuronal dysplasia generally can be distinguished from HD by rectal biopsy. Management Definitive treatment of HD is surgical. In all instances, biopsy of the muscularis propria of the bowel is indicated at the time of surgery to assess for the presence of ganglion cells in the myenteric plexus and to delineate the proximal extension of aganglionosis. All full-term babies with meconium plug in the newborn nursery should be evaluated for HD before discharge, because approximately 15% of children with HD have a history of meconium plug. Discharge of any newborn with undiagnosed HD and consequent delay in operative intervention may result in a greater frequency of enterocolitis, increased morbidity and even mortality. The specific method of surgery is operator dependent. Long-term prognosis varies and may depend on the length of the aganglionic segment. Even in the most common form of HD (short segment), it is usual to see older children continue to have defecatory issues with fecal retention and encopresis. The exact reasons for these continuing problems remain unclear, but the mechanism may involve an intrinsic

A

B Figure 96-23.  Photomicrographs of a rectal biopsy specimen from a patient with intestinal neuronal dysplasia. A, Increased number of enlarged ganglia (arrows). B, Active inflammation of the rectal mucosa with a crypt abscess (arrow). (A and B, Hematoxylin and eosin, ×250.)

abnormality in what is described as normal colon or in the pacemaker system of the colon. In the future, cell therapy using precursor cells from the developing human enteric system might prove to be a therapeutic option.61

Intestinal Neuronal Dysplasia

Intestinal neuronal dysplasia (IND) is a motility disorder that manifests with intestinal obstruction or severe chronic constipation; characteristic biopsy findings include an increased number of enlarged ganglia and neural hypertrophy (Fig. 96-23A).62 In addition, acetylcholinesterase activity is increased in the lamina propria and muscularis mucosae. A full-thickness surgical biopsy specimen is often necessary to diagnose IND. IND has been reported as an isolated lesion affecting especially premature infants, or infants with a history of formula protein intolerance, ileal stenosis, or small left colon–meconium plug syndrome. Three types of IND have been defined. IND type A usually manifests acutely in the neonatal period as severe constipation and enterocolitis. Biopsy features include mucosal inflammation (see Fig. 96-23B), ulceration with hyperplastic neural changes limited to the myenteric plexus, and increased acetylcholinesterase activity in the lamina propria and muscularis mucosae. The submucosal plexus in this type of intestinal neuronal dysplasia is histologically normal.

1639

1640

Section X  Small and Large Intestine IND type B usually is seen in children between six months and six years of age who have chronic constipation and megacolon. Histopathologic findings include hyperplastic submucosal ganglia with increased acetylcholinesterasepositive fibers in the muscularis mucosae and lamina propria. Ectopic ganglion cells in the muscularis mucosae and lamina propria also have been described. No changes are seen in the myenteric plexus. Significant interobserver variation has been documented for the pathologic diagnosis of IND type B by rectal suction biopsies. Some reports have speculated that some of the morphologic features described in type B are normal age-related phenomena. A third, mixed type of IND has an acute presentation and involves both the submucosal and the myenteric plexuses. The pathogenesis of IND is controversial. In some patients it is congenital malformation, whereas in others it is an acquired phenomenon. IND also can be seen in association with other syndromes such as neurofibromatosis or MENIIB, in proximal-segment HD, and with congenital anomalies predominantly of the gastrointestinal tract.63 Other associated conditions include cystic fibrosis, microvillus inclusion disease, congenital anomalies, lipoblastomatosis, and inflammatory bowel disease. Therefore, IND may not represent a well-defined entity but rather may constitute a secondary phenomenon related either to age or to obstruction or inflammation.64 IND can resolve with age.

Chronic Intestinal Pseudo-obstruction

Congenital forms of neuropathic and myopathic pseudoobstruction are rare and sporadic, perhaps representing new mutations (see Chapter 120). In these situations, a family history of pseudo-obstruction is lacking, as are any asso­ ciated syndromes and evidence of other predisposing factors such as toxins, infections, ischemia, or autoimmune disease. Children with chromosomal abnormalities such as Down syndrome, as well as those with MEN-III or with Duchenne’s muscular dystrophy, may suffer from pseudo-obstruction.

MISCELLANEOUS AND GENETIC DEFECTS Microvillus Inclusion Disease

Congenital microvillus atrophy, also known as microvillus inclusion disease, is an autosomal recessive disorder that may manifest with severe diarrhea shortly after birth and is characterized by atrophy of the intestinal villi and characteristic electron microscopic findings.65 Although the prevalence of microvillus inclusion disease is not known, it is reported to be the most common cause of familial intract­ able diarrhea.66 A female gender predominance has been observed, and consanguinity is reported in 20% of cases. The incidence of microvillus inclusion disease may be higher among Navajo Indians and persons from the Middle East. Defective protein trafficking and abnormal cytoskeletal and microfilament function have been proposed as possible etiologies.67 A blockage in the transport pathway from the Golgi apparatus leads to fusion of the small vesicles into microvillus inclusions.68 Secretory diarrhea is severe, with intolerance to oral feeding and unresponsiveness to most therapeutic modalities. Three variants of microvillus inclusion disease are recognized: congenital, the most frequent and severe, manifesting within the first week of life; late-onset, starting at six to eight weeks; and atypical, with either early or late onset. The wall of the small intestine is paper-thin in micro­ villus inclusion disease. The mucosa of the duodenum and small bowel is characterized by villus atrophy, hypoplastic

lp

A

i

B Figure 96-24.  Photomicrographs of the duodenum from a patient with microvillus inclusion disease. A, Villus atrophy with crypt hyperplasia (arrow) and decreased cellularity of the lamina propria (lp). (Hematoxylin and eosin, ×250.) B, On electron microscopy, lack of or shortened microvilli (arrow) and a cytoplasmic inclusion (i), composed of a vesicle lined by microvilli, can be seen. (×15,000.) (Courtesy of S. Teichberg, PhD.)

or normal crypts, and normal or decreased cellularity of the lamina propria (Fig. 96-24A). The absence of the brush border membrane is demonstrated by lack of linear staining with PAS, carcinoembryonic antigen (CEA), and CD10.69 These stains also visualize the microvillus inclusions on light microscopy. Evaluation by electron microscopy reveals ultrastructural abnormalities of the microvillus membrane, including disruption or absence of the brush border membrane, shortening and absence of the microvilli, and microvillus inclusions (see Fig. 96-24B). Although these lesions are most commonly noted in biopsies from the small intestine, micro­ villus inclusions also may be seen in specimens from the rectum and colon. Total parenteral nutrition must be used to prolong survival. The secretory diarrhea persists but becomes less voluminous. Small bowel transplantation should be considered because without it, microvillus inclusion disease is fatal.70 Intestinal Epithelial Dysplasia Intestinal epithelial dysplasia (IED), also known as tufting enteropathy, is a congenital enteropathy with early onset, severe intractable diarrhea, and characteristic microscopic findings.71 In IED, there is a variable degree of villus atrophy. Surface epithelial cells are arranged in tufts with a round apex. Tufts can also been seen in the colonic mucosa. These epithelial cells have an abnormal expression of E-cadherin and do not contain inclusions on electron microscopic examination. In the basement membrane, heparin sulfate proteoglycan is increased, and laminin is faint and irregular.71

Chapter 96  Anatomy, Histology, Embryology, and Developmental Anomalies of the Small and Large Intestine The diarrhea is secretory, malabsorption intractable, and growth is impaired. Several cases of IED have been associated with congenital anomalies.71 Nonspecific punctate keratitis is observed in more than 60% of patients with IED. Most patients with IED have consanguineous parents or affected siblings. In the Middle East, IED is even more common than microvillus inclusion disease. IED is characterized by a basement membrane with abnormal distribution of 2 β1 integrin adhesion molecules along the crypt-villus axis.71 Tufts result from nonapoptotic epithelial cells that are no longer in contact with the basement membrane. Small bowel transplantation is required.

Congenital Glucose and Galactose Malabsorption

Familial glucose and galactose malabsorption, transmitted as an autosomal recessive trait, due to mutation in the SGlLT1 gene, is characterized by an absence of the active transport carrier protein (Na+-glucose cotransporter) for glucose and galactose.72 Ingestion of any formula containing glucose or galactose results in severe life-threatening watery diarrhea in the newborn period. Stools are strongly positive for reducing substances. Neither blood nor white blood cells are present in the stool. Findings on biopsy of the small bowel and colon are normal. Discontinuation of formula containing glucose, galactose, or lactose (lactose is metabolized to glucose and galactose) and institution of a fructosecontaining formula with resultant therapeutic benefit usually are sufficient to make a clinical diagnosis of glucose or galactose malabsorption. Diarrhea abruptly ceases and the newborn begins to thrive when fructose-containing formula feedings are instituted. Some reports indicate that the severity of the diarrhea from glucose or galactose malabsorption diminishes with age because of the increased capacity of the intestinal flora to metabolize glucose.

Congenital Sucrase and Isomaltase Deficiency

Because sucrose is not a common dietary carbohydrate during the first six months of life, watery stools generally do not develop in children with this disorder until sucrose is administered in baby food. An exception to this rule is in the newborn receiving a formula (usually with soy protein or casein hydrolysate) with sucrose as the carbohydrate. Because sucrose itself is not a reducing substance, to make the diagnosis, the stool must be hydrolyzed by boiling it with 1N hydrochloric acid for 20 minutes, thereby changing sucrose to glucose and fructose. Congenital sucrase or isomaltase deficiency, although extremely rare, is the most common congenital disaccharidase deficiency.

Congenital Lactase Deficiency

Congenital absence of lactase is extremely rare. Affected babies receiving a lactose-containing formula develop severe watery diarrhea, which resolves with the institution of a non–lactose-containing formula. Biopsy specimens of the small intestine are normal histologically, but assay for disaccharidases reveals diminished or absent lactase.

Congenital Chloride Diarrhea

Congenital chloride diarrhea is an autosomal recessive disorder of intestinal Cl-HCO3 exchange caused by mutations of the SLC26A3 gene.73 The chloride-bicarbonate exchange mechanism in the ileum and colon is reversed, and chloride is actively secreted, resulting in a chloride-rich diarrhea. The baby with congenital chloride diarrhea often is premature and may present with an ileus or absence of passage of meconium. Watery diarrhea with a high stool chloride content and low stool pH is lifelong. Increased absorption

of bicarbonate may result in dehydration, a hypochloremic metabolic alkalemia, hyponatremia, and marked hypokalemia. The stool contains no blood, no white blood cells, and no reducing substances. Urinary chloride is low. Biopsy specimens of the small intestine and colon are normal. Treatment is fluid and electrolyte replacement. Acid reduction with proton pump inhibitors has been tried with variable results.

Congenital Sodium Diarrhea

Congenital sodium diarrhea is caused by defective sodium or proton exchange.74 Patients have acidemia and hyponatremia. The stool concentration of HCO3 and sodium are increased.

Cystic Fibrosis

Cystic fibrosis is an autosomal recessive disorder of cyclic adenosine monophosphate chloride transport that results from a defect in the cystic fibrosis transmembrane regulator (CFTR) (see Chapter 57). Approximately 10% to 15% of newborns with cystic fibrosis present with neonatal meconium ileus or its complications. Meconium plug syndrome also may occur, resulting in colonic obstruction, rather than small bowel obstruction, as is seen with meconium ileus. Antenatally, small intestinal ischemia and perforation may occur, resulting in meconium cyst, intestinal atresia, or meconium peritonitis with intra-abdominal or scrotal calcifications.

KEY REFERENCES

Abdullah F, Arnold MA, Nabaweesi R, et al. Gastroschisis in the United States 1988-2003: Analysis and risk categorization of 4344 patients. J Perinatol 2007; 27:50-5. (Ref 22.) Amiel J, Sproat-Emison E, Garcia-Barcelo M, et al. Hirschsprung disease, associated syndromes and genetics: A review. J Med Genet 2008; 45:1-14. (Ref 54.) Auclair BA, Benoit YD, Rivard N, et al. Bone morphogenetic protein signaling is essential for terminal differentiation of the intestinal secretory lineage. Gastroenterology 2007; 133:887-96. (Ref 2.) Cho S, Moore SP, Fangman T. One hundred three consecutive patients with anorectal malformations and their associated anomalies. Arch Pediatr Adolesc Med 2001; 155:587-91. (Ref 43.) Fairbank TJ, Sala FG, Kanard R, et al. The fibroblast growth factor pathway serves a regulatory role in proliferation and apoptosis in the pathogenesis of intestinal atresia. J Pediatr Surg 2006; 41:132-6. (Ref 36.) Goulet O, Salomon J, Ruemmele F, et al. Intestinal epithelial dysplasia (tufting enteropathy). Orphanet J Rare Dis 2007; 2:20-29. (Ref 71.) Karnak I, Ocal T, Senocak ME, et al. Alimentary tract duplication in children: Report of 26 years experience. Turk J Pediatr 2000; 42:11825. (Ref 33.) Madison BB, Braunstein K, Kuizon E, et al. Epithelial hedgehog signals pattern the intestinal crypt-villus axis. Development 2005; 132:27989. (Ref 3.) Mastroiacovo P, Lisi A, Castilla EE, et al. Gastroschisis and associated defects: An international study. Am J Genet A 2007, 143:660-71. (Ref 21.) Meier-Ruge WA, Ammann K, Bruder E, et al. Updated results on intestinal neuronal dysplasia (INDB). Eur J Pediatr Surg 2004; 14:384-91. (Ref 62.) Pena A. Imperforate anus. In: Wyllie R, Hyams JS, editors. Pediatric Gastrointestinal Disease. Pathophysiology, Diagnosis, Management. 2nd ed. Philadelphia: WB Saunders; 1999. p 499. (Ref 49.) Penco JM, Murillo JC, Hernandez A, et al. Anomalies of intestinal rotation and fixation: Consequences of late diagnosis beyond two years of age. Pediatr Surg Int 2007; 23:723-30. (Ref 29.) Schonhoff SE, Giel-Moloney M, Leiter AB. Minireview: Development and differentiation of gut endocrine cells. Endocrinology 2004; 145:2639-44. (Ref 8.) Scoville DH, Sato T, He XC, Li L. Current view: Intestinal stem cells and signaling. Gastroenterology 2008; 134:849-64. (Ref 1.) Full references for this chapter can be found on www.expertconsult.com.

1641

CHAPTE R

97 Small Intestinal Motor and Sensory Function and Dysfunction Jane M. Andrews and L. Ashley Blackshaw

CHAPTER OUTLINE Anatomy  1644 Key Elements in Normal Small Intestinal Motor and Sensory Function  1644 Smooth Muscle  1644 Interstitial Cells of Cajal  1644 Neurons  1645 Gastrointestinal Hormones  1648 Integrated Control of Motility  1648 Peristalsis  1648 Interdigestive Motor Cycle  1648 Mechanisms Underlying Abnormal Motor and Sensory Function  1649 Smooth Muscle Dysfunction  1649

The two most important goals of small intestinal motor and sensory function are the efficient absorption of nutrients and the maintenance of orderly aboral movement of chyme and indigestible residues along the small intestine. Small intestinal motility is also critically important in preventing bacterial overgrowth within the intestinal tract. This is achieved by the net aboral flow of luminal contents during both the fed and the fasting states, probably with the assis­ tance of the gatekeeper function of the ileocecal junction, which prevents backflow of cecal contents and keeps small intestinal bacterial concentrations at their usual relatively low levels. Net movement of contents along the small intestine is antegrade, but retrograde flow also occurs normally over short distances. Optimal progression of luminal contents allows the optimal mixing of digested food with intestinal secretions and contact of contents with the epithelium; such contact is important for absorption and sensing of nutrients within the lumen. Both absorption and mucosal sensing of nutrients exert significant feedback control on gastric and small intestinal motor function, an interplay thought to optimize the rate at which additional nutrients are pre­ sented to the absorptive epithelium, and to minimize the amount of nutrients lost to the colon. Preceding emesis, and in association with nausea, gross retrograde movement of small intestinal contents occurs over long distances, when a unique pattern of a strong zone of phasic small intestinal contractions travels in an orad direction over a large portion of the small intestine. These contractions deliver luminal contents back to the stomach for ejection into the esophagus

Intrinsic Neural Dysfunction  1649 Extrinsic Afferent Dysfunction  1650 Measurement of Small Intestinal Motility  1651 Basic Principles  1651 Clinical Approach  1652 Normal in Vivo Small Intestinal Motility Patterns  1654 Contractions at a Fixed Point  1654 Contractions that Travel along the Small Intestine  1654 Patterned Motility  1654 Clinical Consequences of Disordered Small Intestinal Motility  1656 Approach to Patients with Possible Small Intestinal Motor Dysfunction  1657

during emesis. This coordinated motor pattern underscores the versatile modulation of small intestinal motility accord­ ing to physiologic need. The motor function of the small intestine depends directly on smooth muscle in the intestinal wall, which contains the basic control mechanisms that initiate contractions and control their frequency. Overlying these basic control mech­ anisms are the enteric nervous system (ENS) and the auto­ nomic nervous system (ANS). In addition, a number of hormones modulate the frequency and patterning of small intestinal contractions. Each of these factors plays a role in the motility of the small intestine in health; specific damage to each component in some diseases has helped to define their discrete roles. This chapter concentrates on the physiology of normal small intestinal motility. Anatomy is considered first, with discussion of the structural and functional elements that control sensory and motor activity. Neurophysiology, in­ tegrative control, and patterning of small intestinal motility are reviewed next, along with some insights into possible mechanisms underlying motor and sensory dysfunction. Basic, and then clinical, measurement techniques and limi­ tations in the evaluation of motility then are discussed, followed by descriptions of commonly recognized motor patterns. Finally, we present a more clinically directed commentary on specialized tests used to assess small intestinal motility, disease states in which small intestinal motor and/or sensory function is disturbed, and a general approach to the patient with suspected small intestinal motor dysfunction.

1643

1644

Section X  Small and Large Intestine ANATOMY The small intestine is approximately 3 to 7 meters long and extends from the duodenal side of the pylorus to the ileocecal valve. It is divided into three regions—duodenum, jejunum, and ileum—based on structural and functional considerations. Although some structural and functional differences exist among these three regions, they exhibit similar motor characteristics. At each end of the small intes­ tine, however, physiologic sphincters—the pylorus and the ileocecal valve—have distinctly different motor patterns that give them the ability to act as controllers of flow between the antrum and duodenum and between the ileum and colon, respectively. The motor function of the pylorus is discussed in Chapter 48, the ileocecal region is discussed in Chapter 98, and general anatomy is discussed in Chapter 96. The duodenum is a fixed, largely retroperitoneal struc­ ture located in the upper abdomen, and the distal ileum generally is anchored in the right iliac fossa by its attach­ ment to the cecum. Except for these regions, the small intestine is mobile within the peritoneal cavity.

KEY ELEMENTS IN NORMAL SMALL INTESTINAL MOTOR AND SENSORY FUNCTION SMOOTH MUSCLE The wall of the small intestine comprises the mucosa, con­ sisting of the epithelium and lamina propria; submucosa; muscular layer (muscularis); and serosa (Fig. 97-1). The muscularis is composed of inner circular and outer longi­ tudinal layers of smooth muscle, which are present in con­ tinuity along the length of the small intestine. Contractions within these layers are responsible for gross small intestinal motility. A much smaller additional muscular layer, the muscularis mucosae, is present between the mucosa and the submucosa and plays a role in mucosal or villus motility1 but does not contribute to gross motility and is not con­ sidered further in this chapter. The smooth muscle cells within each muscle layer form a syncytium. Myocytes communicate electrically with each other through physically specialized areas of cell-to-cell contact, called gap junctions, which are visible by electron microscopy. This intimate contact between adjacent myo­ cytes gives low-resistance electrical contact or coupling among them, thereby enabling them to be excited as a unit. Mechanical connections among myocytes in each layer enable them to function as a contractile unit. At a cellular level, the mechanical connections are provided by inter­ mediate junctions, and at a tissue level, mechanical connec­

tions are provided by the dense extracellular stroma of collagen filaments between bundles of smooth muscle cells. Within each layer, the smooth muscle cell bodies are arranged in parallel, so that the circular muscle layer encircles the lumen, and the longitudinal layer extends axially along the small intestine; each may be controlled independently. Hence, small intestinal muscle contractions reduce luminal diameter and/or shorten small intestinal length. The myocytes themselves are spindle-shaped cells that derive their contractile properties from specialized cyto­ plasmic filaments and from the attachment of these fila­ ments to cytoskeletal elements. On electron microscopy, condensations of electron-dense, amorphous material are noted around the inner aspect of the cell membrane (dense bands) and throughout the cytoplasm (dense bodies). The contractile filaments—actin and myosin—are arranged in a fashion similar to that in skeletal muscle and insert onto the dense bands and bodies approximately in parallel with the long axis of the cell. Thus, when the contractile filaments are activated to slide over each other, cell shortening results. Most of the Ca2+ required for activating the contractile appa­ ratus enters the cells via L-type Ca2+ channels (Fig. 97-2). Ca2+ entry also can be supplemented to a varying extent by release of Ca2+ from the sarcoplasmic reticulum membrane via IP3 receptor–operated Ca2+ channels. IP3 is generated by phospholipase C, which in turn is activated by G-proteins, coupled to receptors for excitatory transmitters (G-protein– coupled receptors). The increased cytoplasmic Ca2+ binds to the Ca2+ binding protein calmodulin, enabling it to activate myosin light chain kinase, which phosphorylates the 20 kD light chain of myosin (MLC20). Phosphorylation of MLC20 facilitates actin binding to myosin and initiates cross-bridge cycling and development of mechanical force. Phosphorylation of MLC20 is reduced by MLC phosphatase. Dephosphory­ lation of MLC20 reduces cross-bridge cycling and leads to muscle relaxation. The dephosphorylation process is under a complex system of hierarchical control, which is impor­ tant in setting the gain of smooth muscle contractility.2

INTERSTITIAL CELLS OF CAJAL

Interstitial cells of Cajal (ICC) are specialized cells within the smooth muscle layer that are vital for normal small intestinal motor function. ICC are pleomorphic mesenchy­ mal cells that form an interconnecting network via long, tapering cytoplasmic processes. ICC lie in close proximity to both nerve axons and myocytes, with which they form electrical gap junctions.3 ICC serve two roles in control of small intestinal motility: first, they act as pacemakers gen­ erating the electrical slow wave that determines the basic rhythmicity of small intestinal contractions4; second, they

Epithelium Mucosa Lamina propria Muscularis mucosae Submucosa Figure 97-1.  Diagram showing the layers and components of the small intestinal wall. DMP, deep muscular plexus; ICCIM, intramuscular interstitial cells of Cajal; ICCMY, myenteric interstitial cells of Cajal. (Advice from Dr. Elizabeth Beckett is acknowledged.)

Circular muscle Muscularis propria

Submucous plexus ICCIM (DMP) ICCMY Myenteric plexus

Longitudinal muscle Serosa

ICCIM

Chapter 97  Small Intestinal Motor and Sensory Function and Dysfunction G-protein coupled receptors L-type

Ca2+ channel IP3

PLCβ

Ca2+ Calmodulin

Ca2+

Relaxation

MLC 20 MLCK MLCP MLC 20 (P)

Multiple intracellular pathways

Contraction

Figure 97-2.  Diagram of a smooth muscle cell showing pathways that lead to contraction and relaxation. See text for details. MLC 20; 20-kD myosin light chain; (P), phosphorylated; PLC, phospholipase C; MLCK, myosin light chain kinase; MLCP, myosin light chain phosphatase. (Modified from Sanders KM. Regulation of smooth muscle excitation and contraction. Neurogastroenterol Motil 2008;20 Suppl 1:39-53.)

transduce both inhibitory and excitatory neural signals to the myocytes5 and thus can vary the myocyte membrane potential and, in turn, contractile activity. This transduction occurs because ICC are interposed functionally between nerve terminals and the smooth muscle that the nerves supply. The neuroeffector junctions of the small intestine are not just simple contacts between nerve terminals and smooth muscle cells; they are contacts between enteric nerve terminals and ICC, and from there with myocytes by means of electrical gap junctions. Thus, effective neuro­ transmission results from the activation of specific sets of receptors on ICC, rather than by direct action on smooth muscle cells. At least three separate functional groups of ICC exist. They are the myenteric ICC (ICCMY), intramuscular ICC (ICCIM), and ICC in the deep muscular plexus (ICCDMP). Cells of the ICCMY population form a dense, electrically coupled network within the intermuscular space at the level of the myenteric plexus between the circular and longitudi­ nal muscle layers. ICCMY are the pacemaker cells in the small intestine that trigger generation of slow waves in the smooth muscle. These cells possess a specialized mecha­ nism that uses their oxidative metabolism to generate an inward (pacemaker) current, resulting from the flow of cations through nonselective cation channels in the plasma membrane. A primary pacemaker initiates slow waves. This depolarization from the primary event then entrains the spontaneous activity of other ICC within the network. This sequence results in a propagation-like phenomenon by which slow waves spread, without decrement, through the ICC network by means of gap junctions. A specialized type of ICCMY line the septa (ICCSEP) between circular muscle bundles; these cells form a crucial conduction pathway for spreading excitation deep into muscle bundles of the human jejunum, which is necessary for the motor patterns underly­ ing mixing.6 The second main population of ICC, ICCIM, is distributed within the muscle layers. ICCIM are innervated preferentially by intrinsic enteric motor neurons. In the small intestine, a third population, ICCDMP, which may be a specialized type of ICCIM in the small intestine, is concentrated at the

inner surface of the circular muscle layer at the region of the deep muscular plexus; it also receives preferential innervation. Both inhibitory and excitatory enteric nerve terminals selectively target intramuscular ICC. Their responses are transduced, in turn, to smooth muscle cells through gap junctions. Inputs from enteric excitatory motor neurons are mediated by muscarinic acetylcholine receptors (M2 and M3) and NK1 substance P receptors that result in increased inward currents, thereby causing depolarization. When depolarization reaches smooth muscle, it increases the opening of L-type Ca2+ channels during slow waves. These conditions result in greater Ca2+ entry and more forceful phasic contractions. Inputs from inhibitory enteric motor neurons are mediated by neurotransmitters including nitric oxide and vasoactive intestinal polypeptide, which activate both receptor and nonreceptor mechanisms in ICCIM. The result of these inputs is increased opening of K+ channels and, in turn, a stabilizing effect on membrane potential, reduced Ca2+ channel opening, and less forceful contrac­ tions of smooth muscle. Therefore, the mechanical response of small intestinal muscle to the ongoing slow wave activity depends strongly upon regulation of its excitability by the enteric nervous system via ICCIM. ICC, in general, play broadly similar roles in the small intestine and colon, and the reader is referred to Chapter 98 for a discussion of their roles in the large bowel (see also Fig. 98-2), as well as recent reviews by Sanders, Ward, and their colleagues.4,5 Absence or inactivity of ICC has been implicated in a number of clinical disorders that manifest as disturbed intestinal motility (see Chapter 20).

NEURONS

The small intestine is richly innervated with both extrinsic and intrinsic neurons. Intrinsic neurons have their cell bodies within the wall of the small intestine and constitute the ENS. These intrinsic neurons greatly outnumber the neurons of the extrinsic supply, which have their cell bodies outside the gut wall, but they have projections that end within the intestinal wall. Extrinsic neurons can be classi­ fied anatomically according to the location of their cell bodies and the route along which their projections travel. Extrinsic motor neurons belong to the ANS and connect the central nervous system (CNS) with the ENS and, from there, the small intestinal smooth muscle through the ICC. Some extrinsic motor neurons terminate directly in the muscle layers. Extrinsic sensory neurons from the small intestine do not belong to the ANS and are classified as spinal or vagal, depending on the route they follow to the CNS (Fig. 97-3). Neurons supplying the intestine are designated either afferent or efferent, depending on the direction in which they conduct information. By convention, information is conducted centrally by afferent neurons and peripherally by efferent neurons. Thus, the term afferent in regard to neural supply is used to describe pathways conducting informa­ tion that is detected in the intestine; in most texts “afferent” is interchangeable with the “sensory,” although most sensory information from the small intestine is not per­ ceived at a conscious level. The terms efferent and motor in regard to neural supply are used to describe pathways con­ ducting signals toward the effector small intestinal smooth muscle. Although the importance of motor innervation for motility is self-evident, the pivotal role of afferent function in determining motor responses has been less well appreci­ ated. The importance of the extrinsic afferent innervation is emphasized by the observation that at least 80% of vagal fibers are afferent.7

1645

1646

Section X  Small and Large Intestine CNS

Brain Nodose ganglion

Autonomic nervous system and sensory connections

Spinal cord Parasympathetic motor supply

Preganglionic sympathetic motor supply

Spinal sensory supply

Prevertebral ganglia Intestinofugal neuron Vagal sensory supply

Axon collaterals

Postganglionic sympathetic motor supply

Muscularis ICCs

Intestinal wall ENS

Figure 97-3.  Schematic representation of the relationships among components of the small intestinal motor control system. For further details, see text. CNS, central nervous system; ENS, enteric nervous system; ICC, interstitial cells of Cajal.

Intrinsic Neurons

ENS elements of the small intestine can be subdivided into three major functional groups: primary sensory (afferent) neurons, motor (efferent) neurons, and interneurons. Other categories of neurons, including secretomotor and vaso­ motor neurons and motor neurons to endocrine cells, are recognized, but they are not considered further in this chapter. Many distinct groups of enteric neurons are now well characterized both structurally and functionally and are reviewed in detail elsewhere.8,9 The cell bodies of ENS neurons are grouped together in the ganglia (clusters of cell bodies) of two main intramural plexuses. These plexuses lie in the submucosa (submucosal plexus) and between the two muscle layers (myenteric plexus). A deep plexus exists within the circular muscle but does not contain ganglia. The ganglia in the submucosal and myenteric plexuses are connected by interganglionic fasci­ cles. These fascicles are composed predominantly of the axons of motor neurons and interneurons, because sensory nerve processes do not often extend for any distance outside the ganglia. The myenteric plexus consists of ganglia spaced at regular intervals connected by a network of interganglionic fasci­ cles; this major network is known as the primary plexus. Within this main structure, smaller branches of nerve bundles arise from the primary plexus and form the second­ ary plexus, and still smaller branches form the tertiary plexus. The submucosal plexus has two layers, one close to the mucosa and another nearer to the circular muscle layer. These two layers are connected by interganglionic fascicles. The submucosal plexus does not have a hierarchy of subor­ dinate plexuses. Afferent Supply The primary afferent neurons of the ENS morphologically are Dogiel type II neurons (neurons with numerous pro­

cesses).10 Intrinsic primary afferent neurons that respond to mucosal chemical stimuli have their cell bodies in the myenteric plexus, and they project axons toward the mucosa. The myenteric plexus also contains the cell bodies of intrinsic afferent neurons that discharge in response to mechanical stimulation of the muscle layer induced by muscle activity or stretch. Intrinsic afferent neurons that respond to mechanical stimulation of the mucosa also are believed to exist, based on enteric reflexes seen in extrinsically denervated preparations. The cell bodies and processes of these neurons have not yet been identified definitively, although available evidence is consistent with the presence of their cell bodies in the submucosal ganglia.10 Intrinsic sensory neurons synapse in the intramural plexuses with intrinsic motor neurons and interneurons, which they excite mainly by release of acetylcholine and substance P. A more detailed account of the function and role of intrinsic afferent neurons can be found in a review by Furness and coworkers.10 Observations indicate that other classes of enteric neurons also respond to mechanosensory stimuli, suggesting that the ENS behaves as a sensorimotor network rather than as sepa­ rate components.11 Efferent Supply The axons of the intrinsic motor neurons that supply small intestinal smooth muscle exit the intramural ganglia and enter either the circular or the longitudinal muscle layer, where they pass in close proximity to both the myocytes and ICC. No specific neuromuscular junctions are present in small intestinal smooth muscle as in skeletal muscle, although the multiple varicosities along the motor axons probably represent specialized areas of neurotransmission. The motor axons discharge along their length, potentially activating large numbers of myocytes through ICC but pos­ sibly also directly activating them. The lack of exclusive, specific neuromuscular junctions, the electrical gap junc­ tions among myocytes, and the overlap of innervation of myocytes from more than one motor axon mean that func­ tionally discrete motor units in the intestinal smooth muscle do not appear to exist, in contrast with skeletal muscle. The ENS motor supply itself is both inhibitory and excitatory, and intrinsic motor neurons generally contain both a fast and a slow neurotransmitter. The pre­ dominant excitatory transmitters are acetylcholine (fast) and substance P (slow), and the predominant inhibitory trans­mitters are nitric oxide (fast), vasoactive intestinal polypeptide (VIP) (slow), adenosine triphosphate (ATP) (fast), and the nucleotide β-nicotinamide adenine dinucleo­ tide (β-NAD).12 Interneurons Interneurons connect ENS neurons of the same class or of different classes with one another. They permit local com­ munication within limited lengths of intestinal wall (mea­ sured in millimeters or centimeters) and are implicated in simple local responses by means of release of acetylcholine or nitric oxide, depending on their oral or aboral direction of projection. Some evidence also suggests the presence of connections within the intestinal wall along greater dis­ tances, but these neural pathways are not well defined. These connections may be provided anatomically by the ENS or by connections between the ENS and ANS. Inter­ neurons that play an additional sensory role have been identified.11 A special type of interneuron, the intestinofugal neuron, may be important for controlling local reflexes. Intestinofu­

Chapter 97  Small Intestinal Motor and Sensory Function and Dysfunction gal neurons have cell bodies within the myenteric plexus. These cell bodies receive input from several local enteric neurons and project to the prevertebral ganglia, where they synapse with sympathetic motor neurons (see Fig. 97-3).13

Extrinsic Neurons

Afferent Supply The small intestine is innervated by vagal and spinal extrin­ sic afferents. The pathway of small intestinal vagal afferent innervation is relatively straightforward. The vagal afferent neurons have endings in the intestinal wall and cell bodies within the nodose and jugular ganglia, which deliver input directly to the brainstem. Spinal afferent fibers travel along perivascular nerves to the prevertebral ganglia, where neurons do not end but might give off axon collaterals that synapse on postganglionic sympathetic motor neurons; these fibers then pass into the thoracic spinal cord along the splanchnic nerves. Spinal afferent neurons have their cell bodies throughout the thoracic dorsal root ganglia and enter the spinal cord through the dorsal roots; they synapse mainly on neurons of the superficial laminae of the spinal gray matter. These neurons, in turn, can send projections to numerous areas of the brain involved in sensation and pain control. Spinal afferent neurons also can give off axon collaterals closer to the intestinal wall, which synapse on components of the ENS, blood vessels, smooth muscle, or secretory elements (see Fig. 97-3). The different stimulus response profiles of vagal and splanchnic mechanoreceptors are generally interpreted as evidence that vagal afferents subserve physiologic regulation, and splanchnic afferents mediate pain.14-16 Functionally, three distinct and characteristic patterns of terminal distribution can be identified within the intestinal wall. Extraluminal afferent fibers have responsive endings on blood vessels in the outer, serosal layer and in the mes­ enteric connections. Muscular afferents form endings either in the muscle layers or in the myenteric plexus.17 Mucosal afferents form endings in the lamina propria, where they are positioned to detect substances absorbed across the mucosal epithelium or released from epithelial and subepithelial cells, including enterochromaffin and immunocompetent cells.17 These three different populations of afferent endings have different sensory modalities, responding to both mechanical and chemical stimulation.14,18 Serosal and mesenteric affer­ ents are found mainly in the splanchnic innervation and are activated by distortion of the intestine and its attachments; they do not normally signal distention or contraction of the bowel wall unless it is strong enough to cause distortion of the outer layers. Serosal and mesenteric receptors also com­ monly show evidence of chemosensitivity. This observation hints at potential responsiveness to circulating or locally released factors, especially in view of the localization of these receptors on or near blood vessels.19 Muscular afferents respond to distention and contraction with lower thresholds for activation, and they reach maximal responses within levels of distention that are encountered normally during digestion. Muscular afferents show main­ tained responses to distention of the small intestine and signal each contractile event, giving rise to the term in-series tension receptors. Nerve tracing studies have identified vagal afferent terminals in the longitudinal and circular muscle layers described as intramuscular arrays (IMAs), consisting of several long (up to a few millimeters) and rather straight axons running parallel to the respective muscle layer and connected by oblique or right-angled short connecting branches.17,20 IMAs were proposed to be the inseries tension receptor endings, possibly responding to both

passive stretch and active contraction of the muscle, although direct evidence for this proposal is currently lacking. Vagal afferent terminals surrounding the myenteric plexus throughout the gastrointestinal tract have been described as intraganglionic laminar endings (IGLEs). These endings are in intimate contact with the connective tissue capsule and enteric glial cells surrounding the myenteric ganglia, and they have been hypothesized to detect mechan­ ical shearing forces between the orthogonal muscle layers. Evidence for such a mechanosensory function of IGLEs has been elaborated by mapping the receptive field of vagal afferent endings in the esophagus, stomach, and large intes­ tine, showing morphologically that individual hot spots of mechanosensitivity correspond with single IGLEs.21 Func­ tional evidence exists for muscular afferents in both the vagal and the spinal innervation, but the appearance of spinal distention-sensitive afferents in the small intestine is yet to be determined. It is likely to be distinct from that of vagal afferents due to their higher thresholds for distention.22 Small intestinal mucosal afferents have been found in the vagal supply, but their existence in the spinal supply can be inferred only from the fact that they exist in the colon.14,19 Mucosal afferents do not respond to distention or contrac­ tion but are exquisitely sensitive to mechanical deformation of the mucosa, as might occur with particulate material within the lumen.16,23 In the rat duodenum and jejunum, vagal afferent fibers penetrate the circular muscle layer and submucosa to form networks of multiply branching axons within the lamina propria of both crypts and villi.17 Termi­ nal axons are in close contact with, but do not seem to penetrate, the basal lamina and thus are in an ideal position to detect substances including absorbed nutrients and mediators that are released from epithelial cells and other structures within the lamina propria. Efferent Supply The extrinsic efferent pathways to the small intestine are supplied by the parasympathetic and sympathetic divisions of the ANS. The small intestinal parasympathetic supply is cranial and cholinergic, whereas the sympathetic supply is spinal (thoracic) and adrenergic. These two motor pathways are not entirely separate, however, because postganglionic sympathetic fibers arising from cervical ganglia sometimes are found within the vagus nerve. The parasympathetic motor neurons of the small intestine have cell bodies within the dorsal motor nuclei of the vagi in the medulla oblongata. Their axons extend through the vagi to the intestinal intramural plexuses, where they synapse with motor neurons of the ENS. The sympathetic motor supply is more complex: Primary motor neurons within the intermediolateral horn of the thoracic spinal cord synapse with second-order neurons in the prevertebral ganglia, which then synapse with ENS motor neurons within the intestinal intramural plexuses, directly with smooth muscle, or possibly with ICC. Both excitatory and inhibitory extrinsic motor outputs to the small intestine are recognized. Excitatory outputs depo­ larize, and inhibitory outputs hyperpolarize the smooth muscle, thereby facilitating and impeding the development of contractions, respectively. In general, the sympathetic motor supply is inhibitory to the ENS, and this ENS inhibi­ tion leads to decreased smooth muscle activity, with the opposite effect seen in sphincter regions. Direct sympathetic inhibitory and excitatory outputs to smooth muscle also exist. The parasympathetic motor output to the ENS is more diffuse, each primary motor neuron supplying a large area.

1647

1648

Section X  Small and Large Intestine Excitatory parasympathetic motor output occurs to either inhibitory or excitatory ENS motor neurons, through which parasympathetic efferents can selectively inhibit or excite smooth muscle.

Central Connections of Neural Control Elements

Centrally, the sensory and motor supplies to the small intes­ tine are closely interrelated; the vagal sensory input and the parasympathetic motor output are closely located, as are the spinal sensory input and the sympathetic motor output. Both the vagal parasympathetic and the spinal sympathetic supplies have widespread connections to many other areas throughout the CNS that are implicated in feeding, arousal, mood, and other reflex behavior. The proximity of these CNS areas involved in small intestinal regulation, and their interconnections, makes it likely that the vagal parasympa­ thetic and the spinal sympathetic control mechanisms are interconnected and might function less independently than has been previously thought. The parasympathetic primary motor neurons are located bilaterally in the dorsal motor nuclei of the vagus in the medulla, which lie close to and receive substantial input from neurons of the nuclei tracti solitarii (NTS). The NTS is the site of terminals of vagal afferent fibers, which enter through the tracti solitarii and have cell bodies in the nodose ganglia. Each NTS also has extensive connections to other CNS regions, and several of these regions have input to the dorsal motor nuclei of the vagus, thereby influencing vagal motor output to the intestive. The central connections of the spinal and sympathetic supply to the gut are less well described. The spinal sensory neurons enter the spinal cord, where they synapse ipsilaterally on second-order sensory neurons and also provide direct feedback to sympathetic preganglionic motor neurons through axon collaterals. The second-order sensory neurons then ascend the spinal cord either contra­ laterally or ipsilaterally, after which they terminate in numerous areas,15 including the raphe nuclei and periaq­ ueductal gray matter in the brainstem and the thalamus. The thalamus has extensive ramifications throughout the CNS. The central influence on sympathetic motor output to the small intestine is complex and not well understood, but stress and arousal level play a role. These influences have their output through the brainstem and descending tracts to the sympathetic preganglionic motor neurons in the intermediolateral horn of the spinal cord, which send their axons to the prevertebral ganglia, whereupon they synapse with sympathetic postganglionic adrenergic nerves.13

GASTROINTESTINAL HORMONES

Gastrointestinal hormones are dealt with in detail in Chapter 1, but it is important to emphasize here their vital role in modulation of small intestinal motor and sensory function. Gastrointestinal hormones relevant to small intestinal function can act in either a humoral or paracrine fashion on both enteric neurons and myocytes, and generally they are released in response to the presence (or anticipation) of enteral nutrition. The best known of these hormones include CCK, somatostatin, VIP, glucagon-like peptide-1 (GLP-1), gastric inhibitory peptide (GIP), ghrelin, and motilin. Most of the hormones released in response to the presence of food in the lumen lead to slowing of small intestinal transit, signals of satiety, and increased mixing or segment­ ing contractions (see later). For a detailed description of these hormones and their effects, the reader is referred to Chapter 1.

INTEGRATED CONTROL OF MOTILITY So far we have considered the structure and function of individual components of the neuromuscular apparatus of the small intestine. When we consider how these com­ ponents operate together to produce known motility pat­ terns, several gaps are revealed in our knowledge, because the evidence for contribution of specific mechanisms is often circumstantial. Two important examples of motility patterns—peristalsis and the interdigestive motor cycle (IDMC)—are described next. These motor patterns illustrate the involvement of integrated hierarchical levels of control and our current level of understanding of the control systems.

PERISTALSIS

Peristalsis is the fundamental integrated motility pattern of the small intestine and can be coordinated entirely within the ENS and muscular layers. It may be initiated in response to a number of mechanical and chemical stimuli in the lumen and consists of progression of contractile activity usually, but not always, in an aboral direction. Therefore, both sensory and motor aspects to peristalsis are recognized. The populations of intrinsic primary afferent neurons described earlier probably are responsible for detection of luminal stimuli, either directly or following release of medi­ ators from mucosal enteroendocrine cells. Their activation results in transmitter release onto neighboring interneurons and motor neurons whose activity is coordinated subse­ quently as a network to provide synchronous activation of circular and longitudinal muscles on one side of the bolus (usually the oral side) and synchronous inhibition of muscle on the other side. This networked activity normally travels aborally, but the mechanism of propagation is not yet under­ stood. It might result from patterns of activity in interneu­ rons that can project over distances of several millimeters and thus mediate a general descending excitation. The mechanism by which peristalsis is reversed—for example, in conditions of luminal toxicity—is not known, but the fact that reverse peristalsis does occur in the small intestine illustrates that the pattern is not a totally polarized phenomenon. Debate is ongoing about the precise interactions of trans­ mitters and mediators in the normal function of peristalsis, but peristalsis is known to be affected by exogenous activa­ tion of several pre- and postsynaptic mechanisms, some of which also may be active endogenously. Of particular inter­ est are serotoninergic mechanisms, which have been shown to have involvement in initiation of peristalsis and modula­ tion of transmission between subclasses of enteric neurons.

INTERDIGESTIVE MOTOR CYCLE

The IDMC is discussed here because it serves to demon­ strate the extraordinary integrative capacity of the ENS; other aspects of the IDMC are described later in this chapter. The IDMC is a complex series of periods of variable con­ tractile activity with distinct phases showing different con­ tractile amplitudes, propagation, and regularity. The pattern as a whole sweeps slowly down the small intestine in the fasting state and recurs at regular intervals. Although a number of candidate hormones are proposed to be involved in its initiation and recurrence, the switch between quies­ cent and active phases and their orderly migration along the bowel are functions of the ENS; this ENS autonomy is dem­ onstrated by occurrence of the IDMC in extrinsically dener­

Chapter 97  Small Intestinal Motor and Sensory Function and Dysfunction vated or autotransplanted intestine. The ENS therefore is capable of controlling large segments of the small intestine independent of extrinsic input, probably by virtue of its extensive interneuronal connections and constant sensory feedback. Although the ENS has this regulatory capacity, normal function is modulated by ANS efferent output, which in turn may be influenced by locally or centrally processed information gathered from primary spinal or vagal afferents. In particular, synapses outside the CNS in the prevertebral ganglia are capable of subserving inhibitory intestinointestinal reflexes that are potentially important in the minute-to-minute regulatory control of motility.13 Small intestinal neuromuscular function is also influenced by a number of hormones acting in either endocrine or paracrine fashion. Little direct information is available on the precise con­ tribution of each extrinsic pathway to motor function of the small intestine in humans. Vagal reflexes generally are thought to make an important contribution in the integra­ tion of major homeostatic functions, such as motility, secre­ tion, blood flow, and the control of food and water intake.14-16 The role of sympathetic reflexes is thought to be concerned primarily with inhibition of motility and other functions in response to noxious stimuli, rather than in digestive small intestinal functions.

MECHANISMS UNDERLYING ABNORMAL MOTOR AND SENSORY FUNCTION Much of the evidence for the mechanisms involved in dys­ function of the small intestine is derived from animal models in which mucosal inflammation or infection has been induced, after which alterations in physiology, pharma­ cology, and anatomy of motor and sensory elements are assessed. These models provide some clues to the underly­ ing mechanisms involved in motor abnormalities seen clini­ cally; however, because many clinical manifestations are of unknown etiology, this approach is limited in the extent to which basic findings can be translated directly. Infection and inflammation of the intestine can result in long-term changes in all elements, including myocytes, ICC, and intrinsic and extrinsic neurons. Symptoms in func­ tional gastrointestinal diseases such as functional dyspepsia and irritable bowel syndrome (IBS) may be attributable partly to specific sensorimotor abnormalities occurring locally in the intestine, but they also are attributable to alterations in the extrinsic neural control system of the intestine and possibly to alterations in central perception, processing of afferent information, or both (see Chapters 13 and 118). Abnormalities in pain control systems in the brain and disordered processing of affective components of vis­ ceral sensations also have been described in these condi­ tions24 and can produce symptoms through the central connections described in the preceding sections. Some clinical scenarios in which discrete abnormalities have been identified or hypothesized in small intestinal motility are outlined in Table 97-1.

SMOOTH MUSCLE DYSFUNCTION

It is often difficult to separate pathologic changes in the function of smooth muscle from those in neural control mechanisms; however, a number of changes can be attrib­ uted directly to alterations in smooth muscle. Cytokines

play an important role in the abnormal smooth muscle function associated with gastrointestinal inflammation and infection. Different insults induce different patterns of cyto­ kines, which in turn determine the type of infiltrating immune or inflammatory cells, which in turn release spe­ cific mediators. Thus, the resultant effect on smooth muscle function depends on the origin of disease. For example, nematode infection induces mastocytosis and eosinophilia, which lead to activation of intracellular signaling pathways in smooth muscle by IL-4 and IL-13, ultimately resulting in hypercontractility of smooth muscle.25 By contrast, chemically induced inflammation is characterized by the presence of neutrophils and macrophages among other cells. Inflammation and infection can lead to changes at sites in the small intestine distant from the affected site, and the functional effects of inflammation in smooth muscle can persist following recovery from the acute insult as is seen with post-infection IBS. Smooth muscle hyperresponsiveness may be characterized by enhanced responses to cholinergic and noncholinergic excitation and are observable in human inflammatory bowel disease.26

INTRINSIC NEURAL DYSFUNCTION

Several abnormalities of small intestinal intrinsic control are attributable to developmental dysfunction and are dealt with separately in Chapter 96. Changes in the ENS also can occur after a bout of intestinal infection or inflammation. Many of these changes are centered on the intrinsic primary afferent neurons. These neurons become more excitable because of changes in the expression of ion channels that initiate generation of action potentials and those that deter­ mine recovery of membrane potential after an action poten­ tial. Thus, the long after-hyperpolarization that characterizes intrinsic primary afferent neurons from other classes is shortened, and they are able to fire in longer trains. This ability directly affects the responses of other interneurons and motor neurons that receive inputs from these afferent neurons and that therefore are involved in intrinsic (ENS) reflexes. Changes in excitability may be observed during an acute phase of infection or inflammation,26 or for several weeks afterward,27 at least in the large intestine. These longer-term changes are referred to as plasticity and might partly explain the occurrence of exaggerated motor responses to a given stimulus in the acute phase and after recovery of mucosal lesions. Changes can result from alterations in gene expression in enteric neurons that persist beyond the initial insult, from persisting increases in locally released media­ tors following alterations in mucosal cell types, or from both types of responses.28 In animal models of insulin-dependent diabetes mellitus, altered levels of neuropeptides may be seen, which might explain the disordered motility noted clinically in diabetes mellitus. The only reported neuroanatomic human study in a patient with type 1 diabetes mellitus showed that ICC were markedly decreased throughout the entire thickness of the jejunum. A decrease in neuronal nitric oxide synthase, VIP, pituitary adenyl cyclase–activating peptide (PACAP), and tyrosine hydroxylase–immunopositive nerve fibers was observed in the circular muscle layer, and substance P immunoreactivity was increased.29 Although patients with type 1 diabetes mellitus and sympathetic denervation have abnormally slow gastric emptying (see Chapter 48), their transit of a liquid meal through the distal small intestine is more rapid, which might play a part in the production of diarrhea. Diabetic patients also show abnormal duodenal motility patterns such as early recurrence of phase III after

1649

1650

Section X  Small and Large Intestine Table 97-1  Disorders Associated with Abnormal Small Intestinal Motility DISORDER

SMOOTH MUSCLE ABNORMALITIES

Irritable bowel syndrome

None identified

Post-infection and inflammatory changes (subset of patients)

Acute severe illness

Decreased strength of contractions

Altered neurotransmission (? related to metabolic or electrolyte disturbances)

Pregnancy

Decreased strength of contractions None identified

None identified

Metabolic disturbances*

Possible decreased strength of contractions

Altered neurotransmission

Drugs†

Possible decreased strength of contractions

Intestinal obstruction

NEURAL ABNORMALITIES

SENSORY ABNORMALITIES

POTENTIAL OUTCOMES

Altered afferent function Increased visceral sensitivity Heightened sensitivity to neurohumoral feedback loops None identified

Heightened sensation Disordered motility

Altered neurotransmission

Enhanced perception of gastrointestinal stimuli Nausea Altered sensory perception None identified

Hypertrophy, if chronic

None identified

None identified

Pseudo-obstruction syndromes

Myopathy of hollow viscera

None identified

Scleroderma and other connective tissue disorders

Ischemia and fibrosis

Neurologic syndromes‡

NA

Multiple neural abnormalities: neuron loss, plexus abnormalities, altered distribution of neurotransmitters Nerve loss in intestinal wall Extrinsic neural supply may also be damaged by vasculitis Neural absence or loss

Abnormal patterning of contractions Slow or rapid transit Abnormal patterning of contractions Slow or rapid transit Ileus Slow or rapid transit Disordered contractions High-amplitude forceful contractions Feeble contractions Absent phase III of the IDMC Slow or failed transit

Rare myopathies

Myocyte and mitochondrial abnormalities; inadequate contractile force Fibrosis (? related to ischemia)

NA

None identified

Diabetes mellitus

Radiation enteritis

Acromegaly

Altered neurotransmission

Delayed transit Ileus Decreased absorption Slowed transit

None identified

Feeble contractions Thickening of bowel wall Slow transit

Loss of afferent neurons with consequent loss of sensory information for reflux control

Disorganized IDMC Failure to convert to fed pattern Transit failure

Increased muscarimic (m) 3 receptors

None identified

Postulated autonomic nerve dysfunction

None identified

Less mixing, disordered transit Stasis, bacterial overgrowth, diarrhea Delayed orocecal transit time Bacterial overgrowth

Insufficient force for transit and mixing

IDMC, interdigestive motor cycle; NA, not appropriate. *Examples include disturbances of potassium, calcium and magnesium homeostasis, and renal and hepatic failure. † Examples include antidepressants, calcium channel blockers, and beta-blockers. ‡ Examples include dysautonomia and Parkinson’s disease.

a meal (see later). No consistent correlation, however, has been found between changes in manometric parameters and the degree of cardiac autonomic neuropathy, nor has any correlation yet been established between changes in enteric neurotransmitters and ICC and manometric and transit observations.

EXTRINSIC AFFERENT DYSFUNCTION

Mechanisms leading to extrinsic afferent dysfunction after infection or inflammation probably are similar to those involved in intrinsic primary afferent and smooth muscle dysfunction. It is well established that a wide range of chemical mediators can influence mechanosensitivity of extrinsic primary afferents, in addition to evoking direct responses as detailed earlier. These chemical mediators can be released in conditions of inflammation, injury, or

ischemia from a variety of cell types including platelets, neutrophils, lymphocytes, macrophages, mast cells, glial cells, fibroblasts, blood vessels, muscles, and neurons. Each of these specific cells can release several modulating agents, some of which act directly on the sensory nerve terminal; others act indirectly, causing release of other agents from other cells in a series of cascades. The end result of these actions is that the response properties of extrinsic afferents, like their intrinsic counterparts, are subject to plasticity, usually resulting in an increased sensitivity of the afferent endings; this process is described as peripheral sensitization. Some evidence supports the involvement of algesic medi­ ators, including prostaglandins and purines, in changes leading to peripheral sensitization.30 Other endogenous chemical mediators, including somatostatin, can down-

Chapter 97  Small Intestinal Motor and Sensory Function and Dysfunction regulate small intestinal afferent sensitivity such that an imbalance in prosensitizing and antisensitizing mecha­ nisms leads to a disordered sensory signal. Such mecha­ nisms are likely clinically relevant to functional bowel disorders, such as IBS, in which increased perception of mechanical and chemical stimulation is apparent. More­ over, because these afferents also serve to trigger reflex mechanisms that control and coordinate intestinal motor function, their sensitization can contribute to chronic dys­ motility, resulting in a cycle of disordered sensory and motor function.

1.5 cm intervals

Antrum

MEASUREMENT OF SMALL INTESTINAL MOTILITY BASIC PRINCIPLES Spatiotemporal Measurements

The outcomes of small intestinal motor activity depend on the patterning of small intestinal contractions in both space and time: Where and when do the contractions occur with respect to each other? Measurement methods must therefore gather functionally relevant information on the temporospa­ tial organization of small intestinal motility. This presents substantial challenges, especially in humans, because of the length of the small intestine, the spatiotemporal complexity of motor events, and the long time frame (several hours) over which small intestinal motility determines the success­ ful absorption and movement of each meal. In health, the occurrence and patterning of a large number of individual motor events determine the outcomes of absorption and transit, so that whole-animal measures of small intestinal transit and absorption yield a gross, or summary, report. More-detailed descriptions of small intes­ tinal motility report great variability in the patterning of individual contractile events, depending in part on the tech­ nique used, the time frame over which motility is observed, and the temporospatial resolution of the measurement tech­ nique itself. To understand the relationship between individual motor events and transport in the small intestine, the temporal resolution of the measurement technique must be greater than the duration of each discrete motor event. Based on similar principles, the spatial resolution of measurements is also an important parameter to consider if relationships between motor events and intraluminal flow(s) are to be defined. The importance of spatiotemporal resolution can be appreciated by considering Figure 97-4. Direct evalua­ tion of small intestinal motility requires methods of mea­ surement with a time resolution of at least two seconds, because in humans, the intrinsic frequency of duodenal contractions is up to 12 per minute. The optimal spatial resolution for studies of small intestinal motor function has not been determined, but the spatial patterning of pressures is known to vary over relatively small distances,31 with most propagating pressure wave sequences traveling less than 6 cm. Because of practical limitations of data handling and the number of sensors one can place in the small intestine, measurement techniques usually either achieve high tem­ porospatial resolution over a short distance or low tempo­ rospatial resolution over a far greater distance. Realistically this means that data gained from different studies are usually interpreted alongside one another to provide more complete information.

Evaluation of Single Cell Functions

At the cellular level, a number of techniques can be used to yield insights into small intestinal motor physiology.

Distal duodenum

Antrum

4.5 cm intervals

Distal duodenum Antrum

Distal duodenum

6 cm intervals

50 mm Hg 0 mm Hg

30 seconds

Figure 97-4.  Multichannel manometric recordings of the human antrum and duodenum, with recording points placed at varied intervals: 1.5 cm (top panel), 4.5 cm (middle panel), and 6 cm (bottom panel). These data demonstrate some of the limitations of varying the interval between recording points: As a phasic contraction travels along a section of intestine, the associated rise in pressure is detected only at each measurement point. If the interval between recording points is too wide, unrelated pressures may be judged to be related to the propagated pressure wave, or a propagated pressure wave sequence may be judged to be a limited phasic event. Spatial detail is lost as the recording interval is widened. (Courtesy Dr. J.M. Andrews.)

Intracellular recordings of electrical potential can be obtained from a number of cell types within the small intes­ tine and its extrinsic neural control system. These record­ ings give detailed information about the signals received and transmitted by individual cells, with excellent temporal resolution, but generally they cannot be applied concur­ rently over a significant length of intestine and therefore have limited real-time spatial resolution with regard to motor events. A combined functional and neuroanatomic approach whereby imaging of specific neurons with intracellular or extracellular recordings and chemical coding using immu­ nohistochemistry are performed concurrently has allowed important correlations to be made between structure

1651

1652

Section X  Small and Large Intestine and function. In particular, this approach has led to under­ standing of the function of IGLEs and IMAs (see “Extrinsic Afferent Supply” earlier). Although electrophysiologic and anatomic methods provide information on structure, neurotransmitters used, and proximity to other elements, they cannot describe pre­ cisely how these relate to the actual resulting motility and its temporospatial organization. Although these single-cell techniques generally have been applied to animal tissues, the results also probably apply to humans, because a similar structural organization of control elements is seen in human tissue.

Recording of Muscle Contractions

Increased muscle tension generally is directly recorded with strain gauges; these can be used in muscle strips, isolated loops of intestine, and whole-organ preparations or even chronically implanted in animals. Strain gauges are capable of excellent temporal resolution of motor events, but spatial resolution is limited by the size and number of strain gauges that are used concurrently. Over short lengths of intestine, a spatial resolution of approximately 1 cm is possible. Unfortunately, strain gauges are not suitable for use in human subjects, although they have provided much valu­ able information on the organization of motor events in animals.32 Muscle contractions also can be measured by surrogate measurement techniques that record associated phenom­ ena. One such approach is fluorescence measurement of calcium transients (rapid increases in free intracellular calcium) in smooth muscle.33 Over short sections of intes­ tine (1-2 mm), such measurements provide excellent tem­ porospatial resolution and are helpful in elucidating neurophysiologic control rather than describing wholeorgan function. Other measurement techniques that record phenomena resulting from contractions of smooth muscle include luminal manometry (reflecting intraluminal pres­ sure increases), fluoroscopy (showing wall movement and movement of intraluminal contrast), and transit studies per­ formed by a number of approaches. Luminal manometry measures the change in intraluminal pressure that results mainly from lumen-occlusive or near– lumen-occlusive contractions. Fortunately, because the small intestine is tubular, with a relatively small diameter, a large portion of motor events are recognized as pressure rises. Researchers have hypothesized that contractions not resulting in a detectable change in intraluminal pressure are less important in determining flow, and therefore little mechanical information is lost by failure to detect them, but small changes in intraluminal pressure can be pivotal in producing flows in some regions of the small intestine.34 Manometry can be applied in several settings, ranging from short isolated intestinal segments in the laboratory to clinical use in humans. Modern computer-based recording systems allow excellent temporal resolution (~10 Hz), and spatial resolution can be tailored to give either close spatial resolution (intervals of 1-2 cm) over 20 to 40 cm, or wider resolution, while still covering a longer segment of small intestine. Manometric assemblies are either of the perfused side-hole or the solid-state sensor design and are capable of routinely recording at up to 22 sites.

Wall Motion and Transit Studies

Contrast fluoroscopy is the most widely available wall motion study. It yields detailed information on the time and space patterning of motor events in vivo and useful insights into associated movements of luminal contents. When this technique is used in combination with other techniques,

such as manometry, intraluminal impedance, or strain gauges, useful correlations can be made between contrac­ tions or luminal pressures and transit of contents. These insights are likely to lead to improved understanding of pressure patterns, which in turn might enable us to better interpret less-intrusive techniques such as manometry and impedance in humans. Improving the interpretation of these other techniques is important, because risks associated with radiation exposure restrict the use of fluoroscopy in humans. Other in vivo imaging methods for assessing small intes­ tinal wall motion and movement of intraluminal contents include magnetic resonance imaging (MRI), ultrasono­ graphy, and intraluminal impedance recording. These approaches are suitable for human use with good temporal resolution, although they have significant practical limitations. These limitations previously restricted their applications outside of research centers as an alternative to contrast fluoroscopy, but they are now gaining more widespread use. MRI allows prolonged observation but, because of the anatomic complexity of the small intestine, difficulties with spatial resolution can limit views. Additionally, MRI is expensive, and not all centers have sufficient MRI capacity for it to become a routine clinical tool for this indication. Ultrasonography also allows prolonged observation and repeat measurements, but only of short segments and with relatively poor spatial resolution. Ultrasonography is limited in many instances by patient factors, such as body habitus and intestinal gas, and it is operator dependent. Multichannel intraluminal impedance (MII) is a tech­ nique for assessing intraluminal bolus transit rather than motility. The technique is based on the different conductivi­ ties of intraluminal air and liquids compared with those of opposed sections of bowel wall. Voltage is applied to a recording assembly along which several electrodes are sited. The current recorded between electrode pairs depends on the conductivity and thickness of any air or fluid bolus straddling the electrode pair. In this fashion, MII sequen­ tially measures the transit of a conducting bolus between electrode pairs. Recordings in the small intestine can, there­ fore, depend on the state of its filling,35 and motility in an empty bowel might not be assessed accurately. Other transit and absorption measurements demonstrate whether mass transit occurs but give no information on the mechanical pattern by which the transport of contents is achieved. Methodology for transit studies includes breath tests and scintigraphy. Breath tests are based on the exhalation of gases such as H2 or CO2 (labeled with 13C or 14C), which are generated when a test meal reaches the colon and undergoes bacterial degradation. Scintigraphic tests of small intestinal transit visually assess the arrival of a labeled meal at the cecum. These two transit techniques yield the lowest temporospa­ tial resolution in assessing small intestinal motility but are clinically useful and are discussed later in this chapter. In vitro techniques for detailed assessment of small intes­ tinal wall movements reveal subtle motility patterns that cannot be detected with manometry or in vivo wall motion studies. For example, one technique using digitized video recording can measure changes in diameter and length of an immobilized segment of intestine36 and has the unique capacity to appreciate discrete changes in the longitudinal and circular muscle layers.

CLINICAL APPROACH

The broader issues of measurement of small intestinal motor function were considered earlier, and the discussion that

Chapter 97  Small Intestinal Motor and Sensory Function and Dysfunction follows is limited to the clinically relevant techniques used to assess small intestinal motor function.

Small Intestinal Transit Studies

Small intestinal transit time can be measured with breath tests or scintigraphic observation of the movement of intra­ luminal contents. Unless the test substance is delivered past the pylorus by tube, these techniques also include gastric emptying (and thus gastric function) in their measurements; they are, therefore, imprecise about actual small intestinal transit time and are more accurately termed tests of orocecal transit time. Because each technique measures a different aspect of motility, the results obtained from different tech­ niques are not directly comparable. The lactulose breath test is perhaps the best known and most widely used of these techniques. Lactulose is a non­ absorbable disaccharide that is fermented on reaching the bacteria-laden environment of the colon. The H2 gas that is formed is rapidly absorbed and exhaled from the lungs. Samples of exhaled gases are taken at baseline and at regular intervals after the ingestion of lactulose. The orocecal transit time is taken as the time at which a sustained rise in exhaled H2 is seen. An early rise, or a high baseline level, may be evidence of small intestinal bacterial overgrowth, but this measure is relatively insensitive for bacterial overgrowth. The administration of lactulose itself is known to hasten small intestinal transit and so the result is not directly com­ parable to other transit time measures. Similar principles are used in 13C or 14C breath tests, which measure gastric emptying combined with the evalu­ ation of small intestinal absorption of specific nutrients. Acetate, octanoic acid, and triolein have been used in this regard. Acetate appears to be a good liquid marker, octanoic acid is better suited for solids, and triolein is useful in sus­ pected cases of malabsorption. This nutrient-focused assess­ ment of small intestinal function can be combined with the H2 lactulose breath test to measure orocecal transit time as well. The more familiar visual and anatomic scintigraphic measurement of small intestinal transit is also widely avail­ able. The major difficulty with these studies is the lack of a reliable anatomic landmark for the cecum. Either the cecum is defined arbitrarily as the right iliac fossa and a skin marker is used or it is considered retrospectively as the area in which radioisotope accumulates. Two approaches are used to report the scintigraphic orocecal transit time. In the simpler approach, the time of first appearance of isotope in the cecum is given; in the other, the initial activity of the radiolabeled meal is quantified in the stomach, and the orocecal transit time is reported as the time taken for 50% of this initial gastric activity to reach the cecum. Values obtained vary depending on which of these methods is used, and each laboratory should set its own normal range.

Fluoroscopy

Contrast fluoroscopy is useful for detecting mural disease and fixed narrowings of the intestinal lumen that can induce secondary changes in motility, transit, and absorption. Fluoroscopy is insensitive for detecting abnormal nutrient absorption and measuring transit time. Clinical fluoroscopy is limited by the necessarily short observation times because of concern with radiation exposure; therefore, only gross disturbances of motor activity may be detected. Once a substantial amount of contrast has entered the small intes­ tine, the usefulness of fluoroscopy is reduced further, because overlying loops of bowel hinder the interpretation of the movement of contrast.

Manometry

Manometry of the small intestine gives direct measurement of the forces that are applied to luminal contents as a result of motor function. Manometry can be performed over hours or even days and over long or short segments; it is capable of excellent spatial resolution, although it has major practi­ cal limitations. Manometric assemblies can be placed in any part of the human small intestine and are moderately well tolerated, although placement of such an assembly along the small intestine can be demanding even in healthy persons, and it is especially challenging in patients who have major abnormalities of motor function. Manometry allows recognition of some abnormal patterns of pressure over time at individual recording points, but no studies have yet performed a critical evaluation of the best spacing of pressure recording points and of diagnostic cri­ teria for abnormal pressure patterns to distinguish between health and disease. This lack of specific criteria reflects the limited understanding of the relationship between small intestinal intraluminal time-space pressure patterning and the achievement of mixing and propulsion within the small intestine. Because of practical limitations, one must choose between high spatial resolution over a short segment and lower spatial resolution over a longer segment of intestine. Both approaches are likely to be necessary in achieving an accu­ rate understanding of small intestinal motor physiology, perhaps in conjunction with a technique to assess wall motion or intraluminal flow.

Multiple Intraluminal Impedance

Recording assemblies can be used to measure impedance in humans in much the same fashion as manometry. Multiple intraluminal impedance (MII) can measure episodes of bolus transit in any tubular section of the upper intestine. MII gives good spatiotemporal resolution, but owing to tech­ nical limitations of how far apart the sensors can be spaced, it cannot give continuous cover in measuring transit over long lengths of the intestine. MII is increasingly used for the clinical evaluation of esophageal motility, and in recent times it has been applied to the proximal small intestine with success.35 In combination with manometry, MII has the potential to show real-time pressure-flow relationships.

Magnetic Resonance Imaging

Magnetic resonance imaging (MRI) is capable of excellent spatiotemporal resolution of small intestinal wall motion and movement of intraluminal contents. Because the small intestine is not all within one plane, however, it cannot, at present, be viewed routinely all at once; it does not involve a radiation dose and thus is not time limited on this basis. Because it is also an anatomic imaging technique, it has a substantial advantage over other techniques of being able to offer additional information in the assessment of patients with suspected small intestinal motility problems. Wall thickening, fibrosis, inflammatory changes, and stenoses all can be revealed, and this information can help with direct­ ing diagnosis and even therapy.37 MRI has several disadvantages: it is expensive; some claustrophobic subjects find it too confining; patients who have pacemakers or other metallic prostheses cannot be assessed with MRI. At present, its use in assessing small intestinal motility is restricted to units with a research interest in functional gastrointestinal MRI.38 Its use is increasing in clinical gastroenterology, however, especially for small intestinal Crohn’s disease,37 and it also has the potential to encompass clinical motility assessments.39

1653

1654

Section X  Small and Large Intestine NORMAL IN VIVO SMALL INTESTINAL MOTILITY PATTERNS CONTRACTIONS AT A FIXED POINT

The increased smooth muscle tension arising from muscular contractions can result in increased intraluminal pressure, decreased intraluminal diameter, small intestinal shorten­ ing, or a combination of these effects. Smooth muscle con­ tractions can be tonic or phasic, but common usage has labeled tonic contractions as tone and phasic motor events as contractions. Human phasic small intestinal contractions generally last from 0.8 to 6.0 seconds. Small intestinal electrical recordings reveal continuous cyclical oscillations in electrical potential, referred to as the slow wave, basic electrical rhythm, or pacesetter potential. This slow wave is generated by the ICC (see earlier). In humans, the slow-wave frequency decreases from a peak of 12 per minute in the duodenum to approximately 7 per minute in the distal ileum. A small intestinal contraction arises when an electrical action potential, or spike burst, is superimposed on the slow wave (Fig. 97-5). Spike bursts may be caused by the intrinsic motor output from the ENS to the ICC and are likely also to be modulated by the extrinsic motor supply. Except during phase III of the IDMC (interdigestive migrating motor complex), not every slow wave leads to a phasic contraction. The region-specific frequency of the slow wave thus controls small intestinal rhythmicity by determining the timing and maximal fre­ quency of contractions. The rapid increases in free intracellular calcium, or calcium transients, that underlie smooth muscle contraction can be visualized with fluorescence techniques and appear to spread in a coordinated fashion over an area of smooth muscle and to extend over variable distances of the bowel wall. These calcium transients are extinguished by collision with each other or by encountering locally refractory regions.33

Intracellular recording

Extracellular recording

Muscle tension

Time Figure 97-5.  Schematic representation of the relationship among slow waves, spike bursts, and muscle contraction. The top tracing is from an intracellular electrode in the muscle; the middle tracing is from an extracellular electrode; and the bottom tracing shows muscle tension. The cyclical fluctuation in membrane potential in the top tracing is the slow wave. When spike bursts are superimposed on the peak of the slow wave, the muscle depolarizes, and contraction occurs. (From Christensen J. Gastrointestinal motility. In West JB, editor. Best and Taylor’s Physiologic Basis for Medical Practice. Baltimore: Williams & Wilkins; 1990. p 614.)

CONTRACTIONS THAT TRAVEL ALONG THE SMALL INTESTINE

The electrical slow wave migrates along the small intestine in an aboral direction so that each subsequent site along the intestine is depolarized sequentially. When a slow wave results in contraction, the propagation of the slow wave along the small intestine also leads to the contraction pro­pagating along the small intestine. The propagation velocity of the slow wave thus determines the maximal rate at which contractions can travel along the small intes­ tine. Because not every slow wave leads to a contraction, however, contractions do not always travel at this maximal rate. The distance over which muscular excitation or inhibition spreads appears to be determined by ENS influ­ ences mediated through local inhibitory and excitatory circuits.33 Contraction sequences travel aborad (in an antegrade direction) or orad (in a retrograde direction). From animal data and some human studies at high spatial resolution, it is known that a large portion of contractions travel along the small intestine, rather than remaining static, although most contractions are limited to only a few centimeters in extent.40,41 Further data are needed to determine the con­ tribution that these short contraction sequences make to overall transit compared with the less-frequent longer sequences.

PATTERNED MOTILITY

From isolated small intestinal segments, ascending excita­ tion and descending inhibition are the simplest wellrecognized patterns of motility. Ascending excitation refers to the contraction that occurs proximal (oral) to a stimulus, and descending inhibition refers to the inhibition of motor activity that occurs distal to a stimulus. These simple reflexes can be demonstrated in the absence of any extrinsic innervation and are thus entirely attributable to the ENS, although extrinsic influences can modulate their occur­ rence. These two patterns are thought to be responsible for peristalsis and retroperistalsis when they travel in a coordi­ nated fashion along the intestine. Recordings of human small intestinal motility show iso­ lated (stationary) phasic contractions, but often, spatial pat­ terns are more complex. The limited spatial resolution of many recording techniques can lead to over-reporting of the fraction of stationary contractions. Commonly, phasic motor activity consists of a recognizable group of contractions associated along the small intestine in space and time; phase III activity of the IDMC (see later) is a good example of this association. Several other types of grouped small intestinal contractions have been described and include contractions associated with emesis42 and discrete clustered contractions, which are said to be common in IBS (see Chapter 118).43 The most commonly observed motor pat­ terns in the healthy small intestine, however, are described simply as the fed or postprandial pattern and the fasting (interdigestive) pattern, or IDMC (Fig. 97-6). The motor pattern is determined by the presence or absence of a significant amount of nutrient within the small intestine. Despite a large number of studies on fasting motil­ ity, few studies have been performed on human postpran­ dial small intestinal motility; this paucity probably exists because of the difficulty in knowing which aspects of post­ prandial motility to study, in contrast to fasting motility, which has an easily recognized cyclic pattern and thus easily studied parameters. The fed motor pattern ensures transit of small intestinal contents at a rate consistent with normal digestion and absorption. The fasting motor pattern is less involved with orderly luminal transport and

Chapter 97  Small Intestinal Motor and Sensory Function and Dysfunction FASTING STATE Phase II (IDMC)

Phase III Phase I 50 mm Hg

5 min FED STATE

Liquid meal

variable amounts of the same nutrient, generate recogniz­ ably different motor responses.31,32,46,47 In general, the pres­ ence of unabsorbed small intestinal nutrients slows small intestine transit by decreasing the frequency and length of travel of phasic contractions, so that the rate at which a substance is absorbed limits its transit rate. In the absence of sufficient proximal small intestinal nutrient stimulation, the fasting motor pattern re-emerges four to six hours after a meal. In the absence of its interruption by intraluminal nutrients, the IDMC repeats continuously. Distention, intraluminal pH changes, and hyperosmolar contents are capable of stimulating small intestinal motor activity. Hyperosmolar contents and pH changes probably are sensed by receptors in the mucosa, whereas distention is signaled by receptors in the muscle. In the normal course of events, these stimuli occur concurrently with the pres­ ence of nutrients, and the significance of their isolated effects in healthy subjects is unclear. The small intestine also exerts negative feedback control on the rate of gastric emptying through neural and humoral means. This negative feedback is achieved by the release of neural signals and intestinal hormones that suppress phasic gastric motor activity, relax the gastric fundus, and increase tonic and phasic pyloric pressures subsequent to mucosal sensing of small intestinal nutrients.48 This process indi­ rectly also prolongs whole-meal small intestine transit time by slowing the input of small intestinal chyme. The small intestine, in particular the duodenum, also is thought to offer direct mechanical resistance to gastric emptying by acting as a capacitance resistor49 and by reaugmenting gastric contents as a result of duodenogastric reflux.50

Fed Motor Pattern

80 mm Hg 10 min Figure 97-6.  Manometric tracings demonstrating small intestinal motility in the fasting (top) and fed (bottom) state. The three phases of the interdigestive motor cycle (IDMC) (courtesy Dr. J.M. Andrews) and the conversion to a fed motor pattern by infusing a liquid nutrient meal into the small intestine are shown. In the top set of tracings, at a given time point (dashed vertical line), all three phases of the IDMC can be encountered at different points along the small intestine. The similarity of phase II and the fed motor pattern can be appreciated by comparing the top (fasting) and bottom (fed) sets of tracings. (Professor R.J. Fraser provided data for this figure.)

is thought to serve important roles in clearing the upper intestine of solid residues, which otherwise can accumulate and form bezoars; in maintaining relative sterility of the small intestine by keeping it empty; and in preventing net oral migration of colonic bacteria. Within 10 to 20 minutes of consumption of a meal, the IDMC that is in progress at the time of eating is inter­ rupted.44 The presence of intraluminal nutrients is sensed by mucosal nutrient contact, as evidenced by the fact that portally administered or intravenous nutrients do not have the same effects as those consumed orally.45 Several neural and humoral signals result from mucosal nutrient contact and are implicated in the induction of the fed motor pattern, including vagal afferent signals, cholecystokinin, and GLP-1. Moreover, the sensing of intraluminal nutrients is relatively complex, because different types of nutrients, or

Radiologic Observations Early radiologic observations of the small intestine in animals described several different patterns of wall motion and transit of intestinal contents. Walter Cannon42,51 observed both localized contractions over short segments of intestine in association with to-and-fro movement of con­ tents and intermittent episodes of propulsion of contents over greater distances caused by aborally traveling waves of peristalsis. In the fed state, the most common pattern of wall motion consisted of localized circular contractions that recurrently divided and formed short columns of chyme into new aliquots by temporary local occlusion of the lumen over distances of less than 1 to 2 cm, this pattern being labeled rhythmic segmentation.42,51 These contractions did not travel along the small intestine and did not result in much, if any, net oral movement of contents.42,51 Peristalsis also was commonly observed, often in combi­ nation with segmentation. During small intestinal nutrient loading, peristalsis was noted to have two forms: One was a slow advance of chyme over short distances in association with segmentation and the other was a rapid transit of chyme over longer distances, sometimes several loops, of the small intestine. The “fast peristalsis” was often seen in the cat duodenum.51 Similar observations have been made in other animal species32,42 and correlate with some of the motor patterns seen during clinical radiologic studies in humans (although these studies usually are performed when the subject is fasting and show the rapid peristaltic pattern more than the segmenting postprandial activity). Transit Time Observations The small intestinal transit time for a meal varies greatly according to the amount and nature of what is consumed, because both caloric content and physical form of a meal determine the gastric emptying rate and the rate of transport

1655

1656

Section X  Small and Large Intestine along the intestine.41,52-54 Depending on the test and param­ eter used, postprandial orocecal transit time usually is less than six hours. As assessed by lactulose breath testing, however, orocecal transit time can be as rapid as about 70 minutes with low nutrient loads. A systematic evaluation of the optimal conditions for nutrient loading is much needed to reveal abnormal small intestinal motor function, using transit studies. Manometric Observations Postprandial small intestinal motility is characterized by irregular phasic pressure waves without a discernible cycli­ cal pattern. Most small intestinal motility data are quite limited in spatial resolution because of the length of the small intestine. Nevertheless, most phasic pressures (pres­ sure wave sequences) are thought to travel only a short distance31,32 and probably represent the mixing and seg­ menting contractions noted in earlier radiologic studies.42,51 In animal studies, postprandial small intestinal motility is more segmenting than is fasting phase II activity, and phasic pressures occur less frequently and travel shorter distances along the bowel, resulting in slower transit of the contents.32 A similar suppression in the frequency of pressure wave sequences now has been found in the human duodenum.31 This segmenting motor pattern is thought to assist in mixing food with digestive enzymes and in maximizing the expo­ sure of food to the mucosa to optimize absorption.

Fasting Motor Pattern

During fasting, small intestinal motor activity adopts a repetitive cyclic motor pattern, the IDMC. The IDMC is absent in a number of disease states, presumably because of a primary neuropathic process. This absence is associated clinically with stasis of small intestinal contents, malab­ sorption, and small intestinal bacterial overgrowth. For detailed reviews, see the articles by Husebye43 and Sarna.44 Radiologic Observations Contrast agents can stimulate small intestinal mucosal receptors sensitive to pH, caloric content, and osmolarity changes. It is possible, therefore, that radiologic studies of “fasting” motility do not truly represent the fasting state. In general, however, contrast agents appear to move more swiftly through the small intestine during fasting than during the postprandial state and to be associated with more episodes of peristalsis over one or more loops and fewer segmenting contractions. When the phase of the IDMC is assessed concurrently (see later), little net movement of small intestinal contents is seen during phase I, but residual luminal contents are swept through the small intestine and into the terminal ileum during late phase II and phase III of the IDMC. This finding is not surprising because, by definition, phase I is the absence of measurable phasic pres­ sure waves, which are likely to be necessary to generate a sufficient intraluminal pressure gradient to cause intralumi­ nal flow. Transit Time Observations Studies of transit time through the small intestine also prob­ ably do not represent a true assessment of fasting motor function, because most of the substrates used to measure transit also interact with small intestinal mucosal receptors. The lower the caloric content, the more closely fasting motility will be assessed (see earlier). Manometric Observations The IDMC is defined manometrically and comprises three main phases. Phase I is defined as motor quiescence (less

than three pressure waves per 10 minutes at any one site); phase II is characterized by random pressure waves at less than the maximal rate; and phase III is characterized by pressure waves at the maximal rate (for the region) for longer than two minutes and, ideally, extending over a segment longer than 40 cm. Some authors also include a fourth phase (phase IV) as a transitional period between phases III and I, although this approach is not universal. Phases I and III are quite distinctive and easily recognized, whereas phase II can be recognized reliably only when sandwiched between phases I and III, because it superfi­ cially resembles the fed pattern. The phases of the IDMC start proximally and migrate distally over variable dis­ tances; few phase IIIs reaching the ileum.40 Phase III of each IDMC can start at any of a variety of locations; approxi­ mately one third of IDMCs have a gastroduodenal compo­ nent, and most onsets of phase III occur near the proximal jejunum.40 Because of the length of the small intestine and the velocity of travel of the IDMC, one part of the small intestine can be in phase I while other parts are in phase II or III (see Fig. 97-6). The normal periodicity of the IDMC varies greatly both within and between subjects; however, its median duration is 90 to 120 minutes.

CLINICAL CONSEQUENCES OF DISORDERED SMALL INTESTINAL MOTILITY Most of the time, the overall outcome of small intestinal motility is achieved without conscious awareness; a range of symptoms can arise, however, when an optimal outcome is not attained. Fortunately, like other organs, the small intestine has a substantial reserve capacity and copes with many insults, including infection, resection, inflammation, and denervation, before clinical problems become manifest. In IBS, the most common clinical syndrome in which altered motility is implicated, the sufferer’s physical wellbeing rarely is threatened even when symptoms are con­ siderable. Infrequently, the motor disturbances are severe enough to disrupt a person’s ability to maintain oral nutrition. The most important diseases and clinical settings asso­ ciated with abnormal small intestinal motility are listed in Table 97-1. Because these disorders are covered elsewhere in this book, they are mentioned here only with regard to the associated small intestinal motor disturbances. In IBS, a number of abnormalities of visceral sensation have been documented. These sensory abnormalities prob­ ably also lead to disordered motility; however, whereas motor abnormalities have been documented in some patients with IBS, they are absent in others (see Chapter 118). Because it appears likely that IBS is an as-yet-undefined generalized enteric neuropathy or low-grade neuroinflam­ matory disorder,55,56 failure to detect specific motor abnor­ malities might simply reflect our current poor understanding of normal small intestinal motor physiology and the rela­ tively gross measures by which motility has been assessed in patients with IBS. Small intestinal motility is severely disrupted in acutely ill persons and is increasingly recognized as an important factor to consider in postoperative and intensive care unit patients. Such motility disturbances likely result from several factors, including sepsis and drugs, which disrupt the slow wave rhythm; abdominal trauma and surgery, which stimulate reflex motor responses; and inflammatory mediators and cytokines, which affect neurotransmission within the CNS, ANS, and ENS. For a more detailed review,

Chapter 97  Small Intestinal Motor and Sensory Function and Dysfunction see the articles by Ritz and colleagues,57 and Chapman and colleagues.58 Pregnancy is known to alter the function of the lower esophageal sphincter, delay gastric emptying and disturb the frequency of gastric slow waves, and it is often asso­ ciated with constipation. In view of these widespread findings related to altered intestinal motility, it is likely that small intestinal motor function also is altered. In guinea pigs, the strength of the contraction of intestinal circular smooth muscle has been shown to be impaired during pregnancy by down-regulation of Gαq/11 proteins (which mediate contraction) and up-regulation of Gs alpha protein (which mediates relaxation).59 It is intriguing that G protein associations now are also being reported in functional gas­ trointestinal disorders, suggesting a final common pathway for sensorimotor intestinal disturbances.60 Diabetes has widespread effects on gastrointestinal motil­ ity. Acute effects result from changes in blood glucose levels, but they also can result from the autonomic neuropa­ thy that develops in patients with long-standing disease. As indicated predominantly by studies of the stomach, hyperglycemia can alter the rhythm of the slow wave, modulate sensory signaling, lead to changes in the tempo­ rospatial pattern of phasic contractions, and even stimulate inappropriate phase III–like IDMC activity in the small intestine. Metabolic disturbances of potassium, magnesium, and calcium homeostasis are likely to impair small intestinal motor function because these chemicals are vital for normal neuromuscular function. The effects of abnormal levels of these electrolytes on normal human small intestinal func­ tion have not been studied specifically, but in organ bath experiments, their alterations have caused gross distur­ bances in neural and muscular function. In addition, renal and hepatic failure are likely to alter small intestinal motil­ ity because of the multiple homeostatic inputs of the affected organs; however, altered motility usually is not a prominent clinical feature in these conditions. Many drugs affect small intestinal motility, especially those that alter ion transport, such as antidepressants, calcium channel blockers, and beta blockers. Sedatives and narcotic analgesics also alter motility but usually do not cause clinically important small intestinal motor dysfunc­ tion, except in critically ill patients or those with acute severe pain. Pseudo-obstruction, scleroderma and other connective tissue diseases, dysautonomia, visceral myopathies, and other rare diseases in which abnormal small intestinal motor function occur are discussed in detail in other chapters. These diseases may be uncommon causes of dis­ ordered small intestinal motility, but they have increased our understanding of normal motility, because in some cases, the neural and myopathic processes are impaired separately.

APPROACH TO PATIENTS WITH POSSIBLE SMALL INTESTINAL MOTOR DYSFUNCTION Taking a thorough history is a vital first step in approaching a patient who may have abnormal small intestinal motility. A review of exposures to drugs and toxins, family history, and, in the younger patient, milestones of growth and devel­ opment are especially important to consider. Findings on physical examination in this setting often are unremarkable. First-line investigations generally are suggested by the history, physical examination, and age of the patient and

may include a plain abdominal film (to look for dilated small intestinal loops, thickened bowel wall, or air-fluid levels), complete blood count with determination of red blood cell indices (to look for evidence of malabsorption), measurement of serum albumin and electrolyte levels, and random testing of blood glucose or glycosylated hemoglobin level. How much further to proceed with investigation depends on these results and on the severity of the patient’s condition. Special investigations may be indicated to answer par­ ticular questions. No standard approach has been recog­ nized, however, and local interest and expertise often determine which investigations are available. Fluoroscopy is widely available and can help exclude medically or surgi­ cally treatable problems. Endoscopy with small intestinal biopsy or aspiration is useful if celiac sprue, small intestinal bacterial overgrowth, or intestinal infection is considered likely. Analysis of stool may be necessary to exclude mal­ absorptive or secretory causes of small intestinal diarrhea. Small intestinal manometry, if available, can help distin­ guish neuropathic from myopathic forms of disordered motility, although in many settings, the abnormalities asso­ ciated with these two forms overlap (see Table 97-1). Manometry can show features typical of intestinal obstruc­ tion, although abdominal imaging by a variety of radiologic techniques is a better tool to identify an obstruction. In selected cases, full-thickness biopsy of the small intestine is necessary, but such biopsy should be performed only in centers with expertise in immunohistochemistry of intesti­ nal neurons, because standard histologic approaches often yield little useful information. Unfortunately, there are few therapies to date, beyond supportive measures, that can be offered to patients with disordered small intestinal motility. Nutritional status is of prime importance, and where patients can manage this independently, no further specific treatment may be needed. Symptomatic treatment approaches include modifications in diet (small frequent meals, lower fat intake), exercise (which is shown to improve bloating symptoms and expul­ sion of intestinal gas), antinausea agents, antispasmodics, and drugs to modulate sensory function. Thus far, there are no clinically available agents that spe­ cifically modify visceral hypersensitivity, and simple anal­ gesics, opiates, and antidepressants are all used. Apart from the tricyclic antidepressants and selective serotonin reup­ take inhibitors, there is little proof that these offer sig­ nificant benefit, and opiates can even worsen symptoms, leading to the narcotic bowel syndrome. Treatment of psy­ chological comorbidities also is important because anxiety and depression can heighten the perception of, and distress caused by, intestinal symptoms. Prokinetic agents have been limited in their therapeutic benefit, and because of safety concerns, availability of several, such as cisapride and tegaserod, has been restricted. There is hope, however, that prokinetics and visceral-specific analgesics might offer a better balance between safety and efficacy in the future.

KEY REFERENCES

Berthoud HR, Blackshaw LA, Brookes SJ, Grundy D. Neuroanatomy of extrinsic afferents supplying the gastrointestinal tract. Neurogastro­ enterol Motil 2004; 16 Suppl 1:28-33. (Ref 17.) Blackshaw LA, Gebhart GF. The pharmacology of gastrointestinal nociceptive pathways. Curr Opin Pharmacol 2002; 2:642-649. (Ref 14.) Cannon WB. The Mechanical Factors of Digestion. London: Edward Arnold, 1911. (Ref 42.) Furness JB. Types of neurons in the enteric nervous system. J Auton Nerv Syst 2000; 81:87-96. (Ref 8.)

1657

1658

Section X  Small and Large Intestine Hennig GW, Costa M, Chen BN, Brookes SJ. Quantitative analysis of peristalsis in the guinea-pig small intestine using spatio-temporal maps. J Physiol 1999; 517(Pt 2):575-90. (Ref 36.) Hunt JN, Smith JL, Jiang CL. Effect of meal volume and energy density on the gastric emptying of carbohydrates. Gastroenterology 1985; 89:1326-30. (Ref 53.) Husebye E. The patterns of small bowel motility: physiology and impli­ cations in organic disease and functional disorders. Neurogastroen­ terol Motil 1999; 11:141-61. (Ref 43.) Krauter EM, Strong DS, Brooks EM, et al. Changes in colonic motility and the electrophysiological properties of myenteric neurons persist following recovery from trinitrobenzene sulfonic acid colitis in the guinea pig. Neurogastroenterol Motil 2007; 19:990-1000. (Ref 27.) Sanders KM. Regulation of smooth muscle excitation and contraction. Neurogastroenterol Motil 2008; 20 Suppl 1:39-53. (Ref 2.) Sanders KM, Koh SD, Ward SM. Interstitial cells of Cajal as pacemakers in the gastrointestinal tract. Annu Rev Physiol 2006; 68:307-43. (Ref 4.)

Sarna SK. Cyclic motor activity; migrating motor complex: 1985. Gastroenterology 1985; 89:894-913. (Ref 44.) Schwizer W, Steingoetter A, Fox M. Magnetic resonance imaging for the assessment of gastrointestinal function. Scand J Gastroenterol 2006; 41:1245-60. (Ref 38.) Vermillion DL, Huizinga JD, Riddell RH, Collins SM. Altered small intestinal smooth muscle function in Crohn’s disease. Gastroenter­ ology 1993; 104:1692-9. (Ref 26.) Ward SM, Sanders KM. Involvement of intramuscular interstitial cells of Cajal in neuroeffector transmission in the gastrointestinal tract. J Physiol 2006; 576:675-82. (Ref 5.) Wood JD. Enteric nervous system: reflexes, pattern generators and motility. Curr Opin Gastroenterol 2008; 24:149-58. (Ref 9.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

98 Colonic Motor and Sensory Function and Dysfunction Ian J. Cook, Simon J. Brookes, and Philip G. Dinning

CHAPTER OUTLINE Anatomy and Basic Control Mechanisms of the Colon and Anorectum  1660 Macroscopic Structure of the Colon  1660 Structure and Activity of Colonic Smooth Muscle  1660 Ion Channels in Colonic Smooth Muscle  1661 Interstitial Cells of Cajal  1661 Innervation of the Colon  1662 The Enteric Nervous System  1662 Sympathetic Innervation  1664 Parasympathetic Innervation  1664 Extrinsic Afferent Pathways  1664 Anorectal Anatomy and Innervation  1664 Relationships among Cellular Events, Pressure, and Flow  1665 Recognizable Colonic and Anorectal Motor Patterns and Putative Functions  1665 Nonpropagating Motor Patterns  1665 Propagating Motor Patterns  1665 Rectal Motor Complexes  1665 Regional Variation of Propagating Sequences  1665 Regional Linkage among Propagating Sequences  1665

Each day, 1200 to 1500 mL of ileal effluent enter the colon, 200 to 400 mL of which are finally excreted as stool. The colon mixes its contents to facilitate the transmural exchange of water, electrolytes, and short-chain fatty acids and stores stool for extended periods. The mixing process involves rhythmic to-and-fro motions, together with short stepwise movements of contents, resulting in an overall net aboral flow rate that averages approximately 1 cm per hour. When dehydration threatens survival, such as with water deprivation or severe diarrhea, the ability of the colon to reabsorb fluids is of major physiologic significance; appropriate motility patterns are important in achieving this function. For example, the colon has the capacity to increase its fluid absorption five-fold when required, but this ability is greatly impaired when transit is accelerated. Under normal circumstances, viscous contents occasionally are propelled aborally at a rapid rate, and, if circumstances are appropriate, stool is evacuated under voluntary control. Thus, the colon is capable of showing a diverse range of motor patterns that are suited for particular physiologic functions. The generic term motility describes the range of motor patterns and the mechanisms that control them. Common sensorimotor symptoms, such as constipation, diarrhea, bloating, abdominal pain, or rectal urgency, can arise from disturbances of ileocolonic delivery, colonic propulsion, or stool expulsion. Clearly, these symptoms and dysmotility must be linked, although our current under-

Regulation of Colonic Filling and Transit  1666 Role of the Ileocecal Junction  1666 The Colon as a Storage Organ  1666 Relationships between Colonic Motor Patterns and Flow  1667 Defecation  1668 Rectal Filling, Capacitance, and Accommodation and Motility of the Anal Sphincters  1669 Anorectal Motility during Defecation  1670 Modulators of Colonic Motility  1670 Physiologic  1670 Pharmacologic  1671 Nonpharmacologic  1672 Disorders of Colonic Motility  1672 Constipation  1672 Diarrhea  1673 Irritable Bowel Syndrome  1673 Colonic Motility Disturbances Secondary to Nonmotor Intestinal Disorders  1673

standing of such linkages is limited, largely because of technical difficulties involved in studying the human colon. Because of differences among species, care is required in extrapolating data from animal studies to humans. For many years, intraluminal motility recordings in humans were obtained mainly from the rectum and sigmoid, but it is now clear that the motor activity of these distal regions is not representative of the colon as a whole. The contents of the colon become increasingly viscous distally, and this alteration complicates the relationship between propulsion and the contractile activity of the smooth muscle. Colonic movements are much less frequent and transit is considerably slower than in other regions of the gastrointestinal tract. The highly propulsive, stereotypical motor patterns that are associated with stool expulsion generally occur only once or twice daily. Hence, study of the motor patterns in the human cannot be achieved using contrast radiography. Prolonged recording techniques must be used to capture such infrequent motor patterns. Recording of intraluminal pressure, by means of manometric catheters inserted per rectum, requires prior bowel cleansing, which can modify colonic motility. Furthermore, interpretation of intraluminal pressure measurements is complicated, because many contractions of the colonic wall do not occlude the lumen and therefore are detectable manometrically only if they cause significant pressure changes. Measurement of colonic wall tone using a barostat

1659

1660

Section X  Small and Large Intestine provides information on these nonocclusive colonic wall movements, but it tells us nothing of the spatiotemporal patterning of motility. Smooth muscle electromyography provides insight into the patterning of muscle activity but generally requires access to the muscular wall of the colon, which ethically is problematic in humans. Scintigraphy with suitably high frame rates can resolve discrete movements of the contents but is suboptimal for measuring actual wall motion. In vitro study of the cellular basis of motility using isolated specimens of colon faces fewer technical and ethical limitations; however, data obtained at the cellular level, often under highly nonphysiologic conditions, can be difficult to extrapolate to the more complex integrated responses of the entire organ in vivo. Nonetheless, while recognizing the intrinsic limitations of all these measurement techniques, in combination they have allowed us to piece together a number of concepts that have provided important insights into the relationships among muscle activity, wall motion, intraluminal pressure, and flow.

ANATOMY AND BASIC CONTROL MECHANISMS OF THE COLON AND ANORECTUM MACROSCOPIC STRUCTURE OF THE COLON

The human colon is just over one meter long and is divided anatomically into the cecum; the ascending, transverse, descending, and sigmoid colon; and the rectum, which lies between the rectosigmoid junction and the anal canal. The outer longitudinal smooth muscle layer forms three thick, cord-like structures called the teniae coli, which are spaced evenly around the circumference of the colon. Between the teniae, the longitudinal smooth muscle is much thinner, allowing the wall to bulge noticeably. Irregularly spaced circumferential constrictions pinch the colon into a series of pockets, called haustra, which give the colon a sacculated appearance for much of its length. Haustra are not fixed structures and appear to be caused by sustained contractions of the circular muscle. Myogenic activity alone, however, does not seem sufficient to explain haustration, and neural input is likely to contribute to their formation; haustra move, disappear, and re-form during the propulsion of colonic contents. The teniae fuse to form a continuous outer longitudinal smooth muscle layer at the rectosigmoid junction, which then continues down to the distal margin of the anal canal, insinuating itself between the internal and the external anal sphincters. Throughout the length of the colon, the circular smooth muscle layer consists of thick bundles of cells, which are separated by connective tissue septa. The internal anal sphincter consists of a thickening of the circular muscle layer over the last 2 to 4 cm of the anal canal. Macroanatomy of the colon is also discussed in Chapter 96.

STRUCTURE AND ACTIVITY OF COLONIC SMOOTH MUSCLE Structure

Smooth muscle cells in the human colon, as in other muscular organs, are spindle-shaped, nucleolated cells with tapered ends. The surface area of the smooth muscle cell membrane is increased greatly by numerous caveolae, or small pits. Individual smooth muscle cells are connected mechanically to neighboring cells by intermediate junctions and electrically by gap junctions, which allow ions and

small molecules—those with molecular weights up to approximately 1000 kD—to diffuse between the cells, thereby ensuring that the cells are functionally coupled to one another. Therefore, the smooth muscle cells do not contract as individual cells; rather, they contract together in a large, coordinated assembly, a syncytium.

Activity

Like smooth muscle throughout the gastrointestinal tract, colonic smooth muscle shows spontaneous, oscillatory electrical activity, even when all neural activity is blocked. Two types of rhythmic myoelectrical activity occur1: myenteric potential oscillations (MPOs) and slow waves. MPOs are small-amplitude, rapid oscillations, with a frequency of 12 to 20 per minute, that originate from the plane of the myenteric plexus. These small oscillations spread, by means of gap junctions, into both the longitudinal and the circular smooth muscle layers and often reach the threshold potential for generating smooth muscle action potentials in both muscle layers. In the circular muscle layer, MPOs, with superimposed action potentials, generate small phasic contractions of the circular muscle layer. When the muscle is strongly excited by neurotransmitters released by enteric excitatory motor neurons, each MPO evokes an action potential, and the phasic contractions summate into powerful contractions that last several seconds. Although the functions of the colon circular smooth muscle are well understood, the role that the longitudinal muscles play in colonic motility, mixing, and propulsion is a matter of some controversy. The longitudinal muscle probably acts in an antagonistic role to the circular muscle, contracting largely in concert with it and thus preventing excessive lengthening when the circular muscle contracts, which would be mechanically disadvantageous. It might also contribute to propulsion by pulling the colon over its contents, so that circular muscle contractions gain more purchase on them. Some evidence from modeling suggests that it also might play a role in mixing of liquid contents, at least in the small intestine. A second pacemaker region is located at the submucosal border of the circular muscle. This region produces largeramplitude, slower myogenic oscillations in membrane potential called slow waves, which also spread through the thickness of the circular smooth muscle by means of gap junctions. Slow waves also often reach the threshold for triggering smooth muscle action potentials and can evoke strong contractions. Slow waves occur throughout the human colon at a frequency of approximately 2 to 4 per minute. In the small intestine, a gradient of slow wave intrinsic frequencies causes slow waves to propagate predominantly aborad. This is not the case in the colon: Slow waves propagate over short distances up or down the colon, and complex interactions occur as waves coming from different initiation sites collide, leading to mixing of contents with little propulsion. The currents produced by pacemaker cells at the myenteric and submucosal borders decay as they spread through the thickness of the circular muscle layer. Thus, operating in the middle of the circular smooth muscle layer is complex spontaneous electrical activity consisting of a mixture of MPOs and slow waves, with superimposed smooth muscle action potentials. Most of the time, slow waves determine the contractile activity of the smooth muscle and cause nonpropulsive mixing movements. During times of strong enteric neuronal activity, however, MPO-related contractions summate, giving rise to powerful patterned contractions of much longer duration than those produced by slow waves. These contractions can propagate for long distances

Chapter 98  Colonic Motor and Sensory Function and Dysfunction along the colon and are known as propagating sequences. Action potentials in the smooth muscle can be recorded in vivo with electrodes attached to the serosal surface, thereby giving a high-resolution measurement of myoelectric activity or spike bursts.

ION CHANNELS IN COLONIC SMOOTH MUSCLE

The membrane of colonic smooth muscle cells contains a variety of ion channels, including several types of potassium channels, calcium channels, chloride channels, and nonselective cation channels.2 Although the exact physiologic roles of many of these ion channels are unknown, the high-threshold, voltage-operated calcium channels (L-type calcium channels) do play a crucial role in colonic muscle contractility. These channels open when the membrane potential of smooth muscle cells is depolarized beyond a voltage threshold, and they are responsible for the rapid upstroke of smooth muscle action potentials. The influx of calcium through L-type calcium channels during action potentials is a major trigger for activation of the contractile apparatus. It is not surprising that pharmacologic blockade of L-type calcium channels by dihydropyridine drugs such as nifedipine can reduce the contractility of colonic smooth muscle substantially. Release of calcium from intracellular stores, which is triggered by excitatory neurotransmitters, also may play a role in muscle contraction.

and ICCSM form extensive networks along the colon and are electrically coupled to one another and to the smooth muscle layers by gap junctions (Figs. 98-1 and 98-2). ICCMY probably are the pacemakers for the small, rapid (12-20/ min) oscillations in membrane potential (MPOs) of longitudinal and circular smooth muscle layers. ICCSM are the pacemakers for the large-amplitude slow waves (2-4/min) originating in the plane of the submucosal plexus; these slow waves have a powerful influence on the patterning of circular muscle contraction. The exact ionic basis of rhythmicity in ICCMY and ICCSM that gives rise to MPOs and slow waves is not entirely clear; however, oscillations in membrane potential are an intrinsic

Mucosa ICCSM CM Nerve ICCIM CM ICCMP

INTERSTITIAL CELLS OF CAJAL

LM

Since 1991, the interstitial cells of Cajal (ICC) have been shown to play at least two important roles in the control of gastrointestinal motility: control of myogenic activity and mediating or amplifying the effects of motor neurons on the smooth muscle apparatus. ICC are non-neuronal in origin and are derived from common progenitors of smooth muscle cells. Mutant mice and rats that are deficient in ICC have profoundly disturbed intestinal motility, an observation that provides insight into the roles of ICC in the human gastrointestinal tract. In the human colon, three types of ICC are recognized and are named according to their locations: ICC in the plane of the myenteric plexus (ICCMY), ICC near the submucosal plexus (ICCSM), and intramuscular ICC located between the circular and the longitudinal muscle layers (ICCIM). ICCMY

A

B

Figure 98-1.  Schematic cross section of the muscularis externa of the human colon. The outer longitudinal smooth muscle layer (LM) is thickened at the teniae. In the plane of the myenteric plexus (not shown) is a network of interstitial cells of Cajal (ICC), which generate a rapid myenteric potential oscillation (ICCMY; see lower waveform on the right). The circular muscle layer (CM) is innervated by axons of enteric motor neurons with transmitter release sites (clusters of clear vesicles) that are associated with specialized intramuscular ICC (ICCIM). At the outer border of the circular muscle is another network of submucosal ICC, which generate slow waves (ICCSM; see upper waveform on the right). Also present are axons of motor neurons in the longitudinal muscles and ICCIM (not shown in this cross section). The tiny white squares represent gap junctions, which electrically couple cells.

100 µm

Figure 98-2.  Micrographs of interstitial cells of Cajal (ICC) in the human colon, labeled by c-Kit im­munohistochemistry. A, ICC in the plane of the myenteric plexus (ICCMY) have an irregular shape, form a dense network of cells, and probably function as pacemakers. B, A different plane of focus of the same region shows spindle-shaped intramuscular ICC (ICCIM) in the overlying circular muscle layer. These cells probably are involved in neuromuscular transmission to the smooth muscle. (Courtesy of Liz Murphy and David Wattchow.)

1661

1662

Section X  Small and Large Intestine property of both ICCMY and ICCSM. Intramuscular ICC (ICCIM) are a major target of neurotransmitters released from the axons of excitatory and inhibitory enteric motor neurons. Acetylcholine and nitric oxide (and probably several other motor neuron transmitters) evoke changes in the membrane potential of ICCIM, which then spread through the smooth muscle by means of gap junctions. ICCIM also may be involved in amplifying the slow waves as they spread through the muscle layers. Thus, these cells appear to be key players in integrating non-neuronal pacemaker activity and neuronal inputs to the smooth muscle. The discovery that cellular mechanisms long considered to be the properties of smooth muscle cells actually are mediated by ICC may have important clinical implications. For example, in the distal bowel, reduced numbers of ICC, or a reduction in the total volume of ICC, has been associated with anorectal malformations, colonic manifestations of Chagas’ disease, and possibly some cases of slow-transit constipation.3 Some reports have suggested that the density of ICC may be reduced in aganglionic segments of colon in Hirschsprung’s disease and that this deficit might contribute to diminished propulsive activity; this finding, however, has not been consistent between studies.3

INNERVATION OF THE COLON 1 THE ENTERIC NERVOUS SYSTEM

Direct neuronal control of colonic motility is mediated mostly by the enteric nervous system (ENS). Although the ENS is capable of expressing a diverse repertoire of motor patterns, its functions are modulated by sympathetic, parasympathetic, and extrinsic afferent pathways (Fig. 98-3). In terms of numbers of nerve cells, the ENS is by far the largest component of the autonomic nervous system, with considerably more neurons than those of the parasympathetic and sympathetic divisions combined. The nerve cell bodies of the ENS are located in plexuses of myenteric ganglia (Auerbach’s plexus), which lie between the longitudinal and the circular muscle layers of the muscularis externa, or in the submucosal ganglia, which lie between the circular muscle and the mucosa (Fig. 98-4). The submucosal plexus is divisible into at least two networks: Meissner’s plexus, which lies closer to the mucosa, and Schabadasch’s plexus, which lies adjacent to the circular muscle. Some authors have identified an additional intermediate plexus. Internodal strands that contain hundreds of axons run within and between the different plexuses. Finer nerve trunks innervate the various target tissues of the intestinal wall, including the longitudinal muscle layer, circular muscle, muscularis mucosae, mucosal crypts, and mucosal epithelium. Within the ganglia of each plexus, different functional classes of enteric nerve cell bodies are intermingled, although differences in the proportions of cell types among the plexuses have been observed. It has become clear that an exquisite degree of organization is characteristic of the ENS, each class of nerve cell making highly specific and precise projections to its particular target. The ENS uses many transmitters in addition to the major transmitters acetylcholine and nitric oxide, including tachykinins, purines, numerous other modulatory peptides, and some amines. Many other substances, released from neural and non-neural cells, also modulate neuronal and muscular excitability, including gaseous mediators (carbon monoxide and hydrogen sulfide) and, in inflammation, prostanoids, cytokines, purines, bradykinin, H+ ions, and neurotrophins.

Vagus nerve Lumbar colonic nerves

Nodose ganglion DRG

Ascending colon

Transverse colon

Descending colon

Inferior mesenteric plexus Pelvic plexus ganglia

Sigmoid colon Rectum External anal sphincter

Lesser splanchnic nerve

DRG

Pelvic nerves

Pudendal nerve Figure 98-3.  The extrinsic innervation of the human colon. Parasympathetic efferent pathways ( filled cell bodies) arise from the dorsal motor nucleus of the vagus in the brainstem and pass through the vagus nerve and prevertebral sympathetic ganglia, through the lumbar colonic nerves to the proximal colon. Parasympathetic pathways also extend from nuclei in the sacral spinal cord that run through the pelvic nerves and either synapse in the pelvic plexus ganglia or run directly into the bowel wall. Sympathetic pathways (open cell bodies) consist of preganglionic neurons in the thoracic spinal cord that synapse with sympathetic postganglionic neurons either in the inferior mesenteric plexus or in the pelvic plexus. Enteric nerve cell bodies in the colon receive input from both parasympathetic and sympathetic pathways. Viscerofugal enteric neurons project out of the bowel to the prevertebral ganglia. Afferent pathways consist of vagal afferent neurons from the proximal colon with cell bodies in the nodose ganglia. In addition, spinal afferent neurons with cell bodies in lumbar dorsal root ganglia (DRG) run through the lesser splanchnic and colonic nerves to the colon and mediate nociception. Another population of spinal afferents, with cell bodies in the sacral DRG, runs through the pelvic nerves and pelvic ganglia to the rectum; these include sensory neurons that transmit non-nociceptive information about the distention of the rectum. The striated muscles of the pelvic floor (including the external anal sphincter) are supplied by motor neurons with cell bodies in the spinal cord and axons that run in the pudendal nerves. Triangles represent transmitter release sites; combs represent sensory transduction sites. Anus

Primary Afferent Neurons

Much of the motor and secretory activity of the intestine can be conceptualized as a series of reflexes evoked by mechanical or chemical stimuli. These reflexes involve activation of enteric primary afferent neurons, integration by interneurons, and execution of appropriate responses by motor neurons. The first neurons in these reflex circuits are primary afferent neurons (sometimes called “sensory” neurons, although they do not give rise to conscious sensation). These neurons are located in both myenteric and submucosal plexuses and characteristically have several long axonal processes. Some primary afferents fire action potentials in response to stretch or tension in the bowel

Chapter 98  Colonic Motor and Sensory Function and Dysfunction

Mucosa Meissner’s submucosal plexus

Submucosa

Schabadasch’s submucosal plexus

Circular muscle

Auerbach’s myenteric plexus

Longitudinal muscle

Oral Sensory neuron ending

Aboral Motor (output) neuron ending

Interneuron axon projection

Figure 98-4.  Diagram showing the layers and components of the intestinal wall. The lumen is at the top and the longitudinal muscle layer is at the bottom. Auerbach’s myenteric plexus and the submucosal plexuses (Meissner’s and Schabadasch’s plexuses) are shown, along with some of their major classes of enteric neurons. The networks of interstitial cells are shown in Figure 98-1.

wall; others are activated by chemical or mechanical stimuli of the mucosa. These mucosal stimuli probably work, at least in part, by activating specialized enteroendocrine cells in the mucosal epithelium, such as the serotonin-containing enterochromaffin cells. The primary afferent neurons then release synaptic transmitters, such as acetylcholine or tachykinins or other peptides, to excite other classes of enteric neurons in nearby ganglia. Enteric primary afferent neurons also make excitatory synaptic contacts onto other neurons of their own class, so that they fire in coordinated assemblies.

cyclase-activating peptide (PACAP). The varicose transmitter release sites of inhibitory motor neurons also are associated with ICCIM, just as are the release sites of excitatory motor neurons. Interstitial cells probably mediate a large component of the electrical effects on smooth muscle of neurotransmitters released by enteric motor neurons. Inhibitory motor neurons usually are tonically active, modulating the ongoing contractile activity of the colonic circular smooth muscle. Inhibitory motor neurons are particularly important in relaxing sphincteric muscles in the ileocecal junction and the internal anal sphincter.

Motor Neurons

Interneurons

Enteric motor neurons typically have smaller cell bodies than afferent neurons, with a few short dendrites and a single long axon. Separate populations of motor neurons innervate the circular and longitudinal muscle layers. Excitatory motor neurons synthesize acetylcholine, which they release from their varicose endings in the smooth muscle layers; some also release the tachykinin peptides, substance P and neurokinin A, which excite smooth muscle. Typically, axons of excitatory motor neurons project either directly to the smooth muscle close to their cell bodies or orad for up to 10 mm.4 Once in the smooth muscle layers, the axons turn and run parallel to the smooth muscle fibers for several millimeters; they branch extensively and form many small varicosities, or transmitter release sites, closely associated with intramuscular ICC (ICCIM). Inhibitory motor neurons typically are slightly larger than excitatory motor neurons, and there are fewer of them. They also have short dendrites and a single axon but, unlike excitatory motor neurons, they project aborally to the smooth muscle layer for distances of 1 to 15 mm in the human colon.4 Once the axon reaches the smooth muscle, it branches extensively to form multiple varicose release sites. Inhibitory motor neurons release a cocktail of transmitters that inhibit smooth muscle cells, including nitric oxide, adenosine triphosphate (ATP), and peptides, such as vasoactive intestinal polypeptide (VIP) and pituitary adenyl

When a region of colon is stimulated, such as by a bolus that distends it, intrinsic primary afferent neurons (IPANs) are activated. These neurons then activate excitatory and inhibitory motor neurons, which, because of their polarized projections, cause contraction of the muscle orad to the bolus and relaxation aborally. These effects tend to propel the contents aborally. From the new position of the bolus, another set of polarized reflexes is triggered, and peristaltic propulsion results. The ascending excitatory reflex and the descending inhibitory reflex sometimes are called “the law of the intestine.” These reflexes spread farther than is predicted by the projections of the excitatory and inhibitory motor neurons, because interneurons also are involved in these reflex pathways. Ascending cholinergic interneurons in the human colon have axons that project up to 40 mm orad and extend the spread of ascending excitatory reflex pathways. In addition, several classes of descending interneurons are present in the human colon, with axons that project up to 70 mm aborally. Some of these interneurons are involved in spreading descending inhibition along the colon, but others are likely to be involved in the pro­ pagation of migratory contractions. It also is likely that some interneurons are themselves stretch sensitive, thereby functioning as primary afferent neurons. In addition to the sensory neurons, interneurons, and motor neurons, viscerofugal nerve cells project to the sympathetic prevertebral

1663

1664

Section X  Small and Large Intestine ganglia, vasomotor neurons innervate blood vessels, and secretomotor neurons stimulate secretion from the colonic epithelium.

SYMPATHETIC INNERVATION

The major sympathetic innervation of the proximal colon arises from the inferior mesenteric ganglion and projects through the lumbar colonic nerves to the ascending and transverse colon (see Fig. 98-3). A small number of sympathetic neurons in the celiac and superior mesenteric ganglia, in the paravertebral chain ganglia, and in the pelvic plexus ganglia also project to the colon (see Fig. 98-3). These neurons receive a powerful cholinergic drive from preganglionic nerve cell bodies in the intermediolateral column of the spinal cord (segments L2-L5). This is a major pathway by which the central nervous system modifies bowel activity, such as during exercise. Sympathetic efferent neurons also receive input from the enteric viscerofugal neurons and from extrinsic, spinal sensory neurons with cell bodies in the dorsal root ganglia, forming several reflex loops with the distal bowel. Sympathetic nerve fibers from prevertebral ganglia cause vasoconstriction of the mucosal and submucosal blood vessels. Other cells project to the enteric ganglia, where they cause presynaptic inhibition of synaptic activity in the ENS and thus depress reflex motor activity. Another target for sympathetic axons is the circuitry of the submucosal plexus (largely Meissner’s plexus) involved in controlling epithelial secretion. Hence, these pathways inhibit colonic motor activity, reduce blood flow, and inhibit secretion to limit water loss from the body during times of sympathetic activation. In addition, some sympathetic axons innervate the smooth muscle directly, particularly the ileocecal junction and internal anal sphincter, where they cause contraction; these effects also are consistent with closing down enteric motor activity during sympathetic arousal.

PARASYMPATHETIC INNERVATION

The colon receives parasympathetic innervation from both the vagus nerve and pathways in the sacral spinal cord. Branches of the vagus nerve reach the prevertebral ganglia (superior hypogastric plexus) and then run with sympathetic axons to the cecum and the ascending and transverse colon. The distal colon is supplied largely by sacral parasympathetic axons via the pelvic nerves (pelvic splanchnic nerves). Some of these cholinergic spinal efferent neurons synapse first onto nerve cell bodies in the pelvic plexus (inferior hypogastric plexus), and others project directly to the colon. From their point of entry into the colon, many of the axons run in an oral direction and form thick trunks called shunt fascicles. Parasympathetic axons project to the enteric ganglia in the colon, where they make excitatory cholinergic synapses onto enteric nerve cell bodies. Sacral parasympathetic pathways play an important role in increasing the propulsive activity of the distal colon before defecation and also may be involved in triggering the pro­ pagating complexes that start more proximally before defecation.

EXTRINSIC AFFERENT PATHWAYS

Sensation from the colon is mediated by primary afferent neurons with cell bodies outside the bowel wall. Vagal afferent neurons, with nerve cell bodies located in the nodose and jugular ganglia, project to the proximal colon and run with the vagal efferent parasympathetic pathways. Currently, their exact role in reflex control and sensation is not clear, but they are unlikely to be involved in the transmission of pain sensation from the colon.

The entire colon also is innervated by spinal primary afferent neurons with nerve cell bodies in the lumbar dorsal root ganglia. Lumbar spinal afferents project along the lumbar splanchnic nerves, through the prevertebral inferior mesenteric ganglion, and through the lumbar colonic nerves to the colon, where they terminate in sensory endings in the mesentery, serosa, muscular layers, and mucosa throughout the entire colon and rectum. In addition, a population of spinal afferents, with cell bodies in the sacral dorsal root ganglia, projects along the pelvic nerves to the colon and traverses the pelvic plexus en route. Evidence indicates that some of these sacral spinal afferent neurons form a functionally different population from the lumbar spinal afferents, encoding different types of information, particularly from the rectum. Sacral afferents include many mechanoreceptors with a low threshold and wide dynamic range; these mechanoreceptors probably are responsible for graded sensations of rectal filling and for activating defecatory reflexes.5 By contrast, lumbar spinal afferents and some higherthreshold sacral afferents are responsible for generating pain sensations from all regions of the colon and rectum. They respond to gross distention of the bowel wall, traction on the mesenteric membranes, powerful colonic contractions, or chemical stimulation of the mucosa by bile acids, high osmolarity, and other stimuli. It is well established that the sensitivity of many spinal afferents is increased greatly by inflammation in the colon wall. In addition to their role in sensation, spinal afferents also have axon branches (colla­t­ erals) in enteric ganglia and prevertebral sympathetic ganglia and on mucosal blood vessels, where they might play a role in generating peripheral reflex responses to noxious stimuli. In summary, sacral afferent and efferent (parasympathetic) pathways run in parallel and connect the distal bowel with neural circuitry in the sacral spinal cord via pelvic and rectal nerves. The important role of these pathways in both rectal sensation and in generating the enhanced motility required for defecation is clearly demonstrated by the effects of nerve lesions at several levels. Thus, severing of peripheral nerves and distal spinal cord injury can lead to loss of rectal sensation and to severely impaired defecatory ability.

ANORECTAL ANATOMY AND INNERVATION Although the rectum is in direct continuity with the colon, the longitudinal muscle layer within this region is not organized into teniae; rather, it forms a continuous outer layer, uniformly encircling the rectum, and insinuating between the internal and external anal sphincters to the distal end of the anal canal. The narrowed distal rectum, or anorectal junction, is formed by the longitudinal muscle coat of the rectum, which is joined by the sling fibers of the pubo­ rectalis muscle, attachments of the levator ani muscles, and proximal margins of the internal and external anal sphincters. The puborectalis and levator ani muscles have important roles in maintaining continence and in defecation. These striated muscles form part of the pelvic floor and are in a state of constant tone that serves to pull the rectum anteriorly and elevate it, thereby reducing the anorectal angle; this mechanical effect tends to prevent entry of stool into the upper anal canal. The internal anal sphincter is a thickened band of smooth muscle, with relatively high spontaneous tone, that is in

Chapter 98  Colonic Motor and Sensory Function and Dysfunction continuity with the circular smooth muscle of the rectum. By contrast, the external anal sphincter is a striated muscle and is located distal to, but partly overlying, the internal sphincter. The external sphincter also has a high resting tone, but unlike that of its internal counterpart, its tone can be influenced by voluntary efforts, to help maintain continence. As expected, the sources of innervation of the internal and external anal sphincters are different. The internal sphincter directly receives a powerful inhibitory innervation from intrinsic, enteric inhibitory motor neurons and also extrinsic input from lumbar sympathetic and sacral parasympathetic nerves that project via the pelvic plexus ganglia. The external anal sphincter and other pelvic floor muscles are innervated, through the pudendal nerve (S3S4), by motor neurons with cell bodies in the spinal cord. The rectum and proximal anal canal are richly supplied with sensory receptors that respond to rectal stretch and the composition of the intraluminal contents. These receptors are important for detecting rectal filling, triggering sensations of urgency, facilitating rectal accommodation, and differentiating the composition (stool or gas) of rectal content (see Chapters 96 and 125).

ICCSM at the submucosal border of the circular muscle.1 The timing of these nonpropagating contractions is affected relatively little by enteric motor neural activity but is very dependent on the degree of wall distention. This nonpro­ pagating activity also displays a circadian rhythm, being significantly reduced during sleep.

PROPAGATING MOTOR PATTERNS

When excitatory motor neurons are active, contractions evoked by MPOs summate, giving rise to powerful lumenocclusive contractions that can last longer than slow waves and that can propagate substantial distances along the colon. These identifiable colonic motor patterns are commonly referred to as propagating sequences or high-amplitude propagating sequences (also termed propagating contractions or high-amplitude propagating contractions [HAPCs]). Depending upon their direction of propagation, these patterns can be further qualified by the terms antegrade (aboral) or retrograde (orad). In the healthy colon, antegrade pro­ pagating sequences are recorded with a three-fold higher frequency than retrograde propagating sequences.6 As with nonpropagating activity, propagating sequences display nocturnal suppression and can be stimulated by a meal.

RECTAL MOTOR COMPLEXES

RELATIONSHIPS AMONG CELLULAR EVENTS, PRESSURE, AND FLOW Smooth muscle activation often is divided into two com­ ponents. The first component is the tonic, ongoing activation that gives smooth muscle its basal resistance to stretch, its tone. The second component comprises the dynamic, phasic contractions that mix and propel contents. Compliance is a term used to describe the extent to which the bowel wall can stretch to accommodate contents. For example, a muscle that is very distensible—for example, because of powerful inhibitory motor neuron activity—is said to have a high compliance. During phasic contractions, a transient increase occurs in the resistance of the bowel wall to stretch, namely, a decrease in its compliance. If bowel contents are fluid and no downstream resistance is present to impede flow, the smooth muscle rapidly shortens. The contents are then propelled, with a minimal increase in intraluminal pressure. By contrast, if resistance to forward flow of contents is encountered, as by a lumenoccluding contraction occurring distally, the smooth muscle does not shorten significantly, although its tension increases. This increase in tension increases intraluminal pressure, but it does not cause propulsion. In most situations in vivo, smooth muscle contraction causes a mixture of shortening, increased tension, increased pressure, and propulsion. The process of propagation is controlled by pathways intrinsic to the enteric neural circuitry and by triggering sequences of polarized reflexes that cause peristaltic propulsion.

RECOGNIZABLE COLONIC AND ANORECTAL MOTOR PATTERNS AND PUTATIVE FUNCTIONS NONPROPAGATING MOTOR PATTERNS

Nonpropagating, apparently random activity makes up the majority of the recorded colonic motor activity and is presumed to serve segmenting or mixing functions. The frequency of nonpropagating colonic contractions in vivo is generally 2 to 4 cycles per minute, similar to the frequency of the spontaneous myogenic slow waves generated by

Periodic contractile activity predominates in the sigmoid colon and rectum. This activity is commonly termed the rectal motor complex (RMC) or periodic rectal motor activity (PRMA). The mean RMC amplitude ranges from 15 to 60 mm Hg with a duration of 3 to 30 minutes.6 In contrast to all other colonic contractile patterns, the circadian trend for RMCs is reversed, i.e., the RMC is more prevalent during sleep, suggesting the relevance of the extrinsic neural control of this pattern. The relationship between the RMC and flow is still incompletely understood. RMCs can be triggered by propagating pressure waves from the proximal colon and by the arrival of stool or gas from the sigmoid colon,7 suggesting the RMC provides a braking mechanism to keep the rectum empty.

REGIONAL VARIATION OF PROPAGATING SEQUENCES

Contractile activity in the human colon demonstrates marked regional variation. For example, propagating pressure waves originate nearly four times as frequently in the proximal colon than in the distal colon (Fig. 98-5). The mean distance covered by antegrade pressure waves arising from the cecum is 50 cm, compared with only 20 cm for sequences originating in the descending colon. Still, pressure waves arising proximally generally do not propagate beyond the mid-colon (see Fig. 98-5). It is now clear that slower propagation rates favor the effective propulsion of contents. The conduction velocity of pressure waves increases as the waves migrate caudally. Indeed, such events often accelerate to the point of synchronicity, which arrests the progress of content moving ahead of the contractile front. In addition, nonpropagating (segmenting) pressure waves make up a higher proportion of activity in the distal colon than in more proximal regions.6 Thus, most motor activity in the distal colon functions to retard forward flow, thereby minimizing challenges to continence.

REGIONAL LINKAGE AMONG PROPAGATING SEQUENCES

In addition to the regional variation in propagating sequence frequency, these sequential motor patterns are linked in an

1665

Section X  Small and Large Intestine organized spatiotemporal pattern.8 Many of these regionally linked propagating sequences also form series in which three or more consecutive propagating sequences demonstrate a regional shift in the same direction. Each propagating sequence in a linked series originates in either a more proximal or more distal colonic location. Although most single propagating sequences do not span the length the colon, collectively a series of linked propagating sequences can do so (Fig. 98-6), and it is likely that such linkage is

Cecum

Rectum

20

50

10

25

0 Cecum Hepatic flexure

Splenic flexure

Sigmoid colon

0 Rectum

Propagating pressure waves per colonic region (per 24 hr)

Antegrade propagating sequence frequency (per 24 hr)

1666

Colonic regions Figure 98-5.  Regional variation in the frequency of initiation and extent of propagation of antegrade propagating sequences. The blue histogram (bottom) shows the distribution of antegrade propagating sequences grouped according to the site of origin. The green horizontal bars at the top show the mean extent of propagation by sequences originating at the same site. Note that propagating sequences originate significantly more frequently in the proximal than in the distal colon, and the extent of propagation is much greater for sequences originating in the proximal colon. The solid purple line (bottom) is proportional to the propagating sequence frequency shown at the bottom and indicates that the density of component pressure waves is highest in the mid-colon and lowest at the ends of the colon.

important for the transport of content over great lengths of the colon. The mechanisms underlying regional linkage are yet to be determined.

REGULATION OF COLONIC FILLING AND TRANSIT ROLE OF THE ILEOCECAL JUNCTION

In humans, the ileocecal junction regulates colonic filling and prevents coloileal reflux, thereby preventing contamination of the small bowel with colonic bacteria.9 In the fasting state, cecal filling is slow and erratic, and chyme is retained in the distal ileum for prolonged periods.10 The close physical link between the terminal ileum and the cecum by the ileocecal ligaments behaves functionally as a valve and is responsible in part for continence of the ileocecal junction. A specialized band of muscle forms a low-pressure tonic sphincter11 and prominent 6 cyclesper-minute (cpm) phasic contractions contribute to the regulatory function of the ileocecal junction. Phasic and tonic activity are inhibited concurrently with episodic terminal ileal flow or distention of the ileum, and the tone of the ileocecal junction increases in response to cecal distention.11 Phase III of the interdigestive motor cycle (IDMC) (or migrating myoelectric-motor complex [MMC]), a motor pattern that occurs every 90 to 120 minutes in the upper intestine during fasting (see Chapter 97), does not contribute to ileocecal transit, because it rarely reaches the terminal ileum in the human. Most ileal chyme, driven by ileal propagating contractions in synchrony with inhibition of phasic contractions of the ileocecal junction, enters the cecum in a pulsatile fashion within 90 minutes of a meal. Prolonged studies, over several hours, correlating ileocecal movement of isotope with intraluminal pressures show that 72% of episodes of ileocecal transport result from monophasic, ileal propagating pressure waves.9 Furthermore, 93% of cecal propagating pressure waves were temporally associated with episodes of cecal filling, a finding that suggests episodic cecal filling is one of the triggers for proximal colonic propagating contractions (Fig. 98-7).9

THE COLON AS A STORAGE ORGAN

The region of preferential storage of colonic content is not entirely settled. In 1902, Cannon proposed on the basis of radiologic observations that the proximal colon is the site

mm Hg >150 Cecum

100

Splenic flexure

50

Rectum

0

4

2

6

0

Time (hr)

Figure 98-6.  Six-hour spatiotemporal map of colonic propagating sequences in a female healthy control subject. In this map each individual ridge represents an antegrade propagating sequence. The start of each ridge indicates the site of origin and the time the propagating sequence occurred. The length of the ridge indicates the extent of propagation. The shading within the ridge indicates the amplitude of the component pressure waves. The hatched white arrows link the site of origin of sequential propagating sequences. Although no single propagating sequence spans the entire colon, a linked series of propagating sequences can do so.

Chapter 98  Colonic Motor and Sensory Function and Dysfunction 1

7.5 cm

2

Terminal 3 ileum 4 1

2

5 Cecum 6

3

7 8 Hepatic flexure 20 mm Hg

T = 0 sec 20 sec

T = 10 sec

8 7 3 2 1 6 5 4

T = 20 sec T = 40 sec

Ascending colon

Terminal ileum

Recording sites

Figure 98-7.  Propagating pressure wave sequences identified in the terminal ileum and proximal colon during prolonged combined scintigraphic and manometric recordings. The bottom left corner of the far right box shows a scintigraphic image of technetium sulfur colloid in the terminal ileum and ascending colon of a healthy control subject. The pink circles indicate the location of the recording side holes, each spaced at 7.5 cm. The green hatched lines indicate the regions from which the luminal flow was recorded. Four scintigraphic images have been selected to indicate flow across the ileocolonic junction (solid bars 1 and 2) and mid-ascending colon (solid bar 3). The black arrows correspond to the time (horizontal axis) of acquisition of each 10-second scintigraphic frame. Small blue arrowheads on the scintiscans indicate the location of the manometric side hole from which the corresponding pressure tracing was recorded. Corresponding with the scintigraphic frame at T = 0, a cecal pressure wave is recorded. This cecal pressure wave initiates an ascending colonic propagating sequence that was temporally associated with coloileal reflux (solid bar 1) and flow across the mid-ascending colon (solid bar 2). During the coloileal reflux, an ileal propagating sequence is initiated (hatched black arrow), and this ileal propagating sequence is temporally associated with antegrade flow across the ileocolonic junction (solid bar 3). The red circle on the scintiscan images T = 0 to T = 40 follows the direction of the retrograde flow from cecum to ileum (T = 0 and T =10 sec) and then antegrade flow from the ileum to cecum (T = 20 sec and T = 40 sec).

of storage and mixing, whereas the distal colon acts as a conduit for expulsion. Subsequent studies, however, found no difference in the dwell time for radiopaque markers in the proximal, middle, and distal colon: roughly 11 hours in each. Composition of the diet influences regional transit and probably accounts for some of the discrepancies among studies. With a liquid diet, the ascending colon empties rapidly, within one to two hours, whereas the transverse colon retains isotope for 20 to 40 hours.12 Solid diet retards transit through the cecum and ascending colon. With a mixed diet, particulate matter and liquids are stored in both the ascending and transverse colon.13

RELATIONSHIPS BETWEEN COLONIC MOTOR PATTERNS AND FLOW

Emptying of the proximal colon occurs more rapidly when wall tone is increased (e.g., by intraluminal fatty acids) than when the tone is low; the volume and consistency of the contents also affect the rate of emptying. Isotonic fluid infused into the proximal colon stimulates proximal colonic emptying, suggesting that distention, per se, can activate propulsive motor patterns. Irritant laxatives (which act by stimulating mucosal receptors) in the proximal colon, however, trigger propagating contractions much more reliably than distention alone.14 Hence, proximal colonic emptying is influenced by a combination of increased wall tone and the initiation of propagating contractions, probably under the influence of both chemical and mechanical factors.

Mass movements, first detected radiologically, are infrequent movements of stool over long distances. More often, movement of colonic content occurs in a stepwise manner over short distances and in both antegrade and retrograde directions.15 Studies combining manometry with radiography in animals and with high frame rate scinti­graphy in humans have shown that 93% of all propagating sequences in the proximal colon, regardless of amplitude or polarity, are temporally associated with discrete movements of isotope-labeled colonic contents within the unprepared colon (Fig. 98-8).15 The strength of this pressure-flow relationship is region dependent, being stronger in the transverse colon than in the cecum and ascending colon. Most episodic antegrade movements of colonic content, which are not associated with propagating sequences, can be attributed to repetitive, nonpropagating pressure waves (over short distances in either direction), but there remain a number of movements of content that currently cannot be attributed to identifiable changes in intraluminal pressure. This observation might reflect the occurrence of contractions at points remote from the recording sites. Alternatively, propulsion is sometimes caused by motor events that do not significantly affect intraluminal pressure, such as longitudinal muscle shortening, non–lumen-occluding circular muscle contractions, or alterations in regional wall tone. Retrograde movements occur frequently. About half of retrograde contractions follow immediately after an antegrade movement, indicating frequent reflux of content back into the region from which it had just moved. This subtle to-and-fro motion is likely to help maintain maximal absorp-

1667

1668

Section X  Small and Large Intestine Cecum 1

Ascending colon 2

T = 0 sec Hepatic flexure 3

T = 30 sec Transverse colon 4

T = 45 sec Splenic flexure 5

T = 60 sec Proximal descending colon 6 50 mm Hg 15 sec T = 75 sec Figure 98-8.  Intracolonic pressure measurements and corresponding scintiscans that show a clear correlation between a propagating pressure wave sequence and discrete movement of colonic contents from the cecum to the sigmoid colon. This particular movement of contents was not associated with defecation or sensation. The oblique lines ending in vertical arrows (scintiscan images) correspond to the time of acquisition of each 15-second scintigraphic frame. Small arrowheads on the scintiscans indicate the location of the manometric side hole from which the corresponding pressure tracing was recorded. In the proximal colon and mid-colon (channels 2, 3, and 4 from the top), a close temporal relationship exists between movement of the isotope and the onset of the propagating pressure wave upstroke. When the pressure wave reaches the splenic flexure, however, the proximal descending colon is seen to expand to accommodate the isotope, consistent with loss of lumen occlusion at this region. The pressure waves in channels 5 and 6 do not appear to correspond to lumen-occluding contractions. Note also that propagating pressure-wave amplitudes in channels 3 and 4 are only 30 and 39 mm Hg, respectively, yet the motor pattern is clearly propulsive. (From Cook IJ, Furukawa Y, Panagopoulos V, et al. Relationships between spatial patterns of colonic pressure and individual movements of content. Am J Physiol 2000;278:G329.)

tion, retard colonic transit, and, therefore, reduce stool frequency. In summary, outside of the immediate predefecatory phase (see later), the distal colon displays a combination of fewer propagating sequences, shorter extent of propagation, higher conduction velocity, and lower probability of content propulsion than is observed in the proximal colon. In addition, proportionally more nonpropagating (segmenting) pressure waves occur in the distal colon than proximally. Considered together, these features would be expected to result in retardation of flow into the distal sigmoid and rectum, thus minimizing challenges to continence while maximizing the mixing of content more proximally.

DEFECATION Variations in propagating motor activity along the colon, as just described, would limit or might even prevent colonic contents from ever reaching the rectum and being expelled.

Clearly, additional mechanisms must occur from time to time that lead to defecation. Traditionally, defecation was conceptualized as an exclusively anorectal function; however, evidence for the integration of colonic motor activity with defecation has come from several sources. Radio-opaque markers and scintigraphic recordings confirm that the greater proportion of the entire colonic contents is evacuated in some cases. Furthermore, pancolonic manometric studies have demonstrated that the preparatory phase of defecation not only involves the greater part of the colon but also commences up to one hour before stool expulsion.16 In this predefecatory phase, a characteristic progressive increase occurs in the frequency of propagating pressure wave sequences. These sequences start first in the proximal colon, with each successive sequence originating slightly more distal to the preceding one; these priming sequences do not evoke conscious sensation. By contrast, in the 15 minutes leading up to defecation, a dramatic increase occurs in the frequency of these propagating sequences, which leads to a strong defecatory urge. In the last 15 minutes of the predefecatory

Chapter 98  Colonic Motor and Sensory Function and Dysfunction A

Stool expulsion

Terminal ileum

Cecum

Splenic flexure

Sigmoid colon

1

2

3

4 Attempted stool expulsion

B Cecum

Splenic flexure

Rectum

1

2 80 mm Hg

1 min Figure 98-9.  Intracolonic pressures leading to spontaneous defecation by the healthy human colon. Recordings were made with a perfused Silastic catheter passed transnasally to produce 15 recording sites at 7.5-cm intervals. A, Stool expulsion is preceded by four propagating sequences (1-4), the last of which actually was associated with stool expulsion. Each propagating sequence originates from a site more proximal than the preceding sequence. Note also the increase in amplitude and slowing of propagation velocity with successive sequences leading to stool expulsion. B, Two propagating sequences (1-2) precede defecation; however, the attempted stool expulsion is associated with straining only. (From Bampton PA, Dinning PG, Kennedy ML, et al. Spatial and temporal organization of pressure patterns throughout the unprepared colon during spontaneous defecation. Am J Gastroenterol 2000;98:1027.)

phase, propagating pressure waves begin to originate in the distal colon; however, in this late phase, each successive propagating sequence originates from a site proximal to the preceding one. Each sequence also tends to run for a slightly longer distance and has a higher amplitude compared with the preceding propagating sequence (Fig. 98-9). These final sequences provide potent forces to fill and distend the rectum, activating specialized low-threshold sacral spinal afferent mechanoreceptors. These mechanoreceptors then give rise to the defecatory urge, prompting the expulsive phase in which the anorectum comes into play.

RECTAL FILLING, CAPACITANCE, AND ACCOMMODATION AND MOTILITY OF THE ANAL SPHINCTERS

When stool or gas enters the rectum, the rectal wall is stretched, thereby simultaneously activating an enteric descending inhibitory reflex that causes transient relaxation of the internal anal sphincter and an extrinsic reflex pathway that leads to a brief contraction of the external anal sphinc-

ter. The anorectal inhibitory reflex can be demonstrated and tested by balloon distention of the rectum, and its presence reflects the integrity of enteric neural pathways. For example, the rectoanal inhibitory reflex is absent in Hirschsprung’s disease, which is characterized by loss of enteric ganglia in the rectal myenteric plexus. In health, this reflex permits entry of a small amount of content into the upper anal canal, and continence is maintained by the reflexive contraction of the external anal sphincter. This sampling of content by sensory receptors in the proximal anal canal permits the distinction between solid or liquid stool and gas. Sampling reflexes of this kind occur many times each day in response to low-volume rectal distentions, are not registered consciously, and do not cause an urge to defecate. A large-volume rectal distention causes an internal sphincter relaxation of longer duration, which is registered consciously and which necessitates extra voluntary contraction of the external anal sphincter to maintain continence while the person decides how best to deal with the intraluminal content (stool or gas). Suppression of the defecation urge at this time, together with receptive accommodation of

1669

1670

Section X  Small and Large Intestine

A

B

C

D

E

F

Figure 98-10.  Some of the mechanical processes that facilitate stool expulsion, as illustrated by sequential films of a simulated defecation of thickened barium during defecation proctography. A, The rectum at rest with a normal resting angle of approximately 90 degrees; the anal canal is closed. B, On straining, as the anterior rectal wall begins to flatten, the proximal anal canal begins to funnel as barium contrast is forced into it. C, As more pressure is exerted, the anterior rectal wall flattens further, contrast fills the anal canal, and evacuation begins. At this time, the puborectalis muscle and external anal sphincter are relaxing, resulting in the onset of descent of the rectoanal junction. At the same time, the levator ani muscles are activated and help control the descent of the rectoanal junction (note the posterior indentation resulting from contraction of the pubococcygeus muscle). D, The puborectalis is fully relaxed; this, in combination with vigorous straining, has resulted in nearly complete descent of the rectoanal junction. Note the position of the rectoanal junction, which in this frame is well below the horizontal pale artifact (due to the water-filled toilet seat), compared with that in the previous frame, in which the junction is level with this artifact. This descent has now opened up the anorectal angle, thereby further reducing the resistance to outflow through the anal canal. E, Rectal emptying continues, and anterior rectal compression is more obvious. F, After evacuation, the anorectal junction has ascended to its original position, and the anorectal angle has returned to its more acute resting angle. (Courtesy of Prof. D. Z. Lubowski.)

the rectum (see later), results in temporary storage of stool or gas in the rectum or retrograde transport of the stool or gas back to the sigmoid colon. Although the rectum is usually empty, it has the capacity to temporarily store feces until convenient evacuation can be arranged. Moreprolonged rectal storage is made possible by the ability of the rectum to accommodate an increasing volume without a corresponding increase in intrarectal pressure, in a manner similar to gastric fundic relaxation.17 This adaptive increase in rectal compliance, mediated by inhibitory nerves, is important for maintaining continence by permitting prolonged fecal storage without a constant urge to defecate. Such rectal distention also has negative feedback effects on the proximal bowel and inhibits gastric emptying, slows small bowel transit, reduces the frequency of proximal colonic propagating pressure waves, and delays colonic transit.18 Typically, rectal tone is increased following a meal. A pathologic reduction of rectal compliance, such as after pelvic radiotherapy, causes rectal urgency. Conversely, excessive compliance, as in megarectum, attenuates the urge to defecate.

ANORECTAL MOTILITY DURING DEFECATION

If the processes just described give rise to the urge to defecate and the social circumstances are appropriate, the full defecation process is activated. This process involves a combination of pelvic reflexes coordinated in the medulla and pons. Rectal distention by stool stimulates reflexinduced complete relaxation of the internal anal sphincter,

and the stool moves into the upper anal canal, heightening the sense of urge. Postural changes and straining facilitate this process in several ways: Sitting or squatting causes descent of the anorectal junction, and straining produces further rectal descent. Both activities serve to increase the anorectal angle, thereby reducing resistance to outflow. At this point, if the person wishes to proceed to expel stool, the external anal sphincter is relaxed voluntarily. At the same time, the puborectalis muscle is relaxed (further increasing the anorectal angle); the levator ani muscles contract; the perineum descends further; and stool is funneled into the anal canal and expelled by increasing strain-induced, intrarectal pressure (Fig. 98-10). Once the expulsion phase has commenced, evacuation of stool can proceed in some cases without further straining, as a consequence of colonic contractions propagating toward the anus (see Fig. 98-9).16 Expulsion of stool is possible in response to strain alone without rectosigmoid contractions, although a contribution from increased rectal wall tone cannot be excluded.

MODULATORS OF COLONIC MOTILITY PHYSIOLOGIC

Twenty-four hour recordings of myoelectric activity or intraluminal pressure show that colonic phasic and tonic activity predictably are increased one to two hours after a

Chapter 98  Colonic Motor and Sensory Function and Dysfunction meal (the gastrocolonic response) and are markedly suppressed at night.19 The entire colon responds to the meal, with an increase in colonic wall tone, migratory long spikebursts, and propagating and segmenting contractile patterns. A minimum caloric load of approximately 300 kcal is required to generate the colonic response to a meal, and a meal of only 200 kcal increases rectal muscle tone.20 The meal response also is highly dependent on the fat content of the caloric load. For example, 600 kcal of fat induces the response, whereas an equicaloric load of protein or carbohydrate does not. The mechanism of the colonic meal-response remains unclear, although it is known that neither the stomach nor the spinal cord needs to be intact to display the response. Non-nutrient gastric distention, by balloon or water, also can stimulate rectosigmoid motility, yielding a similar response to that following intraduodenal lipid infusion. Both of these responses are markedly attenuated by prior intravenous administration of the 5-hydroxytryptamine-3 (5-HT3) receptor antagonist granisetron, which suggests that 5-HT3 receptors on vagal afferents may be involved in the gastrocolic response.21 Cholecystokinin (CCK), which is released by fats and fatty acids in the duodenum, can replicate the gastrocolic response, but only at doses exceeding those occurring postprandially. The CCK-A antagonist loxiglumide blocks the effects of CCK on the colon but does not abolish the gastrocolic response, thus making CCK an unlikely mediator of the response. Colonic myoelectric and pressure activities are profoundly suppressed at night.19 During stable sleep, colonic motility virtually ceases (except for the antipropulsive rectal motor complexes, which increase), thereby reducing the challenges to continence at a time when anal sphincter

Sleep stage mm Hg 80 Ascending 0 80 Hepatic flexure 0 80 Transverse 0 Splenic 80 flexure 0 80 Descending 0 80 Proximal sigmoid 0 80 Midsigmoid 0 80 Distal sigmoid 0

Stage 2

Stage 1

tone and awareness of colorectal sensations are minimal. If the subject shifts to a lighter level of sleep, even without actually awakening, an immediate increase occurs in propagating and nonpropagating pressure waves (Fig. 98-11). Forced awakening at night and spontaneous early-morning awakening both stimulate an immediate increase in colonic propagating pressure waves. This phenomenon clearly is linked with the readily identifiable habit of defecation soon after awakening in the morning and demonstrates the potential for profound modulation of colonic motor activity by the central nervous system. Stress and emotional factors long have been believed to influence colonic motility, but experimental evidence for this is conflicting, possibly because of a reliance on measurements from the distal colon, which might not be representative. In light of the profound waking-response, it is likely, but unproved, that stress does induce propagating pressure waves. Due to technical difficulties associated with trying to record physical activity and colonic motility simultaneously, data on the colonic response to physical activity are sparse; however, physical exercise, perhaps through increased sympathetic tone, decreases colonic motility.22 The colonic response to stress and exercise highlight the importance of the autonomic nervous system in modulating colonic function. Similarly, autonomic dysfunction, resulting from pelvic surgery, childbirth, or neural degradation, has been implicated in several colonic disorders including slow-transit constipation and irritable bowel syndrome (IBS).23

PHARMACOLOGIC

Laxatives exert their diarrheal actions by increasing mucosal secretion or by stimulating colonic propulsive activity.

Awake

Arousal Arousal

Stage 1

Stage 2

Arousal

1 min Figure 98-11.  Relationship of propagating pressure wave sequences to nocturnal arousals. An arousal represents a lightening of the level of sleep and need not necessarily culminate in awakening. This example demonstrates an arousal-induced event propagating from the proximal to the distal sigmoid colon and followed by another arousal-induced event propagating from the ascending colon to the sigmoid. Only the second arousal culminated in a brief period of wakefulness. Repetitive propagating sequences of this type also are seen on early-morning awakening. (From Furukawa Y, Cook IJ, Panagopoulos V, et al: Relationship between sleep patterns and human colonic motor patterns. Gastroenterology 1994;107:1372.)

1671

1672

Section X  Small and Large Intestine For example, the irritant laxative bisacodyl and the bile acid chenodeoxycholic acid both stimulate high-amplitude colonic propagating pressure wave sequences, thereby leading to mass movements. Bisacodyl exerts its motor effect through mucosal afferent nerve fibers, because the response can be blocked by topical mucosal application of lidocaine. In addition to the local response, these agents, when administered rectally, can stimulate motor activity in the proximal colon, thereby indicating the existence of long reflex pathways between the rectum and proximal colon. Colchicine, a natural alkaloid, is well known to cause diarrhea. Colchicine increases the frequency of spontaneous bowel movements and accelerates colonic transit in patients with chronic constipation24; the mode of action is not yet clear but it has been shown to increase prostaglandin synthesis and to promote intestinal secretion, the latter mediated through cyclic AMP. In the rat, small intestinal colchicine stimulates myoelectric activity. Lubiprostone, a type 2 chloride channel (ClC2) activator, is a member of a new class of compounds known as prostones. Activation of ClC2 increases intestinal chloride secretion resulting in increased intraluminal fluid accumulation, which accelerates intestinal transit, softens stools, and increases spontaneous stool frequency in patients with constipation.25 Serotonin (5-HT) is an important mediator of bowel physiology, and both 5-HT3 and 5-HT4 receptors play a role in colonic peristalsis and transit. For example, the 5-HT3 receptor antagonists granisetron and ondansetron blunt the gastrocolic response and delay colonic transit, respectively.21 Alosetron, another antagonist of the 5-HT3 receptor, exerts a significant constipating affect by slowing colonic transit.26 In contrast, 5-HT4 agonists (e.g., tegaserod, pru­ calopride, renzapride), act on presynaptic receptors and facilitate release of acetylcholine and CGRP (calcitonin gene-related peptide), thereby inducing colonic propagating contractions and accelerating colonic transit. Although this class of drug shows promise for the treatment of constipation,27 tegaserod, a 5-HT4 agonist, was withdrawn from the market because of concern about associated adverse cardiovascular events (see Chapter 118). Other highly selective 5-HT4 agonists, such as prucalopride, might be attractive options because they do not interact with 5-HT3 or 5-HT1B receptors, and prucalopride does improve stool frequency and symptoms in severe constipation.28 Further trials with these agents are awaited. Opiates are well known to have an antidiarrheal effect, but their mechanism of action is less clear. In the human colon, morphine increases phasic segmenting activity, reduces colonic tone, and attenuates the bowel’s response to a meal.29 Opiates are known to inhibit presynaptic and postsynaptic enteric neural circuitry. The reduction in neurally dependent propagating contractions and the enhancement of myogenic mixing movements and fluid absorption contribute to the constipating effect of the drug. Specific constipation syndromes, such as opiate-induced constipation or postsurgical ileus, might respond to opiate antagonists such as methylnaltrexone and alvimopan (see Chapter 120).30 Nitric oxide is a potent endogenous inhibitor of colonic propulsive activity and the human colon appears to be under a state of tonic nitrergic inhibition. For example, infusion of the nitric oxide synthase inhibitor, l-NMMA (NGmonomethyl l-arginine), is a potent stimulator of colonic propagating contractions.31 Alternatively, segmental lengthening of the colon induced by the entry of content triggers nitric oxide release from descending pathways, which in turn inhibits colonic propulsive activity.32

NONPHARMACOLOGIC

Probiotics are living organisms that, when ingested in adequate amounts, are claimed to exert a health benefit to the host. Relatively few rigorously designed studies have been conducted with probiotics but some strains have been shown to have a beneficial effect in IBS (see Chapter 118), ulcerative colitis (see Chapter 112), and diarrhea.33 In the colon, probiotics are likely to modulate the inflammatory response through activation of signals with the epithelium and immune system.33 Probiotics may well inflence colonic motility, but this has not been systematically evaluated. Sacral nerve stimulation modulates the extrinsic nerves innervating the pelvic floor and colon. Electrical stimulation of the S3 sacral root induces a modest increase in external anal sphincter tone and has been used successfully in the management of fecal incontinence. Stimulation of the S3 root also induces propulsive activity throughout the entire colon and has been shown to increase stool frequency in patients with slow-transit constipation.34 Randomized trials of this promising technique for treating slow-transit constipation are in progress; the precise mode of this action remains unknown. The substantial latency between stimulus and pelvic floor or colonic contractile responses is longer than would be expected via a polysynaptic efferent pathway, which suggests possible involvement of extrinsic neural pathways.35 Magnetic stimulation of the sacral nerve S3 also shows promise in modulating colonic and anorectal function.36 Because this approach is less invasive than electrical stimulation of sacral nerves, it may be a reasonable treatment option in children with colonic or anorectal dysfunction. Acupuncture has been shown to have significant effects upon upper gastrointestinal tract disorders such as nausea and vomiting. Only two studies have evaluated its potential in constipation, one in children and one in adults.37 Acupuncture improved stool frequency in children, but these results weren’t replicated in adults; this warrants further study. Acupuncture is known to activate neural, opioid, humoral, and serotoninergic pathways and potentially has a clinical role in treating disorders such as constipation and IBS.37 Biofeedback has been shown to improve stool frequency and rectal evacuation in patients with pelvic floor dyssynergia, and the technique has been shown to accelerate colonic transit in this subset of patients with constipation (see Chapter 18).38 The mode of action of biofeedback is not fully understood, but evidence suggests that extrinsic autonomic efferent pathways mediate the response.39

DISORDERS OF COLONIC MOTILITY Disorders attributable to disturbed colonic motor function are discussed elsewhere in this book (Chapter 120). It is useful, however, to consider how disturbances in the mechanisms of colonic motility described in this chapter might relate to symptoms or pathophysiologic phenomena.

CONSTIPATION

Intuitively, one would expect that constipation and diarrhea should be manifestations of hypomotility and hypermotility, respectively. Sometimes this is true, but in the distal colon, at least, the converse may be true. A paradoxical increase in nonpropagating (segmenting) contractions and myoelectrical short spike-bursts has been reported in the rectosigmoid region in constipated patients. Conversely, patients with diarrhea have hypomotility in this region.

Chapter 98  Colonic Motor and Sensory Function and Dysfunction A

Morning awakening

Healthy female control subject

Lunch (1000 kCal)

mm Hg >150

Cecum Splenic flexure Rectum

B

100 14:00

22:00 Nocturnal 06:00 suppression Defecation

Patient with slow-transit constipation

Morning awakening

Lunch (1000 kCal)

14:00

50

0 Retrograde Antegrade

Cecum Splenic flexure Rectum

14:00

22:00

06:00

14:00

Time of day (24 hr) Figure 98-12.  Twenty-four-hour spatiotemporal maps of colonic propagating sequences in a female healthy control subject (A) and a female patient with slow-transit constipation (B). Within these maps, each individual ridge represents a propagating sequence. Antegrade propagating sequences (green) originate at the orad end of the ridge, and retrograde propagated sequences (red) originate at the anal end of the retrograde ridge. The start of each antegrade and retrograde ridge indicates the site of origin and the time of day the propagating sequence occurred. The length of the ridge indicates the extent of propagation. The shading within the ridge indicates the amplitude of the component pressure waves. In health (top), the maps indicate several physiologic characteristics of colonic propagating sequences. These include increased frequency and amplitude before defecation, immediately after morning awakening, and in response to a high-calorie meal. The map also indicates nocturnal suppression of these motor patterns. In contrast, the patient with severe constipation (B) demonstrates a clear ability to generate propagating sequences, but there are notable differences in the characteristics of these motor events compared with those of healthy control subjects. For example, the patient demonstrates an increased frequency of low-amplitude antegrade and retrograde propagating sequences in the proximal and distal colon and few propagating sequences spanning the mid-colon. This patient also demonstrates a lack of the normal colonic nocturnal suppression of propagating sequences and a lack of the normal colonic response to a highcalorie meal. (Courtesy of Dr. P.G. Dinning, Kogarah, New South Wales.)

It is likely that segmenting activity retards forward flow, whereas suppression of such activity permits unrestricted access of stool to the rectum, where a defecatory urge is initiated. Thus, constipation can result from either infrequent or ineffective propagating pressure waves, or from an increase in segmenting distal colonic pressure waves, or perhaps both. In severe slow-transit constipation, prolonged manometric studies have confirmed a reduction in the overall number of high-amplitude propagating pressure waves19; however, the overall number of propagating pressure waves of any magnitude is often normal or increased. Studies examining the spatiotemporal patterning of propagating sequences have revealed colonic regions in which activity is diminished or absent, particularly within the vicinity of the splenic flexure (Fig. 98-12).40 Furthermore, there appears to be a loss of the normal linkage in patients who have sequential progressive systemic sclerosis and constipation.8 The underlying pathogenesis of severe slow-transit constipation is unclear, but changes in enteric excitatory motor inner­ vation of the smooth muscle in patients with severe slow-transit constipation are likely to contribute to this disorder.41 Constipation is fully discussed in Chapter 18.

entire colon in patients with diarrhea would help explain these observations, but are lacking. A relative lack of distal colonic segmenting activity, perhaps in combination with increased proximal colonic propagating pressure waves, might explain this preferential acceleration of proximal colonic transit, but proof of this hypothesis is awaited. Diarrhea is fully discussed in Chapter 15.

DIARRHEA

COLONIC MOTILITY DISTURBANCES SECONDARY TO NONMOTOR INTESTINAL DISORDERS

Detailed scintigraphic studies in patients with diarrhea have shown the dominant feature to be early and rapid transit through the ascending and transverse colon.42 Normally, propagating sequences are more frequent in these proximal regions than elsewhere. Manometric data from the

IRRITABLE BOWEL SYNDROME

Although colonic transit generally is slower in constipationpredominant IBS and faster in diarrhea-predominant IBS, no colonic motor pattern is specific for IBS.43 Exaggerated responses to stimuli such as meals, CCK, and mechanical stimuli have been reported, but a consistent disturbance has not emerged, probably because of the heterogeneity of the disease and the methodologies used for characterization. In addition, remarkably little study of the proximal colon in IBS has been conducted to date. At present, compelling evidence regarding the pathophysiology of IBS suggests a major contribution by afferent hypersensitivity, in addition to a variable alteration in colonic motor function. IBS is fully discussed in Chapter 118.

Altered motility secondary to underlying inflammation or a hormonal disturbance can contribute to the colonic symptoms of a nonmotor disease. The diarrhea of idiopathic inflammatory bowel disease, for example, results from a

1673

1674

Section X  Small and Large Intestine combination of enhanced secretion, reduced absorption, and altered colonic motor function. In ulcerative colitis, rectosigmoid-segmenting, nonpropagating pressure waves are diminished, whereas postprandial propagating pressure waves are increased.13 Rectal compliance also is reduced, and together, these effects can exacerbate diarrhea, as suggested by studies demonstrating rapid rectosigmoid transit in ulcerative colitis.13 The motility of the healthy colon also can be perturbed by ileal diseases. For example, exposure of the healthy proximal colon to supranormal concentrations of bile salts, such as from terminal ileal disease or resection, not only stimulates net colonic secretion but also initiates high-amplitude propagating pressure waves, thereby accelerating colonic transit.13

KEY REFERENCES

Bampton PA, Dinning PG, Kennedy ML, et al. Spatial and temporal organization of pressure patterns throughout the unprepared colon during spontaneous defecation. Am J Gastroenterol 2000; 95:1027-35. (Ref 16.) Dickson EJ, Spencer NJ, Hennig GW, et al. An enteric occult reflex underlies accommodation and slow transit in the distal large bowel. Gastroenterology 2007; 132:1912-24. (Ref 32.) Dinning PG, Szczesniak MM, Cook IJ. Removal of tonic nitrergic inhi­ bition is a potent stimulus for human proximal colonic propagating sequences. Neurogastroenterol Mot 2006; 18:37-44. (Ref 31.) Dinning PG, Szczesniak MM, Cook IJ. Determinants of postprandial flow across the human ileocecal junction: A combined manometric and scintigraphic study. Neurogastroenterol Mot 2008; 10:1119-26. (Ref 9.)

Di Stefano M, Miceli E, Missanelli A, et al. Meal induced rectosigmoid tone modification: A low caloric meal accurately separates functional and organic gastrointestinal disease patients. Gut 2006; 55:1409-14. (Ref 20.) Farrugia G. Ionic conductances in gastrointestinal smooth muscles and interstitial cells of Cajal. Ann Rev Physiol 1999; 61:45-84. (Ref 2.) Hagger R, Kumar D, Benson M, Grunday A. Periodic colonic motor activity identified by 24-h pancolonic ambulatory manometry in humans. Neurogastroenterol Mot 2002; 14:271-8. (Ref 7.) Hardcastle JD, Mann CV. Physical factors in the stimulation of colonic peristalsis. Gut 1970; 11:41-6. (Ref 14.) Kamath PS, Phillips SF, O’Connor MK, et al. Colonic capacitance and transit in man: modulation by luminal contents and drugs. Gut 1990; 31:443-9. (Ref 29.) O’Brien MD, Phillips SF. Colonic motility in health and disease. Gastroenterol Clin North Am 1996; 25:147-62. (Ref 13.) Porter AJ, Wattchow DA, Brookes SJ, Costa M. The neurochemical coding and projections of circular muscle motor neurons in the human colon. Gastroenterology 1997; 113:1916-23. (Ref 4.) Rae MG, Fleming N, McGregor DB, et al. Control of motility patterns in the human colonic circular muscle layer by pacemaker activity. J Physiol 1998; 510:309-20. (Ref 1.) Rao SS, Sadeghi P, Beaty J, Kavlock R. Ambulatory 24-hour colonic manometry in slow-transit constipation. Am J Gastroenterol 2004; 99:2405-16. (Ref 19.) Sanders KM, Koh SD, Ward SM. Organisation and electrophysiology of intersitial Cajal and smooth muscle cells in the gastrointestinal tract. In: Johnson LR, editor. Physiology of the Gastrointestinal Tract. Amsterdam: Elsevier; 2006. p 533-76. (Ref 3.) Scott M. Manometric techniques for the evaluation of colonic motor activity: current status. Neurogastroenterol Mot 2003; 15:483-513. (Ref 6.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

99 Intestinal Electrolyte Absorption and Secretion Jayashree Venkatasubramanian, Mrinalini C. Rao, and Joseph H. Sellin

CHAPTER OUTLINE Intestinal Architecture and Transport  1675 Basic Epithelial Cell Model  1677 Segmental Heterogeneity of Transport  1677 Movement across the Intestinal Epithelium  1678 Tight and Leaky Epithelia  1678 Transepithelial Transport  1679 Transcellular Transport  1679 Water Movement  1679 Channels, Carriers, and Pumps  1679 Ion Transporters  1681 Apical Sodium Channel  1681 Nutrient-Coupled Sodium Transport  1681 Sodium-Hydrogen Exchangers  1681 Electroneutral Sodium Chloride Absorption  1682 Chloride (Anion) Absorption  1683 Chloride Secretion  1683 Chloride Channels  1683

The gastrointestinal (GI) tract processes 8 to 9 L of fluid daily that is derived from oral intake and endogenous exocrine secretions. Intestinal fluid absorption is a process that functions with 98% efficiency, allowing only 100 to 200 mL to be excreted each day. The intestine also extracts nutrients, vitamins, and minerals; excludes destructive antigens and microbes; and excretes waste (Fig. 99-1). This multitasking is achieved by the unique tissue, cellular, and molecular architecture of the small and large intestine in combination with a complex array of intricate regulatory mechanisms (Fig. 99-2). Regulation is accomplished by cross-talk between endocrine and paracrine hormones, neurotransmitters, immunomodulators, and luminal factors. Remarkably, this orchestration proceeds smoothly on a daily basis; however, when the balance is perturbed, as occurs with an enteric infection, diarrhea ensues. Over the past four decades, our understanding of intestinal ion transport processes has been revolutionized by the elucidation of the molecular basis of two devastating diseases, cholera and cystic fibrosis. Although the two diseases effect opposite ends of the physiologic spectrum—too much versus insufficient fluid secretion, respectively—examination of their underlying physiologic, regulatory, and genetic parameters have vastly advanced our knowledge. This increased insight of how the intestine transports fluid and electrolytes has had significant clinical impact, most notably in the development of oral rehydration therapy (ORT) for

Calcium-Activated Chloride Channels  1684 Potassium Transport  1684 Bicarbonate Transport  1685 Short-Chain Fatty Acid Transport  1685 Paracrine Immunoneuroendocrine System  1686 Extracellular Regulation  1687 Endocrine, Paracrine, Juxtacrine, and Autocrine Regulation  1687 Neural  1687 Immunologic  1688 Systemic Effects  1689 Osmotic Effects  1689 Specific Regulatory Factors  1690 Absorptive  1690 Secretory  1690 Intracellular Mediators  1692 Homocellular Regulation  1693

diarrheal diseases, one of the major health advances of the 20th century. In this chapter, we review the current understanding of the cellular and molecular underpinnings of the trafficking of ions and solutes in different regions of the small and large intestine and their regulation in health and disease states. The functional activities of intestinal transporters have long been recognized; however, only recently has it become apparent that there are a plethora of transport proteins that carry out these specific functions. This understanding is critical for appreciating normal intestinal function, the pathophysiology of intestinal absorptive abnormalities, and the development of therapeutic strategies for specific diseases.

INTESTINAL ARCHITECTURE AND TRANSPORT The structural and functional design of the intestine is optimally geared to its functions of absorbing nutrients and transporting fluids. In the small intestine, a 600-fold amplification of the absorptive surface is achieved by structural features, such as the circular folds of Kerckring (plicae circulares), villus-crypt architecture, and microvilli. Using a cylinder as the model, it has been estimated that the surface area of the small intestine is about 3300 cm2; the plicae

1675

Section X  Small and Large Intestine Duodenum

Jejunum

Ileum

Cecum

Proximal colon

Distal colon

Absorption 7000 mL Diet 1500 mL 1400 mL Salivary 1500 mL Gastric 2500 mL Biliary 500 mL Pancreatic 1500 mL Intestinal 1000 mL

100 mL

1500 mL Secretions 7000 mL

Tissue permeability

Leaky

Mod. leaky

Mod. tight

Tight

3 mV

6 mV

12 mV

20 mV

Na-nutrient Na+-H+

Na-Cl Na-nutrient Na-bile acid

Na+ SCFA

PD Absorptive mechanisms

Na-Cl SCFA

Na+

[Na+] 20–140 mM

–40

3–20 mV

10

[Na+] 140 mM 3Na+

2K+

Zona occludens (Tight junction)

Gap junction

0 mV

Desmosomes

Basolateral

Apical

la u JA dins M O cc lu di ns Zona adherens

Figure 99-1.  Overview of intestinal fluid balance. Approximately 8.5 L of fluid flow into the intestine daily. Salivary, gastric, biliary, pancreatic, and intestinal secretions make up most of this amount. The bulk of this fluid is absorbed in the small intestine, and approximately 1500 mL cross the ileocecal valve. The colon efficiently reabsorbs most of this fluid, with only 100 to 200 mL lost in stool. Permeability can be viewed as a surrogate to conductance, which is a reciprocal of resistance. Based on Ohm’s law, current = potential difference divided by resistance (I = PD/R). If I is a constant, PD increases as R increases. Permeability of the intestinal epithelium decreases down the length of the cephalocaudal axis, the distal colon having a relatively tight epithelium. Thus, the spontaneous PD demonstrates a corresponding rise with resistance along the cephalocaudal axis. Absorptive mechanisms in each segment of the intestine differ; chloride secretion is found throughout the intestine. SCFA, short-chain fatty acid.

C

1676

Spectrin/actin Paracellular Transcellular

Passive Active Passive

Myosin/ terminal web

ZO-1 ZO-2 ZO-3

Rab3b MUPP1 aPKC PAR-3

Figure 99-2.  Architecture of intestinal epithelia. Intestinal epithelial cells are structurally and functionally geared for vectorial transport: The cell membrane is divided into distinct apical and basolateral domains by the tight junctions with an asymmetrical distribution of transporters; the Na+ pump on the basolateral membrane is integral to maintaining an electrochemical profile; this profile permits downhill entry of sodium from either the apical or basolateral side; water and solutes can cross the epithelium either between the cell (paracellular) or through the cell (transcellular). Transcellular transport across the membrane can be passive or active. The paracellular pathway is characterized by a series of structures that are defined by specific molecular distributions. The tight junction, or zona occludens (ZO), is made up of a network of strands and grooves that consist of membrane proteins (e.g., occludins, claudins, and junctional adhesion molecules [JAMs]) that attach to a group of scaffolding proteins (zonula occludens proteins [ZO-1, ZO-2, ZO-3], multi-PDZ domain protein-1 [MUPP1]). These scaffolding proteins are then linked to the cytoskeleton, participate in vesicular transport (via monomeric guanosine triphosphatase [GTPase] of the Ras superfamily (Rab3b) and in the activation of signaling molecules that regulate junction assembly (partition-defective protein, PAR-3 and -6, and atypical protein kinase C, [aPKC]). Cadherins span the paracellular pathway across the zona adherens and are responsible for cell-to-cell attachment and maintenance of cell polarity. Cadherins bind to catenins, which are linked to the actin cytoskeleton through an additional family of molecules, including radixin, vinculin, and α-actinin. Molecules associated with the zona adherens, including rab, src, and yes, are involved in intracellular signaling through second messengers. Desmosomes are cadherin-like molecules that are linked to intermediate filaments. Gap junctions, made by an assembly of membrane spanning proteins called connexins, allows exchange of small molecules between neighboring cells. PD, potential difference.

Chapter 99  Intestinal Electrolyte Absorption and Secretion circulare, villi, and microvilli amplify the surface area by factors of 3, 10, and 20, respectively, ultimately giving a surface area of about 2,000,000 cm2. In the large intestine, the spatial separation of crypts and surface cells allows efficient reabsorption of fluid. The overall architecture of the intestinal musculature can influence bulk fluid flow and transit time via changes in motility patterns (see Chapters 97 and 98), but the work of fluid transport occurs in the epithelia. Most epithelia serve as semipermeable barriers: They act as the first line of defense between the mucosal (luminal) and serosal (blood-side) compartments and are capable of bulk transport of fluid from one compartment to the other. These epithelia, including those of the intestine, share common characteristics. One fundamental property of epithelia is cellular polarity, with molecularly distinct apical (luminal) and basolateral (serosal) membranes demarcated by intercellular tight junctions. The permeability of the tight junctions vary from being relatively leaky in the small intestine to fairly tight in the large intestine, and these differences determine an individual epithelium’s effectiveness as a barrier. A loss of tight junction integrity disrupts the barrier function and the vectorial transport capabilities of the tissue.

BASIC EPITHELIAL CELL MODEL All GI epithelial cells have two fundamental similarities: discrete apical and basolateral membranes, with distinct biochemical and biophysical properties, separated by tight junctions; and a basolateral Na+ pump (ouabain-inhibitable Na+,K+-ATPase [adenosine triphosphatase]) that establishes a specific intracellular electrochemical environment with a low intracellular Na+ concentration ([Na]i) and a negative intracellular voltage. This basic cell model is modified by insertion of transporters into either the apical or basolateral membrane or by the characteristics of tight junctions that determine the unique qualities of a specific epithelial segment. A complex interaction of protein-sorting signals, cytoskeletal elements, and intracellular trafficking processes determines whether a newly synthesized protein is targeted to either the apical or basolateral membrane. For example, proteins with a glycosyl phosphatidyl inositol (GPI) anchor (e.g., alkaline phosphatase or carcinoembryonic antigen) are often associated with lipid rafts, and the GPI anchor serves to direct them toward the apical membrane.1 Membrane proteins destined to be delivered to the basolateral membrane carry specific membrane-sorting signals (amino acid sequences) in their cytoplasmic tails. In contrast, other proteins can insert randomly into either apical or basolateral domain, but they may be retained in the basolateral pole by specific components such as ankyrin.2 Regulation of intracellular trafficking ensures delivery of the right protein to the right membrane and is critical for establishing epithelial polarization and vectorial transport. When tight junctions are disrupted in vitro, diffusion and intermingling of apical and basolateral proteins in the fluid phase of the membrane result in a loss of epithelial cell polarity. There is some evidence that the distribution of Na+ pumps is altered during postischemic injury.3 The most prominent feature of epithelial cell polarity is targeting of the Na+,K+-ATPase pump to the basolateral membrane, for which expression of the beta subunit of Na+,K+-ATPase is critical. The Na+ pump is electrogenic, extruding three Na+ ions in exchange for two K+ ions, and

thereby maintaining relatively low intracellular Na+ and high intracellular K+ concentrations compared with con­ centrations of these electrolytes in the extracellular environment (see Fig. 99-2). There also is greater membrane permeability for K+ over Na+, which favors diffusional exit of K+ from the cell over diffusional cellular entry of Na+. These features, in combination with the large number of intracellular proteins with fixed negative charges, lead to the characteristic negative intracellular potential difference compared with either the mucosal or serosal compartments.* Low [Na+] and electronegativity establish a favorable electrochemical gradient for passive Na+ entry into the cell. Functionally, the epithelial cell uses the energy of the favorable Na+ gradient to transport not only Na+ ions but also a variety of nutrients, vitamins, and electrolytes. These properties provide the basic mechanisms of ion and water transport that apply to all epithelia. In the intestine, differences in transport can be seen along its cephalocaudal length as well as along the surface-crypt axis within a particular segment of intestine. Tissue- and segment-specific nuances arise from structural-functional and regulatory differences of both intracellular and inter­ cellular proteins.4

SEGMENTAL HETEROGENEITY OF TRANSPORT All intestinal segments from the duodenum to the distal colon have mechanisms for absorbing and secreting water and electrolytes. The diverse physiologic functions along the length of the GI tract are supported by the varied array of transporters encountered in its different segments. For example, the glucose- and amino acid-coupled transporters in the jejunum are well suited for absorption of large volumes of nutrients and water. The cecum, proximal colon, and distal colon, however, exhibit distinctly different transporters, with electrogenic Na+ absorption in the distal colon accomplishing the necessary final fluid extraction in preparation of feces.5-8 Different transporter molecules have been identified in specific segments of the GI tract. What is not clear, however, is why an individual transporter is located only in a specific segment of the intestine. For example, the DRA (down-regulated in adenoma) protein is an anion exchanger, and although anion exchange function is recognized in different segments of the intestine, DRA is predominantly expressed in the colon9-10 (see the later discussion of bicarbonate transport). There also is segmental heterogeneity along the cryptvillus axis. Stem cells near the base of the crypt differentiate and migrate upward to form villus enterocytes in the small intestine or surface colonocytes in the large intestine while undergoing important changes in their transport and barrier properties (Fig. 99-3).11,12 As epithelial cells migrate away from the proliferative zone, the complexity of their tight junctions increases, the microvillus architecture of their apical membrane becomes more pronounced, and underlying cytoskeleton and signaling molecules undergo change; there also is increased expression of Na+ nutrient-coupled transporters, apical Na+-H+ exchangers, and brush border membrane hydrolases. In contrast, the levels of the Na+ pumps remain relatively constant and others, such as *There are several potential differences across the epithelium: across the apical membrane into the cell, from the cell interior across the basolateral membrane, across the epithelium, across the mucosa, and across the entire GI tract. By convention, the potential differences across the epithelium, the mucosa, and the entire GI tract generally are considered the same.

1677

1678

Section X  Small and Large Intestine Lumen

Cells at tip undergo apoptosis and slough into lumen

Villus/surface

Na-channel Na-nutrient NHE-3 PAT

CFTR

Na pump DRA NHE-I

Enterocytes

Endocrine cell

Villus region

Turnover 3-5 days

Mucin-secreting goblet cell

Proliferative zone

Crypt region

Cells mature as they migrate up to the villus region

Paneth Submucosa cells Figure 99-3.  Types of epithelial cells of the intestinal mucosa: enterocytes, endocrine cells, goblet cells, and Paneth cells. All of these cell types originate from the proliferative zone near the base of the intestinal crypt. With the exception of Paneth cells, these cells migrate up the villus axis, mature during this process, and eventually undergo apoptosis and slough after three to five days at the tip of the villus. Paneth cells remain at the base of the crypt and make defensins, which are important in host defense.

the signaling molecule adenylate cyclase and the cyclic adenosine monophosphate (cAMP)-associated Cl− channel CFTR (cystic fibrosis transmembrane conductance regulator), decrease in more mature villus cells. This spatial distribution of transporters (Fig. 99-4) is consistent with a model in which secretory function resides primarily in the crypts and absorption occurs in villus or surface cells. This dichotomy between absorptive surface cells and secretory crypt cells, however, is far from absolute; for example, in the colon, crypts absorb Na+ and fluid, and surface cells secrete Cl−.13,14 Thus, depending on their relative position along the crypt-villus axis, the crosstalk between transporters and their signaling molecules can vary and fine tune intestinal function. This segregation of absorptive and secretory functions might explain why, in diseases that selectively damage villi or surface epithelia—such as enteric infection, inflammatory bowel disease (IBD), and celiac disease—secretion predominates.

MOVEMENT ACROSS THE INTESTINAL EPITHELIUM Movement of ions and solutes across the epithelium is bidirectional and occurs via the transcellular and paracellular routes. Paracellular movement is largely passive, in response

Crypt Figure 99-4.  Spatial location of transport protein gradients. There is a significant spatial geometry of transport proteins along the crypt-villus (crypt-surface) axis. Some transport proteins are found at relatively constant concentrations along this axis, whereas some proteins exhibit a greater density in the base of the crypt and others are denser toward the villus or surface. CFTR, cystic fibrosis transmembrane conductance regulator; DRA, down-regulated in adenoma; NHE, sodium-hydrogen exchanger; PAT, putative anion transporter.

to a variety of gradients, including concentration, electrical, osmotic, and hydrostatic; transcellular movement of ions and solutes occurs by active and passive transport mechanisms. Net transport is termed absorptive if the mucosal-toserosal flux (Jms) is greater than the serosal-to-mucosal (Jsm) flux and, and it is termed secretory if Jsm > Jms. Changes in either or both can alter the direction of the net movement; for example, the ileum, which normally exhibits an absorptive flux, responds to cholera toxin with a decrease in Jms and an increase in Jsm for Cl−, resulting in massive fluid secretion. Characteristics of the tight junctions—for example, tight versus leaky—vary along the length of the intestine and dictate the contribution of paracellular fluxes to overall transport. The effectiveness of a transepithelial gradient may be modified by series of physical barriers, including an unstirred layer created by the glycocalyx above the apical membrane, the lipid composition of the apical and basolateral membrane, the tight junctions, the geometry of the basolateral space between cells, and the basement membrane. Generally, movement of an uncharged particle is dictated solely by concentration gradients. In contrast, the transport of an ion is governed by the electrical potential and concentration differences (the electroche­ mical gradient) across the transported surface. Solvent drag, a nonspecific entraining of solutes along with the movement of water across paracellular pathways, is an absorptive mechanism that may be especially important in the small intestine, for example, for Na+-coupled solute absorption.

TIGHT AND LEAKY EPITHELIA The paracellular space and junctional complexes between cells define the barrier function of epithelia. Epithelia with a low transepithelial voltage and low resistance are con­ sidered leaky, and those that exhibit a high transepithelial voltage and high resistance are considered tight. The tight junctions in villi have higher resistance than do those in crypts. Transepithelial resistance increases in a cephalocaudal direction (see Fig. 99-1).13

Chapter 99  Intestinal Electrolyte Absorption and Secretion Since the 1990s, the model of paracellular transport and tight junctions has rapidly evolved from a static rigid barrier to a dynamic complex structure that is finely regulated (see Fig. 99-2). Movement through the space is exclusively passive, but it is influenced by electrical conductivity, charge selectivity, and its ability to be regulated. Cell-to-cell communications along the paracellular pathway occur in several discrete structures: zona occludens (ZO; tight junction), zona adherens (ZA), desmosomes, and gap junctions. The ZO is composed of several families of proteins that determine its physical and biological properties. For example, claudins belong to a family of 24 membranespanning proteins (24-27 kd) that form pores by interactions of the extracellular domains of claudins of adjoining cells; homotypic adhesion claudins are important in determining the charge selectivity of the tight junction.15,16 Additional proteins in the tight junction include occludins, junctional adhesion molecules (JAMs), and scaffolding proteins such as the zona occludens proteins (ZO-1, ZO-2) and multi-PDZ domain protein 1 (MUPP-1). The scaffolding proteins serve to link membrane proteins to an array of protein kinases, phosphatases and, via filamentous actin, to myosin in the terminal web, thereby influencing paracellular permeability.17,18 For example, disruption of tight junctions by enteropathogenic Escherichia coli is specifically associated with protein kinase Cζ activation.19 Another junctional complex that allows cell-to-cell interaction is the zona adherens. In epithelia, the zona adherens primarily is made up of E-cadherins, 120-kd transmembrane glyco­ proteins, with extracellular motifs that engage in calciumdependent homotypic interaction with cadherins of adjoining cells. Intracellularly, cadherins bind to a family of adhesion molecules, the catenins, which in turn anchor to a dense actin-filament network. Alterations in cadherincatenin distribution or function have been implicated in carcinogenesis.20 Desmosomes are junctional complexes that are structurally similar to zona adherens junctions, although instead of actin, they link to intermediate filaments through a dense plaque of intracellular anchor proteins. Gap junctions have a unique function: They bridge gaps between cells, thus allowing neighboring cells to exchange small molecules. They are made up of an assembly of connexins, a four-pass membrane-spanning protein, six of which join to form a hemichannel. When these hemichannels in two adjoining cells are aligned, they form a continuous pore that connects the interior of the two cells.1

TRANSEPITHELIAL TRANSPORT Our current understanding of the movement of ions, solutes, and fluid across epithelia is gleaned from a combination of in vitro studies using reductionist models of cell lines or isolated epithelial sheets, and from complex in vivo metho­ dologies such as the triple-lumen perfusion technique. All these models underscore that transepithelial ion (largely Na+) movement from the mucosa to the serosa drives fluid absorption, whereas net ion (largely Cl−) movement in the reverse direction drives fluid secretion. Although different approaches help elucidate a complex mechanism, at times they give confounding results. For example, some in vitro studies report decreased Cl− secretion and increased Na+ absorption in the jejunum of cystic fibrosis patients, implying that the intestinal manifestations of the disease are due to hyperabsorption of water. In contrast, in vivo studies show decreases in both Cl− secretion and passive Cl− absorp-

tion, suggesting that rather than a hyperabsorption of fluid, the severity of the disease is reflected by decreased fluid absorption.21 The reductionist models allow us to focus on transport processes at the cellular and paracellular level. In the intact intestine, however, things are more complicated. The geometry of the intestinal wall and the unstirred layer influence the distance that an individual molecule must traverse to reach the apical membrane. The extracellular glycosylated domains of apical membrane proteins make up the glyco­ calyx, which contributes to the thickness and permeability of the unstirred layer; this layer can be a diffusive barrier to the movement of large lipophilic molecules in a chiefly aqueous milieu. Physical parameters such as the mixing of luminal contents by peristalsis, villus motility, and the finer movement of the microvilli influence this rate.

TRANSCELLULAR TRANSPORT Transcellular transport of ions and solutes can be passive or active. Because of the semipermeable nature of the lipid membrane, movement through the cell requires the deployment of specialized membrane proteins, such as channels, carriers, and pumps. The negative intracellular potential favors cation entry into, and anion exit from, the cell. This leads to the curious situation in which ions can move passively against their concentration gradient. For example, although the chemical concentration of Cl− in the cell is relatively low (~35 mmol) compared with the outside concentration (~110 mmol), the intracellular electronegativity creates a driving force for Cl− exit out of the cell.

WATER MOVEMENT Although water movement is a major property of the intestine, the mechanism(s) of intestinal water transport have not been clearly delineated. The movement of water is inextricably linked to the movement of solutes, in response to osmotic gradients. The standing-gradient hypothesis of water absorption suggests that even a small increase (2-3 mOsm) in the osmolarity of the intercellular and subepithelial spaces can cause movement of water across the epithelium, both through and around the cells.22 In the early 1950s, “water pores” were postulated to explain transepithelial water movement, but it was not until the remarkable discovery of the aquaporin (AQP) family of water transporters that a role of specific membrane proteins was implicated in erythrocyte and renal water transport.23,24 Although AQP2, AQP3, and AQP7 have been localized to the small intestine and AQP1, AQP3, AQP4, and AQP8 have been localized to the large intestine, the specific intestinal apical water channel, if any, has eluded discovery.25 Wright and associates proposed that the apical Na+-glucose transporter (SGLT) also may be able to transport water, perhaps as much as 5 L/day,26,27 but whether this can compensate for the puzzling lack of functional apical AQPs remains to be determined.

CHANNELS, CARRIERS, AND PUMPS Small hydrophobic and uncharged molecules move across the lipid bilayer of the cell by diffusion, the rate of transport

1679

1680

Section X  Small and Large Intestine A

Lipid-soluble solutes

B

Lipid-insoluble solutes*

C

D

Lipid-insoluble solutes

Water molecules Extracellular space

Membrane

Cytoplasm Simple diffusion across membrane

Channel-mediated facilitated diffusion via channel proteins

Na+

E

Na+

Carrier-mediated facilitated diffusion via carrier proteins

Osmosis: diffusion via a channel protein or across the membrane

Glucose

Na+

Na+ Extracellular space

Na+

Membrane Active transport K+

K+

1. Primary active transport: creates an ionic gradient

Cytoplasm

Na+

Na+ 2. Secondary active + transport: utilizes the Na gradient to transport uphill

Glucose

Figure 99-5.  Channels, carriers, and pumps. Because only nonpolar solutes freely cross a lipid domain by simple diffusion (A), transfer of ions and charged molecules necessitates specific transmembrane proteins to modulate entry and exit. Ion-specific channels mediate membrane transport by facilitated diffusion (B). Carriers permit facilitated diffusion and transfer specific solutes across the membrane by undergoing a conformational change (C). Trans­ cellular transport of water molecules occurs via channel proteins or carrier proteins (D). Active transport occurs against an electrochemical gradient and can be driven by adenosine triphosphate (ATP) (primary active transport; E1) or an ionic gradient (secondary active transport; E2)

determined by the concentration gradients and diffusion coefficients (Fig. 99-5). Oxygen, carbon dioxide, fat-soluble vitamins, and unconjugated bile acids are examples of substances transported by diffusion. Because the majority of ions and solutes cannot cross the phospholipid membrane by diffusion, the cell employs an array of distinct integral membrane proteins, including channels, carriers, and pumps to cross cell membranes (Fig. 99-5).1 Channels are pores that allow the swift (>106 ions/sec) and controlled (by rapid opening and closing) transit of ions across the membrane, driven by the electrochemical gradient. The advent of molecular cloning techniques, patch clamp methodology (which allows the measurement of function of single channels), and membrane protein crystallography has greatly advanced our knowledge of how these proteins function. Channels tend to be ion selective. For example, Na+ channels exclude K+ despite its same charge and smaller size. Selectivity is determined by the hydration radius of the ion and the physiochemical nature of the pore. The overall transport of a particular ion is determined by the electrochemical gradient, the density of channels, and the gating (open-close time) of the channel; gating may be modulated by voltage, ion, concentration, or intracellular regulation. Mutations of critical residues in the channel protein can have dire functional consequences; for example,

in cystic fibrosis, specific mutations of the CFTR affect the ability to transport chloride and bicarbonate.28,29 Carriers are another class of integral membrane proteins responsible for transport of ions and solutes at rates several orders of magnitude lower than channels. Carrier-mediated transport exhibits substrate specificity, saturation, and inhibitory kinetics. Carriers undergo a series of sequential conformational changes to facilitate the transport of substrates across a membrane. When concentration or electrochemical gradients drive carrier-mediated transport, the process is downhill and is termed facilitated diffusion. For example, the entry of fructose into the enterocyte via Glut-5 is by facilitated diffusion. The entering fructose is rapidly isomerized to glucose, maintaining the downhill gradient for fructose. In contrast, other carriers harness the electrochemical energy established by the downhill movement of a second ion, usually Na+, to move a solute or another ion uphill. This process is termed secondary active transport because the specific gradient is indirectly created by a distinct energy-using process. For example, glucose uptake via the apical membrane Na+-dependent glucose transporter, SGLT, is driven by the Na+ gradient generated by the basolateral Na+,K+-ATPase. Carriers exhibit substrate specificity; thus, SGLT transports d-glucose but not l-glucose. Equally impor-

Chapter 99  Intestinal Electrolyte Absorption and Secretion tant, carriers can transport single or multiple substrates and perform the transport in different directions. Uniporters, such as Glut-2 in the basolateral membrane, transport one type of substrate, hexoses. Symporters, such as the Na+/ K+/2Cl− cotransporter, move Na+, K+, and Cl− in the same direction, whereas antiporters, such as the Na+-H+ exchangers, move the two ions in opposite directions. Pumps are the third class of integral membrane proteins and directly use energy, generally adenosine triphosphate (ATP) hydrolysis, to move ions against an electrochemical gradient. This process is termed primary active transport. Na+,K+-ATPase is the quintessential pump in the intestine. Other pumps important in GI epithelial transport include the luminal gastric and colonic H+,K+-ATPases and the basolateral Ca2+-ATPases.

ION TRANSPORTERS APICAL SODIUM CHANNEL

In the GI tract, the surface epithelial cells of the distal colon and rectum exhibit electrogenic Na+ absorption against a fairly steep concentration gradient. The downhill electrochemical gradient created by the Na+ pump drives Na+ entry via an apical membrane Na+-specific ion channel (Fig. 99-6), which belongs to the family of epithelial Na+ channels (ENaCs). Members of the ENaC family are found in many epithelia.30-32 They are multimeric proteins composed of α, β, and γ subunits; they exhibit a high sensitivity to the diuretic amiloride; and they are stimulated by mineralocorticoids and cAMP. Colonic ENaCs are inhibited by increases in intracellular Ca2+. Unlike many other channels that regulate transport by gating, ENaCs modulate transport by varying the channel density in the cell membrane; this variation may be accomplished through changes (increases or decreases) in synthesis (e.g., aldosterone) or exocytosis (e.g., cAMP, vasopressin) of the channels. Additionally, both aldosterone and cAMP block the association of ENaC with Nedd4-2, a ubiquitin protein ligase, which normally flags the protein for degradation; this block of degradation increases ENaC. Mutations in this pathway can result in the Cl−

HCO3− H+

6 Cl−

1 3 Na+

Na+

+ H+ K

Cl− 3

Na+

5

~

Na+ 2

K+

4 Na+

H2O

H2O

~ GLUT2

Na+

–Glucose Figure 99-6.  Apical sodium transporters. Sodium crosses the apical membrane of the epithelial cell down an electrochemical gradient. The mechanisms may be (1) an ion-specific channel that can be blocked by amiloride; (2) a carrier (e.g., SGLT1) that couples the movement of sodium and a nutrient such as glucose; or (3) a carrier (antiport carrier; e.g., NHE3) that allows electroneutral entry of sodium in exchange for intracellular hydrogen. The common exit pathway across the basolateral membrane is the Na+ pump (4). K+ channels help maintain the electrochemical gradient (5). Cl− moves passively through the paracellular pathway or via cellular transporters (6) . Glucose exits the basolateral membrane via the facilitated diffusion hexose transporter, GLUT2. NHE3, sodium-hydrogen exchanger 3; SGLT1, sodium-glucose cotransporter 1.

increased Na+ absorption and hypertension characteristic of Liddle’s syndrome, an autosomal dominant disorder with features of hyperaldosteronism that is a cause of infantile hypertension.

NUTRIENT-COUPLED SODIUM TRANSPORT

By and large, nutrient transporters are the purview of the small intestine. Transport of many hydrophilic nutrients including glucose, amino acids, and some vitamins occurs against their concentration gradients via secondary active transport at the apical membrane and facilitated diffusion across the basolateral membrane. Glucose transport processes, elegantly elucidated by Wright and coworkers, provide a good example of nutrient transport.33 Transport across SGLT1 is electrogenic (2 Na+ to 1 glucose), is stereospecific (d-isomer), and transports galactose but not fructose.33 Glucose exit across the basolateral membrane occurs via a separate family of facilitated diffusion carriers, the glucose transporters (GLUT-2) (see Fig. 99-6). Fructose enters the cell via another member of this family, GLUT-5, and exits via GLUT-2. Although it is clear that Na+ and glucose absorption stimulates water absorption, the mechanism is not fully delineated. The classic explanation is that basolateral exit of glucose creates a hypertonic compartment in the paracellular space, thereby generating an osmotic gradient for the entry of fluid from the lumen. Some enticing evidence suggests that secondary to transcellular transport through SGLT, passive processes triggers the contraction of the actomyosin ring in the terminal web, resulting in increased paracellular permeability to glucose and to water. Additionally, SGLT activation results in a protein kinase-dependent recruitment of GLUT2 to the apical membrane, which then serves as a high-capacity, low-affinity route for sugar entry during feeding.34,35 Evidence suggests that SGLT can serve as a water channel (210 water to 2 Na+ to 1 glucose) and might account for up to 5 L of fluid reabsorption in the fed state.36 How molecular regulation of transporters is translated into net nutrient absorption during feeding in vivo is a critical area of inquiry. For a description of similar advances made in our understanding of amino acid and vitamin transport, see Chapters 100 and 101.

SODIUM-HYDROGEN EXCHANGERS

The exchange of extracellular Na+ for intracellular H+ is a process that is driven by the electrochemical gradient for Na+ and by a pH gradient resulting from a moderately acidic intracellular environment; this process occurs in almost every cell. In mammalian intestine, members of the Na+-H+ exchange (NHE) gene family play an important role in electroneutral sodium absorption. Electroneutral Na+ absorption may be down-regulated during eating and increases upon postprandial nutrient absorption. Ten mammalian isoforms of NHE have been cloned. NHE1-4 and 6-9 exhibit species- and segment-specific distribution in the GI tract.37-39 NHE1 is a plasma membrane protein found in epithelial and nonepithelial cells. It is expressed on epithelial basolateral membranes and functions as the housekeeper regulator of intracellular pH, cell volume, and growth. NHE2 and NHE3 are apical membrane proteins restricted to epithelia and are the major conduits for electroneutral Na+ absorption in the intestine (see Fig. 99-6). NHE2 is expressed throughout the GI tract, but maximal expression is in the proximal colon. NHE3 is considered a marker for the absorptive cells of the small intestine and colon; it is expressed only in the villus or surface cells, and not in the crypts. NHE4 is located in the basolateral membrane and is primarily expressed in gastric parietal and chief cells, where it might have a role in acid

1681

1682

Section X  Small and Large Intestine and NHERF2, which anchors cGMP kinase protein (GKAP). Activation of guanylate cyclase C by guanylin or E. coli– stable toxin A increases cGMP content near the brush border to locally activate cGKII (Fig. 99-7), which then inhibits NHE3 activity.40

secretion. NHE5 and NHE10 are not expressed in the GI tract, and the roles of NHE6-9 remain to be determined. NHE activity is differentially modulated by neural, paracrine, or endocrine stimuli through intricate scaffolding complexes that include the exchanger itself, a family of NHE regulatory factors (NHERFs) that act as a bridge between the exchanger, and a variety of kinases and phosphatases.40,41 Different stimuli use differing scaffolding complexes to exert their effect. For example, glucocorticoids stimulate Na+ absorption and up-regulate NHE3 but not NHE1, 2, or 4, consistent with their respective roles in vectorial transport and housekeeping. Glucocorticoids act via a serum and glucocorticoid inducible kinase, SGK1; SGK1 stimulates the activity of NHE3 by interacting directly with NHERF2. Alternatively, cAMP inhibits NHE3 by activating protein kinase A, which is recruited to the C-terminus of NHE3 by NHERF1, NHERF2, and an additional cytoskeletal protein, ezrin. In this location, PKA induces its inhibitory effect by phosphorylating NHE3. Cyclic guanosine monophosphate (cGMP) can inhibit NHE3 by triggering the formation of a complex between cGMP-dependent protein kinase II (cGKII) Guanylin STa 1

Extracellular space

ELECTRONEUTRAL SODIUM CHLORIDE ABSORPTION

Sodium absorption is coupled to the movement of Cl− through a Cl−-HCO3− anion exchanger, located specifically in the ileum and proximal colon. The rates of the transporters are similar and coordinated by cell pH and HCO3−. Alkalinization of the cell by NHE drives the exit of HCO3− in exchange for Cl−, resulting in electroneutral Na+ and Cl− absorption, maintenance of cell pH, and luminal release of H+ + HCO3− (water and CO2). The coupling of the two exchangers exhibits species and segmental variations. Thus, it is fairly tight in the ileum and proximal colon, whereas in other intestinal segments, although NHE is dependent on Cl− or HCO3−, the linkage between Na+ and Cl− transport is VIP Prostaglandins 1

Somatostatin

Membrane

R

2

4

Active PKG II

GTP Cyclic GMP

Cytoplasm

Gs 2

3

Adenylate cyclase

ATP

GKAP

Cyclic AMP

3

Active PKA

4

Gi

AKAP

Activity of membrane channels/transporters 5 Figure 99-7.  Second messengers: cAMP and cGMP. Five steps are involved in the transduction of an external signal into a change in cellular function: (1) Binding of either a stimulatory or an inhibitory agonist to an appropriate receptor of the membrane-bound adenylate cyclase or guanylate cyclase system. (2) Binding of ligand to receptor, modulates the cyclase activity either within the same molecule in the case of guanylate cyclase, or by activating the corresponding membrane-bound heterotrimeric guanine nucleotide regulatory proteins (G proteins) in the case of adenylate cyclase. (3) An intracellular signal results from production of cAMP from ATP and cGMP from GTP. (4) Increase in [cAMP]i (intracellular cAMP concentration) activates protein kinases such as PKA and increase in [cGMP]i activates protein kinases such as protein kinase G II, which is fixed to the membrane by myristoylation. Involvement of kinase-anchoring proteins such as A kinase-anchoring proteins (AKAPs) and G kinase-anchoring proteins (GKAPs) has been demonstrated in the signaling. (5) Protein kinase phosphorylation of specific target proteins results in change in the activity of channels or transporters such as chloride channel or the Na+-H+ exchanger. In cAMP signaling, binding of stimulatory regulators, such as VIP and prostaglandins, to specific receptors causes activation. Activated receptors couple via Gs to signal adenylate cyclases to catalyze the conversion of ATP to cAMP, which then activates specific cAMP kinases. An inherent GTPase returns Gs to its nascent state; in cholera, the toxin prevents this occurrence by covalently modifying Gs, leaving enterocyte turnover as the only recourse to returning the tissue to its basal state. Other hormones, such as somatostatin, trigger the activation of inhibitory G proteins (Gi) to decrease cAMP. The adenylate cyclase cascade is localized to the basolateral membrane of epithelial cells. In cGMP signaling, cGMP is generated by the activation of membrane or soluble guanylate cyclases (GCs). In contrast to the adenylate cyclases, membrane GCs are single-pass transmembrane proteins for which the extracellular domain serves as the receptor-binding domain and the intracellular domain catalyzes conversion of GTP to cGMP. Thus, the GCs are specific for their ligands, which include the endogenous atrial natriuretic peptides, guanylin and uroguanylin, as well as enterotoxins such as the heat-stable enterotoxin of E. coli. The intestinal cGMP protein kinase (PGII) is tethered to the membrane via a myristoylated N-terminal region. The soluble GCs are the target of nitric oxide (NO) activation; they are minimally expressed in the small intestinal epithelium, but they are present in colonic epithelia, subepithelial elements, and smooth muscle, where they cause muscular relaxation. ATP, adenosine triphosphate; cAMP, cyclic adenosine monophosphate; cGMP, cyclic guanosine monophosphate; Gi, inhi­bitory G protein; Gs, stimulatory G protein; GTP, guanosine triphosphate; PKA, protein kinase A; PKG, protein kinase G; STa, heat-stable toxin; VIP, vasoactive intestinal peptide.

Chapter 99  Intestinal Electrolyte Absorption and Secretion H2O

Na+

Cl− HCO3−

Lumen

cAMP, cGMP Ca2+ K+ ~ K+

Na+

Na+

2Cl− Figure 99-8.  Intestinal chloride secretion. Discrete basolateral entry steps and apical exit steps are integral to chloride secretion. The Na+-K+-2Cl− carrier couples the movement of Na+, K+, and Cl− in a 1 : 1 : 2 stoichiometric relationship and permits Cl− to accumulate in the cell in a concentration greater than its electrochemical equilibrium. Cl− then exits the cell across the apical membrane by means of a chloride channel; Na+ and water follow passively. The Na+ and K+ that entered with the Cl− are recycled by, respectively, the Na+ pump and a basolateral K+ channel, both critical to maintaining the driving force. These transporters can be regulated by second messengers such as Ca2+, cAMP (cyclic adenosine monophosphate), and cGMP (cyclic guanosine monophosphate).

variable. The molecular nature of the anion exchangers are described under HCO3− transport later.

CHLORIDE (ANION) ABSORPTION

The transepithelial, lumen negative potential difference contributes to the passive movement of Cl− and other anions via the paracellular pathway in the jejunum.7 While coupled Cl−-HCO3− and Na-H exchangers govern electroneutral transport in the ileum and proximal colon, sodium-independent Cl−-HCO3− exchange occurs in the distal colon (see Fig. 99-6) (see the discussion of HCO3− transport, later).

CHLORIDE SECRETION

The principal driving force for the secretion of fluid is the transcellular movement of Cl− from the serosal to the luminal compartment. Na+ and water follow passively in response to the ensuing electrical and osmotic gradients (Fig. 99-8). The small and large intestine exhibit a basal rate of Cl− secretion that is maintained by the interplay of cell volume, [Cl]i, and paracrine, autocrine, neuronal, endocrine, luminal, and immune modulators. Disruptions in the balance of these regulatory processes can lead to secretory diarrhea. Several epithelia in the GI tract exhibit electrogenic Cl− secretion. Although there are some tissue-specific regulatory differences, the mechanisms underlying this secretion are remarkably similar. The Na+ pump provides the driving force, Cl− enters the cell across the basolateral membrane via an electroneutral cotransporter (NKCC1) that couples the movement of 1Na+:1K+:2Cl−, and Cl− leaves the cell via specific channels on the apical membrane. The Na+ entering the cell via NKCC1 exits via the Na+ pump, and the K+ leaves via K+ channels either on the apical or the basolateral membrane. This complex interplay of transporters is an elegant demonstration of cellular economy. The NKCC1 cotransporter effectively moves 2 Cl− and 1 K+ uphill for the expenditure of a single Na+ ion. The pump-to-leak relation between K+ channels and the Na+ pump helps to maintain the interior of the cell as electronegative, thereby providing the driving force for Cl− exit. Basolateral K+ exit electrically balances the large Cl− flux across the apical membrane. NKCC cotransporters belong to a superfamily of cation transporters and are characterized by their inhibition

by the loop diuretics bumetanide and furosemide.42-44 NKCC1 is regulated by many kinases, including a unique PASK (proline-alanine–rich STE20-related kinase), phosphatases, actin-myosin interactions, cell volume, and intracellular Cl−. At least three classes of Cl− channels, belonging to distinct protein families and with distinct electrophysiologic characteristics, have been identified in secretory epithelia. The most important of these is the cystic fibrosis transmembrane conductance regulator (CFTR), which is coded for by the gene that is defective in cystic fibrosis. The crucial role of this channel is underscored by the exocrine pathologies that are the hallmark of cystic fibrosis. In addition to abnormalities of the lungs, sweat glands, and pancreas, infants with cystic fibrosis often present with meconium ileus, and 15% of adults with cystic fibrosis exhibit distal intestinal obstructive syndrome. Interestingly, the major pathology of the CFTR-deficient mouse is meconium ileus, which results in early mortality unless treated with an osmotic laxative. CFTR has been localized to the apical membrane of various segments of the small and large intestine, with greater expression in the crypts than the villus or surface cells.45 CFTR is a 250-kd membrane protein belonging to the superfamily of ATP binding cassette proteins. CFTR is a small-conductance (8-10 pS) linear channel, with an ion selectivity of Br− > Cl− > I− > F− that also can transport HCO3−.46,47 It has two membrane-spanning domains, two ATP-binding domains, and a regulatory (R) domain that has many consensus sequences for phosphorylation by kinases, specifically for PKA. Gating of the channel is regulated by sequential binding of ATP to the two domains and phosphorylation and dephosphorylation of the R domain. In secretory diarrhea such as cholera, PKA activates the R domain to increase channel activity while simultaneously stimulating the recruitment of CFTR-bearing endosomes to the apical membrane, increasing the number of channels (Fig. 99-9). Although more than 1000 mutations of CFTR have been identified, approximately 70% of cystic fibrosis patients carry the ΔF508 mutation, a single amino acid deletion that results in improper folding and diversion of the protein to the endoplasmic reticulum-associated degradative pathway rather than to the apical membrane. The protein is pleiotropic and interacts with a variety of other proteins, influencing their expression and regulation, and being modulated, in turn, by mechanisms that are not fully deciphered. Some of this interaction can occur through crosstalk of CFTR with scaffolding proteins it shares with other transporters; for example, the C-terminal of CFTR interacts with NHERF, a protein that also binds to NHE3. Diminished intestinal fluid secretion in CF−/− mice is associated with goblet cell hyperplasia, increased crypt cell proliferation, Paneth cell abnormalities, and increased inflammation.48 Whether all these changes are secondary to decreased fluid secretion or result from more direct effects of CFTR on other protein function remains to be determined (see later). By combinatorial chemistry, a series of compounds have been designed to specifically inhibit or stimulate CFTR, thereby serving as therapeutic strategies for secretory diarrhea or cystic fibrosis.49

CHLORIDE CHANNELS CIC Family Channels

Chloride channels of the distinct ClC family,50,51 in parti­ cular the widely distributed plasma membrane protein ClC2, are of clinical interest as a potential salvage pathway for Cl− transport in cystic fibrosis patients and as the specific

1683

1684

Section X  Small and Large Intestine A

B

Vibrio cholerae

Cholera toxin

A B ZOT Cl−

HCO−3

Na+

H+

GM1 No effect

Cl− −

Na+ −

Gs cAMP

+

2

Na+ H2 O

3 AC

~

1

Retrograde endocytosis

cAMP

4

EC cells

Na+ K+ VIP 5-HT

+

PG

+

Mesenchymal cells Myenteric plexus

Secretory reflex via interneurons

To colon

Figure 99-9.  Mechanism of action of Vibrio cholerae enterotoxin. V. cholerae produces an enterotoxin, and a zonula occludens toxin (ZOT) that disrupts tight junction permeability, and other toxins (not shown here) that are not fully identified. CT induces diarrhea by inhibiting the absorptive processes in the villus and surface epithelial cells (A) and by stimulating the secretory processes in the crypt epithelial cells (B). As shown in A, a CT-mediated increase in cAMP leads to an inhibition of salt absorption (Na+-H+ and Cl−-HCO3− transporters), but does not affect Na-glucose transport. The events that follow enterotoxin binding leading to cAMP generation are similar in villus and crypt cells. As shown in B, the enterotoxin binds to ubiquitous GM1 gangliosides via its B subunits on the intestinal brush border membrane (1) and, by capturing elaborate cellular processes, including lipid rafts, retrograde endocytosis, and endoplasmic reticular proteolysis (2), transmits the A1 peptide of its A subunit to the basolateral membrane. At the basolateral membrane, A1 catalyzes the adenosine diphosphate ribosylation of Gαs (Gs). This permanently activates Gs by covalently inhibiting the nascent GTPase and preventing activated Gs from returning to its nascent state (3). Activated Gs then stimulates adenylate cyclase to produce an unregulated increase in cAMP (4). Although the major pathophysiologic effects are attributed to the enterotoxin, V. cholerae also recruits multiple components of the PINES, including enteric neurons, enterochromaffin cells, prostaglandins, and serotonin (as shown at the bottom of the figure) that together contribute to the copious fluid output in the small intestine. In addition, interneurons between the small and large intestine underlie the ability of cholera toxin in the small intestine to trigger a reflex secretory response in the colon. AC, adenylate cyclase; cAMP, cyclic adenosine monophosphate; CT, cholera toxin; EC, enterochromaffin; GM1, monosialotetrahexosylganglioside; Gs, stimulatory G protein; 5-HT, serotonin; PG, prostaglandin; PINES, paracrine, immunologic, neural, and endocrine systems; VIP, vasoactive intestinal polypeptide.

target of lubiprostone, a novel laxative.52 ClC2 is involved in regulation of epithelial transport, intracellular pH, intracellular chloride and cell volume. Because of significant differences in its localization in a diverse set of human and animal models, however, its exact role in chloride secretion is uncertain. Thus, although lubiprostone activates ClC2 in the apical membrane of the human colon carcinoma cell line T84, it cannot rescue Cl− secretion in the CFTRdeficient mouse. This dichotomy suggests that the channel is located on the basolateral membrane of the CF mouse. Further research will be needed to clarify how these channels and their pharmacologic agonist work in regulating chloride secretion in humans.

infection. The rotavirus toxin NSP4 activates calciumdependent Cl− secretion, especially in the colons of young mammals. Despite an understanding of the regulation of calcium-dependent Cl− secretion in many epithelia, however, the molecular nature of these channels has eluded precise identification; recent evidence points to a group of proteins, the bestrophins, as potential candidates. It remains to be determined if bestrophins are regulated channels per se or regulators of channels or whether they use auxiliary proteins to function as channels.53,54 The CLCA proteins also might play a role in goblet cell function: Mouse CLCA3 is severely reduced in the CF−/− mouse, and its restoration ameliorates the severity of intestinal cystic fibrosis disease.55

CALCIUM-ACTIVATED CHLORIDE CHANNELS

POTASSIUM TRANSPORT

A third class of Cl− channels, the calcium-activated Cl− (CLCA) channels, are involved in the diarrhea of rotavirus

A plethora of K+ transport processes help the intestine cope with its need to balance fluid and electrolyte movement.56,57

Chapter 99  Intestinal Electrolyte Absorption and Secretion Potassium secretion and absorption occur along the length of the intestinal tract, although the specific pathways are segment specific. In the small intestine, the luminal negative potential difference drives the passive absorption of K+. In contrast, K+ absorption in the distal colon occurs by primary active transport via H+,K+-ATPase pumps located on the luminal membrane. These H+,K+-ATPase pumps are P-type ATPases, related to the gastric H+,K+-ATPase, and least two colonic isoforms have been identified: a ouabain-sensitive isoform in the crypt cells and a ouabain-insensitive isoform in the surface cells. Depletion of aldosterone and K+ upregulates the ouabain-insensitive H+,K+-ATPase and stimulates K+ absorption. Potassium channels are the largest group of ion channels in the human genome. A number of K+ channels have been localized to intestinal cells, but only a few are mentioned here (using the nomenclature of the Human Genome Organization). Found primarily on the basolateral membrane of intestinal epithelial cells, K+ channels contribute significantly to intestinal electrolyte homeostasis through several mechanisms. K+ channels modulate the hyperpolarization of the cell interior that is needed for vectorial, voltagedriven transport processes. In the basolateral membranes of the small and large intestine, the cAMP-activated KCNE3/KCNQ1 channels and the Ca2+-calmodulin– activated KCNN4 channels hyperpolarize the membrane and promote Cl− secretion. Basolateral K+ channels contribute to the transepithelial potential difference, which influences paracellular movement; K+ secretion in the colon occurs via apical KCNMA1 (MaxiK) channels, which are regulated by mineralocorticoids. In response to cell swelling, K+ channels are activated and cause a regulatory volume decrease, a critical function for intestinal cells faced with constant fluctuations in osmolarity. K+ channels also play roles in differentiation, apoptosis, and carcinogenesis, functions that involve different channel proteins and that can be regulated by a number of processes including protein modifications (phosphorylation, sumoylation), membrane voltage, cytosolic calcium, pH, cell swelling, and cell metabolism.

BICARBONATE TRANSPORT

Bicarbonate is a metabolic product and a critical anion in fluid homeostasis in the intestine. In clinically significant diarrhea, bicarbonate is the major anion in the stool. It is secreted by electrogenic and electroneutral processes in the duodenum, ileum, and colon. Being a metabolic product, intracellular HCO3− can arise from intracellular metabolism, diffusion of CO2 or the action of transporters such as the basolateral Na+-HCO3− cotransporter. Electrogenic HCO3− secretion can occur via apical anion channels, including CFTR; however, the major mechanism for HCO3− secretion in the small and large intestine is inexorably linked to the inward movement of Cl− and occurs through apical Cl−HCO3− exchangers.58 It is postulated that electrogenic Cl− secretion via CFTR provides luminal Cl−, which is then recycled across the apical membrane in exchange for intracellular HCO3−. SCFA-dependent bicarbonate secretion also has been observed in surface cells of the colon.59 Although the prototype of anion exchangers, the red cell Cl−-HCO3− exchanger (AE1), has been extensively studied, identification of intestinal exchangers is relatively recent. In a reorganization of nomenclature, the more than 360 identified solute carriers (SLC) are now classified in one of 46 families; of these, the SLC4 bicarbonate transporter family encompasses the red cell anion exchanger (AE1) and its epithelial isoforms (e.g., AE2, AE3, and AE4).60-62 The precise role of the AEs in overall intestinal Cl−-HCO3− trans-

port remains to be clarified, and apical and basolateral localizations have been suggested. Structurally distinct from the SLC4 bicarbonate transporter family is the SLC26 multifunctional anion exchange family. SLC26 exchangers can transport Cl−, HCO3−, sulfate, formate, oxalate, hydroxyl ions, and other anions with differing affinities. Their varied distribution along the GI tract provides them with the flexibility to handle a variety of luminal anions. Two members of this family, SLC26A3 and SLC26A6, are of special interest.10,63-65 The former, first identified as DRA (down-regulated adenoma), is expressed abundantly on the apical membranes of colonocytes, but not enterocytes, and transports more than 2Cl−:1HCO3− ion. Mutations in DRA cause congenital chloride diarrhea, which manifests with severe diarrhea, volume depletion, and metabolic alkalosis. In contrast, SLC26A6, also known as the putative anion transporter 1 (PAT1), is expressed abundantly in the apical membrane of villus enterocytes and less so in the colon. PAT1 transports more than 2HCO3−:Cl− ion; a separate HCO3− conductive pathway might mediate bicarbonate secretion into the duodenum. In contrast to the apical NHE isoforms, anion exchangers are present on the apical membranes of crypt and surface cells. A number of regulatory pathways, including cGMPmediated pathways, strongly influence HCO3− secretion. The physiologic implications of the spatial distribution and varied anion transporters in net HCO3− secretion remain to be elucidated. It is, however, intriguing that HCO3− transport is dependent on CFTR and that anion exchange processes are down-regulated in cystic fibrosis.

SHORT-CHAIN FATTY ACID TRANSPORT

The 2- to 4-carbon short-chain fatty acids (SCFAs), such as acetate, propionate, and butyrate, are generated by the fermentative action of the bacterial microflora on poorly absorbed carbohydrates. Unlike the small intestine, in which Cl− and HCO3− predominate, the major anions (60 to 150 mmol/kg) in the colon are SCFAs; thus, there is a large SCFA concentration gradient across the colonic epithelium. The magnitude of the daily colonic load and absorption of SCFAs is comparable to that of colonic Na+. SCFAs are a major source of metabolic fuel for the colonocytes, modulate epithelial growth and differentiation, and have been implicated in the pathogenesis of and therapy for several inflammatory diseases of the colon, such as bypass colitis. SCFAs are rapidly absorbed in the colon and also greatly enhance Na+ and fluid reabsorption through linked transport mechanisms and by upregulating the expression of NHE3 on the apical membrane of colonocytes.66 SCFAs are weak electrolytes and can be ionized or protonated. Ionized SCFAs need specific carriers, whereas nonionized protonated species can diffuse across the colonocyte membrane; at luminal colonic pH, SCFAs are 95% to 99% ionized. A picture of the molecular basis of SCFA transport is beginning to emerge.32,67 First, apical NHEs can create an acidic pH microclimate and enhance the diffusion of protonated SCFAs into the cell. Monocarboxylate transporters (MCTs), members of two different SLC families, are involved in electroneutral carrier-mediated SCFA transport. Members of the SLC16 family, specifically SLC16A1 (MCT1), transport 1H:1SCFA and require an ancillary protein for their function.68 High-affinity (SLC5A8) and low-affinity (SLC5A12) Na+-dependent MCTs have been identified in the colon and intestine. Although their molecular identity is unclear, Cl−-butyrate exchangers and SCFA-HCO3− exchangers, functionally coupled to Na+-H+ exchange, might

1685

1686

Section X  Small and Large Intestine account for SCFA promotion of electroneutral Na+ and Cl− absorption.59,69

PARACRINE IMMUNONEUROENDOCRINE SYSTEM Extracellular factors from classic paracrine, immunologic, neural, and endocrine systems (PINES) as well as autocrine factors and luminal content regulate intestinal ion transport. The borders separating members of PINES are arbitrary at best, because there is considerable overlap and crosstalk of the underlying factors, and patterns can be altered in disease states (Fig. 99-10). Luminal mechanical (stroking and stretch) or chemical (toxins) stimuli can activate mechanoreceptors and chemoreceptors, respectively, to activate one or more arms of PINES. All these interactions are compounded by many of the factors acting through cell-specific multiple receptors and signaling pathways. Within the subepithelium, structural elements of PINES, including blood vessels, are in close proximity (see Fig. 99-10), so release of mast cell mediators can easily target neurons and vice versa; this proximity and interplay contributes to the minute-by-minute local regulation necessary in the intestine. Although it is possible to separate the specific effects of an individual component in vitro, clinically they are inextricably intertwined. For example, VernerMorrison syndrome (pancreatic cholera) is classified as an endocrine-mediated diarrhea, because pancreatic islet cell tumors produce large amounts of vasoactive intestinal

Food

peptide (VIP). In the healthy adult, however, VIP is not found in the pancreas but is a peptidergic neurotransmitter in the enteric nervous system that stimulates epithelial cell secretion and smooth muscle relaxation. Alternatively, a single agonist such as cholera toxin directly acts on epithelial cells while simultaneously stimulating neural, paracrine, and immune responses. In a third example, serotonin (5-hydroxytryptamine [5-HT]) released by mucosal enterochromaffin cells acts via distinct receptors on epithelial cells to directly stimulate secretion; it acts on myenteric neurons to release acetylcholine (Ach) and elicit migratory contractions; or it acts on submucosal neurons to release Ach and calcitonin gene-related peptide to stimulate peristalsis and secretory reflexes. Fluid secretion is the major component in the production of diarrhea, and the organism’s defensive response to intestinal challenge; motility, mucus secretion, and blood flow, all regulated by PINES, are important adjuncts to the process. The involvement of PINES in motility helps to explain diarrhea associated with rapid intestinal transit (e.g., following gastrectomy), altered anorectal motility (e.g., small-volume diarrhea), or decreased motility (e.g., bacterial overgrowth). Decreased motility leads to a increase in the bacteria in the small intestine, which causes diarrhea by a variety of mechanisms (see Chapters 97 and 102). Alternatively, inflammatory mediators such as prostaglandins or bacterial enterotoxins target both the epithelial and muscle layers to elicit a coordinated secretory response, while promoters of absorption such as opiates and enkephalins suppress motility and promote electrolyte absorption. Thus, PINES allow for a coordinated and integrated response to multiple extracellular signals.

Bacterial toxins/ rotavirus Bile acids

Villus

Immune

Submucosal plexus

Neurocrine 2 Myenteric plexus 1

Autocrine

Ach

PG Mesenchymal cells

3

Paracrine

Endocrine 4 Capillary Circular muscle

Crypt Figure 99-10.  Model depicting the integral components of the enteric nervous system and the immune system that regulate intestinal ion transpart. The components include (1) neurons responsive to intraluminal mechanical and chemical stimuli (e.g., food, bile acids, bacterial toxins, rotavirus); (2) interneurons in either the myenteric or submucosal plexuses; (3) secretory neurons that release acetylcholine (Ach) which acts on epithelial cells; and (4) interactions among secretory neurons and blood vessels, immune cells, and paracrine cells. Intestinal cells can release an array of secretory factors, which can act either directly on the epithelium or indirectly by stimulating the mesenchymal cells or enteric neurons to release prostaglandins (PGs) or acetylcholine (Ach). See text for details.

Chapter 99  Intestinal Electrolyte Absorption and Secretion EXTRACELLULAR REGULATION Tables 99-1, 99-2 and 99-3 list the major neurohumoral substances and toxins that modulate intestinal fluid transport. Agents that promote net fluid secretion generally inhibit Na+ absorption and stimulate Cl− secretion, whereas agents that promote net fluid absorption increase Na+ uptake and attenuate Cl− secretion. In a healthy person, net absorption prevails, and when this balance is disrupted, diarrhea

Table 99-1  Agents That Stimulate Intestinal Absorption of Fluid and Electrolytes Endogenous Absorbagogues α-Adrenergic agonists Aldosterone Angiotensin Enkephalins Glucocorticoids Growth hormone Neuropeptide Y Peptide YY Prolactin Short-chain fatty acids Somatostatin Pharmacologic Agents Berberine Clonidine (α2-agonist) Cyclooxygenase inhibitors Glucocorticoids Lithium Mineralocorticoids Octreotide Opiates Propranolol

can ensue. It is unclear if there is a corollary for a pre­ dominant absorptive pattern in a subset of patients with constipation.

ENDOCRINE, PARACRINE, JUXTACRINE, AND AUTOCRINE REGULATION Intestinal transport is modulated by classic endocrine, paracrine, juxtacrine, and autocrine processes. Intestinal endocrine cells are interspersed between epithelial cells and function as sensors that rapidly respond to changes in the luminal environment by releasing secretory granules containing biogenic amines and hormones; these mediators cross the basolateral membrane. These hormones can act either in a classic endocrine manner on distant target cells (via the circulation) or in a local (paracrine) manner by affecting neighboring cells in the intestinal wall. Juxtacrine mediators are those released from nonendocrine cells, such as neural and inflammatory cells, and they affect neigh­ boring cells. Intestinal mesenchymal cells, in particular myofibroblasts, are a rich source of cytokines, chemokines, eicosanoids, and growth factors that can alter intestinal transport. Epithelial cells can self-regulate (autocrine) by secreting factors such as eicosanoids, which act on epithelial cell receptors to alter function.70

NEURAL

Neural input is critical in the regulation of fluid and electrolyte transport (see Fig. 99-10) and involves interactions of the parasympathetic and sympathetic divisions of the autonomic nervous system with the labyrinthine enteric nervous system (ENS) (see Chapters 1, 97, 98 and 120). Cholinergic stimulation of secretion, predominantly through parasympathetic vagal input, and adrenergic stimulation of

Table 99-2  Endogenous Agonists of Intestinal Secretion AGONIST Acetylcholine Adenosine Arachidonic acid Atrial natriuretic peptide Bombesin Bradykinin Calcitonin, calcitonin gene-related peptide Galanin Gastric inhibitory polypeptide Gastrin Guanylin Histamine Leukotrienes Motilin Neurotensin Nitric oxide Peptide histidine isoleucine Platelet activating factor Prostaglandins Reactive oxygen metabolites Secretin Serotonin Substance P Vasoactive intestinal polypeptide

INTRACELLULAR MEDIATOR 2+

Ca cAMP cAMP cGMP Ca2+ cAMP ? Ca2+ ? Ca2+(PKC/MaPK?) cGMP Ca2+ ? Ca2+ Ca2+ cGMP cAMP cAMP cAMP cAMP cAMP Ca2+ Ca2+ cAMP

cAMP, cyclic adenosine monophosphate; cGMP, cyclic guanosine monophosphate; ENS, enteric nervous system.

SOURCE ENS Immune cells Immune cells, cell membranes Heart ?? Immune cells ENS ?? Endocrine cells Goblet, epithelial cells Immune cells Immune cells Endocrine M cells ENS Immune cells, mesenchymal cells ENS Immune cells Immune cells, mesenchymal cells Immune cells Endocrine cells ENS ENS ENS

1687

1688

Section X  Small and Large Intestine Table 99-3  Luminal Agents That Stimulate Intestinal Secretion AGENT Bacterial Enterotoxins Aeromonas sp. Campylobacter jejuni Clostridium difficile (toxin A) Clostridium perfringens Escherichia coli (heat labile toxin) E. coli (heat stable toxin) Rotavirus NSP4 Salmonella sp. Vibrio cholerae   accessory cholera enterotoxin   enterotoxin   zona occludens toxin Vibrio parahaemolyticus Yersinia enterocolitica Miscellaneous Agents Bile salts Laxatives Long-chain fatty acids

INTRACELLULAR MEDIATOR cAMP cAMP Ca2+ ?? cAMP cGMP Ca2+ cAMP ?? cAMP ?? Ca2+ cGMP cAMP/Ca2+ ?? cAMP/Ca2+

cAMP, cyclic adenosine monophosphate; cGMP, cyclic guanosine monophosphate; NSP, nonstructural protein.

absorption through prevertebral and sympathetic ganglia have long been recognized as fundamental neural pathways affecting the intestinal epithelium. The ENS is the end controller of neural activity in the intestinal wall, however, independently integrating the regulation of the epithelia, muscles, and blood vessels, with input and modification from the central nervous system. Like other neural networks, the ENS has reflexes that can have important clinical implications. Target cells for neurons include components of PINES, blood vessels, and, of course, epithelial cells. Sensory input into the ENS comes from changes in the luminal content (e.g., acidity, dietary content, pathogens) or volume (e.g., stretch). Thus, acid or distension can activate TRVP1 vanilloid receptor on capsaicin-sensitive afferent nerves, which in turn evokes secretion by stimulating submucosal neurons and causes vasodilation by directly activating submucosal arterioles.71,72 Endocrine, dendritic, and/or paracrine cells releasing serotonin, adenosine (see later), and other signals are implicated as auxiliary sensors. Primarily cholinergic, interneurons are believed to underlie the ENS-mediated regulation of colonic epithelial responses to distant small intestinal challenges. Motor neurons that innervate epithelial and submucosal cells can be cholinergic or VIPergic, each releasing additional neuroactive substances. A basal cholinergic secretory drive is tempered by the sympathetic tone; loss of adrenergic sympathetic innervation in diabetic neuropathy is associated with the development of “diabetic diarrhea” and may be corrected by α2-adrenergic agonists.73 As in the brain, the number and variety of compounds that act as neuroactive agents has ballooned, and this has added to the complexity of our elucidating the underlying mechanisms. Thus, neurons can release specific combinations of mediators, such as VIP, cholecystokinin, gastrin-releasing peptide, and ATP, rather than a single substance. Individual neurotransmitters can have biphasic effects varying with concentrations. Furthermore, agents can act as classic neurotransmitters; alternatively, they can act as neuromodulators, fine-tuning the neuronal circuits of presynaptic sites of neurons, or, as in the case of serotonin, they can function as paracrine substances.

IMMUNOLOGIC The clinical correlation between intestinal inflammation and diarrhea is obvious (see Chapter 2), with ulceration, exudation of protein, changes in motility, and loss of absorptive surface area implicated as causing the fluid losses of (inflammatory) diarrheas. Immunocompetent cells of the intestine reside chiefly in the lamina propria, form the gutassociated lymphoid tissue (GALT), and often are distinct from their systemic counterparts; in noninflamed intestine, T lymphocytes account for 60%, with smaller numbers of B lymphocytes and plasma cells (25% to 30%), macrophages (8% to 10%), mast cells, and polymorphonuclear cells (2% to 5%) (usually eosinophils).74 These cells secrete a vast array of soluble products (chemokines, cytokines, eicosanoids, nucleotides, biogenic amines) whose actions are intimately intertwined with other aspects of PINES. Intestinal inflammation increases the number of immunocytes, the cause of the inflammation determining the type of inflammatory cells; for example, acute bacterial infections increase polymorphonuclear leukocytes, whereas parasitic infections dramatically enlarge the mast cell population and celiac disease is characterized by intraepithelial lymphocytes. In IBD, there is activation of all components of GALT with an increase in immunoglobulin (Ig)Gsecreting cells.75,76 Thus, the cause of the inflammatory reaction can determine the type of immunocytes recruited, the range of cytokines released, and the specific effects on transport and motility. Many inflammatory mediators are potent secretagogues, including peptides (e.g., cytokines, platelet-activating factor, substance P, interferon gamma, kallikreins, and bradykinin), eicosanoids (e.g., arachidonic acid, leukotrienes, and prostaglandins), and oxidants (e.g., superoxides). These mediators either interact directly with epithelia to alter ion transport and barrier function or elicit these effects indirectly by activating other PINES elements (see Fig. 99-10). Prostaglandins are central to the secretory response associated with inflammation, affecting various PINES components, such as enteric neurons, and with mediators such as bradykinin, which liberates arachidonic acid to stimulate prostaglandin production. In interpreting the effects of the inflammatory mediators in normal model systems, it is

Chapter 99  Intestinal Electrolyte Absorption and Secretion important to recognize that in vivo, cells damaged by the inflammatory process might not be able to function normally. With this caveat, a few examples are provided. Mucosal mast cells are strategically located in close proximity to enteric neurons, blood vessels, and epithelial cells, and they are central to several inflammatory reactions. Mast cell mediators including histamine, eicosanoids, and cytokines elicit secretion by direct effects on the epithelial cells and by indirect neural stimulation and prostaglandin release. The mechanisms of secretion resulting from polymorphonuclear infiltration of the mucosa recently have become better understood. Polymorphonuclear leukocytes, responding to chemoattractants (e.g., fMLP [formyl-Met-Leu-Phe]), uniquely present in the inflamed colon, leave the vasculature and interact with epithelial cells.77 The white cells burrow through the intercellular space of the colonic cells in a complex integrin-dependent process. The migrating leukocytes release 5-AMP, which is converted to adenosine by apical membrane enzymes. Adenosine is a potent secretagogue, and this adenosine-stimulated secretion might serve as a mechanism to cleanse the crypt lumen.78 Thus, complex specific immunocyte-epithelial cell interactions are important in alterations of electrolyte secretion associated with mucosal inflammation. Anti-integrin targeted therapy may be effective in treating Crohn’s disease (see Chapter 111). During inflammation, oxidants such as superoxides, hydrogen peroxide, and hydroxyl radicals released from neutrophils stimulate Cl− secretion; cytokines such as interleukin-1 and interleukin-3 also stimulate secretion.79 In contrast, interferon-γ and TNF-α can cause diarrhea more through an antiabsorptive effect, by down-regulating particular transporters 80-82 or by altering permeability of tight junctions. Similarly, studies on colonic specimens from patients with IBD or ulcerative colitis indicate that diarrhea in these entities results not from stimulated secretion but from increased tissue permeability combined with decreased apical Na+ channel, basolateral K+ channel, and Na+,K+ATPase expression, which decreases NaCl reabsorptive capacity.76 The multiplicity of transport malfunctions seen in IBD might reflect a sick cell syndrome rather than specific alterations modulated by one or two unique cytokines.

SYSTEMIC EFFECTS Acid-base balance modulates intestinal electrolyte transport in vivo and in vitro. Changes consistent with metabolic acidosis are potent stimulators of electroneutral NaCl absorption, whereas metabolic alkalosis inhibits this process.83-84 Intracellular bicarbonate concentrations can modulate basal Cl− secretion, and intracellular pH and Pco2 can alter Na+:H+ exchange. Volume status and intestinal blood flow also alter ion transport. Any decrease in intravascular volume, such as with hemorrhage, elicits a series of responses that increase fluid absorption. Cardiopulmonary mechanoreceptors and carotid baroreceptors increase sympathetic input into the ENS, resulting in decreased secretion. Angiotensin II, antidiuretic hormone, and atrial natriuretic peptide also can contribute to regulation of intestinal fluid transport in these conditions.85-86 The metabolic status of the intestine has an impact on its transport capability; a well-fed intestine transports more effectively.87 There also are segmental preferences for metabolic fuels. Although the entire intestinal tract uses glucose, the small bowel

effectively uses glutamine, and the colon preferentially uses SCFAs, particularly butyrate.88,89

OSMOTIC EFFECTS Unlike the kidney, the intestinal epithelium cannot maintain an osmotic gradient. Under normal physiologic conditions, the duodenum and upper jejunum are subject to major fluid shifts as they adjust to dietary intake of hypertonic foods and liquids. Rapid equilibration usually is accomplished by movement of water into the intestinal lumen, and absorptive processes along the remainder of the intestine steadily decrease the luminal volume. The continued presence of a nonabsorbable solute within the intestinal lumen, however, can negate functioning absorptive pathways in the distal intestine. This is the basis for osmotic diarrhea (see Chapter 15). Carbohydrates, usually disaccharides, are a common source of nonabsorbable solute. Disaccharides must be hydrolyzed to monosaccharides before they can cross the apical membrane of the small intestine, (see Chapters 100 and 101). The most common clinical example of maldigestion resulting from a deficiency of a specific disaccharidase is lactose intolerance, in which the glucose-galactose disaccharide cannot be broken down because of lactase deficiency. Because the human intestine does not naturally possess a lactulase, the disaccharide lactulose reliably increases small intestinal fluid because of luminal hyperosmolarity and bacterial fermentation of the disaccharide. The limited intestinal absorptive capacity for several sugars found in processed foods and drinks, such fructose and sorbitol, can play important, but often overlooked, roles in osmotic diarrhea, bloating, abdominal pain, and irritable bowel syndrome. The physiology of carbohydrate-induced osmotic diarrhea is complicated by the fact that nonabsorbable solutes in the small bowel can be converted into absorbable solutes by colonic bacteria. Almost all classes of carbohydrates not absorbed by the small intestine are rapidly converted to SCFAs once they cross the ileocecal valve and encounter the colonic bacterial flora; these SCFAs can be absorbed and serve as metabolic fuel for the colon. Thus, depending on the rate of carbohydrate conversion to SCFAs and the colonic capacity for SCFA absorption, small intestinal fluid loss may be compensated by colonic fluid absorption. However, if the capacity of the intestinal flora to convert carbohydrates to SCFAs is maximized, or if the SCFA absorptive capacity of the colon is exceeded, additional unmetabolized carbohydrates could pass through the colon and exacerbate the osmotic effects of nonabsorbable solute.90 Cations (e.g., magnesium) or anions (e.g., sulfate and phosphate) are absorbed poorly by the normal intestine. Increased ingestion of these ions easily leads to osmotic diarrhea, which may be intended, such as with laxatives, or unintended, such as with ingestion of magnesiumcontaining antacids or supplements. In clinical situations in which there is malabsorption or a generalized destruction of the epithelium, solutes normally absorbed readily can remain in the intestinal lumen and thereby contribute an osmotic component to an inflammatory diarrhea or a malabsorptive state. Osmolality is an important factor in patients receiving enteral nutrition (see Chapters 4 and 5). For example, complex carbohydrates provide a significant amount of calories with minimal osmolality compared with simple sugars. Absorption of dipep-

1689

1690

Section X  Small and Large Intestine tides and tripeptides instead of amino acids reduces intestinal osmolality. This balance between calories and osmolality becomes clinically relevant in effectively designing appropriate tube-feeding regimens. Osmolality also is an important factor in designing second-generation oral rehydration therapy (ORT) formulations; by replacing glucose with complex carbohydrates such as rice, intestinal absorption is stimulated further by creating a hypotonic luminal environment, thereby enhancing water absorption. In addition to providing numerous sugar molecules per milliosmole, complex carbohydrates such as rice and amylase-resistant starches have another advantage in ORT: They are metabolized by commensal bacteria in the colon to release SCFAs, which in turn promote fluid absorption (see the earlier discussion of SCFAs).91

SPECIFIC REGULATORY FACTORS ABSORPTIVE

Intestinal agents that stimulate absorption are listed in Table 99-1. Mineralocorticoids (e.g., aldosterone) primarily affect electrogenic Na+ absorption in the distal colon and have little effect on the small intestine, which exhibits electroneutral Na+ absorption. Aldosterone increases the activity and numbers of the apical membrane ENaC (see the earlier discussion of Na channels) and stimulates activity of the Na+/K+ pump and SGK1, resulting in an increase in Na+ absorption. Aldosterone increases K+ absorption and K+ secretion.92 Neonates exhibit a correlation between high circulating levels of aldosterone and enhanced colonic Na+ absorption. Clinically, the physiologic role of aldosterone can be seen in the increased colonic absorption after Na+ depletion (aldosterone stimulation) or in the diarrhea associated with Addison’s disease (aldosterone deficiency). Glucocorticoids are also potent stimulators of Na+ absorption in the small intestine and the colon, in addition to having well-documented anti-inflammatory effects. At low concentrations, glucocorticoids stimulate electroneutral Na+ absorption and suppress electrogenic Na+ absorption, whereas at high concentrations they stimulate both processes. The actions of glucocorticoids are complex and are species and segment specific, and they may be directed at the level of apical Na+ transporters and at the Na+ pump. For example, in the rabbit jejunum, ileum, and colon, glucocorticoid-stimulated increases in Na+ absorption are associated with selective increases in NHE3 but not NHE2 or NHE1 mRNA and protein. Both glucocorticoids and aldosterone evoke rapid cellular responses involving the SGK1 pathway as well as genomic transcriptional effects. These effects might account, in part, for the potent antidiarrheal action of glucocorticoids in a wide variety of clinical settings.38 Catecholamines, enkephalins, and somatostatin all stimulate electroneutral Na+ absorption and often decrease HCO3− secretion with similar patterns of action. They bind to specific heptahelical membrane receptors and activate the Gαi cascade, which in turn suppresses the prosecretory cAMP signaling cascade. Catecholamines (e.g., dopamine and epinephrine) acting as α-adrenergic agonists, all have similar absorptive properties. The theoretical basis for the use of clonidine as an antidiarrheal agent, particularly in diabetic diarrhea, is rooted in this adrenergic absorptive pathway.93 The use of plant opiates as antidiarrheal agents dates back two millennia to the early Egyptians and underscores

their effectiveness. Elucidating their therapeutic effect led to the characterization of the mammalian opioid peptides— enkephalins, endorphins, and dynorphins—a classic example of molecular mimicry.94 Acting via one of three main opioid receptor subtypes—mu, delta, and kappa—the opiates and opioid peptides decrease secretion and promote nonpropulsive motility patterns, thereby increasing transit time. They can act directly on the epithelial and smooth muscle cells or can modify the electrical and synaptic behavior of ENS neurons. The constipation associated with morphine intake can be due to hyperpolarization of secretomotor neurons and suppression of secretion or to a centrally mediated stimulation of sympathetic noradrenergic discharge, or both. Direct activation of K+ channels and inhibition of Ca2+ channels via a G protein-mediated mechanism underlie these effects. Chronic treatment with opiates leads to tolerance, and diarrhea ensues upon abrupt withdrawal. Clinically, management of constipation in patients receiving opiates as analgesics can be a clinical challenge. The development of long-acting analogs of somatostatin has transformed this ubiquitously distributed hormone from a physiologically fascinating regulator to a clinically relevant pharmacologic agent (see Chapter 1). In the intestine, enterochromaffin D cells produce somatostatin, which stimulates salt and water absorption in the ileum and colon and blocks the effects of several secretagogues.95,96 Somatostatin analogs such as octreotide are effective in treating several types of diarrheal diseases, particularly endocrine-related secretory diarrhea. Their therapeutic effect is due to a combination of actions, including inhibition of hormone release by tumors, slowing of intestinal transit, and a direct effect on epithelial cells. Paradoxically, elevated somatostatin levels, as encountered in somatostatinomas or with large pharmacologic doses of octreotide, can precipitate diarrhea secondary to steatorrhea.93 Other peptide hormones, including peptide YY, angiotensin II, and insulin, have been implicated as proabsorptive agents, but their physiologic significance is as yet unknown.

SECRETORY

Endogenous agents that stimulate secretion are listed in Table 99-2. Although there are subtle differences in their biologic actions, in general, eicosanoids, including arachidonic acid, prostaglandins, and leukotrienes, inhibit electroneutral NaCl absorption and stimulate electrogenic Cl− secretion. Most intestinal prostaglandins arise from submucosal immunocytes. They have a major autocrine effect on epithelial cells, but they also modulate enteric nerves and affect intestinal motility and blood flow. Depending on the type of prostaglandin and receptor subtype, prostaglandins primarily act via cAMP and, to a lesser extent, via intracellular Ca2+. Prostaglandins can contribute to the basal secretory tone of the epithelium, because cyclooxygenase inhibitors, such as indomethacin or aspirin, increase basal rates of absorption. Increased intestinal production of eicosanoids contribute to the diarrhea of IBD (see Chapters 111 and 112). Glucocorticoids can decrease prostaglandin synthesis. Although the 5-ASA (acetylsalicylic acid) class of medications, which are a mainstay of the treatment of IBD, target cyclooxygenase and decrease prostaglandin production, their clinical efficacy probably depends on additional mechanisms of action, including effects on other inflammatory pathways, including nuclear factor κB (NFκB) and reactive oxygen species (ROS).97,98 Leukotrienes are considered also to have a role; the mechanism of their action is less well understood

Chapter 99  Intestinal Electrolyte Absorption and Secretion but might involve activation of secretomotor neurons in the subepithelium. Acetycholine, serotonin, guanylin, VIP, and other hormones and neurotransmitters have been implicated as stimulators of intestinal secretion, generally acting to inhibit electroneutral NaCl absorption and stimulate Cl− secretion. They are classified by their mechanisms of action. Acetylcholine acts via the muscarinic receptor M3. Serotonin and neurotensin increase intracellular Ca2+, and VIP and related peptide hormones (e.g., secretin, peptide histidine leucine, and peptide histidine methionine) increase intracellular cAMP. Guanylin acts by increasing cell cGMP.

Serotonin

Serotonin plays a critical role in modulating intestinal motility, sensation, and secretion and is responsible for the diarrhea associated with carcinoid tumors. About 95% of the body’s serotonin is produced by enterochromaffin cells, and the remainder is produced by serotoninergic neurons of the myenteric plexus. Sensory receptors on enterochromaffin cells are activated by mechanical stimuli, acidity, invading pathogens, and dietary contents; for example, SGLT-like protein activates enterochromaffin cells, which serve as a glucose sensor to secrete serotonin (see Fig. 99-9).99 Although not completely elucidated, this signaling involves a complex sequence of autocrine and paracrine actions: ATP is released and converted extracellularly to ADP, which in turn activates a purinergic (P2Y) receptormediated calcium signaling cascade in the enterochromaffin cell to release 5-HT.99 5-HT then acts in a paracrine manner to stimulate epithelial cells, intrinsic primary afferent neurons (IPANs), and extrinsic primary afferent neurons (EPANs). Specific 5-HT receptor subtypes on different IPANs modulate the secretory reflex. Thus, 5-HT1PR on submucosal IPANs and amplifying presynaptic 5-HT4 receptors cause the release of acetylcholine and calcitonin generelated peptide, which stimulate peristaltic and secretory reflexes. In contrast, 5-HT3Rs on myentric IPANs trigger the release of acetylcholine to stimulate giant migrating contractions.99-101 Serotonin stimulation of EPANs results in CNSmediated responses of nausea and discomfort. The major mechanism of serotonin inactivation is by a serotonin reuptake transporter (SERT) on enterocytes and neurons. Interestingly, SERT may be decreased in patients with diarrhea-predominant irritable bowel syndrome (IBS-D) and ulcerative colitis and might account for their increased colonic motility and diarrhea. There has been a concerted effort to clinically alter intestinal function by pharmacologic manipulation of specific serotonin receptors. Thus, 5-HT3 receptor antagonists, such as alosetron, are used to treat IBS-D, and tegaserod, a partial 5-HT4 agonist, can alleviate constipation associated with IBS. Unfortunately, significant side effects have limited their clinical use (see Chapter 118).93,101,102 Adenosine and related purine nucleotides play unique and complex roles in modulating secretion in vivo. As described earlier, they can stimulate secretion directly as it occurs in response to polymorphonuclear leukocyte infiltration of the mucosa or indirectly via release of 5-HT.99 Adenosine acting via P1 purinoreceptors, however, has been shown to attenuate secretion evoked by mechanical stimuli. Activation or inhibition of different populations of channels can underlie these seemingly opposite effects.

Guanylin, Nitric Oxide, and Reactive Oxygen Metabolites

The search for an endogenous activator of the E. coli heatstable enterotoxin receptor led to the discovery of guanylin and uroguanylin, another example of molecular mimicry.

These small peptides, synthesized in goblet and columnar cells activate membrane guanylate cyclase to increase intracellular cGMP and elicit fluid secretion. The guanylin family of peptides, in conjunction with the atrial natriuretic peptides, calibrate the intestinal-renal acid-base response axis. In contrast, nitric oxide, a neuroimmune regulator, stimulates soluble guanylate cyclase to increase cGMP.103 This enzyme is far more prevalent in the subepithelium of the small intestine, but it is expressed in colonic epithelia. Reactive oxygen metabolites, including oxygen free radicals and hydrogen peroxide, generally are produced by immune cells, but they can be released by epithelial cells under certain conditions. They stimulate fluid secretion (see the discussion of immunologic regulation). Additional agonists stimulate secretion (see Table 99-2). Interestingly, most of these agents also affect intestinal motility.

Microbial Pathogens

Microbial pathogens including bacteria, viruses, and fungi can alter electrolyte transport, increase intestinal permeability, and trigger inflammation to elicit diarrhea. They do this by a variety of mechanisms, including attaching to epithelial cells to insert their own products and alter host cell machinery, and by elaborating enterotoxins, which may be cytotoxic or can capture cell-signaling mechanisms to elicit secretion or disrupt tight junctions (see Table 99-3; see also Chapter 107).104-108 A few examples are provided here. Shigella causes dysentery by release of Shiga cytotoxins, which enter the epithelial cell, inhibit protein synthesis, impair absorption, and damage the mucosa. In contrast, secretory diarrheas, such as those associated with cholera and traveler’s diarrhea, result from noninvasive pathogens, which elaborate enterotoxins that capture and turn on the secretory machinery of the epithelium. The archetypal enterotoxin-mediated diarrhea is cholera. Vibrio cholerae carry a virulence cassette that produces at least three different molecules: an enterotoxin; a zonula occludens toxin (ZOT), which disrupts tight junction permeability; and a channel-like protein.109 The enterotoxin causes an unregulated increase in cAMP, activates CFTR, and inhibits NHE3, thus resulting in copious fluid secretion (see Fig. 99-9 legend for details). Mice lacking CFTR do not respond to cholera toxin. Despite voluminous secretion, specific intestinal Na+-coupled nutrient absorptive (Na+-glucose, Na+amino acids) pathways are unaltered by the toxin, forming the physiologic basis for ORT. Bacteria such as Salmonella species, Campylobacter jejuni, and E. coli elaborate enterotoxins similar to cholera toxin, thereby employing the cAMP machinery to elicit fluid secretion. Strains of E. coli and Yersinia enterocolitica associated with traveler’s diarrhea elaborate small-molecular-weight, heat-stable enterotoxins, which increase cGMP to stimulate fluid secretion; signaling cascades distinct from cAMP and cGMP also activate CFTR and inhibit NHE3 (see the discussion of intracellular mediators). Vibrio parahaemolyticus elaborates a thermostable direct hemolysin (TDH) and is a major cause of gastroenteritis; its associated intestinal fluid secretion is attributed to an increase in intracellular calcium and activation of the Ca2+-calmodulin and protein kinase C signaling pathways.93,108 Rotavirus, the major cause of infantile gastroenteritis, induces watery diarrhea. The virus predominantly infects the mature enterocyte of the villus and elaborates an enterotoxin NSP4.110,111 NSP4 can inhibit brush border membrane disaccharidases and SGLT1 activity, thereby limiting Na+glucose and fluid absorption, with resultant diarrhea. NSP4 in vitro elicits Cl− secretion via a calcium-phospholipase C pathway that stimulates Ca2+-activated Cl− channels; this

1691

1692

Section X  Small and Large Intestine effects a transient secretion similar to the actions of carbachol. Unlike other secretagogues, NSP4 has no effect on crypt cell secretion, but under favorable electrochemical conditions it stimulates secretion from villus cells. Paradoxically, NSP4 also can stimulate Cl− absorption from villus cells. The role of NSP4 in vivo has yet to be defined. Many bacterial pathogens use different signaling molecules (e.g., kinases and phosphatases) to perturb the delicate balance of tight junctional proteins and cytoskeletal elements, thus disrupting intestinal permeability. Enteropathogenic (EPEC) and enterohemorrhagic (EHEC) E. coli decrease transepithelial resistance using different signaling cascades. The EPEC strains use a fascinating arsenal to alter host cell responses: They adhere to intestinal cells and in the process recruit a complex network of host cytoskeletal elements. Infectious bacteria use a variety of mechanisms, classified as types I to IV, to infect host cells. EPEC strains use a type III secretion apparatus to insert effector molecules into the host cell and use these molecules to co-opt the cell machinery to cause changes in the actin-myosin network, alter tight-junction proteins, and modulate ion transport processes to decrease absorption and increase fluid accumulation.105,107 The anaerobic bacterium Clostridium difficile, which causes antibiotic-associated pseudomembranous colitis, and Clostridium perfringens, which is associated with food-borne illnesses, alter intestinal permeability by using two distinct processes: The C. difficile toxins A and B interact with the Rho family of cellular proteins to disrupt the perijunctional actin-myosin ring, whereas the tight-junction claudin proteins serve as receptors for C. perfringens enterotoxin (CPE), and binding results in a disruption of tight-junctional fibrils. Other bacteria, including Bacteroides fragilis and V. cholerae, elaborate proteases that attack junctional proteins, including occludins, claudins, and cadherin, to disrupt the integrity of tight junctions.93,108 Many bacterial strains, including EPEC, EHEC, ETEC, Salmonella, and Shigella, trigger a highly specialized cascade to stimulate ion secretion. These pathogens induce the expression of receptors for the peptide neurohormone galanin, whereas uninfected cells do not possess galanin receptors. Galanin in turn activates Cl− secretion via Ca2+dependent signaling processes.112-114,117 The ever-expanding spectrum of toxin-induced mechanisms underscores the importance of delineating the intrinsic regulatory processes and the molecular pathophysiology of infectious diarrhea.

Bile Acids

An increase in colonic bile acids secondary to ileal malabsorption or oral supplementation can cause diarrhea (see Chapters 15 and 64).13,14 Only 7α dihydroxy bile acids such as chenodeoxycholic acid (3α; 7α), but not 7β, are associated with diarrhea. At high concentrations, bile acids act as a detergent and increase intestinal permeability. At more physiologic concentrations, bile salts indirectly increase cAMP and activate mast cells and, more importantly, stimulate epithelial cell Cl− secretion via the Ca2+ and PKCδ cascade.115,116 The ability of bile salts to stimulate Cl− secretion occurs only in the adult and is absent in the neonatal animal. Long-chain fatty acids are seen in increased concentration in the colonic lumen in conditions such as sprue, when long-chain triglycerides are digested by lipase but the fatty acids are malabsorbed within the small intestine (see Chapters 101 and 104). Hydroxylated fatty acids are more potent secretagogues than the corresponding long-chain fatty acids and arise from colonic bacterial metabolism; ricinoleic acid is the long-chain fatty acid that is derived from oral castor

oil, which is a nontoxic oil before it is hydroxylated. Specific fatty acid transporters have been identified in the intestine, and their mechanisms of action in electrolyte secretion are similar to those of bile acids.118,119

INTRACELLULAR MEDIATORS The barrage of extracellular stimuli need to be translated into an intracellular language so the cell can regulate its transport machinery. The second messenger cascades of the cell include the cyclic nucleotides cAMP and cGMP, intracellular Ca2+, and the inositol phosphate-diacyl glycerol and tyrosine kinases. These messenger systems are common to several organ systems, and many cell-specific and tissuespecific structural and functional nuances contribute to the net biological response. More detailed descriptions of second messenger systems can be found elsewhere, but an overview is provided here.120 Epithelial cells require rapid response cascades for turning on and turning off ion transport systems. For example, cyclases are poised to synthesize cyclic nucleotides, and phosphodiesterases are set to degrade them. The net biologic response is governed by the relative contributions of the accentuating and attenuating processes. In addition to the burgeoning increase in new signaling molecules awaiting definition of their roles, there are other variables compounding the resulting net biologic response. The molecules at almost every step in the signal transduction cascade, from the activating hormone to the receptors, cyclases, kinases, and phosphatases, and, finally, the transporters themselves, exist as multiple isoforms and variants (see Fig. 99-7). These isoforms exhibit differences in species, tissue, cell type, and subcellular distribution and are subject to regulation during their development as well as in response to routine physiologic demands. The cannonical cascade of stimulus → second messenger → kinase → response is an oversimplification; extensive crosstalk exists between different signaling pathways. Other post-translational modifications such as glycosylation, myristoylation, nitration, and sumoylation are increasingly recognized as important modulators. Within a signaling cascade there are critical feed-forward and feedback regulatory steps. Thus, protein kinases catalyze the transfer of the terminal phosphate from ATP to the hydroxyl group of a serine, threonine, or tyrosine of a target protein, leading to both conformational and functional changes, such as altered affinity for substrate. They exhibit specificity in their activators and substrates, and their action is essentially irreversible in living cells. Phosphoproteins can only be dephosphorylated by protein phosphatases, which are a separate class of enzymes that are also subject to regulation. Phosphoproteins may be the transporters themselves or may be modulator proteins in the membrane or cytosol, or they may be both. Equally important is the recognition that protein phosphorylation is not synonymous with activation; the dephosphorylated protein may be the active form. In general, Ca2+-specific and cyclic nucleotide-specific protein kinases are serine-threonine kinases, whereas tyrosinespecific protein kinases are associated with receptors of cytokines and hormones involved in growth, such as EGF. In addition to specific serine-threonine and tyrosine phosphatases, dual-specificity kinases and phosphatases add to the complexity of crosstalk. Compartmentalization of components of the signaling cascade via cytoskeletal runners, anchoring domains, or

Chapter 99  Intestinal Electrolyte Absorption and Secretion sequestration in vesicles as a means of regulation is especially germane to the polarized enterocyte. Localizing transporters and their signaling systems into specific subcellular domains is the norm rather than the exception. This is a dynamic and highly regulatable process. Scaffolding proteins can promote docking of various proteins, kinases, and phosphatases by protein-protein interactions—to each other, to the cytoskeleton, and/or to the membrane (see Fig. 99-7); for example, guanylate cyclase C and the intestinal protein kinase GII have cytoskeletal and membrane interacting domains that bring them in close proximity to CFTR in the brush border membrane (Fig. 99-7). Cholesterol-rich membrane domains such as lipid rafts influence membrane fluidity and anchor specific transporters and their regulators. It should come as no surprise that many anchoring domains serve as multienzyme signaling complexes. Finally, trafficking of transporters into and out of the membrane via endosomal vesicles is an effective way of rapidly altering the Vmax of the transporter. For example, cAMP increases CFTR translocation to the membrane and NHE3 retrieval from the membrane, resulting in an increase in Cl− secretion and decrease in Na+ absorption (see Fig. 99-9). Thus, the cAMP-stimulated increases in Cl− secretion and decreases in Na+ absorption (see Fig. 99-9), respectively, result from an increase in CFTR translocation to, and NHE-3 retrieval from, the membrane. Although it is not associated with rapid responses, neurohumoral stimulation can cause changes at the transcriptional level, leading to the synthesis of new proteins, such as aldosterone, which increases ENaC synthesis in the distal colon. Diseases such as cystic fibrosis underscore the importance of intracellular quality control in the cell machinery; thus ΔF508 CFTR, the most common mutation, is a misfolded protein and is tagged for degradation. All signaltransduction mechanisms need to be assessed with respect to their physiologic relevance in the intact intestine because reductionist models, although a necessity, do not provide the complete picture. With these caveats, some common themes have emerged. Generally, agents that elevate intracellular cAMP, cGMP, or Ca2+ increase fluid secretion (see Tables 99-2 and 99-3). They can activate one or more transporters associated with electrogenic Cl− secretion: apical Cl− and K+ channels, basolateral K+ conductances, and NKCC1; they also inhibit the apical NHEs, NHE2 and NHE3. Conversely, fluid absorption is associated with a decrease in these messengers or with an activation of some tyrosine kinase pathways. cAMP plays an additional role in promoting trafficking of transporter-bearing vesicles (CFTR in crypts and distal colonic Na+ channels) to the apical membrane.91,111 The cAMP cascade is triggered by a hormone, , such as VIP, binding to a specific member of the superfamily of heptahelical membrane-spanning receptors (7TM-VPAC1 and VPAC2) (see Fig. 99-7).121 Cyclic GMP is generated by the activation of membrane or soluble guanylate cyclases (GCs) by the natriuretic peptides guanylin and uroguanylin (see Fig. 99-7). Guanylins share their receptor GC-C with the heat-stable enterotoxins (see the earlier discussion of guanylin). Hormones and neurotransmitters such as substance P and acetylcholine activate secretion by increasing intracellular Ca2+ (Fig. 99-11; see figure legend for details). A number of growth factors, cytokines, and inflammatory mediators use entirely different signaling pathways, which involve a combination of extracellular regulated kinases (ERKs), dual-specificity kinases, receptor kinases, and receptor-associated tyrosine kinases.

Acetylcholine, neurotensin 1

Bile acids

Substance P Ca2+ 2 G-protein IP3

PLC 3

DAG

PIP2

G-protein

4 PKC

IP3R 5

Ca2+ Ca/CAM kinase

Ca2+ Calcium stores

Activity of membrane channels/transporters

Figure 99-11.  Calcium signaling in intestinal epithelial cells. Certain hormones and neurotransmitters (e.g., substance P, acetylcholine) activate secretion by increasing intracellular Ca2+. Substance P can stimulate Ca2+ channel activity, and acetylcholine binds to M3 muscarinic heptahelical membrane spanning (HHMS) receptors coupled to the Gaq class of G proteins (1). Activated Gaq stimulates phospholipase C-β (PLC) (2) to hydrolyze PIP2 to release DAG and inositol IP3 (3). DAG can also be produced from phosphatidic acid by the activation of phospholipase D by tyrosine kinase receptors. DAG is rapidly metabolized and does not increase intracellular Ca2+; its major action is to stimulate PKC, a family of phosphatidyl serine-dependent enzymes that have far-reaching biological actions (4). In contrast to DAG, IP3 binds to specific receptors to release Ca2+ from intracellular compartments (5). Intracellular free Ca2+ is tightly regulated and maintained at less than micromolar concentrations, in contrast to the 1-2 mmol/L in the plasma. Transient elevations in intracellular Ca2+ are sufficient to elicit a host of biological responses including ion transport. Calcium directly activates target proteins, such as Ca2+ channels, or binds to the ubiquitous Ca2+-binding protein calmodulin to activate specific calcium-calmodulin protein kinases. Ca2+-dependent secretagogues may be responsible for the minute-by-minute regulation needed in the intestine. This is underscored by the transient nature of Ca2+ signaling and its desensitization to Ca2+-dependent secretagogues. Ca2+ is rapidly resequestered by Ca2+-dependent adenosine triphosphatases on the endoplasmic reticulum or effluxed by Na+-Ca2+ exchange on the plasma membrane. The transient receptor potential channels allow the replenishment of intracellular Ca2+ from the extracellular compartment. PLC activation can concomitantly release polyinositol phosphates, such as inositol 3,4,5,6tetrakisphosphate, which function as ileal brakes and dampen Ca2+-induced Cl− secretion. Phenothiazines and loperamide can interfere with Ca2+ metabolism. CAM, calmodulin; DAG, diacylglycerol; IP3, inositol triphosphate; IP3R, inositol triphosphate receptor; PIP2, phosphatidyl inositol bisphosphate; PKC, protein kinase C; PLC, phospholipase C.

HOMOCELLULAR REGULATION Epithelia need to exhibit homocellular regulation. Given the vicissitudes in luminal content and osmolality, intestinal epithelial cells must be prepared for large and rapid changes in the rates of ion and nutrient transport. What enters the cell on one side must exit the cell at the other end at a similar rate. If not, the cell will either shrink or explode, owing to a rapid change in ionic content and osmolality.

1693

1694

Section X  Small and Large Intestine The concept of homocellular regulation is best exemplified by Na+ absorption. Thus, when Na+-glucose cotransport is maximally stimulated, the amount of Na+ flowing through the cell each minute is several orders of magnitude greater than the basal cell Na+ content and needs to be regulated. The rate-limiting step for transepithelial Na+ transport is entry across the apical membrane, for both channel transport and carrier-mediated transport. Changes in rates of Na+ entry initiate a series of coordinated changes in the Na+ pump, basolateral K+ conductances, and the apical entry mechanism itself such that the intracellular environment remains constant. Potential regulators of this dialog between membranes include cell volume, ATP, stretch-sensitive ion channels, and specialized intracellular pools of either Na+ or Ca2+. The ability of the cell to fine-tune discrete events at its opposite borders allows it to function effectively.

KEY REFERENCES

Alberts B, Johnson S, Lewis J, et al. Molecular Biology of the Cell. New York: Garland Science, Taylor & Francis Group; 2007. (Ref 1.) Alper SL. Molecular physiology of SLC4 anion exchangers. Exp Physiol 2006; 91:153-61. (Ref 61.) Barrett KE, Seely SJ. Integrative Physiology and Pathophysiology of Intestinal Electrolyte Transport. San Diego: Academic Press; 2006. pp 1931-51. (Ref 4.) Field M. Intestinal ion transport and the pathophysiology of diarrhea. J Clin Invest 2003; 111:931-43. (Ref 13.)

Gershon MD, Tack J. The serotonin signaling system: from basic understanding to drug development for functional GI disorders. Gastroenterology 2007; 132:397-414. (Ref 101.) Jentsch TJ, Neagoe I, Scheel O. ClC chloride channels and transporters. Curr Opin Neurobiol 2005; 15:319-25. (Ref 51.) Kunzelmann K, Mall M. Electrolyte transport in the mammalian colon: mechanisms and implications for disease. Physiol Rev 2002; 82:24589. (Ref 32.) Kunzelmann K, McMorran B. First encounter: how pathogens com­ promise epithelial transport. Physiology (Bethesda) 2004; 19:240-4. (Ref 108.) Schneeberger EE, Lynch RD. The tight junction: a multifunctional complex. Am J Physiol 2004; 286:C1213-28. (Ref 16.) Thiagarajah JR, Verkman AS. New drug targets for cholera therapy. Trends Pharmacol Sci 2005; 26:172-5. (Ref 109.) Rao, M. Absorption and secretion of water and electrolytes. In: Ratnaike R, editor. Small Bowel Disorders. London: Hodder Headline Group; 2000. pp 116-34. (Ref 120.) Rao MC. Oral rehydration therapy: new explanations for an old remedy. Ann Rev Physiol 2004; 66:385-417. (Ref 91.) Weber CR, Turner JR. Inflammatory bowel disease: is it really just another break in the wall? Gut 2007; 56:6-8. (Ref 82.) Wright EM, Loo DD, Hirayama BA, Turk E. Surprising versatility of Na+-glucose cotransporters: SLC5. Physiology (Bethesda) 2004; 19:370-76. (Ref 33.) Zachos NC, Tse M, Donowitz M. Molecular physiology of intestinal Na+/ H+ exchange. Annu Rev Physiol 2005; 67:411-43. (Ref 38.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

100 Digestion and Absorption of Nutrients and Vitamins James J. Farrell

CHAPTER OUTLINE Digestion and Absorption of Nutrients  1695 An Overview of the Digestive Process  1695 Fat  1698 Dietary Intake  1698 Digestion and Absorption  1699 Carbohydrate  1706 Dietary Intake  1706 Digestion and Absorption  1707 Protein  1712 Dietary Intake  1712 Proteins from Endogenous Sources  1712 Digestion and Absorption  1712 Vitamins  1717 Water-Soluble Vitamins  1717 Fat-Soluble Vitamins  1720 Minerals and Trace Elements  1722 Calcium  1722 Magnesium  1723

Most nutrients are absorbed with remarkable efficiency: Less than 5% of ingested carbohydrate, fat, and protein is excreted in the stool of adults consuming a normal diet.1 Even much of the indigestible dietary fiber is absorbed from the colon as short-chain fatty acids that are liberated by bacterial breakdown of fiber.2 The intestinal tract of neonates is less efficient: infants fail to absorb 10% to 15% of their dietary fat, and in premature infants as much as 25% to 35% of fat may be lost in the stool.3,4 In old age, nutrient absorption remains highly efficient unless the intestine becomes diseased. Despite considerable variations in types of food ingested and nutritional intake across national and racial groups, absorption remains efficient. Absorptive mechanisms adapt to the nature and amount of various nutrients presented to the intestinal tract. Such changes occur not only during early development5 but also throughout life and at times of specific need, such as during pregnancy.6 In achieving the overall objective of nutrient absorption, the different parts of the gastrointestinal tract act in a closely integrated and coordinated manner under the control of neural and humoral regulatory mechanisms. The understanding of intestinal digestion and absorption at a molecular level has improved our knowledge of the integration and coordination of these functions within the gastrointestinal tract. The pharmacokinetics and pharmaco-

Iron  1723 Trace Elements  1725 Adaptation to Changes in Need or Load  1726 Mucosal Hypertrophy  1726 Specific Reversible Adaptation  1727 Vitamins and Trace Elements  1728 Signaling for Intestinal Adaptation and Implications for Therapy  1729 The Neonatal Intestine  1729 Development and Adaptation of Nutrient Digestion and Absorption  1729 Developmental Changes  1729 Triglyceride Digestion  1730 Carbohydrate Digestion and Absorption  1732 Protein Digestion and Absorption  1732 Effects of Bariatric Surgery on Normal Digestion and Absorption  1732

dynamics of several key carbohydrate, fat, peptide, amino acid, vitamin, and nutrient transporters are increasingly understood. In this chapter, integration of intestinal function with the dietary intake, digestion, and absorption of major nutrients (fat, carbohydrate, and protein) and essential micronutrients (vitamins and trace elements) is discussed. The evolving genetic and molecular basis of these functions also is evaluated.

DIGESTION AND ABSORPTION OF NUTRIENTS AN OVERVIEW OF THE DIGESTIVE PROCESS

The cerebral phase of digestion, whether triggered by the sight, smell, or thought of food, initiates the digestive process. Salivary and gastric secretory responses to this type of stimulus are mediated via the autonomic nervous system, and there is modest stimulation of pancreaticobiliary secretion via the vagus nerve.7 The further stimulus of nutrients in the mouth and upper gastrointestinal tract markedly potentiates secretion by humoral and local neural mechanisms (see Chapter 1).8 The rapidity with which food is normally chewed and swallowed affords little time for significant oral digestion of nutrients; however, good mastication and mixing with

1695

1696

Section X  Small and Large Intestine saliva initiates digestion of starch by salivary amylase. In infants, digestion of fat is begun in the stomach by gastric lipase. Gastric acid would soon switch off these enzymes were it not for the buffering capacity of food that allows some digestion to continue. The optimal pH for gastric lipases is 4.5 to 6.0, and it has been suggested that a considerable portion of dietary triglyceride may be digested by these lipases.9,10 Protein digestion begins in the stomach when gastric pepsinogens are converted to pepsins by gastric acid. Pepsins become increasingly active as intraluminal pH falls and therefore the digestive action of pepsins on proteins is restricted to the stomach. During ingestion of food, the stomach may become distended, but intragastric pressure rises little because of neurally mediated receptive relaxation. The mechanisms by which subjects perceive satiety and, therefore, cease eating are complex and explained only partly by the sensation of fullness. Although dozens of enzymes and hormones are secreted by the gastrointestinal tract in response to intraluminal food, only a few are able to influence food intake directly. Satiety signals are relayed to the hindbrain, either indirectly via nerves such as the vagus nerve or else directly via the blood. Most factors that influence how much food is eaten during individual meals act by changing the sensitivity to satiety signals.11 Cholecystokinin (CCK), gastrin-releasing peptide, and apolipoprotein A-IV (apo A-IV) have all been implicated as messengers that transmit the satiety signal to the central nervous system (CNS).12,13,14 They potentiate each other’s actions, and a combination of these agents may participate in the satiety signal. The administration of exogenous CCK or other satiety signals causes smaller meals to be consumed, whereas blocking the action of endogenous CCK and other satiety signals allows larger meals to be consumed.11,15,16 Additional peptides, known as the anorectic peptides, including peptide YY (PYY), pancreatic polypeptide (PP), glucagon-like peptide 1 (GLP-1), and oxyntomodulin also have been shown to decrease appetite and promote satiety in animal and human models.17 Apo A-IV is a glycoprotein synthesized by the enterocytes of human intestine and the hypothalamus, especially the arcuate nucleus. Intestinal apo A-IV synthesis is markedly stimulated by fat absorption and does not appear to be mediated by the uptake or re-esterification of fatty acids to form triglycerides. The local formation of chylomicrons acts as a signal for the induction of intestinal apo A-IV synthesis. Intestinal apo A-IV synthesis is also enhanced by a factor from the ileum (probably peptide tyrosine-tyrosine [PYY]), as well as neuropeptide Y (NPY) and pancreatic polypeptide (PP).18 Inhibition of food intake by apo A-IV is mediated centrally. The stimulation of intestinal synthesis and secretion of apo A-IV by lipid absorption are rapid; thus apo A-IV plays a role in the short-term regulation of food intake. There also is evidence suggesting that apo A-IV may be involved in the long-term regulation of food intake and body weight, because it is influenced by leptin and insulin. Chronic ingestion of a high-fat diet blunts the intestinal and the hypothalamic apo A-IV response to lipid feeding.19 Hypothalamic apo A-IV level is reduced by food deprivation and restored by lipid feeding.20,21 Leptin, a hormone released from fat cells, is an important peripheral signal from fat stores that modulates food intake by acting on receptors in the arcuate nucleus and hypothalamus.22 Leptin deficiency and leptin receptor defects produce massive obesity. Only one gastrointestinal signal, ghrelin, has been shown to increase appetite.11 The major digestive processes are initiated in the duodenum. Delivery of chyme from the stomach is delicately

Particle size

Acid Osmolarity Fatty acids

Calorie content Peptide YY

Nutrient

GLP-1 GLP-2

Ileal brake

Figure 100-1.  Some factors that delay gastric emptying. Receptors for acid, osmolarity, fatty acids, and other nutrients in the duodenum signal gastric delay via neurohumoral mechanisms. Food particles larger than 2 mm in diameter (large circles) are rejected by the antrum. Nutrients in the ileum and colon also influence gastric emptying by the ileal brake mechanism (see Chapter 48). GLP, glucagon-like peptide; peptide YY, peptide tyrosine-tyrosine.

adjusted so that it enters the duodenum at a controlled rate, thus allowing efficient mixing with pancreaticobiliary secretions. Control of gastric emptying is thus critical to ensuring optimal digestion. The characteristics of gastric contents that determine the rate at which the stomach empties include their consistency, pH, osmolality, and lipid and calorie content (Fig. 100-1).23 The pylorus is selective in that it allows rapid passage of liquids while retaining solid particles with diameters of 2 mm or larger.24 Thus, large particles are retained and progressively reduced in size by the gastric mill, a process referred to as trituration. Trituration ensures that particles will be small enough to allow them reasonably close apposition to digestive enzymes once the nutrient is allowed to enter the duodenum. Meals of high viscosity empty more slowly than do those of low viscosity. Duodenal mucosal receptors for pH and osmolality trigger a delay in gastric emptying when the gastric effluent is acidic or hyper- or hypotonic.25,26 When duodenal luminal contents are neutralized by pancreaticobiliary bicarbonate

Chapter 100  Digestion and Absorption of Nutrients and Vitamins and osmolality is adjusted by water fluxes, gastric emptying is encouraged once more. This careful titration in the duodenal lumen ensures that nutrients are presented optimally to the pancreatic enzymes, which function best at neutral pH. The total calorie content of meals also controls gastric emptying rates; on average, the human stomach delivers about 150 kcal/hr to the duodenum.27 An increase in the size or energy density of a meal leads to a corresponding increase in the rate of delivery. Receptors for fatty acids, amino acids, and carbohydrates in the duodenal mucosa are involved in this response, which probably is mediated by both neural and humoral feedback mechanisms.28 Gastric emptying additionally is controlled by a mechanism involving the ileum and colon. If much nutrient escapes digestion and absorption in the jejunum, its presence in the ileum and colon delays gastrointestinal transit, and this again provides more time for digestion and absorption.29,30 This ileal brake probably is mediated by a neurohumoral mechanism, for which various neuro­ transmitters and hormones have been implicated including peptide YY and the glucagon-like peptides-1 and -2 (GLP-1 and GLP-2).31,32 The GLPs are synthesized in and cosecreted from enteroendocrine cells in the small and large intestine in response to luminal carbohydrate and fat. GLP-1 promotes efficient nutrient assimilation by decreasing appetite, slowing gastric empting and enhancing glucose-induced insulin secretion. GLP-2 is cosecreted with GLP-1and also regulates energy absorption by its effects on nutrient intake, gastric acid secretion, gastric emptying, and nutrient absorption. Circulating levels of GLP-1 and GLP-2 are low in the fasted state and increase rapidly following ingestion of nutrients.33-37 The gallbladder is stimulated to contract and the pancreas to secrete simultaneously in response to the presence of nutrients in the duodenal lumen. A range of nutrient receptors stimulates the release of CCK and secretin from mucosal endocrine cells into the portal circulation, and these are largely responsible for this response. Exocrine pancreatic secretion is primarily controlled by cephalic mechanisms (the vagus nerve), gastric mechanisms (acid and pepsin secretion, and nutrients delivered into the duodenum by gastric emptying), and intestinal mechanisms (secretin and CCK) (see Chapter 56). CCK and other enterohormones stimulate the pancreas by excitation of sensory nerves and by triggering of long vagovagal or enteropancreatic reflexes. Numerous neurotransmitters, such as acetylcholine and nitric oxide, and certain neuropeptides, such as gastrin-releasing peptide (GRP), generated by neurons of the enteric nervous system, have been implicated in the regulation of the exocrine pancreas. CCK appears to act via vagal cholinergic pathways to mediate pancreatic enzyme secretion. Human pancreatic acini lack functional CCK-A receptors, explaining why a CCK infusion that produces plasma CCK levels similar to those seen postprandially stimulates pancreatic exocrine secretion by an atropine-sensitive pathway.38 Under physiologic conditions, cholinergic vagal afferent pathways rather than pancreatic acinar cells represent the primary targets on which CCK can act as a major mediator of postprandial pancreatic secretion.38 Serotonin (5-hydroxytryptamine, 5-HT) released from enterochromaffin cells in the intestinal mucosa and nerve terminals of the enteric nervous system and the intrapancreatic nerves may be involved in both stimulation and inhibitory mechanisms through its various receptor subtypes; 5-HT also mediates the actions of secretin and CCK.

A synergistic interaction between CCK and 5-HT at the level of the nodose ganglia might explain the robust postprandial pancreatic secretion despite a modest postprandial increase in plasma CCK. Peptides affecting appetite and originating from the intestine (e.g., leptin and ghrelin) or from the pancreas (e.g., PP and NPY) appear to modulate the exocrine pancreas via hypothalamic centers.39,40 Pancreatic juice provides both positive and negative feedback regulation of pancreatic secretion through mediation of both secretin- and CCK-releasing peptides. Three CCKreleasing peptides have been purified: monitor peptide from pancreatic juice, diazepam-binding inhibitor from porcine intestine, and luminal CCK-releasing factor from rat intestinal secretion. All have been shown to stimulate CCK release and pancreatic enzyme secretion. Pancreatic phospholipase A2 from pancreatic juice and intestinal secretions appears to function as a secretin-releasing peptide.41 The simultaneous release of bile salts, pancreatic enzymes, and bicarbonate provides optimal conditions for further nutrient digestion. The simultaneous release of entero­ peptidase (enterokinase) from duodenal mucosa is critically important to the activation of pancreatic proteolytic enzymes. Enterokinase releases trypsin from trypsinogen, thus encouraging proteolysis within the duodenal lumen rather than inside the pancreatic duct. These three factors— bile, pancreatic enzymes, and enteropeptidase—remain separate until they are mixed in the intestinal lumen, which ensures that they become operative only at the site of nutrient delivery. Adequate lipid digestion is critically dependent on the presence of bile salts and pancreatic lipase and colipase at nearly neutral pH,42 whereas digestion of carbohydrate and protein depends on the combined actions of intraluminal secreted enzymes and then enzymes sited on the brush border membrane and within the intestinal mucosa. The close physical relationship, at the brush border, between the sites for terminal digestion of protein and carbohydrate and the active absorption of digestive products provides a very efficient mechanism for dealing with these nutrients. Two other simultaneous phenomena encourage efficient digestion and absorption. Ingestion of a meal stimulates salt and water secretion by the jejunal mucosa, and this maintains luminal contents in a sufficiently fluid state for proper mixing and digestion (see Chapter 99).43 The other phenomenon is the motor response of the intestine. After feeding, the characteristic repetitive pattern of motility that occurs during fasting is disrupted. Instead, an apparently disordered pattern is seen, which, presumably, ensures that nutrients are well mixed and brought into close contact with intestinal mucosa (see Chapter 97). There is close integration of the neurohumoral control mechanisms involving the motor and secretory responses of the intestine.44 For rapidly absorbed molecules, intestinal blood flow may be the ratelimiting step.45 Efficient conservation and recycling mechanisms ensure that gastrointestinal secretions are not entirely lost. Gastric acid secretion is balanced to a large extent by pancreaticobiliary bicarbonate secretion, so that overall acid-base balance is not disturbed. Although intact digestive enzymes are reabsorbed only in trace amounts, the nitrogen they contain is reabsorbed after their digestion. Finally, efficient enterohepatic circulation recycles bile salts several times each day so that they may be used approximately twice for each meal.46 Although bile salts are passively reabsorbed throughout the small intestine, most reach the terminal ileum, where they are reabsorbed via specific active absorptive mechanisms. Thus, bile salts remain in the lumen where they are needed for lipid digestion, but they are

1697

1698

Section X  Small and Large Intestine largely reabsorbed at the last moment to avoid being lost by the colon (Fig. 100-2; see Chapter 64). Once intestinal chyme leaves the ileum and enters the colon, most nutrients have been digested and absorbed and colonic function largely serves to dehydrate luminal contents through absorption of salt and water and to store the

residuum. Dietary fiber may be digested by bacteria, with release of short-chain fatty acids, which are avidly absorbed; however, short-chain fatty acids do not usually have much nutritional significance, other than in areas of the world where the major source of energy is a high-fiber diet. Shortchain fatty acids, however, are the major source of nutrition for the colonocytes.

FAT

Total pool size 5g

DIETARY INTAKE

In the United States, fat intakes for all age groups and both sexes rose from approximately 34% of total energy consumption in the 1930s to 40% to 42% in the late 1950s to mid-1960s, and then fell steadily to approximately 36% (90 to 100 g/day) in 1984. Saturated and monounsaturated fatty acid intakes fell from 18% to 20% of total energy consumption in the early 1950s to 12% to 13% of energy in 1984; polyunsaturated fatty acid intakes rose from 2% to 4% of energy to 7.5%.47 The majority of fatty acids present in dietary triglyceride are oleate and palmitate (18 : 1 and 16 : 0, respectively).48 In animal triglyceride, most of the fatty acids are long-chained (i.e., longer than 14 carbon chains) and saturated (Fig. 100-3). Polyunsaturated fatty acids such as linoleic and linolenic acid are derived from phospholipids of vegetable origin and, because they cannot be synthesized de novo, they are considered essential fatty acids (Table 100-1). The average range of phospholipid ingestion is between 2 and 8 g/day. The most commonly ingested phospholipid is phosphatidyl choline (lecithin), and the predominant fatty acids in phospholipid are linoleate and arachidonate (see Fig. 100-3). More phospholipid is found in the duode-

Portal vein

Figure 100-2.  Enterohepatic circulation of bile salts. Active transport in the ileum retrieves most bile salts, and the small fraction lost into the colon and eliminated in the feces is compensated for by fresh hepatic synthesis (see Chapter 64).

Glycerol

Ester bonds

Fatty acids

O 1CH 2

C

O

(CH2)n

CH3

(CH2)n

CH3

(CH2)n

CH3

O 2CH

O

C O

3CH 2

A

O

C

Triglyceride

Glycerol

Hydrophilic portion Choline (CH3)3

N

O (CH2)2

O

P

O

Ester bonds O

Fatty acids

CH2

O

C O

(CH2)n

CH3

CH

O

C

(CH2)n

CH3

CH2 Hydrophobic portion

O

B

Phospholipid

Figure 100-3.  General molecular structure of triglyceride (A) and phospholipid (phosphatidyl choline or lecithin) (B).

Chapter 100  Digestion and Absorption of Nutrients and Vitamins nal lumen (10 to 22 g/day) than is ingested, most of which is derived from endogenous sources, particularly bile. Cholesterol intake varies widely but averages about 200 to 250 mg/day.48 Some people consume as much as 500 mg/day. Commercial hydrogenation of unsaturated bonds in the fatty acids of natural oils raises their melting points, thus allowing production of margarines and spreads of variable consistency. Hydrogenation, in addition to saturation, results in isomerization of cis to trans double bonds.49 Although many commercial products contain partially hydrogenated fats, the content of trans-fatty acids in some margarines exceeds 60%, thus raising concerns about their relationship to cancer induction.50

DIGESTION AND ABSORPTION

Most dietary lipid is absorbed by the upper two thirds of jejunum, although its rate and extent of absorption are influenced by the presence of other foods, particularly dietary fiber, which reduces the rate of absorption.51 The types of ingested fat also appear to influence the absorptive process, both by modifying the morphologic structure of the intestinal mucosa and by influencing its absorptive function for other nutrients such as carbohydrate.52 The insolubility of fat in water dominates the mechanisms that have evolved to digest and absorb lipid. Within the lumen, ingested fat has to be physically released and broken down into emulsion droplets. Following digestion, the products have to be transported across the bulk (lumen) water phase to the lipid epithelial cell membrane. Transfer across the lipid membrane is followed, within the epithelium, by reconstitution into larger lipid molecules, pre­ dominantly triglyceride, which then require specialized processing to permit export from the cell. Thus, to be used

Table 100-1  Common Dietary Fatty Acids FATTY ACID

CONFIGURATION*

Saturated Fatty Acids Butyric Caproic Lauric Myristic Palmitic Stearic Mono-unsaturated Fatty Acids Oleic Palmitoleic Polyunsaturated Fatty Acids Arachidonic Linoleic Linolenic

4 : 0 6 : 0 12 : 0 14 : 0 16 : 0 18 : 0 18 : 1 16 : 1 20 : 4 18 : 2 18 : 3

*By convention, the number of carbon atoms in the chain is given by the first figure and the number of double bonds in the chain is given by the second.

after ingestion, lipid must pass through three physical phases: water in the lumen, lipid in the epithelial membrane, and water in the lymphatics and bloodstream. Despite these potential barriers, more than 95% of ingested fat is absorbed by adults.

Triglyceride

Liberation of fatty acids from the glycerol backbone of triglycerides (lipolysis) is achieved by lipases acting at the surface of emulsified droplets (Table 100-2). This process occurs initially in the stomach, but most lipolysis is accomplished in the small intestine.53 Intragastric lipolysis might account for 20% to 30% of total intraluminal lipid digestion.54 Gastric lipase, which is of fundic origin, has been demonstrated in the gastric contents of premature neonates and in mucosal biopsy specimens from adults up to 80 years of age. Gastric lipase does not hydrolyze phospholipids or cholesterol esters. Human gastric lipase is a 379-amino acid protein that shares similar homology with rat gastric lipase but not human pancreatic lipase. For either gastric or small intestinal lipolysis to occur, two conditions are critical: First, a stable emulsion is required of fat droplets of such a size that they present a large surface area to the digestive enzyme; second, a mechanism is required for bringing enzyme and triglyceride into close apposition within the emulsion. Emulsification A number of factors assist in optimal production of an emulsion. Physical release of fat by mastication and the gastric milling of food produces a relatively unstable emulsion that is delivered into the duodenum. To permit its stabilization, the droplets in this emulsion have to be coated, and dietary phospholipid provides one such coat. The ratio of ingested phospholipid to triglyceride is about 1 : 30, and more phospholipid is added in the duodenum from bile.55 In breast milk, emulsion droplets are smaller and have proteins as well as phospholipid incorporated into their surface trilayer.56 Emulsification also is enhanced by the fatty acids liberated by intragastric lipolysis and, within the duodenum, by bile salts (Fig. 100-4). The final product in the duodenum is an emulsion consisting predominantly of triglyceride together with cholesterol esters and some diglyceride and coated by phospholipid, partially ionized fatty acids, monoglyceride, and bile salts. Lipase This stable emulsion is then presented to pancreatic triglyceride lipase. Unlike other soluble enzymes, which can act in a three-dimensional solution, lipase has to act at the two-dimensional surface of the emulsion droplet, and this requirement poses particular problems.57 Certain characteristics of the enzyme itself are important. Thus, the lipolytic zone of the molecule is hydrophobic and lies deep within it, shielded from the aqueous phase. It is revealed to the lipid only on close apposition to its surface. The presence of a coat on the lipid droplet thus poses a barrier to the

Table 100-2  Characteristics of Lipase Activity in Infancy ENZYME

OPTIMAL pH

site of lipase activity

OTHER

Milk-derived lipase Gastric lipase Pancreatic lipase

7.0 (inactivation by acid is reversible) 4.0-6.0 7.0

a-1, -2, and -3 ester bonds a-1 ester bond a-1 and -3 ester bonds

Stimulated by bile salts Inhibited by pancreatic proteolysis —

1699

1700

Section X  Small and Large Intestine Panel B

Product vesicles

Stomach

Emulsification Increased stability of lipid emulsion

TG PL

TG (Oil droplet)

Gastric lipolysis

Mixed micelles

TG PL DG Duodenal lipolysis

MG FA

Duodenum

Aqueous phase MG

Bile salts Lecithin

Lipase

PL

Colipase

BS Panel A

MG Lipid phase

TG

BS

DG Lipid phase

Aqueous phase Panel C FA 1 2 3

+ TG

MG

FA

–

Long section

Cross section

Lipase + HCO3

Figure 100-4.  Steps in lipolysis. The initial step in lipolysis is to increase the stability of the fatty emulsion. Gastric lipase acts on triglycerides to yield fatty acids and diglyceride (diglyceride enhances emulsification). This step is enhanced in the duodenum by bile salts and phospholipid (lecithin), which enable lipase, in the presence of colipase, to act at the surface of the emulsion droplet to bring it close to the triglyceride molecule, whereupon monoglyceride and fatty acids are released. Lipolysis in the duodenum yields fatty acids (from the α1 and α3 positions) and monoglyceride and occurs in a rapid and efficient manner at nearly neutral pH. In panel A are diagrammatic representations of bile salt molecules (top) oriented at an oil-water interface with its hydrophobic sterolic backbone in the oil phase and its hydrophylic hydroxyl and either taurine or glycine conjugates in the aqueous phase. At concentrations above critical micellar concentration, bile salts aggregate as simple micelles in water, with their hydrophylic groups facing into the water. In this diagram, three hydroxyl groups (cholate) are shown as open circles and an additional polar group represents either taurine or glycine. Panel B is a diagrammatic sketch of the dispersion of the products of lipolysis into lamellae at the surface of the oil phase, each about 4 to 5 nm thick, with water spacings up to 8 nm, and from there into vesicles of about 20 to 130 nm in diameter. In panel C, fatty acids and monoglyceride within the vesicles pass into mixed micelles. BS, bile salt; DG, diglyceride; FA, fatty acid; MG, monoglyceride; PL, phospholipid; TG, triglyceride.

action of lipase, and assistance is required to bring it into close contact with the triglyceride. The presence of colipase, cosecreted by the pancreas along with lipase in a molar ratio of 1 : 1, is critical in approximating lipase to triglyceride (see Fig. 100-4). Colipase attaches to the ester bond region of the triglyceride, lipase then binding strongly to colipase by electrostatic interactions.55 Phospholipase A2 digestion of the phospholipid on the surface of the lipid emulsion allows exposure of the triglyceride core to the colipase-lipase complex, further enhancing colipase-dependent anchoring of lipase to the lipid emulsion. Phospholipase A2 digestion requires

bile salts and Ca2+ for activation, which can further assist colipase-lipase–mediated triglyceride lipolysis by providing a mechanism for removing lipolytic products. In the absence of colipase, bile salts on the surface of the emulsion droplet inhibit lipase activity. The colipase gene is located on chromosome 6, and the amino acid sequence of the lipid-binding domain, the lipase-binding domain, and the activation peptide appear to be highly conserved.58 Colipase is secreted by the pancreas as pro-colipase,59 which is activated when trypsin cleaves a pentapeptide from its N-terminus after entering the small intestinal lumen. A valine residue at position 407 and a

Chapter 100  Digestion and Absorption of Nutrients and Vitamins leucine at position 412 are important for the interaction of lipase with colipase and the bile salt micelles.60 Interestingly, the pentapeptide cleaved from the pro-colipase by trypsin, called enterostatin, seems to be a specific satiety signal for the ingestion of fat.61 Because pancreatic lipase is most active at nearly neutral pH, secretion of bicarbonate by the pancreas and biliary tree is critically important and provides the necessary neutralization of gastric acid; however, luminal pH falls to about 6 in the jejunum, and here the fact that bile salts lower the pH optimal for lipase activity from 8 to 6 may be significant. In the presence of colipase and optimal pH, lipase activity releases fatty acids and monoglyceride extremely rapidly and efficiently (see Fig. 100-4). Pancreatic triglyceride lipase also binds strongly to the mucosal brush border membrane,62 where it may participate in lipolysis of cholesteryl esters or triglyceride, releasing fatty acids, monoglyceride, and free cholesterol in proximity to the brush border membrane, where they undergo rapid uptake. In addition to pancreatic triglyceride lipase and its protein cofactor, colipase, pancreatic acini also synthesize two pancreatic lipase-related proteins (PLRP-1 and PLRP-2), which have strong nucleotide and amino acid sequence homology to pancreatic triglyceride lipase. Although PLRP-1 has no known activity, PLRP-2 does have lipase activity and, like pancreatic triglyceride lipase, PLRP-2 cleaves triglycerides but with broader substrate specificity. PLRP-2 also hydrolyzes phospholipids and galactolipids, two fats that are not substrates for pancreatic triglyceride lipase. It is also different from pancreatic triglyceride lipase with respect to sensitivity to bile salts and in response to colipase. A further critical difference is that PLRP-2 mRNA appears before birth and persists into adulthood, whereas pancreatic triglyceride lipase mRNA first appears at the suckling-to-weaning transition. This suggests that PLRP-2 plays a critical role in the digestion of breast-milk fats.63,64 Micelles and Other Lipid-Containing Particles The products of lipolysis are distributed among the aqueous, oil, and intermediate phases in a number of forms prepared for transfer across the lumen to the mucosal brush border membrane. The shuttling of these products depends, in part, on the formation of micelles with bile salts. The concentration of bile salts secreted in bile is about 35 mmol and, in the duodenum, this is further decreased by dilution to 10 to 20 mmol; this concentration lies well above the critical concentration for micelle formation. Mixed micelle production depends on a number of other factors, including pH, presence or absence of lipids, and the types of bile salts that are secreted (see Chapter 64).42 Bile salts are capable of forming micelles because they have a particular three-dimensional structure and they are amphipathic; that is, their molecules have both watersoluble and lipid-soluble portions (see Fig. 100-4). They orient themselves at an oil-and-water interface and thus are ideal emulsifying agents. In addition, micelles are formed when bile salt levels are present above critical concentrations and thus are able to aggregate in disk-like particles with their hydrophobic sterolic backbones oriented toward each other and their hydrophilic polar groups facing outward into the aqueous phase. Bile salt micelles have the capacity to dissolve fatty acids, monoglycerides, and cholesterol, but not triglyceride.65 The mixed micelles thus formed are arranged so that the insoluble lipid is surrounded by bile salts that are oriented with their hydrophilic groups facing outward. Mixed micelles are about 50 to 80 nm in diameter and, unlike emulsion droplets, are too small to

scatter light; thus, micellar solutions are clear. The presence of phospholipid secreted in bile enlarges mixed micelles and makes them more efficient in the dissolution of fat. Other lipid-containing particles participate in the transfer of lipid to the mucosa. As the emulsion droplet shrinks during lipolysis, liquid crystalline structures are formed at its surface.66,67 These vesicular structures with multilamellar and unilamellar forms can be seen under the electron microscope, budding off the surface of emulsion droplets and occasionally close to the brush border membrane of the intestinal mucosa.68 This physical phase of lipid within the lumen might provide a significant mechanism for transfer of lipid to the mucosa, beyond that provided by bile salt micelles, and it could explain the observation that in the absence of bile salts, some 50% or more of dietary triglyceride may be absorbed. In the presence of adequate concentrations of bile salts, however, these vesicles undergo rapid spontaneous dissolution and release their lipid into micelles, which are likely to be the major route for lipid traffic (see Fig. 100-4); numerically, they are much more common than lipid vesicles. Importance of Intraluminal pH Lipid digestion and absorption are highly dependent on intraluminal pH at several steps in the process. Pancreatic lipase operates best in the presence of bile salts and at least pH 6. It therefore functions well at the pH of the luminal duodenum, where most lipid digestion occurs. Glycineconjugated bile salts precipitate below pH 5; fatty acids are in their protonated form below about pH 6 and have limited solubility in bile salt micelles. Thus, in conditions in which intraluminal pH becomes more acid, as for example in the Zollinger-Ellison syndrome, pancreatic lipase is inactive, bile acids precipitate out of solution, and fatty acid partitioning is reduced. It is not surprising, therefore, that steatorrhea (without any other nutrient or hematologic disturbances) is a feature of this syndrome. Biological characteristics of lipases, including effect of pH on activities, are detailed in Table 100-2. Unstirred Water Layer An unstirred water layer is present on the surface of the intestinal epithelium, which in humans is approximately 40  µm deep.69 This layer may be rate limiting for uptake of long-chain fatty acids but not for short- or mediumchain fatty acids, the limiting step for which occurs at the brush border membrane.55 The provision of a high concentration of fatty acid in the microenvironment adjacent to the epithelium depends on the diffusion of micelles into this region. The microclimate here is slightly acidic, owing to activity of a sodium-hydrogen (Na+/H+) exchanger at the brush border membrane, and at pH between 5 and 6, the solubility of fatty acids in micelles decreases, thus encouraging liberation of fatty acids close to the mucosa. The high concentration of fatty acids necessary for diffusion across the mucosal membrane is thus achieved; evidence for this model is increasingly persuasive.70 The low-microclimate pH also encourages the fatty acids to be presented in an undissociated, protonated form. Thus, the pH partition hypothesis predicts that fatty acids could diffuse passively into the cell as protonated species and, at the near-neutral intracellular pH, become trapped in the ionized form. A surfactant-like material has been discovered close to the brush border membrane, although its role in absorption, if any, is uncertain.71 It is secreted by enterocytes, contains phosphatidylcholine and alkaline phosphatase, and appears

1701

1702

Section X  Small and Large Intestine as flat lamellae or vesicles adjacent to the brush border membrane.

Other Lipids

Phosphatidylcholine, the major dietary phospholipid, is hydrolyzed by pancreatic phospholipase A2 (PLA2) to yield fatty acid from the 2-position and lysophosphatidylcholine. Pancreatic PLA2 is secreted as an anionic zymogen that is activated in the small intestine by tryptic cleavage of an N-terminal heptapeptide. It has a molecular weight of approximately 14 kd and requires calcium for activation and bile salts for its activity. It has multiple isoforms and apparently requires a 2 : 1 bile salt-to-phosphatidylcholine molar ratio for optimal activity. Although the bulk of intestinal PLA2 activity is derived from pancreatic juice, there is some contribution from the intestinal mucosa, where the enzyme is concentrated in the brush border.72 Cholesterol esters, in the presence of bile salts and calcium, are hydrolyzed by carboxyl ester lipase (CEL) (also known as pancreatic cholesterol esterase) to release the free sterol, in which form it is absorbed. Cell culture and other in vitro studies have thoroughly defined the potential functions of CEL in the digestion of cholesteryl ester, phospholipids, and triglycerides, but only its cholesteryl ester hydrolytic activity is unique to this enzyme in the digestive tract.73 CEL belongs to the α/β hydrolase family, is well conserved, and shares 78% homology in rats and humans.74 It is secreted primarily by the pancreatic acinar cells and lactating mammary glands. Using site-directed mutagenesis, the serine at position 194, the histidine at position 435, and the aspartic acid at position 320 have been shown to be important for CEL’s catalytic activity.73,75-77 The hydrolysis of water-insoluble substrates by CEL requires bile salt–containing 3α- or 7-α-hydroxy groups (e.g., cholate or chenodeoxycholate and their conjugates).78 The arginine-63 and arginine-423 sites play an important role in this bile salt–dependent process, but not in the bile salt– independent lysophospholipid hydrolytic activity of CEL.79 Both hydrolytic enzymes act on the emulsion phase at the surface of droplets, and the products of digestion are released into multilamellar and unilamellar vesicles and then to mixed micelles. Fatty acids and monoglycerides increase the solubility of cholesterol in micelles, thus encouraging its absorption. The products of phospholipid and cholesterol hydrolysis thus follow the same route to the brush border membrane as the fatty acids and monoglyceride, which originate from dietary triglyceride. Unabsorbed long-chain fatty acids that enter the colon are not absorbed by this organ, and they undergo a series of bacterial modifications, principally hydroxylation. In healthy persons, no undigested triglyceride is found in the stool, and the normal fecal fat estimate of approximately 7 g/day reflects the cumulative total excretion of saponification products (i.e., fatty acids) that arise principally from membrane phospholipid and bacteria.

Transfer across the Brush Border Membrane

Much of the current understanding of the micellar solubilization and uptake of dietary lipids comes from the work of Hofmann and Borgstrom, who described the uptake of lipid digestion products by enterocytes.80 Further work by Carey discovered the coexistence of unilamellar liposomes with bile salt–lipid mixed micelles in the small intestine.81 Although the uptake of lipid digestion products by enterocytes has been accepted as a passive process, recent work has raised the possibility that some lipids may be taken up by enterocytes via carrier-mediated processes that are energy dependent.82

Studies with brush border membrane vesicles suggest that linoleic acid uptake occurs by facilitated diffusion.83 Absorption of oleic and arachidonic acid also appears to occur by a saturable process, suggesting the possibility of active transport. Several membrane proteins that increase the uptake of long chain fatty acids when overexpressed in cultured mammalian cells have been identified, the most prominent and best characterized of which are FAT/CD36, long-chain fatty acyl-CoA synthetases (LACS), and fatty acid transport proteins (FATPs/solute carrier family 27).84-87 The FATPs are transmembrane proteins that have been shown to enhance the cellular uptake of long-chain and very-long-chain fatty acids. In humans, FATPs comprise a family of six highly homologous proteins, hsFATP 1-6, which are found in all tissues of the body that use fatty acids.84,88,89 Although hsFATP1 is the best characterized of the FATPs, hsFATP4 is the only FATP expressed in the small intestine; it is localized to the apical brush border of the epithelial cells, where it is responsible for absorption of dietary lipids. Studies with cell lines and isolated enterocytes that overexpress FATP4 demonstrated that FATP4 is both necessary and sufficient for efficient uptake of longchain and very-long-chain fatty acids.90 Detailed substrate studies based on 14C-labeled fatty acids have been presented for FATP1 and FATP4.87,90 Both studies showed that uptake of fatty acids shorter than 10 carbon atoms, such as butyric acid and octanoic acid, was unaffected by FATP expression, whereas uptake of common long-chain fatty acids, such as palmitate and oleate, was robustly enhanced.90 More recently, a wrinkle-free phenotype has been associated with the spontaneous autosomal recessive mutation of the gene for FATP4, resulting in a very tight and thickskinned phenotype.91 Nutrients, hormones, and cytokines have been reported to regulate FATP expression. Rats fed a high-fat diet showed increased FATP expression in the heart, but not the liver. Several reports have shown a positive regulation of mouse FATP by ligands that activate either PPAR-γ, PPAR-α, or PPAR-γ/RXR heterodimers in hepatoma cell lines, the liver, and the intestine. Further, a PPAR binding site was identified in the murine FATP1 promoter. TNF-α is a negative regulator of FATP expression and down-regulates FATP mRNA in liver and FATP1 and FATP4 proteins in adipocytes.88 The exact mechanism of FATP transport of long-chain fatty acids into the intestinal cell is unknown. It has been postulated that extracellular long-chain fatty acids might directly bind to FATP complexes and be transported into cells. Alternatively, long-chain fatty acids could bind first to CD36, which hands off the long-chain fatty acids to FATP dimers. Intracellular long-chain fatty acids are coupled to CoA by long-chain acyl-CoA synthetase (LACS), preventing their efflux, and fatty acid–binding proteins (FABPs) act as a cytoplasmic buffer for incorporated long-chain fatty acids (Fig. 100-5).88 Cholesterol, unlike β-sitosterol (plant sterol), is well absorbed by the proximal jejunum,92 although both are present in the human diet. The second-order kinetics of cholesterol absorption, its sterol specificity, and its inhibition by drugs such as ezetimibe all suggest that cholesterol absorption is mediated by specific transport proteins at the brush border membrane.93,94 Thurnhofer and colleagues first described the presence of a possible binding protein in the small intestinal brush border that facilitates the uptake of cholesterol by the small intestine95; however, this 14K protein was later identified as sterol carrier protein-2 (SCP2), which is an intracellular protein.96 The adenosine triphosphate (ATP)-binding cassette (ABC) A1 transporter and

Chapter 100  Digestion and Absorption of Nutrients and Vitamins LCFA

FATP CD36

Acyl CoA LACS

they are coregulated by the nuclear hormone receptor’s liver X receptor (LXR)105; however, these transporters provide the apparatus for efficient shunting of sterols away from the transfer pathway directing the production of cholesteryl esters by acyl-CoA cholesterol acyl transferase 2 (ACAT2) and do not appear to be involved in the initial uptake of cholesterol. Niemann-Pick C1-like 1 (NPC1L1) protein has been suggested as a key component of intestinal cholesterol transport. Niemann-Pick C1 (NPC1) is the defective gene in the cholesterol storage disease Niemann-Pick type C (NP-C); NPC1 protein is highly abundant in a variety of tissues, and functions in intracellular cholesterol trafficking.106,107 In contrast, NPC1L1 protein has 50% amino acid homology with NPC1 protein, has several predicted features of a plasma membrane–expressed transporter, and is expressed in high abundance in the small intestine, especially in the brush border membrane of enterocytes. NPC1L1 protein has been suggested as a target for the cholesterol absorption– reducing drug ezetimibe.108 NPC1L1-deficient mice show a substantial reduction in absorbed cholesterol that is unaffected by dietary supplementation of bile acids.109

Intracellular Processing and Absortion

ACBP

FABP Figure 100-5.  Mechanisms of long-chain fatty acid (LCFA) and verylong-chain fatty acid transport. Extracellular LCFA may directly bind to a fatty acid transport protein (FATP) dimer complex and be transported into the cell. Alternatively, LCFA may bind first to CD36, which transfers the LCFA to a FATP dimer. Intracellular LCFA is coupled to coenzyme A (CoA) by long-chain fatty acyl-CoA synthetase (LACS), preventing their efflux, and fatty acid binding protein (FABP) acts as a cytoplasmic buffer for LCFA. ACBP, acyl CoA binding protein. (Adapted, with permission, from Stahl A. A current review of fatty acid transport proteins (SLC27). Pflugers Arch 2004; 447:722-7.)

scavenger receptor type B1 (SR-B1) also were postulated to play a role in cholesterol absorption. Evidence supporting the role of scavenger receptors in cholesterol uptake includes decreased cholesterol uptake by use of scavenger receptor inhibitors such as ezetimibe,97,98 but targeted inactivation of these genes in mice had no effect on cholesterol uptake.99-101 CD36 also has been demonstrated to be important in promoting the uptake of cholesterol by the small intestine and to play an important role in packaging absorbed fatty acids as triglycerides.102 Evidence favoring a cholesterol membrane transporter also is seen in individuals with β-sitosterolemia, a condition in which the intestine fails to discriminate between cholesterol and β-sitosterol. Major findings in patients homozygous for sitosterolemia include xanthomatosis and accelerated, often fatal, premature atherosclerosis. The genetic defect of β-sitosterolemia is linked to chromosome 2p21.103 Seven different mutations in two adjacent genes have been described that are responsible for encoding new members of the ABC transporter family (ABCG5 and ABCG8) in patients with sitosterolemia. These defects cause an increased intestinal absorption and decreased biliary excretion of all sterols (plant sterols and cholesterol), leading to a 50- to 200-fold increase in plasma plant sterol concentrations. Feeding cholesterol to mice up-regulated these genes, suggesting that ABCG5 and ABCG8 work together to limit intestinal cholesterol absorption by cholesterol efflux from small intestinal epithelial cells.104 These two genes are expressed almost exclusively in the liver and intestine, and

Once within the cell, fatty acids bind to specific FABPs, which are found predominantly in the jejunum and more in villus cells than in crypt cells. The small intestine has three distinct proteins belonging to the intracellular lipidbinding protein family: the liver-type FABP (L-FABP), the intestinal FABP (I-FABP), and the ileal lipid-binding protein (ILBP)110,111; all have greater affinity for unsaturated fatty acids than for saturated ones and very little affinity, if any, for short-chain or medium-chain fatty acids.42 Based on nuclear magnetic resonance (NMR) binding studies, it has been suggested that the binding of I-FABP is involved in the intracellular transport of fatty acids, whereas the L-FABP is involved in the intracellular transport of monoglycerides and lysophosphatidylcholine.112 These binding proteins can assist transfer across the cytoplasm to the endoplasmic reticulum for triglyceride resynthesis as well as modulating intracellular lipid metabolism and regulating gene expression. In addition, two sterol carrier proteins, SCP-1 and SCP-2, have been isolated and characterized. SCP-1 is important in the microsomal conversion of squalene to lanosterol,113 whereas SCP-2 participates in the microsomal conversion of lanosterol to cholesterol as well as the intracellular transport of cholesterol from cytoplasmic lipid droplets to mitochondria.114 In the endoplasmic reticulum, during feeding, triglyceride is resynthesized by two processes (Fig. 100-6).115 In the first, monoglyceride is re-esterified with absorbed fatty acid after it has been activated to form acyl coenzyme A (CoA) (the monoglyceride pathway). Microsomal acyl CoA-ligase is necessary to synthesize acyl CoA from the fatty acid before esterification. Diglyceride and then triglyceride are formed sequentially in reactions that favor long-chain fatty acid absorption from the lumen. This route, involving monoglyceride esterification, accounts for the majority of the triglyceride synthesized during the absorptive phase, no more than 4% being formed by acylation of absorbed glycerol. It is thought that the synthesis of triglyceride from diglyceride is catalyzed by the enzyme acyl CoA:diglyceride acyl transferase.116 The gene for this enzyme has been isolated, although a knockout mouse model of this gene still synthesized triglyceride in the intestinal mucosa, suggesting that there may be yet another enzyme involved in forming triglyceride from diglyceride.117,118

1703

1704

Section X  Small and Large Intestine Glucose

Fasting

α-Glycerophosphate Phosphatidic acid

ATP Fatty acid

Fatty acid + CoA

Lipoproteins Triglyceride

Acyl CoA Diglyceride

Monoglyceride

Phospholipid

Triglyceride

Monoglyceride Feeding

Figure 100-6.  Metabolic fate of absorbed fatty acid and monoglyceride in enterocytes. During feeding, triglyceride is resynthesized largely from absorbed fatty acid and monoglyceride. During fasting, triglyceride and phospholipid are synthesized from α-glycerophosphate derived from glucose entering across the basolateral membrane of the enterocyte, and from fatty acids. Unsaturated fatty acids tend to form the phospholipid. ATP, adenosine triphosphate; CoA, coenzyme A.

During fasting, triglyceride (and phospholipid) is synthesized via the second route, which involves acylation of α-glycerophosphate with the formation of phosphatidic acid and, then, triglyceride or phospholipid (see Fig. 100-6). The α-glycerophosphate is synthesized largely in the cytoplasm, from glucose. The relative importance of the monoglyceride-pathway and the α-glycerophosphate pathway depends on the availability of 2-monoacylglycerol and fatty acid. During normal lipid absorption, when 2-monoacylglycerol is sufficiently present, the monoglyceride pathway facilitates the conversion of 2-monoacylglycerol and fatty acid to form triglyceride and aids in inhibiting the α-glycerophosphate pathway. Conversely, when the supply of 2-monoacylglycerol is lacking or insufficient, the α-glycerophosphate pathway becomes the major pathway for forming triglyceride. Some absorbed lysophosphatidylcholine is reacylated to form phosphatidylcholine. The remaining absorbed lysophosphatidylcholine is hydrolyzed to form glycero-3phosphorylcholine. The liberated fatty acids are used for triglyceride synthesis, whereas the glycero-3-phosphorylcholine is readily transported via the portal blood for use in the liver. Absorbed dietary cholesterol enters a free cholesterol pool within enterocytes that also contains cholesterol from endogenous sites (nondietary sources such as biliary cholesterol, cholesterol derived from plasma lipoproteins, and cholesterol synthesized de novo). Cholesterol is transported mainly as esterified cholesterol and almost exclusively by the lymphatic system. Cholesterol esterase and ACAT are thought to be predominantly responsible for cholesterol esterification. ACAT is stimulated by a high-cholesterol diet and appears to play a more important role in mucosal cholesterol esterification than cholesterol esterase does.119 Two ACAT proteins have been identified: ACAT-1 and ACAT-2.120,121 The role of ACAT-2 in intestinal cholesterol absorption is supported by resistance to diet-induced hypercholesterolemia due to defective cholesterol esterification and absorption by the small intestine in the ACAT-2 knockout mouse model.122 Once synthesized, triglyceride, cholesterol and its esters, and phospholipids are packaged for export in the form of chylomicrons and very-low-density lipoproteins (VLDLs). During fasting, VLDLs are the major triglyceride-rich lipoproteins that emerge from the epithelium; after feeding,

chylomicrons predominate. VLDL triglycerides have a different fatty acid composition from those in chylomicrons, different pathways being involved in their formation. Furthermore, the fatty acids derived from dietary tri­ glyceride go predominantly into forming chylomicrons, whereas those derived from phospholipid appear to be used in forming VLDL.55 The diameter of chylomicrons ranges between 750 and 6000 nm; their cores comprise triglycerides, and cholesterol ester and phospholipid form more than 80% of the surface coat. Forming a smaller portion of the surface of chylomicrons is an essential component, apolipoprotein. Apo A is an important apoprotein for all lipoproteins, including chylomicrons, VLDLs, and high-density lipoproteins (HDLs). It is synthesized in the small intestine and is found in bile.123 Apo B probably is synthesized in the Golgi cysterni and is found in the rough endoplasmic reticulum. After feeding, Apo B is found in association with the chylomicrons in the smooth endoplasmic reticulum. The absence of apo B prevents synthesis and secretion of chylomicrons; however, data suggest that the supply of apo B is not the rate-limiting step for chylomicron formation. For example, the apo B output in lymph does not change after intraduodenal infusion of lipid, even though lymphatic triglyceride output increases seven-fold to eight-fold.124,125 Apolipoproteinemia is a rare genetic disorder resulting in complete failure of the liver and intestine to make triglyceride-rich lipoproteins.126 Previously, it had been thought that abetalipoproteinemic patients have a problem synthesizing apo B. Actually, apo B synthesis is reduced, but not abolished, suggesting that failure of the intestine and liver to synthesize apo B might not be the reason abetalipoproteinemic patients do not produce chylomicrons and VLDL,127 a fact that has been confirmed by the finding that the abetalipoproteinemia results from mutations of the microsomal triglyceride transfer protein gene.125,128 This gene’s lipid transfer activity is primarily responsible for the lipidation of the primordial particle—the initial step in chylomicron formation whereby the addition of phospholipids to the apo B molecule is followed by the addition of small amounts of triglyceride. Anderson’s disease, also known as chylomicron retention disorder, is another disorder of formation or secretion of chylomicrons by the small intestine. There is no defect in genes that carry known apoproteins or microsomal triglyc-

Chapter 100  Digestion and Absorption of Nutrients and Vitamins eride transfer protein,129 suggesting that this disease is caused by an unknown factor central to secretion of chylomicrons. Once chylomicrons have formed in the smooth endoplasmic reticulum, they are transferred to the Golgi apparatus. Golgi-derived chylomicron vesicles are then incorporated into the basolateral membrane and secreted by exocytosis into the lymphatic circulation (Fig. 100-7). During absorption, lacteals distend and endothelial cells, which overlap each other in the fasting state, move apart and open gaps through which chylomicrons can readily pass.130 Medium-chain fatty acids are absorbed by way of the portal vein, but as the chain length of saturated fatty acids increases, they are increasingly absorbed via the lymphatics. Polyunsaturated fatty acids may pass directly across the basolateral membrane and into the portal circulation.

conversion of cholesterol into bile acids so as to promote the net excretion of cholesterol. LXR-α up-regulates the transcription of CYP7A1 by directly binding to an LXRE in the promoter of this gene.136,137 The liver-specific expression of CYP7A1 requires LRH-1 (liver receptor homolog-1, also

Intestinal lumen

Lysophospholipids Phospholipids Fatty acids Cholesterol

Brush border SER

Liver X Receptors and Lipid Homeostasis

The liver X receptors (LXR-α and LXR-β) are nuclear receptor transcription factors that are activated by certain derivatives of cholesterol.131 Hence, LXR activity may be up-regulated by cellular lipid load or dietary cholesterol intake. The identification of a large list of LXR target genes and their response to LXR activation (Table 100-3) indicate that the LXRs play an important role in the response to excess cholesterol and that their activation might protect against tissue cholesterol overload.132 LXR-α and LXR-β form obligate heterodimers with the retinoid X receptor (RXR) to result in transcription factors that can be activated by ligands (i.e., lipids) for either RXR or LXR. RXR-LXR heterodimers bind to a specific DNA sequence called the LXR response element (LXRE), which consists of two hexanucleotide sequences separated by four bases (Fig. 100-8).133 A specific group of LXR agonists has been identified: intermediates in cholesterol metabolic pathways. Furthermore, using mouse models lacking LXR-α, LXR-β, or both, the key role of LXRs has been shown in regulating the expression of genes involved in cholesterol catabolism, absorption, and transport, as well as fatty acid synthesis (see Table 100-3). Cholesterol 7α-hydroxylase (CYP7A1) is the rate-limiting enzyme of the classic pathway that converts cholesterol into bile acids.134,135 The soluble bile acids, primarily produced in the liver, promote the secretion of cholesterol into bile for excretion in feces. In some species (e.g., rodents), but not humans, expression of the CYP7A1 gene is induced in response to dietary cholesterol, thereby accelerating the

RER

Golgi

Nucleus

Basement membrane Lamina propria Lymphatic vessels Figure 100-7.  Pathway taken by lipids during passage across an enterocyte. Triglyceride and phospholipid are synthesized in the smooth endoplasmic reticulum (SER) and accumulate there as dense droplets (small blue dots). Apolipoproteins, synthesized in the rough endoplasmic reticulum (RER), assist in the formation of chylomicrons and very-low-density lipoproteins in the tubular endoplasmic reticulum and Golgi apparatus (yellow circles). Chylomicrons are finally released across the basolateral membrane by exocytosis to enter the lymphatics.

Table 100-3  Liver X Receptor Gene Targets and Functions TARGET GENE

TARGET TISSUES

FUNCTION

CYP7A1 ABCA1

Liver Macrophage Intestine Liver, intestine Macrophage Macrophage, adipocyte Liver Mouse macrophage — Liver, fat Intestine Macrophage Liver

Bile acid synthesis Cholesterol efflux

ABCG5/G8 ABCG1 APOE LPL CETP SREBP-1c LXR YZ-2

Sterol transport Cholesterol efflux Component of lipoprotein particles Triglyceride hydrolysis Cholesterol ester transfer Fatty acid synthesis Cholesterol sensor Pyrimidine metabolism(?)

1705

1706

Section X  Small and Large Intestine called CPF and FTF), a monomeric orphan nuclear receptor. The transcription of CYP7A1 also is regulated via feedback inhibition. Specifically, the bile acid receptor FXR (farnesoid X receptor) binds bile acids and induces the expression of SHP (small heterodimer partner), an orphan nuclear receptor that preferentially dimerizes with LRH-1 and represses a number of enterohepatic genes involved in synthesis and transport of bile acids, including CYP7A1.138 In this way, LXR and FXR act together to tightly regulate bile acid homeostasis, respectively functioning, with a cholesterol precursor and cholesterol metabolite, to up-regulate and down-regulate CYP7A1. Dietary and secreted biliary cholesterol enter the intestinal lumen and are absorbed by proximal enterocytes. In the mouse, LXR and RXR agonists, which are metabolic indicators of active cholesterol turnover, decrease cholesterol absorption.139 This net reduction in cholesterol uptake is associated with an RXR- or LXR-mediated up-regulation of LXRE-containing genes that encode ABC transporters.140 In the small intestine, at least three ABC transporters are transcriptionally regulated by LXRs and putatively limit cholesterol absorption by pumping cholesterol back into the lumen of the gut.139 ABCA1 is a full transporter protein containing two symmetric halves, each composed of a six-transmembrane domain and an ABC. Mutations in the ABCA1 gene result in Tangier disease, which is characterized by a low concentration of plasma HDL and the accumulation of cholesterol esters in tonsils, liver, spleen, intestinal mucosa, and macrophage foam cells.141-143 ABCA1 is up-regulated by LXRs in both intestine and macrophages.144 ABCG5 and ABCG8 are half-transporters, each composed of a single transmembrane domain and ATP-binding cassette, and are coexpressed exclusively in the liver and intestine.104 Mutations in either gene cause a rare autosomal recessive disorder called sitosterolemia (see earlier), which

LXRE

LXRE

(A/G)G(G/T)T(C/T)Annnn(A/G)G(G/T)T(C/T)A (n = any nucleotide) Figure 100-8.  The consensus DNA sequence that defines the liver X receptor (LXR) response element (LXRE) to which functional LXRs bind to promote transcription. The LXRE consists of a direct repeat of the indicated hexanucleotide sequence separated by four bases. A/G and G/T signify that either base (A or G and G or T) may be used in the hexanucleotide sequence of the LXRE.

H 4

O

is characterized by increased absorption of cholesterol and toxic plant sterols and by decreased biliary sterol secretion. In vivo and in vitro experiments indicate that ABCG5 and ABCG8 are direct targets of LXRs. These findings strongly support the hypothesis that LXRs promote cholesterol loss by increasing biliary cholesterol secretion and limiting cholesterol absorption.104

CARBOHYDRATE DIETARY INTAKE

In Western societies, about 45% of total energy requirement is provided by carbohydrate, making it the major source of calories at all stages of life.145 The volume of carbohydrate ingestion appears to be declining, however, owing, in part, to a reduction in the intake of purified sugar.146,130 Overall, total calorie intake also is on the decline because of reductions in dietary fat and carbohydrate by affluent and dietconscious Western societies. The proportion of carbohydrate ingested as fruit and vegetables is rising as the intake of raw fiber increases. In adults who consume a Western diet, the amount of glucose produced by digestion is about 180 g/day (~1 mol). A growing amount of fructose had been increasingly added to our diets (often in excess of 50 g/day) through the widespread use of corn syrup as a sweetener, although recently there has been a move to limit this use and return to sugar as the preferred sweetening agent. All ingested glucose and galactose is absorbed normally, but the capacity to absorb fructose is limited in both young children and adults. This was evident in healthy, young adults (medical students in the United States and the United Kingdom) in whom the ingestion of 50 g of fructose produced abdominal pain, bloating, borborygmi, flatus, and a positive hydrogen breath test in 70% of subjects.147 It has been noted that two 12-ounce cans of some popular soft drinks contain about 50 grams of fructose in the form of corn syrup. About half of the digestible carbohydrate in an average Western diet is starch that is derived from cereals and plants and of which it is the major storage form of carbohydrate. Starch (as either amylose or amylopectin) is made up of long chains of glucose molecules. Amylose, a linear polymer in which each glucose molecule is coupled to its neighbor by α-1,4 linkage, has a molecular weight 106. Amylopectin, by contrast, is a branched-chain polymer in which α-1,6 links provide the angulations between adjacent chains of α-1,4 linked glucose molecules (Fig. 100-9); it has a molecular weight greater than 109. The amylose-to-amylopectin ratio

CH2OH O H OH H H

1

O H OH H

4

O

H

OH H 4

Figure 100-9.  Part of an amylopectin molecule indi­ cating the disposition of α−1,4 and α−1,6 linkages between glucose molecules.

CH2OH

H

H

O

CH2OH O

α–1,6 link 1

O

OH

H

H OH H

H

H

1

4

O

6 CH2 O

H

H OH H

H 1

4

O

6 CH2OH 5

OH

H

OH

H

H 1 OH H 3

H

O

α–1,4 link H

2

OH

α O

Chapter 100  Digestion and Absorption of Nutrients and Vitamins + Amylose

Maltotriose

Maltose

α amylase

Amylopectin

α-limit dextrins

Figure 100-10.  Action of pancreatic α amylase on amylose and amylopectin molecules. Because the α−1,6 link in the latter is resistant to amylase, the products include α-limit dextrins. Brown circles represent glucose units, and blue circles represent reducing glucose units. Sugars are classified as reducing or non-reducing based on their reactivity with Tollens’, Benedict’s, or Fehling’s reagents; if a sugar is oxidized by these reagents, it is called reducing since the oxidant (Ag+ or Cu2+) is reduced in the reaction. (From Gray GM. Carbohydrate absorption and malabsorption. In: Johnson LR, editor. Physiology of the Gastrointestinal Tract. 1st ed. New York: Raven Press; 1981. p 1064.)

varies widely, but most starches usually contain more amylopectin than amylose. Although starches are relatively easily digested, food preparation can influence their biologic availability. Use also may be determined by proteins associated with the starch, particularly gluten.1 Other major sources of dietary carbohydrate include sugars derived from milk (lactose), contained within the cells of fruits and vegetables (fructose, glucose, sucrose), or purified from cane or beet sources (sucrose). Processed foods form a major source of dietary sugars, particularly fructose and corn syrup; the latter contains not only fructose but also oligosaccharides and polysaccharides. The sugar alcohol sorbitol is used widely in the manufacture of diabetic sweets and preserves. Sorbitol is formed when the aldehyde group of glucose is hydrogenated to an alcohol group during manufacture, which slows its rate of absorption and thus diminishes its effect on blood sugar concentrations.48 Glycogen is the major storage form of polysaccharide in animals, but the amounts ingested in a normal diet are small. The structure of glycogen is similar to that of amylose and composed of straight chains of α-1,4-linked glucose monomers. Nonstarch polysaccharides form the majority of unavailable carbohydrates. The dietary fiber component of unavailable carbohydrate is found most abundantly in cereals, peas, beans, carrots, and peanuts. In the United Kingdom, 10 to 15 grams of dietary fiber, consisting predominantly of celluloses and hemicelluloses, is consumed by each person every day.146 Cellulose is made up of β-1,4-linked glucose molecules in straight chains, and hemicelluloses are pentose and hexose polymers with both straight and branched chains. Both forms are resistant to digestion in the small intestine because the β-1,4 bond, unlike the α bond in starch, is resistant to amylases. They are, however, broken down to some extent by colonic bacteria to yield short-chain fatty acids, which are avidly absorbed by colonic mucosa.148 The quantity of cellulose and hemicelluloses in vegetables and fruit varies markedly and depends on their age and ripeness. Other unavailable carbohydrates include pectins, gums, and alginates, which are only partially metabolized in the colon. Lignins, elaborated by plants in the process of becoming woody, are completely indigestible.48 It is well recognized that an increased intake of dietary fiber can ease constipation by increasing fecal bulk, mainly as a result of the increase in the mass of fecal flora. Dietary fiber has other roles, however, and also has effects on the absorption of other nutrients. Thus, for example, fiber

delays absorption of sugars and fats and curtails the insulin response to a carbohydrate meal. Some fiber, such as lignins, can lower serum cholesterol by binding bile salts. It may be these effects that have led to the widespread recommendation of a high-fiber diet for management or prevention of such diseases as diabetes mellitus and atherosclerosis. Finally, satiety is achieved more rapidly from a diet rich in fiber than from a low-fiber diet; it also takes longer to ingest a high-fiber meal. Advantage of this is taken in the management of obesity (see Chapter 6).

DIGESTION AND ABSORPTION Salivary and Pancreatic Amylase

Salivary and pancreatic amylases are endoenzymes; that is, they cleave the α-1,4 links internal to, or at the second or third bond from, the end of the polysaccharide chain. The products of amylase digestion therefore are short, linear oligosaccharides of maltotriose and maltose (Fig. 100-10). Because α-1,6 links, and the adjacent α-1,4 bonds, in the branched chains of amylopectin are not hydrolyzed by amylase, the products of amylopectin digestion include short, branched oligosaccharides, termed α-limit dextrins. Amylase proteins are encoded by a clustered gene family located on chromosome 1 of the human genome.149 In humans, the AMY1 gene is expressed in the parotid gland, and the AMY2 gene is expressed in the pancreas.150 The sequences of the pancreatic and salivary complementary DNAs are 94% similar, encoding for polypeptides with the same number of amino acids.151 Salivary amylase depends for its effect on its proximity to the ingested starches and the time they spend within the mouth. Thus, careful, slow chewing affords a good start to digestion, whereas rapid swallowing of poorly chewed foods—often a problem for edentulous persons—can cause suboptimal salivary amylase action. Salivary amylase is rapidly inactivated by gastric acid, but some activity may persist within the food bolus; shortchain oligosaccharides offer further protection for the enzyme against inactivation at acid pH. It is uncertain what fraction of dietary starch is digested before it reaches the duodenum. Pancreatic amylase is the major enzyme of starch digestion and, as with salivary amylase, produces short oligosaccharides, maltotriose, maltose, and α-limit dextrins; glucose monomer is not produced. Most of this hydrolysis occurs within the intestinal lumen, but because amylase also attaches itself to the brush border membrane of enterocytes, some digestion can occur at this site as well. Amylase

1707

1708

Section X  Small and Large Intestine Table 100-4  Characteristics of Brush Border Membrane Carbohydrases ENZYME

SUBSTRATE

PRODUCTS

Lactase

Lactose

Glucose Galactose Glucose

Maltase α–1,4 Linked oligosaccharides (glucoamylase) up to 9 residues Trehalase Trehalose Sucrase-isomaltase (Sucrose-α-dextrinase)   Sucrase Sucrose   Isomaltase   Both enzymes

a-limit dextrin, α–1,6 link α-limit dextrin, α–1,4 link at nonreducing end

Maltase or Isomaltase

Glucose Glucose Fructose Glucose Glucose

Maltase or Isomaltase

Isomaltase

concentration becomes limiting for starch hydrolysis only in severe cases of pancreatic insufficiency, in which luminal amylase activity levels are reduced to less than 10% of normal.152 Human milk contains amylase activity, which may be important for carbohydrate digestion in infants.153

Brush Border Membrane Hydrolases

The terminal products of luminal starch digestion, together with the major disaccharides in the diet (sucrose and lactose), cannot be absorbed intact and are hydrolyzed by specific brush border membrane hydrolases that are maximally expressed in the villi of the duodenum and jejunum. Several types have been identified (Table 100-4).154 Lactase hydrolyzes lactose to produce one molecule of glucose and one of galactose. Sucrase-isomaltase (SI, sucrase-α-dextrinase) possesses two subunits of the same molecule, each with distinct enzyme activity. Sucrase hydrolyzes sucrose to yield one molecule of glucose and one of fructose. Both sucrase and isomaltase remove glucose molecules from the nonreducing end of α-limit dextrins. Critically important is the ability of isomaltase (debrancher enzyme) in hydrolyzing the 1-6 glycosidic linkage in α-limit dextrins. The concerted action of sucrase and isomaltase thus yields monomeric glucose molecules from sucrose and α-limit dextrins (Fig. 100-11). In addition, two other carbohydrases participate in terminal hydrolysis of starch products: maltase-glucoamylase and α-limit dextrins. Maltase-glucoamylase acts on 1-4linked oligosaccharides containing as many as nine glucose residues, liberating glucose monomers. The human maltaseglucoamylase gene (MGAM) is located on chromosome 7 and has a structural homology similar to that of the SI gene.155 The maltase-glucoamylase enzyme does not undergo intracellular or extracellular proteolytic cleavage and is expressed in the brush border membrane as a monomeric protein. Maltase-glucoamylase is expressed prenatally with similar levels after birth and into adulthood.156 It has been suggested that isomaltase hydrolyzes the smallest α-limit dextrin, and another enzyme, α-limit dextrinase, is responsible for rapid hydrolysis of penta- and hexa-α-limit dextrins.157 The combination of SI, maltase and α-limit dextrinase serves to liberate glucose monomers very rapidly and close to hexose carriers, thus encouraging efficient absorption. Because free hexoses are found in the intestinal lumen, it is likely that the transport process is the rate-limiting step for uptake of monomers into the epithelium rather than the actions of the carbohydrases.

Sucrase or Maltase Sucrase or Maltase

+ Figure 100-11.  Actions of brush border membrane hydrolases. The combined actions of maltase, isomaltase, and sucrase yield glucose molecules from α-limit dextrins. Isomaltase is necessary to split the α−1,6 link. Brown circles represent glucose units, and blue circles represent reducing glucose units. Sugars are classified as reducing or non-reducing based on their reactivity with Tollens’, Benedict’s, or Fehling’s reagents; if a sugar is oxidized by these reagents, it is called reducing since the oxidant (Ag+ or Cu2+) is reduced in the reaction.

Trehalose is a disaccharide found predominantly in mushrooms, and so it is an insignificant element of the normal diet; nevertheless, there is a specific brush border enzyme, trehalase, for its hydrolysis to its two glucose molecules. Isolated trehalase deficiency has been reported in Greenland and can result in severe diarrhea after ingestion of mushrooms.156

Disaccharidase Biosynthesis and Regulation

Much has been learned about the gene regulation, biosynthesis, and processing of the disaccharidases.158,159,160,161 The human trehalase gene (TREH) is located on chromosome 11 and encodes a 583-amino acid protein with a molecular mass of about 75 kDa. SI is encoded by a single gene in the human,162 located on human chromosome 3 at locus 3q2526.163 The 5′-flanking region of the SI gene has a number of DNA regulatory regions that control initiation of gene transcription.164,165 Using mouse genetics, all four epithelial cell types in the small intestinal mucosa have the transcriptional machinery to express the SI gene.166 The elements necessary to direct intestinal epithelial cell–specific expression are embodied in a 201-nucleotide, evolutionarily conserved, 5′-flanking regions of the gene.167 At least three types of transcriptional proteins are involved in SI promoter trans­ cription, including hepatocyte nuclear factor 1 (HNF1),168,169 GATA-type zinc-finger transcription factor family members (GATA 4 and 5),170 and caudal-related homeodomain proteins (Cdx).171 The interaction of tissue-specific and tissue-

Chapter 100  Digestion and Absorption of Nutrients and Vitamins N

Transfer vesicle

N

Golgi apparatus N

Rough endoplasmic reticulum

Isomaltase site Sucrase site P Pancreatic protease High mannose oligosaccharides Complex type oligosaccharides

N

P

Plasma membrane

restricted transcription factors facilitates the transcription of genes in a single cell type. It has been suggested that the ratio of HNF-1α to HNF-1β might determine the degree of transcription induced by HNF-1.168,169 Congenital SI deficiency (CSID) is an autosomal recessive intestinal disease that is characterized by the absence of the sucrase and most of the maltase digestive activity within the SI enzyme complex; the isomaltase activity varies from absent to normal. Clinically, the disease is manifested as an osmotic-fermentative diarrhea upon ingestion of disaccharides and oligosaccharides. Analysis of this disorder at the molecular and subcellular levels has unraveled a number of phenotypes of CSID, which are characterized by perturbations in the intracellular transport, polarized sorting, aberrant processing, and defective function of SI.172,173 Changes in diet have a marked effect on the expression of SI. Starvation leads to a decline in brush border proteins and SI activity; this decline in sucrase-isomaltase activity is restored rapidly after refeeding. The type of carbohydrate ingested is important for regulation of SI expression. Starch and sucrose both induce SI activity, although sucrose is a more potent inducer.174 Study of the intestinal cell line Caco-2 has shown that a promoter region of the human sucrase gene (nucleotides −370 to +30) can down-regulate SI transcription in the presence of glucose.175 The human lactase gene is approximately 55 kb long; it has 17 exons and is located on the long arm of chromosome 2.176,177 Studies in intestinal cell lines have identified functional DNA elements in the lactase gene promoter that interact with nuclear transcription factors.178 Cdx proteins, GATA 5, and HNF-1α all have been shown to interact with the human lactase gene promoter and to activate transcription.179 Lactose intolerance is the most common manifestation of disaccharidase deficiency and results from an absence or drastically reduced level of lactase. In humans, lactase is expressed in fetal small intestine at a time in gestation just after the onset of expression of SI. Lactase expression is

Figure 100-12.  Biosynthesis of sucrase isomaltase. The nascent polypeptide (N) is translocated across the rough endoplasmic reticulum membrane after ribosomal mRNA translation. Oligosaccharide side chains join the polypeptide to be transferred to the Golgi apparatus for further processing. After incorporation in the plasma membrane, luminal proteases cleave the molecule into its active subunits. (From Lloyd ML, Olsen WA. Intestinal carbohydrases. Viewpoints Dig Dis 1991; 3:13-8.)

maintained throughout development and during childhood, although sometime during childhood, lactase activity declines to 5% to 10% of early childhood levels in most of the world’s populations. This decline occurs at the same time that intestinal SI activity is increasing. Ingestion of milk or milk products by persons with diminished lactase activity leads to flatulence, abdominal cramping, and diarrhea. This pattern of reduction of lactase activity has been termed late-onset lactase deficiency or adult-type hypolactasia. It initially was thought that the regulation of lactasephlorizin hydrolase (LPH) was post-translational and associated with altered structural features of the enzyme; it is now believed that the major mechanism of regulation of LPH is transcriptional (see later). Other forms of lactose intolerance include the rare congenital lactase deficiency and secondary forms, such as those caused by mucosal injury resulting from infectious gastroenteritis, parasitic infection, celiac disease, drug-induced enteritis, and Crohn’s disease. Differential activation of both the lactase and the SI promoter is effected by multiple similar transcription factors including GATA factors, HNF-1α, and Cdx-2, alone and in combination. This synergistic activation may be a method whereby higher levels of tissue-specific expression might be possible.180 Disaccharidase synthesis occurs within the endoplasmic reticulum, and the proenzymes then follow the path for secretory proteins through the Golgi complex before being inserted into the brush border membrane. All are glycoproteins and all undergo extensive intracellular processing with removal of redundant segments of the molecule. In the case of SI, final processing occurs on insertion into the brush border membrane after exposure to luminal pancreatic proteases (Fig. 100-12), at which point it is cleaved into its two active subunits; in contrast, lactase is already completely processed before its insertion. In their final active form, the carbohydrases project into the intestinal lumen, forming part of the glycocalyx, and they are attached to the membrane by a hydrophobic anchor

1709

1710

Section X  Small and Large Intestine that represents about 10% of the total mass of the molecule. Evidence suggests that MYO1A (brush border myosin I), a group of monomeric actin-based motors that are known to associate with membranes in intestinal villi, are involved in retaining SI within the brush border.181 Disaccharidases are synthesized by both crypt and villus cells but are expressed only on the latter. The expression of these genes in the intestine exhibits a complex spatial pattern along the vertical (crypt-to-villus) and horizontal (proximal-to-distal) axes.182 There is little SI activity in the crypts and villus tip cells, with maximal activity in lowerand mid-portions of the villus.183 The major mechanism for regulating the expression of the SI protein along the cryptvillus axis is the steady-state level of SI mRNA; however, post-transcriptional and post-translational regulation also likely play a role in the expression of the functional SI protein along the intestinal crypt-villus axis.184 A functional difference also exists between the jejunum and the distal ileum that reflects differences in the expression of different genes, or gradients of gene expression, along the proximal-distal axis of the intestine. For example, SI activity is four- to five-fold greater in the jejunum than in the ileum,185 although SI mRNA appears to be similar in the two areas. Although there are minor differences in the pattern of glycosylation in the Golgi apparatus, the major difference in regulation between the jejunum and ileum appears to be at the level of mRNA translation.186 Pancreatic proteolytic enzymes shorten the half-life of the carbohydrases.187 SI half-life can drop to as low as 4.5 hours after meals, compared with more than 20 hours during fasting. Presumably, proteolysis, as largely determined by meals, is responsible for the diurnal variation in carbohydrase activity.48 The levels of SI and other saccharidases also can decrease with infection and inflammation. In some cases, a decline in enzyme activity leads to malabsorption of carbohydrates with resultant diarrhea, flatulence, and weight loss. In most disease processes, however, the diminished levels of SI are associated with global dysfunction of the small intestinal mucosa.

Transport across the Mucosa

The three major diet-derived monosaccharides, glucose, galactose, and fructose, are absorbed by saturable carriermediated transport systems located in the brush border membrane of enterocytes in the proximal and mid small intestine.188 The active transport of glucose and galactose is achieved by the same transport protein that acts as a sodium cotransporter (primarily SGLT1)189; active glucose transport is driven by the sodium gradient across the apical cell membrane (Fig. 100-13). First, a low intracellular sodium concentration is generated by the Na+,K+-ATPase (adenosine triphosphatase) pump located in the basolateral membrane of the enterocyte, which transports 3 Na+ out of the cell and 2 K+ into the cell, resulting in a low intracellular Na+ concentration. Then two Na+ ions bind to the outer face of the transporter, producing a conformational change that permits subsequent sugar binding. The two Na+ ions and the glucose molecule then are transferred to the cytoplasmic face of the membrane through another conformational change involving a coordinated rotation or tilt (or both) of transmembrane helices.190 At the cytoplasmic surface, glucose dissociates first, and then the two Na+ ions dissociate into the cytosol to produce a ligand-free transporter. The low affinity of the cytosolic sites for glucose and Na+, and the low intracellular Na+ concentration relative to the extracellular concentration (10 vs. 140 mEq/L), promote these dissociations. The ligand-

Glucose Galactose

Na+

SGLT1

Na+

~

Na+

Fructose

GLUT5

K+

GLUT2

GLUT5

Glucose Fructose Galactose Figure 100-13.  Sodium gradient hypothesis of glucose absorption. The sodium pump (Na+,K+-ATPase) at the basolateral membrane generates a low intracellular sodium concentration. Sodium passes down the concentration gradient so created across the apical membrane coupled to glucose on a common carrier. The sodium pump thus generates the energy for this system. Glucose leaves the cell via facilitated diffusion across the basolateral membrane. GLUT2, glucose transporter; GLUT5, fructose transporter; SGLT, sodium-glucose cotransporter.

free transporter then relaxes to the outward-facing conformation to complete the cycle. The complete enzymatic turnover of the transporter occurs about 1000 times a second at 37°C. Although some of this glucose fuels cellular metabolism, a sizable fraction passes out of the cell across the basolateral membrane by facilitated diffusion (uniport). The net result is that for every glucose molecule that is transported across the brush border, Na+ ions (and two accompanying anions) also are transported across the epithelium. This, in turn, draws about 1100 water molecules across the epithelium to maintain iso-omolarity of the absorbate. Ion and nutrient absorption across the intestine do not increase the osmolarity of the fluid remaining in the intestinal lumen. The coupling among glucose, salt, and water absorption provides the explanation for the finding that water absorption across the upper and mid intestine is glucose dependent, and it is the rationale for the oral rehydration therapy (ORT) used so effectively to treat patients with secretory diarrhea (see Chapter 107).191 The prevailing opinion is that two types of glucose transporters are found across brush borders: One is a high-affinity Na+-dependent, phlorizin-sensitive transporter (SGLT1), and the other is a low-affinity transporter that might or might not be Na+-dependent and phlorizin-sensitive. Candidates for the latter role in humans include GLUT2, SGLT4, and SGLT6.192 The sodium-glucose cotransporter (SGLT1) has been characterized extensively.193,194,195 Activity of this 73-kDa

Chapter 100  Digestion and Absorption of Nutrients and Vitamins cotransporter in the intestinal brush border membrane rests with the presence of four independent, identical subunits arranged in a homotetramer. SGLT1 resides on chromosome 22 and has been cloned and sequenced. The cloned cDNA encodes for transport activity with the same relative specificity as the previously characterized native transport system: d-glucose > α-methyl-d-glucose > d-galactose > 3-O-methyl-d-glucopyranose >>> l-glucose.196 The cDNA encodes a 662-amino acid protein with a predicted molecular weight that correlates well with the biochemically defined size. SGLT1 is predicted to have 14 membranespanning domains, with one asparagine-linked carbohydrate group on the third extracytoplasmic loop.197 The expression and activity of glucose transport in the intestinal brush border are regulated by short-term and longer-term processes. In the short term, activity of glucose transport is increased by both protein kinase A- and C-dependent processes.198 The mechanism of this enhanced activity is an increase in the number of membrane transporters, mediated by changes in exocytosis and endocytosis of membrane vesicles that contain the transport protein. Longer-term regulation of glucose transport is mediated by changes in the expression of SGLT, which is controlled by changes in the nutrient environment.199 Glucose-galactose malabsorption is characterized by the neonatal onset of severe diarrhea upon the newborn’s ingestion of breast milk or regular infant formula. Several distinct mutations in the SGLT1 gene have been identified (e.g., a missense mutation resulting in a change of amino-acid residue 28 from an aspartate to an asparaginase),9 and most of these mutations are responsible for defective passage of the SGLT1 through the biosynthetic machinery from the endoplasmic reticulum or for poor trafficking from the Golgi apparatus to the brush border membrane. Rarely do mutant SGLT1 proteins reach the brush border at a normal rate, in which case, the glucose transport is defective.192 Fructose absorption occurs by facilitated diffusion: Transport occurs not against a concentration gradient but with a carrier protein to achieve transport rates greater than one would expect from simple diffusion. This process is completely independent of glucose absorption. Studies in humans have shown that there is a saturable, facilitative transport system for fructose in the intestinal epithelium that has a lower activity than that for transport of glucose and galactose. The protein responsible for most apical membrane fructose transport is a member of the facilitative monosaccharide transporter family called GLUT5, encoded by the gene SLC2A5. This 501-amino acid protein in humans has 12 membrane-spanning domains, as do other GLUT molecules, and transports fructose exclusively.200 GLUT2, however, might assist in absorption of excess luminal fructose. Little fructose is metabolized in the enterocytes, and fructose is transported across the basolateral membrane (by GLUT2 and GLUT5) and is taken up and metabolized rapidly by the liver, resulting in low postabsorptive blood levels of fructose. There may be more than one type of fructose transport system. Malabsorption of fructose in humans can be prevented by the simultaneous administration of glucose, suggesting that another glucose-responsive system may be present in the enterocytes. No inherited disorders of fructose transport (GLUT5) have been reported.201 Overall, fructose is not as well absorbed as glucose. High levels of dietary fructose can lead to dietary intolerance. Ingestion of large amounts of fructose can cause diarrhea, excessive intestinal gas, and recurrent abdominal pain. Fructose malabsorption has been associated with similar symptoms.202

Debate has developed over the mechanism of the passive or diffusive component of intestinal glucose absorption and, indeed, whether it exists.203 Pappenheimer and colleagues proposed that paracellular solvent drag contributes a passive component, which, at high concentrations of sugars similar to those in the jejunal lumen immediately after a meal, is several-fold greater than the active component mediated by the Na+-glucose cotransporter SGLT1.204 Other investigators have argued that the kinetics of glucose absorption can be explained solely in terms of SGLT1 and that a passive or paracellular component plays little, if any, part.205 Morerecent data suggest that the passive component of glucose absorption exists but that it is facilitated because it is mediated by the rapid glucose-dependent activation and recruitment of the facilitative glucose transporter GLUT2 to the brush-border membrane. This is regulated through a protein kinase C–dependent pathway activated by glucose transport through SGLT1 and also involves mitogen-activated protein kinase (MAP kinase) signaling pathways.206

Exit from the Epithelium

Most hexoses are exported from the epithelial cell by way of the basolateral membrane, although small amounts are used for intracellular metabolism. Exit across the basolateral membrane depends on facilitated diffusion (not requiring energy) via a specific carrier. Two genes that are expressed in the small intestine—GLUT2, the basolateralmembrane–associated glucose transporter, and GLUT5, an apical membrane fructose transporter—encode these facilitative sugar transport proteins.207 GLUT2 has molecular structural characteristics similar to those of the other members of this family of genes. The protein has 500 amino acids with many hydrophobic residues that predict a total of 12 membrane-spanning domains. There is one long extracellular loop between membrane-spanning domains 1 and 2 that contains an asparagine that is N-glycosylated and one long cytoplasmic loop between membrane-spanning domains 6 and 7. Once the hexoses have entered the interstitial space, they pass onward by diffusion into the portal circulation. A congenital defect in glucose transport by GLUT2 has been identified (Fanconi-Bickel syndrome). Because GLUT2 is normally expressed in the liver, pancreas, and kidney as well as in the intestine, defects in this transporter are expected to have a widespread effect on glucose homeostasis. Indeed, patients with the Fanconi-Bickel syndrome exhibit tubular nephropathy, fasting hypoglycemia, rickets, stunted growth, and hepatomegaly secondary to glycogen accumulation.208,209 The accepted dogma of intestinal glucose absorption at the basolateral membrane by glucose transporters has been challenged by studies of intestinal glucose absorption in GLUT2 null mice and in patients with GLUT2 deficiency; in both cases, glucose absorption was not impaired. Additional work has suggested that there are two separate pathways for the exit of sugar from enterocytes: one that involves GLUT2 and another that requires glucose phosphorylation, the transfer of glucose-6-phosphate into the endoplasmic reticulum and the release of free glucose into the blood. The release mechanism is unclear, but it has been proposed to involve vesicle trafficking. This postulate is supported by oral tolerance tests on a patient with congenital deficiency in glucose-6-phosphate translocase1, in whom glucose absorption was impaired but not eliminated.209 Not all potentially digestible carbohydrate is absorbed in the small intestine. As much as 20% of dietary starch escapes into the colon, particularly that derived from cereals

1711

1712

Section X  Small and Large Intestine and potatoes.2 Most of this unclaimed carbohydrate, however, is metabolized by colonic bacteria, and the shortchain fatty acids thus derived are readily absorbed; hydrogen and methane also are generated and contribute to production of flatus.

PROTEIN DIETARY INTAKE

Dietary proteins are the major source of amino acids, and in the average Western diet they provide about 10% to 15% of energy intake. Affluent populations ingest more protein than they need to maintain their normal balance. An average adult in a Western country consumes 70 to 100 g of protein per day, whereas the poor in Asia and Africa consume 50 g or less per day.48 Recommended dietary requirements vary from 0.75 to 1 g/kg of body weight per day, but deficiency states are rare even with intakes of 0.5 g/kg per day or less. In the United Kingdom, protein intake has remained fairly steady since the mid-1970s, but with the marked decline in fat and carbohydrate ingestion, the ratio of protein to nonprotein energy intake has risen.48 Little harm appears to occur in the unusual subgroups of society who consume very large amounts of protein, although renal function can be impaired by this dietary habit. The Masai tribes of Africa and the Gaucho of South America, who consume 250 to 300 g (largely of animal origin) per day, suffer no obvious untoward effects from this consumption.48 The variety of types of animal and plant proteins is enormous. Generally, plant proteins are less digestible than those derived from animals, but some fibrous animal proteins, such as keratin and collagen, also are relatively indigestible. High-proline proteins such as the glutenins are less thoroughly digested than are others. The quality of proteins depends largely on their amino acid composition; proteins rich in essential amino acids are regarded as being of high quality. Proteins from animal sources have a high content of essential amino acids, unlike proteins from certain specific plant sources, which are said to be incomplete because they lack or contain only certain essential amino acids. Such deficiencies in essential amino acids typically are overcome in a mixed diet; however, the relative contribution of animal and plant protein varies according to geographic region. In developed countries, such as in North America and Europe, animal protein contributes about 70% of the total protein compared with developing nations in the Middle East and Africa where the animal protein contribution can be as low as 20%. Food processing, by heat for example, can cause interand intramolecular bonding in the proteins to produce polymeric forms that are relatively resistant to hydrolysis.1 Other constituents of the diet also can interfere with protein digestion; for example, starch and reducing sugars have the potential to impair digestion.48 Despite these interferences, digestion and absorption of proteins are remarkably complete, and only about 3% to 5% of ingested nitrogen is lost in the stool, probably because of the resistance of some peptide bonds to hydrolysis.1 A few selected proteins are resistant to proteolysis in the small intestine, including secretory immunoglobulin (Ig) A and intrinsic factor. Among the 20 common amino acids that form animal and plant proteins, 8 cannot be synthesized by animals: leucine, isoleucine, lysine, methionine, phenylalanine, threonine, tryptophan, and valine. These eight essential amino acids have to be ingested, usually in plant-derived foods. Histidine also is required for growth in infants.

PROTEINS FROM ENDOGENOUS SOURCES

Almost half of all protein that enters the intestine is derived from endogenous sources. Of this, about 20 to 30 g/day are derived from secretions of salivary, gastric, biliary, pancreatic, and mucosal origin. Another 30 g/day of protein are provided by epithelial cells desquamated from the villus tips, and 2 g of plasma proteins are delivered into the intestinal lumen each day.

DIGESTION AND ABSORPTION Pepsins

Digestion of proteins begins in the stomach with the action of pepsins. Pepsins are released from their precursor pepsinogens with the loss of a small basic peptide by auto­ activation in an acid pH. At neutral or alkaline pH, the pepsinogen amino-terminal region is folded in such a way as to mask a catalytic site. In the acidic environment of the stomach, the catalytic site is uncovered and then proceeds to remove the amino-terminal region, which consists of 40 amino acids, thereby generating the active form of the molecule, pepsin. Pepsinogen release from chief cells is stimulated by gastrin, histamine, and cholinergic stimulation and closely mirrors acid secretion.210 The pepsins are a family of endoproteases that hydrolyze internal peptide bonds in proteins. They act preferentially on peptide bonds formed by the aromatic amino acids phenylalanine and tyrosine and by the branched-chain amino acid leucine. There are two immunologically distinct groups of pepsins (groups 1 and 2), although eight fractions are identified electrophoretically. Both of the immunologically separated species are secreted by chief cells, but group 2 isoforms also are present in the mucus cells in the oxyntic and pyloric areas of the stomach and in Brunner’s glands of the duodenum. Their substrate specificities vary little, but their pH optima differ slightly (between 1.8 and 3.5); all are irreversibly inactivated in alkali. Pepsins remain active at the acid pH of gastric contents to produce a mixture of peptides with a small portion of amino acids. Pepsin activity is therefore confined to the stomach. The completeness of gastric proteolysis depends, in part, on the rate of gastric emptying, the pH of intragastric contents, and the types of protein ingested. Moreover, the products of protein digestion by pepsins in the stomach can further influence acid and pepsinogen secretion as well as gastric emptying. Subjects who are achlorhydric or who have lost control of gastric emptying as a result of pyloroplasty or partial gastrectomy do not appear to have a problem with assimilation of protein, suggesting that gastric proteolysis is not an essential component of digestion.

Pancreatic Proteases

In contradistinction to amylase and lipase, which are secreted in their active forms, each of the pancreatic pro­ teases is secreted as a proenzyme and therefore must be activated within the intestinal lumen. Enterokinase (enteropeptidase) plays a key role in proteolysis. It is liberated from its superficial position in the brush border membrane by the action of bile acids,211 its action being to convert trypsinogen to trypsin by removing its hexapeptide NH2 terminus. Trypsin in turn activates the other proteases and continues to split more trypsin from trypsinogen (Fig. 100-14). The proteases are classified as endo- and exopeptidases, according to the sites of the peptide bonds against which they are most active. Endopeptidases include trypsin, chymotrypsin, and elastase, and exopeptidases include carboxypeptidase A and B (Table 100-5).

Chapter 100  Digestion and Absorption of Nutrients and Vitamins Chymotrypsinogen

Chymotrypsin

Proelastase Trypsinogen

Elastase

Trypsin Procarboxypeptidase A

Carboxypeptidase A

Enterokinase Procarboxypeptidase B

Carboxypeptidase B

3000

ACTION

PRODUCTS

Trypsin

Endopeptidase; cleaves internal bonds at lysine or arginine residues; cleaves other pancreatic proenzymes Endopeptidase; cleaves bonds at aromatic or neutral amino acid residues Endopeptidase; cleaves bonds at aliphatic amino acid residues Exopeptidase; cleaves aromatic amino acids from carboxy terminal end of protein and peptides Exopeptidase; cleaves arginine or lysine from carboxy terminal end of proteins and peptides

Oligopeptides

Chymotrypsin

Elastase Carboxypeptidase A

Carboxypeptidase B

Oligopeptides

Oligopeptides Aromatic amino acids and peptides

Absorption rate (µmol/min/30 cm jejunum)

Table 100-5  Pancreatic Proteolytic Enzymes, Their Sites of Action, and Their Products ENZYME

Figure 100-14.  Activation of pancreatic proteolytic enzymes. Enterokinase (enteropep­ tidase) plays a critical role in activating trypsinogen to form trypsin. Trypsin in turn activates not only more trypsinogen but also the other proteolytic enzyme precursors. Products of enzymatic actions are shown in italics.

2400

Triglycine

1800

Diglycine 1200

600 Free glycine

Arginine, lysine, and peptides

0 0

100

200

300

Amount of glycine in perfusate (mmol)

Trypsin, chymotrypsin, and elastase have specificity for peptide bonds adjacent to certain specific amino acids. They split peptide bonds within the protein molecule, whereas exopeptidases remove a single amino acid from the carboxyl terminal end of the peptide. Trypsin produces short-chain oligopeptides that are further hydrolyzed by the exopeptidases, carboxypeptidase A acting on aromatic and aliphatic carboxyl terminals, and carboxypeptidase B acting on peptides containing basic carboxyl terminals. The final products of intraluminal digestion thus are produced by cooperative activity of endo- and exopeptidases and consist of a number of neutral and basic amino acids together with peptides of two to six amino acids in length. About 30% of luminal amino nitrogen is found in amino acids and about 70% is found in oligopeptides.212 In addition to nutrient protein hydrolysis, pancreatic proteases have other functions: They split vitamin B12 from the R protein to which it is linked so that it then can bind intrinsic factor; they increase the turnover of brush border membrane hydrolytic enzymes, and, as discussed earlier, they initiate the final steps in the processing of the SI complex; finally, they may have a role in the inactivation of some organisms.1

Figure 100-15.  Rates of glycine absorption (mean ± standard error of mean) from perfusion solutions containing equivalent amounts of glycine in free or peptide (diglycine, triglycine) form. Results are from studies in the jejunums of four normal humans. (From Adibi SA, Morse EL, Masilamani SS, Amin PM. Evidence for two different modes of tripeptide disappearance in human intestine. Uptake by peptide carrier systems and hydrolysis by peptide hydrolases. J Clin Invest 1975; 56:1355-63.)

Digestion at the Brush Border Membrane and in the Cytoplasm In contrast to the absorption of carbohydrate, which is largely restricted to uptake of hexose monomers across the brush border membrane, amino acids can be absorbed either as monomers or as di- or tripeptides. Indeed, amino acid absorption is achieved more efficiently in the form of peptides than as single amino acids (Fig. 100-15).213 The fact that the vast majority of the end-products of protein digestion that reach the portal circulation are amino acids, however, speaks strongly in favor of the presence of peptidases in the epithelium. Patients with cystinuria and Hartnup’s disease, who have specific defects in the absorption of basic and neutral amino acids, respectively, do not develop protein deficiency states

1713

1714

Section X  Small and Large Intestine Table 100-6  Peptidases Found on the Brush Border Membrane and in the Cytoplasm of Villus Epithelial Cells PEPTIDASE Brush Border Membrane Peptidases Amino-oligopeptidases (at least two types) Aminopeptidase A Dipeptidase I Dipeptidase III Dipeptidyl aminopeptidase IV Carboxypeptidase P Gamma-glutamyl transpeptidase Folate conjugase Cytoplasmic Peptidases Dipeptidases (several types) Aminotripeptidase Proline dipeptidase

ACTION

PRODUCTS

Cleave amino acids from carboxy terminus of 3-8 amino acid peptides Cleaves dipeptides with acidic amino acids at amino terminus Cleaves dipeptides containing methionine Cleaves glycine-containing dipeptides Cleaves proline-containing peptides with free α-amino groups Cleaves proline-containing peptides with free carboxy terminus Cleaves γ-glutamyl bonds and transfers glutamine to amino acid or peptide acceptors Cleaves pteroyl polyglutamates

Amino acids and dipeptides

Gamma glutamyl amino acid or peptide Monoglutamate

Cleave most dipeptides Cleaves tripeptides Cleaves proline-containing dipeptides

Amino acids Amino acids Proline and amino acids

Table 100-7  Distribution of Peptidase Activity SUBSTRATE Dipeptides Tripeptides Tetrapeptides Higher peptides

BRUSH BORDER MEMBRANE (%) 5-10 10-60 90 98

CYTOPLASM (%) 80-95 30-60 1-10 0

because the absorption of peptides in these patients is normal.214 The discovery that di- and tripeptides are actively transported by the brush border membrane of enterocytes has been valuable in explaining this observation, and it emphasizes the need for critical evaluation of the supposed nutritional advantage provided by elemental diets that consist only of free amino acids. A range of peptidases is present in the brush border membrane and in the cytoplasm of villus epithelial cells for the hydrolysis of oligopeptides up to approximately eight amino acid residues in length (Table 100-6).215-217 The peptidases on the brush border membrane differ in several important respects from those within the cytoplasm (Table 100-7). About 90% of the dipeptidases are found in the cytoplasm and only about 10% in the brush border, whereas the distribution of hydrolases for tetrapeptides is the reverse of this. Peptidases for pentapeptides and larger molecules are confined almost entirely to the brush border membrane. Cytoplasmic enzymes are much more heat labile than those in the brush border, and there are differences in the electrophoretic mobility patterns for the two sets of enzymes.75 Most oligopeptidases appear to be aminopeptidases; that is, they act by removing residues from the amino terminus of the peptide. The chain length of the peptides is an important factor that determines not only whether the site at which hydrolysis occurs is at the brush border or within the cell but also its rate. Thus, rates of brush border membrane hydrolysis for tripeptides are most rapid and for dipeptides is least rapid, whereas tetra- and pentapeptide hydrolysis rates occupy an intermediate position.213

Amino acids Amino acids Amino acids Peptides and amino acids Peptides and amino acids

Distinct from the amino oligopeptidases are at least three other peptidases. Aminopeptidase A has specificity for peptides with acidic amino acids at their amino termini. Aminopeptidases 1 and 3 (distinguished on electrophoretic mobility) have specificities for different substrates with different amino acid peptide bonds.1 Proline-containing oligopeptides are not readily hydrolyzed by most proteases, although many proteins— including collagen, gliadin, and casein—are rich in proline. Two proline-specific carboxypeptidases, however, have been demonstrated in the brush border membrane. They have slightly different substrate specificities218 and together with a cytoplasmic proline dipeptidase, they are likely to be responsible for hydrolysis of proline-rich peptides. A number of other brush border membrane peptidases need to be mentioned. Gamma glutamyl transpeptidase hydrolyzes gamma glutamyl peptide bonds, with the transfer of the gamma glutamyl group to another amino acid to form a gamma glutamyl amino acid or peptide derivative.1 The role of this brush border membrane in the intestine is not yet clear. Folate conjugase, an enzyme concerned with hydrolysis of dietary folate, will be considered later. The recent demonstration of angiotensin I-converting enzyme (ACE) in intestinal mucosa suggests that it, too, might hydrolyze dietary peptides.219 Indirect evidence suggests that endopeptidases also may be present on the brush border membrane, because protein digestion occurs, even in the complete absence of pancreatic function; these enzymes have yet to be isolated. As with other proteins, synthesis of each specific peptidase occurs in the rough endosplasmic reticulum, and following transfer to the Golgi apparatus, the proteins are transported to the brush border membrane, where they are inserted by exocytic fusion.220,221 They are attached to the basement membrane by short anchoring pieces in a manner analogous to the attachment of disaccharidases222; however, unlike the latter enzymes, there is little post-translational processing, either within the cytoplasm or by pancreatic enzymes on the brush border. Of the cytoplasmic dipeptidases, the most abundant appears to be one with broad specificity for neutral amino acid-containing dipeptides. The tripeptidase isolated has broad specificity for amino-terminal residues and high spec-

Chapter 100  Digestion and Absorption of Nutrients and Vitamins ificity toward tripeptides containing proline as the aminoterminal residue, which distinguishes it from the brush border membrane amino oligopeptidase. Other characteristics of the tripeptide that are required for rapid hydrolysis include a free α amino group, an α carboxyl group, and an l-configuration for the two amino acid residues.223

Absorption of Peptides

Substrate inhibition studies indicate that tri- and dipeptides inhibit uptake of either peptide from the lumen, but neither is affected by single amino acids. Such evidence suggests that small peptides use a separate transporter system from those used by single amino acids. By contrast, tetrapeptide absorption is inhibited by single amino acids but not by di- and tripeptides, suggesting that tetrapeptides are split before absorption. The advantage of dipeptide absorption over single amino acid absorption has been largely demonstrated experimentally with single peptides containing a single amino acid, usually glycine.213 Several studies, however, have demonstrated the kinetic advantage of peptides over amino acids, even in complex mixtures of partial digests of proteins.224,225 Absorption was greater from tryptic hydrolysates of proteins than from a mixture of amino acids. Furthermore, the wide variation in rates of absorption seen with different individual amino acids was reduced when they were presented as a tryptic hydrolysate. A number of other factors influence digestion and absorption. The presence of amino acids in the lumen inhibits peptide hydrolysis (product inhibition), whereas luminal glucose and luminal acidification each inhibit amino acid and peptide absorption.213 There is good evidence to suggest that di- and tripeptides are taken up by a single type of transporter with some stereospecificity because the length of the amino acid side chains on the di- or tripeptides is important; the longer the side chain the more preferred the substrate for the absorption site (Table 100-8).226 The l-isomers of the amino acids in dipeptides are much preferred to the d-isomers, whereas the presence of acidic and basic amino acid residues in dipeptides reduces affinity for the transport system, compared with neutral amino acid residues. Affinity is also greater for dipeptides than for tripeptides, at least in the example of peptides that contain glycine. The transporter for peptides is not dependent on sodium, but cotransport with protons may occur instead.227 The peptide transporter for human small intestine has been cloned228,229 and is a member of a superfamily of H+coupled peptide transporters. The human protein consists of 708 amino acids, with a predicted core molecular size of 79 kDa that contains 12 transmembrane domains. The gene is located on chromosome 13. In humans, it is expressed in the small intestine (duodenum, jejunum, and ileum), but not in the esophagus, stomach, or colon. In the small intestine it is expressed only on absorptive epithelium. It recog-

Table 100-8  Relative Specificities of Intestinal Peptide Transporters DIPEPTIDES

TRIPEPTIDES

L-form

d-form

of amino acids Neutral amino acids Long side chains

Acidic or basic amino acids Short side chains

nizes a variety of neutral, anionic, and cationic dipeptides as substrates,230,231 which explains the broad substrate specificity of the intestinal peptide transport system. The most interesting feature of this transport process is that it uses a transmembrane electrochemical H+-gradient rather than a transmembrane electrochemical Na+-gradient as its driving force.232 There is an acid pH microclimate on the luminal surface of the intestinal brush border membrane that creates a H+-gradient across the brush border membrane in vivo. This acid pH microclimate is generated and maintained by the combined action of the Na+-H+ exchanger in the brush border membrane and Na+,K+-adenosine triphosphatase (ATPase) in the basolateral membrane of the enterocyte. The mechanism of the transport process is a simultaneous translocation of H+ and peptide substrate involving a single H+ binding site on the protein (Fig. 100-16).233,234 A multitude of processes are involved in the absorption of peptides. The well-established processes include a Na+H+ exchanger located in the brush-border membrane that maintains an intracellular alkaline pH, a Na+,K+-ATPase located in the basolateral membrane that maintains an inside negative membrane potential, and several cytoplasmic peptidases that prevent intracellular accumulation of absorbed peptides. These enzymes convert most of the

Lumen Dipeptides, tripeptides

H+

H+

Na+-H+ exchanger

Brush-border membrane

~

Pept-1

Peptides

H+

Na+

Peptidase

Na+ Amino acids

Peptides

~ Basolateral membrane Amino acids

ATPase Peptides

K+

Blood Figure 100-16.  Peptide transport across the intestinal epithelium. This transport process uses a transmembrane H+-gradient rather than a transmembrane electrochemical Na+-gradient as the driving force. The acid pH microclimate on the luminal surface of the intestinal brush border membrane is generated and maintained by the combined action of the Na+-H+ exchanger in the brush border membrane and Na+,K+-ATPase (adenosine triphosphatase) in the basolateral membrane of the enterocyte. The mechanism of the transport process is a simultaneous translocation of H+ and peptide substrate involving a single H+ binding site on the protein, Pept-1.

1715

1716

Section X  Small and Large Intestine absorbed oligopeptides to amino acids that are either used by the absorbing cells or are released into the portal circulation via the amino acid transporters located on the basolateral membrane of these cells. The oligopeptides that escape hydrolysis by the cytoplasmic peptidases are transported across the basolateral membrane into the portal circulation by a peptide transporter that appears to be different from the Pept-1 transporter. Oligopeptide transport could be regulated by alteration in activity or abundance of Pept-1, Na+H+ exchanger, Na+,K+-ATPase, cytoplasmic peptidases, and basolateral oligopeptide transporter.235,236 Studies of individual substrates and hormones in cell culture have shown that the membrane population of Pept-1 is increased by dipeptides, certain amino acids, insulin, and leptin and decreased by epidermal growth factor (EGF) and triiodothyronine. In the case of dipeptides, EGF, and thyroid hormone, there are parallel changes in the gene expression brought about by alteration of transcription or stability of Pept-1 mRNA. In contrast, treatment with insulin and leptin does not induce any alteration in the Pept-1 gene expression, and the mechanism of increased protein expression appears to be increased trafficking from a preformed cytoplasmic pool to the apical membrane.235,236

Transport of Amino Acids

Whereas there appears to be only one type of dipeptide transporter in the brush border membrane for the 400 possible dipeptides, there is a multiplicity of transport mechanisms for the 20 amino acids. In adults, these are situated on villus enterocytes and involve carrier-mediated active transport or facilitated diffusion processes, which typically depends on the Na+ gradient as the driving force; a small portion may be absorbed by simple diffusion, independent of any ion gradient. There has been some difficulty in defining the number and types of transporters because of their overlapping specificities; several amino acids use a number of different transport systems (Table 100-9). On the basis of kinetic studies, at least four active processes have been identified for transport of neutral amino acids across the

Table 100-9  Major Amino Acid Transport Systems Detected in Intestinal Epithelial Cells TRANSPORT SYSTEM Brush Border Membrane Neutral Amino Acids SLC6A19 SLC36A1 SLC6A20 SLC6A14 SLC1A5 SLC7A9/SLC3A1 Basic Amino Acids Acidic Amino Acids SLC1A1 (X-GA−) Basolateral Membrane L A SLC1A5 (ASC) N

apical cell membrane. Each is electrogenic and sodium dependent. One has broad specificity for a number of neutral amino acids (NBB system); a second provides another route for phenylalanine and methionine (PHE system); a third provides a mechanism for imino acid absorption (IMINO system); and the fourth transports beta amino acids. Separate sodium-dependent, active transport processes for basic and acidic amino acids also have been demonstrated, and some evidence suggests that facilitated diffusion of these types of amino acids also occurs, although this is likely to be a minor pathway. Genomic advances have allowed most mammalian amino acid transport functions to be attributed to specific gene products: At least 52 amino acid transporter-related gene products are grouped within 12 solute carrier families, with their own new nomenclature.237 The classic Na+-dependent imino acid transporter has been identified as the human PAT1 (human proton-coupled amino acid transporter 1) or solute carrier SLC36A1. This high-capacity imino acid carrier has been localized to the small intestinal luminal membrane and transports imino and amino acids (glycine, proline, alanine, taurine).238,239 Human PAT1 mediates 1 : 1 symport of protons and small neutral amino acids. The acid microclimate of the brush border membrane drives transport of the amino acids into the cytosol. Transport activity is independent of Na+ and Cl− (Fig. 100-17). In addition, the IMINO system is a Na+-dependent transporter with specificity toward the imino acids proline and hydroxyproline. The protein responsible for this transport activity is SIT1(Na+coupled imino acid transporter 1).240,241 System B0,+, also present on the brush border membrane, mediates the Na+ and Cl− coupled electrogenic transport of neutral as well as cationic amino acids across the brush border membrane. The gene encoding the protein respon-

Apical

Basolateral

Gly, Ala, Pro PAT-1 H+

H+

SUBSTRATES

Neutral amino acids Imino acids; proline, hydroxyproline Imino acids Neutral and cationic amino acids Alanine, serine, cysteine, glycine, asparagine Neutral amino acids, cationic amino acids, cysteine Lysine, cysteine, basic amino acids Glutamate, aspartate Broad selectivity Broad selectivity Neutral amino acids, alanine, serine, cysteine Glutamine, histidine, asparagine

H+

H+ Na+-H+ exchanger

Na+

Enterocyte Figure 100-17.  Intestinal amino acid transport. The human proteincoupled amino acid transporter (PAT-1) is involved in the absorption of small amino acids across the apical membrane. The acid microenvironment generated by a Na+-H+ exchanger provides the electrochemical proton gradient that drives amino acids to the cytosol. Ala, alanine; Gly, glycine; Pro, proline. (Boll M, Daniel H, Gasnier B. The SLC36 family: Proton-coupled transporters for the absorption of selected amino acids from extracellular and intracellular proteolysis. Pflugers Arch 2004; 447[5]:776-9).

Chapter 100  Digestion and Absorption of Nutrients and Vitamins sible for this activity, ATB0,+ is located on human chromosome X. A separate Na+-dependent transport system, X−AG, (SLC1A1) is specific for anionic amino acid, aspartate, and glutamate. The glutamate transporter expressed in the intestinal brush border membrane is known as the EAAT3. This transporter is defective in the inherited amino acid transport defect known as dicarboxylic aciduria. Hartnup’s disease is a disorder of renal and gastrointestinal neutral amino acid transport that is inherited as an autosomal recessive trait. The gene causing Hartnup’s disease has been localized to chromosome 5p (it had pre­viously been localized to chromosome 19), and a new gene, SLC6A19, a sodium-dependent and chlorideindependent neutral amino acid transporter, has been suggested as the defective gene by two separate groups.242,243 This transporter has been shown to be expressed in the intestine and has properties of system B0. System B0 refers to a broad range of amino acids with neutral (0) charge. SLC1A5 is the proposed ASC carrier for the neutral amino acids alanine, serine, and cysteine.244 Whereas the SLC1A1 carrier cotransports 3 Na+ and 1 H+ with countertransport of 1 K+, the SLC1A5 transporter mediates Na+-dependent transport. The B0,+ is a Na+-independent transport system that recognizes neutral and cationic amino acids in addition to the disulfide amino acid cystine. It is a heterodimer, consisting of a light and heavy chain, with the genes responsible for each chain being found on chromosomes 2 and 19, respectively. This transporter is defective in cystinuria. Several hormones have been shown to alter the amino acid and peptide transport process in the intestine. Somatostatin and vasoactive intestinal polypeptide decrease these transport processes, whereas EGF, neurotensin, cholecystokinin, and secretin enhance them. Human Pept1 appears to be inhibited by protein kinase C245 and cyclic adenosine monophosphate (cAMP).246 The expression of the intestinal peptide transporter is also modulated by dietary protein content.247 Even though the peptide transporter is expressed along the entire small intestine, diet-induced changes in the expression of the transporter are specific to certain regions. A high-protein diet increases the steady state levels of the transporter-specific messenger RNA in the middle and distal regions of the small intestine. The expression of the brush border peptidases dipeptidylcarboxypeptidase and dipeptidylaminopeptidase IV, which release dipeptides from oligopeptides, also are enhanced by a highprotein diet.

Exit from the Epithelium

Exit through the basolateral membrane operates via a number of different mechanisms that involve active transport and diffusion of both facilitated types.248 Active, sodium-dependent processes exist at this membrane for the uptake of neutral amino acids, which presumably supply nutrients for crypt cells and for villus enterocytes during fasting when a luminal source is unavailable (see Table 100-9). Villus enterocytes normally receive the amino acids necessary for production of their own protein from luminal nutrients; crypt cells obtain their supply from the portal circulation. Of all the amino acids, glutamine appears to be a unique and major source of energy for enterocytes; ammonia is an important metabolic byproduct of this process. Active uptake of glutamine at the basolateral membrane, as well as via apical membrane processes, is therefore of particular importance. It has been estimated that approximately 10% of amino acids are used in the production of enterocyte protein. Some

of these proteins are secreted across the basolateral membrane specifically by villus enterocytes, including apo A-I and apo A-IV, secretion of which increases many-fold after a fatty meal.48 The intestinal basolateral membrane possesses a set of amino acid transport systems that differ from those in the brush border membrane. The amino acid transport systems in the basolateral membrane function to export amino acids from the enterocytes into the portal circulation during feeding. They also participate in the import of amino acids from the portal circulation into the enterocyte for cellular metabolism when amino acids are not available from the intestinal lumen, such as between meals. The intestinal basolateral membrane also possesses a peptide transporter system that is probably identical to that in the brush border membrane and that facilitates the exit of hydrolysisresistant small peptides from the enterocyte into the portal circulation. Several well documented amino acid transport systems have been described in the basolateral membrane. System y+L is the amino acid exchanger that permits Na+independent efflux of cationic amino acids from intestinal cells into the blood coupled to the Na+-dependent influx of neutral amino acids from the blood into intestinal cells. System A is a Na+-coupled transport system for neutral amino acids, including glutamine, that plays a role in the entry of amino acids from the blood into intestinal cells for cellular metabolism. This Na+-coupled neutral amino acid transporter (SNAT) consists of three subtypes, SNAT1, 2, and 4; SNAT2 is expressed in the small intestine.249 Very small amounts of dipeptides have been detected in the portal circulation after a meal, but the great majority of absorbed products of protein digestion that reach the circulation are in the form of single amino acids. A somewhat surprising finding is that digestion of protein continues into the ileum, with approximately 40% of ingested protein undergoing transport in this segment of small intestine.250

VITAMINS WATER-SOLUBLE VITAMINS

Although it had been thought that the absorption of watersoluble vitamins depended simply on passive diffusion across the intestinal mucosa, there has been increasing recognition of the importance of specific carrier-mediated processes for this process (Table 100-10). Several of these vitamins are present in the diet as conjugates or coenzymes that require hydrolysis before or during their absorption.

Ascorbic Acid (Vitamin C)

Although most species synthesize all their vitamin C requirements, primates, guinea pigs, and some birds have lost this capacity and thus depend on diet for their needs. Vitamin C is found in a wide range of foods, but the most abundant sources are fresh fruits and fruit juices. Black currants are particularly rich (200 mg/100 g) in vitamin C; apples and pears less so (5 mg/100 g). Of animal sources, raw liver contains about 20 mg/100 g, and milk contains about 2 mg/100 g; fresh meat contains only traces of vitamin C. The ingestion of as little ascorbic acid as 10 mg/day prevents scurvy, and the recommended daily intake is 40 mg.145 Absorption of ascorbic acid decreases with increased intake and varies from 16% at high (more than 10 g) to 98% at low (less than 20 mg) intakes. From what is known, 80% to 90% of dietary ascorbic acid is absorbed.

1717

1718

Section X  Small and Large Intestine Table 100-10  Water-Soluble Vitamins VITAMIN

Recommended dietary allowances*

Ascorbic acid Folic acid

75-90 mg/day 400 mg/day

Cobalamin (B12)

2.4 µg/day

Thiamine Riboflavin Pantothenic acid Biotin

1.1-1.2 mg/day 1.1-1.3 mg/day 5 mg/day 30 µg/day

Pyridoxine Niacin

1.3-1.7 mg/day 14-16 mg/day

TRANSPORT MECHANISM Active; Na+-dependent process at brush border membrane Hydrolysis of dietary polyglutamates by folate conjugase at brush border membrane; Na+-dependent active transport or facilitated diffusion of monoglutamate at brush border membrane Intrinsic factor binding; uptake of intrinsic factor–B12 complex at ileal brush border membrane by way of specific receptor Na+-dependent active transport Absorption includes hydrolytic and phosphorylation steps Absorption includes hydrolytic and phosphorylation steps Saturable facilitated diffusion (at low concentrations); nonsaturable linear diffusion (at high concentrations) Simple diffusion ?

*Daily Reference Intakes (DRI) were established by the Institute of Medicine between 1997-2001. They are quantitative estimates of nutrient intakes to be used for planning and assessing diets for healthy people. The DRIs include both recommended intakes and tolerable upper intake levels. The RDAs (Recommended Dietary Allowances) are a component of the DRIs and are defined as the daily intake of a nutrient considered sufficient to meet the requirements of 97% to 98% of adults.

Cooking destroys some of the ascorbic acid contained in food, but such destruction can be minimized by shortening cooking times and not keeping foods hot for prolonged periods before they are eaten. Prolonged storage of foods also depletes vitamin C content. With loss of the capacity for hepatic synthesis of vitamin C, a specific absorptive mechanism has developed in humans (and guinea pigs). Transport across the apical membrane of small intestinal enterocytes occurs by an active Na+-dependent process.251 The active absorption mechanism becomes saturated when the mucosal concentration of vitamin C is greater than 6 mmol/L; this might account for the decreased fraction of dietary vitamin C absorbed with increasing intake. Electrically neutral, uphill transport of vitamin C probably occurs in the form of sodium ascorbate, via a concentrative, carrier-mediated, Na+-dependent mechanism. Two distinct isoforms of these Na+-dependent vitamin C transporters have been described in humans: hSVCT1 and hSVCT2.252,253 The intestinal ascorbic acid uptake process is regulated by extracellular substrate levels and by an intracellular protein kinase C–mediated pathway.254 Flavanoids in fruit and vegetables can decrease vitamin C absorption by inhibiting the SVCT1 transporter.255 A variable proportion of luminal vitamin C is present in the oxidized form as dehydroascorbic acid, which also is actively absorbed. Within the intestinal cell, dehydroascorbate is rapidly reduced back to ascorbic acid by the enzyme dehydroascorbate reductase, which requires reduced glutathione. It is through this mechanism that the intracellular level of dehydroascorbate is believed to be maintained at low, nontoxic levels. Ascorbic acid then circulates unbound in plasma and is transported against a gradient into target tissues by SVCT2.

Folic Acid

Folic (pteroyl monoglutamic) acid consists of the complex pterin molecule conjugated to para-aminobenzoic acid and glutamic acid. Although considerable amounts of dietary folate are in the form of polyglutamates with at least six glutamic acid residues, much is present as formyl- and methylhydrofolate. The folates are distributed widely in the diet, particularly rich sources being spinach (200 mg/100 g), liver (140 mg/100 g), and peanuts and beans (100 mg/100 g). Meat, chicken, potatoes, and fruit (except orange juice) are poor sources (less than 15 mg/100 g). Prolonged cooking of

food destroys its folate. Recommended dietary intakes are on the order of 200 mg/day in adults and 400 mg/day during pregnancy.145 Unlike vitamin B12, there is no vast storehouse of folate. The adult body contains only 2 to 3 mg of folate; therefore, in cases of poor intake (or malabsorption of folate) for just a short period, folate depletion occurs relatively rapidly. Absorption of dietary polyglutamates depends on hydrolysis to monoglutamate at the brush border membrane followed by transport into the cytoplasm.256,257 The apical membrane hydrolase (conjugase) in human intestine is expressed predominantly in the proximal jejunum and is a folylpoly-γ-glutamate carboxypeptidase (GCP2, or folate hydrolase) that cleaves off a single glutamic acid residue at a time. This brush border form of folate hydrolase has been cloned and has been shown to be up-regulated in dietary folate deficiency.258,259 GCP2 is a 120-kDa membrane protein that contains a single 5′ membrane-spanning domain. It is an exopeptidase with an optimal pH of 6.5 that releases intermediate products of folylpolyglutamate ending in the folylmonoglutamate derivative.258 Also recognized is a cytoplasmic folate hydrolase, an endopeptidase prominent in several species and present in humans; its role in humans is uncertain. Uptake is achieved by the reduced folate carrier, a specific concentrative, carrier-mediated, Na+-dependent, pH-sensitive process that is active at acid pH.260 It is inhibited by diphenylhydantoin and sulfasalazine, which also reduces hydrolysis. Prolonged exposure to ethanol inhibits hydrolysis (but not uptake), and this may be relevant to the folate deficiency sometimes found in alcoholics. As folate is transported through the enterocyte, pteroylglutamate is reduced and methylated, and 5-methyltetrahydrofolate (5-MTHF) is transported across the intestinal basolateral membrane by the reduced-folate carrier to the portal vein and the liver. The intestine also is exposed to a second source of folate: the folate that is synthesized by the normal microflora of the large intestine. Significant amounts of this folate source have been shown to exist in the absorbable monoglutamate form. The colon is capable of absorbing some of this folate.261

Cobalamin (Vitamin B12)

Cobalamin exists largely as hydroxycobalamin, methylcobalamin, and adenosylcobalamin, and these are found almost entirely in animal sources: Liver, kidney, beef, fish,

Chapter 100  Digestion and Absorption of Nutrients and Vitamins Salivary R Food-bound protein vitamin B12 Gastric parietal cell HCI

Stomach

IF Pancreatic zymogen cell

Pancreas

Trypsin

Liver Ileal enterocyte

Terminal ileum

Portal venous system Transcobalamin II Figure 100-18.  Steps in the chain leading to the binding of vitamin B12 to intrinsic factor (IF). Food-bound B12 is released by pepsin working at an acid pH and is picked up preferentially by salivary R protein in the stomach. Proteolysis of R protein by duodenal trypsin releases B12 for binding to IF. The subsequent binding and uptake of the IF-B12 complex occurs via a specific receptor-mediated process on the brush border membrane of ileal enterocytes. Vitamin B12 is released at an intracellular site, transported across the basolateral membrane, and there taken up by transcobalamin II for transport into the portal circulation.

eggs, and milk provide most of the cobalamin in a normal diet.48 Vegetables are almost entirely lacking in vitamin B12, and therefore the strict vegan’s dietary intake of cobalamins may be inadequate. About 10 to 20 µg is ingested per day in an average diet, and of this, about 1 to 2 µg/day is required to provide for normal needs.262 Because cobalamin absorption declines with age, persons older than 50 years of age are advised to take supplemental vitamin B12. Three types of binding proteins are concerned with the absorption of cobalamin: one in saliva, one in gastric juice, and one in the circulation.262 The vitamin is released by mastication and by gastric acid from the various dietary proteins with which it is associated (Fig. 100-18), after which the first specific binding protein secreted into saliva and also by parietal cells, the R protein (haptocorrin), takes up the free cobalamin and binds it with strong affinity. At intragastric pH values below 3, intrinsic factor (IF) has much weaker affinity for the vitamin than does R protein.263 It is only in the duodenum, where the R protein is hydrolyzed by pancreatic enzymes, that IF can bind the cobalamin that has been released.264 In humans, IF is secreted by parietal cells; in rats and mice it is secreted by chief cells. IF release is stimulated in response to the same agonists that stimulate acid secretion: histamine, gastrin, and cholinergic agonists. Unlike R proteins, which can bind a wide variety of cobalamin analogs, IF is much more selective and specific for cobalamin. It has been suggested that the nonspecificity of binding to the R protein that exists in plasma might offer an advantage in

binding potentially harmful compounds.265 IF has a very strong affinity for cobalamin and binds it tightly by enclosing the vitamin in its cup-like interior. This complex resists pancreatic proteolysis by undergoing molecular conformational changes and glycosylation, passes down the intestine to the terminal ileum, and there binds to specific receptors on ileal enterocytes, called cubulin-amnionless (AML). Distribution of ileal cubulin-AML receptors is patchy,266 and estimates suggest that there are about 300 to 400 receptors per enterocyte, or one per microvillus, located deep between microvilli.267 The number of receptors available determines how much vitamin can be absorbed; absorption doubles during pregnancy by a doubling of the number of available receptors.268 After binding to the receptor, the IF-cobalamincubulin-AML complex probably enters the cell intact by translocation (see Fig. 100-18). B12 accumulates in the mitochondria, and the complex is split at some point within the enterocyte. Free cobalamin leaves the base of the cell, where it is immediately bound to an ileal pool of trans­ cobalamin II, a 43-kDa protein that is synthesized in the enterocyte, and which transports it into the portal circulation. The transcobalamin II–cobalamin complex is essential for the transport of cobalamin into all cells in the body, where, after uptake by endocytosis, the cobalamin is enzymatically released. It is clear that this complicated series of events can be interrupted at a number of different points in the pathway. Lack of pancreatic proteolysis leads to a defect in the release of the vitamin from the R protein for subse-

1719

1720

Section X  Small and Large Intestine quent IF binding, lack of IF fails to provide the complex necessary for binding and absorption at the ileal mucosa, loss of ileal receptors prevents absorption, and defects within ileal enterocytes could prevent release of the vitamin into the circulation.

Other Water-Soluble Vitamins48,269

Thiamine Thiamine (vitamin B1) is distributed widely, but the only important dietary sources are seeds of plants. Germs of cereals, nuts, peas, beans, and lentils are major sources, and green vegetables and fruit are relatively poor ones. White flour or bread and purified rice have virtually no thiamine. Thiamine is water soluble and is readily lost in water used for cooking. Thiamine occurs almost entirely in phosphorylated form in animal foods, but it exists in its free form in foods of plant origin and enriched cereal. Dephosphorylation results from the action of phosphatases, present in the intestinal lumen, before it is absorbed. The rate of thiamine absorption approaches a limit as the dosage increases. This absorption is greatest in the jejunum and ileum and can be active or passive, depending upon concentration of the vitamin. With low intraluminal concentrations, thiamine is absorbed by a Na+-independent active process. Two human thiamine transporters, SLC19A2 and SLC19A3, have been cloned.270-273 At high concentrations (more than 8 mg in a single dose), thiamine is absorbed by means of passive diffusion.274 After absorption, thiamine is phosphorylated in the enterocyte and then transported out of the cell, possibly by the basal membrane Na+,K+ATPase. This exit step is inhibited by ethanol, which, in addition to poor intake, contributes to thiamine deficiency in alcoholics. Niacin Niacin (nicotinic acid) and nicotinamide are distributed widely in foods, but their availability in foods varies. About half of the North American and European dietary intake of niacin is supplied in meat and fish. The niacin content of legumes, however, is largely bound and unavailable, although it can be released by treatment with alkali; unfortunately, the food preparation methods widely employed in Asia and Africa, such as for maize, do not render the niacin available. Nicotinic acid can be synthesized in humans from tryptophan, 60 mg of tryptophan being required for the synthesis of 1 mg of niacin. The concept of the nicotinic acid equivalent has thus arisen, and foods lacking in niacin nonetheless can remain valuable in preventing pellagra because of high tryptophan content; such is the case with milk and eggs. Both nicotinic acid and nicotinamide are absorbed rapidly from the stomach and the small intestine. At low concentrations, absorption occurs through Na+-dependent facilitated diffusion, but at higher concentrations, passive diffusion predominates. Niacin occurs mostly in its coenzyme forms, which are hydrolyzed in the intestinal lumen by pyrophosphatase to nicotinamide; nicotinamide seems to be absorbed as such, without further hydrolysis in the gastrointestinal tract. Riboflavin Riboflavin (vitamin B2) is linked with phosphoric acid in most animal and plant tissues to form flavin mononucleotide and with adenosine monophosphate (AMP) to form flavinadenine dinucleotide. The richest dietary sources are liver, eggs, milk, green vegetables, and beer. Riboflavin also is synthesized by colonic bacteria, but its availability from the colon is uncertain. Cooking does not destroy much ribo-

flavin, but exposure to sunlight might. Riboflavin is presented to the mucosa in the form of coenzymes, so it is necessary for these to be hydrolyzed by intestinal phosphatases at the brush border membrane before active transport occurs into the cell. Studies suggest that riboflavin uptake occurs mainly in the proximal portion of the small intestine and involves a specialized Na+-independent carrier-­mediated system. This system is regulated by extracellular substrate levels and by specific intracellular protein kinase–mediated pathways.275,276 Once within the cell, rephosphorylation occurs. Biotin Biotin is so widely available that spontaneous deficiency states in humans have been described only rarely. Many yeasts and bacteria contain biotin and can provide a sufficient supply in normal foods. Liver, legumes, nuts, and vegetables are reasonable sources. Little is known about the mechanism of absorption of biotin in humans. Animal studies have indicated that most absorption of the free vitamin occurs in the jejunum and upper ileum by two mechanisms: a saturable facilitated diffusion process that operates at low concentrations and a nonsaturable linear diffusion process operating at higher luminal concentrations of the vitamin. Further studies have shown that the intestinal biotin transport system is also used by two unrelated micronu­ trients: the vitamin pantothenic acid and the metabolically important substrate lipoate. For these reasons, the biotin transport system now is referred to as the Na+-dependent multivitamin transporter (SMVT).277 It is unclear how cells regulate transport of the individual vitamin via this common transport system and how the substrate level of the individual substrate affects SMVT function in these cells. Pantothenic Acid Pantothenic acid usually is found as its calcium salt and is derived largely from animal tissues, especially liver, kidney, egg yolk, wheat germ, and peanuts. It is almost completely lacking in many processed foods but is not lost in normal cooking. The intestine also is exposed to bacterial sources of pantothenic acid. In the diet, pantothenic acid exists mainly in the form of coenzyme A, which is hydrolyzed to free pantothenic acid in the intestinal lumen before absorption. Hydrolysis is followed by transport of free pantothenic acid into the absorptive cells via the SMVT (see earlier).277 Pyridoxine Pyridoxine (vitamin B6) occurs in the diet in one of three forms: pyridoxamine phosphate, pyridoxal phosphate, and pyridoxine phosphate. Its presence is widespread in plant and animal tissues; cereals, peanuts, bananas, and liver are good sources. The phosphorylated form must be dephosphorylated before absorption can occur, and absorption is achieved by means of a membrane-bound alkaline phosphatase found in the intestinal brush border. All three forms of pyridoxine are freely absorbed by passive diffusion in the jejunum and ileum.

FAT-SOLUBLE VITAMINS

Vitamins A, D, E, and K are structurally different from one another, but all can be classified as polar, nonswelling, insoluble lipids. Although their chemical structures are known, the retention of a letter to signify their individuality is useful because each consists of a number of closely related compounds with similar properties (Table 100-11).48

Chapter 100  Digestion and Absorption of Nutrients and Vitamins β carotene

Vitamin A

Retinol R1 OH

CH3

R2

CH3

CH3

O R3

CH3 α-tocopherol

CH3 Vitamin E

CH3 7 dehydrocholesterol

O CH3

OH Ultraviolet light

CH2

CH

O

C

CH2

n

H

n = 4 to 13

CH3

Vitamin K2 (Menaquinones) O CH3 CH2

CH3

O

CH

C

CH2

CH2

CH2

CH2

CH2

CH2

3

H

CH2

Vitamin K1 (Phytomenadione)

OH

Vitamin K

Vitamin D

Figure 100-19.  Structural formulae of the fat-soluble vitamins A, D, E, and K. Vitamin D is shown as cholecalciferol (D3).

Table 100-11  Absorptive Mechanisms for Fat-Soluble Vitamins VITAMIN

Recommended dietary allowances*

MECHANISM OF ABSORPTION

A (retinol) D3 (cholecalciferol) E (α-tocopherol) K (phytomenadione [K1]284 and menaquinones [K2])

700-900 µg/day 10-15 µg/day 15 mg/day 90-120 µg/kg/day

Passive diffusion Passive diffusion Passive diffusion K1: carrier-mediated uptake K2: passive diffusion

*Daily Reference Intakes (DRI) were established by the Institute of Medicine between 1997-2001. They are quantitative estimates of nutrient intakes to be used for planning and assessing diets for healthy people. The DRIs include both recommended intakes and tolerable upper intake levels. The RDAs (Recommended Dietary Allowances) are a component of the DRIs and are defined as the daily intake of a nutrient considered sufficient to meet the requirements of 97% to 98% of adults.

Vitamin A

Vitamin A (retinol) is found in the diet in milk and milk products, egg yolk, and fish oils. Carotenoids are defined by their chemical structure: The hydrocarbon carotenoids are known as carotenes, and the oxygenated derivatives of these hydrocarbons are referred to as xanthophylls.

Beta carotene, the principal carotenoid in carrots, consists of two conjoined molecules of retinol and is a precursor of the active vitamin (Fig. 100-19). There are many other carotenoids in the diet, but these contain only one retinol molecule. Carotenoids are found predominantly in green vegetables and carrots, and in the United States and Europe

1721

1722

Section X  Small and Large Intestine these sources account for about half the dietary intake. Retinol and the carotenes are stable in normal cooking. Lycopene is the carotenoid that gives tomatoes and other red fruits and vegetables their color. Although it is similar in structure to carotene, lycopene lacks a β-ionone ring structure and thus cannot form vitamin A or retinal. Its biologic effects in humans are attributed to mechanisms other than those for vitamin A, Retinol and carotene are absorbed in the small intestine, carotene less readily than retinol.278,279 Dietary retinal (vitamin A aldehyde) esters first are hydrolyzed to retinol in the intestinal lumen before they are absorbed into the intestinal mucosa. Carotenes, by contrast, are converted into retinol, primarily in the enterocytes. Before entering the mucosal layer of the intestine, however, carotenes first are solubilized into micellar solution along with other fatsoluble compounds. Hence, bile salts and dietary fats are needed for absorption of carotenes in the upper region of the small intestine. The biosynthetic process for retinol from carotene in the enterocytes involves two soluble mainly cytosolic enzymes. The first enzyme is responsible for the oxidative split of the beta carotene molecule to produce the cleavage product of retinal, which is then reduced to retinol by a second enzyme that depends on reduced nicotinamide adenonine dinucleotide (NADH). Transport across the apical membrane appears to occur by passive diffusion, but facilitated diffusion cannot be excluded. Free retinol (from retinyl ester and carotenes) in the mucosal cells is re-esterified mainly with palmitic acid before incorporation into chylomicrons, which is how vitamin A mostly leaves the mucosa.

Vitamin D

Vitamin D comprises a group of sterols that have antirachitic properties, but the only two nutritionally important members are vitamins D2 (ergocalciferol) and D3 (cholecalciferol). Both are produced by ultraviolet irradiation of their precursor sterols, ergosterol and 7-dehydrocholesterol, respectively (see Fig. 100-19). Ergosterol, found in fungi and yeasts, is an unusual constituent of the normal diet, whereas vitamin D3 is the major dietary form. Vitamin D3 is found in a restricted range of foods, predominantly the oils of fatty fish, which themselves ingest it in plankton found near the surface of the sea.48 Human breast milk contains sufficient vitamin D to prevent rickets, but cow’s milk is a poor source of this vitamin. Most of a person’s requirement for vitamin D, however, is supplied by endogenous synthesis in the skin during exposure to sunlight, and dietary intake becomes critical only when such exposure is inadequate (see Table 100-11). As with vitamin A, vitamin D absorption occurs by simple passive diffusion in the small intestine.280 Bile salts are unnecessary, but luminal pH influences absorption. Absorption is reduced at neutral pH and increased in an acidic milieu.281 Most absorbed vitamin D passes into the lymphatics unchanged in chylomicrons. Both vitamins D2 and D3 are converted in the liver to 25-hydroxycholecal­ciferol, which is then further converted to 1-α,25-dihydroxycholecalciferol (1,25(OH)2D3) by renal 1-hydroxylase. 1,25(OH)2D3 functions to further promote intestinal calcium absorption.

Vitamin E

Vitamin E is still seeking a role in humans. It comprises a group of eight or so tocopherols, the most potent of which in animals is α-tocopherol (see Fig. 100-19).106 It is distrib-

uted widely in the diet: vegetable oils, cereals, eggs, and fruit are good sources. Margarines are particularly rich in vitamin E, and breast milk contains much more than cow’s milk. Although a variety of diseases can result from deficiency of vitamin E in a number of animal species, it has proved difficult to ascribe a human disease to vitamin E deficiency. Vitamin E is absorbed passively across the intestinal mucosa.282 The ester form, in which many vitamin preparations are presented, is hydrolyzed by pancreatic and/or duodenal esterases before absorption, but the ester can be absorbed intact.283 After incorporation into micelles, vitamin E is transported into enterocytes and incorporated into chylomicrons for transfer into lymphatics.

Vitamin K

Vitamin K is found in two forms: K1, derived largely from plants, is phytomenadione; K2 comprises a group of bacteria-produced compounds, the multiprenyl menaquinones. K1, the major dietary form, is found in green vegetables, but beef liver is another good source. K2 is produced by colonic bacteria, and although some K2 may be absorbed from the colon, this alone is an inadequate source if K1 absorption is impaired. Absorption of K1 from the small intestine is dependent on luminal bile salts, and uptake is achieved by a carrier-mediated process,284 whereas K2 absorption is entirely passive.285

MINERALS AND TRACE ELEMENTS Various divalent ions are essential nutrients; some are absorbed in milligram amounts and are major constituents of the body; others are necessary only in trace amounts. Iron, calcium, magnesium, phosphorus, and sulfur are in the former category, and specialized absorptive mechanisms are concerned with their assimilation.

CALCIUM

Milk and other dairy products are the most valuable sources of calcium, accounting for up to 75% of dietary calcium intake; cereals, legumes, and other vegetables contribute lesser amounts. Phytic acid or oxalate in vegetables binds strongly to calcium, thus reducing its availability. Dietary fiber also binds calcium and can interfere with its absorption; by contrast, dietary lactose enhances its absorption.286 Fractional, or true, absorption is only about 20% to 30% of total dietary calcium, the remainder being excreted in stool. Absorption of calcium across the intestinal mucosa is achieved by two parallel processes: an active, transcellular transport process, which dominates with lower levels of calcium intake, and a passive, paracellular diffusive process, which becomes more important at higher levels of calcium intake.287-289 Under normal dietary conditions, the duodenum is the major site for active calcium transport, whereas passive, paracellular transfer occurs throughout the small intestine. Despite this localization of the active transport site, quantitatively more calcium may be absorbed in the jejunum and ileum than in the duodenum because of the relative amounts of time luminal contents spend in these regions of intestine. The human jejunum absorbs calcium faster than does the ileum, and absorption rates in both regions are increased by treatment with vitamin D.290 The paracellular route, via the tight junctions, may be capable of modifying calcium transport because passive transport increases in response to treatment with vitamin

Chapter 100  Digestion and Absorption of Nutrients and Vitamins 1,25-(OH)2 vitamin D

Ca2+ Protein transcription

ADP ATPase ATP

Ca2+ Ca2+

Ca2+- Calbindin

Figure 100-20.  Mechanisms of calcium transport across the intestinal epithelium. A paracellular route allows bidirectional flux. Transport into the epithelial cell occurs via specific channels down an electrochemical gradient. A critical step is the binding to calbindin, which then presents calcium for export via a calcium-dependent adenosine triphosphatase (ATPase) on the basolateral membrane. Each of these processes appears to be influenced by 1,25-(OH)2 vitamin D, although its maximal effect is on synthesis of fresh calbindin. ADP, adenosine diphosphate; ATP, adenosine triphosphate.

D.291 Furthermore, there is evidence suggesting that tight junctional permeability increases during sugar transport, and this might provide another mechanism for control of paracellular transport.292 The transcellular route involves transport across the apical membrane, transfer across the cytoplasm, and exit across the basolateral membrane (Fig. 100-20). Entry pro­ bably occurs via specific non–voltage-gated calcium channels in the apical membrane and down the prevailing electrochemical gradient. Within the cytoplasm, binding to a calcium-binding protein, calbindin D9K (or calbindin 3), is a key step.293 Maximal transport rates correlate closely with calbindin concentrations. This protein, present in concentrations of 0.1 to 0.2 mmol, must rapidly take up the calcium entering the cell because intracellular free calcium concentrations are carefully maintained at very low values (about 10−7 M). Transient rises in intracellular calcium act as key second messenger signals for secretory responses in enterocytes. Absorbed calcium thus is pre­ sumably segregated from the calcium concerned with cell signaling, and calbindin D9K plays a vital role here by bringing calcium to the transporter at the basolateral membrane.294 Another calcium-binding protein, calbindin D28K (CALB1) is induced by vitamin D and binds four calcium ions compared with calbindin D9K, which binds two calcium ions. An active mechanism is then necessary to drive calcium uphill against the electrochemical gradient, for which a calcium-dependent ATPase is responsible.293 Calcium arrives at the basolateral pole bound to a site at the cytoplasmic aspect of the calcium-dependent ATPase that spans the basolateral membrane. There follows a phosphorylation-induced change in the conformation of the calcium-dependent ATPase, and the calcium ion is extruded through the channel formed by the enzyme transmembrane elements.295 The rate-limiting step in the absorption process of calcium is the intracellular calbindin concentration, which is regulated by a metabolite of vitamin D, 1,25-dihydroxyvitamin

D (1,25-[OH]2D), produced in the kidneys from 25-hydroxyvitamin D (25-[OH]D); the latter is converted from absorbed vitamin D by the liver.288 The vitamin also has a modest effect on the calcium-entry step and enhances activity of the basolateral calcium ATPase. Up-regulation of the calbindin gene in response to vitamin D occurs largely in villus cells.296 Some evidence supports colonic absorption of calcium, which also can be enhanced in response to vitamin D.297 Although normally the colon can account for only up to 7% of total calcium absorption, it becomes an important route for calcium absorption in patients with short bowel syndrome.298 Active duodenal calcium absorption is increased in calcium deficiency states and reduced in calcium repletion states. Increased production of the active 1,25-(OH)2 vitamin D metabolite in response to a small drop in plasma calcium concentration is responsible for increasing calcium absorption, and this change occurs within a day of changing from a high-calcium to a low-calcium diet.299 This same mechanism is likely to be the cause of the enhanced calcium absorption seen during late pregnancy and lactation. At birth, the active vitamin D–dependent absorptive mechanisms already are present in the human duodenum. Ingestion of large amounts of calcium, together with lactose, in breast milk ensures adequate intake at this critical stage of life. Calcium absorption declines with age, but this might result in part from a lack of vitamin D, or decreased responsiveness of the intestine to vitamin D.286

MAGNESIUM

An average diet provides about 300 to 500 mg of magnesium per day from a wide range of vegetables. Magnesium absorption has been less thoroughly investigated than that of calcium, but it seems likely that the mechanisms involved are different. In contrast to calcium absorption, magnesium absorption in the basal state is greater in the human ileum than in the duodenum or jejunum.290 Jejunal absorption of magnesium is increased by vitamin D, whereas ileal absorption is not. Ileal transport involves both a paracellular, diffusive pathway and a transcellular, carrier-mediated, saturable process.300 There is some competition from calcium for the diffusive pathway but not for the saturable, presumably carrier-mediated process.300 Quantitatively, magnesium fluxes across ileal mucosa are several-fold greater than those for calcium, but the overall efficiency of magnesium absorption after normal dietary intake ranges from 21% to 27%.

IRON

Meat-eating, affluent societies ingest about 20 to 30 mg of iron per day, largely as myoglobin or hemoglobin. Vegetarian societies in poor countries ingest much less than this in wheat and vegetables, and iron in these foods is less readily available for absorption. A careful balance of absorption and loss is maintained in normal adults: each is about 1 mg/day. Developing children and adolescents need to absorb about 0.5 mg/day more, to build up total body iron to adult values. Iron is present in breast milk in the form of lactoferrin, for which a specific brush border membrane receptor has been demonstrated301,302 that facilitates iron absorption in neonates. During reproductive life, normal women need to compensate for menstrual losses, which are on the order of 5 to 50 mg per month and approximately 500 mg for each pregnancy. Because dietary intake often markedly exceeds the body’s need for iron, it is necessary to absorb only a small portion

1723

1724

Section X  Small and Large Intestine of that ingested. Overall, there is a positive and linear relationship between the amount ingested and that absorbed, but the fraction absorbed decreases as more is taken in.303 Total body iron content is regulated by controlling the level of iron absorbed from the diet. Under normal circumstances, only about 10% (1 to 2 mg/day) of dietary iron is absorbed. Most absorption occurs in the proximal small intestine, and the ferrous (Fe2+) form is absorbed better than the ferric (Fe3+) form. The latter is insoluble at pH values greater than 3, and gastric acid and some sugars and amino acids render it more available for absorption. The presence of some anions, such as oxalate, phosphate, and phytate, precipitate iron out of solution and reduce its absorption. The presence of bile enhances absorption, but the mechanism of this enhancement is unclear. Dietary iron is predominantly found in the ferric form, but Fe3+ is highly insoluble under physiologic conditions. Therefore, during uptake, Fe3+ is converted to the Fe2+ form at the apical cellular membrane before it attaches to an acceptor protein in the membrane. The ability of intestinal mucosa to reduce Fe3+ to Fe2+ has been documented,304 and a ferrireductase activity has been characterized for the intestinal Caco-2 and HuTu-80 cell lines.304,305 A functional role for Fe3+ reduction in iron transport across the brush border is implicated by the fact that inhibition of ferrireductase activity reduces Caco-2 cell apical iron uptake. Increased ferrireductase activity correlates with enhanced iron uptake induced by iron deficiency and hypoxia.305,306 Iron uptake into the body occurs at two interfaces of the intestinal epithelium: the apical and basolateral plasma membranes.307 The apical plasma membrane of the differentiated enterocyte is specialized for transport of heme and ferrous iron into the cell. Three major pathways of iron transport across the apical membrane have been proposed. The best-characterized pathway is via the divalent metal transporter 1 (DMT1, also known as Nramp2 or DCT1),308 which is located in the brush border membrane, primarily in the villus tip cells. There are two splice variants of DMT1 that yield two messenger RNAs (mRNAs): one containing an iron-responsive element (IRE) termed DMT1 (IRE) mRNA, and another without an IRE designated DMT1 (non-IRE) mRNA. DMT1 is a proton symporter that transports ferrous iron and other divalent metals from the intestinal lumen into the enterocyte (Fig. 100-21). It is up-

Fe 2+

Fe 2 + DMT1 Fe 2 +

Fe-BP F e 2+

Fe 3 + Heme

Fe 2+ + Transferrin

Heme-Oxygenase Heme

Figure 100-21.  Mechanisms of iron transport in the intestine. A small amount of inorganic iron may pass through the paracellular route. Inorganic iron (Fe3+) is converted into its ferrous form (Fe2+) at the brush border membrane before transport into the cell. Heme iron is transported into the cell by a separate mechanism. Within the cell, one or more ironbinding proteins take up iron and transfer it to the basolateral membrane for delivery across the membrane and subsequent binding to transferrin. DMT1, divalent metal transporter 1; Fe3+, ferric; Fe2+, ferrous; Fe-BP, ironbinding protein.

regulated during iron deficiency and down-regulated in cases of iron excess. In order of substrate preference, DMT1 can mediate import of Fe2+, Zn2+, Mn2+, Co2+, Cd2+, Cu2+, Ni2+, and Pb2+. The idea that the transporter responsible for dietary iron absorption recognizes other divalent cations agrees well with observations that Zn2+, Mn2+, Cd2+ and Cu2+ all can inhibit this process. DMT1 mRNA is found in many different tissues, but the protein and its mRNA are most abundant in the proximal duodenum, with decreasing absorption along the distal axis, consistent with a function in intestinal iron absorption.309 Iron depletion results in increased DMT1 mRNA levels in the intestine, which suggests that IREs in its 3′ untranslated region bind and stabilize the DMT1 mRNA.309 Although the major route for dietary iron absorption likely is mediated by DMT1, this transporter is found only in the apical surface of enterocytes. Thus, other factors must be involved in the transfer across the intestinal epithelium. DMT1 may be involved in the pathogenesis of hereditary hemochromatosis. HFE is the gene responsible for hereditary hemochromatosis. HFE protein is found in the crypt cells of the duodenum associated with β2-microgobulin and transferrin receptor. It is hypothesized that HFE protein facilitates transferrin receptor–dependent iron uptake into crypt cells and that mutant HFE protein might lose this ability, leading to a relative iron deficiency in duodenal crypt cells. In turn, this might lead to an increase in the expression of DMT1, resulting in increased iron absorption in hereditary hemochromatosis. Up-regulation of DMT1 expression has been confirmed in the HFE-knockout mouse and in humans with hereditary hemochromatosis (see Chapter 74).310 Iron also can be absorbed in the form of heme iron (in hemoglobin and myoglobin), which is readily transported across the brush border of the enterocytes as an intact heme moiety. It is the presence of globin, which increases the absorption of iron in this form. Heme-carrier protein 1 (HCP1) has been isolated from the mouse duodenum. This large hydrophobic transporter is present in the apical membrane during iron deficiency and in the cytoplasm during iron overload, making it a putative candidate for heme-iron transport.311 Once within the cell, heme is broken down by heme oxygenase, and the iron is released into the nonheme pool for incorporation into intracellular ferritin and export out of the cell.312,313 Another iron absorptive pathway has been proposed involving intestinal mucins, a 56-kDa protein designated mobilferrin, an integrin, and a ferric reductase. This pathway is regulated and depends on metabolic energy, and it appears to be encouraged by nonessential fatty acids.305,314,315 A small proportion of the iron crossing the mucosa uses a paracellular route by simple diffusion (see Fig. 100-21). Once within the enterocyte, the expression of the iron-storage protein ferritin is regulated by the intracellular concentration of iron. Ferritin synthesis increases when iron is present in excess and decreases when iron level is low. Iron transport across the basolateral plasma membrane of villus enterocytes involves at least two proteins: a ferroxidase called hephestin, which is associated with the basolateral membrane, and a basolateral iron transporter termed ferroportin 1 (FPN1), iron-regulated protein 1, or metaltransport protein 1. These proteins may work in close conjunction with each other, with ferroportin 1 transporting ferrous iron out of the cell and hephestin oxidizing the ferrous iron to ferric iron, which permits the avid incorporation into circulating apotransferrin (apoTF).307 The transferrin receptor, in combination with the hemochromatosis

Chapter 100  Digestion and Absorption of Nutrients and Vitamins protein (HFE), allows the binding of apoTF-bound iron and its reuptake back into intestinal cells.316 FPN1 is a membrane-bound protein containing iron-responsive elements that is up-regulated during iron deficiency and down-regulated during iron excess. Hepcidin, which is produced in the liver, might influence enterocyte iron transport by binding to FPN1, resulting in the internali­ zation and degradation of FPN1, and so reducing iron absorption.317 Intestinal iron absorption is regulated in at least three ways. The longstanding concept of mucosal block is based on the observation that after a large oral iron dose, enterocytes do not absorb additional iron for several days. A second regulatory mechanism of iron absorption is termed the stores regulator. It acts on a pathway that facilitates a slow accumulation of nonheme dietary iron. The functioning of the stores regulator is of great physiologic importance, because it prevents iron overload after ensuring iron needs are met. The exact molecular mechanism of the stores regulator has not been established, but it has been proposed to involve soluble factors such as transferrin-bound iron, serum ferritin, serum transferrin, or hepcidin. The erythropoietic regulator is a third regulatory mechanism that adjusts intestinal iron absorption in response to the demands of erythropoiesis, independent of body iron stores. This regulator must signal directly between the hematopoietic bone marrow and the duodenum. Although the erythropoietic regulator has been proposed to be a soluble component of the plasma, it is distinct from the stores regulator. This is evidenced by the rate of iron uptake in anemic persons that is much greater (20 to 40 mg/day) than could be produced by the stores regulator alone. The stores and erythropoietic regulators are circulating factors that maintain iron homeostasis of the entire organism. Within individual cells, the iron regulatory proteins (IRPs) IRP-1 and IRP-2 act to control iron availability by translational control of the synthesis of proteins such as transferrin and ferritin. IRPs are cytoplasmic RNA-binding proteins that function on mRNAs that contain IREs. Functional IREs are present in the 3′ untranslated region of mRNAs for transferring in one of the two isoforms of DMT1 (DMT1 IRE) and in the 5′-untranslated region of mRNAs for ferritin, ferroportin 1, mitochondrial aconitase, and the erythroid-specific form of δ-aminolevulinic acid synthase. IRPs functionally connect intracellular iron availability with cellular iron utilization; IRP function also can be altered by inflammation and oxidative stress. Basolateral iron uptake from the plasma by cryptal enterocytes plays an important role in sensing body iron stores. Indeed, there is considerable evidence that the iron concentration within the cryptal enterocyte is an important determinant of iron absorption. The mechanism by which the intracellular iron concentration can respond to body iron needs is poorly understood. It is clear, however, that cells in the crypts of Lieberkühn always express transferrin, and the endocytic mechanism imparts information about body iron storage based on plasma transferrin saturation. It is also well recognized that acute changes in body iron status, whether overload or deficiency, are not reflected by changes in iron absorption for a period of two to three days. This lag response time probably correlates with the migration time for proliferating cells in the crypts to differentiate and migrate into functional, mature enterocytes of villi. Thus, the luminal epithelial cells may be preprogrammed in the crypts based on body iron needs. This preprogramming would, in turn, initiate synthesis of iron transport proteins that are required for dietary iron uptake across the membranes of the villus enterocyte.

TRACE ELEMENTS

The importance of zinc, copper, and iodine in human nutrition has long been recognized, and they have received increasing attention in recent years as their roles in defined enteral and parenteral forms of nutrition have been demonstrated. The value of selenium also has been emphasized, and the need for manganese and chromium is receiving attention. Despite this interest, surprisingly few systematic studies of their absorption have been undertaken.

Zinc

Zinc is present in the body in about half the amount of iron (about 2 g), and largely in a wide variety of enzymes. It also plays important roles in maintaining configuration of gene transcription proteins and the integrity of membranes. It is found particularly in meat, shellfish, cereals, and legumes. Daily requirements are approximately 12 to 15 mg/day in adults. Persons who consume a low-energy diet might take in marginal amounts of zinc, and requirements are increased during pregnancy and lactation. Absorption is impaired by phytates and oxalates in the diet through their chelating properties, and food processing can render zinc less available for absorption.48,318 The protein content of a diet is positively correlated with zinc absorption, likely because of amino acids or small peptides that facilitate enterocyte uptake of zinc. Overall, the efficiency of zinc absorption from regular diet is 15% to 35%. There is enterohepatic circulation of zinc, and reabsorption appears to be maximal in the distal small intestine.319 Studies with vesicles of porcine jejunal brush border membranes have identified two uptake processes: an active saturable, carrier-mediated process (which dominates at low or normal intake) and a nonsaturable, diffusive process (which contributes more to absorption at higher intake).320 The relative importance of each is not known. There are nine members of the ZNT family of zinc transporters, with ZNT1, 2, and 4 occurring mainly in the villus.321,322 ZNT1 is a ubiquitously expressed protein that is present in the villi of the proximal small intestine. In response to manipulation of dietary zinc, however, expression in rats was increased in response to zinc supplemen­ tation but not to zinc restriction.323 These and other observations have led to the consensus that ZNT1 functions mainly as a zinc exporter and might play a role in zinc homeostasis for zinc acquisition and elimination under conditions of excess zinc.321 ZNT2 and ZNT4 are involved in intracellular transport of zinc by the enterocyte. The role of metallothionein (MT), an intracellular metalbinding protein, in the regulation of zinc absorption, particularly in conjunction with the zinc transporters, also remains unclear. This binding protein may be concerned with zinc absorption because changing dietary loads of zinc rapidly affect protein synthesis and alter binding capacity.324 Subjects on a low-zinc diet respond by decreasing their urinary excretion rate of zinc and by increasing its absorption rates.319,325 Absorption increases in pregnancy and during lactation.326 In experiments with knockout and transgenic mice, the rise in serum zinc after a single dose of zinc was much greater than in the control animals. By contrast, the serum zinc response of the MT transgenic animals was blunted compared with that of the control animals. The expression of ZNT1 also was measured and found to be directly related to serum zinc levels but unaffected by MT levels.327 Thus, MT might function in cellular responses to limit free zinc concentrations within narrow ranges and function as a zinc pool.321,327 Another transporter potentially involved in zinc and other metal uptake is DCT1, a transmembrane polypeptide

1725

1726

Section X  Small and Large Intestine that is found in the duodenum in the crypts and lower villi and may be available for the uptake of several metal ions.322 The ZIP (Zrt-, Irt-like protein) family of proteins also are believed to be involved in zinc transport.328 ZIP4 and ZIP5 likely exist on the apical side and basolateral membrane of the enterocyte, respectively, and may be responsible for zinc transport into the circulation.329

Copper

Copper is found in green vegetables and fish, and the average Western diet provides 1 to 3 mg/day, which is adequate for a daily need of about 1 mg/day. Dietary copper is absorbed very efficiently from the stomach and small intestine, especially the duodenum. Although the precise mechanisms involved in copper absorption remain incompletely known, within physiologic ranges of intake, absorption is probably by active transport. Competition for absorption between copper and zinc or iron may be demonstrable with large doses of these elements but not with normal dietary intakes.330 The uptake of copper might increase in pregnancy.331 Active transport and passive diffusion are both responsible for copper absorption in humans. A putative high affinity protein copper transport protein, denoted hCtr1, has been identified by functional complementation of the respiratory defect in yeast cells defective in copper transport because of inactivation of both the CTR1 and CTR3 genes.332 Human Ctr1 is a 190–amino acid protein with three transmembrane domains and significant homology to yeast Ctr1 and Ctr3, suggesting that mammalian high-affinity copper transporters may have evolved from Ctr1 and Ctr3. RNA blotting analysis has demonstrated that that hCtr1 is expressed in all organs and tissues examined: Liver, heart, and pancreas exhibited the highest levels of expression; intestine had intermediate levels of expression; and expression in brain and muscle was low. Whether hCtrl1 plays an important role in copper uptake into intestinal mucosal cells has yet to be firmly established.333,334 Two putative low-affinity mammalian copper transporters, hCtrl2 and Nramp2, have also been identified. It is unclear what role hCtr2 plays in copper homeostasis, because its mRNA levels are highest in the placenta and very low in liver, intestine, and colon.332,334 The Nramp2 protein also has been identified as a proton-coupled metal ion transporter that transports a broad range of metal ions.309 Acting as a permease or by endocytosis, Ctr1 delivers Cu2+ within cells.335 The mechanism for reduction of copper ion before uptake remains unknown. If uncontrolled, this pool of cuprous ions could lead to generation of reactive oxygen species; however, very few, if any, free copper ions exist in the cytoplasm. The delivery of copper to target cuproenzymes depends on an elegant metallochaperone system. Several cytoplasmic chaperones have been described (Atox1, CCS [copper chaperone for Cu, Zn superoxide dismutase], and Cox17) as well as membrane-associated copper-transporting ATPases (ATP7A and ATP7B).336 The copper-transporting Menkes ATPase ATP7A (MNK) is responsible for copper export from the enterocyte, and may be defective in patients with Menkes disease in whom copper accumulates in intestinal cells.337,338 Once entering the plasma, copper is bound with albumin and histidine in the portal blood and rapidly deposited in the liver, where hCtr1 may play a role in this process. Ceruloplasmin, a major copper-containing protein in plasma, is synthesized in the liver with incorporation of

copper by the Wilson disease protein, which has a high homology with MNK and is defective in Wilson disease patients who suffer from copper accumulation in the liver (see Chapter 75).

Iodine

Iodine is present in varying amounts in a wide range of foods, depending on the soil content in the region where animals were reared and vegetation was grown. Seafood is particularly rich in iodine. Iodine is absorbed largely as inorganic iodide, but some iodine also is transported as amino acid complexes.339

Selenium

Selenium is found predominantly in association with amino acids, and about 60% of dietary selenium is absorbed. Selenium deficiency states have been reported from China (Keshan disease), where there is very little selenium in soil and water,340 but not in New Zealand, where intake is equally sparse.341 Absorption of selenium occurs rapidly when it is associated with amino acids, as in selenomethionine, probably by active transport mechanisms operative for the amino acid.342 Inorganic selenium is absorbed more slowly, possibly by simple diffusion.

Others

The mechanisms underlying the absorption of other trace elements, including manganese and chromium, are largely unknown.339 Deficiencies of trace elements are rare in normal persons, even in those with poor protein and calorie intake. Exceptions occur when local geographic availability is suboptimal, as can occur with iodine and possibly with selenium.

ADAPTATION TO CHANGES IN NEED OR LOAD One of the most fascinating aspects of intestinal function is the phenomenon of adaptation. Two specific forms of intestinal adaptation have been identified in the intestine: mucosal hypertrophy6 (leading to a global increase in absorption of all nutrients) and an increase in specific transport mechanisms induced in response to specific dietary needs or availability.343

MUCOSAL HYPERTROPHY

Resection of more than 50% of the human intestine results in increased fecal nitrogen losses, which subsequently slowly return toward normal, thus implying that mucosal adaptation has occurred. Normalization is explained largely by hypertrophy of intestinal mucosa, which manifests with increases in the number of villus enterocytes and in villus height without obvious increase in the absorption rate per individual cell.153 Absorption increases for all nutrients, and absorptive capacity may be enhanced up to five-fold in response to intestinal resection. Jejunal adaptation following ileal resection appears to be less efficient than ileal adaptation in response to jejunal resection. While hypertrophy in response to resection is the bestcharacterized example of adaptation, other causes also have been discerned, at least in experimental animals. Thus, during lactation and pregnancy,344 in diabetes,345 and in the physiologic response to extreme cold,346 hypertrophy is evident, but this may be a result, at least in part, to the hyperphagia that accompanies these conditions.

Chapter 100  Digestion and Absorption of Nutrients and Vitamins High-CHO diet

Amylase output

The mechanisms by which hypertrophy occurs have been the subject of much study. Signals of adaptation might relate to various hormone levels, transcription factors, ATP levels, or changes in the concentration of luminal solutes.347 The signals and mechanisms of the adaptive process may be different for the jejunum and ileum, as well as in the crypt and villus tip, explaining the sitespecific alterations and differences between crypt and villus enterocytes.347,348 The presence of luminal nutrition is a major stimulus to growth,6 in addition to intraluminal pancreaticobiliary secretions349; certain peptide hormones also have been implicated, particularly enteroglucagon and glucagon-like peptides.350 Gastrin, CCK, and other trophic factors (e.g., epidermal growth factor and insulin-like growth factors) display trophic effects on the gastrointe­ stinal tract,351-353 but it is uncertain whether these factors act as local paracrine mediators or as circulating hormones. Polyamines are other important local mediators of mucosal hypertrophy,354 because epithelial production of polyamines follows intestinal resection, and inhibiting their synthesis prevents the hypertrophy usually associated with resection. Polyamines also might play a role in maintaining normal mucosal structure, because their mucosal level in the intestines of experimental animals decreases rapidly in response to a 24-hour fast and increases within a few hours of refeeding.355 Although certain prostaglandins have been shown to enhance cell proliferation in the stomach and intestine, their role in adaptation is uncertain.356

Low-CHO diet

–3

–2

–1

Meal

1

2

3

Hours Figure 100-22.  Pancreatic enzyme adaptation to dietary manipulation. Diagram of typical pancreatic amylase outputs in normal subjects fed either a high- or a low-carbohydrate (CHO) diet for two weeks. Greater amylase secretion rates occur, both during the interdigestive phasic periods (↓) between meals and in response to a standard meal, in subjects given a high-CHO diet.

8

It has long been known that the digestive capacity of pancreatic juice can be altered by changes in nutritional intake, but it is now clear that specific responses occur after different types of manipulation of the diet.357 A high-protein diet enhances proteolytic enzyme production; a highcarbohydrate diet enhances amylase secretion; and a highfat diet stimulates lipase secretion (Fig. 100-22). In part, these changes appear to depend on specific polypeptide hormone release. Prolonged administration of cerulein (an analog of CCK) stimulates trypsinogen and inhibits amylase secretion,357 whereas secretin stimulates lipase secretion (Fig. 100-23).358 Insulin released from pancreatic beta cells, in response to carbohydrate ingestion, appears to be involved indirectly in enhancing amylase secretion.357 A high-fat diet also induces increased capacity to secrete gastric lipase, but the mechanisms involved are not known.359 The underlying molecular biologic events that lead to pancreatic adaptation have been studied, and, as might be expected, responses depend on the period over which a dietary stimulus is applied. Responses to short-term stimulation, as after a single meal, appear to depend on enhanced translation of mRNAs for enzymes.357 Protein synthesis increases within the first two hours of stimulation of hormones and appears to be a result of translational events; however, prolonged stimulation—over several days—increases mRNA production by increased transcription, leading to enhanced biochemical commitment to enzyme secretion.360 A single stimulus after a prolonged period of high intake of protein, therefore, results in much greater proteolytic enzyme output than in persons whose protein intake is low. The polypeptide hormones secretin, CCK, and possibly insulin, liberated in response to a meal, not only cause immediate release of pancreatic enzymes but also stimulate

Absolute rates of lipase synthesis (cpm x 10–2/μg DNA)

SPECIFIC REVERSIBLE ADAPTATION Secretion of Pancreatic Enzymes

6 Secretin

4

Cerulein

2

Saline

3

6

12

24

Infusion time (hr) Figure 100-23.  Pancreatic enzyme responses to prolonged hormone stimulation in rats. The lipase synthesis rate rose in response to secretin (16 units/kg/hr, orange line), but not to the cholecystokinin (CCK) analogue, cerulein (0.25 µg/kg/hr, blue line) or saline (purple line), infused for up to 24 hours. (From Rausch U. Lipase synthesis in the rat pancreas is regulated by secretin. Pancreas 1986; 1:522-8.) cpm, counts per minute.

gene expression over the longer term and thus increase secretory capacity.

Mucosal Responses

Adaptive responses to changes in dietary intake influence mucosal digestive and absorptive processes. Activity of the

1727

Section X  Small and Large Intestine disaccharidase enzymes sucrase and maltase increases in response to high carbohydrate intake over several days but not to manipulation of protein intake.361 Sucrase levels increase first in crypt cells about 24 hours after refeeding sucrose following a period of starvation. Thus, synthesis of the disaccharidases is stimulated, but their breakdown is also diminished. Conversely, lactase is an enzyme that appears not to respond to manipulation of dietary intake of lactose.362 Absorptive function also adapts to dietary manipulation.205 It has been widely assumed that there is a considerable reserve of absorptive function under normal circumstances, but Diamond and colleagues argued eloquently that it would be inefficient and costly in biosynthetic energy for the intestine to have a large spare capacity.205,363 Furthermore, a fairly close match has been demonstrated between absorptive capacity for many nutrients and dietary load. There is a clear need, therefore, for adaptation to occur in response to changes in load, and there is good evidence to suggest that most nutrients regulate their specific mucosal transporter. Three types of adaptive responses are discernible in the mucosa (Fig. 100-24).205 In the first, as exemplified by sugars, peptides, and nonessential amino acids, transport activity rises in response to increased dietary loads. Experimental animals fed diets high and low in glucose increase or decrease their maximum capacity for glucose transport, respectively, over a two-fold range, probably by changing the number of transporters. An analogous response to increased dietary load is seen with protein ingestion, in which peptide transporters and some amino acid transporters are increased. In the second type of mucosal response, as exemplified by a number of vitamins and trace elements, absorptive mechanisms are switched on by low dietary loads and switched off by a large load. Here absorption is enhanced

A

Peak

Relative transporter activity

1728

C

Basal

B

Deficient

Average

Excessive

Nutritional intake Figure 100-24.  Diagram of three types of adaptive responses of intestinal transporters to variation in nutritional intake. The first type of response (A) characterizes hexose and nonessential amino acid transport; the second type of respone (B) typifies elements such as iron and calcium and some vitamins; the third type of respone (C) is a mixed pattern seen with some essential amino acids. (Adapted from Ferraris RP, Diamond JM. Specific regulation of intestinal nutrient transporters by their dietary substrates. Annu Rev Physiol 1989; 51:125-41, with permission.)

in nutrient deficiency but inhibited with nutrient excess, when potentially toxic effects may result. A third and mixed pattern is seen with other types of nutrients, as with essential, predominantly basic, and neutral amino acids, when absorption, presumably in an effort to ensure adequate intake, is enhanced at very low dietary levels. Absorptive mechanisms are at their minimum with average dietary intake, but they rise as dietary ingestion increases above that range.205 The signal for up-regulation of brush border membrane glucose absorption is glucose itself, although an actively transported but unmetabolized sugar also stimulates glucose uptake.363 Fructose stimulates its own absorption by a mechanism separate from the one that stimulates glucose transport.364 Regulation of mucosal transport of acidic amino acids and imino acids follows the same pattern as that for sugars; that is, an increase in dietary load up-regulates transport. There is an interesting cross-induction of transport mechanisms by one type of amino acid with another. Thus, the basic amino acid arginine up-regulates acidic as well as basic amino acid transport, and the acidic amino acid aspartate induces maximal transport of basic amino acids as well as acidic ones.365 This cross-stimulation occurs between amino acids and peptides, each of which stimulates the absorption of the other. Enhanced absorption could result from increased numbers of transporters or from increased activity of each transporter.205 There is evidence in favor of both mechanisms, although the former probably predominates. It is uncertain whether an increase in transporter number is caused by increased synthesis, decreased degradation, or an increase in the insertion of preformed transporters into the brush border membrane. Because dietary regulation of glucose and amino acid transporters takes two to three days, it is likely that regulation occurs at the level of crypt cells. In diabetes mellitus, the persistent hyperglycemia stimulates both basolateral and apical membrane glucose transport, which can be inhibited by protein synthesis inhibitors such as cyclohexamide, suggesting a role for increased synthesis of new glucose transporters.366

VITAMINS AND TRACE ELEMENTS

Deficiencies of vitamins and trace elements are associated with up-regulation of their absorptive mechanisms. This is seen, for example, with biotin, thiamine, and ascorbic acid and with iron, calcium, zinc, and phosphate (Fig. 100-25).205 Because some of these moieties are potentially toxic, most down-regulate their transport mechanisms when present in higher concentrations. Low body stores of iron, zinc, calcium, and phosphate signal enhanced absorptive mechanisms. Zinc deficiency enhances zinc absorption approximately five-fold by increasing transport capacity. Dietary calcium deficiency stimulates calcium uptake in the proximal intestine by a vitamin D–dependent mechanism involving increases in a cytosolic calcium-binding protein and by stimulating transport across the brush border membrane and at the basolateral membrane. These changes occur within a few hours, suggesting that mature enterocytes on the villi are capable of being regulated. By contrast, the effects of dietary regulation of glucose and amino acids take two to three days. The difference in time scales over which gastrointestinal responses to stimulation by various nutrients occur provides fascinating insights into the molecular and biologic events underlying these phenomena. Immediate responses seen within seconds to a few minutes after exposure are likely to involve release, or activation, of preformed proteins; adaptive responses found within two to three hours

Chapter 100  Digestion and Absorption of Nutrients and Vitamins 5 Iron, zinc

Relative absorptive capacity

4

Calcium

3

2 Phosphate 1

0 Deficient

Average

Excessive

Nutritional intake Figure 100-25.  Changes in intestinal absorptive capacity for some nutrients in response to nutritional intake, ranging from deficient to excessive. (Adapted from Ferraris RP, Diamond JM. Specific regulation of intestinal nutrient transporters by their dietary substrates. Annu Rev Physiol 1989; 51:125-41, with permission.)

increase in small-bowel length and function per unit length.372 In hypophysectomized rats, mucosal hypoplasia of the small intestine and a reduced adaptive response develops after resection, and they are restored by GH. Transgenic mice expressing elevated levels of GH experience hypertrophy of the small intestine.372 IGF-1 expression in the small intestine is regulated by GH and is believed to induce enterotrophic effects following resection. Human and rabbit studies have indicated that increased nutrient transport activity devoid of morphologic changes may be the method of GH-induced intestinal adaptation.373 In patients who have short bowel syndrome and are dependent on home parenteral nutrition (HPN), high-dose recombinant human GH (0.4 mg/kg/day) in controlled and uncontrolled studies has led to variable results.374-376 GLP-2 increases the adaptive response to massive intestinal resection in rats.377 The non–placebo-controlled study by Jeppesen and colleagues in eight patients with short bowel syndrome and an end-enterostomy type of anastomosis (six had Crohn’s disease and four were not receiving hyperalimentation) showed a significant increase in the absolute and percentage intestinal absorption of net weight and a significant increase in the percentage but not the absolute absorption of energy and nitrogen in response to the use of GLP-2.378

THE NEONATAL INTESTINE of stimulation are probably due to increased translation; responses that take several hours or days are likely to be due to increased transcription and production of more mRNA.367

SIGNALING FOR INTESTINAL ADAPTATION AND IMPLICATIONS FOR THERAPY

A variety of signals have been described as possibly playing a role in the process of intestinal adaptation. These include growth hormone (GH), epidermal growth factor (EGF) and EGF receptor, prostanoids, uncoupling proteins, peroxisome proliferation-activated receptor-α (PPAR-α), insulinlike growth factor (IGF-1) receptors and IGF-binding proteins, transforming growth factor-α, SPARC (secreted protein, acidic and rich in cystine), Bcl-2, endothelin-1, erythropoietin, keratinocyte growth factor, the GATA family of zinc-finger transcription factors, hepatocyte growth factor, the early-response genes (ERGs), ornithine decarboxylase (ODC), PC4/TIS7, epimorphin, and AP-1, a transcription factor composed of Fos and JUN family proteins.347 Several of these signals might lend themselves to be modified in a clinical setting to enhance the intestinal adaptive response. For example, evidence from animal intestinal resection models suggest that the enhancing influence of glucocorticoids on sugar uptake may be achieved by post-translational processes involving signaling with c-jun, ODC and proglucagon, or other as yet unknown signals.368,369 In an animal model of extensive intestinal resection (50% enterectomy), prednisone had no effect on the uptake of glucose or fructose. By contrast, the locally acting steroid budesonide increased the value of the jejunal maximal transport rate for the uptake of glucose by more than 120% and increased the ileal uptake of fructose by more than 150%.368,370 GH might possess such proadaptive properties.371 In animal models, the administration of GH results in an

DEVELOPMENT AND ADAPTATION OF NUTRIENT DIGESTION AND ABSORPTION

Nutrient requirements vary markedly during early postnatal development, and this is mirrored by alterations in digestive and absorptive capacities. Some of these changes are genetically determined and programmed and do not appear to be greatly influenced by changes in dietary load.5 Thus, for most of the world’s population, excluding whites, the decline in activity of the disaccharidase lactase, which occurs after infancy, cannot be prevented by maintaining a high milk intake.205 Some early postnatal responses and most responses in adult life, however, appear to be purposive and reversible, paralleling changes in dietary intake, for digestive enzyme production and absorptive capacity. Such adaptations may occur in response to changes in dietary load or altered body needs.205,357

DEVELOPMENTAL CHANGES

Approximately 50% of the total calorie requirement of infants is provided by the fat in milk. Breast milk contains 3.5% to 4% lipid, of which 95% is in the form of triglyceride.176 The fatty acid composition is a mixture of mediumand long-chain fatty acids. In neonates, pancreatic lipase secretion is low and the digestion of triglyceride in milk relies on the other lipases present in milk or secreted by the tongue or gastric mucosa. Pancreatic lipase secretion rises after weaning because milk-derived lipase is no longer available (Fig. 100-26).367 Protein digestion is incomplete in infancy, and many proteins, such as human milk protein (whey), partially escape digestion.379 This relative immaturity also might have advantages for the infant because some biologically important peptides and immunoglobulins remain intact. Proteolytic enzymes are derived from a variety of sources, which also are changing during early neonatal life. Thus, there are several specific proteases, including trypsin and elastase, in breast milk.379 Gastric acid and pepsin are

1729

Section X  Small and Large Intestine Ileal bile salt absorption

100

Pancreatic enzymes Milk-derived lipase

Relative activity (%)

1730

Lactase 0 Birth Weaning 1 year

Adult life

Figure 100-26.  Diagram of the relative activities of various digestive enzymes and processes from birth to adulthood.

secreted at birth and increase toward adult values over the following three to four months. Nonetheless, little protein digestion appears to occur in the stomach during the first few weeks of life, possibly owing to the presence of protease inhibitors in milk. Likewise, luminal proteolytic machinery in the small intestine is not fully developed at birth, although enterokinase and pancreatic proteolytic enzymes are detectable. Rates of chymotrypsin and trypsin secretion are slower in infancy than in adult life, and responses to stimulation with CCK are depressed.380 Low rates of pancreatic enzyme secretion at birth may be attributable in part to the retarded display of polypeptide hormone receptors on the basolateral membrane of acinar cells. Digestive enzymes appear at different times after birth, suggesting that the genes that code for these enzymes may be activated at different times during development.380 One of the most characteristic changes in the postnatal period is the decline in lactase activity seen in most of the world’s populations, apart from white persons.381 The human population can be divided into two groups with respect to the ability to digest lactose: lactase-persistent persons, who are able to digest lactose throughout their lifespan, and lactase-nonpersistent (lactose intolerant, adult-type hypolactasia) persons. Lactase nonpersistence is caused by a decrease in the expression of the small intestinal–specific enzyme lactase–phlorizin hydrolase (LPH) at approximately age 5 to 10 years. Persons with lactase nonpersistence have a limited ability to digest lactose as adults; however, both lactase-nonpersistent and lactose-persistent persons have a high LPH expression after birth. The lactasepersistent phenotype is found most often among Northern Europeans, Indians from Punjab, Bedouins, and some nomadic tribes in Africa.381,382 The mechanism behind the developmental downregulation of LPH expression has been shown to be mainly transcriptional in humans, with several transcription factors (Cdx-2, GATA factors, and hepatocyte nuclear factor 1) mutually interacting and activating LPH expression.383,384 The binding sites for these factors are clustered within 100 bp upstream of a TATA box. No differences have been described in the DNA sequence in the LPH gene that are correlated to the lactase phenotype; however, a T/C poly-

morphism at position −13910 and an A/G polymorphism at position −22018 from the start codon of the LPH gene have been identified. Although these nucleotide variants are located in introns 8 and 13 of the neighboring MCM6 gene, the −13910C polymorphism associates 100% and the −22018G polymorphism associates approximately 97% with the lactase-nonpersistent phenotype in the Finnish population.385 It has been shown that the two single nucleotide polymorphisms are associated with the transcriptional regulation of the LPH gene.386 The −13910 T/C polymorphism is located in a transcriptional enhancer sequence, which strongly activates the LPH promoter activity. Furthermore, the −13910T variant isolated from a lactase-persistent person has a significantly higher enhancer activity than the corresponding −13910C variant isolated from a lactase-nonpersistent person. Analyses of the binding of nuclear factors to the −13910 polymorphic sequences show that the −13910T sequence binds nuclear factors with higher affinity than the −13910C sequence. The −22018 region does not possess enhancer activity and, in fact, results in a very small but significant reduction in reporter gene activity. The reduction of the reporter gene level by the −22018 region is greater in the presence of the enhancer of the −13910 region. It has been suggested that the lactase-persistent phenotype is caused by a mutation in the −13910 position, creating a strong enhancer that is able to keep the LPH gene active during adulthood (Fig. 100-27).382 Based on these results, a model explaining the mechanisms behind the postweaning down-regulation of LPH expression and adult-type hypolactasia has been proposed.387,388 Transcription factors necessary for LPH expression are present in excess during childhood and before weaning in mammals. After the weaning period, the expression of some intestinal transcription factors is changed (e.g., HNF1α). Also, the availability of some of these factors may be decreased because genes necessary for digestion of a starch-rich diet (e.g., SI), are up-regulated after weaning. Many of these intestine-specific genes depend on the same transcription factors (Cdx-2, HNF1, and GATA factors) as the LPH promoter. The competition for these transcriptional activators is higher after weaning. These changes result in a lower LPH expression owing to the weak nature of the LPH promoter; however, the strong enhancer effect of the −13910T variant compensates for these changes and can keep the LPH gene active throughout adulthood, giving the phenotype of lactase persistence (see Fig. 100-27). Changes in epithelial membrane transport of nutrients take place when the intestine is suddenly and rapidly expected to assume the role of the placenta in providing nutrients at the moment of birth and immediately there­ after. Brush border membrane glucose and most amino acid transporters are present in the human fetal intestine well before birth,5 when, in contrast to adult intestine, they are found throughout the crypt-villus axis. Fructose absorptive capacity rises rapidly after weaning when this sugar is presented in the diet. Transporters for bile salts are not programmed to appear on ileal enterocytes until weaning.343

TRIGLYCERIDE DIGESTION

Some differences have been recognized between the way infants digest triglyceride, which they receive in milk, and how adults digest this same fatty substrate. In contradistinction to other dietary sources, the triglyceride in milk is

Chapter 100  Digestion and Absorption of Nutrients and Vitamins HNF1/GATA/Cdx

Child

High level of transcription MCM6 gene A

LPH gene T

G

C

–22018

–13910

Lactase-persistent adult –

+ High level of transcription

A –22018

T –13910

Lactase-nonpersistent adult –

+ Low level of transcription

G

C

–22018 –13910 Figure 100-27.  Schematic model showing the interaction between the promoter of the lactase phlorizin hydrolase (LPH) gene and polymorphic −13910 and −22018 regions in the MCM6 gene in children and lactase-persistent and -nonpersistent adults. During childhood and before weaning in other mammals, the level of LPH expression is high because the transcription factors (HNF1α, GATA factors, Cdx-2) known to regulate LPH expression are available in excess. The expression of LPH is therefore not dependent on the −13910 enhancer activity. In adulthood, accessibility to the transcription factors is reduced. The enhancer activity of the −13910T variant ensures an active LPH gene throughout life (lactase persistence). The lower activity of the −13910C variant fails to activate or recruit the transcription factors, which results in low LPH gene activity in lactase-nonpersistent adults. Although the −22018 region represses LPH transcription, the role of the −22018 region is unclear; the repression does not seem to be related to the A/G polymorphism. (From Troelsen JT. An upstream polymorphism associated with lactase persistence has increased enhancer activity. Gastroenterology. 2003; 125:1686-94.)

packaged in smaller emulsion droplets, each surrounded by a trilaminar membrane that includes both phospholipid and proteins (albumin and β-lactoglobulin).367 At a time in its life when the newborn infant relies for more than half of its energy requirements on milk-derived triglyceride, pancreatic lipase secretion rates are only about half those of adults when expressed in terms of body surface area. Only at weaning does pancreatic lipase secretion begin to rise to adult levels.389 Two other lipases are important at this stage in life (see Table 100-2). The first is secreted by the mammary gland but is inactive in milk and requires the presence of bile salts to activate it. It begins to function, therefore, upon entering the duodenum.367 The second lipase is secreted either from serous glands at the base of the tongue or from gastric mucosa (or both), depending on the species. In humans, this preduodenal lipase is largely, if not entirely, derived from chief cells in the gastric mucosa390; in the rat it is derived from the tongue. This lipase has optimal pH of 4 to 6, which is lower than that of pancreatic lipase. It preferentially acts at the α1 position on triglyceride and releases fatty acid and diacylglycerol, as opposed to the monoglyceride and two fatty acids produced by pancreatic lipase. It is not stimu-

lated by bile salts and is released during feeding by autonomic nervous stimulation. Gastric lipase appears to be particularly active at the surface membrane of droplets derived from milk; pancreatic lipase is less so. Although the amount of lipolysis that occurs within the stomach is relatively small, the release of even modest amounts of fatty acids, particularly from the shorter- and medium-chain triglycerides, may be important in the emulsification of fat in the duodenum and in enhancing pancreatic lipase activity. Gastric lipase is rapidly inactivated in the duodenum by proteolytic enzymes. There is, however, conflicting evidence about the importance of intragastric lipolysis, and although some have suggested that as much as 30% of fat may be digested here,56 the rapid product inhibition of gastric lipase activity by released fatty acids makes this unlikely. Each of the three lipases has a different specificity for the ester bonds in triglyceride. Gastric lipase preferentially cleaves the bond in the 1α position, and pancreatic lipase the bonds in the 1α and 3α positions; milk-derived, bile salt–stimulated lipase is nonselective and splits the bonds at the 1-, 2-, and 3α positions. Thus, newborns’ luminal contents contain more fatty acids and less monoglyceride

1731

1732

Section X  Small and Large Intestine and diglyceride than those of adults, and this is probably advantageous for absorption. Smaller amounts of bile salts are available in newborns than in adults, at least in part because active ileal reabsorption is immature.367 Under these conditions, fatty acids are likely to be absorbed more readily than monoglyceride. In addition, the low availability of bile salts makes it likely that transfer of fat to the brush border membrane depends more on unilamellar liquid crystalline vesicles. Gastric lipase persists into adult life, and the amounts found in biopsies of adult gastric mucosa are similar to those found in infant gastric mucosa. Studies suggest it might hydrolyze as many as one in four triglyceride acyl chains during digestion of a meal.10

CARBOHYDRATE DIGESTION AND ABSORPTION

Lactose is the major carbohydrate in breast milk, and the need for an amylase before weaning is therefore minimal. An α amylase is present in milk, however, and an amylase also is secreted in saliva at birth.381 Both of these amylases are inactivated by acid in the stomach, but they may resume their activity at nearly neutral pH on reaching the duodenum. It has been estimated that 15% to 40% of amylase activity in the duodenum of infants is of salivary origin.389 Pancreatic amylase secretion is low, and stimulation with exogenous agonists produces little response, indicating the prematurity of the pancreas at this stage. In any event, the need for any amylase in neonates is minor unless starch is introduced early. Most infant formulas do not contain starches, but some contain glucose polymers. Digestion of any starch ingested during the first two or three months of life relies on salivary amylase and mucosal α glucosidases as well as colonic salvage by fermentation of undigested carbohydrate by bacteria. Mucosal lactase is present at birth in high concentration, as are the other glucosidases.391 Nevertheless, lactose absorption may be incomplete in neonates, particularly in premature neonates. Estimates of the amount of lactose that reaches the colon vary and have been based on indirect measurements of breath hydrogen concentrations. Probably less than 20% of ingested lactose reaches the colon, but a much smaller proportion is lost in the stool because of bacterial hydrolysis and absorption of the products. The glucose/galactose transporter in the apical membrane of villus enterocytes is well developed in full-term infants, but it may be suboptimal in premature ones. This is unlikely, however, to pose a significant barrier to nutrition.

PROTEIN DIGESTION AND ABSORPTION

Although acid and pepsinogens are secreted in neonates, little intragastric proteolysis occurs during the first two to six weeks of life. A renin-like protease is secreted during the first 10 days of life, which causes protein precipitation.379 Pancreatic proteolytic enzymes are secreted at birth, although at slower rates than in adults. Trypsinogen secretion is low and, particularly in preterm infants, does not respond to feeding. Stimulation with pancreozymin has little effect on pancreatic enzyme secretion for the first one to two months of life.177 Enterokinase (enteropeptidase) is present at birth and is capable of activating trypsinogen. Despite the apparent immaturity of the proteolytic machinery, it has been estimated that duodenal proteolysis can cope with as much as 3 or 4 grams of protein per kilogram of body weight of casein, and infants seem not to be prone to defective nitrogen nutrition. A number of proteases have been found in breast milk, including plasmin, which is most active against casein.392

The overall nutritional importance of these milk-derived proteases or the protease inhibitors also found in milk393 is not known. Transport systems for amino acids and small peptides appear to be well developed in newborns. The infant intestine has greater capacity than the adult intestine to absorb intact macromolecules, including proteins. Transport by pinocytosis or receptor-mediated endocytosis probably accounts for the ability of infants to take in biologically important whole proteins, such as the immunoglobulins, during this phase of life. This mechanism disappears after the first three months of life, when closure is said to have occurred394; however, uptake of intact proteins continues throughout life, albeit in trace amounts, and the role of M cells on Peyer’s patches is of major importance in this process. It is likely to be an important mechanism by which dietary antigens are presented for immune surveillance later in life, but it is of little nutritional significance.

EFFECTS OF BARIATRIC SURGERY ON NORMAL DIGESTION AND ABSORPTION The total volume of bariatric surgical procedures in the United States continues to grow rapidly (see Chapter 7). The American Society for Bariatric Surgery estimates that such operations have increased 1431% since 2000 to more than 250,000 annually. At present, the Roux-en-Y gastric bypass (RYGB) is the most widely performed weight loss procedure in the United States, followed by vertical banded gastroplasty or laparoscopic gastric banding and biliopancreatic diversion with or without duodenal switch.395,396 The exact mechanisms of weight loss associated with each treatment modality, however, and how they affect intestinal digestion and absorption is not clearly understood. Operations for weight loss have generally been classified as malabsorptive, restrictive, or both, based on the proposed mechanism for the induction of weight loss. The jejunoileal bypass (JIB) and its more recent modifications, the bilio­ pancreatic diversion (BPD) and the distal gastric bypass with duodenal switch (DGB-DS), are classified as primarily malabsorptive procedures. The gastroplasties (horizontal, vertical, and vertical banded gastroplasty [VBG]) are restrictive. RYGB represents a combination of these two mechanisms. The results of surgical treatment of severe obesity differ considerably based on the mechanisms of weight loss. Malabsorptive procedures, especially the JIB, produce the greatest degree of weight loss but can be associated with serious and potentially life-threatening metabolic and nutritional complications.397 BPD and DGB-DS produce significant weight loss that persists and causes fewer metabolic complications than JIB.396,397 Restrictive procedures produce moderate degrees of weight loss and are associated with the lowest incidence of metabolic and nutritional complications.396 RYGB has consistently been shown to produce greater and more sustained weight loss than VBG and avoids the severe metabolic and nutritional consequences of intestinal bypass. The mechanism of weight loss following VBG appears to be similar to that of other forms of caloric restriction, although the weight loss tends to be more pronounced and to persist for a longer time. The role of the regulation of central satiety mechanisms has been studied for restrictive and malabsorptive types of surgery. Ghrelin is a gastric peptide with potent orexigenic effects. Circulating ghrelin concentrations are high in obese subjects, and they increase

Chapter 100  Digestion and Absorption of Nutrients and Vitamins after weight loss. In patients undergoing RYGB (compared with patients undergoing VBG or BPD), however, a decrease in ghrelin levels has been reported and appears to depend on the surgically induced bypass of the ghrelin-producing cell populations of the gastric fundus.398 Several mechanisms have been suggested to account for the more substantial weight loss after JIB, RYGB and BPD. These can be divided into three groups: caloric restriction, changes in energy metabolism, and alterations in gastrointestinal hormones and nutrient absorption. Caloric restriction seems to play a prominent role in JIB and RYGB, and numerous studies document significantly decreased caloric intake after each type of surgery. Animal studies have shown that rats subjected to intestinal bypass maintained lower weights despite caloric intakes comparable with those of sham-operated controls, suggesting an increase in energy expenditure. Human data suggest a relative increase in the mean resting energy expenditure in patients undergoing RYGB.399,400 Malabsorption leading to a loss of lean body mass, diarrhea, and vitamin, mineral, and electrolyte deficiencies prominently contributes to weight loss after JIB, but less so after BPD and DGB-DS. After JIB, malabsorption results from the drastically reduced intestinal surface in contact with nutrients (common channel) and rapid transit time. It has been suggested that the enhanced weight loss after BPD is related to the combination of a short common channel (approximately 50 cm), which limits fat absorption, and a long afferent limb, which is not in direct contact with food. Therefore, the effect of the alimentary (Roux) limb, where protein and carbohydrates are absorbed, is limited. The documented changes in absorption might result from the alterations in the hormonal milieu of the gastrointestinal tract. Dumping syndrome, evidenced by symptoms of lightheadedness, sweating, palpitations, and diarrhea following a test meal containing a high-carbohydrate load, results from operations that bypass the normal pyloric functions and has been documented in patient with RYGB.401,402 These symptoms are temporally associated with changes in intestinal peptides, especially an increase in enteroglucagon, which is secreted by the ileum. Increase in enteroglucagon and neurotensin following ileal transposition in rats is associated with weight loss equivalent to that achieved with caloric restriction. These findings suggest that rapid transport of undigested and unabsorbed nutrients into the ileum after these procedures stimulate the secretion of intestinal peptides, (e.g., enteroglucagon), which then produce or

serve as a marker for full-blown or subclinical dumping, which in turn contributes to the observed weight loss.401,403 Peptide YY (PYY) is a hormone secreted by the L cells lining the terminal small intestine and colon in response to intestinal nutrients, especially long-chain fatty acids.21,404 PYY decreases gastric and intestinal emptying and improves intestinal absorption. This mechanism is the so-called ileal brake.32 PYY is also a potent anorexigenic hormone that acts by modulating appetite circuits in the hypothalamus, as well as stimulating apo A-IV, which plays an important role in upper gastrointestinal function and satiety.21,405 It has been shown that PYY serum levels are higher in patients who have undergone JIB.406

KEY REFERENCES

Altmann SW, Davis HR, Jr., Zhu LJ, et al. Niemann-Pick C1 Like 1 protein is critical for intestinal cholesterol absorption. Science 2004; 303:1201-4. (Ref 109.) Drobnik W, Lindenthal B, Lieser B, et al. ATP-binding cassette transporter A1 (ABCA1) affects total body sterol metabolism. Gastro­ enterology 2001; 120:1203-11. (Ref 100.) Kuokkanen M, Enattah NS, Oksanen A, et al. Transcriptional regulation of the lactase-phlorizin hydrolase gene by polymorphisms associated with adult-type hypolactasia. Gut 2003; 52:647-52. (Ref 386.) Nemeth E, Tuttle MS, Powelson J, et al. Hepcidin regulates cellular iron efflux by binding to ferroportin and inducing its internalization. Science 2004; 306:2090-3. (Ref 317.) Owyang C, Logsdon CD. New insights into neurohormonal regulation of pancreatic secretion. Gastroenterology 2004; 127:957-69. (Ref 38.) Parkkila S, Niemela O, Britton RS, et al. Molecular aspects of iron absorption and HFE expression. Gastroenterology 2001; 121:1489-96. (Ref 307.) Peet DJ, Turley SD, Ma W, et al. Cholesterol and bile acid metabolism are impaired in mice lacking the nuclear oxysterol receptor LXR alpha. Cell 1998; 93:693-704. (Ref 137.) Ritz V, Alfalah M, Zimmer KP, et al. Congenital sucrase–isomaltase deficiency because of an accumulation of the mutant enzyme in the endoplasmic reticulum. Gastroenterology 2003; 125:1678-85. (Ref 173.) Santer R, Hillebrand G, Steinmann B, Schaub J. Intestinal glucose transport: Evidence for a membrane traffic-based pathway in humans. Gastroenterology 2003; 124:34-9. (Ref 209.) Stahl A. A current review of fatty acid transport proteins (SLC27). Pflugers Arch 2004; 447:722-7. (Ref 88.) Wang DQ. Regulation of intestinal cholesterol absorption. Annu Rev Physiol 2007; 69:221-48. (Ref 144.) Woods SC. Gastrointestinal satiety signals I. An overview of gastrointestinal signals that influence food intake. Am J Physiol Gastrointest Liver Physiol 2004; 286:G7-13. (Ref 11.) Full references for this chapter can be found on www.expertconsult.com.

1733

CHAPTE R

101 Maldigestion and Malabsorption Christoph Högenauer and Heinz F. Hammer

CHAPTER OUTLINE Etiology and Pathophysiology  1736 Fats  1737 Defective Mixing  1737 Reduced Solubilization of Fat  1738 Decreased Lipolysis  1738 Decreased Mucosal Absorption and Chylomicron Formation  1738 Defective Lymphatic Transport of Chylomicrons  1738 Proteins and Amino Acids  1738 Defective Intraluminal Proteolysis  1738 Defective Mucosal Hydrolysis of Peptides and Decreased Absorption of Oligopeptides and Amino Acids  1738 Carbohydrates  1739 Defective Intraluminal Hydrolysis of Carbohydrates  1739 Mucosal Defects of Carbohydrate Digestion and Absorption  1739 Vitamins  1739 Fat-Soluble Vitamins  1739 Water-Soluble Vitamins  1739 Minerals  1739 Calcium  1739 Magnesium  1740 Iron  1740 Zinc  1740 Others  1740 Mechanisms That Compensate for Malabsorption  1740 Role of the Colon  1740 Role of Intestinal Transit in the Salvage of Malabsorbed Nutrients  1741

In the past, it was believed that most malabsorptive diseases manifested clinically with diarrhea and steatorrhea. It is now recognized, however, that many malabsorptive dis­ orders, such as celiac disease, might have subtle clinical presentations, such as bloating or changes in bowel habits, or mainly extraintestinal manifestations, such as anemia, bone loss, or menstrual disturbance, that lead to erroneous diagnoses. Awareness also is increasing that subtle malab­ sorption of single nutrients, such as calcium or vitamin B12, can, if unrecognized, lead to complications that may be difficult to reverse or even irreversible. Therefore, the clini­ cal challenge today is to recognize and treat malabsorption despite its subtle manifestations. Classically, maldigestion is defined as defective hydro­ lysis of nutrients, and malabsorption is defined as defective mucosal absorption. Although this distinction may be

Clinical Features and Evaluation  1741 Suspecting and Confirming the Presence of Malabsorption  1742 Diagnostic Approach  1742 Anatomic Investigations  1746 Endoscopy, Biopsy, and Duodenal Aspiration  1746 Abdominal Imaging  1749 Noninvasive Evaluation of Gastrointestinal Digestive and Absorptive Function  1749 Malabsorption in Specific Situations and Disease States  1754 Lactose Malabsorption and Intolerance  1754 Incomplete Absorption and Intolerance of Fructose  1755 Ileal Bile Acid Malabsorption  1755 Amyloidosis  1756 Malabsorption Caused by Drugs and Food Supplements  1756 Malabsorption after Gastric Resection or Bariatric Surgery  1756 Malabsorption in the Elderly  1758 Connective Tissue Diseases  1758 Congenital Defects That Cause Malabsorption  1758 Primary Immunodeficiency Diseases  1763 Neurofibromatosis Type 1 (von Recklinghausen’s Disease)  1764 Nongranulomatous Chronic Idiopathic Enterocolitis and Autoimmune Enteropathy  1764 Endocrine and Metabolic Disorders  1765 General Approach to Management  1767

useful on pathophysiologic grounds, the clinical presenta­ tion and complications of maldigestion and malabsorption are similar. Moreover, physiologic processes other than digestion and absorption, such as solubilization, intestinal motility, or hormone secretion, contribute to the normal absorption of nutrients, vitamins, and minerals. Therefore, the classic definitions of maldigestion and malabsorption do not cover the pathophysiologic spectrum of the malab­ sorption syndrome. In this chapter, the terms digestion and absorption, or maldigestion and malabsorption, are used separately only in the discussion of pathophysiology. When the distinction between these terms is not of clinical relevance, only the terms absorption and malabsorption are used. Malabsorption can be caused by many diseases of the small intestine and also by diseases of the pancreas, liver,

1735

1736

Section X  Small and Large Intestine Table 101-1  Diseases That Cause Nutrient Malabsorption Gastric Diseases Atrophic gastritis Autoimmune gastritis (pernicious anemia) Gastric resection or bypass surgery Pancreatic Diseases Congenital pancreatic enzyme deficiencies Colipase deficiency Lipase deficiency Trypsinogen deficiency Pancreatic insufficiency Chronic pancreatitis Cystic fibrosis Johanson-Blizzard syndrome Pearson’s marrow-pancreas syndrome Shwachman’s syndrome Pancreatic tumors Liver Diseases Inborn errors of bile acid biosynthesis and transport Cirrhosis and other liver diseases Portal hypertension Obstructive Biliary Diseases Biliary tumors Primary and secondary sclerosing cholangitis Intestinal Diseases Amyloidosis Autoimmune enteropathy Celiac disease Collagenous sprue Congenital intestinal defects (see Table 101-14) Crohn’s disease Enteroendocrine cell deficiency Autoimmune polyglandular syndrome type 1 Enteric anendocrinosis Enterokinase deficiency Eosinophilic gastroenteritis Fistulas Food allergy Graft-versus-host disease Hypolactasia Ileal bile acid malabsorption Intestinal infections AIDS (HIV infection): Cryptosporidiosis, Mycobacterium avium complex infection, viral infections Giardiasis

Helminthic infections Tuberculosis Whipple’s disease Immunoproliferative small intestinal disease Intestinal ischemia Intestinal lymphoma Intestinal resections or bypass Mastocytosis Nongranulomatous chronic idiopathic enterocolitis Postinfection malabsorption Primary immunodeficiency diseases Radiation enteritis Refractory sprue Sarcoidosis Small intestinal bacterial overgrowth Tropical sprue Lymphatic Diseases Primary intestinal lymphangiectasia Secondary intestinal lymphangiectasia Lymphoma Solid tumors Thoracic duct trauma, damage, or obstruction Neuroendocrine Tumors Carcinoid syndrome Glucagonoma Somatostatinoma Zollinger-Ellison syndrome Cardiac and Vascular Diseases Congestive heart failure Constrictive pericarditis Endocrine Causes Addison’s disease Diabetes mellitus Hyperthyroidism Systemic Diseases Cronkhite-Canada syndrome Mixed connective tissue disease Neurofibromatosis type 1 Protein-calorie malnutrition Scleroderma Systemic lupus erythematosus

AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus.

biliary tract, and stomach (Table 101-1). Whereas in some of these diseases, malabsorption may be the presenting feature, in others malabsorption may be only a minor clinical problem or may be detected only as a laboratory abnormality. This chapter provides an overview of basic pathophy­ siologic mechanisms leading to symptoms or complications of maldigestion or malabsorption, reviews the clinical mani­ festations and complications of malabsorption, describes tests that can be used clinically to evaluate digestive and absorptive function, provides a rational diagnostic approach to the individual patient, and discusses malabsorptive diseases and general measures in the treatment of malab­ sorption syndrome that are not covered in other chapters of this book.

ETIOLOGY AND PATHOPHYSIOLOGY From a pathophysiologic point of view, mechanisms causing malabsorption can be divided into premucosal (luminal)

factors, mucosal factors, and postmucosal factors (vascular and lymphatic). For clinical purposes, this approach is of limited value, because the various clinical pictures caused by malabsorption syndromes are determined mainly by the nature of the malabsorbed substrates. We therefore discuss the mechanisms causing malabsorption on the basis of the malabsorbed substrate. A separate section is devoted to the role of mechanisms compensating for the consequences of malabsorption. Normal uptake of nutrients, vitamins, and minerals by the gastrointestinal tract requires several steps, each of which can be compromised in disease. (Normal digestion and absorption are discussed in Chapter 100.) Solubilization is a prerequisite for the absorption of such nutrients as fat or calcium. Fat and fat-soluble vitamins are solubilized by the formation of micelles, and calcium is solubilized through acidification of the gastrointestinal lumen. Alternatively, increased solubilization of the components of intestinal chyme can contribute to the mani­ festations of gastrointestinal diseases, such as increased absorption of oxalate, which can result in the development of kidney stones in patients with short bowel syndrome.

Chapter 101  Maldigestion and Malabsorption Table 101-2  Mechanisms of Malabsorption, Malabsorbed Substrates, and Representative Causes MECHANISM

MALABSORBED SUBSTRATE(S)

representative causes

Fat Fat-soluble vitamins Calcium Magnesium

Hepatic parenchymal disease Biliary obstruction Bacterial overgrowth with bile acid deconjugation Ileal bile acid malabsorption CCK deficiency

Pancreatic insufficiency

Fat Protein Carbohydrate Fat-soluble vitamins Vitamin B12 (cobalamin)

Congenital defects Chronic pancreatitis Pancreatic tumors Inactivation of pancreatic enzymes (e.g., Zollinger-Ellison syndrome)

Reduced mucosal digestion

Carbohydrate Protein

Congenital defects (see Table 101-14) Acquired lactase deficiency Generalized mucosal disease (e.g., celiac disease, Crohn’s disease)

Intraluminal consumption of nutrients

Vitamin B12 (cobalamin)

Small intestinal bacterial overgrowth Helminthic infections (e.g., Diphyllobothrium latum infection)

Fat Protein Carbohydrate Vitamins Minerals

Congenital transport defects (see Table 101-14) Generalized mucosal diseases (e.g., celiac disease, Crohn’s disease) Previous intestinal resection or bypass Infections Intestinal lymphoma

Fat Protein

Intestinal lymphangiectasia Primary Secondary (e.g., solid tumors, Whipple’s disease, lymphomas) Venous stasis (e.g., from congestive heart failure)

Vitamin B12

Pernicious anemia Atrophic gastritis Previous gastric resection

Decreased gastric mixing and/or rapid gastric emptying

Fat Calcium Protein

Previous gastric resection Autonomic neuropathy

Rapid intestinal transit

Fat

Autonomic neuropathy Hyperthyroidism

Maldigestion Conjugated bile acid deficiency

Malabsorption Reduced mucosal absorption

Decreased transport from the intestine

Other Mechanisms Decreased gastric acid and/or intrinsic factor secretion

CCK, cholecystokinin.

Digestion of macromolecular compounds, such as poly­ saccharides, triglycerides, and proteins, to their molecular components—monosaccharides, fatty acids, and amino acids, respectively—is achieved by soluble or membranebound digestive enzymes. Absorption of undigested or par­ tially digested macromolecular compounds occurs to a very minor degree in health and may be increased slightly in various intestinal diseases. Although such absorption does not play a nutritive role, it may be important for the normal function of the immune system and for the pathogenesis of diseases such as food allergy (see Chapter 9). Liberation of substrate, such as vitamin B12, from binding sites in food or, conversely, binding to factors such as intrin­ sic factor allows absorption to take place. Chemical changes to nutrients may be required for absorp­ tion, such as changing the charge of iron. Mucosal absorption can occur by active or passive carriermediated transport or by simple or facilitated diffusion. Postmucosal transport of absorbed substrates also is important.

Intestinal sensory and motor function permits detection of the presence of nutrients, facilitates adequate mixing of nutrients with intestinal secretions and delivery to absorp­ tive sites, and provides adequate time for nutrient absorp­ tion (see Chapter 97). Neural and hormonal functions are required to stimulate and coordinate digestive secretions, mucosal absorption, and intestinal motility. An overview of pathophysiologic mechanisms of mal­ digestion and malabsorption is provided in Table 101-2. This table also shows the ingested substrates primarily affected by the individual pathophysiologic mechanisms and lists examples of etiologic disorders for these mechanisms.

FATS DEFECTIVE MIXING

For sufficient digestion and absorption of lipids, dietary fat must adequately mix with digestive secretions. Gastric

1737

1738

Section X  Small and Large Intestine Table 101-3  Pathophysiologic Mechanisms That Result in Deficiency of Luminal Conjugated Bile Acids PATHOPHYSIOLOGIC MECHANISM

causes

Decreased synthesis and/or secretion of conjugated bile acids

Parenchymal liver diseases (e.g., cirrhosis, primary biliary cirrhosis) Biliary obstruction (e.g., tumors) Biliary fistulas Inborn errors of bile acid synthesis CCK deficiency

Intestinal loss of conjugated bile acids

Ileal resection Severe ileal mucosal disease Congenital defects of the ileal sodium–bile acid cotransporter

Luminal deconjugation of bile acids

Small intestinal bacterial overgrowth

Binding of bile salts or insolubilization of bile salts due to low luminal pH

Cholestyramine (binding) Zollinger-Ellison syndrome (low pH) Exocrine pancreatic insufficiency (low pH)

CCK, cholecystokinin.

resections or gastrointestinal motility disorders that result in rapid gastric emptying or rapid intestinal transit, such as autonomic neuropathy resulting from diabetes mellitus or amyloidosis, can cause fat malabsorption consequent to impaired gastrointestinal mixing of dietary fat.1

REDUCED SOLUBILIZATION OF FAT

Fat malabsorption due to decreased formation of micelles occurs if the luminal concentrations of conjugated bile acids are lower than the critical concentration required for forming micelles.2,3 Table 101-31,4 details the pathophysiologic mechanisms and representative diseases that cause luminal bile acid deficiency.

DECREASED LIPOLYSIS

If exocrine pancreatic function is severely reduced, impair­ ment of pancreatic lipase and colipase secretion results in decreased luminal hydrolysis of dietary fat.5 Chronic pan­ creatitis, cystic fibrosis, pancreatic duct obstruction by pan­ creatic and ampullary tumors, and pancreatic resection are the most common causes of pancreatic insufficiency.1 Even when pancreatic enzyme concentrations are normal, reduced pancreatic lipase activity due to a low luminal pH,6 exces­ sive calcium ingestion,7 or ingestion of the specific lipase inhibitor orlistat8 can cause pancreatic steatorrhea. Selec­ tive congenital lipase or colipase deficiency is a rare cause of pancreatic fat malabsorption.9

DECREASED MUCOSAL ABSORPTION AND CHYLOMICRON FORMATION

Generalized mucosal diseases, such as celiac disease or tropical sprue, often are associated with fat malabsorption. Defective uptake of free fatty acids and monoglycerides results from reduced mucosal surface area because of villus shortening, reduced enterocyte function, and mucosal inflammation.1 Intestinal fat absorption also is impaired in diseases that result in disturbance of intracellular formation of chylomicrons and accumulation of lipids within the enterocytes, including abetalipoproteinemia, hypobetalipo­ proteinemia, and chylomicron retention disease.10

DEFECTIVE LYMPHATIC TRANSPORT OF CHYLOMICRONS

Impairment of lymphatic transport of chylomicrons is a cause for postmucosal malabsorption of dietary fat.

Decreased lymphatic transport can result from congenital diseases such as primary intestinal lymphangiectasia or from obstruction of lymphatic vessels due to metastatic solid tumors, lymphoma, Whipple’s disease, retroperitoneal fibrosis, or trauma6 (see Chapter 28). Usually, lymphatic vessels in the mucosa become dilated (lymphangiectasia), and chylomicrons are lost into the intestinal lumen post­ prandially and also in the fasting state11; steatorrhea in these situations usually is only mild to moderate.10

PROTEINS AND AMINO ACIDS Defective digestion or absorption of dietary proteins has to be differentiated from excessive loss of serum proteins into the gastrointestinal tract, which is termed protein-losing enteropathy (see Chapter 28).

DEFECTIVE INTRALUMINAL PROTEOLYSIS

Protein digestion may be impaired in patients who have undergone partial or total gastric resection, presumably as a result of poor mixing with digestive secretions, although gastric pepsin deficiency could be contributory. Defective proteolysis also occurs with exocrine pancreatic insuffi­ ciency.1,12,13 In congenital diseases, pancreatic proteolysis can be impaired by inborn errors in the synthesis of proteo­ lytic enzymes (trypsinogen deficiency)13 or by defective acti­ vation of pancreatic proenzymes resulting from congenital deficiency of intestinal enterokinase (see later).14

DEFECTIVE MUCOSAL HYDROLYSIS OF PEPTIDES AND DECREASED ABSORPTION OF OLIGOPEPTIDES AND AMINO ACIDS

Generalized mucosal diseases, such as celiac disease and tropical sprue, result in global malabsorption, which includes malabsorption of oligopeptides and amino acids due to lack of mucosal hydrolysis of oligopeptides and defective mucosal absorption.13 Reduction of intestinal absorptive surface, as in short bowel syndrome or jejuno­ ileal bypass, also results in protein and amino acid malab­ sorption.13,15 Congenital defects of amino acid transporters on the enterocytes, such as Hartnup’s disease and lysinuric protein intolerance, can lead to selective malabsorption of a subgroup of amino acids (see later on).

Chapter 101  Maldigestion and Malabsorption CARBOHYDRATES DEFECTIVE INTRALUMINAL HYDROLYSIS OF CARBOHYDRATES

Pancreatic α-amylase normally is secreted in excess into the intestinal lumen. In mild forms of pancreatic insufficiency, carbohydrate digestion usually is at least partially pre­ served,16 but severe pancreatic insufficiency results in clini­ cally apparent carbohydrate malabsorption and diarrhea due to decreased luminal hydrolysis of ingested starch.17

MUCOSAL DEFECTS OF CARBOHYDRATE DIGESTION AND ABSORPTION

The most common cause of carbohydrate malabsorption is late-onset lactose malabsorption due to decreased levels of the intestinal brush border enzyme lactase (adult-type hypo­ lactasia, acquired primary lactase deficiency). Depending on ethnic background, lactase is present in less than 5% to more than 90% of the adult population; its deficiency results in a selective malabsorption of lactose. Acquired malabsorption of carbohydrates occurs commonly after extensive intestinal resections, in diffuse mucosal diseases such as celiac disease or Crohn’s disease, or temporarily after self-limited gastrointestinal infections (postinfection carbohydrate malabsorption).16,17 The pathophysiologic mechanisms of carbohydrate malabsorption are reduction of the intestinal mucosal surface area and a reduced activity or expression of intestinal oligo- and disaccharidases or transport proteins for monosaccharides.16 Congenital disaccharidase deficiencies (lactase, sucrase-isomaltase, and trehalase)18 and congenital deficiency or malfunction of transport molecules as in congenital glucose-galactose mal­ absorption19 can cause early onset of malabsorption of mono- or disaccharides (see later on). Intolerance of fructose is discussed in a subsequent section.

VITAMINS FAT-SOLUBLE VITAMINS

Diseases causing malabsorption of dietary fat commonly cause malabsorption of fat-soluble vitamins, because they require similar absorptive mechanisms. This is especially important in diseases that result in impaired micelle forma­ tion from bile salt deficiency.20 Fat-soluble vitamins also are malabsorbed in diffuse diseases of the mucosal surface area, in diseases affecting chylomicron formation and transport,21 and in exocrine pancreatic insufficiency.22 Some authors have suggested that absorption of fat-soluble vitamins is less affected by exocrine pancreatic insufficiency than by small intestinal diseases resulting in steatorrhea.23

WATER-SOLUBLE VITAMINS Vitamin B 12 (Cobalamin)

Decreased release of dietary vitamin B12 from food sources because of impaired pepsin and acid secretion, as in atro­ phic gastritis24 or use of acid inhibitory drugs such as proton pump inhibitors,25 usually results in only mild cobalamin malabsorption without clinical consequences. By contrast, deficiency of gastric intrinsic factor secretion, as occurs in pernicious anemia or after gastric resections, or secretion of an abnormal intrinsic factor, as in some congenital diseases, results in severe vitamin B12 malabsorption with clinical consequences.24 Autoimmune gastritis of pernicious anemia is the most common cause of vitamin B12 malabsorption.26 Cobalamin

malabsorption in pernicious anemia is caused by decreased intrinsic factor secretion resulting from parietal cell destruc­ tion and by blocking autoantibodies that inhibit intrinsic factor binding to vitamin B12.26 Mild cobalamin malabsorp­ tion may be found in patients with pancreatic insufficiency and in patients with Zollinger-Ellison syndrome, owing to decreased proteolytic release of vitamin B12 from its complex with R-binding protein24,27 (see Chapters 32, 59, and 100). In bacterial overgrowth syndrome (see Chapter 102) or helminthic infection with Diphyllobothrium latum (see Chapter 110), dietary cobalamin is made unavailable to the host or is consumed by the microorganisms or parasites in the intestinal lumen and, therefore, is not available for intes­ tinal absorption.26 Diseases and conditions affecting the ileal mucosa, such as Crohn’s disease or ileal resection, lead to a reduction of specific absorptive sites for the intrinsic factor-vitamin B12 complex.24 Ileal resections of more than 60 cm usually result in clinically significant vitamin B12 malabsorption.28 Imerslund-Gräsbeck syndrome, a disease of autosomal recessive inheritance due to malfunction of the cubilinamnionless (AMN) complex, is characterized by selective ileal malabsorption of the intrinsic factor–vitamin B12 complex despite normal ileal morphology.24,29 Congenital diseases affecting transcobalamin II also result in malab­ sorption of cobalamin.24,30 In previously healthy persons it usually takes several years of vitamin B12 malabsorption before cobalamin defi­ ciency develops, because the body stores contain large amounts of cobalamin and the daily requirement is rela­ tively small.

Folate Folate malabsorption occurs with mucosal diseases affect­ ing the proximal small intestine, such as celiac disease, Whipple’s disease, and tropical sprue.31 Folate deficiency is common in chronic alcoholism, in which it is postulated to be caused by decreased dietary intake as well as decreased intestinal absorption of folate.32 As discussed later, several drugs result in impaired intestinal uptake of folate, and an inherited form of selective folate malabsorption has been described. In contrast with cobalamin, body stores of folate are small relative to the daily requirements; therefore, folate deficiency develops faster than cobalamin deficiency in the setting of malabsorption. Increased serum folate levels resulting from bacterial formation of tetrahydrofolate have been reported in small intestinal bacterial overgrowth states.33

Other Water-Soluble Vitamins Other water-soluble vitamins, such as ascorbic acid and the B-complex vitamins, are absorbed in the small intestine either by carrier-mediated transport or by passive diffusion. Generalized malabsorption syndromes from intestinal causes impair the absorption of these vitamins, thereby leading to deficiency states.34,35 Deficiency of these watersoluble vitamins also occurs in chronic alcoholism, proba­ bly owing to decreased oral intake and reduced intestinal absorption.32

MINERALS CALCIUM

Severe calcium malabsorption can occur in diseases that affect the small intestinal mucosa, such as celiac disease. In these disease states, calcium absorption is impaired directly because of the reduction of the intestinal surface area and

1739

1740

Section X  Small and Large Intestine indirectly because of formation of insoluble calcium soaps with malabsorbed long-chain fatty acids. Therefore, dis­ eases causing malabsorption of long-chain fatty acids by other mechanisms, such as bile acid deficiency, also can result in calcium malabsorption.21 In many of these dis­ eases, malabsorption and deficiency of vitamin D further contribute to intestinal calcium malabsorption.21 Selective intestinal malabsorption of calcium—that is, without fat malabsorption—can occur in renal disease, hypopara­ thyroidism, and inborn defects in formation of 1α,25dihydroxyvitamin D or in the intestinal vitamin D receptor.21 Calcium malabsorption also occurs commonly after gastric resection (see subsequent section, “Malabsorption after Gastric Resection”).

MAGNESIUM

In many generalized malabsorptive disorders, magnesium malabsorption can result in magnesium deficiency.36 Mal­ absorption is due to the reduction in mucosal absorptive surface area and to luminal binding of magnesium by mal­ absorbed fatty acids; a congenital form of selective intestinal magnesium malabsorption also has been reported.37

IRON

Iron deficiency is common in patients with gastric resection or with celiac disease. Reduction in the mucosal surface area of the small intestine as a result of diffuse mucosal disease, intestinal resection, or intestinal bypass also can result in impaired iron absorption, potentially leading to iron deficiency38; a congenital form of iron malabsorption also has been described (see Table 101-14).39 Intestinal loss of iron from chronic gastrointestinal bleeding is, however, the most common gastrointestinal cause of iron deficiency.40 Worldwide, hookworm infection is a common cause of iron deficiency.

ZINC

Zinc, like other minerals, is malabsorbed in generalized mucosal diseases of the small intestine.41 A congenital selective defect of zinc absorption, acrodermatitis entero­ pathica, is caused by a defect in the zinc transport protein hZIP4.42

Colonic Salvage of Incompletely Absorbed Carbohydrates

In healthy people, between 2% and 20% of ingested starch escapes absorption in the small intestine47; pancreatic insuf­ ficiency or severe intestinal disorders further increase this amount.17 Carbohydrates that reach the colon cannot be absorbed by the colonic mucosa, but they can be metabo­ lized by the colonic bacterial flora. Metabolism by anaerobic bacteria results in the breakdown of oligosaccharides and polysaccharides to mono- and disaccharides, which are metabolized further to lactic acid; short-chain (C2 to C4) fatty acids (SCFAs) such as acetate, propionate, and butyr­ ate; and to odorless gases, including hydrogen, methane, and carbon dioxide.48 Studies in normal subjects have suggested that the bacte­ rial metabolism of starch to small carbohydrate moieties is a rapid process in the normal colon. The rate-limiting step in the overall conversion of polysaccharides to SCFAs appears to be the conversion of monosaccharides to SCFAs.17 Colonic absorption of SCFAs results in a reduction of the osmotic load and, as a result, in mitigation of osmotic diar­ rhea.49 In normal subjects, more than 45 g of carbohydrates must reach the colon to cause diarrhea, and up to 80 g of carbohydrates per day can be metabolized by bacteria to SCFAs; approximately 90% of these SCFAs are absorbed by colonic mucosa50 (Fig. 101-1). Chronic carbohydrate malab­ sorption causes adaptive changes in bacterial metabolic activity that result in an even higher efficiency of the bacte­ rial flora to digest carbohydrates,51 although at the expense of increased flatus production (see later).

Diarrhea CHO

Generalized malabsorption can cause deficiency of copper and selenium.43,44 In Menkes disease (kinky hair disease), an inherited disorder of cellular copper transport, selective intestinal copper malabsorption results (see later on). It is uncertain whether malabsorptive diseases result in deficien­ cies of chromium and manganese.41

MECHANISMS THAT COMPENSATE FOR MALABSORPTION

OA

OA H2, CO2

OTHERS

Gas

90% 20-90%

Caloric salvage

(CH4)

CHO

CHO

ROLE OF THE COLON

The colon has the capacity to absorb a limited number but a wide variety of substances and nutrients including sodium, chloride, water, oxalate, short chain fatty acids, calcium, and vitamin K. Although colonic nutrient absorption does not play a major role in health, the nutritive role of the colon in patients with severe malabsorption is clinically rele­ vant.45 Colonic preservation of malabsorbed nutrients also can result in symptoms and complications of malabsorp­ tion,46 such as colonic hyperabsorption of oxalate, which contributes to formation of renal stones (see later on).

Figure 101-1.  Carbohydrate metabolism and absorption of metabolic products in the colon. Up to 80 g of carbohydrate that reaches the colon can be metabolized by colonic bacteria to organic acids—lactic acid and the short-chain fatty acids acetate, proprionate, and butyrate—and to hydrogen, carbon dioxide, and methane. Approximately 90% of the organic acid produced is absorbed by colonic mucosa, which permits salvage of calories. Osmotic diarrhea results when organic acids that escape absorption and carbohydrate that escapes bacterial metabolism accumulate in the colon. Between 20% and 90% of gases produced in the colon is absorbed by the colonic mucosa; the remainder is excreted as flatus. CHO, carbohydrate; OA, organic acid.

Chapter 101  Maldigestion and Malabsorption Long-chain fatty acids bind calcium in the colon, thereby increasing the amount of sodium oxalate that is absorbed.60 Fatty acids with chain lengths longer than 12 carbons can cause diarrhea, because they increase mucosal permeability and inhibit colonic absorption of fluid and electrolytes.61 An increase in colonic permeability due to long-chain fatty acids also may be a contributing factor for the increased colonic oxalate absorption seen in patients with steatorrhea and hyperoxaluria.62 Patients with short bowel syndrome can gain caloric energy from colonic absorption of medium-chain fatty acids, coming from medium-chain triglyceride supplemen­ tation, if they have at least part of the colon in continuity with the remaining small intestine.63 In the rat colon, absorp­ tion of octanoate is not affected by the simultaneous pres­ ence of other luminal substrates.64

100

% Hydrogen in breath

80 r = –0.94 60

40

20

0 0

400

800

1200

Total flatus volume (mL/6 hr) Figure 101-2.  Relationship between flatus volume and colonic hydrogen absorption during fasting (open circles) and after ingestion of 12.5 grams of lactose (closed circles). At high flatus volumes, the fraction of hydrogen that is excreted in breath decreases to approximately 20% of total hydrogen excretion. The remaining 80% is excreted in flatus. (From Hammer HF. Colonic hydrogen absorption: Quantification of its effect on hydrogen accumulation caused by bacterial fermentation of carbohydrates. Gut 1993;34:818.)

Because SCFAs have caloric values between 3.4 and 5.95 kcal/g,52 their colonic absorption can contribute posi­ tively to overall calorie balance. In patients with short bowel syndrome, colonic salvage of malabsorbed carbohy­ drates can save up to 700 to 950 kcal/day, provided that a substantial part of the colon remains in continuity with the small intestine.53 Not all SCFAs are absorbed by the colon, and those that are not absorbed contribute to osmotic diarrhea. The beneficial effects of colonic bacterial carbohydrate metabolism may be accompanied by side effects due to gas production (see Chapter 16). Up to 10-fold differences in the volume of gas produced in the colon have been observed in normal persons.54 The colon also can absorb gas. If intra­ colonic gas volumes are low, up to 90% of the volume of intracolonic gas can be absorbed; if gas volumes are high, however, this proportion can decrease to 20%54 (Fig. 101-2). Therefore, persons who have the disadvantage of producing more gas in their colons have an additional disadvantage of absorbing a smaller fraction of the gas. Gas produced from bacterial carbohydrate metabolism is odorless. The odor of flatus is due to volatile sulfur-containing substrates that result from bacterial metabolism of protein.55 Impaired colonic salvage of carbohydrates has been suggested to contribute to the diarrhea in Crohn’s disease56 and ulcerative colitis.57 Bacterial carbohydrate metabolism may be lessened by antibiotic treatment.58 In some patients, antibiotic-associated diarrhea may be the result of impaired colonic salvage of carbohydrates that normally are not absorbed or the result of dietary fiber that can accumulate in stool because of decreased bacterial fermentation.59

Role of the Colon in Fat Malabsorption

Long-chain triglycerides or fatty acids, which constitute most dietary fat, cannot be absorbed by the human colon.

Colonic Salvage of Calcium

Although most unabsorbed calcium is insoluble when it reaches the terminal ileum,65 preservation of at least half of the colon in patients with extensive small bowel resection improves calcium absorption by about 40%, compared with calcium absorption in patients who have an ileostomy.66 Absorption of calcium requires solubilization of calcium salts. Bacterial metabolism of dietary fiber or incompletely absorbed carbohydrates can help solubilize calcium by causing a decrease in the pH of luminal contents in the colon. Once calcium is solubilized, it can contact the cecal mucosa, which in the rat has been demonstrated to be the site with the highest calcium absorption rate per surface area in the entire intestine.65 Calcium solubilization in the colon from bacterial fermentation of malabsorbed lactose also can occur in patients with lactose malabsorption, because in this condition, the bioavailability of calcium from milk is greater than that from mineral water.67 In addi­ tion to their effect on luminal pH, the SCFAs acetate and propionate, which are products of bacterial metabolism of lactose, have been shown to enhance calcium absorption directly in the human colon.68

ROLE OF INTESTINAL TRANSIT IN THE SALVAGE OF MALABSORBED NUTRIENTS

The lower parts of the gastrointestinal tract do not normally contact nutrients, and when they do, intestinal transit time is prolonged.69,70 This delay in transit could contribute to the compensation mechanisms in malabsorptive diseases; however, nutritional salvage by this mechanism has not been quantitated.

CLINICAL FEATURES AND EVALUATION Diagnosis of malabsorption requires suspecting its presence, confirming its existence, and demonstrating its cause. Malabsorption usually is suspected on the basis of the patient’s history, signs and symptoms, or findings on routine laboratory evaluations. Malabsorption of an ingested nutri­ ent or substrate can be confirmed by measuring its increased stool concentration or its decreased serum concentration or urinary excretion. Finding the cause of malabsorption often requires tests such as endoscopy with small intestinal biopsy; under certain clinical circumstances, noninvasive tests or radiologic imaging are helpful in providing a spe­ cific diagnosis.

1741

1742

Section X  Small and Large Intestine Table 101-4  Symptoms and Signs of Malabsorption and Relevant Pathophysiology SYMPTOM OR SIGN Gastrointestinal Diarrhea

Abdominal distention, flatulence Foul-smelling flatulence or stool Pain Ascites Musculoskeletal Tetany, muscle weakness, paresthesias Bone pain, osteomalacia, fractures Cutoneous and Mucosal Easy bruisability, ecchymoses, petechiae Glossitis, cheilosis, stomatitis Edema Acrodermatitis, scaly dermatitis Follicular hyperkeratosis Hyperpigmented dermatitis Thin nails with spoon-shaped deformity Perifollicular hemorrhage Spiral or curly hair Other Weight loss, hyperphagia Growth and weight retardation, infantilism Anemia Kidney stones Amenorrhea, impotence, infertility Night blindness, xerophthalmia Peripheral neuropathy Fatigue, weakness Neurologic symptoms, ataxia

PATHOPHYSIOLOGIC EXPLANATION Osmotic activity of carbohydrates or short-chain fatty acids Secretory effect of bile acids and fatty acids Decreased absorptive surface Intestinal loss of conjugated bile acids Ileal resection Severe ileal mucosal disease Congenital defects of the ileal sodium–bile acid cotransporter Bacterial gas production from carbohydrates in colon, small intestinal bacterial overgrowth Malabsorption of proteins or intestinal protein loss Gaseous distention of intestine Protein loss or malabsorption Malabsorption of vitamin D, calcium, magnesium, and phosphate Protein, calcium, or vitamin D deficiency; secondary hyperparathyroidism Vitamin K deficiency and vitamin C deficiency (scurvy) Vitamin B complex, vitamin B12, folate, or iron deficiency Protein loss or malabsorption Zinc and essential fatty acid deficiency Vitamin A deficiency Niacin deficiency (pellagra) Iron deficiency Malabsorption of vitamin C Malabsorption of vitamin C Nutrient malabsorption Nutrient malabsorption in childhood and adolescence Iron, folate, or vitamin B12 deficiency Increased colonic oxalate absorption Multifactorial (including protein malabsorption, secondary hypopituitarism, anemia) Vitamin A deficiency Vitamin B12 or thiamine deficiency Calorie depletion, iron and folate deficiency, anemia Vitamin B12, vitamin E, or folate deficiency

SUSPECTING AND CONFIRMING THE PRESENCE OF MALABSORPTION History and Physical Examination

Table 101-4 lists symptoms and signs suggestive of malab­ sorption, although virtually all can have causes other than malabsorption. For example, greasy stools might indicate malabsorption, but a greasy appearance also can be due to mucus in stool. Floating of stool in the toilet water can be due to a high stool fat content, but it also can be caused by high gas content. Nevertheless, such symptoms and signs are helpful in raising the clinician’s index of suspicion and in guiding the physician as to which specific laboratory tests, structural evaluations, or function tests should be ordered. The current obesity epidemic has led to a changing picture of malabsorption; for example, few patients today with celiac disease are underweight at diagnosis and some are even overweight. These patients have been reported to be less likely to present with classic features, such as diarrhea or anemia. In these patients, a further increase in weight after dietary gluten exclusion may be a cause of morbidity.71

Laboratory Findings

Certain blood tests might yield abnormal results in malab­ sorption, but with rare exceptions they are not specific for malabsorptive diseases. Blood tests also can be used as a screening tool to help the physician decide how vigorously

to evaluate malabsorption. Table 101-5 lists blood tests in which abnormal results should raise the suspicion of mal­ absorption and stool tests that should be used to confirm the suspicion of malabsorption. Quantitative fecal fat measurement followed by measure­ ment of fecal chymotrypsin or elastase concentration may be helpful, both in establishing malabsorption and in dif­ ferentiating between pancreatic and intestinal causes of malabsorption. Low levels of serum β-carotene, cholesterol, triglycerides, and calcium and a prolonged prothrombin time suggest malabsorption of fat and fat-soluble vitamins. Low levels of vitamin B12, folate, iron, and albumin suggest malabsorption of water-soluble substances and, therefore, indicate intestinal disease rather than pancreatic or biliary disease. Severe deficiency of fat-soluble vitamins might indicate intestinal or biliary disorders. Low levels of plasma citrulline are associated with destructive small intestinal disease, such as celiac disease, or can follow intestinal resection,72 although fasting plasma citrulline tests are poor predictors of enterocyte dysfunction in clinical practice; an oral citrulline generation test has been proposed to improve its predictive value.73

DIAGNOSTIC APPROACH Clinical Clues to the Presence of Specific Diseases Clinical clues (Table 101-6) or results of laboratory tests (Table 101-7)74 can indicate the presence of a specific

Chapter 101  Maldigestion and Malabsorption Table 101-5  Laboratory Tests That Are Useful in Patients with Suspected Malabsorption and for Establishing Possible Nutrient Deficiencies TEST Blood Cell Count Hematocrit, hemoglobin Mean corpuscular hemoglobin or mean corpuscular volume White blood cells, differential Biochemical Tests (Serum) Triglycerides Cholesterol Albumin Alkaline phosphatase Calcium, phosphorus, magnesium Zinc Iron, ferritin Other Serum Tests Prothrombin time β-Carotene Immunoglobulins Folic acid Vitamin B12 Methylmalonic acid Homocysteine Citrulline Stool Tests Fat Elastase, chymotrypsin pH

COMMENT(S) Decreased in iron, vitamin B12, and folate malabsorption or with blood loss Decreased in iron malabsorption; increased in folate and vitamin B12 malabsorption Decreased in vitamin B12 and folate malabsorption; low lymphocyte count in lymphangiectasia Decreased in severe fat malabsorption Decreased in bile acid malabsorption or severe fat malabsorption Decreased in severe malnutrition, lymphangiectasia, protein-losing enteropathy Increased in calcium and vitamin D malabsorption (severe steatorrhea); decreased in zinc deficiency Decreased in extensive small intestinal mucosal disease, after extensive intestinal resection, or in vitamin D deficiency Decreased in extensive small intestinal mucosal disease or intestinal resection Decreased in celiac disease, in other extensive small intestinal mucosal diseases, and with chronic blood loss Prolonged in vitamin K malabsorption Decreased in fat malabsorption from hepatobiliary or intestinal diseases Decreased in lymphangiectasia, diffuse lymphoma Decreased in extensive small intestinal mucosal diseases, with anticonvulsant use, in pregnancy; may be increased in small intestinal bacterial overgrowth Decreased after gastrectomy, in pernicious anemia, terminal ileal disease, and small intestinal bacterial overgrowth Markedly elevated in vitamin B12 deficiency Markedly elevated in vitamin B12 or folate deficiency May be decreased in destructive small intestinal mucosal disease or intestinal resection Qualitative or quantitative increase in fat malabsorption Decreased concentration and output in exocrine pancreatic insufficiency Less than 5.5 in carbohydrate malabsorption

Table 101-6  Cardinal Clinical Features of Specific Malabsorptive Disorders DISorder

CARDINAL CLINICAL FEATURES

Adrenal insufficiency Amyloidosis

Skin darkening, hyponatremia, hyperkalemia Renal disease, nephrotic syndrome, cardiomyopathy, neuropathy, carpal tunnel syndrome, macroglossia, hepatosplenomegaly Ileal resection or disease, liver disease Flushing, cardiac murmur Variable symptoms: dermatitis herpetiformis, alopecia, aphthous mouth ulcers, arthropathy, neurologic symptoms, and (life-threatening) malnutrition; abnormal liver biochemical test results, mild iron deficiency Arthritis, aphthous mouth ulcers, episcleritis, uveitis, pyoderma gangrenosum, erythema nodosum, abdominal mass, fistulas, primary sclerosing cholangitis (PSC), laboratory signs of inflammation Chronic sinopulmonary disease, meconium ileus, distal intestinal obstruction syndrome (DIOS), elevated sweat chloride Kidney stones, dermatosis Long history of diabetes and diabetic complications Bloating and cramping, intermittent diarrhea Previous intestinal surgery or trauma, Crohn’s disease Migratory necrolytic erythema Symptoms and signs of thyroid disease Recurrent infections Other ischemic organ manifestations; abdominal pain with eating (chronic mesenteric ischemia) Enlarged mesenteric or retroperitoneal lymph nodes, abdominal mass, abdominal pain, fever Urticaria pigmentosum, peptic ulcer Acquired immunodeficiency syndrome History of pancreatitis, abdominal pain; or alcoholism; large-volume fatty, oily stools History of travel to endemic areas Jaundice, itching Dysphagia, Raynaud’s phenomenon, skin tightening Previous intestinal surgery, motility disorder (scleroderma, pseudo-obstruction), small intestinal diverticula, strictures History of travel to endemic area Specific history of exposure, living in or travel to endemic area, immunosuppression, abdominal mass or intestinal obstruction, ascites Lymphadenopathy, fever, arthritis, cerebral symptoms, heart murmur Peptic ulcers, diarrhea

Bile acid deficiency Carcinoid syndrome Celiac disease Crohn’s disease Cystic fibrosis Cystinuria, Hartnup’s disease Diabetes mellitus Disaccharidase deficiency Gastrointestinal fistulas Glucagonoma Hyperthyroidism, hypothyroidism Hypogammaglobulinemia Intestinal ischemia Lymphoma Mastocytosis Mycobacterium-avium complex infection Pancreatic insufficiency Parasitic infection Primary biliary cirrhosis Scleroderma Small intestinal bacterial overgrowth Tropical sprue Tuberculosis Whipple’s disease Zollinger-Ellison syndrome

1743

1744

Section X  Small and Large Intestine Table 101-7  Laboratory Tests That Are Useful in the Differential Diagnosis of Malabsorption TEST Blood Cell Count Acanthocytes Nuclear remnants in erythrocytes (Howell-Jolly bodies) White blood cells, differential Platelets Other Tests ESR, C-reactive protein Ferritin Iron Liver biochemical tests Immunologic Markers Immunoglobulins Allergen-specific IgE Autoantibodies (e.g., ANA) HLA-DQ2 or HLA-DQ8 Antimitochondrial autoantibodies HIV-ELISA/Western blot Neuroendocrine Markers ACTH, cortisol Chromogranin A 5-Hydroxyindoleacetic acid in urine Gastrin* Glucagon* Serum TSH Somatostatin* Tissue transglutaminase antibodies, EMA Stool Tests Occult blood test Ova and parasites Leukocytes

COMMENT Abetalipoproteinemia Splenic atrophy in celiac disease, inflammatory bowel disease, radiation enteritis, amyloidosis Eosinophilia in eosinophilic gastroenteritis and parasitic disease Low lymphocyte count in lymphangiectasia, tuberculosis, protein-losing enteropathy Low CD4+ count in AIDS Increased in inflammatory diseases Increased in Crohn’s disease, Whipple’s disease, lymphoma Increased in inflammatory diseases, lymphoma; decreased in iron deficiency Decreased in celiac disease, chronic occult intestinal bleeding, chronic inflammatory diseases Increased in primary biliary cirrhosis and other liver diseases, celiac disease IgA deficiency, immunodeficiency syndromes IgE-mediated hypersensitivity Connective tissue diseases Celiac disease, refractory sprue Primary biliary cirrhosis AIDS Abnormal values in Addison’s disease Elevated in neuroendocrine tumors Elevated in carcinoid syndrome Elevated in Zollinger-Ellison syndrome Elevated in glucagonoma Decreased in hyperthyroidism; increased in hypothyroidism Elevated in somatostatinoma (normal in duodenal somatostatinoma) Celiac disease Erosive or ulcerative intestinal disease or tumor Repeated samples may be needed to detect Giardia lamblia Present in some inflammatory diseases of the intestine

*Perform this test if there is a strong suspicion of an underlying neuroendocrine tumor.77 ACTH, adrenocorticotropic hormone; AIDS, acquired immunodeficiency syndrome; ANA, antinuclear antibodies; EMA, endomysial antibodies; ESR, erythrocyte sedimentation rate; HIV, human immunodeficiency virus; HLA, human leukocyte antigen; Ig, immunoglobulin; TSH, thyroid-stimulating hormone.

underlying disease or can help in the differential diagnosis. In addition, the following questions may be helpful and should be asked as part of the history before physical examination: Has the patient undergone previous surgery, such as gastric or small bowel resection or a gastrointestinal bypass operation? Is there a family or childhood history of celiac disease? Is there a history of travel to endemic areas of tropical sprue, giardiasis, or other gastrointestinal infections? Is there excessive alcohol consumption? Does the patient have a history of chronic pancreatitis or symptoms suggesting a pancreatic tumor? Does the patient have clinical features of thyrotoxicosis, Addison’s disease, Whipple’s disease, biliary or liver disease, or diabetic neuropathy? Does the patient eat a diet high in poorly absorbable carbohydrates (sweeteners such as sorbitol or fructose) or fat substitutes or an unbalanced diet that could result in malnutrition? Is there a likelihood of human immunodeficiency virus infection? Is the patient receiving treatment with a drug that can cause malabsorption? Does the patient have a history of stem cell or organ trans­ plantation or abdominal radiation? Does the patient have a history of extraintestinal manifesta­ tions of inflammatory bowel disease, celiac disease, or Whipple’s disease

A rational approach to establishing the cause of malab­ sorption can require several diagnostic steps. Depending on the clinician’s background, the availability of different tests, and the patient’s preferences, different diagnostic approaches may be used. If time constraints are not a consideration, a stepwise approach may be used, starting with noninvasive evaluations that can guide further invasive procedures or even provide a diagnosis. In other instances, the physician may choose a more invasive test in the hope of reaching a diagnosis with the fewest possible tests and in the shortest possible time. Diagnostic approaches differ depending on the epidemiologic or ethnic background of an individual patient. For example, if parasitic infections are a likely possibility, stool examination can provide a rapid diagnosis by noninvasive testing. In populations with a very low prevalence of lactose intolerance, a secondary cause of lactose malabsorption is more likely than it would be in populations with a high prevalence of acquired primary lactase deficiency, and therefore, additional tests are appropriate. The sequence of tests thus depends on the affected per­ son’s symptoms and history, as well as results of previous testing (Table 101-8). Tests that can detect the most common causes of malabsorption or are noninvasive or inexpensive usually should be performed initially (first-line tests). In some patients, testing for rarer causes of malabsorption and use of more invasive or more expensive tests may be necessary to establish the diagnosis (second-line tests). For unusually difficult cases, additional tests may be

Chapter 101  Maldigestion and Malabsorption Table 101-8  Tests to Establish the Cause of Malabsorption Based on Main Symptoms Weight Loss, Osteomalacia or Osteopenia, Diarrhea, Suspected Steatorrhea, or Deficiency of Fat-Soluble Vitamins First-line Tests Abdominal ultrasonography Chymotrypsin and/or elastase concentration in stool EGD with small intestinal biopsies Endomysial and tissue transglutaminase antibodies Laboratory tests (complete blood cell count, white blood cell differential, cholesterol, triglycerides, electrolytes, calcium, magnesium, serum ALT, AST, AP, bilirubin levels, prothrombin time, serum albumin level, erythrocyte sedimentation rate and C-reactive protein, TSH) Ova, parasites, and leukocytes in stool Second-line Tests Abdominal CT, MRI Endoscopic examination of the terminal ileum, including ileal biopsies ERCP/MRCP More extensive laboratory investigation (immunoglobulins, human immunodeficiency virus ELISA, antinuclear antibodies, ferritin, food allergen– specific IgE, adrenocorticotropic hormone, cortisol, chromogranin A, gastrin, urinary 5-HIAA) Quantitative fecal fat Quantitative small intestinal culture or breath tests for bacterial overgrowth Small bowel series/small bowel MRI Special staining of small intestinal biopsies (e.g., Congo red for amyloid, PAS for Whipple’s disease) Therapeutic trial of pancreatic enzymes, antibiotics (tetracycline, metronidazole), or a gluten-free diet Video capsule endoscopy Tests in Unusually Difficult Cases (Third-line Tests) Abdominal angiography Antienterocyte antibodies Endoscopic ultrasonography Enteroscopy, including biopsies Magnetic resonance angiography Positron emission tomography Serum or plasma glucagon, somatostatin Somatostatin (octreotide) scan Special tests of intestinal biopsies (e.g., flow cytometry of intraepithelial lymphocytes for lymphoma and refractory celiac disease, PCR for Tropheryma whipplei or other infective organisms, chromogranin A stain for enteroendocrine cells) Spiral CT of the pancreas Tests for bile acid malabsorption Tube test for exocrine pancreatic secretion (secretin, cholecystokinin, or Lundh test) Bloating, with or without Diarrhea First-line Tests Fructose H2 breath test Lactose H2 breath test Lactose tolerance test Second-line Tests Chymotrypsin and/or elastase concentration in stool EGD with duodenal biopsies Endomysial and tissue transglutaminase antibodies Genetic testing for hypolactasia Quantitative small intestinal culture or breath tests for bacterial overgrowth Stool pH (in patients with diarrhea) Anemia and Suspected Malabsorption Microcytic or hypochromic anemia (low MCV, MCH) EGD with duodenal biopsies Endomysial and tissue transglutaminase antibodies Iron, ferritin, and transferrin in serum Ova and parasites in stool Video capsule endoscopy Macrocytic anemia (high MCV, MCH) First-line Tests Folic acid in serum or red blood cells Vitamin B12 in serum Second-line Tests in Cases of Vitamin B12 Deficiency CT, small bowel series, enteroclysis, video capsule endoscopy EGD with gastric and duodenal biopsies Endomysial and tissue transglutaminase antibodies Evaluation of ileum (e.g., double balloon enteroscopy with biopsy, colonoscopy to ileum with biopsy) Ova and parasites in stool Quantitative small intestinal culture or breath tests for bacterial overgrowth Schilling test (with and without intrinsic factor) Second-line Tests in Cases of Folate Deficiency EGD with duodenal biopsies Endomysial and tissue transglutaminase antibodies ALT, alanine aminotransferase; AP, alkaline phosphatase; AST, aspartate aminotransferase; CT, computed tomography; EGD, esophagogastroduodenoscopy; ERCP, endoscopic retrograde cholangiopancreatography; 5-HIAA, 5-hydroxyindoleacetic acid; IgE, immunoglobulin E; MCH, mean corpuscular hemoglobin; MCV, mean corpuscular volume; MRCP, magnetic resonance cholangiopancreatography; MRI, magnetic resonance imaging; PAS, periodic acid–Schiff; PCR, polymerase chain reaction; TSH, thyroid-stimulating hormone.

1745

1746

Section X  Small and Large Intestine required that may be available only in specialized centers (third-line tests). For some disorders, such as bile acid malabsorption, lactose malabsorption, and bacterial overgrowth, it may be difficult to establish a causal link between symptoms and the malabsorbed substrate. In these conditions, observation of the response to therapy may be critical in proving or disproving a causal relationship.

ANATOMIC INVESTIGATIONS

Endoscopic examination of the stomach, duodenum, or ileum and histologic examination of mucosal biopsy speci­ mens can establish a diagnosis of some conditions causing malabsorption. The role of radiologic imaging examinations is limited mostly to answering questions about abdominal regions not easily accessible to endoscopy, such as parts of the small intestine, parenchymatous organs, the peri­ toneal cavity, the mesentery, or the retroperitoneum. Capsule endoscopy and magnetic resonance imaging (MRI) are contributing to making these areas more accessible to diagnostic evaluation. Radiologic studies of the small intes­ tine can show evidence of stasis, blind loops, diverticula, fistulas, rapid transit, and other abnormalities that can assist in diagnosis (see later on).

ENDOSCOPY, BIOPSY, AND DUODENAL ASPIRATION Endoscopy

Endoscopic inspection of the duodenal mucosa can provide clues to some causes of malabsorption. Aphthae suggest Crohn’s disease, and small, diffuse, white, punctate lesions can be seen in primary or secondary lymphangiectasia. Mosaic-like scalloping of duodenal folds (Fig. 101-3) and reduction in the number of duodenal folds are highly suggestive of villus atrophy in celiac disease, although these abnormalities may be seen in other diseases (see Chapter 104).75 Villus atrophy may be seen endoscopically using magnification endoscopy76 and chromoendoscopy with indigocarmine staining; however, a normal duodenal fold pattern should not deter the endoscopist from taking

Figure 101-3.  Endoscopic image showing scalloping of the duodenal folds in a patient with celiac disease.

mucosal biopsy specimens. Endocrine tumors causing mal­ absorption, such as duodenal gastrinomas or somatostatino­ mas or ampullary tumors obstructing the pancreatic duct, also can be detected during endoscopy. If ileal disease is the suspected cause of malabsorption, visual examination and biopsy of the ileal mucosa may be required to establish a diagnosis; this can be accomplished by retrograde intuba­ tion of the ileum at colonoscopy or by double-balloon endoscopy.

Biopsy

Examination of endoscopic biopsy specimens from the duo­ denum may be diagnostic or highly suggestive of a variety of small bowel disorders resulting in malabsorption (Table 101-9); follow-up small intestinal biopsy can be used to assess treatment effects. Duodenal biopsy specimens should be obtained from patients with atypical or nonspecific gas­ trointestinal symptoms, including abdominal pain, bloat­ ing, and weight loss, and should not be limited only to patients with diarrhea.77,78 Endoscopic biopsy is an adequate substitute for jejunal suction biopsy,79 and its advantages over capsule biopsy are that multiple specimens are more easily obtained and focal or patchy lesions can be identified and targeted for sampling.80 Compared with duodenal biop­ sies, endoscopically obtained biopsies from the jejunum to find changes of celiac disease are helpful in only very few patients.81 The adequacy of mucosal biopsy specimens is a function of their size and the number obtained.82 If large specimens can be obtained using jumbo biopsy forceps, they can be oriented on a piece of filter paper before they are put into a fixing solution83; two or three jumbo biopsy specimens usually are sufficient to allow histologic section­ ing parallel to the villi and crypts. Specimens also may be obtained with smaller forceps, although the number of specimens obtained must then be increased to four to six. Specimens can be inspected with a low-power dissecting microscope or by magnification endoscopy to obtain an initial impression of the villus architecture and to ensure proper orientation before they are placed in formalin. The diagnostic yield of biopsy is influenced by the distri­ bution of histologic abnormalities, which in some diseases is diffuse but in other diseases is patchy. Tropical diarrhea malabsorption syndrome (tropical sprue; see Chapter 105), abetalipoproteinemia, and immunodeficiency usually result in a diffuse alteration of small intestinal mucosa. Thus, a completely normal appearance of a duodenal biopsy speci­ men rules out these disorders. Primary lymphangiectasia has a patchy distribution, so that a single mucosal biopsy might not rule out the disorder (see Chapter 28). Patchy distribution also has been described for the histologic changes in some patients with celiac disease, especially when symptoms are subtle, although this disorder usually affects the small intestine diffusely.84 Other possible sources of error and misdiagnosis include poorly oriented speci­ mens and those obtained proximally, where peptic injury can be the cause of mucosal alterations. Additional biopsy specimens from the stomach and duodenal bulb can help the pathologist to establish the extent of peptic injuries in the upper gastrointestinal tract and to interpret inflamma­ tory changes in the duodenum in relation to these lesions. Distortion of villus architecture over Brunner’s glands or lymphoid aggregates, common in the duodenum, should be interpreted with caution. Specific histologic features may be diagnostic for some rare causes of malabsorption (see Table 101-9)85 such as Whipple’s disease (Fig. 101-4), abetalipoproteinemia or hypobetalipoproteinemia, intestinal lymphangiectasia, giar­ diasis (Fig. 101-5), lymphoma, or collagenous sprue. In most

Chapter 101  Maldigestion and Malabsorption Table 101-9  Causes of Malabsorption That Can Be Diagnosed by Small Bowel Biopsy CAUSE OF MALABSORPTION Generalized Histologic Abnormalities Abetalipoproteinemia, hypobetalipoproteinemia Collagenous sprue (Chapter 104) Mycobacterium-avium complex infection (Chapter 33) Whipple’s disease (Chapter 106) Patchy Histologic Abnormalities Amyloidosis (Chapter 35) Crohn’s disease (Chapter 111) Eosinophilic gastroenteritis (Chapter 27) Lymphangiectasia (Chapter 28) Lymphoma (Chapter 29) Mastocytosis (Chapter 35) Parasites (Giardia lamblia, Strongyloides stercoralis, coccidia) (Chapters 109, 110)

MAIN HISTOLOGIC FEATURES Lipid accumulation and vacuolization of enterocytes Collagenous band below atrophic epithelium Acid-fast bacilli, foam cells Foamy macrophages with PAS-positive inclusion bodies Congo red–stained deposits with apple-green birefringence in polarized light Epithelioid granulomas and characteristic focal inflammation Eosinophilic infiltration Ectatic lymph vessels Clonal expansion of lymphocytes Diffuse infiltration with mast cells Some parasites may be seen on histologic examination

PAS, periodic acid–Schiff. Modified from Riddell RH. Small intestinal biopsy: Who? how? what are the findings? In Barkin JS, Rogers AI, editors. Difficult Decisions in Digestive Diseases. Chicago: Year Book; 1989. p 326.

Figure 101-5.  Small bowel biopsy specimen from an immunocompetent patient with giardiasis. A normal-appearing villus and adjacent pearshaped organisms with red-staining nuclei are evident. (Courtesy of Cord Langner, MD.)

A

B Figure 101-4.  Duodenal biopsy specimen from a patient with Whipple’s disease. A, Hematoxylin and eosin staining shows villus blunting. The lamina propria is infiltrated with pale-staining foamy macrophages. (Courtesy of Cord Langner, MD.) B, High-power view demonstrates purple-red macrophages. (Periodic acid–Schiff stain.) (Courtesy of Günter J. Krejs, MD, Graz, Austria.)

patients with small intestinal disorders, however, histologic examination is not diagnostic85 (Table 101-10) and reveals a spectrum of mucosal responses ranging from infiltration by lymphocytic cells to a flat mucosa with villus atrophy and crypt hyperplasia (Fig. 101-6). In many parts of the world, celiac disease is by far the most common cause of this type of histologic alteration, but a definite diagnosis of celiac disease cannot be established by mucosal biopsy alone (see Chapter 104). Some disease states can be identified only with use of special histologic stains, such as Congo red (intestinal amy­ loidosis), periodic acid–Schiff (PAS) (Whipple’s disease), or immunohistochemical techniques for detecting refractory celiac disease, small intestinal lymphoma, or enteroendo­ crine insufficiency (see later on). Polymerase chain reaction analysis of intestinal biopsy specimens for Tropheryma whipplei may be helpful in evaluating patients in whom Whipple’s disease is suspected (see Chapter 106).86 In cases in which these diseases are a possibility, the clinician has to request these specific tests. Measurement of mucosal enzyme activities in a jejunal biopsy can be used to confirm disaccharidase deficiency, although this is not recom­ mended for routine clinical use.

1747

1748

Section X  Small and Large Intestine Table 101-10  Malabsorptive Diseases with Abnormal but Not Diagnostic Small Intestinal Histologic Findings Increased Lymphocyte Infiltration with or without Crypt Hyperplasia AIDS enteropathy (Chapter 33) Celiac disease (Chapter 104) Infection (due to Giardia lamblia, Cryptosporidium [Chapter 109]; viral enteritis [Chapter 107]) Tropical sprue (Chapter 105) Flat Lesion with or without Mucosal Inflammation Celiac disease (Chapter 104) Drug-induced enteropathy (NSAIDs, colchicine, neomycin) Food protein hypersensitivity (rye, barley, egg, fish, rice, poultry) (Chapters 9, 27) Immunodeficiency (hypogammaglobulinemia) (Chapter 2) Immunoproliferative small intestinal disease (IPSID) (Chapters 29, 104, and 115) Infection (due to Giardia lamblia, Cryptosporidium) (Chapter 109) Intestinal transplantation (Chapter 34) Lymphoma (Chapter 29) Nongranulomatous chronic idiopathic enterocolitis Autoimmune enteropathy Prolonged folate or cobalamin deficiency Protein-calorie malnutrition Traumatic injury Tropical sprue (Chapter 105) Atrophic Lesion Chronic radiation damage (Chapter 39) Cicatrizing Crohn’s disease (Chapter 111) Diffuse lymphoma (Chapter 29) Idiopathic diarrhea of infancy (microvillus inclusion disease) (Chapter 96) Unresponsive gluten sensitivity (lymphoma or ulcerative jejunitis) (Chapter 104)

A

AIDS, acquired immunodeficiency syndrome; NSAIDs, nonsteroidal antiinflammatory drugs. Modified from Riddell RH. Small intestinal biopsy: Who? how? what are the findings? In Barkin JS, Rogers AI, editors. Difficult Decisions in Digestive Diseases. Chicago: Year Book; 1989. p 326.

Aspiration

Fluid aspirated from the descending part of the duodenum may be examined microscopically for Giardia lamblia (see Chapter 109) or cultured to detect bacterial overgrowth in patients with diffuse small intestinal motility disorders (see Chapters 97 and 102).

Video Capsule Endoscopy

Video capsule endoscopy (VCE) is an increasingly popular technique for diagnosing diseases of the small intestine. VCE was initially introduced for evaluating suspected bleeding in the small intestine, but subsequently it has been used to diagnose a wider range of diseases such as Crohn’s disease, celiac disease, and other malabsorptive disorders. In several studies, lesions suggesting Crohn’s disease were detected by VCE when they had been missed by conven­ tional diagnostic procedures.87 These reports need to be interpreted carefully, because no biopsy specimens were obtained, and long-term evaluations to confirm the diagno­ sis are lacking. VCE appears to be superior to conventional radiologic imaging of the small intestine and to computed tomogra­ phy (CT) with small bowel enteroclysis to detect subtle mucosal changes, such as aphthous or erosive lesions of the small intestine.87 In celiac disease, the detection of villus atrophy by VCE has a good correlation to villus atrophy seen in duodenal biopsy specimens,88,89 but it is

B Figure 101-6.  Duodenal biopsy specimen from a patient with untreated celiac disease. A, Subtotal villus atrophy, crypt elongation, and lymphoplasmacytic infiltration of the lamina propria can be seen. B, High-power view demonstrates villus blunting with increased intraepithelial lymphocytes. (Hematoxylin and eosin stain.) (Courtesy of Cord Langner, MD.)

questionable if this procedure can detect subtle changes, such as Marsh 1 and 2 lesions. Changes on VCE that suggest villus atrophy are scalloping, mosaic pattern, and fissuring. In a recent study of VCE in patients with celiac sprue, villus atrophy was seen in the duodenum and jejunum in 59% of cases, in the duodenum only in 32%, and in the jejunum only in 3%.89 In refractory celiac disease, VCE can detect changes such as ulcerations and strictures that suggest T-cell lymphoma but that are missed by conventional techniques.90 This test may be used in patients with established malabsorption in whom no diagnosis has been established despite extensive diagnostic workup.

Chapter 101  Maldigestion and Malabsorption ABDOMINAL IMAGING Small Bowel Follow-through and Small Bowel Enteroclysis

The principal role of small bowel radiologic series in evalu­ ating malabsorption is to identify focal or diffuse abnormali­ ties and alterations that predispose to bacterial overgrowth, including diverticula, stagnant loops of intestine, general­ ized intestinal hypomotility or dilatation, intestinal fistulas, and tumors.91 Small bowel enteroclysis is preferred to small bowel follow-through examinations, because distention of the lumen results in better demonstration of the small bowel contour.92 Double-contrast enteroclysis, in which intubation of the upper jejunum is used to instill contrast material directly into the upper jejunum, has a higher sensitivity than small bowel series for detecting mucosal changes, although it is less acceptable to the patient and can miss focal changes in the duodenum, such as diverticula. Use of an intravenous agent such as glucagon to reduce motility enables overlapping loops of small intestine to be separated and imaged more distinctly. Alterations associated with diffuse, localized, or distal mucosal changes that might have been missed by proximal mucosal biopsy also may be identified. Normal findings on small bowel series do not rule out intestinal causes of mal­ absorption and should not dissuade the clinician from per­ forming biopsy of the small intestine. Ulcerations and strictures may be seen in various malab­ sorptive disorders, including Crohn’s disease, radiation enteritis, celiac disease, intestinal lymphoma, and tubercu­ losis. Aphthous ulcers and cobblestoning of the mucosa, either alone or with thickened and distorted folds, are fea­ tures of Crohn’s disease but also can be present in other conditions. Reduced numbers of jejunal folds and an increased number of and thickening of ileal folds can suggest celiac disease.91 Mass lesions can be found with intestinal lymphoma or, rarely, with hormone-producing tumors. The disadvantage of conventional enteroclysis is that direct imaging of the bowel wall and surrounding structures is not possible, and overlapping bowel loops potentially impair complete visualization of the whole small bowel— hence the rationale for combining enteroclysis with CT or MRI scanning.93

Abdominal Computed Tomography

Abdominal CT for small bowel investigation is performed after administration of oral or intravenous contrast agents.94 Small intestinal CT scanning is useful to detect focal intes­ tinal lesions, such as thickening of the small bowel wall in Crohn’s disease or small intestinal lymphoma, intestinal fistula, and dilated bowel loops; however, mild mucosal changes such as aphthae in Crohn’s disease or villus atrophy of various causes are missed by this technique. Diffuse thickening of the small bowel may be seen in Whipple’s disease and in graft-versus-host disease.94 In some cases of celiac disease, reversal of the jejunoileal fold pattern is observed.95 CT is a sensitive test for detecting enlarged abdominal lymph nodes, which can be present in disorders such as Whipple’s disease, small bowel lymphoma, or small intestinal inflammatory diseases such as Crohn’s disease. Evidence for pancreatic disease that may be detected on CT includes calcifications of the pancreas, dilatation of the pan­ creatic duct, and pancreatic atrophy. Tumors obstructing the pancreatic duct or hormone-secreting neuroendocrine tumors also can be located by CT.

Magnetic Resonance Imaging of the Small Intestine

MRI may be used to image the small intestine either with administration of oral contrast solutions or by enteroclysis. Segmental bowel wall thickening with inflammatory involvement of the mesentery, cobblestoning, and ulcer­ ations may be seen in Crohn’s disease; this method is very sensitive for demonstrating complications of Crohn’s disease, such as intestinal fistula formation. In celiac dis­ ease, small bowel MRI with oral contrast can demonstrate small intestinal dilatation, mucosal thickening, and an increased number of folds in the ileum (ileal jejunization) with flattening of the jejunal folds.96 Most of these signs also are found in other inflammatory diseases of the intestine, but the fold pattern abnormalities are most specific for celiac disease.96 This method also is very useful to detect changes suggesting complications, such as lymphoma or carcinoma. With MRI enteroclysis, subtle mucosal changes might be missed and be more evident on conventional small bowel enteroclysis97 or capsule endoscopy.87 Because MRI or CT imaging of the small intestine requires no tube place­ ment, these techniques have largely replaced classic small bowel enteroclysis.

Other Radiologic Studies

A plain film of the abdomen may be helpful to detect pan­ creatic calcifications if exocrine pancreatic insufficiency is suspected. However, morphologic signs of chronic pancre­ atitis alone do not prove a pancreatic cause of malabsorp­ tion, because the function of the exocrine pancreas must be severely impaired before malabsorption becomes evident. A plain film of the abdomen also can document dilated loops of intestine; dilated loops predispose to small bowel bacterial overgrowth or suggest the presence of an obstruction. Endoscopic retrograde cholangiopancreatography (ERCP) can help establish the cause of pancreatic insufficiency (see Chapter 59). It can help distinguish chronic pancreatitis from pancreatic tumor and can document pancreatic duct stones. ERCP and endoscopic ultrasound (EUS) are the methods of choice for documenting various causes of biliary obstruction. Magnetic resonance pancreatography (MRCP) is increasingly being used to replace diagnostic ERCP. If a neuroendocrine tumor (e.g., gastrinoma, somatostatinoma) is the suspected cause of malabsorption, an indium-111 octreotide scintigraphic scan, positron emission tomogra­ phy (PET), or an EUS examination of the pancreas may be helpful in establishing the diagnosis or demonstrating the extent of disease (see Chapters 31 and 32). Transabdominal ultrasound examination has the advan­ tage of no radiation exposure and therefore can be used in pregnant patients. Ultrasonography often is used to investi­ gate the pancreas, although the sensitivity for detecting pan­ creatic tumors is lower than that of ERCP or CT. Nevertheless, obstruction of the biliary tract, pancreatic calcifications, dilatation of the pancreatic duct, or stones within the pan­ creatic duct may be demonstrated. Ultrasound examination also may be used to document thickening of the bowel wall, abscesses, and fistula in Crohn’s disease.

NONINVASIVE EVALUATION OF GASTROINTESTINAL DIGESTIVE AND ABSORPTIVE FUNCTION

Some conditions causing malabsorption can be diagnosed with noninvasive tests, although, as pointed out in Table 101-11, diagnostic accuracy may be limited, and further tests may be necessary to identify underlying diseases or to

1749

Section X  Small and Large Intestine Table 101-11  Malabsorptive Diseases or Conditions in Which Noninvasive Tests Can Establish Malabsorption or Provide a Diagnosis DISEASE OR CONDITION

DIAGNOSTIC TEST(S)*

COMMENT(S)

Lactose malabsorption

Lactose hydrogen breath test Lactose tolerance test Fructose hydrogen breath test 14 C-d-xylose breath test Glucose hydrogen breath test Schilling test with and without antibiotics SeHCAT test, 14C-TCA test

Tests do not differentiate between primary and secondary lactose malabsorption Questionable clinical relevance A predisposing factor should be sought if the result of any of the tests is positive

Fructose malabsorption Small intestinal bacterial overgrowth (see Chapter 102) Bile acid malabsorption Exocrine pancreatic insufficiency Vitamin B12 malabsorption

Does not differentiate between primary and secondary causes To establish malabsorption in chronic pancreatitis Variable sensitivity and specificity, depending on type of test and stage of the disease Test is performed without intrinsic factor and, depending on result with intrinsic factor, with antibiotics or with pancreatic enzymes (see text). Further tests are needed if small intestinal bacterial overgrowth, terminal ileal disease, or pancreatic disease is suspected

Quantitative fecal fat determination Fecal elastase or chymotrypsin, tubeless tests (see Chapters 56 and 59) Schilling test

*See text for diagnostic accuracy of the different tests listed. SeHCAT, selenium-75-homotaurocholic acid test; TCA, taurocholic acid.

differentiate primary and secondary causes. Apart from pro­ viding a diagnosis, tests evaluating gastrointestinal absorp­ tive and digestive function may be helpful in evaluating complex disease presentations. For most or all of the fol­ lowing tests, the potential benefits with regard to the costs of workup or to patient acceptability have not been estab­ lished. Because test procedures and analytical methods can vary among laboratories,98 each laboratory should establish its own reference values for these tests.

Fat Malabsorption

Quantitative Fecal Fat Analysis The van de Kamer method is the quantitative titration of fatty acid equivalents in which the results are expressed as fecal output of fat in grams per 24 hours. This method is considered the gold standard for fecal fat analysis.99 Modi­ fications in which the extracted fats are weighed rather than titrated100 have an excellent correlation with the results of the van de Kamer method. Near-infrared reflectance analy­ sis may be a less-cumbersome method to quantify fecal fat output in stool collections101 because it requires less han­ dling of stool by laboratory personnel, but it still requires a 48- to 72-hour collection to exclude the influence of day-today variability; the stool must be mixed before a sample is obtained for analysis. Fecal fat excretion of less than 7 g per day with a fat intake of 100 g per day usually is considered normal. However, the volume effect of diarrhea by itself increases fecal fat output to levels of up to 14 g per day (secondary fat malabsorption)102 (Fig 101-7); this latter value could be used as the upper limit of normal in patients with diarrhea. Diet is important in considering causes of steatorrhea; for example, elevated fecal fat values can be observed in patients consuming a diet rich in the fat substitute olestra.100 Quantitative fecal fat analysis is available routinely now only in a few centers. Reasons for the limited clinical use of quantitative fecal fat measurements are as follows: (1) If the main symptom of malabsorption is chronic diar­ rhea, measurement of fecal fat might not influence the sub­ sequent workup, because the diagnostic tests performed to establish the etiology of diarrhea are similar to the tests for the workup of steatorrhea. (2) An elevated fecal fat level usually cannot differentiate among biliary, pancreatic, and enteric causes of malabsorption. (3) In many patients with

20

15 Fecal fat (g/day)

1750

10

5

0 0

400

800

1200

1600

2000

Fecal weight (g/day) Figure 101-7.  Graph showing fecal fat output (average of a three-day stool collection) plotted as a function of fecal weight from normal subjects (�) and from subjects with induced diarrhea (). The washout effect of diarrhea increases fecal excretion of fat to levels above the upper limit of normal (7 g/day). With significant diarrhea, therefore, a fecal fat excretion of 14 g/day should be used as the upper limit of normal. (From Fine KD, Fordtran JS. The effect of diarrhea on fecal fat excretion. Gastroenterology 1936;12;102.)

severe steatorrhea, the stools have the characteristic porridge-like appearance, and quantitative studies are not necessary to establish fat malabsorption. (4) Fat absorption may be normal despite malabsorption of other nutrients, so a normal fat balance does not imply normal absorptive function of the gastrointestinal tract. (5) Finally, accuracy depends on quantitative stool collections for 48 to 72 hours, adherence to a diet comprising 80 to 100  g of fat daily, and a diet diary to determine fat intake. Science aside, quantitative fecal fat analysis has never been popular among patients, physicians, or laboratory personnel per­ forming the test. Despite the limitations of quantitative fecal fat analysis, it nevertheless is still useful in several clinical circum­ stances: to establish malabsorption and avoid nutritional

Chapter 101  Maldigestion and Malabsorption deterioration103 when overt features of intestinal or pancre­ atic disorders are lacking, as in some cases of osteoporosis, osteomalacia, anemia, or weight loss; to monitor treatment in patients with established malabsorptive disorders, such as exocrine pancreatic insufficiency or short bowel syndrome; to estimate fecal calorie loss in patients with severe malabsorption syndromes; and to quantitate fecal fat excretion in patients with diarrhea who have undergone ileal resection, thereby distinguishing steatorrhea due to bile acid deficiency from secretory diarrhea caused by bile acid loss.104 Semiquantitative Fat Analysis For the acid steatocrit test,105 a sample of stool is diluted 1 : 3 with distilled water in a test tube. The diluted stool is homogenized, and a 500-µL aliquot is pipetted into a tube. Then 100 mL of 5M HClO4 is added to allow better fat extraction and separation of the lipid layer. An aliquot of the diluted stool-HClO4 mixture is put into a nonheparin­ ized microcapillary tube and sealed on one end. After cen­ trifugation of this aliquot at 13,000 rpm for 15 minutes, the fatty layer (FL) and the solid layer (SL) are measured, and the acid steatocrit (AS) is determined according to the following equation: AS (%) = [FL (FL + SL)] × 100 An acid steatocrit of less than 31% is normal. In a small study, the acid steatocrit for random spot stool samples had a high sensitivity and specificity for detection of steator­ rhea, compared with the van de Kamer method, which is performed on a 72-hour stool collection. A linear correla­ tion also was found between results obtained with the acid steatocrit and those of the van de Kamer method, although results were quite divergent in some patients.105 Because quantitative fecal fat measurements are based on 48- to 72-hour stool collections (to minimize the effect of day-to-day variability in fecal fat excretion), however, the acid steatocrit cannot be expected to replace quanti­ tative measurement of fat output in borderline cases or in cases in which exact measurement of fecal fat loss is required. Qualitative Fecal Fat Analysis Fat analysis by microscopic examination of random stool samples might provide a clue to the presence of steatorrhea, although it cannot be used to exclude steatorrhea; its sole advantage is its ease of performance. A sample of stool is placed on a glass slide to which several drops of glacial acetic acid and Sudan III stain are added. Acidification of stool samples improves fat extraction and separation of the lipid layer.105 The slide is held over a flame-burner and the acidified mixture is heated to boiling and then examined while still warm for presence of orange fat globules. A count of up to 100 globules with a diameter less than 4 mm per high-power field is normal.6 Results of qualitative fat analysis by this method and of quantitative fat analysis do not correlate very well.106 In a small study, Sudan staining of spot stool samples had a sensitivity of 78% and a specificity of 70% for the detection of steatorrhea.105 A quantitative microscopic method of counting and measuring fat globules using the Sudan stain has been shown to correlate well with chemically measured fecal fat output.107 Breath Tests for Fat Malabsorption The principle of the 14C-triolein breath test is to measure 14 CO2 in the breath after ingestion of a triglyceride that has

been radiolabeled with 14C. Fat malabsorption results in decreased pulmonary excretion of 14CO2.108 Because of erroneous results in a variety of metabolic and pulmonary diseases, lack of sensitivity in mild malabsorption, radiation exposure to the patient, cost of the substrate, and the need for expensive equipment, this test has not found widespread acceptance for clinical use; the nonradioactive isotope 13C is used to label triglycerides instead (see later on). Serum Tests for Fat Malabsorption The photometric measurement of β-carotene at 456 nm109 has been suggested as a useful screening test for steatorrhea, although experience with this technique is limited; values less than 100 mg per 100 mL suggest the presence of steator­ rhea, and values less than 47 mg per 100 mL strongly indi­ cate steatorrhea. Concentrations in excess of 100 mg per 100 mL do not exclude mild steatorrhea, although they make steatorrhea with fat losses in excess of 16 g per day very unlikely. Normal values also have been established in the pediatric population.110 β-Carotene can be falsely low in patients with liver disease or in alcoholics who consume a diet deficient in β-carotene. Disorders in lipoproteins or intake of carotene-containing food additives also can influence the results.

Carbohydrate Malabsorption

The hydrogen breath test is a noninvasive test that takes advantage of the fact that in most people, bacterial metabo­ lism of carbohydrate results in accumulation of hydrogen, which then is absorbed by the colonic mucosa and excreted in the breath. Using different carbohydrates, such as lactose or fructose, the hydrogen breath test can be used to detect malabsorption of these carbohydrates. Measurement of breath hydrogen excretion after ingestion of lactulose has been used to assess orocecal transit time, and glucose has been used as a substrate to detect small bowel bacterial overgrowth, although sensitivity and specificity are poor.111 Unfortunately, up to 18% of people are hydrogen nonexcre­ tors,112 and in these persons, hydrogen breath test results may be falsely negative because hydrogen is metabolized by bacteria to methane. Such limitations and pitfalls of breath hydrogen testing have to be taken into account when test results are interpreted.113 The diagnosis of lactose malabsorption is established if an increase in breath hydrogen concentration of greater than 20 parts per million over baseline occurs after ingestion of 20 to 50 grams of lactose. An increase within the first 30 minutes after ingestion of lactose has to be disregarded, because it may be due to bacterial degradation of lactose in the oral cavity. Up to four hours may be required for the increase in breath hydrogen concentration to occur. Breath hydrogen measurements obtained before and at 30, 60, 90, 180, and 240 minutes after ingestion of 50 grams of lactose provide the best diagnostic yield with the fewest possible measurements.112 The lactose hydrogen breath test still is performed by most clinicians for evaluating lactose malabsorption, but this test can miss the disorder in hydrogen nonexcretors. In these patients, a lactose tolerance test—measurement of blood glucose levels before and 30 minutes after ingestion of 50 grams of lactose—can be used. An increase in glucose concentration of less than 20 mg/dL over baseline within 30 minutes of ingestion of 50 grams of lactose indicates lactose malabsorption. The lactose tolerance test has a lower sensitivity than the lactose hydrogen breath test for diagnos­ ing lactose malabsorption.112

1751

1752

Section X  Small and Large Intestine Lactase deficiency in acquired primary lactase deficiency (adult-type hypolactasia) is not caused by mutations in the gene coding for intestinal lactase (LPH gene). It has been shown, however, that a single-nucleotide polymorphism (SNP), either the C or T nucleotide −13910 upstream of the LPH gene, is involved in the regulation of intestinal lactase expression.114 A CC genotype at −13910 C/T is associated with acquired primary lactase deficiency (adult-type hypo­ lactasia), whereas TC and TT genotypes are linked with lactase persistence.115,116 This polymorphism can be used as a diagnostic test for adult-type hypolactasia.116 This SNP is only associated with adult-type hypolactasia in whites; other SNPs are linked to adult-type hypolactasia or lactase persistence in Africans.117 In patients with diarrhea, a stool test to detect a fecal pH lower than 5.5 can serve as a quali­ tative indicator of carbohydrate malabsorption.118 In the research setting, fecal carbohydrates can be determined by the anthrone method, which measures carbohydrates on a weight basis.119 By contrast, the reducing sugar method gives results on a molar basis and, therefore, provides infor­ mation about the osmotic activity of malabsorbed carbohy­ drates.17 Total SCFAs and lactic acid, which are the products of bacterial carbohydrate metabolism, can be measured in stool by titration.120 Individual SCFAs can be determined by gas chromatography.121

Protein Malabsorption

The classic test to quantify protein malabsorption, measure­ ment of fecal nitrogen content in a quantitatively collected stool specimen,12 rarely is used today. For research pur­ poses, a combined 14C–octanoic acid–13C–egg white breath test, accompanied by measurement of the urinary output of phenol and p-cresol, has been used to assess the effect of gastric acid on protein digestion.122 In this method, labeling of the 13C–egg protein test meal with 14C–octanoic acid allows the simultaneous measurement of protein assimila­ tion and gastric emptying rate. Phenol and p-cresol are the quantitatively most important phenolic compounds in the feces and urine and are specific metabolites of tyrosine, produced by bacterial fermentation in the colon. They result from protein that has escaped digestion and absorption in the small intestine and are rapidly absorbed in the colon, detoxified, and excreted in urine. Recovery of higher amounts of urinary phenols observed after omeprazole treatment in the study of this test indicated an increased availability of protein in the colon.

Vitamin B12 (Cobalamin) Malabsorption

The Schilling Test The Schilling test can be used clinically to distinguish between gastric and ileal causes of vitamin B12 deficiency. Because both intrinsic factor and hydrochloric acid are pro­ duced by parietal cells in humans, alternative approaches to diagnosing pernicious anemia are to document atrophic gastritis by endoscopy and biopsy, to confirm achlorhydria by acid secretion analysis and increased serum gastrin levels, and to look for antibodies in the serum directed against parietal cells or intrinsic factor.26,123 Because the intrinsic factor used in the Schilling test is of bovine origin, the test is not commercially available in some countries. The Schilling test is performed by administering a small oral dose of radiolabeled vitamin B12 and, simultaneously or within one or two hours, a large intramuscular flushing dose of nonradiolabeled vitamin B12. The unlabeled B12 saturates vitamin B12 carriers; thus, any radioactive vitamin B12 absorbed by the intestine is excreted in the urine. If less than 7% to 10% of the administered dose is recovered in urine within 24 hours, vitamin B12 malabsorption is con­

firmed. To specify the site of vitamin B12 malabsorption, a second phase of the Schilling test is performed subsequently with oral administration of intrinsic factor. In patients with pernicious anemia, the results of the Schilling test normal­ ize after oral administration of intrinsic factor.24,123 Patients with pancreatic exocrine insufficiency might have an abnormal result on the Schilling test, with or without added intrinsic factor, but results normalize with addition of pancreatic enzymes (Chapter 59). In small bowel bacterial overgrowth the results of the Schilling test can improve after antibiotic therapy (Chapter 102). In ileal disease or following ileal resection, abnormal results of the Schilling test persist despite intrinsic factor. Schilling test results are normal in patients with dietary vitamin B12 defi­ ciency, in protein-bound (food-bound) vitamin B12 malab­ sorption,24 and sometimes in congenital transcobalamin II deficiency.124 False-positive results on the Schilling test can result from renal dysfunction or inadequate urine collec­ tion.123 The value of this test is diminished by the need for accurately timed urine collections. Results in the 5% to 10% excretion range often are difficult to interpret. A varia­ tion of the standard Schilling test is the dual-isotope or single-stage Schilling test, in which two different cobalamin isotopes are given simultaneously, one of them bound to intrinsic factor. This makes it possible to perform the first two phases of the Schilling test in one day; however, the results of this test are not as accurate as those obtained with the standard protocol.24,123 Serum Test for Vitamin B12 and Folate Deficiency Measurements of serum cobalamin and folate concentra­ tions commonly are used to detect deficiency states of these vitamins. The sensitivity and specificity of these tests are unknown because no gold standard test has been estab­ lished and because serum levels do not always correlate with body stores.24,125 Furthermore, results of vitamin B12 levels vary with different commercial tests.126 Several causes of misleading serum cobalamin levels have been established. Serum vitamin B12 levels can be normal despite depleted body stores in small intestinal bacterial overgrowth (as a result of production of inactive cobalamin analogs by the bacteria), liver disease, myeloproliferative disorders, congenital transcobalamin II deficiency, and with high levels of intrinsic factor antibodies. In contrast, oral contraceptives, pregnancy, and folate deficiency can cause low serum cobalamin levels despite normal body stores.123 Therefore, if there is a high suspicion, especially for cobala­ min deficiency, parenteral replacement with monitoring of the clinical response is recommended.126 Measurement of methylmalonic acid, homocysteine, and holotranscobala­ min are of limited clinical use in establishing vitamin B12 deficiency.126 Serum folate concentrations decrease within a few days of dietary folate restriction, even if tissue stores are normal. Feeding also influences serum folate levels; therefore, deter­ mination of folate in the fasting state is recommended. Mea­ surement of red blood cell folate concentration has been considered a better estimate of folate tissue stores than serum folate levels by some authors.123

Small Intestinal Bacterial Overgrowth

Tests for the diagnosis of bacterial overgrowth are covered in more detail in Chapter 102. Briefly, tests used to diagnose bacterial overgrowth are the quantitative culture of a small intestinal aspirate (which is considered to be the gold stan­ dard diagnostic test), measurement of deconjugated bile acids or vitamin B12 analogs in intestinal aspirates, measure­ ment of serum folate, and several breath tests, including the

Chapter 101  Maldigestion and Malabsorption 14

C-glycocholate breath test, the 14C-d-xylose breath test, the lactulose hydrogen breath test, and the glucose hydrogen breath test. The rationale for the breath tests is the produc­ tion by intraluminal bacteria of volatile metabolites (i.e., 14 CO2 or H2), from the administered substances, which can be measured in the exhaled breath.

Exocrine Pancreatic Insufficiency

Pancreatic function tests are discussed in detail in Chapters 56 and 59. Invasive pancreatic function tests require duo­ denal intubation and measurement of pancreatic enzyme, volume, and bicarbonate output after pancreatic stimulation by a liquid test meal (the Lundh test) or by injection of cholecystokinin (CCK) or secretin. Noninvasive tests include measurement of fecal chymotrypsin or elastase concentra­ tion, the fluorescein dilaurate test, and the N-benzoyl-ltyrosyl para-aminobenzoic acid (NBT-PABA) test. Elastase has a higher sensitivity for the detection of exocrine pan­ creatic insufficiency compared with chymotrypsin,127 but the specificity of elastase is low.128 Measurement of pancreatic enzymes and components of pancreatic fluid in duodenal aspirates obtained during endoscopy and after intravenous stimulation with secretin and CCK can have an excellent correlation with the more classic intubation tests for secretory function.129

Bile Salt Malabsorption

In patients with steatorrhea due to ileal disease or resection, bile salt malabsorption usually is present, but measurement of bile acid malabsorption is of limited clinical value. In patients with diarrhea without steatorrhea, bile salt malab­ sorption may be present in the absence of overt ileal disease, and in such cases, measurement of bile salt absorption is helpful. Measurement of Fecal Bile Acid Output Elevated fecal bile acid concentrations or output can indi­ cate intestinal bile acid malabsorption.130 Under steady-state conditions, the increased fecal bile acid output reflects increased hepatic synthesis of bile acids.131 In severe bile acid malabsorption, fecal bile acid output can be reduced if hepatic synthesis of bile acids is impaired. The measure­ ment can be performed by enzymatic methods or by gas chromatography. This test requires a quantitative stool col­ lection, and the analytic techniques are time consuming and require considerable expertise. Enzymatic methods may be unreliable in severe steatorrhea.132 Carbon-14–Taurocholate Bile Acid Absorption Test The 14C-taurocholate bile acid absorption test requires a 72-hour stool collection after ingestion of radioactively labeled bile acid. The rate of intestinal bile acid absorption is calculated from the fecal recovery of 14C-labeled tauro­ cholic acid (14C-TCA). Normal values for this test have been established in normal persons with laxative-induced diar­ rhea, because diarrhea by itself can increase fecal losses of bile acids,131 presumably because of accelerated intestinal transit.133 Clinical limitations of this test are that it requires substantial analytical work, access to a gamma camera, and a time-consuming stool collection. Therapeutic Trial of Bile Acid–Binding Resins (Cholestyramine) A therapeutic trial of cholestyramine or other bile acid– binding resins can be used to diagnose bile acid malabsorp­ tion as a cause of diarrhea. It is, however, controversial to what extent a clinical response to cholestyramine correlates with the presence of bile acid malabsorption, because cho­

lestyramine may have a nonspecific constipating effect in patients with diarrhea from other causes. Failure of diarrhea to remit within three days of initiation of cholestyramine makes bile acid malabsorption an unlikely cause of diar­ rhea; however, some patients respond only to large doses of cholestyramine. In patients with established bile acid malabsorption in whom no improvement is obtained with bile acid–binding resins, it is very unlikely that bile acid malabsorption is the cause of diarrhea. In these patients, bile acid malabsorption is considered a secondary phenomenon due to a washout effect.131 In patients with severe bile acid malabsorption resulting in steatorrhea, cholestyramine might even aggra­ vate fat malabsorption and diarrhea.3 Therefore, without further testing for bile acid malabsorption, neither a positive nor a negative result of a therapeutic trial of cholestyramine constitutes proof of the presence or absence, respectively, of bile acid malabsorption. Selenium-75–Labeled Homotaurocholic Acid Test The radioactive taurocholic acid analog used for this test is resistant to bacterial deconjugation. After it has been admin­ istered orally, the patient undergoes serial gamma scintig­ raphy to measure whole-body bile acid retention or, as suggested by some authors, bile acid retention in the gall­ bladder.134 A limitation of this test is that normal values for bile acid retention, which are used to compare normal and abnormal bile acid absorption, were obtained only in healthy persons without diarrhea.135 However, secondary bile acid malabsorption can be induced by diarrhea itself and is proportional to the stool weight, as demonstrated with the 14C-TCA test.131,133 For this test to be clinically useful, normal values need to be established for patients with diarrhea. Finally, this test is very time consuming, because bile acid retention needs to be measured either four or seven days (depending on the protocol) after the bile acid administration. d-Xylose

Test

Absorption of the pentose d-xylose is facilitated by passive diffusion. Approximately 50% of the absorbed d-xylose is metabolized, and the remainder is excreted in urine. After an overnight fast, a 25-g dose of d-xylose is swallowed, and the patient is encouraged to drink sufficient volumes of fluid to maintain good urine output. Urine is collected for the next five hours. As an alternative, one hour after inges­ tion of d-xylose, a venous sample may be obtained.136 Less than 4 grams (16% excretion) of d-xylose in the urine col­ lection or a serum xylose concentration below 20 mg/dL indicates abnormal intestinal absorption. The traditional urine test appears to be more reliable than the one-hour blood test. False-positive results occur if the duration of urine col­ lection is too short or if the patient is dehydrated or has renal dysfunction, significant ascites, delayed gastric emp­ tying, or portal hypertension. d-Xylose absorption may be normal in patients with only mild impairment of mucosal function or with predominantly distal small intestinal disease. Because d-xylose is susceptible to bacterial metabo­ lism, absorption is diminished in patients with bacterial overgrowth, although the test has a poor sensitivity for detecting this.137 The test is of limited clinical value today and mostly has been replaced by small intestinal biopsy.

Intestinal Permeability Tests

Intestinal permeability tests mostly are used in studies of the pathophysiology of intestinal disorders; they do not provide a specific diagnosis.138

1753

1754

Section X  Small and Large Intestine Most current permeability tests are based on the diffe­ rential absorption of mono- and disaccharides. Damage to the mucosa can result in an increased permeability for disaccharides and oligosaccharides consequent to epithelial injury, and it can result in a decreased permeability of monosaccharides due to reduction of mucosal surface area.139 Absorption is measured by urinary excretion. Expression of results as the absorption ratio of the mono- and disaccharide minimizes the influences of gastric emptying, intestinal transit, renal and hepatic function, and variations in time of urine collections.140 Increased intestinal permeability has been shown to predict the development of Crohn’s disease or relapse in patients with this disease.141 In celiac disease, the finding of considerably increased permeability is a sensitive marker for advanced disease; permeability tests also have been used to judge response to a gluten-free diet142 or to screen firstdegree relatives for celiac disease. Elevated serum amino­ transferase levels in patients with celiac disease correlate with increased intestinal permeability.143 Disturbances of intestinal permeability have been documented in users of nonsteroidal anti-inflammatory drugs,144 in inflammatory joint disease,145 and in diabetic diarrhea.146

Carbon-13 Breath Tests

The increasing availability of stable isotopes has raised interest in replacing radioactive 14C with nonradioactive 13C for breath tests.111,147 In malabsorptive diseases, 13C-labeled substrates have been evaluated in the diagnosis of steator­ rhea,148 small bowel bacterial overgrowth, and exocrine pancreatic insufficiency149 and in the evaluation of the digestibility of egg protein. Because of concerns about diag­ nostic accuracy, cost, and limited availability, these tests have not gained widespread acceptance.

MALABSORPTION IN SPECIFIC SITUATIONS AND DISEASE STATES LACTOSE MALABSORPTION AND INTOLERANCE

Deficiency of the intestinal brush border enzyme lactase can lead to lactose malabsorption, which can result in lactose intolerance. Unlike other intestinal disaccharidases, which develop early in fetal life, lactase levels remain low until the 34th week of gestation.150 Transient lactase deficiency in premature infants can lead to symptoms of lactose malabsorption, such as diarrhea, until normal intestinal lactase activity develops. In rare cases, enzyme deficiency is manifest at the time of birth and is permanent, and con­ genital lactase deficiency (OMIM #223000)* is diagnosed. Reversible lactase deficiency can occur at all ages as a result of transient small intestinal injury associated with acute diarrheal illnesses. Acquired primary lactase deficiency (adult-type hypolac­ tasia, OMIM #223100) is the most common form of lactase deficiency worldwide. Most populations lose considerable lactase activity in adulthood.151 The decline in lactase activ­ ity is a multifactorial process that is regulated at the gene transcription level152 and leads to decreased biosynthesis or retardation of intracellular transport or maturation of the enzyme lactase-phlorizin hydrolase.153 In whites, a single-

*The Online Mendelian Inheritance in Man (OMIM) system assigns numbers to specific diseases according to a continuously updated catalog of human genes and genetic disorders (http://www.ncbi.nlm.nih.gov/omim/).

Table 101-12  Ethnic Groups with High and Low Prevalence Rates of Acquired Primary Lactase Deficiency (Adult-Type Hypolactasia) Lactase Deficiency–Predominant Ethnic Groups (60%-100% of Population is Lactase Deficient) Middle East and Mediterranean: Arabs, Israeli Jews, Greek Cypriots, southern Italians Asia: Thais, Indonesians, Chinese, Koreans Africa: South Nigerian, Hausa, Bantu North and South America: Alaska Natives, Canadian and U.S. Native Americans, Chami Indians Lactase Persistence–Predominant Ethnic Groups (2%-30% of Population is Lactase Deficient) Northern Europeans Africa: Hima, Tutsi, Nomadic Fulani India: Indians from Punjab and New Delhi areas Data from Johnson JD. The regional and ethnic distribution of lactose malabsorption. In: Paige DM, Bayless TM, editors. Lactose Digestion. Clinical and Nutritional Implications, 1st ed. Baltimore: Johns Hopkins University Press; 1981, p 11.

nucleotide polymorphism (SNP) −13910 T/C upstream of the gene coding for the enzyme lactase-phlorizin hydrolase (LPH gene) has been found to be involved in the regulation of lactase-phlorizin hydrolase.114 The CC genotype of the SNP −13910 T/C upstream of the LPH gene is associated with adult-type hypolactasia; TC and TT genotypes are linked with lactase persistence.115 In other populations, such as some African and sub-Saharan African populations, the −13910*T polymorphism is not associated with lactase persistence.117,154 Because it is present in most of the adult human population, this form of lactase deficiency has to be considered normal, rather than abnormal. Lactase deficiency usually produces symptoms only in adulthood, although lactase levels in affected persons start to decline during childhood.155 Lactase activity persists in most adults of Western European heritage156 (Table 101-12). Even in this group, the activity of lactase is only approxi­ mately half the activity of sucrase and less than 20% of the activity of maltase.155 Accordingly, in these persons, lactase activity is much more susceptible to a reduction in function with acute or chronic gastrointestinal illnesses. In lactose malabsorbers, it may be unclear whether lactose malabsorption is from acquired primary lactase deficiency or is the consequence of another small intestinal disorder. Therefore, in the individual lactose malabsorber, especially with an ethnic background associated with a low prevalence of acquired primary lactase deficiency, it may be necessary to exclude other malabsorptive disorders, such as celiac disease. The main symptoms of lactose intolerance are bloating, abdominal cramps, increased flatus, and diarrhea. The development of bloating and abdominal cramps pre­ sumably is associated with increased perception of luminal distention by gas,157 because no clear relation has been observed between the amount of lactose ingested and the severity of symptoms.158 Ingestion of as little as 3 g of lactose to as much as 96 g of lactose may be required to induce symptoms in persons with lactose malabsorption.159 Gastro­ intestinal symptoms, including diarrhea, have been shown to be more severe in adults with shorter small intestinal transit time,160 but no such relation between intestinal transit and symptoms is observed in children.161 Also, in pregnant women and in thyrotoxic patients with Graves’ disease, changes in intestinal motility play a role in the clinical manifestation of lactose malabsorption.162,163 To make a diagnosis of lactose intolerance, and in view of the poor correlation between lactose malabsorption and lactose intolerance, it is very important to monitor symptoms

Chapter 101  Maldigestion and Malabsorption

H2 concentration (ppm × 10), symptoms

10

INCOMPLETE ABSORPTION AND INTOLERANCE OF FRUCTOSE

Lactose malabsorption with intolerance

8

H2 concentration 6 4 Symptoms

2 0 0

60

A 7 H2 concentration (ppm × 10), symptoms

120

180

240

Minutes Lactose malabsorption without intolerance

6 5

H2 concentration

4 3 2

Symptoms

1 0 0

B

60

120

180

240

Minutes

Figure 101-8.  Illustration of the role of symptoms in determining the clinical importance of lactose malabsorption. Assessment of the clinical relevance of an abnormal lactose hydrogen breath test is made by monitoring abdominal symptoms (bloating, cramps, pain) during the test. Breath hydrogen concentration in parts per million (ppm) and gastrointestinal symptoms using an arbitrary scoring system for two different patients are plotted on the graphs. A, The patient has symptoms associated with an increase in breath hydrogen concentration and therefore can be con­sidered to have lactose intolerance. B, The patient has no increase in symptoms, although the breath hydrogen concentration increases con­siderably; therefore, the patient has lactose malabsorption without lactose intolerance.

during a lactose hydrogen breath test and to confirm that any symptoms experienced by the patient during the test are truly those that the patient complains of and that they are associated with a significant increase in breath hydrogen levels (Fig. 101-8). Adult-type hypolactasia also may be a risk factor for developing osteoporosis and bone fractures, owing to patients’ avoidance of dairy products164 or interference with calcium absorption.165 Patients in whom a clear association can be established between symptoms and lactose malabsorption should be educated about a lactose-reduced or lactose-free diet. Yogurt may be tolerated by such patients166 and provides a good source of calcium. Consuming whole milk or chocolate milk, rather than skim milk, and drinking milk with meals can reduce symptoms of lactose intolerance, presumably as a result of prolongation of gastric emptying. Alternatively, supplementation of dairy products with lactase of microbio­ logic origin may be suggested.167 Furthermore, because many carbohydrates other than lactose are incompletely absorbed by the normal small intestine,47 and because dietary fiber also may be metabolized by colonic bacteria, persistence of some symptoms while the patient is on a lactose-free diet is not uncommon. It also must be kept in mind that symptoms arising after ingestion of dairy prod­ ucts may be due to milk protein allergy or to fat intolerance rather than lactose intolerance.

Fructose is found in modern diets either as a constituent of the disaccharide sucrose or as the monosaccharide and it is used as a sweetener in a variety of food items. The average daily intake of fructose varies from 11 to 54 g around the world.168 Fructose as a constituent of sucrose is absorbed by a well-characterized absorptive system that integrates enzy­ matic hydrolysis of the disaccharide sucrose by sucrase and transfer of the resulting two monosaccharides, glucose and fructose, through the apical membrane of the epithelial cell; the absorptive capacity for fructose that is not accompanied by glucose, however, is relatively small.169 The normal absorption capacity of fructose depends on other nutrients as well and is poorly understood. Healthy subjects have the capacity to absorb up to 25 g of fructose, but many have malabsorption and intolerance with intake of 50 g of fructose.170 Ingesting food that contains fructose in excess of glucose can result in symptoms such as abdominal bloating or diar­ rhea171 and especially can provoke symptoms in patients with irritable bowel syndrome172; it has been suggested that as little as 3 g of fructose can precipitate symptoms in patients with functional bowel disorders. Women might complain more often of symptoms and exhibit more fruc­ tose malabsorption than do men; there is no adaptation to regular consumption of fructose.173 Fructose malabsorption usually is identified by a positive result on a hydrogen breath test after ingestion of 25 or 50 g of fructose. Because the fructose content in fruit and in soft drinks usually is less than 8 g per 100 g of fruit or drink, the amounts of fructose used in the hydrogen breath test are not physiologic, and no data are available on how many asymptomatic people would have a positive test result; fructose contents of 30 to 40 g per 100 g can be present in chocolate, caramel, and pralines.174 In a group of patients with isolated fructose malabsorp­ tion, no defect could be detected of the gene encoding the luminal fructose transporter (GLUT5).175 It is therefore unlikely that patients who present with gastrointestinal symptoms have a true defect of intestinal fructose absorp­ tion; it is more likely that they belong to a subset of people in whom ingestion of foods rich in fructose provokes symp­ toms related to other disorders, such as irritable bowel syndrome, or resulting from unique, but not necessarily abnormal, colonic bacterial activity. The latter is suggested by a study in asymptomatic and symptomatic persons with fructose malabsorption in which it was demonstrated that the disappearance rate of fructose in anaerobic, but not aerobic, stool cultures was significantly elevated in the symptomatic group compared with the asymptomatic group.176 Testing for fructose malabsorption by the hydrogen breath test may be useful in identifying patients in whom dietary restriction of foods with excessive fructose content may be beneficial in treating bloating and diarrhea. Symptoms in these persons probably are the result of ingestion of unphys­ iologic amounts of fructose and not the consequence of a defect in fructose absorption.

ILEAL BILE ACID MALABSORPTION

Bile acid malabsorption usually is present in patients who have undergone ileal resection or bypass operations or who have severe disease of the ileum, where specific bile acid transport proteins are located. The clinical consequences of bile acid malabsorption depend on whether bile acid loss can be compensated by increased hepatic synthesis.177 Ileal

1755

1756

Section X  Small and Large Intestine resection of more than 100 cm usually results in severe bile acid malabsorption that cannot be compensated by increased hepatic synthesis; in such cases, steatorrhea results from impaired micelle formation due to decreased luminal con­ centrations of conjugated bile acids.3,177 In ileal resections of less than 100 cm, bile acid malabsorption usually can be compensated by increased hepatic synthesis, and malab­ sorbed bile acids cause secretory diarrhea rather than steat­ orrhea.3,177 Secretory diarrhea caused by or associated with bile acid malabsorption is discussed in detail in Chapter 15. Knowledge of the differing pathophysiology of steator­ rhea and of secretory diarrhea from bile acid malabsorption is important not only for understanding the clinical presen­ tation but also for choosing the appropriate therapy. In patients with compensated bile acid malabsorption, binding of bile salts in the lumen of the intestine by cholestyramine reduces diarrhea. By contrast, in decompensated bile acid malabsorption, cholestyramine further depletes the bile acid pool, thereby worsening steatorrhea. In several cases of decompensated bile acid malabsorption after extensive ileal resections, intestinal fat absorption was improved markedly by oral administration of conjugated bile acids.104,178,179 Cholylsarcosine in a dose of 2 to 3 g per meal has been reported to enhance fat absorption and nutritional status in patients with short bowel syndrome who have residual colon.104,179,180 Natural conjugated bile acids lessen severity of steatorrhea in such patients to a smaller extent.180 Improved fat absorption also was associated with decreased urinary oxalate excretion.179 A syndrome of primary bile acid malabsorption with normal ileal morphology has been reported in children who, at birth, develop severe diarrhea, severe steatorrhea, and failure to thrive and who have reduced plasma choles­ terol levels.181 In an index case, this type of bile acid mal­ absorption was shown to be caused by mutations in the ileal sodium–bile acid cotransporter gene (SLC10A2).4 Adultonset bile acid malabsorption is not caused by SLC10A2 mutations,182 and although its exact pathophysiology is unknown, accelerated intestinal transit may be a causative factor.183

AMYLOIDOSIS

Malabsorption has been reported in AL-type amyloidosis, AA-type amyloidosis, and hereditary amyloidosis (see Chapter 35). Malabsorption occurs in 5% to 13% of patients with AL or AA amyloidosis,184,185 whereas it was present in 58% of Swedish patients with familial amyloidosis.186 Fecal fat excretion can reach levels up to 60 g/day.186 Gastro­intestinal absorption of d-xylose and vitamin B12 can be reduced,186,187 and protein-losing enteropathy can develop.188 Amyloid deposits are found in the muscle layers, the stroma of the lamina propria and the submucosa, the wall of mucosal and submucosal blood vessels in the gastro­intestinal tract, and in enteric and extraenteric nerves.189,190 In many patients with amyloidosis who have diarrhea or malabsorption, or both, symptoms suggesting autonomic neuropathy are present.133,186 Autonomic neuropathy can cause rapid intestinal transit, which in turn can lead to severe diarrhea and malabsorption despite normal transport capacity of the intestinal mucosa.133 Other suggested mecha­ nisms of malabsorption in amyloidosis are decreased absorption from a physical barrier effect of amyloid depos­ its190 and small intestinal bacterial overgrowth, which also might be a consequence of autonomic neuropathy.187 Bile acid malabsorption is found in many patients with amyloidosis associated with autonomic neuropathy191 and

is caused by rapid intestinal transit rather than impaired absorptive transport in the terminal ileum.133 Diarrhea in these patients usually fails to respond to bile acid–binding agents.133 Barium studies in patients with amyloidosis usually are normal but might show thickened folds, nodular lesions, filling defects, dilatation of bowel segments, or altered transit.192 The endoscopic appearance of the gastrointestinal mucosa can show a fine granular appearance, polypoid protrusions, erosions, ulcerations, atrophic changes, and mucosal friability, but in many patients, no macroscopic changes are evident.190,192 Histologic examination demonstrates amyloid deposits in 72% of esophageal, 75% to 95% of gastric, 83% to 100% of small intestinal, and 75% to 95% of colorectal biopsy speci­ mens.185,190 Subcutaneous fat pad aspiration or biopsy can make the diagnosis more safely without having to resort to endoscopic biopsy and the potential risk of intestinal bleed­ ing. Amyloid deposits might not be seen with routine his­ tologic stains but become more evident with Congo red staining. Therapy of diarrhea in patients with amyloidosis includes attempts to prolong intestinal transit time with opioids or octreotide and to avoid further amyloid deposition in the tissue by treating the underlying disorder in AA amyloidosis, the plasma-cell dyscrasia in AL amyloidosis, and by administering colchicine in familial Mediterranean fever.

MALABSORPTION CAUSED BY DRUGS AND FOOD SUPPLEMENTS

Table 101-1321,24,25,32,102,123,193-212 lists drugs and food supple­ ments reported to induce malabsorption of vitamins, miner­ als, or nutrients, as well as the suggested pathophysiologic mechanisms by which this occurs.

MALABSORPTION AFTER GASTRIC RESECTION OR BARIATRIC SURGERY Gastric Resection

Severe steatorrhea after total and partial gastric resections has been a long-observed complication of these operations. Fecal fat excretion rates usually are between 15 and 20 g/ day,1 but values of up to 60 g/day have been reported.208 Suggested mechanisms for steatorrhea include defective mixing of nutrients with digestive secretions, lack of gastric acid and gastric lipase secretion, decreased small bowel transit time, small intestinal bacterial overgrowth, and pan­ creatic insufficiency.1,213 Studies have shown that pancreatic enzyme supplements214 and antibiotic treatments213 neither improve fat absorption nor relieve symptoms after gastric resection. Total and partial gastric resections also can result in significant protein malabsorption, whereas absorption of carbohydrates seems not to be significantly impaired. Nutrient malabsorption in these patients also can result in gastrointestinal symptoms, such as diarrhea and severe weight loss.215 Vitamin E deficiency can occur if food does not pass through the duodenum. The differential diagnosis of neuro­ logic symptoms in postgastrectomy patients should include hypovitaminosis E.216 Loss of parietal cells after total gastric resection results in diminished intrinsic factor secretion, which in turn leads to malabsorption of vitamin B12 and, in approximately 30% of patients, vitamin B12 deficiency; bacterial overgrowth and lack of release of food-bound cobalamin due to diminished

Chapter 101  Maldigestion and Malabsorption Table 101-13  Drugs and Dietary Products That Cause Malabsorption SUBSTANCE

SUBSTRATE MALABSORBED

SUGGESTED MECHANISM

Acarbose Antacids Azathioprine Biguanide (metformin)

Carbohydrate Phosphate, iron, vitamin A Generalized malabsorption Cobalamin, folate, glucose

Carbamazepine Cholestyramine Colchicine

Folate Fat, fat-soluble vitamins, bile acids Fat, xylose, nitrogen, cobalamin, carotene

Contraceptives, oral*

Folate

Ethanol

Xylose, fat, glucose, nitrogen, thiamine, cobalamin, folate

Fiber, phytates Glucocorticoids Histamine H2 receptor antagonists*

Iron, calcium, magnesium, zinc Calcium Cobalamin

Laxatives, irritant type (phenolphthalein, bisacodyl, anthraquinones) Methotrexate

Fat, glucose, xylose

Inhibition of α-glucosidase Luminal binding of substrates Villus atrophy Reduced ileal absorption of intrinsic-factorcobalamin complex; inhibition of intestinal glucose or folate absorption Inhibition of intestinal folate absorption Binding of conjugated bile salts Mucosal damage and villus atrophy at high doses (impaired processing of IF-cobalamin receptor [the cubilin-amnionless complex]) Inhibition of pteroylpolyglutamate hydrolase (folate conjugase) Mucosal damage; decreased disaccharidase activity; decreased pancreatic exocrine function and bile secretion Chelation Inhibition of calcium absorption Impaired release of food-bound B12 owing to reduced gastric acid and pepsin secretion (and reduced IF secretion) Washout effect; toxic effect on the mucosa

†

Methyldopa Neomycin

Folate, fat, cobalamin, xylose

Olestra* Orlistat Para-aminosalicylate Phenytoin

Generalized malabsorption Fat, nitrogen, fat-soluble vitamins, cobalamin, mono- and disaccharides, iron Fat-soluble vitamins Fat, fat-soluble vitamins Fat, cobalamin, folate Folate, calcium

Proton pump inhibitors*

Cobalamin

Pyrimethamine

Folate

Somatostatin analogs (e.g., octreotide)

Fat

Sulfonamides and sulfasalazine Tetracycline Thiazides Triamterene*

Folate Calcium Calcium Folate

Mucosal damage; inhibition of intestinal folate transport Mucosal damage Mucosal damage; disruption of micelle formation Binding of fat-soluble vitamins Inhibition of pancreatic lipase Unknown Inhibition of folate and calcium absorption due to luminal alkalinization; impaired vitamin D metabolism Impaired release of food-bound cobalamin by pepsin owing to reduced gastric acid secretion; small intestinal bacterial overgrowth Competitive inhibition of intestinal folate absorption Inhibition of hepatobiliary bile acid secretion; inhibition of pancreatic enzyme secretion; inhibition of cholecystokinin release Inhibition of pteroylpolyglutamate hydrolase and folate transport Precipitation of luminal calcium Decreased 1,25 dihydroxyvitamin D synthesis Competitive inhibition of intestinal folate absorption

REFERENCE(S) 193 194 195 194, 196, 197 1101 194 24, 194, 199 194 32, 194 200 21 201 102, 194 194, 200 202 194, 200 203, 212 193 24, 194 21, 200, 204 25 205 206, 207 123, 200 209 210 205, 211

*Malabsorption usually does not result in deficiency states. † Findings in case reports.

gastric acid and pepsin secretion have been implicated as additional pathogenetic factors. Iron malabsorption resulting in iron deficiency anemia also is commonly present in patients who have undergone gastric resection, although the mechanisms for iron malab­ sorption are not fully established; lack of acid secretion with resultant decreased solubilization of iron salts has been sug­ gested as a possible cause. Calcium absorption can be severely impaired in patients with gastric resections, resulting in reduced bone density.217 The mechanisms for calcium malabsorption probably are several, including decreased solubilization of calcium salts owing to the loss of gastric acid secretion, rapid

intestinal transit, low calcium intake secondary to milk intolerance, and malabsorption of vitamin D. Studies in gastrectomized rats have suggested that diminished calcium absorption after gastric resections is due mainly, if not entirely, to decreased calcium solubilization.218 By contrast, studies in humans have shown that calcium absorption is normal in patients with atrophic gastritis and in persons in whom acid secretion was inhibited by acid-inhibiting drugs.219 Treatment for patients who have undergone gastric resec­ tion should include the adequate supplementation of mal­ absorbed vitamins and minerals, to prevent serious long-term complications.220

1757

1758

Section X  Small and Large Intestine Bariatric Surgery

The number of patients undergoing bariatric surgery is increasing; indications for and procedural details of the various procedures are described in Chapter 7. These patients need to be monitored for long-term problems, such as changes in bone metabolism. Risk can even increase over time due to poor compliance with supplementation, inad­ equate intake, or ongoing malabsorption. Gastrointestinal long-term problems of bariatric surgery depend on the type of surgical procedure performed. Primarily restrictive procedures such as Roux-en-Y gastric bypass have only a mild component of noncaloric malab­ sorption. Other procedures, like the biliopancreatic diver­ sion, which has been used more extensively in the past, can result in severe malnutrition.221 Roux-en-Y gastric bypass can result in deficiency of proteins, iron, calcium, folate, vitamin B12 and vitamin D. Deficiencies in vitamin B1 are rare but potentially serious.222,223 Iron deficiency after gastric bypass can develop for several reasons, such as intolerance to red meat, diminished gastric acid secretion, and exclusion of the duodenum. Menstruat­ ing or pregnant women may be particularly predisposed to developing iron deficiency after gastric bypass surgery. Postoperatively, oral iron and vitamin C supplementation should be prescribed, because once iron deficiency has developed it may be refractory to oral treatment.224 In Roux-en-Y gastric bypass, colonization of both gastric chambers with aerobic and anaerobic bacteria has been demonstrated, resulting in a positive hydrogen breath test in 41% of subjects; no clinical symptoms such as diarrhea, malabsorption, or pneumonia could be attributed to this bacterial overgrowth.225 It has been suggested that after bariatric surgery patients should have yearly measurements of a basic metabolic panel, magnesium, complete blood count, iron studies, vitamin D, parathyroid hormone, and bone density.221 The routine and lifelong use of multivitamins is considered necessary.226

MALABSORPTION IN THE ELDERLY

Malabsorption in elderly persons should not be ascribed to the aging process; it should be evaluated just like malab­ sorption occurring in younger patients. In healthy elderly persons, small bowel histologic features are normal despite a decline in cell turnover and continual cell renewal.227,228 Malabsorption of fat has been described in chronic conges­ tive heart failure229 and in chronic intestinal ischemia (see Chapter 114), but this is not due to aging per se. Elderly persons may be more susceptible to gastrointestinal insult and subsequent decompensation of gastrointestinal func­ tion.230 Changes in pancreatic anatomy and secretion occur, but only in a minority do they result in overt pancreatic insufficiency.231 Deficiencies of some nutrients, presumably caused by malabsorption, however, may be present in elderly persons with no overt gastrointestinal disease. An increased risk of folate and vitamin B12 deficiency, despite adequate intake of these vitamins, has been reported in the elderly.232 Mal­ nutrition in the elderly can contribute considerably to mor­ bidity and mortality, although it may be difficult to ascertain whether weight loss results from altered appetite, increased catabolism, or malabsorption. Small bowel bacterial overgrowth in elderly persons with gastric hypochlorhydria from atrophic gastritis or treatment with a proton pump inhibitor usually is not associated with clinically significant malabsorption,233 but an improvement in nutritional status after antibiotic treatment has been described in some elderly patients.234 An increased preva­

lence of lactose malabsorption in the elderly may be the result of clinically unapparent small bowel bacterial overgrowth.235

CONNECTIVE TISSUE DISEASES Systemic Sclerosis

The gastrointestinal tract is involved to a variable degree in most patients with systemic sclerosis. Early pathologic changes are characterized by vasculopathy, which results in ischemia and progressive organ dysfunction.236 Typical histologic findings include atrophy of the muscle layers with increased deposition of elastin and collagen in the submucosa and serosa and between smooth muscle bundles of the muscularis externa.237 Small bowel biopsy specimens might reveal an increased number of plasma cells within the lamina propria and collagen deposits around and between lobules of Brunner’s glands in the submucosa of the duodenum.238 Malabsorption in scleroderma usually results from bacte­ rial overgrowth secondary to ineffective motility in the small intestine,239 but other factors, such as decreased mucosal blood flow,240 also can contribute. Malabsorption and bacterial overgrowth are not limited to patients with diffuse disease; they also can occur in patients with longstanding limited cutaneous systemic sclerosis.241 Elevated serum concentrations of motilin and CCK have been described in patients with systemic sclerosis and fat malab­ sorption, but they are thought to result from myogenic or neurogenic disturbances of intestinal or gallbladder con­ traction.242 In addition to antibiotic treatment of bacterial overgrowth, low doses of octreotide (50 µg subcutaneously every evening for three weeks) have been shown to induce intestinal migrating motor complexes, reduce bacterial over­ growth, and relieve abdominal symptoms.239

Lupus Erythematosus and Other Connective Tissue Diseases

Excessive fecal fat excretion associated with abnormalities of d-xylose breath testing may be found in some patients with lupus erythematosus; these findings may be accom­ panied by flattened and deformed villi with an inflamma­ tory infiltrate seen on duodenal biopsy.243 Malabsorption that resolved after treatment with prednisolone also has been described in association with the hypereosinophilic syndrome in lupus erythematosus.244 Malabsorption is an uncommon feature of mixed connective tissue disease and polymyositis.245,246

CONGENITAL DEFECTS THAT CAUSE MALABSORPTION

Table 101-14 lists congenital intestinal diseases that result in malabsorption of specific substrates or in a generalized malabsorption syndrome.*

Amino Acid Transport Defects

Amino acids are absorbed by the enterocyte as oligopep­ tides, dipeptides, and free amino acids. In several inborn diseases, transport defects for different groups of amino acids have been identified in the intestine and kidney (see Table 101-14). In iminoglycinuria, Hartnup’s disorder, and cystinuria, the intestinal transport defect seems to be of no or only Text continued on p. 1763 *See references 4, 10, 14, 24, 29, 37, 39, 42, 124, 247-271.

Malabsorption of Amino Acids Hartnup’s disorder

SLC7A7

?

?

?

?

OCRL1

Lysinuric protein intolerance

Isolated lysinuria*

Iminoglycinuria

Blue diaper syndrome*

Methionine malabsorption syndrome* (Oasthouse syndrome)

Lowe oculocerebral syndrome

Type A: SLC3A1 Type B: SLC7A9

SLC6A19

DISORDER

Cystinuria (types A, B, AB)

CAUSATIVE GENE

XR

AR

AR

AR

?

AR

AR (type A) and incomplete AR (type B)

AR

SUGGESTED MODE OF INHERITANCE

Lysine, arginine

Methionine

Tryptophan

l-Proline

Lysine

Dibasic amino acids (lysine, ornithine, arginine)

Neutral amino acids (tryptophan, leucine, methionine, phenylalanine, tyrosine, valine, ?histidine, ?lysine) Cystine and/or dibasic amino acids (lysine, ornithine, arginine)

MALABSORBED SUBSTRATES

Impaired intestinal lysine and arginine absorption

Intestinal methionine absorption defect

Decreased intestinal absorption of specific free amino acids owing to a defective amino acid transporter at the brush border membrane. Type A: no transport of cystine, lysine, or arginine Type B: reduced or normal cystine transport and reduced or no lysine and arginine transport Defect of the basolateral transporter (y+LAT-1) for dibasic amino acids (also malabsorption of di- and oligopeptides) Decreased intestinal absorption of lysine Impaired intestinal absorption of l-proline in a subgroup of subjects Intestinal tryptophan absorption defect

Decreased intestinal absorption of free neutral amino acids

SUGGESTED MECHANISM OF MALABSORPTION

Table 101-14  Congenital Disorders of the Gastrointestinal Mucosa That Result in Malabsorption271

Blue discoloration of diapers, failure to thrive, hypercalcemia, nephrocalcinosis Mental retardation, convulsions, diarrhea, white hair, hyperpnea; urine has characteristic sweet smell of dried celery Aminoaciduria, mental retardation, cataracts, rickets, choreoathetosis, renal disease

Sparse hair, hyperammonemia, nausea, vomiting, diarrhea, protein malnutrition, failure to thrive, aversion to protein-rich food Mental retardation, malnutrition, failure to thrive Aminoaciduria; benign disorder

Aminoaciduria, cystine stones in the urinary tract

Most patients are asymptomatic; some patients have photosensitive skin rash, intermittent ataxia, psychotic behavior, mental retardation, diarrhea

CLINICAL FEATURES

253

252

251

257

249

249

Continued

248, 250

247

REFERENCE(S)

Chapter 101  Maldigestion and Malabsorption 1759

APOB

SAR1B

LIPA

Familial hypobetalipoproteinemia

Chylomicron retention disease Anderson’s disease

Wolman’s disease, cholesteryl ester storage disease

MTP

AR

AR

Incomplete AD

AR

AR

SLC5A1

Glucose-galactose malabsorption

Malabsorption of Fat Abetalipoproteinemia

AR

?

Trehalase deficiency

AR

Sucroseisomaltose gene

AR

SUGGESTED MODE OF INHERITANCE

Sucrase-isomaltase deficiency

Malabsorption of Carbohydrates Congenital lactase LCT deficiency

DISORDER

CAUSATIVE GENE

Fat

Fat

Fat, fat-soluble vitamins

Fat, fat-soluble vitamins

Glucose, galactose

Trehalose

Sucrose, starch

Lactose

MALABSORBED SUBSTRATES

Deficient activity of hLAL, cholesterol ester hydrolase, causing accumulation of cholesteryl esters and triglycerides in various body tissues; infiltration of intestinal mucosa with foamy cells, intestinal damage

Defective chylomicron formation and accumulation in the enterocyte

Defective lipoprotein assembly owing to a lack of MTP, resulting in triglyceride accumulation in the enterocyte and no chylomicron formation Triglyceride accumulation in the enterocyte in homozygotes owing to formation of a truncated apolipoprotein B

Sucrase activity is absent; isomaltase activity is absent or reduced; reduced maltase activity Lack of intestinal trehalase activity Defect of the brush border sodium-glucose cotransporter (SGLT1)

Permanent very low lactase activity

SUGGESTED MECHANISM OF MALABSORPTION

Table 101-14  Congenital Disorders of the Gastrointestinal Mucosa That Result in Malabsorption—cont’d

Steatorrhea, hepatosplenomegaly, abdominal distention; failure to thrive, adrenal calcifications

Steatorrhea, diarrhea, neurologic symptoms, retinitis pigmentosa, failure to thrive, absence of chylomicrons and VLDL in the blood, acanthocytosis Homozygotes: clinical manifestations as for abetalipoproteinemia Heterozygotes: fat absorption is probably normal; hypolipidemia, neurologic manifestations Steatorrhea, failure to thrive, absence of chylomicrons and reduced LDL levels in the blood; neurologic symptoms in some patients

Diarrhea and/or vomiting after ingesting mushrooms Neonatal onset of osmotic diarrhea, dehydration, intermittent or constant glycosuria

Diarrhea, bloating, and dehydration in the first days of life Osmotic diarrhea after starch or sucrose ingestion; failure to thrive

CLINICAL FEATURES

259, 260

10, 258

10

10

256

255

255

255

REFERENCE(S)

1760 Section X  Small and Large Intestine

Occipital horn syndrome (X-linked cutis laxa)

Isolated magnesium malabsorption (hypomagnesemia with secondary hypocalcemia [HOMG]) Menkes disease

Hereditary folate malabsorption Malabsorption of Minerals Acrodermatitis enteropathica

Imerslund-Gräsbeck syndrome (ileal B12 malabsorption, megaloblastic anemia type I) Transcobalamin II deficiency

Malabsorption of Vitamins Congenital IF deficiency (congenital pernicious anemia)

DISORDER

XR

XR

ATP7A

ATP7A

AR

AR

SLC39A4

TRPM6

AR

AR

TCN2

?

AR

AR

SUGGESTED MODE OF INHERITANCE

CUBN or AMN

GIF

CAUSATIVE GENE

Copper

Copper

Magnesium

Zinc

Folate

Cobalamin (vitamin B12)

Cobalamin (vitamin B12)

Cobalamin (vitamin B12)

MALABSORBED SUBSTRATES

General copper transport disorder; intestinal copper malabsorption with copper accumulation in the intestinal mucosa owing to a defective transmembrane coppertransporting ATPase (MNK) Milder form of same defect as in Menkes disease; low levels of functional MNK

Defective zinc absorption in the small intestine owing to a defect in the zinc transport protein (hZIP4) Selective defect in intestinal magnesium absorption

Defective synthesis of IF or synthesis of an abnormal IF with either reduced affinity for cobalamin or for the ileal IF receptor, or increased susceptibility to proteolysis Impaired ileal absorption of IF-cobalamin complex owing to defects in the cubilin-AMN complex (IF-cobalamin receptor) Defective transport of cobalamin out of enterocytes into the portal blood due to absence or malfunction of transcobalamin II Defective folate transport across the intestinal mucosa

SUGGESTED MECHANISM OF MALABSORPTION

Inguinal hernias, bladder and ureteral diverticula, skin and joint laxity, chronic diarrhea, bone changes

Cerebral degeneration, diarrhea, abnormal hair, hypopigmentation, arterial rupture, thrombosis, hypothermia, bone changes

Diarrhea, scaling erythematous dermatitis, alopecia, neuropsychiatric symptoms; onset after weaning Tetany, convulsion, diarrhea, hypomagnesemia with secondary hypocalcemia

Megaloblastic anemia, diarrhea, neurologic symptoms

Vomiting, diarrhea, failure to thrive, anemia, immunodeficiency, neurologic symptoms

Megaloblastic anemia, neurologic symptoms, proteinuria

Megaloblastic anemia, neurologic symptoms, delayed development

CLINICAL FEATURES

262

262

37

42

267

Continued

30, 124

24, 29, 124

30, 124, 261

REFERENCE(S)

Chapter 101  Maldigestion and Malabsorption 1761

AR

Generalized malabsorption Generalized malabsorption

AR

NEUROG3

EpCAM

Generalized malabsorption

Carbohydrates, fat, cobalamin, electrolytes, water Generalized malabsorption

XR

AR

AR

MYO5B

Bile acids, fat

USH1C, ABCC8, and KCNJ11 FOXP3

AR

SLC10A2

Protein, fat

Calcium

Calcium

Iron

MALABSORBED SUBSTRATES

Intestinal epithelial cell dysplasia and villus atrophy

Lack of enteroendocrine cells

Villus atrophy

Villus atrophy with microvillus inclusions in enterocytes, absent or shortened brush border microvilli Enteropathy with villus atrophy and inflammation

Defective activation of pancreatic proenzymes owing to lack of intestinal enterokinase Defect of the ileal ASBT

Defective 25(OH)D 1αhydroxylase, resulting in 1α,25(OH)2D deficiency and reduced intestinal calcium absorption Malfunction of the vitamin D receptor owing to defective hormone binding, defective receptor translocation to nucleus, or defective receptor binding to DNA, resulting in malabsorption of calcium

Intestinal iron transport disorder

SUGGESTED MECHANISM OF MALABSORPTION

Severe diarrhea , failure to thrive, type I diabetes mellitus Severe diarrhea, failure to thrive

Hyperinsulinism, profound congenital sensorineural deafness, enteropathy, renal tubular dysfunction Polyendocrinopathies, severe diarrhea, hemolytic anemia

Steatorrhea, diarrhea, failure to thrive Severe watery diarrhea and steatorrhea requiring total parenteral nutrition

Diarrhea, failure to thrive, hypoproteinemia, edema, anemia

Bone pain, deformities and fractures, muscle weakness, alopecia

Iron deficient anemia that is unresponsive to oral iron supplementation Bone pain, deformities and fractures, muscle weakness

CLINICAL FEATURES

269

268

270

266

265

4

14, 264

263

263

39

REFERENCE(S)

*Reported in only a few case reports. AD, autosomal dominant; AMN, amnionless; AR, autosomal recessive; ASBT, sodium bile acid cotransporter; CUBN, cubulin; hLAL, human lysosomal acid lipase; IF, intrinsic factor; LDL, low-density lipoprotein; MTP, microsomal triglyceride transfer protein; 1α,25(OH)2D, 1α,25-dihydroxyvitamin D; 25(OH)D, 25-hydroxyvitamin D; VLDL, very-low-density lipoprotein; XR, X-linked recessive.

Congenital tufting enteropathy

Immune dysregulation polyendocrinopathy and enteropathy, X-linked (IPEX) Enteric anendocrinosis*

Hyperinsulinism, with enteropathy and deafness

Congenital bile acid malabsorption Microvillus inclusion disease

AR

AR

VDR

PRSS7

AR

CYP27B1

Hereditary selective deficiency of 1α,25(OH)2D (pseudo– vitamin D deficiency rickets) Hereditary generalized resistance to 1α,25(OH)2D (vitamin D–resistant rickets)

Other Defects Enterokinase deficiency

AR

TMPRSS6

SUGGESTED MODE OF INHERITANCE

Iron-refractory iron deficient anemia

DISORDER

CAUSATIVE GENE

Table 101-14  Congenital Disorders of the Gastrointestinal Mucosa That Result in Malabsorption—cont’d

1762 Section X  Small and Large Intestine

Chapter 101  Maldigestion and Malabsorption minor clinical significance, because the amino acids affected by the transporter defects still can be absorbed as oligopep­ tides and dipeptides, and protein malnutrition can be avoided.247,248,257 Manifestations in these diseases, therefore, are mainly due to amino acid transport defects in the kidney. In Hartnup’s disorder, oral administration of nicotinamide and a high-protein diet have been shown to relieve symp­ toms to some extent.247 In lysinuric protein intolerance, however, the transport defect is located on the basolateral membrane of the entero­ cyte, leading to malabsorption of cationic amino acids in both their monopeptide and dipeptide forms. Patients with lysinuric protein intolerance therefore cannot tolerate highprotein foods, with consequent development of protein mal­ nutrition. Malabsorption of lysine with resultant deficiency of this essential amino acid is thought to be an important factor in the development of several disease manifestations in these patients.249 Treatment consists of protein restriction and supplementation with oral citrulline.

Disaccharidase Deficiency and Transport Defects for Monosaccharides

In sucrase-isomaltase deficiency, affected infants usually become symptomatic after weaning, when starch and sucrose are introduced to the diet. Symptoms and signs include osmotic diarrhea, failure to thrive, excess flatus, and occasional vomiting. Diagnosis can be established by an oral sucrose absorption test. Treatment includes avoidance of dietary starch and sucrose.255 In patients with this disease, symptoms tend to resolve spontaneously with age. Patients with glucose-galactose malabsorption suffer from severe diarrhea, leading to dehydration in the first days of life. The diarrhea stops only if glucose and galactose are eliminated from the diet. Older children and adults tolerate the offending carbohydrates better, but the transport defect is lifelong. Diagnosis can be established with an oral glucose tolerance test or by in vitro glucose absorption tests per­ formed on intestinal biopsy specimens. Therapy consists of a fructose-based diet free of glucose and galactose. After the age of three months, addition of foods containing low quan­ tities of glucose or galactose such as vegetables, fruits, and cheese is considered safe.256

Congenital Disorders of Lipid Absorption

Abetalipoproteinemia is a disorder of autosomal recessive inheritance characterized by triglyceride accumulation in the enterocyte. This disease seems to be caused by muta­ tions in the gene for microsomal triglyceride transfer protein (MTP), resulting in defective assembly of triglyceride-rich lipoproteins.10 Familial hypobetalipoproteinemia, a disorder of auto­ somal dominant inheritance, has clinical manifestations similar to those of abetalipoproteinemia when in the homo­ zygous state. This disease seems to be caused by mutations of the apolipoprotein B gene in most cases.10 Chylomicron retention disease and Anderson’s disease are caused by defective release of chylomicrons by entero­ cytes. The distinction between the two conditions derives from differences in the partitioning of lipid between mem­ brane and cytoplasmic compartments; however, both disease are due to a defect in the same gene (SAR1B). General treatment measures in abetalipoproteinemia, hypobetalipoproteinemia, chylomicron retention disease, and Anderson’s disease include the replacement of triglycerides containing long-chain fatty acids with medium-chain triglycerides and dietary supplementation with tocopherol.10

Wolman’s disease and the milder, late-onset cholesteryl ester storage disease are seemingly caused by mutations in different parts of the LIPA gene, resulting in infiltration of intestinal mucosa with foam cells and intestinal damage.

Congenital Disorders of Cobalamin Absorption

Several congenital diseases can result in vitamin B12 malab­ sorption. Absence of intrinsic factor (IF) synthesis is the most common cause of congenital cobalamin deficiency; abnormal results on Schilling tests normalize with the coadministration of IF.24,124 In some patients, an abnor­ mal (nonfunctional) IF is secreted that has a decreased affinity for cobalamin, a decreased affinity for the ileal IFcobalamin receptor (cubilin-amnionless [AMN] complex), or an increased susceptibility to proteolysis.24,124 Imerslund-Gräsbeck syndrome is a congenital disease characterized by malabsorption of the cobalamin-IF complex despite normal ileal morphology. This syndrome can be caused by mutations in two genes that code for the cubilin and AMN proteins, which are colocalized in the ileal mucosa and form the IF-cobalamin receptor.29 In transcobalamin II deficiency, serum levels of cobala­ min commonly are normal, although in most patients intes­ tinal cobalamin absorption is abnormal.124 Diagnosis is established by demonstrating the absence of transcobalamin II in the plasma.124 All congenital disorders of cobalamin absorption are treated by the parenteral administration of cobalamin, although high-dose oral cobalamin also might be effective.

Intestinal Enterokinase Deficiency

Enterokinase is an enzyme that is secreted by the intestinal mucosa and that initiates the activation of pancreatic proen­ zymes. Several patients have been reported to have an inborn deficiency of this enzyme, with resultant diarrhea, failure to thrive, and hypoproteinemia mainly from protein malab­ sorption. These patients respond well to pancreatic enzyme replacement, and some patients show a tendency to improve with age. Secondary enterokinase deficiency also has been reported in patients with villus atrophy, although patients with celiac disease seem not to be affected.14

PRIMARY IMMUNODEFICIENCY DISEASES

Malabsorption commonly occurs in entities that are charac­ terized by deficiencies in humoral or cellular immunity272 (see Chapter 2). The immunodeficiency syndromes most commonly associated with malabsorption are selective IgA deficiency, common variable immunodeficiency (CVID), and severe combined immunodeficiency. The etiology of the malabsorption varies for the different syndromes.

Selective Immunoglobulin A Deficiency

Selective IgA deficiency (OMIM #609529) is the most common primary immunodeficiency disorder and is char­ acterized by a selective near-absence of secretory and serum IgA, leading to susceptibility to respiratory, urogenital, and gastrointestinal infections. Autoimmune and allergic dis­ eases also commonly develop in patients with this disorder. A 10- to 16-fold increased incidence of gluten-sensitive enteropathy has been reported in patients with IgA defi­ ciency.273 However, at least a subgroup of patients have sprue-like small intestinal lesions, leading to severe diar­ rhea and malabsorption, but are unresponsive to a glutenfree diet.274 Improvement with immunosuppressive therapy has been described in one case report.275 Pernicious anemia, giardiasis, and secondary disaccharidase deficiencies also are seen with increased frequency in patients with selective IgA deficiency.274,276

1763

1764

Section X  Small and Large Intestine Common Variable Immunodeficiency

Common variable immunodeficiency (CVID; OMIM #240500), or CVID-acquired hypogammaglobulinemia, is an immunodeficiency disorder characterized by decreased serum immunoglobulin (Ig)G levels and variably decreased serum levels of other immunoglobulin subclasses with T-cell defects. Familial and sporadic forms are caused by mutations in the TNFRSF13B gene.277 Onset of the disease usually is in adulthood, and it manifests with recurrent respiratory and gastrointestinal infections. Affected patients also are at increased risk for autoimmune and neoplastic diseases. Malabsorption and diarrhea occur in 9% to 40% of patients with CVID274; malabsorption involves dietary fat, carbohydrates, vitamin B12, and folate.272,278 Small intestinal biopsy specimens show either sprue-like features, including villus shortening with increased numbers of lymphocytes in the epithelium and in the lamina propria, or a pattern similar to that in graft-versus-host disease (see Chapter 34).274,279 Some specific histologic features, including a near-absence of plasma cells, are observed. The disease does not respond to a gluten-free diet, and it appears that the sprue-like syndrome in CVID is a distinct entity,278 sometimes referred to as “hypogammaglobulin­ emic sprue.”280 In some patients with CVID, foamy macro­ phages are present, as in Whipple’s disease, but in contrast with Whipple’s disease, the macrophages do not contain material that stains with periodic acid–Schiff.279 In addition, nodular lymphoid hyperplasia can be detected in the gas­ trointestinal tract in a high proportion of CVID patients; the presence of nodular lymphoid hyperplasia does not corre­ late with the presence of malabsorption. The incidence of small bowel lymphoma is increased in CVID and both dis­ orders have to be considered as potential causes of malab­ sorption in these patients. Giardia organisms often are isolated from patients with CVID, and small bowel bacterial overgrowth has been documented in a number of these patients. Unfortunately, only some patients with malabsorption associated with CVID respond to antimicrobial treatment.279 Some patients with sprue-like intestinal changes have ben­ efited from glucocorticoids276 or immunoglobulins. Patients with CVID have a higher prevalence of atrophic gastritis with cobalamin malabsorption, although antibodies against parietal cells and intrinsic factor are absent.274,278

X-Linked Infantile Agammaglobulinemia (Bruton’s Agammaglobulinemia)

X-linked infantile agammaglobulinemia (Bruton’s agamma­ globulinemia; OMIM #300300) is an immunodeficiency disease characterized by lack of mature B lymphocytes and failure of Ig heavy chain rearrangement; it is caused by mutations in the gene for Bruton tyrosine kinase.281 This disease usually manifests after the first six months of life and is characterized by recurrent severe bacterial infections. Severe gastrointestinal problems, such as malabsorption and chronic diarrhea, seem to be less common than in CVID279; the prevalence of chronic gastroenteritis was 10% in one large series.282 In affected patients, the possi­ bility of giardiasis and bacterial overgrowth needs to be evaluated.279,282

Immune Dysregulation-PolyendocrinopathyEnteropathy–X-Linked Syndrome

Immune dysregulation-polyendocrinopathy-enteropathy– X-linked syndrome (IPEX) (OMIM #304790) is a disorder of early childhood characterized by protracted diarrhea,

dermatitis, insulin-dependent diabetes mellitus, thyroiditis, thrombocytopenia, and hemolytic anemia. It is a disorder of X-linked recessive inheritance caused by mutations in the FOXP3 gene.270 Diarrhea and malabsorption are secondary to severe villus atrophy with inflammation. Antienterocyte antibodies commonly are present. The enteropathy usually does not respond to a gluten-free diet, but immunosuppres­ sive therapy has been shown to be of some benefit. IPEX usually is fatal in childhood. Successful bone marrow trans­ plantation with amelioration of enteropathy has been reported in some cases.283

Other Congenital Immunodeficiency Syndromes

In severe combined immunodeficiency (OMIM #300400), diarrhea and malabsorption are common. Symptoms are associated with stunting of intestinal villi or their complete absence. The pathophysiology of malabsorption is unknown, and the syndrome usually fails to respond to antimicrobial treatment.274,276 Malabsorption also has been reported in DiGeorge’s syndrome (thymic hypoplasia, OMIM #188400) and chronic granulomatous disease of childhood (OMIM #306400), but little is known about its cause in these disorders.274

NEUROFIBROMATOSIS TYPE 1 (VON RECKLINGHAUSEN’S DISEASE)

Malabsorption can be an intestinal complication of neuro­ fibromatosis type 1 (OMIM #162200). Mechanisms of malabsorption include periampullary duodenal tumors, which are mainly somatostatin-containing neuroendocrine tumors, and pancreatic carcinomas with resultant pan­ creatic duct obstruction; tumors can cause exocrine pancre­ atic insufficiency and biliary obstruction.284 Duodenal somatostatinomas in von Recklinghausen’s disease usually do not increase plasma somatostatin levels, although one case of somatostatinoma syndrome has been reported.285 Infiltrating mesenteric plexiform neurofibromas and vascu­ lar damage caused by proliferation of nerves can cause lymphatic or vascular obstruction (or both), resulting in abdominal pain, protein-losing enteropathy, diarrhea, steatorrhea, and bowel ischemia.286,287 In patients with von Recklinghausen’s disease, an increased incidence of neuro­ endocrine tumors in other locations has been observed; gastrinomas with Zollinger-Ellison syndrome also have been reported in some of these patients.288

NONGRANULOMATOUS CHRONIC IDIOPATHIC ENTEROCOLITIS AND AUTOIMMUNE ENTEROPATHY

Nongranulomatous chronic idiopathic enterocolitis is an entity that is distinct from refractory celiac disease and inflammatory bowel disease.289 The etiology of this disease is unknown, although chronic infection and an autoimmune cause have been suggested. Severe diarrhea and malabsorp­ tion occur as a result of diffuse villus atrophy, and ulcer­ ations may be present in the small and large intestine. Small intestinal villus atrophy and neutrophilic inflamma­ tion of the mucosa with crypt abscesses may be seen in biopsy specimens from the small intestine and colon (Fig. 101-9); the number of intraepithelial lymphocytes is not increased.289,290 Patients respond dramatically to glucocorticoids, and most require long-term low-dose maintenance therapy.289,290 Improvement with cyclosporine and long-term antibiotic therapy has been reported in one patient each.291 The condition is associated with a high mortality rate.289,290

Chapter 101  Maldigestion and Malabsorption terocyte antibodies and anti–goblet cell antibodies. Absence of antibodies, however, does not exclude the diagnosis.292 Therapy of autoimmune enteropathy is challenging, and some patients have been treated successfully with glucocor­ ticoids and immunosuppressive drugs.

ENDOCRINE AND METABOLIC DISORDERS Adrenal Insufficiency (Addison’s Disease)

Some patients with adrenal insufficiency, independent of its etiology, have fat malabsorption, and fecal fat excretion of up to 30 g/day has been documented.293 Fat malabsorp­ tion also is observed in rats after adrenalectomy.294 The pathophysiologic mechanism of malabsorption in this disease is unknown, but fat absorption normalizes after glu­ cocorticoid replacement. Isolated autoimmune Addison’s disease has been asso­ ciated with pernicious anemia295 and celiac disease.296 An increased incidence of celiac disease and pernicious anemia also is found in autoimmune polyglandular syn­ drome (APS) type 2 (Schmidt’s syndrome), which is char­ acterized by the association of autoimmune Addison’s disease and other autoimmune endocrine disorders except hypoparathyroidism.297

A

B Figure 101-9.  Duodenal biopsy specimen from a patient with nongranu­ lomatous chronic idiopathic enterocolitis. A, Histopathologic features include villus atrophy, diffuse infiltration of the lamina propria with inflammatory cells, and crypt abscesses (arrow). B, High-power view demonstrates crypt infiltration by neutrophils (arrow). (Hematoxylin and eosin stain.) (Courtesy of Cord Langner, MD.)

Nongranulomatous chronic idiopathic enterocolitis shares several clinical and histologic features with adult autoimmune enteropathy.290,292 In many patients with adult autoimmune enteropathy, antienterocyte antibodies, anti– goblet cell antibodies, and other autoimmune disorders are present. Symptoms are chronic severe high-output diarrhea and malabsorption.292 Diagnosis relies on a combination of clinical and histologic findings. Proposed diagnostic criteria require the presence of chronic diarrhea and malabsorption; exclusion of other small intestinal diseases, such as celiac disease; histologic changes on intestinal biopsies such as partial or complete villous blunting, deep crypt lympho­ cytosis, increased crypt apoptotic bodies, and minimal intraepithelial lymphocytosis; and the presence of antien­

Enteroendocrine Deficiency (Autoimmune Polyglandular Syndrome Type 1 and Enteric Anendocrinosis)

Autoimmune polyglandular syndrome type 1 (APS 1) (OMIM #240300) is characterized by multiple endocrine organ failure (especially hypoparathyroidism and adrenal insufficiency) due to autoimmune destruction, with ecto­ dermal dystrophy and susceptibility to chronic Candida infections.297 APS type 1 is inherited as an autosomal reces­ sive disorder and is caused by mutations in the AIRE gene.298 Severe malabsorption, which tends to be recurrent, devel­ ops in approximately 20% of patients with APS type 1. In two patients, malabsorption was caused by a transient and selective destruction of small intestinal enteroendocrine cells, leading to a temporary deficiency of enteroendocrine hormones (especially CCK),299,300 These patients have auto­ antibodies to tryptophan hydroxylase, which are directed against enteroendocrine cells (including CCK-producing cells).301 The long-known association between hypoparathy­ roidism and steatorrhea may be caused by the same mecha­ nism, because in most reports of this association, patients fulfill the diagnostic criteria for APS type 1.302,303 Selective absence of small intestinal enteroendocrine cells can be diagnosed by special immunohistochemical stains for these cells, such as immunohistochemical stains for chromogranin A or CCK (Fig. 101-10) or by measure­ ments of postprandial serum levels of the affected hor­ mones. Patients with APS type 1 also have an increased incidence of vitamin B12 malabsorption as a result of auto­ immune gastritis.297 Lack of enteroendocrine cells also results in congenital malabsorption in a newly described disease resulting from a mutation in the NEUROG3 gene (diarrhea 4; enteric anendocrinosis OMIM #610370) (see Table 101-14).268

Hyperthyroidism and Autoimmune Thyroid Disease

Some reports suggest that up to 25% of hyperthyroid patients have at least some degree of fat malabsorption, but data from large series of patients are lacking. Fecal fat values in hyperthyroid patients can reach 35 g/day.304 The mecha­ nism of steatorrhea in hyperthyroidism has not been estab­ lished. Motility studies in hyperthyroid patients (including

1765

1766

Section X  Small and Large Intestine

A

B

Figure 101-10.  Chromogranin A immunohistochemical staining of enteroendocrine cells in duodenal biopsy specimens obtained from a normal subject (A) and from a patient with malabsorption associated with autoimmune polyglandular syndrome type 1 (B). In B, enteroendocrine cells are absent. See text for details.

patients with and without diarrhea) have demonstrated accelerated small intestinal and whole-gut transit times305; however, fecal fat values were not reported in these patients. It can be hypothesized that more-pronounced disturbances of intestinal transit can lead to decreased mixing of food and digestive secretions and reduced intestinal absorption of nutrients. Some of the steatorrhea in hyperthyroid patients might be due to hyperphagia with increased dietary intake of fat.306 An increased number of lymphocytes and plasma cells and some degree of edema in small intestinal biopsy speci­ mens have been found in patients with steatorrhea and hyperthyroidism; villus architecture is normal.304 Absorp­ tion of glucose and d-xylose is normal in hyperthyroid patients with and without malabsorption.306 Fat malabsorp­ tion tends to normalize when patients attain a euthyroid state.304,306 In patients with autoimmune thyroid diseases, an increased prevalence of celiac disease296 and primary biliary cirrhosis,295 both of which can result in fat malabsorption, has been recognized. The prevalence of celiac disease in patients with autoimmune thyroid disease is approximately 2% to 4%.296 Cobalamin malabsorption resulting from auto­ immune gastritis is found in a considerable number of patients with autoimmune thyroid disease.26,295

bacterial overgrowth,310 and pancreatic insufficiency,313 can be associated with diabetes mellitus. In patients with dia­ betes mellitus type 1, a high prevalence (3% to 8%) of celiac disease has been reported from screening studies; however, most patients identified were asymptomatic.314 Markedly reduced pancreatic exocrine function, as determined by fecal elastase measurement, has been reported in 30% of patients with type 1 diabetes and 17% with type 2 diabetes, compared with 5% of control subjects. In 40% of diabetic patients with reduced fecal elastase levels, fat malabsorp­ tion with fecal fat output of more than 10 g/day was detected.315 Gastrointestinal symptoms and steatorrhea in these patients, however, did not correlate with fecal elastase levels.313,315 In addition, the unresolved specificity of elas­ tase raises the possibility that not all of these patients truly had pancreatic insufficiency.316 The presence of cobalamin malabsorption caused by autoimmune atrophic gastritis is increased three- to five-fold in patients with diabetes mel­ litus type 1 compared with the nondiabetic population.317 Ingested carbohydrates are malabsorbed in patients receiving acarbose as part of their diabetes treatment, which in turn can lead to symptoms of diarrhea and malabsorp­ tion. Foods recommended to diabetics because they contain poorly absorbable carbohydrates such as fructose or sorbitol also can result in bloating and diarrhea.

Diabetes Mellitus

Metabolic Bone Disease

Chronic diarrhea is common in patients with diabetes mel­ litus, especially in those with long-standing diabetes mel­ litus type 1.307 Mild steatorrhea often is present in patients with diabetic diarrhea and in diabetic patients who do not complain of diarrhea.308 Although the pathophysiologic mechanism of malabsorp­ tion and diarrhea in patients with diabetes mellitus is unknown, poor glycemic control is an important cofactor.309 Most of these patients have signs of autonomic neuropathy, such as orthostatic hypotension, impotence, bladder dys­ function, incontinence, inappropriate heart rate variability, and abnormal sweating.310 Therefore, in some patients, the cause of diarrhea and malabsorption has been attributed to rapid gastric emptying and rapid intestinal transit, causing impaired mixing of nutrients with digestive secretions and decreased contact time between nutrients and the intestinal mucosa. The clinician has to be aware, however, that certain treat­ able diseases, such as celiac disease,311,312 small intestinal

Special consideration has to be given to osteoporosis and osteomalacia in malabsorptive diseases. Patients with these metabolic bone diseases usually do not present with sugges­ tive symptoms or abnormalities either on physical examina­ tion or on routine laboratory examinations. Reduced bone mineral density is a common finding in patients with gastric resection,318 celiac disease,319 and lactose malabsorption.164 Osteoporosis has been suggested to result from calcium mal­ absorption or reduced calcium intake, which leads to sec­ ondary hyperparathyroidism, which in turn increases bone turnover and cortical bone loss. Vitamin D malabsorption probably is of lesser importance. Although up to one half of patients on a gluten-free diet have osteoporosis,320 some studies have shown significant improvement in bone mineral density one year after starting a gluten-free diet.321 In inflammatory bowel diseases such as Crohn’s disease, which may be accompanied by malabsorption, other factors such as glucocorticoid use or testosterone deficiency322 may contribute to decreased bone mass.

Chapter 101  Maldigestion and Malabsorption In addition to treating the underlying cause of malabsorp­ tion, calcium supplementation is needed to ensure a daily intake of 1500 mg of calcium. Vitamin D deficiency must be corrected. If osteoporosis is present, bisphosphonate treat­ ment is suggested.319 Nutritional management is discussed in more detail in Chapters 4 and 5.

GENERAL APPROACH TO MANAGEMENT Treatment of malabsorptive diseases must be directed against the underlying condition, if possible. In addition, nutritional deficits must be corrected. The reader is referred to the relevant chapters of this book for discussion about the treatment of specific diseases and their nutritional man­ agement. In patients with abdominal bloating and gasrelated complaints as a result of sugar malabsorption, a diet with reduced content of poorly absorbable carbohydrates such as fructose, sobitol, or fermentable dietary fibers is a long-term effective therapy.323 In pancreatic insufficiency, in disorders of intestinal fat absorption, and in short bowel syndrome, medium-chain triglycerides can be used as a source of dietary calories. In patients with short bowel syndrome and some remaining colon, colonic salvage capacity can be used to regain calo­ ries from carbohydrates324; these patients, therefore, should consume a diet rich in carbohydrates and medium-chain triglycerides. In bile acid malabsorption after extensive ileal resections, intestinal fat absorption can be improved markedly by oral administration of natural conjugated bile acids178 or of syn­ thetic cholylsarcosine.104,179,180 Replacement of conjugated bile acids also reduces urinary oxalate excretion and there­ fore should protect against development of kidney stones.179 Patients with cystic fibrosis or short bowel syndrome who are unable to absorb vitamin D from their diet may benefit from treatment with an ultraviolet lamp, which emits ultra­ violet radiation similar to sunlight.325 In patients with malabsorption and an intact colon, fluid depletion must be avoided to prevent kidney stones associ­ ated with hyperoxaluria.326 In patients with malabsorption syndrome, special care should be given to the replacement of vitamins, iron, calcium, and trace elements to avoid defi­ ciency syndromes (see Chapters 4 and 5).

In patients with diarrhea, symptomatic treatment with opiates or loperamide can increase the time available for absorption of nutrients.

KEY REFERENCES

Fine KD, Fordtran JS. The effect of diarrhea on fecal fat excretion. Gas­ troenterology 1992; 102:1936-9. (Ref 102.) Fine KD, Schiller LR. AGA technical review on the evaluation and management of chronic diarrhea. Gastroenterology 1999; 116:146486. (Ref 98.) Gibson PR, Newnham E, Barrett JS, et al. Review article: Fructose mal­ absorption and the bigger picture. Aliment Pharmacol Ther 2007; 25:349-63. (Ref 168.) Hofmann AF, Poley JR. Role of bile acid malabsorption in pathogenesis of diarrhea and steatorrhea in patients with ileal resection. Gastroen­ terology 1972; 62:918-34. (Ref 3.) Holt PR. Intestinal malabsorption in the elderly. Dig Dis 2007; 25:14450. (Ref 231.) Horton KM, Corl FM, Fishman EK. CT of nonneoplastic diseases of the small bowel: Spectrum of disease. J Comput Assist Tomogr 1999; 23:417-28. (Ref 94.) Longstreth GF, Newcomer AD. Drug-induced malabsorption. Mayo Clin Proc 1975; 50:284-93. (Ref 194.) Online Mendelian Inheritance in Man. McKusick-Nathans Institute of Genetic Medicine, Baltimore: Johns Hopkins University; and Bethesda (Md): National Center for Biotechnology Information, National Library of Medicine [2008 Sep 30]. Available from: http://www.ncbi. nlm.nih.gov/omim/ (Ref 271.) Poitou Bernert C, Ciangura C, Coupaye M, et al. Nutritional deficiency after gastric bypass: Diagnosis, prevention and treatment. Diabetes Metab 2007; 33:13-24. (Ref 222.) Riddell RH. Small intestinal biopsy: Who? How? What are the findings? In: Barkin JS, Rogers AI, editors. Difficult Decisions in Digestive Diseases. Chicago: Year Book Medical Publishers; 1989. p 326-31. (Ref 85.) Romagnuolo J, Schiller D, Bailey RJ. Using breath tests wisely in a gastroenterology practice: An evidence-based review of indications and pitfalls in interpretation. Am J Gastroenterol 2002; 97:1113-26. (Ref 111.) Ryan ER, Heaslip IS. Magnetic resonance enteroclysis compared with conventional enteroclysis and computed tomography enteroclysis: A critically appraised topic. Abdom Imaging 2008; 33:34-7. (Ref 97.) Ryan ME, Olsen WA. A diagnostic approach to malabsorption syn­ dromes: A pathophysiological approach. Clin Gastroenterol 1983; 12:533-50. (Ref 6.) Seetharam B. Gastrointestinal absorption and transport of cobalamin (vitamin B 12). In: Johnson LR, editor. Physiology of the Gastrointes­ tinal Tract. New York: Raven Press; 1994. p 1997-2026. (Ref 24.) Wilson FA, Dietschy JM. Differential diagnostic approach to clinical problems of malabsorption. Gastroenterology 1971; 61:911-31. (Ref 1.) Full references for this chapter can be found on www.expertconsult.com.

1767

CHAPTE R

102 Enteric Microbiota and Small Intestinal Bacterial Overgrowth Seamus O’Mahony and Fergus Shanahan

CHAPTER OUTLINE Composition and Molecular Analysis of the Enteric Microbiota  1769 Interactions between Host and Microbes  1771 Metabolic Activity of the Microbiota  1772 Small Intestinal Bacterial Overgrowth  1772 Etiology and Predisposing Factors  1772

Increasing recognition of the role of the intestinal micro­ biota (flora) in the development and function of the gastro­ intestinal tract has led to a resurgence of scientific and clinical interest in the enteric ecosystem. The contribution of the microbiota to mucosal homeostasis is so profound that it is no longer acceptable to study intestinal patho­ physiology without considering the activities of the indig­ enous bacteria. The lesson of Helicobacter pylori in the genesis of peptic ulcer disease and gastric cancer has been a sobering reminder of the potential pathogenic role of luminal bacteria. Contributions of the intestinal microbiota to health and disease are complex and influenced by a variety of environmental factors, bacterial factors, and host-related factors, including antibiotic exposure, diet, and immunologic and genetic status. As discussed in this chapter, the enteric bacteria are critical to health and host defense, but under conditions such as those favoring bacte­ rial overgrowth, components of the microbiota can become a liability.

COMPOSITION AND MOLECULAR ANALYSIS OF THE ENTERIC MICROBIOTA Most human enteric bacteria cannot be cultured, because of a lack of truly selective growth media. Nonetheless, mole­c­ ular profiling has shown that whereas the microbiota appear distinct in different persons, the composition of each per­ son’s microbiota is relatively stable after infant weaning and throughout adulthood. Evidence from studies of twins sug­ gests that the individuality of human microflora may be genetically controlled,1 but environmental variables includ­ ing diet and sanitation appear to have profound effects on early intestinal colonization with bacteria.2,3 In adulthood, dietary fluctuations appear to induce changes in bacterial enzymes and metabolic activity rather than changes in the relative populations of the microflora.2,4,5

Mechanisms of Malabsorption  1774 Clinical Features  1774 Diagnosis  1775 Treatment  1777

The composition of the microflora varies quantitatively and qualitatively over the longitudinal and the crosssectional axes of the alimentary tract. Beyond the oral cavity, which harbors approximately 200 different bacterial species, the size and diversity of the microflora increase distally along the digestive tract (Fig. 102-1). Gastric acid restricts bacterial numbers within the stomach to fewer than 103 colony-forming units (CFU)/mL. The gradient in bacte­ rial density is greatest across the ileocecal valve, with approximately 108 bacteria per gram of ileal contents and up to 1012 bacteria per gram of colonic contents, comprising more than 1000 different bacterial species.2,4-6 More than 99% of the culturable bacteria in the ileum and the colon are obligate anaerobes, but the composition of the flora at the mucosal surface differs from that within the lumen; ratios of anaerobes to aerobes are lower at mucosal surfaces. Culture-independent methods, such as the various mole­c­ ular approaches described later, suggest that mucosaassociated bacteria differ from those recovered from feces, thus supporting the idea that host-related factors have a role in determining the enteric microflora7 and implying that bacterial aspirates from the lumen may be an incomplete reflection of mucosa-associated bacteria. The microbiota of the proximal small intestine consist predominantly of Gram-positive facultative bacteria— bacteria that can survive under aerobic or anaerobic con­ ditions—although enterobacteria and Bacteroides species also may be present. Peristalsis is the principal factor restricting bacterial numbers in the small intestine. In the distal small intestine, the composition of the microflora resembles that of the colon, with a preponderance of Gramnegative anaerobes. The most prominently represented genera in the distal bowel include Bacteroides, Clostridium, Lactobacillus, Fusobacterium, Bifidobacterium, Eubacterium, Peptococcus, and Escherichia species.2,6 The impact of diet on the composition of the enteric microflora has been studied by several investigators, with varying results. The consensus is that breast-fed babies have

1769

1770

Section X  Small and Large Intestine

Oral cavity 200 species

Stomach Helicobacter pylori

Duodenum and proximal jejunum 2 10 –103 bacteria/mL

Ileum 108 bacteria/mL

Figure 102-1.  Composition and distribution of bacterial flora within the human alimentary tract.

a greater proportion of bifidobacteria in their fecal flora than that observed in infants who are formula fed; highcarbohydrate and high-fiber diets are associated with increased bifidobacteria, whereas a high-fat diet is linked with an increase in Bacteroides species. It is the metabolic activity of the microbiota, however, rather than their composition, that exhibits the greatest response to dietary changes in a given person. Detailed analysis of the enteric microflora has been con­ founded by the limitations of traditional culture-dependent microbiology.8 First, obtaining representative material from different niches within the intestine is problematic; because most of the indigenous microflora are obligate anaerobes, major methodologic difficulties are encountered with sam­ pling, contamination, transport, and storage. Second, the lack of truly selective growth media precludes culture of most components of the microflora. In this respect, it is noteworthy that culture of H. pylori and C. difficile was accomplished only within the past two to three decades. This difficulty has led to a shift in emphasis from conven­ tional bacterial phenotyping toward genotyping and mo­ lecular approaches to study the unculturables.4,9,10 Bacterial nucleic acid extracted from feces or mucosal biopsy samples can provide a profile of the composition of the indigenous microflora. The small ribosomal subunit RNA (16S rRNA in bacteria) contains highly conserved regions of base sequences that reflect an absence of evolu­ tionary change. These conserved sequences are interspersed with hypervariable regions, which contain mutational changes reflecting the evolutionary divergence of different species. Sequencing of 16S rRNA, therefore, represents a method for identification and phylogenetic classification of intestinal bacteria.

Colon 1010 – 1011 bacteria/g 400-500 species including Bacteroides Eubacterium Peptostreptococcus Bifidobacterium Ruminococcus Bacillus Fusobacterium Clostridium Lactobacillus Enterococcus Enterobacter

For rapid profiling of the dominant culturable and non­ culturable organisms within a complex ecosystem such as that in the intestinal tract, 16S rRNA can be amplified by polymerase chain reaction (PCR) with universal primers spanning conserved and variable regions. The mixture of hypervariable RNA fragments can then be separated by a chemical denaturing gradient or a temperature gradient gel electrophoresis (DGGE and TGGE). Complete denaturation of the RNA fragments is prevented by incorporating a GC-rich 5′ end to one of the primers (a GC clamp).11 Varia­ tions in migration distance through the denaturing gradients reflect the diversity of 16S species in the sample (Fig. 1022). Theoretically, the technique is semiquantitative, because the more dominant the organism, the more abundant the specific PCR product. The specific PCR product can be cut from the gel and further amplified, cloned, and sequenced to identify individual bacterial strains without requiring a conventional culture step. Further refinements of the technique can be achieved by using species-specific PCR primers. Other molecular techniques for analysis of specific bacte­ rial species now are possible because of the increasing avail­ ability of genomic sequence data for the major components of the bacterial flora. These techniques include fluorescence in situ hybridization (FISH) flow cytometry (FISH-flow) and bacterial DNA microarrays. In disorders such as inflamma­ tory bowel disease, immunologic reactivity against compo­ nents of the microflora has been used to identify microbes that may be etiologic in the disease. Marker antibodies gen­ erated by hybridoma or phage-display technology have been used as reagents to identify microbial antigens. For example, antineutrophil cytoplasmic antibody (pANCA), which is associated with ulcerative colitis, has been used to identify

Chapter 102  Enteric Microbiota and Small Intestinal Bacterial Overgrowth Patient 1 0

4

8

Patient 2 12

0

4

8

12

Figure 102-2.  Representative examples of molecular fingerprinting profiles. Denaturing gradient gel electrophoresis analysis of the intestinal microbiota in two patients tested over time, at 0, 4, 8, and 12 months. Although the profile for each patient is unique, it also is stable.

colonic bacteria expressing a pANCA-related epitope.12,13 In addition, candidate microbes can be identified by the presence of unique bacterial nucleic acid sequences asso­ ciated with a particular lesion or disease location, by sub­ tractive cloning using genomic representational difference analysis. With this technique, a sequence representing a bacterial transcription factor from an apparent commensal organism, Pseudomonas fluorescens,14-16 was found in lesions of Crohn’s disease but not in adjacent nonlesional mucosa. Serologic expression cloning also has been used to identify bacterial flagellin as a dominant antigen in Crohn’s disease.17 The new science of metagenomics—the sequencing of genes from whole microbial environments at once—prom­ ises to address many of the unresolved questions about the microbiota. The microbiota comprise a repository of genetic information (microbiome) that greatly exceeds that of the host genome. By combining metagenomics with bioinfor­ matics, biochemistry, and traditional bioassays, new insights into the metabolic capacity of the human intestinal micro­ biota can be achieved.4 Major consortia (including the human microbiome project4) around the globe are under way using metagenomics as a tool for bioprospecting the intestinal microbiota in health and disease.

INTERACTIONS BETWEEN HOST AND MICROBES The microbiota exert both positive and negative regulatory effects on the development and function of the intestine. These complex influences first were shown in comparative studies of germ-free and conventionally colonized animals. A sterile intestine is associated with reductions in mucosal cell turnover, digestive enzyme activity, cytokine produc­ tion, lymphoid tissue, lamina propria cellularity, vascular­ ity, muscle wall thickness, and motility, but with an increase in enterochromaffin cell area.18 The molecular events under­ pinning this regulatory signaling from the lumen currently are being explored using modern techniques such as laser capture microdissection and gene array analysis; such

studies promise to reveal new molecular targets to be exploited for the design of novel therapeutics.19,20 Thus, for example, when applied to animals colonized with only a single bacterial strain, Bacteroides thetaiotaomicron, this combined approach has illustrated the impact of bacteria-derived signaling on the expression of host genes controlling mucosal barrier function, nutrient absorption, angiogenesis, and development of the enteric nervous system. Incoming bacterial signals include secreted chemoattrac­ tants, such as the formylated peptide f-Met-Leu-Phe, cellu­ lar constituents such as lipopolysaccharide (LPS) and peptidoglycans, flagellin, and bacterial nucleic acids (i.e., CpG DNA). Detection of bacterial stimuli by the host and discrimination of pathogens from commensals are mediated in part by pattern recognition receptors such as Toll-like receptors (TLRs) that are present on epithelial and immune (dendritic) cells. In health, engagement of TLRs by ligands from the commensal microbiota appears to be required for mucosal homeostasis.21,22 Thus, not only are bacterial signals required for optimal mucosal and immune development, but they also actually are required to maintain and condi­ tion the mucosa for responses to injury.22 The immune system mediates the sense of microbial danger and responses to injury. Although the primary lym­ phoid organs are developed at birth, mucosal immune func­ tions require continual education and fine-tuning of cytokine balances and T-cell responses; this process is achieved by microbial colonization and sporadic mucosal infections. Without the microbiota, mucosal lymphoid tissue is rudi­ mentary, and induction of mucosal immune responses and tolerance is suboptimal.23,24 With a surface area similar to that of a tennis court (approximately 400 m2) and only one cell layer separating the internal milieu from the lumen, the enteric mucosa is well adapted to immunologic sampling of the intraluminal microbial community. Sampling of the microbiota across the epithelial barrier is mediated by M cells, which deliver particulate and microbial antigens to underlying immune cells, and by mucosal dendritic cells, which appear to extend processes into the lumen between the surface entero­ cytes without disrupting tight junctions.25 It appears that intestinal dendritic cells can ingest and retain intact live commensal bacteria and then transit to the mesenteric lymph node, where immune responses to commensals are induced locally.26 Thus, the mesenteric lymph node acts as a gatekeeper, preventing access of commensal bacteria to the internal milieu and protecting the host from harmful sys­ temic immune reactivity. The immunosensory function of dendritic cells is facilitated by their plasticity and versatility of responses,27 depending on whether they are presented with commensals or pathogens; moreover, they appear to exhibit tissue-specific specialization in the intestine.23,24 In addition to specific immune responses to enteric bac­ teria, the surface epithelial cells serve a sensory function to detect microbial danger by producing chemokines that acti­ vate the host immune response and recruit it to any breach in the mucosal barrier caused by pathogenic infection.28 Transduction of bacterial signals into host immune responses after engagement of TLRs may proceed along more than one molecular pathway. The transcription factor nuclear factor-κB (NF-κB) is the pivotal regulator of epithe­ lial responses to invasive pathogens, but nonpathogenic bacteria can attenuate inflammatory responses by delaying the degradation of IκB, which is counter-regulatory to NF-κB.29 Other signal transduction pathways are likely to emerge to account for the anti-inflammatory effects of probiotics and other commensal organisms such as

1771

1772

Section X  Small and Large Intestine Bacteroides thetaiotaomicron. This anaerobic commensal can antagonize the proinflammatory effects of NF-κB within the epithelial cell by enhancing the nuclear export of its transcriptionally active subunit (RelA) in a peroxisome proliferator-activated receptor-γ (PPAR-γ)– dependent manner.30

METABOLIC ACTIVITY OF THE MICROBIOTA The enteric microbiota are tantamount to a hidden meta­ bolic organ (Table 102-1). Although our understanding of indigenous bacterial metabolites is still superficial, coevolu­ tion with this living inner mass of bacteria has several apparent benefits for the host. In addition to the production of regulatory signals for mucosal homeostasis as discussed earlier, the microbiota exhibit important metabolic proper­ ties not possessed by the host. These include biotransforma­ tion of bile acids; degradation of oxalate; breakdown of otherwise indigestible dietary components, such as plant polysaccharides; and production of short-chain fatty acids, a major energy source for colonic epithelium, from ferment­ able carbohydrates. Other activities include synthesis of biotin, folate, and vitamin K.2,20 Clinicians also have exploited enteric bacterial enzymes such as azoreductase to convert prodrugs such as sulfasalazine to active drug metab­ olites (e.g., aminosalicylate). Other examples of bacterial action on drug bioavailability include the metabolism of l-dopa to dopamine and degradation of digoxin. Not all of the metabolic changes induced by the enteric microbiota are beneficial to the host, however, and although bacteria prob­ ably degrade some carcinogens, they also might promote the production of carcinogens from dietary procarcinogens.31 A striking example of the importance of bacterial metabo­ lism is exemplified by the regulatory effect that the enteric microbiota exert on fat storage.32 It has long been known that germ-free animals need a significantly greater caloric intake to sustain a body weight similar to that of normal colonized animals. Thus, the normal host-microbiota rela­ tionship has nutritional benefit, in contrast with the nega­ tive nutritional effect associated with bacterial overgrowth syndromes. Elegant studies with germ-free mice have shown that upon colonization, body weight increases despite a reduced caloric intake. The bacteria in the micro­ biota colonizing the intestine promote storage of dietary calories in fat by increasing absorption of monosaccharides and suppressing epithelial-derived fasting-induced adipo­ cyte factor (FIAF).32 Thus, the composition and activity of the intestinal microbiota should be considered as a dietinfluenced variable that can influence susceptibility to obesity.

Table 102-1  Examples of Metabolic Activities of Intestinal Microbiota Biotransformation of bile acids Breakdown of dietary oxalate Conversion of prodrugs to active metabolites Degradation of polysaccharides of plant origin Production of folate, B vitamins, and vitamin K Production of nutrient short-chain fatty acids Production of regulatory signals for mucosal and immune homeostasis Regulation of fat storage

One of the outcomes of bacterial metabolic activity is gas production. Of the five gases—N2, O2, CO2, H2, CH4—that constitute 99% of flatus, the latter three are produced by the enteric bacteria, and bacteria are the sole source of hydrogen and methane in the intestine. Hydrogen production by bacterial action on carbohydrates, and to lesser extent on protein, normally occurs in the colon. In patients with small intestinal bacterial overgrowth, however, the small intestine also becomes a site of H2 production. Bacterial methanogens occur in the colon and produce methane from H2 and CO2, with significantly detectable excretion in approximately 30% of humans.33-37 The principal gases produced are odor­ less, but bacterial metabolism also is responsible for produc­ ing various trace and odiferous gases in flatus such as hydrogen disulfide.38,39 Qualitative and quantitative vari­ ability in gas production with diet illustrates the fluctua­ tions in bacterial metabolic activity despite the apparent stability of the microbiota in adulthood.

SMALL INTESTINAL BACTERIAL OVERGROWTH Small intestinal bacterial overgrowth (SIBO) is character­ ized by malabsorption and overgrowth of bacteria in the small intestine. The syndrome often is referred to as blind loop syndrome because of recognition of the disorder in patients with predisposing anatomic abnormalities. Other terms that have been used to describe the disorder include stagnant loop syndrome, contaminated small bowel, small intestinal colonization, and small bowel stasis. In 1939, Barker and Hummel40 reported macrocytic anemia in association with intestinal strictures and anastomoses and postulated that the anemia was secondary to bacterial overgrowth, or “putrefaction.” SIBO is not confined to humans and is well recognized in dogs.41 The syndrome is associated with a variety of anatomic disturbances, such as blind loops,42 and motility disorders, such as scleroderma,43 but it can occur in the absence of any specific predisposing factor. It is likely that the condition is underdiagnosed, particularly in the elderly.44 Patients with SIBO do not necessarily present with a florid malabsorption syndrome, and symptoms may be minor and nonspecific. Considerable debate has concerned the relationship between irritable bowel syndrome (IBS) and SIBO (see Chapter 118).45 Bacterial overgrowth has been documented in asymptomatic elderly persons in the com­ munity,46,47 in whom it is debatable whether the phenome­ non is of any significance48; in the absence of malabsorption or related symptoms, such overgrowth probably should not be considered to represent true SIBO. Asymptomatic SIBO may be termed simple colonization and probably results from achlorhydria and abnormal fasting intestinal motility (see later). The diagnosis of SIBO usually is made by noninvasive breath testing,49 even though studies on the accuracy of these tests report very variable results. For this reason, culture of a small intestinal aspirate must be regarded as the diagnostic gold standard. Unfortunately, much of the pub­ lished literature on SIBO is based on breath tests, rather than on culture, and the findings must be interpreted with caution.

ETIOLOGY AND PREDISPOSING FACTORS

The upper small intestine is an environment of relatively low bacterial counts because of the combined effects of gastric acid and peristalsis. Bacterial counts in aspirates

Chapter 102  Enteric Microbiota and Small Intestinal Bacterial Overgrowth Table 102-2  Pathophysiology and Some Conditions Associated with Small Intestinal Bacterial Overgrowth PATHOPHYSIOLOGY

CONDITION

Anatomic abnormalities

Blind loop (Billroth II gastrectomy, end-to-side anastomosis) Small intestinal diverticulosis Small intestinal stricture (Crohn’s disease, radiation enteritis, focal segmental ischemia) Diabetes mellitus Idiopathic intestinal pseudo-obstruction Scleroderma Acid-lowering medication Atrophic gastritis Previous vagotomy Gastrocolic or enterocolic fistula Resection of ileocecal valve

Motility disorders Reduced gastric acid secretion Abnormal connection between colon and proximal bowel Various mechanisms

Celiac disease Cirrhosis Chronic pancreatitis Chronic kidney disease Radiation enteritis Rheumatoid arthritis

from the normal upper small intestine generally are less than 1000/mL. Pathophysiology and predisposing condi­ tions are listed in Table 102-2.

Anatomic Abnormalities

A variety of anatomic abnormalities, including iatrogenic and disease-related abnormalities, lead to stagnation of small intestinal contents, resulting in bacterial overgrowth. The bacteria in SIBO are similar to those found in the normal colon, and certain organisms are common. Common aerobic organisms include Escherichia coli and Streptococcus, Staphylococcus, Micrococcus, Klebsiella, and Proteus species. Common anaerobic species include Lactobacillus, Bacteroides, Clostridium, Veillonella, Fusobacterium, and Peptostreptococcus.50 The classic anatomic cause of SIBO is a blind loop resulting from abdominal surgery, such as Billroth II partial gastrectomy; other anatomic abnormalities that can result in SIBO include intestinal strictures and small bowel diverticulosis.

Motility Disorders

Disorders affecting small intestinal peristalsis, such as scleroderma,43 diabetes mellitus,51 and chronic idiopathic intestinal pseudo-obstruction52 constitute the next most common cause of SIBO after anatomic abnormalities. SIBO is well recognized in scleroderma and occurs mainly in patients with small intestinal involvement72 who have limited cutaneous systemic sclerosis. Diarrhea is the most important symptom. The somatostatin analog octreotide is effective in the management of SIBO associated with scleroderma.73 Although small intestinal dysmotility is thought to be the main predisposing factor in diabetes, SIBO in diabe­ tics is not especially associated with autonomic neuro­ pathy.74 Treatment of SIBO in diabetics improves orocecal transit time.75

Fistula or Ileocecal Valve Resection

The ileocecal valve prevents reflux of colonic bacteria into the small intestine, and resection of the valve or develop­ ment of fistulas between the colon and upper gastrointesti­ nal tract can lead to reflux of colonic contents into the small intestine, with ensuing bacterial overgrowth.53,54

Reduced Gastric Acid Secretion

Achlorhydria is known to be a predisposing factor for SIBO, and SIBO has been described in patients after vagotomy,55 in those with atrophic gastritis, and in those taking acid suppressants.56-59 SIBO occurs more often in patients taking proton-pump inhibitors (PPIs) than in those taking hista­ mine H2 receptor antagonists,58 but clinical malabsorption does not appear to occur in this situation.59

Aging

Advancing age seems to be an independent risk factor for SIBO, but it is not clear if overgrowth results from the aging process itself and age-related changes in intestinal motility or if it is a consequence of achlorhydria. Early studies in this area found that SIBO was a common (and commonly unrecognized) cause of malabsorption in the elderly44,60 and that many such patients did not have an obvious predispos­ ing factor, such as a blind loop. More-recent studies have reported SIBO in asymptomatic elderly persons residing in the community. These patients, although asymptomatic, had lower weights and body mass indices (BMI) than expected, and treatment with antibiotics increased both weight and BMI.46,47 In contrast, a Japanese study reported SIBO (diagnosed by glucose hydrogen breath test) in 25.6% of disabled older adults but in none of the healthy older adults.61

Chronic Liver Disease

SIBO appears to be common in patients with chronic liver disease,62,63 is more common in patients with advanced (Child class C) liver disease,63 and may be an independent risk factor for spontaneous bacterial peritonitis,64 although this association is controversial.65 No association with any particular cause of chronic liver disease has been found,66 but SIBO does not occur in cirrhotic patients who do not have portal hypertension.67 The etiology of SIBO in patients with chronic liver disease is likely to be related to distur­ bances in gastrointestinal motility67 and possibly to the use of antacids,65 both of which can foster proliferation of bacteria. Small intestinal dysmotility is more severe in cirrhotic patients with a history of spontaneous bacterial peritonitis,64 and treatment of SIBO improves motility.68 Liver transplantation improves small bowel dysmotility in cirrhotic patients.68 Antibiotics and prokinetic agents are effective in reducing the SIBO associated with cirrho­ sis.69 SIBO in cirrhosis is associated with systemic endo­ toxemia.70 Oral conjugated bile acids reduce bacterial overgrowth and endotoxemia in cirrhotic rats, suggesting a contributory role for cholestasis in cirrhotic patients with SIBO.71

Other Causes

SIBO is present in many patients with celiac disease who have persistent symptoms despite their adherence to a gluten-free diet.76 It is not clear why this is so, but a motility disturbance seems the most likely explanation. SIBO is common in Crohn’s disease, particularly in patients who have had previous intestinal resection, and orocecal transit time has been reported to be prolonged in Crohn’s patients with SIBO.77 Positive results on glucose

1773

1774

Section X  Small and Large Intestine hydrogen breath tests are particularly associated with the presence of a small bowel stricture.78 SIBO is common in chronic pancreatitis.79 SIBO in this setting may be caused by small bowel dysmotility resulting from chronic opioid use and achlorhydria. Furthermore, pancreatic juice may have an antibacterial effect, so its absence might allow enteric bacteria to proliferate more freely.80 SIBO in rheumatoid arthritis is associated with high disease activity and does not appear to be related to achlorhydria.81 SIBO occurs in late radiation enteritis and appears to be related to intestinal dysmotility.82 SIBO is common in chronic kidney disease, which is associated with neuropathic-type abnormalities of small intestinal motility.83 SIBO occurs commonly in cystic fibrosis (CF). Use of azithromycin is paradoxically associated with an increased risk of a positive breath test for SIBO.84 In a murine model of CF, eradication of SIBO decreased intestinal mucus secretion.85 SIBO has been reported in patient populations with inter­ stitial cystitis,86 acne rosacea,87 morbid obesity,88 fibromyal­ gia,89 acromegaly,90 and focal segmental ischemia (see Chapter 114.) Many patients with SIBO have more than one predispos­ ing factor. For example, SIBO can develop in patients with scleroderma who receive PPI therapy for gastroesophageal reflux.

MECHANISMS OF MALABSORPTION

SIBO classically causes a combination of megaloblastic anemia (due to vitamin B12 deficiency) and steatorrhea (due to fat malabsorption). Megaloblastic anemia was described in association with intestinal strictures as long ago as 1897.91 Vitamin B12 deficiency is caused by bacterial utilization of the vitamin within the intestinal lumen before it can be absorbed across the mucosa.92 Anaerobic organisms mainly are responsible for the vitamin B12 deficiency, and in animal studies, only therapy directed against anaerobes reversed the deficiency.93 Unlike aerobic bacteria, anaerobes can use vitamin B12 both in its free form and complexed with intrin­ sic factor.94 Anaerobic bacteria deprive the host of ingested vitamin B12 and exacerbate B12 deficiency by using the vitamin to produce inactive cobamides, which then can compete with dietary B12 for ileal binding sites, thereby decreasing absorption of the vitamin.95 Deficiencies of thia­ mine96 and nicotinamide also have been reported in SIBO.97 Folate levels tend to be high in SIBO, because the bacteria synthesize folate,98 which then is absorbed and used by the host.99 Much of our knowledge on the mechanisms of malabsorp­ tion in SIBO is derived from animal models of blind loops.100,101 Malabsorption of fat and fat-soluble vitamins results mainly from deconjugation of bile acids,102 and administration of conjugated bile acids has been reported to reverse steatorrhea in human and animal studies.103 Defi­ ciencies of vitamins A, D,104 and E105 have been reported, but vitamin K deficiency is uncommon because production of vitamin K by luminal bacteria offsets any deficiencies attributed to fecal fat loss. SIBO leads to carbohydrate malabsorption by reducing brush border disaccharidase levels.106-108 In animal studies, bacterial extracts of cultures from experimentally created blind loops contain proteases that can remove components of the intestinal surface membrane.109 These proteases appear to have elastase-like substrate specificity and may be etiologic in disaccharidase deficiency. Lactose intolerance

is common and contributes to the diarrhea that typifies SIBO. Bacterial fermentation of carbohydrates contributes to abdominal discomfort and bloating in SIBO and is the basis for the various breath tests used to diagnose the condition. Protein malabsorption in SIBO is caused by a number of factors: decreased absorption of amino acid and peptides, which has been described in animal models and can result from mucosal damage110; low levels of enterokinase, which can impair the activation of pancreatic proteases111; and protein-losing enteropathy.112 Although hypoproteinemia is common in SIBO, manifestations of severe hypoprotein­ emia, such as edema, are rare. Small intestinal histologic findings generally are normal in patients with SIBO, and in one study, morphometric findings in the small intestine also were described as normal.113 Abnormalities of small intestinal mucosa (e.g., villus atrophy, cellular infiltration of the lamina propria, intraepithelial lymphocytosis) have been described in some patients with SIBO, and these changes revert to normal fol­ lowing treatment with antibiotics.114 Electron microscopy studies of experimental animals with SIBO have described enterocyte abnormalities, such as vacuolization of micro­ villus membranes and mitochondrial swelling.115

CLINICAL FEATURES

SIBO may be difficult to diagnose because symptoms associ­ ated with the predisposing disorder can predominate. The classic clinical presentation of SIBO is that of a malabsorp­ tive state characterized by steatorrhea and vitamin B12 defi­ ciency that is not reversible with intrinsic factor. Patients with vitamin B12 deficiency can present with neurologic symptoms, central or peripheral neuropathy, and symptoms of anemia, such as fatigue, breathlessness, and chest pain. Patients with steatorrhea can report weight loss, diarrhea, and abdominal bloating and discomfort. Associated fatsoluble vitamin deficiency can occur, leading to night blind­ ness (in vitamin A deficiency) and metabolic bone disease (in vitamin D deficiency). Osteoporosis is another wellrecognized complication of SIBO.116,117 The clinical presentation of SIBO appears to be changing. Older references to clinical features of SIBO emphasized steatorrhea, megaloblastic anemia, and a history of surgery leading to blind loop syndrome. In a somewhat more modern series, Toskes and Kumar reported data for 100 consecutive, albeit highly selected, symptomatic patients referred for 14C-xylose breath testing118 and found a history of gastrointestinal surgery in only 15%. The three most common associated conditions, which accounted for more than 90% of the positive results on 14C-xylose breath tests for the patients in their referral center, were gastroparesis, chronic pancreatitis, and irritable bowel syndrome. Diar­ rhea, bloating, and flatulence were the most common symp­ toms. More recent studies demonstrate that the clinical presentation of SIBO may be less dramatic than the classic descriptions of SIBO and have milder symptoms. The wide use of breath tests is one reason we see this newer type of SIBO patient.

Small Intestinal Bacterial Overgrowth and Irritable Bowel Syndrome

Many patients with SIBO fulfill the Rome Criteria for IBS, and considerable debate on the relationship between IBS and SIBO has ensued since a study in 2000 reported that 78% of patients with Rome criteria-positive IBS tested posi­ tive for SIBO by lactulose hydrogen breath testing.119 Anti­ biotic therapy led to clinical improvement and normalization of the breath test. This study also provoked criticism of its

Chapter 102  Enteric Microbiota and Small Intestinal Bacterial Overgrowth selection criteria and study design and raised concerns regarding the accuracy of the lactulose hydrogen breath test.120 Two subsequent studies found no difference in lactu­ lose hydrogen breath test positivity between IBS patients and controls.121,122 A Swedish study, using cultures of intes­ tinal aspirate to diagnose SIBO, reported a prevalence of SIBO of 4% in both IBS patients and controls.123 The hy­ pothesis that SIBO is a significant etiological factor in IBS remains unproved. Nonetheless, the putative relationship between SIBO and IBS has stimulated a number of clinical trials of antibiotic therapy in IBS. Three relatively small controlled trials reported a modest benefit of antibiotic therapy in IBS patients,124-126 although the design and statistical analysis of these trials have been criticized.45 The major benefit appeared to be in patients with abdominal bloating, suggest­ ing that the predominant effect of antibiotic therapy is to reduce gas-forming bacteria. Symptom improvement in these studies did not consistently correlate with normaliza­ tion of the lactulose hydrogen breath test, however, suggesting that the benefit of antibiotic therapy might be reduction of colonic, rather than small intestinal, gasforming bacteria. These topics are discussed more fully in Chapter 118.

Small Intestinal Bacterial Overgrowth and Nonalcoholic Steatohepatitis

Considerable interest has arisen in the putative association between SIBO and nonalcoholic steatohepatitis (NASH). It has been postulated that SIBO might play a role in the pathogenesis of NASH,127 because NASH is a common com­ plication of jejunoileal bypass surgery for morbid obesity and can be reversed with metronidazole treatment. Anti­ biotic treatment prevents hepatic and bile duct injury in genetically susceptible rats with surgically created blind loops and SIBO,128,129 although the pattern of hepatic and biliary injury in this experimental situation was histologi­ cally and radiologically more compatible with primary scle­ rosing cholangitis than with NASH. Wigg and colleagues130 postulated that SIBO might lead to NASH by altering small intestinal permeability and thereby increasing absorption of endotoxin. They studied 22 patients with NASH and found SIBO (by lactulose hydrogen breath testing) in 50% of sub­ jects; serum endotoxin levels and small intestinal permea­ bility, measured by the lactulose-rhamnose test, however, were normal in the patients with SIBO.

DIAGNOSIS

The diagnosis of SIBO should be considered in any patient with malabsorption and a predisposing condition. As mentioned earlier, most patients today do not have pre­ disposing surgically induced anatomic abnormalities. It is likely that SIBO is commonly overlooked in patients without known predisposing factors and in patients who have nonspecific symptoms. Blood tests in patients with SIBO typically reveal a macrocytic anemia: Vitamin B12 levels are low, and folate levels may be high. Steatorrhea may be confirmed by three-day quantitative fecal fat collection; this test has understandably fallen from favor with patients and laboratory staff, and qualitative microscopic examina­ tion of fresh stool for fat globules usually is performed.131,132 If an anatomic defect is suspected as the cause of SIBO, appropriate barium studies may be used to define the anatomy.

Aspiration

The gold standard test for the diagnosis of SIBO is aspiration of small intestinal fluid with culture and bacterial counts of

the aspirate; presence of more than 105 CFU/mL of duodenal aspirate is considered diagnostic. Unfortunately, such aspi­ ration is invasive and time-consuming. Moreover, although it still is recommended by most experts, some investigators have raised concerns that the test might miss bacterial overgrowth occurring more distally in the small intestine. Corazza and colleagues,133 however, collected intestinal juice at two different levels of the proximal jejunum and reported a highly significant correlation between the bacte­ rial counts at these sites. Other potential problems with aspiration of small intestinal fluid include contamination of the aspirate with bacteria from the mouth and technical difficulties with transport and culture of the aspirate. Contamination with oropharyngeal bacteria may be con­ trolled for by simultaneous culture of saliva and jejunal aspirate.134 Several techniques for collecting small intestinal contents have been described, including duodenal intubation with fluoroscopic guidance and endoscopic collection of fluid,135 and brushing of the duodenal mucosa with a cytology brush.136 Culture of unwashed small intestinal mucosal biopsy specimens is an alternative to culture of a small intestinal aspirate, although the former method appears to have a lower sensitivity compared with culture of aspi­ rates.137 Aspirate can be collected easily during routine endoscopy, and this is probably the easiest method in routine clinical practice. Small intestinal aspirate is col­ lected by placing a sterile suction catheter inside a sterile overtube, which is passed through the suction channel of the endoscope. The aspirate should be placed immediately in aerobic and anaerobic transport vials, and the aspirate should be plated for aerobic and anaerobic organisms as soon as possible. High levels of jejunal fluid volatile fatty acids, such as acetate and propionate, have been reported in SIBO.138 These acids may be measured by gas-liquid chromatogra­ phy; although the technique is highly specific, the sensitiv­ ity is low,133 and this test is rarely used.

Breath Tests

A variety of noninvasive tests have been developed for diagnosing SIBO. The 14C-glycocholic acid breath test was one of the first breath tests used for this purpose and is based on the ability of bacteria to deconjugate bile salts. 14 C-glycine is produced and metabolized, resulting in a peak of 14CO2 in the expired air. The test has a low sensitivity, because not all bacteria are capable of such deconjugation, and the test has a low specificity, because increased colonic deconjugation of bile salts can occur with ileal disease or following ileal resection.139 The test therefore cannot distin­ guish between SIBO and ileal malabsorption and has largely fallen out of favor. The currently used breath tests are based on the ability of bacteria to produce hydrogen or radiolabeled carbon dioxide after metabolizing a substrate such as glucose, lactulose, or xylose. Breath tests are simple and noninvasive and there­ fore are more attractive than is duodenal intubation or endoscopy for collecting intestinal aspirates. These breath tests, however, do have several potential problems49: About 15% of the population are methane producers (in persons who are colonized with Methanobrevibacter smithii, hydro­ gen reacts with carbon dioxide to form methane, so less hydrogen is produced than in non-methane producers). Both slow and rapid small intestinal transit can affect the accuracy of these tests. An acidic environment in the colon, such as occurs with ingestion of nonabsorbable carbohy­ drates (e.g., lactulose), inhibits bacterial carbohydrate metabolism.140 Several patient-related factors, such as recent

1775

1776

Section X  Small and Large Intestine Table 102-3  Sensitivity and Specificity of Breath Testing in the Diagnosis of Small Intestinal Bacterial Overgrowth: Summary of Clinical Studies NUMBER OF PATIENTS REFERENCE Glucose Hydrogen 133 144 145 146 147 148 Lactulose Hydrogen 133 148 149 150 14 C-D-Xylose 147 152 153 154 155 156 157

Total

With Overgrowth

SENSITIVITY (%)

SPECIFICITY (%)

77 45 30 40 46 81

44 27 20 29 24 32

62 93 75 27-52 58 44

83 78 30 36-80 86 80

77 81 27 28

44 32 9 18

68 31 89 17

44 86 100 70

46 12 60 30 10 20 47

24 12 23 20 6 10 14

42 100 65 95 60 60 79

86 — 59 100 40 90 85

diet, smoking, and exercise, can influence baseline levels of breath hydrogen (see later). The literature on breath tests in SIBO is confusing, with wide variations in sensitivity and specificity (Table 102-3).

colleagues,133 reported sensitivity of 62% and specificity of 83%. Very rapid intestinal transit can lead to a false-positive test result, because glucose can reach the colon before it can be absorbed.

Glucose Hydrogen The glucose hydrogen breath test probably is the most widely used breath test in clinical practice: The substrate is inexpensive, and the hydrogen meter is economical, por­ table, and easy to use. The glucose hydrogen breath test first was reported as a diagnostic test for SIBO in 1972 by Bond and Levitt.141 Normally, glucose is absorbed completely in the upper small intestine; with bacterial overgrowth, however, the glucose is cleaved by bacteria into carbon dioxide and hydrogen. The hydrogen is measured in the exhaled breath (at baseline and then every 30 minutes for 2 hours); a rise of 20 parts per million (ppm) above the base­ line is regarded as diagnostic of SIBO. Fasting breath hydro­ gen levels of more than 20 ppm also are considered positive. High baseline hydrogen levels also are common in untreated celiac disease and normalize after gluten withdrawal for as-yet-unknown reasons.142 Patient preparation is important for the glucose hydrogen test143: Patients must avoid smoking and ingestion of non­ fermentable carbohydrates, such as pasta and bread, the night before the test, because these factors can raise baseline breath hydrogen values. Exercise can induce hyperventila­ tion, thereby reducing baseline breath hydrogen values, and should be avoided for two hours before the test. Some authors recommend an antibacterial mouth rinse before testing to prevent premature hydrogen or carbon dioxide production from the action of the oral flora on the glucose substrate. A number of studies have compared the glucose hydrogen breath test against the gold standard of culture of intestinal aspirate (see Table 102-3). Sensitivity levels from 27% to 52% have been reported, with specificity rates between 30% and 83%133,144-148; the largest study, by Corazza and

Lactulose Hydrogen The lactulose hydrogen breath test is based on a principle similar to that of the glucose hydrogen breath test: Lactulose is a disaccharide that is not absorbed in the small intestine but is metabolized by bacteria in the proximal colon, pro­ ducing a late peak in exhaled hydrogen. In the presence of bacterial overgrowth, an early hydrogen peak is observed. Results of this test may be difficult to interpret with either slow or fast intestinal transit, and sensitivity and specificity have been disappointing133,149,150 (see Table 102-3); Corazza and associates reported sensitivity and specificity rates of 68% and 44%, respectively.133 Sensitivity of the test may be increased by the addition of scintigraphy to correct for abnormalities of intestinal transit,150 but the lactulose hydrogen breath test cannot be recommended for routine clinical use. Xylose The 14C-xylose and 13C-xylose breath tests measure labeled carbon dioxide that is produced by breakdown of labeled substrates by bacteria. The isotope may be radioactive (14C) or stable (13C); the stable isotope has been used in chil­ dren.151 d-Xylose is the most widely used substrate and is a good substrate for breath testing for SIBO because it is absorbed completely in the small intestine, is metabolized minimally, and is catabolized by Gram-negative bacteria. The 14C-d-xylose breath test appears to perform better than the glucose or lactulose hydrogen breath test (see Table 102-3) but, as with these other breath tests, widely differing levels of accuracy have been reported, with sensitivity rates ranging from 42% to 95% and specificity rates between 40% and 100%.147,152-157 The 14C-d-xylose breath test result is con­ sidered positive when the “cumulated dose at four hours

Chapter 102  Enteric Microbiota and Small Intestinal Bacterial Overgrowth exceeds 4.5% of the administered radioactivity.”49 Distur­ bances in intestinal transit particularly affect the perfor­ mance of this test, and accuracy may be improved by the addition of a transit marker (such as barium or diatrizoate meglumine–diatrizoate sodium [Gastrografin]) and radio­ logic imaging.156

Other Tests

Other noninvasive tests described for SIBO include mea­ surement of urinary cholyl-para-aminobenzoic acid (PABA) and serum bile acids. Cholyl-PABA is a synthetic substrate made by conjugating cholic acid with PABA, which is hydrolyzed by the bacterial enzyme cholyl hydrolase to release PABA158; this PABA-based test, however, does not accurately distinguish between SIBO and other causes of malabsorption.159 Elevated free serum bile acids have been reported in SIBO, but the test depends on the presence of bacteria that deconjugate bile salts, such as Bacteroides.160

Choice of Test

It is our view that duodenal intubation with collection and culture of small intestinal aspirate remains the gold stan­ dard test for diagnosing SIBO. Small intestinal juice can be collected easily during standard endoscopy, and we recom­ mend endoscopy with duodenal biopsy and collection of small intestinal juice for culture as a standard diagnostic approach in patients with malabsorption. It is common practice, however, to provide empirical antibiotic treatment for patients with suspected SIBO, without either breath testing or culture of small intestinal aspirate.

TREATMENT

Attention should be given to the patient’s nutritional state, and any vitamin deficiency should be corrected (see Chap­ ters 4, 5, and 100). A lactose-free diet can ameliorate the diarrhea. If possible, any predisposing anatomic or func­ tional abnormality should be corrected, but in practice, this is unlikely to be an option. Acid-lowering medication should be discontinued, if possible. A variety of antibiotics have been reported to be effective in SIBO, but little evidence exists to favor one agent over another. Antibiotics that have been reported to be effective include metronidazole, amoxicillin, amoxicillinclavulanate, ciprofloxacin, tetracycline, and cotrimoxazole. One randomized crossover trial reported that norfloxacin and amoxicillin-clavulanate were effective in SIBO.161 In another study, rifaximin and chlortetracycline normalized results on glucose hydrogen breath testing in 70% and 27%, respectively, of patients with SIBO.162 Both ciprofloxacin and metronidazole were found to be highly effective in SIBO associated with Crohn’s disease, and although these antibiotics have been used for primary therapy in Crohn’s disease, normalization of breath tests occurred in most of the patients in this study.163 There have been several reports on the use of the nonab­ sorbable antibiotic rifaximin in SIBO. Rifaximin at a dose of 1.2 g/day and 1.6 g/day leads to normalization of the glucose hydrogen breath test in 58% and 80% of patients, respectively.164 Metronidazole is more effective than rifa­ ximin, at least in patients with SIBO associated with the blind-loop syndrome.165 Recurrence of SIBO after rifaximin treatment is common.166 Therapy usually is given initially for two weeks, and then clinical response is assessed; it may be useful to repeat a breath test or culture of small intestinal aspirate. Many patients with an underlying anatomic or motility disorder require permanent antibiotic treatment; in such patients, it

is usual to rotate antibiotic treatment every two weeks or, alternatively, to give antibiotics for two of every four weeks. Continuous treatment with a single agent can lead to antibiotic resistance or to side effects associated with long-term use, such as peripheral neuropathy in patients given metronidazole. The somatostatin analog octreotide stimulates intestinal motor activity when administered in low dosage. Given subcutaneously at 50 µg once daily for three weeks, it has been reported to be effective in SIBO associated with scleroderma.73 At higher doses (200 µg three times daily), octreotide paradoxically can cause SIBO by inducing hypo­ motility.167 The prokinetic agent cisapride has been reported to be effective in SIBO associated with cirrhosis,168 but the drug is no longer available in the United States, and its use in several countries is strictly controlled because of risk of drug interactions and cardiac arrhythmias. Probiotic therapy is a logical and attractive approach to the management of SIBO, but it has been examined in only a few studies. Saccharomyces boulardii does not appear to be effective, and in one double-blind crossover study, Lactobacillus fermentum KLD showed no advantage over placebo.169 A small uncontrolled trial showed that Lactobacillus plantarum 299V and Lactobacillus GG bene­ fited children who had SIBO associated with short bowel syndrome.170 In developed countries, SIBO is probably second only to celiac disease as the most common cause of malabsorption. The condition no longer manifests commonly with the classic features of steatorrhea and megaloblastic anemia, and most patients do not have a blind loop or other predis­ posing anatomic abnormalities. Many patients have nonspe­ cific symptoms similar to those of IBS. Although the glucose hydrogen and 14C-xylose breath tests are simple and nonin­ vasive, the gold standard test for diagnosis is culture of small intestinal aspirate. The aspirate can be easily col­ lected at endoscopy, which usually is performed to obtain biopsy specimens of the small intestine during evaluation of malabsorption. Treatment with one of several broadspectrum antibiotics is simple and effective.

KEY REFERENCES

Artis D. Epithelial-cell recognition of commensal bacteria and mainte­ nance of immune homeostasis in the gut. Nat Rev Immunol 2008; 8:411-20. (Ref 28.) Backhed F, Ding H, Wang T, et al. The gut microbiota as an environ­ mental factor that regulates fat storage. Proc Natl Acad Sci U S A 2004; 101:15718-23. (Ref 32.) Bratten J, Spanier J, Jones MP. Lactulose hydrogen breath testing does not discriminate patients with irritable bowel syndrome from healthy controls. Am J Gastroenterol 2008; 103:958-63. (Ref 122.) Castiglione F, Rispo A, Di Girolamo E, et al. Antibiotic treatment of small bowel bacterial overgrowth in patients with Crohn’s disease. Aliment Pharmacol Ther 2003; 18:1107. (Ref 163.) Di Sefano M, Miceli E, Missanelli M, et al. Absorbable vs non-absorbable antibiotics in the treatment of small intestine bacterial overgrowth in patients with blind-loop syndrome. Aliment Pharmacol Ther 2005; 21:985. (Ref 165.) Marchesi J, Shanahan F. The normal intestinal microbiota. Curr Opin Infectious Dis 2007; 20:508-13. (Ref 8.) O’Hara AM, Shanahan F. The gut flora as a forgotten organ. EMBO Rep 2006; 7:688-93. (Ref 2.) Lauritano EC, Gabrielli M, Scarpellini E et al. Small intestinal bacterial overgrowth recurrence after antibiotic therapy. Am J Gastroenterol 2008; 103:2031-5. (Ref 166.) Parlesak A, Klein B, Schecher K, et al. Prevalence of small bowel bacterial overgrowth and its association with nutrition intake in nonhospitalised older adults. J Am Geriatr Soc 2003; 51:768-73. (Ref 47.) Pimentel M, Park S, Mirocha J, et al. The effect of a non-absorbed oral antibiotic (rifaximin) on the symptoms of the irritable bowel

1777

1778

Section X  Small and Large Intestine syndrome: A randomized trial. Ann Intern Med 2006; 145:557-63. (Ref 125.) Rakoff-Nahoum S, Paglino J, Eslami-Varzaneh F, et al. Recognition of commensal microflora by Toll-like receptors is required for intestinal homeostasis. Cell 2004; 118:229-41. (Ref 21.) Riordan SM, McIver CJ, Wakefield D, et al. Small intestinal mucosal immunity and morphometry in luminal growth of indigenous gut flora. Am J Gastroenterol 2001; 96:494-500. (Ref 113.) Romagnuolo J, Schiller D, Bailey RJ. Using breath tests wisely in a gastroenterology practice: An evidence-based review of indications and pitfalls in interpretation. Am J Gastroenterol 2002; 97:2113-26. (Ref 49.) Scarpellini E, Gabrielli M, Lauritano CE et al. High dose rifaximin for the treatment of small intestinal bacterial overgrowth. Aliment Pharmacol Ther 2007; 25:781-6. (Ref 164.) Sharara AI, Aoun E, Abdul-Baki H, et al. A randomized double-blind placebo-controlled trial of rifaximin in patients with abdominal

bloating and flatulence. Am J Gastroenterol 2006; 101:326-33. (Ref 126.) Turnbaugh PJ, Ley RE, Hamady M, et al. The human microbiome project. Nature 2007; 449:804-10. (Ref 4.) Vanner, S. The lactulose breath test for diagnosing SIBO in IBS patients: Another nail in the coffin. Am J Gastroenterol 2008; 103:964-65. (Ref 120.) Vanner S. The small intestinal bacterial overgrowth/irritable bowel syn­ drome hypothesis: Implications for treatment. Gut 2008; 57:1315-21. (Ref 45.) Wigg AJ, Roberts-Thomson IC, Dymock RB: The role of small intestinal bacterial overgrowth, intestinal permeability, endotoxaemia, and tumour necrosis factor alpha in the pathogenesis of non-alcoholic steatohepatitis. Gut 2001; 48:206-11. (Ref 130.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

103 Short Bowel Syndrome Alan L. Buchman

CHAPTER OUTLINE Etiology  1779 Incidence and Prevalence  1779 Pathophysiology  1779 Loss of Absorptive Surface Area  1780 Loss of Site-Specific Transport Processes  1782 Loss of Site-Specific Endocrine Cells and Gastrointestinal Hormones  1782 Loss of the Ileocecal Valve  1783 Intestinal Adaptation to Resection  1783 Medical Management  1783 Limited Ileal Resection  1783 Extensive Small Intestinal Resection and Partial Colectomy  1784

Short bowel syndrome (SBS) is characterized by malabsorption due to insufficient intestinal surface area, such that the affected person is unable to absorb sufficient fluid, energy, or nutrients to sustain life in the absence of specialized nutritional support. This syndrome, also known as intestinal failure, occurs in adults in whom less than 200 cm of small intestine is present. The spectrum of SBS, however, ranges from limited ileocolonic resections with moderate nutritional compromise to extensive small intestinal and colonic resections, with duodenostomy, proximal jejunostomy, or jejunocolonic anastomosis and severe nutritional consequences. SBS also may be a congenital condition or can result from a variety of congenital causes.

ETIOLOGY The major causes of SBS in adults are Crohn’s disease for which multiple intestinal resections have been performed; mesenteric infarction from venous or arterial thrombosis, arterial embolism, or midgut volvulus; massive enterectomy performed to manage traumatic injuries or tumor resection, and radiation injury (Table 103-1). The causes of SBS in the pediatric population are congenital abnormalities (see Chapter 96), including gastroschisis, intestinal atresia, malrotation, aganglionosis, and necrotizing enterocolitis. More than 90% of infants now survive the extensive intestinal resections required for these conditions, and these patients need careful follow-up for their SBS as they mature to adulthood. Intestinal failure also can result from chronic intestinal pseudo-obstruction syndrome in both adults and children (see Chapter 120), as well as from unclassified sprue in adults (see Chapter 104) and congenital villus atrophy in children.

Home Parenteral Nutrition  1786 Complications  1788 Gallstones  1788 Liver Disease  1790 Calcium Oxalate Kidney Stones  1790 d-Lactic Acidosis  1790 Other Complications  1791 Surgical Management  1791 Intestinal Lengthening Procedures  1791 Intestinal Transplantation  1791 Pharmacologic Enhancement of Bowel Adaptation  1793 Survival and Quality of Life  1795

INCIDENCE AND PREVALENCE The incidence of SBS is difficult to assess in the United States, because of a lack of a national registry for affected persons and a lack of prospective studies in defined populations of patients who have undergone extensive intestinal resections. The incidence of severe SBS necessitating longterm parenteral nutrition is estimated to be 2 to 4 cases per 1 million persons per year, based on multinational European data.1 It is estimated that between 10,000 and 20,000 patients in the United States follow a home parenteral nutrition regimen for SBS. Approximately 50% to 70% of patients with SBS who initially require parenteral nutrition can be weaned from this therapy and therefore might not be reflected in the prevalence estimates.2,3 Such patients often still require aggressive nutritional monitoring. The incidence and prevalence of SBS associated with Crohn’s disease are decreasing now that infliximab and strictureplasty have become commonplace.

PATHOPHYSIOLOGY The major consequence of extensive intestinal resection is loss of absorptive surface area, which results in malabsorption of macronutrients, micronutrients, electrolytes, and water.4 The degree of malabsorption is determined by the length of the remnant intestine; the specific portions of small and large intestine resected, along with their sitespecific transport processes and endocrine cells; and the adequacy of adaptive processes in the residual intestine over time. Three types of intestinal resections typically are encountered: limited ileal resection for Crohn’s disease, often with cecectomy or right hemicolectomy; extensive

1779

1780

Section X  Small and Large Intestine Proximal small intestine Fat Sugars Peptides and amino acids Iron Folate Calcium Water Electrolytes

Ileocolonic Jejunocolonic End-jejunostomy anastomosis anastomosis Figure 103-1.  The three common types of intestinal resection and anastomosis observed in patients with short bowel syndrome: ileocolonic anastomosis, jejunocolonic anastomosis, and end-jejunostomy.

Table 103-1  Causes of Short Bowel Syndrome and Intestinal Failure In Adults Catastrophic vascular accidents Superior mesenteric arterial embolism Superior mesenteric arterial thrombosis Superior mesenteric venous thrombosis Chronic intestinal pseudo-obstruction* Intestinal resection for tumor Midgut volvulus Multiple intestinal resections for Crohn’s disease Radiation enteritis* Refractory sprue* Scleroderma and mixed connective tissue disease* Trauma In Children Congenital villus atrophy* Extensive aganglionosis* Gastroschisis Jejunal or ileal atresia Necrotizing enterocolitis *Functional short bowel syndrome also can occur in conditions associated with severe malabsorption, in which the bowel length often is intact.

ileal resection with or without partial colectomy and with jejunocolonic anastomosis; and extensive small intestinal resection and total colectomy resulting in proximal jejuno­ stomy (Fig. 103-1). Patients in the two latter groups commonly suffer from Crohn’s disease or had mesenteric infarction.

LOSS OF ABSORPTIVE SURFACE AREA Nutrient Malabsorption

The length of the small intestine is estimated at 3 to 8 meters, and nutrient absorption is preserved until more than one half of the small intestine is resected.5-9 Most macronutrients (carbohydrate, fat, and nitrogen) are absorbed in the proximal 100 to 150 cm of intestine.10 Specific areas of absorption in the small intestine of nutrients, minerals, vitamins, electrolytes, and trace elements are discussed in Chapters 99 to 101 and are illustrated in Figure 103-2. Enterocytes lining the small intestine appear uniform from the duodenum to the ileocecal valve, but a distinct proximal-to-distal gradient exists in both morphology and function.11 Villi are taller and crypts are deeper in the

Colon Water Electrolytes MCTs Amino acids

Middle small intestine Sugars Peptides and amino acids Calcium Water Electrolytes Distal small intestine Bile salts Vitamin B12 Water Electrolytes

Figure 103-2.  Specific areas of absorption of dietary constituents and secretions in the small intestine and colon. Macronutrients and micronutrients are absorbed predominantly in the proximal jejunum. Bile acids and vitamin B12 (cobalamin) are absorbed only in the ileum. Electrolytes and water are absorbed in both the small and the large intestine. Medium-chain triglycerides (MCTs), calcium, and some amino acids can be absorbed in the colon.

jejunum than in the ileum, and the activity of microvillus enzymes and nutrient absorptive capacity per unit length of intestine are several-fold higher in the proximal than in the distal small intestine; loss of part of the jejunum initially compromises nutrient absorption more than does loss of an ileal segment of similar length, because of these morphologic and functional differences. The ileum, however, eventually is able to compensate for jejunal loss, whereas the jejunum is unable to compensate for ileal absorption of bile salts and vitamin B12. Normal digestion and absorption depend on the gradual gastric emptying of partially digested nutrients, mixing of these nutrients with bile and pancreatic enzymes in the duodenum, and rapid digestion and absorption of the digestive products in the proximal small intestine. Patients with a proximal jejunostomy have rapid gastric emptying of liquids and rapid intestinal transit, which can compromise the gastric phase of digestion and result in inadequate mixing with biliary and pancreatic secretions, insufficient enzymatic digestion, and nutrient maldigestion. Rapid intestinal transit decreases nutrient-enterocyte contact time, and therefore, segmental absorption is decreased. Patients with a high jejunostomy are net secretors of salt and fluid, because jejunal fluid secretion is stimulated by oral intake and subsequent gastric emptying of nutrients; these patients excrete more fluid than they ingest, and accordingly, their fluid management may be challenging.12 Most patients whose jejunum is shorter than 100 cm and who have no colon require long-term parenteral nutrition. Preservation of even some colon at surgery is highly beneficial for nutrient absorption. The ileocecal valve acts as a brake to slow intestinal transit, thereby increasing nutrient-enterocyte contact time and enhancing absorption.

Chapter 103  Short Bowel Syndrome 50-g bread meal

Table 103-2  Daily Stomal or Fecal Losses of Electrolytes, Minerals, and Trace Elements in Severe Short Bowel Syndrome*

24 g of malabsorbed CHO Colonic bacterial fermentation

240 mmol of SCFA (60% acetate), hydrogen, methane, carbon dioxide, and sulfides

72 kcal Figure 103-3.  Colonic absorption of malabsorbed carbohydrate (CHO) in a hypothetical patient with short bowel syndrome following ingestion of a 50-g bread meal. Unabsorbed carbohydrates (∼24 g), nonstarch polysaccharides, and soluble fiber are fermented by colonic bacterial flora to hydrogen, methane, carbon dioxide, sulfides, and ∼240 mmol short-chain fatty acids (SCFAs), including acetate, butyrate, and propionate to generate 72 kcal. By comparison, normal persons absorb 220-720 mmol SCFA from fermentation of 30-60 g nonstarch polysaccharides. 10

100% colon remaining <100% colon remaining 0% colon remaining

Fecal energy loss (MJ/d)

8

6

4

2

0 0

350

300

250

200

150

100

50

Small intestine length (cm) Figure 103-4.  The role of the colon as an important digestive organ in patients with short bowel syndrome. Upwards of 1000 kcal/day (4.2 MJ) can be assimilated by means of the metabolism of short-chain fatty acids (SCFAs) synthesized during carbohydrate fermentation; as more of the colon is sacrificed, more energy is lost in the stool. (From Nordgaard I, Hansen BS, Mortensen PB. Colon as a digestive organ in patients with short bowel. Lancet 1994; 343:373-76.)

In addition, malabsorbed carbohydrates are fermented by bacterial enzymes in the colon to short-chain fatty acids (SCFAs), which are readily absorbed and used by colonocytes (Fig. 103-3). It has been estimated that this intracolonic digestive process can generate up to 1000 kcal (4.2 MJ) per day; in energy supply (Fig. 103-4), 1.0 MJ equals 238.8 kcal.13-15 Small intestine should be anastomosed to colon as soon as the patient is stable.

Water and Electrolyte Malabsorption

Loss of intestinal absorptive surface area can result in significant stomal or fecal losses of electrolytes, water, miner-

COMPONENT

AMOUNT lost

Sodium Potassium Calcium Magnesium Iron Zinc Copper

90-100 mEq/L 10-20 mEq/L 772 (591-950) mg/day 328 (263-419) mg/day 11 (7-15) mg/day 12 (10-14) mg/day 1.5 (0.5-2.3) mg/day

*For sodium and potassium, the average concentration per liter of stomal effluent is given. The values for minerals and trace elements are mean 24-hour losses, with the range in parentheses. See text for details.

als, and trace elements (Table 103-2). The proximal small bowel receives approximately 7 to 9 L daily of water and electrolytes from food and secretions each day, of which 6 to 8 L are reabsorbed (see Chapter 99). On unrestricted diets, patients with a proximal jejunostomy cannot reabsorb such large volumes, a consequence of which is that voluminous diarrhea develops, often complicated by hypovolemia, hyponatremia, and hypokalemia. For example, in one study,12 the diarrheal volume in six jejunostomy patients with a mean jejunal length of 50 cm ranged from 3.2 to 8.3 L/day when they were allowed free access to food and water. All six patients were in negative sodium and water balance, four of the six were in negative potassium balance, and all six required parenteral nutrition with electrolyte replacement and restriction of oral intake of food and water to avoid unacceptable stomal losses. In the same study,12 seven of nine other jejunostomy patients who had a mean jejunal length of 120 cm were able to maintain positive water and sodium balance under the same conditions; absorption of water, sodium, and potassium in these 15 jejunostomy patients was correlated with jejunal length. At least 100 cm of intact jejunum is required to maintain positive water and electrolyte balance, similar to the length of jejunum required for nutrient absorption. In general, patients with a proximal jejunostomy lose 90 to 100 mEq sodium and 10 to 20 mEq potassium per liter of stomal effluent (see Table 103-2).16 Some of these patients require long-term parenteral electrolyte and water supplements, often administered overnight, but others can maintain a positive balance by sipping a glucose-saline oral rehydration solution (ORS) throughout the day. The tight junctions of the jejunum are relatively leaky compared with the tight junctions of the ileum and colon, and therefore, a high NaCl concentration (greater than 90 mmol/L) is required in the glucose-saline solution to achieve net sodium and water absorption.17,18 Actively absorbed solutes (e.g., glucose, glucose polymers, galactose, oligopeptides, or l-amino acids) promote intestinal ion transport, although solutes also may be absorbed passively by means of solvent drag once active electrogenic Na+ absorption occurs. Water transport into the enterocyte is directly proportional to Na+ transport. Na+ also is absorbed by means of an active electrogenic mechanism coupled with Cl− and H+ exchange and solvent drag. Absorptive and secretory processes occur simultaneously. Active Na+ secretion occurs against a concentration gradient from the enterocyte by means of the sodium pump, activated by Na+K+-ATPase (adenosine triphosphatase) in the basolateral membrane (Fig. 103-5). A mixture of 90 to 120 mmol/L NaCl and 50 mmol/L glucose is recommended, although such a solu-

1781

1782

Section X  Small and Large Intestine Na+ D-glucose D-galactose

L-amino

S Na+

S

Na/GLU cotransporter (SGLT-1) Na+ Na+

Serosal glucose carrier (GLUT2)

S

Na+ Na+

ATPase pump K+

S

acids

S

60%

25% S

K+

S

S S

S

S

Diffusion

Figure 103-5.  Mechanism of solute-coupled Na+ cotransport. Sodium contained in oral rehydration solutions replaces diarrhea-associated Na+ losses and promotes water absorption by means of solvent drag. This schematic diagram illustrates the active, Na+-coupled cotransport of solute into the jejunal enterocyte, where the solutes may be glucose (GLU), glucose polymers, galactose, oligopeptides, or l-amino acids. ATP, adenosine triphosphate; GLUT, glucose transporter; S, solute; SGLT, sodiumglucose linked transporter.

tion might not be palatable. This mixture takes advantage of the coupled active transport of sodium with glucose and amino acids in the jejunum (see Chapter 99). Electrolyte and water absorption continue in the colon, and in normal humans only 100 to 150 mL of water is lost in the stool each day. The colon has a large reserve absorptive capacity for electrolytes and water, estimated to be 3 to 4 L of isotonic salt solution per day. Preservation of even part of the colon can reduce fecal electrolyte and water losses significantly in patients with SBS. A comparison of two groups of patients with similar jejunal length and jejunum either ending in a jejunostomy or anastomosis to colon showed that patients in the latter group were less likely to require oral or intravenous supplements.19

LOSS OF SITE-SPECIFIC TRANSPORT PROCESSES

Nutrient absorption potentially takes place at any level of the small intestine, albeit at different rates, owing to the proximal-to-distal gradient in functional activity of microvillus enzymes and transporters. The absorption of some compounds is restricted to certain areas of the small intestine (see Fig. 103-2); thus, calcium, magnesium, phosphorus, iron, and the water- and fat-soluble vitamins are absorbed predominantly in the duodenum and proximal jejunum (see Chapters 99 and 100). Most patients with SBS have an intact duodenum and a variable length of jejunum, so the development of iron, phosphorus, or water-soluble vitamin deficiency, even in patients with a proximal jejunostomy, is relatively uncommon. Calcium absorption was found to be highly variable in a large study of patients with small intestinal resections.20 The net absorption of calcium (intake minus fecal loss) ranged from +573 to −268 mg/day, with a median of +65 mg/ day; 64% of the patients, however, were in a negative calcium balance (intake minus fecal and urinary loss). In a

study of 25 patients with a mean jejunal length of 128 cm, large-volume diarrhea (2 to 6 L/day), and steatorrhea,21 hypocalcemia and hypomagnesemia developed in 13 and 18 patients, respectively, during a trial of enteral hyper­ alimentation—despite supplementation with calcium, magnesium, and vitamin D. Malabsorption of calcium and magnesium is a consequence of fat malabsorption, because these minerals are precipitated intraluminally by unabsorbed long-chain fatty acids. Calcium and magnesium absorption improve on a low-fat diet in patients with small intestinal resections.22 The active absorption of vitamin B12 and bile acids is restricted to the ileum. The B12-intrinsic factor complexes and bile acids are taken up by specific transport proteins in ileal enterocytes (see Chapters 64 and 100). Most patients with SBS have lost part or all of the ileum, as a result of which vitamin B12 and bile acid malabsorption develops. The degree of malabsorption depends on the length of resected ileum. Vitamin B12 malabsorption usually is demonstrable when more than 60 cm of ileum has been resected.4 Resection of less than 100 cm of ileum causes moderate bile acid malabsorption and increased bile acid loss to the colon or in stomal effluents.23 The increased loss of bile acids to the colon induces electrolyte and water secretion and can exacerbate diarrhea. More-extensive ileal resections (greater than 100 cm) cause severe bile acid malabsorption, which, if bile acid loss exceeds hepatic synthesis, can result in a reduced bile acid pool size, with insufficient micellar solubilization of lipolytic products and resultant steatorrhea. Following extensive ileal resection, fat malabsorption develops, as can fat-soluble vitamin deficiency; essential fatty acid (linoleic acid) deficiency is rare. Loss of unabsorbed long-chain fatty acids to the colon can exacerbate diarrhea if the fatty acids are hydroxylated by colonic bacteria, because hydroxylated fatty acids stimulate colonic electrolyte and water secretion.24

LOSS OF SITE-SPECIFIC ENDOCRINE CELLS AND GASTROINTESTINAL HORMONES

The synthesis of gastrointestinal hormones in the intestinal mucosa is distributed in a site-specific manner along the gastrointestinal tract (see Chapter 1). Gastrin, cholecysto­ kinin (CCK), secretin, gastric inhibitory polypeptide, and motilin are produced by endocrine cells in the proximal gastrointestinal tract and regulate secretory processes and motility. The area within which these hormones are synthesized usually is intact in patients with SBS, and hormonal profiles are normal. In approximately 50% of patients with extensive intestinal resections, however, hypergastrinemia and increased gastric acid secretion temporarily develop in the early postoperative phase.25,26 The cause of this postoperative hypergastrinemia is not known but could be loss of inhibitory signals, because it resolves spontaneously. Glucagon-like peptides 1 and 2 (GLP-1 and GLP-2), neurotensin, and peptide YY (PYY) are produced in the ileum and proximal colon, and these intestinal segments are often lost in SBS patients.27 GLP-1 and GLP-2 and PYY are released by intraluminal fat and carbohydrates, cause a delay in gastric emptying, and slow intestinal transit (the ileal brake).28,29 Jejunostomy patients demonstrate impaired release of these hormones in response to a meal, rapid gastric emptying, and rapid intestinal transit of liquids.30,31 Patients with SBS and preserved colon have increased GLP-1 and GLP-2 concentrations and demonstrate normal gastric emptying.32 GLP-1 and GLP-2 and PYY also have been shown to inhibit gastric acid secretion and to promote intestinal growth in animal models.

Chapter 103  Short Bowel Syndrome LOSS OF THE ILEOCECAL VALVE

The primary functions of the ileocecal valve are to separate ileal and colonic contents, thereby minimizing bacterial colonization of the small intestine, and to regulate emptying of ileal contents into the colon. The ileocecal valve is removed in most ileal resections, as a consequence of which intestinal transit time decreases, and bacterial overgrowth is risked if the ileum is anastomosed to the colon. Bacterial overgrowth can worsen nutrient and cobalamin malabsorption (see Chapters 100 and 102), because bacteria compete with enterocytes for nutrient assimilation. Rapid intestinal transit in these patients, however, can counteract the risk of bacterial colonization. Studies are lacking to document the role of bacterial overgrowth in malabsorption in patients with SBS.

INTESTINAL ADAPTATION TO RESECTION Adaptive changes in the remaining intestine after intestinal resection have been studied extensively in animal models and to a limited extent in humans33,34; adaptive changes are more pronounced in the ileum than in the jejunum. After jejunectomy and duodenoileal anastomosis, the ileum attains the morphologic characteristics of the jejunum, with taller villi and deeper crypts35; with time, an increase in ileal diameter and length also occurs. A prospective study of seven patients with jejunoileal bypass operation (20 cm of jejunum anastomosed to 25 cm of ileum) showed an increase in the length and diameter of the jejunum (80% and 40%, respectively) and ileum (128% and 50%, respectively) after 18 months of observation.36 An increase in absorptive capacity was demonstrated in another study of 41 patients with SBS (mean jejunal length, 119 cm) in whom the mean stool volume decreased from 2.5 to 0.9 L per day over a period of 3 months with continuous oral intake37; patients gained weight, and nitrogen balance increased from +3.2 g in the first month to +7.8 g in the second month postoperatively. The same study also demonstrated a gradual increase in intestinal transit time, which was most pronounced for ileal transit. The result of all of these changes is an increase in intestinal absorptive surface area, with an increase in microvillus enzyme activity and absorptive capacity per unit length of intestine.38 An improvement in mineral absorption with time also has been observed in a series of 30 patients with SBS (mean jejunal length, 81 cm) in whom fractional calcium absorption was correlated with time after surgery.39 In humans, these adaptive changes can take 1 to 2 years to develop fully. The younger the patient, the more profound the adaptive response. Adaptive changes depend on the presence of food and biliary and pancreatic secretions in the intestinal lumen40; adaptive hyperplasia of the ileum failed to develop in jejunectomized animals fed only by parenteral alimentation.41 To induce these adaptive processes, patients with SBS are encouraged to start oral intake as early as possible in the postoperative phase. Patients with SBS whose colon is in continuity demonstrate qualitative and quantitative changes in colonic flora that result in an increased capacity to metabolize carbohydrate and in an increased fecal bacterial mass.42 Adaptive hyperplasia is the result of an increase in crypt cell production rate, presumably mediated by growth factors released by the presence of food and secretions in the intestinal lumen. Vascular endothelial growth factor (VEGF), CCK, gastrin, insulin, neurotensin, GLP-2, and l-glutamine have been shown to stimulate intestinal growth in experi-

mental animals;43-45 studies in humans have not indicated any value of supplemental glutamine to enhance intestinal adaptation.46,47 These extracellular growth factors stimulate polyamine synthesis in crypt cells, which in turn induces increased DNA synthesis and mitotic activity.48 Inhibition of polyamine synthesis in jejunectomized animals prevents adaptive changes in the ileum.49 Elucidation of the mediators regulating enterocyte proliferation eventually can lead to development of pharmacologic interventions that can accelerate intestinal adaptation in patients with SBS. The presence of comorbid conditions and the health of the residual bowel and its blood flow are important factors in the prognosis for patients who have undergone massive enterectomy.

MEDICAL MANAGEMENT The initial management of the patient with SBS involves primarily supportive care designed to enhance the potential for survival. This care includes achievement of hemodynamic stability and appropriate fluid and electrolyte management. In the immediate postoperative phase, most patients with extensive intestinal resections are kept fasting and are supported with total parenteral nutrition (TPN). Weight and volume status are carefully monitored and stomal, fecal, and urinary losses of water, sodium, and potassium are measured to ensure optimal electrolyte and water balance. Histamine H2 receptor blockers or proton pump inhibitors are given intravenously to suppress hypergastrinemia-induced gastric acid hypersecretion and to limit volume losses.50,51 Patients with jejunostomies have stomal effluents up to several liters per day in this early phase, with obligatory losses of sodium, potassium, and possibly magnesium. Enteral tube feeding, followed by oral feeding, is begun in the late postoperative phase once the patient is hemodynamically stable, adequate intestinal blood flow has been restored, and postoperative ileus has resolved. Patients with extensive resections are kept fasting up to 5 to 10 days to allow a second-look operation at 24 to 48 hours, for the healing of enteric anastomoses, and to assess basal losses of water and electrolytes.

LIMITED ILEAL RESECTION

Patients with a limited ileal resection (less than 100 cm), with or without right hemicolectomy, may resume intake of solid food in the late postoperative phase. The response to solid food is determined mainly by the length of ileum removed and whether or not the right colon was resected; patients can develop diarrhea or steatorrhea with consumption of a regular diet. Secretory diarrhea without steatorrhea is the typical finding in limited ileal resections. Treatment with a bile acid-binding resin, such as cholestyramine (2-4 g with meals) or colestipol (1-2 g with meals) often ameliorates diarrhea if bile acid malabsorption is the main cause. Colestipol often is tolerated better than is cholestyramine. The diarrhea of some patients with limited ileal resection and right hemicolectomy does not respond to cholestyramine or colestipol despite documented bile acid malabsorption and presumably is due to loss of intestinal absorptive capacity for sodium chloride.52 Patients with documented fat malabsorption on a regular diet might have less severe steatorrhea while on a low-fat (40 g), high-carbohydrate diet; however, oral energy intake also will be reduced, because fat is calorically dense (9 kcal/g). Patients maintained on such a diet experience a decrease in diarrhea and steatorrhea and improve their net

1783

1784

Section X  Small and Large Intestine absorption of calcium, magnesium, and zinc.4 If necessary, medium-chain triglycerides (MCTs), which do not require micellar solubilization, can be added as a source of fat calories. The possibility of vitamin B12 malabsorption should be assessed with a Schilling test, and if the malabsorption is documented, parenteral B12, usually in a dose of 1 mg intramuscularly monthly, is required for life. Malabsorption of fat-soluble vitamins, calcium, and magnesium is a risk in patients with fat malabsorption. Fourteen of 27 patients with ileal resections of 50 to 150 cm and an intact colon were in negative calcium balance when studied on a fixed daily calcium intake of 800 mg supplemented with 400 to 800 IU of vitamin D daily.21 Supplementation with vitamins, calcium, and possibly magnesium should be initiated before overt signs of vitamin deficiency or hypocalcemia and hypomagnesemia develop. Magnesium supplementation by mouth may be unrewarding, because magnesium is a cathartic. Although magnesium gluconate is water soluble and therefore may be the most readily absorbed magnesium salt, some patients still require periodic parenteral replacement. Magnesium deficiency can occur despite a normal serum concentration, because most Mg2+ is present in the intracellular space. Therefore, measurement of 24-hour urine Mg2+ concentration is prudent in patients who have suspected magnesium deficiency but normal serum Mg2+ concentration. Magnesium deficiency can result in calcium deficiency, because the release of parathyroid hormone is impaired in the presence of hypomagnesemia.53 Most patients with SBS already are in a negative calcium balance54 and therefore an oral supplement of calcium at a daily dose of 800 to 1500 mg is recommended. The tests to assess vitamin and mineral balance and recommended dosages in patients with malabsorption are discussed in Chapters 99 and 100. Absorption of water-soluble vitamins, carbohydrates, and proteins is, in general, not compromised in patients with limited ileal resections.

EXTENSIVE SMALL INTESTINAL RESECTION AND PARTIAL COLECTOMY Management of Fluid and Electrolytes

Massive enterectomy is associated with gastric hypersecretion for approximately the first six months postoperatively. These patients benefit from the use of intravenous H2 antagonists or oral or intravenous proton pump inhibitors; absorption of orally ingested medications may be impaired, and more than the usual doses of these agents may be required (Table 103-3). Rapid intestinal transit contributes to malabsorption and diarrhea, and use of antidiarrheal drugs is common (see Table 103-3). These medications should be taken one hour before meals, and their effect on volume of diarrhea should be evaluated before long-term treatment is recommended. Use of antimotility agents is important to control fluid losses; such agents include loperamide hydrochloride (16 to 24 mg/day) and diphenoxylate (10 to 20 mg/day), codeine (30 to 240 mg/day), tincture of opium, and the somatostatin analog octreotide (50 to 100 µg two or three times a day). Most studies have shown that these agents reduce stomal output by up to 50%, but a positive water and electrolyte balance rarely is achieved. Octreotide usually is not necessary except for some patients with a proximal jejunostomy. Octreotide can slow intestinal transit and increase sodium and water absorption,55-58 but it also decreases splanchnic protein synthesis, thereby inhibiting postresectional intestinal adaptation57; the risk of cholelithiasis also is increased

Table 103-3  Therapeutic Agents Used to Decrease Intestinal Transit and Diarrheal Volume AGENT

DOSAGE

Loperamide* Diphenoxylate-atropine* Codeine phosphate* Tincture of opium Ranitidine† Omeprazole‡ Octreotide Clonidine

4-6 mg four times daily 2.5-5 mg four times daily 15-30 mg two to four times daily 0.6 mL (2.5 mg) two to four times daily 300 mg twice daily 40 mg twice daily 50-100 µg SC twice daily 0.3 mg transcutaneous patch once weekly

*The antidiarrheal agents loperamide, diphenoxylate-atropine, and codeine are given 1 hour before meals and at bedtime. Dosages may be increased over those recommended, because of incomplete absorption in patients with short bowel syndrome. † Cimetidine, famotidine, and nizatidine are alternatives. ‡ Esomeprazole, lansoprazole, rabeprazole, and pantoprazole are alternatives. SC, subcutaneously.

with octreotide.58 The α2-adrenergic agonist clonidine also may be useful to decrease diarrhea by its effects on chloride absorption. Transdermal administration avoids the potential for medication malabsorption.59 Glucose polymer-based ORSs should be provided to patients to improve hydration and thereby reduce TPN requirements. Glucose and sodium are absorbed by the same active transport mechanism and stimulate absorption of each other. In addition, glucose promotes sodium and water absorption by means of solvent drag (see Fig. 103-5).60 Therefore, because the jejunum is permeable to both sodium and chloride, passively absorbed solutions that have a high sodium chloride concentration are absorbed to a significant degree; sodium is not as readily absorbed from isotonic or hypotonic solutions. A simple solution developed by the World Health Organization (WHO) can be formulated by dissolving 2.5 g of table salt, 1.5 g of KCl (requires a prescription), 2.5 g of sodium bicarbonate (NaHCO3), and 1.5 g of table sugar (sucrose) in 1 L of water. This solution provides a sodium concentration of approximately 90 mmol/L. Additional salt may be added to increase the osmolarity as tolerated, to 100 to 120 mmol/L or more, which can increase effectiveness.61 Sodium losses actually increase when solutions are consumed that contain less sodium than is in the small bowel effluent (90 mmol/L). The use of ORS is not as critical in patients in whom the colon is intact, provided that sufficient dietary sodium is present, because of the colon’s ability to absorb sodium and water. For patients who have had significant jejunal resections, the addition of glucose to the ORS is not critical, because glucose does not enhance ileal water absorption.62 In addition to sodium losses, significant quantities of bicarbonate and magnesium are lost in feces.

Management with Diet

Patients with SBS should be encouraged to eat substantially more than usual (a hyperphagia diet) to compensate for malabsorption; they might need to consume two to three times as much energy as that normally ingested before their abdominal catastrophe. This may be the single most important dietary intervention to reduce parenteral nutrition requirements. It has been suggested that patients may counterbalance the discomfort associated with increased fecal volume by the satisfaction of recovering relatively normal eating habits and requiring less parenteral nutrition.21 Patients also should be encouraged to eat small portions

Chapter 103  Short Bowel Syndrome Table 103-4  Macronutrient Requirements in Patients with Short Bowel Syndrome COLON PRESENT Carbohydrate Complex carbohydrates/starches 30-35 kcal/kg/day Soluble fiber Fat MCT/LCT 20%-30% of caloric intake Protein Intact protein 1.0-1.5 g/kg/day

COLON ABSENT Variable types 30-35 kcal/kg/day LCT 20%-30% of caloric intake Intact protein 1.0-1.5 g/kg/day

LCT, long-chain triglycerides; MCT, medium-chain triglycerides.

throughout the day, rather than at defined meal times. Separation of liquid and solid portions of meals is imprac­tical and not associated with decreased fecal wet weight loss. Patients with SBS whose colon is in continuity should be provided a high complex-carbohydrate diet that includes starch, nonstarch polysaccharides, and soluble fiber (Table 103-4). These foodstuffs typically are not absorbed by the human small intestine63; however, when they pass undigested into the colon, colonic bacteria ferment these foodstuffs into SCFAs such as butyrate, acetate, and propionate. Approximately 75 mmol of SCFAs are produced from 10 g of unabsorbed carbohydrate (see Fig. 103-3). Butyrate is the preferred fuel for the colonocyte.64 A person with an intact colon can absorb a wide range of up to 310 to 740 kcal (1.3 to 3.1 MJ) daily when fed a 60% carbohydrate diet.15 Other studies have indicated that up to 525 to 1170 kcal (2.2 to 4.9 MJ) daily can be absorbed by an intact colon from fermentation of unabsorbed carbohydrate and soluble fiber.13 The amount of energy absorbed is proportional to the amount of residual colon15,19 (see Fig. 103-4) and can increase as part of the adaptive response to enterectomy.43,65 During this adaptive period, colonic bacteria increase and β-galactosidase and other enzymes appear to increase in concentration or activity.42 Sodium and water absorption are stimulated by SCFAs as well, although decreased fecal fluid and sodium losses have not been documented clinically.15 When more than 100 cm of terminal ileum has been resected, fat maldigestion can develop, because bile salt malabsorption leads to decreased micelle formation, which results in poor fat solubilization. Use of bile salt replacement therapy with ox bile or a synthetic conjugated bile acid (cholylsarcosine) has been reported only in a few patients,66,67 decreasing fecal fat in most, but leaving fecal volume either unchanged or increased. The bile acid-sequestering agent cholestyramine may be useful to decrease bile salt-related diarrhea in patients with less than 100 cm of terminal ileum resected, but it can worsen steatorrhea in patients who have undergone a more-significant resection; because of its binding to dietary lipid,68 fat-soluble vitamin deficiency also can develop. In addition, cholestyramine binds to many medications, including warfarin, antibiotics, beta-blockers, diuretics, oral hypoglycemia agents, and others. Limited data are available to support the use of low-fat diets in patients with massive enterectomy,69 although fat restriction often does lead to decreased steatorrhea in patients with limited terminal ileal resections. Because fat

is concentrated energy, however, dietary fat restriction results in decreased energy intake and can worsen the patient’s energy balance; a low-fat diet also may be unpalatable. A high-fat diet, although having greater energy content than that of a high-carbohydrate diet, is associated with increased loss of divalent cations (Ca2+, Mg2+, and Zn2+),70 slows gastric emptying, and can induce early satiety, leading to reduced total energy intake. Diets high in fat also can lead to water secretion from the colon. Because MCTs (C8 to C10, 8.3 kcal/g) are absorbed in the colon, dietary supplementation with MCTs can lead to increased energy absorption.71 MCT supplementation is of much more limited benefit in patients with an endjejunostomy. MCTs also do not supply essential fatty acids, and excessive MCT intake is associated with nausea, vomiting, and ketosis. Experience with long-term parenteral nutrition mainly has been gained in patients with severe SBS (see Chapter 5). Despite the limited adaptive capacity of the jejunum, approximately 50% of patients on a home parenteral nutrition regimen can discontinue TPN and resume oral intake after one to two years.72 The diet composition for patients with a jejunostomy on oral intake can be more liberal, because the percentages of energy absorption are similar for a low-fat, high-carbohydrate and a high-fat, low-­ carbohydrate diet.73,74 The average daily stomal losses of electrolytes, minerals, and trace elements in severe SBS are listed in Table 103-2.22,70,73,75,76 Water-soluble vitamins, except for vitamin B12, are absorbed in the proximal jejunum, as are macronutrients. It is unusual for deficiencies of these vitamins to develop, except in patients with a proximal jejunostomy or duodenostomy, and these patients invariably require TPN with vitamin supplementation. Loss of the ileum results in bile acid and vitamin B12 malabsorption, but patients who suffer such a loss as well as those with extensive jejunal resection also are at risk for more pronounced malabsorption of nutrients, minerals, vitamins, and electrolytes and water than is seen in patients with limited ileal resections, because of the greater loss of absorptive surface area and rapid intestinal transit. Fat-soluble vitamin (A, D, E, K) deficiency is encountered much more commonly as a result of concurrent fat malabsorption than as a result of loss of absorptive surface area. Most of the human daily vitamin K requirement is synthesized by colonic bacterial flora,77 so patients with any residual colon are a lower risk for developing vitamin K deficiency than are those whose colon has been resected. Conversely, patients who have received broad-spectrum antibiotics also are at risk for vitamin K deficiency. Zinc and selenium are lost in significant concentrations in the feces. The concentration of zinc is 12 mg/L in small bowel effluent and 16 mg/L in stool.78 Oral supplementation of vitamins, minerals, and trace elements generally is required for patients who do not require TPN (Table 103-5). Malabsorption of medication also occurs in patients with SBS.79,80 Many medications are absorbed in the jejunum, but medication malabsorption still can occur in patients who have undergone ileal resections alone because their intestinal transit time is decreased. The loss of the ileocecal valve increases the risk of bacterial overgrowth in the small intestine, which can worsen nutrient absorption and make management more difficult. The ultimate goal is to ensure a stable condition in which all nutritional needs are met, preferably by oral intake alone. In a series of 38 patients with a jejunum shorter than 200 cm and in continuity with the colon, all those with a jejunal length of more than 100 cm could be managed on oral intake alone.19

1785

1786

Section X  Small and Large Intestine Table 103-5  Vitamin and Mineral Requirements* in Patients with Short Bowel Syndrome MICRONUTRIENT

REQUIREMENT

Vitamin A Vitamin B12

10,000-50,000 units daily* 1000 µg subcutaneously monthly for patients with terminal ileal resection or disease 200 mg daily 50,000 units 1,25(OH2)-D3 twice weekly to twice daily 30 International Units daily 10 mg weekly 1000-1500 mg daily See text As needed 60-150 µg daily 220-440 mg daily (sulfate or gluconate form) As needed

Vitamin C Vitamin D Vitamin E Vitamin K Calcium Magnesium Iron Selenium Zinc Bicarbonate

Note: The table lists rough guidelines only. Vitamin and mineral supplementation must be monitored routinely and tailored to the individual patient, because relative absorption and requirements can vary. Supplements may be taken orally unless otherwise indicated. *Use cautiously in patients with cholestatic liver disease, because of the potential for liver toxicity.

In the late postoperative phase, the liquid diet is replaced by solid food, and the absorptive capacity of the remaining intestine is assessed again by measurement of fecal fat, volume, weight, and electrolytes while the patient is on a known nutrient and liquid intake. Fat absorption is, in general, more compromised than are nitrogen and carbohydrate absorption in SBS patients. The optimal diet composition for patients with SBS has been debated, but a low-fat, high-carbohydrate diet is of documented advantage in patients whose colon is in continuity with the remaining small intestine. Nitrogen is the least affected nutrient in SBS. Because absorption of dietary protein in the form of dipeptides and tripeptides occurs in the very proximal bowel, patients with only short segments of residual jejunum can benefit from the use of hydrolyzed protein or free amino acid-based enteral formulas. McIntyre and colleagues compared energy, nitrogen, and fat absorption in seven patients, each of whom had an end-jejunostomy, when they were provided with either a polymeric formula or a peptide-based formula. The length of residual jejunum in these patients ranged between 6 and 150 cm; no differences in nutrient absorption were observed.73 Similar uncontrolled observations were reported by Levy and coworkers.81 Contrary to these results, however, were those of Cosnes and associates, who reported modest improvement in six end-jejunostomy patients (mean residual small bowel length, 90 to 150 cm) who received a peptide-based diet, although energy absorption was unaffected.82 The patient populations in these studies were somewhat heterogeneous, and the peptide chain length and relative concentrations varied among formulas, making it difficult to compare the two studies. Lactose malabsorption due to substantial loss of jejunal length can worsen diarrhea, but a study of 14 patients with SBS on either a lactose-free diet or a diet with 20 g of lactose per day showed no significant dif­ ferences in stool volumes.83 Patients with SBS whose colon is in continuity should receive an oxalate-restricted diet

(see “Calcium Oxalate Kidney Stones” under “Complications”) (Table 103-6).

HOME PARENTERAL NUTRITION For the patient who requires long-term TPN, infusions typically are given on a continuous basis in the hospital until postoperative recovery has progressed and fluid needs and other metabolic issues have stabilized. Patients should be encouraged to adopt a hyperphagic diet while TPN volume and nutrient support are adjusted to maintain reasonable weight, fluid status, and nutrient sufficiency. As a patient gains weight or retains additional fluid, TPN fluid volume and nutrient composition can be decreased. It is important not to suppress the hypothalamic hunger center. Even if patients are unable to increase their oral intake significantly, they still should be encouraged to eat, to stimulate normal gallbladder contraction (if the gallbladder is in situ) and to prevent biliary complications (see later on). To prepare the patient for home TPN, the TPN regimen should be compressed gradually in two- to four-hour daily increments so that the total volume can be infused over a 10- to 12-hour period, typically overnight. Some patients with hyperglycemia or renal or congestive heart failure require a more prolonged infusion. The TPN infusion generally is tapered off over a 30- to 60-minute period to avoid hypoglycemia. Patients with a proximal jejunostomy might require additional fluid before or following completion of home TPN and, in some cases, during the day as well. TPN solutions are hypertonic and therefore must be infused into a central vein, such as the superior or inferior vena cava, through a tunneled catheter, to decrease the risks of infection and thrombosis.84,85 Percutaneously inserted central catheters (PICCs) should be reserved for short-term use (less than six months). For the patient to qualify for Medicare benefits, home TPN must be required for at least three months, and fat malabsorption as well as failed enteral nutritional support must be documented. The patient should be instructed about the indications for TPN, appropriate catheter care and dressing changes, the pump, preparation of TPN solutions, and acute complications of TPN, including air embolism, hypoglycemia, and catheter-related infections. Instruction on glucose selfmonitoring also should be included if hyperglycemia has been a problem or if insulin is required. The patient will need to add multivitamins, insulin, and possibly other additives to the TPN solution each night, because these other medications are not sufficiently stable to be added by the home TPN pharmacy. TPN solution typically is delivered in one- to two-week batches, so the patient will need a refrigerator dedicated to TPN. The patient’s home environment should be assessed. An appropriate location for setting up the TPN infusion and storing supplies, catheter cleaning, and hookup devices should be identified. This should not be a contaminated area such as a bathroom or kitchen. Patients often find it helpful to contact a local support group of the Oley Foundation (1-800-776-OLEY or www. oley.org). This independent, nonprofit organization includes patients and their families, as well as health care providers, and provides information, outreach services, emotional support, and conference activities. Physicians caring for patients on a home TPN regimen also should be familiar with TPN- and catheter-related complications and their recognition and treatment. These topics are beyond the scope of this chapter but have been reviewed in Chapter 5 and elsewhere (Figs. 103-6 and 103-7).86,87

Chapter 103  Short Bowel Syndrome Table 103-6  Dietary Recommendations for Patients Who Require an Oxalate-Restricted Diet FOODS CLASSIFIED BY OXALATE CONTENT Little* or None (<2 mg per Serving) Eat as Desired Beverages Apple or pineapple juice Bottled beer Coffee Colas (12 oz. limit/day) Distilled alcohol Milk, yogurt Orange juice (4 oz.) Tap water Wine (red, rosé) Meats, Fish Lean lamb, beef, pork Poultry Seafood Vegetables Asparagus Avocado Brussels sprouts Cabbage Cauliflower Mushrooms Onions Potatoes Radishes Sweet corn

Fruits Bananas Cherries, Bing Grapefruit Grapes, white Mangos Melons Nectarines Pears Pineapple Plums, green/golden Bread, Pasta, Cereal Macaroni Noodles Oatmeal Rice Spaghetti White bread Miscellaneous Butter Cheese, cheddar Eggs Jelly or preserves (made with allowed fruits) Mayonnaise Salad dressing Soups (made with allowed ingredients) Sugar Vegetable oils *For a low-oxalate diet, restrict to 40-50 mg daily.

Moderate (2-10 mg per Serving) Limit: Two 1/2-cup Servings per day

High (>10 mg per Serving) Avoid Completely

Cranberry juice (4 oz.) Grape juice (4 oz.) Nescafe powder

Cocoa Draft beer Juices containing berries Lemonade or limeade Tea Tomato juice

Sardines

Broccoli Cauliflower Cucumber Eggplant Green peas Lettuce Lima beans Squash Tomato, 1 small Turnips Vegetable soup

Beans Beets Carrots Celery Chives Collards Dandelion greens Endive Escarole French fried potatoes Kale Leeks Okra Parsnips Sweet potato Swiss chard

Apples Apricots Black currants Cherries, red sour Fruit cocktail Orange Peaches Plums, red Prunes

Berries Concord grapes Red currants Tangerines

Cornflakes Spaghetti, canned in tomato sauce Sponge cake

Bran cereal Grits, white corn Soybean crackers Wheat germ

Chicken noodle soup, dehydrated Fruitcake

Chocolate Nuts Peanut butter Pepper (>1 tsp/day) Pretzels Soybean curd (tofu)

1787

1788

Section X  Small and Large Intestine Fever (may occur only during infusion); chills when catheter is flushed or during infusion

Erythema of SQ tunnel tract

Blood cultures from the catheter and peripherally for bacteria and fungi CBC with peripheral blood smear (examine for bacteria, yeast)

Tunnel infection

Remove catheter

Exit site infection

Empiric treatment for MRSA with vancomycin or linezolid; adjust antibiotics based on C&S

Rx for coag. neg. Staph and Gram neg. organisms empirically; adjust antibiotics based on C&S + hold TPN for 24 hrs

Empiric treatment for MSSA and MRSA with vancomycin or linezolid

IV antibiotics × 1–3 wks or antibiotic lock technique × 1 wk

Catheter may be replaced in a different site

Bacteria

Fever resolves in 72 hrs

No

Severe sepsis despite antimicrobial therapy

Remove catheter

Infection present

Fungus

Continue appropriate antibiotic(s)

Continue antibiotic(s)

Tenderness, erythema or purulence of exit site (or on dressing)

Remove catheter Ampho B 100–250 mg total

Rx for an additional 2 wks with IV antibiotics

Replace catheter once patient is afebrile × 48–72 hrs or blood culture is negative in the absence of fever

Infection cured

No further Rx Remove catheter

Replace catheter at a different site

Recurrence within 6 wks of antibiotic discontinuation (same organism)

Retreat Unsuccessful

Remove catheter

Figure 103-6.  Algorithm for the diagnosis and management of catheter-related infection. ampho B, amphotericin B; coag., coagulase; CBC, complete blood count; C&S, culture and sensitivity testing; MSSA, methicillin-sensitive Staph aureus; MRSA, methicillin-resistant Staph aureus; neg., negative; Rx, treatment; SQ, subcutaneous; Staph, Staphylococcus; S. aureus, Staphylococcus aureus; TPN, total parenteral nutrition.

Patients in whom the frequency of TPN infusions can be reduced to fewer than five nights per week should have their micronutrient status monitored two to three times yearly to detect deficiencies. At the clinical visit, particular attention should be paid to the catheter exit site for evidence of erythema, purulent discharge, warmth, or tenderness. A catheter may remain in place indefinitely if it is properly maintained.88

COMPLICATIONS GALLSTONES

Interruption of the enterohepatic circulation of bile acids by ileal resection results in decreased hepatic bile acid secretion and altered composition of hepatic bile in terms of its organic components: bile acid, cholesterol, and phospholipids (see Chapters 64 and 72). Hepatic bile

Chapter 103  Short Bowel Syndrome Slowed rate of TPN and/or lipid emulsion infusion, resistance to flushing of catheter, leakage or swelling of the exit site

Empirically treat for possible thrombosis with tPA (2 mg in 2 mL)

Radiologic contrast study of the catheter to verify or exclude occlusion

Attempt aspiration through catheter in 60 min

Malposition

Attempt to adjust position with a pig-tail catheter (sterile technique by radiologist)

Successful

Unsuccessful

Consider minidose warfarin (1–2 mg by mouth daily)

Radiologic contrast study of the catheter

Successful

Unsuccessful

Thrombosis No further treatment

Replace catheter

tPA (2 mg in 2 mL)

Unsuccessful

Occlusion, but no thrombosis

Remove and replace catheter

HCI (0.1N) or ethanol (70%)

NaOH (0.1N)

Attempt aspiration through catheter after 1 hr

Attempt aspiration through catheter after 6–7 hrs

Successful

Unsuccessful

Review TPN formula for incompatibilities

Replace catheter

Successful

Unsuccessful

Figure 103-7.  Algorithm for the diagnosis and management of thrombotic or nonthrombotic catheter-related occlusion. tPA, tissue plasminogen activator; TPN, total parenteral nutrition.

becomes supersaturated with cholesterol, with subsequent formation of cholesterol crystals and gallstones in gallbladder bile (see Chapter 65). Most gallstones in patients with SBS, however, are composed of calcium bilirubinate; the pathophysiology is unclear. A prevalence of 44% of asymptomatic gallstones was documented in a study of 84 patients with severe SBS who required TPN.19 Formation of biliary sludge and gallbladder hypomotility probably

contribute to the high prevalence of these stones, because many of these patients are on long-term parenteral nutrition.89 Postprandial CCK concentration is decreased in some patients with SBS,90 and injections of CCK have been used experimentally to induce gallbladder contraction, although this therapy is not always successful and results in nausea, vomiting, and abdominal pain in some patients.91,92

1789

1790

Section X  Small and Large Intestine LIVER DISEASE

Liver disease often develops in patients who require longterm TPN. After five years of TPN, more than 50% of these patients will be found to have severe liver disease defined as grade 2 fibrosis, cirrhosis, or one of the following: total serum bilirubin greater than 3.5 mg/dL for longer than one month, ascites, portal hypertension, hepatic encephalopathy, or liver failure with a factor V concentration less than 50% of normal.93 Liver failure develops in approximately 15% of all TPN-dependent patients.94 The incidence, prevalence, and severity of liver disease in young children and infants, in particular, are much greater than in adults.95 The incidence and prevalence of liver disease and liver failure specifically in patients with SBS requiring TPN are unknown. Although these disorders often are referred to as TPN-associated liver disease, the pathogenesis probably is related to malabsorption of nutrients such as choline96 and to the route of nutrient assimilation, namely, through the central axis, rather than the portal circulation.97 Patients with the least amount of residual intestine are at greatest risk for developing liver disease.98,99 Diagnosis of liver disease related to intestinal failure in the patient with SBS requires the exclusion of other potential causative disorders. SBS-related liver disease can manifest as cholestasis, steatosis, or steatohepatitis; cholestasis is more common in infants. Studies have suggested benefit from oral lecithin, although it is poorly absorbed, and from intravenous choline (investigational) and, to a lesser extent, ursodeoxycholic acid.100-105 A recent uncontrolled case series of 18 infants described the substitution of a fish oil-based lipid emulsion for the conventional long-chain triglyceride-based emulsion. Compared with historic controls from a different era, reversal of cholestasis occurred sooner (9.4 vs. 44.1 weeks in the fish oil-supplemented group.106 Dextrose overfeeding (greater than 40 kcal/kg/ day) and excessive fat emulsion infusion (2.5 g/kg/day, possibly only 1.0 g/kg/day) should be avoided.93 A minimum of 2% to 4% of total calories, however, should be provided as linoleic fatty acid (50% of most lipid emulsions), to prevent essential fatty acid deficiency. Carnitine supplementation is not useful.107

CALCIUM OXALATE KIDNEY STONES

Fat malabsorption secondary to bile acid deficiency in patients with extensive ileal resection is associated with an increased risk of oxalate kidney stones if the colon is preserved. Oxalate in food usually precipitates as calcium oxalate in the intestinal lumen and is lost in the stool. Lipolysis in patients with SBS and fat malabsorption is normal, and unabsorbed long-chain fatty acids compete with oxalate for available luminal calcium. Consequently, a larger amount of free oxalate is lost to the colon, where it is absorbed and ultimately excreted by the kidney (Fig. 103-8), manifesting as just hyperoxaluria or with calcium oxalate stone formation. Patients with SBS who do not have a colon in continuity are not at increased risk. In one study, symptomatic kidney stones developed within two years of enterectomy in 9 of 38 patients (24%) with SBS and an intact colon.19 Urinary oxalate excretion should be monitored regularly in these patients. Treatment of hyperoxaluria consists of restriction of oxalate-containing food products (see Table 103-6). If hyperoxaluria persists, then oral administration of calcium citrate should be tried; the extra calcium precipitates dietary oxalate, and the citrate prevents stone growth in the urine. A single case report describes the use of conjugated bile acid supplementation to reduce hyperoxaluria.108 Hyperoxaluria also may be

Normal

Steatorrhea +2 LCFA Ca

Ca+2 Oxalate

Insoluble

Oxalate

Insoluble

Excreted Excreted Figure 103-8.  Mechanism of oxalate hyperabsorption in patients with steatorrhea. Normally, oxalate in food is precipitated as calcium oxalate in the intestinal lumen and lost in the stool (left). Lipolysis is normal in patients with short bowel syndrome with fat malabsorption (right), and unabsorbed long-chain fatty acids (LCFA) compete with oxalate for the available calcium. Consequently, a larger amount of free oxalate passes from the small intestine to the colon, where it is absorbed and ultimately excreted by the kidney, often resulting in hyperoxaluria.

related to the metabolism of the vitamin C in TPN solution in the presence of light.109 D -LACTIC d-Lactic

ACIDOSIS

acidosis is a rare complication of SBS and in this setting is observed only in patients with a preserved colon. The episodes of acidosis usually are precipitated by increased oral intake of refined carbohydrates and can be induced in the patient with SBS by carbohydrate overfeeding.110 Malabsorbed carbohydrate is metabolized by colonic bacteria to SCFAs and lactate, which lower the intracolonic pH. A lower pH inhibits the growth of the predominant Bacteroides species and promotes the growth of acidresistant, Gram-positive anaerobes (Bifidobacterium, Lactobacillus, and Eubacterium), which have the capacity to produce d-lactate. d-Lactate is absorbed from the colon and is metabolized to only a limited extent in humans because of our lack of d-lactate dehydrogenase. The main excretory route for d-lactate is the kidney.111 Absorbed d-lactate results in the development of a metabolic acidosis and characteristic neurologic signs and symptoms of nystagmus, ophthalmoplegia, ataxia, confusion, and inappropriate behavior. Patients with d-lactate acidosis often are suspected of being inebriated, but their blood alcohol levels are normal. The constellation of specific neurologic symptoms and metabolic acidosis in a patient with SBS should raise the suspicion of possible d-lactic acidosis. Blood tests will confirm a metabolic acidosis and a normal lactate level; however, the clinical laboratory should be notified to quantify the d-lactic acid rather than the l-lactic acid concentration. The diagnosis is confirmed by measurement of whole-blood d-lactate concentration, which will be elevated significantly (to greater than 3 mmol/L, compared with the normal level of less than 0.5 mmol/L). Treatment consists of correcting the acidosis with sodium bicarbonate and stopping oral intake, which usually results in rapid abatement of the neurologic symptoms. The poten-

Chapter 103  Short Bowel Syndrome tial benefit of antibiotic treatment to change the colonic flora is debated. Substitution of refined carbohydrates for starch has prevented recurrent d-lactic acidosis in a few patients.112 The mediator of the neurologic symptoms still is unknown, and infusion of d-lactic acid in normal subjects to achieve blood levels commonly observed in patients with d-lactic acidosis does not cause any neurologic symptoms. The neurologic symptoms have a striking resemblance to those of Wernicke’s encephalopathy, and in one patient with SBS, recurrent d-lactic acidosis was prevented by thiamine supplementation.113

intestinal transplantation, and it should be performed only in centers with significant experience in this area. To date, no studies have been conducted to compare medical and surgical therapies. A less complex procedure, the serial transverse enteroplasty (STEP), developed by Kim, is a novel technique during which a linear surgical stapler is applied from alternating and opposite directions along the intestine’s mesenteric border to incompletely staple and divide the dilated intestine (Fig. 103-11).119 This procedure leads to the tapering of the intestine in a zig-zag pattern, which results in nutrients being channeled along a narrower but longer intestine. Rather than an intestinal lengthening procedure, this technique is better described as an intestinal tapering procedure. Results reported from an international registry comprising 38 patients from 19 centers have indicated the procedure increases intestinal length by almost 50%, and it has resulted in close to a 100% increase in nutrient absorption.120,121 These improvements in some patients, however, may have been due in part to increased segmental absorption, observed as part of the natural postenterectomy adap­tation process. Nevertheless, the tapering of a dilated, essentially nonfunctional loop of bowel might decrease bacterial overgrowth and improve nutrient absorption. Furthermore, the STEP might have an advantage over the Bianchi procedure in addition to being less technically demanding, in that by avoiding intestinal transection, it may be easier to preserve the blood supply of the intestine.

OTHER COMPLICATIONS

Renal dysfunction,114 metabolic bone disease,115 memory deficits,116 and neurologic abnormalities117 all have been described in patients with SBS who require long-term TPN.

SURGICAL MANAGEMENT INTESTINAL LENGTHENING PROCEDURES

The most important surgical procedure is reanastomosis of the residual small bowel to the residual colon. This procedure carries relatively low mortality and morbidity rates and allows enhanced energy absorption from SCFAs produced from the bacterial fermentation of unabsorbed carbohydrate. A number of other surgical procedures, such as tapering enteroplasty, construction of intestinal valves, creation of recirculating loops, reversal of a short intestinal segment, or colonic interposition, have been attempted to increase intestinal transit time. These procedures are considered experimental, the experience with each is limited, and outcomes generally are not optimal.1 Longitudinal intestinal lengthening and tailoring (Bianchi procedure) (Fig. 103-9) may be useful in patients who have segmental dilation and nonfunctional intestine due to dysmotility and bacterial overgrowth. In this procedure, the surgeon divides the dilated bowel, creates two hemiloops, and anastomoses the hemiloops in an endto-end fashion, thereby doubling the bowel length (Fig. 103-10).118 Although the surface area is not truly increased, bowel function can improve, allowing reduction or elimination of TPN. Nearly all of the approximately 100 operations reported have been undertaken in children. This procedure should be attempted only as a last resort before

INTESTINAL TRANSPLANTATION

Intestinal transplantation is being performed in an increasing number of centers worldwide. The main indication for transplantation in children and adults is TPN-dependent SBS complicated by progressive liver disease. Combined intestine-liver transplantation is the only alternative for patients in whom end-stage liver disease has developed. Isolated intestinal transplantation may be considered for patients with clinically significant liver disease that has not yet progressed to cirrhosis.122 Patients who have significant fluid losses and who experience frequent episodes of severe dehydration despite appropriate medical management also may be candidates for isolated intestinal transplantation. Medicare has approved other indications, including twomajor-vessel thrombosis, a single episode of fungemia, a single episode of bacterial sepsis with shock, and two lifetime episodes of catheter sepsis, although the preponder-

Y

X

Y

X

A

B

C

Figure 103-9.  The Bianchi procedure for intestinal lengthening depicted in a schematic diagram. A, The bowel is split lengthwise. B, Two hemiloops (X and Y) are created. C, The hemiloops (X and Y) are anastomosed end to end.

1791

1792

Section X  Small and Large Intestine

B A

D

C

Figure 103-10.  The Bianchi procedure shown in intraoperative photographs. A, The tips of the forceps are within the dilated loop of intestine, which has been opened; the beginning of each hemiloop is evident (right side). B, A blood vessel can be seen going to the left hemiloop. C, Completed anastomosis. D, The suture shows the gain in the length of intestine. The first hemiloop extends from the tip of the forceps to the first perpendicular suture line. The distance from that point to the end of the thread represents the gain in intestinal length (approximately 26 cm in this infant). (Photographs kindly provided by Kishore R. Iyer, MD, New York, NY.)

Table 103-7  Patient and Graft Survival Rates (%) for Transplants Performed for Short Bowel Syndrome from January 1988 to March 30, 2007, in the United States 1 Year TRANSPLANT TYPE Isolated intestine (n = 510) Intestine/liver (n = 405) Multivisceral (n = 454)

3 Years

5 Years

10 Years

Patient

Graft

Patient

Graft

Patient

Graft

Patient

Graft

83.2 63.6 71.1

78.4 61.3 67.8

65.8 52.4 54.3

55.7 49.8 54.3

55.0 47.4 58.2

43.2 44.4 47.9

43.6 39.2 52.7

24.1 36.0 NA

Based on Organ Procurement and Transplantation Network (http://optn.transplant.hrsa.gov/) data as of June 18, 2008. This work was supported in part by Health Resources and Services Administration contract 231-00-0115. NA, not available.

ance of evidence does not support these as appropriate indications for transplantation. Survival has improved considerably since intestinal transplantation was initiated, with reported survival and nutritional autonomy of up to 18 years.123 As of May 31, 2007 (most recent data from the International Small Bowel Transplant Registry), 1720 transplantation procedures had been performed worldwide in 1608 patients, 909 of whom were still alive. This experience included 746 isolated intestine, 594 intestine-liver, and 380 multivisceral transplants. Patients who have undergone transplantation more recently, generally have had better survival because of improved technique and optimized immunosuppressive regimens. Mean hospitalization was 56 days for isolated intestine, 76 days for intestine-liver, and 62 days for multivisceral transplant recipients. Additional information can be found on the International Intestinal Transplant Registry website,

http://intestinaltransplant.org, which is updated every 2 years (last time in 2007). Current (2008) patient and graft survival data for the United States are presented in Table 103-7. The mortality rate for patients waiting for an intestinalliver transplant is significantly greater than for those waiting for an isolated liver transplant.1 Therefore, early referral to an intestinal transplantation center at the first sign of liver disease is recommended, even if a transplant does not ultimately become necessary. Intestinal and multiorgan transplantations are expensive and generally cost between $250,000 and $3 million per case. Post-transplantation complications, and the most common causes of death afterward, include acute rejection, chronic rejection, cytomegalovirus infection, sepsis (often complicating rejection), and post-transplantation lymphoproliferative disease (PTLD).124 Antirejection medications amount to another $10,000

Chapter 103  Short Bowel Syndrome 12%, and MELD scores were significantly elevated up to 180 days before death, although they were less reliable when obtained 90 days or more before death in a group of 133 patients with intestinal failure; mortality was 50% at 328 days patients with a MELD score between 15 and 25. Increased CRP was also an independent predictor of death within 90 days before death, and a CRP of four or greater universally predicted mortality.126 One unit increase in CRP was associated with a 20% increased risk of immediate death. Further evaluation of such predictors of poor outcome will be necessary, however, before they can be used reliably to support early intestinal transplantation.

PHARMACOLOGIC ENHANCEMENT OF BOWEL ADAPTATION

A

a

b

B

c

Figure 103-11.  The serial transverse enteroplasty (STEP) procedure shown intraoperatively (A) and in line drawings (B). STEP is a technique in which a linear surgical stapler is applied from alternating and opposite directions along the intestine’s mesenteric border to incompletely staple and divide the dilated intestine. The arrows in a and b point to an invagination produced by the staples. The configuration after recovery is shown in c. (Photograph kindly provided by Kishore R. Iyer, MD, New York, NY.)

yearly, in addition to repeated hospitalizations for infection and rejection. For patients who do well, however, nearly all are successfully weaned from TPN, although a few require some maintenance intravenous fluids. This expense compares with a charge of $100,000 to $150,000 per year for home TPN, in addition to the costs of hospitalization for complications. The actual costs of TPN (including pharmacists’ time) however, are closer to $20 to $30 per day. Intestinal transplantation has reached a stage at which it is a feasible, but not yet practical, alternative to conservative treatment of the patient with SBS. One of the greatest dilemmas facing intestinal transplantation is balancing the avoidance of premature transplantation with late referral for transplantation; the latter often requires addition of a liver graft and often results in a lessoptimal outcome.125 High-risk patients likely to develop complications on home TPN need to be identified early, and every attempt must be made to enhance nutrient and fluid absorption and decrease the need for TPN. Both the Model for End-stage Liver Disease (MELD) score and C-reactive protein (CRP) can be used to predict mortality in patients who require long-term TPN. In a study by Putchakayala and colleagues,126 each point of increase in the MELD (to more than 15) was associated with an increased death risk of

The growing knowledge about growth factors has stimulated several clinical studies in patients with SBS. The promising results with the use of growth hormone and dietary lglutamine in a large uncontrolled study of TPN-dependent patients with SBS127 raised hopes that intestinal mucosal growth could be enhanced beyond the adaptive period.6 Two placebo-controlled studies of identical growth hormone and l-glutamine supplementation failed to show any beneficial effect on absorption,46,47 however, and two other studies showed only marginal improvements in fluid and nutrient retention.128,129 GLP-2 is an intestinotrophic enteric hormone that initially was evaluated in a small uncontrolled study of eight patients with SBS, who received native GLP-2 400 mg subcutaneously twice a day for 35 days.130 The treatment resulted in an increase in several absorptive parameters, body weight, and mucosal growth. Use of a synthetic analog of GLP-2 (teduglutide) was associated with increased villus height and increased fluid absorption, with more modest improvements in energy and nitrogen absorption, that regressed once the medication was discontinued.131 A double-blind, randomized, multicenter study indicated administration of the GLP-II analog resulted in a significant decrease in the requirement for parenteral nutrition, although only a few subjects were able to be fully weaned.132 Increased fluid retention was associated with less chronic dehydration,133 a primary factor in the development of nephropathy in patients who require long-term TPN.134 The rapid advance in our knowledge of epithelial growth factors undoubtedly will lead to discovery of still other growth factors that can stimulate intestinal epithelial growth and thus benefit these patients. A double-blind, randomized, controlled trial of growth hormone (0.1 mg/kg/day for 4 weeks) in 41 TPN-dependent patients showed that TPN requirements in treated patients could be reduced by an additional 2 L per week (or one night weekly) over the reduction with standard therapy described earlier in this chapter.135 It is unclear whether these effects were related to improved absorption or appetite stimulation. This study led to the FDA approval of growth hormone injections for treating TPN-dependent SBS. The benefit from this therapy lasted nearly four months following completion of three weeks of daily growth hormone injections; it is unclear whether booster injections will be required. The benefits of this therapy must be weighed against the potential side effects, which include fluid retention, edema, arthralgias, and carpal tunnel syndrome; it also is unknown whether any of the potential growth factor therapies would be more effective if administered during the adaptive phase following enterectomy.

1793

1794

Section X  Small and Large Intestine Jejunoileal anastomosis, no colon present

Jejunoileal anastomosis

End jejunostomy

Re-anastomosis of colon, if possible

≥100 cm residual small intestine

<100 cm residual small intestine

>100 cm small intestine + colon

<100 cm small intestine + colon or <150 cm small intestine + partial colon

Partial TPN support with weaning

TPN generally not required

TPN likely

Oral rehydration solutions Electrolyte replacement/monitoring Fluid management H2 blocker or PPI for the 1st 6 mos Antidiarrheal therapy High-complexcarbohydrate diet Low oxalate diet Oral calcium supplements MCT

Oral rehydration solutions Electrolyte replacement/monitoring Fluid management H2 blocker or PPl for the 1st 6 mos Antidiarrheal therapy Oral calcium supplements Cycle TPN at home Teach appropriate catheter care

Oral rehydration solutions Electrolyte replacement/monitoring Fluid management H2 blocker or PPI for the 1st 6 mos Antidiarrheal therapy High-complex-carbohydrate diet Low oxalate diet Oral calcium supplements MCT Attempt TPN weaning Teach appropriate catheter care Cycle TPN at home

Continue above measures Monitor vitamin and mineral status Yearly bone densitometry

Monitor for and manage complications Bone

Hepatic

Avoid dextrose overfeeding Use <2.5 g/kg/d lipid, and if possible <1.5 g/kg/d Diagnostic testing

Improvement

Infectious

Biliary

Monitor for signs of catheter infection; manage as per Fig. 103-6

Encourage oral food intake to avoid gallbladder stasis Cholecystectomy for cholecystitis

No improvement

Monitor calcium and vitamin D status Consider IV bisphosphonate therapy

Catheter occlusion Live donor available?

Yes Intestinal/ liver transplant

Renal

Live donor intestinal transplant

See Figure 103-7 No

Cadaveric intestinal transplant

ESLD: end stage liver disease Figure 103-12.  Algorithm for management of the patient with short bowel syndrome. ESLD, end-stage liver disease; MCT, medium-chain triglycerides; PPI, proton pump inhibitor; tPA, tissue plasminogen activator; TPN, total parenteral nutrition.

Chapter 103  Short Bowel Syndrome SURVIVAL AND QUALITY OF LIFE The prognosis for patients with SBS is determined primarily by the type and extent of intestinal resection and by the underlying disease. Patients with limited small intestinal resections in general have an excellent prognosis with careful management of their specific malabsorptive defects. Patients with high jejunostomies and severe malabsorption present difficult management problems, and their long-term care poses a challenge for surgeons, gastroenterologists, and dietitians. The rate of survival, prognosis, and quality of life are, however, steadily improving even in this group of patients because of increasing experience with long-term parenteral nutrition and better methods to assess nutritional needs. The probability of survival and TPN dependence has been assessed in a prospective study of 124 patients with SBS.72 Most of these patients had intestinal resection for either mesenteric infarction or radiation enteritis. The probability of survival was 86% at 2 years and 75% at 5 years. TPN dependence rates were 49% at two years and 45% at five years, suggesting that most patients requiring long-term TPN can be weaned successfully within two years using conventional techniques. In a multivariate analysis, survival was related negatively to high jejunostomy, small bowel length less than 50 cm, and mesenteric infarction as a cause for intestinal resection. TPN dependence was related primarily to small bowel length. Remnant bowel length less than 100 cm was highly predictive of permanent intestinal failure and lifelong TPN dependence. Similar results were reported in a study of 225 patients from the Mayo Clinic.136 Most patients with SBS have a good quality of life and can work full time. Figure 103-12 depicts an algorithm for managing the patient with SBS.

KEY REFERENCES

Buchman AL. Complications of long-term home total parenteral nutrition: Their identification, prevention and treatment. Dig Dis Sci 2001; 46:1-18. (Ref 86.) Buchman AL. Practical Nutrition Support Techniques. Thorofare, NJ: Slack; 2003. (Ref 87.)

Buchman AL, Iyer K, Fryer J. Parenteral nutrition–associated liver disease and the role for isolated intestine and intestine/liver transplantation. Hepatology 2006; 43:9-19. (Ref 96.) Buchman AL, Moukarzel A. Metabolic bone disease associated with total parenteral nutrition. Clin Nutr 2000; 19:217-31. (Ref 115.) Buchman AL, Scolapio J, Fryer J. AGA technical review on SBS and intestinal transplantation. Gastroenterology 2003; 124:1111-34. (Ref 1.) Buchman AL, Sohel M, Dubin M, et al. Choline deficiency causes reversible hepatic abnormalities in patients during parenteral nutrition: Proof of a human choline requirement; a placebo-controlled trial. JPEN J Parenter Enteral Nutr 2001; 25:260-8. (Ref 102.) Cavicchi M, Beau P, Crenn P, et al. Prevalence of liver disease and contributing factors in patients receiving home parenteral nutrition for permanent intestinal failure. Ann Intern Med 2000; 132:525-32. (Ref 93.) Fryer JP. Intestinal transplantation: Current status. Gastroenterol Clin N Am 2007; 36:145-59. (Ref 125.) Jeppesen PB, Hartmann B, Hansen BS, et al. Impaired meal stimulated glucagon-like peptide 2 response in ileal resected short bowel patients with intestinal failure. Gut 1999; 45:559-63. (Ref 31.) Messing B, Crenn P, Beau P, et al. Long-term survival and parenteral nutrition dependence in adult patients with SBS. Gastroenterology 1999; 117:1043-50. (Ref 2.) Nordgaard I, Hansen BS, Mortensen PB. Importance of colonic support for energy absorption as small-bowel failure proceeds. Am J Clin Nutr 1996; 64:222-31. (Ref 13.) Putchakayala K, Polensky S, Fitzhugh J, et al. An evaluation of the model for end-stage liver disease and serum C-reactive protein as prognostic markers in intestinal failure patients on parenteral nutrition. JPEN J Parenter Enteral Nutr 2009; 33(1):55-61. (Ref 126.) Roslyn JJ, Pitt HA, Mann LL, et al. Gallbladder disease in patients on long-term parenteral nutrition. Gastroenterology 1983; 84:148-54. (Ref 89.) Scolapio JS, Camilleri M, Fleming CR, et al. Effect of growth hormone, glutamine, and diet on adaptation in short-bowel syndrome: A randomized, controlled trial. Gastroenterology 1997; 113:1074-81. (Ref 46.) Solhaug JH, Tvete S. Adaptive changes in the small intestine following bypass operation for obesity. Scand J Gastroenterol 1978; 13:401-8. (Ref 36.) Woolf GM, Miller C, Kurian R, Jeejeebhoy KN. Diet for patients with a short bowel: High fat or high carbohydrate? Gastroenterology 1983; 84:823-8. (Ref 69.) Full references for this chapter can be found on www.expertconsult.com.

1795

CHAPTE R

104 Celiac Disease and Refractory Celiac Disease Richard J. Farrell and Ciarán P. Kelly

CHAPTER OUTLINE Definitions  1797 History of Celiac Disease  1797 Epidemiology  1798 Pathology  1799 Pathogenesis  1800 Environmental Factors  1800 Genetic Factors  1802 Immune Factors  1802 Clinical Features  1804 Childhood Presentation  1804 Adulthood Presentation  1804 Gastrointestinal Features  1804 Extraintestinal Features  1805 Diagnosis  1807 Serology  1807 Genetic Testing  1809 Small Intestine Biopsy  1809 Gluten Challenge  1810

DEFINITIONS Celiac disease is characterized by small intestinal malabsorption of nutrients after the ingestion of wheat gluten or related proteins from rye and barley, villus atrophy of the small intestinal mucosa, prompt clinical and histologic improvement following strict adherence to a gluten-free diet, and clinical and histologic relapse when gluten is reintroduced.1 The many other names used to identify patients with this condition, including nontropical sprue, celiac syndrome, adult celiac disease, idiopathic steatorrhea, and primary malabsorption, among others, are testimony to the confusion of the past. The term celiac disease is recognized widely and is used in this chapter; celiac sprue and gluten-sensitive enteropathy are acceptable alternative terms. Celiac disease exhibits a spectrum of clinical presentations (Fig. 104-1). Atypical celiac disease is fully expressed gluten-sensitive enteropathy manifest only by extraintestinal symptoms and signs including short stature, anemia, osteoporosis, and infertility. Silent celiac disease is fully expressed gluten-sensitive enteropathy usually found after serologic screening in asymptomatic patients. The atypical and silent variants are more common than classic or typical celiac disease, which is fully expressed gluten-sensitive enteropathy found in association with the classic gastrointestinal symptoms of malabsorption. A combination of serologic, genetic, and histologic data also has led to the identification of two other types of celiac disease. Patients with latent celiac disease have normal

Stool Examination  1810 Hematology and Biochemistry Tests  1810 Radiology  1811 Differential Diagnosis  1811 Diseases Associated with Celiac Disease  1811 Treatment  1813 Gluten-Free Diet  1813 Dietary Supplements  1814 Glucocorticoids  1815 Nonresponsive Celiac Disease  1815 Refractory Celiac Disease  1816 Ulcerative Jejunoileitis  1817 Collagenous Sprue  1817 Treatment  1817 Complications  1818 Celiac Disease and Malignancy  1818 Prognosis  1819 Future Therapies  1819

villus architecture on a gluten-containing diet but, at another time, have had or will have gluten-sensitive villus atrophy. For example, a patient who had celiac disease in childhood and recovered completely on a gluten-free diet might have latent celiac disease later in life on resumption of a normal diet. Patients with potential celiac disease have never had a biopsy consistent with celiac disease but show immunologic abnormalities characteristic for the disease, such as a positive immunoglobulin (Ig)A antibody to endomysium (or tissue transglutaminase [tTG]) or increased intraepithelial lymphocytes (IELs) in the small intestine. These patients often have a genetic predisposition to celiac disease, especially human leukocyte antigen class II DQ (HLA-DQ2), an affected first-degree relative, or both. The probability of their eventually developing celiac disease is unpredictable.2 Refractory celiac disease, also known as unclassified or intractable celiac sprue, is defined as symptomatic, severe small intestinal villus atrophy that mimics celiac disease but does not respond to at least six months of a strict glutenfree diet. This is a diagnosis of exclusion that is not accounted for by inadvertent gluten ingestion, other causes of villus atrophy, or overt intestinal lymphoma.1,3

HISTORY OF CELIAC DISEASE Celiac disease was recognized as a clinical entity by Aretaeus the Cappadocian in the first century ad.4 The name

1797

1798

Section X  Small and Large Intestine

IgA EMA or tTG antibody positive

Intestinal mucosal abnormality

Classic celiac disease Atypical celiac disease Silent celiac disease Latent celiac disease

Genetic susceptibility HLA-DQ2 or -DQ8 Healthy individual

Figure 104-1.  The celiac iceberg and the spectrum of celiac disease. EMA, endomysial antibody; Ig, immunoglobulin; tTG, tissue transglutaminase.

sprue was coined in the 18th century and is derived from the Dutch word spruw, which means “aphthous disease,” so named because of the high prevalence of aphthous mouth ulcers in these patients. In 1888, Samuel Gee published his paper “On the Coeliac Affection,” which described many of the clinical features of celiac disease in patients of all age groups and concluded, “If the patient can be cured at all it must be by means of the diet.”5 It was not until the middle of the 20th century, however, that the link between certain cereals and celiac disease was made by Willem Karel Dicke, a Dutch pediatrician. He became convinced that the consumption of bread and wheat flour was directly responsible for the deterioration in patients suffering from this condition.6 During World War II, cereals used to make bread were particularly scarce in the Netherlands, and during this time, children with celiac disease improved, only to relapse after the supply of cereal was re-established at the end of the war. It was this serendipitous observation that led to the finding that wheat ingestion exacerbated celiac disease. Subsequent work by van de Kamer and coworkers showed that it was the water-insoluble portion, or gluten moiety, of wheat that produced intestinal injury in patients with celiac disease.7 In 1954, Paulley provided the first accurate description of the characteristic intestinal lesion in patients with celiac disease.8 With the development of effective peroral suction biopsy instruments in the late 1950s, Rubin and coworkers demonstrated that celiac disease in children and idiopathic or nontropical sprue in adults were identical diseases with the same clinical and pathologic features.9 Since the 1980s, we have seen substantial advances in our understanding of the genetic, immune, and molecular mechanisms fundamental to the pathogenesis of celiac disease. In 1986, Howell and associates observed that celiac disease was associated with specific HLA-DQ2 haplotypes.10 In 1993, Lundin and colleagues demonstrated that the DQ2 gene products preferentially present gluten-derived gliadin peptides to intestinal mucosal T cells in celiac patients.11 Subsequently, the enzyme tTG (more specifically tTG type 2 [tTG-2]) was identified as a celiac autoantigen, leading to more accurate serologic diagnostic tests.12 In 1998, Molberg and colleagues reported that modification of gliadin by host tTG enhances gliadin-specific celiac disease T-cell responses.13 The identification of specific tTG-modified deamidated gliadin peptides (DGPs) as dominant α-gliadin T-cell epitopes has highlighted the pivotal role played by tTG in the pathogenesis of celiac

disease.14 This discovery already has led to more accurate antigliadin antibody serologic testing using DGPs as capture antigens, and it might pave the way for antigen-specific immunotherapy. The key role played by IELs in the development of refractory celiac disease and enteropathy-associated T cell lymphoma (EATL) continues to evolve.15 Studies also point to the importance of interleukin (IL)-15, a potent proinflammatory cytokine at the interface between innate and adaptive immunity in the pathogenesis of both celiac disease and refractory celiac disease.16 Epidemiologic studies using endomysial antibody (EMA) and tTG serology have substantially increased estimates of celiac disease prevalence in the United States and elsewhere.17 This in turn has led to renewed interest in potential nondietary treatments including glutenases, modifiers of tight junction function, tTG inhibitors and immunebased interventions, bringing celiac disease therapy into a new era.18,19

EPIDEMIOLOGY The term celiac iceberg was coined to describe the wide variations in the nature and intensity of clinical presentation of which overt celiac disease is only the emerging peak (see Fig. 104-1). The discovery of the large immersed part of the celiac iceberg has transformed the status of celiac disease, long considered a rare disease, particularly in adults, to that of a common health problem. Because we are uncertain of the depth and breadth of the celiac iceberg, the true prevalence of celiac disease remains unknown. Serologic testing has demonstrated that silent celiac disease, characterized by positive serology and villus atrophy with few or no symptoms, is approximately seven times more common than symptomatic celiac disease.20 A Finnish study of 3654 schoolchildren of ages 7 to 16 years, using two serologic screens with antiendomysial and tTG antibodies, demonstrated the heterogeneity of the celiac iceberg, with one of every 99 children having biopsy-proved celiac disease.21 Only 10 of 56 subjects with a positive serology had overt symptoms of celiac disease. Two subjects with positive antibodies and at risk for celiac disease because of HLA-DQ2 haplotype had normal mucosa, but both had increased epithelial expression of HLA-DR suggesting mild intestinal inflammation, and one had high

Chapter 104  Celiac Disease and Refractory Celiac Disease counts of IELs; these patients might represent cases of potential disease susceptible to evolving into overt celiac disease. Five patients who had HLA-DQ2 and positive antibodies when studied in 1994 had negative antibodies on repeat testing in 2001; their intestinal biopsies were normal, but all had increased HLA-DR expression, and four of the five had markedly increased numbers of IELs. This latter finding might indicate a variation in the natural history of celiac disease, occasionally seen in teenagers, in whom gluten sensitivity fluctuates with time. Celiac disease shows a marked geographic variation, with the highest incidence in Western Europe. The condition is more common in Scandinavian and Celtic populations, where the prevalence has been reported to be as high as 1 in 9921 and 1 in 122,22 respectively. The prevalence is similarly high in Italy20 and the southeastern region of Austria.23 The prevalence in Denmark is 40-fold lower than that in Sweden,24 suggesting considerable variation in prevalence among geographically proximate populations. Factors such as predominant HLA haplotype, timing of introduction of gluten into the diet, differences in the gliadin concentration of infant formulas, and interobserver variation in interpreting small intestinal biopsy findings might explain the differences in prevalence.25 Celiac disease also is found in areas to which Europeans have emigrated, notably North America, South America, and Australia. Epidemiologic studies in the United States, where the disease only recently has attracted much attention, underscore the varying clinical presentation of celiac disease and indicate that the prevalence in the United States is comparable with that in Western Europe. A large multicenter study by Fasano and coworkers17 determined the prevalence of antiendomysial antibodies in more than 13,000 at-risk and not-at-risk American subjects and found the prevalence of antiendomysial antibodies to be 1 in 22 and 1 in 39 among first-degree and second-degree relatives of subjects with celiac disease, respectively.17 A prevalence of 1 in 56 was documented among patients with celiac-like gastrointe­ stinal symptoms or with associated disorders. Of most significance, these investigators found a prevalence of anti-

endomysial antibodies of 1 : 133 among 4126 “not-at-risk” subjects. Although celiac disease is rare in the predominantly riceeating area of southern India, it is prevalent in the Bengal and Punjab provinces of northwest India, where wheat rather than rice has, for many generations, has been a staple of the diet. The condition has been reported in blacks, Arabs, Hispanics, Israeli Jews, Sudanese of mixed Arabblack descent, and Cantonese and is particularly high among the Saharawi population in northwest Africa.26 The condition rarely affects people of purely sub-Saharan African, African-Caribbean, Chinese, or Japanese descent. Some authors have noted a female-to-male ratio of 2 : 1, whereas others have reported ratios as low as 1.3 : 1 but still suggesting a female predominance.

PATHOLOGY Celiac disease affects the mucosa of the small intestine; the submucosa, muscularis propria, and serosa usually are not involved. The mucosal lesion of the small intestine in celiac disease can vary considerably in severity and extent.9 This spectrum of pathologic involvement might contribute to the striking variations in the clinical manifestations of the disease. Examination under magnification of the small intestinal mucosal surface in severe untreated celiac disease reveals a flat mucosal surface with complete absence of normal intestinal villi. Histologic examination of tissue sections confirms this loss of normal villus structure (Fig. 104-2A). The intestinal crypts are markedly elongated and open onto a flat absorptive surface. The total thickness of the mucosa is reduced only slightly in most cases, because crypt hyperplasia compensates for the absence or shortening of the villi. These architectural changes decrease the amount of epithelial surface available for digestion and absorption.9 The remaining absorptive cells, which appear columnar in normal biopsy specimens, are cuboidal or, at times, squa-

3 1 4 5

2

A

B

Figure 104-2.  Mucosal pathology in celiac disease. A, Duodenal biopsy specimen of a patient with untreated celiac disease. The histologic features of severe villus atrophy (arrow 1), crypt hyperplasia (arrow 2), enterocyte disarray (arrow 3), and intense inflammation of the lamina propria and epithelial cell layer (arrow 4) are evident. B, Repeat duodenal biopsy after 6 months on a strict gluten-free diet. There is marked improvement, with well-formed villi (arrow 5) and a return of the mucosal architecture toward normal.

1799

1800

Section X  Small and Large Intestine moid in celiac disease biopsy specimens. Their cytoplasm is more basophilic (i.e., RNA rich), the basal polarity of the nuclei is lost, and the brush border is markedly attenuated. When viewed by electron microscopy, the microvilli of the absorptive cells appear shortened and often fused. The number of free ribosomes is increased, reflecting impaired differentiation and resulting in the increase in cytoplasmic basophilia evident on histologic examination. Degenerative changes, including cytoplasmic and mitochondrial vacuo­ lization and the presence of many large lysosomes, are obvious. Structural abnormalities of tight junctions between damaged absorptive cells provide a morphologic explanation for the increased permeability of the mucosal barrier in celiac disease.27 The endoplasmic reticulum is sparse, reflecting the low level of synthesis of digestive enzymes, including disaccharidases and peptidases. Thus, mature absorptive cells are reduced in number and functionally compromised. Unlike the absorptive cells, the undifferentiated crypt cells are markedly increased in number in patients with severe untreated celiac disease, and the crypts are therefore lengthened. Moreover, the number of mitoses in crypts is strikingly increased. Cytologic features and histochemistry of the crypt cells are normal by both light and electron microscopy. Studies of epithelial cell kinetics in untreated celiac disease suggest that “villus atrophy” is a misnomer because there is evidence for an actual increase in enteropoiesis in the crypts. Wright and colleagues28 estimated that intestinal mucosa from patients with celiac disease produces six times as many cells per hour per crypt as does normal small intestine and that the cell cycle time is halved, reflecting premature shedding. The experimental evidence suggests, therefore, that the central mechanism of villus shortening in celiac disease is a gliadin-associated toxic effect on maturing enterocytes that results in their premature loss into the intestinal lumen and a compensatory increase in enterocyte replication in the crypts. Such a mechanism would explain many of the histologic abnormalities described earlier. The cellularity of the lamina propria is increased in the involved small intestine. The cellular infiltrate consists largely of plasma cells and lymphocytes. The number of IgA-, IgM-, and IgG-producing cells is increased two-fold to six-fold, but, as in normal mucosa, IgA-producing cells predominate.29 Polymorphonuclear leukocytes, eosinophils, and mast cells also can contribute substantially to the increased cellularity of the lamina propria. The number of IELs per unit area of absorptive epithelium (often reported as number of IELs per 100 enterocytes) is increased in untreated celiac disease.9 In the normal small intestinal mucosa, lamina propria T cells are predominantly CD4+ (helper/inducer) cells, whereas the IELs are mainly CD8+ (cytotoxic/suppressor) cells. In untreated celiac disease, this distribution of lamina propria T cells is maintained, but the density of cells in both compartments is increased. Marsh30 pioneered the theory of a sequence of progression of the celiac lesion in the small intestinal mucosa. Starting with a normal, preinfiltrative (stage 0) mucosa, the initial observed event is an increase in IELs, followed by infiltration of the lamina propria with lymphocytes (stage 1). Crypt hyperplasia (stage 2) precedes villus atrophy (stage 3) and is observed only in the presence of lamina propria lymphocytosis, suggesting that IELs are not sufficient to induce intestinal architectural changes in celiac disease. Finally, total mucosal atrophy (stage 4) develops and is characterized by complete loss of villi, enhanced apoptosis, and crypt hyperplasia.

In untreated patients, the length of small intestinal involvement by the celiac disease lesion varies among individual patients and correlates with the severity of clinical symptoms. Thus, the patient with a severe lesion that involves the full length of the small intestine has more severe malabsorption than the patient with a severe duodenal lesion, a milder jejunal lesion, and a normal ileum. When the intestinal lesion does not involve the entire length of small bowel, the proximal intestine is usually the most severely involved; sparing of proximal intestine with involvement of the distal small intestine can occur, but it is rare. In some untreated patients with clinically mild celiac disease, even the proximal intestine shows only mild partial villus atrophy.9 It is important to note that an increase in IEL count alone is not sufficient to support the histologic diagnosis of celiac disease. This finding is nonspecific and is seen in many other conditions including bacterial overgrowth, mild peptic duodenitis, H. pylori infection, and in other autoimmune disorders. Thus, some shortening of the villi, crypt hyperplasia, cytologically abnormal surface cells, and increased lamina propria cellularity must be present to establish the diagnosis firmly. Treatment with a gluten-free diet results in significant improvement in intestinal structure (see Fig. 104-2B). The cytologic appearance of the surface absorptive cells improves first, often within a few days. Tall, columnar absorptive cells with basal nuclei and well-developed brush borders replace the abnormal, immature cuboidal surface cells; the ratio of IELs to absorptive cells decreases. Subsequently, villus architecture reverts toward normal, with lengthening of the villi and shortening of the crypts; the lamina propria decreases in cellularity. The mucosa of the distal small intestine improves more rapidly than that of the more severely involved proximal bowel.30,31 In some patients, months or even years of gluten withdrawal may be required before the mucosa reverts to normal; indeed, some residual abnormality, which may be striking or subtle, often persists, possibly because of inadvertent gluten ingestion.32 In the debilitated patient with severe untreated celiac disease and associated nutritional deficiency states, pathologic changes may be present in many other organ systems besides the digestive tract. Finally, the mucosal lesion of celiac disease can be identical histologically to the mucosal response to injury typical of a wide range of other enteropathies (see “Differential Diagnosis”).

PATHOGENESIS The interaction of the water-insoluble protein moiety (gluten) of certain cereal grains with the mucosa of the small intestine in susceptible persons is central to the pathogenesis of celiac disease. Although the exact molecular mechanism by which gluten damages the mucosa has not been established, our knowledge of the pathogenesis of celiac disease has accelerated recently. Celiac disease is now considered an immune disorder that is triggered by an environmental agent (gliadin) in genetically predisposed persons. The wide spectrum of clinical manifestations is the result of a complex interplay of varying environmental, genetic, and immune factors. How these control expression of celiac disease and passage from latent to overt disease remains unknown.

ENVIRONMENTAL FACTORS

Celiac disease is a model for autoimmune diseases with a defined environmental trigger. Early work involving physi-

Chapter 104  Celiac Disease and Refractory Celiac Disease ologic digestion with pepsin and trypsin, followed by separation according to solubility properties, identified several wheat proteins as being responsible for the grain’s toxicity in celiac disease. Wheat protein exists in a number of storage forms that can be categorized into four general groups based on their solubility characteristics: prolamins (soluble in ethanol), glutenins (partially soluble in dilute acid or alkali solutions), globulins (soluble in 10% NaCl), and minor albumins (soluble in water). The term gluten encompasses both the prolamins and the glutenins. Although most toxicity studies have been performed with prolamins, there are data to suggest that glutenins also can damage the celiac intestinal mucosa.33 The prolamins of wheat are referred to as gliadins. Pro­ lamins from other cereals also are considered to be gluten and are named according to their source (secalins from rye, hordeins from barley, avenins from oats, and zeins from corn). The taxonomic relationships of the major cereal grain families provide a framework on which their toxicities in celiac disease can be predicted (Fig. 104-3).34 Wheat, rye, and barley belong to the tribe known as Triticeae, and oats belong to a neighboring tribe known as Aveneae. Avenin is genetically less similar to gliadin than gliadin is to secalin and hordein. Despite their genetic differences, however, prolamins from oats, barley, wheat, and rye still have immunologic cross-reactivity because of their common ancestry.35 Grains that do not activate disease (rice, corn, sorghum, and millet) are separated still further from wheat, rye, and barley in terms of their derivation from the primitive grasses. Gliadin can be separated electrophoretically into four major fractions that range in molecular weight from 20 to 75 kd and exist as single polypeptide chains. These have been designated α-, β-, γ-, and ω-gliadins, and all four fractions appear to be toxic to patients with celiac disease.36 The complete amino acid sequences of several of the gliadins and related prolamins in grains other than wheat are known.33 In 2000, Anderson and colleagues14 identified a partially deamidated peptide, consisting of amino acids 56 to 75 of α-gliadin as a dominant epitope, responsible for activation of T cells in celiac disease. The complexity and diversity of the gliadin-specific T-cell response, however, is far greater than was previously appreciated, and persons with celiac disease can respond to a diverse repertoire of gluten peptides.37 Furthermore, the release of intracellular tTG leads to the deamidation of gluten proteins and an enhancement of T-cell responses to the resulting DGPs.14

In organ cultures, a synthetic peptide corresponding to amino acids 31 to 49 of α-gliadin has been shown to be toxic to intestinal mucosa and to induce epithelial lesions via recruitment of IELs. Peptide 31-49 does not activate intestinal CD4+ T cells from patients with celiac disease in vitro, but a related peptide corresponding to amino acids 31 to 43 is capable of activating peripheral CD4+ T cells isolated from patients with celiac disease and of inducing epithelial cell apoptosis and activating macrophages, thereby indicating a likely role for innate immune responses in disease pathogenesis.38 Gianfrani and colleagues39 reported that the α-gliadin-derived peptide corresponding to amino acids 123 to 132 is recognized by CD8+ T lymphocytes from patients with celiac disease and is associated with cytotoxic activity. By contrast, another peptide corresponding to amino acids 57 to 68 appears to function in adaptive immunity via stimulation of intestinal T cells in vivo but does not appear to be directly toxic to the intestinal mucosa of patients in vitro.37 It also is possible that immunologic similarities between gliadin protein motifs and enteric pathogens may be involved in the pathogenesis of an immunologic response to gluten antigens. This hypothesis was supported by a study in which analysis of α-gliadin demonstrated an amino acid region that was homologous to the 54-kd E1b protein coat of adenovirus 12, suggesting that exposure to the virus in a susceptible person could be involved in pathogenesis of celiac disease.40 Although patients with celiac disease have been reported to have a significantly higher prevalence of past adenovirus 12 infection than do control subjects,41 the role of adenovirus molecular mimicry in the pathogenesis of celiac disease has not been confirmed. The reason oats are tolerated by almost all patients with celiac disease is not obvious, because the prolamin fraction of oats contains the same amino acid sequences (QQQPF, where Q = glutamine, P = proline, and F = phenylalanine) that in wheat gliadin have been shown to be toxic.42 A possible explanation for this paradox is that oats contain a relatively smaller proportion of this toxic prolamin moiety than do toxic gluten-containing cereals. Although a feature common to prolamins of wheat, rye, and barley is a high content of glutamine (∼30%) and proline (∼15%), the prolamins of oats have an intermediate content of these amino acids, and the nontoxic prolamins of rice, corn, and millet have an even lower content of them.43 This hypothesis is supported by collectively considering the studies on oat challenge in patients with celiac disease; these studies

Family

Gramineae

Subfamily

Tribe

Subtribe

Genus

Panicoideae

Festucoideae

Aveneae

Triticeae

Triticinae

Oryzeae

Hordeinae

Andropogoneae

Tripsacinae

Anthraxoninae

Paniceae

Triticum

Secale

Hordeum

Avena

Oryza

Zea

Sorghum

Pennisetum

Wheat

Rye

Barley

Oats

Rice

Corn

Sorghum

Millet

Figure 104-3.  Taxonomic relationships of the major cereal grains. (From Kasarda DD, Okita TW, Bernardin JE, et al. Nucleic acid [cDNA] and amino acid sequences of α-type gliadin from wheat [Triticum aestivum]. Proc Natl Acad Sci U S A 1984; 81:4712.)

1801

1802

Section X  Small and Large Intestine suggest that tolerance to oats might depend at least in part on the total amount consumed.44 Daily oats consumption of less than 40 to 60 g/day by patients whose celiac disease is in remission appears to be well tolerated. The data on oats also highlight the important relationship between the amount of gluten consumed and the severity of disease manifestation. A 5- to 10-fold higher incidence of overt celiac disease in children from Sweden compared with Denmark (two populations with similar genetic backgrounds) has long been cited as evidence of the importance of environmental over genetic factors in pathogenesis of celiac disease. Subsequent studies found as much as a 40-fold difference in the gliadin concentration of Swedish compared with Danish infant formula.25 This finding suggests that early exposure of the immature immune system to significant amounts of gliadin is a prominent cofactor for the development of overt celiac disease, possibly by skewing the intestinal immune response to gliadin toward a T-helper 1 (Th1) T-cell response. The age at which gluten is first introduced into an infant’s diet might also play a pivotal role in facilitating gluten tolerance or intolerance. In one study, early exposure to dietary gluten (within three months of birth) was associated with a five-fold increased risk for celiac disease compared with later gluten introduction (four to six months).45 In the same study, delaying gluten introduction (after 7 months of age) also was associated with a slightly increased risk for subsequent celiac disease (1.9-fold compared with the nadir at introduction at four to six months).

GENETIC FACTORS

Family studies that demonstrate frequent intrafamilial occurrence of celiac disease reflect the importance of genetic factors in its pathogenesis.44 Concordance for celiac disease in first-degree relatives ranges between 8% and 18% and reaches 70% in monozygotic twins.46 Our understanding of the nature of this genetic predisposition began with the significant observation by Howell and coworkers10 that celiac disease was associated with specific HLA-DQ2 haplotypes. HLA class II molecules are glycosylated transmembrane heterodimers (α and β chains) that are organized into three related subregions—DQ, DR, and DP—and encoded within the HLA class II region of the major histocompatibility complex on chromosome 6p. An important link to a genetic predisposition was provided by the isolation of gliadin-specific HLA-DQ2-restricted T-cell clones from celiac disease mucosa.11,47 The HLA class II molecule DQ2 is present in more than 90% of persons with celiac disease compared with approximately 35% of the general white population. DQ2 is a heterodimer composed of either α1*0501 or (less commonly) α1*0201 together with β1*02. The DQ α1*0301,β1*0302 heterodimer, known as HLA-DQ8, is found in almost all of the remaining patients with celiac disease. Occasional cases of celiac disease have been reported in patients who are DQ2 and DQ8 negative but nonetheless carry a single DQ2 allele, such as α1*0201. Thus, in some cases, typing of individual celiac-associated alleles may be helpful in addition to determining DQ2 and DQ8 status. A gene dose effect also has been identified, whereby persons who are homozygous for DQ2 are at greater risk than heterozygotes for developing celiac disease. It is now known that after gluten is absorbed, lamina propria antigen-presenting cells (probably dendritic cells) that express HLA-DQ2 or HLA-DQ8, present gliadin peptides on their α/β heterodimer antigen-presenting grooves to sensitized T lymphocytes expressing the α/β T cell receptor (TCR). These lymphocytes then activate B lymphocytes

to generate immunoglobulins and other T lymphocytes to secrete cytokines, including interferon (IFN)-γ, as well as IL-4, IL-5, IL-6, IL-10, tumor necrosis factor (TNF)-α, and transforming growth factor (TGF)-β.48 These cytokines induce not only enterocyte injury but also expression of aberrant HLA class II cell-surface antigens on the luminal surface of enterocytes, possibly facilitating additional direct antigen presentation by these cells to the sensitized lymphocytes (Fig. 104-4). Only a minority of persons who express DQ2 actually develop celiac disease. HLA-DQ2 is expressed by approximately 35% of Europeans and their descendants, but it is rare in other populations (e.g., in sub-Saharan Africa or far eastern Asia). Thus, much of the genetic predisposition to celiac disease is conferred by genes other than those encoding HLA DQ molecules. The search for other genes that confer susceptibility to celiac disease has revealed numerous loci of interest on several different chromosomes, some of which also are associated with susceptibility to type 1 diabetes.49-52

IMMUNE FACTORS

There is substantial evidence implicating both humoraland cell-mediated immune responses to gliadin and related prolamins in the pathogenesis of celiac disease. There is a two- to six-fold increase in the numbers of immunoglobulinproducing B cells in the lamina propria of the small intestine in untreated celiac disease patients.28 In addition, IgA and IgG serum antibodies to purified gliadin, all major fractions of gliadin, and DGPs can be detected in the sera of most patients with untreated celiac disease.14,53-56 Antigliadin antibodies (AGAs), however, do not appear to be essential for the pathogenesis of celiac disease and might simply reflect a nonspecific response to the passage of incompletely digested antigenic gluten proteins across an abnormally permeable intestinal epithelium. Many normal persons have increased IgA or IgG antigliadin.57 The frequency of elevated IgA or IgG DGP antibodies in healthy controls, however, is very low, possibly reflecting the antigenic potency of DGPs and their more central role in disease pathogenesis.14,54-56 Many persons with celiac disease have increased levels of serum antibodies against other food proteins, such as β-lactoglobulin, casein, and ovalbumin.58 It is unclear whether this reflects a general aberrant immune responsiveness to food antigens in patients with celiac disease or enhanced systemic exposure to these proteins because of increased small intestinal permeability. Gluten can be absorbed across normal epithelium, but it is unclear if this results in immune tolerance in persons who are not genetically predisposed to develop celiac disease. The identification of more-specific autoantibody responses has altered our understanding of the pathogenesis of celiac disease. IgA antibodies to endomysium, a connective tissue structure surrounding smooth muscle, are virtually patho­ gnomonic for celiac disease and are found only rarely in the absence of disease.59 It is now known that the target autoantigen contained within the endomysium is the enzyme tTG-2.12 Gliadin is a preferred substrate for this ubiquitous calcium-dependent intracellular enzyme, and it has been shown that tTG deamidates key neutral glutamine residues in gliadin and converts them into negatively charged glutamic acid residues, which are preferred in positions 4, 6, and 7 of the nonapeptide antigen-binding groove of the HLA-DQ2 heterodimer (see Fig. 104-4),13,14,60 thereby facilitating antigen presentation. Thus, tTG-mediated modification of gliadin to generate DGPs plays a pivotal role in eliciting a stronger proliferative response by gliadin-specific

Chapter 104  Celiac Disease and Refractory Celiac Disease Enterocytes

Gluten

Increased HLA-DQ2 gene expression Enterocyte damage

Plasma cell Antibodies (AGA, tTG antibody, others)

T-helper lymphocyte

Cytokines (IFNγ, IL-4, TNF-α)

Lymphocytes (T, NK, B)

T-cell receptor HLA-DQ2 molecule Lamina propria Antigen-presenting cell

α/β T-cell receptor

Neutral glutamines

α-gliadin peptide T-cell epitope

Activated α/β T cell receptor

Tissue transglutaminase -

Positions 4, 6 and 7 of the antigen-binding groove of the HLA-DQ2 molecule

-

-

Negatively charged glutamic acid residues are preferred in positions 4, 6, and 7 of the HLA-DQ2 antigen-binding groove, thereby eliciting a stronger T cell response

Figure 104-4.  Proposed pathogenesis of celiac disease. Gluten is absorbed by the small intestinal mucosa into the lamina propria and presented in conjunction with HLA-DQ2 (or DQ8) cell-surface antigens by antigen-presenting cells, probably dendritic cells, to sensitized T lymphocytes that express the α/β T-cell receptor (afferent limb). Tissue transglutaminase deamidates gliadin peptides, generating acidic negatively charged glutamic acid residues from neutral glutamines (insert). Because negatively charged residues are preferred in positions 4, 6, and 7 of the antigen-binding groove of HLA-DQ2, deamidated gliadin elicits stronger T-lymphocyte responses. These lymphocytes then activate other lymphocytes to generate immune products (cytokines) that damage the enterocytes, resulting in villus atrophy (efferent limb). Induction of aberrant HLA class II cell-surface antigens on the enterocytes can permit additional gluten antigen presentation by these cells to the sensitized lymphocytes. AGA, antigliadin antibodies; HLA, human leukocyte antigen; IFN-γ, interferon gamma; IL-4, interleukin-4; NK, natural killer; TNF-α, tumor necrosis factor alpha; tTG, tissue transglutaminase.

T-cell clones, or, stated differently, tTG makes gliadin tastier for the T cells. With gliadin serving as the glutamine donor, tTG also can generate additional novel antigenic epitopes by crosslinking molecules of the extracellular matrix with gliadin or with tTG-gliadin complexes.61 As evidence of the fundamental role of tTG in celiac disease pathogenesis, one of the dominant epitopes responsible for the T-cell response contains a deamidated glutamine residue (Q65E) of α-gliadin.14 It also has been observed that tTG is necessary for the bioactivation of TGF-β that is required for epithelial differentiation. In a T84-crypt epithelial cell culture system, autoantibodies to tTG-blocked TGF-β–mediated enterocyte differentiation,62 a finding that suggests that release of tTG from cells during inflammation potentiates gliadin presentation by HLA-DQ2 and HLA-DQ8 and that local pro­ duction of autoantibodies to tTG might contribute to the lack of epithelial differentiation observed in the active celiac lesion. Given the marked infiltration of lymphocytes into the small intestinal mucosal epithelium and lamina propria in active disease, it is not surprising that cell-mediated immune

responses also are important in the pathogenesis of celiac disease. Many findings support interplay between adaptive immunity, characterized by a specific and memory T-cell response to gluten peptides, and innate immunity, involving less-specific mechanisms. Many of the T cells in the small intestinal mucosa are activated in untreated celiac disease and release potent proinflammatory mediators such as IFN-γ, TNF-α, IL-2, IL-6, and TGF-β.48 Activated T lymphocytes, most of which are CD4+ cells, are abundant in the lamina propria of the small intestine.63 In contrast, IELs, which are present in large numbers in untreated celiac disease, are predominantly CD8+ T cells.64 There is an influx of primed memory T cells, marked by high CD45RO expression, in the mucosa of untreated celiac disease patients.65 In healthy persons, more than 90% of IELs express the α/β TCR, whereas expression of the γ/δ TCR by IELs in patients with untreated celiac disease is increased as much as six-fold (to 35%) and is considered a hallmark of the disease.66 These primitive lymphocytes recognize bacterial nonpeptide antigens and unprocessed stress-related proteins. They appear to act as mucosal guardians and might protect the intestinal mucosa from chronic exposure

1803

1804

Section X  Small and Large Intestine to dietary gluten in gluten-tolerant persons by secreting IL-4, which dampens Th1 in favor of Th2 reactivity.67 Their continuous presence in patients on a gluten-free diet might indicate inadvertent gluten ingestion. Patients with refractory celiac disease also have aberrant IELs with restricted γ/δ TCR gene rearrangements indicating oligoclonality. The pathogenetic role of these lymphocytes, compared with lamina propria lymphocytes, continues to evolve (see “Refractory Celiac Disease”).68 Studies suggest that IL-15 plays a key role in bridging the innate and adaptive immune responses in pathogenesis of celiac disease.15,16,69 This enterocyte- and macrophagederived proinflammatory cytokine is increased massively in the mucosa of patients with active celiac disease and refractory celiac disease. Although the mechanisms that lead to its overproduction remain unknown, IL-15 regulates IEL homeostasis by promoting migration, preventing apoptosis, and enhancing the capacity of dendritic cells to function as antigen-presenting cells.69 In response to gliadin peptides, IL-15 triggers an adaptive CD4+ T-cell response in the lamina propria and also is capable of inducing direct epithelial cell injury by inducing IEL secretion of IFN-γ.16

CLINICAL FEATURES Samuel Gee’s classic description, with its evocative account, was concerned largely with the gross manifestations of the disorder.5 This florid presentation, however, is now unusual in the Western world, constituting only the extreme tip of the celiac iceberg. Although some patients still present with severe illness, most have few, subtle, or no symptoms at diagnosis. Such cases may be identified by screening relatives of patients during research studies or from screening patients with associated disorders, such as type 1 diabetes mellitus, autoimmune thyroid disease, or Down syndrome. Incidental hematologic abnormalities (e.g., iron deficiency anemia) or biochemical abnormalities (e.g., elevated serum aminotransferase levels) also can lead to a diagnosis of celiac disease.

CHILDHOOD PRESENTATION

The classic presentation of celiac disease in infancy is not easily missed. The typical history is of steatorrhea with or without vomiting and occasional cramping abdominal pain that can occur anytime after weaning when cereals are introduced into the diet, but especially in the first and second years of life. Classically, the child fails to thrive, is apathetic and irritable, and has muscle wasting, hypotonia, and abdominal distention. Watery diarrhea or occasionally constipation may be reported. Diagnosis is more difficult when gastrointestinal features are less prominent, and the possibility of gluten sensitivity should be considered in all children who present with short stature or failure to thrive, even when there are no other symptoms to suggest an entero­ pathy. Once a gluten-free diet is commenced, catch-up growth is well documented.70 Nutritional deficiencies, particularly anemia, are another common mode of presentation, especially in older children. With earlier diagnosis, clinical rickets now is an uncommon complication but is seen occasionally, especially among Asian children with untreated celiac disease. Many pediatric patients enjoy a temporary, spontaneous remission of symptoms during adolescence, and it is unusual for celiac disease to manifest during the teens. Considerable debate continues as to why celiac disease tends to be diagnosed later and with milder signs and symp-

toms than in the past. A number of studies suggest that breast-feeding can significantly delay the onset of symptoms,71,72 but not all studies support this conclusion.73 In one study of at-risk children, the introduction of gluten into the diet during the first three months of life or after seven months of age was associated with a significantly increased risk for celiac disease (hazard ratio [HR] 23.0; 95% confidence interval [CI]: 4.6-115.9; P = 0.001) compared with introduction at four to six months (HR 3.98; 95% CI: 1.1813.46; P = 0.04).74

ADULTHOOD PRESENTATION

In the past, celiac disease was perceived to be a pediatric disorder, but the diagnosis now is being made increasingly in adults; currently, the overall mean age at presentation is approximately 45 years. Symptoms also have changed during the past 50 years. Diarrhea now is reported less often, and many patients now present with higher body mass indices and even with obesity. The unmasking of asymptomatic disease by surgery that induces rapid gastric emptying (e.g., gastric resection, pyloroplasty) or the finding of the typical lesion in asymptomatic relatives of celiac disease patients suggests that adults can have silent celiac disease for some time. A proportion of these adult patients have short stature or give a history consistent with unrecognized celiac disease in childhood. In many, however, there is nothing to suggest previous disease, and it is possible that celiac disease can develop for the first time in adult life. Celiac disease also is being diagnosed increasingly in later life, with approximately 25% of cases diagnosed in patients older than 60 years.75

GASTROINTESTINAL FEATURES

Clinical manifestations of celiac disease vary tremendously from patient to patient. Because many of the symptoms result from intestinal malabsorption, they are not specific for celiac disease and resemble those seen in other malabsorptive disorders. Many adults present with gastrointestinal symptoms including diarrhea, steatorrhea, flatulence, and weight loss similar to those seen in childhood celiac disease. Diarrhea often is episodic rather than continuous. Nocturnal, early morning, and postprandial diarrhea are common. Patients with extensive intestinal involvement can have more than 10 stools per day. Because of their high fat content, the stools of patients with celiac disease may be light tan or grayish and greasy in appearance, with a tendency to float and to be difficult to flush from the toilet bowl; this pallor of the stools is reflected in the ancient Latin term diarrhea alba, alba meaning “white.” Steatorrhea often is absent in patients with disease limited to the proximal small intestine. Several factors contribute to the diarrhea associated with celiac disease. The stool volume and osmotic load delivered to the colon are increased by the malabsorption of fat,76 carbohydrate, protein, electrolytes, and other nutrients. In addition, the delivery of excessive dietary fat into the large bowel results in the production by bacteria of hydroxy fatty acids, which are potent cathartics. Electrolytes actually are secreted into, rather than absorbed from, the lumen of the severely damaged upper small intestine in symptomatic patients. This secretion further increases luminal fluid in an intestine with an already compromised absorptive capacity. There also is evidence that secretin and cholecystokinin release in response to a meal are impaired in celiac disease, diminishing delivery of bile and pancreatic secretions into the gut lumen and possibly compromising intraluminal digestion.77 Alterations in the secretion of other intestinal peptides have been noted and can contribute to the observed

Chapter 104  Celiac Disease and Refractory Celiac Disease Table 104-1  Extraintestinal Manifestations of Celiac Disease MANIFESTATION Cutaneous Ecchymoses and petechiae Edema Dermatitis herpetiformis Follicular hyperkeratosis and dermatitis Endocrinologic Amenorrhea, infertility, impotence Secondary hyperparathyroidism Hematologic Anemia Hemorrhage Thrombocytosis, Howell-Jolly bodies Hepatic Elevated liver biochemical test levels Muscular Atrophy Tetany Weakness Neurologic Peripheral neuropathy Ataxia Demyelinating central nervous system lesions Seizures Skeletal Osteopenia Osteoarthropathy Pathologic fractures

diarrhea. Finally, if the disease extends to and involves the ileum, patients can experience the direct cathartic action of malabsorbed bile salts on the colon.76 The amount of weight loss in a patient with celiac disease depends on the severity and extent of the intestinal lesion and on the ability of the patient to compensate for the malabsorption by increasing dietary intake. Some celiac disease patients with substantial malabsorption have enormous appetites and lose little or no weight. Rarely, in severe disease, anorexia develops with associated rapid and severe weight loss. In such debilitated patients, some of the weight loss may be masked by fluid retention caused by hypoproteinemia. Malaise, lassitude, and fatigue also are common even when anemia is absent. Occasionally, severe hypokalemia resulting from fecal loss of potassium causes severe muscle weakness. Vague abdominal discomfort and especially abdominal bloating are extremely common and can lead to a mistaken diagnosis of irritable bowel syndrome (IBS). Because of the difficulty in distinguishing celiac disease with mild gastrointestinal manifestations from symptomatic IBS, serologic testing of IgA EMAs or IgA tTG should be considered in patients with symptoms suggesting diarrhea-predominant IBS. In a UK study, Sanders and colleagues78 evaluated 300 consecutive new patients who fulfilled Rome II criteria for IBS and 300 healthy age- and sex-matched controls for celiac disease using IgA AGA (antigliadin antibody), IgG AGA, and EMA; two controls (0.7%) (both EMA positive) and 14 IBS patients (4.6%) had celiac disease (P = 0.004; odds ratio [OR], 7.0; 95% confidence interval [CI]: 1.7-28.0). Severe abdominal pain can occur but is uncharacteristic in uncomplicated celiac disease; its occurrence can suggest the presence of complications such as intussusception, ulcerative jejunitis, or intestinal lymphoma. Abdominal disten-

PROBABLE CAUSE(S) Vitamin K deficiency; rarely, thrombocytopenia Hypoproteinemia Unknown Vitamin A malabsorption, vitamin B complex malabsorption Malnutrition, hypothalamic-pituitary dysfunction Calcium and/or vitamin D malabsorption causing hypocalcemia Iron, folate, vitamin B12, or pyridoxine deficiency Vitamin K deficiency; rarely, thrombocytopenia due to folate deficiency Hyposplenism Unknown Malnutrition due to malabsorption Calcium, vitamin D, and/or magnesium malabsorption Generalized muscle atrophy, hypokalemia Deficiencies of vitamins such as vitamin B12 and thiamine Cerebellar and posterior column damage Unknown Unknown Malabsorption of calcium and vitamin D Unknown Osteopenia

tion with excessive amounts of malodorous flatus is a common complaint. Conversely, nausea and vomiting are uncommon in uncomplicated celiac disease. Recurrent severe aphthous stomatitis affects many celiac patients and may be their sole presenting complaint. It is important to exclude celiac disease in cases of recurrent aphthous stomatitis because a significant proportion of these patients respond well to dietary treatment.79

EXTRAINTESTINAL FEATURES

As patients with celiac disease get older, they tend to present with complaints not directly referable to the gastrointestinal tract. These extraintestinal symptoms and clinical findings often result from nutrient malabsorption and can involve virtually all organ systems (Table 104-1).80 Extraintestinal features, including anemia, osteopenia, neurologic symptoms, and menstrual abnormalities, often prove more distressing to the patient than do the gastrointestinal symptoms.

Anemia

Anemia is a common manifestation of celiac disease in children and adults and usually is caused by impaired iron or folate absorption from the proximal intestine; in severe disease with ileal involvement, vitamin B12 absorption also is impaired. Patients with extensive disease can bleed into the skin or mucous membranes or can develop hematuria, epistaxis, or vaginal or gastrointestinal bleeding. Bleeding can aggravate pre-existing anemia and most often is caused by a coagulopathy resulting from impaired intestinal absorption of fat-soluble vitamin K. Evidence of hyposplenism of unknown cause, with thrombocytosis, deformed erythrocytes, and splenic atrophy, occurs in up to 50% of adults with celiac disease but only rarely is seen in children.81 In

1805

1806

Section X  Small and Large Intestine many patients, evidence of hyposplenism disappears with elimination of gluten from the diet.81

Osteopenia

Osteopenia is the most common complication of celiac disease, and its prevalence increases with age at diagnosis. More than 70% of patients with untreated celiac disease have osteopenia,82 and osteoporosis occurs in more than one quarter of all celiac disease patients.83 Osteopenia develops as a result of impaired calcium absorption (secondary to defective calcium transport by the diseased small intestine), vitamin D deficiency (caused by impaired absorption of this fat-soluble vitamin), and binding of intraluminal calcium and magnesium to unabsorbed dietary fatty acids (forming insoluble soaps, which are then excreted in the feces). Chronic intestinal inflammation also can contribute to bone loss through release of inflammatory mediators. Patients can present with bone pain, especially of the lower back, rib cage, and pelvis. Calcium and magnesium depletion can cause paresthesias, muscle cramps, and even frank tetany. With prolonged calcium malabsorption, patients may develop secondary hyperparathyroidism resulting in mobilization of calcium from the bones, further exacerbating the osteopenia. Osteopenia is less common in patients with silent celiac disease, in whom prevalence rates between 30% and 40% have been reported.84 Whereas bone disease generally is more severe among patients with symptomatic disease, severe osteopenia has been reported in up to one third of symptom-free adults whose celiac disease was diagnosed during childhood and who resumed a normal diet during adolescence.85 A key unanswered question is the functional consequence of osteopenia. An increased risk of fractures was observed in patients with overt celiac disease in one study84 but not in another.86 The fracture risk among patients with silent celiac also remains unclear.

Neurologic Symptoms

Neurologic symptoms caused by lesions of the central or peripheral nervous system occasionally occur in patients with celiac disease and are poorly understood. Celiac disease often is found in patients presenting with non­ hereditary ataxia, and progressive gait and limb ataxia may be the sole manifestations of disease in some patients. These abnormalities, referred to as gluten ataxia, are believed to result from immunologic damage to the cerebellum, posterior columns of the spinal cord, and peripheral nerves.87 Muscle weakness and paresthesias with sensory loss also are encountered occasionally, and pathologic evidence of peripheral neuropathy and patchy demyelinization of the spinal cord, cerebellar atrophy, and capillary proliferation suggestive of Wernicke’s encephalopathy have been described rarely. Although potential causative roles for specific vitamin deficiencies (including vitamin B12, thiamine, riboflavin, and pyridoxine) have not been established, neurologic symptoms have been reported to improve in some patients who are given multivitamins, including vitamins A, B, and E, and calcium. Night blindness is a clear indication for vitamin A therapy. Peripheral neuropathy and ataxia, however, often appear unrelated to specific vitamin deficiency states and usually do not respond to gluten withdrawal.88 The associations of celiac disease and epilepsy, frequently complex partial seizures, and bilateral parieto-occipital cerebral calcification are well recognized.89 In one series, epilepsy was reported in approximately 5% of children and

young adults with celiac disease.90 The cause of the epilepsy remains unclear, and the prognosis might depend on how early in the course of the disease a gluten-free diet is started. Although most patients with celiac disease do not appear psychologically abnormal, many affected subjects report a striking improvement in mood after commencing a glutenfree diet.91

Gynecologic and Fertility Problems

Gynecologic and obstetric problems are common in women with untreated celiac disease.92 Amenorrhea occurs in one third of women of childbearing age and menarche is often delayed, typically by one year, in untreated subjects. Women with untreated celiac disease can present with infertility, and it is common for infertile women with celiac disease to become pregnant shortly after commencing a glutenfree diet.93 A high prevalence of silent celiac disease has been reported in women with recurrent spontaneous abortions, intrauterine fetal growth retardation, and unfavorable outcomes of pregnancy, underlining the need to test for celiac disease in these situations.94 Infertility secondary to impotence or an abnormally low sperm count can occur in men with untreated celiac disease.95 Although malnutrition, including folate deficiency related to malabsorption, can contribute to male infertility, abnormalities in hypothalamic-pituitary regulation of gonadal function and gonadal androgen resistance that disappears on gluten withdrawal also have been incriminated.95

Physical Examination

Physical findings, like symptoms, vary considerably among patients with celiac disease. Patients with mild disease often have a completely normal physical examination. In more severe disease, physical abnormalities usually result from malabsorption and, therefore, are not specific for celiac disease. Growth retardation commonly occurs in children, but when they commence a gluten-free diet before puberty, a compensatory growth spurt occurs so the effect on adult height potentially can be minimized. Persons with celiac disease are, on average, three inches shorter than their peers. Tall patients are seen, however, and a height of more than six feet does not preclude the diagnosis. In patients with severe celiac disease, emaciation with evidence of weight loss, including loose skin folds and muscle wasting, may be prominent. It is common for adults with celiac disease to experience a weight gain of more than six kilograms following institution of a gluten-free diet. Clubbing of the fingers occurs occasionally, and koilo­ nychia may result from long-standing iron deficiency anemia. There may be pitting edema of the lower extremities secondary to hypoproteinemia. Hypotension may be related to fluid and electrolyte depletion, and the skin may be dry with poor turgor, if there is dehydration. Occasionally a low-grade fever associated with anemia is found in untreated celiac disease, but this finding might indicate a concurrent complication, such as infection or malignancy, particularly lymphoma. Increased skin pigmentation may be obvious in severely ill patients. In addition to dermatitis herpetiformis (DH) (see later), other dermatologic findings may include spontaneous ecchymoses related to hypoprothrombinemia, hyperkeratosis follicularis caused by vitamin A deficiency, and pallor caused by anemia. Examination of the mouth may show aphthous stomatitis, angular cheilosis, and glossitis with decreased papillation

Chapter 104  Celiac Disease and Refractory Celiac Disease of the tongue. Dental enamel defects are common.96 The abdomen may be protuberant and tympanitic, with a characteristic doughy consistency, owing to distention of intestinal loops with fluid and gas. Hepatomegaly and abdominal tenderness are uncommon, but ascites may be detected in patients with significant hypoproteinemia. Peripheral lymphadenopathy is unusual in the absence of complicating lymphoma. The extremities may reveal loss of various sensory modalities, including light touch, vibration, and position, usually resulting from peripheral neuropathy and, rarely, demyelinating spinal cord lesions. If neuropathy is severe, deep tendon reflexes are diminished or even absent. Hyperpathia may be present. A positive Chvostek or Trousseau sign may be elicited in patients with severe calcium or magnesium depletion. In such persons, bone tenderness related to osteoporosis may be elicited, especially if collapsed vertebrae or other fractures are present.

DIAGNOSIS Laboratory findings in celiac disease, like the symptoms and signs, vary with the extent and severity of the intestinal lesion. Serum IgA EMA or tTG antibody and small intestinal biopsy are the most reliable diagnostic tests for celiac disease. Stool studies, hematologic and biochemical tests, and radiologic studies may be abnormal, but they seldom provide a specific diagnosis because similar abnormalities often are seen in patients with other malabsorptive diseases (see Chapter 101).

SEROLOGY

In current clinical practice, there are many serologic studies to aid in the diagnosis of celiac disease; however, the most powerful and clinically useful are the IgA EMA and IgA tTG assays. IgA EMA and IgA tTG are based on the target antigen tTG, and IgA and IgG AGAs are based on the target antigen gliadin.97 A second-generation AGA test has been developed based on DGPs with improved diagnostic accuracy.14,54-56 The approximate sensitivity and specificity of commonly used serum antibody tests are outlined in Table 104-2.98-102 A working group of 13 European laboratories has attempted to improve standardization by establishing standard curves on reference sera and protocols for calibration of quality controls. This collaboration has reported that IgA EMA has a sensitivity of 90%, specificity of 99%, and reproducibility of 93% and currently remains the gold standard. IgA antihuman tTG is slightly less reliable (sensitivity 93%, specificity

95%, reproducibility 83%), and the anti-IgG AGA and IgA AGA are the least reliable.98 In addition to laboratory variation, the reported sensitivities and specificities of these tests depend on the prevalence of the disease in the tested population and the severity of the disease. In one study of 101 patients with biopsy-proven celiac disease, the sensitivity of IgA EMA among patients with total villus atrophy was 100% compared with only 31% in those with partial villus atrophy.103

Immunoglobulin A Endomysial Antibody

Serum IgA EMA binds to connective tissue (endomysium), surrounding smooth muscle cells, producing a characteristic staining pattern that is identified by indirect immunofluorescence.59 The target antigen has been identified as tTG. Frozen sections of monkey esophagus initially were used for the assay, but currently most laboratories use sections of human umbilical cord, which are more readily available.104 The test result is reported simply as positive or negative because even low titers of serum IgA EMA are highly specific for celiac disease. IgA EMA has a sensitivity of 90% or greater and a specificity approaching 100% in untreated celiac disease.59,105 Antibody levels fall on a gluten-free diet, the test often becoming negative in treated patients.106 The clinical applications of EMA and other serologic tests are discussed in the following sections.

Tissue Transglutaminase Antibodies

The epitope against which EMA is directed has been identified as type 2 tTG (tTG-2).12 Type 1 tTG, which has different structure, enzymatic activities, and tissue distributions, is not implicated in celiac disease. Type 3 (epidermal) tTG plays a distinct role in DH (discussed later). An IgA enzyme-linked immunosorbent assay (ELISA) that used guinea pig tTG was used initially, and it proved less costly and easier to perform than the immunofluorescence assay used to detect IgA EMA. IgA guinea pig tTG assays are sensitive and specific for the diagnosis of celiac disease.99,100 In one study, anti-tTG was present in 98% of patients with biopsy-proven celiac disease compared with 5% of controls.99 In another study that included 136 patients with celiac disease and 207 controls, the sensitivity and specificity of IgA guinea pig tTG were 95% and 94%, respectively.100 IgA guinea pig tTGs, however, are responsible for some false-positive results, particularly in at-risk persons with autoimmune diseases or liver disease and in patients with other inflammatory bowel diseases.98 ELISAs using human tTG are now preferred because they have proved to be more specific than those using guinea pig tTG.101,102 In contrast to the false-positive results of IgA

Table 104-2  Sensitivity, Specificity, and Positive and Negative Predictive Values of Serologic Tests for Untreated Celiac Disease SEROLOGIC TEST

SPECIFICITY* (%)

POSITIVE PREDICTIVE VALUE (%)

NEGATIVE PREDICTIVE VALUE (%)

85-98 95-98 95-100

97-100 94-95 97-100

98-100 91-95 80-95

80-95 96-98 100

75-90 69-85

82-95 73-90

28-100 20-95

65-100 41-88

SENSITIVITY* (%)

Immunoglobulin A Endomysial Antibody Indirect immunofluorescence assay Guinea pig tTG† ELISA Human tTG‡ ELISA Antigliadin Antibodies (AGAs) IgA IgG

*Wide variations in test sensitivity and specificity rates are reported among different laboratories.98 The guinea pig tissue transglutaminase antibodies data are based on two large studies.99,100 ‡ The human tissue transglutaminase antibodies data are based on two large studies.101,102 AGA, antigliadin antibodies; ELISA, enzyme-linked immunosorbent assay; IgA, immunoglobulin A; tTG, tissue transglutaminase.

†

1807

1808

Section X  Small and Large Intestine guinea pig tTG, a false-positive IgA human tTG result is unlikely (especially at high titer). A false-positive EMA also is highly unlikely and in the setting of a normal biopsy might indicate a future predisposition to development of clinical celiac disease—that is, potential celiac disease. Although IgG and IgM tTG assays are useful in patients with IgA deficiency, they are not as sensitive nor as specific.

Antigliadin Antibodies

Serum IgA and IgG AGA levels often are elevated in untreated celiac disease. Unfortunately, these tests have only moderate sensitivity and their specificity is substantially lower than those of IgA EMA or tTG tests.1,53,97,107 Thus, testing for AGA no longer is recommended as a primary test for untreated celiac disease; IgA EMA or IgA tTG testing is preferable. The sensitivity and specificity of IgA AGA are marginally superior to those of IgG AGA; however, many clinicians test simultaneously for both IgA and IgG AGA, an approach that gives a small incremental increase in sensitivity but reduces specificity even further. IgG AGA testing may be useful in the 2% of patients with celiac disease who have IgA deficiency. The positive predictive value of AGA in a general population is relatively poor. In one series, for example, the positive predictive value for IgG AGA, corrected for the expected prevalence in the general population, was less than 2%.105 AGA test results are reported as a titer—a high titer of AGA is somewhat more specific for celiac disease than a low titer—but some normal persons have high AGA levels.57 It is now known that tTG-2 catalyses the deamidation of gliadin peptides. This increases their binding to the antigen groove of DQ2, thereby increasing their toxicity.14 Based on this knowledge, a second-generation AGA test has been developed using synthetic deamidated gliadin peptides (DGP) that replicate the structure of tTG-modified gliadin antigens. IgA and IgG DGP assays are available with sensitivities, in early studies, similar to those obtained using tTG assays.54-56 The specificity of a positive DGP test result is substantially greater than that of standard AGA assays, presumably indicating that celiac-associated AGAs recognize DGP, whereas AGAs in persons who do not have celiac disease recognize other epitopes less specific for the disease. DGP testing is not yet in widespread use, but potential applications include the confirmation or exclusion of celiac disease in persons who have small bowel biopsy findings consistent with celiac disease but a negative IgA-tTG test result (including those with IgA deficiency where DGP-IgG can be used) and in persons with positive AGA test results but a negative IgA-tTG. DGP concentrations fall on a glutenfree diet in a similar fashion to AGA or anti-tTG concentrations.54 A more controversial application of DGP testing is to use anti-tTG and anti-DGP assays in combination for diagnosis or exclusion of disease without small bowel biopsy.

Clinical Application of Serologic Tests

Serologic tests are used to evaluate patients with suspected celiac disease, monitor adherence and response to a glutenfree diet, and potentially, screen asymptomatic persons for the disease. An approach to diagnosing celiac disease is outlined in Figure 104-5.108 When the index of suspicion is low—the pretest probability is less than 10%—a negative result for IgA EMA or IgA tTG has a high negative predictive value and can obviate the need for small bowel biopsy. Falsely negative IgA EMA and IgA tTG test results are more likely to occur in very young children (<2 years of age), those with

mild celiac enteropathy, and, of course, in IgA deficiency. The AGA tests have lower diagnostic accuracy and are no longer recommended for initial diagnosis of celiac disease.105 Because the specificities of IgA EMA and IgA tTG tests are high, their positive predictive values are high even in lowrisk populations.105,109,110 When the index of suspicion is moderate to high—the pretest probability is greater than 10%—the very high specificity of IgA EMA and IgA tTG has led to debate as to whether a positive result in the appropriate clinical setting can be considered diagnostic and eliminate the need for small bowel biopsy. We recommend that both IgA tTG (or EMA) and a small bowel biopsy be performed before dietary treatment is recommended, because this approach provides the best means of making a definitive diagnosis of celiac disease at the outset. Because AGA tests have a high falsepositive rate, there is no role for a trial of a gluten-free diet for presumed celiac disease based on the finding of an elevated IgA or IgG AGA level. The use of anti-DGP testing was discussed earlier. Similar to IgA AGA levels, IgA tTG levels decrease in the months following a gluten-free diet and are useful in assessing dietary compliance and excluding inadvertent gluten ingestion.53,54,111 IgA AGA and IgA tTG currently are the most widely used tests for monitoring adherence and response to a gluten-free diet among patients whose antibody levels are elevated before therapy.54,107,111 Hence, a pretreatment antibody level should be determined at the time of diagnosis. A normal baseline value typically is reached within 3 to 12 months depending on the pre­ treatment antibody concentration (IgA tTG concentrations fall with a half-life of approximately six to eight weeks) and on the degree of success in avoiding gluten ingestion. If the levels do not fall as anticipated, the patient may be continuing to ingest gluten either intentionally or inadvertently. The advent of highly sensitive and specific serologic tests has changed the epidemiology of celiac disease radically by revealing the high incidence of silent celiac disease; this awareness in turn has led to debate on the merits of mass screening. To date, the benefit of screening for asymptomatic celiac disease, usually using IgA tTG or EMA, has not been demonstrated.112 The potential advantages of screening for asymptomatic celiac disease are a reduction in risk for malignancy including enteropathy-associated T-cell lymphoma (EATL); a reversal of unrecognized nutritional deficiency states; resolution of mild or ignored intestinal symptoms; possible reduction in T-cell activation and “antigenic drift” to other autoantigens, thereby reducing the onset of other autoimmune disorders; and an improvement in general well-being.113,113A All of these hypothetical benefits, however, depend on compliance with a challenging dietary regimen, and asymptomatic patients might not be motivated sufficiently to adhere to a strict gluten-free diet.114 There also may be adverse psychological effects when asymptomatic indivi­ duals are given the diagnosis of celiac disease. Furthermore, the natural history of undetected celiac disease and the consequences of screening and treating silent celiac disease are unknown. For these reasons, mass screening of asymptomatic persons generally is not advocated at this time, even in populations in which the prevalence of celiac disease is high. The current standard of care is a case-finding approach, which targets at-risk subjects such as patients with mild gastrointestinal symptoms, iron deficiency anemia, or IBSlike symptoms, all instances in which the value of serologic testing for celiac disease is accepted widely. Simply by case

Chapter 104  Celiac Disease and Refractory Celiac Disease Low probability of celiac disease

Moderate-to-high probability of celiac disease (e.g., typical GI symptoms and a family history of celiac disease, steatorrhea, unexplained iron deficiency anemia, or failure to thrive in children)

IgA EMA or tTG antibody ± IgA AGA, IgG AGA Positive

IgA EMA or tTG antibody serology and small intestinal biopsy

IgA EMA or tTG antibody positive Histology negative

Review or repeat biopsy

IgA EMA or tTG antibody positive Histology positive Positive

Tr e a t

These diagnoses excluded or unlikely

Small intestinal biopsy

IgA EMA or tTG antibody negative Histology positive

Negative

Diagnosis excluded

IgA EMA or tTG antibody negative Histology negative

Consider other causes of villus atrophy Cow’s milk protein intolerance (children) Post-gastroenteritis syndrome Giardiasis Peptic duodenitis (including Zollinger-Ellison syndrome) Crohn’s disease Small intestinal bacterial overgrowth Eosinophilic gastroenteritis Radiation or cytotoxic chemotherapy Tropical sprue Severe malnutrition Diffuse small intestinal lymphoma Graft versus host disease Hypogammaglobulinemia Alpha chain disease

Figure 104-5.  Diagnosis of celiac disease. A false-positive IgA EMA or tTG is rare; a false-negative IgA EMA and tTG can occur in mild enteropathy, in children younger than 2 years, and in patients with IgA deficiency. AGA, antigliadin antibodies; EMA, endomysial antibody; GI, gastrointestinal; Ig, immunoglobulin; tTG, tissue transglutaminase. (From Farrell RJ, Kelly CP. Diagnosis of celiac sprue. Am J Gastroenterol 2001; 96:3237.)

finding among at-risk subjects in a primary care setting, Hin and colleagues115 observed a four-fold increase in the number of celiac disease diagnoses during a one-year period. Thus, for now, increased awareness of the typical and atypical presentations of celiac disease, coupled with a low threshold for serologic testing in at-risk subjects, can uncover a substantial portion of the submerged iceberg.

GENETIC TESTING

As discussed earlier, almost all patients with celiac disease are positive for HLA DQ2 or DQ8. As a result, HLA testing may be helpful in excluding celiac disease in specific clinical situations, mainly before embarking on a gluten challenge (see later) or evaluating a patient who has a celiac-like enteropathy but negative IgA tTG, EMA, and DGP serologies, in which case a negative result indicates the need to investigate for an alternative diagnosis. Approximately 40% of persons of European ancestry are DQ2 or DQ8 positive, however, so a positive result is of little diagnostic value.97,116-118

SMALL INTESTINE BIOPSY

Although the diagnosis of celiac disease may be suspected on clinical grounds or as a result of abnormal serologic tests, biopsy of the small intestine has remained the standard test to establish the diagnosis. Biopsies usually are performed during endoscopic examination of the upper gastrointesti-

nal tract, an examination that may be indicated for reasons related or unrelated to celiac disease, such as investigation of iron deficiency anemia or upper abdominal discomfort.119 Multiple biopsies should be obtained (e.g., a total of six to eight biopsies from the second and third parts of the duodenum). Biopsies taken from the duodenal bulb also may be diagnostic, but might show mucosal architectural distortion produced by Brunner’s glands and changes caused by peptic duodenitis, both of which can cause difficulty in histopathologic diagnosis.9 Scalloping or absence of duodenal folds has been noted in some patients with celiac disease (Fig. 104-6).120 Scalloping is not specific for celiac disease, however, and may be seen in eosinophilic enteritis, giardiasis, amyloidosis, tropical disease, and human immunodeficiency virus entero­ pathy.121 Other endoscopic features include multiple fissures or a mosaic-like appearance where the fissures circumscribe areas of mucosal nodularity in a manner similar to the grouting around a mosaic tile. The mucosa of celiac disease, however, often appears normal at endoscopy, and absence of the previously described macroscopic features does not obviate the need for biopsy and histologic examination if celiac disease is suspected, based on clinical grounds or serologic testing. Once the biopsy tissue has been obtained, handling should be minimized to avoid artifactual damage. By gently floating the tissue on a small amount of saline, an experi-

1809

1810

Section X  Small and Large Intestine

Figure 104-6.  Endoscopic view of the duodenum showing scalloping of the folds in a patient with celiac disease; if present, this finding should alert the endoscopist to the possibility of this diagnosis. Scalloping, however, is not specific for celiac disease and has been seen in other conditions, including eosinophilic gastroenteritis, giardiasis, amyloidosis, tropical sprue, and human immunodeficiency virus enteropathy.

enced observer can examine even small endoscopic biopsy specimens using a dissecting microscope or a magnifying lens. The villus structure of normal small intestinal mucosa is readily apparent. In severe untreated celiac disease, a flat, featureless mucosa is recognized easily. The mosaic pattern of less complete forms of villus atrophy also may be recognized, although differentiation from the normal is more difficult and requires some experience. Capsule endoscopy and enteroscopy provide opportunities to examine the more distal small intestine, where the macroscopic features just described might also be evident. Because these features are not specific, however, capsule endoscopy cannot replace biopsy and histopathologic examination, and it is not necessary for routine diagnosis and management. Capsule endoscopy may be valuable in patients with complicated or refractory celiac disease in whom ulcerative jejunitis or intestinal lymphoma may be seen.

GLUTEN CHALLENGE

In the past, gluten challenge—discontinuation of the glutenfree diet, followed by repeat biopsy of the small intestine— was considered an important confirmatory step in the diagnosis of celiac disease. In current practice, however, gluten challenge is reserved for the few patients in whom the diagnosis remains in doubt after a period of treatment with a gluten-free diet. Gluten challenge is seldom necessary for patients who present with typical signs or symptoms of celiac disease and have documented abnormalities consistent with a celiac lesion on small bowel biopsy. A positive IgA EMA or IgA tTG test before treatment lends further support to the diagnosis of celiac disease and makes a later gluten challenge superfluous. A gluten challenge should be considered in patients who began a gluten-free diet empirically without documentation of a characteristic intestinal lesion or the presence of IgA

EMA antibody. In such patients, symptomatic response to a gluten-free diet might indicate the presence of glutensensitive enteropathy or simply reflect a change in gastrointestinal function in response to a major dietary change. Mild and nonspecific gastrointestinal symptoms after ingesting gluten are not a reliable way to diagnose celiac disease. For example, patients with IBS might experience improvement in symptoms, including abdominal bloating, cramping, or diarrhea, after beginning a gluten-free diet.122 Gluten challenge also should be considered if a diagnosis of celiac disease was made during childhood based on small intestinal biopsy abnormalities in the absence of a positive IgA EMA or IgA tTG, because a number of transient childhood enteropathies can mimic the celiac lesion (see later). Before embarking on a gluten challenge, patients should be evaluated to determine their current response to treatment with a gluten-free diet. This should include a careful nutritional history noting their symptomatic responses to previous episodes of inadvertent or purposeful ingestion of gluten-containing foods. Patients who experience substantial symptoms following gluten ingestion are unlikely to tolerate formal gluten challenge and might prefer to remain on a gluten-free diet despite diagnostic uncertainty. Serologic studies, such as IgA tTG, should be performed, and HLA typing for DQ2 and DQ8 may be helpful because a negative result virtually excludes celiac disease. A small bowel biopsy should be obtained as a baseline because an abnormal biopsy likely obviates the need for challenge. Gluten challenge must be initiated with caution because occasionally patients are exquisitely sensitive to small amounts of gluten123; other patients might require prolonged challenge before symptoms or significant histologic abnormalities recur. If a small amount of gluten, such as a cracker or one quarter of a slice of bread, is tolerated, the amount can be doubled every three days until the patient is ingesting the equivalent of at least four slices of bread daily. The challenge should be continued for at least six weeks or until more severe symptoms redevelop, at which time both serum IgA EMA or IgA tTG and small bowel biopsy should be performed. A total of 10 g of gluten daily for a period of six to eight weeks is usually sufficient to result in definite histologic deterioration. If serologic tests and small bowel biopsy both are negative, the patient should be monitored for signs and symptoms of celiac disease on a normal diet for at least six months, after which serologic testing should be repeated.

STOOL EXAMINATION

If malabsorption is sufficient to produce significant steatorrhea, a watery or bulky, semi-formed, light tan or grayish, malodorous, greasy-appearing stool results. Microscopic evaluation of the fat content of a stool suspension stained with Sudan III or IV after hydrolysis with glacial acetic acid and heat is a helpful test. To document steatorrhea unequivocally, the amount of fat in a three-day collection of stool may be determined quantitatively, using the reliable van de Kamer chemical method.

HEMATOLOGY AND BIOCHEMISTRY TESTS

A variety of hematologic and biochemical abnormalities may be found in persons with untreated celiac disease, including deficiencies of iron, folic acid, and vitamin D. These abnormalities reflect nutritional deficiency states secondary to enteropathy-induced malabsorption. Iron deficiency anemia is common in both children and adults with celiac disease, and combined iron and folate deficiency is characteristic, especially in children. With the exception of pregnancy, severe anemia is uncommon, usually develops

Chapter 104  Celiac Disease and Refractory Celiac Disease with extensive disease, and should raise the suspicion of a complication, such as lymphoma. The peripheral blood film might reveal target cells, siderocytes, Heinz bodies, crenated red blood cells, and Howell-Jolly bodies, which suggest splenic atrophy.81 Although they are relevant to patient evaluation and management, none of these hematologic or biochemical tests is sufficiently sensitive or specific to serve as useful screening or diagnostic tools. Similarly, although an oral d-xylose test, lactulose-mannitol test, or fecal fat evaluation may be abnormal in untreated celiac disease, they, too, lack sensitivity and specificity and so are not useful as routine investigations in suspected celiac disease. Chronically elevated serum aminotransferase levels in the range of 1.5 to 2 times normal values have been reported in 9% to 40% of patients with untreated celiac disease, and in most patients, the elevated levels resolve on a gluten-free diet.

RADIOLOGY

Barium studies of the small intestine seldom are required in evaluating patients with suspected celiac disease. Abnormal roentgen findings include dilatation of the small intestine and replacement of the normal delicate feathery mucosal pattern with either marked thickening or complete obliteration of the mucosal folds and straightening of the valvulae conniventes. Even with modern, less viscous barium preparations, flocculation, segmentation, and clumping of contrast is occasionally seen in severe cases. In patients with mild or moderate disease, the distorted mucosal pattern usually is confined to the proximal small intestine, whereas patients with severe disease may have an abnormal mucosal pattern through the entire small intestine. Excessive secretion of fluid into the proximal small intestine, coupled with defective absorption of intraluminal contents, causes dilution of the barium, resulting in decreased contrast in the distal small intestine. Small bowel studies are most useful in suggesting diagnoses other than celiac disease, including Crohn’s disease (terminal ileitis), scleroderma (hypotonicity), bacterial overgrowth (small bowel diverticulosis), or collagenous disease (bowel wall rigidity). Patients with mild celiac disease might have a normal small bowel barium study, and these studies are not as sensitive as small intestinal biopsy or serology in providing diagnostic information. Routine small bowel barium studies are unnecessary in most patients with celiac disease and are considered mainly to exclude complications, such as lymphoma, carcinoma, ulcerative jejunoileitis, or stricture. Abdominal computed tomography (CT) or magnetic resonance imaging (MRI) might provide a diagnostic clue to the presence of celiac disease or refractory celiac disease by revealing splenic atrophy, ascites, mesenteric lymphadenopathy, and the presence of cavitating mesenteric lymph nodes. A bone survey might reveal diffuse demineralization with a generalized decrease in bone density. Occasionally, the secondary effects of osteopenic bone disease are seen, including vertebral body compression fractures and pseudofractures (Milkman’s lines).

DIFFERENTIAL DIAGNOSIS The differential diagnosis of celiac disease includes other causes of malabsorption and gastrointestinal disorders that are associated with changes in proximal small intestinal morphology. In both children and adults, the high positive predictive value of a positive IgA EMA or IgA tTG test result means that celiac disease usually can be diagnosed with a

high degree of certainty from the outset, thereby avoiding the need for an in-depth evaluation of alternative diagnoses and for formal gluten challenge. Malabsorption and steatorrhea can result from pancreatic insufficiency, cholestatic liver disease, terminal ileal disease or resection, or small intestinal bacterial over­ growth. In some patients, microscopic colitis or bacterial overgrowth may be present concurrently with celiac disease. It is important to exclude these disorders in patients who do not respond to treatment with a gluten-free diet (see later).118,124 In adults, celiac disease is histologically distinguished easily from Whipple’s disease and from malabsorption because of infiltration of the mucosa with Mycobacterium avium complex. Changes in mucosal morphology can be seen in parasitic infections other than Giardia, including strongyloidiasis, coccidiosis, and hookworm disease, but these changes rarely include villus atrophy. Although villus atrophy is characteristic of untreated celiac disease, it is by no means pathognomonic and may be seen in varying degrees in a wide variety of other enteric disorders (see Fig. 104-2); thus villus atrophy on small intestinal biopsy is not in itself sufficient to diagnose celiac disease. Crypt cell activity, enterocyte characteristics, and the nature of the inflammatory infiltrate also must be examined and in some instances will point toward another diagnosis. For example, patients with hypogammaglobulinemia can have an architectural lesion that resembles celiac disease, but plasma cells may be absent or markedly diminished in the lamina propria—not increased as in celiac disease. Absence of the other histologic features of celiac disease often suggests an alternative diagnosis. After an acute viral gastroenteritis, the morphologic abnormalities may be indistinguishable from those of celiac disease. In infants and young children, cow’s milk or soy protein intolerance also can result in biopsy findings identical to those of celiac disease.125,126 Soy protein often is used as a substitute for milk protein in cow’s milk protein intolerance, but some children also develop mucosal abnormalities resembling those of celiac disease following ingestion of soy protein.126 A rare condition that can cause diagnostic confusion is collagenous sprue. Patients with collagenous sprue might present initially with symptoms and biopsy findings consistent with celiac disease, but their symptoms fail to respond to gluten withdrawal, and with time, extensive deposition of collagen in the lamina propria develops just beneath the absorptive epithelium.127 The relationship between celiac disease and both collagenous sprue and the microscopic colitides (lymphocytic and collagenous colitis) is discussed later.

DISEASES ASSOCIATED WITH CELIAC DISEASE

A large number of diseases occur more commonly among patients with celiac disease and are delineated in Table 104-3.128 In addition to an association with autoimmune disorders, some of the associated diseases also have similar HLA haplotype associations.

Dermatitis Herpetiformis

DH is a skin disease characterized by papulovesicular lesions that occur symmetrically over the extensor surfaces of the extremities and the buttocks, trunk, neck, and scalp. Unlike celiac disease, DH rarely is diagnosed in childhood and usually manifests in early or middle adult life. DH is slightly more common in men (3 : 2), but in patients younger than 20 years, women predominate (3 : 2).129 The rash is intensely pruritic, and scratching off the vesicle relieves the

1811

1812

Section X  Small and Large Intestine Table 104-3  Disorders Associated with Celiac Disease Definite Association Bird-fancier’s lung Dermatitis herpetiformis Diabetes mellitus type 1 Down syndrome Epilepsy with cerebral calcification Fibrosing alveolitis Hypothyroidism or hyperthyroidism Idiopathic pulmonary hemosiderosis Immunoglobulin A deficiency Immunoglobulin A mesangial nephropathy Inflammatory bowel disease Microscopic colitis Recurrent pericarditis Rheumatoid arthritis Sarcoidosis Possible Association Addison’s disease Autoimmune hemolytic anemia Autoimmune liver diseases Cavitary lung disease Congenital heart disease Cystic fibrosis Immune thrombocytopenic purpura Iridocyclitis or choroiditis Macroamylasemia Myasthenia gravis Polymyositis Schizophrenia Sjögren’s syndrome Systemic and cutaneous vasculitides Systemic lupus erythematosus Modified from Mulder CJ, Tytgat GN. Coeliac disease and related disorders. Neth J Med 1987;31:286.

itching; hence, intact vesicles might not be present except for the earliest lesions. The diagnosis of DH requires demonstration by immunofluorescence of granular or speckled IgA deposits in an area of perilesional skin—that is, skin close to a lesion but not affected by blistering.129 DH is associated with a mild patchy enteropathy indistinguishable from celiac disease; because it has a patch distribution, multiple intestinal biopsies may be required for diagnosis.129 DH-associated enteropathy tends to be less severe than celiac disease, and only a minority of patients has intestinal symptoms. An increased risk of intestinal malignancy has been reported, however, as in celiac disease,130,131 and most lymphomas occur in patients whose DH was not controlled by a strict gluten-free diet or in those who had been treated with a gluten-free diet for a period less than 5 years.131,132 Approximately 5% to 15% of patients with DH-like skin lesions have linear IgA deposits along the dermoepidermal junction. This condition has been termed linear IgA disease and is distinguished from DH on the basis of its unique immunofluorescent finding; the presence of circulating IgA anti–basement membrane antibody, which binds to a 97-kd protein found in normal human skin133; the absence of circulating IgA EMA or tTG; different HLA susceptibility genes; and, most important, the lack of any associated gluten-sensitive enteropathy.59,134 Sardy and colleagues shed light on the pathogenesis of DH by demonstrating that epidermal (type 3) transglutaminase (eTG) is the dominant autoantigen in DH (rather than the type 2 tTG autoantigen of celiac disease).135 This helps to explain why DH skin lesions appear in only a minority of patients having celiac disease. They also showed that the IgA precipitates in the papillary dermis of patients with DH,

the defining manifestation of the disease, contain eTG but not tTG or keratinocyte transglutaminase. In DH, the prevalences of HLA-DQ2, circulating AGA, antireticulin, and EMA parallel those observed in patients with celiac disease without DH.129 Although many patients with DH have elevated IgA tTG antibodies, confirming its pathogenic relationship with celiac disease, its prevalence in DH (75%) is lower than that found in celiac disease (95% to 98%).136 Thus, DH and celiac disease are two very closely related gluten-sensitive disorders but nonetheless distinct clinical disease entities. Most, if not all, patients with DH also have at least latent celiac disease, whereas less than 10% of patients with celiac disease have DH. Dapsone treatment at a dose of 1 to 2 mg/kg daily is effective and often diagnostic in its ability to heal the rash of DH and to relieve the pruritus rapidly, but the enteropathy associated with DH does not improve with dapsone. Six to 12 months of gluten withdrawal, however, usually reverses not only the intestinal but also the skin lesions in most patients with DH, and a strict gluten-free diet allows most patients to reduce or discontinue dapsone.137 Iodine can also exacerbate DH and should be avoided, especially in refractory cases. Patients with DH, just like those with celiac disease, can include moderate amounts of oats in their gluten-free diet without deleterious effects to their skin or intestine.138

Other Disease Associations

Autoimmune disease is strongly associated with celiac disease and has a prevalence of approximately 20% in adult patients. The strong association between celiac disease and type 1 diabetes mellitus (T1DM) reflects, in part, the increased frequency of the celiac-associated DQ alleles in patients with T1DM. The frequency of celiac disease in T1DM patients is approximately 5% (ranging from 3% to 8%)139-141 and the frequency of T1DM in celiac disease is also approximately 5%.142 Most patients with T1DM who have celiac disease are asymptomatic from the point of view of their celiac disease, but unexpected episodes of hypoglycemia or diarrhea in patients with T1DM should alert clinicians to the possibility of coexisting celiac disease. Control of diabetes in patients with celiac disease can be difficult because of varying nutrient absorption. There also is a high prevalence of autoimmune thyroid disease among patients with celiac disease, hypothyroidism being more common than hyperthyroidism.143 Celiac disease also can be associated with a variety of other autoimmune connective tissue diseases, including inflammatory bowel disease, chronic hepatitis, sclerosing cholangitis, primary biliary cirrhosis, IgA nephropathy, interstitial lung disease (including chronic fibrosing alveolitis), idiopathic pul­ monary hemosiderosis, systemic lupus erythematosus, Sjögren’s syndrome, and polymyositis.1,52,97,117,128,142,144,145 Although the relationship between celiac disease and many autoimmune disorders has been explained by the sharing of a common genetic factor, Ventura and colleagues146 suggested an increased incidence of autoimmune disease with increased age at diagnosis and lack of diet therapy. The role of a gluten-free diet in preventing the subsequent development of autoimmune disease, however, has been challenged by more recent studies.147 Although many patients with celiac disease exhibit lactose and fructose intolerances at the time of diagnosis, only a small percentage has persistent disaccharidase deficiency following gluten withdrawal. These patients experience abdominal pain and diarrhea with disaccharide intake, and diagnosis can usually be made just by history or by an appropriate hydrogen breath test. Should

Chapter 104  Celiac Disease and Refractory Celiac Disease concom­itant disaccharidase deficiency be present, the relevant disaccharide should be reduced or excluded from the diet. Selective IgA deficiency occurs in approximately 2% of celiac disease patients (20 times the population prevalence). Hyposplenism and splenic atrophy have been noted frequently in patients with celiac disease; the incidence increases with advancing age, duration of exposure to dietary gluten, and disease activity.81 The underlying mechanism is unknown, but affected patients may be at increased risk of developing bacterial infections148 and might benefit from vaccination for pneumonia. There is a well-established relationship of celiac disease with inflammatory bowel disease and the microscopic colitides (see Chapter 124).149-151 Mild to moderate small intestinal lymphocytosis and, occasionally, partial or subtotal villus atrophy are common in both lymphocytic and collagenous colitis,150 and mild colonic lymphocytosis occurs in many patients with untreated celiac disease and has improved on a gluten-free diet.151,152 Rectal gluten challenge in patients with celiac disease has been shown to induce a mild proctitis characterized by lymphocytosis of the rectal lamina propria and epithelium.153 Furthermore, a glutenfree diet has been reported to be an effective therapy in some patients with refractory collagenous colitis.154 The demonstration that patients with celiac disease and microscopic colitis share a set of predisposing HLA-DQ genes155 underscores the overlap between both diseases. Confusion also can arise in patients with refractory celiac disease, who have a higher prevalence of colonic lymphocytosis than patients with responsive celiac disease. Colonic lymphocytosis can be difficult to distinguish from lymphocytic colitis, although most colonic IELs in lymphocytic colitis are CD8+, whereas those in the colonic lymphocytosis of refractory celiac disease rarely are CD8+.124

TREATMENT GLUTEN-FREE DIET

Removal of gluten from the diet is essential for treating patients with celiac disease (Table 104-4).80 The importance of gluten withdrawal was established by Dicke’s, van de Kamer’s, and Weijers’ astute studies in the early 1950s when the toxicity of wheat protein in children with celiac disease was demonstrated.6,7 In 1962, Rubin and colleagues31 showed that instillation of wheat, barley, and rye flour into the histologically normal small intestine of persons with treated celiac disease rapidly induced celiac-like symptoms and that these symptoms were accompanied by the development of the typical celiac lesions in the exposed mucosa. Because a gluten-free diet represents a lifetime commitment for patients with celiac disease, is more expensive than a normal diet, and carries a social liability, it should not be undertaken casually or as a therapeutic trial. In reality, complete dietary elimination of all glutencontaining cereal grains is a challenge for most patients to achieve and maintain. Hidden gluten is present in a wide variety of processed foods, because wheat flour is used widely in the food industry as a thickener and inexpensive filler for many commercial products, precooked meals, and convenience foods, including ice cream, pasta, sausages, fish sticks, cheese spreads, salad dressings, soups, sauces, mixed seasonings, mincemeat for mince pies, and even some medications156 and vitamin preparations (Table 104-

Table 104-4  Principles of Initial Dietary Therapy for Patients with Celiac Disease Avoid all foods containing wheat, rye, and barley gluten. Avoid all oats initially. Avoid malt unless clearly labeled as derived from corn. Use only rice, corn, maize, buckwheat, millet, amaranth, quinoa, sorghum, potato or potato starch, soybean, tapioca, and teff, bean and nut flours. Wheat starch and products containing wheat starch should only be used if they contain less than 20 ppm gluten and are marked “gluten free.” Read all labels and study ingredients of processed foods. Beware of gluten in medications, supplements, food additives, emulsifiers, or stabilizers. Limit milk and milk products initially if there is evidence of lactose intolerance. Avoid all beers, lagers, ales, and stouts (unless labeled gluten free). Wine, most liqueurs, ciders, and spirits, including whiskey and brandy, are allowed. ppm, parts per million. Modified from Trier JS. Celiac sprue and refractory sprue. In: Feldman M, Scharschmidt BF, Sleisenger MH, editors. Gastrointestinal and Liver Disease. 6th ed. Philadelphia: WB Saunders; 1997. p 1557.

Table 104-5  Some Potential Sources of Hidden Gluten Beers, ales, other fermented beverages (distilled beverages are acceptable) Bouillon and soups Candy Communion wafers Drink mixes Gravy and sauces Herbal tea Imitation meat and seafood Lipstick and lip balms Medications (pills and capsules) Nutritional supplements Play-Doh Salad dressings and marinades Self-basting turkeys Soy sauce Toothpaste

5). Furthermore, grains that are naturally gluten-free can become contaminated with wheat, particularly when mills use the same production lines and equipment to process both gluten-containing and gluten-free products. Beers, lagers, ales, and stout should be avoided (apart from the few available specifically gluten-free products), but wines, many liqueurs, and ciders as well as spirits, including brandy, malt, and scotch whiskey, can be consumed (unless glutencontaining flavorings have been added after distillation). Helpful recipes as well as detailed instructions regarding gluten-free diets have been published in excellent, inexpensive books that are of great value to patients with celiac disease.157 National celiac societies in many countries publish regularly updated handbooks that list the available gluten-free products. Food lists are applicable for use only in the country in which they were compiled. Similar foods with well-known brand names may be made under franchise using slightly different recipes in different countries and may be gluten-free in one country and not in others. Consequently, patient education is crucial, and the institution of an effective gluten-free diet requires extensive and repeated instruction of the patient by the physician and dietitian, as well as a motivated and basically suspicious, label-reading patient. The importance of patient education

1813

1814

Section X  Small and Large Intestine Table 104-6  Key Elements in the Management of Celiac Disease Consultation with a skilled dietitian Education about the disease Lifelong adherence to a gluten-free diet Identification and treatment of nutritional deficiencies Access to an advocacy group Continuous long-term follow-up by a multidisciplinary team From the National Institutes of Health Consensus Development Conference Statement on Celiac Disease, June 28-30, 2004. Gastroenterology 2005;128:S1-S9.

and support by a multidisciplinary team of health care providers was emphasized in the recently published National Institutes of Health consensus development conference statement (Table 104-6).158 There is considerable variation among patients with celiac disease in their ability to tolerate gluten. Some patients can ingest small amounts of gluten without developing symptoms. Others are exquisitely sensitive to ingestion of even minute amounts of gluten and can develop massive watery diarrhea reminiscent of acute cholera within hours of eating very small amounts of gluten. Occasionally, the diarrhea is so severe that it can induce acute dehydration, termed gliadin shock or celiac crisis.123 Patients with untreated celiac disease might have accompanying brush border lactase deficiency secondary to damage to surface epithelial cells. Therefore, milk and dairy products should be avoided upon initiation of a gluten-free diet. After the disease responds to the diet, however, these products can be reintroduced, if they are tolerated. It is now clear that moderate amounts of oats are not toxic to the vast majority of patients with celiac disease. In a carefully conducted randomized clinical trial, adults with celiac disease who consumed 50 to 70 g of oats per day for 6 to 12 months did not differ in regard to symptoms, nutritional status, or duodenal mucosal histology compared with patients maintained on an oat-restricted, gluten-free diet159; in a smaller follow-up study, the same investigators showed no harm in celiac disease patients from five years of oat ingestion.160 Patients with DH also can include moderate amounts of oats in their gluten-free diet without deleterious effects to the skin or intestine,130 and an open-labeled study showed that a six-month trial of commercial oat breakfast cereal also was safe for children with newly diagnosed celiac disease beginning a gluten-free diet.161 It should be stressed, however, that oat products obtained from the grocery store shelf can contain significant amounts of other grains, especially wheat, because of contamination in the field, during harvesting or transport, or at the mill. Consequently, oats often are avoided initially in patients with newly diagnosed celiac disease until remission is achieved on a gluten-free diet. Subsequently, up to 2 ounces of oats per day from a reliable source may be introduced and continued if tolerated. A gluten-free diet can be rich in nutrition and conducive to excellent general health. Careful attention is needed, however, to avoid certain pitfalls of the gluten-free diet. The gluten-free diet often contains inadequate iron, calcium, vitamin D, and B vitamins.162 Hence nutritional counseling and monitoring together with a daily gluten-free multivitamin supplement is recommended to avoid deficiency states. Avoiding wheat, barley, and rye often leads to inadequate fiber intake and constipation unless steps are taken to replace these with other sources of dietary fiber. Some weight gain is common after starting a gluten-free diet as malabsorption resolves; however, excess weight gain or obesity can easily ensue, especially if fat- and calorie-

rich gluten-free processed foods and snacks are consumed in excess. Preparing meals using fresh ingredients including gluten-free grains and reducing the use of prepared or processed foods are often the keys to a healthy glutenfree diet. After starting a gluten-free diet, most patients improve within a few weeks. In many, substantial symptomatic improvement is noticed within 48 hours, although it can take weeks or months to achieve full clinical remission. Pink and Creamer163 reported that 70% of patients with celiac disease who started a gluten-free diet returned quickly to normal health and reported improvement in their symptoms within two weeks. The speed and eventual degree of histologic improvement are unpredictable, but they invariably lag behind the clinical response. Although an increase in enterocyte height may be evident within a week of gluten withdrawal, the return of villus architecture toward normal takes considerably longer and might not be evident on rebiopsy for two or three months. In some patients, histologic resolution can take up to two years; the main reason for this slow or partial recovery is inadvertent exposure to gluten.31 Although a return to normal is common in children, in approximately 50% of adults on a gluten-free diet, biopsies show only partial improvement; the less severely damaged distal intestine recovers more rapidly than the maximally damaged proximal intestine.164 The investigation and management of patients who have celiac disease and whose symptoms or signs do not respond to dietary gluten avoidance are discussed later (see the discussions of nonresponsive and refractory celiac disease).

DIETARY SUPPLEMENTS

In addition to the gluten-free diet, patients with newly diagnosed celiac disease should receive appropriate supplemental therapy to help correct nutritional deficiencies caused by malabsorption; iron deficiency is the most common. Deficiencies of vitamin D, vitamin B12, or folic acid are not uncommon. Patients with purpura, bruising, or other evidence of bleeding might have prolongation of their prothrombin time and require supplemental vitamin K. Patients with severe diarrhea and dehydration require vigorous intravenous replacement of fluids and electrolytes. Intravenous calcium gluconate (1 to 2 g) should be administered promptly to a patient who has tetany. If there is no response, the tetany may be caused by hypomagnesemia and require magnesium replacement. The risks of osteopenia and osteoporosis should be explained to all patients with celiac disease and general advice should be given about weight-bearing exercises, dietary calcium and vitamin D intake, and the adverse effects of smoking and excessive alcohol consumption. There is compelling evidence that strict adherence to a gluten-free diet protects against further bone loss and initially is associated with an increase in bone mineral density (BMD).165 A total daily calcium intake of 1500 mg should be ensured—a cup of skim milk provides 300 mg. If dietary calcium is inadequate, 500 to 1500 mg of supplemental calcium should be given. Vitamin D deficiency should be sought and treated (particularly in patients with significant steatorrhea), until the malabsorption has responded to gluten withdrawal, to prevent mobilization of skeletal calcium. One year of gluten withdrawal has been shown to reverse osteopenia in most patients, including postmenopausal women and patients with incomplete mucosal recovery,166 but patients who have secondary hyperparathyroidism at the time celiac disease is diagnosed tend to have more

Chapter 104  Celiac Disease and Refractory Celiac Disease refractory osteopenia, and their BMD might not normalize even after several years of gluten withdrawal.167 Clinicians often rely on serum calcium, phosphate, and alkaline phosphatase measurements, but osteomalacia still can exist even if these tests are normal. If these tests are normal and osteomalacia still is suspected, a serum 25-hydroxy vitamin D level can be determined; this test is expensive, however, and the less expensive parathormone assay may be considered. A low-normal calcium and an elevated parathormone level indicates secondary hyperparathyroidism, and calcium (500 to 1000 mg daily) together with vitamin D (400 to 800 units daily) should be given.168 The serum calcium level must be monitored and supplementation promptly discontinued if hypercalcemia develops. All patients should have their BMD measured one year after diagnosis. Recent guidelines suggest that patients with treated celiac disease who have progressive osteoporosis despite correction of calcium and vitamin D deficiency should be offered oral biphosphonate therapy and have their BMD checked every one to two years.168 Vitamin A, thiamine, riboflavin, niacin, pyridoxine, vitamin C, and vitamin E, all in the form of a multivitamin preparation, probably should be taken daily by all patients with celiac disease in light of their predisposition to malabsorption and the risks for nutritional deficiencies associated with a gluten-free diet.162 Some patients have reported symptomatic improvement with correction of magnesium, copper, and zinc deficiencies.169 Drugs, like nutrients, may be absorbed capriciously by patients with severe celiac disease. Medications considered essential for the patient’s well-being might need to be administered parenterally or monitored closely until absorption improves in response to treatment with a glutenfree diet.

GLUCOCORTICOIDS

In vitro studies have shown that the addition of glucocorticoids prevented the harmful effects of gluten on biopsy specimens from patients with celiac disease.170 Although celiac disease can be treated with glucocorticoids, with rapid improvement in symptoms, the effect rarely persists once treatment is stopped, and significant side effects are common.171 Therefore, glucocorticoids are not indicated in the routine management of celiac disease but are reserved for severely ill patients who present with acute celiac crisis manifested by severe diarrhea, dehydration, weight loss, acidosis, hypocalcemia, and hypoproteinemia.172 These few patients might benefit from a short course of glucocorticoids until the gluten-free diet takes effect. A brief course of glucocorticoids also may be used in the rare instances of gliadin shock that occurs occasionally in treated patients who are subject to gluten challenge.123

NONRESPONSIVE CELIAC DISEASE Nonresponsive celiac disease (NRCD) is a clinical diagnosis defined by the persistence of symptoms, signs, or laboratory abnormalities typical of celiac disease despite adherence to a gluten-free diet for at least six months.118,173,174 Ten percent of patients with celiac disease are nonresponsive either immediately after the initial diagnosis (primary NRCD) or following a period of response to the gluten-free diet (secondary NRCD).173 An approach to the evaluation of NRCD that is based on the early identification and correction of common causes and that culminates in the diagnosis of

Review initial diagnosis of celiac disease Confirmed

Search for gluten exposure by history, nutritionist evaluation, IgA tTG antibody Weight loss present

Dietary treatment

No gluten exposure Trial of lactose restriction if primary NRCD No response to GFD/LFD EGD with biopsy of SI (plus colonoscopy, if diarrhea prominent) Normal or near normal SI histology

Persistently abnormal SI histology

Consider: IBS Microscopic colitis SIBO Eating disorders Food intolerances Peptic ulcer disease Gastroparesis Pancreatic insufficiency

Consider: SIBO Peptic duodenitis Immunodeficiency states including CVID Crohn’s disease Tropical sprue Giardiasis Post-gastroenteritis Eosinophilic gastroenteritis Auto-immune enteropathy Excluded Refractory sprue

Consider: Imaging studies (SBFT, enteroscopy, capsule endoscopy, CT) to evaluate for UJ or EATL T-cell clonality and aberrant IEL markers to differentiate type I and type II refractory celiac disease Figure 104-7.  Diagnostic approach in patients with nonresponsive celiac disease. CT, computed tomography; CVID, common variable immune deficiency; EATL, enteropathy-associated T-cell lymphoma; EGD, esophago­ gastroduodeno­scopy; GFD, gluten-free diet; IBS, irritable bowel syndrome; IEL, intra-epithelial lymphocyte; LFD, lactose-free diet; NRCD, nonresponsive celiac disease; SBFT, small bowel follow-through; SI, small intestine; SIBO, small intestinal bacterial overgrowth; tTG, tissue transglutaminase; UJ, ulcerative jejunitis. (From Abdallah H, Leffler D, Dennis M, Kelly CP. Refractory celiac disease. Curr Gastroenterol Rep 2007; 9:401-405.)

refractory celiac disease (which affects 1% of patients with celiac disease) is outlined in Figure 104-7.118 The first step in evaluating NRCD is to carefully review the primary diagnostic studies, because if the diagnosis of celiac disease is mistaken, a sustained response to the gluten-free diet cannot be expected. This is especially the case for patients who test negative for anti-tTG and EMA at first presentation. In such circumstances, careful review of the biopsy pathology by an expert gastrointestinal patholo-

1815

1816

Section X  Small and Large Intestine gist is warranted to seek alternative diagnoses. Special testing to exclude hypogammaglobulinemia, measurement of DGP, antibodies, and HLA DQ2 and DQ8 typing also may be helpful. The single most common cause for NRCD is continued gluten ingestion, which is often inadvertent and occult. A persisting elevation of anti-tTG is strongly associated with ongoing gluten exposure.173 Intolerance to disaccharides (e.g., lactose, fructose) also is common, especially in primary NRCD. Thus, evaluation by an expert dietitian to seek ingestion of hidden gluten and intolerance of disaccharides is an essential next step.118 If no dietary causes can be identified, a small bowel biopsy should be repeated and the findings compared with the initial pretreatment biopsy. If the enteropathy has healed or is substantially improved, diagnostic considerations for ongoing symptoms and signs include irritable bowel syndrome (22%), small intestinal bacterial overgrowth (6%), other food allergies and intolerances (1%), and pancreatic insufficiency.173,174 If diarrhea is a prominent symptom, colonic biopsy also should be obtained and examined for microscopic colitis (6%). If repeat biopsies show ongoing

changes consistent with active celiac disease, refractory celiac disease becomes likely. Other causes of a celiac-like enteropathy, however, should again be considered, including small bowel bacterial overgrowth, peptic duodenitis, hypogammaglobulinemia, tropical sprue, intestinal infections (e.g., giardiasis), Crohn’s disease, and autoimmune enteropathy.173,174 Patients with NRCD who show substantial weight loss are at a significantly greater risk for refractory celiac disease.173

REFRACTORY CELIAC DISEASE Refractory celiac disease, also known as unclassified or intractable celiac disease, is defined as symptomatic, severe small intestinal villus atrophy that mimics celiac disease but does not respond to at least six months of a strict glutenfree diet and is not accounted for by other causes of villus atrophy or overt intestinal lymphoma.1,3,97 Refractory celiac disease is uncommon in adults, extremely rare in children, and largely a diagnosis of exclusion (Figs. 104-7 and 104-8).

Clinical features and biopsy findings consistent with celiac disease

Response to gluten-free diet

No recurrence of symptoms

Recurrence of symptoms

No response to gluten-free diet

Confirm strict gluten-free diet Exclude inadvertent gluten ingestion Exclude other causes of villus atrophy

Overt lymphoma or carcinoma? No Refractory disease

Response to glucocorticoids

No response to glucocorticoids

Immunostain small intestinal biopsy samples and extract tissue DNA for PCR testing No aberrant IELs No TCR γ rearrangement

Aberrant IELs Clonal TCR γ rearrangement

Cryptic EATL

Low probability of evolution to lymphoma, ulcerative jejunoileitis, or collagenous sprue

High probability of evolution to lymphoma, ulcerative jejunoileitis, or collagenous sprue

Adjuvant immunosuppressive therapy (e.g., azathioprine, cyclosporine, infliximab)

Prognosis poor Novel immunotherapy is possible in the future

Figure 104-8.  Approach to the diagnosis and management of refractory celiac disease. DNA, deoxyribonucleic acid; EATL, enteropathy-associated T-cell lymphoma; IEL, intraepithelial lymphocyte; PCR, polymerase chain reaction; TCR, T-cell receptor.

Chapter 104  Celiac Disease and Refractory Celiac Disease Symptoms can persist in treated celiac disease patients for a variety of reasons as described in the section on nonresponsive celiac disease.118,173-175

ULCERATIVE JEJUNOILEITIS

Ulcerative jejunoileitis, also known as chronic, nongranulomatous ulcerative enterocolitis or nongranulomatous jejunitis, is a rare but serious complication of celiac disease characterized by ulceration and strictures of the small intestine (see Chapter 115). Whether ulcerative jejunoileitis truly is a discrete entity has been questioned, because lymphoma ultimately is diagnosed in many of these patients.176 Indeed, ulcerative jejunoileitis in association with EATL previously was designated malignant histiocytosis. Ulcerative jejunoileitis should be suspected in patients with celiac disease who present with weight loss, abdominal pain, and diarrhea that do not respond to a gluten-free diet. Typically, patients also experience recurrent episodes of intestinal ulceration and obstruction with gradual weight loss despite surgery and strict adherence to a gluten-free diet. Areas of intestinal ulceration and stricture formation typically cause hemorrhage and obstruction (Fig. 104-9); perforation with peritonitis also can occur. Diagnosis is made by enteroscopy, contrast studies of the small intestine, abdominal CT, capsule endoscopy, or laparotomy. Some patients respond to a gluten-free diet, but surgical excision of the worst affected segments of small intestine so far has proved to be the most effective treatment. There is a high risk for transition to diffuse or multifocal EATL, but in a few patients with well-documented celiac disease and localized jejunoileitis, no evidence of malignant disease develops and there is a response to either surgical resection or therapy with glucocorticoids and azathioprine.177 Even in the absence of overt malignant transformation, however, the five-year survival rate for patients with ulcerative jejuno­ ileitis is less than 50%.

COLLAGENOUS SPRUE

Collagenous disease is characterized by the development of a subepithelial collagen band thicker than 10 mm in the small intestine. Although collagenous disease has been regarded as an entity distinct from celiac disease,127 deposition of collagen under the intestinal epithelial cells has been noted in up to 36% of patients with classic celiac disease.178 Furthermore, there are several reports of patients with collagenous disease who have EMA179 and complications of refractory celiac disease, specifically ulcerative jejuno­ ileitis180 and lymphoma.15 Although collagenous disease often is refractory to therapy, the presence of subepithelial collagen does not, a priori, preclude a successful response to gluten withdrawal.178,181 Collagenous disease should be distinguished from collagenous colitis, which rarely accompanies celiac disease and should be considered in the differential diagnosis of nonresponsive celiac disease.154 Compared with both celiac disease and collagenous colitis, the prognosis in collagenous disease is grim, with most reported patients dying from the disease.

TREATMENT

In patients with celiac disease and no demonstrable cause for lack of response to a gluten-free diet, a variety of treatments (based mostly on small, uncontrolled studies) have been described; these include glucocorticoids, immunosuppressive drugs, elimination diets, and dietary supplementation with zinc and copper.182-186 Evidence supporting the use of immunosuppressive therapy in the treatment of refractory celiac disease is based mainly on anecdotal reports; to date, no controlled trials

A

B Figure 104-9.  A small bowel barium contrast study and corresponding histopathology from a patient with ulcerative jejunoileitis complicating celiac disease. A, A segmental area of fixed narrowing with associated mucosal distortion and ulceration in the distal jejunum and proximal ileum is seen. B, Histology of a segment of resected small intestine showing ulcerated mucosa with adjacent diffuse villus atrophy and lymphocytic infiltrate consistent with celiac disease. Lymphocytes within the epithelium and lamina propria were positive for T-cell antigen (CD3); no overt lymphoma was evident, but Southern blot analysis revealed clonal T-cell receptor gene rearrangements in both the involved and uninvolved small intestine and an adjacent mesenteric lymph node consistent with cryptic enteropathy-associated T-cell lymphoma.

have been performed.185 Glucocorticoid treatment also may be necessary in patients with refractory celiac disease (see later). Azathioprine or 6-mercaptopurine may be used as a glucocorticoid-sparing agent if a dose of 10 mg of prednisolone or more per day is required to keep the condition under control.182 In one open pilot study, 13 adult patients with refractory celiac disease were treated for two months with oral cyclosporine, in doses titrated to achieve serum levels of 100 to 200 ng/mL; small intestinal histology improved in eight patients (61%), and villi normalized in five (38%).186 Cyclosporine therapy has been reported to be lifesaving in occasional patients with refractory celiac disease–like disease, and it can result in reversal of glucocorticoid resistance, but its efficacy remains unproved.183 There also has been a report on the efficacy of infliximab, a chimeric antibody to TNF-α, in refractory celiac disease.184

1817

1818

Section X  Small and Large Intestine In a study of patients with refractory celiac disease, 16 of 29 (55%) responded fully to treatment with oral entericrelease budesonide (Entocort, 9 mg daily), and an additional six (21%) showed a partial response.187 Improvements were evident in patients with coexisting microscopic colitis and in those without. Oral budesonide was well tolerated, and no serious side effects were reported. Because of its good side-effect profile, oral budesonide is fast becoming a drug of first choice for treatment of refractory celiac disease. Although a trial of immunosuppressive therapy is worth considering in all patients with refractory celiac disease, caution must be used, because these patients often are malnourished and hyposplenic; hence they may be prone to opportunistic infections. Unfortunately, some patients have only partial or no response to immunosuppressive therapy, and the clinical course is characterized by progressive malabsorption necessitating total parenteral nutrition.

COMPLICATIONS

It has long been appreciated that patients with refractory celiac disease are at high risk for developing fatal complications such as ulcerative jejunoileitis and lymphoma. Until recently, the precise link between refractory celiac disease and these complications as well as between refractory celiac disease and celiac disease remained controversial. The spectrum of autoimmune enteropathy was implicated in a handful of adult refractory celiac disease patients by the presence of antienterocyte antibodies.188,189 It is now becoming clear, however, that refractory celiac disease, EATL, and ulcerative jejunoileitis represent a heterogeneous, but related, group of clinical conditions at the extreme end of the celiac disease spectrum. Moreover, there is now a growing realization that many of these patients have a cryptic intestinal T-cell lymphoma, characterized by phenotypically abnormal IELs that have monoclonal rearrangements of the TCR γ gene.15 Early immunophenotypical studies demonstrated that the normal cell counterpart of EATL was the IEL.190 It was not until 1995, however, that Murray and colleagues191 made the remarkable observation that in patients with overt EATL, lymphocytes from adjacent nonlymphomatous mucosa contained the identical monoclonal TCR gene rearrangement as the overt lymphoma and coined the term cryptic intestinal T-cell lymphoma. Ashton-Key and colleagues192 later confirmed this finding and showed that both the inflammatory ulcers and the intact (nonlymphomatous) mucosa in cases of ulcerative jejunoileitis harbored a monoclonal T-cell population and that the lymphomas that developed in these patients consisted of the identical T-cell clone. Cellier and colleagues68 showed that the IELs in refractory celiac disease patients are abnormal in that they lack expression of CD8, which is consistently found on most normal or celiac disease IELs. Subsequent work confirmed this finding and showed that the abnormal IELs in ulcerative jejunoileitis and nonlymphomatous mucosa in EATL shared not only the genotype but also the immunophenotype of the lymphoma.193 Cellier and colleagues15 detected aberrant clonal IELs (similar to those in most cases of EATL) in 16 (84%) of 19 patients with refractory celiac disease (type II refractory celiac disease). Of these 19, 7 (37%) had collagenous disease, 6 (32%) had ulcerative jejunoileitis, 6 (32%) had mesenteric lymph node cavitation, and 3 (16%) developed overt EATL that was clonally identical to the IELs of the preexisting refractory celiac disease. The three patients (16%) without aberrant clonal IELs (type I refractory celiac disease) all made a complete clinical and histologic recovery with glucocorticoid therapy plus a gluten-free diet.

Thus, the cumulative evidence now points to type II refractory celiac disease being a manifestation of an aberrant clonal IEL-mediated neoplastic process. These cells have destructive properties, possible related to their cytotoxic phenotype,194 which leads to mucosal ulceration and lymph node cavitation, and they sometimes, but not always, undergo further molecular and clinical progression to lymphoma. As noted earlier, the proinflammatory cytokine IL-15 is massively increased in the intestine of patients with refractory celiac disease. IL-15 induces IEL secretion of IFN-γ and increases IEL cytotoxicity against epithelial cells, thereby favoring the severe enteropathy characteristic of type II refractory celiac disease.16 Increasing evidence suggests that IL-15, through its key role in modulating IEL homeostasis, ultimately might lead to lymphomatous transformation because IL-15 provides signals mandatory for the survival or expansion of the abnormal clonal IELs. Based on this evidence, patients with suspected refractory celiac disease should have TCR and monoclonal antibody studies performed on their small intestinal biopsy samples. In the future, early recognition of the malignant potential of the intestinal infiltrate in type II refractory celiac disease might permit curative surgery or chemotherapy. Alternatively, novel anti–IL-15 biological agents might prevent the progression of refractory celiac disease and the development of lymphoma.

CELIAC DISEASE AND MALIGNANCY In the past, patients with celiac disease or DH had been reported to have a 10-fold increased risk for certain gastrointestinal tract malignancies and a 40- to 70-fold increased risk for non-Hodgkin’s lymphoma (NHL).130,195 Recent studies, however, indicate that the increase in risk of malignancy, particularly lymphoma, is much less than initially believed. A large retrospective Swedish study followed almost 12,000 hospitalized patients who had either celiac disease or DH between 1964 and 1994, with a mean followup of 10 years.131 The overall cancer and malignancy lymphoma risks were increased only modestly (standardized incidence ratio [SIR], 1.3; 95% CI: 1.2-1.5; and SIR, 5.9; 95% CI: 4.3-7.9, respectively). In a prospective Italian study, patients with celiac disease had a 3.1-fold increased risk of NHL.196 Conversely, the prospective BioMed European Working Group on Celiac Disease and Malignancy, which reviewed data from 10 countries, reported that the prevalence of celiac disease was increased 2.6-fold in 1446 patients with NHL compared with 9659 control subjects.197 The European study suggested that the risk of NHL is even less evident in silent celiac disease. Small intestinal lymphoma, often multifocal and diffuse, accounts for one half to two thirds of the malignancies complicating celiac disease and typically occurs after 20 to 40 years of disease (see Chapters 29 and 121).131,195 Whereas in the general population, most small intestinal lymphomas are of B-cell origin, intestinal lymphoma in celiac disease is typically of T-cell origin, and the term EATL (enteropathy-associated T-cell lymphoma) was coined to describe both the intestinal and extraintestinal lymphomas that complicate celiac disease. The European multicenter study indicated that intestinal T-cell lymphomas, chiefly EATL, are highly characteristic, occurring almost exclusively in those with celiac disease (odds ratio of 28 compared to population controls). The clinical onset of EATL may be insidious, and its initial presentation and small bowel biopsy appearance can

Chapter 104  Celiac Disease and Refractory Celiac Disease mimic those of untreated celiac disease. EATL commonly is accompanied by mucosal ulceration, as seen in ulcerative jejunoileitis, and these ulcers sometimes are the only endoscopic manifestation of lymphoma (see Chapters 29, 121). Although some patients with EATL have a partial or temporary response to a strict gluten-free diet, most are eventually unresponsive to gluten withdrawal. In patients whose disease was previously controlled on a gluten-free diet, recurrence of gastrointestinal symptoms (e.g., abdominal pain, weight loss, diarrhea) should raise the clinical suspicion of lymphoma. In some patients with lymphoma, mucosal histology adjacent to and distant from the lymphoma is indistinguishable from that of untreated celiac disease, yet the patient’s symptoms do not respond to gluten withdrawal.190 There is long-standing controversy whether such patients had latent celiac disease that became evident after lymphoma developed, refractory celiac disease complicated by lymphoma, or refractory enteropathy induced by primary intestinal T-cell lymphoma and indistinguishable by histologic criteria from celiac disease.15 Molecular and immunohistochemical studies that have advanced our understanding of the relationships among celiac disease, refractory celiac disease, and EATL are discussed in the section on refractory celiac disease. Other features suggesting lymphoma include intestinal obstruction, intestinal bleeding, fever, hypoalbuminemia, lymphadenopathy, and erythrophagocytosis evident in bone marrow or the peripheral blood. Small intestinal radiology, enteroscopy with biopsy of the mucosa at multiple levels, capsule endoscopy, and CT or MR scanning may be helpful. Mesenteric lymphadenopathy with central cavitation has been described in celiac disease, both with198 and without199 lymphoma. If the index of suspicion is high and studies are not diagnostic, full-thickness biopsy specimens of the small intestine should be obtained at laparoscopy or laparotomy with careful examination of the entire length of the small intestine and examination of mesenteric lymph nodes. Even with such an aggressive approach, EATL can be extremely difficult to diagnose. EATL commonly is fatal: Overall one-year and five-year survival rates of 31% and 11%, respectively, were reported in one small series, with long-term survival almost exclusively confined to those treated with chemotherapy.200 Carcinoma, particularly of the oropharynx, esophagus, and small intestine, account for one third of the remaining malignancies complicating celiac disease. The average patient so affected is older than 50 years. The Swedish study reported elevated risks for small intestinal cancer (SIR, 10), oropharyngeal cancer (SIR, 2.3), esophageal cancer (SIR, 4.2), and primary liver cancer (SIR, 2.7).131 Patients with DH also had a slightly increased overall cancer risk (SIR, 1.2) owing to excesses of lymphoma and leukemia, but they had no increases in gastrointestinal carcinomas.131 The mechanisms responsible for the increased prevalence of malignancy in celiac disease are unknown. Increased crypt mitotic activity, increased turnover of lymphoid cells within the mucosa, penetration of the damaged intestinal mucosa by carcinogens, infection with oncogenic viruses, and underlying abnormalities in the mucosal immune system and surface epithelium all are potential factors. In the Swedish study, the excess risk of malignancies, which was confined to adults, disappeared after a 10-year followup.131 This declining risk of malignancies with increased duration of follow-up, and thus with the length of glutenfree diet, supports the results of a previous study, which indicated that a strict gluten-free diet for five years reduced the risk of all malignancies, not just EATL, to that of the general population.130

PROGNOSIS Celiac disease has an excellent prognosis if it is diagnosed early and the patient adheres to a lifelong gluten-free diet. Conversely, if it is not recognized and properly treated, patients can develop marked malnutrition and debilitation and can die of complications such as intercurrent infection, hemorrhage, or malignancy. Although earlier studies reported 1.9-fold201 and 3.4-fold202 increases in mortality, these studies included patients who were not adhering to a gluten-free diet as well as patients with refractory celiac disease and intestinal lymphoma. A study of 335 adults with celiac disease from Finland, at least 83% of whom adhered strictly to a gluten-free diet, showed that the five-year survival was comparable with that of the general population.142 Growth and development in infants and children with celiac disease proceed normally following gluten withdrawal. In adults, absorptive functions usually return, and many of the manifestations of disease disappear after a gluten-free diet is initiated. Certain complications, however, such as peripheral neuropathy, ataxia, or pathologic fractures secondary to severe osteopenic bone disease, particularly in the setting of secondary hyperparathy­ roidism, might not be completely reversible. The potential protective effect of gluten-free diet against the development of other autoimmune diseases and the functional consequences of a gluten-free diet in silent celiac disease remain unknown. Several lines of evidence, gathered during recent years, suggest that celiac disease is not always a lifelong condition. First, the long-term follow-up of children with proven celiac disease shows that 10% to 20% develop latent celiac disease and become tolerant (defined on clinical, biological, and histologic grounds) to gluten during adolescence. Second, it also has been shown, in individual cases, that the mucosal lesions typical of the disease can appear de novo during adulthood.203 The factors leading to the appearance or disappearance of gluten-sensitive enteropathy, however, are still unknown. Although adolescent patients might stray from their gluten-free diet, often without apparent ill effects, their inability to tolerate gluten remains, and many asymptomatic adolescent patients can be shown to have persistent hematologic, biochemical, and morphologic abnormalities.22,85 If gluten ingestion continues into adult life, most patients with celiac disease eventually develop recurrent clinical evidence of celiac disease. Therefore, patients with unequivocal evidence of celiac disease in childhood should be encouraged to remain on a gluten-free diet indefinitely if recurrent clinical disease is to be avoided during adult life.

FUTURE THERAPIES Improved knowledge of celiac disease epidemiology and pathogenesis has encouraged the search for alternatives to the gluten-free diet. Shan and colleagues204 showed that treating gliadin peptides with a combination of digestive enzymes in conditions mimicking the in vivo situation released a protease-resistant 33-amino-acid peptide (33mer) encompassing a cluster of three immunodominant T-cell epitopes. The resistance of this peptide to human digestive enzymes was ascribed to its high proline content. Rapid hydrolysis, however, was achieved in the presence of a bacterial prolyl-endopeptidase.204 Oral glutenases from bacterial or cereal sources are now being studied for their potential therapeutic use in cleaving toxic gliadin peptides

1819

1820

Section X  Small and Large Intestine within the stomach and proximal small intestine, thereby potentially abrogating gluten toxicity.205 An increase in intestinal permeability is a welldocumented feature of celiac disease and can facilitate the passage of gluten across the epithelial layer to be taken up by antigen-presenting cells and activate gliadin-specific T cells. Larazotide acetate is an octapeptide inhibitor of paracellular permeability whose structure is derived from a protein (zonula occludens toxin) secreted by Vibrio cholerae.206 In early clinical trials, larazotide acetate showed promise in preventing symptoms and signs of celiac disease activity during gluten challenge.206 Genetic modification of wheat to delete toxic peptides has been proposed to prevent the activation of celiac disease. This approach is complicated by the large number of T-cell epitopes and by the complexity of wheat genetics. Furthermore, nature has already provided a wealth of nontoxic cereal alternatives. An alternative strategy might be to develop peptide analogs capable of interfering with HLA-DQ binding and T-cell activation to redirect the immune response toward tolerance. In mice, a vaccine based on the intranasal administration of whole gliadin or of one of its isoforms partially inhibited the systemic T-cell response to parenteral challenge by whole gliadin.207 This approach is not easy to transpose into humans, however, because of the risk of enhancing immunization instead of promoting tolerance. As mentioned earlier, tTG modification of gliadin peptides greatly increases their immunogenicity and toxicity in celiac disease.208 Thus, tTG inhibitors are also being evaluated for their ability to reduce gluten toxicity. Another strategy, suggested by Maiuri and colleagues,38 is blockade of signals derived from the cytokine IL-15. This proposal may be premature for uncomplicated celiac disease in the absence of data concerning the consequence of IL-15 blockade in vivo in humans. Blocking IL-15 and its signals, however, is an attractive possibility in refractory celiac disease when patients have become unresponsive to the gluten-free diet and do not experience a response to conventional anti-inflammatory therapy. Because celiac disease is, in most cases, a benign disease effectively treated by a safe, established diet, any alternative treatments that emerge must meet high standards of efficacy

and safety. That stated, many patients with celiac disease would welcome the development of agents that could reduce the daily burden of strict gluten avoidance and also improve the control of their disease.209,210

KEY REFERENCES

Askling J, Linet M, Gridley G, et al. Cancer incidence in a populationbased cohort of individuals hospitalized with celiac disease or dermatitis herpetiformis. Gastroenterology 2002; 123:1428. (Ref 131.) Bernstein CN, Leslie WD, Leboff MS. AGA medical position statement: Guidelines on osteoporosis in gastrointestinal diseases. Gastroenterology 2003; 124:791. (Ref 168.) Cellier C, Delabesse E, Helmer C, et al. Refractory sprue, coeliac disease, and enteropathy-associated T-cell lymphoma. French Coeliac Disease Study Group. Lancet 2000; 356:203. (Ref 15.) Dicke W. Coeliac disease: Investigation of harmful effects of certain types of cereal on patients with coeliac disease. Utrecht: University of Utrecht; 1950. (Ref 6.) Dieterich W, Ehnis T, Bauer M, et al. Identification of tissue transglutaminase as the autoantigen of celiac disease. Nat Med 1997; 3:797. (Ref 12.) Fasano A, Berti I, Gerarduzzi T, et al. Prevalence of celiac disease in at-risk and not-at-risk groups in the United States: A large multicenter study. Arch Intern Med 2003; 163:286. (Ref 17.) Green PH, Cellier C. Celiac disease. N Engl J Med 2007; 357:1731-43. (Ref 97.) Leffler DA, Dennis M, Hyett B, et al. Etiologies and predictors of diagnosis in nonresponsive celiac disease. Clin Gastroenterol Hepatol 2007; 5:445-50. (Ref 173.) Lundin KE, Scott H, Hansen T, et al. Gliadin-specific, HLADQ(α1*0501,β1*0201) restricted T cells isolated from the small intestinal mucosa of celiac disease patients. J Exp Med 1993; 178:187. (Ref 11.) Molberg O, McAdam SN, Korner R, et al. Tissue transglutaminase selectively modifies gliadin peptides that are recognized by gut-derived T cells in celiac disease. Nat Med 1998; 4:713. (Ref 13.) National Institutes of Health Consensus Development Conference Statement on Celiac Disease, June 28-30, 2004. Gastroenterology 2005; 128:S1-S9. (Ref 158.) Schuppan D. Current concepts of celiac disease pathogenesis. Gastroenterology 2000; 119:234. (Ref 43.) Shan L, Molberg O, Parrot I, et al. Structural basis for gluten intolerance in celiac sprue. Science 2002; 297:2218. (Ref 204.) Sollid LM, Khosla C. Future therapeutic options for celiac disease. Nat Clin Pract Gastroenterol Hepatol 2005; 2(3):140-7. (Ref 18.) Stern M. Comparative evaluation of serologic tests for celiac disease: A European initiative toward standardization. J Pediatr Gastroenterol Nutr 2000; 31:513. (Ref 98.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

105 Tropical Diarrhea and Malabsorption B. S. Ramakrishna

CHAPTER OUTLINE Infectious Diarrhea in the Tropics  1821 Tropical Sprue  1822 Definition  1822 History  1822 Epidemiology  1823 Etiology  1823 Clinical Features  1824 Histopathology  1825 Pathophysiology  1825 Diagnosis  1826 Treatment  1827 Tropical Enteropathy and Its Distinction from Tropical Sprue  1828 Protozoan Infections That Cause Malabsorption  1828

Diarrhea and malabsorption are common in the tropics and most often result from infectious causes. Enteric infections that cause diarrhea are common in tropical countries, a result both of deficient sanitation and the ambient temperature that fosters proliferation of infectious organisms in water and food. Tropical diarrhea affects not only the indigenous population but also visitors to the tropics,1 who are more susceptible than residents to symptomatic enteric infection and who commonly develop diarrhea either during or shortly after their visit. It has been shown repeatedly that a significant number of asymptomatic and apparently healthy residents in the tropics transiently harbor pathogenic microbes in their gastrointestinal tracts. Up to 15% of the rural population in some countries is infected at any given time by a variety of pathogens, such as Campylobacter, Cryptosporidium, and Giardia species.2-4 By and large, however, a homeostasis is achieved that is perturbed when the person is infected with an overwhelming number of pathogens or when there is impairment of intestinal immunity; this results in diarrheal illness. Residents of the tropics, possibly by virtue of repeated exposure to these or related pathogens, remain asymptomatic, whereas visitors to the tropics quickly develop symptoms upon exposure. The predominant cause of malabsorption varies with geographic location in the tropics.5-7 In many areas, tropical sprue remains the major cause of malabsorption in adults and a lesser cause of malabsorption in children. Parasitic infections of the small intestine are probably the next most common cause of chronic diarrhea and malabsorption in many areas. In belts of the tropics, celiac disease is now becoming an increasingly diagnosed and appreciated problem and needs to be considered in the differential diagnosis of malabsorption. Intestinal tuberculosis, small intes-

Giardiasis  1828 Other Parasitic Infections  1829 Helminthic Infections That Cause Malabsorption  1829 HIV Infection and AIDS  1830 Intestinal Tuberculosis  1830 Crohn’s Disease  1830 Celiac Disease  1830 Primary Immunodeficiency Syndromes  1830 Immunoproliferative Small Intestinal Disease and Small Bowel Lymphoma  1830 Tropical Pancreatitis and Malabsorption  1831 Approach to the Patient with Suspected Tropical Malabsorption  1831

tinal bacterial overgrowth, and pancreatic steatorrhea are other significant causes of malabsorption in the tropics. This chapter deals mainly with tropical sprue and tropical enteropathy, but it discusses very briefly other specific causes of diarrhea and malabsorption in the tropics.

INFECTIOUS DIARRHEA IN THE TROPICS Acute and chronic infectious diarrhea in the tropics is caused by a variety of bacterial, viral, and parasitic agents (Table 105-1). Although these pathogens affect the indigenous population of the tropics, many of these infected persons remain asymptomatic, probably because of immunity acquired by earlier exposures to the same or related infectious agents. The risk of diarrhea in visitors to the tropics can range up to 55%, depending on their specific travel destination.8 Cholera is the most dramatic form of acute diarrhea, resulting in death from dehydration and electrolyte imbalance, if untreated. The disease occurs in both endemic and epidemic form in the tropics and mainly affects the indigenous population. Only very occasionally does cholera afflict Western travelers.9,10 Cholera is endemic in the Indian subcontinent, particularly in the southern and eastern parts; in the Indonesian islands; in the Philippines; and also in Latin America. In a study of 17,353 returned travelers from tropical countries, acute diarrhea occurred in 22.2% and chronic diarrhea in 11.3%.9 Parasitic diarrhea (giardiasis and amebiasis) was most common overall, whereas bacterial diarrhea (Campylobacter > Shigella > nontyphoidal Salmonella) was more common in travelers to Southeast Asian countries.

1821

1822

Section X  Small and Large Intestine Table 105-1  Causes of Infectious Diarrhea in the Tropics Bacteria Aeromonas hydrophila Campylobacter jejuni Nontyphoidal Salmonella species Pathogenic Escherichia coli: enterotoxigenic, enteroaggregative, enteroinvasive, enterohemorrhagic Plesiomonas shigelloides Shigella species: S. dysenteriae, S. flexneri, S. sonnei, S. boydii Vibrio cholerae Yersinia enterocolitica Helminths Capillaria philippinensis Fasciolopsis buski Heterophyiasis (Metagonimus yokogawai, Haplorchis taichui) Schistosoma mansoni Strongyloides stercoralis Protozoa Blastocystis hominis Cryptosporidium parvum Cyclospora cayetanensis Encephalitozoon intestinalis Enterocytozoon bieneusi Giardia lamblia Isospora belli Leishmania donovani Viruses Astroviruses Caliciviruses: Norovirus and Sapovirus Enteric adenoviruses Human immunodeficiency virus Rotavirus

Diarrhea in travelers is often geographically determined, being caused by enterotoxin-producing Escherichia coli with travel to South America or to Mexico, by Giardia lamblia and Cryptosporidium in southern Central Asia, and by Campylobacter in Southeast Asia.9,11 Infection with the coccidian parasites occasionally causes diarrhea in travelers to the tropics. Cyclospora cayetanensis first was reported to cause outbreaks and prolonged diarrhea in travelers to Nepal and has subsequently been reported from other regions in the tropics including Southeast Asia, Africa, Turkey, and Latin America.12,13 Blastocystis hominis is another opportunistic protozoan that occasionally causes acute or chronic diarrhea in travelers.14

Table 105-2  Causes of Malabsorption Syndrome in the Tropics Small Intestinal Bacterial Overgrowth Following ulcer surgery Secondary to intestinal tuberculosis and Crohn’s disease Infections Bacteria Mycobacterium avium intracellulare complex Mycobacterium tuberculosis Helminths Capillaria philippinensis Strongyloides stercoralis Protozoa Cryptosporidium parvum Cyclospora cayetanensis Encephalitozoon intestinalis Enterocytozoon bieuneusi Giardia lamblia Isospora belli Leishmania donovani Lymphatic Obstruction Intestinal lymphangiectasia Mucosal Diseases Autoimmune enteropathy Celiac disease Eosinophilic gastroenteritis HIV enteropathy Immunoproliferative small intestinal disease Intestinal lymphoma Primary immunodeficiencies Tropical sprue Neonatal Diseases Microvillus inclusion disease Tufting enteropathy Pancreatic Insufficiency Alcoholic pancreatitis Cystic fibrosis Tropical pancreatitis Specific Transport Disorders Abetalipoproteinemia Fructose malabsorption Glucose-galactose malabsorption Hypolactasia Sucrose intolerance HIV, human immunodeficiency virus.

Tropical sprue remains a significant cause of malabsorption in several tropical countries. The etiology of this disease continues to remain obscure, and it is unlikely that it will ever be conclusively settled given its diminishing presence. Tropical sprue needs to be differentiated from a variety of other conditions that also cause malabsorption in residents of the tropics (Table 105-2).

causes of malabsorption had been excluded. Historically, this definition originated at a time when testing for intestinal absorption was commonplace. The availability of specific diagnostic tests for many of the diseases that cause malabsorption, such as abdominal imaging for chronic pancreatitis or antiendomysial and antitissue transglutaminase antibody for celiac disease, led to a general decline in the use of tests for absorption. In the developed world it is now uncommon to find laboratories that perform tests for absorption, particularly fecal fat estimation.16 The diagnosis of tropical sprue, however, continues to require the demonstration of malabsorption and the exclusion of other specific pathologies including celiac disease, chronic pancreatitis, and parasitic infections.

DEFINITION

HISTORY

TROPICAL SPRUE

Tropical sprue is a primary (i.e., not caused by other known disease) malabsorption syndrome that occurs in visitors to or residents of the tropics. Baker and Klipstein,15 working in South India and Central America respectively, defined tropical sprue as an intestinal mucosal disease characterized by malabsorption of two or more unrelated nutrient groups (e.g., fat, carbohydrate, vitamins) where other known

Modern medical history records the first description of tropical sprue from Barbados in the West Indies by William Hillary in 1759.17 He described this disease (aphthoides chronica) as commencing with severe mouth ulcers and glossitis, followed sequentially by diarrhea, marasmus, and death. The word sprue probably originated from the Dutch term sprouw used to describe a condition characterized by

Chapter 105  Tropical Diarrhea and Malabsorption 60

30

Tropic of Cancer

0

Equator

Figure 105-1.  Geographic distribution of tropical malabsorption. Purple indicates areas where overt tropical sprue occurs; yellow indicates areas where a disorder resembling tropical sprue occurs; and green indicates areas where only subclinical abnormalities of small intestinal structure or function have been observed. (Modified from Klipstein FA. Tropical sprue in travelers and expatriates living abroad. Gastroenterology 1981; 80:590-600.)

Tropic of Capricorn 30

60

severe aphthous ulceration of the mouth and stomatitis.18 The Dutch used the term Indische Sprouw to describe what appeared to be the same condition occurring in their colonies in Southeast Asia. A similar illness was noted in Europeans who spent time in the Asian colonies including India, Indochina, and China and that went by many names including “chronic diarrhea of the tropics” before being named “sprue” by Manson in 1880.17 Although now pri­ marily described in Westerners who have spent some time in the tropics, descriptions of a malabsorption syndrome in the indigenous population in India date back to the second century bc. The ancient Indian medical textbook the Charaka Samhita19 described an illness (Grahani vyadhi) characterized by glossitis, diarrhea, malabsorptionlike stools, and wasting and ascribed to a loss of the digestive fire.

EPIDEMIOLOGY

Tropical sprue has been described in South and Southeast Asia, Central America, South America, the Caribbean islands, and parts of Africa (Fig. 105-1). Endemic tropical sprue is most often recognized in visitors who have spent at least a few months in the tropics before returning home,20-22 although tropical sprue or tropical malabsorption also occurs in indigenous residents of the tropics.23-26 It is not clear, however, if the disease that occurs in returning Westerners is the same as the one that occurs in indigenous residents of the tropics. At least in expatriates returning to Europe, it appears that that tropical sprue may be secondary to prolonged infection, or follows an unusual infection, an occurrence that has been called postinfectious malabsorption.27 Tropical sprue also has occurred in epidemic form.5 Epidemics occurred in soldiers and prisoners of war in Asia during the Second World War28,29 and also were reported in South India during the 1960s and 1970s.30 Epidemic tropical sprue affected adults more often than it did children, and exposure during the first wave of an epidemic often conferred protection during the second wave; intrafamilial secondary transmission of disease also was noted. In the early 1960s, epidemic sprue was responsible for the deaths of 30,000 to 40,000 people in South India alone. A seasonal occurrence of tropical sprue was noted in Puerto Rico, where it was common during the first three months of the year.31

Epidemic sprue is no longer reported from South India or other parts of the world. Sporadic tropical sprue also has become relatively rare in some areas of India,7 compared with celiac disease and intestinal tuberculosis, where it once used to represent the major cause of malabsorption and chronic diarrhea. Nonetheless, tropical sprue continues to account for a significant proportion of adult malabsorption and a smaller proportion of childhood malabsorption in South Asia.6,24,25

ETIOLOGY

The etiology of tropical sprue remains unknown. Persistent infection of the small intestine with coliform organisms was postulated to cause the disease in some areas of the world. In Haiti and Puerto Rico, overgrowth of toxin-producing coliforms (Klebsiella spp., Enterobacter cloacae, or E. coli) in the small bowel was associated with feasting and ingesting excessive amounts of long-chain unsaturated fatty acids.32,33 Bacterial overgrowth in the proximal small intestine in patients with tropical sprue also has been noted in other parts of the world, but it has been ascribed to the slow small intestinal transit that characterizes the disease.34 The normal human jejunum contains up to 103 bacteria per milliliter of luminal fluid in Western residents, and bacterial overgrowth is considered to occur when the concentration of bacteria exceeds 105/mL. These criteria, however, do not necessarily hold true for apparently healthy residents of the tropics, and studies in South India have demonstrated that bacterial counts in jejunal luminal fluid ranged up to 105/ mL in healthy asymptomatic persons.35 Patients with trop­ ical sprue and asymptomatic persons from the same backgrounds had similar levels and types of bacteria in their proximal small intestine, indicating that bacterial overgrowth probably was not responsible for the malabsorption of tropical sprue. Toxin-producing bacteria were not found in the intestine of patients with tropical sprue in South India.36 Viral particles resembling human enteric coronaviruses have been identified in the stool and jejunal enterocytes of patients with tropical sprue,37 but they also may be present in apparently normal persons, so they have not been definitely implicated in the etiology of sprue. Acute and reversible flattening of the small intestinal mucosa identical to that seen with tropical sprue has been noted in the absence

1823

1824

Section X  Small and Large Intestine Table 105-3  Clinical Features of Tropical Malabsorption and Their Causes clinical feature(s)

CAUSE(s)

mechanism(s)

Diarrhea

Carbohydrate malabsorption Unabsorbed fatty and bile acids

Osmotic diarrhea Colonic water secretion

Pale, bulky stool

Fatty acid and bile malabsorption

Mucosal disease Pancreatic insufficiency

Foul-smelling stool

Fat and protein malabsorption

Oxidation of malabsorbed fatty acids Production of skatole from tryptophan

Borborygmi, abdominal fullness

Carbohydrate malabsorption

Bacterial fermentation of unabsorbed carbohydrate

Nocturia

Delayed water absorption

Small intestinal disease delays water absorption

Pedal edema, skin changes, leukonychia, muscle wasting

Hypoproteinemia and protein-losing enteropathy

Loss of mucosal surface Ulceration Functional pancreatic insufficiency

Pallor

Anemia due to vitamin B12, folate, iron deficiency

Mucosal disease in duodenum (folate, iron) and ileum (B12)

Koilonychia

Iron deficiency

Mucosal disease in duodenum

Hyperpigmentation of palms, knuckles, mouth

Vitamin B12 deficiency

Disturbed metabolism of melanin

Angular stomatitis, glossitis

Vitamin B deficiency

Mucosal disease of small intestine

Night blindness, corneal xerosis, Bitot’s spots

Vitamin A deficiency

Steatorrhea with malabsorption of vitamin A

Muscle weakness

Hypophosphatemia, hypokalemia, hypomagnesemia

Mucosal disease with malabsorption or secretion

Tetany, carpopedal spasm

Hypocalcemia

Steatorrhea

Hemorrhagic diathesis

Vitamin K deficiency

Steatorrhea with malabsorption of vitamin K

Lack of taste (ageusia)

Zinc deficiency

Steatorrhea with malabsorption of vitamin D

Weight loss

Reduced calorie intake Increased fecal calorie loss

Anorexia secondary to folate and B12 deficiency Malabsorption

Ogilvie’s syndrome

Colonic pseudo-obstruction due to electrolyte disorders (e.g., hypokalemia, hypomagnesemia)

Uncoordinated muscle activity and segmentation

of gluten-sensitive enteropathy and ascribed to a probable viral etiology.38 The morphologic appearances of sprue are nonspecific and can occur with infectious and other causes; the etiology of tropical sprue has not been settled yet. All previous investigations, however, antedated our current appreciation of innate immunity and its relation to inflammation in the intestine, and it is possible that investigations targeted toward these mechanisms will lead to progress in our understanding of what causes the disease. The etiology of tropical sprue in visitors to the tropics may be quite different from that occurring in residents of tropical climates. Longer-term visitors might experience one or more episodes of diarrhea, usually in response to enteric infection with pathogens such as enteropathogenic E. coli, Giardia intestinalis, or Cyclospora species; persistent intestinal inflammation following such infection has been blamed as the cause of tropical sprue in this setting. Although tropical sprue might very well occur following exposure to certain known pathogens, the possibility has never been examined that this disorder also might occur in response to otherwise harmless bacteria, namely to an overload of apparently nonpathogenic bacteria. In most cases, tropical diarrhea in visitors is short-lived and self-limited or is truncated by antibiotic use. Some persons, however, develop chronic diarrhea and evidence of malabsorption, and the intestinal mucosal biopsy specimens reveal villus atrophy and inflammatory infiltration in the lamina propria.

Tropical sprue in visitors to tropical countries is thus sometimes termed postinfectious malabsorption.

CLINICAL FEATURES

Although tropical sprue does occur in children, typically the disease affects adults. Presentation is with chronic diarrhea, steatorrhea, glossitis, abdominal distention, prominent bowel sounds or borborygmi, and weight loss (Table 105-3). In expatriates and during epidemics, the illness often begins with fever and watery or, rarely, bloody diarrhea. These symptoms resolve after a week or so to be followed by a lingering diarrhea or steatorrhea associated with weight loss. The physical signs that may be found in affected patients (see Table 105-3) (see Chapter 100) include pallor from iron and vitamin B deficiency; angular stomatitis, cheilitis, and glossitis due to vitamin B deficiency; and peripheral edema and skin and hair changes secondary to hypoproteinemia. Hyperpigmentation of the buccal mucosa, palms, and knuckles sometimes is noted (Fig. 105-2) and has been ascribed to vitamin B12 deficiency. Vitamin A deficiency with night blindness, Bitot’s spots, and corneal xerosis and subacute combined degeneration of the spinal cord resulting from vitamin B12 deficiency are much rarer. Fever has been noted in a quarter of patients in South India. Patients are grossly emaciated in the later stages of illness, with muscle weakness, particularly in the proximal muscles.

Chapter 105  Tropical Diarrhea and Malabsorption

Figure 105-2.  Hyperpigmentation of the palms in a patient with tropical sprue. This is often a finding in patients with tropical sprue in southern India and has been ascribed to vitamin B12 deficiency.

Figure 105-3.  Finger clubbing in a patient with immunoproliferative small intestinal disease.

secretion manifest on gastric secretory testing, and examination of biopsy specimens indicated that more than half of patients with tropical sprue had atrophic gastritis.39 These studies, however, were performed before the recognition of Helicobacter pylori and, therefore, their true significance is unknown. Small intestinal biopsy specimens originally were taken using a peroral biopsy capsule, and today it is standard to obtain biopsies from the distal duodenum or proximal jejunum during upper gastrointestinal endoscopy. Biopsies obtained using the capsule were examined with a hand lens or a dissecting microscope and characteristically showed blunting and fusion of the villi. The endoscopic parallel of this finding is scalloping of the duodenal mucosa, seen in patients with villus atrophy of the duodenal mucosa.40 Originally described in celiac disease, it is now appreciated that scalloping is nonspecific and indicates villus atrophy. Histologically, small intestinal biopsy specimens show varying degrees of villus shortening (atrophy), whereas the crypts are elongated.41-43 In addition to blunting, the villi sometimes show fusion. The normal villus-to-crypt ratio in the jejunal mucosa is 4 : 1 or 5 : 1; in the partial villus atrophy seen with tropical sprue this ratio is reduced to 2 : 1 or 1 : 1. Complete villus atrophy, noted in some patients with celiac disease, is not seen in tropical sprue. The lamina propria of the small intestine is infiltrated by lymphocytes, and there is an increase in intraepithelial lymphocytes. Intraepithelial lymphocytes are a unique T-cell population seen in celiac disease, but they also may be increased in tropical sprue and a number of other diseases.44 Baker and Mathan41 graded the histologic changes in tropical sprue on a scale of 0 to III (Fig. 105-4), which is somewhat different from the Marsh scoring system that is widely used to score celiac disease biopsy specimens. Electron microscopy reveals degenerating cells in the crypts of the small intestine,45 a finding similar to that seen in radiation enteritis and in enteritis secondary to chemotherapy. The location of the damaged cells in the base of the crypt suggests that there was damage to the stem cells in the small intestine. In addition to changes noted in the stomach and the small intestine, an infiltration of the colonic mucosa by lymphocytes also is seen in tropical sprue,46 similar to the lymphocytic colitis that may be seen in association with celiac disease.47

PATHOPHYSIOLOGY

Peripheral neuropathy may be present, but subacute combined degeneration of the spinal cord, once noted in tropical sprue, is no longer found, perhaps because of earlier diagnosis and treatment. The presence of finger clubbing should suggest other illness such as immunoproliferative small intestinal disease, in which this finding is characteristic (Fig. 105-3). Sprue coma, manifest by listlessness and apathy was described in the past and attributed to a deficiency of divalent cations, particularly magnesium. Rarely, a syndrome similar to the Ogilvie syndrome occurs, characterized by abdominal pain and distention with exaggerated bowel sounds secondary to colonic pseudo-obstruction. This surprising finding likely results from disordered intestinal motility in these patients. The clinical features of florid malnutrition probably are less common in expatriates with tropical sprue than in residents with the disease.

HISTOPATHOLOGY

Although tropical sprue typically is a disease of the small intestine, involvement of the stomach and colon has been reported. Many patients with sprue have reduced acid

The characteristic finding in tropical sprue is villus atrophy in the small intestine accompanied by elongation of the crypts. Ultrastructural studies show the microvilli of the epithelial cells are distorted and grouped. These changes lead to a marked reduction in the total absorptive surface area of the small intestine, which is reflected in reduced xylose absorption.48 Dead epithelial cells at the villus base are extruded more rapidly than normal,49 and this is compensated by increased crypt cell production and enterocyte migration up the villus. In contrast to comparable disorders such as celiac disease, electron microscopy in tropical sprue reveals degenerating and dead cells in the basal region of the crypts, suggesting that there is an intestinal stem cell defect in tropical sprue that prevents adequate restoration of absorptive surface area. Fat malabsorption in tropical sprue is secondary to epithelial cell damage in the small intestine. It also has been shown that the pancreatic enzyme response to an indirect test such as the Lundh meal is impaired, although pancreatic response to secretin is normal. Such functional pancreatic insufficiency is due to inadequate pancreatic stimulation as a consequence of mucosal disease and is reflected in an

1825

1826

Section X  Small and Large Intestine

A

B

C

D

Figure 105-4.  Histopathology of tropical sprue. Endoscopic deep duodenal biopsy specimens showing a gradation of mucosal histology according to the grading system of Baker and Mathan41: A, Grade 0: normal. B, Grade I: shortening of the villi, increase in the depth of the glandular layer, and increase in lymphocytic infiltration of the lamina propria and surface epithelial cells. C, Grade II: further increase in the depth of the glandular layer extending up to half the total distance from the crypt base to the villus tip with more cellular infiltration. D, Grade III: the glandular layer occupies more than half the distance from the crypt base to the villus, together with flattening of surface epithelial cells, disorganization of the brush border, and marked cellular infiltration of the lamina propria and surface epithelial cells. The completely flat mucosal surface that occurs in some cases of celiac disease is not seen in tropical sprue (see Fig. 104-2).

abnormal pancreolauryl test in more than 50% of patients with tropical sprue.49 The fat in the stool is largely in the form of free fatty acids, both saturated and unsaturated, depending on the nature of the ingested dietary fat. Folate and iron deficiency in tropical sprue reflect proximal small intestinal involvement, whereas B12 malabsorption reflects involvement of the terminal ileum. The three-stage Schilling test in tropical sprue evidences mucosal abnormality, because defective vitamin B12 absorption is not corrected either by intrinsic factor or by antibiotic therapy. Bile acid malabsorption also results from terminal ileal involvement and can contribute to diarrhea. Colonic malabsorption of water and electrolytes contributes signi­ ficantly to diarrhea in tropical sprue and can result from the action of unabsorbed bile acids and free unsaturated fatty acids in the stool, which inhibit colonocyte Na+,K+-ATPase.36,50 Small intestinal transit is prolonged in patients with tropical sprue, and this has been ascribed to the effects of unab-

sorbed fat on the ileal brake.34 The ileal brake is considered to be mediated via peptide YY, and elevated serum levels of this peptide have been noted in tropical sprue.51

DIAGNOSIS

There is no single test that allows the diagnosis of tropical sprue. Although electron microscopy of the jejunal mucosa of patients with tropical sprue in South India showed the presence of degenerating cells in the crypts, an apparently specific change, this has not led to its use as a diagnostic test. In practice, tropical sprue continues to be diagnosed by establishing the presence of malabsorption, excluding other known causes of malabsorption, and demonstrating the characteristic mucosal histology. Malabsorption commonly is established by testing for fecal fat, measuring serum levels of vitamin B12 and folate, and testing for d-xylose absorption (see Chapter 101). Two abnormal tests in the appropriate setting are consistent

Chapter 105  Tropical Diarrhea and Malabsorption with tropical sprue in the absence of other causes of malabsorption. The quantitative fecal fat estimation of van de Kamer52 is the gold standard for establishing steatorrhea but is not available in most laboratories. Because of difficulties in performing this test, steatorrhea often is assessed qualitatively using Sudan staining of oil (triglyceride) droplets in stool. The acid steatocrit, a test based on the separation of fat from nonfat components by centrifuging acidified stool in a hematocrit tube, was developed as a simple test for the steatorrhea from pancreatic insufficiency in which fecal fat is in the form of triglycerides.53 In the steatorrhea of mucosal disease, however, excreted fat is in the form of free fatty acids and the steatocrit is not reliable. In tropical sprue in India, d-xylose malabsorption, steatorrhea, and vitamin B12 malabsorption are found in about 99%, 90%, and 60% to 90% of patients, respectively. Xylose absorption is tested either by measuring urinary excretion of the sugar over a period of five hours after ingesting xylose or by measuring blood xylose level one hour after ingesting oral d-xylose. In the tropics, a 5-g dose rather than the traditional 25-g dose of d-xylose is used because the lower dose does not cause vomiting or diarrhea. Xylose absorption is a reflection of surface area of the small intestine, because this pentose sugar is absorbed by passive diffusion. It is a good but nonspecific indicator of reduced mucosal surface area in tropical sprue. In contrast to celiac disease, in which involvement of the proximal small intestine predominates and, therefore, serum folate levels are low and vitamin B12 levels are normal, tropical sprue affects the distal small intestine and terminal ileum; as a result, serum folate levels are normal and vitamin B12 levels are low. Vitamin B12 deficiency is common in tropical sprue and is demonstrated best by the three-stage Schilling test. An increase of breath hydrogen after glucose or xylose ingestion has been interpreted as evidence of small bowel bacterial overgrowth, whereas the abnormal lactulose breath hydrogen test suggests prolonged small bowel transit in tropical sprue.34 Fecal examination for occult blood and for parasites is essential. Protozoan parasites that cause diarrhea must be sought using special stains of the feces including modified acid-fast stains. In tropical sprue, characteristic changes observed on small bowel series include thickening of the mucosal folds of the jejunum, loss of the feathery mucosal pattern in the proximal jejunum, and delayed transit through the small intestine (Fig. 105-5). Currently, however, computed tomographic scanning (CT) of the abdomen is more commonly performed than small bowel series. Dilated, featureless, atonic loops of small intestine and dilution of oral contrast, suggesting a hypersecretory state, are found. The main role of CT, however, is to exclude intestinal masses and lymphadenopathy within the mesentery and retroperitoneum, which are characteristic of other inflammatory diseases, such as tuberculosis, lymphoma, parasitic infestations, and eosinophilic enteritis. Capsule endoscopy of the small bowel is sometimes undertaken in patients with small bowel diarrhea after excluding the presence of strictures; however, enteroscopy using push enteroscopes or double-balloon or single-balloon instruments has the advantage of obtaining biopsies and is likely to become the standard for diagnosis in these patients, especially to exclude secondary causes of malabsorption.54,55 Small intestinal mucosal biopsies are taken from the third or fourth part of the duodenum during upper gastrointestinal endoscopy to exclude specific causes of malabsorption and to diagnose tropical sprue. Endoscopy can reveal a scalloped duodenal mucosa similar to that seen in celiac disease,

Figure 105-5.  Film from a barium follow-through study of the small bowel in a patient with tropical sprue showing loss of the normal feathery mucosal pattern in the small bowel, with dilatation of jejunal loops and thickening of the mucosal folds.

which results from mucosal atrophy and loss of the villus pattern. A similar appearance of the jejunal mucosa also may be found during enteroscopy. Magnification endoscopy, enhanced by application of a 3% acetic acid spray, has been reported to increase the recognition of patchy villus atrophy and to allow targeted mucosal biopsies that can more likely provide a diagnosis in tropical sprue.56 Small intestinal bacterial overgrowth may be demonstrated by quantitative aerobic and anaerobic culture of the small bowel contents aspirated either during endoscopy or by a jejunal tube. Fasting counts exceeding 105/mL of jejunal fluid are considered to indicate bacterial overgrowth.

TREATMENT

Dehydration and electrolyte imbalance must be corrected with appropriate intravenous fluids, such as full-strength lactated Ringer’s solution. Profound hypokalemia occurs in some patients with severe diarrhea and might need to be corrected under electrocardiographic monitoring. Deficiencies of magnesium and calcium need to be corrected with parenteral magnesium and calcium, respectively, in patients with long-standing illness. Deficiency of vitamin D and calcium may be treated with oral supplements, but deficiency of vitamin A, indicated by corneal xerosis or symptoms of night blindness, should be corrected with parenteral administration of the vitamin. Vitamin B12 given parenterally and folic acid given orally quickly lead to resolution of symptoms of anemia, glossitis, and anorexia and result in weight gain. Folate supplementation results in improvement of both macrocytic anemia and villus atrophy. Tetracycline 250  mg four times daily (or doxycycline 100  mg once daily) for 6 months is prescribed as specific therapy for tropical sprue.57,58 There is, however, little experience with use of other antibiotics in this condition. A high-calorie, high-protein, fat-restricted diet usually is given to these patients. Restriction of long-chain fatty acids in the diet, and its substitution by medium chain triglycer-

1827

1828

Section X  Small and Large Intestine ides, is particularly useful in reducing diarrhea and steatorrhea. Colonic pseudo-obstruction, seen in the rare patient, can require colonic decompression. Relapses of disease after cessation of therapy are known to occur in endemic tropical sprue and require investigation to exclude other causes before resuming therapy.

TROPICAL ENTEROPATHY AND ITS DISTINCTION FROM TROPICAL SPRUE The mucosa of the small intestine of residents of the tropics is structurally different from that of residents of temperate countries. This condition, believed to be secondary to environmental factors peculiar to the tropics, has been variously labeled tropical enteropathy and subclinical tropical malabsorption. Typically in tropical enteropathy, intestinal villi are shorter than those of the residents of temperate climates, crypts are more elongated, and there are increased numbers of lymphocytes and plasma cells in the lamina propria. These changes lead to a significant reduction in the surface area of the intestine, which manifests as abnormal d-xylose absorption and increased intestinal permeability. In contrast to persons with tropical sprue, those with tropical enteropathy are asymptomatic. The issue of whether tropical enteropathy is a milder form (subclinical malabsorption) of tropical sprue remains controversial. Baker and Mathan41 championed a clear distinction between tropical sprue and tropical enteropathy, the former causing symptomatic malabsorption and the latter causing an asymptomatic condition with mild abnormalities of absorption. Menzies and coworkers found that 218 healthy volunteer residents of tropical areas had a higher mean lactulose-to-rhamnose ratio and lower mean five-hour recoveries of 3-O-methyl-d-glucose, d-xylose, and l-rhamnose, indicating higher intestinal permeability, lower mucosal surface area, and lower absorptive capacity than 228 healthy volunteer residents of subtropical and temperate regions.59 Investigation of visiting residents in subtropical and temperate climates suggested that differences in intestinal permeability and absorptive capacity were related to their area of residence rather than area of origin.59 Histologic enteropathy continues to be observed even today in residents of tropical countries, with mucosal biopsy specimens showing blunting of villi and increased cellularity of the lamina propria in persons without overt intestinal disease. It has been postulated that tropical enteropathy might represent an adaptation of the intestine to frequent intestinal infection in childhood. There is some evidence that the mucosal damage of tropical enteropathy is mediated by T cells.60 In a large longitudinal study, Kelly and colleagues performed regular monthly bacteriologic and parasitologic examinations of the stool and annual endoscopic jejunal biopsy with morphometry on 238 asymptomatic adults in Zambia. Five-hour urine xylose excretion was low (16.6 ± 6.9%) in 182 members of a cohort drawn from a township compared with 13 staff members undertaking the study (25.5 ± 4.4%), which in turn was significantly less than that of healthy volunteers in London (33.1 ± 0.7%).2 The mucosal biopsies of those from the Zambian township showed an altered crypt-to-villus ratio of approximately 5 : 3. There were marked changes over time, including reduction in villus height (16%), reduction in xylose absorption (16%), and increase in intestinal permeability (28%)

during the peak rainfall months of December and January, in association with an incidence of diarrhea that was higher than in other months.2 Asymptomatic intestinal infections also were noted in approximately half of the participants, and enteropathy was more severe in persons who had been infected with Citrobacter rodentium or hookworm. It has been suggested that the increased intestinal permeability can lead to absorption of endotoxins, systemic inflammation, and growth faltering in Gambian infants.61 However, fecal neopterin excretion, a marker of inflammation, was associated with relative growth failure but did not correlate with intestinal permeability, a marker of tropical enteropathy.62 Attempts to improve tropical enteropathy by oral administration of the probiotic bacterium Lactobacillus GG for 30 days in Malawian children did not reveal any effect of the probiotic on mucosal function.63 Tropical enteropathy is not without clinical consequence. The reduced surface area implies that absorption of energy may be slightly diminished, which might explain a low body mass index in many persons living in the tropics. As an extension of the hygiene hypothesis that relates increased hygiene to an increased prevalence of autoinflammatory diseases, it has been suggested that tropical enteropathy might protect against these inflammatory diseases through poorly understood immune and neuroendocrine mechanisms within the intestinal tract.64 It also has been suggested that tropical enteropathy influences the severity of infectious gastroenteritis and that it has implications for the composition of oral rehydration solution, leading to the use of lower concentrations of sodium and glucose than in the original formulation that was isotonic with plasma.65,66 Ongoing research also addresses the issue of whether tropical enteropathy has an impact on drug absorption, an issue particularly relevant to the management of a number of diseases including infection with the human immunodeficiency virus.

PROTOZOAN INFECTIONS THAT CAUSE MALABSORPTION This topic is discussed in detail in Chapter 109.

GIARDIASIS

Infection with the protozoan parasite G. lamblia is most often asymptomatic in persons who live in the tropics.5 Giardiasis can cause diarrhea in immunocompromised persons or in visitors from Western countries to the tropics, and it can cause self-limited illness in children. Diarrhea is associated with the presence of trophozoites in the stool, and the presence of cysts alone should be interpreted with caution. Decreased brush border surface area in the jejunum leads to carbohydrate malabsorption, and associated small intestinal bacterial overgrowth and consequent bile salt deconjugation leads to steatorrhea. The diagnosis of giardiasis is made by microscopic examination of fresh stool specimens using simple microscopy or by a direct fluorescence test; examination of at least three fecal specimens is recommended. Examination of duodenal or jejunal biopsies does not increase the diagnostic yield over examination of stool specimens. The parasites can be seen in the mucus layer overlying the epithelium, and the mucosa can show atrophy of villi and elongation of crypts with a mononuclear inflammatory cell infiltrate in the lamina propria.

Chapter 105  Tropical Diarrhea and Malabsorption Table 105-4  Specific Therapy of Tropical Malabsorption DISEASE

SPECIFIC THERAPY

DURATION

Capillariasis

Thiabendazole 25 mg/kg twice daily or Albendazole 400 mg twice daily Gluten free diet Human gamma globulin intravenously every 4 to 6 weeks Bone marrow transplantation Nitazoxanide 500 mg twice daily or Paromomycin 500 mg 3 or 4 times daily Trimethoprim-sulfamethoxazole 160 mg/800 mg twice daily or ciprofloxacin 500 mg twice daily Followed by trimethoprim-sulfamethoxazole 160 mg/800 mg every other day or three times a week, or ciprofloxacin 500 mg three times a week Metronidazole 400 mg 3 times daily or Tinidazole 500 mg twice daily Albendazole 400 mg twice daily Tetracycline 250 mg 4 times daily

20 days 7-10 days Lifelong Lifelong

Trimethoprim-sulfamethoxazole 160 mg/800 mg twice daily followed by 400/80 twice daily Albendazole 400 mg twice daily

10 days 3 weeks 2-3 weeks

Nitazoxanide 500 mg twice daily or Fumagillin 20 mg 3 times daily Thiabendazole 25 mg/kg twice daily or Albendazole 400 mg twice daily or Ivermectin 200 µg/kg Isoniazid, rifampicin, pyrazinamide, and ethambutol followed by or Two-drug therapy with isoniazid and rifampicin Tetracycline 250 mg 4 times daily

3-7 days 2 weeks 3 days 3-7 days 1-2 days 2 months 4-7 months 6 months

Celiac disease Common variable immunodeficiency Cryptosporidiosis Cyclosporiasis

Giardiasis Immunoproliferative small intestinal disease stage A Isosporiasis Microsporidiosis due to Encephalitozoon intestinalis Microsporidiosis due to Enterocytozoon bieuneusi Strongyloidiasis Tuberculosis Tropical sprue

Symptomatic giardiasis responds quickly to treatment with metronidazole, tinidazole, or other imidazoles. Albendazole and nitazoxanide also have been used successfully (Table 105-4).

OTHER PARASITIC INFECTIONS

Other protozoa associated with malabsorption include Cryptosporidium parvum, Isospora belli, Cyclospora cayetanensis, and the microsporidia Enterocytozoon bieneusi and Encephalitozoon intestinalis. Infection with these protozoa is widespread in tropical countries.67 Most infected persons are asymptomatic, but some develop self-limited diarrhea. Endoscopy with intestinal biopsy is helpful in diagnosis.68 Malabsorption resulting from infection with these parasites occurs mainly in immunocompromised hosts, such as those with HIV infection, persons with primary immunodeficiency syndromes, and post-transplant immunosuppressed patients, as a consequence of epithelial cell infection. Cryptosporidiosis with malabsorption is treated with paromomycin and nitazoxanide. Cyclospora causes a malabsorption syndrome with villus atrophy and crypt hyperplasia. Infection with cyclospora is treated with cotrimoxazole. Isosporiasis is treated with a 10-day course of cotrimoxazole followed by long-term low-dose cotrimoxazole in immunosuppressed patients. Microsporidiosis has been described in many countries in Asia, tropical Africa, and Central and South America, and its therapy in symptomatic persons depends on the infecting species. Visceral leishmaniaisis, common in some tropical countries, is characterized by the presence of parasitized macrophages in the lamina propria of the small intestine with inflammatory cell infiltration; it

3-7 days 7 days 10 weeks 7-14 days 7-14 days 7-14 days 6 months to 2 years

can cause chronic diarrhea with malabsorption of vitamin A and d-xylose.

HELMINTHIC INFECTIONS THAT CAUSE MALABSORPTION Helminthic infections also can cause a malabsorption syndrome in the tropics.5 The most common helminthic infections are with Strongyloides stercoralis and Capillaria philippinensis. Infection with Strongyloides stercoralis can cause chronic diarrhea and malabsorption in immunocompetent persons, although immunosuppression, particularly with glucocorticoid use, predisposes to hyperinfection with this parasite. Infection with the human T-lymphotropic virus-1 (HTLV-1) is associated with persistent Strongyloides infection and chronic diarrhea. Diarrhea may be intermittent or persistent, and steatorrhea, anemia, and hypoproteinemia are common. Small bowel series can show changes suggesting mucosal infiltration and ulceration in the duodenum and jejunum. Diagnosis usually is made by examination of feces for the larvae of the parasite, although occasionally the infection is only recognized upon small bowel biopsy. Treatment with thiabendazole, albendazole, or ivermectin is effective. Intestinal capillariasis causes a malabsorption syndrome and is common in Southeast Asia, especially Thailand and the Philippines, but is now reported from other countries including Taiwan, Korea, India, Iran, and Egypt. Intestinal capillariasis is associated with protein-losing enteropathy

1829

1830

Section X  Small and Large Intestine as well as malabsorption of fat and d-xylose. It is usually treated with thiabendazole or albendazole. This topic is discussed in more detail in Chapter 110.

HIV INFECTION AND AIDS Since the AIDS epidemic has spread throughout so many tropical countries, it has become necessary to exclude HIV infection in any patient with malabsorption syndrome or chronic diarrhea. Diarrhea and malabsorption lasting for longer than a month is an AIDS-defining condition, is associated with low CD4 counts, and usually is caused by infection of the small intestine with one of the protozoan pathogens mentioned earlier. Studies, however, suggest that chronic diarrhea is neither a very sensitive nor a specific indicator of AIDS in tropical countries, such as India.69 Occurrence of pathogen-negative diarrhea and malabsorption in patients with AIDS has been attributed to possible direct viral (HIV) enteropathy, although this has never been proven. Diarrhea and malabsorption require specific therapy of the opportunistic infection as well as antiretroviral therapy. This topic is discussed in more detail in Chapter 33.

INTESTINAL TUBERCULOSIS Intestinal tuberculosis, comprising ulcerative, hypertrophic, and ulcerohypertrophic varieties, is common in tropical countries. The ulcerative variety of tuberculosis commonly manifests with chronic diarrhea and malabsorption, whereas the hypertrophic variety more commonly causes abdominal pain and intestinal obstruction. In contrast to mucosal diseases such as tropical sprue, abdominal pain is a significant symptom and results from ulceration and partial obstruction of the small bowel. Biochemical evidence of malabsorption can be found in many patients with intestinal tuberculosis, even if the patient does not present with a clinical diagnosis of malabsorption syndrome. The causes of malabsorption in tuberculosis include bacterial overgrowth in a stagnant loop, bile salt deconjugation, and diminished absorptive surface due to ulceration and lymphatic obstruction. The diagnosis is made in most cases by endoscopy with a combination of histology, culture, and polymerase chain reaction tests (see Chapter 107).70

CROHN’S DISEASE Crohn’s disease is increasing in incidence in tropical countries and now is an important item in the differential diagnosis for any patient in whom tuberculosis is a possible diagnosis. Malabsorption in Crohn’s disease may be due to a number of factors. About a third of patients have small intestinal involvement, and this can reduce the absorptive surface area; extensive small bowel resections ultimately have the same effect. Terminal ileal disease or resections can lead to vitamin B12 deficiency and bile salt malabsorption, and ileocecal valve resections can result in bacterial overgrowth with resultant malabsorption. Approximately 9% of cases of malabsorption were found to be due to Crohn’s disease in an unselected case series from northern India (see Chapter 111).6

CELIAC DISEASE Celiac disease (gluten sensitive enteropathy), hitherto considered uncommon in the tropics, is increasingly described from northern India and selected areas of sub-Saharan Africa71 and may be unmasked by intestinal infection. The disease often manifests in infancy around the time of weaning, but presentation at later ages, including adulthood, is not uncommon. Celiac disease is differentiated from tropical sprue by the presence of complete villus atrophy in mucosal biopsies. The diagnosis is confirmed by the presence of IgA antiendomysial and antitissue transglutaminase antibodies, although these tests may be negative in persons with selective IgA deficiency. Clinical and histologic responses to gluten withdrawal are important in confirming the diagnosis (see Chapter 104).

PRIMARY IMMUNODEFICIENCY SYNDROMES Common variable immunodeficiency (CVI) occurs sporadically in residents of the tropics and can manifest as a malabsorption syndrome.72 Recurrent diarrhea, recurrent sinopulmonary infections, and recurrent meningitis are other manifestations of this disease. CVI first may be suspected when a small bowel biopsy shows reduced numbers of plasma cells in the lamina propria or by the finding of nodular lymphoid hyperplasia, which is occasionally associated with the disease. The most common intestinal infection in these patients is giardiasis. Other protozoa, such as Isospora belli, Cryptosporidium parvum, and microsporidia, also can colonize the small bowel and cause malabsorption. Selective IgA deficiency may be associated with a flat mucosa and giardiasis. Bacterial colonization of the upper small bowel occurs in some patients with primary immunodeficiency and causes malabsorption, but it responds quickly to treatment with tetracycline or other antibiotics. Periodic administration of intravenous gamma globulin is the major therapy for patients with CVI.

IMMUNOPROLIFERATIVE SMALL INTESTINAL DISEASE AND SMALL BOWEL LYMPHOMA Immunoproliferative small intestinal disease (IPSID) and small bowel lymphoma, also termed Mediterranean lymphoma, is not uncommon in the tropics.73 It usually affects socioeconomically disadvantaged persons. Patients with IPSID present with chronic diarrhea and malabsorption in the second and third decades of life. Abdominal pain is usually a significant complaint. Weight loss, nutritional deficiencies, and clubbing of the fingers are noted, and abdominal masses may be palpated on physical examination. The disease is caused by a clonal proliferation of cells that produce an abnormal alpha heavy chain immunoglobulin, and it can be diagnosed by immunoassay for the alpha heavy chain in the serum. It is suggested that the clonal expansion is driven by an infectious antigen, in a way similar to the link between H. pylori and mucosa-associated lymphoid tissue (MALT) lymphoma of the stomach. Campylobacter jejuni infection has been causally associated with IPSID.74 Mucosal biopsy of the small intestine reveals a dense cellular lymphoplasmacytic infiltrate in the lamina propria, leading to effacement of the crypts. The disease progresses over several years from a relatively benign infiltration of the entire small intestinal mucosa

Chapter 105  Tropical Diarrhea and Malabsorption (stage A) to the development of lymphoplasmacytic and immunoblastic lymphoma (stage C). The disease is staged as other lymphomas by bone marrow examination, by looking for evidence of disease on either side of the diaphragm, and by computed tomography of the abdomen. In patients with stage A disease diagnosed by mucosal biopsy, it is advisable to perform laparoscopy or laparotomy with full-thickness biopsy of the intestine to exclude transmural lymphoma before commencing antibiotic therapy. Areas of bulky tumor are resected before chemotherapy, and fullthickness biopsies of involved areas of intestine and biopsy of enlarged mesenteric nodes are performed. In the premalignant stage A, long-term therapy with antibiotics such as tetracycline can cure the disease. In the more advanced stages of the disease (B and C), chemotherapy or total abdominal irradiation are used (see Chapter 59).

TROPICAL PANCREATITIS AND MALABSORPTION Idiopathic chronic calcific pancreatitis or tropical pancreatitis is endemic in several tropical regions including the Indian subcontinent and southern Africa. Symptoms of recurrent abdominal pain typically develop in childhood or adolescence and often persist for eight to 10 years. Exocrine pancreatic insufficiency, with a history of passing oil in the stool, eventually develops in more than 25% and diabetes mellitus develops in more than 50% of affected patients. The fecal fat is grossly increased, and d-xylose absorption is normal. Vitamin B12 malabsorption is noted in some persons resulting from lack of proteolytic cleavage of the R protein–vitamin B12 complex. The disease is likely to be genetically determined, and both disease-inducing and disease-protective mutations have been noted. The most common mutation involves a serine protease (SPINK1) gene and occurs in about 40% of the patients.76 Whether the association of diabetes mellitus in these patients requires a different or an additional mutation is controversial. Diagnosis is made by detecting pancreatic calcification on plain films or ultrasound examination of the abdomen. The malabsorption is treated by restricting long-chain fats in the diet and by substituting medium-chain triglycerides. Pancreatic enzymes with a high lipase content usually are administered with each meal and are most effective when ingested about halfway through the meal. Therapy of the pain in this disease includes administration of pancreatic enzymes, celiac plexus block, endoscopic removal of calculi, and surgery with drainage of the pancreas.

APPROACH TO THE PATIENT WITH SUSPECTED MALABSORPTION Patients presenting with chronic diarrhea and malabsorption in tropical countries first should be screened for infection with protozoa and helminths by obtaining three samples of stool for microscopy. The stool samples are examined by microscopy of wet smears, directly and after concentration (sedimentation and flotation) techniques for ova and cysts, and with special stains (such as safranine methylene blue or modified acid-fast bacilli stains) for coccidian parasites. Feces also is tested for occult blood to exclude ulcerative conditions of the gastrointestinal tract. Testing for HIV infection after counseling is performed when indicated. Hematologic and biochemical evaluation is undertaken to establish the presence of specific nutrient deficiencies, including folate, vitamin B12, and iron. Increasingly in spe-

cific populations, it is now necessary to evaluate serum immunoglobulin levels and to test for the IgA antitissue transglutaminase antibody to exclude celiac disease. Because tropical sprue remains a major cause of malabsorption in the tropics and because there is no specific diagnostic test for this disease, it is necessary to establish the presence of steatorrhea by fecal fat estimation and to test for d-xylose absorption as an index of mucosal disease. If these tests are abnormal, endoscopic biopsies are obtained from the third or fourth part of the duodenum or from the jejunum, which often allows specific diagnosis of coccidian or helminthic infection, IPSID, and other infiltrative disorders, or it might show the villus flattening and crypt elongation and inflammatory infiltration associated with tropical sprue. If the duodenal biopsy does not provide a diagnosis, it is necessary to image the small bowel using barium or CT enteroclysis. At this stage, it is also appropriate to perform ileocolonoscopy and to take biopsies from the ileum and different segments of the colon. Double-balloon enteroscopy is sometimes necessary to obtain biopsies from suspected abnormal areas in the jejunum or ileum beyond the reach of the standard endoscopes. Double-balloon enteroscopy has largely replaced laparoscopy or laparotomy and enteroscopy with full-thickness biopsy, which was occasionally necessary in patients with malabsorption in the tropics.

KEY REFERENCES

Behera B, Mirdha BR, Makharia GK, et al. Parasites in patients with malabsorption syndrome: A clinical study in children and adults. Dig Dis Sci 2008; 53:672-9. (Ref 67.) Freedman DO, Weld LH, Kozarsky PE, et al. Spectrum of disease and relation to place of exposure among ill returned travelers. N Engl J Med 2006; 354:119-30. (Ref 9.) Fry LC, Bellutti M, Neumann H, et al. Utility of double-balloon enteroscopy for the evaluation of malabsorption. Dig Dis 2008; 26:134-9. (Ref 55.) Ghoshal UC, Ghoshal U, Ayyagari A, et al. Tropical sprue is associated with contamination of small bowel with aerobic bacteria and reversible prolongation of orocecal transit time. J Gastroenterol Hepatol 2003; 18:540-7. (Ref 34.) Hanson JP. Tropical sprue in Far North Queensland. Med J Aust 2005; 182:536-537. (Ref 26.) Kelly P, Menzies I, Crane R, et al. Responses of small intestinal architecture and function over time to environmental factors in a tropical population. Am J Trop Med Hyg 2004; 70:412-19. (Ref 2.) Lo A, Guelrud M, Essenfeld H, Bonis P. Classification of villous atrophy with enhanced magnification endoscopy in patients with celiac disease and tropical sprue. Gastrointest Endosc 2007; 66:377-82. (Ref 56.) Mathan M, Mathan VI, Baker SJ. An electron-microscopic study of jejunal mucosal morphology in control subjects and in patients with tropical sprue in southern India. Gastroenterology 1975; 68:17-32. (Ref 45.) Morales M, Galván E, Mery CM, et al. Exocrine pancreatic insufficiency in tropical sprue. Digestion 2001; 63:30-4. (Ref 49.) Owens SR, Greenson JK. The pathology of malabsorption: Current concepts. Histopathology 2007; 50:64-82. (Ref 43.) Ramakrishna BS, Venkataraman S, Mukhopadhya A. Tropical malabsorption. Postgrad Med J 2006; 82:779-87. (Ref 5.) Ranjan P, Ghoshal UC, Aggarwal R, et al. Etiological spectrum of sporadic malabsorption syndrome in northern Indian adults at a tertiary hospital. Indian J Gastroenterol 2004; 23:94-8. (Ref 6.) Rickles FR, Klipstein FA, Tomasini J, et al. Long-term follow-up of antibiotic-treated tropical sprue. Ann Intern Med 1972; 76:203-10. (Ref 57.) Thomas PD, Forbes A, Green J, et al. Guidelines for the investigation of chronic diarrhoea, 2nd edition. Gut 2003; 52 Suppl 5:v1-15. (Ref 16.) Wahnschaffe U, Ignatius R, Loddenkemper C, et al. Diagnostic value of endoscopy for the diagnosis of giardiasis and other intestinal diseases in patients with persistent diarrhea from tropical or subtropical areas. Scand J Gastroenterol 2007; 42:391-6. (Ref 68.) Westergaard H. Tropical sprue. Curr Treat Options Gastroenterol 2004; 7:7-11. (Ref 58.) Full references for this chapter can be found on www.expertconsult.com.

1831

CHAPTE R

106 Whipple’s Disease Matthias Maiwald, Axel von Herbay, and David A. Relman

CHAPTER OUTLINE History  1833 Epidemiology  1833 Microbiology and Genomics  1834 Pathogenesis and Immunology  1835 Clinical Features  1836 Small Intestine and Lymphatic System  1836 Central Nervous System  1836 Cardiovascular System  1837

Whipple’s disease is a chronic systemic infection caused by a Gram-positive bacterium, Tropheryma whipplei. The small intestine is affected most often, but a variety of other organs also may be involved, including the joints, the central nervous system (CNS), and the heart. Clinical symptoms and findings are protean and include weight loss, diarrhea, malabsorption, fever, arthralgias, skin hyperpigmentation, and dementia. Whipple’s disease was considered to be uniformly fatal in the preantibiotic era, but today, treatment with antibiotics usually leads to clinical remission. Many open questions still surround its pathogenesis, but host immunologic factors are presumed to influence susceptibility to the disease.

HISTORY In 1907, the pathologist George H. Whipple reported in detail, the case of a 36-year-old male physician-missionary who died after a five-year illness involving arthritis, chronic cough, weight loss, and chronic diarrhea.1 At autopsy, Whipple found lipid deposits in the intestinal mucosa as well as in mesenteric and retroperitoneal lymph nodes. Microscopic examination further revealed a large number of macrophages with foamy cytoplasm in the lamina propria of the small intestine. Whipple suspected a disorder of fat metabolism and proposed the term intestinal lipodystrophy for the disease subsequently bearing his name. In the following decades, only a few cases were reported and the diagnosis uniformly was made at autopsy. The first antemortem diagnosis was made in 1947 based on findings in a mesenteric lymph node removed at laparotomy,2 and the first diagnosis by peroral intestinal biopsy was made in 1958.3 In 1949, Black-Schaffer4 introduced the periodic acid-Schiff (PAS) stain to the histopathologic diagnosis of Whipple’s disease. Inclusions in macrophages stained red using this stain, thus documenting that the intracellular material was glycoprotein rather than lipid.

Musculoskeletal System  1837 Other Manifestations  1837 Pathology  1838 Small Intestine  1838 Extraintestinal Pathology  1839 Diagnosis  1840 Differential Diagnosis  1840 Treatment and Prognosis  1841

The first report of successful antibiotic treatment (using chloramphenicol) was published in 1952.5 In 1961, two groups independently visualized bacteria by electron microscopy in affected tissues6,7; subsequent reports confirmed these observations. The bacteria associated with Whipple’s disease are rod-shaped and of uniform size. Consistent positive therapeutic effects were achieved with antibiotic treatment.8 These findings and the positive PAS reaction4 suggested that the disease was unlikely to be a primary disorder of fat metabolism, but rather that it was a bacterial disease; efforts to cultivate this bacterium before 2000 failed to yield reproducible or consistent results. The nature of the bacterium remained obscure until the early 1990s, when its 16S ribosomal DNA (rDNA) sequence was determined and phylogenetic analysis established the relationship of the bacterium to the actinomycetes.9,10 The name Tropheryma whippelii was introduced,10 and the novel 16S rDNA sequence provided the basis for sensitive diagnostic testing using the polymerase chain reaction (PCR). In situ hybridization experiments showed that the unique bacterial 16S rRNA sequence colocalized with areas of pathology, thus supporting the relevance of the sequence and organism, the presence of which was thereby inferred.11 Further advancement came with successful propagation of the Whipple’s disease bacterium in coculture with human fibroblast cells.12 At that point, the bacterium formally was described as a new species, and its name was modified to Tropheryma whipplei.13 With the availability of adequate amounts of purified genomic DNA, the complete genome sequences of two different bacterial isolates were determined and published in 2003.14,15

EPIDEMIOLOGY Whipple’s disease is a rare disorder. The first comprehensive epidemiologic survey was performed by Dobbins in 1987,8 compiling information on 696 patients and including

1833

1834

Section X  Small and Large Intestine 617 published and 79 unpublished cases recorded through 1986. According to this analysis, Whipple’s disease is a sporadic disorder with a predilection for middle-aged white men. Data on age and sex were available for 664 patients; 86% were male, and the mean age at diagnosis was 49 years. Most patients were white; only 10 were African, one was a Native American, three patients were from India, and one was Japanese. Most of the patients originated from Europe (373 patients) or from the United States (246 patients). Within Europe, Germany (114 patients) and France (91 patients) were strongly represented. Relatively few cases originated from South America (11 patients) and Australia (13 patients). A small epidemiologic study from western Switzerland calculated the incidence of Whipple’s disease to be approximately 0.4 per million of the population per year.16 A similar incidence of 0.4 per million per year was estimated for Germany.17 An epidemiologic analysis of 110 patients with Whipple’s disease in Germany, identified between 1965 and 1995, noted the incidence of cases to be relatively stable over three decades and a relatively even geographic distribution of the patients’ residences.18 There are only a few observations of geographically confined case clusters (up to seven cases).19-21 In recent decades, several studies have indicated a statistically significant increase in the age of patients at diagnosis and an increasing percentage of female patients.18,22,23 Currently, patients are first dianosed at a mean age of 56 years.18 It has been speculated that the increasing use of antibiotics for unrelated complaints may be a contributing factor in delaying the age of onset of Whipple’s disease. In a cohort of 191 patients whose Whipple’s disease was diagnosed between 1992 and 2007, 75% were male and 25% were female.24 There are virtually no cases in children and young adults. One remarkable epidemiologic feature in Dobbins’ analysis8 was the strong representation of patients with occupations in the farming and building trades, involving outdoor work or frequent contact with animals or soil: of 191 patients for whom data were available, 43 (22%) were farmers and 10 (5%) were carpenters; patients in all farming-related trades accounted for 34% of the total; by comparison, the fraction of farm workers among the total workforce in the analyzed countries was approximately 10%.

Figure 106-1.  Photomicrograph of Tropheryma whipplei in a culture of cerebrospinal fluid from a patient with Whipple’s disease. Bacteria were stained with YO-PRO-1 nucleic acid dye. Note the distinctive appearance of small rods arranged in chains. (Scale bar represents micrometers.) (From Maiwald M, von Herbay A, Fredricks DN, et al. Cultivation of Tropheryma whipplei from cerebrospinal fluid. J Infect Dis 2003; 188:801.)

MICROBIOLOGY AND GENOMICS After many unsuccessful attempts to cultivate the bacterium associated with Whipple’s disease, successful propagation of T. whipplei was reported in 2000, using infected heart valve tissue in coculture with human fibroblast cells.12 Seven culture passages were performed, over a period of 285 days, and the 16S rDNA sequence of T. whipplei was detected after each passage. The initial estimate of the doubling time of the bacterium was 18 days, which represents extremely slow growth. Since the initial report, approximately 20 to 30 additional strains of T. whipplei have been isolated from various clinical specimens, including infected heart valves, duodenal biopsy specimens, ocular vitreous fluid, cerebrospinal fluid (CSF), synovial fluid, heparinized blood, mesenteric lymph node tissue, muscle tissue, and feces.25-31 Methods to determine growth and identity of the bacterium in culture include immunofluorescence,25,28 nucleic acid staining (Fig. 106-1),29 endpoint PCR and sequencing,25,28,29 quantitative PCR,27,29 electron microscopy

Figure 106-2.  Scanning electron micrograph of Tropheryma whipplei from cerebrospinal fluid in fibroblast cell culture. Note the small rod-shaped bacteria outside of cells, arranged in cords. Original magnification ×20,000. (From Maiwald M, von Herbay A, Fredricks DN, et al. Cultivation of Tropheryma whipplei from cerebrospinal fluid. J Infect Dis 2003; 188:801.)

(Fig. 106-2),29 and in situ hybridization.29 Subsequent studies arrived at estimates of shorter bacterial doubling times between 28 hours and four days.27,29,32 A cell-free (axenic) medium has been designed, using data from the genome sequence32; it consists of cell culture medium sup-

Chapter 106  Whipple’s Disease plemented with extra amino acids. Despite these experimental advancements, however, culture of T. whipplei at present is feasible only in specialized laboratories and is not suitable for routine diagnostic purposes. Phylogenetic analysis of the T. whipplei 16S rDNA sequence, initially amplified by broad-range PCR from infected tissue, established that the bacterium is an actinomycete, a member of the class Actinobacteria.9,10 A more detailed analysis places the organism in an intermediate phylogenetic position between the genus Cellulomonas (with the common group A peptidoglycan) and a rare group of actinomycetes (with group B peptidoglycan; i.e., a different linkage of cell wall components).33 Both groups of organisms consist predominantly of environmental bacteria that are found in soil and water and on plants. Nevertheless, the relationships of T. whipplei to any of the other known actinomycetes are quite distant (<92% 16S rRNA sequence similarity). Differences among strains of T. whipplei first were observed in the 16S-23S rDNA intergenic spacer sequ­ ence34,35; seven different 16S-23S rRNA spacer sequence types have been described so far.34-36 One study found the two most common types, 1 and 2, in a similar ratio (≈1 : 2) among patients from the United States, Germany, and Switzerland.34 In any given patient, the same spacer type was found in different anatomic compartments—for example, intestine, blood, CSF34—which argues for systemic dissemination of a single bacterial strain in a patient with Whipple’s disease. Additional variability among strains was found in a 23S rRNA insertion sequence,37 in the groEL heat-shock protein gene,38 and at a series of variable number of tandem repeat (VNTR) loci in T. whipplei.39 Genome sequence information has been used to distinguish strains of T. whipplei. One study found that T. whipplei strains were quite heterogeneous when PCR-amplified sequences of four highly variable genomic loci were compared.40 Another study, based on microarray hybridization of DNA from cultivated isolates, found that genomic divergence of strains is due mostly to differences in members of the novel WiSP (T. whipplei surface protein) family but that aside from these differences, genome content is relatively conserved.31 So far, there is no indication that different strain types are associated with different clinical features or with different geographic locations. The genome of T. whipplei is quite small for a bacterium.14,15 It consists of approximately 926,000 base pairs and is the smallest of all known actinomycete genomes. Its guanine + cytosine (G+C) content of 46% is unusually low for actinomycetes, which generally are organisms with high genomic G+C content. Genome size contraction is believed to have resulted from gene loss during the evolution of T. whipplei and is a general feature of bacteria that occupy a host-dependent ecological niche. This bacterium lacks various metabolic capabilities, including deficiencies in carbohydrate and energy metabolism and amino acid biosynthesis, which makes it dependent on supplies from its host environment. Two more features of the T. whipplei genome are quite remarkable. A relatively large fraction of its genes is dedicated to the biosynthesis of cell surface molecules, and features of the genome suggest multiple “built-in” mechanisms for antigenic variation involving the WiSP family. These mechanisms are believed to involve VNTR sequences, which are known to be associated with antigenic phase variation in other organisms. There also are two unusual, large genomic regions of noncoding repetitive DNA that are thought to contribute to genetic plasticity.14 A comparative

analysis revealed that the two sequenced strains are distinguished by inversion of a large segment of the genome (≈57%).15 WiSP family protein genes at each end of the large segment serve as anchoring points for the inversion. Taken together, these features suggest that interaction of T. whipplei with its host and its evasion of a host immune response are major parts of the organism’s lifestyle; these factors might contribute to its ability to sustain a chronic infection.

PATHOGENESIS AND IMMUNOLOGY The exact source of infection and the sequence of events leading to bacterial multiplication and pathologic changes are still unclear. Because of the prominence of intestinal manifestations, an oral route of acquisition is assumed,8 but this is unproven. Current concepts hold that once T. whipplei has been acquired, it enters the proximal small intestine where the bacteria invade the mucosa. Evidence for this is provided by electron microscopy.41,42 Fluorescence in situ hybridization demonstrates colocalization of T. whipplei 16S rRNA with areas of pathologic change and indicates that most viable bacteria are extracellular and located just below the epithelial basement membrane in the lamina propria (see later).11 From the intestinal mucosa, bacteria are thought to spread via lymphatics into mesenteric and mediastinal lymph nodes and into the systemic circulation. Relatively little is known about the natural habitats of T. whipplei. Only humans seem to be affected, with outdoor workers more strongly represented than other professional groups.8 A PCR-based search in effluent from a German sewage treatment plant revealed positive results for T. whipplei DNA in 25 of 38 samples from five different plants43; another study found T. whipplei DNA in 17 of 46 samples from Austrian sewage plants.44 Several studies have reported detection of T. whipplei DNA in saliva, gastric juice, intestinal biopsies, and stool of asymptomatic persons,44-48 whereas several other PCR-based studies of intestinal biopsy samples have provided little or no evidence of infection in persons without the histologic features of Whipple’s disease.49-53 Two studies that reported T. whipplei DNA in the stool of asymptomatic persons found higher rates of positive results in sewage workers than in other groups of people (25% vs. 7% and 12% vs. 4%, respectively).44,48 A concept of asymptomatic healthy carriage of T. whipplei has been proposed but not confirmed.47,48 So far, there is no evidence for person-to-person transmission of T. whipplei,8 and there are only a few reports of the disease in relatives of persons with the disease.54,55 The genome sequence suggests that the organism is highly dependent on nutrients from other sources.14,15 The proposed extracellular location in the villus tips below the intestinal basement membrane,11 a site with rich influx of nutrients, fits these requirements well. Several abnormalities of immune function have been observed in patients with Whipple’s disease,8,56,57 including transient (i.e., during active disease), as well as persistent (i.e., after therapy) abnormalities; the persistent abnor­ malities are presumed to serve as predisposing factors for development of disease. Precisely defined immune defects, however, such as the physical or functional absence of specific cell types, mediators, or receptors, have not been identified. Small case series58,59 have described an over-

1835

1836

Section X  Small and Large Intestine representation of the HLA-B27 haplotype in patients with Whipple’s disease, but others60,61 have not supported this association. Humoral immunity in patients with Whipple’s disease grossly appears to be normal.8 During active disease, reduced CD4/CD8 T-cell ratios (both in the lamina propria and in peripheral blood), reduced proliferation of peripheral T cells to stimulating agents (e.g., phytohemagglutinin, concanavalin A), and reduced delayed-type hypersensitivity reactions to common antigens in skin tests have been obsesrved.56,58,62 This may be a consequence of malnutrition, however, rather than a pre-existing immunological abnormality. One study showed that the monocytes of a patient with Whipple’s disease exhibited an impaired ability to degrade bacterial antigens,63 which is consistent with the prolonged persistence of bacterial remnants in intestinal macrophages after therapy observed in histologic studies of Whipple’s disease.8,64 Other immunologic abnormalities persist after therapy: reduced numbers of peripheral blood monocytes that express the alpha chain of complement receptor 3 (CD11b),56 a reduced capability of peripheral blood monocytes to produce interleukin (IL)-12 on stimulation with bacterial antigens,57 and a dysregulation of mononuclear cell function, such that the components of a Th1-type immune response are reduced and those of a Th2 immune response are increased.65 The latter observation was supported in a study with specific T. whipplei antigen from cultivated bacteria: Duodenal lymphocytes and peripheral blood mononuclear cells from healthy people exhibited robust Th1 type immune reac­tivity, but those of patients with Whipple’s disease showed reduced or absent T. whipplei-specific Th1 responses.66 Furthermore, it was shown that macrophages from duo­denal tissue of a patient with Whipple’s disease exhibited a transcriptional pattern associated with a Th2 immune response.67 Another investigation revealed that IL-16, a cytokine that is constitutively expressed in T cells, mast cells, dendritic cells, and circulating monocytes and that is released during apoptosis, was expressed at high levels and released by macrophages upon infection with T. whipplei; when added to the experimental model, IL-16 promoted T. whipplei replication in both monocytes and macrophages.68 Circulating blood levels of IL-16 and nucleosomes (a marker of apoptosis) also were found to be elevated in patients with active Whipple’s disease compared with patients with treated Whipple’s disease and compared with controls.69 Several reports describe secondary or opportunistic infections in patients with Whipple’s disease,64,70-72 the most common being Giardia lamblia, which is observed in 8% to 12% of patients. Rare cases of infections with Pneumocystis jirovecii, Cryptosporidium parvum, Nocardia spp., Mycobacterium tuberculosis, Serratia marcescens, Candida spp., dermatophytes, and Strongyloides stercoralis also have been recorded. In addition, T. whipplei infection has been detected by PCR in one patient with acquired immunodeficiency syndrome (AIDS).73 A possible role of the immune system in clearing T. whipplei infection was further suggested by the report of a patient without adequate response to antibiotic treatment, who eventually benefited from adjuvant interferon-γ treatment74; however, this effect was not reproduced in other patients. Taken together, all these observations and laboratory findings suggest that there are immunologic factors, including quantitative deficiencies in macrophage activation, microbial phagocytosis, and the regulation of a cellular immune response, that facilitate the occurrence of Whipple’s disease.

CLINICAL FEATURES Whipple’s disease usually is a systemic infection, and almost any organ or organ system can be affected.8 Manifestations in the intestinal tract are reported most commonly and are largely responsible for the classic clinical features of Whipple’s disease.20,75 In many patients, arthralgias precede intestinal symptoms by several years (1 to 10 years; up to 30 years reported), although it is unclear whether joints are infected at that point; in some cases, low-grade intermittent fever also occurs for years before the diagnosis is made.23,76 More recent reports suggest a wider spectrum of extraintestinal manifestations, most likely reflecting advances in diagnostic procedures. Patients tend to have less-advanced disease at the time of diagnosis, possibly as a result of earlier detection.8,23

SMALL INTESTINE AND LYMPHATIC SYSTEM

Bacterial and macrophage-predominant inflammatory cell infiltration of the small intestinal mucosa and obstruction of mesenteric lymph nodes lead to a malabsorption syndrome with weight loss, diarrhea, and abdominal pain as the dominant signs and symptoms.20,23,75-77 Weight loss in amounts of 5 to 20 kg occurs gradually, usually over a period of at least a year, sometimes resulting in severe cachexia in the terminal stage of untreated disease.8,20,76 Diarrhea can consist of voluminous steatorrheic stools or may be watery.20 Occult gastrointestinal bleeding is not uncommon, and in some cases gross gastrointestinal bleeding occurs.8,20 Abdominal (mesenteric and retroperitoneal) and peripheral lymphadenopathy are common,20,23,76,77 and in some instances, enlarged abdominal lymph nodes have raised the suspicion of malignancy.77 In rare instances, malignant lymphomas have occurred in patients with Whipple’s disease.78-80 Barium examination of the intestinal tract can reveal nonspecific abnormalities, such as prominent and edematous duodenal and jejunal folds and intestinal dilatation, that also are found in other malabsorption syndromes (Fig. 106-3).8,76 Computed tomography (CT) (Fig. 106-4) or magnetic resonance imaging (MRI) can detect retroperitoneal or para-aortic lymphadenopathy.76,81 Enlarged abdominal lymph nodes have a hypodense appearance on CT scans and are hyperechoic on ultrasonograms.71,82 Laboratory examinations in patients with intestinal Whipple’s disease often reveal an increased erythrocyte sedimentation rate, decreased serum carotene level, decreased serum iron concentration, anemia, decreased serum protein levels, proteinuria, and elevated stool fat content.8,76

CENTRAL NERVOUS SYSTEM

Symptomatic CNS manifestations have been reported in 10% to 43% of patients with intestinal Whipple’s disease with more recent series reporting lower rates.20,23,76,77 Neurologic disease can occur concurrently with intestinal manifestations at the time of diagnosis, but it is more common at the time of clinical relapse, which can occur during or after treatment.8,83 It is thought that bacteria enter the CNS early in the course of disease and that because most drugs do not penetrate the CNS well, the bacteria persist during medical treatment. The result is the impression that intestinal disease goes into remission initially and neurologic disease develops subsequently, even as antibiotics continue to be given. Relapses affecting the CNS are ominous because they tend to be refractory to renewed antibiotic treatment.8,84 Although rare, instances of isolated primary neurologic

Chapter 106  Whipple’s Disease cases,87,88 and they have not yet been documented in other CNS diseases. Results of neuroimaging (CT or MRI scans) may be normal or reveal mild to moderate brain atrophy or focal lesions without a predilection for specific sites.87,88,90 These abnormalities are not specific for Whipple’s disease, but focal lesions may be used to guide stereotactic biopsies, which in most cases reveal characteristic histology87; MRI appears to be more sensitive than CT scan.88 Results of standard CSF examinations usually are normal, although sometimes there is mild pleocytosis.8,87 CSF cytology often reveals PASpositive sickle-form particle-containing cells, and PCR often yields positive results for T. whipplei DNA, even in a considerable proportion of neurologically asymptomatic patients.91

CARDIOVASCULAR SYSTEM

Figure 106-3.  Barium contrast study of the small intestine from a patient with Whipple’s disease. There is marked thickening of the plicae circulares and a loss of the normal delicate mucosal relief pattern. The small intestine is slightly dilated. (Courtesy of Elihu Schimmel, MD, Boston, Mass.)

Cardiac manifestations of Whipple’s disease include endocarditis, myocarditis, and pericarditis.23 In one autopsy series from the preantibiotic era,75 valvular endocarditis with vegetations was noted in 58% of cases. In contrast, in a more recent series of Whipple’s disease,23 clinically apparent endocarditis was less common (3 of 52 patients). All valves may be affected, but the mitral valve is most commonly pathologically altered, and involvement of the aortic valve leads to the most significant symptoms.8 Some patients require valve replacement. PAS-positive macrophages and bacteria have been documented in native92 and porcine prosthetic93,94 valve tissue and in the myocardium95 by histology and electron microscopy, respectively. Based on PCR testing of excised heart valve tissue, T. whipplei is increasingly being recognized as an agent of “blood culturenegative endocarditis,” even in patients with only minor or no apparent intestinal involvement.94,96-100

MUSCULOSKELETAL SYSTEM

Figure 106-4.  Computed tomography scan showing extensive retroperi­ toneal and mesenteric adenopathy caused by Whipple’s disease, and simulating lymphoma. (Courtesy of Mark Feldman, MD, Dallas, Tex.)

Whipple’s disease have been reported in patients without intestinal or other manifestations.85,86 Two reviews summarized the neurologic findings of Whipple’s disease in 84 and 122 published cases, respectively.87,88 Common findings are progressive dementia and cognitive changes (28% to 71%), supranuclear ophthalmoplegia (32% to 51%), and altered level of consciousness (27% to 50%); less common are psychiatric symptoms, hypothalamic manifestations (e.g., polydipsia, hyperphagia, insomnia),89 cranial nerve abnormalities, nystagmus, seizures, and ataxia. Two signs that are characteristic of CNS Whipple’s disease are oculomasticatory myorhythmia and oculofacial skeletal myorhythmia,87 each consisting of slow rhythmic and synchronized contractions (≈1/sec) of ocular, facial, or other muscles; they occur in less than 20% of

Oligoarthralgias or polyarthralgias, usually involving the ankles, knees, elbows, or fingers, are a common complaint of patients with Whipple’s disease.8 Rheumatoid factor usually is absent. Destructive joint changes or accumulation of synovial fluid are rare, but when they are present, they are accompanied by PAS-positive macrophages (by histology), bacteria (by electron microscopy), or DNA of T. whipplei (by PCR) in synovial tissue or joint fluid.8,101,102 One synovial fluid specimen yielded T. whipplei in culture.30 Sacroiliitis and spondylitis may occur, but ankylosing forms are rare, and there does not seem to be a strong association of these manifestations with HLA-B27.8 Rare manifestations are infectious spondylodiskitis103 and prosthetic joint infection.104

OTHER MANIFESTATIONS

One common feature of Whipple’s disease is skin hyper­ pigmentation, which has been found in 17% to 66% of patients.20,23,75-77 This finding tends to occur in light-exposed areas of the skin and is unrelated to adrenal dysfunction or hyperbilirubinemia. Histopathologic changes in the skin, however, are extremely rare,105 and the pathophysiology of this hyperpigmentation is unknown. Ocular manifestations of Whipple’s disease have been described and are diverse, but rare. These include uveitis, vitritis, retinitis, retrobulbar neuritis, and papilledema.8 They usually are associated with CNS disease, and almost all reported patients also had clinical or histologic evidence of intestinal involvement. PAS-positive macrophages, DNA of T. whipplei, and visible bacilli may be detected in vitrectomy specimens.106,107 One case of uveitis has been reported in which the vitreous fluid and one intestinal biopsy speci-

1837

1838

Section X  Small and Large Intestine men yielded positive PCR results, although intestinal histology was normal.107 Another uveitis case was a source for a positive culture.26 Chronic cough was a symptom in Whipple’s original patient and was reported relatively frequently in earlier series75 but infrequently since then.23 Some patients have pleuritis with effusion or granulomatous pulmonary disease resembling sarcoidosis.8

PATHOLOGY SMALL INTESTINE

On gross inspection, the mucosa of the distal duodenum and jejunum is abnormal in most patients with Whipple’s disease. Whitish-to-yellow plaque-like patches are observed in approximately three quarters of patients (Fig. 106-5); alternatively, the mucosa can appear pale yellow.64,108 Abnormal villus structure and mild mucosal flattening become evident using magnifying optics. The histopathologic features of intestinal Whipple’s disease are quite distinctive. Viewed with light microscopy, the visible patches reflect lipid deposits or lymphangiec­ tasia, whereas villus distention results from infiltration by macrophages in the lamina propria (Fig. 106-6). The swollen cytoplasm of macrophages appears foamy when stained with hematoxylin-eosin, but numerous granular particles become visible when the PAS stain is used (see Fig. 106-6). These particles correspond to phagolysosomes filled with numerous T. whipplei, and the positive reaction with PAS reflects the glycoprotein content of the bacterial cell walls. Single extracellular bacteria are barely visible with conventional light microscopy because of their small size, but they become evident in the mucosal stroma with high-resolution light microscopy and electron microscopy (Fig. 106-7). The number of bacteria varies greatly among patients. Electron microscopy shows uniformity in size and shape of the bacteria, with an external diameter of 0.2 to 0.25 µm

Figure 106-5.  Endoscopic view of the distal duodenum in a patient with untreated Whipple’s disease. The plicae circulares are swollen, and the mucosal surface is intact. Numerous whitish patches, reflecting lipid deposits, are present within the mucosa (see Fig. 106-6). (Courtesy of Hans Jörg Meier-Willersen, MD, Heidelberg, Germany.)

and a length of up to 2.5 µm.42,109 There is an electron-dense outer layer that is not found in other bacteria; some authors have speculated that this unusual membrane may be of host origin.109 Most of the structurally intact bacteria, including dividing forms, are found outside of host cells in the lamina propria (see Fig. 106-7).8,109 In contrast, the intracellular bacteria in macrophages often are found in various stages of degradation. Findings based on fluorescent in situ hybri­ dization using specific T. whipplei 16S rDNA probes support and extend the findings derived from electron microscopy11; the 16S rRNA signal from metabolically active bacteria is found in the intestinal lamina propria, just beneath the basement membrane, but it is absent from the PAS-positive macrophages (Fig. 106-8). Thus, T. whipplei appears to prefer extracellular environments within the host, despite its association with eukaryotic cells. Following the success of T. whipplei laboratory cultivation, antisera have been raised against T. whipplei isolates and have been used for testing small intestine, lymph node, heart, and brain tissue12,110,111; observed staining patterns are similar to those of PAS staining. These T. whipplei-specific antisera did not react with a number of control bacteria and tissues affected by other diseases; however, apart from a few experimental studies, immunohistology has not yet been more widely adopted for routine diagnostic purposes. Mucosal infiltration with PAS-positive cells usually is diffuse, but patchy lesions may be present in some patients. The inflammatory reaction generally is dominated by macrophages, whereas neutrophils, eosinophils, lymphocytes, and plasma cells are scarcer.64,112 This cellular composition is unusual for an invasive bacterial infection, a feature that suggests a disturbance of mobilization and chemotaxis of leukocytes.64 Variants of the usual histologic findings occur in some patients. These include rare cases with PAS-positive macrophages that are located exclusively in the submucosa and rare cases with epithelioid granulomas in the affected

Figure 106-6.  Histology of the small intestinal mucosa in the same patient as in Fig. 106-5. A villus is distended by an infiltrate of macrophages that contain periodic acid–Schiff–positive granular particles (type 1 cells) and by lipid droplets. The epithelial layer is intact. Original magnification ×84.

Chapter 106  Whipple’s Disease N

BM

Figure 106-7.  Electron microscopy of a small intestinal biopsy specimen in a patient with untreated Whipple’s disease. Just beneath the epithelial basement membrane (BM), the lamina propria is densely infiltrated by extracellular rod-shaped bacteria. The bacteria are uniform in size and structure. Some of them are dividing (arrow). N indicates the nucleus of an adjacent enterocyte.

mucosa.64 Taken together, the intestinal histopathology of Whipple’s disease demonstrates some heterogeneity. Although the characteristic lesions are almost invariably present in the proximal small bowel, they might continuously extend as far as to the terminal ileum.75 Occasionally, the diagnosis is first made with ileal biopsies obtained during colonoscopy. During treatment, the histologic findings in the intestinal mucosa change substantially but slowly, over several months or more.41,64,113 Apart from a continuous decrease in the number of PAS-positive macrophages, the pattern of cellular infiltration of the mucosa changes with time from diffuse to patchy; this feature necessitates multiple biopsies during follow-up endoscopic examinations. Mucosal infiltration shifts from the upper part of the mucosa (i.e., villi) to the lower part of the mucosa (i.e., pericryptal lamina propria) and submucosa. More significantly, the cytologic aspects of the PAS-positive macrophages undergo changes.64 Before treatment, most macrophages have numerous granular particles in the cytoplasm that stain intensely red with PAS (type 1 macrophages; see Fig. 106-6). Within one to six months of treatment, the percentage of type 1 macrophages gradually decreases, and, in parallel, cells with only some coarse granular inclusions and a background of diffusely or finely granular, more faintly PAS-positive cytoplasm (type 2 macrophages) increase in number. After six to 15 months, most macrophages that are still present have diffuse and faintly PAS-positive material in their cytoplasm but lack granular inclusions (type 3 macrophages). Type 3 macrophages contain only filamentous remnants of bacteria.64 Thus, their positive PAS reaction reflects the presence of glycoprotein residues of degraded bacterial cell walls. Even in adequately treated patients, some type 3 macrophages usually persist, occasionally for more than 10 years; in fact, the finding of type 3 macrophages alone is consistent with intestinal remission. Despite documented clinical remission of intestinal disease, however, some patients still harbor viable T. whipplei and can later develop extraintestinal Whipple’s disease. Thus, the prognostic value of intestinal histology during the follow-up of patients is limited.64

EXTRAINTESTINAL PATHOLOGY

Figure 106-8.  Fluorescent in situ hybridization of a small intestinal mucosal biopsy specimen in a patient with Whipple’s disease. In this confocal micrograph of a single villus, nuclei of human cells are green, the intracellular cytoskeletal protein vimentin is red, and Tropheryma whipplei ribosomal RNA (rRNA) is blue. The T. whipplei rRNA signal is most intense in the extracellular spaces of the lamina propria, immediately subjacent to the basal membrane. ×200. (Courtesy of David N. Fredricks, MD, Seattle, Wash.)

Autopsy reports of untreated patients with Whipple’s disease have revealed involvement of virtually any organ and tissue.75,114 As with intestinal disease, the histologic hallmark of extraintestinal involvement is the presence of intracellular PAS-positive granular particles; the diagnostic significance of these lesions in extraintestinal tissues is limited, however, and additional evidence is required for the diagnosis of Whipple’s disease. Rod-shaped bacteria have been documented by electron microscopy in many extraintestinal organs, including liver, heart, lung, brain, eye, lymph nodes, bone marrow, and spleen.8 Two different types of lymph node lesions are common in Whipple’s disease. Abdominal nodes generally contain lipid deposits that induce a granulomatous foreign body type of reaction.1 Peripheral lymph nodes (inguinal, axillary, cervical) generally do not contain lipids but feature a toxoplasmosis-like lymphadenitis with small clusters of epithelioid macrophages, some of which have PAS-positive particles that correspond to inclusions with T. whipplei. Rarely, and most commonly in the mediastinum, a third type of lymph node reaction may be observed that resembles sarcoidosis. Whipple’s disease affects diverse regions of the brain. Most commonly, perivascular infiltrates of PAS-positive macrophages are present, as well as tumor-like granulomas

1839

1840

Section X  Small and Large Intestine of variable size, consisting of glial cells with intensely PASpositive granular particles.8 Occasionally, granulomas in the ventricular system cause occlusive hydrocephalus. PASpositive macrophages often can be detected free in the CSF, even in patients without neurologic or psychiatric symptoms (see later).

DIAGNOSIS The vast majority of patients with Whipple’s disease suffer from involvement of the intestinal tract regardless of the presence or absence of gastrointestinal symptoms.23,76,77 Thus, the primary diagnostic approach to a patient with clinically suspected Whipple’s disease is upper endoscopy (see Fig. 106-5) with mucosal biopsy. To avoid sampling errors in patients with patchy lesions, one should obtain approximately five biopsy specimens from regions as far distal as possible within the small intestine.24 Histologic examination with routine hematoxylin-eosin and PAS stains is usually sufficient to reach a diagnosis (see earlier). In some cases, findings may be corroborated with silver stains,1 but the Gram stain is less useful in this infection. Traditionally, electron microscopy has been used as the gold standard for confirming the diagnosis of Whipple’s disease by showing bacteria of characteristic size and shape.8 More recently, PCR analysis has taken over the role as preferred confirmatory test.49,50 Since the molecular characterization of T. whipplei, a number of PCR-based assays have been developed for diagnostic purposes.10,37,49-51,103,107,115 These assays vary by their target sequences and amplification strategy as well as by the amount of validation and diagnostic experience that is available for them. Newer PCR tests increasingly use the real-time PCR platform.94 In almost all patients with a histologic diagnosis of Whipple’s disease, well-designed and well-standardized PCR assays detect T. whipplei DNA in the intestinal mucosa.49,50 PCR from intestinal biopsies appears unsuitable as a primary diagnostic approach, because Whipple’s disease only rarely is diagnosed by PCR in the setting of histologically negative intestinal biopsies,49-53 and T. whipplei DNA has been detected in unaffected control subjects.45-47 In practical terms, normal intestinal histology in the absence of extraintestinal disease that suggests Whipple’s disease excludes the diagnosis, provided that multiple biopsy specimens are examined. Extraintestinal manifestations warrant the examination of specimens from affected sites. Histology and cytology with PAS staining, electron microscopy, and PCR all are useful for this purpose. A small number of cases has been reported with positive PCR results in extraintestinal samples and intestinal histology that is negative for PAS-positive macrophages. Examples of such cases include febrile illness with erythrocyte-associated bacteria, uveitis, endocarditis, and neurologic disease.94,98-100,107,116,117 In an effort to distinguish between true Whipple’s disease and asymptomatic carriage, one study used quantitative real-time PCR on a number of different sample types.117 When both saliva and stool were positive, the sensitivity was 65% and the positive predictive value was 95% for detecting intestinal manifestations. The positive predictive value increased to 100% when the bacterial load was greater than 104 per gram of stool. When PCRs from both stool and saliva were negative, the negative predictive value for intestinal disease was 96% (one of 23 intestinal cases was PCR-negative in both stool and saliva). The sensitivity for diagnosing extraintestinal disease was insufficient.

Similarly, PCR from peripheral blood lacked sensitivity in this study, and its diagnostic value previously has been questioned.118 Considering the systemic nature of the disorder, it is important to evaluate commonly involved organ systems on a routine basis whenever a new diagnosis of Whipple’s disease has been established. Ultrasound examination might reveal enlarged mesenteric lymph nodes that have unusually high echogenicity due to lipid deposits.71 Neurologic examination is indicated, including routine sampling of CSF.91 Based on cytologic or PCR analysis of the spinal fluid, 70% of patients with intestinal Whipple’s disease in one study were found to have CNS infection with T. whipplei, even though they had no neurologic or psychiatric symptoms.91 Imaging studies of the brain generally are not helpful in the absence of neurologic symptoms. In selected patients with Whipple’s disease and anemia, wireless capsule endoscopy might detect a site of bleeding in the small intestine. In one case report, capsule endoscopy revealed diffuse disease and discrete areas of bleeding in the middle and distal portions of the jejunum, although the exact reason for the bleeding was not defined.119 During treatment, diagnostic assessments should be repeated at regular intervals. Endoscopic lesions usually resolve within months but can last for up to a year.108 Intestinal histology improves within several months,64 and PCR assays on intestinal biopsy tissues convert to negative within a time range of about one to 12 months after appropriate therapy is instituted.49 Some PAS-positive macrophages can persist for years,64 even while the patient remains in clinical remission (see earlier). Enlarged abdominal lymph nodes can require more than a year to regress and can result in fibrosis. Follow-up examination of the CSF should include PCR analysis.91 As has been documented in a culture-positive case, T. whipplei can persist in a viable state in the CNS despite prolonged administration of antibiotics.29

DIFFERENTIAL DIAGNOSIS Almost all symptoms and findings of Whipple’s disease are nonspecific. The broad spectrum of possible clinical presentations generates a wide differential diagnosis, involving several subspecialties of medicine: gastroenterology, infectious diseases, rheumatology, cardiology, hematology, neurology, psychiatry, and ophthalmology. Disorders that mimic the histology of Whipple’s disease are uncommon.17 PAS-positive cells in intestinal biopsies can include mucosal smooth muscle cells that are rich in glycogen or plasma cells that contain immunoglobulin (Russell bodies) in the setting of chronic duodenitis; other disorders include macroglobulinemia, intestinal xanthelasmas, and pseudomelanosis duodeni. Rarely, PAS-positive cells reflect intestinal infection with Mycobacterium avium complex and Rhodococcus equi (in HIV coinfected hosts), or Histoplasma capsulatum. Differentiation from Whipple’s disease usually is possible by means of histochemical stains (e.g., stains for acid-fast bacteria and use of diastase) and by immunocytochemistry.17 Sarcoid-like granulomas are rare in Whipple’s disease but can occur in the stomach,120 small intestine,64,121 liver,122 and lymph nodes. A possible relationship between Whipple’s disease and sarcoidosis remains unresolved.123 By means of PCR analysis, thoracic sarcoidosis124 and intestinal sarcoidosis125 tissues were both found to be negative for T. whipplei DNA.

Chapter 106  Whipple’s Disease Table 106-1  Antibiotics Used to Treat Whipple’s Disease DRUGS

DOSe

COMMENTS

REFERENCE(S)

Ceftriaxone

2 g IV once daily

91

Penicillin G + streptomycin Trimethoprim/sulfamethoxazole

6-24 million units IV daily (in divided doses) + 1 g IM once daily 160 mg/800 mg PO twice daily

Induction therapy (first 10-14 days) or salvage therapy Induction therapy (first 10-14 days)

83, 126, 128, 130

Penicillin VK

500 mg PO four times daily

Doxycycline (or tetracycline)

100 mg PO twice daily (500 mg PO four times daily)

Cefixime

400 mg PO twice daily

Rifampin Chloramphenicol

600 mg PO once daily 500 mg PO four times daily

Erythromycin

500 mg PO four times daily

Long-term therapy; first-line drug; good CNS penetration, but CNS relapses can occur Alternative for long-term therapy; limited experience Used for many years, but clinical relapses, including CNS relapses, are well described Alternative drug for long-term therapy; limited experience Second-line drug; good CNS penetration Second-line drug; worrisome side effects Second-line drug; limited experience

8, 20, 83

8, 83 20, 83, 126 130 86, 131 5, 8, 86 8, 127

CNS, central nervous system; IM, intramuscular; IV, intravenous; PO, oral.

Most patients with Whipple’s disease have enlarged abdominal lymph nodes.78 Rare cases have been observed of Whipple’s disease associated with metachronous or synchronous malignant lymphomas,78-80 but the relationship of T. whipplei infection to lymphoma, if any, remains unclear.

TREATMENT AND PROGNOSIS The initial response of Whipple’s disease to antibiotic treatment usually is prompt.126 Diarrhea often resolves within several days, arthralgias often resolve within a few weeks, and significant weight gain occurs within a few months.20 In the 1970s and early 1980s, long-term tetracycline therapy usually was given,23,77 but it became increasingly clear that patients treated in this manner often suffered relapses, many of which affected the CNS.127,128 CNS relapses have a poor prognosis, because they are often refractory to renewed treatment.83 It therefore was suggested that the treatment of Whipple’s disease include antibiotics that cross the blood-brain barrier. Since the mid-1980s, trimethoprimsulfamethoxazole has been used commonly.83,129 Current recommendations for the treatment of Whipple’s disease are based on observations from numerous case reports,8 several clinical series,23,76,77 and retrospective analyses of antibiotic regimens.83,126 In a retrospective analysis of 88 patients,83 relapses were most common after monotherapy with tetracyclines. Of 49 patients treated with tetracycline alone, 21 relapsed, including nine with CNS relapses. Only a small number of relapses (2 of 15 patients treated), none of which involved the CNS, were observed after initial parenteral treatment with penicillin plus streptomycin, followed by long-term oral tetracycline (the Duke regimen). Tetracyclines and trimethoprim-­ sulfamethoxazole were compared in another series of 30 patients.126 Trimethoprim-sulfamethoxazole was superior to tetracyclines in inducing remission (12 of 13 vs. 13 of 22 treatment courses, respectively, including remissions after relapse). CNS relapses occurred in two of 22 patients receiving tetracycline and in one of 13 receiving

trimethoprim-sulfamethoxazole. Despite its clinical efficacy and ability to cross the blood-brain barrier, several reports indicate that relapses, including CNS relapses, also can occur after use of trimethoprim-sulfamethoxazole.130-132 Some patients appear to have benefited from repeated intravenous courses of third-generation cephalosporins.91 Based on these clinical observations, the current recommendation for treatment of Whipple’s disease is to begin with an induction phase using either penicillin G plus streptomycin or a third-generation cephalosporin (e.g., ceftriaxone) followed by treatment with at least one drug that efficiently crosses the blood-brain barrier (e.g., trimethoprim-sulfamethoxazole) for at least one year. An overview of antibiotic treatments, including suggested doses, is given in Table 106-1. Data concerning antibiotic susceptibility have become available from in vitro experiments with T. whipplei in culture as well as from genome analysis. Growth in the presence of antibiotics was assessed with real-time PCR.133,134 In coculture with fibroblasts, the bacterium appears sensitive to doxycycline, macrolides, penicillins, rifampin, teicoplanin, and trimethoprim-sulfamethoxazole; variably sensitive to imipenem; and only moderately sensitive or resistant to cephalosporins, fluoroquinolones, and vancomycin.133 In axenic medium, the results were very similar to those in cell culture, except that T. whipplei was sensitive to ceftriaxone and vancomycin.134 Interestingly, the T. whipplei genome lacks the gene for dihydrofolate reductase, which is the target for trimethoprim action,135 so that the susceptibility to trimethoprim-sulfamethoxazole most likely is based solely on its sulfamethoxazole com­ ponent. Consistent with this, trimethoprim has been found inactive in laboratory experiments.134 Some have proposed treating Whipple’s disease with either sulfamethoxazole or sulfadiazine alone instead of in combination with trimeth­ oprim.136 Another treatment suggestion—to use hydroxychloroquine in combination with doxycycline—came from in vitro experiments showing enhanced bactericidal activity from the known alkalinization of intracellular vacuoles by hydroxychloroquine.133 Clinical experience with this regimen is still limited, and, in addition, the doxycycline component does not penetrate well into the CNS.

1841

1842

Section X  Small and Large Intestine KEY REFERENCES

Bentley SD, Maiwald M, Murphy LD, et al. Sequencing and analysis of the genome of the Whipple’s disease bacterium Tropheryma whipplei. Lancet 2003; 361:637. (Ref 14.) Dobbins WO III: Whipple’s Disease. Springfield, Ill: Charles C Thomas; 1987. (Ref 8.) Fenollar F, Lepidi H, Raoult D: Whipple’s endocarditis: Review of the literature and comparisons with Q fever, Bartonella infection, and blood culture–positive endocarditis. Clin Infect Dis 2001; 33:1309. (Ref 99.) Fredricks DN, Relman DA. Localization of Tropheryma whippelii rRNA in tissues from patients with Whipple’s disease. J Infect Dis 2001; 183:1229. (Ref 11.) La Scola B, Fenollar F, Fournier PE, et al. Description of Tropheryma whipplei gen. nov., sp. nov., the Whipple’s disease bacillus. Int J Syst Evol Microbiol 2001; 51:1471. (Ref 13.) Maiwald M, von Herbay A, Fredricks DN, et al. Cultivation of Tropheryma whipplei from cerebrospinal fluid. J Infect Dis 2003; 188:801. (Ref 29.) Maizel H, Ruffin JM, Dobbins WO III. Whipple’s disease: A review of 19 patients from one hospital and a review of the literature since 1950. Medicine (Baltimore) 1970; 49:175. (Ref 20.) Moos V, Kunkel D, Marth T, et al. Reduced peripheral and mucosal Tropheryma whipplei–specific Th1 response in patients with Whipple’s disease. J Immunol 2006; 177:2015. (Ref 66.)

Raoult D, Birg ML, LaScola B, et al. Cultivation of the bacillus of Whipple’s disease. N Engl J Med 2000; 342:620. (Ref 12.) Raoult D, Ogata H, Audic S, et al. Tropheryma whipplei twist: A human pathogenic actinobacteria with a reduced genome. Genome Res 2003; 13:1800. (Ref 15.) Relman DA, Schmidt TM, MacDermott RP, Falkow S. Identification of the uncultured bacillus of Whipple’s disease. N Engl J Med 1992; 327:293. (Ref 10.) Renesto P, Crapoulet N, Ogata H, et al. Genome-based design of a cellfree culture medium for Tropheryma whipplei. Lancet 2003; 362:447. (Ref 32.) Vital Durand D, Lecomte C, Cathébras P, et al. Whipple disease: Clinical review of 52 cases. Medicine (Baltimore) 1997; 76:170. (Ref 23.) von Herbay A, Ditton HJ, Schuhmacher F, Maiwald M. Whipple’s disease: Staging and monitoring by cytology and polymerase chain reaction analysis of cerebrospinal fluid. Gastroenterology 1997; 113: 434. (Ref 91.) von Herbay A, Maiwald M, Ditton HJ, Otto HF. Histology of intestinal Whipple’s disease revisited: A study of 48 patients. Virchows Arch 1996; 429:335. (Ref 64.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

107 Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Ralph A. Giannella

CHAPTER OUTLINE Susceptibility to Intestinal Infection  1843 Host Defense Factors  1844 Bacterial Factors  1844 Classification of Bacterial Diarrhea  1846 Diagnosis of Infectious Diarrheal Disease  1846 Evaluation of the Patient  1846 Laboratory Diagnosis  1849 Toxigenic Pathogens  1849 Vibrio cholerae  1850 Other Vibrio Species  1852 Aeromonas Species  1853 Plesiomonas shigelloides  1854 Escherichia coli  1854 Invasive Pathogens  1857 Shigella Species  1857 Nontyphoidal Salmonellosis  1861 Typhoid Fever  1864 Campylobacter Species  1867 Yersinia enterocolitica  1868 Viral Pathogens  1869 Rotavirus  1869 Calicivirus  1871 Enteric Adenovirus  1871 Astrovirus  1872 Torovirus  1872 Special Situations  1872 Hospital-Acquired Diarrhea  1872 Acute Diarrhea in Pregnancy  1872

Infectious diarrhea is a major cause of illness throughout the world. Diarrhea is the first or second most common cause of death in most developing countries; its greatest impact is on infants and children. In developing countries, prevalence rates of diarrhea are two to three times higher than in the United States, where there is an average of two episodes of diarrhea per year in children younger than five years of age. In Western countries, diarrhea leads to high morbidity with loss of time from school and work. Overall, physicians in the United States are consulted annually for 8.2 million diarrheal episodes.1,2 A long list of complications, some mild and others life-threatening, can accompany infectious diarrhea. Medical costs and loss of productivity resulting

Traveler’s Diarrhea  1872 Diarrhea in the Elderly  1875 Dysentery versus Ulcerative Colitis  1876 Tuberculosis of the Gastrointestinal Tract  1877 Pathogenesis  1877 Classification and Distribution of Disease  1877 Pathology  1877 Clinical Features  1877 Diagnosis  1877 Treatment  1878 Bacterial Food Poisoning  1878 Staphylococcus aureus  1879 Clostridium perfringens  1881 Listeria Species  1882 Bacillus cereus  1882 Botulism  1882 Bacillus anthracis  1883 Fish Poisoning  1884 Ciguatera  1884 Scombroid  1884 Overview of Treatment  1884 Fluid Therapy  1884 Diet  1885 Antimicrobial Drugs  1885 Nonspecific Therapy  1886

from infectious diarrhea amount to more than $23 billion each year in the United States. Our knowledge of infectious diarrheal disease has expanded exponentially in the past two decades. References 3 to 6 are excellent discussions of various aspects of enteric infections.3-6

SUSCEPTIBILITY TO INTESTINAL INFECTION Acquisition of an enteric infection is the result of the interaction of host factors that protect against infection and microbial virulence factors that function to overcome host defenses.

1843

1844

Section X  Small and Large Intestine HOST DEFENSE FACTORS Gastric Acidity

Ingested pathogenic bacteria and other pathogens first must survive passage through the stomach to infect the small or large intestine. In this regard, gastric acidity is the first line of defense.7 Most bacterial pathogens are highly susceptible to low pH, and thus exposure to gastric acid significantly reduces the number of viable bacteria after their ingestion. Gastric juice with a pH of less than 4.0 is rapidly bactericidal, whereas bacteria survive prolonged exposure to gastric juice from patients who are achlorhydric.8 In experimental studies of cholera in healthy adults, clinical infection did not develop when as many as 1010 cholera bacteria were ingested, whereas as few as 104 Vibrio cholerae were able to produce disease when organisms were administered with sodium bicarbonate9; even fewer bacteria were necessary to produce clinical illness when organisms were directly instilled into the duodenum. Naturally occurring cholera also occurs more often in achlorhydric patients.10 The gastric barrier also may be important in preventing other enteric infections such as salmonellosis11 and shigellosis.12

Intestinal Motility

Organisms surviving the milieu of the stomach enter the small intestine where normal propulsive motor activity clears them. Some bacteria, including Clostridium difficile, Clostridium perfringens, and heat-stable enterotoxinproducing Escherichia coli, elaborate toxins that impair intestinal motility.13 In experimental animals it is often necessary to restrict intestinal motility with pharmacologic agents or with ligatures to allow enteropathogens to establish infection.14

Intestinal Microflora

The normal intestinal microflora, primarily in the colon, resist colonization of the intestine by newly introduced bacteria. Products elaborated by the resident microflora, including lactic acid and short-chain fatty acids, are toxic to many bacterial pathogens, and when the intestinal microflora are altered in instances such as administration of an antibiotic, colonization resistance is lessened and the host may be more susceptible to intestinal infection (e.g., C. difficile).15 Alteration of intestinal flora by antibiotics also increases susceptibility to salmonellosis.16,17

Mucus

Mucus, in concert with intestinal motility, provides a physical barrier to bacterial proliferation and mucosal colonization. Gastric mucus can act in conjunction with gastric acidity as the first line of enteric defense. Differences in the carbohydrate composition of intestinal mucus between immature and mature rats suggest that this difference plays a role in the reduced host defense of the immature animal.18

Systemic and Local Immune Mechanisms

The mucosa’s antibacterial immune response is quite complex and important in combating enteric pathogens. Secretory antibody in the intestine appears before serum antibody does in response to intestinal infection with Shigella.19 In cholera, there is a better correlation between the level of coproantibody and immune protection than there is with serum antibody and resistance to enteric infection with this pathogen20; however, both mucosal and systemic immune systems provide important protection against pathogenic bacteria. These immune responses may be directed against multiple targets. For example, the immune response against cholera may be directed against

the toxin or the bacterium and originate from either the mucosal immune system (secretory IgA) or from the serum (IgG). Regardless, both serum and secretory antibodies exert their protective effects at the intestinal level, even though the serum components are produced outside the intestine.

Others

Breast-feeding also serves as a defense mechanism against bacterial enteropathogens. Breast-fed infants are less susceptible to bacterial diarrhea than are formula-fed infants.21,22 Multiple factors are responsible for this protection. Breast milk contains secretory IgA antibodies against specific enteropathogens that survive passage through the infant’s gastrointestinal tract.23 Other components such as lactoferrin, lysozyme, and lactoperoxidase also have anti-infective properties, and breast milk glycolipids can interfere with toxin or microbial adherence.24

BACTERIAL FACTORS

Bacterial pathogens have evolved various virulence factors and mechanisms that enable them to overcome host defenses, including adherence factors, enterotoxin and cytotoxin elaboration, mucosal invasion, and a variety of others.

Adherence

The ability of bacteria to adhere to host mucosal cells is a critical virulence factor in enterotoxin-producing and invasive bacteria as well as in enteroadherent E. coli (EAEC) and enteropathogenic E. coli (EPEC). Bacterial adherence to host mucosal cells may be the predominant virulence factor, as in the case of EPEC; one of two important factors (adherence plus toxin elaboration), as in the case of enterotoxigenic organisms; or only one of several factors required for expression of full pathogenicity, as seen with invasive organisms. Bacteria that cause disease by adhesion alone do not elaborate any of the traditional enterotoxins, but rather they adhere tightly to the mucosa of both the small and large intestine.25 The classic EPEC and the EAEC26 are typical of this group. Other organisms, including enterotoxigenic E. coli (ETEC) and the invasive Salmonella and Shigella species, also must adhere to the intestinal surface to be fully pathogenic. Studies on the mechanism by which EPEC cause diarrhea show that they attach to the intestinal mucosa in a characteristic manner, producing ultrastructural changes known as attaching-effacing lesions27 that lead to elongation and destruction of microvilli.25,28 This pattern of bacterial binding to enterocytes also has been referred to as attaching and effacing adherence27 and the particular morphologic alteration as pedestal formation (Fig. 107-1).25 The laboratory counterpart of mucosal colonization is adherence in tissue culture to various cell lines such as Hep-2 and HeLa. A characteristic form of localized adherence is observed only with classic EPEC serotypes. These events occur in the following three phases29: 1. Nonintimate attachment of EPEC to intestinal epithelial cells: Attachment is mediated by a bundle-forming pilus associated with a large plasmid common to EPEC isolates 2. A signal transduction event that leads to cytoskeletal changes in the enterocyte via activation of protein kinase and the release of intracellular calcium 3. Intimate attachment of the bacterium to the host cell membrane: Attachment is mediated by an outer membrane protein called intimin, which is encoded by the eaeA gene cluster on the EPEC chromosome30 The presence of a plasmid in EPEC serves to increase intimin production; this process is needed for localized

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning

+

Choleragen (mw = 84,000)

B (mw = 58,000)

(mw = 11,600)

Figure 107-1.  Electron micrograph of enteropathogenic Escherichia coli (EPEC) adherent to the jejunal mucosa and demonstrating the attaching and effacing lesion, also called pedestal formation. Tightly adherent E. coli are obliterating the brush border.

adherence to occur.31 EPEC strains with localized adherence produce acute diarrhea when these strains are administered to normal volunteers.32 The role of the eaeA gene as a virulence factor in human EPEC infection has been confirmed in volunteer challenge studies.33 Enterotoxigenic organisms also require expression of bacterial adherence for proliferation of the organisms and colonization, as well as for full expression of toxicity.34 ETEC adhere to the surface of the small bowel epithelium without penetrating the epithelial layer and do so by mechanisms different from those used by EPEC. The most important mechanism by which enterotoxigenic bacteria adhere to the intestinal mucosa is related to specific protein antigens on the surface of the bacterial cell known as pili or fimbriae, also referred to as adherence antigens or colonization factor antigens.35 These pili bind to specific receptor sites on the surface of the intestinal cell via specific ligand-receptor interactions. The antigenic structure of the adherence pili determines the host specificity of the ETEC strains. For example, those bearing a K88 antigen are pathogenic for piglets, whereas others bearing K99 antigen cause disease in calves and lambs. ETEC adhesion antigens for humans include type 1,3,P and BFP pili. Evidence that these colonization factors (e.g., pili and lectins) are important to the pathogenesis of E. coli diarrheal disease in animals is derived from the observations by Moon36 that loss or gain of fimbriae by genetic manipulation results in the loss or gain of the ability to adhere to and colonize the intestine. Adherence not only permits colonization but also can facilitate the delivery of enterotoxin to the epithelium and might even enhance the ability of the organism to elaborate enterotoxin.34,36

Enterotoxin Production

Enterotoxins are polypeptides, secreted by bacteria, that alter intestinal salt and water transport without affecting mucosal morphology.5,37,38 Many organisms elaborate enterotoxins (e.g., V. cholerae, Shigella, ETEC, and Staphylococcus aureus), and several enterotoxins may be elaborated by a single organism. Although most enterotoxins affect the small intestine, the colon also may be a target organ. Several enterotoxin-producing organisms cause food poisoning and induce disease without requiring intestinal colonization, such as S. aureus and C. perfringens; their toxins

S

S

A (mw = 27,000)

A1 (mw = 22,000)

A2 (mw = 5,000)

Figure 107-2.  Subunit structure of cholera toxin (choleragen) (see text). mw, molecular weight. (From Fishman PH. Action of cholera toxin: Events on the cell surface. In Field M, Fordtran JS, Schultz SG, editors. Secretory Diarrhea. Bethesda, Md: American Physiological Society; 1980. p 86.)

are ingested preformed in food, thus accounting for their characteristic brief incubation period. Whether the enterotoxin is ingested preformed or is first expressed within the intestinal lumen, the toxinenterocyte or toxin-colonocyte interaction begins with the binding of the enterotoxin to a specific mucosal receptor. The toxin-receptor interaction increases the concentration of an intracellular mediator, resulting in alteration of salt and water flux. Thus far, three intracellular mediator systems have been shown to be involved in the pathogenesis of enterotoxigenic diarrhea: adenylate cyclase and cyclic adenosine monophosphate (cAMP), the guanylate cyclase and cyclic guanosine monophosphate (cGMP) systems, and intracellular calcium.39,40 Alterations in these mediator systems have similar effects on transport processes to decrease the coupled influx of sodium and chloride and to stimulate the active secretion of chloride from the cell into the intestinal lumen. Other intracellular mediator systems involved in the pathogenesis of bacterial diarrhea include protein kinase C and arachidonic acid metabolites, among others. Cholera toxin (molecular weight ≈84,000 kDa), which stimulates adenylate cyclase, is a prototypical enterotoxin (Fig. 107-2 and Chapter 99), the toxin-enterocyte interaction of which is well understood. Cholera toxin is composed of an A subunit surrounded by five B subunits that bind the toxin to a ganglioside (GM1) receptor on the brush border membrane of the villus epithelial cell. The A subunit, which consists of two parts (A1 and A2) linked by a disulfide bond, slowly penetrates the brush border membrane and is cleaved into its two component peptides. Reduction of this bond releases the active A1 peptide that traverses the cell to the basolateral membrane, where it stimulates the ribosylation of Gs, the stimulatory subunit of a heterotrimeric G protein. This action results in the irreversible activation of Gs and an increase in cytosolic cAMP. This cAMP in turn activates cAMP-dependent kinases that inhibit NaCl-coupled transport and stimulate chloride secretion. In addition to cholera toxin, other important enterotoxins are those elaborated by E. coli.41,42 Two classes of E. coli enterotoxins are known: heat-labile (LT) and heat-stable (ST) toxins. LT, which exists in two forms (LT-1 and LT-2),42 is a large molecular-weight protein that causes diarrheal disease similar to, but less severe, than, cholera. The subunit structure and mechanism of action of LT-1 and LT-2 also

1845

1846

Section X  Small and Large Intestine are similar to those of cholera toxin; although cholera toxin and LT bind to a glycolipid receptor, specifically GM1, an additional glycoprotein receptor might exist for LT.42 Heatstable enterotoxins (STas) also may be elaborated by E. coli and bind to brush border receptors on enterocytes and colonocytes, a receptor guanylate cyclase, to increase intracellular levels of cGMP. STb is an unrelated enterotoxin elaborated by some E. coli pathogenic for pigs. Other organisms that elaborate highly homologous heat-stable enterotoxins include Yersinia enterocolitica, Citrobacter, and non-O1 vibrios.42

Cytotoxin Production

Cytotoxins are polypeptides that cause cell injury, inflammation, intestinal secretion through inhibition of protein synthesis or via a cascade involving one or more inflam­ matory mediator substances, and cell death. Examples of organisms that produce cytotoxins include C. difficile (see Chapter 108),15 some EPEC, enterohemorrhagic E. coli (EHEC), and Shigella.43 The mechanisms by which cytotoxins cause cell injury, inflammation, and intestinal secretion are numerous and complex and include inhibition of protein synthesis, disruption of cellular actin and tight junction integrity, mitochondrial damage, and adenosine triphosphate depletion among others.

Mucosal Invasion

The mechanism of mucosal invasion involves invasion of enterocytes or colonocytes by the infecting organisms with subsequent intracellular multiplication, resulting in cell injury and possibly cell death. Shigella species are classic examples of invasive enteropathogens. Salmonella species, Campylobacter jejuni, Y. enterocolitica, and some (enteroinvasive) strains of E. coli invade intestinal cells and pass into the lamina propria, where they elicit an inflammatory response and cause mucosal ulceration.4,6 Unlike enterotoxigenic organisms that favor colonization of the small intestine, invasive organisms primarily, but not exclusively, colonize the colon. In salmonellosis, the ileum is colonized in addition to the colon; in shigellosis, the small intestine is colonized transiently early in the course of the disease when watery diarrhea rather than dysentery is the predominant symptom.44 Subsequently, colonization occurs in the colon and bloody diarrhea ensues. In the cases of Shigella and Salmonella species, the ability to invade the gastrointestinal mucosa is of primary importance in establishing the enteric infection.45,46 Bacterial invasion alone is not sufficient to establish disease; other properties of invading organisms also are required. In the case of Shigella species, the organisms must multiply intracellularly. Thus, strains of Shigella flexneri that can invade but cannot multiply do not cause disease when fed to a susceptible host.45 Intracellular multiplication of Shigella organisms also involves lateral spread to adjacent intestinal cells and their death. In the cases of Salmonella species and Y. enterocolitica, however, the organisms penetrate into the lamina propria and can disseminate to extraintestinal sites. As a consequence of mucosal invasion and of the intramucosal multiplication of the organisms, an acute inflammatory reaction develops and mucosal ulceration can occur. Gross ulceration of the colonic mucosa commonly occurs in shigellosis, which accounts for dysenteric stools, but it is much less common with Salmonella and Yersinia infections. Yersinia infection more commonly manifests with microscopic and minute ulcerations involving both the ileum and colon. Some of the mechanisms by which invasive organisms induce intestinal secretion include increased intracellular

calcium and products of inflammation, such as prostaglandin and leukotriene metabolites, serotonin, substance P, interleukin (IL)-1, and reactive oxygen metabolites, among others.

Others

Other bacterial virulence factors, not as yet well defined, can modulate infectivity or the spectrum of illness, including the ability of bacteria to respond to chemotactic signals released by the mucosa, bacterial elaboration of mucolytic enzymes to enable bacterial penetration of the intestinal cell, resistance to phagocytosis, and elaboration of substances that interfere with intestinal motility. The importance of these factors in modifying virulence is uncertain and they are not discussed further.

CLASSIFICATION OF BACTERIAL DIARRHEA Acute bacterial diarrhea can be classified according to the mechanism by which the bacteria cause disease: Toxigenic, in which an enterotoxin is the major, if not exclusive, pathogenic mechanism Cytotoxic, in which a cytotoxin induces acute inflammation and intestinal secretion Invasive, in which the organism penetrates the mucosal surface as the primary event and induces acute inflammation; enterotoxin may be produced as well Adherent, in which the cell cytoskeleton is altered Many organisms elaborate toxins that cause intestinal fluid and electrolyte secretion. As discussed earlier, the recognized diarrheal toxins can be grouped broadly into two categories: enterotoxins, which produce fluid secretion by activation of intracellular enzymes such as adenylate cyclase without any damage to the epithelial surface, and cytotoxins, which cause injury to the mucosal cell and induce fluid secretion. Diarrheal toxins produce two clinically recognizable acute diarrheal syndromes: noninflammatory (enterotoxin-induced) and inflammatory (cytotoxin-induced) diarrheas, both of which are discussed in the following section.

DIAGNOSIS OF INFECTIOUS DIARRHEAL DISEASE EVALUATION OF THE PATIENT

The initial step in the diagnostic evaluation of a patient with acute diarrhea should be a thorough history and physical examination, the goals of which are to identify patients who may be at risk of severe illness or susceptible to complications and those who will benefit from specific therapy. Most patients simply need rehydration therapy. Consideration of the patient’s general health, severity and duration of illness, and the setting in which the illness was acquired should enable the clinician to determine who needs further evaluation (Fig. 107-3). Patients who are debilitated, malnourished, or immunocompromised and those who have severe comorbid illnesses are at increased risk for complications of diarrhea and infection. They can require hospitalization and early diagnostic tests. Other patients who also require a moreaggressive approach include those with systemic signs and evidence of an inflammatory diarrhea, illness lasting more than three to four days, a history or physical examination suggesting a disease process that will benefit from specific

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Fever >103˚ Systemic illness Tenesmus Bloody diarrhea Dehydration No

Yes

No culture or studies Symptomatic therapy: Oral fluids, if necessary No antimicrobial therapy Direct fecal smear

Special circumstances: Outbreak of food poisoning Overseas travel Male homosexual (“gay bowel”) Immunocompromised host Diabetes, HIV/AIDS Hematologic malignancy Immunosuppressive drugs Extremes of age Raw seafood ingestion Antibiotic use

PMNs

No PMNs

Causes: Shigella Campylobacter *E. coli - EIEC - EHEC Yersinia (rare) Salmonella (rare) C. difficile (rare) Aeromonas (rare) Plesiomonas (rare)

Causes: Salmonella Yersinia Aeromonas *E. coli - ETEC - EPEC C. difficile

*Special tests

Consider: Noninfectious causes Oral fluids Presumptive antimicrobial therapy Hospitalization

Parasites

Specific treatment, if needed

Figure 107-3.  Algorithm for the diagnosis and treatment of infectious diarrhea. AIDS, acquired immunodeficiency syndrome; C. difficile, Clostridium difficile; E. coli, Escherichia coli; EHEC, enterohemorrhagic E. coli; EIEC, enteroinvasive E. coli; EPEC, enteropathogenic E. coli; ETEC, enterotoxigenic E. coli; HIV, human immunodeficiency virus; PMN, polymorphonuclear neutrophil.

therapy, and infection with certain specific organisms (Table 107-1).6 Because the number of conditions that cause acute and chronic diarrhea is large, a useful method is to classify the diarrheal illness into one of two clinical syndromes: a watery, noninflammatory diarrheal syndrome and an inflammatory diarrheal syndrome (Table 107-2); a subgroup of the latter is the proctitis diarrheal syndrome. Categorization into one of these syndromes limits the number of potential causes and diagnostic tests that need to be considered. This classification can usually be made on clinical grounds and with simple, inexpensive diagnostic tests. In the United States, most of the cases of watery, noninflammatory diarrhea result in an illness that is self-limited and does not require specific therapy. Evaluation of such patients is generally unrewarding and usually unnecessary. By contrast, many patients with acute inflammatory diarrhea are more ill, a specific pathogen often can be diagnosed, and antibiotic therapy may be beneficial.

Noninflammatory Diarrhea

Patients with noninflammatory diarrhea usually present with large-volume watery stools without blood or pus, or

Table 107-1  Pathogens Indicating Need for Antimicrobial Therapy in Patients with Infectious Diarrhea Bacterial Infection Campylobacter diarrhea (if prolonged) Cholera Clostridium difficile Salmonellosis (if patient is pregnant, toxic, or has extraintestinal spread) Shigellosis Traveler’s diarrhea Protozoal Infection Entamoeba histolytica Giardia lamblia Sexually Transmitted Diseases Chlamydia Gonorrhea Herpes simplex virus Syphilis From Park SI, Giannella RA. Approach to the adult patient with acute diarrhea. Gastroenterol Clin North Am 1993;22:483.

1847

1848

Section X  Small and Large Intestine Table 107-2  Characteristics That Help Distinguish Inflammatory from Noninflammatory Diarrhea CHARACTERISTIC

INFLAMMATORY DIARRHEA

NONINFLAMMATORY DIARRHEA

Clinical presentation

Bloody, small-volume diarrhea; lower quadrant cramps; patients may be febrile and toxic Colon Indicated

Large-volume, watery diarrhea; patients may have nausea, vomiting, cramps Small intestine Indicated only if the patient is severely volume depleted or appears toxic Absent Viruses, Vibrio spp., Giardia lamblia, enterotoxigenic E. coli, enterotoxin-producing bacteria, food-borne gastroenteritis

Site of involvement Diagnostic evaluation Fecal leukocytes Causes

Present Shigella spp., Salmonella spp., Entamoeba histolytica, Campylobacter spp., Yersinia spp., invasive Escherichia coli, Clostridium difficile

From Park SI, Giannella RA. Approach to the adult patient with acute diarrhea. Gastroenterol Clin North Am 1993; 22:483-97.

they present with severe abdominal pain. These patients generally have few systemic signs or symptoms, and fever usually is absent. Abdominal cramping, nausea, and vomiting can occur. The most likely causes are viruses (rotavirus, norovirus), enterotoxigenic E. coli, V. cholerae, staphy­ lococcal and clostridial food poisoning, and Giardia and Cryptosporidium infections. Noninflammatory diarrheas generally do not require extensive evaluation. Pathogens causing noninflammatory diarrhea usually infect the small intestine and merely adhere to the mucosal surface without invading the epithelium or causing acute inflammation. Most of these organisms (e.g., cholera, rotavirus) elaborate enterotoxins that stimulate intestinal secretion, occasionally resulting in profound dehydration.

Inflammatory Diarrhea

Patients with inflammatory diarrhea usually present with numerous small-volume stools that may be mucoid, grossly bloody, or both. Such patients may appear toxic and usually are febrile. Abdominal cramping may be severe. Because of the small stool volumes, these patients are less likely to be dehydrated than those with noninflammatory diarrhea. Physical findings might point to a specific diagnosis (Table 107-3). Organisms causing inflammatory diarrheas usually affect the colon and either invade or elaborate cytotoxins, resulting in an acute inflammatory reaction with disruption of the epithelial barrier, mucus, red blood cells, and white blood cells in the stool (Table 107-2). Microbes causing this syndrome include Shigella, Campylobacter, EHEC, C. difficile, Salmonella, Yersinia, and Entamoeba histolytica. Fecal leukocytes (or positive stool lactoferrin test) indicate an acute inflammatory process, and sheets of polymor­ phonuclear leukocytes (PMNs) usually indicate colitis. The acute inflammatory diarrheal syndrome also can have an noninfectious etiology, for example, ulcerative colitis, Crohn’s disease, radiation or ischemic colitis, and diverticulitis. Table 107-4 lists the organisms that may be associated with the presence of fecal leukocytes.47

Proctitis Syndrome

Proctitis syndrome is characterized by frequent painful bowel movements containing blood, pus, and mucus. The sensation of tenesmus, often with rectal pain, usually is prominent. Infectious causes include Shigella, herpes simplex virus type 2, and Neisseria gonorrhoeae, Treponema pallidum (syphilis), and Chlamydia (lymphogranuloma venereum). The importance of this syndrome is that many of the causes have specific treatments. Noninfectious causes include idiopathic ulcerative proctitis, Crohn’s proctitis, radiation proctitis, and solitary rectal ulcer syndrome.

Table 107-3  Clinical Finding(s) That Suggest the Causative Organisms for Some Inflammatory Diarrheas FINDING

CAUSATIVE ORGANISMS

Hemolytic-uremic syndrome

Shigella spp., enterohemorrhagic E. coli Salmonella spp., Shigella spp., Campylobacter spp., Yersinia spp. C. difficile, enterohemorrhagic E. coli Yersinia spp. Yersinia spp.

Reactive arthritis Peritoneal signs Right lower quadrant tenderness Thyroiditis, pericarditis, glomerulonephritis

From Park SI, Giannella RA. Approach to the adult patient with acute diarrhea. Gastroenterol Clin North Am 1993; 22:483.

Table 107-4  Fecal Leukocytes in Intestinal Infections Usually Present Campylobacter spp. EHEC EIEC Shigella spp. Present or Absent Aeromonas spp. Clostridium difficile (antibiotic-associated colitis) EAEC Salmonella spp. Vibrio parahaemolyticus Yersinia spp. Usually Absent Bacillus cereus Clostridium perfringens DAEC Entamoeba histolytica EPEC ETEC Food poisoning Giardia lamblia Staphylococcus aureus Vibrio cholerae Viruses Calicivirus, including norovirus Rotavirus Other viruses DAEC, diffusely adhering Escherichia coli; EAEC, enteroaggregative E. coli; EHEC, enterohemorrhagic E. coli; EIEC, enteroinvasive E. coli; EPEC, enteropathogenic E. coli; ETEC, enterotoxigenic E. coli.

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning LABORATORY DIAGNOSIS

After taking an adequate medical history and performing a physical examination, the clinician should be able to classify the acute diarrheal illness as inflammatory or noninflammatory; to determine its severity and whether complications are likely; and to determine if diagnostic testing is needed and, if so, which tests should be used. In general, most episodes of acute diarrheal illness in the United States are self-limited; diagnostic testing may be kept to a minimum, and treatment is aimed at preventing dehydration. Investigations should be performed only if it is likely the result will influence management and outcome. Diagnostic testing should be reserved for patients with severe illness, including large-volume dehydrating diarrhea, severe abdominal pain, or a prolonged course (i.e., longer than three days); patients with bloody stools, systemic symptoms such as fever higher than 101°F, or prostration; patients with history of recent travel to high-risk areas; and patients at high risk for complications because of extremes of age, diabetes mellitus, and immunocompromise such as acquired immunodeficiency syndrome (AIDS), hematologic malignancy, or use of immunosuppressive medications (see Fig. 107-3). Diagnostic testing for acute infectious diarrhea in the normal host consists of stool cultures for bacterial pathogens including E. coli O157:H7, stool examination for ova and parasites, stool testing for C. difficile toxin, and stool examination for polymorphonuclear neutrophils (PMNs). More-invasive investigations, including flexible sigmoidoscopy with biopsies and upper gastrointestinal endoscopy with duodenal aspirate and biopsies, are reserved for special situations, such as the immunocompromised host, in whom stool examination has not yielded a diagnosis. Investigations should be focused to diagnose specific agents as suggested by the patient’s history and physical examination.

Fecal Leukocytes

A particularly useful technique to focus the differential diagnosis is microscopic examination of the stool for PMNs (see Table 107-4). Invasive pathogens that primarily affect the colon, such as Shigella and Campylobacter, produce a “sea of polys,” as well as red blood cells. The toxigenic organisms, viruses, and food-poisoning bacteria cause a watery stool that harbors few formed elements. Stool tests for lactoferrin or calprotectin (proteins made by PMNs) in fecal specimens are available and provide a rapid and sensitive alternative to microscopy for identifying PMNs and, by inference, inflammatory diarrhea.48 Salmonella, Yersinia, Vibrio parahaemolyticus, and C. difficile diarrhea have unpredictable and variable numbers of fecal leukocytes, depending on the degree of colonic involvement; most cases show only occasional PMNs. An acute exacerbation of ulcerative or Crohn’s colitis can produce a stool with many PMNs, thereby resembling bacillary dysentery. Although fecal microscopic examination is neither infallible nor even helpful in all cases, it is inexpensive and yields immediate information that can guide antibiotic therapy. Although the fecal lactoferrin and calprotectin tests are slightly more expensive than microscopy for fecal leukocytes, they have several advantages, including ease, rapidity, and the absence of a requirement for a fresh stool specimen.

Stool Cultures

Stool cultures are ordered too frequently. In most microbiology laboratories, routine stool cultures are processed for Shigella, Salmonella, and Campylobacter. Other enteric

pathogens such as Yersinia, Vibrio, and E. coli O157:H7 are not sought routinely. Therefore, if clinical suspicion is high, the microbiology department needs to be notified to search for these pathogens. Because of sporadic shedding of pathogens (nontyphoidal Salmonella spp., Salmonella typhi) and because most episodes of acute diarrhea are caused by viruses, undetectable pathogens, or noninfectious causes, stool cultures are not usually positive. At Massachusetts General Hospital, the isolation rate of bacterial pathogens from 2000 fecal cultures in 1980 was 2.4%.49 In patients with severe diarrhea requiring hospitalization, the bacterial isolation rate from feces is somewhat higher, ranging from 27% to 43%,50,51 and up to 58% in a study that used moreadvanced techniques.52 Even in patients hospitalized for dysentery, the rate of positivity for microbiologic diagnosis is only 40% to 60%. In community patients with severe acute gastroenteritis (more than four fluid stools per day, lasting at least three days and with at least one associated symptom), the yield of a stool culture and ova and parasite examination increased to 88%.53 In outbreaks of gastroenteritis in the United States, only one half of the cases have a confirmed etiology, of which two thirds are bacterial in origin. These figures suggest that many cases of acute diarrhea are caused by unidentified pathogens. When parasitic or protozoal infection is suspected, stool examination for cysts, trophozoites, larvae, or eggs should be performed. When either Giardia or cryptosporidia is suspected, a stool enzyme-linked immunosorbent assay (ELISA) for antigens of these organisms should be requested. ELISA also is done for rotaviruses or noroviruses, but this test rarely is required in adults. It is not appropriate to request stool cultures and stool examination for ova and parasites for every patient with diarrhea. These tests should be performed only when the clinical suspicion is high that parasitic infection is the cause of diarrhea and the results of the stool tests will change management and influence outcome.

Endoscopy

Endoscopy may be useful to obtain aspirates and biopsies of the small intestine to detect Giardia, cryptosporidia, microsporidia, Isospora belli, or Mycobacterium aviumintracellulare. Flexible sigmoidoscopy can be useful in evaluating patients with proctitis, tenesmus, or sexually transmitted diseases or in identifying the pseudomembranes of C. difficile. In HIV-infected patients, colonoscopy with multiple biopsies is used to detect cytomegalovirus ulcers, which may be confirmed by the presence of inclusion bodies on biopsy. An algorithm for the diagnosis of acute diarrhea is presented to help determine which patients should be treated symptomatically and which require further diagnostic studies and treatment (see Fig. 107-3). Approximately 90% of cases of acute diarrhea fall into the “No” category.

TOXIGENIC PATHOGENS The prototypical organisms in this group are V. cholerae and ETEC, both of which elaborate enterotoxins that cause dehydrating diarrhea. The salient characteristic of diarrheal disease caused by V. cholerae and ETEC is that the entire disease results from intestinal fluid loss, which is related to the action of the enterotoxin on the small intestinal epithelial cells. These organisms do not invade the mucosal surface; rather, they colonize the upper small intestine, stick to the epithelial cells, and elaborate an enterotoxin. Mucosal

1849

1850

Section X  Small and Large Intestine architecture remains intact, with no evidence of cellular destruction. The fecal effluent is watery and often voluminous, producing clinical features of dehydration. The origin of the fluid is primarily the upper small intestine, where the enterotoxin has its greatest activity. Bacteremia is not a complication of toxigenic diarrhea.

VIBRIO CHOLERAE

Cholera is a severe diarrheal disorder that can cause dehydration and death within three to four hours of onset. Stool output can exceed 1 L/hr, with daily fecal outputs of 15 to 20 L if parenteral fluid replacement keeps up with losses. The acutely ill patient typically has marked signs of dehydration, poor skin turgor, “washerwoman’s” hands, absent pulses, reduced renal function, and hypovolemic shock. Cholera is the prototypical toxigenic diarrhea. More has been learned about pathophysiology—and normal intestinal function—from cholera than from any other intestinal disease. Treatment programs have been devised, including an oral rehydration regimen used to treat all major diarrheal illnesses; the cholera enterotoxin has been purified; the immunology and epidemiology of cholera have been clarified; the secretory mechanisms and second messenger systems involved have been elucidated; and anticholera vaccines have been developed.

Microbiology

Vibrio cholerae is a Gram-negative, short, curved rod that looks like a comma. It is actively motile by means of a single polar flagellum. Vibrios are strongly aerobic and prefer alkaline and high-salt environments. The terminology and classification of V. cholerae are complex. Agglutination by antisera against the O1 antigen (cell wall polysaccharide) is used to characterize vibrios into O1 or non-O1 groups. They are then classified into serotypes based on the subspecificity of O1 antigen (A, B, C: Ogawa [A, B]; Inaba [A, C]; Hikojima [A, B, C]). Biotype refers to different phenotypic qualities (e.g., production of hemolysins, agglutination of various species of erythrocytes, resistance to polymyxin). Toxigenic V. cholerae that agglutinates in O1 antiserum is the main cause of epidemic cholera. There are two major biotypes of V. cholerae O1: classic and El Tor. The latter strain is responsible for the current pandemic that began in 1961. El Tor vibrios are somewhat hardier than others in nature. Clinical disease is similar with both biotypes, although on average, El Tor infections are milder. The major serotypes associated with clinical disease are Inaba and Ogawa; a rare third type is Hikojima. The El Tor Inaba type was responsible for the 1991 outbreak in South America. There also are unique O1 cholera strains (e.g., V. parahaemolyticus and V. vulnificus) that cause endemic disease along the Gulf Coast of the United States.54 A newly described toxigenic non-O1 strain, now designated V. cholerae O139 Bengal, was responsible for an epidemic that started in southern India and Bangladesh in late 1992 and spread rapidly to many countries in Southeast Asia.55,56 This strain was classified as a new serogroup because it did not react with antisera to the previously identified 138 serogroups.56

Cholera Toxin

All wild strains of V. cholerae, including O139, elaborate the same enterotoxin, a protein molecule with a molecular weight of 84,000 kDa (see Fig. 107-2).57 The structural genes for the cholera toxin are encoded by a filamentous bacteriophage.58 Like the diphtheria toxin, the cholera toxin is composed of two types of subunits. Each toxin molecule

contains five B subunits that encircle a single A subunit. The B subunit is responsible for binding to the receptor on the mucosa. The A subunit is responsible for activation of adenylate cyclase located on the basolateral cellular membrane (see earlier). A second 10 to 30 kDa LT, zonula occludens toxin, has been described that alters intestinal permeability by acting on intestinal epithelial cell tight junctions.59

Epidemiology

For many centuries, the Bay of Bengal had been considered the “cradle of cholera.” Western countries were relatively free of cholera epidemics until the 19th century, but since then, with the worldwide spread of the disease, six pandemics (across continents) have been reported. We are currently in the seventh pandemic, which started in 1961 in Indonesia and then made its way to the Philippines, Hong Kong, Japan, Korea, Thailand, India, Pakistan, and the Middle East, passing across the African continent to engulf the entire region, and, in 1991, spreading to South America. Although the overall number of cases of cholera in Latin America has subsided since 1991, outbreaks of V. cholerae have continued to occur sporadically throughout subSaharan Africa. During 1999, more than 200,000 cases of cholera were reported from Africa, accounting for 81% of the global total of cholera cases.60 Cholera occurs sporadically along the Gulf Coast of the United States, primarily in Texas and Louisiana.61 Among the millions of American travelers to endemic areas in foreign countries, only 41 imported cases of cholera were reported in the United States from 1961 to 1990, and none was associated with secondary spread. The epidemic in South America resulted in 151 cases of cholera in the United States: 26 cases in 1991, 103 in 1992, and 22 in 1993; only one death was reported.62 The organism associated with the current pandemic is an El Tor biotype which generally causes a milder disease than that seen with the classic strains and has a higher frequency of inapparent infection. The South American epidemic that began in Peru in January 1991 reached more than one million cases in its first three years. From 15,000 to 20,000 cases of cholera were reported each week during the peak of the epidemic, for a national incidence of 1 : 1000 persons. Unboiled drinking water, unwashed fruits or vegetables, and food or water from street vendors were implicated risk factors in this explosive outbreak.63,64 The epidemic of V. cholerae O139 Bengal that began in southern India and Bangladesh in late 1992 affected adults predominantly.56 The clinical features of infection with the O139 Bengal strain were virtually indistinguishable from infection caused by V. cholerae O1.65,66 Contaminated water and food are the major vehicles for the spread of cholera. Infection by person-to-person contact is uncommon, and rarely do physicians, nurses, ward attendants, and laboratory workers who come in contact with the microorganism acquire clinical disease. The inoculum required to cause acute cholera is large, approximately 109 organisms. Even this number usually does not cause disease in a healthy person without some agent, such as a proton pump inhibitor or hydrogen pump blocker, to buffer gastric acidity. People with low gastric acidity, often associated with malnutrition, are more easily infected than those with normal acidity. Humans are the only host for cholera vibrios. The carrier rate is approximately 5% after acute exposure, although long-term carriers are much less common. Cholera vibrios are harbored in the gallbladder, like S. typhi.

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Pathogenesis

The clinical syndrome of cholera is caused by the action of the toxin on intestinal epithelial cells. Cholera toxin increases adenylate cyclase activity to result in elevated levels of cAMP in the intestinal mucosa, and this in turn causes intestinal secretion. Fluid loss in cholera originates in the small intestine. The most sensitive areas are the upper intestine, particularly the duodenum and upper jejunum; the ileum is less affected, and the colon usually is in a state of absorption and is relatively insensitive to the toxin. Diarrhea results because the large volume of fluid produced in the upper intestine overwhelms the absorptive capacity of the colon. Attachment of V. cholerae to the intestinal mucosa is mediated by various surface components, including a fimbrial colonization factor known as toxin-coregulated pilus. The toxin-coregulated pilus attachment protein might play an important role in producing naturally occurring protective antibodies against V. cholerae.67 Despite the derivation of the term cholera (Greek: chole, bile), the appearance of choleric stools resembles rice water; that is, the stool has lost all pigment and becomes a clear fluid with small flecks of mucus. The electrolyte composition (Table 107-5) is isotonic with plasma, and the effluent has a low protein concentration. On microscopic examination there are no inflammatory cells, only small numbers of shed mucosal cells. Cholera vibrios do not invade the mucosal surface, and bacteremia is virtually unknown in this disease. A biopsy specimen of the mucosa during acute cholera shows evidence of dehydration, with maintenance of normal architecture, in sharp contrast to the invasive and ulcerating lesions associated with Salmonella and Shigella.

Clinical Features

Like many other infectious diseases, there is a spectrum of clinical manifestations with V. cholerae, from an asymptomatic carrier state to a desperately ill patient with severe dehydration. The initial stage is characterized by vomiting and abdominal distention and is followed rapidly by diarrhea that accelerates over the next few hours to frequent large volumes of rice-water stools. All the clinical symptoms and signs can be ascribed to fluid and electrolyte losses. Patients present with profound dehydration and hypovolemic shock, usually leading to kidney failure. The stool is isotonic with plasma, although there is an inordinate loss of potassium and bicarbonate, with resultant hypo-

kalemic acidosis (see Table 107-5). Mild fever may be present, but there are no signs of sepsis.

Immunologic Responses

After recovery from acute cholera, two serum antibodies can be demonstrated: a vibriocidal antibody directed against somatic antigen and an antitoxin antibody against the enterotoxin. Vibriocidal titers rise and fall rapidly after infection, and by six months only 1% of patients have high serum levels. In areas of high prevalence, such as the Indian subcontinent, the level of vibriocidal titer rises with age; by the 10th year of life, 50% of people have measurable titers. Protection is related to the presence and actual level of vibriocidal antibody. From these observations, it follows that acute cholera in endemic areas is a disease largely of young children, primarily those who lack vibriocidal antibody. Antitoxin titers rise somewhat slowly after acute infection and remain elevated for many months. The susceptibility of adults in areas endemic for the O139 Bengal strain of cholera indicates that the afflicted populations are immunologically naive and that prior exposure to V. cholerae O1 does not provide cross-protective immunity. Nevertheless, volunteer challenge studies indicate that an initial infection with O139 Bengal provides protection against recurrent disease.65 Elevation in titers of vibriocidal antibody may be caused either by actual infection with vibrios or by asymptomatic carriage of vibrios. In field situations, the clinical case rate is approximately 0.26%; that is, for every clinical case of cholera there are approximately 400 asymptomatic people who have had contact with the organism, as demonstrated by an elevation in vibriocidal antibody titers.

Treatment

Treatment of acute cholera is based on the physiologic principles of restoring fluid and electrolyte balance and maintaining intravascular volume. These objectives can be accomplished with intravenous solutions or oral fluids that contain electrolytes in isotonic concentrations (see Table 107-5). Particular attention is paid to administration of bicarbonate and potassium, which are lost excessively in choleric stool. Various oral rehydration solutions (ORSs) have been developed for treating mild to moderate cases; ORS is especially useful in developing countries (Table 107-6).68 The simple therapeutic principles of fluid replacement and antibiotics can save many lives. This knowledge has

Table 107-5  Electrolyte Concentrations of Choleric and Nonspecific Fecal Fluid and of Intravenous Fluids Used to Treat Infectious Diarrheas ELECTROLYTE CONCENTRATIONS (mmol/L) TYPE OF FLUID Cholera Stool   Adult   Child Nonspecific Diarrhea (Child) Intravenous Therapy Lactated Ringer’s solution 5 : 4 : 1 solution†‡ 2 : 1 solution§

Sodium

Potassium

Chloride

Bicarbonate

124 101 56

16 27 25

90 92 55

48 32 14

130 129 141

4 11 —

109 97 94

28* 44 47

*Equivalent concentration after lactate conversion. † Add glucose, 110 mmol/L (20 g/L). ‡ Intravenous solution that is 5 g of sodium chloride, 4 g of sodium bicarbonate, and 1 g of potassium chloride per liter. § Solution that has a carbohydrate-to-sodium ratio of 2 : 1.

1851

1852

Section X  Small and Large Intestine Table 107-6  Compositions of Some Oral Hydration Beverages BEVERAGE Rehydration   WHO solution   Rehydralyte* Maintenance   Infalyte*   Lytren*   Pedialyte*   Resol*   Ricelyte* Other Liquids   Apple juice   Chicken broth   Cola   Ginger ale   Tea

SODIUM (mmol/L)

POTASSIUM (mmol/L)

CHLORIDE (mmol/L)

BASE (mmol/L)

CARBOHYDRATE (mmol/L)

OSMOLARITY (mOsm/L)

90 75

20 20

80 65

10 (C) 10 (C)

111 140

310 305

50 50 45 50 50

20 25 20 20 25

40 45 35 50 45

10 (B) 10 (C) 10 (C) 11 (C) 11 (C)

111 111 140 111 30 (D)

270 290 250 270 200

30 250 2 2 0

0 0 13 (B) 4 (B) 0

690† 0 730† 500† 0

730 450 750 540 5

3 250 2 3 0

28 8 0.1 1 0

*Ready to use. † Combination of glucose and fructose. B, bicarbonate; C, citrate; D, rice-syrup solids (g/L); WHO, World Health Organization. Modified from Avery ME, Snyder JD. Oral therapy for acute diarrhea: The underused simple solution. N Engl J Med 1990; 323:891.

been available only since approximately 1970; before then, the mortality rate for cholera was 50% to 75%. Application of these physiologic principles reduces the mortality rate in adults to less than 1%. Indeed, the mortality rate in the Peruvian epidemic was less than 1%. Children with cholera still have a mortality rate of 3% to 5% because of a lack of fluid reserve in young children. Antimicrobial agents are useful ancillary measures to treat cholera because their use leads to reductions in stool output, duration of diarrhea, fluid requirements, and Vibrio excretion.69 Tetracycline is recommended at a dose of 40 mg/kg per day orally up to a maximum of 4 g/day in four divided doses for two days; there is no proven value in lengthening the duration of treatment to four days. Singledose therapy with ciprofloxacin results in a successful clinical response in 94% of patients infected with V. cholerae.70,71 As a result of rising rates of resistance, tetracycline and doxycycline often are less effective than the fluoro­ quinolones.71,72 Alternative drugs include trimethoprimsulfamethoxazole (TMP-SMX) and furazolidone.

Vaccines

Currently, no vaccines for the treatment of cholera are available in the United States. No cholera vaccination requirements exist for entry or exit in any country. Cholera vaccination is not recommended.

OTHER VIBRIO SPECIES

In addition to the cholera vibrios, at least nine other vibrios have important pathogenic significance.61,73,74 These strains represent a diverse group of organisms that are mor­ phologically and biochemically identical to V. cholerae but that do not agglutinate with the O group antiserum of the three cholera serotypes.74,75 The non-O1 cholera vibrios produce several toxins and cause a wider range of infection than do the cholera vibrios, including watery diarrhea, dysentery, wound infections, ear infections, and septicemia.61,74 The non-O1 cholera vibrios can be isolated from salty coastal waters of the United States, most commonly in the summer and fall when the temperature rises. Mollusks, particularly oysters, have a reported contamination rate of 10%

to 15% and are the major source of non-O1 Vibrio disease; clams, mussels, and crabs also have been implicated. Strains within the same species can produce different enterotoxins, cytotoxins, and hemolysins. The diversity of toxin production is matched by the diversity of clinical symptoms: diarrhea ranges from watery dehydrating diarrhea to frank dysentery; some strains penetrate the intestinal mucosa and produce bacteremia; others have been incriminated in wound infections after exposure to ocean water or handling raw seafood.76 In East Asia, non-O1 cholera vibrios have been associated mainly with severe dehydrating diarrhea. In Peru, serogroups O10 and O12 were isolated from patients with liquid diarrhea associated with mild to moderate dehydration.77 In the United States, reported cases of disease caused by non-O1 cholera vibrios include wound and ear infections, septicemia, and infections of the lung and biliary tract.74 The most common antecedent history is consumption of raw oysters within the previous 72 hours. In outbreaks, there is a high attack rate, with incubation periods that range from as short as six to 12 hours to as long as three days. A one-week course of diarrheal illness is common. Because the gastrointestinal disease is self-limited and relatively benign in the United States, antibiotics are not recommended; however, septicemia, wound infections, and deep organ infections should be treated with appropriate antibiotics. The incidence of Vibrio intestinal infections was studied among participants at an antimicrobial conference in New Orleans, many of whom had consumed raw oysters. Of 479 persons surveyed, 11% had a positive stool culture for vibrios, mainly V. parahaemolyticus, and approximately one third of those with a positive culture had diarrhea. Samples of local seafood, especially oysters, were found to harbor five different species of vibrios.78 In the Chesapeake Bay area, the annual incidence of Vibrio infections related to consuming seafood is estimated to be 1.6 per 100,000 persons.79

Vibrio parahaemolyticus

V. parahaemolyticus causes an acute diarrheal disease after consumption of contaminated raw fish or shellfish. Recog-

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning nized as an important pathogen in Asia, V. parahaemolyticus also has been isolated in the United States, although the exact prevalence is unknown. Strains of V. parahaemolyticus produce a number of distinct hemolysins, the most significant of which appears to be responsible for the Kanagawa phenomenon, which causes hemolysis of human red blood cells in Wagatsuma bacteriologic medium. Kanagawapositive isolates are pathogenic for humans, whereas Kanagawa-negative strains are nonpathogenic and are isolated from marine sources as part of their flora. Pathogenic strains of V. parahaemolyticus produce a number of other toxins, including a lethal toxin that also is hemolytic. In some studies, these organisms produce an enterotoxin that causes fluid accumulation in the rabbit ileal loop model and a cytotoxic toxin that causes damage to HeLa cells. Some strains have the ability to invade the intestinal mucosa and cause bacteremia in experimental animals.80 Epidemiology Many outbreaks of V. parahaemolyticus gastroenteritis have been reported in Japan. During the warm months, when the frequency of this infection is higher, V. parahaemolyticus is responsible for most episodes of bacterial food poisoning in Japan. Infections also have been documented in other countries in Asia, as well as Australia and Great Britain. In the United States, there is a striking geographic association, with most cases occurring in coastal states such as Maryland, Massachusetts, Louisiana, New Jersey, and Washington. The organism is ubiquitous in marine waters and can be found along the coastlines of most countries in which cases have been reported. The attack rate in epidemics varies from 24% to 86% of exposed persons. The mean incubation period for most outbreaks has been 13 to 23 hours, with a range of four to 48 hours. Most infections have been associated with sea fish or seawater ingestion. Occasionally, boiled sardines, salted vegetables (contaminated from salt water), or crabs, shrimp, and oysters (cooked or uncooked) have been incriminated. The common factor in most outbreaks appears to be a hiatus of several hours without proper refrigeration between catching the sea fish or mollusk and eating it. Clinical Features The diversity in toxins and virulence mechanisms is reflected in the variation in symptoms and signs observed in outbreaks in the United States.81,82 Explosive watery diarrhea is the cardinal manifestation in more than 90% of the cases. Abdominal cramps, nausea, vomiting, and headaches are common. Fever and chills occur in 25% to 50% of cases. Clinically, this illness resembles that produced by nontyphoidal Salmonella; however, in some cases a bloody dysenteric syndrome is observed, with fecal leukocytes and superficial mucosal ulcerations seen on sigmoidoscopic examination. The duration of illness generally is short, with a median of six days (range, less than one to 30 days). Fatalities are rare and usually occur in persons with preexisting medical conditions. The diarrhea of V. parahaemolyticus usually is not as profuse as with V. cholerae, but hypotension and shock can occur. Subclinical cases occur in less than 1% of healthy persons. Infection is rare in the winter, suggesting that the carrier state probably is transient. The organism is no longer detectable in the stool once symptoms have resolved. Treatment Although explosive in onset, this disease usually is rather short lived. Patients generally are treated symptomatically.

The organism is sensitive to several antibiotics, including tetracycline, but there is no evidence that antimicrobial therapy has a role in management.

Additional Vibrio Species

V. vulnificus is perhaps the most important noncholera Vibrio species in the United States because of its severity of illness, especially in patients with underlying liver disease. V. vulnificus can be acquired as a wound infection in people swimming in salt water or by direct consumption of seafood, usually raw oysters; the mortality rate of resulting septicemia is 50%. Because this infection can be fatal in patients with underlying liver disease, such persons should be warned to avoid eating raw seafood, especially oysters.83 Vibrio mimicus acquires its name from its similarity to cholera vibrios, even in producing an enterotoxin that resembles cholera toxin.84 The organism has been isolated from patients in the United States with diarrhea, septicemia, or wound infections.73 Vibrio hollisae, also known as enteric EF-13, is a rare isolate from stool and, occasionally, blood cultures. Vibrio furnissii is found in Asia, but its most celebrated outbreak was on an air flight from Tokyo to Seattle, during which 23 passengers developed severe diarrhea, resulting in one death and two hospitalizations. Vibrio fluvialis, previously designated as enteric group EF-6, has been isolated from patients with severe watery diarrhea in Asia and the coastal United States.73,74,79 The isolates produced a range of toxins, including an enterotoxin similar to classic cholera toxin. The organism is found only rarely in other parts of the world, including the United States. Bacteremia caused by Vibrio metschnikovii has been described in a limited number of cases and may be more common in patients with an underlying disease.85 Vibrio alginolyticus is a rare cause of wound or ear infections and gastroenteritis.73 Vibrio damsela is encountered rarely in wound infections.

AEROMONAS SPECIES

Aeromonas species are ubiquitous environmental organisms found principally in fresh and brackish water, especially in the summer months. These Gram-negative organisms often are mistaken for coliforms in the laboratory and, as a result, reported incidence rates are falsely low. Aeromonas species are divided into two groups: psychrophilic (Greek: psychros, cold) aeromonads, which grow optimally at temperatures ranging from 22°C to 28°C, and mesophilic aeromonads, which grow best between 35°C and 37°C.43 Psychrophilic strains usually are isolated from environmental water sources and fish; Aeromonas salmonicida is the most common strain in this group. Based on their phenotypic features, the mesophilic aeromonads are grouped into three complexes: Aeromonas hydrophila, Aeromonas caviae, and Aeromonas veronii. All three of these Aeromonas species have been associated with human infection.86,87 Aeromonas strains produce an array of toxins, including heat-labile enterotoxin, hemolysin, and cytotoxin.88

Epidemiology

Aeromonas infections often are associated with drinking untreated water, such as well water or spring water.89 Several studies have reported a high frequency of isolation of the organism from the stools of children with diarrhea; for example, the incidence of Aeromonas isolations in Western Australia was 10.2% in more than 1000 cases of childhood diarrhea, compared with 0.06% in control subjects.90 Other studies have found a high carrier rate in healthy people, with a range of 0.7% to 3.2% and up to 27% in Thailand. The high carrier rate has raised some question about the pathogenicity of Aeromonas.91

1853

1854

Section X  Small and Large Intestine Clinical Features

Aeromonas has long been recognized as a cause of wound infections after swimming in fresh or brackish water and of bacteremic or deep organ infections in immunocompromised hosts. In recent years, however, most isolates have come from intestinal infections. There is a range of illness, from mild diarrhea, seen mostly in children, to more-severe cases that can require hospitalization. In a study from Western Australia, 22% of patients had blood and mucus in their stools, and one third required hospitalization for severe illness.90 Most cases resolved within one week, but 37% of these children had symptoms for two or more weeks. In adults, chronic diarrhea is even more common, lasting an average of 42 days in the United States.89

Treatment

These organisms are consistently resistant to β-lactam antibiotics, such as penicillin, ampicillin, and first- or second-generation cephalosporins.92 In fact, some cases of Aeromonas diarrhea have been activated apparently by prior treatment with ampicillin. Aeromonas species tend to be sensitive to TMP-SMX, third-generation cephalosporins, fluoroquinolones, tetracycline, and chloramphenicol. There is no convincing evidence that mild cases are improved by antibiotic treatment, but the duration of a chronic infection may be shortened by appropriate use of these drugs.

PLESIOMONAS SHIGELLOIDES

Plesiomonas shigelloides also is a member of the family Vibrionaceae but is isolated less often than Aeromonas in the United States.86,91,93 Most cases have been associated with consumption of raw oysters or recent travel to Mexico or Asia.94 Diarrhea ranges from mild and watery to severe colitis with visible blood. Abdominal pain often is pro­ minent. Antibiotic sensitivity is similar to that of Aero­ monas, but little information is available on the efficacy of treatment.

ESCHERICHIA COLI

E. coli are major components of the normal intestinal microflora in humans and animals. Although most strains are relatively harmless in the bowel, others possess virulence factors that are related to diarrheal disease. At least six types of E. coli intestinal pathogens have been recognized (Table 107-7). Their virulence factors include toxin

production, adherence to epithelial cells, and invasiveness, each encoded by specific genetic elements (plasmids or chromosomal genes) that determine pathogenicity.

Enteropathogenic Escherichia coli

Severe epidemics of diarrhea raged in neonatal nurseries for decades, starting in the 1920s; although uncommon in recent years, such outbreaks had a high mortality. Approximately 14 serotypes were associated epidemiologically with neonatal diarrhea, including the well-known types O55, O111, and O119.26 These organisms adhere to the mucosal surface of the small and large intestine and cause dissolution of the glycocalyx and flattening and dissolution of the microvilli.25,28 This results in a form of localized adherence resulting in an attaching and effacement lesion (see Fig. 107-1). (This process is discussed earlier in “Bacterial Factors Involved in Intestinal Infection.”) The mechanisms by which this attachment results in intestinal secretion is not understood clearly, but they include alterations in enterocyte tyrosine kinase activity and intracellular calcium. EPEC adherence factor is contained within a plasmid of the EPEC and has been used to construct a probe that, in turn, has been used to rapidly identify EPEC strains in stools of patients with diarrhea.95 In a study from São Paulo, Brazil, E. coli adherence factor-positive classic EPEC was found in 26% of children with acute diarrhea; these organisms were the most common pathogens isolated from the children, exceeding rotavirus isolations in frequency.96 EPEC strains are less-common causes of diarrhea in industrialized countries, but they seem to be important pathogens in many developing countries, especially in children in their first two years of life.97,98 Resistance to antimicrobial drugs is common in E. coli adherence factor–positive classic EPEC strains.96 Because most of these infections appear to be self-limited, there is no indication for antibiotic treatment, although nonabsorbable antibiotics such as neomycin have been used in the past for neonates with severe EPEC diarrhea.

Enterotoxigenic Escherichia coli

Inspired by the discoveries in cholera, investigators directed their attention to E. coli as a cause of acute toxigenic diarrheal disease. Originally in India, and thereafter in many parts of the world, strains of E. coli were found to elaborate

Table 107-7  Pathogenic Strains of Escherichia coli in the Intestine STRAINS

PATHOGENIC MECHANISMS

PERSONS AFFECTED

CLINICAL FEATURES

DAEC

Diffuse adherence to Hep-2 cells

Children in developing countries

Watery diarrhea (acute) and persistent diarrhea

EAEC

Aggregative adherence to Hep-2 cells

Children in developing countries

Watery diarrhea (acute) and persistent diarrhea

EHEC

Shiga-like toxin (large quantities) O serogroups (usually O157:H7)

Children and adults Persons who ingest contaminated food, especially hamburger (outbreaks)

Bloody diarrhea Hemolytic-uremic syndrome

EIEC

Shiga-like toxin Epithelial cell invasion

Children and adults Persons who ingest contaminated food and water (outbreaks)

Dysentery

EPEC

Attaching and effacing adherence O serogroups

Children Newborns in a nursery (outbreaks)

Watery diarrhea

ETEC

Heat-labile and/or heat-stable toxin Adherence

Children in developing countries; travelers

Watery diarrhea

DAEC, diffusely adhering E. coli; EAEC, enteroaggregative E. coli; EHEC, enterohemorrhagic E. coli; EIEC, enteroinvasive E. coli; EPEC, enteropathogenic E. coli; ETEC, enterotoxigenic E. coli; RBC, red blood cell; WBC, white blood cell.

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning an enterotoxin similar to that of V. cholerae.99 ETEC is a group of E. coli distinct from EPEC serotypes. ETEC infections mostly are sporadic but can cause large outbreaks. Mechanism of Infection ETEC is acquired by consuming contaminated foods and liquids. The organisms first adhere to the surface of small intestinal enterocytes by mechanisms discussed earlier in the section “Bacterial Factors Involved in Intestinal Infection.” They colonize the surface of the small bowel epithelium without penetrating the epithelial layer. As in cholera, there is neither mucosal damage nor bacteremia. Although a variety of specific pili are involved in adhesion of specific ETEC strains, all ETEC strains, whether pathogenic for humans or animals, elaborate similar enterotoxins. Two types of enterotoxins are produced by ETEC.29 The LT is a protein that is destroyed by heat and acid and has a molecular weight of approximately 80,000 kDa. It acts pathophysiologically like cholera toxin by activating adenylate cyclase, thereby causing secretion of fluid and electrolytes into the small intestinal lumen. LT also shares antigenic components with cholera toxin. The second toxin is heat stable (ST) and is able to withstand heating to 100°C. This toxin has a low molecular weight of approximately 2000 kDa and activates guanylate cyclase; the resultant increase in cGMP induces intestinal secretion from both the small and large intestine. ST is really a family of toxins; the forms that cause disease most commonly in humans are ST1a and ST1b (which differ from each other by a few amino acid residues). ETEC strains elaborate LT only, ST only, or both LT and ST. These toxins cause diarrhea in humans, and similar types of toxigenic E. coli also cause dehydrating diarrhea in domestic animals, including pigs, cows, and sheep. Epidemiology ETEC infections are acquired from other humans; animal strains of ETEC are host-specific. The major vehicles of infection appear to be contaminated foods and beverages. Infection occurs primarily in children, and the highest incidence is in the tropics. There have been varying reports of ETEC infection in the United States, with high incidences in Chicago and Dallas and low figures in other American cities and Canada. In developing countries, the frequency of ETEC infection in children has varied from 15% to 50% of all diarrheal episodes. ETEC is the most common cause of diarrhea in travelers from North America and Northern Europe to areas of the developing world where diarrheal disease is prevalent. ETEC also has become the leading bacterial etiology of gastroenteritis outbreaks on cruise ships; water stored at overseas ports is the probable source of these ETEC infections.100 Clinical Features ETEC infections are among the most common causes of diarrhea in children living in developing countries and travelers to these regions.97 There is nothing distinctive about the clinical presentation of ETEC diarrhea. The incubation period is 24 to 48 hours, and the disease often begins with upper intestinal distress, followed shortly by watery diarrhea. The infection can be mild, with only a few loose bowel movements, or quite severe, mimicking cholera, with severe dehydration and even rice-water stools. Indeed, the initial demonstration of such toxigenic diarrhea came from studies in Calcutta of a serious form of diarrheal disease called acute undifferentiated diarrhea. Affected patients were admitted to the cholera ward until it was determined that vibrios were not present in their stools. ST-only strains cause a milder attack of diarrhea than do LT-producing

strains, but ST-affected patients have more vomiting and constitutional complaints.101 Immunity Antibodies to the enterotoxins and colonization factors occur in persons infected with ETEC. It appears that people residing in areas at high risk for ETEC infection acquire some mucosal immunity over time.29 Thus, the risk that ETEC diarrhea would develop in students at a college in Mexico depended on their country of origin; those from South America had a relatively low risk of ETEC diarrhea, whereas those from North America had a high risk.102 Treatment Most patients with ETEC diarrhea have only mild dehydration, but even small amounts of intestinal purging can have serious consequences in children and older people. The stool electrolyte losses in ETEC diarrhea are similar to those in cholera, and fluid replacement should follow the same principles. Although these organisms often are sensitive to many antimicrobial drugs, including TMP-SMX and quinolones, resistant isolates are increasingly encountered.100 Studies of patients with acute traveler’s diarrhea have demonstrated shortening of the duration of diarrhea when effective antimicrobial therapy is initiated early in the course of illness.103 Nevertheless, because most episodes of ETEC diarrhea are self-limited, treatment with antibiotics generally is not necessary.

Enteroinvasive Escherichia coli

Originally described in Asia, enteroinvasive E. coli (EIEC) is recognized as a rare cause of dysentery. During 1971, there was an EIEC outbreak in the United States that was related to contaminated imported cheese.104 Most episodes of EIEC infections are characterized by watery diarrhea; some patients experience a dysenteric syndrome that manifests as bloody mucoid diarrhea, tenesmus, fever, and intestinal cramps, with multiple PMNs in stool. EIECs have been recognized in at least eight E. coli serogroups, most of which are related biochemically and antigenically to Shigella. Other similarities to Shigella include the ability to invade epithelial cells and the production of two toxins, a Shigalike toxin (STX) and an enterotoxin. Diagnosis of EIEC in a routine bacteriologic laboratory is difficult and generally impractical. Surveys of EIEC in the United States have shown low isolation rates, except in a few celebrated outbreaks. Low rates of infection have been observed in some less-developed countries,97 although in Thailand the organism is common in children with diarrhea.105

Enterohemorrhagic Escherichia coli

Acute hemorrhagic colitis, which first was recognized in two separate outbreaks in Michigan and Ohio in 1982,106 has been associated mainly with a specific serotype of E. coli, O157:H7. This organism is estimated to be responsible for 0.6% to 2.4% of all cases of diarrhea and 15% to 36% of cases of hemorrhagic colitis in Canada, the United Kingdom, and the United States.107,108 The spectrum of disease associated with E. coli O157:H7 includes bloody diarrhea, which is seen in up to 95% of patients, nonbloody diarrhea, hemolytic-uremic syndrome (HUS), and thrombotic thrombocytopenic purpura (TTP). Currently, the class of EHEC includes more than 100 different serotypes.108 Epidemiology EHEC has become the most commonly isolated pathogen from the stools of patients with bloody diarrhea in the United States.109 EHEC disease is most common in northern

1855

1856

Section X  Small and Large Intestine climates such as in Massachusetts, Minnesota, and the Pacific Northwest, but it occurs throughout the United States. It also is well known in Canada, Great Britain, and throughout Europe. Infections occur sporadically or in large outbreaks. The leading vehicle of infection is hamburger meat, although outbreaks have been associated with precooked meat patties, roast beef, salami, fresh-pressed apple cider, lettuce, alfalfa sprouts, and unpasteurized milk.29,110-113 Water-borne outbreaks also have been asso­ ciated with contaminated swimming pools and other recreational water bodies, well water, and municipal water systems.29 Person-to-person transmission probably has played a role in outbreaks in daycare centers, and nursing homes.107,114,115 Infection rates vary seasonally, and peak incidence is from June to September. EHEC strains are found in the fecal flora of a wide variety of animals, including cattle, sheep, pigs, goats, chickens, dogs, and cats. Many of these strains are of serotypes other than E. coli O157:H7. The most important reservoir of infection is cattle, hence, transmission via hamburger meat. Virulence Factors EHEC strains possess at least two virulence factors: an adherence mechanism causing attachment-effacement lesions similar to those seen with EPEC (see earlier) and two Shiga-like toxins (STX cytotoxins I and II).107,108,116 The toxins, which are identified either in stool samples or from culture of the organism itself, cause characteristic lesions in tissue culture lines such as Vero cells and HeLa cells. Some EHEC strains produce only STX I or II, whereas others produce both toxins. Most strains of E. coli O157:H7 possess the eaeA gene, which is associated with intimate attachment to the intestinal mucosa, as in EPEC (see earlier). They also produce enterohemolysin and are capable of using both heme and hemoglobin, a property that may enhance their virulence.117 E. coli O157:H7 toxins can result in colitis that resembles ischemic colitis because they can cause endothelial damage, platelet aggregation, and microvascular fibrin thrombi.107,108,118,118a Clinical Features After an incubation period of one to 14 days (mean, three to four days), watery nonbloody diarrhea begins. It is associated with severe abdominal cramping and often progresses to visibly bloody stools. Other symptoms include nausea, vomiting, low-grade fever, and chills. The development of frankly bloody diarrhea often results in admission to the hospital. Examination of the colon by endoscopy demonstrates a segmental colitis (i.e., friable inflamed mucosa with patchy erythema, edema, and superficial ulcerations). The process usually is most evident in the right colon (Fig. 107-4), but virtually any part of the colon may be affected, just as with idiopathic ischemic colitis. Plain films of the abdomen might show subepithelial edema and hemorrhage (thumbprinting), usually in the ascending and transverse colon. Leukocytosis with a shift to the left usually is present, but anemia is uncommon unless infection is complicated by the development of HUS or TTP.107 Microscopic examination of the stool reveals red and white blood cells in low to moderate amounts. The median duration of diarrhea is three to eight days, with longer durations in children and persons with bloody diarrhea.107 A striking association has been noted between intestinal infection with EHEC and HUS. In Minnesota, the incidence of HUS increased progressively during the 1980s to a current rate of 2.0 cases per 100,000 child-years. E. coli O157:H7 was isolated in 46% of children presenting with HUS. Risk

Figure 107-4.  Colonoscopic appearance of the sigmoid colon in a patient with enterohemorrhagic Escherichia coli (EHEC) infection. The mucosa is edematous and violaceous, with diffuse subepithelial hemorrhage.

factors for HUS include age younger than five years, attendance at a large daycare center, presence of bloody diarrhea, and a high white blood cell count.110 A study from the British Isles showed that 95% of the cases of HUS had a prodromal diarrheal illness. The disease was seen most commonly in the summer. Most EHEC strains were O157:H7 or O157:H− (H not able to be typed) and approximately 30% of the isolates belonged to nine other serogroups of E. coli.119,120 HUS is characterized by acute renal failure, microangiopathic hemolytic anemia, and thrombocytopenia. Diagnosis Several laboratory methods are used to diagnose EHEC infections. Because most isolates of E. coli O157:H7 do not ferment d-sorbitol, screening for this pathogen usually is done with sorbitol-MacConkey agar (SMAC). Sorbitolnegative colonies can then be serotyped with commercially available O157:H7 antisera. Such colonies should be sent to a reference laboratory for confirmation. The chances of obtaining a positive culture in stool depend on the time between the onset of symptoms and collection of the stool. Within two days of onset, virtually all stool specimens from EHEC-infected patients are positive for EHEC, whereas after seven days only one third are positive.121 In contrast, other studies have found that the median duration of excretion of EHEC is 17 to 29 days, with some patients shedding the bacterium for as long as 124 days.29,122,123 Testing for Shiga toxin–producing E. coli (STEC) has been widely used but requires special facilities. Newer tests, including DNA probes, polymerase chain reaction (PCR), and enzyme immunoassays, can detect STEC I and II directly in stool specimens but are not in wide use. One report detailed use of peroxidase-labeled antibody directed against whole E. coli O157:H7, and subsequent immunohistochemical staining to identify the organisms on archival paraffin block tissue sections from patients with hemorrhagic colitis and ischemic colitis.118a Positive staining was found in three of 11 cases of ischemic colitis, suggesting an etiologic role for E. coli O157:H7 in some cases of presumably idiopathic ischemic colitis. Treatment The desire to treat EHEC infections is understandable because of the presence of bloody diarrhea and concern that

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning complications such as HUS will develop. Several reports, however, have raised concern that the risk of HUS is increased by antimicrobial therapy. In a murine model, certain antibiotics, notably ciprofloxacin, caused enhanced STX production by E. coli O157:H7 in vitro via the induction of bacteriophage encoded genes; this occurrence was associated with an increased death rate in antibiotictreated mice.124 Antimicrobial therapy in humans does not appear to provide much benefit and might even be harmful. A randomized, controlled trial of TMP-SMX in children with E. coli O157:H7 enteritis found no effect of therapy on the duration of symptoms, pathogen excretion, or incidence of HUS.125 One prospective cohort study identified 71 children with acute E. coli O157:H7 gastroenteritis, of whom only 9 had been treated with antibiotics. However, 5 of the 10 children in whom HUS developed had received either TMP-SMX or a cephalosporin.126 In this study, antibiotic therapy was associated with a significantly increased risk of HUS, but this conclusion has been challenged by others.127,128 Because antibiotic use has not been shown to decrease morbidity resulting from EHEC and might increase the risk of HUS, routine use of antibiotics is not recommended in the treatment of gastroenteritis if E. coli O157:H7 is the known or suspected cause. In cases of confirmed E. coli O157:H7 infection, patients should be followed closely for manifestations of HUS. Thorough cooking of ground beef is an important preventive measure.

Enteroaggregative Escherichia coli

Unlike the attaching and effacing adherence to Hep-2 cells seen with EPEC, some E. coli strains have been observed to adhere in an aggregative pattern, with the bacteria clumping to the cell surface in a stacked-brick pattern.29 Although some investigations have implicated enteroaggregative E. coli (EAEC) as a cause of acute and persistent diarrhea in children in developing countries,29,129 other investigations have failed to find a significant association with diarrhea.96,97,130 Up to one third of children infected with EAEC have grossly bloody diarrhea. EAEC has been associated with diarrhea in patients infected with HIV,131 and EAEC has been shown to be a cause of traveler’s diarrhea.132 Volunteer challenge studies with different strains of EAEC have yielded mixed results, suggesting that certain strains are more virulent than others.133 As yet, there have been no studies documenting the need for or efficacy of treatment of EAEC infections. EAEC include numerous serogroups that largely are distinct from those of EPEC. Certain serotypes such as O44:H18 appear to be more pathogenic than others. There have been no controlled trials of therapy for EAEC infections in children. One study of HIV-positive patients with diarrhea caused by EAEC found a 50% reduction in stool output, fewer intestinal symptoms, and microbiologic eradication of the organism during treatment with ciprofloxacin.134 Similarly, ciprofloxacin therapy of EAEC resulted in a reduction of the duration of diarrhea in patients with traveler’s diarrhea.132

Diffusely Enteroadherent Escherichia coli

Another type of adherent E. coli is diffuse adhering E. coli (DAEC), which adheres to tissue culture cells in a diffuse pattern. The role of these organisms in diarrheal disease is unclear, but they may be a cause of acute or persistent diarrhea in children.132a

INVASIVE PATHOGENS Invasive organisms make their main impact on the host by invading the intestinal epithelium. Whereas toxigenic organisms characteristically involve the upper intestine, invasive pathogens target the lower bowel, particularly the distal ileum and colon. The main histologic finding with invasive pathogens is mucosal ulceration with an acute inflammatory reaction in the lamina propria. The principal pathogens in this group are Salmonella, Shigella, EIEC, Campylobacter, and Yersinia. There are important differences among these organisms, but they all share the property of mucosal invasion as their initiating event in causing disease. The precise mechanism of fluid production in invasive diarrhea is not known, but three theories have been invoked: An enterotoxin may be responsible for producing fluid, at least in the initial phase of the illness. Most Shigella strains elaborate enterotoxin that differs significantly from cholera toxin, yet still causes secretion of intestinal fluid and electrolytes.135 Salmonella, Campylobacter, and Yersinia strains elaborate enterotoxins. Invasive organisms increase local synthesis of prostaglandins and cytokines at the site of the intense inflammatory reaction. In experimental animals, fluid secretion can be blocked by prostaglandin inhibitors such as indomethacin,136 suggesting that prostaglandins are responsible for fluid secretion and subsequent diarrhea. Damage to the epithelial surface can prevent reabsorption of fluids from the lumen. Transudation of fluid from damaged intestine does not appear to be a significant factor137,138; however, colonic malabsorption of fluid, with a constant level of secretion, could be a contributory factor.

SHIGELLA SPECIES

Shigella organisms cause bacillary dysentery, a disease that has been described since early recorded history. The inhabitants of Athens in the second year of the Peloponnesian War were ravaged by dysentery. In the American Civil War, more than 1,700,000 soldiers suffered from dysentery, with 44,500 deaths. World War I also produced a high incidence of dysentery: 371,000 total casualties in France and up to 486,000 casualties in East Africa. Although dysentery is a disease that becomes more prevalent in wartime, there is a constant endemic incidence in tropical countries and in temperate zones.

Microbiology

Shigella species comprise a group of Gram-negative enteric organisms that are included in the Enterobacteriaceae and most closely resemble E. coli. Unlike E. coli they are nonmotile, do not produce gas from glucose, and are generally lactose negative. The four major subgroups are: A: Shigella dysenteriae, 10 serotypes B: Shigella flexneri, 14 serotypes C: Shigella boydii, 18 serotypes D: Shigella sonnei, 1 serotype Group A (S. dysenteriae 1), also known as the Shiga bacillus, produces the most-severe form of dysentery. An outbreak in Central America in the late 1960s and early 1970s caused more than 10,000 deaths, mostly in young children. This organism has caused outbreaks in many developing countries in recent years. By contrast, S. sonnei produces the mildest disease. There have been shifts in the incidence of dysentery and in the prevalence of specific serotypes. In the tropics, dysentery occurs mostly in late summer. In developed coun-

1857

1858

Section X  Small and Large Intestine tries, such as the United States and countries in Europe, the occurrence of dysentery has increased steadily, and the seasonal prevalence has shifted to winter. S. flexneri is the most common serotype in tropical countries, whereas in the United States and Europe, S. sonnei is the most common serotype. In the United States, for example, 60% to 80% of cases of bacillary dysentery are caused by S. sonnei.

Epidemiology

Shigellosis is a major diarrheal disease, causing 10% to 20% of all cases of diarrhea throughout the world. Dysentery occurs mostly in children between the ages of six months and five years, among whom the disease tends to be less severe than in adults; it is rare in infants younger than six months of age. During infection, shigellae are present in large numbers in the feces. The route of infection is oral. The organisms survive best in alkaline conditions and are highly sensitive to heat and drying. Most transmission is person to person and is related to close human contact. There also have been dramatic epidemics related to ingestion of milk, ice cream, other foods, and occasionally water. A high incidence of infection occurs among laboratory workers who come in contact with this organism. Measurements of inoculum size in volunteers reveal that 105 organisms produce an attack rate of 75%,139 but increasing the inoculum size above this number does not increase the attack rate. There is not a good dose-response curve with Shigella (in contrast with Salmonella); indeed, dysentery can be produced with as few as 200 bacteria. The ability of Shigella species to survive in acidic conditions might account for the small inoculum that can produce disease.140 Person-to-person transmission, facilitated by the low infective dose, accounts for rapid spread of Shigella in daycare centers and among people living in conditions of poor hygiene. These factors also explain the high frequency of dysentery among male homosexuals.

Pathogenicity

All strains of Shigella cause dysentery, a term that refers to a diarrheal stool that contains an inflammatory exudate

A

composed of PMNs, mucus, and blood. The exudative character of the stool is a point to be emphasized: This is not mere watery diarrhea but rather a loose bowel movement that contains pus. The inflammatory exudate is related to the main pathologic event: invasion of the colonic epithelium. Humans are the only natural host for the dysentery organism. Experimental infections can be produced in monkeys and guinea pigs, and disease is made worse by starving them, feeding them antibiotics (thus altering the colonic microflora), or administering opium to reduce intestinal motility. The major site of attack of Shigella is the colon, although scattered ulcerations can be seen in the terminal ileum as well. Invasion by Shigella is associated with a constellation of virulence factors that are related to various stages of invasion and lead eventually to death of the intestinal epithelial cell, focal ulcers, and inflammation of the lamina propria. These virulence factors are encoded by both chromosomal and plasmid genes, all of which are needed for the full expression of virulence. All virulent Shigella, as well as EIEC, contain large 120- to 140-megadalton plasmids, which are related to outer membrane proteins. Various loci encode for an invasion plasmid antigen (ipa); invasion factors (inv); and a series of vir proteins that are involved in cell regulation mechanisms within the cell.141 Having penetrated the mucosal surface, the organisms multiply within the epithelial cells and extend the infected area by cell-to-cell transfer of bacilli. Shigellae rarely penetrate beyond the intestinal mucosa and generally do not invade the bloodstream; however, bacteremia can occur in malnourished children and immunocompromised hosts. The initial lesions are confined to the epithelial layer, and the local inflammatory response is severe, consisting of PMNs and macrophages. There is edema, formation of microabscesses, loss of goblet cells, degeneration of normal cellular architecture, and ulceration of mucosa. These events give rise to the characteristic clinical picture of bloody, mucopurulent diarrhea. As the disease progresses, the lamina propria is involved extensively with the inflammatory response. Crypt abscess formation is a nonspecific but prominent feature (Fig. 107-5).

B

Figure 107-5.  Shigellosis. A, Colonoscopic view of this rectum shows luminal narrowing and mucosal inflammation similar to that seen in ulcerative colitis. B, Histopathologic features include a severe inflammatory infiltrate of polymorphonuclear neutrophils and macrophages in the mucosa. Notice that the glands are straight, without architectural distortion or branching, because the process is acute. (A and B from Wilcox CM. Atlas of Clinical Gastrointestinal Endoscopy. Philadelphia: WB Saunders; 1995.)

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Cytotoxins

Initially, only S. dysenteriae 1 was known to elaborate an enterotoxin. This toxin, first identified by Shiga, has been shown to display a variety of biologic effects, depending on the experimental model used, including cytotoxicity, neurotoxicity in mice (seizures), and enterotoxicity (secretion of fluid and electrolytes); the neurotoxic properties can contribute to the seizures seen in some children with shigellosis. Shiga toxin, which is a 75-kDa protein composed of two subunits, inhibits protein synthesis by irreversible inactivation of the 60S ribosomal subunit.142 Inhibition of protein synthesis causes cytotoxicity and cell death. A toxin with similar antigenic and physiologic effects has been found in strains of S. flexneri and S. sonnei.143

Clinical Features

The classic presentation of bacillary dysentery is cramping abdominal pain, rectal burning, and fever, associated with multiple small-volume, bloody, mucoid bowel movements142; this full array of symptoms, however, is not seen in all patients. The most constant findings are lower abdominal pain and diarrhea. Fever is present in approximately

40% of patients, and the typical dysenteric stool, consisting of blood and mucus, is present in only one third. Approximately one third of patients only have diarrhea without dysentery. Many patients demonstrate a biphasic illness. Initial symptoms are fever, abdominal pain, and watery diarrhea without gross blood; this stage may be related to the action of the enterotoxin on the small intestine. The second phase, which starts three to five days after onset, is char­ acterized by tenesmus and small-volume bloody stools, corresponding to invasion of the colonic epithelium and acute colitis. A few patients have a toxic, highly febrile illness associated with more-severe colitis; even in this setting, bacteremia is distinctly uncommon. Malnutrition, especially in young children, and infection with S. dysenteriae 1 are associated with a more-severe course. Among the intestinal complications of shigellosis are intestinal perforation and severe protein loss. An extensive list of extraintestinal complications of various bacterial enterocolitides including bacillary dysentery is presented in Table 107-8.144 Many patients complain of respiratory symptoms, such as cough and coryza, although

Table 107-8  Complications and Manifestations of Bacterial Enterocolitis Campylobacter fetus Carditis Meningoencephalitis Mycotic aortic aneurysm Pneumonia Relapsing fever, chills, myalgias Septic abortion Septic arthritis Thrombophlebitis Campylobacter jejuni Acute cholecystitis Cystitis Guillain-Barré syndrome Pancreatitis Postinfection irritable bowel syndrome Reactive arthritis* Septic abortion Toxic megacolon Clostridium difficile Ileus and toxic megacolon Leukemoid reaction Postinfection irritable bowel syndrome Protein-losing colopathy Enterohemorrhagic Escherichia coli Hemolytic-uremic syndrome Ischemic colitis Thrombotic thrombocytopenic purpura Toxic megacolon Nontyphi Salmonella Bacteremia Chronic carrier state (1%) Cholecystitis, gallstones Extraintestinal infections Arteritis Arthritis Endocarditis Meningitis Osteomyelitis Pneumonia Postinfection colitis Postinfection irritable bowel syndrome Toxic megacolon

Salmonella typhi Bone marrow suppression Chronic carrier state (2%-4%) Cholecystitis gallstones Hepatosplenomegaly Intestinal and colonic ulceration, bleeding, and perforation Myocarditis Relapse of disease in 8%-12%; may be more common in patients taking antibiotics, especially chloramphenicol Rose spots on the upper abdomen Shigella Species Appendicitis Erythema nodosum Hemolytic-uremic syndrome (associated with Shigella dysenteriae) Leukemoid reaction Meningismus and seizures (children) Myocarditis Pneumonitis Postinfection colitis Postinfection irritable bowel syndrome Protein-losing enteropathy Reactive arthritis* Thrombotic thrombocytopenic purpura Toxic megacolon Vibrio vulnificus and Vibrio alginolyticus Cellulitis Otitis media Sepsis and high mortality rate in patients with cirrhosis or who are immunocompromised Yersinia enterocolitica Erythema nodosum (more common in women) Exudative pharyngitis Glomerulonephritis Pancarditis Polyarthritis Reactive arthritis* Sepsis in patients with immunocompromise, cirrhosis, iron overload (especially patients treated with deferoxamine) Thyroiditis Toxic megacolon

*Associated with human leukocyte antigen B27. Modified from Wolf D, Giannella RA. Antibiotic therapy for bacterial enterocolitis: A comprehensive review. Am J Gastroenterol 1993; 88:1667.

1859

1860

Section X  Small and Large Intestine pneumonia is rare. In young children, hypoglycemia can occur, and several neurologic findings can dominate the clinical picture, even before the diarrheal symptoms. Meningismus and seizures can occur with shigellosis, although there is no direct involvement of the central nervous system.145,146 The meningismus and seizures have been related to the high fever, but they also can occur when the fever is not extraordinarily high. During the acute phase of disease, HUS may occur.147 Thrombocytopenia and a severe leukemoid reaction also have been reported.148 A rash (rose spots) can occur during the acute phase of shigellosis. After an acute attack of dysentery, usually two to three weeks after onset, arthritis can appear. Joint pain or effusion usually is asymmetrical and involves large joints. Joint complaints are present by themselves, not necessarily with other signs of Reiter’s syndrome, and usually occur in patients who are positive for human leukocyte antigen (HLA)-B27; autoantibodies to this antigen cross-react with Shigella proteins, thereby resulting in circulating antibody-antigen complexes.149 The course of shigellosis varies. Children tend to have mild infections lasting no more than one to three days. The average length of symptoms in adults is approximately seven days. In more-severe cases, symptoms can persist for three to four weeks and often are associated with relapses. Untreated bacillary dysentery, particularly when the course is prolonged, can be confused with ulcerative colitis. Chronic carriers of Shigella have been identified; they can pass this organism in their feces for a year or more. Such carriers are distinctly uncommon, and usually they stop shedding the organism spontaneously. Carriers of Shigella are prone to intermittent attacks of the disease, in contrast to Salmonella carriers, who rarely become reinfected with the strain they carry.

Diagnosis

The diagnosis of shigellosis should be suspected by the acute onset of the triad of lower abdominal pain, rectal burning, and diarrhea. Microscopic examination of stool is extremely useful and reveals multiple PMNs and red blood cells. This information should suggest bacillary dysentery, although the identification of the specific bacterial pathogen must await culture, because other microorganisms can cause a dysentery syndrome (e.g., Campylobacter, V. parahaemolyticus, Salmonella). Fecal specimens are the best source of a positive culture; blood and urine rarely are positive. Because Shigella species are fastidious, stool specimens or rectal swabs should be inoculated promptly into appropriate media. Sigmoidoscopy can confirm the diagnosis of colitis but is not necessary in most cases of shigellosis and is extremely uncomfortable in the setting of dysentery. Serologic and molecular tests are not useful for diagnosing acute cases of dysentery, although they are available for epidemiologic investigations. A subacute presentation of dysentery can masquerade as ulcerative colitis (see later). The patient might have endured bloody diarrhea, cramps, and rectal pain for two to four weeks, and sigmoidoscopic findings are indistinguishable from those of idiopathic ulcerative colitis (see Fig. 107-5) A colonic biopsy may be helpful to differentiate the two150: With dysentery, the inflammatory process is acute (PMNs) and involves the entire lamina propria; also, colonic glands are straight and without signs of regeneration. In ulcerative colitis, inflammation is chronic (lymphocytes) and involves mainly the lower one third of the lamina propria; colonic glands show signs of regeneration such as branching. Two major differences between dysentery and idiopathic ulcer-

ative colitis are a positive stool culture for Shigella and dramatic improvement in symptoms after treatment of patients with dysentery with appropriate antimicrobial agents. When in doubt, treatment for shigellosis is recommended. Deaths are rare in healthy persons, particularly adults, with bacillary dysentery; mortality usually is seen in young, often malnourished children or in debilitated patients— either the elderly or those with an immunodeficiency disease. A decreased level of consciousness and documented seizures are associated with a poor outcome in children.145

Treatment

The following general principles apply to the therapeutic approach to bacillary dysentery: Rehydration must be managed appropriately, as in any diarrheal disease. General supportive measures require attention. In the case of dysentery, children can have seizures related to high fever and electrolyte imbalance or meningismus. Antibiotic treatment is indicated for most patients with shigellosis. Fluid and Electrolyte Therapy Most patients with dysentery can be managed with oral rehydration. High-volume diarrhea is seen occasionally with shigellosis and can result in severe dehydration and hypovolemia. Intravenous fluid replacement is indicated in this situation and also when severe vomiting prevents oral replacement. Fluid losses can be replaced within a few hours by intravenous solutions, and oral replacement should be encouraged as soon as possible (see Table 107-6). Antidiarrheal remedies generally are unhelpful, and some believe they might even aggravate bacillary dysentery. Kaolin and pectate and other water-binding agents do not diminish stool volume or frequency. Antimicrobial Agents The major determinant in the decision to use antibiotics is the severity of the illness. In practice, patients with moderate or severe cases of dysentery should receive antibiotic therapy. Mild cases often pass as self-limited events, without a visit to a physician. If such cases are seen in the clinic or in the physician’s office, antibiotic therapy might not be required in light of the relatively benign course of infection. A reappraisal should be made when the culture report returns positive for Shigella. In many cases, diarrhea has already ceased. Patients with diarrhea for longer than one week should receive antibiotics. Ampicillin was previously the preferred antibiotic, with TMP-SMX as an alternative choice; however, many strains, if not most, in the United States and abroad now are resistant to these antibiotics. The quinolone antibiotics, such as ciprofloxacin, ofloxacin, and norfloxacin, are highly active in vitro against Shigella and are the drugs of choice. Singledose therapy with 1 g of ciprofloxacin is as effective as two doses or a five-day standard regimen in patients with Shigella infection; however, single-dose therapy proved less effective than multiple-dose regimens for patients with S. dysenteriae 1.151 Problems with using quinolones in the treatment of Shigella include the high cost of the drugs and concern about cartilage damage in young children. Nalidixic acid is an alternative therapeutic agent that has produced good results, although resistance develops rapidly with widespread use of this drug.152 Because there is now increasing evidence of the skeletal safety of quinolones in children,153 these drugs are being

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning studied increasingly in pediatric populations. With singledose pefloxacin therapy of infected children during an outbreak of multidrug-resistant S. dysenteriae 1 in Burundi, 91% of treated children became symptom free by day five; the remainder were substantially improved.154 None of the children experienced any joint problems during the fourweek period of follow-up. Similarly, a double-blind trial of pivmecillinam compared with ciprofloxacin suspension for childhood shigellosis found that ciprofloxacin resulted in clinical responses in 80% of children, with no associated arthropathy.155 Although early animal and human volunteer studies indicated that the use of antimotility agents in the treatment of invasive diarrhea might lead to prolonged fever and pathogen carriage, a recent study has challenged this dictum. Treatment of dysenteric patients with a combination of the synthetic antidiarrheal agent loperamide and ciprofloxacin resulted in a significantly shortened duration of diarrhea and decreased number of stools when compared with ciprofloxacin alone.156 The use of loperamide did not lead to prolonged fever or excretion of the pathogenic bacilli. Antibiotics for shigellosis must be absorbed from the bowel to reach organisms within the intestinal wall and lamina propria, and the only effective delivery system is the bloodstream.157 Nonabsorbable drugs, such as neomycin, kanamycin, paromomycin, colistin, and polymyxin, are clinically ineffective, despite in vitro sensitivity. Intravenous cefamandole also has proved disappointing. Curiously, amoxicillin, which is well absorbed and achieves higher serum levels than ampicillin, is not effective therapy for shigellosis.158 Chronic carriers of Shigella are rare. Postinfection carriage generally lasts less than three to four weeks and rarely exceeds three to four months. In circumstances in which eradication of the carrier state is deemed necessary, TMP-SMX or a fluoroquinolone should be used, guided by antibiotic sensitivity results. Such treatment eliminates the carrier state in approximately 90% of patients. Mild diarrhea and cramps can continue for days to weeks after treatment of bacillary dysentery, even when the organism is no longer present and the acute episode seems to have passed. These symptoms are not necessarily a cause for alarm, because the bowel might have sustained severe mucosal injury that requires time for repair. Approximately 10% of patients with shigellosis, however, may be left with these symptoms chronically, a condition called postinfection irritable bowel syndrome.159 Shigellosis is highly contagious. Spread within a family is common. Secondary cases can occur in hospitals among other patients, nurses, and physicians. Careful hand washing and stool precautions are important to prevent dissemination of this disease.

NONTYPHOIDAL SALMONELLOSIS

Nontyphoidal salmonellosis refers to disease caused by any serotype of the genus Salmonella, with the exception of S. typhi and S. paratyphi. Approximately 2000 serotypes and variants are potentially pathogenic for animals and humans.

Microbiology

Salmonella species are a group of Gram-negative bacilli, most of which are motile and produce acid and gas from glucose, mannitol, and sorbitol (except S. typhi and rare strains that produce only acid); they are active producers of hydrogen sulfide; and they are closely related to each other by somatic (O) and flagellar (H) antigens. These organisms are primarily intestinal pathogens, although some can be found in the bloodstream and internal organs of inverte-

brates; they are often isolated in sewage, river and sea water, and certain foods. Most Salmonella species have a wide range of hosts. The typing scheme for Salmonella species is based on antigenic structure, but in recent years, the name of the strain has been derived from the city in which it was first isolated (e.g., Montevideo, Heidelberg, Dublin, Newport). Most salmonellae are flagellated; using the proper growth conditions, the H (flagellar) and O (somatic) antigens can be tested for separately. Some strains, notably typhoid bacilli, have an additional somatic antigen associated with virulence (Vi). The Vi antigen prevents agglutination with O antigen. A positive correlation exists between virulence in mice and the amount of Vi antigen in a specific strain; however, this correlation does not carry over completely to humans. More than 70 anti-Vi phage types have been identified. For convenience in the laboratory, a series of serogroups, the Kauffmann-White serotypes, was developed based on shared antigens among the most common Salmonella types. Ninety percent of Salmonella species pathogenic for humans falls into groups A to E, which contain 40 serotypes. The application of newer molecular methods to the taxonomy of Salmonella has revealed that all serotypes of Salmonella belong to one species that includes seven subspecies, which can be differentiated with biochemical tests. To avoid confusion with previous nomenclature, the new species Salmonella enterica has been proposed.160 Using this approach, the typhoid bacillus would be named S. enterica subspecies enterica serotype typhi. Because this lengthy name is cumbersome, however, simpler acceptable versions are S. typhi or S. enterica serotype typhi.

Epidemiology

Salmonella is one of the great food-borne infections.2 The major route of passage is by the five Fs: flies, f ood, f ingers, feces, and f omites. The disease can cause large outbreaks, which often are associated with common-source routes of spread. A typical setting is an institutional supper or barbecue. Community outbreaks can persist for several months. For example, Riverside, California, experienced an epidemic involving 16,000 persons that raged for months and was related to a contaminated municipal water supply. Although approximately 45,000 cases of salmonellosis are reported annually, these numbers reflect vast underreporting, and it is estimated that 1.4 million cases of Salmonella food poisoning occur each year.2 The two most common serotypes in the United States are Salmonella enteritidis and Salmonella typhimurium. Attack rates of Salmonella show a strong relationship to age. Children younger than one year have the highest attack rate, a susceptibility that may be related to immunologic immaturity. There also are high attack rates and increased mortality in elderly persons. Nonhuman reservoirs play a crucial role in the transmission of the disease. In 500 outbreaks investigated over a 10-year period, almost 50% were related to animals or animal products. Poultry, meats, eggs, and dairy products were involved most often when a causative product was identified (Fig. 107-6). Salmonellae have a tendency to colonize domestic animals. Poultry has the highest incidence of Salmonella carriage, particularly hens, chickens, and ducks. Vertical transmission via the transovarian route can occur in chickens, so even normal-appearing eggs can be contaminated with Salmonella. Pigs and cattle also may be heavily contaminated. Many of these animals can cohabit peacefully with salmonellae and usually are asymptomatic. Other

1861

1862

Section X  Small and Large Intestine Dairy Pets products

4%

Table 107-9  Conditions That Predispose to Salmonella Infection

3%

Eggs 6% Unknown 28%

Person-to-person 10%

Meat 13%

Poultry 17%

Miscellaneous* 19%

Achlorhydria Autoimmune gastritis Gastroduodenal surgery Hemolytic anemia Bartonellosis Louse-borne relapsing fever Malaria Sickle cell disease Immunosuppression AIDS Chemotherapy Glucocorticoid therapy Radiation Malignancy Disseminated carcinoma Leukemia Lymphoma Schistosomiasis Ulcerative colitis AIDS, acquired immunodeficiency syndrome.

Figure 107-6.  Source of infection in 500 human salmonellosis outbreaks between 1966 and 1975. *Includes more than 50 sources that individually caused less than 3% of outbreaks. (Redrawn from the Centers for Disease Control. Salmonella Surveillance, Annual Summary, 1976. Washington, DC: U.S. Department of Health, Education and Welfare, Public Health Service; 1977.)

animals known to harbor Salmonella include buffalo, sheep, dogs, cats, rats, mice, guinea pigs, hamsters, seals, donkeys, turkeys, doves, pigeons, parrots, sparrows, lizards, whales, tortoises, house flies, ticks, lice, fleas, and cockroaches. Commercially prepared foods may be contaminated with salmonellae: 40% of turkeys examined in California, 50% of chickens in Massachusetts, and 20% of commercial egg whites have been shown in surveys to harbor these organisms. Large national and international outbreaks have been traced to commercially prepared chocolate balls, precooked roast beef, smoked whitefish, frozen eggs, ice cream, rawmilk cheese, alfalfa sprouts, cantaloupe, peanut butter, and powdered milk. Other commercial products not directly related to foods, such as carmine dye or brewer’s yeast, also can be contaminated. Infected pets, especially turtles and lizards, have been implicated in the transmission of salmonellosis.

Pathogenic Mechanisms

Salmonellae attack the ileum and, to a lesser extent, the colon. They cause mild mucosal ulcerations, rapidly make their way through the epithelial surface to the lamina propria and then to the lymphatics and bloodstream, and then rapidly spread to other organs hematogenously. Histologic sections show edematous, shortened crypts and PMNs in the lamina propria. A series of pathogenic factors, each controlled by plasmids or chromosomal loci, are required for a Salmonella strain to be fully pathogenic. Specific plasmids encode for bacterial spread from Peyer’s patches to other sites in the body161,162 or the ability to survive within macrophages after phagocytosis.163 The outer membrane lipopolysaccharide and the Vi antigen are additional virulence factors. Another virulence factor imparts the ability of salmonellae to elicit transepithelial signaling to neutrophils,164 which contribute to cell damage and secretion. Finally, Salmonella

strains produce enterotoxins that can play a role in diarrhea.161,165,166 The infectivity of a specific strain is related to its serotype and the inoculum size. For example, 105 Salmonella newport produce illness in some volunteers, whereas 109 Salmonella pullorum are unable to do so. The latter strain is poorly adapted to humans, as suggested by its rarity in clinical infections; it is well adapted to chickens, from which it is often isolated. A dose-response curve has been determined for certain strains of Salmonella; an approximately 50% infection rate is seen with 107 organisms, whereas the infectivity rate rises to 90% with 109 organisms. In experimental animals, the number of bacteria required to produce infections can be reduced by pretreating the animals with antibiotics. Antibiotic exposure in humans also increases susceptibility to Salmonella infection.167 In addition, reduced or absent gastric acid is known to increase susceptibility to infection, because acid in the stomach kills many of the challenge organisms.140,141

Predisposing Conditions

A number of conditions increase the risk of salmonellosis (Table 107-9). The relationship between sickle cell anemia and Salmonella osteomyelitis is well known. Indeed, several forms of hemolytic anemia predispose to this infection, including malaria, bartonellosis, and louse-borne relapsing fever. The presumed mechanism of increased susceptibility is blockage of the reticuloendothelial system by macrophages that have ingested breakdown products of red blood cells, thereby reducing their ability to phagocytose salmonellae.168 Patients with sickle cell anemia also have a decreased capacity to opsonize salmonellae because of defective activation of the alternative complement pathway.169 Neoplastic disease is associated with an increased risk of salmonellosis. Leukemia, lymphoma, and disseminated malignancy predispose patients to bloodstream invasion by this organism.170 Use of glucocorticoids, chemotherapy, or radiation therapy also is associated with Salmonella sepsis. In AIDS patients, persistent Salmonella bacteremia, only temporarily yielding to antibiotic therapy, is related to the profound suppression of cell-mediated immunity.171 Gastric surgery appears to be an important predisposing condition in the development of Salmonella infection because destruc-

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Table 107-10  Relative Frequencies of the Clinical Syndromes of Salmonella Infection SYNDROME

FEATURES

Gastroenteritis

Characterized by mild to severe and dehydrating (dysenteric) colitis With or without gastroenteritis, endocarditis, arteritis, or AIDS With or without gastroenteritis May involve meninges, bones and joints, wounds, gallbladder, and may form abscesses —

Bacteremia

Typhoid fever (“enteric fever”) Localized infection

Carrier state (>1 yr)

FREQUENCY (%) 75

5-10

5-10

<1

AIDS, acquired immunodeficiency syndrome.

tion of the gastric acid barrier enhances the host’s susceptibility to infection.7 Schistosomiasis also is associated with invasive salmonellosis.172 Ulcerative colitis also can pre­ dispose to Salmonella infection and the carrier state.173

Clinical Features

Although many serotypes of Salmonella are restricted to animals, some strains are less fastidious and can cause serious human infection. S. typhimurium causes a spectrum of disease ranging from gastroenteritis to bacteremia; S. newport causes septicemia, and S. typhi, S. paratyphi, Salmonella schottmülleri, and Salmonella hirschfeldii (the last three known formerly as S. paratyphi A, B, and C, respectively) cause enteric (typhoid) fever. Five clinical syndromes are seen with Salmonella (Table 107-10), including gastroenteritis, noted in 75% of Salmonella infections; bacteremia, with or without gastrointestinal involvement, seen in approximately 10% of cases; typhoidal or enteric fever, seen with all typhoid and paratyphi strains and in approximately 8% of other Salmonella infections; localized infections (e.g., bones, joints, meninges, and blood vessels), seen in approximately 5%; and a carrier state in asymptomatic people (the organism usually is harbored in the gallbladder).174 The most common syndrome caused by Salmonella is gastroenteritis. The incubation period is usually six to 48 hours but can last as long as seven to 12 days. Initial symptoms are nausea and vomiting, followed by abdominal cramps and diarrhea. The diarrhea usually lasts three or four days and is accompanied by fever in approximately 50% of persons. Diarrhea can vary from a few loose stools to dysentery with grossly bloody and mucopurulent feces to a cholera-like syndrome, in patients who are hypochlorhydric or achlorhydric.11,175 Persistent fever or specific findings on physical examination suggest bacteremia or focal infection. Salmonella bacteremia is similar to sepsis caused by other Gram-negative bacteria. Recurrent Salmonella bacteremia is seen in patients with AIDS.171 Once the organism invades the bloodstream, almost any organ can become involved, (e.g., meningitis, arteritis, endocarditis, osteomyelitis, wound infections, septic arthritis, and focal abscesses).174 Patients can become chronic carriers (defined as persistence for longer than one year) of nontyphoidal Salmonella.

The overall carrier rate is between 2 : 1000 and 6 : 1000 infected persons. Children, especially neonates, and persons older than 60 years of age become carriers more commonly than do others. Also, structural abnormalities in the biliary tract (e.g., cholelithiasis) or the urinary tract (e.g., nephrolithiasis) predispose to and perpetuate the carrier state.176

Salmonella Colitis

Involvement of the colon in the course of Salmonella gastroenteritis probably is common. Although most patients with Salmonella present with mild diarrhea and watery bowel movements, colonic involvement can dominate the clinical picture; toxic megacolon and perforation due to Salmonella can occur.177 Patients with Salmonella colitis typically have diarrhea for 10 to 15 days before the diagnosis is established. In contrast, patients with the usual form of gastroenteritis are symptomatic for five days or less. In the colonic form, diarrhea is more persistent, even though the organism might have disappeared from the feces on clinical presentation. Bowel movements are grossly bloody in approximately one half of the patients with Salmonella colitis. Sigmoidoscopic findings include hyperemia, granularity, friability, and ulcerations. Rectal biopsy specimens reveal mucosal ulcerations, hemorrhage, crypt abscesses, straight glands, and a PMN infiltration that involves the lamina propria. Barium enema films confirm these findings and usually show pancolitis. In the acute period, there is no reliable method to distinguish idiopathic ulcerative colitis from Salmonella colitis, except by a positive stool culture. Any patient with an acute onset of colitis, no past history of colitis, and a duration of symptoms of three weeks or less should be considered to have infectious colitis; Salmonella, as well as EHEC, Shigella, Campylobacter, and C. difficile, are important diagnostic considerations. The course of Salmonella colitis varies and can be as short as one week or as long as two to three months. The average duration of illness is three weeks. Complications include toxic megacolon, bleeding, and overwhelming sepsis. It is important to recognize Salmonella colitis, so that inappropriate therapy is not administered. Glucocorticoids can exacerbate Salmonella colitis and result in silent per­ foration and septicemia. Finally, patients can be reassured of the self-limited course of Salmonella colitis as opposed to the chronic relapsing course of idiopathic ulcerative colitis.

Treatment

Although many antibiotics have been used to treat nontyphoidal Salmonella gastroenteritis, all have failed to alter the rate of clinical recovery. In fact, antibiotic therapy increases the frequency and duration of intestinal carriage of these organisms.178 A review of 12 randomized trials found no differences in the duration of illness, diarrhea, or fever between patients treated with antibiotics and those treated with placebo.179 Relapses were more common in those treated with antimicrobial agents, as were adverse drug reactions. Thus antimicrobial therapy should not be used in most cases of Salmonella gastroenteritis. Despite this general rule, antibiotics should be used when Salmonella gastroenteritis complicates certain conditions (Table 107-11), such as lymphoproliferative disorders, malignant disease, immunosuppressed states (AIDS and congenital or acquired disorders), organ transplantation, known or suspected abnormalities of the cardiovascular system (e.g., prosthetic heart valves, vascular grafts, aneu-

1863

1864

Section X  Small and Large Intestine Table 107-11  Indications for Antibiotic Therapy in Salmonella Gastroenteritis Abnormal cardiovascular system Aneurysms Prosthetic heart valves Valvular heart disease Vascular grafts Extreme ages of life Hemolytic anemia Immunosuppression AIDS Congenital and acquired immunosuppressive disorders Glucorticoid treatment Organ transplants Lymphoproliferative disorders Leukemia Lymphoma Malignancies Pregnancy Prosthetic orthopedic devices Sepsis AIDS, acquired immunodeficiency syndrome.

rysms, and rheumatic or congenital valvular heart disease), foreign bodies implanted in the skeletal system, hemolytic anemia, extreme ages of life, and pregnancy. In addition, patients with Salmonella gastroenteritis should be treated with antibiotics when they exhibit findings of severe sepsis including high fever, rigors, hypotension, decreased kidney function, and systemic toxicity. If a decision is made to initiate therapy, the choice of drug may be problematic because of high levels of antibiotic resistance to ampicillin or TMP-SMX; currently, a fluoroquinolone is the drug of first choice. For patients with strains sensitive to ampicillin or TMP-SMX, these agents can be used (Table 107-12). The quinolones, particularly ciprofloxacin, have shown good results in patients with enteric fever180 and in chronic carriers.181 Ciprofloxacin therapy in patients with uncomplicated Salmonella gastroenteritis, however, has led to a high relapse rate that is associated with more-prolonged fecal excretion of salmonellae than seen in placebo-treated control subjects.178 As might be expected, resistance to ciprofloxacin has been observed during therapy.182 Since the turn of the 21st century there has been an increase in the isolation of quinolone-resistant Salmonella isolates in Europe.183 The use of quinolones for veterinary use has been blamed for this rise in resistance. Resistance to ceftriaxone of a Salmonella strain from livestock also has been observed in the United States.184 As a consequence of the increasing levels of drug resistance, both domestically and internationally, antimicrobial therapy of Salmonella infections must be limited to high-risk patients and should be based on sensitivity testing.

TYPHOID FEVER

Typhoid (Greek typhos, “stupor” or “cloudy”) fever is a febrile illness of prolonged duration, marked by hectic fever, delirium, persistent bloodstream infection, splenic enlargement, abdominal pain, and a variety of systemic manifestations. The illness caused by this pathogen differs from the nontyphoidal Salmonella infections in several respects. Typhoidal disease is not truly an intestinal disease and has more systemic than intestinal symptoms. S. typhi is remarkably adapted to humans, who represent the only

natural reservoir; the other salmonellae are associated with animals. Although S. typhi is the main cause of typhoid fever, other Salmonella serotypes occasionally produce a similar clinical picture, known variously as typhoidal disease, enteric fever, or paratyphoid fever. These serotypes are S. paratyphi (formerly S. paratyphi A), S. schottmülleri (formerly S. paratyphi B), and S. hirschfeldii (formerly S. paratyphi C), as well as others, including S. typhimurium.

Microbiology

S. typhi is biochemically similar to other Salmonella species and is distinguished primarily by its specific antigens. As a rule, this organism produces little or no gas from carbohydrates, elaborates only small amounts of hydrogen sulfide, and bears the Vi antigen on its surface. These markers should alert the laboratory to the possibility of this pathogen; confirmation of S. typhi is accomplished by serotyping.

Pathogenic Mechanisms

The pathologic events of typhoid fever are initiated in the intestinal tract after oral ingestion of typhoid bacilli.16 The organism penetrates the small intestinal mucosa and makes its way rapidly to the lymphatics, the mesenteric nodes, and the bloodstream. There is a paucity of local inflammatory findings, which explains the lack of intestinal symptoms at this early stage. This sequence of events is in marked contrast to that of other forms of salmonellosis and shigellosis, in which intestinal findings are prominent at the onset. After the initial bacteremia, the organism is sequestered in macrophages and monocytic cells of the reticuloendothelial system. It undergoes multiplication and re-emerges several days later in recurrent waves of bacteremia, an event that initiates the symptomatic phase of infection. Now in great numbers, the organism is spread throughout the host and infects many organ sites. The intestinal tract may be seeded by direct bacteremic spread to Peyer’s patches in the terminal ileum or via drainage of contaminated bile from the gallbladder, which often harbors large numbers of organisms. Hyperplasia of the reticuloendothelial system, including lymph nodes, liver, and spleen, is characteristic of typhoid fever. The liver contains discrete micronodular areas of necrosis surrounded by macrophages and lymphocytes. Inflammation of the gallbladder is common and can lead to acute cholecystitis. Patients with preexisting gallbladder disease have a tendency to become carriers, because the bacillus becomes intimately associated with the existing chronic infection and may be incorporated within gallstones. Lymphoid follicles in the intestine, such as Peyer’s patches, become hyperplastic and infiltrated with macrophages, lymphocytes, and red blood cells. Subsequently, a follicle may ulcerate, penetrate through the submucosa to the intestinal lumen, and discharge large numbers of typhoid bacilli. As the bowel wall progressively is involved, it becomes paper-thin (most commonly in the terminal ileum), and is susceptible to perforation into the peritoneal cavity. Erosion into blood vessels produces severe intestinal hemorrhage.

Epidemiology

Improvements in environmental sanitation have reduced the incidence of typhoid fever in industrialized nations. Approximately 400 to 500 cases occur each year in the United States, chiefly in young people. Large-scale epidemics of typhoid occur on a regular basis in developing countries and usually are traced to contaminated food that is

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Table 107-12  Antibiotic Therapy for Bacterial Enteropathogens in Adults ORGANISM OR SYNDROME

RECOMMENDED ANTIBIOTIC(S)

ALTERNATIVE ANTIBIOTIC(S)

Shigella Species   Shigella infection

Ciprofloxacin 500 mg twice daily × 3 days

Norfloxacin 400 mg twice daily × 3 days TMP-SMX 160 mg/800 mg twice daily × 3 days Nalidixic acid 1 g three or four times daily × 5 days

None Ciprofloxacin 500 mg twice daily × 7 days

None TMP-SMX 160 mg/800 mg twice daily × 7 days Ceftriaxone 1 g IV twice daily × 7 days Tetracycline 500 mg four times daily × 7 days

Chloramphenicol 50 mg/kg/day in four divided doses (not for use in vascular infections) Ampicillin 1 g every 4-6 hr Amoxicillin 1 g every 6-8 hr Chloramphenicol 500 mg orally or IV every 8 hours for at least 14 days

Cefotaxime, especially for meningitis or vascular infection, 2 g every 4-6 hr Ciprofloxacin 750 mg twice daily, especially long term for prevention of recurrent bacteremia in patients with AIDS

Salmonella Species   Gastroenteritis, uncomplicated   Gastroenteritis, severe; immunocompromised patients; pregnant women; prostheses; cancer   Bacteremia and localized infection*

Typhoid and enteric fevers

Campylobacter Species   C. jejuni    Mild gastroenteritis    Severe or prolonged gastroenteritis   Campylobacter fetus infection Yersinia Species   Yersinia enterocolitica infection

Escherichia coli†   Enteroinvasive

  Enterotoxigenic   Enteropathogenic

  Enterohemorrhagic Vibrio Species   Vibrio cholerae   Vibrio parahaemolyticus   Vibrio vulnificus and Vibrio alginolyticus

None Erythromycin 250 mg twice daily × 5 days Ampicillin and gentamicin IV

Amoxicillin 1 g four times daily × 2 wk Ciprofloxacin 500 mg twice daily × 7-14 days TMP-SMX 160 mg/800 mg 1 or 2 tablets twice daily × 14 days None Ciprofloxacin 500 mg twice daily Norfloxacin 400 mg twice daily Erythromycin 0.25-1 g every 6 hr

Not usually required Tetracycline 500 mg four times daily TMP-SMX 160 mg/800 mg twice daily Ciprofloxacin 500 mg twice daily

Chloramphenicol 50 mg/kg/day in four divided doses An aminoglycoside

No controlled studies; consider empiric antibiotic therapy Ciprofloxacin 500 mg twice daily Norfloxacin 500 mg twice daily Ciprofloxacin 500 mg twice daily Norfloxacin 500 mg twice daily Rifaxamin 500 mg twice daily × 3 days No controlled studies; consider empiric antibiotic therapy Ciprofloxacin 500 mg twice daily Norfloxacin 500 mg twice daily Unclear if antibiotics are effective; may be harmful

TMP-SMX 160 mg/800 mg twice daily

Tetracycline 500 mg every 6 hr × 3 days

Ciprofloxacin 500 mg twice daily × 3 days Doxycycline 300 mg once or 100 mg once daily × 3 days

Supportive therapy Tetracycline 500 mg four times daily or 0.5-1 g IV every 12 hr

TMP-SMX 160 mg/800 mg twice daily TMP-SMX 160 mg/800 mg twice daily

Chloramphenicol 50 mg/kg/day four times daily Penicillin G 20 million units/day IV

*Bacteremia is treated for 10-14 days; endocarditis and osteomyelitis are treated for ≥4-6 wk. † Enteroaggregative and diffusely enteroadherent E. coli are omitted from this table because these types are defined in research laboratories and are not diagnosed in routine clinical practice. Note: All antibiotics are administered orally unless otherwise indicated. AIDS, acquired immunodeficiency syndrome; IV, intravenous; TMP-SMX, trimethoprim-sulfamethoxazole. From Giannella RA. Treatment of acute infectious diarrhea. In Wolfe M, editor. Therapy of Digestive Disorders: A Companion to Sleisinger and Fordtran’s Gastrointestinal and Liver Disease. 2nd ed. London: Elsevier; 2006.

imported from an endemic area or to contaminated water supplies.185 Because S. typhi cohabits exclusively with humans, the appearance of a case could indicate the presence of a carrier. An investigation by public health authorities should be instituted to determine the source and the presence of other cases. As they are discovered, chronic carriers are registered with the health authorities, and the particular microorganism is phage-typed so that it can be traced in the event of an outbreak. Registered carriers represent only some of the potential reservoir, however, and do not take into account

imported cases of typhoid, which represent more than 70% of the acute infections in the United States.186

Clinical Features

In its classic form and without treatment, typhoid fever lasts about four weeks and evolves in a manner consistent with the pathologic events. The illness is described traditionally as a series of one-week stages, although this pattern may be altered in mild cases and by antibiotic treatment.187 The incubation period generally is seven to 14 days, with wide variations.

1865

1866

Section X  Small and Large Intestine During the first week, high fever, headache, and abdominal pain are common. The pulse often is slower than would be expected for the degree of fever, a finding referred to as Faget’s sign. Abdominal pain is localized to the right lower quadrant in most cases but can be diffuse. In approximately 50% of patients, there is no change in bowel habits; in fact, constipation is more common than diarrhea in children with typhoid fever. Near the end of the first week, enlargement of the spleen is noticeable, and an evanescent classic rash (rose spots) becomes manifest, most commonly on the chest. During the second week, the fever becomes more continuous, and the patient looks sick and withdrawn. During the third week, the patient’s illness evolves into the “typhoidal state,” with disordered mentation and, in some cases, extreme toxemia. It is from this altered mental state that the term typhoid derives. In this period there is often intestinal involvement, manifested clinically by greenish pea-soup– like diarrhea and the dire complications of intestinal perforation and hemorrhage. The fourth week brings slackening of the fever and improvement in the clinical status, if the patient survives and recovers. Typhoid fever is a less-severe illness in previously healthy adults who seek medical attention for the earliest symptoms of fever, lassitude, and headache than it is in those who wait. Prompt diagnosis and appropriate therapy interrupt the classic four-week scenario and produce an aborted illness consisting of little more than a few days of fever and malaise. Because the typhoid bacillus is disseminated widely through recurrent waves of bacteremia, many organ sites can be involved. Patients with typhoid fever can have pneumonia, pyelonephritis, and metastatic infections of bone, large joints, and the brain. The gallbladder and liver are involved with inflammatory changes. Acute cholecystitis can occur during the initial two to three weeks, and jaundice, resulting from diffuse hepatic inflammation, has been observed in some patients. The preeminent complications are intestinal hemorrhage and perforation.188 These events are most likely to occur in the third week and during convalescence and are not related to the severity of the disease; they tend to occur in the same patient, with bleeding serving as a harbinger of possible perforation. Bleeding may be sudden and severe or a slow ooze. Before the availability of antibiotics, the incidence of hemorrhage was as high as 20%; it is less common since specific treatment has become available. Approximately 3% of patients with typhoid fever experience intestinal perforation, most commonly in the ileum.189 Onset of perforation may be sudden, with signs of an acute abdomen, or there may be a leak of intraluminal contents to form an abscess in the lower quadrant or pelvis, producing a more insidious course. After defervescence has occurred and the patient has apparently ridden through the storm, a potential for recurrence remains. Relapse generally occurs eight to 10 days after cessation of drug therapy and consists of a reenactment of the major manifestations. The organism is the same as the one that caused the original infection, with the identical antimicrobial susceptibility pattern.

Carrier State

After six weeks, approximately 50% of typhoid victims still shed organisms in their feces. This figure declines progressively with time, and after three months only 5% to 10% are excreters; by one year the frequency is 1% to 3%.190 The chronic carrier is identified by positive stool cultures for S. typhi at least one year after the acute episode or, in some

cases, positive stool cultures without a documented history of disease. The probability of spontaneously aborting the carrier state is highly unlikely after this time. Chronic carriers are more common in older age groups, women (a 3 : 1 ratio of women to men), and persons with biliary disease. The organism usually is harbored in the gallbladder, although occasionally it is carried in the large intestine without involvement of the biliary tract.

Diagnosis

The diagnosis of typhoid fever is established by isolating the organism. Blood culture is the primary diagnostic test, is positive in 90% of patients during the first week, and remains positive for several weeks thereafter if the patient is not treated. Bone marrow culture also has a high yield, even in treated patients.191 Stool cultures become positive in the second and third weeks. Sampling duodenal contents by a string test yields a positive culture in 70% of patients. By the third week, urine cultures reveal the organism in approximately 25% of patients. The titer of agglutinins against somatic (O) antigen (Widal test) rises during the second and third weeks of illness. A serum O titer of 1 : 80 or more in a nonimmunized person is suggestive of typhoid fever, and a titer of 1 : 320 usually is diagnostic in the appropriate clinical setting; a fourfold rise in titer provides stronger evidence. Although the H antigen is less specific than the O titer and is likely to be elevated from prior immunization or by infection with other enteric bacteria, an initial H antigen titer of 1 : 640 is strongly suggestive of typhoid fever. There are many false-positive and occasional false-negative Widal reactions, so that diagnosis based on a rise in titer alone is tenuous. Other serologic tests have become available and permit rapid diagnosis of typhoid fever with a higher sensitivity and specificity relative to blood culture than the Widal test; PCR assays for the diagnosis of S. typhi have been developed.

Treatment

Drug resistance, mediated by plasmids, occurs among typhoid bacilli. Most strains are susceptible to chloramphenicol and ampicillin, although notable epidemics with strains resistant to either or both of these drugs have been reported in recent years. Hence, a great effort should be made in each case to isolate the organism and perform drug susceptibility tests. Antibiotic treatment is listed in Table 107-12. Chloramphenicol has high activity against most clinical isolates of typhoid bacilli. The response to therapy is remarkably constant, and defervescence regularly occurs three to five days after treatment is begun.188 The clinical condition improves within one to two days, with decreased toxemia and slowly declining fever. In adults, chloramphenicol should be given in a total daily dose of 2 g, administered in four equally divided doses by mouth. Occasionally, in very sick patients it may be necessary to give the drug by the intravenous route in the same total daily dose. Oral medication can be given upon clinical improvement. Chloramphenicol is well absorbed from the intestinal tract but is rather poorly absorbed from intramuscular sites; the intramuscular route is to be avoided. The duration of treatment is two weeks; prolongation of this treatment does not reduce the frequency of complications or carriers. Intestinal perforation and hemorrhage can occur during apparently successful treatment. Relapse can follow an otherwise uneventful course and should be treated with the same drug as was used initially. Ampicillin has been recommended as alternative therapy but it has been disappointing in comparison with chloram-

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning phenicol.187,192 Amoxicillin, a closely related drug, provides better absorption and increased efficacy. Several studies have shown that amoxicillin, in doses of 4 g/day in four divided doses, has good activity. TMP-SMX also has been used with good results in typhoid fever. The advent of plasmid-mediated multidrug resistance and newer potentially more effective antimicrobial agents, such as the quinolones and third-generation cephalo­ sporins, has led to a reevaluation of the treatment of typhoid fever. The fluoroquinolones, ciprofloxacin and ofloxacin, have been found to be highly effective therapy for infections caused by multidrug-resistant S. typhi and S. paratyphi. Long-term fecal carriage of S. typhi is rare in patients treated with quinolones. A 10- to 14-day course of a quinolone has proved highly effective for treating enteric fever, with cure rates consistently close to 100%.72 The only exception has been norfloxacin, which has provided slightly lower cure rates of 83% to 90%.72 Defervescence generally occurs within three to five days of initiating therapy. The optimal length of fluoroquinolone therapy for typhoid fever has not been fully elucidated. A number of studies have shown that courses of therapy ranging from seven to 14 days provide a high degree of success. Courses of therapy shorter than five to six days have been associated with unacceptable levels of failure.72 The duration of fever before treatment, severity of infection at the time of presentation, and time to defervescence are factors that must be considered when determining the duration of fluoroquinolone therapy in a patient with enteric fever. The frequency of resistance of S. typhi to ciprofloxacin has been increasing gradually, especially in the Indian subcontinent, central Asia, and Vietnam.185,193,194 Third-generation cephalosporins, such as cefotaxime, ceftriaxone, and cefoperazone, also have been used successfully to treat typhoid fever; courses as short as three days have been shown to be as effective as the usual 10- to 14-day regimens.195,196 By contrast, one trial that compared a fiveday course of ofloxacin with a seven-day course of cefixime found that the median fever clearance time was significantly longer and the rate of treatment failure was higher in children treated with cefixime.197 Glucocorticoids are administered for severe toxemia and fever and can produce a dramatic response in patients with profound sepsis.198 Glucocorticoids should be given in high doses—for example, 60 mg/day of prednisone divided into four doses—and tapered rapidly over the following three days. The wide experience with glucocorticoid treatment has failed to show any adverse effects, although the potential for masking intestinal perforation always is present. Glucocorticoids should be reserved for patients with severe toxicity. Studies have emphasized the importance of aggressive surgical intervention in typhoid fever.189,199,200 Indications for surgery include progressive peritoneal signs or formation of abscesses. Closure of the intestinal perforation coupled with broad-spectrum antibiotics has resulted in reduced mortality from this dreaded complication200; the ileum may be riddled with multiple perforations, and resection may be required. A chronic carrier who has been discharging S. typhi for longer than one year can be treated with antimicrobials in an attempt to eliminate the infection. The quinolone antibiotics, including ciprofloxacin and norfloxacin, have become the treatment of choice in eradicating the carrier state.201 Reappearance of the carrier state after such treatment generally is associated with gallbladder disease. Cholecystectomy eliminates the carrier state in 85% of carriers with gall-

stones or chronic cholecystitis, but it is recommended only for persons whose profession is incompatible with the typhoid carrier state, such as food handlers and health care providers.

Vaccines

Two types of vaccine are available currently. The acetoneinactivated injectable vaccine affords 55% to 85% protection for two to five years. A live attenuated S. typhi strain, Ty21a, which is given by mouth, produced 96% protection in an initial field trial in Alexandria, Egypt,202 although subsequent studies showed less-impressive results.203,204 Because of its low toxicity and ease of administration, this vaccine should be used for travelers to high-risk areas, particularly in the developing world.

CAMPYLOBACTER SPECIES

The most important Campylobacter species found in human infections are C. jejuni, a major cause of diarrhea; C. fetus, which is generally found in immunocompromised patients; C. coli, a rare cause of gastroenteritis; and two new species, C. cinaedi and C. fennelliae, which are found in male homosexuals. Other species that cause diarrhea on rare occasions are C. hyointestinalis, C. upsaliensis, and C. laridis. The incidence and importance of human Campylobacter gastroenteritis have been recognized increasingly in recent years. It is estimated that 4% to 11% of all cases of diarrhea in the United States are caused by C. jejuni, and the isolation of Campylobacter often exceeds that of Salmonella and Shigella.205 The organism is isolated only rarely from fecal samples of asymptomatic persons, except in the tropics, where the incidence of Campylobacter infections is higher and there are many asymptomatic carriers.

Epidemiology

Transmission to humans appears to occur most commonly from infected animals and their food products. The reservoir for Campylobacter is enormous, because many animals can be infected, including cattle, sheep, swine, birds (poultry and others), and dogs. Most human infections are related to consumption of improperly cooked or contaminated foods. Just as for Salmonella, chicken seems to be the major source, accounting for 50% to 70% of infections.206

Clinical Features

The incubation period is 24 to 72 hours after organisms are ingested, but it can extend as long as 10 days. There is a wide spectrum of clinical illness, from frank dysentery to watery diarrhea to asymptomatic excretion.207 Diarrhea and fever are almost invariable (90%). Abdominal pain is usually present (70%), and the patient may note bloody stools (50%). Constitutional symptoms such as headache, myalgia, backache, malaise, anorexia, and vomiting are common. Stool examination suggests colitis on the basis of fecal leukocytes and occult blood.207 Endoscopy might reveal an inflammatory colitis (Figs. 107-7 and 107-8). The duration of illness usually is less than one week, although symptoms can persist for two weeks or more, and relapses occur in as many as 25% of patients. Prolonged carriage of Campylobacter for two to 10 weeks after the onset of illness occurs in 16% of patients.208 Infections rarely are complicated by gastrointestinal hemorrhage, toxic megacolon, pancreatitis, cholecystitis, HUS, bacteremia, meningitis, and purulent arthritis.207 Postinfection complications include reactive arthritis, usually in patients with the HLA-B27 phenotype, Guillain-

1867

1868

Section X  Small and Large Intestine The most reliable way to diagnose Campylobacter gastroenteritis is by stool culture. A selective isolation medium containing antibiotics must be used because Campylobacter organisms grow more slowly than do other enteric bacteria; the plates are grown at 42°C under CO2 and reduced oxygen conditions. Dark-field or phase-contrast microscopy of fresh diarrheal stool shows the organism as a curved, highly motile rod, with darting corkscrew movements.

Treatment

Figure 107-7.  Colonoscopic appearance of Campylobacter proctitis. The proctitis is patchy, with areas of erythema and erosion (seen on the left). The rectal mucosa on the right is hyperemic, but without loss of the mucosal vascular pattern. (From Wilcox CM. Atlas of Clinical Gastrointestinal Endoscopy. Philadelphia: WB Saunders; 1995.)

Figure 107-8.  Photomicrograph of a patient with acute self-limited colitis, the cause of which could have been an infection with Campylobacter, Salmonella, Escherichia coli species, or any one of a number of other bacteria. The presence of inflammatory cells throughout the lamina propria and straight glands without architectural distortion or branching helps differentiate acute colitis from chronic colitis. (Courtesy of Feldman’s Online Atlas, Current Medicine.)

Barré syndrome,207,209 and immunoproliferative small intestinal disease.210

Diagnosis

Although certain clinical features suggest the diagnosis of Campylobacter rather than other pathogens, the diagnosis only can be established by culture. Features suggesting Campylobacter infection are a prodrome consisting of constitutional symptoms with coryza, headache, and generalized malaise; a prolonged, often biphasic diarrheal illness manifesting initially with diarrhea, followed by slight improvement and then by increasing severity; and many white and red blood cells on microscopic examination of the stool.

Antibiotic treatment is listed in Table 107-12. Although C. jejuni is sensitive to erythromycin in vitro, three controlled therapeutic trials with this drug have shown no effect on the clinical course compared with placebo.211 One study showed some clinical benefit when the antibiotic was started within three days of the onset of symptoms. If therapy is delayed beyond four days, it yields no clinical improvement; fecal excretion of the organism, however, is reduced by erythromycin. The fluoroquinolone antibiotics, such as ciprofloxacin, also are active against these organisms.212 Resistance to fluoroquinolones has been observed during the course of treatment for Campylobacter diarrhea,213,214 and such resistance is a major problem in some parts of the developing world. A study of U.S. military personnel in Thailand found that 50% of isolates were resistant to ciprofloxacin, whereas none was resistant to azithromycin.215 A large study of human Campylobacter isolates in Minnesota found a rise in quinolone resistance from 1.3% to 10.2% between 1992 and 1998.216 Factors associated with quinolone resistance of Campylobacter species include foreign travel and local patterns of fluoroquinolone use, especially if these agents are used in animal husbandry.216,217 In locales where quinolone resistance is common, azithromycin has been shown to be superior to ciprofloxacin in decreasing the excretion of Campylobacter species and equivalent in terms of reducing the duration of symptoms.215 Mild cases of Campylobacter do not benefit from antibiotic therapy, but treatment should be given, early if possible, to patients with dysentery and those with high fever suggesting bacteremia. Because of the difficulty in making an etiologic diagnosis on clinical grounds, a quinolone antibiotic should be used empirically because of its activity against Campylobacter, Shigella, and other enteric pathogens.

YERSINIA ENTEROCOLITICA

Y. enterocolitica causes a spectrum of clinical illnesses from simple gastroenteritis to invasive ileitis and colitis.218 It is a non–lactose-fermenting, urease-positive, Gram-negative rod. More than 50 serogroups and five biotypes have been identified.218 Pathogenic mechanisms include the ability to invade epithelial cells and the production of a heat-stable enterotoxin, which is elaborated at 25°C but not at 37°C. Not all strains have these pathogenic properties. The organism targets the epithelium overlying Peyer’s patches for invasion, after which it proliferates within the follicles and spreads to the lamina propria. The abilities to attach to and penetrate epithelial cells are determined by the inv gene, which encodes a 103-kDa protein known as invasin.219

Epidemiology

Yersinia gastroenteritis has been reported more often in Scandinavian and other European countries than in the United States. Several epidemics have been related to the consumption of contaminated milk and ice cream. The

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning organism can be isolated from many animals, including puppies, cats, cows, pigs, chickens, and horses. Animals, either as pets or food sources, are believed to be involved in the transmission of this disease.220 The serotypes most commonly involved in Scandinavia and Europe are O:3 and O:9, and Canada has many serotype O:3 isolates.221 Most of the isolates in the United States are serotype O:8. Serogroups O:8 and O:5,27 have been responsible for most episodes of invasive disease in the United States.

querading as appendicitis. For the chronic relapsing form of diarrhea, antibiotic therapy has not proved useful. Septicemia in the immunocompromised patient is associated with a high mortality, with no apparent benefit from antibiotics, although treatment is mandatory in this setting. Antibiotic treatment is listed in Table 107-12.

Clinical Features

The major causes of acute gastroenteritis in the United States and in the rest of the world are viruses, which account for 30% to 40% of acute episodes of diarrhea. The leading human pathogens can be grouped into five categories: rotavirus, calicivirus including Norwalk virus (renamed Norovirus), enteric adenovirus, astrovirus, and torovirus (Table 107-13).229

Several clinical syndromes have been described with Yersinia and tend to vary with the age of the patient and the underlying disease state.218,222 Enterocolitis is the most common clinical condition and accounts for two thirds of all reported cases. This illness occurs most commonly in children younger than five years.219,222 Presentation is nonspecific, with fever, abdominal cramps, and diarrhea, usually lasting one to three weeks. Microscopic examination of stool reveals leukocytes and red blood cells in most instances. Profuse watery diarrhea, possibly related to the enterotoxin, also can occur. Diarrhea can persist for several weeks and can raise the possibility of inflammatory bowel disease. Radiologic findings, particularly in prolonged cases, are most intense in the terminal ileum and can resemble those of Crohn’s disease223; most patients, however, have normal findings on endoscopy, intestinal biopsy, and barium studies.221 In children older than five years, mesenteric adenitis and associated ileitis have been described. Accompanying symptoms include nausea, vomiting, and oral aphthous ulcers. Affected children often undergo a laparotomy, at which time enlarged mesenteric nodes and an ulcerated ileitis are observed. Clinically, the condition may be confused with acute appendicitis, although ultrasonography can be useful in separating these processes.224 Yersinia is less likely to cause severe disease in adults, in whom acute diarrhea may be followed two to three weeks later by joint symptoms and rash (erythema nodosum or erythema multiforme), reminiscent of Reiter’s syndrome. Reactive polyarthritis occurs in 2% of patients with yersiniosis, usually in persons seropositive for HLA-B27. Yersinia antigens can be detected in synovial fluid cells225 and in intestinal mucosal biopsies, and specific IgA antibodies are found in the blood.226 Yersinia bacteremia is a relatively uncommon condition that is seen in patients with underlying diseases such as malignancy, diabetes mellitus, anemia, and liver disease. Metastatic foci can occur in bones, joints, and lungs. The diagnosis of yersiniosis is established by culture of stool or body fluids. Because the organism resembles coliforms and therefore is easily misdiagnosed on the culture plate, the laboratory should be advised that this infection is clinically suspected. Serologic tests have proved useful in Europe and Canada227 but have not provided much help for the cases reported in the United States.228

Treatment

Y. enterocolitica strains are susceptible to several antimicrobial agents, including chloramphenicol, gentamicin, tetracycline, TMP-SMX, and fluoroquinolones, but they are resistant to penicillins and cephalosporins. There is no substantial evidence that antibiotics alter the course of the gastrointestinal infection218,228; indeed, the diagnosis often is established late in the course when the patient is already improving spontaneously. Antibiotics should be used in more severe intestinal infections, particularly those mas-

VIRAL PATHOGENS

ROTAVIRUS

Rotavirus is a group of viruses that was discovered in 1973 in studies of Australian children with diarrhea in whom viral particles were identified in duodenal biopsy specimens by electron microscopy. Rotavirus infection is now recognized to be worldwide in distribution.

Microbiology

The virus measures 70 to 75 nm in diameter, has a doublewalled outer capsid (Fig. 107-9B), and contains segmented double-stranded RNA. The virus has an icosahedral structure resembling the spokes of a wheel, hence “rota.” Highly stable to heat, ether, and mild acids, the virus can be maintained in prolonged storage. Three groups of rotavirus, A, B, and C, cause disease in humans. Group A rotavirus contains four common serotypes, which are the leading pathogens and cause severe gastroenteritis in young children worldwide.229,230 Group B rotavirus is responsible for large outbreaks of diarrhea in children and adults in China but is otherwise a rare isolate. Group C rotavirus infrequently causes disease in various parts of the world.

Pathology and Pathogenesis

Duodenal biopsy specimens of young children with rotavirus infection have demonstrated a patchy abnormality of the upper intestine.231 In its severe form, the infection can produce denuded villi and flattening of the epithelial surface that can persist for three to eight weeks. The morphologic changes are accompanied by physiologic abnormalities such as decreased xylose absorption and reduced brush border levels of disaccharidases. Rotavirus causes diarrhea by elaborating an enterotoxin and by activating the enteric nervous system, both of which stimulate intestinal secretion and diarrhea.232,233

Epidemiology

Rotavirus is responsible for childhood diarrhea in 35% of hospitalized and 10% to 30% of community-based cases.230,234 The virus appears to be spread by the fecal-oral route. In temperate zones, the disease is more common in winter, but in the tropics it is endemic year round. Within a family, often the young child is afflicted with clinical illness and older siblings and adults excrete the virus asymptomatically.

Clinical Features

Rotavirus causes a range of clinical illness from asymptomatic carriage to severe dehydration and death.235 The disease

1869

1870

Section X  Small and Large Intestine Table 107-13  Medical Importance and Epidemiologic and Clinical Features of Human Gastroenteritis Viruses

VIRUS

MEDICAL IMPORTANCE DEMONSTRATED

Rotavirus   Group A

Yes

  Group B

Partially

  Group C

Partially

Calicivirus

Yes

  Norovirus

Yes

  Norwalk-like viruses (small, round structured viruses) Enteric adenovirus

Partially

LABORATORY DIAGNOSTIC TESTS*

EPIDEMIOLOGIC FEATURES

CLINICAL FEATURES

A major cause of endemic severe diarrhea in infants and young children worldwide (in winter in temperate zone) Large outbreaks in adults and children in China Sporadic cases in young children worldwide A cause of diarrhea in children; associated with ingestion of contaminated shellfish and other foods in adults Epidemics of vomiting and diarrhea in older children and adults; occurs in families, communities, and nursing homes; often associated with ingestion of shellfish, other food, or water Similar to characteristics of Norovirus-like illness

Dehydrating diarrhea for 5-7 days; vomiting and fever are common

Immunoassay, electron microscopy, PAGE

Severe watery diarrhea for 3-5 days Similar to features of group A rotavirus Rotavirus-like illness in children; Norovirus-like in adults

Electron microscopy, PAGE

Acute vomiting, diarrhea, fever, myalgias, and headache lasting 1-2 days

Immunoassay, immune electron microscopy

Acute vomiting, diarrhea, fever, myalgias, and headache lasting 1-2 days

Immunoassay, immune electron microscopy

Yes

Endemic diarrhea of infants and young children

Immunoassay, electron microscopy with PAGE

Astrovirus

Yes

A cause of diarrhea in children; reported in nursing homes

Torovirus

Yes

A cause of acute and persistent diarrhea in children; increased risk in immunocompromised children; occurs in community and hospital settings

Prolonged diarrhea lasting 5-12 days; vomiting and fever Watery diarrhea, often lasting 2-3 days, occasionally longer Dehydrating, watery, occasionally bloody diarrhea with vomiting and abdominal pain; usually lasts 5-7 days

Electron microscopy, PAGE Immunoassay, electron microscopy

Immunoassay, electron microscopy Immunoassay, electron microscopy

*Laboratory diagnostic tests, other than those for rotavirus group A, are usually available only in specialized research or diagnostic referral laboratories. Immunoassays are usually enzyme-linked immunosorbent assays or radioimmunoassays. PAGE, polyacrylamide-gel electrophoresis and silver staining of viral nucleic acid in stool. Modified from Blacklow NR, Greenberg HB. Viral gastroenteritis. N Engl J Med 1991; 325:252.

A

occurs principally in children three to 15 months of age; infections continue into the second year of life but are less common thereafter. Mild infections with group A rotaviruses can develop in adults and usually are acquired from a sick child in the household. Vomiting often heralds the illness and is followed shortly by watery diarrhea. The incubation period is one to three days, and the average duration of illness is five to seven days, although some instances of chronic diarrhea have been noted. Excretion of virus for as long as three to eight weeks occurs in approximately one third of infected children.236

Diagnosis

Rapid diagnosis is achieved by detection of rotavirus antigen in the feces with several commercial immunoassays. PCR and nucleic acid probes also are available to detect the virus and identify its serogroups.237,238

B Figure 107-9.  A, Electron micrograph of Norovirus particles from the stool of a volunteer to whom the Norwalk agent was administered. (Bar = 100 nm). B, Human rotavirus particles from the stool of an infant with gastroenteritis. The particles appear to have a double-shelled capsid. Occasional empty particles are seen. (Bar = 100 nm). (A and B courtesy of A. Kapikian, MD. Previously published in Lennete EH, Schmidt NJ. Diagnostic Procedures for Viral, Rickettsial, and Chlamydial Infections. 5th ed. New York: American Public Health Association; 1979. p 933.)

Immunity

After infection with rotavirus, antibody develops in serum and intestinal secretions.239 The antibody is active against the specific serotype but crosses over to other serotypes.240,241 Natural rotavirus infection has a protective efficacy of 93% against recurrent rotavirus disease.241 Levels of antibody in either serum or intestinal fluids do not correlate precisely with protection, a finding that has raised the possibility that cellular immunity is important.

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Infants are protected for at least the first three months of life by maternal antibodies, although there is no firm evidence that breast-feeding provides complete protection against this infection.242 In addition to the possible protective role of maternal antibodies in breast milk, human milk mucin appears to have a potent antirotaviral activity in vitro and in a mouse model of infection.243 The glycoprotein lactadherin binds specifically to all human rotavirus strains and inhibits their infectivity. Higher breast milk concentrations of lactadherin have been associated with protection against symptomatic rotavirus infection.244

Pathology and Pathogenesis The upper small intestine is the focus of attack, and intestinal biopsy specimens reveal patchy mucosal lesions. Because the virus is so small, in contrast to rotavirus, Norovirus particles cannot be observed in electron microscopic sections of intestine. Among the physiologic abnormalities observed during illness are malabsorption of fat and xylose, diminished activity of intestinal disaccharidases, and delayed gastric emptying. Morphologic and physiologic abnormalities reverse within one to two weeks after infection.

Treatment

Clinical Features The disease has a spectrum of symptoms and signs, all mild. In one Norovirus outbreak, diarrhea was noted in 92% of proved cases, nausea in 88%, abdominal cramps in 67%, vomiting in 66%, and muscle aches in 56%.255 Generally, the clinical illness lasts no longer than 24 to 48 hours. Diagnosis can be established by identifying viral antigen in the stool, but such tests currently are available only in research laboratories. The virus also can be seen in fecal effluent by using immune electron microscopy with the aid of serum from a convalescent subject (Fig. 107-9A). A monoclonal antibody–based ELISA and a PCR assay have been developed and can detect Norovirus in stool specimens.256,257

Because loss of fluids and electrolytes appears to be the main pathophysiologic event, rehydration is the mainstay of therapy for this infection. Field studies have established that the ORSs consisting of glucose and electrolytes are effective in restoring fluid balance.245 Rotavirus vaccines are derived from genetic reassortants of various human and animal strains.246 Two reassortant rotavirus vaccines (Rotateq and Rotarix), have been approved and are recommended for all infants except those with immunosuppressive illnesses; they provide moderate protection against all degrees of rotavirus gastroenteritis and excellent protection against severe dehydrating disease. Antirotavirus immunoglobulin of bovine colostral origin has been found to be effective in reducing the duration of rotavirus infection and the amount of oral rehydration therapy required.247

CALICIVIRUS

Caliciviruses (Greek kalyx, the cup of a flower) are singlestranded RNA viruses that are responsible for human and animal infections. Recent molecular studies have reclassified various enteric viruses (e.g., Norovirus, Snow Mountain, Montgomery, Sapporo) and on the basis of genetic composition placed them in the taxonomic family of caliciviruses.248 The typical caliciviruses cause disease mainly in infants and young children.249,250 Calicivirus disease is particularly common in daycare centers, where it accounts for more diarrhea than is attributed to bacterial infections.249 The illness generally is mild and indistinguishable from that of rotavirus or even epidemic Norovirus disease.

Norovirus and Norwalk-like Virus

Norwalk virus, renamed Norovirus, was named for a 1968 outbreak of “winter vomiting disease” in Norwalk, Ohio. The group includes many small viruses, each named after the site of an outbreak of gastroenteritis: Norwalk, Hawaii, Snow Mountain, Montgomery, Taunton, Otofuke, and Sapporo. These small viral agents measure 27 to 35 nm,251 and all contain a single structural protein with singlestranded RNA. This group of viruses is identified as the pathogen in approximately 40% of nonbacterial epidemics of diarrhea in the United States. The virus also has been encountered in Hawaii, England, Australia, and Japan.229 Norovirus causes explosive epidemics of diarrhea that sweep through a community and have a high attack rate. The virus shows no respect for age and preys on virtually all age groups, except infants. Transmission occurs by person-to-person contact, primarily by the fecal-oral route. Raw shellfish also is a major source of infection; during an eight-month period in 1982, 103 outbreaks of Norovirus infection in New York State were related to ingestion of raw clams or oysters.252 Norovirus is a major cause of outbreaks of gastroenteritis in camps, cruise ships, nursing homes, and hospitals.253 It also can contaminate drinking water supplies.254

Immunity Serum antibody titers are low in children, increase in adolescents, and are present in 60% of adults. Volunteer studies have revealed an unusual form of immunity that apparently is not related to formation of antibodies.258 Volunteers who became sick during initial challenge were the same ones who became ill when rechallenged 24 to 42 months later. In contrast, those who resisted the initial challenge also resisted the subsequent challenge. Measurement of antibody in serum and intestinal juice showed higher levels of antibody in the volunteers who became ill on both the initial and subsequent challenges. This antibody had some protective value, albeit short-lived, because early rechallenge at six to 14 weeks after the initial dose produced protection in the subjects with antibody. Protection did not persist, however, because the same group with antibody became ill when rechallenged several months later. Thus, it is postulated that nonimmune mechanisms in the intestine resist infection by this virus; repeated infections produce some protection, which is not permanent. Treatment No specific treatment for Norovirus infection is available. The disease is usually mild, but it can produce dehydration in elderly patients, who may then require hospitalization.

ENTERIC ADENOVIRUS

Most adenoviruses cause upper respiratory infections, but a new group of fastidious strains, known as serotypes 40 and 41, which constitute subgenus F, are responsible for gastroenteritis in children younger than two years259 and in patients with AIDS; 5% to 10% of childhood diarrhea is associated with enteric adenovirus. There is no seasonal occurrence with enteric adenovirus230 and unlike infection with rotavirus or Norovirus, enteric adenovirus infection has a long incubation period of eight to 10 days; the illness can be prolonged for up to two weeks. Nosocomial and daycare center outbreaks are common and associated with high rates of asymptomatic infections.260,261 Adults generally are protected from this infection, although it has been documented to cause diarrhea in persons with AIDS. The virus cannot be

1871

1872

Section X  Small and Large Intestine cultured in available cell lines but can be visualized in stool by electron microscopy or with dot-blot hybridization or immunoassays. An enzyme immunoassay licensed by the U.S. Food and Drug Administration (Adenoclone) is available for detecting enteric adenoviruses in stool specimens.

ASTROVIRUS

Astrovirus is a small, nonenveloped, single-stranded RNA virus similar in structure to calicivirus;262 there are at least seven viral serotypes. In adults, the disease has relatively low infectivity, but in children it is a major cause of diarrheal illness. In a study from Thailand, astrovirus was second only to rotavirus as a cause of diarrhea in children.263 Antibody develops to many of the astrovirus serotypes by four years of age, indicating that they probably cause infections frequently in childhood. Astroviruses are responsible for outbreaks of diarrhea in daycare centers and communities with children younger than 12 months.261,264,265 The disease is characterized by watery or mucoid stools, nausea, vomiting, and occasionally fever, and tends to be milder than rotavirus diarrhea, with less than 6% of children becoming dehydrated.263,265 Coinfections with other pathogens are common, and repeated infections can occur as a result of a lack of crossprotective immunity to the multiple serotypes of astrovirus. The virus can be recognized in stool specimens by electron microscopy, specific immunoassays, RNA probe hybridization, and PCR methodology, although all these tests remain research tools.264,266 Treatment is supportive and emphasizes oral rehydration.

TOROVIRUS

Toroviruses are enveloped single-stranded RNA viruses that cause enteric infections in animals267 and that, in casecontrol studies, have been demonstrated to cause diarrhea in children.268,269 In a large prospective study of pediatric viral diarrhea, toroviruses accounted for 3% of episodes, a greater percentage than that for either the caliciviruses or astrovirus.230 Although most diarrhea from toroviruses occurs in children younger than two years, older children, especially those who are immunocompromised, are at risk for symptomatic infections.269 Toroviruses have been associated with acute diarrhea and persistent (lasting more than 14 days) diarrhea in children.268,269 When torovirus has been encountered in children with persistent diarrhea, it is found often in association with other potential pathogens such as EAEC. Children infected with torovirus have less vomiting and more bloody diarrhea than those with rotavirus infection, although bloody diarrhea occurs in only 11% of patients. Toroviruses can be detected in stool specimens by electron microscopy or ELISA. As in other viral diarrheas, treatment is supportive. Fluid replacement often is required for as long as a week.269

SPECIAL SITUATIONS HOSPITAL-ACQUIRED DIARRHEA

Acute diarrhea developing in the hospitalized patient may be due to infection with C. difficile, a side effect of various medications, ingestion of elixirs containing sorbitol or mannitol, or as a consequence of tube feedings. Antibioticassociated diarrhea, from C. difficile infection, is the most common cause of acute diarrhea in hospitalized patients (see Chapter 108).270,271 Diarrhea can range from mild illness to life-threatening disease from pseudomembranous colitis.

C. difficile infection can follow treatment with almost any type of enteral or parenteral antibiotic. Because C. difficile infection is the single most common cause of diarrhea in hospitalized patients and Salmonella, Shigella, Campylobacter, and parasites are rarely the cause of diarrhea in hospitalized patients, evaluation of patients in this setting with routine stool examinations for enteric pathogens and ova and parasites is unrewarding and inappropriate.

ACUTE DIARRHEA IN PREGNANCY

Diarrhea in the pregnant patient is potentially a serious problem, and infection must be excluded. Appropriate therapy is imperative for the health of both mother and fetus. Dehydration must be avoided because it can be deleterious to placental blood flow. The pregnant woman is at risk for infection with any enteric pathogen and perhaps is predisposed to listeriosis. In addition to the usual bacterial and viral pathogens, diarrhea and abdominal pain in the pregnant patient can result from malaria,272 amebiasis,273 or Giardia.274 In addition to dehydration, certain infectious agents, such as Campylobacter and Shigella species, can cause serious periparturitional complications. Campylobacter infections in the pregnant patient can cause spontaneous abortion, prematurity, neonatal sepsis, neonatal enterocolitis, and death.275 The incidence of complications associated with infectious diarrhea in the pregnant patient is unknown. One case series identified Campylobacter as the most common bacterial enteric pathogen isolated from stool cultures of pregnant women with enteritis. Of nine patients infected with Campylobacter, two with nontyphoidal Salmonella, and one with E. coli O157:H7, one infection resulted in premature birth with neonatal sepsis and death, and one neonate developed C. jejuni enterocolitis.275 Gastrointestinal infections can result in premature labor or spontaneous abortion, probably as a result of bacteremia and placental infection. In other cases, infants can acquire the infection intrapartum by exposure at the introitus of asymptomatic carrier mothers. Shigella sonnei can occur in neonates born to mothers with asymptomatic shigellosis276 and manifests with diarrhea, failure to feed, and gastroenteritis, sometimes preceded by a respiratory illness, two to five days after birth. Septicemia in neonates is more common than in the adult form of this disease. If shigellosis is diagnosed in the pregnant patient, treatment should probably be instituted (although there are no data on which to base this recommendation). Fluoroquinolones or TMP-SMX are not recommended in pregnancy, making ampicillin the choice with least risk to mother and fetus; ceftriaxone also may be used. Infections also can occur from an ascending transcervical route, possibly resulting in a chorioamnio­ nitis.277 Salmonellosis in the pregnant patient also should be treated with antibiotics to protect against placental infection. Treatment during pregnancy is problematic. Drug therapy poses a potential risk to the developing fetus, and such risk must be considered carefully before initiating treatment. Sulfa drugs, fluoroquinolones, metronidazole, and tetracyclines should be avoided. Ampicillin and erythromycin are used widely and apparently are safe. If giardiasis is mild, treatment can be delayed until after delivery; however, the infant should be closely observed for signs of failure to feed and diarrhea because infection of the fetus can occur during delivery.274

TRAVELER’S DIARRHEA

Diarrhea occurring among persons traveling from developed to developing countries is common and affects millions of

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Table 107-14  Relative Frequencies of Microbial Causes of Traveler’s Diarrhea FREQUENCY, % PATHOGEN Enterotoxigenic E. coli Enteroadherent E. coli Campylobacter spp. Shigella spp. Rotavirus Invasive E. coli Salmonella spp. Vibrio spp. Aeromonas spp. Giardia lamblia Entamoeba histolytica Cryptosporidium Cyclospora cayetanensis Hafnia alvei No pathogen identified

Average 40-60 15 10 10 5 <5 <5 <5 <5 <5 <5 <5 <5 <5 40

Range 0-72 NA 0-41 0-30 0-36 0-5 0-15 0-30 0-30 0-6 0-6 NA NA 0-16 22-83

NA, not available.

people each year. Travelers from the United States to Mexico alone number more than 15 million annually, and attack rates in such travelers can be as high as 25% to 50%. Of those afflicted, nearly 30% are ill enough to stay in bed and another 40% must alter their scheduled activities.278 Fortunately, 90% of cases are brief and self-limited, but 5% to 10% of patients develop dysentery, and 1% to 2% of travelers have persistent diarrhea lasting longer than one month. In some cases, diarrhea can last longer than six months and become a chronic illness (see later).

Microbiology

Traveler’s diarrhea is a syndrome, not a specific illness, and is caused mainly by infectious microorganisms that are acquired from food and drink.279 Infectious organisms can be identified in 60% to 80% of cases. Although an array of pathogens has been found, the leading culprits are various forms of E. coli, particularly ETEC and EAEC (Table 107-14). Shigella species can occur in approximately 10% of cases, although the rate of isolation varies from 0% to more than 20%. Traveler’s diarrhea caused by Shigella tends to be more severe than the usual form of traveler’s diarrhea. Strains of Campylobacter are isolated in as many as 41% of cases, with higher rates during cooler seasons. Salmonella organisms are found in less than 5% of cases, although the frequency is higher among travelers to Asia than elsewhere. Rotavirus has been identified in approximately 10% of episodes of traveler’s diarrhea when this pathogen is sought.279,280 Among the parasites, Giardia lamblia and Cryptosporidium are the more common hazards to travelers. The protozoan parasite Cyclospora cayetanensis is responsible for traveler’s diarrhea in visitors to a number of less-developed countries, especially Nepal.281 More than one pathogen may be found in travelers with acute diarrhea (up to 15% in Mexico and 33% in Thailand). To confuse the issue further, no pathogens have been identified, despite careful laboratory study, in approximately 30% to 40% of cases from all parts of the world; a portion of these respond to antibiotics, suggesting that unidentified bacteria may be responsible.

Epidemiology

The risk of developing traveler’s diarrhea depends on the host’s susceptibility, travel and eating habits, length of stay,

and, most important, destination. Estimates of attack rates reveal substantial differences based on destination. Areas of the world having diarrhea attack rates of approximately 40% to 50% include Mexico, South and Central America, the Middle East, Southeast Asia, and Africa. Countries considered to be of moderate risk, with attack rates ranging from 10% to 20%, include most Caribbean islands, China, Japan, Mediterranean countries, eastern Europe, and republics of the former Soviet Union. Areas with attack rates of less than 8% include Canada, northern Europe, Australia, the United States, South Africa, the United Kingdom, Germany, and France.282,283 The national origin of the traveler is another important factor. At an international conference in Mexico held in 1968, participants from the United States and northern Europe had a 36% attack rate, compared with only 8% for their colleagues from developing countries and 2% for local Mexicans. Longer residence in a tropical country also leads to increased resistance to traveler’s diarrhea, although a high risk of diarrhea persists during the first two years of residence.281 The purpose of travel and the style of eating also are important factors that influence the risk of developing this illness. The greatest frequency of diarrhea occurs in people traveling as students or itinerant tourists, the lowest risk is in those visiting relatives, and an intermediate risk exists in business travelers.280 Young travelers, particularly those ages 20 to 29 years, have the highest risk, and rates are lowest in persons older than 55 years. Traveler’s diarrhea is acquired through ingestion of fecally contaminated foods or beverages. Younger travelers, missionaries, and tourists are at higher risk because of exposure to native foods, commercial eating establishments, and contaminated food and water. Travelers eating at private homes have the lowest risk. Dietary discretion is the first line of defense against acquiring traveler’s diarrhea. Especially risky foods include uncooked vegetables, meat, and seafood; foods served steaming hot are the safest. Tap water, ice, unpasteurized milk and dairy products, and unpeeled fruits also are associated with an increased risk. Bottled carbonated beverages (especially flavored beverages), beer, wine, hot coffee, hot tea, and water boiled or appropriately treated with chlorine are relatively safe.

Clinical Features

In most cases, traveler’s diarrhea is self-limited, and evaluation by a physician is not necessary. In cases that are prolonged; that are accompanied by fever, systemic manifestations, or bloody stool; or that occur in immunocompromised persons, evaluation should be done. The presence of dysentery, occult or gross blood in the stool, or fecal leukocytes on microscopic examination should prompt further diagnostic tests. Many laboratories merely culture for Salmonella, Shigella, and Campylobacter, and enteric pathogens, such as Yersinia, ETEC, and EHEC are not sought routinely. If there is concern for a specific pathogen, the laboratory should be alerted to look for it. The disease does not begin immediately after the traveler’s arrival but generally has its onset two to three days later.280 Although most people have 3 to 5 loose stools per day, about 20% can have 6 to 15 watery bowel movements.284 The average duration of illness in untreated subjects is one to three days, but a few unfortunate persons have persistent diarrhea throughout their stay. Watery loose stools are the most common complaint, with an array of associated symptoms (Table 107-15). From 2% to 10% of patients have fever, bloody stools, or both, and they are more likely to have shigellosis. Despite the impressive list of symptoms, less than 1% of travelers are admitted

1873

1874

Section X  Small and Large Intestine Table 107-15  Symptoms in Patients with Traveler’s Diarrhea SYMPTOM Gaseousness Fatigue Cramps Nausea Fever Abdominal pain Anorexia Headache Chills Backache Dizziness Vomiting Malaise Arthralgias

PERCENT AFFECTED 79 74 68 61 56 55 53 39 38 35 34 29 24 23

to a local hospital, and no reports of death from diarrhea were recorded among many thousands of travelers from Switzerland.280 Diarrhea persists in 1% to 3% of travelers for 30 days or longer.284

Treatment

In most cases of traveler’s diarrhea, symptomatic therapy is adequate. As for all forms of diarrhea, treatment entails fluid replacement and appropriate drugs. In most cases of traveler’s diarrhea, severe dehydration seldom is encountered, and fluid losses generally can be replaced with fruit juices, bottled water or carbonated beverages, hot broth, and soups. With severe diarrhea, ORSs should be used (see Table 107-6). Milk and dairy products should be avoided because lactose intolerance is common. Improvement in diarrhea should occur within 48 hours, and if bloody stools or systemic symptoms develop, travelers should seek medical attention. Drug treatment is directed at either suppressing the pathogen with antibiotics or reducing fluid and electrolyte losses with antimotility or antisecretory agents. Antimotility drugs have enjoyed considerable popularity, and their use is supported by good scientific studies.285 Loperamide induces rapid improvement that is demonstrable even on the first day of therapy, with results that are significantly better than those of either placebo or bismuth subsalicylate.285 Bismuth subsalicylate also is an effective alternative treatment for mild to moderate traveler’s diarrhea. The most effective relief has been provided by a combination of an antimicrobial drug and an antimotility drug. In a study of travelers to Mexico, the combined use of loperamide and TMP-SMX curtailed diarrhea in one hour, compared with 30 hours when either drug was used alone or 59 hours with placebo.286 Bismuth subsalicylate has antisecretory, antibacterial, and anti-inflammatory properties and also decreases the number of stools. It should not be used in combination with other antibiotic drugs because it reduces the bioavailability of tetracyclines and may do so with other antibiotics. Table 107-12 lists the antimicrobial therapy of specific enteric pathogens. Several antibiotics have been used successfully to treat traveler’s diarrhea. Ciprofloxacin was as effective as TMPSMX287; results with a single dose of fluoroquinolones are encouraging.288 The development of ciprofloxacin resistance in patients with Campylobacter enteritis has been associated with clinical relapse after treatment and in areas where fluoroquinolone-resistant C. jejuni has become more common, azithromycin, rather than ciprofloxacin, should

be used to treat traveler’s diarrhea.215 Several studies have shown that the nonabsorbable antibiotic rifaximin is as effective as a fluoroquinolone.289,290 Because it is nonabsorbable, side effects are rare. Current recommendations for treatment are as follows: for mild to moderate diarrhea (generally fewer than four bowel movements per day without blood or fever), either loperamide or bismuth subsalicylate can be used effectively. For more-severe diarrhea, optimal therapy seems to be a combination of an antimotility drug and an effective antimicrobial drug, generally a fluroquinolone.291

Prevention

Several approaches to preventing traveler’s diarrhea are available: avoidance of unsafe foods and beverages, use of anti-infective drugs, and use of other medications such as antimotility and antisecretory agents. Precautions about eating habits should be observed so as to prevent not only diarrhea but also other food- and water-borne diseases. Bottled beverages generally are safe, although some epi­ demics have been associated with contaminated bottled drinks.292 Carbonated beverages are safer than noncarbonated ones, due to their low pH (generally 4.0 to 5.0), which has antibacterial properties. Tea or coffee prepared with boiling water generally is safe if it is consumed while it is still hot. Travelers are advised not to eat food from street vendors. The high incidence of bacterial pathogens as a cause of traveler’s diarrhea makes the use of prophylactic antibiotics tempting; however, safe-eating and safe-drinking habits are the traveler’s best methods of prophylaxis. Even in high-risk areas, 50% to 60% of travelers do not get diarrhea. Prophylaxis should be recommended, however, for certain groups of people: those with severe kidney, liver, or heart disease; insulin-dependent diabetes; inflammatory bowel disease; gastrectomy, achlorhydria, or ileostomies; also persons who are taking glucocorticoids or who have immunosuppressive illnesses. In addition, people for whom the trip will be ruined, people whose important business plans will be disrupted, or travelers who will not follow careful dietary practices benefit from prophylaxis. Bismuth subsalicylate (Pepto-Bismol) significantly decreases the risk of developing traveler’s diarrhea and prevents approximately 52% to 65% of the cases of diarrhea in high-risk areas. The most effective dosage is 2.1 g/day, and this agent should be given four times a day (at meals and at bedtime). Whereas prophylactic antibiotic therapy has a better success rate, bismuth subsalicylate has a narrower and safer side-effect profile. Side effects of this drug include transient blackening of the tongue and stool and constipation. The salicylate portion of the molecule is absorbable and can result in toxicity, but this usually does not occur with the short courses given for preventing traveler’s diarrhea. Salicylate toxicity, however, should be considered in the elderly and in those already taking salicylates. Salicylates are best avoided in patients with peptic ulcer disease or chronic renal failure. Table 107-16 lists drugs used to prevent traveler’s diarrhea. Various antibiotic regimens can prevent up to 70% to 80% of the episodes of traveler’s diarrhea but do so at the expense of significant possible side effects. Various antibiotic regimens have been found to be effective. Recent data suggest that rifaximin is an effective agent to prevent traveler’s diarrhea with few side effects. Use of antiparasitic agents or hydroxyquinone, however, has never been shown to be effective and should not be considered for prophylaxis. Side effects of antibiotic therapy vary and include skin rashes, photosensitivity, the rarer Stevens-Johnson

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Table 107-16  Drugs Used to Prevent Traveler’s Diarrhea in Adults DRUG

DOSE*

COMMENTS

Bismuth subsalicylate

Two 262-mg tablets chewed well four times daily (with meals and at bedtime) 100 mg daily 200 mg with meals

Not as effective as antimicrobial drugs but fewer side effects than other agents Resistance is found in many areas of the world Because rifaximin is not absorbed in the intestine, serious side effects are rare Resistance is common in tropical areas

Doxycycline Rifaximin Trimethoprim-sulfamethoxazole Fluoroquinolones   Norfloxacin   Ciprofloxacin   Ofloxacin   Fleroxacin

160/800 mg daily 400 mg 500 mg 300 mg 400 mg

daily daily daily daily

The most predictably effective antimicrobial drugs when susceptibilities are not known

*All drugs should be taken orally beginning on the day of arrival in the country one is visiting and continuing for 1-2 days after returning home, but none should be taken for more than 3 weeks. Modified from DuPont HL, Ericsson CD. Prevention and treatment of traveler’s diarrhea. N Engl J Med 1993; 328:1821.

Table 107-17  Causes of Prolonged Diarrheal Illness after Travel Antibiotic-associated colitis Celiac disease Dietary intolerances Altered diet with osmotic-induced diarrhea Disaccharidase deficiency Inflammatory bowel disease Lymphocytic/collagenous colitis Onset of chronic (presumably viral) enteritis or colitis Persistent bacterial infection Aeromonas Campylobacter Escherichia coli (enteroinvasive) Salmonella Shigella Yersinia Persistent protozoal infection Cryptosporidium hominis Entamoeba histolytica Giardia lamblia Postinfection irritable bowel syndrome Tropical sprue From Chak A, Banwell JG. Traveler’s diarrhea. Gastroenterol Clin North Am 1993; 22:549.

exposure to antibiotics, a stool assay for C. difficile toxin should be requested. After infectious etiologies have been excluded, the possibility of malabsorption or tropical sprue should be considered. An abnormal complete blood count, prothrombin time, or Sudan stain for fecal fat or reduced serum levels of albumin, vitamin B12, folate, or carotene might suggest this diagnosis. If upper gastrointestinal endoscopy is done to search for small bowel pathogens, aspirates and biopsies can be obtained to exclude giardiasis, tropical sprue, or bacterial overgrowth. In addition to tropical sprue, postinfection syndromes developing after an acute diarrheal illness include the unmasking of previously subclinical illnesses such as celiac disease, inflammatory bowel disease, or lymphocytic colitis. A syndrome called postdysenteric colitis might represent a postviral enteritis or colitis or a previously undiagnosed ulcerative colitis; endoscopically this entity appears similar to ulcerative colitis. Despite extensive evaluation, some patients have no identifiable cause for their chronic diarrhea. In such a case, postinfection irritable bowel syndrome may be implicated.159,293,294 Patients should be reassured, treated, and managed similarly to patients with irritable bowel syndrome.

DIARRHEA IN THE ELDERLY

syndrome, antibiotic-associated colitis, and vaginal can­ didiasis. In the individual patient, the broad use of anti­ biotics risks development of antimicrobial resistance, bacterial overgrowth, or persistent salmonellosis. Antidiarrheal agents and lactobacillus preparations are not useful prophylactically.

Chronic Diarrhea

In an unknown but small percentage of patients with traveler’s diarrhea, the diarrhea becomes persistent and chronic. Such patients require evaluation. The differential diagnosis of chronic diarrhea in a patient whose diarrhea began with travel includes chronic infections as a result of bacteria, protozoa, or parasites, as well as postinfection complications. Table 107-17 lists the potential causes of this syndrome.19 Infectious etiologies should be evaluated first. G. lamblia, C. cayetanensis, and, rarely, Shigella, Salmonella, I. belli, or C. jejuni may be responsible for rare cases of persistent diarrhea in travelers.284 In travelers with any

Diarrheal disease is a major cause of morbidity and mortality in adults older than 65 years.295 The elderly are at increased risk for diarrhea and associated complications as a consequence of hypochlorhydria, intestinal motility disorders, underlying chronic medical diseases, immune senescence, and exposures to multiple medications, including antibiotics. A review of national mortality data found that most diarrheal deaths in the United States between 1979 and 1987 occurred in people older than 74 years (51%) followed by deaths in adults aged 55 to 74 years (27%).295 Elderly people living in long-term care facilities are at greater risk for death from diarrhea than are the independent elderly.295,296 Infectious diarrhea was the fourth most common infectious disease in residents of long-term care facilities.297

Microbiology

A number of bacterial, viral, and parasitic agents can cause diarrhea in the elderly. Noninfectious causes of diarrhea also need to be considered, including medications such

1875

1876

Section X  Small and Large Intestine as laxatives or antacids, intestinal tumors that cause obstruction or produce secretory hormones, malabsorption, and systemic illnesses including diabetes mellitus and thyrotoxicosis.296 A consistent association has been noted between advanced age and C. difficile infection.298 Given the frequent need for hospitalization and the use of antimicrobial agents in the elderly, it is not surprising that C. difficile is the most commonly identified cause of diarrhea in the elderly. In addition to age older than 65 years, a white blood cell count greater than 15,000/cm3, rising creatinine, and falling albumin all presage an increased mortality in patients infected with C. difficile.299 Older adults are also more likely than younger people to have a severe event when exposed to EHEC300 and C. perfringens.301 In an outbreak of E. coli O157:H7 in a long-term care facility in Toronto in 1985, one third of the nursing home residents were infected, whereas only 13% of the employees developed diarrhea.115 The elderly residents also were at substantially increased risk for mortality from EHEC in this study; 35% of the infected residents died, primarily from complications of HUS, whereas none of the infected staff members died. Elderly persons represented the only fatalities in a large outbreak of E. coli O157:H7 in Missouri.300 As was demonstrated in Washington state,126 the use of antibiotics for treatment of EHEC infections in the nursing home residents was associated with a three-fold increased risk of death.115 Noroviruses are a common source of nursing home outbreaks and are responsible for substantial morbidity in the institutionalized elderly.302 The elderly also are more likely than younger persons to have severe infections and to die from nontyphoidal salmonellosis and campylobacteriosis. Among the parasitic causes of diarrhea, Cryptosporidium parvum is a cause of morbidity in the elderly. In a Rhode Island hospital, a retrospective chart review of stool studies for C. parvum over a five-year period showed that 36% of positive smears were in elderly patients, whereas 50% were in patients with HIV.303 Most of the elderly with positive stool specimens acquired the infection in an institutional setting, and nearly one half were coinfected with C. difficile.

Clinical Features

Although the clinical manifestations of gastrointestinal infections in the elderly vary by pathogen, there are few major differences between young and old. Noteworthy is that the elderly are more likely to experience complications resulting from volume depletion. Dehydration can exacerbate other age-related complications, such as delirium, electrolyte disturbances, renal insufficiency, malnutrition, and micronutrient deficiencies. If hemorrhagic colitis develops, the new blood loss superimposed on preexisting anemia can precipitate congestive heart failure, angina, or myocardial infarction.

Diagnosis

Because many episodes of infectious diarrhea in the elderly are self-limited, supportive therapy often is all that is necessary. Signs of dehydration that are useful for evaluating young adults and children are often less reliable indicators of an elderly patient’s hydration status because older adults often have decreased skin elasticity, dry oral mucosa because of mouth breathing, and sunken eyes. Although it is important to take orthostatic vital signs, orthostatic changes generally do not occur until there has been sub­ stantial volume loss (more than 10%). Laboratory tests also are not especially helpful for diagnosing dehydration in the elderly. Although the ratio of blood urea nitrogen

to creatinine, if elevated, suggests dehydration, this ratio is only a crude indicator of a patient’s underlying volume status. Noninfectious causes of diarrhea, such as use of magnesium-containing antacids, laxatives, stool softeners, and dietary supplements, should be eliminated before embarking on a diagnostic evaluation; fecal impaction also should be excluded. Although stool cultures often are not helpful, except in outbreaks, stool should be evaluated for C. difficile toxin, especially if the patient recently received chemotherapy or antimicrobial treatment. Endoscopic procedures should be done if fecal occult blood tests are positive and invasive pathogens have been excluded or if symptoms are persistent or recurrent.

Treatment

If an elderly patient with diarrhea and dehydration is not severely dehydrated and is able to tolerate fluids by mouth, oral rehydration therapy is preferred over intravenous treatment. Patients who are severely dehydrated or unable to tolerate oral therapy should be rehydrated by the parenteral route. Antimicrobial therapy generally should be reserved for the treatment of specific infections such as shigellosis, invasive salmonellosis, or C. difficile.

DYSENTERY VERSUS ULCERATIVE COLITIS The clinical distinction between bacillary or amebic dysentery and ulcerative colitis can be difficult. Two features distinguish dysentery, a condition also known as acute selflimited colitis, from an acute attack of idiopathic ulcerative colitis: a pathogen revealed on stool culture or microscopic analysis and a self-limited course without relapse. Positive bacteriologic culture is encountered in only 40% to 60% of reported cases.52,150 Histopathologic examination of colonic mucosa obtained by endoscopic biopsy can be helpful. Both the microbial form (dysentery) and the idiopathic form of acute colitis show edema, neutrophils throughout the lamina propria, and superficial cryptitis with preservation of the normal crypt pattern. Idiopathic ulcerative colitis, however, shows signs of chronicity such as crypt distortion and lymphoplasmacytosis in the lamina propria, which typically involves the base of the mucosa. Focal cryptitis, crypt abscesses, and a mild increase in the cellularity of the lamina propria are found in both the microbial and idiopathic forms and can lead to confusion.150 In clinical practice, the main diagnostic problem is illustrated by a patient who has had severe acute colitis for several days and who has not responded to antibiotic therapy. Presumptive treatment should include a fluoro­ quinolone for bacterial pathogens and metronidazole for protozoa. The decision to use other treatment, such as glucocorticoids, rests on the distinction between these diseases, and it may not be possible to make this distinction based on just culture or histopathologic findings. Certain pathogens involve principally the lower small bowel but can invade the colon as well; Salmonella and Yersinia make up this group. Although watery diarrhea is the usual manifestation, depending on the focus of infection, the spectrum extends from dehydrating diarrhea to frank colitis. Vibrios produce varying clinical manifestations, apparently related to the virulence factors in each infecting strain. E. histolytica attacks the large bowel and produces an invasive disease. Curiously, there is a paucity of PMNs in the stool of a patient with amebiasis.

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning TUBERCULOSIS OF THE GASTROINTESTINAL TRACT Any region of the gastrointestinal tract can be involved with tuberculosis, and such involvement still is prevalent in developing countries where tuberculosis is a common health problem. In recent years there has been an upsurge of gastrointestinal tuberculosis in the United States as a result of the influx of immigrants and the AIDS epidemic.

PATHOGENESIS

Mycobacterium tuberculosis is the pathogen responsible for most cases of intestinal tuberculosis. In some parts of the world, Mycobacterium bovis, an organism found in dairy products, still is responsible for some cases; however, M. bovis is an uncommon human pathogen in Western countries. The usual route of infection is direct penetration of the intestinal mucosa by swallowed organisms. In the past, intestinal tuberculosis was associated with active pulmonary infection and especially with active laryngeal involvement. Autopsies of patients with pulmonary tuberculosis, before the era of effective treatment, demonstrated intestinal involvement in 55% to 90% of cases. There also was a higher risk of intestinal involvement in patients with pulmonary cavitation and positive sputum smears, again reflecting the risk of a high inoculum of swallowed organisms. In modern series, pulmonary involvement is seen in less than 50% of patients with intestinal tuberculosis.304,305 Indeed, the chest film is unremarkable in most patients now seen with intestinal tuberculosis.

CLASSIFICATION AND DISTRIBUTION OF DISEASE

The ileum and cecum are the most common sites of intestinal involvement and are affected in 75% of cases. Both sides of the ileocecal valve usually are involved, leading to incompetence of the valve, a finding that helps distinguish tuberculosis from Crohn’s disease. Other locations of involvement, in order of descending frequency, are the ascending colon, jejunum, appendix, duodenum, stomach, esophagus, sigmoid colon, and rectum. Multiple areas of the bowel can be affected.

PATHOLOGY

The gross appearance of intestinal tuberculosis has been divided into the following three categories304,305: Ulcerative lesions are seen in 60% of patients and consist of multiple superficial lesions confined largely to the epithelial surface. The process is highly virulent and in the past was associated with a high mortality rate. Hypertrophic lesions occur in 10% of patients and manifest as scarring, fibrosis, and heaped-up mass lesions that can mimic carcinoma. Ulcerohypertrophic lesions are seen in 30% of patients and, in this type, mucosal ulcerations are combined with healing and scar formation. At surgery, the bowel wall appears thickened, and there typically is an inflammatory mass surrounding the ileocecal region. Active inflammation is apparent, as are strictures and even fistulas. The serosal surface is covered with multiple tubercles. The mesenteric lymph nodes typically are enlarged and thickened. The mucosa itself is hyperemic, cobblestoned, edematous, and, in some cases, ulcerated. In contrast to Crohn’s disease, the ulcers tend to be circumferential and perpendicular to the longitudinal axis of the bowel. When these ulcers heal, the associated fibrosis causes stricture and stenosis of the lumen.

Figure 107-10.  Photomicrograph of the colon of a patient with colonic tuberculosis, showing granulomas in the mucosa and submucosa. (Hematoxylin and eosin.)

Histologically, the distinguishing lesion is a granuloma (Fig. 107-10). Caseation is not always seen, especially in the mucosa, although caseating granulomas are found with regularity in regional lymph nodes. The muscularis usually is spared. With the Ziehl-Neelsen stain, sections of involved intestine show acid-fast bacilli in approximately one third of patients. The organism also can be recovered by culture of the involved tissues.

CLINICAL FEATURES

The most common complaint is nonspecific chronic abdominal pain, reported in 80% to 90% of patients. Weight loss, fever, diarrhea or constipation, and blood in the stool may be present.304-306 An abdominal mass, usually deep and posterior in the right lower quadrant of the abdomen, is found in approximately two thirds of patients. Laboratory findings include mild anemia with a normal white blood cell count. Complications include intestinal hemorrhage, perforation, obstruction, fistula formation, and malabsorption.305,307 Perforation is uncommon but can occur even during treatment. Intestinal obstruction is a more common finding and typically results from segmental, stenotic disease. Surgical intervention may be required to relieve obstruction despite appropriate drug therapy. Malabsorption can occur when obstruction leads to proximal bacterial overgrowth.

DIAGNOSIS

A presumptive diagnosis of intestinal tuberculosis can be established in a patient with active pulmonary tuberculosis and radiologic and clinical findings that suggest intestinal involvement. Colonoscopic findings, although nonspecific, consist of superficial areas of ulceration and a nodular friable mucosa.306,308 Linear or circumferential ulcers that are oriented perpendicular to the longitudinal axis of the bowel more strongly suggest the diagnosis. Definitive diagnosis is made by identification of the organism in tissue by direct visualization with an acid-fast stain, by culture of resected tissue, or by a PCR assay.308 The tuberculin skin test is not

1877

1878

Section X  Small and Large Intestine TREATMENT

Figure 107-11.  Film from a barium enema of a patient with colonic tuberculosis showing extensive involvement of the cecum, ascending, and transverse colon. The ulcerated, narrowed ahaustral appearance is typical of colitis. (Courtesy of H. I. Goldberg, MD, San Francisco, Calif.)

very helpful, because a positive test does not necessarily mean active disease. In addition, many patients, especially older persons with weight loss and inanition and those with AIDS, have a negative skin test in the face of active intestinal tuberculosis. Radiologic examination of the bowel reveals a thickened bowel wall with distortion of the mucosal folds, ulcerations, varying degrees of bowel stenosis, and formation of pseudopolyps (Fig. 107-11).309,310 Computed tomography can show thickening of the ileocecal valve and medial wall of the cecum, extending to the terminal ileum, and massive lymphadenopathy with central necrosis.309 The cecum is contracted with disease on both sides of the valve, and the valve itself often is distorted and incompetent; conification of the cecum, as seen on barium enema, is characteristic of tuberculosis and is referred to as Stierlin’s sign. Tuberculosis tends to involve short segments of the intestine with stenosis and fistula formation. In the hypertrophic form, a mass can be seen that may resemble a cecal carcinoma. Calcified mesenteric lymph nodes and an abnormal chest film are other findings that aid in the diagnosis of intestinal tuberculosis. Several diseases can resemble intestinal tuberculosis. Crohn’s disease can give virtually all of the changes of intestinal tuberculosis, except for the presence of the organism. Y. enterocolitica can produce mesenteric adenopathy, ulcerations, and thickening of the bowel mucosa. Usually, this infection has a shorter history and resolves spontaneously. Involvement of the cecum with carcinoma, lymphoma, amebiasis, syphilis, and lymphogranuloma venereum also can resemble tuberculosis. Colonoscopy with biopsy is critical in differentiating mimicking disorders.

Standard antituberculosis treatment gives a high cure rate for intestinal tuberculosis. There are no controlled studies to determine the optimal therapy or duration, but extrapolation from other forms of extrapulmonary tuberculosis suggests that a three-drug regimen for a period of 12 months is adequate treatment, using isoniazid (300 mg/day), pyrazinamide (15 to 30 mg/kg/day), and rifampin (600 mg/day). In AIDS patients, either ethambutol (15 mg/kg) or streptomycin (15 mg/kg) should be added. The course of intestinal tuberculosis in AIDS is more prolonged and can require treatment with second-line drugs because of the high prevalence of resistant organisms. In the past, surgical intervention often was required for intestinal tuberculosis, especially in cases involving the ileocecal region.311,312 Obstruction and fistulas were the leading indications for surgery. In the current era, most fistulas and ulcerative complications respond to medical management, but mass lesions associated with the hypertrophic form still can necessitate an operative approach, because they can lead to luminal compromise with complete obstruction.304,307 Surgery also may be necessary when free perforation, confined perforation with abscess formation, or massive hemorrhage occur. Because of its similarity to carcinoma of the cecum, undiagnosed ileocolonic tuberculous disease can prompt exploratory laparotomy and right hemicolectomy, although colonoscopy has reduced the need for diagnostic laparotomy. Less-dramatic resections are required for tuberculous disease because the condition often improves significantly with appropriate drug therapy. Postoperative complications are more common in patients with than without concomitant active pulmonary tuberculosis312 and include obstruction, enterocutaneous fistula, perforation, wound infection, and bleeding.304,312

BACTERIAL FOOD POISONING Food poisoning is defined as an illness caused by the consumption of food contaminated with bacteria, bacterial toxins, parasites (e.g., trichinosis), viruses (e.g., hepatitis), or chemicals (e.g., amanitin with ingestion of mushrooms). Food poisoning caused by bacteria constitutes 75% of the outbreaks and 86% of cases in the United States for which an etiology can be determined313,314; however, only 42% of such outbreaks fulfill the microbiologic standards for a confirmed etiology. The Centers for Disease Control and Prevention (CDC) recorded 2751 food-borne outbreaks occurring from 1993 to 1997 that affected 86,000 persons314; however, unreported cases and unidentified microorganisms probably increase the incidence by 10- to 100-fold. A food-borne disease outbreak is defined by two criteria: similar illness, usually gastrointestinal, in two or more persons; and epidemiologic or laboratory investigation that implicates food as the source. An extensive list of agents has been associated with food-borne illness syndromes (Table 107-18).313-315 The major recognized causes of bacterial food poisoning are C. perfringens; S. aureus; Vibrios, including V. cholerae and V. parahaemolyticus; Bacillus cereus; Salmonella; Clostridium botulinum; Shigella; E. coli O157:H7; and certain species of Campylobacter, Yersinia, Listeria, and Aeromonas. Other bacteria, such as group A Streptococcus and Listeria monocytogenes, have been implicated in some outbreaks (see Table 107-18). Salmonella outbreaks predominate and constitute more than half of confirmed cases of food-borne illness, in part because of its ease of recognition and general awareness of

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Table 107-18  Estimates of Rates of Food-Borne Illnesses and Associated Mortality in the United States, 1999 PATHOGEN Bacteria Brucella Campylobacter Escherichia coli   O157:H7 (EHEC)   Non-O157:H7 (non-EHEC) Listeria monocytogenes   Salmonella typhi   Nontyphoidal Salmonella Shigella Noncholera Vibrio   Vibrio vulnificus   Yersinia enterocolitica Toxins Bacillus cereus Clostridium botulinum (food botulism) Clostridium perfringens Food Poisoning   Staphylococcal   Streptococcal Parasites Cryptosporidium parvum Cyclospora cayetanensis Giardia lamblia Toxoplasma gondii Trichinella spiralis Viruses Astrovirus Norovirus and Norwalk-like viruses Rotavirus Hepatitis A virus Total

ESTIMATED TOTAL NO. OF CASES

FOOD-BORNE TRANSMISSION, %

NO. OF DEATHS

CASE-FATALITY RATE

1,554 2,453,926

50 80

11 124

0.0071 0.0000

73,480 36,740 2,518 824 1,412,498 448,240 7,880 94 96,368

85 85 99 80 95 20 65 50 90

61 30 504 3 582 70 20 37 3

0.0008 0.0008 0.2001 0.0036 0.0004 0.0002 0.0025 0.3936 0.0000

27,360 58 248,520

100 100 100

0 4 7

0.0000 0.0690 0.0000

185,060 50,920

100 100

2 0

0.0000 0.0000

300,000 16,264 2,000,000 225,000 52

10 90 10 50 100

66 0 10 750 0

0.0002 0.0000 0.0000 0.0033 0.0000

3,900,000 23,000,000 3,900,000 83,391 38,629,641

1 40 1 5 —

10 310 30 83 1,809

0.0000 0.0000 0.0000 0.0010 0.0000

EHEC, enterohemorrhagic Escherichia coli. From Mead PS, Slutsker L, Dietz V, et al. Food-related illness and death in the United States. Emerg Infect Dis 1999; 5:607.

physicians and the public about the organism. E. coli is the next most common cause of food-borne outbreaks, followed closely by C. perfringens, S. aureus, and Shigella. Several pathogens rarely are reported, namely B. cereus and V. parahaemolyticus, but they have have been well studied in certain parts of the world. Their contribution to food-borne diarrheal illness in the United States has been recognized only recently, and their recovery from stool or food requires special laboratory procedures.

Approach to the Patient

A thorough history should provide clues to the etiology of the food-borne illness (Table 107-19). Details to be elicited should include the food ingested (Table 107-20); the time period between ingestion and onset of symptoms; the number of people who ingested the food and how many became ill; and the means of preparation and storage of the suspected food (e.g., picnic, home canning, restaurant). Some food-borne illnesses are more common during certain seasons.313,314 For example, during the summer months, illnesses due to bacteria such as Salmonella, Shigella, and S. aureus are prevalent. Illnesses due to C. jejuni are more common in the spring and fall. C. perfringens outbreaks occur least often in the summer. Infections due to B. cereus and Norovirus occur year round. In addition to considering the organism and its vector, one must also take into account the susceptibility of the host. Persons with liver disease have an annual rate of illness from V. vulnificus 80 times greater, and a death rate

200 times greater, than those of adults without liver disease.316,317 Patients with compromised immune symptoms such as chronic renal insufficiency, malignancy, diabetes, or iron overload states and patients taking glucocorticoids also probably are at increased risk for infection and death.317 Gastric acidity, as discussed earlier, is a natural defense mechanism against infection that may be compromised by prior gastric surgery or use of proton pump inhibitors. The presenting symptom complex can give a clue to the etiologic organism. Symptom complexes may be classified as nausea and vomiting, noninflammatory diarrhea, inflammatory diarrhea, neurologic symptoms, and systemic or miscellaneous symptoms.313 This section deals with S. aureus, C. perfringens, Listeria monocytogenes, B. cereus, and botulism; the other bacterial agents are discussed in previous sections. Table 107-19 lists the characteristics of the more common types of bacterial food poisoning.

STAPHYLOCOCCUS AUREUS

Coagulase-positive S. aureus is a common cause of food poisoning in the United States; before 1973, it was the leading cause.

Microbiology

Five immunologically distinct enterotoxins have been associated with food-poisoning strains of S. aureus. These enterotoxins, termed A, B, C, D, and E, are heat-resistant polypeptides. When they are tested in a rat intestinal loop

1879

Fried rice, vanilla sauce, cream, meatballs, boiled beef, barbecued chicken Milk, chicken, beef Beef, turkey, chicken

Salads, beef

Milk, raw vegetables, cole slaw, dairy products, poultry, beef Eggs, meat, poultry

Milk, salads (potato, tuna, turkey)

Ham, pork, canned beef, cream-filled pastry

Seafood (rarely saltwater) or salted vegetables Chocolate milk or raw milk, pork

Bacillus cereus

Campylobacter jejuni Clostridium perfringens

Escherichia coli spp.

Listeria monocytogenes

Shigella spp.

Staphylococcus aureus

Vibrio parahaemolyticus 72 (2-144)

12 (2-48)

3 (1-6)

24 (7-168)

24 (5-72)

?

24 (8-44) 96 (24-120)

48 (24-240) 12 (8-22)

2 (1-16) 9 (6-14)

Heat stable

Role of toxin unclear

Heat stable

Role of toxin unclear

Role of toxin unclear

Unknown

Heat labile Heat stable Verotoxin

Unknown Heat labile

Heat stable Heat labile

PRIMARY TOXIN

F, C, D, V, pharyngitis, arthritis, mesenteric adenitis, rash

D, C, N, V, H, F, B (rare)

V, N, C, D, F (rare)

C, F, D, B, H, N, V

D, C, N, V, F, H, B (rare), enteric fever

D, C, N, H, F, M F, M, D, C B, C, F, hemolyticuremic syndrome D, F, C, N, V, B

D, F, C, B, H, M, N, V D, C (N, V, F rare)

V, C, D D, C, V

CLINICAL FEATURES

7 (2-30)

3 (2-10)

1 (0.3-1.5)

3 (0.5-14)

3 (0.5-14)

?

3 (1-4)

7 (2-30) 1 (0.3-3)

0.4 (0.2-0.5) 1 (1-2)

MEDIAN DURATION, DAYS (RANGE)

B, bloody diarrhea; C, cramping abdominal pain; D, diarrhea; F, fever; H, headache; M, myalgia; N, nausea; NA, not available; V, vomiting. From Snydman DR. Food poisoning. In Gorbach SL, Bartlett JG, Blacklow NR, editors. Infectious Diseases. Philadelphia: WB Saunders; 1992. p 771.

Yersinia enterocolitica

Salmonella spp.

COMMON VEHICLES

ORGANISM

MEDIAN INCUBATION (HR) (RANGE)

Table 107-19  Features of Bacterial Food Poisoning

20

0

0

40-60

30-50

10

0

25 0

0

SECONDARY ATTACK RATE, %

Stool, rectal swab from patients and food preparers; raw food Stool, rectal swab from patients, and food preparers; implicated food Stool, vomitus; food or food-contact surfaces; nose, hands, purulent lesion on food preparer Stool, rectal swab; food, food-contact surfaces; seawater Stool from food preparer

Stool, rectal swab

Stool, rectal swab Stool, rectal swab; food, food-contact surfaces Stool, rectal swab

Vomitus, stool, implicated food

SOURCES OF DIAGNOSTIC MATERIAL

1880 Section X  Small and Large Intestine

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Table 107-20  Organisms and Food-Borne Diseases Associated with Specific Foods or Beverages Beef and Pork Salmonella spp. Staphylococcus aureus Clostridium perfringens EHEC Bacillus cereus Yersinia enterocolitica Listeria monocytogenes Brucella spp. Trichinella spiralis Chinese Food Bacillus cereus (in fried rice) Monosodium glutamate poisoning Eggs Salmonella spp. Staphylococcus aureus Fish Clostridium botulinum Ciguatera poisoning Scombroid poisoning Diphyllobothrium latum Anisakiasis Honey Clostridium botulinum Milk and Cheese Salmonella spp. Campylobacter spp. EIEC and EHEC Yersinia enterocolitica Group A streptococci Brucella spp. Listeria monocytogenes Poultry Salmonella spp. Staphylococcus aureus Campylobacter Clostridium perfringens Listeria monocytogenes Shellfish Vibrio parahaemolyticus Vibrio cholerae (O1 and non-O1) Hepatitis A Norovirus and Norwalk-like viruses Paralytic shellfish poisoning Neurotoxic shellfish poisoning Vegetables Clostridium botulinum Salmonella spp. Shigella spp. Bacillus cereus Norovirus EHEC, enterohemorrhagic Escherichia coli; EIEC, enteroinvasive E. coli. From Bishai WR, Sears CL. Food poisoning syndromes. Gastroenterol Clin North Am 1993; 22:579.

model, net secretion of water and electrolytes is observed.318 Staphylococcal enterotoxins induce vomiting when fed to monkeys or human volunteers.

Epidemiology

Staphylococcal food poisoning has a short incubation period of about three hours, with a range of one to six hours. The disease usually is clustered within a family or group and has a high attack rate. Many foods have been implicated in this form of food poisoning; however, foods with a highsalt concentration (e.g., ham or canned meat) or a high-sugar content (e.g., custard and cream) selectively favor the growth of staphylococci. The major mode of transmission is from a food handler to the food product. Involved foods usually

have been cut, sliced, grated, mixed, or ground by workers who are carriers of toxin-producing strains of S. aureus.

Pathogenic Mechanisms

Three requisites for staphylococcal food poisoning are contamination of a food with enterotoxin-producing staphylococci, suitable growth requirements of the food for the organism, and suitable time and temperature for the organism to multiply. The emetic dose of enterotoxin A or B for humans has been estimated to be between 1 and 25 µg.

Clinical Features

Symptoms of staphylococcal food poisoning are primarily profuse vomiting, nausea, and abdominal cramps, often followed by diarrhea occurring one to four hours after ingestion of a suspect meal. Vomiting is the dominant initial symptom and can lead to a severe metabolic alkalosis. Fever is unusual. Rarely, hypotension and marked prostration occur. Fatalities are unusual, and recovery is complete within 24 to 48 hours. Diagnosis is made based on the typical presentation occurring a few hours following the ingestion of typical foods. Most people with staphylococcal food poisoning do not consult a physician. More-severe cases can require supportive care, particularly rehydration and correction of alkalosis. No specific therapy is available.

CLOSTRIDIUM PERFRINGENS

C. perfringens is a major food-borne pathogen that produces vomiting and diarrhea. The disease is caused by an enterotoxin elaborated by strains of C. perfringens type A. A moresevere and often fatal food-borne illness, known variously as enteritis necroticans (Darmbrand) and pigbel, is caused by C. perfringens type C (see later).

Microbiology

Clostridia are Gram-positive, spore-forming, obligate anaerobes that can be found in the normal intestinal flora of humans and animals and in the soil. Although an anaerobe, C. perfringens is remarkably aerotolerant and survives exposure to oxygen for as long as 72 hours. C. perfringens produces 12 toxins that are mostly active in tissues, as well as several enterotoxins. The food-poisoning syndrome is caused by a heat-labile protein enterotoxin, better termed a secretory cytotoxin, which is a structural component of the spore coat and is formed during sporulation; like other enterotoxins, it causes fluid accumulation in the rabbit ileal loop model, presumably by altering membrane ion permeability.319,320 Unlike cholera and E. coli enterotoxins, clostridial enterotoxin has its maximum activity in the ileum, inhibits glucose transport, damages the intestinal epithelium, and causes protein loss into the intestinal lumen.320

Epidemiology and Pathogenic Mechanisms

Epidemics of C. perfringens are characterized by high attack rates and a large number of affected persons, usually 40 to 50 per outbreak. The incubation period varies from eight to 14 hours but can be as long as 22 hours. In almost every outbreak of clostridial food poisoning, roasted, boiled, stewed, or steamed meats or poultry is the vehicle of infection. Usually, the meat is cooked in bulk so that heat gain and internal pressure are insufficient to kill the spores. The implicated food invariably undergoes a period of inadequate cooling, which allows the spores to germinate. The organism proliferates rapidly at temperatures between 15°C and 50°C. Unless the food is reheated to a very high temperature, it will contain many viable organisms.

1881

1882

Section X  Small and Large Intestine Clinical Features

C. perfringens food poisoning is characterized by watery diarrhea, severe cramping abdominal pain, and, often, vomiting that begins eight to 24 hours after the incriminating meal. Fever, chills, headache, or other signs of infection usually are absent. The illness is of short duration, usually less than 24 hours. Rare fatalities have been recorded in debilitated or hospitalized patients. No specific treatment is required for this illness.

Enteritis Necroticans

Enteritis necroticans is a disease that originally was described in post–World War II Germany, in an outbreak affecting more than 400 people who consumed rancid meat. Similar outbreaks, associated with the consumption of poorly cooked pork, have been described in Papua New Guinea and are labeled pigbel.321,322 The disease is caused by strains of C. perfringens type C. A beta toxin is produced that in malnourished patients, especially children, cannot be inactivated by the usual intestinal proteases and causes transmural intestinal wall necrosis; protease inactivation, such as consuming large amounts of sweet potatoes, a food staple in Papua New Guinea, can facilitate the damage. Intestinal perforation, sepsis, and hemorrhage result in a 40% mortality rate. Fortunately, this disease is rare. In the uncomplicated case, treatment is symptomatic and supportive. Outbreaks of pigbel have been related to consumption of pig in large native feasts. The pig is improperly cooked, and large quantities are consumed over three or four days. Other cases, most often in children younger than 10 years, occur in villages remote from the site of the cooking. Enteritis necroticans associated with the consumption of chitterlings is encountered rarely in the United States.322

LISTERIA SPECIES

Listeria are Gram-positive highly motile bacilli that are relatively heat resistant. They have been isolated from the intestinal tracts of humans and animals and from sewage and well water. Cases can occur as part of an outbreak or on a sporadic basis. In reported epidemics, the vehicles of infection have been raw and pasteurized milk, soft cheeses, cole slaw, shrimp, rice salad, pork dishes, and raw vegetables.323,324 Listeria can be cultured from raw poultry, beef, or pork; prepackaged meat products; cheeses; and raw vegetables.325 Listeriosis usually is a systemic disease associated with bacteremia that can seed the meninges, heart valves, or body organs. Intestinal symptoms such as diarrhea and cramping often precede fever and bacteremia. Immunocompetent hosts occasionally develop gastroenteritis characterized by fever, headache, abdominal pain, nausea, and diarrhea; this form of listeriosis usually is not complicated by bacteremia.326 Among the food-borne pathogens, Listeria have been associated with the highest mortality rates; 70 deaths were reported from 1983 to 1987, for a case-fatality rate of 27%. Neurologic sequelae can occur in a sizeable proportion of survivors of central nervous system listeriosis. The propensity of the organism to attack immunosuppressed persons and pregnant women might account for the severity of the infection. The reason for this propensity is not known.

BACILLUS CEREUS

B. cereus is an aerobic, spore-forming, Gram-positive rod that has been associated with two clinical types of food poisoning: a diarrhea syndrome and a vomiting syndrome.327 The organisms responsible for the two syndromes produce different toxins and have different epidemiologies.

Diarrhea Syndrome

Diarrhea results from an enterotoxin that causes intestinal secretion by activation of adenylate cyclase in intestinal epithelial cells, similar to the action of cholera toxin. The median incubation period appears to be nine hours, with a range of six to 14 hours. The clinical illness is characterized by diarrhea (96%), generalized cramps (75%), and vomiting (23%).327 Fever is uncommon. The duration of illness ranges from 20 to 36 hours, with a median of 24 hours. The original report of B. cereus as a cause of diarrheal disease was associated with consumption of contaminated meatballs, but strains of B. cereus associated with diarrhea have been found in approximately 25% of many foods sampled, including cream, pudding, meat, spices, dried potatoes, dried milk, vanilla sauces, and spaghetti sauces, all of which are contaminated before cooking.328 If the food is prepared so that the temperature is maintained at 30°C to 50°C, vegetative growth is permitted. Spores can survive extreme temperatures, and when allowed to cool relatively slowly, they germinate, multiply, and elaborate toxin. There is no evidence that human carriage of the organism or other means of contamination plays a role in transmission. Whether the diarrhea-causing heat-labile enterotoxin actually is ingested or produced in vivo is not known; however, the incubation of diarrheal illness is too long for preformed toxin, and a large inoculum (106) is required to cause illness, observations that suggest that intestinal colonization is required. Usually, no therapy is required because of the short duration of symptoms.

Vomiting Syndrome

Although the organism associated with the vomiting disease appears to be the same as the one causing diarrhea, a different type of toxin has been implicated.329 The emetic syndrome results from ingestion of a preformed enterotoxin that is stable to heat. The emetic syndrome is more common than the diarrheal syndrome. When fed to rhesus monkeys, cell-free culture filtrates from these strains do not produce intestinal secretion and diarrhea, but they produce vomiting. The vomiting syndrome has a short incubation period of approximately two hours. Virtually all affected persons have vomiting and abdominal cramps. Diarrhea is present in only one third. The duration of illness ranges from eight to 10 hours, with a median of nine hours; the illness usually is mild and self-limited, so no specific therapy is required. Nearly all reported cases involving the vomiting toxin have implicated fried rice as the vehicle.327 In England, almost 90% of uncooked rice was found to be colonized by B. cereus, although the number of organisms was relatively low.330 The disease has been ascribed to the common practice in Chinese restaurants of allowing large portions of boiled rice to drain unrefrigerated to avoid clumping. Flash-frying during the final preparation of the fried rice does not produce enough heat to destroy apparently preformed heat-stable toxin. It appears that the emetic illness is caused by preformed toxin, because the incubation period is short, and there is an extremely high attack rate, approaching 100% in outbreaks.

BOTULISM

Botulism is a rare food-borne disease resulting from exposure to neurotoxins secreted by strains of Clostridium botulinum. Between 1993 and 1997, there were 13 outbreaks of botulism in the United States, accounting for 56 cases and one death.313,314 Although food-borne botulism is relatively uncommon, it is the most lethal of all the bacterial toxinmediated food-borne diseases and the only one for which specific effective therapy is available.

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Epidemiology

During the past few decades, food-borne botulism has become the least common form of botulism, trailing after wound and infant botulism. Food-borne botulism develops after the ingestion of preformed toxin in improperly preserved canned vegetables, salsas, meats, and fish. A disproportionate number of recent cases have occurred in the Pacific Northwest and Alaska, and cases have been associated with Native American foods such as whale or seal that have been fermented or preserved with traditional methods. Outbreaks in the United States have been associated with baked potatoes, cheese sauce, beef stew, and garlic cooking oil.315 Infant botulism develops in infants whose gastrointestinal tract becomes colonized with live C. botulinum bacteria, which secrete small amounts of botulinum toxin. Absorption of low concentrations of the toxin leads to lethargy, poor feeding, constipation, diminished muscle tone, and a weak cry. The source of the botulinum toxin is not clear; household dust, soil, and honey in feedings have been suggested as possible sources. It is recommended that honey not be given to infants.

Pathogenic Mechanisms

C. botulinum and closely related species of Clostridia produce heat-resistant spores that are capable of surviving food preservation techniques that destroy nonsporulating organisms. The seven serologically distinct botulinum toxins are designated with the letters A to G. Neutralization by type-specific serologic reagents is used to differentiate the serotypes. Types A, B, and E are responsible for most human cases of botulism.331 Neurotoxin-producing strains of Clostridium butyricum and Clostridium baratii are less commonly responsible for human botulism. Toxin production occurs in the presence of anaerobic, low-solute, and low-acid conditions. C. botulinum usually is unable to replicate in the mature human intestine, although the toxin is acid stable and easily traverses the stomach intact. After absorption, botulinum toxin binds irreversibly to presynaptic cholinergic nerve endings of the cranial and peripheral nerves, thereby resulting in inhibition of the release of acetylcholine and the characteristic clinical syndrome that results from blockade of voluntary motor and autonomic cholinergic junctions.

Clinical Features

Ingestion of botulinum toxin initially results in gastrointestinal symptoms, including nausea, vomiting, abdominal pain, and diarrhea, usually within 18 to 36 hours after toxin ingestion.315 Once neurologic symptoms develop, constipation is common. Dry mouth, diplopia, and blurred vision are followed by dysarthria, dysphonia, dysphagia, and peripheral muscle weakness. The typical symmetrical descending paralysis starts with the cranial nerves and then affects the upper extremities, respiratory muscles, and, finally, lower extremities. Respiratory muscle paralysis can result in respiratory failure and death if mechanical ventilation is not instituted; higher cortical functions are unaffected.

Diagnosis

Botulism should be suspected in any patient with the acute onset of gastrointestinal, autonomic nervous system, and cranial nerve dysfunction, especially if the patient recently has consumed home-canned foods. Magnetic resonance imaging or computed tomography of the brain and results of lumbar puncture are normal in patients with botulism, but electromyography can show characteristic abnormalities. If food-borne botulism is suspected, stool, serum, and

implicated foods should be tested for botulinum neurotoxin. These tests are performed at the CDC.

Treatment

Supportive therapy with mechanical ventilation has helped to greatly reduce mortality rates from botulism. The diagnosis of botulism must be considered early in any case of unexplained paralysis, and antitoxin should be administered if the diagnosis is credible. The trivalent equine botulinum antitoxin is available only through the CDC, which maintains supplies of antitoxin at sites around the country for immediate release in case of an emergency. To obtain the antitoxin, physicians need to contact their state health department’s emergency hotline or the CDC directly (telephone 800-232-4636). Speed is of the essence, because the antitoxin cannot displace the toxin once toxin has bound to the presynaptic nerve terminal; at this point, antitoxin serves only to bind free circulating toxin. Once symptoms have developed, the usefulness of the antitoxin is greatly reduced. In a large retrospective analysis of 134 cases of botulinum toxin A-mediated disease, patients who received antitoxin therapy early in the course had a mortality rate of 10%, as opposed to a mortality rate of 15% in those who received the antitoxin more than 24 hours after the onset of symptoms and 46% in those who did not receive antitoxin at all.332 Patients who received antitoxin stayed in the hospital an average of 10 days, compared with 56 days for the untreated group. The current recommendation is to administer a single 10-mL dose of intravenous antitoxin to each exposed person. This recommendation is based on the calculation that each vial has enough neutralizing antibody (for types A, B, and E) to bind a titer of toxin that is 100 times greater than the highest titer documented to date by the CDC.

BACILLUS ANTHRACIS

Although most anthrax infections are the result of cutaneous exposure to or inhalation of infected spores, ingestion of infected animal tissue can lead to gastrointestinal disease.

Microbiology

Bacillus anthracis is an aerobic, Gram-positive, sporeforming, nonmotile bacillus that is found in soil. Endospores can remain dormant in soil for many years. Anthrax spores germinate in nutrient-rich environments. Vegetative anthrax bacilli elaborate an antiphagocytic polyglutamyl capsule and a toxin complex that is composed of protective antigen, lethal factor, and edema factor.333 Protective antigen acts as the binding site for the lethal and edema factors. The lethal factor stimulates macrophages to release tumor necrosis factor (TNF)-α and IL-1, which contribute to death from toxemia in anthrax infections characterized by high-grade bacteremia.

Epidemiology

The consumption of endospore-contaminated meat from infected animals is the primary mode of transmission of gastrointestinal anthrax. In 2000, an outbreak of gastrointestinal illness characterized by diarrhea, abdominal pain, and fever was identified in two family members who had consumed meat from a carcass that was found to be contaminated with B. anthracis.334 Fortunately, the infections were mild and self-limited.

Pathogenic Mechanisms

Entry of endospores through the gastrointestinal mucosa initiates infection. Macrophages phagocytose ingested endospores, which then germinate to form vegetative bacteria in mesenteric lymph nodes. The bacteria are then

1883

1884

Section X  Small and Large Intestine released from the macrophages, multiply in the local lymphatic systems, and enter the bloodstream. The release of the exotoxin complexes results in local tissue damage with massive edema, mucosal ulcerations, and the development of systemic toxemia.

Clinical Features

Approximately one to seven days after the ingestion of raw or undercooked meat from infected animals, nausea, vomiting, abdominal pain, and fever develop. Patients often rapidly develop worsening symptoms characterized by bloody diarrhea, diffuse abdominal pain with rebound tenderness, and, occasionally, hematemesis. Ascites, which may be purulent, develops two to four days later. More than 50% of episodes are fatal, with death occurring as a consequence of toxemia, intestinal perforation, or shock from hemorrhage and fluid losses. Oropharyngeal anthrax is a less-common form of infection that develops when spores are deposited in the oropharynx. Symptoms include fever, a severe sore throat, and dysphagia, which can progress to respiratory distress. Examination often reveals swelling of the neck, lymphadenitis, and pharyngeal ulcers covered by a pseudomembrane. Despite the relatively severe symptoms, this form of infection tends to be milder than the gastrointestinal disease and rarely is fatal.

Treatment and Prevention

Because some strains of B. anthracis contain an inducible β-lactamase, initial therapy should consist of ciprofloxacin. Because there is a higher risk of mortality in severe cases, the addition of rifampin or clindamycin, or both, is recommended in these situations. Penicillin and doxycycline are both highly active against B. anthracis in the absence of resistance. An anthrax vaccine, consisting of a sterile filtrate of an attenuated strain of the organism, is available to the U.S. military but not to civilians.

FISH POISONING Food poisoning from marine toxins in fish is a common worldwide problem. The commonest of these are ciguatera and scombroid poisoning.335,336

CIGUATERA

Ciguatera (Spanish: cigua, sea snail) poisoning manifests with a combination of gastrointestinal and neurologic symptoms and findings. It accounts for approximately half of fish-related outbreaks in the United States. Ciguatera poisoning occurs most commonly in the coastal regions of the United States; however, in this age of air transport of foods, cases can occur inland as well. It is especially common in the tropics and subtropics. More than 400 species of fish have been associated with ciguatera poisoning, including grouper, red snapper, amberjack, and dolphin. The illness is caused by the consumption of fish containing toxins produced by dinoflagellates. Fish consume the dinoflagellates that elaborate the toxin(s) that subsequently are stored in fish flesh and viscera. The toxin is concentrated up the food chain as small fish are consumed by larger fish. The fish are not affected by the toxins; they do not appear spoiled and they taste normal. The commonest toxin involved is ciguatoxin, a marine saponin, but a number of toxins can be involved.337 The toxins are both heat- and acid-resistant and survive preservation (freezing) and preparation (cooking) procedures.

Gastrointestinal symptoms occur three to six hours after eating the contaminated fish and usually involve a combination of nausea, vomiting, abdominal cramps, and diarrhea. Sweating and headaches also can occur. A variety of neurologic symptoms occur three to 72 hours after ingestion and include paresthesias, blurred vision, nerve palsies, and cold dysesthesia. Cardiovascular symptoms also can occur and include bradycardia, heart block, and hypotension. Variations in the symptom complex and in the severity of individual symptoms can occur depending on the type of fish eaten and presumably on the type and quantity of toxin or toxins consumed. Diarrhea is the result of toxin-stimulated intestinal secretion mediated by changes in intracellular calcium. Neurologic symptoms are a consequence of alterations in voltage-dependent neural sodium channels. The illness can persist up to one month and, rarely, can last up to one year. Approximately 3% to 20% of patients have chronic effects, such as fatigue, myalgias, and headaches. Chronic symptoms may be aggravated or triggered by ingestion of caffeine or alcohol. Fatalities are rare and usually result from cardiovascular collapse. The diagnosis is a clinical one based on clinical suspicion and compatible signs and symptoms. There is no available confirmatory test and no specific treatment; treatment is supportive. Intravenous mannitol may be helpful in severe cases.

SCOMBROID

Scombroid poisoning is a common but under-reported illness that often is misdiagnosed as a fish “allergy.” It occurs after the consumption of fish that has been poorly refrigerated or improperly stored, allowing bacterial proliferation. Bacteria decarboxylate histidine in the muscle of fish, producing high levels of histamine. The fish do not appear spoiled but might taste peppery; histamine is not destroyed by cooking or freezing. The illness can occur after ingestion of either fresh or canned fish or consumption of foods such as tuna salad or tuna burgers. The most common fish involved are dark meat fish such as tuna, mackerel, and bonito, but scombroid poisoning can also occur with ingestion of mahi-mahi, bluefish, swordfish, or salmon. The usual clinical presentation begins as soon as one hour after ingestion of the contaminated fish. Symptoms and signs include flushing, warmth, erythematous skin rash, pruritus, palpitations, and tachycardia. Patients also may experience headache, blurred vision, and respiratory distress. Occasionally, respiratory distress occurs from facial and lingual swelling. The illness usually resolves spontaneously within 12 hours without any sequelae. Diagnosis is a clinical one based on the signs and symptoms just described. Plasma histamine levels, if measured acutely, may be elevated. Treatment mainly is supportive, but H1 antihistamines may be helpful.

OVERVIEW OF TREATMENT FLUID THERAPY

Potentially devastating consequences of acute infectious diarrhea can result from fluid losses. Rehydration, prevention of ongoing gastrointestinal fluid loss, and replacement of electrolytes, therefore, are the initial goals of therapy. Toxigenic organisms, such as V. cholerae, certain strains of E. coli (ETEC), and rotavirus infection in children can cause extreme dehydration resulting from production of large amounts of isotonic fluid in the small bowel that overwhelm

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning the ability of the colon to reabsorb it. Nonspecific diarrhea, usually caused by viruses, causes less fluid loss and a lower electrolyte concentration in the fecal effluent than does toxigenic diarrhea. The traditional route of fluid administration has been intravenous, but ORSs, used widely in developing countries, can be used to rehydrate patients with moderate volume depletion.68,69,338 Even in the United States, ORS is the treatment of choice for mild to moderate diarrhea in children and adults and can be used in severe diarrhea after initial parenteral replacement of fluid.339 The effectiveness of ORS is a function of its electrolyte content, which has been formulated specifically to replace stool losses. ORS also contains an actively transported substrate, commonly glucose. The use of ORS is based on the physiologic principle that glucose and other substrates enhance sodium absorption in the small intestine, even in the presence of secretory losses caused by bacterial toxins. A variety of ORSs are available and are effective (see Table 107-6).338,340 Even in patients who are vomiting, small increments of ORS can be given effectively. Patients with hypovolemic shock need intravenous hydration. Although there is agreement on the value of ORS in treating dehydrating diarrhea, the specific formulation of electrolytes remains in dispute, particularly in treating well-nourished children in developed countries.340-344 The two areas of dispute concern the appropriate sodium concentration and the osmolality of the solution. Some authorities have voiced concern that the concentration of sodium (90 mmol) in the standard ORS formulation may be too high and that it could cause hypernatremia and seizures.338,342 This issue was examined in a study from Scotland in which children with acute diarrhea and mild dehydration were treated in a randomized fashion with solutions containing sodium concentrations of 35, 50, or 90 mmol and dextrose concentrations of 200, 111, and 110 mmol, respectively; all three formulations proved to be equally safe and effective.343 Several authorities have recommended lower concentrations of sodium and a reduced osmolarity in ORS for children with diarrhea in developed countries.338,342 A multicenter study of children with acute noncholera diarrhea found that treatment with a reduced-osmolarity, reducedsodium ORS was associated with a lower total stool output, less total ORS intake, and a shortened duration of diarrhea than was treatment with the standard ORS.344 Additional criticisms of the traditional glucose-based ORS are that it fails to decrease the quantity and duration of diarrhea. An inexpensive alternative to glucose-based ORS is the substitution for glucose of starch derived from rice or cereals. Treatment with rice-based salt solutions produces smaller stool losses, a shorter duration of diarrhea, and greater fluid and electrolyte absorption than does glucose-based ORS in children and adults with diarrhea.345 Recently, the addition of amylase-resistant starch to ORS has been demonstrated to be even more effective in reducing the duration of diarrhea and fecal weight than was rice flour–based ORS in patients with cholera.346

DIET

The traditional approach to any diarrheal illness is dietary abstinence, which restricts the intake of necessary calories, fluids, and electrolytes. Certainly, during an acute attack, the patient often finds it more comfortable to avoid highfiber foods, fats, and spices because any oral consumption can provide a stimulus to defecation. Although giving the bowel a rest provides relief from symptoms, the patient must maintain intake with oral fluids containing calories and some electrolytes. On balance, it is better to eat judi-

ciously during an attack of diarrhea than to restrict oral intake severely. In children, it is particularly important to restart feeding immediately after the child is able to accept oral intake. It is wise to avoid milk and dairy products during an acute episode of diarrhea because it may be complicated by secondary lactase deficiency. Ingestion of such items in this setting could potentiate fluid secretion and increase stool volume. Caffeine and methylxanthine-containing products should be avoided because the inhibition of phosphodiesterase can increase intracellular cAMP and result in increased intestinal secretion. Thus, coffee, strong tea, cocoa, and soft drinks such as colas can potentiate abdominal cramps and diarrhea. Alcohol can irritate the intestine, and abstinence is recommended. In addition to the oral rehydration therapy outlined earlier, acceptable beverages for mildly dehydrated adults include fruit juices and various bottled soft drinks. It is advisable to defizz a carbonated drink by stirring it or letting it stand in a glass before it is consumed. Soft, easily digestible foods are most acceptable to a patient with acute diarrhea. Secondary lactase deficiency and intolerance to lactose-containing foods can occur during and after various small intestinal infections and can persist for up to one year.

ANTIMICROBIAL DRUGS

Less than 10% of cases of acute infectious diarrhea benefit from treatment with antimicrobial drugs (see Table 10712).347 Infections in adults that benefit from specific therapy include Shigella, EIEC, C. difficile, traveler’s diarrhea, V. cholerae, E. histolytica, Giardia, and some cases of Salmonella. There are conflicting reports concerning the efficacy of antimicrobial drugs in several important infections, such as those caused by Campylobacter; data are insufficient to allow definitive recommendations for infections caused by Yersinia, Aeromonas, vibrios, and several forms of E. coli. Yersinia infections probably do not benefit from treatment unless systemic illness is suspected. Antibiotic therapy is recommended for prolonged or severe cases of Salmonella, Campylobacter, Aeromonas, or Plesiomonas infections (see Tables 107-1 and 107-12).71 Empiric therapy may be recommended for persons with traveler’s diarrhea and for patients with a febrile, dysenteric illness. If Shigella or C. jejuni is suspected, a fluoroquinolone is the drug of choice in adults. The issue of when antimicrobial therapy is appropriate for the management of acute diarrhea has been a vexing problem. Several studies have addressed this issue. Dryden and associates53 studied empiric treatment in patients with severe, acute, community-acquired gastroenteritis (patients with more than four stools per day for more than three days and at least one associated symptom) and found that treatment with ciprofloxacin, 500 mg twice a day for five days, was associated with a reduction in the duration of diarrhea and other symptoms by more than two days, fewer failures, and significant clearing of pathogens when compared with placebo. Six weeks later, there was no difference in stool carriage of the pathogen (12%) and no demonstrable antibiotic resistance. Goodman and colleagues348 studied adults with acute diarrhea comparing ciprofloxacin, TMP-SMX, and placebo and found that ciprofloxacin, but not TMPSMX, shortened the duration of diarrhea.349 Similar findings have been noted in other studies of the empiric therapy of diarrhea.350 Based on these studies, a patient with severe communityacquired diarrhea should receive an antimicrobial drug, preferably a fluoroquinolone. Severe community-acquired diarrhea is defined as more than four watery stools per day,

1885

1886

Section X  Small and Large Intestine lasting at least three days, and including one or more of abdominal pain, fever, vomiting, myalgia, or headache in a previously healthy person. In such patients, there is a high likelihood that a bacterial pathogen will be isolated and treatment with an antibiotic will provide prompt relief of symptoms. In patients with bloody diarrhea, it is not possible to distinguish among Shigella, Campylobacter, and EHEC on clinical grounds. If symptoms of dysentery predominate, quinolone therapy is indicated. If dysentery is not present and if there is a reasonable possibility, based on epidemiologic evidence, that EHEC is the responsible pathogen, then antimicrobial therapy should be withheld until a microbiologic diagnosis can be established.126 The choice of antimicrobial drugs, when indicated, is based on in vitro sensitivity patterns. The fluoroquinolones, including norfloxacin, ciprofloxacin, and ofloxacin, possess broad-spectrum activity against virtually all important diarrheal pathogens (except C. difficile) and thus represent one of the best choices for treatment. Recommendations for antimicrobial therapy of specific infections are given in Table 107-12.349 The optimal duration of antimicrobial therapy has not been defined with precision. Some authors recommend three days of treatment for diarrhea, others five days, and others 10 days; however, several studies of patients with severe diarrhea suggest that a single dose is as effective as more prolonged therapy in cholera and shigellosis.151,351-353

NONSPECIFIC THERAPY

Literally hundreds of antidiarrheal nostrums can be found in pharmacies and assorted medical establishments throughout the world. Many products contain a combination of drugs, most of them therapeutically worthless and others potentially dangerous. Because most patients with infectious diarrhea, even with a recognized pathogen, have a mild self-limited course, specific treatment generally is not required. For more-severe cases, as defined earlier, empiric antimicrobial therapy with a fluoroquinolone should be instituted pending results of stool and blood cultures. A variety of over-the-counter and prescription preparations are available for symptomatic relief of diarrhea and abdominal cramping (Table 107-21). Anticholinergics (dicy-

Table 107-21  Nonspecific Therapy for Infectious Diarrhea Effective Antimotility drugs Codeine, paregoric, tincture of opium Diphenoxylate-atropine Loperamide Bismuth subsalicylate Fluid replacement Intravenous Oral rehydration solution Food Avoid lactose, caffeine, and methylxanthines Lactobacillus GG* Zaldaride maleate† Not Effective Anticholinergics Cholestyramine Hydroxyquinolones Kaolin, pectin, charcoal Lactobacilli *May be effective in children with rotaviral diarrhea. † An antisecretory drug.

clomine [Bentyl], hyoscyamine [Levsin]) decrease intestinal motility and might provide relief of abdominal cramps but do not significantly alter diarrhea. Adsorbents such as attapulgite (Donnagel), kaolin, pectin, and activated charcoal decrease stool liquidity, but there is no evidence that these preparations decrease intestinal fluid loss or number of bowel movements. The opiate derivatives loperamide and diphenoxylateatropine are particularly useful in controlling moderate to severe diarrhea. These agents decrease intestinal motor activity, decrease fluid secretion, and enhance mucosal absorption. The overall effect is to enhance fluid transport, slow transit time, reduce fluid losses, and ameliorate abdominal cramping. They are effective and generally safe. Loperamide is arguably the best agent for acute diarrhea because it does not cross the blood-brain barrier, thereby reducing the risk for habituation or other central nervous system side effects. It also has the additional property of increasing anal sphincter tone. Treatment with loperamide produces rapid improvement, often within the first day of therapy. The concern that an antimotility drug might exacerbate a case of dysentery354 largely has been dispelled by clinical experience. Patients with shigellosis, even S. dysenteriae 1, have been treated inadvertently with loperamide as the only drug and have had normal resolution of disease without evidence that the illness was prolonged or excretion of the pathogen was delayed.285 These drugs, however, generally should not be used in a patient with acute severe colitis, either infectious or noninfectious in origin. Bismuth subsalicylate is an antisecretory agent that has a low incidence of side effects and that is recommended to decrease stool liquidity and frequency. The drug possesses both antimicrobial and antisecretory properties on the basis of the bismuth and salicylate moieties, respectively. In various trials of diarrhea among travelers in Mexico or West Africa, bismuth subsalicylate reduced the frequency of diarrhea significantly over placebo; results generally were better when a high dose (4.2 g/day) was used.355 The combination of an antimicrobial drug and an antimotility drug provides the most rapid relief of diarrhea. In a study of travelers to Mexico, the combined use of loperamide and TMP-SMX curtailed diarrhea in one hour, compared with 30 hours with either drug alone or 59 hours with placebo.286 Even with the severest diarrhea with fecal leukocytes or blood-tinged stool, the median duration of illness was 4.5 hours, a remarkable result in this setting. In a similar study involving U.S. military personnel in Egypt, loperamide added little to the efficacy of ciprofloxacin except in the initial 24 hours, when the combination was slightly better than the antibiotic alone. Addition of loperamide to ciprofloxacin for the treatment of invasive diarrhea has led to a significantly shorter duration of diarrhea and a reduction in the median number of diarrheal stools.156

ACKNOWLEDGMENT

This chapter succeeds the chapter in the seventh edition by Davidson H. Hamer and Sherwood L. Gorbach, entitled “Infectious Diarrhea and Bacterial Food Poisoning.” I wish to acknowledge that I have drawn heavily on the material from their chapter, and many portions of the original chapter remain.

KEY REFERENCES

DuPont HL. New insights and directions in traveler’s diarrhea. Gastroenterol Clin North Am 2006; 35:337-53. (Ref 278.) Fasano A. Toxins and the gut: Role in human disease. Gut 2002; 50:iii9iii14. (Ref 5.) Field M. Intestinal ion transport and pathophysiology of diarrhea. New Engl J Med 2003; 111:931-43. (Ref 40.)

Chapter 107  Infectious Enteritis and Proctocolitis and Bacterial Food Poisoning Giannella RA. Treatment of intestinal infections. In: Wolfe M, editor. Therapy of Digestive Disorders, 2nd ed. Philadelphia: Elsevier; 2006. pp 721-31. (Ref 349.) Grassi GA, Finlay BB. Pathogenesis of enteric Salmonella infections. Curr Opin Gastroenterol 2008: 24:22-6. (Ref 161.) Guerrant RL, Van Gilder T, Steiner TS et al. Practice guidelines for management of infectious diarrhea. Clin Infect Dis 2001; 32:331-50. (Ref 4.) Kelly CP, LaMont JT. Antibiotic-associated diarrhea, pseudomembranous enterocolitis, and Clostridium difficile–associated diarrhea and colitis. In: Feldman M, Friedman LS, Brandt LJ, editors. Sleisenger and Fordtran’s Gastrointestinal and Liver Disease. 8th ed. Philadelphia: Saunders; 2006. pp 2393-2412. (Ref 15.) Landzberg BR, Connor BA. Persistent diarrhea in the retunring traveler: Think beyond persistent infection. Scand J Gastroenterol 2005; 40:112-14. (Ref 284.) Nataro JP, Kaper JB. Diarrheagenic Escherichia coli. Clin Microbiol Rev 1998; 11:142-201. (Ref 29.) Olsen SJ, Aucott JN, Swerdlow DL. Food poisoning. In: Blaser M, Smith P, Ravdin J, editors. Infections of the Gastrointestinal Tract. 2nd ed.

Philadelphia: Lippincott Williams & Wilkins; 2002. pp 199-214. (Ref 313.) Park SI, Giannella RA. Approach to the adult patient with acute diarrhea. Gastroenterol Clin North Am 1993; 22:483-97. (Ref 6.) Sears CL, Acheson DWK. Enteric bacterial toxins.In: Blaser M, Smith P, Ravdin J, editors. Infections of the Gastrointestinal Tract. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 2002. pp 89-107. (Ref 38.) Snyder JD. Oral therapy for diarrrhea. In: Blaser M, Smith P, Ravdin J, editors. Infections of the Gastrointestinal Tract. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 2002. pp 1241-9. (Ref 338.) Spiller RC. Postinfectious irritable bowel syndrome. Gastroenterology 2003; 124:1662-71. (Ref 293.) Thielman NM, Guerrant RL. Acute infectious diarrhea. New Engl J Med 2004; 350:38-47. (Ref 3.) Full references for this chapter can be found on www.expertconsult.com.

1887

CHAPTE R

108 Antibiotic-Associated Diarrhea, Pseudomembranous Enterocolitis, and Clostridium difficileAssociated Diarrhea and Colitis Ciarán P. Kelly and J. Thomas Lamont

CHAPTER OUTLINE Antibiotic-Associated Diarrhea  1889 Etiology  1889 Treatment  1890 Pseudomembranous Enterocolitis  1890

ANTIBIOTIC-ASSOCIATED DIARRHEA ETIOLOGY

Diarrhea is a common side effect of antibiotic use and can result from a variety of mechanisms.1 The most common type of diarrhea, often called simple antibiotic-associated diarrhea (AAD), is believed to result from a disturbance of the normal colonic microflora, leading to alterations in bacterial degradation of nonabsorbed carbohydrates and bile salts. Colonic bacteria normally ferment the complex carbohydrates in dietary fiber and other carbohydrates that are not absorbed in the small intestine, and the fermentation products are then metabolized and absorbed in the colon. Disruption of this process by antibiotic therapy is believed to cause osmotic diarrhea. Some, but not all, bacteria can deconjugate bile salts, and unconjugated bile salts are known to stimulate fluid secretion by the colonic mucosa; another mechanism for AAD may be reduced bacterial degradation of bile salts within the colonic lumen. Other mechanisms that can account for AAD include stimulation of intestinal motility through the motilin-like effect of erythromycin, an allergic reaction, or infection with microorgan-

Clostridium difficile-Associated Diarrhea and Colitis  1891 Pathogenesis and Epidemiology  1891 Clinical Features  1895 Diagnosis  1896 Treatment  1898

isms other than Clostridium difficile, including Clostridium perfringens type A, Staphylococcus aureus, and Salmonella enterica.2-4 The genotype of C. perfringens that causes AAD appears to be distinct from those that induce food poisoning.3,5 Type A strains isolated from patients with AAD carry the C. perfringens enterotoxin (CPE) gene on a plasmid, whereas those that cause food poisoning have a chromosomal CPE gene. S. aureus was identified as a cause of severe AAD and enterocolitis before C. difficile-associated diarrhea was identified.2,6 Since the advent of sensitive and specific testing for C. difficile, however, very few cases of S. aureus AAD have been confirmed, and the true role played by this pathogen in AAD is unclear. Antibioticassociated infection with Klebsiella oxytoca has been described in patients with right-sided hemorrhagic colitis. This rare pathogen releases several potent toxins, and it appears to colonize the bowel after the indigenous flora has been altered by exposure to antibiotics.7 AAD complicates 2% to 5% of antibiotic treatment courses, but the incidence varies depending on the antibio­ tic used; it is more common, for example, during therapy with ampicillin (5%-10%), amoxicillin-clavulanate (10%-

1889

1890

Section X  Small and Large Intestine Table 108-1  Differences between Antibiotic-Associated Diarrhea from Clostridium difficile Infection and from Other Causes CHARACTERISTIC

AAD FROM C. DIFFICILE INFECTION

AAD FROM OTHER CAUSES

Most commonly implicated antibiotics

Clindamycin, cephalosporins, penicillins, fluoroquinolones Usually no history of antibiotic intolerance

Clindamycin, cephalosporins, ampicillin, or amoxicillin-clavulanate History of diarrhea with antibiotic therapy is common

May be florid; evidence of colitis with cramps, fever, and fecal leukocytes is common Evidence of colitis is common; pseudomembranes often are present Hypoalbuminemia, anasarca, toxic megacolon; relapse can occur after treatment with metronidazole or vancomycin Positive May be epidemic or endemic in hospitals or long-term care facilities

Usually moderate in severity (nuisance diarrhea) without evidence of colitis Usually normal

History Clinical Features   Diarrhea   Findings on CT or colonoscopy   Complications   Results of assay for C. difficile toxin   Epidemiologic pattern Treatment   Withdrawal of implicated antibiotic   Antiperistaltic agents   Oral metronidazole or vancomycin

Condition can resolve but often persists or progresses Contraindicated Prompt response

Usually none except occasional cases of volume depletion Negative Sporadic Condition usually resolves Often useful Not indicated

From Bartlett JG. Clinical practice: Antibiotic-associated diarrhea. N Engl J Med 2002; 346:334. AAD, antibiotic-associated diarrhea; CT, computed tomography.

25%), or cefixime (15%-20%) and less common during therapy with fluoroquinolones (1%-2%) or trimethoprimsulfamethoxazole (<1%).8-9 Most cases of AAD are mild and self-limited. Pseudomembranous colitis is absent, and significant complications are rare. C. difficile infection accounts for less than 10% of AAD cases but is an important pathogen to identify because it often requires specific antimicrobial therapy and can lead to life-threatening complications, as discussed in the following section. A comparison between the clinical features of AAD caused by C. difficile and AAD from other causes is presented in Table 108-1.

TREATMENT

The management of simple AAD consists of discontinuing the inciting antibiotic if symptoms are moderately severe or poorly tolerated. If necessary, antiperistaltic agents (e.g., loperamide) may be used to relieve symptoms; these are traditionally avoided for C. difficile-associated diarrhea because of concern for exacerbating toxin-mediated colonic mucosal injury or precipitating toxic megacolon. Because AAD is believed to result from an alteration of the normal colonic microflora, a variety of probiotic agents has been evaluated for its treatment and prevention. In a double-blind controlled clinical trial, oral capsules containing viable Saccharomyces boulardii, a nonpathogenic yeast, were coadministered with antibiotics; this combination treatment reduced the incidence of AAD in hospitalized patients from 22% in the placebo group to 9.5% in the S. boulardii group (P = 0.04).10 Another randomized, placebocontrolled trial, however, failed to demonstrate a beneficial effect for S. boulardii in an elderly population of antibiotic recipients.11 Lactobacillus species, in particular Lactobacillus rhamnosus GG, also have been studied in clinical trials of AAD. In one study, Lactobacillus GG was effective in reducing the incidence of AAD to 5% in children being treated for respiratory tract infections compared with a 16% incidence in the placebo group12; other clinical trials of Lactobacillus GG have yielded negative results.13 A meta-analysis examined the results of randomized, double-blind, placebo-controlled trials of probiotic therapy

for AAD published between 1966 and 2000.14 Nine studies were analyzed, including four using S. boulardii and four using Lactobacillus GG. The combined odds ratio for AAD in the probiotic-treated groups was 0.37 compared with placebo (95% confidence interval [CI]: 0.26-0.53; P < 0.001). For S. boulardii, the odds ratio in favor of active treatment over placebo was 0.39 (95% CI: 0.25-0.62, P < 0.001) and for lactobacilli the odds ratio was 0.34 (95% CI: 0.19-0.61, P < 0.01). A second meta-analysis yielded similar results.15 Thus, the weight of published evidence suggests that probiotic agents such as S. boulardii and lactobacilli, when used in combination with antibiotics, reduce the risk for AAD. Such therapy may be especially advantageous in patients with a history of susceptibility to AAD.

PSEUDOMEMBRANOUS ENTEROCOLITIS Pseudomembranous enterocolitis was a rare entity in the medical literature before the widespread use of antibiotics. In recent decades, pseudomembranous colitis emerged as a common complication of antibiotic use, and almost all cases are now caused by infection with toxin-producing strains of C. difficile. A case report by Finney published in 1893 is considered to be the first description in the medical literature of pseudomembranous enterocolitis.9,16 In that instance, fatal pseudomembranous inflammation of the small intestine followed surgery in a debilitated young woman with gastric outlet obstruction caused by peptic ulcer disease. The presence of an inflammatory pseudomembrane overlying the intestinal mucosa characterizes pseudomembranous colitis (when the colon alone is involved) or pseudomembranous enterocolitis (when the small intestine also is involved).9 The pseudomembrane comprises inflammatory and cellular debris and forms distinctive patches of yellow or gray exudate that obscure the underlying mucosa. In early lesions, a 1- to 2-mm area of punctate ulceration may be visible. Classic pseudomembranes consist of ovoid plaques of 2 to 10 mm in diameter separated by areas

Chapter 108  Antibiotic- and Clostridium difficile-Associated Diarrhea and Colitis of normal or hyperemic mucosa. Histologically, pseudomembranes can be seen to emanate from central areas of epithelial ulceration to form the mucosal plaques. In more-severe cases, the areas of ulceration and the overlying pseudomembranes can coalesce to cover large areas of mucosa. Risk factors for the development of pseudomembranous enterocolitis in the absence of C. difficile infection include intestinal surgery, intestinal ischemia, and other enteric infections. During the 1940s to the 1970s, most reported cases of pseudomembranous enterocolitis occurred following abdominal or pelvic surgery.17,18 Bartlett has identified numerous descriptions of pseudomembranous enterocolitis in the medical literature associated with a wide variety of other intestinal disorders including Shigella infection, Crohn’s disease, neonatal necrotizing enterocolitis, intestinal obstruction, Hirschsprung’s disease, and colonic carcinoma.9 Intestinal ischemia can result in histologic changes similar to those observed in severe C. difficile colitis, although well-defined characteristic patchy pseudomembranes usually are not seen. Severe systemic insults including shock, advanced renal failure, spinal fracture, extensive burns, heavy metal poisoning, and hemolytic-uremic syndrome also have been associated with pseudomembranous enterocolitis. A potential common etiologic factor shared by many of these disorders is hypoperfusion of the intestinal mucosa leading to ischemic necrosis and ulceration. Other infectious agents have been implicated as causes of pseudomembranous colitis in the absence of C. difficile infection, most notably S. aureus.2,3,6 Before C. difficile was identified as the most common cause of pseudomembranous colitis, S. aureus often was identified in stool cultures of patients with postoperative pseudomembranous enterocolitis, and oral vancomycin proved to be an effective therapy.6 In retrospect, it is difficult to ascertain whether the efficacy of vancomycin reflected its activity against staphylococcal infection or against unrecognized infection with C. difficile. Currently, 2% to 3% of patients with antibiotic-associated pseudomembranous colitis have negative tests for C. difficile and its toxins in stool specimens despite use of the most sensitive available assays; it remains unclear what proportion of these patients have false-negative tests for C. difficile or instead are infected with an as-yet-unidentified infectious agent.

dence of community-acquired C. difficile AAD is substantially lower, ranging from 8 to 12 cases per 100,000 person-years.30,31 C. difficile infection also appears to be accompanied by greater morbidity and mortality in the past decade, owing in part to the emergence of increasingly virulent strains. One such strain, designated NAP1/BI, has a mutation in a bacterial gene called txcD, which allows the organism to produce more toxins. The mutant strain also produces a third toxin (called binary toxin) and is resistant to fluoroquinolones, making it more prevalent in patients receiving this class of antibiotics.32

PATHOGENESIS AND EPIDEMIOLOGY

The pathogenesis of C. difficile infection requires the following conditions: alteration of the normal colonic microflora by antibiotics or, rarely, chemotherapeutic agents; oral ingestion of C. difficile or its spores with resultant colonization of the large intestine; release of toxins A and B into the colonic lumen; binding and internalization of toxins by colonocytes; and subsequent colonic damage (colitis). Several host factors, particularly the immune response to C. difficile toxins, determine whether a patient remains an asymptomatic carrier or develops colitis (Fig. 108-1).33

Alteration of the Colonic Microflora

Alteration of the resident colonic microflora, as a consequence of antimicrobial therapy, occurs in nearly all patients who develop C. difficile infection. The protective barrier provided by the intestinal microflora often is referred to as colonization resistance; its impairment by antibiotics and subsequent infection with C. difficile can be demonstrated in animal models.34,35 C. difficile also can colonize the intestines of germ-free mice but is eliminated after these animals are inoculated with fecal flora from normal mice, clearly confirming the importance of the normal flora in preventing colonization.34 Colonization resistance can be demonstrated in vitro by growth inhibition of C. difficile by fecal extracts from healthy adults but not by sterilized extracts.36 Human neonates and infants have poor colonization resistance because they have not yet developed a stable complex colonic microflora.37,38 Colonization rates with C. difficile of 25% to 80% have been reported in healthy infants and children up to 24 months of age, who, despite large concentrations of toxins in the feces, rarely develop C. difficileassociated diarrhea.39 Absence of toxin receptor expression

CLOSTRIDIUM DIFFICILE-ASSOCIATED DIARRHEA AND COLITIS Antibiotic therapy

C. difficile, an anaerobic, Gram-positive, spore-forming, toxigenic bacillus, was first isolated in 1935 from the fecal flora of healthy neonates.19 The organism passed into obscurity until 1978, when the association between toxins released by this organism and antibiotic-induced pseudomembranous colitis first was reported.20,21 Since that time, the incidence of C. difficile infection has increased dramatically, and the organism is now recognized as the primary cause of nosocomial infectious diarrhea in developed countries.22-27 The reported incidence of C. difficile-associated diarrhea has risen steadily over the past decade. For example, in the United States, the rate of nosocomial C. difficile infection per 100,000 population rose from 31 in 1996 to 61 in 200028 to 84 in 2005 (personal communication, L. Clifford McDonald, Centers for Disease Control and Prevention, Atlanta, Ga.). Similarly, C. difficile as a primary or contributing cause of death in the United Kingdom rose from 1000 cases in 2000 to 6500 cases in 2006.29 The reported inci-

Altered colonic microflora C. difficile exposure and colonization Toxin production Protective immune response Asymptomatic carriage

No protective immune response Diarrhea and colitis

Figure 108-1.  Pathogenesis of Clostridium difficile-associated diarrhea and colitis. (From Kyne L, Farrell R, Kelly CP. Clostridium difficile. Gastroenterol Clin North Am 2001; 30:753.)

1891

1892

Section X  Small and Large Intestine Table 108-2  Antimicrobial Agents That Predispose to Clostridium difficile-Associated Diarrhea and Colitis Most Frequently Ampicillin and amoxicillin Cephalosporins Clindamycin Fluoroquinolones Less Frequently Macrolides (including erythromycin) Other penicillins Sulfonamides Trimethoprim/sulfamethoxazole Rarely or Never Bacitracin Carbapenems Chloramphenicol Daptomycin Metronidazole Parenteral aminoglycosides Rifampin Rifaximin Tetracyclines Tigecycline Vancomycin Adapted from Kelly CP, LaMont JT. Treatment of Clostridium difficile diarrhea and colitis. In: Wolfe MM, editor. Gastrointestinal Pharmacotherapy. Philadelphia: WB Saunders; 1993. p 199; and Bouza E, Burillo A, Muñoz P. Antimicrobial therapy of Clostridium difficile–associated diarrhea. Med Clin North Am 2006; 90:1141-63.

on the immature colonic epithelium has been suggested as a mechanism to explain the carrier state in infants and children.40 Almost all antimicrobial agents can predispose to C. difficile diarrhea and colitis, including vancomycin and metronidazole,41,42 but the precise risks associated with individual agents are difficult to establish.43-45 The frequency of association of specific antibiotics is related to their frequency of use, their route of administration, and their effect on the colonic microflora.43,45 Antibiotics commonly associated with C. difficile infection and diarrhea include clindamycin, cephalosporins, ampicillin, amoxicillin, and (more recently) the fluoroquinolones (Table 108-2).32,44,46-48 Cancer chemotherapy agents that possess antibacterial properties and bowel preparation regimens (e.g., before colonoscopy or colonic surgery) rarely can result in sufficient disturbance of the intestinal microflora to allow subsequent colonization with C. difficile.49 One report indicated that healthy adults without known exposure to antibiotics or other modifiers of the colonic microflora also occasionally develop C. diffile colitis.30

Hospital Epidemiology of Clostridium difficile Infection

Chronic intestinal carriage rates of C. difficile in healthy adults are low (0% to 3% in American and European populations) and might represent intestinal transit without true colonization.27,50,51 It also is unclear whether carriage is a temporary or permanent state.42 In contrast, hospital inpatients treated with antibiotics have reported colonization rates of 10% to 21%.23,27,52-55 The hospital environment is a major source of C. difficile infection; not only infected stool, but also environmental surfaces, soiled bedding, bedpans, toilet seats, and the hands and stethoscopes of health care workers are potential sources of nosocomial C. difficile infection.23,54 In one study, C. difficile was acquired on average in 3.2 days by patients who shared a room with an infected roommate compared with 18.9 days by patients in single rooms or with room-

Table 108-3  Practice Guidelines for the Prevention of Clostridium difficile Diarrhea Limit the use of antimicrobial drugs Wash hands between contacts with all patients Use enteric (stool) isolation precautions for patients with C. difficile diarrhea Wear gloves when contacting patients with C. difficile diarrhea or their environment Disinfect objects contaminated with C. difficile with sodium hypochlorite, alkaline glutaraldehyde, or ethylene oxide Educate the medical, nursing, and other appropriate staff members about the disease and its epidemiology From Fekety R. Guidelines for the diagnosis and management of Clostridium difficile-associated diarrhea and colitis. American College of Gastroenterology, Practice Parameters Committee. Am J Gastroenterol 1997; 92:739.

mates whose stool cultures were culture negative for C. difficile.52 In the same study, C. difficile was cultured from the hands of 59% of hospital workers caring for patients with positive C. difficile cultures. The organism also was frequently cultured from bedrails, toilets, floors, call buttons, and other surfaces in the rooms of infected patients. Asymptomatic carriers rarely develop C. difficileassociated diarrhea, but they serve as an important reservoir of nosocomial infection.55,56 In one study, 29% of environmental cultures taken from the hospital rooms of symptomfree carriers were positive for C. difficile, compared with only 8% of cultures from rooms of patients who were culture-negative for C. difficile.52 In antibiotic-treated animals, the infective dose of toxigenic C. difficile may be as low as two organisms.21 If human susceptibility is similar, control of C. difficile infection in hospitals will continue to be a major challenge because up to 109 organisms per gram are excreted in liquid feces.57,58 Highly resistant spores of C. difficile can persist for many months in the hospital environment and can result in infection if ingested by a susceptible host.57 Although it is not possible to eradicate C. difficile and its spores from the hospital environment, certain control measures have been recommended to reduce the prevalence of C. difficile-associated diarrhea (Table 108-3).59,60 Infected inpatients should be bedded in private rooms whenever possible to reduce patient-to-patient spread. Strict enteric precautions and regular hand washing after patient contact should be observed, because C. difficile can be cultured from the hands of health care workers after as many as 60% of contacts with infected patients.52 The use of alcoholbased hand gels is not as effective as washing with soap and running water in removing C. difficile spores.60 A controlled trial of using vinyl disposable gloves during patient contact also reduced the transmission of infection.61 After discharge of infected patients, surface environmental disinfection is best performed using a cleaning agent (e.g., hypochlorite solution) containing at least 5000 ppm available chlorine.60 Hospital outbreaks of C. difficile-associated diarrhea are common and likely result from the close approximation of susceptible persons (elderly and infirm patients) who are taking antibiotics and who are then exposed to the pathogen either in the hospital environment or through person-toperson spread. Some recent reports suggest that hospital and community outbreaks are related to the emergence of mutated hypervirulent strains, which are highly toxigenic and resistant to numerous antibiotics including fluoroquinolines.32,48 Prophylactic therapy with metronidazole or vancomycin is not effective as a disease control measure,62 and C. difficile diarrhea is prevented best by avoiding the

Chapter 108  Antibiotic- and Clostridium difficile-Associated Diarrhea and Colitis unnecessary use of broad-spectrum antibiotics, especially in hospitalized patients, and by careful attention to hand hygiene. In the future, increasing individual and herd immunity to C. difficile by vaccination or by passive immunotherapy may become a viable approach to reducing the prevalence of this common nosocomial disease.59,63-65 Prophylactic measures such as the use of bacterial and yeast probiotics or toxin binders in high-risk hospital patients also warrant further investigation.66-69

Clostridium difficile Toxins

Pathogenic strains of C. difficile produce two structurally similar protein exotoxins, toxin A and toxin B, which are the only known virulence factors. The genes encoding toxin A and toxin B reside in a 19.6-kb chromosomal region, the C. difficile pathogenicity locus, which contains the genes encoding toxin A (tcdA) and B (tcdB) as well as two putative regulatory genes (tcdC and tcdD, also called tcdR) (Fig. 108-2).70-75 The tcdD gene product appears to up-

regulate toxin transcription by complexing with RNA polymerase that binds to the toxin promoter regions. The tcdC gene is transcribed in the opposite direction to tcdA, tcdB, and tcdD, and its gene product appears to decrease toxin production.70-75 Mutations of tcdC are associated with increased virulence that may be related to increased toxin production.76,77 The fifth gene of the pathogenicity locus, tcdE, encodes a protein of undetermined function, although some data support the theory that it acts to lyse bacterial cell walls, thereby releasing toxins A and B.78 Toxins A (308 kd) and B (220 kd) are members of the large clostridial cytotoxin family, share a number of structural features, and are 49% identical at the amino acid level.79-81 Both toxins carry an N-terminal enzymatic domain that mediates their toxic effects on mammalian cells, a central hydrophobic region that might act as a transmembrane domain to facilitate entry into the cytoplasm, and a Cterminal domain consisting of a series of repeated sequences that mediate toxin binding (Fig. 108-3). Both toxins function as uridine diphosphate glucose (UDP-glucose) hydrolases

RNA polymerase TcdD

TcdD

TcdC

TcdB

TcdE

TcdA

Toxin B

TcdE

Toxin A

Lysis of bacterial cytoplasmic membrane

Toxin release Figure 108-2.  Clostridium difficile pathogenicity locus. The pathogenicity locus of C. difficile is a 19.6-kb segment carrying five genes (TcdA-E), including the genes encoding toxin A (TcdA) and toxin B (TcdB). TcdD (also called TcdR) appears to encode a positive regulator of toxin A and toxin B transcription. The TcdD gene product forms complexes with RNA polymerase that bind to the TcdA and TcdB promoter regions. TcdC appears to act as a negative regulator of toxin production. TcdE may mediate toxin release through its ability to disrupt the bacterial cytoplasmic membrane. (From Warny M, Kelly CP. Pathogenicity of Clostridium difficile toxins. In: Hecht G, editor. Microbial Pathogenesis and the intestinal epithelial cell. Washington, DC: ASM Press; 2003. p 503.)

C. difficile toxin A: 2710 aa (308 kd) Enzymatic domain 102–TRP

Hydrophobic region ?

958 1130

CROPs (861 aa) 1848

2710

N

C CROPs (515 aa)

C. difficile toxin B: 2366 aa (270 kd) 102–TRP N

546

956 1128

1851

2366 C

Figure 108-3.  Structure of Clostridium difficile toxins. Toxin A and toxin B share three similar domains: an N-terminal enzymatic domain responsible for cytotoxicity that carries a conserved tryptophan residue (Trp-102) probably involved in binding to uridine diphosphate glucose (UDP-glucose); a central, major hydrophobic region of 172 amino acids (aa) that is highly conserved and might act as a transmembrane domain to facilitate exit from endosomes into the cytoplasm; a C-terminal binding domain composed of contiguous repeating units also known as clostridial repetitive oligopeptides (CROPs). Toxin A carries 30 CROPs whereas toxin B carries 19. They include sequences of 50 amino acids (represented in blue) and 21 amino acids (represented in white), respectively. (From Warny M, Kelly CP. Pathogenicity of Clostridium difficile toxins. In: Hecht G, editor. Microbial pathogenesis and the intestinal epithelial cell. Washington, DC: ASM Press; 2003. p 503.)

1893

Section X  Small and Large Intestine and glucosyltransferases, a requirement for their cellular toxic effects. Following internalization into the host cell cytoplasm, the toxins catalyze the transfer and covalent attachment of a glucose residue from UDP-glucose to a conserved threonine amino acid on small (20 to 25 kd) guanosine triphosphate– binding rho proteins. Rho proteins are part of the Ras superfamily, are expressed in all eukaryotic cells, and act as intracellular signaling molecules to regulate cytoskeletal organization and gene expression. The rho proteins, RhoA, Cdc42, and Rac, are substrates for both toxins A and B, whereas Rap is a substrate for toxin A only.74,82-84 Glucosylation of rho proteins by the toxins leads to disordered cell signaling, disorganization of the cytoskeleton, disruption of protein synthesis, cell rounding, and cell death.74,85 Both toxins also activate nuclear factor κB (NF-κB), mitogenactivated protein (MAP) kinases, and cyclooxygenase (COX)-2 in target cells, leading to the release of proinflammatory cytokines including interleukin (IL)-1β, tumor necrosis factor (TNF)-α, and IL-8.85,86 These cellular proinflammatory effects contribute to the marked intestinal inflammatory response evident in C. difficile-associated diarrhea and pseudomembranous colitis. Toxin A initially was thought to be the only enterotoxin based on studies in animals85,87,88 whereas toxin B, an extremely potent cytotoxin, had minimal enterotoxic activity in animals. This suggested that toxin B did not contribute to diarrhea and colitis in humans,87,89-91 although studies on human colon show that, in fact, toxin B is 10 times more potent than toxin A in inducing in vitro colon injury.92,93 Furthermore, toxin A−/toxin B+ strains of C. difficile have been isolated from patients with diarrhea and pseudomembranous colitis,94-97 confirming that toxin B is a major virulence factor in human disease. A minority (less than 10%) of C. difficile clinical isolates produce the third toxin—binary toxin—that is analogous to the iota toxin of Clostridium perfringens and is encoded at a site distant from the pathogenicity locus that encodes toxins A and B. Binary toxin is composed of two parts: a 48-kd enzymatic protein and a 99-kd binding protein. Although binary toxin shows some enterotoxic activity in animal models, its role, if any, in the pathogenesis of C. difficile-associated diarrhea and colitis is unclear. Most pathogenic strains of C. difficile lack binary toxin but nonetheless cause substantial colonic inflammation and injury. The NAP-1/BI or epidemic strain is binary toxin positive, however, thereby raising renewed suspicion that this toxin might enhance the effects of toxins A and B.

The Immune Response to Clostridium difficile

Serum IgG and IgA antibodies against C. difficile toxins are found in more than 50% of healthy children and adults.98-102 Mucosal IgA antitoxin antibodies also are detectable in more than 50% of human colonic secretions and might inhibit receptor binding of toxin A.100,102 Immunization against C. difficile toxins protects animals from C. difficile colitis but does not protect against colonization—a situation that may be similar to the asymptomatic carrier state in humans.53,103 High serum IgG antitoxin A antibody concentrations are associated with protection against C. difficile-associated diarrhea and colitis.63-65 Recurrent C. difficile diarrhea has been associated with low serum antitoxin antibody concentrations in children and in adults.98,102,104,105 In one study, adult inpatients with C. difficile diarrhea and a low level of serum IgG against toxin A had a 48-fold greater risk of recurrent disease after successful treatment compared with patients who had high antitoxin concentrations

3 Serum IgG antitoxin A (ELISA units)

1894

Asymptomatic carriage

2 Single episode of C. difficile diarrhea

1 Recurrent C. difficile diarrhea –3

1 3 6 9 12 (Admission (Discharge to hospital) from hospital) Days after colonization by C. difficile

Figure 108-4.  Serum immunoglobulin G (IgG) antitoxin A antibody response and clinical outcome of infection with Clostridium difficile. Patients with nosocomial C. difficile diarrhea were studied prospectively, and serum IgG antitoxin A antibody concentrations were measured by enzyme-linked immunosorbent assay (ELISA) at regular intervals. A correlation was observed between the IgG response to toxin A and the clinical outcome of infection. Asymptomatic carriers mounted an early memory immune response to toxin A. By contrast, no significant increase was found in serum IgG antitoxin A of patients who experienced recurrent C. difficile diarrhea. In those who had a single episode of diarrhea, IgG antitoxin A levels generally were increased on day 12 of their first episode. Thus, a serum antibody response to toxin A during C. difficile infection is associated with protection against symptoms or against recurrent diarrhea. (Adapted from Kyne L, Warny M, Qamar A, Kelly CP. Association between antibody response to toxin A and protection against recurrent Clostridium difficile diarrhoea. Lancet 2001; 357:189.)

(Fig. 108-4).106 High serum IgG antitoxin A concentrations also have been identified in asymptomatic carriers of toxigenic C. difficile. In a prospective study of nosocomially acquired C. difficile, the 51% of infected patients who were asymptomatic carriers had serum IgG antitoxin A concentrations that were three times higher than those in patients with diarrhea (see Fig. 108-4).53 The immune response to toxin B has not yet been clearly correlated with specific clinical outcomes. Nonetheless, toxin B is both pathogenic and immunogenic in humans, and antibody responses to toxin B might contribute to immune protection against C. difficileassociated diarrhea.63-65

Other Risk Factors for Clostridium difficile Infection

In addition to antimicrobial therapy, increasing age and increased comorbidity are important risk factors for C. difficile infection.107 In England and Wales, 75% of reported C. difficile infections between 1992 and 1996 occurred in patients aged 65 years or older.50 Data from the United States also demonstrate that age is an independent risk factor for this infection.108,109 The elderly particularly are predisposed to infection with C. difficile because of increased nosocomial exposure and frequent courses of antibiotics and a reduced ability of their polymorphonuclear leukocytes to phagocytose these organisms.110 In one study of antibiotic recipients, patients with severe underlying disease at the time of hospital admission were eight

Chapter 108  Antibiotic- and Clostridium difficile-Associated Diarrhea and Colitis times more likely to develop C. difficile infection compared with patients who were less severely ill.53 Other reported risk factors for C. difficile infection include the use of a nasogastric tube, gastrointestinal procedures, intensive care unit stay, and length of hospital stay.43 The strengths of the associations of these risk factors with C. difficile vary from study to study. These factors often are markers of disease severity, older age, or both, and the significance of their association with C. difficile can decline or be lost after controlling for these confounding variables.53,107,109 The role of acid suppression in C. difficile infection is unclear. In theory, reduction of gastric acid could allow a greater number of viable spores to reach the colon; however, because spores are generally acid resistant, the importance of this effect is unclear. Some studies have shown an increased risk of C. difficile infection with acid suppression,109 but others have not confirmed this after adjusting for confounding variables.32,111 Patients undergoing cytotoxic chemotherapy for malignancy are at risk for C. difficile-associated diarrhea infection because of frequent antibiotic use, nosocomial exposure to C. difficile, and severe comorbidity.111,112 Even in the absence of antibiotic use, antineoplastic chemotherapy, especially with methotrexate, predisposes to C. difficile infection, reflecting the ability of these drugs to alter the colonic microflora and reduce C. difficile colonization resistance.49 C. difficile-associated diarrhea also has been reported in patients undergoing immunosuppressive therapy in the setting of solid organ or bone marrow transplantation.113,114 Patients with human immunodeficiency virus (HIV) infection also are at risk for C. difficile-associated diarrhea because of multiple risk factors, including frequent pro­ phylactic and therapeutic antibiotic use, hospitalization, and immunocompromise.115-118 C. difficile colitis behaves the same in patients with acquired immunodeficiency syndrome (AIDS) as it does in control groups118 and testing for C. difficile should be a routine part of the diagnostic evaluation in patients with HIV infection, diarrhea, and a history of current or recent antibiotic treatment. Patients with inflammatory bowel disease (IBD) are at increased risk for C. difficile infection.119-122 Infection with a broad range of enteric pathogens including C. difficile, Campylobacter, and Salmonella species can precipitate or mimic disease relapse in IBD. C. difficile is the most commonly identified specific pathogen in IBD patients in North America and Europe, however, and is present in as many as 5% to 19% of patients with relapse in some case series.119-122 Some IBD patients with C. difficile infection do not have a history of recent antibiotic use, suggesting that IBD itself might impair colonization resistance. The possibility of enteric infection with C. difficile or other pathogens should be considered in patients with an increase in IBD disease activity. If C. difficile infection is identified, antimicrobial therapy with metronidazole or vancomycin is indicated in combination with other IBD therapies. In one study, IBD inpatients with coexisting C. difficile infection were more likely to have severe disease and to require colectomy than similar patients without coexisting infection.122

CLINICAL FEATURES

Clinical manifestations of C. difficile infection range from asymptomatic carriage to mild or moderate diarrhea to lifethreatening pseudomembranous colitis. Asymptomatic carriage of C. difficile is common in hospitalized patients. Several large epidemiologic studies indicate that 10% to 21% of hospital inpatients receiving antibiotics in high-risk

units are carriers.28,52,53,55,123 Although most of the C. difficile isolates from carriers are toxin producing, the carriers do not develop symptomatic disease, perhaps as a result of protective immunity.28,55,104 In patients who develop diarrhea with C. difficile, symptoms usually begin soon after colonization. The incubation period is usually less than a week, with a median time of onset of approximately two days.28,52,53,124 Colonization can occur during or after antibiotic treatment. Olson and associates26 reported that 96% of patients with symptomatic C. difficile infection had received antibiotics within 14 days of the onset of diarrhea and that all had received an antibiotic within the previous three months. C. difficile diarrhea typically is associated with the frequent passage of loose or watery bowel movements. Mucus or occult blood may be present, but melena or hematochezia is uncommon and, if present, should suggest the presence of IBD, colon cancer, or another source of bleeding. Some patients present with fever, leukocytosis, and cramping abdominal pain.125 Because C. difficile is not an invasive pathogen, extraintestinal manifestations of C. difficile infection such as septic arthritis, bacteremia, or splenic abscess are extremely rare.126-129 An oligoarticular, asymmetrical, nondeforming large-joint arthropathy, similar to that seen in other infectious colitides, sometimes is seen.130 Patients with more severe disease can develop a colonic ileus or toxic dilatation and present with minimal or even no diarrhea.125 In the absence of diarrhea, the only clues to the diagnosis may be high fever, moderate or marked (e.g., leukemoid) polymorphonuclear leukocytosis, lower or diffuse abdominal pain, tenderness, and distention. Abdominal plain films might reveal a dilated colon (more than 7 cm in its greatest diameter), toxic megacolon, or small bowel ileus with air-fluid levels mimicking intestinal obstruction or ischemia. In such cases, a computed tomographic scan of the abdomen may reveal nonspecific features common to ischemic, infectious, and inflammatory colitides (Fig. 108-5).131 Radiologic features of pseudomembranous colitis include mucosal edema, a thickened colonic wall, pancolitis, and pericolonic inflammation with or without ascites, usually without any small bowel involvement other than ileus.132 Flexible sigmoidoscopy or colonos-

Figure 108-5.  Computerized tomography scan of the abdomen in a patient with Clostridium difficile colitis. Marked thickening of the colonic wall in the sigmoid colon and an accordion-like pattern, produced by a series of broad edematous colonic haustral folds, are evident (arrows). (From Linevsky JK, Kelly CP. Clostridium difficile colitis. In: Lamont JT, editor. Gastrointestinal infections: diagnosis and management. New York: Taylor & Francis Group; 1997. p 293.)

1895

1896

Section X  Small and Large Intestine Table 108-4  Stool Tests for the Diagnosis of Clostridium difficile Infection TEST

DETECTS

ADVANTAGES

DISADVANTAGES

Cytotoxin assay

Toxin B

Enzyme immunoassay

Toxin A and/or B

Latex agglutination assay Culture

Glutamate dehydrogenase Glutamate dehydrogenase Toxigenic and nontoxigenic C. difficile

Gold standard; highly sensitive and specific Fast (2-6 hr); easy to perform; highly specific Fast (2-6 hr); easy to perform; sensitive Fast; inexpensive; easy to perform Sensitive; allows strain typing in epidemics

Polymerase chain reaction assay

Toxin A or B genes in isolates or directly in feces

Requires tissue culture facility; takes 24-48 hr Not as sensitive or as specific as the cytotoxin assay Not as specific as the cytotoxin assay Lacks sensitivity and specificity Requires anaerobic culture; not specific for toxin-producing strains; takes 2-5 days Requires expertise in molecular diagnostic techniques

Highly sensitive and specific

Adapted from Linevsky JK, Kelly CP. Clostridium difficile colitis. In: Lamont JT, editor. Gastrointestinal infections: Diagnosis and management. New York: Taylor & Francis; 1997. p 293.

copy is sometimes indicated to identify pseudomembranous colitis (see later) when the diagnosis remains unclear after initial evaluation. Complications of severe C. difficile colitis include dehydration, hypoalbuminemia, electrolyte disturbances, renal failure, hypotension, toxic megacolon, systemic inflammatory response syndrome, bowel perforation, and death.59,125

DIAGNOSIS

The diagnosis of C. difficile diarrhea or colitis is based on a history of recent or current antimicrobial therapy, development of diarrhea or other evidence of acute colitis, and demonstration of infection by toxigenic C. difficile, usually by detection of toxin A or B, or both, in a stool sample.48,59

Tests for Clostridium difficile Infection

The diagnosis of C. difficile diarrhea should be considered in any patient with acute diarrhea who has received anti­ biotics within the previous three months and especially in anyone whose diarrhea began 72 hours or more after hospitalization. Approximately 40% of patients with C. difficile diarrhea at tertiary referral centers are symptomatic on admission to the hospital, and most have had a recent prior hospitalization.52,53,133 Although a history of recent antibiotic use is common, it is not a requirement for diagnosis.30 Testing of solid or formed stools for C. difficile toxin is not recommended because only patients with diarrhea require treatment.52,53,59,62,123 Treatment of asymptomatic carriers with antimicrobial agents against C. difficile is not recommended because it might only prolong the carrier state beyond the usual two to six weeks.62 Follow-up stool testing is not indicated in an asymptomatic patient, even in patients discharged to chronic care facilities, because asymptomatic carriage is already highly prevalent in these facilities. Stool carriage of C. difficile can persist for up to six weeks after cessation of symptoms and does not require therapy.134 Because asymptomatic carriers can act as hidden reservoirs for C. difficile infection, especially in hospitals and nursing homes, universal precautions should be followed for all patients to reduce the likelihood of patientto-patient spread of nosocomial infectious disease. If C. difficile diarrhea is suspected, a freshly passed stool sample should be submitted immediately to the laboratory in a clean watertight container to be tested for the presence of fecal toxin A or B. Anaerobic storage or the use of transport media is not necessary, but storage at ambient temperatures can result in denaturation of fecal toxin; samples

should therefore be tested immediately or refrigerated or frozen, pending later testing.54,135 A variety of laboratory tests are available to diagnose infection with toxigenic C. difficile, but enzyme immuno­ assays (EIAs) to detect toxin antigens in stool currently are used most commonly (Table 108-4). These tests have the advantages of being relatively inexpensive, quick (2 to 12 hours), and highly specific, although their relatively low sensitivity (about 90%) leads to some false-negative results. The tissue culture cytotoxicity assay is more sensitive and has greater diagnostic accuracy, but it is more costly and time consuming (24 to 72 hours). Tissue Culture Cytotoxicity Assay The current gold standard test to identify C. difficile toxins in stool is the tissue culture cytotoxicity assay.136-138 Toxins A and B inactivate rho proteins, causing a disintegration of the actin cytoskeleton and a characteristic rounding of cells in tissue culture. A suspension of diarrheal stool diluted in phosphate-buffered saline is centrifuged, filtered, and then added to a monolayer of cultured cells, such as fibroblasts or Chinese hamster ovary cells. The monolayer is examined at 24 hours and again at 48 hours for cell rounding. The specificity of a positive result is established by preincubating an aliquot of the patient’s stool specimen with specific neutralizing antitoxin antibody; stool that gives a positive result when tested alone and becomes negative after incubation with antitoxin antibody is a true positive. False-positive results can be seen with pathogens other than C. difficile and with non– C. difficile enterotoxins. The cytotoxicity assay is highly sensitive (67% to 100%) and specific (85% to 100%) if it is performed under optimal conditions; however, sensitivity may be reduced by inactivation of toxins during transport and storage, by the age and type of cell line used, and by dilution of the stool.54,135,139-141 Therefore, a negative cytotoxicity test does not completely exclude C. difficile as the cause of diarrhea. Disadvantages of the cytotoxicity assay are that it is relatively expensive, requires technical expertise and a cell culture facility, and takes 24 to 72 hours for the test to be completed. Enzyme-Linked Immunoassays Commercially available EIAs are widely used to detect toxin A or toxins A and B of C. difficile in stool specimens.136-138,142 Toxin is detected by its interaction with either a monoclonal antibody or polyclonal antiserum that specifically recognizes toxin epitopes. EIAs are easier to perform than is the

Chapter 108  Antibiotic- and Clostridium difficile-Associated Diarrhea and Colitis cytotoxicity test, are relatively inexpensive, and are fast; results may be available within two to six hours. Although they have high specificity (75% to 100%) for toxins, their main drawback is that they are less sensitive than the cytotoxicity test (63% to 99%). In addition, some EIA kits detect only toxin A, in which case diarrhea due to a toxin A−/B+ strain of C. difficile has a falsely negative test result.96 For this reason, commercial kits that detect both toxins A and B have a slight advantage over those that detect toxin A alone.142 Distinct from tests for C. difficile toxins A and B, immunoassays also have been used to detect C. difficile common antigen (glutamate dehydrogenase, GDH) in stool. The initial latex agglutination assay method lacked diagnostic accuracy and is not recommended. More recent EIAs for fecal GDH have shown improved sensitivity (85% to 95%) and specificity (89% to 99%), are rapid, and are not expensive. These changes have led to the use of EIA for GDH as an initial screening test, with confirmation of positive results using another test such as the tissue culture cytotoxicity assay.143

A

Clostridium difficile Culture Culture of stool for C. difficile is sensitive (89% to 100%) but is not specific for toxin-producing strains of the bacterium. Therefore, cultured isolates then must be tested in vitro for toxin production to improve test specificity, but this is costly and time consuming. One advantage of culturing C. difficile is that it permits strain typing of individual isolates, and therefore it is useful in tracking hospital outbreaks for epidemiologic studies.56,57,75-77 Polymerase Chain Reaction for Detection of Toxin Genes Polymerase chain reaction (PCR), using specific primers based on the genes for toxins A and B, can detect toxigenic C. difficile in clinical isolates.144-152 PCR is highly sensitive (100%) and specific (96.7% to 100%), and PCR methods for detecting toxin genes directly in feces have been described.147-152 Using a nested PCR assay on stool samples, Alonso and colleagues152 reported 99% concordance with the cytotoxicity assay and a sensitivity and specificity of 96.3% and 100%, respectively. Application of PCR methods in the clinical laboratory might help to overcome some of the limitations of the currently used EIA and tissue culture assays. Sigmoidoscopy and Colonoscopy Neither sigmoidoscopy nor colonoscopy is required for diagnosis in most patients with C. difficile diarrhea.59 Endoscopy is helpful, however, when the diagnosis is in doubt or when disease severity demands rapid diagnosis. Sigmoidoscopy may be normal in patients with mild diarrhea or might demonstrate nonspecific colitis in moderate cases. The finding of colonic pseudomembranes in a patient with AAD is virtually pathognomonic for C. difficile colitis (Fig. 108-6).152,153 Pseudomembranes appear as yellow, gray, or white plaques 2 to 5 mm in diameter, and in some areas they can coalesce to cover large portions of the mucosal surface. Sigmoidoscopy might not be sufficient to identify all patients with pseudomembranous colitis, because approximately 15% to 20% only have pseudomembranes in the more proximal areas of the colon.154 Other nonspecific endoscopic findings include erythema, edema, friability, small ulcerations, and erosions. In mild disease, colonic mucosal biopsies may be normal or demonstrate only mild and nonspecific acute inflammatory changes with neutrophil infiltration. In more severe

B Figure 108-6.  Colonoscopic appearance of pseudomembranous colitis (A) and a colon resection specimen from a patient with severe, refractory Clostridium difficile diarrhea and colitis (B). Characteristic raised adherent yellow plaques that vary in size from 2 to 5 mm are visible on the colonic mucosa. In some areas, coalescing pseudomembranes are evident. There is some erythema of the colonic mucosa between the pseudomembranes, but the epithelium is intact. (A, From Kwon JH, Kelly CP. Clostridium difficile and antibiotic-associated diarrhea. In Bayless RM, Diehl AM, editors. Advanced Therapy in gastroenterology and liver disease, 5th ed. Hamilton, Ontario, Canada: BC Decker; 2005, p 302. B, From Kelly CP, Pothoulakis C, LaMont JT. Clostridium difficile colitis. N Engl J Med 1994; 330:257.)

cases, colonic histology shows focal ulceration of the mucosa associated with the eruption of inflammatory cells and necrotic debris that covers the area of ulceration, the so-called summit or volcano lesion (Fig. 108-7).152,155

Miscellaneous Laboratory Tests

Many patients with acute C. difficile diarrhea develop a polymorphonuclear leukocytosis with a left shift. Occasionally a leukemoid reaction with an extremely high white blood cell count of more than 50,000 cells/mm3 is seen. A peripheral white blood cell count of greater than 15,000 cells/mm3 is associated with negative clinical outcomes and a count of greater than 25,000 cells/mm3 is associated with an increased mortality risk.156 Decreased serum albumin and elevated creatinine levels also are markers of severe disease. Patients with protein-losing colopathy and severe hypoalbuminemia can also develop peripheral edema, ascites, or anasarca.

1897

1898

Section X  Small and Large Intestine Table 108-5  Treatment of Clostridium difficile Diarrhea and Colitis Discontinue the inciting antibiotic if possible Provide supportive therapy (see text) Confirm the diagnosis Prescribe specific therapy if symptoms are moderately severe or persistent: Metronidazole orally for 10-14 days (drug of choice for mild-tomoderate disease) Vancomycin orally for 10-14 days if Diarrhea and colitis are severe Diarrhea does not improve during metronidazole treatment Patient cannot tolerate metronidazole Patient is pregnant or younger than 10 yr of age If the patient cannot tolerate oral medication, prescribe metronidazole 500 mg, every 6 hours intravenously Adapted from Fekety R. Guidelines for the diagnosis and management of Clostridium difficile-associated diarrhea and colitis. American College of Gastroenterology, Practice Parameters Committee. Am J Gastroenterol 1997; 92:739.

Figure 108-7.  Histologic image of an endoscopic biopsy specimen from a patient with pseudomembranous colitis showing a summit or volcano lesion. Focal ulceration of the colonic mucosa is evident (lower arrow), with exudation of a pseudomembrane made up of inflammatory cells, fibrin, and necrotic debris (upper arrow). The adjoining mucosa is intact. (From Kelly CP, Pothoulakis C, LaMont JT. Clostridium difficile colitis. N Engl J Med 1994; 330:257.)

TREATMENT Mild to Moderately Severe Clostridium difficile Diarrhea and Colitis

The first step in the management of C. difficile diarrhea and colitis is to discontinue the precipitating antibiotics if possible (Table 108-5).59 Diarrhea resolves in approximately 15% to 25% of patients without specific anti-C. difficile therapy.26,157 Conservative therapy alone, however, is not appropriate in patients who are severely ill or who have multiple other active medical problems. In patients with active infections elsewhere (e.g., pneumonia and urinary tract infection) and in whom antibiotic therapy must be continued, the antibiotic regimen should be switched, if possible, to agents with a relatively low likelihood of exacerbating C. difficile diarrhea, for example, parenteral aminoglycosides, trimethoprim, or erythromycin (see Table 108-2).46 Antimotility agents such as diphenoxylate plus atropine (Lomotil), loperamide (Imodium), or narcotics often are avoided because of concern for impaired toxin clearance or precipitation of ileus and toxic dilatation, but the data supporting this practice are limited and contradic-

Table 108-6  Comparison of Metronidazole and Vancomycin for the Treatment of Clostridium difficile Diarrhea VARIABLE

METRONIDAZOLE

VANCOMYCIN

Dose Frequency Duration Route Response rate Disadvantages

250-500 mg Three or four times daily 10-14 days Oral or intravenous 87% Systemic side effects; rare resistant strains of C. difficile

125-500 mg Four times daily 10-14 days Oral only 97% Can encourage proliferation of nosocomial VRE

VRE, vancomycin resistant enterococci. Adapted from Kelly CP, LaMont JT. Treatment of Clostridium difficile diarrhea and colitis. In: Wolfe MM, editor, Therapy of digestive disorders. Philadelphia: WB Saunders; 2000. p 513; Aslam S, Hamill RJ, Musher DM. Treatment of Clostridium difficile-associated disease: old therapies and new strategies. Lancet Infect Dis 2005; 5:549-57; and Zar FA, Bakkanagari SR, Moorthi KM, Davis MB. A comparison of vancomycin and metronidazole for the treatment of Clostridium difficileassociated diarrhea, stratified by disease severity. Clin Infect Dis 2007; 45:302-7.

tory.59,158-159 Treatment of asymptomatic carriers with antimicrobial agents against C. difficile is not recommended because it can prolong the carrier state.62 Many antimicrobial agents show activity against C. difficile in vitro,160-166 but resistance to cephalosporins is so widespread that cefoxitin is used in selective media to culture C. difficile.167 Clindamycin resistance is seen in some clinical isolates and has been associated with nosocomial outbreaks.168 There is increasing evidence of fluoro­ quinolone resistance among nosocomial C. difficile isolates, and the epidemic NAP1/BI strain that has caused several outbreaks shows high-level fluoroquinolone resistance.32,169 Fortunately, resistance to metronidazole is rare, and resistance to vancomycin is essentially nonexistent. In one study of 186 clinical isolates, all were sensitive to both metronidazole and vancomycin, with minimum inhibiting concentrations (MICs) of 0.5 to 4 mg/mL.167 In another series from Spain, 6% of 415 isolates showed intermediate sensitivity to metronidazole (MIC > 16 mg/mL),170 but this partial resistance pattern was not clonal and was not sustained in serial culture. These findings suggest an acquired tolerance rather than genetically determined metronidazole resistance. Many antimicrobial agents, such as ampicillin or amoxicillin, which have in vitro activity against C. difficile, are common causes of C. difficilea-ssociated diarrhea in clinical practice.161,171 This observation illustrates the fact that in vitro sensitivity testing alone is a poor predictor of therapeutic efficacy in this disease. The preclinical efficacy of therapeutic agents for C. difficile-associated diarrhea has been assessed in clindamycin-exposed Syrian hamsters.66,162,165,172,173 These animals develop a fulminant and fatal cecitis when exposed to toxigenic C. difficile after administration of clindamycin. Historically, this animal model has provided a reliable indication of the effectiveness of therapeutic agents for C. difficile infection and diarrhea. Specific antibiotic therapy to eradicate C. difficile is required in patients with severe symptoms or in those whose symptoms persist despite discontinuation of anti­ biotic treatment. The most effective antimicrobials for the treatment of C. difficile diarrhea are metronidazole (250 to 500 mg three or four times a day for 10 to 14 days) and vancomycin (125 to 500 mg four times a day for 10 to 14 days) (Table 108-6).174 Bacitracin, teicoplanin, and fusidic acid also have been used to treat acute infection but have

Chapter 108  Antibiotic- and Clostridium difficile-Associated Diarrhea and Colitis few if any advantages over metronidazole or vancomycin. In a systematic review, none of these alternative antibiotics were superior in terms of response rates.175 The advantages and disadvantages of specific therapeutic agents are discussed in the following sections. Metronidazole Metronidazole generally is recommended as the drug of first choice for mild to moderately severe C. difficile diarrhea and colitis.48,59,176 It is inexpensive ($0.50 per 250-mg tablet) and is usually effective. Several clinical studies before 2000 indicated that metronidazole therapy resulted in the resolution of diarrhea and colitis in more than 95% of patients treated.26,160,175 For example, in a prospective, randomized trial of acute C. difficile infection, metronidazole (250 mg four times a day for 10 days) was as effective as vancomycin (500 mg four times daily for 10 days) in terms of response and recurrence rates.157 Studies published between 2004 and 2007, however, report an average failure rate of 19% for metronidazole (range, 7% to 38%) compared with only 4% for vancomycin (range, 3% to 6%).48,177,178 In one trial, subjects with acute C. difficile infection were stratified according to disease severity and then randomized to receive either metronidazole 250 mg or vancomycin 125 mg each given four times per day. In mild disease, both treatments yielded similar response rates (90% and 98%, P = 0.36). In severe disease however, metronidazole was less efficacious (76% versus 97%, P = 0.02).178 Thus oral vancomycin should be used as the first-line agent in severe disease. Metronidazole, unlike vancomycin, is well absorbed in the upper intestine following oral administration. Fecal concentrations are low or absent in healthy persons or asymptomatic carriers of C. difficile, but higher fecal concentrations are observed in patients with C. difficile colitis because metronidazole is secreted through the inflamed intestinal mucosa.179 Intravenous metronidazole (500 mg four times per day) may be used in patients who cannot tolerate oral medication, because it accumulates in bactericidal levels within the inflamed colon.179 Oral metronidazole therapy usually is well tolerated but can be associated with systemic side effects.179 In one report of more than 600 patients receiving metronidazole for C. difficile diarrhea, only 1% experienced significant side effects.26 Adverse effects include nausea, a metallic taste, a disulfiram-like reaction with alcohol, and a peripheral sensory neuropathy with prolonged therapy. Metronidazole can potentiate the action of warfarin, resulting in prolongation of the prothrombin time. Metronidazole usually is not used in pregnant and nursing women because of chromosomal abnormalities, unknown effects on fetal organogenesis, and reports of tumorigenicity in rodents. Its safety in children has not been documented. Enigmatically, metronidazole has been identified as the antibiotic agent responsible for causing some cases of C. difficile diarrhea, demonstrating the importance of reduced colonization resistance in the pathophysiology of C. difficile-associated diarrhea.180-182 Vancomycin Vancomycin was introduced for treating C. difficileassociated diarrhea and colitis in 1978,183 and its phar­ macokinetic properties make it an ideal agent for treating this infection. When given orally, vancomycin is neither absorbed nor metabolized significantly, and as a result, high concentrations in the colonic lumen are achieved. The efficacy of vancomycin in treating C. difficile colitis has been demonstrated in controlled trials.51,157,177,178,183,184

Improvement in diarrhea usually is evident within 72 hours of initiating therapy, and complete resolution of symptoms occurs in most patients (96% overall) by the end of a 10-day treatment course.26,175,177 Fekety and coworkers185 demonstrated that vancomycin at a dose of 125 mg four times a day is as effective as vancomycin 500 mg four times a day. The lower dose is recommended for patients with mild to moderate colitis, and the higher dose is recommended for critically ill patients (i.e., those with ileus, colonic dilatation, or fulminant pseudomembranous colitis). Vancomycin may be administered by mouth, nasogastric tube, or even by enema,26,59 but it should not be given intravenously to treat C. difficile infection because effective colonic luminal concentrations are not obtained following parenteral administration.186,187 Oral vancomycin is not absorbed appreciably, and as a result, systemic side effects are rare. Despite its many advantages, oral vancomycin now is considered a secondline agent for the treatment of mild to moderately severe C. difficile infection because of its higher cost (a 10-day course can cost as much as $800) and concerns regarding the spread of vancomycin-resistant enterococci.176 Vancomycin therapy is recommended, however, for patients with severe infection and for patients who fail to respond to metronidazole, are intolerant of metronidazole, are pregnant, or are younger than 10 years.59,176 Other Antimicrobial Agents Bacitracin (25,000 units four times daily for seven to 10 days) is less effective than metronidazole or vancomycin for treating C. difficile diarrhea, with an overall response rate of only 80% and a relapse rate of 30%.184,188-190 In randomized therapeutic trials, teicoplanin, 100 mg twice a day for 10 days, was as effective as vancomycin for treating C. difficile diarrhea.191,192 Teicoplanin, however, is relatively expensive and is not available for oral administration in the United States. Fusidic acid has been tested in a limited number of patients but appears to be less effective than metronidazole or vancomycin and is associated with a relapse rate of approximately 28%.192,193 Treatment with colestipol, an ion exchange resin that binds toxins (10 g four times daily), is associated with a low response rate (36%) and is not recommended as primary therapy.51

Severe Pseudomembranous Colitis

Severe or fulminant pseudomembranous colitis occurs in only a minority of patients with C. difficile infection but is associated with a mortality rate of up to 65%.129,194,195 Many patients who develop fulminant C. difficile disease already have substantial comorbid disease and often are critically ill.196,197 Diarrhea may be minimal or absent because of ileus, and patients can present with abdominal pain, peritoneal signs, colonic dilatation, leukocytosis, and a clinical picture of progressive sepsis.88,157,198 Prompt diagnosis and aggressive therapy are necessary to avoid substantial morbidity and mortality. The first step is to discontinue precipitating antibiotics if possible and start therapy with high-dose oral vancomycin (500 mg four times per day), although there are no data to demonstrate that this higher dose is more effective than the standard dose of 125 mg four times per day.59 Intravenous metronidazole should be given if oral medication is not tolerated well. Intravenous vancomycin is not recommended, for the reasons mentioned earlier. In the presence of ileus, vancomycin (500 mg every six hours) may be administered via nasogastric tube with intermittent clamping of the tube.26 Intracecal infusion of vancomycin has been reported but is not recommended because of the risks

1899

1900

Section X  Small and Large Intestine associated with placement of a narrow-bore tube over a guidewire at colonoscopy in patients with severe, active colitis.199 For critically ill patients, a combination of antibiotics administered by various routes may be indicated. Six of eight patients with severe ileus were treated successfully using a combination of vancomycin administered by nasogastric tube, intravenous metronidazole, and vancomycinretention enemas (vancomycin 500 mg in 100 mL of normal saline administered every six hours via a Foley catheter inserted into the rectum). Patients treated with this regimen responded within 5 to 17 days.26 Passive immunization with pooled human immunoglobulin has been used empirically in patients with severe colitis who were not responsive to metronidazole or vancomycin. As discussed earlier, patients with severe or prolonged C. difficile diarrhea have low serum and fecal concentrations of antibody against C. difficile toxins.53,98,102,104-106 Intravenous infusion of normal pooled human immunoglobulin (400 mg/kg body weight) increases serum IgG antitoxin concentrations and has been used to treat patients with severe C. difficile colitis, although its efficacy has not been tested in controlled trials.105,200-202 Emergency surgery sometimes is required in patients with severe colitis not responding to medical therapy and in whom bowel perforation is impending or has already occurred.132,195,197,198,202 A surgical consultation should be sought immediately in patients with fulminant disease. The operation of choice usually is a subtotal colectomy with temporary ileostomy. Surgical intervention in this setting is associated with a high perioperative mortality rate, making the decision to operate difficult. Grundfest-Bronitowski and associates197 reported an overall mortality rate of 42% in a series of patients undergoing surgery for fulminant, severe C. difficile infection. In another series of five patients with toxic megacolon, subtotal colectomy and ileostomy were successful in only one patient.203

Recurrent Clostridium difficile Diarrhea

One of the most difficult clinical problems in treating patients with C. difficile infection is the high incidence of recurrences.59,204,205 Multiple episodes of recurrent C. difficile-associated diarrhea are not uncommon, and some patients have experienced more than 10 bouts of recurrence. Approximately 20% of patients successfully treated with vancomycin or metronidazole relapse after completing their initial antibiotic therapy.69,106,175,205,206 The clinical features of recurrence are similar to the initial attack, with watery diarrhea, cramping abdominal pain, or fever occurring 2 to 14 days after discontinuing therapy. Late recurrences are less common but can occur more than two months after stopping antibiotic treatment. The diagnosis of recurrent C. difficile-associated diarrhea is best confirmed by stool toxin assay whenever possible or in certain instances by colonoscopy and biopsy (see later). In patients with typical symptoms of recurrence, therapy can be reinstituted while awaiting stool assay results (Table 108-7). Prompt therapy is especially important in patients whose initial attack of C. difficile diarrhea was severe, because they are more likely to suffer from severe and recurrent disease, possibly because of their inadequate immune response to C. difficile toxins.65,106 Some patients with persistent symptoms following successful therapy of C. difficile infection develop diarrhea as a result of postinfection irritable bowel syndrome.207 Frequent watery diarrhea and cramping lower abdominal pain may be partially responsive to antibiotic therapy. Patients with post-C. difficile irritable bowel syndrome have normal

Table 108-7  Approach to Management of Recurrent Clostridium difficile Colitis First Relapse Confirm diagnosis Symptomatic treatment if symptoms are mild 10- to 14-day course of metronidazole if symptoms are moderate 10- to 14-day course of vancomycin if symptoms are severe Second Relapse Confirm diagnosis Vancomycin-taper regimen 125 mg every 6 hr for 10 to 14 days 125 mg every 12 hr for the next seven days 125 mg daily for the next seven days 125 mg every other day for the next eight days 125 mg every three days for the next 15 days Third Relapse 10- to 14-day course of vancomycin followed by a 14-day course of oral rifaximin 400 mg twice a day Additional Options Therapy with microorganisms, e.g., bacteriotherapy, Saccharomyces boulardii, or Lactobacillus spp. in combination with and following metronidazole or vancomycin or Intravenous immunoglobulin 400 mg/kg two or three times with a three-week interval between doses or Vancomycin 125 mg every 6 hr plus cholestyramine 4 g twice daily* or Vancomycin 125 mg every 6 hr and rifampicin 600 mg twice daily *Because cholestyramine binds vancomycin, oral doses of these two agents must be separated by two to three hours, making this regimen difficult to implement. Adapted from Linevsky JK, Kelly CP. Clostridium difficile colitis. In: Lamont JT, editor. Gastrointestinal Infections: Diagnosis and management. New York: Taylor & Francis; 1997. p 293.

colonoscopy and biopsy, and their stools are negative for toxin. Other diarrheal conditions that require differentiation from recurrent C. difficile infection include IBD, microscopic colitis, celiac disease, and food (e.g., lactose) intolerance. Bacteriologic typing studies demonstrate that symptomatic recurrence can result from reinfection with either the same strain that caused the initial episode or a different strain of C. difficile.208,209 Resistance to metronidazole or to vancomycin is seldom if ever an important factor in recurrence. For example, Bartlett and colleagues were unable to demonstrate in vitro vancomycin resistance in 23 isolates of C. difficile from relapsing patients.210 In some patients, C. difficile can be cultured from the stools during successful vancomycin therapy, and these patients may be more likely to relapse than those in whom eradication of the pathogen occurs during therapy208; however, C. difficile can also be cultured from the stools during and after successful antibiotic treatment in patients who do not relapse.134 Culture positivity during symptomatic improvement might reflect the persistence of antibiotic-resistant spores. In one study, 18 of 22 patients with recurrence were noted to have colonic diverticula, leading to the speculation that spores might survive in diverticula where they escape the normal cleansing action of diarrhea and might not be exposed to the high luminal concentration of antibiotics204; however, reinfection by bacterial spores through the usual fecal-oral route is a more likely mechanism of recurrence.208,209 In a study of 569 patients who had a positive C. difficile assay, 135 of whom had diverticulosis and 434 who did not, there was a non–statistically significant trend to a higher rate of recurrence in those with diverticulosis (15.6% vs. 12.0%).211

Chapter 108  Antibiotic- and Clostridium difficile-Associated Diarrhea and Colitis Conservative Therapy In a report of 20 patients with clindamycin-associated pseudomembranous colitis, published before the discovery of vancomycin as effective therapy, all patients eventually recovered when clindamycin was stopped.212 An important advantage to this form of management is that recurrence of diarrhea or colitis does not occur, probably because stopping all antimicrobial agents allows restoration of the colonic microflora and C. difficile colonization resistance. Many patients with mild symptoms of recurrence can be managed conservatively without specific antibiotic treatment, thereby avoiding subsequent recurrences; this approach might not be appropriate for elderly or infirm patients and is not advised for those with moderate or severe symptoms. Repeat Treatment with Vancomycin or Metronidazole Patients with recurrence typically are treated with a second course of the same antibiotic used to treat the initial attack, but treatment is usually for 14 days; this has a success rate of about 40%. Patients with one recurrence have a substantial risk of further recurrences, and in two inde­ pendent studies, patients with one or more previous recurrences had a subsequent recurrence rate of greater than 50% following standard therapy with metronidazole or vancomycin.69,106 Prolonged or Tapering and Pulsed Vancomycin Therapy Tedesco and colleagues204 treated 22 patients who had multiple recurrences of C. difficile colitis using tapering doses of vancomycin for a three-week period, followed by every-other-day therapy for one week and every third day for an additional week. All patients responded symptomatically and remained well during a mean follow-up period of six months. Although data from randomized, controlled trials are not available, one subsequent study that examined various physician-selected antibiotic regimens to treat recurrent C. difficile infection found that regimens incorporating prolonged or pulsed-dose oral vancomycin were the most effective.213 Overall, 73 of 163 patients (45%) treated for recurrent C. difficile infection had a subsequent recurrence. Of all the regimens used, only those incorporating prolonged-dose vancomycin (9 of 29 recurred [31%], P = 0.01) or pulsed-dose vancomycin (one of seven recurred [14%], P = 0.02) showed significantly lower recurrence rates. The mechanism whereby this treatment approach is effective is unknown and might simply reflect prolonged treatment. Toxin production by C. difficile does not occur during the early exponential growth phase of the bacterium but rather in the subsequent stationary phase.72,75 Hence, after active C. difficile toxin-induced diarrhea and colitis have been controlled by treatment with vancomycin, a period of 24 to 72 hours is needed for the bacteria to reinitiate production and release of toxin. Thus, pulsed dosing might prevent toxin production and release while also facilitating restoration of resistance to colonization. Binding Resins Binding resins, which bind to toxins in the bowel lumen, have been proposed as a possible alternative to antimicrobial therapy. Clinical studies have been performed using colestipol, cholestyramine, and tolevamer. For colestipol, the symptomatic response in patients with acute C. difficile colitis was a disappointing 36%, compared with a placebo response rate of 22%.51,175 Cholestyramine therapy yielded

a somewhat better overall response rate of 68%,51,175,214 but this still compares poorly with response rates of more than 95% with vancomycin or metronidazole. Therefore, binding resins are not used as primary therapy for C. difficile colitis but may be beneficial in treating recurrence as an adjunctive agent (see below). Tedesco treated 11 patients who had relapsing C. difficile colitis with tapering doses of vancomycin plus colestipol 5 g every 12 hours.215 Because anion-exchange resins bind vancomycin and other drugs, they must be taken at least two or three hours apart from the vancomycin, making such combination therapy cumbersome. Tolevamer is a soluble anionic polymer specifically developed to bind C. difficile toxins. In preclinical studies, tolevamer strongly inhibited the cytotoxicity and enterotoxicity of C. difficile toxins and was superior to metronidazole in protecting hamsters from death caused by C. difficile cecitis.66 In a phase II human clinical trial, results with tolevamer were similar to those of vancomycin when used as primary treatment for mild or moderately severe infection.67 In two larger phase III studies, however, response rates with tolevamer were substantially lower than with either vancomycin or metronidazole.216 Interestingly, recurrence rates after successful tolevamer treatment were substantially lower than after antibiotic treatment with metronidazole or vancomycin, suggesting its potential use for primary or secondary disease prevention.67 Probiotic Therapy In contrast to treatment with antimicrobial agents that further delay recolonization by normal colonic bacteria, probiotic agents are an attractive therapeutic option for recurrent C. difficile infection, because restoration of colonization resistance can lead to permanent eradication of C. difficile from the colon. Bacteriotherapy has been reported in patients with recurrent infection using enemas of fresh feces from a healthy relative, administration of fecal filtrates via a nasogastric tube, or by rectal infusions of a mixture of 10 different aerobic and anaerobic bacteria.217-219 The defined bacterial mixture led to bowel colonization with Bacteroides species, as well as prompt elimination of C. difficile, suggesting that Bacteroides may be one of the organisms that normally protects against pathogenic colonization with C. difficile. Another probiotic therapy for C. difficile diarrhea is the oral administration of a nontoxigenic strain of C. difficile that was reported to be effective in two patients with relapsing C. difficile diarrhea.220 Preclinical studies are under way to characterize a nontoxigenic strain of C. difficile that may be suitable for administration as a prophylactic agent to prevent infection with toxigenic C. difficile in hospital patients who are receiving antibiotics.68 Lactobacillus species have been used widely as pro­ biotics. In an open-label study, Lactobacillus strain GG was reported to be effective in preventing diarrhea in patients with recurrent C. difficile colitis.221 A subsequent controlled clinical trial, however, did not demonstrate that Lactobacillus GG was effective in protecting against AAD in hospital patients.222 Another placebo-controlled study of hospital patients receiving antibiotics used a probiotic drink mixture containing Lactobacillus casei, Lactobacillus bulgaricus, and Streptococcus thermophilus (DanActive) and found that simple AAD was reduced from 34% to 12% (placebo versus active; P = 0.007), and C. difficile infection was reduced from 17% to 0% (placebo versus active; P = 0.001).223 These dramatic positive results need to be confirmed in additional, multicenter studies.

1901

1902

Section X  Small and Large Intestine The yeast S. boulardii is used widely as a probiotic agent in continental Europe and is now available in the United States without prescription.10,224 In a double-blind, controlled clinical trial, coadministration of oral capsules containing viable S. boulardii with antibiotics significantly reduced the incidence of AAD in hospitalized patients (from 22% on placebo to 9.5% in the S. boulardii group; P = 0.04).10 In that study, however, few patients had C. difficile-associated diarrhea. A second randomized, placebo-controlled trial examined the efficacy of S. boulard­ii in combination with either vancomycin or metroni­dazole in patients with C. difficile diarrhea.69 Diarrhea recurrence rates were similar in subjects treated during their first episode of C. difficile diarrhea (19% in the S. boulardii group vs. 24% in the placebo group; P = 0.86). In contrast, patients with a history of recurrent C. difficile diarrhea who received S. boulardii had fewer recurrences than the placebo group (35% and 65%, respectively; P = 0.04). In a subsequent study, S. boulardii (500 mg twice daily for 28 days) only reduced relapse rates (from 50% to 17%; P = 0.05) in patients treated with high-dose vancomycin (500 mg four times a day for 10 days) but not in patients on other antibiotic regimens.225 These controlled clinical trials indicate that S. boulardii is safe and effective in some patients with a history of recurrent C. difficile-associated diarrhea, but its protective effects are not uniform. S. boulardii should not be administered to immunocompromised patients because of the risk of fungemia. Bacteriotherapy It has long been recognized that alteration of the bacterial flora of the colon by antibiotics is the major predisposing cause of C. difficile infection, and that restoration of the normal flora eventually eliminates this pathogen. Based on this knowledge, measures to hasten reconstitution of the normal fecal bacteria are an attractive and logical mode of treatment. Therapy with microorganisms including stool transplantation and colonization with nontoxigenic strains of C. difficile have been reported. Several reports have documented efficacy of stool transplants from healthy relatives and intimate contacts delivered by colonoscopy or through a nasogastric tube in recurrent 217,225 and acute226 C. difficile infection. However, concern for the inadvertent spread of pathogenic microorganisms from donor to host as well as low patient and physician palatability for the procedure has limited the widespread acceptance of this approach. Oral administration of nontoxigenic strains of C. difficile to compete with and limit the growth of toxin-producing strains also has been reported to lead to remission of recurrent C. difficile infection in two patients.220 Immunization Against Clostridium difficile Toxins As described earlier, there is considerable evidence that some persons have protective immunity against C. difficileassociated diarrhea and that protection is associated with higher antitoxin antibody concentrations in serum, intestinal secretions, or both.53,64,65,102,104-106 Leung and coworkers reported on six children with multiple relapses of C. dif­ficile-associated diarrhea who had low concentrations of serum IgG antibody to toxin A.105 Five of these children were treated with intravenous immune globulin at a dose of 400 mg/kg, which contained high-titer IgG antitoxin A. Symptoms resolved following treatment. Similar results have been reported by other investigators.201 A C. difficile vaccine has been produced containing inactivated toxoid A and B. In early clinical trials, this vaccine was immunogenic,227,228 and in a small case series, vaccination was associated with resolution of recurrent C. difficile

diarrhea in three subjects.229 Further studies are needed to determine whether passive or active immunization against C. difficile and its toxins can be effective in treating patients with refractory or recurrent disease. If effective, these therapeutic approaches also may be useful in preventing C. difficile-associated diarrhea in high-risk persons, such as elderly and infirm patients receiving antibiotic therapy in the hospital. Overall Approach The management of a first episode of recurrent C. difficileassociated diarrhea does not differ greatly from treatment of an initial episode (Table 108-7).131 Stool samples should be obtained to reconfirm infection with toxigenic C. difficile. Patients with mild symptoms of recurrence may be able to be managed conservatively without additional antibiotic treatment, just like patients with a primary episode. If symptoms persist or are severe, a 14-day course of metronidazole or vancomycin should be administered. If a second recurrence occurs, other treatment approaches should be considered. Tedesco and associates proposed a tapering and pulsed antibiotic regimen that is well tolerated and often successful.204,212 In two recent case series, treatment with oral rifaximin 600 mg to 1200 mg daily for two to four weeks after completion of a standard course of vancomycin was effective in preventing recurrence in 12 of 14 patients with multiple previous recurrences.230,231 If this method fails, a wide range of other approaches have been described, some of which are summarized in Table 108-7. Unfortunately, with the exception of the probiotic agent S. boulardii, none of these treatment options has been evaluated in randomized, controlled trials. In some instances, multiple recurrences develop, and a variety of different regimens must be used before the organism is finally eradicated. In such cases, prolonged therapy with oral vancomycin (125 mg once or twice daily) is a pragmatic and effective means to prevent further recurrences. This approach is indicated in high-risk patients in whom symptomatic recurrence might be life-threatening or in patients in whom other measures have failed consistently, including multiple antibiotic regimens and probiotics.

KEY REFERENCES

Aslam S, Hamill RJ, Musher DM. Treatment of Clostridium difficile– associated disease: Old therapies and new strategies. Lancet Infect Dis 2005; 5:549-57. (Ref 177.) Bartlett JG, Chang TW, Gurwith M, et al. Antibiotic-associated pseudomembranous colitis due to toxin-producing clostridia. N Engl J Med 1978; 298:531-4. (Ref 20.) Fekety R. Guidelines for the diagnosis and management of Clostridium difficile–associated diarrhea and colitis. American College of Gastroenterology, Practice Parameters Committee. Am J Gastroenterol 1997; 92:739-50. (Ref 59.) Kelly CP, LaMont JT. Clostridium difficile: more difficult than ever. N Engl J Med 2008; 359; 1932-40. (Ref 48.) Kyne L, Warny M, Qamar A, Kelly CP. Asymptomatic carriage of Clostridium difficile and serum levels of IgG antibody against toxin A. N Engl J Med 2000; 342:390-7. (Ref 53.) Kyne L, Warny M, Qamar A, Kelly CP. Association between antibody response to toxin A and protection against recurrent Clostridium difficile diarrhoea. Lancet 2001; 357:189-93. (Ref 106.) McDonald LC, Owings M, Jernigan DB. Clostridium difficile infection in patients discharged from US short-stay hospitals, 1996 to 2003. Emerg Infect Dis 2006; 12:409-15. (Ref 28.) McDonald LC, Killgore GE, Thompson A, et al. An epidemic, toxin gene-variant strain of Clostridium difficile. N Engl J Med 2005; 353:2433-41. (Ref 169.) McFarland LV, Mulligan ME, Kwok RY, Stamm WE. Nosocomial acquisition of Clostridium difficile infection. N Engl J Med 1989; 320:20410. (Ref 52.)

Chapter 108  Antibiotic- and Clostridium difficile-Associated Diarrhea and Colitis Mylonaki M, Langmead L, Pantes A, et al. Enteric infection in relapse of inflammatory bowel disease: Importance of microbiological examination of stool. Eur J Gastroenterol Hepatol 2004; 16:775-8. (Ref 121.) Pepin J, Valiquette L, Alary ME, et al. Clostridium difficile–associated diarrhea in a region of Quebec from 1991 to 2003: a changing pattern of disease severity. CMAJ 2004; 171:466-72. (Ref 156.) Tedesco FJ, Gordon D, Fortson WC. Approach to patients with multiple recurrences of antibiotic-associated pseudo-membranous colitis. Am J Gastroenterol 1985; 80:867-8. (Ref 204.) Viscidi R, Laughon BE, Yolken R, et al. Serum antibody response to toxins A and B of Clostridium difficile. J Infect Dis 1983; 148:93-100. (Ref 101.)

Warny M, Pepin J, Fang A, et al. Toxin production by an emerging strain of Clostridium difficile associated with outbreaks of severe disease in North America and Europe. Lancet 2005; 366:1079-84. (Ref 76.) Zar FA, Bakkanagari SR, Moorthi KM, Davis MB. A comparison of vancomycin and metronidazole for the treatment of Clostridium difficile–associated diarrhea, stratified by disease severity. Clin Infect Dis 2007; 45:302-7. (Ref 178.) Full references for this chapter can be found on www.expertconsult.com.

1903

CHAPTE R

109 Intestinal Protozoa Christopher D. Huston

CHAPTER OUTLINE Entamoeba histolytica  1905 Epidemiology  1905 Pathogenesis, Pathology, and Immunology  Clinical Features  1908 Diagnosis  1909 Treatment  1910 Control and Prevention  1910 Other Intestinal Amebae  1911 Giardia intestinalis  1911 Epidemiology  1911 Pathogenesis, Pathology, and Immunology  Clinical Features  1913 Diagnosis  1913 Treatment  1914 Control and Prevention  1914 Dientamoeba fragilis  1914 Blastocystis hominis  1914 Cryptosporidium Species  1914 Epidemiology  1914 Pathogenesis, Pathology, and Immunology  Clinical Features  1915 Diagnosis  1915 Treatment  1915 Control and Prevention  1916 Cyclospora cayetanensis  1916 Epidemiology  1916 Pathogenesis, Pathology, and Immunology 

1907

1913

1915

1916

The intestinal protozoa traditionally have been considered important pathogens in the developing world, where food and water hygiene are poor. A basic knowledge of the intes­ tinal protozoa that cause human disease is of growing importance to physicians practicing medicine in the United States, Canada, and Europe, however, as a result of increas­ ing world travel, globalization of the world’s economy, and the growing number of chronically immunosuppressed people. For example, in patients with the acquired immuno­ deficiency syndrome (AIDS) and organ transplant recipi­ ents, microsporidia, Cryptosporidium species, Isospora belli, and Cyclospora cayetenensis are the leading causes of chronic diarrhea worldwide. Cryptosporidium species, I. belli, and C. cayetenensis have been recognized as common pathogens in immunocompetent persons as well, and foodand water-borne outbreaks in the United States and Canada raise questions about the safety of our increasingly complex food and water supplies. Our understanding of the biology of these organisms often is still rudimentary, but is rapidly changing; for example, it has only recently been recognized that Entamoeba histolytica, the cause of amebic dysentery, and the nonpathogenic intestinal ameba Entamoeba dispar are distinct species; and the Cryptosporidium species of medical importance were reclassified in 2002.

Clinical Features  1916 Diagnosis  1916 Treatment  1916 Control and Prevention  1916 Isospora belli  1917 Epidemiology  1917 Pathogenesis, Pathology, and Immunology  1917 Clinical Features  1917 Diagnosis  1917 Treatment  1917 Control and Prevention  1917 Microsporidia  1917 Epidemiology  1917 Pathogenesis, Pathology, and Immunology  1917 Clinical Features  1917 Diagnosis  1917 Treatment  1918 Control and Prevention  1918 Trypanosoma cruzi (American Trypanosomiasis or Chagas’ Disease)  1918 Epidemiology  1918 Pathogenesis, Pathology, and Immunology  1918 Clinical Features  1918 Diagnosis  1918 Treatment  1919 Control and Prevention  1919

The emergence of these pathogens as major causes of disease in the developed world has stimulated a growing number of basic science studies of parasite biology and rapid development of new diagnostic tests, treatments, and attempts at vaccination. This chapter summarizes major recent advances in our understanding of the intestinal protozoa, with an emphasis on clinical epidemiology, disease characteristics, and optimal approaches to accurate diagnosis, and treatment.

ENTAMOEBA HISTOLYTICA EPIDEMIOLOGY

Entamoeba histolytica was first linked causally to amebic colitis and liver abscess by Lösch in 1875, and it was named by Schaudinn in 1903 for its ability to destroy host tissues. In 1925, Emil Brumpt proposed the existence of a second, morphologically identical but nonpathogenic Entamoeba species, Entamoeba dispar, to explain why only a minority of people infected with what was then termed E. histolytica develop invasive disease. Although Brumpt’s hypothesis was not accepted during his lifetime, it is now clear that

1905

1906

Section X  Small and Large Intestine he was correct and E. histolytica (Schaudinn, 1903) was recently reclassified to include two morphologically indis­ tinguishable species: E. histolytica, the cause of invasive amebiasis, and E. dispar, a nonpathogenic intestinal com­ mensal parasite (see later section).1 Entamoeba histolytica is a parasite of global distribution, but most of the morbidity and mortality from amebiasis occurs in Central and South America, Africa, and the Indian subcontinent.2 Fortunately, the majority of the 500 million persons worldwide previously believed to be asymptomatic E. histolytica cyst passers are actually infected with E. dispar. The best current estimate is that E. histolytica causes 34 to 50 million symptomatic infections annually world­ wide, resulting in 40,000 to 100,000 deaths each year.3,4 In Dhaka, Bangladesh, where diarrheal diseases are the leading cause of childhood death, 80% of children studied prospec­ tively were infected with E. histolytica at least once during four years of follow-up.5 Furthermore, E. histolytica–associ­ ated diarrhea in these children was associated with signifi­ cantly low weight and height for age.6 E. histolytica has a simple, two-stage life cycle consisting of an infectious cyst and a motile trophozoite (Fig. 109-1). The cyst form measures 5 to 20 µm in diameter and contains four or fewer nuclei. The ameboid trophozoite, which is responsible for tissue invasion, measures 10 to 60 µm (Fig. 109-2) and contains a single nucleus with a central karyosome (Fig. 109-3). The cysts are relatively resistant to

chlorination and desiccation, and they can survive in a moist environment for several weeks. Infection occurs following ingestion of cysts in fecally contaminated food or water. Within the lumen of the small intestine, the quadrinucleate cyst undergoes nuclear then cytoplasmic division, giving rise to eight trophozoites.7 Only about 10% of infected persons develop invasive disease characterized by invasion of the colonic epithelium by tro­ phozoites.1 Trophozoites that gain access to the bloodstream can spread hematogenously to establish infection at distant sites (most commonly liver abscess, as discussed in Chapter 82). Why some persons develop invasive disease and others remain asymptomatic remains a mystery; parasite and host differences are likely to be important in this regard. A molecular epidemiologic study that used the polym­ erase chain reaction (PCR) to amplify a polymorphic region of the E. histolytica genome and assign a genotype to differ­ ent clinical isolates has demonstrated a correlation between different E. histolytica strains and the outcome of infection.8 The specific underlying genetic differences among ameba strains that are responsible for altered virulence, however, remains unknown. Furthermore, amebic liver abscess is pri­ marily a disease of men, and studies suggest that suscepti­ bility to both intestinal and hepatic amebiasis is linked to human leukocyte antigen (HLA) class II alleles.9,10 For example, the HLA DQB1*0601 allele may be associated with protection from intestinal amebiasis.9 As is the case for

Liver

Excystation occurs in the small intestine, following exposure to gastric acid and intestinal proteases, releasing trophozoites

Hematophagous trophozoites invade the colon or migrate to the liver or elsewhere, attracting PMNs and lysing cells on contact causing:

Colon

Possible outcomes Amebic colitis with flask-shaped ulcers Liver abscess Other abscesses (brain, etc.) Ameboma

Cysts are ingested in fecally contaminated food or water Hematophagous trophozoites and hardy cysts are passed in the stool Figure 109-1.  Life cycle of Entamoeba histolytica. PMNs, polymorphonuclear neutrophils. (From Petri WA, Sing U, Ravdin JI. Enteric amebiasis. In: Guerrant RL, Walker DH, Weller PF, editors. Tropical Infectious Diseases: Principles, Pathogens, and Practice. Philadelphia: WB Saunders; 1999.)

Chapter 109  Intestinal Protozoa

Entamoeba histolytica

Human pathogen

Estimated frequency

Yes

1–10%

Trophozoite morphology and usual size range in µm

10–20 (Range: 10–60) E. coli

No

Cyst morphology and usual size range in µm

5–20

3–20%

10–30 15–25 (Range: 10–50) E. hartmanni

No

?

No

Central punctate karyosome, erythrophagocytosis Indistinguishable from E. dispar and E.moshkovskii Large size, 5 to 8 nuclei; splinter-like chromatoid bodies Eccentric karyosome distinguishes this trophozoite from E. histolytica and E. dispar Small size

4–10

<10 E. gingivalis

Characteristic features

10–90% (mouth)

none

Oral trophozoite only

12–14

Uninucleate cyst with large karyosome May represent multiple species

15 (Range: 3–35) E. polecki

Uncertain

rare

16–18 Endolimax nana

No

10–33%

Vesiculate nucleus 6–10 8–12

Iodamoeba butschlii

No

“I” cyst (see text)

5–8% 6–15 9–20

Dientamoeba fragilis* Uncertain

none

4–10%

Binucleate trophozoites with a connecting thread

4–12 *Dientamoeba fragilis was initially classified as an ameba, but it is more closely related to the flagellates (trichomonads) based on morphologic studies and phylogenetic analyses. Figure 109-2.  Amebae that infect the human gastrointestinal tract. E., entamoeba. (From Ravdin Jl, Guerrant RL. Current problems in the diagnosis and treatment of amebic infections. Curr Clin Trop Infect Dis 1986; 7:82.)

genetic differences among E. histolytica strains, however, there is no evidence of a direct causal role for different HLA types; rather, these HLA types are likely to be in linkage disequilibrium with genes in the nearby vicinity that encode the causal factors.

PATHOGENESIS, PATHOLOGY, AND IMMUNOLOGY

Both amebic factors and the host’s inflammatory response contribute to tissue destruction during invasive amebiasis. Microscopy studies have defined a stepwise progression of disease.11-13 After excystation within the lumen of the

small intestine, trophozoites adhere to colonic mucins and epithelial cells, largely via an amebic galactose/N-acetyl-dgalactosamine inhibitable surface lectin.14-16 Secreted cyste­ ine proteinases then facilitate tissue invasion by degrading human colonic mucus and extracellular matrix proteins.17-20 Further disruption of the colonic epithelium results directly from contact-dependent cytolysis of epithelial and immune cells and from an acute epithelial cell inflammatory response with recruitment of neutrophils and immune-mediated tissue damage.14,21-25 The cecum and ascending colon are affected most com­ monly, although in severe disease the entire colon may be

1907

1908

Section X  Small and Large Intestine

A

B

Figure 109-3.  A, An Entamoeba histolytica trophozoite in a stool specimen. Note the nucleus with a prominent central karyosome. B, Giardia intestinalis cyst in stool. (Original magnification ×400.)

involved. On gross examination, pathology can range from mucosal thickening to multiple punctate ulcers with normal intervening tissue (Fig. 109-4) to frank necrosis. For unknown reasons, the downward invasion of amebic tro­ phozoites often is halted at the level of the muscularis mucosa. Subsequent lateral spread of amebae undermines the overlying epithelium, resulting in the clean-based, flaskshaped ulcers that characterize amebic colitis.26,27 Early in infection, an influx of neutrophils is typical, but in wellestablished ulcers, few inflammatory cells are seen.13,26-28 Organisms may be seen ingesting red blood cells (erythro­ phagocytosis) (Fig. 109-5). At distant sites of infection (e.g., liver abscess), similar pathologic characteristics include central liquefaction of tissue surrounded by a minimal mononuclear cell infiltrate.27-29 Because more than 90% of persons colonized with E. histolytica spontaneously clear the infection within a year, an effective immune response to amebiasis seems to develop.30 Children with fecal anti-amebic lectin immuno­ globulin (Ig)A have short-lived protection from subsequent intestinal infection.5,31,32 The protective role of secretory IgA is not certain, however, and the contributions of humoral and cellular immunity to protection from amebiasis remain unknown. Nearly everyone with invasive amebiasis devel­ ops a systemic and a mucosal humoral immune response.33-38 Antibodies alone are unable to clear established infection, however, because asymptomatic cyst passers remain infected for months after anti-amebic antibodies develop.30,33 Passive immunization experiments in a severe combined immunodeficient (SCID) mouse model of liver abscess do suggest an important role for preexisting humoral immunity in protection from infection.39 Reports that patients receiv­ ing glucocorticoids may be at increased risk for severe amebic colitis suggest that cellular immunity also plays an important role in control of E. histolytica infection40,41; despite this concern, no increase in disease severity in patients with AIDS has been observed. In fact, in a mouse model of amebic colitis, disease was exacerbated by CD4+ T cells.42

CLINICAL FEATURES

Infection with E. histolytica results in one of three out­ comes. Approximately 90% of infected persons remain asymptomatic. The other 10% of infections result in inva­

Figure 109-4.  Colonoscopic findings in a patient with amebic colitis. Multiple punctate ulcers are visible.

Figure 109-5.  Amebic colitis. This high-power view of a colon biopsy specimen shows multiple amebic trophozoites, many of which have ingested red blood cells (erythrophagocytosis). Nonpathogenic ameba do not exhibit erythrophagocytosis. (From the photo collection of the late Harrison Juniper, MD.)

Chapter 109  Intestinal Protozoa Table 109-1  Comparison of Amebic Colitis and Invasive Bacterial Dysentery FEATURE

AMEBIC COLITIS

BACTERIAL DYSENTERY*

Travel to or from an endemic area Usual duration of symptoms Diarrhea Fecal occult blood Abdominal pain Weight loss Fever >38°C

Yes >7 days 94-100% 100% 12-80% Common Minority

Sometimes 2-7 days 100% 40% ~50% Unusual Majority

*See Chapter 107. Adapted from Huston CD, Petri WA. Amebiasis. In: Rakel RE, Bope ET, editors. Conn’s Current Therapy, 2001. Philadelphia: WB Saunders; 2001. pp 50-4.

sive amebiasis characterized by dysentery (amebic colitis) or, in a minority of cases, extraintestinal disease (most com­ monly amebic liver abscess; see Chapter 82).1,30 In the United States, immigrants from or travelers to endemic regions, male homosexuals, and institutionalized persons are at greatest risk for amebiasis. In addition, mal­ nourished patients, infants, the elderly, pregnant women, and patients receiving glucocorticoids may be at increased risk for fulminant disease.2,40,41 When one or more of these epidemiologic risk factors are present, amebic dysentery should be considered in the differential diagnosis of occult or grossly bloody diarrhea. The major diagnostic challenge for the clinician seeing a patient with amebic colitis is to distinguish the illness from other causes of bloody diarrhea. The differential diagnosis includes the causes of bacterial dysentery, such as Shigella, Salmonella, and Campylobacter species and enteroinvasive or enterohemorrhagic Escherichia coli, and noninfectious diseases, including inflammatory bowel disease, and isch­ emic colitis.2,43 In contrast to bacterial dysentery, which typically begins abruptly, amebic colitis begins gradually over one to several weeks (Table 109-1). Although more than 90% of patients with amebic colitis present with diar­ rhea, abdominal pain can occur without diarrhea; abdomi­ nal pain, tenesmus, and fever are highly variable. Weight loss is common because of the chronicity of the illness. Microscopic blood is present in the stool of most patients with amebic dysentery.2,43,44 The most feared complication of amebic dysentery, acute necrotizing colitis with toxic megacolon, occurs in 0.5% of cases. This complication manifests as an acute dilatation of the colon, and 40% of patients die from sepsis unless it is promptly recognized and treated surgically.45,46 Unusual complications include the formation of enterocutaneous, rectovaginal, and enterovesicular fistulas and ameboma. Ameboma, due to intraluminal granulation tissue, can cause bowel obstruction and mimic carcinoma of the colon.2,43 Although a history of dysentery early in the illness is common, dysentery has resolved in most patients by the time of presentation.47-49 Extraintestinal sites of infection are involved and typically result either from direct extension of liver abscesses (e.g., amebic pericarditis or lung abscess) or from hematogenous spread of disease (e.g., brain abscess).2,50

DIAGNOSIS

Because amebiasis patients erroneously treated for inflam­ matory bowel disease with glucocorticoids can develop fulminant colitis, accurate initial diagnosis is critical.40,41 The gold standard for diagnosis of amebic colitis remains colonoscopy with biopsy, and colonoscopy should be per­ formed whenever infectious causes of bloody diarrhea are strong considerations in the differential diagnosis of ulcer­ ative colitis. Because the cecum and ascending colon are affected most often, colonoscopy is preferred to sigmoidos­

copy. Classically, multiple punctate ulcers measuring 2 to 10  mm are seen with essentially normal intervening tissue (see Fig. 109-4); however, the colonic epithelium might simply appear indurated with no visible ulcerations; appear like ulcerative colitis with a myriad of ulcerations and granular, friable mucosa; or, in severe cases where the ulcers have coalesced, the epithelium may appear necrotic. Histologic examination of a biopsy specimen taken from the edge of an ulcer reveals amebic trophozoites and a variable inflammatory infiltrate (see Fig. 109-5).27 Identi­ fication of amebae can be aided by periodic acid– Schiff staining of biopsy tissue, which stains trophozoites magenta. Stool examination for ova and parasites, the traditional method for diagnosing amebiasis, should not be relied upon. Although the presence of amebic trophozoites with ingested erythrocytes strongly correlates with E. histolytica infection, these rarely are present,51 and in the absence of hematophagous trophozoites, microscopy cannot distin­ guish E. histolytica from E. dispar. Difficulty in distinguish­ ing other nonpathogenic amebae (see later) and white blood cells from E. histolytica also limits the specificity of stool microscopy.52 The sensitivity of microscopy for identifica­ tion of amebae is at best 60% and it may be reduced by delays in processing of stool samples.52,53 The primary utility of stool microscopy for ova and parasites in a patient with diarrhea, therefore, is to evaluate the stool for other parasitic causes of diarrhea. Noninvasive methods to accurately differentiate E. histolytica from E. dispar include stool culture with isoenzyme analysis, serum amebic-antibody titers, PCR, and an enzymelinked immunosorbent assay (ELISA) that detects the amebic lectin antigen in stool samples.54-64 Of these, only serum amebic-antibody titers and the stool ELISA are widely avail­ able for clinical use. Because serum anti-amebic antibodies do not develop in patients infected with E. dispar, serologic tests for amebiasis accurately distinguish E. histolytica and E. dispar infection. From 75% to 85% of patients with acute amebic colitis have detectable anti-amebic antibodies on presentation, and con­ valescent titers develop in more than 90% of patients.34,35,65 For amebic liver abscess, 70% to 80% of patients have detectable antibody titers on presentation, and convalescent titers develop in more than 90% of patients. Because anti­ amebic antibodies can persist for years, however, a positive result must be interpreted with caution.34 For persons with known epidemiologic risks (e.g., emigration from or prior travel to an endemic region), a positive result might simply represent infection in the distant past. In the setting of recent travel to an endemic region and a positive antibody titer, diagnosis is confirmed by an appropriate symptomatic response to anti-amebic treatment. The most specific clinically available test for diagnosis of amebiasis is a stool ELISA to detect the E. histolytica adher­

1909

1910

Section X  Small and Large Intestine ence lectin. Only one of the many ELISA tests developed thus far (the E. histolytica II test, TechLab, Blacksburg, Va.) accurately distinguishes E. histolytica from E. dispar.53,60,61 This test’s specificity, when compared with the gold stan­ dard of stool culture followed by isoenzyme analysis, was greater than 90%, and it was greater than 85% sensitive for diagnosis of intestinal amebiasis when fresh fecal samples were analyzed without delay.61 In other studies, the sensi­ tivity of this method has been less impressive, emphasizing the need for rapid processing of stool samples.66,67 It also may be possible to use this antigen detection test to diag­ nose amebic liver abscess, because before treatment is initi­ ated, amebic lectin antigen can be detected in the serum of greater than 90% of patients who have amebic liver abscess.68

TREATMENT

Drugs for treatment of amebiasis are categorized as luminal or tissue amebicides on the basis of the location of their anti-amebic activity (Table 109-2). The luminal amebicides include iodoquinol, diloxanide furoate, and paromomycin.69,70 Of these, paromomycin, a nonabsorbable aminoglycoside, is preferred because of its safety, short duration of required treatment, and superior efficacy. Its major side effect is diarrhea. Approximately 85% of asymptomatic patients are cured with one course of paromomycin, and, because it is nonabsorbable and has moderate activity against trophozoites that have invaded the colonic mucosa, it might also be useful for single-drug treat­ ment of mild invasive disease during pregnancy.71,72 The tissue amebicides include metronidazole, tinidazole, nitazoxanide, erythromycin, and chloroquine.70,73 Of these, metronidazole and tinidazole are the drugs of choice, with cure rates greater than 90%.74 Nitazoxanide, a new antiparasitic agent, appears to be efficacious, with similar cure rates in several randomized, placebo-controlled trials.73,75-77 Erythromycin has no activity against amebic liver disease, and chloroquine has no activity against intes­ tinal disease.78 Because approximately 10% of asymptomatic cyst passers develop invasive amebiasis, E. histolytica carriers should be

treated.1,4 For noninvasive disease, treatment with a luminal agent alone is adequate (e.g., paromomycin 25-35 mg/kg/ day in three divided doses for seven days).70 Patients with amebic colitis should first be treated with an oral nitro­ imidazole (either metronidazole [500-750 mg three times daily for 10 days] or tinidazole [2 grams once daily for three to five days]) to eliminate invasive trophozoites. Metronida­ zole and tinidazole are believed to be less effective against organisms in the colonic lumen, and subsequent treatment with a luminal agent such as paromomycin is recommended to prevent recurrent disease.70,74 It is also for this reason that the familiar tissue amebicides (e.g., metronidazole) are not recommended as first-line agents for treatment of asymptomatic infection. At the recommended doses of metronidazole and tinidazole, gas­ trointestinal side effects including nausea and vomiting develop in approximately 30% of patients.74 Because of severe gastrointestinal side effects, simultaneous treatment with a nitroimidazole and a luminal agent generally is not recommended. Most patients with colitis respond promptly with resolu­ tion of diarrhea in two to five days.2 Despite conflicting reports on the safety of the nitroimid­ azoles for the developing fetus during pregnancy, women with severe disease during pregnancy should probably be treated without delay. As discussed in Chapter 82, metro­ nidazole (750 mg three times a day for 10 days) followed by a luminal agent is also the treatment of choice for amebic liver abscess.70,78

CONTROL AND PREVENTION

Prevention and control of E. histolytica infection depends on interruption of fecal-oral transmission. Water can be made safe for drinking and food preparation by boiling it for one minute, by halogenation (with chlorine or iodine), or by filtration.7 In the United States and Europe, modern water treatment facilities effectively remove E. histolytica. The importance of safe drinking water is highlighted by an outbreak of amebiasis in Tblisi, Republic of Georgia, where there was a water-borne epidemic due to decay of the water

Table 109-2  Amebicidal Agents Currently Available in the United States AMEBICIDAL AGENT For Luminal Amebiasis Paromomycin (Humatin) Iodoquinol (Yodoxin)

ADVANTAGES

DISADVANTAGES

7-day treatment course; may be useful during pregnancy Inexpensive and effective

Frequent gastrointestinal side effects; rare ototoxicity and nephrotoxicity 20-day treatment course; contains iodine; rare optic neuritis and atrophy with prolonged use Available in the United States only from the CDC; frequent gastrointestinal side effects; rare diplopia

Diloxanide furoate (Furamide) For Invasive Intestinal Disease Only Tetracyclines, erythromycin For Both Invasive Intestinal and Extraintestinal Amebiasis Metronidazole (Flagyl) Drug of choice for amebic colitis and liver abscess Tinidazole (Tindamax) Alternative to metronidazole; once daily dosing; now approved for distribution in the United States Nitazoxanide (Alinia) Useful alternative if the patient is intolerant of metronidazole or tinidazole For Extraintestinal Amebiasis Only Chloroquine (Aralen) Useful only for amebic liver abscess

Not effective for liver abscess; frequent gastrointestinal side effects; tetracyclines should not be administered to children or pregnant women Anorexia, nausea, vomiting, and metallic taste in nearly one third of patients; disulfiram-like reaction with alcohol; rare seizures Side effects are similar to those with metronidazole Limited clinical data for amebiasis; rare and reversible conjunctival icterus Occasional headache, pruritus, nausea, alopecia, and myalgias; rare heart block and irreversible retinal injury

CDC, Centers for Disease Control and Prevention. Adapted from Huston CD, Petri WA. Amebiasis. In: Rakel RE, Bope ET, editors. Conn’s Current Therapy, 2001. Philadelphia: WB Saunders; 2001. pp 50-4.

Chapter 109  Intestinal Protozoa treatment facilities following the demise of the Soviet Union.79 In the vast majority of the developing world, however, no modern water treatment facilities exist and none are likely to be constructed in the foreseeable future. Naturally acquired immunity to intestinal amebiasis pro­ vides short-lived protection against reinfection, giving hope that a vaccine may be feasible.5,31,32 Because humans and some higher nonhuman primates are the only known hosts for E. histolytica, a vaccine that successfully prevents colo­ nization might enable eradication of the disease.80

OTHER INTESTINAL AMEBAE Eight species of commensal amebae commonly infect the human gastrointestinal tract (see Fig. 109-2). These include Entamoeba dispar, Entamoeba moshkovskii, Entamoeba coli, Entamoeba hartmanni, Entamoeba gingivalis, Entamoeba polecki, Endolimax nana, and Iodamoeba butschlii. Dientamoeba fragilis (discussed in the following section), previously thought to be an ameba, is more closely related to the flagellated protozoan Trichomonas vaginalis than to the true amebae.7 With the exception of E. gingivalis, which has no known cyst stage, all of these true amebae have simple two-stage life cycles, consisting of an infectious cyst form and a motile trophozoite form.7 All but E. dispar and E. moshkovskii can be differentiated from E. histolytica using light microscopy based on characteristic features of the cyst and trophozoite forms. E. dispar must be differenti­ ated from E. histolytica based on biochemical, antigenic, or genetic differences.1 E. dispar is a nonpathogenic protozoan parasite that is morphologically indistinguishable from Entamoeba histolytica by light microscopy.1 An estimated 450 million people worldwide are infected with E. dispar, and infection with E. dispar is approximately 10 times more prevalent than E. histolytica infection.1,3,4 Although E. dispar has been demonstrated to cause mucosal ulcerations in animal models, it has not been demonstrated to cause human disease and does not require treatment.1 Entamoeba moshkovskii, which is primarily thought to be a free-living ameba, also has cysts and trophozoites indistinguishable from E. dispar and E. histolytica except that trophozoites of E. histolytica might show erythrophagocytosis. Although in some studies a high prevalence of E. moshkovskii infection in humans has been demonstrated, reports are conflicting regarding its pathologic potential.81-83 The primary clinical significance of E. dispar or E. moshkovskii is that they must be distinguished from E. histolytica to enable accurate diag­ nosis of invasive amebiasis. PCR to amplify small ribosomal RNA (not clinically available), and ELISAs using monoclo­ nal anti-amebic antibodies to detect specific E. histolytica antigens, make accurate diagnosis possible (see the section on E. histolytica diagnosis).55-64 Besides E. dispar, Entamoeba coli is the intestinal com­ mensal most commonly mistaken for E. histolytica. Entamoeba coli trophozoites contain a single nucleus with a prominent karyosome that usually is eccentric in location, distinguishing them from E. histolytica and E. dispar tro­ phozoites, which have a centrally located karyosome. In addition, the cyst form of Entamoeba coli typically contains five to eight nuclei (see Fig. 109-2). Entamoeba coli is non­ pathogenic and requires no specific treatment; however, it is a valuable marker of fecal-oral exposure, and it can be found concurrently with E. histolytica in 10% to 30% of patients in endemic regions.7 Entamoeba hartmanni was classified as “small race” E. histolytica for many years. The trophozoites resemble those

of E. histolytica except for their small size (less than 10 µm).7 Entamoeba hartmanni now is recognized as a nonpathogen that requires no treatment. Entamoeba gingivalis is the only ameba found in the oral cavity, where it lives in the anaerobic environment of the gingival crease. The trophozoite is identical in size to that of E. histolytica and contains a single nucleus with a promi­ nent central karyosome (see Fig. 109-2). No cyst form of E. gingivalis has been identified, and oral-oral contact is believed to be its mode of transmission.7,84 Entamoeba gingivalis is associated with poor dental hygiene and periodon­ tal disease, but no causal relationship to periodontitis has been proven.84 The increased frequency of colonization in this setting might simply reflect a more hospitable host environment. E. gingivalis is often associated with periodontal disease in AIDS patients, however, in whom treatment with metronidazole has been reported to be effective.85 Entamoeba polecki, a parasite characterized by a uni­ nucleated cyst, is primarily a parasite of pigs and monkeys that sometimes infects humans. It has been suggested that several distinct uninucleated cyst-producing Entamoeba species can infect humans and it has been proposed that these organisms collectively be termed “E. polecki-like.”86 Infection with E. polecki is rare except in Papua New Guinea where as many as 30% of children were found to be colon­ ized in one study.87 At present, specific treatment of E. polecki–like infections is not routinely recommended, but persons with heavy burdens of this parasite can develop nonspecific gastrointestinal symptoms and might benefit from treatment. Good clinical responses to metronidazole and diloxanide furoate have been reported.88 Endolimax nana is a nonpathogenic intestinal ameba that often infects humans.7 The distribution of E. nana is worldwide, but it is most common in the tropics, where 5% to 33% of persons are infected.89,90 Infection requires no specific treatment, but it serves as a useful marker for fecal-oral exposure. E. nana trophozoites can be distin­ guished from E. histolytica by their vesiculate nucleus, large irregular karyosome, and relatively small size (8 to 12  µm).7 Iodamoeba butschlii is a nonpathogenic intestinal ameba passed by the fecal-oral route. Trophozoites of I. butschlii contain a single nucleus with a large karyosome (which is distinct from the punctate karyosome of E. histolytica); its cysts contain a single nucleus, and a large, eccentric gly­cogen mass that stains with iodine (hence the name Iodamoeba). I. butschlii infection requires no treatment.7

GIARDIA INTESTINALIS EPIDEMIOLOGY

Giardia intestinalis (also called G. lamblia and G. duodenalis) is a ubiquitous flagellated intestinal protozoan. Van Leeuwenhoek accurately described its motile trophozoite form in his own stools in 1681, but it was not until 1915 that Stiles named the species.7 The life cycle of Giardia consists of an infectious cyst form and a motile trophozoite (Fig. 109-6). The cyst is oval (8 to 12 µm long by 7 to 10 µm wide), contains four nuclei, and has a rigid outer wall that protects it from dehydration, extremes of temperature, and chlorination (Fig. 109-7). Giardia cysts can survive in cold water for several weeks.7,91 Ingestion of as few as 10 to 25 cysts can result in infection.91 After ingestion, excystation occurs following exposure to stomach acid and intestinal proteases, each cyst giving rise to two trophozoites. Giardia trophozoites (see Fig. 109-3B)

1911

1912

Section X  Small and Large Intestine

Excystation follows exposure to gastric acid and intestinal proteases, releasing trophozoites which multiply by binary fission and reside in the upper small intestine adherent to enterocytes

Cysts are ingested (≤10–25 cysts) in contaminated water or food or by direct fecal-oral transmission (as in daycare centers)

Possible outcomes Asymptomatic infection, acute diarrhea, or chronic diarrhea and malabsorption Small intestine may demonstrate villus blunting, crypt hypertrophy, and mucosal inflammation

Cysts can survive in the environment (up to several weeks in cold water) They may also infect nonhuman mammalian species

Encystation occurs when the bile salt concentration increases and the luminal pH becomes alkaline Smooth-walled cysts can contain two trophozoites

Cysts and trophozoites are passed in the stool into the environment

Figure 109-6.  Life cycle of Giardia intestinalis. (From Hill DR, Nash TE. Intestinal flagellate and ciliate infections. In: Guerrant RL, Walker DH, Weller PF, editors. Tropical Infectious Diseases: Principles, Pathogens, and Practice. Philadelphia: WB Saunders; 1999.)

Figure 109-7.  Giardiasis. High-power view of a duodenal biopsy specimen showing many trophozoites near the surface of the epithelium between villi. (Courtesy of the Carlo Denegri Foundation, Turin, Italy.)

are pear-shaped (10 to 20 µm long by 7 to 10 µm wide), contain two nuclei, have eight flagellae for locomotion, and replicate by binary fission. The trophozoites live in the duodenum, where they adhere to enterocytes. Eventually they encyst, following exposure to alkaline conditions or bile salts, and are excreted in the stool to complete their life cycle.91 G. intestinalis, which was defined originally as a species by morphology, is more accurately defined as a species complex with at least seven major genotypes (assemblages A through G).92 Of these, only assemblages A and B are known to infect humans. Both of these genotypes also com­ monly infect cats and dogs, highlighting the importance of these pets as reservoirs for human disease.92 New data suggest that assemblage A isolates may be more virulent than assemblage B isolates.93,94 G. intestinalis is the most commonly identified intestinal parasite in the United States and was identified in 7.2% of stool samples examined by state health departments in 1987.95 Giardiasis occurs in both endemic and epidemic forms via water-borne, food-borne, and person-to-person

Chapter 109  Intestinal Protozoa transmission.96-102 Worldwide, Giardia infects infants more commonly than adults, and in highly endemic regions, essentially all children are infected by two to three years of age.103,104 In the developing world, it is likely that recurrent infantile diarrhea from giardiasis contributes significantly to malnutrition.103 In the United States, children in daycare and sexually active homosexual men have the greatest risk of infection.95,105 During a year-long longitudinal study at a U.S. daycare center, Giardia cysts were identified at some time in the stool of more than 30% of children.102 Addi­ tional risk factors for infection include drinking untreated surface water, a shallow well as a residential water source, swimming in any natural body of fresh water, and contact with a person who has giardiasis or contact with a child in daycare.97

PATHOGENESIS, PATHOLOGY, AND IMMUNOLOGY

Giardia causes malabsorptive diarrhea by an unknown mechanism. Trophozoites adhere (perhaps by suction) to the epithelium of the upper small intestine using a disk structure located on their anterior ventral surface.7,91 There is no evidence that trophozoites invade the mucosa,106 but electron microscopy has shown they damage the mucosal brush border.91,107 On biopsy, pathologic changes range from an entirely normal-appearing duodenal mucosa (except for adherent trophozoites), as was found in more than 96% of biopsy specimens in one large study, to severe villus atrophy with a mononuclear cell infiltrate that resembles celiac sprue.106,108,109 The severity of diarrhea appears to correlate with the severity of the pathologic change.91 The host immune response plays a critical role in limit­ ing the severity of giardiasis. When infected with Giardia, persons with common variable immunodeficiency develop severe, protracted diarrhea and malabsorption with spruelike pathologic changes that resolve with treatment of the Giardia.109 Both systemic and mucosal humoral immune responses can be measured consistently following Giardia infection. High serum anti-giardia IgM, IgG, and IgA titers can be detected, and anti-giardia secretory IgA (s-IgA) can be detected in saliva and in breast milk of infected mothers.110-112 Animal studies suggest that both early and late immune responses are important for control of Giardia infections. Interleukin-6 (IL-6) is important in the early immune response to Giardia in mice, as are mast cells, which might function as IL-6 producers or via another mechanism.113-115 In a B-cell-deficient transgenic mouse model, infection with Giardia does not resolve, confirming the importance of the humoral immune response for clear­ ance of established infections.116 In culture, Giardia tro­ phozoites vary expression of a group of cysteine-rich surface proteins termed variant surface proteins, and in experimental human infections, G. intestinalis isolates have been shown to undergo antigenic variation after approximately two weeks, roughly the time required to mount an initial antibody response.117 Although the role of the variant surface proteins remains unproved, antigenic variation might enable Giardia to evade the host immune response.118 The importance of a cellular immune response also is clear from animal studies. Athymic nude mice are unable to control Giardia muris infection, but reconstitution with immune spleen cells results in partial control. Upon immune reconstitution, however, severe inflammatory changes and villus atrophy develop in the intestine, suggesting that the immune response to infection might contribute to patho­ logic findings.119

Table 109-3  Frequency of Symptom(s) in Patients with Giardiasis100,106,110 SYMPTOM(S) Diarrhea Fatigue Abdominal pain, cramps Flatulence, bloating Weight loss Anorexia Vomiting Fever

FREQUENCY (%) 32-100 22-97 75-83 58-79 60 45 17-26 12-21

CLINICAL FEATURES

The clinical manifestations of Giardia are highly variable, ranging from asymptomatic infection to severe, chronic diarrhea with malabsorption. In one large study of biopsyproven giardiasis, only 32% of patients had diarrhea; most had nonspecific gastrointestinal complaints.106 Reported symptoms, in order of decreasing frequency, include diar­ rhea, fatigue, abdominal cramps, bloating, malodorous stool, flatulence, weight loss, fever, and vomiting (Table 109-3).100,110 During a food-borne outbreak, the mean dura­ tion of diarrhea was 16 days, but symptoms resolved spon­ taneously in nearly half of infected patients after seven to eight days.100 Many patients with clinically apparent giardiasis suffer from lactose intolerance, malabsorption, or both for months following cure of infection.120 As mentioned earlier, the severity of illness depends upon host and parasite factors. Different Giardia isolates have dramatically different abilities to cause disease during experimental human infections.121 Furthermore, certain populations, including children younger than two years and patients with hypogammaglobulinemia, are more likely to develop serious disease.103,109 Despite the importance of cel­ lular immunity for controlling infection in animal models and the increased risk of Giardia infection among sexually active homosexual men, giardiasis is not more common, or severe, or resistant to treatment in patients with AIDS,122 except perhaps when AIDS is advanced.123,124

DIAGNOSIS

Examination of concentrated, iodine-stained, wet stool preparations and modified-trichrome-stained permanent smears has been the conventional approach to identifying Giardia infections (see Fig. 109-3B). Because cysts and tro­ phozoites are present only intermittently in the stool, however, the sensitivity is only about 50%, even with exam­ ination of multiple specimens.108 With direct sampling of duodenal contents, such as duodenal aspiration or the string test, sensitivity can be improved to approximately 80%.108 On small intestinal biopsy specimens, identification of tro­ phozoites requires careful examination of multiple micro­ scope fields to ensure accuracy (see Fig. 109-7).106 Numerous molecular tests based on ELISAs or direct immunofluorescent antibody (DFA) microscopy are now widely commercially available to diagnose giardiasis in stool samples.125-127 These assay kits all work well and have sensitivities greater than 90% and specificities approaching 100%.127 Given the inability to exclude giardiasis even with repeated conventional stool examinations and the difficul­ ties of duodenal sampling, the first diagnostic test performed to evaluate for Giardia infection should be a stool ELISA or DFA. The primary role of endoscopy is evaluation for other pathologic conditions.

1913

1914

Section X  Small and Large Intestine TREATMENT

Metronidazole (250 mg orally three times a day for five days) is the preferred treatment for giardiasis.70 At this rela­ tively low dosage, metronidazole is generally well tolerated and is 80% to 95% effective at eradicating Giardia.128 The most common side effects of metronidazole are nausea, a metallic taste, and a disulfiram-like reaction with alcohol. Nitazoxanide appears to be at least as effective as metro­ nidazole and has the advantage of being available in a liquid formulation for use in pediatric patients. The recommended dosage in children is 100 mg (ages 12 to 47 months) or 200 mg (age older than four years) twice daily, and in adults is 500 mg twice daily for three days.73,76,129 Alternative regimens include tinidazole (2 g orally for one dose), quinacrine (2 mg/kg three times a day for five days; maximum 300 mg per day), furazolidone (100 mg orally four times a day for seven to 10 days), or paromomycin (25-35 mg/kg/day in three divided doses for seven days). Single-dose treatment with tinidazole has been used for years in Europe and the developing world and is approved by the U.S. Food and Drug Administration.70 Because paro­ momycin is not absorbed and there have been conflicting reports regarding the safety of metronidazole and tinidazole for the developing fetus, paromomycin may be especially useful for treatment of giardiasis during pregnancy.70 As noted earlier, many patients have prolonged lactose intolerance following Giardia infection, which can mimic ongoing infection.120 Therefore, the diagnosis should be reconfirmed before repeating therapy. For people in whom therapy fails, repeat therapy with the same drug (e.g., with higher doses of metronidazole) or combination therapy with metronidazole and quinacrine might work.124,128 Nitazoxanide alone also may be effective.130,131 Patients in whom treatment repeatedly fails should be evaluated for common variable immunodeficiency.109,128

CONTROL AND PREVENTION

Control of giardiasis relies on interruption of fecal-oral transmission. Water can be made safe for drinking and food preparation by boiling (for one minute), halogenation (with chlorine or iodine preparations), or filtration.7,128 Because of the low infectious dose of Giardia cysts and the poor hygiene of infants and children, person-to-person spread in daycare centers is much more difficult to control. Temporarily removing infected ill children from daycare is ineffective, perhaps because many infected children remain asymptom­ atic and go unrecognized.132 In the developing world, endemic giardiasis is unlikely to be controlled until facili­ ties become available for adequate filtration of water and disposal of sewage. A Giardia vaccine composed of killed G. intestinalis tro­ phozoites has been licensed for use in cats and dogs, but there have been few studies addressing human vaccination for giardiasis.133 Reduced susceptibility of some people living in endemic areas suggests that vaccination may be possible.134

DIENTAMOEBA FRAGILIS Dientamoeba fragilis is a binucleate organism with an ameboid trophozoite that measures 4 to 12 µm in diameter (see Fig. 109-2). No cyst form has been identified. The organism initially was classified as an ameba, but it is more closely related to the flagellates (trichomonads) based on morphologic studies and phylogenetic analyses of smallsubunit rRNA gene sequences. The mode of its transmission remains unknown. The absence of a cyst form makes direct

fecal-oral transmission unlikely, because the trophozoite is killed by stomach acid. Because of an association with Enterobius vermicularis (pinworm), some have hypothe­ sized that it is carried in pinworm eggs.135 D. fragilis infec­ tion is common throughout the world. D. fragilis was identified in 0.5% of all stool samples examined in a large U.S. study, and the prevalence is as high as 20% to 50% in selected populations.95,136-139 The role of D. fragilis as a pathogen had been controver­ sial. D. fragilis trophozoites do not invade tissue, and many persons infected with D. fragilis are asymptomatic.135 Fur­ thermore, the organism often is identified in the presence of other intestinal parasites, making its role in disease unclear.136,138,139 However, several studies of patients infected only with D. fragilis have found an association with diar­ rhea, abdominal pain, nausea, weight loss, anorexia, flatus, and malaise that resolved only after eradication.137,140,141 Based on these studies, it now is believed that D. fragilis is pathogenic and that it should be treated. Treatment with iodoquinol (650 mg orally three times a day for 20 days), metronidazole (500 mg to 750 mg three times a day for 10 days), paromomycin (25 to 35 mg/kg/day orally in three divided doses for seven days), or tetracycline (500 mg orally four times a day for 10 days) has been effective.137,141-143

BLASTOCYSTIS HOMINIS Blastocystis hominis is an intestinal protozoan that com­ monly infects the human colon. It is of uncertain taxonomic classification. Diameter ranges from 3 to 30 µm. In culture, B. hominis has ameboid, vacuolated, granular,7,144 and cystic forms.145 The distribution of B. hominis is worldwide, but infection is most common in the tropics.90,95,146-148 In a large study of intestinal parasitism in the United States, B. hominis was identified in 2.6% of stool specimens submit­ ted to state health departments; more than 70% of positive samples were from California.95 Among American travelers and expatriates, the prevalence often exceeds 30%.90,148 The significance of B. hominis as a pathogen remains controversial. Several studies have suggested an association with irritable bowel syndrome, but neither cause nor effect has been established and in most series, B. hominis infec­ tion is not more common among patients with gastrointes­ tinal complaints than among asymptomatic control subjects.90,148-151 In addition, the parasite burden does not correlate with symptoms.90,146 Nevertheless, multiple studies have used iodoquinol (650 mg orally three times a day for 20 days) or metronidazole (750 mg orally three times a day for 10 days) for treatment of symptomatic patients, with an overall improvement rate of about 50%.89,152 Clinical improvement actually may be due to treatment of unrecog­ nized infections with other organisms, because many people infected with B. hominis simultaneously harbor known pathogens.136,153,154 In one series of patients with B. hominis infection, 84% of patients were found to have at least one recognized pathogen other than B. hominis (E. histolytica, G. intestinalis, or D. fragilis) when repeated stool examina­ tions were obtained.136

CRYPTOSPORIDIUM SPECIES EPIDEMIOLOGY

First recognized in 1907 by Tyzzer as a gastric infection in mice, Cryptosporidium species are tiny intracellular proto­

Chapter 109  Intestinal Protozoa zoan parasites (2 to 5 µm) belonging to the phylum Apicom­ plexa; other medically important Apicomplexan parasites include Plasmodium species, which cause malaria, and Toxoplasma gondii, which causes toxoplasmosis. Cryptosporidium species infect the gastrointestinal epithelium of a wide range of vertebrates. Based on genetic and biologic differences, Cryptosporidium parvum human genotype (genotype 1 or genotype H) was renamed Cryptosporidium hominis in 2002. The name Cryptosporidium parvum was retained for the bovine genotype (genotype 2). Together, C. hominis and C. parvum cause most human infections, but occasionally other species cause human disease, especially in immunocompromised patients.155 Cryptosporidia were brought to prominent medical atten­ tion only in the early 1980s because of the devastating disease they caused in patients with advanced HIV infec­ tion. However, Cryptosporidium species increasingly are recognized as a cause of self-limited diarrhea, usually lasting one to four weeks, in immunocompetent persons.156,157 In developing countries, children younger than five years of age are affected most frequently.158 In a cohort of children two to five years of age in Dhaka, Bangladesh, 25.7% had at least one symptomatic Cryptosporidium infection during three years of follow-up.159 In industrialized countries, because Cryptosporidium oocysts are small and highly chlo­ rine-resistant, cryptosporidiosis has been associated with water-borne epidemics including numerous chlorinated swimming pool outbreaks. Cryptosporidia were responsible for the largest waterborne outbreak ever recorded that infected more than 400,000 Milwaukee residents in 1993.160,161 A low infectious dose and ready person-toperson spread also has resulted in epidemics in hospitals and daycare centers.162,163 In rural areas, zoonotic infections from direct contact with farm animals have been reported.164 Asymptomatic infection is common and infection appears to be more common in children who carry the DQB1*0301 HLA class II allele or the B*15 HLA class I allele.159

PATHOGENESIS, PATHOLOGY, AND IMMUNOLOGY

Upon ingestion of an infectious dose that may be as low as 1 to 10 oocysts, excystation and release of sporozoa occur in the presence of bile salts in the small intestine. The spo­ rozoites then attach to the intestinal epithelium, which trig­ gers elongation of epithelial cell microvilli on either side of the point of attachment. Fusion of the elongated microvilli with one another encloses the sporozoite within a vacuole located just underneath the brush border inside the epithe­ lial cell. The sporozoites then develop into merozoites, which replicate asexually. After several rounds of asexual replication, the merozoites exit the host cell and invade uninfected neighboring cells. In immunocompetent persons, second-generation merozoites undergo meiosis to yield the male and female micro- and macrogametocytes, respec­ tively. The microgametocytes then divide and exit the cell, whereupon they fertilize the macrogametocytes, forming oocysts that are shed in the feces. Rarely, multiplication has been seen in biliary, respiratory, or even conjunctival epi­ thelium in immunocompromised patients.7 Animal and human studies suggest that both humoral and cellular immune responses aid in the control of Cryptosporidium infections. Cryptosporidial diarrhea is clearly much more severe, if not intractable, in patients with immuno­ globulin deficiency, lymphocytic malignancies, or low CD4 counts associated with HIV infection.157,165,166 For unknown reasons, many more rounds of asexual reproduction typi­ cally occur in immunocompromised persons before devel­

opment of the merozoites into micro- and macrogametocytes, which at least in part explains the increased chronicity of infection that is seen in this context.

CLINICAL FEATURES

Following a one-week incubation period (range, 2 to 14 days), a watery, relatively noninflammatory diarrheal illness typically lasts for 10 to 14 days in immunocompetent hosts. Nausea, vomiting, abdominal pain, and mild fever may also be seen. Rarely, respiratory symptoms, pancreatitis, and biliary tract involvement have been reported, the latter in HIV-infected patients (see Chapter 33). Brief recurrence of diarrhea may be seen after improvement.161,167 In immunocompromised patients, particularly those with very low CD4 lymphocyte counts, the diarrheal illness with cryptosporidial infection can be cholera-like, protracted (often for the duration of severe immune compromise), and fatal.157

DIAGNOSIS

Because cryptosporidial infection usually is not identified in the laboratory except on specific request, the most impor­ tant element in diagnosis is to consider it in patients with diarrhea lasting longer than five to seven days and to request the appropriate special fecal studies. Because Cryptosporidium is spread in water, it is reasonable to consider cryptosporidiosis whenever the diagnosis of giardiasis is considered. In addition, it should be considered as a cause of persistent diarrhea in immunocompromised patients. Traditionally, cryptosporidial oocysts have been detected with a modified acid-fast stain of the stool (which can also detect Cyclospora and Isospora).168 As with giardiasis, ELISA or direct fluorescence antibody tests of the stool have replaced microscopy as the diagnostic test of choice. Numer­ ous commercial kits using either of these two methods have been developed that have sensitivities and specificities in excess of 90%. These immunodiagnostic tests may be of limited value in testing environmental samples, however, because there is some cross-reactivity with nonhuman cryp­ tosporidial oocysts.127 Occasionally, cryptosporidiosis is diagnosed with intestinal biopsy. Serologic tests are helpful primarily in epidemiologic studies, especially because they may be negative at the time of initial clinical presentation and positivity persists after infection has resolved. Finally, abdominal ultrasound, CT scans, and endoscopic retrograde cholangiopancreatog­ raphy (ERCP) may be helpful in diagnosis of acalculous cholecystitis and cholangiopathies, especially in immuno­ compromised patients.

TREATMENT

Nitazoxanide, an antiparasitic agent with broad-spectrum antiprotozoan and anthelminthic activity, is the only known drug with consistent efficacy for treating cryptosporidiosis in immunocompetent patients.73,75,169-171 Unfortunately, failure is common in immunocompromised patients, such as those with advanced HIV infection, although some studies have shown benefit.169,171 The recommended dosage in children is 100 mg (ages 12 to 47 months) or 200 mg (age greater than 4 years) twice daily and in adults is 500 mg twice daily for three days. Nitazoxanide generally is well tolerated. It is converted to the active metabolite tizoxanide, which undergoes conjugation to tizoxanide glucuronide and is excreted in the urine, bile, and feces. Nitazoxanide and tizoxanide are yellow, resulting in yellow urine and, in some patients on prolonged therapy, in yellow discoloration of the eyes that resolves after the drug is discontinued. Additional treatment options in immunocompromised

1915

1916

Section X  Small and Large Intestine patients not responsive to nitazoxanide include the nonab­ sorbable aminoglycoside paromomycin or paromomycin in combination with azithromycin.172 Most important in treating HIV-infected patients with cryptosporidiosis is highly active antiretroviral therapy (HAART), because ultimately, improvement of a cryptospo­ ridial illnesses depends on improvement in the immune compromise and the CD4 lymphocyte counts. Finally, papillotomy may be required for biliary obstruction with papillary stenosis from cryptosporidiosis in patients with AIDS (see Chapter 33).

CONTROL AND PREVENTION

Most important in control and prevention of this difficult protozoan parasitic infection is education regarding boiling or careful filtration of water; filter pores must be less than 1 µm in diameter. In addition, scrupulous enteric precau­ tions are required in institutions such as hospitals, daycare centers, or extended-care facilities for the elderly. These precautions are especially important because chlorine is ineffective in reducing viability of oocysts. Other means for disinfection that are being studied include ultraviolet light and irradiation. In one study, 2.5% glutaraldehyde was effective at inactivating Cryptosporidium oocysts, but only when a relatively low number of oocysts (15,000) were present and after treatment for 10 hours. Due to the corrosiveness of glutaraldehyde, the need to immerse instruments with their contaminating oocysts for this length of time presents a challenge for sterilization of endoscopes.173 Finally, because of the potential substantial long-term impact of cryptosporidial infection on childhood growth and development, control of cryptosporidiosis is critical in developing areas and must receive appropriate high priority in programs directed at improved water and sanitation worldwide.174,175

CYCLOSPORA CAYETANENSIS EPIDEMIOLOGY

Cyclospora was first reported as a cause of human disease by Ashford in 1979, who described infection of three patients in Papua New Guinea by what was at that time, an unnamed coccidian parasite.176 This parasite came to wider attention when documented as a cause of protracted diarrhea in AIDS patients and persistent diarrhea in non­ immunocompromised patients in New York City and the Caribbean, among expatriates in Nepal, and in an outbreak among house staff in a Chicago hospital.177-179 Finally, in 1993, Ortega and colleagues at Cayetano Heredia University in Peru demonstrated formation of sporozoites (sporulation) within immature oocysts and excystation in vitro and used electron microscopy to demonstrate oocysts containing spo­ rozoites with organelles characteristic of coccidians of the phylum Apicomplexa. They classified the organism as a member of the genus Cyclospora, and it has now been officially named Cyclospora cayetanensis to acknowledge this work.180 Ribosomal DNA analysis of phylogenetic relationships suggest that Cyclospora is closely related to Eimeria.181 Like Cryptosporidium, Cyclospora is being increasingly recognized in immunocompetent as well as immunocom­ promised persons. The infection is usually highly seasonal (in summer or wet months) and is probably spread via fecal contamination of water and vegetables.182,183 Cyclospora was brought to prominent attention throughout the United States and Canada with repeated outbreaks of diarrheal illnesses

occurring in more than 2000 patients every year from 1996 through 2000 in association with consumption of the late spring shipment of Guatemalan raspberries.184,185

PATHOGENESIS, PATHOLOGY, AND IMMUNOLOGY

The pathogenesis, pathology, and immunology of Cyclospora, although not as thoroughly studied as for Cryptosporidium, appear to be similar for the two organisms. One important distinction, however, is that unlike Cryptosporidium, which is promptly infectious when it is excreted in the stool, Cyclospora requires development outside the host before it becomes infectious. Consequently, secondary person-to-person spread, which is very common with cryp­ tosporidial infections, is not described with Cyclospora infections. An additional difference is that unlike the numerous mammalian hosts for cryptosporidial infections that also can infect humans, the animal reservoirs for Cyclospora are very poorly understood at present. The histopathologic changes seen in Cyclospora infec­ tions are similar to those seen with cryptosporidiosis, with villus blunting and a mild inflammatory infiltrate in the lamina propria predominantly in the small intestine.186

CLINICAL FEATURES

The clinical presentation of Cyclospora infection is indis­ tinguishable from that described with Cryptosporidium infections except perhaps for more severe generalized fatigue and malaise with Cyclospora infections, even in immunocompetent patients. Cyclospora diarrhea typically lasts for one to three weeks and may be associated with significant weight loss. Also as seen with cryptosporidiosis, protracted diarrhea and acalculous cholecystitis can occur with Cyclospora infection in HIV-infected persons.

DIAGNOSIS

As with Cryptosporidium, one must consider the diagnosis of Cyclospora in patients with protracted diarrhea. Diagno­ sis is best made at present with the acid-fast stain. Cyclospora oocysts measure 7 to 10 µm, nearly twice the size of those of Cryptosporidium, which are 4 to 5 µm.187 Cyclospora exhibits striking blue-green autofluorescence when exam­ ined under fluorescence microscopy, a characteristic that might have contributed to its initial confusion with cyano­ bacteria.188 Improved diagnostic methods using PCR have been developed, but they are not currently available for clinical use.189

TREATMENT

In contrast to Cryptosporidium infections, Cyclospora infec­ tions are readily treatable, even in immunocompromised patients. The drug of choice is trimethoprim-sulfamethoxa­ zole at a dosage of 160/800 mg twice daily for one week. Treatment promptly eradicates the organism and relieves symptoms.190,191 This treatment is similarly effective in patients with AIDS, although maintenance therapy with a single dose of trimethoprim-sulfamethoxazole three times per week may be needed to prevent relapse.192 Recent data show that ciprofloxacin provides a reasonable alternative in patients unable to tolerate trimethoprim-sulfamethoxa­ zole.193 Nitazoxanide also appears to be effective.75

CONTROL AND PREVENTION

Although readily treatable, Cyclospora infections are extremely difficult to control or prevent because of our limited ability to detect low infectious doses (for humans) of oocysts, which can contaminate products such as rasp­ berries, from which it is very difficult to eradicate. From

Chapter 109  Intestinal Protozoa limited studies, the organism also appears to be relatively chlorine-resistant and thus poses challenges to effective water treatment, much like Cryptosporidium. Elucidation of the reservoir of Cyclospora undoubtedly will enhance our ability to prevent and control the spread of this highly infec­ tious parasite. For example, it remains unclear why it is only the spring rather than the fall shipment of raspberries from Guatemala that has consistently posed problems with spread of Cyclospora infections. Whether this is related to migration of an avian reservoir has been postulated, but not proved.194 Consistent with this, several studies have reported isolation of Cyclospora oocysts from chickens.182,183

ISOSPORA BELLI EPIDEMIOLOGY

A relative of Cyclospora and Eimeria, I. belli is much larger, with elliptical oocysts measuring 20 to 30 µm long and containing two visible sporocysts that are acid-fast. Like Cyclospora, Isospora oocysts appear to require sporulation outside of the human host before they become infectious. There are no known nonhuman hosts for I. belli, and its distribution appears to be throughout tropical areas around the world. It appears to be a less common cause of diarrhea in children in developing areas than Cryptosporidium and is seen more often in older children, immunocompro­ mised patients, and in institutionalized children in North America.195,196

PATHOGENESIS, PATHOLOGY, AND IMMUNOLOGY

The pathogenesis, pathology, and immunology of Isospora infections appear to be similar to that of Cryptosporidium and Cyclospora infections, although less thoroughly studied.

CLINICAL FEATURES

Similar to Cryptosporidium and Cyclospora infections, Isospora characteristically produces a self-limiting diarrheal illness in immunocompetent persons and in travelers to tropical areas, with watery diarrhea and abdominal pain lasting two to four weeks. In immunocompromised patients, Isospora can produce a protracted sprue-like illness with malabsorption, weight loss, and prolonged diarrhea.196 As with Cryptosporidium and Cyclospora, acalculous chole­ cystitis also has been reported in patients with AIDS and Isospora infections.

DIAGNOSIS

The diagnosis of Isospora should be suspected in immuno­ competent patients with diarrhea lasting longer than five to seven days, especially following travel to tropical or develop­ ing areas, and in immunocompromised patients with persis­ tent diarrhea. Unlike other protozoan infections, Isospora infections may be associated with peripheral eosinophilia and with Charcot-Leyden crystals in the stool. The diagnosis of Isospora relies on identification of the large, oval oocysts (20 to 30 µm by 10 to 19 µm) on microscopic examination of concentrated fecal specimens by acid-fast staining. Oocysts also may be seen in biopsy specimens from the small intestine. In contrast to Cryptosporidium and Cyclospora infections, Isospora organisms have been observed invading beyond the epithelium into the lamina propria.155,197

TREATMENT

As with Cyclospora, Isospora infections are readily treated with trimethoprim-sulfamethoxazole; the dosage is

160/800 mg orally four times a day for 10 days, and then two times a day for three weeks, with both symptomatic and parasitologic responses, even in patients with AIDS.70 As previously described with Cyclospora, maintenance of sup­ pressive therapy may be required in patients with AIDS.191 Alternatives to trimethoprim-sulfamethoxazole may include ciprofloxacin.193

CONTROL AND PREVENTION

Prevention and control of Isospora infections will likely require improved sanitation in tropical areas.

MICROSPORIDIA Microsporidia infections are discussed in detail in Chapter 33.

EPIDEMIOLOGY

Microsporidia, the nontaxonomic term for Enterocytozoon bieneusi, Encephalitozoon (old Septata) intestinalis, and several other nonintestinal members of the phylum Microspora, are important causes of diarrhea, primarily in patients with impaired cell-mediated immunity due to AIDS or organ transplantation.198,199 E. bieneusi causes approxi­ mately 90% of cases.200 Microsporidia are identified in as many as 50% of AIDS patients with chronic diarrhea and are the most commonly identified pathogen in most series.199 The prevalence of infection is strongly correlated with decreasing CD4 T lymphocyte counts, although cases are not uncommon in persons with CD4 cell counts greater than 200 cells/mL.157,200,201 Microsporidia is distinctly less common in immunocompetent persons. The reservoir and modes of transmission are not certain.202-204 Epidemiologic data suggest that waterborne, person-to-person, and possi­ bly sexual transmission occur.

PATHOGENESIS, PATHOLOGY, AND IMMUNOLOGY

E. bieneusi enters only the cytoplasm of enterocytes, but E. intestinalis forms a parasitophorous vacuole in enterocytes, endothelial cells, fibroblasts, and macrophages, and it can disseminate to the kidney, prostate gland, and upper respi­ ratory tract. Typically, intestinal pathology is marked by villus atrophy, crypt hyperplasia, and mild inflammation in the lamina propria.198 The importance of cellular immunity in determining both infection and illness with intestinal microsporidia is indicated by its striking predominance in immunocompromised persons after organ transplantation or in those with AIDS.

CLINICAL FEATURES

Although primarily limited to immunocompromised patients, microsporidia cause chronic watery, relatively noninflammatory diarrhea and weight loss, occasionally with abdominal pain, nausea, vomiting, fever, and acalcu­ lous cholecystitis or even sclerosing cholangitis.198 E. intestinalis also can cause colitis and disseminate especially to the kidneys or less often to sinuses, bronchi, conjunctivae, or prostate.205 Rarely, cases have been reported of selflimited diarrhea in travelers or health professionals.202-204

DIAGNOSIS

Most laboratories use a modified trichrome stain to identify microsporidia in stool specimens.206 This method requires considerable skill and has limited sensitivity because of the small size of the spores (E. bieneusi measures 1 by 1.5 µm;

1917

1918

Section X  Small and Large Intestine E. intestinalis is slightly larger). Sensitivity can be improved by initially screening samples with fluorescent chitin stains such as Fungi-Fluor chitin stain (Polysciences, Warrington, Pa.) or Uvitex 2B (Ciba Geigy, Rueil Malmaison, France), and confirming positive results by modified trichrome stain­ ing.207,208 In addition, Gram stain and electron microscopy can identify the organisms in intestinal biopsy specimens. Sensitive PCR methods have been developed that enable species differentiation, but use of these methods is currently limited to research applications.204,209

TREATMENT

The response of E. bieneusi to albendazole is poor, but recent data indicate that oral fumagillin (20 mg three times daily for two weeks) may be effective for treatment of intestinal E. bieneusi infection in immunocompromised patients.210 Side effects including neutropenia and thrombo­ cytopenia are common and, given the limited clinical data available, this treatment should be undertaken with caution. E. intestinalis infections (approximately 10% of cases) respond well to albendazole, 400 mg twice daily for three weeks.70 As with all opportunistic infections in patients with AIDS, effective antiretroviral therapy (HAART) is essential for controlling microsporidial infections.

CONTROL AND PREVENTION

Because the reservoir and transmission of microsporidia remain unclear, control measures are primarily directed toward appropriate sanitary precautions and hand washing.

TRYPANOSOMA CRUZI (AMERICAN TRYPANOSOMIASIS OR CHAGAS’ DISEASE) EPIDEMIOLOGY

Although symptomatic Chagas’ disease has been confined to South and Central America, at least four autochthonous (indigenous) cases, as well as occasional laboratory-acquired and imported cases of acute Chagas’ disease, have occurred in the United States. Increasing numbers of immigrants are presenting with chronic Chagas’ disease, however, and pose distinct risks for transmission of Chagas’ disease.211 In patients surviving acute infection with T. cruzi in whom the chronic form of illness develops, myocardial disease is the most common manifestation. Megaesophagus and megaco­ lon are the most common intestinal manifestations of American trypanosomiasis. Small intestinal dilatation and aperistalsis also are seen. At postmortem examination, even in patients with asymptomatic T. cruzi involvement of the intestine, the small intestine has a significant reduction in submucosal and myenteric autonomic plexuses. American trypanosomiasis could prove to be a significant health problem in the United States because of the large reservoir of T. cruzi infection detected in animals in Arizona, California, New Mexico, Texas, Louisiana, Georgia, Florida, and Maryland. The epidemiologically important insects, the reduviid bugs of the Triatominae group, also have the same wide geographic distribution. Infection is transmitted when the reduviid bug infected with T. cruzi bites the victim. On biting, the arthropod discharges feces. The parasite is then introduced through the skin when the patient scratches the bite.

PATHOGENESIS, PATHOLOGY, AND IMMUNOLOGY

Metacyclic trypanosomes, an extracellular, flagellated form of the parasite that develops in the triatomine bug’s intesti­

nal tract, are deposited with the feces of the bug during the time it is taking a blood meal. Characteristically, deposition occurs on or near mucous membranes, particularly on the outer canthus of the eye or around the nose or lips. The invading organisms are phagocytosed by histiocytes in the corium and invade the adipose and subcutaneous muscle cells. They then develop into the intracellular form, the amastigote, which replicates by binary fission within the host’s cells. At variable intervals, the intracellular amasti­ gotes differentiate into trypomastigotes, a flagellated form that emerges into the blood and lymphatic circulation. The trypomastigotes then either invade other cells in diverse areas of the body (where they transform back to amastigotes and multiply) or are taken up into another reduviid bug’s midgut during a blood meal. The signs and symptoms of Chagas’ disease are caused by the intracellular amastigote forms. When the host cell rup­ tures, large numbers of amastigotes escape and temporarily enter the circulation as trypanosome forms. In the intestine, tissue injury can occur acutely or can trigger autoimmune damage to cardiac or nerve epitopes that cross-react with T. cruzi antigens to destroy the submucosal and the myenteric plexuses. The end result is enteromegaly, which at times may be massive. Immunosuppression as a consequence of chemotherapy or AIDS can reactivate chronic T. cruzi infec­ tion, causing acute disease or brain abscesses.

CLINICAL FEATURES

Acute Chagas’ disease occurs most often in children. It is characterized by high fever and marked edema, particularly with a periorbital distribution and often involving the entire body. In patients with acute Chagas’ disease, the periorbital edema of one or both eyes is striking. The victim can appear to be suffering from myxedema. There usually is hepato­ splenomegaly and enlargement of the thyroid gland, lymph nodes, and salivary glands. The acute stage lasts about 20 to 30 days. The manifestations of chronic Chagas’ disease depend on the major organ involvement within the body. Most com­ monly, symptoms are cardiac, manifested primarily as arrhythmias and congestive heart failure. With megaesopha­ gus, the history, barium esophagram, and esophageal motil­ ity tracing are indistinguishable from those of achalasia. With megacolon (Fig. 109-8), infrequent bowel movements and chronic constipation are the cardinal symptoms. With dilatation of the small intestine, diarrhea or constipation may be part of the picture. There may be evidence of weight loss and abdominal distention caused by the markedly dilated bowel.

DIAGNOSIS

Routine laboratory data provide no clue to the diagnosis of Chagas’ disease. Diagnosis of acute disease depends on demonstration of the trypanosome forms on blood smears during periods when the amastigotes rupture cells. During febrile periods, if the blood smear results are negative, inoc­ ulation of a patient’s blood into a guinea pig leads to pro­ liferation of trypanosomes that often can be recovered and identified. Amastigote forms may be detected in bone marrow, the spleen, or enlarged lymph nodes. Diagnosis of chronic Chagas’ disease depends on the pres­ ence of a typical clinical and epidemiologic history and on serologic tests, because parasites are only rarely identifiable in the blood or on biopsies of affected organs. Serum anti­ bodies to T. cruzi can be detected either by complement fixation or ELISA, and antibody testing can be requested from the U.S. Centers for Disease Control and Prevention (CDC). Xenodiagnosis has been used but is relatively insen­

Chapter 109  Intestinal Protozoa disease and might ameliorate dysphagia.215,216 Most patients with Chagas achalasia are best treated with either balloon dilation of the esophagus or esophagomyotomy. Occasion­ ally, nonperistaltic segments of intestine that are responsi­ ble for symptoms need to be resected.

CONTROL AND PREVENTION

Control and prevention require improved housing, use of insecticides and netting, and screening of blood for anti­ body in endemic areas.

KEY REFERENCES

Figure 109-8.  Film from a barium enema examination revealing megarectum and megasigmoid in a patient with Chagas’ disease. This complication, caused by autoimmune destruction of the submucosal and myenteric nerve plexuses, is believed to be a consequence of a cross-reaction of nerve epitopes with an antigen from Trypanosoma cruzi.

sitive, identifying less than 50% of patients infected with chronic Chagas’ disease. In this technique, trypanosomefree laboratory reduviid bugs are allowed to bite suspected victims. The trypanosomes multiply rapidly in the intesti­ nal tract of the insect, and examination of the intestine reveals flagellated trypanosomes in 10 to 30 days. Sensitive and specific PCR-based assays have been developed for diagnosis of acute and chronic Chagas’ disease, but these assays are not available for clinical use.212,213

TREATMENT

Nifurtimox (8 to 10 mg/kg daily in four divided doses for 90 to 120 days) or benznidazole (5 to 7 mg/kg daily in two divided doses for 30 to 90 days) can be used for treatment of Chagas’ disease.70 Both are subject to availability prob­ lems, have limited efficacy, and are associated with signifi­ cant side effects, including gastrointestinal symptoms in 40% to 70% of patients (nausea, vomiting, abdominal pain, and anorexia) and frequent neurologic sequelae.214 Isosor­ bide dinitrate has been shown to increase esophageal emp­ tying in patients with achalasia resulting from Chagas’

Ali IKM, Mondal U, Roy S, et al. Evidence for a link between parasite genotype and outcome of infection with Entamoeba histolytica. J Clin Microbiol 2007; 45:285-9. (Ref 8.) Amadi B, Mwiya M, Musuku J, et al. Effect of nitazoxanide on morbidity and mortality in Zambian children with cryptosporidiosis: A ran­ domised controlled trial. Lancet 2002; 360:1375-80. (Ref 171.) Duggal P, Haque R, Roy S, et al. Influence of human leukocyte antigen class II alleles on susceptibility to Entamoeba histolytica infection in Bangladeshi children. J Infect Dis 2004; 189:520-6. (Ref 9.) Haque R, Mollah NU, Ali IKM, et al. Diagnosis of amebic liver abscess and intestinal infection with the TechLab Entamoeba histolytica II antigen detection and antibody tests. J Clin Microbiol 2000; 38: 3235-9. (Ref 68.) Haque R, Mondal D, Duggal P, et al. Entamoeba histolytica infection in children and protection from subsequent amebiasis. Infect Immun 2006; 74:904-9. (Ref 5.) Haque R, Roy S, Kabir M, et al. Giardia assemblage A infection and diarrhea in Bangladesh. J Infect Dis 2005; 192:2171-3. (Ref 93.) Johnson EH, Windsor JJ, Clark CG. Emerging from obscurity: Biological, clinical, and diagnostic aspects of Dientamoeba fragilis. Clin Micro­ biol Rev 2004; 17:553-70. (Ref 135.) Kirkpatrick BD, Haque R, Duggal P, et al. Association between Cryptosporidium infection and human leukocyte antigen class I and class II alleles. J Infect Dis 2008; 197:474-8. (Ref 159.) Mackenzie WR, Hoxie NJ, Proctor ME, et al. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public water supply. N Engl J Med 1994; 331:161-7. (Ref 161.) Monis PT, Andrews RH, Mayrhofer G, Ey PL. Genetic diversity within the morphological species Giardia intestinalis and its relationship to host origin. Infect Gen Evol 2003; 3:29-38. (Ref 92.) Osterholm MT, Forfang JC, Ristinen TL, et al. An outbreak of foodborne giardiasis. N Engl J Med 1981; 304:24-8. (Ref 100.) Rossignol JF, Ayoub A, Ayers MS. Treatment of diarrhea caused by Cryptosporidium parvum: A prospective randomized, double-clind, placebo-controlled study of nitazoxanide. J Infect Dis 2001; 184: 103-6. (Ref 170.) Rossignol JF, Ayoub A, Ayers MS. Treatment of diarrhea caused by Giardia intestinalis and Entamoeba histolytica or Entamoeba dispar: A randomized, double-blind placebo-controlled study of nitazoxanide. J Infect Dis 2002; 184:381-4. (Ref 76.) Verdier RI, Fitzgerald DW, Johnson WD, Pape JW. Trimethoprimsulfamethoxazole compared with ciprofloxacin for treatment and prophylaxis of Isospora belli and Cyclospora cayetanensis infection in HIV-infected patients. A randomized, controlled trial. Ann Intern Med 2000; 132:885-8. (Ref 193.) Xiao L, Fayer R, Ryan U, Upton SJ. Cryptosporidium taxonomy: Recent advances and implications for public health. Clin Microbiol Rev 2004; 17:72-97. (Ref 155.) Full references for this chapter can be found on www.expertconsult.com.

1919

CHAPTE R

110 Intestinal Infections by Parasitic Worms David E. Elliott

CHAPTER OUTLINE Nematodes  1921 Ascaris lumbricoides  1921 Strongyloides stercoralis  1924 Capillaria (Paracapillaria) philippinensis  1925 Hookworms (Necator americanus, Ancylostoma duodenale, and Ancylostoma caninum)  1925 Whipworm (Trichuris trichiura)  1927 Pinworm (Enterobius vermicularis)  1928 Trichinella Species  1929 Anisakis simplex  1930

Parasitic worms are found worldwide. Modern travel, emigration,1 and consumption of “exotic” cuisines allow intestinal helminths to appear in any locale. People now acquire tropical helminths without leaving their industrialized temperate cities. Travel history is a critical, but often overlooked, aspect of the patient interview. Many helminths survive for decades within a host, so even a remote history of visits to or emigration from countries where they are endemic is important. Fresh food is flown around the world and often consumed raw. Physicians need to remain alert to the possibility of infection with these organisms because some cause severe disease that requires years to develop or occurs only under special circumstances. For example, patients might have occult Strongyloides stercoralis until treatment with glucocorticoids causes fulminant disease, occult Clonorchis sinensis until they develop cholangiocarcinoma, or occult Schisto­ soma mansoni until they develop portal hypertension and bleeding from esophageal varices. In developed countries, we usually diagnose an intestinal helminth because we stumble across it rather than because we actively pursue it. Helminths are complex organisms well adapted to their hosts; like quiet house guests, most cause no symptoms. Worms rarely cause diarrhea, but many medical laboratories do not assay formed stool routinely for parasite eggs. Physicians need to communicate their concerns of possible helminthic infection to laboratory personnel. A telephone call to the local laboratory before a sample is sent can improve diagnostic results dramatically. Occasionally, alarmed patients bring proglottids or whole worms that they passed with their stools. These specimens should be fixed in 5% aqueous formalin and sent for identification.2 All specimens should be handled carefully with full precautions to avoid accidental exposure. Some helminthic infections are difficult to diagnose, especially when the worm burden is light. Diagnosis can Videos for this chapter can be found on www.expertconsult.com.

Cestodes  1931 Diphyllobothrium Species  1931 Taenia saginata and Taenia solium  1931 Hymenolepis nana and Hymenolepis diminuta  1932 Dipylidium caninum  1933 Trematodes  1933 Intestinal Flukes (Fasciolopsis buski, Heterophyes Species, and Echinostoma Species)  1933 Liver Flukes (Clonorchis sinensis, Opisthorchis Species, and Fasciola Species)  1934 Blood Flukes (Schistosoma Species)  1935

require serologic evaluation, analysis of multiple stools, or use of concentration techniques in addition to a high level of physician awareness. For example, S. stercoralis eggs do not appear in the stool, and diagnosis is best made serologically. Ancyclostoma caninum causes eosinophilic enteritis but does not lay eggs when infecting people. Some helminths can cause severe disease, but this is unusual. Most persons colonized with helminths have no symptoms or illness attributable to the parasites. Only with heavy infections does disease result. Well-adapted worms usually act more as commensals than as pathogens. It is even possible that exposure to helminths affords some protection against disease due to excessive immune reactions.3,4 Helminths induce immune regulatory pathways.5 Recent studies in mice and rats show that exposure to helminths can be used to prevent or treat colitis,3,6-8 insulin-dependent diabetes,9 and autoimmune encephalitis.10,11 Studies in humans show that helminth exposure improves ulcerative colitis12 and probably Crohn’s disease13,14 and that helminth eradication increases atopy.15 Although it remains important to treat helminthic infections when they are discovered, further research on these organisms can enable discovery of new approaches to treat immune-mediated disease. This chapter is divided into three sections: nematodes (roundworms), cestodes (tapeworms), and trematodes (flukes or flatworms). For the most part, each worm is addressed separately, noting its epidemiology, life cycle, clinical manifestations, diagnosis, and treatment.

NEMATODES ASCARIS LUMBRICOIDES

Ascaris lumbricoides is the largest of the nematode parasites that colonize humans. Females can grow to 49  cm (19 inches).16 The name “lumbricoides” alludes to its resemblance to earth worms (Lumbricus sp.). The parasite is

1921

1922

Section X  Small and Large Intestine acquired by ingesting its eggs. Ascaris can cause intestinal obstruction and pancreaticobiliary symptoms. Treatment is albendazole.

Epidemiology

A. lumbricoides has a worldwide distribution, although these parasites are most numerous in less-developed countries and in areas with poor sanitation. About 25% of the world’s population (1.2 billion people) harbor A. lumbri­ coides.17,18 Children acquire the parasite by playing in dirt contaminated with eggs, whereas adults most often are infected by farming or eating raw vegetables from plants fertilized with untreated sewage. Pigs harbor Ascaris suum, which is closely related to A. lumbricoides, but crossinfection is rare.19

Life Cycle

Humans acquire the parasite by ingesting embryonated eggs that contain third-stage larvae. Freshly deposited fertilized eggs incubate in the soil for 10 to 15 days while the embryo develops and molts twice. The eggs become infective after this incubation period. The eggs are remarkably stable, can survive freezing, and can remain viable for seven to 10 years. The eggs are resistant to most chemical treatments including pickling, but they rapidly die in boiling water. Once ingested, eggs hatch in the duodenum and release their larvae, which penetrate the intestinal wall and enter the mesenteric venules and lymphatics. Larvae migrating with portal blood pass to the liver, through the sinusoids to the hepatic veins, and then through the right side of the heart to enter the lungs. Larvae migrating via the lymphatics pass through mesenteric lymph nodes to the thoracic duct and enter the superior vena cava to arrive in the lungs. The larvae then lodge in the pulmonary capillaries and break into the alveoli, where they molt twice while growing to 1.5 mm in length. Larvae then ascend the tracheobronchial tree, and arrive in the hypopharynx, they are again swallowed, and pass into the small intestine, where they molt again and finally mature. Mature male A. lumbricoides are smaller (10 to 30  cm) than females (20 to 49 cm). Worms mate in the small intestine and females deposit about 200,000 eggs a day. Adult worms live for about one year (six to 18 months). Because their eggs require incubation in the soil to become infective, Ascaris does not multiply in the host. Continued infestation requires repeat ingestion of embryonated eggs.

Clinical Features and Pathophysiology

A. lumbricoides produces no symptoms in most infected persons. Often, worms are found unexpectedly on endoscopy20,21 (Video 110-1) or are seen on radiologic imaging,22 or eggs are identified in stool specimens of patients with symptoms not directly attributable to the worms. Disease usually develops only in those with heavy worm burdens: pulmonary, intestinal, and hepatobiliary ascariasis are well described. Pulmonary ascariasis (Ascaris pneumonia) develops four to 16 days after ingesting infective eggs. The larvae migrate into the alveoli and elicit an inflammatory response that can cause consolidation. The pneumonia usually is self-limited but can be life-threatening. Large numbers of mature worms can cause severe intestinal symptoms including abdominal pain, distention, nausea, and vomiting. The most common complication of intestinal ascariasis is partial or complete small bowel obstruction; such patients often have a history of passing mature worms in their stool or vomitus. Patients with intestinal obstruction generally have more than 60 worms,23 and the rare

Figure 110-1.  Small intestinal obstruction caused by Ascaris lumbricoides. (From Wasadikar PP, Kulkarni AB. Intestinal obstruction due to ascariasis. Br J Surg 1997; 84:410.)

patients with fatal cases often have more than 600 worms. Fatality results from intestinal necrosis caused by obstruction, intussusception, or volvulus (Fig. 110-1).24 Most cases of obstruction, absent signs of peritonitis or perforation, can be managed conservatively. A. lumbricoides are highly motile. Mature worms can enter the ampulla of Vater (Fig. 110-2) and migrate into the bile or pancreatic ducts, causing biliary colic, obstructive jaundice, ascending cholangitis, acalculous cholecystitis, or acute pancreatitis.16 Pregnancy can promote biliary trespass.25 The worms can move in and out of the papilla, producing intermittent symptoms and fluctuating laboratory tests. Recurrent ascending cholangitis or acute pancreatitis from ascariasis is rare in highly developed Western countries but can be fatal if the diagnosis is not entertained.26

Diagnosis

Often it is an alarmed patient who discovers Ascaris after passing a motile adult worm with a bowel movement. The worms, however, usually do not cause diarrhea. Most patients do not have specific symptoms or eosinophilia. Ascaris eggs are visible in direct smears of stool (Fig. 1103). The eggs begin to appear in the stool about two months after initial exposure. Fertilized eggs are 35 by 55 µm and have a thick shell and outer layer; females also lay unfertilized eggs that are larger (90 by 44 µm) and have a thin shell and outer layer. Ascaris eggs that lose their outer layer resemble the eggs of hookworms. Adults worms may be seen at endoscopy,21 or identified on upper gastrointestinal series as long, linear, filling defects within the small intestine.22 The worms retain barium after it has cleared from the patient’s gastrointestinal tract, producing linear opacities. Similar findings are seen on endoscopic retrograde cholangiopancreatography (ERCP) if a worm is in the bile or pancreatic duct (Fig. 110-4). Ascaris also has a characteristic appearance on ultrasound examination of the biliary tree or pancreas: They appear as long, linear echogenic strips that do not cast acoustic shadows.22

Treatment

Asymptomatic colonization with A. lumbricoides is treated easily with a single 400-mg oral dose of albendazole. Alben-

Chapter 110  Intestinal Infections by Parasitic Worms

A

B

Figure 110-2.  A, Endoscopic view of Ascaris lumbricoides partially within the ampulla of Vater. B, Ascaris lumbricoides removed. (From Esser-Kochling BG, Hirsch FW. Images in clinical medicine. Ascaris lumbricoides blocking the common bile duct. N Engl J Med 2005; 352:e4.)

C

D

A

B A

Figure 110-3.  Stool specimen containing helminth eggs. A, Ascaris lumbricoides. B, Hookworm. C, Trichuris trichiura. D, Fasciolopsis buski. (A to D, Courtesy of Mae Melvin, MD, Atlanta, Ga.)

dazole inhibits glucose uptake and microtubule formation, effectively paralyzing the worms. Albendazole is poorly absorbed but is teratogenic, and therefore it should not be used in pregnant women. When possible, treatment with this agent should be delayed until after delivery. Singledose mebendazole also is efficacious for Ascaris.27 A study of 1042 pregnant women in Peru found no adverse effect of a single 500-mg oral dose of mebendazole on birth outcomes.28 Patients with pulmonary ascariasis should be treated with glucocorticoids to reduce the pneumonitis and be given two 400-mg doses of albendazole one month apart. Because albendazole is poorly absorbed, ascaricidal tissue concentrations are not achieved. The first dose kills mature worms that finished migrating to the intestine, and the second dose kills worms that were in transit when the first dose was given. Albendazole can cause nausea, vomiting, and abdominal pain.

Figure 110-4.  Endoscopic retrograde cholangiogram showing several Ascaris lumbricoides in the bile duct. (From van den Bogaerde JB, Jordaan M. Intraductal administration of albendazole for biliary ascariasis. Am J Gastroenterol 1997; 92:1531.)

Intestinal ascariasis with obstruction often can be treated conservatively with fluid resuscitation, nasogastric decompression, antibiotics, and one dose of albendazole. Surgery is not required unless the patient develops signs of volvulus, intussusception, or peritonitis. If the bowel is viable, an enterotomy allows intraoperative removal of worms. Albendazole may be held until after the obstruction resolves and then used to eradicate any remaining organisms. Hepatobiliary ascariasis also can be treated conservatively with fluid resuscitation, bowel rest, and antibiotics.29 Worms in the bile duct are not effectively treated with albendazole because it is poorly absorbed and not concentrated in the

1923

1924

Section X  Small and Large Intestine bile. This feature of albendazole is advantageous because were paralyzed worms unable to pass through the sphincter of Oddi, they could become trapped in the bile duct. Patients with hepatobiliary ascariasis should be treated with albendazole each day for several days because the worms only become susceptible when they migrate out of the bile duct. Worms also can invade the pancreatic duct and can be treated conservatively, as for hepatobiliary ascariasis.30 Ascending cholangitis, acute obstructive jaundice, or acute pancreatitis requires emergent ERCP with worm extraction from the ducts by balloon, basket, or forceps—preferably without sphincterotomy. Ampullary sphincterotomy permits worms easier access to the ducts and can increase the risk of recurrent pancreaticobiliary ascariasis.31

STRONGYLOIDES STERCORALIS

S. stercoralis is a free-living tropical and semitropical soil helminth, the filariform larvae of which can penetrate intact skin. As a parasite, Strongyloides lives in the intestine and lays eggs that hatch while still in the bowel. Filariform larvae develop within the intestine, migrate along defined paths, and mature to increase the number of adult parasites in the host. Immunosuppression and glucocorticoid treatment cause a fulminant reproduction of parasites that can prove fatal. Treatment is ivermectin.

Epidemiology

S. stercoralis is endemic in tropical and semitropical regions, but it can also be acquired in rural southeastern United States and northern Italy. Strongyloides exists as a free-living organism that does not require a host to replicate. Improved sanitation does not remove the risk of acquiring the parasite from soil. Patients from endemic areas, military veterans who served in Asia, and prisoners of war are at high risk for subclinical strongyloidiasis.

exit the anus during autoinfection. A study of prisoners of war found this creeping eruption to be a far more common symptom of chronic strongyloidiasis than were gastro­ intestinal complaints.32 Occasionally, patients have nausea, abdominal pain, or unexplained occult gastrointestinal blood loss from S. stercoralis. The parasite also can cause colonic inflammation that resembles ulcerative colitis but is more right-sided and strongly eosinophilic.33-35 While the parasite burden remains balanced, symptoms are minimal or absent. Immunosuppression or gluco­ corticoid administration upsets this balance. Previously asymptomatic, but chronically infested, patients develop fulminant, potentially fatal strongyloidiasis due to massive autoinfection.36,37 The mechanisms that permit massive autoinfection are unknown, but events that inhibit Th2directed immune responses can release eosinophil-mediated control of the parasites. In addition, glucocorticoids can act directly on the parasites to increase the development of infective filariform larvae.38 Fulminant disseminated strongyloidiasis rarely complicates HIV and AIDS.39 Massive autoinfection produces disseminated fulminant strongyloidiasis. Migrating filariform larvae injure the intestinal mucosa and carry luminal bacteria into the bloodstream, resulting in polymicrobial sepsis with enteric organisms. Streptococcus bovis endocarditis or meningitis40 also can result. Numerous larvae migrating through the lungs cause pneumonitis, and worms can arrive in unusual locations such as the brain. Fulminant strongyloidiasis often is fatal.

Diagnosis

Adult male and female S. stercoralis live in the soil and lay eggs that hatch rhabditiform larvae. Rhabditiform larvae develop in the soil into mature adults to complete the life cycle of this worm. Rhabditiform larvae (250 µm) also can develop into longer (500 µm) infective filariform larvae that can penetrate any area of skin contacting soil, after which they migrate through the dermis to enter the vasculature. The larvae circulate with the venous blood until they reach the lungs, where they break into the alveoli and ascend the bronchial tree. The worms then are swallowed with bronchial secretions and pass into the small intestine, where they embed in the jejunal mucosa and mature. Female S. stercoralis can lay fertile eggs by parthenogenesis and therefore do not require males to reproduce. The eggs hatch within the small intestine, and rhabditiform larvae migrate into the lumen. Rhabditiform larvae, not eggs, are passed in the stool. A critical feature of S. stercoralis infestation is that some rhabditiform larvae sporadically develop into infective filariform larvae within the intestine. Filariform larvae are able to reinfest (autoinfect) the patient, thereby increasing the parasite burden and permitting prolonged colonization so that subclinical strongyloidiasis can exist for many decades after the host has left an endemic area.

A recent survey of United States physicians-in-training demonstrated very poor ability to identify or even consider strongyloidiasis.41 Patients with chronic strongyloidiasis often are asymptomatic. Peripheral blood eosinophils may be elevated, but a normal eosinophil count does not argue against infestation with the parasite. Currently, the best method for detecting exposure is enzyme-linked immunosorbent assay (ELISA) for immunoglobulin (Ig) G antibodies against S. stercoralis. This assay is performed by the Centers for Disease Control and Prevention (CDC) in the United States and is 95% sensitive,42 sensitivity being highest for immigrants with prolonged exposure and lowest for returning visitors with lower-level recently acquired infestation.43 False-positive reactions can occur in patients exposed to other helminthic parasites,44 and serologic positivity can indicate prior exposure to S. stercoralis, not necessarily active infestation. Because chronic strongyloidiasis can remain subclinical and difficult to detect for decades, however, treatment of seropositive patients is warranted. Indeed, some argue that patients with only suspected strongyloidiasis, such as immigrants from endemic countries who have elevated eosinophil counts, should be treated empirically before glucocorticoid therapy.45 Active infestation can be diagnosed by finding rhabditiform larvae in direct smears of the stool, though this is an insensitive method. A 10-fold more sensitive technique is to spread stool on an agar plate and look for serpentine tracks left by migrating larvae.46 Intestinal biopsy is also an insensitive means of diagnosis.

Clinical Features and Pathophysiology

Treatment

Life Cycle

Most patients with S. stercoralis have no abdominal symptoms. Patients might have a serpiginous urticarial rash (larva currens) caused by the rapid (5 to 10 cm/hour) dermal migration of filariform larvae. This rash often occurs on the buttocks from larvae entering the perianal skin after they

Chronic strongyloidiasis is best treated with one dose of ivermectin (200 µg/kg) given orally; this dose is used in both adult and pediatric patients. Ivermectin is better tolerated than thiabendazole. Ivermectin paralyzes the intestinal adult worms but not the larvae migrating through tissue,

Chapter 110  Intestinal Infections by Parasitic Worms and therefore patients can develop recurrent infestation from migrating larvae; a repeat dose after two weeks helps to prevent this outcome. Successful treatment causes a fall in antibody titer by six months in most (about 90%) patients.42 Immunocompromised patients require repeat doses given 2, 15, and 16 days after the first dose.37

CAPILLARIA (PARACAPILLARIA) PHILIPPINENSIS

Capillariasis is acquired by eating raw fish that are infested with the parasite.47 The nematode causing capillariasis has been renamed from Capillaria philippinensis to Paracapil­ laria philippinensis,48 but by any name, it is deadly. The parasite replicates in the host, producing an ever-increasing number of intestinal worms. Patients develop proteinlosing, sprue-like diarrhea with progressive emaciation and anasarca, which ultimately leads to death. Treatment is albendazole.

Symptomatic patients have detectable eggs in their stool. The eggs are easily confused with those of Trichuris trichi­ ura, but T. trichiura eggs have prominent bipolar plugs that appear cut off in P. phillipinensis.47

Treatment

Capillariasis requires extended anthelminthic treatment with albendazole 200 mg orally twice daily for 10 days or mebendazole 200  mg orally twice daily for 20 days to prevent recurrence. Albendazole is better tolerated than mebendazole, which can cause headache, diarrhea, and abdominal pain. Extended treatment is necessary because larvae are resistant to these agents.

HOOKWORMS (NECATOR AMERICANUS, ANCYLOSTOMA DUODENALE, AND ANCYLOSTOMA CANINUM)

Epidemiology

The first known human case of capillariasis was reported in 1964. It remains a rare but deadly parasitic infestation. From 1965 through 1968, an epidemic in the rural Philippines involved 229 cases, with an overall mortality rate of 30%.49 As the name implies, Paracapillaria phillippinensis is endemic to the Philippines, but it also is endemic in Thailand and cases occur in Japan, Taiwan, Egypt, and Iran. Modern travel transports cases worldwide.50

Worldwide, an estimated 740 million people are infested with hookworm,17 usually by Necator americanus, Ancylos­ toma duodenale, or a mixture of the two. Hookworm is acquired by skin contact with contaminated soil. Moderate infestation contributes to iron deficiency. Hookworm should be suspected in patients with eosinophilia and iron-­ deficiency anemia. The dog and cat parasite Ancylostoma caninum is a cause of eosinophilic enteritis. Treatment is albendazole.

Life Cycle

Necator americanus and Ancylostoma duodenale

Birds, not humans, are the natural hosts for P. philippinen­ sis. In the avian small intestine, the larvae mature into adults. The adults are very small, measuring up to 3.9 mm for males and 5.3 mm for females. Adult worms mate and produce eggs. Eggs are deposited in bird droppings into ponds and rivers and are swallowed by fish to complete the life cycle. People become infested with the worm by eating raw or undercooked freshwater or brackish-water fish that contain the parasitic larvae. Some female adult P. philippinensis are larviparous, producing infective larvae instead of eggs. These larvae then mature in the small intestine and increase the parasite burden. This pathway of autoinfection permits a massive increase in parasite numbers. A rhesus monkey originally fed 27 larvae had more than 30,000 worms by 162 days of infection.51

Clinical Features and Pathophysiology

Capillariasis produces a progressive sprue-like illness. Symptoms begin with vague abdominal pain and borborygmi. Two or three weeks after infection, patients begin to have diarrhea. Initially intermittent, diarrhea becomes persistent and increasingly voluminous. Patients rapidly waste from escalating steatorrhea and protein-losing enteropathy. Eventually they manifest emaciation, anasarca, and hypotension; diarrhea produces severe hypokalemia. If untreated, patients die from cardiac failure or secondary bacterial sepsis usually about two months after the initial onset of symptoms. The progressive disease is believed to result from an everincreasing number of poorly adapted intestinal parasites. In autopsy studies, the jejunal intestinal mucosa showed flattened, denuded villi with numerous plasma cells, lymphocytes, macrophages, and neutrophils infiltrating the lamina propria.47

Diagnosis

Diagnosis is made by finding eggs and larvae in stool specimens. No serologic tests for capillariasis are available.

Epidemiology The geographic distribution of N. americanus and A. duo­ denale extensively overlap, but N. americanus predominates in the Americas, South Pacific, Indonesia, southern India, and central Africa, whereas A. duodenale is more common in North Africa, the Middle East, Europe, Pakistan, and northern India. Hookworm infestation is acquired by contacting soil contaminated with human waste. Hookworm is endemic in tropical to warm temperate areas that lack sufficient sewage facilities. Indigenous hookworm infestation largely has been eradicated in the United States, although small pockets of transmission still exist. Life Cycle Infective third-stage hookworm larvae penetrate intact skin, such as between the toes of bare feet while walking on contaminated ground. Larvae migrate through the dermis to reach blood vessels. This migration can cause a pruritic, serpiginous rash, cutaneous larva migrans (Fig. 110-5). Ancylostoma braziliense normally infests dogs and cats, but it produces a similar rash during infective dermal wandering in humans and is the usual cause of cutaneous larva migrans. Larvae of N. americanus and A. duodenale enter blood vessels in the skin and migrate with venous flow through the right side of the heart to the lungs. A. duodenale larvae can arrest their migration and become dormant for many months before proceeding to the lungs.52 In the lungs, larvae penetrate the alveoli and enter the air spaces, after which they migrate up the pulmonary tree, are swallowed with saliva, and pass into the small intestine, where they mature. Patients also can acquire A. duodenale by directly ingesting larvae crawling on contaminated fresh vegetables. Adult worms develop large buccal cavities and graze on the intestinal mucosa, ingesting epithelial cells and blood (Figs. 110-6 and 110-7). Adults are about one centimeter long and can live for up to 14 years. Mature worms mate and lay eggs. Each female N. americanus lays about 10,000 eggs a day, and each female A. duodenale lays about 20,000 eggs a day. Eggs are deposited with feces in moist, shady soil, where

1925

1926

Section X  Small and Large Intestine Table 110-1  A Comparison of Daily Physiologic Iron Losses and Iron Losses Due to Hookworm Infection in Women* CONDITION Physiologic Losses Menstruation Pregnancy Lactation Losses Due to Hookworm Infection Necator americanus (60-200 worms) Ancylostoma duodenale (20-100 worms)

IRON LOSS (mg/day) 0.44 2.14 0.23 1.10 2.30

*Losses shown are in addition to the basal iron loss of 0.72 mg/day. Adapted from Stoltzfuss RJ, Dreyfuss ML, Chwaya HM, Albonico M. Hookworm control as a strategy to prevent iron deficiency. Nutr Rev 1997; 55:223-32.

Figure 110-5.  Serpiginous rash caused by hookworm larvae migrating through the dermis. (Courtesy of the University of Iowa Department of Dermatology, Iowa City, Ia.)

Figure 110-6.  Scanning electron micrographic view of the buccal cavities of Ancylostoma duodenale (left) and Necator americanus (right). (From Hotez PJ, Pritchard DI. Hookworm infection. Sci Am 1995; 272:70.)

they hatch to release larvae. The larvae molt twice after which they move to the soil surface and seek a suitable host. Clinical Features and Pathophysiology Light infestations with N. americanus and A. duodenale cause no symptoms.53 The major consequence of moderate and heavy hookworm infestation is iron deficiency. Adult worms feed on intestinal epithelial cells and blood. The closely related A. caninum (see later) secretes anticoagulant peptides that inhibit clotting factors54 and platelet aggregation,55 thereby preventing hemostasis and permitting the hematophagous parasites to feed on host blood. Intestinal blood loss is estimated to be 0.01 to 0.04 mL/day per adult N. americanus and 0.05 to 0.3 mL/day per adult A. duode­ nale.56 With a moderate number of worms, this blood loss becomes appreciable (Table 110-1). Iron deficiency results when iron loss outstrips iron absorption. The average North American diet is high in iron so anemia might not develop, and men with a diet high in iron (more than 20 mg/day) can tolerate up to 800 adult hookworms without developing anemia. Infestation with hookworm can modulate immune responses.57 Clinical trials are under way to determine if subclinical infestation with hookworm inhibits immunemediated disease such as Crohn’s disease and asthma.58,59 Dose-ranging studies on healthy volunteers suggested that low-level hookworm infestation (10 larvae) is well tolerated.58 Diagnosis Hookworms can be visible endoscopically (Fig. 110-8),60 but diagnosis is made by identifying eggs on direct smears of formalin-fixed stool (see Fig. 110-3). Evaluation of three stool specimens obtained on separate days should permit diagnosis of hookworm,61 but light infestations can require concentration techniques. Eggs mature rapidly at room temperature and can hatch to release larvae. It is difficult to distinguish N. americanus eggs from those of A. duodenale simply by morphology.

Figure 110-7.  Longitudinal section of a hookworm grazing on intestinal mucosa. (Courtesy of Wayne M. Meyers, Washington, DC.)

Treatment Albendazole 400  mg given orally as a single dose is adequate treatment for hookworm. Mebendazole 100 mg given orally twice daily for three days also is effective but not as well tolerated. A. duodenale larvae can remain in a dormant state for months before maturing and causing relapse, a situation that is treated with a repeat course of albendazole or mebendazole.

Chapter 110  Intestinal Infections by Parasitic Worms

Figure 110-9.  Trichuris species: Adult male () and female () whipworms.

Figure 110-8.  Endoscopic view of Necator americanus in the duodenum (arrow). (From Reddy SC, Vega KJ. Endoscopic diagnosis of chronic severe upper GI bleeding due to helminthic infection. Gastrointest Endosc 2008; 67:990.)

Ancylostoma caninum

Epidemiology and Life Cycle A. caninum is a common hookworm of dogs and cats. It has worldwide distribution and is prevalent in the northern hemisphere. The parasite exists in areas with adequate sanitation, because dogs and cats indiscriminately defecate in yards, parks, and sandboxes. The life cycle of A. caninum is similar to that of A. duodenale, and the worm can be acquired orally; however, A. caninum does not fully mature in the human host, so no eggs are produced. Clinical Features and Pathophysiology A. caninum is a well-recognized cause of cutaneous larva migrans, a distinctive serpiginous rash caused by an abortive migration of the parasite in an unsupportive host. A. caninum also can cause eosinophilic enteritis, although not all eosinophilic enteritis is caused by this parasite (see Chapter 27). Patients with eosinophilic enteritis from A. caninum often are dog owners and present with colicky mid-abdominal pain and peripheral eosinophilia,62 but they do not recall having cutaneous larva migrans. Intestinal biopsies show high numbers (>45/high-power field) of mucosal eosinophils,63 and eosinophilic inflammation is most prevalent in distal small bowel. Unlike eosinophilic gastroenteritis, tissue eosinophilia is not present in the stomach. On endoscopy of the terminal ileum, patients might have scattered small superficial aphthous ulcers and mucosal hemorrhage.64 Serologic evidence suggests that A. caninum also may be a cause of abdominal pain without eosinophilia or eosinophilic enteritis.62 Diagnosis Diagnosis of A. caninum infestation is difficult. The parasite never fully matures, does not lay eggs, and is hard to detect. Serologic tests for A. caninum are research tools not routinely available. Therefore, treatment for A. caninum is empirical. Treatment Patients with distal small intestinal eosinophilic enteritis not attributable to another cause might benefit from empirical treatment for A. caninum. Albendazole 400  mg as a

single oral dose or mebendazole 100 mg orally twice daily for three days is adequate to treat A. caninum infestation. Given for brief periods, these drugs are quite safe.

WHIPWORM (TRICHURIS TRICHIURA)

T. trichiura, commonly called whipworm, has worldwide distribution. People acquire Trichuris by ingesting embryonated parasite eggs. Most persons have no symptoms, although heavy infestations are associated with a dysenterylike syndrome. Treatment is mebendazole.

Epidemiology

An estimated 800 million people harbor T. trichiura. It occurs in temperate and tropical countries and remains prevalent in areas with suboptimal sanitation. In one equatorial Cameroon province, 97% of the school-age children had T. trichiura.65 Whipworm eggs are sensitive to desiccation, so prevalence is low in desert climates.

Life Cycle

T. trichiura has a simple life cycle. Colonization occurs by ingesting the parasite egg, each of which contains one developed larva. The eggs hatch in the intestine, and larvae migrate to the cecum, where they mature, mate, and lay eggs. This process takes about eight to 12 weeks. Adult worms are approximately three centimeters long and have a thin tapered anterior region so that the worm resembles a whip (Fig. 110-9, Video 110-2).66 A mature female worm lays about 20,000 eggs a day and can live for three years. Eggs are deposited with feces into the soil. Over the next two to six weeks, one larva develops within each egg, but the egg is not infective until it has fully embryonated. Therefore, T. trichiura does not multiply in the host and is not directly transmitted to other persons.

Clinical Features and Pathophysiology

Most persons with T. trichiura infestation have no symptoms attributable to the parasite. Most people in an endemic area are colonized by small numbers (less than 15) of worms and for them, the parasite is a commensal organism rather than a pathogen. Some people harbor hundreds or even thousands of worms,67 and they are the ones who develop symptoms68; this bimodal distribution of infestation persists after patients are treated and then become reinfected naturally, suggesting that unique host factors (genetic or behavioral) contribute to determining an individual patient’s worm burden. Rectal prolapse can occur in children with extremely high numbers of T. trichiura worms.69 Some persons with numer-

1927

1928

Section X  Small and Large Intestine ous worms have mucoid diarrhea and occasional bleeding, a combination of symptoms called the Trichuris dysentery syndrome (TDS). Children with this condition have growth retardation,70 but studies attributing these symptoms to T. trichiura are complicated, because persons with TDS often are socioeconomically deprived and may be coinfected with other pathogens. Colonic biopsy specimens from children with TDS show few or no abnormalities compared with healthy local children,71 other than an increase in mast cells72 and in the number of cells that express TNF-α and calprotectin.73 A different but closely related species, Trichuris muris, infests mice. Mouse strains that react to the parasite with a strong Th2 response, characterized by production of interleukin (IL)-4, IL-5, and IL-13, are able to expel the worms, whereas strains that respond with a Th1 response (interferon [IFN]-γ) have difficulty expelling the worms.74 Blocking IL-4 makes resistant strains susceptible, and blocking IFN-γ makes susceptible strains resistant to chronic infestation with T. muris.75 The type of immune response developed by inbred mice to T. muris is an important factor in determining length and intensity of infestation. A similar response in humans might explain why some people repeatedly acquire heavy infestations whereas others carry only a few worms.

Pinworm remains common in many areas, but it appears to be decreasing in prevalence. A survey of positive cellophane tape tests (see later) in New York City documented a sharp decline in positivity from 57 of 248 tests in 1971 to 17 of 165 in 1978 to 0 of 38 in 1986.83 Similar trends are reported from California.

Diagnosis

Clinical Features and Pathophysiology

Diagnosis is made by identifying T. trichiura eggs in stool specimens. Trichuris eggs are 23  µm by 50  µm and have characteristic plugs at each end (see Fig. 110-3).

Treatment

T. trichiura is treated with mebendazole 100 mg twice a day for three days; alternatively, patients can take albendazole 400  mg each day for three days. Heavily infested patients might require seven days of treatment.76 Single-dose treatment with albendazole is ineffective27 but one treatment with a combination of albendazole (400 mg) and ivermectin (200 µg/kg) appears quite effective, with cure rates of up to 80% and egg reduction rates of 94%.77,78

Life Cycle

E. vermicularis has a simple life cycle with a “hand to mouth” existence. The worm is acquired by ingesting parasite eggs. Most often these eggs are on the hands of the host; however, the small eggs also may become airborne, inhaled, and then swallowed. Eggs hatch in the duodenum, releasing larvae that molt twice as they mature and migrate to the cecum and ascending colon (Fig. 110-10, Video 110-3).84 The parasites are small: adult males measure 0.2 mm by 2 to 5 mm, and adult females measure 0.5  mm by 8 to 13  mm. After mating, gravid females migrate to the rectum. During the night, eggladen females migrate out of the anal canal and onto the perianal skin. Each female deposits up to 17,000 eggs, which mature rapidly, becoming infective within six hours. Pinworm infestation typically causes perianal itching, and scratching gathers eggs onto the hands, promoting reinfection and transmission to others. E. vermicularis is an extremely well adapted parasite that produces no specific symptoms in the vast majority of colonized persons. Most symptoms are minor, such as pruritus ani and restless sleeping. Rarely, pinworm causes eosinophilia or eosinophilic enteritis.85 Vulvovaginitis is more common in girls with pinworm than in girls without this infection. Vulvovaginitis may be caused by migration of the worms into the introitus and genital tract. Dead worms and eggs encased in granulomas have been found in the cervix, endometrium, fallopian tubes, and peritoneum, attesting to the migratory effort of female worms.86 Ectopic enterobiasis is rare and causes no or very little overt pathology.

PINWORM (ENTEROBIUS VERMICULARIS)

E. vermicularis, commonly called pinworm, is the most common helminthic parasite encountered by primary care providers in developed nations. It is acquired by ingesting parasite eggs, and most people remain asymptomatic after being colonized. Diagnosis is made by the cellophane tape test. Treatment is mebendazole for the affected patient and for all family members.

Epidemiology

E. vermicularis is a quintessential intestinal parasite with no geographic constraints. It is transmissible by close contact with colonized persons. People have had pinworm for thousands of years, and before modern sanitation, colonization by pinworm probably was universal. E. vermicu­ laris eggs were identified in a 10,000-year-old human coprolite found in Utah.79 The pinworm Enterobius gregorii, originally thought to be a separate species of pinworm,80,81 actually may be just a young adult form of E. vermicularis.82 People of every socioeconomic group can acquire pinworm and it remains quite prevalent. School-age children are most often colonized, compelling other household members to acquire the parasite. Crowding and institutionalization promotes acquisition. Eggs can survive in the environment for approximately 15 to 20 days and are resistant to chlorinated water (e.g., swimming pools).

Figure 110-10.  Pinworm (Enterobius vermicularis, arrows) found on screening colonoscopy of an institutionalized man.

Chapter 110  Intestinal Infections by Parasitic Worms Infestation with E. vermicularis can influence mucosal immune responses. One case report described a 12-year-old girl with pinworm and apparently latent ulcerative colitis, who developed severe ulcerative colitis after treatment with pyrantel to remove the worms.87 While she was colonized with E. vermicularis, intestinal biopsies showed increased expression of mRNA for IL-4, transforming growth factor (TGF)-β, IL-10, and FOXP3 compared with biopsy specimens taken after anthelminthic treatment; these transcripts are associated with immune regulatory pathways that suppress inflammation.

Diagnosis

E. vermicularis eggs are not plentiful in stool, an observation that might explain the low prevalence rates found in studies that only use stool specimens for diagnosis. The NIH cellophane tape test is the classic diagnostic test for pinworm. A two- to three-inch piece of clear tape is applied serially to several perianal areas in the morning before washing. The tape is then applied to a glass slide. Microscopic evaluation demonstrates parasite eggs that measure 30 by 60 µm, have a thin shell, and appear flattened on one side. Three to seven daily samples are needed to exclude pinworm infestation.

Treatment

Pinworm actually requires no treatment unless the patient is symptomatic. It is highly transmittable, however, and for that reason should be expunged. E. vermicularis is readily treated with a single 100-mg dose of mebendazole or a 400-mg dose of albendazole. Reinfestation is common, and patients should receive a second treatment after 15 days. All members of the family should be treated and clothes and bed linens should be washed. Albendazole and mebendazole are potentially teratogenic. Because E. vermicularis has very low pathogenicity, treatment of pregnant women should be postponed until after delivery.

TRICHINELLA SPECIES

Trichinosis is a systemic illness caused by any of the eight closely related Trichinella species. People acquire the parasite by ingesting larvae present in raw or undercooked meat such as pork. Trichinosis has both intestinal and systemic phases characterized sequentially by nausea and diarrhea, fever, myalgia, and periorbital edema. Intense exposure can cause death due to severe myositis, neuritis, and thrombosis. Treatment is albendazole and glucocorticoids.

babwe, Ethiopia, and Mozambique; and T. papuae only in Papua New Guinea. Each of the Trichinella species can infest any mammal. T. nativa is resistant to freezing for up to five years. Trichinosis was much more common in the United States than it is now. In the late 1940s, about 400 cases per year of symptomatic trichinosis were reported to various health agencies, and this number dropped to an average of 14.4 cases per year in the time period 1997 to 200189; reports from Germany show a similar pattern.90 This decrease is explained by two major factors: First is the strong admonition to thoroughly cook all pork products; second is a change in farming practice to now feed pigs only grain. Industrialized pig farms in North America have been free of trichinosis for more than 50 years, but trichinosis is a reemerging illness in eastern Europe, related to relaxed enforcement of regulations.91 Currently, most reported cases involve a discrete exposure. For example, a 1991 outbreak in Wisconsin involved 40 people who ate pork sausage from one shop. A 1995 outbreak in Idaho involved 10 people who ate cougar jerky.92 A 2005 outbreak in Canada involved at least 14 people who ate frozen then stewed black bear meat.93 In France, several outbreaks have resulted from eating raw horse meat.94 This emphasizes that all mammals including herbivores can transmit Trichinella.

Life Cycle

The same host harbors both the adult and larval form of Trichinella.95 People acquire the parasite by eating raw or undercooked meat that contains encapsulated parasite larvae. Each cyst dissolves in the digestive tract, releasing one larva that invades the small intestinal mucosa and lives within the cytoplasm of about 45 villus cells (Fig. 110-11). Larvae mature rapidly and mate within 30 hours. Adults are minute: Male worms measure 60 µm by 1.2 mm and female worms measure 90 µm by 2.2 mm. Females are viviparous and begin releasing larvae about one week after their initial ingestion. Adults are short-lived, producing larvae for only four weeks, by which time they are expelled by the host.

Epidemiology

Trichinosis is acquired by eating raw or undercooked meat that contains parasite larvae of Trichinella species. Worldwide, domestic pigs are the most common carriers. Trichi­ nella species are divided into two groups,88 one that forms encapsulated muscle cysts and only infests mammals (Trichinella spiralis, Trichinella britovi, Trichinella nelsoni, Trichinella native, Trichinella murrelli), and one that does not form encapsulated cysts and infests mammals and birds (Trichinella pseudospiralis) or mammals and reptiles (Trichinella papuae, Trichinella zimbabwensis). To date only T. zimbabwensis has not been implicated in human disease. These species are closely related, morphologically nearly identical, and distinguished using molecular approaches. Trichinella has worldwide distribution, with T. nativa and T. murrelli in the Arctic and subarctic regions; T. spiralis and T. pseudospiralis in the Americas, Europe, and Russia; T. britovi in Europe, north Africa, the Middle East, and Asia; T. nelsoni in equatorial Africa; T. zimbabwensis in Zim­

Figure 110-11.  Illustration of Trichinella spiralis coiled through enterocytes in the small intestine. Each Trichinella larva lives within the cytoplasm of approximately 45 villus cells.

1929

1930

Section X  Small and Large Intestine The larvae live much longer than the adult worms. Larvae measure six by 100 µm and enter the intestinal blood and lymphatic vessels. They are distributed by the circulatory system through the body but develop only within striated muscle. The larva enters a striated muscle fiber but does not kill the myocyte. Instead, it induces the cell to transform into a novel nurse cell that houses and feeds the parasite. The larva grows and develops into the infective stage in about five weeks. The coiled larvae remain viable for many years awaiting ingestion by another animal.

Clinical Features and Pathophysiology

Although most infestations with Trichinella are asymptomatic, significant exposure produces illness and even death.96 Clinical trichinosis has two phases caused by the enteral (adult) and parenteral (larval) stages of the parasite. Intestinal symptoms result from enteritis due to adult worms that have embedded themselves in the intestinal epithelium. Enteritis produces abdominal pain, nausea, vomiting, diarrhea, and low-grade fever. Intestinal symptoms begin about two days to one week and peak at two weeks after ingestion of contaminated meat. The timing and severity of symptoms vary with intensity of exposure. The intestinal phase of trichinosis often is misdiagnosed as viral gastroenteritis or food poisoning. T. spiralis also infests mice and rats, permitting detailed study of the intestinal phase of infection.97 Mice begin to expel adult worms about two weeks after initial infestation. Type 2 (Th2) cytokines (IL-4 and IL-5) promote worm expulsion. Expulsion of adult worms results from focal immune attack, increased secretions, and enhanced intestinal motility; T lymphocytes, eosinophils, and mast cells assist this primary response. Rats previously exposed to T. spiralis rapidly expel the parasite upon rechallenge, a protection likely resulting from an immediate-type hypersensitivity response to the parasite triggered by IgE-armed mast cells. The parenteral phase of trichinosis begins with the birth of migratory larvae about one week after ingestion of the contaminated meat. Larvae migrate into muscle and other organs such as the brain, spinal cord, and heart, evoking inflammatory responses; high fever, myalgia, periorbital edema, dysphagia, headache, and paresthesia result. Symptoms peak about four to five weeks after initial exposure and can take months to resolve. The severity and timing of symptoms vary with the intensity of exposure. Many patients develop systemic complaints without prior intestinal symptoms. The inflammatory response to migrating larvae produces myositis. Patients have eosinophilia and an elevated serum level of creatine phosphokinase (CPK). An intense exposure can cause fatal myocarditis, neuritis, and vasculitis or thrombosis. Patients are at highest risk of death between the third and sixth week after exposure. Because trichinosis is rare, index cases often are misdiagnosed initially. Numerous persons presenting in a narrow time frame and with similar and compatible symptoms should prompt consideration of trichinosis as the diagnosis.

Diagnosis

Trichinella cannot by diagnosed by stool examination or intestinal biopsy. Trichinella species do not lay eggs, and no larvae are present in stool specimens. Even with heavy infestations, adult worms are too uncommon to be found by random biopsy. Diagnosis is made by muscle biopsy demonstrating larvae within nurse cells. Diagnosis also can be made by serology. Acute and convalescent serum samples confirm a rise in anti-Trichinella antibody.

Treatment

Although adults are short-lived, treatment with albendazole 400 mg twice a day or mebendazole 5 mg/kg/day for 10 to 15 days98 is warranted and abbreviates the production of larvae by adult worms. Addition of glucocorticoids reduces inflammation and systemic symptoms; however, glucocorticoids given in the absence of a benzimidazole can prolong the intestinal phase, increasing the number of larvae released.

ANISAKIS SIMPLEX

A. simplex and another anisakid, Pseudoterranova decipi­ ens, can infect people transiently, causing abdominal pain, hematemesis, or intestinal inflammation. A. simplex is also a potent allergen that might explain some cases of fish allergy. Anisakidosis is acquired by eating raw or undercooked fish. No treatment is usually required.

Epidemiology and Life Cycle

A. simplex and P. decipiens infest fish and marine mammals.99 People become accidental hosts by eating raw or pickled fish. Anisakidosis has become more common with the increased popularity of eating raw fish (e.g., sushi). Many species of saltwater fish harbor A. simplex larvae including herring, mackerel, salmon, plaice, and squid. The parasite larvae initially infest crustaceans that are consumed by fish. The larvae migrate to the fish musculature and, if a parasitized fish is eaten by another fish, the larvae again migrate to the musculature of their new host. Eventually, a parasitized fish is eaten by a marine mammal that serves as the definitive host. In the marine mammal, the parasite larvae mature into adult intestinal worms and lay eggs that are passed with feces, the eggs hatch to release larvae that infest crustaceans, and the life cycle is thus renewed.

Clinical Features and Pathophysiology

A. simplex and P. decipiens cause transient infestations in humans. They do not reach full maturity in humans and therefore produce no eggs. The most common gastrointestinal symptom is acute severe stomach pain with nausea and hematemesis shortly after eating larva-infested raw fish. Endoscopy may demonstrate a small larva partially penetrating the gastric or intestinal wall.100,101 Rarely, A. simplex can enter the intestinal wall and cause a strong inflammatory reaction that can mimic acute appendicitis102 or Crohn’s disease. Human infestations with either A. simplex or P. decipiens is termed anisakidosis after the family name (Anisakidae) for these parasites. A. simplex is a potent allergen, and many cases of seafood (fish) allergy actually may be reactions to A. simplex,103 including anaphylaxis from well-cooked marine fish.99,104 In Spain, 12% to 22% of persons are seropositive for IgE against A. simplex.105,106

Diagnosis and Treatment

A history of recent (within three days) ingestion of raw fish suggests anisakidosis in the appropriately symptomatic patient. Diagnosis is made by finding the larvae on endoscopy or in surgically excised specimens. Gastric anisakidosis is diagnosed by endoscopy, and endoscopic removal of the anisakid alleviates symptoms. Intestinal anisakidosis can prompt surgery for patients presenting with symptoms of acute small bowel obstruction or peritonitis,107 but surgery may be avoidable if a recent history of eating raw fish is elicited and conservative treatment is tolerated.108 A. simplex and P. decipiens infestations are transient because the parasites do not survive in humans. Therefore, treatment with an anthelminthic is not needed.

Chapter 110  Intestinal Infections by Parasitic Worms CESTODES DIPHYLLOBOTHRIUM SPECIES

Fish tapeworm (Diphyllobothrium species) is the largest parasite of humans, reaching lengths of up to 40 feet (12 meters). People acquire the parasite by eating raw or undercooked freshwater fish. Diphyllobothrium latum absorbs dietary cobalamin and can cause vitamin B12 deficiency over time. Treatment is albendazole.

Epidemiology

D. latum is most common but other Diphyllobothrium species (e.g., Diphyllobothrium dendriticum, Diphylloboth­ rium nihonkaiense) can colonize humans.109,110 D. latum is endemic in northern Europe, Russia, and Alaska, but fish tapeworm has been reported in Africa, Japan, Taiwan, Australia, South America, North America, and Canada.111

Life Cycle

Fish tapeworm has a complex life cycle with two intermediate hosts. Parasite eggs that reach fresh water embryonate and then release free-swimming larvae called coracidia. Coracidia are ingested by water fleas (Cyclops and Diapto­ mus) and develop into procercoid larvae. Freshwater fish eat these small crustaceans, and the parasite changes into the infective plerocercoid form. The plerocercoid larva embeds in fish muscle and organs, growing to two centimeters in length. If an infected fish is consumed by another fish, the plerocercoid larva simply migrates into the flesh of the second fish. Trout, salmon, pike, perch, and whitefish all can harbor D. latum. People acquire the parasite by eating raw or undercooked fish. D. latum also can colonize many other mammals such as dogs, cats, bears, and seals. In mammals, the ingested plerocercoid larva attaches to the wall of the small intestine and matures into an adult worm. A long chain of proglottids, called the strobila, develops off of the scolex (Fig. 110-12). D. latum is the largest parasite of

humans, reaching 12 meters (40 feet) in length. The proglottids release eggs into the lumen that pass with the feces.

Clinical Features and Pathophysiology

Fish tapeworm is not invasive and causes no direct symptoms. The worm obtains nutrients by absorbing luminal contents through its surface. D. latum produces a substance that splits B12 from intrinsic factor in the intestine,112 interfering with host absorption of the vitamin. The tapeworm also avidly absorbs B12, effectively competing with its host’s use of the vitamin. D. latum is long-lived and, over time, can cause significant B12 deficiency in patients with limited dietary cobalamin. Rarely, severe B12 deficiency results in megaloblastic anemia and neurologic symptoms.

Diagnosis and Treatment

Fish tapeworm is diagnosed by identifying D. latum eggs in stool specimens. Occasionally, diagnosis is made because the patient passes proglottids and brings them in for identification or the worm is seen on endoscopy.113 Praziquantel is effective in a single oral dose of 10 mg/kg. Patients should be warned that they might pass a rather long worm two to five hours after taking the medication. Albendazole 400 mg each day for three days also kills the tapeworms.

TAENIA SAGINATA AND TAENIA SOLIUM

An estimated 80 million people are colonized with beef (Taenia saginata) or pork (Taenia solium) tapeworm. Colonization occurs by eating raw or undercooked meat infested with cysticerci. Tapeworms usually cause no symptoms and can surprise an endoscopist who finds the unsuspected jejunal or colonic inhabitant (Videos 110-4 and 110-5).114,115 Ingestion of T. solium eggs causes cysticercosis, a potentially fatal disease. Treatment is praziquantel or albendazole.

Epidemiology

Beef and pork tapeworm occur where livestock are exposed to untreated human waste and people eat raw or undercooked meat. Both parasites have a worldwide distribution, although infestations originating in the United States and Europe are rare. Beef tapeworm is endemic in Africa, the Middle East, Eastern Europe, Asia, and Latin America. Pork tapeworm is endemic in Africa, India, China, Asia, and Latin America. T. solium is rare in Muslim countries, where pork consumption is prohibited. T. solium is considered an eradicable parasite,116 though progress in such eradication is hindered by socioeconomic barriers.117

Life Cycle

Figure 110-12.  Endoscopic view of the cecum in a middle-aged woman with watery diarrhea after a fishing trip in Northern Canada during which she also frequently ate sushi. Part of a fish tapeworm is seen, and the strobila with maturing proglottids is visible. The worm, which was several feet long, was residing in the small intestine and was retrieved by suction. She was treated successfully with praziquantel. (Courtesy of Dr. Roy Joseph, MD, Denton, Tex.)

Adult tapeworms release gravid proglottids each containing up to 100,000 eggs that are released when the proglottid degenerates. Proglottids and eggs are passed with the stool. Proglottids of T. saginata remain motile and can crawl out of the feces, alarming the patient. Untreated human waste used to fertilize fields allows cattle to eat infective eggs on vegetation. Free-ranging pigs are coprophagous and directly consume poorly disposed human waste. Ingested eggs release an embryo (oncosphere) that penetrates the intestinal wall and enters the blood vessels or lymphatics. The oncospheres are carried to subcutaneous tissue, muscle, and organs, where they develop into cysticerci that can live for several years awaiting human consumption of infected meat. Once in the human intestine, the cysticercus evaginates to form a scolex that serves as the anterior attachment point of the tapeworm to the mucosa of the proximal jejunum. The worm develops over several months as proglottids form and mature in a chain behind

1931

1932

Section X  Small and Large Intestine the scolex, referred to as the strobila. Beef tapeworms can reach four to 10 meters in length, and pork tapeworms attain lengths of two to four meters. Mature gravid proglottids break away from the distal end of the worm, pass with the stool, and rupture to release eggs to complete the life cycle. Adult tapeworms can live in the small intestine for 25 years.

Clinical Features and Pathophysiology

Most people colonized with adult T. saginata or T. solium are asymptomatic; those with symptoms complain of mild abdominal discomfort, loss of appetite, or change in stool pattern. Colonization usually is limited to one worm that obtains nutrients by absorbing luminal contents through its surface. Motile proglottids can crawl out of the anus or “swim” in the toilet, eliciting immediate concern. Rarely, acute biliary or pancreatic duct obstruction can occur if proglottids migrate into these sites. The most feared complication of T. solium infestation is cysticercosis,118 which occurs when people inadvertently consume T. solium eggs. Just as in pigs, the eggs release oncospheres that penetrate the intestinal wall, disseminate through the body, and form cysticerci. Cysticerci produce localized inflammation in the brain, spinal cord, eye, and heart, with dire consequences. Neurocysticercosis is a common cause of epilepsy in countries where T. solium is endemic. Worldwide, an estimated 50,000 people die of neurocysticercosis each year. In the United States, 221 people died from cysticercosis between 1990 and 2002.119 Because the disease occurs after ingestion of parasite eggs, neurocysticercosis in a patient who has not visited or emigrated from an endemic country should prompt an effort to identify local carriers.

Diagnosis

Beef and pork tapeworm are diagnosed by identifying eggs or proglottids in stool specimens. The eggs of the two species are indistinguishable microscopically. The proglottids of T. saginata are two centimeters long and have more than 12 uterine branches; those of T. solium measure 1.2 cm and have fewer than 10 uterine branches.112 Egg and pro­ glottid production can be sporadic, necessitating repeated stool tests. Cysticercosis usually is diagnosed by computed tomography (CT) or magnetic resonance imaging (MRI) and confirmed by serology using a larval cyst antigen-specific immunoblot.120

Treatment

Praziquantel is effective in a single oral dose of 10 mg/kg. Albendazole 400 mg daily for three days also kills the tapeworms. The worms usually break apart and are passed as sections of disintegrating strobila. Patients with cysticercosis should be treated with albendazole 15  mg/kg daily for eight days to kill the cysticerci. Local inflammation transiently increases as cysticerci die. The addition of glucocorticoids prevents exacerbation of neurocysticercosis during therapy.

HYMENOLEPIS NANA AND HYMENOLEPIS DIMINUTA

Hymenolepis nana (dwarf tapeworm) is the smallest but most common tapeworm that colonizes people. It can be transmitted directly from person to person. Self-inoculation or internal autoinfection permits accumulation of a large number of worms that can cause anorexia, abdominal pain, and diarrhea. Hymenolepis diminuta (rodent tapeworm) is larger and rarely colonizes people. It is acquired by ingesting infested insects and usually causes no symptoms. Treatment is praziquantel.

Epidemiology

H. nana is the most common tapeworm of humans. Unlike other tapeworms, it can be transmitted from person to person without an intermediate host. Dwarf tapeworm has a worldwide distribution, with highest prevalence in warm and arid regions. A survey of Egyptian children found that 16% carried H. nana.121 In the United States, a 1987 survey of state diagnostic laboratories found that 900 of 216,000 submitted stool specimens demonstrated H. nana, with 34 states reporting positive specimens.122 H. nana also colonizes mice and rats; however, the strains that colonize people appear to differ from those of rodents. Human colonization with H. diminuta is rare, but it too has worldwide distribution. Rats and mice are the parasite’s usual hosts. People acquire rodent tapeworm by ingesting fleas, grain beetles, mealworms, or cockroaches infested with larval forms of the parasite. Most cases involve young children. The incorporation of beetles in traditional oriental medications also permits transmission.123

Life Cycle

H. nana does not require an intermediate insect host. Ingested eggs release oncospheres that invade the mucosa of the small intestine. They lodge within the lymphatics of the villi and develop into cysticercoid larvae. Each cysticercoid larva then ruptures into the lumen and evaginates a scolex that attaches to the mucosa of the ileum. The worms mature, growing a strobila or chain of developing proglottids. Adult worms average two centimeters in length and have about 200 proglottids, each containing about 150 eggs. The most distal proglottids disintegrate to release eggs into the lumen. About 20 to 30 days after initial ingestion, the worm begins to shed eggs in the stool. H. nana adults live for only four to six weeks, but eggs shed in the stool are immediately infective. Self-inoculation or internal autoinfection allows colonization to persist for years. Ineffective sanitation or poor hand washing permits transmission to others. Like other Hymenolepis species, H. nana can infest insects, forming cysticercoid larvae. Ingestion of infested fleas, beetles, mealworms, or cockroaches allows transmission of H. nana; however, acquisition by this pathway is rare, and most transmission is by direct ingestion of eggs. H. diminuta requires intermediate insect hosts. Insects ingest eggs as they consume rodent droppings. The eggs release oncospheres that penetrate into the insect’s viscera and form cysticercoid larvae. Rats and mice that eat infested insects acquire the tapeworm. People acquire rodent tapeworm the same way, by eating infested insects. Once in the intestine, the cysticercoid larva evaginates to form a scolex that attaches to the ileal mucosa. The worm matures, growing a strobila of proglottids and reaching a length of up to 90 cm. The most distal proglottids disintegrate, releasing eggs into the intestinal lumen.

Clinical Features and Pathophysiology

Most people colonized with H. nana or H. diminuta have no symptoms, but self-inoculation or internal autoinfection can cause heavy infestations with H. nana, resulting in anorexia, abdominal pain, and diarrhea. Mice can harbor Hymenolepis, permitting investigation of the mechanisms that limit worm density. It appears that a Th1-mediated IFN-γ response provides protective immunity against cysticercoid larvae,124,125 and a Th2 response involving IgE and mast cells assists in the expulsion of adult worms.126,127 The mucosal immune response to the tapeworm can alter intestinal inflammation elicited by other agents; for example, mice colonized with H. dimunita are protected from dinitrobenzene sulfonic acid (DNBS)-

Chapter 110  Intestinal Infections by Parasitic Worms induced colitis128 but are more susceptible to oxazaloneinduced colitis.129

Diagnosis and Treatment

Dwarf and rodent tapeworm are diagnosed by finding parasite eggs in the stool. H. nana eggs measure 30 to 47 µm in diameter. The eggs of the much less prevalent H. diminuta are larger, measuring 56 to 86 µm in diameter. Examination of several stool specimens taken on different days is needed to identify low-level colonization. Adults of both parasites can be killed with a single oral dose of praziquantel at 25  mg/kg, although eggs escape this treatment. Therefore, patients with H. nana infestation should be retreated one week after initial treatment. Family members also should be examined and considered for treatment.

DIPYLIDIUM CANINUM

D. caninum (dog tapeworm) is a common parasite of household pets that rarely colonizes children. It is acquired by eating fleas that contain parasite cysticercoid larvae. Dog tapeworm causes no symptoms in humans, but parents who find proglottids crawling in their child’s diaper understandably seek medical evaluation. Treatment is praziquantel. Echinococcus species also are tapeworms of dogs. Ingestion of Echinococcus granulosus, Echinococcus multilocu­ laris, or Echinococcus vogeli eggs causes severe disease due to formation of hydatid cysts (see Chapter 82).

Epidemiology

D. caninum is the most common tapeworm of domesticated dogs and cats, and it has a worldwide distribution. People acquire dog tapeworm by inadvertently ingesting fleas infested with the parasite. Most cases involve infants and young children who have close contact with their pets.

Life Cycle

Parasite eggs are ingested by the larval form of fleas that inhabit dogs or cats. Each egg releases an oncosphere that penetrates the gut wall and develops into a cysticercoid larva within the flea larva’s viscera. The insect larva then develops into an adult flea that can distribute the cysticercoid larva to other animals. Dogs, cats, and occasionally children ingest infested adult fleas. Once in the mammalian intestine, the cysticercoid larva evaginates to form a scolex that attaches to the mucosa of the small intestine. The worm matures, forming a strobila or chain of developing proglottids that trails the scolex. The adult worm measures 10 to 70 cm in length. Gravid proglottids detach from the distal end of the worm and pass with the stool. The proglottids look like cucumber seeds (12 by 3  mm), are motile, and occasionally move out of the anus. They can be mistaken for maggots. As they dry, they release small packets, each of which contain five to 15 eggs.

Clinical Features and Pathophysiology

Because people don’t often eat fleas, colonization is limited. Low numbers of dog tapeworms cause no symptoms. D. caninum is discovered when children or their parents find motile proglottids crawling in a diaper, underwear, or stool.

Diagnosis and Treatment

D. caninum is identified by its characteristic proglottid that looks like a moving cucumber seed. Often the proglottids of D. caninum are mistaken for adult pinworms (E. vermicu­ laris) because the latter is much more common than the former. Stool examination for egg packets usually is unrewarding. D. caninum causes a self-limited colonization that spontaneously clears, and dog tapeworm requires no treatment.

Most patients and their families, however, prefer that the parasite be expunged actively, and so treatment is given with a single oral dose of praziquantel 10  mg/kg, or niclosamide 500 mg (chewable tablet).

TREMATODES INTESTINAL FLUKES

Most intestinal trematodes have a broad host range, and more than 50 different species are capable of colonizing humans.130 Many of these are geographically restricted and are acquired because of specific indigenous dietary behavior. The more common intestinal trematodes are Fasciolop­ sis buski, Heterophyes species, and Echinostoma species. These parasites are acquired by ingesting larval metacercariae encysted on freshwater plants (F. buski) or in freshwater fish (Heterophyes, Echinostoma). The parasites usually cause no specific symptoms, but heavy infestations can cause diarrhea and abdominal pain. Treatment is with praziquantel.

Fasciolopsis buski

F. buski is the largest intestinal trematode that colonizes humans. Adults measure 7.5  cm long and 2  cm wide. F. buski is endemic in southeast Asia and Indonesia131 and is acquired by ingesting metacercariae encysted on freshwater plants. The metacercariae excyst in the duodenum and attach to the small intestinal mucosa. Within three months, they mature to adult flatworms and begin to lay eggs. The eggs pass with feces, and if they are deposited into fresh water, they embryonate. Each egg releases a ciliated miracidium that seeks a suitable snail to infect. The miracidium enters the snail and develops into a sporocyst that asexually multiplies, releasing numerous cercariae. The cercariae swim to freshwater plants, and each encysts to form a metacercaria on the plant’s surface, awaiting ingestion by a mammal. Adult F. buski live for about one year and cause no symptoms in most people.132 Histology of jejunal biopsy specimens along with carbohydrate, fat, and protein absorption were normal in one study of patients harboring F. buski133; however, in 1952, a 15-year-old Thai girl, hospitalized for diarrhea and abdominal pain, died of anasarca with more than 470 adult worms in her small intestine.134 Diagnosis is by finding parasite eggs in the stool (see Fig 110-3). Rarely the large flatworm is found on endoscopy (Video 110-6).135 Treatment is one dose of praziquantel 15 mg/kg given orally.

Heterophyes Species

Heterophyes species and the closely related Metagonimus yokogawai are small, flat worms measuring about 1.0 to 1.7  mm long by 0.3 to 0.6  mm wide. Heterophyes hetero­ phyes is endemic in west Africa, Egypt, Israel, Turkey, China, Japan, Taiwan, and the Philippines. Heterophyes nocens is endemic to Japan and Korea. M. yokogawai is endemic in Siberia, the Balkans, China, Korea, and Japan. People acquire these parasites by eating raw or undercooked fish that contain metacercariae. In the United States, a case of H. heterophyes involved a Pennsylvanian woman who ate sushi flown in from Asia.136 The metacercariae ingested in raw fish excyst in the intestine, attach to the small intestinal mucosa, and develop into adults. The adults lay eggs that are deposited with feces. If passed into fresh or brackish water, the eggs release miracidia that swim in search of a suitable snail. A miracidium enters a snail and develops into a sporocyst that asexually multiplies, releasing numerous cercariae. The cercariae swim away from the

1933

1934

Section X  Small and Large Intestine snail in search of a fish to infect. Either freshwater fish or saltwater fish feeding in brackish outlets can become infected. These parasites produce no specific symptoms in most people. Occasional heavy infections cause mild abdominal pain and diarrhea. The worms attach at the villus crypts and produce a localized eosinophilic inflammation. Rarely, parasite eggs enter blood vessels and lymphatics, producing distant granulomatous reactions. Diagnosis is by finding eggs in the stool, which can require concentration techniques. The eggs of Heterophyes species appear similar to those of M. yokogawai. Treatment of the trematodes is a single 20-mg/kg oral dose of praziquantel.

Echinostoma Species

There are at least 16 species of Echinostoma that can colonize humans.137 Adults are 2 to 6 mm long and 1 to 1.5 mm wide, depending on the species. Echinostoma species are endemic in Taiwan, Korea, Thailand, Japan, Indonesia, and the Philippines. One outbreak of probable echinostomiasis involved 18 of 20 American travelers returning from Kenya.138 People acquire Echinostoma by eating raw or undercooked freshwater mollusks or fish infected with metacercariae. The ingested metacercariae excyst in the intestine, attach to the small intestinal mucosa, and develop into adults. The adults lay eggs that, if deposited into fresh water with feces, embryonate and hatch to release miracidia that swim in search of suitable snails. A miracidium enters a snail and develops into a sporocyst that asexually multiplies, releasing numerous cercariae. Depending on the species, the Echinostoma cercariae swim away from the snail in search of another mollusk or fish to infect. Echinostoma species produce no symptoms in most people, but these parasites can cause epigastric pain, abdominal cramps, and diarrhea.138 Diagnosis is by finding eggs in the stool or adults on endoscopy.139 Echinostoma eggs resemble those of F. buski but are smaller. Treatment is one 25-mg/kg dose of praziquantel given orally.

LIVER FLUKES

These trematodes reside in the bile ducts and are acquired by ingesting larval metacercariae encysted in freshwater fish (Clonorchis sinensis, Opisthorchis) or on freshwater plants (Fasciola). Most infections are asymptomatic, but these parasites can cause recurrent cholangitis. People chronically infected with C. sinensis or Opisthorchis viverrini can develop cholangiocarcinoma. C. sinensis or Opisthorchis infections are treated with praziquantel. Fasciola infections are treated with triclabendazole.

Clonorchis sinensis, Opisthorchis viverrini, and Opisthorchis felineus

C. sinensis and Opisthorchis species are closely related parasites that have similar life cycles and cause similar disease. C. sinensis is endemic to China, Hong Kong, Taiwan, the Republic of Korea, and North Vietnam. O. viver­ rini is endemic to Thailand and Laos.140 O. felineus is endemic to Russia and the Ukraine. Infection with C. sinen­ sis and other food-borne trematodes is increasing in prevalence, possibly due to fish farming.141 People acquire these parasites by eating metacercariae present in raw or undercooked fish such as grass carp (Ctenopharyngodon idellus) or pond smelt (Hypomesus olidus). Studies in Korea show that at least 80 species of freshwater fish can harbor metacercariae.142 The metacercariae excyst in the stomach and duodenum as the meat is digested. The worms migrate along the mucosa

to the ampulla of Vater and into the biliary tree, where they grow into adults. Leaf-shaped adult C. sinensis measure 5 mm wide by 2.5 cm long by 1 mm thick. Opisthorchis is smaller. The adult parasites lay eggs that pass with the bile into the intestinal lumen to be excreted. The excreted eggs are ingested by freshwater snails in which they hatch, releasing miracidia that develop into sporocysts. Each sporocyst asexually reproduces within the snail, eventually producing numerous cercariae. The cercariae exit the snail and swim in search of a suitable fish to invade. The parasites encyst as metacercariae in the muscles of the fish, awaiting ingestion by a mammalian host. Most infections with C. sinensis or Opisthorchis are asymptomatic. With heavy exposures, patients develop fever, malaise, hepatic tenderness, and eosinophilia,143 symptoms and signs that abate as the worms mature and begin laying eggs in the bile ducts (Video 110-7).144 In a minority of patients, these parasites can cause relapsing cholangitis (see Chapter 82). The worms elicit a fibrotic and adenomatous reaction in the smaller branches of the biliary ducts, which can produce a localized obstruction and hepatic abscess. The flukes also can migrate into the pancreatic duct and cause pancreatitis. The most important complication of chronic infection with C. sinensis or O. viverrini is cholangiocarcinoma (see Chapter 69).145 Infection with these parasites dramatically increases the risk of developing this otherwise rare cancer (Table 110-2)146,147: Parasites damage the bile duct, causing cellular desquamation followed by hyperplasia, adenomatous hyperplasia, periductal fibrosis, dysplasia, and finally cholangiocarcinoma.151 Cancer can result from increased sensitivity to carcinogens. Hamsters infected with O. viver­ rini develop cholangiocarcinoma when treated with subcarcinogenic doses of dimethylnitrosamine. C. sinensis and O. viverrini can sensitize patients to dietary or endogenously produced N-nitroso compounds and thereby increase the risk for cholangiocarcinoma.152 This is an important consideration in Western countries as well. A 1977 study found that 26% of Chinese immigrants relocating to New York had C. sinensis.153 Because of the increased cancer risk associated with these parasites, it is advisable to look for them in any patient from an endemic area.154 Diagnosis is by finding parasite eggs in the stool or duodenal aspirate. Symptomatic patients might have curvilinear lucencies in the biliary and pancreatic ducts on ERCP.155 Ultrasound findings include increased periductal echogenicity and floating echogenic foci in the gallbladder.156 The recommended treatment is praziquantel 25 mg/kg every eight hours for three doses. Heavy infections may require Table 110-2  Relative Risks of Cholangiocarcinoma in Patients with Clonorchis or Opisthorchis Infestation reference Clonorchis sinensis 146 147 148 Opisthorchis viverrini 149 152*   Light   Medium   Heavy

RELATIVE RISK

95% CI

3.1 6.5 6.0

0.13-8.4 3.7-12 2.8-13

5.0

2.3-11.0

1.7 3.2 14.0

0.2-16.3 0.4-30 1.7-119

CI, confidence interval. *Light , <1500 eggs/g stool; medium, 1501-6000 eggs/g stool; heavy, >6000 eggs/g stool.

Chapter 110  Intestinal Infections by Parasitic Worms two days of therapy.157 An alternative treatment is albendazole 10  mg/kg twice a day for seven days. Albendazole is teratogenic and should not be given to pregnant women.

Fasciola hepatica and Fasciola gigantica

Fasciola hepatica has a worldwide distribution, whereas Fasciola gigantica is endemic in Hawaii, Asia, India, the Middle East, and Africa. Both species infect sheep, goats, and cattle as their normal hosts. Humans acquire these parasites by ingesting metacercariae encysted on freshwater plants such as watercress. Ingested metacercariae excyst in the small intestine, penetrate through the bowel wall, and enter the peritoneal cavity, where they migrate to the liver, penetrate the capsule, and travel through the hepatic parenchyma in search of a bile duct. They reside within the bile ducts, reaching maturity within three or four months, after which they lay eggs. Adult F. hepatica are 1.3  cm by 4.0  cm, and F. gigantica grow up to 7.0 cm in length. Adults of both species are only one millimeter thick and resemble leaves. Fasciola are longlived; one documented infection persisted for 16 years.158 Adults lay eggs that pass with the bile into the intestinal lumen, from which they are excreted. Upon reaching fresh water, Fasciola eggs embryonate, hatch, and release miracidia that swim in search of a suitable snail. A miracidium enters a snail and develops into a sporocyst that asexually multiplies, eventually releasing numerous cercariae. The cercariae swim to a freshwater plant and encyst on the wall, awaiting ingestion by a mammal. Fasciola infestations usually are asymptomatic. In the acute phase, patients can have abdominal pain and hepatomegaly as the parasites penetrate the intestinal wall and hepatic capsule. Abdominal CT scan may show low-density areas in the periphery of the liver. Patients also develop symptoms from migration of the parasites to other sites such as subcutaneous fat.159 Acute symptoms wane as the parasites enter the bile ducts. During the chronic phase of fascioliasis, patients can have symptoms of intermittent biliary obstruction and cholangitis. Rarely, patients develop pancreatitis. ERCP may show curvilinear lucencies in the bile duct (Fig. 110-13).160 Diagnosis is by finding eggs in the stool. Fasciola release low numbers of eggs, however, making this test insensitive. Duodenal or bile aspirates also can demonstrate eggs. The most sensitive method to detect Fasciola infection is ELISA for antibodies against the worms161; antibody titer drops after successful drug treatment. Unlike other trematodes, Fasciola are resistant to prazi­ quantel. Triclabendazole is the drug of choice for fascioliasis. In one study, a single oral dose of triclabendazole (10 mg/kg) cured 79% of patients as measured by fecal egg counts and ELISA.162

BLOOD FLUKES

Visceral (hepatosplenic and intestinal) schistosomiasis is caused by Schistosoma mansoni, Schistosoma japonicum, Schistosoma mekongi, and Schistosoma intercalatum. Schistosomes (including Schistosoma hematobium, which affects the urinary tract) infest more than 200 million people worldwide. People acquire the parasite through contact with contaminated water. Visceral schistosomiasis can cause colitis and fibrosis of the portal venous system, producing portal hypertension. Treatment is praziquantel.

Epidemiology

Schistosomes are tropical parasites with a worldwide distribution. S. mansoni is endemic in regions of Africa, the Middle East, Puerto Rico, the Dominican Republic, Central

Figure 110-13.  Fasciola hepatica on an endoscopic retrograde cholan­ giogram, appearing as curvilinear lucencies (arrows) in the distal bile duct. A leaf-shaped fluke was extracted from the bile duct. (From Veerappan A, Siegel JH, Podany J, et al. Fasciola hepatica pancreatitis: Endoscopic extraction of live parasites. Gastrointest Endosc 1991; 37:473.)

America, and South America. S. japonicum is endemic in China, Indonesia, the Philippines, and Thailand. S. mekongi is endemic in Laos and Cambodia. S. intercalatum is endemic in Africa. In most countries in which schistosomes are endemic, some regions have a high prevalence of infection; in other areas the parasite is absent. Schistosomes live in tropical snails for part of their life cycle. It is the distribution of these snails that helps define the geographic limits of schistosomes. Construction of water reservoirs and irrigation canals has expanded the snail habitat in many countries, a practice that has increased the risk of acquiring schistosomiasis. Mice and other mammals can harbor schistosomes and might allow spread of the parasite even were sanitation to be improved,163 thereby making it difficult to eradicate. Nonetheless, Japan successfully eradicated S. japonicum, and S. mansoni is vanishing from areas of Puerto Rico.164

Life Cycle

Schistosome worms are acquired by contacting fresh water infested with parasite cercariae. Cercariae are fork-tailed, microscopic larvae that swim through the water in search of a suitable mammalian host. Upon finding this host, they penetrate through intact skin, shed their tails, and transform into schistosomules that are covered with a double lipidbilayer tegument; this tegument thwarts most immunologic attacks (see later). Schistosomules migrate into blood vessels, where they are swept with the venous flow through the right side of the heart into the lungs. They migrate through the pulmonary capillaries, flow through the left side of the heart into the systemic circulation, and eventually reach the liver, where they mature, mate, and migrate against venous flow in the portal system. The twocentimeter female is partly ensheathed by the shorter male, and the “couple” reside together within the mesenteric veins. S. mansoni and S. intercalatum prefer to dwell in the

1935

1936

Section X  Small and Large Intestine vessels drained by the inferior mesenteric vein, whereas S. japonicum and S. mekongi prefer the vessels drained by the superior mesenteric vein. The worms remain in the mesenteric vessels, consuming blood and nutrients and depositing eggs. S. mansoni lays 250 eggs and S. japonicum lays 3500 eggs per worm pair each day. Many of the eggs pass through the intestinal wall and enter the lumen of the bowel. The eggs are excreted with the stool, and if deposited in fresh water, they hatch to release ciliated miracidia. Miracidia swim in search of a suitable tropical snail to infect. After penetrating into the snail’s foot process, a miracidium transforms into a primary (mother) sporocyst. Secondary sporocysts bud off of the primary sporocyst, migrate to the snail’s liver, and mature. Cercariae bud off the secondary sporocysts, exit the snail, and swim in search of a permissive mammalian host.

Clinical Features and Pathophysiology

Dermal invasion and migration by infecting cercariae usually produce no symptoms. Patients with repeated contact can develop a mild papular rash, in contrast to the intensely pruritic papular rash that develops after exposure to avian schistosomes such as Trichobilharzia ocellata. These avian trematodes infect water fowl but are unable to live in mammals, and so the cercariae and schisosomules die in a person’s skin, eliciting an immunologic response that produces swimmer’s itch. Swimmer’s itch is common in the Great Lakes region and has been found as far north as Iceland.165 Swimmer’s itch is not dangerous, but repetitive scratching can cause secondary cellulitis. Schistosomules migrate through the body without producing symptoms. Juvenile and adult worms evade immune attack elegantly: Their tegument is coated with histocompatibility and blood group antigens derived from the host.166 The tegument contains immunoglobulin receptors and proteases that might help cleave any bound antibody. Moreover, schistosomes produce several proteins that prevent complement, neutrophils, macrophages, or lymphocytes from injuring them.167 Such immune evasion allows adult worms to survive in the blood vessels without causing much direct damage. The average life span of worms is thought to be about five years, but there are documented cases of adult worms surviving for more than 35 years after persons had left an endemic area.168 Schistosome worms release eggs each day throughout their long life, and it is the parasite’s eggs that cause disease. Whereas the adult worms evade an immune response, the schistosome eggs invite one, exuding antigens that trigger a strong cell-mediated Th2 immune response. Katayama fever is the classic presentation of acute schistosomiasis. It results from a brisk early immune response to schistosome eggs that occurs within the first two weeks of egg deposition or from about 35 to 50 days after contacting water that is heavily infested with cercariae. Symptoms are caused by circulating immune complexes and resemble those of serum sickness. Patients have fever, malaise, arthralgia, myalgia, cough, and diarrhea, with the additional finding of marked eosinophilia. Serum aminotransferases are normal, and eggs usually are absent from the stool. S. japonicum releases the largest number of eggs and causes the most intense acute schistosomiasis reaction, with fatality rates approaching 25%. Most people don’t develop acute schistosomiasis, but in those who do and survive, symptoms resolve as the infection enters the chronic phase. Each schistosome egg secretes antigens that provoke a focal granulomatous inflammatory reaction that helps move the egg from the inside of a capillary, through the intestinal wall, and out into the lumen.169 Thus, inflammation actually

Figure 110-14.  Film from a barium enema in a 20-year-old Egyptian man with bloody diarrhea and tenesmus. Multiple polypoid lesions due to Schistosoma mansoni are seen throughout the rectosigmoid colon, which is displaced out of the pelvis by a large pericolic abscess. (From Reeder MM, Hamilton LC. Radiologic diagnosis of tropical diseases of the gastrointestinal tract. Radiol Clin North Am 1969; 7:57.)

benefits the parasite. Passage of eggs through the bowel wall causes intestinal schistosomiasis with guaiac-positive stools or even bloody diarrhea. Patients also can have tenesmus and tenderness over the sigmoid colon. Patients with S. mansoni can develop colitis with inflammatory pseudopolyps (Fig. 110-14) composed of numerous eosinophils and occasional eggs170—a picture that can resemble Crohn’s disease or ulcerative colitis. S. japonicum prefers to dwell in veins drained by the superior mesenteric vein and lays thousands of eggs at a time. S. japonicum can produce upper abdominal pain unrelated to meals, gastric bleeding, and pyloric obstruction due to inflammation and fibrosis. About half of the eggs pass out of the body; the other half lodge in the host’s tissues and cause the pathology of chronic schistosomiasis. Eggs are carried by the portal flow and some lodge in the liver. Other eggs lodge in the mesenteric and portal veins or remain in the intestinal wall. In these locations, the eggs elicit granulomatous inflammation with eosinophils, macrophages, lymphocytes, fibroblasts, and mast cells (Fig. 110-15). Eosinophils account for 50% of the schistosome egg granuloma cell population. When eosinophils degranulate, they deposit major basic protein that produces an eosinophilic halo around the eggs, termed the Splendore-Hoeppli phenomenon. This phenomenon is nonspecific and can be seen with bacterial, fungal, and parasitic infections. Eosinophils likely assist in killing the miracidia protected by the tough egg shell. After one or two weeks the miracidium dies, antigen release wanes, and the granuloma involutes to leave a fibrotic scar. Over the years, the daily production of eggs, granulomas, and scars accumulates enough damage to produce disease. Eggs that lodge in the hepatic and portal vessels produce a unique pattern of scarring called Symmers’ pipe stem fibro-

Chapter 110  Intestinal Infections by Parasitic Worms

Figure 110-16.  Histopathology of a colonic biopsy specimen showing viable Schistosoma mansoni eggs. A 20-year-old woman who emigrated five years earlier from the Democratic Republic of the Congo was evaluated for persistent iron-deficiency anemia. She had normal stool habits but occasional hematochezia. The mucosa was normal on colonoscopy except for some areas suggestive of neovascularization. Random biopsy specimens demonstrated viable Schistosoma mansoni eggs (20×; inset, 60×). She was treated with praziquantel and the anemia resolved. (Courtesy of P. Kirby and F. Mitros, Iowa City, Ia.)

Figure 110-15.  Histopathology of a colonic biopsy specimen from a patient with schistosomiasis. A schistosome egg granuloma is seen.

sis, in which the vessels become fibrotic and resemble clay pipe stems on cross section; this process causes the presinusoidal venous obstruction and portal hypertension characteristic of hepatosplenic schistosomiasis (see Chapter 82). Patients typically have an enlarged left hepatic lobe, splenomegaly, and thrombocytopenia due to platelet sequestration. Hepatocellular function remains normal because the blood supply to the liver is maintained by increased hepatic artery flow. Patients have normal serum aminotransferase levels and mildly elevated serum levels of alkaline phosphatase and gamma glutamyl transpeptidase. Patients with hepatosplenic schistosomiasis do not develop cirrhosis unless they are coinfected with hepatitis B or C, and so they lack stigmata of chronic liver disease. The classic presentation of decompensated hepatosplenic schistosomiasis is variceal hemorrhage. Hepatosplenic schistosomiasis results from accumulated injury and requires prolonged, moderately intense infection. Patients with hepatosplenic schistosomiasis typically range in age from adolescence to late 20s and have had schistosomiasis for five to 15 years. Compensated disease improves, however, after schistosomes are killed by drug therapy, permitting the portal tributaries to heal and remodel.171,172 Schistosome eggs also can lodge in other sites besides the intestine, liver, spleen, and splanchnic venous circulation. Eggs can percolate through portocaval collateral vessels, lodge in the pulmonary capillaries, and over time cause pulmonary hypertension and cor pulmonale. Eggs can enter the vertebral venous plexus and embolize the spinal cord or brain. Granulomatous inflammation in the CNS can result in conus equinus syndrome, transverse myelitis, or schistosomal cerebritis. Patients with schistosomiasis can present with recurrent bacteremia. Adult schistosome worms can ingest enteric bacteria transiently present in the portal circulation, harbor these bacteria, and serve as reservoirs for infection. Recurrent salmonella infection is particularly common in patients with schistosomiasis.173

Schistosomiasis can cause membranoproliferative glomerulonephritis or focal glomerulosclerosis with proteinuria, nephrotic syndrome, and end-stage renal disease. Schistosomal nephropathy results from deposition of immune complexes of parasite antigens and antibodies, and the renal disease can be progressive even if the parasites are killed with drug therapy.174

Diagnosis

Schistosome eggs are present in stool, but not in high numbers. The classic method for detecting eggs is the KatoKatz thick smear.175 This technique is not performed as part of the standard ova and parasite test, and standard evaluation is not sensitive enough to find the relatively rare schistosome eggs. Even Kato-Katz thick smears are not highly sensitive and are unlikely to detect eggs at very low levels of infection. The vast majority of patients with intestinal schistosomiasis are asymptomatic; patients come to medical attention during evaluation of mild anemia, positive fecal occult blood tests, or unexpected variceal hemorrhage. On endoscopy, a patient might have inflammatory polyps that contain eggs,170 but usually, the intestinal mucosa appears normal. Subtle changes in the vascular pattern can result from egg emboli producing a terminal curling of small blood vessels.176 Occasionally, histopathology of random biopsies of the colonic mucosa show schistosome eggs (Fig. 110-16), but this is an insensitive means of diagnosis. Biopsy of the rectum can demonstrate eggs, especially when the biopsy is crushed between two glass slides and the whole biopsy specimen is surveyed microscopically. Evaluation of six crush biopsies is more sensitive than two Kato-Katz smears for S. mansoni.177 Although eggs lodge in the liver and cause portal hypertension, liver biopsy is an insensitive method for detecting schistosomiasis. Liver biopsy should not be used solely to test for schistosomiasis but rather to stage comorbid disease such as viral hepatitis B or C. Present or past exposure to schistosomes is detectable by serology. Antischistosome antibodies are detected by ELISA using adult microsomal antigens. Sensitivity varies depending on whether the infecting schistosome is the same species as that used to prepare the antigens. The ELISA uses S.

1937

Normal structure Patterns observed in schistosomiasis* “Starry sky” (diffuse echogenic foci) Highly echogenic “ring echoes,” which correspond to the “pipestems” seen in a scan perpendicular to the one in which rings are seen (see Figure 82-4) Highly echogenic “ruff” around the portal bifurcation and main stem Highly echogenic “patches” extending from the main portal vein and branches into the parenchyma Highly echogenic “bands” and “streaks” extending from the main portal vein and its bifurcation to the liver surface, where they retract the organ surface. Patterns indicating pathology different from periportal fibrosis. If these are present, no score is given. Diffusely coarse liver texture, irregular liver surface, distorted hepatic veins, rounded caudal liver edge Diffusely increased liver echogenicity, loss of highly reflective edges of peripheral portal branches, possibly distal sound extinction, rounded caudal liver edge Other liver abnormalities

0 1 1 1 4, 8 4, 8 1, 4, 8 1, 4, 8 1, 4, 8

0 0 0 0 0 0 0 0 3-13

No sign of periportal fibrosis Incipient periportal fibrosis not excluded Periportal fibrosis possible Periportal fibrosis possible Periportal fibrosis Periportal fibrosis Advanced periportal fibrosis Advanced periportal fibrosis Advanced periportal fibrosis + portal hypertension

—

—

—

8

6

4

1 2

0

increase 0 to ≤2 SD increase 2 to ≤4 SD increase >4 SD

0-6 0-4 0-3 0-13

+ Ascites score = Portal Hypertension score

score = 0 score = 3

score = 0 score = 4

score = 0 score = 4 score = 6

score = 0 score = 3 score = 7

Portal vein score + Collateral vein score

This is the sum of the above three scores:

Calculate the final Portal Hypertension (PH) score

Ascites None vessels detected Ascites present

Collateral veins No collateral vessel detected Collateral vessels detected

Normal Dilatation Marked dilatation

Portal vein diameter: Adjust the value for height

Portal Hypertension (PH) Score

Add the preliminary PT score to the intermediate PT score This gives a final PT score in the range 1 (1 + 0) to 8 (1 + 7)

Calculate the final PT score

Adjust results for body height Normal range 2 SD or less above mean Increased >2 SD but ≤4 SD above mean Much increased >4 SD above mean This gives the intermediate PT score

Calculate the mean wall thickness (both walls) for the two (or three) vessels measured.

Measure the thickness of the wall of the second order portal branches

  Continue the examination

  Assign a preliminary PT score of 1

If liver parenchyma shows indications of periportal fibrosis

Periportal Thickening (PT) Score

*Staging is determined by assessing the degree of parenchymal changes, periportal thickening, and portal hypertension. † Combined patterns can exist and are assigned an IP score corresponding to the highest IP score of the two or three patterns. Adapted from Abdel-Wahab MF, Esmat G, Milad M, et al. Characteristic sonographic pattern of schistosomal hepatic fibrosis. Am J Trop Med Hyg 1989; 40:72.

0 1 2 4 2 4 6 8 4-8

Interpretation of the final score IP score PT score PH score Interpretation

Z

Y

X

F

E

D

B C

A

Liver Parenchymal patterns on ultrasonography and Image Pattern (IP) scores Pattern Sonographic Appearance IP Score

Table 110-3  World Health Organization Criteria for Staging Hepatosplenic Schistosomiasis*

1938 Section X  Small and Large Intestine

Chapter 110  Intestinal Infections by Parasitic Worms mansoni microsomal antigens, and immunoblot tests using antigens from S. japonicum and S. hematobium (the schistosome responsible for urinary schistosomiasis) also can be performed.178 The antibody assay also is useful to diagnose acute schistosomiasis (Katayama fever) because there are few or no eggs in the stool during the peak of the reaction. The ELISA does not distinguish active from prior infections, and therefore it is most useful for diagnosis in recent travelers rather than in expatriates. Because schistosomes can be long-lived, one-time treatment of antibody-positive patients is reasonable. Active infection can be demonstrated by detecting circulating schistosome gut-associated protein antigens CCA (circulating cathodic antigen) and CAA (circulating anodic antigen) in the patient’s serum.179 Serologic detection of CCA and CAA has an equivalent or higher sensitivity than the Kato-Katz thick smear, but each test misses some lowlevel infections.180 Measurement of circulating antigens also can prove useful to document response to treatment,181 but these tests are not yet commercially available in the United States. Abdominal ultrasound is an important additional test in hepatosplenic schistosomiasis. Ultrasound evaluation documents periportal fibrosis, splenomegaly, portal blood flow, and collateral vessels. Periportal fibrosis has a characteristic appearance: multiple echogenic areas, each with central echolucency.182,183 A scoring system exists that uses a liver parenchyma and image pattern (IP), a portal thickening (PT), and a portal hypertension (PH) score to stage the disease (Table 110-3).184

Treatment

Praziquantel is the drug of choice to treat schistosomiasis. It is the safest schistosomicide in current use. Praziquantel given orally in three doses of 20  mg/kg, each four hours apart (total dose, 60 mg/kg), gives the best cure rates of 60% to 98%, depending on the series. Eggs continue to be shed in the stool for up to two weeks after drug treatment, because eggs that were deposited before treatment can take this long to work through the intestinal wall. Patients who are not cured with a single course of praziquantel have a dramatic decrease in egg counts and respond to a second course of

treatment. Periportal fibrosis improves after the worms are killed, halting the daily deluge of eggs and permitting the portal tributaries to heal and remodel.171,172

KEY REFERENCES

Audicana MT, Kennedy MW. Anisakis simplex: From obscure infectious worm to inducer of immune hypersensitivity. Clin Microbiol Rev 2008; 21:360-79. (Ref 99.) Boulware DR, Stauffer WM, Hendel-Paterson BR, et al. Maltreatment of Strongyloides infection: Case series and worldwide physicians-intraining survey. Am J Med 2007; 120:545-8. (Ref 41.) Crompton DW. How much human helminthiasis is there in the world? J Parasitol 1999; 85:397-403. (Ref 18.) Das CJ, Kumar J, Debnath J, Chaudhry A. Imaging of ascariasis. Australas Radiol 2007; 51:500-6. (Ref 22.) Dick TA, Nelson PA, Choudhury A. Diphyllobothriasis: Update on human cases, foci, patterns and sources of human infections and future considerations. Southeast Asian J Trop Med Public Health 2001; 32(Suppl 2):59-76. (Ref 109.) Elliott DE, Summers RW, Weinstock JV. Helminths and the modulation of mucosal inflammation. Curr Opin Gastroenterol 2005; 21:51-8. (Ref 5.) Garcia HH, Del Brutto OH. Neurocysticercosis: Updated concepts about an old disease. Lancet Neurol 2005; 4:653-61. (Ref 120.) Kaewpitoon N, Kaewpitoon SJ, Pengsaa P, Sripa B. Opisthorchis viver­ rini: The carcinogenic human liver fluke. World J Gastroenterol 2008; 14:666-74. (Ref 140.) Keiser J, Utzinger J. Emerging foodborne trematodiasis. Emerg Infect Dis 2005; 11:1507-14. (Ref 141.) Keiser PB, Nutman TB. Strongyloides stercoralis in the immunocompromised population. Clin Microbiol Rev 2004; 17:208-17. (Ref 37.) Loukas A, Constant SL, Bethony JM. Immunobiology of hookworm infection. FEMS Immunol Med Microbiol 2005; 43:115-24. (Ref 57.) Pozio E, Zarlenga DS. Recent advances on the taxonomy, systematics and epidemiology of Trichinella. Int J Parasitol 2005; 35:1191-204. (Ref 88.) Richter J, Hatz C, Campagne G, et al. Ultrasound in Schistosomiasis: A Practical Guide to the Standardized Use of Ultrasonography for the Assessment of Schistosomiasis-related Morbidity. World Health Organization: Second International Workshop, Niamey, Niger 2000. (Ref 184.) Sorvillo FJ, DeGiorgio C, Waterman SH. Deaths from cysticercosis, United States. Emerg Infect Dis 2007; 13:230-5. (Ref 119.) Varkey P, Jerath AU, Bagniewski S, Lesnick T. Intestinal parasitic infection among new refugees to Minnesota, 1996-2001. Travel Med Infect Dis 2007; 5:223-9. (Ref 1.) Full references for this chapter can be found on www.expertconsult.com.

1939

CHAPTE R

111 Crohn’s Disease Bruce E. Sands and Corey A. Siegel

CHAPTER OUTLINE History of Crohn’s Disease  1941 Epidemiology  1942 Etiology and Pathogenesis  1942 Initiating Events  1942 Genetics  1943 Environment  1946 Adaptive Immune Response and Inflammation  1946 Pathology  1947 Early Findings  1948 Later Findings  1948 Other Findings  1949 Clinical Features  1949 Disease Location  1949 Clinical Presentation  1949 Disease Behavior  1950 Classification of Disease  1952 Pathophysiology of Common Symptoms and Signs  1952 Extraintestinal Manifestations  1953 Differential Diagnosis  1955

Idiopathic inflammatory bowel disease (IBD) comprises conditions characterized by chronic or relapsing immune activation and inflammation within the gastrointestinal tract. Crohn’s disease and ulcerative colitis (UC) are the two major forms of idiopathic IBD; less common, but increasingly recognized, are the microscopic colitides, primarily collagenous colitis and lymphocytic colitis (see Chapter 124). Other chronic inflammatory conditions of the intestine share some features of presentation and pathogenesis with idiopathic IBD, but they have identifiable etiologies. These disorders include diversion colitis, bypass enteropathy, radiation colitis, and drug-induced colitides. The two major forms of IBD share many clinical and epidemiologic characteristics, suggesting that underlying causes may be similar. Indeed, more than occasionally, Crohn’s disease cannot be distinguished from UC on clinical grounds, yet the two diseases are distinct syndromes with divergent treatment and prognosis. Crohn’s disease is a condition of chronic inflammation potentially involving any location of the alimentary tract from mouth to anus, but with a propensity for the distal small bowel and proximal large bowel. Inflammation in Crohn’s disease often is discontinuous along the longitudinal axis of the intestine and can involve all layers from mucosa to serosa. Affected persons usually experience diarrhea and abdominal pain, often accompanied by weight loss. Common complications include strictures and fistulas, which often necessitate surgery. Numerous extraintestinal manifestations also may be present. The etiology of Crohn’s disease is incompletely understood, and therapy, although generally effective in alleviating the symptoms, is not curative.

Establishing the Diagnosis and Evaluating Disease Activity  1956 Differentiating Crohn’s Disease from Ulcerative Colitis  1958 Measuring Disease Activity  1959 Treatment  1960 Goals  1960 Medical Therapy  1960 Nutritional Therapy  1970 Surgical Therapy  1971 Costs of Care  1971 Crohn’s Disease in the Life Cycle  1971 Children and Adolescents  1971 Sexuality, Fertility, and Pregnancy  1972 The Aging Patient  1972 Prognosis  1972 Morbidity  1972 Cancer  1972 Mortality  1973 Coping with Crohn’s Disease  1973

HISTORY OF CROHN’S DISEASE Although the eponym Crohn’s disease has gained general acceptance in recent decades, clear clinicopathologic reports of the same process date back at least two centuries. Morgagni provided a description of intestinal inflammation characteristic of Crohn’s disease in 1761.1 Only after the identification of the tubercle bacillus by Koch in 1882 was it possible to describe persons with ileocecal disease similar to intestinal tuberculosis but lacking the organism. Such reports were provided by Fenwick (1889),1 Dalziel (1913),2 Weiner (1914), Moschcowitz and Wilensky (1923 and 1927), and Goldfarb and Suissman (1931).3 In 1932, the landmark publication of Crohn, Ginzburg, and Oppenheimer called attention to “terminal ileitis” as a distinct entity and chronic disease.4 This term was soon deemed unsuitable, however, when it became apparent that the disease process might involve the colon. Patients, too, misunderstood and were frightened by the “terminal” nature of their illness. The term “regional enteritis” embraced the focal nature of the process, but failed to incorporate knowledge of the possibility of disparate sites of involvement within the gastrointestinal tract, including the small and large bowel in combination5 and large bowel in isolation.6 The term “granulomatous enterocolitis” lost acceptance when it became clear that granulomas were not a sine qua non of the diagnosis. In the end, the name “Crohn’s disease” has been adopted to encompass the many clinical presentations of this pathologic entity. But for the alphabetic priority these authors chose, Crohn’s disease might well have been Ginzburg’s or Oppenheimer’s disease.

1941

1942

Section X  Small and Large Intestine EPIDEMIOLOGY Accurate comparisons of epidemiologic data on the incidence and prevalence of Crohn’s disease are hampered by a lack of gold-standard criteria for diagnosis and inconsistent case ascertainment. Moreover, the invasiveness and expense of diagnostic modalities ensures that diagnosed cases represent only a fraction of the diseased population. Finally, studies relying on the observations of large referral centers may be biased toward reporting more-aggressive forms of the disease, thereby underestimating its true incidence. Misclassification of disease is problematic. Historically, unidentified infections, later recognized by improved culture and diagnostic techniques, might have accounted for some portion of cases, particularly among persons with a single episode of disease. At times, differentiating Crohn’s disease from UC may be difficult, especially at the time of diagnosis, and before the passage of time has allowed distinctive disease characteristics to become manifest. Reassignment of a diagnosis of Crohn’s disease or UC may be as high as 9% in the first two years after diagnosis.7 Despite these methodologic limitations, distinct and reproducible geographic and temporal trends in incidence have been observed. In both North America and Europe, higher incidence rates have been noted in more northern latitudes. For example, age-adjusted annual incidence rates of 9 per 100,000 persons were reported in Olmsted County, Minnesota,8 and Copenhagen County, Denmark,9 and as high as 20 per 100,000 in Nova Scotia.10 Comparatively, estimates of incidence rates reached only 0.9 per 100,000 in Spain11 and 3.4 per 100,000 in Italy.12 A north-south gradient similar to that observed in Europe13 has been noted in the United States14 and even within the state of California itself, with estimated incidence rates of 7.0 and 3.6 per 100,000 in northern and southern California, respectively.15,16 In Asian countries, the incidence rate has remained low, with a mean estimated incidence of 0.5 per 100,000 persons in Korea17 and a similar incidence in Japan,18,19 whereas in Australia and New Zealand, incidence rates have ranged from 1.75 to 2.1 per 100,000.20,21 Crohn’s disease is thought to be extremely rare in much of South America and Africa22 with the exception of South Africa, where the most recent estimate of the incidence rate for the white population is 2.6 per 100,000; it is considerably lower among the nonwhite population.23 In some regions of the world where Crohn’s disease was rare, although still low compared with western countries, incidence is rising dramatically. For example, in Seoul, South Korea, the incidence increased from 0.05 per 100,000 in the early 1980s to 1.34 per 100,000 between 2001 and 2005.17 This trend has been seen throughout other regions as well.24-27 Estimates from less-affluent nations may be influenced by increasing access to health care, and therefore, genetic and environmental factors in these regions are difficult to disentangle. Incidence rates have continued to rise in some regions, such as in Denmark,9,28 whereas in others they appear to be stabilizing. In Olmsted County, Minnesota, rates had been steadily increasing from approximately 3 per 100,000 (19541963) to nearly 8 per 100,000 (1964-1973), but since the late 1970s these rates have not changed significantly.8,29 Mortality trends for Crohn’s disease have followed a similar pattern, with rising mortality until the mid-1970s and stable rate since.30 Although improved diagnostic capabilities might have played some role in the rising incidence leading up to the mid-1970s, the fact that Crohn’s-related mortality

was rising in parallel argues against the theory that rising Crohn’s diagnoses merely represented detection bias involving mild cases. Most recently, the prevalence of Crohn’s disease in the United States is estimated to be 201 per 100,000 adults and 43 per 100,000 in children, adolescents, and adults younger than 20 years.14 Studies throughout the world have shown a small excess risk of Crohn’s disease among women. Most reports show a female-to-male ratio in adult patients between unity and 1.3 : 1.10,14,28 In the pediatric population this is reversed, with more boys having Crohn’s disease than girls.14 This slight difference in risk in adult-onset disease may be explained by hormonal or life-style factors and stands in contrast to the nearly equal or even slight male predominance seen in UC. Crohn’s disease is diagnosed most often among persons 15 to 30 years of age, although the age of diagnosis can range from early childhood throughout an entire lifespan. Population-based studies have shown the median age of diagnosis to be approximately 30 years.8,10 Conflicting information may be found regarding trends in the age of diagnosis. In Olmsted County, Minnesota, younger age groups have had a fairly stable incidence over the past 20 years, with rising rates in patients aged 60 and older.8,10 This trend of older-age diagnoses also was seen in population-based studies in Copenhagen, Denmark,31 and in Stockholm, Sweden,32 the median age of diagnosis increased from 25 years in 1960 to 1964 to 32 years in 1985 to 1989. These findings reflect diagnosis in a larger proportion of patients older than 60 years. Indeed, many, though not all,10 studies have shown a smaller second peak in incidence later in life, generally in the seventh decade.33 This second peak may be the result of ascertainment bias because of more frequent contact with medical care and more frequent evaluation of older patients. Differences in clinical presentation among younger and older patients suggest that distinct risk factors are operative at different ages at onset.34 The pathologic findings in young and old patients are not different, although some studies have identified a greater proportion of colonic and distal disease among older patients,33 compared with a predominance of ileocolonic disease in younger patients.35

ETIOLOGY AND PATHOGENESIS INITIATING EVENTS

In light of the nature of the pathologic findings in Crohn’s disease (see later) and UC, it long has been clear that IBD represents a state of sustained immune response. The question arises as to whether this is an appropriate response to an unrecognized pathogen or an inappropriate response to an innocuous stimulus. Many infectious agents have been proposed as the cause of Crohn’s disease including chlamydia, Listeria monocytogenes, cell-wall–deficient Pseudomonas species, reovirus, and many others. Paramyxovirus (measles virus) has been implicated etiologically in Crohn’s disease as a cause of granulomatous vasculitis and microinfarcts of the intestine,36 but a proposed association between early measles vaccination and Crohn’s disease largely has been disproved.37 Another suggestion has been that the commensal flora, although normal in speciation, possess more subtle virulence factors, such as enteroadherence, that cause or contribute to IBD.38 Among the most enduring hypotheses is that Mycobacterium paratuberculosis is the causative agent of Crohn’s disease. This notion dates to Dalziel’s observation in 1913

Chapter 111  Crohn’s Disease that idiopathic granulomatous enterocolitis in humans is similar to Johne’s disease, a granulomatous bowel disease of ruminants caused by M. paratuberculosis.39 M. paratuberculosis is extremely fastidious in its culture requirements, and some proponents of this hypothesis have speculated that the presence of M. paratuberculosis as a spheroplast may have confounded efforts to confirm this hypothesis by culture of the organism; demonstrating it by immuno­ histochemistry, in situ hybridization, and polymerase chain reaction methodology; and empirical treatment with antimycobacterial antibiotics (see “Medical Therapy,” later). Most investigation in this area has been inconclusive, providing insufficient evidence to prove or reject the hypothesis. In light of the diversity of substances and bacteria within the intestinal lumen, it is remarkable that the intestine is not perpetually inflamed. The presence of low-level physiologic inflammation within the healthy intestinal mucosa represents a state of preparedness to deal with potentially harmful agents. A more vigorous response would not be appropriate if directed toward the innocuous commensal flora of the intestine. Experiments in animal models of IBD suggest that in a genetically susceptible host, a classic pathogen is not necessary to cause IBD, but rather nonpathogenic commensal enteric flora are sufficient to induce an inappropriate chronic inflammatory response. In diverse models, animals raised under germ-free conditions show diminished or delayed expression of the IBD phenotype.40 On introduction of defined bacterial flora, however, the expected phenotype of bowel inflammation becomes manifest.40 Such models suggest that a diversity of genetic alterations, including those that affect intestinal barrier function and regulation of mucosal immunity, can result in intestinal inflammation. As in the animal models of IBD, evidence in patients with Crohn’s disease also points to an over-responsiveness of mucosal T cells to the enteric flora, manifest in part by the presence of antibodies against an array of bacterial antigens, including Escherichia coli outer membrane porin C (OmpC) and flagellin. Advances in the genetic bases of Crohn’s disease confirm that diverse bacterial agents are capable of fueling the inflammation of Crohn’s disease (see later).

GENETICS

The argument for a genetic predisposition to IBD begins with the observation that family members of affected persons are at greatly increased risk for developing IBD. The relative risk among first-degree relatives is 14 to 15 times higher than that of the general population.41 Roughly one of five patients with Crohn’s disease report having at least one affected relative. Many families have more than one affected member, and although there is a tendency within families for either UC or Crohn’s disease to be present exclusively, mixed kindreds also occur, suggesting the presence of some shared genetic traits as a basis for both diseases, as has recently been confirmed. Ethnicity also plays a role. Eastern European (Ashkenazi) Jews are at a two- to four-fold higher risk of developing IBD than non-Jews of the same geographic location, and they are at greater risk of having multiple affected family members. Studies of monozygotic and dizygotic twins suggest that genetic composition is a more powerful determinant of disease for Crohn’s disease than for UC: The concordance rate among monozygotic twins is as high as 67% for Crohn’s disease but only 13% to 20% for UC. Most studies have suggested that concordance of disease location and disease behavior are higher than one would expect by chance.42 Finally, some subclinical markers of Crohn’s disease,

including anti-OmpC antibodies, are more common among apparently healthy family members of Crohn’s disease probands than among the general population.43 Early studies of IBD genetics were limited by the slow speed of DNA sequencing and an incomplete understanding of the human genome’s structure. Studies of IBD genetics depended upon selecting a candidate gene based upon the then-current understanding of pathogenesis. With automated, rapid DNA sequencing, and mapping of the common genetic variants that occur in humans (the HapMap, or haplotype map), genome-wide association (GWA) studies became feasible. GWA studies can simultaneously explore many hundred thousands of genetic markers, providing a broad and unbiased approach to assessing the association of genomic loci to a specific disease without prior hypothesis about a candidate gene. GWA studies have been more successful in Crohn’s disease than in any other complex disease, and they have accelerated the pace of gene dis­ covery, providing unexpected insights into pathogenesis.44 More than 30 genetic loci have been convincingly associated with Crohn’s disease in a large and powerful GWA study, confirming some known loci and discovering many loci that had not been described previously (Table 111-1).45 Thus far, genetic studies have highlighted three pathways of fundamental importance in the pathogenesis of Crohn’s disease. The first susceptibility locus for Crohn’s disease was identified in 2001 as the NOD2 (nucleotide-binding oligomerization domain 2) gene, also known as CARD15 (caspase-recruitment domain 15).46,47 The allelic variants most commonly associated with Crohn’s disease in European and American populations include one frameshift insertion leading to early truncation of the protein (Leu1007­ fsinsC) and two missense mutations (Arg702Trp, Gly908­ Arg). Carriage of disease-associated allelic variants on both chromosomes confers an odds ratio for Crohn’s disease of 17.1 (95% confidence interval [CI], 10.7-27.2), and heterozygous persons have an odds ratio of 2.5 (95% CI, 2.0-2.9) for the disease.48 Studies have associated genetic polymorphisms of NOD2/CARD15 with younger onset and ileal location of disease and increased likelihood of stricture formation.49-52 It has been estimated that as many as 20% to 30% of patients with Crohn’s disease bear abnormal NOD2/ CARD15. Nevertheless, the penetrance of NOD2/CARD15 is estimated to be less than 1%53; that is, disease-related allelic variants of the gene may be found in a large number of persons who do not have Crohn’s disease; this strongly suggests that environmental factors, as yet incompletely elucidated, play a significant role in the expression of the Crohn’s disease phenotype (see later). The discovery of the association of NOD2/CARD15 with Crohn’s disease has opened a remarkable window into the pathogenesis of Crohn’s disease. The gene product of NOD2/ CARD15 is a cytosolic protein that functions as an intracellular sensor of bacteria. Specifically, the protein binds to muramyl dipeptide (MDP; MurNAc-L-Ala-D-isoGln), a component of bacterial peptidoglycan, found in Gram-positive and Gram-negative bacteria.54,55 The NOD2/CARD15 protein is expressed in monocytes and enterocytes, specifically in Paneth cells,56 which lie within the crypts and produce the endogenous antimicrobial peptides called defensins. The NOD2/CARD15 gene consists of two CARD domains, a nucleotide binding domain (NBD), and 10 leucine-rich repeats (LRR). NOD2/CARD15 variants associated with Crohn’s disease lie within the LRR and interfere with binding to MDP. In mononuclear cells, mutations in NOD2 result in decreased activation of nuclear factor (NF)-κB, whereas an excess of NF-κB expression is observed in tissue

1943

1944

Section X  Small and Large Intestine Table 111-1  Risk Loci Implicated in Genome-Wide Association Studies of Crohn’s Disease

CHROMOSOME

GENES OR LOCI

DESCRIPTION

1p13

PTPN22

1p31

IL23R

1q23

ITLN1

1q24 1q32 2q37

— 3 genes ATG16L1

3p21

35 genes

5p13

—

5q31 5q33

7 genes IRGM

5q3

IL12B

6p22

CDKAL1

6q21 6q27

— CCR6

7p12 8q24 9p24

— — JAK2

9q32

TNFSF15

10p11 10q21 10q24

3 genes ZNF365 NKX2-3

11q13

C11orf30

Protein tyrosine phosphatase, nonreceptor type 22 Recognized as a risk locus in rheumatoid arthritis and type 1 diabetes mellitus but as conferring protection in Crohn’s disease Interleukin (IL)-23 receptor IL-23 induces IL-17, which promotes the inflammatory response in the mucosa; the pathway also is implicated in maintaining intestinal barrier function Intelectin-1 A lectin expressed in small and large bowel that recognizes galactofuranosyl residues present in microbial cell walls No recognized gene product Significance unknown Involved in autophagy, by which there is clearance of some intracellular pathogens, as well as integration of innate and adaptive immune responses MST1, which encodes macrophage stimulating protein 1 (hepatocyte growth factor-like protein) and regulates production of proinflammatory mediators, is a gene of interest This region is devoid of genes (a gene “desert”) Studies suggest the risk locus modulates expression of PTGER4 (prostaglandin E receptor 4), which promotes epithelial restitution Significance unknown Immunity-related GTPase family M Involved in autophagy The gene product is the p40 subunit common to the heterodimeric cytokines IL-12 and IL-23 Function of the encoded protein is not known A different allele of this gene is associated with type 2 diabetes mellitus No recognized gene product Chemokine receptor 6, expressed by immature dendritic cells and memory T cells and, in Crohn’s disease, by IL-17–producing Th17 cells and IFN-γ–producing Th1 cells The ligand is macrophage inflammatory protein 3 α No recognized gene product No recognized gene product Janus kinase 2 Gene product is involved in signal transduction in the IL-23 pathway in response to certain cytokines and growth factors TNF superfamily 15, also called TL1A The gene product is a cytokine induced by TNF and capable of activating NF-κB Significance unknown Encodes zinc finger protein 365 NK2 transcription factor–related, locus 3 Targeted disruption of this transcription factor gene causes defects in intestinal anatomy and disruption of secondary lymphoid organs in mice Mice with a mutation in this gene lack MAdCAM-1, which is necessary for lymphocyte homing to MALT and spleen The gene product of chromosome 11, open reading frame 30, is a regulatory protein capable of repressing transcription and likely participating in DNA repair processes

RISK ALLELE FREQUENCY IN CONTROLS

ODDS RATIO IN CASE-CONTROL STUDY

0.899

1.31

0.933

2.50

0.682

1.14

0.243 0.697 0.533

1.19 1.18 1.28

0.271

1.20

0.125

1.32

0.425 0.090

1.25 1.33

0.708

1.11

0.780

1.21

0.289 0.463

1.17 1.21

0.678 0.619 0.348

1.20 1.08 1.12

0.677

1.22

0.345 0.387 0.478

1.16 1.25 1.20

0.386

1.16

Chapter 111  Crohn’s Disease Table 111-1  Risk Loci Implicated in Genome-Wide Association Studies of Crohn’s Disease—cont’d

CHROMOSOME

GENES OR LOCI

DESCRIPTION

12q12

3 genes

13q14 16q12

3 genes NOD2

17q21 17q21

17 genes 4 genes

18p11

PTPN2

21q21 21q22

— ICOSLG

Genes of interest in this region include LRRK2, which encodes leucine-rich repeat kinase 2, and MUC19, a member of the mucin gene family Autophagy is believed to be induced by mutant LRRK2 Significance unknown Role in innate immune response through binding of intracellular muramyl dipeptide, a component of bacterial cell walls ORMDL3 is a gene of interest STAT3 is a gene of interest, and the gene product, STAT3, is downstream from JAK2 in the JAK-STAT signaling pathway In animal models of colitis, activation of STAT3 in innate immune cells leads to improved mucosal barrier function, whereas activation of STAT3 in T cells exacerbates colitis Protein tyrosine phosphatase, nonreceptor type 2 Associated with various immune and inflammatory conditions No recognized genes Gene product is inducible T-cell costimulator ligand expressed on intestinal epithelial cells It may have a role in antigen presentation and regulation of T cells

RISK ALLELE FREQUENCY IN CONTROLS

ODDS RATIO IN CASE-CONTROL STUDY

0.017

1.54

0.221 0.152

1.25 3.99

0.473 0.565

1.12 1.18

0.152

1.35

0.565 0.389

1.18 1.13

GTPase, guanosine triphosphatase; IFN, interferon; IL, interleukin; JAK, Janus kinases; MAdCAM-1, mucosal addressin adhesion molecule 1; MALT, mucosa-associated lymphoid tissue; NF-κB, nuclear factor-κB; STAT, signal transducers and activators of transcription; Th1, type 1 T-helper cell; TNF, tumor necrosis factor.

inflamed by Crohn’s disease. This apparent paradox has yet to be unraveled completely, but it is clear that defects in NOD2 impair antibacterial responses, particularly to oral exposure to pathogens. Notably, the production of β-defensins, which are antibacterial proteins produced by Paneth cells, is defective in Crohn’s patients with variant NOD2.57 These findings strongly implicate defects in innate immunity—the immediate and nonspecific immune responses to microbial infection—in a subset of patients with Crohn’s disease, with subsequent chronic activation of adaptive immunity, the antigen-specific responses mediated by antigen presenting cells (APCs) and T cells. Subsequent discoveries have implicated defects in multiple genes in the autophagy pathway in the pathogenesis of Crohn’s disease.58,59 Autophagy is an ancient cellular process, highly conserved in evolution, by which segments of cytoplasm are isolated within a membrane and delivered to lysosomes by mechanisms that do not involve transport through endocytic or vacuolar sorting pathways. This unique process plays a role in cellular homeostasis by clearing proteins that are long-lived, misfolded, or aggregated, and by clearing apoptotic bodies, which might otherwise trigger inflammation and autoimmunity. Autophagy has been shown to contribute directly to innate immunity through direct killing of pathogens, activation of Toll-like receptors and Nodd-like receptors, and elaboration of immunomodulatory cytokines such as interferon (IFN)-γ. Autophagy also stands at the interface of innate and adaptive immune responses, delivering antigen to human leukocyte antigen (HLA) class II molecules in APCs for antigen-specific binding.59

GWA studies have identified variants that predispose to Crohn’s disease in at least two autophagy-related genes. The first, the autophagy-related 16-like 1 (ATG16L1) gene, was noted as having a disease-associated single nucleotide polymorphism (SNP) encoding an amino acid substitution in exon 8, resulting in a change from alanine to threonine58-60; this minor allele is protective against Crohn’s disease. ATG16L1 is expressed by intestinal epithelial cells, APCs, and various subsets of human T cells. The second autophagy gene associated with Crohn’s disease is the IRGM (immunity-related GTPase [guanosine triphosphatase] family member M) gene on chromosome 5q33.1.61 Careful study suggests that the disease-associated variants of this gene do not affect the amino acid sequence of its product, but they more likely alter its expression.61 IRGM appears to be important in resistance to intracellular pathogens such as mycobacteria, Listeria monocytogenes, and Toxoplasma gondii.59 A third pathway associated with Crohn’s disease is interleukin (IL)-23 and other gene products associated with this protein.62 IL-23 is a heterodimeric cytokine comprising two linked subunits (p19 and p40). IL-23 is produced by many cell types, including dendritic cells and macrophages, in response to diverse microbial signals. Naïve CD4+ T cells up-regulate IL-23 receptor when exposed to IL-6 and transforming growth factor (TGF)-β, completing an autocrine loop in the generation of Th17 T cells, effector T cells that produce IL-17.63,64 A rare variant of the IL23R gene leading to a glutamine at position 381 rather than an arginine is strongly protective for Crohn’s disease, with an odds ratio of 0.26 to 0.45; other, more common SNPs are associated

1945

1946

Section X  Small and Large Intestine with increased risk for Crohn’s disease and UC.65 In the same pathway, variants of the IL12B gene, encoding the p40 subunit common to IL-12 and IL-23, and of the JAK2 and STAT3 genes, with roles in IL23R signaling, as well as in Th17 differentiation in the case of STAT3, also have been associated with Crohn’s disease susceptibility.45 Together, these findings support the pivotal role of this pathway in maintaining mucosal homeostasis in the normal intestine. As the functional alterations associated with the many other identified genetic risk loci are elucidated, it is certain that new insights into the causes of Crohn’s disease will arise.

ENVIRONMENT

Although the greatest relative risk of Crohn’s disease is found among first-degree relatives of affected persons, particularly siblings of the proband, environmental factors also are important. As noted earlier, the rising incidence of Crohn’s disease over many decades highly suggests an environmental contribution to the expression of disease. Epidemiologic studies have examined numerous risk factors for Crohn’s disease. Most studies have found breast-feeding to be protective for IBD, presumably by playing a role in early programming of immune responses in the developing gastrointestinal tract. Occupations associated with outdoor physical labor are relatively under-represented among Crohn’s patients. Crohn’s disease has been associated with higher socioeconomic status,66 presumably because of relative underexposure to diverse environmental antigens in the course of childhood—the hygiene hypothesis as it relates to intestinal mucosal immunity in IBD. Many, but not all, studies have discerned an increased risk of Crohn’s disease among women who use oral contraceptives. Nonsteroidal anti-inflammatory drugs (NSAIDs) have been implicated not only in exacerbations of IBD but also as a potential precipitant of new cases, perhaps by increasing intestinal permeability. Increased intake of refined sugars and a paucity of fresh fruits and vegetables in the diet have been associated with the development of Crohn’s disease. It is conceivable that this observation may be confounded by exacerbation of symptoms in patients with mild disease because of increased dietary fiber intake and subsequent avoidance of these food items before diagnosis. A virtually unexplored area is the connection between environmental and genetic factors in the expression of disease. It is presumed that the most important environmental factors may be those that determine the composition of the intestinal flora, given its necessary role in the onset of IBD. Such factors might include breast-feeding, composition of the diet, hygiene, use of antibiotics, and many others. Smoking is one of the more notable environmental factors for IBD. UC is largely a disease of ex-smokers and non­ smokers, whereas Crohn’s disease is more prevalent among smokers. In addition, smokers have more surgery for their disease and a greater risk of relapse after resection. The reasons for the divergent effect of smoking on Crohn’s disease and UC are poorly understood, but they might include effects on intestinal permeability, cytokine production, and clotting in the microvasculature. More recently, studies have focused on the role of carbon monoxide in stimulating immunosuppressive effects mediated by heme oxygenase-1.67 Many patients report a correlation between disease exacerbations and stress. Although depression and anxiety are a common reaction to illness, Crohn’s disease has not been shown to be caused by stress or by an anxious personality. The mind-body connection between emotional states or stress and intestinal inflammation in IBD is slowly being

revealed, however, and studies indicate that stress may be associated with risk of relapse in Crohn’s disease.68

ADAPTIVE IMMUNE RESPONSE AND INFLAMMATION

The interaction between effector T cells and APCs is critical to the pathogenesis of Crohn’s disease (Fig. 111-1). The antigens that perpetuate the inflammatory response are taken up by APC. Degradation of antigen within proteasomes results in presentation of an epitope in the context of major histocompatibility complex (MHC) class II. Interaction between MHC class II and the T-cell receptor (CD3) results in antigen-specific interaction between the macrophage and the CD4+ T cell. This event is necessary, but not sufficient, to activate the T cell. A second costimulatory signal is needed as well, because binding of CD3 to MHC class II without a costimulatory signal can result in anergy or apoptosis. Important costimulatory signals include binding of tumor necrosis factor (TNF) to TNF receptor, CD40 to CD40 ligand, and B7 to CD28. Activation of T cells leads to production of IL-2, an important growth factor for T cells. Inflammation normally is kept in check through an active process termed immune tolerance. The nature of the costimulatory signal, the type of APC, and the cytokine milieu influence the differentiation of T cells into populations of effector T cells, involved in harmful immune responses, and regulatory T cells, which ameliorate the immune response. Dendritic cells in the lamina propria actively sample the luminal contents and play a particularly vital role as key APCs capable of shaping the immune response. As noted earlier, the p40 subunit is common to IL-12 and IL-23, each of which, in turn, is critical in shaping the Th1 and Th17 responses that characterize Crohn’s disease. In addition to IL-23, the presence of TGF-β and IL-6 facilitate differentiation of naïve T cells into pathogenic Th17 cells.63 Activated APCs further shape and amplify the immune response by producing the T cell growth factor IL-2 and the proinflammatory cytokines IL-1 and TNF. Within mononuclear cells, the key nuclear transcription factor is NF-κB, which regulates the transcription of IL-1, IL-6, IL-8, TNF, and other peptides central to the inflammatory response.69 NF-κB is regulated tightly within the cell. In the inactive state, NF-κB is held in the cytoplasm, bound to inhibitory κBα. During cell activation after receptor binding, various kinases phosphorylate inhibitory κBα, thereby leading to its degradation. NF-κB is then released, permitting trans­ location to the nucleus, where it binds to the promoter regions of numerous genes that support the inflammatory response. Such genes include those that encode proinflammatory cytokines such as TNF, adhesion molecules, and chemokines.69 In addition to being essential to the formation of granulomas, TNF causes neutrophil activation and, along with IFN-γ, induces the expression of MHC class II on intestinal epithelial cells. Finally, TNF and other proinflammatory cytokines contribute to the expression of adhesion molecules on the endothelial cells of the intestinal vasculature. Expression of adhesion molecules is critical to amplify the immune response, because the resident populations of granulocytes and mononuclear cells alone do not account for the vigorous inflammatory reaction characterizing IBD. Adhesion molecules on the leukocyte surface and their ligands on the endothelium of venules in the lamina propria interact in a coordinated multistep process that permits trafficking of inflammatory cells into the mucosa. First, a weak interaction between selectins on the leukocyte surface

Chapter 111  Crohn’s Disease

Bacteria

IL-23

IL-23R Innate immunity

Th17

Defensins

Paneth cells

Barrier function

DC

PTGER4 Th1

IL-12 IL-23

DC Th0

Mϕ NOD2

Th2

TNF Adaptive

immunity

Leukocyte trafficking α4β7 MAdCAM

Treg

Autophagosome Bacteria Mϕ ATG16L1 IRGM

ICAM-1

α4β1

Endothelium Figure 111-1.  Pathogenesis of Crohn’s disease. Animal models of inflammatory bowel disease (IBD), studies of human genetics, and clinical trials with targeted therapeutic agents suggest that Crohn’s disease is a complex, polygenic disease driven by disturbances in distinct physiologic pathways; however, not all defects or pathways appear to be aberrant in all patients. Patients with allelic variants in NOD2 have defective sensing of intracellular bacteria, as well as reduced production of defensins, which are natural antimicrobial products produced by Paneth cells in the base of the intestinal crypts. The net result is excessive activation of adaptive immune responses to compensate for defective innate immunity. Similarly, variant loci of the ATG16L1 and IRGM genes are associated with defective autophagy, a process that is involved in defense against microbes and that stands at the interface of innate and adaptive immunity in the processing of intracellular pathogens and presentation of antigens to T cells. Adaptive immunity also may be deranged along the interleukin (IL)-12/IL-23 pathway, thereby shaping the expression of the helper T-cell response toward the spectrum of T-helper Th17. Together, Th17 and Th1 responses, most closely associated with Crohn’s disease, account for upregulation of effector T-cell responses in Crohn’s disease. Defects in regulatory T cells (Treg) of a few varieties also may be a factor in expression of disease. Dendritic cells (DCs) are active participants in maintaining immunologic tolerance within the intestine, continuously sampling luminal contents via podocytes extending through the epithelium. Activation of DCs and macrophages (Mϕs) result in the expression of tumor necrosis factor (TNF) within the mucosa. This pleiotropic cytokine has many downstream, proinflammatory effects that contribute to disease, and anti-TNF antibodies are effective in treating Crohn’s disease. Antigen presenting cells (APCs), including macrophages and DCs, also lead to activation of T cells when antigen is presented to the T cell in the context of major histocompatibility complex (MHC) class II, along with a costimulatory signal. Defects in the barrier function of the intestinal mucosa, such as through variant PTGER4, can lead to increased microbial and antigenic penetration of the mucosa, also leading to immune activation. Leukocyte trafficking is a necessary element in amplification of the mucosal immune response. Integrins are heterodimeric proteins that facilitate adhesion of leukocytes to the endothelium and recruitment into tissue. Integrins containing α4, such as α4β1 and α4β7, bind respectively to intercellular adhesion molecule 1 (ICAM-1) in the endothelium of inflamed tissues throughout the body and to mucosal addressin cellular adhesion molecule 1 (MAdCAM-1), which is specific to the intestinal endothelium. Blocking these interactions interferes with adhesion and recruitment of inflammatory cells, thereby disrupting inflammation.

and the endothelium leads to rolling of the leukocytes along the endothelium. Second, in the presence of chemokines such as IL-8, activation occurs, and integrins are expressed on the leukocyte surface. Third, interactions between leukocyte integrins and immunoglobulin-like cellular adhesion molecules on the endothelial surface lead to spreading of the cell and diapedesis.70 Specificity is conferred by the presence of tissue-specific cellular adhesion molecules. The integrins α4β7 and αEβ7 are of special importance in IBD, because the corresponding ligands—mucosal addressin cellular adhesion molecule and E-cadherin—are intestine specific. Mucosal addressin cellular adhesion molecule is expressed constitutively on the endothelium of venules in the lamina propria,70 whereas binding of αEβ7 on intestinal lymphocytes to E-cadherin on intestinal epithelium permits localization of intraepithelial lymphocytes. Antibodies to the α4 subunit of integrin have proved to be therapeutic in Crohn’s disease.71 Once recruited to the lamina propria, mononuclear cells and granulocytes elaborate a variety of injurious and pro­

inflammatory substances that ultimately cause tissue destruction. These substances include prostaglandins, reactive oxygen metabolites, nitric oxide, leukotrienes, and pro­ teases. Collagenase and matrix metalloproteinases play a pivotal role in the tissue destruction seen in IBD.72 Counterbalancing these destructive substances are other substances that promote epithelial restitution and repair, including IL-11, trefoil peptides, and growth factors such as epidermal growth factor and keratinocyte growth factor.

PATHOLOGY Focal intestinal inflammation is the hallmark pathologic finding in Crohn’s disease. This tendency for focal inflammation is evident in focal crypt inflammation, focal areas of marked chronic inflammation, the presence of aphthae and ulcers on a background of little or no chronic inflammation, and the interspersing of segments of involved bowel with

1947

1948

Section X  Small and Large Intestine segments of uninvolved bowel. Even within a single biopsy specimen one can see a pronounced variability in the degree of inflammation. The presence of focally enhanced gastritis, characterized by a focal perifoveolar or periglandular lymphomonocytic infiltrate, is a common finding that occurs in 43% of unselected patients with Crohn’s disease.73 This finding underscores the focal nature of the inflammation, despite the strong potential for inflammation to occur anywhere along the entire longitudinal axis of the gastrointestinal tract. To a certain extent, the nature of the findings and the depth of inflammatory changes depend on the chronicity of the inflammation.

EARLY FINDINGS

Because of the variable and often long delay between the onset of the disease process and its diagnosis, it rarely is possible to observe the evolution of pathology from the earliest events. Studies of recurrent Crohn’s disease after ileal resection have offered a window into the sequence of pathologic changes in the disease.74

Aphthous Ulcers

The earliest characteristic lesion of Crohn’s disease is the aphthous ulcer. These superficial ulcers are minute, ranging in size from barely visible to three millimeters, and are surrounded by a halo of erythema. In the small intestine, aphthous ulcers arise most often over lymphoid aggregates, with destruction of the overlying M cells. In the colon, aphthae can occur without an endoscopically visible central erosion and may be associated with lymphoepithelial complexes. Crohn’s aphthae typically occur in normal mucosa, although villus blunting may be seen in the surrounding small intestinal mucosa.75 Aphthous ulcers represent focal areas of immune activation. The M cells and underlying lymphoid aggregates are primary locations for antigen sampling and antigen presentation, so it is not surprising that human leukocyte antigen (HLA)-DR is strongly expressed on the follicle-associated epithelium of the aphthous ulcer.76 Contact with luminal contents is a key factor in the development of aphthous ulcers in Crohn’s disease. Aphthous ulcers heal in bowel excluded from the fecal stream by ileostomy, whereas re-establishing intestinal continuity leads to their recurrence77; these observations provide strong evidence for the role of luminal factors in the early pathogenesis of Crohn’s disease.

Granulomas

The presence of granulomas (Fig. 111-2), while highly characteristic of Crohn’s disease, is neither unique to Crohn’s disease nor universally found.78 Noncaseating granulomas, like aphthous lesions, are believed to be an early finding. Estimates of the prevalence of granulomas in Crohn’s disease have varied greatly, ranging from 15% in endoscopic series to as high as 70% in surgical series.79 Whether granulomas are found appears to be, in part, a matter of how hard one looks and how much tissue is available to examine; the more tissue sampled, the larger the specimen, and the more levels taken for histopathology, the more likely granulomas will be found. Granulomas may be discovered in involved and uninvolved bowel, in any layer of the intestine, and in mesenteric lymph nodes. Granulomas also may be found outside the intestinal tract—for example, in skin, eye, and liver— but extraintestinal granulomas are rare; occasionally, they may be recognized as millet seed-like nodules on the serosal surface of the bowel at laparotomy. The granulomas of Crohn’s disease are sarcoid-like, consisting of collections of

Figure 111-2.  Photomicrograph of a typical Crohn’s disease granuloma found in an endoscopic biopsy specimen. Note the loosely formed collection of cells, consisting of multinucleated giant cells (not always observed) and mononuclear cells, including T cells and epithelioid macrophages. Central caseation is not noted. (Courtesy of Dr. Gregory Lauwers, Boston, Mass.)

epithelioid histiocytes and a mixture of other inflammatory cells, including lymphocytes and eosinophils; giant cells occasionally are seen. The granulomas usually are sparse, scattered, and not well formed. In contrast to the granulomas of tuberculosis, there is little or no central necrosis, and acid-fast stains and mycobacterial cultures are negative. It also is important to distinguish the granulomas of Crohn’s disease from those that can occur in association with an injured crypt. The latter represent a response to mucin released from injured goblet cells and may be found in UC and other conditions.79 Regardless of whether granulomas are found, the granulomatous inflammation of Crohn’s disease represents a particular process involving characteristic cell types and regulation by specific cytokines and adhesion molecules. TNF is the key cytokine in the formation of granulomas. Appreciation of this fact led to the concept of anti-TNF therapies as a treatment for Crohn’s disease (see later).

LATER FINDINGS

Resected specimens of intestine may show localized foci of architectural distortion unaccompanied by chronic inflammation, an observation that suggests early superficial lesions such as aphthae may be transient and reversible. When the disease becomes chronic, however, aphthae can coalesce into larger ulcers with a stellate appearance. Linear or serpiginous ulcers can form when multiple ulcers fuse in a longitudinal direction. The classic cobblestoned appearance of Crohn’s disease results when linear and transverse ulcers intersect and networks of ulcers surround areas of relatively normal mucosa and prominent submucosal edema. Ulcers also can extend down to the muscularis propria. A prevailing generalization is that intestinal inflammation in Crohn’s disease is a transmural process, in contrast to the more superficial inflammation of UC. The transmural nature

Chapter 111  Crohn’s Disease of the inflammation, however, cannot be appreciated on superficial endoscopic biopsy, and in resected specimens it tends to be focal. Transmural involvement is observed less commonly than is disease of the mucosa and submucosa, but to the extent that transmural disease is noted, it is highly consistent with a diagnosis of Crohn’s disease. Dense lymphoid aggregates can enlarge the submucosa. At times, lymphoid aggregates also may be seen just outside the muscularis propria. The presence of lymphoid aggregates in the submucosa and external to the muscularis propria is a reliable sign of Crohn’s disease even when granulomas are not seen.79 Lymphoid aggregates occasionally may be seen within the muscularis propria, most often adjacent to the myenteric plexus. Large ulcers, sinus tracts, and strictures are late features of Crohn’s disease. Sinuses and fistulas represent extensions of fissures; sinus tracts end blindly, and fistulas enter epithelial-lined organs, such as bowel, skin, bladder, or vagina. Intramural sinus tracts are recognized easily on barium studies. With penetration of inflammation to the serosa, serositis can occur, resulting in adhesion of bowel to loops of small intestine, colon, or other adjacent organs. As a result of the chronicity of the inflammatory process and adhesions, free perforation is much less common than walled-off or contained intra-abdominal abscesses or fistula formation. Fissures and fistulas are lined by neutrophils and surrounded by histiocytes and a mononuclear cell infiltrate; partial epithelialization also is often observed, perhaps reflecting partial healing. Fibrosis is another transmural aspect of the disease. Fibrosis may be evident grossly as irregular thickening of the bowel wall and, along with hypertrophy of the muscularis mucosa, can contribute to the development of strictures. TGF-β is released locally in the presence of inflammation and is a cytokine critical for restitution and healing. In Crohn’s disease, however, TGF-β may be a double-edged sword. Fibroblasts isolated from the lamina propria produce primarily type III collagen in response to TGF-β1, and in the inflamed tissues of Crohn’s disease, significantly greater amounts of type III collagen are produced in response to this cytokine.80 Thus, a cytokine essential to the healing process also is implicated in fibrogenesis in Crohn’s disease.

OTHER FINDINGS

At the anatomic level, one of the most characteristic findings of Crohn’s disease is the presence of fat wrapping, a term that refers to the creeping of mesenteric fat onto the serosal surface of the bowel. Surgeons have long taken fat wrapping as a reliable indicator of the presence of diseased tissue. Mesenteric adipose tissue hypertrophy and creeping fat are recognized early in the course of disease at laparotomy or laparoscopy. Locally, fat wrapping correlates with the presence of underlying acute and chronic inflammation, as well as transmural inflammation in the form of lymphoid aggregates.81 Expression of peroxisome proliferatoractivated receptor γ (PPARγ), a pivotal mediator in the regulation of adipose tissue homeostasis, is increased greatly in Crohn’s tissues.82 In turn, adipocytes may participate in the inflammatory process of Crohn’s disease by producing TNF and other inflammatory mediators. At the microscopic level, the finding of pyloric metaplasia, normally a response to peptic ulcer disease when found in the duodenum, strongly suggests a diagnosis of Crohn’s disease when found in the terminal ileum. Careful descriptive immunopathology of areas of pyloric metaplasia reveals the presence of an ulcer-associated cell lineage. Bud-like glandular structures arise adjacent to areas of ulceration and are distinguished by production of epidermal growth factor

in acinar cells of the nascent gland and by trefoil proteins (see Chapter 1) in the more superficial cells lining the tract. Epidermal growth factor and trefoil proteins, in turn, can promote restitution of the epithelium in adjacent mucosal ulceration.

CLINICAL FEATURES DISEASE LOCATION

Crohn’s disease has a predilection for the distal small intestine and proximal colon. One third to one half of all patients have disease affecting both ileum and colon. Another one third have disease confined to the small intestine, primarily the terminal ileum, and there may be an increasing group with isolated colonic disease.83-85 Isolated jejunal involvement is rare. Gross involvement of the esophagus, stomach, or duodenum also is rare and almost always is seen in association with disease of the distal small intestine or colon. Focally enhanced acute and chronic inflammation may be seen in gastric biopsies in patients with Crohn’s disease either with or without gross involvement of the stomach.86 The discontinuous nature of the disease makes possible many variations in disease location, leading to considerable differences in clinical presentation. The disease usually stays confined to the segment in which it begins, but anatomic localization can vary over time, generally by involvement of additional segments of the alimentary tract, reflecting gross involvement with a disease that has the potential to affect any segment of the gastrointestinal tract.

CLINICAL PRESENTATION

The presentation of Crohn’s disease may be subtle and varies considerably. Factors contributing to this variability include the location of disease, the intensity of inflammation, and presence of specific intestinal and extraintestinal complications. Compared with UC, abdominal pain is a more frequent and persistent complaint. Pain is attributable to inflammation, abscess, or obstruction and may be intermittent and colicky or sustained and severe. Some patients experience symptoms that are mild but long-standing or that are atypical. Such patients are more likely to experience a delay in diagnosis in excess of a year. In the past, a mean delay in diagnosis of 3.3 years from the onset of symptoms was reported,87 but with improved diagnostic methods, and perhaps heightened awareness of the disease, more recent series have described delays of less than one year.84 A prodromal period is common in Crohn’s disease (not typically seen in UC) and might contribute to a delayed diagnosis,88 as does a prior diagnosis of irritable bowel syndrome and older age at onset of symptoms.89 Occasionally, radiologic and endoscopic findings are subtle, precluding definitive diagnosis even among patients with typical symptoms. Fecal occult blood may be found in approximately one half of patients, but in contrast to UC, gross rectal bleeding is uncommon, and acute hemorrhage is rare.90 Constitutional symptoms, particularly weight loss and fever, or growth retardation in children, may be pro­m­ inent and occasionally are the sole presenting features of Crohn’s disease.

Typical Presentations

Disease of the ileum, often accompanied by involvement of the cecum, can manifest insidiously. Some patients present with a small bowel obstruction, perhaps precipitated by impaction of indigestible foods, such as raw vegetables or

1949

1950

Section X  Small and Large Intestine fruit. Many years of subclinical inflammation can progress to fibrotic stenosis, with the subsequent onset of intermittent colicky pain, sometimes accompanied by nausea and vomiting. Physical examination can reveal fullness or a tender mass in the right hypogastrium during obstructive episodes. Patients with an active inflammatory component to their disease more often present with anorexia, loose or frequent stools, and weight loss; their examination might reveal fever or evidence of malnutrition. Occasionally, a patient presents with acute right lower quadrant pain, mimicking appendicitis. Colonic disease can involve mainly the right colon or can extend distally to involve most or all of the colon (extensive or total colitis). In patients with Crohn’s colitis, tenesmus is a less-common complaint than in patients with UC, because the rectum often is not involved or may be less severely inflamed than other colonic segments. The typical presenting symptom of colonic disease is diarrhea, occasionally with passage of obvious blood. The severity of the diarrhea tends to correlate with both the extent of colitis and the severity of inflammation, and the presentation may range from minimally altered bowel habits to fulminant colitis. Abdominal pain may be present to a greater extent than is seen in UC. Systemic manifestations such as weight loss and malaise also may be prominent. Although most patients with Crohn’s colitis have relative or complete sparing of the rectum, proctitis may be the initial presentation in some cases. Among a series of 96 patients with idiopathic proctitis, 13.6% were diagnosed with Crohn’s disease, usually within three years of initial presentation.91 Perianal disease is another common presentation of Crohn’s disease. In as many as 24% of patients with Crohn’s disease, perianal disease precedes intestinal manifestations, with a mean lead time of four years.92 More often, however, perianal disease occurs concomitantly with or after the onset of symptoms of luminal disease. Perianal findings may be categorized as skin lesions, anal canal lesions, and perianal fistulas.93 Skin lesions include maceration, superficial ulcers, abscesses, and skin tags. Skin tags are generally of two types: type 1 (elephant ears) are typically soft and painless and can be quite large; type 2, which often arise from healed fissures, ulcers or hemorrhoids, are typically edematous, hard, and tender.94 Anal canal lesions include fissures, ulcers, and stenosis. The anal fissures of Crohn’s disease tend to be placed more eccentrically than the usual idiopathic fissures, which generally occur in the midline. In most cases, anal stricture is asymptomatic, but occasionally obstruction occurs, particularly if stool consistency improves in the course of treatment. Deeper abscesses can arise secondary to fistulas, especially when the internal os is located high in the rectum.

Unusual Presentations

Upper gastrointestinal tract Crohn’s disease is uncommon in the absence of disease beyond the ligament of Treitz. Approximately one third of patients with proximal Crohn’s disease do not have evidence of distal Crohn’s disease at the time of diagnosis, but virtually all develop distal disease in time.95 Patients with proximal Crohn’s disease tend to be younger at the time of diagnosis and more often present with abdominal pain and malaise95; they do not undergo surgery more often than do patients with lower tract disease alone, but the length of bowel that is resected tends to be greater.95 Gastroduodenal Crohn’s disease manifests as Helicobacter pylori-negative peptic ulcer disease, with dyspepsia or epigastric pain as the primary symptoms. When outflow

obstruction occurs because of stricture formation or edema, early satiety, nausea, vomiting, and weight loss can predominate. Esophageal Crohn’s disease is rare, occurring in less than 2% of patients. Presenting symptoms can include dysphagia, odynophagia, substernal chest pain, and heartburn. These symptoms may be progressive and lead to profound weight loss.96 Aphthous ulcers sometimes are found in the mouth and posterior pharynx. Esophageal stricture and even esophagobronchial fistula can complicate the course. An intriguing observation is that HLA-DR expression often is seen in the esophageal epithelium of patients with Crohn’s disease, even when the disease is located more distally in the gastrointestinal tract, perhaps indicating widespread immunologic activation of the gastrointestinal mucosa.97 Crohn’s disease confined solely to the jejunum and ileum is unusual and may be impossible to differentiate from ulcerative jejunoileitis, a distinct condition that occasionally responds to a gluten-free diet (Chapter 115). Frank malabsorption and steatorrhea often occur. When the disease is confined to a short segment of intestine or has features consistent with Crohn’s disease, initial management should be based on the presumed diagnosis of Crohn’s disease. Controversy continues to surround the diagnosis of Crohn’s disease of the appendix. When idiopathic granulomatous inflammation is confined to the appendix, the presentation most often resembles that of acute appendicitis and occasionally periappendiceal abscess. The condition is rare, but the lack of disease in other locations of bowel portends a favorable prognosis, with a postoperative recurrence rate as low as 6%.98 Some authors suggest that granulomatous appendicitis should be considered an entity separate from Crohn’s disease.98

DISEASE BEHAVIOR

Clinical observation suggests that disease behavior in Crohn’s disease may be divided roughly into two categories: aggressive fistulizing disease and indolent cicatrizing disease99; a third subset of patients appear to develop neither of these behaviors over long periods of observation. Moreover, these distinctions are not always neat. Both fistula and stricture can occur simultaneously in the same patient, such as in the patient with a fistula arising behind a terminal ileal stricture, or at different times. Genetic factors are important in determining disease behavior, with NOD2 variants being associated with fibrostenotic disease.49 In addition, serologic antibody responses to microbrial antigens and carbohydrates are associated with certain disease phenotypes.100-105 Specifically, the pres­ ence of antiglycan antibodies to mannan (a constituent of the cell wall of baker’s yeast anti-Saccharomyces cerevisiae antibody [ASCA]) correlates with small intestinal disease; identification of anti-CBir1 (antiflagellin) is associated with internal penetrating and stricturing disease; and antiEscherichia coli outer membrane porin C (anti-OmpC) predicts internal perforations. When perinuclear antineutrophil cytoplasmic antibodies (pANCA) are present in a patient with Crohn’s disease, the phenotype is often that of an inflammatory “UC-like” Crohn’s disease.

Fistula and Abscess

Fistulas are frequent manifestations of the transmural nature of Crohn’s disease. Immune activation triggers the release of a variety of proteases and matrix metalloproteinases106 that can contribute directly to tissue destruction, sinus tract formation, and, finally, penetration to adjacent tissues. Perianal fistulas are common and are estimated to occur

Chapter 111  Crohn’s Disease

A

B

Figure 111-3.  Perianal fistulas in Crohn’s disease. A, Multiple complex fistulas in a man with Crohn’s disease. Several are active and draining. The scrotum, perianal skin, and buttocks are discolored and hardened by healed fistulas and abscesses. B, A simple fistula in a woman with Crohn’s disease. The purplish discoloration surrounding the fistula is from an abscess that drained spontaneously through the fistula. (Courtesy of Dr. Lawrence Brandt, Bronx, New York.)

in 15% to 35% of patients (Fig. 111-3). When the fistula arises from an anal gland, a low-lying perianal fistula is the most common result. Such fistulas often are minimally symptomatic and can resolve with local care alone. Sur­ prisingly, not all perianal fistulas occur in the setting of active rectal inflammation. In some cases, perianal fistulization may be extensive, forming a network of passages and extending to multiple openings that can include not only the perianal region but also the labia or scrotum, buttocks, or thighs. Fistulas from one segment of the gastrointestinal tract to another also occur commonly. Enteroenteric, enterocolonic, and colocolonic fistulas often are asymptomatic. Much more rarely, colonic disease penetrates normal duodenum or stomach to form a coloduodenal or cologastric fistula, respectively. Affected patients might have feculent vomiting. If the fistula tracks posteriorly from the terminal ileum to the retroperitoneum, the ensuing phlegmon can ensnare the ureter (usually the right ureter), causing noncalculous hydronephrosis; such patients often present with thigh pain or a limping gait. Deeper penetration yields the classic, but fortunately rare, circumstance of a psoas abscess. Affected patients typically present with right flank discomfort, fever, and a gait similar to those with ureteral entrapment. Fistula to the vagina can occur with penetration from a severely inflamed rectal vault anteriorly (rectovaginal fistula) or from the small intestine. Enterovaginal fistulas tend to occur among women who have had a hysterectomy, permitting direct extension to the adjacent vaginal cuff without the interfering presence of the uterus. Patients present with foul, persistent vaginal discharge and occasionally with passage of flatus or frank stool per vagina. Patients also might complain of dyspareunia or perineal pain. The vaginal os of the fistula may be difficult to identify, but palpation might elicit tenderness of the posterior vaginal wall. Fistulas arising from terminal ileal disease often occur in the setting of an ileal stricture, back pressure and stasis perhaps contributing to the process. Enterovesicular or colovesicular fistulas can manifest as recurrent polymicrobial urinary tract infection or as frank pneumaturia and fecaluria. These fistulas are notoriously difficult to

heal by nonsurgical means, although the resulting cystitis may be controlled with antibiotics. Enterocutaneous fistulas to the anterior abdomen, often occurring after surgery, may be especially troublesome. A classic presentation of Crohn’s disease is the onset of an enterocutaneous fistula after appendectomy for what had been presumed to be appendicitis. Often the tract of the fistula follows the planes of dissection to the abdominal surface. It has been estimated that as many as one fourth of all patients with Crohn’s disease present with an intraabdominal abscess at some time in their lives107; this figure is much less than one would imagine in light of the high incidence of fistulas. For the most part, inflamed serosal surfaces adhere to innocent serosa, thereby containing what would be an otherwise free perforation. Another common scenario is a perforation and abscess around the site of a surgical anastomosis. The classic presentation of an intraabdominal abscess is that of a patient with spiking fevers and focal abdominal tenderness or localized peritoneal signs. Unfortunately, many of the patients at highest risk for perforation or abscess also are taking glucocorticoids, which are notorious for suppressing peritoneal signs and fever and masking the presentations of infection; therefore, a high level of suspicion must be maintained. When free perforation and peritonitis occur, the situation is life-threatening.

Stricture

Stricture is another characteristic complication of Crohn’s disease. Strictures represent long-standing inflammation and can occur in any segment of the gastrointestinal tract in which inflammation has been active. Strictures do not develop in all patients with inflammatory disease, but are likely to recur, most often at the anastomosis, in patients who undergo bowel resection because of a stricture. These observations suggest that additional unidentified factors play a role in stricture formation. Strictures usually are silent until the luminal caliber is small enough to cause relative obstruction. Symptoms can include colicky, postprandial abdominal pain and bloating, punctuated by more-severe episodes, and often culminating in complete obstruction.

1951

1952

Section X  Small and Large Intestine

A

B

Figure 111-4.  Films from an upper gastrointestinal series and small bowel follow-through in a patient with Crohn’s disease. A, Multiple areas of narrowed small bowel are noted (arrows), with a classic cobblestoned appearance of the mucosa. Note also the separation of bowel loops. B, Small bowel followthrough that demonstrates a string sign in the right lower quadrant. The classic radiologic string sign (arrows) of a markedly narrowed bowel segment amidst widely spaced bowel loops is a result of spasm and edema associated with active inflammation rather than fibrostenosis; the typical string sign transiently resolves with administration of glucagon, which relieves smooth muscle spasm. (Courtesy of Dr. Jack Wittenberg, Boston, Mass.)

However, not all obstructive presentations are caused by fibrotic strictures. The classic radiologic string sign of a markedly narrowed bowel segment amid widely spaced bowel loops (Fig. 111-4) is a result of spasm and edema associated with active inflammation rather than fibro­ stenosis; the typical string sign transiently resolves with administration of glucagon, which relieves smooth muscle spasm. Short of demonstrating a clear response to antiinflammatory therapy or reviewing a surgical specimen, the clinician may find it extremely difficult to differentiate a fibrostenotic from an inflammatory stricture. All strictures must be considered with suspicion, and biopsies of a stricture need to be pursued vigorously, because some strictures harbor cancer.

CLASSIFICATION OF DISEASE

A major need in the clinical investigation of Crohn’s disease is the ability to define subgroups of patients with distinctive, if not unique, characteristics. The ability to define such subgroups could add tremendous power to the investigation of new therapies and to genetic studies. In light of the wide heterogeneity of demographic, anatomic, and disease behavior characteristics, however, distilling the numerous possible phenotypes into simple categories is a formidable task. The Vienna Classification of Crohn’s Disease is one proposed scheme that incorporates the patient’s age at onset, disease location, and disease behavior into a schema with 24 potential subgroups.108 It is not surprising that in this scheme, significant associations are noted between age at

diagnosis and location, and between disease behavior and location, along with a trend toward an association between age at diagnosis and disease behavior.109 Increasingly, subclinical characteristics such as serologic markers and genetic profiles will be used for their prognostic value in projecting outcomes in this heterogeneous disease. The Montreal classification is a first attempt at integrating serotype, genotype, and clinical phenotype as a structure to subclassify disease, and systems such as this will be used more as further supportive data emerge.110

PATHOPHYSIOLOGY OF COMMON SYMPTOMS AND SIGNS Diarrhea

Diarrhea is the most common complaint among patients with Crohn’s disease. Increased stool frequency and decreased stool consistency arise through alterations in mucosal function and intestinal motility. In any given patient, multiple factors are likely to contribute to diarrhea. Altered fluid and electrolyte absorption and secretion can decrease stool consistency. Increased mucosal permeability from mucosal inflammation can result in exudation of protein and fluids. Increased production of prostaglandins, biogenic amines, cytokines, neuropeptides, and reactive oxygen metabolites all contribute to these alterations. An imbalance in the luminal concentration of bile salts relative to dietary fat can result in either bile salt-induced diarrhea or steatorrhea in the setting of ileal dysfunction or resection (see Chapter 101). Bacterial overgrowth can occur behind strictured bowel and

Chapter 111  Crohn’s Disease contribute to malabsorption (Chapter 102). Disordered colonic motility is seen in the setting of chronic inflammation and also contributes to diarrhea. Occasionally, medications used to treat Crohn’s disease can exacerbate diarrhea. Secretory diarrhea can occur with olsalazine, and any of the 5-aminosalicylates rarely can induce a paradoxical increase in diarrhea, usually from salicylate sensitivity.

Abdominal Pain

The pathophysiology of abdominal pain in Crohn’s disease is not well understood. Numerous lines of investigation have provided tantalizing clues about the connection between the nervous system and Crohn’s disease, although the relationship among the enteric nervous system, inflammation, and immune activation in Crohn’s disease is quite complex. Stretch receptors in the intestinal wall may be stimulated as a food bolus passes through stenotic bowel, leading to abdominal pain and possibly vomiting. Visceral pain can result from serosal inflammation. The ganglia of the myenteric plexuses in the intestine in Crohn’s disease are increased in size and number, possibly indicating neural dysfunction.111 Substance P receptors have been found in increased numbers around the lymphoid follicles, in the microvasculature, and on enteric neurons in Crohn’s disease, even in locations distant from active inflammation,112 and there is increased binding of substance P to its receptors in the setting of an inflamed mucosa.113,114 Substance P binding can participate in the expression of pain. Enteroglia, support cells of the enteric nervous system, express MHC class II antigens in Crohn’s disease, raising the possibility that they also participate in the inflammatory process as APCs.115

Weight Loss and Malnutrition

Weight loss and malnutrition often are seen in patients with Crohn’s disease and contribute to the complaints of weakness, irritability, malaise, and easy fatigability that are so common. In children, malnutrition can manifest as growth retardation. A host of specific nutritional deficiencies may be found even among patients in long-standing remission,116,117 including iron, folic acid, vitamin B12, calcium, magnesium, zinc, and, particularly in the setting of malabsorption from small intestinal disease, fat-soluble vitamins. Potential contributing factors for these deficiencies are numerous and include inadequate intestinal absorption among patients with extensive small intestinal disease or resection and increased protein losses through exudation from inflamed intestine. Specific medications can cause absorption problems, including decreased calcium absorption with glucocorticoids; malabsorption of fat, fat-soluble vitamins, and calcium with cholestyramine; and folate malabsorption with sulfasalazine. The catabolic state induced by intense inflammation can increase energy and protein requirements. Unrecognized infection can be a major contributing factor beyond the catabolism induced by the disease itself. Bypassing of small intestine by enteroenteric or enterocolonic fistulas also can contribute to undernutrition. The most important factor in weight loss, however, is poor oral intake. Most often, poor intake results from fear of eating induced by postprandial abdominal pain or diarrhea. This concern may be accentuated when in social situations, and it can contribute to a lack of interest in being outside of the home. Decreased intake occasionally may be a consequence of unnecessarily restrictive diets imposed by the physician or the patient in an effort to control symptoms. Weight loss disproportionate to the burden of disease should raise the suspicion of occult malignancy.

Anorexia, nausea, and vomiting also can contribute to weight loss and poor nutrition. As with other symptoms of Crohn’s disease, diverse mechanisms may be contributory. TNF originally was discovered as a cytokine capable of inducing cachexia in patients with malignancy and sepsis. Indeed, serum levels of TNF in severely ill patients with Crohn’s disease may be high enough to contribute to anorexia. Delayed gastric emptying may be a causative factor for their symptoms in as many as one third of children with Crohn’s disease,118 and it reflects an unexpectedly high rate of gastroduodenal involvement. Anorexia, nausea, or vomiting also may be caused by drugs used to treat the disease, including metronidazole, sulfasalazine, 6mercaptopurine, azathioprine, and methotrexate.

Fever

Fever associated with active Crohn’s disease usually is low grade and occasionally is the presenting complaint, especially in children; increased production of proinflammatory cytokines, including IL-1, IL-6, and TNF, likely are contributory. When spiking, high, or persistent fevers occur, the clinician needs to consider an infectious etiology and undertake an evaluation appropriate to the clinical picture. Rarely, such fever patterns are manifestations of Crohn’s disease activity alone without superimposed illness or even abscess formation.

Anemia

Anemia is found in one third of patients with Crohn’s disease, primarily as a consequence of iron deficiency from blood loss. Macrocytic anemia can result from vitamin B12 deficiency because of ileal disease or resection, from bacterial overgrowth or, less commonly, from folate deficiency because of proximal small intestinal disease or sulfasalazine therapy. Overproduction of IFN-γ, TNF, or IL-1 can inhibit erythropoietin production, contributing to anemia resistant to iron supplementation.119

EXTRAINTESTINAL MANIFESTATIONS

In addition to penetrating and cicatrizing complications that can arise in patients with Crohn’s disease, numerous complications can occur distant from the bowel. Depending on the definition, it is estimated that between 6% and 25% of all patients with Crohn’s disease have an extraintestinal manifestation of IBD.120-122 Many of these complications are common to Crohn’s disease and UC and indeed to other nonidiopathic inflammatory conditions of the intestine. For example, patients with ileal Crohn’s disease are at increased risk for cholelithiasis, but patients with extensive UC are at nearly the same risk.123 In Crohn’s disease, however, the major risk factor for this complication appears to be the number of prior ileal resections.116 In large series, extrain­ testinal manifestations are found to occur more often in Crohn’s disease than in UC and are more common among patients with colonic involvement than in patients with no colonic inflammation. One fourth of those affected have more than one manifestation.83,124 Some complications occur as a direct result of the bowel disease, such as nephrolithiasis resulting from oxalate malabsorption. In the case of inflammatory mucocutaneous, joint, and ocular manifestations, the pathogenesis is an influx of mononuclear cells activated in the intestine but homing aberrantly to the involved extraintestinal organs.125

Musculoskeletal

Among the most common extraintestinal manifestations are disorders of the bones and joints. Clubbing of the fingernails is a common and innocuous finding. More consequential

1953

1954

Section X  Small and Large Intestine are arthritic manifestations, which are observed more commonly in patients with Crohn’s disease than in those with UC. In a study of 976 patients with UC and 483 patients with Crohn’s disease, pauciarticular arthropathy (type I, affecting four or fewer joints) occurred in 3.6% of patients with UC and in 6.0% of those with Crohn’s disease.126 In most patients, joint symptoms occurred in the setting of a relapse of intestinal symptoms. Polyarticular arthropathy (type II, with five or more joints affected) occurred in 2.5% of patients with UC and 4.0% of those with Crohn’s disease.126 Among patients with Crohn’s disease, nearly one half had joint symptoms in association with a relapse in intestinal disease. Distinct HLA genotypes are associated with these two types of peripheral arthropathy. Type I is associated with HLA-DRB1*0103, B*35, and B*27; type II is associated with HLA-B*44.127 Other reports indicate that peripheral arthralgias occur in 16% to 20% of patients with Crohn’s disease,128 most strongly in association with colonic disease.83 Patients tend to have waxing and waning joint pain and stiffness in association with flares of intestinal disease. Joints may be involved in an asymmetrical or migratory fashion. With rare exception, the disease is nondeforming and often is accompanied by skin (erythema nodosum) and eye (uveitis) complications. Rheumatoid factor typically is negative. Knee and ankle joints often are affected first, but elbows, wrists, proximal interphalangeal, metacarpophalangeal, and metatarsophalangeal joints may be involved subsequently.128 Patients who have undergone ileocecal resection for their disease tend to have fewer arthritic complications after their surgery.129 Axial arthropathies are less common than peripheral arthropathies and occur in 3% to 10% of patients with IBD.130 Spondylitis associated with IBD, like idiopathic ankylosing spondylitis, manifests as insidious low back pain and morning stiffness that is improved by exercise. As many as 75% of patients with Crohn’s disease and spondylitis are positive for HLA-B27. Iritis can occur in association with this manifestation. Bilateral symmetrical sacroiliitis without progression to spondylitis is more common than spondylitis and is reported to occur in 4% to 18% of patients.128 In one study, radiologic findings of sacroiliitis were detected in 29% of patients with Crohn’s disease, although only 3% had symptoms of sacroileitis.131 Rarer rheumatologic complications include granulomatous vasculitis,132 periostitis, and amyloidosis. In addition, a septic joint, although a rare complication of Crohn’s disease, should be kept in mind. A septic hip joint is a striking, devastating, and fortunately rare complication of a psoas abscess that extends directly to the acetabular capsule. Glucocorticoids used to treat Crohn’s disease may be a cause of joint pain. Withdrawal of glucocorticoids can lead to pseudoarthritis, with diffuse joint aches that gradually resolve; adrenal insufficiency should be considered in such patients. Aseptic necrosis of the hip and other joints can occur with or without the use of glucocorticoids and may be disabling.133 Osteomyelitis can occur as a result of direct extension by a fistula, usually to the pelvis, or it may be a recurrent problem distant to the site of inflammation, presumably through hematogenous spread of bacteria.134 Metabolic bone disease is common in Crohn’s disease; osteopenia (T score on dual energy x-ray absorptiometry between −2.49 and −1.0) or osteoporosis (T score no more than −2.5) occurs in 30% to 60% of patients. Morbidity, as a consequence of increased susceptibility to bone fractures, includes debilitating and painful vertebral crush fractures, which can occur even in children with Crohn’s disease.

Although glucocorticoid use is the main risk factor for this metabolic bone disease in UC, low bone mineral density is a feature of Crohn’s disease even at diagnosis in both adults and children.135,136 Contributing factors include malabsorption of calcium and vitamin D, smoking,137 and perhaps the effects of proinflammatory cytokines such as TNF, IL-1, and IL-6 on osteoclasts, some of which may be genetically determined.138 Low body mass index may be the most important risk factor for developing osteoporosis.139 Sarcopenia (decreased muscle mass) is closely associated with decreased bone density and is seen in up to 60% of patients with Crohn’s disease.140

Mucocutaneous

The most common skin lesions associated with IBD are pyoderma gangrenosum and erythema nodosum. Neither condition is found solely in IBD, and the finding of one or the other lesion is not specific for either major form of IBD.141 Pyoderma gangrenosum appears first as a papule, pustule, or nodule. It can occur virtually anywhere on the body but most often it occurs on the leg or occasionally around a stoma, and progresses to an ulcer with undermined borders. The ulcer typically has a violaceous rim and crater-like holes pitting the base. The phenomenon of pathergy, or the development of large ulcers in response to minor trauma, is characteristic of pyoderma gangrenosum and the skin lesions of Behçet’s syndrome.142 Healing classically is associated with a cribriform, or pocked, scar. In Crohn’s disease, pyoderma gangrenosum often occurs without an associated flare of intestinal symptoms. Erythema nodosum is seen much more commonly in women than in men. Like pyoderma gangrenosum, many other diseases are associated with erythema nodosum, including Streptococcus or Yersinia infection, tuberculosis, leprosy, fungal infections, Behçet’s syndrome, and sarcoidosis. The classic appearance is of tender subcutaneous nodules with an erythematous or dusky appearance, most often on the pretibial region. There is a strong association with arthropathy. Erythema nodosum often manifests during exacerbations of intestinal disease and tends to improve with treatment of the underlying bowel disease. When possible, erythema nodosum lesions should not be biopsied because biopsied lesions tend to scar; spontaneously resolving lesions heal without forming a scar. Aphthous ulcers of the mouth are common among patients with Crohn’s disease and UC but also are often seen among otherwise healthy persons.143 As the most cephalad part of the gastrointestinal tract, the mouth rarely may be involved directly by the granulomatous inflammation of Crohn’s disease. Angular cheilitis is seen in nearly 8% of patients with Crohn’s disease.143 A rare manifestation is metastatic Crohn’s disease, granulomatous inflammation of the skin remote from the gastrointestinal tract but histologically identical to the primary intestinal lesion.144 Described cases have included lesions behind the ears, in the perineum, or on the feet, legs, penis, and vulva. Other rare skin manifestations of Crohn’s disease include leukocytoclastic vasculitis,145 Sweet’s syndrome (neutrophilic dermatosis),146 cutaneous polyarteritis nodosa, and epidermolysis bullosa acquisita. Some reports have suggested an increased occurrence of psoriasis among patients with Crohn’s disease.147

Ocular

Ocular manifestations are estimated to occur in 6% of patients with Crohn’s disease.148 Episcleritis is more

Chapter 111  Crohn’s Disease common in Crohn’s disease than in UC, consists of injection of the sclera and conjunctiva, and does not affect visual acuity. Episodes tend to occur in association with active intestinal disease. Scleritis involves deeper layers of the eye and also occurs most often in parallel with active intestinal disease, but it can cause lasting damage if untreated. Uveitis usually manifests with headache, deep eye pain, lacrimation, blurred vision, and photophobia, as a consequence of iridospasm. Physical examination findings include meiosis and ciliary flush. Visual acuity is preserved unless the posterior segment becomes involved. In contrast to the uveitis associated with ankylosing spondylitis, the presentation of uveitis in patients with IBD often is insidious, with bilateral involvement and extension to the posterior segment.149 Slitlamp examination demonstrates an inflammatory flare in the anterior chamber. At least one report suggests that children with Crohn’s colitis often have asymptomatic anterior chamber inflammation.150 Other ocular complications of Crohn’s disease include a particular corneal injury referred to as keratopathy and night blindness resulting from malabsorption of vitamin A.

Hepatobiliary

Gallstones are found in over 25% of men and women with Crohn’s disease, representing a relative risk of 1.8 compared with the general population.151 Asymptomatic and mild elevations of liver biochemical tests often are seen in Crohn’s disease, but few of these patients develop clinical evidence of cirrhosis. Primary sclerosing cholangitis more often is associated with UC, but it occurs in 4% of patients with Crohn’s disease, usually those with colonic involvement.152 No genetic risk factors have yet been identified.153 In patients with Crohn’s disease, the inflammatory changes most often are confined to the small biliary radicals, and therefore the presentation is usually one of abnormal liver biochemical tests, pericholangitis on liver biopsy, and a normal cholangiogram.152 Other hepatobiliary complications of Crohn’s disease include fatty liver and autoimmune hepatitis.

Renal and Genitourinary

In addition to the direct complications of perforating Crohn’s disease with encroachment on the bladder and other genitourinary structures, and inflammatory entrapment of the ureter, uric acid and oxalate stones are common in patients with Crohn’s disease.154 In the setting of fat malabsorption resulting from intestinal resection or extensive small intestinal disease, the malabsorbed free fatty acids bind luminal calcium, thereby decreasing the calcium that is available to bind and clear oxalate. Increased oxalate is absorbed as the sodium salt, resulting in hyperoxaluria and calcium oxalate stone formation. Uric acid stones are believed to result from volume depletion and a hypermetabolic state. Rare intrinsic renal complications include membranous nephro­pathy, glomerulonephritis, and renal amyloidosis. Interstitial nephritis has been associated with mesalamine use, but it is not clear if it is a direct result of the medication or of the disease itself. Penile and vulvar edema also have been reported, but the mechanism for these occurrences is unknown.

Vascular

A prothrombotic tendency has been noted in both major forms of IBD. Patients might present with venous thromboembolism or, much less commonly, arterial thrombosis.155 The hypercoagulable state can arise from many possible causes. Contributing factors can include thrombocytosis, increased levels of fibrinogen, fibrinopeptide A, factor V,

and factor VIII, antithrombin III deficiency, and free protein S deficiency, all related to active bowel inflammation. Circulating immune complexes, increased levels of plasminogen activator inhibitors, decreased levels of tissue plasminogen activator, and spontaneous platelet aggregation may be present independent of bowel inflammation. Defective methylenetetrahydrofolate reductase (MTHFR), along with folate and vitamin B12 deficiency, is linked to hyperhomocysteinemia, which in turn predisposes to thrombosis. Increased prevalence of the factor V Leiden mutation156 and MTHFR157 have been observed by some but not other investigators.158 In more than half of patients who experience thrombosis, no predisposing factor can be identified.159

Other

Clinically significant disease of the lungs,160 heart, pancreas, and nervous system161 in association with Crohn’s disease is unusual but reported. Subclinical lung involvement may be much more common than is apparent, perhaps reflecting the commonality of bronchus-associated lymphoid tissue and gut-associated lymphoid tissue.160 Patients with IBD are more at risk to develop asthma, and there also may be an association with chronic obstructive pulmonary disease.147,162 Cardiomyopathy can result from a variety of nutrient deficiencies in patients with marked malabsorption. Pleuropericarditis, myocarditis, and endocarditis occur rarely.163 Acute pancreatitis,164 granulomatous pancreatitis,165 and pancreatic insufficiency166 also have been reported.

DIFFERENTIAL DIAGNOSIS Establishing a diagnosis of Crohn’s disease usually is straightforward once it is considered. Nevertheless, a large number of alternative diagnoses may be considered during various stages of the evaluation. Reports are legion of other diseases mistakenly diagnosed as Crohn’s disease and of Crohn’s disease mistaken for other diseases. Misdiagnoses may be attributed to the protean presentations of Crohn’s disease, which include considerable variability among patients with distinct anatomic distributions of disease, different degrees of inflammation, and the variable presence of intestinal complications and extraintestinal manifestations. There are a number of clinical situations in which Crohn’s disease should enter the differential diagnosis, including diarrhea or abdominal pain, especially when localized to the right lower quadrant; evidence of intestinal inflammation on radiologic or endoscopic studies; discovery of an intestinal stricture or fistula arising from the intestine; and evidence of inflammation or granulomas on intestinal histology. Categories of causation that overlap with Crohn’s disease in clinical presentation include functional bowel disorders, primarily irritable bowel syndrome; immunemediated diseases, particularly other colitides and most importantly UC; medications, especially NSAIDs; vascular disorders, notably ischemic bowel disease and collagen vascular diseases; neoplasia, including carcinoma and lymphoma; infectious diarrheas, intestinal inflammation, or granulomas; and miscellaneous other diseases and syndromes, including diverticular disease. Once the presence of bowel inflammation has been confirmed, the differential diagnosis may focus on presentation according to the anatomic location of the findings (Table 111-2).

1955

1956

Section X  Small and Large Intestine Table 111-2  Differential Diagnosis of Crohn’s Disease Differential Diagnosis of Ileitis Backwash ileitis in ulcerative colitis Drug-related Ischemia (oral contraceptives, ergotamine, amphetamines, phenylephrine, cocaine) NSAID-related ulcer or stricture Gynecologic disorders Ectopic pregnancy Endometriosis Ovarian cyst or tumor Ovarian torsion Pelvic inflammatory disease Tubo-ovarian abscess Ileitis associated with spondyloarthropathy Infection Actinomycosis israelii Anisakis simplex Cryptococcus neoformans Cytomegalovirus Histoplasma capsulatum Mycobacterium avium complex Mycobacterium tuberculosis Neutropenic enterocolitis Salmonella spp. Yersinia enterocolitica Yersinia pseudotuberculosis Infiltrative disorders Amyloidosis Eosinophilic gastroenteritis Other inflammatory disorders Appendicitis/appendiceal abscess Cecal diverticulitis Lymphoid nodular hyperplasia Neoplasms Carcinoid tumor Cecal or ileal adenocarcinoma Lymphoma Metastatic cancer Radiation enteritis Torsion of the appendiceal epiploica Vascular disorders Behçet’s disease Intestinal ischemia: focal segmental ischemia: acute enteritis, chronic enteritis, stricture; chronic mesenteric ischemia

Vasculitis: Henoch-Schönlein purpura, polyarteritis nodosa, Churg-Strauss syndrome, systemic lupus erythematosus, Takayasu’s arteritis, Wegener’s granulomatosis, lymphomatoid granulomatosis, giant cell arteritis, rheumatoid vasculitis, thromboangiitis obliterans Differential Diagnosis of Colitis Acute self-limited colitis Behçet’s disease Chronic granulomatous disease Diversion colitis Diverticulitis Drug-related intestinal inflammation (NSAIDs, gold, penicillamine) Eosinophilic gastroenteritis Graft-versus-host disease Indeterminate colitis Infections Aeromonas spp. Campylobacter spp. Clostridium difficile Cytomegalovirus Entamoeba histolytica Escherichia coli (enterohemorrhagic, enteroinvasive) Mycobacterium tuberculosis Salmonella spp. Schistosoma mansoni Shigella spp. Strongyloides stercoralis Yersinia enterocolitica Ischemic colitis Chronic ischemic colitis Ischemic stricture Ischemic colitis with toxic megacolon Transient ischemic colitis Microscopic colitis Collagenous colitis Lymphocytic colitis Radiation colitis Sarcoidosis Segmental colitis associated with diverticular disease (SCAD) Solitary rectal ulcer syndrome Ulcerative colitis

NSAID, nonsteroidal anti-inflammatory drug. From Sands BE. From symptom to diagnosis: Clinical distinctions among various forms of intestinal inflammation. Gastroenterology 2004; 126:1518, with permission.

ESTABLISHING THE DIAGNOSIS AND EVALUATING DISEASE ACTIVITY No single symptom, sign, or diagnostic test establishes the diagnosis of Crohn’s disease. Rather the diagnosis is established through a total assessment of the clinical presentation with confirmatory evidence from radiologic, endoscopic, and, in most cases, pathologic findings. The initial evaluation includes a thorough history-taking, physical examination, and simple laboratory tests. Historytaking focuses on the key symptoms and their severity and duration. Specific points to be covered should include recent travel history, use of antibiotics and other medications, diet, and sexual preference and activity. Family history of IBD can raise the level of suspicion but does not guarantee the diagnosis. The review of systems should focus on eliciting extraintestinal manifestations and weight loss. Fever may be associated with the underlying disease or a suppurative complication. A careful examination of the abdomen for signs of obstruction, tenderness, or a mass should be undertaken. Thorough inspection of the perineum

and a rectal examination might disclose findings highly suggestive of the underlying diagnosis or gross or occult blood. Laboratory data may be normal. Anemic patients should undergo further evaluation to define the contributions of iron, folate, or vitamin B12 deficiencies. The white blood cell count may be normal or elevated; an increased number of band forms suggests the possibility of a pyogenic complication. In the patient with vague symptoms suggesting irritable bowel syndrome, an elevated C-reactive protein or erythrocyte sedimentation rate, although not specific for IBD, can prompt further investigation. Stool studies should include culture, examination for ova and parasites, and testing for Clostridium difficile toxin and should be performed before endoscopy or barium studies. Serology for Entamoeba histolytica should be considered in selected patients. Ultimately, the diagnosis of Crohn’s disease is confirmed by findings on imaging studies, endoscopy, and usually histopathology. Barium studies had been the mainstay of Crohn’s imaging for many years, and they accurately define

Chapter 111  Crohn’s Disease the anatomic location of disease and can reveal evidence of active inflammation. Although a small bowel follow-through study is being replaced at many centers with computed tomography (CT) or magnetic resonance imaging (MRI) (see later), it is still often the primary modality when small intestinal disease is suspected (see Fig. 111-4). Barium studies are especially useful to delineate the late transmural complications of Crohn’s disease, but typical findings may be seen early in the disease as well. Early findings include aphthous ulcers, a coarse villus mucosal pattern, and thickened folds.167 Submucosal edema may be evident as thickening or flattening of the valvulae conniventes, whereas transmural edema manifests as widening of the separation between bowel loops. Ulcers most often occur on the mesenteric border, with consequent pseudosacculation of the antimesenteric border because of shortening of the mesenteric portion.167 Later findings include a cobblestone appearance resulting from edema and inflammation of relatively spared islands of mucosa separated by intersecting longitudinal and transverse knife-like clefts of ulceration.167 Still later, one can discern fistulas, sinus tracts, and fixed strictures. Standard CT studies do not demonstrate mucosal detail and often appear normal early in the course of the disease. The advent of CT enterography, however, has allowed fine mucosal changes to be evaluated along with extraluminal features (and complications of Crohn’s disease). CT enterography varies from routine CT by the use of a high-resolution multidetector scanner, intravenous contrast, and large volumes of oral contrast (either dilute barium or negative water-based contrast) to improve visualization of the small intestinal wall and reveal luminal details (Fig. 111-5). Radiologic findings that are significantly correlated with endoscopic evidence of Crohn’s activity include mural enhancement (segmental enhancement of all or part of the small intestinal wall); increased density of perienteric fat (focal increased inhomogeneous attenuation in the perienteric fat, compared with the appearance of subcutaneous or perienteric fat in adjacent noninflamed intestinal loops), and the comb sign (segmental dilatation of the vasa recta involving an intestinal loop).168 Mural enhancement may be the most useful finding and can be quantified in a semi­ automated fashion using dedicated software.168,169 When compared to a consensus diagnosis of Crohn’s disease based on clinical presentation and four different imaging modalities, the sensitivity of CT enterography was 82%, specificity was 89%, and accuracy was 85%.170 The safety of radiation

Figure 111-5.  Computed tomography enterography in a patient with Crohn’s disease showing an intestinal stricture with prestenotic dilatation. The stricture is partly inflammatory, with hyperenhancement, mural thickening, and perienteric inflammation. (Courtesy of Dr. Edward Loftus and Dr. Joel Fletcher, Rochester, Minn.)

exposure associated with the routine use of CT is a matter of much debate, but safety needs to be taken into careful consideration if this technology is to replace other diagnostic modalities, especially in children.171,172 As an alternative to CT, MRI has made substantial strides in evaluating the intestine. MR enterography has the advantages of providing high soft tissue contrast, obtaining static and dynamic images, and avoiding ionizing radiation.173 Similar to CT enterography, patients drink an oral contrast agent before the procedure. Some European centers incorporate enteroclysis with nasoduodenal intubation to administer the contrast, which might increase the yield for subtle mucosal lesions but is likely to be less acceptable to most patients.173 Findings of intestinal wall thickening, submucosal edema, vasa recta engorgement, and lymphadenopathy are signs of active disease (Fig. 111-6). Using dynamic FIESTA (fast imaging employing steady state acquisition), images can add information regarding the functional status of fibrotic segments. A scoring system was developed for assessing small intestinal Crohn’s disease and gives higher scores for details such as increased wall thickness and contrast enhancement, stenosis and mucosal abnormalities, absence of peristalsis and distensibility, and extraintestinal findings.174 Using these criteria, compared with the gold standard of ileocolonoscopy with biopsies, the MRI images yielded a diagnostic accuracy of 91%. Other potentially useful diagnostic modalities include ultrasound and scintigraphy. Transabdominal ultrasound is used mainly to exclude other causes of abdominal pain, including biliary and gynecologic causes, but it can also be effective in evaluating disease activity of luminal Crohn’s disease.175 There has been some interest in using endoscopic ultrasound to differentiate Crohn’s disease (transmural) from UC, but its major value is still to help evaluate and guide therapy of perianal disease.176,177 Doppler vascular

Figure 111-6.  Magnetic resonance enterography with gadolinium contrast in a patient with Crohn’s disease. This coronal view shows mural hyperenhancement, mural thickening, and the comb sign (engorged perienteric vasculature) involving the terminal ileum. The vessels are seen to the left of the inflamed loop and resemble the teeth of a comb. (Courtesy of Dr. Edward Loftus and Dr. Jeffrey Fidler, Rochester, Minn.)

1957

1958

Section X  Small and Large Intestine flow studies to evaluate Crohn’s disease activity have been investigated with mixed results.178,179 Ultrasound- and CTguided percutaneous drainage of intra-abdominal abscesses is a safe and effective alternative to surgical drainage in well-selected patients.180 A growing body of evidence suggests that leukocyte scintigraphy may be a useful diagnostic study in Crohn’s disease. Among children with suspected IBD, 99mTc leukocyte scintigraphy was highly sensitive in identifying abnormalities in patients with just mild inflammation on biopsy and a normal small bowel followthrough.181 Compared with the gold standard of intraoperative findings, it had an accuracy of 84%.182 Early data suggest a possible role for positron emission tomography (PET)/CT scan to help in evaluating the activity level and distribution of Crohn’s disease.183,184 Because of its ability to visualize the mucosa directly and permit biopsy for histopathology, endoscopy complements radiologic techniques. Typical mucosal features recognized on endoscopy include aphthous ulcers, mucosal edema, cobblestoning, and luminal narrowing (Fig. 111-7). The visual impression of demarcated lesions on a background of normal mucosa is most easily recognized in early or mild disease. Rectal sparing is more specific before treatment has been initiated. The discontinuous segmental nature of the disease is an important clue to the diagnosis and has a high positive predictive value.185 Intubation and biopsy of the terminal ileum should be attempted in all patients having colonoscopy and greatly increase the sensitivity and specificity of the examination.186 In general, the diagnostic accuracy of colonoscopy and histologic

interpretation is increased substantially by obtaining multiple biopsies from both involved and uninvolved sites. The use of jumbo forceps should be considered to improve sampling of the submucosa. Balloon dilation of strictures is another application of endoscopy in Crohn’s disease that might delay surgery or eliminate the need for it, but balloon dilation is associated with a measurable complication rate. Wireless capsule endoscopy has become routine for detecting the small intestinal lesions of Crohn’s disease. Although wireless capsule endoscopy may be very sensitive in identifying lesions (even if standard endoscopy has been unrevealing),187,188 its low specificity limits its use as a firstline study to diagnose small intestinal Crohn’s disease.170 The presence of significant bowel stricture should be excluded radiologically before attempting capsule endoscopy, because the rate of capsule retention may be as high as 25% and obstruction is possible.189 Nonetheless, capsule retention can be clinically useful to localize occult strictures pre- and intra-operatively.190

DIFFERENTIATING CROHN’S DISEASE FROM ULCERATIVE COLITIS

When IBD is confined to the colon, the main diagnostic distinction is between Crohn’s colitis and UC. As noted earlier, UC and Crohn’s disease share many similarities in epidemiology and clinical manifestations, and the distinction between them is becoming increasingly important with regard to choices of surgical and medical therapies. Patients

A

B

C

D

Figure 111-7.  Endoscopic appearance of Crohn’s disease. A wide variety of findings may be visualized on endoscopy, in part depending on the duration and severity of the inflammation. A, Typical aphthous ulcers (arrows), consisting of a central white depression surrounded by a slightly elevated, erythematous rim only a few millimeters in diameter. B, Findings more typical of advanced disease, with erythema, edema, and a cobblestone appearance. C, Stellate ulcers (arrows) in the terminal ileum. D, Discrete ulcers (arrows) with normal intervening mucosa, typical of the patchy inflammation seen in Crohn’s disease.

Chapter 111  Crohn’s Disease with features of both diseases are said to have indeterminate colitis (also referred to as IBD, unclassified), a vague term applied in various ways among different centers. As many as 10% of patients presenting with IBD are considered to have indeterminate colitis. A diagnosis of indeterminate colitis has particular implications for surgical therapy. Patients undergoing ileoanal pouch construction for indeterminate colitis have a relatively high likelihood of developing Crohn’s-like complications of the pouch, although the rate of pouch failure is not significantly different from those with UC.191,192 Histology, when applied without attention to clinical features, is highly likely to be unable to differentiate Crohn’s disease from UC.193,194 Therefore, the entire clinical picture must be considered for accurate diagnosis (Table 111-3). Discriminating features for Crohn’s disease include the presence of small intestinal disease, predominantly right-sided colonic disease, rectal sparing, fistulization (with the exception of rare rectovaginal or perianal fistulas in UC), major perineal complications, and granulomas. In cases initially labeled as indeterminate, the true diagnosis usually becomes clear with the passage of time. With an incomplete understanding of the environmental and genetic determinants that produce a clinical phenotype of Crohn’s disease or UC, the immunologic markers noted previously are being explored as a means of differentiating the two diseases. pANCA and ASCA were the first such markers shown to correlate with the diagnosis of UC and Crohn’s disease, respectively.103,195,196 To improve the test characteristics, the prediction model has added anti-OmpC99 and anti-CBir1 to pANCA and ASCA.104 When combined, and applying a diagnostic algorithm in a high-prevalence (59%) group, the sensitivity for diagnosing Crohn’s disease is 88% and for diagnosing UC is 93%, with a specificity greater than 95% for both diseases. The positive predictive value (PPV) and negative predictive value (NPV) for Crohn’s

disease is 96% and 93%, respectively. The PPV and NPV for UC is 89% and 98%, respectively. The pretest probability is high in this group, and when applying this algorithm to a lower-prevalence group (15%), which is more appropriate for the pretest probability associated with indeterminate colitis, the PPV and NPV for Crohn’s disease drop to 74% and 73%, respectively.197 Thus, serologic testing for pANCA, ASCA, OmpC, and anti-CBir1 currently is an adjunct to diagnosis in selected cases—one additional piece of evidence to be considered but not definitive in establishing the diagnosis. Genetic research has led to great strides over the past few years, and CARD15/NOD2 testing is available commercially; due to its low diagnostic accuracy, however, this test is not currently recommended as part of the diagnostic algorithm for Crohn’s disease.

MEASURING DISEASE ACTIVITY

In daily practice, usually it is sufficient to follow the patient’s symptoms and signs with treatment. Rarely is it necessary to subject the patient to repeated radiologic studies or colonoscopies to ascertain disease activity; disease location tends to be stable over time. Repeat studies are undertaken when symptoms have increased substantially or have changed and are suspected to arise not from persistent intestinal inflammation but from other causes, such as infection, complication, or functional disorder. In clinical research, however, more quantitative evaluations are needed. Composite scoring systems, most commonly the Crohn’s Disease Activity Index (CDAI, Table 111-4), are used in an attempt to integrate the many possible features of the disease. Other disease activity indices include the van Hees index,198 the Cape Town index,199 the HarveyBradshaw index,200 the International Organization of IBD (or Oxford) index,201 the St. Marks Crohn’s index,202 De

Table 111-3  Differentiating Crohn’s Colitis from Ulcerative Colitis feature

CROHN’S COLITIS

ULCERATIVE COLITIS

Mucosal lesions

Aphthous ulcers are common in early disease; late disease is notable for stellate, rake, bear-claw, linear, or serpiginous ulcers and cobblestoning Often discontinuous and asymmetrical, with skipped segments of normal intervening mucosa, especially in early disease

Micro-ulcers are more common, but larger ulcers are possible Pseudopolyps are more common Continuous, symmetrical, and diffuse, with granularity or ulceration found in entire involved segments; however, periappendiceal inflammation (cecal patch) is common, even when the cecum is not involved Typically involved with variable proximal distribution Not involved, except as backwash ileitis in panulcerative colitis Mucosal; transmural only in fulminant disease Absent except in severe colitis or toxic megacolon

Distribution

Rectum Ileum

Complete, or more often relative, rectal sparing Often involved (approximately 75% of cases)

Depth of inflammation Serosal findings

Mucosal, submucosal, and transmural Marked erythema and creeping fat (the latter is virtually pathognomonic) Often prominent, including large anal skin tags, deep fissures, perianal fistulas, that are often complex Often present Perianal, enterocutaneous, rectovaginal, enterovesicular, and other fistulas may be present Granulomas are present in 15%-60% of patients (higher frequency in surgical specimens than in mucosal pinch biopsies) Crypt abscesses may be present Focally enhanced inflammation, often on a normal background, is the hallmark pANCA in 20%-25% ASCA in 41%-76%

Perianal complications Strictures Fistulas Histopathology

Serology

Not prominent (fissure or fistula if present, should be uncomplicated) Rarely present; suggests adenocarcinoma Absent, except for rare occurrence of rectovaginal or perianal fistula Granulomas should not be present (microgranulomas may be associated with ruptured crypt abscess) Crypt abscesses and ulcers are the defining lesion Ulceration on a background of inflamed mucosa pANCA in 60%-65% ASCA in 5%

ASCA, anti-Saccharomyces cerevisiae antibody; pANCA, perinuclear antineutrophil cytoplasmic antibody. From Sands BE. From symptom to diagnosis: clinical distinctions among various forms of intestinal inflammation. Gastroenterology 2004; 126:1518, with permission.

1959

1960

Section X  Small and Large Intestine Table 111-4  Crohn’s Disease Activity Index (CDAI)* VARIABLE

SCALE

Liquid or very soft stools Abdominal pain

Stool count summed daily for 7 days Sum of 7 days of daily ratings as: 0 = none, 1 = mild, 2 = moderate, 3 = severe Sum of 7 days of daily ratings as: 0 = generally well, 1 = slightly below par, 2 = poor, 3 = very poor, 4 = terrible Any of the following present during the 7 days: a. Arthritis or arthralgia b. Skin or mouth lesions, including pyoderma gangrenosum, erythema nodosum, aphthous stomatitis c. Iritis or uveitis d. Anal fissure, fistula, or perirectal abscess e. Other external fistula f. Fever > 100°F 0 = no, 1 = yes 0 = none, 2 = questionable, 5 = definite Males: 47 – hematocrit Females: 42 – hematocrit 100 × [1 – (body weight/standard weight)]

General well-being Features of extraintestinal disease

Opiates for diarrhea Abdominal mass Hematocrit value: % Body weight below standard

WEIGHT 2 5 7 20 each

30 10 6 1

*To calculate the CDAI, the scale is multiplied by the weighting factor for each variable, and then all eight weighted variables are added. From Best WR, Becktel JM, Singleton JW, et al. Development of a Crohn’s disease activity index. National Cooperative Crohn’s Disease Study. Gastroenterology 1976; 70:439, with permission.

Dombal’s index,203 the Talstad index,204 and a Crohn’s disease activity index for survey research.205 Specialized indices have been developed for use in children with Crohn’s disease.206,207 All these indices vary in the features included in the scoring system, but most include a combination of subjective symptoms and objective findings on examination and laboratory testing. A great deal of interobserver and methodologic variation has been noted, even among experienced researchers,208 and in fibrostenotic disease, indices that rely more heavily on subjective measurements can poorly reflect bowel inflammation as a cause of symptoms.209 Other approaches have included use of disease-activity indices that focus on a specific outcome, such as perianal disease,210,211 endoscopic findings,212,213 or achieving an individual goal of therapy.214 Each of these approaches has advantages and disadvantages, but all have their application in research rather than in clinical practice. Another approach with some merit is the measurement of biologic markers of disease inflammation. The erythrocyte sedimentation rate and serum acute phase response proteins (e.g., C-reactive protein and orosomucoid) may be useful in tracking disease activity, but they lack sensitivity and specificity. Another approach is to measure the serum levels of cytokines, cytokine receptors, and adhesion molecules that are proximate to the expression of acute phase reactants; examples include IL-6, IL-1, soluble IL-2 receptor, and soluble intercellular adhesion molecule-1.215 As with the acute phase reactants, these tests also lack sensitivity and specificity. Direct measurements of intestinal immune activation in a mucosal sample could enhance sensitivity and specificity but are inconvenient, invasive, and, if dependent on biopsy, subject to variability and poor standardization. Quantification of radiolabeled leukocytes in stool appears to be a sensitive and fairly specific indicator of mucosal inflammation, but it is cumbersome to perform and exposes the patient to radiation. Fecal excretion of calprotectin (a calcium- and zinc-binding protein found in neutrophils) and lactoferrin (an iron-binding glycoprotein secreted by most mucosal membranes), has been shown to be sensitive markers of intestinal inflammation216 that also

might correlate with relapse of quiescent disease217 and response to biologic therapy.218 Ultimately, it is desirable to measure the patient’s overall state of well being, or subjective health status. Healthrelated quality of life may be measured with generic instruments, which focus on various domains of health common to many disease states, or with disease-specific instruments, which focus on specific domains relevant to the disease of interest. The Inflammatory Bowel Disease Questionnaire is the most widely accepted disease-specific instrument and measures separate domains for bowel, social, systemic, and emotional function.219

TREATMENT GOALS

Because neither medical nor surgical therapy provides a cure for Crohn’s disease, the primary goals of therapy are to induce and maintain remission. In achieving these goals, the intention is to ameliorate symptoms and improve the patient’s quality of life. Therefore, it is essential to consider the adverse consequences of therapy, particularly with regard to any durable consequences of short-term treatment and adverse effects of maintenance therapy. Other goals may be specific to the individual patient, such as healing a fistula or achieving normal growth in a child. Maintaining adequate nutrition can at times be a challenge and is an important goal in all patients.

MEDICAL THERAPY Aminosalicylates

In the United States, aminosalicylates (ASAs) are used in the treatment of UC and in mild to moderate Crohn’s disease. Sulfasalazine, the parent compound of all ASAs used in IBD, was developed by the Swedish physician Nana Svartz in 1938-1939 as a treatment for rheumatoid arthritis. In 1941-1942 sulfasalazine was serendipitously found to improve the intestinal symptoms of patients with colitis

Chapter 111  Crohn’s Disease who were being treated for associated arthropathy. Later, a classic experiment by Azad Khan and colleagues implicated 5-aminosalicylate (5-ASA, mesalamine) rather than sulfapyridine as the therapeutic moiety of sulfasalazine.220 Most of the adverse effects of sulfasalazine are due to its sulfapyridine moiety, which serves as a carrier for the 5-ASA portion of the molecule; release of 5-ASA from sulfapyridine depends on azoreductase activity of the colonic flora, thereby ensuring its delivery to the colon. Because almost the entire dose of a bare 5-ASA compound is absorbed in the upper gastrointestinal tract when given orally, and the beneficial effects of 5-ASA depends on its topical delivery to affected mucosa, diverse means of overcoming the proximal intestinal absorption of 5-ASA were developed. These delivery systems include enemas or suppositories, which provide the drug to the rectum and left colon; coating with protective materials that release the drug in a pHdependent manner to achieve controlled (Pentasa) or delayed (Asacol and Lialda) delivery; and diazo-bonding the drug to a second 5-ASA molecule (olsalazine) or to an inert carrier (balsalazide). The site of delivery of coated preparations depends largely on the properties of the coating material used and its pH sensitivity. Some preparations (Pentasa) release half of the dose in the small intestine and the remainder in the colon, and other preparations release mesalamine in the distal terminal ileum and beyond. Diazobonded preparations have release profiles closely resembling that of sulfasalazine. Numerous therapeutic mechanisms have been attributed to ASAs, including inhibition of T-cell proliferation, presentation of antigen to T cells, and antibody production by B cells; inhibition of macrophage and neutrophil adhesion; and decreased production of IL-1 and TNF. The ASAs are excellent free-radical scavengers and inhibit cyclooxygenase and production of prostaglandin E2.221 Many of these effects appear to be mediated through down-regulation of NF-κB activity.222 Most studies have shown sulfasalazine to be superior to placebo in inducing remission in active Crohn’s disease, when the colon is the primary site affected.223 Efficacious doses, as used in the National Cooperative Crohn’s Disease Study (NCCDS), are in the range of 4 to 6 g/day (1 g/15 kg body weight).223 The European Cooperative Crohn’s Disease Study (ECCDS) found sulfasalazine 3 g/day to provide no significant benefit in achieving remission.224 Early studies with controlled-release mesalamine (Pentasa) at doses less than 2 g/day failed to show efficacy in the treatment of mild to moderately active Crohn’s disease.225,226 A much larger study of 466 patients with mild to moderate Crohn’s disease compared daily doses of 1, 2, and 4 g with placebo for 16 weeks. The 43% remission rate on 4 g mesalamine was statistically and clinically superior to the placebo response rate of 18%.227 Notably, patients responding best to the 4 g/day dose were those with ileumonly disease, suggesting that mesalamine provides a potential benefit over sulfasalazine in treating this subgroup of patients. Subsequent trials of similar design, however, failed to show benefit over placebo; although the treatment effect was of similar magnitude, the placebo response was larger than the originally observed 18%. A meta-analysis failed to demonstrate a clinically significant benefit of Pentasa 4 g/day in patients with mild to moderate Crohn’s disease.228 Numerous studies with a variety of preparations have failed to demonstrate prevention of relapses of Crohn’s disease with 5-ASA compounds.229 Therefore, although maintenance therapy with mesalamine often is prescribed in Crohn’s disease, little data justify the expense and inconvenience of this practice, and mesalamine-based products

have been excluded from recent evidence-based treatment algorithms.230 In summary, sulfasalazine 4 to 6 g/day may be useful for inducing remission of mild to moderate colonic Crohn’s disease, whereas the role of mesalamine is uncertain. The small margin of benefit and relatively slow onset of effect (four to eight weeks) must be weighed against the excellent safety profile of these agents (Table 111-5).

Antibiotics

Antibiotics have a clear role in treating pyogenic complications of Crohn’s disease. On the basis of relatively little evidence, antibiotics also are used to treat perineal disease, fistulas, and active luminal Crohn’s disease. The largest reported experience has been with metronidazole. The anaerobic flora affected by metronidazole might have particular importance in the pathogenesis of Crohn’s disease.231 Perhaps the clearest demonstration of this principle is a study of postsurgical prophylaxis after ileal resection. In this disease model, which in some ways might replicate the earliest events in the initiation of Crohn’s disease, high-dose metronidazole (20 mg/kg/day for three months) demonstrated a prophylactic effect on endoscopic and clinical recurrence at one year, with numerical but not statistical advantages at two and three years of follow-up.232 In this study, as in clinical use, side effects, including gastrointestinal upset, nausea, dysgeusia, and peripheral neuropathy, were common. A study using ornidazole, also a nitroimidazole antibiotic, showed similar results.233 Ornidizole might have fewer side effects than metronidazole, but toxicity was still a problem with this formulation. Ongoing studies with other antibiotics that have improved long-term tolerance are ongoing. Open-label experience suggests that metronidazole 20 mg/kg/day is beneficial in healing perineal fistulas.234 Fistulas tend to recur with cessation of therapy, but longterm use is limited by side effects. Studies of metronidazole in active Crohn’s disease generally have not demonstrated benefit, but they have suggested better outcomes in subgroups of patients with colonic involvement.235,236 Ciprofloxacin is increasingly used to treat Crohn’s disease. For perianal fistulas, ciprofloxacin may be similar in efficacy to metronidazole and with fewer side effects,237 and its addition might improve the response to infliximab.238 In the treatment of luminal disease, one study found ciprofloxacin 1 g/day to be equivalent to mesalamine 4 g/day in achieving remission of mild to moderately active Crohn’s disease at week 6, with more than half of the patients in each group achieving remission.239 Another study compared the combined use of ciprofloxacin and metronidazole, 1 g each, against methylprednisolone for active Crohn’s disease. The antibiotic combination was comparable with glucocorticoids in achieving remission over 12 weeks.240 A morerecent study failed to detect additional efficacy of the same dual antibiotic regimen over placebo when added to controlled ileal-release budesonide; however, a trend toward benefit was noted in the subgroup of patients with colonic disease.241 Preliminary evidence suggested that clarithromycin monotherapy was useful in treating active disease,242 but a follow-up randomized trial did not confirm the open label experience.243 Additional interest in clarithromycin is sparked by its role as part of a highly effective treatment regimen for atypical mycobacterial infection. As reviewed earlier, M. paratuberculosis has been the most extensively studied infectious agent thought to be important in the etiology of Crohn’s disease. Studies of antimycobacterial therapy,

1961

1962

Section X  Small and Large Intestine Table 111-5  Safety Profiles of Agents Used to Treat Crohn’s Disease AGENT 5-Aminosalicylates (5-ASA) Sulfasalazine

Sulfa-free 5-ASAs (mesalamine, olsalazine, balsalazide) Antibiotics Metronidazole Ciprofloxacin

Glucocorticoids Classic

Novel

Immune Modulators 6-Mercaptopurine, azathioprine

Methotrexate Cyclosporine

ADVERSE EFFECTS

PREGNANCY*

NURSING*

Anorexia, dyspepsia, nausea and vomiting; hemolysis, neutropenia, agranulocytosis; folate deficiency; reversible male infertility; neuropathy; see also sulfa-free 5-ASAs Headache; drug fever; rash; paradoxical disease exacerbation; pancreatitis; hepatitis; pericarditis; pneumonitis; nephritis; secretory diarrhea (olsalazine)

No evidence of teratogenicity; normal fetal growth; give with folic acid

Negligible amounts; safe for term neonates

No evidence of teratogenicity, normal fetal growth

Found in breast milk in low concentrations; rare watery diarrhea in breast-fed infants

Anorexia, nausea and vomiting, dysgeusia; disulfiram-like effect; peripheral neuropathy; reversible neutropenia Nausea and vomiting; headache; restlessness; rash; pseudomembranous colitis; elevated serum aminotransferase levels; spontaneous tendon rupture

Questionable teratogenicity, normal fetal growth

Found in breast milk; with rare exception, should not be used Found in breast milk, should not be used

Sleep and mood disturbance; acne, striae, hirsutism; adrenal suppression; proximal myopathy; glucose intolerance; hypertension; narrow-angle glaucoma, cataracts, pseudotumor cerebri; infection; edema; impaired wound healing; growth retardation; bone loss, aseptic necrosis Controlled ileal-release budesonide: Adrenal suppression at doses of 9 mg/day and higher in two divided doses, but occurrence of classic glucocorticoid adverse effects similar to placebo

No evidence of teratogenicity in humans, more frequent stillbirths and reduced fetal birth weight when used for other diseases; may be used as indicated by severity of disease

Safe for breast-feeding

No data available

No data available

Nausea; drug fever, rash, arthralgias; leukopenia, thrombocytopenia, bone marrow suppression; pancreatitis; hepatitis; infection; lymphoma?

Teratogenic in animals, but large series in renal transplantation and other diseases do not show increase in birth defects; evidence for fetal growth retardation and prematurity; isolated cases of neonatal immune and bone marrow suppression; outcomes appear favorable in limited series of patients with IBD; may be used when indicated because of disease severity Highly teratogenic, particularly in the first trimester; abortifacient

Small amounts excreted in breast milk; not recommended

Anorexia, nausea and vomiting; leukopenia, megaloblastic anemia; alopecia; hepatic fibrosis; interstitial pneumonitis; neuropathy Reversible or irreversible decrease in renal function; hypertension; tremor, headache, paresthesias, seizure; hypertrichosis; hepatotoxicity; infection; lymphoma; gingival hyperplasia

Biological Response Modifiers Anti-TNF antibodies Upper respiratory tract and other infections; (infliximab, adalimumab, disseminated tuberculosis; increased risk of certolizumab pegol) systemic fungal infection and other intracellular pathogens; acute or delayed hypersensitivity reactions; antinuclear antibodies, anti–double-stranded DNA antibodies, lupus-like reaction; demyelinating disease; contraindicated in heart failure because of increased mortality; lymphoma Natalizumab Headache, flushing, infections; progressive multifocal leukoencephalopathy; jaundice, liver failure

Theoretical teratogenic potential; insufficient data in humans

Significant levels in fetal circulation; does not appear to be teratogenic; intrauterine growth retardation and premature delivery increased, especially at higher doses; little reported experience in IBD

Small amounts excreted in breast milk; not recommended Excreted in breast milk; not recommended

Limited data in humans

Unknown safety in nursing. Early data suggest minimal infliximab levels in breast milk

Teratogenic in animals

Unknown safety in nursing

IBD, inflammatory bowel disease; TNF, tumor necrosis factor. *From Connell WR. Safety of drug therapy for inflammatory bowel disease in pregnant and nursing women. Inflam Bowel Dis 1996; 2:33, with permission. Adapted from Sands BE. Therapy of inflammatory bowel disease. Gastroenterology 2000; 118(2 Suppl 1):S72, with permission.

Chapter 111  Crohn’s Disease however, have not shown consistent benefit. A large recent study of triple antibiotic therapy (clarithromycin, rifabutin, and clofazimine) combined with prednisolone showed an early clinical benefit, but when followed over two more years, this benefit was not sustained.244 The authors concluded that the treatment regimen might have contributed to nonspecific antibacterial effects and improvement during the course of therapy, but the findings did not support a significant role for M. paratuberculosis in the pathogenesis of Crohn’s disease. Rifaximin is a nonabsorbable oral rifamycin antibiotic that is approved for the treatment of traveler’s diarrhea, and it has shown success in the treatment of irritable bowel syndrome. A recent randomized, controlled trial for treating mildly to moderately active Crohn’s disease showed a numerical advantage for rifaximin 800 mg twice daily compared with once-daily dosing or placebo, but this difference was not statistically significant.245 Treatment failures in the placebo group were significantly higher, and those with an elevated C-reactive protein did have a significantly better response. Therefore, this medication shows promise as a future choice in treating Crohn’s disease. In summary, antibiotics can play an adjunctive role in the treatment of Crohn’s disease and, in selected patients, they may be useful in treating perineal disease, enterocutaneous fistulas, or active colonic disease. As the antigenic determinants of the intestinal flora are elucidated further, more directed antibiotic approaches might be feasible.

Glucocorticoids

Glucocorticoids play a central yet vexing role in the treatment of Crohn’s disease. Early favorable series of glucocorticoid treatment led to the validation of their short-term efficacy in the NCCDS (prednisone 0.5 to 0.75 mg/kg/day for initial treatment of active disease, with the dose adjusted according to CDAI)223 and the ECCDS (6-methylprednisolone 48 mg/day in the first week, tapered to 12 mg by week 6, and held at 8 mg for remission up to two years).224 In usual practice, patients with mild to moderate disease that does not respond to primary therapy and patients with moderately severe symptoms are treated initially with 40 to 60 mg of prednisone, the dose then being tapered off over 6 to 12 weeks. Response rates are approximately 80% by one month.246 When doses are pushed as high as 1 mg/kg/ day for up to seven weeks, 92% of patients can achieve clinical remission.247 The onset of response is rapid, usually within the first three weeks of treatment. Patients with severely active disease usually respond to intravenous administration of glucocorticoids.248 Options for glucocorticoid formulation include hydrocortisone (100 mg intravenously every 8 hours), prednisolone (30 mg intravenously twice daily) or methylprednisolone (16 to 20 mg intravenously every 8 hours). Hydrocortisone can cause slightly more salt retention and sodium wasting, but it is likely to be equally effective. In a randomized double-blind trial with ulcerative colitis patients, continuous infusion was no better than divided dosing for efficacy and safety.249 Numerous anti-inflammatory and immunosuppressive effects have been attributed to glucocorticoids to account for their efficacy, including inhibition of the expression of proinflammatory cytokines, adhesion molecules, MHC class II molecules, leukotrienes, elastase, collagenase, and nitric oxide synthase. Glucocorticoids bind to a cytoplasmic receptor found in all cells and then enter the nucleus to bind glucocorticoid-response elements on the chromosomal DNA, thereby producing a variety of downstream physiologic effects.250 The anti-inflammatory effects of glucocorticoids might follow from down-regulation of NF-κB and

induction of inhibitory κB.251 Direct cellular effects also can occur, with reduced phagocytic activity of neutrophils and, in some situations, apoptosis of lymphocytes.250 Unfortunately, the beneficial effects of glucocorticoids come at the expense of frequent and often severe adverse effects (see Table 111-5). The most common side effects are troubling neuropsychiatric symptoms, including mood disturbance and insomnia, and cosmetic effects, including acne, cushingoid appearance, hair loss, and hirsutism. Still more serious are metabolic consequences, including adrenocortical suppression, glucose intolerance, myopathy, and bone loss. The risk of infectious complications is increased, particularly at doses of prednisone higher than 40 mg; doses lower than 10 mg confer no appreciable increased risk of infection.252 Among patients taking immunomodulators or infliximab, the concomitant use of prednisone appears to lead to more frequent serious infections and higher rates of mortality than when these agents are used alone.253,254 The unfavorable risk profile of glucocorticoids makes their prolonged use hazardous. Glucocorticoids are not effective as long-term therapy. A meta-analysis of maintenance glucocorticoid therapy in Crohn’s disease failed to detect benefit in the prevention of relapse at 6, 12, or 24 months.255 Conversely, once glucocorticoids are introduced, they cannot be discontinued without recurrent symptoms in many patients, even with gradual tapering; this problem is referred to as glucocorticoid dependence. Among patients with Crohn’s disease who received glucocorticoids for the first time, no response (glucocorticoid resistance) was seen in 20% in the first 30 days.246 Among the 80% who were complete or partial responders, 55% had a prolonged response, and 45% relapsed or could not have treatment tapered off within one year.246 Similar results were seen in a cohort from Olmsted County, Minnesota.256 Clinical factors associated with glucocorticoid dependence include smoking, younger age at onset, colonic location, and non-fibrostenotic disease.257 Mechanisms that can contribute to glucocorticoid resistance include up-regulation of the multidrug resistance (mdr) gene258 and increased serum levels of glucocorticoid-binding glob­ ulin.259 Moreover, only 29% of patients who achieve clinical remission on glucocorticoids also achieve endoscopic remission.247 This finding suggests that the effect of glucocorticoid treatment in most patients is to suppress symptoms when given in doses above a threshold that can vary among patients and even in the same patient over time. There are several principles of glucocorticoid use in Crohn’s disease: Use an effective dose. Underdosing at the start of therapy typically leads to dose escalation and prolonged dosing to achieve a response. Do not overdose. Patients who do not benefit from 40 to 60 mg are unlikely to benefit from increased or prolonged oral dosing. Such patients require intravenous dosing or treatment with another rapidly acting agent, such as an anti-TNF agent (see later). Do not treat for excessively short periods. Doses should not be tapered too quickly once symptoms have been controlled. Very brief courses of glucocorticoids (three weeks or less) are likely to result in a rebound flare. Do not treat for excessively long periods. Patients in whom a glucocorticoid taper fails should be considered candidates for glucocorticoid-sparing immune modulators. Glucocorticoids should not be begun without a strategy in mind for terminating treatment. Anticipate side effects. Bone loss in particular may be anticipated with even short-term use. (See later, “Adjunctive Therapies.”)

1963

1964

Section X  Small and Large Intestine In an attempt to limit the unintended systemic effects of glucocorticoid therapy, novel glucocorticoids have been developed. Budesonide possesses glucocorticoid receptor affinity superior to that of traditional glucocorticoids and also takes advantage of enhanced first-pass metabolism by the liver to limit systemic exposure. A controlled ilealrelease formulation of budesonide targets the terminal ileum and right colon. Studies have demonstrated that 9 mg/day of this preparation are superior to placebo and mesalamine and about 15% less effective than prednisolone in achieving remission, but with fewer side effects.260 Pushing the dose higher results in better efficacy but at the expense of increasing adrenocortical suppression and side effects.261 Budesonide treatment for maintaining remission was reported in a pooled analysis of four studies. Results showed that although a lower dose of budesonide (6 mg daily) was more effective than placebo in maintaining remission at three and six months, this was not sustained at 12 months.262 A subsequent randomized, controlled trial compared 6 mg to 9 mg for maintenance at 12 months.263 Both doses were associated with relatively low relapse rates (24% and 19%, respectively) that were not significantly different, although a placebo comparison group was not included. Adverse events were not different in these two dosage groups, supporting the safety of the 9-mg dose over a one-year period. Similarly, budesonide 3 or 6 mg/day after ileal or ileocecal resection did not improve the recurrence rate at one year.264-265 Therefore, lack of a maintenance effect is consistent for both novel and traditional glucocorticoids. In light of the superior response in comparison to mesalamine and its relative safety, budesonide may be considered as first-line therapy for patients with active ileal, ileocecal, or right colonic disease. In addition, some patients who are dependent upon conventional glucorticoids may be switched successfully to budesonide, with the potential benefits of decreased systemic glucocorticoid exposure.266 In summary, glucocorticoids are effective for the shortterm control of symptoms of Crohn’s disease, but they are neither effective nor safe for long-term maintenance of response. In patients with disease that is refractory to or dependent on glucocorticoids, steroid-sparing strategies should be considered, including immune modulators or surgery.

Thiopurine Agents

The thiopurine antimetabolites azathioprine and 6-mercaptopurine (6-MP) have been used to treat Crohn’s disease since the initial report of Brooke and colleagues describing healing of fistulas with azathioprine.267 Another decade would pass, however, before the efficacy of this class of drugs was demonstrated in a randomized, controlled trial by Present and colleagues214; earlier studies were marred by either insufficient power or incomplete understanding of adequate dosing and the slow onset of action of these agents. A Cochrane meta-analysis of studies of azathioprine and 6-MP in Crohn’s disease has provided the best summary of the effects of these drugs.268 For active disease, treatment produced a response rate of 54% compared with a 33% placebo rate, and corresponding to an odds ratio for response of 2.36 (95% CI, 1.57 to 3.53); the number needed to treat (NNT) was 5. The odds ratio for response increases after 17 weeks of therapy, suggesting the minimum duration for a trial of 6-MP or azathioprine. A steroid sparing-effect also was significant (OR, 3.86; 95% CI, 2.14 to 6.96), and the NNT was 3. A limited number of patients with fistulas was included, but a 55% rate of fistula healing was noted, com-

DNA RNA

6-thioguanine nucleotides 6-thiouric acid IMPDH

XO AZA

6-MP TPMT 6-methylmercaptopurine

HPRT

6-thioinosine 5´-monophosphate TPMT

6-methyl-mercaptopurine ribonucleotides

Purine synthesis Figure 111-8.  Metabolism of azathioprine (AZA) and 6-mercaptopurine (6-MP). AZA is converted to 6-MP nonenzymatically. HPRT, hypoxanthine phosphoribosyltransferase; IMPDH, inosine monophosphate dehydrogenase; TPMT, thiopurine methyltransferase; XO, xanthine oxidase. (From Dubinsky MC, Lamothe S, Yang HY, et al. Pharmacogenomics and metabolite measurement for 6-mercaptopurine therapy in inflammatory bowel disease. Gastroenterology 2000; 118:705-13, with permission.)

pared with a 29% healing rate on placebo; this was not statistically significant (OR, 4.58; 95% CI, 0.49 to 52.82), most likely the result of a type II error. Convincing evidence of the benefit of thiopurines also was seen in an analysis of the maintenance of remission.269 The remission rate was 67% compared with a 52% rate on placebo (OR, 2.16; 95% CI, 1.46 to 3.47); the NNT was 7.269 The OR for azathioprine maintaining remission increased from 1.20 at 1 mg/kg up to 4.13 (95% CI, 1.59 to 10.71) at 2.5 mg/kg, demonstrating the importance of appropriate dosing. Overall, approximately one half to two thirds of patients may respond to thiopurine therapy. In contrast to glucocorticoids, mucosal healing often is seen with adequate dosing of these agents.270 In clinical practice, azathioprine and 6-MP are used virtually interchangeably, with the exception of dosing. Azathioprine generally is used in doses of 2.0 to 2.5 mg/kg/day, and 6-MP is given in doses of 1.0 to 1.5 mg/kg/day. Much is known about the metabolism of 6-MP and azathioprine (Fig. 111-8). Azathioprine is a prodrug that is converted in part to 6-MP through nonenzymatic means and into a variety of other immunologically active and inert metabolites. Xanthine oxidase converts 6-MP to 6-thiouric acid, in competition with hypoxanthine phosphoribosyltransferase. The former pathway accounts for an important drug reaction with allopurinol, a xanthine oxidase inhibitor (see later). Thiopurine methyltransferase (TPMT) plays a key role in the metabolic pathway. Persons who are homozygous for a recessive mutation that results in inactivation of TPMT (approximately one in 300 persons) produce exceedingly high levels of 6-thioguanine (6-TG) nucleotides. These persons are unlikely to tolerate thiopurine agents and tend to develop profound leukopenia and other

Chapter 111  Crohn’s Disease limiting adverse effects. In contrast, persons who are TPMT heterozygous (approximately 11% of the population) are likely to have moderately high levels of 6-TG nucleotides.271 They usually require lower doses of drug but are much more likely to respond. A steady state in the production of erythrocyte 6-TG nucleotides is reached two weeks after dosing.272 There have been mixed results reported in studying the correlation between 6-TG nucleotide levels and response to therapy, but a meta-analysis of six studies found an overall significant relationship.273 A threshold of 230 to 260 pmol/ 8 × 108 red blood cells corresponded to a 62% rate of remission, compared with a rate of 36% in those with lower levels (OR, 3.27; 95% CI, 1.71 to 6.27) of 6-TG. Correlations between higher levels of 6-TG nucleotides and leukopenia, and between metabolite levels and response to therapy, might explain the clinical observation that patients who achieve mild leukopenia are more likely to respond to such therapy.274 Conversely, however, leukopenia is not necessary to achieve a therapeutic response. It is not clear if routine measurement of thiopurine metabolite levels and directed dose adjustment would contribute to improved management of Crohn’s disease, as opposed to the standard weight-based dose approach. When a patient is not responding to thiopurine therapy after three to four months, it is useful to measure metabolite levels to identify patients who are noncompliant, underdosed, or shunting. Shunting refers to high TPMT activity resulting in low 6-TG levels and high 6-MMP levels (see Fig. 111-8); a 6-MMP:6-TG ratio of greater than 10 has been suggested as a profile of metabolism that is unlikely to lead to clinical benefit.275 In these patients, it may be possible to add allopurinol and take advantage of the drug interaction noted earlier. A study testing this hypothesis showed that by decreasing the thiopurine dose to 25% to 50% of the original dose and adding a low dose of allopurinol, 6-TG levels rose significantly, with a coincident drop in 6-MMP levels and an improvement in clinical outcomes.276 This strategy needs to be tested further to prove both safety and efficacy, but it may be an alternative treatment approach for patients in whom thiopurine therapy fails. The relevant mechanisms of action of azathioprine and 6-MP are not clearly understood. Metabolites of both agents inhibit de novo synthesis of purine ribonucleotides and thereby inhibit cell proliferation. Azathioprine and its metabolites might have immunosuppressive properties beyond those of the metabolites produced in common with 6-MP. Both drugs inhibit cell-mediated immunity. The thiopurines cause a reduction in the number of circulating natural killer cells over many months, perhaps accounting for the slow onset of action of these agents.277 Azathioprine and 6-MP have been shown to activate Rac1 when costimulation of CD4+ T lymphocytes occurs via CD28, effecting apoptosis of the cell and diminishing the immune response.278 A study of intravenous loading of azathioprine found no acceleration of the time-to-response over oral dosing.272 Unexpectedly, the oral dosing arm of the study with azathioprine 2 mg/kg/day achieved a maximum rate of remission (discontinuation of prednisone and CDAI less than 150) of 24% by week 8. The remission rate did not increase over an additional eight weeks of follow-up, thus contradicting the long-held notion of a prolonged time to response. The study did not, however, address improvement beyond the 16th week of treatment. In the Cochrane analysis described earlier, adverse events severe enough to result in drug withdrawal were seen in 9.3% of patients, corresponding to a number needed to

harm (NNH) of 14.268 Nausea often is reported in the first weeks of treatment but gradually subsides. Allergic reactions consisting of fever, rash, or arthralgias are seen in 2% of patients, usually within a few weeks of introducing the drug. Pancreatitis, observed in 3% to 7%, is another idiosyncratic reaction and usually occurs in the first month of therapy. The presentation may be subtle, with nausea and vague dyspepsia, or it may be classic, with epigastric pain that radiates to the back. If serum amylase and lipase levels are followed closely, symptoms sometimes may be noted to precede the discovery of laboratory abnormalities. When symptoms are recognized promptly, discontinuation of the drug leads to resolution of pancreatitis. Rechallenge with either drug should not be attempted because recurrent pancreatitis is certain to occur. Elevated serum aminotransferase levels develop in as many as 9% of patients and have been correlated with the presence of very high levels of 6-methylmercaptopurine.279 Cholestatic hepatitis is rare, occurring in less than 1% of patients.280 Bone marrow suppression is another concern with thiopurine agents. A 27-year, retrospective, single-center study of 739 IBD patients treated with azathioprine found 28 patients (3.8%) developed leukopenia (white blood cell count less than 3 × 109 cells/L [less than 3000 cells/mm3], 9 of whom (1.2%) had severe leukopenia (white blood cell count less than 2 × 109 cells/L [less than 2000 cells/mm3])281; three of these patients became pancytopenic, and two died of sepsis. In another retrospective report of 396 patients treated with 6-MP, 2% experienced leukocyte counts below 3.5 × 109/L (less than 3500 cells/mm3).282 Although leukopenia occurs early among patients with low TPMT activity, it might not be related solely to TPMT genotype and can occur at any time during therapy.282 For this reason, it is advisable to continue monitoring the complete blood count every one to three months for the duration of therapy and more frequently (every two weeks if TPMT activity is normal, weekly if it is heterozygous) in the weeks after introducing the drug or increasing dosage. Temporary cessation of therapy for a week or two and an adjustment in dose usually suffice to bring the leukocyte count back within normal range. Careful monitoring of the leukocyte count also should be performed during a tapering regimen of glucocorticoids. Concurrent treatment with glucocorticoids can raise the leukocyte count, but as the glucocorticoid is discontinued, leukopenia can develop. Infections can occur in the setting of thiopurine therapy. With long-term treatment, as many as 1.8% of patients experience a severe infection, not necessarily in the setting of leukopenia.280 Patients treated concurrently with glucocorticoids may be at greater risk of serious infection, including cytomegalovirus. Treatment should be interrupted when serious infections occur, although the effect of the drug will endure for weeks. The question of whether patients with IBD who are treated with 6-MP or azathioprine are at increased risk for malignancy is unresolved. Immunosuppressive regimens given to patients after organ transplantation and for other immunemediated conditions are associated with an excess risk of malignancy, particularly non-Hodgkin’s lymphoma (NHL). Such regimens have included azathioprine, often administered in high doses and in conjunction with other immunosuppressive agents. The combined long-term experience at two centers involving treatment of 1151 patients with IBD failed to reveal an excess risk of cancer, but it did include a single case of diffuse histiocytic lymphoma of the brain.280,283 A meta-analysis of IBD patients treated with 6-MP and azathioprine found a relative risk of 4.18 (95% CI, 2.07 to 7.51) compared with IBD patients not treated

1965

1966

Section X  Small and Large Intestine with these agents.281 A possible confounding factor is an increased baseline rate of NHL in patients with Crohn’s disease,285 but data are conflicting on this issue.286 Nevertheless, if there is an increased risk of lymphoma with immunosuppressive therapy, studies suggest that the absolute risk is small.287 For the properly selected patient, the small risk of lymphoma does not outweigh the benefits of improved quality of life and possibly of increased life expectancy.288 Azathioprine and 6-MP should be considered for patients with active Crohn’s disease in whom first-line therapies fail or who cannot successfully taper off glucocorticoids. Patients who are treated with antibiotics for a fistula and who cannot tolerate or do not respond to these agents also may be considered for treatment with a thiopurine agent. The introduction of thiopurine medications should be timed with their slow onset of action in mind; many patients require a tapering regimen of glucocorticoids to bridge the time period until the thiopurines have taken effect. Thiopurine therapy also may be considered for the postsurgical prophylaxis of Crohn’s disease. Although conflicting data exist about the efficacy of this approach,289,290 the benefits in preventing recurrence might outweigh the risks of these agents in patients who are deemed to have a high likelihood of future resections. Given the favorable efficacy of thiopurine agents and their relative safety, there is increasing interest in introducing these agents earlier in the course of the disease. Evidence suggests that the addition of 6-MP is advantageous for children who require even a first course of glucocorticoids shortly after diagnosis.291 Furthermore, recent pediatric data have shown that initiating thiopurines within three months of diagnosis decreases exposure to glucocorticoids and rate of hospitalizations.292 It is unclear whether thiopurine therapy prevents or delays the occurrence of fibrostenotic complications of Crohn’s disease; it is possible to see the slow progression of a stricture in patients who otherwise have been asymptomatic on such therapy for many years. Once treatment with a thiopurine agent has proved to be effective, the question of how long to continue such therapy inevitably arises. A randomized, controlled trial demonstrated a clinical relapse rate of 21% 18 months after withdrawal of azathioprine in patients who had been in remission for at least 3.5 years on the drug,293 compared with a relapse rate of only 8% in the group who continued azathioprine. The authors concluded, and most authorities agree, that azathioprine maintenance therapy should be continued longer than 3.5 years. The decision to withdraw thiopurine therapy should only be undertaken after discussion between doctor and patient of possible risks and benefits.

Methotrexate

Methotrexate has long been used to treat psoriasis and rheumatoid arthritis. A randomized, controlled trial studied patients with chronically active Crohn’s disease despite at least three months of prednisone (at least 12.5 mg/day) and with at least one failed attempt to taper off treatment.294 All patients were brought to a 20 mg/day dose of prednisone to standardize therapy, with separate stratification for patients in whom the dose of prednisone was increased and for those in whom the dose had dropped to 20 mg before entry. Subjects then received either weekly injections of methotrexate 25 mg intramuscularly or placebo while executing a tapering prednisone regimen over 16 weeks. Overall, 39.4% of patients assigned to methotrexate achieved remission off prednisone compared with 19.1% of placebotreated patients.294 Most patients responded by the eighth week of treatment. Although the remission rates in the

methotrexate-treated high- and low-prednisone group were nearly equal (39.0% and 40.0%, respectively), the remission rate for placebo-treated patients in the high-prednisone dose group was 10.0%, compared with 35.3% in the low-dose group.294 This result often is misconstrued as showing that methotrexate works well for patients on high doses of prednisone but not for those on low doses of prednisone, but it merely shows an unexpectedly high placebo response rate among patients dependent on low doses of glucocorticoids. Methotrexate also is beneficial in maintaining remission. A follow-up study randomized patients who achieved remission by week 16 on methotrexate 25 mg intramuscularly once weekly to receive either weekly injections of placebo or methotrexate at a dose of 15 mg. At week 40, 65% of patients treated with methotrexate were still in remission, compared with 39% of placebo-treated patients (P = 0.04).295 Treatment was well tolerated. Among patients who relapsed on the maintenance dose, more than half were able to achieve remission again with resuming a 25-mg dose. If the 16 weeks of induction therapy were included, the combined duration of therapy was nearly one year, with some patients in selected practices treated successfully for more than four years. Although 15-mg intramuscular dosing was studied for maintenance, many continue on 25 mg weekly without dose reduction. Pharmacokinetic studies in rheumatoid arthritis have shown equivalency for subcutaneous and intramuscular dosing, and therefore most gastroenterologists administer methotrexate subcutaneously.296,297 Although oral dosing would be more convenient for longterm administration, this route is unreliable because of variable intestinal absorption,298 particularly in the presence of small intestinal disease. Although methotrexate is a folate antagonist, the drug often is given with folic acid (1 to 2 mg/day) to prevent nausea and stomatitis, and so other modes of action are likely responsible for its efficacy. The drug possesses a variety of immune-modulating and anti-inflammatory effects, including inhibition of IL-1, IL-2, IL-6, and IL-8 and induction of adenosine, which has direct immunosuppressive properties.299 In addition to stomatitis and nausea, diarrhea, hair loss, and mild leukopenia can occur with methotrexate. Serum aminotransferase elevations sometimes may be seen, but they correlate poorly with the more serious complication of hepatic fibrosis. Liver biopsy is performed routinely in patients with psoriasis after cumulative doses of 1.5, 3, and 5 g have been administered, but these guidelines have not been widely adopted in patients with rheumatoid arthritis, in whom the risk of hepatic fibrosis appears to be lower. In one series of IBD patients who received a mean cumulative dose of methotrexate greater than 2.5 g and had liver biopsy, there was only minimal hepatic toxicity.300 Obesity, diabetes, and alcohol intake correlate with hepatic fibrosis. Methotrexate interacts with sulfa medications and with azathioprine and 6-MP to cause severe leukopenia. Rare but potentially life-threatening interstitial pneumonitis can manifest as cough and dyspnea of insidious onset. Early detection, cessation of methotrexate, and prompt treatment with glucocorticoids is essential. Methotrexate is toxic to sperm, and men should wait three months after stopping methotrexate before trying to conceive.301 Finally, methotrexate is a potent abortifacient and is strongly teratogenic. Women of childbearing capacity must use methotrexate only with highly effective contraception. Methotrexate may be considered as an alternative to the thiopurine analogs, particularly among patients who do not tolerate these drugs. Some patients who do not respond to

Chapter 111  Crohn’s Disease 6-MP might respond to methotrexate.302 In addition to its proven role as a glucocorticoid-sparing agent, methotrexate may be considered as a treatment for active disease, although its value for this indication is less clear.303

Other Immune Modulators

There appears to be little role for cyclosporine in Crohn’s disease. Series of uncontrolled and randomized, controlled trials have shown high doses of cyclosporine to be efficacious in treating inflammatory disease and fistulas but at an unacceptably high cost of adverse effects. Moreover, lower doses, although somewhat safer, are not effective in maintaining remission304; for virtually all indications, equally effective and less-hazardous medications are available. Tacrolimus is absorbed more reliably from the intestine than is cyclosporine and has a similar mode of action via inhibition of calcineurin, thereby diminishing T-cell activation.305 Preliminary data suggest that tacrolimus may be useful in treating glucocorticoid-resistant disease, and a randomized, controlled trial has demonstrated efficacy in healing fistulas.306,307 The drug also may be effective as a topical agent for oral and perianal ulcerating disease.308,309 Mycophenolate mofetil, like azathioprine and 6-MP, inhibits purine synthesis. Because of the similarity of these agents in mode of action, mycophenolate mofetil had been considered primarily as an alternative treatment for patients intolerant of or resistant to treatment with azathioprine. Studies have shown mixed results, however, and the role of mycophenolate mofetil in Crohn’s disease remains unclear.310,311

Biological Therapy

Anti–Tumor Necrosis Factor Agents Infliximab is the first biological response modifier shown to be effective in Crohn’s disease. This chimeric monoclonal TNF antibody had an unsuccessful history as an investigational antisepsis agent before its use in Crohn’s disease was explored. Despite conflicting reports regarding the importance of TNF in IBD, the Dutch investigator van Deventer posited that in light of the critical role of TNF in granuloma formation, an anti-TNF agent might prove efficacious for granulomatous bowel disease. Open-label trials subsequently demonstrated rapid and prolonged improvement in disease activity, accompanied in many cases by mucosal healing.312-314 A randomized, controlled trial provided strong confirmation of the initial impression of efficacy. Patients with moderate to severe Crohn’s disease were randomized to an initial infusion of placebo or 5, 10, or 20 mg/kg infliximab, then called cA2.315 Qualifying patients had moderate to severe Crohn’s disease (CDAI, 220 to 400) despite treatment with aminosalicylates (59%), oral glucocorticoids (59%), or 6-MP or azathioprine (37%). Approximately half of the studied patients had had prior segmental intestinal resections, and the group had a mean duration of disease in excess of 10 years. The major end point was clinical response, defined as a decrease in the CDAI of 70 or more points at week four. All treatment groups had results significantly better than those with placebo (placebo response rate of 17%). The highest rate of response was seen at four weeks in the 5-mg/kg group (81%).315 A smaller but still significant fraction of patients had a clinical response by week 12 (48% for 5 mg/kg vs. 12% for placebo). The fraction of patients in clinical remission (CDAI less than 150 and a decrease in CDAI of 70 or more points) at week four also was significantly higher among the 5-mg/kg group (33%) compared with the placebo group (4%). Time to response for nearly

all patients was two weeks. Clinical improvement was accompanied by improvement in health-related quality of life and decreases in serum C-reactive protein levels. Coincidental healing of enterocutaneous fistulas in some patients led to a separate randomized, controlled trial of infliximab for this indication. Patients with draining enterocutaneous fistulas were enrolled and followed for closure of 50% or more of the fistulas at two successive visits one month apart. More than one half of the patients had more than one fistula, and 90% of the fistulas were perianal. In­ fliximab in a dose of 5 or 10 mg/kg or placebo was infused at weeks 0, 2, and 6. Among patients assigned to infliximab 5 mg/kg, 68% achieved the primary end point, compared with 26% of those given placebo (P = 0.0002).316 Complete closure of all fistulas was observed in 55% of patients given infliximab 5 mg/kg but in only 13% of placebo-treated patients. In patients who achieved the primary end point, the median duration of response was three months. Limited information regarding the efficacy of repeated dosing was available before the commercial release of in­ fliximab in the United States. Maintenance dosing every eight weeks was demonstrated to maintain response in patients with fistulizing and nonfistulizing disease.317,318 In these studies, although significant differences were seen at the primary endpoints, the treatment effect was not as robust as with initial therapy. A trend emerged that continues to hold with the newer biological agents as well: Approximately 60% of patients with luminal disease initially respond to therapy and 40% of those maintain that response at one year. Other important observations from these studies included demonstrating the steroid-sparing effect of infliximab, and sustained improvement in quality of life out to the 54-week duration of these trials. In addition, patients who failed to respond to induction dosing were unlikely to respond to additional repeated dosing. Infliximab was approved for the treatment of moderate to severe pediatric Crohn’s disease by the United States Food and Drug Administration in 2006. In a study design similar to that of the adult maintenance trials, 112 pediatric patients received an open-label infliximab induction regimen at weeks 0, 2, and 6.319 The responders then were randomized to receive 5 mg/kg infliximab every eight weeks or every 12 weeks. There was no placebo group. At week 10, 88% of patients had responded to therapy and 58.9% were in remission. At week 54, in the every-eight-weeks dosing group, 63.5% of the responders were still responding, and 55.8% were in remission; this is compared with 33.5% and 23.5% response and remission rates in the every-12-weeks dosing group (P = 0.002 and P < 0.001, respectively). These very high response and remission rates likely reflect the early nature of disease in children compared with adults and a different phenotype (e.g., inflammatory) at young age. Since the success of infliximab, two additional anti-TNF agents have been approved for treating Crohn’s disease. Adalimumab is subcutaneously administered human immunoglobulin G1 (IgG1) monoclonal antibody targeting TNF. After success in rheumatoid arthritis, an early open-label study was performed of patients with Crohn’s disease who had lost response to or become intolerant to infliximab. In this study, 59% responded to adalimumab therapy and 29% of patients were in remission at 12 weeks.320 Importantly, no patients experienced acute or delayed hypersensitivity reactions to adalimumab. This paved the way to larger randomized, controlled trials that showed efficacy for inducing remission321 and maintenance at one year.322 Based on results from these studies, the recommended adult dosing is 160 mg subcutaneously at week 0, 80 mg at week two and then 40 mg every other week.

1967

1968

Section X  Small and Large Intestine Certolizumab pegol, initially known as CDP870, is a polyethylene-glycolated Fab′ fragment of a humanized antiTNF antibody. An exploratory study of a single dose in patients with Crohn’s disease was promising323 and led to a series of large randomized, controlled trials. In the first, although clinical benefit clearly was evident, the primary endpoint of clinical response at 12 weeks missed statistical significance.324 Post hoc analysis exposed a high placebo rate in subjects with low CRP levels, and subsequent recalculation limiting analysis to patients with an elevated CRP showed a significant difference. In two follow-up studies that stratified patients based on CRP levels, certolizumab pegol 400 mg, administered subcutaneously at weeks 0, 2, and 4 weeks and then every four weeks, proved to be effective to induce and maintain clinical response and remission out to 26 weeks.325,326 Interestingly, patients responded equally regardless of their CRP status. Infliximab, adalimumab, and certolizumab pegol are all effective for treating moderately to severely active Crohn’s disease. Although it is impossible to compare the response and remission rates of these agents because of different study designs and patient populations, all appear to have similar rates of initial response (approximately 60%) and maintenance of that response between six and 12 months (approximately 40% of the initial responders). Differences may be found over the longer term based on specific modes of action, pharmacokinetics, development of antibodies, and side effects. Although these three medications are similar, all anti-TNF agents are not equivalent, as illustrated by etanercept (a human soluble TNF receptor Fc fusion protein) and CDP571 (a humanized monoclonal antibody to TNF), both of which failed to show significant efficacy for Crohn’s disease.327,328 The mode of action of anti-TNF agents is likely to be more involved than just its nominal binding of TNF. The antibody can bind and clear soluble TNF, but it also binds to cellbound TNF. Through the latter mechanism, infliximab and adalimumab have been shown to induce apoptosis of cells expressing membrane TNF. Etanercept does not induce apoptosis, an observation thought to be the explanation for its lack of efficacy;331,332 However, certolizumab pegol also does not induce apoptosis,333 and, as mentioned earlier, has efficacy similar to that of these other agents. Therefore, the mechanism of action is most likely multifactorial, all pointing toward the ability to control the mucosal immune response. Treatment with the anti-TNF agents usually is well tolerated. In the largest and longer-term clinical trials, between 4% and 16% of patients withdrew from the study due to an adverse event.318,334 Injection site and infusion reactions occurred at variable rates, and they were highest with infliximab and lowest with certolizumab pegol. Infusion reactions with infliximab typically are associated with antibodies to infliximab (ATI), also referred to as HACA (human antichimeric antibodies). ATI developed in 13% of infliximab-treated patients with Crohn’s disease. Patients in whom ATI develop are more likely, although not uniformly so, to experience acute infusion reactions, which can consist of chest tightness, dyspnea, rash, and hypotension. ATI are less likely to develop in patients treated concomitantly with glucocorticoids or immune modulators, providing a justification for continuing methotrexate, azathioprine, or 6-MP, even when these treatments have failed. Delayed hypersensitivity reactions, consisting of severe polyarthralgia, myalgia, facial edema, urticaria, or rash, are unusual complications that can occur two to 12 days after an infusion.332 High ATI concentrations appear in such patients after the occurrence of such reactions, but they are

not necessarily found before reinfusion. The major risk factor for delayed hypersensitivity appears to be a long delay (probably six months or more) between infusions, thereby priming an amnestic antibody response.332 Delayed hypersensitivity appears to be less common when a standard induction regimen is used and when an immune modulator is given concurrently.332 Antibody formation is not unique to infliximab. For ada­ limumab, human antihuman antibodies (HAHAs) can form. They were seen in only 2.6% of patients in a one-year maintenance study of adalimumab in Crohn’s disease334 but in up to 17% in patients with rheumatoid arthritis.335 Anti­ bodies to certolizumab were seen in approximately 10% of patients in the induction and maintenance trials.324,326 The clinical significance of the presence of antidrug antibodies is a matter of debate, but data demonstrate an association with lower infliximab serum levels in the setting of episodic therapy, where ATI formation is highest,336 and a decreased response rate to adalimumab among patients with HAHAs.337 Antinuclear antibodies are common and appear in approximately 50% of patients receiving infliximab after two years. Of the patients who develop antinuclear anti­ bodies, approximately 30% develop anti–double-stranded DNA.336 Actual drug-induced lupus is rare, but it can occur. Treatment with certolizumab pegol has been observed to induce a lower rate of antinuclear antibodies than does infliximab or adalimumab, perhaps a result of this agent’s inability to induce T-cell apoptosis. The clinical significance of these autoantibodies is unclear. Infections occur more commonly during treatment with anti-TNF agents. In clinical trials, infections were reported in up to 57% of anti-TNF treated patients.334 Serious infections fortunately were unusual, occurring in only 2% to 4% of patients, which was consistent across the three anti-TNF agents.317,322,325 In the course of treatment of patients with enterocutaneous fistulas, perianal abscesses can arise from superficial healing and closure of an infected pocket. Any patient with a suspected pyogenic complication of Crohn’s disease or any serious infection should undergo adequate drainage and treatment with antibiotics before starting or continuing infliximab. A systematic review found that sepsis can be fatal, with a mortality rate of 0.4% in antiTNF–treated patients.339 Patients with sepsis typically were older and had comorbidities. Primary contributing factors included the concomitant use of narcotics, prednisone and other immune suppressing agents, which can independently increase the risk of serious infections and death.254,340 Reactivation of tuberculosis has been observed with antiTNF therapies, including infliximab, and has resulted in disseminated disease and death. An estimated rate of tuberculosis occurrence from the rheumatology literature is 5 per 10,000 anti-TNF–treated patients.341 All patients should be screened for pulmonary tuberculosis before starting therapy. In a meta-analysis of patients with Crohn’s disease treated with anti-TNF agents, the rate of NHL was 6.1 per 10,000 patient-years.342 Compared with the background rate in the general population of 1.9 per 10,000 patient-years, this corresponds to an standardized incidence ratio of 3.23 (95% CI, 1.5 to 6.9). Compared with the rate of NHL in Crohn’s patients seen in the meta-analysis of immunomodulators noted earlier,284 the standardized incidence ratio is 1.7 (95% CI, 0.5 to 7.1). Because most of the anti-TNF treated patients also were exposed to immunomodulators, it is not possible to determine the magnitude of risk contributed by the antiTNF treatment alone as opposed to the combination of both agents. In addition, as noted with immunomodulators, the influence of Crohn’s disease on NHL risk is not resolved. A

Chapter 111  Crohn’s Disease reasonable estimate is that both immunomodulators and anti-TNF agents increase the absolute risk of NHL to a real but small extent. Monotherapy probably is safer than combination therapy. Hepatosplenic T-cell lymphoma (HSTCL) has been described in patients with Crohn’s disease who have been treated with anti-TNF agents in combination with immunomodulators and with immunomodulators alone (6-MP and azathioprine).337 This nearly universally fatal form of NHL predominantly affects young men, but its incidence is unknown; fortunately it appears to be rare, and more often occurs in the setting of combination therapy. This, in addition to the uncertain magnitude of benefit of combination therapy, has led to an emphasis on immunomodulator or anti-TNF monotherapy, particularly in young male patients.343 Infliximab, adalimumab, and certolizumab pegol are class B agents for use in pregnancy. Most of the clinical data in Crohn’s disease comes from experience with infliximab, and there are early data with adalimumab. Infliximab crosses the placenta in the third trimester and is detectable in the infant for several months after birth.344 Certolizumab pegol theoretically does not cross the placenta in the third trimester, potentially preventing neonatal exposure to the drug, but it can cross the placenta in the first trimester in low levels. The rate of birth defects does not appear to be elevated in the pregnancies that have occurred while the mother is on anti-TNF therapy, but safety is not clearly established and use during pregnancy has not been routine. Optimizing Anti–Tumor Necrosis Factor Response.  Proper selection of patients is the key to using anti-TNF agents safely, effectively, and appropriately. Patients without objective findings of inflammation or with fibrostenotic disease are unlikely to benefit, and treating patients who have an undrained abscess is likely to be unsafe. In addition to proper patient selection, opportunities to optimize the response to anti-TNF treatment include intensifying the dose, avoiding smoking, using concomitant immunomodulators, and performing early aggressive therapy. Dose intensification often is required to regain response. For infliximab, this means increasing the dose from 5 mg/kg to 10 mg/kg or increasing the maintenance frequency from every eight weeks to every six weeks. This tactic works in the majority of patients and allows therapy to be continued after an attenuated response.345 Increasing adalimumab dosing to 40 mg weekly (or 80 mg every two weeks) or giving an additional certolizumab pegol dose can yield similar results.322,346 Because the response rate to a second anti-TNF agent is less than the first, dose intensification is favored before moving on to a different anti-TNF agent. Smoking is a strong negative predictor of response to anti-TNF therapy, and all patients need to be informed of the poor initial and long-term outcomes associated with smoking.347 The value of the concomitant use of immunomodulators (6-MP, azathioprine, or methotrexate) to improve the antiTNF initial response and maintenance effect is controversial. As noted earlier, immunomodulators appear to decrease the production of antidrug antibodies, thereby increasing serum drug levels, but the clinical significance of such increased levels is uncertain. Furthermore, the additional risk of infections and lymphoma (including HSTCL) resulting from combination therapy make the decision to use these agents a difficult one. The first prospective study addressing this question (i.e., treating patients with infliximab alone versus infliximab in combination with methotrexate) showed no difference in outcomes at the end of one

year.348 Although post hoc analyses of randomized, controlled trials support this result, one study did show a statistically significant benefit of combination azathioprine and infliximab therapy compared with either azathioprine or infliximab alone.349 This large study favors combination therapy, but the long-term benefit weighed against the risks remains uncertain. The strategy of early aggressive therapy for Crohn’s disease has gained much attention because of the enthusiasm of this approach in rheumatoid arthritis, and favorable results have been reported in Crohn’s disease.350 In the landmark study by D’Haens, 133 patients with recent-onset active Crohn’s disease were randomized to receive early combined immunosuppression with azathioprine and in­fliximab (followed by maintenance azathioprine and ondemand infliximab) or conventional treatment with the sequential use of prednisone, azathioprine, and then infliximab.350 At the end of one year, 62% of patients in the earlycombination therapy group were in remission compared to 42% in the conventional group. By study design, all patients in the conventional group received glucocorticoids, but no patients in the early-combination group required glucocorticoids. Finally, at two years, 73% of patients in the earlycombination group had complete mucosal healing, compared with 30% in the conventional group. The authors concluded that more-intensive treatment early in the course of Crohn’s disease could lead to better outcomes. There are significant limitations in this open-label study, including the lack of an arm of early immunomodulator treatment alone. Also, the use of episodic infliximab therapy is not standard of care, although it must be acknowledged that this strategy might yet prove effective in disease of recent onset. Despite these and other limitations, an important point is made that aggressive treatment early in the course of Crohn’s disease can prevent the need for glucocorticoids and lead to a high rate of mucosal healing. Further studies are needed to confirm these data, but this work has highlighted the potential benefit of early aggressive therapy. Natalizumab Natalizumab is humanized monoclonal antibody against α4integrin that inhibits leukocyte adhesion and migration into inflamed tissue. Also used for the treatment of multiple sclerosis (MS), natalizumab is the first new class of drug approved for the treatment of Crohn’s disease since infliximab’s approval in 1998. It is administered intravenously at a dose of 300 mg every four weeks. An early dose-ranging study of 248 patients did not show a significant difference for remission at six weeks compared with placebo, but it did show promising results that led to further studies.351 A large randomized, controlled trial, Efficacy of Natal­ izumab as Active Crohn’s Therapy (ENACT)-1, had as a primary endpoint the percentage of patients with a clinical response at week 10.71 ENACT-2 randomized responders from ENACT-1 to continued treatment with either natal­ izumab or placebo through week 56.71 The primary endpoint of ENACT-2 was the proportion of responders from ENACT-1 who sustained a clinical response through week 36. Natalizumab-treated patients in ENACT-1 had a response rate of 56% at week 10 compared with a 49% placebo response (P = 0.05). Sixty-one percent of natalizumabtreated responders maintained their response in the ENACT-2 trial, compared with 28% of the placebo-treated patients (P < 0.001), and 44% of natalizumab-treated responders were in remission compared with 26% of placebo-treated patients (P = 0.003). A follow-up induction study of 509 patients with moderately to severely active Crohn’s disease showed a clear benefit of natalizumab with

1969

1970

Section X  Small and Large Intestine a significantly higher response rate than placebo at eight weeks that was sustained until the end of the study at week 12 (48% versus 32%, P < 0.001).352 Conclusions from these studies were that natalizumab is an effective induction and maintenance agent for the treatment of moderately to severely active Crohn’s disease. Unfortunately, in the open-label extension portion of the ENACT studies, one patient died from progressive multi­ focal leukoencephalopathy (PML).353 PML is a progressive degenerative neurologic disease caused by infection with JC virus; it is typically seen in patients with acquired immune deficiency syndrome (AIDS). This information, together with two additional reports of PML in natalizumab-treated patients with MS led to a temporary withdrawal of natal­ izumab from the market for MS (for which it had already been approved) and suspension of further clinical studies in Crohn’s disease. After careful scrutiny and follow-up of all treated patients,354 natalizumab was returned for use in MS and approved by the FDA in 2008 for treating patients who have moderate to severe Crohn’s disease with evidence of active inflammation refractory to prior treatment, including anti-TNF. A mandatory patient registry, Tysabri Outreach Unified Commitment to Health (TOUCH), tracks all treated patients in the United States. Since return to the market, 10 additional cases of PML were reported in patients with MS, but to date there have been no further cases in Crohn’s disease (personal communication, Elan Pharmaceuticals, Inc., San Diego, Calif). In addition to PML, severe hepatic toxicity that could lead to death or the need for liver transplantation has been reported, but this also appears to be very rare.354,355 Typically, natalizumab is very well tolerated: approximately 10% of patients in the treatment arm of ENACT withdrew because of an adverse event, and serious infections were reported at a rate of 2% to 3%, which is no different from the rate in placebo-treated patients. The future of natalizumab for the treatment of Crohn’s disease is promising, but careful attention will be paid to its continued safety record.

Adjunctive Therapies

Many other therapies are used to control the symptoms and adverse consequences of Crohn’s disease. Antidiarrheal and anticholinergic agents can help to alleviate symptoms. Patients with ileal disease or resection can require parenteral vitamin B12 supplementation or the addition of cholestyramine (1 to 4 g/day) or colesevelam (625 to 3800 mg/ day) to control bile salt diarrhea. Iron supplementation also may be needed. Smoking cessation should be vigorously pursued as a means of improving long-term outcomes.356 Bone loss should be anticipated as a potentially serious complication in all patients. Bone density should be checked at diagnosis and at regular intervals thereafter, with appropriate medical management of bone loss. Strategies to preserve bone density include smoking cessation and, at a minimum, daily supplementation of calcium and vitamin D; 1000 mg/day of elemental calcium is enough for younger men and premenopausal women, and men and women older than 50 years should have as a goal elemental calcium intake of 1500 mg/day. For most people, 400 to 800 IU of vitamin D daily is adequate.357 In patients at the highest risk and without contraindications, bisphosphonates should be considered at the time corticosteroids are initiated.

Novel Therapies

Progress in understanding the pathogenesis of Crohn’s disease has borne fruit in the development of a wide variety

of novel therapeutic agents. In addition to holding real promise for safer and more effective therapy in the future, clinical trials of novel therapeutics offer access to these agents for patients whose disease has exhausted approved therapies. Promising agents under investigation include antibodies directed against the shared p40 subunit of IL-12 and IL-23. Anti-IL12 (ABT-874) is a recombinant, human IgG1 antibody directed against p40, a component of both IL-12 and IL-23. Early data suggest that patients with active Crohn’s disease respond to this molecule and, in fact, do have a decrease in the secretion of Th1-mediated cytokines in response to treatment.358 The human anti IL-12/23 monoclonal antibody ustekinumab was studied in a blinded crossover trial with 104 patients with moderate to severe Crohn’s disease.359 Although the primary endpoint of clinical response at eight weeks was not achieved, significant differences compared with placebo were seen at other time points, with the largest differences seen in patients who had previously received infliximab. Treatment was well tolerated. Serious adverse events were rare and primarily related to active Crohn’s disease. Open-label studies with thalidomide, which has antiangiogenic properties in addition to destabilizing TNF mRNA, are promising, but the potent teratogenicity and frequent neuropathy of this agent precludes widespread use.360,361 Lenalidomide, an analog of thalidomide with similar immunomodulatory properties but with less toxicity (specifically neuropathy) was studied in a randomized, controlled trial, but it did not show a response significantly different from placebo.362 Probiotic therapies have been examined as a safe means of modulating the intestinal immune response in IBD, but studies in pediatrics and in the setting of postoperative prophylaxis have been disappointing.363,364 Some novel agents defy conventional approaches to the treatment of Crohn’s disease. Porcine whipworm (Trichuris suis) has been administered as a possible treatment for both Crohn’s disease and UC, with promising effect and excellent safety and tolerability.365 Theoretically, this iatrogenic helminthic infestation might prove effective through the induction of regulatory T cells. Recombinant granulocyte-macrophage colony stimulating factor (GM-CSF, or sargramostim) has been tried as a therapeutic agent for Crohn’s disease on the hypothesis that defective innate immune responses, rather than defects in the adaptive immune response, are a contributing factor to Crohn’s disease.366 Administering sargramostim can, in theory, bolster the innate immune response by stimulating neutrophil function. Adult mesenchymal stem cells are being studied for the treatment of refractory Crohn’s disease. The inherent antiTNF activity of these cells may be responsible for the response observed in a short-term, open-label study with good safety results.367 Although the hypotheses underlying these unusual therapies have yet to be proved, the excitement over the novel mechanisms and how they inform us of the pathophy­siology of Crohn’s disease continue to stimulate enthusiasm.

NUTRITIONAL THERAPY

Nutritional therapy in Crohn’s disease conceivably has two purposes: repletion of nutrients and treatment of the primary disease (see Chapters 4 and 5). Specific deficits should be identified and corrected. Protein-calorie malnutrition should be addressed, preferably with enteral supplementation. Many, but not all, patients with Crohn’s disease are lactose intolerant and can require increased calcium

Chapter 111  Crohn’s Disease supplementation. Total parenteral nutrition may be considered for patients with severe malnutrition before surgery or for selected patients with severe Crohn’s disease as a primary therapy in combination with bowel rest.117 Patients with short bowel syndrome from numerous small bowel resections can require enteral nutrition with defined diets; rarely, patients with severe short bowel syndrome require life-long total parenteral nutrition. A meta-analysis has found defined enteral diets to be inferior to glucocorticoids in achieving clinical response,368 but defined enteral or polymeric diets still may be useful in some children for whom glucocorticoids are undesirable.117,368,369 Elemental diets do not appear to be superior to polymeric diets. Children may be taught to receive nocturnal feedings after self-intubation with a nasogastric tube. Long-term tolerance may be poor, however, and disease tends to recur when the patient’s usual diet is reintroduced. Self-reported food intolerances are common among patients with Crohn’s disease,370 but exclusion diets have not been shown to be beneficial.371 Eliminating multiple food items can lead to serious malnutrition. Patients with a stricture might tolerate roughage poorly or might experience complete intestinal obstruction. Increasingly, specific nutrients are being considered for their therapeutic properties. An example includes the use of omega-3 free fatty acids. After initial enthusiasm, two randomized, controlled trials did not show a difference with omega-3 free fatty acids as monotherapy compared with placebo for maintaining remission of Crohn’s disease.372

SURGICAL THERAPY

Surgery plays an integral role in the treatment of Crohn’s disease to control symptoms and to treat complications. By the 20th year from the onset of symptoms, roughly three fourths of patients have had surgery.373 Depending on the prevalent medical culture in the country of study, the rate of surgery within three years of diagnosis varies from 25% to 45%. From 25% to 38% of patients require a second surgery by five years after the first, and about one third of patients who need a second surgery eventually require a third.374 Because of the high likelihood of recurrence after segmental resection, the guiding principle of surgery in Crohn’s disease is preservation of intestinal length and function. Taking wide margins does not reduce the likelihood of recurrent disease and, with repeated resection, can contribute to intestinal failure. Surgical procedures may be categorized as resections with or without anastomosis, external (i.e., stoma) or internal bypass surgery, and a variety of surgical approaches for repair or resection of a fistula.375 Stricturoplasty rather than resection may be appropriate for patients with multiple fibrotic nonphlegmonous strictures of the small intestine and for patients who have short bowel syndrome or who are at risk for this complication because of prior resections.375 Although 80% of patients who undergo ileocecal resection have evidence of endoscopic recurrence in the neoterminal ileum within one year of surgery,74 the time to symptomatic recurrence usually is several years. Recurrent disease almost always occurs proximal to the anastomosis. Small intestinal recurrence after proctocolectomy and permanent ileostomy for Crohn’s disease of the colon is relatively uncommon. Indications for surgery include complications such as intra-abdominal abscess, medically intractable fistula, fibrotic stricture with obstructive symptoms, toxic megacolon, hemorrhage, and cancer.376 Patients with symptoms refractory to medical therapy also should be considered for

surgery, particularly when the patient remains dependent on or refractory to glucocorticoids despite optimal medical therapy. Some patients prefer to consider a limited small bowel resection as opposed to a trial of immunomodulator or biologic therapy.377 Among children, a well-timed bowel resection may be indicated for growth failure. In patients with indeterminate colitis for whom colectomy is required, ileal pouch-anal anastomosis (IPAA) can be considered, but pouch-related complications are seen at a higher rate than in patients with UC.191 In patients with Crohn’s disease, however, there is a high rate of pouch failure and IPAA typically is avoided. In selected cases with rectal sparing and lack of fistulizing behavior, IPAA or ileorectal anastomosis may be considered (see Chapter 112).378,379 Increasing facility with laparoscopic approaches to Crohn’s disease might reduce operative morbidity and improve the safety of surgery.380

COSTS OF CARE Substantial medical and societal costs are incurred in the course of Crohn’s disease. A study from Sweden showed that although UC is twice as prevalent as Crohn’s disease, the total combined medical and societal costs of Crohn’s disease are twice that of UC.381 An analysis of data from a population-based cohort in the United States found that projected lifetime costs of medical care exceeded $40,000 when median charges were applied.382 Surgery generated the largest proportion of costs (44%), but nearly two thirds of patient time was spent off medical therapies, largely reflecting surgically induced remission.382 Short of a cure, safe, well-tolerated, effective, and inexpensive means of maintaining remission should provide the greatest economic impact in the care of patients with the disease. The biologic agents can prevent hospitalizations and operations and thereby lead to overall cost savings, despite their significant expense.383

CROHN’S DISEASE IN THE LIFE CYCLE CHILDREN AND ADOLESCENTS

Approximately 25% of new Crohn’s disease diagnoses are made in persons younger than 20 years. In most respects, Crohn’s disease has the same pathophysiology and clinical features in children as it does in adults. The special consequences of Crohn’s disease in children and adolescents relate to the vulnerability of this population to disturbances in physical growth, sexual maturation, and psychosocial development. Deceleration of growth velocity can precede gastrointestinal symptoms in as many as 20% of children.384 Correction of nutritional deficits and vigorous treatment of inflammation will lead to normal growth and development in most children with Crohn’s disease. The potential for glucocorticoids to cause mood disturbances and cosmetic side effects can have dire implications for a child’s psychosocial development, but their role in decreasing height velocity has been questioned.385 Increasingly, immune modulators are being incorporated into pediatric treatment regimens as a means of minimizing glucocorticoid use.291,386 In evaluating disease, medical personnel must be particularly sensitive to the trauma of intrusive and sometimes painful examinations and procedures that are done routinely in adults. Sadly, children with Crohn’s disease also are subject to the vicissitudes of their social circum-

1971

1972

Section X  Small and Large Intestine stances: underinsured children are more likely to present with more-severe disease, undoubtedly because of poor access to medical care and prolonged delays in diagnosis.387 As much as possible, children with Crohn’s disease should be permitted to function normally in school and extracurricular activities, although special accommodations sometimes need to be arranged.

SEXUALITY, FERTILITY, AND PREGNANCY

Crohn’s disease affects many persons in the peak of their reproductive years. Studies have varied in the assessment of female fertility in Crohn’s disease, showing either no difference from the general population or a slight decrease. Studies that have detected diminished fertility generally have correlated this finding with increased disease activity. Contributing factors can include true infertility or a conscious decision to avoid childbearing. In men and women, decreased libido because of symptoms such as diarrhea, abdominal pain, and fatigue are not uncommon; in women, dyspareunia and rectovaginal fistulas can play additional roles. Except for reversible sperm abnormalities caused by sulfasalazine and methotrexate, men with Crohn’s disease have normal fertility. The effect of pregnancy on the course of Crohn’s disease depends on the status of the disease at conception. Women with quiescent disease at conception have the same rate of relapse during pregnancy as nonpregnant women. Among women with active disease at conception, the one-third rule applies: one third improve, one third worsen, and one third have unchanged symptoms during their pregnancy.388 HLA disparity at DRB1 and DQ between mother and fetus is strongly associated with improvement in IBD symptoms during pregnancy.389 Postpartum relapses are uncommon and tend to be mild. Most pregnancies carried by women with Crohn’s disease are normal, but there appears to be an increased rate of adverse conception outcomes (spontaneous abortion), adverse pregnancy outcomes (preterm birth, child small for gestational age, and stillbirth) and pregnancy-related complications.390 Other than preterm birth, the contribution of disease activity to these outcomes may be lower than previously thought.390,391 The rate of cesarean section is not increased compared with that of the general population, and perineal complications occur infrequently among women who deliver vaginally with an episiotomy if there is not active perineal disease at the time of delivery.392,393 For a review of the safety of medical therapies in pregnancy and nursing, see Table 111-5.

THE AGING PATIENT

Compared with younger patients with Crohn’s disease, older patients are more likely to have disease of the colon, particularly of the distal colon. As with children, the presentation may be subtle; extraintestinal symptoms can predominate, and diagnosis may be delayed. Medical management is essentially no different for the older patient, but the clinician more often must consider the variety of other conditions prevalent among older patients when choosing therapies. Thus, for example, older persons treated with glucocorticoids are at increased risk for hypertension, hypokalemia, and confusion.394 Glucocorticoids also tend to complicate the management of diabetes. Antibiotics can diminish vitamin K production by intestinal flora and cause an excessively prolonged prothrombin time with warfarin therapy. Anticholinergic drugs can induce urinary retention, altered mental status, or glaucoma. As expected, older IBD patients who are hospitalized have a higher morbidity and mortality than younger patients.395

PROGNOSIS MORBIDITY

The natural history of Crohn’s disease is a moving target, continuously changing as therapeutic strategies improve.84 The course of disease is highly variable and difficult to predict for a given patient. Population-based studies from Scandinavia provide the best information regarding the course of disease. In the first year after diagnosis, the cumulative relapse rate is high, approaching 50%, with 10% of patients having a chronic relapsing course.396 Thereafter, patients generally are true to their own history: The rate of relapse in the first two years of the disease correlates with the risk of relapse in the ensuing five years.397 Symptomatically active disease in the preceding year yields a high likelihood of active disease in the next year. Conversely, a year in which symptoms are quiescent has an 80% probability of being followed by another year without exacerbation.397 Over a four-year period, the same analysis has shown that 22% of patients remain in remission, 25% experience chronically active symptoms, and 53% have a course that fluctuates between active and inactive disease.397 Although most persons continue to lead productive lives, the course of the disease may be punctuated by periods of poor productivity. Over time, approximately 10% of patients are disabled by their disease. Increasingly, serologic markers are recognized as providing prognostic information. Pediatric patients with Crohn’s disease who have higher immune responses to the microbial antigens ASCA, anti-CBir, and anti-OmpC have higher rates of complicated disease.398 Furthermore, the highest antibody sum group has the most rapid disease progression. Similar results also have been seen in adults, and with other serologic antibody markers (antibodies to carbohydrates).102,399 There have been mixed results regarding the utility of genetic polymorphisms in predicting the natural history of disease, perhaps reflecting our currently incomplete understanding of the genetic and environmental factors that shape disease expression.398,400

CANCER

The estimated risk of colorectal cancer (CRC) in Crohn’s disease has varied widely, ranging from no more risk than that of the general population to an estimated standardized incidence ratio as high as 26.6.401 When Crohn’s disease involves the large bowel, the excess risk of CRC appears to be similar to that in UC of similar extent.401 The characteristics and prognosis of CRC in Crohn’s disease also are similar to those for CRC in UC.402 For these reasons, surveillance colonoscopy has been recommended as a means of early detection.403 A thinner-caliber colonoscope may be required to traverse narrowed bowel. Segments of bowel excluded by diversion procedures are at greatly increased risk for developing CRC and present a great challenge to early detection. Little controversy surrounds the increased risk of small bowel adenocarcinoma associated with long-standing disease or in bypassed loops of small intestine. Small intestinal cancers are rare in Crohn’s disease, but by comparison to the extremely low incidence of this disease in the general population, there is a high associated relative risk.404 The association between Hodgkin’s and non-Hodgkin’s lymphomas and Crohn’s disease remains unclear. Studies relying on cases at referral centers have found an increased risk of lymphoma, whereas population-based studies have not.285,286,405 The most likely explanations are either a referral bias or an increased risk confined to patients with

Chapter 111  Crohn’s Disease more-severe disease. Squamous cell carcinomas can arise in association with a chronic fistula to the skin and in anal Crohn’s disease. Some studies also have found an association between Crohn’s disease and respiratory cancers,406 perhaps attributable to smoking behavior.

MORTALITY

Population-based studies generally have shown a modestly increased mortality rate in Crohn’s disease.407-411 The cause of death is disease related more often than not, or it is indirectly related as a result of gastrointestinal malignancies.412,413 A recent population-based study demonstrated a 40% increase in the standardized mortality rate among patients with Crohn’s disease compared with controls, but the absolute difference in death rates was still small.414 Specifically, the standardized mortality rate among Crohn’s disease patients was 66.9 deaths per 10,000 person-years versus 49.7 deaths per 10,000 person-years in the general population, representing an excess of 17 deaths per 10,000 person-years attributable to Crohn’s disease.

COPING WITH CROHN’S DISEASE Although the old myths surrounding psychopathology as an underlying cause of IBD have long been debunked, coping with diarrhea, pain, malaise, and decreased energy takes a toll on all persons who suffer from Crohn’s disease as well as on their families. Depression and anxiety often diminish daily functioning that already may be impaired by the physical manifestations of the disease; psychosocial functioning, of course, has a large impact on the patient’s quality of life, even in the absence of active disease.415,416 Patients cite concerns about lack of energy, loss of control, body image, fear and isolation, feeling unclean, and not reaching their full potential.417 The medical provider can help greatly in alleviating these concerns by providing accurate and plentiful information. Lay organizations such as the Crohn’s and Colitis Foundation of America provide valuable resources in support of affected persons and their families (http:// www.ccfa.org). An attitude of hopefulness is warranted, as an astounding number of therapeutic innovations—and someday perhaps a cure—unfold.

KEY REFERENCES

Barrett JC, Hansoul S, Nicolae DL, et al. Genome-wide association defines more than 30 distinct susceptibility loci for Crohn’s disease. Nat Genet 2008; 40:955-62. (Ref 45.) Colombel JF, Sandborn WJ, Rutgeerts P, et al. Adalimumab for maintenance of clinical response and remission in patients with Crohn’s disease: The CHARM trial. Gastroenterology 2007; 132:52-65. (Ref 322.) Crohn BB, Ginzburg L, Oppenheimer GD. Regional ileitis, a pathological and clinical entity. JAMA 1932; 99:1323-9. (Ref 4.) D’Haens G, Baert F, van Assche G, et al. Early combined immunosuppression or conventional management in patients with newly diagnosed Crohn’s disease: An open randomised trial. Lancet 2008; 371: 660-7. (Ref 350.) Feagan BG, Fedorak RN, Irvine EJ, et al. A comparison of methotrexate with placebo for the maintenance of remission in Crohn’s disease. North American Crohn’s Study Group Investigators. N Engl J Med 2000; 342:1627-32. (Ref 295.) Hanauer SB, Feagan BG, Lichtenstein GR, et al. Maintenance infliximab for Crohn’s disease: The ACCENT I randomised trial. Lancet 2002; 359:1541-9. (Ref 317.) Hanauer SB, Stromberg U. Oral Pentasa in the treatment of active Crohn’s disease: A meta-analysis of double-blind, placebo-controlled trials. Clin Gastroenterol Hepatol 2004; 2:379-88. (Ref 228.) Hyams J, Crandall W, Kugathasan S, et al. Induction and maintenance infliximab therapy for the treatment of moderate-to-severe Crohn’s disease in children. Gastroenterology 2007; 132:863-73. (Ref 319.) Loftus CG, Loftus EV Jr, Harmsen WS, et al. Update on the incidence and prevalence of Crohn’s disease and ulcerative colitis in Olmsted County, Minnesota, 1940-2000. Inflamm Bowel Dis 2007; 13:254-61. (Ref 8.) Pearson DC, May GR, Fick G, Sutherland LR. Azathioprine for main­ taining remission of Crohn’s disease. Cochrane Database Syst Rev 2000:CD000067. (Ref 269.) Punati J, Markowitz J, Lerer T, et al. Effect of early immunomodulator use in moderate to severe pediatric Crohn disease. Inflamm Bowel Dis 2008; 14:949-54. (Ref 292.) Sandborn WJ, Colombel JF, Enns R, et al. Natalizumab induction and maintenance therapy for Crohn’s disease. N Engl J Med 2005; 353:1912-25. (Ref 71.) Sands BE, Anderson FH, Bernstein CN, et al. Infliximab maintenance therapy for fistulizing Crohn’s disease. N Engl J Med 2004; 350:87685. (Ref 318.) Schreiber S, Khaliq-Kareemi M, Lawrance IC, et al. Maintenance therapy with certolizumab pegol for Crohn’s disease. N Engl J Med 2007; 357:239-50. (Ref 326.) Seow CH, Benchimol EI, Griffiths AM, et al. Budesonide for induction of remission in Crohn’s disease. Cochrane Database Syst Rev 2008:CD000296. (Ref 260.) Full references for this chapter can be found on www.expertconsult.com.

1973

CHAPTE R

112 Ulcerative Colitis Mark T. Osterman and Gary R. Lichtenstein

CHAPTER OUTLINE Epidemiology  1975 Etiology and Pathogenesis  1977 Genetics  1977 Environmental Factors  1978 Immune Factors  1979 Psychogenic Factors  1980 Pathology  1981 Clinical Features  1982 Symptoms  1983 Signs  1983 Laboratory Findings  1984 Natural History and Prognosis  1984 Diagnosis  1985 Endoscopy  1985 Radiology  1987 Differential Diagnosis  1988 Crohn’s Disease  1988 Infection  1989 Other Causes  1990

Ulcerative colitis (UC) is a chronic idiopathic inflammatory disease of the gastrointestinal tract that affects the large bowel and is a major disorder under the broad group of conditions termed inflammatory bowel diseases (IBDs). Dr. Samuel Wilks is credited with being the first to describe UC in 1859 when he wrote on “idiopathic colitis” and recognized it as distinct from the then more common bacillary dysentery.1 He also reported the pathologic finding of dilated and thinned colon with severe pancolonic inflammation in a patient with this condition.2 In 1909, Hawkins described the chronic and relapsing nature of the disease course and the “stealthy hemorrhage” onset of distal disease, in which bleeding often occurred in the presence of constipation.3 In that same year, Sir Arthur Hurst gave a more complete description of UC, including its sigmoidoscopic appearances and differentiation from bacillary dysentery.4 The etiology of UC remained controversial, however, and an infectious or psychosomatic origin was considered its primary cause. The discovery of the double-helix structure of DNA by Watson and Crick in the 1950s paved the way for the era of genetic research, and in the last several decades, immunology has taken the central stage of research in attempting to unravel the pathogenesis of UC. At present, the precise etiology of UC still is unknown, but it is thought to be multifactorial, involving genetic, immunologic, and environmental factors. It also is now evident that patients with UC can have a broad spectrum of clinical presentations and extraintestinal manifestations. Although UC is not associated with an increased mor­ tality compared with the general population, it can have substantial morbidity and can lead to sizable direct

Assessment of Disease Activity  1991 Treatment  1992 Medical  1992 Algorithms for the Treatment of Ulcerative Colitis  2002 Surgical  2002 Specific Complications  2005 Toxic Megacolon  2005 Strictures  2005 Dysplasia and Colorectal Cancer  2005 Pouchitis  2008 Extraintestinal Manifestations  2010 Cutaneous/Oral  2010 Ophthalmologic  2011 Musculoskeletal  2011 Hepatobiliary  2012 Hematologic  2012 Others  2012

and indirect health care costs. Fortunately, significant advances have been made regarding management of UC and its associated complications, including medical therapies and surgical techniques.

EPIDEMIOLOGY The incidence and prevalence of UC vary with geographic location and ethnicity. Rigorous epidemiologic studies have been limited by several potential issues: Diagnosis of UC may be difficult due to its varied clinical manifestations and, in some regions, the common occurrence of infectious colitis that can mimic UC. Differences in health care systems also contribute to inaccurate estimation of cases, and reliable determination of the epidemiology of UC for a particular population might not be possible. With improved diagnostic techniques and increased awareness, UC is now recognized worldwide, although despite this increased recognition, most of the available epidemiologic data derive from population- or hospital-based studies conducted in North America and northern Europe. In general, there has been a distinct north-south gradient in risk. The areas with the highest rates of reported incidence and prevalence of UC include North America, England, northern Europe, and Australia (Tables 112-1 and 112-2). In North America, incidence rates range from 6.0 to 15.6 cases per 100,000 person-years, and the prevalence ranges from 38 to 246 cases per 100,000 persons.5 In Europe, incidence rates range from 1.5 to 20.3 cases per 100,000

1975

1976

Section X  Small and Large Intestine Table 112-1  Incidence of Ulcerative Colitis in Various Geographic Regions

REGION

PERIOD OF STUDY

North America Alberta 1981 Manitoba 1987-1996 California 1980-1981 Minnesota 1984-1993 Europe Scandinavia 1980-1999 Great Britain 1985-1994 Northern Europe 1988-1994 Southern Europe 1980-1994 Asia India 1999-2000 Japan 1991 Korea 1992-1994 Africa South Africa 1980-1984 Central and South America All countries 1987-1993

INCIDENCE (PER 100,000 PERSON-YEARS) 6.0 14.3-15.6 10.9 8.3 9.2-20.3 13.9 3.2-11.8 1.5-9.6 6.0 1.9 1.2 0.6-5.0 1.2-2.2

Adapted from Loftus EV. Clinical epidemiology of inflammatory bowel disease: Incidence, prevalence, and environmental influences. Gastroenterology 2004; 126:1504.

Table 112-2  Prevalence of Ulcerative Colitis in Various Geographic Regions REGION North America Alberta Manitoba Minnesota Europe Scandinavia Great Britain Northern Europe Southern Europe Asia Israel India Japan Korea Singapore

PERIOD OF STUDY 1981 1994 2001

PREVALENCE (PER 100,000 PERSONS) 37.5 169.5 246

1987 1995-1996 1984 1988-1992

161.2 122-243 24.8 21.4-121

1980-1985 1999 1991 1997 1985-1996

55.2-70.6 44.3 18.1 7.6 6.0

Adapted from Loftus EV. Clinical epidemiology of inflammatory bowel disease: Incidence, prevalence, and environmental influences. Gastroenterology 2004; 126:1504.

person-years, with prevalence of 21 to 243 cases per 100,000 persons.5 The disease initially had been considered more common in northern Europe, although studies now suggest that the incidence of UC in southern Europe is comparable to that in northern Europe.6 In contrast, studies have reported significantly lower incidence rates of 0.6 to 6 per 100,000 person-years in other parts of the world, including Asia, Africa, and Latin America.5 The overall incidence of UC has remained relatively stable since the 1980s; however, there appears to be geographic variation in this time frame. Although the incidence rates in North America, England, and Sweden have remained

unchanged or have declined, those in southern Europe have increased.7 These time trends contrast with those of Crohn’s disease, which has shown an increase in incidence across geographic regions. There appears to be a marked ethnic variation in the incidence of UC. One ethnic group with high incidence of this disease is the Jewish population. Incidence rates of UC among Jews have been shown to be several-fold higher than in the non-Jewish population across various geographic regions. In the United States, for example, the annual incidence of UC among Jews is 13 cases per 100,000 personyears compared with 3.8 per 100,000 among non-Jewish whites.7 Within Israel, Ashkenazi Jews have a higher incidence than do Sephardic Jews, but they have a lower incidence than in the United States and northern Europe, suggesting that environmental factors also play an etiologic role. UC traditionally has been considered extremely uncommon in minority populations, but recent studies have challenged this notion. Early studies reported that UC was rare in blacks. Most of these studies, however, were conducted in regions with limited black populations, and morerecent studies suggest an increasing incidence of UC among African Americans. By the late 1970s, incidence rates were comparable between whites and nonwhites in the United States.8,9 An increase in incidence of UC also has been observed among blacks in South Africa, although the incidence rate still remains lower than that for South African whites.10 In Asia, UC is generally is less common than in the Western countries. The prevalence and annual incidence rates for UC in Japan have remained relatively stable at about 5.5 cases per 100,000 persons and 0.36 to 0.5 cases per 100,000 persons, respectively11; this stability is in contrast with the rising incidence of Crohn’s disease in the Japanese population. Limited data also suggest similar findings in the Chinese and Korean populations.12,13 The prevalence of UC in India has been reported to be substantially lower than that among Europeans,14 although an accurate assessment of its epidemiology is hampered by several of the previously mentioned issues. In contrast to the limited data on indigenous South Asians, several studies have demonstrated that South Asian immigrants in England are more likely to have UC than are European natives.14-16 This changing epidemiology with a population emigrating from a lowrisk to a high-risk geographic region supports the concept of environmental influences on development of disease. UC can occur at any age, although diagnosis before the age of five years or after 75 years is uncommon. The peak incidence of UC occurs in the second and third decades of life. Studies have reported a second, smaller peak in the elderly, between the ages of 60 and 70 years. This second peak of disease incidence is less pronounced than that for Crohn’s disease. Most studies have not shown any gender difference in the occurrence of UC, and a male-to-female ratio of nearly 1 : 1 applies to all age groups. Certain lifestyle and socioeconomic factors have been associated with the development of UC. It is more common in industrialized than in less-developed countries and among urban than rural populations. Within a defined population, there may be a slightly higher incidence of disease among those of higher socioeconomic status. As mentioned before, studies of immigrants moving to high-risk geographic regions have shown increases in their incidence rates of UC compared with the incidence rate for the same ethnic groups living in their native countries. Together, these observations support the notion that environmental factors influence the development of UC.

Chapter 112  Ulcerative Colitis ETIOLOGY AND PATHOGENESIS The etiology of UC is currently unknown but is likely multifactorial. The currently held paradigm involves a complex interaction of three elements: genetic susceptibility, host immunity, and environmental factors. Dysregulation of the enteric immune response in genetically predisposed persons leads to the development of acute and chronic inflammation and the pathologic feature of mucosal damage. The specific inciting antigens for the inflammatory process have yet to be identified, but several sources have been suggested, including pathogenic and commensal microorganisms, metabolic byproducts of these agents, and normal epithelial structures.

GENETICS Family History

Genetic factors have been linked to the development of UC, supported largely by the observation that family history is one of the most important risk factors for developing the disease. A familial incidence of UC has been recognized for many years, and although figures vary widely in different studies, about 10% to 20% of patients have at least one other affected family member.17 Familial associations generally occur in first-degree relatives. The relative risk of the same disease in a sibling of a person with UC has been estimated to be between 7% and 17% based on North American and European studies. Parents, offspring, and seconddegree relatives appear to be at a lower risk for developing UC than are first-degree relatives. Data from the United States suggest a preponderance of parent-sibling combinations, but in the United Kingdom, the disease is shared more commonly by siblings. Indeed, the strongest evidence of a genetic influence for UC is derived from twin studies. In three large European twin pair studies, approximately 6% to 16% of monozygotic twin pairs had concordant UC compared with 0% to 5% of dizygotic twin pairs.18-20 These concordance rates are substantially lower than those for Crohn’s disease, suggesting that genetic determinants, although important, play a less-significant role for UC than for Crohn’s disease. No twin pair demonstrated both UC and Crohn’s disease, further supporting the genetic basis of these disorders. Familial association is greater in persons of Jewish descent, a heritage known to have a higher incidence of IBD. The lifetime risk of developing disease is three-fold higher among first-degree relatives of Jewish patients compared with relatives of non-Jewish patients.21 A similar increase in risk also has been observed in relatives of patients who had early onset of disease. This familial association contrasts with the low incidence of UC among spouses of patients with IBD in most series. Although reports of IBD in both spouses are rare, a study of 30 conjugal instances of IBD (17 of which were concordant for Crohn’s disease, 3 for UC, and 10 mixed) found that the majority of these couples developed IBD after cohabitation, thus suggesting a shared environmental exposure.22 For all affected first-degree relatives within a family, there is a high concordance for type of disease (UC vs. Crohn’s disease) and occurrence of extraintestinal manifestations.23 In fact, epidemiologic studies in families with multiple affected members have demonstrated disease-type concordance rates of 75% to 80%, the other 20% to 25% of families being mixed (i.e., having members with both UC and Crohn’s disease).24 The overlap in inheritance patterns among the mixed families suggests that a subset of genes associated with IBD confers susceptibility to both UC and

Crohn’s disease, whereas others are specific to one disease or the other. Numerous studies have examined the effect of family history on disease location. In UC, no consistent correlation has been observed in the studies that examined disease extent namely, left-sided colitis versus extensive colitis and family history of UC.25

Genetic Mutations

The inheritance of UC cannot be described by a simple mendelian genetics model. It is likely that multiple genes are involved and that different genes confer susceptibility, disease specificity, and phenotype. Linkage studies have suggested that there are susceptibility genes for UC on chromosomes 1, 2, 3, 5, 6, 7, 10, 12, and 17.26-29 The IBD2 locus on chromosome 12 appears to have strong linkage demonstrated in studies involving large numbers of families with UC.28 The NOD2/CARD15 gene mutations located on chromosome 16 associated with Crohn’s disease have not been associated with UC, although UC patients from families with a history of Crohn’s disease and UC might possess NOD2 variants.30 In contrast, the C3435T polymorphism for the human multidrug resistance 1 (MDR1) gene is linked to susceptibility for UC but not Crohn’s disease.31 The MDR1 gene product, P-glycoprotein, is highly expressed in intestinal epithelial cells and serves an important barrier function against xenobiotics. In contrast to NOD2/CARD15, the frequency of this polymorphism in patients with Crohn’s disease is similar to that in control subjects. Genome-wide association studies have looked for overlap of Crohn’s disease genes in patients with UC by using a nonsynonymous single nucleotide polymorphism (SNP) scan.32,33 The Crohn’s disease autophagy genes ATG16L1 on chromosome 2q37 (which encodes autophagy-related 16-like protein) and IRGM on chromosome 5q33 (which encodes immunity-related GTPase [guanosine triphosphatase] family, M), both of which are involved in bacterial processing and the protection of cells from various bacterial pathogens and their toxins (i.e., autophagy), are not seen in patients with UC.32 A number of Crohn’s disease loci or genes (or both) have been identified in UC, however, and include IL-23R on chromosome 1p31, which encodes the interleukin[IL]-23 receptor; chromosome 3p21, which encodes MST1 and other potential genes of interest; IL-12β on chromosome 5q33, which encodes the IL-12 receptor β1 subunit (also known as p40) that constitutes part of both the IL-23 and IL-12 receptors; NKX2-3 on chromosome 10q24, which encodes NK2 transcription factor related, locus 3; and chromosome 17q21, which encodes STAT3 and other potential genes of interest. Another genome-wide association study has observed a strong association between the gene encoding IL-23R and both Crohn’s disease and UC.33 Several polymorphisms of this gene have been identified, most notably the Arg381Gln polymorphism. Heterozygous carriage of the glutamine allele is associated with a three-fold decreased risk of Crohn’s disease and a more-modest reduction in the risk of UC in non-Jewish populations; this reduction is not seen in UC within the Jewish population. IL-23R is important because it plays a key role in the differentiation of a relatively newly discovered subset of T cells called Th17 cells (see later). There also are genes that appear to influence disease behavior independently of susceptibility genes. The best studied of these genes are the human leukocyte antigen (HLA) alleles. One allele of HLA-DR2 (DRB1*1502) appears to be involved in disease susceptibility in Japanese and Jewish populations. Several centers have reported an asso-

1977

1978

Section X  Small and Large Intestine ciation between severe disease and a rare allele of HLA-DR1 (DRB1*0103). In some studies, the HLA-DR3,DQ2 haplotype is associated with extensive colitis, especially in women. Among the Jewish population, the perinuclear antineutrophil cytoplasmic antibody (pANCA) is a marker for the DRB*1502 allele of HLA-DR2, but in non-Jewish whites, this antibody is associated with the HLA-DR3 DQ2-tumor necrosis factor (TNF)-α2 haplotype.

ENVIRONMENTAL FACTORS

It is now almost universally accepted that the pathogenesis of IBD is a result of continuous antigenic stimulation by commensal enteric bacteria, fungi, or viruses, which leads to chronic inflammation in genetically susceptible hosts who have defects in mucosal barrier function, microbial killing, or immunoregulation. Several infectious organisms, including mycobacteria and viruses, have been implicated in the pathogenesis of IBD. No specific infective organism, however, has been isolated consistently from patients with UC, and therefore it is unlikely that the disease is caused by a single common infectious agent. Numerous clinical and experimental observations have suggested involvement of intestinal bacterial microbiota in the pathogenesis of IBD. The most obvious observation perhaps is that Crohn’s disease and UC preferentially occur in regions of the bowel that contain the highest concentration of bacteria, namely, the terminal ileum and the colon, where bacterial concentrations approach 1012 organisms per gram of luminal contents. Interestingly, diverting the fecal stream in patients with Crohn’s disease can treat and even prevent disease, whereas reinfusion of ileostomy contents leads to new inflammatory changes within only one week.34 Other human data have shown that antibiotics are useful in the treatment or postoperative prevention of Crohn’s disease and pouchitis. Finally, probiotics (discussed later in this chapter) have been shown to have efficacy in the primary and secondary prevention of pouchitis. The most glaring evidence of the necessary role of bacteria in the pathogenesis of IBD from rodent data is that genetically susceptible mice or rats in a gnotobiotic (germ-free) environment do not have intestinal inflammation; however, these same rodents rapidly develop intestinal inflammation after bacterial colonization.35-38 Just as in humans, rodent gut inflammation can be treated and prevented with antibiotics and probiotics.39,40 With respect to the human gastrointestinal microbiome, much has been learned in the last few years. Complex mathematical models have estimated that the human gastrointestinal microbiome contains at least 1800 genera and between 15,000 and 36,000 species of bacteria41-43; three studies have identified more than 45,000 bacterial small-subunit (SSU) rRNA genes, a technique that captures at most only 50% of the predicted species-level biodiversity.41-43 Of the bacterial genes identified to date, almost all (more than 98%) can be grouped into four phyla: Firmicutes accounts for 64% of the total and includes the family Lachnospiraceae (e.g., Clostridium groups XIVa and IV) and the subgroup Bacillus (e.g., Streptococcaceae and Lactobacillales). Bacteroidetes account for 23% of the total, Proteobacteria account for 8% of the total and include the family Enterobacteriaceae (e.g., Escherichia coli), and Actinobacteria account for 3%.43 Four general mechanisms have been postulated to explain how components of the normal intestinal microbiome might initiate or contribute to the development of the chronic inflammatory state.44 First, microbes can induce intestinal inflammation, either by adhering to or invading intestinal epithelial cells, thereby causing downstream proinflammatory cytokine production or by producing enterotoxins.

Second, a breakdown in the balance between protective and harmful intestinal bacteria, termed dysbiosis, can lead to disease. Most studies comparing the intestinal microbiome in IBD with that in healthy controls show biodiversity in the IBD populations is decreased by 30% to 50%. One study found that this reduction in biodiversity was due to decreased concentrations of Firmicutes (specifically Lach­ nospiraceae) by 300-fold and Bacteroides by 50-fold.43 The loss of these organisms is important because they are known to produce short-chain fatty acids, such as butyrate, which nourish colonocytes. As a result of their decrease, the relative concentrations of Proteobacteria and Actinobacteria increase in IBD patients relative to controls, although quantitative PCR analysis showed that the absolute numbers of Enterobacteriaceae were not higher in IBD patients than in controls. Loss of protective bacteria, however, could set the stage for overgrowth of pathogenic bacteria. Another interesting finding in this study is that the microbiota of patients with Crohn’s disease and UC were similar to each other. In addition, the study did not identify any individual species that was particularly prevalent in grossly abnormal diseased tissue, and thus no active bacterial etiologic agent was suspected of causing or propagating disease. The third and fourth ways bacteria could play a role in the pathogenesis of IBD deal with the host itself. Genetic defects in host microbial killing or impaired mucosal barrier function can lead to immune hyper-responsiveness to intestinal bacteria, as the microbes have more exposure to epithelial cells and can trigger the production of high levels of proinflammatory cytokines. Finally, genetic defects in host immunoregulation can lead to a heightened immune response to even nonpathogenic bacteria, such as abnormal antigen processing or presentation, loss of tolerance, or overly aggressive T-cell responses. In addition to infectious agents, several other environmental factors have been proposed as contributing etiologic factors of UC. The best characterized environmental factor associated with UC is cigarette smoking. Numerous studies have consistently shown that UC is more common among nonsmokers than among current smokers, with the relative risk of UC in nonsmokers ranging from two to six45; this association is independent of genetic background and gender. Furthermore, there may be a dose-response relationship, with the disease more common in current light smokers than in heavy smokers. This relationship is consistent with observations that clinical improvement with nicotine therapy in patients with UC appears to be limited to those treated with high doses of nicotine and not those receiving lower doses. This risk of developing UC with smoking is particularly high for former smokers, especially within the first two years of smoking cessation. The rebound effect also is higher for former heavy smokers than for former light smokers.46 Smokers also appear to have reduced rates of hospitalization for UC and reduced rates of pouchitis following colectomy.47 Studies on the role of passive smoking in UC have yielded conflicting results. A recent metaanalysis of 10 studies examining this issue found no association between passive smoking and future development of UC.48 Several mechanisms have been postulated to account for the apparent protective effect of active smoking on UC. These include modulation of cellular and humoral immunity, changes in cytokine levels, increased generation of free oxygen radicals, and modification of eicosanoid-mediated inflammation. Smoking also might have an effect on mucus production by the colonic mucosa, and might alter colonic mucosal blood flow and intestinal motility. No single mechanism, however, can explain the clinical observation of the

Chapter 112  Ulcerative Colitis beneficial influence of smoking on UC and its adverse effect on Crohn’s disease (see Chapter 111). Other environmental risk factors that have been suggested to influence the development of UC include diet (wheat, maize, cow’s milk, refined sugar, fruits and vegetables, alcohol), oral contraceptives, food additives (silicon dioxide), toothpaste, and breast-feeding7; none, however, has been shown conclusively to be associated with UC. Studies have suggested an inverse relationship between appendectomy and the subsequent development of UC49,50; the mechanisms for this protective effect of appendectomy on UC are unknown. It is possible that removal of appendiceal-associated lymphoid tissues abrogates certain pathologic alterations in mucosal immune responses and therefore prevents the onset of UC.

IMMUNE FACTORS

The prevailing theory of the pathogenesis of UC emphasizes the role of the enteric immune response. The physiologic state of the intestine is one of constant low-grade inflammation in response to environmental stimuli such as bacterial products or endogenous factors. Breaches in this wellregulated mucosal immune system lead to the chronic, uncontrolled mucosal inflammation observed in UC. In this regard, immunologic mechanisms in the pathogenesis of UC involve both humoral and cell-mediated responses.

Humoral Immunity

Histologic examination of the inflamed colon indicates a marked increase in the number of plasma cells. This increase is not uniform among cells producing different classes of immunoglobulins. The largest proportional increase occurs in immunoglobulin (Ig)G synthesis, which has the highest pathogenic potential among antibody classes. The increase in IgG synthesis in UC is most pronounced in the IgG1 and IgG3 subclasses, in contrast to Crohn’s disease, in which the increase in IgG2 synthesis is more prominent.51,52 This disparity in the local IgG subclass response likely reflects differences in antigenic stimuli or host immunoregulatory responses between the two groups of IBD patients. The increased IgG synthesis in IBD may represent polyclonal stimulation; patients with UC often have circulating antibodies to dietary, bacterial, and self antigens that are mostly of the IgG isotype, usually the IgG1 subclass. Many of these antibodies are thought to be epiphenomena because the serum antibody titers do not correlate with clinical features. Nevertheless, the known cross-reaction between enterobacterial antigens and colonic epithelial epitopes may be an important triggering event, even though, later in the course of the disease, the serum antibody titer to either the bacterial or the colonic antigen may be unimportant. The concept that UC is an autoimmune disease is supported by its increased association with other autoimmune disorders, including thyroid disease, diabetes mellitus, and pernicious anemia.53 Patients with UC have varying levels of autoantibodies to lymphocytes, ribonucleic acid, smooth muscle, gastric parietal cell, and thyroid; these are specific for neither tissue nor disease. Antibodies to epithelial cell-associated components, which specifically recognize intestinal antigen, also have been described.53 The bestcharacterized intestinal autoantigen is a 40-kDa epithelial antigen found in normal colonic epithelium.54 This autoantigen is recognized by IgG eluted from the inflamed colonic mucosa of patients with UC and is a component of the tropomyosin family of cytoskeletal proteins.55 The antibody response to this 40-kDa protein appears to be unique to UC and is not found in Crohn’s disease nor in other inflammatory conditions. This autoantigen shares an epitope with

antigens found in the skin, bile duct, eyes, and joints, sites often involved in the extraintestinal manifestations of UC. The precise pathogenic significance of this autoantibody in UC, however, remains unclear at present. An autoantibody that has received significant attention in UC patients is pANCA.56 This autoantibody is present in 60% to 85% of patients with UC.57,58 It is synthesized within the lamina propria and is of the IgG1 subclass. The antigen to which the pANCA is directed has not yet been determined with certainty, and a variety of putative antigens have been proposed, including nuclear histone and nonhistone proteins. The most recent evidence suggests that the antigen is a 50-kd nuclear envelope protein that is specific to myeloid cells.59 Just as with other autoantibodies found in patients with UC, the pathogenic relevance of pANCA in this disorder is unknown. In fact, the prevailing thought is that pANCA has no pathogenic role in UC but that it might serve as a potential marker of susceptibility and genetically distinct subsets of UC. The level of pANCA titer does not correlate with disease activity, but it might decline in patients with longstanding remission or in patients who have had colectomy at least 10 years previously. Studies have suggested that pANCA may be associated with a more-aggressive disease course60 and with the development of pouchitis after ileal pouch-anal anastomosis (IPAA) in patients with UC.61,62 Also of interest is that Crohn’s disease patients with colonic disease have higher rates of pANCA seroreactivity than those without colonic disease, suggesting a possible UC-like Crohn’s disease phenotype.63 Another type of autoantibody commonly seen in IBD, especially in Crohn’s disease, is that against bacterial antigens. Antibodies to bacterial antigens seen in UC include anti-CBir1 and anti-OmpC. Anti-Cbir1 is an antibody to flagellin from Clostridium species; it is found in about 6% of UC patients64 and also appears to be associated with the development of pouchitis.65 Anti-CBir1 also is found in 50% of patients with Crohn’s disease, in which it is associated with more complicated disease. 64 Anti-OmpC (outer membrane porin C of E. coli) is seen more often in UC patients who have a mixed family history of Crohn’s disease and UC rather than those with a family history of only UC.66

Cellular Immunity

Immune dysregulation in UC also involves cell-mediated immunity. Cell-mediated immunity consists of two com­ ponents, innate immunity and adaptive immunity. The innate immune system, which involves largely monocytemacrophages and dendritic cells, is nonspecific and untrained and acts as the first line of defense against foreign antigens, particularly bacterial antigens. Bacteria prompt immune responses largely through pattern-recognition receptors (PRRs), which include the 11 Toll-like receptors (TLRs) and 23 nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) that have been identified to date. Activation of the TLRs and NLRs results in downstream activation of nuclear factor-κB (NF-κB), which then stimulates the transcription of genes coding for various proinflammatory cytokines (including TNF, IL-1, IL-6, and IL-8), chemokines, adhesion molecules, and costimulatory molecules. In addition, activation of NF-κB stimulates the maturation of dendritic cells, which are involved in antigen presentation. Defects in any of the PRR pathways can lead to abnormal bacterial processing and possibly IBD.67 The adaptive immune system, which is governed by T cells and B cells, is specific. Mucosal lymphocytes often are divided into two groups based on location: lamina propria

1979

1980

Section X  Small and Large Intestine lymphocytes and intraepithelial lymphocytes (IELs). Lamina propria lymphocytes express surface adhesion molecules, α4β7, that provide a homing signal for peripheral immune cells to the mucosal sites.68 Most investigators have found a similar distribution of T-cell subsets (CD4+, CD8+) within the lamina propria in patients with UC compared with that in controls.69 Also, the absolute number of IELs is normal or reduced in UC, most of which are CD8+ cell. The function of IELs has not been well characterized, but it has been suggested that they are cytotoxic and perhaps active in suppressing local immune response. Regardless of their functional status, mucosal T cells within the lamina propria and epithelium, as well as peripheral blood T cells, display a variety of activation markers, suggesting an activated memory phenotype.70 Studies have suggested that the T-cell receptor repertoire is altered in active IBD.71 Although T-cell–mediated immunity has attracted the most attention in the pathogenesis of UC, nonspecific cellular immunity also is altered. In patients with active disease, there is an overproduction of circulating monocytes as well as mucosal macrophages.72 The inflamed mucosa of patients with UC also exhibits infiltration of substantial numbers of granulocytes.

Epithelial Cells

Intestinal epithelial cells serve barrier functions and play a role in enteric immunity. Colonocytes express class II major histocompatibility complex (MHC) antigens and can function as antigen-presenting cells.73 In addition, they also express cytokine receptors, secrete various cytokines and chemokines, and express leukocyte adhesion molecules.74-77 Thus, abnormalities in colonic epithelial cells can contribute to the development of UC. Patients with UC have an increased turnover rate of colonic epithelium78 and other abnormalities of epithelial cells including reduced metabolism of short-chain fatty acids, especially butyrate, abnormal membrane permeability,79 and altered composition of glycoprotein mucus produced by the colonic epithelium.80 Specifically, the mucus layer in UC appears to be thinner than normal.81 These and other abnormalities can lead to the finding of increased numbers of adherent bacteria, in both the mucus layer and even at the epithelial surface, in patients with UC.82-84 The role of epithelial cells in the pathogenesis of IBD is supported further by animal models of colitis produced by disruption of colonic epithelium.85

Consequences of Immune Activation

Activation of macrophages, lymphocytes, and colonic epithelial cells leads to the release of a variety of cytokines and mediators that further amplify the immune and inflammatory response of UC and result in tissue damage. Based on the cytokines they produce, proinflammatory CD4+ T cells have been divided into three major immune phenotypes: T helper 1 (Th1), T helper 2 (Th2), and a newly recognized subset called T helper 17 (Th17). The Th1 response is characterized by cell-mediated immunity and is associated with the production of interleukin (IL)-2 and interferon (IFN)-γ. The differentiation of T cells along a Th1 pathway is stimulated by IL-12 generated in response to exposure to infectious agents. The Th2 response is characterized by the production of cytokines IL-4, IL-5, IL-10, and IL-13, which amplify the humoral immune response. Th1 and Th2 subsets reciprocally downregulate each other through cytokine production.86 Th1 and Th2 pathways can be regulated by unique regulatory T cell (Treg) subsets that produce IL-10 and transforming growth factor (TGF)-β and down-regulate inflammation.87

Historically, the oversimplified view of adaptive immunity in IBD is that Crohn’s disease is mediated by Th1 cells whereas UC is mediated by Th2 cells; there is considerable evidence that the story is much more complex. Macrophages in the inflamed colon of patients with active UC synthesize IL-1β, TNF, and IL-6, whereas lamina propria T cells probably produce IL-2 and IFN-γ. This immune response can be up-regulated further by presentation of antigen to CD4+ lymphocytes by colonic epithelial cells that express HLA class II antigens.73 Studies have implicated a specialized type of T cell, the natural killer (NK) T cell, which seems to mediate the Th2 response in UC.88,89 These NK T cells, which are not classical NK T cells in that they do not express the typical NK T-cell receptors seen with classical NK T cells, secrete large amounts of IL-5 and IL-13 and are actually cytotoxic for intestinal epithelial cells. A novel T-cell–mediated inflammatory pathway, which appears to be involved in the pathogenesis of both Crohn’s disease and UC, has been discovered and centers on the Th17 cell lineage. Th17 cells have been shown to produce a variety of cytokines, most notably IL-6 and IL-17. IL-17 is a potent proinflammatory cytokine that not only facilitates T-cell activation but also stimulates a variety of cells, including fibroblasts, macrophages, epithelial cells, and endothelial cells, to produce an array of proinflammatory cytokines, including IL-1, IL-6, TNF-α, and chemokines.90 Th17 cell development is inhibited by Th1 and Th2 cells but is promoted by IL-6, TGF-β, IL-21, and IL-23R.91 IL-23R, which is highly expressed by activated Th17 cells, also is expressed by NK cells, NK T cells, other CD4+ T cells, and CD8+ T cells.29 The interaction of IL-23 with its receptor has been shown to have a central role in the development of inflammation in various mouse models of colitis.92,93 Antibodies to IL-23 thus represent another potential therapeutic strategy for the treatment of IBD. Release of these various cytokines from the T-cell inflammatory pathways also might lead to other abnormalities seen in UC, such as increased epithelial cell permeability and collagen synthesis. Alteration of endothelium by a variety of cytokines can result in local ischemia. Increased expression of endothelial adhesion molecules in response to inflammatory mediators recruits circulating granulocytes and monocytes to the inflamed tissues, thus further perpetuating the inflammatory response. Elevated cytokine levels within the mucosa also stimulate the release of metalloproteinase from fibroblasts with subsequent matrix degradation. Mucosal concentrations of many mediators have been shown to be elevated in patients with active UC, including leukotrienes, thromboxane, platelet-activating factor, nitric oxide, and reactive oxygen metabolites.94 These mediators, which are mostly released from active macrophages and neutrophils, contribute to inflammation and mucosal injury and alter epithelial cell permeability, thereby further contributing to diarrhea. Diarrhea in UC also is caused by complement activation and the release of kinins and other inflammatory mediators from mast cells and eosinophils (see Chapter 2).

PSYCHOGENIC FACTORS

Psychosomatic factors first were implicated in the pathogenesis of UC in the 1930s,95 but there is no good direct evidence to support this concept. Since the introduction of glucocorticoids for the treatment of patients with UC and the focus on immunologic aspects of the pathogenesis of the disease in the 1950s, this previously widely held notion has diminished in popularity. Experimental studies have helped identify mechanisms of the proinflammatory potential of stress in animal models

Chapter 112  Ulcerative Colitis of colitis.95 When rats are exposed to stress before proinflammatory stimuli are introduced, the severity of colonic inflammation is increased. This particular response has been shown not to be mediated by either vasopressin or corticotropin-releasing factor. In addition, stress has been shown to directly increase intestinal permeability in rats, an action mediated by cholinergic nerves, and to potentiate intestinal inflammation in this particular situation. There are indeed studies reporting that psychosocial stress increases the risk of relapse in patients with quiescent UC.96,97 Conversely, many of the psychological features observed in patients with UC are likely secondary to this chronic disease process, a phenomenon physicians must be aware of when managing these patients.

PATHOLOGY At the time of initial presentation, approximately 45% of patients with UC have disease limited to the rectosigmoid, 35% have disease extending beyond the sigmoid but not involving the entire colon, and 20% of patients have pancolitis.98 The disease typically is most severe distally and progressively less severe more proximally. In contrast to Crohn’s disease, continuous and symmetrical involvement is the hallmark of UC (Fig. 112-1), with a sharp tran­sition between diseased and uninvolved segments of the colon. There are a few exceptions to this general rule. First, medical therapy can result in areas of sparing. For example, topical enema therapy can lead to near-complete mucosal healing in the rectum and distal sigmoid colon. Second, up to 75% of patients with left-sided UC have periappendiceal inflammation in the colon and patchy inflammation in the cecum,99 resembling the skip pattern characteristic of Crohn’s disease. These patterns of rectal sparing and skip lesions can lead to a misdiagnosis of Crohn’s disease.

Figure 112-1.  Total colectomy specimen from a patient with ulcerative colitis. The colon shows diffuse mucosal inflammation that extends proximally from the rectum without interruption to the transverse colon. The mucosal pattern in the terminal ileum and cecum (arrow) is normal. The distal mucosa is erythematous and friable, with many ulcers and erosions. (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine, 1996.)

Macroscopically, the mucosa in UC appears hyperemic, edematous, and granular in mild disease. As disease progresses, the mucosa becomes hemorrhagic, with visible punctate ulcers. These ulcers can enlarge and extend into the lamina propria. They often are irregular in shape with overhanging edges or may be linear along the line of the teniae coli. Epithelial regeneration with recurrent attacks results in the formation of pseudopolyps, which is typical of long-standing UC but which also may be seen in acute disease (Fig. 112-2). Another characteristic appearance of long-standing disease is atrophic and featureless colonic mucosa, associated with shortening and narrowing of the colon. Patients with severe disease can develop acute dilatation of the colon, also characterized by thin bowel wall and grossly ulcerated mucosa with only small fragments or islands of mucosa remaining. With perforation of the colon, a fibrinopurulent exudate may be seen on the serosal surface of the bowel. Microscopically, the early stage of UC is marked by edema of the lamina propria and congestion of capillaries and venules, often with extravasation of red blood cells. This is followed by an acute inflammatory cell infiltrate of neutrophils, lymphocytes, plasma cells, and macrophages, often accompanied by increased numbers of eosinophils and mast cells. Neutrophilic infiltration of colonic crypts gives rise to cryptitis and ultimately to crypt abscesses with neutrophilic accumulations in crypt lumens. This migration of neutrophils from the circulation into the lamina propria occurs in response to a variety of chemoattractants, including chemotactic peptides of colonic bacteria, IL-8, activated complement, platelet-activating factor, and leukotriene B4. The cryptitis is associated with discharge of mucus from goblet cells and increased epithelial cell turnover. Thus, the acute inflammatory infiltration results in the characteristic histopathology of goblet cell mucin depletion, formation of exudates, and epithelial cell necrosis. None of these histologic findings, however, is specific for UC. Inflammation in UC characteristically is confined to the mucosa, in contrast to the transmural involvement of Crohn’s disease. The inflammatory changes typically end at

Figure 112-2.  Surgical specimen of resected colon from a patient with severe ulcerative colitis showing numerous inflammatory polyps (pseudopolyposis). Pseudopolyps are most common in ulcerative colitis but also may be seen in Crohn’s disease, ischemia, and other ulcerative conditions of the colon. These blunt or finger-like lesions develop as byproducts of ulcers that penetrate into the submucosa, leaving islands of adjacent regenerative mucosa. Although the intervening areas of colonic mucosa are ulcerated, pseudopolyps can persist even when inflammation has abated and the mucosa has healed. (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine, 1996.)

1981

1982

Section X  Small and Large Intestine the luminal aspect of the muscularis mucosa. With increasing inflammation, however, the surface epithelial cells become flattened, eventually ulcerate, and can become undermined if the ulcers are deep. At this stage of the disease, some inflammation and vascular congestion may be present in the submucosa, and ulceration can extend into the muscularis mucosa. This deeper involvement may be confused with Crohn’s disease, but it usually pre­ sents diffusely rather than with the segmental fissuring pattern of transmural inflammation that characterizes Crohn’s disease. During the healing phase of UC, the inflammatory infiltrate subsides and epithelial regeneration takes place. Epithelial cells undergoing regenerative changes become cuboidal with eccentric, large nuclei, and prominent nucleoli. These features may be confused with dysplasia. Thus, a diagnosis of dysplasia in UC should be made with caution in the presence of acute inflammation. Accordingly, surveillance colonoscopy (see “Dysplasia and Colorectal Cancer”) should be performed during a period of remission. A classic histologic feature of chronic quiescent UC is crypt architectural distortion or actual dropout of glands (Fig. 112-3). Architectural changes include branching or bifid glands, wide separation among glands, and shortened glands that do not extend down to the muscularis mucosa. Architectural alteration is a prominent feature of chronic quiescent UC, but the histologic abnormalities can revert to normal after mild flares early in the course of disease. Another characteristic feature of chronic quiescent UC is Paneth cell metaplasia, with Paneth cells located distal to the hepatic flexure, where they normally are absent. Other

nonspecific chronic changes seen in UC include neuronal hypertrophy and fibromuscular hyperplasia of the muscularis mucosa. Varying degrees of acute or chronic inflammation of the lamina propria may be present in chronic quiescent disease. A thin band of predominantly lymphocytic inflammation occasionally may be seen deep to the muscularis mucosa, presenting diagnostic challenges. Most of these pathologic findings are not specific for UC. Features that reflect chronicity and thus argue against a diagnosis of infectious or acute self-limited colitis include distorted crypt architecture, crypt atrophy, increased intercrypt spacing to fewer than six crypts per millimeter, an irregular mucosal surface, basal lymphoid aggregates, and a chronic inflammatory infiltrate.100,101 The histologic severity of inflammation does not necessarily correlate with clinical disease activity in patients with UC, because patients may be relatively symptom free although histology reveals significant inflammation.

CLINICAL FEATURES Patients with UC can present with a variety of symptoms. Common symptoms include diarrhea, rectal bleeding, passage of mucus, tenesmus, urgency, and abdominal pain. In more-severe cases, fever and weight loss may be prominent. The symptom complex tends to differ according to the extent of disease.102 Patients with proctitis often have local symptoms of tenesmus, urgency, mucus, and bleeding, whereas patients with extensive colitis usually have more diarrhea, weight loss, fever, clinically significant blood loss,

A

C

B

Figure 112-3.  Photomicrographs of a colonic biopsy specimen showing the histology of ulcerative colitis. A, Diffuse chronic inflammation of the lamina propria and crypt distortion are present. These features are important in differentiating ulcerative colitis from acute self-limited colitis. B, The base of a single distorted colonic crypt. There are many plasma cells between the crypt and the muscularis mucosae, another important finding that helps differentiate acute from chronic colitis. C, A single crypt abscess. The bottom of this distorted crypt has been destroyed by an aggregate of polymorphonuclear neutrophils. This finding is not specific for ulcerative colitis and may be seen in Crohn’s disease and other types of colitis. (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine, 1996.)

Chapter 112  Ulcerative Colitis and abdominal pain. In general, the severity of the symptoms correlates with the severity of the disease; however, active disease may be found at colonoscopy in patients who are otherwise asymptomatic. Additionally, patients with known UC can have severe symptoms that are not necessarily due to UC, such as those caused by bacterial (e.g., Clos­ tridium difficile) or viral (e.g., cytomegalovirus) infections or a host of other similar disorders. The onset of UC typically is slow and insidious. Symptoms usually have been present for weeks or months by the time the typical patient seeks medical attention. The median interval between the onset of symptoms and diagnosis of UC is approximately nine months.103 Some patients with UC present much more acutely, with symptoms mimicking acute infectious colitis. Indeed, it is not uncommon to find a patient whose UC began after a documented gastrointestinal infection, such as Salmonella or C. difficile. This observation raises the question whether the infection revealed preexisting but silent disease or whether it was actually the initiating factor.

SYMPTOMS Rectal Bleeding

Rectal bleeding is common in UC, its characteristics determined by the distribution of disease. Patients with proctitis usually complain of passing fresh blood, either separately from the stool or streaked on the surface of a normal or hard stool.104 This symptom often is mistaken for bleeding from hemorrhoids. In contrast to hemorrhoidal bleeding, however, patients with ulcerative proctitis often pass a mixture of blood and mucus and might even be incontinent. Patients with proctitis also often complain of the frequent and urgent need to defecate, only to pass small quantities of blood and mucus without fecal matter. When the disease extends proximal to the rectum, blood usually is mixed with stool or there may be grossly bloody diarrhea. When disease activity is severe, patients typically pass liquid stool containing blood, pus, and fecal matter. This stool often is likened to anchovy sauce, and some patients with this symptom do not actually recognize that they are passing blood. Unless the patient has severe disease, passage of blood clots is unusual and suggests other diagnoses such as a tumor. Active UC that is sufficient to cause diarrhea almost always is associated with macroscopically evident blood. The diagnosis needs to be questioned if visible blood is absent.

Diarrhea

Diarrhea is common but not always present in patients with UC. Up to 30% of patients with proctitis or proctosigmoiditis complain of constipation and hard stools.104 Most patients with active disease complain of frequent passage of loose or liquid stools and may have nocturnal diarrhea. Fecal urgency, a sensation of incomplete fecal evacuation, and fecal incontinence also are common, especially when the rectum is severely inflamed. Diarrhea in this setting often is accompanied by passage of large quantities of mucus, blood, and pus. The pathophysiology of diarrhea in UC involves several mechanisms, but failure to absorb salt and water is the predominant factor105 and results from reduced Na+,K+ATPase (adenosine triphosphatase) pump activity, increased mucosal permeability, and altered membrane phospholipids. High mucosal concentrations of lipid inflammatory mediators, which are detected in UC, have been shown to stimulate chloride secretion in normal colon, and it is possible that these mediators also contribute to diarrhea by

increasing mucosal permeability. Urgency and tenesmus, which are common symptoms when the rectum is inflamed, are caused by decreased rectal compliance and loss of the reservoir capacity of the inflamed rectum.106 With severe inflammation, the urgency can be sufficiently acute to cause incontinence. Colonic motility is altered by inflammation, and there is rapid transit through the inflamed colon. With left-sided disease, distal colonic transit is rapid, but there is actual slowing of proximal transit,107 which might help explain the constipation that is commonly seen in patients with distal colitis. Prolonged transit in the small intestine also occurs in the presence of active colonic inflammation.107

Abdominal Pain

Many patients with UC complain of abdominal pain with active disease, although pain generally is not a prominent symptom unless disease activity is severe. Patients can experience vague lower abdominal discomfort, an ache in the left iliac fossa, or intermittent abdominal cramping that precedes bowel movements and often persists transiently after defecation. Severe cramping and abdominal pain can occur in association with severe attacks of the disease. The cause of the pain is unclear but might relate to increased tension within the inflamed colonic wall during muscular contraction. Patients with active proctitis also often complain of tenesmus and urgency associated with painful straining and passage of mucus and blood with only scanty stools.

Others

Disease of moderate or severe activity often may be associated with systemic symptoms. Patients can develop anorexia and nausea and, in severe attacks, might actually vomit. These symptoms, as well as protein loss through inflamed mucosa, hypercatabolism, and down-regulation of albumin synthesis caused by the inflammation, account for weight loss and hypoalbuminemia that may be profound. Fever, an added catabolic factor, usually accompanies severe attacks but is typically moderate. Patients also might complain of symptoms from anemia and hypoalbuminemia, including fatigue, dyspnea, and peripheral edema. Patients can present with extraintestinal manifestations, including acute arthropathy, episcleritis, and erythema nodosum, that typically parallel the activity of colitis.

SIGNS

Patients with mild or even moderately severe disease exhibit few abnormal physical signs. These patients are usually well nourished and well appearing and show no signs of chronic disease. Caution should be exercised because these patients can appear deceptively well. Weight always should be recorded and, for children and adolescents, both height and weight should be plotted on developmental growth charts. The affected portion of the colon may be tender on abdominal palpation, but tenderness usually is mild and not associated with rebound or guarding. Bowel sounds are normal. Digital rectal examination also is often normal, but the rectal mucosa might feel velvety and edematous; the anal canal may be tender; and blood may be seen on withdrawal of the examining finger. Patients with severe attacks also might appear well, but most are ill with tachycardia, fever, orthostasis, and weight loss. The abdomen typically is soft, with only mild tenderness over the diseased segment. Abdominal tenderness may become diffuse and moderate with more severe disease. Bowel sounds may be normal or hyperactive but diminish with disease progression. In fulminant colitis, the abdomen

1983

Section X  Small and Large Intestine often becomes distended and firm, with absent bowel sounds and signs of peritoneal inflammation. There may be aphthoid ulceration of the oral mucosa. Clubbing of the fingernails is a manifestation of chronic disease. Peripheral edema can occur secondary to hypoalbuminemia. Minor perianal disease may be present but is never as severe as is seen in patients with Crohn’s disease. Signs of extraintestinal manifestations also may be present.

LABORATORY FINDINGS

Laboratory findings in UC are nonspecific and reflect the severity of the underlying disease. Patients with active proctitis and proctosigmoiditis often have normal laboratory test results. Patients with limited distal disease often pass visible blood in the stool, but the amount of blood loss typically is small and anemia, if present, is mild. Patients with active extensive disease or severe distal disease can demonstrate laboratory abnormalities. Hematologic changes, including anemia, leukocytosis, and thrombocytosis, reflect active disease. In contrast, patients with quiescent UC typically manifest no laboratory abnormalities. Iron deficiency anemia may be present because of chronic blood loss. Anemia also may be present secondary to bone marrow suppression resulting from chronic inflammation or medications, including azathioprine, 6-mercaptopurine (6-MP), and sulfasalazine. Mild or moderate attacks rarely are associated with any biochemical disturbance. Hypokalemia, metabolic alkalosis, and elevated serum levels of blood urea nitrogen and creatinine may be present in severe flares of UC, reflecting volume depletion. Hypoalbuminemia may be seen with acute and chronic disease. Minor elevations in serum levels of aspartate aminotransferase or alkaline phosphatase also are commonly associated with severe disease, but these changes are transient and return to normal when the disease enters remission; these abnormalities probably reflect a combination of fatty liver, sepsis, and poor nutrition. Persistently elevated liver biochemical tests, especially serum alkaline phosphatase, are seen in about 3% of patients with UC and should lead to further investigation, particularly to exclude primary sclerosing cholangitis (PSC) (see Chapter 68). Serum inflammatory markers including erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) may be elevated in active disease. These abnormalities are typically absent or minimal in patients with mildly to moderately active disease. Elevation in these inflammatory parameters is neither sensitive nor specific for UC; measuring them, however, may be useful in clinical practice to assess disease activity in individual patients, particularly if these values are normal during periods of inactive disease. For following clinical changes, CRP is more sensitive than ESR because of the shorter half-life of CRP.

NATURAL HISTORY AND PROGNOSIS Most (80%) patients with UC have a disease course characterized by intermittent flares interposed between variable periods of remission. The duration of relapse-free periods varies greatly from patient to patient. More than 50% of patients present with mild disease at their first attack, and 6% to 19% of patients have severe disease at presentation.108,109 Following the initial flare, 40% to 65% of patients have an intermittent course, and 5% to 10% of patients have a chronic continuous course.110,111 Up to 10% of patients have a severe first attack ultimately requiring colectomy.107

100

80

Percent

1984

60

40

20

0 0

2

4

6

8

10 12 14 16 18 20 22 24

Years after diagnosis of ulcerative colitis Remission Active Colectomy Figure 112-4.  Percentage of patients with ulcerative colitis who were in remission, had active disease, or have had colectomy each year after diagnosis. After a few years, the fraction of patients in remission remains relatively constant, with approximately 50% of all patients in remission at any time point during follow-up. The fraction of patients with active disease gradually decreases to about 30%, and approximately 20% of patients undergo colectomy within 25 years after diagnosis. (Adapted from Langholz E, Munkholm P, Davidsen M, et al. Course of ulcerative colitis: Analysis of changes in disease activity over years. Gastroenterology 1994; 107:3.)

In population-based studies, the proportion of patients with active disease remains relatively constant over years, with approximately 50% of all patients being in remission at any time point during follow-up (Fig. 112-4).98,110 Twentyfive years after the diagnosis of UC, 90% of patients still have a relapsing course (Fig. 112-5)97; however, disease activity in the preceding years predicts the subsequent chance of disease activity. The probability of remaining in remission for one year after a relapse has been estimated at 30%. After being in remission for one year, the risk of relapse decreases to 20% for the following year. Few patients (1%) with UC have only one attack followed by a relapsefree course,110 and they likely have misdiagnosed infectious colitis. Factors influencing disease relapse and remission include bacterial and viral infections, the use of nonsteroidal antiinflammatory drugs (NSAIDs) and antibiotics, smoking, seasonality, and psychosocial stress. Both the severity and extent of disease are important prognostic factors for the first attack of UC. In general, patients with disease that is limited to the distal colon do better than those with extensive colitis. UC diagnosed in the elderly generally has been thought to manifest with more-severe initial attacks, but this pattern has not been observed consistently. No one factor has been consistently identified as predicting future disease activity. In patients initially presenting with proctitis or proctosigmoiditis, disease extension occurs in approximately 10% to 30% of patients at 10 years after diagnosis.108,112,113 Less commonly, extensive colitis regresses over time with treatment. Colectomy rates vary in different studies, in part related to the different proportions of patients with extensive versus limited disease, but rates seem to be highest in the first year of diagnosis. In one Scandinavian study, the colectomy rate

Chapter 112  Ulcerative Colitis 100

features, endoscopic appearances, and histologic findings. Stool cultures should be obtained to exclude infection with routine bacterial pathogenic organisms; assay for toxins A and B of C. difficile, and examinations for ova and parasites also should be performed. Infection with E. coli O157:H7 should be considered and requires special stool cultures (or molecular probes). Similarly, special cultures for gonococcus or Chlamydia may be necessary in selected cases. In immunosuppressed patients, the possibility of opportunistic infection of the colon must be excluded (see “Differential Diagnosis”). The diagnosis of UC should be questioned if there is only a single episode of acute illness or if the histopathology findings are nonspecific and lack signs of chronicity.

Relapsing course

Probability (%)

80

60

40 Single flare

20

Continuous activity 0 0

2

4

6

8

10 12

14

16 18 20

22 24

Years after diagnosis of ulcerative colitis Figure 112-5.  Cumulative probabilities of various disease courses after a diagnosis of ulcerative colitis. At 25 years after the diagnosis, 90% of patients had a relapsing course with remission and relapses. The cumulative probability of a completely relapse-free course (single flare) decreased rapidly with time to approximately 10% after 25 years. Similarly, the cumulative probability of a continuously active course was low: 1% after 5 years and 0.1% after 25 years. (Adapted from Langholz E, Munkholm P, Davidsen M, et al. Course of ulcerative colitis: Analysis of changes in disease activity over years. Gastroenterology 1994; 107:3.)

was 10% within the first year of diagnosis, 3% in the second year, and approximately 1% per year thereafter.110 In general, the overall colectomy rate is 24% at 10 years and 30% at 25 years (see Fig. 112-4).98 The probability of colectomy was higher in patients with extensive colitis. A more-recent European study reported a lower colectomy rate of 9% at 10 years, with a four-fold higher risk of colectomy in patients with extensive colitis.114 Despite the burden of a chronic illness, more than 90% of patients with UC are able to maintain capacity for work after 10 years of disease, and data suggest that the overall quality of life is not impaired significantly, including marital issues, physical activities, and social function.98,109,115 The disease can affect the quality of life to some degree during acute flares, however, and even during periods of remission, patients might remain anxious for fear of relapse and alter their lifestyle accordingly. Despite the morbidity of UC, mortality associated with the disease has dropped dramatically since the late 1950s and 1960s. The mortality rate for a severe attack of UC was approximately 35% before the introduction of glucocorticoid therapy and now is less than 2%. Long-term survival does not differ significantly from that expected for agematched controls, even with the risk of colorectal cancer that attends long-standing colitis. It is now generally believed that patients with UC have life expectancies comparable to those of the general population, although studies have reached conflicting conclusions.108,111,116-119 Mortality risk is greatest in the elderly and in those with extensive colitis, mostly related to postoperative complications within the first two years of disease and to comorbidities.

DIAGNOSIS Currently, there is no single test that allows the diagnosis of UC with acceptable sensitivity and specificity. Thus, diagnosis relies on a combination of compatible clinical

ENDOSCOPY

The diagnosis of UC can be strongly suggested by sigmoidoscopy in most cases. In patients presenting with their first attack of UC, sigmoidoscopy with biopsies usually is sufficient to confirm the diagnosis, thereby allowing initiation of therapy. In patients with active flares, sigmoidoscopy is best performed in unprepared bowel so the earliest signs of UC can be detected without the hyperemia that is often present because of preparative enemas. Colonoscopy is not recommended in patients with severely active disease for fear of perforation; care must be taken to avoid excessive distention. After active disease has been controlled in a patient with newly diagnosed UC, colonoscopy should be performed to establish the extent of the disease and to exclude Crohn’s disease or other disease states that can complicate UC. Multiple biopsy specimens should be taken from throughout the colon to map the histologic extent of disease and to confirm the diagnosis if there is concern about Crohn’s disease.120 In addition, intubation and biopsy of the terminal ileum should be attempted to exclude the presence of Crohn’s disease or other disease states that can mimic IBD. In patients with an established diagnosis of UC who present with a typical flare, sigmoidoscopy usually is not necessary, although it may be indicated for the rapid diagnosis of pseudomembranous colitis. Sigmoidoscopy combined with histologic evaluation, however, may be useful for assessing disease severity, particularly when therapeutic response is in question. Colonoscopy may be similarly useful, especially in patients whose symptoms seem out of proportion to the known extent of disease. Additionally, colonoscopy is essential for colorectal cancer surveillance (see below). The hallmark of UC is symmetrical and continuous inflammation that begins in the rectum and extends proximally without interruption for the entire extent of disease. The earliest endoscopic sign of UC is a decrease or loss of the normal vascular pattern, with mucosal erythema and edema (Fig. 112-6); distortion or loss of vascular markings may be the only endoscopic evidence of UC in patients with quiescent disease. As disease progresses, the mucosa becomes granular and friable. With more-severe inflammation, the mucosa may be covered by yellow-brown mucopurulent exudates associated with mucosal ulcerations. In UC, mucosal ulcerations occur in areas of inflammation, vary in size from a few millimeters to several centimeters, and may be punctate, annular, linear, or serpiginous. Finally, severe UC is associated with mucosa that bleeds spontaneously, and, with diffuse colitis, there may be extensive areas of denuded mucosa from severe mucosal ulcerations (see Fig. 112-6). Marked edema can at times lead to narrowing of the lumen. In patients with long-standing UC, pseudopolyps may be present. Inflammatory pseudopolyps develop in active

1985

A

B

C

D

F

E Figure 112-6.  Spectrum of severity of ulcerative colitis. A, Colonoscopic findings in mild ulcerative colitis demonstrating edema, loss of vascularity, and patchy subepithelial hemorrhage. B, Colonoscopic findings in severe ulcerative colitis demonstrating loss of vascularity, hemorrhage, and mucopus. The mucosa is friable, with spontaneous bleeding as well as bleeding after the mucosa is touched by the endoscope. C, Histopathology showing a severe acute and chronic inflammatory process, with multiple crypt abscesses. D, The colonic architecture is distorted, with a loss of crypts and abnormal branching of the crypts. Recognition of the disordered architecture is useful in differentiating acute from chronic colitis. E, Colonoscopy findings in a patient with chronic ulcerative colitis undergoing surveillance for colorectal cancer. The ascending colon (top left), transverse colon (top right), and descending colon (bottom left) are normal, but the sigmoid colon shows active inflammation (bottom right). F, A biopsy specimen of the normal-appearing colon demonstrates abnormal architecture with shortened crypts and no active colitis.

Chapter 112  Ulcerative Colitis disease and result from inflamed, regenerating epithelium that is interposed among ulcerations. These inflammatory pseudopolyps may give the colonic mucosa a cobblestoned appearance. With repeated inflammation that is followed by healing, these pseudopolyps remain during the quiescent phase of disease and usually do not regress with treatment. Endoscopically, pseudopolyps typically are small, soft, pale, fleshy, and glistening; however, they may be large, sessile, or pedunculated and may have surface ulcerations. Differentiation of these benign pseudopolyps from neo­ plastic polyps may be difficult and require histologic confirmation. There is a loss of normal colonic architecture with longstanding inflammation that is characterized by muscular hypertrophy, loss of the normal haustral fold pattern, decreased luminal diameter, and shortening of the colon; a resultant featureless appearance of the colon in chronic UC gives rise to the lead pipe appearance seen on barium enema. Strictures can occur in patients with chronic UC and result from focal muscular hypertrophy associated with inflammation. Malignancy must be excluded in patients with UC who have strictures, particularly long strictures without associated inflammation and strictures proximal to the splenic flexure (see later).

RADIOLOGY Plain Films

Figure 112-7.  Plain abdominal film of a patient with severe ulcerative colitis. The transverse colon is dilated (arrow), the colon wall is thickened, and mucosal islands are visible. In addition, distended loops of small bowel are apparent.

Barium Enema

Figure 112-8.  Plain abdominal film of a patient with mild left-sided ulcerative colitis showing a stool-filled proximal colon.

Patients with a severe attack of UC should have a supine plain film of the abdomen. The presence of intraperitoneal air may be missed on plain abdominal films, however, and CT has demonstrated a better diagnostic yield than plain abdominal radiography for detecting disease complications and extent. In the presence of severe disease, the luminal margin of the colon—the interface between the colonic mucosa and the luminal gas—becomes edematous and irregular. Thickening of the colonic wall often is apparent on a plain film, and prognostic signs such as islands of residual mucosa surrounded by extensive deep ulcerations, distention of the small bowel, and dilatation of the colon can be detected (Fig. 112-7). Plain films also are useful for detecting the presence of fecal material. Inflamed colons seldom contains feces, and no fecal material is present when the whole colon is involved. It is common, however, for a patient with leftsided disease to have proximal constipation (Fig. 112-8). Thus, a plain film can give considerable information with respect to the extent of disease. The presence of marked colonic dilatation suggests fulminant colitis or toxic megacolon. A plain abdominal film also can detect unsuspected free air and is especially useful in following the daily progress of a patient on high-dose glucocorticoid therapy in whom such a complication may be masked. With the advent of endoscopy, barium studies have been used less often in the care of patients with UC. Barium studies of the colon remain important, however, and may be superior to colonoscopy for certain specific scenarios, such as evaluation of colonic strictures; barium enema provides information on their location, length, and diameter and allows visualization of the entire colon when the presence of strictures precludes advancement of the colonoscope. Upper gastrointestinal barium study and small bowel follow-through with air-contrast visualization of the terminal ileum should be performed to exclude Crohn’s disease. The earliest radiologic change of UC seen on barium studies is fine mucosal granularity (Fig. 112-9). The mucosal

line becomes irregular and is not as sharp as that of a normal colon. With increasing severity, the mucosal line becomes thickened and irregular, and superficial ulcers are well shown en face. Deep ulceration can appear as collar-stud or collar-button ulcers in tangent, which indicates that the ulceration has extended through the mucosa to the muscularis propria (Fig. 112-10). Haustral folds may be normal in mild disease but become edematous and thickened as disease progresses. Loss of haustrations also can occur, especially in patients with long-standing disease (Fig. 112-11). Because the left colon

1987

1988

Section X  Small and Large Intestine

Figure 112-9.  Film from a double-contrast barium enema examination in a patient with long-standing ulcerative colitis as indicated by a marked loss of haustration. The mucosa is finely granular throughout the colon, consistent with mildly active disease. The terminal ileum is normal.

Figure 112-11.  Film from a barium enema examination showing postinflammatory polyposis in a shortened sigmoid and descending colon in a patient with active ulcerative colitis.

DIFFERENTIAL DIAGNOSIS A variety of inflammatory and noninflammatory diseases of the colon can mimic UC and need to be considered in establishing the correct diagnosis. This differential diagnosis can be grouped broadly into three categories: Crohn’s disease, infections, and noninfectious causes (Table 112-3).

CROHN’S DISEASE

Figure 112-10.  Film from a double-contrast barium enema examination in a patient with active ulcerative colitis. This localized view of the splenic flexure shows multiple ulcers. At the flexure itself there is deep ulceration appearing as a collar-button ulcer (arrow).

may normally lack haustration, this sign is relevant for only the ascending and transverse colon. With long-standing disease, loss of haustration can lead to a featureless and tubular appearance of the colon. Other chronic changes are shortening of the colon and widening of the presacral (retrorectal) space as seen on a lateral film of the rectum. Pseudopolyps may be present and often are filiform. In the presence of active changes, these pseudopolypoid changes can resemble a cobblestone pattern (see Fig. 112-11).

Crohn’s disease should be excluded in all patients given a diagnosis of UC, and colonoscopy with multiple biopsies is important in this regard. The presence of skip lesions or granulomas supports the diagnosis of Crohn’s disease. It is important to recognize that muciphage granulomas may be present in patients who do not have Crohn’s disease. Other endoscopic features distinguishing UC from Crohn’s disease are listed in Table 112-4. It is not uncommon for patients with ileal Crohn’s disease also to have rectal involvement, and these patients might present with symptoms of proctitis rather than symptoms of small intestinal involvement. In patients with proctitis or diffuse pancolitis, it may be impossible to differentiate UC and Crohn’s disease. Thus it is advisable to obtain radiologic assessment of the small intestine in all patients with colonic disease, particularly in those with pancolitis or proctitis on colonoscopy and elevated inflammatory markers or hypoalbuminemia on laboratory testing. A definitive diagnosis of Crohn’s disease may not be able to be made until the development of small bowel disease or perianal complications. “Indeterminate” colitis is diagnosed in approximately 10% of patients when the distinction between Crohn’s disease and UC cannot be made. These patients generally are initially managed as if they had UC, until more characteristic features of Crohn’s disease appear. The key points

Chapter 112  Ulcerative Colitis in the differential diagnosis between UC and Crohn’s disease are summarized in Tables 112-4 and 112-5.

INFECTION

Another major category of differential diagnosis for UC is infection (see Chapter 107). As mentioned previously, newly diagnosed UC can manifest as part of a well-documented episode of infectious colitis. It is unknown if the infection prompts the UC or simply unmasks underlying UC that previously had subclinical activity. Patients with documented UC in clinical remission also can develop acute infectious colitis and present with symptoms of a flare of UC. Thus, infections need to be excluded with each episode of disease exacerbation. The most common organisms causing infectious colitis are Salmonella, Shigella, and Campylobacter. Patients with

Table 112-3  Differential Diagnosis of Ulcerative Colitis Infectious Causes Aeromonas hydrophila Campylobacter jejuni Chlamydia spp. Clostridium difficile Cytomegalovirus Entamoeba histolytica Escherichia coli O157:H7, other EHEC Herpes simplex virus Listeria monocytogenes Neisseria gonorrhoeae Salmonella spp. Schistosomiasis Shigella spp. Yersinia enterocolitica Noninfectious Causes Acute self-limited colitis Behçet’s disease Crohn’s disease Diversion colitis Diverticulitis Drugs and toxins Chemotherapy Gold Penicillamine Eosinophilic colitis Graft-versus-host disease Ischemic colitis Microscopic colitis Collagenous colitis Lymphocytic colitis Neutropenic colitis (typhlitis) Nonsteroidal anti-inflammatory drugs Radiation colitis Solitary rectal ulcer syndrome EHEC, enterohemorrhagic E. coli.

infectious colitides usually have a more acute onset of symptoms than do patients with a flare of UC, and they have a prominence of abdominal pain; they also might report diarrheal illness in one or more of their contacts. The sigmoidoscopic appearance of infectious colitis may be indistinguishable from that of UC, but the histologic appearance often is helpful in differentiating infectious acute colitis from a more chronic condition. The presence of a chronic inflammatory infiltrate, architectural disturbances, and basal lymphoid aggregates favors a diagnosis of UC, and these features distinguish infectious colitis from UC with a probability of 80%, albeit with considerable interobserver variation. One would not be able to identify bacterial superinfection on a background of chronic UC on histology, however, because the additional changes would be nonspecific; only viral (especially cytomegalovirus) or amebic superinfection would be readily identifiable on biopsy specimens by the presence of intranuclear inclusion bodies or the actual organisms themselves, respectively (see later). Other bacterial infectious colitides include infection with E. coli O157:H7, which can occur in outbreaks or sporadically. Patients with this infection, particularly children and the elderly, usually present with bloody diarrhea and can develop associated hemolytic-uremic syndrome or thrombotic thrombocytopenic purpura. Because the diagnosis requires a special culture medium and cannot be made on routine stool cultures, clinicians need to have a high index of suspicion and specifically request such a test. Development of molecular probes might facilitate the ability to establish this diagnosis. Yersinia infections can cause enteritis, enterocolitis, or colitis and can last for several months before resolving spontaneously. The diagnosis is made on the basis of stool culture or a rising titer of serum antibody. Other, less-common bacterial infections causing colitis include Aero­monas hydroph­ ila and Listeria monocytogenes; the former is usually associated with drinking untreated water and the latter is often associated with consumption of unpasteurized milk. A history of antibiotic use suggests pseudomembranous colitis associated with C. difficile (Chapter 108). This infectious colitis, which is among the most common infections in UC patients, manifests with diarrhea and may be superimposed on or even lead to a relapse of UC. In a national survey of hospitalized patients in the United States, the prevalence of C. difficile among UC patients (3.7%) was eight times higher than that of either non-IBD gastroenterology patients or all hospital-discharge patients.121 This study also showed that prevalence of C. difficile infection in UC patients has doubled over the last seven years. C. difficile infection can occur in the absence of antibiotics, especially in the elderly and in patients who are taking proton pump inhibitors or who have had placement of a postpyloric tube. Thus, it is important to exclude C. difficile infection in

Table 112-4  Endoscopic Differentiation of Ulcerative Colitis and Crohn’s Disease FEATURE

ULCERATIVE COLITIS

CROHN’S DISEASE

Distribution

Diffuse inflammation that extends proximally from the anorectal junction Diffuse erythema, early loss of vascular markings with mucosal granularity or friability Small ulcers in a diffusely inflamed mucosa; deep, ragged ulcers in severe disease Often narrowed in long-standing chronic disease; tubular colon; strictures are rare

Rectal sparing, frequent skip lesions

Inflammation Ulceration Colonic lumen

Focal and asymmetrical, cobblestoning; granularity and friability less commonly Aphthoid ulcers, linear or serpiginous ulceration; intervening mucosa is often normal Strictures are common

1989

1990

Section X  Small and Large Intestine Table 112-5  Features That Distinguish Ulcerative Colitis from Major Differential Diagnoses RADIOLOGIC AND COLONOSCOPIC FEATURES

DIAGNOSIS

CLINICAL FEATURES

Ulcerative colitis

Bloody diarrhea

Extends proximally from rectum; fine mucosal ulceration

Crohn’s colitis

Perianal lesions are common; may be associated with ileitis; frank bleeding is less common than in ulcerative colitis Occurs in the elderly; sudden onset, often painful; usually resolves spontaneously in several days

Segmental disease; rectal sparing; strictures, fissures, ulcers, fistulas; small bowel involvement

Ischemic colitis

Segmental splenic flexure and sigmoid involvement are most common, with thumbprinting early and ulceration after 24-72 hr; rectal involvement is rare Usually normal

Microscopic colitis

Watery diarrhea; normal-appearing mucosa at colonoscopy

Infectious colitis

Sudden onset; identifiable source in some cases (e.g., Salmonella spp.); pain may predominate (e.g., Campylobacter spp.); pathogens are present in stool History of travel to endemic area; amebae may be detected in a fresh stool specimen but ELISA for ambic lectin antigen is preferable Rectal pain; pus

Nonspecific findings

Often a history of antibiotic use; characteristic pseudomembranes may be seen on sigmoidoscopy; Clostridium difficile toxin is detectable in stools

Edematous; shaggy outline of colon; pseudomembranes may be identified radiologically or seen at colonoscopy

  Amebic colitis

  Gonococcal proctitis Pseudomembranous colitis

HISTOLOGIC FEATURES Distortion of crypts; acute and chronic diffuse inflammatory cell infiltrate; goblet cell depletion; crypt abscesses; lymphoid aggregates Focal inflammation; submucosal involvement; granulomas; goblet cell preservation; transmural inflammation; fissuring Mucosal necrosis with ghost cells; congestion with red blood cells; hemosiderin-laden macrophages and fibrosis (when disease is chronic) Chronic inflammatory infiltrate; increased intraepithelial lymphocytes (lymphocytic colitis) and subepithelial collagen band (collagenous colitis) Crypt architecture is usually normal; edema, superficial neutrophilic infiltrate, crypt abscesses

Discrete ulcers; ameboma or strictures

Similar to ulcerative colitis; amebae present in lamina propria or in flask-shaped ulcers, identified by periodic acid–Schiff stain

Granular changes in rectum

Intense polymorphonuclear neutrophil infiltration; purulent exudate; Gram-negative diplococci May resemble acute ischemic colitis; shows summit lesions of fibrinopurulent exudate

ELISA, enzyme-linked immunosorbent assay.

hospitalized patients whose colitis does not respond to medical therapy as expected or who suddenly lose their initial response to treatment. Patients with C. difficile infections often present with watery diarrhea (usually without rectal bleeding) and have characteristic pseudomembranes at colonoscopy. This infection can cause severe colitis that progresses to toxic megacolon and bowel perforation. Thus, appropriate stool studies for toxin analysis are necessary to exclude superimposed C. difficile infection, even in patients with established UC who present with an exacerbation.122 In patients from endemic areas, certain protozoan and parasitic infections need to be considered (see Chapters 109 and 110). Amebic colitis tends to have a more prolonged course than do most bacterial colitides, but amebiasis is not a cause of chronic colitis. Schistosomal colitis may be chronic and diffuse, exhibit pseudopolyps, and involve the rectum. The presence of characteristic ova in a biopsy specimen confirms the diagnosis. Other infectious causes of a bloody diarrhea include opportunistic infections of the colon in immunosuppressed patients (see Chapters 33 and 34). Cytomegalovirus (CMV) infection has been reported in patients with UC, typically those with long-standing disease who are being treated with glucocorticoids or immunosuppressants; the diagnosis of CMV infection should be con­ sidered whenever patients who have UC and are taking

glucocorticoids either fail to respond as expected or lose their response to treatment. CMV colitis in steroid-naive patients also has been described.123 Patients with CMV colitis often present with abdominal pain and bloody diarrhea and have discrete deep ulcers on colonoscopy; however, CMV colitis can manifest with diffuse inflammation and resemble UC. Because the clinical presentation of CMV colitis may be indistinguishable from a flare of UC, a high index of suspicion is needed to make the diagnosis. Endoscopic biopsies should be obtained from both the ulcer bed and adjacent mucosa; careful histologic examination for giant cells with intranuclear inclusion bodies is important to confirm the diagnosis. Mycobacterium avium complex usually causes patchy rather than diffuse inflammation. Sexually transmitted causes of proctitis, including gonorrhea, Chlamydia, and lymphogranuloma venereum, usually do not cause diarrhea and are associated with large volumes of watery pus, especially gonorrhea. These diagnoses are suspected clinically and confirmed by appropriate cultures as well as histologic appearance on rectal biopsy specimens.

OTHER CAUSES

Noninfectious causes of colitis that should be considered in the differential diagnosis of UC include diverticulitis, ischemia, radiation, collagenous colitis, lymphocytic colitis,

Chapter 112  Ulcerative Colitis and drug-induced colitis. Diverticulitis and ischemic colitis usually present acutely or (less commonly) subacutely, but most of the noninfectious colitides have prolonged presentations that can extend for several months. Acute diverticulitis most commonly occurs in the sigmoid colon and does not involve the rectum (see Chapter 117). When the inflammation does extend to the rectum, it tends to be patchy and involves only the upper rectum. This appearance is more likely to be confused with Crohn’s disease than UC. Ischemic colitis usually occurs in the elderly (see Chapter 114). The classic distribution is segmental involvement in the watershed areas around the splenic flexure or sigmoid colon; ischemic proctitis also has been described. Radiation colitis usually occurs in patients who have been given radiation therapy for uterine, cervical, or prostate cancer. The location of disease depends on the sites irradiated but typically involves the rectosigmoid. The onset of symptoms often temporally corresponds to the radiation therapy but can develop years afterward (see Chapter 39). Microscopic colitis, including lymphocytic and collagenous colitis, manifests with diarrhea and should be distinguished readily from UC by lack of rectal bleeding, the normal endoscopic appearance, and characteristic histopathology (see Chapter 124). A drug history must always be taken in a patient with colitis, because NSAIDs, gold, and penicillamine among many other drugs may induce colonic inflammation. Patients with UC can present with symptoms similar to those of IBS (Chapter 118), specifically diarrhea, abdominal pain, fatigue, and poor general well-being. Lack of rectal bleeding and laboratory markers of inflammation, as well as a normal endoscopic and histologic appearance, helps to distinguish IBS from active UC. Patients with quiescent UC can have concomitant symptomatic IBS in the absence of active inflammation. The physician must be able to recognize these subtleties in order to provide appropriate management for the patient. Patients with UC also can present with symptoms mimicking colonic neoplasm (Chapter 123), solitary rectal ulcer syndrome (Chapter 124), diverticular disease (Chapter 117), and factitious diarrhea (Chapter 15). These diagnoses also do not give rise to diffuse inflammation in the colon and, therefore, should be distinguished easily from UC on colonoscopy.

love and Witts.124 This purely clinical classification categorizes disease as mild, moderate, or severe based on a combination of clinical findings and laboratory parameters, including frequency of bowel movements, rectal bleeding, fever, tachycardia, anemia, and elevated ESR (Table 112-6). The Truelove and Witts classification is reliable and simple to use in clinical practice, although it is most applicable for patients with extensive colitis and might not adequately reflect disease severity in patients with limited colitis. A numerical disease activity instrument that is more useful for patients with limited disease and for conducting clinical trials is the Ulcerative Colitis Disease Activity Index (UCDAI, also known as the Sutherland Index).125 This index, which combines clinical and endoscopic assessments, is the sum of scores from four components: stool frequency, rectal bleeding, sigmoidoscopic findings, and physician’s global assessment (Table 112-7). This disease activity index ranges from 0 to 12, with the higher total scores representing more-severe disease. In general, a patient is considered to be in remission if the UCDAI score is 2 or less and to have

Table 112-6  Truelove and Witts Classification of the Severity of Ulcerative Colitis Mild <4 stools/day, without or with only small amounts of blood No fever No tachycardia Mild anemia ESR < 30 mm/hr Moderate Intermediate between mild and severe Severe >6 stools/day, with blood Fever > 37.5°C Heart rate > 90 beats/min Anemia with hemoglobin level < 75% of normal ESR > 30 mm/hr ESR, erythrocyte sedimentation rate. Adapted from Truelove SC, Witts LJ. Cortisone in ulcerative colitis: Final report on a therapeutic trial. Br Med J 1955; 2:1041.

Table 112-7  Ulcerative Colitis Disease Activity Index* SCORE

ASSESSMENT OF DISEASE ACTIVITY Assessment of disease activity is important for prognostication and therapeutic decision making. Unlike the case with Crohn’s disease (where there is one well-accepted and validated instrument to measure disease activity), there exist numerous instruments to measure disease activity for UC, some of which are purely clinical, endoscopic, or histologic, and others that combine clinical and endoscopic assessments. When evaluating these various activity indices, it is important to recognize that generally these measures are nonspecific, and thus patients with other disorders, such as IBS, could attain high scores even in the absence of any inflammation. Also, almost none of the outcome measures in UC have been prospectively validated. Moreover, the scores of the indices typically are derived by incorporating various signs and symptoms for which there are no standardized definitions. For these reasons, one should exercise appropriate caution when interpreting the results of a study of UC patients that uses an activity index. Although none of these indices is accepted universally as standard, one of the most commonly used is that of True-

Stool Frequency 0 1 2 3 Rectal Bleeding 0 1 2 3 Mucosal Appearance 0 1 2 3 Physician Global Assessment 0 1 2 3

CRITERIA Normal 1-2 stools/day > normal 3-4 stools/day > normal >4 stools/day > normal None Streaks of blood Obvious blood Mostly blood Normal Mild friability Moderate friability Exudation, spontaneous bleeding Normal Mild Moderate Severe

*Sutherland index: Range, 0-12. (From Sutherland LR, Martin F, Greer S, et al. 5-Aminosalicylic acid enema in the treatment of distal ulcerative colitis, proctosigmoiditis, and proctitis. Gastroenterology 1987; 92:1894.)

1991

1992

Section X  Small and Large Intestine Table 112-8  Endoscopic and Histologic Assessment of Disease Activity in Ulcerative Colitis

Table 112-9  Induction Therapy for Ulcerative Colitis Based on Disease Severity

SCORE

Mild Disease 5-Aminosalicylates Topical (distal colitis) Oral (distal/extensive colitis) Combination Moderate Disease 5-Aminosalicylates Topical (distal colitis) Oral (distal/extensive colitis) Combination Glucocorticoids Topical (distal colitis) Oral (distal/extensive colitis) Combination Azathioprine or 6-mercaptopurine Severe Disease IV glucocorticoids IV cyclosporine IV infliximab

CRITERIA

Endoscopic Assessment 0 Normal mucosa 1 Loss of vascular pattern 2 Granular, nonfriable mucosa 3 Friability on rubbing 4 Spontaneous bleeding, ulceration Histologic Assessment 0 Normal 1 No significant inflammation: Possibly architectural changes of chronic disease and small foci of lymphocytes but no acute inflammation, crypt abscesses, or epithelial destruction 2 Mild to moderate inflammation: Edema, vascularity, increased acute and chronic inflammatory cells but intact epithelium 3 Severe inflammation: Heavy infiltrate of acute and chronic inflammatory cells, crypt abscesses, ulceration of surface epithelium, purulent exudate

severe disease if the score is greater than 10. Clinical response generally is accepted to be reflected by a decrease of three points from the patient’s initial baseline score. An index very similar to the UCDAI that has been used extensively in recent randomized, controlled trials (RCTs) is the Mayo score, which incorporates the same four components as the UCDAI.126 Other scales also have been developed, many of which are modifications of the Truelove and Witts classification and the UCDAI. None of these disease activity instruments has ever been formally validated. There also exist many endoscopic and histologic scales for grading the severity of colitis (Table 112-8).127,128 Endoscopic findings do not always correlate with clinical symptoms, and such correlations generally are more consistent within individuals. Thus, although therapeutic decisions are based primarily on clinical status, it may be useful to follow the sigmoidoscopic mucosal appearance over time in an individual patient, if the clinical response to treatment is uncertain. In addition to the typical categorization of disease activity into mild, moderate, and severe, an important subgroup is fulminant colitis. Patients with severe colitis who appear toxic, with fever higher than 38.3°C (101°F), tachycardia, abdominal distention, signs of localized or generalized peritonitis, and leukocytosis, are considered to have fulminant colitis. Toxic megacolon is said to occur when there is radiologic evidence of transverse colon dilatation to greater than 6 cm in an acutely ill patient. Fulminant colitis and toxic megacolon are clinical diagnoses, and complete co­ lonoscopic examination should be avoided in patients with severe or fulminant colitis because of the risk of inducing megacolon or perforation. In this patient population, a limited flexible sigmoidoscopy is appropriate to ensure that the etiology of the symptoms is UC itself and not other conditions.

TREATMENT MEDICAL

The goals of therapy of UC are to induce remission, to maintain remission, to maintain adequate nutrition, to minimize disease- and treatment-related complications, and

IV, intravenous.

Table 112-10  Maintenance Therapy for Ulcerative Colitis 5-Aminosalicylates Topical (distal colitis) Oral (distal/extensive colitis) Azathioprine or 6-mercaptopurine Infliximab

Disease distribution Distal Extensive Prior therapy

Severity of disease Mild Moderate Severe

Medication choice

Response Side effects Compliance

Figure 112-12.  Factors that should be considered in the choice of medical therapies for ulcerative colitis.

to improve the patient’s quality of life. Current management strategy focuses on using appropriate medical therapy and optimizing timing of surgery. Several factors should be considered in determining optimal therapy for patients with UC (Fig. 112-12). Current therapeutic strategies can be classified broadly, based on disease activity, into those that treat active disease (induction therapy) (Table 112-9) and those that prevent recurrence of disease once remission is achieved (maintenance therapy) (Table 112-10). This concept of induction and maintenance of remission forms the basis of our evaluation of the efficacy of a specific therapy. The extent of disease is an important consideration that helps determine the route of administration of medication. Thus, for example, proctitis may be treated with suppositories or foam preparations as well as with oral therapy, and enema preparations may be used alone or in combination with systemic therapy for patients with

Chapter 112  Ulcerative Colitis left-sided disease. Other important factors to consider are a patient’s prior response to or side effects from a specific medication and compliance with medication. These factors might favor or preclude the use of a specific agent. Given the chronic nature of UC, medications need to be efficacious and well accepted by patients from the standpoints of safety and ease of administration. The mainstay of medical therapy focuses on regimens that alter host response to decrease mucosal inflammation. Therapies that target other aspects of the systemic inflammatory process or manipulate the enteric flora also have been developed to treat UC. One important consideration when evaluating the efficacy of a particular medication (e.g., in a RCT that compares a novel therapy to placebo) is the placebo response rate. Even though placebos often are thought of as inert agents, they have been noted to lead to improvement in a variety of both subjective and objective outcome measures in a number of different medical conditions, such as anxiety, depression, insomnia, pain, asthma, obesity, hypertension, and even myocardial infarction.129 In some of these disorders, placebo response rates of up to 40% have been reported. With respect to placebo response rates in UC, a meta-analysis of 40 randomized, placebo-controlled trials, in which the most commonly used activity indices were the Mayo Score or the UCDAI, reported a pooled placebo response rate of 28% and remission rate of 13%.130 Studies using less-stringent outcome definitions were noted to have higher placebo response and remission rates. Univariate analysis suggested that longer follow-up duration, higher number of follow-up visits, longer disease duration, and lower disease severity at study entry were positively associated with the placebo remission rate.

Aminosalicylates

Oral Sulfasalazine consists of an antibacterial component, sulfapyridine, bonded by an azo bond to a salicylate, 5-aminosalicylic acid (5-ASA, mesalamine) (Fig. 112-13).131 The drug was synthesized by Nana Svartz in 1938-1939 and its benefit for the treatment of IBD was discovered serendipitously in 1941-1942 by her when patients with UC receiving this medication for a presumed diagnosis of rheumatoid arthritis noted improvement in colitis symptoms132; in retrospect, these patients had peripheral arthropathy associated with their IBD. Research subsequently established that 5-ASA is the principal therapeutic moiety of sulfasalazine in IBD and that the sulfapyridine component of the parent drug serves as an inactive carrier, largely preventing absorption of 5-ASA in the small intestine and allowing it to be released in the colon.133,134 Approximately 90% of sulfasalazine reaches the colon, and only a small amount is absorbed in the small intestine. On reaching the colon, the enzyme azoreductase, which is elaborated by colonic bacteria, cleaves the azo bond to release the active constituent moiety, 5-ASA. After 5-ASA is absorbed from the colon, 20% of the compound undergoes hepatic acetylation, forming N-acetyl 5-ASA, and is excreted in the urine. Sulfasalazine is one of several agents in the class of 5-ASA compounds that is considered to be the first line of therapy for inducing remission in patients with mild to moderate UC.131,135 Mesalamine derivatives have not been evaluated in a randomized, controlled fashion in patients with severely active disease. At a dose of 3 to 6 g/day, sulfasalazine induces remission in 39% to 62% of patients with mild to moderate UC, about twice the remission rate of placebotreated patients.136,137 Various formulations and controlled-release systems (Table 112-11; see Fig. 112-13) have been developed to

Sulfasalazine HOOC

O N

HO

S

N

NH

O 5-Aminosalicylate

N

Sulfapyridine

Olsalazine 5-ASA

N

N

5-ASA

Mesalamine (Pentasa brand) HOOC NH2

HO

Ethylcellulose

Mesalamine (Asacol brand) HOOC Eudragit-S NH2

HO

Balsalazide HOOC HO

O N

N

Figure 112-13.  Molecular preparations.

structures

CH2

COOH

5-aminosalicylate

(5-ASA)

NH

C of

CH2

deliver 5-ASA to specific sites of the gastrointestinal tract without the sulfapyridine moiety, which is thought to be responsible for most of the side effects. Olsalazine (Dipentum) is a 5-ASA dimer linked by an azo bond and is formulated in gelatin capsules. Balsalazide (Colazal) consists of a 5-ASA monomer linked to a biologically inactive carrier molecule, 4-aminobenzoyl-β-alanine. Similar to sulfasalazine, 5-ASA is released from olsalazine and balsalazide in the colon upon cleavage of the azo bond via the bacterial enzyme azoreductase. Approximately 99% of the drug is delivered intact to the colon, and its metabolites are cleared rapidly in the urine. Three commonly used mesalamine preparations allow delivery of 5-ASA before the drug reaches the colon: Pentasa, Asacol, and Lialda. Pentasa uses ethyl cellulose-coated microgranules that release mesalamine from the duodenum throughout the small bowel and the colon; about 50% of 5-ASA is released in the small intestine, and the remainder is released in the colon. Asacol is a Eudragit-S-100–coated mesalamine tablet that is released at a pH greater than 7, usually in the distal ileum and the colon. With Asacol, about 15% to 30% of mesalamine is released in the small intestine. Lialda (MMx mesalamine) is a novel mesalamine formulation that uses a multimatrix structure composed of an inner lipophilic matrix and an outer hydrophilic matrix. It is coated with a pH-dependent polymethacrylate film to

1993

1994

Section X  Small and Large Intestine Table 112-11  Oral 5-Aminosalicylic Acid Preparations and Sites of Delivery in the Gastrointestinal Tract

Table 112-12  Side Effects of Sulfasalazine and 5-Aminosalicylates

DRUG

Dose-Related Alopecia Anorexia Back pain Folate malabsorption (with sulfasalazine) Headache Nausea, vomiting, dyspepsia Non–Dose-Related Agranulocytosis, aplastic anemia Arthralgia Colitis Fever Fibrosing alveolitis, pulmonary eosinophilia Hemolytic anemia (Heinz bodies) Hepatitis Hypersensitivity skin rashes (occasionally with photosensitivity) Male infertility (with sulfasalazine) Pancreatitis Pericarditis, myocarditis

Prodrugs Sulfasalazine Olsalazine Balsalazide

FORMULATION

Sulfapyridine + 5-ASA 5-ASA dimer 4-aminobenzoyl β-alanine + 5-ASA Mesalamine Preparations Asacol, Claversal, pH sensitive, Salofalk resin-coated; delayed release Rowasa Enema Canasa Suppository Pentasa Ethylcellulose-coated microgranules; controlled release Lialda pH sensitive, multi-matrix and polymethacrylate coated; delayed and slow release

SITE OF DELIVERY Colon Colon Colon Distal ileum, colon Distal colon Rectum Duodenum to colon Distal ileum, colon

5-ASA, 5-aminosalicylate.

allow the delayed release of mesalamine in the terminal ileum and colon at a pH greater than 7. This technology also allows mesalamine to be released slowly and in close proximity to the colonic mucosa. These oral 5-ASA derivatives (mesalamines) have been shown to be superior to placebo for mildly to moderately active UC.137-142 Meta-analyses have demonstrated that the mesalamines are as efficacious as sulfasalazine, and the various mesalamine preparations appear to be comparable in efficacy.137,140 Balsalazide has been shown to have superior efficacy and a more rapid response compared with traditional mesalamine agents.143,144 In a RCT, balsalazide 6.75 g/day, a dose equivalent to mesalamine 2.4 g daily, achieved higher rates of remission and had better tolerance compared with pH-dependent mesalamine 2.4 g/day.143 It has been suggested that the greatest benefit of balsalazide is in patients with newly diagnosed left-sided UC.144 More important than the specific 5-ASA preparation is the dose-dependent response when 5-ASA is used as an induction therapy for active UC.137,140 For this indication, mesalamine is not effective at doses lower than 2 g daily, and there is an increased response at doses of 4 to 4.8 g daily. The ASCEND I and II trials showed that mesalamine at doses of 2.4 and 4.8 g/day have similar efficacy for patients with mildly active disease, but the higher dose (4.8 g/day) was more efficacious in patients with moderately active disease.145,146 This dose of mesalamine is comparable to 12 g/day of sulfasalazine, which is impractical in clinical practice because of the high probability of intolerance. No RCT has evaluated the use of aminosalicylates for severely active UC, but these agents are generally thought not to be effective in severely active disease. Once remission is achieved, sulfasalazine and other 5-aminosalicylates are effective in maintaining it.147-150 This benefit appears to be dose dependent for sulfasalazine, with a dose of 2 g/day often used to balance efficacy and adverse side effects.147 Such a dose-dependent response, however, has not been found with the other 5-ASA preparations, and at doses of 1.5 to 4.8 g/day, remission can be maintained in more than 50% of patients.151 One meta-analysis has suggested that sulfasalazine might have a slight but statistically

significant therapeutic superiority relative to the newer 5-ASAs in maintaining remission when considering trials of six months’ duration; however, when these trials were combined with those of 12 months’ duration, this statistically significant benefit was lost.151 A double-blind RCT comparing two doses of balsalazide (1.5 g twice daily and 3 g twice daily) with mesalamine 0.5 g three times daily for six months reported a remission rate of 77.5% with the higher dose of balsalazide compared with remission rates of 56.8% and 43.8% with mesalamine and the lower dose of balsalazide, respectively.152 In general, the same dose of 5-ASA derivative that induces remission is recommended for maintenance therapy, although this recommendation has not been formally tested in a randomized, placebocontrolled fashion. Common side effects of sulfasalazine include fever, rash, nausea, vomiting, and headache (Table 112-12). Other, lesscommon but important side effects of sulfasalazine include hypersensitivity reactions, reversible sperm abnormalities, and impairment of folate absorption. Approximately 15% of patients taking sulfasalazine develop significant side effects that require discontinuing the medication. Up to 90% of patients who are intolerant to sulfasalazine, however, can tolerate mesalamine.153 In clinical trials, the newer 5-ASA preparations and balsalazide have been shown to be better tolerated than sulfasalazine,140,154,155 although the adverse event profiles during maintenance therapy appear to be similar for 5-ASA preparations and sulfasalazine.151 Sulfasalazine can impair folate absorption (by competitively inhibiting the jejunal enzyme, folate conjugase) thereby contributing to anemia, and folate supplementation should be prescribed to patients receiving sulfasalazine. Olsalazine is associated with drug-induced diarrhea in up to 10% of patients, which often limits its use. It has been noted that if olsalazine is ingested with meals and is continued despite the diarrhea, the incidence of this side effect can be lessened substantially to 3%. A systematic review of oral 5-ASA for maintenance of remission in UC found olsalazine to be significantly inferior to sulfasalazine, and this reduced efficacy was related mostly to a significantly higher rate of withdrawals because of adverse events.151 Oral mesalamine preparations do not appear to have significant dosedependent toxicity.

Chapter 112  Ulcerative Colitis Topical Topical aminosalicylates can be administered in the form of 5-ASA enemas, 5-ASA suppositories, and, in Europe, 5-ASA foam. The use of enemas allows the medication to be delivered up to the level of the splenic flexure in about 95% of patients, and suppositories can be used to treat disease up to 15 to 20 cm from the anal verge. Topical mesalamine derivatives may be used as an alternative monotherapy or as an adjunctive therapy to oral agents in patients with left-sided colitis or pancolitis. They are effective for inducing remission in patients with mildly to moderately active distal UC, without a clear dose-response effect156,157 in nonrefractory patients. The standard dosing regimens used to induce remission are 1 to 4 g of 5-ASA in the form of an enema nightly, or mesalamine suppositories 1 to 1.5 g either nightly or in divided doses throughout the day. Mesalamine enemas have been shown to be comparable to oral sulfasalazine in the treatment of active distal UC, with fewer side effects.155 Similar efficacies have been demonstrated for mesalamine enemas regardless of whether the 1-, 2-, or 4-g formulation is used for inducing remission in patients with mild to moderate left-sided UC not requiring concurrent glucocorticoids or immunomodulators. In fact, mesalamine enemas are perceived to be even more effective than topical glucocorticoid enemas in this setting.157,159 A combination of topical and oral mesalamine also may be more effective than either agent alone in patients with left-sided colitis or pancolitis, suggesting a dose-response effect.160,161 In patients with proctitis, mesalamine suppositories, 500 mg administered twice daily, have been shown to be beneficial for treating active disease.162 Mesalamine foam has a more uniform distribution and longer persistence in the distal colon compared with mesalamine enemas. The foam preparation has been shown to have better patient acceptance than the enema preparation,163 but mesalamine foams currently are not available in the United States. Topical mesalamine preparations also are effective for maintaining remission in left-sided UC or proctitis.156,157 The effective maintenance dosing interval ranges from nightly to every three days. Topical mesalamine is as effective as oral mesalamine,164 and the combination of topical and oral mesalamine may be more effective than oral mesalamine alone as a maintenance regimen.165

Glucocorticoids

Systemic At doses equivalent to 40 to 60 mg/day of oral prednisone, glucocorticoids are effective first-line therapy for moderate or severe flares of UC.124,166-170 The use of doses higher than 60 mg/day is associated with increased side effects without appreciable clinical benefit and thus should be avoided. The addition of sulfasalazine to corticosteroids in moderately to severely active UC does not offer any incremental benefit. Although no study has directly compared the efficacy of oral and parenteral glucocorticoids, the latter commonly are used in severe disease. No adequately designed controlled study has been performed to confirm the clinical impression that continuous infusion of parenteral glucocorticoids is superior to pulse therapy. The use of adrenocorticotropin (ACTH) has been suggested as an alternative to conventional glucocorticoid therapy of active UC in small studies.171 One double-blind RCT suggested that intravenous ACTH was more effective than intravenous hydrocortisone for the treatment of severely active UC only in steroid-naïve patients172; this observation has not been confirmed. Because most patients with severely active flares have been treated previously with glucocorticoids, ACTH rarely is used in clinical practice. A

noteworthy complication of ACTH therapy is bilateral adrenal hemorrhage. Glucocorticoids have no maintenance benefits in patients with UC. Steroid-dependent patients, or patients who are unable to taper off glucocorticoids without experiencing disease exacerbation, benefit from the addition of steroidsparing agents. There has been no trial to date assessing mesalamine therapy and its efficacy in maintaining remission induced with glucocorticoids. The long-term remission rate in patients who require parenteral glucocorticoids for severe UC is approximately 50%.173 Immunomodulatory agents, as discussed, should be considered in patients who are dependent on steroids, who require two courses of glucocorticoids for induction of clinical response or remission within one year, or who require parenteral glucocorticoids to induce remission. In addition to the use of immunomodulatory agents, one should consider using infliximab for steroid-dependent patients. Glucocorticoids are associated with many mild and serious side effects in patients with IBD (Table 112-13). These side effects occur commonly and involve nearly every organ system. Every effort should be made to minimize glucocorticoid use and exposure. Budesonide is a glucocorticoid preparation that is structurally different from prednisone. The presence of 16α,17αacetyl side chains allows enhanced topical anti-inflammatory activity and affinity for glucocorticoid receptors compared with prednisone.174 In addition, budesonide has an approximately 90% first-pass metabolism in the liver and erythrocytes and is converted to metabolites that have little or no biological activity. The resultant low systemic bioavailability translates to significantly less toxicity compared with

Table 112-13  Side Effects of Glucocorticoids Cutaneous Acne Impaired wound healing Purpura, ecchymoses, petechiae Striae Endocrine Adrenal insufficiency Cushingoid appearance Gastrointestinal Dyspepsia Dysphagia/odynophagia (candidiasis) Infectious complications Numerous pathogens Metabolic Electrolyte imbalance, hypokalemia Fluid retention Growth retardation Hyperglycemia, secondary diabetes mellitus Hyperlipidemia, altered fat distribution Hypertension Musculoskeletal Myopathy Osteonecrosis Osteoporosis Neuropsychiatric Anxiety, mood swings Depression Insomnia Psychosis Ocular Cataracts Glaucoma

1995

1996

Section X  Small and Large Intestine traditional glucocorticoids. Entocort is a controlled-ilealrelease oral budesonide preparation consisting of EudragitL-100–coated microgranules with an internal ethyl cellulose component; it releases budesonide at pH greater than 5.5, and about 50% to 80% of budesonide is absorbed in the ileocecal region. There currently is no oral formulation of budesonide that provides optimal release characteristics for the entire length of the colon. A small uncontrolled study has suggested that Budenofalk, which is not available in the United States, may be effective for prednisone-dependent UC.175 Controlled studies have not shown the benefit of oral budesonide for the treatment of active UC.176 Topical Topical glucocorticoids in liquid and foam formulations are effective short-term therapy for active UC distal to the splenic flexure.177,178 Foam preparations often are tolerated better by patients and may be easier to retain than liquid preparations. Topical glucocorticoids have been found to be less effective than topical mesalamine for inducing remission of distal UC159; however, the combination of topical corticosteroids and topical mesalamine has been more efficacious than either alone in the short-term treatment of distal UC.179 Whereas systemic absorption of glucocorticoids with topical therapy is significantly less than that with oral administration, prolonged treatment with topical glucocorticoids still may be associated with steroid-related side effects and should be avoided. As mentioned previously, budesonide is a potent corticosteroid with a rapid first-pass metabolism. Budesonide enemas, which currently are neither available nor approved in the United States, have been shown to be effective for the treatment of active distal UC in several controlled trials. In a double-blind RCT of patients with active distal UC, budesonide, 2 mg/100 mL for six weeks, resulted in a remission rate of 19% compared with 4% in patients receiving placebo therapy (P < 0.05).180 Subsequent trials have shown budesonide enema to be as efficacious as or even superior to prednisolone enema without resultant depression of endogenous cortisol levels.181-183 Budesonide enema perhaps is inferior in efficacy to me­ salamine enema,184 but it clearly presents an alternative topical glucocorticoid for treatment of distal UC. The optimal dose for budesonide enema consistently has been shown to be 2 mg/100 mL once daily.180,181,185 Budesonide in foam preparation also has been shown to have comparable efficacy with traditional hydrocortisone foam for the treatment of active proctosigmoiditis.186 Additional studies are needed to determine the effect of longer-term topical budesonide use. As with other glucocorticoid preparations, budesonide enema is not effective for maintaining remission in UC.185

Immunomodulators

Azathioprine and 6-Mercaptopurine Of the various immunomodulatory agents, the most widely used are azathioprine and 6-MP. These two agents are purine analogs that interfere with nucleic acid metabolism and cell growth and exert cytotoxic effects on lymphoid cells. They are inactive prodrugs with subtle structural differences. Azathioprine is nonenzymatically converted to 6-MP, which is then metabolized through a series of enzymatic pathways to active and inactive metabolites (see Fig. 111-8). The two primary metabolites of 6-MP are 6thioguanine nucleotides (6-TGNs) and 6-methylmercaptopurine (6-MMP). The 6-TGN metabolites are thought be responsible for the immunomodulatory action of azathio-

prine and 6-MP and their bone marrow suppression property, whereas hepatotoxicity is thought to be related to 6-MMP. One key enzyme involved in the biotransformation of 6-MP is thiopurine methyl­transferase (TPMT), which converts 6-MP to its inactive metabolites, 6-MMP and 6-methylmercaptopurine ribonucleotides. There is a population polymorphism in the TPMT gene: 89% of the population have homozygous wild-type TPMT, and 11% and 0.3% of the population have heterozygous and homozygous mutations, respectively.187 Persons with heterozygous and homozygous TPMT mutations have decreased to absent enzyme activity. The clinical significance of this genetic polymorphism is that inherited differences in TPMT may be responsible for most of the variability in drug response observed among individual patients. The efficacy of azathioprine in the treatment of UC is a matter of debate. Four RCTs have evaluated azathioprine for inducting remission in active UC (Table 112-14).188-191 These four studies were small, heterogeneous in design, used different outcome definitions for response, and reached different conclusions. Two of the studies involved steroiddependent patients,190,191 one other study used steroids for induction,188 and two studies used 5-ASAs as a comparator group rather than placebo.189,191 Only one study showed a significant benefit with azathioprine compared with 5-ASA for induction therapy in steroid-dependent disease.191 With respect to the use of azathioprine for maintenance of remission in UC, four RCTs have been performed (see Table 112-14).188,192-194 Just as with studies of induction therapy, these four studies also had small sample sizes, used heterogeneous designs with different outcome definitions of response, allowed for various cotherapies, and again reached different conclusions. One of the studies was in steroiddependent disease,192 another allowed the use of steroids for relapse,188 one study used 5-ASA as a comparator group rather than placebo,194 and another included patients who were mostly taking 5-ASAs and was actually a study of azathioprine withdrawal.193 Only this withdrawal study showed a benefit with continued azathioprine. Thus, for the purpose of induction or maintenance therapy for UC, our use of azathioprine is largely based on its established efficacy in Crohn’s disease rather than any proven benefit in UC. One subset of patients, however, has been shown to obtain benefit with the use of azathioprine, specifically patients who have severely active UC and who are able to attain induction of remission with intravenous followed by oral cyclosporine. In these patients, maintenance therapy with azathioprine has been reported to decrease colectomy rates (see later). The optimal dose of azathioprine or 6-MP for treating UC is unclear, and no formal dose-ranging study has been reported in the literature. The effective doses for 6-MP and azathioprine generally are 1 to 1.5 mg/kg/day and 2 to 3 mg/ kg/day, respectively.195 At these doses, however, there still may be nonresponders and, for them, higher doses may be necessary. Induction of leukopenia had been advocated for dose optimization,196 but this practice was not supported by subsequent studies.197-199 Monitoring metabolite levels may be beneficial in determining the optimal dose of azathioprine or 6-MP. To date, at least 13 studies examining response in IBD with respect to 6-TGN level have been published. A metaanalysis of the first 12 of these studies found that the studies were similar in that they were retrospective and the majority of patients were adults with Crohn’s disease, but they were heterogeneous with respect to sample size, the proportion of patients in remission, and the activity indices used to

Chapter 112  Ulcerative Colitis Table 112-14  Randomized, Controlled Trials of Azathioprine for Ulcerative Colitis REFERENCE Induction 188 189 190 191 Maintenance 188 192 193 194

N

DOSE (MG/KG/D)

DURATION OF THERAPY (months)

RESPONSE (AZA)

RESPONSE (control)

P-VALUE

1 3 6 6

78% 60% NR 53%

68% 80% NR 19%

NS NS NS .006

Glucocorticoids in all None; control = 5-ASA None None; control = 5-ASA Glucocorticoids for relapse None 5-ASA in most AZA withdrawal Glucocorticoid induction Control = 5-ASA

80 20 44* 72*

2.5 2.5 2-2.5 2

80 30* 67

1.5-2.5 1.5 NR

11 6 12

40% NR 64%

23% NR 41%

NS NS .039

25

2.5

18

42%

62%

NS

CO-THERAPY

*All patients in this study were glucocorticoid-dependent. 5-ASA, 5-aminosalicylates; AZA, azathioprine; mo, months; N, number of patients; NR, not reported; NS, not significant.

assess response.200 Of the seven studies that reported data on 6-TGN threshold levels, a pooled analysis of the first six studies showed a three-fold significantly higher rate of remission among patients with a 6-TGN level of greater than 230 to 260 pmol/8 × 108 red blood cells. Incorporation of 6-TGN metabolite measurement into the management regimen of patients receiving azathioprine or 6-MP therapy for IBD is not mandatory and it is a subject of continuing controversy. Currently, 6-TGN measurement appears to be most useful for identifying reasons for nonresponse to therapy and for suspected noncompliance. If used, metabolite levels should be determined at least two weeks following any dose adjustment to allow sufficient time for the metabolites to reach steady-state. Currently, it is recommended in the package insert and by the U.S. Food and Drug Administration (FDA) to determine TPMT genotype or phenotype before initiating therapy. The active metabolites, 6-TGNs, also are responsible for myelosuppression with therapy, and patients with TPMT mutation or decreased TPMT enzyme activity are more likely to experience this toxicity because of preferential shunting of 6-MP metabolism toward the excessive production of 6-TGN.201 Thus, identifying TPMT polymorphism before initiating azathioprine or 6-MP therapy can decrease the risk of myelotoxicity. Patients with homozygous wildtype TPMT or normal (to high) TPMT enzyme activity level may receive these agents starting at the weight-based optimal dose of 2.5 mg/kg/day for azathioprine or 1.5 mg/kg/day for 6-MP. It has been suggested by some investigators that in patients with heterozygous TPMT mutation or intermediate enzyme activity level, 6-MP or azathioprine should be started at 50% of the weight-based optimal dose. Alternative therapy should be considered in patients with homozygous mutations for TPMT. Regardless of whether a patient’s TMPT genotype or phenotype is known, continued frequent monitoring of complete blood counts remains necessary, because only 27% of all patients with leukopenia have TPMT mutations.202 In addition, two studies have reported that TPMT testing may be cost effective.203,204 Azathioprine and 6-MP therapy have a delayed onset of action. The mean time to clinical response with azathioprine or 6-MP therapy in patients with UC has been reported to be three to four months in uncontrolled studies,205,206 a

Table 112-15  Side Effects of Azathioprine and 6-Mercaptopurine Abnormal liver biochemical test results Bone marrow suppression Hypersensitivity reactions (fever, rash, arthralgia) Infections Lymphoma Nausea, abdominal pain, diarrhea Pancreatitis

figure that is similar to the 17 weeks’ response time to clinical benefit in placebo-controlled trials of azathioprine or 6-MP therapy for active Crohn’s disease.207 Intravenous loading of azathioprine at 40 mg/kg for 36 hours does not shorten the time required for a therapeutic response in patients with Crohn’s disease.202 Such practice presumably would have the same results if attempted in patients with UC. Because azathioprine or 6-MP therapy is associated with a number of potentially significant toxicities, its duration of therapy should be determined by weighing clinical benefit against these potential toxicities. The optimal duration of maintenance therapy with azathioprine or 6-MP currently is unknown in patients with UC. In patients with Crohn’s disease, the maintenance benefit of azathioprine or 6-MP can be observed for at least five years.208,209 Based on these data in Crohn’s disease and the paucity of alternative maintenance therapies, in patients with UC in whom remission is maintained with azathioprine or 6-MP, treatment generally is continued indefinitely as long as there is no significant adverse side effect. Common side effects of azathioprine and 6-MP therapy include nausea, vomiting, bone marrow suppression, pancreatitis, allergic reactions, and infections (Table 11215).210,211 Bone marrow suppression occurs in 2% to 5% of patients.210,212 It is dose dependent and manifests primarily as leukopenia, although all three cell lines may be affected. This hematologic toxicity can increase with concurrent use of sulfasalazine or mesalamine compounds.198,213-215 It is

1997

1998

Section X  Small and Large Intestine known that mesalamine can interact with the enzyme TPMT, leading to increased levels of 6-TGN, and that this interaction has been associated with leukopenia. Bone marrow suppression is managed by reducing the dosage of immunomodulator or withdrawing the medication. Routine monitoring of complete blood count with differentials is necessary for patients receiving azathioprine or 6-MP and should be continued for the entire duration of therapy. Allergic reactions to azathioprine or 6-MP usually manifest as fever, rash, and arthralgia and resolve following discontinuation of these medications.212,216 Recurrence of similar reactions occurs with medication challenge, although patients who develop allergic reactions to one agent may be able to tolerate subsequent challenge with the other.217 Pancreatitis also is idiosyncratic and independent of dosage.212,218,219 It usually occurs during the first month of therapy and is reversible upon withdrawal of the drug. Patients using azathioprine or 6-MP therapy can have abnormal liver biochemical tests, but these usually resolve following drug withdrawal.220 Because liver biopsy is not performed routinely in these patients, their pattern of hepatic injury, if any, is unknown. Cholestasis with inflammation, nodular regenerative hyperplasia, and peliosis hepatis have been reported with azathioprine and 6-MP therapy.212,220 As is the case for complete blood counts, routine monitoring of liver biochemical tests is recommended. An increased risk of malignancy, primarily lymphoma, has been reported, but not consistently. A metaanalysis of six studies examining this risk reported a fourfold elevated risk of lymphoma with 6-MP/azathioprine.221 The lymphoma that develops in patients who have IBD and receive these immunomodulatory agents appears to be associated with Epstein-Barr virus.222 Cyclosporine Cyclosporine A is a potent inhibitor of cell-mediated immunity. Its use in UC is primarily in patients with severe, steroid-refractory disease. There has only been one randomized, placebo-controlled trial evaluating the efficacy of intravenous cyclosporine in severe UC. In this study of 20 patients who did not respond to at least seven days of intravenous hydrocortisone, nine (82%) of the 11 patients receiving continuous intravenous infusion of cyclosporine at 4 mg/kg/day responded, compared with none of the nine patients receiving placebo therapy.223 The time to clinical response was rapid, at a mean of seven days. After the intravenous route of therapy was converted to oral cyclosporine, 44% of those patients who responded initially required colectomy during the sixmonth follow-up period. Intravenous cyclosporine monotherapy may be as effective as intravenous glucocorticoids in patients with severely active UC; its use thus potentially minimizes the toxicities of combination therapy.224 The addition of azathioprine or 6-MP in patients who have responded to intravenous cyclosporine has been shown in other studies to reduce the rate of relapse or colectomy.225,226 Thus, cyclosporine can be considered a bridge therapy to control active disease in patients with steroid-refractory UC while waiting for elective surgery or the onset of action of azathioprine or 6-MP. With the addition of azathioprine, long-term remission at one year may be more likely in patients who initially respond to intravenous cyclosporine monotherapy than in those who respond to intravenous glucocorticoids. A European retrospective cohort study of 142 patients who were treated with cyclosporine, 118 of whom responded initially, reported the probability of avoiding colectomy to be 63% at one year, 41% at four years, and 12% at seven years; overall,

54% of patients required colectomy at some point.227 Patients who were already taking 6-MP or azathioprine at the time cyclosporine was initiated continued taking their current dose, and those who were naïve to 6-MP or azathioprine were started at target doses at the time of response to cyclosporine during their hospitalization. The authors found that 59% of patients previously taking 6-MP or azathioprine required eventual colectomy, compared with 31% for patients naïve to these drugs (P < 0.05). Because most of the serious adverse effects associated with the use of cyclosporine are dose-dependent, intravenous doses lower than 4 mg/kg that still can achieve efficacy are desirable. One RCT has shown that a dose of 2 mg/kg is as effective as 4 mg/kg given intravenously in patients with severely active UC, judged by clinical response rates, time to response, and short-term colectomy rates.228 The mean plasma cyclosporine levels were 237 ng/mL in patients receiving the 2 mg/kg dose and 332 ng/mL in patients receiving the 4 mg/kg dose. Thus, initiating therapy at 2 mg/kg may be reasonable, but regardless of the dose used, careful monitoring of plasma cyclosporine trough levels is necessary. Cyclosporine has been associated with many adverse effects, including paresthesias, tremors, headache, hyper­ trichosis, and gingival hyperplasia (Table 112-16). Other potentially serious toxicities include hypertension, seizures, electrolyte and liver biochemistry abnormalities, nephrotoxicity, anaphylaxis, and opportunistic infections. These complications are mostly dose-dependent. Severe complications have been reported with cyclosporine in up to 12% of patients with UC,229 and two large series have reported death rates of 1.8% to 2.8% with cyclosporine, more than half of which were due to infections acquired while taking the drug.227,229 Careful monitoring for adverse effects is critical during cyclosporine therapy. Baseline serum electrolytes, creatinine, cholesterol, and liver biochemical values should be measured. Cyclosporine therapy should be avoided in patients with an impaired creatinine clearance to minimize the risk of severe nephrotoxicity. Patients with serum cholesterol lower than 120 mg/dL should receive nutritional support to improve the level before initiating cyclosporine therapy, because a low cholesterol level is associated with an increased risk of seizures. During intravenous therapy, cyclosporine levels should be monitored daily, and the dose should be adjusted to achieve a trough concentration (measured one hour before dosing) between 200 and 400 ng/mL, determined by high-pressure liquid chromatography. Serum electrolytes and serum creatinine levels should be monitored daily or every other day. The dose of cyclosporine also

Table 112-16  Side Effects of Cyclosporine Anaphylaxis Diarrhea Electrolyte abnormalities Gingival hyperplasia Headache Hepatotoxicity Hirsutism Hypertension Infections Nausea, vomiting Opportunistic infections Paresthesia Renal insufficiency Seizure Tremor

Chapter 112  Ulcerative Colitis should be decreased when the serum creatinine increases by 20% to 30% over baseline. If patients respond to intravenous cyclosporine, the route of administration can be changed to oral therapy with 2 mg of oral agent for each 1 mg of intravenous cyclosporine. The drug can be administered in two divided doses daily. Drug monitoring during oral cyclosporine therapy includes weekly trough cyclosporine levels and weekly to biweekly electrolyte and creatinine levels. Oral cyclosporine should be continued for three to six months, while waiting for surgery or for azathioprine or 6-MP to take effect. Patients on long-term cyclosporine therapy should receive Pneumo­ cystis carinii pneumonia prophylaxis with trimethoprimsulfamethoxazole. Methotrexate Methotrexate is a folic acid antagonist and has antime­ tabolite and anti-inflammatory properties. Although early reports suggested potential benefit of methotrexate administered intramuscularly or orally in UC, the only randomized, placebo-controlled trial failed to demonstrate its efficacy for the treatment of active UC.230 In this study of 67 patients with chronic active UC, oral methotrexate at 12.5 mg/wk for nine months was comparable to placebo therapy in the rate of achieving first remission, time to first remission, relapse following remission, and the mean glucocorticoid dose. It is unknown if methotrexate at higher doses administered intramuscularly or subcutaneously may be beneficial in inducing or maintaining remission in UC. Given the absence of data supporting its efficacy, metho­ trexate cannot at this time be considered a standard therapy for UC. Other Immunomodulators Alternative immunomodulators have been explored for patients who do not tolerate or have not responded to the previously mentioned immunosuppressants. Mycophenolate mofetil has pharmacodynamic properties similar to those of azathioprine and 6-MP but a more rapid onset of action. A pilot study of patients with chronic active UC receiving concomitant prednisolone found azathioprine to be superior to mycophenolate mofetil throughout the oneyear study period, with remission rates at one year of 100% and 88%, respectively.231 Uncontrolled studies reported less than 50% remission rates with mycophenolate mofetil therapy in patients with steroid-dependent UC232,233 and the intolerance rate was high.232 A substantial number of patients developed adverse effects necessitating drug withdrawal, including recurrent upper respiratory tract infection, bacterial meningitis, depression, and migraine headache.231,233 Tacrolimus is another immunosuppressant with actions similar to those of cyclosporine. In contrast to cyclo­ sporine, it has a 100-fold greater potency and a more-rapid onset of action. A number of small uncontrolled studies have suggested benefit of oral or intravenous tacrolimus for the treatment of patients with refractory UC. The only randomized, placebo-controlled trial of tacrolimus in UC involved 63 Japanese patients with either steroiddependent or steroid-refractory disease who were randomized to receive either initial oral tacrolimus at 0.05  mg/kg or placebo twice daily.234 Patients in the high-trough concentration (10 to 15  ng/mL) tacrolimus group had a significantly higher rate of response and nonsignificantly higher rate of remission than those in the placebo group at week two, and a number of patients demonstrated response or remission (or both) after an additional 10 weeks of openlabel therapy. As with cyclosporine, tacrolimus can result in a number of toxicities including nephrotoxicity, electro-

lyte abnormalities, nausea, diarrhea, headache, tremors, paresthesias, insomnia, alopecia, hirsutism, and gingival hyperplasia.235,236 Thus, given the limited data and potential for harmful adverse events, the use of these alternative immunomodulators currently is not incorporated into standard practice.

Antibiotics

Antibiotics have a limited role in the management of UC, and most controlled studies have not demonstrated their benefit either in active disease or maintenance of remission.237-241 The most commonly used antibiotics in this setting are metronidazole and ciprofloxacin. One RCT found oral tobramycin to be superior to placebo as a short-term adjunctive therapy to glucocorticoids for active UC.242 Another RCT reported a modest benefit for the addition of ciprofloxacin for six months in patients with UC refractory to mesalamine and corticosteroids.243 At present, the data showing efficacy of antibiotics for treatment of patients with UC are not as convincing as are the data for antibiotic treatment of Crohn’s disease. Thus, at present the primary role of antibiotics in the treatment of UC is in the management of its suppurative complications.

Probiotics, Prebiotics, and Synbiotics

Probiotics are living organisms in foods and dietary supplements that might beneficially affect the host in a number of ways, including improving its intestinal microbial balance, blocking adhesion sites on colonocytes (which might improve mucosal barrier function), and enhancing local immune response.39,244 A probiotic can be a specific nonpathogenic strain of a bacterial species or a mixture of multiple species and strains, most commonly including Lactobacillus or Bifidobacterium species; sometimes they contain fungal antigens as well. An example of a common probiotic is VSL#3, which contains four strains of Lactoba­ cillus (Lactobacillus acidophilus, Lactobacillus delbrueckii subspecies bulgaricus, Lactobacillus plantarium, and Lactobacillus casei), three strains of Bifidobacterium (Bifidobacterium infantis, Bifidobacterium longum, Bifido­ bacterium breve), and one strain of Streptococcus (Strepto­ coccus salivarius subspecies thermophilus). Prebiotics are nondigestible food ingredients that selectively stimulate the growth or activity of one or more organisms of the intestinal microbiota, such as Lacto­ bacillus or Bifidobacterium species, thereby potentially conferring beneficial effects to the host.245,246 The majority of prebiotics are nondigestible oligosaccharides, with galacto-oligosaccharide, fructo-oligosaccharide, lactulose, and inulin being the most commonly used agents. Because probiotics have the challenge of competing with indigenous microbiota for nutrients, scientists have developed synbiotics, which are combinations of probiotics and prebiotics, in the hope of facilitating the survival of probiotics in the intestines. With respect to the use of these agents for inducing remission in mildly to moderately active UC, four RCTs have been performed using different agents.247-250 Two of three studies that measured rates of remission found no benefit of probiotics (VSL#3 in one study, fermented milk in the other) added to 5-aminosalicylates247,248; the third study found that E. coli Nissle 1917 combined with glucocorticoids had efficacy similar to that of mesalazine combined with glucocorticoids.249 The fourth study, which used a synbiotic, reported a nonsignificant improvement in disease activity when the synbiotic was combined with standard therapy.250 With respect to the use of these agents for the maintenance of remission in mildly to moderately

1999

2000

Section X  Small and Large Intestine active UC, six RCTs have been published.249,251-255 Two of these studies reported significantly lower rates of relapse for patients receiving a probiotic (Bifidobacterium in one study, fermented milk in the other) after medically induced remission compared with those receiving placebo,251,252 and the other four studies (using E. coli Nissle in three studies and Lactobacillus rhamnosus strain GG in the fourth) found no difference in rates of relapse.250,253-255 Nontraditional probiotic therapies that also have been evaluated include Saccharomyces boulardii and Trichuris suis.256,257 A small, uncontrolled study of 24 patients with mild to moderate active UC suggested a potential benefit of Saccharomyces boulardii when used in addition to mesalamine.256 The use of helminths in active UC was investigated by Weinstock and colleagues, who randomized 54 patients with active disease to receive 2500 T. suis ova or placebo orally every 2 weeks for 12 weeks and reported that rates of improvement were significantly higher in the active treatment group at week 12 (43% vs. 17%, P = 0.04); significant improvement was seen as early as week six.257 In summary, at present, there is no convincing evidence to support the use of probiotics, prebiotics, or synbiotics for the treatment of UC. However, future large well-designed RCTs are necessary to address this issue more definitively.

Nutritional Therapy

Short-chain fatty acids, especially butyrate, have been shown to be the main energy substrate for colonocytes. Butyrate metabolism accounts for approximately 70% of colonocyte oxygen use. The suggestion that there is an impairment of colonocyte oxidation of short-chain fatty acids in UC led to therapeutic investigations of this form of nutritional therapy. Indeed, placebo-controlled studies have found butyrate enemas to be beneficial in treating mildly active, left-sided colitis.258-260 Fish oil containing eicosapentaenoic acid has been found to attenuate colitis in animal models of colitis, probably by protecting the integrity of colonic mucosa, suppressing the inflammatory response, or both.261-263 In a small, placebocontrolled, crossover study of patients with mild to moderate UC, treatment with fish oil resulted in a 56% reduction in disease activity compared with a 4% reduction in controls (P < 0.05).264 This benefit has not been confirmed in other studies, and a benefit in maintaining remission has not been observed.265-267 Furthermore, compliance is limited because of side effects and the odor of the fish oil preparation. In contrast to Crohn’s disease, where bowel rest and total parenteral nutrition can improve disease, multiple studies have not found total parenteral nutrition with or without bowel rest to have any therapeutic advantage in patients with UC.268,269 Parenteral nutrition, however, can offer nutritional benefit in these patients. In general it is important to provide adequate nutrition to patients with UC who are about to undergo surgery. Nutrition is no more effective than placebo, however, for use as primary therapy of active UC.

Nicotine

Based on the observation that smoking is associated with a decreased risk of developing UC and that a former smoker with active colitis may gain clinical benefit on resuming smoking, nicotine has been used to treat patients with this disease. RCTs have shown some benefit of transdermal nicotine in the treatment of active UC.270-273 When administered at the highest tolerated dosage of 22 mg/day or less for four weeks in patients with mildly to moderately active UC, transdermal nicotine resulted in clinical improvement in 39% of patients compared with 9% of patients who received

placebo therapy (P = 0.007).270 As a single therapy, however, transdermal nicotine was not as effective as low-dose prednisolone.274 Common side effects included nausea, lightheadedness, itching, and tremor. Topical nicotine therapy has fewer side effects and may be an alternative. Pilot studies have shown topical nicotine to be beneficial in patients with distal UC, but no large RCT has been performed,275,276 and transdermal nicotine has not been found to be effective as a maintenance therapy.277 Thus, based on available data on clinical efficacy and the overall poor patient tolerability, nicotine cannot be considered part of the standard treatment for patients with UC.

Heparin

Heparin, a group of sulfated glycosaminoglycans, has antiinflammatory and immunomodulatory properties in addition to its well-known anticoagulant activity. The exact mechanism whereby heparin might ameliorate UC remains uncertain. An anticoagulant benefit, however, might not be responsible, because similar efficacy has not been observed in patients with IBD when treated with warfarin. Because of their negative charge, the glycosaminoglycans that constitute heparin have varied biological effects, including significant anti-inflammatory actions and augmentation of the peptide growth factors involved in intestinal mucosal repair and regeneration. Based on reports of fortuitous improvement in patients with UC receiving heparin for treatment of deep venous thromboses, pilot studies have suggested that unfractionated heparin may be effective for inducing remission in patients with severe, refractory UC.278,279 Compared with glucocorticoids as a first-line therapy, however, small RCTs have reported conflicting results.280,281 Intravenous heparin therapy has been associated with substantial bleeding complications. Low-molecular-weight heparin (LMWH) offers advantages over unfractionated heparin in its subcutaneous route of administration, and preliminary studies suggested a benefit of LMWH in the treatment of active UC.282,283 Unfortunately, this finding was not confirmed in a large, placebo-controlled trial of patients with mildly to moderately active UC receiving LMWH for six weeks.284 At this time, the use of either unfractionated heparin or LMWH cannot be advocated as primary therapy for patients with active UC.

Biological Therapy

Recent advances in our understanding of the pathogenesis of IBD have resulted in the development of therapies targeted at specific molecules or mediators involved in the inflammatory processes of these diseases. Most studies evaluating the efficacy of these agents have been performed in patients with Crohn’s disease, and only limited data are available for patients with UC. Anti-Tumor Necrosis Factor Antibodies TNF is a key proinflammatory cytokine that has been demonstrated to play a role in several disease states, including IBD. Elevated TNF concentrations have been found in inflamed intestine in patients with Crohn’s disease and UC, and stool and mucosal concentrations of TNF in patients with IBD have been shown to correlate with clinical disease activity. Infliximab (Remicade) is a chimeric monoclonal antibody of IgG1 subclass directed against human TNF-α. It consists of 75% human and 25% murine components (Fig. 112-14). The efficacy of infliximab in Crohn’s disease is well established, and it is approved by the FDA to treat Crohn’s disease and UC. Infliximab is thought to operate in Crohn’s disease via a multitude of mechanisms, including

Chapter 112  Ulcerative Colitis Human IgG antibody

Chimeric human/mouse antibody

Variable

Constant

Fc

Fc portion

infliximab Mouse component Figure 112-14.  Structural diagram of anti–tumor necrosis factor antibody. Normal human immunoglobulin G (IgG) antibody is shown on the left and infliximab is shown on the right. Infliximab is 75% human and 25% murine.

antagonizing the activity of TNF-α,285,286 initiating cytotoxicity on immune cells,287 and inducing T-cell apoptosis.288,289 Results from two large, multicenter, randomized, doubleblind, placebo-controlled trials (ACT 1 and 2) showed efficacy of infliximab therapy in UC.290 In these two similarly designed trials, 728 patients with moderately to severely active UC who failed conventional therapy with glucocorticoids alone or in combination with thiopurines (ACT 1) or glucocorticoids alone or in combination with thiopurines and 5-aminosalicylates (ACT 2) were randomized to placebo, infliximab 5 mg/kg, or infliximab 10 mg/kg at weeks 0 and 2 and then every eight weeks through week 46 (ACT 1) or week 22 (ACT 2). With respect to clinical response at week 8, in ACT 1 69% and 61% of patients receiving infliximab at 5 and 10 mg/kg, respectively, had a clinical response, compared with 37% of patients receiving placebo (P < 0.001 for both comparisons). In ACT 2 at week 8, 64% and 69% of patients receiving infliximab at 5 mg/kg and 10 mg/kg, respectively, had a clinical response, compared with 29% of patients receiving placebo (P < 0.001 for both comparisons). With respect to clinical remission at week 8 in ACT 1, 39% and 32% of patients receiving infliximab at 5 mg/kg and 10 mg/kg, respectively, attained remission, compared with 15% of patients receiving placebo (P < 0.003 for both comparisons). In ACT 2 at week 8, 34% and 28% of patients receiving infliximab at 5 mg/kg and 10 mg/kg, respectively, attained remission, compared with 6% of patients receiving placebo (P < 0.001 for both comparisons). The results for clinical remission at week 30 (ACT 1 and 2) and week 54 (ACT 1) were very similar for all groups, with highly significant greater than two-fold higher remission rates for the infliximab-treated patients. The proportions of patients with a sustained clinical response or remission also were sig­ nificantly higher in the infliximab groups. Treatment with in­fliximab also was shown to have steroid-sparing and mucosal healing properties. These data have led to the approval of infliximab by the FDA for patients with moderately to severely active UC who have had an inadequate response to conventional therapy. Infliximab is now accepted as part of the standard treatment options in patients with UC. Two other anti-TNF agents, adalimumab and certolizumab pegol, have shown efficacy for the induction and maintenance of remission in Crohn’s disease but have not yet been studied in patients with UC.

Anti-Adhesion Molecules Several agents directed at blocking small adhesion molecules have been evaluated for the treatment of UC. These molecules are glycoproteins expressed on the surfaces of endothelial cells and lymphocytes. Adhesion molecules are important in cellular trafficking in IBD and other diseases, in which immune and inflammatory cells from the periphery are recruited into sites of inflammation. Among these, natalizumab is a humanized IgG4 monoclonal antibody against lymphocyte adhesion molecules, α4 integrins. A pilot study of 10 patients with active UC suggested clinical benefit with a single infusion of 3 mg/kg of natalizumab.291 Natalizumab currently is approved for treating patients who have Crohn’s disease and in whom anti-TNF therapy has failed; its use in UC is now undergoing evaluation. Another anti-adhesion molecule agent is MLN-02 (formerly called LDP-02), a humanized IgG1 monoclonal antibody to α4β7 integrin. In a phase 2 study, two infusions of 0.5 mg/kg of MNL-02 administered 29 days apart were found to be effective in achieving clinical remission and response at six weeks after the initial infusion in patients with moderately active UC.292 Others Although historically considered more important in the inflammation of Crohn’s disease, as it is produced by Th1 cells, IL-2 also has been implicated in UC inflammation (see earlier). Two agents designed to block the binding of IL-2 to its receptor have been examined for potential efficacy in UC. Daclizumab, a humanized monoclonal antibody against the IL-2 receptor (CD25), was suggested to be beneficial in patients with refractory UC in a small open-label pilot study.293 A potential clinical benefit also has been reported with basiliximab, a chimeric monoclonal antibody to the IL-2 receptor, in a small, uncontrolled study of patients with steroid-refractory UC.294 Along with the emphasis on T-cell– mediated immune response in the pathogenesis of UC, a humanized monoclonal antibody to CD3, visilizumab, has shown promise in an open-label phase 1 study of 32 hospitalized patients with UC whose disease failed to respond to intravenous glucocorticoids.295 Other biological therapies include agents targeted at tissue repair and restitution following mucosal injury. In this regard, epidermal growth factor (EGF) is a potent mitogenic peptide that stimulates cell proliferation in the gastrointestinal tract. A preliminary study showed that EGF enemas at a dose of 5 µg daily for two weeks was effective in treating mild to moderate left-sided UC when administered along with oral mesalamine.296 In contrast, another potent stimulant of intestinal epithelial cells, repifermin (keratinocyte growth factor 2) was not found to be more effective than placebo when administered intravenously in patients with active UC in a phase 2 dose-ranging study.297 Further studies clearly are necessary to confirm some of these early promising findings.

Cytapheresis

Active UC is characterized by activation and infiltration of leukocytes in the colonic mucosa. Because leukocytederived inflammatory cytokines play an important role in initiating and perpetuating the inflammatory process, reduction of peripheral blood levels of leukocytes has been proposed as a therapeutic option for treating UC. Several methods of depleting peripheral blood leukocytes have been developed and have been shown to hold promise in the treatment of severely active UC in small controlled and uncontrolled studies.298-305

2001

2002

Section X  Small and Large Intestine Granulocyte/monocyte apheresis using the Adacolumn Apheresis System (JIMRO, Ltd., Takasaki, Japan) in patients with moderate to severe active UC was formally evaluated in a large multicenter randomized, double-blind, shamcontrolled North American pivotal trial of 168 patients and a smaller identically designed companion trial of 47 patients in Europe and Japan.306 In this study, patients were required to have active disease despite concomitant therapy with 5-aminosalicylates, glucocorticoids, or 6-MP and azathioprine. Neither the pivotal nor the companion study showed a difference in clinical remission or response rates (as defined by the Mayo score) at week 12 between the Adacolumn and sham treatment groups after a total of 10 apheresis sessions over a nine-week period. Thus, at this time, Adacolumn apheresis in similar patients cannot be recommended.

Peroxisome Proliferator Receptor Agonists

Peroxisome proliferator-activated receptor (PPAR)-γ is a nuclear hormone receptor best known for its roles in regulating metabolism and adipocyte differentiation. It also has been shown to have immunomodulatory and antiinflammatory properties in multiple sites, including the colon.307-310 A randomized double-blind, placebo-controlled trial assessed the efficacy of a 12-week treatment with the

PPAR-γ ligand rosiglitazone in 105 patients with mildly to moderately active UC.311 The authors reported a significant benefit of rosiglitazone over placebo with respect to clinical response (44% vs. 23%, P = 0.04) and clinical remission (17% vs. 2%, P = 0.01) at 12 weeks based on the Mayo score. However, given the reports of a potential for increased risk of myocardial infarction and long-bone fractures with ro­ siglitazone (based on studies in diabetic patients, in whom this medication is mostly used), use of rosiglitazone should be restricted to patients who have failed or cannot tolerate standard medical therapy.

ALGORITHMS FOR THE TREATMENT OF ULCERATIVE COLITIS

An algorithm for treating patients with active UC of mild to moderate severity is outlined in Figure 112-15. The treatment of severely active UC is shown in Figure 112-16.

SURGICAL

Removal of the colon and rectum cures UC. Common indications for surgical therapy of UC are medically refractory disease, intractable disease with impaired quality of life, and unacceptable side effects from medical therapy (Table 112-17). Other indications include uncontrolled bleeding,

Topical and/or oral 5-ASA

Good response Maintenance therapy with 5-ASA

Poor response

Good response

Add topical and/or oral glucocorticoids Poor response

Taper glucocorticoids Successful taper Maintenance therapy with 5-ASA

IV glucocorticoids and/or 6-MP or AZA Unsuccessful taper

Good response

Poor response

More prolonged taper; add 6-MP or AZA; maintain with 6-MP or AZA (±5-ASA)

Consider IV cyclosporine or infliximab

Poor response Surgery

Taper glucocorticoids; maintain with 5-ASA and/or 6-MP or AZA

Good response

Oral cyclosporine or infliximab and oral glucocorticoids

Taper glucocorticoids: add 6-MP or AZA

Maintain on 6-MP or AZA (± 5-ASA); discontinue cyclosporine within 6 mo; continue infliximab Figure 112-15.  Algorithm for the management of mildly to moderately active ulcerative colitis. 5-ASA, 5-aminosalicylate; AZA, azathioprine; IV, intravenous; 6-MP, 6-mercaptopurine.

Chapter 112  Ulcerative Colitis Hospitalization; oral glucocorticoids or IV glucocorticoids

Good response

Poor response

Taper glucocorticoids; maintenance therapy with 5-ASA; consider adding 6-MP or AZA

Successful taper Continue maintenance therapy

IV glucocorticoids

Poor response

Good Unsuccessful response taper More prolonged taper; add 6-MP or AZA; maintain with 6-MP or AZA (±5-ASA)

Convert to oral glucocorticoids followed by tapering; add 5-ASA; consider adding 6-MP or AZA

IV cyclosporine or IV infliximab Poor response Surgery

Unsuccessful taper Longer glucocorticoid taper; maintain with 6-MP or AZA (±5-ASA)

Successful taper Maintain with 6-MP or AZA (±5-ASA)

Good response Oral cyclosporine and oral glucocorticoids; or maintain with infliximab

Taper glucocorticoids; add 6-MP or AZA

Maintain on 6-MP or AZA (±5-ASA); discontinue cyclosporine within 6 mo. Figure 112-16.  Algorithm for the management of severely active ulcerative colitis. 5-ASA, 5-aminosalicylate; AZA, azathioprine; IV, intravenous; 6-MP, 6-mercaptopurine.

Table 112-17  Indications for Surgery in Patients with Ulcerative Colitis Colonic dysplasia or carcinoma Colonic hemorrhage, uncontrollable Colonic perforation Growth retardation Intolerable or unacceptable side effects of medical therapy Medically refractory disease Systemic complications that are recurrent or unmanageable Toxic megacolon

toxic megacolon, perforation, dysplasia or carcinoma, systemic complications, and growth retardation. The goals of surgical treatment are to remove the entire diseased colon while preserving continence and sexual function. Elimination of the risk of colorectal cancer is also important. The role of prophylactic proctocolectomy in patients with long-standing extensive UC is controversial. Whereas most clinicians do not routinely recommend proctocolectomy solely for the purpose of prophylaxis against colorectal cancer, patients should be informed of the limitations of our current colonoscopic surveillance program (see Chapters 122 and 123). There are multiple surgical options for UC (see Chapter 113), including subtotal colectomy with ileostomy, colectomy with ileorectal anastomosis, proctocolectomy with Brooke ileostomy, proctocolectomy with continent ileostomy, restorative proctocolectomy with ileal pouch-anal anastomosis (IPAA), and proctocolectomy with ileal pouchanal transition zone anastomosis (Fig. 112-17). The choice of operation is based on several factors, including the indica-

tion for and urgency of surgery, the age and general health of the patient, the status of the patient’s anal function, and the patient’s preference of functional outcome and lifestyle.

Colectomy

Subtotal colectomy with ileostomy is the least extensive of these operations. Most of the colon is removed, a Hartman pouch or a mucus fistula is created for the remaining colon, and an end-ileostomy is created. This surgery typically is performed in patients requiring emergent surgery for severe or fulminant colitis, and it has the advantage of allowing restorative surgery in the future. Colectomy with ileorectal anastomosis is similar to subtotal colectomy with ileostomy, but it maintains bowel continuity. Many patients continue to have attacks of proctitis and the retained rectal stump is at risk for developing colorectal cancer. Thus, lifelong endoscopic surveillance of the rectum is necessary for patients who elect this type of operation. Total proctocolectomy with a permanent Brooke endileostomy was one of the earliest operations performed for UC. Removal of the entire colon and rectum eliminates any future disease and risk of colorectal cancer. The primary disadvantage of this operation is the presence of the permanent ileostomy, which might not be acceptable from the standpoint of quality of life for some patients. This is the operation of choice for elderly patients, those with anal dysfunction, and those who do not wish to have a restorative proctocolectomy. Proctocolectomy with continent ileostomy (Koch pouch) was developed as an alternative to the conventional end-ileostomy.312 In this operation, loops of small bowel are used to create an intra-abdominal pouch with an intussuscepted (nipple) valve. This pouch allows storage of stool contents and is attached to the abdominal

2003

2004

Section X  Small and Large Intestine

A

B

C

D

E

Figure 112-17.  Schematic diagrams of various surgical options for the management of ulcerative colitis. A, Conventional (Brooke) ileostomy with a subtotal colectomy and a Hartman pouch. B, Subtotal colectomy with ileorectal anastomosis. C, Conventional (Brooke) ileostomy with a total proctocolectomy. D, Continent ileostomy (Koch pouch) with total proctocolectomy. E, Restorative proctocolectomy with ileal pouch anal anastomosis (see Chapter 113). (A-E, Adapted from Blumberg D, Beck DE. Surgery for ulcerative colitis. Gastroenterol Clin North Am 2002; 31:219.)

wall with a flush ostomy opening. The stool contents in the pouch are emptied by inserting a suction-catheter through the stoma. Because of technical challenges associated with this operation (e.g., slippage of the nipple valve) and the development of restorative procedures, proctocolectomy with continent ileostomy rarely is performed.

Proctocolectomy with Ileal Pouch-Anal Anastomosis

Restorative proctocolectomy with IPAA currently is the operation of choice for most patients with UC who require elective colectomy. In this procedure, the entire colon and rectum are removed, the anal sphincters are preserved, and a pouch is constructed from approximately 20 cm of the distal ileum (see Fig. 112-16). Bowel continuity is established by anastomosing this pouch with the anal canal. An IPAA usually is performed as a two-stage operation, during the first stage of which a temporary diverting ileostomy is created to allow the ileal pouch to heal. This operation can be performed as a single-stage operation; however, there have been reports suggesting a higher rate of bowel obstruction and sepsis when this is done.313 The ileostomy is reversed after approximately two to four months. Proctocolectomy with IPAA presents technical challenges and might not be suitable or technically feasible for all patients. Most reports suggest satisfactory quality of life following IPAA surgery.314 Mean stool frequency ranges from four to nine bowel movements per day, including one or two nocturnal stools. Nocturnal seepage occurs in approximately 20% of patients in the early postoperative period but is infrequent after the first year. Rates of complications following IPAA surgery vary widely. In a series from the Cleveland Clinic of more than 1000 patients undergoing restorative proctocolectomy and IPAA, most of whom had UC, the overall morbidity rate was 63% (early complications 28%, late complications 51%).314 A larger case series from the Mayo Clinic included 1885 patients who had proctocolectomy with IPAA for UC and who were followed for up to 20 years (mean follow-up 11 years).315 Pouch success rates were reported of 96.3%, 93.3%, 92.4%, and 92.1% by 5, 10, 15, and 20 years, respectively, however, complication rates also were high and included pouchitis (48% by 10 years, 70% by 20 years), small bowel obstruction (42% by 20 years), anastomotic stricture (39% by 20 years), abscess (16% by 20 years), and

fistula (14% by 20 years), as well as pelvic sepsis. Other complications of proctocolectomy with IPAA are covered in Chapter 113 and include fecal incontinence and sexual and urinary dysfunction. This surgery is best performed in centers with considerable experience with the operation and with managing pouch dysfunction. These high rates of complications will likely decrease over time as experience grows with this type of surgery. A widely performed modification of proctocolectomy with IPAA is proctocolectomy with ileal pouch-anal transition zone anastomosis. This technically less-complex surgery involves stapling the ileal pouch to the distal rectum in close proximity to the dentate line (1 to 4 cm), thereby eliminating the need to perform rectal mucosectomy. This type of surgery might carry a lower risk for fecal incontinence and may be performed as a single-stage operation without a temporary diverting ileostomy. Laparoscopic approaches to the proctocolectomy with IPAA have become popular among both physicians and patients. Because this modality is relatively new, comparison data versus open proctocolectomy with IPAA are limited. A case-matched series from the Mayo Clinic, however, which included 100 consecutive laparoscopic IPAA case patients matched to 200 open IPAA control patients, reported that although the median operative time was significantly longer for the laparoscopic group by 103 minutes, laparoscopic patients had significantly improved recovery, as witnessed by lower median time to regular diet (three vs. five days), time to ileostomy output (two vs. three days), length of stay (four vs. seven days), and intravenous narcotic use (all P < 0.05).316 In addition, postoperative morbidity and hospital readmission rates were similar between the two groups and no mortalities were observed. Fewer patients required reoperation within three months with the laparoscopic approach (3% vs. 6.5%, P < 0.2). A follow-up survey in this same cohort of patients one year after operation showed that patients reported high cosmetic, body image, quality of life, and sexual function scores irrespective of the type of operation.317 A smaller randomized trial from Amsterdam of 60 patients who underwent laparoscopic or open IPAA found significantly higher operative time for the laparoscopic approach (by 77 minutes) but similar morbidity (20% vs. 17%), length

Chapter 112  Ulcerative Colitis of stay (10 vs. 11 days), time to regular diet, narcotic requirement, and quality of life at three months after surgery in both groups.318 Larger RCTs will need to be conducted before any definitive conclusions can be made regarding the superiority of one approach over another. The experience of the particular surgeon and surgical center is likely to be an important determinant of success rates irrespective of procedure.

SPECIFIC COMPLICATIONS TOXIC MEGACOLON

Toxic megacolon is defined as acute colonic dilatation with a transverse colon diameter of greater than 6 cm (on radiologic examination) and loss of haustration in a patient with a severe attack of colitis.319,320 Maximal colonic dilatation most commonly is observed in the transverse colon. This complication of UC results from extension of colonic inflammation beyond the mucosa to the underlying tissues, including the muscularis propria. Loss of contractility from the inflammatory reaction leads to the accumulation of gas and fluid within the lumen and subsequent colonic dilatation. Toxic megacolon occurs in approximately 5% of severe flares of UC. It often is encountered early in the course of the disease and may be the initial presentation of UC. Nearly 50% of patients with toxic megacolon develop this com­ plication within three months of their diagnosis.321 Toxic megacolon usually occurs in patients with extensive colitis, but patients with disease limited to the left colon also can develop it. Precipitating factors for toxic megacolon include electrolyte imbalance (particularly hypokalemia), use of antimotility drugs including anticholinergic agents and narcotics, and procedures such as barium enema and colonoscopy performed during a severe attack.322 These procedures should be avoided in the presence of a severe flare of UC. With clinical deterioration, patients can develop fever, tachycardia, hypotension, diffuse abdominal distention and tenderness, and decreased bowel sounds. Other laboratory parameters reflecting progressive severe systemic inflammation include marked leukocytosis, metabolic alkalosis, and electrolyte disturbances. Medical management for toxic megacolon is directed at treating the underlying inflammation, restoring colonic motility, and preventing free colonic perforation. If colonic dilatation occurs during the initial presentation of UC, intravenous glucocorticoids and fluid replacement should be initiated and electrolyte abnormalities should be corrected. Reduction of fluid and air within the gastrointestinal tract may be achieved through bowel rest and nasogastric decompression. Other conservative management approaches include maneuvers to reduce abdominal distention by allowing the redistribution or passage of colon gas. Gas tends to rise and fill the most superiorly located bowel segment (the transverse colon) if the patient is in the supine position. Encouraging the patient to move about, rotating a patient who is bed-bound, using the knee-elbow position when the patient is prone, and inserting a rectal tube all have been suggested to be helpful in reducing bowel distention.323,324 Systemic antibiotics often are empirically administered, because mortality from toxic megacolon correlates with the development of sepsis.325 Approximately 50% of acute dilatation resolves with medical therapy.324,326 However, because the presence of colonic perforation is the most important predictor of mortality (44% in patients undergoing emer-

gent colectomy after perforation compared with 2% in patients undergoing colectomy without perforation),327 an important aspect of management is to determine the optimal time for surgical intervention. In general, patients who do not improve after 48 to 72 hours of medical therapy should undergo surgery.324 Close clinical observation for signs of impending perforation is critical. Patients with progressive abdominal distention, development of rebound tenderness, or hemodynamic instability should undergo immediate colectomy. For patients who achieve remission on medical therapy, subsequent management is controversial. In one series, nearly 50% of patients treated successfully for toxic megacolon eventually required colectomy for intractable disease.328 Thus, some clinicians recommend elective colectomy following resolution of toxic megacolon. Free perforation also can develop in the absence of colonic dilatation. This complication is rare, occurring in 1% of patients with UC without toxic megacolon.326,329 Classic physical findings of peritonitis may be absent, largely because of the masking effect of administered glucocorticoids, but most patients have a marked deterioration in overall clinical condition after perforation. It is important to examine the abdomen for hepatic dullness every day in patients who have severe colitis and are taking high-dose glucocorticoids because they might have a free perforation and not have classic signs of peritonitis. A daily plain film of the abdomen also is recommended for the same purpose. As with toxic megacolon, patients with extensive colitis appear to be at greatest risk for this complication. The segment most at risk for free perforation is the sigmoid colon. The mortality associated with free perforation in UC patients without toxic megacolon has been reported to be more than 50%.329 Thus, the possibility of free perforation must be considered in patients with fulminant UC, particularly if there is deterioration in general condition, even in the absence of colonic dilatation.

STRICTURES

Colonic strictures complicate UC in approximately 5% of patients, most commonly in those with extensive and longstanding colitis.330 Patients with colonic strictures usually present with alterations in bowel habits, both constipation and diarrhea. Clinically significant obstruction is rare. Colonic strictures complicating UC typically are short (2 to 3 cm), occur distal to the splenic flexure, and represent hypertrophy and thickening of muscularis mucosa rather than fibrosis.330 There needs to be a high index of suspicion of malignancy in patients with colonic strictures associated with UC, especially when the strictures are located proximal to the splenic flexure.331 One series reported malignancy in 24% of colonic strictures in patients with UC.332 Moreover, in patients with UC, cancer associated with strictures tends to be more advanced than cancers not associated with strictures. Endoscopic appearance cannot reliably distinguish benign strictures from malignant strictures, and multiple biopsies are recommended at colonoscopy. Because carcinoma might not be detected on mucosal biopsies, surgical resection of the stricture is advised, particularly in patients with longstanding disease.

DYSPLASIA AND COLORECTAL CANCER

(see Chapter 123) Patients with UC have an increased risk of colorectal cancer. This risk depends on several factors, the most important being the duration and extent of the disease. Other risk

2005

2006

Section X  Small and Large Intestine factors include PSC, family history of colon cancer, age at diagnosis of disease, severity of inflammation, presence of pseudopolyps, and possibly backwash ileitis.332-337 The incidence of colon cancer in UC varies depending primarily on the duration and extent of the disease, but it has been estimated at approximately 7% to 10% at 20 years of disease and as high as 30% after 35 years of disease.338 Thus, in general, the risk of colorectal cancer may be estimated to increase within the range of 0.5% to 1.0% per year after 8 to 10 years of disease in patients with extensive UC.338 Although prophylactic colectomy can virtually eliminate the risk of colorectal cancer, patients often are unwilling to undergo surgery, particularly if there is no other indication for colectomy. Thus, colonoscopic surveillance programs have been developed in an effort to reduce the risk of colorectal cancer associated with UC. The primary goal of surveillance colonoscopy is to detect dysplasia, defined as unequivocal neoplastic epithelium, because currently it is the most important marker to detect concurrent or subsequent cancer. Dysplasia can be classified histologically and endoscopically into several groups, each with different prognostic implications. Histologic assessment is classified as negative, indefinite, and positive for dysplasia. Two grades of dysplasia are recognized: low (Fig. 112-18A) and high (see Fig. 112-18B).339 Endoscopically, dysplasia can be characterized as flat or raised based on the appearance of the surface of the dysplastic area. Raised dysplasia is also termed dysplasia-associated lesion or mass (DALM) and can be found as a polypoid lesion, mass, plaque, or stricture.

A

B Figure 112-18.  Photomicrograph of a colonic biopsy specimen showing the histologic features of dysplasia. A, Low-grade dysplasia is characterized by nuclear enlargement, crowding, and hyperchromasia in the colonic epithelial cells. Nuclei are stratified but remain in the basal half of the cells. There is some depletion of mucin. B, In high-grade dysplasia, the changes are more pronounced. Nuclei are stratified to the surface, and there is a marked increase in nuclear pleomorphism. Branching of the glands in a cribriform pattern and scattered cell necrosis are shown. No mucincontaining goblet cells are evident. (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine, 1996.)

Historically, it had been thought that flat dysplasia represented most of the dysplasia detected in patients with UC. These flat dysplastic lesions are difficult to see on standard white-light colonoscopy, but they may be detected with greater accuracy by the addition of magnification chromoendoscopy or narrow band imaging (see later). However, three small retrospective case series have suggested that a surprising proportion of dysplastic or cancerous lesions may be visible on standard white-light colonoscopy with a variety of appearances, including polypoid lesions or other masses, mucosal irregularities, ulcerations, and strictures.340-342 In a landmark study, Blackstone and associates identified 12 patients with DALMs during surveillance colonoscopies, of whom seven (58%) subsequently were found to have colon cancer.343 Subsequent studies have confirmed this high incidence of developing invasive carcinoma,344-346 including the finding of carcinoma on immediate colectomy in up to 50% of patients with DALMs.347 A systematic review of 10 surveillance studies of 1225 patients reported that 10 (42%) of 24 patients with high-grade dysplasia who underwent immediate colectomy had synchronous cancer, and 15 (32%) of 47 patients who were found to have highgrade dysplasia after an initially normal colonoscopy subsequently were found to have cancer.347 Thus, the presence of a DALM or high-grade dysplasia appeared to be highly predictive of concurrent or subsequent colon cancer, and colectomy was initially recommended for these patients. DALMs, however, consist of a heterogeneous group of dysplastic lesions and can be categorized into adenoma-like DALMs and nonadenoma-like DALMs based on their endoscopic and histologic appearances. Several small series reported that adenoma-like DALMs carry a low risk for developing cancer and that they may be managed conservatively without immediate colectomy.348-350 A prospective follow-up study compared the outcomes of 24 patients in three groups: UC and adenoma-like DALMs (located within an area of colitis, but with no other areas of flat dysplasia or cancer); patients with UC and coincidental sporadic adenomas (located proximal to the extent of colitis); and patients without UC but with sporadic adenomas.351 All patients were managed conservatively with complete polypectomy followed by colonoscopic surveillance. During the follow-up period, the fraction of patients with adenoma-like DALMs developing further adenomas (58%) was similar to those in the other two groups of patients, and no cancer was found, suggesting that these polypoid, adenoma-like DALMs follow a relatively benign course comparable to that of sporadic adenomas unrelated to colitis. Another prospective study also supported the practice of polypectomy for dysplastic polypoid lesions without flat dysplasia adjacent to the polyps or throughout the colon in patients with UC.352 Thus, conservative management with complete polypectomy and endoscopic surveillance has become an increasingly accepted alternative to the traditional recommendation of colectomy in patients with UC and adenoma-like DALMs. The predictive value of low-grade dysplasia for the development of more-advanced lesions varies among studies. In the previously mentioned systematic review of 10 studies in the literature, the risk of synchronous cancer at immediate colectomy was 19% (3 of 16) in patients with low-grade dysplasia.347 In patients with newly diagnosed low-grade dysplasia after a previously normal surveillance colonoscopy, 16% later progressed to high-grade dysplasia, DALM, or cancer (8%), and 29% of patients with untreated low-

Chapter 112  Ulcerative Colitis grade dysplasia found at the initial surveillance colonoscopy progressed to high-grade dysplasia, DALM, or cancer (13%). To date, more than 20 studies evaluating the risk of colorectal cancer, high-grade dysplasia, or both in patients with low-grade dysplasia have been performed and have reached different conclusions. In a single center study of 46 UC patients with flat lowgrade dysplasia on surveillance colonoscopy, unexpected advanced neoplasia occurred in 4 (24%) of 17 patients who underwent immediate colectomy.353 Five additional cases of cancer at stage II or higher occurred despite surveillance examinations. These findings suggest that patients with low-grade dysplasia are at relatively high risk of developing more-advanced lesions or cancer. In two European studies with a total of 89 patients, however, only 3% to 10% of patients with low-grade dysplasia later developed high-grade dysplasia, DALM, or cancer after 10 years of follow-up.354,355 Thus, some experts have advocated a more conservative approach for low-grade dysplasia. A meta-analysis of 20 studies published through July 2005 included 508 patients with low-grade dysplasia out of more than 2677 patients with UC who were enrolled in a colorectal cancer surveillance program.356 Data from eight of the studies reported that of 98 patients with low-grade dysplasia who underwent colectomy, 36 had a concurrent advanced lesion (colorectal cancer in 25, high-grade dysplasia in 11). The authors reported incidence rates of 1.4% and 3% per person per year after diagnosis of low-grade dysplasia, which translated into a nine-fold higher risk of colorectal cancer and a 12-fold higher risk of high-grade dysplasia compared with patients without dysplasia. Therefore, a finding of low-grade dysplasia on colonoscopy should prompt serious discussion about colectomy given the potentially high risk of future advanced lesions. Most authorities recommend annual to biennial colonoscopy with biopsies in patients who have UC extending beyond the rectum and who have had disease for 8 to 10 years.357,358 Colonoscopic examination with biopsies should be performed during periods of inactive disease so as not to allow inflammation and reactive change to be potentially

Flat dysplasia

DALM

Adenoma-like DALM

confused with dysplasia. Four-quadrant biopsies should be obtained every 10 cm throughout the colon, and targeted biopsies should be obtained from any raised or potentially dysplastic lesions. Thus, performing proper surveillance requires extensive biopsies. In fact, it has been estimated that 33 biopsies are required to achieve a 90% probability of identifying dysplasia or cancer if it is present, and nearly twice as many biopsies are necessary to exclude dysplasia or cancer with 95% certainty.359 Because the goal of surveillance colonoscopy is to detect dysplasia before cancer develops, a minimum of 64 biopsies are required to detect the highest rate of dysplasia anywhere in the colon with 95% confidence, but unfortunately, data from questionnaire surveys show that many fewer biopsies are obtained in routine clinical practice.360,361 A successful surveillance program also depends on appropriate management based on colonoscopic and biopsy findings. Colectomy generally is recommended for flat dysplasia, either low-grade or high-grade, or invasive carcinoma (Fig. 112-19).337,347,362 Unifocal low-grade dysplasia may be managed by continued surveillance colonoscopy, although all patients should be made aware of the potential high risk of developing high-grade dysplasia or colorectal cancer in the future and thus should be offered the option of colectomy; colectomy is more routinely recommended for patients with multifocal low-grade dysplasia. The management of polypoid lesions identified on surveillance colonoscopy is summarized in Figure 112-20. In patients of the appropriate age to develop sporadic adenomas, polypoid lesions identified outside of underlying colitis are considered coincidental sporadic adenoma and can be treated with complete polypectomy, if it is technically feasible. These patients should continue to receive surveillance colonoscopy annually or biennially. For pedunculated or sessile polypoid lesions within regions of colitis, endoscopic polypectomy should be performed. In addition to multiple random biopsies at regular intervals throughout the entire colon as mentioned earlier, biopsies should be obtained from the area around the polyp site. If the polyp­ ectomy is not complete, surgical resection is necessary. If

Non–adenoma-like DALM

Low-grade multifocal Colectomy dysplasia

High-grade dysplasia

unifocal Follow algorithm for the management of polypoid lesions (see Figure 112–21)

Colectomy

Indefinite dysplasia

No dysplasia

Repeat colonoscopy within 6 months

Continued surveillance

Colectomy Repeat colonoscopy within 6 months

Colectomy

Dysplasia? Yes Colectomy

No Repeat colonoscopy at frequent intervals*

*Ideal frequency has not been determined.

Figure 112-19.  Algorithm for the management of dysplasia-associated lesion or mass (DALM), flat dysplasia, indefinite dysplasia, or no dysplasia found on surveillance colonoscopy in patients with ulcerative colitis.

2007

2008

Section X  Small and Large Intestine Polypoid lesion

Outside area of colitis

Within area of colitis

Polypectomy* and biopsies of surrounding mucosa (in addition to random biopsies, from throughout the colon)

Dysplasia in the surrounding mucosa or the rest of colon? Yes Colectomy

Polypectomy*

Repeat surveillance in 1–2 years

No Repeat surveillance at frequent intervals†

*If polypectomy is incomplete, patients should undergo surgery. †Ideal frequency has not been determined. Figure 112-20.  Recommended algorithm for management of polypoid lesions.

the polypectomy is complete and there is dysplasia in the surrounding mucosa or the rest of the colon, colectomy is advisable because of the high association with synchronous carcinoma and subsequent development of invasive cancer.343,347 Alternatively, if polypectomy is complete and there is no dysplasia in the surrounding mucosa or the rest of the colon, this polypoid dysplastic lesion can be considered an adenoma-like DALM. In this scenario, patients should be advised of the uncertainty regarding optimal management of adenoma-like DALMs.363 Studies have suggested that conservative management with complete endoscopic polypectomy followed by careful colonoscopic surveillance may be an acceptable alternative to the traditional recommendation of colectomy.351,352 If the polyp is pedunculated with a well-defined stalk, endoscopic resection with continued surveillance at an increasing interval (e.g., six months initially) generally is considered adequate. If the polyp is sessile, the same practice may be used provided that the patient understands it is controversial.352 Decisions should be individualized. Other factors that should be taken into consideration include age (lesions in younger patients are more likely to be disease-related rather than sporadic adenomas), duration of disease, severity and course of disease, family history of colorectal cancer, and personal history of PSC.351,363 Colonoscopic surveillance is fraught with many potential pitfalls, including sampling error, interobserver variability for determining dysplasia, and difficulty detecting and differentiating dysplastic lesions from other lesions. Thus, techniques have been developed to enhance the diagnostic accuracy of surveillance colonoscopy in patients with UC. These techniques include the use of magnifying colonoscopy combined with either chromoendoscopy, in which tissue stains are applied to the gastrointestinal mucosal surface at endoscopy to better enhance or characterize specific findings,364-366 or narrow band imaging (NBI), which uses short wavelengths of the hemoglobin absorption bands to enhance mucosal vasculature and surface contrast.367,368

These techniques do not currently represent the standard of practice. The use of magnification and chromoendoscopy as an adjunct to conventional colonoscopy has been shown, in comparison with conventional colonoscopy, to increase the rate of detecting dysplasia with a sensitivity and specificity of 93% for differentiating neoplastic and non-neoplastic lesions.364 The use of NBI for detecting dysplasia in UC is still in its preliminary stages, because it is unclear exactly which vascular patterns are associated with dysplasia. In addition, research has investigated the use of molecular, genetic, and immunohistochemical markers to enhance diagnostic accuracy of colitis-associated dysplasia. Although sporadic colon cancer and IBD-associated malignancy follow a similar pathway from dysplasia to carcinoma,369 studies have shown differences in the prevalence and timing of certain molecular events between these two neoplastic groups. Some of these differences in colitis-associated dysplasia include infrequent and late mutations in the APC and β-catenin genes, more frequent and early abnormalities in the 3p (von Hippel-Lindau) gene locus, p53, and p16 loci, and higher prevalence of Sialyl-Tn antigen expression.370-377 Another potential strategy to reduce the risk of colorectal cancer in UC is medical chemoprevention. Three different medications have shown promise in this regard and include folic acid, 5-aminosalicylates, and ursodeoxycholic acid. The results of two observational studies examining the effects of folate have been published, both of which showed that folate supplementation was associated with a non­ significant reduction in risk of dysplasia or colorectal cancer378,379; 5-aminosalicylates have been studied more extensively in this regard. A meta-analysis of nine studies (three cohort and six case-control studies) totaling 1932 patients, and including 334 cases of colorectal cancer and 140 cases of dysplasia, showed a significant protective effect of 5-aminosalicylates for the prevention of colorectal cancer (pooled OR 0.51; 95% CI: 0.37-0.69) and colorectal cancer or dysplasia (pooled OR 0.51; 95% CI: 0.38-0.69).380 Two studies have shown that ursodeoxycholic acid might have significant benefit for preventing neoplasia in patients with UC and PSC.381,382 Even though proctocolectomy with IPAA is the most efficient way to prevent dysplasia and colorectal cancer in UC, there is still a potential risk of dysplasia or cancer to develop in the rectal cuff-anal transition zone and even in the ileal pouch. Since 1990, there have been at least 14 reported cases of cancer arising in the anorectal mucosa and nine cases of cancer in the pouch.383,384 A recent systematic review of 23 uncontrolled observational studies (most of which were retrospective case series), including a total of 2040 patients in whom no cancers were found, reported a pooled prevalence rate of 1.13% for dysplasia in the anorectal mucosa or pouch, with roughly equal prevalence in the anorectal mucosa and the pouch.383 The authors noticed that dysplasia and cancer identified preoperatively or during surgery were significant predictors of the development of dysplasia, whereas pouch­ itis and length of follow-up were not. Currently, no guidelines exist for endoscopic surveillance after IPAA, but given the unexpectedly high prevalence of dysplasia in the residual anorectal mucosa and pouch, periodic endoscopy with biopsy should be considered.

POUCHITIS (see Chapter 113) Pouchitis is the most common long-term complication of colectomy with IPAA for UC (see Chapter 113) but, curiously, it is rarely seen in patients undergoing the same surgery for familial adenomatous polyposis. Pouchitis is said to occur when there is nonspecific inflammation of the

Chapter 112  Ulcerative Colitis ileal reservoir, resulting in variable clinical symptoms resembling those of UC. Pouchitis occurs in 7% to 51% of patients undergoing restorative proctocolectomy for UC.385,386 The incidence is highest during the first six months after loop ileostomy closure.

Pathophysiology

The pathophysiology of pouchitis is not well understood but is likely multifactorial. Several etiologic mechanisms have been postulated including fecal stasis with bacterial overgrowth, recurrent UC following colonic metaplasia of the ileal epithelial cells of the pouch, Crohn’s disease, mucosal ischemia, and viral infection; another postulated mechanism is an ongoing immune process that may be recurrent UC, misdiagnosed Crohn’s disease, or an overlap or other form of IBD.387,388

Signs and Symptoms

Characteristic symptoms of pouchitis include increased bowel frequency, rectal bleeding, abdominal cramping, rectal urgency, tenesmus, and fecal incontinence.389 Other associated symptoms are fever, malaise, arthralgias, and erythema nodosum. Endoscopic evaluation, which is essential to distinguish pouchitis from cuffitis (continued rectal cuff inflammation due to UC), Crohn’s disease in the small bowel, or irritable pouch syndrome (a condition of increased small bowel motility in the absence of mucosal abnormalities), may reveal mucosal erythema, edema, granularity, friability, petechiae, loss of vascular pattern, erosions, and superficial ulcerations. Deep and irregularly shaped ulcers similar to those seen in Crohn’s disease may be present. Histologically, pouchitis is marked by an acute inflammatory infiltrate with mucosal ulceration and crypt abscesses in addition to chronic inflammation, villus atrophy, and crypt hyperplasia. Similar to UC, the diagnosis of pouchitis is based on a constellation of clinical symptoms, endoscopic appearances, and histologic features.380,390 There appear to be two distinct clinical forms of pouch­ itis, acute pouchitis and chronic pouchitis. Acute pouchitis is defined by symptom duration of less than four weeks and prompt response to antibiotics (also classified as antibioticresponsive), and chronic pouchitis is defined by symptom duration more than four weeks and need for long-term antibiotics or other therapeutic agents (also classified as antibiotic-dependent or antibiotic-refractory).388,390 More than 20% of patients have chronic continuous symptoms but less than 10% have severe chronic pouchitis requiring long-term maintenance therapy.386

Risk Factors

A number of studies examining risk factors for pouchitis have been performed. Unfortunately, many of these studies were based in referral centers, were retrospective, and had relatively short follow-up periods; they also did not distinguish acute pouchitis and chronic pouchitis, which at present seem to be very different disease entities. As a result, it should not come as a surprise that these studies identified different risk factors for developing pouchitis, with only a few studies agreeing on the same risk factors. Among these studies, risk factors for pouchitis that have been variably identified include the presence of extraintestinal manifestations and especially PSC, smoking, use of NSAIDs, various serologic markers, and preoperative thrombocytosis.

Acute versus Chronic Pouchitis

There is growing evidence that acute pouchitis and chronic pouchitis are in fact distinct disease entities. For instance,

fecal stasis with bacterial overgrowth presumably occurs in the pouch after IPAA in all UC patients,391,392 yet only some develop acute pouchitis or chronic pouchitis. Also, chronic pouchitis typically appears sooner after stoma closure than does acute pouchitis.393,394 In addition, the serologic profiles associated with acute pouchitis and chronic pouchitis seem to be different.62,65,394,395 Finally, patients with acute pouch­ itis appear to react to endoluminal antigens only transiently, whereas those with chronic pouchitis have more persistent inflammation. Two prospective studies from Cedars-Sinai Medical Center have examined risk factors for acute pouchitis and chronic pouchitis separately in the same group of patients.65,394 The first study, which examined clinical risk factors for developing pouchitis, found that preoperative use of glucocorticoids and smoking were independently associated with the development of acute pouchitis, whereas extraintestinal manifestations, preoperative thrombocytosis, and increased length of follow-up were independently associated with the development of chronic pouchitis on multivariate analysis.394 Interestingly, smoking was found to be significantly protective against the development of chronic pouchitis on multivariate analysis, thus again supporting the concept that acute pouchitis and chronic pouchitis may be distinct disease entities. The second study, with longer follow-up, followed 238 patients for a median of 47 months (range, three to 142 months) and reported incidence rates of acute pouchitis and chronic pouchitis of 18% and 12%, respectively.65 Although the median time to the diagnosis of chronic pouchitis was shorter than that for acute pouchitis (six vs. nine months), these differences did not attain statistical significance. The second study also looked at serologic profiles, specifically the presence or absence of pANCA or anti-CBir1 antibodies and the risk of pouchitis, and found that both antibodies were seen significantly more often in patients who developed pouchitis.394 In addition, because the authors previously had shown that high pANCA levels (more than 100 U/mL) were associated with the development of chronic pouchitis but not acute pouchitis, they stratified the pANCA+ patients into high-level and low-level pANCA+ groups and found that acute pouchitis occurred with the same frequency in patients with high-level versus low-level pANCA expression, whereas chronic pouchitis occurred significantly more often in patients with high-level pANCA seroreactivity, thus confirming the results from their prior study. They also observed that anti-CBir1 expression influenced the onset of acute pouchitis only in patients with low-level pANCA expression, and the onset of chronic pouchitis only in patients with high-level pANCA expression.

Treatment

The mainstay of therapy for pouchitis is antibiotics. In one of the few placebo-controlled trials of antibiotics for the treatment of chronic active pouchitis, metronidazole at 1200 mg/day for one week resulted in an overall response rate of 73% compared with 9% for placebo.396 Ciprofloxacin at 1 g/day for two weeks also is effective for the treatment of acute pouchitis and has been shown to be superior to metronidazole in efficacy and tolerability in an RCT.397 Alternative regimens include topical metronidazole, amoxicillin/ clavulanic acid, and combinations of ciprofloxacin with metronidazole, rifaximin, and tinidazole.398-402 For patients on chronic therapy, cycling of multiple antibiotics at weekly intervals might help overcome bacterial resistance. The second-line options for treating pouchitis include topical and oral mesalamines,390,403 although the use of

2009

2010

Section X  Small and Large Intestine these agents is based on uncontrolled studies and anecdotal experiences. Topical and systemic glucocorticoids, including topical and oral budesonide, can be used for patients with pouchitis who do not respond to antibiotics and me­ salamines.404,405 Immunosuppressive and biological therapies, including cyclosporine enemas, azathioprine, and infliximab, also have been reported to be beneficial and may be considered in patients whose pouchitis is refractory to conventional therapies.406,407 Probiotic therapy is another option. The majority of data regarding the use of probiotics for pouchitis involves VSL#3. Three randomized, placebo-controlled trials of VSL#3 for either primary (one study) or secondary (two studies) prophylaxis of pouchitis showed significant benefit with the probiotic.408-410 An open-label study from the Cleveland Clinic, however, showed that VSL#3 was ineffective for secondary prophylaxis of pouchitis.411 Thus, it remains unclear whether probiotics are beneficial in patients with pouchitis, and larger randomized studies are needed to help resolve this issue. Two small studies have suggested a possible benefit of prebiotics (inulin) or synbiotics (probiotic L. rhamnosus GG combined with prebiotic fructo-oligosaccharide) for managing pouchitis.412,413 Larger randomized studies are needed before any recommendations can be made regarding the use of these agents in pouchitis. Other nonconventional therapies include nutritional replacement using known energy substrates for the bowel and bismuth carbomer enema. Bismuth has been used to treat pouchitis with varying results. Although an open-label study reported benefit of bismuth carbomer enemas in patients with chronic pouchitis,414 a placebo-controlled study failed to show benefit.415 Another uncontrolled study reported improvement in patients with chronic pouchitis treated with chewable bismuth subsalicylate tablets,416 but this finding has not been confirmed in controlled trials. In the only randomized comparison of glutamine and butyrate suppositories in patients with chronic pouchitis after withdrawal of all conventional therapies, 60% of patients treated with glutamine 1 g twice daily for three weeks entered remission compared with 33% of patients receiving butyrate 40 mmol twice daily.417 Surgical options such as ileal pouch excision or reconstruction should be considered in cases refractory to medical therapies or cases in which the frequency or chronicity of pouchitis compromises the patient’s quality of life.418

EXTRAINTESTINAL MANIFESTATIONS Patients with UC commonly present with a wide range of systemic and local problems that can add to the complexity of treatment. These extraintestinal manifestations can affect any organ system, but they most commonly involve the skin, eyes, mouth, joints, and liver (Table 112-18). These complications often are classified by their relations to the activity of the colitis, but they can occur before, during, or following exacerbations of bowel disease. Manifestations that parallel disease activity usually improve upon successful treatment of the colitis.

CUTANEOUS/ORAL

The most common dermatologic manifestations of UC are complications of drug treatment. These include hypersensitivity, photosensitivity, and urticarial rashes related to sulfasalazine and less commonly to mesalamine. Patients receiving glucocorticoids often develop acne, which can be

Table 112-18  Common Extraintestinal Manifestations of Ulcerative Colitis Cutaneous/Oral Angular stomatitis Aphthous stomatitis Erythema nodosum Oral ulcerations Psoriasis Pyoderma gangrenosum Pyostomatitis vegetans Sweet’s syndrome (acute febrile neutrophilic dermatosis) Ophthalmologic Conjunctivitis Episcleritis Retinal vascular disease Scleritis Uveitis, iritis Musculoskeletal Ankylosing spondylitis Osteomalacia Osteonecrosis Osteopenia Osteoporosis Peripheral arthropathy Sacroiliitis Hepatobiliary Autoimmune hepatitis Cholangiocarcinoma Pericholangitis Primary sclerosing cholangitis Hepatic steatosis Hematologic Anemia of chronic disease Autoimmune hemolytic anemia Hypercoagulable state Iron deficiency anemia Leukocytosis or thrombocytosis Leukopenia or thrombocytopenia

distressing cosmetically. Other common dermatologic manifestations associated with UC are erythema nodosum and pyoderma gangrenosum. Erythema nodosum occurs in 2% to 4% of patients with UC. Its activity typically parallels the activity of the underlying bowel disease. Erythema nodosum also can occur as a drug reaction to the sulfapyridine component of sulfasalazine. It classically manifests as single or multiple tender, raised, erythematous nodules on the extensor surfaces of the lower extremities. If possible, the diagnosis should be made clinically without biopsy, because biopsy is associated with increased tendency to form scars. Erythema nodosum usually responds to treatment of the UC. Severe or refractory cases can require systemic glucocorticoids or immunosuppressive therapy. Pyoderma gangrenosum is less common than erythema nodosum and occurs in 1% to 2% of patients. It usually is related to the activity of colitis but can manifest or persist despite inactive bowel disease. Lesions may be single or multiple and usually occur on the trunk or extremities but can develop on the face, breast, or sites of trauma, including stoma and intravenous sites.419 The classic lesion begins as erythematous pustules or nodules that break down, ulcerate, and coalesce into a larger, tender, burrowing ulcer with irregular, violaceous edges.420 Although the appearance can be dramatic, the ulcers are sterile. Histopathologically,

Chapter 112  Ulcerative Colitis Table 112-19  Types of Peripheral Arthropathy Associated with Ulcerative Colitis FEATURE Characteristics Frequency Duration of attacks Association with bowel disease activity Joints Affected Number Type Prevalence

TYPE 1 (PAUCIARTICULAR)

TYPE 2 (POLYARTICULAR)

35% <10 wk (median, 5 wk) Parallel

24% Months to years (median, 3 yr) Independent

<5 Mainly large joints Knee > ankle > wrist > elbow > MCP > hip > shoulder

≥5 Mainly small joints MCP > knee > PIP > wrist > ankle > elbow > shoulder

MCP, metacarpophalangeal joint; PIP, proximal interphalangeal joint; UC, ulcerative colitis. Adapted from Su C, Judge TA, Lichtenstein GR. Extraintestinal manifestations of inflammatory bowel disease. Gastroenterol Clin North Am 2002; 31:307.

pyoderma has the features of a sterile abscess with a marked neutrophilic infiltration. Pyoderma gangrenosum can resolve with treatment of the underlying colitis. Most cases usually respond to intralesional glucocorticoid injections or topical therapy with cromolyn sodium, me­ salamine, glucocorticoids, or tacrolimus.420-422 More-severe cases can require systemic glucocorticoids, immuno­ suppressants (e.g., cyclosporine, azathioprine, methotrexate, and tacrolimus), dapsone, or anti-TNF therapy (e.g., infliximab).423,424 Less-common skin manifestations associated with UC include Sweet’s syndrome (acute febrile neutrophilic dermatosis), and pyodermite végétante of Hallopeau. The latter has a presentation similar to that of pyoderma gangrenosum but also involves the mouth. At least 10% of patients with UC develop oral aphthous ulcers. These lesions usually occur with flares of colitis and resolve on control of the bowel disease. Angular stomatitis and a sore tongue may be seen in patients with deficiencies of iron, B vitamins, or other micronutrients (see Chapters 4 and 100). A rare oral lesion that may be seen in patients with UC is pyostomatitis (pyoderma) vegetans, which appears as a pustular eruption of the oral mucosa resulting in a cobblestone appearance.425

OPHTHALMOLOGIC

The two most common ocular manifestations associated with UC are episcleritis and uveitis, occurring in 5% to 8% of patients. Episcleritis is characterized by painless hyperemia of the sclera and conjunctiva without loss of vision. It typically parallels the activity of bowel disease and usually responds to anti-inflammatory therapy. In contrast, uveitis presents as an acute or subacute painful eye with visual blurring often accompanied by photophobia and headache. Temporal correlation of uveitis with the activity of the colitis is less predictable than with episcleritis. Patients with uveitis should receive prompt ophthalmologic consultation and treatment with local glucocorticoid ocular drops to prevent progression to blindness. The occurrence of uveitis increases with the dose and duration of glucocorticoid use. Glucocorticoid therapy also can lead to posterior subcapsular cataracts. Thus, patients receiving glucocorticoid therapy should be advised to undergo annual ophthalmologic examination.

MUSCULOSKELETAL

Musculoskeletal abnormalities associated with UC can be grouped broadly into rheumatologic disorders and metabolic bone diseases.

Peripheral Arthropathy

Peripheral arthropathy occurs in 5% to 20% of patients with UC. The risk of arthropathy increases with the extent of colonic disease. Peripheral arthropathy can be classified into two distinct types (Table 112-19).426 Type 1 is asymmetrical and pauciarticular, affecting fewer than five joints and typically involving the large joints (knees, elbows, ankles). It usually manifests with acute, self-limited episodes that parallel the underlying bowel disease activity. Type 2 arthropathy is symmetrical and polyarticular, affecting five or more joints and typically involving the small joints. This type manifests with persistent symptoms independent of the colitis activity. Both forms are nondeforming and seronegative. The involved joints are swollen, erythematous, and hot. Peripheral arthropathy usually responds to treatment of colitis. Rest, physical therapy, intra-articular glucocorticoid injection, and therapeutic arthrocentesis also can help control symptoms.

Axial Arthropathy

Axial arthropathy occurs less often than does peripheral arthropathy in patients with UC and includes sacroiliitis and spondylitis. Isolated sacroiliitis occurs in 10% to 15% of patients, but the incidence may be higher based on MR imaging. It usually does not parallel the activity of the bowel disease. The typical symptom is low back pain, but some patients are asymptomatic. Most patients with sacroiliitis are HLA-B27 negative and do not progress to ankylosing spondylitis. Ankylosing spondylitis occurs in 1% to 2% of patients with UC, and most of these patients are HLA-B27 positive. Symptoms of ankylosing spondylitis can appear long before or after the onset of the intestinal symptoms and are independent of the activity of colitis. Patients often experience onset of severe back pain at a young age, usually associated with morning stiffness and exacerbated by periods of rest. The course of ankylosing spondylitis is progressive, resulting in permanent skeletal damage. Radiologic films in early stages may be normal or show only minimal sclerosis. Advanced stages are characterized by a bamboo spine, with squaring of vertebral bodies, bony proliferation, and ankylosis. Treatment of axial arthropathy is similar to that for peripheral arthropathy, except that control of the underlying colitis does not alter the progressive nature of ankylosing spondylitis. Clinical success with anti-TNF therapy has been reported, with significant improvements in pain scales, range of motion, physical function, and overall quality of life.427

2011

2012

Section X  Small and Large Intestine Low Bone Mineral Density

Patients with UC can develop low bone mineral density owing to several factors, including glucocorticoid therapy, low physical activity, and possibly inflammatory cytokines. In general, patients with UC who have not been treated with glucocorticoids are not considered to be at increased risk for osteopenia or osteoporosis. Bone densitometry (dual-energy x-ray absorptiometry [DEXA] scan) at one- to two-year intervals is recommended for patients who have received glucocorticoids at high doses or for a long duration. Osteonecrosis, also known as avascular or aseptic necrosis of bone or osteochondritis dissecans, is a less-common but serious complication in patients with UC. Patients typically present with joint swelling and pain exacerbated by motion. Most cases involve the hips and are bilateral, but the knees and shoulders may be affected. The most important risk factor for osteonecrosis is glucocorticoid therapy, with one series reporting a 4% incidence of osteonecrosis within six months of glucocorticoid therapy in patients with IBD428; concurrent use of total parenteral nutrition is another risk factor. Early diagnosis by MR imaging or bone scan is essential for proper treatment, including medical management, cortical decompression, and arthroplasty. Management of decreased bone density includes calcium (typically calcium citrate at 500 mg three times daily) and vitamin D supplementation (at 1000 IU daily) initially. It also is useful to check a 25-OH-D (vitamin D2) level initially to determine if a higher dose of vitamin D is needed. The addition of bisphosphonates should be considered for worsening osteopenia or osteoporosis.

HEPATOBILIARY

A wide range of hepatobiliary complications is associated with UC. Mild elevations in serum aminotransferase and alkaline phosphatase levels are common in severe attacks of UC. In most cases, serum enzyme levels return to normal once remission is achieved. These abnormalities are thought to be related to a combination of factors, including malnutrition, sepsis, and fatty liver. An excess of fat in the hepatocytes is found in 60% of patients who undergo urgent colectomy for severe colitis. The most important hepatobiliary complication associated with UC is PSC, which occurs in approximately 3% of patients (see Chapter 68). PSC is a chronic inflammatory disease of the biliary tree resulting in fibrosis and, eventually, cirrhosis and hepatic failure. Intrahepatic or extrahepatic ducts (or both) may be involved. PSC is characterized radiologically by beading, irregularity, and stricturing of the bile ducts. The diagnosis of PSC is made on either endoscopic retrograde cholangiography or MR cholangiography. A liver biopsy might support the diagnosis but rarely is diagnostic. The classic histopathology of PSC on liver biopsy is the onion skin pattern, with concentric fibrosis around the small bile ducts and, ultimately, obliteration of the ducts. The histologic appearance can be variable, however, ranging from chronic inflammatory infiltration in the portal tracts to cirrhosis. PSC should be excluded in patients with UC who have persistently abnormal liver biochemical tests or evidence of chronic liver disease. PSC is independent of the underlying colitis, and it usually follows a progressive course after many years of stable disease. Unfortunately, no treatment has been shown definitively to be effective. Ursodeoxycholic acid can slow disease progression. Patients with a dominant extrahepatic biliary stricture can benefit from endoscopic dilatation or stent placement. Patients who develop end-stage liver disease require liver transplantation. Additionally, patients with

PSC are at a significantly increased risk for cholangiocarcinoma and colon carcinoma.

HEMATOLOGIC Hypercoagulability

The occurrence of hypercoagulability is a well-recognized complication of UC. The incidence of thromboembolic events in patients with UC varies widely in different studies, but most commonly they manifest as deep venous thrombosis or pulmonary embolism. Renal artery thrombosis, cerebrovascular accidents, coronary artery thrombosis, and venous thrombosis of mesenteric, portal, and hepatic vessels all have been reported.429-431 The cause of the thromboembolism is multifactorial. Physiologically, microvascular activation of coagulation is present in the inflammatory states of colitis. A variety of coagulation and platelet abnormalities are present in patients with UC, particularly those with severe disease, and include thrombocytosis; increased levels of fibrinogen, coagulation factors V and VIII, and plasminogen activator inhibitor; and decreased levels of antithrombin III, proteins C and S, factor V Leiden, and tissue plasminogen activator. Although there is no increased incidence of specific coagulation abnormalities in patients with UC, there may be an increased incidence of factor V Leiden mutation in patients with thromboembolic complications associated with UC.432-434 Patients with these complications should be treated with anticoagulants, just as in other patient populations. Although there may be concerns of an increased risk of gastrointestinal bleeding with anticoagulation, it generally is safe and rarely is complicated by colonic bleeding.

Anemia

Another common hematologic complication is anemia. The anemia in patients with UC may be a result of acute or chronic gastrointestinal blood loss, chronic disease, folate deficiency from sulfasalazine therapy, or autoimmune hemolysis. Autoimmune hemolytic anemia, usually Coombs positive, may be related to sepsis or glucose-6-phosphate dehydrogenase deficiency in patients taking sulfasalazine.

OTHERS

Secondary systemic amyloidosis is a rare but serious complication associated with UC.435 Amyloidosis in these patients usually affects the kidney and manifests with proteinuria followed by the nephrotic syndrome and subsequent renal insufficiency. Diagnosis is made with a fat pad aspiration or alternatively, biopsies from the liver, rectum, or kidney. Pericarditis, pleuropericarditis, and constrictive pericarditis have been reported in patients with UC436-439; this complication also may be related to mesalamine therapy.440,441 The pathogenesis of these complications is unknown, and their true association with UC is uncertain. Patients with UC can also develop abnormalities in pulmonary function, including an increase in functional reserve capacity and a decrease in diffusion capacity.442 Other pulmonary diseases that have been described in patients with UC include bronchiectasis, bronchiolitis, fibrosing alveolitis, pulmonary fibrosis, and pulmonary vasculitis.443-445

KEY REFERENCES

Cho JH. The genetics and immunopathogenesis of inflammatory bowel disease. Nature Rev Immunol 2008; 8:458. (Ref 29.) Duerr RH, Taylor KD, Brant SR, et al. A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science 2006; 314:1461. (Ref 33.)

Chapter 112  Ulcerative Colitis Fleshner P, Ippoliti A, Dubinsky M, et al. A prospective multivariate analysis of clinical factors associated with pouchitis after ileal pouch-anal anastomosis. Clin Gastroenterol Hepatol 2007; 5:952. (Ref 394.) Frank DN, St. Amand AL, Feldman RA, et al. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proc Natl Acad Sci U S A 2007; 104:13780. (Ref 43.) Hahnloser D, Pemberton JH, Wolff BG, et al. Results at up to 20 years after ileal pouch-anal anastomosis for chronic ulcerative colitis. Br J Surg 2007; 94:333. (Ref 315.) Kandiel A, Fraser AG, Korelitz BI, et al. Increased risk of lymphoma among inflammatory bowel disease patients treated with azathioprine and 6-mercaptopurine. Gut 2005; 54:1121. (Ref 221.) Lichtiger S, Present DH, Kornbluth A, et al. Cyclosporine in severe ulcerative colitis refractory to steroid therapy. N Engl J Med 1994; 330:1841. (Ref 223.) Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease: Incidence, prevalence, and environmental influences. Gastroenterology 2004; 126:1504. (Ref 5.) Moskovitz DN, Van Assche G, Maenhout B, et al. Incidence of colectomy during long-term follow-up after cyclosporine-induced remission

of severe ulcerative colitis. Clin Gastroenterol Hepatol 2006; 4:760. (Ref 227.) Rubin PH, Friedman S, Harpaz N, et al. Colonoscopic polypectomy in chronic colitis: Conservative management after endoscopic resection of dysplastic polyps. Gastroenterology 1999; 117:1295. (Ref 352.) Rutgeerts P, Sandborn WJ, Feagan BG, et al. Infliximab for induction and maintenance therapy for ulcerative colitis. N Engl Med J 2005; 353:2462. (Ref 290.) Sartor RB. Microbial influences in inflammatory bowel diseases. Gastroenterology 2008; 134:577. (Ref 44.) Shih DQ, Targan SR. Immunopathogenesis of inflammatory bowel disease. World J Gastroenterol 2008; 14:390. (Ref 67.) Sutherland L, MacDonald JK. Oral 5-aminosalicylic acid for induction of remission in ulcerative colitis. Cochrane Database Syst Rev 2006; (2):CD000543. (Ref 140.) Sutherland L, Macdonald JK. Oral 5-aminosalicylic acid for maintenance of remission in ulcerative colitis. Cochrane Database Syst Rev 2006; (2):CD000544. (Ref 151.) Full references for this chapter can be found on www.expertconsult.com.

2013

CHAPTE R

113 Ileostomy, Colostomy, and Pouches Robert R. Cima and John H. Pemberton

CHAPTER OUTLINE Pathophysiologic Consequences of Proctocolectomy  2016 Fecal Output after Proctocolectomy  2016 Functional Sequelae  2016 Clinical Consequences of Proctocolectomy  2016 The Conventional Brooke Ileostomy  2017 Continent Ileostomy (Kock Pouch)  2017 Ileal Pouch-Anal Anastomosis  2018 Long-Term Results  2018 Controversies  2022 Additional Issues  2022 Abdominal Colectomy and Ileorectal Anastomosis  2023 Patient Selection  2024 Complications  2024

Proctocolectomy and permanent ileostomy return most patients with chronic ulcerative colitis (UC) to excellent health and remove premalignant mucosa in patients with UC or familial adenomatous polyposis (FAP). Many of the inconveniences and dangers formerly associated with an ileal stoma have been eliminated by improved surgical techniques, a wider range of better stomal appliances, and more effective education of patients.1 Between 1930 and 1950, the metabolic consequences of ileostomy became apparent, as did the frequent mechanical complications caused by ileostomy dysfunction. Better understanding of fluid, electrolyte, and blood replacement lessened the former problem, and newer techniques of ileostomy construction mitigated the second.2,3 Before these advances, ileostomies were made by withdrawal of the intestine through the abdominal wall, the serosal surface of ileum then being sutured to the skin. Ileostomy dysfunction resulted from the serositis following exposure of the serosal surface to the stomal effluent. The mucosa of the ileum, however, is not susceptible to inflammation after a similar exposure, and a solution, therefore, became conceptually simple: evert the mucosal surface of the bud and suture the mucosa to the skin. This modification was described simultaneously early in the 1950s in the United Kingdom and United States and is commonly referred to as a Brooke ileostomy (Fig. 113-1).1 Development of new ileostomy appliances quickly led to better acceptance by patients and, ultimately, to excellent longterm results.4 Enterostomal therapy was introduced in the 1960s as an additional allied health support, and ileostomy societies have blossomed in most countries, providing a lay component of support to patients with stomas.

Physiology  2024 Colostomy in the Management of Ulcerative Colitis  2024 Laparoscopic Approach  2024 Risk-Benefit Analysis  2025 Conventional Ileostomy  2025 Ileorectal Anastomosis  2025 Continent Ileostomy  2025 Ileal Pouch-Anal Anastomosis  2025

Brooke ileostomies are incontinent by definition, and during the 1960s, Nils Kock, a Swedish surgeon, developed the first effective alternative to this incontinent ileostomy.5 The Kock pouch procedure featured an ileal pouch, a nipple valve, and an ileal conduit, which led to a cutaneous stoma that, because this ileostomy was continent and therefore an appliance was not needed, could be made flush with the skin. The Kock pouch was used in selected patients with chronic UC and FAP.6 Stimulated by patients’ poor acceptance of the ileostomy, surgeons explored other alternatives to the incontinent ileal stoma with its ever-present external appliance. The ileoanal pull-through operation was resurrected, with an important technical modification: the addition of an ileal reservoir.7,8 This procedure offered the advantages of a normal exit for stool and preservation of the anal sphincters. Indeed, the use of this procedure in thousands of patients has revealed ileal pouch-anal anastomosis (IPAA) to be the procedure of choice in most patients requiring proctocolectomy for chronic UC or FAP. Although the Brooke ileostomy had become the usual operation after a colectomy in the United Kingdom and the United States, ileorectostomy was standard in continental Europe and South America. Indeed, these different attitudes continue to influence approaches to the newer operations. This chapter details the pathophysiologic and clinical implications of colectomy per se and describes the options for control of enteric content. There are three surgical options after total colectomy in patients with chronic UC and FAP: proctocolectomy with the terminal Brooke ileostomy, IPAA, and ileorectostomy. The Kock pouch and IPAA are contraindicated in patients with Crohn’s disease, but the option of segmental colectomy is available.

2015

2016

Section X  Small and Large Intestine

Ileal flow Figure 113-1.  Anatomy of the Brooke ileostomy. The mucosa is everted and sewn to the skin. Therefore no serosal surface is exposed to intestinal content.

PATHOPHYSIOLOGIC CONSEQUENCES OF PROCTOCOLECTOMY FECAL OUTPUT AFTER PROCTOCOLECTOMY

After a colectomy with any type of ileostomy, the absence of the colon obviously prevents its reabsorption of electrolytes and water. Usually, this creates no major pathophy­ siologic disturbance, but some important principles should be remembered. A normal colon absorbs at least 1000 mL of water and 100 mEq of sodium chloride each day, and the healthy colon can increase absorption more than 5 L/day when presented with increased amounts of fluid.9,10 Also, the colon has a greater capacity than the small intestine to conserve sodium chloride when a person is salt depleted. For example, under conditions of extremely low salt intake, sodium losses in normal stool can be reduced to 1 or 2 mEq/day, whereas patients with ileostomies have obligatory sodium losses of 30 to 40 mEq/day.11-13 The majority of patients adapt to these daily losses through minor changes in salt and water intake and physiologic compensation.14 Well-functioning conventional (Brooke) ileostomies discharge 300 to 800 g of material daily, 90% of which is water.12,13 Continent ileostomies and IPAAs have similar volumes of effluent.15 Foods containing substantial unabsorbable residue increase the total ileostomy output by increasing the amount of solids discharged. Although many anecdotes are reported on the effect of foods on the volume and consistency of stomal effluents, the response to specific foods varies from one patient to another, and changes usually are minimal.16

FUNCTIONAL SEQUELAE

When oral intakes of sodium, chloride, and fluid are adequate, patients with ileostomies do not become depleted in volume or electrolytes; negative sodium balance, however, can follow periods of diminished oral intake, vomiting, or excess perspiration.17 In addition, chronic oliguria is to be anticipated, even with established ileostomies, because

normal stools contain approximately 100 mL of water, whereas ileostomies lose 500 to 600 mL/day.14 Patients with ileostomies also have lower urinary Na+/K+ ratios because of compensatory renal conservation of sodium and water. These changes in the composition of urine presumably contribute to the increased frequency of urolithiasis (about 5%) in patients with ileostomies, whose stones are predominantly composed of urate or calcium salts18; these patients have relatively narrow tolerances for change in their volume and electrolyte status, and even minor changes potentially result in life-threatening electrolyte disturbances.19 When the terminal ileum is resected and a proximal ileostomy is constructed, there can be abnormalities of bile acid reabsorption, malabsorption of vitamin B12 (see Chapters 64, 100, and 101), steatorrhea, and more than expected losses of fluid (1 L/day). These abnormalities usually do not follow a colectomy that is performed for chronic UC or FAP because the ileum, being free of disease, is preserved. Resection for Crohn’s colitis can require removal of additional diseased ileum with the possible consequences of malabsorption and even short bowl syndrome, depending on the length of small bowel removed (see Chapters 101, 103, and 111). Colectomy also reduces the exposure of bile acids to the metabolic effects of the fecal flora, and after ileostomy, secondary bile acids largely disappear from bile; no metabolic consequences of significance have been recognized in this situation.20,21 The flora of ileostomy effluents have quantitative (104 to 107 colony-forming units [CFUs] per 100 milli­ liters) and qualitative characteristics that are intermediate between those of feces and those of normal ileal contents, whereas the flora in an IPAA or Kock pouch are more similar to feces.22-24 The principal pathophysiologic sequelae of colectomy with ileostomy are mainly the potential consequences of a salt-losing state; patients should be advised to use salt liberally and to increase their fluid intake, especially at times of stress, in hot weather, and after vigorous exercise. A balanced salt solution (Gatorade or Powerade) is a good source of balanced electrolytes. The limited ability of the small intestine to absorb sodium and water, however, means that stomal volumes also increase when the oral intake is increased.13

CLINICAL CONSEQUENCES OF PROCTOCOLECTOMY After successful proctocolectomy, life expectancy is slightly below normal for the first few years owing to complications of the stoma and to intestinal obstruction; after ileorectostomy for FAP or chronic UC, particularly the former, cancer can develop in the retained rectum. In general, however, the long-term mortality rate in patients after proctocolectomy and conventional ileostomy is the same as for a matched normal population.25 Ninety percent of patients with conventional ileostomies who responded to a survey rated the results of their operation excellent and claimed little inconvenience.4 Overall, there is no real difference in the reported quality of life of patients with conventional ileostomies, continent ileostomies, or an ileal pouch.26 Almost all were able to lead normal lives and enjoy normal sexual relationships; a few patients avoided certain strenuous physical activities. The metabolic consequences of a proctocolectomy per se should be the same regardless of whether a conventional ileostomy or an alternative procedure is performed. Patients in whom an ileostomy alternative achieves an excellent

Chapter 113  Ileostomy, Colostomy, and Pouches result have a better quality of life than do patients with a stoma because the former do not need to wear an ileostomy appliance. Indeed, when the Brooke ileostomy and IPAA were compared, patients with IPAA experienced significant advantages in performing daily activities and appeared to enjoy a better quality of life.27 There are certain unique complications of the newer operations, however, including incontinence (Kock pouch), pelvic infections and sepsis, and pouchitis (IPAA), which are discussed later.

THE CONVENTIONAL BROOKE ILEOSTOMY Major long-term complications relate to malfunctioning ileostomies, prestomal ileitis, and irritation of the peristomal skin. If the ileostomy was improperly constructed (a less-common problem with newer techniques), the stoma can become obstructed. Obstruction leads to cramping abdominal pain, increased ileal discharge (up to 4 L/day), and fluid and electrolyte depletion. Excessive ileal output arises, at least in part from increased intestinal secretion as the result of dilatation of the intestine proximal to the obstructed stoma. Stomal obstruction usually can be demonstrated by examining the stoma with the little finger or by endoscopy with a narrow endoscope. Radiologic studies reveal a dilated ileum proximal to the point of obstruction. Many obstructed ileostomies require reconstruction, and at operation, ulcerations often are found in the resected terminal ileum; the pathogenesis is unclear but probably relates in some way to the mechanical consequences of obstruction. Prestomal ileitis is a much less common problem than is stomal obstruction. Patients with this complication exhibit the features of mechanical obstruction, and, in addition, they have signs of systemic toxicity (e.g., fever, tachycardia, anemia).28,29 In prestomal ileitis, the ileum has numerous punched-out ulcers, sometimes extending to the serosa. It is not clear whether prestomal ileitis has a different pathogenesis from the changes that follow simple mechanical obstruction of the stoma; both complications involve ileum that was normal histologically at the time of colectomy. Backwash ileitis, seen typically in chronic UC, does not predispose to either prestomal ileitis or obstruction. Conversely, patients who have had colectomy and ileostomy for Crohn’s disease experience problems with the ileal stoma more often, perhaps because transmural inflammation involves the new terminal ileum. In some instances, it is difficult to determine whether stomal dysfunction results from mechanical obstruction or recurrent Crohn’s disease. Most people with an ileostomy lead a normal life and eat a normal diet; poorly digestible foods (e.g., nuts, corn, some fruits, lightly cooked vegetables) can obstruct the stoma and should be eaten in moderation and with careful chewing.4 Some patients experience continuing difficulties managing their ileostomy. These problems vary in severity, some being minor inconveniences and others being significant drawbacks to the success of the operation. Mechanical difficulties because of a poorly fitting stomal appliance can cause excoriation of the skin around the ileostomy and can even erode the stoma to produce a fistula. Some patients complain of unpleasant odors arising from the ileostomy bag, especially after eating certain foods, such as onions and beans. Because most odor arises from bacterial action on the contents of the appliance, however, the problem may be offset by frequent emptying of the appliance or by adding sodium benzoate or chlorine tablets to the appliance. Oral bismuth subgallate also controls odor, but doubts exist as to

whether its long-term use is justified, because questions of neurotoxicity and encephalopathy have been raised.30,31 A review of the long-term outcomes associated with ostomies has demonstrated a high rate of complications. The most common problems related to ostomies are skin irritation and parastomal hernias, both of which contribute to difficulty with appliance pouching, a term used by entero­ stomal therapists and referring to the fitting of an ostomy device. Several risk factors are associated with development of parastomal hernias, including obesity, malnutrition, chronic respiratory disorders that are associated with increased intra-abdominal pressure, chronic use of glucocorticoids or other immunosuppressive agents, malignancy, advanced age, and wound infection.32-35 Several techniques have been described for parastomal hernia repair, including primary fascial repair, stoma relocation, and mesh repair; mesh repair achieves the best outcomes.32 The simplest approach, primary repair, is associated with very poor outcomes and up to 100% recurrence rates. Stoma relocation may be an effective approach when the initial stoma site is unsatisfactory, but parastomal hernias occur at the new site in up to 76% of patients.35 The use of prosthetic mesh for parastomal hernia repair has significantly improved outcomes, with recurrence rates reported as low as 10%.36,37 To address the concerns of possible mesh infection and erosion of mesh into the bowel, the use of biologic materials, such as human acellular dermal matrix, has been reported, with results that are comparable to mesh in small series.38 Trained stomal therapists and lay societies of ileostomy patients can help with numerous aspects of postoperative care. Education of the patient is best started before surgery; meetings with others who have undergone ileostomy and referral to specialized publications can allay many fears and uncertainties. The United Ostomy Associations of America (UOAA, P.O. Box 66, Fairview, TN 37062-0066; www.uoaa .org) publishes an excellent series of booklets dealing with all aspects of life for the ostomy patient. These materials also are of great help to patients and nursing staffs in the absence of a registered enterostomal therapist (wound ostomy and continence nurse [WOCN]). The location of a registered therapist can be obtained from the Wound Ostomy and Continence Nurses Society (WOCN Society National Office, 15000 Commerce Parkway, Suite C, Mt. Laurel, NJ 08054; www.wocn.org).

CONTINENT ILEOSTOMY (KOCK POUCH) Clearly, one of the major (social) drawbacks to ileostomy could be eliminated if a continent stoma were possible. Nils Kock reasoned that a pouch and nipple valve constructed of terminal ileum could store ileal content internally until emptied voluntarily by the patient passing a large, soft catheter into the pouch several times daily, thereby obviating the need for an external appliance (Fig. 113-2). The first such operations were reported in 1969, and the results were promising; however, the nipple valve sometimes failed, usually because it slipped out of the pouch, resulting in incontinence.5 Techniques gradually improved, and the most recent approaches have been more successful, providing continence in most patients. In two series, more than 90% of patients were continent for gas and feces (i.e., never requiring an appliance).34,39 This high success rate, however, is achieved at the price of additional operations in most patients for nipple or pouch dysfunction, fistula, or stricture. Wasmuth and colleagues reported a 50% rate of reoperation by 14 years after

2017

2018

Section X  Small and Large Intestine ILEAL POUCH-ANAL ANASTOMOSIS

Ileal pouch Valve or nipple Stoma

Figure 113-2.  The continent ileostomy. The pouch is formed from a loop of ileum, folded on itself as a U, and sutured along the antimesenteric borders. The two limbs making up the pouch are then incised, exposing the mucosa, and the nipple valve is fashioned. The pouch is closed and positioned, as shown, underneath the abdominal wall. Note that the stoma is flush with the skin. (© Copyright 1991, Mayo Clinic, Rochester, Minnesota.)

continent ileostomy construction.40 Furthermore, in a series reported by Lepisto, 59% of patients with a continent ileostomy required reoperation, with a total of 85 pouch reconstructions being performed: 42 patients had one reconstruction, nine had two reconstructions, three had three reconstructions, one had four reconstructions, and two had six reconstructions.41 Others have reported similar findings: patients generally did well after initial continent ileostomy construction, but a sizable minority required repeated sur­ gical intervention either to salvage pouch function or remove the pouch.42,43 Despite requiring numerous reoperations, the majority of Kock pouch patients are satisfied with the outcomes of their functioning pouch. In a recent comparative study on quality of life in patients with standard ileostomies, ileal pouch, and Kock pouch, the Kock pouch patients did not fare significantly better or worse than those with a conventional ileostomy or IPAA; 56% of the continent ileostomy patients, however, did require reoperation to maintain function of the continent ileostomy.26 The fundamental mechanical problem of the nipple valve design in the Kock pouch has prevented widespread acceptance of the procedure. Another continent ileostomy, the T pouch, has been developed to combat this problem44; its design prevents slippage of the intussusceptive nipple valve constructed in the traditional Kock pouch. In the T pouch, the valve mechanism is made by securing an isolated distal ileal segment into a serosal-lined trough formed by the base of two adjacent ileal segments. The high volume-low pressure reservoir is fashioned around this isolated valve segment. Once constructed, the distal end of the valve mechanism is brought up through the skin as a stoma. T pouches have been constructed in only a few patients, and the results are promising, but long-term follow-up is required to assess the structural integrity and clinical success of the new valve design. Given the wide acceptance of the IPAA, continent ileostomy operations are performed rarely and used mainly in patients who have had a proctocolectomy and ileostomy and who desire enteric continence.

IPAA is now the procedure of choice for most patients who require proctocolectomy for chronic UC or FAP. IPAA is not considered suitable for patients with Crohn’s disease, although this recommendation now is being questioned.45,46 The operation has several major advantages: nearly all mucosal disease is removed in contrast to ileorectostomy; the normal route for elimination is maintained (a permanent stoma is not required); the anal sphincters are undisturbed; and the pelvic dissection, being less extensive than in cancer operations, should not endanger innervation of the sexual organs. The general principle of ileoanal anastomosis was first described in 1947, and its revival was influenced by the success of pediatric surgeons in children with Hirschsprung’s disease.7 Early approaches used a straight pull-through and sutured the ileum directly to the anal verge.47 Although results in children were encouraging, excessive stool frequency and anal seepage were unacceptable to many adult patients. Subsequently, the operation was modified to include one of several forms of ileal pouch. The basic surgical steps are as follows: a proctocolectomy is performed; the distal rectal mucosa is divided at the top of the anal canal, which leaves a small cuff of residual rectal and anal canal mucosa; an ileal pouch is fashioned and then stapled or sutured to the cuff of remaining rectal and anal canal tissue. A diverting ileostomy is usually required for two to three months until the anastomosis heals completely. At a second operation eight to 12 weeks later, the diverting ileostomy is closed. The Mayo Clinic has acquired considerable experience with IPAA, having performed more than 2200 of these operations.48-50 Although pouches of different configurations have been advocated by various surgical groups in the past, the pouch routinely used today is the J pouch because of its ease of construction and reliable function (Fig. 113-3).

LONG-TERM RESULTS

IPAA is a complex operation and complications occur frequently. The overall morbidity rate still hovers between 25% and 30%.49-51 Failure, however, is rare, even in those who suffer a postoperative complication. At the Mayo Clinic, overall pouch success rate was 92% in patients who had had their IPAA for up to 20 years.52 In this series of 1885 IPAA operations performed for UC over a 20-year period with a mean follow-up of 11 years, the overall rate of pouch success at 5, 10, 15 and 20 years was 96.3%, 93.3%, 92.4% and 92.1%, respectively. Over time, the mean daytime stool frequency increased from 5.7 times at one year to 6.4 times at 20 years; nighttime stool frequency also increased from 1.5 to 2.0. The incidence of frequent daytime fecal incontinence increased from 5% to 11% during the day (P < 0.001) and from 12% to 21% at night (P < 0.001). This series demonstrated that IPAA is a reliable surgical procedure for patients requiring proctocolectomy for UC and indeterminate colitis. Furthermore, it showed that the clinical and functional outcomes are excellent and durable. When performed in centers that have experience with IPAA, the outcomes are very good; even more important, these centers understand the potential complications and their management in IPAA patients. Although it is not discussed any further here, the key to a successful outcome is a surgeon who performs the operation comfortably; the operation struggled through is the one fraught with complications and sometimes with failure.

Chapter 113  Ileostomy, Colostomy, and Pouches

Figure 113-4.  Computed tomography (CT) scan showing a peripouch abscess (arrows) identified 8 days after surgery. This complication is best managed by placement of a percutaneous drain under CT guidance.

Figure 113-3.  The anatomy of the most commonly used type of ileal pouch, the J pouch. A pouch approximately 12 to 15 cm long is constructed by opening the common wall between the two limbs of the J formed from the distal terminal ileum. The apex is then anastomosed to the upper anal canal.

Complications

Pelvic sepsis, an ominous development, occurs in 5% to 24% of patients after IPAA.49,50,53,54 Computed tomography (CT) is useful for demonstrating pelvic fluid collections or phlegmon. Patients with pelvic phlegmon usually respond to conservative treatment with broad-spectrum antibiotics and bowel rest, whereas patients with a pelvic abscess ideally should undergo CT-guided drainage, if technically feasible, or laparotomy and drainage (Fig. 113-4). The most commonly cited risk factor for pelvic sepsis is chronic or high-dose glucocorticoid use in the perioperative period.55 Pelvic sepsis can, in the short-term, lead to pouch excision, which fortunately is rare; long-term functional results of the pouch are worse, however, and there is a higher rate of pouch loss compared with patients who did not experience pelvic sepsis.52 A diverting temporary ileostomy, while minimizing the impact of pelvic sepsis, is associated with a number of complications.56 Closure of temporary ileostomies also may be associated with complications. Peritonitis occurred in 4% of patients and postoperative intestinal obstruction in 12% of patients. Proximal and distal serosal tears during stoma mobilization, in addition to anastomotic leaks, are important causes of peritonitis. If all extraperitoneal bowel (afferent and efferent limbs and the stoma itself) is resected, however, the chance of leaving an unrecognized perforation is nearly eliminated. Almost all patients have a web-like stricture of the ileoanal anastomosis before ileostomy closure (Fig. 113-5). This stricture generally can be dilated digitally without difficulty, but narrowing can recur and is the most common indication for surgical intervention after an IPAA.57 If the pouch retracts under anastomotic tension, heavy scarring can result in a long, fibrotic stricture. This type of stricture

Figure 113-5.  Film from a water-soluble radiocontrast enema performed before ileostomy closure demonstrating a mild anastomotic narrowing (arrows). This stricture was dilated in the operating room before the ileostomy was closed.

is manifested by increased straining to empty the pouch, a sensation of incomplete pouch evacuation, or a high stool frequency (more than 10 to 12 stools per day). Repeated anal dilation can prevent progression of the stricture.

Clinical Results

Following an IPAA, the average stool frequency is six stools during the day, with one stool at night.48-50,54 Daytime and nocturnal stool frequency and the ability to discriminate flatus from stool remain relatively stable over time, whereas the need for stool bulking and hypomotility agents declines. The lower stool frequencies six months after surgery compared with the frequency in the early postoperative period are likely attributable to increased pouch capacity over time.

2019

2020

Section X  Small and Large Intestine In the Mayo Clinic experience, major fecal incontinence (more than twice per week) occurs in 5% or less of patients during the day and 12% of patients during sleep.50 In contrast, minor episodes of nocturnal incontinence occur in up to 30% of patients at least one year after the operation. A pad must be worn by 28% of patients for protection against seepage. Minor perianal skin irritation is reported by 63% of patients. Patients older than 50 years have a higher daytime stool frequency (eight per day) than do patients younger than 50 years (six per day). Men and women have similar stool frequencies postoperatively, but women have more episodes of fecal soilage during the day and night; this is thought to be related to a shorter average anal canal length in women. Seventy-eight percent of patients report excellent continence one year after surgery, which remains unchanged at 10 years; 20% experience minor incontinence; and 2% have poor control. Of patients with minor incontinence at one year, 40% remain unchanged, 40% improve, and 20% worsen by 10 years.52 Nocturnal fecal spotting increases during the 10-year period, but not significantly.

Pouchitis and Cuffitis

A wide range of reported incidences suggests that the level of clinical suspicion and the diagnostic criteria for pouchitis vary greatly.58-60 An early experience in IPAA patients demonstrated that patients who have preoperative extra­ intestinal manifestations of UC had significantly higher rates of pouchitis than patients without such manifestations (39% with preoperative symptoms and 26% without).58 A recent study by Hoda and colleagues, however, demonstrated that whereas extraintestinal manifestations might indicate a predisposition to episodes of acute pouchitis, they are not predictors of chronic pouchitis.60 Furthermore, they report that patients at highest risk for developing chronic pouchitis suffered from postoperative complications, more specifically anastomotic and septic complications. Other investigators have suggested that there is a biological predisposition for pouchitis and that better risk stratification can be obtained by the preoperative use of serum markers of IBD.61-63 In the study by Hui and colleagues, 63% of patients who went on to develop chronic pouchitis had a positive perinuclear-staining antineutrophil cytoplasmic antibody (pANCA) preoperatively, and only 17% had negative serologic results.61 There remain questions about these findings, because not all authors have demonstrated such a relationship.64 Furthermore, although the serologic information is intriguing, it is unclear if it provides any information that would change surgical decision making before surgery, because the surgical options are limited. As more surgeons have begun to perform a double stapled anastomosis, as opposed to a mucosectomy and hand-sewn anastomosis, there is a remnant cuff of rectal mucosa. This cuff of epithelial tissue can experience intermittent or chronic activity of UC, which has been termed cuffitis.65 Cuffitis can cause symptoms similar to pouchitis. In a study of 61 IPAA patients with symptoms of pouchitis, 7% had cuffitis.66 Most patients with apparent pouchitis or cuffitis have intermittent symptoms or respond well to therapy. In a few, however, symptoms are severe and persistent enough to lead to surgical removal of the pouch. Patients present with increased volumes of output, bleeding, discomfort from the pouch, and general symptoms similar to those of the initial disease. Fever, anemia, and dehydration as a result of diarrhea are variably present; fecal incontinence also is common. Extraintestinal dermatologic and rheumatologic manifestations are seen occasionally.58

Table 113-1  Pouchitis Disease Activity Index* CRITERIA

SCORE

Clinical Postoperative Stool Frequency Usual 0 1 or 2 stools/day more than usual 1 2 ≥3 stools/day more than usual Rectal Bleeding None or rare 0 Present daily 1 Fecal Urgency or Abdominal Cramps None 0 Occasional 1 Usual 2 Fever (>100°F) Absent 0 Present 1 Endoscopic Edema 1 Granularity 1 Friability 1 Loss of vascular pattern 1 Mucoid exudate 1 Ulceration 1 Histologic Polymorphonuclear Leukocyte Infiltration None 0 Mild 1 Moderate + crypt abscess 2 Severe + crypt abscess 3 Percent of Mucosa That Is Ulcerated per Low-Power Field (Average) 1 <25 25-50 2 3 >50 *Pouchitis is defined as a total score of 7 or greater. The total score is the sum of the individual scores. Adapted from Sandborn WJ, Tremaine WJ, Batts KP, et al. Pouchitis after ileal pouch-anal anastomosis: A pouchitis disease activity index. Mayo Clin Proc 1994; 69:409.

In a patient with pouchitis, endoscopy shows the pouch mucosa to be reddened, swollen, and often ulcerated. The mucosa is friable and bleeds readily from minor trauma during endoscopy; inflammatory changes usually are confined to the pouch but also can be seen in the adjacent ileum. Biopsy specimens show a range of acute and chronic inflammatory changes depending on severity, and a pouch­ itis disease activity index combining clinical, endoscopic, and histologic features has been developed (Table 113-1).59 Gradual escalation of medical therapy for acute or chronic pouchitis is now the standard of care.67 Exclusion of possible etiologies that may require specific treatment is essential before initiating treatment for pouchitis. The following points must be remembered: Patients with ileal pouches are not immune to superimposed specific enteric infections; stool culture and microscopy for parasites are appropriate. Unsuspected Crohn’s disease in the pouch is always a major concern. The clinical features of the underlying colitis that led to the operation should be reviewed carefully, including gross and histologic examination of the resected bowel. A small but definite proportion of colitis falls into an unclassifiable group, and some patients with previously unrecognized Crohn’s disease present with pouchitis, which is actually a manifestation of (recurrent) Crohn’s disease. Surgically treatable problems such as an anastomotic stricture resulting in partial pouch outflow obstruction need to be looked for carefully and treated with dilation.

Chapter 113  Ileostomy, Colostomy, and Pouches Pathogenesis Acute, nonspecific inflammation of ileal pouches apparently reflects the propensity of the patient for IBD. Thus, pouchitis is much more common in patients with IBD than it is in patients with familial adenomatous polyposis (FAP). Still, patients operated on for FAP are not completely protected from pouchitis. The nonspecific pathologic features of pouchitis reflect the fact that pouchitis can result from different causes. Histopathology of healthy and diseased pouches has shown that chronic inflammation is usual, even when the patient is asymptomatic.68,69 Villus architecture is distorted, and colonic metaplasia is present in biopsies from most pouches, even in the absence of severe acute inflammation. Thus, these changes are considered natural sequelae of the altered anatomy, just as the histologic changes of experimental and clinical blind-loop syndrome have been attributed to bacterial overgrowth. Other possible causes of pouchitis have little support, including damage by bile acids or their bacterial metabolites and lack of short-chain fatty acids (SCFAs).59 Normal colonic mucosa uses SCFAs as a source of nutrition, and some authors have proposed that IBD can result when the colon is deprived of SCFAs.70 The clearest clinical experiment that tests this hypothesis is diversion colitis. Harig and coworkers proposed that diversion colitis is caused by deprivation of SCFAs,71 support for which is provided by the observation that diversion colitis improves in response to SCFA enemas. Ileal pouches contain high concentrations of SCFAs (≥100 mmol), however, and so a state of deprivation seems unlikely. Indeed, pouchitis has worsened or shown no predictable response to SCFA enemas.72 In a detailed evaluation of luminal factors, including fecal concentrations of bacteria, bile acids, and SCFAs, there were no differences between patients with or without pouchitis.59 Alteration in butyrate metabolism has been linked to pouchitis.73 At present, there is no definitive cause for pouchitis, but it most likely does involve an interaction between the pouch microenvironment and the patient’s underlying immune response to that environment. In the Mayo Clinic experience in patients who have had their IPAA for nearly 20 years, the rate of having at least one episode of pouchitis was 48% at 10 years, and it rose to 78% at 20 years.52 In this cohort, less than 5% of patients developed chronic pouchitis, and only 2% required pouch removal or permanent diversion. Treatment If diarrhea alone is the major complaint, treatment with simple antidiarrheal measures may be all that is required. For more severely symptomatic patients, a variety of empiric treatments have emerged. When the condition was first encountered in continent ileostomies, anecdotal evidence was that constant drainage would help, based on the assumption that stasis was an important predisposing factor.23 Stasis should be a lesser factor after IPAA, although incomplete emptying (e.g., with the S pouch) or a persistent anastomotic stricture might need to be excluded. Metronidazole, 500 mg twice daily for 28 days, has been used often as a first line of treatment. The response to metronidazole or other broad-spectrum antibiotics usually is dramatic. Some patients relapse after initial therapy with antibiotics, and they require subsequent courses of treatment. In general, antibacterial agents with a spectrum of activity against anaerobes have been most successful in treating pouchitis. When antibiotics are ineffective, treatment should consist of regimens that are effective in IBD: glucocorticoid enemas, aminosalicylates, mesalamine enemas, and even systemic

glucocorticoids. In some cases of antibiotic resistant pouch­ itis, bismuth subsalicylate (Pepto-Bismol) 270 mg daily for three weeks is an effective treatment. Most patients unresponsive to antibiotics improve on one of these programs. Severe recurrent disease or major extraintestinal symptoms rarely require pouch removal. Sequelae Although the prevalence of chronic pouchitis is low, the possible consequences of chronic inflammation of the neorectum, especially dysplasia and malignant change, arouse concern. Morphologic and biochemical changes occur in the ileal mucosa of pouches, including villus blunting, chronic inflammatory infiltrates, variable transition to production of a colonic type of mucus (sulfomucins), and increased cellular proliferation.74 Observations based on long-term follow-up (mean of 6.3 years) reveal three patterns of mucosal adaptation: approximately one half of the patients showed mild villus atrophy and minimal inflammation, slightly fewer had transient moderate or severe atrophy and inflammation with intervals of recovery, and approximately 10% had permanent subtotal or total villus atrophy with chronic inflammation.68 In this study, three of eight patients developed low-grade dysplasia; in one patient it developed two years postoperatively. In patients having a double-stapled anastomosis, the remaining rectal cuff and anal canal transition zone tissue represent at-risk tissue. In one series of 225 patients with a stapled IPAA, 238 rectal cuff and anal canal biopsies were obtained; 202 (84.9%) had histologically confirmed chronic inflammation, 11 (4.6%) had acute inflammation, and 25 (10.5%) were read as normal.75 Interestingly, 9 of the 11 patients with acute inflammation were asymptomatic.

Pouch Failure

Large series have reported failure rates between 2% and 12%, and in the Mayo clinic series, 8% of patients ultimately required pouch excision or construction of a permanent ileostomy.52 The most common causes of pouch failure, either alone or in combination, include pelvic sepsis, high stool volumes, Crohn’s disease, and uncontrollable fecal incontinence; pouchitis is the sole cause in 2% of all patients. Of the patients in whom the pouch fails, failure occurs within 1 year in 75%, by 2 years in 12%, and by 3 years in 12%. Fortunately, pouch failure is relatively uncommon. Early failures are almost always related to technical issues or complications related to the original operation, whereas late failures are more commonly related to chronic pouchitis or Crohn’s disease in the pouch.

Quality of Life

Often, quality of life is the deciding factor for patients choosing a particular operation for UC. Several studies that analyzed the outcome of surgery for UC have demonstrated that most patients are satisfied with the operation and have a normal lifestyle regardless of the procedure. In one study of quality of life after a Brooke ileostomy or IPAA for UC and FAP, patients were highly satisfied with either operation (Brooke ileostomy, 93%; IPAA, 95%).27 Daily activities (e.g., sexual life, participation in sports, social interaction, work, recreation, family relationships, travel), however, were more likely to be adversely affected with a Brooke ileostomy than by IPAA.

Sexual Dysfunction

Impotence and retrograde ejaculation developed in 1.5% and 4% of men, respectively. Dyspareunia developed in 7% of women postoperatively.49,50 Early studies of IPAA focused

2021

2022

Section X  Small and Large Intestine on physiologic assessment, but recent studies have concentrated on more multidomain quality-of-life assessments. A prospective evaluation of sexual function in patients with IPAA was performed using validated survey instruments, including the International Index of Erectile Function in men and the Female Sexual Function Index in women. Overall quality of life was assessed using the Short Inflammatory Bowel Disease Questionnaire. Preoperative scores were compared with scores at six and 12 months postoperatively. Of the 59 patients who completed the study, male sexual function and erectile function scores remained high 12 months after surgery and female sexual function improved 12 months after surgery. Quality of life significantly improved after IPAA in both men and women.76 Others have reported similar improvements after IPAA.77,78

CONTROVERSIES Double-Stapled versus Hand-Sewn Anastomosis

Much of the debate whether to staple the anastomosis or not has evolved because functional outcomes should be improved if the anal transition zone (ATZ) is preserved. In nonrandomized trials, a stapled anastomosis has been equated with better outcome, which in turn has been attributed to less injury of the anal sphincters; preservation of the ATZ and hence anal sensory discrimination; and preservation of the rectoanal inhibitory reflex. To determine if stapled IPAA conferred any advantage over hand-sewn IPAA, we conducted a randomized study at the Mayo Clinic in which 41 patients were randomized to double-stapled (17 patients) or hand-sewn (15 patients) technique.79 In the stapled group, 1.5 to 2.0 cm of ATZ was preserved, whereas complete mucosectomy was performed in the hand-sewn group. Overall, complications were the same in the two groups. Stool frequency and rates of fecal incontinence during the day and night were similar between the groups; however, fewer patients treated with the doublestapled technique had nocturnal incontinence. Similar findings have been reported by other groups.80 In a meta-analysis of more than 4000 patients, Silvestri and colleagues concluded that both techniques had similar early postoperative outcomes; stapled IPAA offered improved nocturnal continence, however, which was reflected in higher anorectal physiologic measurements.81

Role of Defunctioning Ileostomy

The most feared complication of IPAA is pelvic sepsis; therefore, a defunctioning ileostomy after pouch construction usually is performed to minimize its occurrence.82 In the literature, the reported rate of pelvic sepsis after IPAA varies between 0% and 25%. Although the incidence of pelvic sepsis is relatively low (6%) at the Mayo Clinic, when it occurs, it is responsible for a significant proportion of the failed pouches.52,57 Proponents of defunctioning ileostomies argue that diverting stomas allow the anal sphincter and ileal mucosa to recover before intestinal continuity is restored and that patients have a short-lived experience of a stoma to fully appreciate the ultimate benefit of IPAA. Use of a loop ileostomy does not appear to protect the patient fully from pelvic sepsis; however, its presence makes it easier to manage a patient with this complication. Supporters of a one-stage procedure believe that an IPAA can be performed without increased risk of pelvic sepsis.83-87 A one-stage procedure avoids an ileostomy and a second hospitalization and operation, lowers the total cost, and results in a shorter hospital stay and perhaps a decreased incidence of small bowel obstruction. In the large single-surgeon study reported by Sugerman and associates, there were no differences in the complica-

tion rates and functional outcomes of patients who had not had a diverting ileostomy compared with those who had a diverting ileostomy; there also was no relationship to gluco­ corticoid use.83,88 Whereas there might be no significant difference in the complication rate in patients without a diverting ileostomy, one study has suggested that the severity of complications was greater in patients without a protecting ileostomy.89 Although it is our practice to perform a diverting ileostomy in all patients undergoing an IPAA, in properly selected patients who have uncomplicated procedures performed by experienced surgeons, a one-stage IPAA might be appropriate. The surgeon and patient care team must be attentive to the early signs of pelvic sepsis, aggressively investigate the possibility of a pouch leak, and intervene as needed.

ADDITIONAL ISSUES Risk of Cancer

Patients with UC are at risk for developing adenocarcinoma of the colon, which increases with the duration of disease and extent of colonic involvement (see Chapters 112 and 123). Figures suggest that the risk of colon cancer for people with IBD increases by 0.5% to 1.0% yearly, eight to 10 years after diagnosis.90 Any surgery that leaves behind diseased colonic mucosa puts the patient at risk for developing dysplasia or neoplasia in this residual tissue. The risk of developing a carcinoma in the residual colonic mucosa may be directly related to the amount of residual mucosa remaining in situ. Complete excision of the rectal mucosa during IPAA substantially decreases the risk of dysplasia. With the more commonly used stapled IPAA, however, a small amount of residual rectal and anal canal tissue is retained. Early studies by Tsunoda and colleagues demonstrated the presence of dysplasia in mucosectomy specimens, which they believed supports the use of a mucosectomy and hand-sewn anastomosis.91 Even performing a mucosectomy, however, does not ensure that all rectal and at-risk anal canal mucosa is removed. In one study that evaluated anal canal specimens, islands (rests) of mucosa were present despite “complete” mucosal resection.92 In a study with long-term follow-up, however, the risk of dysplasia was quite low.93 A cohort of 289 patients with stapled IPAAs was followed and had multiple biopsies of the rectal cuff and anal transition zone performed over a 10-year period. Dysplasia was identified in eight patients, including four with low-grade and four with high-grade dysplasia. No cancer in the ATZ was found during the study period. The authors concluded that ATZ dysplasia after stapled IPAA was infrequent and usually self-limiting. ATZ preservation did not lead to the development of cancer in the ATZ with a minimum of 10 years of follow-up, although long-term surveillance is recommended to monitor dysplasia. Detection of neoplastic change in the pouch itself is another reason to perform follow-up in patients with IPAA. A subgroup of patients has been identified in whom the mucosa of the pelvic pouch develops severe villus atrophy.68 These patients seem to have a significantly higher incidence of dysplasia compared with patients without villus atrophy (71% vs. 0%). The former group may be at greater risk for developing carcinoma and might require more intensive follow-up with regular pouch endoscopy and biopsy. Despite these findings, dysplasia in a pouch is a rare event. In group of 45 patients followed for a median of six years (one to 28 years), dysplasia of any type was found in 4% of pouch biopsies and there was no evidence of malignancy.94 To date, there have been a small number of case reports of carcinomas arising in ileal pouches or in the region of the

Chapter 113  Ileostomy, Colostomy, and Pouches anastomosis.95-98 Surprisingly, many of these cancers have arisen in patients who have undergone a complete mucosectomy with hand-sewn anastomosis. In a recent review, Branco and colleagues found that the occurrence of adenocarcinoma following IPAA for UC is an infrequent event.99 Their conclusions were that post-IPAA cancer can occur following either mucosectomy or stapled anastomosis; that this malignancy can occur after IPAA performed for UC either with or without neoplasia; and that this complication is seen whether or not the initial cancer or dysplasia had involved the rectum. Given the known occurrence of dysplasia in pouch mucosa and the rare reports of cancers arising in pouches, routine clinical and endoscopic surveillance should be performed in patients after their IPAA.

Fertility and Pregnancy

Many patients with UC are in their childbearing years. Therefore, the impact of surgery on fertility is an important consideration when discussing surgery with a young woman. A number of studies have evaluated fertility and the course of a subsequent pregnancy after surgery.100,101 Patients who have had a proctocolectomy and end ileostomy or Kock pouch can expect to have a normal pregnancy and delivery; however, these women often have temporary stoma or Kock pouch dysfunction.102 A similar disturbance of pouch function is seen in IPAA patients, in whom a slight increase in stool frequency, incontinence, and pad use is reported during the pregnancy.100,101 Fortunately, this is temporary, and patients return to their baseline pouch function after the pregnancy. There is a reported higher rate of cesarean sections in IPAA patients, but there appears to be no contraindication to vaginal delivery, and the decision to proceed to a cesarean section should be based upon obstetric considerations. Previous studies evaluated the course of pregnancies after IPAA, but the specific issue of fecundity after IPAA had not been considered. A Swedish population-based study, however, demonstrated a significant reduction in fecundity after IPAA.103 More important, of the post-IPAA patients who became pregnant, 29% of pregnancies occurred only after in vitro fertilization compared with the expected 1% of all births in Sweden during the study period. The basis of this decreased fertility is unknown, but the authors hypothesize that changes in pelvic anatomy resulting from removal of the rectum and scarring from the pelvic dissection are major contributors to the problem. A case-control study comparing women with IPAAs and female controls who had previous abdominal surgery found that women who had an IPAA had significantly more infertility evaluations and need for infertility treatments.104 Furthermore, an analysis of infertility treatments for post-IPAA women demonstrated that they suffered a reduction in the probability of conception rather than complete infertility. This reduction in fecundity is not seen in women who have undergone ileorectal anastomosis for FAP, which further the supports the idea that postoperative adhesions or altered pelvic anatomy contribute to this problem.105 Given the growing evidence that IPAA reduces a woman’s fecundity, women need to be counseled regarding this prospect during the informed consent before IPAA. If this is a major concern in a young woman suffering from medically refractory UC, a subtotal colectomy and ileostomy to control the disease without disturbing the pelvic anatomy may be offered with a planned IPAA after completion of childbearing.

Ileal Pouch-Anal Anastomosis and Indeterminate Colitis

Among 1519 consecutive patients with UC undergoing IPAA between January 1981 and December 1995, 82 patients (5%) had features of indeterminate colitis, including unusual dis-

tribution of inflammation, deep linear ulcers, neural proliferation, transmural inflammation, fissures, and creeping fat.106 We found that 12 (15%) of the 82 patients with indeterminate colitis eventually developed Crohn’s disease during follow-up, compared with only 26 (2%) of 1437 patients with UC. The probability of remaining free of Crohn’s disease at 10 years was 98% in patients with UC and 81% in the indeterminate colitis patients. After IPAA, patients with indeterminate colitis who did not develop Crohn’s disease experienced long-term outcomes nearly identical to those of patients with UC; that is, nearly 85% had functioning pouches 10 years after the operation. Crohn’s disease, however, regardless of whether it develops and is diagnosed after IPAA operations for UC or indeterminate colitis, is associated with poor long-term outcomes. Similar to our findings, other institutions have shown that although pouch complications are higher in patients with indeterminate colitis, the functional results after IPAA for indeterminate colitis are identical to those after UC.107

Impact of Biological Medical Therapy

The inclusion of newer biological therapies, such as infliximab, in the treatment of UC has raised concerns regarding their impact on the surgical outcomes after IPAA. In a study by Selvasekar and colleagues, 47 UC patients received preoperative (IPAA) infliximab therapy and 254 did not.108 Patients who received infliximab were statistically more likely to have postoperative infectious complications and pelvic abscesses. After multivariate adjustment for disease severity and other medication use, infliximab remained independently associated with an increased risk of ileal pouch-related and infectious complications. Another study of 85 patients with UC who received infliximab preoperatively found that they were at increased risk for postoperative septic complications as well as late complications compared with patients who did not receive infliximab.109 The authors noted that patients who received infliximab were more likely to have undergone a three-stage IPAA, likely due to surgeon reluctance to perform an IPAA in the setting of preoperative infliximab administration. Both of these studies are limited in their ability to draw any conclusion regarding the exact role of infliximab in the increased postoperative infection rate due to the retrospective nature of the analyses and possible selection bias. Most likely, patients who require infliximab for treatment of their UC represent a sicker and thus higher-risk group of patients at the time of surgery. Further prospective studies need to be performed to clarify this important issue. As previously discussed, pelvic sepsis and abscess are devastating postoperative complications and are the leading risk factor for ileal pouch loss.52 Given the uncertainty about the role of biologic therapy contributing to this problem or the possibility of much sicker patients presenting for surgery, a three-stage approach in patients who have received preoperative infliximab may be considered.

ABDOMINAL COLECTOMY AND ILEORECTAL ANASTOMOSIS The aim of a colectomy with an ileorectal anastomosis (IRA) is to extirpate most of the diseased colonic mucosa, thus reducing the risks of hemorrhage, megacolon and its complications, and malignant degeneration, while allowing the rectum to retain continence for stool and gas. The rationale for an IRA is that the operation avoids a permanent stoma, minimizes or eliminates injury to the pelvic nerves, and is easy to perform. Other operations, if they become necessary,

2023

2024

Section X  Small and Large Intestine are not precluded.110-113 The rationale against the operation, however, is nearly as convincing. Subsequent proctectomy is required in 6% to 37% of patients, poor results have been reported in up to 50%, and the risk of developing carcinoma in the retained rectum approaches 17% after 27 years.113,114 In a study by Pastore and colleagues, who reviewed the course of 48 patients with UC and 42 patients with Crohn’s disease who underwent an IRA, 84% of the UC and 91% of the Crohn’s disease patients reported an improvement in their quality of life.115 One patient with UC developed carcinoma of the rectal stump 11.5 years after the colectomy and IRA (cumulative probability of remaining free of cancer, 85.7% at 12 years). In patients with Crohn’s disease and minimal or no rectal involvement, IRA with excision of the diseased colon is an appropriate operation. In addition, IRA, as a sphincter-saving procedure, continues to have a place in the surgical treatment of UC for high-risk or older patients who are not good candidates for IPAA and have relatively mild disease. It is also important that these patients understand the need for continued screening of the rectum because of possible development of a malignancy. The quality of life after IRA has been reported to be good; patient satisfaction is high, and an active, productive lifestyle can be preserved.115,116

PATIENT SELECTION

Patients are candidates for IRA if the rectum is distensible, if the disease (UC, Crohn’s) involves the rectum only minimally, and if patients are willing to undergo follow-up screening for rectal cancer. Patients of any body habitus may undergo IRA. Although there is no maximum age that contraindicates an ileorectostomy, functional results related to poor sphincter function must be considered in older patients, especially in older female patients.

COMPLICATIONS

Operative mortality for elective IRA has been reported to vary between 2% and 8%.116,117 In the series of UC patients reported by Leijonmarck and colleagues, complications occurred in seven of 43 patients (16%) undergoing an elective, one-stage procedure.116 There were two postoperative deaths (4%). Twenty-two patients (43%) had a functioning IRA at the time of follow-up, with a mean time of observation of 13 years. The cumulative probability of having the IRA in function at 10 years was 51%. The causes of total excision were recurrent inflammation in the retained rectum (N = 23), dysplasia (N = 3), and postoperative complications (N = 3). No rectal carcinoma occurred.

PHYSIOLOGY

The primary attraction of an IRA is that the major anatomic mechanisms responsible for maintaining continence are retained: The rectal reservoir, the pelvic floor, and the internal and external anal sphincters are preserved. The absorptive capacity of the proximal colon is lost, however, and ileal content is continuously presented to the rectal remnant. The rectum should be large and compliant to allow passive accommodation. Compliance depends on rectal wall elasticity, and in active inflammatory disease, compliant accommodation is impaired due to the inflammation. Therefore, the less the mucosal inflammation, the better the compliance and the fewer stools per day. Sphincteric function in patients with IRA differs little from that in normal people. After an IRA, patients with quiescent rectal disease should be able to absorb water and sodium in the rectal segment.

Moreover, with quiescent disease, the rectum is capacious and distensible, resulting in low stool frequency and little or no incontinence. Conversely, if the rectal mucosa is inflamed, absorption is impaired and fecal volume is greater; moreover, the more inflamed the mucosa, the less capacious and distensible will be the rectum, and increased stool frequency, urgency, and incontinence will follow. The quality of life after IRA is good; satisfaction is high and a productive lifestyle is achieved. Overall satisfaction is tempered, however, by the fact that patients often know they have not been cured and because they must undergo frequent follow-up examinations.

COLOSTOMY IN THE MANAGEMENT OF ULCERATIVE COLITIS

For patients with UC, partial colectomy or colostomy has no role in management, either electively or emergently. Partial colectomy of predominantly left-sided colitis leaving the uninvolved remaining proximal colon often results in recurrent disease activity. Resection of the rectosigmoid, coloanal anastomosis, and cecoanal anastomosis are operations for UC mentioned here only to be condemned. Urgent or emergent intervention for decompensating or fulminating UC is best managed by abdominal colectomy, Brooke ileostomy, and either oversewing the rectum or establishing a rectal mucous fistula.

LAPAROSCOPIC APPROACH The most important change in surgical practice related to all of these procedures is the application of laparoscopic techniques. Minimal-access colon surgery began in the early 1990s; however, improvements in image technology and instrumentation have only recently facilitated complex colorectal procedures. Laparoscopic approaches for IPAA were developed to reduce the impact of this procedure on patients who are already physiologically stressed, decrease length of hospital stay, reduce morbidity, and improve cosmesis. The initial reports of laparoscopic IPAA were discouraging, however, because of very long operative times and few observed postoperative benefits.118-120 Subsequent reports, from our institution and others, have clearly demonstrated the benefits of a minimally invasive approach in regard to postoperative decreases in length of stay, need for narcotics, and overall morbidity.121-123 At the Mayo Clinic, the minimally invasive approach has become the preferred surgical operation for IPAA. Indications for operative intervention are not changed by the laparoscopic approach. In a case-matched series, 40 patients undergoing laparoscopic IPAA (LAP) were matched to open controls. The patients were matched for disease, age, gender, body mass index (BMI), and date of operation. The LAP group exhibited significant benefits in time to ingesting clear liquids (1 vs. 3 days; P < 0.001), eating a regular diet (3 vs. 4 days; P < 0.001), and regaining bowel function (2 vs. 3 days; P < 0.001).93 Duration of narcotic use was shorter in the LAP group (P < 0.001), and length of stay was reduced (4 vs. 7 days; P < 0.001). LAP patients had longer operative times (270 vs. 192 minutes; P < 0.001), but operative time decreased with experience and now averages 180 to 210 minutes. Subsequent studies within our institution have continued to demonstrate these short-term patient benefits.124

Chapter 113  Ileostomy, Colostomy, and Pouches Table 113-2  Comparison of Surgical Options after Colectomy

MORTALITY, %

OVERALL MORBIDITY, %

SMALL BOWEL OBSTRUCTION, %

PERINEAL WOUND COMPLICATION, %

STOOLS PER 24 HOURS

FAILURE, % —

ALL DISEASE REMOVED?

CANCER RISK, %

Yes

0

SURGERY

STOMA

CONTINENCE

Brooke ileostomy Continent ileostomy IPAA IRA

Yes

No

<1.0

19-70

15

33

NA

Yes

Yes

<1.0

15-60

7

35

3-5

50

Yes

†

No No

Yes Yes

<1.0 2.5-8.0

30-50 16-20

22 15

NA NA

5-7 1-3

8 24-60

Yes* No

* 15 (30 yr)

DISEASE INDICATION CD (?UC, FAP) UC, FAP UC, FAP CD, UC

Note: Maximum follow-up time is 13 years. *Ten instances of neoplasia in the cuff or pouch after ileal pouch-anal anastomosis have been reported in more than 12,000 cases. † Two cancers in Kock pouches (continent ileostomies) have been reported. CD, Crohn’s disease; FAP, familial adenomatous polyposis; IPAA, ileal pouch-anal anastomosis; IRA, ileorectal anastomosis; NA, not applicable; UC, ulcerative colitis.

RISK-BENEFIT ANALYSIS CONVENTIONAL ILEOSTOMY

The Brooke ileostomy is safe and reliable and has broad applicability to patients with IBD who require proctocolectomy. It is not, however, entirely free of complications (Table 113-2). Up to 30% of patients have a septic complication, 20% to 25% require revision of the stoma, 15% have recurrent small bowel obstruction, and stomal dysfunction can occur in up to 30% of cases.

ILEORECTAL ANASTOMOSIS

The primary benefit of an IRA is that the rectum is undisturbed by the operative dissection, the normal pathway of defecation is left in situ, and the incidence of bladder or sexual problems is low. There also is no perineal wound (see Table 113-2). In many patients, the overall functional results are reasonably good. The major problem with an IRA is that actual or potentially diseased mucosa is left in situ. In a few patients, inflammatory changes do resolve, but in most, the disease process continues unabated. The sequelae of leaving disease behind include poor anastomotic healing, which is responsible for the relatively higher mortality after IRA than after continent ileostomy and IPAA; continued need for antiinflammatory therapy; continued bleeding and mucus discharge; incontinence and high stool frequency when inflammation flares up; and the possibility of malignant degeneration.

CONTINENT ILEOSTOMY

A major benefit of the Kock pouch is that although a stoma is constructed, discharge is controlled without the need for an external appliance. The principal problem with continent ileostomy is the high rate of complications, usually involving the nipple valve, and leading to incontinence or to complete outflow obstruction; these complications, in turn, almost always require another operation. This operation is rarely performed today because the success of IPAA has made it obsolete.

ILEAL POUCH-ANAL ANASTOMOSIS

The major benefit of IPAA is that it successfully restores fecal continence in most patients; the major problem is the high complication rate of approximately 30%. Occasional incontinence appears early in almost all patients after operation, particularly at night. Major episodes of daytime incontinence affect approximately 10% of patients, but this frequency declines to almost zero after four years. Other

complications are pelvic infection, stricture, fistulas, sinus tracts, pouch leakage, and small bowel obstruction. As surgeons’ experiences with the operation have broadened, these surgical complications have decreased in frequency. Although nonspecific inflammation of the pouch (pouchitis) is the most important current drawback, in most patients it is treated effectively and simply with antibiotics. When severe and recurrent, pouchitis can lead to failure of the operation, but this is uncommon. Despite these problems, the benefits of IPAA are clear: All disease is removed, the patient does not have a stoma, and anal defecation is voluntary and controlled.

KEY REFERENCES

Farouk R, Pemberton JH, Wolff BG, et al. Functional outcomes after ileal pouch–anal anastomosis for chronic ulcerative colitis. Ann Surg 2000; 231:919-26. (Ref 50.) Hahnloser D, Pemberton JH, Wolff BG, et al. The effect of ageing on function and quality of life in ileal pouch patients: A single cohort experience of 409 patients with chronic ulcerative colitis. Ann Surg 2004; 240:615-21. (Ref 51.) Hahnloser D, Pemberton JH, Wolff BG, et al. Results at up to 20 years after ileal pouch–anal anastomosis for chronic ulcerative colitis. Br J Surg 2007; 94:333-40. (Ref 52.) Israelsson LA. Preventing and treating parastomal hernia. World J Surg 2005; 29:1086-9. (Ref 34.) Larson DW, Cima RR, Dozois EJ, et al. Safety, feasibility, and short-term outcomes of laparoscopic ileal-pouch–anal anastomosis: A single institutional case-matched experience. Ann Surg 2006; 243:667-72. (Ref 123.) Olsen K, Joelsson M, Laurberg S, et al. Fertility after ileal pouch–anal anastomosis in women with ulcerative colitis. Br J Surg 1999; 86:493. (Ref 103-5.) Pemberton JH, Phillips SF, Ready RR, et al. Quality of life after Brooke ileostomy and ileal pouch–anal anastomosis. Ann Surg 1989; 209:62028. (Ref 27.) Remzi FH, Fazio VW, Delaney CP, et al. Dysplasia of the anal transitional zone after ileal pouch–anal anastomosis: Results of prospective evaluation after a minimum of ten years. Dis Colon Rectum. 2003; 46:6-13. (Ref 93.) Selvasekar CR, Cima RR, Larson DW, et al. Effect of infliximab on shortterm complications in patients undergoing operation for chronic ulcerative colitis. J Am Coll Surg 2007; 204:956-62. (Ref 108.) Shen B, Fazio VW, Remzi FH, et al. Comprehensive evaluation of inflammatory and non-inflammatory sequelae of ileal pouch–anal anastomosis. Am J Gastroenterol 2005; 100:93-101. (Ref 65.) Weise WJ, Serrano FA, Fought Gennari FJ. Acute electrolyte and acidbase disorders in patients with ileostomies: A case series. Am J Kidney Dis 2008; 52:494-500. (Ref 19.) Yu E-D, Shao Z, Shen B. Pouchitis. World J Gastroenterol 2007; 13:5598604. (Ref 67.) Ziv Y, Church JM, Fazio VW, et al. Effect of systemic steroids on ileal pouch–anal anastomosis in patients with ulcerative colitis. Dis Colon Rectum 1996; 39:504-8. (Ref 55.) Full references for this chapter can be found on www.expertconsult.com.

2025

CHAPTE R

114 Intestinal Ischemia Lawrence J. Brandt and Paul Feuerstadt

CHAPTER OUTLINE Anatomy of the Splanchnic Circulation  2027 Celiac Axis  2027 Superior Mesenteric Artery  2027 Inferior Mesenteric Artery  2028 Collateral and Anastomotic Circulation  2028 Pathophysiology and Pathology  2028 Acute Mesenteric Ischemia  2029 Incidence  2030 Clinical Features  2030 Laboratory Features and Diagnosis  2030 Treatment  2032 Mesenteric Venous Thrombosis  2036 Incidence  2036 Predisposing Conditions  2036 Pathophysiology  2036 Clinical Features  2036 Diagnosis  2036 Treatment  2037 Prognosis  2038 Focal Segmental Ischemia of the Small Intestine  2038 Colon Ischemia  2038 Incidence  2038 Pathophysiology and Causes  2038 Pathology  2040

Intestinal ischemia produces a broad spectrum of disorders, depending on the onset, duration, and cause of the injury; the area and length of bowel affected; the vessel involved; and the degree of collateral blood flow. Variability in these factors influences not only the presentation of the ischemic event but also its treatment and outcome. Ischemic injury may be acute or chronic. It may be caused by a disturbance in the arterial supply or venous drainage of the bowel and involve the small intestine, the colon, or both. Since the development and widespread use of colonoscopy, angiography, computed tomography (CT), and other imaging modalities, various types of ischemic injury to the gastrointestinal tract have been recognized and increasingly appreciated (Table 114-1; see also Tables 114-2 and 114-4). Our concepts of their pathogenesis, diagnosis, and management have been so altered since the 1980s that much of what has been written in the past is no longer applicable. In this chapter we describe the spectrum of ischemic damage to the gastrointestinal tract and discuss the management of these conditions in light of recent advances.

ANATOMY OF THE SPLANCHNIC CIRCULATION The celiac axis (CA), superior mesenteric artery (SMA), and inferior mesenteric artery (IMA) supply almost all of the

Clinical Features and Diagnosis  2041 Clinical Course and Treatment  2042 Special Clinical Problems  2043 Chronic Mesenteric Ischemia (Intestinal Angina)  2044 Clinical Features  2044 Diagnosis  2045 Treatment  2045 Vasculitis and Angiopathy of the Splanchnic Circulation  2046 Allergic Granulomatous Angiitis (Churg-Strauss Syndrome)  2046 Behçet’s Disease  2046 Buerger’s Disease  2046 Cogan’s Syndrome  2046 Fibromuscular Dysplasia  2046 Henoch-Schönlein Purpura  2047 Hypersensitivity Vasculitis  2047 Kawasaki’s Disease  2047 Köhlmeier-Degos Disease (Malignant Atrophic Papulosis)  2047 Polyarteritis Nodosa  2047 Rheumatoid Vasculitis  2047 Systemic Lupus Erythematosus  2047 Takayasu’s Disease  2047

blood flow to the digestive tract.1 There is marked variability of vascular anatomy among individuals, but typical patterns have emerged from anatomic dissections and abdominal angiographic studies.

CELIAC AXIS

The CA (Fig. 114-1) arises from the anterior aorta and typically gives rise to three major branches: the left gastric artery, the common hepatic artery, and the splenic artery. The common hepatic artery gives rise to the gastroduodenal, right gastroepiploic, and anterior superior pancreaticoduodenal arterial branches. The splenic artery gives off pancreatic and left gastroepiploic arterial branches. The CA and its branches supply the stomach, duodenum, pancreas, and liver.

SUPERIOR MESENTERIC ARTERY

The SMA (Fig. 114-2) has its origin from the anterior aorta near the neck of the pancreas. It gives rise to five major vessels: the anterior and posterior inferior pancreaticoduodenal vessels, middle colic, right colic, and ileocolic arteries, as well as to a series of jejunal and ileal branches, all of which supply their named portions of intestine. These intestinal branches typically form a series of arcades, and from the terminal arcade, numerous straight vessels arise that enter the intestinal wall.

2027

2028

Section X  Small and Large Intestine A H

MC

LG S

CA

PM

A

CP

DP

GD

RC

SMA

TP PIPD

PIPD AIPD

ASPD AIPD

JEJ

SMA IC

RGE COL Figure 114-1.  Diagram of typical celiac axis (CA) anatomy and its anastomoses with the superior mesenteric artery (SMA). A, aorta; AIPD, anterior inferior pancreaticoduodenal artery; ASPD, anterior superior pancreaticoduodenal artery; CP, caudal pancreatic artery; DP, dorsal pancreatic artery; GD, gastroduodenal artery; H, common hepatic artery; LG, left gastric artery; PIPD, posterior inferior pancreaticoduodenal artery; PM, pancreata magna; RGE, right gastroepiploic artery; S, splenic artery; TP, transverse pancreatic artery. (From Nebesar RA, Kornblith PL, Pollard JJ, Michels NA. Celiac and superior mesenteric arteries: a correlation of angiograms and dissections. Boston: Little, Brown; 1969.)

Table 114-1  Types and Approximate Frequencies of Intestinal Ischemia TYPE Colon ischemia* Acute mesenteric ischemia* Focal segmental ischemia* Chronic mesenteric ischemia

FREQUENCY (%) 75 25 <5 <5

*Includes mesenteric venous thrombosis. Mesenteric venous thrombosis can manifest as colon ischemia, acute mesenteric ischemia, or focal segmental ischemia.

INFERIOR MESENTERIC ARTERY

The IMA (Fig. 114-3) arises 3 to 4 cm above the aortic bifurcation close to the inferior border of the duodenum. It branches into the left colic artery, gives off multiple sigmoid branches, and terminates as the superior rectal artery. The IMA and its branches supply the large intestine from the distal transverse colon to the proximal rectum. The distal rectum is supplied by branches of the internal iliac (hypogastric) artery.

COLLATERAL AND ANASTOMOTIC CIRCULATION

Abundant collateral circulation to the stomach, duodenum, and rectum accounts for the paucity of ischemic events in these areas. The major anastomosis between the CA and the SMA is formed from the superior pancreaticoduodenal branch of the CA and the inferior pancreaticoduodenal branch of the SMA. These vessels constitute the pancreaticoduodenal arcade and provide blood to the duodenum and the pancreas. The splenic flexure and sigmoid colon have limited anastomoses, and ischemic damage is more common in these locations. There are three potential paths of communication between the SMA and IMA: the marginal artery of Drummond, which

IL

Figure 114-2.  Diagram of typical superior mesenteric artery (SMA) anatomy. AIPD, anterior inferior pancreaticoduodenal artery; COL, colic branches; IL, ileal branches; IC, ileocolic artery; JEJ, jejunal branches; MC, middle colic artery; PIPD, posterior inferior pancreaticoduodenal artery; RC, right colic artery. (From Nebesar RA, Kornblith PL, Pollard JJ, Michels NA. Celiac and superior mesenteric arteries: a correlation of angiograms and dissections. Boston: Little, Brown; 1969.)

is closest to and parallel with the wall of the intestine; the central anastomotic artery, a larger and more centrally placed vessel; and the arc of Riolan, an artery in the base of the mesentery. In the presence of SMA or IMA occlusion, a large collateral termed the meandering artery may be identified angiographically and represents a dilated central anastomotic artery or arc of Riolan (Fig. 114-4). It is critical to determine the direction of flow within a meandering artery before sacrificing the IMA, such as during aortic aneurysm surgery, lest the IMA be the main vessel supplying blood to the small bowel because of an occluded SMA.

PATHOPHYSIOLOGY AND PATHOLOGY Ischemic injury of the intestine results from deprivation of oxygen and nutrients necessary for cellular integrity. Remarkably, the bowel can tolerate a 75% reduction of mesenteric blood flow and oxygen consumption for 12 hours with no changes on light microscopy, because only one fifth of the mesenteric capillaries is open at any time, and when oxygen delivery is decreased, the bowel adapts by increasing oxygen extraction.2 Below a critical level of blood flow, however, these compensatory mechanisms are overwhelmed and no longer protective. When a major vessel is occluded, collaterals open immediately in response to the drop in arterial pressure distal to the obstruction and remain open as long as pressure in the vascular bed distal to the obstruction remains below systemic pressure. After several hours of ischemia, however, vasoconstriction develops in the obstructed bed, elevating its pressure and reducing collateral flow. If sustained for a prolonged period, the vasoconstriction can become irrevers-

Chapter 114  Intestinal Ischemia

LMC

RMC

MA

CA MC

AOR SMA

ASC IMA

DSC S S S

SR

Figure 114-3.  Diagram of typical inferior mesenteric artery (IMA) anatomy and its anastomoses with the superior mesenteric artery (SMA). AOR, arc of Riolan; ASC, ascending branch of the left colic artery; CA, central artery; DSC, descending branch of the left colic artery; LMC, left branch of middle colic artery; MA, marginal artery; MC, middle colic artery; RMC, right branch of middle colic artery; S, sigmoid branches; SR, superior rectal artery. (From Nebesar RA, Kornblith PL, Pollard JJ, Michels NA. Celiac and superior mesenteric arteries: a correlation of angiograms and dissections. Boston: Little, Brown; 1969.)

ible and persist even after correction of the cause of the ischemic event. Such persistent vasoconstriction explains the operative findings of progressive bowel ischemia after cardiac function has been optimized and in the absence of arterial or venous obstruction. Blood flow is affected by a variety of systemic, humoral, local, and neural influences. The sympathetic nervous system, mainly via α-adrenergic receptors, is of primary importance in maintaining resting splanchnic arteriolar tone; other vasoactive substances, including angiotensin II, vasopressin, and prostaglandins, also have been implicated in the pathogenesis of ischemic injury. Ischemic damage results both from hypoxia during the period of ischemia and reperfusion injury when blood flow is re-established. More reinjury from brief ischemia appears during reperfusion, but as the ischemic period lengthens, hypoxia becomes more detrimental than reperfusion3; the injury after three hours of ischemia and one hour of reperfusion is more severe than that after four hours of ischemia. Reperfusion injury has been attributed to many factors, including reactive oxygen radicals. When molecular oxygen is reduced in univalent steps, superoxide, hydrogen peroxide, and hydroxyl radicals are formed. These oxygen radicals damage an array of molecules found in tissues, including nucleic acids, membrane lipids, enzymes, and receptors; such widespread damage can result in cell lysis, impaired

Figure 114-4.  Film from a flush aortogram of a patient with superior mesenteric artery (SMA) occlusion. The presence of a prominent meandering artery indicates that the collateral channels have been present for some time and that the occlusion is not acute. The arrows show the direction of flow from the inferior mesenteric artery to the SMA. (From Boley SJ, Brandt LJ, Veith FJ. Ischemic disorders of the intestines. Curr Probl Surg 1978; 15:29.)

cell function, and necrosis on reperfusion of ischemic tissues. A potent source of oxygen radicals in ischemic, reperfused tissue is the enzyme xanthine oxidase (XO), the ratelimiting enzyme in nucleic acid degradation. In non-ischemic tissue, this enzyme exists as a dehydrogenase (XDH) that uses nicotinamide adenine dinucleotide (NAD) rather than O2 as the electron acceptor during purine oxidation; as a result, it does not produce oxygen radicals. During is­ chemia, XDH is converted to XO with production of reactive oxygen radicals. Inhibition of XO by allopurinol dramatically attenuates the epithelial cell necrosis and the increased microvascular permeability seen during reperfusion. Neutrophils are another source of reactive oxygen me­ tabolites. During reperfusion, XO-derived oxidants initiate the production and release of leukotriene B4 and plateletactivating factor, which lead to neutrophil adherence and migration. The adherent leukocytes mediate microvascular injury by release of proteases and physical disruption of the endothelial barrier. Oxygen radical scavengers (superoxide dismutase, dimethyl sulfoxide), XO inhibitors, and agents that inhibit leukocyte adherence and migration have been shown experimentally to protect various organs against reperfusion injury, but are not yet used clinically because, in large measure, they must be given before or coincident with the ischemic injury to have protective effects.3

ACUTE MESENTERIC ISCHEMIA Intestinal ischemia can be classified as acute or chronic and of venous or arterial origin. In the acute forms, intestinal viability is threatened, whereas in the chronic forms, blood

2029

2030

Section X  Small and Large Intestine Table 114-2  Causes and Approximate Frequencies of Acute Mesenteric Ischemia CAUSE SMA embolus Nonocclusive mesenteric ischemia SMA thrombosis Mesenteric venous thrombosis Focal segmental ischemia

FREQUENCY (%) 50 25 10 10 5

SMA, superior mesenteric artery.

flow is inadequate to support the functional demands of the intestine. Acute mesenteric ischemia (AMI) is much more common than the chronic type, and arterial disease is more common than venous disease. Arterial forms of AMI include SMA embolus (SMAE), nonocclusive mesenteric ischemia (NOMI), SMA thrombosis (SMAT), and focal segmental ischemia (FSI) (Table 114-2). Acute mesenteric venous thrombosis (MVT) and FSI are the venous forms of AMI. AMI results from inadequate blood flow to all or part of the small intestine and can involve the right half of the colon because its blood supply is from the SMA. Regardless of the cause of the ischemic insult, the end results are similar: a spectrum of bowel injury that ranges from transient alteration of bowel function to transmural gangrene. Clinical manifestations vary with the extent and severity of ischemic injury and, to a lesser degree, with its cause.

INCIDENCE

AMI accounts for about 0.1% of admissions to our tertiary care center. This figure has risen since the early 1980s owing to increased recognition of the disorder, an aging population, and the widespread use of intensive care units with the salvage of patients who previously would have died from cardiovascular conditions but who now survive to develop AMI as a delayed consequence of their primary disease. Most series of AMI reported in the late 1970s and early 1980s showed that SMAE was responsible for 40% to 50%, NOMI for 20% to 30%, and SMAT for 10% to 20% of cases. The incidence of NOMI has now declined, however, likely because intensive care unit monitoring enables prompt correction of hypotension and blood volume deficits, and the widespread use of calcium channel blockers and other systemic vasodilators might protect the vascular bed from spasm. Today, SMAE is the most common cause of AMI. In one study of autopsies on patients who died from acute mesenteric thromboembolic occlusion the embolusto-thrombus ratio was 1.4 : 1.4

CLINICAL FEATURES

Early identification of AMI requires a high index of suspicion, especially in patients older than 50 years who have long-standing congestive heart failure (particularly if the heart failure is poorly controlled), cardiac arrhythmias, recent myocardial infarction, or hypotension. The development of sudden abdominal pain in a patient with any of these risk factors should suggest the diagnosis of AMI. Younger patients, however, are not without risk of AMI, especially if they are taking vasoactive medications (e.g., phenylephrine, amphetamines, triptans), are using cocaine, or have underlying thrombophilia. Hence, unexplained, persistent, and severe abdominal pain should prompt consideration of AMI as an explanation for the pain. A history of postprandial abdominal pain in the weeks to months

preceding the acute onset of severe abdominal pain is associated only with SMAT. Almost all patients with AMI have acute abdominal pain. Early in the course of disease, the pain of AMI is far more impressive than the physical findings. Initially, the pain is severe, but the abdomen usually is flat, soft, and most often not tender, or it is less tender than expected based on the magnitude of the pain. Sudden, severe abdominal pain accompanied by rapid and often forceful bowel evacuation, especially with minimal or no abdominal signs, strongly suggests SMAE. A more indolent and less striking onset is more typical of MVT, whereas with NOMI, appreciation of abdominal pain may be overshadowed by the precipitating disorders, such as hypotension, acute congestive heart failure, acute hypovolemia, or cardiac arrhythmias. Pain is absent in as many as 25% of patients with NOMI. Unexplained abdominal distention or gastrointestinal bleeding may be the only indications of AMI when pain is absent, especially when AMI is due to NOMI. Distention, although absent early in the course of AMI, is often the first sign of intestinal infarction. The stool contains occult blood in 75% of patients. Right-sided abdominal pain associated with the passage of maroon or bright red blood in the stool, although characteristic of colon ischemia, also may be seen with AMI, because the blood supply to the right colon and small intestine originates from the SMA. Elderly patients with AMI have been reported to develop mental confusion acutely in as many as 30% of cases.5 In patients who survive cardiopulmonary resuscitation and who then develop recurrent bacteremia or sepsis, the suspected cause of sepsis should be NOMI that resulted in a segment of bowel with subacute ischemic injury, acting as a portal for bacterial translocation.6 Although episodes of sepsis may be treated successfully with antibiotics, the length of damaged bowel must be removed to prevent recurrent sepsis. Although abdominal findings early in the course of intestinal ischemia may be minimal or absent, increasing tenderness, rebound tenderness, and muscle guarding reflect progressive loss of intestinal viability. Such abdominal findings strongly indicate the presence of infarcted bowel. The rate of progression from the onset of abdominal pain to intestinal infarction varies not with the specific cause of ischemia but with the severity of the ischemic insult; MVT generally has a more indolent course than do the arterial causes of AMI.

LABORATORY FEATURES AND DIAGNOSIS

On admission to the hospital, approximately 75% of patients with AMI have leukocytosis greater than 15,000 cells/mm3 and about 50% have metabolic acidemia. A normal white blood cell (WBC) count cannot be used to exclude early AMI, just as a high WBC count does not make the diagnosis. Elevated levels of serum phosphate, d-lactate, amylase, and other enzymes have been noted, as have high peritoneal fluid amylase and intestinal alkaline phosphatase activity, but the sensitivity and specificity of these markers of intestinal ischemia have not been established.7 More-specific intestinal enzymes including diamine oxidase, hexos­ aminidase, glutathione S-transferase,8 and intestinal fatty acid-binding protein9 also lack sufficient sensitivity and specificity to diagnose AMI. Moreover, serum markers, when elevated, usually indicate late-stage disease. Although they are poorly sensitive (30%) and nonspecific, plain films of the abdomen still are obtained in evaluating patients with suspected AMI. Plain films of the abdomen usually are normal in AMI before infarction. Later on, formless loops of small intestine, ileus, thumbprinting of the small bowel or right colon (Fig. 114-5), or still later,

Chapter 114  Intestinal Ischemia pneumatosis and portal or mesenteric vascular gas may be seen. In one study, the mortality rate of patients with normal plain film studies was 29%, whereas it was 78% in those with abnormal findings.10 The primary purpose of plain films (or CT scans) is to exclude causes of abdominal pain other than ischemia that might mandate a different therapeutic approach. Duplex scanning and Doppler flowmetry can be used to evaluate patients with suspected AMI, but these techniques are limited in their clinical use by the following factors: Only the proximal portions of the major splanchnic vessels can be studied reliably, not the peripheral vasculature. Vessel occlusions are not diagnostic of intestinal ischemia because complete occlusions can be seen in asymptomatic patients. Blood flow though the SMA is highly variable, which makes interpretation difficult. NOMI, which accounts for approximately 25% of AMI, cannot be diagnosed reliably by ultrasound studies.

Figure 114-5.  Plain film of the abdomen showing an ileus and a formless, fixed loop of small intestine (arrows) in a patient with acute mesenteric ischemia from a superior mesenteric artery embolus.

A

CT has largely replaced plain film study of the abdomen for diagnosis and is used to identify arterial and venous thromboses as well as ischemic bowel.11-14 Findings on CT include colon dilatation, bowel wall thickening, abnormal bowel wall enhancement, lack of enhancement of arterial vasculature with timed intravenous contrast injections, arterial occlusion, venous thrombosis, engorgement of mesenteric veins, intramural gas and mesenteric or portal venous gas (Fig. 114-6), infarction of other organs, ascites, and signs related to the cause of the infarcted bowel such as hernia.11 There are three relatively specific findings of AMI that are better depicted on CT scans compared with plain films: gas in the bowel wall or portal system, acute embolic infarction of other intra-abdominal organs, and thrombi in the mesenteric vessels.12 Unfortunately, the early signs on CT are nonspecific and the late signs reflect necrotic bowel. In a study of 26 patients with AMI who had a preoperative multislice CT scan followed by exploratory laparotomy, CT scanning identified mesenteric arterial thrombosis in 16 of 17 patients and mesenteric vein thrombosis in 7 of 7 patients, all confirmed at operation. In this study, the sensitivity and specificity of CT scanning for occlusive AMI was 92% and 100%, respectively.15 The predictive value of CT scanning in the community might not be as high as in this report, because this study used only highly trained radiologists; improved CT scanner technology, however, likely will yield higher detection rates than in the past. CT angiography has been shown to be promising in the diagnosis of AMI and, in one study, the added CT angiographic findings were believed to alter clinical management in 19% of 62 patients by making the diagnosis of AMI when CT alone did not.16 Magnetic resonance (MR) angiography and venography are newer imaging techniques used to diagnose AMI; they not only can image the vasculature but might be useful in determining metabolic consequences of inadequate blood flow.17,18 Laparoscopy may be useful, but it also can be misleading, because early in ischemic injury, blood may be shunted to the serosa, giving a normal appearance to the outside of the bowel even if the mucosa is necrotic. Moreover, laparoscopy is potentially dangerous because SMA blood flow decreases when intraperitoneal pressure exceeds 20 mm Hg.

B

Figure 114-6.  Computed tomography (CT) scans of a patient with acute mesenteric ischemia showing gas (arrow) in the portal veins (A) and gas (arrows) in the wall of the intestine as well as the mesentery and its vessels (B). Pneumatosis intestinalis is a late sign of ischemic injury, connotes bowel necrosis, and mandates explorative laparotomy.

2031

2032

Section X  Small and Large Intestine Selective mesenteric angiography, often with papaverine infusion, currently is the mainstay of diagnosis and initial treatment of both occlusive and nonocclusive forms of AMI, and it should be performed promptly if AMI is suspected or diagnosed on other imaging tests. Sensitivity and specificity of mesenteric angiography for diagnosing AMI in most studies are 90% to 100% and 100%, respectively.19 Opponents of routine angiography for patients with suspected AMI cite several problems with this approach: Difficulties in performing these studies in critically ill patients can make angiography impractical and contribute to inordinate delays in surgery. The large number of negative examinations done to identify patients with AMI early in the course of disease offsets the value of the study. A critical delay in surgical correction of vascular insufficiency is possible if angiography is not available; this is the most serious potential drawback. Proponents of angiography accept that the large number of negative angiographic studies is necessary if diagnoses are to be made early enough to improve survival. Prompt laparotomy, however, is indicated in patients with suspected AMI if angiography cannot be performed expeditiously. More controversial is the need for angiography in a patient with suspected AMI and signs of peritonitis. Because such signs usually connote infarcted bowel, the most compelling reason for angiography, namely, diagnosis of AMI while the effects of intestinal ischemia are still reversible, is no longer relevant. Angiography nonetheless still plays an important role in this situation because it can diagnose AMI and its cause and provide a roadmap for revascularization and access for serial postoperative angiographic studies.

TREATMENT

Our approach to the management of AMI is based on several observations. First, if the diagnosis is not made before intestinal infarction, the mortality rate is 70% to 90%. Second, diagnosis of both the occlusive and nonocclusive forms of AMI can be made in most patients by angiography. Third, vasoconstriction, which can persist even after the cause of the ischemia is corrected, is the basis of NOMI and a contributing factor in the other forms of AMI. Finally, vasoconstriction can be relieved by vasodilators infused into the SMA. The cornerstones of our approach, therefore, are the earlier and more liberal use of angiography and the incorporation of intra-arterial papaverine in the treatment of both occlusive AMI and NOMI. Duration of symptoms parallels mortality, and therefore early diagnosis and treatment is paramount to increase the chance for survival.20 AMI should be suspected in patients older than 50 years who have the risk factors previously described and in younger patients—especially those with atrial fibrillation, vasculitis, a coagulation disorder, and those on vasoactive medications—who seek medical attention for sudden, severe abdominal pain that lasts longer than several hours. These patients should be managed according to the algorithm shown in Figure 114-7. Less-absolute indications for inclusion into this protocol consist of unexplained acute abdominal distention, colonoscopic evidence of isolated right-sided colonic ischemia, and acidosis without an identifiable cause. Initial management of patients with suspected AMI includes resuscitation and diagnostic imaging studies. Resuscitation includes relief of acute congestive heart failure and correction of hypotension, hypovolemia, and cardiac arrhythmias. Broad-spectrum antibiotics (e.g., levofloxacin, metronidazole, piperacillin-tazobactam) are given immediately because of the high incidence of positive blood

cultures in AMI and because they reduce the extent and severity of ischemic injury in experimental animals.21 There are no randomized, controlled trials showing the benefit of antibiotics in AMI, and it is unlikely that such trials will ever be done. After resuscitation, plain films or CT scan of the abdomen are obtained, not to establish the diagnosis of AMI but rather to exclude other causes of abdominal pain. A normal plain film or CT scan does not exclude AMI; ideally, patients are studied before radiologic signs appear because these signs connote irreversibly damaged bowel. If no alternative diagnosis is made on these studies, selective SMA angiography is performed. Based on the angiographic findings and the presence or absence of peritoneal signs, the patient is treated according to the algorithm in Figure 114-7. Even when the decision to operate has been based on clinical grounds, preoperative angiography should be performed to manage the patient properly at and after laparotomy. Relief of mesenteric vasoconstriction is essential to the treatment of emboli, thromboses, and the nonocclusive low-flow states. Infusion of the phosphodiesterase inhibitor papaverine, through the angiography catheter in the SMA, is used to relieve mesenteric vasoconstriction preoperatively and postoperatively. The papaverine is infused by pump at a constant rate of 30 to 60 mg/hr; papaverine concentrations may vary with the need for fluid restriction. Although most of the papaverine infused into the mesenteric bed is cleared during one pass through the liver, blood pressure and cardiac rate and rhythm must be monitored constantly. Some patients with liver disease exhibit a drop in blood pressure with this dose of papaverine, but the most common cause of hypotension during the papaverine infusion is dislodgment of the catheter. In patients who have a sudden drop in blood pressure, the papaverine should be replaced with saline or glucose solution and a plain film image of the abdomen should be promptly performed to confirm the catheter’s location. If the catheter has come out of the SMA, it should be replaced and the papaverine should be restarted. The patient’s clinical and angiographic responses to the vasodilator determine the duration of therapy. Laparotomy is performed in AMI to restore arterial flow obstructed by embolus or thrombosis or to resect irreparably damaged bowel, or both. Embolectomy, thrombectomy, or arterial bypass precedes evaluation of intestinal viability because bowel that initially appears infarcted can show surprising recovery after adequate blood flow is restored. In the operating room, intestinal viability can be assessed clinically, by qualitative or quantitative surface fluorescence or by Doppler ultrasonography.22 Animal models show that administration of intravenous glucagon, intravenous heparin-binding epidermal growth factor (HB-EGF)-like growth factor or intraluminal nitroglycerin after revascularization of an acute arterial occlusion can improve mucosal viability and minimize reperfusion damage.23,24 In practice, glucagon is sometimes used because of its accessibility and anecdotal beneficial effects despite lack of strong supporting data from human trials. Short segments of bowel that are nonviable or questionably viable after revascularization are resected, and a primary anastomosis is performed. If extensive portions of the bowel are of questionable viability, however, only the clearly necrotic bowel is resected and re-exploration (second look) is planned for within 12 to 24 hours. The interval between the first and second operations is used both to allow better demarcation between viable and nonviable bowel and to attempt to improve intestinal blood flow by using intra-arterial papaverine and by maximizing cardiac output.

Long-term parenteral nutrition

Warfarin

Continuous papaverine infusion, if possible

Observe

Thrombectomy, heparin, papaverine

Thrombolytic agent

Second-look operation ± resection

Heparin, papaverine

Main vein occluded

Absent or poor filling of SMA

Good filling of SMA

Second-look operation if appropriate

Laparotomy or laparoscopy, arterial reconstruction, and/or bowel resection

No angiographic evidence of collaterals

Persistent signs of peritonitis

Angiographic evidence of collaterals

No persistent signs of peritonitis

Viable

Main vein open or reconstituted

Close

No

Major arterial occlusion (non-embolic)

History of DVT or hypercoagulable state

Stop the infusion and remove the catheter

Laparotomy or laparoscopy and local resection

Continuous papaverine infusion

Persistent signs of peritonitis

Thrombolytic agent or heparin

Observe

Angiogram normal

Repeat angiogram

Observe

Continuous papaverine infusion

No persistent signs of peritonitis No persistent signs of peritonitis

Thrombolytic agent

Repeat angiogram

Continuous papaverine infusion

Repeat angiogram and consider a second-look operation

Continuous papaverine infusion postoperatively

Embolectomy and/or resection

Repeat angiogram

Repeat angiogram and consider a second-look operation

Continuous papaverine infusion postoperatively

Laparotomy or laparoscopy ± resection

Continuous papaverine infusion

Continuous papaverine infusion Observe

Persistent signs of peritonitis

No persistent signs of peritonitis

Splanchnic vasoconstriction (no occlusion)

Laparotomy or laparoscopy

Observe

Continuous papaverine infusion preoperatively

Persistent signs of peritonitis

Major embolus

Persistent signs of peritonitis

No persistent signs of peritonitis

Selected cases: contraindication to surgery good perfusion of the distal mesenteric vascular bed after bolus of a vasodilator

Abnormal

Normal

Minor arterial occlusion or embolus

Abdominal angiogram

Abnormal

CTA

CTA with venous protocol

Figure 114-7.  Algorithm for the diagnosis and treatment of intestinal ischemia. CTA, computed tomographic angiogram; DVT, deep venous thrombosis; SMA, superior mesenteric artery. Solid lines show conventional management plan; dotted lines show alternative management plan. (Modified from Brandt LJ, Boley SJ. AGA technical review on intestinal ischemia: American Gastrointestinal Association. Gastroenterology 2000; 118:954; corrected version in Gastroenterology 2000; 119:281.)

Resect, if possible

Nonviable

Heparin

Resect

Extensive ischemic involvement

Laparotomy or laparoscopy

Heparin ± thrombolytic agents

Short ischemic segment

Persistent signs of peritonitis

No persistent signs of peritonitis

Mesenteric venous thrombosis

Resuscitate the patient and correct any predisposing or precipitating factors

Yes

Chapter 114  Intestinal Ischemia 2033

2034

Section X  Small and Large Intestine formed in patients with major emboli after papaverine infusion is begun. Nonoperative therapy using only papaverine infusion is attempted if there are significant contraindications to surgery, no peritoneal signs, and adequate perfusion of the vascular bed distal to the embolus after a bolus of vasodilator into the SMA. Exploratory laparotomy is mandatory when peritonitis is present; embolectomy and bowel resection are performed as necessary. If possible, intra-arterial papaverine is begun before surgery and is continued during surgery. If no “second-look” operation is planned, infusion is continued for 12 to 24 hours postoperatively; persistent vasospasm is excluded by angiography before the catheter is removed (see Fig. 114-8). If a second operation is planned, the infusion is continued through the second procedure until angiography shows the vasoconstriction is ceased. Recognition of persistent vasoconstriction has prompted some authorities to recommend routine use of intra-arterial papaverine in all patients with SMAE; the best survival rates are seen in patients treated by this approach.19 Use of transcatheter thrombolytic therapy (e.g., alteplase or urokinase) can be considered before exploratory laparotomy if the patient does not have signs of peritonitis. Prospective studies and meta-analyses have shown that thrombolysis may be effective in resolving thrombi, improving symptoms, and avoiding surgery in patients with lesions amenable to such therapy.25,26 Thrombolytic therapy is most likely to be successful when the embolus is partially occluding or is minor and when the study is performed within 12 hours of the onset of symptoms.27 A canine study showed that intra-arterial streptokinase was more effective than intra-arterial papaverine in lysing clots implanted into the SMA, although greater ischemic damage occurred with

The use of anticoagulants in the management of AMI is controversial. Anticoagulation with heparin can cause intestinal or intraperitoneal hemorrhage and, except for MVT, should not be used routinely in the immediate postoperative period; 48 hours after embolectomy or arterial reconstruction, when thrombosis is common, anticoagulation is appropriate.

Specific Types of Acute Mesenteric Ischemia

Superior Mesenteric Artery Embolus SMAE is responsible for 40% to 50% of AMI episodes. Emboli usually originate from a left atrial or ventricular mural thrombus. Many patients with SMAE have had previous peripheral artery emboli, and approximately 20% have synchronous emboli. SMAEs lodge at points of normal anatomic narrowing, usually immediately distal to the origin of a major branch. Angiography typically reveals a rounded filling defect with nearly complete obstruction to flow. Mesenteric atherosclerosis is usually not as severe as in SMAT. Emboli proximal to the origin of the ileocolic artery are considered major emboli. Minor emboli are those that lodge in the SMA distal to the takeoff of the ileocolic artery or in the distal branches of the SMA (Fig. 114-8). Various therapeutic approaches have been proposed for SMAE, depending on the presence or absence of peritoneal signs, whether the embolus is partially or completely occluding, and whether the embolus is above the origin of the ileocolic artery or more distal. Therapy for SMAE has included surgical revascularization, intra-arterial perfusion with vasodilators or thrombolytic agents, and anticoagulation.19 In the absence of peritoneal signs, minor SMA emboli have been treated successfully with all of these agents without the need for surgery. Exploration is usually per-

A

B

Figure 114-8.  A, Superior mesenteric artery (SMA) angiogram in a 71-year-old man with abdominal pain showing an embolus occluding the SMA at the level of the origin of the right colic artery (arrow). Vasoconstriction is noted distal to the embolus. B, Repeat angiogram done 54 hours after SMA embolectomy and preoperative and postoperative papaverine infusions into the SMA. Vasodilatation is seen, and all vessels are patent except for a distal jejunal branch, which contains a piece of the inciting embolus (arrowhead) that broke off during the course of vasodilator therapy because of endogenous thrombolysis. Papaverine protected the bowel within the distribution of the embolized vessel by enabling vasodilatation and maintenance of adequate blood flow.

Chapter 114  Intestinal Ischemia streptokinase than with papaverine because of papaverine’s action to cause vasodilation and open collateral pathways for blood flow around the obstructing clot. When streptokinase and papaverine were administered simultaneously, neither medication functioned as well as it did alone and intestinal damage was intensified.28 Given the shortage of supporting evidence for thrombolytics in AMI and the high complication rate attending their use, this treatment remains controversial.27

When papaverine infusion is used as the only treatment for NOMI in patients without signs of peritonitis, it is continued for 24 hours, and repeat angiography is performed 30 minutes after changing the papaverine infusion to normal saline. Papaverine infusion is maintained and angiography repeated daily until there is no roentgenographic evidence of vasoconstriction and the patient’s clinical findings resolve. Infusions, usually discontinued after 24 hours, have been given for as long as 5 days.

Nonocclusive Mesenteric Ischemia NOMI is responsible for 20% to 30% of AMI and usually is due to splanchnic vasoconstriction consequent to a preceding cardiovascular event. AMI can appear hours to days after the event, and vasoconstriction, which initially is reversible, can persist even after the precipitating event has been corrected. Precipitating causes for NOMI include acute myocardial infarction, congestive heart failure, arrhythmias, shock, cirrhosis, medications (e.g., digitalis), cardiopulmonary bypass surgery, and chronic kidney disease, especially when patients are on either hemodialysis or peritoneal dialysis. When presenting with abdominal pain, patients on peritoneal dialysis may be thought to have peritonitis, thereby delaying the diagnosis of NOMI and resulting in a poor outcome.29 NOMI is diagnosed by angiography using four criteria: narrowing of the origins of SMA branches, irregularities in the intestinal branches, spasm of the arcades, and impaired filling of intramural vessels. Patients with these signs who are neither in shock nor on vasopressors and who do not have pancreatitis can be considered to have NOMI (Fig. 114-9). SMA infusion of papaverine is begun as soon as the diagnosis is made. Operation is performed if peritoneal signs are present, and the infusion is continued during and after exploration. Necrotic bowel is resected; it is better to leave bowel of questionable viability and perform a second-look operation than to perform massive enterectomy, because compromised but viable bowel often improves with supportive measures. The infusion is continued as for secondlook operations following embolectomy.

Acute Thrombosis of the Superior Mesenteric Artery Acute SMAT occurs in areas of severe atherosclerotic narrowing, most often at the origin of the SMA. The acute ischemic episode may be superimposed on chronic mesenteric ischemia (CMI), and 20% to 50% of patients have a history of postprandial abdominal pain and weight loss during the weeks to months preceding the acute event. Evidence of coronary, cerebrovascular, or peripheral arterial insufficiency is common. SMAT is demonstrated on flush aortography, which usually shows occlusion of the SMA 1 to 2  cm from its origin. Some distal filling of the SMA via collaterals is common. Branches proximal and distal to the obstruction can show localized or diffuse vasoconstriction. In patients with abdominal pain, no abdominal tenderness, and complete occlusion of the SMA on aortography, it is important, though difficult, to distinguish between acute thrombosis and long-standing, coincidental chronic occlusion. Prominent collaterals between the SMA and other major splanchnic vessels indicate chronic SMA occlusion. If there is good filling of the SMA, the occlusion is considered chronic and the abdominal pain is considered unrelated to mesenteric vascular disease (see Fig. 114-4). The absence of collateral vessels or the presence of collaterals with inadequate filling of the SMA indicates an acute occlusion and demands prompt intervention. If possible, an angiographic catheter is placed in the proximal SMA, and papaverine infusion is begun before surgery is undertaken. At surgery, necrotic bowel is resected and remaining bowel is revascularized. Papaverine infusion is continued throughout the operative period, and management is the same as for SMAE. There are only a few reports of use of thrombolytic agents or percutaneous angioplasty for SMAT.

Complications

Complications of angiography and prolonged infusion of vasodilator drugs include transient acute tubular necrosis following angiography, local hematomas at the arterial puncture sites, catheter dislodgment, and fibrin clots on the arterial catheter. Infusion for more than 5 days has not had significant systemic effects.

Results

A

B

Figure 114-9.  Superior mesenteric angiogram in a patient with nonocclusive mesenteric ischemia (NOMI) following a bout of gastrointestinal hemorrhage and shock. A, The pretreatment film shows the diffuse vasoconstriction of NOMI. B, Marked vasodilatation is evident on the repeat study after 48 hours of an intra-arterial papaverine infusion. (From Brandt LJ, Boley SJ. Ischemic intestinal syndromes. Adv Surg 1981; 15:1.)

Although mortality rates of 70% to 90% were reported through the 1980s for patients whose AMI was diagnosed and treated conventionally, the approach described here can reduce these catastrophic figures. The best survival is reported in series in which angiography has been used routinely in the management of AMI.30-35 In our tertiary medical center, more than 50% of the patients with AMI treated according to our approach survived, and more than 75% have lost less than one meter of intestine. The importance of early diagnosis is emphasized by the survival of 90% of patients who had AMI but no signs of peritonitis, and who had angiography early in their course. Ideally, all patients with AMI should be studied when plain films of the abdomen and CT scanning are normal and before signs of an acute surgical abdomen and

2035

2036

Section X  Small and Large Intestine laboratory evidence of infarction appear. Diagnosis before intestinal infarction occurs is the most important factor in improving survival of patients with AMI.

MESENTERIC VENOUS THROMBOSIS MVT occurs as an acute, subacute, or chronic disorder. It is only since the development of recent imaging techniques that these various forms of MVT have been recognized; previously, only acute MVT was known, and diagnosis was made at laparotomy or autopsy.

INCIDENCE

In early studies, MVT was believed to be the major cause of AMI, but most of these cases probably represented NOMI. Today, only 5% to 10% of patients with AMI have MVT. The mean age at presentation with MVT is in the mid-60s.36 A Swedish study showed that the highest incidence of MVT was 11.3 per 100,000 person years among those 70 to 79 years old.37

PREDISPOSING CONDITIONS

Previously, a cause of MVT was identified in fewer than half of patients. The discoveries of the primary and secondary hypercoagulable states and the use of estrogens for contraception and hormone replacement have led to identification of the cause in almost 90% of patients.38 Arterial hypertension is the most commonly associated medical comorbidity with this disorder, and neoplasms (e.g., acute lymphocytic leukemia, adenocarcinoma of the pancreas, stomach) and coagulation disorders (e.g., lupus anticoagulant, factor V Leiden, and protein S deficiency) also are commonly seen.36,39 A list of predisposing conditions for MVT is given in Table 114-3.

Table 114-3  Conditions Associated with Mesenteric Venous Thrombosis Hypercoagulable states Activated protein C resistance Antithrombin deficiency Protein C deficiency Protein S deficiency Methyltetrahydrofolate deficiency Estrogen use (oral contraceptives, hormone replacement therapy) Polycythemia vera Essential thrombocytosis Neoplasms Inflammation Appendicitis Diverticulitis Pancreatitis Peritonitis Inflammatory bowel disease Pelvic or intra-abdominal abscess Portal hypertension Cirrhosis Congestive splenomegaly After sclerotherapy of esophageal varices Blunt abdominal trauma Splenectomy and other postoperative states Peripheral deep venous thrombosis Pregnancy Decompression sickness

PATHOPHYSIOLOGY

The location of the initial thrombus within the mesenteric venous circulation varies with the cause. MVT secondary to cirrhosis, neoplasm, or operative injury begins at the site of obstruction and extends peripherally, whereas thrombosis of primary hypercoagulable states starts in smaller branches and propagates into the major trunks. Intestinal infarction is rare unless the branches of the peripheral arcades and the vasa recta are involved. When collateral circulation is inadequate and venous drainage from a segment of bowel is compromised, the affected intestine becomes congested, edematous, cyanotic, and thickened with intramural hemorrhage. Serosanguineous peritoneal fluid heralds hemorrhagic infarction. Arterial vasoconstriction can be marked, but pulsations persist up to the bowel wall. Transmural infarction can make it impossible to differentiate venous from arterial occlusion.

CLINICAL FEATURES

MVT can have an acute, subacute (weeks to months), or chronic onset. Except for late complications, chronic MVT is asymptomatic. As many as 60% of patients have a history of peripheral vein thromboses.40 Acute MVT manifests with abdominal pain in more than 90% of patients and, as with acute arterial ischemia, the pain initially is out of proportion to the physical findings. The mean duration of pain before admission is five to 14 days, but it is more than one month in as many as 25% of patients.41 Other symptoms, including nausea and vomiting, occur in more than 50%. Lower gastrointestinal bleeding, bloody diarrhea, or hematemesis occur in 15% and indicate bowel infarction. Fecal occult blood is found in more than half of instances during the course of MVT. Initial physical findings vary at different stages and with different degrees of ischemic injury, but guarding and rebound tenderness develop as bowel infarction evolves. Most patients have a temperature higher than 38°C (100.4°F), and 25% exhibit signs of septic shock. Subacute MVT describes the condition in patients who have abdominal pain for weeks to months but no intestinal infarction. Subacute MVT can be due either to extension of thrombosis at a rate rapid enough to cause pain but that permits collaterals to develop, thus preventing infarction, or to acute thrombosis of venous drainage sufficient to permit recovery from ischemic injury. The diagnosis usually is made on imaging studies ordered to evaluate the cause of undiagnosed pain. Nonspecific abdominal pain usually is the only symptom of subacute MVT, and physical examination and laboratory tests are normal. Some patients who present with subacute MVT ultimately develop intestinal infarction; this blurs the distinction between the acute and subacute forms of MVT. At autopsy, coexistent new and old thromboses have been found in nearly half of the patients. Chronic MVT is seen in patients who are asymptomatic at the time of thrombosis but who may develop gastrointestinal bleeding from varices.42 Most patients who bleed do so from gastroesophageal varices secondary to portal or splenic vein thrombosis, and they have physical findings of portal hypertension. Laboratory studies may show secondary hypersplenism with pancytopenia or thrombocytopenia.

DIAGNOSIS Acute Mesenteric Venous Thrombosis

The absence of specific symptoms, signs, or laboratory results and the typical variability in the course of the disease make it difficult to diagnose acute MVT preoperatively. Abdominal plain film signs of MVT are similar to those of other forms of AMI and almost always reflect the presence

Chapter 114  Intestinal Ischemia of infarcted bowel. Barium enemas are of little diagnostic value, because MVT unusually involves the colon. Characteristic findings on small bowel series include marked thickening of the bowel wall due to congestion and edema, with separation of loops and thumbprinting. Selective mesenteric arteriography can establish a definitive diagnosis before bowel infarction, can differentiate venous thrombosis from arterial forms of ischemia, and can provide access for vasodilator therapy. Angiographic findings of acute MVT include thrombus in the SMV with partial or complete occlusion, failure to visualize the SMV or portal vein, slow or absent filling of the mesenteric veins, arterial spasm, failure of the arterial arcades to empty, reflux of contrast medium into the artery, and prolonged blush in the involved segment.32 Ultrasonography, CT, and MR imaging all have been used to demonstrate thrombi in the SMV and the portal vein.43,44 CT can diagnose MVT in more than 90% of patients and is the diagnostic study of choice. Specific findings include thickening and enhancement of the bowel wall, enlargement of the SMV, a central lucency in the lumen of the vein (representing a thrombus), a sharply defined vein wall with a rim of increased density, and dilated collateral vessels in a thickened mesentery (Fig. 114-10). CT also is very accurate in differentiating transmural infarction from nontransmural ischemia.45 When MVT is diagnosed on CT scanning, angiography might not be necessary, but in selected symptomatic patients it better delineates thrombosed veins and provides access for intra-arterial vasodilators. Magnetic resonance venography also appears to be highly sensitive for diagnosing MVT and is a diagnostic alternative to CT.46 Esophagogastroduodenoscopy and colonoscopy rarely are helpful, because the duodenum and colon are infrequently involved. As in other forms of AMI, laparoscopy may be useful either when other studies are contraindicated or in concert with imaging tests.47 The diagnosis of MVT usually has been made at laparotomy, where its hallmarks are serosanguineous peritoneal fluid, dark red to blue-black edematous bowel, thickening of the mesentery, good arterial pulsations in the involved segment, and thrombi in cut mesenteric veins. At this stage, some degree of intestinal infarction invariably has occurred. When persons with suspected AMI exhibit clinical features suggesting MVT, contrast-enhanced CT is performed before

A

SMA angiography, although this recommendation is becoming moot as CT angiography is increasingly used as the initial imaging test.

Chronic Mesenteric Venous Thrombosis

Because chronic MVT is asymptomatic or manifests as gastrointestinal bleeding, the diagnostic evaluation is directed toward determining the cause of the bleeding. Endoscopy and appropriate imaging studies should identify the cause and site of bleeding and the extent of thrombosis.

TREATMENT Acute Mesenteric Venous Thrombosis

Most patients with acute MVT initially are believed to have some form of AMI and are treated as discussed in earlier sections and as outlined in the algorithm of Figure 114-7. In asymptomatic persons in whom the diagnosis is made on a CT scan done for other than abdominal pain, either three to six months of anticoagulation or no therapy, in some cases, is reasonable. In symptomatic patients, treatment is determined by the presence or absence of peritoneal signs; signs of peritonitis mandate laparotomy and resection of infarcted bowel. If long segments of questionably viable bowel are found, papaverine is infused, and if arterial spasm is relieved and the SMV or portal vein is visualized, thrombectomy or a second look may be attempted to determine whether resection should be performed. Following surgery, heparin should be administered. Immediate heparinization for seven to 10 days has been shown to diminish recurrence and progression of thrombosis and to improve survival.48,49 In the absence of peritoneal signs, immediate heparinization followed by a three- to six-month course of warfarin may be all that is needed. A comparison of patients who were treated surgically with those who were managed medically suggested that nonoperative management is a reasonable option provided the diagnosis on CT scan is certain and there is no transmural necrosis or perforation.50 A few case reports have documented the use of thrombolytic agents in the treatment of acute MVT. Current recommendations for the duration of anticoagulation are not supported by evidence-based data, but rather are based on conventional practice. If an underlying hypercoagulable state is found, lifelong treatment with warfarin is advised. If no underlying thrombophilic state is docu-

B

Figure 114-10.  Superior mesenteric venous thrombosis. A, Abdominal computed tomography (CT) with intravenous contrast demonstrating an enlarged superior mesenteric vein with a central lucency in the lumen, representing the thrombus. The vein wall is sharply defined, with a rim of increased density surrounding the thrombus (arrows). B, Abdominal CT with intravenous contrast agent showing thickening and persistent enhancement of the bowel wall (black arrows) and dilated collateral vessels within a thickened mesentery (white arrows). (From Boley SJ, Brandt LJ. Ischemic disorders of the intestines. Surg Clin North Am 1992; 72:194.)

2037

2038

Section X  Small and Large Intestine mented, a three- to six-month course of therapy is thought to be sufficient.

Chronic Mesenteric Venous Thrombosis

Treatment of chronic MVT is aimed at controlling bleeding, usually from esophageal varices. Sclerotherapy, variceal banding, portosystemic shunts, transjugular intrahepatic portosystemic procedures, devascularization procedures, and bowel resection all have a place in treating selected patients. Use of beta blockers and anticoagulation was found to be associated with improved survival in these patients.51 No treatment is indicated for patients with asymptomatic chronic MVT.

Table 114-4  Types and Approximate Frequencies of Colon Ischemia in Patients Seen at a Tertiary Referral Hospital TYPE Reversible colopathy and transient colitis Transient colitis Chronic ulcerating colitis Stricture Gangrene Fulminant universal colitis

FREQUENCY (%)* >50 10 20 10 15 <5

*Because of the approximate nature of the frequencies, the total of the frequencies of all types of colon ischemia is not 100%.

PROGNOSIS

Mortality associated with acute MVT is lower than that for other forms of AMI and is approximately 20%. Intestinal infarction, not having a CT scan performed, and treatment on a nonsurgical ward all are associated with increased mortality.37 Recurrence rates of 20% to 25% fall to about 15% if heparin therapy is begun promptly. The natural history of chronic MVT is not known, but from postmortem studies it appears that almost 50% of patients with MVT have no bowel infarction and many have no symptoms.

FOCAL SEGMENTAL ISCHEMIA OF THE SMALL INTESTINE Vascular insults to short segments of small intestine produce a broad spectrum of clinical features without the lifethreatening complications associated with more extensive ischemia. The causes of FSI include atheromatous emboli, strangulated hernias, immune complex disorders and vasculitis, blunt abdominal trauma, segmental venous thrombosis, radiation therapy, and oral contraceptives, among others. With FSI there is usually adequate collateral circulation to prevent transmural infarction; the most common lesion is partial bowel wall necrosis with invasion by intestinal bacteria. FSI can manifest as acute enteritis, chronic enteritis, or a stricture. In the acute pattern, abdominal pain often simulates acute appendicitis. Physical findings are those of an acute abdomen, and an inflammatory mass may be palpated. The chronic enteritis pattern can resemble Crohn’s disease, with cramping abdominal pain, diarrhea, fever, and weight loss. Roentgenographic findings also can resemble those of Crohn’s disease except that FSI occurs anywhere in the small bowel, whereas Crohn’s disease mainly affects the terminal ileum. The most common presentation is chronic small bowel obstruction from a stricture with intermittent abdominal pain, distention, and vomiting. Bacterial overgrowth in the dilated loop proximal to the obstruction can produce a blind loop syndrome. Radiologic studies typically reveal a smooth tapered stricture of variable length with an abrupt change to normal bowel distally and dilated bowel proximally. Treatment of FSI is resection of the involved bowel.

COLON ISCHEMIA Colon ischemia (CI) is a common disorder of the large bowel in older persons and is the most common form of intestinal

ischemic injury. It comprises a spectrum (Table 114-4) that includes reversible colopathy (submucosal or intramural hemorrhage), transient colitis, chronic colitis, stricture, gangrene, and fulminant universal colitis. The initial presentation usually is the same among these types and does not necessarily predict the course of disease, with the exception of ischemia involving the ascending colon. This latter pattern can simultaneously involve the small intestine, usually is caused by SMAE or NOMI, can have associated shock, and carries a mortality rate of more than 50%.52-54

INCIDENCE

The incidence of CI is underestimated, because many patients suffer only mild or transient damage and do not seek medical attention. Also, CI is commonly misdiagnosed and confused with other disorders, notably inflammatory bowel disease. In our tertiary care hospital, CI accounts for approximately 1 in 2000 hospital admissions and is seen in approximately 1 in 100 flexible sigmoidoscopies and colonoscopies. A study using medical claims data from a large health care organization calculated a crude incidence rate of 7.2 cases per 100,000 person-years of observation in the general population, in contrast to 42.8 cases per 100,000 person-years for patients with irritable bowel syndrome (IBS). After adjustment for age, sex, and calendar year, the incidence of CI in people with IBS was 3.4 times higher than it was in persons without IBS.55 CI has female gender predilection, and more than 90% of patients with CI of non-iatrogenic causes are older than 60 years of age. CI affecting young persons has been documented in case reports or series of a few patients and usually has been due to vasculitis, coagulation disorders, illicit use of cocaine, and a variety of iatrogenic causes, including a wide variety of medications such as estrogens, serotoninergic agonists and antagonists, sumatriptan, and methamphetamine.

PATHOPHYSIOLOGY AND CAUSES

CI can result from alterations in the systemic circulation or from anatomic or functional changes in the mesenteric vasculature, and it is thought to result from local hypoperfusion and reperfusion injury. In most cases, no specific cause for the ischemia is identified, and such episodes are viewed as localized nonocclusive ischemia, likely a result of smallvessel disease. An increasing variety of causes of CI is being defined, including hematologic disorders, thrombophilic states, and medications (see later) (Table 114-5). Abnormalities on angiography rarely correlate with clinical manifestations of disease, but age-related abnormalities

Chapter 114  Intestinal Ischemia Table 114-5  Causes of Colon Ischemia Acute pancreatitis Allergy Amyloidosis Heart failure or cardiac arrhythmias Hematologic disorders and coagulopathies Activated protein C resistance Antithrombin deficiency Paroxysmal nocturnal hemoglobinuria Polycythemia vera Protein C and S deficiencies Prothrombin G20210A mutation Sickle cell disease Infection Bacteria (Escherichia coli O157:H7) Parasites (Angiostrongylus costaricensis) Viruses (hepatitis B and C viruses, cytomegalovirus) Inferior mesenteric artery thrombosis Long-distance running Medications and toxins Alosetron Cocaine Danazol Digitalis compounds Ergots Estrogens Flutamide Glycerin enema Gold salts Immunosuppressive agents Interferon-α Methamphetamine Nonsteroidal anti-inflammatory drugs Penicillin Phenylephrine Polyethylene glycol 3350 colon lavage solutions Pit viper toxin Progestins Pseudoephedrine Psychotropic drugs Saline laxatives Sumatriptan Tegaserod Vasopressin Pheochromocytoma Ruptured ectopic pregnancy Shock Strangulated hernia Surgery/Procedures Aortic aneurysmectomy Aortoiliac reconstruction Barium enema Colectomy with inferior mesenteric artery ligation Colon bypass Colonoscopy Exchange transfusions Gynecologic operations Lumbar aortography Thromboembolism Cholesterol (atheroembolism) Myxoma (left atrial) Trauma (blunt or penetrating) Vasculitis and vasculopathy Buerger’s disease Fibromuscular dysplasia Kawasaki’s disease Polyarteritis nodosa Rheumatoid vasculitis Systemic lupus erythematosus Takayasu’s arteritis Volvulus

in the splanchnic vessels, including narrowing of small vessels, tortuosity of the long colic arteries, and fibromuscular dysplasia (FMD) of the superior rectal artery, can contribute to CI. The colon is particularly susceptible to ischemia, perhaps owing to its relatively low blood flow, its unique decrease in blood flow during periods of functional activity, and its sensitivity to autonomic stimulation. What triggers the episode of CI, however, usually is not known.

Medications as a Cause of Colon Ischemia

Medications should always be considered as a possible etiology for CI.56 Antibiotics Antibiotic-associated hemorrhagic colitis (AAHC) is believed to be mediated by CI. The penicillins and their derivatives, including amoxicillin and ampicillin, most commonly have been associated, although macrolides, cephalosporins, chloramphenicol, fluoroquinolones, and tetracyclines also are known precipitants. AAHC typically manifests two to seven days after antibiotics are initiated, beginning with lower abdominal pain and loose stools and followed by hematochezia several hours later.57 Some believe the mechanism of AAHC is a hypersensitivity or allergic reaction, and others believe it results from a cross reaction between the infecting agent and the antibiotic. Regardless of mechanism, if AAHC is suspected, it is recommended that the antibiotic be stopped and an alternative regimen started. Chemotherapeutic Agents When associated with chemotherapy, CI usually is asso­ ciated with the alkaloid and taxane classes of chemotherapeutic agents, such as vinorelibine tartrate (alkaloid) and paclitaxel and docetaxel (taxanes).58 Although it is not proved, the mechanism of injury of CI with these compounds is believed to be either a direct toxic injury of the colonic epithelium or anti-angiogenic toxicity.59 If patients develop CI while taking these medications, the clinician must reconsider the risks and benefits of their use. Constipation-Inducing Agents Patients taking medications that have constipation as a known adverse effect are potentially at increased risk for CI.60 More than 250 different medications from a variety of medication classes fit this category.56 One potential mechanism of CI caused by constipation-inducing agents is based upon the observation that patients with idiopathic colonic slow transit have reduced baseline rectal and mucosal blood flow, possibly as a result of impaired efferent vagal cholinergic activity. Impaired cholinergic innervation is a side effect of many constipation-inducing medications, and the resultant unopposed sympathetic input leaves the colon susceptible to ischemic injury. Decongestants Pseudoephedrine is used for relief of nasal congestion because direct application to α1-adrenergic receptors constricts vessels and relieves symptoms. The drug can be absorbed, however, and cause mesenteric vasoconstriction with resultant CI. Women are affected more often than men, and the splenic flexure watershed region seems to be most susceptible to injury by this agent. Patients with a history of CI or vasculitis or with a known thrombophilic state should be alerted to this potential complication.56

2039

2040

Section X  Small and Large Intestine Diuretics

Diuretics, such as furosemide, have been implicated in NOMI and CI. The presumed mechanism for CI with these agents is a decrease in extracellular fluid volume and reduction in peripheral resistance, prompting a steal of blood from the bowel to the limbs.61 Improved fluid balance should treat this condition. Hormonal Therapies Oral contraceptive pills (OCPs) usually are combinations of low-dose ethinyl estradiol and progestin, whereas hormone replacement therapy (HRT) often consists of ethinyl estradiol, with progestin added if the patient has an intact uterus. Young women taking OCPs are at a six-fold greater risk of CI compared with an age-matched cohort not taking these agents.62 OCPs and HRT are known to result in a hypercoagulable state with ensuing microthromboembolic events and resultant CI. In any patient with other risk factors for CI, the use of OCPs or HRT, if clinically appropriate, should be monitored closely. Patients experiencing CI while taking these medications should stop taking them and consider other options. Controlled or Illicit Pharmacologic Agents Amphetamines are sympathomimetic vasoconstricting medications used for medicinal and recreational purposes. Significant increases in morbidity from a variety of ischemic insults attributed to their use have been reported, including myocardial ischemia and intestinal gangrene.63 Amphetamine-induced CI is rare and usually manifests with hematochezia and mild to moderate abdominal pain. These medications tend to affect the ascending colon, possibly as a result of selective vasoconstriction of the SMA.56 Cocaine is known to induce CI.64 The main mechanisms of injury with this agent include mesenteric vasoconstriction, a hypercoagulable state, and direct toxicity to the vasculature.65 Patients present 12 to 24 hours after intranasal or intravenous cocaine use with rectal bleeding with or without abdominal pain. Involvement of the descending colon, sigmoid, and occasionally the rectum are typical of cocaine-induced CI. Laxatives Sodium polystyrene sulfonate or sorbitol (SPS, Kayexelate) commonly is used to treat hyperkalemia. Case series have shown that patients who have uremia and are given SPS are at increased risk for intestinal and colonic necrosis.66 The exact mechanism of disease is not clear, although some believe that sorbitol-induced osmotic mucosal injury combined with elevated renin levels predisposes patients to NOMI via angiotensin-mediated vasoconstriction.67 When using SPS, creatinine and urine output should be closely monitored.56 Magnesium citrate and sodium phosphate have been shown to cause CI,68 which can manifest with abdominal pain and diarrhea as quickly as within one hour of taking the medication. It is believed that rapid fluid shifts from the mesenteric circulation to the colonic lumen results in transient hypoperfusion and ischemia. All reported cases have been self-limited. Bisacodyl, a stimulant laxative, has been associated with CI in young healthy patients.69 It is believed that the mechanism of this disease is enhanced colonic motility resulting in decreased mucosal perfusion. Patients usually develop hematochezia and abdominal pain several hours after ingesting bisacodyl. In patients with a history of CI, this medication is relatively contraindicated.

Glycerin enemas have been reported to cause hematochezia six hours after their use, with a flexible sigmoidoscopic appearance consistent with CI.70 One possible mechanism of this insult is inferior mesenteric artery spasm (seen on mesenteric angiography) resulting from increased colonic intraluminal pressure. Nonsteroidal Anti-inflammatory Drugs NSAID-induced colitis manifests with abdominal pain, diarrhea, and hematochezia, with occasional fevers and weight loss.71 It is associated with an endoscopic and histologic appearance similar to CI. Injury induced by NSAIDS is believed to result from long-term use of the medication, with a consequent decrease of vasodilating prostaglandins and increase of vasoconstricting leukotrienes. Serotonin Agonists and Antagonists Serotonin (5-hydroxytryptamine, 5-HT) plays a critical role in modulating enteric neurotransmission and central nervous system signaling. The mechanisms for CI associated with serotonin agonists and antagonists are unproved, although several possibilities have been put forth, including cross-talk among 5-HT receptors and 5-HT3-modulated neurotransmitter release, which possibly is aggravated in atherosclerosis.72 Sumatriptan, a 5-HT1 serotonin receptor agonist used to treat migraine headaches, is an uncommon cause of CI. Its primary mechanism is intracranial vasoconstriction, but in patients with CI, it is believed there is vasoconstriction of the colonic vascular bed.73 Women are more commonly affected than men and typically present with abrupt onset of abdominal pain and hematochezia approximately two days after starting the medication. If CI is suspected, the medication should be discontinued. Alosetron, a 5-HT3 antagonist used to treat diarrhea-­ predominant irritable bowel syndrome (IBS-D) in women, was removed from the United States market in 2000 as a result of its observed association with CI. Alosetron subsequently was reintroduced in 2002 for use in patients with IBS-D who had not responded to conventional therapies. The estimated incidence rate of CI during alosetron treatment (since its reissuance in 2002) is 1.53 cases per 1000 patient-years.74 Approximately one quarter of patients present within one week and one half present within one month of beginning the drug; abdominal pain and bloody diarrhea are typical.74 Alosetron-induced CI is usually reversible and rarely has caused stricture or gangrene. As a result of the association between serotoninergic medications and CI, these medications should not be used in any patient at an increased risk of CI or in women with a history of an ischemic event in any vascular bed. Relative contraindications to its use include a history of hyperactive vascular disorders (migraine headaches) and history of deep vein thrombosis.

PATHOLOGY

Morphologic changes after CI vary with the duration and severity of the injury. The mildest injury is mucosal and submucosal hemorrhage and edema, with or without partial necrosis and ulceration of the mucosa. With more severe injury, chronic ulcerations, crypt abscesses, and pseudopolyps develop, changes that can mimic inflammatory bowel disease (Fig. 114-11).75 Pseudomembranes also may be seen. Iron-laden macrophages and submucosal fibrosis are characteristic of ischemic injury. With severe ischemia, the muscularis propria is replaced by fibrous tissue, forming a stricture. The most-severe form of ischemic damage causes transmural infarction.

Chapter 114  Intestinal Ischemia

Figure 114-11.  Colonoscopic views of deep ulcerations in a patient with colon ischemia whose illness was misdiagnosed as Crohn’s disease.

CLINICAL FEATURES AND DIAGNOSIS

CI usually manifests with sudden cramping, mild, left lower abdominal pain; an urgent desire to defecate; and passage within 24 hours of bright red or maroon blood or bloody diarrhea. Bleeding is not sufficient to require transfusion. Mild to moderate abdominal tenderness usually is present over the involved segment of bowel. Patients with ischemia isolated to the right side of the colon more often present with lower abdominal pain than they do with rectal bleeding or bloody diarrhea.54 A large retrospective study of patients with biopsy-proven CI showed that no region of the colon is spared from involvement. A segmental pattern of involvement is seen most commonly and the sigmoid is affected most often (22.9%), followed by the descending-to-sigmoid colon segment (11.0%), the cecum-to-hepatic flexure segment (8.0%), the descending colon alone (8.0%), and a pancolonic pattern (6.6%).76 Although no specific etiology was associated with any specific anatomic distribution, pancolitis and isolated right-sided colonic disease were seen frequently in patients with sepsis.76 In older reports, certain causes were believed to affect particular segments: local nonocclusive ischemic injuries, the watershed areas (the splenic flexure and rectosigmoid); ligation of the IMA, the sigmoid. The length of affected bowel can depend on the cause of CI: atheromatous emboli involve short segments, and nonocclusive injuries involve longer portions of colon. If CI is suspected, a CT scan might support the clinical suspicion and also diagnose potential complications.77 If the CT scan shows only nonspecific findings such as a thickened segment of colon, or if the abdominal plain film appears normal, colonoscopy should be performed on the unprepared colon within 48 hours of the onset of symptoms. During colonoscopy and barium enema examination, care should be taken not to overdistend the colon because high intraluminal pressure diminishes intestinal blood flow and can aggravate ischemic damage, particularly in patients with vasculitis.78 Colonoscopy is preferable to barium enema because it is more sensitive in diagnosing mucosal abnormalities, and biopsy specimens may be obtained. Hemorrhagic nodules seen at colonoscopy represent bleeding into the submucosa and are equivalent to thumbprints on barium enema studies (Fig. 114-12). Segmental distribution of these findings, with or without ulceration, is highly suggestive of CI, but the diagnosis of CI cannot be made conclusively on the basis of a single examination unless mucosal gangrene is seen (Fig. 114-13). A colonoscopic finding called the colon singlestripe sign has been described in patients with CI, referring to a single line of erythema with erosion or ulceration ori-

Figure 114-12.  Colonoscopic equivalent of a radiologic thumbprint, caused by subepithelial hemorrhage and edema, in a patient with colon ischemia.

Figure 114-13.  Colonoscopic view of mucosal gangrene in a patient with colon ischemia. The necrotic epithelium appears black against the relatively healthier tissue.

ented along the longitudinal axis of the colon; it had a 75% histopathologic yield in making the diagnosis of ischemic injury and signified a milder course than did a circumferential ulcer.79 Segmental disease, rectal sparing, and rapid spontaneous evolution usually resulting in resolution of disease are characteristics of CI. The initial diagnostic study should be performed within 48 hours, because thumbprinting disappears within days as the submucosal hemorrhages are resorbed or the overlying mucosa sloughs. Studies performed one week after the initial study should reflect evolution of the injury—either normalization of the colon or replacement of the thumbprints with a segmental ulcerative colitis-type pattern (Fig. 114-14). Universal colonic involvement, however, favors true ulcerative colitis, whereas fistula formation suggests Crohn’s disease. Occasionally, an abundant inflammatory response can produce heaping up of mucosa and submucosa that resembles a stricture or neoplasm (Fig. 114-15). At the time of symptom onset, colon blood flow typically has returned to normal; therefore, mesenteric angiography

2041

2042

Section X  Small and Large Intestine

A

B

C

Figure 114-14.  Films from serial barium enema examinations in a patient with reversible ischemic damage to the transverse colon and splenic flexure. A, Initial study shows dramatic thumbprints (arrows) throughout the area of involvement. B, Eleven days later, the thumbprints have resolved, and the involved colon has the appearance of segmental colitis. C, Five months after the onset, the colon has returned to normal. The patient was asymptomatic by three weeks after the onset of her illness. (From Boley SJ, Schwartz SS. Colonic ischemia: Reversible ischemic lesions. In Boley SJ, Schwartz SS, Williams LF, editors. Vascular disorders of the intestines. New York: Appleton-Century-Crofts; 1971. p 589.)

or throughout the colon imply CI, whereas nodules isolated to the ascending colon suggest the possibility of otherwise silent SMA disease and the need to evaluate the mesenteric vasculature.

Figure 114-15.  Colonoscopic view of colon ischemia resembling neoplasia in a patient with metastatic cancer treated with interleukin-2 and interferon-α. The inflamed, edematous mass was thought to be due to cecal neoplasia. The lesion resolved spontaneously after just five days. (From Sparano JA, Dutcher JP, Kaleya R, et al. Colonic ischemia complicating immunotherapy with interleukin-2 and interferon-α. Cancer 1991; 68:1538.)

usually is not indicated. An exception to this rule is when the clinical presentation does not allow a clear distinction to be made between CI and AMI or perhaps when only the ascending colon is involved. Administration of air during flexible sigmoidoscopy or a limited colonoscopy can subsequently and immediately be used to reveal thumbprinting not otherwise visible on abdominal plain films; thumbprints stand out as relatively radiodense nodules against the radiolucency of the administered air. Nodules in the left colon

CLINICAL COURSE AND TREATMENT (Fig 114-16) When CI is diagnosed and physical examination does not suggest gangrene or perforation, the patient is treated expectantly. Parenteral fluids are administered and the bowel is placed at rest. Broad-spectrum antibiotics are given to cover the fecal flora because in experimental models, antibiotics reduce the extent and severity of bowel damage. No randomized, controlled, blinded trials have been done to prove the validity of this recommendation. An electrocardiogram, Holter monitoring, and transthoracic echocardiogram should be obtained to exclude or confirm a cardiac source of embolism. Patients with segmental, nongangrenous CI undergoing such evaluation are 2.5 times more likely to have their cardiac risk factor identified compared with other patients with CI.80 Cardiac failure and arrhythmias are treated, and medications that can cause mesenteric vasoconstriction are withdrawn. If the colon appears distended, it is decompressed with a rectal tube. Serial imaging tests of the colon and continued monitoring of the hemoglobin level, WBC count, and electrolyte levels are indicated until the patient’s condition stabilizes. Increasing abdominal tenderness, guarding, rebound tenderness, rising temperature, and paralytic ileus indicate colonic infarction and demand immediate laparotomy and colon resection if appropriate. At operation, mucosal injury may be extensive despite normal-looking serosa, and the extent of resection should be guided by the distribution of disease as seen on preoperative studies rather than the appearance of the serosal surface of the colon at the time of operation. In more than half of patients with CI, the disease is reversible. Generally, the symptoms of CI resolve within 48 to 72

Chapter 114  Intestinal Ischemia Diagnosis of colon ischemia by colonoscopy or BE

Evolving signs of peritonitis or clinical deterioration

IV fluids, antibiotics NPO for 48–72 hr Maximize cardiac output Avoid vasopressors

Condition stable or improving

Continued diarrhea or bleeding, >2–3 wk

Consider repeat BE or colonoscopy after 1–2 wk

Laparotomy or laparoscopy Symptomatic Resection of involved bowel

Recurrent fever or sepsis

Segmental colitis

Normal

Asymptomatic

Observe

Observe

Stricture formation

Consider colonoscopic dilation Unsuccessful

Successful

Observe

hours and the colon heals in one to two weeks. With severe injury, it can take one to six months for the colon to heal; however, during this time the patient is usually asymptomatic. A retrospective study of 350 patients with biopsyproven CI showed that those with isolated right-sided colon ischemia had a worse outcome than those with CI isolated to other segments, including a five-fold increase in the need for surgery and a two-fold increase in mortality.54 Another retrospective study showed that patients’ ages, leukocyte counts, lactate dehydrogenase levels, blood lactate levels, and absence of vascular flow to the colonic wall on abdominal Doppler ultrasonography were independent predictors of complicated CI; only absence of arterial flow was a significant predictor of complicated disease when confounding for other factors.81 Symptoms that persist for more than two weeks also are associated with a higher incidence of acute complications and irreversible disease: gangrene and perforation, segmental ulcerating colitis, or stricture.

Gangrene

Abdominal tenderness with fever and signs of peritonitis suggests infarction and the need for emergent laparotomy.

Segmental Ulcerating Colitis

Segmental ulcerating colitis may be seen with recurrent fevers and sepsis, continuing or recurrent bloody diarrhea, and persistent or chronic diarrhea with protein-losing colopathy. Patients who are asymptomatic or who are minimally symptomatic but have endoscopic evidence of persistent disease should undergo follow-up colonoscopy to determine whether the colitis is healing, becoming chronic, or forming a stricture. Recurrent fever, leukocytosis, and septicemia suggest unhealed segmental colitis and, if found, mandate resection of the ischemic segment of bowel. Patients with persistent diarrhea, bleeding, or protein-losing

Treat as IBD?

Figure 114-16.  Algorithm for the management of colon ischemia. Solid lines indicate conventional management plan; dashed line indicates alternative management plan. BE, barium enema; IBD, inflammatory bowel disease; IV, intravenous; NPO, nothing by mouth; PLC, protein-losing colopathy. (From Brandt LJ, Boley SJ. AGA technical review on intestinal ischemia: American Gastrointestinal Association. Gastroenterology 2000; 118:954.)

colopathy of more than two weeks’ duration are at high risk for perforation, and resection is indicated. Patients who present with segmental ulcerating colitis are often given a misdiagnosis of inflammatory bowel disease. Response to oral steroid therapy usually is poor and may be associated with an increased incidence of perforation. Success has been achieved with fatty acid enemas and corticosteroids given per rectum (Dr. L. Brandt, personal observation). Patients whose symptoms cannot be controlled medically should have a segmental resection, which usually is curative.

Ischemic Stricture

Ischemic strictures that produce no symptoms can be observed. Some disappear over 12 to 24 months with no therapy. Of course, resection is required for those that cause obstruction. There is limited experience with endoscopic dilation of ischemic strictures, although in a few cases, this technique has been successful.

Universal Fulminant Colitis

Sudden onset of a toxic universal colitis picture with signs of peritonitis and a rapidly progressive course are typical of universal fulminant colitis, a rare variant of CI. Total abdominal colectomy with ileostomy usually is required.

SPECIAL CLINICAL PROBLEMS Isolated Ischemia of the Right Colon

Ischemia that only involves the right side of the colon (IRCI) recently has been shown to occur in 23% of cases, an incidence more than twice the conventionally accepted incidence of 8% to 10%.54,76 Such a pattern of involvement is important to document, because these patients require surgery twice as often as and have a mortality rate five times greater than patients with involvement of other areas of the

2043

2044

Section X  Small and Large Intestine colon, including those in whom the right side is involved synchronously with other segments (surgery: 54.9% vs. 10.9%; 30-day mortality 22.5% vs. 11.9%).76 Because the SMA supplies blood to the right side of the colon—as well as to the small intestine—and because a substantial number of patients with IRCI have it as the heralding presentation of otherwise silent SMA obstructive disease, we currently recommend evaluating the splanchnic vasculature in these patients. A CTA will identify patients at risk for AMI and, if an SMA occlusion is found, revascularization should be strongly considered. IRCI is an exception to our general practice of not evaluating the splanchnic vascular system in a patient with CI.

Colon Ischemia in Patients with Carcinoma of the Colon and Other Potentially Obstructive Lesions

Less than 5% of patients with CI have a distal and potentially obstructing lesion or disorder, including carcinoma of the colon, diverticulitis, volvulus, fecal impaction, postoperative stricture, prior ischemic stenosis, or radiation stricture. Typically, the associated lesion is distal, and there is a segment of normal colon between the distal lesion and the proximal colitis (Fig. 114-17). The mechanism of this association may involve increased intracolonic pressure proximal to the lesion with resultant decreased colon blood flow.

Colon Ischemia in Irritable Bowel Syndrome

CI occurs 3.4 to 3.9 times more frequently in the presence of IBS than without it.55,82 Although some authors believe

that patients with IBS visit their doctor more frequently and therefore are more likely to have CI diagnosed than the general population, others hypothesize there is a common pathophysiology in patients with IBS, such as hyper­ sensitivity of the colonic vasculature, autonomic hyper-­ responsiveness, or differences in the sensitivity of serotonin receptors. Clinical studies and experimental models have shown that alosetron increases risk for CI (see earlier73,83,84). The incidence of CI, however, has been shown to be higher in IBS patients regardless of therapy used.85 More studies, better data, and a greater understanding of the mechanisms of actions of serotoninergic agents are needed to further elucidate any such association.

Colon Ischemia Complicating Aortic Surgery

CI complicates elective aortic surgery in up to 7% of cases and surgery for ruptured abdominal aortic aneurysms in up to 60% of cases.86 CI is responsible for approximately 10% of the deaths after aortic replacement. Factors that contribute to postoperative CI include rupture of aneurysm, hypotension, operative trauma to the colon, hypoxemia, arrhythmias, prolonged cross-clamp time, and improper management of the IMA during aneurysmectomy. Tonometric determination of intramural pH of the sigmoid before and after cross-clamping the aorta has been used successfully to predict which patients will develop CI after aneurysmectomy.87 Because postoperative CI is serious and difficult to diagnose early, colonoscopy should be performed within two to three days after surgery for a ruptured abdominal aortic aneurysm or in patients with a prolonged cross-clamping time, a patent IMA on preoperative aortography, nonpulsatile flow in the hypogastric arteries during surgery, or postoperative diarrhea. If CI is identified, oral feeding and liquids are stopped and antibiotic therapy is begun. Clinical deterioration requires reoperation. At surgery, all ischemic colon must be resected.

CHRONIC MESENTERIC ISCHEMIA (INTESTINAL ANGINA) CMI is uncommon, accounting for less than 5% of intestinal ischemic diseases; it almost always is caused by mesenteric atherosclerosis, although rare causes such as collagen vascular disease and inflammatory vasculopathy are known. There is no specific association between CMI and smoking, although 75% of patients with CMI have a history of smoking.88 Abdominal pain is likely caused by ischemia in the small intestine as blood is stolen from this organ to meet the increased demand for gastric blood flow as food enters the stomach.89 This rationale for why the pain occurs so soon after eating, when food still remains in the stomach, is preferable to the traditional explanation that a fixed and limited blood supply is incapable of meeting the increased metabolic demands of the small intestine during digestion.

CLINICAL FEATURES

Figure 114-17.  Film from a barium enema demonstrating a narrowed segment of colon ischemia (upper arrow) proximal to a carcinoma in the distal sigmoid (lower arrow). The area of colon between the carcinoma and the ischemic segment is normal. (From Boley SJ, Brandt LJ, Veith FJ. Is­ chemic disorders of the intestines. Curr Probl Surg 1978; 15:1.)

The cardinal clinical feature of CMI is abdominal cramping discomfort that usually occurs within 30 minutes after eating, gradually increases in severity, and then slowly resolves over one to three hours. Although minimal at first, abdominal pain progressively increases in severity over weeks to months. The association of pain with meals leads to fear of eating, with resultant weight loss. Nausea, bloating, episodic diarrhea, and malabsorption or constipation

Chapter 114  Intestinal Ischemia can occur, but it is the weight loss and relation of the abdominal pain to the meals that characterize this syndrome. Early in the course of disease, if patients do not eat, they remain pain free; pain occurs only after eating or during a meal. Later, pain can become continuous, and this portends intestinal infarction. Uncommon presentations of CMI include antral ulcerations that are unassociated with Helicobacter pylori and that do not heal on therapy with proton pump inhibitors; gastroparesis that resolves after revascularization; and acalculous cholecystitis. Approximately one third to one half of patients have evidence of cardiac, cerebral, or peripheral vascular disease. Physical findings are usually limited, but patients with advanced disease can appear cachectic. The abdomen typically remains soft and nontender even during painful episodes, although distention may be appreciated. An abdominal bruit is common but nonspecific.

DIAGNOSIS

Diagnosis of CMI is difficult because of the vague nature of the complaints and the lack of a specific diagnostic test. Abdominal plain films and CT scans are usually normal, although vascular calcification may be present. Endoscopic inspection of the gastrointestinal tract usually reveals it to be normal, and random biopsies of the upper tract might show only nonspecific abnormalities; a diagnostic clue may be antral ulcerations that are unassociated with H. pylori and that do not heal on acid-suppression therapy. Barium studies are normal or show nonspecific evidence of either malabsorption or a motility disturbance. Rarely, radionuclear emptying tests might show delayed gastric emptying. Ultrasonography, MR angiography, and even traditional mesenteric angiography all merely reveal morphologic abnormalities and anatomic limitations of splanchnic blood flow; they do not establish the presence or absence of intestinal ischemia. Duplex ultrasonography can be used to identify splanchnic artery stenoses but not to establish the diagnosis of CMI.90 Elevated peak systolic velocity in the SMA and CA of 275 and 200 cm/sec, respectively, is a reliable sign of at least 70% stenosis of these vessels.91 Duplex ultrasonography and phase-contrast cine MR imaging of the SMA and CA have been used to measure the effect of eating on mesenteric blood flow, all based on the principle that eating normally increases blood flow to the small intestine, whereas in CMI, this fails to occur; however, postprandial studies are no better than fasting examinations, especially at lesser degrees of vascular stenosis.91 Gastric tonometry exercise testing (GET) uses a nasogastric tube and peripheral arterial catheter to obtain gastric juice and arterial blood from a fasting patient who has been given acid-suppression medication. Determination is made of the gastric-arterial Pco2 gradient before, during, and after exercise. An increase in the gradient after exercise is an indicator of gastrointestinal ischemia. In a study of the diagnostic accuracy for chronic gastrointestinal ischemia, GET and duplex sonography, alone and in combination, were compared with the gold standard, angiography. Combined GET and duplex sonography would not have missed any patients with symptomatic gastrointestinal ischemia; therefore, they are currently believed to constitute the best diagnostic workup strategy.92 More experience with these provocative tests is needed, however, before firm conclusions can be made about their diagnostic usefulness. In the absence of any specific, reliable diagnostic test, the diagnosis of CMI is based on clinical symptoms, in combination with radiologic demonstration of an occlusive process of the splanchnic vessels, and, to a great measure, the exclusion of other gastrointestinal disorders. Angiogra-

phy should show occlusion of two or more splanchnic arteries to allow the diagnosis of CMI; however, such occlusions, even of all three vessels, do not by themselves make the diagnosis of CMI, because they may be present with no corresponding clinical symptoms. In most patients with CMI, at least two of the three splanchnic vessels either are completely obstructed or severely stenotic. In a review of series of patients with CMI,88 91% had occlusion of at least two vessels and 55% had involvement of all three; 7% and 2% had isolated occlusion of the SMA and CA, respectively.

TREATMENT

CMI is not considered to require urgent therapy, although acute complete occlusion of the gastrointestinal blood supply can occur if thrombosis is superimposed on already narrowed arteries. A patient with the typical pain of CMI and unexplained weight loss whose diagnostic evaluation has excluded other gastrointestinal disease and whose angiogram shows occlusive involvement of at least two of the three major arteries should undergo revascularization (Fig. 114-18). Surgical revascularization has been the traditional method of therapy for patients with CMI. Since the early 1980s, percutaneous transluminal mesenteric angioplasty (PTMA) alone or with stent insertion has been used as alternative therapy. The results of surgical revascularization for CMI vary in different reports, depending on the nature of the operations used, the number of vessels revascularized, and whether concurrent operations such as aortic reconstruction are performed.

Suspicion of chronic mesenteric ischemia by clinical criteria Exclusion of other diseases Screening tests: Doppler ultrasound, MRA, or spiral CTA

Normal

Splanchnic angiography

Abnormal

Observe

Abnormal

Normal

Treat

Observe Good-risk patient

Poor-risk patient

Angioplasty with or without stent Successful

Surgical revascularization

Unsuccessful

Observe Figure 114-18.  Algorithm for the management of chronic mesenteric is­chemia. Solid lines indicate conventional management plan; dashed lines indicate alternative management plan. CTA, computed tomographic angiography; MRA, magnetic resonance angiography. (Modified from Brandt LJ, Boley SJ. AGA technical review on intestinal ischemia: American Gastrointestinal Association. Gastroenterology 2000; 118:954.)

2045

2046

Section X  Small and Large Intestine The true efficacy of surgical revascularization and PTMA is difficult to determine because of the varied criteria used by different investigators to define a successful outcome. Thus, some authors use graft or vessel patency rates, whereas others define success by relief of symptoms, recurrence rates, or long-term survival. A tabulation of 17 series of surgical revascularization for CMI totaling 614 patients yielded perioperative mortality rates that ranged from 0%93-98 to 16%,99 success rates of 59%100 to 100%,94-98 and recurrence rates of 0%85,93,94,99,101,102 to 26.5%.103 Most recent series have reported mortality rates less than 10%, success rates of more than 90%, and recurrence rates generally less than 10%.19 Several long-term studies have shown that patients surviving surgical revascularization have cumulative 5-year survival rates of approximately 80% to 90%.19 The rates of success for PTMA are similar to those for surgical revascularization. The experience with PTMA is more limited but has been achieved in patients often considered too high risk for a surgical procedure. In 10 representative series of PTMA for CMI, totaling 128 patients, clinical success rates (i.e., relief of symptoms) have varied from 63%104 to 100%,105 with little mortality.19 Recurrence of symptoms, however, has been much higher than after surgical revascularization, varying from 10% to 67% in the larger series.19 Intraluminal stenting is an adjunctive treatment to PTMA and attempts to decrease the incidence of recurrent stenoses. Percutaneous transluminal mesenteric angioplasty with stenting (PTMAS) results in the same short-term patency rates for both stenotic and occluded vessels.106 Long-term studies of PTMAS patency rates indicate that approximately 70% are patent at seven years and 56% of patients are free from recurrent symptoms. When compared with open surgical bypass, PTMAS shows a decreased primary patency, decreased primary assisted patency, and earlier required reinterventions.107 Patients with CMI who are otherwise relatively healthy probably should be treated by surgical revascularization; patients at poorer risk should have an initial attempt at PTMA or PTMAS to relieve symptoms.

VASCULITIS AND ANGIOPATHY OF THE SPLANCHNIC CIRCULATION Inflammation and necrosis can affect splanchnic blood vessels of all sizes: arteries, veins, the vasa recta, arterioles, and venules.108 Symptoms depend on the size of the involved vessel and may be indistinguishable from AMI caused by emboli or thromboses; associated systemic features such as renal failure, cutaneous nodules, and pulmonary infiltrates suggest a systemic disorder. The vasa recta and intramural arteries and arterioles may be affected in systemic vasculitides. With vasculitis, the ischemic injury typically involves short segments of the intestine. Abdominal pain, fever, gastrointestinal bleeding, diarrhea, and intestinal obstruction are common. Ulceration and stricture formation are common, but with small-vessel involvement perforation is less common. Typically, vasculitis is caused by immune complex deposition in the walls of vessels, which leads to activation of the complement system and an inflammatory reaction; aneurysm formation, vessel rupture and bleeding, vascular occlusion, thrombosis, or fibrosis can ensue. A variety of vasculitides are discussed in the following sections.

ALLERGIC GRANULOMATOUS ANGIITIS (CHURG-STRAUSS SYNDROME)

Allergic granulomatous angiitis is a disorder that is typified by asthma, glomerulonephritis, eosinophilia, and granulomatous inflammation associated with antineutrophil cytoplasmic autoantibodies.109 Necrotizing vasculitis affects small and medium-sized vessels and involves the gastrointestinal tract in almost half the patients. As in other vasculitides, abdominal pain and bleeding secondary to ischemia are the usual manifestations. Glucocorticoid therapy usually is effective.

BEHÇET’S DISEASE

Behçet’s disease (see Chapter 35) is characterized by oral and genital ulcers, recurrent iritis or chorioretinitis, and skin lesions. It is most often seen in Eastern Mediterranean men and is strongly associated with the B51 allele. Smallvessel vasculitis accounts for much of the damage, but largevessel involvement of arteries and veins is not uncommon. Gastrointestinal disease, present in 50% of patients, typically involves the ileocecal area, although involvement of the esophagus and small intestine has been reported.110 Attacks are recurrent and usually self-limited except for the uveitis which may be chronic. The most common gastrointestinal symptoms are abdominal pain, diarrhea, and bleeding; deep ulcers are responsible for the most common intestinal complications: severe bleeding and perforation. Mortality is low in Behçet’s disease; however, intestinal perforation is one of the common causes of death. Therapy with glucocorticoids, immunosuppressive agents, and colchicine has been tried, with varying success.

BUERGER’S DISEASE

Also called thromboangiitis obliterans, Buerger’s disease involves medium-sized and small peripheral arteries and veins, especially the infrapopliteal vessels; foot claudication and rubor are the most common symptoms. It is largely a disease of men, especially those who have smoked cigarettes from an early age, and typically has its onset before the age of 50 years; there is a distinct absence of other atherosclerotic risk factors. Intestinal involvement is unusual, but most common is involvement of the vessels supplying the small intestine.111 In the acute lesion, inflammation spreads outward from the thrombus-endothelium interface through the vessel wall. Later, microabscesses, necrotizing granulomas, and multinucleated giant cells occur in the thrombus, after which the thrombus organizes and becomes occlusive. Intestinal involvement usually requires resection.

COGAN’S SYNDROME

Cogan’s syndrome is a rare disorder of young people characterized by vasculitis of the conjunctiva, cornea, and cochlea.112 Although this vasculitis usually is localized, it is considered to be a hypersensitivity reaction to an unknown viral agent, and the disease can become disseminated. Three percent to 10% of patients develop gastrointestinal symptoms, with diarrhea and bloody stools. High-dose glucocorticoids, and occasionally cytotoxic agents, are required. Vascular surgery may be needed after inflammation is controlled.

FIBROMUSCULAR DYSPLASIA

FMD is a rare angiopathy that is neither related to atherosclerosis nor to inflammation.113 Its cause is unknown, although genetic factors might play a role and there is an association with cigarette smoking and hypertension. There

Chapter 114  Intestinal Ischemia are several types of FMD depending on which arterial layer is involved: intima, media, or adventitia. The renal arteries are most commonly involved, followed by the carotid and vertebral arteries and then other vasculature including the mesenteric arteries. Splanchnic involvement can manifest with symptoms of AMI or CMI.114 Diagnosis is based on the same techniques used to image other vascular disorders; the classic string-of-beads appearance is typical of only the medial type of FMD, whereas aneurysms and dissection are known complications of all types. Therapy consists of percutaneous transluminal angioplasty, surgical revascularization, and resection of necrotic bowel as indicated.

HENOCH-SCHÖNLEIN PURPURA

Henoch-Schönlein purpura (see Chapter 35) typically affects children aged 4 to 7 years. It is characterized by IgA immune complexes deposited within the small vessels of the skin, gastrointestinal tract, joints, and kidneys and is often preceded by an upper respiratory infection. The classic clinical triad consists of palpable purpura (usually below the waist), arthritis (knees and ankles), and abdominal pain; the gastrointestinal tract is involved in up to 75% of patients.115 Abdominal pain and gastrointestinal bleeding are the most common gastrointestinal symptoms and are caused by mucosal and submucosal hemorrhage; a submucosal hematoma may be the lead point of an intussusception. Gastrointestinal involvement may be documented by endoscopy116 or by CT study.117 The disease is usually self-limited, but the outlook may be less favorable in adults, in large measure because of the development of renal failure.

HYPERSENSITIVITY VASCULITIS

Hypersensitivity vasculitis uncommonly involves the splanchnic vasculature, and, in contrast with necrotizing vasculitis, which involves arteries, affects mainly the postcapillary venules. A large variety of causes are known to trigger this disorder, including infections (Streptococcus, Staphylococcus, hepatitis B virus, influenza virus, cyto­ megalovirus, mycobacteria, and rickettsiae), drugs, and chemicals.

KAWASAKI’S DISEASE

Kawasaki’s disease, also called infantile febrile mucocutaneous lymph node syndrome, is a necrotizing vasculitis of medium-sized arteries.118 It manifests as fever, rash on the palms and soles, desquamation, conjunctival congestion, strawberry tongue, and cervical lymphadenopathy in infants and children. Many patients have nausea, vomiting, abdominal pain, and diarrhea, and they can suffer ileus, small bowel obstruction, bleeding, and perforation. Death may be due to coronary artery aneurysms and myocardial infarction. Treatment is aspirin for the acute phase and large intravenous doses of gamma globulin for the prevention of coronary artery aneurysms.

KÖHLMEIER-DEGOS DISEASE (MALIGNANT ATROPHIC PAPULOSIS)

Köhlmeier-Degos disease is a rare form of progressive occlusive vascular disease of young men that affects the small and medium-sized arteries, mainly those of the skin and intestine.119 Typically, skin lesions of porcelain-white punctate scars with erythematous borders are found on the trunk and upper extremities. The rash is followed, within months to years, by the development of abdominal pain and spontaneous intestinal perforation. Thrombosis of small and medium-sized vessels is found, without inflammatory

cell infiltration. There is no known therapy for this disease, and it is generally fatal.

POLYARTERITIS NODOSA

Polyarteritis nodosa (see Chapter 35) is a necrotizing vasculitis of medium-sized and small arteries characterized by aneurysms at branch points. Abdominal symptoms are reported in up to half of patients with the disorder, the most common of which is abdominal pain, usually from is­ chemia.120 Involvement of the small intestine is most common, followed by lesions of colon, liver, and pancreas. Diagnosis is suggested by typical angiographic findings of aneurysms in the mesenteric, renal, and hepatic vasculature. Treatment with glucocorticoids and cyclophosphamide or azathioprine has improved survival greatly. Vasculitis resembling polyarteritis also is associated with hepatitis B and C virus infections.121,122 Fifty percent of patients with classic polyarteritis are hepatitis B surface antigen positive, but unlike in classic polyarteritis, only small arteries are involved. Patients develop a polyarteritis picture following the viral infection, presumably from deposition of antigen-associated immune complexes on the vessel wall.

RHEUMATOID VASCULITIS

Rheumatoid vasculitis affects the gastrointestinal tract in approximately 10% of patients, usually those who have subcutaneous nodules and who are seropositive for rheumatoid factor.123 As with all vasculitides, ischemia manifests with abdominal pain, bleeding, perforation, and gangrene. Other diseases have been noted in association with rheumatoid arthritis, including atrophic gastritis, gastric antral vascular ectasia, inflammatory bowel disease, collagenous colitis, and amyloidosis.

SYSTEMIC LUPUS ERYTHEMATOSUS

Systemic lupus erythematosus (see Chapter 35) affects the gastrointestinal system in about half of cases and can involve any of the hollow and solid gastrointestinal organs.108 The most common symptoms are nausea, vomiting, and abdominal pain, but diarrhea, malabsorption, pseudo-obstruction, peritonitis, pancreatitis, protein-losing enteropathy, and ascites are also well-known occurrences. The systemic nature of the disorder makes differential diagnosis complicated, but vasculitis-induced ischemia underlies many of the presentations. The vasculitis typically involves small vessels and causes FSI and gastrointestinal bleeding, both of which are associated with high mortality rates if not diagnosed promptly.

TAKAYASU’S DISEASE

Takayasu’s disease (pulseless disease) is an idiopathic chronic inflammatory disorder that most often affects the aorta and its branches in young women of Asian heritage; it is unusual that the splanchnic vessels are involved.124 Fibrotic occlusion of the involved vessels is the end result of the inflammatory process. Takayasu’s disease rarely has been associated with Crohn’s disease and ulcerative colitis, and in the serum of some of these patients, antibodies to colonic mucosa and aorta have been detected. Treatment is large doses of glucocorticoids before reconstructive surgery. The five-year survival rate is higher than 90%.

ACKNOWLEDGMENT

I am indebted to Dr. Scott J. Boley, who for decades has been laying the groundwork for this chapter and who played a critical role in preparing earlier editions.

2047

2048

Section X  Small and Large Intestine KEY REFERENCES

Acosta S, Ogren M, Sternby NH, et al. Clinical implications for the management of acute thromboembolic occlusion of the superior mesenteric artery: autopsy findings in 213 patients. Ann Surg 2005; 241:516-22. (Ref 4.) Acosta-Merida MA, Marchena-Gomez J, Hemmersbach-Miller M, et al. Mesenteric venous thrombosis. Associated systemic disorders and hypercoagulability status of 21 surgical patients. Hepatogastroenterology 2007; 54:1080-4. (Ref 36.) Atkins MD, Kwolek CJ, LaMuraglia GM, et al. Surgical revascularization versus endovascular therapy for chronic mesenteric ischemia: a comparative experience. J Vasc Surg 2007; 45:1162-71. (Ref 107.) Boley SJ, Feinstein FR, Sammartano R, et al. New concepts in the management of emboli of the superior mesenteric artery. Surg Gynecol Obstet 1981; 153:561-9. (Ref 31.) Boley SJ, Sprayregan S, Siegelman SS, Veith FJ. Initial results from an agressive roentgenological and surgical approach to acute mesenteric ischemia. Surgery 1977; 82:848-55. (Ref 30.) Brandt LJ, Boley SJ. AGA technical review on intestinal ischemia. American Gastrointestinal Association. Gastroenterology 2000; 118:95468. (Ref 19.) Cole JA, Cook SF, Sands BE, et al. Occurrence of colon ischemia in relation to irritable bowel syndrome. Am J Gastroenterol 2004; 99:486-91. (Ref 55.) Hass DJ, Kozuch P, Brandt LJ. Pharmacologically mediated colon ische­ mia. Am J Gastroenterol 2007; 102:1765-80. (Ref 56.)

Higgins PD, Davis KJ, Laine L. Systematic review: The epidemiology of ischaemic colitis. Aliment Pharmacol Ther 2004; 19:729-38. (Ref 82.) Kougias P, Lau D, El Sayed HF, et al. Determinants of mortality and treatment outcome following surgical interventions for acute mesenteric ischemia. J Vasc Surg 2007; 46:467-74. (Ref 20.) Lee SS, Ha HK, Park SH, et al. Usefulness of computed tomography in differentiating transmural infarction from nontransmural ischemia of the small intestine in patients with acute mesenteric venous thrombosis. J Comput Assist Tomogr 2008; 32:730-7. (Ref 45.) Otte JA, Geelkerken RH, Huisman AB, Kolkman JJ. What is the best diagnostic approach for chronic gastrointestinal ischemia? Am J Gastroenterol 2007; 102:2005-10. (Ref 92.) Poole JW, Sammartano RJ, Boley SJ. Hemodynamic basis of the pain of chronic mesenteric ischemia. Am J Surg 1987; 153:171-6. (Ref 89.) Sotiriadis J, Brandt LJ, Behin DS, Southern WN. Ischemic colitis has a worse prognosis when isolated to the right side of the colon. Am J Gastroenterol 2007; 102:2247-52. (Ref 54.) Zandrino F, Musante F, Gallesio I, Benzi L. Assessment of patients with acute mesenteric ischemia: multislice computed tomography signs and clinical performance in a group of patients with surgical correlation. Minerva Gastroenterol Dietol 2006; 52:317-25. (Ref 15.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

115 Ulcers of the Small and Large Intestine Caroline Loeser and Deborah Denise Proctor

CHAPTER OUTLINE Isolated Ulcers  2049 Nonspecific or Idiopathic Small Intestinal Ulceration  2049 Solitary Rectal Ulcer Syndrome  2050 Stercoral Ulcers of the Colon  2052 Ulceration Induced by Nonsteroidal Anti-inflammatory Drugs  2052

Ulcers of the small and large intestine are rare, yet they are responsible for a broad spectrum of disease. Ulcers can occur singly, as in solitary rectal ulcer syndrome, or diffusely, as in enteropathy-associated T-cell lymphoma. Clinical presentations vary widely with location and degree of intestinal involvement, ranging from anemia and hypoproteinemia to abdominal pain, hemorrhage, obstruction, and perforation. This chapter is divided into two sections. The first section covers isolated intestinal ulcers, including nonspecific solitary ulcers of the small intestine, solitary rectal ulcer syndrome, stercoral ulcers, and ulcerations induced by nonsteroidal anti-inflammatory drugs (NSAIDs). The second section covers syndromes of diffuse intestinal ulceration, including ulcerative enteritis, refractory celiac disease types I and II, and enteropathy-associated T-cell lymphoma (EATL). Because of the length and relative inaccessibility of the small intestine, diagnosis of small intestinal ulcerative diseases has been challenging. Currently, however, videocapsule endoscopy enables the entire small intestine to be viewed with clarity, and double-balloon enteroscopy offers the opportunity to evaluate the small intestine visually, obtain biopsy specimens, and provide a variety of endoscopic therapies.

ISOLATED ULCERS NONSPECIFIC OR IDIOPATHIC SMALL INTESTINAL ULCERATION

Solitary ulcers of the small intestine result from a wide variety of causes (Table 115-1). Radiation injury is a known cause of small intestinal ulceration and is discussed in Chapter 39. Solitary ulcers beyond the duodenum that cannot be explained on the basis of any known etiology are referred to as nonspecific or idiopathic intestinal ulcers. Such solitary nonspecific ulcers are rare, with an incidence of 4 per 100,000.1

Diffuse Ulcerations  2054 Definitions  2054 Background  2054 Clinical Features  2055 Diagnosis  2055 Treatment and Prognosis  2056

Clinical Features

Patients with nonspecific ulcers of the small intestine can present with acute or chronic gastrointestinal bleeding, symptoms of small bowel obstruction, abdominal pain, or perforation. Symptoms may be present from a few days to many years before diagnosis. In a review of the Mayo Clinic’s experience with 59 cases of small intestinal ulcers over a 22-year period ending in 1979, Boydstun and associates1 showed that 53 (89.8%) patients had no identifiable cause of ulceration. Patients ranged in age from 17 to 77 years, with most presenting in the fifth and sixth decades of life; no gender predominance was found. The most common presenting symptom was intermittent small bowel obstruction (63%). Physical findings ranged from nonspecific abdominal tenderness and distention to an acute abdomen resulting from intestinal perforation. Laboratory evaluation was notable only for anemia in one half of the patients. Radiologic studies localized the ulcer in a minority of patients.

Pathology

In the Mayo Clinic series,1 the ileum was the most common location of nonspecific ulceration (78%), and perforation (13 cases, 22%) occurred most commonly in the jejunum (78%). At surgery, 41 patients were found to have solitary ulcers, five patients had two ulcers, and six patients had more than three ulcers. Ulcer size varied between 0.3 and 5 cm. On pathologic examination, the ulcers were predominantly on the antimesenteric border of the small intestine and, in some cases, were associated with a fibrous scar that narrowed the lumen. Microscopy revealed nonspecific chronic inflammation that ended abruptly at the ulcer edge. The intervening bowel and vasculature were normal.1

Treatment

All patients were treated with segmental resection; only two patients had recurrent ulceration, 2 and 10 years after initial diagnosis and resection. Ischemia, central nervous system disease, infection, trauma, and hormonal influences all have been put forth as possible causes of primary nonspecific

2049

2050

Section X  Small and Large Intestine Table 115-1  Causes of Small Intestinal Ulceration Acid-Related Disorders Heterotopic gastric mucosa Zollinger-Ellison syndrome Celiac Disease Refractory celiac disease Ulcerative enteritis Collagen-Vascular and Other Immunological Diseases Autoimmune enteropathy Churg-Strauss syndrome Giant cell arteritis Henoch-Schönlein purpura Mixed connective tissue disease Polyarteritis nodosa Polymyositis-dermatomyositis Reactive arthritis Sjögren’s syndrome (microscopic polyangiitis) Systemic lupus erythematosus Thrombotic thrombocytopenic purpura Vasculitis Congenital Duplications Stenoses Drugs Antibiotics Antimetabolites Aspirin Bacillus Calmette-Guérin Chemotherapeutic agents Nonsteroidal anti-inflammatory drugs Slow-release (enteric-coated) potassium Hypersensitivity Food allergies Hypogammaglobulinemia Infections Bacteria Fungi Parasites Protozoa Viruses Worms Inflammatory Diseases Behçet’s syndrome Crohn’s disease Cryptogenic multifocal ulcerous stenosing enteritis Eosinophilic gastroenteritis Granulomatous enteritis Lymphocytic enterocolitis Nongranulomatous chronic idiopathic enterocolitis Sarcoidosis Tropical sprue Ischemia Incarcerated hernia Intussusception Mesenteric ischemia Vascular abnormalities Metabolic Disorders Malnutrition Uremia Neoplasms Angiocentric T-cell lymphoma Enteropathy-associated T-cell lymphoma Primary neoplasms Metastatic neoplasms Radiation Accidental Therapeutic Toxins Heavy metal poisoning Trauma and Mechanical Injury Foreign body ingestion Stomal ulceration

ulcerations, but the cause or causes still remains unknown. In the absence of more recent reviews, it is impossible to determine the current incidence rate of these ulcers.

SOLITARY RECTAL ULCER SYNDROME

Solitary rectal ulcer syndrome (SRUS) is an uncommon or under-reported disorder of defecation that affects patients of all ages. The term is a misnomer: Patients can present with hyperemic mucosa only, a solitary ulcer, multiple ulcers, or even a polypoid lesion resembling carcinoma.2 Regardless, the histology of SRUS is typical, showing fibromuscular obliteration of the lamina propria and smooth muscle fibers extending from a hypertrophied muscularis mucosa to the lumen.2 The diagnosis of SRUS often is delayed because of its varied endoscopic appearance and a lack of awareness of the disorder.

Pathogenesis

SRUS is a disorder of defecation, but its pathogenesis is uncertain, and it has a spectrum of disease presentations. A large subgroup of patients with SRUS strain excessively during defecation, and some have a behavioral disorder. Occult or overt rectal prolapse with paradoxical contraction of the pelvic floor during defecation appears to be involved in most patients3; evidence of inappropriate pelvic floor contraction has been shown in electromyographic and video-proctographic studies.3 It has been suggested that the rectal mucosa can be traumatized from the pressure of being prolapsed against a closed anal canal4 and that straining during defecation results in prolapse and high fecal voiding pressures that reduce local blood flow, causing ischemia and ulceration.4 The mucosa of the anterior rectal wall, 7 to 10 cm above the anal verge, is the most common area of such prolapse into the anal canal, and this corresponds to the usual location of ulceration in SRUS. SRUS also has been associated with the use of ergotamine suppositories and is well known after radiotherapy, further supporting a pathogenic role for ischemia.5,6 Successful treatment of SRUS using biofeedback has been associated with an increase in local blood flow, additionally suggesting that SRUS may be associated with reduced rectal blood flow from impaired extrinsic autonomic cholinergic nerve activity.7 The association of SRUS and rectal prolapse, however, is neither pathogenically clear nor universal; the incidence of associated rectal prolapse varies from 13% to 94%.2 It is assumed that the ulcer develops as a result of local trauma to the apex of the prolapse, either because of manual attempts to reduce the prolapse digitally or because of contractions of the external anal sphincter when the mucosa prolapses through the anal canal.4 Du Boulay and colleagues have shown that the histology of the rectal mucosa in patients with SRUS is similar to that seen at other sites of mucosal prolapse, suggesting that prolapse of the mucosa alone rather than the entire rectal wall is important in SRUS pathogenesis.8 Ischemia results in fibromuscular obliteration of the lamina propria and the formation of an ulcer. Once the ulcer is formed, it can further intensify the urge to defecate; this urge combined with straining and changes in local blood flow causes persistent symptoms and chronic ulceration.

Clinical Features

Patients with SRUS present with varied symptoms, but most patients typically complain of passage of mucus and blood per rectum on defecation.9,10 Some patients also complain of tenesmus, straining, altered bowel habits, and the sensation of incomplete evacuation. Men and women are affected

Chapter 115  Ulcers of the Small and Large Intestine

* Figure 115-1.  Endoscopic photograph obtained during colonoscopy in a 40-year-old patient with solitary rectal ulcer syndrome. A large (4 × 3 cm) ulcer is seen in the distal rectum 3 cm above the anal verge. The ulcer margins have a polypoid appearance. Black arrows show exudative material amid the nodularity of the ulcer; the white star is the colon lumen.

equally, and they usually present in the third and fourth decades, respectively.11 The mean duration of symptoms is long, ranging from 3.5 to 5 years, possibly reflecting a delay in diagnosis.

Diagnosis and Pathology

Diagnosis of SRUS is based on clinical symptoms, physical examination, endoscopic findings, and typical histology. Physical examination can demonstrate tenderness in the left lower quadrant. On digital rectal examination, there may be reduced anal sphincter tone and an indurated area, or thickened folds may be palpated.12 Overt rectal prolapse may be demonstrated by having the patient assume a squatting position and straining as if to have a bowel movement. Sigmoidoscopy may demonstrate single or multiple ulcers or a patch of erythematous mucosa on the anterior rectal wall within 10 cm of the anal verge. The lesion has a polypoid appearance (Fig. 115-1) in 25% to 44% of patients.2,12 Differential diagnosis includes inflammatory bowel disease, malignancy, ischemic colitis, stercoral ulcer, medication-induced ulceration, trauma, and infections, including amebiasis and secondary syphilis.11 Biopsy specimens always should be taken from the ulcer margin and from any abnormal-appearing mucosa. In 1969, Madigan and Morson9 first described the histologic features of SRUS. There is fibromuscular obliteration of the lamina propria by collagen from fibroblasts and smooth muscle fibers derived from the muscularis mucosae. The muscularis mucosae is often hypertrophied, and its fibers are in continuity with those in the lamina propria. There is no significant increase in the number of inflammatory cells. The polypoid variant is similar to the ulcerative variant except for regenerative hyperplastic changes, such as cystic dila­ tation and mucus cell production.11 Epithelial elements and lamina propria can be displaced into the submucosa (Fig. 115-2). This displaced tissue can then undergo cystic dilatation because of mucus retention. The misplaced and dysplastic-appearing glands may be misdiagnosed as adenocarcinoma, especially when the histologic and macroscopic features of SRUS are not recognized; at times SRUS is present in association with an carcinoma, further confusing the issue.

Figure 115-2.  This high-power photomicrograph of a rectal biopsy specimen illustrates the typical histopathologic features of solitary rectal ulcer syndrome. The findings include disorganized crypts with reactive epithelium, mild lamina propria inflammation, and smooth muscle fibers abnormally present in the mucosa (arrows). The smooth muscle fibers represent hyperplasia of the muscularis mucosa, a common histologic finding in this condition. (Courtesy of Marie E. Robert, MD, New Haven, Conn.)

Histology typically provides a definitive diagnosis; however, defecography may be useful to shed light on the pathophysiology of SRUS, especially if surgery is being considered. Defecography may be used to demonstrate mucosal prolapse, intussusception, rectal prolapse, a nonrelaxing puborectalis muscle, and incomplete or delayed evacuation.11 Endorectal ultrasound can demonstrate the presence and components of rectal wall thickening, particularly the muscularis propria, and may be useful to distinguish SRUS from other conditions such as invasive cancer.11

Treatment

Asymptomatic patients might not require any treatment, and in some patients, SRUS resolves spontaneously. Treatment includes improving bowel habits; consuming a highfiber diet; using bulk laxatives, local agents, and biofeedback; and undergoing surgery. The addition of fiber as a bulking agent along with bowel habit training to reduce straining can improve symptoms in patients with mild disease. Local agents such as topical glucocorticoids and aminosalicylates are not effective. Sucralfate enemas and human fibrin sealant have been effective in small studies.13 Argon plasma coagulation (APC) has been used to treat hemorrhage from SRUS; continued treatment with APC has been associated with symptomatic and endoscopic improvement.14 Behavioral therapy or biofeedback is the first line of therapy for those with more-severe disease, and it improves symptoms in more than 50% of patients; ulcer healing, however, is seen in a minority of patients. Behavioral therapy aims at bowel habit training with normalization of pelvic floor coordination. Jarrett and associates demonstrated that biofeedback resulted in improved rectal blood flow, which was associated with a successful clinical outcome.7 Surgery is indicated in symptomatic patients who have severe disease and who do not respond to medical or biofeedback therapy, but the best surgical option depends upon

2051

2052

Section X  Small and Large Intestine the underlying anatomic pathology, and every patient must be assessed individually.15 Surgical procedures include excision of the ulcer, low anterior resection, colostomy, or anterior resection with rectopexy.10 It is difficult to compare surgical treatments for SRUS because of the small number of patients in surgical series and the variety of anatomic pathology that underlies SRUS.10

STERCORAL ULCERS OF THE COLON

Stercoral ulcers result from pressure necrosis of the mucosa caused by the direct effect of a hard fecal mass (scybalum). Over time, the pressure of the scybalum results in local ischemic necrosis and ulceration and can eventuate in perforation. These ulcers are rare and usually asymptomatic until they manifest with lower gastrointestinal bleeding or colonic perforation. Fecal disimpaction occasionally precipitates rectal hemorrhage when the scybalum is removed, along with an adherent blood vessel in the sub­ jacent ulcer crater. Maurer and coworkers observed that 3.2% of colonic perforations in their series were caused by stercoral ulcers.16 Chronic constipation is the major risk factor for stercoral ulceration and, although described in patients of all ages, it is more common among elderly patients with clinical features that may be associated with constipation.17 Although constipation and fecal impaction are observed commonly, complications of stercoral ulceration are relatively uncommon. Serpell and Nicholls reviewed 64 cases of stercoral perforation of the colon. The median age of these patients was 60 years, and 23% of them were nursing home residents. Factors that increase constipation and formation of a scybalum, such as antacids containing aluminum hydroxide, use of narcotic analgesics, constipating sedatives and psychiatric medications, and chronic renal failure, were observed in patients who developed stercoral ulceration.18 Why a stercoral ulcer develops is unclear, although implicated factors predisposing the left colon to ulceration include dehydrated and hard feces, a narrow-diameter colon with high pressure, and relatively poor blood supply. Patients with perforated ulcers usually present with peritonitis and findings of an acute abdomen.18 The scybalum sometimes is palpable as an abdominal mass. Plain films of the abdomen might demonstrate pneumoperitoneum, fecal loading, or calcified scybala. Nonperforating ulcers can manifest with lower gastrointestinal bleeding. Caution must be used in performing disimpaction in patients with hard fecal masses in the rectum, because removing the mass can result in severe hemorrhage if the underlying blood vessel in the ulcer base is torn during removal. The antimesenteric border of the colon is most commonly involved, usually in the sigmoid or proximal rectum. Ulcers usually are large, irregular, and sharply demarcated, and they may be single or multiple. Ulcers conform to the contour of the impacted scybala and result from ischemic pressure necrosis. A rounded or ovoid perforation may be seen in the center of the ulcer. Necrotic colonic mucosa with acute and chronic inflammation is noted on histology.17 Differential diagnosis includes spontaneous colonic perforation, malignancy, ischemia, and infection. Perforated stercoral ulcers require emergency laparotomy with resection of the affected colonic segment. A Hartmann’s operation is the preferred procedure, and along with extensive peritoneal lavage, it is associated with a lower mortality than other surgical procedures.16,18 Nonperforating stercoral ulcers might respond to antibiotics and aggressive treatment of constipation, although surgical resection remains the only definitive treatment.

ULCERATION INDUCED BY NONSTEROIDAL ANTI-INFLAMMATORY DRUGS

NSAIDs are among the most frequently administered drugs in the world, and their adverse side effects involve not only the stomach and duodenum but also the distal portions of the small intestine and colon. Gastroduodenal risks of NSAIDs are well known, but lower gastrointestinal tract risks of NSAIDs occur with similar frequency.19,20 The gastroduodenal effects of NSAIDs are discussed in Chapter 51.

Clinical Features

NSAIDs cause small intestinal and colonic inflammation and ulceration and have a wide spectrum of manifestations from clinically silent subtle mucosal changes to significant ulceration and overt bleeding or intestinal obstruction.20,21 Inflammation secondary to NSAIDs can cause subtle changes known as NSAID enteropathy, in which there is increased intestinal permeability, inflammation, and subtle bleeding. The clinical picture often is silent and undiagnosed unless it progresses to manifest with anemia and hypoalbuminemia. More-overt presentations include weight loss, anemia, diarrhea, overt bleeding, and perforation.20,22 Symptoms of partial small bowel obstruction such as vomiting and colicky abdominal pain may be seen secondary to the development of diaphragm-like strictures. Laboratory evaluation often is notable for hypoalbuminemia and iron deficiency anemia.

Pathology

Autopsy findings of NSAID users established that small intestinal ulcerations distal to the duodenum were prevalent.22 Of 713 patients studied, NSAIDs had been prescribed to 249 in the six months before death; 8.4% of the NSAID users had ulcerations of the small intestine compared with only 0.6% of the NSAID nonusers.22 Although no information is available regarding morbidity caused by NSAIDs during life, three of the NSAID users died of small intestinal perforation.22 The pathologic appearance of NSAID-induced ulceration is nonspecific: Ulcerations can be single or multiple and range from tiny punched-out ulcers to confluent areas of deep ulcer with stricture formation. Capsule endoscopy has shown abnormalities ranging from reddening of the mucosa to erosions, ulcers, and active bleeding (Fig. 115-3A).23 The intervening mucosa is normal. NSAID ulcers cannot be distinguished from nonspecific or idiopathic intestinal ulcerations on the basis of their gross or microscopic pathologic appearance. Ulcerations rarely are associated with diaphragm-like strictures in patients with longstanding NSAID use, an association referred to as diaphragm disease; diaphragms are thin, 2- to 4-mm thick septae, concentric strictures that comprise mucosa and submucosa with or without submucosal fibrosis (Fig. 115-3B).

Pathogenesis

The mechanisms of NSAID-induced injury to the small intestine are not completely understood, but may involve both systemic and local events (Fig. 115-4). NSAIDs appear to cause a disturbance in the microcirculation of the villi, leading to the loss of epithelial cells.24,25 Initially, endothelial projections develop at the tip of the villus vascular arcade. Blood flow slows and stasis develops in the capillaries at the tip of the villus, occluding vascular lumens. Endothelial cells become vacuolated, with further stasis of blood flow and sloughing of the overlying epithelial cells.24 It is not clear if this is mediated through cyclooxygenase (COX)-1 or COX-2 pathways. The topical effects of NSAIDs appear to require prolonged exposure of the drug to the

Chapter 115  Ulcers of the Small and Large Intestine

A

B

Figure 115-3.  A, Video capsule endoscopy photograph of a nonsteroidal anti-inflammatory drug (NSAID)-induced small intestinal ulcer. B, Video capsule endoscopy photograph of a membranous eccentric stricture in the small intestine caused by NSAID use. (From Cave DR. Iatrogenic diseases. In: Halpern M, Jacob H, editors. Atlas of Capsule Endoscopy. 1st ed. Haifa: Given Imaging; 2002. p 65. Copyright Given Imaging Ltd, 2004. Used with permission.)

NSAID

Mitochondrial uncoupling and oxygen radical formation Inhibition of reparatory prostaglandin synthesis

Initial damage Increased intestinal permeability Increased mucosal susceptibility to further damage Chemotactic events Endogenous aggressors: Polymorphonuclear neutrophils Inflammatory mediators

Luminal aggressors: Bile Bacteria Hydrolytic enzymes Proteolytic enzymes

measured by scintigraphic assessment of accumulation of indium-111–labeled white blood cells in the small intestine and by fecal excretion of indium-111.27 Nineteen of the 32 patients who also underwent simultaneous scanning with 99m Tc-labeled red blood cells showed blood loss at sites identical to where intestinal inflammation was demonstrated. Loss of mucosal integrity allows luminal contents, including bile acids, pancreatic secretions, bacteria, and food antigens, to enter the mucosa. This process results in neutrophil chemotaxis with nonspecific inflammation and ulceration as a systemic response to the initial injury. Although the role of COX-1 and COX2 has not been clearly defined in NSAID-induced injury to the small intestine, it is clear that COX-2 inhibitors cause less deleterious clinical effects on the small intestine than do COX-1 inhibitors. Using capsule endoscopy, Goldstein and colleagues found significantly less small intestinal injury in healthy subjects taking a selective COX-2 inhibitor (celecoxib) than in those who were taking ibuprofen plus omeprazole.28

Diagnosis

Secondary damage with ulceration and strictures Blood loss Protein loss Figure 115-4.  Hypothetical sequence of events involved in the pathogenesis of nonsteroidal anti-inflammatory drug (NSAID) enteropathy. (From Aabakken L. Small-bowel side-effects of non-steroidal anti-inflammatory drugs. Eur J Gastroenterol Hepatol 1999; 11(4):383-388. Copyright 1999, Lippincott Williams & Wilkins. Used with permission.)

mucosa, and this may be effected by the enterohepatic circulation of NSAIDs. The initial injury progresses to mucosal barrier dysfunction. Indeed, Bjarnason demonstrated that NSAIDs cause increased intestinal permeability in humans by showing loss of chromium-51–labeled proteins into the intestine.26 In another study, 33 of 49 (67%) patients taking oral NSAIDs were found to have intestinal inflammation,

Despite the use of esophagogastroduodenoscopy (EGD), colonoscopy, and barium contrast studies, no source of blood loss is found in one half of patients with iron deficiency anemia who are taking NSAIDs. Capsule and double balloon enteroscopy, however, have significantly changed the diagnosis and management of small intestine ulceration and diaphragm disease (see Fig. 115-3).23,29 Tibble and associates demonstrated that fecal excretion of calprotectin, a nondegraded neutrophil cytosolic protein, can be used to assess intestinal inflammation, and therefore that this test might be practical to diagnose NSAID enteropathy. Fecal calprotectin levels correlate with fecal excretion of indium111.30 Maiden showed that after two weeks of diclofenac, 75% of volunteers had increased fecal calprotectin levels and 68% showed injury on capsule enteroscopy.23 Thus, NSAIDs can cause injury and ulceration throughout the gastrointestinal tract. This observation has led to the recommendation that NSAIDs be discontinued before extensive evaluation of patients with obscure gastrointestinal bleeding because of the likelihood that these agents are responsible for the blood loss.

2053

2054

Section X  Small and Large Intestine Treatment

Although avoidance of NSAIDs is the most effective therapy for NSAID enteropathy, experimental studies show that metronidazole reduces inflammation and occult blood loss without changing intestinal permeability in patients with NSAID enteropathy.31 Sulfasalazine also has been shown to reduce intestinal inflammation, as measured by fecal indium-labeled neutrophil excretion, suggesting that active mediators of inflammation as well as anaerobic organisms normally found in the small intestine might play a role in the pathogenesis of NSAID enteropathy.32 Recommendations on the use of such antibiotics to treat patients with NSAID-induced enteropathy have not been formalized.

DIFFUSE ULCERATIONS DEFINITIONS

Diffuse small intestinal ulcerations can complicate celiac disease or can occur in patients not known to have celiac disease but who present with sprue-like symptoms, have either flat or normal intervening intestinal mucosa, and are refractory to gluten withdrawal. These syndromes have been variously termed ulcerative jejunitis, chronic ulcerative (nongranulomatous) jejunoileitis, or idiopathic chronic ulcerative enteritis. The term ulcerative enteritis is used in this discussion. Refractory celiac disease itself encompasses a heterogeneous group of patients and is discussed more fully in Chapter 104. For the purposes of this discussion, refractory celiac disease refers to patients with small intestinal histology or antibodies consistent with celiac disease; severe, persistent malabsorption, often with diffuse small intestinal ulcerations (ulcerative enteritis) despite strict adherence to a gluten-free diet for longer than one year, or when severe symptoms require intervention independent of the duration of a gluten-free diet; and no overt lymphoma. Other causes of villus atrophy, malabsorption, and diarrhea must be excluded.33,34 Two types of refractory celiac disease (RCD) are recognized: RCD I, where there is a normal expression of T-cell lymphocyte surface markers, and RCD II, where the T cells are aberrant. The T cells in RCD II are immunophenotypically abnormal but cytologically normal and do not form tumor masses.35 Continued clonal expansion of the abnormal T-cell population in RCD II eventually leads to enteropathy-associated T-cell lymphoma (EATL). EATL can complicate established celiac disease or can manifest de novo with multiple intestinal ulcerations and malabsorption in patients without previously known underlying celiac disease but with small intestinal biopsies that demonstrate villus atrophy. EATL is recognized by the World Health Organization International Classification Project as a specific disorder and is discussed in greater depth in Chapter 29.

BACKGROUND

The association of malabsorption and lymphoma was reported first in 1937. In the 1980s, evidence in favor of underlying celiac disease as the predisposing factor for ulcerative enteritis, refractory celiac disease, and EATL in many patients emerged from registries, genotyping studies, and literature reviews. As technologic advances occurred, molecular biology studies, polymerase chain reaction (PCR) amplification, immunophenotypical analysis, and other methods were used to link celiac disease, refractory celiac disease, ulcerative enteritis, and EATL via intraepithelial T-cell lymphocyte abnormalities.36,37

The repertoire of intraepithelial and lamina propria lymphocytes is polyclonal before and shortly after birth. Because of interactions with the intestinal microenvironment, including food antigens and microbes, the repertoire becomes oligoclonal by the time adulthood is reached.38 In RCD I, the intraepithelial T-cell phenotype is similar to that of uncomplicated celiac disease: polyclonal with normal expression of cell surface markers. Less than 10% of the T cells are aberrant in intestinal biopsy specimens of RCD I, and when more than 20% of the T cells become immunophenotypically abnormal, RCD I is said to have progressed to RCD II.39 The expansion and hyperplasia of T cells in active celiac disease and RCD is at least partially regulated by interleukin-15 (IL-15). In active celiac disease, IL-15 synthesis is up-regulated in the intestinal mucosa. Similarly in RCD, IL-15 is markedly overexpressed in intestinal epithelial cells and in the lamina propria, inducing clonal expansion of intraepithelial lymphocytes and initiating T-cell cytotoxicity against intestinal epithelial cells.40 IL-15 has been shown to deliver signals that impair SMAD3–dependent transforming growth factor-β (TGF-β) signaling in T lymphocytes, which leads to the promotion and continuation of intestinal inflammation.41 Cytogenetic studies have demonstrated that the intraepithelial T lymphocytes in RCD II possess a partial trisomy of 1q22q44.42 This gain on chromosome 1q, also shown to be present in 16% of cases of EATL, may be an early event on the way to malignant transformation of T cells in RCD. EATL tumor cells have chromosomal imbalances in 87% of cases, 58% showing gains on chromosome 9q and 16% on chromosome 1q; recurrent genetic losses occur on chromosomes 8p, 13q, and 9p. Chromosomes 9q and 8p are of particular importance because 9q has a gene implicated in lymphomagenesis, and 8p has a gene implicated in apoptosis. With gains and losses of important genetic material, increased production and decreased apoptosis of intraepithelial T cells might contribute to the accumulation in RCD and subsequent development of EATL.43 RCD II is said to have developed once more than 20% of the T cells are aberrant. The intraepithelial lymphocytes in RCD II are phenotypically abnormal but not yet cytologically abnormal. Intestinal ulcers often are present, especially in the jejunum, defining the clinical syndrome of ulcerative enteritis, but no tumor masses are seen. RCD II with or without ulcerative enteritis is considered a cryptic lymphoma and represents a continuum with EATL. Monoclonal expansion continues, and once tumor masses are present or there is radiologic or bone marrow evidence of cytologically abnormal T cells, EATL is said to have developed.33 Celiac disease is associated with the HLA-DQ2 and HLA-DQ8 alleles, and both RCD II and EATL have been associated with HLA-DQ2 homogyzosity.44 Genetic variants in the MYO9B gene on chromosome 19 also have been linked to celiac disease, and genetic studies have reported a single nucleotide polymorphism, rs7259292 T allele in the MYO9B gene, to be present more often in patients with RCD II and EATL.45 Therefore, chronic gluten antigenic exposure in patients with an underlying genetic predisposition for celiac disease induces clonal expansion of a subset of intraepithelial T cells. Amidst ongoing inappropriate immunologic stimulation, the loss of cell surface markers and the acquisition of abnormal genetic material all lead to complicated celiac disease, including RCD with or without intestinal ulcerations, and finally, EATL.

Chapter 115  Ulcers of the Small and Large Intestine CLINICAL FEATURES

The age of patients with ulcerative enteritis, RCD, and EATL ranges from 18 to 80 years; most patients present in their forties or later. RCD II and EATL patients present 10 years later than do RCD I patients. Women are affected slightly more commonly than are men in a ratio of 1.6-2.0 : 1.0. Patients with either new-onset celiac disease or with longstanding, proven celiac disease can present with worsening malabsorption and abdominal pain that are increasingly unresponsive to a gluten-free diet. Other patients present with unexplained malabsorption and abdominal pain and have either a flat or otherwise normal appearing small intestinal biopsy, but never respond to a gluten-free diet. Malabsorption, diarrhea, and weight loss are conspicuous and often severe, and they may be present for years. Alternatively, some patients present more acutely with complications, including gastrointestinal hemorrhage, perforation, or intestinal obstruction secondary to stricture formation. Physical examination reveals profound weight loss, cachexia and signs of severe malabsorption, steatorrhea, and protein-losing enteropathy. Abdominal tenderness may be mild, diffuse, or severe. Hepatomegaly and splenic atrophy may be present. Peripheral lymphadenopathy is unusual, but patients who have developed lymphoma can have an abdominal mass. Dermatitis herpetiformis, a condition usually associated with celiac disease, rarely may be observed with RCD. Signs and symptoms of anemia can occur as a result of chronic nutritional deficiency or acute gastrointestinal hemorrhage. Intestinal perforation typically leads to signs of peritonitis. Patients with intes­ tinal obstruction might have acute vomiting and abdominal distention.35,37,44,46

DIAGNOSIS Laboratory

Laboratory abnormalities reflect the diarrhea, severe malabsorption, and complications of the disease. Findings include iron deficiency or macrocytic anemia, prolongation of the prothrombin time, electrolyte abnormalities consistent with the degree of dehydration, hypoalbuminemia, hypocalcemia, hypomagnesemia, hypocholesterolemia, and low serum carotene levels. Stool abnormalities include increased volume, mild to severe steatorrhea, increased fecal α1antitrypsin excretion, and a positive fecal occult blood test. The d-xylose test usually is abnormal.33,34,37 Except for de novo EATL, where serology typically is negative at the time of diagnosis, serum antibodies for antigliadin and tissue transglutaminase (tTG) usually are present in ulcerative enteritis, RCD, or EATL. A positive serologic assay for tTG antibodies reverts to negative on a gluten-free diet, similar to the typical serologic pattern of uncomplicated celiac disease.33,35 Although not necessary for diagnosis, HLA-DQ2 or DQ8 genes are found in most patients with one of these clinical syndromes. HLA-DQ2 homozygosity has been linked to an increased risk of developing RCD II and EATL.44,45

Radiology

Radiologic abnormalities are common in these clinical entities. Computed tomography (CT) scanning can demonstrate characteristic findings of celiac disease, such as decrease or loss of jejunal folds, thickening and separation of small bowel loops, increased ileal folds or reversal of the jejunoileal fold pattern, small bowel dilatation and stricturing, intussusception, increased number of small mesenteric vessels, enlarged or cavitated mesenteric lymph nodes, hyposplenism, or metastatic disease. Splenomegaly tends to

Figure 115-5.  Film from a barium study of the small intestine showing diffuse small intestinal ulceration in a patient with refractory celiac disease (ulcerative enteritis). There is diffuse involvement of the small intestine, with multiple ulcerations and separation and thickening of the loops of jejunum and ileum. (Courtesy of Christophe Cellier, MD, PhD, Paris, France.)

occur in celiac disease and RCD I.47 CT enterography additionally may show jejunal or ileal wall thickening with mucosal hyperenhancement. The tracer fluorine-18 fluorodeoxyglucose (18F-FDG), a glucose analog taken up by rapidly growing tumor cells and used during positron emission tomography (18F-FDG PET), was shown to be more sensitive and specific than CT in detecting EATL in patients with RCD.48 Barium contrast study of the small bowel or enteroclysis can show jejunization of the ileum, separation and thickening of small intestinal loops, intussusceptions, ulcerations, or a mass (Fig. 115-5). Strictures may be single or multiple and appear as areas of luminal narrowing alternating with more dilated portions of small bowel. Abnormalities tend to be more noticeable in the proximal small intestine.37

Endoscopy

Because the ulcers usually are located distal to the duodenum, direct visualization of the jejunum and ileum is necessary and can best be accomplished via videocapsule endoscopy or double-balloon enteroscopy. The major advantages of videocapsule endoscopy are ease of use and visualization of the entire small intestine in 80% to 90% of patients during the eight-hour acquisition time. Highquality videocapsule endoscopy images can demonstrate villus atrophy, loss of folds, scalloping, whitish villi, and small intestinal erosions, ulcerations, nodularity, masses, or strictures (Fig. 115-6). The capsule needs to be used cautiously in patients with known or suspected strictures.49 The major advantage of double-balloon enteroscopy is examination of the entire small intestine with the ability to take biopsies. Double-balloon enteroscopy can detect or exclude RCD I and RCD II, with or without ulcers (ulcerative enteritis), and EATL. Findings are similar to those seen during videocapsule endoscopy.50 During enteroscopy,

2055

2056

Section X  Small and Large Intestine

A

B

Figure 115-6.  A, Video capsule endoscopy photograph of a patient with refractory celiac disease and multiple erosions (ulcerative enteritis) in the duodenum and jejunum. The capsule image shows villus atrophy and an erosion in the mid-jejunum. (Courtesy of Jeffrey P. Baker, MD, Toronto, Ontario, Canada.) B, Video capsule endoscopy photograph of another patient with refractory celiac disease and multiple ulcerations (ulcerative enteritis) in the small intestine. The image shows an ulceration with exudate, located in the ileum. Duodenal biopsies revealed a clonal T-cell population. (Courtesy of Moshe Rubin, MD, and Peter Green, MD, New York.)

biopsies should be taken of abnormal as well as endoscopically normal-appearing intervening mucosa. Occasional patients with ulcerative enteritis, RCD, or EATL also have gastric or colonic ulcers that can be identified on routine endoscopy. Biopsies of normal-appearing gastric and colonic mucosa should be obtained, because abnormal lymphocytes have been demonstrated in these disorders throughout the gastrointestinal tract.51

Pathology

Ulcerative enteritis is more common in RCD II than in RCD I. The ulcers in ulcerative enteritis are diffuse, more commonly located in the jejunum and ileum than in the duodenum, range in size from 1 mm to 3.5 cm, are rarely solitary, and are well circumscribed. Some of the ulcers are superficial, extending only to the muscularis mucosa, but usually they extend to the muscularis propria and occasionally through it, thus causing perforation.35,36 Histologically, they are benign. The intraepithelial T lymphocytes have a widespread distribution and may be present in intestine that appears normal or abnormal, ulcerated or nonulcerated. T cells in uncomplicated celiac disease tend to be CD3+, CD4−, and CD8+. In RCD I, the T-cell phenotype is similar, with a polyclonal immunophenotypic expression of CD3 and CD8 and a normal expression of the T-cell receptor (TCR). Less than 10% of the T cells are abnormal in RCD I. Once more than 20% of the intraepithelial T lymphocytes in biopsy specimens are immunophenotypically abnormal, as determined by flow cytometry quantification, RCD II is said to have developed (median 52%, range 27% to 94%).39 In contrast to RCD I, the aberrant T cells in the majority of RCD II patients express intracytoplasmic CD3 and surface CD103, but they lack cell surface expression of the TCR complex, CD3, CD4, and CD8; rarely, the abnormal T cells are surface CD3− and CD8+.39 PCR analysis of the T-cell receptor γ-chain gene further demonstrates monoclonality of the T lymphocytes consistent with RCD II being a cryptic T-cell lymphoma. The T cells in RCD II appear cytologically normal.33,51,52 Once the abnormal intraepithelial clone has been identified, the risk of progression to

EATL is very high and should warrant a thorough investigation for EATL, including endoscopic and radiologic studies52 (Fig. 115-7). When cytologic abnormalities appear, RCD II has progressed onto EATL. Overt T-cell lymphoma can manifest as multiple ulcers, as an ulcerated mass, or as extraintestinal disseminated disease. Tumor cells can be demonstrated in both ulcerated and nonulcerated areas of the intestine. In EATL, the majority of abnormal T lymphocytes are large-tomedium-sized T cells; less commonly they are small and monomorphic. Cell markers are similar to those seen in RCD II.33,43,53 Normal-appearing intestinal mucosa adjacent to or distant from either a benign- or malignant-appearing ulcer or mass can show normal villus architecture or partial or total villus atrophy with crypt hyperplasia and an increased number of intraepithelial lymphocytes with immunophenotypic characteristics as described earlier (Fig. 115-8). Other causes of villus atrophy such as tropical sprue, giardiasis, common variable immunodeficiency syndrome, or autoimmune enteropathy must be excluded.33 Features of inflammatory bowel disease are absent, and there is no evidence of an infectious process. Histology of endoscopically normal gastric and colonic mucosa in patients with RCD II demonstrates widespread distribution of abnormal lymphocytes throughout the gastrointestinal tract; associated lymphocytic gastritis and lymphocytic and collagenous colitis have been reported. In patients with RCD II and concurrent lymphocytic gastritis or lymphocytic colitis, gastric and colonic T cells are monoclonal, immunophenotypically abnormal, and identical to T cells in the duodenum and jejunum. T cells in gastric and colonic biopsies in patients with RCD II without lymphocytic gastritis or colitis tend to have a polyclonal gene rearrangement.51

TREATMENT AND PROGNOSIS

With the recognition that RCD I and II with or without ulcerative enteritis are a continuum on the spectrum of developing aberrancy in T cells that eventually leads to EATL if unchecked, recent therapy has targeted the abnor-

Chapter 115  Ulcers of the Small and Large Intestine Sprue-like symptoms Prior history of celiac disease Yes

No Endoscopy - villus atrophy ± ulcers Proof of underlying celiac disease Antibodies, HLA typing Exclude other causes of villus atrophy ± ulcers Ensure GFD

Ensure GFD

Response to GFD Yes No (persistent symptoms)

Continue GFD

EGD and colonoscopy with biopsies Exclude concomitant disorder Present

Absent

Treat disorder

VCE/DBE Villus atrophy ± ulcers

RCD II ± UE: T lymphocytes >20% monoclonal, phenotypically abnormal, cytologically normal Further investigations to exclude EATL: CT chest/abd/pelvis/CT enterography PET/Bone marrow examination

RCD I ± UE: T lymphocytes >90% polyclonal, phenotypically normal, cytologically normal Consider TPN Glucocorticoids/budesonide ± azathioprine Consider infliximab if no response to immunosuppressant therapy or lifethreatening RCD I is present

EATL absent (RCD II only)

EATL present (cytological abnormalities present, bone marrow or other metastases present)

Consider: TPN 2–CDA Chemotherapy/ASCT

Consider: TPN Chemotherapy/ASCT

mal T-cell population at an earlier stage with better outcomes over the past five years.35 All patients initially should be treated with fluid and electrolyte replacement. Antibiotics should be used if infectious complications or sepsis occur. Total parenteral nutrition (TPN) can improve nutritional status in severely malnourished patients. If small intestinal biopsies show villus atrophy and there is no evidence of EATL, patients should be placed on a strict gluten-free diet with the assistance of a nutritionist to ensure adherence. After gluten withdrawal failure in patients with RCD I with or without ulcerative enteritis that is not life threatening, glucocorticoids achieve a response in 50% to 75% of patients.54 In patients who do not initially respond to glucocorticoids or who become dependent on them, the addition of azathioprine usually is effective therapy to induce or maintain a response35,55 (see Fig. 115-7). Case reports using infliximab in patients with RCD I and life-threatening symptoms have shown improvement in symptoms, with duodenal histology reverting to near normal.56,57

Figure 115-7.  Algorithm for the diagnosis and management of ulcerative enteritis, refractory celiac disease, and enteropathy-associated T-cell lymphoma. abd, abdomen; ASCT, autologous stem cell transplant; 2-CDA, cladribine; CT, computed tomography; DBE, enteroscopy, preferably double-balloon enteroscopy; EATL, enteropathyassociated T-cell lymphoma; EGD, esophago­ gastroduodenoscopy; GFD, gluten-free diet; HLA, human leukocyte antigen; PET, positron emission tomography; r/o, rule out; RCD, refractory celiac disease (see text for differences between RCD I and II); TPN, total parenteral nutrition; UE, ulcerative enteritis; VCE, video capsule endoscopy.

Once aberrant monoclonal T cells are present, RCD II is considered to have developed. Therapy should be more aggressive and directed against the abnormal T-cell popu­lation. Cladribine, 2-chlorodeoxyadenosine (2-CDA), a cytotoxic chemotherapeutic agent with activity against lymphocytes, was used in an open-label, prospective trial in 17 patients with RCD II. Six patients had a clinical improvement, 10 patients (five with ulcerative enteritis) had histologic improvement, and six patients had a significant decrease in aberrant T cells. Seven patients developed EATL and subsequently died. This medication showed promise in stabilizing patients but did not prevent the development of EATL. There also is the additional concern of accelerating the development of EATL or inducing a secondary malignancy.35,58 After conditioning with the chemotherapeutic agents fludarabine and melphalan, seven patients with RCD II underwent autologous stem cell transplant (ASCT) in an open-label, prospective study. After a mean follow-up of 15.5 months, six of the seven patients showed significant improvement in their symptoms along with a reduction in

2057

2058

Section X  Small and Large Intestine

A

Elective surgical resection usually is neither indicated nor possible, especially in a severely malnourished patient. In selected EATL patients with localized involvement, however, surgical resection followed by chemotherapy or ASCT might offer a reasonable chance for survival.35,61 In the future, novel therapeutic options likely will include allogeneic stem cell transplantation, better preconditioning regimens, and biologic agents directed against proinflammatory mediators, such as interleukin-1540 or T-cell markers such as CD52.62 Given the rarity of these diseases and the lack of prospective trials, the treatment of RCD I and RCD II patients with or without ulcerative enteritis and EATL should be undertaken at medical centers with knowledge and expertise in the treatment of these serious complications of celiac disease.

KEY REFERENCES

B Figure 115-8.  Full-thickness biopsy of the jejunum in a patient with longstanding, severe refractory celiac disease. A, Low-power view. The mucosal architecture shows complete villus flattening with crypt hyperplasia (severe mucosal lesion). There is an increase in the number of intraepithelial lymphocytes. The lamina propria is expanded by inflammatory cells that extend beneath the base of the crypts. B, High-power view. Neutrophils are destroying an epithelial crypt. In other areas (not shown), small ulcers were present. (Hematoxylin and eosin.) (Courtesy of Marie E. Robert, MD, New Haven, Conn.)

the aberrant T cells in duodenal biopsies. This pilot study suggests that treating RCD II patients with ASCT might result in long-term improvement.46 Once EATL has developed, prognosis is poor. Options are limited to chemotherapy with or without bone marrow transplantation. Chemotherapy is often poorly tolerated secondary to poor nutritional status and complications during treatment. Aggressive chemotherapy followed by ASCT in patients with early-stage disease resulted in better survival59,60 than did treatment of patients with later stage disease.61 Due to widespread T-cell involvement of the gastrointestinal tract in RCD II and EATL, surgical options are limited. Exploratory laparotomy may be necessary if a complication, such as perforation, obstruction, or hemorrhage occurs.

Abdallah H, Leffler D, Dennis M, Kelly CP. Refractory celiac disease. Curr Gastroenterol Rep 2007; 9:401-5. (Ref 34.) Al-Toma A, Verbeek WH, Hadithi M, et al. Survival in refractory coeliac disease and enteropathy-associated T-cell lymphoma: Retrospective evaluation of single-centre experience. Gut 2007; 56:1373-8. (Ref 35.) Al-toma A, Visser OJ, van Roessel HM, et al. Autologous hematopoietic stem cell transplantation in refractory celiac disease with aberrant T cells. Blood 2007; 109:2243-9. (Ref 46.) Biagi F, Lorenzini P, Corazza GR. Literature review on the clinical relationship between ulcerative jejunoileitis, coeliac disease, and enteropathy-associated T-cell. Scand J Gastroenterol 2000; 35: 785-90. (Ref 36.) Bishton MJ, Haynes AP. Combination chemotherapy followed by autologous stem cell transplant for enteropathy-associated T cell lymphoma. Br J Haematol 2007; 136:111-13. (Ref 59.) Boydstun JS Jr, Gaffey TA, Bartholomew LG. Clinicopathologic study of nonspecific ulcers of the small intestine. Dig Dis Sci 1981; 26: 911-16. (Ref 1.) Daum S, Cellier C, Mulder CJ. Refractory coeliac disease. Best Pract Res Clin Gastroenterol 2005; 19:413-24. (Ref 33.) Daum S, Wahnschaffe U, Glasenapp R, et al. Capsule endoscopy in refractory celiac disease. Endoscopy 2007; 39:455-8. (Ref 49.) Fortun PJ, Hawkey CJ. Nonsteroidal antiinflammatory drugs and the small intestine. Curr Opin Gastroenterol 2007; 23:134-41. (Ref 21.) Gekas P, Schuster MM. Stercoral perforation of the colon: Case report and review of the literature. Gastroenterology 1981; 80:1054-8. (Ref 17.) Hadithi M, Al-toma A, Oudejans J, et al. The value of double-balloon enteroscopy in patients with refractory celiac disease. Am J Gastroenterol 2007; 102:987-96. (Ref 50.) Laine L, Connors LG, Reicin A, et al. Serious lower gastrointestinal clinical events with nonselective NSAID or coxib use. Gastroenterology 2003; 124:288-92. (Ref 20.) Maiden L, Thjodleifsson B, Theodors A, et al. A quantitative analysis of NSAID-induced small bowel pathology by capsule enteroscopy. Gastroenterology 2005; 128:1172-8. (Ref 23.) Sharara AI, Azar C, Amr SS, et al. Solitary rectal ulcer syndrome: Endoscopic spectrum and review of the literature. Gastrointest Endosc 2005; 62:755-62. (Ref 11.) Tjandra JJ, Fazio VW, Church JM, et al. Clinical conundrum of solitary rectal ulcer. Dis Colon Rectum 1992; 35:227-34. (Ref 2.) Whittle BJ. Mechanisms underlying intestinal injury induced by anti-inflammatory COX inhibitors. Eur J Pharmacol 2004; 500:427-39. (Ref 24.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

116 Appendicitis Kfir Ben-David and George A. Sarosi, Jr.

CHAPTER OUTLINE Historical Perspective  2059 Epidemiology  2059 Anatomy and Embryology  2060 Pathology  2060 Pathogenesis  2060 Clinical Features  2061 Diagnosis  2062 Laboratory Studies  2062 Imaging Studies  2062 Clinical Scoring Systems and Computer-Aided Diagnosis  2065 Laparoscopy  2066 Diagnostic Accuracy  2066

HISTORICAL PERSPECTIVE The first anatomic mention of the appendix was made by Leonardo da Vinci in the early 15th century. The first clearly recognizable case report of appendicitis was recorded in 1711 by the German surgeon Lorenz Heister,1 but it was not until 25 years later that the first inflamed appendix was removed by Claudius Amyand, a Sergeant Surgeon to Queen Ann, King George I, and King George II. Amyand operated on an 11-year-old boy with a perforated appendix within a scrotal hernia that he was able to excise and repair, respectively.2 Throughout the 18th and 19th centuries, the prevailing medical opinion was that acute abdominal pain and right lower quadrant inflammation was a consequence of inflammation of the cecum or its surrounding tissues. The modern description of the pathophysiology of appendicitis and the role of the appendix in acute abdominal syndromes dates to 1886, the year Reginald Fitz presented a paper to the Massachusetts Medical Society in which he coined the term appendicitis and espoused early surgical intervention as its appropriate treatment.1 The first now-customary appendectomy for classic acute appendicitis actually had been performed by Lawson Tait in 1880, but it was not reported until Charles McBurney, one of the great contributors to our understanding of appendicitis, published his recommendation for early laparotomy for the treatment of appendicitis in 1889.4 It is in this paper that what subsequently became known as McBurney’s point is described as the point of “maximum tenderness, one half to two inches inside the right anterior spinous process of the ilium on a line drawn from the umbilicus.”2,3

Complications  2067 Treatment  2068 Outcomes  2069 Special Topics  2069 The Appendix and Ulcerative Colitis  2069 Crohn’s Disease of the Appendix  2069 Recurrent and Chronic Appendicitis  2069 Diverticulitis of the Appendix  2070 Epithelial Malignancies of the Appendix  2070 Incidental or Prophylactic Appendectomy  2070

Almost a century later, the first laparoscopic approach to appendectomy was described by Kurt Semm,4 and with development of natural orifice transluminal endoscopic surgery (NOTES), the first successful transvaginal appendectomy was reported by Sanntiago Horgan and Mark A. Talamini in early 2009.5

EPIDEMIOLOGY Appendicitis is the most common acute abdominal emergency seen in developed countries. The crude incidence rate of appendicitis in the United States for all age groups is 11/10,000 persons per year,6 and similar rates are noted in other developed countries. Inexplicably, the rates of appendicitis are as much as 10 times lower in many less-developed African countries.7 The incidence rate of the disease peaks between 15 and 19 years of age at 48.1/10,000 population per year and falls to about 5/10,000 population per year by age 45 years, after which it remains constant.6 Men are at greater risk than women, with a case ratio in most series of 1.4 : 1. The lifetime risk of appendicitis has been estimated at 8.6% in men and 6.7% in women.6 Approximately 250,000 appendectomies are performed each year in the United States; data from most developed countries suggest that the incidence of appendicitis is decreasing. Between 1989 and 2000, a 15% decrease in the overall incidence of appendicitis was noted in an English study8; similar trends have been noted in Canada and Greece.9,10 One paper, however, suggests that, at least in the United States, the number of appendectomies per-

2059

2060

Section X  Small and Large Intestine formed for acute appendicitis has been increasing since 1995.11 Regardless of the direction of the epidemiologic trend, appendicitis remains the most common indication for emergency abdominal surgery.

ANATOMY AND EMBRYOLOGY The vermiform appendix and the cecum are best thought of as a single anatomic unit. Developmentally part of the midgut, the appendix and cecum form between the 8th and 12th weeks of gestation as a bud arising from the midgut loop, before the ascending colon has become delineated. Congenital malformations of the appendix such as agenesis and duplication are very rare. With an average length of 9 cm,12 the origin of the appendix varies and the appendix may assume any of the positions of a clock hand, with the center considered the appendiceal origin. Unlike the rest of the colon, the appendix has a complete longitudinal muscle layer. The blood supply of the appendix is found in a separate mesentery, the mesoappendix, and consists of an appendicular branch of the ileocolic branch of the superior mesenteric artery. The lymphatic drainage of the appendix occurs via the ileocolic lymph nodes, which are shared with the terminal ileum and right colon. Although the right colon is fixed in the retroperitoneum, the appendix and cecum have a more variable position within the abdomen. The position of the appendix depends upon a number of factors: the degree of cecal descent and peritoneal fixation, the configuration of the cecum, appendiceal length, associated adhesions, and the habitus of the person.13 Typically, the location of the appendix is described as retrocecal, pelvic, subcecal, or para-ileal (Fig. 116-1). The position of the appendix has important clinical implications: For example, the classic progression of symptoms requires irritation of the parietal peritoneum by a mobile appendix; as many as 60% of people have a retrocecal or pelvic-positioned appendix, resulting in an atypical clinical presentation of acute appendicitis. The classic surface anatomy of the appendix was described by McBurney in 1889, and as mentioned earlier, McBurney’s point is located at the junction of the lateral and middle thirds of a line drawn from the right anterior superior iliac spine to the umbilicus.4 Classically, this surface marking has been important in both the diagnosis and treatment of acute appendicitis; however, investigators have shown that the appendix is located within 5 cm of McBurney’s point in less than 50% of the cases.14 This anatomic finding helps explain why pain or tenderness located at McBurney’s point is not found in all cases of appendicitis.

PATHOLOGY Acute appendicitis is classified as acute, gangrenous, or perforated. The earliest gross findings of acute appendi-

Posterior

Retrocecal

Anterior

Postileal Preileal

Subcecal

Pelvic

Figure 116-1.  Positions of the appendix. Five different positions of the appendix are illustrated; variations in position can affect the clinical presentation of the appendix (see text). (From Buschard K, Kjaeldgaard A: Investigation and analysis of the position, fixation, length, and embryology of the vermiform appendix. Acta Chir Scand 1973; 139:293.)

citis are injection of the serosal blood vessels and edema of the appendiceal wall. In more-advanced cases, the serosal surface appears dull and is covered by fibrinopurulent exudates. Over time, focal areas of gangrene develop, marked by greenish and black discoloration of the wall; with perforation, focal necrosis of the appendiceal wall develops and adjacent abscesses form.15 Microscopically, each of these forms of appendicitis has distinctive characteristics. In acute or suppurative appendicitis, a neutrophilic infiltrate involves the muscularis propria layer circumferentially, accompanied by acute inflammation and ulceration of the mucosa, edema and microabscesses in the appendicular wall, and vascular thrombosis. The hallmarks of gangrenous appendicitis are transmural inflammation of the appendix with focal areas of mural necrosis. Vascular thrombosis is more prominent in gangrenous than in suppurative appendicitis. The presence of mucosal inflammation alone (“catarrhal” inflammation) is more characteristic of infectious enteritis or colitis and is not considered evidence of acute appen­dicitis; for the microscopic diagnosis of appendicitis to be made, inflammation must extend to the muscularis propria.15

PATHOGENESIS Despite more than 100 years of study, there still is no single explanation for all cases of appendicitis. The

Chapter 116  Appendicitis classic hypothesis is that obstruction of the appendiceal lumen by either a fecalith or lymphoid hyperplasia produces an increase in intraluminal pressure, which in turn results in venous hypertension, ischemia of the appendiceal wall, and subsequent bacterial invasion of the appendix with necrosis and perforation. Experimental evidence in animal models exists to support this hypothesis of the etiology of acute appendicitis.16 This hypothesis, however, does not explain all cases of appendicitis. Careful review of pathologic series shows that luminal obstruction is found in a minority of cases. Fecaliths are present in only 8% to 44% of cases of acute appendicitis, with most series at the lower end of the range,15,17 and lymphoid hyperplasia is more common in noninflamed appendices than in acute appendicitis.18 Other causes of luminal obstruction such as foreign bodies, tumors, and fibrous bands are uncommon. Direct measurement of intraluminal pressure at appendectomy for appendicitis reveals an elevated pressure in only a minority of cases.19 An alternative hypothesis for the etiology of appendicitis is based on the concept that either bacterial or viral enteric infection leads to mucosal ulceration of the appendix and subsequent bacterial invasion from the normal colonic flora. The finding that up to 75% of cases of appendicitis demonstrate well-defined superficial mucosal ulceration supports this theory. Furthermore, mucosal ulceration is a more consistent finding than dilatation of the appendix or fecaliths and is found earlier in the course of appendicitis.20 One report found human cytomegalovirus (HCMV) early antigen expression in 64% of cases with acute appendicitis and no HCMV antigens in normal appendices, suggesting that in some cases CMV, infection might produce mucosal ulcerations leading to acute appendicitis.21 Additional support for the role of infection in the etiology of appendicitis is found in two lines of epidemiologic evidence. The first is based in the hygiene theory of appen­dicitis advocated by Barker in the mid 1980s.22 According to this hypothesis, changes in sanitation tied to the Industrial Revolution resulted in a decrease in enteric infections in infants, with subsequent decreased immunity to the same infections in childhood and young adulthood. Acquisition of these infections later in life was believed to predispose people to appendicitis, explaining the rise in the incidence rates of appendicitis in the first half of the 20th century. The decrease in the overall rate of enteric infections during the last half of the 20th century explains the overall decline in appendicitis. The second line of epidemiologic evidence supporting the role of infection in the etiology of appendicitis is the seasonal variance in incidence of appendicitis and the occurrence of temporal and spatial clusters of appendicitis, both hallmarks of infectious diseases.23 It is important to recognize, however, that no specific infectious agent has been linked with all cases of appendicitis, suggesting that infection is not the complete story. A decrease in dietary fiber intake (the fiber hypothesis) also has been proposed as a cause of appendicitis.

According to this hypothesis, decreased dietary fiber causes firm stool and an increased enteric transit time, resulting in more fecaliths and more appendicitis. This hypothesis was felt to explain both the rise in appendicitis rates in the early 20th century and the marked differences in appendicitis rates between more-developed Western countries and less-developed African countries. Doubt has been cast upon this hypothesis, however, for several reasons. First, although dietary fiber ingestion has been falling in urban Africans, appendicitis rates have not risen markedly.24 Second, rates of appendicitis in the Western world have fallen without changes in dietary fiber intake. Finally, a prospective series from Africa demonstrated continued high fiber intake even in patients with appendicitis.25 It is likely that any one of several different inciting events such as luminal obstruction, infection, or trauma can initiate breakdown of the appendiceal mucosa, resulting in bacterial invasion; the end result is appendicitis.

CLINICAL FEATURES A detailed history and careful physical examination remain cornerstones of the diagnosis of acute appendicitis. Although no single item of the history, in isolation, allows the diagnosis to be made reliably, combination of the classic symptoms and the typical progression of symptoms coupled with right lower quadrant tenderness allows good diagnostic accuracy. In the classic presentation of acute appendicitis, patients first note vague, poorly localized epigastric or periumbilical discomfort, which typically is not severe and often is attributed to “gastric upset.” Patients commonly report feeling that a bowel movement should make the pain better, a sensation known as the downward urge.26 Diarrhea sometimes is seen early on with appendicitis, but this is not common. Within 4 to 12 hours of the onset of pain, most patients also note nausea, anorexia, vomiting, or some combination of these three symptoms. The nausea usually is mild to moderate, and most patients have only a few episodes of emesis. If vomiting is the major symptom, the diagnosis of appendicitis should be questioned. Likewise, emesis that occurs before the onset of pain should suggest other diagnoses.27 Many patients report a mild fever or chills; high fevers or significant rigors are uncommon. The patient’s abdominal pain typically increases in intensity, and a characteristic shift in the pain to the right lower quadrant occurs over 12 to 24 hours. The character of the pain becomes achy and more localized. Localization of the pain to the right lower quadrant is a valuable finding when present and occurs in more than 80% of patients with appendicitis.27 On physical examination, most patients appear slightly ill. Tachycardia is uncommon with simple appendicitis, but it may be seen with complicated appendicitis. Most patients with simple appendicitis have a temperature less than 100.5°F; temperature

2061

2062

Section X  Small and Large Intestine greater than 100.5°F is most often associated with perforated or gangrenous appendicitis.17 Patients with appendicitis, like other patients with peritonitis, tend to lie still rather than move about. Right lower quadrant tenderness and rigidity, both voluntary and involuntary, are common findings. Localized right lower quadrant tenderness is an important finding when present, but its absence does not rule out appendicitis. A variety of methods exist to elicit localized right lower quadrant peritonitis, including the cough sign (the presence of point tenderness with a cough), percussion tenderness, and formal elicitation of rebound tenderness. Although all of these techniques are reasonably sensitive, one small study showed rebound tenderness to be the most accurate predictor of the localized peritonitis associated with appendicitis.28 Additional findings that may be helpful in diagnosing appendicitis include the psoas sign, the obturator sign, Rovsing’s sign, and rectal tenderness. The psoas sign is sought by having a supine patient actively flex the right hip against resistance, or by the examiner flexing and extending the patient’s right hip with the patient in the left lateral decubitus position. Pain with either of these maneuvers is thought to result from irritation of the underlying psoas muscle by an inflamed retroperitoneal appendix. The obturator sign is elicited by internally and externally rotating the flexed right hip. Pain is thought to arise when the inflamed pelvic appendix irritates the adjacent obturator internus muscle. Rovsing’s sign is the finding of right lower quadrant pain during palpation of the left side of the abdomen or when left-sided rebound tenderness is elicited. All of these findings are valuable when present, but their absence does not exclude appendicitis.27 Appendicitis can be easy to diagnose when the presentation is typical, but a typical presentation is encountered in only 50% to 60% of cases. An atypical presentation of appendicitis occurs for a variety of reasons. The classic migration of periumbilical pain to the right lower quadrant is thought to occur when the parietal peritoneum in the right lower quadrant becomes irritated by the inflamed appendix. In cases of retrocecal or pelvic appendicitis, this site might not become irritated. Atypical presentations of appendicitis are particularly common in patients who are at the extremes of age, pregnant, or immunosuppressed, including those with acquired immunodeficiency syndrome (AIDS) and a low CD4 cell count. Appendicitis in infants and young children remains a difficult diagnostic challenge because of difficulties in obtaining an accurate history. In young patients, the characteristic history of pain is difficult to elicit, and nonspecific findings of vomiting, lethargy, and irritability tend to predominate. Physical examination is difficult to perform because of poor patient cooperation and because localized right lower quadrant tenderness is found in less than 50% of patients.29 In addition, the characteristic laboratory findings often are not present. Leukopenia is as common as leukocytosis in young infants.30 As a result, errors in diagnosis are common,

and the frequency of complicated appendicitis is as high as 40% to 70%.31 The diagnosis of appendicitis in elderly patients also may be a challenge. In the elderly, the classic pattern of pain migration, right lower quadrant tenderness, fever, and leukocytosis are observed in only 15% to 30% of cases.30,32 Older patients also tend to present to medical attention in a delayed time frame relative to younger patients. For all of these reasons, the complication and perforation rates can be as high as 63% in patients older than 50 years.33 The presentation of appendicitis during pregnancy also is associated with an atypical clinical presentation, particularly in the later stages of pregnancy. In one series, only 57% of pregnant women with appendicitis had the classic progression of pain.34 Nausea and vomiting tend to be more common in pregnant women with appendicitis, but they also are common occurrences during normal pregnancy. Fever and leukocytosis are less commonly seen in pregnant woman than in other patient groups, and the value of leukocytosis is obscured by the physiologic leukocytosis of pregnancy. Although right-sided abdominal pain and tenderness are found in more than 90% of pregnant women with appendicitis, pain is located in the right lower quadrant only 75% of the time.34 Immunocompromised patients in general, and patients with AIDS in particular, represent a challenging group in which to diagnose appendicitis. Abdominal pain is reported in 12% to 45% of AIDS patients with appendicitis. The range of diagnoses responsible for this pain is significantly greater than in patients without HIV and includes opportunistic infections and malignancies, although in most cases, the pain is related to a diagnosis not associated with HIV.35 Research suggests that appendicitis occurs more often in HIV-infected patients than in HIV-negative patients, with as much as a four-fold increase in incidence.36 Although patients with AIDS usually present with the classic symptoms of appendicitis, there often is a history of chronic abdominal pain. Diarrhea also is a more common presenting symptom of appendicitis in HIV-positive patients, and leukocytosis is relatively uncommon. Declining CD4 counts are associated with delays in presentation to medical attention and increased perforation rates.37 Despite the challenges of diagnosing appendicitis in patients with HIV, the surgical outcomes with appropriate treatment are quite good; the largest series to date had no mortalities and a 13% complication rate, which is comparable to outcomes in patients without HIV.37

DIAGNOSIS Diagnosis of appendicitis remains a significant clinical challenge because of the many different entities that manifest with acute abdominal pain and the relatively nonspecific initial presentation of the disease. Because the natural history of appendicitis is a time-dependent progression to perforation, there is some urgency in

Chapter 116  Appendicitis Table 116-1  Differential Diagnosis of Appendicitis DIAGNOSIS

FINDINGS THAT HELP DIFFERENTIATE FROM APPENDICITIS

Bacterial or viral enteritis Mesenteric adenitis

Nausea, vomiting, and diarrhea are severe; pain usually develops after vomiting Duration of symptoms is longer; fever is uncommon; RLQ physical findings are less marked; WBC count is usually normal Pain is more likely to be felt in the right flank; high fever and rigors are common; marked pyuria or bacteriuria and urinary symptoms are present; abdominal rigidity is less marked Pain radiates to the right groin; significant hematuria; character of the pain is clearly colic Pain and vomiting are more severe; tenderness is less well localized; serum amylase and lipase levels are elevated History of similar attacks; diarrhea is more common; palpable mass is more common; extraintestinal manifestations may occur History of prior attacks is common; pain and tenderness are greater; radiation of pain is to the right shoulder; nausea is more marked; liver biochemical tests are more likely to be abnormal Very difficult to distinguish preoperatively from appendicitis Difficult to distinguish preoperatively from appendicitis; symptoms are milder and of longer duration; CT scan is helpful Usually occurs in older patients; changes in bowel habits are more common; radiation of pain is to the suprapubic area, not RLQ; fever and WBC count are higher History of abdominal surgery; colicky pain; vomiting and distention are more marked; RLQ localization is uncommon History of menstrual irregularities; characteristic progression of symptoms is absent; syncope; positive pregnancy test Occurs in the middle of the menstrual cycle; pain is of sudden onset; nausea and vomiting are less common; WBC count is normal Vomiting is more marked and occurs at the same time as the pain; progression of symptoms is absent; abdominal or pelvic mass often is palpable Longer duration of symptoms; pain begins in the lower abdomen; often there is a history of STDs, vaginal discharge, and marked cervical tenderness often are present

Pyelonephritis Renal colic Acute pancreatitis Crohn’s disease Cholecystitis Meckel’s diverticulitis Cecal diverticulitis Sigmoid diverticulitis Small bowel obstruction Ectopic pregnancy Ruptured ovarian cyst Ovarian torsion Acute salpingitis or tubo-ovarian abscess

CT, computed tomography; RLQ, right lower quadrant; STD, sexually transmitted disease; WBC, white blood cell.

making a prompt and accurate diagnosis. Not all causes of acute abdominal pain, however, require surgical intervention, and a negative appendectomy carries some risks for the patient, including adhesion formation, infection, and postoperative disability. Table 116-1 illustrates common diagnoses that can mimic acute appendicitis. Compounding this diagnostic challenge, there is no single symptom, finding, or laboratory test that is completely sensitive or specific for appendicitis.27

LABORATORY STUDIES

Laboratory findings in acute appendicitis include a variety of markers of acute inflammation. An elevated white blood cell count (WBC) in the range of 11,000 to 17,000/mm3 is seen in approximately 80% of patients, but the specificity of this finding for acute appendicitis versus other causes of acute abdominal pain is poor.38 An elevated proportion of granulocytes in the total white count or an elevated total neutrophil count (left shift) also is seen in the vast majority of patients with appendicitis, but is not specific for appendicitis.38 C-reactive protein (CRP), an acute phase reactant synthesized by the liver, is thought to rise within 12 hours of the development of an acute inflammatory process. Although CRP is elevated in 50% to 90% of cases of appendicitis, CRP is nonspecific when cutoff values of 5 to 25  mg/L are used.39 A urinalysis often is obtained in patients with acute appendicitis to exclude urinary tract infections, but mild abnormalities, either pyuria

or hematuria, are present in about 50% of cases of appendicitis.40 The value of laboratory investigations in diagnosing acute appendicitis has been a matter of some debate. In patients with a classic presentation by history and physical examination, many authors think that little additional information is obtained from laboratory studies. When all cases of appendicitis are considered, however, adding laboratory studies such as WBC, left shift, and CRP has been shown to improve diagnostic accuracy.41 When clinical findings are compared with inflammatory markers, inflammatory markers are stronger predictors of appendicitis than individual history or physical findings. Direct comparison of WBC and CRP suggests that total WBC or total granulocyte count is more sensitive and accurate than CRP for detecting acute appendicitis.39,41 The diagnostic performance of inflammatory markers is even better in identifying patients with perforated appendicitis. All patients with suspected acute appendicitis should have a CBC. A pregnancy test should be obtained in women of childbearing age. The value of other laboratory tests such as amylase, liver biochemical tests, or urinalysis lies in helping to exclude other diagnoses that can mimic acute appendicitis (see Table 116-1).

IMAGING STUDIES

Traditionally, there has been little role for routine imaging studies in patients with suspected acute appen-

2063

2064

Section X  Small and Large Intestine

Ap

A

Ap

Ap

B

C

Figure 116-2.  Transverse (A and B) and longitudinal (C) ultrasonograms of the right lower quadrant demonstrating a swollen, noncompressible appendix (Ap) proved at operation to be acute appendicitis. (Courtesy of Roy A. Filly, MD.)

dicitis. As is stated in the classic text Cope’s Early Diagnosis of the Acute Abdomen, “Over reliance on laboratory tests and radiological evaluations will very often mislead the clinician, especially if the history and physical examination are less than diligent and complete.”28,27 In 50% to 60% of cases, the diagnosis of appendicitis requires no imaging studies and can be made on clinical grounds alone.42,43 When diagnosis is less certain, a variety of imaging tests has been used to help confirm or exclude the diagnosis of acute appendicitis: plain abdominal films, abdominal ultrasound, radionuclide scans, and abdominal and pelvic computed tomography (CT).

Plain Abdominal Films

Plain films of the abdomen often are the initial imaging test for patients with acute abdominal pain. Findings on plain films of the abdomen consistent with appendicitis include a radiopaque right lower quadrant coprolith; focal right lower quadrant ileus or a sentinel loop; loss of the right psoas shadow; and a right lower quadrant soft tissue mass. All of these findings are suggestive of but not definitive for appendicitis. In a prospective study in which plain abdominal films were ordered on all patients with suspected appendicitis, the films changed clinical management in only 6% of cases.44 Plain abdominal films should be used in the evaluation of acute appendicitis only when bowel obstruction or gastrointestinal perforation is thought to be likely; their role in the routine evaluation of appendicitis should be discouraged.

Ultrasonography

Abdominal ultrasound has been used to image the acute abdomen since the 1980s. Although ultrasound is considered the imaging test of choice for biliary and gynecologic diseases, its importance in the diagnosis of acute appendicitis remains controversial. The ultrasound characteristics of appendicitis are well defined. Using a 5 or 7.5 MHz transducer, the technique of graded compression is used to displace the mobile loops of bowel in the right lower quadrant of the abdomen.

The diagnosis of appendicitis can be made with confidence if a 7-mm or thicker noncompressible blind-ended loop of bowel is identified (Fig. 116-2). A shadowing appendicolith, pericecal inflammation, or a localized pericecal fluid collection all suggest appendicitis.45 Appendicitis is excluded during ultrasound study by demonstration of a normal appendix. A normal appendix, however, is demonstrated in less than 50% of cases even by experienced sonographers, thus reducing the value of a “negative” ultrasound study.46 The reported sensitivity of ultrasound in the diagnosis of appendicitis in adults is 86% and its specificity is 81% in a collected review.47 Ultrasound appears to be more sensitive and specific in children than in adults, with sensitivity and specificity greater than 90% in most series and detection of a normal appendix in up to 90% of cases.48,49 There are some important limitations to the usefulness of ultrasound in the diagnosis of appendicitis. All ultrasound-based techniques are operator-dependent. The excellent results just mentioned were achieved in dedicated trials performed by interested and experienced ultrasonographers. In one multicenter trial focused on diagnosis of the acute abdomen, the real world sensitivity of ultrasound fell to 55%.50 Ultrasound also is less sensitive in patients with a body mass index (BMI) greater than 25 and in those with perforated appendicitis.51 Finally, ultrasound is more useful in confirming than in excluding the diagnosis of appendicitis, reducing its clinical utility in patients with a low pretest probability of appendicitis.

Radionuclide Scanning

Radionuclide scanning has been advocated when the diagnosis of appendicitis is uncertain. Two major techniques are used: either HMPAO (99mTc-hexamethylpropyleneamine oxime) labeling of the patient’s leukocytes, or 99Tc-labeled antigranulocyte antibodies. In both of these techniques, an accumulation of the radionuclide in the right lower quadrant is considered positive for appendicitis. The reported sensitivity of radionuclide

Chapter 116  Appendicitis

A

B

Figure 116-3.  Computed tomography for the diagnosis of acute appendicitis. A, Diffuse inflammatory changes in the mesentery surrounding the distal ileum and cecum in a patient with appendicitis. B, A fecalith is shown in the appendiceal lumen (arrow). (Courtesy of William R. Brugge, MD, Boston, Mass.)

scanning is 91% to 94%, and specificities are in the 82% to 94% range.52 Limitations of these techniques remain their lack of availability in all hospitals, the relatively long time required to perform them, and operator dependence in their interpretation.

scanning for appendicitis include the time required for enteric contrast to fill the bowel, decreased sensitivity in patients with low body fat, allergic reactions to intravenous contrast agents, exposure to ionizing radiation, and cost.

Computed Tomography

Overall Approach

Abdominal CT scans are considered the imaging study of choice in nonclassic cases of appendicitis. With the development of rapid helical and multidetector CT scanners, CT imaging is used increasingly to evaluate patients with acute abdominal pain. CT has long been considered valuable in making the diagnosis of appendiceal abscess, and CT-based therapy of these abscesses has become common.53 Since the 1990s, a number of authors have advocated broadening the use of CT scans to assist in the diagnosis of atypical appendicitis. A wide variety of techniques has been used for appendiceal protocol CT scans, which differ in terms of the amount of the abdomen scanned, the thickness of the individual cuts, and the types of contrast used. Several conclusions have emerged from these studies: thin (5 mm) cuts are better than thick (10 mm) cuts,54 and enteric contrast improves accuracy. CT findings consistent with appendicitis include an inflamed, distended (more than 6 mm) appendix that fails to fill with contrast or air (Fig. 116-3), often accompanied by an appendicolith or appendiceal wall thickening; periappendiceal inflammation, cecal apical thickening, and pericecal fluid collections are associated findings in appendicitis.55 Identification of a normal appendix or the finding of alternative intra-abdominal pathology constitutes a negative study. The performance of CT scanning for appendicitis has been impressive, with sensitivity rates of 94% and specificity rates of 95% in one collected review of multiple studies.47 The best results occur when enteric contrast is administered both by mouth and by rectum and contrast opacification of the cecum occurs. Limitations of CT

What constitutes the best imaging study has not been conclusively determined for patients in whom a diagnosis of appendicitis cannot be made confidently after clinical history, physical examination, and review of laboratory findings. Based on current evidence, however, it would appear that CT scanning is more sensitive, more specific, and less operator-dependent than ultrasound in adults.56 In pregnant women and in very thin patients, especially in institutions with experienced ultrasonographers, abdominal ultrasound is probably an alternative first imaging study in atypical cases of appendicitis. In pediatric patients, when the diagnosis of appendicitis cannot be made confidently after evaluation by a pediatric surgeon, ultrasound should be the first imaging test selected. This recommendation is based on the increased sensitivity of ultrasound in children and on the theoretical 10-fold increase in lifetime cancer risk engendered by exposure of children to ionizing radiation.57 In patients of any age, the initial step in evaluating patients with suspected acute appendicitis should be evaluation of the patient by an experienced surgeon, because this diagnostic evaluation is at least as accurate as any imaging study.58

CLINICAL SCORING SYSTEMS AND COMPUTER-AIDED DIAGNOSIS

Based upon data suggesting that examiner experience improves diagnostic accuracy in acute appendicitis, a variety of scoring systems has been devised since the 1990s to aid in the diagnosis of appendicitis. Most of these scoring systems assign numerical weights to find-

2065

2066

Section X  Small and Large Intestine ings from history, physical examination, and laboratory values in an attempt to predict the probability of appendicitis. More than 10 different scoring systems have been published, all of which purport to reduce errors in diagnosis and negative appendectomy rates. In an examination of the performance of multiple, published scoring systems on a single, well defined patient data set, the ability of all scoring systems to predict appendicitis was disappointing.59 The ability of scoring systems to perform well when applied to patient populations other than the population for which they had been developed remains a problem with these systems; other studies have reported similar results looking at individual scores.60 At this point there is no universally applicable scoring system for the diagnosis of acute appendicitis.

LAPAROSCOPY

Diagnostic laparoscopy has been proposed to assist in diagnosing equivocal cases of acute appendicitis. Inserting a laparoscope into the abdomen allows direct inspection of the appendix without appendectomy, if the appendix is found to be normal. The appeal of this approach is greatest in woman of childbearing age in whom gynecologic causes of acute abdominal pain can cloud diagnosis and who often are amenable to laparoscopic treatment. Two prospective studies of diagnostic laparoscopy in cases of possible appendicitis revealed gynecologic causes of pain in 48% to 73% of women with a normal appendix.61,62 Because there is some, albeit weak, evidence to suggest appendectomy might pre­ dispose women to tubal infertility,63 avoidance of unnecessary appendectomies is desirable in women of child­bearing age. Diagnostic laparoscopy has been used in two prospective series to nearly eliminate negative appendectomies in women of childbearing age.61,62 Despite these promising results, some cautionary notes must be sounded. Most studies of diagnostic laparoscopy report examinations performed under general anesthesia, making this a resource-intensive test compared with radiologic imaging studies. Although diagnostic laparoscopy can be performed under local anesthesia, inherent technical constraints reduce its success rate. For example, gynecologic pelvic laparoscopy performed under local anesthesia fails to obtain complete visualization of the pelvis in up to 15% of cases64; this incomplete examination rate compares poorly with CT scanning. Currently, diagnostic laparoscopy cannot be recommended over appendiceal protocol CT scanning as an initial test, but it probably should be used as a supplement to CT or ultrasound evaluations in which the results are equivocal.

DIAGNOSTIC ACCURACY

The concept of diagnostic accuracy refers to the fact that not all patients with a preoperative diagnosis of appendicitis are found to have acute appendicitis at operation. Because of the time-dependent risk of appendiceal perforation with its resultant increase in complications, it is important to make the diagnosis of appendicitis as

quickly as possible.65 As a result, treatment decisions often are made in the presence of incomplete clinical information. An appendectomy is termed negative when a normal appendix is found at exploration for acute appendicitis. Traditionally, an inverse relationship has been found between the frequency of negative appendectomies and the frequency of perforation at operation. Studies have shown that an increased diagnostic accuracy at operation carries an increased perforation rate,66 a tradeoff believed to be a consequence of the increased time required to confirm the etiology of acute abdominal pain in the absence of any specific test for appendicitis. In the interests of avoiding complications, standard teaching has been to accept a certain negative appendectomy rate to improve patient outcomes. Without diagnostic imaging, a negative appendectomy rate of 10% to 30% with a perforation rate of 10% to 25% is felt to represent a “good” balance6,17,42; in these series, the negative appendectomy rate was higher in women than men. In recent years, the use of imaging studies has improved the diagnostic accuracy for appendicitis without concomitant increases in perforation risk. In several series in which CT scanning was used selectively or universally in cases of presumed appendicitis, negative appendectomy rates have been reduced to 2% to 8% without an increase in perforation rates.42,43,67-70 This improvement in diagnostic accuracy has been observed in all patient groups, but most notably in women and children. These results suggest that with diagnostic imaging, it is possible to increase our diagnostic certainty without exposing patients to an increased risk of perforation. Whether a policy of increased use of imaging studies in the diagnosis of appendicitis will prove to be cost-effective is not yet clear, but early data suggest it might be, if enough negative explorations can be avoided.71 As a result of all of these recent diagnostic modalities, a new approach is emerging to the patient with acute abdominal pain and suspected appendicitis (Fig. 116-4). The goal of this new approach relies on imaging techniques and laparoscopy to eliminate in-hospital observation as a tool for improving diagnostic accuracy, thereby reducing the time required to increase diagnostic certainty and decreasing the likelihood of time-dependent complications. A strong incentive exists for avoiding negative appendectomies beyond diagnostic pride. Complication rates between 5% and 15% have been reported with the removal of a normal appendix, the majority of which are infectious, including wound, pulmonary, and urinary tract infections. A 1.3% risk of small bowel obstruction is reported in the series with the longest follow-up.72 Of patients found to have a normal appendix at operation, about 12% are found to have another surgical disease. An additional 18% to 20% have intra-abdominal findings that can explain their symptoms but are nonsurgical, the most common causes being ileitis or ileocolitis, mesenteric adenitis, or right ovarian cystic disease.73 An additional advantage of using CT scan in cases of atypical appendicitis is that many of these diagnoses can be

Chapter 116  Appendicitis Patient with presumed appendicitis

History, PE, CBC, UA, pregnancy test, serum amylase level High probability of appendicitis Appendectomy

Intermediate or low probability of appendicitis Child or pregnant woman

All others Appendiceal CT scan

Abdominal ultrasonography

Positive Appendectomy

Negative Discharge patient or perform diagnostic laparoscopy depending on clinical suspicion

Positive Appendectomy

made at CT scan.42 In as many as 60% of patients with negative appendectomy, no diagnosis can be made, even on subsequent evaluation.

COMPLICATIONS The major complication of untreated appendicitis is perforation, with resultant peritonitis, abscess, and portal pylephlebitis. Overall, the perforation rate in most series is between 10% and 30%, but the rate of perforation varies widely with age; perforation is most common at the extremes of age. Perforation rates as high as 90% have been reported in children younger than two years,29 and patients older than 70 years have perforation rates between 50% and 70%.33,74 Patients between the ages of 10 and 30 years have the lowest perforation rates, generally between 10% and 20%. The risk of perforation appears to increase as the duration of illness increases, particularly after 24 hours. Perforation of the appendix is a consequence of delay in diagnosis, and several studies have shown that patients with perforation have symptoms that average 30 hours longer than do patients with simple appendicitis.75 Much of this delay appears to be a result of delays in presentation to medical attention rather than delays in medical decision-making, but patients with perforation often have atypical presentations of their appendicitis. Patients with perforation are more likely to have sig­ nificant fever, leukocytosis, and physical findings of peritonitis than are patients with uncomplicated appendicitis. Although perforation often can be predicted preoperatively based on the presence of these findings, not

Negative Consider diagnostic laparoscopy depending on clinical suspicion

Figure 116-4.  Algorithm for managing the patient with presumed appendicitis. CBC, complete blood count; CT, computed tomography; PE, physical examination; UA, urinalysis.

all patients with these findings have perforation.76 Free perforation into the peritoneal cavity results in findings of diffuse peritonitis and can be associated with free intraperitoneal air on abdominal plain films. Patients with generalized peritonitis from appendicitis are difficult to distinguish preoperatively from patients with other causes of diffuse peritonitis. An abscess will develop after perforation if the perforated appendix is walled off from the remainder of the peritoneal cavity because of its retroperitoneal location or if it is walled off by loops of small intestine or omentum. A localized collection of inflammatory tissue (phlegmon) initially forms, and subsequently a true abscess develops. On physical examination, patients with an abscess resulting from appendicitis often have a palpable right lower quadrant abdominal mass. The most severe complication of appendiceal perforation is septic thrombophlebitis of the portal vein, also known as portal pylephlebitis. Although more common early in the 20th century, there are still cases of this disease today, although diverticular disease is now the most common cause. This rare complication should be considered in a patient with appendicitis who presents with high fever and mild jaundice. Treatment of pylephlebitis is control of the inciting infection and longterm (four to six weeks) antibiotic therapy. The major organisms causing pylephlebitis are Gram-negative enteric aerobes and anaerobes. Even with aggressive therapy, the incidence of hepatic abscesses following pylephlebitis is 50%, and mortality rates are 30% to 50%.77 A long-term complication of pylephlebitis is portal vein thrombosis with cavernous transformation of the portal vein and esophagogastric varices.

2067

2068

Section X  Small and Large Intestine TREATMENT Treatment of acute appendicitis is and remains appendectomy, despite the advent of sophisticated diagnostic and therapeutic modalities. Little has changed since Fitz and McBurney advocated early operative treatment of appendicitis in the late 19th century. Appendectomy is recommended, even though some cases of appendicitis resolve spontaneously. Although small studies have demonstrated that the vast majority of cases will improve with intravenous antibiotics alone, more than 35% of these patients have a relapse in one year.78 We currently lack the ability to identify prospectively self-limited cases, however, and to wait for resolution places patients at risk for perforation with its resultant life-threatening complications. Thus, appendectomy is a surgical urgency, and not a true emergency. Patients with appendicitis should be given appro­priate intravenous fluids to correct intravascular volume depletion and electrolyte imbalances, and intra­venous antibiotics to decrease wound infection rates; they should be taken to the operating room when they are stable. Brief periods of time may be taken to optimize the patient’s concomitant medical conditions before operation, but long delays increase the rate of perforation and compromise outcome. Two standard operative approaches exist for performing an appendectomy, either open appendectomy or laparoscopic appendectomy. Open appendectomy is performed though a muscle-splitting right lower quadrant incision; either an oblique or a transverse skin incision may be used. The appendix is identified and removed even if it is found to be normal. If normal, it is removed primarily to prevent future diagnostic confusion, and an exploration is carried out to identify other intra-abdominal causes of the patient’s symptoms. If other surgical pathology is found at exploration, the initial incision may be extended or a separate incision performed to address the problem. In advanced cases with severe inflammation, cecectomy may be required.79 Any abscesses are drained, and the abdomen is irrigated and closed. The other common approach to appendectomy is the laparoscopic appendectomy. First described by Semm in 1983,4 this procedure has been the subject of considerable study since that time. The technique of laparoscopic appendectomy has become standardized, and typically it is performed via a three-trocar technique. After gaining access to the abdomen, the appendix and then the entire abdomen are inspected. If the appendix is inflamed, an appendectomy is performed. If other intra-abdominal surgical pathology is found, it can be treated laparoscopically, or an appropriate open surgical procedure can be performed. It remains controversial whether laparoscopic appen­ dectomy is superior to open appendectomy. There have now been more than 20 randomized, controlled trials and five recent meta-analyses, including an analysis by the Cochrane collaboration, comparing the two procedures80; all of these studies have remarkably similar conclusions. Both procedures are safe and effective in the treatment of nonperforated appendicitis. After laparo-

scopic appendectomy, however, patients require less pain medication and return to normal activity about one week sooner than after open appendectomy. The wound infection rate is 50% lower than after open appen­ dectomy, but there may be an increased rate of intraabdominal abscess formation. The hospital course after laparoscopic appendectomy is 0.7 days shorter, and patients resume a normal diet at about the same time as after open appendectomy. Laparoscopic appendectomy takes more time to perform, and is associated with higher equipment costs. At this point, it is not possible to say that one procedure is superior to another for all patients; however, for some patient groups, especially young women, employed patients who need to return to work as soon as possible, and those with an uncertain diagnosis, laparoscopic appendectomy is preferable.80 Additionally, many consider the laparoscopic appendectomy the procedure of choice for the treatment of acute appendicitis in the morbidly obese population because it has been associated with shorter length of stay and lower morbidity.81,82 An exception to the statement that all patients with appendicitis require urgent appendectomy is the patient with perforation and a palpable right lower quadrant mass. These patients usually have extensive periappendiceal inflammation or abscess formation. In patients with a palpable mass who do not have diffuse peritonitis or toxicity, initial management can be operative or nonoperative. Although data are limited, there is a suggestion that early operative intervention may be associated with a higher complication rate.53 With initial nonoperative management, patients are placed on bowel rest and given intravenous fluids and antibiotics, and a CT scan of the abdomen is obtained. If a single abscess 3 cm or larger is discovered, percutaneous drainage of the abscess under CT guidance is performed. If multiple abscesses are found or the patient does not improve within 24 to 48 hours of conservative therapy, operative drainage is performed. Success rates of 88% to 95% have been reported with initial nonoperative management.83,84 Following resolution of the acute illness in older patients in whom a perforated cecal cancer is a possibility, colonoscopy, barium enema, or virtual colon­ography should be performed, because the incidence of appendiceal or cecal cancer in patients older than 60 years who present with acute appendicitis can exceed 20%.85 Some authors recommend interval appendectomy when the acute inflammation has resolved (6 to 12 weeks later), but the role of interval appendectomy remains controversial because the rate of recurrent appendicitis is less than 20% at one year.86 Natural orifice transluminal endoscopic surgery (NOTES) is an emerging field in minimally invasive surgery that is driving the development of new technology and techniques for procedures such as transluminal appendectomy. Although mostly investigational, NOTES procedures are performed using a transgastric, transcolonic, or transvaginal access point to the peritoneum. In 2004, Kalloo and his colleagues were able to demonstrate that natural orifices provide a port of entry via the gas-

Chapter 116  Appendicitis trointestinal tract to the peritoneal cavity. This approach requires the creation of a per-oral transgastric perforation into the peritoneal cavity using conventional endoscopes.87 Theoretically, this approach could reduce postoperative abdominal wall pain, wound infection, hernia formation, and adhesions. A number of case reports describing the removal of the appendix using a NOTES technique have been reported in registries established by the Natural Orifice Surgery Consortium for Assessment and Research (NOSCAR) or other organized groups such as EuroNOTES and the Brazilian Registry. In the United States, Sanntiago Horgan and Mark A. Talamini were the first to report the successful removal of an inflamed appendix through a patient’s vagina.5 With a number of rigorous laboratory and clinical research studies under way, it is hoped that as data are accumulated and instrumentation improves, NOTES may play an integral role in the future of abdominal surgery, including transluminal appendectomy.

OUTCOMES The modern treatment of simple acute appendicitis is associated with excellent outcomes. Factors responsible for these outcomes are advances in anesthesia, antibiotics, intravenous fluids, and blood products. The mortality rate from acute appendicitis in one large series was 0.08% with a complication rate of 5%.17 Older series have reported mortality rates of 0.2% with a complication rate of 6%.6 Patients typically are hospitalized for 24 to 48 hours after open appendectomy and 24 to 36 hours after laparoscopic appendectomy. Patients usually return to full activity two weeks after laparoscopic appendectomy and three weeks after open appendectomy.88 Morbidity and mortality attributable to appendicitis increase markedly with complicated appendicitis and in particular with perforation. Mortality rates of 1% to 4% and complication rates of 12% to 25% have been reported for perforated appendicitis.83 In patients older than 70 years, in whom perforation and significant medical comorbidity are common, mortality has been reported to be as high as 32%.33 Death in these circumstances usually is attributable to uncontrolled gram-negative sepsis or peritonitis, and patients with perforated appendicitis often have a stormy postoperative course, with intraabdominal abscesses and need for operative or percutaneous abscess drainage. Wound infection and dehiscence also are common in patients who have had open appendectomy, but these often promptly respond to wound drainage and antibiotics. These complications are minimized when a laparoscopic approach is chosen.

SPECIAL TOPICS THE APPENDIX AND ULCERATIVE COLITIS

A number of epidemiologic studies suggest that appendectomy protects against the development of ulcerative colitis (UC),89 particularly when performed for appendi-

citis; a similar relationship is not seen with Crohn’s disease. A meta-analysis of 17 case-control studies suggests that the relative risk of developing UC after appendectomy is about 0.3 times that of controls.90 Although these data come from case-control studies and questions can be raised about the appropriateness of the controls, this conclusion has been supported in one of the two large cohort studies performed.91 Some researchers have suggested that appendectomy also attenuates the course of active UC.92,93 Several investigators have also reported the improvement of UC after appendectomy, especially in young patients.94 In a mouse model of autoimmune colitis similar to UC, removal of the appendix early in life prompted significant attenuation of colonic inflammation.95 Although these findings are far from conclusive, they provide potential insights into both UC and the potential normal function of the appendix.

CROHN’S DISEASE OF THE APPENDIX

Although the appendix is often involved in patients with Crohn’s disease of the ileum or colon, isolated Crohn’s disease of the appendix is quite rare.96 Crohn’s appendicitis is difficult to distinguish from acute appendicitis preoperatively, although patients with Crohn’s appendicitis commonly have a longer history of pain. The treatment of appendiceal Crohn’s disease is appendectomy, which can be accomplished with a low rate of postoperative fistula formation.97 The clinical course of Crohn’s disease isolated to the appendix appears to be much more benign than that of typical Crohn’s disease. Because isolated Crohn’s disease of the appendix is quite rare, any patient found to have Crohn’s appendicitis should undergo evaluation of the remainder of the gastrointestinal tract for evidence of Crohn’s disease.

RECURRENT AND CHRONIC APPENDICITIS

Recurrent appendicitis is the clinical scenario in which a patient with pathologically confirmed acute appendicitis relates one or more prior episodes with identical symptoms, which resolved without surgical intervention. This diagnosis remains somewhat controversial but has been documented in clinical series.98 The diagnosis of recurrent appendicitis presupposes that some cases of appendicitis can resolve without medical intervention. Series of such cases exist in the radiologic literature, where patients with imaging findings consistent with appendicitis had rapid resolution of their symptoms without treatment. The percentage of cases of appendicitis that resolve spontaneously is unknown, but it is estimated at 6% to 8%. In small series of patients with spontaneous resolution of appendicitis, the recurrence rate is approximately 40%.99 No prospective means of identifying spontaneously resolving appendicitis have been identified, and therefore all cases of appendicitis should be treated surgically. The existence of recurrent appendicitis serves as a reminder not to discount the diagnosis of appendicitis in patients just because of prior episodes of similar abdominal pain. Chronic appendicitis is diagnosed when pathologic findings of fibrosis and chronic inflammation are found with a clinical syndrome consistent with appendicitis.

2069

2070

Section X  Small and Large Intestine Many of these patients report previous episodes of pain and relief of their symptoms after appendectomy.100 This is not a common problem, and caution should be used in applying this diagnosis to patients with poorly characterized chronic abdominal pain, because many of these patients are unlikely to improve with appendectomy.

DIVERTICULITIS OF THE APPENDIX

Diverticula of the appendix are uncommon, with a reported incidence in appendectomy specimens between 0.004% and 2.1%.101 Two forms of diverticula exist: congenital and acquired. True congenital diverticula are quite rare, but acquired diverticula are found in 1% to 2% of appendectomy specimens.102 Although the etiology of acquired appendiceal diverticula is unclear, they are thought to be pulsion diverticula, like colonic diverticula. Appendiceal diverticula typically are diagnosed incidentally on barium enema or CT scan or at surgical exploration.103 Acute inflammation of appendiceal diverticula (diverticulitis) produces a clinical picture that mimics acute appendicitis, making diverticulitis of the appendix a difficult preoperative diagnosis. Appendiceal diverticulitis, however, typically occurs in patients in the fourth decade of life rather than in the first or second decades, and it tends to manifest with a more insidious course, with many days of pain before presentation.104 CT scan can readily make the diagnosis. Appendiceal diverticulitis is more likely to be complicated by perforation than the usual case of appendicitis, making surgery, rather than nonoperative management, the treatment of choice.

EPITHELIAL MALIGNANCIES OF THE APPENDIX

Tumors of the appendix are rare and are found in approximately 1% of appendix specimens submitted for pathologic examination. The vast majority of appendiceal tumors are carcinoid, but this tumor is a rare cause of appendicitis because it usually arises from the tip of the appendix, not the base. (see Chapter 31). The incidence of epithelial malignancies of the appendix has been estimated to be 0.12 per 1 million persons per year.105 There are two types of epithelial malignancies: mucinous adenocarcinoma or cystadenocarcinoma of the appendix and colonic type (non-mucinous) adenocarcinoma of the appendix. Mucin-producing tumors are roughly twice as common as non–mucin-producing tumors.106 Non–mucin-producing tumors of the appendix typically manifest with a clinical picture indistinguishable from that of acute appendicitis, with acute right lower quadrant pain and tenderness. On CT scan, findings of a soft tissue mass or an appendix more than 15 mm in diameter should raise the suspicion of an appendiceal cancer.107 In contrast, less than one third of mucinous appendiceal adenocarcinomas manifest as acute appendicitis. More commonly these lesions are found incidentally on imaging studies as a cystic right lower quadrant mass or in a patient with increasing abdominal girth secondary to pseudomyxoma peritonei. The optimal treatment of all adenocarcinomas of the appendix is right hemicolectomy, either as a primary operation or as a secondary operation after adenocarci-

noma of the appendix is noted on pathologic examination of an appendectomy specimen. Additionally, patients with appendiceal adenocarcinoma have a significant risk of synchronous and metachronous neoplasms, which often originate from the gastrointestinal tract.105 Overall survival of patients with adenocarcinoma of the appendix is roughly 60% at five years and is a function of tumor stage at presentation. Appendiceal lymphoma is extremely uncommon, and primary lymphoma of the appendix accounts for 1% to 3% of all gastrointestinal lymphomas (see Chapter 29).108 Patients with appendiceal lymphoma usually present with acute appendicitis and an appendiceal diameter more than 2.5 cm, with surrounding soft-tissue thickening. Management of appendiceal lymphoma is appendectomy; right hemicolectomy is indicated only if there is extension of tumor beyond the appendix onto the mesentery or cecum.109 Appendiceal mucoceles are uncommon entities arising from a variety of different pathologic processes, of which only a small subset are associated with development of pseudomyxoma peritonei. Hence, the pathologic diagnosis determines further management.110 Perforation of a mucocele results in intraperitoneal dissemination of mucoid material, which can be acellular or can contain cells with varying degrees of dysplasia; cellular spread to the peritoneal surfaces leads to pseudomyxoma peritonei. These tumors usually are less aggressive than colorectal cancer, however, and they rarely manifest with lymph node or liver metastasis.111 The combination of surgery and complete cytoreduction should be followed by intraperitoneal rather than intravenous chemotherapy because the mainstay goal is prevention of locoregional recurrence, not prevention of systemic spread of disease.

INCIDENTAL OR PROPHYLACTIC APPENDECTOMY

The lifetime risk of appendicitis at birth is about one in 12, and declines to one in 35 by age 35 years. The greatest risk of appendicitis in a given year occurs over the second decade of life with a risk of about 0.25% per year.6 Although appendicitis is the most common cause of emergent abdominal surgery, given the low lifetime risk of appendicitis, elective prophylactic appendectomy cannot be recommended. Incidental appendectomy, the removal of a normal appendix at the time of other abdominal surgery was, at one time, the leading cause of appendectomy in women. In light of the falling incidence of appendicitis, enthusiasm for incidental appendectomy has declined. In operations where it will not add morbidity, however, a case may exist for incidental appendectomy in patients younger than 30 years. In older patients, the low residual lifetime risk of appendicitis makes incidental appendectomy difficult to defend.

ACKNOWLEDGMENT

The authors would like to acknowledge the significant contributions of Richard H. Turnage, MD, to prior editions of this chapter.

Chapter 116  Appendicitis KEY REFERENCES

Addiss DG, Shaffer N, Fowler BS, Tauxe RV. The epidemiology of appendicitis and appendectomy in the United States. Am J Epidemiol 1990; 132:910-25. (Ref 6.) Andersson RE. Meta-analysis of the clinical and laboratory diagnosis of appendicitis. Br J Surg 2004; 91:28-37. (Ref 41.) Carr NJ. The pathology of acute appendicitis. Ann Diagn Pathol 2000; 4:46-58. (Ref 15.) Ditillo MF, Dziura JD, Rabinovici R. Is it safe to delay appendectomy in adults with acute appendicitis? Ann Surg 2006; 244:656-60. (Ref 65.) Golden RL, Reginald H. Fitz, appendicitis, and the Osler connection—a discursive review. Surgery 1995; 118:504-9. (Ref 1.) Rao PM, Rhea JT, Rattner DW, et al. Introduction of appendiceal CT: impact on negative appendectomy and appendiceal perforation rates. Ann Surg 1999; 229:344-9. (Ref 42.) Rothrock SG, Pagane J. Acute appendicitis in children: emergency department diagnosis and management. Ann Emerg Med 2000; 36:3951. (Ref 29.)

Sauerland S, Lefering R, Neugebauer EA. Laparoscopic versus open surgery for suspected appendicitis. Cochrane Database Syst Rev 2004:CD001546. (Ref 80.) Schumpelick V, Dreuw B, Ophoff K, Prescher A. Appendix and cecum. Embryology, anatomy, and surgical applications. Surg Clin North Am 2000; 80:295-318. (Ref 12.) Terasawa T, Blackmore CC, Bent S, Kohlwes RJ. Systematic review: computed tomography and ultrasonography to detect acute appendicitis in adults and adolescents. Ann Intern Med 2004; 141:537-46. (Ref 47.) van Randen A, Bipat S, Zwinderman AH, et al. Acute appendicitis: meta-analysis of diagnostic performance of CT and graded compression US related to prevalence of disease. Radiology 2008; 249:97-106. (Ref 56.) Wagner JM, McKinney WP, Carpenter JL. Does this patient have appendicitis? JAMA 1996; 276:1589-94. (Ref 27.) Full references for this chapter can be found on www.expertconsult.com.

2071

CHAPTE R

117 Diverticular Disease of the Colon Jeffrey M. Fox and Neil H. Stollman

CHAPTER OUTLINE Epidemiology  2073 Pathologic Anatomy  2074 Etiology and Pathogenesis  2075 Colonic Wall Structure  2075 Motility  2075 Dietary Fiber  2075 Uncomplicated Diverticulosis  2076 Asymptomatic Diverticulosis  2076 Symptomatic Uncomplicated Diverticular Disease  2076 Complicated Diverticulosis  2078 Uncomplicated Diverticulitis  2078 Complicated Diverticulitis  2081

The earliest description of the pathology of diverticulosis traditionally has been attributed to Cruveilhier in 1849, although an earlier description by Sir Erasmus Wilson was noted in an editorial comment in Lancet in 1840.1 Occasional reports of the condition appear in the literature of the 19th century. The role of surgery in the treatment of acute diverticulitis was discussed by Mayo and associates in 1907. The presence of uncomplicated pseudodiverticula, herniations of the mucosa and submucosa through the muscularis of the colon, was defined as diverticulosis on radiologic studies by Case in 1914. The role of diet in the pathogenesis of diverticulosis was advanced in the landmark paper by Painter and Burkitt in 1971.2 Our knowledge of the epidemiology and clinical behavior of diverticular disease has grown rapidly over the past few decades, in large measure because of advances in imaging technology. The incidence of this disorder is increasing in the Western world.3-5 In 1998, diverticular disease resulted in total medical costs of $2.499 billion (direct costs: $2.358 billion; indirect costs: $141 million) and accounted for 230,058 hospital stays (with diverticular disease as the primary diagnosis only), 147,785 outpatient hospital visits, 165,343 emergency department visits, and 2,216,519 physician office visits.6 Despite the morbidity observed in symptomatic patients, most patients (80%) with diverticulosis remain asymptomatic throughout their lifetime. This observation raises the question whether a largely asymptomatic disorder, present in a majority of the elderly population, is in fact a “disease.” Moreover, because more elderly people have diverticulosis than do not, one can question whether it is even an abnormality and not just a normal phenomenon of aging in the Western world.

Special Topics Related to Diverticulitis  2083 Recurrent Diverticulitis  2083 The Young Patient  2083 The Elderly Patient  2084 The Immunocompromised Patient  2084 Right-Sided Diverticulitis  2084 Segmental Colitis Associated with Diverticulitis  2084 Diverticular Hemorrhage  2085 Epidemiology  2085 Pathophysiology  2085 Clinical Features  2085 Diagnosis and Treatment  2085

This chapter reviews the epidemiology, pathophysiology, clinical presentations, and treatments of diverticular disease of the colon.

EPIDEMIOLOGY Because most patients are asymptomatic, the true incidence and prevalence of diverticulosis are difficult to ascertain. Autopsy series can underestimate prevalence if small diverticula are missed or not commented upon by the pathologist, whereas series using barium enema for diagnosis can overestimate the condition and lead to selection bias because the study usually is done to investigate symptoms.7 Recent studies report overall prevalence rates for diverticulosis of 12% to 49%.8 The prevalence of diverticular disease clearly increases with age, ranging from less than 10% in those younger than 40 years of age to an estimated 50% to 66% of patients 80 years of age and older.9 Diverticulosis appears to be just as common in men and women,10 although men may have a higher incidence of diverticular bleeding and women may have more episodes of obstruction or stricture.11 Diverticulosis, with its striking geographic variation, has been termed a disease of Western civilization. The disorder is extraordinarily rare in rural Africa and Asia; conversely, the highest prevalence rates are seen in the United States, Europe, and Australia.2 Within a given country, the prevalence of colonic diverticula also can vary among ethnic groups. In Singapore, the annual incidence of diverticulitis in Chinese inhabitants was 0.14 cases per million, whereas in European inhabitants, the rate was 5.41 per million.12 Japanese-born persons who migrated to and lived in Hawaii

2073

2074

Section X  Small and Large Intestine Table 117-1  Factors That Influence the Risk for Diverticulosis Increased Risk Increasing age Dietary meat intake Living in Western countries (e.g., United States, Western Europe, Australia) Connective tissue diseases Decreased Risk High dietary fiber intake Living in predominantly rural Asian or African countries (e.g. Kenya, Jordan, Thailand) Equivocal or No Risk Gender Alcohol Smoking Colorectal cancer Polycystic kidney disease

had diverticulosis at autopsy in 52% of cases—much more frequently than those remaining in Japan.13 A registry of residents of Sweden revealed that immigrants to that country from low-prevalence regions, including Asia, Africa, and the Middle East, had hospitalization rates for diverticular disease that were 30% to 50% lower than those of Swedish natives and immigrants from Western countries.14 The magnitude of this gap narrowed as time from immigration to Sweden increased, however, and the gap nearly disappeared after 10 or more years in the country. As an individual country becomes more urbanized, an increase in diverticulosis seems to follow over time, as has been shown in Singapore and Israel.15,16 This observation may be attributable in part to a Westernization of diet with an increase in meat ingestion and a diminution of fiber intake as a country becomes more industrialized.10 Aside from age, geography, and ethnicity, other inherited and acquired risk factors have been associated with the presence of diverticulosis (Table 117-1). The role of dietary fiber is detailed later in this chapter. There is no conclusive evidence that diverticular disease is associated with colorectal cancer. Much of the sentinel data on the natural history of diverticulosis was reported by Parks in Belfast in the 1960s and 1970s,9,17 although these data suffer from the selection bias of studying only symptomatic patients. Parks observed that patients with many diverticula were, on average, older than those with few diverticula, suggesting that the number of diverticula in a patient increases with age. In contrast, patients with total colonic involvement were, on average, younger than those with segmental disease, suggesting that the pattern of colon involvement may be determined early on and remain more or less fixed. A study of barium enemas performed an average of 4.4 years apart in patients with diverticulosis demonstrated no apparent progression of disease in most patients.18

PATHOLOGIC ANATOMY Diverticula do not arise randomly around the circumference of the colon. Rather, they originate in four distinct rows that correspond to the four sites of penetration of the bowel wall by the major branches of the vasa recta: on either side of the taenia mesocolica and on the mesenteric side of

Antimesenteric taeniae

Diverticula

Mesenteric Mesentery taenia Arterial branch Figure 117-1.  Diagram showing colonic diverticula and their relationship to the taenia coli.

the taenia omentalis and taenia libera (Fig. 117-1). The diverticula point to the mesenteric border, and no bona fide diverticula arise from the antimesenteric intertaenial area. Diverticula maintain this fixed anatomic relationship to the taenia and are conspicuously absent from the portion of colon between the two antimesenteric taenia. Because diverticula do not involve all the layers of the colon wall, but rather are herniations of the mucosa and submucosa through a defect in the muscularis, colonic diverticula are, strictly speaking, pseudodiverticula. In this chapter, the incorrect, but traditional, terms diverticulum (singular) and diverticula (plural) are used rather than pseudodiverticulum and pseudodiverticula. Diverticula can vary in number from one to literally hundreds. The typical size of a diverticulum is 3 to 10 mm in diameter, but they can exceed 2 cm. Giant colonic diverticula have been described, with sizes up to 25 cm. Most giant diverticula are discovered incidentally, are single, are located in the sigmoid colon, and are asymptomatic. Rarely, giant sigmoid diverticula exhibit a valve-like phenomenon by which gas can enter but not leave; they get progressively larger over a relatively short period of time, can obstruct or perforate, and thereby require surgery.19 Diverticula can occur anywhere in the colon, and the segment typically involved depends highly on geography. In Western countries, diverticula occur mainly in the left colon, with up to 90% of patients having involvement of the sigmoid; only 15% have right-sided with or without left-sided involvement.1,9,20 In contrast, right-sided involvement is predominant in Asian countries. The ascending colon is involved in about 75% of patients in Singapore15,21 and only 25% of patients have sigmoid disease. In Japan, the prevalence of right-sided diverticulosis found on barium enema doubled (from 10% to 20%) from 1982-1997, a period over which left colon involvement remained the same (∼4%).22 The structural morphology of the diverticula found on either side of the colon appears to be identical worldwide. Although the precise factors causing the segmental predominance of left colon and right colon involvement in the West and East, respectively, are not known, environmental (e.g., dietary) and genetic factors are believed to play roles. Additionally, a blending of genetic and cultural lineages could explain why many people in the United States develop pancolonic diverticulosis.

Chapter 117  Diverticular Disease of the Colon ETIOLOGY AND PATHOGENESIS Investigations directed at elucidating the etiology of colonic diverticulosis have focused mainly on three areas: altered colonic wall structure, abnormal motility producing increased intraluminal pressures, and the role of dietary fiber; all three mechanisms are likely contributory.

COLONIC WALL STRUCTURE

Early gross descriptions of colons bearing diverticula reported thickening of the muscle wall and shortening of the taenia, with a resulting accordion-like pleating of the folds. This appearance, called myochosis (Greek myo, “muscle,” and chosis, “a heaping up”), is corroborated by the colonoscopic appearance in which markedly thickened folds are seen. Routine histology, however, does not reveal muscle hypertrophy. Electron microscopic studies confirm that the colonic walls in patients with diverticulosis have structurally normal muscle cells but, compared with controls, contain a greater than 200% increase in elastin deposition between the muscle cells in the taenia.23 The elastin is laid down in a contracted form, resulting in shortening of the taenia and bunching of the circular muscle. An increase in type III collagen synthesis in patients with diverticulosis also has been described, raising the possibility that agerelated changes in collagen composition play an etiologic role.24 In addition to an overall increase in the collagen content, an overexpression of a tissue inhibitor of metalloproteinases has been identified in colons with diverticula.25,26 Because matrix metalloproteinases are believed to regulate deposition of extracellular matrix proteins, an increase in their regulatory molecule, tissue inhibitor, might explain the increase in elastin and collagen deposition found in diverticular colons. The importance of intestinal wall connective tissue also is underscored by the higher rate of diverticulosis reported in patients with connective tissue disorders, such as Ehlers-Danlos syndrome, Marfan’s syndrome, and scleroderma.5

MOTILITY

Early investigations using colonic manometry demonstrated higher resting, postprandial, and neostigmine-stimulated luminal pressures in patients with diverticulosis compared with controls.27,28 Based on simultaneous manometry and cineradiography, Painter proposed a theory of segmentation, postulating that contraction of the colon at haustral folds caused the colon to act not as a continuous tube but as a series of discrete “little bladders,” which led to excessively high pressures within each segment.28,29 He further suggested that the Western diet might alter colonic motility to augment hypersegmentation, thereby increasing the tendency to form diverticula. More recently, using flexible endoscopy to accurately place manometric catheters within the sigmoid colon, the motility abnormalities previously described have been confirmed.30 Further, patients with symptomatic diverticular disease have been reported to have higher motility indices than either asymptomatic patients or persons without diverticula.31 In addition to increased contraction amplitude, another study found retropropagation of contractile waves in diverticular segments of colon, indicating that motility in these patients may be abnormal in magnitude and direction.32 An Asian study has confirmed elevated resting and stimulated luminal pressures in the presence of proximal colon diverticulosis.33

Although the demonstration of abnormal motility and elevated intraluminal pressures in the diverticular colon has been consistent, the physiologic basis for these abnormalities is less clear. Ion transport across the epithelial membrane of diverticular colons is the same as in controls.34 The number of myenteric and submucosal plexus neurons in diverticular colons are normal; however, the number of interstitial cells of Cajal (enteric pacemaker cells) is reduced.35 Increased activity of excitatory cholinergic nerves and a decreased activity of nonadrenergic, noncholinergic inhibitory nerves, the latter in part mediated by nitric oxide, have been demonstrated in diverticular colons compared with control colons.36 The magnitude of electrically stimulated contraction in diverticulosis-affected sigmoid colon is markedly reduced by antagonists of cholinergic and tachykinin neurotransmitters.37 In contrast, a study of the tachykinin neurotransmitter system showed a decreased contractility of circular muscle induced by substance P in diverticular colons compared with normal ones.38 The exact details of the neurochemical derangements underlying diverticulosis still need to be clarified, but there does appear to be an imbalance in the normal excitatory and inhibitory influences that results in the increased tonicity observed in colons with diverticulosis. The abnormal pressures and tone might contribute to formation of diverticula and to bowel dysfunction in patients with diverticulosis.

DIETARY FIBER

The wide geographic variation of diverticular disease, with higher prevalence in countries with a Westernized diet, has long suggested a dietary factor in its pathogenesis. Low intake of dietary fiber has been strongly suspected to be the main dietary factor behind these geographic differences. Burkitt and Painter were early proponents of this theory, labeling diverticulosis a “deficiency disease” that, like scurvy, should be avoidable with dietary changes.2 In one important study, they demonstrated that persons in the United Kingdom who consumed a refined, low-fiber Western diet had stool transit times more than twice as slow and stool weights significantly less than those of rural Ugandans eating a diet very high in fiber.39 The long intestinal transit time and smaller-volume stools were believed to allow an increase in intraluminal pressure, thus predisposing to diverticular herniation, whereas bulkier stools were asso­ ciated with less colonic contraction and lower wall pressures. Although more recent studies in Western cohorts have failed to confirm this finding in humans consistently, corroborative animal data do exist. Wistar rats fed low-fiber diets developed diverticula in 45% of instances, compared with only 9% of those on high-fiber diets.40 In humans, other observational evidence exists with respect to the etiologic role of fiber in diverticulosis. In the United States, fiber intake decreased by 28% from 1909 to 1975,5 a period of dramatic increase in the prevalence of diverticular disease. In a British study, a group of vegetarians on a high-fiber diet had a lower prevalence of diverticulosis than did non­ vegetarians (12% vs. 33%).41 Dietary influences for diverticulosis may have different effects on the right and left sides of the colon. Right-sided diverticular disease was shown in an Asian study to have no association with intake of fruits and vegetables or supplemental fiber, but it was strongly associated with meat consumption.42 Whether these associations apply to Westerners with right-sided diverticulosis is not known. The potential role of fiber in the treatment of diverticular disease is addressed later in this chapter.

2075

2076

Section X  Small and Large Intestine UNCOMPLICATED DIVERTICULOSIS Most patients with diverticulosis remain asymptomatic throughout their lifetimes, and many cases go undetected. Others have symptoms attributed to their disease but lack serious or life-threatening features; these patients with symptomatic uncomplicated diverticular disease (SUDD) are discussed later. Patients with diverticulitis and associated complications such as abscess, free perforation, fistula, stricture, or obstruction are described in the section on symptomatic complicated diverticular disease. Diverticular hemorrhage is described separately.

ASYMPTOMATIC DIVERTICULOSIS

Asymptomatic diverticulosis usually is an incidental finding in patients undergoing evaluation for other indications, such as occult blood loss or colon cancer screening. With an increasing number of people undergoing colonoscopic screening for colon cancer, it is likely that more such asymptomatic patients will be found. There is no clear indication for any specific therapy or special follow-up in these patients. A possible prophylactic benefit of a high-fiber diet has been suggested in two publications of 47,888 male health professionals who were followed over four years and in whom 385 (0.75%) new cases of symptomatic diverticular disease were identified.43,44 A dietary review found an inverse association between dietary fiber intake and the risk of subsequently developing clinically evident diverticular disease. The study also noted that fruit and vegetable fiber, or insoluble fiber, had a greater protective effect than did cereal fiber. Conversely, diets high in fat and red meat were associated with an increased risk of diverticular disease. Although prospective randomized trials are lacking, these findings suggest that patients with asymptomatic diverticulosis and SUDD (see later) might benefit from increasing their fruit and vegetable fiber intake while decreasing their fat and red meat consumption, a sensible lifestyle change that likely also provides other salutary health benefits. For decades, it has been a widely held but weakly supported belief that patients with diverticulosis should avoid consumption of seeds, nuts, and popcorn to avoid possible plugging of diverticula and precipitating an episode of diverticulitis; this theory also was investigated in the aforementioned cohort of male health professionals.45 Not only was no increase in risk of diverticular complications found as a result of consuming nuts, corn, popcorn, or seeded fruit (strawberries or blueberries), but consumption of nuts and popcorn were inversely associated with the risk of developing diverticulitis. Although there probably is not enough evidence to warrant actively encouraging patients with diverticulosis to eat large quantities of nuts, popcorn, or seeds, neither should these foods be categorically avoided.

SYMPTOMATIC UNCOMPLICATED DIVERTICULAR DISEASE Clinical Features

Some patients come to clinical attention because of nonspecific abdominal complaints and are found to have diverticulosis coli. A causal relationship between the diverticulosis and the abdominal symptoms often is difficult to establish. If there are features that are consistent with a diverticular source, however, and there is no evidence of a serious inflammatory condition, the disease may be defined as SUDD. Most patients with SUDD present with left lower quadrant pain; the British refer to this condition as painful

Figure 117-2.  Colonoscopic view of the sigmoid colon in a patient with symptomatic uncomplicated diverticular disease.

diverticular disease. The pain often is exacerbated by eating and diminished by defecation or the passage of flatus. Patients also may report other symptoms of colonic dysfunction, including bloating, constipation, diarrhea, or the passage of mucus per rectum. Physical examination may be normal or may reveal fullness or mild tenderness in the left lower quadrant, but frank rebound or guarding is absent. Because rates of occult bleeding in diverticulosis are similar to those in healthy controls, a positive fecal occult blood test never should be attributed to diverticulosis.46 Because a barium enema can characterize the number, size, and location of diverticula, it previously had been a commonly used initial study in such patients. Barium enema, however, may be insufficient to rule out competing or associated diagnoses such as malignancy in patients with diverticulosis. In symptomatic patients in whom a doublecontrast barium enema showed sigmoid diverticulosis, subsequent colonoscopy confirmed only 55% of neoplastic lesions that were suspected on the barium enema, whereas eight polyps and three malignancies were identified on colonoscopy that were missed on the barium enema (24% false-negative rate for barium enema).47 Although the barium enema may remain useful in certain cases, particularly if a colonoscopy cannot be performed safely or completely, endoscopic evaluation (Fig. 117-2) has assumed a primary role in the evaluation of most patients, particularly to exclude neoplasia. It once was believed to be unsafe to perform colonoscopy in patients with diverticulosis because of an increased risk of perforation; however, it subsequently was demonstrated that manometrically measured burst pressures for diverticula far exceed the usual pressures encountered during routine sigmoidoscopy or colonoscopy, even with the endoscope pressing against the wall or with heavy air insufflation.48 These data and many years of clinical experience have demonstrated the relative safety of using endoscopy to evaluate patients with abdominal symptoms. Caution should be exercised in patients with suspected or proved diverticulitis, however, because of a theoretical risk of perforating the wall of a diverticulum that has lost its integrity due to inflammation. In all patients, air insufflation should be minimized and excessive force in advancing the endoscope should be avoided.

Chapter 117  Diverticular Disease of the Colon The diverticula-laden colon can be challenging for the endoscopist to navigate because of spasm, luminal narrowing, fixation from prior inflammation and fibrosis, or confusion between luminal and diverticular openings. A number of solutions have been proposed to alleviate this problem. The use of a smaller-diameter pediatric colonoscope can be useful for difficult colons. One group has reported a success rate of more than 90% with a pediatric colonoscope in cases in which an adult colonoscope could not be passed through the sigmoid; 44% of these patients had diverticulosis.49 When colonoscopy with standard and pediatric colonoscopes were compared, the reason for failure to complete the examination was thought to be stenosing diverticular disease in 12 of 14 patients with the standard colonoscope, compared with two of eight with the pediatric colonoscope.50 A technique involving distention of the lumen with 100 to 300 mL of water, called the sigmoid floatation maneuver, was said to have facilitated colonoscopy in six technically difficult cases of severe diverticular disease.51 Occasionally, an endoscopist encounters an inverted diverticulum, where a diverticular dome protrudes into the lumen instead of out from it. These inversions often resemble polyps endoscopically, although they may be distinguished by their normal overlying mucosa, broad base, and location within a bed or row of diverticula. They are softappearing when manipulated with the endoscope tip or a biopsy forceps (pillow sign) and may be reducible. On barium enema, inverted diverticula appear as broad-based sessile polyps with a characteristic central umbilication,52 although it is not always possible to distinguish such a diverticulum from a polyp. When inverted diverticula are encountered, their removal should be avoided, although inadvertent colonoscopic diverticulectomy has been reported,53 and these patients had uneventful recoveries with conservative therapy. The presenting symptoms of SUDD overlap considerably with those of irritable bowel syndrome (IBS). Some authorities have postulated that diverticula are, in fact, a late consequence of IBS. In a Danish cohort of IBS patients, one third of whom had diverticula, no difference in symptoms or prognosis was detected between those with diverticula and those without diverticula over more than five years of follow-up.54 This finding led the investigators to conclude that there is no basis to consider SUDD as a separate entity from IBS. Ritchie reported that there was a similarity of pain sensation from rectal balloon distention in patients with IBS and those with diverticulosis.55 Whether these two disorders are distinct entities is unknown and probably not clinically important, because both are treated in a similar nonspecific fashion with equally good prognoses.

Treatment

Fiber Aside from the reported preventive effect of dietary fiber described earlier, fiber also is a mainstay of treatment for SUDD. Many uncontrolled trials of fiber in SUDD have been reported, all limited by a high placebo response rate. A randomized, double-blind trial from Oxford University showed a statistically significant decrease in bowel symptoms relative to controls in patients with SUDD who were placed on a high-fiber diet56; the separation between treatment and control groups, however, was not noted until the three-month follow-up evaluation. It is important to instruct patients to start fiber supplementation at a low dose and slowly increase the dose, because patients initially can worsen from diarrhea, gas, or bloating if the fiber dose is started too high or quickly. Because it often can take months to improve, as demonstrated in the Oxford study, the initial

adverse symptoms can discourage patients from adhering to the fiber supplements if they are not counseled properly. In contrast, another study of fiber supplementation in patients with diverticular disease showed no significant improvement in overall symptoms, although decreased transit time and increased stool frequency were documented.57 Despite these conflicting data and the certainty that diverticula do not regress with an increased fiber intake, some amelioration of symptoms in patients with SUDD often can be seen with a high-fiber diet. 5-Aminosalicylic Acid Additional medical therapies for SUDD have been studied since 2000, drawing on approaches that are effective in other colonic diseases such as inflammatory bowel disease (IBD) and IBS. 5-Aminosalicylate (5-ASA) compounds, a well established therapy for ulcerative colitis and Crohn’s disease, have been evaluated as a potential treatment for SUDD. Although patients with SUDD by definition lack severe or overt inflammation, in some patients subtle inflammation is suspected even without gross signs of diverticulitis; it is these patients that the anti-inflammatory properties of 5-ASA might benefit. Published studies that have examined the role of 5-ASA were randomized to either daily or cycled (e.g., 10 days per month) 5-ASA but were not placebo controlled.58-60 Results of each study show a significantly reduced symptom score relative to pretreatment scores, but the lack of a control arm and the known high placeboresponse rates in functional bowel syndromes make the results difficult to interpret. The doses of 5-ASAs used in these studies were lower than generally are used for IBD. These results, combined with the relative safety of this medication class, make 5-ASAs a promising therapy for SUDD, although placebo-controlled trials supporting its efficacy will be needed before widespread use can be advised. Antibiotics and Probiotics The role of pathogenic and nonpathogenic bacteria in intestinal disease is being increasingly scrutinized. Some have postulated that a disturbance in the local microflora in and around diverticula61 might predispose to diverticulitis. If this were true, medications that alter this flora might help treat or prevent attacks of diverticulitis. Although conventional antibiotics against coliform bacteria clearly are indicated in overt diverticulitis, it is difficult to justify their use in a condition such as SUDD, in which evidence of gross infection is scant. The considerable risks of antibiotic therapy with regard to systemic side effects, potential development of Clostridium difficile colitis, and potential for bacterial resistance appear to outweigh any theoretical benefits when it is difficult or impossible to prove the presence of active infection. In contrast, rifaximin, a nonabsorbable antibiotic with broad-spectrum activity, mitigates some of these concerns by its solely luminal activity and by its use as an emerging potential therapy for C. difficile infection. Rifaximin has been studied in patients with SUDD and has shown promise in reducing frequency and severity of symptoms.62,63 Rifaximin has also been shown to be effective for IBS,64 possibly by eradicating concomitant small bowel bacterial overgrowth (SIBO). Whether inadvertently treating SIBO is the actual reason for the effectiveness of rifaximin in SUDD is not known. Based on the same theory that altered local microflora is present in these patients, probiotics also have been studied in SUDD.59,65 Although some benefit has been shown, such trials are small and lack a placebo group. Although higher-

2077

2078

Section X  Small and Large Intestine quality evidence needs to be produced to support this approach, the microflora may become an important target for therapy in SUDD in the coming years; a number of trials are under way evaluating anti-inflammatory agents or probiotics. Anticholinergics and Antispasmodics Hypermotility of the colon in diverticulosis suggests that anticholinergic or antispasmodic medications such as dicyclomine or hyoscyamine might improve symptoms by diminishing muscular contraction. Nonetheless, there are no adequately controlled therapeutic trials documenting such a benefit. Intravenous glucagon has been reported to offer short-term relief of pain, presumably as a result of smooth muscle relaxation. There is no rationale for the use of narcotic analgesics in uncomplicated diverticular disease. Role of Surgery Surgical intervention generally is not considered for patients with truly uncomplicated diverticulosis, because the risks of surgery outweigh the benefits in most cases. Some patients with subclinical or smoldering diverticulitis present with pain characteristic of diverticulitis but show no signs of systemic inflammation, such as fever or leukocytosis. In a cohort of such patients from the Mayo Clinic who underwent sigmoid resection with primary anastomosis for their symptoms without signs or laboratory markers of systemic inflammation, 76% of the resected specimens had evidence of acute or chronic diverticular inflammation.66 Sigmoid colectomy in these patients resulted in resolution of pain in 88% and complete resolution of symptoms in 76% after one year or more of follow-up. This finding underscores the importance of clinical follow-up and an open mind regarding patients with apparently uncomplicated disease whose symptoms do not improve with conservative treatment.

COMPLICATED DIVERTICULOSIS Diverticulitis, defined as inflammation, infection, or both, associated with diverticula, is probably the most common clinical manifestation of this disorder, affecting an estimated 10% to 25% of patients with diverticula.9 It generally is believed to be the result of perforation of a single diverticulum.67 When this results in a localized phlegmon, the term uncomplicated diverticulitis is used. Complicated diverticulitis refers to cases associated with abscess, free perforation with peritonitis, fistula, or obstruction.68 Besides diverticulitis, the other major form of complicated diverticular disease is bleeding, which is discussed later in this chapter.

UNCOMPLICATED DIVERTICULITIS Pathophysiology

The process by which a diverticulum becomes inflamed has been likened to that causing appendicitis, in which the diverticular sac becomes obstructed by inspissated stool in its neck; the fecalith abrades the mucosa of the sac, causing low-grade inflammation and further blocking drainage. Histologically, one of the earliest signs of inflammation is hyperplasia of the mucosal lymphoid tissue, with lymphoid tissue aggregation at the apex of the involved sac.69 The obstructed diverticulum predisposes to expansion of the normal bacterial flora, diminished venous outflow with localized ischemia, and altered mucosal defense mecha-

Table 117-2  Hinchey Classification of Colonic Diverticular Perforation STAGE

DEFINITION

I II III

Confined pericolic abscess Distant abscess (retroperitoneal or pelvic) Generalized peritonitis caused by rupture of a pericolic or pelvic abscess (not communicating with the colonic lumen because of obliteration of the diverticular neck by inflammation) Fecal peritonitis caused by free perforation of a diverticulum (communicating with the colonic lumen)

IV

nisms. One such alteration is a defective CD2 pathwayinduced apoptosis, which has been found in lamina propria lymphocytes in patients with diverticulitis, possibly leading to an up-regulation of the local immune response in these patients similar to that seen in patients with IBD.70 Evidence also suggests that cytomegalovirus (CMV) reactivation might contribute to local inflammatory activity, because active CMV replication was found in tissue from the affected bowel segments of more than two thirds of patients with diverticulitis.71 The cascade of events initiated by fecalith obstruction, and possibly enhanced by underlying innate or acquired abnormalities, allows bacteria to breach the mucosa and extend the process transmurally, ultimately leading to perforation.72 The extent and localization of the perforation determine its clinical behavior. Microperforations can remain very well localized, contained by the pericolic fat and mesentery, and cause small pericolic abscesses. A larger perforation can allow a more extensive abscess to form, which can track longitudinally around the bowel wall. This process can lead to a large inflammatory mass, fibrosis, extension to other organs, or fistula formation. Free perforation into the peritoneum causing frank bacterial or fecal peritonitis can be life threatening, but fortunately it is uncommon, with a population incidence of 4 cases per 100,000 population per year.3,73 Hinchey and associates have described a staged grading system reflecting the severity of perforation (Table 117-2).74

Clinical Features

Patients with acute diverticulitis typically present with left lower quadrant abdominal pain, reflecting the propensity for this disorder to occur in the sigmoid colon in Western countries; a redundant sigmoid colon, however, can manifest with suprapubic or right-sided pain. In contrast, Asian patients with diverticulitis have predominantly right-sided symptoms, corresponding to the location of their diverticula.75 The pain may be intermittent or constant and frequently is associated with a change in bowel habits, either diarrhea or constipation.76 Anorexia, nausea, and vomiting also can occur. Dysuria and urinary frequency can result from bladder irritation caused by the adjacent inflamed sigmoid colon. Physical examination usually discloses localized tenderness, generally in the left lower quadrant; however, as noted, right-sided signs do not preclude the possibility of diverticulitis. Guarding and rebound tenderness may be present, as may a tender, cylindrical, palpable mass. Bowel sounds typically are depressed but may be normal in mild cases or increased in the presence of obstruction. Rectal examination can disclose tenderness or a mass,

Chapter 117  Diverticular Disease of the Colon particularly with a low-lying pelvic abscess. Fever is present in most patients, whereas hypotension and shock are unusual. The white blood cell (WBC) count commonly is elevated, although one study reported a normal WBC count with no left shift in 46% of patients.76 No other laboratory abnormalities are routinely helpful, although they can help to rule out other diagnostic possibilities in select patients. The differential diagnosis for diverticulitis is extensive. Acute appendicitis is the misdiagnosis most often made in patients with diverticulitis, particularly with right-sided disease. In Hong Kong, where awareness of the predominance of right-sided diverticulosis presumably is high, 34 of 35 patients with right-sided diverticulitis initially were believed to have acute appendicitis.75 Although appendicitis is, on average, a disease of younger patients than is acute diverticulitis, there is a wide range of ages for both. Clinical suspicion for one must remain high when diagnosing the other on clinical grounds. Other common diagnoses that need to be considered include IBD; other forms of colitis (infectious or ischemic); colorectal cancer; and gynecologic conditions such as pelvic inflammatory disease, ovarian cyst rupture, and ovarian torsion. Occasionally, diver­ ticulitis can occur concomitantly with other diseases; in one study of patients admitted to the hospital with diverticulitis, 7% were found later also to have a colon malignancy.77

Diagnosis

Most patients with acute diverticulitis present with signs and symptoms sufficient to justify the clinical diagnosis and institution of empiric therapy. Clinical diagnosis can, however, occasionally be inaccurate, and emergency surgery for presumed diverticulitis, without the benefit of radiologic confirmation, carries a misdiagnosis rate as high as 34% to 67%.78 Therefore, radiologic studies to confirm the diagnosis of diverticulitis should be employed, particularly if invasive intervention may be required. Plain Films An erect chest film, together with erect and supine abdominal films, should be performed on patients with significant abdominal pain. The erect chest film has the dual purpose of detecting pneumoperitoneum, which has been reported to be present in up to 11% of patients with acute diverticulitis,79 and of assessing cardiopulmonary status in a generally elderly population with common comorbid illness. Plain abdominal films are abnormal in 30% to 50% of patients with acute diverticulitis,79,80 with findings that include bowel dilatation from obstruction or ileus, or a soft tissue density suggesting an abscess. Contrast Enema Examinations Contrast barium enema (Fig. 117-3) had been the diagnostic standard for diverticulitis and its complications for many years. Because the use of barium in the setting of an intestinal perforation carries a risk of barium peritonitis, only water-soluble contrast enemas, such as Gastrografin, should be used in the setting of suspected diverticulitis. A gentle, single-contrast study should be performed and terminated once findings of diverticulitis are discovered, with visualization of the entire colon deferred to a later date; air (double)-contrast studies are not indicated. Findings considered diagnostic of diverticulitis include demonstration of extravasated contrast material with or without the outlining of an abscess cavity, an intramural sinus tract, or a fistula.1,81 Extensive diverticulosis, spasm, mucosal thickening or spiking, or deformed sacs, although suggestive of

Figure 117-3.  Film from a barium enema in a patient with pancolonic diverticulosis. There is also marked redundancy and overlapping of the sigmoid colon, which obscures the identification of intraluminal lesions.

Figure 117-4.  Computed tomography scan of a patient with acute diverticulitis showing colon wall thickening, and stranding of the pericolic fat.

diverticulitis, are not conclusive. An extraluminal mass compressing the colon is said to be the most common finding in severe diverticulitis,82 although this finding is not specific for this diagnosis. Obviously, in the absence of diverticula or associated findings, the diagnosis must be reconsidered. Contrast enema has been shown to have a sensitivity of 62% to 94% for detecting acute diverticulitis, with false-negative results in 2% to 15%.68,83 Computed Tomography Because diverticulitis is mainly an extraluminal disease, luminal contrast studies may be inaccurate. Computed tomography (CT) (Fig. 117-4) has now replaced contrast

2079

2080

Section X  Small and Large Intestine enema as the diagnostic procedure of choice for acute diverticulitis and has the ability to image mural and extraluminal disease and to enable therapy with percutaneous drainage of abscesses. Abdominal and pelvic scanning ideally is performed with water-soluble contrast, given both orally and rectally, and with intravenous contrast when it is not contraindicated. CT criteria for diverticulitis include the presence of diverticula with pericolic infiltration of fatty tissue (often appearing as fat stranding), thickening of the colon wall, and formation of abscesses. The earliest large series of CT findings in diverticulitis reported the finding of pericolic fat inflammation in 98%, diverticula in 84%, a colonic wall thickness greater than 4 mm in 70%, and an abscess in 35%.84 Contrast enemas in the same patients underestimated the extent of disease in 15 (41%) of 37 cases. Numerous subsequent trials comparing these two modalities in patients with suspected diverticulitis consistently have reported CT sensitivities of 93% to 98% and specificities of 75% to 100%, significantly more accurate than contrast enemas.81,85,86 CT also has been found to be highly sensitive and specific for rightsided diverticulitis and in helping to differentiate diver­ ticulitis from colorectal cancer of the ascending colon and cecum.87,88 Although some reports show a lower sensitivity of CT for diverticulitis,89,90 it is increasingly becoming clear that, when diagnosis is in doubt or clinical deterioration occurs, CT is the best primary radiologic diagnostic modality. Conversely, in patients with mild disease and in whom the diagnosis is straightforward, CT scanning might not be necessary. Ultrasonography Based on its relatively low cost, convenience, and noninvasive nature, ultrasonography (US) has been advocated as a potentially useful diagnostic modality in diverticulitis. Characteristic findings implying active inflammation include bowel wall thickening, presence of diverticula or abscesses, and hyperechogenicity of the bowel wall. US has a reported sensitivity of 84% to 98% and specificity of 80% to 93%.91,92 One study of 71 patients with suspected diverticulitis who underwent US reported a negative predictive value of 100%.93 A trial comparing US with CT revealed equally good accuracy.94 US also is useful in female patients to exclude gynecologic pathology. Despite these encouraging data, US remains highly operator dependent, especially for detecting abscesses between loops of bowel, air-filled abscesses, or disease complications posterior in the abdomen that are hard to visualize. US therefore remains a second-line diagnostic tool to be used in select circumstances. Magnetic Resonance Imaging With limited resolution secondary to motion artifact introduced by peristalsis and respiration, the potential role of MRI remains to be demonstrated. Whether decreased scan times and intraluminal contrast agents will overcome these limitations remains to be seen. Endoscopy Because of the risk of perforation, either from the instrument itself or from air insufflation, endoscopy generally is avoided in the initial evaluation of patients with suspected acute diverticulitis. A limited sigmoidoscopy with minimal air insufflation may be helpful to exclude alternative diagnoses, such as IBD, carcinoma, or ischemic colitis in unclear cases. As noted earlier, once the acute phase has passed, a colonoscopy should be electively performed one to three

months later to exclude competing diagnoses, particularly neoplasia.

Treatment

One of the initial decisions regarding patients with diverticulitis involves a determination of the need for hospitalization. Factors to be considered include the severity of illness, ability to tolerate oral intake, comorbid diseases, and available outpatient support systems, such as a reliable family. An appropriate patient for outpatient management is one with mild symptoms, no peritoneal signs, the ability to take oral fluids, and a supportive home network. These patients should be treated with a clear liquid diet and antibiotics. When cultured, most diverticular abscesses grow mixed aerobic and anaerobic organisms, the most common single organisms being Escherichia coli, Streptococcus species, and Bacteroides fragilis.95 Therefore, oral antibiotics with broad-spectrum coverage (see later) are recommended.96 Patients treated as outpatients should have close followup. They should be instructed to call the physician for increasing pain, fever, or inability to tolerate oral fluids, each of which could indicate the development of com­ plications and the need for hospitalization. Symptomatic improvement generally should be evident within two to three days, at which time the diet may be slowly advanced. Antibiotic treatment should be continued for seven to 10 days. Patients with uncomplicated diverticulitis who are elderly or immunosuppressed, have severe comorbidities, or demonstrate high fever or significant leukocytosis should be hospitalized. Although data imply that early consultation with a gastroenterology subspecialist improves quality of care for the inpatient management of diverticulitis,97 this is probably not necessary for many straightforward cases. Bowel rest with either clear liquids or nothing by mouth should be instituted. Intravenous fluid therapy to restore intravascular volume, balance electrolytes, and ensure adequate urinary output should be initiated. Broad-spectrum intravenous antibiotics should be started. Recommended combination regimens include anaerobic coverage with metronidazole or clindamycin and Gram-negative coverage with an aminoglycoside, monobactam, or a third-generation cephalosporin.96 Single-agent coverage with intravenous second-generation cephalosporins or beta-lactamase inhibitor combinations, such as ampicillin/sulbactam or ticarcillin/clavulanate, are reasonable alternatives. Symptomatic improvement with decreasing fever and leukocytosis should be observed within two to four days, at which point diet may be advanced. If improvement continues, patients may be discharged, but they should complete a seven- to 10-day course of oral antibiotics. Failure to improve with conservative medical therapy warrants a diligent search for complications, consideration of alternative diagnoses, and surgical consultation. Most patients hospitalized with acute diverticulitis respond to conservative medical therapy, but it has been estimated that 15% to 30% require surgery during the hospital admission.1,5,17,68 Surgery may be necessary when pain, fever, and leukocytosis do not respond to antibiotics and supportive care. If surgery is planned and complicated disease, such as abscess, is ruled out, resection of the diseased segment of bowel with primary anastomosis is the most commonly performed operation, usually in a single-stage procedure. The main requirements in performing a single-stage procedure are the ability to perform bowel preparation and technical feasibility, usually determined by the extent of extramural inflammation. The entire sigmoid colon should

Chapter 117  Diverticular Disease of the Colon

A

B

C Figure 117-5.  Computed tomography (CT) scans with CT-guided percutaneous drainage of an abdominal abscess complicating acute diverticulitis. A, Arrows point to the abscess arising from the sigmoid colon. B, A drainage catheter has been inserted into the abscess. The patient also was treated with intravenous antibiotics. C, The catheter has been removed, and the abscess has resolved.

be removed (in left-sided disease) because this is the most common location for initial disease and recurrence. The distal resection margin should be the proximal rectum, and the anastomosis should be free of tension to decrease risk of anastomotic leak.98 For uncomplicated diverticulitis, laparoscopic sigmoid colectomy has gained increasing enthusiasm99 and has significant advantages over open techniques with respect to length of stay and postoperative in-hospital morbidity.100 Discussion of the medical and surgical management of complications such as perforation, abscess, fistula, or obstruction follows.

COMPLICATED DIVERTICULITIS Abscess

When perforation of a colonic diverticulum occurs, the ability of the pericolic tissues to control the spread of the inflammatory process determines the subsequent clinical course of disease and its treatment. A localized phlegmon initially develops with limited spread. Further spread can lead to the formation of larger local or distant abscesses (Hinchey stages I and II, respectively). When abscess contents spread diffusely in the peritoneum, causing purulent or fecal peritonitis (Hinchey stages III and IV, respectively), it can lead to sepsis and death if the patient does not undergo urgent surgical intervention. Clinical signs suggesting abscess formation include a tender abdomen and sometimes a tender mass on physical examination, and persistent fever, or leukocytosis despite an adequate trial of appropriate intravenous antibiotics. Once an abscess is suspected, radiologic evaluation with a CT scan is the best modality

for confirming the diagnosis and following its course over time. Small pericolic abscesses (Hinchey stage I) often can be treated conservatively with broad-spectrum antibiotics and bowel rest.98 In one series of patients with diverticulitis, seven of 10 patients with pericolic abscesses responded successfully to conservative treatment.101 This favorable prognosis can result from maintenance of a fistula between the abscess and the colon lumen, thus permitting spontaneous internal drainage. Continued noninterventional management of abscesses should be considered only in stable patients who demonstrate unequivocal improvements in pain, fever, tenderness, and leukocytosis over the first few days of therapy. CT-guided percutaneous drainage of abdominal abscesses has assumed a prominent complementary role to surgery (Fig. 117-5). The immediate advantage of percutaneous catheter drainage is rapid control of sepsis and patient stabilization without the need for general anesthesia. It often eliminates the need for a multiple-stage surgical procedure with colostomy,102 instead allowing temporary palliative drainage and subsequent single-stage resection in three to four weeks. Success rates of CT-guided drainage for stabilizing patients and safely allowing subsequent single-stage procedures range from 74% to 80%.103,104 An urgent surgical procedure is required in the 20% to 25% of patients in whom the abscess is multiloculated, anatomically inaccessible, or not resolving with percutaneous drainage. The approach to surgical intervention has evolved over the last half century. Historically, three-stage management was preferred for complicated diverticulitis of any severity:

2081

2082

Section X  Small and Large Intestine stage I involved a proximal colostomy and drainage of abscesses, stage II was for resecting the diseased bowel (colostomy maintained to protect anastomosis), and stage III was to restore bowel continuity. Although the multiplestage surgical approach to diverticular disease was effective for managing abscesses, generally it is desirable to limit the number of laparotomies to minimize morbidity and duration of hospitalization. Additionally, only about half of patients undergoing proximal-end colostomy have the colostomy reversed because of the technical difficulties and the risks of anastomotic leakage and mortality.105 In the last few decades, two-stage management (e.g., primary resection of diseased bowel with proximal end colostomy and oversewing of the distal stump, also referred to as the Hartmann procedure, followed by reanastomosis) and, increasingly, management with a single operation (resection with primary anastomosis) have become the preferred surgical approaches. The switch to fewer operations has been made without a discernable compromise in overall outcomes relative to three-stage approaches.106 Single-stage management, which is becoming the standard for elective management of uncomplicated diverticulitis, also is being used increasingly in appropriate patients with complicated diverti­culitis. To do so, bowel preparation is believed to be necessary. When free perforation is ruled out, bowel preparation is performed before the operation with either a one-day polyethylene glycol (PEG)-electrolyte lavage or a traditional two- or three-day mechanical preparation; both types of preparation have been shown to have similar outcomes in a randomized trial.107 In the setting of urgent or emergent surgery when free perforation is a concern, bowel lavage may be performed “on table,” thereby allowing a primary anastomosis to be performed in a single stage.108,109 Regardless of approach, surgery for complicated diverticulitis has evolved into a relatively safe procedure with one of the highest success rates of any of the common gastrointestinal surgical procedures.110

Free Perforation

Peritonitis (Hinchey stages III or IV) is a surgical emergency and requires urgent operative intervention. Although uncommon in the antibiotic era, mortality from generalized peritonitis associated with diverticulitis has been reported in the 12% to 26% range.111 Early identification of free perforation is critical. CT scan can confirm the diagnosis in ambiguous cases, but an abdominal plain-film series showing free intraperitoneal air plus high clinical suspicion is sufficient to justify surgical exploration. Broad-spectrum intravenous antibiotics, such as a second- or thirdgeneration cephalosporin and metronidazole, should be instituted immediately. For more-confined abscesses, surgical options are primary or secondary resection. Primary resection of the diseased sigmoid colon uses either the Hartmann procedure (twostage) or primary colorectal anastomosis (single-stage). In secondary resection, the initial operation creates a diverting colostomy and sutures the perforation, if possible, and the subsequent procedure(s) resects the diseased colon and restores bowel continuity. There is no clear consensus whether primary or secondary resection is more beneficial for peritonitis from diverticulitis. Retrospective studies suggest an advantage for primary over secondary resection with respect to morbidity and mortality.111,112 Two randomized trials comparing primary and secondary resection yielded conflicting results, neither appearing to be clearly superior.113,114 One decision analysis suggested that primary anastomosis with a proxi-

mal defunctioning stoma, rather than primary resection and anastomosis or the Hartmann procedure, resulted in the highest quality of life overall.115 The primary anastomosis with a proximal defunctioning stoma group had higher stoma-reversal rates than the rate of reanastomosis in Hartmann procedure patients and a lower complication rate than the primary resection and anastomosis patients. Practically, the decision of whether to perform a primary or secondary resection is made intraoperatively based on the extent of disease, the difficulty of bowel mobilization, the degree of peritoneal contamination, and the surgeon’s expertise. In most cases of free perforation, at least two separate operations are necessary regardless of whether primary or secondary resection is performed. As noted previously, in many cases, restoration of the anastomosis is not possible and the colostomy is left indefinitely. Cases of peritonitis also require pelvic drainage, clearance of the rectum of fecal material when possible, and mobilization of the splenic flexure to perform a tension-free anastomosis. Placement of ureteral stents prophylactically can help prevent accidental ureteral injury during the operation.110 With the increasing use of the laparoscopic approach to uncomplicated diverticular disease, some centers have extended this approach to perforated disease with promising results.116,117

Fistula

When a diverticular phlegmon or abscess extends or ruptures into an adjacent organ, a fistula results. Fistulas are believed to develop in fewer than 5% of patients with diverticulitis but are present in about 20% of those who require surgery for diverticulitis.118 In a Cleveland Clinic review of 84 patients seen over 26 years with internal fistulas caused by diverticular disease, 65% were colovesical.119 There was a 2 : 1 male predominance, attributed to the protection of the bladder in women by the uterus. In one series, pneumaturia was present in 57% and fecaluria in 42%120; the latter is pathognomonic for colovesical fistula. Cystoscopy, cystography, and barium enema can be useful, although demonstration of the actual fistula often is difficult. In both of these series, single-stage operative resection with fistula closure and primary anastomosis could be performed in 75% of patients. Presumably, single-stage management is possible so often in the presence of fistulas to extracolonic organs because the fistula has effectively decompressed the inflammatory process. Colovaginal fistulas are the next most common internal fistula after colovesical fistula, representing approximately 25% of all cases.119 The passage of stool or flatus per the vagina is pathognomonic. Frequent vaginal infections or copious or feculent vaginal discharge should prompt consideration of this complication. Many patients with colovaginal fistula have undergone a previous hysterectomy. Treatment is surgical resection of the diseased segment of colon with repair of the contiguous organ, which generally can be performed as a single-stage procedure.98 Coloenteric, colouterine, and coloureteral fistulas occur much less commonly than colovesical and colovaginal fistulas. Spontaneous colocutaneous fistulas are rare and more commonly follow surgical repair. Other rare fistula types that have been reported as a presumed or documented complication of diverticulitis include colocholecystal, coloappendiceal, and colosalpingal. Although diverticular disease is a common cause of fistula from the colon to adjacent organs, other conditions, including IBD, pancreatitis, radiation enteritis and colitis, and malignancy, also can cause fistulas, and these diagnoses must be considered when a fistula involving the colon is discovered.

Chapter 117  Diverticular Disease of the Colon Obstruction

Obstruction can accompany diverticular disease either acutely or chronically. During an attack of acute diverticulitis, partial colonic obstruction can occur because of luminal narrowing from the pericolic inflammation, compression from abscess formation, or both. Obstruction can be confirmed with a gentle water-soluble contrast enema in a patient not suspected to have free perforation,118 while simultaneously excluding an obstructing sigmoid neoplasm. A CT scan with oral and rectal contrast also can give useful information about diverticular obstruction while also assessing for extraluminal disease. Complete obstruction is unusual. Colonic ileus or pseudo-obstruction also can occur, as can small bowel obstruction if a loop of small intestine becomes incorporated into the inflammatory mass. These conditions usually improve with effective medical therapy including antibiotics, bowel rest, and nasogastric suction. Surgical intervention may be required for persistent obstruction from acute diverticulitis not responding to medical therapy. Ideally, a modified bowel preparation with gentle irrigation enemas or low-dose oral laxatives given over a period of a few days can be performed preoperatively,118 thereby allowing the possibility of primary anastomosis in some cases. In cases when bowel preparation is not possible, a Hartmann procedure usually is performed. Recurrent attacks of diverticulitis, which may be subclinical, can initiate chronic stricturing of the colonic wall without ongoing inflammation. In such cases, high-grade or complete obstruction can occur. A contrast enema can help distinguish benign stricture from neoplasm. Colonoscopy also plays an important diagnostic role, and one group investigating strictures with colonoscopy was able to distinguish a benign from malignant etiology in 67% of patients.121 Strictures in which malignancy cannot be excluded despite colonoscopic and radiologic examinations should be treated by surgical resection. A trial of endoscopic dilation therapy can reasonably be attempted in patients in whom neoplasm is believed to be sufficiently excluded and in whom acute diverticulitis is not a concern. Success rates for balloon dilation of benign colonic strictures have been reported in the 67% to 79% range.122,123 Colonic metal stents may have a role in treating obstruction complicating diverticular disease, particularly in providing temporary decompression to allow bowel preparation and subsequent single-stage resection without diversion.124

SPECIAL TOPICS RELATED TO DIVERTICULITIS RECURRENT DIVERTICULITIS

For patients who respond well to conservative therapy, the issues of likelihood of recurrence and elective prophylactic surgical resection arise. The risk of recurrent symptoms following an attack of acute diverticulitis has been reported to range from 7% to 45%,1,5,17,68 with half of second attacks occurring within one year. Historical evidence had suggested that recurrent attacks were less likely to respond to medical therapy and had a higher mortality rate.17,68 Predicated on the notion that diverticulitis was a progressive disease, elective resection had been traditionally recommended after two attacks of diverticulitis.98,99,125 This approach is now being challenged. Studies such as the 13-year experience of the Mayo Clinic have suggested that the risk of poor outcomes is not higher with recurrent diverticulitis (more than two attacks) than it is with the first one or two attacks.126 In fact, the mortality rate from diverticulitis in this cohort was higher in those

with no prior history of diverticulitis (10%) than in those who had had prior attacks (2.5%). This difference may be due to the higher rate of free perforation in initial attacks relative to subsequent attacks. Thus, multiple recurrences do not appear to predict less-favorable outcomes. Additionally, there are some emerging medical therapies for those with diverticulitis to attempt to reduce likelihood of future attacks. As noted in the discussion of treatment for SUDD, studies have begun to show a decrease in recurrence rates of diverticulitis and symptomatic diverticular disease in patients treated with mesalamine, rifaximin, probiotics, and combinations of these agents.58-63,65,127,128 Finally, a recent decision analysis predicted that performing colectomy after the fourth attack of diverticulitis rather than after the second attack would result in fewer deaths and colostomies while having a superior cost-effectiveness.129 Recognizing this trend, published guidelines on the decision to perform colectomy after an attack of diverticulitis are now advocating evaluation on a case-by-case basis rather than empirically performing elective surgery after the second attack.130 If the choice to operate is made, most patients report having a good functional outcome and low rates of recurrent disease after elective resection for recurrent diverticulitis.131 In considering an elective sigmoid resection, relevant variables include the severity and responsiveness of the attack(s), the general health of the patient, the risk to the patient of a subsequent attack, and the risk of the resection itself. The latter factor may be lessened by the increasing use of the laparoscopic approach in many patients. Additionally, up to 10% of patients have symptomatic recurrent diverticulitis after surgical resection, and reoperation may be required in 2% to 3%.68,98,131,132 Patients undergoing resection for diverticulitis have higher recurrence rates when the sigmoid colon is used for the distal resection margin, rather than the rectum133; it is recommended to resect the entire distal colon whenever possible, forming the distal anastomosis with the proximal rectum and the proximal anastomosis with a noninflamed portion of colon.68,98

THE YOUNG PATIENT

Diverticulitis is relatively uncommon in patients younger than 40 years (2% to 5% of all patients with diverticulitis1,17) but the incidence in this age group may be rising.134 Nevertheless, because diverticulitis is uncommon in younger patients, it is often missed or mistaken for other diagnoses, such as appendicitis or IBD. Like diverticulitis in older patients, the disease is mainly sigmoid in location, although one report has described a right-sided predominance in young Israelis.135 There seems to be a significant male predominance in young patients.1,136 Attacks often are more severe, and 40% to 88% of younger patients require urgent surgery during their initial attack; recurrence and complication rates are also higher than in older patients.1,136,137 When patients with acute diverticulitis are managed nonoperatively, youth is an independent risk factor for poor outcome in subsequent course,138 possibly due to delay in diagnosis. For these reasons, some authors have advocated elective segmental colectomy in a healthy young person after one well-documented episode of diverticulitis68,125,138; others have questioned this approach.98,135,139 In the largest series to date of young patients with diverticulitis, the authors suggested that the higher incidence of surgical management in these patients relative to their older counterparts was not due to worse outcomes but rather to a higher rate of elective procedures done to prevent the expected poor outcomes.134 Thus, the latest surgical guidelines are advocating more of a case-by-case approach to elective resections for all patients with diverticulitis.130

2083

2084

Section X  Small and Large Intestine THE ELDERLY PATIENT

Because diverticular disease is more prevalent in older populations, diverticulitis in the elderly warrants special mention. Diverticulitis can manifest with more subtle symptoms and signs in the elderly, making the diagnosis more challenging. Distinguishing diverticulitis from colorectal cancer becomes a much more important issue in older patients, because the incidence of colorectal cancer also increases with age. Though the risk of a severe initial attack of diverticulitis may be lower in older patients than in younger ones, the risk of death when diverticulitis results in perforation is more than three times higher in the elderly.140 However, it appears that the number of comorbidities is a stronger predictor of mortality from perforated diverticulitis than age.141 Because subsequent complicated disease is uncommon if the initial attack was mild,138 and because older patients are less likely than younger patients to have recurrent disease after nonoperative management,142 surgery is often delayed or set aside in older patients. Not only is surgery more risky in the elderly, but colostomy reversal is also less often performed143 and, when attempted, tends to have a higher rate of morbidity and mortality than in younger patients.144

THE IMMUNOCOMPROMISED PATIENT

Diverticulitis can manifest more subtly in immunocompromised patients and represents a more difficult diagnostic challenge than in those with a normal immune system. Although diverticulitis in such patients does not appear to be more common, it appears to have graver consequences. One study reported that 24% of nonimmunosuppressed patients needed surgery for diverticulitis, whereas 100% of immunosuppressed ones required surgery.145 Immunocompromised patients have a higher rate of free perforation (43% vs. 14%), need for surgery (58% vs. 33%), and postoperative mortality (39% vs. 2%) than do noncompromised patients.146 In solid organ (e.g., heart, lung, kidney) transplant populations, mortality from diverticulitis has been found to be extremely high, ranging from 25% to 100%.147-149 Because of this high risk, many authorities advocate elective resection after an initial episode of diverticulitis in an immunosuppressed patient.7,125

RIGHT-SIDED DIVERTICULITIS

In Western countries, diverticulitis of the ascending colon or cecum is uncommon, because of the relatively low prevalence of diverticula in these portions of the colon. Nonetheless, right-sided diverticulitis should be part of the differential diagnosis for any patient with right-sided abdominal symptoms. In Asia, right-sided diverticulitis is the predominant form of diverticulitis. Especially in younger patients, the diagnosis of right-sided diverticulitis is more difficult to make than is left-sided disease and is virtually indistinguishable clinically from acute appendicitis. Clinical factors that might be helpful to distinguish diverticulitis from appendicitis in a person of Asian ethnicity who might therefore have rightsided diverticulitis are that patients with diverticulitis tend to be older and have a lower frequency of nausea and vomiting than patients with appendicitis; the characteristic progression of symptoms seen with appendicitis is absent150 (see Chapter 116). Radiologically, right-sided diverticulitis and appendicitis also are easily confused, especially when a local abscess is present (Fig. 117-6). There is an estimated preoperative misdiagnosis rate in right-sided colon inflammatory con­ ditions of 40% to 92%.150,151 Even with excellent imaging, the diagnosis of right-sided diverticulitis often is made at laparotomy.

Figure 117-6.  Computed tomography scan of a patient with a right lower quadrant abscess (arrow). The differential diagnosis of this finding includes right-sided diverticulitis and appendicitis.

When the proper diagnosis is made preoperatively, treatment of right-sided disease is the same as for left-sided diverticulitis. One study has suggested better overall responsiveness of right-sided diverticulitis to medical therapy alone, even after multiple attacks.152 The much more common complication associated with right colonic diverticula is hemorrhage, discussed later in this chapter.

SEGMENTAL COLITIS ASSOCIATED WITH DIVERTICULOSIS

A subset of patients with diverticulosis is found to have mucosal inflammation within the segment of colon containing the diverticula. Segmental colitis associated with diverticulosis (SCAD) initially was thought to be a rare form of Crohn’s colitis, but it is now increasingly recognized as a distinct, but poorly understood, manifestation of diverticular disease.153-155 SCAD primarily affects the sigmoid colon of certain patients with diverticulosis and mimics IBD in its clinical presentation. In limited study, there does not appear to be a higher risk of diverticulitis or colon cancer in these patients.155 Presenting symptoms generally are left lower quadrant cramping pain, diarrhea, and rectal bleeding. Endoscopically, in addition to diverticula, the sigmoid colonic mucosa shows erythema, friability of varying degrees, and mucosal erosions. The remaining segments of colon, including the rectum, are not visibly or histologically involved. Biopsy specimens can show chronic lymphocytic infiltration, cryptitis, crypt abscesses, and even granulomas; many cases are histologically indistinguishable from IBD.156 In fact, a subset of patients when followed endoscopically through time appear to evolve into a picture similar to ulcerative proctosigmoiditis or Crohn’s colitis. This observation should prompt a low threshold for endoscopic re-evaluation if a patient with presumed SCAD develops persistent or progressive symptoms. Patients with SCAD generally are responsive to 5aminosalicylic acid (5-ASA) compounds, with one series reporting over 80% of patients achieving clinical remission on 5-ASA.155 In most cases, the clinical course tends to be benign and self-limited,157 although there are reports of cases requiring sigmoid colectomy for bleeding or stricture complications.158

Chapter 117  Diverticular Disease of the Colon DIVERTICULAR HEMORRHAGE Diverticulosis, angioectasias, colitis, neoplasms, and hemorrhoids are responsible for most cases of lower gastrointestinal bleeding in adults (see Chapter 19).20,159-162 Diverticular hemorrhage is the most common identifiable cause of significant lower gastrointestinal bleeding, accounting for 30% to 40% of cases with confirmed sources.163,164 It often is difficult, however, to make a precise determination of the source of bleeding, and conclusive evidence of the cause of bleeding is available only in a minority of cases,159,160 either by seeing an actively bleeding lesion endoscopically or by identifying extravasation from a specific site angio­ graphically. More often, circumstantial evidence is used to suggest diverticulosis as the source of hemorrhage. Clinical features suggesting diverticular hemorrhage include copious bright red or maroon blood per rectum, the presence of diverticulosis on colonoscopy or radiologic studies, exclusion of an upper gastrointestinal source, and exclusion of alternative colonic sources. These clinical criteria, suggested by Quinn in 1960, however, are nonspecific and are met by various lesions other than diverticula, most notably angioectasias.

EPIDEMIOLOGY

Severe hemorrhage has been reported to occur in 3% to 5% of patients with diverticulosis.20,165,166 Although most diverticula are in the left colon in Western patients, the site of bleeding diverticula has been believed to be in the proximal colon in more than one half of patients.167-169 A large series of 180 Asian patients with diverticular hemorrhage reported a higher bleeding rate and greater need for surgery with right-sided compared with left-sided disease.169 Patients with pancolonic diverticulosis appear to have a higher hemorrhage rate than those with diverticula on only one side of the colon.170

PATHOPHYSIOLOGY

To study the pathogenesis of diverticular bleeding, Meyers and colleagues used sophisticated microangiographic techniques on resected colon specimens from patients with arteriographically documented bleeding diverticula.171 Three-dimensional histologic reconstructions demonstrated consistent findings of intimal thickening and medial thinning of the vasa recta as it coursed over the dome of the diverticulum. They proposed that these changes led to a segmental weakening of the artery, thus predisposing to its rupture. This arterial lesion was absent in diverticula that did not bleed. What predisposes to this arterial change and what precipitates its rupture are unknown. Inflammation does not appear to be a contributing factor, because it is not found histologically in bleeding diverticula that were resected. This absence might explain why bleeding rarely complicates diverticulitis. Nonsteroidal anti-inflammatory drugs (NSAIDs) have been implicated in lower intestinal, and specifically diverticular, bleeding. A large prospective series of patients with lower intestinal bleeding reported a risk of bleeding with NSAIDs equal to that of duodenal ulcer.172 An overall increased risk of diverticular bleeding also was found at the four-year follow-up of a large study of health professionals who had been free of diverticulosis at baseline.173 Both aspirin and nonaspirin NSAIDs increase bleeding risk, although the magnitude of this risk appears to be higher with nonaspirin NSAIDs than with aspirin.170 In addition to having a higher rate of complications from diverticulosis, including bleeding and perforation, patients taking NSAIDs also appear to have more severe complications.174-176 Hence,

the necessity for NSAIDs in patients who have an episode of diverticular bleeding should be re-evaluated. Whether patients with diverticulosis without a prior bleeding episode should be counseled as well to avoid NSAIDs is less clear. In a multicenter trial of patients randomized to naproxen (conventional NSAID) or rofecoxib (cyclooxygenase [COX]-2 selective agent), the relative risk of lower intestinal bleeding from all causes was 0.46 in patients taking rofecoxib relative to the naproxen group.177 Whether hemorrhage specifically from diverticula is reduced by replacing NSAIDs with COX-2 selective agents is unknown, and the cost-effectiveness of this approach has not been established.

CLINICAL FEATURES

Diverticular hemorrhage typically manifests as abrupt, painless hematochezia. Because the bleeding is arterial, the volume of blood usually is moderate or large, an observation that can help distinguish diverticular hemorrhage from other common causes of rectal bleeding. Patients often pass red or maroon clots; melena is unusual. Because the bleeding is overt, neither a positive fecal occult blood test nor iron-deficiency anemia should be attributed to diverticular hemorrhage. Natural history studies report that bleeding ceases spontaneously in 70% to 80% of patients, and rebleeding rates range from 22% to 38%.165,166,169 The chance of a third bleeding episode can be as high as 50%,165 leading some to recommend surgical resection after a second bleeding episode.101

DIAGNOSIS AND TREATMENT

The diagnosis and treatment of patients with lower intestinal bleeding in general have been reviewed comprehensively elsewhere (see Chapter 19).130,178 The algorithm in Figure 117-7 summarizes the management of patients with bleeding diverticula. In unstable patients, volume and blood product resuscitation require immediate attention. Excluding an upper gastrointestinal source by nasogastric lavage or upper endoscopy is warranted because 10% to 15% of patients with hematochezia have an upper tract cause. If bleeding is massive or if the patient remains unstable after attempted resuscitation, early angiography to attempt bleeding localization and surgical consultation should be obtained. A stable patient with suspected active or recent diverticular bleeding should undergo bowel preparation for a co­ lonoscopy. The ability to identify a diverticular source, to exclude alternative diagnoses, and to provide therapy of actively bleeding lesions support colonoscopy as a primary investigation in this setting. Rapid (over three to four hours) oral or nasogastric purge with a balanced electrolyte solution provides a safe and effective bowel preparation; a dose of metoclopramide before initiation can improve tolerance to the lavage.179 If diverticulosis is not found on colonoscopy, alternative diagnoses should be entertained. If diverticula are found but bleeding has stopped and no other colonic causes are found, a presumptive diagnosis of diverticular hemorrhage is made and the patient should be instructed to avoid NSAIDs and anticoagulants, if possible. As noted, most patients with diverticular hemorrhage do not rebleed. The endoscopic identification of active bleeding or stigmata of recent hemorrhage from a specific site (Fig. 117-8) is evident in about 10% to 20% of colonoscopic examinations for diverticular bleeding. The appearance of nonbleeding stigmata—visible vessel or adherent clot within a diverticulum—can permit one to estimate the future risk of hemorrhage from that site,180 although this concept has not been validated as well in the colon as it has for peptic ulcer bleeding. Identification of a bleeding site allows endoscopic

2085

2086

Section X  Small and Large Intestine Patient with suspected diverticular hemorrhage Hemodynamically stable

Patient stabilizes

Expeditious colonoscopy Diverticulosis Consider alternative causes of GI bleeding (see Chapter 19)

Hemodynamically unstable Resuscitation Treat coagulopathy Exclude UGI source by NG aspirate ± EGD

Diverticulosis

(source not confirmed) 

Patient remains unstable, actively bleeding

Angio 

Stigmata of active or recent bleeding

Urgent angiogram Surgical consultation



Angio 

Endoscopic therapy Bleeding stops Discontinue NSAIDs if applicable Observe for rebleeding

Superselective embolization

Bleeding persists

Successful

Tagged RBC scan 

Bleeding stops Supportive care Blood transfusions





Bleeding continues Subtotal colectomy

Brisk bleed, patient unstable, ongoing transfusion requirement



Angiogram Surgical consultation 

Slow bleed, patient stable



Discontinue NSAIDs if applicable Observe for rebleeding

Unsuccessful

Bleeding stops

Supportive care Blood transfusions

Bleeding continues, ongoing transfusion requirement Segmental colectomy

Repeat tagged RBC scan

Figure 117-7.  Algorithim for the management of patients with bleeding colonic diverticula. Angio, angiogram; EGD, esophagogastroduodenoscopy; GI, gastrointestinal; NSAIDs, nonsteroidal anti-inflammatory drugs; NG, nasogastric; RBC, red blood cell; UGI, upper gastrointestinal.

A

B

Figure 117-8.  Colonoscopy in a patient with lower intestinal bleeding from a diverticulum in whom a site of active bleeding was identified (A) and treated successfully with placement of two hemoclips (B). (Courtesy of Janak Shah.)

Chapter 117  Diverticular Disease of the Colon therapy to be applied. The use of epinephrine injection alone181 or in combination with other therapies such as heater probe coagulation,182 bipolar coagulation,183-185 endoclips,186-188 fibrin sealant,189 and band ligation190 all have been shown in small case series to achieve hemostasis safely in patients with diverticular bleeding.177 If endoscopic therapy is not effective or durable, localizing the site facilitates directed therapy with angiography or segmental surgical resection. With a paucity of high-quality evidence, the true effectiveness of endoscopic treatment for diverticular hemorrhage is not known. In a retrospective study, Jensen and colleagues were able to identify and treat (with epinephrine and bipolar cautery) definite bleeding sources in 10 of 48 patients with suspected diverticular bleeding.185 None of the endoscopically treated patients had recurrent bleeding or required surgery. These results were compared with 17 historical controls who had bleeding stigmata but no endoscopic therapy, nine of whom rebled and six of whom required surgery, lending indirect support to the use of endoscopic therapy for diverticular hemorrhage. Stigmata of active or recent hemorrhage are not common findings at endoscopy, however, and, therefore, endoscopic therapy only is relevant to a minority of patients. Another group at the Mayo Clinic with an approach to lower intestinal bleeding similar to Jensen’s group reported

that urgent colonoscopy (performed less than 12 hours after admission) was no more effective at identifying bleeding stigmata than colonoscopy performed later in the hospitalization.191 Another cohort of 100 patients with acute lower intestinal bleeding was randomized to urgent colonoscopy (within eight hours of hospitalization) or standard care (radiologic bleeding studies for rapid bleeds, elective co­ lonoscopy for slow or inactive bleeds). A definite bleeding source was found in more cases in the urgent colonoscopy group than in those who received standard care, although important outcomes such as mortality, hospital stay, rebleeding, and need for surgery were the same between groups.192 Despite having little firm evidence that expedited colonoscopy improves outcomes, it still seems reasonable to perform colonoscopy within the first 12 to 48 hours of admission in most cases. When active bleeding is present but colonoscopy fails to allow localization or treatment of a bleeding source, further evaluation with nuclear scintigraphy (tagged red blood cell scan) or angiography can be undertaken, taking into account local availability and expertise. A third radiographic option, enhanced CT, also has shown promise for identifying active lower intestinal bleeding sources, although its role has yet to be defined.193 Nuclear scintigraphy (Fig. 117-9) has many theoretical advantages in the evaluation of lower intestinal bleeding194: It is noninvasive, technically simple, relatively

A

B

C

D

Figure 117-9.  Nuclear scintigraphy scan in a patient with lower intestinal bleeding from right-sided diverticulosis. Early scan (A) shows pooling of radiolabeled blood in the right upper quadrant corresponding to a bleeding site in the hepatic flexure. Subsequent images (B, C, D) show radiolabeled blood progressing through the transverse colon over time.

2087

2088

Section X  Small and Large Intestine

A

B

Figure 117-10.  Selective angiography of the superior mesenteric artery in a patient with lower intestinal bleeding from diverticulosis. The bleeding site is identified in the ascending colon by a “blush” of contrast material (A, arrow). The bleeding site is embolized with microcoils, which are seen in the bleeding vessel (B, arrow), and the extravasation of contrast material has stopped.

inexpensive, and sensitive to bleeding rates as low as 0.1 mL/min. Scintigraphy, however, can identify only the site, not the etiology of the bleeding, and it has no therapeutic potential. Furthermore, some studies have questioned its accuracy195 and have underscored a lack of proven impact on mortality, transfusion requirement, and eventual need for surgery.196 Given its sensitivity and relative simplicity, however, many centers use scintigraphy before angiography to minimize the chance of a negative angiogram and to help select a specific artery for injection of contrast.197 Angiography has a sensitivity for lower intestinal bleeding rates of 0.5 mL/min. Advantages of angiography are identification of the site of bleeding accurately enough to direct segmental surgical resection and therapeutic cap­ ability. Although angiographic embolization had previously been felt to carry a risk of bowel infarction, superselective embolization (Fig. 117-10), or embolization of distal arterial branches, has been demonstrated to be effective (67% to 100% with lasting hemostasis) and relatively safe (ischemia rates less than 20%).198-201 Where available, arterial embolization increasingly is becoming the nonsurgical therapy of choice when endoscopic methods are not possible. Surgery for lower intestinal bleeding usually is avoided unless endoscopic or angiographic therapies are unavailable or fail. Because diverticular bleeding stops spontaneously in most patients, surgical management is required infrequently. The primary indications for operative management are large transfusion requirements, recurrent hemorrhage that is refractory or not amenable to therapy, or hemo­ dynamic instability unresponsive to resuscitation. When surgery is necessary, a partial colectomy is preferred to a subtotal colectomy whenever possible. Segmental resection can be performed if the bleeding site is clearly identified from a therapeutically unsuccessful angiographic or endoscopic procedure or when the extent of diverticulosis is proven to be confined to a specific segment of the colon. The rebleeding rate was 6% in seven series of patients who underwent segmental resections for angiographically

documented bleeding sites.202 As noted, however, it is often difficult to identify an extravasation site angiographically. One series found that angiography resulted in a successful directed resection in only 12% of patients.203 In patients with persistent, life-threatening bleeding and no identification of a likely bleeding site, a subtotal or blind colectomy may be required as a last resort. These patients have had an extremely high morbidity and mortality,204 possibly because of the multiple invasive tests leading to that end and the resultant delay in definitive management. Additionally, a blind colectomy runs the risk of not resecting the bleeding lesion when the source is more proximal and has a high risk of anastomotic failure.205 More recent literature, however, has shown morbidity and mortality rates of a subtotal colectomy not to differ from those of a blind hemicolectomy, when the site of the bleeding is not identified.205,206 Clearly, a close collaborative relationship between the gastroenterologist and the surgeon is paramount in managing such patients.

ACKNOWLEDGMENT

Dr. Stollman gratefully acknowledges the contributions of Jeffrey B. Raskin, MD, who inspired his interest in this field and has provided ongoing guidance, mentoring, and insight.

KEY REFERENCES

Aldoori WH, Giovannucci EL, Rockett HR, et al. A prospective study of dietary fiber types and symptomatic diverticular disease in men. J Nutr 1998; 128:714-19. (Ref 44.) Burkitt DP, Walker AR, Painter NS. Effect of dietary fibre on stools and the transit-times, and its role in the causation of disease. Lancet 1972; 2:1408-12. (Ref 39.) Delvaux M. Diverticular disease of the colon in Europe: epidemiology, impact on citizen health and prevention. Aliment Pharmacol Ther 2003; 18(Suppl 3):71-4. (Ref 8.) Hjern F, Johansson C, Mellgren A, et al. Diverticular disease and migration—the influence of acculturation to a Western lifestyle on diverticular disease. Aliment Pharmacol Ther 2006; 23:797-805. (Ref 14.)

Chapter 117  Diverticular Disease of the Colon Jensen DM, Machicado GA, Jutabha R, Kovacs TO. Urgent colonoscopy for the diagnosis and treatment of severe diverticular hemorrhage. N Engl J Med 2000; 342:78-82. (Ref 185.) Laine L, Connors LG, Reicin A, et al. Serious lower gastrointestinal clinical events with nonselective NSAID or coxib use. Gastroenterology 2003; 124:288-92. (Ref 177.) Longstreth GF. Epidemiology and outcome of patients hospitalized with acute lower gastrointestinal hemorrhage: a population-based study. Am J Gastroenterol 1997; 92:419-24. (Ref 163.) Painter NS, Burkitt DP. Diverticular disease of the colon: a deficiency disease of Western civilization. Br Med J 1971; 2:450-4. (Ref 2.) Painter NS, Truelove SC, Ardran GM, Tuckey M. Segmentation and the localization of intraluminal pressures in the human colon, with special reference to the pathogenesis of colonic diverticula. Gastroenterology 1965; 49:169-77. (Ref 29.) Parks TG. Natural history of diverticular disease of the colon: a review of 521 cases. BMJ 1969; 4:639-42. (Ref 17.)

Parks TG. Natural history of diverticular disease of the colon. Clin Gastroenterol 1975; 4:53-69. (Ref 9.) Rafferty J, Shellito P, Hyman NH, et al. Practice parameters for sigmoid diverticulitis. Dis Colon Rectum 2006; 49:939-44. (Ref 130.) Salem L, Veenstra DL, Sullivan SD, et al. The timing of elective colectomy in diverticulitis: a decision analysis. J Am Coll Surg 2004; 199:904-12. (Ref 129.) Stemmermann GN, Yatani R. Diverticulosis and polyps of the large intestine: a necropsy study of Hawaii Japanese. Cancer 1973; 31:126070. (Ref 13.) Strate LL, Liu YL, Syngal S, et al. Nut, corn, and popcorn consumption and the incidence of diverticular disease. JAMA 2008; 300:907-14. (Ref 45.) Full references for this chapter can be found on www.expertconsult.com.

2089

CHAPTE R

118 Irritable Bowel Syndrome Nicholas J. Talley

CHAPTER OUTLINE Definitions  2091 Clinical Features  2092 History  2092 Physical Examination  2093 Epidemiology  2093 Prevalence  2093 Gender and Race  2093 Subgroups  2093 Incidence and Disappearance of Symptoms  2094 Risk Factors  2094 Health Care Seeking  2094 Excess Abdominal Surgery  2094 Impact on Quality of Life and Costs  2094 Pathophysiology  2095 Altered Colonic and Small Bowel Motility  2095 Visceral Hypersensitivity  2095

Irritable bowel syndrome (IBS) is an important disease entity because of its high prevalence, substantial morbidity, and enormous costs.1-3 In the United States, approximately 12% of patients seen by primary care physicians have IBS, but it is likely that this frequency is an underestimate.4-6 In gastrointestinal practices, more than one third of patients have functional gastrointestinal disorders, IBS being the most common diagnosis.7 Because a substantial proportion of gastroenterology practice comprises patients with IBS or other functional gastrointestinal disorders, it is essential that clinicians develop expertise in their diagnosis and treatment. The diagnosis of IBS rests on making a positive clinical diagnosis from the history; that tests often are not needed represents an important conceptual advance.8 There is increasing evidence that at least a subset of IBS has an organic basis in the gastrointestinal tract.9 Nonetheless, only symptom-directed therapy rather than disease-modifying treatments are available; the evidence base for current therapy has strengthened considerably with the publication of well-performed meta-analyses. In this chapter, current knowledge of the epidemiology and pathophysiology of IBS is reviewed to provide a rational basis for its diagnosis and therapy.

DEFINITIONS IBS is characterized by the presence of abdominal discomfort or pain associated with disturbed defecation.3 Bloating or visible abdominal distention often is present in patients

Abnormal Gas Propulsion and Expulsion  2096 Local Inflammation  2097 Role of Food  2098 Abnormal Colonic Flora and Small Intestinal Bacterial Overgrowth  2098 Central Dysregulation  2098 Psychological Factors  2098 Genetic Factors  2099 Diagnosis  2099 Treatment  2100 Education and Support  2100 Diet  2101 Medication  2102 Psychological Treatments  2103 Alternative Treatments  2103 Prognosis  2103

with IBS but are not considered essential symptoms for diagnosis.3 Furthermore, individual symptoms are neither sensitive nor specific enough on their own to diagnose IBS.10 In a classic study from the United Kingdom, Manning and associates first reported that six symptoms were more common in patients in whom IBS was subsequently documented, although only four were statistically significant in the initial report (Table 118-1).11 Later studies showed that these symptoms were specific, but not sensitive, for identifying IBS and were of greater diagnostic value in women.10,12 The Kruis scoring system is based on the presence and duration of symptoms, negative physical examination findings, and normal simple laboratory tests, and it has modest diagnostic utility (see Table 118-1).13 In an effort to build on the diagnostic utility of the Manning and Kruis criteria, the Rome (I, II, and III) criteria were created following a formal consensus process to provide a standard for clinical research (see Table 118-1).3 The Rome criteria are useful in clinical practice and can be used to make a positive clinical diagnosis.1,3 The sensitivity and specificity of the Rome I criteria have been reported to be 71% and 85%, respectively,14 and although adequate validation data for Rome III are lacking, the main criteria for Rome II and III are very similar.10 Comparisons of the criteria have shown that Rome I and II criteria are specific and identify similar patient populations, although, compared with Rome I criteria, the Rome II criteria appear to identify fewer cases in some studies.15-17 The Manning criteria identify additional patients with IBS-like symptoms who do not fulfill any of the Rome criteria but arguably also should be classified as having true IBS.15-17

2091

2092

Section X  Small and Large Intestine Table 118-1  Manning, Kruis, and Rome III Criteria for Irritable Bowel Syndrome Manning Criteria* Abdominal pain eased after bowel movement Looser stools at onset of abdominal pain More frequent bowel movements at onset of abdominal pain Abdominal distention Mucus per rectum Feeling of incomplete emptying Kruis Criteria Patient’s History Abdominal pain Flatulence Irregularity of bowel movements Symptoms more than 2 years Mixed diarrhea and constipation Pellet-like stools or mucus Physician’s Assessment† Abnormal physical findings Erythrocyte sedimentation rate >20 mm/2 hr Leukocytosis (>10,000 cm3) Hemoglobin (female <12 g/dL; male <14 g/dL) History of blood in stool Rome III Criteria‡ Recurrent abdominal pain or discomfort§ at least three days/month in the last three months associated with two or more of the following: Improvement with defecation Onset associated with a change in frequency of stool Onset associated with a change in form (appearance) of stool *Diagnostic cut-off: three or more of the six symptoms listed. † If any abnormal physical findings or any of the laboratory parameters assessed by the physician are present, IBS is excluded. ‡ Criteria fulfilled for the previous three months, with symptom onset at least six months before diagnosis. § “Discomfort” means an uncomfortable sensation not described as pain. In pathophysiology research and clinical trials, a pain or discomfort frequency of at least two days a week during screening evaluation is recommended for subject eligibility. Manning criteria adapted from Manning AP, Thompson WG, Heaton KW, Morris AF. Towards positive diagnosis of the irritable bowel. BMJ 1978; 2:653-4. Kruis criteria adapted from Kruis W, Thieme C, Weinzierl M, et al. A diagnostic score for the irritable bowel syndrome. Its value in the exclusion of organic disease. Gastroenterology 1984; 87:1-7. Rome III criteria from Longstreth GF, Thompson WG, Chey WD, et al. Functional bowel disorders. In: Drossman DA, editor. Rome III: The Functional Gastrointestinal Disorders. 3rd edition. McLean, VA: Dagnon Associates. 2006, pg 491. Used with permission from the Rome Foundation.

CLINICAL FEATURES HISTORY Abdominal Discomfort or Pain

IBS should not be diagnosed in the absence of abdominal discomfort or pain.3 Distinguishing discomfort from pain can be problematic for both the patient and physician, however, because of the strong influence of cultural issues. In the United States, what a physician might label as mild pain often is considered discomfort by the patient. The pain or discomfort in IBS typically is relieved by defecation, or its onset is associated with an increase or decrease in stool frequency, or looser or harder stools. The pain often is poorly localized, waxes and wanes, may be aggravated by eating, and can occur in any part of the abdomen, although it more typically is located in the lower abdomen; it may be referred to different areas in the abdomen or to the chest or back. Exacerbation of pain by life events or difficult life situations is common. Abdominal discomfort or pain that is continuous or unrelated to defecation or induced by menstruation, urination, or physical activity is unlikely to be caused by IBS.

Constipation and Diarrhea

Patients with IBS experience constipation, diarrhea, or a mixture of these symptoms.3 Symptom predominance has led some authors to attempt to classify IBS patients by their predominant symptom: constipation (IBS-C), diarrhea (IBSD), and mixed (IBS-M), although these symptoms often are variable and intermittent and patients can change their bowel habits from one pattern to another (see later). An irregular stool consistency (abnormal stool form) is characteristic. The terms “constipation” and “diarrhea” can reflect a wide variety of different symptom experiences to different patients, and so whenever a patient uses these terms, an exploration of their meaning is required.18 Any combination of infrequent defecation, passage of hard stools, excessive straining, feelings of incomplete rectal evacuation, or rectal discomfort may be referred to as constipation, whereas increased stool frequency, urgency, or the passage of liquid or watery stools, or even more frequent small hard stools, may be referred to as diarrhea by the patient. Stool form can be measured objectively and graded by patient or physician; the Bristol stool form scale (Fig. 118-1) now is routinely used in clinical trials, and changes in stool form (at the extreme ends of the scale) roughly correlate with colonic transit time.19,20

Bloating and Visible Distention

A feeling of bloating is almost universal in patients with IBS, and its site can be difficult for the patient to localize. Visible abdominal distention is characteristic but less common21; it can be objectively measured and usually is not imagined.22 Gas can mean excess bloating, belching, flatus, or even reflux symptoms to the patient, and so it is important to ask patients to explain the meaning of the terms they are using to describe their symptoms.

Noncolonic Symptoms

Other clinical features can help support the diagnosis of IBS but in themselves are not diagnostic. Nausea is common, and at least one third of patients with IBS have epigastric discomfort or pain (dyspepsia).1,8,23 Gastroesophageal reflux disease (GERD) occurs more commonly in IBS than would be expected by chance, affecting up to one in three persons with IBS.24 Extraintestinal symptoms including headache (and migraine), backache, impaired sleep, chronic fatigue, increased urinary frequency or urgency, pelvic pain, and dyspareunia are more common in patients with IBS but have no accepted diagnostic value.1,25 Musculoskeletal pain syndromes including fibromyalgia23 and temporomandibular joint disorder also are associated with IBS.2,23

Inflammatory Bowel Disease and Irritable Bowel Syndrome

Typical IBS symptoms are common in patients with documented inflammatory bowel disease (IBD) in remission; in one study, 33% with ulcerative colitis and 42% with Crohn’s disease fulfilled Rome II criteria for IBS.26 Clinically these conditions can be difficult to distinguish. IBS symptoms appear to be more prevalent before a diagnosis of IBD is made.27

Chronicity

For a confident diagnosis of IBS, symptoms should have been present for at least six months3; IBS may accompany other chronic disorders. For example, IBS is present in one third or more of patients with IBD in remission.26 A number of different conditions can cause transient bowel symptoms including pregnancy, dietary indiscretion, food poisoning, traveler’s diarrhea, bed rest, weight loss, and acute stress

Chapter 118  Irritable Bowel Syndrome Bristol Stool Scale Form Rome III IBS Subtypes

Type I

100 25% of BM is the threshold for classification

Type II

75

Type IV

Type V

Type VI

% BM hard or lumpy

Type III

Type I Type I Type II

50

Type VI

IBS-C Types I,II

25

IBS-M

IBS-D Types VI,VII

IBS-U Type VII

0

0

25

Type VI

50 % BM loose or watery

Type VII

75

100

Figure 118-1.  The Bristol stool form scale and classification of subtypes of IBS. BM, bowel movement; C, constipation; D, diarrhea; IBS, irritable bowel syndrome; M, mixed. (Adapted from Longstreth GF, Thompson WG, Chey WD, et al. Functional bowel disorders. In: Drossman DA, editor. Rome III: The Functional Gastrointestinal Disorders. 3rd edition. McLean, VA: Dagnon Associates. 2006; pg 492. Used with permission from the Rome Foundation.)

(nervous diarrhea); these must be distinguished from the chronic, recurrent symptoms of IBS.

PHYSICAL EXAMINATION

The physical examination in IBS usually is normal, although deep tenderness over the colon may be appreciated.10 Abdominal wall pain should be excluded clinically. Tensing the abdominal wall by flexing the chin on the chest or sitting up partially lessens tenderness that is caused by an intraabdominal process. If tensing the abdominal wall muscles increases abdominal tenderness, a point of localized abdominal wall tenderness should be sought with a probing finger (Carnett’s test); identification of such a point might enable the tenderness to be treated with an injection of lidocaine and triamcinolone.28,29 The painful rib syndrome (point tenderness on springing the rib cage) also may be confused with IBS pain.30 Ovarian cancer needs to be considered in any middle-aged or older woman presenting with newonset IBS-like symptoms.31 A pelvic examination therefore may be relevant to determine if there is any irregular, fixed pelvic mass.

EPIDEMIOLOGY IBS is a common disorder all over the world.1,17,32 Epidemiologic studies have defined the prevalence and identified potential risk factors for IBS.

PREVALENCE

Prevalence estimates for IBS have varied anywhere from 3% to 20% in the United States, with similar results reported elsewhere; however, prevalence estimates are influenced substantially by the definition applied. For example, in Olmsted County, Minnesota, the prevalence of IBS varied from 8% to 22% depending on the criteria used.33 Younger people have a higher prevalence of IBS in the community. Generally, it is believed that IBS is uncommon

in the elderly, but population-based studies indicate that IBS increases with advancing age. Thus, for example, using three or more of the Manning criteria to define IBS, the prevalence of IBS in Olmsted County ranged from 8% in those 65 to 74 years of age to more than 12% in those older than 85 years.34 Obviously, organic disease is more prevalent in elderly persons and could account for some of the reported IBS-like symptoms, but it seems likely that IBS in the elderly is often underdiagnosed or misdiagnosed, for example, as diverticular disease.35

GENDER AND RACE

Gender-specific prevalence rates for IBS are approximately two female to one male in most studies, and all populationbased studies have reported a female predominance.17,36 Healthy women have greater rectal sensitivity, slower colonic transit, and smaller stool outputs than do men, which might explain why certain symptoms, such as straining and passage of hard stools, seem to be more common in women.37,38 In clinical practice in the United States, women outnumber men, which partly is explained by increased health care–seeking behavior among women; this appears to be culturally derived, because data from India indicate more men than women present for care of IBS in that part of the world.39 The prevalence of IBS generally is similar in whites and blacks, although some data have suggested it may be lower in Hispanics than in non-Hispanic whites in the United States.40,41 IBS is common in China, Japan, South America, and the Indian subcontinent. Indeed, IBS is common and its prevalence comparable in all countries where it has been studied.39,42-45

SUBGROUPS

Subdividing IBS based on the predominant symptom pattern is attempted commonly, but few data are available on IBS prevalence by symptom subgroup. Moreover, it is unclear if those with one predominant symptom—diarrhea or constipation—if followed long enough, eventually develop the

2093

2094

Section X  Small and Large Intestine other, namely, constipation in patients with IBS-D or diarrhea in patients with IBS-C; some data from primary care support this contention.46 The Rome III definition uses stool form to subclassify IBS, but definitions of IBS subgroups remain arbitrary, and different definitions have been used in different studies (see Fig. 118-1).3 Nonetheless, in a study from Olmsted County, Minnesota, 5.5% of the population had IBS-D and 5% IBS-C; both diarrhea and constipationtype symptoms occurred in 5% (IBS-M), and 4% did not meet strict criteria for either constipation or diarrhea.47 Another population-based study found higher rates of constipation in community subjects with IBS.48 In a study of 317 patients recruited for a clinical trial, at baseline 36% had IBS-D, 34% had IBS-C, and 31% had IBS-M49; more than 75% switched type over a year of follow-up, usually to and from the IBS-M pattern, but less than a third changed from IBS-D to IBS-C or vice versa.49

INCIDENCE AND DISAPPEARANCE OF SYMPTOMS

The onset rate of IBS was 67 per 1000 person-years by applying the Manning criteria to a cohort in Olmsted County that was surveyed at a 12- to 20-month interval.50 This study did not exclude people with a past history of IBS, however, and hence this is not the true incidence.50 Another study reported that the incidence of a clinical diagnosis of IBS in Olmsted County was 0.2% per year; this figure reflects the lower end of the incidence rate, because people with IBS symptoms who did not seek consultation could not be included in this calculation.51 Over a 10-year follow-up, 15% of community subjects free of baseline IBS symptoms developed the syndrome.32 In a follow-up study in Olmsted County, 38% of subjects meeting the definition of IBS at entry did not meet these criteria 12 to 20 months later50; they lost their symptoms. The actual prevalence of IBS did not change from year to year, however, because the disappearance of symptoms in some patients with IBS was balanced by others who developed IBS. Among those losing IBS in the community, there is a subset in whom symptoms evolve to reflect another functional gastrointestinal disorder52,53; hence, IBS usually is a chronic disorder, although symptoms often are variable.

RISK FACTORS The best-accepted risk factor for IBS is bacterial gastro­ enteritis.54-57 The risk of postinfection IBS has been reported to be increased with depression,58 adverse life events and hypochondriasis,59 female gender, younger age, and prolonged duration of diarrhea following the initial attack.60 Bacterial virulence factors also may be important,61 but IBS can follow nonbacterial enteritis, including Norovirus gastroenteritis, or infection with trichinella.62 Other risk factors for IBS include an affluent childhood environment,63 estrogen use, postmenopausal estrogen use,64 recent antibiotic use,65 food intolerance,66,67 extrain­testinal somatic symptoms,67 and poor quality of life.32 IBS runs in families,68 and low birth weight is also a risk factor for IBS, even after controlling for genetic influences.69 In contrast, oral glucocorticoid users may be at a lower risk for IBS.70 IBS is associated with an approximately three-fold increased risk of ischemic colitis71; however, a cause-and-effect relationship has not been established and the absolute risk remains very small (43 per 100,000 person-years).

HEALTH CARE SEEKING Understanding why a patient is presenting for care is important in terms of planning appropriate management strategies. Burden of illness studies estimate that there are 3.6 million physician visits for IBS in the United States annually.72 The rate of health care seeking for IBS may, in part, be affected by health care access; consulting rates in the United States have varied between 25% and 46%, but up to 40% of patients do not have easy access to health care in this country.73 In Australia, where health care access essentially is universal, consulting rates have been 73%.73 Drivers of health care seeking remain poorly documented.73 The severity and chronicity of symptoms, in particular abdominal pain, partly promote health care seeking.74,75 IBS patients are more concerned about their health and more fearful of illness, suggesting that anxiety about their illness may be another factor.73 Children of a parent with IBS may see physicians more often than those who do not have a parent with IBS.76 As adults, they also are more likely to report poorer health in childhood as well as having received greater parental attention and gifts or rewards for being ill, suggesting there may have been early childhood programming of abnormal illness behavior.77 Those who seek medical attention tend to be more disturbed psychologically than those who do not seek such consultation,78-81 and those who consult for IBS also are more likely to consult about relatively minor complaints as well as other nongastrointestinal somatic symptoms.73 There remain other unknown factors that must be important, however, because these psychological factors still seem to poorly explain observed health care–seeking rates.73

EXCESS ABDOMINAL SURGERY

There is evidence that patients with IBS are at risk for undergoing excess surgery.82-84 In a large health maintenance organization study, patients with IBS, compared with controls, reported having had more cholecystectomies (12% vs. 4%), appendectomies (21% vs. 12%), and hysterectomies (33% vs. 17%); IBS was associated independently with these operations.82 A full explanation for these findings is uncertain, but presumably, some of this excess surgery reflects misdiagnosis and inappropriate intervention.85 It also is possible that IBS could predispose to an excess of certain diseases that lead to surgery. For example, constipation is associated with an increased risk of gallstones,86 but whether this association applies to IBS-C is uncertain. Alternatively, a history of a biliary event—identification of gallstones or a cholecystectomy—is associated with an increased risk of new-onset IBS.87 Some surgeons continue to believe that patients with IBS-type symptoms respond favorably to intra-abdominal surgery, although the evidence is anecdotal and probably reflects a placebo response.85

IMPACT ON QUALITY OF LIFE AND COSTS

A systematic review concluded that there was good evidence for a decrease in health-related quality of life in patients with moderate to severe IBS88 and that the quality of life in IBS is impaired to a degree comparable with other chronic disorders such as depression or GERD. Rather than IBS causing impaired quality of life, an alternative explanation for this association is that poor quality of life predisposes to a higher risk of IBS, and some evidence supports this contention.32 Regardless, the presence of impaired quality of life in IBS indicates that IBS deserves serious attention and therapeutic intervention.88 IBS is associated with substantial costs because of days lost from work, excess physician visits, diagnostic testing, and use of medications.89-91 Patients with IBS miss three

Chapter 118  Irritable Bowel Syndrome times as many days from work as do those without bowel symptoms.92 IBS is the sixth leading physician diagnosis in outpatients in the United States, and this is likely an underestimate.93 A comprehensive burden-of-illness study in the United States estimated that IBS cost $1.6 billion in direct costs and a staggering $19.2 billion in indirect costs.72 Moreover, patients with IBS consume over 50% more health care resources than do matched controls without IBS.89

PATHOPHYSIOLOGY A number of different mechanisms have been implicated in the pathogenesis of IBS, including abnormal motility, visceral hypersensitivity, low-grade inflammation, and stress.1,8,94 Genetic factors could modulate the processing of gastrointestinal signals centrally and the inflammatory and immune responses locally, possibly predisposing to IBS. It seems reasonable to postulate that for IBS to manifest, several abnormalities—multiple hits—might need to occur. Some authors, therefore, conceptualize IBS as “a discrete collection of organic bowel diseases,”8 whereas other experts are concerned about “organification” of IBS because it might reduce the emphasis on the biopsychosocial model1,95 and imply that biological factors are sufficient to cause the disease. It seems likely that in IBS, an understanding of the individual, including his or her psychosocial nature and response to environmental factors, influences the expression of any biological determinants (Fig. 118-2). Regardless, further major therapeutic advances in the field seem unlikely to occur until the specific biological basis for symptoms is better identified.

ALTERED COLONIC AND SMALL BOWEL MOTILITY

In IBS, diarrhea can occur from multiple colonic mechan­ isms including increased high-amplitude propagated contractions (HAPCs), an enhanced gastrocolic response (prolonged rectosigmoid motor activity in response to a meal), or rectal hypersensitivity.96-98 Constipation may be secondary to increased segmental (nonpropulsive) contractions, decreased HAPCs, or reduced rectal sensation.99-101 Colonic and small bowel transit has been documented to be delayed in IBS-C and accelerated in IBS-D.1,101,102 Abdominal pain in IBS also may be associated with HAPCs.103 A greater increase in phasic contractions in the terminal ileum and colon has been observed following dis-

Early life Genetics Environment

tention, fatty meals, and cholecystokinin in patients with IBS (Fig. 118-3A).104 Discrete clusters of jejunal contractions also have been noted with increased frequency and duration in IBS (Fig. 118-3B), and they have been associated with pain in limited numbers of patients with IBS.104 Colonic motility in IBS can be increased by stress, anger, or instillation of deoxycholic acid, but this increase, although greater than in controls, is not specific for IBS1; a greater increase in colonic phasic contractions has been observed after administration of corticotropin-releasing hormone (CRH),105 and this increase is reduced by a CRH antagonist.106 Patients with IBS also have greater small bowel motor stimulation than do controls after cholecystokinin infusion, a fatty meal, or ileal distention.107 Autonomic dysfunction also has been reported in IBS patients with sympathetic adrenergic dysfunction associated with diarrhea and vagal dysfunction with constipation.108 Unfortunately, none of these motility parameters can be used as a diagnostic marker for IBS, and there remains no consensus on the exact patterns of motor derangement that actually induce constipation or diarrhea. It is possible that the motor abnormalities observed in IBS are secondary rather than primary.

VISCERAL HYPERSENSITIVITY

In the 1970s, balloon distention in the rectum was shown to induce pain at lower volumes in patients with IBS109; this has been confirmed in multiple studies using the barostat balloon that controls for changes in compliance, leading to the suggestion that colonic hypersensitivity is a useful biological marker of IBS.110,111 Visceral hypersensitivity might explain the fact that IBS patients seem more likely than controls to be aware of the presence of gas or intestinal contractions after meals or stress. Visceral hypersensitivity probably is confined to the intestine because somatic pain thresholds are normal, although not all studies agree on this.112-114 Visceral hypersensitivity is not a universal finding in patients with IBS and affects about 60% of patients (Fig. 118-4).115 In contrast to control subjects, patients with IBS and normal baseline visceral hypersensitivity might have rectal hypersensitivity induced by repeated distention of the sigmoid colon.116 This suggests that in IBS there is abnormal sensitization within the dorsal horn of the spinal cord or higher up in the central nervous system. Putative neurotransmitters that are of relevance to visceral hypersensitivity include serotonin, neurokinins, and calcitonin gene-related peptide.117 The capsaicin (red pepper) receptor on nerve fibers, also called transient recep-

Psychosocial factors Life stress Psychological state Coping Social support

Physiology Motility Sensation

IBS Symptom experience Behavior

Outcome Medications Healthcare visits Daily function Quality of life

Figure 118-2.  A conceptual model depicting the relationship between early life, psychosocial factors, physiology, symptom experience, and behavior and outcome. IBS, irritable bowel syndrome. (Adapted from Drossman D, Camilleri M, Mayer E, Whitehead W. AGA technical review on irritable bowel syndrome. Gastroenterology 2002; 123;2108.)

2095

2096

Section X  Small and Large Intestine A

I1

Abdominal cramps

I2

I3 50 mm Hg I4

1 min C

B

50 mm Hg

1 min

Abdominal discomfort Figure 118-3.  A, Motility recording from four sites between the terminal ileum (I1 to I4) and the cecum (C) in a patient with irritable bowel syndrome (IBS). Note the coincidence of abdominal cramps with high-pressure peristaltic waves propagated from the ileum to the colon. B, Fasting small intestinal motility recordings from four separate sites in a patient with IBS. Note the clusters of contractions (minute rhythm) associated with abdominal discomfort; this pattern is seen more frequently in patients with IBS than in control subjects. (From Kellow JE, Phillips SF: Altered small bowel motility in irritable bowel syndrome is correlated with symptoms. Gastroenterology 1987; 92:1885.)

tor potential vanilloid-1 (TRPV1), appears to be increased in the rectosigmoid colon in IBS and might mediate visceral pain.118 The N-methyl-d-aspartate (NMDA) receptor also may be important, because it modulates central (spinal cord) neuronal excitability.119 Visceral sensitivity, at least in the esophagus, can be reduced by an NMDA receptor antagonist.119 Serine proteases are thought to act as signaling molecules via the activation of proteinase-activated receptors (PARs). Extracts derived from colonic mucosal biopsies of patients with IBS (but not controls) have been observed to sensitize murine nerves in culture; this was blocked by a serine protease inhibitor.120 A significant increase in serine proteases has been observed in the stools of patients with IBS-D.121 Serine proteases could potentially damage tight junctions and increase intestinal permeability via PAR activation.122 The origin of stool serine proteases is uncertain, but they might derive from mast cells or the fecal microbiota.122 It is possible that inflammation is responsible for the sensitization in a subset of patients with IBS, as discussed later; however, some of this decreased threshold to balloon

distention may be attributed to hypervigilance or excessive attention to, or fear of, a painful stimulus.123

ABNORMAL GAS PROPULSION AND EXPULSION

Ambulatory monitoring of abdominal girth has revealed that the abdomen normally swells during the day, peaking in the late evening, but decreasing on lying down; this phenomenon often is exaggerated in IBS.124 Retention of gas following gas infusion into the small intestine is greater in patients with IBS than it is in healthy controls.125 Furthermore, in those with IBS, intestinal gas infusion induces more discomfort than it does in controls when subjects are asked to voluntarily suppress passing the gas.125 During gas infusion, IBS patients, in contrast to healthy controls, involuntarily suppress their abdominal wall muscle contraction, which might explain their tendency to become distended; this could reflect an abnormal intestinal-somatic reflex response.126 Small intestinal bacterial overgrowth (SIBO) has been speculated to contribute to bloating in IBS, but this is not established with certainty.127,128 Although modest increases

Chapter 118  Irritable Bowel Syndrome

Rectal pain threshold (mm Hg)

48

40

32

24

16

Digestive organic disease (n=7) Digestive functional disease (n=16) Psychiatric disease (n=8)

8 Controls

IBS

Functional constipation

Functional dyspepsia

Miscellaneous

Figure 118-4.  Distribution of rectal pain thresholds (distention pressure in mm Hg inducing pain) for each subject in the following groups: asymptomatic controls, patients with irritable bowel syndrome (IBS), patients with functional constipation, patients with functional dyspepsia, and a miscellaneous group as defined in the key. Black bars and boxes represent median ±25% of pain thresholds and interquartile ranges. At the level of 40 mm Hg, the sensitivity of the rectal barostat for separating IBS patients from normal subjects and non-IBS patients was 95% and the specificity was 71.8%. (From Bouin M, Plourde V, Boivin M, et al. Rectal distention testing in patients with irritable bowel syndrome: Sensitivity, specificity, and predictive values of pain sensory thresholds. Gastroenterology 2002; 122:1771.)

of bacteria have been documented in the proximal small intestine in IBS, SIBO rates do not differ between patients with IBS and controls.129 Similarly, lactulose breath testing, a surrogate marker of SIBO, often is abnormal in IBS using published criteria, but careful studies in controls suggest the rates of abnormality do not differ.130 Methane gas production by methagenic fecal flora, which occurs in the minority of the population, is now well established to be associated with constipation,128,131 possibly via slowing of intestinal transit in predisposed persons.132 Intravenous neostigmine has been demonstrated to clear retained intestinal gas and to reduce abdominal symptoms in patients with IBS and functional bloating.133 Physical activity might also enhance gas transit,134 and thus is to be encouraged.

LOCAL INFLAMMATION

The normal intestine is chronically in a state of inflammation, which occurs because of the balance between commensal enteric organisms and the host immune system. Inflammatory cells, including mast cells,135-138 and activated T lymphocytes139,140 are increased above normal in the mucosa in a subset of patients with IBS, suggesting that a low-grade inflammatory bowel disease may be present. Furthermore, lymphocytic infiltration of the myenteric plexus associated with neuron degeneration has been observed in severe IBS,141 as have increased mast cells in the muscularis externa.142 The cause of these abnormalities is unknown, but infections, abnormal bacterial flora, bile, or food antigens all could be contributors. From 7% to 30% of patients who have recovered from a proved episode of bacterial enteritis develop IBS.54-56,58,60,61 One study, however, has suggested that those with preexisting IBS who develop gastroenteritis may be more likely to seek medical consultation, thereby inflating the apparent risk estimates of this group.143 If the illness lasts more than three weeks or there are organisms involved that are toxigenic, then the risk of postinfection IBS is increased.61 Fur-

Figure 118-5.  Enteroendocrine (enterochromaffin) cell hyperplasia in a rectal biopsy specimen obtained from a patient with postinfection irritable bowel syndrome three months after a bout of Campylobacter enteritis. Synaptophysin-positive endocrine cells are brown, with blue hematoxylin counterstain. (Courtesy of S. Dunlop, MD, and R. Spiller, MD, Nottingham, UK.)

thermore, those with psychological distress might have a further increased risk of postinfection IBS.55,58,59 In those who develop postinfection IBS, there are increases in CD3, CD4, and CD8 T lymphocytes, macrophages, and enteroendocrine (enterochromaffin) cells (Fig. 118-5).58,139 Increased small intestinal permeability as demonstrated by the lactulose-mannitol test also has been reported to occur in postinfection IBS61,139; however, this test is confounded by intestinal transit and bacterial overgrowth, and whether abnormal intestinal permeability occurs remains speculative. Mast cells may play a central role in IBS. Activated mast cells release tryptase and histamine and have been observed to lie in close proximity to colonic nerves in patients with IBS; this finding has been correlated with abdominal pain.144 Supernatants prepared from colonic mucosal biopsies of

2097

2098

Section X  Small and Large Intestine patients with IBS have been shown to excite rat nociceptive visceral sensory nerves, suggesting that mast cell mediators, including tryptase, histamine, and prostaglandin E2, might represent another mechanism inducing visceral hypersensitivity in IBS.145 Colonic inflammation is associated with the production of a number of important mediators including 5-hydroxytryptamine (5-HT), prostaglandins, bradykinins, adenosine, and nerve growth factors.8 Abnormal release of 5-HT might have a central role in the manifestations of IBS.146 More than 95% of 5-HT is located in the enteroendocrine cells of the intestine and 5-HT is released from these cells following stroking or increased pressure, such as after a meal; 5-HT then acts on primary intrinsic afferent neurons to initiate the peristaltic reflex by activation of ascending excitation and descending inhibition.8,147 5-HT is taken up again by a specific serotonin transporter (SERT) expressed in the enterocytes. There is some evidence that an exaggerated release of 5-HT in IBS can occur after a meal.148 It has been observed in rectal biopsy specimens that 5-HT molecular signaling may be abnormal in IBS. In one study, 5-HT reuptake was reduced in IBS and also in ulcerative colitis compared with controls, although 5-HT release was unaffected and the numbers of enteroendocrine cells were unchanged.149 The findings were similar in IBS patients with constipation or diarrhea, leading to the hypothesis that in IBS there is increased availability of mucosal 5-HT that can induce diarrhea, but if there is desensitization of 5-HT receptors, this leads to constipation or an alternating bowel pattern149; confirmatory data currently are not available.150

ROLE OF FOOD

Many patients with IBS attribute their symptoms to certain foods, with wheat, dairy products, citrus fruits, potatoes, onions, and chocolate most commonly implicated.151,152 In an uncontrolled trial, one half of the patients with IBS reported improvement with elimination diets.66

Wheat Intolerance or Allergy

Substantial amounts of wheat are eaten in Western countries, 10% to 15% of which is not digested by human enzymes.8 Furthermore, subtle forms of gluten intolerance may be present in some people with IBS who do not have any overt evidence of celiac disease (see Chapter 104).153 A six-month gluten-free diet improved symptoms in 70% of patients with IBS-D who were HLA DQ2 positive,154 compared with 20% who were negative. In the absence of a definitive diagnosis of celiac disease, however, a gluten-free diet remains of unproven benefit.8

Sugar Malabsorption

Symptoms of IBS can be confused with those of lactose intolerance.155 Lactose intolerance occurs with varying prevalence depending on one’s ethnic group and is seen in 10% of populations of northern European descent, 40% to 60% of those of Asian descent, 90% of Chinese, and 60% to 80% of Africans (see Chapter 101). In acquired hypolactasia, there is some residual ability to digest small amounts of lactose and because most people do not ingest more than 12.5  g of lactose a day, they do not suffer from this ingestion.156 Unless a lactose-­intolerant patient regularly ingests substantial amounts of lactose, lactose intolerance cannot be the explanation for the symptoms.157 Fructose and sorbitol malabsorption might contribute to IBS symptoms in some patients; however, fructose and sorbitol malabsorption, with a prevalence of 30% in those with IBS, may be no more common in IBS than in the background population.158-162 In a double-blind rechallenge trial, 25

patients who had responded to fructose withdrawal were challenged with fructose or fructans; nearly 80% developed symptoms compared with less than 15% who were given glucose.161

ABNORMAL COLONIC FLORA AND SMALL INTESTINAL BACTERIAL OVERGROWTH

It has been suggested that the colonic flora could be abnormal in a subset of patients with IBS, resulting in increased colonic fermentation, production of excess gas, and development of symptoms.163 This has led to an interest in preand probiotic therapy for IBS. Others have reported that there is a high prevalence of SIBO in IBS, based on hydrogen breath testing and the clinical response to nonabsorbable antibiotics.164,165 Abnormal hydrogen breath test results, however, can occur because of transit abnormalities, and any association with IBS remains to be established.130,166,167

CENTRAL DYSREGULATION

Visceral afferent signals from the intestine reach the brainstem and thalamus and are consciously perceived only occasionally, although there may be some subliminal registration of low-intensity signals.3 Abnormal modulation of visceral afferent signals can occur at multiple levels in visceral, spinal, and central regions. Based on cerebral blood flow changes, functional brain imaging studies (functional magnetic resonance imaging or positron-emission tomography [PET]) have suggested that there are alterations in the brain response to visceral stimuli in IBS. In IBS patients, greater activation of the mid-cingular cortex, an area that processes visceral signals, has been reported following delivered or anticipated rectal distention.168-170 These observations could explain why anxiety or stress can enhance perception of visceral pain, whereas relaxation or distraction decreases pain in IBS. Using hypnosis to selectively alter noxious stimuli, PET scanning revealed significant changes in anterior cingulate cortex activity.171 Sex differences in brain networks concerned with antinociceptive and autonomic responses following rectal distention in IBS also have been observed.172

PSYCHOLOGICAL FACTORS (see Chapter 21) Depression, anxiety, and somatization are the most common psychiatric conditions that coexist in IBS; in referral practice, 40% to 94% of patients with IBS are so affected.173 Some have suggested that consultation (referral) bias explains the higher rates of psychological and psychiatric comorbidity in IBS compared with controls,1,79-81,174-176 but other data suggest the association is real.174-176 In patients with IBS, a history of sexual, physical, or emotional abuse also is reported more often than in those without IBS.177-179 Abuse has not been shown to alter rectal sensation180 but it might modulate central brain responses to pain.181,182 Patients with IBS are more likely to report greater lifetime and daily stressful events than are those with organic disease or healthy controls and may be more susceptible to stressaltering gastrointestinal function.1 Childhood stress may be particularly important.94 Gastric suction at birth was associated with a subsequent three-fold increased risk of having a hospital admission for unexplained abdominal pain (or GI symptoms) compared with sibling controls who had not undergone gastric suction.183 In rats, maternal separation in the perinatal period induces an anxiety state and is associated with visceral hypersensitivity.94 Furthermore, in rats, moderately severe stress leads to the release of corticotropin-releasing factor and ac­

Chapter 118  Irritable Bowel Syndrome celeration of colonic transit.184 Stress in healthy volunteers changes intestinal secretion and permeability responses.185 Sustained stress could, therefore, be important in both the onset and persistence of IBS. Anxiety and depression, rather than being a primary problem, might occur secondary to production of proinflammatory cytokines.186 In IBS, immune activation of the intestine has been linked to elevated TNF-α levels and anxiety, suggesting that anxiety in IBS might occur secondary to intestinal inflammation in some cases.187 Mast cells communicate with the enteric and central nervous systems; an excess of mast cells in the colon appears to be associated with depression and fatigue in IBS.138

GENETIC FACTORS

Limited, but increasing, evidence points to at least a small hereditary component of IBS. There is clustering of IBS in families.68,188,189 Twin studies generally have shown that there is a greater concordance of IBS in monozygotic compared with dizygotic twins, suggesting a modest genetic component, although the environmental component probably is much greater.69,190-193 Potential candidate genes have been reported to be associated with IBS, but further confirmatory evidence is required and their functional significance must be unraveled.194 For example, patients with IBS have been reported to have significantly lower frequencies of the high-producer genotypes of interleukin (IL)-10. A lower amount of this antiinflammatory cytokine might predispose to greater inflammation in response to an infectious insult in IBS195; others have failed to confirm these observations.196 A specific sodium channel mutation has been identified in IBS (SCN5A).197 Associations with polymorphisms of the promoter region of the serotonin transporter gene in IBS have not been consistent.197-199 A functional variant in the serotonin type III receptor gene may be associated with IBS-D in women.200 Common genetic factors do not appear to explain the association of IBS with depression.201

DIAGNOSIS In patients who present with IBS-like symptoms, there are a number of alarm features (or red flags) that clearly warrant prompt investigation. These include any history of bleeding or evidence of anemia, a history of unexplained weight loss, unexplained vomiting, progressive dysphagia, a family history of malignancy, and new-onset symptoms in older age (Table 118-2).155,202,203 The traditional alarm features, however, have been documented to have poor diagnostic utility.1 For example, night-time symptoms are common in IBS and appear not to discriminate IBS from organic disease,202 although most physicians would still investigate patients who were being awakened in the middle of the night by pain or those with nocturnal diarrhea. Any patient 50 years of age or older requires a structural evaluation of the colon, if one has not been previously performed. The preference remains colonoscopy to exclude other disease and, in particular, colon cancer. Although older persons can develop IBS, risk of organic disease increases with age. Furthermore, colon cancer screening in those 50 years of age and older, even with no symptoms, currently is recommended in the United States. Patients who meet the Rome criteria for IBS and who have no alarm features are less likely to have a cause for their presentation other than IBS. For example, among 98 patients who met the Rome I criteria and had no alarm features, of

Table 118-2  Alarm Features Considered Potentially Relevant in the Diagnosis of Organic Disease as Opposed to Irritable Bowel Syndrome History Blood in the stool Family history of colon cancer, inflammatory bowel disease, or celiac disease Fever Onset after age 50 years Night-time symptoms (awakening the patient from sleep) Chronic diarrhea Progressive dysphagia Recurrent vomiting Severe chronic constipation Short history of symptoms Travel history to locations endemic for parasitic diseases Weight loss Physical Examination Abdominal mass Arthritis (active) Dermatitis herpetiformis or pyoderma gangrenosum Occult or overt blood on rectal examination Signs of anemia Signs of intestinal obstruction Signs of intestinal malabsorption Signs of thyroid dysfunction Adapted from Olden KW. Diagnosis of irritable bowel syndrome. Gastroenterology 2002; 122:1701-14.

Table 118-3  Frequency of Organic Disease in Patients Meeting Symptom-Based Criteria for Irritable Bowel Syndrome ORGANIC DISEASE Inflammatory bowel disease Colorectal cancer Celiac sprue Gastrointestinal infection Thyroid dysfunction Lactose malabsorption

IBS PATIENTS (%)

GENERAL POPULATION (%)

0.51-0.98

0.3-1.2

0-0.51 4.7 0-1.5 6 22-26

0-6 (varies with age) 0.25-0.5 NA 5-9 25

NA, not applicable. Data from Cash B, Schoenfeld P, Chey W. The utility of diagnostic tests in irritable bowel syndrome patients: A systematic review. Am J Gastroenterol 2002; 97:2812-19.

whom 50% had been referred because of diagnostic uncertainty, the Rome I criteria had a positive predictive value of up to 100% in this setting, although their sensitivity was only 65%.204 Systematic reviews have evaluated the value of diagnostic tests in IBS.9 The results, based on limited numbers of referred patients, suggested that IBS patients do not have an increased risk of most organic diseases compared with non-IBS controls (Table 118-3). Extensive investigations to exclude most possibilities are expensive and carry the danger of reinforcing abnormal illness behavior. There also is the real risk of uncovering findings that are irrelevant to the diagnosis but that precipitate yet more expensive and even dangerous investigations. Traditional screening tests that have low yields in IBS patients include a full blood count, renal and liver biochemical testing, thyroid function testing, and evaluation of three fresh stool samples for parasites; these are inexpen-

2099

2100

Section X  Small and Large Intestine New patient meeting Rome criteria for IBS

Yes

Alarm features No Screen for celiac disease

Screening blood and stool tests including tests for celiac disease Structural colonic evaluation Other tests depending on the predominant complaints, e.g., pelvic floor function in patients with constipation; further evaluation of diarrhea Figure 118-6.  Algorithm for diagnostic testing in patients with possible irritable bowel syndrome (IBS) according to the Rome criteria. (Adapted from Cash BD, Chey WD. Irritable bowel syndrome: An evidence-based approach to diagnosis. Aliment Pharmacol Ther 2004; 19:1235.)

sive tests and they can be reassuring for both patient and physician if they are negative or normal.8 Crohn’s disease can be missed. An elevated C-reactive protein, although nonspecific, can indicate the presence of undiagnosed Crohn’s disease but often is negative in this clinical setting. Fecal calprotectin has been shown to discriminate IBS from Crohn’s disease with excellent sensitivity and specificity.14,205 In contrast, evaluation of the small intestine either radiologically or via capsule endoscopy has a very low yield in the setting of typical IBS symptoms without alarm features; only if there is concern should such investigations be considered. Hydrogen breath testing to identify lactose intolerance or SIBO cannot be endorsed routinely.8 Bile salt malabsorption causing diarrhea in the setting of IBS occurs uncommonly, and a therapeutic trial of cholestyramine is probably more useful than a 75-seleno-homocholic acidtaurine (Se-CAT) test.8,206,207 If diarrhea is persistent, co­ lonoscopy with biopsy may be considered, although the yield of colonic biopsy remains low, and this test is probably not cost-effective.9 In a setting of severe constipation, exclusion of pelvic outlet obstruction with anorectal manometry including balloon expulsion testing should be considered because the result might alter management.1,2,8 Routine testing for celiac disease is now recommended in patients with typical symptoms of IBS-D.1 Data from the United Kingdom have shown a seven-fold increased risk of celiac disease in patients with IBS; up to 5% of patients with symptoms consistent with IBS had celiac disease compared with 0.5% of controls,208 but much lower rates have been observed in the United States.209 Decision analysis suggests that testing is cost-effective unless the prevalence of celiac disease falls to less than 1% in those with IBS-like symptoms.210 A screening test (tissue transglutaminase antibody) in those consuming a normal diet, followed by duodenal biopsies in those with a positive test, should be considered unless the background prevalence of celiac disease is very low in the patient population being seen. Latent celiac disease (antibody positive, biopsy normal) might respond to a gluten-free diet, but its prevalence in IBS is unclear.153 Celiac disease can manifest with atypical features including bloating and constipation without diarrhea (see Chapter 104).211 In summary, the diagnosis of IBS can be made based on the history (with particular attention to presence or absence

 First-line therapy for IBS

Failure of therapy: reconsider diagnosis and testing

of the Rome criteria) and in the absence of any red flags. A normal physical examination can be reassuring for patient and physician. In this setting, the patient who responds to an empiric trial of therapy for IBS does not require any further diagnostic evaluation (Fig. 118-6).203 Those who fail to respond should undergo more-extensive evaluation, depending on the predominant symptoms.

TREATMENT EDUCATION AND SUPPORT

IBS tends to be a life-long disorder, and establishment of a strong physician-patient relationship is key to providing the best clinical care (Table 118-4).8,212 Indeed, patients with IBS often perceive their physician as having a highly negative medical belief about the disorder, and this perception itself impedes best care.213,214 Other common perceptions of the care they receive include being mislabeled as psychologically disturbed and that they have not been provided with adequate medical information or support.213,214 A good physician-patient relationship has been associated with reduced use of medical services.215 It is important to discover why the patient has decided to visit the health care provider at the time he or she did. The reasons can vary: new life stressors, exacerbating factors in the diet or changes in medications, increased fear of serious disease, and the development of treatable psychiatric comorbidity. A hidden agenda such as seeking disability or new narcotic abuse sometimes explains the consulting behavior. In terms of providing optimal reassurance, it is important first to educate patients and then to actively reassure them. Patients typically want to understand why their symptoms have occurred; they also want to obtain validation that their symptoms are real. Specific education classes designed for those with IBS appear to be useful therapeutic interventions,216,217 although randomized, controlled trials have not been done to prove the effectiveness of these interventions. A stepped-care approach depending on the severity of the presenting symptoms provides a useful guide for considering therapies (Tables 118-5 and 118-6).218

Chapter 118  Irritable Bowel Syndrome DIET

The standard of care for IBS typically has been a high-fiber diet.219-225 Data from available randomized, controlled trials indicate that fiber is of global benefit, and the number needed to treat [NNT] is 11.226,227 NNT is used to assess the effectiveness of a health-care intervention and is the number of patients who need to be treated to prevent one additional bad outcome. Fiber supplements often are better tolerated than introduction of dietary fiber. The best evidence for a benefit of fiber supplements comes from the studies of ispaghula (psyllium hydrophilic mucilloid; ispaghula husk).1 Wheat bran has been no better than placebo in IBS.226,227 Fiber is not helpful for pain, but it can benefit constipation and can sometimes firm up loose stools. The key to taking fiber supplements is to begin at a low dose and increase very slowly. This reduces the problems of excessive bloating, gas, and pain that typically are aggravated by increased fiber. For example, most of the powder fiber supplements contain about 3 to 6 g of fiber. If the goal

Table 118-4  Management Recommendations for Irritable Bowel Syndrome Make a positive diagnosis based on symptoms and the absence of alarm features. Determine the effect of illness on the patient and on the patient’s psychosocial resources (e.g., family support). Determine if there is a comorbid psychiatric disease or an unresolved major loss or trauma. Assess the patient’s expectations and hidden fears (e.g., find out why he or she has presented now despite long-standing symptoms), and try to address all concerns. Provide education, including an understandable explanation of why symptoms might arise, emphasizing that the patient is not alone in his or her suffering and the prognosis is benign. Provide firm reassurance, emphasizing that the symptoms are known to be real (not just “in the patient’s head”) and that irritable bowel syndrome is a recognized bowel disease. Avoid giving mixed messages (e.g., by reassuring the patient and then ordering extensive tests without an adequate explanation). Avoid repeated tests unless the new development of structural disease is suspected (e.g., presentation with new alarm features). Base treatment on the principle of patient-based responsibility for care. Set realistic treatment goals. Consider referral to a patient support group. Organize a continuing care strategy if symptoms have been chronic or disabling. Consider psychological treatments for patients with moderate to severe symptoms. Adapted from Talley N, Spiller RC. Irritable bowel syndrome: A little understood organic bowel disease? Lancet 2002; 360:555-64.

is to supplement approximately 10 to 15 g of fiber in total, the amount should be increased by 3  g every one to two weeks. Many patients with IBS suspect that food intolerance may be relevant to their symptoms. It is useful to determine the amounts of milk and milk products being consumed to decide whether lactose intolerance testing should be considered. Clinical experience suggests that even in the setting of a diagnosis of lactose intolerance and typical IBS symptoms, more often than not, IBS symptoms persist despite withdrawal of all lactose in the diet, indicating that this is the chance overlap of common conditions. Fructose consumption has increased dramatically in the United States and in other developed countries; excess fructose could lead to some IBS-like symptoms that might be relieved by exclusion of this sugar.158-161 Reducing fatty foods, gas-producing foods, caffeine, or alcohol also may be helpful for some patients, but randomized, controlled trials to support these recommendations are lacking. Exclusion diets can be useful in some cases. One randomized trial measured IgG antibodies to foods and then excluded those IgG positive foods from the diet in the active arm of the trial; in the sham arm, other foods from the diet were excluded in a blinded fashion.228 The exclusion of foods with a positive IgG antibody response provided therapeutic benefit in patients with IBS-C and IBS-D. A systematic review of food-exclusion diets has suggested that in IBS, 12% to 67% of patients will respond, but most of the data are uncontrolled.229 Cromoglycate did show some benefit in one study of patients with positive skin prick tests to foods, but these results remain to be confirmed.230

Table 118-6  Suggested Sequence of Treatment of the Irritable Bowel Syndrome PREDOMINANT SYMPTOM

FIRST STEP

SECOND STEP

Bloating

Adjust diet Treat constipation

Constipation

Fiber supplement, e.g., ispaghula Polyethylene glycol Loperamide

Probiotic Nonabsorbable antibiotic Tricyclic antidepressants, SSRIs Lubiprostone

Diarrhea Abdominal pain

Antispasmodic, peppermint oil

5-HT3 antagonist (alosetron) Tricyclic antidepressants, SSRIs Psychological therapy

SSRIs, selective serotonin reuptake inhibitors.

Table 118-5  Treatment of Irritable Bowel Syndrome: A Stepped-Care Approach STEP

SEVERITY OF SYMPTOMS

LEVEL OF CARE

CLINICAL AND PSYCHOLOGICAL ISSUES

1

Mild

Primary

Fear of serious disease, anxiety, worry, stress

2

Moderate

Secondary

3

Severe

Tertiary

Uncertainty regarding diagnosis; disrupted lifestyle Coexistent psychiatric disease; possible secondary gain; disability, chronic pain

MANAGEMENT Positive diagnosis; explanation; reassurance; dietary management; regular follow-up Reinforce above measures Avoid over-testing; low-dose TCA or SSRI; alosetron (for severe diarrhea); treat depression, anxiety; refer to pain clinic

SSRI, selective serotonin reuptake inhibitor; TCA tricyclic antidepressant. Adapted from Drossman DA, Thompson WG. The irritable bowel syndrome: Review and a graduated multicomponent treatment approach. Ann Intern Med 1992; 116(12 Pt 1):1009-16.

2101

2102

Section X  Small and Large Intestine MEDICATION Antispasmodics and Anticholinergics

In the United States, anticholinergics (dicyclomine, pro­ pantheline, belladonna, and hyoscyamine) continue to be used commonly for IBS.223,224 A meta-analysis of randomized, controlled trials concluded that antispasmodics were superior to placebo in the treatment of IBS.224 Overall, there was an improvement of abdominal pain and IBS global symptoms in the pooled analyses224; however, the quality of most of these trials was low, the results were mixed, and publication bias could not be excluded. Moreover, only the anticholinergic antispasmodics are available in the United States.1 Non-anticholinergic antispasmodics, unavailable in the United States, that appear to be efficacious include otilonium (imetropium) and certain selective calcium channel blockers (e.g., pinaverium).224 Anticholinergics clinically seem most useful for patients with postprandial pain when taken 30 minutes before eating. No advantage of sublingual or suppository over oral anticholinergic preparations has been documented in patients with IBS. Peppermint oil does appear to be efficacious in IBS for abdominal pain, and it is usually well tolerated; the NNT is 2.5.1 The usual dose is 0.2 mL three times a day 30 minutes before meals (swallowed not chewed). Hot peppermint tea also may be useful if there are associated upper GI symptoms, but no controlled trial data exist regarding this agent.231

Laxatives

The efficacy of this class of drugs for IBS-C is uncertain. No randomized, controlled trials of laxatives in IBS are available. Osmotic laxatives often are prescribed but can aggravate bloating and pain. Polyethylene glycol, although not approved for this indication, does seem useful in practice, particularly if constipation is troublesome despite other therapy.232,233 Stimulant laxatives are probably safer than has been appreciated, but they often induce abdominal cramping or pain and generally seem unsatisfactory for patients with IBS.234 Lubiprostone is a chloride channel activator that stimulates intestinal fluid secretion. In IBS a lower dose of lubiprostone is currently approved for women with IBS-C than is used for chronic constipation (8 µg twice daily); the global benefit over placebo in IBS-C, however, is modest.235

Antidiarrheals

Loperamide is established to be efficacious based on randomized, controlled trials in IBS-D, but it does not improve abdominal pain or bloating.226,236-238 Loperamide is most effective when taken prophylactically, rather than being taken after diarrhea has occurred; doses of loperamide range from 2 to 16 mg/day, and high doses seem safe. Diphenoxy­ late has not been tested in IBS but may be similarly efficacious. Codeine phosphate, because of its side effects (dizziness, nausea, sedation) and high risk of inducing dependence, should be avoided in IBS. Anecdotally, bile salt–sequestering agents, such as cholestyramine, seem to help diarrhea in IBS in some cases, possibly because of underlying bile salt malabsorption. It is best begun at a low dose and built up slowly to four times daily if needed. Bismuth subsalicylate anecdotally is also useful for diarrhea control in IBS but must be used intermittently because of concerns about bismuth toxicity.

Serotonin-Receptor Drugs

Alosetron is a 5-HT3 antagonist that is efficacious in women with severe IBS-D.239 The NNT is 8,239 and it has been shown to improve quality of life.240 The starting dose is 0.5 to 1 mg

daily. In the United States, it is available only via a restricted prescribing program because of concerns about ischemic colitis and severe constipation.241 The dose can be increased to 1  mg twice daily after four weeks if symptoms are not controlled and there have been no side effects. Ischemic colitis occurs in 0.1% of alosetron-treated patients and is drug-related but dose independent; the ischemia is usually transient and without irreversible consequence, although up to 50% of patients with alosetron-associated ischemic colitis required hospitalization.242 Constipation occurs in one third of alosetron-treated patients. The prescription of the drug is absolutely contraindicated in IBS patients with any history of constipation, thrombotic tendency, or ischemic colitis. Pharmacogenomic data suggest that alosetron may be more efficacious in those with the long homozygous polymorphism of the SERT transporter gene, but this needs confirmation.243

Antidepressants and Anxiolytics

The tricyclic antidepressants appear to be efficacious in IBS but might improve global well-being more than symptoms.244 Meta-analyses have reported the NNT to be between 3 and 4, although these studies included a number of low-quality and potentially flawed trials.1,223,245 A large, high-quality randomized, controlled trial of desipramine (at a dose of 50 to 150 mg) versus placebo in female patients showed that 60% were responders to the TCA versus 47% to placebo; this difference failed to reach significance in the intention-to-treat analysis but was significant in the per-protocol analysis.246 When using a TCA in IBS, it is recommended to start it at a low dose (e.g., 10 to 25 mg of desipramine or nortriptyline) and increase the dose by 10 to 25 mg weekly, aiming for a dose of 50 mg initially. Many patients do not require full antidepressant dosing, unless comorbid depression is present. TCAs tend to be constipating, and therefore they may be of most benefit in IBS-D, although the data supporting this postulate are unclear. Adverse events with TCAs are a problem. Approximately one in three to one in four treated patients develops side effects, with one in 22 having potentially serious reactions; up to 40% discontinue use or change therapy because of intolerance.247 The selective serotonin reuptake inhibitors (SSRIs) cause fewer side effects than the TCAs, and a meta-analysis of the randomized, controlled trials in IBS has reported a global benefit of SSRIs with an NNT of 3.5.1 In one trial that lacked a placebo arm, paroxetine was compared with usual care in IBS patients and shown to improve quality of life but not abdominal pain.248 Other studies showed some benefit for pain or no effect on pain249,250; the findings remain inconsistent. It is possible that SSRIs may be more beneficial in IBS-C because they accelerate small intestinal transit.251 Benzodiazepines might have a small benefit over placebo in IBS, but the evidence for this observation is very weak.221 Because of habituation, this class of drugs generally should be avoided.

Antibiotics

Nonabsorbable antibiotics such as rifaximin appear to be superior to placebo in IBS in short-term treatment studies, especially for bloating and diarrhea,1 but long-term results are unavailable.164,165,252 The daily dose of rifaximin typically is 400 mg for 10 days. Any benefit can last at least 10 weeks, but treating a recurrence of IBS symptoms with another course cannot currently be recommended.1

Probiotics

It has been suggested that abnormal colonic flora could be relevant in the pathogenesis of IBS, which has led to great

Chapter 118  Irritable Bowel Syndrome interest in using probiotics to try to naturally alter the flora. Some initial small trials produced promising results with Bifidobacterium infantis and combination products,253,254 but the findings have been variable.194 Such variation might reflect the strain and dose used, whether live or dead organisms were given, and other unknown factors. A capsule containing live organisms (more than 1 billion) given once a day is the usual dosing recommendation.

Other Drugs

Anticonvulsants have analgesic properties. Phenytoin was not shown to be beneficial in a randomized trial.255 Gaba­ pentin might have some therapeutic value for pain of IBS anecdotally. Pregabalin is approved for neuropathic pain and fibromyalgia; anecdotally, some patients with IBS respond.256 Beta blockers can inhibit colonic motor function, but neither atenolol nor timolol was better than placebo in small trials.257,258 Leuprolide is a gonadotropin-releasing hormone analog that was superior to placebo in terms of reducing abdominal pain and nausea,259,260 but the entry criteria for these studies were vague and side effects of chemical castration were significant; this drug cannot be routinely recommended. Colchicine increases spontaneous bowel movements and decreases colonic transit time,261 but its role in IBS with constipation is unknown. Misoprostol may be helpful for refractory constipation,262 but no controlled trials in IBS are available. Domperidone and erythromycin are prokinetic agents used predominantly to treat upper gastrointestinal tract disorders and probably are not efficacious in IBS.221 Pyridostigmine is a parasympathomimetic that enhances intestinal motility and might benefit patients with slowtransit constipation, but data in IBS are lacking.222 Octreotide reduces intestinal transit time, secretion, and sensation in IBS263 but is impractical to use for diarrhea in IBS. Clonidine, an α2-receptor agonist, may be useful in IBS-D because it enhances rectal compliance and reduces fasting colonic motor activity264; however, significant side effects limit its use, particularly postural hypotension, sedation, fluid retention, and depression. There is increasing interest in testing anti-inflammatory drugs for IBS. Three weeks of oral prednisone (30 mg/day) failed to improve postinfection IBS symptoms in one placebo-controlled trial.265 The role of 5-aminosalicylic acid compounds in IBS is unknown, but these agents might have antinociceptive and mast cell-stabilizing actions.266 There is no convincing evidence that either simethicone267 or charcoal works to relieve the symptoms of excessive gas in IBS.268,269 Beano (α-galactosidase) might reduce flatus but not other IBS symptoms.270 Malodorous flatus might respond to zinc acetate, bismuth subsalicylate, a bulky charcoal cushion device, or peppermint oil.271-273

PSYCHOLOGICAL TREATMENTS

Psychotherapy, hypnotherapy, and cognitive behavioral therapy (CBT) have been proposed as useful treatments for IBS.1 A systematic review concluded that these therapies were superior to wait-list controls (who are followed similarly but are not treated).1 Hypnotherapy can improve cognition in IBS.274 Excellent efficacy data exist for CBT,95,248 albeit not in all controlled trials.275 There are no head-tohead studies comparing the different psychological interventions or combination therapy. Based on the available literature, IBS patients with abdominal pain, diarrhea, and psychological distress appear most likely to have a beneficial response to such intervention, particularly if the symptoms have been of short duration and have waxed and waned.1 Patients with constant abdominal pain do poorly

with psychological treatment.1 Indeed, symptoms tend not to improve; rather, the ability to cope with IBS seems to drive any global benefit. The major advantage of psychological treatment is that despite the initial expense, long-term benefits might offset the cost.250

ALTERNATIVE TREATMENTS

Many different alternative remedies are used by patients with IBS.233,276 In one high-quality randomized, controlled trial, Chinese herbal medicine (comprising a combination of 20 herbs) was superior to placebo, although this requires confirmation277 and the risks of using multiple herbal concoctions continue to be of concern.278 An ayurvedic preparation also was superior to placebo,279 but the data here remain limited. Ginger (especially for nausea) and aloe are used often, but there are no randomized, controlled trials in IBS.276 St. John’s wort probably is not efficacious. Whether pancreatic enzymes help remains unknown, although a small double-blind cross-over study suggests some benefit for postprandial bloating.280 Acupuncture might reduce rectal hypersensitivity,281,282 but efficacy is not established in randomized, controlled trials.

PROGNOSIS There is no evidence for even a small increase in mortality in IBS, despite referrals for invasive testing, excess abdominal and other surgery rates, and a link to ischemic colitis.283 In clinical practice, once a diagnosis of IBS has been made, it usually requires no revision despite prolonged follow-up. In 112 consecutive IBS patients followed through their medical records for a median of 29 years, survival in IBS was not different from expected, although 9% developed organic disease a median of 15 years after diagnosis.215 Among 75 patients with a clinical diagnosis of IBS followed up for 10 to 13 years, none had another explanation uncovered for their symptoms, yet symptoms did not resolve in 92% and 47% had undergone a repeat structural colonic evaluation to no avail.284 Some IBS patients have spontaneous improvement over time, but IBS usually is a relapsing disorder. The presence of excessive psychological distress or anxiety, as well as a long duration of complaints, tends to indicate a poorer prognosis.1

KEY REFERENCES

American College of Gastroenterology Task Force on IBS. Brandt LJ, Chey WD, Foxx-Orenstein AE, et al. Systematic review of the management of irritable bowel syndrome in North America. Am J Gastroenterol 2009; 104(Suppl 1):S1-35. (Ref 1.) Barbara G, Stanghellini V, De Giorgio R, et al. Activated mast cells in proximity to colonic nerves correlate with abdominal pain in irritable bowel syndrome. Gastroenterology 2004; 126:693-702. (Ref 144.) Bengtson MB, Ronning T, Vatn MH, et al. Irritable bowel syndrome in twins: Genes and environment. Gut 2006; 55:1754-9. (Ref 69.) Bijkerk CJ, de Witt NJ, Muris JW, et al. Soluble or insoluble fibre in irritable bowel syndrome in primary care? Randomised placebo controlled trial. BMJ 2009; 339:b3154. Camilleri M, Atanasova E, Carlson PJ, et al. Serotonin-transporter polymorphism pharmacogenetics in diarrhea-predominant irritable bowel syndrome. Gastroenterology 2002; 123:425-32. (Ref 245.) Drossman DA, Corrazziari E, Delvaux M, et al. Rome III: the functional Gastrointestinal Disorders. 3rd ed. McLean, Va: Degnon Associates; 2006. (Ref 3.) Drossman D, Toner BB, Whitehead WE, et al. Cognitive-behavioral therapy versus education and desipramine versus placebo for moderate to severe functional bowel disorders. Gastroenterology 2003; 125:19-31. (Ref 248.)

2103

2104

Section X  Small and Large Intestine Ford AC, Talley NJ, Spiegel BM, et al. Efficacy of fibre, antispasmodics and peppermint oil in irritable bowel syndrome: systematic review and meta-analysis. BMJ. 2008; 337:a2313. (Ref 224.) Ford A, Talley NJ, Veldhuyzen van Zanten S. Will the history and physical examination reveal the cause of my patient’s lower gastrointestinal symptoms? JAMA 2008; 300:1793-805. (Ref 10.) Gecse K, Roka R, Ferrier L, et al. Increased faecal serine protease activity in diarrhoeic IBS patients: A colonic lumenal factor impairing colonic permeability and sensitivity. Gut 2008; 57:591-9. (Ref 121.) Halder SL, Locke GRI, Schleck CD, et al. Natural history of functional gastrointestinal disorders: A 12-year longitudinal population-based study. Gastroenterology 2007; 133:799-807. (Ref 53.) Hammer J, Eslick G, Howell S, et al. Diagnostic yield of alarm features in irritable bowel syndrome and functional dyspepsia. Gut 2004; 53:666-72. (Ref 202.) Hiatt RB, Katz L. Mast cells in inflammatory conditions of the gastrointestinal tract. Am J Gastroenterol 1962; 37:541-5. (Ref 142.) Pimentel M, Lin HC, Enayati P, et al. Methane, a gas produced by enteric bacteria, slows intestinal transit and augments small intestinal

contractile activity. Am J Physiol Gastrointest Liver Physiol 2006; 290:G1089-95. (Ref 132.) Pimentel, M., Park S, Mirocha J, et al. The effect of a nonabsorbed oral antibiotic (rifaximin) on the symptoms of the irritable bowel syndrome: A randomized trial. Ann Intern Med 2006; 145:557-63. (Ref 254.) Ringel Y, Drossman DA, Leserman JL, et al. Effect of abuse history on pain reports and brain responses to aversive visceral stimulation: An FMRI study. Gastroenterology 2008; 134:396-404. (Ref 182.) Saito YA, Talley NJ. Genetics of irritable bowel syndrome. Am J Gastroenterol 2008; 103:2100-4. (Ref 194.) Spiegel BM, DeRosa VP, Gralnek IM, et al. Testing for celiac sprue in irritable bowel syndrome with predominant diarrhea: A cost-­ effectiveness analysis. Gastroenterology 2004; 126:1721-32. (Ref 210.) Talley N, Spiller RC. Irritable bowel syndrome: A little understood organic bowel disease? Lancet 2002; 360:555-64. (Ref 8.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

119 Intestinal Obstruction Richard H. Turnage and Maureen Heldmann

CHAPTER OUTLINE Definitions  2105 Small Bowel Obstruction  2105 Etiology  2105 Pathophysiology  2106 Clinical Features  2106 Laboratory Findings  2107 Radiologic Findings  2107 Treatment and Outcome  2110 Special Considerations  2113

DEFINITIONS Impairment to the aboral passage of intestinal contents can result from either a mechanical obstruction of the bowel or failure of normal intestinal motility in the absence of an obstructing lesion (ileus). Intestinal obstruction may be categorized according to the degree of obstruction to flow (partial or complete), the absence or presence of intestinal ischemia (simple or strangulated), and the site of obstruction (small bowel or colonic). These distinctions have important prognostic and therapeutic relevance. For example, complete or strangulated obstruction requires urgent operative management, whereas partial small bowel obstruction, in selected cases, may be managed successfully without laparotomy. A closed-loop obstruction is a mechanical obstruction in which both the proximal and distal parts of the involved bowel are occluded. This condition is associated with a particularly high risk of strangulation, necrosis, and perforation.

SMALL BOWEL OBSTRUCTION ETIOLOGY

The most common cause of small bowel obstruction (SBO) is intra-abdominal adhesions following laparotomy; this accounts for about three fourths of all cases.1,2 Peritoneal adhesions are common after laparotomy and are exacerbated by intra-abdominal infection, the tissue ischemia attending wound closure (e.g., an anastomosis), external beam radiation, and the presence of foreign material, such as sutures. Although SBO can occur any time after laparotomy, the risk is greatest in the first few postoperative years. Lower abdominal or pelvic operations have a higher risk of adhesive SBO than do upper abdominal procedures, such as cholecystectomy. Several population-based and large case studies have defined the risk of adhesive obstruction after common abdominal operations. For example, the risk of adhesive SBO after appendectomy is about 1% with 30 years of follow-up; the risk is greater for patients with perforated

Colonic Obstruction  2116 Etiology  2116 Pathophysiology  2117 Clinical Features  2117 Diagnostic Evaluation  2117 Treatment and Outcome  2117

appendicitis (2.76% at 30 years) than nonperforated appendicitis (0.75% at 30 years).3 The incidence of adhesive SBO after gynecologic operations is similar to that of appendectomy except for cesarean delivery, which has a lower risk of subsequent SBO. Using Medicare data, Beck and colleagues found that the incidence of adhesive obstruction after intestinal resection and anastomosis was 14.3% within two years.4 About one fourth of patients who present with SBO have an etiology other than intraperitoneal adhesions. Of these, the most common causes are Crohn’s disease (7%) (see Chapter 111), intra-abdominal neoplasms (5%), and abdominal wall hernias (2%).2 A comprehensive list of causes of intestinal obstruction is shown in Table 119-1. Abdominal wall hernias (e.g., umbilical) and incisional hernias, as well as inguinal and femoral hernias, are much more common causes of SBO than internal or intraabdominal hernias (e.g., paraduodenal hernias). Less than 10% of patients presenting to the hospital with SBO have a hernia as the cause of their obstruction,2,5 and up to 30% of patients requiring an operation for SBO have an incarcerated hernia as the etiology of their obstruction.1 Intestinal obstruction caused by a hernia has a particularly high risk of strangulation, failure to resolve spontaneously, and recurrence when it is not surgically corrected. In one study of 877 patients, three fourths of patients presenting with incarcerated hernias and SBO had ischemic bowel at the time of operation; the bowel was necrotic in 27%.1 In contrast, only 29% of patients presenting with SBO due to adhesions had a strangulated obstruction, and only 11% had nonviable bowel.1 The increased risk of obstruction and strangulation is due, at least in part, to the rigid fascial defect through which the herniated intestine passes. Femoral hernias, in particular, pose a high risk of intestinal strangulation. Although SBO from a groin hernia can occur at any age, it is particularly prevalent in the elderly; advanced age, concomitant chronic illnesses, and treatment delay are associated with unfavorable outcomes in this subset of patients presenting with SBO.6 Internal hernias may be congenital (e.g., paraduodenal) or acquired (e.g., hernias through mesenteric defects created

2105

2106

Section X  Small and Large Intestine Table 119-1  Causes of Intestinal Obstruction Intrinsic Bowel Lesions Congenital atresia or stenosis Inflammatory causes Diverticulitis Inflammatory bowel disease (e.g., Crohn’s disease) Ischemic injury Radiation injury Intussusception Obturation Bezoars Feces Foreign bodies Gallstones Polypoid neoplasms Neoplastic strictures Surgical anastomosis Extrinsic Bowel Lesions Abscess Adhesions Carcinomatosis Congenital bands Endometriosis Hernias Volvulus

in the performance of intestinal anastomoses). The 3% incidence of internal herniation of the Roux limb after gastric bypass for weight loss is a particularly important example of an internal hernia, given the frequency with which this procedure is performed today.7 Congenital and acquired internal hernias are discussed in greater detail in Chapter 24. Various authors have reported herniation of a portion of the bowel wall (Richter’s hernia) or a whole segment of the intestine through a laparoscopic trocar site with resultant intestinal obstruction.8,9 The incidence of trocar site hernias is 1% to 2%, and intestinal obstruction is significantly less common.8 In these instances, the hernias usually occur at 10-mm port sites positioned at or close to the midline.9 Intestinal obstruction after laparoscopic transabdominal preperitoneal herniorrhaphy usually is due to herniation of the bowel through a defect in the peritoneal closure. In these instances, the bowel is tethered by adhesions between the partially peritoneal-covered prosthesis and the intestine, with formation of a kink or a point of torsion. Neoplasms are a relatively unusual cause of SBO, accounting for less than 10% of cases2,10; adenocarcinoma accounts for more than 50% of instances of colonic obstruction by neolasms. In more than 90% of such neoplastic SBO, the small intestine becomes obstructed by extrinsic compression or local invasion from advanced gastrointestinal or gynecologic malignancies. Advanced colorectal cancer and ovarian adenocarcinoma are the two most common malignancies associated with SBO. Hematogenous metastases from breast cancer, melanoma, or Kaposi’s sarcoma also can involve the small intestine, with subsequent obstruction. Primary neoplasms of the small intestine, of which adenocarcinoma and carcinoid tumors are most common, are the cause of SBO in less than 3% of cases (Chapter 121).

PATHOPHYSIOLOGY

The duration and degree of obstruction and the presence and severity of ischemia determine the local and systemic consequences of intestinal obstruction. The intestinal mucosa is an important and early site of injury in both

simple and strangulated intestinal obstruction. Microscopic evidence of epithelial injury occurs within the first four to six hours of simple intestinal obstruction and progresses to focal epithelial necrosis within 8 to 12 hours.11 Strangulated obstruction exacerbates the injury, causing extensive mucosal necrosis and sloughing. Intestinal obstruction causes the profound accumulation of fluid, swallowed air, and gas within the intestinal lumen proximal to the site of obstruction. Fluid accumulates because of impaired water and electrolyte absorption and also enhanced secretion, which results in the net movement of isotonic fluid from the intravascular space into the intestinal lumen. The accumulation of swallowed air, and to a lesser extent gases generated by bacteria within the obstructed bowel (e.g., hydrogen, carbon dioxide, and methane), contributes to intestinal distention. Failure of normal intestinal motility causes bacterial overgrowth within the small intestine and loss of the normally increasing concentration gradient of bacteria from the jejunum to the ileum. Disruption of the ecologic balance of the normal enteric microflora is associated with the translocation of bacteria to mesenteric lymph nodes and systemic organs. In a study by Deitch, enteric bacteria, particularly Escherichia coli, were cultured from mesenteric lymph nodes in nearly 60% of patients with simple intestinal obstruction compared with only 4% of controls.12 These observations are consistent with experimental studies that described the translocation of bacteria into the submucosa within 36 minutes of simple SBO.13 Together, these data are consistent with the hypothesis that translocating enteric bacteria contribute to the septic consequences of SBO. The systemic manifestations of SBO are related to hypovolemia and the inflammatory response incited by bacterial translocation, with or without the presence of ischemic or gangrenous intestine. Hypovolemia primarily results from the loss of fluid into the intestinal lumen, the bowel wall, and the peritoneal cavity. When combined with anorexia and vomiting, a marked reduction in intravascular volume results. Intestinal ischemia markedly exacerbates the loss of intravascular fluid locally into the intestine as well as systemically through a generalized microvascular leak. The generation and activation of proinflammatory mediators, including neutrophils, complement, cytokines, eicosanoids, and oxygen-derived free radicals, also has been linked to remote organ failure and mortality caused by intestinal is­chemia and reperfusion injury.

CLINICAL FEATURES History

Patients with SBO classically present with the acute onset of cramping mid-abdominal pain, vomiting, obstipation, and abdominal distention. The magnitude of symptoms depends upon the degree (complete or partial), site, and duration of the obstruction. Typically, patients describe paroxysms of periumbilical pain occurring at four- to fiveminute intervals for proximal obstructions and less frequently for more-distal obstructions. With prolonged obstruction, the cramping pain subsides as the motility in the distended intestine decreases. Patients with proximal intestinal obstruction have profuse vomiting, more-frequent pain, and minimal abdominal distention, whereas patients with distal obstruction present with less-frequent vomiting and much more abdominal distention. The emesis of patients with SBO is often feculent because of the increased bacterial count in the obstructed intestine. Although obstipation is an important symptom of SBO, patients with partial obstruction can continue to pass flatus and stool.

Chapter 119  Intestinal Obstruction Even patients with complete SBO evacuate the bowel contents distal to the obstruction. Continuous severe pain, particularly when localized, strongly suggests the presence of strangulated obstruction. Closed-loop obstructions are associated with the sudden onset of severe unremitting abdominal pain.

Physical Examination

In general, patients with simple SBO appear to be acutely ill, with abdominal distention and systemic evidence of intravascular volume depletion. Auscultation of the abdomen reveals periods of increased bowel sounds separated by intervals of relative quiet. The quality of the bowel sounds usually is described as high-pitched or musical. Borborygmi are pronounced rumbling bowel sounds that correspond with paroxysms of cramping abdominal pain. In the setting of prolonged obstruction, bowel sounds disappear as intestinal motility decreases. As alluded to earlier, the abdomen generally is distended and only minimally tender. Abdominal tenderness with guarding or other evidence of peritonitis suggests strangulation of the obstruction and necessitates urgent laparotomy. Patients with proximal SBO can have minimal abdominal distention if they have been vomiting. Patients with closed-loop obstructions can present with pain out of proportion to the physical findings, much like that of other causes of acute mesenteric ischemia, such as embolus in the superior mesenteric artery (SMA) (see Chapter 114). The presence of a tender mass at the site of an inguinal, femoral, or umbilical hernia strongly suggests that this is the etiology of the obstruction.

LABORATORY FINDINGS

Patients with suspected SBO should have blood drawn for a complete blood cell count (CBC) and serum electrolyte and creatinine concentrations. The CBC in patients with simple intestinal obstruction often reveals a slight leukocytosis as well as evidence of hemoconcentration. Significant neutrophilia and immature white blood cellular forms

suggest strangulated obstruction, although the predictive value of this parameter is too low to be useful as a sole determinant of strangulation. Vomiting and the profound loss of fluid and electrolytes from the intravascular space that accompanies SBO cause hemoconcentration, increased blood urea nitrogen and serum creatinine concentration, and abnormalities of serum sodium, potassium, and chloride concentrations. The presence of a metabolic acidosis suggests severe intravascular volume depletion with or without intestinal ischemia.

RADIOLOGIC FINDINGS Abdominal Plain Films

After history and physical examination, plain abdominal films are an inexpensive and reasonable first step in evaluating patients with suspected intestinal obstruction. Films taken with the patient in supine and upright positions may confirm the diagnosis of intestinal obstruction, localize the obstruction to the small intestine or colon, provide evidence of the degree of obstruction (partial or complete), and, if the upright image includes the diaphragm, detect pneumoperitoneum, thereby suggesting intestinal perforation. Classically, the abdominal films of patients with SBO demonstrate multiple dilated gas- or fluid-filled loops of small intestine with a decompressed colon (Fig. 119-1). The finding of dilated small bowel loops containing air-fluid levels is insufficient to distinguish SBO from ileus; however, when combined with an absence of colonic gas, the diagnosis of SBO becomes very likely. When dilated small bowel loops are accompanied by colonic distention, ileus or large bowel obstruction become more likely (Fig. 119-2). A gasless abdomen may be seen in patients with a very proximal SBO or those in whom the intestine is filled with fluid (Fig. 119-3). Lappas and colleagues reviewed 12 radiologic findings associated with SBO and found that the combination of airfluid levels of different heights in the same bowel loop and

*

A

B

Figure 119-1.  Supine (A) and erect (B) abdominal plain films in two patients with adhesive small bowel obstruction (SBO). A, Air-filled distended small bowel loops (arrows) with collapsed colon (*). Note that the small bowel folds (valvulae conniventes) typically extend completely across the intestinal loops. B, Multiple air-fluid levels in dilated small bowel loops (arrows) in the context of nondistended colon.

2107

2108

Section X  Small and Large Intestine SBO have equivocal or normal studies.16,17 False-negative plain films are most likely to occur with low-grade, proximal, or closed-loop obstructions, and in such cases further imaging may be diagnostic (see Fig. 119-3).

a mean air-fluid level diameter of 2.5 cm were most predictive of a high-grade partial or complete SBO.14 Thompson and coworkers corroborated the predictive value of these types of air-fluid levels and reported their sensitivity for SBO to be 59% to 93%.15 The limitations of abdominal plain films in determining the presence of intestinal obstruction are well recognized: 20% to 30% of patients with proven

Computed Tomography

Many studies support the use of abdominal computed tomography (CT) to evaluate patients with suspected intestinal obstruction.18,19 Advances in CT hardware and software provide high-resolution reconstructed images in any plane, thus enhancing image resolution and diagnostic confidence.20,21 Overall, CT is 90% to 95% sensitive, 96% specific, and 95% accurate in determining the presence of complete or high-grade SBO, and it provides important information regarding the site of obstruction and etiology in up to 95% of instances. CT findings of mechanical SBO are listed in Table 119-2 and illustrated in Figures 119-4 to 119-7. The demonstration of dilated, fluid- or gas-filled loops of proximal bowel and collapsed loops of distal bowel supports the diagnosis of intestinal obstruction. A transition point between bowel loops with disparate calibers may be identified (see Fig. 119-4A) and the degree of obstruction estimated by the amount of enteral contrast passing through the obstruction. Tapered bowel at the transition point can form a beak (see Fig. 119-3B), and a thorough search of this area can suggest the cause of obstruction. Although peritoneal adhesions are not usually seen on imaging studies, the presence of a transition point without another identifiable cause strongly favors adhesive obstruction. CT, however, can demonstrate tumors and other nonadhesive causes of bowel obstruction (see Fig. 119-6). Many classic CT signs have been proposed to distinguish simple from strangulated SBO (see Table 119-2)22-28; these include circumferential bowel wall thickening and edema (see Fig. 119-4B), ascites (see Fig. 119-5D), mesenteric engorgement, abnormal vessel course (see Fig. 119-5A),

* *

Figure 119-2.  Supine abdominal plain film in a patient with an ileus. The small intestine (arrows) and the colon (*) are significantly distended.

St

D

A

B

Figure 119-3.  Supine abdominal plain film (A) and coronal computed tomography (CT) image (B) in a patient with recalcitrant vomiting, obstipation, and a remote history of laparotomy for a gunshot wound to the abdomen. In this case, the abdominal plain film demonstrates a gasless abdomen. CT demonstrates proximal jejunal dilatation, with an abrupt caliber change (arrow) and no passage of contrast beyond the obstruction. Proximal small bowel obstruction was relieved by adhesiolysis. D, duodenum; St, stomach.

Chapter 119  Intestinal Obstruction altered enhancement of the bowel wall (see Fig. 119-7), and a bowel configuration suggesting a closed-loop obstruction or volvulus (see Fig. 119-4B). The small bowel feces sign refers to the presence of a mottled admixture of particulate matter and gas within the dilated bowel proximal to a lowgrade obstruction or in the setting of intestinal ischemia (see Fig. 119-5C).24,27 A closed-loop obstruction or small bowel volvulus is suggested by U- or C-shaped dilated bowel loops

Table 119-2  Computed Tomography Findings in Patients with Small Intestinal Obstruction Simple Complete Intestinal Obstruction

Proximal bowel dilatation; discrete transition zone, with collapsed distal small bowel and no passage of oral contrast beyond the transition zone Colon with little gas or fluid Small bowel feces sign

Closed-loop Obstruction

Bowel Wall Changes U-shaped, distended, fluid-filled bowel loop Whirl sign: tightly twisted mesentery around a collapsed bowel segment Beak sign: fusiform tapering in the longitudinal section of bowel at the site of obstruction Two adjacent collapsed round, oval, or triangular loops of bowel at the site of obstruction Mesenteric Changes Fixed radial distribution of several dilated bowel loops with stretched and thickened mesenteric vessels converging toward the point of obstruction

Strangulated Intestinal Obstruction

Bowel Wall Changes Bowel wall thickening with increased attenuation on unenhanced images Target or halo sign: concentric rings of slightly different densities Pneumatosis intestinalis linearis Poor enhancement or lack of enhancement of the bowel wall with intravenous contrast Serrated beak configuration of the obstructed bowel loop Mesenteric Changes A spectrum ranging from haziness and blurring of the mesenteric vessels to obliteration of the fatty mesentery and its vessels caused by mesenteric congestion and hemorrhage Diffuse engorgement of the mesenteric vasculature Unusual course of the mesenteric vasculature Other Changes Large amount of ascites High-density ascites

and a radial distribution of stretched mesenteric vessels converging toward a point of torsion (see Fig. 119-4B).22 The whirl sign25 also suggests intestinal torsion or volvulus and refers to a swirling mass of soft tissue and fat density that occurs as the bowel rotates on its mesentery (see Fig. 119-5B). Decreased bowel wall enhancement is the most specific finding of intestinal ischemia (see Fig. 119-7), but its sensitivity is relatively low in most series (34% to 56%).23,24,28 Porto-mesenteric venous gas, pneumoperitoneum, and pneumatosis intestinalis linearis (see Fig. 119-5D) may be seen very late in the natural history of strangulated obstruction and suggest the presence of extensive necrosis. Despite these many signs, the early diagnosis of strangulated SBO remains a challenge with reported sensitivity, specificity, and accuracy of these various criteria to detect strangulated obstruction ranging from 14% to 95%.24 Detection of low-grade or intermittent bowel obstruction may be extremely difficult using standard radiologic and CT approaches; diagnostic accuracy varies from 48% to 66%.29 In these settings, CT enteroclysis (Fig. 119-8) combines the advantages of active luminal distention with the mural and extraenteric evaluation of cross-sectional imaging; this technique has been shown to raise the accuracy for detecting small bowel diseases to nearly 100%.30,31 Magnetic resonance imaging also has been used to detect SBO and to characterize benign and malignant causes,32 although its greater cost, lower spatial resolution, and lack of incremental diagnostic gain over CT has limited its widespread implementation.33

Barium and Water-Soluble Small Bowel Contrast Studies

Fluoroscopic studies of the gastrointestinal tract with enteral contrast agents (barium sulfate or Gastrografin) have long been used to evaluate patients with suspected SBO, particularly when the clinical presentation is atypical, abdominal plain films are nondiagnostic, and lower grades of bowel obstruction are suspected. Contrast radiography with barium sulfate has been shown to provide useful information—definite diagnosis, no obstruction, high-grade or complete obstruction—in 50% to 80% of patients examined,34 and several studies suggest that passage of orally administered Gastrografin into the colon within six to 24 hours can identify patients most likely to respond to nonoperative management.35-37 The disadvantages of luminal study approaches, including prolonged examination times, the need to ingest large volumes of contrast agent, and limited visibility of mucosa at transition point, have caused

*

* *

*

*

* A

B

Figure 119-4.  Axial computed tomography images of simple (A) and closed loop (B) small bowel obstructions. A, The small bowel is dilated proximal to a transition point (arrow) and collapsed distally. Rectal contrast identifies the colon (*). B, Dilated, fluid-filled small bowel loops (*) are radially arranged, and several demonstrate concentric rings of wall thickening and submucosal edema. There is complete loss of mesenteric fat (arrow). Infarcted ileum was resected at surgery.

2109

2110

Section X  Small and Large Intestine

*

A

B

C

D

*

Figure 119-5.  Computed tomography signs in small bowel obstruction. A, Abnormal position of mesenteric vessels: The superior mesenteric artery (*) is anterior to the superior mesenteric vein. B, Whirl sign (arrow). C, Small bowel feces sign (arrow). D, Pneumatosis (arrow) and large-volume ascites. The colon (*) is collapsed (see Table 119-2).

fluoroscopic small bowel studies to be largely replaced with CT or CT enteroclysis. In patients with chronic partial SBO, fluoroscopic imaging of the site of obstruction may be facilitated by barium enteroclysis, which can characterize the site of obstruction accurately in 86% to 100% of instances (Fig. 119-9).17 After intubating the duodenum or proximal jejunum, contrast material is delivered directly into the small intestine, and the small bowel then is evaluated for distensibility, stenoses, masses, and inflammatory changes. The degree of intestinal obstruction may be determined by challenging luminal distention and by identifying the arrival and passage of contrast through an area of narrowing. This study is, of course, contraindicated in patients presenting acutely with obstructive symptoms or evidence of a complete or highgrade SBO. In patients with combined small bowel and colonic distention, a Gastrografin contrast enema or CT with rectal contrast may be used to distinguish ileus from a distal large bowel obstruction.

TREATMENT AND OUTCOME Medical Management

The initial step in the management of patients with SBO is the restoration of intravascular volume by the infusion of isotonic fluid. A Foley catheter should be placed to facilitate assessment of the patient’s functional intravascular volume status and hence the adequacy of resuscitation. Abnormalities of serum electrolyte concentrations must be corrected as rapidly as possible, and metabolic acidosis should be treated by the restoration of normal intravascular volume.

Persistent acidosis suggests the presence of ischemic bowel and hence strangulated obstruction. A nasogastric (NG) tube should be placed to decompress the stomach and minimize further intestinal distention. In addition to reducing the discomfort associated with gastric and intestinal distention, NG decompression also can reduce the risk of aspiration. Although there are no data supporting the use of systemic antibiotics in the nonoperative management of patients with simple SBO, patients who are to undergo laparotomy should receive broad-spectrum antibiotics directed toward Gram-negative aerobes and anaerobes. Second-generation cephalosporins such as cefoxitin or cefotan are used commonly. In the absence of clinical evidence of strangulated obstruction or an incarcerated hernia, 64% to 73% of patients with SBO can be managed successfully with fluid and electrolyte resuscitation and NG aspiration.5,38 Foster and colleagues reported that 76% of 32,583 patients with SBO admitted to acute care hospitals in California in 1997 were managed without operation.5 In contrast, less than half of patients presenting with a complete SBO are managed successfully nonoperatively.38 Most patients who can be successfully managed nonoperatively have substantial improvement within the first 48 hours of treatment. Clinical deterioration, or even the failure to improve, mandates urgent operation. Of those patients in whom nonoperative management fails, almost 20% have strangulated obstruction.38 As alluded to earlier, patients whose SBO is most likely to resolve without operation may be identified by the appearance of oral contrast within the

Chapter 119  Intestinal Obstruction

*

*

N

*

A

*

Figure 119-8.  Image from coronal computed tomography enteroclysis with gastrointestinal and intravenous contrast. Neutral enteral contrast instilled via a transduodenal tube (*) aids in the detection of active mucosal hyperemia (arrow) and lower grades of luminal narrowing in this patient with Crohn’s disease. N

B Figure 119-6.  Axial computed tomography image (A) and intraoperative photograph (B) of partial small bowel obstruction caused by carcinoid tumor. The small bowel is dilated (*) proximal to the partially calcified primary lesion (arrow), and a metastatic nodal mass is evident in the mesentery (N).

*

Figure 119-7.  Computed tomography image revealing a strangulated obstruction of the small intestine. A long segment of obstructed small bowel wall does not enhance (arrows) after intravenous contrast, indicating poor perfusion. Note the high-density mesenteric fluid (*). Small bowel volvulus and impaired perfusion were found at surgery.

cecum on abdominal films within four to 24 hours of administration (sensitivity, 0.96; specificity, 0.96).39 CT, with enteral and parenteral contrast, can provide the same prognostic information with regard to the passage of oral contrast into the cecum; in addition, this modality also can provide radiologic evidence of ischemic bowel. Although some studies have suggested that orally administered Gastrografin may enhance the likelihood of successful nonoperative management of SBO,37 the preponderance of data suggests that it has no therapeutic effect.39 Similarly, few data suggest that placement of a long tube into the small intestine improves the likelihood of successful nonoperative management compared with a standard NG tube for decompression. It is our opinion that if long intestinal tubes have a place in the management of patients with SBO, it is in patients with prior episodes of intestinal obstruction from adhesions, radiation enteritis, or Crohn’s disease.

Surgical Management

The operative management of SBO entails relief of the obstruction and resection of any gangrenous bowel. The point of obstruction often can be identified by a transition zone of dilated intestine proximal to the point of obstruction and decompressed bowel distal to it. In the absence of frankly necrotic intestine, viability should be assessed several minutes after release of the obstruction. Return of normal color and peristalsis and return of arterial pulsation in the vasa recta suggest that the involved segment is viable. The detection of blood flow at the antimesenteric surface with a Doppler flowmeter and inspection of the bowel with a Wood’s lamp following intravenous injection of fluorescein are other useful techniques to assess the viability of a segment of obstructed intestine. Although laparoscopy has revolutionized the management of many gastrointestinal diseases, its use for patients with SBO, first reported in 1991, has not been implemented

2111

2112

Section X  Small and Large Intestine

A

The morbidity and mortality rates associated with the laparoscopic management of patients with SBO compare favorably with those of open approaches. In a systematic review of 19 studies that included 1061 patients, Ghosheh and Salameh reported that the overall complication rate for laparoscopic treatment of SBO was 15.5%; the mortality rate was 1.5%.43 The incidence of iatrogenic perforation of the bowel was 8.4% in the largest series of patients undergoing laparoscopic treatment of SBO (308 patients from 35 surgical centers);40 this figure is similar to the average incidence of 6.5% reported in Ghosheh and Salameh’s collective review.43 The risk of intraoperative bowel injury can be mitigated by using endoscopic scissors rather than monopolar electrocautery to lyse adhesions; beginning the bowel exploration from the cecum and collapsed distal small bowel loops, and working proximally toward the transition zone and dilated intestine; dividing only those adhesions relevant to the obstruction; gently manipulating the dilated bowel using atraumatic intestinal clamps placed on the mesentery instead of the intestinal wall; and using an open Hasson technique by which entry into the peritoneal cavity is performed under direct visualization, entering the abdomen at a site remote from any previous incisions.

Outcome

B Figure 119-9.  Nonadhesive small bowel obstruction. An abnormal but nonspecific bowel gas pattern on plain film (A) is confirmed to be from a chronic ischemic stricture (arrow) when further evaluated by tube enteroclysis (B).

as widely as for other conditions. The advantages of this operative approach include a quicker resolution of postoperative ileus, a reduction in the length of hospital stay, and fewer complications.40-42 Laparoscopic treatment of SBO, however, is limited by the technical difficulties associated with exploring the abdomen in the presence of markedly distended loops of small intestine and the perceived heightened risk of injury to the distended and thin-walled bowel. In experienced hands, laparoscopic treatment of SBO is successful in about one half to two thirds of patients,40,41,43 especially when there has been only one or two previous abdominal operations, and if the antecedent procedure was an appendectomy. Patients who present to the hospital relatively early after the onset of symptoms and are operated on within 24 hours of admission do better than those who have a longer trial of nonoperative management. The most common reasons for conversion to laparotomy are the presence of dense adhesions,40 the need for intestinal resection, an iatrogenic injury to the bowel, or an inability to identify the site of the obstruction.43

The outcome of patients with SBO may be considered in terms of immediate risk of morbidity and mortality and the long-term risk of recurrent obstruction. The risk of death associated with intestinal obstruction ranges from 2% to 8% in studies published since 2000 (Table 119-3).5,38,44-46 The most commonly cited risk factors for increased risk of death are advanced age and American Society of Anesthesiologists (ASA) class 3 or higher. Chronic illnesses, strangulated obstruction, and treatment delay also have been associated with an increased mortality risk.1,45,46 The incidence of major complications after operative treatment of SBO ranges from 12% to 47% depending upon the procedure performed; bowel resection poses a greater risk of complications than does simple lysis of adhesions. The most common medical complications include pneumonia and respiratory failure, pulmonary embolism, cardiac complications, and prolonged ileus. The most common surgical complications are wound infections, intra-abdominal infections, intra-abdominal bleeding, and intestinal necrosis and perforation.1,44-46 Once a patient has developed adhesive SBO, the rate of recurrent obstruction is high. In one retrospective study of 500 patients who underwent operative treatment of SBO, the rate of recurrent obstruction was 18% with 10 years of follow-up and 29% at 30 years.47 These figures are similar to those of Miller and coworkers, who found that one third of 410 patients presenting to the hospital with adhesive SBO subsequently were readmitted with recurrent obstruction.48 Similarly, Foster and colleagues reported that 18% of 32,583 patients admitted to California hospitals in 1997 with a diagnosis of SBO had a subsequent admission for the same diagnosis within five years.5 Although obstruction can recur at any time in the patient’s life, most (50% in Fevang’s study) occur within the first five years after the index operation for obstruction.47 Some evidence suggests that operative treatment of adhesive SBO is associated with a lower rate of recurrent obstruction47 and a longer interval between admissions than nonoperative management.5,48 Miller and coworkers related the likelihood of recurrent obstruction to the pattern of peritoneal adhesions: Obstruction from a single adhesive band had a 25% risk of recurrence, compared with a 49% incidence of recurrence when adhesions were dense and matted.48

Chapter 119  Intestinal Obstruction Table 119-3  Morbidity and Mortality in Patients with Small Bowel Obstruction

STUDY POPULATION

YEARS OF TREATMENT

1

877 patients who underwent surgery

1961-1995

5%

Retrospective chart review

25.6%

5

32,583 patients managed operatively (7,935) and nonoperatively (24,648) 166 patients with SBO

1997

Entire group: 7% Operated group: 5% Nonoperated group: 8%

Longitudinal, population-based

—

1994-1995

Prospective

12%

NSQIP database

37% of patients having lysis of adhesions and 47% of patients having bowel resection had >1 complication 14%

REFERENCE

44

MORTALITY RATE

45

2002 patients who underwent surgery

1991-2002

Total: 4% Death with partial SBO: 1% Death with complete SBO: 5% 7.7%

46

286 patients who underwent surgery

1997-2001

3%

STUDY DESIGN

Prospective, multicenter

COMPLICATION RATE

OUTCOMES AND RISK FACTORS FOR POORER OUTCOMES Patient’s age Premorbid illness Nonviable strangulation Treatment delay >24 hr Patients managed nonoperatively had a greater mortality rate (8%) than those managed operatively (5%)

Dirty or infected wounds ASA 4 or 5 Age > 80 years Dyspnea at rest Age > 75 years ASA ≥ 3

ASA, American Society of Anesthesiologists class; NSQIP, National Surgical Quality Improvement Program; ref, reference; SBO, small bowel obstruction.

SPECIAL CONSIDERATIONS Early Postoperative Obstruction

The most important clinical feature differentiating early postoperative SBO from postoperative ileus is the occurrence of obstructive symptoms after an initial return of bowel function and resumption of oral intake. Generally, patients with early postoperative SBO develop nausea and vomiting, abdominal distention, and abdominal pain as early as the fourth to eighth postoperative day. Plain films of the abdomen usually demonstrate dilated loops of small intestine with air-fluid levels and a paucity of gas in the colon. In contrast, abdominal plain films of patients with an uncomplicated postoperative ileus often have gas in both the large and small intestine. CT confirms the presence of a partial SBO and can provide important evidence of an internal hernia, volvulus, or strangulated obstruction. Furthermore, CT can identify other complications of the index operation that may be related to the occurrence of the obstruction, such as intra-abdominal abscess. As with other patients with SBO, initial management of postoperative SBO begins with intravascular volume resuscitation and NG aspiration. In one series, 20 of 23 such cases resolved with nonoperative management alone, all within six days of treatment.49 Although strangulated postoperative SBO is unusual, evidence of clinical deterioration with the development of severe abdominal pain, hemodynamic instability, or evidence of peritonitis suggests strangulation and mandates immediate laparotomy. In patients who develop SBO early after laparoscopy, the surgeon must suspect that a portion of the intestine herniated through a trocar site, and urgent operative relief of the obstruction is required.

Small Bowel Obstruction in Patients with Malignancies

Peritoneal spread of gastrointestinal and gynecologic malignancies is an important cause of SBO. In both colorectal and ovarian cancer, the risk of developing malignant obstruction is strongly linked to the initial stage of the disease. For example, Ellis and colleagues found that only 18% of patients who developed SBO after undergoing colectomy for early-stage colon cancer had recurrent cancer as the cause of their obstruction, whereas carcinomatosis was the cause of obstruction in 82% of patients with more-advanced disease.50 Similarly, Tunca and colleagues found that only 15% of patients with stage 1 ovarian cancer eventually developed malignant SBO compared with 35% of patients with more-advanced disease.51 In the absence of recurrent malignancy, intestinal obstruction may be due to postoperative intraperitoneal adhesions or fibrosis from radiation therapy. Gadolinium-enhanced MRI and positron emission tomography can assist in differentiating malignant from benign causes of intestinal obstruction in patients with a history of malignancy. Because most instances of SBO from recurrent malignancy represent incurable disease, the goal of treatment is to improve the quality of life over the remainder of a limited life expectancy. In such instances, the purpose of surgery is to relieve the symptoms of obstruction. Factors to consider in planning operative therapy include the chance of successful palliation, the risk of repeat obstruction, the quality of life for the patient after surgery, the ability to administer future chemotherapy, and the risk of operative morbidity and mortality. Surgical options include resection with reanastomosis, surgical bypass with an enteroenterostomy, or diverting ileostomy. In many instances, symptoms can be

2113

2114

Section X  Small and Large Intestine palliated successfully by operation, at least for a time.52 In one retrospective study of 64 patients with recurrent ovarian cancer, the obstruction was relieved surgically in 84% of cases, and 71% of patients were able to tolerate a diet for at least 60 days postoperatively53; the median survival of those patients in whom the obstruction was relieved was about 12 months. Unfortunately, the incidence of operative and postoperative complications is high for patients undergoing an operative approach to malignant intestinal obstruction.52

V

*

Intussusception

Intussusception is the invagination of a proximal segment of bowel (intussusceptum) into an adjacent distal segment (intussuscipiens) (Figs. 119-10 and 119-11). Although more often recognized as a cause of SBO in children, about 5% of cases occur in adults. In contrast to children, in whom there is rarely an anatomic abnormality of the intestine, small bowel intussusception in adults is associated with an underlying pathologic process in 80% to 90% of cases. Benign tumors are the most common lead point in the small intestine, whereas adenocarcinoma is the most common lead point for ileocolic and colocolic intussusceptions. The incidence of intussusception peaks in children between the ages of four and seven months, and most cases occur before the child is one year of age. Most children present with the acute onset of vomiting, bloody stools, and abdominal pain; only about 20% have all three of these classic symptoms. In one study of 244 children with intussusception, only 6% had all three of these symptoms and an abdominal mass on examination.54 Adults with intussusception present with cramping abdominal pain and symptoms of intestinal obstruction that often are several weeks in duration. An abdominal mass and evidence of gastrointestinal bleeding or anemia are present in a minority of cases.55-57 In most medical centers, ultrasonography (US) is the initial procedure of choice for diagnosing intussusception in the pediatric population. The accuracy of this technique is greater than 90% in experienced hands,58 and its noninvasive nature and lack of ionizing radiation represent sig-

A

B

A

* V

B Figure 119-10.  Abdominal ultrasonography (A) and axial computed tomography (B) images in a patient with intussusception. The intussuscepted bowel, or intussusceptum (arrow), is collapsed and carries mesenteric fat (*) and vasculature into the dilated intussuscipiens. Note the dilated mesenteric vessels (V).

C

Figure 119-11.  Air reduction of an ileocolic intussusception in a child. Retrograde instillation of gas under closely monitored pressure and time may be successful in progressive reduction (A through C) of the intussusceptum (arrows). To ensure complete reduction, the goal of the procedure is reflux of air into the ileum.

Chapter 119  Intestinal Obstruction nificant advantages over contrast enemas.59 Most practices reserve the air or liquid enema for reduction of intussusceptions found by US in suitable candidates. In nonpregnant adults, the procedure of choice for diagnosing intussusception is CT. CT and US detect the double bowel of the intussusceptum and intussuscipiens along with variable amounts of mesenteric fat (see Fig. 119-10). Concentric rings of inner and outer bowel appear hypoechoic on US (intermediate density on CT), with the characteristic hyperechoic fat (low density on CT) between the intussusceptum and intussuscipiens. Images obtained in long axis or obliquity might appear reniform or sausage shaped. Most pediatric ileocolic intussusceptions occur in the subhepatic region. Seventy percent to 80% of intussusceptions in the pediatric population may be successfully managed nonoperatively by reduction with air- or liquid-contrast enemas under fluoroscopic or US guidance (see Fig. 119-11). The successful reduction of pediatric intussusception is higher with air-contrast enemas than with liquid enemas, and the success of this technique diminishes substantially as the duration of symptoms exceeds 24 hours.60 The overall rate of bowel perforation is low, and recurrent intussusception occurs in about 10% of cases.59 Obvious contraindications to enema reduction include peritonitis, shock, sepsis, or pneumoperitoneum. In adults, the association of neoplasms and other intestinal pathology with intussusception mandates resection of the involved bowel and makes hydrostatic or pneumatic reduction unreasonable. Primary resection without attempting reduction is the preferred treatment for colonic intussusception, including ileocecal intussusception. For lesions involving the right colon, a right hemicolectomy is performed. When the intussusception involves only the small bowel, resection is the preferred operative approach, although manual reduction of the intussusception with careful palpation of the intestinal wall might allow the surgeon to limit the amount of bowel resected.

Gallstone Ileus

Gallstone ileus is an unusual cause of intestinal obstruction, accounting for about 1% to 4% of all cases (see Chapters 65 and 66).61 This complication of cholelithiasis is more common in the elderly, most patients presenting in the

A

B

seventh or eighth decades of life. The term “gallstone ileus” is a misnomer, because this condition represents a true mechanical obstruction of the intestine by a gallstone or gallstones within the lumen of the bowel. Most commonly, gallstones large enough to cause obstruction enter the bowel via a cholecystoduodenal fistula. As the stone migrates through the intestinal tract, it produces intermittent obstruction, with resultant waxing and waning of symptoms, thereby confounding early diagnosis. The most common site of obstruction is the ileum, accounting for about 60% of cases; this is followed by the jejunum (15%), stomach (15%), and colon (5%). In the absence of an intestinal stricture, a gallstone of at least 2 cm is required to cause intestinal obstruction (Fig. 119-12). The diagnosis of gallstone ileus is delayed in up to half of the patients because of nonspecific and inconsistent symptoms. Only 50% to 70% of patients have clinical features of SBO. The classic radiologic features of gallstone ileus include pneumobilia, intestinal obstruction, aberrant gallstone location, and a change in the location of a previously observed stone. Only about 10% of gallstones are sufficiently calcified to be identified on abdominal plain films, although the increasing use of radiologic imaging in the emergency department has led to descriptions of gallstone ileus on CT, MR, and US. In 2005, Lassandro and colleagues reported on the use of CT in 40 patients with surgically proven gallstone ileus; pneumobilia and the biliary-enteric fistula were identified in nearly 90% and 12% of cases, respectively.62 Treatment of gallstone ileus is focused on removing the obstructing stone, usually by operative enterolithotomy; endoscopic removal of the stone with or without lithotripsy also has been reported. In general, enterolithotomy alone is the appropriate initial treatment given the emergent nature of this procedure, the advanced age of many of these patients, and the common presence of a complex right upper quadrant mass containing the cholecystoenteric fistula. Together, these factors argue against identification and repair of the fistula at the time of emergent laparotomy for intestinal obstruction. Elective cholecystectomy with repair of the fistula generally is performed after the patient has recovered from the initial operation because up to 17% of patients develop recurrent gallstone ileus or other biliary complications after enterolithotomy alone.63 Lobo and

C

Figure 119-12.  Gallstone ileus in a 78-year-old woman with a two-week history of nausea and vomiting. A, Plain film of the abdomen revealing characteristic features of gallstone ileus, including air in the biliary tree (arrowheads) and dilated loops of small intestine (arrows); an obstructing stone was not seen on this plain film examination. B, At exploratory laparotomy, two obstructing gallstones (arrows) were identified in the jejunum, with adjacent perforations and fecal contamination of the peritoneal cavity. C, The stones were removed via enterotomy and segmental resection of the jejunum. In addition, the cholecystoduodenal fistula was repaired and cholecystectomy was performed. (From Besselink MG, Kroeze J. Gallstone ileus. Mayo Clin Proc 2005; 80:699.)

2115

2116

Section X  Small and Large Intestine coworkers, however, reported 11 elderly patients who presented with gallstone ileus and were treated with enterolithotomy alone; none of these patients developed subsequent biliary symptoms with nearly four years of follow-up.61

*

Midgut Volvulus

Midgut volvulus due to intestinal rotational anomalies is an important cause of intestinal obstruction, particularly in neonates. The anatomy and embryology that underlies the development of duodenal obstruction from Ladd’s bands and midgut volvulus is discussed in Chapter 96. It is worth reiterating, however, that the most common anomaly associated with midgut volvulus is nonrotation, in which there is inadequate counterclockwise rotation of the midgut loop around the SMA. This limited rotation results in the duodenojejunal junction and the entire small intestine being located to the right of the midline, while the colon resides in the left abdomen and the cecum is positioned near the midline. The narrow mesenteric pedicle predisposes the patient to midgut volvulus, and the peritoneal attachments (Ladd’s bands), which extend anterior and lateral to the duodenum to fix the cecum to the posterior body wall, can obstruct the duodenum. Fifty percent to 75% of malrotations are detected during the first month of life and more than 90% are discovered within the first year. The discovery of malrotation at this time usually results from clinical evidence of duodenal obstruction from Ladd’s bands or midgut volvulus. Infants with duodenal obstruction from Ladd’s bands present with the clinical signs and symptoms of gastric and proximal duodenal obstruction, including bilious vomiting with minimal abdominal distention. Midgut volvulus also pre­ sents with obstructive symptoms and symptoms of intestinal ischemia. Occult bleeding is common, and if transmural necrosis develops, acidosis, thrombocytopenia, and frank sepsis can ensue. The high risk of intestinal ischemia and necrosis from midgut volvulus and the associated very high mortality rate mandates early diagnosis and aggressive management of neonatal intestinal obstruction. An abdominal plain film that demonstrates a distended stomach and proximal duodenum with a paucity of small bowel gas suggests obstruction, which can obviate further radiologic evaluation. A normal plain film does not exclude malrotation, however, and direct imaging of the stomach and small intestine with an upper gastrointestinal series (UGIS) is needed to confirm or exclude mechanical causes of bilious vomiting. A diagnostic UGIS demonstrates duodenojejunal junction malposition and can illustrate the characteristic corkscrew or coiled appearance of the duodenum and proximal jejunum when malrotation is complicated by midgut volvulus (Fig. 119-13). The use of a contrast enema to demonstrate malrotation is less direct than UGIS and can have false positive and false negative rates of 15% and 20%, respectively. Treatment of intestinal malrotation, whether manifested by duodenal obstruction from Ladd’s bands or midgut volvulus, is operative. In the case of midgut volvulus, the diagnosis should be followed immediately by laparotomy because a delay of even hours could mean the difference between viable or infarcted intestine. Operative repair of malrotation is achieved by the performance of the Ladd procedure. The first part of the Ladd procedure consists of relieving the midgut volvulus and dividing the peritoneal bands that tether the cecum, small bowel mesentery, mesocolon, and duodenum around the base of the SMA. This maneuver allows the mesenteric leaves to open widely and is associated with a very low incidence of recurrent volvu-

Figure 119-13.  Upper gastrointestinal series via nasogastric tube (*) in a patient with midgut volvulus. The third and fourth portions of the duodenum are abnormally vertical, and the intestine spirals around the superior mesenteric artery axis to produce a corkscrew sign (arrows).

lus. Then Ladd’s peritoneal bands are divided to relieve the extrinsic compression and obstruction of the distal duodenum. This is accomplished by meticulous and complete mobilization of the entire duodenum with division of all anterior, lateral, and posterior attachments. Finally, the bowel is repositioned and fixed, with the small intestine on the right side and the colon on the left side of the abdomen.

COLONIC OBSTRUCTION ETIOLOGY

The three most common causes of colonic obstruction are adenocarcinoma of the colon and rectum, volvulus, and benign stricture from diverticulitis. These three conditions account for about 90% of cases of colorectal obstructions. Adenocarcinoma causes more than half of all cases of colonic obstruction. About 20% of patients with colorectal cancer present with obstruction, half of whom require emergency treatment. Most obstructing colon cancers occur distal to the splenic flexure, where the colon diameter is relatively small and the stool is relatively solid; scirrhous tumors cause progressive luminal narrowing. Cancers occurring in the right colon also can cause obstruction when they reach a size that occludes the lumen or by acting as the lead point of an intussusception. Colonic volvulus is the axial twisting of the colon on its vascular pedicle. The anatomic factors necessary for the development of volvulus include a redundant segment of bowel that is freely movable within the peritoneal cavity, a long movable mesocolon, and close approximation of two points of fixation of the colon. Typically, a closed-loop obstruction is produced that causes ischemia of the bowel wall due to twisting of the vascular pedicle and increased wall tension from colonic distention. The sigmoid colon and cecum are the most common sites of colonic volvulus, accounting for about 75% and 22% of all cases, respectively. Rarely, volvulus occurs at the transverse colon (2%)

Chapter 119  Intestinal Obstruction or the splenic flexure (1%). Although colonic volvulus causes only 15% of colonic obstructions in the United States and other western countries, it is a significantly more common cause of colonic obstruction in parts of Africa and the Middle East. Less than 10% of cases of colonic obstruction are due to a fibrotic inflammatory stricture caused by acute diverticulitis. The complications of diverticulitis are discussed further in Chapter 117.

PATHOPHYSIOLOGY

The competency of the ileocecal valve is of great importance in the pathophysiology of colonic obstruction. When the ileocecal valve is competent, the obstructed colon cannot decompress fluid and gas into the small intestine; a closedloop obstruction results, ultimately with colonic ischemia and perforation. In the setting of colonic obstruction, the cecum is most susceptible to ischemia because of the direct relationship between wall tension and radius as defined by Laplace’s law (T = P × R, where T is wall tension, P is pressure, and R is radius). The cecum has a greater diameter than any other segment of colon, as a result of which, as intraluminal pressure and the diameter of the obstructed colon increase, the increase in wall tension is greatest in the cecum. When wall tension exceeds capillary perfusion pressure, the bowel wall becomes ischemic. Generally, acute dilatation of the cecum to 10 cm suggests colonic wall ischemia, and a diameter greater than 13 cm implies imminent perforation.

CLINICAL FEATURES

Many of the clinical manifestations of colonic obstruction are similar regardless of etiology. Periumbilical or hypogastric pain and abdominal distention are the two most common presenting features. The pain varies from a vague discomfort to the excruciating pain of peritonitis. Severe unremitting pain suggests gangrenous bowel and requires urgent laparotomy. Patients can experience either diarrhea (reflecting the passage of liquid stool around an obstructing lesion) or obstipation, depending upon the degree and location of the obstruction. Patients with left-sided colonic tumors or benign fibrotic strictures often have noted a change in stool caliber over the recent months. Blood in the stool, or an iron-deficiency anemia, is highly suggestive of carcinoma, as is the occurrence of weakness, weight loss, and anorexia. Vomiting, when present, is usually a late finding. Symptoms of malignant and diverticular-associated obstruction often are insidious in onset, with a median duration of three months. In one study, 25% of patients with obstruction from colo­rectal cancer had symptoms for six to 24 months before diagnosis.64 Patients with sigmoid volvulus are usually in their sixth to eighth decades of life and often have concomitant chronic illnesses. Acute abdominal distention is the most common presentation for patients with colonic volvulus, occurring in about two thirds of patients. About 20% of patients have abdominal pain, nausea, vomiting, and constipation. The duration of symptoms of colonic volvulus is significantly less than that seen with a malignant or diverticular stricture. In one series of 228 patients with colonic volvulus, the average duration of symptoms before presentation was 73 hours.65 Abdominal tenderness is present in less than one third of patients with volvulus, and peritonitis suggests impending or actual colonic necrosis and perforation. Patients with cecal volvulus tend to be younger than patients with sigmoid volvulus and often have a history of prior abdominal operations. Nearly one third of these patients have a concomitant partially obstructing lesion

located more distal in the colon. A history of chronic constipation and laxative use is also common in patients with cecal volvulus.

DIAGNOSTIC EVALUATION

The initial diagnostic approach to patients with suspected colonic obstruction is similar to that of patients with SBO. Plain abdominal films in the supine and upright positions should be obtained to determine the site of the obstruction and whether the obstruction is partial or complete. As alluded to earlier, small bowel distention may or may not be present in patients with colonic obstruction depending upon the competency of the ileocecal valve (Fig. 119-14). Abdominal plain films also may be diagnostic, or at least highly suggestive, of sigmoid and cecal volvulus. As described by Agrez and Cameron,66 the standard radiologic feature of sigmoid volvulus is a distended ahaustral sigmoid loop—a bent inner-tube appearance—whose apex is directed toward the right shoulder (Fig. 119-15A). Radiologic features of cecal volvulus include a massively dilated cecum located in the epigastrium or left upper quadrant, a distended kidney bean-shaped cecum, distended loops of small bowel suggesting SBO, and a single long air-fluid level present on upright or decubitus films (see Fig.119-15B). Although these classic radiographic findings of colonic volvulus are seen in only 40% to 60% of cases, the diagnosis of colonic volvulus can be made with abdominal films alone in up to 85% of cases. Performing further diagnostic studies in patients with suspected colonic obstruction is predicated on the presence or absence of peritonitis and the degree of obstruction (partial or complete). Patients with peritonitis should undergo resuscitation and urgent laparotomy without additional diagnostic procedures. In contrast, patients whose abdominal plain film suggests a distal obstruction and who do not have evidence of peritonitis or strangulated obstruction should undergo a water-soluble contrast enema or CT with rectal contrast to confirm and localize the site of obstruction (Figs. 119-16 and 119-17). In patients with suspected sigmoid or cecal volvulus, a water-soluble contrast enema may be helpful by demonstrating a point of torsion (a mucosal spiral pattern or beak sign) without the risk of barium peritonitis in the case of unrecognized perforation (Fig. 119-18). In patients with abdominal plain films suggesting a sigmoid volvulus, initial rigid or flexible proctosigmoidoscopy can allow prompt decompression. Although colonoscopy may be useful in patients with partial colonic obstruction, it has little role in the initial evaluation of patients with suspected complete obstruction. Insufflation of air or carbon dioxide through the endoscope into the obstructed bowel can exacerbate colonic distention, further reduce colonic blood flow, and precipitate perforation.

TREATMENT AND OUTCOME

Initial management of patients with colonic obstruction consists of fluid and electrolyte resuscitation and NG aspiration. As alluded to earlier, clinical evidence of peritonitis or strangulated obstruction mandates emergent laparotomy. In the absence of these findings, the therapeutic goals for managing patients with colonic obstruction are prompt decompression of the obstructed colon, definitive treatment of the obstructing lesion, and re-establishment of intestinal continuity.

Malignant and Benign Colonic Strictures

Numerous studies have documented the successful use of colonoscopically placed self-expanding metal stents to

2117

2118

Section X  Small and Large Intestine

A

B

Figure 119-14.  Abdominal plain film in a patient with a high-grade colonic obstruction and competent ileocecal valve. A, Note the markedly distended, gas-filled colon without dilatation or gas in the small intestine. The etiology of the patient’s colonic obstruction was a left inguinal hernia, which is demonstrated on the water-soluble contrast enema (B). Although a common cause of small intestinal obstruction, inguinal hernias rarely obstruct the colon. The arrow demonstrates the point of obstruction of the sigmoid colon as it traverses the inguinal canal.

*

Cecum

Sigmoid

Figure 119-15.  Abdominal plain films of sigmoid (A) and cecal (B) volvulus. A, In sigmoid vol­vulus, the right, transverse, and left colon are distended (*) upstream from the point of sigmoid obstruction (arrow). B, In cecal volvulus, note the bean shape and left upper quadrant location of the dilated twisted cecum, and collapsed distal colon (*).

*

A

relieve malignant colonic obstruction before definitive resection or to palliate obstructive symptoms in patients with advanced disease. Successful management has been shown in up to 80% to 90% of patients in whom colonoscopic stent placement has been attempted.67-69 This approach allows mechanical bowel preparation with subsequent partial colectomy and primary anastomosis using either laparoscopic or open operative techniques or treatment of significant concomitant medical illnesses.67 In Khot’s systematic review of 262 cases in which colonic

*

*

B

stents were used as a bridge to definitive resection, 223 were successful (85%).69 Other authors have placed selfexpanding colonic metallic stents in patients with obstructing colon cancers and nonresectable metastases, allowing earlier admini­stration of chemotherapy70 and earlier discharge from the hospital.71 Complications associated with colonic stent placement include perforation, stent migration, and recurrent obstruction. If colonic stenting is unsuccessful, patients who are medically fit and have relatively small tumors or inflammatory

Chapter 119  Intestinal Obstruction

Figure 119-16.  Representative film from a contrast enema of a patient with high-grade sigmoid obstruction from a stricture (arrows identify the proximal and distal extent of the stricture). Although this patient has multiple diverticula within the sigmoid colon, differentiation of this benign diverticular stricture from a malignant stricture is not possible on the basis of this study alone.

Figure 119-18.  Contrast enema in a patient with a sigmoid volvulus demonstrating a characteristic bird’s beak obstruction at the junction between the sigmoid colon and the rectum.

*

A

B

Figure 119-17.  Carcinoma of the ascending colon. Contrast enema (A) demonstrates the typical apple-core lesion (arrow), and coronal computed tomography (CT) with antegrade gastrointestinal contrast (B) further shows the irregular luminal narrowing and mucosal disruption (arrow). Ileocolic nodal tumor spread (*) is noted on this staging CT.

2119

2120

Section X  Small and Large Intestine masses of the left colon or sigmoid should undergo the appropriate segmental colectomy with either an end colostomy or primary anastomosis. The classic operative approach to such patients is a left colectomy with an end colostomy and Hartmann’s pouch, although numerous investigators have reported on the safety of a primary anastomosis after emergent left colectomy for obstructing colorectal cancer and diverticular strictures72-74 or sigmoid volvulus.75 These studies have shown rates of anastomotic leakage (1% to 8%), wound infection (5% to 16%), and mortality (1% to 6%) comparable to earlier data with resection and end colostomy.64 Lim and colleagues compared manual decompression and primary anastomosis with intraoperative colonic irrigation and found that intraoperative irrigation offered no improvement in outcome compared with manual decompression alone.73 Colonic obstructions located proximal to the splenic flexure are most often adenocarcinomas and are treated with partial colectomy and primary ileocolic anastomosis in all but the most unfit patients. The presence of nonviable colon necessitates resection in even the most severely compromised patient. In such an instance, resection with end ileostomy and distal mucous fistula obviates the risk of performing an anastomosis in a grossly contaminated field or in a critically ill patient in whom the risk of anastomotic dehiscence may be particularly high. Colonoscopically placed self-expanding metal stents also have been used with good results in the management of patients with obstructing proximal colon cancers, although most are treated with primary resection and anastomosis.76

Volvulus

Decompression of a sigmoid volvulus may be accomplished with a flexible or rigid sigmoidoscope and the placement of a rectal tube. In a classic study, Ballantyne compiled 19 American series totaling 595 patients and found that sigmoidoscopy, either alone or combined with a rectal tube, successfully reduced the volvulus in 70% to 80% of attempts.77 Placement of a rectal tube for 48 hours can minimize the possibility of early recurrence. Endoscopic reduction of sigmoid volvulus alone is associated with a recurrence rate of 25% to 50%,65,77 and therefore, sigmoid resection and coloproctostomy or, in medically compromised patients, end colostomy, should follow endoscopic decompression and mechanical preparation of the bowel. Recurrence rates with this approach are 3% to 6%.65,77 Patients requiring emergent laparotomy for strangulated sigmoid volvulus should have sigmoid resection with end colostomy and a Hartmann’s pouch or primary anastomosis as discussed earlier.75 Right colectomy with primary ileotransverse colostomy is the procedure of choice for cecal volvulus. Regardless of the approach, emergent operations to relieve colonic obstruction are associated with significantly greater morbidity and mortality risks than those performed elec-

tively. Grossmann and colleagues found that the mortality rate was 24% for patients undergoing emergency operation for sigmoid volvulus compared with 6% for those undergoing elective resection.65 Similarly, Sjo and coworkers reported that the early postoperative mortality rate was 19% after emergency operation for colorectal cancer compared with 3.5% for patients undergoing elective resection.78 Emergency operation, American Society of Anesthesiologists class 3 or 4, necrotic bowel, and advanced tumor stage have been consistently identified as independent predictors of early postoperative mortality.65,78-80 The long-term outlook for patients who present with obstructing colon cancers is significantly worse than that for patients without obstruction. Wang and coworkers found that despite similar tumor stages, patients who presented with obstructing right-sided colon cancers had significantly higher rates of tumor recurrence than did patients who presented without obstruction (49% vs. 22%, respectively). Similarly, the five-year survival was worse for patients presenting with obstruction than for those patients without obstruction (36% vs. 77%, respectively).81 Even in the setting of a potentially curative resection, the cancerspecific survival at five years was 75% for patients without obstruction compared with 52% for those patients presenting with obstruction.82

KEY REFERENCES

Abbas S, Bissett IP, Parry BR. Oral water soluble contrast for the management of adhesive small bowel obstruction. Cochrane Database Syst Rev 2007; (3):CD004651. (Ref 39.) Fevang B-T, Fevang J, Lie SA, et al. Long-term prognosis after operation for adhesive small bowel obstruction. Ann Surg 2004; 240:193-201. (Ref 47.) Fevang BT, Fevang J, Stangeland L, et al. Complications and death after surgical treatment of small bowel obstruction. A 35-year institutional experience. Ann Surg 2000; 231:529-37. (Ref 1.) Foster NM, McGory ML, Zingmond DS, Ko CY. Small bowel obstruction: a population-based appraisal. J Am Coll Surg 2006; 203:170-6. (Ref 5.) Grossman EM, Longo WE, Stratton MD, et al. Sigmoid volvulus in Department of Veterans Affairs Medical Centers. Dis Colon Rectum 2000; 43:414-18. (Ref 65.) Kaiser AD, Applegate KE, Ladd AP. Current success in the treatment of intussusception in children. Surgery 2007; 142:469-77. (Ref 54.) Khot UP, Lang AW, Murali K, Parker MC. Systematic review of the efficacy and safety of colorectal stents. Br J Surg 2002; 89:1096-102. (Ref 69.) Levard H, Boudet M-J, Msika S, et al. Laparoscopic treatment of acute small bowel obstruction: a multicentre retrospective study. ANZ J Surg 2001; 71:641-46. (Ref 40.) Margenthaler JA, Longo WE, Virgo KS, et al. Risk factors for adverse outcomes following surgery for small bowel obstruction. Ann Surg 2006; 243:456-64. (Ref 45.) McArdle CS, McMillan DC, Hole DJ. The impact of blood loss, obstruction, and perforation on survival in patients undergoing curative resection for colon cancer. Br J Surg 2006; 93:483-8. (Ref 82.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

120 Ileus and Pseudo-obstruction Amy E. Foxx-Orenstein

CHAPTER OUTLINE Epidemiology  2123 Neural Control of Small Bowel and Colonic Motility  2123 Ileus  2124 Pathophysiology  2124 Clinical Features  2126 Treatment  2126 Acute Colonic Pseudo-obstruction  2128 Pathophysiology  2128 Clinical Features  2129 Prognosis  2129 Prevention  2129 Treatment  2130

Ileus and pseudo-obstruction both refer to intestinal dys­ motility syndromes that have symptoms, signs, and the radiologic appearance of bowel obstruction in the absence of a mechanical cause. There has been nomenclatural con­ fusion because of the use of these two terms as synonyms; they are not. The term ileus is used when the contents of the small intestine are acutely unable to transit through because of impermanent neural or muscular inadequacy; ileus often is accompanied by sudden and extreme pain, abdominal distention, and vomiting, but ultimately it resolves completely. Pseudo-obstruction describes a chronic neuromuscular disorder that can involve the small intestine alone, the colon alone, or other regions of the gastro­ intestinal tract in combination, such as intestine and stomach or intestine, stomach, and colon. The onset of pseudo-obstruction can follow an acute event or develop insidiously. To add to the confusion, acute ileus events can be superimposed on established chronic pseudoobstruction, destabilizing an often fragile nutritional balance. Acute colonic pseudo-obstruction (the term seems inter­ nally contradicting) occurs most often in aged patients with severe underlying disorders that may be responsible for the acuteness of the situation; it has a particularly high rate of complicating ischemia with perforation. With aggressive treatment, the massive colon dilatation can resolve and normal function can return. Presentation dictates manage­ ment: Critically impaired absorption of fluid and nutrients requires prompt investigation and intensive management incorporating medical, radiologic, and surgical opinions; intervention for subacute and chronic indications can range from just observation to frequent monitoring of nutritional and medical needs or even elective surgery. The functions of the small intestine and colon are regu­ lated through the integrated activities of intestinal smooth

Chronic Intestinal Pseudo-obstruction  2130 Pathophysiology  2131 Primary Causes  2133 Secondary Causes  2135 Clinical Features  2140 Complications  2140 Natural History  2141 Diagnosis  2141 Treatment  2142 Megacolon and Megarectum  2143

muscle, the interstitial cells of Cajal (ICC), intrinsic and extrinsic nerves, neurohumoral peptides, and gastrointesti­ nal hormones (see Chapters 97 and 98). Neurons located in the intestine wall constitute the intrinsic neural network known as the enteric nervous system (ENS) (Fig. 120-1). The extrinsic network consists of visceral sensory afferents in the vagus, thoracic, and pelvic splanchnic nerves (Fig. 1202) and visceral motor efferents of the autonomic nervous system. These extrinsic neurons synapse with the ENS and connect it to the central nervous system. The small intestine and colon transport contents over long distances by the action of propulsive and nonpropul­ sive motor activity, and they display divergent fasting and fed patterns of absorption, motility, and transit. Between meals, the migrating motor complex (MMC) propels food distally in a characteristic sequence that cycles every one to two hours. There are four phases to this intercibal fasting pattern: 1) oscillating smooth muscle without con­ tractions; 2) intermittent smooth muscle contractions; 3) continuous sweeping contractions increasing to a maximum rate (11/min in the duodenum); and 4) quiescent phase with cessation of all contractions. Feeding interrupts the MMC and is immediately followed by the fed pattern, which consists of continuous, low, varying-amplitude, ungrouped phasic contractions, the activity of which depends on the quantity and composition of the ingested food. The small intestinal motor activity facilitates gastric emptying; mixes chyme with digestive enzymes, bile, and intestinal secre­ tions; propels material distally; helps absorb nutrients and resorb fluids; and delivers residue to the colon by the action of intermittent, giant phase 3 contractions. Discontinuous emptying from the ileum ensures time for salvage of remain­ ing nutrients in the small intestine. The small intestine uses feedback inhibitory reflexes, including the duodenal and ileal brake, that delay functions such as gastric emptying

2121

2122

Section X  Small and Large Intestine Proximal contraction (Ach/SP)

Opioid/SS/GABA interneurons

Distal relaxation (VIP/NO)

5-HT4

Myenteric plexus

IPAN Submucosal plexus 5-HT3

5-HT1P

5-HT (serotonin)

Enterochromaffin cells in GI tract release 5-HT

Extrinsic innervation Smooth muscle

Intrinsic innervation

Figure 120-1.  Intrinsic pathways that govern intestinal motility. The enteric nervous system (ENS) includes intrinsic motor neurons (efferent), extrinsic motor neurons (efferent), interneurons, and sensory neurons, which together make the ENS capable of carrying reflexes and acting as a center of integration in the absence of input from the central nervous system. Ach, acetylcholine; CGRP, calcitonin gene-related peptide; GABA, γ-aminobutyric acid; GI, gastrointestinal; 5-HT, serotonin; 5-HT1P, serotonin 1p isotype; 5-HT3, serotonin 3 isotype; 5-HT4, serotonin 4 isotype; IPAN, intrinsic primary afferent neuron; NO, nitric oxide; PACAP, pituitary adenylate cyclase-activating peptide; SP, substance P; SS, somatostatin; VIP, vasoactive intestinal peptide. (See text and Chapters 97 and 98.)

and intestinal secretion while influencing the digestive process and ingestive behavior. These negative-feedback mechanisms include hormones, such as glucagon-like pep­ tides, peptide YY, and oxyntomodulin; neural mediators; and food constituents such as fat. Colonic motor activity exhibits two primary components: propulsive and segmental. Propulsive activity includes high-amplitude propagated contractions (HAPC) responsi­ ble for mass movements that commonly precede bowel movements and periodic low-amplitude contractile activity that move along fluid and flatus. Segmental activity is the main motor response to eating, and it results in more gradual displacement of intraluminal contents, thereby allowing

optimal absorption of fluid, nutrients, short-chain fatty acids, and bacterial metabolites. The colon facilitates emp­ tying of the small intestine into a microbial-rich environ­ ment; serves as a reservoir; absorbs water and limited nutrients; propels intestinal contents distally via coordi­ nated sensorimotor activity; and stores and expels intestinal residue. This chapter focuses on the acute and chronic forms of intestinal pseudo-obstruction including ileus, which pri­ marily affect the small intestine; acute colonic pseudoobstruction, which is associated with large bowel dilatation; chronic intestinal pseudo-obstruction, which typically affects the small intestine yet is commonly associated with

Chapter 120  Ileus and Pseudo-obstruction Vagus nerve Sympathetic

Parasympathetic

Celiac ganglion

Spinal nerve

Thoracic splanchnic nerve Spinal dorsal root

Sympathetic chain Pelvic splanchnic nerve

Figure 120-2.  Extrinsic pathways governing intestinal motility. Enteric control is modulated by the parasympathetic and sympathetic nerves of the autonomic nervous system, which respectively stimulate and inhibit nonsphincteric muscle. (See text and Chapters 97 and 98.)

a generalized motility disorder involving other regions of the intestinal tract, including the colon; and isolated mega­ colon and megarectum.

EPIDEMIOLOGY The prevalence of intestinal dysmotility varies according to the underlying pathology and pathophysiology. Postopera­ tive ileus is an unavoidable adverse response to abdominal or retroperitoneal surgery that accounts for delayed refeed­ ing, prolonged hospitalization, and increased costs. It is the main determinant of length of hospital stay after uncompli­ cated abdominal surgery.1 The economic impact of postop­ erative ileus in the United States is estimated to be more than $7.5 billion annually, not including lost work expense. Acute colonic pseudo-obstruction is estimated to occur in 0.1% of surgical patients, 0.05% of patients admitted for trauma,2,3 and 0.3% of critically ill patients with burn injury.4 More than 95% of patients with acute colonic pseudo-obstruction have factors that predispose them to develop the condition,3 including nonsurgical trauma, pelvic or hip surgery, cardiovascular disease, or infection.5 Spontaneous perforation rates in this condition range from 3% to 15%, with an associated 50% mortality rate. Most cases of chronic pseudo-obstruction result from primary acquired defects in nerves, ICC, and smooth muscle cells, or they are secondary to metabolic disorders, inflamma­

tion, infiltrating disease, autoimmune conditions, or cancer, namely paraneoplastic chronic pseudo-obstruction.6,7 Of the developmental neuropathies, the prototypical congenital dysmotility disorder, Hirschsprung’s disease, occurs in one in 5000 live births (see Chapter 96). The literature documents several heritable degenerative smooth muscle disorders— familial visceral myopathies (FVMs) and childhood visceral myopathies (CVMs)—and inherited degenerative disorders of the myenteric plexus (familial visceral neuropathies [FVNs]) that are rarely encountered, even in tertiary centers. A small number of families, mostly whites and less often African Americans8 and Latin Americans,9 has been docu­ mented with these conditions.

NEURAL CONTROL OF SMALL BOWEL AND COLONIC MOTILITY The enteric nervous system (ENS) consists of extensive ganglionated plexuses located in the muscular wall of the gastrointestinal tract10; functionally most important are the myenteric and submucosal plexuses (see Fig. 120-1). In association with the muscle layers, the networks of ICC are recognized as pacemakers that activate neuromuscular func­ tion. The ENS consists of approximately 100 million neurons in higher mammals, roughly equal to the number of neurons in the spinal cord. Histologic and physiologic studies11-13 of the intestinal tract have characterized the properties of the

2123

2124

Section X  Small and Large Intestine neurons and transmitters mediating its functions, including the peristaltic reflex, and the neuroimmune interactions between neurons and inflammatory cells. The ENS develops in utero by migration of neural crest cells to the developing alimentary canal. Migration, coloni­ zation, differentiation, and maintenance or survival of neural crest cells are regulated by specific signaling mole­ cules that include transcription factors (e.g., Mash1), neurotrophic factors (e.g., the glial-derived neurotrophic factor [GDNF] and its receptor subunits), and the neuregulin signaling system. Precursor cells arise from three axial levels of the neural crest: vagal, rostral-truncal, and lumbosacral14 levels. The enteric neurons mainly arise from the vagal neural crest of the developing hindbrain and colonize the bowel by migra­ tion in a rostrocaudal direction. Vagal crest cells are not restricted to a particular intestinal region. Some enteric neurons arrive in the hindgut from the lumbosacral level via a caudorostral wave of colonization. Rarely, these migrating cells do not reach the entire bowel; usually this affects the terminal portion of the colon, as in the classical forms of Hirschsprung’s disease.15 Neural crest cells that migrate and colonize the bowel become neuroblasts or neuronal support cells called glio­ blasts. Differentiation into neurons and glial cells takes place after the neural crest cells reach their final destination in the intestine. Movement through the intestinal mesen­ chyme, survival in the bowel, and differentiation into mature cells are influenced by contacts of precursor cells with the microenvironment, other cells in the mesenchyme, neural crest-derived cells, and the extracellular matrix. The latter provides directional and differentiation signals, such as GDNF, ensure survival of committed neuroblasts. Enteric serotoninergic neurons appear so early that they coexist in primordial enteric ganglia with still-dividing neural precursors. Serotonin (5-hydroxytryptophan [5-HT]) strongly promotes development of neurons at specific times and affects the development of late-arising enteric neurons. Neural control of the intestines is covered in Chapters 97 and 98.

ILEUS Ileus refers to the inhibition of propulsive intestinal motor activity in the absence of a mechanical obstruction. Abdom­ inal or retroperitoneal surgery is the most common cause of ileus, which develops in essentially all such cases16 and is the focus of this section. Other causes of ileus are listed in Table 120-1 and include general anesthesia, medications, inflammation, infection, ischemia, and various metabolic and neurologic processes. The autonomic nervous system plays a key role in regulat­ ing gastrointestinal motility.17 Parasympathetic nerve activ­ ity stimulates intestinal motility by inducing the release of acetylcholine from excitatory neurons in the myenteric plexus. Sympathetic nerve activity inhibits activity by blocking the release of acetylcholine from the same excit­ atory fibers while also providing the efferent limb of mul­ tiple reflex pathways. Sympathetic neural input appears to be the main physiologic mechanism involved with impaired intestinal motility in the postoperative period. Afferent neural input caused by irritation (incision) or inflammation of the peritoneum results in an increase in inhibitory sym­ pathetic efferent neural activity via the splanchnic nerves and an overall decrease in bowel activity.18 Sympathetic

Table 120-1  Factors That Contribute to Ileus Infection Abscess Infected ascites Pneumonia Sepsis Inflammation Local tissue trauma Pancreatitis Peritonitis Retroperitoneal hemorrhage Metabolic Hypercalcemia Hyperphosphatemia Hypocalcemia Hypokalemia Hypomagnesemia Hyponatremia Neurohumoral Endogenous opiates Nitric oxide Vasoactive intestinal peptide Pharmacologic Anesthetic agents Anticholinergic agents Opioids Tricyclic antidepressants Surgery Gynecologic Laparoscopy Laparotomy Nonabdominal Miscellaneous Anxiety Excessive perioperative intravenous hydration Immobility Myocardial infarction Pain

blockade does not entirely prevent the delayed intestinal motility induced by abdominal surgery, and so alternate mechanisms such as nonadrenergic, noncholinergic (NANC) nerves are believed to contribute to peripheral autonomic neuroeffector transmission and to influence postoperative ileus.19 Uncomplicated postoperative ileus resolves in a predict­ able fashion. The small intestine resumes activity within 24 hours, the stomach in 24 to 48 hours, and the colon gener­ ally within 72 hours.20,21 Ileus is an important cause of prolonged hospitalization and escalating costs associated with surgery.1 Determining when ileus has resolved depends on the endpoint chosen, and all endpoints have relative weaknesses. Return of bowel sounds requires frequent examination, and their presence does not ensure propulsive activity.21 Passage of flatus requires verbal feedback and a willingness of the patient to report such activity. Bowel movements may be the most reliable end point, but they represent distal bowel activity rather than global gastroin­ testinal function. Resolution of postoperative ileus is com­ plete when oral intake is tolerated without pain, bloating, or emesis.

PATHOPHYSIOLOGY

The pathophysiology of postoperative ileus involves several mechanisms including disruption in autonomic neural pathways, release of neurohumoral stress mediators, intes­ tinal inflammatory response, perioperative fluid excess, and various pharmacologic elements.19,22,23

Chapter 120  Ileus and Pseudo-obstruction Disrupted Neural Pathways

Parasympathetic (vagal) neural input stimulates intestinal motor activity, whereas sympathetic (splanchnic) input is inhibitory. Sympathetic activity is the predominant in­ hibitory influence on intestinal motility and provides the efferent limb of multiple reflex pathways. Sympathetic stimulation is a key factor in the development of postopera­ tive ileus. Local stimuli trigger an inhibitory reflex that results in intestinal dilatation, for which the somatic and splanchnic nerves provide both the afferent and efferent pathways.24 Local stimuli include incising the abdominal wall and handling of the intestine; blocking sympathetic stimuli during surgery does not fully prevent such inhibi­ tion of intestinal motility. NANC neurotransmitters also play a role in postopera­ tive ileus. Nitric oxide (NO) may be the predominant NANC neurotransmitter that mediates the inhibitory influ­ ence of ENS neurons on smooth muscle. Other NANC neurotransmitters, including vasoactive intestinal peptide (VIP) and neuropeptide Y act through inhibitory pathways similar to the NO pathway to relax intestinal smooth muscle; substance P and adenosine triphosphate (ATP) act through excitatory motor pathways to influence postopera­ tive ileus.25,26 NO inhibitors, and VIP and substance P antagonists decrease postoperative ileus in animal studies, but the potential usefulness of these agents in humans is unclear.18,25

Release of Neurohumoral Stress Mediators

Surgical stress and circulating inflammatory mediators activate pathways that trigger release of hypothalamic corticotrophin-releasing factor (CRF).27 CRF modulates feeding behavior under stressful conditions—explaining why anorexia is a common reaction to stress—and is con­ sidered essential in orchestrating the stress response.28 Experimental studies have shown that administration of CRF or CRF-related peptides delays gastric emptying and inhibits gastric motility29,30 similar to postoperative gastric ileus, whereas administration of the CRF1 antagonist, CP-154,526, blocks surgery-induced delay in gastric emptying.19 Abdominal surgery also stimulates capsaicin-sensitive afferent fibers, activating the inhibitory efferent pathways that disrupt coordinated intestinal motility.31 These fibers have been shown to play a crucial role in acute gastroprotec­ tion. Release of neurotransmitters such as calcitonin generelated peptide (CGRP) and the related increase in mucosal blood flow have been identified as key factors in the protec­ tive effect of the stimulation of these fibers by capsaicin.32 Capsaicin-sensitive afferent nerves can influence ulcer healing by mediating the hyperemic response through the release of CGRP and facilitating acid disposal in the mucosa.32 Studies are ongoing to determine mechanisms of CRF activity and other stress mediators in the inhibition of gastrointestinal motility.

Intestinal Inflammation

Handling the intestine during surgery results in a marked inflammatory reaction,33-35 leading to induction of inducible nitric oxide synthase (iNOS) and cyclooxygenase (COX)-2 and the release of mast cells, monocytes, and neutrophils within the muscularis propria, but not the mucosa. Bowel handling also leads to secretion of numerous proinflamma­ tory cytokines including interleukin (IL)-1β, IL-6, tumor necrosis factor (TNF)-α, and monocyte chemotactic protein (MCP)-1 and up-regulates intercellular adhesion molecule (ICAM)-1.36 Thus, extensive manipulation of the intestine or exposure of the peritoneum to irritants (e.g., pus or blood)

increases the probability of developing postoperative ileus. The effect on intestinal transit appears to be proportional to the magnitude of the inflammatory response.37 In addition to autonomic signaling pathways, extrinsic noradrenergic nerves contribute to the regulation of various digestive functions, including intestinal propulsion, mucosal secretions, and bowel sensation, through activation of α2-adrenoceptors. In the presence of intestinal inflam­ mation, enteric α2-adrenoceptors are up-regulated, giving enhanced inhibitory control of cholinergic and noradrener­ gic transmission.38,39 Located on neurons and inflammatory cells, α2-adrenoreceptors may be important in the pathogen­ esis of postoperative ileus, because their location provides a link between the inflammation induced by intestinal handling and activation of neuronal pathways.40

Perioperative Fluid Excesses

Intravenous fluid overload during or after surgery delays recovery of gastrointestinal function41 and is associated with poor survival and complications.42 In patients undergoing colon resection, use of perioperative and postoperative restricted fluid regimens has resulted in reduced hospital stays, faster return of gastrointestinal function, and fewer postoperative complications including reduced nausea and vomiting compared with higher-volume strategies.43

Pharmacologic Mechanisms

Anesthesia All anesthetic agents have an inhibitory effect on intestinal motility, although the technique of administration can sig­ nificantly influence the duration of postoperative ileus.44 Mid-thoracic (T6-T9) epidural anesthesia often is used as an adjunct to general anesthesia in abdominal operations. The technique allows local placement of the anesthetic for sympathectomy, which substantially reduces the severity and incidence of postoperative ileus44 and theoretically blocks afferent and efferent inhibitory reflexes associated with abdominal surgery.19,45 Local epidural anesthesia appears to increase splanchnic blood flow to the intestine, disrupt afferent inhibitory effects, and further reduce sym­ pathetic neural input.18,46,47 Use of systemic opioid therapy or epidural opioids, regardless of whether administered by a low-thoracic or a low-lumbar route, does not shorten the duration of ileus compared with nonopioid epidural anesthetics.48,49 Opiates Endogenous and exogenous opiates contribute significantly to the development of postoperative ileus.50,51 Of the three main classes of opiate receptors (µ, κ, δ) located in the central nervous system (CNS) and gastrointestinal tract,52 the CNS µ receptors mostly modulate analgesia within the brain and the spinal cord. Endogenous opioids released from neurons within the submucosal and myenteric plexus of the gastrointestinal tract modulate sphincteric and peri­ staltic activity in a coordinated fashion.53 Activation of intestinal µ receptors suppresses release of acetylcholine from cholinergic neurons, resulting in delayed intestinal motility.54 Exogenous opioids increase antral and proximal duode­ nal tone with an overall inhibitory effect on motility. The effect of morphine on the small intestine is biphasic, initially stimulating MMC activity followed by atony, which impedes propulsion and delays intestinal transit.55 In the colon, morphine increases the tone and amplitude of nonpropagating contractions, thus reducing propulsive activity and slowing transit. The overall effect of opiates is to decrease intestinal motility.

2125

2126

Section X  Small and Large Intestine CLINICAL FEATURES

Typical clinical features of postoperative ileus include abdominal distention, poorly localized abdominal pain, obstipation, and absence of bowel sounds. Tympany may be present on percussion. Symptoms can range from none to anorexia, nausea, and vomiting. Ileus in the early post­ operative period generally does not require diagnostic evaluation. Factors that contribute to the development of ileus are listed in Table 120-1. Plain abdominal films can show airfilled stomach or distended loops of intestine or colon. Abdominal computed tomography (CT) can confirm the diagnosis while delineating extraluminal findings including abscess, retroperitoneal hematoma, pancreatitis, ascites, and inflammation, which might contribute to nonmechani­ cal obstruction.

TREATMENT

Historically, treatment for postoperative ileus has included bowel rest, nasogastric tube decompression, and early ambu­ lation. These interventions were thought to shorten recovery time by lowering the incidence of complications and improving outcomes; critical review of existing evidence, however, does not support such conclusions. Meanwhile, perioperative care has improved with newer anesthetic tech­ niques and analgesic agents, minimally invasive surgery, goal-directed fluid management, and medications that reduce the stress of surgery. Fast-track, enhanced postopera­ tive recovery protocols combine effective evidence-based treatments with the aim of lowering organ dysfunction, enhancing recovery, and shortening hospital stays.

Standard

Bowel rest neither shortens the time to first bowel move­ ment nor decreases the time to oral intake. Large meals can induce nausea and bloating, but small meals can stimulate gastrointestinal motility and reduce the duration of postop­ erative ileus.56,57 Nasogastric tube decompression, once con­ sidered standard of care, is not recommended in routine abdominal operations because it does not hasten recovery from ileus.58,59 In fact, routine use of nasogastric decom­ pression has been associated with a higher postoperative complication rate of atelectasis and aspiration pneumonia and prolonged hospitalization. Early mobilization after surgery is recommended to reduce complications of atelec­ tasis, pneumonia, and deep venous thrombosis; however, there is little evidence that ambulation hastens resolution of ileus.60

efferent mechanisms that trigger inhibitory reflexes and the CRF-mediated stress response that can restore metabolic activities and reduce postoperative morbidity.65 A signifi­ cant reduction in the duration of ileus occurs in patients who receive epidural bupivacaine compared with those receiving epidural opioid.66-68 Epidural bupivacaine alone was superior to epidural bupivacaine combined with opioid in reducing the duration of ileus without compromising pain relief. Delivery of local anesthetic via a mid-thoracic (T6-T9) epidural catheter for abdominal surgery signifi­ cantly reduces duration of ileus and hospitalization,18 whereas low thoracic and lumbar epidural deliveries do not show similar benefit.46,47 Local anesthesia may be more important in reducing postoperative ileus when used as an adjunct to open abdominal surgery rather than to laparo­ scopic surgery.69 The optimal duration of epidural anes­ thetic appears to be two to three days beyond the operation.70 Epidural anesthesia might not be necessary in laparoscopic colorectal surgery and, in some cases, it may be replaced by opioid-sparing multimodal analgesia, including oral acet­ aminophen, NSAIDs, systemic local anesthetics, or continu­ ous infusion of the wound with local anesthetic.71,72 Perioperative Fluid Management.  Several large trials have shown that excessive hydration in the perioperative period increases morbidity. Fluid excess, which can cause bowel edema42 and pulmonary compromise, is linked to prolonged postoperative ileus and extended hospital stay.73 Fluid management by perioperative optimization of hemody­ namic function, known as goal-directed therapy, maximizes cardiac stroke volume by using small fluid challenges43,74 and has been shown to improve patient outcome. Avoiding fluid overload by using goal-directed therapy can reduce postoperative complications and facilitate rapid functional gastrointestinal recovery.43 Postoperative Period Postoperative nausea and vomiting (PONV) occurs in 20% to 30% of low-risk surgical patients and in up to 80% of high-risk surgical patients.75 General anesthesia increases the risk of PONV nine-fold greater than the use of regional anesthesia.76 Reducing risk factors (Table 120-2) when pos­ sible and using prophylactic agents in patients at high risk for PONV can shorten the duration of ileus; low-risk patients are less likely to receive benefit. Dexamethasone, droperi­ dol, and the 5-HT3 antagonist ondansetron, alone or in com­ bination, can provide effective antiemetic prophylaxis against PONV77 in high-risk patients. Aprepitant is a sub­

Preventive

Operative Period Nature of Surgery.  When compared with open surgical pro­ cedures, laparoscopy decreases the duration of postopera­ tive ileus.61 Theoretically, reducing the mechanisms known to cause postoperative ileus (see earlier) by making small incisions and minimizing tissue handling should improve surgical outcome. Circulating levels of cytokines (e.g., IL-1β and IL-6) and C-reactive protein are significantly lower after laparoscopy than after open abdominal procedures and may reflect reduced tissue injury.62 Return to a normal interdi­ gestive pattern occurs 40% faster after laparoscopic colec­ tomy compared with conventional colectomy, and time to first bowel movement is reduced 27%.63 Also, tolerance of oral feedings and improved transit occur earlier after lapa­ roscopic resection.64 Anesthesia.  Epidural anesthesia, but not general anesthe­ sia, interrupts sympathetic outflow, blocking afferent and

Table 120-2  Risk Factors for Postoperative Nausea and Vomiting Anesthesia Moderate-acting to long-acting agents Nitrous oxide Use of perioperative opioids Volatile anesthetics Patient Female gender History of motion sickness or postoperative nausea and vomiting Nonsmoker Surgery Prolonged duration Type: laparotomy, laparoscopy, gynecologic Modified from Gan TJ, Meyer TA, Apfel CC, et al. Society for Ambulatory Anesthesia guidelines for the management of postoperative nausea and vomiting. Anesth Analg 2007; 105(6):1615-28.

Chapter 120  Ileus and Pseudo-obstruction stance P antagonist that acts on the neurokinin 1 receptor and is FDA approved for prevention of PONV. Chewing Gum.  Several studies show gum chewing in the postoperative period significantly reduces the time to first flatus and the time to first passage of feces compared with standard treatment alone,78,79 although the length of hospital stay is only marginally reduced. Chewing gum does not increase complications or readmissions and is a low-risk, inexpensive, first-line approach to resolving ileus after colorectal surgery.80 Early Oral Intake.  Several randomized, controlled trials support early postoperative feeding (within 24 hours) in patients undergoing gastrointestinal surgery.81 Although early postoperative feeding can increase vomiting, mortality is reduced with a trend toward reduction of risk of post­ surgical complications and shorter hospitalization. Early oral nutrition attenuates catabolism and reduces intestinal permeability that can lessen infectious complications.82,83 Despite ongoing controversy, studies suggest early post­ operative feeding may be of benefit.

Drug Therapy

Opioid-Sparing Analgesia Nonsteroidal anti-inflammatory drugs (NSAIDs) and COX-2 selective agents are prostaglandin inhibitors used exten­ sively in postoperative pain management to avoid opioidrelated side effects, including constipation, abdominal distention, PONV, and gastroesophageal reflux as well as urinary retention, respiratory depression, and sleep distur­ bance, among others. Ketorolac tromethamine is an NSAID that can be admin­ istered parenterally, thereby circumventing the time lag accompanying postoperative oral intake. Ketorolac is equiv­ alent to morphine in analgesic effect after major abdominal surgery without the functional gastrointestinal side effects of morphine.84 Ketorolac has anti-inflammatory and anti­ pyretic activity and reduces postoperative ileus and mor­ phine requirements.85 Continuous epidural infusion of lidocaine86 after colonic resection can reduce morphine use, improve pain relief, and hasten recovery. Ketamine, gabapentin, pregabalin, and neural blockade are recognized opioid-sparing agents or techniques with the potential to accelerate postoperative recovery; additional study is required because results of their effects are inconsistent.87 In addition, use of post­ operative patient-controlled analgesia (PCA) that allows on-demand intravenous delivery may be as effective as continued epidural administration and can reduce medica­ tion dose and duration.88 Opioid Antagonists Stimulation of µ-opioid receptors in the brain and intestine by morphine and other opioids imparts a potent central analgesic effect plus dose-limiting debilitating constipation and other unwelcome side effects.51 Therefore, effective blockade of peripheral opioid receptors in the gastrointesti­ nal tract should manage opioid-related bowel dysfunction. Alvimopan and methylnaltrexone are selective µ-opioid antagonists with activity that is restricted to the peripheral receptors. Alvimopan (12 mg, orally two hours before surgery, then twice daily for seven days) shortened the time to gastroin­ testinal recovery (passage of flatus, tolerance of solid food, first evacuation) by 15 to 18 hours and the hospital stay by one day following bowel resection.89 The effects of alvimo­ pan are achieved without adverse effects on abdominal pain

scores or patient-administered analgesia.89-92 The alvimopan studies used general (not epidural) anesthesia and nasogas­ tric tubes up to 24 hours postoperatively; oral feeding was initiated at 48 hours postoperatively. Avoiding known risk factors and implementing measures to prevent postopera­ tive ileus as it occurs in fast-track postoperative surgery can alter outcome. Methylnaltrexone has been shown to reverse morphineinduced delay in gastric and orocecal transit time in healthy subjects.93 In patients with opioid-induced constipation with advanced illness, methylnaltrexone induced laxation and reduced symptoms of constipation.93 Naloxone, a nonselective opioid antagonist, crosses the blood-brain barrier and acts on central opioid receptors to reverse analgesia and to elicit opioid withdrawal, both unwelcome side effects in postoperative management. Prokinetic Agents There are no available effective prokinetic agents for the treatment of postoperative ileus. Cisapride is a serotonin agonist that facilitates acetylcholine release from the intrin­ sic neural plexus with inconsistent effect on resolution of postoperative ileus. Cisapride was removed from the market because of potentially dangerous cardiac side effects, but it is available under a limited-access program through Janssen Pharmaceutica.94 Neostigmine and lidocaine require addi­ tional studies with clinically relevant outcomes to prove treatment effectiveness.92 Erythromycin has demonstrated a consistent absence of effect in hastening postoperative recovery in several prospective randomized clinical trials. Metoclopramide, a mixed cholinergic agonist and dopamine antagonist, shows only small or insignificant benefit in the treatment of postoperative ileus.92 Laxatives There is some evidence to support the use of laxatives in the postoperative period. Osmotic and stimulant (supposi­ tory) laxatives appear safe and beneficial in producing early bowel evacuation and might shorten the duration of post­ operative ileus. Small trials have examined the effect of bisacodyl suppository alone,95 bisacodyl in combination with milk of magnesia,96 and magnesium oxide combined with disodium phosphate,97 the latter used in a fast-track rehabilitation protocol following abdominal hysterectomy. No significant adverse events have been reported with any agent. High-quality prospective trials are needed. Emerging Drugs Lubiprostone, a chloride-2 channel activator that stimulates intestinal fluid secretion and increases intestinal transit, is under investigation as an agent to treat postoperative ileus. Other drugs emerging from trials or preclinical testing as potential agents for the treatment of postoperative ileus include linaclotide (a guanylate cyclase-C agonist), motilin agonists (KOS-2187, GM-611, GSK-326416), a neurokinin-2 receptor antagonist (MEN-11420), and growth hormone secretagogues (ipamorelin, TZP-101).31

Multimodal Approach to Preventing Postoperative Ileus

The advent of laparoscopic surgery, opioid-sparing analge­ sics, regional anesthesia, and other factors that hasten recovery from postoperative ileus (Table 120-3) has led to improved care of patients undergoing abdominal surgery. First described in the 1990s, enhanced recovery manage­ ment strategies—fast-track surgery methods—combine treatment modalities that individually improve outcome following major surgery.98 An overarching goal of fast-track surgery is to lower rates of organ dysfunction, thereby

2127

2128

Section X  Small and Large Intestine Table 120-3  Factors That Contribute to Enhanced or Delayed Recovery from Postoperative Ileus

Factor

RANDOMIZED CONTROLLED STUDIES (RCTs)*

Enhances Recovery Chewing gum Early oral feeding† Goal-directed fluid therapy and avoidance of fluid overload Intravenous or wound (local) anesthetics Laparoscopic vs. open surgery|| Laxatives† Length and/or type of incision‡ Mobilization‡ Peripheral opioid antagonists PONV antiemetic agents Thoracic epidural local anesthetics Delays Recovery Administration of excess perioperative fluid Nasogastric tubes Opioids Restrictions on oral intake†

x x x x

x x x x x x

*Based on two or more RCTs or meta-analyses. † Preliminary studies are positive, but further studies are required before general recommendations are made. ‡ Data are limited or inconclusive. || Data are difficult to interpret because of the use of nasogastric tubes, use of opioid analgesia, and restrictions for oral intake in “open” groups. PONV, postoperative nausea and vomiting. Used with permission and adapted from Kehlet H. Postoperative ileus—an update on preventive techniques. Nat Clin Pract Gastroenterol Hepatol 2008; 5(10):552-558.

Table 120-4  Benefits of Fast-Track* Surgery Decreased period of convalescence Early oral intake of energy substrates Improved muscle strength and exercise capacity Reduced cardiopulmonary morbidity Reduced costs Reduced duration of ileus Shorter hospital stay *Fast track refers to the use of enhanced recovery management strategies to improve outcomes after surgery.

reducing morbidity, hastening recovery, and shortening hospital stay.69 Some benefits of fast-track surgery are listed in Table 120-4. The concept has proved valid across all surgical specialties, but the most physiologic data are avail­ able for colonic surgery. Fast-track surgery has been shown to enhance recovery from postoperative ileus: More than 90% of patients have a normal oral intake, defecate within 48 hours, and have a hospital stay of two to four days (reduced from five to 10 days) after uncomplicated open colonic surgery.71,99,100

ACUTE COLONIC PSEUDO-OBSTRUCTION Acute colonic pseudo-obstruction, also known as Ogilvie’s syndrome,101 is characterized by acute massive colon dilata­ tion involving primarily the right side of the colon and without a mechanical cause. It is most often diagnosed in

Table 120-5  Conditions Commonly Associated with Acute Colonic Pseudo-obstruction Cardiovascular Heart failure Myocardial infarction Stroke Drugs Antidepressants Antiparkinsonian agents Opiates or narcotics Phenothiazines Inflammation Acute cholecystitis Acute pancreatitis Pelvic abscess Sepsis Metabolic Alcohol abuse Electrolyte imbalance Liver or kidney failure Neoplasia Disseminated Leukemia Retroperitoneal Neurologic Alzheimer’s disease Multiple sclerosis Parkinson’s disease Spinal cord disease Post-surgical Cesarean section Hip surgery Kidney transplantation Pelvic surgery Post-traumatic Femur fracture Pelvic trauma Spinal cord injury Respiratory Mechanical ventilation Pneumonia Used with permission from De Giorgio R, Stanghellini V, Barbara G, et al. Prokinetics in the treatment of acute intestinal pseudo-obstruction. IDrugs 2004; 7(2):160-165.

hospitalized, debilitated medical or surgical patients with a wide array of medical conditions (Table 120-5). Ogilvie’s syndrome is estimated to occur in 0.1% of all surgical patients,3 and patient outcome depends on the severity of the underlying illness, the person’s age, the maximum diameter of the cecum, the delay until colonic decompres­ sion, and the presence of colonic ischemia.

PATHOPHYSIOLOGY

Disrupted autonomic nervous system activity is considered a key factor in the pathogenesis of acute colonic pseudoobstruction. Parasympathetic (excitatory, causing contrac­ tion) innervation of the colon is from the vagus nerve, which supplies the right colon and extends to the splenic flexure; parasympathetic innervation of the distal colon and rectum is from the spinal supply sacral plexus. Sympathetic (inhib­ itory, causing relaxation) innervation of the colon is from the celiac and mesenteric ganglia (see Fig. 120-2). The current proposed mechanisms of acute colonic pseudoobstruction are summarized in Table 120-6. Local stimuli effect an inhibitory reflex that results in dilatation of the colon, for which the splanchnic nerves provide both the afferent and efferent pathways. Local stimuli, including peritoneal inflammation, infection, and

Chapter 120  Ileus and Pseudo-obstruction Table 120-6  Proposed Mechanisms for Acute Colonic Pseudo-obstruction Intestine Fails to Contract Excess sympathetic motor input Decreased parasympathetic motor input Stimulation of peripheral µ-opioid receptors (endogenous or exogenous) Reflex motor inhibition through splanchnic afferents Intestine Fails to Relax Excess parasympathetic motor input Inhibition of nitric oxide release Modified from Delgado-Aros S, Camilleri M. Pseudo-obstruction in the critically ill. In: Scholmerich J, editor. Bailliere’s best practice & research in clinical gastroenterology: gastrointestinal Disorders in the Critically Ill, vol. 17. London: Elsevier Science; 2003. pp. 427-44.

handling of the colon, result in the release of mast cells, leukocytes, and monocytes.33-35 Inflammatory mediators such as ICAM-1, MCP-1, iNOS, and COX-2 are up-regulated during the first 18 hours after abdominal surgery, and the degree of postoperative intestinal impairment correlates with the intestinal inflammatory response. The release of endogenous opioids after surgery has been related to the inflammation and impaired motor activity that characterize the physiologic response to surgery. Adminis­ tration of antidepressant, phenothiazine, anti-parkinsonian, or narcotic medications can induce acute colonic pseudoobstruction.102 Opioids inhibit release of NO from inhibitory motor neurons in vitro and delay transit in vivo.

Figure 120-3.  Acute colonic pseudo-obstruction. Left, Upright plain film of the abdomen revealing diffuse, but predominantly right, colonic distention, with a cecal diameter of 13 cm, that developed after placement of a right hip prosthesis for a fractured femur. Right, Colonic distention resolved shortly after administration of intravenous neostigmine. (From Bharucha AE, Camilleri M. Common large intestinal disorders. In: Hazzard WR, Blass JP, Halter JB, et al, editors. Principles of Geriatric Medicine and Gerontology. New York: McGraw-Hill; 2003. p 652.)

usually a radiologic sign of intestinal perforation, although when free air occurs with pneumatosis intestinalis, it might not be associated with perforation. A water-soluble contrast enema or CT scan can exclude mechanical obstruction if gas and distention are present throughout all colonic segments, including the rectum and sigmoid colon.

CLINICAL FEATURES

Differential Diagnosis

Symptoms and Signs

PROGNOSIS

Acute colonic pseudo-obstruction typically occurs in older (mean age, 60 years) men who are hospitalized or institu­ tionalized (60%) with severe underlying medical or surgical conditions (see Table 120-5). The most characteristic feature of acute colonic pseudoobstruction is abdominal distention, which can develop gradually over three to seven days or more acutely within 24 hours. About 60% of patients experience nausea and vomiting. Abdominal pain (80%) tends to be mild and constant, with occasional slight rebound tenderness and painless distention.103 Low-grade fever may be present. Disturbances of electrolyte levels occur, notably hypokale­ mia, hypocalcemia, hyponatremia, and hypomagnesemia. New abdominal pain or tenderness, increasing fever, and increasing white blood cell count are features of ischemia or per­foration. About 90% of patients have abnormal bowel sounds, which vary from absent to hyperactive. In many cases, the abnormal bowel sounds are high-pitched and suggest mechanical obstruction.2

Radiology

The most distinctive feature of acute colonic pseudoobstruction on a plain abdominal film is dilatation of the colon that preferentially affects the right side of the colon. The maximal diameter of the cecum typically ranges from nine to 25  cm, often with a cutoff sign at the hepatic or splenic flexure (Fig. 120-3). The left colon, including the rectosigmoid, and the small bowel also may be dilated. Airfluid levels can be seen in the small intestine but usually do not occur in the colon. Haustral folds often are visualized despite severe distention. Air throughout all colonic segments helps differentiate this condition from mechanical obstruction. Free air is

The differential diagnosis includes mechanical obstruction, toxic megacolon, and ischemic colitis. Frequent clinical evaluation including white blood count and plain abdomi­ nal radiography are vital to identifying bowel perforation. The mortality rate of patients with acute colonic pseudoobstruction varies from 0% to 32% and is partly determined by their comorbidity.3,104 Older patients, poor clinical condi­ tion, and surgical treatment for acute colonic pseudoobstruction are associated with an increased risk of mortality. There are no randomized clinical trials compar­ ing surgical and medical treatment to clarify whether surgery itself or selection bias influences mortality associ­ ated with surgical treatment. Intestinal ischemia or perfora­ tion, which supervenes in about one in every six or seven cases, is associated with a 40% increase in the risk of death.3 The diameter of the colon is also a risk factor for mortal­ ity. When surgical decompression is utilized in mechani­ cally obstructed patients with a cecal diameter greater than nine centimeters, there is a dramatic reduction in mortal­ ity,105 which is the basis for the nine-centimeter cutoff as a sign of impending perforation in patients with acute colonic pseudo-obstruction. In one study, perforation rates for cecal diameters less than 12 cm, 12 to 14 cm, or larger than 14 cm were 0%, 7%, and 23%, respectively.3 Mortality also was associated with delay in decompression: 15% in those decompressed less than four days after onset of dilata­ tion; 27% when the decompression occurred after four to seven days; and 73% after day seven.

PREVENTION

Minimally invasive surgery, thoracic epidural anesthesia, and nominal use of opioid analgesia have led to improved care of surgical patients. Enhanced operative recovery pro­

2129

2130

Section X  Small and Large Intestine Table 120-7  Some Methods Used in Fast-Track* Rehabilitation Programs Preoperative patient education Epidural anesthesia Nonopioid analgesia Avoidance of perioperative fluid overload or use of goal-directed therapy No preoperative bowel cleansing No routine use of nasogastric tubes Early oral refeeding Well-defined daily care regimens and discharge criteria *Fast track refers to the use of enhanced recovery management strategies to improve outcomes after surgery. Modified from Kehlet H. Fast-track colorectal surgery. Lancet 2008; 371(9615):791-3.

grams, referred to as multimodal or fast-track programs, combine modalities that individually have a lower morbid­ ity compared with standard treatment.72,106 Methods used in some fast-track programs are listed in Table 120-7.69 Studies have shown various fast-track programs can reduce the duration of ileus and the costs and duration of hospital stay while decreasing cardiopulmonary morbidity and postop­ erative convalescence.

TREATMENT

The goal of management is to achieve colonic decompres­ sion. With correction of reversible potential causes or asso­ ciated imbalances (e.g., infection, hypovolemia, hypoxemia, electrolyte levels), discontinuation of medications that can induce ileus (e.g., anticholinergics, opiates), the ileus disap­ pears by day six in 83% to 96% of patients.3 Intravenous saline and glucose solutions suffice for hydration because of the short-lived and reversible nature of the dysfunction in most cases; in patients with prolonged acute colonic pseudo-obstruction, parenteral or enteral nutrition may be necessary.107 Nasogastric aspiration for nausea and vomiting may be beneficial, but often it is ineffective because the functional obstruction affects the colon. Rectal tubes may be useful if the sigmoid colon is dilated. Enemas might “cleanse” the colon, but only Gastrografin enemas have shown efficacy in anecdotal reports. Enema use has been complicated by colonic perforation.108 When the diameter of the cecum is greater than nine centimeters and has not responded to treatment within the first 72 hours after diag­ nosis, decompression should be performed to reduce the risk of ischemia, perforation and death.3 A treatment algo­ rithm is shown in Figure 120-4.

Medical Decompression

Adrenergic blockers and the acetylcholinesterase inhibitor neostigmine have been tested in open-label studies. Only neostigmine has been tested in a randomized, controlled trials and has an effective initial treatment response rate of 60% to 90%.109,110 Abdominal cramping with the passage of gas and decom­ pression of the colon occurs promptly after administration of neostigmine, but such volume decompression has not been shown to reduce the risk of perforation and mortality in these patients. Nonetheless, if there are no contraindica­ tions to its use, neostigmine is a safe choice in patients whose cecal diameter is greater than nine centimeters for 72 hours. Contraindications include bradycardia, hypoten­ sion, active bronchospasm, serum creatinine concentration greater than 3 mg/dL, mechanical obstruction of the gastro­

intestinal or genitourinary tract, and peritonitis. Adverse events described after intravenous administration of 2.0 to 2.5  mg of neostigmine include abdominal cramps (17%), excessive salivation (13%), sweating (4%), nausea or vomit­ ing (4%), and transient bradycardia (6%); a starting dose of 1 mg rather than 2 mg reduces the likelihood of bradycardia. A second dose of neostigmine should be considered if there is partial or no response to the first trial or if ileus recurs after an initial response. 5-HT4 receptor agonists, motilin receptor agonists, and muscarinic receptor agonists have been the subject of anecdotal reports, but none have been formally tested in the setting of acute colonic pseudo-obstruction.108 Metoclo­ pramide has shown very small or insignificant benefit in the treatment of postoperative ileus.92 The peripherally acting µ-opioid receptor antagonist alvimopan significantly short­ ens the duration of postoperative ileus, but neither metoclo­ pramide nor alvimopan has been sufficiently tested as treatment for acute colonic pseudo-obstruction.89-92

Endoscopic Decompression

Randomized, controlled trials of endoscopic decompres­ sion are lacking. Colonoscopic decompression can be achieved technically in 80% of patients with acute colonic pseudo-obstruction, albeit with a high risk of cecal perfora­ tion.110 The complication rate ranges from about 1% to 5%,5 with a 3% perforation rate.104 Colonic decompression has not been shown to improve the outcome of these patients. It has a high recurrence rate and colonoscopy of unprepared bowel can distend the colon further and result in perforation.3 Placement of a decompression tube into either the right or transverse colon has a similar technical success rate.104 Endoscopic decompression should be con­ sidered in patients with a high risk of cecal perforation when conservative and pharmacologic maneuvers have failed.

Percutaneous Cecostomy

In view of the high rate of recurrence of colonic dilatation after endoscopic decompression, alternative techniques to decompress the colon have been proposed to avoid surgery. Percutaneous cecostomy has been reported to be successful in several case reports and one small case series, but it is associated with significant morbidity. Percutaneous cecos­ tomy can be performed endoscopically, radiographically, or surgically.111,112

Surgical Decompression

Surgical decompression, which includes cecostomy, colos­ tomy, or resection, is associated with a 6% morbidity and a 30% mortality overall.3 It is used for patients with moresevere disease who have failed to respond to conservative or other measures. In cases of ischemia or perforation, seg­ mental or subtotal colectomy is indicated.111

CHRONIC INTESTINAL PSEUDO-OBSTRUCTION Chronic intestinal pseudo-obstruction is a syndrome that manifests insidiously with symptoms of intestinal obstruc­ tion in the absence of an anatomic lesion obstructing the flow of intestinal contents; its course is marked by intermit­ tent acute episodes. It is primarily a disorder of the small intestine but can involve more than one region of the gas­ trointestinal tract. Rare familial and some secondary causes of chronic intestinal pseudo-obstruction can have character­ istic extraintestinal manifestations. The clinical presen­

Chapter 120  Ileus and Pseudo-obstruction Cecal diameter >9 cm Mechanical obstruction ruled out

Response

Resume standard care

Treat reversible causes (infection, electrolyte disturbances, medications, volume depletion) Keep patient NPO No response after 72 hr

Neostigmine contraindicated*

Neostigmine not contraindicated

Consider gentle watersoluble enema

Infuse neostigmine 1-2 mg over 3-5 min Monitor cardiac rhythm and vital signs Repeat neostigmine in 4 hr if no response or dilatation recurs Give atropine 0.5-1 mg for bradycardia

No response Endoscopic decompression Gentle saline enema preparation or no preparation Do not overinflate colon If a decompression tube is placed, advance at least to transverse colon Response

No response

Resume standard care

↑ WBC, fever, abdominal pain, signs of peritonitis or free air

Present * Contraindications to neostigmine: Systolic BP <90 mm Hg Bradycardia Bronchospasm Serum creatinine >3 mg/dL Perforated bowel History of intolerance to neostigmine

Surgery

Brief response Repeat colonic decompression. Use decompression tube if not used initially.

Not present Cecostomy Endoscopic Radiologic Surgical

Figure 120-4.  Algorithm for the treatment of acute colonic pseudo-obstruction. NPO, nothing by mouth; WBC, white blood cell count.

tation is variable but is characterized by progressively disabling gastrointestinal symptoms that eventually persist even between acute episodes; difficulty in maintaining ade­ quate oral nutrition and a normal body weight; and a poor quality of life. Having a high index of suspicion for this disorder is important to reduce complications and avoid potentially harmful surgical procedures.

PATHOPHYSIOLOGY

Causes of chronic pseudo-obstruction are classified as either primary or secondary, but in most cases a cause is never determined. When a cause is found, it most often is from an underlying and potentially reversible condition affecting the intrinsic or extrinsic enteric neural pathways (neuropa­ thies), the ICC (mesenchymopathies), or the intestinal smooth muscle cells (myopathies) (Table 120-8).113 Overlap, such as neuropathy with features of myopathy or neuropa­

thy plus mesenchymopathy, can occur. Familial causes are rare and are far outnumbered by sporadic and secondary causes. When a thorough diagnostic evaluation fails to reveal a cause, obtaining a full-thickness tissue specimen by laparoscopy with subsequent specific pathologic testing, including detailed light microscopic evaluation and ultrastructural analysis, may aid in making a diagnosis (Table 120-9).114

Enteric Neuropathies

Inflammatory neuropathies that cause chronic pseudoobstruction may be primary or secondary to a variety of infectious, paraneoplastic, and neurologic disorders. Inflam­ matory neuropathies are characterized by an intense inflam­ matory infiltrate of CD4 and CD8 lymphocytes that are CD3 positive and confined to the myenteric plexus.115,116 The juxtaposition of lymphocytes that are CD3 positive with

2131

2132

Section X  Small and Large Intestine Table 120-8  Classification of Chronic Intestinal Pseudo-obstruction According to Etiology Primary chronic intestinal pseudo-obstruction Familial Familial autonomic dysfunction Familial visceral myopathies Familial visceral neuropathies Sporadic Visceral myopathies Visceral neuropathies Secondary chronic intestinal pseudo-obstruction (acquired) Connective tissue disorders Scleroderma Dermatomyositis Polymyositis Systemic lupus erythematosus Endocrine disorders Diabetes mellitus Hypoparathyroidism Hypothyroidism Pheochromocytoma Idiopathic myenteric ganglionitis Infections Trypanosoma cruzi (Chagas’ disease) Viral (cytomegalovirus, Epstein-Barr virus) Neuromuscular disorders Amyloidosis (primary and secondary) Muscular dystrophy (myotonic, Duchenne’s, and oculopharyngeal muscular dystrophies) Paraneoplastic syndrome Parkinson’s disease Medications Anticholinergic agents Anti-Parkinson medications Opiates Tricyclic antidepressants From Sutton DH, Harrell SP, Wo JM. Diagnosis and management of adult patients with chronic intestinal pseudoobstruction. Nutr Clin Pract 2006; 21(1):16-22.

myenteric neurons is the foundation by which immunologic interactions affect ganglion cell structure and survival.114,117 In the gastrointestinal tract, inflammation and immune activity have been shown to significantly influence the mor­ phology and function of the enteric nerves. Many patients with intestinal dysfunction secondary to paraneoplastic visceral neuropathy exhibit circulating anti­ neuronal autoantibodies, such as anti-Hu antineuronal antibodies; autoantibodies reinforce a probable role of the immune system in neuronal dysfunction.112 Experimentally, these autoantibodies can elicit neuronal hyperexcitability118 and modify the ascending reflex pathway of peristalsis.113 When incubated with a primary culture of myenteric neurons, anti-Hu neuronal antibodies have been shown to trigger activity of caspase-3 and apaf-1 together with apoptosis,119 which implies that anti-Hu antibodies contrib­ ute in a direct way to the lymphocytic infiltrate in ENS dysfunction and degeneration observed in patients with inflammatory neuropathy. Indication of viral infection, such as herpes virus DNA, has been isolated deep in the myen­ teric plexus of some patients with severe neuropathic intestinal dysmotility, lending evidence to viral infectious agents as a cause of inflammatory ganglionitis.120 Eosino­ philic ganglionitis has been recognized in some pediatric patients with pseudo-obstruction,121 although in these cases, eosinophilic ganglionitis did not appear to cause neuronal deterioration or loss.113 Also, mast cell ganglionitis has been described in patients with severe intestinal dysfunction122 and is associated with a reduced nitric oxide synthase expression, suggesting impaired inhibitory innervation of the ENS. Noninflammatory (degenerative) neuropathies may be familial or sporadic and the result of dysfunctional mito­ chondria, altered calcium signaling, and accumulation of free radicals that leads to eventual degeneration and loss of neurons.123 Sporadic cases can be primary or secondary to a wide range of conditions including radiation, diabetes mellitus, amyloidosis, myxedema, and drug toxicity.113 Two

Table 120-9  Markers for and Targets of Intestinal Neuropathy MARKERS

CELL TARGETS AND SITES

DESCRIPTION

PGP9.5, NSE, MAP-2 NFs, tubulins, Hu C/D

Neurons: membrane and cytoplasmic

B-S-100, GFAP C-Kit Substance P, VIP, PACAP, CGRP, neuropeptide Y, galanin, 5-HT, NOS, ChAT, somatostatin, calbindin, NeuN; NK1, NK2, NK3 Bcl-2, TUNEL, caspase-3, caspase-8, APAF1

Glial cells: cytoplasmic ICCs: membrane and cytoplasmic Subclasses of enteric neurons; ICC: membrane and cytoplasmic

Identification of the general structure of the ENS Detection of enteroglial cells Different ICC networks Characterization of neurochemical coding and enteric neuron subclasses; subsets of ICC Assessment of apoptosis and related pathways Assessment of smooth muscle integrity Evaluation of B (CD79α) and T lymphocytes (CD3), T-helper (CD4) cells, T-suppressor (CD8) cells, macrophages (CD68) in enteric ganglionitis; MIP-1α is a chemokine; TNF-α and IFN-γ are inflammatory cytokines

Actin, myosin, desmin, smoothelin CD3, CD4, CD8, CD79α, CD68; MIP-1α, TNF-α, IFN-γ

Apoptosis and related mechanisms: nuclear and cytoplasmic Smooth muscle cells: cytoplasmic Immune cells, chemokines, and cytokines: membrane and cytoplasmic

APAF1, apoptotic peptidase activating factor 1; Bcl-2, B cell lymphoma-2 protein; ChAT, choline acetyltransferase; ENS, enteric nervous system; CGRP, calcitonin gene-related peptide; GFAP, glial fibrillary acidic protein; 5-HT, 5-hydroxytryptamine (serotonin); Hu C/D, Hu C/D molecular antigen; ICC, interstitial cells of Cajal; IFN-γ, interferon γ; MAP-2, microtubule associated protein-2; MIP1-α, macrophage inflammatory protein 1-α; NeuN, neuronal-specific nuclear protein; NF, neurofilament; NK, neurokinin; NOS, nitric oxide synthase; NSE, neuron-specific enolase; PACAP, pituitary adenylate cyclase activating polypeptide; PGP9.5, protein gene product 9.5; TNF-α, tumor necrosis factor α; TUNEL, terminal deoxynucleotidyl transferase–mediated doxyuridine triphosphate nick-end labeling; VIP, vasoactive intestinal polypeptide. From Antonucci A, Fronzoni L, Cogliandro L, et al. Chronic intestinal pseudo-obstruction. World J Gastroenterol 2008; 14(19):2953-61.

Chapter 120  Ileus and Pseudo-obstruction predominant pathologic patterns have emerged in visceral neuropathy from degenerative sporadic causes. One pattern shows a reduction in the number of intramural cells that occur in relation to swollen processes and nerve cell bodies, an increase in glial cells, and fragmentation and loss of axons; in the second pattern, in the absence of dendritic swelling or glial proliferation there is a loss of the normal staining that occurs in subsets of enteric neurons.113,114,124 Severe forms of idiopathic intestinal intrinsic neuropathy have been associated with a decreased expression of the protein encoded by Bcl-2, a gene related to one of the intra­ cellular pathways leading to apoptosis.125-127

Enteric Mesenchymopathies

The ICC are derived from mesenchymal cells that express c-kit. ICC are important for pacing electrical slow-wave activity and for gastrointestinal motor contractions128,129 (see Chapters 97 and 98). Spindle-shaped ICC are distributed within the muscularis (ICC-IM), and network-forming ICC are closely associated with the myenteric plexus (ICC-MY) of the ENS. The division of morphologically distinct classes of ICC within different layers of the GI musculature suggests that different ICC can perform discrete physiologic roles in intestinal motility.130 ICC are closely apposed to nerve terminals and electri­ cally coupled via gap junctions to neighboring smooth muscle cells. Studies indicate they play a fundamental role in the reception and transduction of both inhibitory and excitatory enteric motor neurotransmission.131 Confocal electron microscopy shows abnormalities of ICC in patients with chronic intestinal pseudo-obstruction that include irregular cell surface markings,132 damaged intracellular organelles and cytoskeleton, and decreased ICC density.113,133 As a result, it has been proposed that ICC involved in pace­ maker activity and neurotransmission to smooth muscle can contribute to the enteric motility abnormalities detectable in patients with chronic pseudo-obstruction.

Enteric Myopathies

Smooth muscle fibrosis and vacuolization are histologic abnormalities that have been observed in the circular and longitudinal muscle layers of some patients with primary visceral myopathies.124,134 In a large series of wellcharacterized patients with chronic pseudo-obstruction, a deficiency of one isoform of the cytoskeletal smooth muscle protein actin, α-actin, was described in approximately 25% of patients with the myopathic phenotype.135 The authors proposed that a selective decrease or absence of α-actin staining in the circular muscle of the small bowel wall could be a biologic marker of chronic intestinal pseudoobstruction,135 although controlled studies have not yet been done.

PRIMARY CAUSES Sporadic Nonfamilial Visceral Neuropathy

Also known as chronic idiopathic intestinal pseudoobstruction (CIIPO), sporadic, nonfamilial visceral neuropa­ thy can result from injury to the myenteric plexus from various causes including drug toxicity, ischemia, radiation, or viral infections such as cytomegalovirus, Epstein-Barr virus, and herpes simplex virus.136 CIIPO is the most common diagnosis given to cases of intestinal pseudoobstruction, followed by various secondary causes that are covered later. Delayed gastric emptying and intestinal dys­ motility have been observed in patients after viral gastroen­ teritis, which is associated with permanent damage to the myenteric plexus in some patients. Kamm and colleagues

have documented such a syndrome in association with herpes viral elements in the myenteric plexus of patients with pseudo-obstruction.120 Patients with CIIPO often have disturbed motility of the entire gastrointestinal tract. The intestine may be dilated. Histologic examination of the myenteric plexus may reveal a reduction in the number or an abnormal morphology of neurons, which may be enlarged with thick, irregular clubbed processes. Hypertrophy of one or both muscle layers of the muscularis propria has been described. Intes­ tinal phasic pressure studies are abnormal, but routine his­ tology findings often were normal in patients whose tissue was available for analysis.137

Familial Intestinal Pseudo-obstruction

Primary familial cases of intestinal pseudo-obstruction result from abnormalities of smooth muscle cells of the muscularis propria (familial visceral myopathies) or abnor­ malities of the enteric neuronal structures (familial visceral enteropathies). These are rare genetic disorders with auto­ somal dominant, autosomal recessive, or X-linked transmis­ sion.115 Abnormal genes and loci have been recognized in various syndromic forms of CIIPO, including the DNA po­ lymerase gamma gene (POLG) on chromosome 21 (21q17), the transcription factor SOX10 on chromosome 22 (22p12), and a locus on chromosome 8.113,115,116,130 X-linked recessive forms of CIIPO have been mapped to Xq28I, and a large cytoskeletal protein filamin A has been identified as vital for proper enteric neuron development.136 Some familial disorders have been well characterized based on the pattern of inheritance and predominant abnormalities. Familial Visceral Myopathies FVMs are uncommon genetic disorders causing chronic pseudo-obstruction and are characterized by degeneration and fibrosis of gastrointestinal smooth muscle; one type also involves urinary smooth muscle. There are at least two phenotypes. Type I is autosomal dominant and usually is diagnosed after the first decade of life, with esophageal dilatation, megaduodenum, megacystis, and mydriasis. Patients respond favorably to regional resection or bypass of affected tissue segments, and the prognosis generally is good. Type II, called MNGIE (mitochondrial DNA neurogastro­ intestinal encephalopathy) is the best-characterized myo­ pathic phenotype, although the pathogenetic mechanisms causing its intestinal dysmotility remain unclear. It is an autosomal recessive mitochondrial encephalopathy that can lead to chronic pseudo-obstruction because of dysfunction of the mitochondrial respiratory chain.138 MNGIE is charac­ terized by exacerbations with severe symptoms on a back­ ground of chronic intestinal pseudo-obstruction with mildly dilated small intestine and dilated stomach, lactic acidosis, ptosis, ophthalmic paralysis, peripheral neuropathy, altera­ tions in the white matter on magnetic resonance imaging (MRI) of the brain, skeletal muscle ragged red fibers (Fig. 120-5), and specific mitochondrial changes at the ultrastruc­ tural level.138,139 Patients manifest skeletal muscle pain and cramps and lactic acidosis with elevated circulating muscle enzyme levels, including creatine phosphokinase (CPK), alanine transferase (ALT), and aldolase. Small intestinal diverticulosis complicated by inflammation and perforation is the cause of death in a majority of these patients in early adulthood.138 MNGIE syndrome results from mutations in the thymi­ dine phosphorylase gene, which leads to markedly reduced

2133

2134

Section X  Small and Large Intestine

Figure 120-5.  Histologic section of a biopsy specimen taken from skeletal muscle showing ragged red fibers (center) containing subsarcolemmal accumulations of mitochondria. Most fibers are morphologically normal. (Used with permission from Mueller LA, Camilleri M, Emslie-Smith AM. Mitochondrial neurogastrointestinal encephalopathy: manometric and diagnostic features. Gastroenterology 1999; 116:959-963.)

thymidine phosphorylase activity that, in turn, results in accumulation of thymidine (dThd) and deoxyuridine (dUrd) in blood and tissues.140 High levels of dThd and dUrd lead to nucleoside pool imbalances that cause abnormalities in DNA including depletion, point mutations, and multiple deletions.140,141 A link has been established between DNA depletion and myopathic changes involving the external layer of the muscularis propria.138 Screening tests for MNGIE include measurements of serum lactic acid, circulating muscle enzymes, and thymidine phosphorylase in circulat­ ing leukocytes.142 Familial Visceral Neuropathies FVNs are rare genetic diseases characterized by degenera­ tion of the neural structures of the myenteric plexus. There are at least two distinct phenotypes of FVN.143 Type I is autosomal dominant, results in segmental dilatation of varying lengths of the small intestine, megacolon, and gas­ troparesis and has its onset of gastrointestinal symptoms at any age; more than 75% of patients with FVN type I have gastrointestinal symptoms. Histology demonstrates degen­ eration of argyrophilic neurons and reduced numbers of nerve fibers. Type II FVN is autosomal recessive and its findings include hypertrophic pyloric stenosis, malrotation of the small intestine, and short and dilated small intestine; CNS malformation and patent ductus arteriosis are seen in some cases. Onset of symptoms is in infancy, and histologi­ cally there is a deficiency of argyrophilic neurons and increased numbers of neuroblasts. There is no effective treatment (medical or surgical) for type II FVN, and progno­ sis is poor. Several syndromic congenital neuropathies exist that can be classified broadly144 as: Disorders of colonization by migrating neural crestderived neurons (e.g., Hirschsprung’s disease) related

to abnormalities in the RET gene (the gene for tyrosine kinase) and GDNF (glial derived neurotrophic factor), or the disorder of ET-3 (endothelin-3) and its receptor, ETB (endothelin B receptor); Disorders of differentiation of enteric nerves (e.g., intesti­ nal ganglioneuromatosis) related to a specific germline point mutation in RET at codon 918 of exon 16 (M918T) or codon 883 (A883F) in multiple endocrine neoplasia (MEN) 2B syndrome; Disorders of the survival or maintenance of enteric nerves (e.g., hypoganglionosis and possibly congenital achala­ sia) related to one of several derangements of ligands (e.g., 5-HT or neurotrophin-3) and their receptors (e.g., 5-HT2B and tyrosine kinase C) or transcription factors (e.g., SOX10). Disturbances in these mechanisms result in syndromic dysmotilities such as Hirschsprung’s disease, WaardenburgShah syndrome, MEN type 2A or B, and idiopathic hyper­ trophic pyloric stenosis. Hirschsprung’s disease, perhaps the most common con­ genital neuropathy, occurs in one in 5000 live births (see Chapter 96). It is often associated with chromosomal abnor­ malities, the most frequent being trisomy 21 (Down syn­ drome) and affecting 2% to 10% of patients. Mutations in gene-encoding receptors or ligands, such as GDNF-Ret or ET-3-ETB, can occur. Mutations in the receptors are much more commonly encountered than mutations in the ligands. Different mechanisms can cause the terminal colon to become aganglionic: A deficiency of GDNF-Ret that is less severe and does not cause the entire bowel to become agan­ glionic; a deficiency of ET-3-ETB in which neural crestderived cells differentiate prematurely and precursors cease dividing prematurely, leaving the last segment uncolonized; and other syndromes that result from mutations of trans­ cription factors. Waardenburg syndrome (WS) is an autosomal dominant familial disorder that affects body pigments and has a sen­ sorineural hearing loss.145 It occurs with an incidence of 1/50,000 live births and accounts for 2% to 5% of congeni­ tal hearing loss. There are at least four types of the syn­ drome, and four genes have been identified in this condition: PAX3, MITF, EDNRB, and EDN3. The combination of Hirschsprung’s disease with WS defines the WS4 type.146 Patients with a Sox10 mutation and WS4 can present with ataxia, central and peripheral demyelinating neuropathies, and megacolon.147 Multiple endocrine neoplasia type 2B syndrome (MEN 2B) is a severe congenital familial neuropathic condition transmitted as an autosomal dominant trait and character­ ized by medullary carcinoma and tumor development in the neuroendocrine system (see Chapter 122). Presenting symptoms can include severe constipation or megacolon, diarrhea (when associated with enterocolitis), or obstruc­ tion, typically occurring shortly after birth.144 External stigmata of MEN 2B include a characteristic facies with “blubbery” lips due to mucosal neuromas that involve the lips, tongue, face, and eyelids (Fig. 120-6)148; marfanoid habitus; medullated corneal nerve fibers; and medullary thyroid carcinoma. Biopsies characteristically show find­ ings of transmural ganglioneuromatosis with massive pro­ liferation of neural tissue, including neurons, supporting cells; nerve fibers appear as thickened nerve trunks among mature nerve cells.

Childhood Visceral Myopathies

CVMs have been recognized as two distinctive forms of disease that differ from FVM in their mode of inheritance and clinical presentations (Table 120-10). Degeneration and

Chapter 120  Ileus and Pseudo-obstruction Table 120-10  Classification of Childhood Visceral Myopathies CHARACTERISTIC(S)

TYPE I

TYPE II*

Mode of transmission Clinical features   Age of onset   Gender   Symptoms Extragastrointestinal features Gross lesions

Autosomal recessive (?)

Autosomal recessive (?)

Infancy and young childhood Both Constipation, distention ± chronic pseudo-obstruction Megacystis and megaureters Dilatation of entire GI tract

Histologic changes

Degeneration and fibrosis of GI and urinary smooth muscle cells No effective Rx; prognosis poor

Infancy Predominantly female Obstipation, intestinal pseudo-obstruction Megacystis and megaureters Short, malrotated small intestine and malfixation of microcolon Vacuolar degeneration of GI and urinary smooth muscle cells No effective Rx; prognosis poor

Treatment, prognosis

*Megacystis-microcolon-intestinal hypoperistalsis. GI, gastrointestinal; Rx, treatment. Camilleri M. Acute and chronic pseudo-obstruction. In: Feldman M, Friedman LS, Brandt LJ, editors. Sleisenger and Fordtran’s Gastrointestinal and liver disease: pathophysiology, Diagnosis, Management. 8th ed. Philadelphia: Saunders Elsevier; 2006. pp 2679-2702.

Figure 120-6.  Typical mucocutaneous feature of multiple endocrine neoplasia type IIB. Clinical photograph shows multiple ganglioneuromas of the tongue, lips (arrows), and face. (Used with permission from Tan FLS, Tan YM, Lim DTH. The significance of cystic adrenal lesions in multiple endocrine neoplasia IIB syndrome. Singapore Med J 2004; 45[10]:495.)

fibrosis of gastrointestinal and urinary smooth muscle can be detected in both types of CVM and result in obstructive symptoms, bowel dilatation, and hydroureteronephrosis or megacystis, the latter resulting from degeneration of bladder muscle.149

SECONDARY CAUSES Systemic Sclerosis (Scleroderma)

Systemic sclerosis (SSc) is a generalized disorder of the microvasculature and connective tissue that results in increased deposition of collagen and other matrix elements, leading to thickening and fibrosis of the skin, with involve­

ment of synovia and fibrosis of intestinal organs. The highest incidence of SSc occurs between 45 and 55 years of age. The gastrointestinal tract is the second most common target organ after the skin and the esophagus is the most com­ monly affected gastrointestinal organ, followed by the small intestine.150 Small bowel dysmotility leads to bacterial overgrowth resulting in steatorrhea, malabsorption, and weight loss. There are two forms of SSc, diffuse cutaneous and limited cutaneous. Gastrointestinal symptoms are associated with diffuse cutaneous SSc and occur along with other manifes­ tations of systemic involvement, including interstitial lung disease, kidney failure associated with malignant hyperten­ sion crisis, and myocardial disease. Intestinal and gastric involvement can be identified radio­ logically in up to 40% of patients. Contrast studies can reveal dilatation of the duodenum and jejunum, along with fixed, narrow, tightly packed valvulae conniventes, despite bowel dilatation, producing an accordion appearance (Fig. 120-7) or a “hide-bound” appearance.151 Wide-necked diver­ ticula are seen more often in the colon than in the small intestine. Disturbances of small bowel motility152,153 result in delayed gastric or small bowel transit154 and include absence or hypomotility of the interdigestive MMC, low-amplitude clusters of propagated and nonpropagated contractions, a prolonged MMC cycle, diminished activity of phase III,153 and antral hypomotility (characterized by low-amplitude contractions, typically less than 40  mm  Hg). Intestinal involvement usually causes fasting and postprandial con­ tractile amplitude to fall below 10 mm Hg.155 Gastric empty­ ing may be delayed by resistance to flow in the hypomotile small intestine, even if the stomach itself is unaffected. SSc patients without gastrointestinal involvement demonstrate normal small bowel manometry155 and normal intestinal transit time.

Dermatomyositis and Polymyositis

The gastrointestinal tract is involved in half of the cases of dermatomyositis and polymyositis. Dysphagia is often the presenting symptom, associated with weakness and atrophy of skeletal muscles of the esophagus. Megaduodenum and delayed intestinal transit are prominent features of chronic pseudo-obstruction. Histopathology demonstrates atrophy and fibrosis of intestinal smooth muscle, most consistent with a visceral myopathy.155

2135

2136

Section X  Small and Large Intestine transit,154 by manometry,157 or sometimes by evidence of a dilated intestine. Motility studies of the small intestine in patients with diabetes have varied significantly. Normal MMCs are found in many patients with documented gastroparesis, but absence of intestinal phase III has been demonstrated in some. Other abnormalities include MMCs that originate in the distal duodenum or jejunum and uncoordinated bursts of nonpropagated contractions (Fig. 120-8).158 The clinical relevance of these abnormal findings is uncertain. Demye­ lination of the proximal vagus nerve and sympathetic nerves supplying the bowel occurs in patients with long-standing diabetes. The intrinsic nervous system of the intestine does not appear to be affected because no morphologic abnor­ malities of the myenteric or submucosal plexuses have been observed; however, animal models and a single case report of a patient requiring transplantation of the pancreas and kidneys was associated with degeneration of the ICC.157 Although thickening of the small bowel muscle layers and eosinophilic hyaline bodies in smooth muscle cells have been observed,159 most authorities believe that myopathy is not a cause of gastrointestinal dysmotility in diabetic patients.

Parkinson’s Disease

Figure 120-7.  Select view from an upper gastrointestinal series in a patient with systemic sclerosis showing a markedly dilated duodenum and proximal jejunum, with an increased number of small bowel folds crowded together (despite luminal distention), producing a “hide-bound” appearance. Also note multiple outpouchings (arrows) due to asymmetrical fibrosis with sacculation of the opposite bowel wall. (Used with permission from Levine MS, Rubesin SE, Laufer I. Pattern approach for diseases of mesenteric small bowel on barium studies. Radiology 2008; 249:451.)

Systemic Lupus Erythematosus

Abdominal pain is the most common gastrointestinal com­ plaint, which occurs in up to 20% of patients. Autopsy studies reveal a 65% prevalence of peritoneal inflammation, although documented cases of serositis as a cause of abdom­ inal pain are rare. Smooth muscle dysfunction related to ischemia can lead to dilatation of the small intestine. Diar­ rhea occurs in 10% of lupus patients. Chronic intestinal pseudo-obstruction can occur as the presenting symptom and is potentially reversible with early treatment.156 Lupus enteritis is the term used to describe the bowel changes in this disease that result from vasculitis of small blood vessels. Ischemia leads to intestinal mucosal ulcer­ ation, edema, and hemorrhage. The diagnosis often can be made radiologically, and small bowel changes include dila­ tation, thick folds, and thumbprinting related to subepithe­ lial edema and hemorrhage. Unusual intestinal manifestations can cause diagnostic difficulties. Lupus can simulate Crohn’s ileitis or can cause a severe protein-losing enteropa­ thy in association with vasculitis and basement membrane thickening that can be diagnosed on full-thickness jejunal biopsy.

Diabetes Mellitus

Patients with diabetes mellitus often have diarrhea, which may be due to bacterial overgrowth, pancreatic exocrine insufficiency, bile salt malabsorption, impaired absorption or secretion, loss of adrenergic sympathetic innervation, or disturbed motility. In patients presenting with symptoms of gastroparesis, the small intestine also may be affected, and such involvement can be documented by prolonged

Patients with Parkinson’s disease often have symptoms of gastrointestinal dysfunction. Proximal dysphagia, bloating, constipation, and difficulty with evacuation of stool occur commonly.160 Dilatation of the small bowel may be seen radiologically.161 Small intestinal dysmotility does occur, but its frequency is not known, and the contribution of these abnormalities to symptoms is not clear. Compared with controls, manometric studies in patients with Parkinson’s disease reveal infrequent or absent MMCs, hypomotility in the fed state, and an increased incidence of retrograde and tonic contractions.162 The pathogenesis of small intestinal dysmotility in Parkinson’s disease has not been determined. Lewy bodies are neurons containing cytoplasmic hyaline inclusions that originally were identified in the brain of parkinsonian patients. They have not yet been reported in the small intes­ tine but have been found in the myenteric plexus of the esophagus and colon.163 Dopaminergic neurons are reported to be deficient in colons from patients with Parkinson’s disease and constipation.164

Spinal Cord Injury

Spinal cord injury usually produces only mild and probably insignificant dysmotility in the small intestine. The only documented changes in manometric studies performed in a group of patients with spinal cord injury were a greater number of phase III contractions that began in the duode­ num rather than in the antrum in patients with high spinal cord lesions.165 Patients with injury to the lower spinal cord demonstrated no abnormalities, a finding consistent with the preservation of innervation to the intestine from the vagus and third thoracic sympathetic levels. One report described a woman with cervical spinal stenosis and para­ plegia who displayed normal MMCs but a two-fold to threefold prolonged interval between MMCs. Immediately after spinal cord injury, a state of spinal shock develops in which there is complete loss of all sensory, motor, and reflex function below the level of injury. Paralytic ileus with abdominal distention follows and usually resolves in a few days. In the long term, however, postprandial abdominal distention and discomfort occur in more than 40% of patients with spinal cord injuries; these symptoms are likely due to the more significant problem of

Chapter 120  Ileus and Pseudo-obstruction Control

Diabetes Mellitus

Antroduodenum 1 2 3 Descending duodenum 50 mm Hg Distal duodenum Jejunum 1 2 3

5 min

Figure 120-8.  Gastroduodenal motility tracings in a healthy control subject and in a patient with diabetes mellitus. These postprandial tracings show normal amplitude and irregular contractility in the healthy control and a paucity of antral contractions, tonic contractions at the pylorus (third tracing), and development of a fasting-like migrating motor complex pattern in the small intestine in a patient with diabetes mellitus. (Adapted from Camilleri M. Acute and chronic pseudo-obstruction. In: Feldman M, Friedman LS, Brandt LJ, editors. Sleisenger and Fordtran’s Gastrointestinal and Liver Disease: pathophysiology, Diagnosis, Management. 8th ed. Philadelphia: Saunders Elsevier; 2006. p 2687.)

constipation. Many stable patients with spinal cord injury exhibit increased amounts of gas in the small and large intestine on routine abdominal plain films. Colonic dysmotility is well recognized and is responsible for the common problems of constipation and difficulty of evacuation in patients with spinal cord injury.166 Spinal cord injury decreases colonic motility; though the postpran­ dial colonic response is present, it is suboptimal and con­ fined to the descending colon.167 In these patients, rectal compliance and resting anal sphincter pressures are lower than normal values, and ramp rectal inflation demonstrates patterns of sphincter activity similar to those recorded in the patients’ cystometrograms. There is no definite relation­ ship of bowel function to the findings on anorectal manom­ etry in patients with spinal cord injury. Rehabilitation goals include continence of stool, simple voluntary independent defecation, and prevention of gastro­ intestinal complications. Individualized person-centered bowel care includes diet, laxatives, enemas, suppositories, and scheduling of bowel care to initiate defecation and accomplish fecal evacuation. Digital-rectal stimulation is a technique used during bowel care of patients with spinal cord injury to open the anal sphincter and facilitate reflex peristalsis.168

Neurofibromatosis (von Recklinghausen’s Disease)

Gastrointestinal involvement in neurofibromatosis is estimated to occur in 11% to 25% of patients.169 Small bowel dysmotility and intestinal pseudo-obstruction due to neurofibromatosis, however, are rare. Involvement of the intestine occurs in three principal forms: hyperplasia of the submucosal and myenteric nerve plexuses and mucosal ganglioneuromatosis, which can lead to disordered intesti­ nal motility and chronic pseudo-obstruction; GIST tumors that show varying degrees of neural or smooth muscle dif­ ferentiation and typically manifest with bleeding; and a glandular, somatostatin-rich carcinoid of the periampullary

region of the duodenum, which also has been associated with pheochromocytoma.170

Paraneoplastic Visceral Neuropathies

Paraneoplastic neurologic syndrome is a remote effect of cancer that results in (visceral) neuropathy. It is not caused by the tumor or metastases, and it is not caused by infection, ischemia, or metabolic disruptions.171 Paraneoplastic neuro­ logic syndrome is caused by autoimmune processes trig­ gered by the cancer and directed against antigens common to both the cancer and the nervous system, designated as onconeural antigens. Paraneoplastic neurologic syndrome is rare, occurring in less than 0.01% of patients with cancer172; the Lambert-Eaton myasthenic syndrome is more common, occurring in nearly 1% of patients with small-cell lung cancer.173 Less than 50% of patients have detectable anti­ bodies and up to 10% have an atypical antibody that is not well characterized.171 Paraneoplastic neurologic syndrome is considered to be mediated by the immune system, so suppression of the immune response represents another potential treatment approach.171 Associated neo­ plasms include small cell lung cancer (SCLC), thymoma, gynecologic and breast tumors, Hodgkin’s lymphoma, mul­ tiple myeloma, and colon cancer. The most commonly associated tumor is SCLC, and the most common antibodies are Hu-Ab or CV2-Ab.171 Experi­ mental evidence suggests that anti-Hu antibodies can exert a direct pathogenic role or can contribute, in association with the lymphocytic infiltrate, to ENS dysfunction in patients with chronic intestinal pseudo-obstruction related to an inflammatory neuropathy.113 Some patients with sub­ acute parasympathetic and sympathetic autonomic failure and prominent gastrointestinal dysfunction also have antibodies directed against the neuronal autonomic gan­ glion type of acetylcholine receptors (nAchR antibodies). In these cases, the main associated tumors are thymoma and SCLC.

2137

2138

Section X  Small and Large Intestine Patients with paraneoplastic neurologic syndrome can experience weight loss, persistent bloating, and abdominal distention from damage to the neurons of the enteric plex­ uses.171 Some patients present with dysphagia, nausea, and vomiting due to esophageal dysmotility or gastroparesis, or, more commonly, they present with symptoms of severe constipation. Radiologic studies can show dilatation of the small bowel, colon, or stomach.171 Esophageal manometry can reveal spasm or achalasia. These inflammatory neuropathies are characterized by a dense inflammatory infiltrate with CD3-positive CD4 and CD8 lymphocytes that almost invariably are confined to the myenteric plexus.115,116,130

Myotonic Dystrophy

Myotonic muscular dystrophy is a slowly progressive disease characterized by myotonia, or difficulty in muscle relaxation. Diarrhea and abdominal cramping occur in up to one third of affected persons, malabsorption and steator­ rhea have been reported in a few cases, and constipation is frequent and can alternate with diarrhea. The small intes­ tine can demonstrate abnormal but nonspecific radiologic changes, including dilatation, diminished motility, and delayed transit.174 Dysmotility of the small intestine can play a significant role in the production of intestinal symptoms. Manometric findings included low-amplitude contractions in fasting and fed states, retrograde propagation of phase III waves, inter­ruption of phase III, and increased incidence of tonic contractions.175 Spontaneous pneumoperitoneum, megaco­ lon, and low-amplitude small intestinal contractions have been documented. Histologically, smooth muscle cells in the small intestine show changes similar to those found in dystrophic skeletal muscle: They are swollen, partially destroyed, decreased in size, and replaced by fat. Using silver stain, degenerative changes of the myenteric plexus of the colon were found in a patient with megacolon,176 indicating that intestinal dys­ motility may be caused by smooth muscle as well as enteric nerve dysfunction. In most patients, the predominant cause of dysmotility appears to be smooth muscle damage.

Duchenne’s Muscular Dystrophy

Duchenne’s muscular dystrophy is the most common of the inherited muscular dystrophies. A sex-linked disease, it affects one in 3500 boys; without respiratory support, death can occur before age 25 years. It is caused by mutations in the gene for dystrophin, a protein that helps stabilizes muscle cell membrane.177 In some cases, the mutations lead to an aberrant messenger RNA and result in a highly trun­ cated dystrophin or no protein at all. Without functional dystrophin, muscle cell membranes leak, followed by muscle fiber necrosis and regeneration, and progressive replacement of muscle by fibrosis and fat. The muscles steadily waste away, and paralysis eventually occurs.177,178 In a milder form with onset in the teens or early adult­ hood (Becker’s muscular dystrophy) the muscle cells produce dystrophin that is not normal but retains enough activity for normal function. As with Duchenne’s, muscle weakness begins in the hips and pelvic girdle. The rate of degeneration is variable in Becker’s muscular dystrophy with many patients living to old age. Gastrointestinal symptoms usually are related to esopha­ geal and gastric dysmotility, which are more severely affected than is the small intestine. Dysphagia is the pre­ dominant gastrointestinal symptom (36% in one series), followed by vomiting, diarrhea, and constipation.179 Gastric

emptying delay and acute gastrointestinal dilatation charac­ terize the clinical course.180 Orocecal transit time can be normal in asymptomatic subjects, although severe bowel dysmotility can occur.

Amyloidosis

Amyloidosis is a mixed group of disorders characterized by extracellular deposits of abnormal protein fibrils with a β-sheet fibrillar structure. The abnormal structure can be identified by x-ray diffraction studies and visualized with electron microscopy, but clinically, amyloid is identified by viewing the intestine with polarized light after staining it with Congo red.181 In patients with amyloidosis, diarrhea and constipation often are present for years and are followed or accompanied by a myriad of disparate problems including gastrointestinal bleeding, steatorrhea, protein loss, perforation, obstruction, intussusception, ischemia, pneumatosis, and pseudoobstruction.182 A variety of mechanisms might explain the diarrhea: delayed motility and small bowel bacterial overgrowth,183 bile salt diarrhea,184 a sprue-like condition with amyloid deposition in the tips in the villi,185 and rapid orocecal transit.186 The severity of gastrointestinal dysmotility is correlated to the quantity and distribution of deposited amyloid. Amyloid is slowly deposited between muscle fibers, causing pressure atrophy of adjacent fibers so that eventually, the whole muscle layer is replaced by amyloid.182 Neuromuscu­ lar infiltration initially affects the intrinsic nervous system and results in a neuropathic process187 characterized by contractions that have normal amplitude but are uncoordi­ nated.182,188 At a later stage, tissue wall infiltration results in a myopathic process with low-amplitude contractions that typically are associated with prolonged transit as in other systemic disorders, such as scleroderma. The vasculature often is involved, with amyloid deposition in the subintima or adventitia, and often involving the submucosa. When the vessel wall thickens, the lumen narrows and eventually obstructs, resulting in ischemia, infarction, and perfora­ tion.189 Mucosal architecture remains normal until massive deposits of amyloid destroy the mucosal structures. There are several forms of amyloidosis: primary (AL), secondary (AA), hemodialysis-related (Aβ2MG), hereditary (ATTR), senile, and localized. AL is most common and has the most significant gastrointestinal involvement. AL amyloid reflects a generalized deposition of excess light chains associated with plasma cell dyscrasia, and 15% of patients with AL have multiple myeloma.190 Amyloid protein is deposited in the small intestine in all forms of amyloidosis.191 Secondary (AA) amyloidosis with acute-phase reactant serum amyloid A protein (A) is associated with infectious, inflammatory, or, rarely, neoplastic disorders.182,192 The inci­ dence of amyloidosis in patients with rheumatoid arthritis ranges from 7% to 21%, the highest prevalence derived from an autopsy study.182 Other disorders associated with AA are Crohn’s disease, ankylosing spondylitis, Reiter’s syndrome, psoriasis, progressive systemic sclerosis, primary biliary cirrhosis, and systemic lupus erythematosus. Inherited forms of amyloidogenic proteins are rare. The most common of this type is with variant transthyretin (TTR), which is produced by the liver,182,183 and the result­ ing amyloidosis (ATTR) is called familial amyloidotic poly­ neuropathy. Besides familial amyloidotic polyneuropathy, there are hereditary non-neuropathic systemic amyloidoses associated with mutations in genes for apolipoprotein AI, lysozyme, and fibrinogen α-chain.

Chapter 120  Ileus and Pseudo-obstruction Senile amyloidosis is found in 10% to 36% of patients older than 80 years and mainly involves the heart, but it also can be seen throughout the gastrointestinal tract. Amyloid has been observed in the subserosal veins of the large and small intestine in 41% to 44% of elderly patients.193

Chagas’ Disease

In nearly one third of patients, the late phase of infection with Trypanosoma cruzi leads to destruction of the sub­ mucosal and myenteric plexuses along the length of the gastrointestinal tract resulting in dilatation (see Chapter 109). Megacolon and megaesophagus are the most common presentations, although megaduodenum and megajejunum also can occur. Swallowing difficulties may be the first symptom of digestive disturbances and can lead to malnu­ trition. Some patients are entirely asymptomatic despite significant destruction of neurons. The early phase of gastric emptying of liquids is accelerated in patients who have Chagas’ disease and megaduodenum, suggesting that increased duodenal receptivity has a significant effect on the gastroduodenal transfer of liquids.194

Thyroid Disease

The earliest manifestation of a thyroid disorder might be gastrointestinal dysfunction. Hypothyroid patients often complain of constipation, and their gastric emptying time may be significantly delayed.195 Intestinal dysmotility (asso­ ciated with change in the frequency of the slow wave oscil­ lations of smooth muscle electrical potential) results from the altered thyroid state and has been considered to be the cause of symptoms. In hypothyroidism, small intestinal transit is significantly slowed.196 Many hypothyroid patients develop constipation, which may be the result of colonic dysmotility. In one patient, manometry revealed decreased amplitude of small intestinal contractions and an overall decreased motility index.197 With severe hypothy­ roidism (myxedema), paralytic ileus and intestinal pseudoobstruction can occur. These abnormalities return to normal after the thyroid disorder is corrected.

Hypoparathyroidism

Cases of pseudo-obstruction and malabsorption have been observed in patients with hypoparathyroidism, although the mechanism is not known. Calcium is essential for smooth muscle contraction, and hypocalcemia can impair intestinal contractile activity. Symptoms improve with administration of calcium.

Medications

Many drugs profoundly affect gastrointestinal motility.198 Although the colon is considered the principal target organ for drug-induced dysmotility, the small intestine often is similarly affected. Tricyclic antidepressants are noted for causing ileus. Phenothiazines and some antiparkinsonian drugs decrease colonic and small intestinal motility and can cause constipation, colonic pseudo-obstruction, and adynamic ileus. The anticholinergic agents atropine and scopolamine and related belladonna alkaloids decrease intestinal tone along with the amplitude and frequency of peristaltic contractions. Opiate analgesics act on µ-opiate receptors throughout the intestinal tract to suppress motility; this effect is most pro­ nounced in the colon. Loperamide, a predominantly periph­ eral opioid antagonist, causes chronic pseudo-obstruction by this antikinetic mechanism.198 Morphine enhances the amplitude of nonpropulsive small intestinal contractions and markedly decreases propulsive contractions. The duo­

denum and jejunum are more prone to these effects than is the ileum, and the overall effect is a delay in transit through the small intestine. Calcium channel antagonists, particularly verapamil, by their direct stimulation of smooth muscle relaxation, cause constipation in up to 20% of patients and can also cause chronic pseudo-obstruction.199 Small intestinal transit time in subjects taking verapamil was unchanged from pretreat­ ment values, although transit through the colon was slowed and this effect likely accounts for the constipation seen with this drug.

Celiac Disease

Intestinal pseudo-obstruction has been documented in asso­ ciation with celiac disease,200 but the mechanism is unclear. Dilated loops of small intestine with delayed transport of barium can be observed radiologically. In one patient who underwent exploratory laparotomy with full-thickness jejunal biopsy, the nerves and muscle cells appeared normal on both light and electron microscopy.200

Jejunal Diverticulosis

Diverticula can occur anywhere in the small intestine, but diverticulosis is most common in the jejunum. Like their counterparts in the colon, diverticula represent herniations through the mesenteric side of the bowel and usually are acquired. Jejunal diverticulosis is associated with many diseases, including scleroderma, celiac disease, MNGIE cytopathy, and Cronkhite-Canada syndrome.201 Patients can present with symptoms of small intestinal bacterial overgrowth, including diarrhea, steatorrhea, weight loss, and megaloblastic anemia. Local complications of diver­ ticula also can occur, such as diverticulitis, bleeding, and perforation.

Acute and Chronic Irradiation

Ionizing radiation results in damage to all structures of the small intestine, including the mucosa, blood vessels, con­ nective tissue, enteric nerves, and smooth muscle (see Chapter 39). Radiation damage to the intestine can be sepa­ rated into acute and chronic injury. Acute radiation injury is associated with symptoms of nausea, vomiting, abdominal pain, and diarrhea, which subside soon after exposure is discontinued. Small intesti­ nal dysmotility can play a significant role in acute radiation enteropathy. Changes in small bowel absorption are consid­ ered the main cause of diarrhea. Delayed intestinal fasting and fed motility patterns can persist up to four weeks fol­ lowing the last exposure.202,203 One study has shown acceler­ ated small intestinal transit in a group of patients undergoing abdominopelvic irradiation compared with their pretreat­ ment values.204 More than 75% of the patients exhibited diarrhea during treatment. Chronic radiation injury can cause gastrointestinal com­ plications decades after the initial exposure. The neuro­ pathic and myopathic injury results in intestinal dysmotility that leads to bacterial overgrowth, diarrhea, and malabsorp­ tion. Contrast studies often show dilated, thickened loops of bowel with air-fluid levels. Atrophy and fibrosis of smooth muscle fibers is a characteristic histologic finding, and proliferation of submucosal neurons of the myenteric plexus with extension into the circular muscle coat has been observed.205

Diffuse Lymphoid Infiltration

Diffuse lymphoid infiltration is a rare cause of pseudoobstruction characterized by a diffuse infiltration of small

2139

2140

Section X  Small and Large Intestine lymphocytes mixed with mature polyclonal plasma cells that reflects a pseudolymphoma rather than a true neo­ plasm.179 It affects all the layers of the intestinal wall206 and is associated with extensive damage to the submucosal and myenteric nerve plexus.179

Idiopathic Myenteric Ganglionitis

Ganglionitis refers to an inflammatory neuropathy charac­ terized by a dense lymphoplasmacytic infiltrate of the myenteric plexus134,205 and a marked reduction in the number of ganglia. Other cases have shown an eosinophilpredominant infiltrate, and neurons in the myenteric ganglia expressed the potent eosinophil chemoattractant IL-5.121,179 Ganglionitis is typically associated with paraneoplastic, infectious, or neurologic disorders, although some cases remain idiopathic. The diagnosis of enteric ganglionitis is supported by detecting circulating antineuronal antibodies against certain targets, such as Hu and Yo proteins, neu­ rotransmitter receptors, and ion channels.134 A brief course of glucocorticoids or other immunosuppressive agent has been shown to be beneficial in some cases, thus reinforcing the value of establishing a correct diagnosis in the early stages of the disease process.134,207

Anorexia Nervosa and Bulimia

Delayed gastric emptying of solids and prolonged orocecal transit are well-established abnormalities in patients with anorexia or bulimia.208 Electrolyte abnormalities, such as hypokalemia, may be responsible in some cases and be related to nutritional deficiencies or concomitant diuretic abuse. Typically, the motility disorder is less significant than the underlying psychological and nutritional manifes­ tations of the disease.

CLINICAL FEATURES

Patients can experience cramping, generalized and epigas­ tric abdominal pain or discomfort; bloating and distention; anorexia; early satiety; nausea and vomiting; and weight loss. Symptoms typically are aggravated by eating. Diarrhea can occur in patients as a result of bacterial overgrowth and malabsorption. A history of repeated abdominal operations without evidence of mechanical obstruction is common; no surgeon wants to miss a true obstruction. Depression can develop as a result of chronic disabling digestive symptoms, deconditioning, and poor quality of life. Physical findings vary according to the severity of symptoms. In less-symptomatic patients, physical exami­ nation may be normal. Patients with more-severe symp­ toms may be dehydrated, cachectic, and malnourished because they are unable to take in adequate fluid or nutri­ ents, or they might have malabsorption due to bacterial overgrowth, which leads to weight loss. The abdomen may be distended and mildly tender. Borborygmi (loud sounds of fluid and gas transfer) may be audible. During an acute obstructive episode of chronic intestinal pseudoobstruction, abdominal examination may be indistinguish­ able from that in true mechanical obstruction, which is why these patients’ abdomens often have multiple surgical scars. Dilatation of the small intestine or of isolated gastrointes­ tinal segments is not a consistent finding in chronic pseudoobstruction.124 During acute exacerbation, abdominal plain films can show air-fluid levels and dilatation of the small intestine. A chronically dilated, atonic intestine and bacterial overgrowth are more common in chronic pseudo-obstruction of intestinal myopathy and less likely in cases of neuropathic origin. Patients with intestinal dilatation often have malabsorption and steatorrhea, which

causes additional weight loss. The prevalence and severity of recurrent obstructive episodes varies from patient to patient and from episode to episode in the same patient. Extraintestinal features can occur, depending on the underlying disease. Megaduodenum, megacystis, and mega­ ureter are commonly seen in type I FVM; these patients can have urinary retention and frequent infections. Type II FVM (MNGIE), discussed earlier, is defined clinically by two or more of the following extraintestinal conditions: progres­ sive external ophthalmoplegia, ptosis, polyneuropathy, leu­ koencephalopathy, or a family history with no other known cause of pseudo-obstruction.209 MNGIE also is associated with hearing loss. Type III FVM is notable for marked dila­tation of the entire gastrointestinal tract without extragastrointestinal features. In chronic intestinal pseudo-obstruction due to secondary causes, patients may or may not have systemic mani­ festations of the underlying condition. Polymyositis and dermatomyositis are characterized by proximal muscle weakness; scleroderma is often associated with skin changes. Cardiac problems (e.g., cardiomyopathy) are common in patients with Chagas’ disease. Causes of paraneoplastic syn­ drome include lung, breast, and ovarian cancer; Hodgkin’s lymphoma; and multiple myeloma. In patients in whom an underlying diagnosis has not been determined, extrain­ testinal symptoms and findings can provide important clini­ cal clues. A careful review of systems is essential in all patients.

COMPLICATIONS Malnutrition

Malnutrition occurs in patients with severe dysmotility of the small intestine as a result of insufficient food intake, vomiting, bacterial overgrowth, diarrhea, and malabsorp­ tion. Anemia can occur due to deficiencies of iron, folate, and vitamin B12. Serum cholesterol, calcium, and albumin levels may be low, especially in patients with malabsorp­ tion secondary to bacterial overgrowth. In severe cases, especially in patients with intestinal dysfunction due to myopathy, long-term parenteral nutrition may be necessary to provide sufficient nutrients.

Total Parenteral Nutrition–Related Disorders

Patients with disabling gastrointestinal symptoms and irre­ versible intestinal failure often require total parenteral nutrition (TPN) (see Chapter 5). These patients are at risk for developing complications such as catheter-related sepsis, venous thrombosis, loss of venous access, and liver failure. Long-term TPN may be complicated by progressive cholestatic liver disease that can become irreversible The onset of hepatic dysfunction in a patient who has intestinal failure and is on chronic TPN is an indication for trans­ plantation of intestine, in isolation or combined with a liver graft.210

Small Intestinal Bacterial Overgrowth

Common in chronic pseudo-obstruction, small bowel bacte­ rial overgrowth syndrome is a condition associated with the proliferation of colon-type bacteria within the small intes­ tine (see Chapter 102). The syndrome is characterized by steatorrhea, flatulence, abdominal discomfort, and bloating, with symptoms of macrocytic anemia, malabsorption of nutrients and vitamins, and weight loss; some patients have no symptoms.211 Direct quantitative cultures of jejunal con­ tents is the diagnostic gold standard,212 although in the clini­ cal setting noninvasive hydrogen breath tests are much more commonly used.211

Chapter 120  Ileus and Pseudo-obstruction Pneumatosis Cystoides Intestinalis

Pneumatosis cystoides intestinalis is a rare condition char­ acterized by multiple gas-filled cysts in the submucosa or subserosa of the colon and, less often, the small intestine.213 The mechanism is thought to involve increased gas produc­ tion by intestinal bacteria, thereby altering the partial pres­ sure of nitrogen in the intestinal wall.214 Pneumatosis cystoides intestinalis is more common in patients with chronic obstructive pulmonary disease, but it occurs in cases of systemic sclerosis and has been described in patients with polymyositis, diabetes, duodenal and gastric ulceration, regional enteritis, and gastrointestinal malig­ nancy. Intestinal peristaltic hypofunction associated with chronic pseudo-obstruction leads to raised intraluminal pressure, allowing the gas-producing bacteria to invade the intestinal mucosa through breaks in the mucosal integrity, forming pneumocysts. This condition can be diagnosed radiologically and endoscopically (Fig. 120-9). Subserous pneumocysts, in particular, are liable to rupture, releasing free gas into the peritoneal cavity (pneumoperitoneum), making it important to distinguish this condition from bowel perforation. Pneumoperitoneum from rupture of pneumatosis cysts does not per se mandate operation, and is not complicated by peritonitis.

A

NATURAL HISTORY

Chronic intestinal pseudo-obstruction is a severely debili­ tating and progressive disorder. Familial cases have their onset at birth or typically in the first few years of life.215 Childhood onset is characterized by a particularly severe course with high mortality rates.215,216 The first occlusive episode in adults is often preceded by years of nonspecific progressive digestive complaints. The diagnosis is made a median of 8.8 years after the initial onset of symptoms, and 88% of adults undergo a mean of 2.96 unnecessary opera­ tions217 before the diagnosis is established.113 The principal causes of death relate to complications of surgery, paren­ teral nutrition, transplantation, and septic shock of gastro­ intestinal origin. Intestinal myopathy and intestinal hypomotility predict poor outcome. MNGIE has a particu­ larly poor prognosis, with death occurring at about age 40 years.113

DIAGNOSIS

The diagnosis is based primarily on clinical symptoms and is supported by radiologic, manometric, laboratory, histo­ pathologic, and endoscopic investigations.

Radiologic Studies

Plain abdominal films are very useful in patients who com­ plain of abdominal distention and bloating because they might show gaseous distention of the areas affected by severe dysmotility; dilatation is greater in cases of pseudoobstruction related to myopathy than when the cause is neuropathic dysmotility. Enteroclysis can be useful for detecting lesions in the small intestine and ruling out mechanical obstruction. Entero-CT scanning allows simul­ taneous internal and external views of the intestinal wall. Abdominal CT and MR scans are important in investigating possible causes of bowel compression, and MR angiography can noninvasively identify congenital or acquired vascular abnormalities. Excretory urograms should be performed in patients with urinary symptoms.113 Radiopaque polyethylene markers that move with the colonic contents are robust measures of intestinal transit and are used to evaluate whole-gut transit. Segmental colonic transit can be measured with daily abdominal plain films after a single dose of radiopaque markers. Metcalf and

B Figure 120-9.  Pneumatosis cystoides intestinalis. A, Endoscopic view showing multiple cystic lesions with an overlying normal mucosal layer in the right colon. B, CT colonography image of the same patient in A showing multiple air-containing polypoid lesions in the ascending colon. (From Kim BN, Jeong JY, Sohn KD, et al. Pneumatosis cystoides coli of the ascending colon: Colonoscopic and CT colonoscopic features. Endoscopy 2007; 39:E73-4.)

colleagues218 simplified the method so that only one film with high kV technique is necessary and radiation exposure is minimized. By this method, radiopaque markers are taken in fixed numbers (24 per day), at the same time (arbitrarily, 9:00 a.m.) each day for three days. On the fourth day, again at the same time, plain film is taken. The method works on the assumption that a 24-hour sampling interval approxi­ mates continuous observation. Rapid transit can cause all the markers to be lost in the feces before filming on the fourth day; conversely, in slow transit, all 72 markers may be present on the single plain film; a film on day seven then gives more information.

Laboratory Tests

A complete blood count can reveal anemia and macrocyto­ sis as a result of malnutrition and bacterial overgrowth (see Chapters 101 and 102). Blood chemistries also reflect malnutrition and malabsorption. Diabetic patients have hyperglycemia, and hypoparathyroid patients can have hypocalcemia. Patients with connective tissue disease can have a positive antinuclear antibody or SCL-70. Patients

2141

2142

Section X  Small and Large Intestine with thyroid disease have changes in serum T3, T4, and TSH levels. Muscular dystrophy or mitochondrial cytopathy patients can have elevated creatine phosphokinase (CPK) and isoenzymes. Hemagglutination and complement fixa­ tion for Chagas’ disease may be positive in patients with a history of living in Central or South America. Antineuronal nuclear antibody (ANNA), SCLC, and anti-Hu and anti-CV2 antibodies should be sought in patients with an acute presentation of symptoms to exclude a para­ neoplastic cause of pseudo-obstruction. Blood lactate, pyru­ vate (signs of acidosis), CPK, ALT (muscle damage), and leukocyte thymidine phosphorylase are screening tests for mitochondrial cytopathy.

Endoscopy

The main indication for upper gastrointestinal endoscopy is to obtain aspirates from the small intestine in patients with clinically suspected bacterial overgrowth and to exclude mechanical occlusion of the gastroduodenal and gastrojejunal regions. Mucosal biopsies of the small intes­ tine should be taken to rule out celiac disease. Colonoscopy can be used to decompress the colon (see earlier).

Surgery

Full-thickness biopsies should be obtained from dilated and nondilated areas of the intestinal tract in all patients with suspected chronic pseudo-obstruction who undergo surgery for unexplained occlusive episodes; biopsy should be con­ sidered in patients with acute symptoms. Biopsies should be processed by traditional staining and by specialized immunohistochemistry techniques in dedicated laborato­ ries with a specific interest in this area.125

Manometry

Small intestine manometry lacks the specificity to diagnose the underlying disease but can provide information about the pathophy­siologic process (see below). Myopathic Pattern In patients with myopathy, manometry demonstrates low-amplitude contractions in the affected segment; this pattern generally is found during both fasting and fed periods. During fasting, the MMC usually is present but is diminished in amplitude. The fed pattern also is associated with reduced frequency of contractions. The poorly devel­ oped intestinal fed pattern is in part a result of reduced gastric emptying. Weston and associates showed that in myopathic disorders, antral amplitudes are usually less than 40  mm  Hg and duodenal amplitudes are less than 10 mm Hg.152 Neuropathic Pattern Neurologic disorders tend to produce uncoordinated bursts of intense contractions of variable duration but normal amplitude. The MMC often is absent or abnormal in these patients. An abnormal rate of migration, and retrograde propagation of the activity front (phase III), also may be noted. Activity fronts can appear to be normal proximally and then arrest or disappear in the more distal segments of the small intestine. In neuropathic disorders, the normal fed pattern might not replace the fasted pattern, MMC-like activity persists postprandially (normally, the MMC activity should be abolished for one hour per 200 kcal ingested), and the frequency of antral contractions in the first hour is typi­ cally less than one per minute, in contrast to the average two (range one to three) per minute in healthy controls.155 These abnormalities reflect dysregulation by the enteric or extrinsic nervous system.

Mechanical Obstruction The manometric pattern of mechanical obstruction (as opposed to pseudo-obstruction) is characterized by giant propagated or nonpropagated contractions lasting at least 10 seconds or clustered contractions lasting one to five seconds followed by more than one minute of absent motor activ­ ity.124 Manometry is not the usual way to establish the diag­ nosis of mechanical obstruction, but it should alert the physician to the possibility of mechanical obstruction and mandate careful small bowel barium examination, such as by enteroclysis. Anorectal manometry is important to rule out Hirschsprung’s disease, particularly in patients with intrac­ table constipation and marked distention of the large intestine.

TREATMENT

Treatment of chronic pseudo-obstruction is suboptimal. Disabling digestive symptoms combined with dysfunction of the alimentary tract commonly lead to weight loss and malnutrition. Anemia can result from deficiencies of iron, folate, or vitamin B12 and low serum cholesterol, calcium, and albumin levels can be seen, especially in patients with malabsorption secondary to bacterial overgrowth. Therapies are primarily aimed at correcting the underlying processes when possible, preventing malnutrition and controlling symptoms.

Secondary Causes

A few types of secondary small intestine dysmotility, such as hypothyroidism, celiac disease, hyperglycemia associated with diabetes, and drug-induced dysmotility, can be treated with thyroid replacement, a gluten-free diet, improving glycemic control, and discontinuing the offending drugs, respectively. There are no specific treat­ ments for most of the secondary causes of small intestine dysmotility.

Pain

Long-term narcotic use should be discouraged because patients can become addicted to narcotics, and narcotics can further disturb gastrointestinal motility. Tramadol, gabapentin, and pregabalin sometimes are prescribed for such pain, but these too can retard intestinal transit. Unfortunately, there are conflicting data on transit effects of these medications in otherwise healthy people and virtually no such data in patients with intestinal pseudo-obstruction.219-221

Constipation

Constipation is common in patients who also have colon involvement. It is important to make certain that these patients have a good bowel movement at least once every few days, because constipation tends to increase symptoms of intestinal dysmotility. Milk of magnesia can be effective in doses of 30 to 60 mL per day (or two tablets three times daily). Enemas may be useful if the patients have had no bowel movement for three days. Bulk-forming laxatives should be avoided in patients with severe small intestine dysmotility, because they often exacerbate gastrointestinal symptoms.

Acute Subocclusive Episodes

Acute episodes are treated similarly to acute mechanical obstruction. Fluid, electrolyte, and caloric support should be given intravenously. Nasogastric decompression may be beneficial for nausea and vomiting; rectal tubes often are ineffective, but placement of a colonoscopic decompression

Chapter 120  Ileus and Pseudo-obstruction tube or cecostomy can be attempted for acute refractory dilation (see earlier). Pharmacotherapy using prokinetic agents can be used to reduce duration of the acute episode (see earlier). Prolonged subocclusive episodes necessitate appropriate nutritional support and intravenous or poorly absorbable antibiotics to prevent bacterial overgrowth.

Nutrition Management

In patients with gastrointestinal dysmotility, abdominal pain, bloating, nausea, and vomiting often are related to eating. Symptoms usually can be minimized by manipulat­ ing the amount, constituents, and frequency of meals. Patients need sufficient calories without overloading the inefficient bowel: a useful general rule is 25  cal/kg of the patient’s ideal body weight per day taken in small, equalsize feedings four to six times daily. Multivitamins and salt supplementation generally is recommended. Because a liquid meal empties faster from the stomach and probably progresses more readily through the small intestine than a solid meal does, the liquid meal or perhaps an emulsified solid meal may be better tolerated than a solid meal. Supple­ mental formulas contain the daily vitamin and nutrient requirements, and many are lactose free. There is no advan­ tage to using elemental formulas. Hypercaloric solutions are available, although generally formulas that have a lower caloric content (1  kcal/mL) are better tolerated than those with greater caloric density. Parenteral nutrition may be required to meet caloric and nutritional needs, but oral intake should continue as tolerated. TPN-related complica­ tions including catheter-related sepsis or thrombosis, liver insufficiency, pancreatitis, and glomerulonephritis113 sig­ nificantly increase the morbidity associated with chronic intestinal pseudo-obstruction.

Pharmacotherapy

The aim of treatment is to reduce digestive symptoms and lower the risk of complications. Disordered motility of the small intestine leads to bacterial overgrowth and malnutri­ tion. Poorly absorbed antibiotics, such as rifaximin, are the preferred treatment for bacterial overgrowth. Cycled treat­ ments using metronidazole, ciprofloxacin, and doxycycline can limit resistance. Antiemetics and antispasmodic agents should be tried, although their benefit may be short-lived. Improving intestinal motor function is another means of reducing symptoms. Two controlled trials showed cisapride can improve symptoms and enhance emptying in patients with chronic intestinal dysmotility.222-224 The positive effect of cisapride is not as apparent in patients with vagal dys­ function, such as in diabetes and following vagotomy.224 Other prokinetic agents, including erythromycin, meto­ clopramide, domperidone, neostigmine, and bethanechol, have been used to treat chronic pseudo-obstruction, but results are anecdotal and high quality studies are lacking. Domperidone is not FDA approved and is unavailable in the United States. In a small, short-term study, the soma­ tostatin analog octreotide stimulated intestinal motility, possibly reduced bacterial overgrowth, and improved abdominal symptoms in patients with scleroderma.152 Other open-labeled studies confirmed the long-term effectiveness of octreotide with erythromycin in the treatment of chronic pseudo-obstruction, but some data show that octreotide retards gastric225 and small bowel transit in health, and many use the drug (50 µg dose) to induce MMC-like contrac­ tions at least two hours after the last meal of the day to sweep residue out of the small intestine toward the colon and prevent bacterial overgrowth.226 Oral pyridostigmine (60 mg/5 mL, 1 2 to 2 teaspoons one to three times daily with gradual titration as needed 10-15 minutes before meals) may

be effective in reducing symptoms of bloating and constipa­ tion, but clinical trials with this agent are needed. Treatment of inflammatory neuropathies is centered on the use of immunosuppressive therapies. The diagnosis should be suspected because of the presence of circulating antibodies and proved through specialized tissue analysis, if available.124 When a diagnosis of an enteric ganglionitis or inflammatory neuropathy has been established, glucocor­ ticoids, such as prednisolone (60-100 mg/day), methylpred­ nisolone (up to 600  mg/day), or beclomethasone (200  µg three times daily), alone or in combination with other immunosuppressive treatments, such as azathioprine or cyclophosphamide, have been associated with significant clinical improvement.134

Surgical Therapy

In carefully selected patients, such as those with megaduo­ denum of the small intestine, resection or bypass of dys­ functional segments can have a good prognosis.227 Venting gastrostomy and enterostomy may be used to reduce abdom­ inal distention and vomiting, which can have a positive effect on conveying intestinal contents through the gastro­ intestinal tract.113 Because chronic intestinal pseudoobstruction tends to be a progressive disorder, surgery might provide only temporary benefit. Unnecessary surgery should be avoided because it can create adhesions and future addi­ tional difficulties. Small intestine and multiorgan transplantation are gaining recognition as standard treatment for patients with irrever­ sible intestinal failure in whom TPN fails or those who have high morbidity and poor disease-related quality of life despite optimal parenteral nutrition.228 Isolated intestinal transplant is technically less complex than transplantation of combined or multiorgan grafts, and it is preferred in cases of liver disease associated with intestinal failure because the liver failure often reverses if transplantation is successful.229 Immunosuppressive management, such as tacrolimus, represents a landmark in the treatment of patients undergo­ ing transplantation for intestinal failure. Advances in immunosuppressive protocols, including induction therapy with monoclonal antibodies (daclizumab, alemtuzumab, or thymoglobulin), has lowered immunosuppressive require­ ments and reduced the risk of renal failure and sepsis. Advances in surgical techniques and graft monitoring via graft ileoendoscopy and intestinal biopsy have led to improved graft and patient survival rates.229 In one series of isolated intestinal transplants in adult patients, the survival rates were similar to those for other solid organ transplanta­ tions.229 Three year patient survival rates of 70% for isolated intestinal transplantations and 41% for multivisceral trans­ plantations are reported, with mortality rates of 32.5% from sepsis (63%) or rejection.229

MEGACOLON AND MEGARECTUM Megacolon and megarectum are descriptive terms without etiologic or pathophysiologic implications. Megacolon has been defined as a diameter of the rectosigmoid region or descending colon on abdominal plain film of greater than 6.5  cm, of the ascending colon of greater than 8  cm, or of the cecum greater than 12 cm. Megacolon can be caused by aganglionosis (Hirschsprung’s disease), can be idiopathic (complicating chronic constipation of any cause), or may be a manifestation of a generalized gastrointestinal dysmotility (intestinal pseudo-obstruction).

2143

2144

Section X  Small and Large Intestine In congenital megacolon (Hirschsprung’s disease), a con­ genital absence of the intramural neural plexus that medi­ ates relaxation (aganglionosis) causes narrowing (most often of the rectum) of a variable distance proximally and results in functional obstruction (see Chapter 96). Short segment involvement (rectum or rectosigmoid) is most common, although an ultrashort segment of aganglionosis involving only the internal anal sphincter or long segment involvement (less than 20% of patients) can occur. Phy­siologic testing may be necessary to define ultrashort segment as morphologic, and radiologic confirmation of the diagnosis may be difficult. Hirschsprung’s disease is detected more commonly in children than in adults, but ultrashort segment disease is usually not diagnosed until adulthood.230 Acquired megacolon describes any of the many causes of constipation that are associated with colonic dilatation that was not present at some earlier examination. A common background for acquired megacolon is colonic inertia, which can occur at both extremes of life. In children, this form can be confused with the congenital condition. Infec­ tion with Trypanosoma cruzi (Chagas’ disease) is the most common cause of acquired megacolon worldwide. In this condition, the dilated segment of colon is abnormal, owing to destruction of the enteric nervous system by the organ­ ism’s neurotoxin. Chagas’ disease is most common in persons who have lived in South America.

KEY REFERENCES

Antonucci A, Fronzoni L, Cogliandro L, et al. Chronic intestinal pseudoobstruction. World J Gastroenterol 2008; 14(19):2953-61. (Ref 113.) Apfel CC, Korttila K, Abdalla M, et al. A factorial trial of six interven­ tions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004; 350:2441-51. (Ref 77.) Bauer AJ, Boeckxstaens GE. Mechanisms of postoperative ileus. Neuro­ gastroenterol Motil 2004; 16(Suppl 2):54-60. (Ref 17.) Bauer TM, Schwacha H, Steinbruckner B, et al. Diagnosis of small intestinal bacterial overgrowth in patients with cirrhosis of the liver: poor performance of the glucose breath hydrogen test. J Hepatol 2000; 33:382-6. (Ref 212.) De Giorgio R, Guerrini S, Barbara G, et al. Inflammatory neuropathies of the enteric nervous system. Gastroenterology 2004; 126:1872-83. (Ref 134.) Ebert EC, Nagar M. Gastrointestinal manifestations of amyloidosis. Am J Gastroenterol 2008; 103:776-87. (Ref 182.) Kreiss C, Toegel S, Bauer AJ. α2-Adrenergic regulation of NO production alters postoperative intestinal smooth muscle dysfunction in rodents. Am J Physiol Gastrointest Liver Physiol 2004; 287:G658-66. (Ref 40.) Luckey A, Livingston E, Tache Y. Mechanisms and treatment of post­ operative ileus. Arch Surg 2003; 138:206-14. (Ref 19.) Pironi L, Spinucci G, Paganelli F, et al. Italian guidelines for intestinal transplantation: potential candidates among the adult patients managed by a medical referral center for chronic intestinal failure. Transplant Proc 2004; 36:659-61. (Ref 228.) Vanek VW, Al-Salti M. Acute pseudo-obstruction of the colon (Ogilvie’s syndrome). An analysis of 400 cases. Dis Colon Rectum 1986; 29:20310. (Ref 3.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

121 Tumors of the Small Intestine Emmy Ludwig and Robert C. Kurtz

CHAPTER OUTLINE Epidemiology  2145 Pathology  2145 Etiology  2147 Risk Factors  2147 Small Intestinal Adenocarcinoma  2148

EPIDEMIOLOGY Small intestinal neoplasms are a diverse and rare group of tumors that affect the gastrointestinal tract from the duodenum to the ileum. Although the small intestine represents 75% of the length and 90% of the absorptive area of the gastrointestinal tract,1 less than 5% of all primary gastrointestinal tract tumors originate in the small intestine. Many types of tumors arise from the small intestine, and about 65% of them are malignant.2 The four most common types of malignancies—adenocarcinomas, carcinoids, lymphomas, and sarcomas (including gastrointestinal stromal tumors or GISTs)—account for 95% of all small intestinal tumors. In most case series of Western patients, adenocarcinoma is the most common small intestinal neoplasm, followed closely by carcinoid tumor.3,4 Adenocarcinoma (Fig. 121-1) most commonly arises in the duodenum, and carcinoid tumor and lymphoma predominate in the jejunum and ileum. Sarcoma distributes equally among all three segments of the small intestine.5 A review of 144 cases of malignant tumors of the small intestine showed that 64% of patients were men, with a median age of 55.7 years; 47% of the tumors were adenocarcinomas, 28% were carcinoid tumors, 13% were sarcomas, and 12% were lymphomas. Overall five-year survival was 57% and median survival was 53 months. Survival was highest for early-stage tumors and when curative resection was performed successfully.6 Survival was not related to location of tumor. A recent study using the Connecticut Tumor Registry data reviewed the epidemiologic and clinical characteristics of 1260 cases of small intestinal tumors. Of these tumors, 49.8% were found in men; the mean age at presentation was 65.2 years. As for location, 25.4% occurred in the duodenum, 15.3% were in the jejunum, and 29.7% arose from the ileum; the remainder of tumors were either multifocal or their location was unspecified. By type, 33.1% were carcinoids, 30.1% were adenocarcinomas, 16.5% were lymphomas, and 7.1% were GISTs. Virtually all (99%) of the tumors were invasive. Surgery was the primary treatment in 87.7% of patients.7

Clinical Features  2148 Diagnosis  2149 Treatment  2151 Prognosis  2152 Other Malignant Small Intestinal Neoplasms  2152

Surveillance Epidemiology and End Results (SEER) data estimate that in the United States in 2008, small intestinal cancers were diagnosed in 3200 men and 2910 women, of whom 1110 people will die of these tumors. From data for 2001 to 2005, the median ages at diagnosis and death were 67 years and 71 years respectively. The overall five-year relative survival rate for 1996 to 2004 from 17 SEER geographic areas was 57.8%; 31% of patients had localized disease at diagnosis, 33% had locally advanced disease, and 29% had metastatic disease at presentation. The corresponding five-year relative survival rates were 77.0% with localized disease, 62.2% with locally advanced disease, 36.0% with metastatic disease, and 40.2% for patients whose stage of disease was not known.8 Worldwide, rates of small intestinal tumors vary; rates are higher in North America and Western Europe and lower in Asia. The Maori of New Zealand have an unusually high incidence of small intestinal tumors.9 Within the United States, incidence rates appear to be somewhat higher in African Americans and in men. Some have suggested an increasing incidence of small intestinal cancer over time.10 Carcinoid tumor (see Chapter 31) is the second most commonly diagnosed small intestinal malignancy, with an annual incidence of approximately three cases per million persons in the United States. The average age at diagnosis is 55 years. The most common sites of extranodal lymphomas (Fig. 121-2) are the stomach and small intestine (see Chapter 29). GISTs can be found throughout the small intestine (see Chapter 30).

PATHOLOGY Neoplasms can arise from any of a wide range of epithelial and nonepithelial cells of the small intestine. Adenomas and adenocarcinomas arise from glandular mucosa; carcinoid tumors originate from argentaffin cells; lymphomas develop from clonal proliferation of lymphocytes; and GISTs have been identified as originating from the interstitial cells of Cajal, which help control the motility of the intestine and have elements of both smooth muscle and

2145

2146

Section X  Small and Large Intestine Table 121-1  TNM Classification CATEGORY

DESCRIPTION

Tumor (T) Tx T0 Tis T1 T2 T3

Primary tumor cannot be assessed No evidence of primary tumor Carcinoma in situ Tumor invades lamina propria or submucosa Tumor invades muscularis propria Tumor invades through muscularis propria into subserosa T4 Tumor invades through visceral peritoneum or directly invades other organs or structures Regional Lymph Nodes (N) Nx Regional lymph nodes cannot be assessed N1 No regional lymph node metastases N2 Regional lymph node metastases Distant Metastases (M) Mx Distant metastases cannot be assessed M1 No distant metastases M2 Distant metastases Figure 121-1.  Endoscopic image of a mucosal neoplasm in the jejunum of a patient with metachronous small intestine adenocarcinomas. (Courtesy of Mark A. Schattner, MD, New York.)

Table 121-2  Staging of Small Intestinal Adenocarcinoma

STAGE 0 1 2 3 4

percentage TUMOR NODE METASTASIS of cases at (T) (N) (M) presentation Tis T1 or T2 T3 or T4 Any T Any T

N0 N0 N0 N1 Any N

M0 M0 M0 M0 M1

2.7 12 27 26 32.3

Note: Tumor stages are defined in Table 121-1.

Figure 121-2.  Endoscopic image of a mucosal neoplasm in the duodenum in a patient with anemia and overt gastrointestinal bleeding. Pathology revealed an Epstein-Barr virus and associated B-cell lymphoma. (Courtesy of Satish Nagula, MD, New York.)

neural differentiation.11 Leiomyomas arise from smooth muscle cells in the muscularis propria or muscularis mucosa, can be intra- or extraluminal, and account for about 40% of benign intestinal neoplasms. Leiomyomas with more than two mitoses per high-power field are reclassified as leiomyosarcomas. Lipomas represent approximately 20% of benign small bowel neoplasms, are most commonly intraluminal, and are found in the ileum; they rarely bleed, but they can obstruct or intussuscept (Fig. 121-3).12 The remainder of benign neoplasms include desmoid tumors (see Chapter 122) and hemangiomas (see Chapter 36). Small intestinal adenocarcinoma is staged using the TNM system as shown in Tables 121-1 and 121-2. Subtypes of

this cancer include adenocarcinoma in situ, mucinous adenocarcinoma, signet cell carcinoma, squamous cell carcinoma, adenosquamous cell carcinoma, small cell carcinoma, medullary carcinoma, and undifferentiated carcinoma. Typical patterns of distribution are seen with benign lesions of the small intestine just as with malignant lesions. The most common benign tumor is the adenoma, which can be found throughout the small intestine but is slightly more common in the duodenum and ileum. Adenomas account for about 30% of all benign small intestinal neoplasms. They are more likely to be polypoid in the ileum and sessile and villous in the duodenum. Villous adenomas tend to be somewhat larger and often occur in the second portion of the duodenum; 40% to 45% can have pathologic features of malignancy. Benign lesions of the small intestine can often be removed by endoscopic means, after being localized using some or all of the methods described later. Techniques for resection and destruction include snare polypectomy, endoscopic mucosal resection (EMR), and argon plasma coagulation. Early duodenal cancers also have been removed using EMR with good results. Because the duodenal wall is thin, the risk of perforation with EMR is high; careful submucosal injection of the duodenum is recommended before removing a duodenal lesion.13 Two studies have shown EMR to be safe and effective, with the potential of reducing the need for surgical intervention for these relatively rare conditions.14,15

Chapter 121  Tumors of the Small Intestine

A

B

Figure 121-3.  A, Film from an enteroclysis study that demonstrates a smooth submucosal lesion that was found to be a lipoma (arrow). B, Surgical resection specimen of a lipoma from another patient who presented with intussusception and bleeding. (Courtesy of Igor Laufer, MD, Philadelphia.)

Table 121-3 shows the variety of tumors that can arise in the small intestine.

ETIOLOGY Despite similarities in genetic mechanisms underlying adenocarcinoma of the small intestine and colon, the incidence of small intestinal adenocarcinoma is far lower than that of colonic adenocarcinoma. Given the identical genetic makeup and similar tissue anatomy of the small intestine and colon, the marked difference in cancer incidence between the two sites is difficult to explain. Both tumors in both locations likely develop via an adenoma-carcinoma sequence, and mutated K-ras, Ki-ras, TP53, DPC-4 (Smad4), β-catenin, and mismatch repair genes are implicated in tumor progression (see Chapter 123).16 Factors that may reduce the incidence of small intestinal adenocarcinoma include higher intraluminal pH, liquidity of small intestinal content, rapid transit time of chyme, faster turnover rate of small intestinal epithelium, lower load of carcinogenproducing bacteria, and increased amounts of immuno­ globulin (Ig)A-secreting lymphoid tissue surrounding the small intestinal lumen.7

RISK FACTORS Despite the low incidence of malignant small intestinal tumors, several risk factors have been identified that contribute to their development (Table 121-4). A dietary study found that increased ingestion of refined carbohydrates, sugar, and red meat increased the risk of small intestinal adenocarcinoma, whereas a diet high in fish, fruits, and vegetables was protective. Alcohol consumption did not affect risk.17 A similar study also found a significantly increased risk with red meat consumption, as well as saltcured or smoked foods, but did not show fruit and vegetable consumption to be protective.18 Tobacco use and alcohol consumption did not affect development of small intestinal cancer.18

Table 121-3  Some Benign and Malignant Tumors of the Small Intestine Epithelial Tumors Adenoma Tubular Tubulovillous Villous Carcinoid Mixed carcinoid-adenocarcinoma Carcinoma Adenocarcinoma Adenosquamous carcinoma Medullary carcinoma Mucinous adenocarcinoma Signet-ring cell carcinoma Squamous cell carcinoma Undifferentiated carcinoma Gangliocytic paraganglioma Intraepithelial neoplasia associated with chronic inflammation High-grade Low-grade Non-Epithelial Benign Tumors Fibroma Hemangioma Leiomyoma Lymphangioma Neurofibroma Neurolemmoma Non-Epithelial Malignant Tumors Fibrosarcoma Gastrointestinal stromal tumor (GIST)* Hemangiosarcoma Kaposi sarcoma Leiomyosarcoma Liposarcoma Lymphangiosarcoma Neurofibrosarcoma Malignant Lymphomas Burkitt’s lymphoma Diffuse large B-cell lymphoma Immunoproliferative small intestinal disease (IPSID) Mantle cell lymphoma T-cell lymphoma Western type B-cell lymphoma (MALT) *GIST may be a benign or malignant tumor, although all GISTs have malignant potential (see Chapter 29).

2147

2148

Section X  Small and Large Intestine Table 121-4  Risk Factors for Small Intestinal Tumors African American ethnicity Male gender Celiac disease Crohn’s disease Increasing age Inherited adenomatous and hamartomatous polyposis syndromes Familial adenomatous polyposis Hereditary nonpolyposis colorectal cancer Peutz-Jeghers syndrome

Other etiologic factors for the development of small intestinal cancer include Crohn’s disease (see Chapter 111), in which the risk for both colon cancer and small intestinal adenocarcinoma is elevated.19-21 A recent meta-analysis pooling 34 studies of 60,122 patients with Crohn’s disease found a relative risk of 28.4 (95% confidence interval [CI]: 14.46 to 55.66) for the development of small intestinal adenocarcinoma.22 Celiac disease (see Chapter 104) also is thought to be a risk factor for developing non-Hodgkin’s lymphoma (NHL) and small intestinal adenocarcinoma.23 Another study showed an increase in esophageal cancer, melanoma, and NHL in patients with celiac disease; only the risk of NHL persisted despite patient adherence to a gluten-free diet.24 Familial adenomatous polyposis (FAP) predisposes to small intestinal adenomas with associated malignant potential (see Chapter 122). Most of these FAP-associated lesions are found in the periampullary duodenum, although they can occur throughout the length of the duodenum. More than 90% of patients with FAP have a varied number of duodenal adenomas, but only 5% to 10% develop duodenal cancer.25,26 Upper gastrointestinal polyps in patients with FAP might follow a different genetic pathway from that of colorectal cancer.27 FAP mice models show overexpression of the β-catenin gene and deregulation of its signaling pathway. In these animals, a diet of chenodeoxycholate results in more significant duodenal polyposis secondary to the overexpression of β-catenin.28 Bile has been proposed as a potential risk factor for small intestinal cancer,29,30 and there is some suggestion it may be synergistic with the genetic mutations seen in patients with FAP.31 Hereditary nonpolyposis colon cancer (HNPCC) patients also are at increased risk for developing small intestinal cancer (see Chapter 122). HNPCC patients can present with a small intestinal adenocarcinoma as their first and only tumor.32 Screening recommendations for patients with these inherited syndromes might change with the advent of newer imaging techniques for the small intestine.33,34 All adenomas of the small intestine, like adenomas of the colon, should be considered precancerous. Adenomas in both locations appear to progress through an adenomato-adenocarcinoma sequence. Malignant features are found in approximately one third of all small intestinal adenomas.35 For reasons that are not clear, adenomas at the ampulla of Vater are typically villous, large, and more likely to be malignant than adenomas found elsewhere in the small intestine.2 Patients with Peutz-Jeghers syndrome develop hamartomatous polyps of the small intestine and colon (see Chapters 22 and 122), and adenocarcinoma can arise within adenomatous foci in the hamartoma, although the overall risk of this is unknown.36 Other hamartomatous polyposis syndromes (see Chapter 122) include juvenile polyposis, which has an increased risk of gastrointestinal tract malignancy, also probably from adenocarcinoma arising from hamartomatous polyps. Other syndromes include

Bannayan-Ruvalcaba-Riley syndrome, Cowden’s syndrome, Cronkhite-Canada syndrome, and Devon family syndrome. The gastrointestinal cancer risk in Cowden’s syndrome is well described, the others are less so.37

SMALL INTESTINAL ADENOCARCINOMA CLINICAL FEATURES

The symptoms and signs of small intestinal adenocarci­ noma are directly related to the location of the lesion. As an example, adenocarcinoma of the duodenum rarely manifests with symptoms of intestinal obstruction; instead, patients tend to develop frank or occult gastrointestinal blood loss, abdominal pain, and biliary obstruction. In contrast, patients with adenocarcinoma of the jejunum and ileum have vague nonspecific complaints, including abdominal pain, nausea, vomiting and abdominal distention, coupled with overt or occult bleeding. Those who present with iron-deficiency anemia may have improved survival because of their earlier presentation. Most neoplasms of the small intestine are not associated with symptoms and are diagnosed either late in their course or incidentally at laparotomy or autopsy. The general absence of symptoms can be attributed to the distensibility of the small bowel wall and the liquid nature of its luminal contents. If a lesion leads to symptoms, the presentation depends upon the pathology of the neoplasm and its location. At least 50% of benign lesions remain asymptomatic, whereas 70% to 90% of malignant lesions are associated with symptoms. No symptoms or signs are specific for either benign or malignant tumors, however, and even though the duration of symptoms tends to be shorter for patients with a malignancy than for those with benign lesions, months can elapse before the diagnosis is made. If a lesion becomes large enough, a patient can present with cramping periumbilical pain, bloating, and nausea and vomiting resulting from mechanical small bowel obstruction. Small bowel obstruction is the most common presentation for benign lesions and occurs in as many as 70% of such cases. Obstruction can result from either luminal constriction or intussusception. In fact, a benign small bowel neoplasm is the most common cause of intussusception in adults. Adenomas, adenocarcinomas, and lymphomas tend to grow into the lumen, but they can grow through the submucosa and muscle, often becoming large before a diagnosis is made; these lesions also can cause volvulus. As many as 80% of malignant tumors are associated with abdominal pain, although the pain is not caused by obstruction. Back pain in a patient with a primary malignant small bowel lesion suggests spread to the retroperitoneum. Mechanisms of back pain include hemorrhage into the tumor, invasion of enteric ganglia, ischemia, and involvement of the serosa. Gastrointestinal bleeding, usually chronic, is the second most common symptom of neoplasms of the small intestine. Bleeding occurs in 20% to 50% of patients with benign lesions (e.g., leiomyomas) and, less commonly, in patients with malignant lesions. Massive hemorrhage is more common with sarcomas (e.g., GISTs) than it is with carcinomas, carcinoid tumors, or lymphomas. Weight loss is rare in patients with benign lesions, but, along with anorexia, it is noted in 50% of patients with malignancies. Intestinal perforation also is rare in patients with benign lesions but occurs in 10% of patients with sarcomas and lymphomas. Periampullary lesions can result in jaundice or pancreatitis. In patients with carcinoid tumors, symptoms of the carcinoid syndrome typically do not develop in the absence of

Chapter 121  Tumors of the Small Intestine hepatic metastases, and even with hepatic lesions, 28% to 50% of patients with such tumors remain free of the carcinoid syndrome (see Chapter 30). Physical examination might not reveal a small inte­stinal malignancy, although up to 25% of such patients have a palpable abdominal mass. Twenty-five percent of patients with malignancy also present with the findings of obstruction: distention, borborygmi, a palpable mass, and diffuse mild to moderate abdominal tenderness. Some patients have a positive fecal occult blood test, although the incidence of this positivity is highly variable and ranges from 10% to 70%. Other patients present with jaundice secondary to either biliary obstruction or hepatic replacement by metastases. Cachexia, hepatomegaly, and ascites may be present in patients with advanced metastatic disease. Laboratory studies performed in patients presenting with early-stage disease may be unremarkable, which might in part explain why the diagnosis is often delayed. Gastrointestinal blood loss can lead to a microcytic iron-deficiency anemia. The presence of a duodenal lesion can block the biliary system, leading to an increase in serum alkaline phosphatase and bilirubin levels. Serum carcinoembryonic antigen (CEA) is an unreliable tumor marker in small intestinal adenocarcinoma, much as it is in other gastrointestinal cancers, and is rarely elevated unless hepatic metastases are present. 5-Hydroxyindole acetic acid (5-HIAA) is a useful marker for carcinoid tumors, but because it undergoes first-pass metabolism by the liver, urinary levels of 5-HIAA are not elevated other than in the setting of significant hepatic metastatic disease.

DIAGNOSIS Plain Films

Plain abdominal films rarely are useful for localization of small intestinal tumors, but they may be helpful clinically in the setting of a suspected intestinal obstruction or perforation. In 50% to 80% of patients with small intestinal neoplasms, an upper gastrointestinal series with a small intestinal follow-through (UGIS-SBFT) reveals an abnormality38 and specifically demonstrates a tumor in 30% to 44% of cases (Fig. 121-4); this figure increases to 90% if an enteroclysis study (small bowel enema) is performed (see Fig. 121-3).39 A barium enema can reveal lesions in the

Figure 121-4.  Film from a small bowel follow-through examination demonstrating an apple-core appearance caused by a metastatic lesion to the small intestine from a scirrhous gastric cancer. (Courtesy of Igor Laufer, MD, Philadelphia, Pa.)

terminal ileum if there is sufficient reflux of barium through the ileocecal valve.

Video Capsule Endoscopy

Video capsule endoscopy (VCE) uses a capsule that contains a radiotransmitter to generate photographic images of the small intestine. The patient swallows the capsule, which is propelled by peristalsis through the small intestine. The capsule generates two images per second and has a battery life of eight hours. Images are captured by a digital recorder worn by the patient, and the images are downloaded in a computer to be reviewed later by the physician.40,41 This technology is useful for detecting tumors of the small intestine, especially when they are beyond the reach of push enteroscopy (Fig. 121-5). In 2002, a prospective trial compared traditional barium UGIS-SBFT with VCE in 22 patients with suspected small intestinal abnormalities. SBFT was found to be diagnostic in four (20%) patients; VCE was diagnostic in nine (45%) patients and suspicious in eight (45%).42 VCE was well tolerated. A much larger study published in 2006 looked at 562 patients who underwent VCE from August 2001 to November 2003 for a variety of indications; 79% (443/562) had evidence of occult gastrointestinal bleeding. A diagnosis was made by VCE in 277 patients (49.3%). Of the 562 patients studied, 50 (8.9%) were found to have a small intestinal tumor, 48% of which were malignant.43 VCE also has been used to locate the primary small intestinal lesion in patients with metastatic neuroendocrine tumors. In a study performed in 2006, 20 such patients were evaluated by CT scan, enteroclysis, nuclear imaging (111In pentetreotide scintigraphy) and VCE.44 Nuclear imaging revealed an abnormality in the abdominal region in 13 patients but could not localize the abnormality further. In contrast, VCE found nine small intestinal tumors, of which five were carcinoid tumors, for a diagnostic yield similar to that of nuclear imaging; VCE findings were felt to be more clinically helpful and relevant.

Figure 121-5.  Capsule endoscopy view of an ulcerated mass in a patient who presented with gastrointestinal bleeding. Four ulcerated, bleeding masses were found throughout the small bowel; these were confirmed at surgery and found to be sarcomas. (Courtesy of Ann Marie Joyce, MD, Philadelphia, Pa.)

2149

2150

Section X  Small and Large Intestine Other studies have shown the superiority of VCE in detecting small intestinal cancer over enteroclysis, push enteroscopy, colonoscopy, and CT scans. VCE was instrumental in diagnosis or management in 55% to 77% of patients evaluated.45-48 A study reported in 2006 by Bailey and colleagues looked at the diagnosis and outcome of small intestinal tumors found by VCE.49 In their study, 416 VCEs were performed in Australia between May 2001 and May 2004, identifying 27 tumors in 26 patients. Twenty-one of the 26 evaluations were for occult gastrointestinal bleeding. Twenty-three patients had both colonoscopy and upper endoscopy, which were unrevealing,23 and had prior radiologic or nuclear evaluation (SBFT, enteroclysis, abdominal CT, tagged RBC study); eight (34.7%) of these patients had only a possible finding. VCE revealed nine benign lesions, eight of which were resected. Four were hamartomatous polyps.17 Thirteen patients were found to have small intestinal malignancies. Adenocarcinoma was identified in five, carcinoid tumor in six, and GIST in two. Most importantly, 23 of 26 patients underwent surgery, and in 12 of these 23 patients, surgery was regarded as curative, including six patients with malignant small intestinal tumors. None of the patients who underwent a curative resection developed evidence of recurrent disease at a follow-up interval that ranged from 26 to 51 months. Although VCE is a useful technology for detecting small intestinal tumors, it is not without its limitations. As yet, no tissue for definitive diagnosis can be obtained by VCE. Tumors that can be seen on endoscopic push enteroscopy can be missed by VCE because views are often rapidly fleeting.50 Push enteroscopy, however, is limited by its ability to proceed only to about 50 cm beyond the ligament of Treitz.51

Double-Balloon Endoscopy

The development of double-balloon endoscopy (DBE) has allowed evaluation of the entire small intestine, much like VCE, but with the important added ability to biopsy tissue and potentially to perform interventional maneuvers, such as coagulation of bleeding lesions and polypectomy. Yamamoto and colleagues first described DBE in 200152 and elaborated on the clinical experience and outcome using it in 2004.53 In this report, 66 of 123 patients were evaluated for occult gastrointestinal bleeding; 10 patients who presented

A

with occult bleeding had polyps or tumors identified by the procedure.51 Another study of DBE demonstrated similar findings in investigating 64 patients with occult gastroin­ testinal bleeding: 13% had evidence of tumors or polyps.54 Other studies have shown the efficacy and utility of DBE for the investigation of the small intestine.55,56 A case series showed that some small intestinal lesions missed on VCE are seen on DBE, suggesting that the two technologies are complementary in the evaluation and management of patients with small intestinal diseases.57

Computed Tomography

Computed tomography (CT) in conjunction with newer modalities has been used with success to better image the small intestine. A study by Pfannenburg and colleagues showed that anatomic functional image fusion using combined transmission and emission tomography (SPET/CT) had a diagnostic accuracy of 99% in classifying neuroendocrine lesions and improved tumor localization and characterization in patients with neuroendocrine tumors.58 The advent of multidetector CT enteroclysis (MDCT-E) has enabled demonstration of various findings including wall thickening, hypervascularity, and mural and extramural abnormalities.59 MDCT-E requires nasojejunal intubation to enable rapid delivery of an enteral contrast medium to distend the small intestinal loops homogenously; water has been found to be the most useful neutral contrast medium.60 Intravenous contrast also is necessary. MDCT-E has been shown to be a useful technology for detecting small intestinal abnormalities (Fig. 121-6). A prospective study reported by Boudiaf and colleagues in 2004, looked at 107 consecutive patients with suspected small intestinal pathology.61 Results were compared with the results of endoscopy, enteroscopy, VCE, pathology findings, and clinical follow-up. MDCT-E revealed 21 patients with small intestinal mass lesions, nine patients with active Crohn’s disease, two patients with small intestinal tuberculosis, four patients with small intestinal lymphoma com­ plicating celiac disease, and 12 patients with low-grade small intestinal obstruction.58 Sensitivity and specificity of MDCT-E were 100% and 95%, respectively, compared with the appropriate gold standard (endoscopy, enteroscopy, VCE, or pathology).61

B

Figure 121-6.  A, An image from a CT enteroclysis study that revealed abnormal focal wall enhancement and multiple discrete masses in the ileum. B, Surgical resection specimen from the same lesion in A, which was found to be a carcinoid tumor. (A, Courtesy of Ilan Weisberg, MD, New York. B, Courtesy of William R. Jarnagin, MD, New York.)

Chapter 121  Tumors of the Small Intestine tinue to play complementary roles in evaluating the small intestine.

Another study looked MDCT-E in symptomatic patients with suspected small intestinal neoplasm; researchers used endoscopic, pathology, and clinical follow-up findings as reference standards.62 After having normal endoscopy and colonoscopy, 219 patients underwent MDCT-E. Fifty-five patients had positive findings on MDCT-E, for an overall sensitivity and specificity of 84.7% and 96.9%, respectively. Pathologic evaluation revealed 50 patients with small intestinal tumors, including 19 with carcinoid, seven with adenocarcinoma, five with lymphoma, nine with jejunal adenoma, five with GIST, two with ectopic pancreas, two with angiomatous masses, and one with a metastatic lesion. Five patients had false-positive evaluations.62

TREATMENT Benign Neoplasms

Therapy of benign neoplasms of the small intestine usually consists of resection, either via an endoscope or at laparotomy, as determined by the size, growth pattern, and location of the lesion. Most adenomas may be adequately removed at surgery by segmental resection, although if there is any evidence of malignant degeneration, the margins of resection should be extended and draining lymph nodes should be excised. Pedunculated duodenal adenomas may be removed endoscopically by snare cautery, whereas sessile duodenal adenomas are often managed by endoscopic mucosal resection. If an adenoma involves the ampulla, it may be possible to remove it endoscopically or by sphincterotomy with biliary and/or pancreatic stenting. Larger periampullary or ampullary lesions, which are more likely to harbor malignancy, may require pancreaticoduodenectomy (Whipple’s procedure), the decision for which is facilitated by histopatho­logy of frozen sections. Patients must undergo periodic surveillance following removal of an adenoma, especially if it is a villous adenoma, because there is a significant chance of recurrence; however, survival is excellent for these patients.

Endoscopic Ultrasonography

Endoscopic ultrasonography (EUS) is potentially helpful in determining the origin of intestinal tumors and whether they are benign or malignant. To date, however, the role of EUS in the small intestine has mainly been in the evaluation of pancreatic and ampullary tumors,63 although its role is sure to evolve with more experience (Fig. 121-7).

Angiography

Angiography is of limited value except to detect tumors with a significant vascular component and for localizing a site of severe bleeding. A technetium-labeled red blood cell scan (tagged RBC scan) is more sensitive than angiography to determine if a patient has active gastrointestinal bleeding and can be used to localize the lesion, but it cannot definitively diagnose the cause.

Malignant Neoplasms

Surgery remains the only known curative treatment for small intestinal adenocarcinoma. Curative surgical resection is successful in 40% to 65% of patients.6,64 Surgical treatment for adenocarcinoma of the duodenum, because of anatomic location, is a pancreaticoduodenectomy (Whipple’s procedure). In patients with resectable adenocarcinoma of the jejunum, removal of the tumor with wide

Future Prospects

It is likely that more comparison studies will be reported over the next few years to better elucidate the optimal algorithm for the diagnostic evaluation of patients with suspected small intestinal lesions. All of these new technologies—VCE, EUS, DBE, and MDCT-E—are likely to con-

A

B

Figure 121-7.  A, Endoscopic ultrasonography in two patients with a gastrointestinal stromal tumor (GIST). A, A benign GIST appears as a small (<3 cm), homogeneous, hypoechoic lesion arising, most commonly, within and contained within the muscularis propria layer. B, A malignant GIST appears as large (>3 cm), inhomogeneous lesions arising within the muscularis propria and with possible extension into other wall layers and surrounding structures. (A and B, Courtesy of Gregory Ginsberg, MD, Philadelphia, Pa.)

2151

2152

Section X  Small and Large Intestine margins and adjacent lymph nodes is performed. Terminal ileal adenocarcinoma is managed similar to cecal cancer, with a terminal ileal resection, right hemicolectomy, and lymph node resection. Dabaja and colleagues presented a large retrospective review of 217 patients with small intestinal adenocarcinoma in 2004.65 Surgery with curative intent was performed in 146 patients (67%). Adjuvant chemotherapy was administered in 59 patients (27%) and palliative chemotherapy in 48 patients (22%). The median overall survival time was 20 months. The five-year overall survival rate was 26%. Cancer-directed surgery, early-stage disease, and lymph node involvement ratio were significantly associated with overall survival by univariate analysis. Cancer-directed surgery and lymph node involvement ratio were independent predictors of overall survival in a multivariate analysis (adjusted rate ratio, 0.14; 95% CI: 0.04 to 0.46; P = 0.001). In a study reported from British Columbia, 47 patients with small bowel adenocarcinoma were identified between 1990 and 2000; 21 of the 47 were given chemotherapy, with the majority receiving 5-fluorouracil (5-FU)-based therapy. Of 19 patients treated with curative intent, the median overall survival was 38.6 months. Sixteen of 37 patients with advanced disease received chemotherapy. The median overall survival for those who received palliative chemotherapy was 15.6 months, compared with 7.7 months for those who did not.66 There have been no large prospective studies of chemotherapy for adenocarcinoma of the small intestine because this tumor is uncommon. Most of the chemotherapy data, therefore, come from retrospective analyses, case reports, and case series. One such large retrospective review of continuous infusion 5-FU in advanced small intestinal adenocarcinoma showed an overall response rate of 37.5% and a median survival of 13 months.67 In another retrospective study, combination chemotherapy with 5-FU plus cisplatin, carboplatin, or oxaliplatin in advanced small intestinal adenocarcinoma had an overall response rate of 21% and an overall survival of 14 months. Second-line chemotherapy using 5-FU plus irinotecan has shown promise, but use of continuous infusion 5-FU or 5-FU with cisplatin is not effective.68 Other regimens with activity in gastric and colon cancers have been tried with varying benefit, including 5-FU alone or in combination with doxorubicin, mitomycin-C, cis­ platin, and carmustine. A phase II study of 5-FU, doxorubicin, and mitomycin-C in 38 patients with unresectable, biopsy-proven adenocarcinoma of the small intestine showed a response rate of 18.4%, which was no better than that of previous studies of 5-FU-containing regimens.69 In a small case series of three patients who had not responded to 5-FU, irinotecan was used as salvage therapy and appeared to have some efficacy, with marked improvement in symptoms, in two of the three patients.70 Another study reported on six patients with carcinomatosis from small intestinal adenocarcinoma. The authors demonstrated that patients treated with cytoreductive surgery followed by intraperitoneal hyperthermic chemotherapy with mitomycin C given intraoperatively had a median survival of 45.1 months compared with various historical controls of 3.1 months and 12 months.71 A retrospective review of chemotherapy in 113 patients with advanced small bowel adenocarcinoma looked at newer agents compared with 5-FU-based regimens. Fortyfour patients received palliative chemotherapy, with an overall response rate of 36% during a first- or second-line regimen (9% complete responses and 27% partial responses). Palliative chemotherapy predicted for overall survival in a multivariate analysis (hazard ratio, 0.47; P = 0.035).72

A large multicenter prospective study likely will be required to determine the best adjuvant and palliative regimens for treating small intestinal adenocarcinoma. Radiation therapy has no proven role in treatment of small intestinal adenocarcinoma and can result in significant morbidity from radiation enteritis.

PROGNOSIS

Long-term survival generally is poor for patients with adenocarcinoma of the small intestine. Curative surgical resection is the mainstay of treatment and improves the five-year survival from 0% after palliative treatment to 54%. The absence of involved lymph nodes at the time of surgical resection improves the five-year survival further up to about 63%. Anemia at the time of initial presentation dramatically improves survival, likely because it leads to evaluation at an earlier stage.73 Several studies have looked at prognostic indicators for survival with small intestinal adenocarcinoma. Brucher and colleagues looked a series of 94 patients with primary tumors of the small intestine, 62 of which were malignant; 22 of these were adenocarcinomas. They found a five-year overall survival of 45%.74 Univariate analysis showed degree of resection (R status) and advancing tumor stage; the presence of lymph node metastases, distant metastases, carcinomatosis, and vascular invasion were poor prognostic indicators. Dabaja and colleagues also looked at survival in 217 cases of small intestinal adenocarcinoma.65 The median overall survival of these patients was 20 months (95% CI: 16 to 24 months). The five-year overall survival of all patients was 26%. Outcome, as expected, was closely related to stage of disease, with stage IV patients having the worst five-year overall survival (5%) and median survival period of 11 months. The five-year overall survival was significantly shorter for patients with positive lymph nodes compared with patients with negative lymph nodes (32% vs. 52%; median survival, 22 months vs. 78 months; P = 0.0039). The five-year overall survival rate was significantly higher for patients who underwent cancer-directed surgery than for patients who did not, and it was particularly high for those who underwent pancreaticoduodenectomy compared with other types of surgical resection (82 months vs. 29 months; P < 0.0001). Adjuvant chemotherapy did not appear to significantly affect overall survival when given after resection, but chemotherapy for patients with unresectable (stage IV) disease did appear to afford some benefit (12 months vs. 2 months, P < 0.02). Wu and colleagues retrospectively reviewed prognostic factors in 80 patients with primary small intestinal adenocarcinoma; 60 of 80 underwent surgical management, and 45 of 60 had a curative resection. The five-year survival for patients undergoing resection was 27%; overall survival was 17.5%. In multivariate analysis, earlier tumor stage (I and II) and curative resection were two independent factors with a positive impact on survival. Conversely, the presence of lymph node metastases was an independent variable for poor disease-free survival in patients who underwent a curative resection.74

OTHER MALIGNANT SMALL INTESTINAL NEOPLASMS For diagnosis, management, and prognosis of lymphomas, GISTs, and carcinoids please refer to Chapters 29, 30, and 31, respectively.

Chapter 121  Tumors of the Small Intestine Metastatic tumors are more commonly found in the small intestine than are primary tumors. Melanoma is the most common cancer to spread to the small intestine (Fig. 121-8). Patients with metastatic melanoma can present with occult or overt gastrointestinal bleeding. Surgical intervention may be performed as a useful palliation. Other cancers such as breast, lung, colon, stomach, and ovary can directly invade or metastasize to the small intestine (see Fig. 121-9). In these situations, surgery can be useful for palliation, and systemic chemotherapy can help control disease progression. Prognosis remains poor, however.

KEY REFERENCES

Figure 121-8.  Endoscopic image of an ulcerated, pigmented lesion in the jejunum; pathology showed metastatic melanoma. (Courtesy of Mark A. Schattner, MD, and Satish Nagula, MD, New York.)

Figure 121-9.  Film from a contrast examination of the small intestine in a patient with partial small bowel obstruction and extensive extramucosal disease caused by metastatic lung cancer. Contrast has been delivered through a Miller-Abbot tube. (Courtesy of Igor Laufer, MD, Philadelphia, Pa.)

Bailey AA, Debinski HS, Appleyard MN, et al. Diagnosis and outcome of small intestinal tumors found by capsule endoscopy: A threecenter Australian experience. Am J Gastroenterol 2006; 101:2237-43. (Ref 49.) Boudiaf M, Jaff A, Soyer P, et al. Small-intestinal disease: Prospective evaluation of multi-detector row helical CT enteroclysis in 107 consecutive patients. Radiology 2004; 233:338-44. (Ref 61.) Brucher BL, Stein HJ, Roder JD, et al. New aspects of prognostic factors in adenocarcinomas of the small bowel. Hepatogastroenterology 2001; 48:727-32. (Ref 74.) Cobrin GM, Pittman RH, Lewis BS. Increased diagnostic yield of small intestinal tumors with capsule endoscopy. Cancer 2006; 107:22-7. (Ref 43.) Czaykowski P, Hui D. Chemotherapy in small intestinal adenocarcinoma: 10-year experience of the British Columbia cancer agency. Clinical Oncology 2007; 19:143-9. (Ref 66.) Dabaja BS, Suki D, Pro B, et al. Adenocarcinoma of the small bowel: presentation, prognostic factors, and outcome of 217 patients. Cancer 2004; 101:518-26. (Ref 65.) Egberts J-H, Scharrer M-L, Hinz S, et al. Small intestinal cancer: singlecentre results over a period of 12 years. Hepatogastroenterology 2007; 54:129-34. (Ref 3.) Filippone A, Cianci R, Milano A, et al. Obscure gastrointestinal bleeding and small intestinal pathology: comparison between wireless capsule endoscopy and multidetector-row CT enteroclysis. Abdom Imaging 2008; 33:398-406. (Ref 59.) Ginsberg GG, Barkun AN, Bosco JJ, et al. Wireless capsule endoscopy. Gastrointest Endosc 2002; 56:621-4. (Ref 40.) Haselkorn T, Whittemore AS, Lilienfeld DE. Incidence of small intestinal cancer in the United States and worldwide: geographic, temporal, and racial differences. Cancer Causes Control 2005; 16:7817. (Ref 9.) Hatzaras I, Palesty A, Abir F, et al. Small bowel tumors: epidemiologic and clinical characteristics of 1260 cases from the Connecticut tumor registry. Arch Surg 2007; 142:229-35. (Ref 7.) Locher C, Malka D, Boige V, et al. Combination chemotherapy in advanced small intestinal adenocarcinoma. Oncology 2005; 69:290-4. (Ref 68.) Schwartz GD, Barkin JS. Small-intestinal tumors detected by wireless capsule endoscopy. Dig Dis Sci 2007; 52:1026-30. (Ref 45.) Wheeler JM, Warren BF, Mortensen NJ, et al. An insight into the genetic pathway of adenocarcinoma of the small intestine. Gut 2002; 50:21823. (Ref 26.) Yamamoto H, Kita H, Sunada K, et al. Clinical outcomes of double-balloon endoscopy for the diagnosis and treatment of smallintestinal diseases. Clin Gastroenterol Hepatol. 2004; 2:1010-16. (Ref 53.) Full references for this chapter can be found on www.expertconsult.com.

2153

CHAPTE R

122 Colonic Polyps and Polyposis Syndromes Steven H. Itzkowitz and Jonathan Potack

CHAPTER OUTLINE Colonic Polyps  2155 Neoplastic Polyps  2155 Non-neoplastic Polyps  2172 Submucosal Lesions  2176

A gastrointestinal polyp is a discrete mass of tissue that protrudes into the lumen of the bowel. A polyp may be characterized by its gross appearance and overall size, whether or not it has a stalk, and whether it is one of multiple similar masses occurring elsewhere in the gastrointestinal tract. Regardless of these features, however, specific definition rests on histologic characteristics. Because of their protrusion into the bowel lumen and the stresses of the fecal stream to which they are exposed, polyps can cause symptoms: They can ulcerate and bleed; abdominal pain can result when a peristaltic wave propels a polyp downstream, thereby stretching its blood supply and nerve fibers; and large polyps rarely may even obstruct the intestine. Symptoms from polyps are uncommon, however, and the greatest concern with polyps is their silent potential to become malignant. The bulk of evidence supports the hypothesis that most colonic cancers arise within pre­viously benign adenomatous polyps. Only a small percentage of all colonic adenomas progress to carcinoma, however, and because colonic polyps are so common in the industrialized world, universal detection and removal pose practical and economic problems. To manage colonic polyps appropriately, therefore, the physician must understand the differences in pathogenesis and natural history of the distinct pathologic categories of these lesions.

COLONIC POLYPS Colonic polyps may be divided into two major groups: neoplastic (the adenomas and carcinomas) and non-neoplastic (Table 122-1). The adenomas and carcinomas share a characteristic—cellular dysplasia—but they may be subdivided according to the relative contribution of certain microscopic features. The non-neoplastic polyps may be grouped into several distinct categories: hyperplastic polyps (including serrated polyps), “mucosal polyps,” juvenile polyps, PeutzJeghers polyps, inflammatory polyps, and others. Submucosal lesions also can impart a polypoid appearance to the overlying mucosa and therefore are briefly mentioned even though they are not true polyps.

Gastrointestinal Polyposis Syndromes  2176 Inherited Polyposis Syndromes  2176 Noninherited Polyposis Syndromes  2188

NEOPLASTIC POLYPS Pathology Histologic Features Adenomatous polyps are tumors of benign neoplastic epithelium that can either be pedunculated (i.e., attached by a stalk) or sessile (i.e., attached by a broad base with little or no stalk). The neoplastic nature of adenomas is apparent by histologic examination of their glandular architecture. Tubular adenomas are the most common subgroup and are characterized by a complex network of branching adenomatous glands (Fig. 122-1A). In villous adenomas, the adenomatous glands extend straight down from the surface to the center of the polyp, thereby creating long, finger-like projections (see Fig. 122-1B). Tubulovillous (villoglandular) adenomas manifest a combination of these two histologic types. A polyp is assigned a histologic type on the basis of its predominant glandular pattern, and in practice, pure villous adenomas are quite rare. According to the World Health Organization, adenomas are classified as tubular if at least 80% of the glands are of the branching, tubule type and as villous if at least 80% of the glands are villiform.1 Of all adenomatous polyps, tubular adenomas account for 80% to 86%, tubulovillous for 8% to 16%, and villous adenomas for 3% to 16%.2,3 Tubular adenomas usually are small and exhibit mild dysplasia, whereas villous architecture is more often encountered in large adenomas and tends to be associated with more severe degrees of dysplasia (Table 122-2). By definition, all colorectal adenomas are dysplastic. Adenomatous epithelium is characterized by abnormal cellular differentiation and renewal, resulting in hypercellularity of colonic crypts with cells that possess variable amounts of mucin and that are hyperchromatic, with elongated nuclei arranged in a picket-fence pattern. These cytologic alterations confer an increased basophilic appearance to the adenomatous epithelium on conventional hematoxylin-eosin staining. Although the predominant cell type is an immature goblet cell or columnar cell, adenomas can contain other cell types, such as neuroendocrine cells, Paneth cells, squamous morules, and, rarely, melanocytes. On cross section, the inner contour of an adenomatous

2155

2156

Section X  Small and Large Intestine gland lumen usually is smooth, in contrast to the serrated appearance of a hyperplastic gland lumen (see later). The dysplasia exhibited by all adenomas can be graded subjectively on the basis of certain cytologic and architectural features into three categories: mild, moderate, and severe. Some polyps contain the entire spectrum from mild to severe dysplasia, but in all cases, the adenoma is classified according to its most dysplastic focus. In cells that exhibit mild dysplasia, the nuclei in the cell maintain their basal polarity but are hyperchromatic, slightly enlarged, and elongated, yet uniform in size, without prominent nucleoli (Fig. 122-2A). There often is loss of goblet cell mucin. Architecturally, the glands manifest branching and budding and become more crowded. With moderate dysplasia, nuclei become stratified and pleomorphic, with prominent nucleoli, along with further loss of goblet cell mucin and increased glandular crowding. Severe dysplasia (see Fig. 122-2B) is characterized by further stratification and pleomorphism of nuclei, more numerous and prom­ inent nucleoli, increased nucleus-to-cytoplasm ratios, and extreme glandular crowding. With further cell proliferation within the crypt, cells pile up, lose polarity, and create glands within glands, giving a disorderly cribriform appearance termed carcinoma in situ

Table 122-1  Classification of Colorectal Polyps Neoplastic Mucosal Polyps Benign (Adenoma) Tubular adenoma Tubulovillous adenoma Villous adenoma Malignant (Carcinoma) Noninvasive carcinoma Carcinoma in situ Intramucosal carcinoma Invasive carcinoma (through muscularis mucosae) Non-Neoplastic Mucosal Polyps Hyperplastic polyp (including serrated polyps) Mucosal polyp (normal mucosa in a polypoid configuration) Juvenile polyp (retention polyp) Peutz-Jeghers polyp Inflammatory polyp Submucosal Lesions Colitis cystica profunda Pneumatosis cystoides coli Lymphoid polyps (benign and malignant) Lipoma Carcinoid Metastatic neoplasms Other rare lesions

Table 122-2  Histologic Types of Adenomas and Their Features DEGREE OF DYSPLASIA† (%)

SIZE OF ADENOMA* (%) TYPE OF ADENOMA Tubular Tubulovillous Villous

<1 cm

1-2 cm

>2 cm

Mild

Moderate

Severe

77 25 14

20 47 26

4 29 60

88 58 41

8 26 38

4 16 21

*Adapted from Muto T, Bussey HJR, Morson BC. The evolution of cancer of the colon and rectum. Cancer 1975; 36:2251. † Adapted from Konishi F, Morson BC. Pathology of colorectal adenomas: A colonoscopic survey. J Clin Pathol 1989; 35:830.

A

B

Figure 122-1.  Comparison of tubular and villous histology. A, Tubular adenomas consist of branched, crowded glands arranged in a complex cerebriform pattern. B, Villous adenomas consist of glands that are long, finger-like fronds typically projecting from the polyp stroma to the surface without much branching. (Courtesy of Noam Harpaz, MD, PhD, New York.)

Chapter 122  Colonic Polyps and Polyposis Syndromes

A Figure 122-3.  Histopathology of a malignant polyp. This pedunculated adenoma demonstrates tubulovillous histology on the surface. The entire stalk of the polyp (center, arrows) contains numerous malignant glands representing well-differentiated adenocarcinoma. Unlike the adenomatous glands on the surface, many of the invasive glands demonstrate sharply angulated edges. (Courtesy of Noam Harpaz, MD, PhD, New York.)

Figure 122-2.  Comparison of low-grade dysplasia (LGD) and high-grade dysplasia (HGD). A, LGD is characterized by branching crypts lined by cells with long, thin nuclei that begin to stratify, resulting in an increased nucleus-to-cytoplasm ratio and a loss of normal goblet cells. These changes occupy the entire crypt, including the surface epithelium. B, Adenoma with HGD and intramucosal carcinoma. This photomicrograph shows a section of adenoma with crypts demonstrating HGD (right, arrowheads). There is nuclear pleomorphism and abnormal cellular polarity resulting in a cribriform appearance (i.e., glands within glands). A focus of intramucosal carcinoma is seen in the center (arrows), characterized by cells with pleomorphic nuclei and poor gland formation that are located in the lamina propria. (Courtesy Noam Harpaz, MD, PhD, New York.)

rounding the gland. If a focus of neoplastic cells grows beyond the basement membrane and into the lamina propria of the mucosa, the lesion is termed intramucosal carcinoma (see Fig. 122-2B). Both carcinoma in situ and intramucosal carcinoma are noninvasive lesions without metastatic potential, because lymphatics are not present in the colonic mucosa above the level of the muscularis mucosae.4 Because clinical confusion often arises on encountering these two entities, it has been suggested that both carcinoma in situ and intramucosal carcinoma be reported as “noninvasive carcinoma” to avoid unnecessarily aggressive management. Only when a focus of neoplastic cells has spread through the muscularis mucosae is the lesion considered invasive carcinoma (Fig. 122-3). An adenoma that contains a focus of invasive carcinoma commonly is referred to as a malignant polyp (see later). Of all adenomatous polyps, mild dysplasia is found in 70% to 86% moderate dysplasia in 18% to 20%, severe dysplasia (carcinoma in situ) in 5% to 10%, and invasive carcinoma in 5% to 7%.3,5,6 Higher grades of dysplasia are more common in adenomas of larger size and greater villous content,2 and adenomas with severe dysplasia are more likely to contain foci of invasive cancer.

(see Fig. 122-2B). Most pathologists group severe dysplasia and carcinoma in situ, considering them both high-grade dysplasia2; one reason for this grouping is to avoid using the term carcinoma for these lesions because they often can be managed endoscopically rather than surgically (see later). Indeed, it is now common practice to categorize dysplasia in colorectal adenomas into only two grades: low-grade dysplasia, which includes mild and moderate dysplasia, and high-grade dysplasia, which comprises severe dysplasia and carcinoma in situ. Carcinoma in situ is characterized by intracryptal cell proliferation that leaves intact the basement membrane sur-

Adenoma Size Adenomas are categorized into three size groups: less than 1 cm, 1 to 2 cm, and greater than 2 cm.5 Overall, most adenomas are smaller than 1 cm, but the size distribution of adenomas can vary greatly among studies, depending on study design, age of the study population, and location of the adenomas within the colon. Thus, in autopsy series, which describe a presumably asymptomatic population dying of other causes, only 13% to 16% of adenomas are larger than 1 cm,7-9 whereas surgical and colonoscopic series that include symptomatic or higher-risk patients report a higher prevalence (26% to 40%) of adenomas larger than 1 cm.2,3,5 In countries where the prevalence of colon cancer is high, adenomas tend to be larger than in low-

B

2157

2158

Section X  Small and Large Intestine prevalence countries.10,11 Adenoma size increases as a function of age,8,12,13 even in low-prevalence countries,10 and larger adenomas are more common in distal colonic segments.2,5,8 Diminutive Polyps Diminutive polyps measure 5 mm or less in diameter and are commonly encountered during endoscopy. An earlier concept that these lesions were almost always non-neoplastic has been revised based on several flexible sigmoidoscopic and colonoscopic studies in which 30% to 50% of diminutive polyps were found to be adenomatous14-18; despite the frequency of adenomatous change, however, they represent little if any threat of cancer. Earlier studies found that less than 1% of diminutive polyps were villous or contained a focus of severe dysplasia and that they almost never harbored invasive carcinoma.14-16,18 In an analysis of 4381 diminutive polyps, 4.4% contained severe dysplasia or villous components, although still only 0.1% had invasive carcinoma.19 Moreover, in a retrospective study of predominantly asymptomatic people with diminutive adenomas found on flexible sigmoidoscopy, full co­­ lonoscopy identified a synchronous proximal adenoma in only 33% of subjects, and most of the proximal lesions were smaller than 5 mm.20 Likewise, prospective colonoscopic studies confirm only a 24% to 34% prevalence of proximal adenomas in asymptomatic patients with distal diminutive polyps (of all histologic types)21,22; the likelihood of finding proximal adenomas is greater when the distal polyp is larger than 5 mm.21 Diminutive adenomas manifest little, if any, appreciable growth over time.23,24 A population-based study that involved fulgurating small polyps (even those up to 1 cm) without obtaining initial histologic identification

reported that the subsequent risk for colorectal cancer and overall survival was no worse than in the general population.25 Thus, taken together, these observations indicate that diminutive polyps, even when they prove to be adenomas, have little biologic or clinical significance. Nonetheless, the fact that these tiny polyps often are missed by computed tomographic (CT) colonography (or virtual colonoscopy; see later) has provided some fuel to the debate over the safety of leaving these lesions undetected. An important exception to the rule of the innocuous nature of diminutive adenomas is in the setting of hereditary nonpolyposis colorectal cancer (HNPCC), in which even small adenomas can display advanced pathologic features such as villous histology or high-grade dysplasia (see later). Malignant Potential of Adenomatous Polyps The three principal features that correlate with malignant potential for an adenomatous polyp are size, histologic type, and degree of dysplasia (Table 122-3). Although higher rates of malignant transformation are found when the source of the pathologic material is mainly from surgical polypectomies5 rather than colonoscopic polypectomies,6 the malignant potential is correlated directly with larger adenoma size, more villous histology, and higher degrees of dys­ plasia. These three histopathologic criteria usually are interdependent, so it is difficult to assign a primary pre­ malignant role to any one of them. For example, although only 1.3% of all adenomas smaller than 1 cm harbor a cancer (see Table 122-3), if these small lesions have a predominant villous component or contain a focus of severe dysplasia, the cancer rate rises to 10% or 27%, respectively (Table 122-4). A small (<1 cm), tubular, mildly dysplastic

Table 122-3  Malignant Potential of Adenomatous Polyps Removed Surgically or Colonoscopically COLONOSCOPIC POLYPECTOMIES†

SURGICAL POLYPECTOMIES* VARIABLE Adenoma Size <1 cm 1-2 cm >2 cm Histologic Type Tubular Tubulovillous Villous Degree of Dysplasia‡ Mild Moderate Severe

Total No.

No. with Carcinoma (%)

Total No.

No. with Carcinoma (%)

1479 580 430

19 (1.3) 55 (9.5) 198 (46.0)

1661 2738 1387

8 (0.5) 125 (4.6) 150 (10.8)

1880 383 243

90 (4.8) 86 (22.5) 99 (40.7)

3725 1542 519

104 (2.8) 130 (8.4) 49 (9.5)

1734 549 223

99 (5.7) 99 (18.0) 77 (34.5)

NA NA NA

NA NA NA

*Adapted from Muto T, Bussey HJR, Morson BC. The evolution of cancer of the colon and rectum. Cancer 1975; 36:2251. † Adapted from Shinya H, Wolff VI. Morphology, anatomic distribution and cancer potential of colonic polyps. Ann Surg 1979; 190:679. ‡ This category refers to the most extensive degree of dysplasia outside the area of carcinoma but within the polyp. However, by convention, because an adenoma is classified according to the most severe grade of dysplasia, if carcinoma is present, it is considered a malignant polyp regardless of the degree of surrounding dysplasia. NA, not available.

Table 122-4  Relation of Adenoma, Histology, and Degree of Dysplasia to the Incidence of Invasive Carcinoma, by Adenoma Size DEGREE OF DYSPLASIA (% with invasive cancer)

HISTOLOGy (% with invasive cancer) ADENOMA SIZE (CM) <1 1-2 >2

Tubular

Tubulovillous

Villous

Mild

Moderate

Severe

1 10 35

4 7 46

10 10 53

0.3 3 42

2 14 50

27 24 48

Adapted from Muto T, Bussey HJR, Morson BC. The evolution of cancer of the colon and rectum. Cancer 1975; 36:2251.

Chapter 122  Colonic Polyps and Polyposis Syndromes adenoma is highly unlikely to harbor a focus of invasive cancer. Nonetheless, although this type of lesion is innocuous in itself, once removed, it often is considered a marker of a person who is at (low) risk for developing a recurrent adenoma (discussed later). Because adenomas that are larger than 1 cm, have villous architecture, or manifest high-grade dysplasia or carcinoma represent a more biologically hazardous group, the term adenoma with advanced pathology (AAP) often is applied to adenomas that display any of these features. Other Adenoma Variants Flat Adenomas.  A subset of adenomas, termed flat adenomas by Muto and coworkers,26 is receiving increasing attention as a potentially important lesion. Macroscopically, a flat adenoma is either completely flat or slightly raised and can contain a central depression. By definition of the Japanese Society for Cancer of the Colon and Rectum, the diameter of this polyp is more than twice its thickness. Typically smaller than 1 cm in diameter, these lesions can be easily missed at endoscopy. This potential risk has prompted investigators, particularly in Japan, to adapt better methods of detection that involve dye-spraying (chromoendoscopy) to generate a contrast relief-map image of the mucosa, or magnification colonoscopy, for enhanced visualization.27 In studies without such specialized endoscopic techniques, flat adenomas accounted for 8.5% to 12% of all adenomas and could be multiple.28 Prospective studies of Western populations aided by the use of chromoendoscopy found that 6.8% to 36% of all detected adenomas were flat. Compared with lesions that were polypoid, these flat polyps tended to be smaller and to have increased rates of high-grade dysplasia and early cancer.29-31 The largest study to date looked at over 1800 veterans undergoing colonoscopy. The prevalence of flat or nonpolypoid neoplasms was 9.4%. These lesions were 10 times more likely to harbor a carcinoma, although the rate of carcinoma was quite low.32 Indeed, it has been suggested that flat adenomas can have distinct biologic and chromosomal profiles.30,33 In contrast, re-evaluation of adenomas removed during the National Polyp Study found no increased risk of high-grade dysplasia in polyps classified as flat, based on histologic features.34 Future studies might help define whether broader acceptance of advanced endoscopic techniques such as chromoendoscopy or narrow-band imaging by endoscopists in Western countries will result in higher detection rates of flat adenomas, lower colorectal cancer incidence, or both following colonoscopy.35 A hereditary flat adenoma syndrome in four families described by Lynch and colleagues36 subsequently was confirmed to be a variant of familial adenomatous polyposis (FAP) (see later). The natural history of flat adenomas is not known. It is possible that they give rise to typical polypoid adenomas. Alternatively, the facts that residual flat adenoma tissue can be found adjacent to flat carcinomas, that some studies have observed a substantial incidence of high-grade dysplasia in these small lesions, and that flat adenomas have a lower incidence of K-ras mutations compared with polypoid adenomas, suggest that malignant progression from flat adenomas might not necessarily involve a polypoid phase.37 It is possible that flat adenomas are the precursors of the long-recognized, but uncommon, small de novo colon carcinomas.38 Serrated Adenomas.  Serrated adenomas are polyps that share features of both adenomatous and hyperplastic polyps. These lesions are characterized by colonic crypts with a

saw-tooth, serrated configuration resembling that of hyperplastic polyps; because of nuclear atypia, they are considered adenomas. This is discussed further in the section on hyperplastic polyps. Aberrant Crypts.  Investigations of human and carcinogen-treated rat colonic mucosa have disclosed a putative preneoplastic lesion called the aberrant crypt.39 Found within macroscopically normal mucosa, aberrant crypts can occur individually or as small, slightly raised foci. They can be identified in methylene blue-stained whole mounts of colonic mucosa using a low-power lens or with a magnifying endoscope.40 When viewed from above, the lumina of aberrant crypts are elliptical and irregular rather than circular. Aberrant crypt foci have become useful biomarkers in animal studies of colon carcinogenesis and chemoprevention. Human aberrant crypts often are hyperplastic; however, when they are dysplastic, they can represent the earliest detectable preneoplastic lesions. This notion is supported by molecular studies indicating that dysplastic, but not hyperplastic, aberrant crypts manifest mutations in the adenomatous polyposis coli (APC) gene (see later).41 In one study, patients who had aberrant crypt foci in the rectum underwent repeat colonoscopy in one year42; at which time 60% of patients had aberrant crypt foci, but less than half had the same foci re-identified and 50% had new foci; this suggests that the progression of aberrant crypt foci is a dynamic process. In a large study using highmagnification chromoendoscopy to identify aberrant crypt foci in the rectum, the number of aberrant crypt foci increased in a stepwise fashion in normal patients, patients with a flat adenoma, and patients with a flat carcinoma.43 Despite these findings, it remains to be fully elucidated if aberrant crypt foci are useful biomarkers for adenomas.

Pathogenesis

Histogenesis Adenomatous polyps are thought to arise from a failure in a step, or steps, of the normal process of cell proliferation and cell death (apoptosis). The initial aberration appears to arise in a single colonic crypt in which the proliferative compartment, instead of being confined to the crypt base, is expanded throughout the entire crypt. This disturbance results in a unicryptal adenoma. The DNA-synthesizing cells at the surface are not sloughed into the lumen, as normally occurs, and they accumulate in a downward infolding manner, interposing themselves between normal preexisting crypts. New adenomatous glands then are created either by further infolding or by branching. Thus, the unicryptal adenoma is believed to arise from a monoclonal expansion of an abnormal cell, and as the adenoma enlarges, the adenomatous cell population becomes polyclonal. Evidence for this concept comes from studying intestinal tissues from an extremely rare patient with FAP who was an XO/XY mosaic.44 Analysis of Y chromosome expression in the intestinal mucosa of this patient revealed that normal crypts of the small and large intestine and even unicryptal adenomas were monoclonal (either XO or XY), whereas at least 76% of very small microadenomas were polyclonal. Whether the same situation also applies to sporadic adenoma development is not clear at present. Adenoma-Carcinoma Hypothesis It is generally accepted that most, if not all, colon cancers originate within previously benign adenomas. Rarely, colon cancers develop de novo in apparently flat, nonadeno­ matous epithelium although, as noted earlier, even these

2159

2160

Section X  Small and Large Intestine lesions might conceivably arise from preexisting flat adenomas, or serrated polyps. Evidence in support of the adenoma-carcinoma sequence comes from epide­ miologic, clinical, pathologic, and molecular studies. Epidemiologic Evidence.  The prevalence of adenomas within a population, and the prevalence of people with multiple adenomas, geographically parallel the prevalence of colon cancer.10 Indeed, adenoma prevalence increases in migrants from low-risk to high-risk colon cancer regions (see Chapter 123). The prevalence rates for both adenomatous polyps and cancer increase with age, and age distribution curves indicate that the development of adenomas precedes that of carcinomas by 5 to 10 years.5,45 Clinicopathologic Evidence.  The most compelling evidence for the adenoma-carcinoma sequence is the fact that in patients with FAP who have hundreds to thousands of adenomas, the development of colorectal cancer is inevitable. For persons in the general population without an inherited predisposition to colon cancer, perhaps the best evidence that adenomas give rise to carcinomas comes from endoscopic intervention studies. The National Polyp Study (see later) demonstrated that colonoscopic removal of adenomas results in a much lower than expected incidence of subsequent colorectal cancer.46 In addition, screening proctosigmoidoscopy can lower the expected incidence of47 and mortality from48 rectal cancer. Pathology-based studies often describe the presence of remnant adenoma tissue within colon cancers. Conversely, small foci of cancer are extremely rare in normal mucosa but commonly are found in adenomas, particularly in those that are larger, more dysplastic, and more highly composed of villous elements (see Tables 122-3 and 122-4). Furthermore, the site distribution within the colon is similar for large adenomas and colon cancers. In addition, adenomatous polyps are found in one third of surgical specimens that contain a single colon cancer and in more than two thirds of specimens that contain more than one synchronous cancer. Molecular Genetic Evidence.  Molecular genetic studies provide some of the strongest experimental support for the adenoma-carcinoma hypothesis. The progression from adenoma to carcinoma results from an accumulation of molecular genetic alterations involving, among other changes, activation of oncogenes, inactivation of tumor suppressor genes, and participation of stability genes (see Chapter 123).49 The K-ras oncogene commonly undergoes point mutations at particular sites within the gene, thereby endowing it with the ability to transform cells. Only 9% of small adenomas exhibit K-ras gene mutations, compared with 58% of adenomas larger than 1 cm and 47% of colon cancers50; therefore, K-ras activation can act at an intermediate stage in tumorigenesis, perhaps contributing to a poly­ poid growth pattern. The fact that a large number of adenomas and cancers do not have K-ras gene mutations indicates that other genetic events also must play a role. Tumor suppressor genes that normally function to suppress tumor development commonly are inactivated in colorectal neoplasms by mutation or allelic deletion, thereby promoting tumorigenesis. The loss of function of tumor suppressor genes on chromosomes 5q, 18q, and 17p is critical for colorectal tumorigenesis. The APC (adenomatous polyposis coli) gene, which resides on the long arm of chromosome 5, is considered the gatekeeper for the process of colon carcinogenesis.51 Mutation or loss of this gene is believed to be the crucial first step that confers susceptibility to colonic adenomas in patients with FAP as well as in

people with sporadic adenomas. The APC protein plays an important role in colonic epithelial cell homeostasis (see later). Other tumor suppressor genes are located on chromosome 18q, in a region where the DCC (deleted in colon cancer) gene resides. Loss of function of DCC, or other nearby tumor suppressor genes, seems to contribute to later stages of adenoma progression, because allelic deletion at this locus occurs in only 11% to 13% of small tubular or tubulovillous adenomas, but increases to 47% of adenomas with foci of cancer and 73% of frank colon cancers.52 Allelic deletion of chromosome 17p, at the locus that contains the TP53 gene, is the most common region of allelic loss in colorectal cancers. Because adenomas seldom manifest 17p deletion,52 this alteration probably occurs as a late step in the adenoma-carcinoma progression. Perhaps the most compelling molecular evidence that colon carcinomas arise from previous adenomas is that when cancer cells arise in a malignant adenoma, their pattern of molecular alterations is identical to that of the neighboring adenoma cells, but in addition, they have acquired further mutations that are presumably critical for malignant behavior.52 Oncogenes and tumor suppressor genes enhance the adenoma-carcinoma process by directly stimulating cell proliferation and inhibiting cell death; however, stability genes, or caretakers, normally keep genetic alterations to a minimum, and thus, when they are inactivated by mutation or loss, they permit mutations in other target genes to occur at a higher rate.49 Examples of stability genes include the DNA mismatch repair (MMR) and base-excision repair (BER) genes responsible for repairing subtle mistakes that are made during DNA replication. Germline mutations of DNA MMR genes (such as hMLH1, hMSH2, hMSH6) occur in persons with HNPCC, whereas inheritance of a mutated BER gene (e.g., MUTYH, also known as MYH) is responsible for a type of attenuated adenomatous polyposis (see later). Pathways of Colon Carcinogenesis It is useful to consider the process of colon carcinogenesis in two general stages: the formation of the adenoma, termed tumor initiation, and the progression of the adenoma to carcinoma, termed tumor progression (Fig. 122-4). It is believed that most, if not all, adenomas arise from an initial loss of APC gene function, and for that to happen, epithelial cells must lose the function of both APC alleles (two hits). In patients with FAP, one APC allele is inherited in a mutated form (germline mutation) from the affected parent. Adenomas arise when the second, normal copy of the APC gene from the unaffected parent either is lost or mutated (somatic mutation). Because persons with FAP are born with the first hit, they develop polyps at a much younger age and in much greater number than does the general population; thus, FAP can be considered a condition of accelerated tumor initiation. Despite this abnormal initiation rate, once adenomas form in patients with FAP, it is believed that each adenoma tends to display a normal progression to carcinoma. Thus, the inevitable progression to cancer in FAP is more a consequence of the numerous polyps than of any increased premalignant potential of the individual adenoma. In the general population, sporadic adenomas arise as a consequence of two acquired somatic mutations of the APC gene. Because two acquired hits are statistically less likely than one acquired hit, sporadic adenomas tend to occur later in life and to be fewer than the polyps in patients with FAP. Another major molecular pathway for colon carcinogenesis involves mutations in DNA MMR genes (see Chapter 123); this is the predominant pathway in patients with

Chapter 122  Colonic Polyps and Polyposis Syndromes Table 122-5  Prevalence of Adenomatous Polyps in Various Populations

PREVALENCE OF COLON CANCER

POPULATION Hawaiian-Japanese U.S. (New Orleans)   Whites   Blacks Brazil (São Paulo) Japan   Akita   Miyagi Costa Rica (San José) Colombia (Cali)

PREVALENCE OF POLYPS IN MEN (%)

PREVALENCE OF POLYPS IN WOMEN (%)

20-39 yr

40-59 yr

60+ yr

20-39 yr

40-59 yr

60+ yr

Very high

50

69

64

0

71

58

High High Intermediate

0 19 5

39 26 14

47 52 30

0 0 8

10 27 14

35 41 23

Intermediate Low Low Low

21 1 0 2

31 9 6 7

46 23 13 18

0 4 2 2

8 9 4 10

37 17 9 15

Adapted from Correa P. Epidemiology of polyps and cancer. In Morson BC, editor. The pathogenesis of colorectal cancer. Philadelphia: WB Saunders; 1978. p 126.

Sporadic colon carcinoma

FAP

HNPCC

Tumor initiation

Normal

Accelerated

Normal

Tumor progression

Normal

Normal

Accelerated

Normal epithelium APC Adenoma K-ras DCC p53 Carcinoma

Figure 122-4.  Pathways of colon carcinogenesis. Adenomas develop as a consequence of factors involved in initiating tumors, and they progress to carcinomas because of factors that act as tumor promoters. A simplified theory comparing the two hereditary colon cancer syndromes suggests that the adenoma-initiation phase is accelerated in patients with familial adenomatous polyposis (FAP), accounting for numerous polyps. In contrast, the adenoma progression phase is accelerated in hereditary nonpolyposis colorectal cancer (HNPCC) patients, accounting for the often rapid progression of adenomas to carcinoma. APC, adenomatous polyposis coli; DCC, deleted in colon cancer.

HNPCC. Mutations in these enzymes result in a characteristic molecular phenotype termed microsatellite instability (MSI), a phenomenon that is observable in colon cancer cells from approximately 85% of HNPCC colon cancers but in only 15% of sporadic colon cancers. Although its name implies a lack of polyps, HNPCC colon cancers do arise from preexisting adenomas. It is believed that the numbers of adenomas that occur in patients with HNPCC are similar to those in the general population but that HNPCC is marked by an accelerated tumor progression stage, so the few adenomas that do arise often manifest advanced pathology (villous features, high-grade dysplasia) even at small sizes.53 Indeed, adenomas in patients with HNPCC often manifest MSI54 even in their earliest stages of formation. Because these adenomas tend to progress more rapidly to carcinoma,55,56 surveillance intervals for colonoscopy following removal of adenomas in HNPCC patients should be shortened (see later; see also Chapter 123).

Epidemiology and Etiology

Prevalence The prevalence of adenomatous polyps is affected by four major factors: the inherent risk for colon cancer in the population, age, sex, and family history of colorectal cancer. The frequency of colon adenomas varies widely among popula-

tions, but it tends to be higher in populations at greater risk for colon cancer (Table 122-5).10 One illustrative example is to compare the very high adenoma prevalence in Japanese living in Hawaii (a very high risk area for colon cancer) with the much lower adenoma prevalence in Japanese who still reside in Japan, an area of much lower risk. Even within Japan itself, adenoma prevalence correlates quite well with colon cancer prevalence in different prefectures of the country. Data from autopsy series provide an approximation of adenoma prevalence. In populations at low risk for colon cancer, adenoma prevalence rates are lower than 12% (see Table 122-5), whereas in most intermediate- and high-risk populations, adenomas are found in 30% to 40% of the population, and rates as high as 50% to 60% have been observed.7,57 One half to two thirds of people older than 65 years in high-risk areas can harbor colonic adenomas.7,57 The true prevalence rate of adenomatous polyps within an asymptomatic population is only now being elucidated because, until recently, colonoscopy was not performed on healthy persons in the absence of gastrointestinal symptoms. Approximately 27% to 32% of asymptomatic averagerisk persons older than 50 years undergoing screening colonoscopy will have an adenoma, and 6% to 10% will have an AAP (Table 122-6).58-61 By comparison, colonoscopic screening of asymptomatic persons between 40 and 49 years of age revealed prevalence rates of only 8.7% for tubular adenomas and 3.5% for AAP or cancer.62 Colonoscopic series indicate that men have a 1.5 relative risk of adenomas compared with age-matched women,62,63 thus confirming earlier observations in autopsy series.8,11 Men also have slightly higher rates of advanced adenomas than do women, with a relative risk of approximately 1.5.64 A large database study found a higher rate of polyps larger than 9 mm in African Americans undergoing screening colonoscopy compared with age-matched whites.65 The prevalence of adenomas is higher in older people, particularly those older than 60 years. In fact, age is the single most important independent determinant of adenoma prevalence7,11,12,57,62,63,66,67 in high-risk and low-risk regions of the world (see Table 122-5). Not only is advancing age associated with a higher prevalence rate of adenomas, but it also correlates with a greater likelihood for multiple polyps, adenomas with more severe degrees of dysplasia, and, in some studies, larger adenomas. Adenoma prevalence also is higher in persons with a family history of colorectal cancer and adenomas,67,68 particularly if more than one relative is affected with colorectal neoplasia and if the affected relative is young.

2161

2162

Section X  Small and Large Intestine Table 122-6  Prevalence of Adenomas and Cancer in Asymptomatic Persons Older Than 50 Years Screening Method COLONOSCOPY features No. of patients Mean age, yr Male gender (%) Family history of CRC (%) Adenomas (%)   Any   Advanced pathology   High-grade dysplasia   ≥10 mm   Villous Cancer (%)

CT COLONOGRAPHY

FECAL DNA OR FOBT*

Ref 58

Ref 59

Ref 60

Ref 61

3121 62.9 96.8 13.9

1994 59.8 58.9 —

1233 57.8 59 2.6

4404 68.8 44.6 14

31.9 9.6 1.6 4.9 2.9 1

— 6.1 — 2.5 — 0.6

13.6† — — 3.9 — 0.2

27 9.7 0.9 5.2 3.2 0.7

*Patients with positive results were evaluated by colonoscopy. † Includes adenomas ≥6 mm only. CRC, colorectal cancer; FOBT, fecal occult blood test.

Table 122-7  Anatomic Distribution of Colorectal Adenomas Based on Autopsy and Colonoscopy Studies ANATOMIC SITE OF COLORECTAL POLYPS (%) SERIES type Autopsy   All adenomas*,†   Adenomas >1 cm* Colonoscopy   Asymptomatic‡   Symptomatic§

Cecum and Ascending Colon

Transverse Colon

Descending Colon

Sigmoid

Rectum

34 34

26 11

10 14

19 16

10 21

23 8

24 14

24 19

24 47

5 12

*Data from Rickert RR, Auerbach O, Garfinkel L, et al. Adenomatous lesions of the large bowel: An autopsy survey. Cancer 1979; 43:1847-57. † Data from Arminski TC, McLean DW. Incidence and distribution of adenomatous polyps of the colon and rectum based on 1,000 autopsy examinations. Dis Colon Rectum 1964; 7:249-61. ‡ Data from Johnson DA, Gurney MS, Volpe RJ, et al. A prospective study of the prevalence of colonic neoplasms in asymptomatic patients with an age-related risk. Am J Gastroenterol 1990; 85:969. § Data from Konishi F, Morson BCJ: Pathology of colorectal adenomas: A colonoscopic survey. J Clin Pathol 1982; 35:830-41.

Incidence Estimating the incidence of new adenomas requires examining the colon at more than one point in time. Two types of endoscopic studies lend themselves to this analysis: surveillance studies following polypectomy (or following cancer resection) and interval examinations in persons who initially had a negative examination. Of course, for both types of studies, the small but measurable miss rate of adenomas can influence the rate of apparent incident adenomas (see “Detection,” later). For the purposes of this discussion, adenomas found in persons after poly­ pectomy are considered recurrences (see “Postpolypectomy Management”), whereas those found in persons after an initial negative colonoscopy are considered incident adenomas. The incidence of new adenomas varies from 24% to 41%.63 In one study, patients underwent colonoscopy twice on the same day to clear the colon of all potentially missed adenomas, and 38% were found to have new adenomas at colonoscopy two years later.69 Three-year follow-up sigmoidoscopy after an initial negative examination in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial found a 3.1% incidence of all adenomas and 0.8% incidence of advanced adenomas or cancer.70 A large surveillance study looked at findings on follow-up colonoscopy performed within 5.5 years of the index colonoscopy.

Patients who had one or two tubular adenomas less than 10 mm were no more likely to have advanced neoplasia than patients with negative baseline colonoscopies. In average-risk, asymptomatic persons with no adenomas at baseline colonoscopy, repeat colonoscopy within five years detects an adenoma in approximately 16% to 27%71,72 and an AAP in approximately 1% to 2.4%.71-73 Anatomic Distribution The distribution of adenomatous polyps within the colon differs, depending on the method of investigation (Table 122-7). In autopsy series in which the normal distribution is approximated in presumably asymptomatic subjects, adenomas are distributed uniformly throughout the colorectum; this even distribution has been confirmed in colonoscopic investigations of asymptomatic subjects.23,57 Large adenomas in autopsy series have a distal predominance, in the region where most colon cancers arise, thereby supporting the adenoma-carcinoma hypothesis. Likewise, adenomas detected in surgical and colonoscopic studies of symptomatic people also display a left-sided predominance, indicating that distal adenomas are more likely to come to clinical attention. In older persons, particularly those older than 60 years, distribution of adenomas demonstrates a shift toward more proximal colonic locations. This phenomenon, which is based on autopsy11,12 and on colonoscopic

Chapter 122  Colonic Polyps and Polyposis Syndromes studies of symptomatic74 and asymptomatic22,58,59,66,75 subjects, has importance for choosing appropriate colon cancer screening approaches (see Chapter 123). Some data suggest that African Americans have a greater proportion of proximal adenomas compared with whites, especially in persons older than 60 years.65

Risk Factors for Susceptibility to Adenomas

Evidence is mounting to suggest that both heredity and environment contribute to susceptibility to colonic adenomas. Indeed, the interplay between genetic predisposition and environmental factors supports a hypothesis proposed by Hill many years ago concerning adenoma causation, which was based mainly on epidemiologic and histopathologic observations.76 He postulated that genetic susceptibility to colonic adenomas is extremely common throughout the world. For adenomas to form and then progress to cancer, several environmental factors, presumably dietary, must act in concert. One factor would be responsible for the initial development of adenomas, another would enhance the growth of adenomas, and one or more carcinogens or tumor promoters would finally give rise to cancer. Inherited Susceptibility to Adenomatous Polyps There is a strong genetic component to the well-defined hereditary polyposis (FAP) and nonpolyposis (HNPCC) colon cancer syndromes that exhibit a Mendelian pattern of inheritance (see later); however, 95% of common (sporadic) adenomas and carcinomas arise in people who do not have these syndromes. In the past, this observation was interpreted to mean that genetic predisposition played only a minor role in most colonic neoplasms. Epidemiologic studies, however, have revealed a two- to three-fold increased risk for colon cancer in probands who have a first-degree relative affected by colonic cancer or adenoma67; a similarly increased risk also has been found for adenomas in first-degree relatives of people with adenomas.67 Moreover, data from the National Polyp Study indicate that siblings and parents of patients with adenomatous polyps are at an increased risk for colon cancer, particularly when the adenoma proband was younger than 60 years at diagnosis.77 There is even a suggestion that adenomas in patients with a family history of colorectal cancer have faster growth rates.78 Of patients with HNPCC, 11% were found to have at least one adenoma on first colonoscopy; 5% had at least one hyperplastic polyp; and the frequency of adenomas, but not hyperplastic polyps, increased with older age.79 It is now estimated that as much as 10% to 30% of colon cancers are familial, implying the possibility of susceptibility genes that give rise to common colon cancers.72 Despite epidemiologic data suggesting an increased risk in patients with family members who have colon cancer or adenomas, however, it has been difficult to fully elucidate the causative genetic basis. Several genes have been identified that can contribute to common familial risk. These include a germline mutation in the APC gene at codon 1307 (I1307K) that appears to predispose Ashkenazi Jewish populations to colon cancer; mutations of the DNA MMR gene hMSH6, a type I transforming growth factor (TGF)-β receptor allele TbR-I(6A); and polymorphisms of certain genes involved in the metabolism of nutrients and environmental agents, such as methylenetetrahydrofolate reductase and N-acetyltransferase-1 and -2.67,80 A specific mutation in methylenetetrahydrofolate reductase has been found to be protective against colon cancer risk.81 Germline mutations of APC and other genes involved in the APC pathway, such as β-catenin and AXIN1, as well as the DNA mismatch repair genes hMLH1 and hMSH2, have been implicated in the predispo-

sition to multiple adenomas.82 The identification of genes responsible for common susceptibility to colonic adenoma and carcinoma, particularly using polymorphisms of candidate genes, is an area of considerable research.83 Dietary and Lifestyle Risk Factors Although genetic predisposition clearly plays a role in colorectal carcinogenesis, diet and life-style factors also contribute. It is estimated that as much as a third to a half of colon cancer risk and a fourth to a third of distal colon adenoma risk might be avoidable by modification of dietary and life-style habits.84 For the most part, dietary factors that correlate with a predisposition to colon cancer (see Chapter 123) also are associated with a risk for colonic adenomas.84,85 Factors that have each been correlated with an increased adenoma risk include excess dietary fat, excess alcohol intake, obesity, and cigarette smoking.86 Curiously, low calcium intake, despite being associated with increased risk for colon cancer, does not appear to confer risk for adenoma (although as discussed later, calcium supplementation does seem to lower adenoma recurrences). Factors that have shown the most consistent protective effect against adenomas in epidemiologic studies include dietary fiber, plant foods, and carbohydrate. Indeed, analysis of dietary questionnaires from patients in the prospective PLCO Trial found that those with the highest fiber intake had a 27% lower risk of adenomas compared with those who had the lowest fiber intake.87 Other protective measures include increased physical activity, increased intake of calcium, and high intake of folate. An analysis of an asymptomatic, predominantly male veteran population undergoing screening colonoscopy found that smoking and moderate-to-heavy alcohol use increased risk, whereas cereal fiber intake, vitamin D intake, and use of nonsteroidal anti-inflammatory drugs (NSAIDs) decreased risk for advanced colonic neoplasia (AAPs and colon cancer).88 Unfortunately, the rather attractive hypothesis that “proper” diets would reduce colon cancer risk has not been substantiated when tested in prospective, interventional studies.89 Dietary changes maintained over two to four years have failed to significantly reduce recurrent or incident adenomas in several studies that tested reductions in fat with increases in fiber, fruits, and vegetables; combinations of low-fat with wheat bran or beta-carotene supplements, or both; supplements of wheat bran fiber with vitamins C and E; and a complex supplement of calcium, vitamin C, vitamin E, and selenium. Unlike these null studies, four classes of chemopreventive compounds have shown protective effects against colon adenomas or cancers: NSAIDs, calcium, hormone replacement therapy (HRT), and selenium. Of these, NSAIDs including aspirin are the most well-established agents. Greater than 90% of the more than 110 studies of various animal models and more than 35 epidemiologic studies confirm a significant reduction in colorectal adenomas, cancers, and cancer-associated mortality among users of aspirin or NSAIDs.89 For example, the Nurses’ Health Study, a prospective cohort study of 27,077 nurses, found that regular use of aspirin was associated with a lower risk of developing colorectal adenomas.90 Three prospective randomized trials investigating the use of aspirin to prevent colorectal adenomas have shown decreased rates of adenoma recurrence in the study groups compared with placebo. In one trial, which was terminated early because of significant results, 635 patients with a history of curative resection of colorectal cancer and randomized to 325 mg/day of aspirin (or placebo) were found to have a significant reduction in incident adenomas after a

2163

2164

Section X  Small and Large Intestine mean of 12.8 months.91 Another trial in patients with prior colorectal adenomas compared 81 mg/day or 325 mg/day of aspirin with placebo and found that the 81-mg dose reduced the relative risk of developing adenomas and advanced adenomas by 19% and 41%, respectively, with a similar, but not significant, trend found with the higher dose.92 The one-year results of a trial comparing two doses of lysine acetylsalicylate with placebo found a significant reduction in recurrence of adenomas larger than 5 mm in both treatment groups but a greater effect with the higher dose.93 In contrast, in the Physicians’ Health Study, administration of aspirin 325 mg every other day over five years showed no reduction compared with placebo in advanced adenomas or colon cancers.94 A meta-analysis pooled data from four randomized trials of aspirin at doses from 81 mg to 325 mg daily compared with placebo. More than 2900 patients with a history of an adenoma were included. At a median followup of 33 months, the rate of recurrent adenoma was 37% in the placebo group and 33% in the aspirin group. The absolute risk reduction with aspirin use was 6.7%.95 NSAIDs act by inhibiting cyclooxygenase (COX)-1 and COX-2 enzymes, which thereby reduces cellular proliferation, enhances apoptosis, and reduces angiogenesis; other COX-independent effects also are operative. Based on these observations as well as studies showing that selective COX-2 inhibition reduces the number of adenomas in patients with FAP (see later), three large-scale prospective studies of COX-2 inhibitors (celecoxib and rofecoxib) have been undertaken.96 The APPROVe, APC, and PreSap trials randomized more than 5000 patients with a history of adenomas to treatment with celecoxib or rofecoxib or placebo. In all three studies, celecoxib or rofecoxib was superior to placebo in preventing recurrent adenomas and, specifically, advanced adenomas; enthusiasm for these agents was tempered by the finding of increased cardiovascular adverse events in the COX-2 arms. The APPROVe study was terminated early and rofecoxib was withdrawn from the market, but the studies illustrated proof of principle that COX-2 inhibition leads to decreased polyp burden.97 Calcium supplements have been shown in two randomized, placebo-controlled phase III studies to reduce adenoma recurrence by approximately 19% to 34%, with effects noticed even after one year of supplementation.98,99 It has even been suggested that calcium supplements might have a more pronounced effect on advanced adenomas.100 The mechanism for this protective effect likely is multifactorial, because calcium has been shown to decrease proliferation of colonic epithelial cells and to inhibit mucosal injury induced by bile acids and carcinogens in the fecal stream. Calcium can act by neutralizing the mutagenic effects of bile acids on the colonic mucosa or by directly inhibiting epithelial cell proliferation. Data suggest that the benefit of calcium in reducing the adenoma burden is maintained for at least five years after stopping calcium therapy.101 In a study investigating chemoprevention of skin cancer, selenium supplements were associated with a 58% reduction in colon cancer incidence as a secondary end-point.89 This result has prompted additional trials to evaluate the effects of selenium on recurrent formation of adenoma. HRT has been associated in many studies with an approximately 20% reduction in colon cancer risk and a protective effect against colonic adenomas.89 As with NSAIDs, the adverse side effects of HRT often outweigh the beneficial chemopreventive effects of these agents. A randomized, placebocontrolled trial in patients with a history of adenomas found that use of ursodeoxycholic acid at a dose of 8 to 10 mg/kg over three years did not lead to a statistically significant decrease in recurrent adenomas; however, there was a

statistically significant 39% decrease in adenomas with high-grade dysplasia.102 Initial evidence from multiple cohort studies, including the Nurses’ Health Study and Canadian National Breast Cancer Screening study, suggested a 40% lower risk of colorectal adenomas in women with the highest levels of folate intake. A randomized, controlled trial of folate alone or in combination with aspirin, however, did not show a decreased risk of polyps or advanced neoplasia in the folate arms.97 Currently, it is unclear if folate has a true benefit in reducing polyp burden. Difluoromethylornithine (DFMO) suppresses polyamine levels in rectal mucosa likely by modulating ornithine decarboxylase, which can be abnormally expressed in the presence of an APC mutation. DFMO in combination with sulindac has been compared with placebo in patients with a history of an adenoma. At the three-year colonoscopic follow-up, the rate of adenomas was 41% in the placebo group and 12% in the treatment group. The rate of advanced adenoma was 8% in the placebo group and 0.7% in the treatment group.103 DFMO is relatively well tolerated, although ototoxicity is a concern at higher doses. Further work is needed to confirm the efficacy of this agent. Other agents that continue to be examined as chemo­ prevention agents include 3-hydroxy-3-methylglutarylcoenzyme A reductase inhibitors, vitamin E, and mesalamine compounds. Overall, given the need for long-term therapy and the potential adverse events associated with many of these agents, they are likely best studied in patients at high risk for colorectal neoplasia. Conditions Associated with Adenomatous Polyps A variety of clinical circumstances have been associated with adenomatous polyps. Of the conditions discussed here, the predisposition to have or to develop adenomas is strongest for ureterosigmoidostomy, acromegaly, and Streptococcus bovis bacteremia. Patients with any of these three conditions should undergo a thorough colorectal examination and, in the former two conditions, periodic surveillance should be considered (although the frequency of such examinations is not well defined). As for the other conditions, either data are conflicting or the risk is not strong enough to recommend a policy of surveillance. Ureterosigmoidostomy Sites.  Patients who have undergone a urinary diversion procedure with implantation of the ureters into the sigmoid colon are at particularly high risk for developing neoplastic lesions at the ureterosig­ moidostomy sites.104 At least 29% of such patients develop colonic neoplasms, usually close to the stoma, after this procedure. Adenomatous polyps and carcinomas have been found after mean latent periods of 20 and 26 years, respectively. Lesions that resemble juvenile polyps and inflammatory polyps also have been reported at ureterosigmoidostomy sites. It has been suggested that these lesions are produced by the generation of N-nitrosamines from urinary amines in the presence of fecal flora. In view of the extremely high frequency of neoplasia in this setting, these patients should undergo lifelong colonoscopic surveillance, recognizing the long latent period between the implantation of the ureters and the subsequent development of colonic neoplasia. Acromegaly.  Patients with acromegaly have an increased tendency to develop colon cancers and adenomas.105-107 Although these studies involved few subjects, consistently high prevalence rates of 5% to 25% for colon cancer and 14% to 35% for adenomatous polyps have been observed in patients with acromegaly. A meta-analysis found the

Chapter 122  Colonic Polyps and Polyposis Syndromes pooled odds ratio for adenomas was 2.4 and for colorectal cancers was 4.3 in patients with acromegaly compared with controls.108 The risk for colonic neoplasia may be higher in younger acromegalics105 and in those with a family history of colon cancer,106 multiple skin tags (acrochordons),107 or previous colorectal adenomas.109 The mechanism for enhanced colonic neoplasia in this disease is not clear but probably relates to increased growth hormone and/or insulin-like growth factor (IGF)-1 levels. In patents without acromegaly, high serum levels of IGF-1 have been correlated with an increased risk of colorectal cancer.110 In patients with acromegaly, high serum IGF-1 levels have been correlated with increased epithelial cell proliferation111 and increased recurrence rates of colorectal adenomas.109 Other studies in acromegalics, however, have not found that blood levels of growth hormone or IGF-1 correlated with the presence of neoplasms107 and that the risk of neoplasia actually was greater in cured acromegalics than in those with active disease.106 Streptococcus bovis Bacteremia and JC Virus.  Bacteremia and endocarditis caused by S. bovis have been associated with colorectal carcinoma, adenomatous polyps, and even FAP.112,113 The fecal carriage rate of this organism is higher in people with adenomas or carcinomas than in those with benign colon diseases or in normal controls.114 It has been suggested that patients with S. bovis bacteremia undergo thorough colonic examination to exclude a neoplasm. Endocarditis caused by Streptococcus agalactiae (an organism that seldom is pathogenic in adults) has been reported in two patients, each of whom had a rectal villous adenoma with foci of carcinoma.115 JC virus, an oncogenic polyomavirus that blocks tumor suppressor genes, also has been associated with colonic adenomas and carcinomas.114 Cholecystectomy.  In some studies, cholecystectomy has been associated with an increased risk for colon cancer, although this increase is only modest and applies mainly to women and to lesions in the proximal colon.116 It is postulated that in the absence of the gallbladder, there is enhanced delivery of bile acids to the colon and possibly a shift from primary to secondary bile acids that enhances the proliferative activity of the colonic mucosa. In general, however, case-control studies have not found an increased risk for adenomatous polyps among patients who have had cholecystectomy,117 nor was this a risk factor for advanced adenomas or colon cancer among asymptomatic male veterans.88

Clinical Features

Most patients with colonic polyps either have no symptoms referable to the gastrointestinal tract or have nonspecific intestinal symptoms. In persons with symptoms that can be attributed to colonic polyps, the most common presenting symptom is occult or overt rectal bleeding. Histopathologic observations suggest that in contrast to colonic carcinomas, which exhibit considerable surface erosion, the generally less rigid adenomas maintain the integrity of their surface epithelium but can bleed into the polyp stroma.118 These findings help explain the clinical impression that bleeding from polyps is intermittent and usually does not cause fecal occult blood loss or anemia. Other symptoms that have been attributed to colonic polyps are constipation, diarrhea, and flatulence. Constipation or decreased stool caliber is more likely to be caused by bulky lesions in the distal colon. Large colonic polyps may be associated with cramping lower abdominal pain from intermittent intussusception. Unless these widely

prevalent symptoms disappear with the removal of the polyp, they must be attributed to other causes. A syndrome of secretory diarrhea with considerable and sometimes life-threatening water and electrolyte depletion occasionally has been observed in patients with villous adenomas.119 Tumors that produce this syndrome are typically larger than 3 to 4 cm in diameter and are almost always located in the rectum or rectosigmoid, with little surface area distal to the tumor for reabsorption of fluid and electrolytes. In contrast to the absorption of water and sodium and secretion of potassium exhibited by normal colonic mucosa, secretory villous adenomas have a net secretion of water and sodium and an exaggerated secretion of potassium.120

Detection

Colorectal polyps usually are clinically silent. They typically are detected either in asymptomatic people being screened for colorectal neoplasia or incidentally during investigation for symptoms ostensibly referable to the colon or evaluation of unexplained iron deficiency anemia. A more complete discussion of colorectal cancer screening can be found in Chapter 123. This section addresses the issue of adenoma detection using the available screening modalities. Fecal Occult Blood Testing The actual frequency of bleeding from adenomas is difficult to determine. A significant adenoma (i.e., >1 cm or carcinoma in situ) is the cause in less than 10% of people who report frank rectal bleeding.121 In general, polyps smaller than 1 cm do not bleed. This dictum is supported by quantitative measurements of fecal blood loss in people with known adenomas; these measurements indicate that only patients with adenomas larger than 1.5 to 2.0 cm lose more than the usual amounts of blood, regardless of the location of the polyp within the colon.122 Thus, less than 40% of patients with known adenomas show positive fecal occult blood test (FOBT) results, and the higher rates occur primarily in patients with larger and distal polyps.122 When asymptomatic persons older than 40 years of age undergo colon cancer screening with guaiac-based FOBTs, about 1% to 3% will have a positive result.123 Upon colon­ oscopic evaluation, less than half of these people will have a colorectal neoplasm, and among the lesions found, adenomas outnumber carcinomas by 3 : 1. Thus the proportions of all positive guaiac tests attributable to colonic neoplasms (i.e., positive predictive values) are 30% to 35% for adenomas and 8% to 12% for cancer.124 Despite the predominance of adenomas among lesions detected, 75% of adenomas still may be missed by guaiac testing (i.e., false-negative values) unless they are large or located in the distal portion of the colon. Positive test results for occult blood one to two years after a negative search will detect some of these missed polyps.125 Because small polyps seldom bleed and their rate of detection by occult blood testing is low, sigmoidoscopy or colonoscopy has been recommended to complement FOBTs. Fecal Immunochemical Testing Guaiac-based testing relies on a peroxidase reaction; false positives can occur if the patient recently had ingested vegetable peroxidases or rare red meat, and false negatives can occur if they had ingested high doses of antioxidants such as vitamin C. To avoid some of these drawbacks, fecal immunochemical testing (FIT) was developed using an immunoassay to detect human globin in stool. The advantages of FIT are that it has increased specificity for human

2165

2166

Section X  Small and Large Intestine blood over guaiac-based testing, it does not require dietary restrictions, and it can be quantitative. Two types of FIT studies have been performed: comparison studies to FOBT, in which all patients received a co­­ lonoscopy, and in vitro studies to determine at what level FIT can detect blood in various samples. Like the guaiacbased FOBT, FIT detects more cancers than adenomas. Most studies report sensitivities of 30% to 60% for detecting either cancer or advanced neoplasia.126,127 Few studies have looked at the sensitivity for only adenomas. One large study analyzed seven different FIT tests in a screening population of more than 1300 patients. FITs were better than guaiacbased FOBTs.128 The two best-performing assays had sensitivities of 25% to 27% for advanced adenomas. Specificity was high (97% for all adenomas and 93% for advanced adenomas). When quantitative FIT tests were set at detecting 50 ng Hb/mL in stool, the sensitivity for advanced adenomas was 55%, and higher rates of detection correlated with larger polyps and multiple adenomas.129 At present, although FIT is an improvement over guaiac-based FOBT, it remains inadequate for adenoma screening and is best used as a screening test primarily for colorectal cancer. Sigmoidoscopy For several decades, sigmoidoscopy was the mainstay of endoscopic colon cancer screening. Rigid sigmoidoscopy would detect polyps (of all histologic types) in about 7% of asymptomatic persons older than 40 years,130 whereas the flexible sigmoidoscope would find polyps in 10% to 15%, principally because a greater length of bowel could be examined.131 Screening sigmoidoscopy reduced mortality from distal rectosigmoid cancers by as much as 60% to 75%, in several retrospective case-control studies.68 Conclusions regarding mortality reduction from three large-scale prospective studies of screening sigmoidoscopy—the PLCO Study in the United States,70 the United Kingdom Flexible Sigmoidoscopy Screening Trial,132 and a “once-only” sigmoidoscopy trial in Italy133—are awaited. Baseline findings of the U.K. study were that 12% and 0.3% of subjects between ages 55 and 64 years were found to have distal adenomas and distal cancers, respectively.132 The Italian study found an 11% prevalence of distal adenomas in the same age group. In the meantime, the increasing use of colonoscopy in the United States has resulted in a marked reduction in screening sigmoidoscopy examinations. Barium Enema Detecting adenomas by barium enema depends on their size. In the National Polyp Study, the detection rates of adenomas smaller than 6 mm, 6 to 10 mm, and larger than 10 mm were 32%, 53%, and 48%, respectively.134 Common sources of error include inadequate cleansing of the colon, which contributes to the 5% to 10% false-positive rate, and diagnostic difficulty caused by the presence of diverticulosis, redundant bowel, or poor mucosal coating, which results in a 10% false-negative rate. Because of these issues, as well as the fact that barium enema never has been formally tested as a colon cancer screening tool, the use of barium enema for colon cancer screening has all but been abandoned in favor of colonoscopy or CT colonography (see later). Colonoscopy Colonoscopy is preferred to double-contrast barium enema examination for detecting adenomas because it has enhanced diagnostic accuracy as well as therapeutic capability. This diagnostic superiority has been demonstrated in studies of patients with known polyps134,135 as well as in symptom-

atic patients who have negative findings on proctosigmoid­ oscopic and barium enema examinations.136 Colonoscopy has become the preferred colon cancer screening test in many settings.137 Despite its reputation as the gold standard for detecting adenomas, colonoscopy does have some limitations.138 Colonoscopy fails to reach the cecum in up to 10% of cases, it usually requires sedating the patient, and it carries a higher cost than FOBT, FIT, or sigmoidoscopy. Colonoscopy also can miss neoplasms, especially those located at flexures or behind folds. In general, adenomas that are missed tend to be small. Studies using a tandem colonoscopy design demonstrate adenoma miss rates of 0% to 6%, 12% to 13%, and 15% to 27% for adenomas larger than 1 cm, between 6 and 9 mm, and less than 6 mm, respectively.138 CT colonography reveals that colonoscopy can miss 12% to 17% of adenomas larger than 1 cm.138 Given the concern about polyp miss rates, there has been increasing attention to quality measures for colonoscopy. Key measures of high-quality colonoscopy include adequacy of preparation, cecal intubation rate, withdrawal time, and adenoma detection rate. Inadequate preparation contributes to prolonged procedure times, decreased detection of lesions, and the need for repeat colonoscopy before recommended surveillance intervals.139 Colonoscopy is not considered complete unless cecal intubation is accomplished. The majority of screening colonoscopy studies report a cecal intubation rate greater than 95%. Current guidelines suggest that cecal intubation rates should be greater than 90% for all colonoscopies and greater than 95% in screening colonoscopies.139 Most screening colonoscopy studies report adenoma detection rates of 25% to 40%. Men have been consistently found to have a higher burden of adenomas than women. Current guidelines suggest that adenoma detection rates should be at least 15% in women and 25% in men.139 A key factor in adenoma detection rate is colonoscopic withdrawal time. A large study examined the effect of withdrawal time in more than 7800 colonoscopies performed by 12 endoscopists. The adenoma detection rate was 28.3% among endoscopists with a withdrawal time of six minutes or more compared with 11.8% when the withdrawal time was shorter than six minutes. The respective detection rates for advanced adenomas were 6.4% and 2.6%140; slower withdrawal time has been validated by the same investigators in a follow-up study of over 2300 colonoscopies.141 Current recommendations suggest that a withdrawal time of at least six minutes is necessary to maximize detection of adenomas. Continued emphasis on adhering to quality measures and detailed elucidation of the reasons lesions are missed can serve to improve colonoscopy further. Computed Tomography Colonography Also known as virtual colonoscopy, CT colonography involves scanning the colon with a helical or spiral CT scanner to produce both two-dimensional and threedimensional images of the colon and rectum. Patients undergo a standard bowel preparation, and the colon is distended with air or carbon dioxide while images are taken with the patient in the supine and prone positions without sedation. In this rapidly emerging field, a number of studies have compared the performance characteristics of CT colonography with standard optical colonoscopy for detecting polyps.142 Factors affecting detection rates include the polyp prevalence rate in the population being studied, the experience of the radiologist, and technical aspects including bowel preparation techniques, software, and the use of

Chapter 122  Colonic Polyps and Polyposis Syndromes

A

C

B

D

Figure 122-5.  Findings on computed tomographic colonography (CTC) in a man at average risk for colorectal cancer. A, An endoluminal threedimensional CTC and colonoscopy image shows a 33-mm lobulated rectal polyp (arrow) and a 13-mm polyp (arrowhead) near the rectosigmoid junction. B and C, Two-dimensional coronal (B) and sagittal (C) CTC images confirm the presence and soft-tissue composition of the larger polyp (arrows). D, A digital photograph from same-day optical colonoscopy shows the endoscopic capture of the polyp immediately before resection. Pathologic evaluation revealed a large benign tubulovillous adenoma with high-grade dysplasia. The second lesion also had benign tubulovillous histologic characteristics but without high-grade dysplasia. (From Kim DH, Pickhardt PJ, Taylor AJ, et al. CT colonography versus colonoscopy for the detection of advanced neoplasia. N Engl J Med 2007; 357:1403-12, with permission).

single-row or multi-row scanners. In high-prevalence populations that included symptomatic patients, the sen­ sitivity for detecting polyps ranged from 29% to 59% for small polyps, 47% to 82% for medium polyps, and 63% to 92% for large polyps (Fig. 122-5). Studies of cohorts with low polyp prevalences fared less well, with sensitivities ranging from 32% to 58% and specificities of 90% for polyps larger than 6 mm in diameter. In the first large study to involve a pure asymptomatic screening population, CT colonography had a sensitivity of 86% for polyps 5 to 9 mm and 92% for polyps larger than 10 mm in the hands of highly skilled radiologists and with patients ingesting oral contrast prior to the study (fecal tagging) to facilitate distinguishing stool from mucosal abnormalities60; a subsequent meta-analysis found very similar results.143 A large multicenter study described a sensitivity of 90% for polyps larger than 10 mm and 78% for polyps between 6 mm and 9 mm.144 In all studies, the detection of polyps smaller than 5 mm by CT colonography has been consistently low. Several important questions remain unanswered regarding CT colonography. The management of polyps smaller than 10 mm is debated. The sensitivity of CT colonography for polyps smaller than 5 mm is quite low, and, even if polyps of this size are detected, they are often not reported. Various strategies have been proposed for polyps from 6 to 9 mm including referral for polypectomy and CT colonog-

raphy surveillance. There is a small but definite risk of advanced neoplasia in these polyps. Given the paucity of natural history data on polyps of this size, the safety of CT colonography surveillance of these polyps is unclear. Concerns have been raised about radiation exposure in healthy screening populations. It has been estimated that the lifetime risk of cancer in any organ associated with a screening CT at age 50 years is 0.14%.137 Given that CT colonography images the entire abdomen, incidental extracolonic findings are often encountered; the rate is as high as 70%, and up to 11% of these incidental findings are clinically significant.137 These incidental findings can prompt additional testing, cost, and anxiety. Lastly, questions remain about the frequency of CT colonography surveillance intervals. Newer Methods for Detecting Adenomas Based on our knowledge of molecular genetic alterations in colon carcinogenesis, a noninvasive method for detecting altered human DNA in stool has been developed.145 In initial studies of patients with colon cancer, the sensitivity of the test was approximately 65%. In a large multicenter study of average-risk, asymptomatic persons older than 50 years, the sensitivity for colon cancer was 52% and specificity was 95%.61 This study demonstrated that stool DNA analysis was three to four times more sensitive than guaiac-based FOBT for detecting invasive cancers and adenomas with high-grade dysplasia. A second-generation stool DNA assay was found to have a sensitivity of 83% and a specificity of 82% for cancer.146 A similar stool DNA test detected 46% of adenomas larger than 1 cm compared to a detection rate of 10% to 17% with guaiac-based FOBT.147 Currently, stool DNA testing, like all noninvasive stool-based screening tests, has a low rate of adenoma detection. New technology, however, is already detecting 59% of advanced adenomas148 and is expected to improve in the years ahead.

Treatment

Proper management of the patient with adenomatous polyps requires an understanding of the natural history of untreated adenomas, the relationship between multiple adenomas and carcinomas, and the course of patients after treatment (polypectomy). Natural History Little is known about adenoma growth rate, primarily because polyps now are removed readily at endoscopy, thereby interrupting the natural history of their growth. Thus, our limited knowledge about polyp growth rate has been pieced together from two main types of studies: longitudinal follow-up studies on patients with nonresected (i.e., untreated) polyps and studies that compare the age distribution of people with adenomas with that of people with carcinomas. The Untreated Adenoma.  Longitudinal studies of people with untreated adenomas afford the most direct picture of the natural history of adenomas. In general, however, studies of this kind have been retrospective and either involved few patients or suffered from a lack of histologic confirmation of the index polyp. Despite these drawbacks, it appears that the adenoma-to-carcinoma progression is rather slow, requiring several years to unfold. Muto and associates reported that in 14 persons with nonresected polyps, it took at least five years and often more than 10 years for histologically proven adenomas to progress to cancer.5 The size of the index polyp affects the interval to carcinoma, because larger adenomas are more likely than

2167

2168

Section X  Small and Large Intestine smaller ones to develop or already contain a focus of cancer. But even starting with a 1-cm polypoid tumor of unknown histology, serial barium enema measurements have suggested that it can take two to five years for cancer to develop149 and that the cumulative risk of cancer at the polyp site is 2.5% at five years, 8% at 10 years, and 24% at 20 years after diagnosis.150 Other radiologic studies indicate that in adenomas with growth rates that are as rapid as those of cancer, doubling times are still longer than four to six months.151 Smaller polyps are likely to require even more time to progress to cancer, and even after several years, many adenomas do not enlarge. For example, in a study involving 213 asymptomatic people with rectal polyps of unknown histology ranging in size from 0.2 to 1.5 cm, serial rigid sigmoidoscopies over three to five years detected only two cancers, and only 4% of the polyps grew larger; the other 96% of polyps remained unchanged, got smaller, or disappeared.152 Also, over a three-year period, adenomas smaller than 1 cm did not significantly change size, and those that were 5 to 9 mm actually showed slight regression.153 In another endoscopic study, histologically proven diminutive adenomas were left in place for two years, after which time only half of them enlarged, but none grew to more than 0.5 cm or developed severe dysplasia or cancer.23 By contrast, other investigators reported that of 30 rectosigmoid polyps 3 to 9 mm that were left in place but measured every six months for two years, none regressed in size, although only three showed a fast growth rate (2 to 4 mm per year).154 A mathematical model, using assumptions based on doubling-time calculations from serial barium enemas, predicts that a diminutive polyp (<0.5 cm) requires two to three years to reach 1-cm size.155 A computational analysis of data from the National Polyp Study, where initial adenoma rates were high but incidence rates were low, also supports the notion that adenomas can regress.156 Despite evidence of polyp regression or slow growth rates, from a clinical management standpoint, any polyp that is detected should be removed. Age Distribution Studies.  Additional, albeit indirect, support of the rather slow growth of adenomas comes from studies that have compared the mean age of people with adenomas with that of people with carcinomas. For instance, studies from St. Mark’s Hospital in London and from the National Polyp Study in the United States indicate that the mean age of people with a single adenoma is about four to five years younger than those with a colon cancer.5,157 A similar analysis in FAP patients has shown that patients with adenomas are about 12 years younger than those with colon cancer.5 The mean age of male veterans with AAP or colon cancer found by screening colonoscopy was 65.1 years, compared with 62.7 years for those without any

polyps.88 Kozuka and colleagues estimated that the transition period for adenomas with mild dysplasia to cancer is eight years, whereas the time for adenomas with severe dysplasia to become malignant is 3.6 years.45 Finally, Eide calculated that, over a 10-year period, there is only a 2.5% risk that an adenoma-bearing person will develop colon cancer, but this risk would be greater if the adenoma were large or villous.158 Multiple Adenomas and Carcinomas.  For proper management of patients and design of cancer screening and surveillance programs, it is important to know the frequency with which adenomas coexist with other adenomas or carcinomas. The term multiple adenomas (or carcinoma) simply means two or more neoplasms and should not be confused with the multiple adenomatous polyposis syndromes that are characterized by tens to hundreds of polyps (see later). An adenoma or carcinoma that is diagnosed at the same time as an index colorectal neoplasm is called a synchronous lesion, whereas one that is diagnosed at least six months later (a somewhat arbitrary limit) is considered metachronous. The adenomatous polyp itself commonly is regarded as a marker of a neoplasm-prone colon. Indeed, 30% to 50% of colons with one adenoma contain at least one other synchronous adenoma, especially in the older age groups.5,12,16 The risks of colon cancer and of high-grade dysplasia both rise with the number of adenomas present (Table 122-8) and approach 100% in people with FAP. A synchronous adenoma can be found in 30% of colons that harbor a carcinoma5,159-161 and in 50% to 85% of those that harbor two or more synchronous cancers.5,159,162 If the synchronous adenoma is diagnosed preoperatively and is distant from the carcinoma, the surgical approach might have to be adapted to the particular circumstances.159 For this reason, before surgical resection of any colorectal carcinoma, a thorough examination of the colon by pre­ operative colonoscopy or CT colonography is strongly recommended. Moreover, the presence of a synchronous adenoma in a patient with colon cancer increases the risk for developing subsequent colon cancer.5,160 Likewise, a synchronous adenoma in a patient with a colonic polyp places that person at greater risk for developing metachronous polyps163 and cancer.163,164 Initial Treatment If a polyp is detected by barium enema or CT colonography, a colonoscopy is recommended to afford the simultaneous opportunities to remove the polyp and search for synchronous neoplasms. There is some debate as to whether rectosigmoid adenomas found at sigmoidoscopy are markers for proximal colonic neoplasia and thereby prompt the need for a full colonoscopy. This controversy applies particularly to

Table 122-8  Correlation between the Number of Adenomas per Patient and Associated Carcinomas or High-Grade Dysplasia NO. OF ADENOMAS PER PATIENT 1 2 3 4 5 ≥6

ST. MARK’S HOSPITAL5

NATIONAL POLYP STUDY2

No. of Patients

% of Patients with Carcinoma

No. of Patients

1331 296 83 40 13 25

30 52 57 50 77 80

1093 430 166 83 40 55

% of Patients with High-Grade Dysplasia 7 10 19 17 20 20

Chapter 122  Colonic Polyps and Polyposis Syndromes Table 122-9  Frequency of Advanced Proximal Neoplasia Related to Findings in the Distal Colon Frequency of Advanced Proximal Neoplasia REF 58* FINDINGS IN DISTAL COLON No distal polyps Distal hyperplastic polyp Distal adenoma Distal advanced neoplasm   Tubular adenoma >1 cm   Villous features   High-grade dysplasia   Invasive cancer

Percent

Odds Ratio (95% CI)

2.7 2.8 6.8

1.0 1.1 (0.6-2.1) 2.6 (1.7-4.1)

8.6 12.5 11.4 25.0

3.2 (1.5-6.8) 4.7 (2.1-10.4) 4.5 (1.5-13.4) 9.8 (3.6-26.4)

REF 75†

REF 59* Percent

Relative Risk (95% CI)

1.5 4.0 7.1 11.5 — — — —

1.0 2.6 (1.1-5.9) 4.0 (1.9-8.3) 6.7 (3.2-16.6) — — — —

Percent

Odds Ratio (95% CI)

5.3 — 5.0

1.0 — 1.26 (0.81-1.98)

4.5 12.1 — —

1.66 (1.10-2.52) 2.46 (1.60-3.77) — —

*Screening colonoscopy study. † Screening sigmoidoscopy followed by colonoscopy study. CI, confidence interval.

patients who have only a single small tubular adenoma with low-grade dysplasia. Some studies indicate that in this subset of patients, colonoscopy does not discover a substantial number of proximal AAPs or cancers,165-167 nor is there a subsequent risk of developing proximal cancer among those who do not undergo colonoscopy.168,169 In contrast, two screening colonoscopy studies suggest that the odds of finding advanced proximal neoplasia, even with a single small tubular adenoma, are 2.6-fold to 4-fold that of finding no distal pathology (Table 122-9).58,75 Some authors have estimated that by not doing colonoscopy for a distal nonadvanced adenoma, 36% of advanced proximal neoplasms would be missed.170 Furthermore, 52% of patients with advanced proximal neoplasia have no distal adenoma,58 and 70% of proximal colon cancers lack a distal marker lesion.171 These observations support the use of colonoscopy as a primary screening modality. Because risk factors for finding advanced proximal neoplasia are increasing age,58,59,75 a positive family history of colorectal cancer,58,75 and male gender,59 performing colonoscopy in these persons on the basis of finding a distal small tubular adenoma would seem prudent. A risk index that stratified patients into low, intermediate, or high risk based on the type and size of distal colonic polyps as well as patient age and male sex was able to predict the risk of proximal advanced colonic neoplasia.172 There is little debate that if sigmoidoscopy reveals more than one adenoma or a distal AAP, a full colonoscopy is warranted because of the higher likelihood of finding synchronous proximal advanced neoplasms (see Table 122-9). Because negative biopsy results from a fractional sample of a polyp cannot possibly rule out cancer, total excision of a polyp is the only method of providing a thorough and accurate histologic diagnosis. For larger polyps, this can require piecemeal excision; for sessile growths, injection of saline into the polyp base can assist with complete endoscopic resection. After apparent complete removal of a large adenoma, it is advisable to repeat the colonoscopy in three to six months to document the completeness of the excision. If a polyp cannot be completely excised after two or three endoscopic sessions, surgical therapy is recommended. Based on reasons listed in the previous discussion, screening colonoscopy is gaining acceptance in the United States as a primary screening tool, although many other countries with more limited resources do not yet subscribe to this philosophy. Two large-scale European trials are

under way to evaluate the use of selective colonoscopy in patients based on the findings of a single sigmoidoscopy performed at about age 60 years.132,133 After screening sigmoidoscopy, 5% to 8% of patients were referred for co­­ lonoscopy based on the finding of high-risk lesions or multiple polyps. Of these patients, 15% to 18% had proximal adenomas, 5% had advanced proximal adenomas, and fewer than 1% had cancer. Follow-up data regarding the incidence and mortality of colorectal cancer in these patients, as well as the results of the PLCO Cancer Screening Trial,70 will provide additional data from prospective trials on the utility of screening sigmoidoscopy. Management of the Malignant Polyp The term malignant polyp refers to an adenoma in which a focus of carcinoma has invaded beyond the muscularis mucosae into the submucosa (see earlier). This term is not applied to adenomas containing either carcinoma in situ or intramucosal carcinoma because these lesions are not invasive and carry no metastatic potential. Rarely, polyps consist entirely of carcinoma. These polypoid carcinomas usually are considered a subset of malignant polyps, and they most likely represent a previous adenoma that has been completely replaced by carcinoma. Sometimes islands of benign adenomatous epithelium are found beneath the muscularis mucosae, and care must be taken not to mistake such pseudocarcinomatous invasion for true invasive carcinoma; this important distinction may be particularly difficult in the rare instance when the ectopic benign epithelium exhibits features of high-grade dysplasia. Ectopic benign epithelium is seen more often in larger pedunculated polyps, especially in the sigmoid colon. Because the distinction between carcinoma in situ and invasive carcinoma influence both management and prognosis, it is crucial that tissue be properly oriented for pathologic examination and that close communication takes place among endoscopist, surgeon, and pathologist. Endoscopic removal of the vast majority of colorectal polyps raises two central and related questions: Is endoscopic polypectomy alone adequate therapy for the malignant polyp? If not, what features of the polyp can predict the presence of residual disease or subsequent recurrence? The answers are vital because they determine the decision for subsequent surgical resection of bowel. Complete endoscopic removal of an adenoma with noninvasive carcinoma is curative. The decision regarding

2169

2170

Section X  Small and Large Intestine therapy becomes much more difficult when polyps contain invasive carcinoma. Although most of these lesions are treated adequately by endoscopic polypectomy, approximately 10% of patients experience an adverse outcome,173 defined as residual cancer in the bowel wall or in lymph nodes either at the time of polypectomy or on follow-up examination. This rate of failure is comparable to that of the precolonoscopy era when adenomas were larger and patients underwent more complete surgical resections.173 Because malignant polyps account for only 5% of all adenomas, and because not all patients who have had colonoscopic removal of malignant polyps have had surgical resection or are available for follow-up, conclusions often are based on small numbers. Notwithstanding these limitations, combined experience has identified certain favorable and unfavorable histopathologic features of a colonoscopically resected malignant polyp that can be used to classify a patient as being at low or high risk for an adverse outcome (see later) (Table 122-10). If none of these unfavorable risk factors is found, the patient is considered to have been cured by the endoscopic polypectomy. This principle applies even to endo-

Table 122-10  Favorable and Unfavorable Features for Adverse Outcomes in Patients with Malignant Polyps Feature

FAVORABLE

UNFAVORABLE

Degree of differentiation Invasion of veins or lymphatics Polypectomy margin

Well or moderate Absent

Poor Present

Clear or >2-mm margin Absent

Involved

Invasion of submucosa of bowel wall

Present

scopically removed polypoid carcinomas, which have been associated with a surprisingly good outcome when not complicated by any unfavorable histopathologic features. If one or more unfavorable features is found in a malignant polyp, the chance of an adverse outcome rises to about 10% to 25%.174 In such cases, surgical resection usually should be performed, taking into account the risk of operative mortality in elderly patients with comorbid illnesses. Several aspects of defining these risk factors demand close collaboration between endoscopists and pathologists. First, some malignant polyps contain only a small focus of poorly differentiated carcinoma, and therefore meticulous pathologic analysis is essential. Second, identification of vessel invasion by cancer cells in malignant polyps may be difficult, in which case special stains of vascular endothelium can be used for clarification. Third, although the polypectomy margin may be found on microscopy to contain cancer cells, some studies suggest that if the endoscopist believes that a complete excision was achieved, surgical resection might not be necessary because the electrocautery might have effectively destroyed residual tumor in the bowel wall.175 Finally, when judging the adequacy of endoscopic polypectomy, the issue of submucosal invasion is important. A pedunculated polyp differs anatomically from a sessile polyp in that the submucosa of the former projects up into the stalk, whereas the submucosa of the latter is in direct continuity with the bowel wall proper (Fig. 122-6). If cancer in a pedunculated polyp is confined to the submucosa of the stalk and all other histologic features are favorable, surgery is not indicated because the chance of an adverse outcome from endoscopic polypectomy is less than the operative mortality. Once the submucosa of the bowel wall is involved with cancer (a situation that occurs more readily in sessile polyps), the chance of an adverse outcome often outweighs the operative mortality, thereby justifying surgical resection. Furthermore, because endoscopic technique purposely avoids cutting deep into the bowel wall submu-

Carcinoma

2

1

Adenoma Carcinoma

Figure 122-6.  Carcinoma in situ versus invasive cancer. Carcinoma is considered intramucosal or carcinoma in situ, as indicated by the area labeled “1,” either in a pedunculated adenoma (left) or in a sessile adenoma (right). This lesion, as a rule, does not metastasize. Carcinoma in an adeno­ matous polyp is considered invasive when it breaches the muscularis mucosae, as indicated by the areas labeled “2.” Invasive cancer in a pedunculated polyp (left) is unlikely to metastasize, and it is managed differently from invasive cancer in a sessile polyp (right), which often requires surgical resection.

Epithelium Muscularis mucosae Submucosa Muscularis propria

2

1

Chapter 122  Colonic Polyps and Polyposis Syndromes

Postpolypectomy Management Polyp Recurrence Rates.  Although patients in whom a colorectal adenoma has been excised completely are likely to develop subsequent (metachronous) neoplasms, the frequency and time course of these occurrences are not well understood. In long-term retrospective studies, the cumulative risk of adenoma recurrence is nearly linear, being 20% at five years after polypectomy and rising to 50% after 15 years (Fig. 122-7).177 Information on this subject is inexact because earlier studies were primarily retrospective, involved short follow-up periods, and differed in endoscopic indications. Moreover, recurrence rates tend to be somewhat inflated because lesions missed during the index colonoscopy might be considered recurrences. With these caveats in mind, it is estimated that one third of people who have undergone polypectomy develop recurrent adenomas.178-180 The recurrence rate at one year is as low as 5% to 15%177,180 but, more realistically, it ranges from 30% to 45% based on prospective colonoscopy studies.178,181-184 In the National Polyp Study, the overall adenoma recurrence rate was 42% for patients who underwent surveillance co­­ lonoscopy at one and three years after index polypectomy compared with 32% in the group that was examined only at three years.46 There is general consensus that recurrent adenomas typically are smaller and less likely to harbor advanced pathology than are index adenomas. Can histopathologic features of adenomas at the time of the index polypectomy be used to help predict recurrence of adenomas? Virtually all studies agree that the presence of multiple index adenomas is an important predictor of subsequent recurrence of adenoma (and carcinoma) (see Fig. 122-7).177,178,180,185,186 This dictum applies despite negative findings on colonoscopy one year after polypectomy.181 Some studies suggest that polyp size greater than 1 cm,179,180,186 severe dysplasia,187 villous histology,180,185 and older age178,180 also are risk factors for adenoma recur-

90 80

>1 adenoma removed

70 Cumulative risk (%)

cosa, there are few published examples of sessile lesions that have been completely excised endoscopically with clear margins and no other unfavorable features. Deciding on the optimal plan of management after poly­ pectomy involves weighing the risks of morbidity and mortality from potential residual or recurrent cancer against the risks of morbidity and mortality from a surgical attempt to cure any such residual disease or lymph node metastasis. A few general recommendations, however, can be offered. If excision of an adenoma is complete, endoscopic polypectomy alone is adequate therapy for adenomas that contain carcinoma in situ, pedunculated adenomas that harbor welldifferentiated or moderately differentiated invasive carcinoma, and polypoid carcinomas. Resectional surgery is indicated for malignant polyps in which the invasive carcinoma is poorly differentiated, involves endothelium-lined channels (lymphatics, blood vessels), extends to or within 2 mm of the polypectomy margin, or involves the submucosa of the colonic wall (including all sessile adenomas). Studies have examined the depth of submucosal invasion as a predictor of lymph node metastasis. Two studies found the rate of lymph node metastasis was zero when depth of submucosal invasion was less than 1 mm in one study and 2 mm in the other.176 Clearly, the ultimate plan of therapy must be individualized according to each patient’s medical condition. For most patients with malignant polyps, polypectomy without sur­ gical resection seems adequate, with the caveat that the polyp was deemed to be completely resected endoscopically, and that postpolypectomy endoscopic surveillance will be incor­porated into the patient’s health care regimen.

60

Recurrent adenomas

50 40 30

1 adenoma removed

20

>1 adenoma removed

10

Cancer 1 adenoma removed

0 0

1 2 3

4 5

6 7

8

9 10 11 12 13 14 15

Years of observation Figure 122-7.  Risk of recurrent adenomas and of colon cancers after removal of adenomas. The risk of developing recurrent adenomas (orange, yellow) or cancer (blue, green) is higher in patients who initially underwent removal of multiple (>1) adenomas (orange, blue) than removal of a single adenoma (yellow, green). (Modified from Morson BC, Bussey HJR. Magnitude of risk for cancer in patients with colorectal adenomas. Br J Surg 1985; 72[Suppl]:S23.)

rence, but the independent relative importance of each of these factors is uncertain. It can be argued that the most clinically important recurrence is an AAP. Some studies have reported a 6.3% to 7.0% recurrence rate for AAP over a four-year follow-up period.183 In the National Polyp Study, the recurrence rate of advanced adenomas was 3.3%, regardless of whether patients underwent colonoscopy at one and three years, or only three years, after polypectomy. The cumulative incidence of AAP was 4% at three years and 8% at six years of follow-up. Independent predictors for AAP at follow-up included more than three adenomas at the index colonoscopy, and age older than 60 years at the time of adenoma diagnosis, with a parent who had colorectal cancer.188 In the presence of these two risk factors, cumulative AAP recurrence rates rose to 10% at three years and 20% at six years of follow-up. The persons at lowest risk for developing AAP were those with a single adenoma diagnosed when they were younger than 60 years. Another colonoscopic followup study reported that multiple adenomas at the index colonoscopy increased the risk of AAP, but it also found that adenoma size greater than 1 cm and proximal location were additional risk factors.186 A pooled analysis of eight prospective trials with more than 9000 patients found that more than five adenomas, size greater than 20 mm, proximal location, and the presence of villous features were the strongest predictors of recurrent advanced adenomas.189 Interestingly, the presence of high-grade dysplasia was not associated with recurrent adenomas. Effect of Polypectomy on Colorectal Cancer Incidence.  If adenomas are the precursor to colon cancer, then removing them should decrease the subsequent incidence of colon cancer. Indeed, both uncontrolled and case-control series strongly suggest that distal colorectal cancers can be prevented and mortality reduced by screening proctosigmoid-

2171

2172

Section X  Small and Large Intestine Table 122-11  Surveillance Recommendations after Colonoscopic Polypectomy FINDINGS ON COLONOSCOPY

NEXT COLONOSCOPY

COMMENT

Small rectal hyperplastic polyps One or two small (<1 cm) tubular adenomas with only LGD (low-risk adenoma)

10 years 5-10 years

3-10 adenomas, any adenoma ≥1 cm, or any adenoma with villous features or HGD >10 adenomas at one examination Suggestive of HNPCC* Piecemeal removal of a sessile adenoma

3 years

Considered equivalent to normal colonoscopy Specific follow-up interval is determined by prior colonoscopic findings, family history, patient preference, and clinical judgment If the 3-year examination is normal, or shows only a low-risk adenoma, repeat again in 5 years Use clinical judgment; consider familial syndrome Consider more frequent intervals Follow-up interval based on clinical judgment

<3 years 2-3 years 2-6 months

*Features suggestive of HNPCC include right-sided cancers that are poorly differentiated and mucus-producing tumors with intense lymphocytic infiltrates; tumors demonstrate microsatellite instability. HGD, high-grade dysplasia; HNPCC, hereditary nonpolyposis colorectal cancer; LGD, low-grade dysplasia. Modified from Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: A consensus update by the US Multi-Society Task Force on Colorectal Cancer and the American Cancer Society. CA Cancer J Clin 2006; 56:143-59.

oscopy.47,48 Other studies found that removing adenomas by polypectomy was associated with an increased incidence of colon cancer.190-192 Because these retrospective studies did not establish a polyp-free colon (or always consider adenoma size and histology), however, the higher colon cancer rate might have reflected malignant progression of other adenomas that had not been removed; two studies from the Mayo Clinic also did not establish a polyp-free colon at the time of polypectomy.25,164 Despite these observations, if the removed polyps were small (<1 cm), there was no greater risk for subsequently developing colon cancer; in contrast, removal of large adenomas was associated with a greater risk of subsequent colorectal cancer, again supporting the concept that advanced neoplasms at index polypectomy are predictors of subsequent neoplasia. Important observations also come from the St. Marks Hospital study in which rectal adenomas were removed without any subsequent examinations.168 If the index polyp was a small, tubular adenoma, there was no greater risk of subsequent cancer anywhere in the colon, whereas the risk was significantly increased if the index adenoma was large or contained villous elements. Other studies have shown a marked reduction in colon cancer incidence related to polypectomy.193 The most valid study design to address this issue is a prospective colonoscopic study in which an adenoma-free colon is established and patients are followed for subsequent development of colonic neoplasms. The National Polyp Study is a landmark study in this regard. This was a prospective, multicenter, randomized trial in which a cohort of 1418 patients underwent a clearing colonoscopy to remove one or more adenomas; patients were then followed at specific intervals for a mean of 5.9 years. During the follow-up period, five early asymptomatic cancers were detected, representing only 10% to 24% of the expected incidence compared with three historical reference groups.46 Two other studies (from Norway and Italy) confirm that colonoscopic polypectomy was associated with a 75% to 80% reduction in subsequent colorectal cancer incidence.194,195 In a smaller study, the Funen Adenoma Followup Study, it was found that although the incidence of subsequent colon cancers was not reduced by polypectomy, the mortality was reduced.196 In summary, these studies indicate that the adenoma is a marker of a neoplasm-prone colon and that colonoscopic clearing of all adenomas is of considerable benefit; however, the data are heterogeneous in describing both the level and duration of protection following a clearing colonoscopy. Furthermore, the number, size, and location of removed polyps plays in a role in determining the level of protection afforded by a clearing colonoscopy. The uneven benefit in preventing subsequent

colorectal cancers afforded by colonoscopy is a major impetus in the drive to enhance the quality of colonoscopic examinations. Frequency of Surveillance Colonoscopy.  The National Polyp Study has taught us that performing follow-up co­­ lonoscopy at one year is a low-yield proposition, as others have observed using sigmoidoscopy.193 It also is worth realizing that an endoscopic examination of the colorectum can afford protection against colon cancer for 6 to 10 years.47,48 Moreover, it has been proposed that a single lifetime sigmoidoscopy at about age 55 to 60 years could be standard policy for colon cancer screening, with referral of patients for colonoscopy and subsequent surveillance only if the sigmoidoscopy discloses a distal adenoma with advanced features.131 A consensus statement for postpolypectomy surveillance guidelines has been generated by the U.S. Multi-Society Task Force on Colorectal Cancer and the American Cancer Society.197 (Table 122-11). A complete colonoscopy should be performed at the time of polypectomy, clearing the colon of all existing adenomas; this can take more than one session for large or multiple polyps. The interval before the next surveillance colonoscopy is based on the patient’s category of recurrent adenoma risk and features of the adenomatous polyps.

NON-NEOPLASTIC POLYPS

Pathologically, whereas neoplastic polyps are part of an identifiable spectrum, non-neoplastic polyps fall into several distinct and unrelated groups, including hyperplastic polyps, mucosal polyps, juvenile polyps, Peutz-Jeghers polyps, inflammatory polyps, and many other submucosal lesions (see Table 122-1).

Hyperplastic Polyps and Serrated Polyps

The most common non-neoplastic polyp in the colon is the hyperplastic polyp, referred to in older literature by some pathologists as “metaplastic polyps.” Hyperplastic polyps usually are small; their average size is less than 5 mm and they seldom are larger than 10 mm. When larger hyperplastic polyps are encountered, this should raise concern that the lesion is part of the serrated polyp pathway (see below). Histologic Features Hyperplastic polyps typically are small sessile lesions that are grossly indistinguishable from small adenomatous polyps. Microscopically, the colonic crypts are elongated and the epithelial cells assume a characteristic papillary configuration (Fig. 122-8). The epithelium is made up of

Chapter 122  Colonic Polyps and Polyposis Syndromes well-differentiated goblet and absorptive cells. The cytologic atypia that is characteristic of adenomatous polyps is not seen. Mitoses and DNA synthesis are limited to the base of the crypts, and orderly cell maturation is preserved. The epithelial cell and attendant pericryptal sheath fibroblast make up an epithelial-mesenchymal unit that migrates up the colonic crypt. In contrast to adenomatous polyps, in which the epithelium and fibroblast appear to be immature, this tissue is more differentiated, and abundant collagen is synthesized in the basement membrane.198 It is thought that the migration of epithelial cells up the colonic crypt is slow and that hyperplastic polyps develop from the failure of mature cells to detach normally. For the most part, true sporadic hyperplastic polyps are considered to have little if any intrinsic malignant potential; however, it is important to consider that hyperplastic polyps and neoplastic lesions can appear in the same colon, suggesting that the two may be pathogenetically related. Indeed, a germline mutation of the APC gene has been associated with an unusually large number of colorectal hyperplastic

Figure 122-8.  Hyperplastic polyp. This high-power photomicrograph demonstrates the crypts of a hyperplastic polyp, consisting of elongated epithelial cells with nuclei that retain their basal orientation and demonstrate no atypia. The surface of the crypts assumes a serrated appearance. (Courtesy Noam Harpaz, MD, PhD, New York.)

polyps in association with adenomas.199 Coexisting hyperplastic and adenomatous polyps also are common in the setting of a strong family history of colorectal cancer, including HNPCC.200 It appears, however, that in patients with sporadic adenomas, the coexistence of hyperplastic polyps does not confer an increased risk of recurrent adenomas during a three-year follow-up period.201 Prevalence The prevalence of hyperplastic polyps is not known with precision, but these growths are common. In colonoscopic examinations of asymptomatic patients older than 50 years, hyperplastic polyps were found in 9% to 10%,66 although this frequency was higher (30% to 31%) among male veterans.202 Sigmoidoscopic screening of asymptomatic relatives of adenoma-prone kindreds revealed 26% with hyperplastic polyps, a prevalence that was essentially identical (28%) to that of asymptomatic spouse controls.203 Autopsy data report a prevalence rate of 20% to 35%.8,12 The incidence of hyperplastic polyps depends largely on the site of the colon being examined and on the age of the patient. Autopsy studies repeatedly observe a distal predominance of hyperplastic polyps.8,11,12 Of course, sigmoidoscopic studies, which focus on the distal colon and rectum, find many hyperplastic polyps, but even screening colonoscopy studies detect more hyperplastic polyps in the distal than in the proximal colon.62 Among all diminutive polyps (<5 mm) removed during colonoscopy, hyperplastic polyps outnumber adenomatous polyps in the rectum and sigmoid, and adenomas predominate in the remainder of the colon.15,204 The prevalence of hyperplastic polyps increases with age.10,53 There also is an association between prevalences of hyperplastic polyps and colon cancer (Table 12212), although this correlation is not as firm as the association between adenomas and colon cancers, nor does it necessarily imply any premalignant potential of the hyperplastic polyp itself. Data suggest that current smoking is a risk factor for hyperplastic polyps.88 Treatment Hyperplastic polyps remain small, usually are sessile, and seldom, if ever, cause symptoms. Inasmuch as they are not likely to give rise to cancer, little is gained by removing them, but because they cannot be distinguished from neoplastic or serrated polyps simply by gross examination, they usually are removed. Given their usual predominance in the distal colorectum, finding hyperplastic polyps in this location is not an alarming finding, particularly in the elderly.

Table 122-12  Prevalence of Hyperplastic Polyps in Various Populations

POPULATION Hawaiian-Japanese U.S. (New Orleans)   Whites   Blacks Brazil (São Paulo) Japan   Akita   Miyagi Costa Rica (San José) Colombia (Cali)

PREVALENCE OF COLON CANCER

PREVALENCE OF POLYPS IN MEN (%)

PREVALENCE OF POLYPS IN WOMEN (%)

20-39 yr

40-59 yr

60+ yr

20-39 yr

40-59 yr

60+ yr

Very high

50

69

84

0

57

73

High High Intermediate

11 10 14

19 18 26

13 14 40

— 6 12

25 7 23

30 9 31

Intermediate Low Low Low

0 1 0 6

2 2 7 17

2 2 7 11

0 0 5 2

2 0 1 9

8 3 9 16

From Correa P. Epidemiology of polyps and cancer. In Morson BC, editor. The pathogenesis of colorectal cancer. Philadelphia: WB Saunders; 1978. p 126.

2173

2174

Section X  Small and Large Intestine Because sigmoidoscopic findings sometimes are used to decide the need for a full colonoscopy, the importance of the distal hyperplastic polyp as a marker of proximal neoplasia has been questioned. A systematic review of the subject found that in asymptomatic persons, a distal hyperplastic polyp is associated with a 21% to 25% risk for any proximal neoplasia and a 4% to 5% risk of advanced proximal neoplasia, thus justifying colonoscopy.205 Thus, although most patients with a distal hyperplastic polyp likely will undergo colonoscopy, current guidelines suggest that if no adenomas are found, such patients can subsequently be followed at intervals similar to those without any polyps.197 Serrated Polyps and the Serrated Neoplasia Pathway Evidence is mounting to support the concept of a distinct serrated neoplasia pathway.206,207 Polyps that display features of both hyperplastic and adenomatous transformation have been described; these mixed hyperplastic-adenomatous serrated polyps account for about 13% of hyperplastic polyps.208 It has become clear that polyps with a serrated phenotype, instead of representing simple hyperplastic polyps, might represent more clinically significant lesions. Serrated polyps are now classified into hyperplastic polyps (with two subtypes: goblet cell and microvesicular), and two types of serrated adenomas: traditional serrated adenoma (TSA) and sessile serrated adenoma (SSA, sometimes referred to as sessile serrated polyp) (Fig. 122-9).209 SSAs are considered the precursor lesions for the large hyperplastic polyps found in the proximal colon of patients with hyperplastic polyposis.209 When sporadic, SSAs are found more commonly in the proximal colon. By contrast, TSAs look and behave like conventional adenomas: They are often pedunculated, have unequivocal adenomatous dysplasia (albeit with crypt serration), and are found more often in the distal colon. At the molecular level, SSAs often have extensive BRAF mutations and extensive DNA methylation, whereas TSAs usually do not have BRAF mutations and have only infrequent mutations of APC, TP53, and MSI, but they can have K-ras mutations.209 Serrated adenomas are much less common than hyperplastic polyps, accounting for less than 1% of all polyps and between 1% and 11% of adenomas.207 The prevalence of high-grade dysplasia and cancer in these lesions, however, may be as high as 5% to 16%.207 The risk of colorectal cancer may be as high as 1 in 17 in a patient with a proximal sessile serrated adenoma. There are no guidelines for the management of SSAs. It is generally recommended that surveillance intervals for serrated adenomas, although not yet defined, should follow that of other adenomas, bearing in mind their possibly more rapid progression.209

Mucosal Polyps

The colon commonly harbors excrescences or mammillations of tissue that histologically are normal mucosa. In these instances, the submucosa has elevated the normal tissue overlying it. These lesions may be termed mucosal polyps, and their presence has no clinical significance. Mucosal polyps always are small and can constitute 8% to 20% of the material recovered in a collection of colonoscopic biopsies.

Juvenile Polyps

Juvenile polyps (Fig. 122-10) are mucosal tumors that consist primarily of an excess of lamina propria and dilated cystic glands, rather than an overabundance of epithelial cells as seen in adenomatous and hyperplastic polyps;

A

B Figure 122-9.  A, Sessile serrated polyp (sessile serrated adenoma). The crypts at the base of the polyp are broad and flattened, with nuclear pleomorphism and prominent nucleoli. The mid and upper portion of the same crypts show maturation toward more normal appearing nuclei but with a serrated appearance on the luminal surface and dystrophic goblet cells. B, Traditional serrated adenoma. The long fronds of this adenoma are lined by dysplastic epithelial cells characterized by nuclear pleomorphism interspersed with cells that have dystrophic goblet cell vacuoles. (Courtesy Noam Harpaz, MD, PhD, New York.)

therefore they are classified as hamartomas. The appearance of distended, mucus-filled glands, inflammatory cells, and edematous lamina propria has prompted some observers to call these lesions retention polyps. Juvenile polyps appear to be acquired lesions because they seldom are seen in the first year of life and are most common from ages one to seven years; occasionally they are found in adults. Juvenile polyps more often are single than multiple, usually are pedunculated, and tend to range in size from 3 mm to 2 cm. Because these polyps tend to be in the rectum and to develop a stalk, they can prolapse during defecation and even slough. In addition, the stroma contains a generous vascular supply, which explains the considerable blood loss suffered by some patients with juvenile polyps. Because of the high likelihood of bleeding and prolapse, removal of juvenile polyps is suggested. Juvenile polyps have essentially no malignant potential when single,210 and they tend not to recur. Although approx-

Chapter 122  Colonic Polyps and Polyposis Syndromes imately 20% of individual juvenile polyps in the rectum may be associated with proximal polyps, proximal adenomas are rare, and the subsequent rate of dying of or developing colorectal cancer is no greater than these rates for the general population, even without specific surveillance.210 When juvenile polyps are multiple (see “Juvenile Polyposis Syndrome,” later), however, the risk of developing cancer is present, either because adenomatous epithelium may be present in some juvenile polyps or because of a coexistent adenoma.

Peutz-Jeghers Polyps

The Peutz-Jeghers polyp is a unique hamartomatous lesion characterized by glandular epithelium supported by an arborizing framework of well-developed smooth muscle that is contiguous with the muscularis mucosae (Fig. 12211). The smooth muscle bands fan out into the head of the polyp and become progressively thinner as they project toward its surface. A Peutz-Jeghers polyp differs from a juvenile polyp in that the lamina propria is normal, and the characteristic architecture of the lesion derives chiefly from its abnormal smooth muscle tissue. These polyps almost always are multiple, and their distinctive appearance, in association with the extraintestinal manifestations, makes Peutz-Jeghers syndrome easily identifiable. This type of polyp seldom is found in the colon in the absence of generalized polyposis (see later).

Inflammatory Polyps (Pseudopolyps)

Figure 122-10.  Juvenile polyps. Top, Resection specimen of a colon harboring multiple juvenile polyps. Bottom, Histologic photomicrograph revealing the characteristic large cystic spaces of this lesion. (From Demetris AJ, Finkelstein SD, Nalensnik MA, et al. Slide carousel of GI pathology course for medical students. Available at: http://www.pathology.pitt.edu/lectures/ gi/. Copyright © Department of Pathology, University of Pittsburgh School of Medicine.)

A

Inflammatory polyps are found in the regenerative and healing phases of inflammation. They usually are formed by full-thickness ulceration of epithelium followed by a regenerative process that leaves the mucosa in bizarre polypoid configurations. Less commonly, inflammatory polyps might represent relatively normal mucosa sitting amidst reepithelialized ulcerations. Inflammatory polyps may be large and solitary, mimicking a neoplastic mass, or they can form mucosal bridges that span the lumen. Multiple lesions can mimic a polyposis syndrome. The term pseudopolyp is used to distinguish them from neoplastic lesions, but in reality these are true polypoid protuberances. Histologically, inflammation and exuberant granulation tissue may be seen in the early postinflammatory period, but later the polyp surface can resemble that of normal mucosa. Any form of severe colitis, including chronic inflammatory bowel disease (ulcerative colitis, Crohn’s disease),211 amebic colitis,212 ischemic colitis, or bacterial dysentery, can give rise to inflammatory polyps. In chronic schistosomiasis, multiple inflammatory polyps that contain granulation tissue, eggs, or adult worms commonly are seen.213 The significance of these polyps, which have no intrinsic neoplastic potential, is that they often appear in diseased colons that are at high risk for developing colon cancer (ulcerative colitis, schistosomiasis); therefore, they must be distinguished from neoplastic lesions that do carry premalignant potential. Giant or grouped pseudopolyps can cause colonic obstruction.

B

Figure 122-11.  Low-power (A) and high-power (B) photomicrographs of a Peutz-Jeghers polyp. In this type of polyp, the glandular epithelium is supported by an arborizing framework of well-developed smooth muscle that is contiguous with the muscularis mucosae. (From Boland CR. The colon, rectum, and anus. In: Feldman M, editor. Gastroenterology and Hepatology: The Comprehensive Visual Reference. Philadelphia: Churchill Livingstone; 1996.)

2175

2176

Section X  Small and Large Intestine Rare cases of multiple and recurrent inflammatory gastrointestinal polyps that produce pain and obstruction have been reported on a sporadic and even a familial basis.214 These lesions are found primarily in the ileum, may be very large, and can even cause intussusception. Cap polyposis is another rare condition, characterized by inflammatory polyps with elongated crypts, a mixed inflammatory infiltrate in the lamina propria, and a surface cap of fibri­ nopurulent exudate.215 Cap polyposis may be confused endoscopically with pseudopolyps in inflammatory bowel disease. Mucosal prolapse has been suggested as a possible underlying etiology.

SUBMUCOSAL LESIONS Colitis Cystica Coli

Colitis cystica profunda (see Chapter 124) is a rare lesion consisting of dilated, mucus-filled glands in the submucosa that can form solitary or multiple polyps. The typical lesion is a solitary polyp smaller than 3 cm that most commonly is found in the rectum in the setting of chronic inflammation. Prior surgical procedures and ulcerative proctitis have been linked to the pathogenesis of this abnormality. The involved epithelium shows no evidence of dysplasia. The primary significance of this lesion is that it must be recognized and distinguished from colloid carcinoma, which can look similar histologically, because a mistaken diagnosis of colloid carcinoma could lead to inappropriate radical surgery.216 The lesion is presumably caused by displacement of normal colonic glands to beneath the epithelium during the healing of a surgical wound or inflammation. Rarely, the polyps become large or recur and can even obstruct the colon. It has been suggested that the pathologic picture and clinical presentation of colitis cystica profunda are similar to those of the solitary rectal ulcer and that both may be produced by rectal prolapse.216

Pneumatosis Cystoides Coli

Multiple gas-filled cysts are occasionally encountered within the submucosa of the colon (and small intestine) and can produce a polypoid appearance. The diagnosis of pneumatosis cystoides intestinalis (see Chapter 124) may be made on full-thickness pathologic sections or by the characteristic radiologic or endoscopic appearance of the intramural gas-filled cysts. The diagnosis is substantiated at endoscopy if the cysts collapse after aspiration with a sclerotherapy needle or by unroofing them with a biopsy forceps. This condition can produce a variety of symptoms, some of which suggest colitis, but it also may be associated with vague symptoms or can remain asymptomatic.217 Two forms of pneumatosis intestinalis have been recognized. One type (pneumatosis linearis) may be associated with a fulminant mucosal process, such as inflammatory or ischemic bowel disease in adults or necrotizing enterocolitis in children. In these fulminant settings, in which the condition often is fatal, it is thought that the cysts result from invasion of the submucosa by gas-forming bacteria. The more common type, pneumatosis cystoides intestinalis, is seen in adults and is more typically a chronic or incidental finding; it even may be associated with an asymptomatic pneumoperitoneum. Pneumatosis cystoides intestinalis is associated with chronic obstructive pulmonary disease and may be seen in patients with scleroderma. The genesis of the gas-filled cysts in these benign settings is incompletely understood, but it has been demonstrated that gas within the bowel lumen diffuses into the cysts, which can contribute to their maintenance. Oxygen therapy results in the

resolution of these cysts,218 but the pathophysiologic basis of this response is by no means clear. The natural history can be deduced from only a small number of cases, but the disease can persist for months.217-219 A single course of oxygen therapy (often as little as 5 to 6 L/min of oxygen) can result in resolution of symptoms for a long time. Antibiotic treatment is of no benefit.

Other Submucosal Lesions

Any lesion beneath the colon mucosa can elevate the overlying epithelium to produce a polypoid appearance. Lymphoid tissue is present throughout the colon, and hypertrophied follicles may be mistaken for a pathologic mucosal process. Benign lymphoid polyps can grow large enough to produce symptoms (pain, bleeding) or can become pedunculated. Multiple benign lymphoid polyps may be found as normal variants, particularly in children. The principal importance of benign lymphoid polyps is in their distinction from malignant lymphoid lesions. Malignant lymphoma220 and chronic lymphocytic leukemia221 can manifest as multiple colonic polyposis. The colon is the most common gastrointestinal site for lipomas, which tend to be solitary but may be multiple submucosal lesions. Lipomas usually are asymptomatic and detected incidentally. The low density of fat can give the lesions a characteristic radiologic appearance, and their soft, deformable nature is helpful to the colonoscopist in making the diagnosis grossly. Colonic lipomas are most common in the right colon and tend to occur on or near the ileocecal valve.222 Removal of these lesions usually is not necessary. Important tumors such as carcinoids, metastatic neoplasms (especially melanoma), and other rare cancers can produce submucosal lesions without distinctive iden­ tifying characteristics. Other submucosal lesions may be detected incidentally, including fibromas, neurofibromas, leiomyomas, granular cell tumors, hemangiomas, and endometriosis.

GASTROINTESTINAL POLYPOSIS SYNDROMES Gastrointestinal polyposis refers to the presence of numerous polypoid lesions throughout the gastrointestinal tract. The polyposis syndromes are distinct entities clinically and pathologically and they have been sorted into recognizable categories over the past century (Table 122-13). Most of these syndromes are inherited, and most are associated with an increased colon cancer risk, but all are classified on the basis of the histologic type of polyp and the clinical presentation. Advances in genetics have permitted a more accurate understanding of the relationships among these syndromes, and the genes responsible for these conditions, particularly the adenomatous polyposis syndromes, have provided insight into the genetic basis of sporadically occurring colon cancer.

INHERITED POLYPOSIS SYNDROMES

The inherited adenomatous polyposis syndromes include several entities that are characterized by large numbers of adenomatous polyps in the colon. The identification of the APC gene set the stage for understanding how colon cancers arise, and it permitted proper classification of Gardner’s syndrome, attenuated familial adenomatous polyposis (AFAP), and many cases of Turcot’s syndrome as variants of classic FAP (Table 122-14). The MUTYH (also known as MYH) base-excision repair gene also has

Chapter 122  Colonic Polyps and Polyposis Syndromes Table 122-13  Classification of Gastrointestinal Polyposis Syndromes Inherited Polyposis Syndromes Adenomatous polyposis syndromes Familial adenomatous polyposis Variants of familial adenomatous polyposis Gardner’s syndrome Turcot’s syndrome Attenuated adenomatous polyposis coli Familial tooth agenesis syndrome Bloom’s syndrome MUTYH polyposis (MYH polyposis) Hamartomatous polyposis syndromes Peutz-Jeghers syndrome Juvenile polyposis PTEN hamartoma tumor syndromes Cowden’s disease Bannayan-Ruvalcaba-Riley syndrome Rare hamartomatous polyposis syndromes Hereditary mixed polyposis syndrome Intestinal ganglioneuromatosis and neurofibromatosis Devon family syndrome Basal cell nevus syndrome Noninherited Polyposis Syndromes Cronkhite-Canada syndrome Hyperplastic polyposis syndrome Lymphomatous polyposis Nodular lymphoid hyperplasia

been identified as an important cause of multiple colonic adenomas.

Familial Adenomatous Polyposis

Genetics FAP is the most common adenomatous polyposis syndrome. It is inherited as an autosomal dominant disease with 80% to 100% penetrance and has an estimated prevalence of 1 in 5000 to 7500.223 In this condition, one mutated APC allele is inherited as a germline mutation from the affected parent, and adenomas develop when the second allele (from the unaffected parent) becomes mutated or lost. The identification of the gene responsible for FAP began in 1986 with the investigation of a patient who had multiple congenital malformations and a deleted portion of the long arm of chromosome 5 that was identified cytogenetically.224 Genetic mapping studies and restriction fragment length polymorphism (RFLP) analysis in 1987 led to the localization of the gene responsible for FAP in the 5q21-q22 region.225,226 At the same time, RFLP analysis suggested that one of the two alleles for this gene often was lost in sporadically occurring colorectal cancers. The fact that a lost gene might contribute to tumor progression suggested that the FAP locus might encode for a tumor suppressor gene.227 In 1991, the APC gene responsible for FAP was cloned by two collaborating groups and reported simultaneously.228-230 The large size of this gene (encoding 2844 amino acids)

MUTYH (mutY homolog [E. coli]) is a gene encoding MUTYH glycolase, an enzyme involved in oxidative DNA damage repair; PTEN (phosphatase and tensin homolog) acts as a tumor suppressor gene.

Table 122-14  Adenomatous Polyposis Syndromes SYNDROME

POLYPS

EXTRAINTESTINAL ABNORMALITIES

GENE MUTATION

Classic FAP

Colonic adenomas (thousands) Duodenal, periampullary adenomas Gastric fundic gland polyps Jejunal and ileal adenomas Ileal lymphoid polyps

Mandibular osteomas Dental abnormalities

APC (usually truncated protein)

Gardner variant of FAP

Same as FAP

Osteomas (mandible, skull, long bones) CHRPE Desmoid tumors Epidermoid and sebaceous cysts Fibromas, lipomas Thyroid, adrenal tumors

APC

Turcot variant of FAP

Colonic adenomas (sometimes fewer than in classic FAP)

Medulloblastoma Glioblastoma multiforme CHRPE

APC DNA MMR*

Attenuated FAP

Colonic adenomas (<100; proximal colon) Duodenal, periampullary adenomas Gastric fundic gland polyps

Mandibular osteomas (rare)

APC 5′ and 3′ regions

Familial tooth agenesis

Colonic adenomas Hyperplastic polyps

Agenesis of teeth

Axin2 (APC pathway)

Bloom’s syndrome

Colonic adenomas

Small stature Facial erythema/telangiectasia Male sterility Adenocarcinomas, leukemia, lymphoma

BLM

MUTYH polyposis

Colonic adenomas (5-100) Duodenal polyposis Gastric cancer

CHRPE Osteomas

MUTYH (MYH)

*May be more appropriately classified under hereditary nonpolyposis colon cancer (HNPCC) (see Chapter 123). APC, adenomatous polyposis coli; CHRPE, congenital hypertrophy of the retinal pigment epithelium; FAP, familial adenomatous polyposis; MMR, mismatch repair; MUTYH (mutY homolog [E. coli]) is a gene encoding MUTYH glycolase, an enzyme involved in oxidative DNA damage repair.

2177

2178

Section X  Small and Large Intestine Codon 500

Codon 1000

5' Gene (exons) 9/ 10/ 1 23 45 678 11 12 13 14 9a 10a

Codon 2000

Codon 1500

Codon 2500 3'

15 Mutation cluster region

Mutations Papillary thyroid carcinoma (140-1309) Hepatoblastoma (141-1578) CHRPE (453-1445) Osteomas (767-1578) Desmoids (1399-1580) AFAP

AFAP

Protein

Oligomerization

β-catenin binding

Microtubule association

Axin binding Figure 122-12.  Schematic diagram of the APC gene, its mutations, and functional domains of the protein. The gene consists of 15 exons. Mutations associated with attenuated familial adenomatous polyposis (AFAP) occur in the 5′ and 3′ regions of the gene. Congenital hypertrophy of the retinal pigment epithelium (CHRPE) lesions are seen only with mutations downstream of exon 9. The mutation cluster region is in the center of the gene, where most mutations give rise to florid polyposis. Papillary thyroid carcinoma, hepatoblastoma, osteomas, and desmoid tumors are associated with the regions shown. Domains of the APC protein responsible for oligomerization, β-catenin binding, and microtubule association are shown along the bottom. (Modified from Goss KH, Groden J. Biology of the adenomatous polyposis coli tumor suppressor. J Clin Oncol 2000; 18:1967; and Groen EJ, Roos A, Muntinghe FL, et al. Extra-intestinal manifestations of familial adenomatous polyposis. Ann Surg Oncol 2008; 15:2439-50.)

might account for the relatively high frequency of new mutations. Germline mutations are found in patients with FAP and Gardner’s syndrome, and in most instances, the mutations create a stop codon resulting in a truncated protein. The germline mutations are dispersed throughout the 5′ half of the gene, whereas somatic mutations of APC tend to accumulate in the mutation cluster region near the center of the gene (Fig. 122-12).231 The APC protein is a multifaceted regulator of colonic epithelial cell homeostasis and participates in processes of cell proliferation, migration, differentiation, apoptosis, and chromosomal segregation.232 The proximal portion of the APC protein contains regions that enable oligomerization as well as binding to proteins that regulate the actin cytoskeleton, thereby affecting cell morphology, polarity, and migration. In its central portion, the APC protein binds to b-catenin, a protein that normally maintains cell-cell junctions by anchoring a cell surface adhesion molecule, Ecadherin. In normal cells, APC forms a complex with other proteins (axin, conductin, and the GSK3b serine-threonine kinase) to bind b-catenin. This binding results in phosphorylation of b-catenin, which subsequently undergoes down-regulation in the cytoplasm. If mutations occur in the b-catenin-binding region, however, b-catenin no longer is down-regulated, allowing it to enter the nucleus, where it acts in conjunction with other transcription factors to upregulate target genes that then promote formation of adenomas. Similarly, germline mutations of axin were discovered to cause a multiple colonic adenoma phenotype (see later). The carboxy-terminal portion of APC contains a domain that contributes to proper chromosomal segregation and cytoskeleton regulation. Because the vast majority of APC mutations truncate the protein, this has the dual effect of disrupting the constitutive breakdown of b-catenin, leading to activation of cancer-associated genes and creating abnormal chromosomal segregation with resultant chromosomal instability.

Clinical Features Colonic.  Classic FAP is characterized by the progressive development of hundreds to thousands of adenomatous polyps in the large intestine. A patient who inherits an APC mutation usually does not develop adenomas until approximately 10 to 12 years of age; rarely, however, polyps appear in the first decade of life. In one early series of FAP cases, the average age at onset of polyps was 25 years, but symptoms did not appear until the age of 33 years. The average age for the diagnosis of adenomas was 36 years, for cancer 39 years, and for death from cancer, 42 years; 90% of FAP cases have been diagnosed by the time the patient is 50 years of age.233 A study that focused on early screening reported that 50% of FAP gene carriers have polyps at sigmoidoscopy by approximately 15 years of age.234 FAP begins with a small number of polyps, and the number increases progressively until the colon becomes studded with adenomas. All varieties of adenomatous polyps may be seen, including tubular, tubulovillous, and villous adenomas. The number of macroscopic polyps in a colectomy specimen averages 1000 but may be tens of thousands. Germline mutations of APC gene between codons 1250 and 1464 have been associated with more profuse carpeting of the colon (Fig. 122-13), whereas mutations elsewhere in the gene result in fewer colorectal polyps (Fig. 122-14).235 Histologic examination of the colon reveals numerous microscopic adenomas as well, the smallest of which can involve a single colonic crypt. The size and number of polyps correspond to the latent period between the onset of clinical disease and the time of detection; tumors tend to be more numerous in symptomatic probands than in asymptomatic younger relatives discovered by screening. Most polyps are small (<1 cm), and, individually, these polyps are identical to adenomatous polyps found in the general population. Colorectal cancer should be considered an inevitable consequence in the natural history of FAP, appearing approxi-

Chapter 122  Colonic Polyps and Polyposis Syndromes mately 10 to 15 years after the onset of polyposis. Colorectal cancer is unusual in adolescence, but it has been diagnosed as early as nine years of age.236 The cancers have the same pathologic grades of malignancy and the same distribution within the colon as are seen in the general population, except that multiple simultaneous cancers are much more frequent (48% of cases).233 Despite attention to screening and surveillance, as many as 25% of patients with FAP have colorectal cancer at the time of colectomy.237 Upper Gastrointestinal.  Because FAP patients are born with a germline APC mutation in all cells of the body, tumors often develop in other organs besides the colon. For

Figure 122-13.  Surgical resection specimen from a patient with familial adenomatous polyposis (FAP). Patients with FAP have multiple adenomatous polyps that often carpet the colon, as demonstrated in this specimen. This close-up view demonstrates the presence of innumerable polyps, all of which are adenomas that can contain villous elements or carcinoma. A total proctocolectomy is the only reasonable management in this situation. (Courtesy of Arnold Markowitz, MD, New York.)

A

example, polyps in the stomach and small intestine are present in almost all FAP patients (Fig. 122-15).238 Gastric polyps occur in 30% to 100% of patients, and curiously, most polyps in the stomach are non-neoplastic fundic gland polyps. These polyps are typically 1- to 5-mm sessile growths characterized microscopically by hyperplasia of fundic glands and microcysts. They can appear in the first decade of life, even before other gastrointestinal adenomas develop. Fundic gland polyps can be found in persons without FAP who take proton pump inhibitor therapy for prolonged periods, but these growths are rarely dysplastic. Epithelial dysplasia, however, can be seen in approximately 25% to 41% of fundic gland polyps in patients with FAP and results from mutations of the APC gene.239,240 Dysplasia in FAP-associated fundic gland polyps has been directly associated with larger polyp size, increased severity of duodenal polyposis, and antral gastritis, and it is inversely associated with use of acid-suppressive therapy and the presence of Helicobacter pylori.240 Gastric adenomas are uncommon and occur in approximately 5% of FAP patients, usually in the gastric antrum. The development of gastric adenocarcinoma in FAP patients is still rare in the United States239,241 but is more common in Japan, where the gastric cancer rate in the general population also is higher.242 A prolonged life expectancy for FAP patients after preventive proctocolectomy has been associated with an increased incidence of gastric cancer (0.6% to 4.2%) in European and Asian FAP registries.243,244 Microcarcinoids also may be found in the stomach of patients with FAP.245 Duodenal adenomas occur in 60% to 90% of FAP patients, and the incidence increases with age.246 There is a propensity for adenomas to involve the periampullary region and even to obstruct the pancreatic or biliary ductal system, rarely resulting in pancreatitis. As many as 50% to 85% of FAP patients manifest adenomatous change of the papilla of Vater (see Fig. 122-15).238,246 As a consequence, a 4% to 12% lifetime incidence of duodenal cancer (usually periampullary) has been reported244 with relative risks of 124 to

B

Figure 122-14.  Colonoscopic views of familial adenomatous polyposis (FAP) and attenuated FAP. A, In this patient with classic FAP, polyposis is not florid; there were hundreds of polyps in the colon. B, In this patient with attenuated FAP, the polyps are fewer and smaller than in classic FAP. (Courtesy of Arnold Markowitz, MD, New York.)

2179

2180

Section X  Small and Large Intestine

A

B

Figure 122-15.  Endoscopic appearance of upper gastrointestinal polyps in a patient with familial adenomatous polyposis. A, Gastric fundic gland polyposis with numerous, small polyps distributed diffusely throughout the stomach. B, Adenomatous change on the papilla of Vater (arrow). These lesions may be subtle and require biopsy confirmation. (Courtesy of Arnold Markowitz, MD, New York.)

331.246,247 Collectively, these adenocarcinomas are the major cause of cancer death after prophylactic colectomy in FAP patients. The risk estimates for duodenal adenocarcinoma may be somewhat inflated by older prevalence studies, because under endoscopic surveillance a rather low rate of duodenal and ampullary adenoma progression to carcinoma has been observed.248 Regardless, it is advisable to perform screening and surveillance of the stomach and duodenum. Development of duodenal carcinoma is extremely rare before age 30 years.249 The Spigelman Classification for duodenal polyposis classifies patients with duodenal polyposis into various stages depending upon polyp number, size, histology, and severity of dysplasia.249 Patients with stage III and IV duodenal polyposis are more likely to progress to duodenal adenocarcinoma than those with earlier-stage disease. Although firm screening guidelines are not established, a suggested approach based on the Spigelman score has been proposed by a European consortium and is outlined in Table 122-15. Jejunal adenomas have been detected in 40% and ileal adenomas in 20% of FAP patients. Fortunately, malignant transformation at these sites is rare,238 but clinical vigilance is warranted and imaging with small bowel series, capsule endoscopy, or double balloon enteroscopy should be considered to examine the entire small intestine.247,250 After subtotal colectomy or total proctocolectomy with ileal pouch-anal anastomosis, attention should be given to surveillance of the distal ileum for developing neoplasia. Lymphoid hyperplasia may be present in the ileum in FAP patients and can be distinguished from adenomatous polyps by biopsy. Extraintestinal Features.  Gardner’s syndrome is a familial disease consisting of gastrointestinal polyposis and osteomas associated with a variety of benign soft tissue tumors and other extraintestinal manifestations (see Table 122-14). Both FAP and Gardner’s syndrome are variable manifestations of a disease caused by mutations of the APC gene. Bone abnormalities include osteomas of the mandible,

skull, and long bones; exostoses; and various dental abnormalities including mandibular cysts, impacted teeth, and supernumerary teeth. When carefully sought, mandibular osteomas can be seen in up to 90% of patients with FAP even without other stigmata of Gardner’s syndrome.251 Radiologic examination of the mandible is a simple and noninvasive means to screen for young carriers of the FAP gene, but it is crucial to distinguish nonspecific sclerotic lesions in the mandible from true osteomas. Mandibular osteomas in FAP tend to be multiple, whereas nonspecific sclerotic bony lesions usually are single and located close to a diseased tooth. Osteomas can occur in children before they develop colonic polyposis. Because osteomas have no malignant potential, they are removed only for symptomatic or cosmetic reasons. Congenital hypertrophy of the retinal pigmented epithelium (CHRPE) has been reported in some families with FAP or Gardner’s syndrome.252,253 More than 90% of patients with Gardner’s syndrome have pigmented lesions of the ocular fundus (vs. 5% of controls), which are likely to be multiple (63% have four or more lesions) and are bilateral in 87% of those affected.252 Pigmented ocular fundus lesions are found in approximately half of the unaffected but at-risk first-degree relatives and have been identified in infants as young as three months old, suggesting that they are probably congenital. The presence of multiple bilateral lesions appears to be a reliable marker for gene carriage in FAP, and their absence predicts lack of carriage if carrier relatives show CHRPE.253 These marker lesions are asymptomatic curiosities that need not be sought in patients with an established diagnosis of FAP. CHRPE perhaps reflects the most accurate genotype-phenotype correlation in FAP patients; these lesions occur in patients with APC gene mutations distal to exon 9 up through the proximal portion of exon 15 (see Fig. 122-12).254 A particularly serious complication of the adenomatous polyposis syndromes is the development of diffuse mesenteric fibromatosis, also called desmoid tumors. Desmoid tumors are reported in 4% to 32% of patients and rank second, after metastatic carcinoma, among lethal complica-

Chapter 122  Colonic Polyps and Polyposis Syndromes Table 122-15  Cancer Risks and Screening Recommendations in the Hereditary Polyposis Syndromes syndrome

LIFETIME RISK

Familial Adenomatous Polyposis; Gene Carriers Colon cancer Near 100% 5-12% Duodenal or periampullary cancer

SCREENING RECOMMENDATIONS

Gastric cancer In families of an affected member Pancreatic cancer Thyroid cancer CNS cancer Hepatoblastoma Peutz-Jeghers Syndrome All GI cancers

~0.5%

Sigmoidoscopy annually; start at age 10-12 yr* EGD with side-viewing endoscope; start at age 25-30 Spigelman Stages 0 and 1: every 5 yr; stage II: every 3 yr; stage III: every 1-2 yr; stage IV (see text): consider surgery Same as duodenal

~2% ~2% <1% 1.6%

Possibly periodic abdominal US after age 20 Annual thyroid examination; start at age 10-12 yr Annual physical examination; periodic head CT/MRI Annual physical examination/hepatic US/serum α-fetoprotein for first decade of life

2-13%

Small intestine

RR, 13

Pancreatic cancer Breast cancer Uterine cancer; ovarian cancer Sertoli cell tumor (testis) Juvenile Polyposis Colon cancer

RR, 100 RR, 8.8 RR, 8.0; 13 Uncommon

Colonoscopy at symptom onset, or late teens if the patient is asymptomatic, interval is determined by the number of polyps; at least every 3 yr Upper GI endoscopy every 2 yr; start at age 10 yr Annual Hb; small bowel series or video capsule endoscopy every 2 yr; start at age 10 yr Endoscopic or abdominal US every 1-2 yr; start at age 30 yr Annual breast examination; mammogram every 2-3 yr; start at age 25 yr Annual pelvic examination, Pap smear, pelvic US; start at age 20 yr Annual testicular examination beginning at age 10 yr; testis US if feminizing features

Gastric, duodenal cancer Cowden’s Syndrome Colon cancer Thyroid cancer Breast cancer Uterine and ovarian cancer

Rare

Colonoscopy: start at symptom onset or early teens if no symptoms; interval is determined by number of polyps, but is performed at least every 3 yr Upper GI endoscopy every 3 yr; start in early teens

Little to none 3-10% 25-50% Increased?

No recommendations given Annual thyroid examination; start in teens Annual breast examination at age 25 yr; annual mammogram at age 30 yr No recommendations given

<50%

*Sigmoidoscopy is used to identify a child with the FAP phenotype (i.e., polyps). If polyps are detected, colonoscopy usually is then performed to exclude proximal neoplasia. CNS, central nervous system; CT, computed tomography; EGD, esophagogastroduodenoscopy; FAP, familial adenomatous polyposis; GI, gastrointestinal; Hb, hemoglobin level; MRI, magnetic resonance imaging; Pap, Papanicolaou; RR, relative risk; US, ultrasonography. Adapted from Burt RW. Colon cancer screening. Gastroenterology 2000; 119:837; and Vasen HF, Möslein G, Alonso A, et al. Guidelines for the clinical management of familial adenomatous polyposis (FAP). Gut 2008; 57:704-13.

tions of the disease.255 The absolute risk of desmoids in FAP patients has been estimated at 2.56/1000 person-years, which is 825 times the risk in the general population.256 Desmoid tumors often display familial aggregation; FAP patients who are first-degree relatives of a patient with FAP and a desmoid have a 2.5-fold greater risk for developing desmoid tumors than patients with FAP in general.255 Because a strong family history poses an increased risk of desmoids,256 in this subset of patients it would be prudent to incorporate abdominal imaging studies into their overall surveillance regimen, even though firm guidelines for desmoid tumor surveillance are not yet established. Desmoids occur when the disease-causing mutation is distal to codon 1444 (or in some studies, codon 1399) of the APC gene (see Fig. 122-12). Curiously, recurrent desmoid tumors can manifest a somatic mutation of APC gene that is different from that of the initial tumor.231 Commonly, desmoid tumors result from progressive growth of mesenteric fibroblasts that occurs after laparotomy (e.g., after prophylactic proctocolectomy), but they occasionally appear spontaneously. The fact that desmoids seem to occur more often in women than men and more often after early colectomy (younger than 18 years) has prompted some experts to suggest considering delaying colectomy for some young female FAP patients.257 Desmoids cause gastrointestinal obstruction; constrict arteries, veins, or ureters; and are associated with a 10% to 50% mortality rate. Additional operative procedures usually are of no avail in this condition. Desmoid tumors can respond to radiation when local-

ized and accessible.258 Unfortunately, most tumors are in the mesentery in these patients, making radiation therapy impractical. Attempts at medical therapy have had some modestly encouraging results. The NSAID sulindac, which can often cause regression of colonic adenomas in FAP (see later), has resulted in partial tumor shrinkage in some patients but no response in others.259,260 The anti-estrogen drug tamoxifen has been effective in a few patients, as has progesterone261,262; a combination of sulindac and tamoxifen is commonly used.249 A staging system for FAP-associated desmoid tumors has been proposed which should help with medical decisionmaking (Table 122-16).263 It appears that patients with earlystage disease have a longer colectomy-to-desmoid interval, a somewhat reduced polyp burden, a higher likelihood that their desmoid will remain stable or disappear, and a much better prognosis. Multimodality therapy, including chemotherapy,264 seems appropriate for stage III and IV disease because of the more aggressive course of these tumors. For desmoids that significantly compromise the small bowel mesentery, small intestinal transplantation should be considered. In addition to desmoid tumors, other soft tissue tumors are well described in FAP and Gardner’s syndrome, including epidermoid cysts, fibromas, and lipomas. The epidermoid cysts, also called inclusion cysts, have erroneously been referred to as sebaceous cysts in the past. Epidermoid cysts are lined with normal epithelium and contain no sebaceous glands. When multiple epidermoid cysts appear before puberty in these kindreds, it is a harbinger of polyp­

2181

2182

Section X  Small and Large Intestine Table 122-16  Staging and Prognosis of Desmoid Tumors in Familial Adenomatous Polyposis STAGE

DEFINITION

I

Asymptomatic; <10 cm maximum diameter, and not growing Mildly symptomatic; <10 cm maximum diameter, and not growing Moderately symptomatic or with bowel or ureteral obstruction; or 10-20 cm, or slowly growing Severely symptomatic; or >20 cm, or rapidly growing

II III IV

NO. OF PATIENTS

TIME FROM COLECTOMY TO DESMOID (YR)

RAPID GROWTH

DIED OF DESMOID

21

7.5

0

0

36

5.8

3 (8%)

0

26

2.4

3 (12%)

4 (15%)

18

1.4

6 (33%)

8 (44%)

From Church J, Lynch C, Neary P, et al. A desmoid tumor-staging system separates patients with intra-abdominal, familial adenomatous polyposis–associated desmoid disease by behavior and prognosis. Dis Colon Rectum 2008; 51:897-901.

osis. Neoplasms of the biliary tree, liver, and adrenal glands also occur in these syndromes, and papillary carcinoma of the thyroid occurs in 1% of patients with FAP, pre­ dominantly in female patients.265 Hepatoblastoma can affect young children in FAP families. The extraintestinal manifestations of FAP have been reviewed in detail elsewhere.265 Genotype-Phenotype Correlations Drawing precise genotype-phenotype correlations in FAP often is difficult because the identical APC gene mutation can give rise either to isolated colonic polyposis or to extracolonic manifestations.229 Moreover, an identical APC gene mutation can manifest quite different colonic and extracolonic phenotypic features among unrelated families.266 Even within a single family, the disease can express itself variably in different persons, including skipped generations and discordance in identical twins.267 There are even some families that appear phenotypically to have FAP but that do not have mutations of the APC gene.268 Additional genetic or environmental disease-modifying factors appear to be responsible for generating phenotypic variation. For example, in an animal model of human FAP in which the mouse APC gene has a germline mutation, a second gene, phospholipase A2, was found to modify the number of polyps.268 So far, studies do not substantiate phospholipase A2 as a genetic modifier of human FAP. In other animal experiments, crossing mice with FAP with those who lack the COX-2 gene resulted in a substantial decrease in intestinal polyposis.269 Indeed, COX-2 inhibitors cause adenoma regression in patients with FAP (see later). Thus, the FAP phenotype can be modified genetically and environmentally. Despite the discrepancies in genotype-phenotype correlations, some general patterns have emerged (see Fig. 122-12). Profuse polyposis is found in the mid-portion of the gene (between codons 1250 and 1464, but especially around codon 1300), whereas a mild colonic phenotype is observed for mutations that affect the extreme proximal (5′) and distal (3′) ends of the APC gene responsible for attenuated FAP (see later). Desmoid tumors often, but not always, are seen with mutations just distal to the profuse polyposis region. CHRPE lesions are present with mutations distal to exon 9. Papillary carcinoma of the thyroid is associated with mutations proximal to the mutation cluster region. Genetic Testing and Counseling Genetic testing is an important component of the overall care of patients with FAP and their families, not so much for the management of the affected person but to detect

mutant gene carriers.270 Approximately 20% of patients with FAP have a negative family history and represent new mutations at the APC locus.233 Genetic testing is performed by extracting DNA from peripheral blood leukocytes. Sequencing offers the best sensitivity for mutations of the APC gene, but it is expensive and might detect variants of unknown clinical significance. Because most mutations of the APC gene result in a truncated protein product, the in vitro protein truncation test offers a useful method for detecting gene carriers. This assay is successful in about 80% of families tested and has the advantage of only requiring one affected person. If it is successful in one family member, this test is nearly 100% accurate for identifying other gene carriers in that family. An affected person is tested first. Absence of a mutation in the affected person suggests that genetic testing of at-risk relatives is not likely to yield clinically useful information and that the family should be screened by clinical tests. A positive gene test allows at-risk relatives to be tested in a more focused manner and at lower cost than with genetic testing. It is recommended that testing at-risk children be delayed until age 10 to 12 years, when clinical screening usually begins. Genetic testing of other family members is performed best within the context of a comprehensive genetic counseling program because it raises many issues such as psychological denial, survivor guilt, premature worrying if testing is performed at too young an age, intrafamily strife, employment discrimination, and medical insurability.271 Diagnosis and Screening Patients with FAP can present with nonspecific symptoms, such as hematochezia, diarrhea, and abdominal pain. The key to the diagnosis and management of this disease, however, is to identify the presymptomatic person, and this objective is achieved by the assiduous pursuit of the diagnosis in the relatives of affected patients. The diagnosis is made easily by sigmoidoscopy, given the often diffuse distribution of polyps, but colonoscopy is preferred so that the full phenotype can be appreciated while excluding the presence of carcinoma elsewhere in the colon. The presence of more than 100 polyps and the confirmation that these are adenomas establish the phenotypic diagnosis of FAP. Studies from St. Mark’s Hospital in London on the natural history of FAP suggest that approximately 10 years elapses between the appearance of polyps and the development of cancer233; however, it is not advisable to delay surgery once the diagnosis is made, even in presymptomatic patients,

Chapter 122  Colonic Polyps and Polyposis Syndromes except in persons who have not completed puberty. Performing genetic testing at approximately age 10 to 12 years for at-risk persons helps to streamline the clinical evaluation. In a family with a known mutation, children who test positive can then undergo a screening sigmoidoscopy to determine the status of their disease. If the gene test is negative, the child can be spared sigmoidoscopy, although it still might be prudent to perform sigmoidoscopy after adolescence just to offset the rare possibility of laboratory error. Treatment Surgery.  Surgery is the only reasonable management option for colonic polyposis in FAP. The timing and extent of surgery are the major clinical considerations. Because any rectal mucosa that is left behind is at risk for developing carcinoma, the optimal treatment is to perform total proctocolectomy either with a conventional ileostomy or as a restorative proctocolectomy with ileal pouch-anal anastomosis. For the most part, the latter operation in skilled hands is associated with little morbidity and is preferred by patients, who must nonetheless be advised about the risk of decreased fecundity among women undergoing this procedure.272 A meta-analysis concluded that compared with restorative proctocolectomy, ileorectal anastomosis was associated with less bowel frequency, nocturnal defecation, and use of incontinence pads, but more fecal urgency.273 The two types of operation demonstrated comparable rates of sexual dysfunction, dietary restriction, and postoperative complications. For some patients, total proctocolectomy with conventional ileostomy is unacceptable, and they also do not want to risk the complications of an ileal pouch. In such cases, the decision to perform a subtotal colectomy with ileorectal anastomosis can be considered, bearing in mind that the rectal segment will remain at risk for carcinoma and the patient will have to comply with periodic surveillance examinations. In contrast to older patients who undergo ileorectal anastomosis, about one fifth of younger patients (median age of 35 years) with many rectal polyps develop cancer in 5 to 23 years. According to the Mayo Clinic experience, among patients who are followed for more than 20 years, about three fifths develop carcinoma in the rectal segment despite semi-annual sigmoidoscopic surveillance and fulguration of all polyps.274 The prognosis in patients who develop rectal cancers in this setting is quite poor; the five-year diseasefree survival rate has been reported to be 25%. In patients with ileorectal anastomosis, the risk of subsequent rectal cancer was higher in patients who had an APC mutation between codons 1250 and 1464, a finding that awaits confirmation in other studies.275 These data provide a strong case for total proctocolectomy for FAP patients. In spite of this ominous warning, others have advocated rectum-sparing operations and have achieved a reasonable degree of success. The Memorial Sloan-Kettering group in New York has asserted that a subtotal colectomy is safe for patients whose rectums are free of polyps. They also spare the rectum in patients with rectal polyps, carefully follow the patients, and perform additional surgery as soon as malignant change is found.276 The St. Mark’s group in London reports satisfaction with rectum-sparing procedures for all patients with FAP and fulgurates only adenomas 5 mm or more in diameter at three- to six-month intervals. This group reported that 11 of 173 of their patients developed carcinoma in the rectum but that only three of the 11 died of rectal cancer.277 The Cleveland Clinic has advocated the use of colectomy with ileorectal anastomosis and reported an actuarial survivorship rate of 80% in 133

patients after 20 years, despite the presence of rectal polyps.278 Other groups in the United States and Europe also prefer subtotal colectomy and ileoproctostomy, but approximately one quarter of patients treated this way have required a total proctectomy at a later date for cancer or intractable benign polyps.279 It appears, therefore, that patients might elect the more limited procedure if they are willing to comply with rigorous follow-up (sigmoidoscopy every three to six months) and accept a risk of malignancy in the rectum of approximately 10%. Because of decreased fecundity rates associated with restorative proctocolectomy, young women with FAP might elect to undergo primary subtotal colectomy with ileorectal anastomosis with plans to convert to a restorative proctocolectomy after they are finished with childbearing. However, they should be aware that desmoid tumors arising after the first operation might prevent secondary restorative proctocolectomy in a substantial minority of them.280 Medical therapy has provided some options for the surgical management of FAP, but as discussed in the next section, rectal cancer can still occur despite adenoma regression. Medical Treatment.  Small adenomatous polyps in the rectum can be reversible lesions. Spontaneous regression of rectal polyps has been reported after subtotal colectomy and ileorectal anastomosis for FAP,281 and this must be taken into account when one is evaluating the response of a rectalsparing surgical procedure or medical treatment for this disease. Because of its antioxidant characteristics and its effects in experimental colon cancer, ascorbic acid (vitamin C, 3 g/day) was tried in patients with FAP who had undergone subtotal colectomy with ileorectal anastomosis at least one year earlier. A modest effect was observed, but it was neither consistent nor strong enough to advocate for general use.282 Supplemental dietary calcium also was ineffective in polyposis patients.283 A more ambitious trial has been reported in which 58 patients with FAP were treated with ascorbic acid (4 g/day), α-tocopherol (vitamin E, 400 mg/ day), and supplemental fiber (22.5 g/day); a modest effect was seen after two years of therapy.284 A higher degree of enthusiasm has been generated for the use of NSAIDs in the treatment of colorectal polyps in FAP. Sulindac has been shown in both uncontrolled and controlled trials to decrease the number and size of colorectal adenomas in patients with FAP75 who had intact colons as well as those with subtotal colectomy and ileorectal anastomosis. Unfortunately, maintaining these patients on sulindac does not protect them from developing rectal cancer, and the drug’s effect on reducing adenomas is reversible on its discontinuation. Sulindac is less successful for controlling upper gastrointestinal neoplasia, and it does not appear to prevent the initial onset of adenomas in children who are genotypically affected with FAP.285 The mechanism by which sulindac causes colorectal adenoma regression in patients with FAP might relate in part to its ability to inhibit COX (prostaglandin synthase) activity and thereby to interfere with arachidonic acid metabolism. Because colorectal tumors (but not normal colonocytes) have high levels of COX-2 expression, it is possible that COX-2 inhibition by sulindac is responsible for adenoma regression.286 Indeed, when mice carrying an APC mutation were bred with mice carrying a disrupted COX-2 gene, fewer polyps developed.269 Sulindac also is capable of restoring the deficient apoptosis seen in colonocytes of patients with FAP, even without affecting colonocyte proliferation.287 Because COX-2 overexpression can prevent apoptosis, it is possible that the benefit from sulin-

2183

2184

Section X  Small and Large Intestine dac relates to its ability to inhibit COX-2. FAP patients with rectal adenomas treated with a selective COX-2 inhibitor also demonstrated a significant reduction in the number and size of adenomas.288 The sulfone derivative of sulindac has no inhibitory effect on either COX-1 or COX-2 enzymes, yet it too has been shown to cause regression of rectal adenoma in FAP patients.289 Celecoxib, a selective COX-2 inhibitor, reduced the number of duodenal adenomas in one study,290 but further confirmation of this effect is awaited. Screening of Extracolonic Organs Upper gastrointestinal screening should be performed at the time colonic adenomas are diagnosed or at least by age 25 years (see Table 122-15). A full evaluation of the entire small intestine should be performed at baseline. This can be done by small bowel series, capsule endoscopy, double balloon enteroscopy, or if necessary by performing intraoperative enteroscopy at the time of initial proctocolectomy. Upper gastrointestinal polyps are rare before the onset of colonic disease, but side-viewing upper endoscopy should be performed in addition to conventional forward-viewing endoscopy because of its better visualization of the duodenal ampulla. The overall approach to upper gastrointestinal polyps is one of conservatism. Gastric polyps should be sampled to see if they are adenomas or fundic gland polyps with dysplasia. In the duodenum, villous adenomas, adenomas with high-grade dysplasia, large adenomas, and symptomatic adenomas, regardless of histology, should be removed. Endoscopic ablation of periampullary adenomas can be performed relatively safely by endoscopists skilled in this procedure, but regrowth of adenomatous tissue is common.282 If duodenal polyps are small or few, surveillance can be performed every one to three years. The presence of worrisome duodenal adenomas or adenomatous change of the duodenal papilla warrants endoscopic inspection at more frequent intervals. Surgical resection of the duodenum, whether by local excision or pancreaticoduodenectomy, may be required in selected patients. Screening of other organs at risk for cancer is summarized in Table 122-15.

Variant Familial Adenomatous Polyposis Syndromes

Turcot’s Syndrome (Glioma-Polyposis) The term Turcot’s syndrome applies to a syndrome of familial colonic polyposis with primary tumors of the central nervous system (see Table 122-14).291,292 The phenotypic spectrum is broad, with colonic manifestations ranging from a single adenoma to profuse adenomatosis coli and brain tumors of different histopathologic types. Controversy exists as to whether this syndrome is inherited in an autosomal dominant or autosomal recessive manner. A comprehensive molecular diagnostic study of 14 Turcot’s syndrome families has clarified that Turcot’s syndrome kindreds fall into two groups based on their types of brain tumor and particular genetic alteration.293 The more common group has germline mutations of the APC gene, and these patients tend to have medulloblastomas. In several cases, the brain tumor preceded the diagnosis of polyposis. The APC mutations were heterogeneous, with no association between specific mutations and the development of brain tumors. The inactivation of both APC alleles in brain tumor tissue implicates the APC gene in the pathogenesis of these neoplasms. The risk of cerebellar medulloblastoma in FAP was calculated to be 92 times that of the general population. In contrast, the second group of patients, including the family originally described by Turcot, had glioblastoma multiforme tumors. These persons were found

to have germline mutations in DNA MMR genes typical of HNPCC. Thus, Turcot’s syndrome can be considered a true variant of FAP, although as with Gardner’s syndrome, maintaining a separate designation may be superfluous. Because of familial clustering, once a person with Turcot’s syndrome has been identified, screening for affected family members should include colonoscopy as well as imaging studies of the brain (see Table 122-15). Attenuated Adenomatous Polyposis Syndromes Patients with classic FAP syndromes typically have tens to thousands of colonic adenomas; however, an attenuated form of FAP (AFAP) has been identified in which persons manifest fewer adenomas that often have a flat rather than polypoid growth pattern and tend to cluster in the proximal colon.294 AFAP is associated with germline mutations of the very proximal and distal portions of the APC gene.295 Like classic FAP, patients with AFAP are prone to develop multiple fundic gland polyps, duodenal and gastric adenomas, and even periampullary carcinoma296; however, colorectal cancers arise at a later age (~55 years) in AFAP patients than in those with classic FAP. In AFAP patients, colonoscopic surveillance is recommended every two years starting at age 18 to 20 years.249 It is becoming increasingly clear that germline mutations not only in the APC gene but in other genes of the APC pathway, such as for b-catenin and axin, can give rise to a multiple adenomatous polyposis phenotype.83 Multiple colonic adenomas, colorectal cancer, and even hyperplastic polyps have been described in a Finnish family with familial tooth agenesis as a result of a germline mutation of Axin2.297 Even more intriguing is the observation that germline defects in genes involved in DNA repair also can result in a phenotype that mimics AFAP. For example, numerous colonic adenomas have been described in a patient with Bloom’s syndrome, a condition characterized by growth retardation, male sterility, facial erythema, and multiple cancers resulting from increased chromosomal breakage.298 It appears that the genomic instability induced by mutations in the causative BLM gene affects the APC gene, thereby resulting in the polyposis phenotype. Mutations of the MUTYH (also called MYH) gene are a common cause of the multiple colorectal adenoma phenotype.299 Among patients with 15 to 100 adenomas, up to 30% have germline (usually biallelic) MUTYH mutations.249 When DNA is damaged, MUTYH acts in concert with other DNA base-excision repair enzymes to prevent mutations from occurring (Fig. 122-16). When MUTYH is defective, G:C → T:A transversions occur, and if this affects the APC gene, the resulting loss of APC function gives rise to multiple adenomas. Clinical features of MUTYH polyposis include multiple adenomas (usually between 5 and 100, rarely a florid polyposis), frequent development of colorectal cancer, and even some hyperplastic polyps. The phenotype occurs at a somewhat later age than does FAP. Gastric cancer, duodenal adenoma, osteoma, and CHRPE lesions have been described in patients with MUTYH mutations, but unlike FAP, MUTYH polyposis is an autosomal recessive disorder. Thus, genetic testing and counseling is directed more toward siblings and spouses than to parents or children. It is recommended that patients with more than 10 adenomas should be tested for MUTYH gene mutations.249 Biallelic, but not monoallelic, MUTYH mutations have been associated with increased risk of colorectal cancer. Like AFAP patients, colonoscopic surveillance of patients with biallelic MUTYH mutations is recommended every two years starting at age 18 to 20 years.249

Chapter 122  Colonic Polyps and Polyposis Syndromes Other Adenoma Syndromes Other familial syndromes in which a high incidence of colonic cancer is associated with a small number of colonic adenomas, such as Muir-Torre syndrome, are probably part of the spectrum of HNPCC and should not be confused with FAP or Gardner’s syndrome.300 A single family with gastric hyperplastic polyposis and a high incidence of gastric cancer has been described.301 This rare condition is of inter-

Oxidative damage G C

est because the gastric fundus can develop hyperplastic polyposis in FAP, but it is not known if this syndrome is linked in any way to FAP. Other patients have been described with multiple hyperplastic polyposis of the colon, some of whom had coexisting adenomas or a vague family history of polyposis, but again any pathologic relationship to FAP is speculative.302 Six patients with multiple serrated adenomas have been described, four of whom had associated adenocarcinoma.303 Further studies are needed to establish any genetic connection between serrated adenomatous polyposis and FAP.

Hamartomatous Polyposis Syndromes G° C

OGG1

First replication G° A

OGG1

Second replication + MYH

– MYH

T A Figure 122-16.  Function of the MYH gene in DNA base excision repair. Oxidative damage gives rise to 8-oxoguanine triphosphate, which is incorporated into DNA (Go). With the next DNA replication, DNA polymerase will preferentially insert adenine (A) instead of cytosine (C) opposite Go, giving rise to Go:A mispairs, with Go in the template strand. In the presence of normal MYH gene function (+MYH), with the subsequent replication, adenine is removed and subsequent base-excision repair will convert Go:A to Go:C, which can then be repaired by 8-oxoguanine glycolase 1 (OGG1) to GC. In the absence of MYH function (−MYH), Go:A mispairs remain uncorrected, giving rise to G:C → T:A transversion mutations. G° C

Several discrete familial syndromes have been described that are characterized by multiple hamartomatous polyps of the gastrointestinal tract. These include Peutz-Jeghers syndrome, juvenile polyposis syndrome, PTEN (phosphatase and tensin homolog) hamartoma tumor syndromes (Cowden’s syndrome and Bannayan-Riley-Ruvalcaba syndrome), and other rare syndromes (Table 122-17). Most of these syndromes are associated with an increased risk of colorectal cancer. The progression of these polyps to cancer is not well understood but represents a different mechanism than that seen with adenomatous polyposis. This mechanism has been called the “landscaper phenomenon” because changes that predominantly affect the lamina propria are what lead to the epithelial cancers. Peutz-Jeghers Syndrome Peutz in 1921 and Jeghers in 1949 described the familial syndrome consisting of mucocutaneous pigmentation and gastrointestinal polyposis that now bears their names. Peutz-Jeghers syndrome appears to be inherited as a single pleiotropic autosomal dominant gene with variable and incomplete penetrance.304,305 Germline mutations of the STK11/LKB1, a serine-threonine kinase gene on chromosome 19p, cause this syndrome,306,307 but not all families

Table 122-17  Familial Hamartomatous Polyposis Syndromes SYNDROME

POLYPS

location of OTHER GI POLYPS

other FEATURES

MUTATED GENE

Peutz-Jeghers syndrome

Hamartomas with bands of smooth muscle in the lamina propria

Small intestine Stomach Colon

Pigmented lesions (mouth, hands, feet) Ovarian sex cord tumors Sertoli tumors of the testes Airway polyps Pancreatic cancer Breast cancer Colon and esophageal cancer

STK11/LKB1

Juvenile polyposis

Juvenile polyps; also adenomas and hyperplastic polyps

Colon Small intestine Stomach

Colon cancer in some families Congenital abnormalities

MADH4, BMPR1A

Cowden’s disease

Hamartomas with disorganized muscularis mucosae

Stomach Colon

Trichilemmomas and papillomas Other hamartomas Benign and malignant breast disease Benign and malignant thyroid disease

PTEN

Bannayan-RuvalcabaRiley syndrome

Juvenile polyps

Colon

Macroencephaly; developmental delay Penile pigmentation

PTEN

Neurofibromatosis

Neurofibromas

von Recklinghausen’s disease MEN 2B

NF1 RET

Small intestine

GI, gastrointestinal; MEN, multiple endocrine neoplasia.

Small intestine Stomach Colon

2185

2186

Section X  Small and Large Intestine

Figure 122-17.  Mucocutaneous pigmentation of Peutz-Jeghers syndrome. (From www.orl.cz/chorroby/ustni/vestibulum/vrozena.)

with Peutz-Jeghers syndrome are linked to this gene locus, suggesting genetic heterogeneity. Early in infancy, the characteristic mucocutaneous pigmentation of Peutz-Jeghers syndrome may be noted. These deposits of melanin are found most commonly around the mouth, nose, lips, buccal mucosa, hands, and feet, and they also can occur in the perianal and genital regions (Fig. 12217). The macular lesions are brown to greenish-black, are smooth and hairless, and, except for the buccal pigmentation, tend to fade at puberty. The clinician must distinguish these melanin deposits from ordinary freckles. Freckles are sparse near the nostrils and mouth, are absent at birth (but can occur in infancy), and never appear on the buccal mucosa. The presence of this pigmentation should alert the clinician to this syndrome, but the skin lesions and intestinal lesions occasionally are inherited separately. Peutz-Jeghers polyps can increase in size progressively and cause small intestinal obstruction or intussusception that can occur as early as infancy. The polyps may be found in the stomach, small intestine, or colon, but they tend to be most prominent in the small intestine. Acute upper gastrointestinal bleeding and chronic fecal blood loss can complicate the disease. The average age of diagnosis in Peutz-Jeghers syndrome is 23 to 26 years old. Carcinomas of the colon, duodenum, jejunum, and ileum have been reported in patients with Peutz-Jeghers syndrome. Although it has been assumed that these cancers arise from rare foci of adenomatous epithelium that can develop within the Peutz-Jeghers polyps, evidence for loss of STK11/LKB1 expression in Peutz-Jeghers polyps even without dysplastic epithelium raises the possibility that the STK11/LKB1 gene itself might be the gatekeeper to carcinogenesis in this syndrome, much like APC is the gatekeeper in FAP.308,309 Colon polyps should be removed for histologic examination, but the relative inaccessibility of small intestinal polyps and the unpredictability of neoplastic complications make it difficult to have a routine surveillance program for small intestinal cancer. Cancers throughout the gastrointestinal tract and other organs are quite common in familial Peutz-Jeghers syndrome (see Table 122-17).310 The mean age of diagnosis of cancer is approximately 40 to 50 years, with a 93% overall cumulative risk of developing cancer between 15 and 64 years of age. The most common cancers include breast (54%), colon (39%), pancreas (36%), stomach (29%), ovary (21%), and small bowel (13%). Ovarian cysts and distinc-

tive ovarian sex cord tumors are seen in 5% to 12% of female patients with this syndrome.311 The ovarian tumors are histologically unique and can occur in young patients. Hormonally active Sertoli cell testicular tumors with feminizing features can occur in young boys with PeutzJeghers syndrome.312 Breast cancers may be found in young women and may be bilateral,313 and the magnitude of breast cancer risk in this syndrome is similar to that for other hereditary forms of breast cancer caused by germline mutations of BRCA1 and BRCA2. Other tumors that can occur in this syndrome include pancreatic cancers in young patients and polyps or cancers of the biliary tree and gallbladder.314 Thus, Peutz-Jeghers syndrome confers an increased risk for cancer in a number of gastrointestinal and nonintestinal organs. Guidelines for screening are difficult to make but should be directed toward at-risk organs for which early detection and treatment are reasonable, such as the entire gastrointestinal tract, gonads (in both sexes), and breasts (in women). Video capsule endoscopy often reveals more polyps than do barium studies.250,315 COX-2 inhibition can reduce the polyp burden in Peutz-Jeghers syndrome patients,316 and rapamycin seems to reduce the polyp burden in an animal model of Peutz-Jeghers syndrome317; the role of these agents in clinical practice is not yet defined. With respect to the luminal gastrointestinal tract, endoscopic polypectomy is the mainstay of treatment. Double balloon enteroscopy, or intraoperative endoscopy of the entire gastrointestinal tract, should be considered. Surgery is reserved for large, difficult to remove, or recurrent polyps; attention is focused on polypectomy while trying to avoid bowel resection as much as possible. Tuberous Sclerosis Tuberous sclerosis is characterized by the presence of hamartomatous lesions, with the classic triad of mental retardation, epilepsy, and adenoma sebaceum.318 Hamartomatous polyps resembling Peutz-Jeghers polyps, as well as adenomatous polyps, can occur in this disease and often are located in the distal colon. Juvenile Polyposis Syndrome Juvenile polyps are distinctive hamartomas that usually are solitary and are located principally in the rectums of children and occasionally in adults. They have a smooth surface and are covered by normal colonic epithelium. Juvenile polyposis (i.e., the presence of multiple juvenile polyps) is a diagnosis of exclusion. The PTEN hamartoma syndromes, which also can be associated with juvenile polyps such as Cowden’s disease and Bannayan-Ruvalcaba-Riley syndrome, should be excluded (see next section). Once that is done, juvenile polyposis syndrome can be defined by any one of the following criteria: five or more juvenile polyps of the colon and rectum, juvenile polyps throughout the gastrointestinal tract, or any number of juvenile polyps in the gastrointe­stinal tract with a family history of juvenile polyps.319 Extracolonic manifestations can affect the CNS (macroce­phaly, hydrocephalus), thorax (coarctation of aorta, atrial septal defects, tetralogy of Fallot), urogenital tract (undescended testes, bifid uterus and vagina, unilateral renal agenesis), and gastrointestinal tract (Meckel’s diverticulum, malrotation).319 Juvenile polyposis typically causes gastrointestinal bleeding, intussusception, and obstruction. The clinical presentations of juvenile polyposis syndrome and FAP differ. Juvenile polyps produce symptoms in childhood, whereas the adenomatosis syndromes rarely manifest in childhood

Chapter 122  Colonic Polyps and Polyposis Syndromes and usually become evident in early adult life. The average age of a patient with juvenile polyposis syndrome is 4.5 years in the nonfamilial form and 9.5 years in the familial form.320 The risk of colon and upper gastrointestinal cancer is increased in familial juvenile polyposis.319 In a large Iowa kindred, 38% of affected persons developed colon cancer and 21% developed upper gastrointestinal cancer.321 Others report a somewhat lower risk estimate of 17% incidence of gastrointestinal malignancy.322 Although juvenile polyps per se are not considered neoplastic, the synchronous adenomatous polyps and mixed juvenile-adenomatous polyps of these patients are what give rise to concern.323,324 Thus, these polyps must be scrutinized by the pathologist for evidence of a mixed adenomatous appearance, coexisting adenomas must be excluded, and kindreds with colorectal cancer should be subjected to careful colonoscopic surveillance. Juvenile polyposis syndrome manifests autosomal dominant inheritance. Approximately 18% of cases have a germline mutation of MADH4 (SMAD4), a tumor suppressor gene involved in TGF-β signaling.325 Knockout mouse models confirm that MADH4 germline mutations predispose to gastrointestinal polyps and cancer.319 Another 21% of cases are caused by germline mutations of bone morphogenetic protein receptor 1A (BMPR1A), another member of the TGF-β superfamily that depends on MADH4 for signal transduction. A family history of juvenile polyposis syndrome occurs in 33% of cases (the other 66% are sporadic juvenile polyposis syndrome). The diagnosis of juvenile polyposis syndrome is made by endoscopy (see Table 122-15). Screening colonoscopy and upper endoscopy usually begins after 15 years of age if symptoms have not occurred already. The cumulative risk of colorectal and gastric cancer is approximately 50% and 21%, respectively. Asymptomatic relatives also should be screened. Identification of a MADH4 or BMPR1A mutation in a family member helps guide screening, analogous to APC mutation testing. In general, juvenile polyps should be removed because of their tendency to bleed and obstruct. With a small number of polyps, periodic endoscopic polypectomy may be adequate. It has been suggested that surveillance endoscopy with polypectomy of the upper and lower gastrointestinal tracts be performed yearly until the patient is polyp free and then every three years.319 For persons with numerous juvenile polyps, colectomy should be considered. If subtotal colectomy with ileorectal anastomosis is chosen, the rectal segment must remain under surveillance. Family history must be defined in patients with multiple juvenile polyps to determine the sites of involvement and the history of neoplastic lesions. Gastric polyposis can be quite diffuse and cause anemia, posing a difficult management problem. The risk of gastric cancer may be as high as 21%.319 PTEN Hamartoma Tumor Syndromes Germline mutations of PTEN, a tyrosine phosphatase protein that functions as a tumor suppressor, account for 81% and 57% of Cowden’s disease and Bannayan-Ruvalcaba-Riley syndrome cases, respectively.319 Cowden’s Disease.  Although reported in only a very small number of families, Cowden’s disease, or the multiple hamartoma syndrome, consists of hamartomatous polyps of the stomach, small intestine, and colon along with extraintestinal manifestations that include orocutaneous hamartomas, fibrocystic disease and cancer of the breast, nontoxic goiter,

and thyroid cancer.319 The hallmark of this autosomal dominant condition is the presence of multiple facial trichilemmomas, which arise from follicular epithelium and typically occur around the eyes, nose, and mouth. Gastrointestinal symptoms and colorectal cancer appear to be uncommon in this syndrome. The colorectal polyps in Cowden’s disease are distinctive lesions characterized by disorganization and proliferation of the muscularis mucosae, with nearly normal overlying colonic epithelium.326 Ganglioneuromatosis of the colon and glycogenic acanthosis of the esophagus have been reported in association with Cowden’s disease.327 There does not appear to be an increased risk of gastrointestinal cancer; the major complication is cancer of the breast, uterus, and thyroid. Thus, there are no specific screening recommendations for the colon. Bannayan-Ruvalcaba-Riley Syndrome.  A rare autosomal dominant syndrome, Bannayan-Ruvalcaba-Riley syndrome consists of hamartomatous gastrointestinal polyposis with macrocephaly, developmental delay and other developmental abnormalities, and pigmented spots on the penis.328,329 Thyroiditis also has been described. Rare Inherited Hamartomatous Polyposis Syndromes Hereditary Mixed Polyposis Syndrome.  A large kindred with a tendency to develop colonic polyps of mixed histologic types has been identified.330 The disease appears to be inherited in an autosomal dominant fashion and is confined to the colon. The earliest age of onset of polyps was 23 years; the median age of symptoms was 40 years; and the median age of colon cancer diagnosis was 47 years. The characteristic polyp was an atypical juvenile polyp, although some persons had polyps of mixed histology and others had more than one histologic type of polyp, including serrated adenomas. Linkage to chromosome 15q14-q22 has been described in a region that overlaps with a possible colonic cancer gene, CRAC1.331,332 Colon cancer risk is increased, supporting the recommendation of surveillance colonoscopy every one to two years.319 Intestinal Ganglioneuromatosis and Neurofibromatosis.  Approximately 25% of patients with von Recklinghausen’s syndrome (caused by NF1 gene mutations) have neurofibromatosis involving the upper digestive tract with multiple submucosal neurofibromas or, less commonly, ganglioneuromas, that can cause dyspepsia, abdominal pain, or hemorrhage.333 Gastrointestinal involvement usually is incidental and asymptomatic. Severe, uncontrolled symptoms have required surgical treatment in a few cases. Multiple intestinal ganglioneuromas also have been observed in families and individual cases unrelated to von Recklinghausen’s disease.334 Ganglioneuromas throughout the gastrointestinal tract can occur in patients with multiple endocrine neoplasia type 2B, related to mutations of the RET gene.335 Devon Family Syndrome.  Multiple and recurrent inflammatory fibroid polyps of the stomach and intestine have been reported in a family.336 These lesions, histologically distinct from juvenile polyps, can cause gastrointestinal obstruction, with symptoms beginning in adult life. Basal Cell Nevus Syndrome.  Basal cell nevus syndrome is another syndrome that has been associated with multiple gastric hamartomatous polyps337; however, several kindreds have been reported without mention of gastrointestinal lesions.

2187

2188

Section X  Small and Large Intestine NONINHERITED POLYPOSIS SYNDROMES Cronkhite-Canada Syndrome

In 1955, Cronkhite and Canada reported the first examples of an acquired nonfamilial syndrome that now bears their names.338 It is characterized by the presence of diffuse gastrointestinal polyposis, dystrophic changes in the fingernails, alopecia, cutaneous hyperpigmentation, diarrhea, weight loss, abdominal pain, and complications of malnutrition.339 Patients typically are middle-aged or older (average, 62 years) and present fairly acutely with a rapidly progressive illness consisting of chronic diarrhea and protein-losing enteropathy with the associated integumentary abnormalities. The diarrhea is attributable primarily to diffuse small intestinal mucosal injury, but bacterial overgrowth may be contributory. Gastrointestinal polyps are found in 52% to 96% of patients and range in location from the stomach to the rectum.339 These polyps are hamartomas similar to the juvenile (retention) type, but unlike juvenile polyposis, the mucosa between polyps is histo­ logically abnormal, with edema, congestion, and inflammation. As is the case with juvenile polyps, there may be foci of adenomatous epithelium, which can confer a risk of carcinoma. It is estimated that the risk of colon cancer is approximately 9%, and the risk of adenomas or adenomatous change is 40%.319 Gastric cancer risk is also increased. Thus, screening of the colon and stomach should be considered. The malabsorption syndrome is progressive in most patients, and the prognosis is poor because there is no specific therapy. It has been suggested that complete sympto­ matic remission occasionally may be achieved with supportive management. In some cases, a variety of medical and surgical measures have been employed, making it difficult to identify the essential therapeutic modality(s). Glucocorticoids, anabolic steroids, antibiotics, and surgical resections have been tried in many of the patients in whom remissions have been reported. Despite this dilemma, aggressive nutritional support appears to be the most important factor in effecting a favorable outcome. Enteral feeding (if possible) or parenteral feeding (if necessary) with sources of calories, nitrogen, and lipids, in addition to appropriate fluids, electrolytes, vitamins, and minerals, has resulted in complete symptomatic remissions with resolution of all the ectodermal aberrations. Antibiotics may be beneficial when bacterial overgrowth contributes to the malabsorption. Although glucocorticoids have been used in some of the cases of symptomatic remission, the evidence to support their use is weak. Surgical therapy offers less and is risky in these malnourished patients. One case of complete remission has been reported in a patient managed only with enteral administration of a nutritionally balanced complete liquid diet.340 Attention should be paid to the possibility of secondary lactose or other disaccharide intolerance or protein-losing enteropathy in patients with diffuse small intestinal disease. Specific management awaits a better understanding of this perplexing syndrome.

Hyperplastic Polyposis Syndrome

Hyperplastic polyposis syndrome (HPS) is characterized by multiple hyperplastic polyps in the colon. There is no defined germline mutation to explain this syndrome. Most cases are sporadic, with only 5% of patients having a family history of this disorder. The World Health Organization definition of HPS is at least five histologically diagnosed hyperplastic polyps proximal to the sigmoid colon, of which

two are larger than 10 mm; or any number of hyperplastic polyps occurring proximal to the sigmoid colon in a person who has a first-degree relative with HPS; or more than 30 polyps, but distributed throughout the colon.339 The hyperplastic polyps in this syndrome tend to predominate in the distal colon, but the larger ones more commonly are located in the proximal colon. The polyps themselves often demonstrate mutation of the BRAF gene and extensive DNA methylation. The latter phenomenon turns off key tumor suppressor genes (including DNA repair genes like MLH1), resulting in colorectal cancer. Sessile serrated adenomas have been described in this syndrome, and these polyps may be the factor that predisposes to colorectal cancer. Adenomas commonly coexist, but colorectal cancer apparently is uncommon.341 Colonoscopic surveillance every one to three years has been suggested. Shorter intervals might be considered in patients with AAPs.

Lymphomatous Polyposis

Lymphoma can manifest as multiple lymphomatous polyps of the gastrointestinal tract.342 A variety of pathologic variants of Hodgkin’s and non-Hodgkin’s lymphomas can manifest this way, including immunoproliferative small intestinal disease. One variant worthy of mention is the mantle zone lymphoma (MZL), which can produce nodular collections of proliferating lymphocytes in wide mantles that surround benign-appearing germinal centers.343 The MZL has a characteristic histologic appearance and can have an indolent clinical course. These lesions are important because they require an adequate biopsy to distinguish them from true epithelial polyps and because of the possibility that MZL can have a prolonged clinical course.

Nodular Lymphoid Hyperplasia

Nodular lymphoid hyperplasia is a rare lymphoproliferative condition that is not related to a specific disease. It can be seen in healthy children and also has been described in the terminal ileum of some patients with Gardner’s syndrome and in some immunodeficiency syndromes, particularly IgA deficiency. These polyps, which are more common in the small intestine and measure approximately 3 to 6 mm, typically do not cause symptoms.

KEY REFERENCES

Arber N, Levin B. Chemoprevention of colorectal neoplasia: The potential for personalized medicine. Gastroenterology 2008; 134:1224-37. (Ref 97.) Calva D, Howe JR. Hamartomatous polyposis syndromes. Surg Clin North Am 2008; 88:779-817. (Ref 319.) Cooper HS. Pathologic issues in the treatment of endoscopically removed malignant colorectal polyps. J Natl Compr Canc Netw 2007; 5:991-6. (Ref 176.) East JE, Saunders BP, Jass JR. Sporadic and syndromic hyperplastic polyps and serrated adenomas of the colon: Classification, molecular genetics, natural history, and clinical management. Gastroenterol Clin North Am 2008; 37:25-46. (Ref 209.) Groen EJ, Roos A, Muntinghe FL, et al. Extra-intestinal manifestations of familial adenomatous polyposis. Ann Surg Oncol 2008; 15:243950. (Ref 265.) Kahi CJ, Rex DK, Imperiale TF. Screening, surveillance, and primary prevention for colorectal cancer: A review of the recent literature. Gastroenterology 2008; 135:380-99. (Ref 86.) Kudo S, Lambert R, Allen JI, et al. Nonpolypoid neoplastic lesions of the colorectal mucosa. Gastrointest Endosc 2008; 68(4 Suppl):S3-47. (Ref 27.) Levin B, Lieberman DA, McFarland B, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: A joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. Gastroenterology 2008; 134:1570-95. (Ref 136.)

Chapter 122  Colonic Polyps and Polyposis Syndromes Martínez ME, Baron JA, Lieberman DA, et al. A pooled analysis of advanced colorectal neoplasia diagnoses after colonoscopic polypectomy. Gastroenterology 2009; 136:832-41. (Ref 189.) Rex DK. Maximizing detection of adenomas and cancers during co­­ lonoscopy. Am J Gastroenterol 2006; 101:2866-77. (Ref 137.) Vasen HF, Möslein G, Alonso A, et al. Guidelines for the clinical management of familial adenomatous polyposis (FAP). Gut 2008; 57:70413. (Ref 249.)

Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: A consensus update by the US Multi-Society Task Force on Colorectal Cancer and the American Cancer Society. CA Cancer J Clin 2006; 56:143-59. (Ref 197.) Winawer SJ, Zauber AG, Gerdes H, et al. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med 1993; 329:197781. (Ref 46.) Full references for this chapter can be found on www.expertconsult.com.

2189

CHAPTE R

123 Colorectal Cancer Robert S. Bresalier

CHAPTER OUTLINE Epidemiology  2191 Etiology  2193 Fat, Bile Acids, and Bacteria  2193 Fiber  2194 Carcinogens and Fecal Mutagens, Vitamins, and Micronutrients  2195 Calcium and Vitamin D  2195 Arachidonic Acid, Eicosanoids, and Cyclooxygenase-2  2196 Chemoprevention  2196 Biology  2199 Abnormal Cellular Proliferation  2199 Molecular Biology and Biochemical Changes  2200 Familial Colorectal Cancer  2204 Predisposing Factors  2207 Age  2208 Prior Adenoma and Carcinoma  2208 Family History  2208 Inflammatory Bowel Disease  2209 Other Associations  2210 Pathology, Natural History, and Staging  2211 Gross Pathology  2211 Histology  2211 Natural History and Staging  2212

Cancer of the colon and rectum (colorectal cancer [CRC]) is a major cause of cancer-associated morbidity and mortality in North America, Europe, and other regions with similar lifestyles and dietary habits. CRC is the fourth most common newly diagnosed internal cancer overall in the United States, after cancers of the lung, prostate, and breast, and currently constitutes 10% of new cancers in men and women. In 2009, there were an estimated 147,000 new CRC cases in the United States and 50,000 related deaths (a rate second only to that of lung cancer).1 In the United States, CRC incidence rates in men and women are similar, although there appears to be a slight male predominance worldwide. Approximately 6% of the American population eventually will develop invasive colon or rectal cancer, and more than 6 million Americans who are alive today will die of the disease; the lifetime risk of dying from CRC in the United States is 2.5%. Globally, CRC is the fourth most common cancer in men and the third most common in women, with mortality paralleling incidence. Despite evidence that fiveyear survival is 90% when CRC is diagnosed at an early stage, less than 40% of cases are diagnosed when the cancer is still localized.2 Rapid growth of knowledge about the molecular and biological characteristics of CRC has provided useful insights

Prognosis  2215 Surgical-Pathologic Staging  2215 Tumor Morphology and Histology  2215 Clinical Predictors of Prognosis  2216 Clinical Features  2218 Diagnosis and Screening  2219 Tests When Colorectal Cancer Is Suspected  2219 Principles of Screening  2219 Screening Techniques  2221 Carcinoembryonic Antigen and Other Tumor Markers  2225 Fecal DNA and Genetic Testing  2226 Approach to Screening  2226 Insurance Coverage for Screening  2227 Screening Capacity, Screening in Underserved Populations, and Quality Assurance  2229 Treatment  2230 Surgery  2230 Chemotherapy  2232 Immunotargeted Therapy and Immunotherapy  2237 Radiotherapy  2237 Endoscopic Therapy  2237 Other Malignant Colonic Tumors  2238

into the pathogenesis of these neoplasms and cancer in general. New insights also have been gained in regard to primary prevention. Because CRC arises over long periods as a result of interactions between genetic predisposition and environmental insults, it has become possible to identify preneoplastic and early neoplastic lesions better and to improve survival rates. Rapidly evolving knowledge of CRC pathogenesis, especially in high-risk groups, is allowing the development of new tools to identify those who will benefit most from cancer surveillance and from adjuvant therapy following potentially curative surgery. After decades with limited options for treating advanced disease, new options for chemotherapy are now available.

EPIDEMIOLOGY The frequency of CRC varies remarkably among different populations (Fig. 123-1).3 Incidence rates are highest in the developed countries in North America and in Australia and New Zealand; intermediate in Europe; and low in Asia, South America, and especially sub-Saharan Africa. Internationally, the incidence of colon cancer in men differs by a

2191

2192

Section X  Small and Large Intestine A

0

5

10 15 20 25 30 35

US, Detroit, Black 35.0 US, Los Angeles, Black 34.8 US, Hawaii, Japanese 34.4 US, Connecticut, Black 30.9 US, Connecticut, White 30.4 France, Bas-Rhin 30.2 US, Los Angeles, Other White 28.9 Australia, New South Wales 28.4 Italy, Varese 27.4 Canada 26.9 US, Los Angeles, Japanese 26.7 Germany, Saarland 25.5 Switzerland, Geneva 25.2 Israel, All Jews 24.9 Japan, Miyagi 24.9 Singapore, Chinese 24.2 Ireland, Southern 24.2 UK, Scotland 23.7 Netherlands, Eindhoven 23.7 China, Hong Kong 22.5 Norway 22.2 New Zealand, Maori 21.5 Denmark 20.6 US, Hawaii, Hawaiian 19.9 UK, England & Wales 19.9 US, Los Angeles, Chinese 18.3 Sweden 17.7 Spain, Navarra 16.5 Slovenia 15.7 Germany, Eastern 15.5 US, Puerto Rico 14.8 China, Shanghai 12.2 Poland, Cracow 11.4 Singapore, Indian 7.6 Colombia, Cali 6.6 Costa Rica 6.0 India, Bombay 3.7 Kuwait, Kuwaitis 3.5

B

0

5

10 15 20 25 30 35

US, Detroit, Black 27.9 US, Los Angeles, Black 26.5 US, Connecticut, Black 25.2 US, Hawaii, Japanese 22.6 US, Connecticut, White 21.6 Canada 21.3 Australia, New South Wales 21.0 Singapore, Chinese 21.6 US, Los Angeles, Other White 20.6 US, Los Angeles, Japanese 20.6 Germany, Saarland 20.4 Ireland, Southern 20.3 Israel, All Jews 19.9 Denmark 19.9 UK, Scotland 19.4 Netherlands, Eindhoven 19.1 Norway 19.1 France, Bas-Rhin 18.8 China, Hong Kong 18.8 Italy 18.7 US, Hawaii, Hawaiian 16.6 New Zealand, Maori 16.0 UK, England & Wales 15.9 Sweden 15.9 Japan, Miyagi 15.7 Switzerland, Geneva 15.4 Germany 13.9 Spain 13.0 US, Los Angeles 12.3 US, Puerto Rico 12.1 China, Shanghai 10.8 Slovenia 10.1 Poland, Cracow 8.6 Costa Rica 6.5 Colombia 6.3 Kuwait, Kuwaitis 4.8 Singapore, Indian 4.7 India, Bombay 3.0

Figure 123-1.  Age-standardized incidence of colon cancer per 100,000 population in various populations for men (A) and women (B). (Data from Parkin DM, Whelen SL, Ferlay J, et al. Cancer incidence in five continents. [IARC Sci. Publ. No. 143]. Lyon: International Agency for Research on Cancer; 1997; and Curado MP, Edwards SB, Shin HR, et al. Cancer incidence in five continents, vol. IX [IARC Sci. Publ. No. 160]. Lyon: International Agency for Research on Cancer; 2007.)

factor of almost 90 between areas with the extreme lowest and highest rates; the incidence of rectal cancer (cancer within 11  cm of the anus) differs by a factor of 13. CRC incidence also differs within countries, depending on region and population (Fig. 123-2). These differences most likely are due to differences in environmental factors, including dietary patterns (discussed later). Although the incidences of colon and rectal cancer overall are parallel, geographic variation is more pronounced for colon than for rectal cancer. High ratios of colon to rectal cancer (≥2 : 1) prevail in high-risk areas such as North America, whereas ratios below unity are often found in lowrisk Asian and African populations. There is a steeper rise in the incidence of colon cancer for each unit increase in the incidence of rectal cancer in women compared with men, suggesting that colon and rectal cancer have related, but not identical, causes. In the United States, the incidence of CRC also varies regionally. In general, rates in the southern and western United States (except the San Francisco Bay area and Hawaii) are lower than the U.S. average, whereas rates are highest in the northeastern and north central states. CRC incidence rates also are moderately higher for urban residents, although socioeconomic status is not a consistent risk

factor for CRC in studies of the U.S. population. These regional differences in the United States have persisted over the long term, but gradually they are fading, perhaps because of the increasing homogeneity of dietary patterns across the country. Between 1950 and the mid-1980s, the incidence of colon cancer in the United States rose in the white population, whereas that of rectal cancer remained fairly stable. Mortality rates from CRC were stable among white men but decreased in white women. Both incidence and mortality rates for CRC increased substantially among the nonwhite population during this period. CRC incidence and mortality have declined since 1985 in American adults at an average annual rate of 1.6% and 1.8%, respectively2; these trends are more evident in whites than blacks. Overall death rates from CRC declined between 1990 and 2004 by almost 30% in men and 25% in women.1 Currently, incidence and mortality rates for CRC are higher in the African American population compared with the white population, and these rates in the U.S. Latin American population are slightly lower.1 The risk of CRC rises rapidly in populations that migrate from areas of low risk to areas of high risk. This pattern was demonstrated clearly in Japanese immigrants to Hawaii and

Chapter 123  Colorectal Cancer 0 Quidong Tianjin

10

20

30

40

50

0

20

40

80

100

All Jews

Colon Rectum

Jews born in Israel

Shanghai Hong Kong Singapore

Jews born in America or Europe Jews born in Africa or Asia

Hawaii

Colon

San Francisco

A

60

Los Angeles

Rectum

Non-Jews

B

Figure 123-2.  Age-standardized incidence of colon and rectal cancers per 100,000 (1988-1992) in ethnic Chinese (A) and populations in Israel (B). Colorectal cancer incidence differs considerably within countries, depending on region and population, and in ethnic groups that migrate to areas with different diets and lifestyles. (Data from Parkin DM, Whelen SL, Ferlay J, et al. Cancer Incidence in Five Continents. [IARC Sci. Publ. No. 143]. Lyon: International Agency for Research on Cancer; 1997.)

to the continental United States during the 1950s and 1960s. Cancer rates for Issei (the migrating generation) rose over a short period to exceed those of native Japanese living in Japan, and the incidence rates for Nissei (their U.S.-born offspring) rose progressively to approximate those of the native white population. A similar upward displacement of CRC risk was noted in Europeans who migrated to Australia after World War II and in Jews who migrated to Israel from low-risk areas in Yemen and North Africa. Longitudinal studies reveal that in many countries where CRC mortality rates were low before 1950, rates have increased sharply, whereas in countries where rates were high or moderate, they have decreased, stabilized, or increased slightly. Japan is a good example of this change: Once a low-risk region for colon cancer, incidence rates have risen to equal or exceed those in North America and Europe.3 Studies of temporal trends by subsite location of large bowel cancer demonstrate that for both sexes, incidence rates have increased for cancers of the right colon (cecum, ascending colon) and sigmoid colon and have decreased for cancers of the rectum; this change might reflect differing susceptibilities to neoplastic transformation in the proximal and distal colon. Currently in the United States, the prevalence of CRCs in whites is higher in the cecum and ascending colon (22% in men, 27% in women) and in the sigmoid colon (25% in men, 23% in women) than elsewhere in the large bowel (Fig. 123-3). Descriptive epidemiology, including the study of temporal trends in CRC incidence, has played an important role in formulating hypotheses about the causes and pathogenesis of these lesions. Alterations in the subsite location of these tumors also have implications for clinical cancer detection; for example, the proportion of tumors beyond the reach of the sigmoidoscope increases with age. Subsite distribution also can differ according to race. These issues are discussed in subsequent sections.

ETIOLOGY Inter-regional differences in the incidence of CRC, including differences among population groups living in geographic proximity but with different lifestyles, strongly suggest that

Transverse colon 15%

Ascending colon and cecum 25%

Descending colon 5%

Sigmoid colon 25% Rectosigmoid colon 10%

Rectum 20% Figure 123-3.  Distribution of colorectal cancers within the colon.

environment plays a role in the development of this disease. Migrant studies and rapid changes in incidence in countries assimilating Western practices support this concept. Strong circumstantial evidence exists for a link between diet and CRC. Population studies and animal studies have attempted to delineate the effects of various fats and proteins, carbohydrates, vegetable and fiber components, and micronutrients on the genesis of cancer of the large bowel (Table 123-1).

FAT, BILE ACIDS, AND BACTERIA

Several lines of evidence suggest that diets containing large percentages of fat predispose to CRC, especially in the descending and sigmoid colon. Colon cancer rates are high in populations whose total fat intake is high and are lower in those who consume less fat. On average, fat (saturated plus unsaturated) constitutes 40% to 45% of total calorie intake in Western countries with high rates of CRC, whereas

2193

2194

Section X  Small and Large Intestine Table 123-1  Factors That May Influence Carcinogenesis in the Colon and Rectum Probably Causative High-fat and low-fiber diet (adjusted for energy intake)* Red meat consumption Possibly Causative Beer and ale consumption (especially for rectal cancer) Cigarette smoking Diabetes mellitus Environmental carcinogens and mutagens Heterocyclic amines (from charbroiled and fried meat and fish) Low dietary selenium Probably Protective Aspirin, NSAIDs, and cyclooxygenase-2 inhibitors Calcium Hormone replacement therapy (estrogen) Low body mass Physical activity Possibly Protective† Carotene-rich foods High-fiber diet Vitamins C and E Vitamin D Yellow-green cruciferous vegetables *Dietary fats and fiber are heterogeneous in composition, and not all fats or fiber components play a role in cause or protection. † Based on limited data. NSAIDs, nonsteroidal anti-inflammatory drugs.

in low-risk populations, fat accounts for only 10% to 15% of dietary calories. Case-control and cohort studies also have suggested that the incidence and mortality rates from colon cancer, and in some cases rectal cancer, are positively correlated with dietary fat, but these findings are less convincing than data from descriptive epidemiologic studies. Early trials also often failed to take into account total energy intake. A prospective study assessed the relationship of meat, fat, and fiber intake among 88,751 women aged 34 to 59 years. After adjustment for total energy intake, the intake of animal fat was significantly correlated with the risk of colon cancer. The intake ratio of red meat to chicken and fish was strongly associated with increased incidence of colon cancer, perhaps owing to differences in their fat composition. Similar findings have been reported that correlate the intake of saturated fat and the ratios of red meat to chicken and fish intake with both the incidence and recurrence of colorectal adenomas in women. Cohort studies and a combined analysis of 13 case-control studies that adjusted for total energy intake, however, failed to provide clear-cut evidence for the association between dietary fat and CRC observed in earlier studies. Studies that specifically examine the association between intake of saturated or animal fat and CRC suggest a stronger association than with total fat. Animal studies lend additional support for the role of dietary fat in the development of colon cancer. These studies usually involve the injection of carcinogens such as 1,2-dimethylhydrazine (DMH) or azoxymethane into rodents fed various diets. Animals fed a variety of polyunsaturated and saturated fats develop greater numbers of carcinogeninduced colonic adenocarcinomas than do those on low-fat diets. The amount and source of dietary fat might affect tumor development in such studies; fatty acids derived from polyunsaturated fish oils (ω-3 fatty acids) and monosatu-

rated olive oil might not promote tumors to the extent that other polyunsaturated fats do. It has been proposed that dietary fat enhances cholesterol and bile acid synthesis by the liver, thereby increasing the amounts of these sterols in the colon. Colonic bacteria convert these compounds to secondary bile acids, cholesterol metabolites, and other potentially toxic metabolic compounds. Population studies demonstrate increased excretion of sterol metabolites and fecal bile acids in groups that consume a high-fat, low-fiber Western diet compared with other groups, and high fecal bile acid levels are found in some patients with CRC. Dietary fat also has been shown to increase the excretion of secondary bile acids in carcinogen-treated rats; secondary bile acids do not act as primary carcinogens but as potent promoters of colon carcinogenesis in such animal models. Little is known about how lipid and sterol metabolites promote tumors, but both bile acids and free fatty acids have been shown to damage the colonic mucosa and increase the proliferative activity of its epithelium. Dietary consumption of high amounts of corn oil and beef fat increase colonic ornithine decarboxylase levels, which are associated with rapidly proliferating mucosa. Activation of protein kinase C (PKC) by bile acids in colonic mucosa also might represent an important intracellular event by which bile acids provoke a proliferative response. Mucin alterations are a common feature of colonic neoplasia, and alterations in MUC2 mucin have been asso­ ciated with tumor progression in the colon. Bile acids induce mucin expression in human colon carcinoma cells by increasing MUC2 transcription through a process involving mitogen-activated protein (MAP) kinase-independent, PKC-dependent activation of the transcription factor AP-1.4 Bile acids can, in addition, induce release of arachidonate and conversion of arachidonic acid to prostaglandins in the mucosa, which can enhance cell proliferation.5 Preclinical and clinical evidence indicate that nonsteroidal antiinflammatory drugs (NSAIDs), which reduce prostaglandin synthesis, reduce the incidence of large bowel cancer (discussed later); inhibition of the inducible enzyme cyclooxygenase (COX)-2 may be particularly important in this regard.6 Certain fatty acids could promote carcinogenesis by altering membrane fluidity after being incorporated into cell membranes. Bacterial enzymes such as 7-α-dehydroxylase (which converts cholic to deoxycholic acid), β-glucuronidase, nitroreductase, and azoreductase may be induced by a highfat diet and also could convert compounds ingested in the diet to active carcinogens (see later). An inverse relationship has been reported between physical activity and risk for CRC in men; obesity is associated with elevated risk of CRC. Serum cholesterol and β-lipoprotein levels have been positively correlated with the development of colorectal adenomas and CRC, but this association has not been demonstrated consistently, and serum cholesterol levels can decline before the development of colon cancer. Adiponectin is a hormone secreted by adipose tissue, serum levels of which are inversely correlated with obesity and hyperinsulinemia. Variants of adiponectin and adiponectin receptor genes have been correlated with differences in CRC risk.7

FIBER

Epidemiologic, case-control, and animal studies suggest that dietary fiber protects against the development of colon cancer. Dietary fiber is plant material that resists digestion and that is composed of a heterogeneous mix of carbohydrates (e.g., cellulose, hemicellulose, pectin) and noncarbo-

Chapter 123  Colorectal Cancer hydrates (e.g., lignin). Although the protective role of fiber is not completely clear, epidemiologic studies correlate high-fiber intake with a lower incidence of colorectal neoplasia.8,9 The majority of observational-epidemiologic and case-control studies support the protective effect of fiberrich diets; however, these data do not define the relationship between fiber-rich food and the importance of nonfiber vegetable components, nutrients, and micronutrients in fruits and vegetables. The effect of fiber components on different portions of the large bowel also can vary. This might explain, in part, the inability to demonstrate a protective effect of fiber in several randomized, controlled trials that have examined the ability of fiber supplementation to prevent adenoma recurrence.10,11 Two large prospective observational studies, including one from the Prostate, Lung, Colon and Ovarian Cancer Trial (PLCO) found that increased fiber intake is significantly associated with reduced risk of colorectal neoplasia.8 Another analysis of 13 prospective cohort studies included in the Pooling Project of Prospective Studies of Diet and Cancer demonstrated that dietary fiber intake was inversely associated with risk of CRC in age-adjusted analyses, but that after accounting for other dietary factors, high intake of dietary fiber was not associated with a reduced risk of CRC.12 Investigators postulate that fibers such as cereal bran exert their protective role by increasing stool bulk, thereby diluting carcinogens and promoters of carcinogenesis, enhancing their elimination, and minimizing their duration of mucosal contact by decreasing intestinal transit time. Increased fiber intake, in the form of whole wheat and rye bread, also reduces the concentration of fecal secondary bile acids and fecal mutagens in healthy subjects. Animal studies also have demonstrated a decreased incidence of colonic tumors in DMH-treated rats fed diets high in fiber and fiber components, such as wheat bran, cellulose, and hemicellulose. Cellulose and hemicellulose decrease the levels of bacterial metabolic enzymes, such as β-glucuronidase, in experimental animals and can diminish the activation of carcinogens or cocarcinogens. Furthermore, some fiber components can bind to toxic or carcinogenic substances, perhaps decreasing their contact with the colonic mucosa. Fiber components also are fermented by fecal flora to short-chain fatty acids, thereby decreasing colonic pH and potentially inhibiting carcinogenesis.

CARCINOGENS AND FECAL MUTAGENS, VITAMINS, AND MICRONUTRIENTS

The possibility that specific genotoxic carcinogens might play a role in the genesis of CRC was raised when it was noted that the stools of certain persons exhibited mutagenic activity for bacteria in vitro. Mutagenic activity often is present in the feces of populations at high risk for large bowel cancer and is low or absent in low-risk populations. A specific group of highly unsaturated reactive compounds synthesized by colonic bacteria, fecapentaenes, might play a role in large bowel carcinogenesis. It also has been recognized that “charbroiled” meat and fish, and to a lesser extent fried foods, contain powerful mutagenic compounds. The structures of these compounds resemble a heterocyclic amine known to cause colon cancer in rodents. Metabolites similar to those of fried meat and fish are being sought as mutagens in human stools. A possible association between rectal cancer and beer- and ale-drinking also has been noted. A two-fold-to-three-fold increase in CRC also has been observed in automotive pattern and model makers, but the specific carcinogenic agent in this industrial environment has not yet been identified.

The exact nature of genotoxic carcinogens that might act in the human colon remains speculative, but the identification of such compounds could provide a basis for intervention and primary prevention of CRC. Limited data suggest that foods rich in carotene (vitamin A) and vitamin C could act as antioxidants in the chemoprevention of colon cancer, but prospective trials have failed to demonstrate such an effect. Other areas that merit further exploration in the prevention of CRC include the role of yellow-green cruciferous vegetables and the role of micronutrients including selenium salts, vitamin E, and folic acid. A good deal of attention has been given to a possible role for dietary calcium in preventing colon cancer (discussed next).

CALCIUM AND VITAMIN D

Epidemiologic, clinical, and laboratory evidence suggest that calcium intake might protect against carcinogenesis in the colon. Calcium has numerous biological effects that might reduce colon carcinogenesis, including actions on the cell cycle, cyclic adenosine monophosphate (cAMP), calmodulin, tyrosine kinases, ornithine decarboxylase, and e-cadherin. The calcium-sensing receptor expressed in the intestine senses extracellular calcium with effects on differentiation and proliferation. The potential chemopreventive activity of calcium was suggested originally by epidemiologic studies reporting an inverse relationship between CRC and intake of vitamin D and calcium. Dietary calcium supplementation in the form of low-fat dairy foods can affect a variety of intermediate biomarkers thought to be associated with tumor progression in the colon, and supplemental calcium plus vitamin D alters preneoplastic features of colorectal adenomas.13 Although the relationship between calcium intake and a lower incidence of colonic adenomas and carcinomas has not been uniformly demonstrated, observational studies overall suggest a protective effect. A pooling project of 10 cohort studies strongly suggested that calcium intake is inversely related to the risk of CRC.14 Further support for the beneficial effect of calcium in preventing large bowel cancer comes from numerous animal studies. Abnormal proliferation occurs in neoplastic and preneoplastic lesions in the colon. The increase in colonocyte proliferation stimulated by intrarectal instillation of deoxycholate and free fatty acids or by dietary supplementation with cholic acid may be ameliorated by oral calcium supplementation in laboratory animals. Studies in rodents fed high-fat diets also demonstrate a reduction in the number of carcinogen-induced tumors in animals receiving supplemental calcium in their diet. This reduction may be true especially for tumors containing K-ras mutations, a finding also suggested by a study in humans. Ornithine decarboxylase, an enzyme involved in polyamine biosynthesis and elevated in preneoplastic states, is reduced in rat colonic mucosa incubated with calcium in vitro, and supplemental calcium suppresses elevated levels of this enzyme in the mucosa of elderly patients with adenomatous polyps. It has been suggested that dietary calcium binds to ionized fatty acids and bile acids in the intestine, converting them to insoluble calcium compounds incapable of stimulating epithelial proliferation. Calcium increases fecal excretion of both phosphate and bile acids and modifies relative amounts of bile acids in bile. In addition, calcium in milk products is capable of precipitating luminal cytotoxic surfactants, thereby inhibiting their effects on colonic mucosa. These potential beneficial effects of calcium have not been observed uniformly, however, and studies of the effects of calcium on the rectal mucosa have not always demonstrated a reduction in the proliferation rates. In other studies, calcium

2195

2196

Section X  Small and Large Intestine supplementation normalized the distribution of proliferating cells in the colonic crypt without affecting the rate of proliferation in the colorectal mucosa. Vitamin D3 metabolites and analogs have been shown to play an important role in the regulation of a number of important cellular processes, including proliferation, differentiation, and apoptosis, in addition to their established role in mineral homeostasis. These steroid compounds have rapid effects that do not involve gene transcription or protein synthesis, as well as genomic effects involving the vitamin D receptor and other transcription factors.15,16 Vitamin D can modulate more than 200 genes involved in cell cycle regulation, growth factor signaling, protection against oxidative stress, bile acid and xenobiotic metabolism, cell adhesion, DNA repair, angiogenesis, inflammation, and immune function. Effects of vitamin D and its metabolites have been demonstrated in normal and malignant colonocytes, and several potential mechanisms have been suggested by which these compounds might prevent carcinogenesis in the colon. Dietary supplementation with calcium and vitamin D in rodents fed colon tumor-inducing Western diets significantly reduced tumor incidence and multiplicity in addition to altering expression of a variety of genes linked to initiating formation of colon tumors.17

ARACHIDONIC ACID, EICOSANOIDS, AND CYCLOOXYGENASE-2

Clinical case-control and cohort studies have shown a 40% to 50% reduction in CRC-related mortality in persons taking aspirin and other NSAIDs on a regular basis compared with those not taking these agents. The exact mechanism for cancer protection with these agents is unknown, but it might relate to altered synthesis of arachidonic acid metabolites (eicosanoids) including prostaglandins, thromboxanes, leukotrienes, and hydroxy-eicosatetraenoic acids. These compounds modulate a number of signal transduction pathways that affect cellular adhesion, growth, and differentiation. COX (or prostaglandin-endoperoxide synthase) oxidizes arachidonic acid to prostaglandin G2, reduces prostaglandin G2 to prostaglandin H2, and is the key enzyme responsible for production of prostaglandins and other eicosanoids. This enzyme exists in two isoforms: COX-1 and COX-2. COX-1, the constitutive form of the enzyme, is present in most tissues and is involved in the physiologic production of prostaglandins to maintain normal homeostasis. COX-2 is induced by cytokines, mitogens, and growth factors, and its level has been shown to be elevated in both murine and human CRCs.18-22 Expression of COX-2 is increased markedly in 85% to 95% of CRCs and in experimental models of CRC. COX-2 inhibition prevents cancer from developing during the initiation and the promotion and progression stages of carcinogenesis.21 Knockout of COX-2 results in suppression of intestinal polyposis in animal models of familial adenomatous polyposis (FAP).19 15-Hydroxy­ prostaglandin dehydrogenase (15-PGDH) is a prostaglandindegrading enzyme that is lost in human colon cancers and has been shown to be a physiologic antagonist of the prostaglandin-synthesizing activity of COX-2.22 It has been speculated that NSAIDs might reduce formation of colon tumors by inhibiting prostaglandin-mediated proliferation, but other evidence suggests that part of their effect might result from inducing apoptosis. Overexpression of COX-2 has been demonstrated to decrease apoptosis, whereas inhibition of COX-2 leads to an increase in apoptosis. One potential mechanism by which NSAIDs can induce apoptosis is through elevating the prostaglandin precursor

arachidonic acid. Increases in arachidonic acid after NSAID inhibition of COX stimulate conversion of sphingomyelin to ceramide, which is a mediator of apoptosis. NSAIDs also might inhibit the activation of genes by the nuclear hormone receptor peroxisome-proliferator-activated receptor δ (PPARδ) by disrupting the ability of this receptor to bind DNA. PPARδ expression is elevated in CRCs and repressed by the APC gene product, which is altered in CRC cells. The inhibition of PPARδ function enhances the ability of NSAIDs to induce apoptosis in colon cancer cells. PPARδ activates a variety of genes, including those involved in cellular growth and differentiation after exposure to a variety of ligands, such as eicosanoids. COX-2 inhibition could prevent production of these ligands and thereby prevent activation of PPARδ. Other potential mechanisms by which COX-2 inhibition might affect tumor formation include alterations of cell adhesion to extracellular matrix proteins, inhibition of tumor neovascularization (angiogenesis), and reduction in carcinogen activation. A study using the ApcΔ716 mouse, an animal model of FAP, demonstrated that treatment with the COX-2 specific inhibitor rofecoxib (Vioxx) was associated with a significant dose-dependent reduction in size and number of polyps as well as alterations in polyp morphology. COX-2 inhibition was associated with decreased levels of vascular endothelial growth factor (VEGF) and with lower rates of DNA replication.20 In summary, environment and diet might affect the genesis of CRC, but their exact roles remain unclear. Their complex nature renders definition of the influence of individual environmental and dietary components difficult.

CHEMOPREVENTION

Chemoprevention refers to the use of natural or synthetic agents to reverse, suppress, or prevent progression or recurrence of cancer23,24; this is a cornerstone of primary prevention. Data on chemoprevention of CRC come from studies in laboratory animals (see earlier), observational epidemiologic studies, case-control studies, and randomized controlled trials (RCTs). Because the natural history of CRC is protracted, clinical RCTs often have concentrated on preventing colorectal adenomas, which represent a form of intraepithelial neoplasia and are the precursors to carcinoma. The duration of the studies required, sample sizes necessary, cost, and ethical considerations make the use of cancer as an end point impractical. This difficulty has led to an increasing use of surrogate biomarkers to study chemoprevention of CRCs,25 with the hope that use of such markers will lead to shorter, smaller, and less expensive trials.26 To be valid, however, such biomarkers need to accurately represent the events involved in the process of carcinogenesis. When an intervention such as a chemopreventive agent is tested, there should be a clear relationship among the agent, modulation of the biomarker, and the development of cancer. Surrogate end points for cancer ideally should be validated in the context of clinical studies that use cancer as the ultimate end point. This is a difficult task because these are the very trials that such markers are designed to complement or replace. There has been interest in the use of magnifying endoscopy to study aberrant crypt foci of the colon as possible markers in chemoprevention trials.24 These foci consist of large, thick crypts that can be detected by chromoen­ doscopy using agents such as methylene blue or indigo carmine (Fig. 123-4). Aberrant crypt foci, particularly large crypts with dysplastic features, are thought to be precursors of adenomas in the colon. Standardization of techniques to identify and quantify these lesions is crucial to the success-

Chapter 123  Colorectal Cancer

A

B

C Figure 123-4.  A-C, Three magnification views of aberrant crypt foci at different magnifications. Aberrant crypt foci consist of large thick crypt, and are thought to be precursors of adenomas in the colon.

ful interpretation of data from these trials, because it is unclear whether these lesions, which appear to be precursors to neoplasia in animal models, play a similar role in the human colon.27 Studies suggest that the majority of aberrant crypt foci in the human colon may be hyperplastic rather than dysplastic (see Chapter 122). The potential benefit of low-fat, high-fiber diets based on descriptive epidemiology and case-control studies already has been discussed, but current data from prospective human trials are thus far equivocal or negative. Two large RCTs examined the effects of fiber supplementation on recurrence of adenomas. The Polyp Prevention Trial11 randomized 2079 subjects with a history of colorectal adenomas to receive counseling together with a low-fat, high-fiber diet rich in fruits and vegetables or to receive their usual diet alone. The incidence of recurrent adenomas at one and four years, as determined by colonoscopy, was similar in both groups. In a study conducted by the Phoenix Colon Cancer Prevention Physicians’ Network,10 1429 patients with a history of colorectal adenoma were randomized to receive 2.0 g or 13.5 g of supplemental wheat bran per day. Colonoscopy failed to show a difference in the incidence of recurrent adenomas at a median follow-up of 34 to 36 months. A large body of observational and laboratory studies suggests a role for dietary calcium supplementation in chemoprevention. A prospective double-blind placebo-controlled trial showed that supplemental calcium (3000 mg of calcium carbonate per day, equivalent to 1200  mg of elemental calcium) reduced the incidence and number of recurrent adenomas in subjects chosen for a recent history of such lesions.26 The effect of calcium was modest: 19% reduction in recurrence of adenomas and 24% reduction in the number of adenomas over three years, independent of age, sex, or dietary intake of calcium, fat, or fiber. The protective effect

of calcium supplementation on the risk of colorectal adenoma recurrence extended up to five years after cessation of active treatment, even in the absence of continued supplementation.28 Analysis of subjects’ serum vitamin D status suggested that calcium supplementation and vitamin D status appear to act together to reduce the risk of adenoma recurrence.29 The results of a Japanese study, the Fukuoka Colorectal Cancer Study,30 support the joint action of calcium and vitamin D in preventing CRC. Human trials using antioxidant vitamins A, C, and E have provided equivocal results, and current data do not support their routine use for colon cancer prevention in average-risk persons. Folic acid and its metabolites play an important role in DNA synthesis, strand integrity, and methylation. Epidemiologic studies have found a lower incidence of CRC among those with high compared with low dietary intake of folate.31 This protective effect also was suggested by the Nurses’ Health Study, in which high doses of folate (as part of multivitamin supplementation) given over several years were protective against CRC. A large prospective RCT32,33 failed, however, to demonstrate a protective effect of 1 mg/ day of folate supplementation on recurrence of adenoma compared with placebo and suggested that folate supplementation in persons with prior adenomas actually might increase adenoma risk. Analysis of baseline dietary and serum folate levels in these subjects supports the idea that although moderate doses of folate may be protective compared with deficiency, at some point of sufficiency, supplementation provides no additional benefit.34 The lack of efficacy of folate (0.5  mg/day) supplementation to prevent adenoma recurrence also was found in another RCT.35 Epidemiologic, case-control, and prospective cohort trials suggest a protective effect against the development of CRC

2197

2198

Section X  Small and Large Intestine in women taking hormone (estrogen) replacement therapy. It has been postulated that estrogen might protect against colon cancer by decreasing production of secondary bile acids, by decreasing levels of insulin-like growth factor-1, or through as yet undetermined direct effects on colonic mucosal epithelial cells. Data from the Women’s Health Initiative demonstrated that short-term use of estrogen plus progestin was associated with a decreased risk of CRC; however, CRCs in women who took estrogen plus progestin were diagnosed at a more advanced stage than those in women who took placebo.36 Cigarette smoking has been associated with incidence of and mortality from CRC in observational studies,37 but the long-term effects of smoking cessation on CRC have not been studied. The most promising results for CRC prevention come from trials using aspirin and NSAIDs. Case-control and cohort studies have suggested that the risk for development of adenoma and carcinoma may be reduced substantially (40% to 50%) among aspirin and NSAID users, compared with controls.6,24,38 A prospective cohort study among male health professionals demonstrated that persons who take aspirin more than twice per week were at lower risk for CRC (relative risk [RR], 0.68) than were controls. An RCT that assessed the effect of low-dose aspirin in an average-risk population demonstrated no significant reduction in the number of CRC cases during the first six years of follow-up; longer follow-up may be necessary to demonstrate a significant effect of aspirin on development of cancer, because the Nurses’ Health Study demonstrated that the benefits of aspirin might not be evident until after at least a decade of regular aspirin consumption.6 Three prospective adenoma prevention trials (see later) now provide compelling evidence that aspirin use reduces the risk of developing colorectal adenoma in persons with a personal history of adenoma or carcinoma.39-41 Given the long natural history of CRC, preventing recurrence of adenoma after endoscopic removal often is used as an intermediate or surrogate end point in chemoprevention trials.24 In FAP, in which hundreds of adenomas occur in the colon and rectum, chemoprevention trials often use reductions of the number and size of adenomas as end points in short-term studies. Such trials have suggested a potential role for NSAIDs as chemopreventive agents in this setting. There is a significant decrease in the mean number and size of polyps in patients treated with the NSAID sulindac. In a small, double-blind RCT of 22 patients with FAP, treatment with sulindac reduced the number and mean diameter of colorectal polyps by 44% and 35% of respective baseline values after nine months; three months after treatment was stopped, however, the number and size of polyps had increased. A subsequent prospective cohort study42 confirmed that long-term use of sulindac is effective in reducing the number of adenomas in patients with FAP. A 76% reduction in the number of polyps was seen at one year and was sustained (74% reduction) through 63 months of follow-up. Sulindac might not be effective, however, for primary prevention of development of adenomas in genetically disposed persons with FAP who have not yet developed macroscopic adenomas.43 A double-blind RCT studied the effects of celecoxib (Celebrex), a selective COX-2 inhibitor, on colorectal polyps in patients with FAP.44 Treatment with high doses of celecoxib for six months was associated with a significant reduction (28%) in the number of colorectal polyps compared with placebo (4.5%). This drug is now approved in the United States as an adjunct to standard therapy in patients with FAP.

Data are now available on the role of nonselective NSAIDs and COX-2 inhibitors as chemopreventive agents for preventing sporadic adenomas. Four published trials have demonstrated a reduction in recurrence of adenoma in chemoprevention trials involving aspirin. An RCT of aspirin versus placebo40 demonstrated that daily use of 81  mg of aspirin was associated with a 19% reduction in occurrence of adenoma and a 41% reduction in occurrence of advanced neoplasms (adenomas measuring at least 1 cm or with tubulovillous or villous features, severe dysplasia, or invasive cancer) compared with results of placebo treatment three years after removal of an index sporadic adenoma. In a similar trial, aspirin (300  mg/day) was associated with a 21% overall risk reduction for all adenomas and a 37% reduction in advanced adenomas at three years.35 A prospective RCT in a higher-risk group for adenoma recurrence, those with a previous history of sporadic CRC, demonstrated a 45% risk reduction in those taking 325  mg of aspirin daily compared with those taking placebo with a mean follow-up of almost three years.39 Given biological plausibility, preclinical in vitro and animal data, and data on regression of adenoma in patients with FAP, three randomized trials were undertaken to examine the effect of COX-2 selective inhibitors on formation of new adenomas in patents with a history of sporadic adenomas. The Adenoma Prevention with Celecoxib (APC) trial45 randomized 2035 subjects to celecoxib at a dose of 200 mg or 400 mg twice daily. Use of celecoxib was associated with a dose-dependent 33% to 45% reduction in the development of new adenomas by three years, with a 57% to 66% reduction in the number of patients developing advanced adenomas. The Prevention of Sporadic Adenomatous Polyps (PreSAP) trial46 randomized 1561 patients at a ratio of two to three to receive 400 mg celecoxib once daily or placebo. Use of celecoxib was associated with a 36% overall reduction in formation of new adenomas and a 51% reduction in advanced adenomas over three years. The Adenomatous Polyp Prevention on Vioxx (APPROVe) Trial47,48 was a double-blind RCT of the efficacy of oral rofecoxib, 25  mg daily, to prevent colorectal adenomas in 2587 subjects. The risk of adenoma recurrence over three years was lower for subjects on rofecoxib than on placebo, with a 24% overall reduction and a 30% reduction in advanced adenomas. The effect of drug was more pronounced in the first year (RR 0.65; 95% confidence interval [CI]: 0.57-0.73) than in the subsequent two years. Thus, three well-conducted prospective RCTs demonstrated a significant reduction in formation of new adenomas with the use of a COX-2 selective inhibitor in those with a history of colorectal adenomas. Unfortunately, adverse thrombotic cardiovascular events were associated with COX-2 inhibition in at least two of these trials. Data suggest that an increased risk of cardiovascular events may be associated with most NSAIDs, and not just COX-2 inhibitors. Other agents currently undergoing study for chemo­ prevention of colorectal neoplasia include the ornithine decarboxylase inhibitor difluoromethylornithine (DFMO); the bile acid ursodiol; 3-hydroxy-3-methylglutarylcoenzyme A (HMG-CoA) reductase inhibitors, such as pravastatin and lovastatin; epidermal growth factor receptor (EGFR) inhibitors; and matrix metalloproteinase (MMP) inhibitors.24 The combination of DFMO and the NSAID sulindac were studied in an RCT to assess their efficacy in preventing sporadic adenoma recurrence in 375 subjects. Use of this regimen was associated with a 70% reduction in new adenomas at three years compared with placebo.49 Larger studies are needed to confirm this result and to fully

Chapter 123  Colorectal Cancer Table 123-2  Efficacy of Chemoprotective Agents for Colorectal Neoplasia OBSERVATIONAL STUDIES OF COLON CANCER INCIDENCE AGENT

RANDOMIZED HUMAN TRIALS

Animal Studies

Case Control

Cohort Studies

Reduction in Mucosal Proliferation

+ +

+ NA

+ NA

NA +

+

+ + + + + + NA NA

+ + + +

+ +

Aspirin or other NSAID Cyclooxygenase-2 inhibitors Vitamins A, C, E Folate Calcium Fiber Selenium Fish oil Organosulfur Difluoromethylornithine

∼

+ + + NA + +

∼

∼

NA NA NA

+ NA NA NA +

Reduced Number of Polyps in Patients with FAP

Reduced Number of Sporadic Adenomas

+ +

+ +

∼

∼

NA NA

∼

− + NA +

− + − NA NA NA +*

*In combination with sulindac. +, Most studies are positive for efficacy; −, most studies are negative for efficacy; ∼, studies are equivocal for efficacy; FAP, familial adenomatous polyposis; NA, not available; NSAID, nonsteroidal anti-inflammatory drug.

Folic acid Vit.C + vit.E Low fat Low fat Low fat

DFMO+sulindac Celecoxib 1.6 Celecoxib 1.4 Celecoxib 1.2 1 0.8 0.6 0.4 0.2 0

Low fat Calcium +vit.D3 Fruits and vegetables Beta-carotene Beta-carotene Wheat bran

Rofecoxib Calcium Calcium Aspirin Aspirin Aspirin

Ursodeoxycholic acid Wheat bran Wheat bran Figure 123-5.  Diagram depicting the relative risk of colorectal adenoma recurrence after a clearing colonoscopy in different colorectal adenoma prevention studies using various potential chemopreventive agents. DFMO, difluoromethylornithine; vit., vitamin. (Courtesy of Asad Umar, DVM, PhD, Bethesda, Md.)

assess the toxicity of this combination. A population-based case-control study of persons with CRC and matched controls demonstrated a 47% relative reduction of CRC associated with statin use, but further investigation is needed to assess this benefit.50 Table 123-2 and Figure 123-5 summarize the status of current studies that examine the effect of chemopreventive agents on colorectal neoplasia.

BIOLOGY Current concepts concerning environmental causes for CRC have been discussed in the preceding sections. It has been suggested that carcinogens introduced into the bowel act in

concert with other luminal factors (e.g., bile acid and other tumor promoters) to affect epithelial cells in the colonic mucosa; however, carcinogenesis is a multistage process. Cells must be genetically primed (through either hereditary disposition or genotoxic events) and induced to proliferate, after which they must pass through a series of stages en route to immortalization and uncontrolled growth. Our knowledge of this sequence of events in the colon, although fragmentary, is growing rapidly (see Chapter 3).

ABNORMAL CELLULAR PROLIFERATION

Abnormal cellular proliferation is a hallmark of neoplasia (see Chapter 3). Actively proliferating cells are more susceptible to initiators of carcinogenesis (primary carcinogens) and genetic alterations than are resting cells. In the normal colon, DNA synthesis occurs and cells divide and proliferate only in the lower and middle regions of the crypts. As cells normally migrate upward from deeper in the crypt, the number of cells that continue to proliferate decreases, and, on reaching the upper crypt region, cells become terminally differentiated and can no longer divide. A substantial body of literature indicates that this sequence of events is disordered during the evolution of neoplastic lesions in the colon. Increased proliferative activity and characteristic differences in the distribution of tritiated thymidine-labeled cells (i.e., those that actively synthesize DNA) within the colonic crypts have been demonstrated and distinguish at-risk and affected members of kindreds with familial poly­ posis, as well as nonpolyposis-inherited colon cancer from lower-risk groups. Correlations between rectal mucosal proliferation and the clinical and pathologic features of nonfamilial colorectal neoplasia also have been demonstrated. Conversely, populations at low risk for developing colon cancer, such as Seventh Day Adventist vegetarians, have relatively quiescent proliferative activity in their colonic mucosa. Disordered proliferative activity can be found in the colonic mucosa of rodents treated with a variety of chemical carcinogens, and increased proliferative activity is seen in animals whose colonic or rectal mucosa is exposed to tumor promoters such as secondary bile acids. Colonic epithelial cells also fail to repress DNA synthesis during epithelial renewal in ulcerative colitis (UC), a condition associated with an increased risk of CRC. Ornithine decarboxylase, an

2199

2200

Section X  Small and Large Intestine enzyme marker of rapid cellular proliferation, is present at high levels in the mucosa of members of familial polyposis kindreds, and levels increase in the colonic mucosa during chemically induced colonic carcinogenesis in rats. Ornithine decarboxylase increases in the colonic mucosa with age and is elevated in elderly patients with colonic adenomas. Experimental evidence also suggests that PKC may be involved in the stimulation of colonic epithelial cell pro­ liferation by tumor promoters. These findings have led to speculation that inhibitors of cellular proliferation might prove useful as anticancer drugs. NSAIDs and COX-2 in­­ hibitors, for example, decrease proliferation and increase apoptosis. An explosion of knowledge in the field of molecular genetics has demonstrated how alterations in protooncogenes and tumor suppressor genes might lead to disruption of mechanisms that regulate the normal cell cycle and cell proliferation. In some cases, the cell is predisposed to abnormal proliferation by germline mutations, such as FAP, whereas in other cases, somatic mutations occur as the result of complex interactions with environmental factors,51 as detailed earlier.

MOLECULAR BIOLOGY AND BIOCHEMICAL CHANGES Molecular Genetics

Tumor cells in the colon, as elsewhere, are characterized by heritable phenotypic changes that are the result of quantitative or qualitative alterations in gene expression (see Chapters 3 and 122). A large body of evidence demonstrates that CRCs are associated with an accumulation of such genetic

alterations (Table 123-3; Fig. 123-6). Genetic changes that might lead to the development of CRC traditionally have been categorized into three major classes: alterations in proto-oncogenes, loss of tumor suppressor gene activity, and abnormalities in genes involved in DNA mismatch repair (MMR).52-54 Adenomas and carcinomas arise in the context of genomic instability, by which epithelial cells acquire the number of mutations needed to attain a neoplastic state. Destabilization of the genome is a prerequisite to tumor formation and most commonly involves chromosomal instability (CIN), which is found in 80% to 85% of colorectal tumors with subsequent allelic loss; chromosomal amplifications and translocations; or increased rates of intragenic mutation in tandemly repeated DNA sequences known as microsatellites (microsatellite instability [MSI]).54 Cellular proto-oncogenes are evolutionarily conserved human genes that contain DNA sequences homologous to those of acute transforming retroviruses. Many of these genes play a role in signal transduction and the normal regulation of cell growth. Inappropriate activation of these genes leads to abnormal transmission of regulatory messages from the cell surface to the nucleus that results in abnormal proliferation and, eventually, tumor formation. Three human ras genes—K-ras, N-ras, and H-ras—encode guanine nucleotide-binding proteins that regulate intracellular signaling pathways. Approximately 65% of sporadic CRCs have activating point mutations in a ras gene, most in K-ras. Most ras mutations appear to occur during intermediate stages of adenoma growth (see Chapter 122). Ras gene mutations occur in 47% of carcinomas, in 58% of adenomas larger than 1 cm, but in only 10% of adenomas smaller than 1 cm, suggesting that earlier events must contribute to for-

Table 123-3  Genes Altered in Sporadic Colorectal Cancer

CHROMOSOME

frequency of TUMORS WITH GENE ALTERATIONS (%)

GENE CLASS

FUNCTION of GENE product

K-ras

12

50

Proto-oncogene

APC

5

70

Tumor suppressor

DCC

18

70

Tumor suppressor?

SMAD4 (DPC4, MADH4)

18

?

Tumor suppressor

TP53

17

75

Tumor suppressor

Encodes guanine nucleotide-binding protein that regulates intracellular signaling Regulation of β-catenin that is involved in activation of WnT/TcF signaling (activates c-myc, cyclin D1)*; regulation of proliferation and apoptosis; interaction with E-cadherin (cell adhesion?) Netrin-1 receptor; caspase substrate in apoptosis; cell adhesion Nuclear transcriptase factor in transforming growth factor (TGF-β1) signaling; regulation of angiogenesis; regulator of WAF1 promoter; downstream mediator of SMAD2 Transcription factor; regulates cell cycle progression after cellular stress; regulates apoptosis, gene expression, and DNA repair Maintains fidelity of DNA replication Maintains fidelity of DNA replication Maintains fidelity of DNA replication Receptor for signaling in the TGF-β1 pathway; inhibitor of colonic epithelial proliferation, often mutated in tumors with MSI

GENE

hMSH2 hMLH1 hMSH6 TGF-β1 RII

2 3 2 3

† † † ‡

DNA mismatch repair DNA mismatch repair DNA mismatch repair Tumor suppressor

*β-Catenin mutations (downstream of APC) are found in 16%-25% of MSI colon cancers but not in MSS cancers. † Approximately 15% of sporadic colorectal cancers demonstrate MSI associated with alterations in mismatch repair genes (principally hMSH2 and hMLH1 but also hMSH3, hMSH6, hPMS1, and hPMS2). ‡ Mutated in 73%-90% of MSI colon cancers. Up to 55% of MSS colon cancer cell lines demonstrate a TGF-β signaling blockage distal to TGF-β1 RII. MSI, microsatellite instability; MSS, microsatellite stable; RII, type II receptor; TGF-β, transforming growth factor-β.

Chapter 123  Colorectal Cancer

A

B

D

C

E

Normal colon

APC hMSH2 hMLH1 abnormalities (hereditary syndromes)

F

Hyperproliferative epithelium

Methylation abnormalities

APC hMSH2 hMLH1 inactivation

Adenoma

K-ras mutation

18q deletion

Carcinoma

TP53 deletion

Further accumulation of genetic abnormalities

Figure 123-6.  Proposed sequence of molecular genetic events in the evolution of colon cancer. Carcinomas arise from an accumulation of events, the sequence of which has been defined. Alterations in APC or DNA mismatch repair genes may be inherited in the germline (familial adenomatous polyposis, hereditary nonpolyposis colorectal cancer) or may be acquired after birth (somatic mutations). Upper part, The top row (A-C) shows colonoscopic photographs and the bottom row (D-F) shows the histology. From left to right: Dysplastic aberrant crypt focus (A, with methylene blue staining; D, E, and E with hematoxylin and eosin staining). adenomatous polyp (B and E); and invasive carcinoma (C and F). (Aberrant crypt focus [A and D] reproduced from Takayama T. Katsucki S, Takahashi Y, et al. Aberrant crypt foci of the colon as precursors of adenoma and cancer. N Engl J Med 1998; 339:1277.)

mation of neoplasms. Alterations in signal transduction might lead to abnormal cell growth and thus participate in neoplastic transformation, but activation of ras alone is not sufficient for progression to carcinoma. The exact functional relationship between ras mutations and carcinogenesis remains to be established, but understanding the role of ras mutations in stimulating proliferation might lead to development of anti-tumor therapies aimed at interrupting signals that alter tumor cell growth. Chromosomal abnormalities have been reported in CRCs for more than a decade, and abundant evidence has shown that allelic losses, particularly at chromosome locations 5q, 17p, and 18q, play major roles in the genesis of large bowel

tumors.52-55 A deletion within chromosome 5 in patients with FAP led to the identification of the APC gene on the long arm of this chromosome (5q21). Positional cloning identified a single tumor suppressor gene, which is mutated in both the germline of FAP patients and in sporadic colorectal tumors. The protein encoded by APC consists of 2843 amino acids and is located at the basolateral membrane of colorectal epithelial cells; expression is increasingly pronounced as cells migrate up through the colonic crypt. Somatic mutations of the APC gene occur in 60% to 80% of sporadic CRCs and adenomas, including the smallest dysplastic lesions. These mutations result in truncation of

2201

2202

Section X  Small and Large Intestine A GSK3-β APC CK1 α/ε β-catenin

B

LRP Frizzled

LRP

Diversin Axin

C

Wnt

GBP

Frizzled Dishevelled

LRP

Frizzled APC

β-catenin

APC

Axin

LEF/TCF

X Wnt target genes

Wnt target genes

Wnt target genes

Figure 123-7.  A model of Wnt signaling in normal (A and B) and cancer (C) cells. A, In the absence of Wnt signaling, APC, axin, and GSK3-β form a complex that results in β-catenin phosphorylation and degradation by a ubiquitin-dependent mechanism. B, Binding of Wnt to its cell surface receptor results in stabilization of β-catenin. Unphosphorylated β-catenin is able to translocate to the nucleus to form a complex with members of the LEF/TCF (lymphoid enhancer factor/T-cell factor) family and activates Wnt target genes. Frizzled and Dishevelled refer to gene products that participate in this pathway. C, Loss or mutation of APC results in lack of β-catenin degradation and high levels of this protein in the cytoplasm and nucleus. Strong evidence exists that misregulation of Wnt target gene expression is crucial to transformation in colon cells. GBP, glycogen synthase kinase-3 binding protein; LRP, lipoprotein receptor-related protein. (From Waterman ML. Lymphoid enhancer factor/T cell factor expression in colorectal cancer. Cancer Metast Rev 2004; 23:41.)

the APC protein in more than 98% of cases, a finding that has led to the development of clinically useful tests for genetic screening of FAP families. Inactivation of both copies of the APC gene appears to be the gatekeeping event for the initiation of colorectal neoplasia. The APC gene product interacts with at least six other proteins, including glycogen synthetase 3β (GSK-3β) and axin in the cytoplasm. Inactivation of this gene is required for net cellular proliferation and initiation of neoplasia in the colon. APC functions to modulate extracellular signals that are transmitted to the nucleus through the cytoskeletal protein β-catenin, as part of the Wnt signaling pathway (Fig. 123-7). Nuclear β-catenin binds to transcription factors in the nucleus that are members of the lymphoid enhancer factor/ T-cell factor (LEF/TCF) family including Tcf-4, which in turn activate various target genes (e.g., c-myc, cyclin D1)56 that affect cell cycling and growth. APC is a tumor suppressor gene that binds to β-catenin and causes its degradation through phosphorylation. Loss of APC function, therefore, leads to accumulation of β-catenin and unopposed stimulation through the Wnt-Tcf signaling pathway, which in turn leads to increased and unregulated proliferation and decreased programmed cell death (apoptosis). APC gene abnormalities also might lead to disruption of normal cell-cell adhesion through altered association with the cellular adhesion molecule E-cadherin. Disruption of APCmediated regulation of transcriptional activation is critical for colorectal tumorigenesis and is achieved most commonly through inactivating mutations of both APC alleles; disruption also can occur through dominant mutations of the β-catenin gene that render β-catenin-Tcf-regulated transcription insensitive to the regulatory effects of normal wild-type APC. Other genetic changes occur later in the adenoma-tocarcinoma sequence. Stepwise tumor progression is asso­ ciated in more than 75% of cases with loss of the tumor suppressor gene activity located on chromosome 18q. Several candidate genes are present on this chromosome, and loss of chromosome 18 is associated with a poor pro­g­ nosis.57 One gene, DCC (deleted in CRC), originally was thought to be important because its loss from a stage II (Dukes B) cancer was associated with a worse prognosis in

some studies; its role as an important tumor suppressor gene has been questioned. DPC4 (SMAD4) is another candidate tumor suppressor gene, inactivation of which might play a role in development of CRC. DPC4 belongs to the SMAD gene family involved in signal transduction pathways activated through the transforming growth factor (TGF)-β family receptors. Experimental inactivation of the homologue Dpc4 in a mouse model of adenomatous polyposis coli results in malignant progression of intestinal and colonic polyps initiated by loss of the Apc gene (the mouse homologue of APC). Mutations in SMAD4 and a related gene, SMAD2, have been reported in some sporadic CRCs. Deletions of chromosome 17p involve the p53 tumor suppressor gene, the product of which normally prevents cells with damaged DNA from progressing from the G1 to the S phase in the cell cycle. Deletions within chromosome 17p are present in approximately 75% of CRCs. Loss of TP53 also may be associated with reduced apoptosis of damaged cells. Inactivation of the TP53 gene mediates the conversion from adenoma to carcinoma, a late and important event in colon carcinogenesis. Distant metastases from CRCs are associated with high fractional allelic loss and deletions of 17p and 18q. A distinct set of metastasissuppressor genes also has been postulated.58 Genomic instability creates a permissive state in which a cell acquires sufficient mutations to be transformed to a cancer cell; this is a common mechanism central to the development of most, if not all, colon cancers.53 Several forms of genomic instability are common in colon cancer, including chromosome instability (CIN) and chromosome translocations, and MSI in which subtle sequence changes, including base substitutions, deletions, or insertions, lead to a hypermutable state (Fig. 123-8). Candidates responsible for CIN include genes responsible for the human mitotic spindle checkpoint (hBUB1 and hBUBR1), genes involved in the DNA damage checkpoint (ATM, BRCA1 and BRCA2, TP53, and hRad17), and genes that control centrosome number. Inactivation of genes involved in base excision repair (BER) resulting from oxidative damage are found in a subset of CRCs. Inactivation of one of the BER genes (MYH) is a cause of an autosomal recessive form of FAP.

Chapter 123  Colorectal Cancer Chromosome unstable (CIN) pathway CIN gene mutation, e.g., BUB1 –18q, +13q, –4 +7p, +7q, +20q K-ras

APC Normal epithelium

APC

Dysplastic ACF

Early adenoma

Late adenoma

K-ras BRAF1

CTNNB1

TGFBR2 TP53 PI3CA Cancer

–8p

PRL3

Metastatic cancer

TGFBR2 BAX, MSH3, E2F4, MSH6

–13q

MMR inactivation

Number of genetic alterations

Microsatellite unstable (MSI) pathway

Time during adenoma-carcinoma progression Figure 123-8.  Model of colon cancer formation in tumors that progress through the adenoma-carcinoma sequence along pathways marked by chromosomal instability (CIN) or microsatellite instability. ACF, aberrant crypt focus; MMR, mismatch repair; TGF, transforming growth factor. (Modified from Grady WM. Genomic instability and colon cancer. Cancer Metast Rev 2004; 23:11.)

Although tumors arising as the result of MYH germline mutations demonstrate chromosomal instability, they appear to have a unique pathogenesis compared with sporadic CRC. The significance of genomic instability in the pathogenesis of a subset of colon cancers became evident with the discovery of MSI in colon cancers associated with hereditary nonpolyposis colorectal cancer (HNPCC). Alterations in genes that help maintain DNA fidelity during replication are characteristic of patients with HNPCC.53,59 Alterations in MMR genes designated hMLH1, hPMS1, hPMS2, hMSH2, hMSH3, and hMSH6 might lead to the inability to repair base pair mismatches and result in DNA replication errors or MSI. Inactivation of the MMR system causes genomic instability by increasing the rate of polymerase-generated replication errors and degrading the fidelity of DNA replication, particularly at microsatellite repeat sequences.53 MSI involves mutations or instability in short, tandemly repeated DNA sequences such as (A)n, (CA)n, and (GATA)n. Such DNA sequences are found in several key genes that are important for maintaining normal cellular function (Table 123-4). The receptor for TGF-β (TGF-βRII), for example, often is mutated as the result of MSI. Multiple lines of evidence suggest that the TGF-β pathway is an important tumor-suppressing pathway in the colon and that alterations in this pathway lead to tumor development. Less frequently targeted genes include the insulin growth factor 2 receptor; Bax and caspase 5, proteins that regulate apoptosis; E2F4, a transcription factor; and MSH3 and MSH6, DNA MMR proteins. β-Catenin mutations are present in up to 25% of MSI colon cancers. MSI, therefore, leads to accumulation of mutations in vulnerable genes, eventually resulting in the acquisition of the malignant

phenotype. Although a high frequency of MSI (instability at more than 40% of microsatellite loci) is characteristic of HNPCC, similar alterations can be found in about 15% of sporadic CRCs and also in premalignant lesions. MSI tumors remain diploid. Patients whose tumors demonstrate MSI might have a better prognosis60 and respond differently to chemotherapy,54,61,62 than those whose tumors are charac­ terized by chromosomal instability. Most patients with MSI colon cancers do not possess mutations in the known MMR genes, and evidence indicates that MSI in these tumors might arise through epigenetic mechanisms. Epigenetics is the study of clonal changes in gene expression without accompanying changes in DNA coding sequences. Epigenetic silencing is recognized now as an important mechanism in the evolution of a subgroup of CRCs. DNA methylation within promoters and alterations in histone modifications appear to be primary mediators of epigenetic inheritance in cancer cells, and hypermethylation of the hMLH1 promoter has been reported in up to 70% of sporadic MSI tumors. In the large intestine, aberrant methyl­ ation may be an important early event in the age-related field defect observed in sporadic colorectal neoplasia. Aberrant methylation also contributes to tumor progression through a hypermethylator phenotype (CPG island methylator phenotype [CIMP]) responsible for most cases of MSI related to hMLH1 inactivation (associated with about 15% of sporadic CRCs).54,63 The hallmark of CIMP is abnormal methylation of several tumor promoters. In one model, hyperplastic aberrant crypt foci may be the initial lesions in a pathway leading to the development of serrated adenomas (see Chapter 122). Methylated promoters of the MGMT, EVL, HLTF, SFRP2, SLC5A8, and MINT1 genes develop during tumor initiation, and hMLH1 promoter

2203

2204

Section X  Small and Large Intestine Table 123-4  Frequency of Some Target Gene Mutations in Colon Cancers with Microsatellite Instability (MSI) TARGET GENE TGF-β R2 ACVR2 IGFIIR BAX hMSH 3 hMSH 6 E2F-4 PTEN MBD4 (MED1) TCF4 CHK1 STK11 (LKB1) BLM Caspase 5 (ICErel-III) CDX2 TPB RIZ hRAD50 SEC63 AIM2

FREQUENCY of mutation IN COLON CANCERS WITH MSI (%) 90 86 10 50 50 33 65 40 39 10 <2 <18 62 <2 83 26 31 49 48

19-34

NORMAL FUNCTION of gene product TGF-β signaling Activin signaling Insulin-like growth factor and TGF-β signaling Apoptosis DNA mismatch repair DNA mismatch repair Cell cycle control Growth regulation DNA repair and binding to methylated DNA Growth regulation G2 cell cycle checkpoint Signal transduction Chromosome stability, DNA repair; helicase Apoptosis Homeobox protein TATA binding protein Interacts with RB DNA repair ER chaperone protein Interferon-inducible protein

Note: Most mutations cause frame shifts that prematurely truncate the protein, leading to inactivation of the affected allele. ER, endoplasmic reticulum; RB, retinoblastoma protein; TGF, transforming growth factor. Modified from Grady WM, Carethers J. Genomic and epigenetic instability in colorectal cancer progression. Gastroenterology 2008; 135:1079-99.

hypermethylation corresponds to the development of a serrated adenoma, with methylated TSP1 and TIMP3 helping to drive tumor progression.54 DNA methylation and histone H3 lysine 9 hypoacetylation and methylation appear to form a mutually reinforcing loop that contributes to tumor suppressor gene inactivation in CRCs. The BRAF gene encoding a downstream component of the RAS/RAF/MAPK pathway is often mutated in sporadic MSI tumors but not in tumors from patients with HNPCC (see later). MicroRNAs are 18- to 25-nucleotide noncoding RNA molecules that regulate translation of many genes. Expression patterns of microRNAs are altered in colon cancers, and they may be associated with survival and therapeutic outcome.64

Biochemical and Other Changes

Chapter 3 deals in depth with the biological and biochemical changes that occur during the development of colorectal neoplasia. Alterations in cell surface and secreted proteins and glycoproteins, including a number of important cell adhesion molecules, are characteristic of CRC.65 Interactions between tumor cells or between tumor cells and their environment may be homotypic (involving like molecules) or heterotypic (involving different adhesion molecules). Homotypic interactions often maintain the integrity of primary tumors by fostering adhesion between neighboring tumor cells, whereas heterotypic interactions might occur among tumor cells and platelets, lymphocytes, vascular endothelial cells, and components of the basement membrane matrix. Most tumor-associated molecules represent quantitatively or qualitatively altered molecules found either on normal tissues or during development, such as oncofetal antigens—for example, carcinoembryonic antigen [CEA]). Many of these molecules appear to play a role in maintaining normal tissue homeostasis or targeting bloodborne cells to specific sites. Altered expression, therefore, might contribute to tumor invasion and metastasis.

MMPs (matrix metalloproteinases) are a family of enzymes that degrade extracellular matrix. Overexpression of MMP-1, -2, -3, -7, -9, and -13 and MT1-MMP has been demonstrated in human CRCs. The degree of overexpression of some MMPs correlates with stage of disease, prognosis, or both.66 Metastasis is a multistage process by which tumor cells escape the primary tumor and establish secondary foci at distant sites (Fig. 123-9). Cells in the primary tumor must become vascularized (angiogenesis via vascular endothelial growth factors). Then they must escape the primary tumor by overcoming adhesive interactions (e.g., loss of E-cadherin) and by disrupting basement membranes (metalloproteinases such as type IV collagenase, matrilysin, loss of tissue inhibitors of collagenase). Finally they must enter lymphatics, the circulation, or both. In the bloodstream, they must survive interactions with blood components and the immune system and be transported to distant organ sites (principally the liver). At distant sites, tumor cells adhere to target endothelia via specific interactions (e.g., tumor-associated sialoglycoproteins and endothelial selectins) (Fig. 123-10), extravasate, interact with the microenvironment (e.g., growth factors), and establish secondary tumor foci. Tumor cell subpopulations with different metastatic potentials exist within the same primary tumor, and metastases result from the selective dissemination of tumor cells that possess the ability to participate in all stages of this complex process. Several carbohydrate antigens have been studied in relation to their potential usefulness as markers of metastatic potential and for their possible role in determining prognosis.65-68

FAMILIAL COLORECTAL CANCER It has become increasingly clear that genetic predisposition plays a role in a substantial number of CRCs. Although it is

Chapter 123  Colorectal Cancer Primary malignant neoplasm

Vascularization

Colon

Invasion

Embolism

Lymphatics Venules Capillaries

Interaction with platelets, lymphocytes, and other blood components

Establishment of microenvironment

Liver

Secondary tumor colonies (metastases)

Extravasation

Adherence of tumor cells

Heart Liver

Deposition in organs

Portal vein

Transport Figure 123-9.  Colon cancer metastasis. Cancer cells metastasize through a complex, multistage process. For tumor cells to form metastatic foci at distant sites, they must complete all stages of this process.

rbc

T S E c

E

H

A

B

c

Figure 123-10.  Colon cancer metastasis. On reaching the liver, colon cancer cells adhere to the sinusoidal endothelium through specific interactions and then invade the parenchyma. A, Photomicrograph showing tumor cells invading the liver after extravasation from a blood vessel. (Hematoxylin and eosin stain.) B, Electron micrograph showing collagen bundles (c) and tumor cells (T) adherent to sinusoidal endothelium (E) and invading between hepatocytes (H). rbc, red blood cell; S, sinusoid.

2205

2206

Section X  Small and Large Intestine Table 123-5  Amsterdam Criteria for Hereditary Nonpolyposis Colorectal Cancer

Table 123-6  Bethesda Guidelines for Testing of Colorectal Tumors for Microsatellite Instability71

At least three relatives with colorectal cancer (one must be a first-degree relative of the other two) Colorectal cancer involving at least two generations One or more colorectal cancer cases before age 50 years

Persons with cancer in families that meet the Amsterdam criteria (see Table 123-5) Persons with two HNPCC-related cancers, including synchronous and metachronous colorectal cancers or associated extracolonic cancers* Persons with colorectal cancer and a first-degree relative with colorectal cancer and/or HNPCC-related extracolonic cancer* and/or a colorectal adenoma; one of the cancers must be diagnosed before age 45 years and the adenoma diagnosed before age 40 years Persons with colorectal cancer or endometrial cancer diagnosed before age 45 years Persons with right-sided colorectal cancer with an undifferentiated pattern (solid/cribriform)† on histopathology diagnosed before age 45 years Persons with signet ring cell-type colorectal cancer‡ diagnosed before age 45 years Persons with colorectal adenomas diagnosed before age 40 years

Criteria defined by the International Collaborative Group on Hereditary Nonpolyposis Colorectal Cancer.

convenient to categorize CRCs into hereditary (or familial) and nonhereditary (or sporadic) types, it is more appropriate to assume that all cancers have genetic components that may be inherited or acquired to varying degrees. Accordingly, persons with familial colon cancer are born with an altered genome, and the environment might contribute additional genotoxic events, leading to the malignant phenotype. In the case of sporadic cancers, multiple somatic mutations are contributed by the environment (see Chapter 3). The role of heredity in the genesis of colon cancer is manifested most obviously in those with the heritable polyposis syndromes (Chapter 122). These syndromes are inherited in an autosomal dominant manner and are characterized by the presence of hundreds to thousands of colonic adenomas, with or without extracolonic tumors. Adenomas develop approximately a decade before the appearance of cancer, and virtually all affected persons eventually develop large bowel cancer if the colon is not removed. Nevertheless, these dramatic syndromes account for less than 1% of all cases of CRC. HNPCC is an inherited disease in which colon cancers arise in discrete adenomas, but polyposis—hundreds of polyps—does not occur.59,69,70 HNPCC accounts for approximately 6% of colonic adenocarcinoma. It is an autosomal dominant disorder with high penetration; approximately 80% are caused by germline mutations in genes responsible for repair of DNA errors, called mismatches, that occur during DNA replication (discussed earlier). During DNA synthesis, DNA polymerase might create single base-pair mismatches, resulting in structural abnormalities (loop-outs) involving unpaired bases. These errors tend to occur at repetitive DNA sequences termed microsatellites, and they are repaired by enzymes coded for by MMR genes. The majority of reported germline mutations in DNA MMR genes have been associated with the hMSH2 gene on chromosome 2 (40% to 50%) and the hMLH1 gene on chromosome 3 (20% to 30%). Mutations in hMSH6, hPMS1, and hPMS2 also have been reported in a small number of patients with HNPCC; no locus has been identified, however, for many HNPCC families. The definition of HNPCC was standardized and most strictly defined by the International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer. These Amsterdam criteria (Table 123-5) include at least three relatives with histologically verified CRC, one of them a first-degree relative of the other two (FAP excluded); at least two successive generations affected; and, in one of the individuals, diagnosis of CRC before age 50 years. Because these criteria do not account for the frequent occurrence of extracolonic cancers in such families, or for small kindreds, broader clinical criteria have been developed, including the Bethesda guidelines published by a National Cancer Institute-sponsored workshop on HNPCC (Table 123-6).70

*Endometrial, ovarian, gastric, hepatobiliary, or small intestinal cancer or transitional cell carcinoma of the renal pelvis or ureter. † Poorly differentiated or undifferentiated carcinoma composed of irregular solid sheets of large eosinophilic cells and containing small gland-like spaces. ‡ Composed of >50% signet ring cells. HNPCC, hereditary nonpolyposis colorectal cancer.

HNPCC families include members whose heritable cancer is limited to the colon and rectum (site-specific HNPCC, HNPCC type a, Lynch’s syndrome I) and families whose members also are prone to cancer of the female genital tract and other sites (cancer family syndrome, HNPCC type b, Lynch’s syndrome II) (Fig. 123-11). In HNPCC syndromes, discrete polyps, but not polyposis, might antedate the cancers. Adenomas in the proximal colon sometimes are flat or slightly raised lesions with foci of adenomatous change confined to the upper half of crypts (flat adenomas). HNPCC tumors have a tendency to involve proximal sites of the colon and to be multiple (synchronous and metachronous), with a higher incidence of mucinous carcinomas (Table 123-7). These CRCs usually appear at age 40 to 50 years, two decades earlier than CRC in the general population. In a study of a small, defined population in central Finland, HNPCC accounted for 4% to 6% of CRCs identified. This figure is similar to estimates of Lynch and others for the U.S. population. Patients with HNPCC and some unaffected family members have biological markers that resemble those in patients with FAP syndromes, including abnormal proliferative activity of colonic crypt cells; increased tetraploidy (twice the normal DNA content) in skin fibroblasts cultured in vitro; decreased degradation of fecal cholesterol; and cellmediated immune defects in vitro that might interfere with the recognition of killing of incipient tumor cells in vivo. The genetic defect in HNPCC, however, results from loss of the hMSH2, hMLH1, and other genes, leading to increased susceptibility to mutation from failure to repair base-pair mismatches.59,71 Although CRC syndromes with readily apparent patterns of inheritance currently account for only a small portion of total colon cancer cases, hereditary factors may be present in a larger proportion of cases.72 Genetic susceptibility to CRC in the general population is suggested by the two-fold to three-fold increase in CRC in first-degree relatives of patients with sporadic adenomas and CRCs. The relative risk is even stronger when cancer occurs in family members

Chapter 123  Colorectal Cancer Carcinoma documented pathologically Clinical history of carcinoma Multiple primary malignancy C-50 Age at diagnosis Site of tumor C - colon O - ovary B - brain E - endometrium S - stomach L - lung P - adenomatous polyp of the colon 2 Totally unaffected offspring

Generation I 1

2

3 C-41

4

II 1

3

2

4

5

6

8

7

9

11

10

12

14

13

C-27,45 C-62

III 1 2 E-45 L-49

3 E-33

4

5

6 O-37 B-60 C-64 S-68

7

8

C, P-50

9

10 C, P-65

5

IV 1

V

2

4

3

2

5

5

6

2

P

12

3

P-45 S-53

4 19

18

4

13 C-26, 49 P-49

5

7

5

11 C-27

2

3

20 21

2

2

Figure 123-11.  Pattern of inheritance of cancer in a familial aggregate with Lynch’s syndrome II (HNPCC type b). Affected members were found in generations I, II, and III; members of generations IV and V were still young and at risk for developing carcinomas when the pedigree was obtained. (From Boland CR. Familial colonic cancer syndromes. West J Med 1983; 139:351.)

Table 123-7  Comparison of HNPCC and Sporadic Colorectal Cancer

CLINICAL FEATURE

HNPCC

SPORADIC COLORECTAL CANCER

Mean age at diagnosis (years) Multiple colon cancers Synchronous colon cancers Metachronous colon cancers Proximal location* Increased risk of malignant tumors at other sites Mucinous and poorly differentiated colon cancers Prognosis

45 35% 18% 24% 72% Yes

67 4-11% 3-6% 1-5% 35% No

Common

Infrequent

Favorable†

Variable

*Proximal to the splenic flexure; location of the initial cancer. † Patients whose tumors demonstrate microsatellite instability have a more favorable prognosis than those with microsatellite-stable tumors. HNPCC, hereditary nonpolyposis colorectal cancer.

younger than 50 years. The precise role of genetic factors in this group and their interaction with the environment in the evolution of CRC remains to be defined. The identification of susceptibility genes in this group is of great interest and has been aided by genome-wide scanning techniques in sibling pairs with colorectal neoplasia.73

PREDISPOSING FACTORS The risk of developing CRC depends on a number of demographic factors (Table 123-8). The probable influence of diet and other environmental factors has been discussed (see Tables 123-1 and 123-2). Other factors include age, personal history of adenoma or of carcinoma, existence of predisposing diseases (particularly inflammatory bowel disease [IBD]), and family history (discussed earlier). It has been proposed that risk scores may be developed to aid in identifying and treating susceptible persons.74

2207

Section X  Small and Large Intestine Table 123-8  Risk Factors for Colorectal Cancer Age ≥50 years High-fat, low-fiber diet* Personal history of Colorectal adenomas (synchronous or metachronous) Colorectal carcinoma Family history of a polyposis syndrome: Familial adenomatous polyposis Turcot’s syndrome Muir-Torre syndrome Peutz-Jeghers syndrome Familial juvenile polyposis Hereditary nonpolyposis colorectal cancer First-degree relative with colorectal cancer Inflammatory bowel disease Ulcerative colitis† Crohn’s disease† *Based on descriptive epidemiology; data from case-control, cohort, and randomized trials are less convincing. † Especially with high-grade dysplasia or dysplasia-associated mass lesions.

30 Percent of all colorectal cancers

2208

mas grow larger, they progressively dedifferentiate, become dysplastic, and then become malignant. With increasing size or increasing villous architecture, the incidence of nuclear atypia, dysplasia, and in situ or invasive carcinoma increases. Despite the potential for adenomas to evolve to carcinomas, however, the actual risk of this phenomenon is unknown. Adenomatous polyps are common, especially after age 50 years, in populations that consume a Western diet, but the prevalence of adenomas is high, compared with the incidence of cancer. It has been estimated that 29% of the living population older than 35 years in Norway have colorectal adenomas, with an annual conversion rate of adenomas to carcinoma (based on cancer incidence from multiple tumor registries in Norway) of 0.25. The malignancy rate is higher in large adenomas, adenomas with villous architecture, and adenomas with cytologic nuclear atypia or dysplasia (advanced adenomas). The estimated annual rate of conversion to invasive cancer in persons with adenomas larger than 1  cm, villous components, or severe dysplasias has been reported to be 3%, 17%, and 37%, respectively. Data from a national colonoscopy data base in Germany suggested that the annual transition rates from advanced adenoma to cancer increase with age.75

Carcinoma 20

10

<40 40–49 50–59 60–69 70–79

>80

Age at diagnosis Figure 123-12.  Frequency of colorectal cancer by age at diagnosis in the United States. Ninety percent of cancers occur after age 50 years.

AGE

The risk of developing CRC rises sharply after age 40 years in the general population, 90% of cancers occurring in persons aged 50 years and older (Fig. 123-12). A 50-year-old person has approximately a 5% chance of developing CRC if he or she survives to age 80 years and a 2.5% risk of dying from the disease; this has important implications for screening (see later). Sporadic CRCs arise in other age groups, such as the third and fourth decades, however, and the diagnosis must be considered in younger persons with signs and symptoms characteristic of this disease, especially if they have a family history of colorectal neoplasia.

PRIOR ADENOMA AND CARCINOMA Adenoma

Present evidence strongly indicates that most CRCs arise from preexisting adenomas (see Chapter 122). The risk of CRC increases with the number of adenomas, the most extreme example being the familial polyposis syndromes. Clinical and morphologic evidence suggest that as adeno-

People with CRC have an increased risk of harboring a second carcinoma (synchronous carcinomas) or of developing another one subsequently (metachronous carcinomas). The frequency of more than one carcinoma in the same person ranges between 2% and 6% (0.7% to 7.6% for synchronous cancers and 1.1% to 4.7% for metachronous ones). Most patients with simultaneous cancers have one in the proximal colon and the other in the distal colon; in the minority of patients with synchronous cancers, the two lesions are located in the same colonic segment. The degree of invasiveness of synchronous cancers often differs, and prognosis depends on the worst-stage lesion. Five-year survival rates for patients with synchronous cancers whose cancers have been resected are similar to those with single lesions. The interval between an initial cancer and a metachronous one may be considerable (lesions separated by as much as 23 years have been reported), but several studies note that 50% of metachronous cancers arise within five to seven years of the index lesion. Second cancers often occur at a site remote from the initial lesion.

FAMILY HISTORY

The risk of CRC in first-degree relatives of those with sporadic CRC is increased two-fold to three-fold (see earlier). The risk is higher when adenoma or carcinoma has occurred in a relative at an early age or when more than one relative has had carcinoma. These factors have been taken into account in screening guidelines that stratify patients according to potential cancer risk.72-83 FAP and its variants are inherited in an autosomal dominant manner, and without colectomy virtually all affected members with polyps eventually develop carcinoma. HNPCC also has an autosomal dominant pattern of inheritance (see Chapter 122). Turcot’s syndrome is a rare combination of inherited adenomatous polyposis and malignant brain tumors. Inheritance is autosomal dominant, and germline mutations in the APC gene or mutations of hMLH1 and hPMS2 characteristic of HNPCC have been described (see Chapter 122). Muir-Torre syndrome is a rare variant of HNPCC that manifests with multiple skin lesions, including sebaceous adenomas and carcinomas, basal cell and squamous cell

Chapter 123  Colorectal Cancer

INFLAMMATORY BOWEL DISEASE

Patients with idiopathic IBD (Crohn’s disease, UC) are at increased risk for developing adenocarcinoma of the colon (see Chapters 111 and 112).88 Actuarially derived (life table) cumulative cancer incidences from tertiary referral centers (Fig. 123-13) agree that the risk of cancer in patients with UC begins after a disease duration of seven years and rises about 10% per decade, reaching approximately 30% at 25 years. Nonetheless, difficulties related to sources of referral, sampling, recognition and characterization of disease, differences in follow-up procedures, and methods used to detect neoplastic disease cloud many such studies. The assessed cancer risk for patients with UC followed in one private practice has been reported to be only 6.6% 26 years after disease onset and 11.4% at 32 years. The true risk of developing CRC in UC must lie between predictions from tertiary centers and the primary care setting. The risk of CRC for patients with UC correlates most closely with duration of disease. In a large group of patients with extensive disease who were followed prospectively, the risk of carcinoma per patient-year was zero before 10 years and one in 86 after 20 years. The risk is greatest with universal colitis. It has been reported that the risk of cancer in left-sided disease (i.e., distal to the splenic flexure) begins approximately a decade later than with universal colitis, but at least one surveillance study found no difference between these groups in the temporal development of preneoplastic

100

Cumulative probability of cancer (%)

carcinomas, and keratoacanthomas in addition to colonic adenomas and adenocarcinomas. MSI and loss of hMLH1 and hMSH2 expression has been noted in a high percentage of tumors occurring in this syndrome. Peutz-Jeghers syndrome and the familial form of juvenile polyposis both have an increased risk of small and large bowel cancer (see Chapter 122). Peutz-Jeghers syndrome is an autosomal dominant disease, which, in most families, has been mapped to chromosome 19 p13.3 and the STK11 gene (serine threonine kinase 11). Adenomatous changes have been reported in 3% to 6% of hamartomas from these patients. Extracolonic cancers are common and have been reported to occur in 50% to 90% of patients with PeutzJeghers syndrome (relative risk for all cancers was 15.2).84 A significant increase in a variety of cancers (esophagus, stomach, small intestine, pancreas, lung, breast, uterus, and ovary) has been reported. Familial juvenile polyposis is a rare autosomal dominant disease associated with polyps that may be limited to the colon or the stomach or that occur throughout the gastrointestinal tract (see Chapter 122).85 The genetic basis of this syndrome is not fully understood, but germline mutations in a gene (SMAD4) located on chromosome 18q21.1 that encodes an intracellular mediator in the TGF-β signaling pathway have been identified in some affected patients, as have mutations in the BMPR1A (bone morphogenetic protein receptor type 1A) gene; each accounts for approximately 20% of cases.86 The bone morphogenetic receptors (BMPs) are a family of transmembrane serine-threonine kinases whose ligands are members of the TGF-β super­ family. BMPs repress WNT signaling, and BMPR1A plays a role in apoptosis. The PTEN (phosphatase and tensin homologue deleted on chromosome 10) gene located on chromosome 10 also has been linked to some cases of juvenile polyposis. PTEN encodes a dual phosphatase, loss of function of which leads to protection of cells from various apoptotic stimuli. Overexpression of PTEN leads to cell cycle arrest or induction of apoptosis, or both. In patients with juvenile polyposis, the cumulative lifetime risk for CRC might approach 40%.87

80 The Mount Sinai Hospital

60

40

Mayo Clinic

20

10

20

30

40

Years after onset of colitis Figure 123-13.  Cumulative probability of developing colorectal carcinoma in patients with ulcerative colitis seen at two tertiary referral centers. Data from primary care settings indicate a similar pattern but lower incidence rates (see text and Chapter 112).

dysplasia. The risk for patients with ulcerative proctitis is only slightly increased compared with that in the general population. The risk of cancer is not related to severity of the initial attack of colitis or age at onset of colitis, independent of disease duration. The relationship to disease activity per se is unclear, although chronic inflammation is a common theme in epithelial carcinogenesis. CRC in persons with UC is a risk factor for CRC in their relatives without colitis and CRC in relatives without UC is a risk factor for those with colitis. There is an association of backwash ileitis with CRC in patients with UC who undergo proctocolectomy.89 Cancer arising in the setting of UC traditionally has been thought to be a highly malignant lesion with a poor prognosis, but studies using matched controls from colon cancer populations without colitis have failed to show a significant difference in survival between the two groups. The increased risk for CRC in patients with Crohn’s disease or ileocolitis has been reported to be as much as 4 to 20 times that in the general population, although one cohort study failed to confirm an increased incidence of colon cancer in these patients. Cancer can arise at an earlier age in these patients than in the general population. Many of these cancers are mucinous carcinomas, and they often occur in surgically bypassed or strictured segments of colon. Carcinomas do not develop de novo from normal mucosa but from mucosa that has undergone a sequence of morphologic changes that culminate in invasive carcinoma. As in precancerous adenomas, dysplasia is a precursor to carcinoma in IBD. Dysplasia comprises abnormalities in crypt architecture and cytologic detail (Figs. 123-14 and 123-15). Epithelial crypts are reduced in number, irregularly branched, and crowded together (“back-to-back” glands). Cell nuclei may be enlarged and hyperchromatic, may have increased numbers of mitoses, and may be located at different levels in the cell, producing a picket-fence appearance (pseudostratification). Dysplasia is classified by grade as mild (or low grade) to severe (or high grade).

2209

2210

Section X  Small and Large Intestine Retrospective analyses report that 90% of resected colons from patients with UC and cancer contain dysplastic mucosa somewhere in the colon and that 30% of patients with severe rectal or colonic dysplasia on resection or biopsy have coexisting carcinoma. Colonoscopic studies suggest that 25% of colons that demonstrate high-grade dysplasia on biopsy harbor a carcinoma. Dysplasia often is patchy, and it may be present in the colon but absent from the rectum. Because the lack of uniformity of the definition of dys­ plasia can make interpretation of such data difficult, the multidisciplinary Inflammatory Bowel Disease Dysplasia Morphology Study Group developed a standardized classification for dysplasia arising in the setting of IBD. Several large prospective studies have attempted to determine the true risk of cancer in patients with colonic dysplasia and UC and the impact of screening programs for dysplasia. Results suggest that biopsy surveillance programs can be effective in helping control the risk of carcinoma in patients with long-standing UC. The risk of cancer appears highest in those with high-grade dysplasia that arises in

visible plaques or masses (dysplasia-associated lesion or mass [DALM], as illustrated in Fig. 123-15). A computer cohort decision analysis suggested that surveillance should increase life expectancy for patients with UC. Most investigators believe that patients who have UC for more than seven to eight years should undergo colonoscopy with multiple mucosal biopsies annually to identify areas of dysplasia; they advocate colectomy for severe dysplasia or DALM. Because the significance of low-grade or moderate dysplasia is less clear, immediate resection for patients with these levels of dysplasia is controversial. Prophylactic colectomy also has been recommended as an option for those with disease of at least 10 years’ duration. Studies employing flow cytometry have detected aneuploid cell populations in colons resected for UC with dysplasia or early cancer. Chromosomal alterations can occur early in UC-related neoplastic progression and appear to precede the histologic development of dysplasia; relative loss of chromosome 18q also may be important in neoplastic progression. Both dysplasia and increased risk of colon carcinoma have been reported in patients with Crohn’s disease. As in UC, dysplasia appears in diseased colon segments, and its presence correlates with duration of disease. In one screening trial,90 the finding of definite dysplasia was associated with age older than 45 years and increased symptoms. By life-table analysis, the probability of detecting dysplasia or cancer after a negative screening colonoscopy was 22% by the fourth surveillance examination.

OTHER ASSOCIATIONS

Figure 123-14.  Photomicrograph of dysplasia in the setting of ulcerative colitis. The number of goblet cells is decreased. Glands are branched, irregular, and crowded together. Cell nuclei are hyperchromatic and occur at different levels, producing a pseudopalisading or picket-fence appearance. (Hematoxylin and eosin stain.)

A

Diverting bile and bile acids to the lower small intestine, either surgically or by administration of cholestyramine, increases the yield of proximal colon tumors in carcino­ gen-treated animals. There is no evidence, however, that chronic use of cholestyramine is associated with an increased risk of colon cancer in humans. The possibility of an increase in colon cancer following cholecystectomy has been raised, although the clinical evidence for such an association is contradictory. Increased proliferative activity in the distal colonic mucosa has been demonstrated in patients who have undergone cholecystectomy, and an increased incidence of tubular adenomas has been observed in patients older than 60 years with a postcholecystectomy interval of more than 10 years, but an increased risk of colon cancer for these patients has been both supported and refuted.

B

Figure 123-15.  Plaque-like dysplasia–associated lesion or mass in a patient with long-standing ulcerative colitis. A, Lesion as seen through the colonoscope. B, Biopsy specimen revealing high-grade dysplasia. (Hematoxylin and eosin stain.)

Chapter 123  Colorectal Cancer A retrospective population-based cohort study that used the Swedish Inpatient Register evaluated almost 23,000 persons who had had cholecystectomy up to 31 years previously.91 Patients having had cholecystectomy had an increased risk of proximal intestinal adenocar­ cinoma, which declined with increasing distance from the common bile duct. The risk was significantly increased for adenocarcinoma of the small bowel and ascending colon but not the remaining colon or rectum.

PATHOLOGY, NATURAL HISTORY, AND STAGING GROSS PATHOLOGY

The gross morphologic features of CRC depend on the tumor’s site. Carcinomas of the proximal colon, particularly those of the cecum and ascending colon, tend to be large and bulky, often outgrowing their blood supply and undergoing necrosis (Fig. 123-16); this polypoid configuration, however, also may be found elsewhere in the colon and rectum. In the more distal colon and rectum, tumors more commonly involve a greater circumference of the bowel, producing an annular constriction or napkin-ring appearance (Fig. 123-17). The fibrous stroma of these tumors accounts for the constriction and narrowing of the bowel lumen, whereas the circular arrangement of colonic lymphatics is responsible for their annular growth. These tumors also can ulcerate, and occasionally they have a flatter appearance with predominantly intramural spread (Fig. 123-18); the latter are seen most often in the setting of

A

C

IBD. Morphologic features of CRCs have clinical, diagnostic, and prognostic implications.

HISTOLOGY

Carcinomas of the large bowel characteristically are adenocarcinomas that form moderately differentiated to welldifferentiated glands and secrete variable amounts of mucin (Fig. 123-19). Mucin, a high-molecular-weight glycoprotein, is the major product secreted by both normal and neoplastic glands of the colon, and may be seen best with histochemical stains such as periodic acid–Schiff (PAS). In poorly differentiated tumors (see Fig. 123-19), gland formation and mucin production are present but less prominent. Signet ring cells, in which a large vacuole of mucin displaces the nucleus to one side, are a feature of some tumors (Fig. 12320A). In approximately 15% of tumors, large lakes of mucin contain scattered collections of tumor cells (see Fig. 12320B). Such mucinous or colloid carcinomas are most common in patients with HNPCC or UC and in patients whose carcinomas occur at an early age. Scirrhous carcinomas are uncommon and are characterized by sparse gland formation, with marked desmoplasia and fibrous tissue surrounding glandular structures. Sometimes tumors demonstrate a mixed histologic picture, with glands of varying degrees of differentiation. Cancers other than adenocarcinomas account for less than 5% of malignant tumors of the large bowel. Tumors arising at the anorectal junction include squamous cell carcinomas, cloacogenic or transitional cell carcinomas, and melanomas (see Chapter 125). Primary lymphomas and carcinoid tumors of the large bowel account for less

B

D

Figure 123-16.  Carcinomas of the cecum seen at colonoscopy. Carcinomas of the proximal colon are often large and bulky polypoid lesions (A and B), may involve the ileocecal valve (C), and can outgrow their blood supply and become necrotic (D).

2211

2212

Section X  Small and Large Intestine

A

Figure 123-17.  Obstructing carcinoma of the sigmoid colon. A, Colonoscopic view. B, Applecore lesion seen on barium enema examination. C, Surgical specimen demonstrating an annular constriction or napkin-ring appearance.

B

C

A

B

Figure 123-18.  Colonoscopic view of flat, plaque-like carcinomas of the colon (A) and rectum (B, arrows) in patients with inflammatory bowel disease.

than 0.1% of all large bowel neoplasms (see Chapters 29 and 31).

NATURAL HISTORY AND STAGING

CRCs begin as intramucosal epithelial lesions, usually arising in adenomatous polyps or glands. As cancers grow, they become invasive, penetrating the muscularis mucosae

of the bowel and invading lymphatic and vascular channels to involve regional lymph nodes, adjacent structures, and distant sites. Although CRCs grow at varying rates, they most often have long periods of silent growth before producing bowel symptoms. The mean doubling time of colon cancers determined radiologically in one study was 620 days. Comparative lesion sequencing using modern molecu-

Chapter 123  Colorectal Cancer

A

B

Figure 123-19.  Histopathology of adenocarcinoma of the colon. A, Well-differentiated adenocarcinoma of the colon. Sections stained with hematoxylin and eosin demonstrate crowded neoplastic glands containing variable amounts of mucin. B, Poorly differentiated adenocarcinoma of the colon.

A

B

Figure 123-20.  Histopathology of mucinous carcinomas of the colon. Histologic types include signet ring cell carcinoma in which a large vacuole of mucin displaces the nucleus (A) and colloid carcinoma with scattered nests of tumor cells floating in lakes of mucin (B) (Hematoxylin and eosin stain.)

lar techniques combined with clinical observation suggests that it can take approximately 17 years for a large benign tumor to evolve to advanced cancer but less than two years for cells within that cancer to acquire the ability to metastasize.92 Patterns of spread depend on the anatomy of the individual bowel segment as well as its lymph and blood supplies. Cancers of the rectum advance locally by progressive penetration of the bowel wall. Extension of the primary tumor intramurally and parallel to the long axis of the bowel most often is limited, and lymphatic and hematogenous spread is unusual before penetration of the muscularis mucosae. Exceptions appear to be poorly differentiated tumors, which can metastasize lymphatically or hematogenously before completely penetrating the bowel. Because the rectum is relatively immobile and lacks a serosal covering, rectal cancers tend to spread contiguously to progressively involve local structures. Transrectal ultrasonography is useful in staging depth of rectal cancers. Because of the dual blood supply of the lower one third of the rectum, tumors arising here can metastasize hematogenously via the superior hemorrhoidal vein and portal system to the liver or by way of the middle hemorrhoidal vein and inferior vena cava to the lungs. The veins of the upper and middle thirds of the

rectum drain into the portal system, and tumors in these segments first spread hematogenously to the liver. Occasionally, lumbar and thoracic vertebral metastases result from hematogenous spread via portal-vertebral communications (i.e., Batson’s vertebral venous plexuses). Colon cancers can invade transmurally and involve regional lymphatics and then distant nodes; lymphatic drainage generally parallels the arterial supply to a given bowel segment. The liver is the most common site of hematogenous spread from colon tumors via the portal venous system. Pulmonary metastases from colon cancer result, in general, from hepatic metastases. On the basis of observations of what he believed to be an orderly progression of locoregional invasion by rectal cancers, Cuthbert Dukes proposed a staging classification in 1929, which has since been modified many times in attempts to increase its prognostic value for cancers of both the rectum and the colon. The most commonly employed modification of the Dukes system is that of Astler and Coller (Table 123-9). This classification uses the following designations: A: Tumors limited to the mucosa B1: Tumors extending into, but not through, the muscularis propria

2213

—

Regional nodal metastases —

—

—

—

B1 B2

B3

C

C1

C2

C3

D

—

Regional nodal metastases near primary lesion Proximal node involved at point of ligation —

—

—

— —

Through bowel wall

Limited to bowel wall

GABRIEL, DUKES, BUSSEY, 1935 (RECTUM)

GITSG, Gastrointestinal Tumor Study Group.

B

Limited to bowel wall Through bowel wall — —

A

STAGE

DUKES, 1932 (RECTUM)

—

—

—

Regional nodal metastases —

—

Into muscularis propria Through muscularis propria

—

Limited to mucosa

KIRKLIN ET AL, 1949 (RECTUM + SIGMOID)

—

—

Same as B2 + regional nodal metastases

Same as B1 + regional nodal metastases

—

—

Into muscularis propria Through muscularis propria (and serosa)

—

Limited to mucosa

ASTLER-COLLER, 1954 (RECTUM + COLON)

Table 123-9  Dukes Staging of Carcinoma of the Rectum and Its Modifications for Colorectal Carcinoma

Distant metastases (liver, lung, bone) or parietal peritoneum or adjacent organ invasion

—

—

Regional nodal metastases —

—

Tumor extension into pericolic fat — —

Limited to mucosa

TURNBULL ET AL, 1967 (COLON)

Same as B3 + regional nodal metastases —

Same as B2 + regional nodal metastases

Same as B1 + regional nodal metastases

Into muscularis propria Through serosa (m = microscopic; m + g = gross) Adherent to or invading adjacent structures —

—

Limited to mucosa

MODIFIED ASTLER-COLLER (GUNDERSON, SOSIN), 1974 (RECTUM + COLON)

—

—

>4 regional nodes positive

1-4 regional nodes positive

—

—

Into muscularis propria Through serosa

—

Limited to mucosa

GITSG, 1975 (RECTUM + COLON)

2214 Section X  Small and Large Intestine

Chapter 123  Colorectal Cancer B2: Tumors penetrating the muscularis propria but without lymph node involvement C: Tumors with regional lymph node involvement Stage C tumors are divided further into primary tumors limited to the bowel wall (C1) and those that penetrate the bowel wall (C2). By contrast, in the system proposed by the Gastrointestinal Tumor Study Group (GITSG), C1 lesions are those in which one to four regional lymph nodes contain tumor, and C2 lesions are those in which more than four lymph nodes contain tumor (see Table 123-9). Another modification, by Turnbull and associates, added a D category referring to distant metastases. In an attempt to provide a uniform and orderly classification for CRCs, the American Joint Committee on Cancer (AJCC) has introduced the tumor-node-metastases (TNM) classification for CRC.93 This system classifies the extent of the primary tumor (T), the status of regional lymph nodes (N), and the presence or absence of distant metastases (M). Cases are assigned the highest value of TNM that describes the full extent of disease and are grouped into five stages (0 through IV). The five stages have become important in uniformly randomizing patients for therapeutic trials (Table 123-10) and have, in most cases, replaced the Dukes classification for CRC.

PROGNOSIS Clinical and pathologic variables that can affect the prognosis of patients with CRC are outlined in Table 123-11. These variables are important in predicting clinical outcome and in designing optimal strategies for treatment and follow-up. Their identification has led to a progressive modification of the staging classifications for CRC. The roles of histologic differentiation; tumor size, location, configuration, and degree of invasion; and lymph node status must be evaluated on the basis of prospective analyses of patients who undergo curative resections for CRC.

SURGICAL-PATHOLOGIC STAGING

The depth of transmural tumor penetration and the extent of regional lymph node spread are the most important determinants of CRC prognosis (Fig. 123-21). The degree of bowel wall penetration also affects prognosis, independently of lymph node status, and correlates with the number of involved nodes as well as with the incidence of local recurrence after surgical resection. The number of involved regional lymph nodes also correlates independently with outcome (see Fig. 123-21B).

TUMOR MORPHOLOGY AND HISTOLOGY

Table 123-10  AJCC TNM Staging of Colorectal Cancer STAGE*

CRITERIA†

0

Carcinoma in situ: intraepithelial or invasion of lamina propria‡ (Tis N0 M0) Tumor invades submucosa (T1 N0 M0) [Dukes A] Tumor invades muscularis propria (T2 N0 M0) Tumor invades through the muscularis propria into subserosa or into nonperitonealized pericolic or perirectal tissues (T3 N0 M0) [Dukes B] Tumor perforates the visceral peritoneum or directly invades other organs or structures and/or perforates visceral peritoneum§ (T4 N0 M0) Any degree of bowel wall perforation with regional lymph node metastasis N1: metastasis in 1-3 regional lymph nodes N2: metastasis in ≥4 regional lymph nodes Any T N1 M0 [Dukes C] Any T N2 M0 Any invasion of bowel wall with or without lymph node metastasis, but with evidence of distant metastasis Any T Any N M1

I II

III

IV

*Stage II is subdivided into IIA (for T3 tumors) and IIB (for T4 tumors). Stage III is subdivided into IIIA (T1-T2 N1 M0), IIB (T3-T4 N1 M0), and IIIC (any T N2 M0). N1 lesions have 1-3 positive nodes; N2 tumors have ≥4 positive nodes. Smooth metastatic nodules in the pericolic or perirectal fat are considered lymph node metastases. Irregularly contoured metastatic nodules in the peritumoral fat are considered vascular invasion. † Dukes B, which corresponds to stage II, is a composite of better (T3 N0 M0) and worse (T4 N0 M0) prognostic groups, as is Dukes C, which corresponds to stage III (any T N1 M0 and any T N2 M0). ‡ Tis includes cancer cells confined within the glandular basement membrane (intraepithelial) or lamina propria (intramucosal) with no extension through the muscularis mucosae into the submucosa. § Direct invasion in T4 includes invasion of other segments of the colorectum by way of serosa; e.g., invasion of the sigmoid colon by a carcinoma of the cecum. AJCC, American Joint Committee on Cancer; N0, no regional lymph node metastasis; M0, no distant metastasis; M1, distant metastasis. From the American Joint Committee on Cancer. AJCC Cancer Staging Manual: Colon and Rectum, 6th ed. New York: Springer-Verlag; 2002.

The TNM classification (see Table 123-10) is based in part on the observation that for most cancers, tumor size correlates with local and distant spread and, thus, indirectly with prognosis. Numerous studies suggest that CRC is an exception, however, and that the size of the primary tumor per se does not correlate with prognosis. In fact, patients with exophytic or polypoid tumors appear to have a better prognosis than those with ulcerating or infiltrating tumors. Tumor prognosis correlates with histologic grade: poor differentiation confers a worse prognosis than does a high degree of differentiation. Mucinous65 and scirrhous car­ cinomas appear to be biologically more aggressive, and patients with these tumors do not survive as long as those who have other types of adenocarcinoma. Mucin-associated antigens might play a role in tumor progression and metastasis of colon cancer cells. Signet ring carcinomas have been reported to present at an advanced stage and to be highly invasive tumors. Venous invasion by CRC (Fig. 123-22A) usually correlates with local recurrence after resection, visceral metastases, and decreased survival, but it might not have independent prognostic value for tumors confined to the bowel wall. Although lymphatic invasion is associated with decreased survival, it is not clear whether this variable is independent of depth of tumor invasion and regional nodal metastasis. Perineural invasion (see Fig. 123-22B) also is linked to increased local recurrence and decreased survival, but such data are limited. The degree of inflammatory response and lymphocytic infiltration in and around a cancer may be related to outcome; increased inflammation and immune reaction appear to confer a better prognosis, but data are limited. The prognosis of patients with CRC may be related to the DNA content of the primary tumor, because survival is shorter for patients with nondiploid or aneuploid tumors than for those whose tumor cells have a normal or diploid DNA content. Although the DNA content of the primary tumor might correlate with the potential for local or distant recurrence after primary resection, the value of routine flow cytometric measurements or image analysis of the DNA content of tumor cells for assessing clinical prognosis and planning treatment remains to be determined. Deletions

2215

2216

Section X  Small and Large Intestine Table 123-11  Pathologic, Molecular, and Clinical Features That May Affect Prognosis in Patients with Colorectal Cancer FEATURE OR MARKER Pathologic Surgical-pathologic Stage Depth of colon wall penetration Number of regional nodes involved by tumor Tumor Morphology and Histology Degree of differentiation Mucinous (colloid) or signet ring cell histology Scirrhous histology Invasion Venous Lymphatic Perineural Other Features Local inflammation and immunologic reaction Tumor morphology Tumor DNA content Tumor size Molecular Loss of heterozygosity at chromosome 18q (DCC, DPC4) Loss of heterozygosity at chromosome 17p (TP53) Loss of heterozygosity at chromosome 8p Increased labeling index for p21WAF/CIP1 protein Microsatellite instability Mutation in BAX gene Clinical Diagnosis in asymptomatic patients Duration of symptoms Rectal bleeding as a presenting symptom Colon obstruction Colon perforation Tumor location Age <30 yr Preoperative CEA level Distant metastases

EFFECT ON PROGNOSIS

Increased penetration diminishes prognosis 1-4 nodes is better than >4 nodes Well-differentiated is better than poorly differentiated Diminished prognosis Diminished prognosis Diminished prognosis Diminished prognosis Diminished prognosis Improved prognosis Polypoid or exophytic is better than ulcerating or infiltrating Increased DNA content (aneuploidy) diminishes prognosis No effect in most studies Diminished prognosis Diminished prognosis Diminished prognosis Improved prognosis Improved prognosis Diminished prognosis Possibly improved prognosis No demonstrated effect Improved prognosis Diminished prognosis Diminished prognosis May be better for colonic than for rectal tumors May be better for left colonic than right colonic tumors Diminished prognosis Diminished prognosis with a high CEA level Markedly diminished prognosis

CEA, carcinoembryonic antigen.

in chromosomes 18q and 17p (TP53) may be important indicators of prognosis, independent of stage.55,57 As indicated in Table 123-11, a growing number of other molecular markers also might predict prognosis or response to therapy.57

CLINICAL PREDICTORS OF PROGNOSIS

Whereas screening programs for CRC suggest that tumors diagnosed in asymptomatic patients are less advanced, assessment of the impact of early diagnosis on survival of asymptomatic persons awaits the results of prospective randomized, controlled studies such as the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial.94 Duration of symptoms might not correlate directly with prognosis, and some presenting symptoms, such as rectal bleeding, may be associated with better rates of survival. Bowel obstruction or perforation has been linked with poor prognosis. Patients who present with obstructing lesions might not be candidates for curative surgery and have higher rates of operative morbidity and mortality. Recurrence following “curative” surgery also is higher in patients who present with obstruction or perforation. The location of the primary tumor can influence outcome. Disease-free survival at three years appears to be 2% to 14% higher after surgery for tumors of the left than of the right colon. Some studies suggest a survival advantage for patients with colon compared with rectal cancers.

As many as 3% of CRCs develop before 30 years of age, and only 11% of such persons have a predisposing con­ dition such as FAP or UC. The prognosis is worse than for older patients and is particularly poor in the pediatric age group. Poor prognosis may be related to a higher percentage of more-advanced cancers and mucinous adenocarcinomas in these young patients. Alternatively, patients with tumors demonstrating MSI appear to have a better prognosis irrespective of age.60 Thus, whereas CRCs occur at a younger age in those with HNPCC, these patients have a better prognosis than those with microsatellite-stable cancers. Outcome also is related to preoperative serum CEA levels. Tumor recurrence is higher, and the estimated mean time to recurrence is shorter, in patients with Dukes B and C cancers who have high preoperative CEA levels. The preoperative CEA level may be of prognostic value only in patients with Dukes C CRCs who also have four or more involved lymph nodes (stage C2), but not in patients with Dukes A and B lesions or Dukes C lesions with fewer than four nodes involved. Expression of mucin-associated carbohydrate antigens other than CEA, such as sialyl Lewisx, also can correlate with prognosis.65 Expression of the carbohydratebinding protein galectin-3 correlates with tumor progression in the colon and can confer a poor prognosis.65,67 Approximately one fourth of patients with CRC exhibit clinical evidence of hematogenous spread when seen initially, and one half of patients with CRC eventually develop

Chapter 123  Colorectal Cancer 100

80

Percent surviving

B

Dukes A (N=21)

Dukes B2 (N=110)

60

Dukes C (N=89)

40

Dukes D (N=47)

20

100

Dukes C, 1-3 nodes (N1) (N=59)

80

Dukes B1 (N=52)

Percent surviving

A

60

Dukes C, ≥ 4 nodes (N2) (N=30)

40

20

0

0 1

2

3

4

5

6

0

1

2

3

4

5

Years after start of treatment Years after start of treatment Figure 123-21.  A, Survival probabilities according to the Dukes stage as modified by Astler-Coller (see Table 123-9) in patients undergoing potentially curative surgery for colorectal cancer. Expected survival among age- and sex-matched general populations is indicated by the straight line. B, Survival probabilities according to the number of nodes involved in patients with stage C colorectal carcinoma. Data using TNM staging confirm a substantial difference in five-year survival among patients with lesions having one to three positive nodes (N1) and four or more positive nodes (N2). (From Moertel CG, O’Fallon JR, Go VL, et al. The preoperative carcinoembryonic antigen test in the diagnosis, staging, and prognosis of colorectal cancer. Cancer 1986; 58:603.)

A

B Figure 123-22.  Pathologic features that can influence prognosis adversely include vascular invasion (A) and lymph node invasion (B). The arrow in A points to a lymphatic vessel containing tumor cells. In B, high-magnification microscopy demonstrates a lymph node that contains adenocarcinoma cells. (Hematoxylin and eosin stain.)

2217

2218

Section X  Small and Large Intestine metastases to a distant site, usually the liver; such metastases carry a poor prognosis at all times in the clinical course. The most important determinant of survival time for patients who present with liver metastases is the extent of hepatic involvement by tumor.

CLINICAL FEATURES CRCs grow slowly and may be present for as long as five years before symptoms appear. Asymptomatic persons with cancer often have occult blood loss from their tumors, and the bleeding rate increases with tumor size and degree of ulceration (Fig. 123-23). Symptoms depend to some extent on the site of the primary tumor. Cancers of the proximal colon usually grow larger than those of the left colon and rectum before they produce symptoms. Constitutional symptoms, such as fatigue, shortness of breath, and angina, secondary to microcytic hypochromic anemia, may be the principal presentation of right colon tumors. Less often, blood from right colon cancers is admixed with stool and appears as mahogany feces. As a tumor grows, it can produce vague abdominal discomfort or manifest as a palpable mass. Obstruction is uncommon with right-sided tumors because of the large diameters of the cecum and ascending colon, although cecal cancers can block the ileocecal valve and cause small bowel obstruction. The left colon has a narrower lumen than the proximal colon, and cancers of the descending and sigmoid colon often involve the bowel circumferentially and cause obstructive symptoms. Patients can present with colicky abdominal pain, particularly after meals, and a change in bowel habits. Constipation can alternate with increased frequency of defecation as small amounts of retained stool move beyond the obstructing lesion. Hematochezia is seen more often with distal than proximal lesions, and bright red blood passed per rectum or coating the surface of the stool is common with cancers of the left colon and rectum. Rectal cancers also cause obstruction and changes in bowel habits, including constipation, diarrhea, and tenesmus. Rectal cancers can invade the bladder, vaginal wall, or surrounding nerves,

A

resulting in perineal or sacral pain, but this is a late occurrence. CRC is often misdiagnosed in symptomatic patients. Symptoms are ascribed to benign conditions such as diverticular disease (abdominal pain, bleeding, change in stool caliber), irritable bowel syndrome (abdominal pain, change in bowel habits), or hemorrhoids (rectal bleeding) (Table 123-12). CRC should be considered when a patient, espe-

Table 123-12  Differential Diagnosis of Colorectal Cancer Mass Lesion Benign tumor (mucosal and submucosal) Endometriosis Inflammatory mass Diverticulitis Infections (tuberculosis, amebiasis, fungal infections) Inflammatory bowel disease Ischemic colitis Solitary rectal ulcer Stricture Crohn’s colitis Ischemia Radiation (late sequela) Rectal Bleeding Diverticulosis Hemorrhoids Infectious colitis Ischemic colitis Solitary rectal ulcer Inflammatory bowel disease Abdominal Pain Diverticulitis Inflammatory bowel disease Irritable bowel syndrome Ischemic colitis Changes in Bowel Habits Infectious diarrhea Inflammatory bowel disease Irritable bowel syndrome Medications This list includes common clinical situations that initially may be confused with symptoms or signs of colorectal cancer, but it is not meant to be inclusive.

B

Figure 123-23.  Colonoscopic view of bleeding carcinomas of the sigmoid colon (A) and cecum (B). Carcinomas of the colon often bleed intermittently. Patients can present with evidence of microcytic anemia, hematochezia, or both, depending on the tumor site and the pattern of bleeding.

Chapter 123  Colorectal Cancer cially one older than 40 years, presents with hypochromic microcytic anemia or rectal bleeding. Too often, anemia in elderly people is ascribed to “chronic disease,” only to be diagnosed later as a sign of advanced CRC. Abdominal pain and bleeding of any type also merit evaluation for cancer in this age group. CRC can affect younger patients, particularly those with IBD or a strong family history for CRC and other cancers. Judicious evaluation of younger patients for CRC is warranted when history and clinical presentation suggest CRC.

DIAGNOSIS AND SCREENING TESTS WHEN COLORECTAL CANCER IS SUSPECTED

When CRC is suspected because of clinical signs and symptoms or when screening suggests the possibility of a colon tumor (discussed later), prompt endoscopic or radiologic diagnostic evaluation should be undertaken (Fig. 12324). Colonoscopy is more accurate than air-contrast barium enema (ACBE), especially for detecting small lesions, such as adenomas smaller than 1 cm; up to one half of adenomas larger than 1 cm may be missed by ACBE.95 If colonoscopy is unavailable, technically difficult, or refused by the patient, an ACBE traditionally has been performed following sigmoidoscopy. ACBEs are more accurate than fullcolumn barium enemas not only for diagnosing cancers but also for detecting small adenomas, which often are present intercurrently. Neoplasms in the rectum and sigmoid colon are sometimes difficult to diagnose radiologically, and proctosigmoidoscopy should be used to complement ACBE. Flexible sigmoidoscopy is superior to rigid sigmoidoscopy. CT colonography (“virtual colonoscopy”) has, in many centers in the United States, replaced ACBE for patients who cannot undergo colonoscopy (see later). If a carcinoma is detected radiologically or by sigmoidoscopy, a full colonoscopic examination should be done because of the high incidence of synchronous lesions and possible implications of the colonoscopic findings for the surgical treatment plan. As many as half of the patients with

A

proven cancers of the colon and rectum harbor additional lesions, and for almost 10%, the operative plan has to be modified as a result of preoperative colonoscopy.

PRINCIPLES OF SCREENING

Cancer prevention may be categorized as primary or secondary. Primary prevention refers to identifying genetic, biological, and environmental factors that are etiologic or pathogenetic and subsequently altering their effects on tumor development. Although several areas of study have been identified that can lead to primary prevention of large bowel cancer, available data do not yet provide a firm basis for the practical application of primary preventive measures. The goal of secondary prevention is to identify existing preneoplastic and early neoplastic lesions and to treat them thoroughly and expeditiously. The assumption is that early detection improves prognosis. In symptomatic patients it is important to minimize diagnostic delay. When the clinical setting suggests CRC (e.g., iron deficiency anemia in an elderly patient), prompt diagnostic evaluation should be undertaken; this approach pertains to individual patients and small groups of patients seen in daily practice and is known as case finding. Screening pertains to large populations. Screening an asymptomatic population for any disease is worthwhile if the disease represents a major health problem, effective therapy is available if the disease is found, a sensitive and specific screening test is available that is readily acceptable to patients and physicians, and the screening test is cost effective. CRC fulfills these conditions, because it represents a major health problem and localized lesions are curable by surgical resection. Furthermore, the prolonged natural history of CRC affords time to detect and eliminate early neoplastic lesions before they reach an advanced, incurable stage. The challenge that remains is to develop effective, easily administered, and cost-effective screening tests for the disease. Current evidence indicates that screening for CRC reduces related mortality. This finding has resulted in a recommendation by the U.S. Preventive Services Task Force (USPSTF) that screening for CRC should be performed in all persons aged 50 years to 75 years.96 Almost all major health-related agencies have endorsed screening for CRC

B

Figure 123-24.  Carcinoma of the cecum. The tumor is seen infiltrating a cecal fold at colonoscopy (A) and air-contrast barium enema (B).

2219

2220

Section X  Small and Large Intestine Table 123-13  Guidelines for Screening Average-Risk Persons for Colorectal Cancer

SCREENING TOOL High sensitivity FOBT (guaiac-based or immunochemical) Flexible sigmoidoscopy Colonoscopy Double-contrast barium enema Computed tomographic colonography Stool DNA testing

U.S. PREVENTIVE SERVICES TASK FORCE*

AMERICAN CANCER SOCIETY, U.S. MULTI-SOCIETY TASK FORCE, and AMERICAN COLLEGE OF RADIOLOGY JOINT GUIDELINES†

Recommended annually as an option

Recommended annually as an option

Recommended every 5 yr + high-sensitivity FOBT every 3 yr as an option Recommended every 10 yr as an option Not recommended Not recommended Not recommended

Recommended every 5 yr as an option Recommended every 10 yr as an option Recommended every 5 yr as an option Recommended every 5 yr as an option Recommended (interval uncertain)

*The U.S. Preventive Services Task Force recommends screening for adults of ages 50 to 75 years. Screening for adults of ages 76 to 85 years is not routinely recommended, and for adults older than 85 years, screening is not recommended. † Testing options are divided into those that detect adenomatous polyps and cancer (flexible sigmoidoscopy, colonoscopy, double-contrast barium enema, computed tomography colonography), and those that primarily detect cancer (FOBT, stool DNA testing). FOBT, fecal occult blood test. U.S. Preventive Services Task Force. Screening for colorectal cancer: U.S. Preventive Services Task Force recommendation statement. Ann Int Med 2008; 149:627-637. Levin B, Lieberman DA, McFarland B, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps 2008: a joint guideline from the American Cancer Society, the U.S. Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. CA Cancer J Clin 2008; 58:130-60.

(Table 123-13), but the key questions of who, how, and how often remain a source of debate. In 1997, the Agency for Health Care Policy Research (AHCPR) authored a 48-page document, in which they set forth clinical practice guidelines for CRC screening, offering a variety of options for CRC screening (fecal occult blood test [FOBT], sigmoidoscopy, colonoscopy, ACBE) and providing a lengthy rationale for each; these guidelines were updated in 2003.97 The American Cancer Society (ACS) followed shortly thereafter with its own set of guidelines for CRC screening and surveillance. These guidelines provided recommendations in three major categories based on risk (average, moderate, and high). Moderate- and high-risk categories were subdivided further according to personal and family history of adenoma, carcinoma, or predisposing disease. Average risk is defined as all people 50 years or older who do not fall into moderate- or high-risk categories. This category encompasses 70% to 80% of the American population. In 2001, the ACS updated its recommendations to offer a broader set of screening choices for different levels of CRC risk,76 thereby allowing greater flexibility in achieving screening goals. This change was deemed necessary because evidence showed little progress in improving CRC screening rates. Screening options included annual FOBT, flexible sigmoidoscopy every five years, annual FOBT plus flexible sigmoidoscopy every five years, ACBE every five years, or colonoscopy every 10 years. These recommendations were modified slightly in 2002 with the addition of immunochemical FOBTs. In 2008, a joint guideline on screening and surveillance for early detection of CRC and adenomatous polyps was issued jointly by the American Cancer Society, the U.S. Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology (see Table 123-13).81 This update of previous guidelines is notable in that it grouped screening tests into those that primarily detect cancer (annual FOBT including those that are guaiac-based or immunochemical tests, and stool DNA tests [interval not specified]), and those that can detect early cancer and adenomatous polyps (flexible sigmoidoscopy every five years, colonoscopy every 10 years, ACBE every five years, or CT colonography every five years). In November 2008, the U.S.

Preventive Services Task Force (USPSTF) also issued updated guidelines (see Table 123-13).96 Based on a targeted evidence-based review and a decision analytic modeling analysis, the USPSTF recommended screening of averagerisk persons aged 50 to 75 years with high-sensitivity FOBT annually, sigmoidoscopy every five years plus FOBT every three years, or colonoscopy every 10 years. Notably, the USPSTF indicated that although the benefits of screening outweigh the potential harms for 50- to 75-year-olds, the likelihood that detection and early intervention will yield a mortality benefit declines after age 75 because of the long average time between adenoma development and cancer diagnosis. Routine screening was therefore not recommended for adults age 76 to 85 years, and screening was not recommended at all in adults older than 85 years. These guidelines also indicated that for all populations there is insufficient evidence to assess the benefits and harms of screening with CT colonography or fecal DNA testing. Although each of the choices has inherent characteristics related to accuracy, prevention, potential costs, and risks, the concept has long been that any one of the tests is better than no test at all. Multiple options can be confusing, however, to both patients and physicians. Furthermore, the test options are not of equal efficacy, and such guidelines might lead to coverage of suboptimal testing by third-party payers. The willingness of patients and physicians to comply with recommendations for screening programs has a major impact on the effectiveness of CRC screening. Compliance by potential screenees and physicians historically has been poor, and interventions to increase screening adherence have been disappointing. The year 2000 goals set forth by the National Cancer Institute in 1996 called for 50% of the population older than 40 years to have had FOBT tests within the prior two years, a goal that was not met. Compliance rates generally are higher for FOBT than for sigmoidoscopy. Clinical trials report compliance rates of 50% to 80% for FOBT among volunteers,98,99 but much lower rates (15% to 30%) are reported from community screening programs. Data from the 1992 National Health Interview Survey indicated that only 17% and 26% of the population older than 49 years reported having an FOBT within the past one and three years, respectively, whereas the median adher-

Chapter 123  Colorectal Cancer Table 123-14  Best Case Values for Screening Procedures for Colorectal Polyps and Cancers* SENSITIVITY SPECIFICITY (%)

Adenomas ≤5 mm (%)

Adenomas 6-9 mm (%)

Adenomas ≥10 mm (%)

Cancer (%)

MORTALITY RATE

Whole colorectum Whole colorectum Whole colorectum

98.0 92.5 95.0

2.0 7.5 5.0

5.0 12.4 10.0

12.0 23.9 22.0

40.0 70.0 70.0

0 0 0

To splenic flexure

92.0

75.0

85.0

95.0

95.0

0

To cecum

90.0

75.0

85.0

95.0

95.0

1/10,000

TEST

REACH

Hemoccult II Hemoccult Sensa Fecal immunochemical test Flexible sigmoidoscopy (within reach) Colonoscopy

*Best case value assumes 100% adherence to screening tests, follow-up of positive findings, and surveillance of persons found to have adenomas. Data from Zauber AG, Landsdorp-Vogelaar I, Knudsen AB, et al. Evaluating test strategies for colorectal cancer screening: a decision analysis for the U.S. Preventive Services Task Force. Ann Intern Med 2008;149:659-69.

ence rate to programmatic offerings of the FOBT was 40% to 50%, depending on the type of population. Adherence to recommended follow-up testing after an initial positive FOBT result also may be lower in the community setting than in larger screening trials. Up to one third of people who test positive might not respond to requests for follow-up. An analysis of diagnostic testing after a positive FOBT result in elderly Medicare recipients indicated that not only was compliance poor but follow-up diagnostic testing often was inadequate or improper. Unfortunately, compliance is often poorer among elderly persons, who are at greatest risk for CRC, and among minorities in whom mortality is high. Despite the availability of flexible sigmoidoscopy, most elderly patients are reluctant to have this test because of cost, discomfort, and fear. Although most physicians agree in principle with guidelines for screening, many do not follow them with all patients. Reluctance to perform what is perceived as an uncomfortable and invasive procedure in asymptomatic persons, requirements for training, and limitations of time and resources contribute to reluctance on the part of primary care physicians. Compliance also is extremely important in any determination of cost effectiveness. Compliance with CRC screening has a major impact on the cost effectiveness of such programs. In one model, the cost per death prevented as the result of performing FOBTs increased from $225,000 to $331,000 as compliance decreased from 100% to 50%.100 The FOBT is especially sensitive to the impact of compliance compared with other tests.101 In the absence of firm clinical data indicating which screening strategy provides the best balance of sensitivity, specificity, logistic feasibility, and cost, various mathematical models have been employed to examine this issue.101-103 The cost-effectiveness of three screening strategies (annual FOBT, flexible sigmoidoscopy every five years, or colonoscopy every 10 years) was compared using a computer model of 100,000 persons 50 years of age.101 This model takes into account the costs of follow-up events. Positive FOBT results or adenomatous polyps found at flexible sigmoidoscopy are worked up using colonoscopy. After polypectomy, colonoscopy is repeated every three years until no polyps are found. This study indicated that colonoscopy represented a costeffective means of screening for CRC because it reduced mortality at relatively low incremental costs. Compliance rates render colonoscopy every 10 years the most cost-effective primary screening strategy for CRC, according to this study. Previous studies reviewed by the USPSTF indicated that CRC screening is likely to be cost effective (<$30,000

Table 123-15  Costs of Screening Tests for Colorectal Cancer PROCEDURE Fecal occult blood test Flexible sigmoidoscopy Colonoscopy Colonoscopy with polypectomy Air-contrast barium enema

COST (U.S.$)* 0 186.00 586.00 808.00 194.00

*Costs are based on Medicare payments in 2008 at the University of Texas MD Anderson Cancer Center and may vary by institution. Charges to patients or other third-party payers are likely to be substantially higher.

per additional year of life gained) regardless of the strategy chosen.79,103,104 A decision analysis commissioned by the USPSTF used microsimulation models from the Cancer Intervention and Surveillance Modeling Network to assess life-years gained and colonoscopy requirements for screening strategies.100 This group concluded that their findings support CRC screening with colonoscopy every 10 years, annual screening with a sensitive FOBT, or flexible sigmoidoscopy every five years with a mid-interval FOBT from age 50 to 75 years. This was part of the basis for modifications to the USPSTF guidelines. The use of screening modalities for detecting adenomatous polyps is discussed in Chapter 122. Table 123-14 presents some of the characteristics of tests used to screen for colorectal neoplasms. Table 123-15 lists the costs of these tests.

SCREENING TECHNIQUES Fecal Occult Blood Testing

Qualitative chromogen tests, which rely on the oxidative conversion of a colorless compound to a colored one in the presence of the pseudoperoxidase activity of hemoglobin, have been standardized using guaiac-impregnated paper and developing solutions (hydrogen peroxide in denatured alcohol). These solutions (e.g., Hemoccult, Hemoccult II) have been widely studied, are available commercially, are convenient to use, and are inexpensive. Their effectiveness in detecting occult blood in the stool, however, depends on the degree of fecal hydration (increases sensitivity), amount of hemoglobin degradation during storage or by focal flora (decreases sensitivity), and the absence of interfering

2221

2222

Section X  Small and Large Intestine substances (e.g., ascorbic acid) that can either enhance or inhibit oxidation of the indicator dye. Any foodstuff that contains compounds with pseudoperoxidase or peroxidase activity can produce a positive FOBT reaction. Red meat and peroxidase-containing foods (broccoli, turnips, cauliflower, radishes, cantaloupe), therefore, should be avoided for three days before and during testing. Although a drop of water added to the slide before development increases sensitivity, this is not recommended for screening average-risk populations, because it gives too many false-positive results. If rehydration is considered, dietary restriction to exclude peroxidase- and heme-rich foods is especially important. Some studies reported falsepositive FOBT results in patients who took supplemental iron preparations, but other studies have reported few or no false-positive FOBT results in these patients. Recommendations for proper performance of these tests are listed in Table 123-16, and the advantages and limitations of the FOBT with the Hemoccult-type slide guaiac tests are outlined in Table 123-17.

Table 123-16  Proper Performance of the Slide Guaiac Test for Fecal Occult Blood For three days before and during testing, patients should avoid the following: Rare red meat Peroxidase-containing vegetables and fruit (e.g., broccoli, turnip, cantaloupe, cauliflower, radish) Certain medications (e.g., iron supplements, vitamin C, aspirin and other NSAIDs) Two samples of each of three consecutive stools should be tested. It is proper to sample areas of obvious blood. Slides should be developed within 4 to 6 days. Slides should not be rehydrated before developing (for average-risk screening). If slides are rehydrated, the patient must have avoided eating red meat; otherwise, too many false-positive results will occur. NSAID, nonsteroidal anti-inflammatory drug.

Table 123-17  Features of the Slide Guaiac Test for Fecal Occult Blood Advantages Readily available Convenient Inexpensive Good compliance in motivated patients Disadvantages Depends on the degree of fecal hydration Affected by storage (hemoglobin degradation can occur) Affected by tumor location Causes of False-Positive Results Exogenous peroxidase activity Red meat (nonhuman hemoglobin) Uncooked fruits and vegetables (vegetable peroxidase: e.g., broccoli, turnip, cantaloupe, cauliflower, radish) Any source of GI blood loss (e.g., epistaxis, gingival bleeding, upper GI tract pathology, hemorrhoids) Certain medications (e.g., iron supplements, vitamin C, aspirin and other NSAIDs) Causes of False-Negative Results Storage of slides for a prolonged period Degradation of hemoglobin by colonic bacteria Ascorbic acid (vitamin C) ingestion Improper sampling or developing Non-bleeding lesion at the time of stool collection GI, gastrointestinal; NSAID, nonsteroidal anti-inflammatory drug.

CRCs and adenomas bleed intermittently, and detecting fecal occult blood by Hemoccult testing depends on the degree of blood loss. In general, 2 mL of blood in the stool is necessary to produce a positive result. Sampling multiple stool specimens therefore is likely to result in fewer falsenegative evaluations. Sampling one specimen yields a 40% to 50% false-negative rate, which improves progressively as more stools are sampled. Two samples of each of three consecutive (daily) stools should therefore be tested. Location of the lesion also affects the ability to detect a cancer by Hemoccult testing. Right-sided cancers produce fewer false-negative tests than cancers elsewhere in the colon, because large bulky tumors bleed frequently. Potential blind spots of the Hemoccult test with high falsenegative rates include the transverse and descending colon. The value of a positive FOBT result performed in conjunction with a digital rectal examination has been disputed, and this FOBT method is not recommended, largely because a single examination done in this way is not nearly as sensitive as the recommended method whereby three consecutive spontaneously passed stools are tested.105 There also will be some false-positive results because digital rectal examination usually is not done in conjunction with dietary restrictions. Rehydration of Hemoccult cards increases sensitivity but reduces specificity, and this method is not recommended for screening.97 In studies that have examined the potential benefit of FOBTs for detecting colorectal neoplasms in large populations, compliance has been in the range of 50% to 70%, although elderly patients—those at substantial risk for colon cancer—tend to be less compliant. The overall positivity rate ranges from 2% to 6% of those tested, and the positive predictive value is about 20% for adenomas and 5% to 10% for cancers. Most studies report that a large percentage of detected cancers are Dukes A and B lesions. Large controlled studies of Hemoccult testing of asymptomatic patients in the general population have been reported from the United States,99,106 Great Britain,98 Scandinavia, and France (Table 123-18).107 These studies cite a rate of test positivity of 1% to 2.6% on first screen for nonhydrated slides and a predictive value for colonic neoplasms (adenomas plus carcinomas) of 22% to 58%. The positive predictive value for carcinomas alone is substantially less (5.6% to 18% for nonhydrated slides). Rehydration of slides with a drop of water before processing results in an increase in positivity and sensitivity but in a decrease in specificity and positive predictive value. An 18-year follow-up in the Minnesota trial106 demonstrated a marked reduction in Dukes stage D cancers in the screened groups compared with the control group. Long-term follow-up of patients tested with Hemoccult in a large group practice setting (Kaiser-Permanente) yielded similar results. The predictive value of a positive test for CRC was 8% at one year, 10% at two years, and 11% at four years. Predictive value depends on what group is screened, and it may be increased in older age groups. Mortality data are available from the Minnesota Study, a randomized, controlled trial that has provided the best evidence for the effectiveness of screening with FOBT.99,106 After 13 years of follow-up, data indicated a 33% reduction in CRC-associated mortality with annual screening but an insignificant reduction of only approximately 5% with biannual screening.99 Approximately 80% of samples were rehydrated, yielding a high positivity rate of 9.8% (compared with 2.4% for nonhydrated slides). These findings resulted in a 38% rate of colonoscopy, leading some to suggest that a substantial portion of the mortality reduction resulted

Chapter 123  Colorectal Cancer Table 123-18  Controlled Trials of Fecal Occult Blood Testing in Screening Asymptomatic People for Colorectal Cancer (CRC) TRIALS VARIABLE

Minnesota99

Nottingham98

Goteborg

Funen

New York

Size of study population Age range Study design

46,000

152,850

28,000

61,933

22,000

50-80 yr Randomized: annual vs. biennial control

50-74 yr Randomized

60-64 yr Randomized

≥40 yr Allocation by month assigned

Rehydration of test cards* Compliance Positivity rate

Yes, most

No

Yes, most

45-74 yr Randomized: biennial vs. control No

Annual 75%; biennial 78% 2.4% (nonhydrated) 9.8% (rehydrated)

50% 1st screen: 2.1% 2nd screen: 1.2%

56% 1st screen: 1.0% 2nd screen: 0.8% 3rd screen: 0.9% 4th screen: 1.3% 5th screen: 1.8%

— Regular attendees: 1.4% 1st screen: 2.6%

PPV for CRC

2.2% (rehydrated) 5.6% (nonhydrated)

1st screen: 9.9% 2nd screen: 11.9%

1st screen: 17.7% 2nd screen: 8.4%

10.7%

CRC mortality†

18-yr follow-up: 33% reduction for the annual group, 21% reduction for biennial group Mortality ratio:   Annual: 0.67   Biennial: 0.79

7- to 8-yr follow-up: 15% reduction in cumulative CRC mortality Mortality ratio: 0.85

— 1st screen: 1.9%(nonhydrated)) 5.8% (rehydrated) 2nd screen: 4.8% (prev. rehydrated) 8.0% (prev. nonhydrated) 1st screen: 5.0% (nonhydrated) 2nd screen: 4.2% (rehydrated) Not yet available

10-yr follow-up: 18% reduction in CRC mortality in the screened group Mortality ratio: 0.82

10-yr follow-up: 43% reduction in CRC mortality in the screened group

No

*Hemoccult test cards were used, rehydrated or nonhydrated. † Reductions in mortality are relative risk reductions. A French trial107 using biennial FOBTs yielded a 16% reduction in CRC-related mortality in the screened group. The mortality ratio was 0.84 (95% confidence interval, 0.71-0.99) with 11 years of follow-up. PPV, positive predictive value; prev., previously.

from chance detection through colonoscopy of nonbleeding cancers. This challenge has been refuted by the investigators, who found that only 6% to 11% of the mortality reduction was explained by chance detection. Results of 18 years of follow-up also have been reported,106 and cumulative 18-year CRC mortality remains 33% lower in the annual group than in the control group. The group tested with biennial screening now demonstrates a 21% lower CRC mortality than did the control group. Other randomized studies reported similar results. Data from Funen, Denmark, suggest an 18% decrease in CRC mortality during a 10-year study period, and data from Nottingham, England, indicate a 15% reduction in mortality at 7.8 years’ followup.98 Data from New York suggest a 43% reduction in mortality in the screened group at 10 years. A randomized French trial106 also demonstrated a reduction in CRC mortality with biennial FOBT screening compared with a control population (mortality ratio, 0.84; 95% CI: 0.71-0.99) at 11 years of follow-up; reduction is mortality was more pronounced in compliant patients (mortality ratio, 0.67; 95% CI: 0.56-0.81). Methods that can decrease the false-positive FOBT rates while maintaining or increasing sensitivity currently are being refined and compared for efficiency with Hemocculttype slide tests. Fecal immunochemical tests (FITs) are designed to detect human globin and are not affected by diet or drugs. One FIT, HemeSelect, showed good performance characteristics compared with standard heme-based FOBT tests in early studies and was used in a combination test

to confirm positive Hemoccult Sensa (a sensitive guaiacbased test similar to Hemoccult) in a large managed care setting. A FIT using a brush-based sampling technique and an immunogold membrane that employs a dual-antibody system specific for human hemoglobin has undergone initial evaluation. Strategies that use an immunochemical-based FOBT have been shown to be cost-effective when used for CRC screening in Japan. The quantitative immunochemical FOBT has been shown to have good sensitivity and speci­ ficity for detection of clinically significant neoplasia in studies of asymptomatic and symptomatic patients,108,109 but test performance in prospective screening programs has been less well studied. Fecal immunochemical tests have, however, been included in screening guidelines.

Proctosigmoidoscopy

The benefit of proctosigmoidoscopy in screening programs for CRC was suggested by several uncontrolled studies that used rigid proctosigmoidoscopy. Those studies suggested that proctosigmoidoscopy in asymptomatic average-risk persons might detect early-stage cancers and that detection and removal of adenomas could result in a lower than expected frequency of rectosigmoid cancers in the screened population. Two case-control studies provided strong evidence that sigmoidoscopy can reduce CRC mortality. A study from the Kaiser-Permanente Medical Care Program110 compared 261 members who died of cancer of the rectum or distal colon with 868 age- and sex-matched control subjects. Only 8.8%

2223

2224

Section X  Small and Large Intestine of case subjects had undergone screening by rigid sigmoidoscopy, compared with 24.2% of controls. Rigid sigmoidoscopy had no effect on mortality in another group whose lesions were beyond the reach of the sigmoidoscope. Furthermore, the beneficial effect of sigmoidoscopy extended 10 years. This and a second case-control study indicate that sigmoidoscopy can result in a 70% to 80% reduction in mortality from cancers within reach of the sigmoidoscope. Because approximately 50% of all CRCs can be detected using the 60-cm flexible sigmoidoscope (see Fig. 123-3), these data suggest that periodic sigmoidoscopic screening could reduce overall CRC-related mortality by about one third. Because the flexible sigmoidoscope is superior to rigid instruments in detecting lesions, flexible sigmoidoscopy has replaced rigid sigmoidoscopy for CRC screening. A casecontrol study using flexible sigmoidoscopy and polypectomy demonstrated a 60% reduction in CRC incidence associated with this procedure. Randomized, controlled trials are now under way to measure the effect of screening with flexible sigmoidoscopy on CRC mortality.94,111 Flexible sigmoidoscopy can be learned by nonphysicians and has been used successfully in screening programs that employ nurse practitioners; wide variations in adenoma detection rates were observed, however, in the UK Flexible Sigmoidoscopy Screening Trial.111 The PLCO study has enrolled 155,000 subjects in a prospective randomized trial that compares flexible sigmoidoscopy to a usual-care control group.94,112 Follow-up is planned through 2015 with cancer-related mortality as the major end point.

Colonoscopy, Barium Enema, and Computed Tomography Colonography

Colonoscopy may well be the most effective tool for CRC screening, but data from prospective randomized trials are lacking. The National Polyp Study of polypectomy and surveillance strongly suggested a reduction in CRC mor­ tality as the result of removing adenomatous polyps compared with historic reference populations.113 A Canadian population-based study compared the risk of developing CRC after a negative colonoscopy in all Ontario residents with a history of a complete negative colonoscopy with controls consisting of the Ontario population without a history of colonoscopy.114 In the negative colonoscopy cohort, the relative risk of distal CRC was significantly lower than the control group in each of the 14 years of follow up, and the relative risk for proximal CRC was significantly lower mainly during the last seven years of follow-up. A second Canadian case-control study demonstrated that complete colonoscopy also was associated with fewer deaths from left-sided CRC but not from right-sided cancer.115 These findings are of interest in light of arguments that colonoscopy is preferable to sigmoidoscopy, because there may be a substantial incidence of proximal colonic cancers and advanced adenomas beyond the reach of the sigmoidoscope. Some of these persons might not have had distal findings on sigmoidoscopy that would have triggered a subsequent colonoscopy. Two trials116,117 suggested that approximately 50% of persons with advanced proximal neoplasms (adenoma >1 cm; adenoma with villous features or dysplasia; cancer) have no distal neoplasms. Less than 2% of those who did not have distal neoplasms, however, had an advanced proximal lesion.116 Given the need for colonoscopic follow-up should FOBT or sigmoidoscopy be positive, colonoscopy also might be

cost-effective.101 A decision analysis commissioned by the USPSTF supports colonoscopy every 10 years as a screening option measured in life-years gained. High-contrast endoscopy using dye or stain solutions combined with colonoscopy (chromoendoscopy), or highresolution optical methods (e.g., narrow-band imaging, laser confocal endoscopy) has been suggested as a means of identifying lesions in high-risk groups or as an adjunct to co­­ lonoscopy where flat lesions (flat adenomas) are suspected. Evidence suggests that flat or depressed neoplasms are more common than previously appreciated and that they carry a high relative risk of containing in situ or invasive carcinoma.118 ACBE has been included as an option in a variety of screening guidelines. No studies, however, have directly addressed the effectiveness of barium enema for CRC screening. Several studies have indicated that the sensitivity of ACBE is less than that of colonoscopy,95 especially for detecting lesions smaller than 1  cm. A population-based study119 suggested that if a cancer is present, there is approximately a one in five chance that it will be missed by ACBE. Computed tomography (CT) colonography, or virtual colonoscopy, uses helical CT to generate high-resolution, two-dimensional images of the abdomen and pelvis. Threedimensional images of the colon can be reconstructed by computer generation offline (Figs. 123-25 and 123-26).120,121 CT colonography has the potential advantage of being a rapid and safe method of providing full structural evaluation of the entire colon. Low sensitivity and specificity and the need for rapid high-resolution helical CT scanners originally precluded its wide application for routine CRC screening, but software and techniques designed to improve the speed, accuracy, and reproducibility of results are now available. Recognition of the importance of CRC screening has raised concerns over the ability of existing resources to handle the ensuing volume of expected procedures such as colonoscopy. Colonography using CT or magnetic resonance imaging (MRI) could represent an alternate method with promise for the future. The accuracy and potential of CT colonography as a screening tool for colorectal neoplasia has been debated because initial studies yielded a wide range of sensitivities.120-125 Two large published multicenter trials fueled this controversy. One trial included 1233 asymptomatic persons who underwent same-day virtual and optical colononoscopy.124 This study employed multidetector CT scanners, three-dimensional endoluminal displays, and solid stool tagging and opacification of luminal fluid (optical cleansing). It demonstrated sufficiently high sensitivity (89% to 94%) and specificity (80% to 96%) for detecting polyps across a broad range of size categories (>6 to >10  mm) to warrant serious consideration as an option for screening. A second study of 600 subjects125 reported a sensitivity for detecting even large polyps (55%) far below that of optical colonoscopy, but it used different technology and methods of analysis compared with the first study. Two published trials provide evidence that CT colonography may be a valid alternative for primary colon cancer screening. The National CT Colonography Trial126 directed by the American College of Radiology Imaging Network (ACRIN) was a multicenter study that employed CT colonography and same-day colonoscopy using a standard matching protocol in 2600 asymptomatic persons. Perpatient sensitivity of CT colonography for adenomas larger than 10 mm was 90%, with a negative predictive value of 99%. A second trial127 compared CT colonography and

Chapter 123  Colorectal Cancer

A

B

Figure 123-25.  Computed tomography (CT) colonography (virtual colonoscopy). An 8-mm sigmoid polyp identified on an axial two-dimensional CT image of the colon (A, arrow) and on an endoluminal three-dimensional reconstruction (B, arrow).

A

B

Figure 123-26.  Representative views of a 2-cm sessile colon lesion (arrows) seen on a fold. Lesion seen at colonoscopy (A) and on three-dimensional reconstruction computed tomography colonography (B).

optical colonoscopy in parallel screening cohorts and demonstrated similar rates of detection of advanced neoplasia in both groups. Several key issues need to be addressed as the use of CT colonography becomes more widespread, principal among which is determination of the acceptable size cut-off of a lesion detected by CT colonography that will necessitate a follow-up colonoscopy. Other issues include the need for bowel preparation, the logistics of same-day colonoscopy, the ability to detect flat lesions, the significance of extracolonic lesions detected by CT, the impact on compliance, and cost-effectiveness. Methodologies that employ CT colonography without cathartic preparation and with fecal tagging120 might make this a more attractive option for screening. CT colonography also might aid in detecting lesions located behind folds or near the anal verge.128

CARCINOEMBRYONIC ANTIGEN AND OTHER TUMOR MARKERS

A great deal of effort has been spent in search of serologic markers that would permit early detection and diagnosis of CRC. A variety of proteins, glycoproteins, and cellular and humoral substances have been studied as potential tumor markers, but none has been found specific for CRC. The most widely studied marker, CEA, may be useful in the preoperative staging and postoperative follow-up of patients with large bowel cancer, but it has a low predictive value for diagnosis in asymptomatic patients. The relatively low sensitivity and specificity of CEA combine to make it unsuitable for screening large asymptomatic populations. Several new protein and carbohydrate antigens are being examined and hold some promise in terms of specificity for preneoplastic and early neoplastic lesions in the colon.65,68

2225

2226

Section X  Small and Large Intestine Their effectiveness for screening, however, remains to be determined. The development of sensitive and specific markers that can be used for early detection of cancer is the focus of the National Cancer Institute-sponsored Early Detection Research Network (EDRN). Promising approaches using genomic or proteomic techniques applied to biomarker discovery are being pursued that might result in practical clinical tests. One such approach involves the study of candidate genes or proteins assembled into panels of markers. Another discovery-based approach uses high-throughput techniques that allow simultaneous assessment of tens of thousands of genes or proteins.129

FECAL DNA AND GENETIC TESTING

A great deal of knowledge has been accumulated about genetic alterations that occur during colon carcinogenesis (see earlier), but specific genetic tests are not yet available for most patients at risk for sporadic CRC. A molecular approach to CRC screening is attractive because it targets biological changes that are fundamental to the neoplastic process. The feasibility of detecting altered DNA in stool has been demonstrated using a multi-target assay panel of molecular markers.130 One multicenter study compared fecal DNA testing using such a panel with FOBT and colonoscopy.131 The fecal DNA panel consisted of 21 mutations: three in the K-ras gene, 10 in the APC gene, eight in the TP53 gene; the MSI marker BAT-26; and a marker of long DNA thought to reflect dis­ ordered apoptosis of cancer cells sloughed into the colonic lumen. Although most of the lesions identified by colonoscopy were not detected by either of the fecal tests, multitargeted fecal DNA testing detected a higher proportion of important lesions compared with Hemoccult. A second study132 compared stool DNA and FOBT for detecting screen-relevant neoplasia (curable-stage cancer, high-grade dysplasia, or adenomas >1  cm). This blinded, multicenter cross-sectional study used two different methodologies for detecting alterations in stool DNA: a 23-marker panel and a new test targeting three broadly informative markers (point mutations on K-ras, a scanned mutator cluster region of APC, and methylated vimentin). Although the multipanel test provided no improvement over FOBT (Hemoccult Sensa) for detecting screen-relevant neoplasms, the new test showed promise by detecting significantly more neoplasms than FOBT detected. Genetic testing is now a reality for families with FAP (see Chapter 122).72 Testing for altered products of the APC gene allows early and accurate identification of family members at risk who require intensive surveillance. Proper genetic counseling, however, must be incorporated into the screening process. Genetic testing for mutations in the hMSH2 and hMLH1 genes is appropriate when HNPCC is suspected, but it presents more difficulty than screening for FAP, because not all the genes involved have been identified, and the preferred method by which families should be screened has yet to be determined.69 A generally accepted approach in persons with suspected HNPCC based on clinical criteria is first to perform MSI testing on the affected person’s tumor using a panel of microsatellite markers. Germline mutation testing for hMLH1 and hMSH2 is performed if the tumor is MSI-high, suggesting a mutation in an MMR gene. In cases in which HNPCC is strongly suspected based on clinical criteria, or when a mutation is established in a family member, germline testing is performed as the first step. If testing for hMLH1 and hMSH2 is negative, but HNPCC still is strongly suspected, germline testing for hMSH6 can be performed.

An allele of APC designated I1307K133 is relatively infrequent in the general population, but it is common in the Jewish population of Ashkenazi descent. There is a modest increase in the relative risk for CRC in those with this allele, but the penetrance for CRC is low compared with carrier frequencies, and genetic testing for I1307K is not recommended.

APPROACH TO SCREENING

Screening and case-finding approaches are different for patients in average-risk and high-risk groups. The former group consists of patients older than 50 years and the latter group includes patients with long-standing UC, previous CRC, previous adenomas, female genital cancer, familial polyposis, HNPCC, and familial colon cancer. Data on the risk of CRC in women with a history of breast cancer are too limited at present to draw firm conclusions regarding appropriate screening intervals.

Average-Risk Group

Patients registered in a health care system should be categorized according to risk, so that appropriate screening can be added to their proper medical evaluation. Relative risk should be assessed by family and personal history. A variety of options are available for screening average-risk patients (∼50 years old with no personal or family history of colo­ rectal adenoma or carcinoma and no personal history of IBD). These have been discussed previously, including guidelines from various health care agencies (see Table 12313). Although yearly FOBTs or flexible sigmoidoscopy every five years is an individual option, it has been suggested that combining the two tests can increase the benefits of either test alone. The tests are complementary because the FOBT has the potential for detecting occult blood from a lesion anywhere in the colon, whereas flexible sigmoidoscopy can detect bleeding and nonbleeding lesions distal to the splenic flexure. Colonoscopy every 10 years has the advantages of examining the entire colon and rectum and providing the opportunity to biopsy or remove lesions should they be found. Growing evidence indicates that colonoscopy is a cost-effective option with an acceptable risk profile. Several trials are under way to examine colonoscopy for averagerisk screening. A diagnostic evaluation is indicated for persons with a positive FOBT or distal neoplasm (adenoma, carcinoma) found at sigmoidoscopy. Colonoscopy is the diagnostic modality of choice. If colonoscopy is unavailable, not feasible, or not desired by the patient, CT colonography or ACBE in combination with flexible sigmoidoscopy is an acceptable alternative to evaluate a positive FOBT result. Screening should be accompanied by programs that educate patients and heighten physicians’ awareness of the concepts and technologies involved in screening, diagnosis, treatment, and follow-up. The popular misconceptions that CRC is an incurable disease and that surgical intervention invariably leads to an impaired life-style, owing to a colostomy, must be discredited.

High-Risk Groups

Familial Adenomatous Polyposis and Familial Cancer Screening of family members in kindreds with familial polyposis is discussed in Chapter 122, and guidelines are available.81 Screening should include genetic testing to detect abnormal APC gene products if a diagnosis can be made by this method in one family member. Those who test positive should have annual or biennial flexible sigmoidoscopy, beginning at age 10 to 12 years, to assess for emer-

Chapter 123  Colorectal Cancer gence of adenomas and to plan appropriate timing for colectomy. If genetic testing is unavailable, annual flexible sigmoidoscopy should begin at age 10 to 12 years. Genetic testing always should be combined with education and counseling of the patient and family members. Colorectal surveillance (sigmoidoscopy) should begin at age 18 to 20 years in persons with attenuated FAP (AFAP) and in persons with MUTYH mutations (in whom colonoscopy is recommended). Patients with a family history of HNPCC must be examined colonoscopically, beginning at age 20 to 25 years, or at an age 10 years younger than that of the index case, because one cannot rely only on the FOBT in these very-high-risk patients. A reasonable approach here is to perform colonoscopy every two years, searching for the scattered adenomas that antedate carcinomas in HNPCC; colonoscopy is more sensitive than radiologic contrast studies. Genetic testing for HNPCC should be offered to first-degree relatives of those with a known MMR gene mutation or to those who meet the modified Bethesda criteria (see Table 123-6).59,69 Genetic testing should be accompanied by counseling of the patient and family members. The benefits of colonoscopic surveillance in patients with HNPCC mutations are suggested by screening trials.134 The approach to patients with a suggestive family history (e.g., one first-degree relative with colon cancer) is not firmly established, but existing data suggest that these patients should be monitored more rigorously than average-risk persons. The joint guidelines from the ACS and US MultiSociety Task Force on Colorectal Cancer recommend that if CRC or adenomatous polyps occurred in any first-degree relative before age 60 years, or in two or more first-degree relatives at any age, then colonoscopy should be performed every five years, beginning at age 40 years or beginning 10 years before the youngest case in the immediate family.81 If either CRC or adenomatous polyps occurred in a first-degree relative 60 years of age or older, or if CRC occurred in two second-degree relatives, then screening should begin at age 40 years using screening options recommended for averagerisk persons. In those with more than two affected firstdegree relatives, special care should be taken to exclude HNPCC, and periodic colonoscopy is advised. Prior Adenomas or Colon Cancer Table 123-19 lists the updated (2008) ACS-Multi-SocietyACR guidelines for screening, surveillance, and early detection of colorectal adenomas and cancer for persons at increased risk or at high risk of disease. These guidelines suggest that those whose index lesion consists of one or two small tubular adenomas with low-grade dysplasia should have a follow-up colonoscopy 5 to 10 years after the initial polypectomy. The precise timing within this interval should be based on clinical factors, such as prior findings, family history, and patient and physician preferences. One study examined the relative risk for advanced neoplasia within 5.5 years of a baseline colonoscopy.135 There was a strong association between the results of baseline screening colonoscopy and the rate of serious incident lesions during surveillance. This study confirmed that patients with one or two small tubular adenomas represent a low-risk group compared with other patients with colorectal neoplasia. In those with a large (>1 cm) adenoma, multiple (3 to 10) adenomas, or adenomas with high-grade dysplasia or villous change, colonoscopy should be repeated within three years of the initial polypectomy. Although the risk for recurrence of advanced adenomas at this follow-up interval is greater in patients with high-risk adenomas than those with low-risk adenomas, the incremental risk is

small.136 If the repeat examination is normal or shows only one or two small tubular adenomas with low-grade dysplasia, then the interval for the subsequent examination should be five years. Patients with more than 10 adenomas on a single examination should have a follow-up colonoscopy less than three years after the initial polypectomy, and the presence of an underlying familial syndrome should be considered. Patients with sessile adenomas that are removed in a piecemeal fashion should have follow-up colonoscopy in two to six months to verify complete removal. Patients with colon or rectal cancer should have highquality perioperative clearing. Colonoscopy should be performed preoperatively, intraoperatively, or within three to six months after cancer resection. Those who have had a colon cancer resected should have colonoscopy performed one year after surgery or the original clearing colonoscopy. If the examination performed at one year is normal, then the interval before the next colonoscopy should be three years; if that examination is normal, the next colonoscopy should be at five years (see Table 123-19). Periodic examination of the rectum to identify local recurrence usually is performed at three- to six-month intervals for the first two or three years after low anterior resection for rectal cancer. Serum CEA levels should be measured at regular intervals because postoperative CEA determinations may be costeffective for detecting recurrent cancers. How long an asymptomatic patient who has had multiple negative examinations should be tested by various modalities is at present unclear. It should be noted that these recommendations are to some extent educated guesses, and not all are based on prospective randomized trials. Inflammatory Bowel Disease The appropriate surveillance schedule for patients with IBD has not been determined in long-term prospective trials. Colonoscopy combined with mucosal biopsy may be effective in detecting preneoplastic and early neoplastic lesions in patients with UC (see Chapter 112). The current recommendation is for colonoscopy every one or two years for patients who have had universal colitis for eight years or left-sided UC for 12 to 15 years (see Table 123-19) (Fig. 12327). Biopsies should be taken throughout the colon at 10-cm intervals, with special attention to areas that suggest a dysplasia-associated lesion or mass (DALM). Although this biopsy procedure enables histology of only a small area of the colon, the short-term risk of carcinoma for patients with negative biopsy specimens is low. If dysplasia is high grade or asso­ciated with a macroscopic lesion or mass, colectomy is recommended. A histologic diagnosis of low-grade dysplasia merits endoscopic follow-up at short intervals, such as three to six months, as does an indeterminate reading resulting from active inflammation. Colectomy has been advocated by some for confirmed low-grade dysplasia. Patients with Crohn’s disease of the colon should be evaluated endoscopically as dictated by symptoms, and special attention should be paid to strictured areas. Studies suggest a role for surveillance colonoscopy in patients with Crohn’s colitis (see Chapter 111).90

INSURANCE COVERAGE FOR SCREENING

Based on evidence from several randomized trials, the Health Care Financing Administration (HCFA) decided to provide coverage for colon cancer screening procedures to Medicare beneficiaries beginning January 1, 1998. After intense lobbying by several groups, Medicare now provides coverage for screening colonoscopy in average-risk persons every 10 years or at an interval four years from a previous sigmoidoscopy. This bill was signed December 21, 2000,

2227

2228

Section X  Small and Large Intestine Table 123-19  Surveillance Guidelines for People at Increased or High Risk of Colorectal Cancer (CRC) RISK CATEGORY

AGE TO BEGIN SURVEILLANCE

Patients with Adenomas at Colonoscopy Patients with 1 or 2 small tubular 5-10 yr after initial adenomas with low-grade dysplasia polypectomy Patients with 3 to 10 adenomas or 1 3 yr after initial adenoma >1 cm or any adenoma polypectomy with villous features or high-grade dysplasia Patients with >10 adenomas on a <3 yr after initial single examination polypectomy Patients with sessile adenomas that 2 to 6 months to verify are removed piecemeal complete removal Patients with CRC 3 to 6 months after cancer resection or intraoperatively

Patients undergoing curative resection for CRC

Patients with a Family History of CRC CRC or adenomatous polyps in a first-degree relative before age 60 yr or in 2 or more first-degree relatives at any age Either CRC or adenomatous polyps in a first-degree relative ≥age 60 yr or in 2 second-degree relatives with CRC Patients at High Risk Genetic diagnosis of FAP or suspected FAP without genetic testing evidence

RECOMMENDed tests

COMMENT

Colonoscopy Colonoscopy

Precise timing is based on clinical factors and on patient and physician preferences If follow-up examination is normal or shows 1 or 2 small tubular adenomas, subsequent examination at 5 yr

Colonoscopy

Consider familial syndrome

Colonoscopy

Surveillance individualized based on endoscopist’s judgment

Colonoscopy

1 year after resection (or 1 year after clearing colonoscopy)

Colonoscopy

Patients with CRC should undergo high-quality perioperative clearing of the colon. For nonobstructing tumors, examination can be done preoperatively; for obstructing cancer, CTC or DCBE can be used to detect proximal neoplasms If exam at 1 year is normal, subsequent exam at 3 yr. If that exam is normal, then subsequent exam at 5 yr Periodic exam of the rectum (3- to 6-month intervals for the first 2-3 yr) may be considered after low anterior resection of rectal cancer

Age 40 yr, or 10 yr before the youngest case in the immediate family

Colonoscopy

Every 5 yr

Age 40 yr

Screening options at intervals recommended for average-risk persons

Screening should begin at an earlier age, but patients may be screened with any recommended form of testing

Age 10 to 12 yr

Annual FSIG to determine if the patient is expressing the genetic abnormality and counseling to consider genetic testing Colonoscopy every 1 to 2 yr and counseling to consider genetic testing (see Chapter 122)

If the genetic test is positive, colectomy should be considered

Genetic or clinical diagnosis of HNPCC or persons at increased risk of HNPCC

Age 20 to 25 yr or 10 yr before the youngest case in the immediate family

Inflammatory bowel disease (ulcerative colitis and Crohn’s colitis)

Cancer risk begins to be significant 8 yr after the onset of pancolitis or 12-15 yr after the onset of left-sided colitis

Colonoscopy with biopsies for dysplasia (see Chapters 111 and 112)

Genetic testing for HNPCC should be offered to first-degree relatives of persons with a known inherited DNA MMR gene mutation. It should also be offered when the family mutation is not known but when 1 or more of the first 3 of the modified Bethesda Criteria (see Table 123-6) is present Every 1-2 yr

CTC, computed tomography colonography; DCBE, double-contrast barium enema; DNA, deoxyribose nucleic acid; FAP, familial adenomatous polyposis; FSIG, flexible sigmoidoscopy; HNPCC, hereditary nonpolyposis colorectal cancer; MMR, mismatch repair. Derived from Levin B, Lieberman DA, McFarland B, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps 2008: a joint guideline from the American Cancer Society, the Multi-Society Task Force on Colorectal Cancer and the American College of Radiology. CA Cancer J Clin 2008; 58:130-60.

and coverage was initiated July 1, 2001. Legislation requiring private insurers to cover CRC screening for any participant or beneficiary older than 50 years or those younger than 50 years who are at high risk for developing CRC has been adopted by several states. The American Gastroenterological Association and the Entertainment Industry Foundation’s National CRC Research Alliance issued the 2004 CRC Screening Legislation Report Card that analyzes the varied and complex state laws gov-

erning insurance for preventive CRC screening. The frequency of screenings complies with current Medicare CRC screening regulations. Data suggest that expansion of Medicare reimbursement to cover CRC screening was associated with an increased use of colonoscopy for Medicare beneficiaries and for those who received diagnoses of colon cancer, and it was associated with an increased probability of diagnosis at an early stage.137 National and regional CRC screening programs have been adopted by several countries

Chapter 123  Colorectal Cancer Extensive colitis for 8-10 years Left-sided colitis for 15 years

Colonoscopy with biopsies × 4 every 10 cm + biopsy suspicious lesions

Low-grade dysplasia

Flat mucosa

Polypoid lesion

High-grade dysplasia

Cancer

Repeat colonoscopy and biopsies at 3-6 months

Negative

High-grade dysplasia or cancer

Repeat colonoscopy in 1 year

Negative for dysplasia or cancer

Repeat colonoscopy in 1 year

Repeat colonoscopy in 2 years

Flat mucosa

Completely removed; no adjacent dysplasia Incompletely removed, or invasive cancer or adjacent dysplasia

Indefinite for dysplasia

Histologic confirmation

Persistent low-grade dysplasia ? Colectomy

Figure 123-27.  Algorithm for colonoscopic surveillance of patients with ulcerative colitis. “Histologic confirmation” refers to agreement by a second experienced pathologist that the biopsy specimen meets the criteria for dysplasia as defined by the Inflammatory Bowel Disease Dysplasia Morphology Study Group. (See Chapter 112.) (Modified from Ahnen DJ. Dysplasia and chronic ulcerative colitis. In Rustgi A, editor. Gastrointestinal Cancers: Biology, Diagnosis, and Therapy. Philadelphia: JB Lippincott; 1995. p 399.)

including Canada (Ontario 2007), New Zealand, and countries in the European Union. The primary modality for screening varies among countries.

SCREENING CAPACITY, SCREENING IN UNDERSERVED POPULATIONS, AND QUALITY ASSURANCE

It has been estimated that only half of the eligible U.S. population has been screened for CRC according to recommended guidelines. Efforts to increase compliance and screening recommendations should take into account the capacity to use various tests for screening and surveillance. A survey by the Centers for Disease Control and Prevention estimated that approximately 2.8 million flexible sigmoidoscopies and 14.2 million colonoscopies were performed in the United States in 2002138; physicians reported the capacity to perform an additional 6.7 million flexible sigmoidoscopy procedures and 8.2 million colonoscopies in one year. A forecasting model139 using data from the U.S. Census Bureau and Centers for Disease Control and Prevention survey indicated that capacity currently exists for widespread screening with the FOBT. The capacity for screening flexible sigmoidoscopy or colonoscopy depends on the pro-

portion of available capacity used for CRC screening. Surveillance colonoscopy needs to be used appropriately as the availability of endoscopic resources decreases. A national survey of colorectal surveillance after polypectomy suggests that resources are being taxed by inappropriate surveillance practices that do not conform to current guidelines.140 Risk stratification will become increasingly necessary as resources become limited.141 Alternative screening modalities such as CT colonography might reduce demand for endoscopic procedures when used to screen low-risk groups, but they will result in increased demand for persons trained in these techniques. The ability of CT colonography to reduce the demand for colonoscopy will also depend on the polyp size that generates a follow-up colonoscopy. Fecal DNA testing could increase compliance and reduce the need for screening colonoscopy if tests were sensitive and specific.142 African Americans in the United States have a higher rate of incidence and mortality from CRC compared with the white population. Compared with white persons, black men and women undergoing screening colonoscopy have a higher risk of larger polyps.143 African Americans who have first-degree relatives with colon cancer are less likely to undergo colonoscopy screening compared with whites who have affected relatives.144 Adherence to screening guidelines

2229

2230

Section X  Small and Large Intestine in urban minority populations may be aided by programs that assist in access and navigation through the health care system.145 Colonoscopy is now the most common endoscopic procedure performed in the United States. As the number of colonoscopies (and colonoscopists) increases, qualityassurance measures will need to be adopted. One measure of quality assurance relates to adequate visualization of the colonic mucosa. One study from a community-based practice146 suggests that detection of overall and advanced neoplasia may be related to withdrawal time during colonoscopy. After implementing a protocol of inspection during a minimum withdrawal time of eight minutes, greater rates of detection were observed. Others, while agreeing that adequate visualization of the mucosa is an important quality assurance parameter, have suggested that adequate examination relates more to the experience and quality of the endoscopist than to the withdrawal time per se. A physician performance measurement set for endoscopy and surveillance has been proposed in a joint document by the American Society for Gastrointestinal Endoscopy, the American Gastroenterological Association, the Physician Consortium for Performance Improvement, and the National Committee for Quality Assurance.

TREATMENT SURGERY

Surgical resection is the treatment of choice for most CRCs. Preoperative colonoscopy should be performed, if possible, to rule out synchronous lesions (see earlier), and serum CEA should be measured for staging and informed postoperative follow-up. Preoperative CT can be valuable for evaluating focal hepatic metastases if partial hepatectomy or chemotherapy is contemplated. It has become standard practice in some centers to provide a trial of systemic chemotherapy in those with hepatic metastases before resecting the primary tumor, because if there is no response, colonic resection will not be curative. CT also is useful for postoperative detection of pelvic recurrence in patients with rectosigmoid tumors. Transrectal ultrasonography or MRI is of value in the preoperative assessment of patients with rectal cancer. The goal of surgery is wide resection of the involved segment of colon, together with removal of its lymphatic drainage (Fig. 123-28). The extent of resection is determined by the blood supply and distribution of regional lymph nodes. The resection should include a segment of colon at least 5  cm on either side of the tumor, although wider margins often are included because of obligatory ligation of the arterial blood supply. The number of lymph nodes recovered during colon cancer surgery has been identified as a potentially important measure of quality cancer care. An analysis of 17 studies suggested that the number of lymph nodes evaluated after surgical resection is positively associated with survival of patients with stage II and stage III colon cancer.147 Sentinel lymph node mapping by dye injection during surgery has not, however, been shown to improve staging accuracy for colon cancer. Minimally invasive laparoscopically assisted surgery may be an acceptable alternative to open surgery for colon cancer in selected patients.142 The approach toward rectal cancers depends on the location of the lesion. For lesions of the rectosigmoid and upper rectum, low anterior resection can be performed through an abdominal incision and primary anastomosis can be accomplished (see Fig. 123-28F ). Surgical treatment of rectal

cancer should employ total mesorectal excision. This technique involves sharp dissection to create an avascular plane between the rectum, the mesorectum (tissue surrounding the rectum which contains lymphatics and vascular structures), and the pelvic side wall. Using sharp dissection, the rectum and mesorectum can be delivered as a single unit. Mesorectal excision is associated with a lower local recurrence rate compared with blunt dissection of the rectum away from surrounding structures. Even for low rectal lesions, a sphincter-saving resection can be performed safely if a distal margin of at least 2  cm of normal bowel can be resected below the lesion, a goal now facilitated by end-to-end stapling devices. Tumor recurrence and survival for rectal cancer are similar after sphincter-saving resections and abdominoperineal resection (APR), if a 2-cm distal margin can be preserved in the former. The inability to obtain an adequate distal margin, the presence of a large, bulky tumor deep within the pelvis, and extensive local spread of rectal cancer all dictate the need for APR, by which the distal sigmoid, rectum, and anus are removed through a combined abdominal and perineal approach and a permanent sigmoid colostomy is established. In a patient with CRC, the primary tumor generally has been resected, even in the presence of distant metastases, to prevent obstruction or bleeding. In patients with advanced disease and multiple medical problems, repeated palliative fulguration of rectal tumors may be preferable to surgery. Modalities such as laser photoablation, argon plasma coagulation, or endoscopic placement of expandable stents represent alternative means of palliation in these patients. Polypoid carcinomas may be removed endoscop­ ically by snare polypectomy techniques (endoluminal resection). Several studies indicate that although the age and phys­ iologic status of a patient can affect operative mortality, advanced age per se does not affect tumor-associated mortality after surgery. Therefore, resection of cancer should not be limited or denied on the basis of age alone.

Follow-up

The incidence of recurrent colon cancer after surgical resection is high in persons who have serosal penetration or lymph node involvement by tumor; the incidence of metachronous CRC is 1.1% to 4.7%. Optimal strategies for surveillance after curative-intent surgery remain uncertain. It is not clear how often, or by what means, a patient should be evaluated following an apparently successful resection for cure. Colonoscopy is beneficial for detecting and removing synchronous and metachronous adenomatous polyps in high-risk groups. History and physical examination, combined with CEA determinations at regular intervals, can provide a costeffective benefit for detecting recurrent cancers. The sensitivity for detecting early recurrences is about 61% using either CT or CEA, but CT can be especially useful in examining the pelvis for recurrence after resection of recto­ sigmoid tumors. CT portography is an accurate method for detecting liver metastases. Immunoscintigraphy after administration of radiolabeled monoclonal antibodies raised against various tumor antigens, including CEA, might provide clinically significant information in staging patients before surgery or in detecting recurrent disease, but use of this modality has not been standardized. The role of positron emission tomography (PET) currently is being evaluated. MRI ultimately might produce the clearest delineation of hepatic metastases. Intraoperative ultrasonography (IOUS) now is being used to increase the ability to detect

Chapter 123  Colorectal Cancer

A

D

B

E

C

F

Figure 123-28.  A-F, Surgical resection (purple wedges) of colorectal cancer based on location of the primary tumor (circles), blood supply, and lymphatic drainage. Insets show anatomy after resection of the tumor and anastomosis.

small and deep hepatic lesions that are not palpable during surgery. Serial CEA determinations have been used to direct second-look surgical procedures. Measuring CEA levels at least every two months for the first two years after resection, and then every four months for the next three years, yields a small percentage of patients (<5%) for whom CEA-directed second-look operations for recurrent carcinoma may be indicated. Survival following second-look procedures is high when surgeons have specialized training in oncologic surgery, but other surgeons have had more limited success, and long-term survival data are lacking. The concept of CEA-directed second-look laparotomy has been applied to resection of localized hepatic metastases. Guidelines for CRC surveillance after primary surgery with curative intent have been produced by a number of

agencies. All emphasize the importance of bowel surveillance with colonoscopy preoperatively, perioperatively, and at subsequent intervals.

Resection of Hepatic Metastases

The liver is the most common site of distant metastases from CRC. Synchronous metastases to the liver are evident at initial presentation in 10% to 25% of patients with colon cancer, and 40% to 70% of those whose cancers disseminate have hepatic involvement. From 70% to 80% of hepatic metastases appear within two years after primary resection. The uniformly poor prognosis for patients with untreated hepatic metastases underlies an aggressive treatment approach.148,149 Hepatic resection is recommended for those whose primary tumor has been resected with curative intent and in whom there is no evidence of extrahepatic disease.

2231

2232

Section X  Small and Large Intestine The extent of liver involvement that is deemed resectable varies from tumor involving one lobe of the liver to focal disease in multiple lobes. The percentage of resectable liver metastases therefore varies in different series from 4.5% to 11% (5% to 6% in most series). Modern techniques of anatomic dissection and hemostasis have resulted in an operative mortality of about 2% in highly trained hands. Dissections along nonanatomic lines have permitted the resection of multiple lesions that previously might have been considered unresectable. Improved preoperative imaging, routine use of IOUS, application of new surgical techniques, and improved perioperative care have increased the number of patients undergoing successful hepatic resections for isolated hepatic metastases. Overall five-year survival rates range from 20% to 45% in selected patients. The literature is difficult to interpret, however, because staging often is not uniform, and prospective controls are lacking. Furthermore, reported two- and three-year survival rates may not be valid, because data suggest that patients with unresected solitary liver lesions live at least three years. Long-term survival for those who undergo surgical resection of hepatic metastases depends on the absence of extrahepatic disease and the ability to achieve adequate surgical margins. In some series, the stage of the primary lesion also is a significant prognostic variable. It is not evident whether patients with a solitary focus of metastasis live longer after resection than those who undergo resection of multiple metastases in the same lobe. It is clear, however, that patients with bilobar metastases are at increased risk for recurrence of metastasis in the liver after resection and that resection should not be attempted when more than four hepatic lesions are present. In patients whose tumor recurs after hepatic resection, the liver is the initial site of recurrence in about 35%. Repeat hepatic resection for isolated metastases can result in long-term survival in selected patients. Improved survival after pulmonary resection of metastases from CRC also has been reported. Patients with up to three nodules in the lung who are surgical candidates should be considered for resection. Combined pulmonary and hepatic resection of metastatic disease has been used in selected cases. Cryotherapy is a technique by which rapid freezing results in crystal formation with significant cellular damage and cell death. Tumors are frozen rapidly by means of a probe with IOUS guidance, so that malignant lesions can be ablated while the remaining liver tissue is preserved. Radio­ frequency ablation involves the use of radiofrequency energy to produce tissue destruction150 and often is performed during an open surgical procedure employing ultrasoundguided needle electrodes that are inserted into the tumor. These are alternative approaches to treatment in patients whose liver metastases are unsuitable for surgical resection. In patients with normal hepatic parenchyma, preservation of a perfused segment of liver accounting for 25% of total hepatic volume has been considered sufficient to prevent postoperative hepatic insufficiency. Preoperative portal vein embolization has been proposed as a means of initiating hypertrophy in segments of liver that would remain following a major liver resection and is under investigation.151

CHEMOTHERAPY Adjuvant Chemotherapy

The prognosis for patients with CRC who undergo potentially curative surgery is correlated strongly with the stage

of the primary tumor at surgery. Despite resection of all macroscopic tumor, patients whose primary tumor has penetrated the serosa or is associated with regional lymph node metastases at the time of surgery have high recurrence rates (see Tables 123-9 and 123-10). The risk of relapse following surgery ranges from 20% to 30% for stage II disease to 50% to 80% for stage III disease. Patients who undergo aggressive surgical resection of isolated hepatic or pulmonary metastases also have high tumor recurrence rates in the liver, lung, and elsewhere. An effective adjuvant program to eradicate microscopic tumor foci is needed for such high-risk patients, who number 35,000 to 40,000 each year in the United States and 200,000 worldwide.152 The principle behind such adjuvant therapy is that treatment is most effective when tumor burden is minimal and cell kinetics are optimal. Data from numerous studies have now demonstrated delays in tumor recurrence and increases in survival for specific groups of patients with CRC who have received adjuvant therapy within eight weeks of surgery. The major advance in the adjuvant treatment of CRC came with the results of trials that explored the combination of 5-fluorouracil (5-FU) and levamisole. A large study assessed the benefit of this regimen in 1296 patients with resected colon cancer that either was locally invasive (Dukes B2; stage II) or had regional lymph node involvement (Dukes C; stage III). Therapy with 5-FU plus levamisole reduced the relative risk of cancer recurrence by 42% and reduced the overall death rate by 33% compared with surgery alone in patients with stage III disease. The results in patients with stage II disease were equivocal and too preliminary to allow firm conclusions to be drawn. Levamisole alone, the mechanism of action of which is not clearly understood, had no detectable effect. Data on all 929 eligible patients followed for at least five years confirm that 5-FU plus levamisole reduced the recurrence rate by 40% and the death rate by 33%. The major effect was on reduction in recurrence at distant sites, such as liver and lungs. Based on these data, adjuvant therapy with 5-FU and levamisole is offered to patients with Dukes C (stage III) colon cancer. The success of combinations of 5-FU and leucovorin for the treatment of advanced CRC led to a trial of this regimen in the adjuvant setting. Several trials suggested that this combination was successful in prolonging disease-free and overall survival.153 Comparison of 5-FU plus levamisole and 5-FU plus leucovorin for the adjuvant treatment of CRC in randomized clinical trials indicated an advantage in disease-free and overall survival in favor of 5-FU plus leucovorin. Review of the combined data suggests that 5-FU plus levamisole given for one year is an effective regimen, but 5-FU plus leucovorin given for six months after curative surgery is superior with regard to convenience and efficacy. Capecitabine, an oral fluoropyrimidine, was approved in 2001 for treatment of metastatic disease (see later). The X-ACT trial was designed to compare capecitabine as an alternative to bolus 5-FU plus leucovorin as first-line therapy for metastatic disease. This trial demonstrated that disease-free survival was at least equivalent with the two regimens.154 Other combined adjuvant chemotherapies for colon cancer currently are being studied, including regimens containing oxaliplatin and irinotecan. The use of combination therapies with either irinotecan and 5-FU plus leucovorin (IFL) or oxaliplatin and 5-FU plus leucovorin (ROX) has demonstrated improved response and survival in advanced CRC (see later). The European MOSAIC trial155 documented significant improvement in three-year disease-free survival when oxaliplatin was added to infused 5-FU plus leucovorin in the FOLFOX regimen for patients with stage II and III

Chapter 123  Colorectal Cancer 100

100

90

90

70 60 50 40

18q LOH

30 20

60 50 40

MSS with 18q LOH

30 p = 0.009

10

0

0 0

1

2

3

A

4

5

6

7

8

9

10

11

0

1

3

100

90

90

80

80

MSI

70 60 50

MSS

40 30 20

6

7

8

9

10

11

70 60 50 40

MSI without TGF-β1 RII mutation

30

p = 0.04

10

0

5

MSI with TGF-β1 RII mutation

20 p = 0.20

4

Years after surgery

100

10

2

B

Years after surgery

Survival (%)

Survival (%)

70

20 p = 0.005

10

0 0

C

MSS without 18q LOH

80

No 18q LOH Survival (%)

Survival (%)

80

1

2

3

4

5

6

7

8

9

10

0

11

Years after surgery

D

1

2

3

4

5

6

7

8

9

10

11

Years after surgery

Figure 123-29.  Molecular markers that influence survival of patients with colorectal cancer after adjuvant chemotherapy with 5-fluorouracil-based regimens. LOH, loss of heterozygosity; MSI, microsatellite instability; MSS, microsatellite stability; TGF-β1 RII, transforming growth factor-β1 receptor II. (From Watanabe T, Wu T-T, Catalano PJ, et al. Molecular predictors of survival after adjuvant chemotherapy for colon cancer. N Engl J Med 2001; 344:1196.)

colon cancer; six-year follow-up data demonstrating significant overall survival in patients with stage III disease are now available. Similar results have been obtained by the Surgical Adjuvant Breast and Bowel Project (NSABP) using bolus 5-FU and leucovorin. These results have led to recommendations favoring combinations of oxaliplatin with intravenous 5-FU and leucovorin as optimal adjuvant therapy.156 Two U.S. cooperative trials will evaluate the addition of antiangiogenesis therapy with bevacizumab (Avastin) to chemotherapy. It is not clear whether patients with stage II, nodenegative colon cancer should receive adjuvant chemotherapy, because the risk-to-benefit ratio in this case has not been established. The Quasar study suggested that adjuvant chemotherapy with 5-FU and leucovorin could improve survival in patients with stage II colon cancer, but the benefit was small.157 Currently, the standard of care is to treat all patients with stage III disease and high-risk patients with stage II disease with adjuvant therapy, although such treatment of the latter group is controversial.158 Anatomic or biological features could, in the future, define subsets of patients with stage II colon cancer who will benefit from adjuvant therapy. Such features might include colloid, signet-ring, or poorly differentiated cancers; high preoperative CEA cell levels; aneuploid DNA content or high S phase; alterations in molecular markers; and the expression of certain tumor-associated antigens (e.g., sialyl-Tn, sialyl Lewisx) or other genetic determinants. One analysis57 indicated that retention of 18q alleles in micro-

satellite-stable cancers and mutation of the gene for TGF-β1 in cancers with high levels of MSI indicate a favorable outcome after adjuvant therapy with 5-FU–based regimens in patients with stage III colon cancer (Fig. 123-29). Tumor MSI status also has been shown to predict benefit from 5-FU–based adjuvant therapy for colon cancer.62 Patients who have completed resection of isolated liver or lung metastases also should be offered adjuvant chemotherapy.149 Portal infusion of chemotherapeutic agents as adjuvant therapy reduces liver metastasis, but this approach has been limited to investigational use. Additional options include systemic regimens with activity in disseminated metastatic disease. Adjuvant therapy for rectal cancer should be considered separate from that for colon cancer, because patterns of failure are different. Local recurrence for stage II rectal cancer after primary resection approaches 25% to 30%, with a 50% or greater local recurrence rate in those with stage III tumors. Local recurrence is associated with significant morbidity, and patients with locally invasive rectal cancer are at high risk for systemic relapse. Surgery for rectal cancers usually includes complete pelvic extirpation with total mesenteric excision (TME). Studies since the 1980s have shown a significant decrease in local recurrence of rectal cancer in patients who receive moderate to high doses of preoperative and/or postoperative radiation (40 to 50 Gy) but little impact on systemic recurrence and survival. Combined adjuvant radiation and chemotherapy has been used to address this potential for local and systemic

2233

Section X  Small and Large Intestine

Patients without recurrence (%)

100 80

Radiation + chemotherapy (N = 104)

60

40

Radiation alone (N = 100)

20

P = 0.0016

0 0

1

A

3

2

5

4

Years after randomization 100 Radiation + chemotherapy (N = 104)

80 Patients surviving (%)

2234

60

40

Radiation alone (N = 100) P = 0.025

20

0 0

B

1

2

3

4

5

6

7

Years after randomization

Figure 123-30.  Combined-modality adjuvant therapy of stages II and III rectal cancer. Effects of radiation and chemotherapy on tumor recurrence (A) and survival (B). (From Krook JE, Moertel CG, Gunderson LL, et  al. Effective adjuvant therapy for high-risk rectal carcinoma. N Engl J Med 1991; 324:709.)

recurrence. Encouraging results were seen in early prospective randomized trials to evaluate the efficacy of combinedmodality adjuvant therapy in patients with modified Dukes B2 and C (stages II and III) rectal cancer following curative surgery. A trial by the North Central Cancer Treatment Group evolved from this earlier work and strongly suggested that postsurgical combined-modality therapy decreases tumor relapse and improves survival over surgery alone or full-dose postoperative radiation therapy. This trial randomized patients with stage II or stage III rectal cancer to receive postoperative radiation alone or radiation plus 5-FU and methyl-CCNU. After a median follow-up of more than seven years, combined therapy significantly reduced local and overall recurrence and distant metastasis and improved patient survival over that with radiation alone (Fig. 123-30); cancer-related deaths and overall death rate were reduced by 36% and 29% respectively. Combining protractedinfusion 5-FU with radiation therapy improved the effect of combined-treatment postoperative adjuvant therapy in patients with high-risk rectal cancer. Based on these data, patients with resected rectal cancer and transmural extension (stage II; Dukes B2) or with positive lymph nodes (stage

III; Dukes C) have been considered for such combinedmodality therapy. Subsequent intergroup trials have attempted to identify the optimal chemotherapeutic agents and their best method of delivery.159 These trials have compared a variety of chemotherapeutic agents in combination with postoperative radiation as adjuvant treatment of rectal cancer. Regimens include 5-FU alone, 5-FU plus leucovorin, 5-FU plus levamisole, 5-FU plus leucovorin plus levamisole, and oxaliplatin plus irinotecan. Comparisons also have included the relative benefits of continuous infusion 5-FU during preradiation and postradiation therapy phases versus bolus 5-FU during the nonradiation therapy portion of the regimen. The term neoadjuvant therapy refers to the use of chemotherapeutic agents or radiation therapy before surgery in patients with advanced but locally confined malignancy. Neoadjuvant therapy allows radiation to be delivered in a nonoperated abdomen, reducing the chance of postoperative complications, such as adhesions and bowel damage; higher doses of preoperative (vs. postoperative) radiation can be delivered. Several trials are ongoing to evaluate the efficacy of preoperative versus postoperative multimodality adjuvant therapy (radiation and chemotherapy) for rectal cancer. Studies from Europe and the United States have suggested less toxicity from preoperative radiation therapy. A prospective, randomized European trial160 comparing pre­ operative and postoperative combined-modality therapy demonstrated a significant reduction in local tumor relapse and less toxicity from preoperative combined-modality therapy compared with similar treatment given postoperatively. Preoperative combined chemoradiation was compared with radiation alone in a prospective randomized trial from the European Organization for Research and Treatment of Cancer (EORTC). Combined preoperative therapy was associated with a better pathologic response and a higher rate of conservative surgery compared with radiation alone, but there was no difference in disease-free or overall survival. Oral capecitabine (Xeloda) is being studied as an enhancer of radiation therapy in the preoperative setting. Accurate endorectal ultrasound and MRI staging has allowed the appropriate use of preoperative therapy, enabling the exclusion of patients with early-stage disease. Adjuvant chemotherapy of approximately four months’ duration currently is recommended for all patients with stage II/III rectal cancer following neoadjuvant chemoradiation and surgery,161 although few studies have evaluated the role of adjuvant therapy for rectal cancer, and its role is not well defined.

Chemotherapy for Advanced Disease

Patients presenting with operable CRCs have benefited from improvements in surgical techniques and advances in adjuvant chemotherapy. Approximately 30% to 40% of patients with CRC, however, have locoregionally advanced or metastatic disease on presentation. Furthermore, the five-year survival rates for patients with stages II and III CRC (82% and 57%, respectively) indicate that a significant portion of these patients will have postsurgical recurrences and die. Systemic chemotherapy therefore is required for a large number of patients with advanced CRC. 5-FU is a fluoropyrimidine that, since the 1970s, has remained the mainstay of systemic chemotherapy for advanced CRC. 5-FU interacts with thymidylate synthetase and inhibits the methylation of deoxyuridylic to thymidylic acid, thereby inhibiting DNA synthesis. It has been admin-

Chapter 123  Colorectal Cancer 1980

1985

1990

5-FU

1995

2000

2005

Irinotecan Capecitabine Oxaliplatin Cetuximab Panitumumab Targeted therapies Bevacizumab

Cytotoxic therapies

Therapeutic concepts Palliative chemotherapy Adjuvant chemotherapy Neoadjuvant chemotherapy

istered orally, intravenously in bolus doses, or by continuous infusion, and it is associated with response rates of 15% to 20% in most studies when used as a single agent. Responses are often short-lived (four to five months) and have not been associated with long-term survival. Toxicity of 5-FU includes myelosuppression, vomiting, diarrhea, and stomatitis and varies according to dose and mode of administration. Despite the approval of a number of new drugs for treatment of metastatic CRC, 5-FU remains a component of most regimens. Various regimens combine 5-FU with high-dose leucovorin (tetrahydrofolate), because leucovorin potentiates the binding of 5-FU to thymidylate synthetase, and the combination is more effective than 5-FU alone. Combined data from numerous trials indicate a two-fold increase in tumor response rates with 5-FU plus leucovorin compared with 5-FU alone (23% vs. 12%), and a small increase in survival.162 The optimal doses of 5-FU and leucovorin and the optimal mode of administration (bolus vs. infusion) remain controversial, but continuous-infusion 5-FU appears to be superior to bolus regimens in terms of response rates, toxicity, and survival compared with bolus regimens.163 Capecitabine is an oral fluoropyrimidine that is converted to 5-FU in tumor tissues. Two large phase III trials that compared capecitabine with bolus 5-FU suggest similar efficacy but fewer side effects with the oral agent.164 Several phase III trials currently are comparing continuous-infusion 5-FU and capecitabine-based regimens. UFT, an oral 5-FU prodrug composed of a 1 : 4 fixed molar ratio of tegafur and uracil, currently is not approved for use in the United States. The advent of a variety of new agents such as irinotecan (Camptosar), oxaliplatin (Eloxatin), and capecitabine, and molecular-targeted agents, such as cetuximab (Erbitux), panitumab, and bevacizumab has led to a rapid evolution in the systemic treatment of CRC (Figs. 123-31 and 123-32). Second-line chemotherapy also has become standard for appropriate patients in whom first-line therapies have failed. The choice of therapy is individualized, based on performance status, the type and timing of prior therapy, and the differing toxicity profiles of the drugs to be used in various regimens. Irinotecan (CPT-11) is a potent inhibitor of topoisomerase 1, a nuclear enzyme involved in the unwinding of DNA during replication. Weekly treatment with irinotecan and 5-FU plus leucovorin was shown to be superior to the widely used regimen of 5-FU plus leucovorin for stage IV metastatic colon cancer in terms response rate (39% vs. 21%), progression-free survival (7 vs. 4.3 months), and

Figure 123-31.  Historical timeline for advances in chemotherapy for colorectal cancer. 5-FU, 5-fluorouracil.

overall survival (14.4 vs. 12 months).165 A second phase III trial confirmed these results,166 and the combination of irinotecan, 5-FU, and leucovorin (IFL) was approved as firstline treatment for advanced CRC. This regimen is associated with significant side effects because those of irinotecan (diarrhea) and 5-FU (nausea, diarrhea, hematologic toxicity) overlap and may be especially severe with bolus 5-FU. A bimonthly regimen that combines leucovorin and 5-FU bolus and a 46-hour infusion at a high dose plus irinotecan (FOLFIRI) is another first-line therapy option for advanced or metastatic disease. Oxaliplatin is a diaminocyclohexane platinum that, unlike other platinum compounds, does not cause nephrotoxicity and has activity against CRC. Several different combinations of biweekly bolus and infusional 5-FU, leucovorin, and oxaliplatin, collectively called FOLFOX, have been studied for treatment of advanced CRC. FOLFOX4 is a common regimen used worldwide that has been compared with IFL and IROX (irinotecan and oxaliplatin) in phase III trials.163 The response rate with FOLFOX4 (45%) was significantly higher than that observed with IFL (31%) or IROX (35%). The median overall survival durations for patients treated with IFL, IROX, and FOLFOX4 were 15.0, 17.4, and 19.5 months, respectively, suggesting that regimens containing oxaliplatin may be options for first-line treatment of patients with advanced CRC. Guidelines from the National Comprehensive Cancer Network recommend that primary therapy for metastatic disease in a patient with good tolerance consist of fluoropyrimidines (either 5-FU plus leucovorin or capecitabine) and either irinotecan or oxaliplatin plus bevacizumab. Bevacizumab is a recombinant humanized monoclonal IgG1 antibody that acts by binding all isoforms of circulating vascular endothelial growth factor-A (VEGF-A), thus decreasing VEGF-A–mediated angiogenesis and vascular permeability. A phase III trial163 compared IFL with IFL plus bevacizumab and showed that the regimen containing bevacizumab increased overall response rate from 35% to 45% and extended median survival from 15.6 to 20.3 months. Bevacizumab currently is approved for use in combination with intravenous 5-FU–based chemotherapy as first-line treatment of patients with metastatic CRC. Cetuximab is a chimeric antibody directed against EGFR, an important molecule involved with cell cycling, survival, invasion, and metastasis. It has been used, primarily in combination with irinotecan, in irinotecan-refractory patients.165 In an analysis of tumor samples from patients with advanced CRC who were randomized to receive cetuximab plus best supportive care or best supportive care alone,

2235

2236

Section X  Small and Large Intestine Initial therapy

FOLFOX ± bevacizumab or CapeOX ± bevacizumab or FOLFOX or CapeOX ± cetuximab1 Patient can tolerate intensive therapy

or FOLFIRI + bevacizumab or FOLFIRI + cetuximab

Therapy after first progression

Therapy after second progression

or

Cetuximab + irinotecan For patients not able to tolerate cetuximab + irinotecan, consider either cetuximab or panitumumab (not as combination)

or

Cetuximab + irinotecan For patients not able to tolerate cetuximab + irinotecan, consider either cetuximab or panitumumab (not as combination)

Cetuximab + irinotecan For patients not able to tolerate cetuximab + irinotecan, consider either cetuximab or panitumumab (not as combination)

Irinotecan or FOLFIRI

Capecitabine ± bevacizumab or Infusional 5-FU + leucovorin ± bevacizumab or Cetuximab or Panitumumab

FOLFOX or CapeOX

Irinotecan

FOLFOXIRI

Patient cannot tolerate intensive therapy

Clinical trial or best supportive care

FOLFOX or CapeOX or cetuximab + irinotecan For patients not able to tolerate cetuximab + irinotecan consider either cetuximab or panitumumab (not as combination) FOLFOX or CapeOX

5-FU/leucovorin + bevacizumab

Cetuximab + irinotecan For patients not able to tolerate cetuximab + irinotecan, consider either cetuximab or panitumumab (not as combination)

FOLFIRI or Irinotecan or FOLFIRI + cetuximab or cetuximab + Irinotecan

Improvement in functional status

Consider initial therapy as above

No improvement in functional status

Best supportive care

FOLFOX – Infusional 5-FU/leucovorin/oxaliplatin FOLFIRI – Infusional 5-FU/leucovorin/irinotecan CapeOX – Capecitabine/oxaliplatin Figure 123-32.  Options for chemotherapy for advanced or metastatic colon cancer. (From National Comprehensive Cancer Network. Clinical Practice Guidelines in Oncology. Colon Cancer, V.1.2010.) 1 All regimens containing cetuximab or panitumumab should be used only in individuals whose tumors express wild-type K-ras.

cetuximab was found to be of benefit in patients whose tumors expressed wild-type K-ras but not in those whose tumors expressed mutated K-ras167 (Fig. 123-33). Results suggesting that EGFR inhibition is only effective in treating CRCs containing wild-type K-ras have been reported in studies using panitumumab, a humanized monoclonal antibody directed against EGFR. Selective infusion of chemotherapeutic agents into the hepatic arterial system may be employed to treat hepatic metastases. This method delivers more concentrated drug into the tumor capillary bed than do conventional means of delivery. The infusion catheter usually is implanted in the common hepatic artery (via the gastroduodenal artery) at the time of laparotomy. The development of implantable infusion pumps has led to increasing use of such therapy in major centers. Fluorinated pyrimidines, such as 5-FU and floxuridine (FUDR), have high hepatic extraction (80% to 95%), and it is thought that high concentrations of these drugs can be delivered with low systemic toxicity by direct

hepatic arterial infusion. FUDR has received the most attention, and its continuous hepatic arterial infusion to treat hepatic metastases from CRC in patients not previously treated can achieve response rates of 54% to 83%. Criteria for response vary, however, and it is still unclear whether an impact on survival will be observed. Randomized trials of systemic versus intrahepatic infusion of FUDR in patients with liver metastasis have shown significantly higher response rates for intrahepatic therapy, but again, the impact on survival remains unclear. Complications of the procedure, including arterial occlusion, local infection, and catheter leak, occur in a small number of patients. Morbidity of treatment consists of gastrointestinal tract inflammation and ulceration, hepatic injury with elevation in serum bilirubin and aminotransferases, and biliary ductal sclerosis, all of which may be substantial. It has been suggested that alternating hepatic intra-arterial FUDR and 5-FU may produce less toxicity than FUDR alone. Some investigators have combined

Chapter 123  Colorectal Cancer Progression-free survival (%)

MUTATED K-ras

A

100 80 60

Cetuximab plus best supportive care

40

Best supportive care alone

20 P = 0.96

0 0

2 4 Months after randomization

No. at risk Cetuximab plus 75 best supportive care Best supportive 76 care alone

19

7

26

6 3

9

4

Progression-free survival (%)

WILD-TYPE K-ras

B

100 80

Best supportive care alone

60

Cetuximab plus best supportive care

40 20 P <0.001

0 0

No. at risk Cetuximab plus 110 best supportive care Best supportive 105 care alone

2 4 6 8 Months after randomization

10

68

44

24

8

5

41

13

2

1

1

Figure 123-33.  Kaplan-Meier curves of progression-free survival in patients with metastatic colon cancer treated with cetuximab plus best supportive care or best supportive care alone, stratified according to mutation status of the K-ras gene. (From Karpetis CS, Khambata-Ford S, Jonker DJ, et al. K-ras mutations and benefit of cetuximab in advanced colorectal cancer. N Engl J Med 2008; 359:1757-65.)

hepatic artery occlusion or embolization with chemotherapeutic agents (chemoembolization) in an attempt to achieve better response rates in patients with extensive hepatic tumor.

IMMUNOTARGETED THERAPY AND IMMUNOTHERAPY

Use of monoclonal antibodies designed to modulate bio­ logical processes key to tumor growth and behavior such as bevacizumab (directed against circulating VEGF), and cetuximab (directed against EGFR) already have been discussed. Both agents have been approved for treatment of advanced CRC. Advances in immunology, molecular biology, and imaging have led to the development of radiolabeled monoclonal antibodies that can be used to detect CRC metastases (radioimmunodetection). These same antibodies can be linked to cytotoxic agents such as the A subunit of the plant toxin ricin, the toxin A chain of diphtheria, lymphokine-activated killer cells, and chemotherapeutic agents for immune-

targeted therapy. Liposomes containing chemotherapeutic agents can be linked to monoclonal antibodies and delivered in a similar fashion. Most patients treated thus far with such therapy have had advanced disease, and further studies using these agents in adjuvant therapy are needed. Attempts to modulate the immune system of patients with metastatic disease also have been reported.168 A large body of preclinical and clinical evidence has suggested that the immune system can be stimulated against malignant cells by means of active specific immunotherapy strategies. These approaches, including experimental cancer vaccine strategies, currently are limited to clinical trials but might hold promise for the future.

RADIOTHERAPY

Patients with rectal cancers whose tumors have penetrated the bowel wall or who have regional lymph node involvement are at 40% to 50% risk for local recurrence following resection of the primary tumor. Radiation therapy is used preoperatively or postoperatively to decrease local recurrence in those with high-risk rectal and rectosigmoid cancers (stages II and III lesions) or in a combined preoperative and postoperative sandwich approach. This approach also is used to convert unresectable large tumors and those fixed to pelvic organs to resectable lesions. Radiation therapy occasionally is useful for palliation of bleeding and pain resulting from advanced rectal disease. The possible advantages of radiation therapy must be balanced against its potential complications of radiation proctitis and small bowel damage (see Chapter 39). Preoperative radiation alone reduces local recurrence in patients with rectal and rectosigmoid cancers, but there is no convincing evidence that it improves survival. Postoperative radiation therapy generally is restricted to patients at high risk for local recurrence of rectal cancer (penetration of the bowel wall, positive lymph nodes). Prospective, but nonrandomized, series show a substantial reduction in local recurrence for those receiving postoperative radiation therapy (6% to 8% for those receiving radiation vs. 40% to 50% for those receiving surgery alone). A randomized study169 also demonstrated results favoring radiation (an overall reduction in locoregional recurrence from 25% to 16%). Distant metastases remain a problem, however, and it is not clear whether survival is altered substantially. Given the demonstration of decreased recurrence and increased survival in patients with rectal cancer receiving combined preoperative or postoperative radiation and combination chemotherapy, this should be considered the treatment of choice for high-risk patients with transmural tumor extension or lymph node metastases.

ENDOSCOPIC THERAPY

Endoscopic therapy using the neodymium : yttriumaluminum-garnet (Nd : YAG) laser or argon plasma coagulation (APC) has been used to recanalize the rectum as palliative therapy in patients with obstructing rectal cancers who are poor surgical risks or who have advanced stages of malignant disease; such palliation generally has been satisfactory.170,171 Reported complications are bleeding and perforation, but they are fewer than would be anticipated after surgery in these high-risk patients. Electrofulguration using a heater probe device has been reported under similar circumstances. Endoscopy with the use of snare cautery also may be used to remove polypoid lesions (Fig. 123-34), often in a piecemeal fashion. Photodynamic therapy (PDT) has been used to treat patients who are poor surgical risks.170 PDT involves the use of a photosensitizer that is taken up by the tumor and

2237

2238

Section X  Small and Large Intestine

A

B

C

Figure 123-34.  Colonoscopic views demonstrating removal of a polypoid carcinoma by snare cautery in a patient at high operative risk because of intercurrent illness. A, Polypoid carcinoma. B, Piecemeal removal by snare cautery. C, Site of lesion after removal.

administered before photoradiation using a tuneable dye laser delivered through a flexible optical fiber. The photosensitizer, porfimer sodium (Photofrin) has been approved by the U.S. Food and Drug Administration for palliation of esophageal cancers, but its use for rectal cancers has been more limited, given other available options. Limitations of PDT include cost and skin photosensitization; the availability of newer porphyrin derivatives associated with shorter periods of photosensitivity could lead to greater use of this modality. Although strictures occur commonly after PDT for esophageal lesions, it is unclear how common strictures are seen after treatment of rectal lesions. Palliation of obstruction from colorectal lesions also may be accomplished by the use of expandable metal stents.171 Intraluminal stents are being used with increasing frequency for palliation and for relieving colorectal obstruction preoperatively, because repetitive treatments as required for ablative therapies such as argon plasma coagulation and PDT are obviated.

OTHER MALIGNANT COLONIC TUMORS Malignant tumors other than adenocarcinomas rarely originate in the large bowel. These include lymphomas, malignant carcinoid tumors, gastrointestinal stromal tumors (GISTs), and leiomyosarcomas. In addition, lymphomas, leiomyosarcomas, malignant melanomas, and cancers of the breast, ovary, prostate, lung, stomach, and other organs can metastasize to the colon. Lymphomas, GISTs, and malignant carcinoid tumors are discussed in Chapters 29, 30, and 31, respectively; carcinomas of the anal canal are discussed in Chapter 125.

KEY REFERENCES

Ahlquist DA, Sargent DJ, Loprinzi CL, et al. Stool DNA and occult blood testing for screen detection of colorectal neoplasia. Ann Intern Med 2008; 149:441-50. (Ref 132.) Baron JA, Cole BF, Sandler RS, et al. A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 2003; 348:891-9. (Ref 40.)

Baron JA, Sandler RS, Bresalier RS, et al. A randomized trial of rofecoxib for the prevention of colorectal adenomas. Gastroenterology 2006; 131:1674-82. (Ref 47.) Grady WM, Carethers JM. Genomic and epigenetic instability in colo­ rectal cancer pathogenesis. Gastroenterology 2008; 135:1079-1099. (Ref 54.) Imperiale TF, Wagner DR, Lin CY, et al. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000; 343:169-74. (Ref 116.) Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2009. CA Cancer J Clin 2009; 59:225-49. (Ref 1.) Johnson DJ, Chen M-H, Toledano AY, et al. Accuracy of CT colon­ ography for detection of large adenomas and cancer. N Engl J Med 2008; 359:1207-17. (Ref 126.) Karpetis CS, Khambata-Ford S, Jonker DK, et al. K-ras mutations and benefit from cetuximab in advanced colorectal cancer. N Engl J Med 2008; 359:1757-65. (Ref 167.) Kelloff GJ, Lippman S, Dannenberg AJ, et al. Progress in chemoprevention drug development: the promise of molecular biomarkers for prevention of intraepithelial neoplasia and cancer—a plan to move forward,. Clin Cancer Res 2006; 12:3661-97. (Ref 25.) Levin B, Lieberman DA, Mcfarland B, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. CA Cancer J Clin 2008; 58:130-60. (Ref 81.) Lieberman DA, Weiss DG, Harford WV, et al. Five year colon surveillance after screening colonoscopy. Gastroenterology 2007; 133:107785. (Ref 135.) National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology. Colon Cancer, 2008. [Internet] [cited 2009 Jun 23] Available at http://www.nccn.org/professionals/physician_ gls/f_guidelines.asp. (Ref 156.) Soetikno RM, Kaltenbach T, Rouse RV, et al. Prevalence of nonpolypoid (flat and depressed) colorectal neoplasms in asymptomatic and symptomatic adults. JAMA 2008; 299:1027-35. (Ref 118.) Steinbach G, Lynch PM, Phillips RK, et al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 2000; 342:1946-52. (Ref 44.) U.S. Preventative Services Task Force. Screening for colorectal cancer: U.S. Preventative Services Task Force recommendation statement. Ann Int Med 2008; 149:627-37. (Ref 96.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

124 Other Diseases of the Colon and Rectum Arnold Wald

CHAPTER OUTLINE Lymphocytic and Collagenous Colitis  2239 Background  2239 Epidemiology  2240 Pathology  2240 Etiology and Pathogenesis  2240 Clinical and Laboratory Features  2241 Differential Diagnosis  2241 Treatment  2242 Diversion Colitis  2242 Background and Epidemiology  2242 Pathology  2242 Pathogenesis  2243 Diagnosis  2243 Treatment  2243 Nonspecific Colonic Ulcers  2243 Pathogenesis  2244 Clinical Features  2244 Diagnosis  2244 Treatment  2244 NSAID-Induced Colonic Lesions  2244 Pathogenesis  2244 Clinical Features  2245 Diagnosis  2245 Treatment  2245 Dieulafoy-Type Colonic Ulceration  2245 Cathartic Colon  2245 Radiologic and Pathologic Features  2246 Changing Role of Laxatives in Colon Damage  2246

LYMPHOCYTIC AND COLLAGENOUS COLITIS BACKGROUND

Lymphocytic and collagenous colitis are uncommon disorders characterized by chronic watery diarrhea and histologic evidence of chronic mucosal inflammation in the absence of endoscopic or radiologic abnormalities of the colon. The two disorders are distinct histologically but have been grouped under the term microscopic colitis. They differ principally by the presence or absence of a thickened collagenous band, which is located in the colonic subepithelium in collagenous colitis. The term collagenous colitis was used first in 1976 by Lindstrom to describe the findings in a middle-aged woman with chronic diarrhea in whom evaluation was normal except for colonic biopsies that showed a thickened band of subepithelial collagen and increased lymphocytes in the lamina propria.1 Histologically, the subepithelial collagen

Clinical Features  2246 Treatment  2246 Pseudomelanosis Coli  2246 Chemical Colitis  2247 Etiology and Pathogenesis  2247 Prevention and Treatment  2248 Pneumatosis Coli (Pneumatosis Cystoides Intestinalis)  2248 Etiology  2248 Clinical Features and Diagnosis  2249 Pathology  2249 Treatment  2250 Malakoplakia  2250 Etiology  2250 Clinical Features and Diagnosis  2251 Treatment  2251 Colitis Cystica Profunda and Superficialis  2251 Etiology  2251 Clinical Features and Diagnosis  2252 Treatment  2252 Neutropenic Enterocolitis (Typhlitis)  2252 Etiology  2252 Clinical Features and Diagnosis  2253 Treatment  2253 Endometriosis  2253 Etiology and Pathogenesis  2254 Clinical Features  2254 Diagnosis  2254 Treatment  2255

deposits resembled those in the small intestine of patients with collagenous sprue. The term microscopic colitis was used first in 1980 by Read and associates, who detailed a group of patients with chronic idiopathic diarrhea, a subset of which had a normal-appearing colon at colonoscopy but abnormal histopathology on biopsy.2 Subsequent review showed that most of these patients had collagenous colitis, but some had increased lymphocytes in the lamina propria in the absence of a thickened collagen band. The term lymphocytic colitis was proposed in 1989 by Lazenby and associates3 as a more-specific histopathologic diagnosis to distinguish this entity from patterns of microscopic colitis in which other cellular elements such as eosinophils, mast cells, or neutrophils predominate. Whether collagenous and lymphocytic colitis represent two ends of the spectrum of a single disorder or different entities remains uncertain, but their clinical presentations, evaluations, and treatments are similar.

2239

2240

Section X  Small and Large Intestine

Figure 124-1.  Histopathology of lymphocytic colitis. The arrows point to surface epithelial damage with increased numbers of intraepithelial lymphocytes. In addition, there is a superficial plasmacytosis without crypt distortion. Although a few intraepithelial neutrophils may be seen, widespread cryptitis is not a feature of lymphocytic colitis. (Hematoxylin and eosin.) (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine; 1996.)

EPIDEMIOLOGY

Collagenous and lymphocytic colitis occur most commonly between ages 50 and 70 years. Both have a strong female predominance and frequent association with arthritis, celiac disease, and autoimmune disorders. In a large populationbased study in Spain, the demographic features of collagenous and lymphocytic colitis were similar: The disorders were found in 9.5 of every 100 patients with chronic watery diarrhea and normal-appearing mucosa on colonoscopy, of whom 61% had lymphocytic colitis; the incidence rates in the general population of lymphocytic colitis and collagenous colitis were 3.1/100,000 and 1.1/100,000, respectively.4 This latter observation contrasts strikingly with published reports of more than 400 cases of collagenous colitis compared with more than 60 cases of lymphocytic colitis, a finding that suggests there may be a publication bias to explain the discrepancy. The overall mean annual incidence of both colitides was 4.2 per 100,000 inhabitants in Spain, similar to the rates from an epidemiologic study conducted in Sweden5 but lower than the 8.6 cases for 100,000 person years in the United States.6 Although the incidence of these microscopic colitides clearly is higher in older age groups, both entities have been reported in children and teenagers, in whom the clinical presentation is similar to that of adults.

PATHOLOGY

In both collagenous and lymphocytic colitis, there is a modest increase of mononuclear cells within the lamina propria and between crypt epithelial cells, consisting mainly of CD8+ T lymphocytes, plasma cells, and macrophages.3 There may be flattening of the surface epithelial cells, a mild decrease in the number of goblet cells, hyperplasia of Paneth cells, and an increased number of intra­ epithelial lymphocytes (Fig. 124-1). Neutrophils are not prominent, and cryptitis and crypt distortion are unusual. In collagenous colitis, there is a thickened subepithelial collagen layer, which may be continuous or patchy (Fig. 124-2). In normal colon, the width of this collagen band is less than 4 to 5 µm and consists predominantly of type IV collagen, whereas in collagenous colitis, it is greater than 10 µm, averages 20 to 60 µm,7 and is composed of type VI collagen and tenascin as well as lesser amounts of types I

Figure 124-2.  Histopathology of collagenous colitis. A thickened irregular subepithelial collagen table (open arrow) with patchy surface epithelial damage is shown. The surface epithelium also contains increased numbers of intraepithelial lymphocytes (closed arrow). In addition, there is a superficial plasmacytosis with prominent numbers of eosinophils. Crypt distortion and crypt abscesses are not usually seen in collagenous colitis. (Hematoxylin and eosin.) (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine; 1996.)

and III collagen. Tenascin is a glycoprotein that is a marker of matrix remodeling and is a product of intestinal subepithelial myofibroblasts.8 These changes are absent in lymphocytic colitis and suggest that the two main forms of microscopic colitis should be considered as separate disease entities. Although inflammatory changes occur diffusely throughout the colon in collagenous colitis, the characteristic collagen band thickening is highly variable, occurring in the cecum and transverse colon in more than 80% of cases and in the rectum in less than 30% of cases. Although involvement of the left colon appears to be less intense than involvement of the right colon, multiple biopsies of the left colon taken above the rectosigmoid during flexible sigmoidoscopy are sufficient to make the diagnosis in approximately 90% of cases.7 The diagnosis of collagenous colitis requires both mucosal inflammation and a thickened collagen band; artifact resulting from poor orientation can give the mistaken appearance of a thickened basement membrane. It has been suggested that tenascin immunohistochemistry be used as a routine test in the diagnosis of microscopic colitis.8

ETIOLOGY AND PATHOGENESIS

The cause(s) of collagenous and lymphocytic colitis is (are) unknown. The most widely held hypothesis is that they are inflammatory disorders arising from epithelial immune responses to intraluminal dietary or bacterial contents. This hypothesis is supported by the regression of inflammation following diversion of the fecal stream and recurrence of inflammation following restoration of intestinal continuity in three patients.9 The identity of the inciting antigenic factors is uncertain, although medications,10,11 bile salts, toxins, and infectious agents12 have been postulated. The strong association of rheumatologic diseases with microscopic colitis has raised the possibility that non­

Chapter 124  Other Diseases of the Colon and Rectum steroidal anti-inflammatory drugs (NSAIDs) might play an etiologic role. One well-controlled study13 found that chronic NSAID use occurred more frequently in patients with collagenous colitis than in age- and gender-matched controls (61% vs. 13%; P < 0.02), a finding confirmed by a case-control study.10 One postulated mechanism by which NSAIDs might damage the colon is by increasing colonic permeability to allow luminal antigens to enter the lamina propria and promote inflammation. Other medications implicated include selective serotonin reuptake inhibitors (SSRIs), specifically sertraline, for collagenous colitis; and SSRIs, beta-blockers, statins, and bisphosphonates for lymphocytic colitis.10 Because many patients with collagenous or lymphocytic colitis have not used these medications, and because use of these drugs in older adults is common but these disorders are uncommon, other causes have been invoked, including genetic susceptibility. Genetic susceptibility is supported by the finding that 12% of patients with microscopic colitis have a family history of inflammatory bowel disease (IBD).11 Approximately 20% to 30% of patients with celiac disease have been reported to have lymphocytic colitis, raising the possibility of similar pathogenetic mechanisms.14 In one study, 40% of patients with collagenous colitis had small intestinal biopsies that were compatible with celiac disease,15 although in another study, the frequency of celiac disease was only 2% of 45 patients with collagenous colitis16 and 9% of 199 patients with lymphocytic colitis.11 Furthermore, patients with microscopic colitis do not respond to a gluten-free diet, and neither collagenous nor lymphocytic colitis is associated with human leukocyte antigens (HLAs) B8 and DR3, as is celiac disease. Finally, CD8+ T intraepithelial cells are predominant in both collagenous and lymphocytic colitis, in contrast to celiac disease, in which CD3 and CD8 predominate. Because autoimmune disorders such as arthritis and thyroid abnormalities have been described in patients with collagenous and lymphocytic colitis,11 there have been continued efforts to associate microscopic colitis with various autoimmune HLA haplotypes and serum markers. One small study showed that HLA-A1 antigens were expressed with increased frequency in lymphocytic but not collagenous colitis,17 and another study showed similar abnormal expressions of HLA-DR antigens by mucosal epithelial cells in both conditions.18 Whether such abnormalities are the cause or the result of these disorders is unknown. Another study found similarities in HLA-DQ loci between patients with celiac disease and patients with either collagenous or lymphocytic colitis.19 Although gluten is not the inciting antigen in microscopic colitis, similar immune mechanisms may be involved in celiac disease and microscopic colitis. The pathogenesis of the increased collagen band in collagenous colitis is unclear. Initially, it had been assumed that collagen synthesis is increased,1 but colonic biopsies from patients with the disease showed decreased levels of interstitial collagenase, suggesting that reduced matrix degradation might contribute to the accumulation of matrix proteins.20 The mechanism of diarrhea in microscopic colitis is related to the severity of inflammation and not the extent or thickness of the collagen band. Perfusion studies have demonstrated defective active and passive absorption of sodium and chloride and reduced chloride-bicarbonate exchange in the colon21; two of six subjects had coexisting abnormalities of small intestinal fluid and electrolyte absorption. Other investigations have correlated colonic fluid absorption with the severity of inflammation.22 A potential role for soluble

mediators is suggested by a report that diarrhea was resolved by a histamine H1 antagonist in a patient with microscopic colitis characterized by increased numbers of mast cells.23 It has been suggested that bile acid malabsorption might contribute to diarrhea in patients with collagenous colitis and that treatment with a bile acid-binding resin such as cholestyramine might lead to a reduction in diarrhea. Bile acids are unlikely to cause the histologic changes observed in collagenous colitis, however, and a reduction in diarrhea with cholestyramine was not associated with a decrease in colitis. Successful treatment of collagenous colitis with budesonide, however, was associated with increased bile acid absorption and normalization of the 75 SeHCAT (selenium-75–labeled homocholyltaurine) test for bile acid malabsorption.24

CLINICAL AND LABORATORY FEATURES

Patients with collagenous and lymphocytic colitis usually present with chronic watery diarrhea, with an average of eight stools each day ranging in volume from 300 to 1700 g per 24 hours21 and associated with occasional fecal incontinence and abdominal cramps. Symptoms decrease with fasting.5 Nausea, weight loss, and fecal urgency are variably present. Diarrhea generally is long-standing, lasting from months to years, with a fluctuating course of remissions and exacerbations. In one series of 172 patients, the median time from the onset of symptoms to diagnosis was 11 months,16 whereas in another smaller series, the median time to diagnosis was 5.4 years.25 Physical examination usually is unremarkable, and blood is not detected in the stool. Routine laboratory studies also are normal. Examination of fresh stools showed fecal leukocytes in 55% of 116 patients with collagenous colitis.16 Mild steatorrhea, mild anemia, low serum vitamin B12 levels, and hypoalbuminemia have been reported in varying numbers of patients and are not characteristic. Autoimmune markers that have been identified in patients with collagenous colitis include antinuclear antibodies (in up to 50%), perinuclear antineutrophil cytoplasmic antibodies (pANCAs) (in 14%), rheumatoid factor, and increased C3 and C4 complement levels,22 but none of these markers is of diagnostic value. Colonoscopic examination usually is normal. Nonspecific abnormalities including patchy edema, erythema, friability, and an abnormal vascular pattern were reported in one study,26 whereas mucosal lacerations in the ascending and transverse colons have been reported in a few patients with collagenous colitis.27

DIFFERENTIAL DIAGNOSIS

Infectious agents should be excluded by testing the stool for enteric pathogens, ova and parasites, and Clostridium dif­ ficile toxin. In many patients incorrectly diagnosed with irritable bowel syndrome (IBS), IBS can be excluded if colonic biopsies are abnormal and if stool volume is increased—both of which are uncharacteristic of IBS. Many other diseases can produce colitis, but they should be distinguishable on histologic grounds. Acute infectious colitis is characterized by neutrophilic inflammation and decreased intraepithelial lymphocytes. Eosinophilic enterocolitis of the mucosal type is characterized by eosinophilic infiltration, shortened crypts, inflammation of the deeper parts of the lamina propria, and absence of increased intraepithelial lymphocytes. Amyloidosis has been mistaken for collagenous colitis, but its distribution includes the basement membranes of the crypts and blood vessels as well as the epithelium; confirmation can be made by histochemical staining. Mild cases of ulcerative colitis and

2241

2242

Section X  Small and Large Intestine Crohn’s disease should present no diagnostic confusion in view of the characteristic endoscopic and histologic findings and absence of increased intra-epithelial lymphocytes. Hormone-producing tumors, surreptitious laxative abuse, and hyperthyroidism can be excluded on clinical and biochemical grounds.

Collagenous or lymphocytic colitis

Antidiarrheal drugs or bismuth subsalicylate

TREATMENT

There have been few controlled trials of treatment for either collagenous or lymphocytic colitis, and therapy is largely empiric. Evaluation of therapy is difficult, because both disorders usually exhibit a relapsing and remitting course over many years. No single agent works in all cases.28 About one third of patients respond to antidiarrheal agents, such as loperamide or diphenoxylate with atropine, as well as bulking agents such as psyllium or methylcellulose; clinical response is not associated with improvement of inflammation or collagen thickness. In an open-label trial of bismuth subsalicylate (eight chewable tablets per day for eight weeks) in 12 patients, diarrhea resolved and stool weight was reduced within two weeks; in nine patients colitis resolved with disappearance of the collagen band thickening.29 Over a seven- to 28-month follow-up, nine patients remained well, two were well but required retreatment, and one had persistent diarrhea. Both collagenous and lymphocytic colitis responded similarly, and there were no side effects of treatment; a subsequent controlled trial by the same investigators published only in abstract form confirmed these findings. Although the basis for its efficacy is unknown, bismuth subsalicylate possesses anti­ diarrheal, antibacterial, and anti-inflammatory properties; bismuth enemas have been reported to be effective in ulcerative colitis and chronic pouchitis.30 Other treatment trials for collagenous and lymphocytic colitis have studied 5-aminosalicylate (mesalamine) compounds, glucocorticoids, and bile acid resins, alone or in combination; these agents appear to improve diarrhea and inflammation in some, but certainly not all, patients.28 Although glucocorticoids given by either the oral or the rectal route provide symptomatic improvement and decrease inflammation in more than 80% of cases, relapse usually occurs quickly after the drug is stopped.5,31 Moreover, long-term use of glucocorticoids has undesirable effects, especially in older patients. Other immunosuppressants, such as azathioprine and 6-mercaptopurine, have been reported to be effective, but there are no sizable studies using these agents.29,32 Budesonide has been reported to be highly effective over a six- to eight-week period in three placebo-controlled trials in patients with collagenous colitis.33-35 Budesonide is a topically acting synthetic corticosteroid with both a high receptor-binding affinity in the mucosa and a high first-pass effect in the liver. In view of its proven efficacy, budesonide should be considered over 5-aminosalicylates or bile acid resins in patients who do not respond to antidiarrheal agents and bismuth subsalicylate.36,37 The only report of surgery for collagenous colitis involved nine patients who underwent ileostomy for disabling refractory collagenous colitis, after which all had symptomatic and histologic remission.28 In patients in whom intestinal continuity was restored, the disease recurred, and of three patients who underwent proctocolectomy with ileal pouchanal anastomosis, problematic diarrhea occurred. Ileostomy should be considered only as a last resort, but it appears to be effective in patients with disabling and refractory symptoms. Based on available data, the treatment algorithm shown in Figure 124-3 is proposed.

No response

Response

Budesonide

Response

No response

Mesalamine

Response

No response

Cholestyramine

No response

No response

Azathioprine or 6-mercaptopurine

Diverting ileostomy – if symptoms are severe

Response

Response

Figure 124-3.  Algorithm for the treatment of collagenous or lymphocytic colitis.

DIVERSION COLITIS BACKGROUND AND EPIDEMIOLOGY

Diversion colitis is an inflammatory process that occurs in the diverted segment of colon and rectum after surgical diversion of the fecal stream. The entity was first reported in 1981 by Glotzer and colleagues in 10 patients who had undergone ileostomy or colostomy for various indications other than IBD.38 Since then, diversion colitis has been found in patients who have undergone surgical diversion for many indications, although it has been reported to occur more commonly in patients with IBD (87%) than in those with noninflammatory conditions (28%).39,40 The prevalence of diversion colitis has been underestimated because many patients are asymptomatic, even though histologic changes are likely to occur in diverted segments of the colon within months of surgical diversion.

PATHOLOGY

A spectrum of histologic changes has been described in diversion colitis, ranging from lymphoid follicular hyper-

Chapter 124  Other Diseases of the Colon and Rectum plasia and mixed mononuclear and neutrophilic infiltration to severe inflammation with crypt abscesses, mucin granulomas, and Paneth cell metaplasia41,42; however, large ulcers and transmural changes are absent and crypt architecture generally is preserved. Endoscopic findings include erythema, friability, nodularity, edema, aphthous ulcerations, exudates, and frank bleeding, as in idiopathic IBD. After extended periods following diversion, inflammatory pseudopolyps and strictures can develop.

PATHOGENESIS

Diversion colitis appears to be caused largely by luminal nutrient deficiency of the colonic epithelium. The principal nutrient substrates of the colonocytes are luminal shortchain fatty acids (SCFAs), which are metabolic products of carbohydrate and peptide fermentation by anaerobic bacteria.43,44 Roediger demonstrated that SCFAs are the major and preferred energy source for colonic epithelium and that the distal colon is more dependent on SCFAs for its metabolic needs than is the proximal colon.45 Butyrate supplies the bulk of oxidative energy to the distal colon, and acetate, glutamine, and ketones provide alternative sources of energy. Harig and associates demonstrated that the excluded segments of colon contain negligible amounts of SCFAs and that infusion of glucose results in no appreciable anaerobic fermentation.46 The number of obligate anaerobes is reduced in the excluded colon, consistent with reduced production of SCFAs.47 Further support for the SCFA nutrient deficiency hypothesis is from the report that instillation of enemas containing SCFAs resulted in disappearance of endoscopic changes within four to six weeks in four patients with diversion colitis, although resolution of histologic abnormalities was slower and incomplete.46 Although SCFA deficiency has been widely accepted as the cause of diversion colitis, other observations suggest that this might not be the entire etiologic explanation. First, studies in children indicate that SCFA enemas are not universally successful in treating diversion colitis.48 Second, in germ-free rodents with surgical diversion and in patients receiving long-term parenteral nutrition or elimination diets (circumstances in which luminal SCFA concentrations are low), mucosal atrophy occurs rather than inflammation.49 Third, inflammation does not occur in urinary colon conduits from which the fecal stream has been diverted, and urine does not contain measurable SCFAs.50 Finally, in a prospective, randomized, double-blind study of 13 patients with diversion colitis, butyrate enemas given for 14 days provided no improvement in either endoscopic or histologic parameters.51 In a subsequent study by the same group, administration of SCFAs did not affect the bacterial population in the excluded colon.52 Other luminal elements besides SCFA deficiency are likely to play a role, but the nature of such factors is unknown.

prominent in diversion colitis.54 If rectal involvement with Crohn’s disease is absent before diversion, rectal inflammation is more likely to be caused by diversion than Crohn’s disease.2,39,55

TREATMENT

The preferred treatment of diversion colitis is surgical restoration of colonic continuity, which rapidly reverses symptoms and histologic changes. If symptoms are moderate to severe and reanastomosis is not feasible, SCFA enemas in a volume of 60 mL and containing a mixture of 60 mmol/L of acetate, 30 mmol/L of propionate, and 40 mmol/L of butyrate with 22 mmol/L of sodium chloride per liter are administered through the anus or mucous fistula twice daily for four weeks and then decreased to once or twice weekly.46 Such preparations are not commercially available and must be formulated by compounding pharmacies, making it the most expensive of the nonsurgical options.56 There are anecdotal reports that 5-aminosalicylate and hydrocortisone retention enemas are effective as well.57 Because they are available commercially, these agents are considered firstline therapies for most patients. One report suggested that intraluminal irrigation with soluble and insoluble fiber solutions improved endoscopic and histologic abnormalities and might be useful to reduce inflammation before surgical restoration of bowel continuity.58

NONSPECIFIC COLONIC ULCERS Benign nonspecific ulcers of the colon are uncommon. The most recent large review of the literature encompassed 127 patients and indicated that such nonspecific ulcers occur at any age, with a peak incidence in the 4th and 5th decades and a slight female predominance.59 Most of these ulcers occur in the proximal colon, virtually all are solitary and located on the antimesenteric side of the colon, and most are round and sharply demarcated from relatively normal surrounding mucosa (Fig. 124-4).59,60 Histologically, there is nonspecific acute and chronic inflammation.59

DIAGNOSIS

The diagnosis of diversion colitis is based on the clinical picture, endoscopic findings, and histology. Diagnosis is relatively straightforward in a patient without preexisting IBD, and stool specimens for C. difficile toxin,53 ova and parasites, and cultures usually are adequate to exclude other etiologies. In patients with a preoperative diagnosis of Crohn’s disease, diversion colitis must be distinguished from recurrent IBD. Colonoscopic findings such as linear ulcers and possibly strictures are said to favor Crohn’s disease, as do transmural inflammation, marked crypt architectural abnormalities, and epithelioid granulomas.39 Lymphoid hyper­ plasia occurs in both disorders but tends to be more

Figure 124-4.  Endoscopic appearance of a nonspecific ulcer of the colon.

2243

2244

Section X  Small and Large Intestine PATHOGENESIS

The causes of nonspecific colonic ulcers are unknown. Potential etiologies that have been suggested, albeit with little or no supporting evidence, include ischemia and cecal diverticulosis. Correlations with the use of drugs such as glucocorticoids, NSAIDs,61,62 oral contraceptives, and oxyphenbutazone have been suggested, but cause has not been established and these drugs have not been implicated in most of the ulcer cases reported. It is likely that no single causative agent explains all cases. There have been reports of associations of nonspecific colon ulcers with chronic renal failure and renal transplantation,63 Churg-Strauss syndrome,64 Wegener’s granulomatosis,65 Behçet’s disease, essential mixed cryoglobulinemia,66 and systemic lupus erythematosus. Perhaps a mechanism common to all exists, but none has yet been identified.

CLINICAL FEATURES

The most common presenting symptoms are abdominal pain and bleeding. More than half of patients with nonspecific colon ulcers present with acute or chronic abdominal pain, often in the right lower abdomen and mimicking appendi­ citis.59 One third have lower gastrointestinal bleeding with hematochezia, and 16% present with an abdominal mass, most often when the ulcer is located in the left or sigmoid colon. A cecal ulcer should be suspected in a patient with gastrointestinal bleeding when the clinical picture is otherwise consistent with appendicitis or in a patient with symptoms suggesting pelvic inflammatory disease, ovarian disease, or Crohn’s disease in the absence of these diseases.

Figure 124-5.  Nonspecific ulcer of the colon. Film from a barium enema examination showing the lesion in Figure 124-4 (arrow). The initial interpretation was carcinoma of the ascending colon.

DIAGNOSIS

Historically, nonspecific colonic ulcers usually were diagnosed at laparotomy after complications occurred. With the advent of endoscopy, many colonic ulcers are now diagnosed preoperatively and, in many cases, are managed conservatively. Colonoscopy currently is the diagnostic test of choice.54 Flexible sigmoidoscopy is inadequate because most colonic ulcers are beyond the reach of the instrument. Abnormalities have been described in up to 75% of air-contrast barium enemas59 and include mucosal irregularities, intraluminal filling defects or narrowing, a mass effect, or localized colonic spasm (Fig. 124-5). Roentgen findings are nonspecific, however, and are diagnostically inferior to direct inspection by colonoscopy. Computed tomographic (CT) scans are most helpful in the presence of perforation or associated abscess formation. The key to the diagnosis of these ulcers is the exclusion of diseases that are associated with ulceration, including Crohn’s disease; infections such as tuberculosis, Entamoeba histolytica, cytomegalovirus, and Salmonella typhi; stercoral ulcers; and solitary rectal ulcer syndrome. Amyloidosis and neoplastic causes such as carcinoma and lymphoma are distinguished on histologic grounds but might not be distinguishable from nonspecific colonic ulcers on the basis of endoscopic appearance alone.

TREATMENT

Surgery is recommended for patients with ulcers complicated by perforation or by significant gastrointestinal bleeding and for those with persistent symptoms and failure of the ulcer to heal. In uncomplicated cases, however, an expectant approach has been advocated, with colonoscopy every six weeks to monitor healing. The most common surgical procedures are local excision of the ulcer, oversewing of the ulcer if there is significant bleeding, and occasionally, segmental colectomy.54,67

NSAID-INDUCED COLONIC LESIONS In contrast to nonspecific colonic ulcers, lesions that are now considered to be pathognomonic for NSAID-induced colitis are known as diaphragm disease68 of the colon. Although it is well known that small bowel strictures are associated with use of NSAIDs, only a small number of cases have been reported since their first published description in 1989.69 The presence of diaphragms with ulcers distinguish these lesions from other nonspecific ulcers. It is believed that the diaphragm-like strictures are due to scarring following an ulcerating injury. Strictures are concentric, often with a pinhole-size lumen. Lesions in the colon are similar to those found in the small intestine (see Chapter 115). The mucosa between diaphragms (if multiple) is normal. Strictures are characterized by submucosal fibrosis with normal overlying epithelium.

PATHOGENESIS

The pathogenesis of these uncommon lesions is unknown but is thought to involve direct contact of the NSAID with colonic mucosa. The vast majority of cases involve oral intake of NSAIDs, and most of the patients in one study were known to have taken extended-release formulations.68 Half had used diclofenac, and most patients had lesions in the proximal colon. The injury is presumed to involve a high local concentration of active NSAID, which increases intestinal permeability, a prerequisite for NSAID-induced enterocolopathy. This is in contrast to gastroduodenal injury, which is thought to be mediated more often by systemic effects. Postulated mechanisms of injury include prostaglandin inhibition alterations in mucosal blood flow and increased permeability to injurious luminal substances. The risk of lower gastrointestinal toxicity is reduced by the use of selective COX-2 inhibitors.70

Chapter 124  Other Diseases of the Colon and Rectum

A

B

Figure 124-6.  A, Colonoscopic photograph showing a diaphragmatic stricture (arrow) associated with chronic use of nonsteroidal anti-inflammatory drugs. B, After balloon dilation there is no residual stricture. (Courtesy of Deepak Gopal, MD, Madison, Wisc.)

CLINICAL FEATURES

The most common clinical presentations include occult blood in the stool, iron-deficiency anemia or frank bleeding, abdominal pain, or change in bowel habits. Other findings can include intermittent bowel obstruction, diarrhea, or colonic perforation with an acute abdomen. Many patients have an underlying rheumatologic disorder and have been using NSAIDs chronically, although lesions have been reported after using NSAIDs for only a few days.71

DIAGNOSIS

As with nonspecific colonic ulcers, colonoscopy is currently the diagnostic test of choice; flexible sigmoidoscopy is inadequate because most diaphragm disease colon ulcers are proximal to the descending colon (Fig. 124-6). Whereas the differential diagnosis of nonspecific colonic ulcers and NSAID-associated colitis and erosions is broad, the presence of characteristic diaphragmatic webs with normal intervening mucosa in a setting of NSAID use is virtually pathognomonic.

TREATMENT

For nonstricturing NSAID lesions, discontinuing the NSAID often is curative and is essential to management; obstructive symptoms associated with strictures, however, require more-aggressive management. For strictures that are easily accessible, endoscopic dilation with a through-the-scope (TTS) balloon has been reported to be safe and effective.72 If endoscopic dilation is not possible, the area should be marked endoscopically to guide the surgeon, because there may be no serosal abnormalities or palpable areas of transition to guide resection. Surgery also is indicated for significant bleeding or perforation and when carcinoma cannot be excluded with confidence.

DIEULAFOY-TYPE COLONIC ULCERATION In 1897, Georges Dieulafoy described massive gastrointestinal bleeding emanating from a relatively enlarged (“persistent calibre”) submucosal artery by way of a minute mucosal ulcer at the most superficial point of the vessel.73 Although originally described in the stomach and most commonly occurring in the gastric fundus, identical lesions have been

described in other gastrointestinal organs, including the colon and rectum.74-76 In the colon, Dieulafoy-type lesions appear to have a strong male predominance and have been reported in all age groups. Histologically, colonic Dieulafoy-type lesions are identical to those found elsewhere in the gastrointestinal tract. The submucosal artery is tortuous and hypertrophic, curving toward the mucosa. Inflammation is absent, and the solitary mucosal ulceration extends no deeper than the upper submucosal layer. The clinical picture is one of acute and massive bleeding. Colonoscopy can identify the lesion in some cases, but in most cases identification is often difficult or impossible, especially when bleeding continues or thorough cleansing of the colon cannot be accomplished. Selective mesenteric angiography is the diagnostic study of choice, and surgical resection has been the principal form of therapy. Even after angiographic detection of the bleeding site, precise localization of the lesion is usually difficult, and extended resection often is required. In some cases, colonic lesions appear as pseudopolyps,74 and successful treatment with sclerotherapy, electrocautery, or endoscopic hemoclipping might obviate the need for surgery.76,77

CATHARTIC COLON Cathartic colon is an uncommon and severe sequela of chronic irritant laxative abuse. In 1943, Heilbrun first described radiologic abnormalities of the colon and terminal ileum associated with prolonged abuse of irritant cathartics.78 Fewer than 50 cases have been reported in the literature, all in women with a duration of laxative abuse ranging from 10 to 70 years. It is important to emphasize that the term cathartic colon is based on barium enema characteristics and is not synonymous with prolonged use of laxatives or with laxative abuse. Indeed, misuse of the term cathartic colon has led to inappropriate concerns over the chronic use of laxatives which, when appropriate, is not associated with structural or functional damage to the colon. Cathartic colon is not the inevitable consequence of chronic laxative abuse, which may be associated with a variety of reversible symptoms as well as fluid and electrolyte abnormalities. In a review of 240 cases of chronic

2245

2246

Section X  Small and Large Intestine laxative abuse published in more than 70 reports, no case of cathartic colon was demonstrated.79,80

RADIOLOGIC AND PATHOLOGIC FEATURES

Heilbrun originally described the following characteristics in his original case report: loss of haustrations, pseudostrictures, dilated colon and terminal ileum, and gaping of the ileocecal valve78; similarity to the radiologic appearance of chronic ulcerative colitis was noted in this and sub­ sequent studies. Characteristic changes are not always found throughout the colon, and there is a predilection for involvement of the ascending colon. Pathologic changes in resected specimens of cathartic colon have included mucosal atrophy, chronic inflammation with thickening of the muscularis mucosae, submucosal fatty infiltration, and mild fibrosis. Irreversible strictures and degenerative changes in intestinal neurons are absent. Neuronal changes have been found in patients with chronic laxative abuse, but these patients did not exhibit cathartic colon as defined here.81

CHANGING ROLE OF LAXATIVES IN COLON DAMAGE

The original suggestion that irritant laxatives, predominantly anthraquinones, damage the colon was based on studies in laboratory animals and in colons resected from laxative abusers.82 Although mucosal atrophy and abnormalities of the enteric nervous system were described, the identities of the laxatives were not documented, nor was there information concerning preexisting conditions that might have prompted chronic use of laxatives. Subsequent studies have reported changes in colonic epithelial cells and the submucosa in patients with long-term laxative abuse, and both anthraquinones and bisacodyl have been implicated.81 The unclear nature and duration of laxative use and the inability to exclude preexisting conditions, however, make the significance of these observations uncertain. Studies in rodents and in chronically constipated women do not support the deleterious effect of anthraquinones on the ultrastructure of colonic nerves,83,84 nor is there evidence to suggest that sennosides, bisacodyl, or related substances cause significant morphologic damage to the colonic enteric nervous system in either experimental animals or humans. Perhaps one or more laxatives that are

A

no longer in use, such as podophyllin, might have accounted for cases of cathartic colon, because no case of cathartic colon has been reported in persons who began to use or abuse irritant laxatives after 1960.85

CLINICAL FEATURES

Habitual laxative users and abusers often complain of abdominal discomfort, bloating, fullness, or inability to defecate completely without using laxatives. In the more-severe cases, electrolyte and fluid abnormalities such as hypokalemia and hypovolemia are associated with excessive thirst and weakness; protein-losing enteropathy is uncommon. All symptoms are reversible on withdrawal of laxatives or conversion to a more-appropriate regimen of laxative use, if there is no underlying diagnosis such as IBS.

TREATMENT

Treatment of cathartic colon and symptoms of chronic laxative use is focused on reducing or eliminating irritant laxatives, substituting bulking or osmotic agents, and retraining the bowel. Although cathartic colon is often thought to be irreversible, there is evidence that it can partially or completely reverse after withdrawal of the offending laxatives. In severe or refractory cases, subtotal colectomy or proctocolectomy has been effective. The cathartic colon is of historic interest and is unlikely to be identified in current clinical practice. There is no evidence that currently used laxatives can produce this entity. The term cathartic colon should not be confused with “chronic laxative abuse syndrome,” nor should the term imply that current laxatives are dangerous if used chronically but appropriately.

PSEUDOMELANOSIS COLI Melanosis coli is a brownish discoloration of the colon mucosa caused by the accumulation of pigment in macrophages within the lamina propria (Fig. 124-7). First described in the early 19th century, the term melanosis coli was coined by Virchow in 1857, because the pigment was thought to be melanin or a melanin-like substance. Subsequently, the pigment proved to be lipofuscin, both histochemically and ultrastructurally.86,87 The term pseudo­

B

Figure 124-7.  A colonoscopic view of pseudomelanosis coli associated with the chronic use of senna laxatives in a patient with ulcerative proctitis. There is little or no pseudomelanin pigment in the distal 30 cm of colon in the presence of active mild colitis (A), in contrast to the heavy pigmentation in the remaining colon (B). (Courtesy of Miguel Regueiro, MD, Pittsburgh, Pa.)

Chapter 124  Other Diseases of the Colon and Rectum Table 124-1  Chemical Agents That Cause Colitis When Given as Enemas

Figure 124-8.  Colonoscopic photograph of a pale colon cancer easily seen against the dark background of pseudomelanosis coli. The neoplasm is pale because it lacks pigment-containing macrophages. (Courtesy of Juergen Nord, MD, Tampa, Fla.)

melanosis coli, though more accurate than melanosis coli, has not been adopted widely. The association between pseudomelanosis coli and chronic use of anthraquinone laxatives is established firmly and is supported further by the development of characteristic pigmentation in laboratory animals after administration of anthraquinones.88 Pseudomelanosis develops in more than 70% of persons who use anthraquinone laxatives (cascara sagrada, aloe, senna, rhubarb, and frangula), often within four months of use, with an average of nine months. The condition is widely regarded as benign and reversible, and the pigment generally disappears within one year of stopping laxatives.89 Pseudomelanosis coli probably can result from other factors or exposure to other laxatives, however, and so its presence is not specific for anthra­ quinone use. The pigment in pseudomelanosis coli is thought to originate from either macrophages or organelles within epithelial cells after cell damage and apoptosis; such a sequence of damage has been demonstrated in guinea pigs exposed to anthraquinones.90 Histologically, the number and size of macrophages within the lamina propria are increased, and the greatest amount of pigment is found in macrophages farthest from the lumen. Abnormalities of colonic epithelial cells are noted on electron microscopy but not on light microscopy.91 Concern about a possible relationship between pseudomelanosis coli and the development of colonic neoplasia has not been substantiated in a prospective case-control study.92 Other confounding factors such as chronic constipation or dietary intake might account for the increased risk of colon cancer suggested by earlier studies.93,94 Colonic neoplasms lack pigment-containing macrophages and therefore are identified more easily in patients with pseudomelanosis coli (Fig. 124-8).95 Biopsies should be taken of any nonpigmented area of the colon in a patient with pseudomelanosis coli who undergoes colonoscopy.

CHEMICAL COLITIS ETIOLOGY AND PATHOGENESIS

Damage to the colon has been reported after exposure to a number of rectally administered agents (Table 124-1), the

Acetic acid Alcohol Ammonia Caustic soda (sodium hydroxide) Chloro-m-xylenol (Dettol) Ergotamine Formalin Glutaraldehyde Herbal medicines Henna Hydrofluoric acid Hydrogen peroxide Lye Potassium permanganate Radiocontrast agents (Hypaque, Renografin-76) Soap Sulfuric acid Vinegar

better known of which are soaps and detergents used as “cleansing” enemas.96,97 Other offending substances include hydrogen peroxide,98 water-soluble contrast agents such as sodium diatrizoate (Hypaque, Gastrografin),99 vinegar, potassium permanganate, herbal medications,100 glutaraldehyde,101,102 formalin,103 and alcohol.104 Milder damage to the mucosa occurs after use of monobasic or dibasic sodium phosphate enemas,105 bisacodyl suppositories, and oral henna.106 Colonic damage presumably occurs from a detergent, hypertonic, or direct toxic effect on the mucosa. The severity of the reaction depends on the type and concentration of the substance, the duration and extent of its contact with the mucosa, and perhaps the presence of underlying colonic disease.107 Soaps consist of a number of substances, including strong alkali, potash, phenol, and sodium and potassium salts of long-chain fatty acids. These agents produce liquefaction necrosis with mild to severe inflammation and saponification of the layers of the colon wall. Acute histologic changes include necrosis, which leads, in more severe cases, to ulceration and formation of granulation tissue. Acute colitis can heal without residua or with fibrosis and scarring, or it can progress to transmural necrosis and perforation. Severity of damage probably is related to the concentration of soap and duration of contact with the mucosa. Endoscopic findings have ranged from loss of the normal mucosal vascular pattern to aphthae to mucosal sloughing and ulceration. Hydrogen peroxide enemas no longer are commonly used, but at one time they were employed to relieve meconium ileus and to remove fecal impactions. There are reports of severe damage associated with use of hydrogen peroxide, including severe colitis, pneumatosis coli, perforation, sepsis, and death.98 Within minutes of contact, diffuse mucosal emphysema occurs, and after about an hour, the colon can become ischemic and eventually ulcerate (Fig. 124-9). Acute colitis also has been reported after using a colonoscope that had been disinfected with glutaraldehyde (Fig. 124-10).101,102 Colitis also has followed the use of several hyperosmolar water-soluble contrast materials that often are employed to opacify the colon in cases of partial obstruction and to treat fecal impactions in adults.99 Damage is believed to occur because of the hypertonicity of these agents, but the addition of Tween 80 to hyperosmolar agents to improve mucosal

2247

2248

Section X  Small and Large Intestine contrast might have contributed to the mucosal damage because of its detergent properties. Most reports of injury have occurred in the colon proximal to an obstruction and mainly in the right colon, suggesting that prolonged contact with these agents predisposes to mucosal injury.

PREVENTION AND TREATMENT

Proper cleaning and rinsing of endoscopes are required to minimize exposure of the patient to injurious disinfecting chemicals. Protocols require strict adherence to proper maintenance and adjustments in the rinse cycle of disinfecting machines.102 Forced-air drying and rinsing of endoscope channels and the exterior of the instrument should ensure a chemical-free procedure. Rectal instillation of substances other than commercially available enemas for medicinal purposes should be discouraged. Patients and health care professionals should be cautioned that soapsuds enemas should not be used. Treatment of chemical colitis is largely supportive, with intravenous fluids and broad-spectrum antibiotics. Surgery is indicated in severe cases of bowel necrosis leading to gangrene or perforation. Most patients recover completely four to six weeks after injury.

PNEUMATOSIS COLI (PNEUMATOSIS CYSTOIDES INTESTINALIS)

Figure 124-9.  Colonoscopic photographs of chemical colitis in a patient to whom hydrogen peroxide was given to help remove a fecal impaction. Both panels show the snow-white sign, referring to the appearance of the stark white necrotic mucosa. (Courtesy of Lawrence J. Brandt, MD, Bronx, NY.)

The term pneumatosis coli is synonymous with pneumatosis cystoides intestinalis when the latter disorder is limited to the colon. This uncommon entity is characterized by multiple gas-filled cysts located in the submucosa and subserosa of the intestine. Pneumatosis cystoides intestinalis must be distinguished from pneumatosis linearis, or gas within the wall of the bowel, which usually is associated with bowel necrosis, signifies loss of bowel viability, and mandates surgery. In acquired immunodeficiency syndrome (AIDS), pneumatosis linearis may be associated with opportunistic infections of the colon and can resolve without surgery if the infection is treated successfully.108 Most cases of pneumatosis cystoides intestinalis occur in the jejunum and ileum, and only 6% of cases involve the colon. There is a propensity for involvement of the left side of the colon. Numerous conditions have been associated with pneumatosis coli, including appendicitis, Crohn’s disease,109 ulcerative colitis, diverticular disease, necrotizing enterocolitis, pseudomembranous colitis,110 ileus,111 and sigmoid volvulus. Pneumatosis also has been associated with nongas­ trointestinal conditions, including emphysema, collagen vascular diseases,111,112 transplantation,112 AIDS,108 glucocorticoid use, chemotherapy,111 and certain medications. In approximately 20% of cases, there are no identified asso­ ciated medical conditions, and pneumatosis is considered primary.113

ETIOLOGY

Figure 124-10.  Colonoscopic photograph of chemical colitis in a young man who had had a normal colonoscopy three hours before this photograph was taken. Colitis was caused by glutaraldehyde that had been used to disinfect the flexible sigmoidoscope after its previous use but that had not been sufficiently cleansed from the instrument. He complained of “agonizing” pain during the sigmoidoscopic examination and began to pass bright red blood two hours later. (Courtesy of Lawrence J. Brandt, MD, Bronx, NY.)

Several theories have been suggested to explain the large and varied number of conditions associated with pneumatosis cystoides intestinalis. The most plausible theories are the mechanical theory and the bacterial theory. According to the mechanical theory, intraluminal gas enters the bowel wall under pressure through a defect or potential defect in the intestinal mucosa. The mucosal defect can result from direct trauma or increased intraluminal pressure. This hypothesis could account for reports of pneumatosis after sigmoidoscopy without biopsy; in cases

Chapter 124  Other Diseases of the Colon and Rectum of colitis, perforated duodenal ulcers, and jejunal diverticula; and after intestinal anastomoses. The plausibility of this theory is diminished by the absence of a connection between the mucosa and the cysts and the presence of elevated levels of hydrogen gas in the cysts.114 The bacterial theory suggests that the cystic gas collections are the by-products of bacteria, specifically those that produce hydrogen. This theory has been supported by clinical observations and laboratory experiments. In laboratory animals, pneumatosis coli can be induced by injecting gasforming bacteria into the bowel wall. In addition to local invasion of the intestinal wall, bacteria can produce gas cysts by manufacturing large amounts of hydrogen gas as a result of the fermentation of carbohydrates. Levitt and Olsson theorized that the high hydrogen tension in the colonic lumen leads to rapid diffusion of this gas into an intramural gas bubble and can cause N2, O2, and CO2 to diffuse from the circulation into the bubble.115 According to their theory, the gas bubble enlarges if there is continued diffusion of hydrogen into it. Indeed, high hydrogen content in the cysts has been documented,116 and cysts regress in patients fed an elemental diet to decrease carbohydrate substrate for colonic bacteria. Two major observations argue against the bacterial theory of pneumatosis: bacteria are not cultured from cysts, and with cyst rupture and pneumoperitoneum, peritonitis is not seen. It has been hypothesized that gas cysts can form by counterperfusion supersaturation of H2 gas in which super H2 production by colonic bacteria provides the condition for H2 tension in the colonic lumen to approach the level of N2 tension in the blood.117 One such mechanism is by exposure to certain drugs, such as chloral hydrate, that inhibit the growth of H2-consuming methanogenic bacteria in the colon and thereby increase net H2 production.117 Another possible etiologic setting is the administration of a nonabsorbable carbohydrate such as lactulose, thereby increasing colonic hydrogen production, in a setting where bacteria metabolizing H2 are deficient.118 Successful treatment with antibiotics116,118 and colonic washouts also supports a bacterial etiology for pneumatosis coli.119 In one study, stools from patients with pneumatosis coli were demonstrated to lack two major species of hydrogen-consuming bacteria.114 Because hydrogen normally is produced only in the colon and not in the small intestine, pneumatosis coli may differ from pneumatosis intestinalis with respect to pathogenic mechanisms.

Figure 124-11.  Film from a single single-contrast barium enema demonstrating the presence of numerous gas-filled cysts (arrowheads) in the wall of the colon characteristic of pneumatosis coli. (From Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine; 1996.)

CLINICAL FEATURES AND DIAGNOSIS

The incidence of pneumatosis coli is highest in the sixth decade, with equal frequency in men and women.113 In most cases, pneumatosis is an unexpected finding on abdominal plain films. The most common symptoms are diarrhea (68%), mucus discharge (68%), rectal bleeding (60%), and constipation (48%).113 Approximately 3% of patients pre­sent with a complication of pneumatosis coli, including pneumoperitoneum, volvulus, intestinal obstruction, intussusception, tension pneumoperitoneum, and intestinal perforation. Physical examination might detect an abdominal mass, and rectal examination might reveal the cystic lesions. A plain abdominal film can show radiolucent clusters or streaks along the bowel wall with pneumoperitoneum, if a cyst has ruptured. A markedly redundant sigmoid colon as well as the outline of the cysts or linear streaks may be seen on barium enema (Fig. 124-11).113 Endoscopic examination with biopsy is necessary for definitive diagnosis, to exclude carcinoma and to differentiate pneumatosis from familial adenomatosis polyposis120 and from the thumbprinting of

Figure 124-12.  Resection specimen of pneumatosis coli. Numerous gasfilled cysts are seen in this surface view of the mucosa (arrows). (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine; 1996.)

colon ischemia. The endoscopic appearance is of multiple cysts, which vary in size from a few millimeters to several centimeters (Fig. 124-12) and which, on puncture with a needle, rapidly deflate. Endoscopic ultrasonography also has been used to establish the diagnosis in pneumatosis.121

PATHOLOGY

The cysts of pneumatosis cystoids resemble soap bubbles. They usually are thin-walled and unilocular and can occur separately or in clusters. They do not communicate with the intestinal lumen or with each other and have a spongy

2249

2250

Section X  Small and Large Intestine of hydrogen within the cysts and a corresponding reduction in the size of the cysts. Because cysts can recur after oxygen therapy,122 a minimum of 48 hours of oxygen therapy is recommended to maximize the success rate. Metronidazole also has been used to treat pneumatosis coli, an observation that suggests that anaerobic bacteria play a role in the genesis of the disorder. Because cysts have been reported to recur after short courses of metronidazole,119 treatment should continue until complete resolution of the cysts is documented. In general, colonic resection is reserved for patients with complications such as intestinal obstruction and massive bleeding.

MALAKOPLAKIA

Figure 124-13.  Histopathology of a large gas-filled cyst in the colonic submucosa. These cysts are usually lined by histiocytes and multinucleated giant cells. Similar cysts also may be present in the subserosa. (Hematoxylin and eosin.) (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine; 1996.)

consistency that pops like a balloon when compressed. On cross section they appear shiny and honeycombed, and they range in size from a few millimeters to several centimeters. Microscopically, the cysts have an endothelial lining that tends to gather and coalesce, forming multinucleate giant cells that increase in number as the cysts collapse, undergo fibrosis (Fig. 124-13), and eventually are sloughed, leaving the cysts without a lining. Progressive fibrosis leads to a decrease in the size of the cysts and ultimately to their obliteration. The connective tissue surrounding the cysts can show a granulomatous inflammatory reaction made up of eosinophils, lymphocytes, macrophages, and plasma cells. Subserosal cysts are surrounded by fibrous connective tissue and can produce adhesion of adjacent bowel loops. The mucosa in pneumatosis may be normal, or it may be thinned and with or without ulcerations and inflammation where it is stretched over a cyst. Mucosal changes vary from mild focal abnormalities to extensive changes including granulomas, abnormal crypts with branching, shortening, cryptitis and abscesses, dilatation, and rupture.

TREATMENT

Because the natural history of pneumatosis is one of spontaneous regression in up to 50% of cases and because cysts can reappear after surgery, specific treatment is not recommended in asymptomatic patients. Symptomatic patients may be treated successfully by breathing high-flow oxygen for several days122 or by using hyperbaric oxygen, especially in resistant cases121; high oxygen levels lead to replacement

Malakoplakia is a rare chronic granulomatous disease first named by von Hansemann in 1903123 after being reported by Michaelis and Gutmann in 1902.124 The term malakopla­ kia is derived from the Greek malakos, “soft,” and plakos, “plaque” and reflects its usual appearance as a friable yellow mucosal lesion. Microscopically, coliform bacteria are located in the cytoplasm of macrophages (von Hansemann cells), and laminated intracytoplasmic inclusion bodies (Michaelis-Gutmann bodies) are considered the diagnostic features of this disorder.125 Malakoplakia can affect many organs, including lung, brain, adrenal glands, pancreas, bone, and the genitourinary tract. The most common site of gastrointestinal involvement is the colon, with the rectum, sigmoid, and right colon, involved in descending order of frequency.126

ETIOLOGY

The pathogenesis of malakoplakia is unknown. Proposed etiologies are infection, immunosuppression, systemic illness, neoplasia, and a genetic disorder. Evidence for an infectious etiology is based on the finding that some patients with malakoplakia have associated chronic infections. This was first described in the urologic form of malakoplakia in which more than 75% of patients were infected with Esch­ erichia coli. This finding led to the belief that E. coli might be a primary cause of malakoplakia; however, other organisms also have been isolated, including Klebsiella, Proteus, Mycobacterium, Staphylococcus, and fungi,127 suggesting that one infection is not the primary cause of the disease. Other evidence points to a defect in macrophage killing as the cause of malakoplakia. Nondigested microorganisms are found within the lysosomes of macrophages in affected persons. Macrophages from these patients show a decrease in cyclic guanosine monophosphate, resulting in impaired bactericidal activity.128 Peripheral blood monocytes also are found to have decreased bactericidal activity in malakoplakia. The defect in macrophage dysfunction may be reversed with the addition of a cholinergic agonist, both in vitro and in vivo.129 Malakoplakia has been reported in patients receiving chemotherapy and immunosuppressive therapy for organ transplantation130; reversal of macrophage abnormalities and resolution of clinical symptoms has been documented after discontinuation of glucocorticoids and azathioprine.131 Malakoplakia has been reported in various immune deficiency states such as primary hypogammaglobulinemia and AIDS,132,133 and it also has been associated with chronic systemic diseases such as systemic lupus erythematosus, ulcerative colitis, and sarcoidosis.126 There have been a substantial number of cases associated with colorectal cancer, to perhaps suggest a neoplastic etiol-

Chapter 124  Other Diseases of the Colon and Rectum

Figure 124-14.  Histopathology of malakoplakia. Sheets of large pale macrophages characterize the histologic changes in malakoplakia. One of the macrophages shows the characteristic ring-like Michaelis-Gutmann body (arrow) consisting of a central core of partially digested bacteria coated with iron and calcium phosphate. (Hematoxylin and eosin.) (Courtesy of Lawrence J. Brandt, MD, Bronx, NY.)

ogy for at least one form of malakoplakia.134,135 A possible genetic etiology was suggested by one report of colonic malakoplakia that clustered in a family.136

CLINICAL FEATURES AND DIAGNOSIS

Patients with malakoplakia usually present with abdominal pain, diarrhea, hematochezia, and fever.126 Physical findings may include a rectal mass on digital examination, abdominal mass, and weight loss. Diagnosis is by colonoscopy (and biopsy), which generally reveals one of the following three patterns of disease: Isolated rectosigmoid involvement, in which lesions appear as yellowish plaques that may be sessile, poly­ poid, and ulcerated. The colonic lumen may be strictured, and fistulas can form, suggesting a diagnosis of cancer or Crohn’s disease Diffuse colonic involvement, which is characteristic of malakoplakia in immunosuppressed patients Focal lesions, which may be associated with an adenomatous polyp or cancer Biopsy is essential to confirm the diagnosis and to exclude an underlying colonic malignancy (Fig. 124-14). Histology reveals the characteristic macrophages with voluminous cytoplasm (von Hansemann cells) containing the classic Michaelis-Gutmann bodies (intracytoplasmic concentric laminated inclusion bodies). The histiocytes (von Hanse­ mann cells) must be distinguished from those found in fungal disease, leprosy, Whipple’s disease, reticulum cell sarcoma, and macrophages harboring Mycobacterium avium complex.

TREATMENT

Patients with newly diagnosed malakoplakia should undergo a thorough medical evaluation to determine if they are taking immunosuppressive medications or have coexisting medical illnesses. Tests of immune function and screening for associated bladder malakoplakia and colorectal cancer are prudent. Patients receiving immunosup­ pressive medications might improve after these medications are discontinued. Antibiotics such as trimethoprimsulfamethoxazole and ciprofloxacin have been successful

Figure 124-15.  Resection specimen of colitis cystica profunda. Several submucosal cysts are filled with mucinous material (arrows). (Courtesy of Feldman M, Boland CR, editors. Slide Atlas of Gastroenterology and Hepatology. Philadelphia: Current Medicine; 1996.)

in treating malakoplakia.137,138 Both antibiotics appear to kill the bacteria associated with malakoplakia and can penetrate the defective host macrophages. Cholinergic agents also may be useful in treating children with malakoplakia.129 Surgical resection of the involved colon is recommended for cases associated with carcinoma or severe bleeding.

COLITIS CYSTICA PROFUNDA AND SUPERFICIALIS Colitis cystica profunda is a rare disease characterized by mucin-filled cysts located in the submucosa of the large intestine (Fig. 124-15). Colitis cystica profunda is to be distinguished from colitis cystica superficialis in which there are numerous small cysts, located superficially in the mucosa of the colon and not penetrating beyond the muscularis mucosa. Colitis cystica superficialis occurs mainly in patients with pellagra or advanced celiac disease, usually causes no symptoms, and resolves following nutritional repletion. Colitis cystica profunda was first described in 1766 by Stark, who reported two cases associated with dysentery.139 There are three patterns of disease: localized with a polypoid lesion, diffuse with multiple polypoid lesions, and diffuse with a confluent sheet of cysts.

ETIOLOGY

The etiology of colitis cystica profunda is unknown, but several theories have been proposed. A possible congenital etiology is supported by several findings. In embryologic examinations, submucosal cysts have been found in multiple gastrointestinal locations. The occurrence of colitis cystica profunda in children and its association with other congenital conditions such as Peutz-Jeghers syndrome140 also support a congenital origin for this disease; the absence of submucosal cysts in large autopsy series of infants and children reduces the plausibility of this etiology. Colitis cystica profunda has been associated with acquired diseases that predispose to mucosal ulceration and inflammation, including ulcerative colitis,141 Crohn’s disease,142 and infectious colitis.143 Submucosal cysts also have been reported in areas exposed to local trauma, such as an intestinal anastomosis or colostomy.144 Proctitis cystica profunda developed in rats treated with irradiation145 and at small bowel stomas also created in rats.146

2251

2252

Section X  Small and Large Intestine Colitis cystica profunda has been found in association with adenocarcinoma of the colon, suggesting a neoplastic etiology. Several cases of adenocarcinoma of the stomach associated with gastritis cystica profunda have been reported.147 In some reports, there is strong evidence of a causal link between cancer and colitis cystica profunda, because the submucosal cysts are often found adjacent to the adenocarcinoma, whereas adjacent benign mucosa is devoid of submucosal cysts. The localized form of colitis cystica profunda is associated with rectal prolapse and solitary rectal ulcer syndrome.139 Mucosal prolapse has been found in more than 50% of patients with the localized form of the disease. Trauma or ischemia caused by chronic traction on the mucosa and intramural vessels might play a role in the development of the submucosal cysts. Microscopic features of the localized form of the disease often include fibrosis of the lamina propria and hypertrophic muscle fibers, changes that are characteristic of solitary rectal ulcer syndrome (see Chapter 115).148

Figure 124-16.  Histopathology of colitis cystica profunda showing the mucus-filled cysts and misplaced epithelium (arrow). (Hematoxylin and eosin, ×5.) (From Mitros FA, editor. Atlas of Gastrointestinal Pathology. New York: Gower Medical Publishers; 1988.)

CLINICAL FEATURES AND DIAGNOSIS

Colitis cystica profunda affects men and women equally. The most common symptoms are rectal bleeding, mucus discharge, and diarrhea139; less common are tenesmus, abdominal pain, and rectal pain. Rarely, the patient presents with intestinal obstruction caused by the cysts.149 At endoscopy, most lesions are located on the anterior rectal wall 6 to 7 cm from the anal verge. The lesions appear as polyps with overlying mucosa that may be normal, inflamed, or ulcerated. Endoscopy might disclose an associated rectal prolapse in some cases. The endoscopic appearance of the lesions may be indistinguishable from a variety of lesions, including adenocarcinoma, adenomatous polyps, submucosal lipoma, neurofibroma, inflammatory pseudopolyps, pneumatosis coli, and endometriosis.139 Barium enema can reveal radiolucent filling defects. Transrectal ultrasound may be useful in differentiating this disease from cancer and reveals hypoechoic cysts that may be surrounded by intact submucosa, unlike invasive cancer.150 A characteristic finding consisting of noninfiltrating submucosal cysts also has been described on magnetic resonance imaging (MRI).151 Biopsy is necessary to differentiate this lesion from a variety of inflammatory, neoplastic, and infectious conditions. On biopsy, the submucosa is seen to be thickened by the presence of the mucus-filled cysts (Fig. 124-16). The cysts usually communicate with the lumen through small openings in the mucosa. Although usually confined to the submucosa, cysts involving the muscularis propria and serosa have been reported. The surrounding connective tissue often shows chronic inflammation, and there may be extensive replacement of the lamina propria by fibroblasts.148 As with solitary rectal ulcer syndrome, misplaced glands can cause confusion with invasive adenocarcinoma (see Fig. 124-16).

TREATMENT

A high-fiber diet and bowel retraining to avoid straining have led to regression of this disease in a few cases.150 If fiber is not effective, polyethylene glycol solutions may be tried.152 Glucocorticoid enemas also have been used with some success.139 Most patients have been treated with surgery. In patients with associated rectal prolapse, repair of the prolapse alone may treat the colitis cystica profunda successfully, whereas for disease localized to the rectum and in the absence of procidentia, local excision through a transanal approach is efficacious.152 When the disease is

localized to the rectum but is circumferential, total excision may be accomplished by mucosal sleeve resection and coloanal pull-through.152 More-diffuse lesions have been removed by segmental resection. Segmental resection also may be necessary for large obstructing lesions and for lesions that cause hypokalemia, hypoalbuminemia, or severe anemia from chronic blood loss. A diverting colostomy can lead to regression of this disease and may be the best option for a patient with significant comorbidities.

NEUTROPENIC ENTEROCOLITIS (TYPHLITIS) Neutropenic enterocolitis (typhlitis) is a potentially lifethreatening condition described by Wagner in 1970 in children undergoing chemotherapy for leukemia.153 Neutropenic enterocolitis subsequently has been described after organ transplantation, with AIDS, in patients with leukemia treated with cytosine arabinoside, and in patients with solid tumors treated with combination chemotherapy154-158; the frequency in persons at risk varies from 1% to 46%.159 The disease commonly affects the ileum and cecum and can result in intestinal perforation.

ETIOLOGY

The cause of neutropenic enterocolitis may be multifactorial. The initial injury is an ulceration of the bowel mucosa with no associated inflammatory response. Mucosal injury can occur from leukemic infiltration, stasis of bowel contents, or mucosal ischemia from splanchnic vasoconstriction resulting from sepsis.155,157,158 Certain drugs also can contribute to mucosal damage. Cytosine arabinoside can cause necrosis and delayed regeneration of intestinal glandular epithelium.160 Vinca alkaloids used to treat leukemia also can contribute to cecal distention by damaging the myenteric plexus of the intestine. With mucosal injury in the setting of impaired host defenses, infectious colitis subsequently occurs. The infection is often polymicrobial; causative bacteria include Escherichia coli and Staphy­ lococcus, Streptococcus, Enterococcus, and Klebsiella species161-162; fungal organisms such as Aspergillus and Candida also have been isolated.161 In addition to transmural infection of the intestine, the cecum can become gangre-

Chapter 124  Other Diseases of the Colon and Rectum nous and perforate as a result of increased distention and ischemia. The process can involve the ileum alone or both the ileum and cecum.161

CLINICAL FEATURES AND DIAGNOSIS

The most common presentation is with fever, diarrhea, nausea, vomiting, and abdominal pain in a patient who is receiving antineoplastic drugs.162 Abdominal tenderness typically is localized to the right lower quadrant of the abdomen, but it may be absent or masked by drugs such as prednisone; localized tenderness can progress rapidly to diffuse signs of peritonitis as a result of intestinal perforation. Shock can occur as a result of bacteremia or intestinal perforation. On occasion, the sigmoid colon may be affected, further complicating the diagnosis.163 Neutropenia is common but was absent in 12% of children in a review of typhlitis in childhood cancer162; blood cultures are positive in up to 50% of cases.155,161 Differential diagnosis includes appendicitis, pseudomembranous colitis, ischemic colitis, volvulus, diverticulitis, and drug-induced diarrhea. Diagnostic evaluation should include a radiologic evaluation to exclude other diseases, confirm the diagnosis, and determine the severity of illness. Abdominal films can demonstrate dilated loops of small bowel with decreased air in the right lower quadrant and free intraperitoneal air if intestinal perforation has occurred (Fig. 124-17).155 CT scans are most sensitive for establishing the diagnosis and help to exclude other conditions such as appendicitis and diverticulitis. The CT scan might reveal a thickened bowel wall, pneumatosis intestinalis, ascites, and free air.164 One study suggested that ultrasonography is superior to CT scan in predicting the outcome of typhlitis.162 Barium enema should be avoided because of the potential risk of perforation. Stool assay for C. difficile toxin should be performed routinely to exclude pseudomembranous C. difficile colitis.

TREATMENT

Management of neutropenic enterocolitis has varied and is controversial because there are no prospective or highquality retrospective studies concerning medical or surgical therapies.165 Approaches have included supportive mea-

sures alone, aggressive initial surgical resection, and combined medical and surgical treatment; successes and failures have been documented with all these approaches. In two studies, all patients treated medically recovered, whereas in another similar series, all patients managed medically died.166 Clearly, successful management of patients with neutropenic enterocolitis needs to be individualized to optimize outcome. In general, medical management includes broad-spectrum antibiotics, nasogastric suction, and bowel rest. Fluid resuscitation with isotonic solutions is critical for main­ taining renal perfusion in the face of decreased systemic vascular resistance from sepsis and intra-abdominal fluid sequestration. Close observation and serial abdominal and radiologic examinations are necessary to monitor the response to medical treatment. Antibiotics should have activity against enteric Gram-negative organisms, Grampositive organisms, and anaerobes. Causative microor­ ganisms include Enterococcus species, S. aureus, E. coli, group A Streptococcus, and Klebsiella.161,162 For patients who do not respond to antibacterial agents, amphotericin should be considered, because fungemia is common.161 Blood transfusions may be necessary because the diarrhea often is bloody. Granulocyte-macrophage colony-stimulating factor to correct the neutropenia may be a useful adjunct to medical therapy.167 Early surgical intervention has been recommended in persons with a rapidly deteriorating course despite maximal medical therapy. Two series have shown a decreased mortality rate in patients with severe disease who are treated surgically compared with those treated medically.166 For patients with complications such as frank gangrene, intestinal perforation, and shock despite vasopressor support, surgical intervention is mandatory. Controversy surrounds the choice of operation. Gangrenous or perforated bowel should be resected. When the bowel is edematous with no vascular compromise and no signs of perforation, successful management has included no resection,168 intestinal diversion with no resection,166 and resection of the involved bowel. If resection is performed, construction of an ileostomy and mucous fistula may be the safest option, because intestinal anastomoses may be prone to breakdown in patients with neutropenia.169 Because recurrences of neutropenic enterocolitis are common when chemotherapy is restarted, right hemicolectomy is recommended before chemotherapy is resumed.157

ENDOMETRIOSIS

Figure 124-17.  Left lateral decubitus plain film in a patient with neutropenic enterocolitis. The colon is dilated with a prominent intraluminal gasfluid interface and free peritoneal air (arrowheads) resulting from cecal perforation. (From Hunger TB, Bjelland UJC. Gastrointestinal complications of leukemia and its treatment. AJR Am J Roentgenol 1984; 142:513.)

Endometriosis, defined as the presence of endometrial tissue outside the uterine cavity and musculature, was first described by von Rokitansky in 1860. Most often, ectopic endometrial tissue lies in the vicinity of the uterus. Endometriosis occurs in up to 15% of menstruating women and up to 30% of infertile women.170 The initial description of nonpigmented endometriosis in 1986 resulted in increased recognition and a much higher prevalence of this disorder than appreciated previously.171 In contrast to endometrial involvement of the female reproductive organs, gastrointestinal involvement by endometriosis is less common, usually asymptomatic, and clinically less important.172 The intestinal organs most commonly involved are the rectosigmoid (96%), appendix (10%), and ileum (5%), with other organs involved uncommonly.173 Intestinal endometriosis can mimic a wide variety of inflammatory, infectious, and neoplastic digestive disorders.174

2253

2254

Section X  Small and Large Intestine ETIOLOGY AND PATHOGENESIS

Several hypotheses have been advanced to explain the ectopic location of endometrial tissue.175-177 The most commonly accepted explanation is that of retrograde passage of endometrial tissue, which then implants and grows on pelvic organs and the peritoneum. From these sites, more distant implants arise via hematogenous or lymphatic dissemination; further dissemination can occur during surgical interventions. A less-accepted hypothesis, with fewer supporting data, is that of endometrial metaplasia, in which multipotential peritoneal mesothelial cells are induced by unknown factors to undergo metaplastic transformation to endometrial tissue. Once implanted, endometrial tissue appears to be regulated by hormonal influences, so that estrogen promotes and progesterone inhibits its growth. These repetitive cycles of growth and sloughing of tissue can lead to serosal irritation and progressive endometrial invasion of intestinal muscle with resultant fibrosis and muscle hypertrophy. Thus, pain can arise from nerve impingement or serosal inflammation, whereas obstruction can result from narrowing of the lumen or kinking of the intestine.

air-contrast barium enema, which demonstrates submucosal polypoid masses or areas of noncircumferential narrowing of the lumen (Fig. 124-18). Diagnostic yield and accuracy may be enhanced by performing these tests just before the onset of menses. CT scans (Fig. 124-19), ultrasonography, and MRI all have been reported to assist with the diagnosis or assessment of the extent of endometrial involvement. High-resolution transvaginal and transrectal ultrasonography also may be useful in detecting small endometrial implants, particularly in the retroperitoneal pelvis.180 Definitive diagnosis of endometriosis often is made by laparoscopy or laparotomy with biopsy and is especially

CLINICAL FEATURES

Endometriosis is found almost exclusively in women of childbearing age, with clinical onset usually between the ages of 20 and 45 years.173 Women who experience symptoms or who undergo surgery for endometriosis beyond menopause presumably have chronic fibrosis or exacerbations induced by exogenous estrogen. Although most women with endometrial implants on intestinal structures have no symptoms, those with serosal implants may complain of localized tenderness, low backache, or abdominal pain. Penetration of endometrial tissue into the bowel wall can produce constipation, diarrhea, and partial obstruction, resulting in intermittent abdominal pain. Contrary to popular thinking, symptoms are not always cyclical and might not fluctuate with hormonal levels; nor are gastrointestinal symptoms necessarily asso­ ciated with gynecologic symptoms. Rarely, hematochezia occurs when endometrial implants penetrate to the mucosa or when severe colonic fibrosis results in ischemia.172 Less common presentations occur with involvement of the more proximal colon or small intestine and include acute appendicitis caused by an obstructing endometrioma, small bowel intussusception, and volvulus.174,178

Figure 124-18.  Film from a single-contrast barium enema demonstrating a large, nodular, partially obstructing endometrioma in the rectosigmoid colon (arrow). (Courtesy of Mark Peterson, MD, Pittsburgh, Pa.)

DIAGNOSIS

The clinical diagnosis of intestinal endometriosis may be difficult because symptoms often are nonspecific and there might not be a close relationship between the symptoms and the menstrual cycle. Endometriosis should always be considered in women with recurrent abdominal pain and bowel symptoms, especially if they are in their reproductive years and have gynecologic complaints. Diagnosis is especially difficult because IBS is so common in women. An important component of the evaluation is a careful pelvic examination that includes combined rectovaginal palpation. Finding tender nodules or irregularities in the cul-de-sac is highly suggestive of endometriosis. Because findings can vary considerably during the menstrual cycle, the pelvic examination should be performed immediately before and again after menstruation if no abnormalities were found initially. It is rare to see endometrial implants on the colonic mucosa except when there is hematochezia. Thus, colonoscopy is often normal except for areas of extrinsic compression or strictures with intact mucosa.179 More helpful is an

Figure 124-19.  Computed tomography scan in the same patient as in Figure 124-18 showing the endometrial mass in the cul-de-sac (arrow) extending into the colon. (Courtesy of Mark Peterson, MD, Pittsburgh, Pa.)

Chapter 124  Other Diseases of the Colon and Rectum

Figure 124-20.  Histopathology of endometriosis. Endometrial stoma is seen around the glandular structures in the muscularis propria of the colon in a woman presenting with obstructive symptoms. The diagnosis of endometriosis was proved at surgery. (Hematoxylin and eosin, ×10.) (From Mitros FA, editor. Atlas of Gastrointestinal Pathology. New York: Gower Medical Publishers; 1988.)

follicle-stimulating hormone and luteinizing hormone. Although both are effective in decreasing pelvic pain associated with endometriosis and appear to decrease the size of endometrial implants, there are no studies of these agents in endometriosis-associated intestinal disease, and there is some concern that treatment can result in increased fibrosis and inadequate resolution of symptoms.185 Ablation of endometrial implants on surfaces that can be visualized laparoscopically can be accomplished using carbon dioxide laser.186 For endometriosis that causes partial obstruction of the colon or small intestine, segmental resection of the involved area provides the best results and also serves to exclude an underlying carcinoma.187-189 Resection can be performed by laparoscopy or by open surgery, according to available expertise.190 If the patient is postmenopausal or if future pregnancies are not wanted, hysterectomy and bilateral salpingo-oophorectomy can be done at the time of resective surgery to minimize the risk of symptomatic disease in the future. Similar surgery also can be performed in premenopausal women who, despite medical therapy, have intractable symptoms.

KEY REFERENCES useful in patients with intestinal implants without pelvic involvement. The classic peritoneal implant appears as a bluish-black “powder-burn” lesion with variable degrees of pigmentation and surrounding fibrosis, the dark coloration resulting from hemosiderin deposition. Most peritoneal implants, however, appear as subtle, nonpigmented lesions. The appreciation that endometrial tissue may be non­ pigmented has increased the yield of these procedures considerably.171 The differential diagnosis of intestinal endometriosis includes inflammatory disorders such as Crohn’s disease and ulcerative colitis with stricture, diverticulitis, infectious diseases such as ileocolonic tuberculosis and schistosomiasis, benign and malignant neoplastic disorders, and colon ischemia.174 It is important to emphasize that no radiologic or imaging finding is pathognomonic of endometriosis; mucosal abnormalities that permit positive biopsies are rare; and tissue for a definitive diagnosis usually is obtained only at laparotomy (Fig. 124-20).

TREATMENT

In general, when a diagnosis of serosal intestinal endo­ metriosis is made, hormonal therapy is often the first therapeutic option, similar to the standard approach to pelvic endometriosis.181,182 Low-dose estrogen-progestin compounds cause a pseudopregnancy state that results in deci­dualization of endometrial tissue and often relieves dysmenorrhea. Their use in more-severe disease is questionable, however, and they generally are not recommended for symptomatic intestinal disease. The most effective agents currently available are the synthetic androgen danazol and the gonadotropinreleasing hormone agonists.183,184 Both act to decrease ovarian steroid synthesis by inhibiting pituitary release of

Abdou N, NaPombejara C, Sagawa A, et al. Malacoplakia: Evidence for monocyte lysosomal abnormality correctable by cholinergic agonist in vitro and in vivo. N Engl J Med 1977; 297:1413-19. (Ref 129.) Beck DE. Surgical treatment for colitis cystica profunda and solitary rectal syndrome. Curr Treat Options Gastroenterol 2002; 5:231-37. (Ref 152.) Cappell MS. Colonic toxicity of administered drugs and chemicals. Am J Gastroenterol 2004; 99:1175-90. (Ref 107.) Chande N, McDonald JWD, MacDonald JK. Interventions for treating lymphocytic colitis. Cochrane Database Sys Rev 2008; (2):CD006096. (Ref 36.) Chande N, McDonald JWD, MacDonald JK. Interventions for treating collagenous colitis. Cochrane Database Sys Rev 2008; (2):CD003575. (Ref 37.) Davila ML. Neutropenic enterocolitis: Current issues in diagnosis and management. Curr Infect Dis Rep 2007; 9:116-20. (Ref 165.) deOliviera-Neto JP, de Aguilar-Nascomento JE. Intraluminal irrigation with fibers improves mucosal inflammation and atrophy in diversion colitis. Nutrition 2004; 20:197-99. (Ref 58.) Eggenberger JC, Farid A. Diversion colitis. Curr Treat Options Gastroenterol 2001; 4:255-9. (Ref 56.) Fernandez-Banares F, Esteve M, Espinos JC, et al. Drug consumption and the risk of microscopic colitis. Am J Gastroenterol 2007; 102:32430. (Ref 10.) Florin THJ. Alkyl halides, super hydrogen production, and the pathogenesis of pneumatosis cystoides coli. Gut 1997; 41:778-84. (Ref 117.) Gadenstatter M, Wetscher G, Crookes PF, et al. Dieulafoy’s disease of the large and small bowel. J Clin Gastroenterol 2002; 27:169-72. (Ref 72.) Gopal DV, Katon RM. Endoscopic balloon dilation of multiple NSAID-induced colonic strictures: case report and review of literature on NSAID-related colopathy. Gastrointest Endosc 1999; 50:12023. (Ref 72.) McCarville MB, Adelman CS, Chenghong L, et al. Typhlitis in childhood cancer. Cancer 2005; 104:380-87. (Ref 162.) Muller-Lissner SA, Kamm MA, Scarpignato C, Wald A. Myths and misconceptions about chronic constipation. Am J Gastroenterol 2005; 100:232-42. (Ref 85.) Olive DL, Pritts EA. Treatment of endometriosis. N Engl J Med 2001; 345:266-75. (Ref 181.) Full references for this chapter can be found on www.expertconsult.com.

2255

CHAPTE R

125 Diseases of the Anorectum Peter W. Marcello

CHAPTER OUTLINE Anatomy  2257 Examination of the Anus and Rectum  2258 Inspection  2258 Palpation  2259 Endoscopy  2259 Hemorrhoids  2260 Internal Hemorrhoids  2260 External Hemorrhoids and Anal Tags  2262 Anal Fissure  2263 Etiology  2264 Symptoms, Signs, and Diagnosis  2264 Treatment  2265 Abscesses and Fistulas  2266 Abscesses  2266 Fistula-in-Ano  2267 Special Fistulas  2268

ANATOMY The functional anal canal is 3 to 4 centimeters long, beginning at the top of the anorectal ring (at the puborectalis sling) and extending down to the anal verge (anal orifice).1 The upper anal canal is lined mostly with columnar epithelium, a continuation of the same type of tissue that lines the rectum. Some squamous epithelium starts to be intermixed with the columnar epithelium in the rectum at about one centimeter above the dentate line. This change to squamous epithelium is gradual, and the area 1 to 1.5 cm proximal to the dentate line is termed the transitional zone.2 The dentate line is located in the mid-anal canal and is seen as a wavy line. Distal to the dentate line, the tissue is squamous epithelium, but it is unlike true skin because it has no hair, sebaceous glands, or sweat glands: it is commonly referred to as anoderm. The anoderm is thin, pale, and delicate, with the appearance of shiny, stretched skin. At the anal verge, the epithelium becomes thicker, and hair follicles begin to be seen (Fig. 125-1).3 Embryologically, the dentate line represents the junction between endoderm and ectoderm. Proximal to the dentate line, there is sympathetic and parasympathetic innervation; distally, the nerve supply is somatic.3 Therefore, above the dentate line, pain sensation is negligible; a biopsy can be done painlessly above the dentate line and without the need for local analgesia. Below the dentate line, however, the anoderm is highly sensitive, an important point to note when examining the anal canal or applying hemorrhoidal bands. The arterial supply of the anal area is from the superior, middle, and inferior hemorrhoidal arteries, which are continuations or branches of the inferior mesenteric, hypogastric, and internal pudendal arteries, respectively. Blood flow is not uniform to the entire circumference of the anus and

Anal Malignancies  2269 Anal Margin Cancers  2269 Anal Canal Cancers  2270 Melanoma  2270 Anal Intraepithelial Neoplasia  2270 Paget’s Disease  2270 Anal Warts  2270 Pruritus Ani  2271 Anal Stenosis  2272 Unexplained Anal Pain  2272 Coccygodynia  2272 Levator Ani Syndrome and Proctalgia Fugax  2272 Hidradenitis Suppurativa  2273 Pilonidal Disease  2273 Rectal Foreign Body  2274

is relatively less posteriorly. This differential flow is an important factor in the role postulated for ischemia in determining the chronicity of anal fissure.4 The venous drainage from the anal canal is by both the systemic and portal systems. The internal hemorrhoidal plexus drains into the superior rectal veins, which drain into the inferior mesenteric vein and then into the portal vein. The distal part of the anal canal drains via the external hemorrhoidal plexus through the middle rectal and pudendal veins into the internal iliac vein (i.e., the systemic circulation).2 The lymphatic drainage of the anus changes at the dentate line. Proximally, the lymphatic vessels accompany the blood vessels and drain into the inferior mesenteric and periaortic nodes.1-3 Distal to the dentate line, the lymphatics drain into the inguinal nodes. Therefore, inguinal adenopathy can be seen with inflammatory and malignant disease of the lower anal canal. Immediately proximal to the dentate line, the mucosa appears to have 6 to 14 pleats, called the columns of Morgagni. This configuration represents the funneling of the rectum as it narrows into the anal canal. Located at the base of the columns of Morgagni are anal crypts that lead to small rudimentary anal glands.3 These glands can extend through the internal anal sphincter; if their ducts are blocked, an anal abscess or fistula can develop (see Fig. 125-1). The muscles surrounding the anal canal are important for maintaining fecal continence. The internal anal sphincter is the thickened continuation of the circular smooth muscle of the rectum. This muscle is involuntary and because it is composed of smooth muscle, it has a black hypoechoic appearance on anal ultrasonography. The internal sphincter ends above the external sphincter.3 The internal sphincter is important for passive continence, but partial division of the internal sphincter is possible without causing significant fecal incontinence. Division in the anterior or posterior

2257

2258

Section X  Small and Large Intestine Puborectalis muscle Column of Morgagni

Internal hemorrhoidal plexus Internal anal sphincter

Dentate line

External anal sphincter

Anal crypt

Anal gland

External hemorrhoidal plexus

Anoderm

Figure 125-1.  Schematic depiction of the anatomy of the anal region. The vertical line with arrowheads denotes the transition zone that extends 1 to 1.5 cm proximal to the dentate line.

position, however, can lead to stool leakage by creating an oval-shaped configuration in the distal anal canal, termed a keyhole deformity because the shape was reminiscent of the opening for the key in old-fashioned door locks. The external anal sphincter is under voluntary control and is composed of a sheet of skeletal muscle arranged as a tube surrounding the internal sphincter. It appears as a broad cylinder with mixed echogenicity on anal ultrasound. Proximally, the external sphincter is fused with the puborectalis muscle, and distally, it ends slightly past the internal sphincter.1,3 Therefore, a groove can be palpated between the internal sphincter and external sphincter on digital examination, which is referred to as the intersphincteric groove. The puborectalis is U-shaped and forms a sling that passes behind the lower rectum and attaches to the pubis. Therefore, it is prominent laterally and posteriorly but absent anteriorly. The anorectal ring is palpable at the top of the anal canal, where the canal meets the rectum. It is composed of the puborectalis and the upper external and internal sphincters. The nerve supply to the external sphincter and puborectalis muscle is from the inferior rectal branch of the internal pudendal nerve (S2, S3, S4) and also from fibers of the fourth sacral nerve.1,3

EXAMINATION OF THE ANUS AND RECTUM All routine adult physical examinations should include a digital examination of the anus. When patients present with problems focused on the anorectal region or colon, a more comprehensive examination is indicated. Examination begins with a thorough history, allowing the patient to describe his or her symptoms and concerns, and the physician to develop rapport with the patient. This time spent is invaluable because it will help alleviate the patient’s potential apprehension and embarrassment during the examination. It is important to remember that many patients delay coming to a physician, even when they have sig­

nificant problems, solely out of fear and embarrassment. Therefore, when doing the examination, explanation of each step before it is done, along with a gentle touch, helps alleviate discomfort. It is equally important to avoid causing pain during the examination, particularly when looking for a fissure or anal abscess. When an examination is too painful, the patient will likely require an examination under anesthesia. Patients can be examined in the office in any of several positions, including the left lateral position, knee-chest, or prone jackknife position. In prone jackknife, the patient is on his or her knees with the arms folded and the shoulders and head on the examining table; some patients may be more comfortable resting on the left shoulder. The examiner may use a special hydraulic table that has a shelf upon which the patient may kneel and then drape his or her chest over the main table. The table then is raised and its head is tipped down, propelling the buttocks forward and elevating them. This position allows the buttocks to be splayed apart for a clear view of the anus.

INSPECTION

The examination begins with inspection of the skin. In some cases, looking at the patient’s underwear will give a clue to the character of anal drainage or the presence of stool incontinence. As the buttocks are gently retracted, scars, skin abnormalities, stool, discharge of blood or pus, anal tags, warts, hemorrhoids, or lesions adjacent to or prolapsing from the anal canal are noted. The anus is inspected for gaping or scars. The patient is asked to squeeze the anus to evaluate movement of the anal muscles. There should be concentric movement of the anus and perianal skin. This maneuver is important when assessing for fecal continence. Next, the patient should be asked to strain so that the anal area can be examined for abnormal descent below its resting level (perineal descent syndrome). Prolapse of the vagina or rectum, bulging hemorrhoids, or leakage of urine also should be noted during straining. In some instances, rectal

Chapter 125  Diseases of the Anorectum

Figure 125-2.  Inspection of an acute anal fissure with a cotton tip swab. Once the acute fissure is identified, no internal examination is needed until the fissure is healed.

prolapse may be seen only if the patient is asked to strain when sitting on the toilet. Traction applied laterally to each side of the anal orifice with a gauze pad allows eversion of the distal anus for further inspection. This technique is particularly helpful in viewing a fissure without causing undue pain (Fig. 125-2). Some examiners stroke the perianal skin with a cottontipped applicator to look for reflex contraction of the anal muscles (anal wink, anocutaneous reflex) or check perianal sensation with a pinprick; these maneuvers give a crude determination of sphincter innervation and are important in evaluating for sensory neuropathies. The nearby skin of the buttocks, perivaginal region, base of the scrotum, and up to the tip of the coccyx should be viewed. Adenopathy in the inguinal region may be seen when certain infectious or neoplastic lesions are found distal to the dentate line.

PALPATION

Using a gloved and lubricated index finger, the examiner palpates the anal canal and perianal skin. Slow insertion and gentle pressure are appropriate. The index finger is swept all around the anal canal. Anal tone is noted, as are any scars, masses, or tenderness. Internal hemorrhoids are not palpable unless they are thrombosed. If insertion of the index finger is too painful, an attempt to apply pressure with the insertion finger on the wall opposite the area of tenderness might allow insertion of the finger. If the examination is still too painful, use of sedation or anesthesia may be warranted. Approximately 80% of the resting anal canal pressure is contributed by the internal anal sphincter. The external sphincter is evaluated by having the patient voluntarily squeeze the anus around the examining finger. Abnormalities sometimes appreciated in the anal canal include fistulous tracts, which feel like a cord or induration; the internal opening of a fistula, which may be appreciated as a knob of tissue in an otherwise smooth area of mucosa; cancers, which may be firm and hard; and ulcers, which feel uneven and craterous. Palpation anteriorly in a woman

may reveal a rectocele or anterior defect in the sphincter complex. Palpation of the distal rectum allows the detection of mass lesions, including polyps and cancers. Attention should be directed to the exact location of the lesion (anterior, posterior, right, left, or in between), and its size, mobility, and character (soft, ulcerated, hard, or pedunculated). Lesions outside the rectal wall also may be appreciated. The cervix can be felt through the anterior rectal wall in women, and the prostate should be examined in men. The character of the prostate should be noted, along with any hard nodularity that could represent cancer; further studies are then ordered if needed. The mucosa should be assessed for its texture; in patients with proctitis, for example, the mucosa can feel rough and gritty; in patients with severe hypoalbuminemia, it can feel wet and slippery. The levator muscles should be palpated. In some patients with unexplained anal pain, these muscles may be in spasm and tender (the levator ani syndrome). Similarly, the coccyx should be palpated between the examining internal index finger and the index finger of the opposite hand pressed over the external skin at the level of the coccyx. This maneuver is done to look for pain with motion (coccygodynia), as might be present with a coccygeal fracture. The contents of the rectum should be assessed regarding the character and amount of stool. When the index finger is removed, any stool, blood, pus, or mucus on the glove should be noted.

ENDOSCOPY

The decision to perform endoscopy depends on the findings on history and physical examination. Endoscopy usually is necessary for the evaluation and exclusion of organic disease in patients with fecal incontinence, constipation, unexplained anal pain,5 anemia, diarrhea, and rectal bleeding.

Anoscopy

Anoscopy allows visualization of the anal canal, dentate line, internal hemorrhoids, and distal rectum. This is the best method of viewing the anal canal. The anoscope is a short metal or plastic tubular device, usually with a beveled end. An anoscope should never be inserted or turned without the obturator in place. Most adult anoscopes have a 2-cm diameter, but smaller ones are available. A fiberoptic light attachment provides optimal illumination; however, external lighting is used with some models. The lubricated anoscope is inserted slowly as the examiner applies gentle pressure on the end of the obturator until the instrument has been fully advanced. The obturator is then removed, and the entire anal canal region is examined. This includes the distal rectum, followed by the upper anal canal, down to the anoderm. To reinsert the scope and view another area, the obturator must be re-placed. Internal hemorrhoids can be seen bulging above the dentate line or prolapsing downward. Internal fistulous openings may be viewed, particularly along the dentate line. When the external skin is compressed, pus may be seen to bubble from the internal opening of a fistula. Rarely, the anoscope is needed to remove a low rectal polyp that cannot be removed by flexible endoscopy because of the low position of the polyp in the rectum or a difficult angulation on retroflexion of the endoscope.

Rigid Proctoscopy

Rigid proctosigmoidoscopy is performed mainly by colon and rectal surgeons today. The rigid proctoscope is 25 centimeters long and 11 to 20 millimeters in diameter. It requires fiberoptic light for visualization.

2259

2260

Section X  Small and Large Intestine In the modern era of video sigmoidoscopes, there are still some instances when a rigid proctoscope is advantageous. The proctoscope can measure the exact distance of a rectal tumor from the anal verge. It also can give the precise location of a lesion on the wall of the rectum. Such measurements might be inaccurate with flexible instruments and are important for planning operative strategies. The rigid proctoscope is sometimes quicker and easier to use than a flexible instrument when evaluating the rectum and doing a biopsy or aspirating fecal contents. The biopsy forceps used with flexible endoscopes also can be used through the rigid proctoscope, although typically a rigid, alligator-toothed forceps is preferred.

Flexible Sigmoidoscopy

The flexible sigmoidoscope is simply a shorter version of a colonoscope, measuring 60 cm in length. Most endoscopists use a gastroscope for sigmoidoscopy because it is thinner, better tolerated by the patient than a sigmoidoscope or colonoscope, and easier to maneuver. One to two enemas are given before the examination and sedation typically is not used, which is the reason patients who have undergone both colonoscopy and flexible sigmoidoscopy report that the latter was more difficult.6 The goal is to examine the left colon, which should be reached at least 80% of the time.7 Lesions can be biopsied, but the presence of polyps mandates full colonoscopy after a bowel preparation to exclude synchronous polyps or cancer. The use of electrocautery and argon plasma coagulation should be avoided during sigmoidoscopy, even if enemas have just been given and the preparation appears optimal, because intracolonic bowel explosions have occurred from ignition of bowel gas that has passed from the stool-containing proximal colon distally to the operative site.8 The exact role of flexible sigmoidoscopy is evolving. Flexible sigmoidoscopy can be used to enhance the diagnostic capability of barium enema, which at times fails to visualize the distal rectum optimally because of the obscuring effect of the balloon needed to distend the colon. Lesions of the rectum and sigmoid seen on radiologic studies also can be evaluated by flexible sigmoidoscopy. Flexible sigmoidoscopy permits serial examinations and treatments of diseases located in the rectosigmoid and left colon, such as proctosigmoiditis and radiation proctitis.

HEMORRHOIDS Hemorrhoids are perhaps the most misunderstood anorectal problem for patients and physicians alike. In clinical practice, patients use the term hemorrhoid to describe almost any anorectal problem, from pruritus ani to cancer.9 In fact, hemorrhoids are a normal part of human anatomy,10,11 in contrast to hemorrhoidal disease, which is manifested by prolapse, bleeding, and itching.10 Hemorrhoids are dilated vascular channels located in three fairly constant locations: left lateral, right posterior, and right anterior. Internal hemorrhoids originate above the dentate line and are covered with columnar or transitional mucosa. External hemorrhoids are located closer to the anal verge and are covered with squamous epithelium. Traditionally, internal hemorrhoids are classified into four grades: first-degree hemorrhoids, which bleed with defecation; second-degree hemorrhoids, which prolapse with defecation but return spontaneously to their normal position; third-degree hemorrhoids, which prolapse through the anal canal at any time, but especially with defecation, and can be replaced manu-

ally; and fourth-degree hemorrhoids, which are prolapsed permanently.11 Although the exact incidence of hemorrhoidal disease is unknown, it is thought to be present in 10% to 25% of the adult population.12

INTERNAL HEMORRHOIDS Symptoms and Signs

It is speculated that internal hemorrhoids become sympto­ matic when their supporting structures become disrupted and prolapse of the vascular cushions occurs.13 Hemorrhoids occur more commonly in people with constipation who have hard, infrequent stools.14 They also can occur in patients who have frequent loose stools or if prolonged periods of time are spent sitting on the toilet, leading to vascular congestion. Bleeding is typically painless and the patient describes bright red blood usually seen on the toilet tissue, dripping into the toilet bowl, or streaking the outside of a hard stool. If the bleeding is more substantial, the blood can accumulate in the rectum and be passed later as dark blood or clots.11 If the patient has chronic hemorrhoidal prolapse, blood or mucus might stain the patient’s underwear, and the mucus against the anal skin might lead to itching.10

Diagnosis

The diagnosis of internal hemorrhoids is best made with the beveled anoscope or by flexible sigmoidoscopy and retrograde view of the anorectal junction. The vascular cushions can be seen to bulge into the anal lumen, or the tissue might prolapse out through the anal canal. It is important to note that hemorrhoids are dynamic structures and may be symptomatic only intermittently. If the patient is seen outside of a flare, the hemorrhoids can appear entirely normal.

Treatment

Treatment is based on the grade of the hemorrhoids. Grade 1 and some early grade 2 internal hemorrhoids usually respond to manipulation of the diet, along with avoidance of medications that promote bleeding, such as nonsteroidal anti-inflammatory drugs (NSAIDs). A high-fiber diet, with 20 to 30 g of fiber daily introduced gradually into the diet, should be accompanied by six to eight glasses of nonalcoholic, noncaffeinated fluid daily. Patients are encouraged to read the package regarding the amount of fiber per serving; for instance, a bowl of raisin bran cereal can have eight grams of dietary fiber per serving whereas a similar portion of corn flakes has only one gram.15 Fiber supplementation with psyllium or hydrophilic colloid may be added to achieve the optimal amount of daily fiber, if the patient’s daily dietary fiber is insufficient. In my experience, patients who can achieve adequate dietary fiber intake without supplements have better long-term relief of hemorrhoidal disease with dietary change than those who need daily supplementation. Patients are urged to avoid straining during defecation and reading while on the toilet. Excessive scrubbing of the anus when showering or bathing or excessive wiping after a bowel movement is discouraged. Most over-the-counter agents are not efficacious, even though many patients report some relief of their symptoms with use of these products.14 Sometimes a stool softener such as docusate sodium or a lubricant such as mineral oil can be prescribed if the stool is hard and does not respond to increased intake of fiber and fluid; laxatives and enemas rarely are needed.15 Even patients who require more-aggressive treatment of their hemorrhoids should be advised to increase their dietary

Chapter 125  Diseases of the Anorectum fiber and fluids and to avoid straining during defecation to prevent recurrence after treatment. When manipulation of the diet does not work, moreaggressive treatment may be needed; these measures can apply to grades 1, 2, and 3 internal hemorrhoids. Unless the patient has fourth-degree internal hemorrhoids, aggressive nonsurgical treatment usually is tried; the majority of patients with fourth-degree hemorrhoids require surgical intervention. Most nonoperative treatments are designed to affix the vascular cushion to the underlying sphincter. Options to achieve such fixation include sclerotherapy, rubber band ligation, cryotherapy, and infrared photocoagulation. Sclerosing Agents Injection therapy for hemorrhoids has been practiced for more than 100 years. The goal is to inject an irritant into the submucosa above the internal hemorrhoid at the anorectal ring (the area that does not have somatic innervation) to create fibrosis, tack down the hemorrhoid and prevent hemorrhoidal prolapse.16 Usually less than one milliliter of sclerosant is needed to create a raised area. Many substances have been used, but sterile arachis oils containing 5% phenol are the most popular.16 This approach usually is advocated for first- and second-degree hemorrhoids. Sclerotherapy can produce a dull pain for up to two days after injection. A rare but severe complication is lifethreatening perineal sepsis, which can occur three to five days after injection and usually is manifested by any combination of perianal pain or swelling, watery anal discharge, fever, leukocytosis, and other signs of sepsis. Prompt sur­ gical intervention and intravenous antibiotics are mandatory.16,17 Approximately 75% of patients with second-degree hemorrhoids improve after injection therapy.18 In patients with active acquired immunodeficiency syndrome (AIDS), injection therapy may be favored over surgical treatments because of concerns about the patient’s poor overall general condition. There also may be problems with wound healing for these patients, but successful treatment of second-, third-, and fourth-degree hemorrhoids without complications has been reported in patients with AIDS,19 some of whom did require repeat treatment to manage persistent symptoms. Rubber Band Ligation Rubber band ligation (RBL) has become the most common office procedure for the treatment of second- and thirddegree hemorrhoids.20 Generally, this approach cannot be used with first-degree hemorrhoids, because there is insufficient tissue to pull into the bander; this treatment is not appropriate for fourth-degree hemorrhoids.18 Rubber bands are applied to the hemorrhoidal complex and rectal mucosa just proximal to the internal anal cushion. To avoid severe pain, bands are never placed below the dentate line, which is innervated by somatic fibers. The number of bands that can be safely placed in one setting remains controversial. Several studies have shown that triple RBL is safe and effective at one sitting,20,21 but many authorities believe that the severity of pain and risk of complications are less if one band is applied per visit. I typically place a single band at the first visit, and if that is well tolerated, I place two bands during the next and subsequent visit. Most grade 2 or 3 hemorrhoids can be managed successfully with two or three RBL procedures. Patients can experience discomfort after RBL; soaking in a sitz bath and taking acetaminophen usually constitute sufficient treatment. Immediate severe pain usually signals

that the band has been placed too close to the dentate line and that it must be removed. After RBL, patients are instructed to increase the fiber in their diet and modify their bowel habits, as discussed previously. When RBL can be performed, success is reported in 75% of patients with first-degree and second-degree hemorrhoids and in 65% of those with third-degree hemorrhoids. Repeated RBL is an option for patients who continue to have prolapsing tissue. Major complications from RBL include bleeding, sepsis, cellulitis, and death. Bleeding when the band and necrotic hemorrhoidal tissue comes off four to seven days after application may be severe and even life-threatening; severe bleeding occurs in about 1% of patients21 and usually can be tamponaded by placing a large-caliber Foley catheter in the rectum, filling the balloon with 25 to 30 mL or more of fluid, and pulling the balloon tightly against the top of the anal ring. If this approach fails, epinephrine can be injected at the bleeding site, but sometimes a suture is required to stop the bleeding. A more serious complication is sepsis. There have been five recorded deaths, two additional patients with lifethreatening sepsis, and three cases of severe pelvic cellulitis following RBL of hemorrhoids.17 The onset of sepsis usually is two to eight days after RBL in otherwise healthy people. New or increasing anal pain, sometimes radiating down the leg, or difficulty voiding may be the first indications of a life-threatening infection. Immediate intravenous antibiotics and surgical débridement are required. Cryotherapy Cryotherapy freezes tissue, thereby destroying the hemorrhoidal plexus. Once a popular treatment, its use has declined because of the profuse, foul-smelling discharge resulting from tissue necrosis. The procedure also can be painful, and healing can be prolonged.22 Infrared Photocoagulation Infrared photocoagulation uses infrared radiation to coagulate the tissue, leading to fibrosis. The device is applied for 1.5 seconds in two or three sites proximal to the hemorrhoidal plexus. Reported results for first- and second-degree hemorrhoids are as good as those reported for RBL or sclerotherapy.18 One study reported a 10% relapse rate at three years in patients who had third-degree hemorrhoids and were treated with RBL at recurrence.23 Pain and other complications are rare with infrared photocoagulation. This technique also can be useful when there is a small amount of residual friable hemorrhoidal tissue that cannot be banded, either because there is too little tissue to be incorporated into the band or the tissue is too distal to be safely banded without causing discomfort. In my opinion, this technique generally is less effective than RBL. Surgical Therapy Based on the common finding of increased resting anal canal pressure in patients with hemorrhoids,24 methods to reduce internal anal sphincter pressure in these patients, including internal sphincterotomy and manual dilation of the anus (Lord’s procedure) had been advocated in the past; today, these procedures are mainly of historical interest. One study of the Lord procedure to treat second- and thirddegree internal hemorrhoids with a median follow-up of 17 years found a nearly 40% recurrence rate and a 52% rate of incontinence.25 Lateral internal sphincterotomy occasionally is performed if patients are found to have both a fissure and extensive hemorrhoids at the time of surgery for hemorrhoids.

2261

2262

Section X  Small and Large Intestine Hemorrhoidectomy is the surgical procedure of choice for fourth-degree and some third-degree hemorrhoids and rarely is needed for first- or second-degree hemorrhoids. Hemorrhoidectomy can be done with local, regional, or general anesthesia. Whether the edges of the mucosa are closed or left open after excision of the hemorrhoidal tissue is a matter of preference, because results and severity of postoperative pain are similar with both approaches.26 In one of the few long-term studies of hemorrhoidectomy, recurrent hemorrhoids were found in 26% at a median follow-up of 17 years,25 but only 11% of patients needed an additional procedure. Postoperative pain is the major drawback of hemorrhoidectomy. In an effort to reduce postoperative pain, topical and oral metronidazole have been used with success, although the mechanism of this action is not known.27 Additionally, a new procedure was introduced in 1998 by Longo,28 in which a circular stapler is used to fix the anal cushions in their correct positions. The mucosa is excised circumferentially just above the anorectal ring, thereby interrupting the vascular supply to the cushion (Fig. 125-3). This procedure is called the procedure for prolapse of hemorrhoids (PPH) or stapled hemorrhoidopexy. It is used for third- and fourth-degree hemorrhoids. Results of the randomized multicenter U.S. experience, which compared PPH with traditional excisional hemorrhoidectomy, showed that PPH-treated patients experienced significantly less pain.29 Another study comparing PPH with RBL found patients reported more pain and an increased risk of postoperative bleeding with PPH; however, more patients in the RBL group required excisional hemorrhoidectomy for persistent symptoms.30 PPH can have significant postoperative complications, of which bleeding and urinary retention are most common.29

Severe persistent postoperative pain occurs in one third of patients and may be related to placing the staple line too close to the dentate line.31 Additionally, defecation urgency can be persistent in up to 28%. Perhaps the most feared complication is pelvic sepsis leading to death.32 Several studies analyzing the long-term results of the prospective randomized studies comparing excisional hemorrhoidectomy and PPH suggest a higher rate of recurrent hemorrhoidal symptoms in the PPH group.33-34 Further longterm studies are needed to define the durability of the PPH procedure. Table 125-1 summarizes the treatment options for internal hemorrhoids.

EXTERNAL HEMORRHOIDS AND ANAL TAGS Symptoms and Signs

External hemorrhoids are visible at the anal verge and actually represent residual redundant skin from previous episodes of external hemorrhoidal inflammation and edema. These skin tags typically occur in young and middle-aged adults and are easily seen when the buttocks are parted. They usually cause no symptoms and do not bleed, because they are covered with squamous epithelium; some people are uncomfortable during anal wiping due to irritation or the uncomfortable sensation of redundant tissue. Rarely, external hemorrhoids can interfere with perianal hygiene and cause itching and irritation. External hemorrhoids can be associated with acute pain from thrombosis.13 The level of pain is variable, but patients might notice a rapidly increasing throbbing or burning pain accompanied by a new lump in the anal region. Sometimes the lump has a bluish discoloration caused by the clot. With time, a small area of necrosis can form over the lump, followed by extrusion of the clot, with relief of the pain.

Treatment

Figure 125-3.  Stapled hemorrhoidopexy procedure. A purse-string suture is placed at the top of the hemorrhoidal column around which a circular stapler is applied to resect the upper hemorrhoidal tissue, disrupt the hemorrhoidal blood supply, and restore the prolapsing distal hemorrhoidal tissue back into the anal canal.

Treatment of external hemorrhoids usually is reassurance and proper anal hygiene, including delicate washing of the anal area and avoidance of aggressive wiping with harsh toilet tissue. Rarely is excision recommended because it is painful and results in swelling of the site, which can create redundant tissue that persists after healing. The patient might feel the redundant tissue again and believe it to be a recurrence. When surgical excision is undertaken for internal hemorrhoids, as discussed earlier, any external component usually is excised at the same time. Treatment of thrombosed hemorrhoids depends on the associated symptoms. With time, the pain associated with the acute thrombosis subsides. If the patient has minimal or moderate pain, sitz baths and analgesics are prescribed. For severe pain, the clot is removed under local anesthesia. Because of the high rate of recurrence with simple enucleation alone, most surgeons recommend excising the entire thrombosis and overlying skin. This procedure also can be done in the office with scissors and local anesthesia (Fig. 125-4).35 The skin edges may be closed or left open to heal by second intention. Another successful therapy has been topical application of 0.3% nifedipine cream.36 It is speculated that the success of nifedipine cream in reducing pain from thrombosed hemorrhoids results from its anti-inflammatory and smooth muscle-relaxing properties. Its use can preclude the need for surgery.

Chapter 125  Diseases of the Anorectum Table 125-1  Treatment Options for Internal Hemorrhoids TYPE OF TREATMENT

HEMORRHOID GRADEs

SUCCESS RATE

COMMENTS

1-4

Unknown

Endoscopic Sclerosing agent

Patients with all grades of hemorrhoids should use these measures (see text) Patients with high grades of hemorrhoids will need additional therapy

1-4

75%

Rubber band ligation

2 and 3

65-75%

Infrared coagulation

1 and 2

Surgical Excisional hemorrhoidectomy

Less than for rubber band ligation

May be the favored treatment of patients with acquired immunodeficiency syndrome (successful results even with grade 3 and 4 hemorrhoids) Life-threatening sepsis rarely complicates therapy Grade 1 hemorrhoids are too small and grade 4 hemorrhoids are usually too large for this procedure Most commonly performed office procedure for hemorrhoids Life-threatening sepsis rarely complicates therapy; 1% risk of severe hemorrhage when band sloughs Equipment for the procedure is expensive Rare complications

3 and 4

>85% on 10-year follow-up

3 and 4

>75%; several studies show higher long-term recurrence rates than with excisional hemorrhoidectomy

General Diet (increase in fiber and fluids) and bowel habit modification

Procedure for prolapsing hemorrhoids*

External tags may be removed in the same surgical procedure Postoperative pain is pronounced Newer procedure Overall, significantly less postoperative pain than with excisional hemorrhoidectomy, but some patients experience severe, persistent postoperative pain or defecation urgency Pelvic sepsis and death have been reported after this procedure

*Also called stapled hemorrhoidopexy.

Special Considerations

Large, edematous, shiny perianal skin tags should alert the physician to the possibility of Crohn’s disease. These tags can have a waxy bluish discoloration (Fig. 125-5), are often tender, and are best described as “funny looking.” Careful history taking for symptoms of Crohn’s disease is indicated, and further testing is needed if there is any suspicion that the tags are other than ordinary. Surgical excision of these Crohn’s disease tags is to be avoided in almost every situation, because after surgery the patient usually is left with unhealed anal ulcers. Patients who are infected with the human immunodeficiency virus (HIV) usually are treated as if they do not carry HIV, unless their immune status is significantly compromised. Previously mentioned was the use of sclerotherapy in this population.19 In another study of 11 HIV-positive men with a mean CD4+ count of 420 cells/mm3 and a mean follow-up of six months, no complications occurred after RBL of hemorrhoids, and symptoms improved in all.37 Pregnant women with problematic hemorrhoids usually are managed medically. Simply increasing the fiber and fluid intake and at times adding a stool softener usually is all that is needed. If a complication develops, such as acute prolapse with strangulation, surgical intervention may be necessary. Patients on anticoagulation, including warfarin and other medications, such as aspirin, low-molecular-weight heparins such as enoxaparin, and antiplatelet agents such as

clopidogrel present a special challenge. There is no standard approach to treatment of this group of patients. Certainly it would be optimal to stop the medications, if possible, for five to seven days before any procedure and plan the treatment choice based on the grade of the hemorrhoid. For patients who have symptomatic hemorrhoids and who take warfarin and cannot be off the medication for an extended period, transition to enoxaparin and cessation 12 hours before the procedure is an option. I favor excisional therapy for this group because of the more controlled setting in the operating room; this is especially true if the most definitive treatment is desired. Rubber band ligation can be performed cautiously while a patient is on anticoagulation therapy if the patient is properly informed of the risk of bleeding and if the patient agrees to seek help promptly if bleeding occurs.

ANAL FISSURE An anal fissure is a longitudinal cut in the anoderm; it starts at the anal verge and can extend to the dentate line. More than 90% of anal fissures are located in the mid-posterior position of the anus; 10% are anterior. Any fissure not located in the anterior or posterior position or one that does not heal should alert the physician to the possibility of other diagnoses, such as Crohn’s disease, tuberculosis, syphilis,

2263

2264

Section X  Small and Large Intestine

A

B

C

D

Figure 125-4.  Office excision of a thrombosed external hemorrhoid. A, Thrombosed right anterior external hemorrhoid. B, Injection of 1% lidocaine with epinephrine. C, Excision of external component. D, Closure with 2 or 3 interrupted absorbable sutures.

squamous cell cancer, a leukemic ulcer, or an HIV-related ulcer, which may be of viral or neoplastic origin.

ETIOLOGY

The etiology of a chronic anal fissures is unknown. Using laser Doppler flowmetry, it has been shown that the posterior area of the anoderm is less well perfused than other areas of anoderm. There is speculation that increased tone in the internal sphincter muscle further reduces the blood flow to this area, especially in the posterior midline.4 Based on these findings, fissures are thought to represent ischemic ulceration.38 Trauma during defecation, especially with passage of a hard stool or explosive diarrhea, is believed to initiate formation of a fissure.

SYMPTOMS, SIGNS, AND DIAGNOSIS

Fissures usually are exquisitely tender, and the act of defecation is reported by patients to feel like passing “razor blades” or “cut glass.” The classic history is severe pain during defecation, but patients can experience continued pain or burning after defecation for up to several hours. Bright red blood may be seen on the tissue. On examination, a tender, edematous skin tag may be seen just distal to the fissure. This is often mistaken for a painful hemorrhoid. Simply spreading the buttocks usually increases the pain and leads to anal sphincter muscle spasm. A digital examination causes extreme pain and increases the spasm, and it should be avoided. Once the fissure is healed or the pain has lessened, a thorough examination may be done to

Chapter 125  Diseases of the Anorectum exclude associated problems. If an examination is done and the anal canal is visualized, a hypertrophied anal papilla may be seen in patients with a chronic fissure. If the diagnosis is in doubt or the patient does not respond to treatment, an examination under anesthesia is indicated. Fissures can be acute or chronic. Acute fissures are simply a split or crack in the anoderm without exposed internal sphincter fibers. Chronic fissures show rolled edges, fibrosis of the edges, deep ulceration with exposure of the underlying internal sphincter muscle, enlargement of the tissue at the dentate line (hypertrophied anal papilla), and edematous skin tags at the distal anal verge (see Fig. 125-2; Fig. 125-6).

TREATMENT

Treatment of acute and chronic anal fissures (Table 125-2) starts with modification of the diet similar to the conservative treatment of hemorrhoids described previously. Patients follow a high-fiber diet with fiber supplements, and they

increase their fluid intake. Soaking in a sitz bath or warm tub relaxes the sphincter and can provide significant relief.13 If needed, stool softeners are added. Acute fissures respond better to these measures than do chronic fissures, but such treatment is still a critical component to the management of a chronic fissure.

Table 125-2  Treatment of Acute and Chronic Anal Fissure TREATMENT

COMMENTS

Acute Increase oral fluids, high-fiber diet, fiber supplements, sitz baths, and stool softeners if needed

Chronic As for acute, usually with one of the following treatments added   0.2-0.4% nitroglycerin cream applied to the anal area

  Calcium channel blockers (topical nifedipine or topical 2% diltiazem cream)

  Botulinum toxin A injected into anal muscle

  Lateral internal sphincterotomy Figure 125-5.  Anal skin tags associated with perianal Crohn’s disease. Note the waxy bluish appearance of the tags.

Avoid digital rectal examination until the fissure is healed unless the diagnosis is in doubt (then examine in the operating room if needed) Usually responds to these measures Avoid digital examination unless the diagnosis is in doubt Headache is a common side effect Long-term success has been questioned Seem promising, but long-term success has been questioned Side effects (especially headache) may be less common than with nitroglycerin cream Dose and optimal injection site are not clear Expensive Long-term success is unknown Standard treatment Best results, with >90% long-term healing rate Durable Can lead to fecal incontinence

Hypertrophied anal papilla

Acute fissure

Skin tag

Chronic fissure

Figure 125-6.  Schematic depiction of acute and chronic anal fissure. An acute fissure is depicted in the inset on the left as simply a split in the anoderm. A chronic fissure can show signs of chronicity with rolled edges, fibrosis, a hypertrophied anal papilla proximally, a tender distal skin tag, and exposed internal anal muscle.

2265

2266

Section X  Small and Large Intestine Medical Therapy

The traditional surgical treatment of chronic anal fissure, lateral internal sphincterotomy, has come under intense scrutiny since the 1990s. Because of concerns about fecal incontinence, pharmacologic therapy has gained popularity. One of the initial treatments was topical nitroglycerin ointment applied to the anal region. A pea-sized amount of 0.2% to 0.4% nitroglycerin ointment with gradual escalation of the dose applied three times daily is the most common way to use the medication. Headache is a significant adverse complaint,39 and a gradual increase in dose might reduce this drawback. Patients should be cautioned to wear a finger cot or surgical glove when applying the ointment to prevent absorption through the digital skin. Despite early successful clinical trials, long-term success rates have been questioned. A large systematic review found that healing of fissures with nitroglycerin was only marginally better than with placebo.39 One randomized trial found overall healing in about only one third of patients at six months.40 Calcium channel blockers (nifedipine and diltiazem) have been advocated because of their ability to relax the internal anal sphincter. Although these topical agents have been studied, far fewer studies are published for them than for nitroglycerin. Topical 2% diltiazem gel has been used to treat nitroglycerin-resistant anal fissures. In one study of 39 fissures unhealed by nitroglycerin, 49% were healed at eight weeks after use of diltiazem gel.41 This treatment modality generally is preferred to nitroglycerin because it does not commonly cause headaches, which are often seen with topical nitroglycerin. Topical 2% diltiazem, applied two to three times daily, is my preferred topical treatment for a chronic anal fissure. The injection of botulinum toxin A, which inhibits acetylcholine release, into the internal sphincter close to a fissure, was described first by Jost and Schimrigk in 1993.42 Since then, this modality has gained popularity as a treatment for chronic anal fissure. In a study from Italy, 73% of anal fissures were healed at eight weeks,43 with no recurrences a mean of 16 months later. In one study comparing botulinum toxin injection with topical nitroglycerin ointment, at eight weeks, fissures were healed in 96% of patients injected with botulinum toxin and in 60% of those treated with nitroglycerin. In neither group was there a recurrence at a mean follow-up of 15 months.44 Comparison of botulinum toxin with lateral internal sphincterotomy showed that at one year, more fissures were healed after surgery (94%) than with the injection (75%).45 Long-term results, however, might not be as optimal as hoped with botulinum toxin injection. In one study of 57 patients, only 22 (41%) had healed fissures a median of 42 months after botulinum injection.46 Two meta-analyses comparing botulinum toxin to lateral internal sphincterotomy demonstrated that longterm healing was better with surgery but that side effects and complications were far less with botulinum toxin injection.47,48

Surgical Therapy

The standard treatment for chronic anal fissures has been lateral internal sphincterotomy, and it remains the standard by which all other treatments must be measured.49 One long-term study found that among 2108 patients with anal fissures who underwent outpatient surgery with local anesthesia and intravenous sedation, the results were very good to excellent in 96% at a follow-up of 4 to 20 years. The rate of recurrent fissures was 1%. Permanent incontinence did not occur.49 When performed correctly, lateral internal sphincterotomy still appears to be superior (and probably

cheaper in the long term) to any currently available medical treatment.50,51

Special Considerations

Postpartum women require special consideration. In one study, painful fissures developed in 29 (9%) of 313 primigravid women in the immediate postpartum period. Anal sphincter hypertonia was not seen in these patients,52 and because vaginal delivery is associated with anal trauma, division of the anal sphincter in this setting may be detrimental. Of the 29 fissures, 27 healed with medical treatment. The researchers suggested that if surgical intervention is needed in a postpartum woman with an anal fissure, an advancement flap (which raises a flap of healthy tissue to cover the open area) may be preferable to a lateral sphincterotomy.53

ABSCESSES AND FISTULAS Almost all anorectal suppurative disease results from infection of the anal glands that extend from the anal crypts, located along the dentate line at the base of the columns of Morgagni. Acute infection can cause an abscess and lead to a chronic fistula-in-ano. Other causes of such infection include Crohn’s disease, fissures that bore into the anal muscle, hematologic malignancies, tuberculosis, actinomycosis, trauma, foreign bodies, and anal surgery. The differential diagnosis also includes a pilonidal sinus, hidradenitis suppurativa, Bartholin’s gland abscess, carcinoma, and lymphoma.

ABSCESSES

Because nearly all of the anal glands terminate in the intersphincteric plane, abscesses tend to originate in the intersphincteric space, from where they can travel up, down, or circumferentially around the anus. Abscesses are classified according to where they extend, and may be perianal, ischiorectal, intersphincteric, or supralevator (Fig. 125-7). The most common type is the perianal abscess (40% to 50%), and the least common type is the supralevator abscess (2% to 9%).54 The diagnosis of anorectal abscess is based on typical symptoms and signs. Swelling, throbbing, and continuous pain are the most common symptoms. On examination, erythema, swelling, and marked tenderness may be present. If the abscess is in the intersphincteric space, however, there may be no abnormal findings externally. Nevertheless, a digital examination may be impossible because of pain, or a boggy area may be felt over the internal anal sphincter adjacent to the abscess. An ischiorectal abscess produces pain in the buttock, but no abnormality may be appreciated on examination. The ischiorectal space is a potential space that can be large, and pus can move upward rather than toward the skin.54 Symptoms of a supralevator abscess may be intra-abdominal or urinary, and patients can have lower abdominal discomfort and urgency without any anorectal complaints or clinical findings. If the patient cannot be evaluated because the pain is severe or if no abnormality is found on examination, the patient should undergo an examination under anesthesia. Treatment of an abscess in the perineal region requires incision and drainage. Antibiotics alone are not adequate. Failure to drain an abscess promptly can result in spread to adjacent spaces, and some necrotizing infections can be mutilating and life-threatening. Small abscesses can be drained in the office. The external opening should be made

Chapter 125  Diseases of the Anorectum

Supralevator abscess

Ischiorectal abscess

Intersphincteric abscess

Intersphincteric abscess

Perianal abscess

as close to the anal sphincter complex as possible without injuring it. Therefore, if a fistula develops, its tract will be short. The incision should be large enough or made in a cruciate fashion so that it will not close over before the inflammatory process has resolved. Packing should not be used, although some surgeons prefer to place a small mushroom catheter into the cavity to facilitate drainage. Large or high abscesses require drainage in the operating room. It usually is not necessary to culture the pus. Hospitalization and intravenous antibiotics are reserved for patients who are immunocompromised or diabetic or who have signs of systemic infection, such as high fever. Patients should be followed closely to ensure that the process resolves.

FISTULA-IN-ANO

A fistula-in-ano is a tunnel that connects an internal opening, usually at an anal crypt at the base of the columns of Morgagni, with an external opening, usually on the perianal skin. A fistula-in-ano develops in half of patients who undergo incision and drainage of an abscess. A fistula is diagnosed upon seeing blood, pus, and sometimes stool drain from its external opening. Some patients complain of perianal itching. If the tract seals over, pus can accumulate within the tract and pain can develop. Long tracts can have secondary openings along their course. If the tract is chronic, it may be palpated as a cord under the skin. Pus sometimes may be expressed from the opening when the cord is palpated. Typically, anoscopy reveals the internal opening of the fistula. The differential diagnosis of fistula-in-ano is the same as for anal abscesses. Treatment of fistula-in-ano is surgical intervention. The course of the fistulous tract influences the type of surgical treatment (Fig. 125-8). The most common treatment is a fistulotomy, or unroofing of the tunnel. Fistulotomy should not be performed if the tract traverses a substantial portion of the external sphincter, in which case its division will result in incontinence. Fistulas that are appropriate for fistulotomy can be unroofed, however, and the base curetted and allowed to heal from the bottom up. The cure rate for uncomplicated fistulas not associated with Crohn’s disease approaches 100%.

Figure 125-7.  Schematic depiction of the classification of abscesses of the anal region based on their location.

A fistula that involves a substantial portion of the anal sphincter requires special treatment to avoid incontinence. The options are transanal advancement flap, skin advancement flap, fibrin glue injection, or collagen plug insertion. Transanal advancement flaps are a common surgical repair for these complex fistulas. Although initial success rates were higher, more-recent studies found success rates varying from 67% to 77%.55,56 Rates of incontinence after a flap repair range from 9% to 35%.55-57 Alterations in continence tend to occur more often in patients older than 50 years.57 Fibrin adhesive made from commercial fibrin sealant has been used to close anal fistulas. The durable healing rate for cryptoglandular fistula was 23% in one study.58 Other studies, however, have reported rates of 59% to 92%,59,60 and most groups now report a long-term success rate of less than 33%. The fistulous tract is curetted aggressively, and the fibrin product is injected via the external opening until it is seen to emerge in the anal canal. Few complications have been reported with this method, and continence should not be affected. Retreatment for failures can be successful. The tract must be at least one centimeter long to allow sufficient length for the fibrin plug to adhere to the tract. This technique however, for the most part has been replaced by a collagen-plug technique because of high failure rates. An anal fistula plug made from collagen has been used increasingly since the turn of the century. In the initial report by Armstrong, healing occurred in 38 of 46 (83%) patients, with a median follow-up of 12 months.61 However, in 2008, five other groups reported short-term success in only 30% to 60% of patients.62-66 My success rate is approximately 50% (Fig. 125-9). Although these later published results are disappointing, they are still improved over the results of fibrin glue injection, and the procedure is being used widely in the United States. Direct closure of the internal opening of a transsphincteric fistula also has been reported, and in one series of 106 patients only seven had persistent fistula, for an overall closure rate of 93.4%67; 24 of the 106 patients, however, did require more than one operative procedure to obtain definitive fistula closure. This technique is not widely used.

2267

2268

Section X  Small and Large Intestine Extrasphincteric fistula

Transsphincteric fistula

Figure 125-8.  Schematic depiction of the classification of fistula-in-ano. Fistulotomy is not appropriate for extrasphincteric fistulas because it would leave the patient incontinent.

Intersphincteric fistulas

Figure 125-10.  Blue setons in a patient with anal Crohn’s disease.

Figure 125-9.  Anal fistula plug being placed through an anterior anal fistula tract. Once it is pulled snugly through the tract, the plug is secured internally and the internal opening is closed. Excess plug projecting externally is also trimmed.

A seton is a rubber band-like material that is threaded through a fistula so that its two ends can be tied loosely external to the outside skin (Fig. 125-10).68 Setons may be used to drain a fistulous tract before surgical repair to prevent the accumulation of pus; however, they also can be therapeutic if used as a cutting seton. In this latter approach, after the seton is placed, the skin is incised over the fistulous tract, and the seton is tightened gradually over weeks to months until it gradually cuts through the muscle. This technique allows the fistulous tract to be unroofed gradually

as the cut ends of the muscle scar close to their usual location, thereby lessening the chance of incontinence. Setons also may be used as a prolonged treatment of an anal fistula to prevent (recurrent) formation of abscesses. In this situation, the seton is tied loosely (typically using a sterile rubber or vessel loop) around the fistulous tract. The patient predictably has some minor continuous drainage but is unlikely to develop further abscess formation. This technique is considered when the patient has anal Crohn’s disease (see later), has multiple fistulas, has failed multiple procedures in the past, or is a poor candidate for surgery.

SPECIAL FISTULAS Fistulas in Crohn’s Disease

Fistulas due to Crohn’s disease might not be cryptoglandular in origin but may be related to increased lymphoid tissue

Chapter 125  Diseases of the Anorectum in the intersphincteric plane. They usually are complex, with curved, multiple tracts. Any unusual fistula or nonhealing fistulotomy site should raise the possibility of Crohn’s disease. Anal disease may be the first or the only manifestation of Crohn’s disease. Anorectal abscesses and suppuration may be less tender in patients with Crohn’s disease than in some patients with cryptoglandular fistulas.69 Treatment is tailored to the overall activity of the Crohn’s disease (see Chapter 111). The goal is to improve the patient’s quality of life; curing the Crohn’s disease is not realistic, and healing the fistula might not be possible. In patients with severe colonic or rectal Crohn’s disease, the placement of a loose seton in an anal fistula can decrease symptoms; the seton may be left in indefinitely.69 Some patients improve simply with the addition of oral metronidazole with or without ciprofloxacin.70 The immunomodulator 6-mercaptopurine also has been used with some success, A meta-analysis in which fistula closure was a secondary end point revealed healing in 54% of patients treated with azathioprine or 6-mercaptopurine compared with 21% who healed on placebo.71 In two randomized, controlled trials, infliximab was shown to be effective in healing perianal Crohn’s fistulas.72,73 In the first, closure of all fistulas for four weeks was seen in 38% of patients treated with infliximab at a dose of 5 mg/kg, in 55% of those treated with 10 mg/kg, and in 13% of patients receiving placebo.72 In another randomized, controlled trial of 296 patients, 87% of whom had perianal fistulas, complete closure of fistulas was seen at 54 weeks in 36% of patients treated with infliximab (5 mg/kg) compared with 19% of those treated with placebo. Of patients who lost response during maintenance treatment, 61% of patients who had been maintained on placebo responded to infliximab (5 mg/kg) and 57% of patients maintained on infliximab at 5 mg/kg responded to an increased dosage (10 mg/kg).73 In 10 uncontrolled series using cyclosporine, a rapid response was seen in 83% of patients, but a high relapse rate occurred when patients were switched from intravenous to oral therapy.74 Fistulotomy may be done for low fistulas with minimal anorectal involvement69; however, a nonhealing wound can persist after such therapy. Advancement flaps may be used successfully to close Crohn’s fistulas, as long as Crohn’s disease in the anorectum is only minimally active at the time of attempted repair. Despite the array of approaches, some patients have severe perianal sepsis and require fecal diversion with an ileostomy. In many, proctectomy is required later because of persistent discomfort or other problems. In general, surgical repair must be used with caution in anal fistulas in patients with Crohn’s disease, and chronic, indwelling, loose setons are preferred to control symptoms of sepsis while preserving anal function.

Anovaginal Fistulas

In women with Crohn’s disease, anovaginal fistulas present additional challenges. Traditionally, proctectomy or longterm seton drainage have been the accepted methods of treatment. Most such fistulas are too short to be amenable to fibrin glue. Surgical treatment with various flap repairs, however, has been successful: In one study of 35 women, the initial success rate in selected women was 54%, with 68% healing after multiple flap procedures.75,76 Anovaginal fistulas not associated with Crohn’s are also difficult to treat. Almost all are related to obstetric injury, and a few are cryptoglandular in origin. The local tissue usually is scarred and too rigid for successful treatment with advancement flaps. The anovaginal septum is usually very

thin. Repair of the sphincter in conjunction with repair of the fistula to give more tissue bulk to the anovaginal septum is the procedure of choice in this situation.77

Fistulas after Radiation Treatment

Fistulas associated with radiation treatment, usually for cervical or prostate cancer, also are challenging (see Chapter 39). The first issue is to exclude recurrent cancer as the cause of the fistula. If the output of stool and gas per vagina is great, a stoma may be needed. A stoma also allows the bypassed tissue to soften, so additional treatment options can be evaluated. It can take up to a year for the area to become pliable enough for a reasonable attempt at surgical intervention to be undertaken. Repair with a flap can be attempted if the tissue looks normal and is pliable, but this rarely is the case. Resection of the rectum with anastomosis of healthy colon to the anus at the dentate line can be successful (coloanal anastomosis). Interposition of the gracilis muscle between the wall of the anorectum and vagina or urethra brings healthy tissue to the area and can lead to healing.

ANAL MALIGNANCIES Anal cancers are rare. They account for 1.5% of gastrointestinal cancers in the United States with 3500 new cases each year.78 The incidence has increased 2% to 3% every year in the United States since the early 1980s.79 Almost 80% are squamous cell cancers, 16% are adenocarcinomas, and 4% are other types.80 Adenocarcinoma of the anal canal behaves like adenocarcinoma of the rectum and is treated as such. Patients undergo abdominal-perineal resection with pre­ operative or postoperative chemoradiation therapy when the tumor is large, invades other organs, or has nodal involvement. Tumors arising in the distal anal canal usually are keratinizing squamous cell carcinomas. Those arising in the transitional mucosa often are nonkeratinizing squamous cell carcinomas. In the past, the two nonkeratinizing subtypes were referred to as transitional cell and cloacogenic. All anal carcinomas today are considered to be variants of squamous cell carcinoma, with various degrees of glandular and adnexal differentiation. The type previously called transitional cell carcinoma is composed of large cells, and the type previously referred to as cloacogenic carcinoma is composed of small cells,81 but microscopically, these tumors may be heterogeneous, and it is not useful to subclassify them. Keratinizing and nonkeratinizing squamous cell cancers exhibit similar behavior. Anal bleeding is the most common symptom (45%), followed by the sensation of a mass (30%); 20% of patients have no symptoms.81 The development of anal cancer has been associated with human papillomavirus (HPV) and HIV infections, history of receptive anal intercourse, history of sexually trans­mitted diseases, history of cervical cancer, and use of immunosuppressive medication after solid organ transplantation.82

ANAL MARGIN CANCERS

Cancers arising distal to the anal verge (anal margin) are considered skin cancers and are treated as such. Small lesions (<4 cm2) with no fixation to deeper tissues are excised widely. Treated patients are then followed closely for five years. If the disease recurs, chemoradiation therapy is started. Squamous cell cancer of the anal margin that is deeply invading is treated by chemoradiation.

2269

2270

Section X  Small and Large Intestine ANAL CANAL CANCERS

In the past, standard treatment of anal canal cancers was abdominal-perineal resection with a permanent colostomy. In 1974, Nigro and colleagues presented the results of combined radiation and chemotherapy and showed that cure was possible without abdominal-perineal resection.83 This led to a regimen of external-beam radiation with fluorouracil and mitomycin as the treatment of choice; surgery was reserved for residual cancer seen in the scar after treatment. Cisplatin has been substituted for mitomycin in treatment trials, and complete response rates to combination treatment have been seen in 94% of patients. In one study, at a follow-up of 37 months, only 14% of patients required a colostomy for residual or recurrent disease.84 With a followup mean of 5.6 years, another study found 30% of patients needed a stoma for recurrent or persistent disease.85 Patients with persistent or recurrent squamous cell carcinoma of the anal canal are treated with an abdominalperineal resection. About 50% of such patients who undergo such surgery can be cured.81 Success also has been reported with an additional boost of radiation therapy combined with cisplatin-based chemotherapy.81

MELANOMA

Melanoma is as deadly in the anal region as elsewhere. Surgical excision offers the only chance for cure, and even this treatment has poor results. Because of the rare success with any treatment, many investigators have questioned the use of abdominal-perineal resection for anal melanoma instead of wide local excision. In one study, 2 of 19 (10%) patients survived, both of whom had wide local excision alone.86 The authors commented that anal ultrasound may be a valuable guide to treatment and recommended that if ultrasound shows that the lesion can be excised, wide excision should strongly be considered to avoid a permanent stoma, because most patients have regional or distant metastasis at presentation.86

ANAL INTRAEPITHELIAL NEOPLASIA

Anal intraepithelial neoplasia (AIN) is described as a dysplastic condition of the epithelium of the anal canal and perineal skin. There are several other commonly used terms for this condition, such as Bowen’s disease, squamous cell carcinoma in situ, or anal squamous intraepithelial lesion (SIL). Anal intraepithelial neoplasia is the more commonly used term now. There seems to be a progression from lowgrade (AIN I) to high-grade (AIN III) lesions, similar to that observed in the cervix. It appears that Bowen’s disease is AIN III. AIN may be discovered at the time of pathologic examination of anal tissue obtained for an unrelated surgery.87 Disagreement continues regarding treatment of high-grade AIN.87 One approach is by high-resolution anoscopy with directed therapy: 3% acetic acid is applied to the distal rectal mucosa, anal mucosa, and perianal skin to allow aceto-whitening. Examination is then done with an operating microscope, looking for patterns suggesting high-grade dysplasia such as punctation and mosaicism. Then the tissues are painted with 10% iodine (Lugol’s solution). Normal tissue will appear dark, and tissue that has highgrade intraepithelial changes appears mustard-yellow to tan. Directed destruction of tissue that has the characteristics of high-grade dysplasia is performed with electrocautery88; using this approach, no HIV-seronegative patient developed recurrent high-grade AIN; however, 79% of HIVpositive patients had recurrent or persistent lesions. Another approach reasons that because the condition stems from HPV infection (anal warts; see later) as a predis-

posing factor,89 the virus will remain in the “normal” tissue that is not excised. Accordingly, the best treatment option for patients with AIN III (including HIV-positive patients) may be close observation, with regular biopsy of any suspicious areas to exclude invasive malignancy.90,91

PAGET’S DISEASE

Paget’s disease is a rare intraepithelial mucinous adenocarcinoma that appears as an erythematous, eczematoid plaque and probably arises from the dermal apocrine sweat glands. The disease is more common in women than men and manifests with intractable itching. Diagnosis is by biopsy, and wide local excision is the treatment if invasion is not found. For invasive cancer, abdominal-perineal resection is the treatment of choice.

ANAL WARTS Anal warts, or condylomata acuminata, are caused by HPV. It is believed that most squamous cell cancers of the anal area also are caused by this virus. Condylomata affect 5.5 million Americans yearly, and the prevalence in the United States is about 20 million persons.92 Condylomata are caused by sexual transmission of HPV, although nonsexual transmission is possible.93 These lesions occur more commonly in men who have sex with men, in immunosuppressed patients, and in persons who have large numbers of sexual partners.87 Numerous treatment modalities exist (Table 125-3), but even when bulky lesions have resolved, the virus remains. Podophyllin is a topical agent that is antimitotic. It requires repeated applications in the office.94 As a single agent, it results in cure rates of 50%.94,95 Podophyllin cannot be used in the anal canal, however, and is poorly absorbed by lesions that are keratinized, a characteristic of longstanding warts. The drug can cause skin irritation and is teratogenic in animals. Trichloroacetic and dichloroacetic acid cause sloughing of tissue. These acids must be used with care to control the depth and size of the wound. They can be used in the anal canal. Cure rates of 75% have been reported with their use.94,96 Cryotherapy can be used in the anal canal. The depth and width of the wound must be monitored carefully. Success rates are similar to those associated with trichloroacetic acid.94,97 Topical 5-fluorouracil (5-FU) penetrates the skin and is used in a 5% cream. Success rates for 5-FU have ranged from 50% to 75%. Perhaps the best use of topical 5-FU is its biweekly application after surgical removal of warts to decrease rates of recurrence.98 Imiquimod cream is an immune response modifier that stimulates monocytes and macrophages to produce cytokines that affect cell growth and have an antiviral effect.99 The cream is applied to the warts at bedtime three times a week for a total of 16 weeks. Imiquimod cannot be used for anal canal warts. The drug seems to work better for women than for men, with one study reporting clearance of warts in 72% of women compared with 33% of men.100 Few side effects have been reported, although local skin irritation is common. Surgical excision and cautery yields the highest success rate, with cure rates of 63% to 91%; laser seems to offer no advantage over cautery. Disadvantages of cautery include the need for local or other forms of anesthesia and the presence of bioactive HPV in cautery-induced

Chapter 125  Diseases of the Anorectum Table 125-3  Treatment Options for Anal Warts TREATMENT

SUCCESS RATE

COMMENTS

Podophyllin

20-50%

Trichloroacetic or dichloroacetic acid

75%

Cryotherapy

75%

Topical 5-fluorouracil

50-75%

Imiquimod Surgical excision (usually combined with cautery)

75% in women, 33% in men 60-90%

Intralesional interferon-α

>70%

HspE7

Experimental

External-beam radiation therapy

Variable

May need repeat applications Skin irritation can occur Not used in the anal canal Poorly absorbed by keratinized lesions (most chronic warts are keratinized) Can be used in the anal canal Care is required to control the size of the slough Can be used in the anal canal Care is required to limit the size of the wound Fumes from the therapy can contain active HPV Probably best if used after surgical excision to decrease the frequency of recurrence Cannot be used in the anal canal; works better in women Fumes from the cautery may contain HPV May need to be done in more than one session if a thick carpeting of lesions is present to avoid excising or burning excessive anoderm Injected into base of up to five warts three times a week for 3-8 weeks Approved by U.S. FDA for refractory condyloma Promising treatment involving subcutaneous injections Fusion protein that combines immune-stimulating properties and a target antigen from HPV Reserved for giant cavitating condyloma (Buschke-Lowenstein lesions) Used as the last resort, usually when bleeding or tissue invasion cannot be controlled

FDA, Food and Drug Administration; HPV, human papillomavirus.

fumes.94,101 Immune status seems to influence recurrence rates of condyloma. One study found significantly more recurrences in a shorter period for those with an immunocompromised status (defined as persons who are HIVpositive, or who have leukemia, idiopathic lymphopenic syndrome, a transplanted organ, or who are on chemotherapy).102 Interferon-α is approved by the U.S. Food and Drug Administration for injectional therapy of refractory condylomata acuminata. The dose is 106 units injected beneath a maximum of five lesions up to three times weekly for three to eight weeks. Recurrence rates are 20% to 40%.103 Topical interferon cream is ineffective.104 Surgery to eradicate the largest lesion, combined with topical or intralesional therapy, can decrease recurrence rates, and I have used it successfully when patients have a rapid recurrence of disease following surgical excision alone. One promising therapy appears to be HspE7. This is a novel fusion protein that combines immune stimulatory properties with an appropriate target antigen from HPV. In one study, HspE7 was broadly active for multiple HPV types.105 The warts were found to improve substantially, but they usually did not disappear totally at the six-months follow-up. Patients were HIV-negative and received three subcutaneous injections of HspE7. More studies are under way to further evaluate these findings. Anogenital warts can affect immunocompromised patients, including those who are HIV-positive and transplant recipients. In such persons, the warts are more aggressive, recur earlier, and are more often dysplastic than in immunocompetent persons. In HIV-positive patients, dysplasia and histologic evidence of HPV can occur in the absence of gross warts.106 Gross warts should be treated as discussed earlier. Topical 5-FU cream and serial examina-

tions, rather than extensive excision, have been advocated for HIV-positive patients with dysplasia.106 Excision is reserved for patients with obvious lesions of the skin. Anal cytology (the “anal Pap test”) has been recommended for this group of patients, but as yet there are no uniform guidelines for its use. Buschke-Löwenstein tumors are a rare variant of anal warts. These lesions appear as giant condylomata that grow rapidly, invade locally, and cause extensive destruction of surrounding tissue. Treatment is surgical excision, if possible. These lesions also have responded to radiation therapy.107 Table 122-4 outlines the treatment options for anal warts.

PRURITUS ANI Pruritus ani is an itch localized to the anus and perianal skin. Most patients presenting with pruritus ani believe their symptoms are from hemorrhoids, but most often this is not the case. Pruritus ani is categorized as either idiopathic (primary) or secondary. Idiopathic pruritus ani is diagnosed when no underlying etiology is found. Secondary pruritus ani results from an underlying disorder, and specific treatment leads to resolution of symptoms. Because pruritus ani is poorly understood, and an underlying premalignant lesion such as Bowen’s disease or Paget’s disease may be the cause of symptoms, all patients with pruritus ani deserve a thorough investigation. In a prospective study, 109 patients were referred over a period of two years to a colorectal clinic with a primary symptom of itching. On evaluation, 26 patients were found to have some form of colorectal or anal dysplasia, 56 had

2271

2272

Section X  Small and Large Intestine anorectal disease (most commonly fissures or hemorrhoids), and 27 had no identifiable etiologic disorder.108 After surgical treatment of all patients with colorectal neoplasia, their itching resolved. Of the 56 patients with anorectal disease, six continued to have itching after medical or surgical treatment. All 27 of the patients with no obvious cause for their itching underwent conservative medical treatment, and six continued to have persistent symptoms.7 A thorough history and physical examination are the starting points for evaluating patients with anal itching. Examination of the rectum and sigmoid colon should be included. The pattern of irritation should be noted, and consideration should be given to biopsy any abnormal skin. Usually, diet-induced pruritus is symmetrical, whereas infectious causes lead to an asymmetrical pattern of anal irritation. Leakage of stool due to fecal incontinence and leakage of mucus are common sources of pruritus ani and may be associated with prolapsing hemorrhoids or true rectal prolapse. Other causes include contact dermatitis, infections (e.g., candidiasis), parasites, systemic diseases (e.g., diabetes mellitus), diet (caffeine in coffee, cola, or chocolate; dairy; citrus; beer; and others), and some medications. It has been said that dietary factors, especially any form of caffeine, may be the most common culprit.109 The exact mechanism whereby caffeine might act as an irritant to the perianal skin is unclear. Any underlying disorder should be treated; however, in many patients the cause of pruritus ani cannot be identified. These patients should be advised to modify their cleansing habits and to eliminate potential caffeine-related dietary culprits such as coffee, tea, and chocolate during a trial period. Most important is to convince the patient to stop “polishing” his or her anus. Frequently, the patient feels that the anal area is unclean, and he or she rubs the area vigorously, both for comfort and to try to clean the area more completely. Wiping gently with wet facial tissue or baby wipes is recommended. Avoiding soap and a washcloth in the shower may help. The patient should be instructed to use plain water and his or her hand to wash the perineum in the tub or shower. Creams or emulsifying ointments may be used instead of soap.110 Perfumed soaps and astringents (such as witch hazel) should be avoided. Tissue paper or a cotton ball placed by the anus and changed several times a day will absorb moisture and create a drier environment. A diet high in fiber with plenty of fluids (similar to the diet described for hemorrhoids) is recommended. A limited amount of 1% hydrocortisone cream can be used. Patients should be warned that chronic use of hydrocortisone will thin the anal skin and can lead to more problems. Most patients respond to this regimen. Relapse is common and can require re-education of the patient. The physician should be aware that a previously overlooked underlying problem may be the cause of the pruritus and should take a fresh look at a patient who has a relapse. Assistance from a dermatologist also may be helpful and should be considered initially.110 Intradermal injection of methylene blue has been used successfully for the treatment of intractable idiopathic pruritus ani that has not responded to any other measures.111

ANAL STENOSIS Anal stenosis is a narrowing of the anal canal. The condition can develop after anorectal surgery, usually radical hemorrhoidectomy. Surgeons in the modern era have modified their surgical technique in an attempt to prevent this

problem. Other causes include chronic anal fissure, inflammatory bowel disease (especially Crohn’s disease), chronic diarrheal disease, infections (tuberculosis, lymphogranuloma venereum, syphilis, and others), cancer, and irradiation (see Chapter 39).112 Treatment of anal stenosis depends on the degree of stenosis and associated symptoms. In mild stenosis, the examiner’s finger can be placed just through the narrowing. Patients consume a high-fiber diet and bulking agents to produce stools that pass more easily, but patients must be cautioned to drink sufficient fluids so that they do not become constipated. Additionally, the bulky stools provide natural dilation. Gradual dilation by the patient (usually daily in the shower) with a commercial medical dilator or a smooth white candle also can produce improvement.113 The candle should be devoid of fragrance and coloring, both of which could potentially be irritating to the anal skin. In moderate stenosis, more force is required to insert the index finger. Patients might respond to dietary changes and graded dilations. The initial dilation may need to be done under anesthesia. If improvement is insufficient, surgery is indicated. Release of a stricture or a sphinc­ terotomy might suffice; however, some patients require an advancement flap.113 In severe stenosis, even the tip of the examiner’s fifth finger might not be able to be inserted into the anal canal. There is loss of the anoderm, and surgical intervention is needed to deliver healthy tissue into the anal canal to compensate for this loss. Various advancement flaps are successful, and the type used depends on the patient’s anatomy and the surgeon’s preference.113

UNEXPLAINED ANAL PAIN Unexplained anal pain refers to pain in the anorectal region in the absence of an underlying anatomic abnormality. Diagnosis is based almost entirely on the patient’s symptoms. Confusing nomenclature has clouded the issue further.

COCCYGODYNIA

Coccygodynia is a pain or ache in the coccyx, commonly referred to as the tailbone, and typically results from trauma or arthritis. Movement of the coccyx on digital rectal examination can reproduce the pain. Treatment in the acute setting includes sitz baths, NSAIDs, and stool softeners.114 Glucocorticoid injections have been used in an attempt to reduce the pain. Rarely, removal of the coccyx is necessary.115

LEVATOR ANI SYNDROME AND PROCTALGIA FUGAX

Levator ani syndrome and proctalgia fugax are probably not the same entity. In the literature, however, the terms have been used interchangeably, but incorrectly so.

Symptoms and Signs

Levator ani syndrome typically affects women younger than 45 years of age. Episodes of pain are chronic or recurring and, by definition, occur during 12 weeks in the preceding year, with each episode lasting 20 minutes or more.116 The discomfort is described as a vague tenderness or aching sensation high in the rectum. Discomfort usually does not awaken the patient from sleep and usually is worse after defecation and with sitting; walking or lying down seems to relieve the pain.117 Symptoms have been attributed to spasm of the levator muscles. Physical examination may be

Chapter 125  Diseases of the Anorectum normal, or the levator muscle might feel like a tight, tender band on rectal examination.114 Proctalgia fugax occurs in young men and persons with perfectionistic-type personalities. It is seen in early adulthood and subsides by middle age. The pain lasts seconds or minutes and then disappears. Pain is described as a sharp cramp or stabbing and can awaken the patient from sleep.118,119 The pathogenesis is thought to involve anal smooth muscle dysfunction, perhaps triggered by stressful events.117 The frequency of proctalgia fugax may be increased in patients with other functional gastrointestinal complaints, such as unexplained abdominal pain, bloating, irritable bowel syndrome, and a sensation of incomplete evacuation after defecation. Physical examination is normal.114,116,117

Treatment

Treatment of the levator ani syndrome and proctalgia fugax is controversial, and no single treatment works for all patients. Treatments for these disorders are grouped together, in part because of our incomplete understanding of the two conditions. Initial treatment is reassurance, sitz baths, perineal strengthening exercises, regulation of bowel habits, and nonsteroidal anti-inflammatory medications. These measures are usually effective.114 Many other treatments have been proposed, including electrogalvanic stimulation,118 levator massage, biofeedback,119 drug therapies, acupuncture, and psychiatric evaluation. In the past, the drug treatment of choice was benzodiazepines; however, because of their addictive potential and the chronic nature of these problems, the popularity of these drugs has declined. In the only randomized, controlled trial for proctalgia fugax, a salbutamol inhaler was used successfully.120 Two puffs at the onset of pain were reported to lead to rapid relief. Clonidine, an α2-adrenergic agonist, relaxes smooth muscle and has been used successfully.121 The dose is 150 mg twice daily for three days, tapered to 75 mg twice daily for two days, and then 75 mg daily for two days. Diltiazem, a calcium channel blocker, relaxes smooth muscle and also has been effective in a dose of 80 mg twice daily.122 Topical nitroglycerin, 0.2% or 0.3%, applied at the onset of proctalgia fugax also has been used successfully.123 I have had good results using amitriptyline, a tricyclic antidepressant, a side effect of which is muscle relaxation. The initial dose is 25 mg nightly and can be increased to 50mg after two weeks, if needed. A newer treatment involves linearly polarized nearinfrared irradiation therapy.124 One study used this treatment for patients with strongly tender points in each side of the rectum rather than in the anal canal. Of 35 patients, 33 had good or excellent improvement in pain after an average of 2.5 20-minute treatments.124

HIDRADENITIS SUPPURATIVA Hidradenitis suppurativa is an inflammatory disease of the apocrine sweat glands and adjacent connective tissue. The initiating event is occlusion of an apocrine duct by a keratinous plug, leading to ductal dilatation and stasis, and then to secondary bacterial infection. The infection often ruptures into the surrounding soft tissue (Fig. 125-11). The chronic, cyclic nature of the disease ultimately leads to fibrosis and hypertrophic scarring of the skin. The axilla is the most common site of disease, followed by the anogenital region. Clinical features include recurrent abscesses, chronic draining sinuses, and indurated, scarred skin and subcuta-

A

B

C

Figure 125-11.  Pathogenesis of hidradenitis suppurativa, an inflammatory disease of the apocrine sweat glands and adjacent connective tissue. A, The initiating event is occlusion of the apocrine duct by a keratinous plug. B, Bacteria are trapped beneath the keratinous plug and multiply to form an abscess, which can rupture into adjacent tissue. C, The end result is recurrent abscesses, chronic draining sinuses, and indurated scarred skin and subcutaneous tissues. Often, multiple tracts are interconnected and lead to the skin.

neous tissues.124-128 The spectrum of severity ranges from mild disease with spontaneous regression to severe involvement at multiple sites. Coexistent Crohn’s disease occurs more commonly than expected, with an incidence of nearly 40% in one study.127 Diagnosis of hidradenitis suppurativa is based on the clinical presentation. A firm, pea-sized nodule may be detected and may rupture spontaneously. The lesion heals with fibrosis but can recur adjacent to the original area. Over time, multiple abscesses and sinus tracts develop in the subcutaneous region to form a honeycomb-like pattern.126 Discharge from the open areas may be serous or purulent. Treatment consists initially of antibiotics active against Staphylococcus. Oral isotretinoin has been used successfully in mild cases.128 Intralesional glucocorticoids, antiandrogen therapy, and topical clindamycin129 also have been used. Surgery is required for chronic disease. Apparent cure can be achieved by extensive excision of the involved area down to the soft tissue and with wide margins. The area is then closed with a skin graft or left to granulate.130,131 New lesions can develop in untreated skin, and close follow-up is needed.

PILONIDAL DISEASE Pilonidal disease is an acquired problem that generally affects young adults after puberty. It occurs in the intergluteal cleft and may be confused with fistula-in-ano. The prevailing theory of pathogenesis is that hair in the cleft, along with desquamated epithelium, is propelled into the base of the cleft, where the barbs of the hair shafts prevent them from being expelled, thereby setting up a granulomatous reaction, which creates a sinus.132 Free hair is found in the cyst or abscess cavity and also may be seen protruding from pits in the intergluteal cleft. The diagnosis is made by identifying the abscess, which characteristically is several centimeters cephalad to the anus, or tiny openings in the intergluteal cleft over the sacrum (Fig. 125-12). In mild cases, successful treatment has been achieved simply by shaving the hair on a regular basis (usually monthly) to prevent the hair from embedding in the intergluteal cleft. Any abscess requires immediate drainage. In patients with chronic disease, more extensive surgical treat-

2273

2274

Section X  Small and Large Intestine

A

B

Figure 125-12.  Pilonidal disease. On examination, there may be multiple pits or external openings in the intergluteal cleft proximal to the anus (A). The openings often communicate with each other (B). A probe can be passed between the communicating network of tracts. Trapped loose hair is usually found in these tracts. One successful treatment option is to unroof all the tracts.

ment is needed. Such surgery usually consists of unroofing all sinus tracts, with marsupialization; excision of the area, with or without closure of the skin edges; or creation of advancement flaps, musculocutaneous flaps, or some variant of these.132-136 When extensive surgical débridement is needed, vacuum-assisted closure devices have been shown to decrease length of time until the wound is healed.137 Afterward, it is still necessary to shave the surrounding hair periodically to prevent recurrence. As with any other chronic draining fistula or sinus, squamous cell carcinoma may arise with long-standing disease.138

RECTAL FOREIGN BODY Most foreign bodies in the rectum are not the result of oral ingestion, but rather from placement into the rectum via the anus. Treatment (removal) often requires skill, trial and error, and luck to avoid a laparotomy. Most rectal foreign bodies result from sexual (erotic) stimulation or criminal assault. The history obtained from the patient is notoriously unreliable139 as to what exactly the object is and how it was placed. If the patient indicates that the object was placed as a result of sexual assault, a full sexual assault examination is needed, including perianal brushings and sampling for sperm.140 An abdominal film is useful to determine the size and location of the foreign body. If free air is seen, a laparotomy is almost unavoidable. To remove the object, relaxation and sedation often are needed. If removal cannot be accomplished in the office or emergency department with local anesthesia, the patient should be taken to the operating room for regional or general anesthesia. An important principle in removing the object via the anus is that the rectum exerts a traction vacuum effect on the foreign body. Therefore, passage of a well-lubricated Foley catheter will break the vacuum air seal and aid in removing the object. Imagination is needed at times to remove the object via the anus. Obstetric forceps and an obstetric vacuum extractor have been used with success.141 Another innovative approach is to put plaster of Paris into the foreign body (if it is a hollow object) and place a string into the plaster of

Paris. The plaster then is allowed to harden, and the string is pulled out with the foreign body.139 Having the patient in the lithotomy position allows pressure to be applied simultaneously over the lower abdomen to facilitate expulsion of the foreign body. After any foreign body has been removed from the anus, sigmoidoscopic examination is needed to rule out a perforation. If the sigmoidoscopic examination is negative, a water-soluble contrast enema study or CT scan with contrast should be done as well to exclude a perforation, even if the foreign body does not look capable of causing injury.142 Laparotomy is reserved for cases of colonic perforation and for objects that cannot be removed otherwise. After opening the abdomen, attempts still should be made to milk the object distally and expel it via the anus, by means of intra-abdominal manipulation. If this approach fails, a colotomy is done to retrieve the object. In many instances, simple closure of the colotomy is risky because of trauma from the attempts at removal and the additional trauma caused by the grinding of the object against the rectal wall. In these situations, a temporary colostomy may be needed. Following removal of the object via laparotomy, other perforations that might have occurred when the object was inserted still must be excluded.

KEY REFERENCES

Buchanan GN, Owen HA, Torkington J, et al. Long-term outcome following loose-seton technique for external sphincter preservation in complex anal fistula. Br J Surg 2004; 91:476-80. (Ref 68.) Champagne, BJ, O’Connor LM, Furguson M, et al. Efficacy of anal fistula plug in closure of cryptoglandular fistulas: long-term follow-up. Dis Colon Rectum 2006; 49:1817-21. (Ref 61.) Chang GJ, Berry JM, Jay N, et al. Surgical treatment of high-grade anal squamous intraepithelial lesions: A prospective study. Dis Colon Rectum 2002; 45:453-8. (Ref 88.) Christoforidis D, Etzioni D, Goldberg SM, et al. Treatment of complex anal fistulas with the collagen fistula plug. Dis Colon Rectum 2008; 51:1482-7. (Ref 63.) Hull TL, Fazio VW. Surgical approaches to low anovaginal fistula in Crohn’s disease. Am J Surg 1997; 173:95-8. (Ref 75.) Jayaraman S, Colquhoun PHD, Malthaner RA. Stapled hemorrhoidopexy is associated with a higher long-term recurrence rate of internal hemorrhoids compared with conventional excisional hemorrhoid surgery. Dis Colon Rectum 2007; 50:1297-1305. (Ref 33.) Jorge JMN, Wexner SD. Anatomy and physiology of the rectum and anus. Eur J Surg 1997; 163:723-31. (Ref 3.) Keighley MRB, Williams NS. Surgical anatomy. In: Keighley MRB, Williams NS, editors. Surgery of the anus, rectum and colon. 2nd ed. Philadelphia: WB Saunders; 1997. p 7. (Ref 1.) MacRae HM, McLeod RS. Comparison of hemorrhoidal treatment modalities: A meta-analysis. Dis Colon Rectum 1995; 38:687-94. (Ref 18.) McGuinness JG, Winter DC, O’Connell PR. Vacuum-assisted closure of a complex pilonidal sinus. Dis Colon Rectum 2003; 46:274. (Ref 137.) Nelson R. A systematic review of medical therapy for anal fissure. Dis Colon Rectum 2004; 47:422-31. (Ref 39.) Nigro ND, Vaitkevicius VK, Considine B Jr. Combined therapy for cancer of the anal canal: A preliminary report. Dis Colon Rectum 1974; 17:354-56. (Ref 83.) Sajid MS, Hunte S, Hippolyte S, et al. Comparison of surgical vs chemical sphincterotomy using botulinum toxin for the treatment of chronic anal fissure: A meta-analysis. Colorectal Dis 2008; 10:547-52. (Ref 47.) Sands BE, Anderson FH, Bernstein CN, et al. Infliximab maintenance therapy for fistulizing Crohn’s disease. N Engl J Med 2004; 350:87685. (Ref 73.) Senagore A, Singer M, Abcarian H. A prospective, randomized, controlled, multicenter trial comparing stapled hemorrhoidopexy and ferguson hemorrhoidectomy: One-year results. Podium presentation at the American Society of Colon and Rectal Surgeons meeting, Chicago, June 3-8, 2002. (Ref 29.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

126 Palliative Care for Patients with Gastrointestinal and Hepatic Disease V. S. Periyakoil

CHAPTER OUTLINE What Is Palliative Medicine?  2277 Hospice versus Palliative Care  2277 Exploring the Goals of Care  2278 Prognostication  2278 Key Prognostic Variables and Tools in Liver Disease  2279 Model for End-Stage Liver Disease Scores  2279 Hepatorenal Syndrome  2280 Malignant Ascites  2280 Hepatic Encephalopathy  2280 Common Themes in Palliating Gastrointestinal and Hepatic Diseases  2280

The purpose of this chapter is to review the physical and emotional care of patients with advanced, chronic, and often life-limiting illnesses. Common gastrointestinal symptoms, important aspects of setting appropriate life goals of care, and communication skills are reviewed.

WHAT IS PALLIATIVE MEDICINE? Palliative care is interdisciplinary care that aims to improve the quality of life of patients and their families facing the problems associated with life-threatening illness through preventing and relieving suffering by identifying, assessing, and treating pain and other problems, including physical, psychosocial, and spiritual ones.1 Although palliative care often is provided to patients with life-threatening illness that is not responsive to curative treatment,2 it can and should be offered concurrently with all other appropriate medical therapies including cure-oriented treatments. Palliative medicine recognizes that dying is a normal life cycle event and seeks neither to unnecessarily hasten nor postpone death, but rather to focus care on patient-defined goals as death nears.2 Most patients want to die with comfort and dignity and are able to articulate some or all of the following goals3: Freedom from pain and other distressing symptoms Having a sense of personal control over end-of-life decisions Avoiding inappropriate prolongation of the dying process Finding meaning and purpose in life Saying goodbye to friends and families

Abdominal Pain  2280 Nausea and Vomiting  2282 Anorexia and Cachexia  2282 Constipation  2284 Diarrhea  2285 Intestinal Obstruction  2285 Jaundice  2285 Ascites  2286 Hepatic Encephalopathy  2286 Gastrointestinal Bleeding  2286

Quality care near the end of life for complex physical and psychological problems cannot be provided by a single clinician. Care ideally is provided by an interdisciplinary team that includes physicians, nurses, social workers, chaplains, and bereavement counselors. The interdisciplinary palliative care team works in concert with, and does not replace, the primary medical team. Thus, the ultimate goal of palliative care is to prevent and relieve suffering and to optimize quality of life for patients and their families, regardless of the stage of the disease or the need for curative or palliative therapies. The American Academy of Hospice and Palliative Medicine provides a list of palliative medicine and hospice physicians on their website (www.aahpm.org).

HOSPICE VERSUS PALLIATIVE CARE Palliative medicine and hospice care share the same philosophy of care. Unlike palliative medicine, however, hospice care in the United States is also a financial reimbursement system for the terminally ill patients, largely defined by the Medicare Hospice Benefit (MHB). As such, hospice care has defined regulations and admission criteria. Under the MHB, patients are eligible for a specific set of services if the physician certifies that the patient has six months or less to live, if the disease follows its usual course, and if the goal of treatment is palliative rather than curative.4 MHB does not require patients to relinquish heroic life-prolonging measures, future hospitalizations, or participation in research. Prognosis is based on the attending physician’s clinical judgment regarding the normal course of the person’s illness; if the patient lives beyond the

2277

2278

Section XI  Palliative, Complementary, and Alternative Medicine Table 126-1  What Is Palliative Care? Palliative care: Provides relief from pain and other distressing symptoms Affirms life and regards dying as a normal process Intends neither to hasten nor postpone death Integrates the psychological and spiritual aspects of patient care Offers a support system to help patients live as actively as possible until death Offers a support system to help the family cope during the patient’s illness and in their own bereavement Uses a team approach to address the needs of patients and their families, including bereavement counseling, if indicated Enhances quality of life, and may also positively influence the course of illness Is applicable early in the course of illness, in conjunction with other therapies that are intended to prolong life, such as chemotherapy or radiation therapy, and includes those investigations needed to better understand and manage distressing clinical complications Adapted from World Health Organization, 2008.

expected six months, the hospice care may be continued by recertifying the patient under MHB. Although hospice care can be provided in special residential facilities or in longterm care facilities, currently most hospice care in the United States is provided in the patient’s home. In contrast to hospice care, there are no specific eligibility criteria for palliative care because there is no specially defined reimbursement mechanism. The palliative care movement in the United States has developed largely in acute care hospitals and, more recently, in long-term care facilities. Therefore, all hospice care is palliative care, but not all palliative care is hospice care (Table 126-1). To truly provide seamless care of patients who are dying, communities need both palliative care and hospice services. There are 65 Hospice and Palliative Medicine fellowship programs in the country, and board certification is available through the American Board of Medical Specialties for the subspecialty of hospice and palliative medicine.5

EXPLORING THE GOALS OF CARE Exploring patient-defined goals of care is the first step in determining the most appropriate therapeutic interventions for a specific disease or symptom. An organized approach to goal setting can help the clinician and patient achieve clearly articulated goals. Goal setting is best accomplished through meeting with the patient and family or surrogate decision-maker (Table 126-2). Before a goal-setting meeting, the clinician should review the disease course, response to prior treatments, and potential for further diseasemodifying treatments and develop a realistic short- and long-term vision for the future clinical course, including a general sense of prognosis. With this in mind, the clinician can begin to review treatment options and help the patient decide which treatments are most likely to help meet his or her specific goals. All therapeutic options should be examined in light of the question, “Does the intervention match or assist with the patient’s treatment goals?”6 As the burden of decision-making increases near the end of life, it is important for physicians to understand their central role in helping patients make decisions. A model of shared decision-making, in which the physician provides guidance and recommendations, generally is preferred to a patern­

alistic approach, or, at the opposite extreme, to one in which options are presented with no guidance.7

PROGNOSTICATION Physicians’ estimates of patient survival are important to physicians as well as to patients and families in all phases of a patient’s life because they influence both medical and nonmedical decisions. This is especially true at the end of life when patients and families have to make numerous personal, fiscal, and social arrangements in anticipation of impending death. Despite the great need for accurate prognoses, prognostication remains an elusive clinical art. Numerous empirical studies have revealed a consistent optimistic bias,8 with most physicians overestimating anticipated life span by a factor of three. Physicians’ prognostic radars are further skewed when they are more connected with the patient. Consequently, patients who are in great need of quality palliative care are referred too late to palliative care services and thus are deprived of access to good symptom management. There are two key tasks in prognosticating: 1) formulating the prognosis by reviewing patient-specific medical factors and disease-specific actuarial estimation of survival; and 2) communicating the prognosis to patients and families while providing ongoing support. Given that prognostication is challenging, many phy­ sicians are uncomfortable with this task and often avoid providing realistic prognostic information, or they continue to offer futile treatment options that can convey false hope and unrealistic expectations to patients and families. When pressed by patients, physicians then typically overestimate survival. If the imminence of death is not discussed honestly, patients might be more likely to accept costly, burdensome, and futile treatments.9 Prognostication guidelines are well established for cancer.10,11 The single best prognostic variable in cancer is performance status.10,11 For example, patients with a Karnofsky Performance Status (KPS) of 40 (disabled; requires special care and assistance) live on average less than 50 days; and patients with a KPS of 20 live an average of only 10 to 20 days (Table 126-3).10 Put another way, patients who are spending more than 50% of the day resting or in bed generally have a prognosis of three months or less. Specific symptoms provide further information: symptoms with an independent predictive value for a poor prognosis are shortness of breath, anorexia, difficulty swallowing, and weight loss.10 Prognostic criteria for noncancer diagnoses have been published and are especially useful to help physicians know when to refer patients for hospice services.12 Specific to gastroenterology, for example, general criteria have been established for chronic liver disease (Table 126-4). Beyond guidelines, some clinicians have advocated a simple test to determine when hospice services are appropriate by asking, “Would I be surprised if this patient died in the next six months?”13,14 No matter what type of cancer or noncancer fatal illness a person has, a “common final clinical pathway” occurs in most patients.10 Signs and symptoms that predict a prognosis of hours to days are decreased or fluctuating levels of consciousness, a precipitous clinical decline, decreased oral intake, and inability to turn over in bed.11 Patients close to death typically exhibit periods of apnea, retained oro­ pharyngeal secretions (the death rattle), fever, and cool or mottled extremities.11

Chapter 126  Palliative Care for Patients with Gastrointestinal and Hepatic Disease Table 126-2  Process Steps for a Goal-Setting Family Meeting 1. Determine the reasons for convening a family conference. Clarify the goals of care and review the patient’s medical condition. Decide on future levels of care and treatments, and resolve family conflicts. 2. Determine who should and who will be present for the conference. Include appropriate health care providers (e.g., nurse, chaplain, social worker, physician consultants, primary care physician). Ask the patient, or health care power of attorney, who he or she would like to participate (e.g., the designated health care power of attorney, appropriate family members, clergy, lawyer, friend). Family and other participants should be given significant time (often 48 hours) to prepare for the meeting. 3. Determine whether the patient has decision-making capacity. Able to understand information about diagnosis and treatment? Able to evaluate alternatives and compare risks and benefits? Able to communicate a choice—verbally, in writing, or with a nod or gesture? 4. Choose the proper physical setting. 5. Introduce yourself, explain your relationship to the patient, and invite all participants to do the same. 6. Identify the legal decision maker, if available.* 7. Review the goals and purpose of the meeting. 8. Establish ground rules. Everyone will have the opportunity to talk. No interruptions are permitted. 9. Review the patient’s current medical condition. “What is your understanding of ____’s present condition?” or “What have you been told about ____’s condition?” Review with the patient and family the current medical condition (e.g., expected prognosis and potential treatment plans). Avoid medical jargon. Invite questions. Defer treatment decision-making until all questions about the patient’s medical status have been asked to the extent possible. 10. Family discussion guidelines include the following: When the patient can speak for himself or herself: Ask the patient what he or she is considering. Ask the patient what type of support he or she would like from family members and from the health care team. Invite discussion from other family members. When the patient cannot speak for himself or herself, Describe the goal of substituted decision-making (i.e., to speak on behalf of the patient by making those choices we believe the patient would make if he or she could speak). Ask each family member what he or she believes the patient would choose if he or she were able to speak on his or her own behalf. Ask each family member what his or her own wishes are for the patient. Allow patients and families and other important people to the patients (e.g., clergy, friends, lawyer) time alone, if they wish to talk before making a decision. If there is a clear consensus of opinion, the meeting can be concluded. If there is no consensus, see No. 11. 1 1. Follow these guidelines when there is no consensus. Ask the family to discuss the issue on their own and schedule a follow-up meeting (use time as an ally). Ask each family member on what values his or her decision is based and how the decision will affect the patient and the other family members. Review again the goals you are trying to reach: What would the patient say if he or she were able to speak? Discuss other resources to support decision-making. 1 2. Bring the conference to conclusion. Summarize the meeting for the family, including areas of agreement and disagreement. Decide if the decisions made lead to related issues that should be addressed while the family is present (e.g., “do not resuscitate” order, continuation or withdrawal of treatments, discharge planning). Provide a plan for follow-up, and offer to schedule further meetings with the family. Document a summary of the meeting in the medical record. Discuss relevant issues with all health team members. *Laws governing surrogate decision-making vary from state to state. Adapted from Weissman DE, Ambuel B. Establishing treatment goals, withdrawing treatments. In: Weissman DE, Ambuel B, Hallenback J, editors. Improving End-of-Life Care. 3rd ed. Milwaukee: Medical College of Wisconsin; 2000. p 101.

Discussing prognosis with patients and families is a key skill in palliative care. Physicians are advised to start by asking patients if they previously have been given prognostic information, if they have a sense of how much time is left, and whether they would like to discuss prognosis. If a patient indicates that he or she wishes to discuss prognostic information, provide a broad estimate, a few days to a few weeks or a few weeks to a few months, rather than, “Mr. Jones, you have only three weeks to live.”4 Once the time frame is presented, important future goals can be determined by asking, “What do you want or need to do in the time that is left?” (e.g., important events, saying goodbye to loved ones). This allows the clinician to aim the information at the level of the patient and family.11

KEY PROGNOSTIC VARIABLES AND TOOLS IN LIVER DISEASE MODEL FOR END-STAGE LIVER DISEASE SCORES

The Model for End-Stage Liver Disease (MELD) (see Chapter 95) is a numeric scale used to prioritize patients for liver transplantation.15-18 It is calculated by a formula using the serum bilirubin, international normalized ratio (INR), and the serum creatinine. The MELD has been validated for short-term and intermediate-term mortality in a heterogeneous group of patients with liver disease. Modifications of the MELD score using serum sodium values have been suggested; these might enhance its prognostic ability for patients with cirrhosis.

2279

2280

Section XI  Palliative, Complementary, and Alternative Medicine Table 126-3  Karnofsky Performance Scale %

LEVEL OF FUNCTIONAL CAPACITY

100 90

Normal, no complaints, no evidence of disease Able to carry on normal activity, minor symptoms or signs of disease Normal activity with effort, some symptoms or signs of disease Cares for self, unable to carry on normal activity or to do active work Requires occasional assistance, but is able to care for most needs Requires considerable assistance and frequent medical care Disabled, requires special care and assistance Severely disabled, hospitalization is indicated although death is not imminent Hospitalization is necessary; patient is very sick; active supportive treatment is necessary Moribund, fatal processes progressing rapidly Dead

80 70 60 50 40 30 20 10 0

From Karnofsky, DA, Burchenal, JH. The clinical evaluation of chemotherapeutic agents in cancer. In: MacLeod CM, editor. Evaluation of Chemotherapeutic Agents. New York: Columbia University Press; 1949. p 196.

Table 126-4  Hospice Eligibility Criteria for End-Stage Liver Disease The patient is not a transplant candidate and opts for comfort care and The patient has: Prothrombin time > 5 sec over control or INR > 1.5 and Serum albumin < 2.5 g/dL and one or more of the following conditions: Ascites refractory to treatment, or the patient is not compliant with treatment Spontaneous bacterial peritonitis Hepatorenal syndrome, elevated creatinine and BUN, oliguria (<400 mL/day), urine sodium concentration < 10 mEq/L, cirrhosis and ascites Hepatic encephalopathy refractory to treatment, or the patient is not compliant with treatment Recurrent variceal bleeding despite intensive therapy Supporting documentation (as applicable): Progressive malnutrition Muscle wasting with reduced strength and endurance Continued active alcoholism (>80 g ethanol/day) Hepatocellular carcinoma HBsAg positivity BUN, blood urea nitrogen; HBsAg, hepatitis B surface antigen; INR, international normalized ratio. Adapted from Standards and Accreditation Committee. Medical Guidelines for Determining Prognosis in Selected Noncancer Diseases. 2nd ed. Arlington, Va: National Hospice Organization; 1996.

HEPATORENAL SYNDROME

Hepatorenal syndrome (see Chapter 92) refers to functional renal failure in end-stage liver disease and is classified into two types, each with a different prognosis. Type 1 hepatorenal syndrome exhibits a rapidly progressive reduction of renal function, manifests clinically as acute renal failure, and patients have a median survival of two weeks. Type 2 hepatorenal syndrome exhibits moderate renal failure with steady or slowly progressive renal failure. These patients present with refractory ascites and have a median survival of four to six months.

MALIGNANT ASCITES

Only 10% patients who have ascites have malignancy as the primary cause; epithelial malignancies (ovary, breast, endo-

metrial, colon, gastric, and pancreatic) cause 80% of cases. Ascites is an indicator of poor prognosis, and patients with malignant ascites have a median survival of four months. If the malignant ascites and the underlying malignancy are chemoresponsive, survival may be prolonged. Ascites is very distressing to patients and should be managed aggressively to optimize patients’ quality of life (see Chapter 91).

HEPATIC ENCEPHALOPATHY

Survival in patients with hepatic encephalopathy (see Chapter 92) is very short: approximately 20% to 40% at one year and 15% at three years of follow-up.

COMMON THEMES IN PALLIATING GASTROINTESTINAL AND HEPATIC DISEASES A complete review of symptoms commonly experienced by patients being cared for by palliative medicine physicians is beyond the scope of this chapter. Some of the more common symptoms experienced by patients with advanced gastrointestinal and hepatic disease, and common gastrointestinal symptoms experienced by all patients at end of life, are discussed briefly in the following sections. As with all interventions, the benefit and burden of each treatment need to be evaluated in light of each patient’s goals of care. If an intervention does not help advance a patient’s individual goals of care, it should be withheld or withdrawn.14

ABDOMINAL PAIN

Freedom from pain is a central goal of palliative medicine. Abdominal pain can occur as one of three types: somatic, visceral, or neuropathic.

Diagnosis

Somatic pain typically is described as dull and achy; abdominal wall stretching from ascites is a common example. Patients describe visceral abdominal pain as poorly localized: diffuse, such as with peritonitis; referred, such as pain referred to the shoulder from hepatic disease; or colicky, such as with bowel or bile duct obstruction.19 Neuropathic abdominal pain arises from direct damage to peripheral or autonomic nerves near the spinal cord, the celiac or lumbar plexus, or more peripheral nerves. Neuropathic pain usually is described as tingling, burning, throbbing, sharp, or lancinating; abdominal pain from pancreatic cancer invading the celiac plexus is a common example.19 It is important to make the distinction between pain types because treatments for each type of pain may be different. A thorough pain assessment is the first step. Questions to be asked include location, duration, temporal pattern, and pain modifiers.20 Patients should be asked to rate the pain on a numeric or visual analog scale or from mild to severe (Fig. 126-1). Other important questions are drug and nondrug treatments that have been used, including their efficacy and toxicity. Finally, clinicians should assess the impact of the pain on the patient’s activities of daily living, the patient’s understanding of the cause of pain, and the patient’s specific goals for pain relief (e.g., improved sleep).21

Treatment

Nonpharmacologic therapies (e.g., relaxation exercises, imagery, and judicious use of heat and cold) and over-thecounter medications (e.g., nonsteroidal anti-inflammatory drugs [NSAIDs] and acetaminophen) often are adequate for mild pain. Opioids are the drugs of choice for moderate to

Chapter 126  Palliative Care for Patients with Gastrointestinal and Hepatic Disease severe pain.19 Short-acting opioids taken orally (morphine sulfate immediate release, oxycodone, hydrocodone, and hydromorphone) reach their peak effect in 60 to 90 minutes and provide analgesic benefit for two to four hours (Fig. 126-2, Table 126-5).19 There is no ceiling dose of opioids, but combination products containing acetaminophen, aspirin, or NSAIDs have a ceiling based on the toxicity profile of the nonopioid.19 Long-acting oral opioid preparations (morphine or oxycodone continuous release) provide 8 to 12 hours of analgesia.19 Transdermal fentanyl offers 72 hours of pain relief and is a good option for patients with stable opioid requirements who cannot take oral medications.21 Transdermal fentanyl takes 12 to 24 hours to reach full effect, and analgesia and toxicities last for up to 24 hours after removal of a patch.21

A small percentage of patients need to replace the patch before 72 hours; cases of withdrawal occurring between 48 and 72 hours have been reported.18,22 All patients receiving a long-acting opioid preparation also need a short-acting opioid for breakthrough pain. One method for calculating the size of the dose for breakthrough pain is to take 10% to 20% of the 24-hour daily opioid dose, given every 1 to 4 hours as needed.19 For example, if a patient is using 120  mg a day of long-acting morphine, an appropriate breakthrough dose would be morphine sulfate immediate release, 15 mg orally every 1 to 4 hours as needed.

By mouth

When giving pain medications, the oral route is always preferred over other routes, such as transdermal, intravenous, or subcutaneous routes

By the clock

Chronic basal pain usually is best treated with scheduled long-acting pain medications, plus short-acting pain medications on an as-needed basis for incidental or breakthrough pain

Simple Descriptive Pain Intensity Scale* Mild pain

No pain

Moderate pain

Worst possible pain

Very severe pain

Severe pain

0–10 Numeric Pain Intensity Scale* By the ladder 0 No pain

1

2

3

4

6 5 Moderate pain

7

8

9

10 Worst possible pain

IN

CR

EA

N SI

2

Visual Analog Scale (VAS)† No pain

Pain as bad as it could possibly be

* If used as a graphic rating scale, a line 10 cm long is recommended. † A line 10 cm long is recommended for VAS scales. Figure 126-1.  Pain intensity scales. (From Agency for Health Care Policy and Research. Management of cancer pain: Adults. Rockville, Md: US Department of Health and Human Services. Public Health Services; 1994.)

G

PA

IN

Opioid for moderate to severe pain +Nonopioid +Adjuvant

3

Opioid for mild to moderate pain +Nonopioid +Adjuvant

Nonopioid +Adjuvant

1

Pain Ladder from the World Health Organization Figure 126-2.  Opioid therapy for patients with chronic severe pain. (From Periyakoil VS. Opioid conversion. Stanford Palliative Care Online Curriculum 2008. [cited 2009 June 23] Available at: http://endoflife. stanford.edu.)

Table 126-5  Common Opioids DRUG

ORAL ROUTE*

PARENTERAL ROUTE*

CONVERSION RATIO TO ORAL MORPHINE

Morphine sulfate Oxycodone

30 mg (oral) 20 mg (oral)

10 mg NA

Hydrocodone

20 mg (oral)

NA

Hydromorphone

7 mg (oral)

1.5 mg

Fentanyl

NA Fentanyl has very poor oral bioavailability Transmucosal forms are available 300 mg (oral)

15 µg/hr

Parenteral morphine is 3 times as potent as oral morphine Oral oxycodone is about 1.5 times more potent than oral morphine Oral hydrocodone is about 1.5 times more potent than oral morphine Oral hydromorphone is about 4-7 times as potent as oral morphine Parenteral hydromorphone is 20 times as potent as oral morphine Transdermal fentanyl is about 80 times as potent as morphine†

Meperidine‡

75 mg

Oral morphine is about 10 times more potent than oral meperidine and about twice as potent as parenteral meperidine (mg for mg)

*Doses of opioids shown here are common usual doses; however, the actual effective dose varies from patient to patient and depends on many individual factors. † Based on studies converting from morphine to fentanyl; currently, there are no observational studies converting fentanyl to morphine. ‡ Meperidine is not a recommended drug in a palliative care setting and should be avoided. If a patient with chronic pain is on meperidine, convert the patient to an equivalent analgesic dose of one of the other opioids listed in this table. NA: not applicable. From Periyakoil VS. Opioid conversion. Stanford Palliative Care Online Curriculum 2008. [cited 2009 June 23] Available at: http://endoflife.stanford.edu.

2281

2282

Section XI  Palliative, Complementary, and Alternative Medicine The oral route is preferred for opioid administration based on availability, cost, and ease of use.19 For those with difficulty swallowing, many of the short-acting oral opioids come in concentrated solutions.19 Fentanyl also comes in a candy matrix on an applicator stick that can be twirled against the buccal mucosa for absorption, and it comes in a buccal tablet; it has a quick onset and is an option for breakthrough pain.19,23 Morphine and hydromorphone can be administered rectally, in the same dosages as used orally, with analgesic benefit similar to that in the oral route.19 Most long-acting oral morphine and oxycodone preparations are too large to put into nasogastric or gastrostomy tubes, and crushing them destroys their long-acting properties. Long-acting opioid analgesia for patients with gastrostomy tubes can be provided using methadone (liquid formulation or by crushing the solid pill), using a special formulation of morphine in which the tiny pellets in the capsule are coated with a polymer that gives them extended release, or by using a fentanyl transdermal patch. Patients unable to take medications by oral, sublingual, transdermal, or rectal routes or patients in need of rapid pain control benefit from an intravenous or subcutaneous infusion.19 Dosage by the subcutaneous route is equivalent to that by the intravenous route.19 Morphine or hydromorphone can be given subcutaneously, either as periodic injections or by a continuous infusion.19 Intramuscular injections are never indicated; they are unnecessarily painful, and absorption can be erratic.19 Opioid toxicities are predictable and often resolve without treatment. Opioids can cause nausea when the drug is initiated or the dose is increased, but this almost always resolves within a few days. Nausea that occurs when patients have been on a stable dosage of opioid for a prolonged period (more than one week) rarely is related to the opioid but can occur in patients on very high doses.24 There is no proven best treatment for opioid-induced nausea; starting with a low-cost dopamine antagonist (e.g., prochlorperazine) is a reasonable first step. Pain is the antidote to the sedative effects of opioids, and respiratory depression represents the most serious endresult of opioid-induced sedation. Tolerance develops rapidly to opioid-induced sedation, and patients with severe pain can easily tolerate many grams per day of morphine or other opioids with no sedation or respiratory depression. Risk factors for respiratory depression include rapid intravenous bolus dosing, high doses in opioid-naïve patients, simultaneous use of other sedating medications, and poor hepatic or renal function. Patients with hepatic or renal insufficiency might need lower opioid doses given at longer intervals; short-acting opioids are preferred over long-acting agents in these patients. A physician’s fear of causing respiratory depression in patients who are dying is not sufficient justification for withholding opioids; pain or palliative care experts should be consulted to assist in patient care. Tolerance never develops to opioid-induced pruritus and constipation. Prophylaxis for opioid-induced constipation should begin at the onset of opioid therapy; preferred first treatment is a senna preparation. Pruritus is largely an idiosyncratic reaction, best treated by rotating opioids in hope of finding a less itch-producing product. True allergic reactions are rare and are heralded by bronchospasm. Tolerance to the analgesic effects of opioids is uncommon; in most patients, increasing pain indicates increasing pathology.25 Physical dependence is universal: all patients on chronic opioids for more than several weeks can be expected to develop some signs of opioid withdrawal if the opioid is discontinued or an antagonist is administered.26 A

difficult diagnostic dilemma can arise in patients with underlying abdominal pain, because cramping abdominal pain is a symptom of opioid withdrawal. Psychological dependence (i.e., addiction) is rare in patients who have chronic pain and no history of addiction; the hallmark of addiction is the use of drugs despite harmful effects and loss of control of drug use.26 A physician’s fear of causing addiction in a dying patient is not a rational reason to withhold opioids.26 Many patients benefit from adjuvant analgesics, defined as nonopioid drugs used to enhance the effectiveness of opioids or drugs specifically used for neuropathic pain.19 Examples include NSAIDs or corticosteroids for somatic pain23,27 and tricyclic antidepressants and antiseizure drugs for neuropathic pain.28 When oral and intravenous medications fail to provide adequate analgesia, anesthetic procedures should be explored, including autonomic blockade (e.g., celiac plexus block29) or epidural or intrathecal anesthesia.30 All patients can benefit from nonpharmacologic treatments, including relaxation exercises, imagery, and judicious use of heat or cold.31 An abbreviated list of adjuvant analgesics for neuropathic pain is provided in Table 126-6. Acute and chronic abdominal pain are covered in Chapters 10 and 11, respectively.

NAUSEA AND VOMITING

Nausea is a highly prevalent symptom24,32 and a challenging one to manage in palliative care patients. Causes of nausea and vomiting in this population are often multifactorial and due to gastrointestinal causes (gastroparesis, gastric compression, dysmotility, constipation, bowel obstruction), CNS causes (brain metastases, anxiety, vestibular dys­ function), hepatic or renal dysfunction, and medications (opioids). Cancer chemotherapy and radiation therapy also often precipitate nausea. Many palliative care experts favor a mechanistic approach33 to management based on the likely etiology and target therapy based on the specific neurotransmitter involved (Table 126-7); others recommend an empirical approach. Evidence supporting the existing consensus-based guidelines for managing nausea and vomiting in patients with advanced cancer is sparse. Welldesigned studies of the impact of standard management and novel agents on nausea and vomiting in palliative care populations are needed.34 Nausea and vomiting are covered in Chapter 14.

ANOREXIA AND CACHEXIA

Anorexia is defined as loss of appetite, especially from disease; cachexia is defined as weight loss and wasting, particularly of muscle mass, and is accompanied by the general debility that can occur with chronic illness.35 These entities often are combined into the anorexia-cachexia syndrome, a term usually referring to cancer-related anorexia and cachexia,36 although this symptom complex can occur in patients with serious illness from any cause. Anorexiacachexia syndrome is reported in up to 80% of cancer patients and represents a poorly understood neuroendocrine and metabolic disorder.36,37 Involuntary weight loss is associated with decreased functional status, decreased response to chemotherapy, and decreased survival. Although the anorexia-cachexia syndrome itself is not painful, patients and families experience emotional suffering given the cultural significance of food intake; the loss of weight and inability to eat are a daily reminder of the illness experience and impending death.37,38 The syndrome contributes to fatigue, increased complications of antineoplastic treatments, and decreased survival.

75 mg/day in divided doses (25 mg/day in elderly)

25 mg at bedtime

STARTING DOSE

Max 3 patches/day for a maximum of 12 hr

Use equianalgesic dosages for other opioids 50 mg once or twice daily

400 mg daily (100 mg four times a day); in patients older than 75 yr, 300 mg daily Maximum of 3 patches daily for a maximum of 12-18 hr

None needed

As for morphine

No maximum dose with careful titration; use longer intervals for renal or hepatic insufficiency

600 mg daily (200 mg three times a day or 300 mg twice daily); reduce if impaired renal function

3600 mg (1200 mg three times a day); reduce if impaired renal function

225 mg daily

60 mg twice daily

150 mg daily (Therapeutic range is 50-150 ng/mL serum) 300 mg/day in adults and 100-150 mg/day in elderly Therapeutic range is 50-300 ng/mL serum, but elderly can manifest toxicity at higher end of the therapeutic range

MAXIMUM DOSAGE

Increase by 50-100 mg daily in divided doses every 3-7 days as tolerated

After 1-2 week, convert total daily dose to long-acting opioid analgesic and continue short-acting med as needed As for morphine

Low risk patient: 0.3 × wt (kg) High risk patient: 0.1 × wt (kg)

50 mg 3 times a day or 75 mg twice daily

Increase by 100-300 mg three times a day every 1-7 days as tolerated Increase to 300 mg daily after 3-7 days, then by 150 mg/day every 3-7 days as tolerated

100-300 mg at bedtime, or 100-300 mg three times a day

37.5 mg once or twice daily

Increase to 60 mg/day after 1 week Increase by 75 mg each week

Increase by 25 mg daily every 3-7 days as tolerated Increase to 100 mg/day in single or divided doses

TITRATION

3 weeks

4 weeks

4-6 weeks

4-6 weeks

4 weeks

3-8 weeks for titration + 2 weeks at maximum dosage

4-6 weeks

4 weeks

6-8 weeks with at least 2 weeks at maximum tolerated dosage 6-8 weeks with at least 2 weeks at maximum tolerated dosage

DURATION OF ADEQUATE TRIAL

Adapted from Periyakoil VS, Pan XP, Lee S, Scheufler J. The American Geriatrics Society Annual Scientific Meeting, Preconference Session: Latest Updates in Hospice & Palliative Medicine: A Skill Based Hands-on Workshop for Interdisciplinary Clinicians Booklet, April 29-May 2, 2009, Chicago, IL.

Topical Lidocaine patch (5%)

Tramadol

Oxycodone, methadone, others

Opioid and Opioid Agonists Morphine

Pregabalin

Venlafaxine Calcium Channel α2δ Ligand Gabapentin

Serotonin/Norepinephrine Reuptake Inhibitors Duloxetine 30 mg/day

Tricyclic Antidepressants Nortriptyline (best tolerated in the elderly) Desipramine

MEDICATION CLASS

Table 126-6  Adjuvant Therapies for Neuropathic Pain

Chapter 126  Palliative Care for Patients with Gastrointestinal and Hepatic Disease 2283

2284

Section XI  Palliative, Complementary, and Alternative Medicine Table 126-7  Targeted Treatment of Nausea MEDIATOR/receptor

TRIGGERING EVENT

SITE INVOLVED

ANTI-EMETIC MEDICATION

Neurokinin 1 (NK1) (substance P) Dopamine (D2)

Chemotherapy

CTZ

Aprepitant

Opioids

CTZ

5-HT3

Radiotherapy

CTZ

5-HT4

Bowel obstruction, radiation enteritis, chemotherapy Anxiety

Various sites in intestine Cortex

Perchlorperazine, haloperidol, chlorpromazine, domperidone Ondansetron, granisetron, dolasetron, palonosetron High-dose metoclopramide Mirtazapine Metoclopramide Dronabinol, nabilone

Dysequilibrium Stasis Tumors, infection, inflammation Tumors, infection, inflammation

Vestibular apparatus Intestine Immune system Immune system

Cyclizine, promethazine, dimenhydrinate Metoclopramide, domperidone Megestrol, glucocorticoids Thalidomide, lenalinomide

Cannabinoid receptors: CB1 in CNS, CB2 in PNS Histamine Muscarinic acetylcholine IL-1, IL-2, TNF-α TNF-α

5-HT, 5-hydroxytryptamine (serotonin); CNS, central nervous system; CTZ, chemoreceptor trigger zone; IL, interleukin; PNS, peripheral nervous system; TNF, tumor necrosis factor. From Periyakoil VS. Opioid conversion. Stanford Palliative Care Online Curriculum 2008. [cited 2009 June 23] Available at: http://endoflife.stanford.edu.

Weight loss itself is an independent risk factor for mortality.37 Aggressive nutritional supplementation in patients with advanced malignancy provides no survival benefit, no improvement in tumor shrinkage, and only minimal decrease in toxicity from antineoplastic treatments.37 Patients and families often want to continue enteral or parenteral nutrition and hydration to provide the sense of “offering comfort” and to prevent their loved one from “starving to death.”39 Attempts to provide such hydration and nutrition, however, can make the goal of comfort harder to achieve in the dying patient. Nasogastric and gastrostomy tubes are uncomfortable at the end of life; intravenous delivery of nutrition necessitates intravenous access; restraints are commonly needed to maintain access; and such nutrition and hydration increase the patient’s fluid volume, which in patients close to death can worsen ascites, pleural effusions, and peripheral edema.40 Appetite stimulants have a limited role in patients with advanced cancer. Indications for starting an appetite stimulant include a prognosis of longer than four weeks combined with the patient’s interest in regaining appetite. No drug has shown efficacy in prolonging survival or improving quality of life in population-based research, although individual patients might gain weight and have a greater sense of wellbeing.38 The progestational agent megestrol acetate has been well studied.41 At doses of 160 to 800  mg/day, megestrol results in a greater than 5% weight gain in 20% to 30% of patients with advanced cancer, the weight gain being adipose tissue, not muscle41; if improvement is not seen in two to four weeks, discontinuation of the drug is appro­ priate.41 Glucocorticoids can lead to weight gain, but the well-known toxicities of these agents often preclude anything but their short-term use.38 Other agents currently under investigation include eicosapentaenoic acid, dro­ nabinol, testosterone, thalidomide, adenosine triphosphate, and NSAIDs.41,42

CONSTIPATION

Constipation43-47 is very common and highly distressing symptom in palliative care patients. It is reported in more than 50% of patients43-44 on a palliative care or hospice unit,

a percentage that probably underestimates its true prevalence.45 In seriously ill and dying patients, the most common precipitants of constipation are poor fluid intake, immobility, autonomic failure, and use of opioid and anticholinergic medications. More than 90% of patients on opioid medications experience constipation.47 Opioids bind to mu receptors on the smooth muscle of the bowel, suppressing peristalsis and raising anal sphincter tone. In one study, patients using transdermal fentanyl used laxatives less than patients using oral morphine, indicating that transdermal fentanyl may be less constipating than other opioids.48 Although a stool softener can help lubricate hard stool, a bowel stimulant such as senna usually is necessary as both a prophylactic drug for opioid constipation and for initial therapy of established constipation. A commonly recommended regimen in the hospice literature is to start with senna, escalating as necessary; if no bowel movement is obtained in three days, bisacodyl or an enema (phosphosoda) is recommended.46 More refractory constipation can be relieved with magnesium citrate, a sorbitol-based product, or small doses of poly­ ethylene glycol. Fiber supplements can lead to obstipation or obstruction and are not recommended for palliative care unless patients also are typically able to ingest large quantities of water. Orally administered opioid antagonists may be effective in reversing opioid-related constipation in patients for whom other therapies are not effective.45 Parenteral opioid antagonists that do not cross the blood-brain barrier are most effective in reversing the opioid effects on the intestine. Methylnaltrexone49,50 has been approved by the Food and Drug Administration (FDA) for chronic opioid-induced constipation. The methylation of naltrexone prevents it from crossing the blood-brain barrier and thereby reversing the central analgesic effects of opioids. When patients are actively dying—that is, death is expected within one to two weeks—and they have no abdominal symptoms, aggressive drug therapy to cause laxation on a regular basis is not indicated. In fact, laxatives that cause cramping or any suppositories or enemas usually increases patient discomfort and can be withheld (Table 126-8). Constipation is covered in Chapter 18.

Chapter 126  Palliative Care for Patients with Gastrointestinal and Hepatic Disease Table 126-8  Laxatives in Palliative Care CLASS

LAXATIVE

SIDE EFFECTS

RELEVANCE TO PALLIATIVE CARE

Bulk-forming agents

Bran, psyllium, calcium polycarbophil, methyl cellulose

Impaction above strictures, bloating

Emollient, softener

Docusate sodium

Rash

Osmolar agents

Sorbitol, lactulose

Abdominal bloating, flatulence

Polyethylene glycol

Nausea, bloating, cramping

Stimulant agents

Magnesium citrate, magnesium sulfate Bisacodyl, senna

Magnesium toxicity (with renal insufficiency) Colicky pain

Patients with serious illnesses are often unable to consume a large amount of fiber Because these patients often eat and drink little, fiber supplements can worsen constipation Constipation in a palliative-care setting is multifactorial and emollients per se are often ineffective in reversing the underlying malfunction Many patients dislike the sickly sweet taste, which can exacerbate their nausea Patients might be unable to drink large volumes of the constituted solution Unpleasant taste can exacerbate nausea

Chloride channel activator

Lubiprostone

Nausea

First-line laxatives in all patients taking opioids Easy to swallow; small capsule; well tolerated

From Periyakoil VS. Opioid conversion. Stanford Palliative Care Online Curriculum 2008. [cited 2009 June 23] Available at: http://endoflife.stanford.edu.

DIARRHEA

Diarrhea is less common than constipation and is reported in up to 10% of hospice patients.43 Although the differential diagnosis for diarrhea is broad, two common causes in the palliative care patient are overuse of laxative therapy and diarrhea secondary to leakage of stool around a fecal impaction.44,45 Treatment is directed at the cause of diarrhea. Octreotide51 is useful for refractory diarrhea (particularly with endocrine tumors such as VIPomas and gastrinomas), although it is quite costly. Replacement of pancreatic enzymes is helpful in cases of malabsorption following pancreatic cancer surgery. Cholestyramine (binding of bile salts) and aspirin (reduction of prostaglandin synthesis thus decreasing water and electrolyte secretion) may be beneficial in secretory diarrhea induced by radiation colitis.52 Opioids or opioid-like medications (e.g., loperamide) might offer the best symptomatic relief. Diarrhea is covered in Chapter 15.

INTESTINAL OBSTRUCTION

Malignant bowel obstruction is seen most commonly as a complication of cancers of the colon, ovary, pancreas, and stomach.53 Obstruction results from either intraluminal or extraluminal tumor. Benign causes, such as adhesions, radiation bowel damage, inflammatory bowel disease, opioid-induced impaction, or hernia also can cause obstruction.53-55 Intestinal obstruction can occur in the small or large bowel or in both sites; multifocal obstructions are common.54,55 The symptoms of intestinal obstruction are nausea, vomiting, abdominal distention, constipation, and colicky abdominal pain.56 Nasogastric suction is used to provide temporary relief; in many cases, just a few days of nasogastric suction may convert a complete obstruction to a partial one or might even resolve an obstruction.57 Surgery with a primary anastomosis, stoma (e.g., colostomy, ileostomy), or endoluminal stenting are the preferred options for managing obstruction; the morbidity from major abdominal surgery in cancer patients with bowel obstruction is high.58 Nasogastric tubes are uncomfortable and therefore not desirable for long-term use. Although a venting gastrostomy

tube may be a less-desirable option than directly bypassing the obstruction, it can provide great symptomatic relief and avoid the need for prolonged nasogastric intubation, especially in patients who have at least weeks to live and are eating for pleasure Although partly digested food passes out of the tube, patients can still enjoy tastes and experience the psychological and communal benefit of eating. Despite the advances in surgery and the growing use of endoluminal stents, such procedures are not possible for many patients.53 Medical management using drugs is effective at reducing the symptom burden of inoperable obstructions. The goals of pharmacologic therapy are reducing distressing symptoms, avoiding the need for parenteral hydration, and removing nasogastric suction. Pharmacologic treatments consist of a cocktail of drugs including opioids for abdominal pain, dopamine antagonist antiemetics such as phenothiazines or butyrophenones for nausea, and antisecretory drugs (octreotide) to reduce colicky pain and reduce intestinal secretions.59-63 A prokinetic agent such as metoclopramide can provide relief for an incomplete or functional obstruction but should be avoided in complete obstruction.57 Intestinal obstruction is covered in Chapter 119.

JAUNDICE Although jaundice64 is usually an ominous sign in patients with cancer, it might have a correctable cause. The goal of pursuing an evaluation of jaundice is to determine if there are conditions amenable to treatment in which the burdento-benefit ratio is favorable, given the underlying extent of cancer. Interventional procedures65 for biliary obstruction, such as surgical bypass of the biliary tract or endoscopic placement of biliary stents, may be useful early in the disease course when patients have a good performance status, but their utility becomes less as performance status declines and the burden of stent placement with its potential for various complications increases. Pruritus commonly is associated with jaundice. Cool temperatures, lower humidity, and topical agents like astringents, moisturizers, and steroid creams can provide relief. Both H1 and H2 antihistamines, phenothiazines,

2285

2286

Section XI  Palliative, Complementary, and Alternative Medicine and bile acid resins have been used with some effectiveness. Opioid antagonists also have been used to treat pruritus, but their systemic use reverses analgesia in many patients receiving opioids.66-70 Gabapentin and butorphanol are effective in palliating uremic pruritus, which usually does not respond to antihistamines. Rifampin has been used to palliate pruritus, but its use may be limited by its risk for causing hepatitis.Jaundice is covered in Chapter 20.

ASCITES

Between 15% and 50% of patients with cancer develop ascites, most commonly from ovarian, endometrial, breast, large bowel, stomach, and pancreatic cancers.71,72 Malignant ascites portends a poor prognosis, with a one-year survival of 40% and three-year survival of less than 10%.64 The wellknown list of nonmalignant causes of ascites, including portal vein thrombosis, congestive heart failure, nephrotic syndrome, and pancreatitis, should be considered before assuming malignancy is the reason for the patient’s ascites. Symptoms of ascites include an increase in abdominal girth, bloating, abdominal wall pain, nausea, anorexia, and dyspnea. Paracentesis should be performed to evaluate new-onset ascites if the intervention will lead to a change in therapy or if the paracentesis will provide symptomatic benefit. Repeated therapeutic paracentesis, with or without an indwelling peritoneal catheter, is the most commonly used invasive treatment for malignant ascites.64 Peritoneovenous shunts73 have a better success rate in patients with nonmalignant ascites than in patients with malignant ascites but are infrequently used today. Octreotide has been reported to provide symptomatic benefit in malignant ascites, but its cost may be prohibitive.74,75 Ascites is covered in Chapter 91.

HEPATIC ENCEPHALOPATHY

Hepatic encephalopathy is associated with poor prognosis. Two studies76-77 have shown that the cumulative survival in these patients was very short: approximately 20% to 40% at one year and 15% at three years of follow-up. Hepatic encephalopathy due to hepatic metastases is uncommon unless there is an overwhelming liver tumor burden and is usually a terminal event. Aggressive attempts at reversal usually are futile except for a very short-term benefit, which for some patients may be needed to help bring family closure. When death is near, clinicians should use opioid analgesics and other sedating medication liberally for control of distressing symptoms, even if such treatments worsen the encephalopathy. Family and staff counseling at this time are important to ensure that all parties share the same goals of care. Hepatic encephalopathy is covered in Chapter 92.

GASTROINTESTINAL BLEEDING

Few events are as traumatic for families caring for their loved one as observing a massive gastrointestinal hemorrhage.78 If gastrointestinal bleeding is considered a likely future event, the key is to discuss care options and to develop a plan to provide patients and families with a sense

of control for what can seem to be an out-of-control situation. If the patient has advanced disease and is dying, the focus should be toward comfort rather than toward diagnostic and therapeutic interventions. Having dark-colored towels and sheets available to camouflage the bleeding is helpful, along with a rapidly acting sedating medication for emergency use. Some experts recommend chlorpromazine 25  mg given as a slow intravenous push or a 50  mg suppository rectally. Education and support of the family is of great importance, especially when the patient is dying at home. Gastrointestinal bleeding is covered in Chapter 19.

KEY REFERENCES

Abrahm JL. Nonpharmacologic strategies for pain and symptom management. In: Abrahm JL, editor. A Physician’s Guide to Pain and Symptom Management in Cancer Patients. Baltimore: Johns Hopkins University Press; 2000. p 247. (Ref 31.) American Academy of Hospice and Palliative Medicine. Fellowship programs. [cited 2009 June 23] Available from: http://www.aahpm .org/fellowship/index.html. (Ref 5.) Baron TH. Palliation of malignant obstructive jaundice. Gastroenterol Clin North Am 2006; 35:101-12. (Ref 65.) Bruera E, Sweeney C. Chronic nausea and vomiting. In: Berger AM, Portenoy RK, Weissman DE, editors. Principles and Practice of Palliative Care and Supportive Oncology, 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 2002. p 222. (Ref 33.) Chappell D, Rehm M, Conzen P. Methylnaltrexone for opioid-induced constipation in advanced illness. N Engl J Med 2008; 359:1071. (Ref 49.) Christakis NA, Lamont EB. Extent and determinants of error in physicians’ prognoses in terminally ill patients: Prospective cohort study. West J Med 2000; 172:310-13. (Ref 8.) Farrar JT, Portenoy RK. Neuropathic cancer pain: The role of adjuvant analgesics. Oncology (Williston Park) 2001; 15:1435-42, 45. (Ref 28.) Huo TI, Lin HC, Huo SC, et al. Comparison of four model for end-stage liver disease–based prognostic systems for cirrhosis. Liver Transpl 2008; 14:837-44. (Ref 18.) Huo TI, Wu JC, Lin HC, et al. Evaluation of the increase in model for end-stage liver disease (DeltaMELD) score over time as a prognostic predictor in patients with advanced cirrhosis: risk factor analysis and comparison with initial MELD and Child-Turcotte-Pugh score. J Hepatol 2005; 42:826-32. (Ref 17.) Kichian K, Bain VG. Jaundice, Ascites and Hepatic Encephalopathy. In: Doyle D, Hank SG, Cherny N, Calman K, editors. Oxford textbook of palliative medicine, 4th ed. Oxford: Oxford University Press; 2004. p 507. (Ref 64.) Lamont EB, Christakis NA. Prognostic disclosure to patients with cancer near the end of life. Ann Intern Med 2001; 134:1096-105. (Ref 11.) Ripamonti CI, Easson AM, Gerdes H. Management of malignant bowel obstruction. Eur J Cancer 2008; 44:1105-15. (Ref 75.) Said A, Williams J, Holden J, et al. Model for end stage liver disease score predicts mortality across a broad spectrum of liver disease. J Hepatol 2004; 40:897-903. (Ref 15.) Srikureja W, Kyulo NL, Runyon BA, et al: MELD score is a better prognostic model than Child-Turcotte-Pugh score or Discriminant Function score in patients with alcoholic hepatitis. J Hepatol 2005; 42:700-6. (Ref 16.) World Health Organization. Definition of palliative care. Geneva: World Health Organization; 2008. [cited 2009 June 23] Available from: http://www.who.int/cancer/palliative/definition/en/. (Ref 1.) Full references for this chapter can be found on www.expertconsult.com.

CHAPTE R

127 Complementary and Alternative Medicine David J. Hass

CHAPTER OUTLINE Definition and Epidemiology  2287 Types of Therapies  2287 Demography of CAM Users  2287 Rationale for Use  2288 Gastrointestinal Disorders Addressed by Complementary and Alternative Therapies  2288 Nausea and Vomiting  2288 Functional Dyspepsia  2289

DEFINITION AND EPIDEMIOLOGY Complementary and alternative medicine (CAM) is defined broadly as medical practices neither taught widely in medical schools nor generally available in U.S. hospitals.1 The prevalence of CAM therapies has increased at an exponential rate both in national and international medical communities. A study by Eisenberg and associates1 demonstrated that among the U.S. population, CAM use increased from 33.8% to 42.1% from 1990 to 1997. Estimated annual expenditures for CAM therapies are in excess of $27 billion, a sum that is equivalent to patients’ out-of-pocket expenditures for all U.S. physician-based services.1 In a study by Ganguli and colleagues, at least 50% of gastroenterology outpatients in a community setting were shown to have implemented CAM therapies to help ameliorate their symptoms.2 Given the widespread use of these modalities and the continuing trend of their increased use, an understanding of CAM therapies, including their potential risks and benefits, is necessary for the practicing gastroenterologist. A thorough knowledge of these practices allows physicians to provide comprehensive medical care and can help further a therapeutic rapport between physicians and their patients.

TYPES OF THERAPIES There are a wide variety of CAM therapies, and those most commonly employed for gastrointestinal and hepatic disease are defined in Table 127-1. Regardless of the therapy employed, the overall philosophy of CAM takes a uniform holistic approach that all disease results from disturbances at a combination of physical, psychological, social, and spiritual levels. Thus, a CAM modality is used to restore balance and to facilitate the body’s own healing responses, thereby ameliorating troublesome symptoms.3

Irritable Bowel Syndrome  2291 Inflammatory Bowel Disease  2292 Diarrhea and Constipation  2293 Liver Disease  2295 Gastrointestinal Malignancies  2297 Safety and Regulation of Complementary and Alternative Therapies  2298

The National Center for Complementary and Alternative Medicine (NCCAM) divides CAM therapies into four major domains. The first domain is mind-body medicine, which includes hypnosis, meditation, biofeedback, and cognitive behavioral therapy. Biologically based practices constitute the second domain within CAM therapy and includes substances within our natural environment that are used to strengthen and heal the human body, such as probiotics, prebiotics, and dietary supplements. Manipulative and body-based practices encompass the third domain, and involve the manipulation and movement of one or more parts of the body as a means of achieving healing; examples include massage, chiropractic manipulation, and reflexology. The final domain within CAM therapy comprises “energy medicine,” namely acupuncture, magnetic therapy, and Reiki. Two additional disciplines of note include traditional Chinese medicine (TCM) and Ayurvedic medicine. TCM has a heritage some 2000 years old and concerns itself with bringing a patient into balance through practices affecting the opposing forces of yin and yang. Ayurvedic medicine is a traditional Indian practice, also based on the premise of balance; it is a comprehensive medical discipline aimed at integrating mind, body, and spirit, in the hope of achieving contentment, prevention of disease, and good health. Practitioners use naturally occurring substances such as oils and herbs, as well as various treatments including fasting, yoga, and meditation, to achieve harmony in an individual patient.

DEMOGRAPHY OF CAM USERS Certain patients are more likely to use CAM therapies than are others. Women and whites tend to use CAM more often than do men and African Americans, respectively. Patients with higher levels of education, higher annual incomes, and

2287

2288

Section XI  Palliative, Complementary, and Alternative Medicine Table 127-1  Common Complementary and Alternative Therapies for Gastrointestinal and Hepatic Diseases Acupuncture Based on the principles of Chinese medicine, qi is energy, which circulates among organs along channels called meridians. Through placement of needles at specifically defined locations (points), the flow of qi is restored to appropriate levels and the health of specific organs is improved. Ayurveda Holistic system of medicine from India that provides diet and lifestyle recommendations to improve overall health. Colonic Irrigation Therapy Cleansing of the colon through various oral and enema preparations to improve “digestive health.” Herbal Medicine Ingestion of various herbal therapies, supplements, or probiotics to improve physiologic function. Homeopathy Based on the principle “like should be cured with like.” Administration of a diluted solution that, when given to a healthy person in an undiluted form, causes symptoms identical to those experienced by the ill person. Hypnosis Induction of a deeply relaxed state during which therapeutic suggestions are made to alter behavior and enhance relief of symptoms. Meditation, Relaxation A process of reflection and contemplation allowing one to focus thoughts to help alleviate symptoms. Reflexology Areas on the feet correspond to organs of the body. Massage and pressure applied to these regions can improve symptoms throughout the body.

comorbid medical conditions also are more likely to use CAM therapies.1,2 Knowledge of these demographics assists the gastroenterologist in determining which patients are likely to be using these therapies, but it is also important to understand each patient’s rationale and motivation for choosing a particular therapeutic modality.

RATIONALE FOR USE Digestive disorders rank among the most common disease states for which people seek the advice of complementary practitioners. The attraction of CAM therapies is multifaceted. First, they provide patients who might not have a medical background with a sense of control over their own bodies and health. Second, they provide patients with therapeutic alternatives when conventional medical therapies have failed to alleviate their symptoms or cure diseases such as terminal cancer. Lastly, complementary therapies are attractive to patients who feel dissatisfied with the ways their physicians demonstrate understanding of their illnesses or handle their complaints.

GASTROINTESTINAL DISORDERS ADDRESSED BY COMPLEMENTARY AND ALTERNATIVE THERAPIES This chapter focuses on the seven areas in gastroenterology and hepatology that are addressed most often by CAM therapies: nausea and vomiting; functional dyspepsia; irritable

bowel syndrome (IBS); inflammatory bowel disease (IBD); diarrhea and constipation; liver disease, specifically hepatitis B and C and alcohol-induced liver injury; and gastrointestinal malignancies. For each area, the data supporting the most commonly used CAM modalities are reviewed, along with their potential benefits and adverse effects.

NAUSEA AND VOMITING

Nausea and vomiting have a wide array of causes ranging from viral gastroenteritis to pregnancy. These symptoms can be quite distressing, and patients often resort to CAM therapies to seek symptomatic improvement. In one study of pregnant women with nausea and vomiting, 61% reported using CAM therapies for relief.4 Several complementary modalities have been used to help ameliorate nausea and vomiting, ranging from herbal medicines to relaxation techniques (Table 127-2).

Ginger

Ginger (Zingiber officinale) is the herbal supplement most commonly employed to relieve nausea and vomiting and derives its name from the Sanskrit word for “horn,” which describes the twisted, gnarled shape of its roots. Several mechanisms have been postulated to explain the antiemetic effect of ginger. Animal studies have demonstrated that one component of the herb, 6-gingerol, improves gastrointestinal motility,5 and another component, galanolactone, is a 5-hydroxytryptamine (HT)3 antagonist,6 similar to ondansetron, an antiemetic agent used to treat chemotherapyinduced nausea and vomiting. The antiemetic effect of ginger has been studied in various clinical conditions, including morning sickness, seasickness, chemotherapy-induced nausea, and postoperative nausea. Although no more effective than placebo for preventing experimentally induced motion sickness, ginger has been documented to reduce vertigo induced by caloric stimulation of the vestibular apparatus within the inner ear.7-9 In a systematic review of randomized clinical trials evaluating the efficacy of ginger for nausea and vomiting, Ernst and Pittler demonstrated that ginger is superior to placebo and equal in efficacy to metoclopramide for postoperative nausea and emesis. Furthermore, ginger relieved symptoms better than did placebo agents for the treatment of seasickness, morning sickness of pregnancy, and chemotherapy-induced nausea and vomiting.10 The dose of ginger prescribed in most of these studies ranged from 0.5 to 1 g/day. Although ginger appears to be a natural supplement with therapeutic effect, its potential adverse reactions must be taken into consideration before advocating it for relief of symptoms. First, ginger has been shown to inhibit platelet aggregation by inhibiting thromboxane synthase. Therefore, if patients are taking warfarin, aspirin, nonsteroidal antiinflammatory drugs (NSAIDs), or clopidogrel concurrently, the risk of bleeding is increased.11 Second, although not proved in animal studies, ginger has been documented to be potentially mutagenic in laboratory assays, thereby raising questions about the safety of the herbal supplement in pregnancy.12,13

Pyridoxine (Vitamin B6)

The water-soluble vitamin pyridoxine is another CAM therapy used to relieve the nausea and vomiting associated with pregnancy,14 and it was one of the most commonly employed CAM agents in a survey of pregnant Canadian women with nausea and vomiting, 29% of whom reported using it.4 Vutyavanich and colleagues reported a significant reduction in nausea, but no statistically significant reduc-

Chapter 127  Complementary and Alternative Medicine Table 127-2  Complementary and Alternative Therapies for Nausea and Vomiting THERAPY

PROPOSED MECHANISM OF ACTION

Ginger (Zingiber officinale)

Enhances GI motility; 5-HT3 receptor antagonist5,6

Pyridoxine (vitamin B6)

Unclear

Acupuncture

Placement of needles at specifically defined locations (points) to restore the flow of qi and improve the health of specific organs Process of reflection and contemplation allows one to focus thoughts to help alleviate symptoms

Relaxation therapy

EVIDENCE

ADVERSE EFFECTS

Efficacy over placebo in RCTs for postoperative and chemotherapy-induced nausea, morning sickness, seasickness10 RCTs reveal mixed results in the treatment of morning sickness15

Inhibits thromboxane synthase, thereby causing an increased risk of bleeding with concurrent antithrombotic or antiplatelet agents11 Questionable safety in pregnancy Decreases serum levels of levodopa, phenobarbital, and phenytoin16 Allergic reactions Taken in excess (>250 mg/day), can cause peripheral neuropathy, dermatoses, photosensitivity, and dizziness17 Infectious complications, perforations of internal organs, and spinal cord injury24-32

RCTs demonstrate efficacy for relief of symptoms following chemotherapy, surgery, and morning sickness20-23 Effective as an adjunctive therapy to standard antiemetic agents for chemotherapy-induced symptoms33

None reported

GI, gastrointestinal; 5-HT, 5-hydroxytryptamine; RCT, randomized controlled trial.

tion in vomiting, with pyridoxine, 30 mg daily, in a randomized controlled trial (RCT).15 Although the mechanism of action of pyridoxine is not established, certain drug interactions and adverse effects have been noted. Pyridoxine has been documented to decrease serum levels of levodopa, phenobarbital, and phenytoin when administered with these agents.16 Allergic reactions to pyridoxine also have been documented, and when taken in excess (more than 250 mg/day), pyridoxine has been reported to cause peripheral neuropathy, dermatoses, photosensitivity, and dizziness.17,18

Acupuncture and Acupressure

Acupuncture is another CAM modality commonly used to treat nausea and vomiting. In Chinese subjects, the P6 acupuncture point stimulated for relief of these symptoms is named neiguan, meaning “medial pass.” This acupuncture point is anatomically located three fingerbreadths above the proximal palmar crease on the volar aspect of the wrist in the midline. To date, more than 30 published trials have evaluated the role of stimulating the P6 acupuncture point for relief of nausea and vomiting.19 In a systematic review, acupuncture was demonstrated to be superior to placebo in ameliorating nausea and vomiting; the results were consistent despite numerous investigators, diverse patient populations, and various forms of acupuncture point stimulation.18 Trials have demonstrated that acupuncture effectively relieved nausea and emesis associated with chemotherapy,20 surgery,21,22 and pregnancy.23 The data supporting use of acupuncture and acupressure are impressive, however, gastroenterologists must recognize the difficulties of applying the traditional placeboRCT methodology to test the efficacy of acupuncture. The nature of this complementary modality is such that each patient’s regimen is individualized for relief of his or her specific symptoms, thereby precluding standardization of the treatment and calling into question the validity of the studies.

Several adverse events have been noted with acupuncture, mainly from infection secondary to improper handling of needles or their reuse without sterilization.24 Such infections have included hepatitis B virus (HBV), hepatitis C virus (HCV), and human immunodeficiency virus (HIV)25-27; bacterial endocarditis secondary to Propionibacterium acnes28; and bacteremia from Staphylococcus aureus and Pseudomonas aeruginosa with a consequent psoas abscess.29 Two fatalities have been documented in which acupuncture was thought to have led to sepsis with staphylococcal organisms.30 Although improperly sterilized needles seem to be the only risk factor for the aforementioned infections, it is difficult to prove such infections were a direct result of acupuncture, because patients might not have divulged other personal potential risk factors such as sexual preference or intravenous drug use.31 Other risks reported to be associated with acupuncture therapy include perforation of an organ during placement of the needles with resultant pneumothorax, hemopericardium with tamponade, and spinal cord injury.24,31,32

Relaxation Therapy

Relaxation therapies have been suggested as a CAM therapy for chemotherapy-induced nausea and vomiting. It has been reported that side effects related to chemotherapy are somewhat conditioned and are developed as a form of associative learning33; the anxiety experienced during chemotherapy sessions can serve as conditioning cues that lead to phys­ iologic reactions. Through progressive muscle relaxation therapy, a patient’s anxiety can be alleviated and physical symptoms averted. Relaxation therapies often are used as an adjunct to standard antiemetic medications.31

FUNCTIONAL DYSPEPSIA

Functional dyspepsia is defined as pain or discomfort in the epigastric area in the absence of demonstrable structural or physiologic abnormalities. Because symptoms tend to be short in duration and relatively mild, dyspepsia often is

2289

2290

Section XI  Palliative, Complementary, and Alternative Medicine Table 127-3  Herbal Supplements for Functional Dyspepsia HERBAL SUPPLEMENT

PROPOSED MECHANISM OF ACTION

EVIDENCE

ADVERSE EFFECTS

Banana (Musa sapientum)

Antiulcerogenic; promotes gastric mucus secretion35,36 Selectively impairs activity of nociceptive C-type pain fibers in CNS37 Contains alkaloids, which have a spasmolytic effect on smooth muscle39

Open-label trial demonstrates efficacy36 RCT demonstrates efficacy over placebo37 RCT demonstrates efficacy over placebo38

Pruritus

Increases gastric emptying, plasma somatostatin and gastrin levels; promotes gastric relaxation34,41 Inhibits gastric smooth muscle contraction43,44

RCT demonstrates efficacy over placebo41

Shenxiahewining

Unknown

STW 5 (Iberogast)

Alters gastrointestinal motility; smooth muscle relaxant43

RCT demonstrates efficacy over control medication (unclear if control was placebo)34 RCTs demonstrate efficacy over placebo53-55

Turmeric (Curcuma longa)

Increases biliary secretion, promotes contraction of gallbladder; antispasmodic agent49

Capsaicin (Capsicum annuum) Celandine (Chelidonium majus) Liu-jun-zi-tang Peppermint and caraway

RCTs demonstrate efficacy over placebo34,45

RCT documents efficacy over placebo34

Abdominal pain, diarrhea Xerostomia, insomnia, diarrhea, fatigue, hepatotoxicity40 Interstitial pneumonitis42 Diarrhea, nausea, vomiting, allergic contact dermatitis, contact urticaria, asthma exacerbations, and atrial fibrillation46,47,48 None reported Potential for hepatotoxicity, increased bleeding risk, potentiation of sedatives and anxiolytics, altered metabolism of drugs metabolized by CYP3A4 and uridine diphosphoglucuronosyl transferase40,66-68 Nausea, vomiting, fatigue, headache

CNS, Central nervous system; CYP3A4, cytochrome P450 3A4; RCT, randomized controlled trial. Data adapted from Coon JT, Ernst E. Systematic review: herbal medicinal products for non-ulcer dyspepsia. Aliment Pharmacol Ther 2002; 16:1689-99.

self-managed34; therefore, CAM therapies clearly are appealing. Herbal therapy has been a mainstay of CAM treatments for functional dyspepsia. The most common supplement therapies for functional dyspepsia, including their active ingredients, proposed mechanisms of action, and adverse effects, are listed in Table 127-3. Banana (Musa sapientum) has been evaluated for the treatment of functional dyspepsia in prospective open trials. This supplement is thought to promote gastric mucus secretion and has been documented to have antiulcerogenic properties in animals.35 In a study by Arora and Sharma,36 treatment with banana powder resulted in a reduction in symptoms in 75% of patients in the treatment group compared with 25% of those in the placebo group (P < 0.05). Causes of organic dyspepsia were excluded through various endoscopic and laboratory methods before subjects were included in the study. The only adverse effect reported was pruritus in the treatment group.36 Capsaicin, derived from the dried fruit of Capsicum annuum (red pepper), is an herbal supplement. Its mechanism of action is selective impairment of pain (C-type) fibers, which carry pain sensation from the abdominal viscera to the central nervous system.37 In one study, 2.5 mg of red pepper powder given daily improved epigastric pain, nausea, and bloating, whereas placebo did not.37 Although abdominal pain and diarrhea occurred initially in patients treated with capsaicin, these adverse effects were selflimited and of no serious clinical consequence. Greater celandine (Chelidonium majus) was investigated in functional dyspepsia by Ritter and colleagues in a randomized, double-blind, placebo-controlled trial.38 Celan-

dine accounted for a 34% greater reduction in symptoms compared with placebo (P = 0.003).38 This agent is thought to contain a variety of alkaloids that also have a spasmolytic effect on smooth muscle.39 Despite its apparent efficacy, celandine has many adverse effects, including xerostomia, insomnia, diarrhea, and fatigue. Idiosyncratic hepatotoxicity also has been described with celandine, but it resolved without complication in most cases when the supplement was discontinued.40 Liu-jun-zi-tang, also known as TJ-43, is a Chinese herbal medicine that has been used for relief of functional dyspepsia. The agent is a combination of several extracts including Actractylodis laneae rhizoma, Ginseng radix, Pinelliae tuber, Hoelen, Zizyphi fructus, Aurantii nobilis pericar­ pium, Glycyrrhizae radix, and Zingiberis rhizoma.34 Multiple mechanisms of action have been proposed, including increased gastric emptying, increased serum levels of gastrin and somatostatin,41 and relaxation of gastric smooth muscle.42 An RCT compared TJ-43, 2.5 g three times per day, with placebo for seven days in patients with functional dyspepsia. The treatment group displayed greater reductions in epigastric fullness, reflux, and nausea compared with the group treated with placebo (P < 0.05).41 The only adverse event noted with TJ-43 was one case of druginduced interstitial pneumonitis, which resolved after therapy was discontinued.34 Peppermint (Mentha piperita) and caraway (Carum carvi) are the supplements that have been investigated most thoroughly for treating functional dyspepsia. Their proposed mechanism of action is thought to be inhibition of smooth muscle contractions by direct blockade of smooth muscle

Chapter 127  Complementary and Alternative Medicine calcium channels.43,44 Several placebo-controlled trials have compared variable, fixed doses of these agents ranging from 180 to 270 mg for peppermint and 100 to 150 mg daily for caraway. A statistically significant improvement in symptoms such as bloating and epigastric pain was demonstrated when treatment groups were compared with placebo groups in several trials.34,45 Adverse effects seen with these supplements include diarrhea, nausea, vomiting, allergic contact dermatitis, contact urticaria,46 asthma exacerbations, and atrial fibrillation.47,48 Shenxiahewining is a mixture of Chinese herbs, specifically Ginseng radix, Pinelliae tuber, Coptidis rhizoma, Zingiberis rhizoma exsiccatum, and Glycyrrhizae radix, in a 3 : 9 : 3 : 3 : 3 ratio.34 In an RCT performed in China, 92% of patients treated with shenxiahewining reported improvement in symptoms compared with 20% of a control group. No important adverse events were noted.34 Turmeric (Curcuma longa) is an agent that also has been documented to have therapeutic efficacy in alleviating functional dyspepsia. This agent is thought to increase biliary secretion, promote contraction of the gallbladder, and act as an antispasmodic.49 In a placebo-controlled trial performed in Thailand, turmeric (2 g/day) was found to significantly improve dyspeptic symptoms (P = 0.003).34 Another agent that has been studied for the treatment of functional dyspepsia is STW 5, also known as Iberogast. This agent is an herbal preparation composed of bitter candytuft (Iberis umbellata), chamomile (Matricaria chamomilla), peppermint, caraway fruit, licorice root (Glycyrrhiza glabra), lemon balm leaves (Melissa officina­ lis), celandine (Chelidonium majus), angelica root (Angelica archangelica), and milk thistle (Silybum marianum). In a meta-analysis performed by Melzer and colleagues,50 evaluating three randomized, double-blind, placebo-controlled trials, STW 5 in a dosage of 1 mL three times daily for four weeks was noted to be more effective than placebo with regard to improving the study patients’ most bothersome dyspeptic symptoms. Twenty-six percent of patients in the placebo group compared with 7% in the STW 5 group reported that their symptom remained “severe” or “very severe” after treatment. Specifically, STW 5 appeared more effective in providing symptomatic relief to patients with predominant epigastric pain and gastroesophageal reflux symptoms.50 von Armin and colleagues51 also have demonstrated in the largest randomized, double-blind, placebo-controlled trial evaluating STW 5 that a significantly higher percentage of patients with functional dyspepsia who are prescribed STW 5 are free from their symptoms when compared with a placebo group.51 Pilichiewicz and coworkers showed that STW 5 affects gastric motility in a region-dependent manner, inducing gastric fundic relaxation and antral contraction.52 Although no adverse events have been reported for STW 5, individual components of the preparation are known to have potential toxicities, details of which are addressed in the next section. The data on these supplement therapies suggest that some of those studied could be useful for patients with functional dyspepsia. Peppermint, caraway, and STW 5 are the most extensively evaluated to date and, given their encouraging safety profiles, warrant further study.34

IRRITABLE BOWEL SYNDROME

IBS is defined as abdominal discomfort and altered bowel function in the absence of structural and biochemical abnormalities (see Chapter 118).53 Symptoms include pain, bloating, cramping, constipation, and diarrhea. Gastroenterologists encounter this disease entity quite commonly. In a system-

atic review citing only studies that used Rome criteria for diagnosis, IBS prevalence was found to vary between 5% and 10%, with a pooled prevalence of 7%.53 Patients with IBS often are frustrated that laboratory, radiologic, and endoscopic examinations fail to reveal an “organic” source of their discomfort, and they therefore often employ CAM therapies to help ameliorate their symptoms. Many CAM therapies have been investigated for the treatment of IBS (Table 127-4). Herbal supplement therapy and the use of probiotics have been evaluated most extensively. Psyllium (Plantago isphagula) is the most commonly prescribed dietary supplement for patients with IBS. This fiber product acts as an osmotic bulking agent and decreases bowel transit time. There have been three placebo-controlled trials of psyllium use in IBS,54-56 but only one fulfilled the five Rome criteria for appropriate study methodology (randomization, concealed allocation, placebo control, double blinding, and appropriate follow-up of study patients). Two additional trials compared psyllium with “active” agents, but neither trial was of high quality.57,58 In general, the evidence that stool frequency, consistency, and ease of passage were better with psyllium than with placebo was modest. There were no statistically significant differences in side effects among psyllium, lactulose, and placebo. Although psyllium appears to be fairly harmless, allergic hypersensitivity reactions have been documented.59 Impaired absorption of certain medications taken con­ comitantly, such as lithium and carbamazepine, also has been reported.60,61 Cases of acute esophageal obstruction also have occurred with psyllium-based agents, suggesting that, in certain cases, dysphagia might preclude its use.62 Other supplemental therapies have been used for IBS. Peppermint oil, as previously discussed, is prescribed for its smooth muscle relaxant capabilities. In a meta-analysis by Pittler and Ernst,63 peppermint oil improved symptoms of IBS compared with placebo treatment. Although statistical significance was demonstrated in this study, flaws in the methodology of the trials studied preclude evidence-based acceptance of the efficacy of peppermint oil in the treatment of IBS. STW 5 also has been used to treat patients with IBS. Placebo-controlled trials have demonstrated that STW 5 improves symptoms of IBS and reduces the severity of abdominal pain.64 Multiple mechanisms of action of STW 5 are postulated; certain components are thought to alter gastrointestinal motility, and others are thought to act as smooth muscle relaxants.60 Although no adverse events have been reported for STW 5, individual components of the preparation are known to have potential toxicities. Specifically, celandine is known to be hepatotoxic at certain doses (greater than 10 mg/day).40 Chamomile is known to contain a coumarin derivative, which increases the risk of bleeding if prescribed concurrently with warfarin, aspirin, or NSAIDs. Chamomile also has been noted to potentiate the central nervous system depressant effects of benzodiazepines and barbiturates,65,66 which often are prescribed to patients with IBS. Milk thistle is known to inhibit cytochrome P-450 3A4 (CYP3A4) and uridine diphosphoglu­ curonosyl transferase and thus could alter the metabolism of many pharmacologic agents.67 Chinese herbal medicine has been used for IBS symptoms, and RCTs have demonstrated a statistically significant benefit of its use over placebo. Patients treated with Chinese herbal medicine reported improvement in their symptoms and less interference in their daily lives from IBS, with an overall improvement in their quality of life.68 Probiotics are microorganisms that promote health effects through alterations of intestinal microflora.69 Patients with

2291

2292

Section XI  Palliative, Complementary, and Alternative Medicine Table 127-4  Complementary and Alternative Therapies for Irritable Bowel Syndrome THERAPY

PROPOSED MECHANISM OF ACTION

EVIDENCE

ADVERSE EFFECTS

Acupuncture

Placement of needles at specifically defined locations (points) to restore the flow of qi and to improve the health of specific organs

Infectious complications, perforations of internal organs, spinal cord injury reported24-32

Ayurvedic medicine

Chinese herbal medicine

Holistic system of Indian medicine that provides dietary and lifestyle recommendations to improve overall health Unknown

RCT demonstrates superiority of acupuncture to sham77 Data (only in abstract form) demonstrate greater efficacy than psyllium for relief of constipation78 RCT demonstrates efficacy over placebo but with significant dropout rates in the trials80

None reported

Homeopathy

Principle of “like should be cured with like”

Hypnotherapy Peppermint oil

Gut-related imagery; patient’s thoughts are focused toward inhibition of gastric acid secretion75 Inhibits smooth muscle contraction43,44

RCT demonstrates efficacy over placebo69 Favorable trend toward efficacy, but no statistical significance reached79 RCT demonstrates efficacy over placebo75 RCTs demonstrate possible efficacy over placebo, but methodologic flaws noted in studies64

Probiotic therapy

Alters intestinal microflora

Psyllium

Osmotic bulking agent; decreases bowel transit time

RCT demonstrates efficacy over placebo70,71 RCT demonstrates efficacy for relief of constipation53

Diarrhea, nausea, vomiting, allergic contact dermatitis, contact urticaria, asthma exacerbations, and atrial fibrillation46-48 None reported

STW 5 (Iberogast)

Alters gastrointestinal motility; smooth muscle relaxant43

RCT demonstrates efficacy over placebo65

None reported

None reported None reported

Allergic reactions, impaired absorption of medications (lithium, carbamazepine), acute esophageal obstruction56-63 Potential for hepatotoxicity, increased bleeding risk, potentiation of sedatives and anxiolytics, altered metabolism of drugs metabolized by CYP3A4 and uridine diphosphoglucuronosyl transferase40,66-68

CYP3A4, cytochrome P450 3A4; RCT, randomized controlled trial.

infectious and inflammatory disease states such as pseudomembranous colitis and IBD as well as patients with IBS also have had benefit from these agents. Evidence from RCTs has demonstrated that ingestion of Lactobacillus plan­ tarum resulted in significant reductions in abdominal pain and flatulence in patients with IBS.70,71 A report of fermented milk containing Bifidobacterium animalis demonstrated that this probiotic improves abdominal distention and associated symptoms in patients with constipation-predominant IBS.72 In addition, a systematic review of probiotics demonstrated that Bifidobacterium infantis 35624 (Align) significantly improves abdominal pain, bloating, distention, and the sensation of incomplete evacuation.73 Hypnosis has been documented to have a significant therapeutic effect on symptoms of IBS. Through the use of gut-directed imagery, whereby patients imagine they are inhibiting gastric secretion, an overall symptom improvement rate of 80% has been reported.74 Clinical remission for up to three months has been documented in patients with IBS treated with hypnotherapy. Women and those younger than 50 years seem to respond well to this modality. Hypnotherapy has proved effective in the pediatric IBS population as well, one study having documented that successful treatment of functional abdominal pain or IBS was accomplished in 85% of children treated with hypno-

therapy compared with 25% of patients treated with placebo.75 Acupuncture has been shown to be superior to sham therapy and fiber supplementation in patients with IBS.76,77 Homeopathy has demonstrated a trend toward efficacy in IBS.7 Lastly, Ayurvedic medicine (see Table 127-1) also has demonstrated efficacy in relief of symptoms from IBS. The trial of Ayurvedic medicine, however, had a large dropout rate and therefore should be interpreted with caution.78

INFLAMMATORY BOWEL DISEASE

The pathophysiology of Crohn’s disease is not completely understood despite decades of research (see Chapter 111). An overactive intestinal mucosal immune system driven at least in part by a reaction to normal luminal flora is thought to be involved in the pathogenesis,79 facilitated by failure of the mucosal epithelium to serve as an effective barrier to potential dietary and environmental toxins. Given the chronic and persistent nature of Crohn’s disease, many patients turn to CAM therapies when conventional therapies fail. Probiotics are often employed by certain subgroups of patients with Crohn’s disease, those having had total proctocolectomy and creation of an ileal pouch-anal anastomosis (IPAA) claiming the greatest benefit from these agents. Pouchitis in these patients occurs with a frequency of

Chapter 127  Complementary and Alternative Medicine approximately 50% after 10 years.80 Although the cause of pouchitis remains unknown, alteration in enteric bacterial flora appears to play an important role.80 VSL #3 is a probiotic agent consisting of four strains of Lactobacillus, three strains of Bifidobacterium, and one strain of Streptococcus. VSL #3 is thought to act by increasing tissue levels of interleukin-10 and decreasing levels of proinflammatory cytokines, such as interleukin-1 and tumor necrosis factor (TNF).81 In RCTs, administration of VSL #3 reduced the frequency of pouchitis in IBD patients after IPAA, and it decreased the number of flares of pouchitis in patients known to have chronic pouchitis.77,82 In another RCT, VSL #3 appeared effective compared with placebo for preventing endoscopic recurrence of IBD.83 No adverse events have been noted with administration of VSL #3 in any of the studies evaluated. Saccharomyces boulardii is a nonpathogenic yeast originally isolated from the litchi fruit, and it is another probiotic that has been shown to decrease the relapse rate of Crohn’s disease.84 S. boulardii exerts its beneficial effect on the intestine by acting as a trophic agent on intestinal mucosa and triggering the release of immunoglobulin (Ig) A.79 In one study, clinical relapses over a six-month period were observed in 37% of patients who received mesalamine alone and in 6% of patients treated with mesalamine plus S. boulardii.79 Another supplemental therapy employed in treating IBD is fish oil. Fish oil contains high amounts of omega-3 fatty acids, which serve as precursors of less proinflammatory cytokines than do other fatty acids commonly found in many foods. In patients with ulcerative colitis (UC) who have frequent disease exacerbations, disease activity scores were improved to a greater extent in patients who received fish oil than in those who consumed other forms of fat85; however, trials have failed to demonstrate that fish oil is effective in maintaining remission of UC.86 Clinical trials evaluating the efficacy of fish oil in the treatment of Crohn’s disease have been disappointing, because no trial has documented fish oil to be effective in maintaining disease remission or lowering an individual patient’s Crohn’s disease activity index (CDAI).85 Turmeric, traditionally used in Indian and Chinese herbal medicine, is a spice that originates from the root Curcuma longa and is a member of the ginger family. This is another complementary agent that has been reported to be effective in the treatment of IBD. A pure curcumin preparation was administered in an open-label study to patients with ulcerative proctitis or Crohn’s disease. All proctitis patients improved, with reductions in concomitant medications in four patients. Four of five patients with Crohn’s disease had lowered CDAI scores and erythrocyte sedimentation rates. This encouraging study warrants further evaluation of this agent with double-blind, placebo-controlled follow-up studies.87 The mechanism of action of turmeric is multifactorial. First, it protects lipids from peroxidation and thereby prevents the formation of free radical species. In addition, it inhibits lipopolysaccharide-induced nitric oxide synthase (iNOS) gene expression, thereby decreasing TNF-α and IL-1β production, as well as inhibiting nuclear factor (NF)-κB activation, cytokines thought to be integral to the pathophysiology of IBD. Lastly, turmeric is thought to inhibit the synthesis of proinflammatory prostaglandins and leukotrienes through inhibiting arachidonic acid uptake by macrophages.88 In safety and toxicity studies of turmeric, curcumin has been found in most studies to be safe even in high doses. In addition, in animal studies, there has been no evidence of mutagenicity or chromosomal damage.89 Turmeric does exhibit an inhibitory effect on platelet aggregation, and therefore,

patients who are maintained on agents such as aspirin, clopidogrel, or ticlopidine should be closely monitored.89

DIARRHEA AND CONSTIPATION

Altered bowel habits often lead to the use of CAM therapies. Within the discipline of CAM, practitioners often group the symptoms of diarrhea and constipation together under the term colonic health. Although several CAM modalities have been reported to promote and improve colonic health, herbal supplements are considered the mainstay of treatment. These supplements range from anthraquinone-based stimulant laxatives, such as aloe (Aloe barbadensis) to osmotic laxatives, such as magnesium citrate, to extracts of papaya (Carica papaya) and raspberry (Rubus udaeus). Those most commonly employed, including their proposed mechanisms of action, possible medication interactions, and reported adverse events, are listed in Table 127-5. As discussed later, because the supplement industry is not regulated, the content and potency of many of these agents are not standardized, which should give practitioners pause before prescribing them. Probiotics also are employed often to prevent diarrhea. A meta-analysis performed by D’Souza and colleagues described the clinical efficacy of various strains of Lactoba­ cillus (Lactobacillus bulgaricus, Lactobacillus acidophilus, Lactobacillus casei, and Lactobacillus GG) and S. boulardii in the prevention of antibiotic-associated diarrhea.90 Castagliuolo and colleagues described the protective effects of S. boulardii on Clostridium difficile-induced diarrhea in humans; the mechanism of this action is the proteolytic digestion of toxin A and B molecules by a protease secreted by S. boulardii.91 Another study reported that S. boulardii stimulates intestinal IgA in response to infection with C. difficile.92 S. boulardii also has been shown to prevent relapses of pseudomembranous colitis and to maintain intestinal mucosal barrier function against enteropathogenic Escherichia coli.93,94 VSL #3 has been demonstrated in an RCT to be effective in preventing radiation-induced diarrhea.95 Ernst and colleagues have reviewed the various CAM therapies that have been used to treat constipation. Biofeedback, a treatment technique in which people are trained to use signals from their own bodies to help recognize a relaxed state, has demonstrated clear efficacy for the treatment of constipation.96 Pelvic floor dyssynergia, an often-neglected cause of chronic constipation, is a result of the inappropriate contraction or failed relaxation of the puborectalis and external anal sphincter muscles during defecation. Pelvic floor dyssynergia is considered a form of maladaptive learning,97 and biofeedback is thought to help retrain the body to alleviate symptoms. Sensory training involves simulated defecation through the use of a water-filled balloon that is inserted into the rectum and then slowly withdrawn as the patient is asked to concentrate on relaxing the muscles that are behaving inappropriately.97 Anal manometry and electromyography, which records muscle activity either from intraluminal probes or perianal surface electrodes, are alternative means by which sensory feedback can be provided to the patient with pelvic floor dyssynergia. More than 70% of adult patients with this disorder improve following biofeedback training.97 Further studies are needed, however, to assess the long-term efficacy of biofeedback. Abdominal massage therapy has shown mixed results in the treatment of constipation. Ernst98 reviewed the data from an RCT for abdominal massage as a treatment for constipation and found that, although some data suggested a significant increase in the number of days with bowel movements,

2293

2294

Section XI  Palliative, Complementary, and Alternative Medicine Table 127-5  Herbal Supplements Claimed to Promote “Colonic Health” SUPPLEMENT

PROPOSED MECHANISM OF ACTION

Aloe (Aloe barbadensis)

Anthracene stimulant laxative; increases bowel motility

Apple pectin

Dietary fiber; binds bile acids (unknown mechanism) Anthracene stimulant laxative; increases bowel motility

Cascara sagrada (Rhamnus purshianus) Chamomile (Matricaria chamomilla)

Unknown mechanism

Clove (Syzygium aromaticum)

Spasmolytic and local anesthetic Promotes immune function through increased cytokine activity, increasing T-helper cell population Stimulates gastrointestinal motility; antispasmodic Unknown mechanism

Echinacea

Fennel (Foeniculum vulgare) Fenugreek (Trigonella foenum-graecum)

Ginger (Zingiber officinale)

POSSIBLE INTERACTIONS

ADVERSE EVENTS

Potentiation of cardiac glycosides; reduced action of glucocorticoids None reported

Gastrointestinal spasm; bloating; hypokalemia; arrhythmias; pseudomelanosis coli None reported

Potentiation of cardiac glycosides; reduced action of glucocorticoids Increased risk of bleeding with concurrent warfarin, aspirin and other NSAIDs; potentiation of CNS depressant medications None reported

Gastrointestinal spasm; bloating; hypokalemia; arrhythmias; pseudomelanosis coli Increased risk of bleeding; lethargy with potentiation of CNS depressants

Acetaminophen (secondary to glutathione depletion)

Allergic reactions have been reported; potential hepatotoxicity if administered with acetaminophen Allergic reactions have been reported Increased risk of bleeding; galactorrhea (secondary to interaction at dopamine receptors) Increased risk of bleeding

None reported Increased risk of bleeding with concurrent warfarin, aspirin and other NSAIDs

Enhances GI motility; 5-HT3 receptor antagonist; inhibits platelet aggregation by inhibiting thromboxane synthase Laxative (unknown mechanism) Osmotic laxative

Increased risk of bleeding with concurrent warfarin, aspirin and other NSAIDs, clopidogrel

Immune stimulant, anti-inflammatory Decreases cholesterol and prostaglandin synthesis Source of vitamin C; stimulates immune system; stimulates intestinal peristalsis Promotes healing of gastrointestinal ulceration and improves pancreatic function Osmotic bulking agent; decreases bowel transit time; generates short-chain fatty acids

Can delay absorption of other medications None reported

Raspberry (Rubus idaeus) Rhubarb (Rheum palmatum)

Unknown Anthracene stimulant laxative; increases bowel motility

Senna (Cassia senna)

Anthracene stimulant laxative; increases bowel motility

Spirulina Valerian (Valeriana officinalis)

Green algae, contains Mg Spasmolytic agent

Yellow dock (Rumex crispis)

Laxative (unknown mechanism)

None reported Potentiation of cardiac glycosides; reduced action of glucocorticoids Potentiation of cardiac glycosides; reduced action of glucocorticoids None reported Interacts at GABA receptor: potentiation of CNS depressant medications None reported

Hibiscus (Hibiscus sabdariffa) Magnesium citrate Marshmallow (Althaea officinalis) Oat bran (Avena sativa) Oregon grape (Berberis vulgaris)

Papaya (Carica papaya)

Psyllium (Plantago isphagula)

None reported

None reported

None reported

None reported

Hypermagnesemia: prolonged QT interval on ECG, hypotension, hyporeflexia None reported None reported

None reported

None reported

Increased risk of bleeding with concurrent warfarin, aspirin and other NSAIDs

Increased risk of bleeding; decreased testicular weight; interrupted estrous cycle in mice

Impaired absorption of medications

Allergic reactions have been reported; impaired vitamin B12, lithium, carbamazepine absorption; esophageal obstruction None reported Gastrointestinal spasm; bloating; hypokalemia; arrhythmias; pseudomelanosis coli Gastrointestinal spasm; bloating; hypokalemia; arrhythmias; pseudomelanosis coli None reported Lethargy; withdrawal from valerian is similar to withdrawal from benzodiazepines None reported

CNS, central nervous system; ECG, electrocardiogram; GABA, γ-aminobutyric acid; GI, gastrointestinal; 5-HT, 5-hydroxytryptamine3; NSAID, nonsteroidal anti-inflammatory drug. Derived from Hass DJ, Lewis JD. Quality of manufacturer provided information on safety and efficacy claims for dietary supplements. Pharmacoepidemiol Drug Saf 2006; 15:1-9.

Chapter 127  Complementary and Alternative Medicine Table 127-6  Complementary and Alternative Therapies for Chronic Liver Disease THERAPY

PROPOSED MECHANISM OF ACTION

Chinese herbal medicine

Antioxidant; inhibition of stellate cell activation in animal studies; increased interferon production119,120; inhibition of HBV DNA polymerase106

Licorice (Glycyrrhiza glabra)

Activates CYP450 phase I detoxification reactions; enhances endogenous interferon production; inhibits TNF-α129 Antioxidant, anti-inflammatory, possibly antifibrotic effects106-108

Milk thistle (Silybum marianum)

Picrorrhiza kurroa

Antioxidant; anti-inflammatory; antiviral effects

S-Adenosyl-l-Methionine (SAMe)

Participates in the synthesis of glutathione; acts as an antioxidant114

Thymic extracts

Unknown

EVIDENCE

ADVERSE EFFECTS

Trials suggest that combination therapy of interferon and CHM might help increase clearance of HBsAg, HBeAg, HBV DNA; methodologic flaws preclude recommendation for use without further investigation RCTs demonstrate a biochemical effect; no demonstrable morbidity or mortality benefit130,131

Hepatotoxicity; interstitial pneumonitis, autoimmune hepatitis, immune thrombocytopenic purpura

Trials detailing benefit over placebo for improving serum aminotransferase levels No data to support morbidity or mortality benefit112,114

Nausea, diarrhea, dyspepsia, headache, arthralgias, skin reactions, impotence, anaphylaxis; inhibits CYP3A4 and uridine diphosphoglucuronosyl transferase112,114 None reported

Only 1 trial details efficacy in acute HBV infection; paucity of data precludes recommendations139 Methodologic flaws in trials assessing efficacy No definitive data to support its use in alcoholic hepatitis (see Chapter 84) Trials suggest that combination therapy with interferon yields higher virologic response rate than interferon alone or placebo127

Pseudoaldosteronism effects: hypokalemia, sodium retention, hypertension; potential digitalis toxicity if taken concurrently132

Dry mouth; nausea; akathesia; reported to block platelet aggregation in vitro Nausea, vomiting, rare thrombocytopenia128

CHM, Chinese herbal medicine; CYP450, cytochrome P450; HBV, hepatitis B virus; HBeAg, hepatitis Be antigen; HBsAg, hepatitis B surface antigen; RCT, randomized, controlled trial; TNF, tumor necrosis factor.

and decrease in the number of episodes of fecal incontinence and number of enemas given, the trials were of poor quality and were methodologically flawed. The trials were not blinded and were subject to observer bias; only one study was randomized. Therefore, further RCTs are needed to determine whether massage is effective in patients with chronic constipation. Homeopathy has been suggested to have clinical efficacy in the treatment of postoperative ileus. A meta-analysis of studies of patients with ileus after abdominal and gynecologic surgery revealed that homeopathic treatment with agents such as opium poppy (Papaver somniferum L.) and chaparral (Raphanus sativus) significantly reduced the time to normal intestinal peristalsis compared with placebo treatment.99 The underlying principle of homeopathy is “like cures like,” and these supplements, in diluted doses, are thought to ameliorate slowed intestinal transit because they themselves are known to cause constipation. This metaanalysis, however, did not yield definitive conclusions, because several of the trials included were reported in publications that are not peer reviewed, thereby raising suspicion as to the quality of the data. Lastly, colonic irrigation therapy has gained popularity among patients interested in complementary therapies. “Colonics,” as they are colloquially termed, differ from enemas because they are not self-administered but rather are given by a practitioner who has a certain degree of training, via a device that controls water flow, temperature, and pressure.100 The rationale for this practice comes from the concept of “autointoxication,” a notion popularized by Sir

Arbuthnot Lane in the late 1800s and early 1900s that toxins originating in the intestine can enter the circulation and poison the body. There are no large RCTs thoroughly evaluating the efficacy of colonic irrigation therapy. Adverse events of colonic irrigations have been reported, including amebiasis and rectal perforation.101,102

LIVER DISEASE

CAM therapies are commonly used to treat conditions such as hepatitis B, hepatitis C, and alcoholic liver disease. One study of U.S. outpatients with chronic liver disease reported that 41% had used some form of CAM therapy in the preceding four weeks.103 As with the other gastrointestinal conditions already discussed, most therapies used for chronic liver disease have been herbal supplements. Table 127-6 details the commonly employed supplements, their mechanisms of action, levels of evidence to support their use, and adverse events.

Silymarin

Milk thistle (Silybum marianum), the CAM compound most commonly used for liver disease, has been employed for many disorders, including alcoholic liver disease, chronic viral hepatitis, and drug-induced hepatitis. Silymarin, the active ingredient, is derived from various parts of the milk thistle plant. Its mechanism of action is not defined fully but appears to be multifaceted. First, it is thought to act as an antioxidant that prevents glutathione depletion.104 Second, it has anti-inflammatory activity and decreases formation of leukotrienes, prostaglandins, and TNF-α.105

2295

2296

Section XI  Palliative, Complementary, and Alternative Medicine Lastly, in several animal studies, silymarin has been shown to block the proliferation of hepatic stellate cells and production of procollagen III, suggesting a role to slow fibrosis in chronic liver disease.106 Silymarin has been evaluated in several trials of alcoholic liver disease. Ferenci and colleagues,107 in an RCT of cirrhotic patients treated with 420 mg of silymarin or placebo, demonstrated an improved four-year survival in the treatment group compared with the placebo-treated group. Patients with alcoholic liver disease and early cirrhosis (Child-Turcotte-Pugh class A) were more likely to benefit than were those with Child-Turcotte-Pugh class B or C. This trial, however, did not confirm a clear benefit of silymarin, because patients were not randomized properly; the placebo group contained patients with more advanced cirrhosis (Child-Turcotte-Pugh class C) than did the treatment group. In addition, the degree of abstinence from alcohol among the study participants was not followed, and the dropout rate was high. A larger, more rigorously defined study by Pares and colleagues108 failed to demonstrate a survival benefit in alcoholic cirrhotic patients treated with 450 mg of daily silymarin compared with a group treated with placebo. The Hepatitis C Antiviral Long-term Treatment against Cirrhosis (HALT-C) Trial was designed to determine whether maintenance interferon therapy could slow disease progression in patients who had failed to eradicate hepatitis C virus (HCV) during prior interferon treatment (nonresponders). Seeff and colleagues109 examined the use and potential effects of silymarin in the HALT-C patient population. Among all participants, 67% had never used silymarin, 16% used it in the past, and 17% used it at baseline. Silymarin use varied widely with gender and ethnicity; men were more frequent users than women; non-Hispanic whites were more frequent users than African Americans and Hispanics. Silymarin use correlated strongly with higher education. No beneficial effect of silymarin was found on serum alanine aminotransferase or HCV RNA levels. Univariate analysis showed significantly fewer liver-related symptoms and better quality-of-life parameters in users than nonusers. A systematic review of RCTs of silymarin in various hepatic disease states (hepatitis B, hepatitis C, alcoholic liver disease) also drew no firm conclusions about its therapeutic efficacy. Approximately one half of the trials demonstrated a significant biochemical response to silymarin, specifically a decrease in serum aminotransferase levels; however, this response did not translate into a statistically significant mortality or morbidity benefit. Favorable trends toward a decrease in the frequency of hepatic encephalopathy and gastrointestinal bleeding were suggested by these trials, but a statistically significant difference was not reached between those treated with silymarin and those treated with placebo.110,111 Reddy and colleagues are evaluating the use of silymarin in the SyNCH trial, a randomized, multicenter, doubleblind, placebo-controlled trial funded by the National Center for Complementary and Alternative Medicine (NCCAM). This is a phase I/II trial, the objectives of which are to examine the impact of silymarin on noncirrhotic patients who are also nonresponders to traditional antiviral medications for HCV, and to evaluate its impact on patients with nonalcoholic steatohepatitis who have failed to respond to conventional treatment. It is hoped that this study will define the role of silymarin in the treatment of these difficult-to-treat clinical conditions for which conventional medicine has failed to yield therapeutic results.112 The reported adverse effects of silymarin include nausea, diarrhea, dyspepsia, headache, arthralgias, skin reactions,

impotence, and anaphylaxis. Most important, milk thistle has been shown to inhibit CYP3A4 and uridine diphosphoglucuronosyl transferase, thereby leading to interactions with traditional prescription medications such as quinine, lidocaine, certain calcium channel–blocking agents, and cyclosporine, all of which are metabolized in part by CYP3A4.110,111

S-Adenosyl-l-Methionine

S-Adenosyl-l-methionine (SAMe) acts as a methyl donor for many biochemical reactions and participates in the synthesis of glutathione, the predominant biochemical antioxidant.113 This compound has been studied best in the treatment of alcoholic liver disease. A systematic review of eight placebo-controlled trials of patients treated for alcoholic liver disease revealed that SAMe had no statistically significant effect on mortality, liver-related mortality, or rate of liver transplantation,111 and the methodologic quality of these trials was poor. Further evaluation of SAMe in more properly designed trials is needed.111 SAMe also has been evaluated in the treatment of cholestasis of pregnancy. In several controlled trials, SAMe reduced pruritus and serum bilirubin levels during pregnancy, thereby suggesting possible efficacy.114 The safety of this agent in pregnancy has been demonstrated in RCTs.115

Chinese Herbal Medicine

Chinese herbal medicine (CHM) is the most common CAM therapy employed for treating HBV, and it is the therapy that has been evaluated most rigorously. HBV is a significant global health problem.116 Given the large number of people affected and high rate of endemic HBV infection in some parts of the world such as Asia, it is not surprising that CAM therapies are often used to treat illness associated with HBV. Many different herbal combinations have been employed to treat HBV infection. For example, TJ-9, known as xiaochai-hu-tang in China, is a combination of seven herbs: Bupleurum root, Pinellia tuber, Scutellaria root, jujube fruit, ginger rhizome, ginseng root, and Glycyrrhiza root. This agent is thought to act as an antioxidant as well as an inhibitor of stellate cell fibrosis.117,118 Another example of CHM is Phyllanthus amarus, whose mechanism of action appears to be inhibition of HBV DNA polymerase.119 A systematic review of nine RCTs that evaluated CHM revealed that compared with placebo treatment, the CHM compound fuzheng jiedu tang significantly increased the rate of clearance of hepatitis B surface antigen (HBsAg), hepatitis B e antigen (HBeAg), and HBV DNA. P. amarus and kurorinone were comparable to interferon treatment in clearing these serologic markers.120 The quality of the aforementioned trials was poor, however, and thus no definitive conclusion can be reached at present regarding the efficacy of these agents for chronic HBV infection. A review of the effects of CHM on asymptomatic HBsAgpositive carriers with normal aminotransferase levels evaluated three RCTs, all of which were of poor methodologic quality. The compound Jianpi Wenshen recipe was found to have beneficial effects on clearance of HBsAg and HBeAg and on seroconversion of HBeAg to antibody to HBeAg.121 Given the flaws in the methodology of the trials evaluated, however, a recommendation for use of this agent cannot be endorsed without further investigation. A meta-analysis of 27 RCTs compared CHM alone, CHM combined with interferon, and interferon alone for chronic HBV infection.122 The absence of a strict placebo group in these trials is of concern. In China, where most CHM is used for HBV infection, CHM often is used as an adjunct or alter-

Chapter 127  Complementary and Alternative Medicine native to interferon therapy. Therefore, these trials were designed to assess the efficacy of CHM in conditions that replicate common clinical practice.118 Patients who received CHM alone were more likely to clear HBsAg than were those treated with interferon alone. CHM was equivalent in efficacy to interferon in achieving clearance of HBeAg and HBV DNA. Patients who received combined therapy were more likely than those treated with interferon alone to achieve seroconversion for HBsAg and HBeAg and to clear HBV DNA.118 Although these trials appear to favor the use of CHM as a potential adjunct therapy to interferon, most of the trials were of poor methodologic quality. In addition, most of the studies had a follow-up of only three months for assessing treatment outcomes. The studies that were reviewed were published in Chinese journals, and many details regarding blinding and randomization of the subjects in the trials were omitted from the publi­cations, raising additional concerns regarding methodologic quality.118 CHM also has been studied for the treatment of HCV infection. A systematic review of 10 randomized trials evaluated the efficacy of CHM in patients with chronic HCV infection. The results of the trials were disappointing in that none of the herbal agents employed was found to increase the rate of HCV RNA clearance. In addition, 9 of the 10 trials showed no improvement in serum aminotransferase levels.123 Adverse effects of CHM include hepatotoxicity; however, given the lack of manufacturing uniformity in content and potency of these agents, definitive causality has not been established.124 Cases of interstitial pneumonitis, autoimmune hepatitis, and acute thrombocytopenic purpura also have been reported.111

Thymic Extract

The efficacy and safety of thymic extract in treating HCV infection has been evaluated in five RCTs. Patients who received thymosin-1, a synthetic polypeptide, in combination with interferon therapy were more likely to have complete virologic response than were those patients treated with interferon alone or with placebo.125 Reported adverse events included nausea and vomiting and one case of thrombocytopenia.126

Licorice (Glycyrrhiza glabra)

Licorice (G. glabra) has been evaluated as a possible CAM therapy for chronic HCV infection. The active component of licorice, glycyrrhizin, is thought to activate cytochrome P-450 phase I detoxification reactions, stimulate endogenous interferon, and inhibit TNF-α.111,127 Several RCTs have evaluated this compound for treating HCV. Suzuki and colleagues128 demonstrated that daily injections of Stronger Neo-Minophagen C, a compound of glycyrrhizin, glycine, and cysteine, decreased serum aminotransferase levels compared with placebo. A morbidity or mortality benefit was not demonstrated. Furthermore, the follow-up period in this trial was only one month, making it extremely difficult to assess any long-term adverse effects of G. glabra. Given that this study was published in 1983, the presence of HCV was not determined in the study population; inclusion criteria merely necessitated histologic evidence of chronic hepa­ titis. It is not clear, in fact, that the study population had HCV infection. Another RCT evaluated the effects of G. glabra in patients with chronic HCV infection.129 The efficacy of ursodeoxycholic acid combined with glycyrrhizin was compared with glycyrrhizin alone. There was a statistically significant biochemical improvement in the treatment group, but the trial

lacked a placebo arm, and biochemical improvement might not have been the most clinically meaningful parameter to assess. Adverse events with G. glabra are thought to be secondary to the active metabolite of licorice root, glycyrrhizin, which inhibits 11-β-hydroxysteroid dehydrogenase. This inhi­ bition leads to a pseudoaldosterone effect, resulting in hypokalemia, sodium retention, and hypertension130; hypokalemia can increase the risk of toxicity from some drugs, such as digitalis.

Ayurvedic Medicine

Picrorrhiza kurroa is an Indian herb commonly used in traditional Ayurvedic medicine. It has been used for many gastrointestinal conditions and also is often used for hepatic disease. The active ingredients, picroside and kutkoside,131 are thought to act as antioxidants, anti-inflammatory agents, and inhibitors of proinflammatory cytokines.111 Various studies have described possible cancer chemopreventive and antiviral qualities of these agents.132,133 One trial demonstrated a beneficial biochemical effect of P. kurroa in reducing serum aminotransferase levels in patients with acute HBV infection.134 Clearly, more data are needed before any recommendation can be made regarding this agent. Liv 52 is an Indian ayurvedic medication that has been marketed specifically for the treatment of liver disease. It is composed of Capparis spinosa (capers), Cichorium intybus (wild chicory), Terminalia arjuna (arjuna), Solanum nigrum (black nightshade), Achillea millefolium (yarrow), and Tamarix gallica (tatarisk).135 This agent has demonstrated efficacy in protecting rats from carbon tetrachloride and alcohol-induced liver injury, as well as improving liver function in patients with acute viral hepatitis. This agent was withdrawn from the U.S. market, however, after an RCT demonstrated decreased survival in patients with alcoholic hepatitis compared with placebo.136

Acupuncture

Acupuncture has been demonstrated in a study by Li and colleagues to have gained popularity in patients with chronic hepatitis. Roughly 9% of patients report implementing this therapy. There are no convincing data detailing the effectiveness of acupuncture in treatment of either chronic hepatitis B or C. Li and colleagues have demonstrated that in patients with pain from hepatocellular carcin­ oma, acupuncture is useful as an adjunctive therapy for ameliorating postoperative pain.137

GASTROINTESTINAL MALIGNANCIES

Estimates are that up to 64% of adult oncology patients have employed CAM therapies at some point during their treatment.138 The motivation for using these therapies in oncology patients is similar to the rationale cited by other patient populations: CAM therapies are appealing as a result of a failure of conventional medicine to control or cure disease, and they provide patients with a mechanism for feeling in more control of their therapeutic plan. Textbooks have been dedicated to this subject, and the discussion that follows highlights the most commonly employed CAM therapies. A systematic review of the beneficial effect of green tea consumption in reducing the incidence of gastrointestinal malignancy139 demonstrated that green tea did help prevent colonic adenomatous polyp formation and chronic atrophic gastritis. No definitive supporting evidence, however, was found to conclude that green tea had a similar beneficial effect on the incidence of gastrointestinal malignancy.

2297

2298

Section XI  Palliative, Complementary, and Alternative Medicine Garlic is thought to inhibit the development of gastric cancer through several proposed mechanisms. An antibacterial effect against Helicobacter pylori has been demonstrated and is attributable to the thiosulfinate component of this agent.140 Kaempferol, a flavonol present in high concentration in garlic, also contributes to the detoxification of carcinogenic compounds.141 Published studies suggest that garlic might protect against the development of gastric and colonic carcinomas. Most of the literature, however, consists of observational studies that cannot be used to confirm a therapeutic effect of garlic. Additional therapeutic intervention trials are needed to substantiate the claim that garlic is chemopreventive.142 Vitamins C and E are antioxidants that might reduce the incidence of colorectal cancer. In an epidemiologic study of colorectal cancer patients, long-term use of vitamins C and E did not provide a mortality benefit. In a subgroup analysis, however, use of vitamin C for more than 10 years was associated with a decreased risk of death from colorectal cancer before 65 years of age and a decreased risk of rectal cancer mortality at any age.143 Other dietary factors also could play a role in preventing malignancy. In Mediterranean countries there is a lower incidence of breast, endometrial, colorectal, and prostate cancer compared with Western countries. These cancers have been postulated to have a relationship to diet, in that a low consumption of fruits and vegetables and a high consumption of red meat correlate with cancer incidence. A traditional Mediterranean diet contains low amounts of red meat and high amounts of fruits, vegetables, and olive oil. Some epidemiologists estimate by statistical modeling that up to 25% of colorectal cancer could be prevented in Western countries if diets were changed to reflect Mediterranean practices.144 Several CAM therapies have been implemented to help ameliorate pain in patients with metastatic disease. Acupuncture has shown promise for the treatment of the pain associated with gastric cancer.145 Lycopodium clavatum, a type of fern moss, has been reported to be effective as a homeopathic treatment for rectal cancer pain.146 Meditation and relaxation therapies are practiced commonly by many cancer patients, not only to ameliorate physical pain but also to help cope with the depression that commonly accompanies malignant disease.

SAFETY AND REGULATION OF COMPLEMENTARY AND ALTERNATIVE THERAPIES With the increasing popularity of CAM therapies, it is important that physicians understand their mechanisms of action as well as the data supporting their efficacy. Although some medical schools and residency programs are increasingly offering education programs in CAM therapies, curriculums should be mandated to include information on these modalities so as to familiarize medical professionals with these practices early in their careers. A study by Mikail and colleagues147 reported that only 16% of medical residents surveyed routinely asked their patients about their use of herbal therapy. It is equally important, however, that health care professionals also understand the regulatory mechanisms, or lack thereof, that are in place for these modalities, so that effective safety measures can be employed to protect the welfare of patients. Total yearly sales of herbal supplements are approximately $13.9 billion and steadily increasing in the United

States. An estimated 15 million adults take prescription medications concurrently with herbal supplements.1 Therefore, the safety of concurrent administration of herbal supplements and traditional allopathic medications is a concern to many physicians. In 1994, the U.S. Congress implemented the Dietary Supplement Health and Education Act (DSHEA). This legislation was developed to prevent the U.S. Food and Drug Administration (FDA) from regulating dietary supplements “excessively” and to ensure that safe and appropriately labeled supplements remain available to those persons who wish to use them. DSHEA officially defines a “dietary supplement” as A product (other than tobacco) that is intended to sup­ plement the diet that bears or contains one or more of the following dietary ingredients: a vitamin, a mineral, an herb or other botanical, an amino acid, a dietary substance for use by man to supplement the diet by increasing total daily intake, or a concentrate, metabolite, constituent, extract or combination of these ingredients.148 Additional FDA guidelines specify that supplement manufacturers themselves are responsible for determining the safety of their products and for providing the evidence, if asked, to substantiate the claims made by their individual products. Approval from the FDA is not required before marketing of most of these agents unless the supplement is deemed “new.” A new supplement is defined as an agent not marketed before October 15, 1994; however, no definitive list of products marketed before this date exists. Therefore, the responsibility rests with the manufacturer to determine if its product is in fact a new supplement. This rule creates an obvious conflict of interest in that new supplements require clinical research and capital expenditures to substantiate their efficacy, and manufacturers would prefer that supplement therapies be considered previously marketed. Additionally, supplement manufacturers are not required to report adverse events that occur with use of their products. It is the responsibility of the FDA to prove that products are unsafe before their use can be restricted.148 The FDA relies on physicians and other health care professionals to report suspected adverse events for an inquiry to be established for a particular agent. Therefore, it is of utmost importance that all health care professionals be aware of their patients’ use of supplements both to provide safe care and to know when to suspect adverse effects or medication interactions. Suspected adverse events or medication interactions can be reported online at http://www.fda.gov/ medwatch. Resources for health care professionals regarding CAM therapies have become much more accessible. Information regarding definitions of CAM therapies, as well as efficacy of the various modalities, can be found at the following websites: www.nccam.nih.gov www.fda.gov www.cochrane.org www.NaturalDatabase.com All health care professionals should be aware of these resources, so as to effectively and safely prescribe medications and care for patients from a comprehensive standpoint. It is hoped that through continued educational efforts at all levels of medical training, the data regarding the efficacy of CAM therapies, as well as their potential benefits and dangers, will be increasingly understood. Health care professionals and their patients will then be able to main-

Chapter 127  Complementary and Alternative Medicine tain a healthy therapeutic rapport while at the same time incorporating these therapeutic modalities, if they are clinically appropriate, in a safe and efficacious manner.

KEY REFERENCES

Brandt LJ, Bjorkman D, Fennerty MB, et al. Systematic review on the management of irritable bowel syndrome in North America. Am J Gastroenterol 2002; 97(Suppl 11):S7. (Ref 53.) Brenner DM, Moeller M, Chey WD. Schoenfeld P. The utility of probiotics in the treatment of irritable bowel syndrome: a systematic review. Am J Gastroenterol 2007; 104:1033-49. (Ref 73.) Chainani-Wu N. Safety and anti-inflammatory activity of curcumin: a component of turmeric (curcuma longa). J Alt Comp Med 2003; 9: 161-8. (Ref 89.) Coon JT, Ernst E. Systematic review: herbal medicinal products for non-ulcer dyspepsia. Aliment Pharmacol Ther 2002; 16:1689-99. (Ref 34.) Eisenberg DM, Davis RB, Ettner SL, et al. Trends in alternative medicine use in the United States, 1990-1997. JAMA 1998; 280:1569-75. (Ref 1.) Ernst E, Pittler MH. Efficacy of ginger for nausea and vomiting: a systematic review of randomized clinical trials. Br J Anaesth 2000; 84:367-71. (Ref 10.) Forster HB, Niklas H, Lutz S. Antispasmodic effects of some medicinal plants. Planta Med 1980; 40:309-19. (Ref 44.)

Holt PR, Katz S, Kirshoff R. Curcumin therapy in inflammatory bowel disease: a pilot study. Dig Dis Sci 2005; 50:2191-3. (Ref 87.) Koretz RL, Rotblatt M. Complementary and alternative medicine in gastroenterology: the good, the bad, and the ugly. Clin Gastroenterol Hepatol 2004; 2:957-67. (Ref 7.) Levy C, Seeff LD, Lindor KD. Use of herbal supplements for chronic liver disease. Clin Gastroenterol Hepatol 2004; 2:947-56. (Ref 111.) Madisch A, Holtmann G, Plein K, Hotz J. Treatment of irritable bowel syndrome with herbal preparations: results of a double-blind, randomized, placebo-controlled, multi-center trial. Aliment Pharmacol Ther 2004; 19:271-9. (Ref 64.) Melzer J, Rosch W, Reichling J, et al. Meta-analysis:phytotherapy of functional dyspepsia with the herbal drug preparation STW 5 (Iberogast). Aliment Pharmacol Ther 2004; 20:1279-87. (Ref 50.) Seeff LB, Curto TM, Szabo G, et al. Herbal product use by persons enrolled in the Hepatitis C Anti-viral Long term Treatment against Cirrhosis (HALT-C) trial. Hepatol 2008; 47:605-12. (Ref 109.) Verma S, Thuluvath PJ. Complementary and alternative medicine in hepatology: review of the evidence of efficacy. Clin Gastro Hep 2007; 5:408-16. (Ref 135.) Whorwell PJ, Prior A, Colgan SM. Hypnotherapy in severe irritable bowel syndrome: further experience. Gut 1987; 28:423-25. (Ref 74.) Full references for this chapter can be found on www.expertconsult.com.

2299

Index Note: Page numbers followed by f and t indicate figures and tables, respectively.

A

A cells, 915 A-like cells, 817-819 Abacavir, hepatotoxicity of, 1430 ABC mnemonic, in acute abdominal pain, 154 ABC transporters in biliary lipid secretion, 1096-1097, 1096f in cholesterol absorption, 1706 in gallstone disease, 1102, 1103t in hepatic drug elimination, 1418-1419 Abdomen abscess of. See Abdominal abscess. acute, in leukemia, 567 distention of endoscopy and, 656 fiber intake and, 276 intestinal gas causing, 238-240, 239f in irritable bowel syndrome, 238-240, 239f, 2092 in small intestinal obstruction, 2106-2107 examination of in acute abdominal pain, 155-156 in acute pancreatitis, 969, 970f in chronic abdominal pain, 163-164 open, 417 pain in. See Abdominal pain. upper, pain or discomfort centered in, 183. See also Dyspepsia. Abdominal abscess, 411-419 antibiotics for, 417-419, 418t bacteriology of, 412-413 computed tomography in, 413-414, 413f-414f in Crohn’s disease, 1951 diagnostic imaging in, 413-415, 413f-414f in diverticulitis, 2080-2082, 2081f drainage of, 415-416 percutaneous, 415, 416f-417f surgical, 417 magnetic resonance imaging in, 414-415 management of, 413-419 outcome in, 419 pathophysiology of, 411-412, 412t radiography in, 414 radionuclide imaging in, 414 risk factors for, 411, 412t ultrasonography in, 414, 414f Abdominal accommodation reflex, 237 Abdominal adhesions after peritonitis, 618 small intestinal obstruction in, 2107f, 2108 Abdominal aortic aneurysm, 606-607 mycotic, 607, 607f ruptured, 154t, 160 Abdominal compartment syndrome, 160 Abdominal distress, after bariatric surgery, 117-118 Abdominal epilepsy, 578-579 Abdominal massage therapy, for constipation, 2293-2295 Abdominal migraine, 578 Abdominal pain in abdominal aortic aneurysm rupture, 154t, 160, 607 in abdominal compartment syndrome, 160 acute, 151-162 age and, 161 aggravating and alleviating factors in, 155 anatomy of, 151-153, 152f-153f approach to acute care of, 154

Abdominal pain (Continued) associated symptoms in, 155 character of, 154t, 155 in children, 161 chronology of, 154-155, 154f, 154t clinical evaluation of, 153-156, 154f, 154t, 157f computed tomography in, 156 definition of, 163 etiology of, 154t, 156-161, 157t, 160t extra-abdominal causes of, 160, 160t history in, 154-155, 154t imaging studies in, 156 in immunocompromised hosts, 161 intensity of, 154t, 155 laboratory data in, 156 location of, 154t, 155 past medical history in, 155 physical examination in, 155-156 in pregnancy, 161 treatment of, 161 ultrasonography in, 156 vomiting in, 200 in acute biliary disease, 154t, 157-158 in acute mesenteric ischemia, 154t, 159-160, 2030 in acute pancreatitis, 154t, 158-159, 969 in anterior cutaneous nerve entrapment, 164-165 in appendicitis, 154t, 157, 2061 assessment of, 2280, 2281f biliary, 157-158 in celiac disease, 1805 after cholecystectomy, 1135-1137, 1137t in cholecystitis acute, 154t, 157-158 in children, 1065 chronic, 163-172. See also Functional abdominal pain syndrome. acute exacerbation of, 163-164 clinical evaluation of, 163-164 definition of, 163-164 differential diagnosis of, 164t imaging studies in, 164 laboratory tests in, 164 laparoscopic adhesiolysis for, 170 in chronic pancreatitis, 995-996, 1004-1009, 1039. See also Pancreatitis, chronic, abdominal pain in. in colonic obstruction, 2117 in colonic ulcer, 2244 in Crohn’s disease, 1949, 1953 in diabetes mellitus, 575 in distal intestinal obstruction syndrome, 945 in diverticulitis, 154t, 158, 2078 in eating disorders, 128-130 in ectopic pregnancy rupture, 154t functional. See Functional abdominal pain syndrome. in gallstone disease, in children, 1064 in gastric motility disorders, 810-811 in gastroenteritis, 154t after hematopoietic stem cell transplantation, 552, 553t in hepatocellular carcinoma, 1570 in HIV/AIDS, 530-531, 530t-531t in intestinal angina, 2044-2045 in irritable bowel syndrome, 2092, 2095 in mesenteric venous thrombosis, 2036

Abdominal pain (Continued) in myofascial pain syndromes, 164-165 neuropathic, 578-579, 2280, 2283t peripheral, 579 palliative care for, 2280-2282, 2281f, 2281t, 2283t in pancreatic cancer, 1020 in pelvic inflammatory disease, 154t in peptic ulcer perforation, 154t, 159, 159f in peritonitis, 614 referred, 152-153, 153f, 163 in slipping rib syndrome, 165 in small intestinal obstruction, 154t, 158, 2106-2107 after solid organ transplantation, 538t, 542-543 somatic-parietal, 152, 2280 subacute, 163 in thoracic nerve radiculopathy, 165 transmission of, 151-153, 152f-153f types of, 2280 in ulcerative colitis, 1983 visceral, 151-152, 152f-153f, 2280. See also Visceral pain. in Zollinger-Ellison syndrome, 501-502 Abdominal paracentesis. See Paracentesis. Abdominal surgery ascites after, 1528 blind loop after, small intestinal bacterial overgrowth and, 1773 ileus after. See Ileus, postoperative. irritable bowel syndrome and, 2094 Abdominal wall defects, 1626-1628, 1627f-1628f Abdominal wall hernia, 379 in ascites, 1535 Abdominal wall pain, chronic, 164-165 Abernethy malformation, 1205, 1205f Abetalipoproteinemia gastrointestinal manifestations of, 577, 577f malabsorption in, 1759t-1762t, 1763 Abscess abdominal. See Abdominal abscess. in acute pancreatitis, 960-961 anorectal, 2266-2267, 2267f in appendicitis, 2067 crypt, in radiation enteritis, 642-643, 643f in diverticulitis, 2080-2082, 2081f hepatic, 1366-1369 amebic, 1368-1369, 1368f, 1369t. See also Amebiasis, hepatic. fungal, after hematopoietic stem cell transplantation, 547t, 549, 554 pyogenic, 1366-1368, 1366t, 1367f, 1369t pancreatic, 960-961, 1038 pelvic, 416, 416f periappendiceal, 413-414, 413f, 416, 417f peridiverticular, 413-414, 414f, 416 postoperative, Seprafilm and, 412 subphrenic, 416, 416f tubo-ovarian, versus appendicitis, 2063t Absorption. See Digestion and absorption; Malabsorption. Abuse, substance colonic ischemia in, 2040 hepatotoxicity related to, 1454 liver transplantation and, 1596 Abuse history in biopsychosocial model, 341 in functional abdominal pain syndrome, 168

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

i

ii

Index Abuse history (Continued) in irritable bowel syndrome, 167-168 visceral hypersensitivity and, 345 Acanthosis nigricans malignant, 364, 364f in obesity, 105-106 Acarbose diarrhea from, 574 for hepatic encephalopathy, 1546 malabsorption with, 1757t Accommodation, gastric functional dyspepsia and, 187-188 in gastric motility, 794-796, 796f-797f Acebutolol, hepatotoxicity of, 1433 Aceruloplasminenia, hereditary, 1253 Acetaldehyde, ethanol conversion to, 1384 Acetaminophen hepatotoxicity of, 1427-1429, 1438t acute liver failure and, 1558-1559, 1558t-1559t, 1559f, 1563, 1563t-1564t alcoholic liver disease versus, 1392 clinical course and prognosis in, 1427-1428 management of, 1428-1429, 1428f prevention of, 1429 risk factors for, 1427, 1427t jaundice from, 327 Acetic acid, in esophageal cancer, 754 Acetyl-coA, 1220 Acetyl-coA acetyltransferase, 1704 Acetyl-coA carboxylase, 1220 Acetylcholine in gastric acid secretion, 821f-822f, 822-823 in gastrointestinal tract, 5, 10-11 in intestinal ion transport, 1691, 1693f receptors for, 10-11 N-Acetylcysteine for acetaminophen poisoning, 1428 for acute liver failure, 1565 Achalasia candidal esophagitis in, 741-742, 742f versus Chagas’ disease, 692-693 clinical manifestations of, 691-692 differential diagnosis of, 692 dysphagia in, 174 epidemiology of, 686-687 herpes simplex virus 1 in, 690 pathogenesis of, 689-690 versus pseudoachalasia, 693 squamous cell carcinoma and, 703, 747-748 subtypes of, 695, 697f, 701 treatment of, 701-703 botulinum toxin injection in, 701 failure of, 702-703 Heller myotomy in, 702 pharmacotherapy in, 701-702 pneumatic dilation in, 701-702 vigorous (spastic), 689-690 Achlorhydria, small intestinal bacterial overgrowth and, 1773 Acid(s) bile. See Bile acid(s). caustic injury from, 406, 407f duodenal hypersensitivity to, in functional dyspepsia, 188 gastric. See Gastric acid. Acid-base balance, in intestinal ion transport, 1689 Acid steatocrit test, for fecal fat, 1751 Acidosis lactic from nucleoside analogue reverse transcriptase inhibitors, 532-533 in short bowel syndrome, 1790-1791 metabolic in glycogen storage disease type I, 1263 in short bowel syndrome, 1790-1791 Acinar cell(s), 912-915, 912f-914f, 921 in chronic pancreatitis, 986, 986f feeding response of, 912 gap junctions in, 913, 921 Golgi complex in, 913-914, 914f granular endoplasmic reticulum in, 913, 913f microvilli in, 913, 913f mitochondria in, 913

Acinar cell(s) (Continued) receptors of, 925, 925f secretory products of, 923, 923t tight junctions in, 913, 921 Acinar cell carcinoma, 1027t, 1033 Acinetobacter lwoffi infection, gastritis in, 851 Acinus, of liver, 1204, 1204f Acitretin, hepatotoxicity of, 1433 Ackee tree, hepatotoxicity of, 1455 Acne, in ulcerative colitis, 2010 Acotiamide, for functional dyspepsia, 193 Acquired immunodeficiency syndrome (AIDS). See HIV/AIDS. Acrochordon. See Skin tags. Acrodermatitis enteropathica malabsorption in, 1759t-1762t in zinc deficiency, 368-369, 368f Acrokeratosis, paraneoplastic, 365 Acromegaly colorectal adenoma and, 2164-2165 gastrointestinal manifestations of, 573t, 576 in GRFoma, 513-514 ACTH. See Corticotropin. ACTHoma. See Cushing’s syndrome. Actinomycosis gastritis in, 850 hepatic manifestations of, 1352 Activating protein-1, in Helicobacter pylori infection, 836 Acupuncture, 2288t colonic motility and, 1672 for constipation, 283 for irritable bowel syndrome, 2292, 2292t for liver disease, 2297 for nausea and vomiting, 208, 2289, 2289t Acustimulation, for gastric motility disorders, 812-814, 813t Acute fatty liver of pregnancy. See Fatty liver disease, pregnancy-related. Acute respiratory distress syndrome, in acute pancreatitis, 963 Acyclovir for herpes simplex esophagitis, 742 for herpes zoster ulcers, 357 for orolabial herpes, 357 Adalimumab, for Crohn’s disease, 1967-1969 Addison’s disease. See Adrenal insufficiency. Adductor pollicis electrical stimulation, in nutritional assessment, 67 Adefovir for hepatitis B, 1303 in liver transplantation, 1599-1600 plus lamivudine, 1305 in pregnancy, 1306 resistance to, 1303 HBV DNA polymerase mutation and, 1292, 1292f Adenocarcinoma ampullary. See Ampulla of Vater, carcinoma of. appendiceal, 2070 biliary. See Cholangiocarcinoma. colorectal. See Colorectal cancer. esophageal. See Esophageal cancer, adenocarcinomatous. gallbladder. See Gallbladder carcinoma. gastric. See Gastric cancer. rectal. See Colorectal cancer. small intestinal. See Small intestinal adenocarcinoma. Adenoid cystic carcinoma, esophageal, 767 Adenoma. See also Cystadenoma. colorectal. See Colorectal adenoma. esophageal, 768 gallbladder, 1149, 1150t gastric in familial adenomatous polyposis, 2179 malignant transformation of, 896 hepatocellular. See Hepatocellular adenoma. rectal. See Colorectal adenoma. small intestinal in familial adenomatous polyposis, 2148 pathology of, 2146 progression of, 2148

Adenoma-carcinoma hypothesis, 2159-2160 Adenomatosis, in hyperinsulinism, 495-496 Adenomatous polyposis, familial. See Familial adenomatous polyposis. Adenomatous polyposis coli gene. See APC gene. Adenomyoma, 1146, 1147f, 1149, 1150t, 1183-1184 Adenomyomatosis, 1146-1149 clinical features of, 1147-1148, 1148f definition of, 1146 diagnosis of, 1148-1149, 1148f epidemiology of, 1146 gallbladder carcinoma and, 1178 pathogenesis of, 1146-1147 pathology of, 1146, 1147f segmental, 1146, 1147f treatment of, 1149 Adenosine in gastrointestinal tract, 12-13 in intestinal ion transport, 1691 Adenosine deaminase, in ascites, 1527 Adenosine monophosphate, cyclic, in intestinal ion transport, 1682, 1682f, 1693, 1693f S-Adenosyl-l-methionine for alcoholic liver disease, 1397 for cholestasis of pregnancy, 631-632 deficiency of, in alcoholic liver disease, 1386-1387, 1386f for liver disease, 2295t, 2296 Adenovirus infection, enteric, 1870t, 1871-1872 Adenovirus 12, in celiac disease, 1801 Adherence factors, bacterial, 1844-1845, 1845f Adhesiolysis, laparoscopic, for chronic abdominal pain, 170 Adhesion molecules antibodies against, for ulcerative colitis, 2001 in cellular proliferation, 32 in Crohn’s disease, 1946-1947 Adhesions formation of, 412 intra-abdominal after peritonitis, 618 in small intestinal obstruction, 2105, 2107f, 2108, 2112 Adiponectin in eating disorders, 123 insulin release and, 103 in nonalcoholic fatty liver disease, 1403 Adipose tissue excess. See Obesity. in hepatic energy metabolism, 1216-1217 mesenteric, hypertrophy of, in Crohn’s disease, 1949 Adiposity, central. See Central adiposity. Adiposity rebound, 102 Adolescents, Crohn’s disease in, 1971-1972 Adrenal insufficiency in cirrhosis, 1554 familial, with alacrima, 686-687 gastrointestinal manifestations of, 573t, 576 β-Adrenergic blockers. See Beta blockers. Adrenergic receptor, 11 alpha2-, in postoperative ileus, 2125 Adrenergic symptoms, in insulinoma, 496 Adrenocorticotropic hormone. See Corticotropin. Adrenoleukodystrophy, X-linked, 1275 Aeromonas infection, 1853-1854 Afferent (sensory) neurons, extrinsic primary, 5 Afferent pathways, in vomiting, 197, 198f Afferent (ascending) visceral pain transmission, 151-152, 153f, 166-167, 166f, 343, 343f amplification of. See Visceral sensitization. Aflatoxin, in oncogenesis, 41-42 Aflatoxin B1 hepatocellular carcinoma and, 1577 hepatotoxicity of, 1455 African iron overload, 1240 Agammaglobulinemia, infantile, X-linked, malabsorption in, 1764 Aganglionosis. See Hirschsprung’s disease. Age. See also Children; Elderly; Infants. colorectal cancer and, 2208, 2208f, 2216 gallstone disease and, 1090-1091

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Age (Continued) gastroesophageal reflux disease and, 705, 706f Helicobacter pylori infection and, 833 hepatitis C and, 1317, 1325 hepatocellular carcinoma and, 1570 liver transplantation and, 1597 ulcerative colitis and, 1976 Agent Orange, hepatotoxicity of, 1452 AIDS. See HIV/AIDS. Air swallowing, intestinal gas from, 233, 235f, 237 Airway obstruction, during endoscopy, 657 Akt protein kinase, in esophageal cancer, 750 Alagille’s syndrome, 1060-1062, 1061f-1062f, 1206, 1206f Alanine aminotransferase (ALT), 1229-1231 in acetaminophen poisoning, 1427 in acute pancreatitis, 973 in alcoholic liver disease, 1390, 1391t in cystic fibrosis, 946-947 in eating disorders, 130-131 elevation in, evaluation of, 1230-1231, 1230t, 1231f in hepatitis C, 1320, 1320f, 1326 in hepatitis E, 1340, 1340f in HIV/AIDS, 535 in jaundice, 331 in nonalcoholic fatty liver disease, 1405 normal values for, 1229-1230 in primary biliary cirrhosis, 1480 in sickle cell anemia, 571 in Wilson disease, 1253 Albendazole for ascariasis, 1360, 1922-1923 for capillariasis, 1925 for cysticercosis, 1932 for Encephalitozoon intestinalis, 1918 for fish tapeworm infection, 1931 for hookworm infection, 1926-1927 for hydatid cyst, 1365 for liver fluke infection, 1934-1935 for pinworm infection, 1929 for strongyloidiasis, 1361 for trichinosis, 1361, 1930 for Trichuris trichiura infection, 1928 Albumin in bile, 1093 bilirubin binding to, 1227 dialysis with for acute liver failure, 1568 for hepatic encephalopathy, 1546 for hepatorenal syndrome, 1549 intravenous for ascites, 1539-1540 for hepatorenal syndrome, 1548, 1548t for spontaneous bacterial peritonitis, 1533 as marker of hepatic synthetic function, 1234 serum in alcoholic liver disease, 1390, 1391t enteric loss of, 437. See also Protein-losing gastroenteropathy. in nutritional assessment, 67-68, 67t, 77-78 Alcohol abuse of, diagnosis of, 1389, 1390f cessation of for chronic pancreatitis, 1005 gastroesophageal reflux disease and, 721 esophageal cancer and, 747-748 gastric cancer and, 892 gastropathy and, 856-857, 857f hepatitis C and, 1325 level of, in various beverages, 1383, 1384t liver disease and. See Alcoholic liver disease. liver transplantation and, 1596 malabsorption with, 1757t metabolism of, 1384-1385 methotrexate hepatotoxicity and, 1444 Oriental flush syndrome with, 1384-1385 pancreatitis and. See Alcoholic pancreatitis. peptic ulcer disease and, 863 Alcohol dehydrogenase, 1384

Alcoholic liver disease, 1383-1400 versus acetaminophen hepatotoxicity, 1392 versus acute viral infection, 1392 versus amiodarone hepatotoxicity, 1391 ascites in, 1518. See also Ascites. AST/ALT ratio in, 1230 versus Budd-Chiari syndrome, 1391 cirrhosis in hepatitis C and, 1392, 1392f histology of, 1383-1384, 1384f, 1390-1391 optimal management of, 1399-1400 prognosis in, 1394, 1394f protein-energy malnutrition in, 64 clinical deterioration in, differential diagnosis of, 1392 decompensated, nutritional support in, 70 diagnosis of, 1389-1392, 1390t-1391t differential diagnosis of, 1391 epidemiology of, 1383, 1384t fatty liver in, 1383-1384, 1384f gender and, 1383, 1388 hepatic encephalopathy in, 1390, 1390t, 1393, 1393f hepatic fibrosis in, 1389 hepatitis C and, 1392, 1392f hepatitis in cholestasis in, 329 histology of, 1383-1384, 1384f jaundice in, 327 nonalcoholic fatty liver disease versus, 1408-1409 optimal management of, 1399 prognosis in, 1393, 1393f versus hepatocellular carcinoma, 1392 hepatorenal syndrome in, 1393, 1393f versus hereditary hemochromatosis, 1391 histologic spectrum of, 1383-1384, 1384f histopathology of, 1384f, 1390-1391 history in, 1389-1392 laboratory findings in, 1390, 1391t malnutrition in, 1395 versus nonalcoholic fatty liver disease, 1391 obesity and, 1392-1393, 1395 pathogenesis of, 1384-1389 autoantibodies in, 1388 cytokines in, 1387-1388, 1388f emerging mechanisms in, 1389 ethanol metabolism in, 1384-1385 gender and, 1388 genetic factors in, 1389 hypoxia in, 1385-1386 immune and inflammatory mechanisms in, 1387-1388, 1388f Kupffer cell activation in, 1387-1388, 1388f methionine metabolism in, 1386-1387, 1386f mitochondrial dysfunction in, 1385 oxidative stress in, 1385 proteasome dysfunction in, 1386 S-adenosylmethionine deficiency in, 1386-1387, 1386f physical findings in, 1390, 1390t portal hypertension in, 1498-1499 prognosis in, 1393-1394, 1393f progression of, cofactors influencing, 1392-1393, 1392f smoking and, 1392-1393 treatment of, 1394-1400 abstinence for, 1394-1395 anabolic steroids for, 1398 antioxidants for, 1397-1398 anti–tumor necrosis factor therapy for, 1397 colchicine for, 1398 glucocorticoids for, 1396-1397, 1396f-1397f lifestyle modification for, 1394-1395 liver transplantation for, 1398-1399, 1399f, 1598-1599 nutritional therapy for, 1395-1396, 1395f optimal, 1399-1400 pentoxifylline for, 1396-1397, 1397f polyenylphosphatidylcholine for, 1398 propylthiouracil for, 1398 S-adenosylmethionine for, 1397 silymarin for, 1398

Alcoholic liver disease (Continued) ursodeoxycholic acid for, 1398 vitamin E for, 1398 Alcoholic pancreatitis acute, 964-965, 970 chronic pancreatitis in, 989 versus gallstone pancreatitis, 973 with hyperlipidemia, 966 chronic, 964-965, 988-989 pathophysiology of, 987-988 prognosis in, 989 Aldosterone, in intestinal ion transport, 1690 Alendronate, for osteoporosis, in primary biliary cirrhosis, 1485-1486 Alexithymia, 342 Alkali, caustic injury from, 406, 406f, 408 Alkaline phosphatase, 1231-1234 in alcoholic liver disease, 1390, 1391t in choledocholithiasis, 1116 in cholestasis, 1232, 1232f, 1233t in cystic fibrosis, 946-947 elevation in evaluation of, 1232-1234, 1232f, 1233t gamma glutamyl transpeptidase level and, 1232, 1232f 5′-nucleotidase level and, 1232, 1232f in HIV/AIDS, 535 in jaundice, 331 in primary biliary cirrhosis, 1480 in primary sclerosing cholangitis, 1160 serum, physiologic variations in, 1231 in Wilson disease, 1231 Allele, deletion of, 38, 39f Allergen exposure, in eosinophilic esophagitis, 426-427, 426f Allergic angiitis, granulomatous. See ChurgStrauss syndrome. Allergic eosinophilic esophagitis, 143-144, 146t Allergic eosinophilic gastroenteritis, 144, 146t Allergic eosinophilic proctocolitis, 144-145, 146t Allergic gastritis, 856 Allergic gastroenteropathy, 440 Allergy. See also Hypersensitivity reaction. drug, 1422, 1431-1434, 1431t food. See Food allergy. gastrointestinal, 143 oral, 143 pollen-food, 143 Allergy testing, in eosinophilic esophagitis, 430 Allopurinol, granulomatous hepatitis from, 1437t Allopurinol loading test, in urea cycle defects, 1273 Aloe, for colonic health, 2294t Alopecia, in Cronkhite-Canada syndrome, 364 Alosetron colonic ischemia from, 2040 for fecal incontinence, 253 for irritable bowel syndrome, 2102 Alpha-chain protein, for small intestinal immunoproliferative disease, 455, 457 Alpha fetoprotein in hepatoblastoma, 1582 in hepatocellular carcinoma, 1571-1572, 1572t Alpha heavy chain disease, gastrointestinal manifestations of, 569 Alpha-tocopherol. See Vitamin E. Alpidem, hepatotoxicity of, 1436 Alprazolam, for gastric motility disorders, 812, 813t ALT. See Alanine aminotransferase (ALT). Alveolar-arterial oxygen gradient, in hepatopulmonary syndrome, 1550 Alveolar bone loss, in HIV/AIDS, 355 Alveolar hydatid disease, 1364 Alverine, hepatotoxicity of, 1436-1437 Alvimopam for constipation, 281 ileus and, 2127 Amanita phalloides hepatotoxicity of, 1450t, 1454 jaundice from, 327 Ambrisentan, for portopulmonary hypertension, 1552

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

iii

iv

Index Amebae, intestinal, 1907f Amebiasis hepatic, 1368-1369 clinical features of, 1368 diagnosis of, 1369, 1369f pathogenesis of, 1367f, 1368 versus pyogenic liver abscess, 1369t treatment of, 1369 intestinal, 1990, 1990t clinical presentation in, 1908-1909, 1909t diagnosis of, 1908f, 1909-1910 epidemiology of, 1905-1907, 1906f-1907f immune response to, 1908 necrotizing colitis with toxic megacolon in, 1909 pathogenesis of, 1907-1908 pathology of, 1907-1908, 1908f in pregnancy, 1910 prevention and control of, 1910-1911 treatment of, 1910, 1910t Amebicidal agents, 1910, 1910t Amenorrhea in celiac disease, 1806 in eating disorders, 128 Amifostine for radiation enteritis, 649-650 for radiation-induced esophagitis, 641 Amiloride, for ascites, 1537 Amine precursor uptake and decarboxylation tumors, 492-493 Amino acid transporters, 1716-1717, 1716f, 1716t defects of, malabsorption in, 1758-1763, 1759t-1762t Amino acids. See also Protein(s). absorption of, 1716-1717, 1716f, 1716t decreased, 1738 dietary sources of, 1712 essential, 1712 malabsorption of, congenital disorders of, 1759t-1762t in protein, 49-50, 79-80 Aminolevulinic acid dehydratase deficiency, 1267, 1267t 5-Aminolevulinic acid synthase in heme synthesis, 1266, 1266f in porphyrias, 1268 Aminopeptidase(s), 1714, 1714t Aminopeptidase A, 1714, 1714t Aminopyrine breath test, 1235 Aminosalicylates for colorectal cancer prophylaxis, in ulcerative colitis, 2008 for Crohn’s disease, 1960-1961, 1962t for diverticulosis, 2077 topical, for ulcerative colitis, 1995 for ulcerative colitis, 1993-1995, 1993f, 1994t Aminosidine, for amebic liver abscess, 1369 Aminotransferases. See Alanine aminotransferase (ALT); Aspartate aminotransferase (AST). Amiodarone hepatic storage of, phospholipidosis from, 1442 hepatotoxicity of, 1442 versus alcoholic liver disease, 1391 Amitriptyline for fecal incontinence, 253 for gastric motility disorders, 812, 813t for levator ani syndrome, 2273 Amlodipine, hepatotoxicity of, 1433 Ammonia, elevated. See Hyperammonemia. Ammonium tetrathiomolybdate, for Wilson disease, 1257 Amodiaquine, hepatotoxicity of, 1431 Amoxicillin, for Helicobacter pylori infection, 841-842, 842t-843t Amoxicillin-clavulanic acid for ascitic fluid infection, 1533 cholestatic hepatitis from, 1440 hepatotoxicity of, 1414-1415 Amphetamines, colonic ischemia from, 2040 Ampicillin, for typhoid fever, 1866-1867 Ampulla of Vater, 779 anatomy of, 1048-1049, 1048f carcinoid tumors of, 479-480

Ampulla of Vater (Continued) carcinoma of, 1181-1183 clinical features and diagnosis of, 1182, 1182f epidemiology of, 1181 etiology of, 1181 pathogenesis of, 1182 pathology of, 1181-1182, 1182f staging of, 1182 treatment of, 1182-1183 sphincter of, 1049, 1067, 1068f. See also Sphincter of Oddi. stenosis of, 1067-1068 Amylase in ascitic fluid, 1525 in carbohydrate digestion, 1707-1708, 1707f functions of, 923-924 pancreatic, 923-924 in carbohydrate digestion, 1707-1708, 1707f salivary, 923-924 in carbohydrate digestion, 1707 serum, in acute pancreatitis, 970-971 Amylin, insulin regulation by, 18-19 Amyloidosis, 584-588 chronic intestinal pseudo-obstruction in, 2138-2139 classification of, 584-587, 584t gastrointestinal, 585-586, 585t, 586f hepatic amyloid deposits in, 586, 586f clinical and laboratory findings in, 587 diagnosis of, 587-588 treatment and prognosis in, 588 hereditary, 2138 jaundice in, 327-328 malabsorption in, 1756 in multiple myeloma, 568-569 oral-cutaneous, 361-362 primary, 2138 radiography in, 586, 586f secondary, 2138 senile, 2139 sites of amyloid deposition in, 585 in ulcerative colitis, 2012 Amylopectin, 1706-1707, 1706f Amylopectinosis, 1264 Amylose, 1706-1707 Amyotrophic lateral sclerosis dysphagia in, 688 gastrointestinal manifestations of, 581 Anabolic steroids for alcoholic liver disease, 1398 hepatotoxicity of, 1440 jaundice from, 328-329 Anal. See also Anorectum; Anus. Anal canal anatomy of, 1616-1617, 1622, 1622f, 2257-2258, 2258f blood supply of, 2257 carcinoma of, 2270 Crohn’s disease of, 1950 dentate line of, 2257, 2258f electrostimulation of, for fecal incontinence, 255 endoscopy of, 2259 examination of, 2258-2260, 2259f imaging of, in fecal incontinence, 248, 248f innervation of, 2257 keyhole deformity of, 2257-2258 lesions of, in Crohn’s disease, 1950 lymphatic drainage of, 2257 palpation of, 2259 radiofrequency energy application to, for fecal incontinence, 256 vascular supply to, 1617 Anal cancer, 2269-2270 Anal crypt, 2257, 2258f Anal endosonography, in fecal incontinence, 248, 248f Anal fissure, 2263-2266 bleeding in, 314 clinical manifestations and diagnosis of, 2264-2265, 2265f in Crohn’s disease, 1950

Anal fissure (Continued) etiology of, 2264 inspection of, 2259, 2259f postpartum, 2266 treatment of, 2265-2266, 2265t Anal fistula, 2267-2268, 2268f anorectal anomalies with, 1634f in Crohn’s disease, 2268-2269 after radiation therapy, 2269 Anal intraepithelial neoplasia, 2270 Anal margin cancer, 2269 Anal melanoma, 2270 Anal motor evoked potentials, in fecal incontinence, 249, 250f Anal pain, unexplained, 2272-2273 Anal plug, 254-255, 2267, 2268f Anal skin tags, 2262-2263, 2265f Anal sphincter bulking of, for fecal incontinence, 255 in continence mechanism, 242, 242f electromyography of in constipation, 273 in fecal incontinence, 249 external, 1616-1617, 1664-1665, 2258, 2258f internal, 1664-1665, 2257-2258, 2258f motility of, 1669-1670 myopathy of, constipation in, 269 obstetric injury to, 243-244, 248 pressure in, measurement of, 247-248, 247f, 251t repair of, for fecal incontinence, 255, 255t weakness of, in fecal incontinence, 243-244 Anal sphincteroplasty, for fecal incontinence, 255-256, 255t Anal stenosis, 1635, 2272 constipation in, 269 Anal strictures, in Crohn’s disease, 1950 Anal valves, 1616-1617 Anal warts, 2270-2271, 2271t Analgesics adjuvant, for neuropathic pain, 2282, 2283t for chronic pancreatitis, 1004 for esophageal cancer, 767 narcotic. See Opioid(s). in palliative care, 2280-2282, 2281f, 2281t, 2283t visceral, for functional dyspepsia, 193 Anaphylaxis, gastrointestinal, oral tolerance during, 23 Anastomotic circulation, splanchnic, 2028, 2029f Ancylostoma caninum infection, 1927 Ancylostoma duodenale infection, 1925-1926, 1926f, 1926t. See also Hookworm infection. Anderson’s disease, 1704, 1759t-1762t, 1763 Anemia in celiac disease, 1805-1806 in Crohn’s disease, 1953 in esophageal cancer, 753 in glucagonoma, 507 hemolytic, salmonellosis and, 1862 iron deficiency differential diagnosis of, 321 iron-refractory, malabsorption in, 1759t-1762t occult gastrointestinal bleeding and, 321 in malabsorption, 1745t megaloblastic, in small intestinal bacterial overgrowth, 1774 pernicious autoimmune metaplastic atrophic gastritis in, 847-848 cobalamin deficiency in, 830 sickle cell hepatic manifestations of, 570-572, 570f-571f salmonellosis and, 1862 in ulcerative colitis, 2012 Anendocrinosis, enteric, malabsorption in, 1759t-1762t, 1765, 1766f Anesthetics hepatotoxicity of, 1447-1449, 1448t-1450t postoperative ileus and, 2125 topical, methemoglobinemia from, 657 Aneurysm abdominal aortic, 606-607 mycotic, 607, 607f ruptured, 154t, 160

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Aneurysm (Continued) hepatic artery, 1381-1382 mycotic, 607, 607f Aneurysmal dilatations, in polyarteritis nodosa, 562, 562f Angiitis, allergic, granulomatous. See ChurgStrauss syndrome. Angina intestinal, 2044-2046, 2045f microvascular, chest pain in, 178 Angiodysplasia definition of, 593 gastrointestinal, in renal failure, 578 Angioectasia colonic, 593-599 angiography in, 595-596, 597f, 598 aortic stenosis in, 595 bleeding from, 314, 593-594, 599 colonoscopy in, 597, 598f, 598t diagnosis and management of, 594f, 595-596, 597f, 598 pathogenesis of, 595-597, 596f-597f pathology of, 595, 595f-596f gastrointestinal bleeding from, 314-317, 317f Angiogenesis in hepatopulmonary syndrome, 1550 in metastasis, 42-43 Angiography. See also Cholangiography. in acute abdominal pain, 156 in acute mesenteric ischemia, 2032 in angioectasia, 595-596, 597f, 598 in colonic diverticular bleeding, 312 computed tomography, in acute mesenteric ischemia, 2031 in diverticular bleeding, 2088, 2088f in gastrointestinal bleeding, 292 in hepatocellular carcinoma, 1573 in lower gastrointestinal bleeding, 310 magnetic resonance, in portal vein thrombosis, 1378-1379, 1378f in nonocclusive mesenteric ischemia, 2035, 2035f in obscure gastrointestinal bleeding, 319 in peptic ulcer bleeding, 302 pulmonary, for hepatopulmonary syndrome, 1552 in small intestinal tumors, 2151 in superior mesenteric artery embolus, 2034, 2034f in superior mesenteric artery thrombosis, 2035 Angiomatosis, bacillary hepatic infection in, 1353 and HIV/AIDS, 1353 Angioplasty for Budd-Chiari syndrome, 1374 percutaneous transluminal mesenteric, for intestinal angina, 2045-2046 Angiosarcoma hepatic, 1582-1583 drug-induced, 1451-1452 liver transplantation for, 1596 Angiotensin in hepatic vasoconstriction, 1490-1491 in portal hypertension, 1493 Angiotensin-converting enzyme (ACE) inhibitors, hepatotoxicity of, 1433 Angiotensin receptor blockers, hepatotoxicity of, 1433 Angle of His, antireflux function of, 708, 708f Anion exchanger, 1685 Anisakiasis gastric, 852 intestinal, 1930 Anisakis simplex infection, 1930 Ankylosing spondylitis gastrointestinal manifestations of, 563 in ulcerative colitis, 2011 Ann Arbor staging system, for gastrointestinal lymphoma, 447, 447t Annular pancreas, 780t, 786-787, 917-918, 917f Anocutaneous fistula, 1634-1635 Anocutaneous reflex, 246 Anoderm, 2257, 2258f

Anorectal angle, 242, 242f Anorectal contractile response, 242-243 Anorectal junction, 1664 Anorectal manometry in constipation, 273 in fecal incontinence, 247-248, 247f, 251t in Hirschsprung’s disease, 1638-1639, 1638f Anorectal motility during defecation, 1670, 1670f rectal filling, capacitance, and accommodation in, 1669-1670 Anorectal ring, 2258 Anorectal sampling in continence mechanism, 242-243 impaired, fecal incontinence and, 245 Anorectal sensation, impaired, fecal incontinence and, 244-245 Anorectal varices, 1494, 1513, 1513f Anorectum, 2257-2274. See also Anal; Anus; Perianal; Rectum. abnormalities of, in fecal incontinence, 243-245, 243f, 243t abscess of, 2266-2267, 2267f anatomy and physiology of, 242-243, 242f, 1664-1665, 2257-2258, 2258f disease of, in HIV/AIDS, 531, 531t endoscopy of, 2259-2260 examination of, 2258-2260, 2259f innervation of, 1665 malformations of, 1633-1636, 1634f, 1635t1636t anomalies associated with, 1635-1636, 1636t classification of, 1634, 1635t Anorexia in Crohn’s disease, 1953 after hematopoietic stem cell transplantation, 545-546, 546f, 555 palliative care for, 2282-2284 after solid organ transplantation, 538t, 541 Anorexia-cachexia syndrome, 2282-2284 Anorexia nervosa binge-eating–purging type, 124 chronic intestinal pseudo-obstruction in, 2140 clinical features or complications of, 127, 129t-130t constipation in, 271 diagnosis of, 122f, 122t, 124 differential diagnosis of, 125-126, 126t epidemiology of, 121 etiology of, 121-123 gastrointestinal manifestations of, 128-131, 129t-130t medical management of, 135-137, 136t laboratory evaluation of, 128 nutritional assessment in, 125-126, 126t onset and course of, 123 pharmacotherapy for, 133 psychotherapy for, 132 restricting type, 124 treatment of, 131-137, 131f weight management for, 134 Anoscopy, 2259 in lower gastrointestinal bleeding, 309 Anovaginal fistula, 2269 Antacids adverse effects of, 869-870 for gastroesophageal reflux disease, 721 malabsorption with, 1757t mechanism of action of, 869-870 for NSAID ulcer prophylaxis, 873 for peptic ulcer disease, 869-870 in pregnancy, 629 Anterior cingulate cortex in functional abdominal pain syndrome, 166-168, 166f-167f in psychosocial distress and visceral pain, 344-345 Anterior cutaneous nerve entrapment, abdominal pain in, 164-165 Anthraquinones for constipation, 277t, 279 pseudomelanosis coli and, 2246f, 2247

Anthrax infection gastrointestinal, 1879t-1880t, 1883-1884 oropharyngeal, 1884 Anthropometry, in nutritional assessment, 65-66, 65t-66t, 77, 78f Anti-neutrophil nuclear antigens, in autoimmune hepatitis, 1461 Anti-tissue transglutaminase antibodies, in celiac disease, 1798, 1807-1809, 1807t Antiangiogenic factors, for colonic angioectasia, 599 Antibiotics for abdominal abscess, 417-419, 418t abdominal distention and bloating and, 239 for acute pancreatitis, 977-978 for Campylobacter spp. infection, 1865t, 1868 for cholangitis, 1118 in cirrhotic patients with variceal bleeding, 307 Clostridium difficile-associated diarrhea and colitis from, 1892, 1892t for Crohn’s disease, 1961-1963, 1962t for diarrhea, 226 diarrhea from, 217-218, 229-230, 1889-1890, 1890t for diverticular disease, 2077-2078, 2080 empiric, for infectious diarrhea, 1885 esophagitis from, 736f, 737 for gastric diffuse large B cell lymphoma, 453 for gastric MALT lymphoma, 450-451 for Helicobacter pylori infection, 841-842, 842t-843t hemorrhagic colitis from, 2039 for hepatic encephalopathy, 1545-1546 for infectious diarrhea, 1847t, 1865t, 1885-1886 for irritable bowel syndrome, 2102 for neutropenic enterocolitis, 2253 for peritonitis, 161, 615 for pouchitis, 2009, 2021 in pregnancy, 1872 prophylactic in acute liver failure, 1566 for endoscopy, 654, 654t for esophageal variceal bleeding, 1509 after liver transplantation, 1610-1611 for spontaneous bacterial peritonitis, 1534, 1535t for traveler’s diarrhea, 1874, 1875t resistance to, Helicobacter pylori infection treatment and, 843 for salmonellosis, 1863-1864, 1864t-1865t for shigellosis, 1860-1861, 1865t for small intestinal bacterial overgrowth, 1777 for small intestinal immunoproliferative disease, 457 for small intestinal obstruction, 2110 for typhoid fever, 1865t, 1866 for ulcerative colitis, 1999 for Whipple’s disease, 1841, 1841t Anticholinergics for abdominal distention and bloating, 240 for diverticulosis, 2078 for infectious diarrhea, 1886 intestinal dysmotility from, 2139 for irritable bowel syndrome, 2102 for malabsorption-related diarrhea, 72 Anticoagulation in acute mesenteric ischemia, 2034 endoscopy and, 654 hemorrhoids and, 2263 in mesenteric venous thrombosis, 2037 in portal vein thrombosis, 1379 in sinusoidal obstruction syndrome, 1377 Anticonvulsants for functional abdominal pain syndrome, 170 hepatotoxicity of, 1432 for irritable bowel syndrome, 2103 Antidepressants for functional dyspepsia, 193 hepatotoxicity of, 1436 for irritable bowel syndrome, 2102 tricyclic in chronic pancreatitis, 1004 for esophageal hypersensitivity, 703-704

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

v

vi

Index Antidepressants (Continued) for functional abdominal pain syndrome, 170 for functional dyspepsia, 193 for gastrointestinal symptoms, 349 hepatotoxicity of, 1436 intestinal dysmotility from, 2139 for irritable bowel syndrome, 2102 for neuropathic pain, 2283t taste disorders from, 354 Antidiabetic drugs, hepatotoxicity of, 1435-1436 Antidiarrheal agents for irritable bowel syndrome, 2102 for microscopic colitis, 2242 for short bowel syndrome, 1784, 1784t Antiemetics for gastric motility disorders, 812, 813t in pregnancy, 201 for vomiting, 207-208 Antiendomysial antibodies, in celiac disease, 1799, 1802-1803, 1803f, 1807, 1807t Antifungal agents, hepatotoxicity of, 1434-1435 Antigen(s) altered, in drug-induced liver disease, 1422 intestinal transport of, 141 oral tolerance to, mechanisms of, 22-23 trafficking of, across intestinal epithelial cells, 25-26 Antigen-presenting cells in IgE-mediated food allergy, 141 immune function of, 26-27, 27f Antigliadin antibodies, in celiac disease, 1798, 1802, 1807t, 1808 Antihistamines for eosinophilic gastrointestinal disorders, 434 for gastric motility disorders, 812, 813t for vomiting, 207 Antihypertensives, in liver transplant recipients, 1609-1610 Anti-inflammatory drugs nonsteroidal. See Nonsteroidal antiinflammatory drugs (NSAIDs). steroidal. See Glucocorticoids; specific drugs, e.g., Prednisone. Antimigraine drugs, for cyclic vomiting syndrome, 202 Antimitochondrial antibodies, in primary biliary cirrhosis, 1478, 1480 Antimotility drugs for infectious diarrhea, 1886 for short bowel syndrome, 1784, 1784t for traveler’s diarrhea, 1874 Antimuscarinic agents, for vomiting, 207 Antinuclear antibodies in autoimmune hepatitis, 1461, 1465 in primary biliary cirrhosis, 1478 Antinucleating proteins, in gallstone disease, 1098-1099 Antioxidants for alcoholic liver disease, 1397-1398 for chronic pancreatitis, 1005 for gastric cancer prevention, 898 for nonalcoholic fatty liver disease, 1409-1410 for Wilson disease, 1257 Antiparkinsonian drugs, intestinal dysmotility from, 2139 Antiporter, 1680-1681 Antiproliferative pathways, attenuation of, in esophageal cancer, 750 Antipsychotic agents, for gastrointestinal symptoms, 349 Antireflux barriers, 707-708, 708f Antireflux stents, for esophageal carcinoma, 766 Antireflux surgery for Barrett’s esophagus, 730 failure of, 725 for gastroesophageal reflux disease, 723f, 724-725 Antiretroviral therapy, highly active (HAART) diarrhea from, 530 esophagitis from, 525 GBV-C virus coinfection and, 1344-1346, 1345f hepatitis B and, 533, 1298 hepatitis C and, 534

Antiretroviral therapy, highly active (HAART) (Continued) hepatotoxicity of, 532, 1417, 1429-1430 impact of, 523 lactic acidosis syndrome from, 532-533 Antispasmodics for diverticulosis, 2078 for irritable bowel syndrome, 2102 Antithymocyte globulin (ATG), diarrhea from, 542 Antitoxin, botulinum, 1883 α1-Antitrypsin clearance of, in protein-losing gastroenteropathy, 442 deficiency of, 1260-1262 clinical features of, 1260-1261 diagnosis of, 1261 histopathology of, 1261 pathophysiology of, 1260 peptic ulcer disease and, 581, 864 screening for, 1261 treatment of, 1261-1262 purified, replacement therapy with, 1261-1262 Antituberculosis drugs hepatotoxicity of, 1434-1436 for intestinal tuberculosis, 1878 Anti-tumor necrosis factor antibody for alcoholic liver disease, 1397 for Crohn’s disease, 1962t, 1967-1969 dose intensification for, 1969 plus immunomodulator, 1969 structure of, 2000-2001, 2001f for ulcerative colitis, 2000-2001 Antiviral agents esophagitis from, 737 for hepatitis B, 1302-1307 hepatitis B flare from, 1297-1298 for hepatitis C, 1329 in liver transplant recipients, 1600 Antral glands, 778 Antrectomy, gastric acid secretion after, 828 Antroduodenal manometry, for gastric motility assessment, 795f, 804 Antrum, 775, 775f atresia of, 780-781, 780t electrophysiologic characteristics of, 790-791, 792f gastritis of, 846, 846f, 863 strictures of, after caustic injury, 408, 408f Anus. See also Anal; Anorectum. anatomy and physiology of, 242-243, 242f examination of, 2258-2260, 2259f imperforate, 1634f, 1635 inspection of, 2258-2259, 2259f itching localized to, 2271-2272 palpation of, 2259 radiation injury to, 650 Anxiolytics, for gastrointestinal symptoms, 349 Aortic aneurysm, abdominal, 606-607 mycotic, 607, 607f ruptured, 154t, 160 Aortic stenosis in angioectasia, 316, 595 gastrointestinal manifestations of, 584 Aortic surgery, colonic ischemia and, 2044 Aortoenteric fistula, 306, 608, 608f APACHE-II scoring system, for acute pancreatitis, 974 APC gene, 38-40 in attenuated familial adenomatous polyposis, 2184 in colorectal cancer, 2200t, 2201-2202 in colorectal tumorigenesis, 2160, 2163 in familial adenomatous polyposis, 2177-2178, 2178f, 2180-2182 in Turcot’s syndrome, 2184 Apheresis, granulocyte/monocyte, for ulcerative colitis, 2002 Aphthous stomatitis, recurrent, in celiac disease, 1805 Aphthous ulcers in Crohn’s disease, 359, 359f, 1948, 1954 oral, 356, 356f, 356t

Apical sodium-dependent bile acid transporter (ASBT), 1081, 1082f, 1083t, 1084 inhibitors of, 1088 Apnea, obstructive sleep after bariatric surgery, 118 obesity and, 105-106 Apocrine sweat glands, hidradenitis suppurativa of, 2273, 2273f Apolipoprotein(s) in bile, 1093 in gallstone disease, 1099 movement of, between HDL and chylomicrons, 1224 synthesis of, 1221-1222 Apolipoprotein A, 1704 Apolipoprotein A-IV, in satiety, 802, 1696 Apolipoprotein B, 1704 Apolipoprotein B-100, in very-low-density lipoprotein, 1223 Apolipoproteinemia, 1704 Apoptosis, 32, 33f in alcoholic liver disease, 1387 in esophageal cancer, 750 during hepatic regeneration, 1214 in radiation injury, 639 Apoptotic bodies, 1214 Apoptotic enteropathy, 217-218 Appendectomy, 2068-2069 incidence of, 2059-2060 laparoscopic, 2068 open, 2068 outcomes of, 2069 ulcerative colitis and, 1979, 2069 Appendicitis, 2059-2072 abdominal pain in, 154t, 157, 2061 abscess in, 2067 acute, 2060 versus cholecystitis, 1115 versus diverticulitis, 2079 anatomic considerations in, 2060 atypical, 2062 chronic, 2069-2070 clinical presentation in, 2061-2062 complications of, 2067 computed tomography in, 2065, 2065f in cystic fibrosis, 943 diagnosis of, 2062-2067, 2063t, 2067f accuracy in, 2066-2067 computer-aided, 2065-2066 imaging studies in, 2063-2065, 2064f-2065f laboratory studies in, 2063 in elderly, 2062 epidemiology of, 2059-2060 gangrenous, 2060 historical perspective of, 2059 in immunocompromised patients, 2062 incidence of, 2059 incidental, 2070 in infants and children, 2062 laparoscopy in, 2066 McBurney’s point in, 2060 natural orifice transluminal endoscopic surgery (NOTES) in, 2068-2069 nausea and vomiting in, 2061 obturator sign in, 2062 with palpable right lower quadrant mass, 2068 pathogenesis of, 2060-2061 pathology of, 2060 perforated, 2060, 2067-2069 peritonitis in, 2061-2062 portal pyelophlebitis in, 2067 in pregnancy, 161, 630, 2062 prophylactic, 2070 psoas sign in, 2062 radiography in, 2064 radionuclide scanning in, 2064-2065 recurrent, 2069 Rovsing’s sign in, 2062 scoring systems for, 2065-2066 treatment of, 2068-2069. See also Appendectomy. ultrasonography in, 2064, 2064f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Appendicostomy, catheterizable, for constipation, 283 Appendix adenocarcinoma of, 2070 anatomy of, 2060, 2060f carcinoid tumors of, 480-481, 481f, 486, 487f-488f, 2070 Crohn’s disease of, 1950, 2069 in cystic fibrosis, 943 diverticulitis of, 2070 embryology of, 2060 lymphoma of, 2070 mucoceles of, 2070 positions of, 2060, 2060f tumors of, 2070 vermiform, 1616 Appetite in eating disorders, 123 regulation of, 1696 by gastrointestinal peptides, 18, 18t stomach and, 802-803, 803f Appetite stimulants, for anorexia-cachexia syndrome, 2284 Apple pectin, for colonic health, 2294t APRI (AST-to-platelet ratio index), 1323, 1324t APUDomas, 492-493 Aquaporins, 1679 Aquaretics, in ascites, 1538 Argentaffin cells, 778, 1618-1619, 1620f Arginase, deficiency of, 1272f, 1273t, 1274 Argininosuccinase, deficiency of, 1271-1274, 1272f, 1273t Argon plasma coagulation for colorectal cancer, 2237 for esophageal cancer, 765 for gastric vascular ectasia, 1515-1516, 1515f for peptic ulcer bleeding, 880 Argyrophilic cells, 778 Arrhythmias, during endoscopy, 655 Arsenic, hepatotoxicity of, 1453 Arteriography. See Angiography. Arteriohepatic dysplasia, 1060-1062, 1061f1062f Arteriovenous fistula, splanchnic, portal hypertension in, 1501 Arteriovenous malformations, congenital, 606, 606f Arteriovenous shunt, in parenteral nutrition, 87 Arthralgia in Crohn’s disease, 1954 in Whipple’s disease, 1837 Arthritis in hepatitis A, 1282 in hepatitis B, 1295 reactive, gastrointestinal manifestations of, 558t, 563 rheumatoid. See Rheumatoid arthritis. in shigellosis, 1860 Arthropathy in Crohn’s disease, 1953-1954 in hereditary hemochromatosis, 1243 in shigellosis, 1860 in ulcerative colitis, 2011, 2011t Artificial bowel sphincter, for fecal incontinence, 255-256, 255t Artificial neural network instrument for lower gastrointestinal bleeding, 309 for upper gastrointestinal bleeding, 293 ASCA, in Crohn’s disease, 1959 Ascariasis clinical features of, 1922, 1922f-1923f diagnosis of, 1922, 1923f epidemiology of, 1922 gastritis in, 852 hepatobiliary, 1356t-1357t, 1360, 1922-1924, 1923f intestinal, 1921-1924, 1922f-1923f pancreatic, 1922, 1924 in children, 938-939 in pregnancy, 1922-1923 pulmonary, 1922-1923 treatment of, 1922-1924 Ascaris lumbricoides, 1921-1924, 1922f-1923f

Ascites, 1517-1542. See also Bacterascites; Peritonitis. abdominal wall hernias in, 1535 in alcoholic liver disease, 1390, 1390t ascitic fluid analysis for, 1521-1527, 1522f algorithm for, 1521, 1522f amylase in, 1525 bilirubin in, 1526 cell count in, 1522f, 1523 culture in, 1524 cytology in, 1526 exudate/transudate determination in, 1523 glucose in, 1525 gram stain in, 1525-1526 gross appearance in, 1521, 1522f lactate dehydrogenase in, 1525 neutrophil count in, 1521 polymorphonuclear leukocyte count in, 1522f, 1523 red blood cell count in, 1521 serum-ascites albumin gradient in, 1522f, 1523-1524, 1523t smear and culture for tuberculosis in, 1526 tests for, 1521-1527, 1522f, 1523t tests that are seldom helpful in, 1526-1527 total protein in, 1524-1525, 1525f triglyceride in, 1521, 1526 white blood cell count in, 1522f, 1523 biliary, 1517, 1526, 1528 cancer-related, 1517-1518, 1524-1527, 1527t, 2280 cardiac, 1518-1519, 1527 cellulitis in, 1534-1535 in chlamydial peritonitis, 1517, 1528, 1536 chylous, 1526, 1528 cirrhotic, 1517-1518, 1518f, 1524-1525. See also Cirrhosis, ascites in. clinical features of, 1518-1519 in coccidioidomycosis, 1517 complications of, 1528-1535 diagnosis of, 1519-1527 dialysis-related, 1528, 1536 differential diagnosis of, 1518t, 1522f, 1527-1528 in heart failure, 1517 in hepatocellular carcinoma, 1517-1518, 1526, 1570-1571 hepatorenal syndrome in. See Hepatorenal syndrome. history in, 1518 infection in, 1528-1534 bacteriology of, 1530, 1531t classification of, 1528-1529, 1528t clinical setting for, 1529 diagnosis of, 1531-1532, 1531t frequency of, 1530 pathogenesis of, 1529-1530, 1529f risk factors for, 1530-1531 signs and symptoms of, 1530, 1530t treatment of, 1531t-1532t, 1532-1534 in liver metastases, 1517 in lupus serositis, 1536 mixed, 1524, 1527 nephrogenic, 1528 in nephrotic syndrome, 1517, 1528 neutrocytic, 1525, 1525f culture-negative, 1528-1529 clinical setting for, 1529 pathogenesis of, 1529f, 1530 signs and symptoms of, 1530, 1530t treatment of, 1532, 1532t noncirrhotic, 1517-1518, 1518t palliative care for, 2280, 2286 pancreatic, 1517, 1527-1528, 1536 paracentesis for, 1519-1521 contraindications to, 1519 diagnostic, 1520 follow-up, 1533-1534 indications for, 1519 needle type and entry site for, 1519-1520 patient position in, 1519-1520 serial, 1538-1539

Ascites (Continued) techniques in, 1520-1521 therapeutic, 1520-1521 pathogenesis of, 1517-1518, 1518f, 1518t in peritoneal carcinomatosis, 618, 1517, 1526, 1536 physical examination in, 1518-1519 pleural effusion in, 1535 prevention of, 1534, 1535t prognosis in, 1534, 1539f, 1540 refractory, 1538-1540, 1539f in ruptured viscus, 1528, 1531 serum-ascites albumin gradient in classification by, 1522f, 1523-1524, 1523t high, 1536-1538 low, 1536 tense, 1535, 1539 treatment of, 1536-1540 aquaretics in, 1538 bed rest in, 1536 colloid replacement in, 1539-1540 diet education in, 1536 diuretics in, 1537-1538 fluid restriction in, 1536 in high albumin-gradient ascites, 1536-1538 hospitalization and, 1536 liver transplantation in, 1538 in low albumin-gradient ascites, 1536 novel methods for, 1540 outpatient, 1538 paracentesis for. See Ascites, paracentesis for. peritoneovenous shunt in, 1540 precipitating cause and, 1536 in refractory disease, 1538-1540, 1539f sodium bicarbonate in, 1538 transjugular intrahepatic portosystemic shunt for, 1540 urinary bladder catheters in, 1537 urine sodium excretion and, 1536-1537 urine sodium-potassium ratio and, 1537 in tuberculous peritonitis, 616-617, 1517-1518, 1526-1527, 1536 of unknown origin, laparoscopy in, 622 Ascorbic acid. See Vitamin C (ascorbic acid). Asialoglycoprotein receptor antibodies, in autoimmune hepatitis, 1470 Asimadoline, for functional dyspepsia, 193 L-Asparaginase, hepatotoxicity of, 1431 Aspartate aminotransferase (AST), 1229-1231 in acute pancreatitis, 973 in alcohol abuse, 1389, 1390f in alcoholic liver disease, 1390, 1391t in autoimmune hepatitis, 1468, 1469t in eating disorders, 130-131 elevation in, evaluation of, 1230-1231, 1230t, 1231f in hepatitis B, 1295 in HIV/AIDS, 535 in ischemic hepatitis, 1379, 1380f in jaundice, 331 in nonalcoholic fatty liver disease, 1405 normal values for, 1229-1230 in primary biliary cirrhosis, 1480 in sickle cell anemia, 571 in Wilson disease, 1253 Aspergillosis, gastritis in, 852 Aspiration complications of, 661 duodenal in malabsorption, 1748 in small intestinal bacterial overgrowth, 1775 in pancreatic cancer, 1022-1023 after percutaneous endoscopic gastrostomy, 90-91 risk of, in oropharyngeal dysphagia, 700 Aspiration pneumonia in enteral nutrition, 658 after upper endoscopy, 657 with vomiting, 206 Aspirin. See also Nonsteroidal anti-inflammatory drugs (NSAIDs). colorectal adenoma risk and, 2163-2164 colorectal cancer and, 2196, 2198, 2199f, 2199t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

vii

viii

Index Aspirin (Continued) after endoscopic hemostasis of peptic ulcer bleeding, 302 esophagitis from, 738 for gastric cancer prevention, 898 gastropathy from, 856 hepatotoxicity of, 1430-1431 low-dose, NSAID use and, 877 peptic ulcer bleeding and, 294-295, 303 peptic ulcer disease and, 862 AST. See Aspartate aminotransferase (AST). Asthma peptic ulcers in, 581 reflux-induced, 179-180, 714 Astrocytoma, oropharyngeal dysphagia in, 688 Astrovirus infection, 1870t, 1872 Ataxia, gluten, 1806 ATG15L1 gene, in Crohn’s disease, 1944t-1945t, 1945 Atherosclerosis, hepatic, 1382 Atorvastatin hepatotoxicity of, 1433 for nonalcoholic fatty liver disease, 1410 Atovaquone, for babesiosis, 1358 ATP-binding cassette (ABC) transporters. See ABC transporters. ATP7A (Menkes disease) gene, 1249 ATP7B (Wilson disease) gene, 1249-1251, 1251f, 1254-1255, 1255f ATP8B1 gene, 1276 Atrial-esophageal fistula, after radiofrequency catheter ablation, 739 Auerbach’s plexus. See Myenteric plexus. Autoantibodies in alcoholic liver disease, 1388 in autoimmune hepatitis, 1461, 1470 in autoimmune pancreatitis, 991 drug-induced, 1422 in enteric neuropathies, 2132 in hepatitis C, 1321 in primary biliary cirrhosis, 1478 in primary sclerosing cholangitis, 1157, 1160 in ulcerative colitis, 1979 Autocrine signaling, 5 Autocrine transmitter, 3, 4f Autoimmune disorders, in celiac disease, 1812-1813 Autoimmune enteropathy, malabsorption in, 1765 Autoimmune hepatitis. See Hepatitis, autoimmune. Autoimmune liver disease, in pregnancy, 637 Autoimmune metaplastic atrophic gastritis, 846f, 847-848 Autoimmune pancreatitis acute, 969 chronic, 986-987, 990-992, 990f, 992f, 993t Autoimmune polyendocrinopathy-candidiasisectodermal dystrophy (APECED), 1471 Autoimmune polyglandular syndrome type 1, malabsorption in, 1765, 1766f Autoimmune sclerosing cholangitis, 1467 Autoimmune thyroid disease in celiac disease, 1812 malabsorption in, 1765-1766 Autoimmunity, drug-induced, 1422 Autonomic function tests, in vomiting, 206 Autonomic nervous system, in small intestinal motility, 1647-1648 Autonomic neuropathy in amyloidosis, 1756 gastrointestinal manifestations of, 579t, 580 Autophagic-lysosomal pathway, in proteolysis, 1216 Autophagy, chaperone-mediated, 1216 Ayurvedic medicine, 2287, 2288t for irritable bowel syndrome, 2292, 2292t for liver disease, 2295t, 2297 Azathioprine for aphthous ulcers, 356 for autoimmune hepatitis, 1471-1475, 1472t, 1474f bone marrow suppression with, 1965

Azathioprine (Continued) for Crohn’s disease, 1962t, 1964-1966 hepatotoxicity of, 543, 1445 for liver transplantation, 1606t lymphoma and, 1965-1966 malabsorption with, 1757t metabolism of, 1964-1965, 1964f, 1996 oncogenicity of, 1473 for perianal Crohn’s fistula, 2269 in pregnancy, 630, 1473 side effects of, 1472, 1997-1998, 1997t teratogenicity of, 1473 for ulcerative colitis, 1996-1998, 1997t Azithromycin, for babesiosis, 1358 Azotemia, prerenal, hepatorenal syndrome and, 1546

B

B cell(s), 915 in IgE-mediated food allergy, 141 migration of, 25, 26f in mucosa-associated lymphoid tissue, 445, 446f proliferation of, in gastric MALT lymphoma, 448 B-cell lymphoma diffuse large gastric, 450f, 451-453, 452t small intestinal, 453 marginal zone (MALT) gastric, 447-451 small intestinal, 453 B-raf gene, in colon cancer, 37 BabA2 gene, in Helicobacter pylori infection, 891 Babesiosis, hepatic manifestations of, 1356t1357t, 1358 Bacillary angiomatosis hepatic infection in, 1353 and HIV/AIDS, 1353 Bacillary dysentery. See Shigellosis. Bacillary peliosis, 1381 Bacillary peliosis hepatis, in HIV/AIDS, 534 Bacillus anthracis infection gastrointestinal, 1879t-1880t, 1883-1884 oropharyngeal, 1884 Bacillus cereus infection, in food poisoning, 1879t-1880t, 1882 Bacitracin, for Clostridium difficile-associated diarrhea and colitis, 1899 Backyard barbecue syndrome, 401 Baclofen for alcohol abstinence, 1394-1395 for gastroesophageal reflux disease, 722 Bacterascites monomicrobial non-neutrocytic bacteriology of, 1530, 1531t clinical setting for, 1529 diagnosis of, 1528 pathogenesis of, 1529-1530, 1529f signs and symptoms of, 1530, 1530t treatment of, 1532, 1532t polymicrobial clinical setting for, 1529 diagnosis of, 1529 pathogenesis of, 1529-1530, 1529f signs and symptoms of, 1530, 1530t treatment of, 1532t, 1534 Bacteremia in salmonellosis, 1863, 1863t in Yersinia enterocolitica infection, 1869 Bacteria adherence factors of, 1844-1845, 1845f cytotoxins of, 1846 enterotoxins of, 1845-1846, 1845f gas-consuming, 235-236 gas-forming, pneumatosis coli and, 2249 in ileostomy effluent, 2016 intestinal (commensal) colonization resistance of, 1844 composition of, 1769-1771, 1770f-1771f in Crohn’s disease, 1943

Bacteria (Continued) culture of, 1770 diet and, 1769-1770 distribution of, 1769, 1770f host interactions with, 1771-1772 immune system interactions with, 1771 metabolic activity of, 1772, 1772t metagenomic analysis of, 1771 molecular analysis of, 1770-1771, 1771f in oral tolerance induction, 140-141 in peritonitis, 613 signaling pathways of, 1771-1772 in ulcerative colitis, 1978 mucosal invasion by, 1846 virulence factors of, 1844-1846, 1845f Bacterial cholangitis. See Cholangitis, bacterial. Bacterial infection. See also Sepsis. in acute liver failure, 1562 bile stasis and, 1104 in cirrhosis, 1530 in diarrhea classification of, 1846 irritable bowel syndrome after. See Irritable bowel syndrome, postinfection. pathogens in, 1843-1846 in food poisoning, 1878-1884, 1879t-1881t in gastritis, 849-851, 851f in glycogen storage disease type I, 1263 after hematopoietic stem cell transplantation, 549 hepatic, 1351-1355 in HIV/AIDS diarrhea, 526t-527t, 528, 529f jaundice and, 1353 in recurrent pyogenic cholangitis, 1167 transition from contamination to, factors influencing, 411, 412t Bacterial overgrowth, small intestinal, 1772-1777 achlorhydria and, 1773 anatomic abnormalities associated with, 1773 aspiration and culture in, 1775 asymptomatic, 1772 bile acids in, serum, 1777 breath tests for, 1775-1777, 1776t in chronic intestinal pseudo-obstruction, 2140 in cirrhosis, 1773 clinical features of, 1774-1775 diagnosis of, 1775-1777 diarrhea in, 223-224 in elderly persons, 1773 enteric fistula causing, 1773 etiology of, 1772-1774, 1773t in HIV/AIDS, 527t, 529 after ileocecal valve resection, 1773 in irritable bowel syndrome, 1774-1775, 2096-2098 malabsorption in, 1774 in motility disorders, 1773 nonalcoholic steatohepatitis and, 1775 in scleroderma, 1758 tests for, 1752-1753 treatment of, 1777 tropical sprue and, 1823 urinary cholyl-PABA in, 1777 xylose breath test for, 1776t, 1777 Bacterial peritonitis. See Peritonitis, bacterial. Bacteriotherapy, for Clostridium difficileassociated diarrhea and colitis, 1902. See also Probiotics. Bacteroides fragilis, in abdominal abscess, 412-413 Bacteroides spp., for Clostridium difficileassociated diarrhea and colitis, 1901 Balloon dilation, for gastric outlet obstruction, 884-885, 885f Balloon distention studies, esophageal, 699 Balloon expulsion test in constipation, 273 in fecal incontinence, 249 Balloon manometry, in fecal incontinence, 248 Balloon sphincteroplasty, for choledocholithiasis, 1191 Balloon tamponade, for variceal bleeding, 307

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Balsalazide, for ulcerative colitis, 1993-1994, 1993f, 1994t Banana, for functional dyspepsia, 2290, 2290t Band ligation endoscopic for esophageal variceal bleeding, 1509-1510, 1510f for gastric variceal bleeding, 1512 for gastrointestinal bleeding, 292 for variceal bleeding, 307-308, 1503-1504, 1503f rubber, for hemorrhoids, 2261, 2263t Bannayan-Ruvalcaba-Riley syndrome, 2185t, 2187 Barbiturates, hepatotoxicity of, 1432 Barcelona Clinic Liver Cancer staging classification, 1575, 1576f Bardet-Biedl syndrome, obesity in, 102 Bariatric surgery, 115-120 advantages and disadvantages of, 116t for binge-eating disorder, 135 comorbidity resolution after, 117-118 complications of, 117t, 118-119 digestion and absorption after, 1732-1733 efficacy of, 116-118 gallstone disease after, 1106 for gastroesophageal reflux disease, 721 malabsorption after, 1758 morbidity and mortality after, 116-117 preoperative evaluation for, 115-116 pseudoachalasia after, 693 satiety and hunger regulation by, 18 types of, 115, 116f volume effect and center of excellence movement in, 119 Barium enema in colonic obstruction, 2117, 2119f in colonic ulcer, 2244, 2244f in colorectal adenoma, 2166 in colorectal cancer, 2219, 2219f, 2221t, 2224 in constipation, 272 in diverticular disease, 2076, 2079, 2079f in Hirschsprung’s disease, 1638, 1638f in intestinal endometriosis, 2254, 2254f in intestinal tuberculosis, 1878, 1878f in lower gastrointestinal bleeding, 311 in pediatric intussusception, 2114f, 2115 in pneumatosis coli and, 2249, 2249f in ulcerative colitis, 1987-1988, 1988f Barium enteroclysis in obscure gastrointestinal bleeding, 318 in small intestinal obstruction, 2109-2110, 2112f in small intestinal tumors, 2147f, 2149 Barium esophagography in esophageal cancer, 753, 753f in esophageal diverticula, 373f, 374 in esophageal intramural pseudodiverticula, 374-375, 374f in esophageal motility disorders, 693-694, 694f in gastroesophageal reflux disease, 719 in pill esophagitis, 736, 736f in Zenker’s diverticula, 372, 372f Barium radiography in celiac disease, 1811 in constipation, 272 in Crohn’s disease, 1952f, 1956-1957 in dysphagia, 174, 176 in extraluminal duodenal diverticula, 376, 376f in gastric cancer, 900 in gastrointestinal bleeding, 292-293 in midgut volvulus, 2116, 2116f in obscure gastrointestinal bleeding, 318 in small intestinal obstruction, 2109-2110, 2112f in small intestinal tumors, 2147f, 2149, 2149f in superior mesenteric artery syndrome, 203, 203f in tropical sprue, 1827, 1827f in ulcerative enteritis, 2055, 2055f in vomiting, 204-205 in Whipple’s disease, 1836, 1837f

Barley, 1801, 1801f Barostat tests, in gastric motility assessment, 796f, 805 Barrett’s esophagus, 727-734 adenocarcinoma in, 727, 748-749 classification of, 727 diagnosis of, 727, 728f dysplasia in, 729-730, 729f, 752 endoscopic surveillance for, 730 treatment of, 731-733 endoscopic mucosal resection for, 732 endoscopy in diagnostic, 727, 728f screening, 730 therapeutic, 731 epidemiology of, 727-728 esophagectomy for, 731 gastroesophageal reflux disease in, 728, 728t, 730 Helicobacter pylori infection and, 838 management of, 730-733 neoplasia in, molecular biology of, 729, 729f neoplastic progression of, molecular events in, 751 pathogenesis of, 728, 728t recommendations for, 732-733 screening for, 756 surveillance in, 756 Bartonellosis, hepatic manifestations of, 1353 Basal cell nevus syndrome, 2187 Basaloid cancer, esophageal, 767 Basilixamab, for liver transplantation, 1605-1606 Basolateral membrane carbohydrate transfer across, 1711-1712 iron transport across, 1724-1725 protein transfer across, 1717 Bcl-2 gene, in esophageal cancer, 750 Beano, 238 Becker muscular dystrophy, intestinal pseudoobstruction in, 2138 Beclere model, in alcoholic cirrhosis, 1394 Beclomethasone for graft-versus-host disease, 545-546 for inflammatory visceral neuropathy, 2143 Bed elevation, for gastroesophageal reflux disease, 721 Bed rest, in ascites, 1536 Beef tapeworm infection, 1931-1932 Beger operation, for chronic pancreatitis, 1008 Behavioral therapy for bulimia nervosa, 134-135 for eating disorders, 131-132 for gastrointestinal symptoms, 349 for solitary rectal ulcer syndrome, 2051 Behçet’s syndrome gastrointestinal manifestations of, 558t, 563, 563f vasculitis in, 2046 Belching, repetitive, 237 Benign migratory glossitis, 354 Benign prostatic hypertrophy, inguinal hernia and, 388 Benign recurrent intrahepatic cholestasis, 328, 1085-1086, 1085t, 1276 Bentazepam, hepatotoxicity of, 1436 Bentiromide (NBT-PABA) test, for pancreatic function testing, 929 Benzamides, for vomiting, 207 Benzene, hepatotoxicity of, 1452 Benzimidazole derivatives, for vomiting, 207 Benznidazole, for Chagas’ disease, 1919 Benzocaine, methemoglobinemia from, 657 Benzodiazepines, for vomiting, 208 Benzphetamine, for obesity, 110t, 111 Berne procedure, for chronic pancreatitis, 1008 Bernstein test, in esophageal motility disorders, 699-700 Beryllium, hepatotoxicity of, 1453 Beta blockers for cyclic vomiting syndrome, 202 for esophageal variceal bleeding, 1508, 1510 hepatotoxicity of, 1433 for irritable bowel syndrome, 2103

Beta blockers (Continued) for portal hypertension, 1502 for portal hypertensive gastropathy, 602 Beta-carotene, 1721-1722, 1721f in fat malabsorption, 1751 Betaine in alcoholic liver disease, 1386-1387 in nonalcoholic fatty liver disease, 1409-1410 Bethanechol for constipation, 281 for gastroesophageal reflux disease, 721-722 Bethesda guidelines, for hereditary nonpolyposis colorectal cancer, 2206t Bevacizumab for carcinoid tumors, 489 for colonic angioectasia, 599 for colorectal cancer, 2235, 2236f for esophageal cancer, 764 for hereditary hemorrhagic telangiectasia, 601 Bezoar, 404-406 clinical presentation in, 405 diagnosis of, 405, 405f epidemiology of, 404-405, 405f, 405t treatment of, 405-406 Bianchi procedure, for short bowel syndrome, 1791, 1791f-1792f Bicarbonate secretion of, 1080 in cystic fibrosis, 943 gastric, 830-831 intestinal, 1685 pancreatic, 922f-923f, 926 secretin- or CCK-induced, 999 sodium, for ascites, 1538 Biguanides, malabsorption with, 1757t Bile, 1075-1088 cholesterol in, solubility of, 1093-1095, 1095f composition of, 1075, 1076f, 1076t, 1093 in pregnancy, 626 flow of. See also Bile acid(s), transport of. bile acid–dependent, 1080 bile acid–independent, 1080 impairment of. See Cholestasis. leak of endoscopic retrograde cholangiopancreatography for, 1191-1192, 1193f percutaneous transhepatic cholangiography for, 1188 lipids in hepatic secretion of, 1095-1097, 1096f physical states of, 1093-1094 micelles in, 1093-1095, 1095f physical chemistry of, 1093-1095, 1095f reflux of gastropathy from, 857-858 heartburn from, 714 saturation of, 1095 secretion of hepatic, 1079f, 1080-1083, 1082f, 1083t interdigestive, 926 stasis of. See Cholestasis. vesicles in, 1094 Bile acid(s) cholehepatic shunting of, 1080-1081 choleretic activity of, 1080 cholesterol, 1705-1706 circulation of, disturbances in, 1085t, 1086-1087 colorectal cancer and, 2194 conjugation of, 1078 deconjugation of, 1078 disturbances in, 1086 deficiency of, pathophysiologic mechanisms of, 1738, 1738t dehydroxylation of, disturbances in, 1086 enterohepatic circulation of, 1079-1080, 1079f disorders of, 1084-1087, 1085t. See also Cholestasis. transport proteins in, 1079f, 1081, 1082f, 1083t fecal, in bile acid malabsorption, 1753

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

ix

x

Index Bile acid(s) (Continued) functions of, 1075-1076 glucuronidation of, 1078-1079 hydroxylation of, 1078-1079 in intestinal ion transport, 1692 malabsorption of 14 C-taurocholate bile acid test for, 1753 in amyloidosis, 1756 cholestyramine therapeutic trial for, 1753 congenital, 1759t-1762t diarrhea in, 226, 229, 1087, 1755-1756 fecal bile acid output for, 1753 gastrointestinal disorders in, 1087 in microscopic colitis, 2241 primary, 1085t selenium-75–labeled homotaurocholic acid test for, 1753 in short bowel syndrome, 1782 steatorrhea in, 1755-1756 tests for, 1753 in tropical sprue, 1826 metabolism of in cystic fibrosis, 947 with ileostomy, 2016 secretion of. See Bile acid(s), transport of. serum in liver disease, 1236 in small intestinal bacterial overgrowth, 1777 synthesis of, 1076-1079, 1077f, 1077t-1078t defects in, 1084-1085, 1274-1276, 1275t therapeutic, 1087-1088 in diarrhea, 226-227 transport of canalicular, 1083 cholehepatic shunt pathway in, 1080-1081 defects in, 1085-1086, 1085t, 1275t, 1276-1277 hepatic, 1079f, 1080-1083, 1082f, 1083t sinusoidal bile acid efflux in, 1083 sinusoidal Na+-dependent bile acid uptake in, 1081 sinusoidal Na+-independent bile acid uptake in, 1081-1083 inhibitors of, 1088 intestinal, 1083-1084 receptors in, 1078, 1078t renal, 1083-1084 transporter proteins in, 1079f, 1081, 1082f, 1083t Bile acid dissolution therapy, for gallstone disease, 1121-1122, 1122t, 1123f Bile acid sequestrants, 1088. See also Cholestyramine; Colestipol; Tolevamer. for cholestasis of pregnancy, 631-632 for irritable bowel syndrome, 2102 Bile canaliculus, 1047, 1204, 1207, 1208f Bile duct(s) accessory, 1050 anatomy of, 1047-1049, 1047f-1048f arterial supply of, 1049 atresia of. See Biliary atresia. benign lesions of, 1183-1184 carcinoma of. See Cholangiocarcinoma. common anatomy of, 910, 911f, 1048, 1048f stones in. See Choledocholithiasis. compression of, jaundice in, 330 congenital anomalies of, 1050 development of, 1201 disorders of in children, 1052-1063 jaundice in, 329-330 duplication of, 1050 epithelial cells of. See Cholangiocytes. extrahepatic congenital anomalies of, 1050 development of, 1046f, 1047 inflammation of. See Cholangitis. injury to during cholecystectomy, 1128-1129, 1129f, 1134-1135, 1136f drug-induced, 1441

Bile duct(s) (Continued) percutaneous transhepatic cholangiography for, 1188-1189 innervation of, 1049 interlobular anatomy of, 1047-1048, 1047f paucity of, 1206, 1206f nonsyndromic, 1060 syndromic, 1060-1062, 1061f-1062f intrahepatic anatomy of, 1048, 1048f congenital dilatation of, 1059-1060, 1060f development of, 1046-1047 small, cholangiocarcinoma of, 1580-1582 lymphatic drainage of, 1049 obstruction of. See also Biliary strictures. bile acid circulation disturbances in, 1086-1087 in chronic pancreatitis, 1012f, 1013-1014 after hematopoietic stem cell transplantation, 547t, 549 jaundice in, 325t, 329, 334-335 after liver transplantation, 1608-1609 symptoms of, 158 perforation of, spontaneous, 1056 scintigraphy of. See Cholescintigraphy. strictures of. See Biliary strictures. vanishing, jaundice in, 328 Bile ductule(s), 1047, 1204 Bile pigments, 1093 Bile plug syndrome, 1056 Bile salt(s), 1093. See also Bile acid(s). enterohepatic circulation of, 1697-1698, 1698f micelle formation with, 1700f, 1701 secretion of, 1096-1097, 1096f solubility of, 1093-1094 synthesis of, 1096 Bile salt export pump (BSEP), 1082f, 1083, 1083t defect in, 1086, 1276 in hepatic drug elimination, 1419 Bilharziasis. See Schistosomiasis. Biliary ascites, 1517, 1526, 1528 Biliary atresia, 1052-1056 anatomic variants in, classification of, 1053-1054 clinical features of, 1054 epidemiology of, 1052-1053 extrahepatic anomalies in, 1053, 1206 pathogenesis of, 1053 pathology of, 1053-1054, 1054f prognosis in, 1055-1056, 1055f reovirus in, 1053 treatment of, 1054-1055, 1055f Biliary bypass, laparoscopic, for recurrent pyogenic cholangitis, 1169 Biliary cirrhosis, primary, 1233, 1477-1488 antimitochondrial antibodies in, 1478, 1480 antimitochondrial antibody–negative, 1487-1488 antinuclear antibodies in, 1478 asymptomatic, 1479-1480 autoantibodies in, 1478 autoimmune hepatitis with, 1466, 1467t cholestasis in, complications of, 1485-1487, 1486t clinical features of, 1479-1480, 1479t cutaneous manifestations of, 365 diagnosis of, 1480-1482, 1481f, 1481t disease associations of, 1480, 1480t epidemiology of, 1477-1478 fat-soluble vitamin deficiency in, 1486 genetic factors in, 1478-1479 histopathology of, 1480-1481, 1481f, 1481t HLA (human leukocyte antigen) phenotypes in, 1479 hyperlipidemia in, 1486 imaging studies of, 1481-1482 jaundice in, 328 liver biochemical tests in, 1480 liver biopsy in, 1481 molecular mimicry in, 1479 natural history of, 1482, 1482t-1483t osteopenic bone disease in, 1485-1486 portal hypertension in, 1498-1499

Biliary cirrhosis, primary (Continued) in pregnancy, 637 versus primary sclerosing cholangitis, 1155 prognosis in, 1482, 1482t-1483t pruritus in, 1479-1480, 1486-1487, 1486t serologic diagnosis of, 1480 steatorrhea in, 1487 symptomatic, 1479-1480, 1479t, 1482t treatment of, 1482-1485 colchicine for, 1484 combination therapy for, 1485 glucocorticoids for, 1484 liver transplantation for, 1487, 1595, 1601-1602 methotrexate for, 1484-1485 ursodeoxycholic acid in, 1483-1484, 1484f xanthelasmas in, 1486 xenobiotics in, 1479 Biliary dyskinesia. See Sphincter of Oddi, dysfunction of. Biliary pain, 157-158 acalculous, 1139-1141, 1150t. See also Cholecystitis, acute (acalculous). versus acalculous cholecystitis, 1150t cholescintigraphy in, 1141 clinical manifestations of, 1139-1140 definition of, 1139-1140 diagnosis and treatment of, 1140-1141 epidemiology of, 1140 gallbladder ejection fraction in, 1141 Meltzer-Lyon test in, 1141 pathophysiology of, 1140 calculous, 1107t, 1112-1113, 1130. See also Cholecystitis, acute (calculous). differential diagnosis of, 1113 in gallbladder dyskinesia, 1067 in sphincter of Oddi dysfunction, 1068-1069, 1068t. See also Sphincter of Oddi, dysfunction of. vomiting in, 200 Biliary pseudolithiasis, in children, 1064 Biliary sludge acute pancreatitis and, 963-964 gallstone disease and, 1091-1092 Biliary stent for benign strictures, 1193, 1195f for bile leak, 1191-1192, 1193f for cholangiocarcinoma, 1177 for distal strictures, 1189-1190, 1193-1194, 1196f for hilar biliary obstruction, 1194-1196 for malignant biliary obstruction, 1189-1190 for pancreatic cancer, 1025 in primary sclerosing cholangitis, 1192-1193, 1194f Biliary strictures. See also Bile duct(s), obstruction of. benign causes of, 1188t endoscopic therapy for, 1193, 1195f in chronic pancreatitis, 1039 distal endoscopic stent placement for, 1193-1194, 1196f percutaneous stent placement for, 1189-1190 endoscopic retrograde cholangiopancreatography for benign, 1193, 1195f indeterminate, 1193 malignant, 1193-1196, 1196f hilar, endoscopic therapy for, 1194-1196 indeterminate, 1193 after liver transplantation, 1188 malignant, 1193-1196, 1196f postoperative, percutaneous transhepatic cholangiography for, 1188 in primary sclerosing cholangitis, 1153-1154, 1155f, 1160, 1160f, 1188, 1189f, 11921193, 1194f in recurrent pyogenic cholangitis, 1167-1170, 1168f treatment of, 1134-1135, 1136f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Biliary tract. See also individual anatomic parts, e.g., Bile duct(s). anatomy of, 1047-1050, 1047f-1048f cancer of, 1171-1184. See also Ampulla of Vater, carcinoma of; Cholangiocarcinoma; Gallbladder carcinoma. in anomalous pancreaticobiliary unions/ malunions, 919 gallstone disease and, 1137 choledochal cysts of, 1205 combined percutaneous and endoscopic interventions for, 1197, 1197f in cystic fibrosis, 947-948, 1277, 1277t development of, 1045-1047, 1046f fistula of, bile acid circulation disturbances in, 1087 imaging of, 1185-1186, 1186f radiologic interventions for, 1186-1190, 1187f, 1189f-1190f scintigraphy of. See Cholescintigraphy. simple cysts of, 1205 tumors of, differential diagnosis of, 1183-1184, 1183t Biliary tract disease. See also Hepatobiliary disorders. acute, abdominal pain in, 154t, 157-158 dyspepsia in, 184 in HIV/AIDS, 534, 534f in infants and children, 1050-1052, 1051t1052t. See also Cholestasis, pediatric. after solid organ transplantation, 540, 544 in somatostatinoma, 512 Biliary tree, extrahepatic, 1203 Biliopancreatic-diversion–duodenal switch, 115, 116t-117t. See also Bariatric surgery. Bilirubin, 1227-1229. See also Hyperbilirubinemia. conjugation of, decreased, 325-326 degradation of heme to, 323 delta, 1228 direct versus indirect, 324 metabolism of, 323-326, 324f, 325t, 1227-1228 hereditary disorders of, 326t isolated disorder(s) of, 325-326, 325t-326t conjugated or mixed hyperbilirubinemia as, 326, 326t unconjugated hyperbilirubinemia as, 325-326, 326t production of, increased, 325 serum in alcoholic liver disease, 1390, 1391t in ascitic fluid, 1526 in cholangitis, 1118 in choledocholithiasis, 1116 in jaundice, 324 measurement of, 324, 1228 normal values for, 1228 in primary biliary cirrhosis, 1480 transport of, 323-324, 324f uptake of, decreased, 325 Billroth II reconstruction, for peptic ulcer bleeding, 882 Binding resins, for Clostridium difficileassociated diarrhea and colitis, 1901 Binge eating in anorexia nervosa, 124 in bulimia nervosa, 124 in eating disorders, 126t Binge-eating disorder bariatric surgery for, 135 clinical features or complications of, 127, 129t-130t diagnosis of, 122f, 125 differential diagnosis of, 125-126, 126t epidemiology of, 121 etiology of, 121-123 onset and course of, 123 pharmacotherapy for, 134 psychotherapy for, 133 treatment of, 131-137, 131f weight management for, 134-135 Bioartificial liver devices, for acute liver failure, 1568

Biofeedback training, 253-254, 253f, 254t colonic motility and, 1672 for constipation, 282-283, 2293 for fecal incontinence, 253-254, 253f, 254t for solitary rectal ulcer syndrome, 2051 Bioimpedance analysis, in nutritional assessment, 66 Biologic response modifiers, for Crohn’s disease, 1962t, 1967-1970 Biologic therapy, for ulcerative colitis, 20002001, 2023 Biologically based practices, 2287 Biomedical model, 337-338, 338f Bioprosthetic valves, for carcinoid heart disease, 483 Biopsy colonic. See Colonic biopsy. esophageal in esophageal cancer, 753-754, 754f in gastroesophageal reflux disease, 716, 717f hepatic. See Liver biopsy. rectal in Hirschsprung’s disease, 1639 in solitary rectal ulcer syndrome, 2051, 2051f small intestinal. See Small intestinal biopsy. Biopsychosocial model, 338-339, 339f abuse history in, 341 versus biomedical model, 337-338, 338f brain-gut axis in, 342-346, 343f-345f. See also Brain-gut interactions. clinical applications of, 346-348 in diagnostic decision making, 347-348 in history-taking, 347 in patient evaluation, 347, 347t coping strategies in, 342 cultural/family factors in, 340 developmental aspects in, 339 of functional abdominal pain syndrome, 165-166, 166f of irritable bowel syndrome, 2095, 2095f life stress in, 341 outcome in, 346 physiologic conditioning in, 339-340 psychological factors in, 341-342 psychosocial environment in, 340-342 social support in, 342 symptom experience and behavior and, 346 treatment approach based on healthy behavior reinforcement in, 348 patient adaptation to chronic illness in, 348 physician-related issues in, 349-350 psychopharmacologic, 348-349 psychosocial factors in, 348 psychotherapy and behavioral therapy in, 349 reassurance in, 348 therapeutic relationship in, 348 Biotin (vitamin B7), 53t-55t, 81t, 82 absorption of, 1718t, 1720 deficiency of, cutaneous manifestations of, 368-369 reference nutrient intake for, 1718t Bipolar neurons, in gastrointestinal tract, 5 Bisacodyl colonic ischemia from, 2040 colonic motility and, 1671-1672 for constipation, 277t, 279-280 BISAP scoring system, for acute pancreatitis, 974-975 Bismuth for functional dyspepsia, 193 for peptic ulcer disease, 871-872 Bismuth carbomer enema, for pouchitis, 2010 Bismuth-Corlette classification of hilar cholangiocarcinoma, 1171, 1172f Bismuth subsalicylate for diarrhea, 226 for Helicobacter pylori infection, 842, 842t for infectious diarrhea, 1886 for microscopic colitis, 2242 prophylactic, for traveler’s diarrhea, 1874, 1875t for traveler’s diarrhea, 1874

Bisphosphonates esophagitis from, 737-738 gastropathy from, 856 for osteoporosis, in primary biliary cirrhosis, 1485-1486 peptic ulcer disease and, 862 Black cohosh, hepatotoxicity of, 1458 Blastocystis hominis infection, 1914 Blatchford Score for upper gastrointestinal bleeding, 293 Bleeding in colonoscopy and sigmoidoscopy, 658-659, 659f diverticular, 2085-2088, 2086f-2088f in diverticulosis. See Diverticulosis, colonic, bleeding in. gastrointestinal. See Gastrointestinal bleeding. intraperitoneal, 620 mesenteric, 620 in peptic ulcers. See Peptic ulcer bleeding. postsphincterotomy, 1197 rectal. See also Gastrointestinal bleeding, lower. in Dieulafoy’s lesions, 314 in Meckel’s diverticulum, 1628-1629 in rectal ulcers, 314-315, 315f in ulcerative colitis, 1983 retroperitoneal, 620 variceal. See Varices, bleeding from. Bloating in celiac disease, 1805 in eating disorders, 128-130 intestinal gas causing, 238-240, 239f in malabsorption, 1745t Blood culture of, in typhoid fever, 1866 in stool, 218-219, 222. See also Fecal occult blood test (FOBT). transfusion of. See Blood transfusion. Blood alcohol test, 1389 Blood banks, hepatitis C diagnosis in, 1322 Blood flukes, 1935-1939, 1936f-1937f, 1938t. See also Schistosomiasis. Blood pressure. See also Hypertension; Hypotension. in obesity, 107 Blood tests in acute pancreatitis, 971 in celiac disease, 1810-1811 in Chagas’ disease, 1918 in malabsorption, 1742, 1743t Blood transfusion for gastrointestinal bleeding, 288 hepatitis C and, 1317 jaundice from, 325 Blood urea nitrogen (BUN) in acute pancreatitis, 974 in gastrointestinal bleeding, 288 Bloom’s syndrome, 2177t, 2184 Blue diaper syndrome, malabsorption in, 1759t-1762t Blue rubber bleb nevus syndrome, 605-606, 605f cutaneous manifestations of, 361, 361f gastrointestinal bleeding in, 317 Blunt objects, as foreign bodies, 402-404 Bmp signaling in hereditary hemochromatosis, 1241 in intestinal development, 1618 BMPR1A gene, in juvenile polyposis, 2187 Boceprevir, for hepatitis C, 1334 Bochdalek hernia clinical manifestations and diagnosis of, 381 etiology and pathogenesis of, 379, 380f incidence and prevalence of, 380-381 treatment and prognosis in, 383 Body compartment analysis, in nutritional assessment, 65, 66t Body composition, protein-energy malnutrition and, 61, 61f, 61t Body image distorted, in anorexia nervosa, 124

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xi

xii

Index Body image (Continued) excessive concern with in bulimia nervosa, 125 in eating disorders, 126t Body mass index in anorexia nervosa, 126 energy requirements for hospitalized patients and, 49, 50t gastroesophageal reflux disease and, 706-707 mortality and, 103-104 in nutritional assessment, 65, 65t-66t, 77 in obesity, 100, 106 Body stuffers or packers, 404 Body weight. See Weight (body). Boerhaave’s syndrome, 206, 740-741 Bone disease, metabolic. See also Osteomalacia; Osteopenia. in Crohn’s disease, 1954 Bone loss. See also Osteoporosis. alveolar, in HIV/AIDS, 355 in eating disorders, 128 Bone marrow in protein-energy malnutrition, 64 in Shwachman-Diamond syndrome, 953-954 transplantation of. See Hematopoietic stem cell transplantation. in typhoid fever, 1866 Bone mineral density in Crohn’s disease, 1954 in ulcerative colitis, 2012 Bone scintigraphy, in celiac disease, 1811 Borborygmi, in eating disorders, 128-130 Borchardt’s triad, 384 Bosentan, for portopulmonary hypertension, 1552 Botanical hepatotoxins, 1454-1455, 1457t Botulinum toxin, 1883 for achalasia, 701 for anal fissure, 2265t, 2266 for constipation, 281 for gastric motility disorders, 812, 813t Botulism food-borne, 1879t-1880t, 1882-1883 infant, 1883 Bougie dilators, for distal esophageal spasm, 703 Bourneville’s disease. See Tuberous sclerosis. Bouveret’s syndrome, 1119 Bowel gas. See Gas, intestinal. Brachytherapy for esophageal cancer, 767 esophageal radiation injury and, 640-641 Bradycardia, during endoscopy, 655 BRAF gene, in colorectal cancer, 2204 Brain disorders of. See Neurologic disorders. functional imaging of, in functional abdominal pain syndrome, 167-168 lesions of, in Whipple’s disease, 1839-1840 tumors of, in Turcot’s syndrome, 2184 Brain-gut interactions, 342-346 cytokines and, 346 in digestion and absorption, 1695-1698, 1696f, 1698f disease susceptibility and, 345-346 in food intake, 802-803, 803f in functional abdominal pain syndrome, 165-168, 166f-167f during illusory self-motion, 802 immune function and, 345-346 neurotransmitters in, 343 stress and, 342-343 in visceral pain regulation, 343-345, 343f-345f Brain-gut-liver axis, in glucose homeostasis, 1217 Brainerd diarrhea, 231 Bran. See also Fiber, dietary. colonic transit time and, 262-263 BRCA1 gene, in pancreatic cancer, 1017 Bread tabs or bread bag clips, ingestion of, 403-404 Breast cancer, obesity and, 105 Breast-feeding Crohn’s disease and, 1946 enteric infection and, 1844 hepatitis C from, 1318

Breast-feeding (Continued) obesity and, 102 oral tolerance and, 141-142 Breast milk fat content of, 1729 lactose in, 1732 oral tolerance and, 141-142 proteases in, 1729-1730, 1732 Breath, gas excretion in, 236 Breath test alcohol, 1389 aminopyrine, 1235 carbon for fat malabsorption, 1751 for malabsorption, 1754 for protein malabsorption, 1752 in small intestinal motility assessment, 1653 for fat malabsorption, 1751 in gastric motility assessment, 804 hydrogen for carbohydrate malabsorption, 1751 for fructose malabsorption, 1755 glucose, for small intestinal bacterial overgrowth, 1776, 1776t for lactose malabsorption, 1751, 1754-1755, 1755f lactulose for small intestinal bacterial overgrowth, 1776, 1776t in small intestinal motility assessment, 1653 for small intestinal bacterial overgrowth, 224 for small intestinal bacterial overgrowth, 1775-1777, 1776t in small intestinal motility assessment, 1652-1653 in tropical sprue, 1827 urea, for Helicobacter pylori infection, 840t, 841, 841f xylose, for small intestinal bacterial overgrowth, 1776-1777, 1776t Bristol Stool Scale, 264, 265f, 2092, 2093f Bromfenac, hepatotoxicity of, 1437 Bromoalkanes, hepatotoxicity of, 1452 Bronchitis, chronic, peptic ulcers in, 581 Brooke ileostomy. See Ileostomy, Brooke. Brucellosis, hepatic manifestations of, 1353 Brunner’s glands, 779-780, 1621, 1621f Brush border membrane carbohydrate transfer across, 1708, 1708f, 1708t hydrolases of, 1708, 1708f, 1708t protein transfer across, 1713-1715, 1713f, 1714t triglyceride transfer across, 1702-1703, 1703f Bruton’s agammaglobulinemia, malabsorption in, 1764 BSEP. See Bile salt export pump (BSEP). BSEP disease, 1086, 1276 Budd-Chiari syndrome, 1371-1375 acute, 1373 versus alcoholic liver disease, 1391 chronic, 1373 clinical features of, 1372-1373 diagnosis of, 1373-1374 etiology of, 1371-1372, 1372f, 1372t fulminant, 1373 liver transplantation for, 1603 membranous obstruction of inferior vena cava in, 1371-1372, 1372f, 1374 pathology of, 1373, 1373f in pregnancy, 637-638 subacute, 1373 treatment of, 1374-1375, 1374f Budesonide for celiac disease, 1818 for collagenous colitis, 228 for Crohn’s disease, 1964 for eosinophilic gastrointestinal disorders, 434 for microscopic colitis, 2242 for primary biliary cirrhosis, 1484 for primary sclerosing cholangitis, 1164-1165 topical, for ulcerative colitis, 1996 for ulcerative colitis, 1995-1996

Buerger’s disease, vasculitis in, 2046 Bulimia nervosa chronic intestinal pseudo-obstruction in, 2140 clinical features or complications of, 127, 129t-130t constipation in, 271 diagnosis of, 122f, 122t, 124-125 differential diagnosis of, 125-126, 126t epidemiology of, 121 etiology of, 121-123 gastrointestinal manifestations of, 128-131, 129t-130t, 135-137, 136t laboratory evaluation of, 128 onset and course of, 123 pharmacotherapy for, 133 psychotherapy for, 132-133 rumination in, 204 treatment of, 131-137, 131f weight management for, 134-135 Bupropion, for obesity, 112 Burkholderia pseudomallei infection, hepatic manifestations of, 1353 Burkitt’s lymphoma, small intestinal, 454-455, 455f Burns acute pancreatitis after, 969 electrosurgical in endoscopy, 656 in polypectomy, 660 Buschke-Lowenstein tumors, 2271 “Bush tea,” hepatotoxicity of, 1456 Butane, hepatotoxicity of, 1452 Butyrate, topical, for ulcerative colitis, 2000 Butyrophenones, for vomiting, 207 Byler’s disease, 1085-1086, 1276 Byler’s syndrome, 1276 Bypass procedures biliary, for recurrent pyogenic cholangitis, 1169 bowel bypass syndrome after, 360 cardiopulmonary, pancreatitis after, 968 gastric diet after, 95 malabsorption after, 1758 Roux-en-Y, 115, 116t-117t. See also Roux-en-Y gastric bypass. jejunoileal, intestinal adaptation to, 1783 for radiation enteritis, 646-647

C

C/EBP, in hepatic regeneration, 1213 C-met, in hepatic regeneration, 1214 c-Myc protein, in gastrointestinal tumors, 37 C-peptide, in insulinoma, 497-498 C-reactive protein in acute pancreatitis, 975 in appendicitis, 2063 C282Y homozygosity clinical penetrance of, 1243t in hereditary hemochromatosis, 1242, 1243t, 1244 CA 19-9 marker in cholangiocarcinoma, 1162, 1173 in intrahepatic cholangiocarcinoma, 1581 in pancreatic cancer, 1023 CAA (circulating anodic antigen), in schistosomiasis, 1939 Cachexia cancer, 84 palliative care for, 2282-2284 E-Cadherin in gastric cancer, 893-894 in metastasis, 42 Cadherin(s), in intestinal epithelium, 1676f, 1679 Cadmium, hepatotoxicity of, 1453 Caffeine diarrhea from, 217-218 fecal incontinence and, 252 Caffeine clearance test, 1235 Cag pathogenicity island, in Helicobacter pylori infection, 835, 891 CagA, in Helicobacter pylori infection, 836

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Calbindin D9K, 1723 Calbindin D28K, 1723 Calcitonin gene-related peptide (CGRP), in postoperative ileus, 2125 Calcium. See also Hypercalcemia; Hypocalcemia. absorption of, 1722-1723, 1723f adaptive changes in, 1728, 1729f in short bowel syndrome, 1781t, 1782, 1784 colonic salvage of, 1723, 1741 dietary, 52t, 80 aging and, 58 colorectal cancer and, 2195-2197, 2199f, 2199t sources of, 1722 in gallstone disease, 1099 in intestinal ion transport, 1693, 1693f malabsorption of, 1739-1740 after gastric resection, 1757 in metabolic bone disease, 1766 in pancreatic enzyme secretion, 925, 925f in short bowel syndrome, 1786t supplemental in celiac disease, 1814 for cholestasis in children, 1062 in Crohn’s disease, 1970 for malabsorption, 74 in primary biliary cirrhosis, 1485-1486 in short bowel syndrome, 1784 in trypsin regulation, 933 Calcium-activated chloride (CLCA) channels, 1684 Calcium channel blockers for achalasia, 701 colorectal adenoma risk and, 2164 hepatotoxicity of, 1433 for hypertension, in liver transplant recipients, 1609-1610 intestinal dysmotility from, 2139 topical, for anal fissure, 2265t, 2266 Calcium channels, in colonic motility, 1661 Calcium citrate, in ulcerative colitis, 2012 Calcium oxalate kidney stones, in short bowel syndrome, 1790, 1790f Calcium polycarbophil, for constipation, 276 Calcium provocation test, in gastrinoma, 503 Calcium-sensing receptor gene, in pancreatitis, 934-935, 934f Calcium transients, in small intestinal motility, 1654 Calculi. See Stones. Calicivirus infection, 1870t, 1871 Calipers, handheld, in nutritional assessment, 78f Caloric assessment disease severity and, 85, 85t in nutritional assessment, 78-79, 79t Calories, requirements for. See Energy requirements. Calorimetry, indirect, in nutritional assessment, 79, 79f Calprotectin, fecal, in NSAID enteropathy, 2053 Cameron ulcer, in sliding hiatal hernia, 381, 381f, 383 Cameron’s lesions, upper gastrointestinal bleeding in, 305, 305f Campylobacter jejuni infection, small intestinal immunoproliferative disease and, 455 Campylobacter spp. infection, 1867-1868 clinical features of, 1867-1868, 1868f diagnosis of, 1868 in elderly, 1876 epidemiology of, 1867 in food poisoning, 1879t-1880t in HIV/AIDS diarrhea, 527t, 529 in pregnancy, 1872 in traveler’s diarrhea, 1873 treatment of, 1865t, 1868 Canaliculus, bile, 1047 Canals of Hering, 1047, 1204 Cancer. See also Neoplasia; specific sites and types. anorexia-cachexia syndrome in, 2282 ascites in, 1517-1518, 1524-1527, 1527t, 2280, 2286

Cancer (Continued) in celiac disease, 1818-1819 chemotherapy for. See Chemotherapy. Crohn’s disease and, 1972-1973 cystic fibrosis and, 945 de novo, in liver transplant recipients, 1610 development of chemical carcinogenesis in, 41 clonal expansion in, 34 dietary factors in, 41-42 environmental agents, 41-42 genomic instability in, 34-35, 35f multistep nature of, 34-35, 35f signaling pathways in, 40-41, 41f gallstone disease and, 1105-1106, 1137 intestinal obstruction in, 2285 metastatic. See Metastatic disease. nutritional therapy for, 84 obesity and, 105 pain in, complementary and alternative medicine for, 2298 palliative care in. See Palliative care. percutaneous endoscopic gastrostomy for, 90 portal hypertension and, 1501 prognostication guidelines for, 2278, 2280t radiation therapy for. See Radiation therapy. salmonellosis and, 1862-1863 Cancer stem cell model, for gastric cancer, 895 Candesartan, hepatotoxicity of, 1433 Candidiasis abdominal abscess in, 418 esophageal squamous cell carcinoma and, 747-748 esophagitis in, 354-355, 741-742, 742f gastritis in, 851 hepatic, 1365, 1365f in HIV/AIDS, 524, 526 oral, 354-355, 354f in HIV/AIDS, 355 Cannabinoid(s) endogenous, in alcoholic liver disease, 1389 for vomiting, 208 Cannabinoid CB1 antagonists, for obesity, 111 Cannabinoid CB1 receptors, in emetic reflex, 197-198 Cannabis abuse cyclic vomiting syndrome and, 202 hepatitis C and, 1325-1326 Cantlie line, 1202 Capecitabine for colorectal cancer, 2232, 2234-2235 for esophageal cancer, 764 Capillaria (Paracapillaria) philippinensis, 1925 Capillariasis, 1925 hepatic, 1356t-1357t, 1360, 1360f intestinal, malabsorption in, 1829-1830, 1829t Capsaicin, for functional dyspepsia, 2290, 2290t Capsule endoscopy. See Endoscopy, capsule. Capsule technology, in gastric motility assessment, 797f, 804 Captopril, hepatotoxicity of, 1433 Caraway, for functional dyspepsia, 2290-2291, 2290t Carbamazepine granulomatous hepatitis from, 1437t malabsorption with, 1757t Carbamyl phosphate synthase, deficiency of, 1271-1274, 1272f, 1273t Carbohydrases, brush border membrane, 1708, 1708f, 1708t Carbohydrate, 1706-1712 dietary, 51, 79, 1706-1707 intake of, 1706-1707 restriction of, 108-109 digestion and absorption of, 1707-1712 amylase in, 1707-1708, 1707f brush border membrane hydrolases in, 1708, 1708f, 1708t disaccharidase in, 1708-1710, 1709f exit from epithelium in, 1711-1712 mucosal defects of, 1739 mucosal transport in, 1710-1711, 1710f in neonates and infants, 1732

Carbohydrate (Continued) gastric emptying of, 800, 800f hydrolysis of, defective, 1739 malabsorption of colon in, 1740-1741, 1740f-1741f congenital disorders of, 1759t-1762t diarrhea from, 224 intestinal gas and, 234-235, 234t, 237 intestinal ion transport in, 1689 mechanisms of, 1739 in small intestinal bacterial overgrowth, 1774 tests for, 1751-1752 metabolism of in cirrhosis, 1219-1220 colonic, 1689, 1711-1712 hepatic, 1217-1220, 1218f Carbohydrate antigen, in gastric cancer, 900 Carbohydrate-deficient transferrin, in alcohol abuse, 1389, 1390f Carbon breath test for fat malabsorption, 1751 for malabsorption, 1754 for protein malabsorption, 1752 in small intestinal motility assessment, 1653 Carbon dioxide, in intestinal gas, 233, 235 Carbon tetrachloride, hepatotoxicity of, 1450-1451, 1451t Carboxyl ester lipase, 956, 1702 Carboxylesterase, 924 Carboxypeptidases, 1712-1714, 1713f, 1713t-1714t Carcinoembryonic antigen (CEA) in ascites, 1526 for cancer detection, 44 in colorectal cancer, 2193-2194, 2216, 2230-2231 in gastric cancer, 900 Carcinogenesis. See also Cancer, development of. chemical, 41 in colorectal cancer, 2160-2161, 2161f, 2195, 2200-2204, 2200t, 2201f-2203f, 2204t Carcinoid crisis, 484 Carcinoid syndrome, 482-484 cardiac involvement in, 482-483, 483f diarrhea in, 482 flushing in, 482, 482f, 484 pancreatic endocrine tumors in, 515 pathophysiology of, 483-484, 483f Carcinoid tumors, 475-490 ampullary, 479-480 appendiceal, 480-481, 481f, 486, 487f-488f, 2070 biliary, 1183 biochemical markers of, 484-485 classification of, 475, 476t, 478-482, 479t clinical presentation in, 476, 477f colonic, 481 cutaneous, 365 diagnosis of, 484-486 duodenal, 477-480, 479t-480t esophageal, 478 gastric, 478-479, 479t, 486-488, 487f-488f, 904, 905f ileal, 480, 480t incidence of, 475 jejunal, 480, 480t liver metastasis of, 488 localization of, 485-486, 485f-486f molecular genetics of, 478 molecular pathogenesis of, 494 nomenclature for, 493 pancreatic, 479, 515. See also Pancreatic endocrine tumors. pathology of, 476-478, 477f race and, 475-476 rectal, 481-482, 486, 487f-488f sites of, 475-476 small intestinal, 480, 480t, 486, 487f-488f, 2145 treatment of, 486-489, 487f-488f chemotherapy in, 488 hepatic artery embolization and chemoembolization in, 488

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xiii

xiv

Index Carcinoid tumors (Continued) interferon-alpha in, 489 new biologic agents in, 489 peptide receptor radionuclide therapy in, 489 somatostatin analogs in, 489 surgical, 486-488, 487f-488f Carcinoma. See also Adenocarcinoma. ampullary. See Ampulla of Vater, carcinoma of. biliary. See Cholangiocarcinoma. in celiac disease, 1819 colorectal. See Colorectal cancer. esophageal. See Esophageal cancer. gallbladder. See Gallbladder carcinoma. gastric. See Gastric cancer. hepatic. See Hepatocellular carcinoma. pancreatic. See Pancreatic cancer. small intestinal. See Small intestinal adenocarcinoma. squamous cell, esophageal. See Esophageal cancer, squamous cell. Carcinomatosis, peritoneal, 618-619 ascites in, 1517, 1526, 1536 clinical features of, 618 prognosis in, 619 pseudomyxoma peritonei, 619 tissue of origin in, 618f treatment of, 618-619 intraperitoneal chemotherapy for, 619 in ovarian versus nonovarian cancer, 619 paracentesis for, 618-619 surgery for, 619 Carcinosarcoma, esophageal, 767 Cardiac. See also Heart. Cardiac ascites, 1518-1519, 1527 Cardiac cirrhosis, in congestive hepatopathy, 1380, 1381f Cardiac gland, 776 Cardiac output, in hepatorenal syndrome, 1547 Cardiomyopathy cirrhotic, 1553-1554, 1553f gastrointestinal manifestations of, 584 Cardiovascular system in protein-energy malnutrition, 63 in Whipple’s disease, 1837 Carditis, 848 Carney-Stratakis syndrome, gastrointestinal stromal tumors and, 473-474 Carney triad, gastrointestinal stromal tumors in, 473-474 Caroli’s disease, 1058-1059, 1058f, 1060f hepatic, 1590 portal hypertension in, 1500 Caroli’s syndrome, 1059 Carotene, 1722 Carvedilol, hepatotoxicity of, 1433 Cascara sagrada, for colonic health, 2294t Caspases activation of, in regulation of apoptosis, 32, 33f, 1214, 1420 inhibition of, 1420 CASR (calcium-sensing receptor) gene, in pancreatitis, 934-935, 934f Castleman’s disease of mesentery, 620 Castor oil, for constipation, 277t, 279 Catecholamines in gastrointestinal tract, 11 in intestinal ion transport, 1690 Cathartic colon, 2245-2246 Catheter central venous complications of, 87 peripherally inserted, in parenteral nutrition, 87 Foley, for foreign body removal, 400 infection of, 87, 1788f occlusion of, 1789f percutaneous, for pancreatic pseudocyst drainage, 981 tunneled Silastic, in parenteral nutrition, 87 urinary bladder, in ascites, 1537

Cauda equina, damage to, fecal incontinence in, 244 Caustic injury, 406-408 acidic, 406, 407f alkali, 406, 406f, 408 clinical features of, 406 complications of, 408, 408f diagnosis of, 407, 407f, 407t epidemiology of, 406 esophageal squamous cell carcinoma in, 747-748 pathophysiology of, 406, 406f-407f treatment of, 407-408 Caveolar internalization, 1210 Cavernous hemangioma, 604, 605f hepatic, 1586-1587, 1586f rectal, 604, 605f CCA (circulating cathodic antigen), in schistosomiasis, 1939 CCL5, 29 CCL20, 30 CCL25, 29 CCR9, 29 CD4+ count of, in AIDS definition, 523 in Helicobacter pylori infection, 891 in oral tolerance, 23 CD8+, in oral tolerance, 23 CD25, in oral tolerance, 23 CD45, tyrosine phosphatase region on, 15 CD117, in gastrointestinal stromal tumors, 463 Cecostomy, percutaneous, for acute colonic pseudo-obstruction, 2130 Cecum, 1616 anatomy of, 2060 carcinoma of, 2211, 2211f, 2219, 2219f diverticulitis of, versus appendicitis, 2063t ulcer of, 2244 volvulus of, 2116-2117, 2118f Cefixime, for Whipple’s disease, 1841t Cefotaxime, for ascitic fluid infection, 1532t, 1533 Ceftriaxone gallstone disease from, 1092 in children, 1064 prophylactic, for spontaneous bacterial peritonitis, 1534, 1535t for Whipple’s disease, 1841t Cefuroxime, prophylactic, in acute liver failure, 1566 Celandine, for functional dyspepsia, 2290, 2290t Celecoxib cardiovascular risk of, 875 in colorectal cancer, 2198 hepatotoxicity of, 1432, 1437 for NSAID ulcer prophylaxis, 874-875 Celiac artery compression syndrome, 608-609, 609f Celiac axis anastomotic circulation of, 2028, 2029f anatomy of, 2027, 2028f Celiac disease, 1797-1820 acute pancreatitis in, 969 in adults, 1804 anemia in, 1805-1806 antiendomysial antibodies in, 1799, 1802-1803, 1803f, 1807, 1807t antigliadin antibodies in, 1798, 1802, 1807t, 1808 atypical, 1797, 1798f autoimmune disorders in, 1812-1813 carcinoma in, 1819 in children, 1804 chronic intestinal pseudo-obstruction in, 2139 clinical features of, 1798f, 1804-1807, 1805t collagenous sprue in, 1817 colonic lymphocytosis in, 1813 definitions in, 1797 dermatitis herpetiformis and, 367-368, 368f, 1811-1812 diabetes mellitus in, 1812 diagnosis of, 1807-1809 antiendomysial antibody in, 1807, 1807t antigliadin antibodies in, 1807t, 1808

Celiac disease (Continued) capsule endoscopy in, 1810 DGP testing in, 1808 genetic testing in, 1809 gluten challenge in, 1810 hematologic and biochemical tests for, 1810-1811 physical examination in, 1806-1807 radiology in, 1811 serologic tests in, 1807-1809, 1807t, 1809f small intestinal biopsy in, 1799-1800, 1799f, 1803f, 1809-1810, 1810f stool examination in, 1810 tissue transglutaminase antibodies in, 1807-1808, 1807t differential diagnosis of, 1811-1813, 1812t disaccharidase deficiency in, 1812-1813 diseases associated with, 1811-1813, 1812t enteropathy-associated T cell lymphoma in, 457-458, 458f, 1798, 1818 environmental factors in, 1800-1802, 1801f eosinophilia in, 433 epidemiology of, 1798-1799, 1798f extraintestinal manifestations of, 1805-1807, 1805t as gastrointestinal food allergy, 145-146 gastrointestinal manifestations of, 1804-1805 genetic factors in, 1802, 1803f geographic variation in, 1799 gliadin in, 1798, 1801-1802, 1808 gynecologic problems in, 1806 history of, 1797-1798 HLA-DQ2/DQ8 haplotypes in, 1798, 1802, 1803f, 1809 hyposplenism and splenic atrophy in, 1813 immune factors in, 1802-1804, 1803f immunoglobulin A deficiency in, selective, 1813 in infants, 1804 infertility in, 1806 inflammatory bowel disease in, 1813 interleukin-15 in, 1798, 1804 intraepithelial lymphocytes in, 1798, 1800, 1803-1804 iron deficiency anemia in, 321 irritable bowel syndrome and, 2100 latent, 1797, 1798f lymphoma in, 1818 malabsorption in, 1829t, 1830 malignant disease in, 1818-1819 microscopic colitis and, 228, 1813, 2241 neurologic symptoms in, 1806 nonresponsive, 1815-1816, 1815f nutritional therapy for, 84 osteopenia in, 1806, 1814-1815 pathogenesis of, 1800-1804, 1803f pathology of, 1799-1800, 1799f prognosis in, 1819 refractory, 1797, 1798f, 1816-1818, 1816f1817f, 2054-2058 background on, 2054 clinical presentation in, 2055 complications of, 1818 definition of, 2054 diagnosis of, 2055-2056, 2055f-2058f pathology of, 2056, 2058f treatment of, 1817-1818, 2056-2058, 2057f screening for, 1808 silent, 1797-1799, 1798f, 1808 small intestinal bacterial overgrowth in, 1773 small intestinal tumors in, 2148 spectrum of, 1797, 1798f thyroid disease in, autoimmune, 1812 tissue transglutaminase antibodies in, 1798, 1807-1808, 1807t tissue transglutaminase in, 1802-1803, 1803f treatment of, 1813-1815, 1814t future directions in, 1819-1820 glucocorticoids for, 1815, 1817 gluten-free diet for, 1813-1814, 1813t immunosuppressive therapy for, 1817 for refractory disease, 1817-1818 supplemental therapy for, 1814-1815

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Celiac disease (Continued) ulcerative jejunoileitis and, 1817-1818, 1817f. See also Ulcerative enteritis. villous atrophy in, in celiac disease, 17991800, 1799f Celiac plexus block, for chronic pancreatitis, 1009 Celiac plexus neurolysis, for chronic pancreatitis, 1009 Cell cycle phases of, 31, 32f regulation of, 31, 32f Cell cycle genes, in hepatic regeneration, 1213 Cell death apoptotic (programmed), 32, 33f, 1420, 1421f in drug-induced liver disease, 1420-1421, 1421f necrotic, 1420-1421, 1421f oxidative stress and, 1421 Cell proliferation, normal cell cycle in, 31-32, 32f programmed cell death (apoptosis) and senescence in, 32, 33f signaling pathways regulating, 32-33, 34f Cellophane tape test, for pinworm infection, 1929 Cellular communication, in gastrointestinal tract, 3-5, 4f, 4t Cellular oncogenes, 35 Cellulase, for bezoar dissolution, 405 Cellulitis, in ascites, 1534-1535 Center of excellence movement, in bariatric surgery, 119 Central adiposity Barrett’s cancer and, 728 disorders related to, 104-105 gastroesophageal reflux disease and, 707 in obesity, 100, 103, 103t, 106-107 Central nervous system. See also Brain. in colonic motility, 1664 in functional abdominal pain syndrome, 165-168, 166f-167f in irritable bowel syndrome, 2098 in schistosomiasis, 1937 in small intestinal motility, 1648 visceral pain transmission to. See Afferent (ascending) visceral pain transmission. in Whipple’s disease, 1836-1837 Central pontine myelinolysis, after liver transplantation, 1608, 1608t Central venous catheter complications of, 87 peripherally inserted, in parenteral nutrition, 87 Centroacinar cells, 912, 913f-914f, 914, 921 Cephalosporins, for typhoid fever, 1867 Cereal grains, taxonomic relationships among, 1801, 1801f Cerebral edema, in acute liver failure, 1561-1562, 1561t, 1565-1566 Cerebrotendinous xanthomatosis, 1085, 1274-1275 Certolizumab pegol, for Crohn’s disease, 1967-1969 Ceruloplasmin in copper transport, 1249 in Wilson disease, 1253, 1253t Cervical spinal cord injury, gastrointestinal problems after, 579-580 Cestode (tapeworm) infection. See also specific disorders, e.g., Diphyllobothrium spp. infection. beef/pork, 1931-1932 dog, 1933 dwarf, 1932-1933 fish, 1931, 1931f hepatic, 1356t-1357t, 1363-1365 intestinal, 1931-1933, 1931f rodent, 1932-1933 Cetuximab for colorectal cancer, 2235-2236, 2236f for esophageal cancer, 764 Cevilimine, xerostomia, 353 CFTR gene in chronic pancreatitis, 987-990 in cystic fibrosis, 934f, 935-937, 936t, 952

CFTR gene (Continued) mutations of, classification of, 936, 936t regulation of, 935-936 structure of, 935, 935f Chagas’ disease, 1918-1919 versus achalasia, 692-693 chronic intestinal pseudo-obstruction in, 2139 clinical presentation in, 1918, 1919f constipation in, 270 diagnosis of, 1918-1919 differential diagnosis of, 692-693 epidemiology of, 1918 esophagitis in, 743 pathogenesis and pathology of, 1918 prevention and control of, 1919 treatment of, 1919 Chamomile, for colonic health, 2294t Chaparral hepatotoxicity of, 1457 for postoperative ileus, 2295 Chaperones autophagy mediated by, 1216 in protein folding, 1216 Charcoal, activated, for odoriferous gas, 238 Charcot-Marie-Tooth disease, gastrointestinal manifestations of, 581 Charcot’s triad in cholangitis, 1118 in recurrent pyogenic cholangitis, 1168 Chaso, hepatotoxicity of, 1458 Cheilitis, granulomatous, 359 Chemical agents, hepatotoxicity of, 1450-1453, 1450t-1451t Chemical carcinogenesis, 41. See also Oncogenesis. Chemical colitis, 2247-2248, 2247t, 2248f Chemical messengers of gastrointestinal tract hormonal, 6-10, 6f, 8f nonhormonal, 10-13, 12f Chemoembolization hepatic artery for carcinoid tumors, 488 for pancreatic endocrine tumors, 521 for hepatocellular carcinoma, 1579 liver transplantation and, 1602 Chemokines in gastrointestinal tract, 13 in gut-associated lymphoid tissue, 29-30 Chemoprevention, of colorectal cancer, 2196-2199, 2197f, 2199f, 2199t Chemoradiotherapy for anal canal cancer, 2270 for colorectal cancer adjuvant, 2233-2234, 2234f neoadjuvant, 2234 for esophageal cancer, 763-764 for gastric cancer, 903-904 intensity-modulated, 651 for pancreatic cancer, 1025-1026, 1026f radiation injury after colonic, 647-650, 648f esophageal, 640-641 gastric, 642 small intestinal, 644-645 Chemoreceptor trigger zone, in vomiting, 197, 198f Chemotherapy for carcinoid tumors, 488 for cholangiocarcinoma, 1177 Clostridium difficile infection and, 1895 colonic ischemia with, 2039 for colorectal cancer, 2232-2237 adjuvant, 2232-2234, 2233f-2234f in advanced disease, 2234-2237, 2235f-2237f in rectal cancer, 2233 diarrhea from, 217-218 for enteropathy-associated T cell lymphoma, 458, 2058 for esophageal cancer, 762-764 esophagitis from, 738 for gallbladder carcinoma, 1181 for gastric cancer, 903-904

Chemotherapy (Continued) for gastric diffuse large B cell lymphoma, 452, 452t for gastric MALT lymphoma, 451 for gastrointestinal stromal tumors, 468-469, 473 gastropathy from, 856 hepatitis B flare from, 1297 for hepatocellular carcinoma, 1579 intraperitoneal, for gastric cancer, 904 nausea and vomiting from, 199t, 200 for pancreatic cancer, 1026-1027, 1026f, 1027t for pancreatic endocrine tumors, 520-521 for peritoneal carcinomatosis, 619 sinusoidal obstruction syndrome from, 1376 for small intestinal adenocarcinoma, 2152 for small intestinal immunoproliferative disease, 457 targeted for hepatocellular carcinoma, 1579 for pancreatic cancer, 1027, 1027t Chenodeoxycholic acid, 1093. See also Bile acid(s). colonic motility and, 1671-1672 synthesis and metabolism of, 1076, 1077f, 1077t, 1078 Chest pain in achalasia, 691-692 in distal esophageal spasm, 692 esophageal, 177-178 in gastroesophageal reflux disease, 714, 725 in pill esophagitis, 736 in radiation-induced esophagitis, 640 Chicago classification, of esophageal motility disorders, 698, 699t Chief cells, 778, 817-819 Child classification, in primary sclerosing cholangitis, 1159 Child-Turcotte-Pugh score in alcoholic cirrhosis, 1394, 1394f for liver disease severity assessment, 1236 liver transplantation risk and, 1595 in transjugular intrahepatic portosystemic shunt (TIPS), 1506, 1506t Children. See also Infants; Neonates. acute abdominal pain in, 161 appendicitis in, 2062 bile duct diseases in, 1052-1063 biliary tract disease in, 1050-1052, 1051t-1052t celiac disease in, 1804 cholecystitis in acalculous, 1065 calculous, 1064-1065 choledochal cysts in, 1058-1059, 1058f-1059f cholestasis in, 1050-1052, 1051t-1052t. See also Cholestasis, pediatric. constipation in, 262 Crohn’s disease in, 1971-1972 fecal incontinence in, 257-258 foreign bodies in, 397, 400 functional fecal retention in, 266 gallbladder disease in, 1063-1066 gallstone disease in, 1063-1064, 1132 gastrointestinal stromal tumors in, 473 genetic testing of, 951-952 hepatoblastoma in, 1582 hydrops of gallbladder in, 1065-1066 interlobular bile duct paucity in, 1206, 1206f nonsyndromic, 1060 syndromic, 1060-1062, 1061f-1062f kwashiorkor in, 61-62, 62t liver transplantation in indications for, 1594 lymphoproliferative disorders after, 1610 marasmus in, 61, 62t nutritional dwarfism in, 62-63, 62t pancreatitis in, 937-939, 938t polyethylene glycol in, 278 protein-energy malnutrition in, 61-63, 62t sclerosing cholangitis in, primary, 1056-1058 short bowel syndrome in, 1779, 1780t Yersinia enterocolitica infection in, 1869

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xv

xvi

Index Chinese herbal medications hepatotoxicity of, 1457-1458 for irritable bowel syndrome, 2103, 2291, 2292t for liver disease, 2295t, 2296-2297 Chlamydial infection, in Fitz-Hugh–Curtis syndrome, 1353 Chlamydial peritonitis, 617, 617f ascites in, 1517, 1528, 1536 Chloramphenicol for typhoid fever, 1865t, 1866 for Whipple’s disease, 1841t Chlordecone, hepatotoxicity of, 1452 Chloretone, hepatotoxicity of, 1452 Chloride absorption of, intestinal, 1682-1683 congenital diarrhea and, 1641 secretion of intestinal, 1683, 1683f pancreatic, 922f Chloride-bicarbonate anion exchanger, 16821683, 1685 Chloride channels activator of, for constipation, 280 calcium-activated, 1684 CFTR, 1683 ClC family of, 1683-1684 Chloridorrhea, congenital, diarrhea from, 213 Chloroform, hepatotoxicity of, 1451 Chloroquine for amebiasis, 1910t for porphyria, 1269 Chlorpromazine cholestatic hepatitis from, 1440 for gastrointestinal bleeding, 2286 Cholangiocarcinoma, 1171-1177 classification of, 1171, 1172f clinical features and diagnosis of, 1173-1175, 1173t, 1174f-1176f epidemiology of, 1171-1172 etiology of, 1172, 1176t extrahepatic, 1171, 1172f, 1173, 1175f surgery for, 1176, 1176t gallstone disease and, 1137 hilar (perihalar), 1171, 1172f, 1174f, 11751176, 1190, 1580-1582 intrahepatic, 1171, 1172f, 1173, 1175f, 1176, 1580-1582 clinical features of, 1580-1581 diagnosis of, 1581 epidemiology of, 1580 pathogenesis of, 1580 pathology of, 1581 prognosis in, 1581-1582 treatment of, 1581-1582 liver fluke infection and, 1934, 1934t liver transplantation for, 1176-1177, 1596, 1602-1603 molecular changes in, 1172-1173, 1173t pathogenesis of, 1172-1173, 1173t pathology of, 1172 in primary sclerosing cholangitis, 1154-1155, 1162-1163, 1163f, 1166, 1172, 1176 in recurrent pyogenic cholangitis, 1170 risk factors for, 1172, 1176t staging of, 1175-1176, 1176t treatment of biliary stent placement in, 1177 chemotherapy in, 1177 palliative, 1177 photodynamic therapy in, 1177, 1196 surgical, 1176-1177, 1176t targeted, 1177 Cholangiocytes development of, 1046-1047 injury to, jaundice in, 328 physiology of, 1210 Cholangiography in Caroli’s disease, 1060, 1060f computed tomography in acute mesenteric ischemia, 2031, 2031f in gallstone disease, 1108t, 1111-1112, 1111f multidetector, 1186, 1186f

Cholangiography (Continued) during laparoscopic cholecystectomy, 1126-1128 magnetic resonance, in gallstone disease, 1108t, 1111-1112, 1112f percutaneous transhepatic. See Percutaneous transhepatic cholangiography. Cholangiopancreatography. See Endoscopic retrograde cholangiopancreatography; Magnetic resonance cholangiopancreatography. Cholangiopathy, AIDS, 534, 534f, 1154 from cytomegalovirus, 1347-1348, 1348f Cholangitis ascending, symptoms of, 158 bacterial, 1107t, 1117-1118 clinical features of, 1118 diagnosis of, 1118 etiology of, 1117-1118 treatment of, 1118 after endoscopic retrograde cholangiopancreatography, 661 fibro-obliterative, 1161 immunoglobulin G4–associated, 1155 after Kasai hepatoportoenterostomy, 1056 recurrent pyogenic, 1167-1170, 1168f sclerosing. See Sclerosing cholangitis. Cholangitis lenta, in bacteremia, 582, 582f Cholate, 1093 Cholecystectomy abdominal symptoms after, 1135-1137, 1137t for acalculous cholecystitis, 1131, 1143 in children, 1065 for asymptomatic gallstone disease, 1129 bile acid circulation disturbances after, 1087 for biliary pain, 1130 for calculous cholecystitis, 1130-1132, 1130t, 1131f in children, 1065 cancer and, 1137 choledocholithiasis after, 1134-1137 choledocholithiasis identified during, 1134, 1134f colorectal adenoma and, 2165 colorectal cancer and, 2210-2211 cystic duct remnant after, 1137 diarrhea after, 229 esophageal adenocarcinoma and, 749 for gallbladder carcinoma, 1180-1181 for gallbladder dyskinesia, 1067 in children, 1066 for gallbladder polyp, 1151 for gallstone disease, 1124-1129 in children, 1064 for hydrops of gallbladder, 1066 incidental, 1133 laparoscopic, 1126-1129 bile duct injury during, 1134-1135, 1136f cholangiography during, 1126-1128 results of, 1128-1129, 1128t, 1129f technique for, 1126, 1127f open, 1125-1126 results of, 1125-1126, 1125f, 1125t technique for, 1125 in pregnancy, 630 in sickle cell anemia, 572 sphincter of Oddi dysfunction after, 1072, 1137 Cholecystitis abdominal pain in, acute, 154t, 157-158 acute (acalculous), 1141-1144 versus acalculous biliary pain, 1150t in children, 1065 cholecystostomy for, 1191 cholescintigraphy in, 1143 clinical features of, 1142 computed tomography in, 1143 definition of, 1141 diagnosis of, 1142-1143, 1142t epidemiology of, 1141 in HIV/AIDS, 534 pathogenesis of, 1141-1142 prevention of, 1144

Cholecystitis (Continued) treatment of, 1131, 1143-1144 ultrasonography in, 1142-1143, 1143f acute (calculous) in children, 1064-1065 cholecystectomy for, 1130-1132, 1130t, 1131f cholecystostomy for, 1131, 1131f, 1190f, 1191 cholescintigraphy in, 1108t, 1110-1111, 1110f clinical features of, 1114, 1114f in diabetes mellitus, 1131 diagnosis of, 1114-1115 differential diagnosis of, 1115 in gallstone disease, 1107t, 1114f, 1130-1132, 1130t, 1131f natural history of, 1114 pathogenesis of, 1113 pathology of, 1113-1114 treatment of, 1116, 1130-1132 ultrasonography in, 1108t, 1109, 1109f versus appendicitis, 2063t chronic, 1112-1113 emphysematous, 1111-1112, 1111f, 1118, 1119t, 1131-1132 after endoscopic retrograde cholangiopancreatography, 661 after hematopoietic stem cell transplantation, 549 necrotizing, 1141 in sickle cell anemia, 572 suppurative, 1114-1115 Cholecystoenteric fistula, in gallstone disease, 1118-1119, 1119t Cholecystokinin (CCK), 7 abdominal pain and, 996 actions of, 7 location and, 4 clinical application of, 7 in eating disorders, 122-123 forms of, 7 in gallstone disease, 1100 in gastric acid secretion, 823 in gastric emptying, 799-800 growth-promoting effects of, 17 in intestinal phase of pancreatic enzyme secretion, 927, 927f in lower esophageal sphincter relaxation, 685 in pancreatic enzyme secretion, 1697 receptors for, 7 regulation of, by intraluminal releasing factors, 17, 17f in satiety, 5, 18, 802, 1696 Cholecystokinin-1 receptor, 7 Cholecystokinin-1R, in gallstone disease, 1102, 1103t Cholecystokinin-2 receptor, 7 Cholecystokinin-releasing peptides, 1697 Cholecystokinin test, 928-929, 928t in chronic pancreatitis, 999-1000 in sphincter of Oddi dysfunction, 1069 Cholecystoses, hyperplastic, 1144 Cholecystostomy for acute acalculous cholecystitis in children, 1065 endoscopic, 1144 percutaneous, 1143 surgical, 1143 for acute calculous cholecystitis, 1131, 1131f, 1190f, 1191 in children, 1065 for hydrops of gallbladder, 1066 Choledochal cysts, 1058-1059, 1058f-1059f, 1205 in Caroli’s disease, 1590 clinical features of, 1059 diagnosis of, 1059, 1059f etiology of, 1058 incidence and classification of, 1058, 1058f pathology of, 1058-1059 treatment of, 1059 Choledochoduodenal junction, 1048f Choledocholithiasis, 1089, 1107t, 1116-1117 clinical features of, 1116 diagnosis of, 1117

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Choledocholithiasis (Continued) differential diagnosis of, 1117 endoscopic retrograde cholangiopancreatography in, 1108t, 1111, 1111f, 1191, 1192f endoscopic ultrasonography in, 1108t, 1109-1110, 1110f etiology of, 1116 extracorporeal shock wave lithotripsy for, 1124 hematopoietic stem cell transplantation and, 545 identified after cholecystectomy, 1134-1137 identified during cholecystectomy, 1134, 1134f jaundice in, 329 known preoperatively, 1134 magnetic resonance cholangiography in, 1108t, 1112, 1112f natural history of, 1116 percutaneous transhepatic cholangiography for, 1189 in primary sclerosing cholangitis, 1154-1155, 1162 in recurrent pyogenic cholangitis, 1169 treatment of, 1117, 1133-1134, 1134f ultrasonography in, 1108t, 1109 Cholehepatic shunting of bile acids, 1080-1081 Cholelithiasis. See Gallstone disease. Cholera, 1821, 1850-1852 O1 (classic and El Tor) vibrios in, 1850 O139 Bengal vibrio in, 1850 clinical features of, 1851 diarrhea in, 214, 215f, 1851, 1851t electrolyte composition of, 1851, 1851t epidemiology of, 1850 immunologic responses in, 1851 intestinal ion transport in, 1691 microbiology of, 1850 non-O1 vibrios in, 1852-1853 pathogenesis of, 1851, 1851t toxin in, 214, 1845, 1845f, 1850 treatment of, 1851-1852, 1851t-1852t, 1865t vaccines for, 1852 Cholescintigraphy in acalculous biliary pain, 1141 in acute acalculous cholecystitis, 1143 in acute calculous cholecystitis, 1115 in cystic fibrosis, 946 in gallstone disease, 1108t, 1110-1111, 1110f morphine-augmented, in acute acalculous cholecystitis, 1143 in neonatal cholestasis, 1052 in sphincter of Oddi dysfunction, 1069 Cholestasis. See also Jaundice. in Alagille’s syndrome, 1061 alkaline phosphatase in, 1232, 1232f, 1233t atypical presentations of, 329 autoimmune hepatitis with, 1467, 1467t bacterial infection and, 1104 chronic, medical management of, 1062-1063 drug-induced acute, 1439-1441 with bile duct injury, 1441 chronic, 1441 diagnosis of, 1439 with hepatitis, 1440 jaundice in, 328-329 without hepatitis, 1439-1440 dyslipidemia in, 1225 extrahepatic, 1084, 1232-1233, 1233t after hematopoietic stem cell transplantation, 547t, 548, 548f in hepatitis A, 1282 in hepatitis E, 1341 intrahepatic, 1084, 1232-1233, 1233t benign recurrent, 328, 1085-1086, 1085t, 1276 familial, 1085-1086, 1085t, 1275t, 1276-1277 idiopathic, in Hodgkin’s disease, 565 of pregnancy, 1085t jaundice in, 329 jaundice in, 327-329 liver transplantation for, 1601-1602 low phospholipid–associated, 1085t

Cholestasis (Continued) pediatric, 1050-1052, 1051t-1052t in biliary atresia, 1054. See also Biliary atresia. in cystic fibrosis, 947 diagnosis of, 1051-1052, 1052t medical management of, 1062-1063 pruritus in, 1063 postoperative, 583-584, 583t, 1450t in pregnancy, 631-632 in primary biliary cirrhosis, 1485-1487, 1486t in primary sclerosing cholangitis, 1161-1162 prominent, jaundice in, 327-329 in sarcoidosis, 589-590 in sepsis, 582-583, 582f after hematopoietic stem cell transplantation, 547t, 548, 548f jaundice in, 329 ursodeoxycholic acid for, 335 Cholesterol in ascites, 1526 in bile, 1093, 1141 dietary, intake of, 1698-1699 digestion and absorption of, 1702, 1705-1706 brush border membrane transfer in, 1702-1703, 1703f esterification in, 1704 liver X receptors in, 1705-1706, 1705t, 1706f estrogen effects on, 1092 in gallstone disease, 1097-1101, 1097f hepatic regulation of, 1095-1097, 1096f nucleation and crystallization of, 1094, 1098-1099 solubility of, 1093-1095, 1095f supersaturation of, 1097 synthesis of, 1095-1096 hepatic, 1222 trafficking of, 1221 Cholesterol 7α-hydroxylase (Cyp7A1) in bile acid synthesis, 1076, 1078, 1705-1706 in gallstone disease, 1102, 1103t Cholesterol ester transfer protein, 1222 Cholesterol polyps, 1145-1146, 1149, 1150t Cholesterol stones, in children, 1064 Cholesterolosis, 1144-1146, 1144f clinical features of, 1145-1146 definition of, 1144 diagnosis of, 1146 epidemiology of, 1144-1145 pathogenesis of, 1145 pathology of, 1145, 1145f treatment of, 1146 Cholestyramine, 1088 adjunctive, in Crohn’s disease, 1970 for bile acid malabsorption, 1753 for Clostridium difficile-associated diarrhea and colitis, 1901 colorectal cancer and, 2210 malabsorption with, 1757t in primary biliary cirrhosis, 1486, 1486t for pruritus, 1063 for short bowel syndrome, 1783, 1785 Cholic acid, 1076, 1077f, 1077t. See also Bile acid(s). Choline deficiency, in parenteral nutrition– induced liver disease, 86 Cholinergic agents, for constipation, 281 Cholinergic neurons, activation of, mealstimulated, 822f, 825 Cholyl-PABA, urinary, in small intestinal bacterial overgrowth, 1777 Cholylsarcosine for bile acid malabsorption, 1756 for short bowel syndrome, 1785 CHOP regimen, for gastric diffuse large B cell lymphoma, 452, 452t Chromium, dietary, 56t-57t, 80, 80t Chromoendoscopy in colorectal cancer, 2196-2197, 2197f, 2224 in esophageal cancer conventional, 754 electronic, 754-755, 754f Chromogranins in carcinoid tumors, 484-485 in pancreatic endocrine tumors, 493

Chromosomal abnormalities in colorectal cancer, 34-35, 35f, 2202-2203, 2203f in esophageal cancer, 751 in oncogenesis, 34-35, 35f Chromosomes, 1208 Chronic obstructive pulmonary disease, peptic ulcers in, 581 Churg-Strauss syndrome gastrointestinal manifestations of, 562 vasculitis in, 2046 Chylomicron(s), 1221 formation and transport of, 1222-1224, 1223f, 1704-1705 defects in, 1704-1705, 1738 lymphatic transport of, defective, 1738 Chylomicron remnant receptor, 1224 Chylomicron retention disease. See Anderson’s disease. Chylous ascites, 1526, 1528 Chymotrypsin, 1712-1713, 1713f, 1713t fecal, 928t, 929 in chronic pancreatitis, 1000 in diarrhea, 226 functions of, 924 Cidofovir, for cytomegalovirus colitis, in HIV/ AIDS, 528 Cigarette smoking. See Smoking. Ciguatera poisoning, 1884 Cimetidine. See also Histamine H2 receptor antagonists. adverse effects of, 870 for aphthous ulcers, 356 drug interactions with, 870 hepatotoxicity of, 1433 mechanisms of action of, 870 for peptic ulcer disease, 870 pharmacokinetics of, 870 Cinitapride, for vomiting, 208 Ciprofloxacin for cholera, 1852 for Crohn’s disease, 1961, 1962t empiric, for infectious diarrhea, 1885 enteral nutrition and, 95 for pneumatosis cystoides intestinalis, 240 for pouchitis, 2009 for salmonellosis, 1864, 1865t for shigellosis, 1860, 1865t for spontaneous bacterial peritonitis, 1533 for traveler’s diarrhea, 1874 Circulation enterohepatic. See Enterohepatic circulation. pancreatic, 911 portal, 1489-1491, 1490f splanchnic. See Splanchnic circulation. Cirrhosis adrenal insufficiency in, 1554 alcoholic. See also Alcoholic liver disease. hepatitis C and, 1392, 1392f histology of, 1383-1384, 1384f, 1390-1391 optimal management of, 1399-1400 prognosis in, 1394, 1394f protein-energy malnutrition in, 64 in α1-antitrypsin deficiency, 1261 ascites in, 1517-1518, 1518f, 1524-1525. See also Ascites. diagnosis of, 1527 hepatic hydrothorax in, 1535 in hepatorenal syndrome, 1547 pleural effusions in, 1535 refractory, 1538-1540, 1539f treatment of, 1536-1540, 1539f AST/ALT ratio in, 1230 in autoimmune hepatitis, 1469 bacterial infection in, 1530 biliary. See Biliary cirrhosis, primary. carbohydrate metabolism in, 1219-1220 cardiac, in congestive hepatopathy, 1380, 1381f cardiomyopathy in, 1553-1554, 1553f coagulation disorders in, 1554-1555 in cystic fibrosis, 945-947, 946f cytochrome P450 expression in, 1418 drug-induced, 1441-1444 dyslipidemia in, 1225

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xvii

xviii

Index Cirrhosis (Continued) endocrine disorders in, 1554 in glycogen storage disease type IV, 1264, 1265f after hematopoietic stem cell transplantation, 554 hepatic mass lesions in, 1591 hepatic microcirculation in, 1490f in hepatitis B, 1288, 1293-1294, 1306 hepatocellular carcinoma in, 1577 in hereditary hemochromatosis, 1242 hypogonadism in, 1554 jaundice in, 327 in nonalcoholic fatty liver disease, 1401, 1406-1407. See also Fatty liver disease, nonalcoholic. osteoporosis in, 1554 portal hypertension in, 1498-1499 portal vein thrombosis in, 1377 in pregnancy, 637 in primary biliary cirrhosis, 1480-1481, 1481f in primary sclerosing cholangitis, 1157, 1161 renal disease in. See Hepatorenal syndrome. small intestinal bacterial overgrowth in, 1773 thyroid dysfunction in, 1554 tuberculous peritonitis in, 617 in tyrosinemia, 1270 well-compensated, risk of complications for, 1594-1595, 1595f Cisapride for abdominal distention and bloating, 239 for chronic intestinal pseudo-obstruction, 2143 for constipation, 280 for functional dyspepsia, 191-193 for gastric motility disorders, 812, 813t for gastroesophageal reflux disease, 721-722 for gastroparesis, 574 for small intestinal bacterial overgrowth, 1777 for vomiting, 208 Cisplatin, for metastatic pancreatic endocrine tumors, 521 Cisplatinum, for esophageal cancer, 764 Citric acid cycle, 1209 Citrin, in urea cycle defects, 1273 Citrulline, in urea cycle defects, 1273 Cladribine, for celiac disease, 2057 Clarithromycin for Crohn’s disease, 1961-1963 for Helicobacter pylori infection, 841-842, 842t Clathrin, 1210 Clear cells, in hepatocellular carcinoma, 1574-1575 Clebopride, for vomiting, 207 Clevudine, for hepatitis B, 1304 Clindamycin, for babesiosis, 1358 Cloaca, 1625 persistent, 1635 Clobetasol, for aphthous ulcers, 356 Clofibrate, gallstone disease and, 1092 Clonal expansion, in oncogenesis, 34 Clonidine for diabetic diarrhea, 575 for diarrhea, 227, 227t in intestinal ion transport, 1690 for irritable bowel syndrome, 2103 for levator ani syndrome, 2273 for short bowel syndrome, 1784, 1784t Clonorchiasis, hepatic, 1356t-1357t, 1363, 1363f Clonorchis sinensis infection, 1934-1935, 1934t Clopidogrel drug interactions of, with proton pump inhibitors, 871 after endoscopic hemostasis of peptic ulcer bleeding, 302 hepatotoxicity of, 1433-1434 Closed eyes sign, in functional abdominal pain syndrome, 168-169 Clostridium botulinum infection, 1879t-1880t, 1882-1883. See also Botulinum toxin. Clostridium difficile biotherapy, with nontoxigenic strains, 1901-1902

Clostridium difficile infection. See also Clostridium difficile-associated diarrhea and colitis. acid suppression and, 1895 in antibiotic-associated diarrhea, 1890, 1890t chemotherapy and, 1895 in elderly, 1876, 1894-1895 in HIV/AIDS, 527t, 529, 1895 hospital-acquired, 1872 hospital epidemiology of, 1892-1893 in inflammatory bowel disease, 1895 intestinal ion transport in, 1692 in pseudomembranous enterocolitis, 18901891, 1897, 1897f, 1899-1900 toxins in, 1893-1894, 1893f vaccine against, 1902 Clostridium difficile-associated diarrhea and colitis, 230, 1891-1902 antibiotics associated with, 1892, 1892t asymptomatic carriage of, 1895 clinical features of, 1895-1896, 1895f colonic microflora in, 1891-1892 colonoscopy in, 1897, 1897f-1898f complications of, 1896 computed tomography in, 1895-1896, 1895f culture in, 1896t, 1897 diagnosis of, 1896-1897, 1896t, 1897f-1898f enzyme immunoassay in, 1896-1897, 1896t epidemiology of, 1892-1893 after hematopoietic stem cell transplantation, 552 immune response in, 1894, 1894f pathogenesis of, 1891-1895, 1891f, 1892t, 1893f-1894f polymerase chain reaction assay in, 1896t, 1897 prevention of, 1892, 1892t risk factors for, 1894-1895 sigmoidoscopy in, 1897 after solid organ transplantation, 541-542 stool tests for, 1896-1897, 1896t tissue culture cytotoxicity assay in, 1896, 1896t toxins in, 1893-1894, 1893f treatment of, 1898-1902, 1898t bacteriotherapy for, 1902 binding resins for, 1901 immunization, 1902 metronidazole for, 1898-1901, 1898t in mild to moderately severe disease, 1898-1899 probiotics for, 1901-1902 in recurrent disease, 1900-1902, 1900t in severe disease, 1899-1900 vancomycin for, 1898-1901, 1898t Clostridium perfringens infection antibiotic-associated diarrhea in, 1889 in food poisoning, 1879t-1880t, 1881-1882 hepatic manifestations of, 1351-1352 intestinal ion transport in, 1692 Clotrimazole for candidal esophagitis, 742 for oral candidiasis, 354 Clove, for colonic health, 2294t Clubbing, digital in Crohn’s disease, 1953-1954 in hepatopulmonary syndrome, 1550 in immunoproliferative small intestinal disease, 1825f Coagulation disorders in acute liver failure, 1561t, 1562, 1566 in acute pancreatitis, 982 ascites and, 1519 in cirrhosis, 1554-1555 in Crohn’s disease, 1955 gastrointestinal manifestations of, 569-570, 569f, 569t after polypectomy, 660 Cobalamin. See Vitamin B12 (cobalamin). Cocaine colonic ischemia from, 2040 gastropathy from, 857 hepatotoxicity of, 1454 peptic ulcer disease and, 862

Coccidioidomycosis ascites in, 1517 in HIV/AIDS, 527t, 530 Coccygodynia, 2259, 2272 Codeine, for short bowel syndrome, 1784, 1784t Cogan’s syndrome gastrointestinal manifestations of, 562 vasculitis in, 2046 Cognitive-behavioral therapy for eating disorders, 132-133 for functional abdominal pain syndrome, 170 for gastrointestinal symptoms, 349 Coin, in gastrointestinal tract, endoscopic management of, 402-403 Colchicine for alcoholic liver disease, 1398 for aphthous ulcers, 356 colonic motility and, 1672 for constipation, 281-282 for irritable bowel syndrome, 2103 malabsorption with, 1757t for primary biliary cirrhosis, 1484 for primary sclerosing cholangitis, 1165 Colectomy. See also Proctocolectomy. for colonic angioectasia, 599 for colonic obstruction, 2118-2120 for constipation, 283 for diverticular disease, 2078, 2080-2082, 2088 for familial adenomatous polyposis, 2183 with ileorectal anastomosis. See Ileorectostomy. partial. See also Colostomy. for ulcerative colitis, 2024 short bowel syndrome after, 1784-1786 surgical options after, 2015-2026, 2025t. See also Ileorectostomy; Ileostomy. for ulcerative colitis, 1984-1985, 1984f, 2003-2004 Colesevelam, 1088 adjunctive, in Crohn’s disease, 1970 for pruritus, 1063 Colestipol, 1088 for Clostridium difficile-associated diarrhea and colitis, 1899, 1901 for short bowel syndrome, 1783 Colic infantile, IgE-mediated hypersensitivity in, 143 renal, 200, 2063t Coliforms, toxin-producing, tropical sprue and, 1823 Colipase, 1700-1701, 1700f congenital absence of, 955-956 Colitis. See also Enterocolitis. in allergic eosinophilic proctocolitis, 144-145, 146t amebic. See Amebiasis, intestinal. bleeding in, 312-313 in Campylobacter spp. infection, 1867, 1868f chemical, 2247-2248, 2247t, 2248f Clostridium difficile-associated. See Clostridium difficile-associated diarrhea and colitis. in Crohn’s disease, 313, 1950, 1990t cytomegalovirus, 1990 differential diagnosis of, 1956t, 1990t diversion, 2242-2243 in diverticulitis, 1991 drug-induced, 1991 in enterohemorrhagic Escherichia coli infection, 1856, 1856f gold, 559 gonococcal, 1990t hemorrhagic, antibiotic-associated, 2039 indeterminate, 1958-1959, 1988-1989 ileal pouch–anal anastomosis in, 2023 infectious, 313, 1989-1990, 1990t versus ulcerative colitis, 1989-1990, 1990t with inflammatory pseudopolyps, in schistosomiasis, 1936, 1936f ischemic, 312-313, 1990-1991, 1990t universal fulminant, 2043

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Colitis (Continued) microscopic (collagenous and lymphocytic), 228, 1990t, 1991, 2239-2242 background on, 2239 in celiac disease, 1813 clinical and laboratory features of, 2241 differential diagnosis of, 2241-2242 epidemiology of, 2240 etiology and pathogenesis of, 2240-2241 pathology of, 2240, 2240f treatment of, 2242, 2242f necrotizing, with toxic megacolon, in amebic dysentery, 1909 noninfectious causes of, 1990-1991 NSAID use and, 2040 postdysenteric, 1875 pseudomembranous, 1989-1990, 1990t radiation-induced, 647-650, 648f, 1991. See also Radiation enteritis. in salmonellosis, 1863 segmental in diverticulosis, 2084 ulcerating, in colonic ischemia, 2043 ulcerative. See Ulcerative colitis. Colitis cystica profunda, 2176, 2251-2252, 2251f-2252f Colitis cystica superficialis, 2251 Collagen, in extracellular matrix, 1212 Collagenous colitis, 228, 2239-2242 clinical and laboratory features of, 2241 differential diagnosis of, 2241-2242 epidemiology of, 2240 etiology and pathogenesis of, 2240-2241 pathology of, 2240, 2240f treatment of, 2242, 2242f Collagenous gastritis, 853, 853f Collagenous sprue, 1811, 1817 Collateral circulation in portal hypertension, 1494-1495 splanchnic, 2028, 2029f Colloid replacement, for ascites, 1539-1540 Colocutaneous fistula, after percutaneous endoscopic gastrostomy, 90-91 Colon. See also Cecum; Rectum. aberrant crypt foci of, colorectal cancer and, 2196-2197, 2197f adenoma of. See Colorectal adenoma. aganglionosis of. See Hirschsprung’s disease. anatomy of, 262-264, 265f, 1615-1616, 1616f, 1622, 1660-1662, 1661f angioectasia of. See Angioectasia, colonic. ascending, 1616 atresia of, 1632 biopsy of. See Colonic biopsy. calcium salvage in, 1741 in carbohydrate malabsorption, 1740-1741, 1740f-1741f carbohydrate salvage by, 1780-1781, 1781f carcinoid tumors of, 481 carcinoma of. See Colorectal cancer. cathartic, 2245-2246 cells of, 1622 complementary and alternative medicine for, 2293-2295, 2294t decompression of, for acute colonic pseudoobstruction, 2130, 2131f descending, 1616 diameter of, 263 diaphragm disease of, 2244-2245, 2245f in digestion and absorption, 1698 diverticular disease of, 2073-2090. See also Diverticula; Diverticulitis; Diverticulosis. duplications of, 1632 dysfunction of, in cystic fibrosis, 943 dysplasia of after ileal pouch–anal anastomosis, 2022-2023 in ulcerative colitis, 2005-2008, 2006f-2007f enlarged. See Megacolon. in fat malabsorption, 1741 fluid and electrolyte transport in, 1675-1694. See also Fluid and electrolyte transport. foreign bodies in, 399, 404, 404f

Colon (Continued) ganglia of, congenital absence or reduction in. See also Hirschsprung’s disease. constipation in, 269-270 gas in, 234-236, 234f-235f hydrogen absorption and, 1741, 1741f haustra of, 1615-1616, 1660 hemangioma of, 604 hypersensitivity of, in irritable bowel syndrome, 2095 inflammation of. See Colitis. innervation of, 263-264, 1623, 1623f, 1662-1664, 1662f-1663f intrinsic, 1662-1664, 1663f parasympathetic, 1662f, 1664 spinal afferents in, 1662f, 1664 sympathetic, 1662f, 1664 interstitial cells of Cajal of, 264 ischemia of. See Colon ischemia. length of, 263 luminal contents of, 262-263 lymphatic drainage of, 1617, 1623 lymphocytosis of, in celiac disease, 1813 lymphoma of, 459 in malabsorption, 1740-1741, 1740f-1741f malakoplakia of, 2250-2251, 2251f microflora of, in Clostridium difficile-associated diarrhea and colitis, 1891-1892 motility of. See Colonic motility. motor function in, 263 mucosa of, 1617-1619, 1618f myopathy of, constipation in, 269 nutrient absorption in, 1740-1741 nutrient presence in, gastrointestinal transit and, 1697 obstruction of. See Colonic obstruction. pain transmission in, 343-345, 343f perforation of in colonoscopy and sigmoidoscopy, 659-660, 659f in ulcerative colitis, 2005 polyps of. See Colonic polyps. pseudo-obstruction of. See Pseudo-obstruction, colonic. pseudomelanosis of, 2246-2247, 2246f-2247f radiation effects on, 647-650, 648f. See also Radiation enteritis. resection of. See also Short bowel syndrome. diarrhea after, 228-229 short-chain fatty acid transport in, 1685-1686 sigmoid, 1616 smooth muscle of, 1660-1661, 1661f ion channels in, 1661 nocturnal suppression of, 1671, 1671f sodium absorption in, 263 as storage organ, 1666-1667 strictures of. See Colonic strictures. stromal tumors of, 465-466 taeniae of, 1615-1616, 1660 transit time of. See Colonic transit time. transverse, 1616 tuberculosis of, 1877, 1877f-1878f ulcers of. See Colonic ulcers. varices of, 1513 vascular supply to, 1617, 1623 wall structure of, in diverticulosis, 2075 water absorption in, 263 Colon ischemia, 2038-2044 aortic surgery and, 2044 causes of, 2038-2040, 2039t clinical features of, 2041-2042 colonoscopy in, 2041, 2041f colorectal cancer and, 2044, 2044f diagnosis of, 2041-2042, 2041f-2042f drug-induced, 2039, 2039t gangrene in, 2041, 2041f, 2043 incidence of, 2038 in irritable bowel syndrome, 2044 pathology of, 2040, 2041f pathophysiology of, 2038-2040 right-sided, 2043-2044 segmental ulcerating colitis in, 2043

Colon ischemia (Continued) single-stripe sign in, 2041 spectrum of, 2038, 2038t strictures in, 2043 thumbprinting in, 2041, 2041f-2042f treatment of, 2042-2043, 2043f universal fulminant colitis in, 2043 Colonic biopsy in Clostridium difficile-associated diarrhea and colitis, 1897, 1898f in colitis cystica profunda, 2252, 2252f in diarrhea, 223-224 in intestinal endometriosis, 2255, 2255f in malakoplakia, 2251, 2251f in pneumatosis coli, 2249-2250, 2250f in schistosomiasis, 1937, 1937f Colonic inertia, 265. See also Constipation. Colonic irrigation therapy, 2288t, 2295 Colonic motility, 1659-1674, 2122 anatomy and, 1660-1662, 1661f calcium channels in, 1661 disorders of, 1672-1674 constipation and, 1672-1673, 1673f diarrhea and, 1673 irritable bowel syndrome and, 1673 secondary to nonmotor disorders, 1673-1674 in diverticulosis, 2075 enteric nervous system in, 263-264, 1662-1664, 1663f afferent neurons of, 1662-1663 efferent neurons of, 1663 interneurons of, 1663-1664 extrinsic afferent pathways in, 1664 interstitial cells of Cajal in, 1661-1662, 1661f in irritable bowel syndrome, 2095, 2096f modulators of, 1670-1672 nonpharmacologic, 1672 pharmacologic, 1671-1672 physiologic, 1670-1671, 1671f myenteric potential oscillations in, 1660 neural control system for, 2121, 2122f-2123f, 2123-2124 nonpropagating motor patterns in, 1665 parasympathetic innervation in, 1662f, 1664 propagating sequences in, 1660-1661, 1665 regional linkage among, 1665-1666, 1666f regional variation of, 1665, 1666f rectal motor complexes in, 1665 regulation of, 1666-1668 defecation and, 1668-1670, 1669f ileocecal junction in, 1666, 1667f pressure patterns and flow in, 1665, 1667-1668, 1668f storage and, 1666-1667 slow waves in, 1660 smooth muscle in, 1660-1661, 1661f in spinal cord injury, 2137 sympathetic innervation in, 1662f, 1664 types of propulsions in, 263 Colonic obstruction, 2116-2120 abdominal radiography in, 2117, 2118f-2119f clinical presentation in, 2117 in colonic strictures, 2116-2120 in colorectal cancer, 2116-2120, 2119f constipation in, 269 contrast enema in, 2117, 2119f diagnostic evaluation of, 2117, 2118f-2119f in diverticulitis, 2083, 2116-2117 etiology of, 2116-2117 pathophysiology of, 2117 treatment of, 2117-2120 in volvulus, 2116-2117, 2118f-2119f, 2120 Colonic polyps, 2155-2176. See also Polyposis. adenomatous. See Colorectal adenoma. bleeding in, 313 classification of, 2155, 2156t in Cronkhite-Canada syndrome, 2188 diminutive, 2158 in familial adenomatous polyposis, 2178-2179, 2179f. See also Familial adenomatous polyposis.

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xix

xx

Index Colonic polyps (Continued) hyperplastic, 2172-2174, 2188 histology of, 2172-2173, 2173f prevalence of, 2173, 2173t treatment of, 2173-2174 inflammatory, 2175-2176 juvenile, 2174-2175, 2175f, 2186-2187 lymphoid, benign, 2176 malignant, 2156-2157, 2169-2171, 2170f, 2170t. See also Colorectal cancer. mucosal, 2174 neoplastic. See Colorectal adenoma; Colorectal cancer. non-neoplastic, 2172-2176 pedunculated, 2170-2171 Peutz-Jeghers, 2175, 2175f, 2185t, 2186 recurrent, 2171, 2171f retention, 2174 serrated, 2174, 2174f sessile, 2170-2171, 2170f submucosal, 2176 in ulcerative colitis, 2007, 2007f-2008f Colonic propulsions, types of, 263 Colonic strictures colonic obstruction in, 2116-2120 in diverticulitis, 2083 ischemic, 2043 NSAID-induced, 2244-2245, 2245f in ulcerative colitis, 2005 Colonic transit time in constipated patient, 264, 265t diet and, 262-263 measurement of in constipation, 272-273, 273f with radiopaque markers, 272-273, 273f with wireless capsule technique, 273 normal, 263 in scleroderma, 560 Colonic ulcers Dieulafoy-type, 2245 nonspecific, 2243-2244, 2243f-2244f NSAID-induced, 2052-2054 stercoral, 2052 Colonization factors, in Helicobacter pylori infection, 834-835 Colonization resistance, impairment of, in Clostridium difficile-associated diarrhea and colitis, 1891 Colonography, computed tomography in colorectal adenoma, 2166-2167, 2167f in colorectal cancer, 2224-2225, 2225f in gastrointestinal bleeding, 310 Colonopathy, fibrosing, from pancreatic enzyme replacement, 942 Colonoscopy in amebic colitis, 1909 in angioectasia, 597, 598f, 598t bleeding in, 658-659, 659f in chemical colitis, 2248f in Clostridium difficile-associated diarrhea and colitis, 1897, 1897f-1898f in colonic ischemia, 2041, 2041f in colonic ulcer, 2243f, 2244 in colorectal adenoma screening, 2166 after sigmoidoscopy, 2169, 2169t surveillance, 2172, 2172t complications of, 658-660 in constipation, 272 decompression with, for acute colonic pseudo-obstruction, 2130 in diaphragmatic ulcer, 2245, 2245f in diarrhea, 220, 223-225 in diverticular bleeding, 2085-2087, 2086f in diverticulosis, 2076-2077, 2076f hernia after, 394 in lower gastrointestinal bleeding, 290-291, 310 patient preparation for, complications of, 660 perforation in, 659-660, 659f in pseudomelanosis coli, 2246f-2247f screening, 2219, 2219f, 2221t, 2224-2225 sigmoid flotation maneuver with, in diverticulosis, 2077

Colonoscopy (Continued) surveillance, in ulcerative colitis, 2007-2008 in ulcerative colitis, 1985-1987, 1986f Colony-stimulating factors, in gastrointestinal tract, 13 Colopathy portal, 602, 603f portal hypertensive, 1514 Colorectal adenoma, 2155-2172 aberrant crypts and, 2159 acromegaly and, 2164-2165 adenoma-carcinoma hypothesis of, 2159-2160 with advanced pathology, 2158-2159 anatomic distribution of, 2162-2163, 2162t in attenuated familial adenomatous polyposis, 2184 barium enema in, 2166 carcinogenesis pathways in, 2160-2161, 2161f carcinoma in situ in, 2156-2157, 2170-2171, 2170f cholecystectomy and, 2165 colonoscopy in screening, 2166 after sigmoidoscopy, 2169, 2169t surveillance, 2172, 2172t computed tomography colonography in, 2166-2167, 2167f conditions associated with, 2164-2165 diagnosis of, 2165, 2167f diminutive, 2158 dysplasia grading of, 2155-2156, 2156t, 2157f epidemiology of, 2161-2163 in familial adenomatous polyposis, 2178-2179, 2179f fecal immunochemical testing in, 2165-2166 fecal occult blood testing in, 2165 flat, 2159 histogenesis of, 2159 histologic characteristics of, 2155-2159, 2156f-2157f, 2156t incidence of, 2162 intramucosal carcinoma in, 2156-2157, 2157f invasive carcinoma focus in, 2157, 2157f. See also Colorectal cancer. in JC virus, 2165 malignant potential of, 2158-2159, 2158t metachronous, 2168 multiple, 2168, 2168t natural history of, 2167-2168 with noninvasive carcinoma, 2156-2157, 2169-2171, 2170f, 2170t number of, 2168, 2168t pathogenesis of, 2159-2161 pathology of, 2155-2159, 2156f-2157f, 2156t prevalence of, 2161, 2161t-2162t progression of, 2167-2168 risk factor(s) for, 2163-2165 diet and lifestyle as, 2163-2164 inherited susceptibility as, 2163 serrated, 2159, 2174, 2174f sigmoidoscopy in, 2166 size of, 2156t, 2157-2158, 2158t stool DNA analysis in, 2167 in Streptococcus bovis bacteremia, 2165 synchronous, 2168 treatment of, 2167-2172 initial, 2168-2169, 2169t for malignant polyp, 2169-2171, 2170f, 2170t postpolypectomy, 2171-2172, 2171f tubular, 2155, 2156f, 2156t, 2158t tubulovillous, 2155, 2156t, 2158t unicryptal, 2159 untreated, 2167-2168 ureterosigmoidostomy and, 2164 villous, 2155, 2156f, 2156t, 2158t Colorectal cancer, 2191-2238 aberrant crypt foci markers in, 2196-2197, 2197f adenoma-carcinoma hypothesis of, 2159-2160 age and, 2208, 2208f, 2216 bile acids and, 2194 biology of, 2199-2204 biochemical changes in, 2204, 2205f cellular proliferation in, 2199-2200

Colorectal cancer (Continued) genes altered in, 2200t molecular genetics in, 2200-2204, 2200t, 2201f-2203f, 2204t bleeding in, 313, 2218, 2218f bowel obstruction or perforation in, 2216 calcium intake and, 2195-2197, 2199f, 2199t carcinoembryonic antigen in, 2193-2194, 2216, 2230-2231 carcinogenesis pathways in, 2160-2161, 2161f carcinogens and fecal mutagens in, 2195 chemoprevention of, 2196-2199, 2197f, 2199f, 2199t cholecystectomy and, 2210-2211 cholestyramine and, 2210 chromosomal instability in, 34-35, 35f, 2202-2203, 2203f clinical features of, 2216-2219, 2218f colonic ischemia and, 2044, 2044f colonic obstruction in, 2116-2120 COX-2 inhibitors and, 2196, 2198 in Crohn’s disease, 1972, 2209-2210, 2227 diagnosis of, 2219-2230. See also Colorectal cancer, screening for. differential diagnosis of, 2218-2219, 2218t difluoromethylornithine (DFMO) and, 2198-2199, 2199f, 2199t distribution of intracolonic, 2193, 2193f prognosis and, 2216 DNA tumor content in, 2215-2216 Dukes classification of, 2213-2215, 2214t epidemiology of, 2191-2193, 2192f-2193f familial, 2204-2207, 2206t-2207t, 2207f in familial adenomatous polyposis, 2178-2179, 2208-2209, 2226-2227 family history in, 2208-2209 fat intake and, 2193-2194 fiber intake and, 2194-2195, 2197, 2199f, 2199t folate intake and, 2197, 2199f, 2199t gallstone disease and, 1137 genomic instability in, 34-35, 35f hereditary nonpolyposis, 2204-2207, 2206t2207t, 2207f Bethesda guidelines for, 2206t clinical features of, 2206, 2207t colon carcinogenesis in, 2160-2161 definition of, 2206, 2206t DNA mismatch repair genes in, 2206 gastric cancer in, 893 genetic testing in, 2226 microsatellite instability in, 40, 43-44, 2200t, 2203, 2203f, 2204t, 2206 small intestinal tumors in, 2148 type a (Lynch’s syndrome I), 2206 type b (Lynch’s syndrome II), 2206, 2207f hormone replacement therapy and, 2197-2198 hyperplastic polyps and, 2173, 2173t after ileal pouch–anal anastomosis, 2022-2023 incidence of, 2191-2193, 2192f-2193f after polypectomy, 2171-2172 in inflammatory bowel disease, 2209-2210, 2209f-2210f, 2227, 2229f in juvenile polyposis, 2209 lymphatic invasion by, 2215, 2217f matrix metalloproteinases in, 2204 metachronous, 2208 metastatic, 2204, 2205f, 2213, 2216-2218 treatment of, 2231-2232 micronutrients and, 2195 microsatellite instability in, 2160-2161, 2200t, 2202-2203, 2203f, 2204t, 2233, 2233f molecular genetics of, 2200-2204, 2200t, 2201f-2203f, 2204t mucinous, 2211, 2213f, 2215 in Muir-Torre syndrome, 2208-2209 multistep nature of, 34-35, 35f natural history of, 2212-2215 nonadenocarcinomatous, 2238 NSAIDS/aspirin and, 2196, 2198, 2199f, 2199t pathogenesis of, 2193-2199, 2194t bile acids and, 2194 calcium intake and, 2195-2197, 2199f, 2199t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Colorectal cancer (Continued) carcinogens and fecal mutagens in, 2195 difluoromethylornithine (DFMO) and, 2198-2199, 2199f, 2199t fat intake and, 2193-2194 fiber intake and, 2194-2195, 2197, 2199f, 2199t folate intake and, 2197, 2199f, 2199t micronutrients and, 2195 NSAIDS/aspirin and, 2196, 2198, 2199f, 2199t vitamin D and, 2195-2197, 2199f, 2199t vitamins and, 2195 pathology of gross, 2211, 2211f-2212f histologic, 2211-2212, 2213f prognosis and, 2215-2216, 2217f personal adenoma/carcinoma history in, 2208, 2227, 2228t in Peutz-Jeghers syndrome, 2209 in primary sclerosing cholangitis, 1163-1164 prognosis in, 2215-2218, 2216t, 2217f pseudomelanosis coli and, 2247, 2247f risk factors for, 2207-2211, 2208t scirrhous, 2211, 2215 screening for, 2219-2230 approach to, 2226-2227 in average-risk group, 2226 barium enema in, 2221t, 2224 best case values in, 2221t capacity for, 2229-2230 carcinoembryonic antigen in, 2225-2226 chromoendoscopy in, 2224 colonoscopy in, 2221t, 2224-2225 computed tomography colonography in, 2224-2225, 2225f costs of, 2221, 2221t in familial adenomatous polyposis, 2226-2227 fecal DNA testing in, 2226 fecal occult blood testing in, 2221-2225, 2221t-2223t. See also Fecal occult blood test (FOBT). genetic testing in, 2226 in hereditary nonpolyposis colorectal cancer, 2226-2227 in high-risk groups, 2226-2227 in inflammatory bowel disease, 2227, 2229f inguinal hernia and, 388 insurance coverage for, 2227-2229 principles of, 2219-2221, 2220t in prior adenoma or colorectal cancer, 2227, 2228t proctosigmoidoscopy in, 2221t, 2223-2224 quality assurance in, 2230 after surgical resection, 2230-2231 in underserved populations, 2229-2230 signet ring, 2215 small intestinal obstruction and, 2113 sporadic, 2207t staging of, 2213-2215, 2214t-2215t prognosis and, 2215, 2217f synchronous, 2208 treatment of, 2230-2238 argon plasma coagulation in, 2237 chemoradiotherapy for adjuvant, 2233-2234, 2234f neoadjuvant, 2234 chemotherapy for, 2232-2237 adjuvant, 2232-2234, 2233f-2234f in advanced disease, 2234-2237, 2235f-2237f in rectal cancer, 2233 endoscopic therapy for, 2237-2238, 2238f immunotargeted therapy and immunotherapy for, 2237 laser therapy for, 2237 in metastatic disease, 2231-2232 photodynamic therapy for, 2237-2238 radiation therapy for, 2237 second-look procedures in, 2231 snare cautery in, 2237, 2238f

Colorectal cancer (Continued) stent therapy for, 2238 surgical, 2230-2232, 2231f in Turcot’s syndrome, 2208 in ulcerative colitis, 2005-2008, 2006f-2007f, 2209-2210, 2209f-2210f cumulative risk of, 2209, 2209f dysplasia in, 2209-2210, 2210f prevention of, 2008 screening for, 2227, 2229f venous invasion by, 2215, 2217f vitamin D and, 2195-2197, 2199f, 2199t vitamins and, 2195 Wnt signaling in, 40-41 Colostomy for constipation, 283 for fecal incontinence, 256 for ulcerative colitis, 2024 Columns of Morgagni, 2257, 2258f Coma after hematopoietic stem cell transplantation, 549 hyperammonemic, in urea cycle defects, 1274 sprue, 1825 Comfrey, sinusoidal obstruction syndrome from, 1376 Common variable immunodeficiency hepatitis C in, 1326 malabsorption in, 1764, 1829t, 1830 Complementary and alternative medicine (CAM), 2287-2300 for constipation, 283, 2293-2295, 2294t definition of, 2287 demographics of, 2287-2288 for diarrhea, 2293-2295, 2294t epidemiology of, 2287 for functional dyspepsia, 2289-2291, 2290t for gastrointestinal cancer, 2297-2298 for inflammatory bowel disease, 2292-2293 for irritable bowel syndrome, 2103, 2291-2292, 2292t for liver disease, 2295-2297, 2295t for nausea and vomiting, 2288-2289, 2289t rationale for use of, 2288 safety and regulation of, 2298-2299 types of, 2287, 2288t Complete blood count (CBC) in eating disorders, 128 in jaundice, 331 Computed tomography in abdominal abscess, 413-414, 413f-414f in acute abdominal pain, 156 in acute acalculous cholecystitis, 1143 in acute calculous cholecystitis, 1115 in acute pancreatitis, 960f, 972, 973f, 975-976, 975t in adenomyomatosis, 1148-1149 in amebic liver abscess, 1369f in appendicitis, 156-157, 2065, 2065f in autoimmune pancreatitis, 990f, 991 in carcinoid tumors, 485 in celiac disease, 1811 in cholangiocarcinoma, 1174 in chronic pancreatitis, 1000t, 1001, 1001f in Clostridium difficile-associated diarrhea and colitis, 1895-1896, 1895f in diverticulitis, 2079-2080, 2079f in esophageal cancer, 753, 756-758, 758f in focal nodular hyperplasia of liver, 1588, 1588f in gallbladder carcinoma, 1179-1180, 1179f in gastric cancer, 901 in gastric motility assessment, 804 in gastrointestinal bleeding, 292 lower, 310 obscure, 318 in gastrointestinal stromal tumors, 466, 466f in hepatocellular carcinoma, 1573, 1573f in hereditary hemochromatosis, 1245-1246, 1246f in hydatid cyst, 1364, 1364f in intestinal endometriosis, 2254, 2254f in intussusception, 2114f, 2115

Computed tomography (Continued) in jaundice, 332, 333t in malabsorption, 1749 in mesenteric venous thrombosis, 2037, 2037f in neonatal cholestasis, 1052 in pancreatic cancer, 1020-1021, 1021f in pancreatic endocrine tumors, 515-516, 516t in pancreatic intraductal papillary mucinous tumor, 1031, 1031f in peritonitis, 614-615 in portal hypertension, 1497, 1498f in pregnancy, 627 in pyogenic liver abscess, 1367, 1367f in recurrent pyogenic cholangitis, 1168, 1168f single photon emission, in gastric motility assessment, 805 in small intestinal obstruction, 2106t, 2108-2109, 2108f-2111f in small intestinal tumors, 2150-2151 in tropical sprue, 1827 in ulcerative enteritis, 2055 in vomiting, 204-205 in Whipple’s disease, 1836, 1837f Computed tomography angiography, in acute mesenteric ischemia, 2031 Computed tomography cholangiography in acute mesenteric ischemia, 2031, 2031f in gallstone disease, 1108t, 1111-1112, 1111f multidetector, 1186, 1186f Computed tomography colonography in colorectal adenoma, 2166-2167, 2167f in colorectal cancer, 2224-2225, 2225f in gastrointestinal bleeding, 310 Computed tomography enteroclysis in gastrointestinal bleeding, 318 in small intestinal obstruction, 2109, 2111f in small intestinal tumors, 2150, 2150f Computed tomography enterography in Crohn’s disease, 1957, 1957f in gastrointestinal bleeding, 318 in ulcerative enteritis, 2055 in vomiting, 204-205 Computed tomography gastrography, in gastric cancer, 900 Condensing vacuoles, 914 Conditioning, physiologic, in biopsychosocial model, 339-340 Condylomata acuminata, 2270-2271, 2271t Confocal endomicroscopy, in esophageal cancer, 755, 755f Congenital anomalies of bile duct, 1050 of duodenum, 780t, 785-787 of enteric nervous system, 1636-1640 of esophagus, 668-675, 669t-670t fecal incontinence associated with, 258 of gallbladder, 1050 internal hernias in, 392-395, 393f of intestines, 1626-1641, 1627t of liver, 1205-1206, 1205f-1206f of pancreas, 917-919 of stomach, 780-785, 780t Congestive hepatopathy, 1380, 1381f Connective tissue disorders cutaneous manifestations of, 363 eosinophilia in, 432 malabsorption in, 1758 mixed, gastrointestinal manifestations of, 558t, 561 peritonitis in, 618 Connective tissue growth factors in carcinoid heart disease, 482-483 in esophageal cancer, 749-750 in radiation injury, 640 Connexons, 1208 Consciousness, loss of, constipation in, 268 Constipation, 259-284, 1672-1673, 1673f in children, 262 chronic, after spinal cord injury, 580 in chronic intestinal pseudo-obstruction, 2142 classification of, 263t, 264, 265t clinical assessment of, 271, 272t colonic anatomy and function in, 262-264, 265f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxi

xxii

Index Constipation (Continued) complementary and alternative medicine for, 2293-2295, 2294t defecatory disorders and, 265-266, 265t-266t, 266f, 283 definition of, 259-260, 260t in depression, 270 in descending perineum syndrome, 266-267 in diabetes mellitus, 268, 575 diagnostic testing in of colonic transit time, 272-273, 273f of defecation physiology, 273-274 to exclude structural disease of the gut, 272 to exclude systemic disease, 272 for physiologic measurements, 272-274, 273f diet and, 262, 271, 275-276, 275t in diminished rectal sensation, 267 drug therapy causing or aggravating, 262, 263t, 270 colonic ischemia and, 2039 in eating disorders, 128-130, 135, 271 education and, 262 in elderly persons, 260-262 in enteric nerve disorder, 269-270 epidemiology of, 260, 261t ethnicity and, 262 exercise and, 262 functional, 264, 265t gender and, 260 history in, 271 in hypercalcemia, 268 in hypothyroidism, 268 incidence of, 260 in irritable bowel syndrome, 2092 after laxative abuse, 137 in loss of conscious control, 268 in muscular dystrophy, 268 normal-transit, 264-265, 265t in obstructive disorders, 269 opioid-induced, 2284, 2285t palliative care for, 2284, 2285t in Parkinson’s disease, 268 pathophysiology of, 264-266, 266f, 266t physical examination in, 271, 272t prevalence of, 260, 261t psychological disorders causing or aggravating, 270-271 public health perspective on, 260 in rectal prolapse, 267-268 in rectocele, 266, 267f risk factors for, 260-262, 262t secondary, 263t, 264 slow-transit, 265, 265t colonic inertia in, 265 interstitial cells of Cajal in, 264 in smooth muscle disorders, 269 socioeconomic status and, 262 after solid organ transplantation, 541-542 in solitary rectal ulcer syndrome, 267-268 after spinal cord lesions, 268-269 stercoral ulcers and, 2052 in structural disorders of colon, rectum, anus, and pelvic floor, 269-270 symptoms of, 259-260, 260t treatment of, 274-283, 274f anorectal biofeedback training in, 282-283 botulinum toxin in, 281 chloride channel activator in, 280 cholinergic agents in, 281 colchicine in, 281-282 commercial fiber products in, 275t, 276 complementary and alternative medicine in, 283 defecation training in, 282 diet in, 275-276, 275t enemas and suppositories in, 277t, 279-280 fluid intake in, 275 general, 274-276 laxatives in, 275t, 276-282, 277t lifestyle changes in, 274-275 linaclotide in, 281-282 neurotrophins in, 281 osmotic laxatives in, 276, 277t

Constipation (Continued) peripheral mu-opioid antagonists in, 281 prokinetic drugs in, 280-281 psychological support in, 275 reassurance in, 274 sacral nerve stimulation in, 283 stimulant laxatives in, 277t, 279 stool softeners and emollients in, 277t, 279 surgical, 283 Contiguous gene syndrome, 1275 Continence, fecal. See Fecal incontinence. Continence mechanism anal sphincter in, 242, 242f anorectal sampling in, 242-243 pudendal nerve in, 242 Continuous ambulatory peritoneal dialysis, peritonitis with, 616 Contractile front velocity, in esophageal pressure topography, 695-696, 696f, 698f Cooking oils, adulterated, hepatotoxicity of, 1453-1454 COP (coatamers), 1209 Coping strategies, in biopsychosocial model, 342 Coporphyria, hereditary, 363 Copper absorption of, 1726 in short bowel syndrome, 1781t dietary, 56t-57t, 80-81, 80t, 1726 hepatic, in Wilson disease, 1253t, 1254 hepatotoxicity of, 1453 lysosomal aggregates of, 1252 malabsorption of, 1740 serum, in Wilson disease, 1253, 1253t toxicity of, in Bedlington terriers, 1250-1251 transport of, 1249-1251, 1250f. See also Wilson disease. urinary, in Wilson disease, 1253t, 1254 Copper transporters, 1726 Coproporphyria, hereditary, 1267, 1267t Cor pulmonale, gastrointestinal manifestations of, 584 Corn syrup, 1707 Cornstarch, for glycogen storage disease, type 1, 1263 Coronary artery disease, liver transplantation and, 1596 Corticosteroids. See Glucocorticoids. Corticotropin tumors secreting, 515 for ulcerative colitis, 1995 Corticotropin-releasing factor in brain-gut interactions, 343 in postoperative ileus, 2125 in stress-immune response, 345-346 Cotrimoxazole, hepatotoxicity of, 1432 Cough, in gastroesophageal reflux disease, 179-180, 715 Cowden’s syndrome, 2181t, 2185t, 2187 cancer risks and screening recommendations for, 2181t cutaneous manifestations of, 364 Coxiella burnetii infection, hepatic manifestations of, 1353 Creatinine, serum in gastrointestinal bleeding, 288 in hepatorenal syndrome, 1547 Creatinine-height index, for nutritional assessment, 68, 68t CREST (calcinosis, Raynaud’s phenomenon, esophageal dysfunction, sclerodactyly, and telangiectasia) syndrome, cutaneous manifestations of, 361 Cricopharyngeal bar dysphagia in, 689, 689f treatment of, 700-701 Cricopharyngeal myotomy for oropharyngeal dysphagia, 700 for Zenker’s diverticula, 700 Cricopharyngeus, 677-678, 678f Crigler-Najjar syndrome hyperbilirubinemia in, 1228-1229 jaundice in, 325-326, 326t

Crisis intervention, for chronic gastrointestinal disorders, 349 Critical flicker frequency test, in hepatic encephalopathy, 1545 Critically ill patients, jaundice in, 329 Crohn’s disease, 1941-1974 abdominal abscess in, 1951 abdominal pain in, 1949, 1953 acute pancreatitis in, 969 anal, 1950 anal skin tags in, 2263, 2265f anemia in, 1953 aphthous ulcers in, 1948, 1954 of appendix, 1950, 2069 arthralgia in, 1954 arthropathy in, 1953-1954 ATG15L1 gene in, 1944t-1945t, 1945 barium radiography in, 1956-1957 cancer risk in, 1972-1973 capsule endoscopy in, 1958 in celiac disease, 1813 in children and adolescents, 1971-1972 classification of disease in, 1952 clinical manifestations of, 1949-1955 Clostridium difficile infection in, 1895 coagulation in, 1955 colitis in, 313, 1950, 1990t colorectal cancer in, 2209-2210, 2227 complementary and alternative medicine for, 2292-2293 computed tomography enterography in, 1957, 1957f coping with, 1973 costs of, 1971 cutaneous manifestations of, 359-360, 359f, 1950, 1954 diagnosis of, 1956-1960, 1957f-1958f age at, 1942 diarrhea in, 214-215, 1950, 1952-1953 differential diagnosis of, 1955, 1956t disease activity measurements in, 1959-1960, 1960t disease behavior in, 1950-1952 disease location in, 1949 in elderly, 1972 endoscopic ultrasonography in, 1957-1958 endoscopy in, 1958, 1958f environmental factors in, 1946 eosinophilia in, 432-433 epidemiology of, 1942 episcleritis in, 1954-1955 erythema nodosum in, 1954 esophageal, 1950 ethnicity and, 1943 etiology and pathogenesis of, 1942-1947 environmental factors in, 1946 genetics in, 1943-1946, 1944t-1945t immune factors in, 1946-1947, 1947f initiating events in, 1942-1943 extraintestinal manifestations of, 1953-1955 fat wrapping in, 1949 fertility in, 1972 fever in, 1953 fibrosis in, 1949 fistula in, 422-423, 1947-1951, 1947f, 1951f anal, 2268-2269 colovesicular, 1951 enterocutaneous, 1951 enterovaginal, 1951 enterovesicular, 1951 gastrointestinal, 1951 perianal, 1950-1951, 1951f rectovaginal, 1951 gallstone disease in, 1101, 1955 gastritis in, 852-853, 852f, 855-856 gastroduodenal, 1950 gastrointestinal ulcers in, 864, 1948 gender and, 1942 genetic findings in, 1943-1946, 1944t-1945t genitourinary manifestations of, 1955 granuloma in, 1948, 1948f heart in, 1955 hepatobiliary manifestations of, 1955

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Crohn’s disease (Continued) history of, 1941 ileal disease in, 1949-1950 after ileal pouch–anal anastomosis, 2020 immune response in, 1946-1947, 1947f incidence of, 1942 inflammation and, 1947-1949, 1947f infliximab for, 1962t, 1967-1969 interleukin-12/23 in, 1944t-1945t, 1945-1946 irritable bowel syndrome and, 2092 jejunoileal, 1950 laboratory data in, 1956 lung in, 1955 magnetic resonance enterography in, 1957, 1957f malabsorption in, 1830 malnutrition in, 1953 metabolic bone disease in, 1954 metastatic, 359, 1954 morbidity in, 1972 mortality in, 1973 mucocutaneous manifestations of, 1954 musculoskeletal manifestations of, 1953-1954 nephrolithiasis in, 1955 NOD2/CARD15 gene in, 1943-1945, 1944t-1945t ocular manifestations of, 1954-1955 outcome of, psychosocial factors and, 346 pathology of, 1947-1949, 1948f perianal, 1950 in pregnancy, 629-630, 1969, 1972 primary sclerosing cholangitis in, 1156, 1955 in infants and children, 1057 probiotics for, 96 proctitis in, 1950 prognosis in, 1972-1973 protein-losing gastroenteropathy in, 441 proximal, 1950 pyloric metaplasia in, 1949 pyoderma gangrenosum in, 1954 recurrent, versus diversion colitis, 2243 renal manifestations of, 1955 scintigraphy in, 1957-1958 serologic tests in, 1959 sexuality in, 1972 sinus tracts in, 1949 small intestinal bacterial overgrowth in, 1773-1774 small intestinal tumors in, 2148 stress and, 346, 1946 strictures in, 1951-1952, 1952f treatment of, 1960-1971 adjunctive therapies for, 1970 aminosalicylates for, 1960-1961, 1962t antibiotics for, 1961-1963, 1962t anti–tumor necrosis factor agents for, 1962t, 1967-1969 biologic response modifiers for, 1962t, 1967-1970 glucocorticoids for, 1962t, 1963-1964 immune modulators for, 1962t, 1966-1967 medical, 1960-1970, 1962t methotrexate for, 1962t, 1966-1967 natalizumab for, 1962t, 1969-1970 novel therapies for, 1970 nutritional therapy for, 83, 1970-1971 surgical, 1971 thiopurine agents for, 1962t, 1964-1966, 1964f versus ulcerative colitis, 1958-1959, 1959t, 1988-1989, 1989t-1990t ultrasonography in, 1957-1958 uveitis in, 1954-1955 vascular complications of, 1955 weight loss in, 1953 Cronkhite-Canada syndrome, 364, 2188 Cryoglobulinemia in hepatitis B, 1296-1297 in hepatitis C, 1320 mixed IgG-IgM, gastrointestinal manifestations of, 563 Cryotherapy for anal warts, 2270, 2271t for colorectal cancer metastasis, 2232

Cryotherapy (Continued) for gastric vascular ectasia, 1515-1516 for hemorrhoids, 2261, 2263t Crypt-villus axis, of intestinal epithelium, 1677-1678, 1678f Cryptococcus, in HIV/AIDS, 527t, 530 Cryptosporidiosis, 1914-1916 clinical presentation in, 1915 diagnosis of, 1915 epidemiology of, 1914-1915 gastritis in, 852 after hematopoietic stem cell transplantation, 544 in HIV/AIDS, 526, 527t, 1916 immune response in, 1915 malabsorption in, 1829, 1829t pathogenesis and pathology of, 1915 prevention and control of, 1916 treatment of, 1915-1916 Cryptosporidium parvum, in elderly, 1876 Crypts of Lieberkühn, 1619f, 1621-1622 aberrant, colorectal adenoma and, 2159 in celiac disease, 1800 in malabsorption, 1746-1747, 1748f in radiation enteritis, 642-643, 643f radiation-induced apoptosis of, 639 in tropical sprue, 1825, 1826f in ulcerative colitis, 1982, 1982f CT. See Computed tomography. Cubulin-amnionless, 1719 Cuffitis, in ileal pouch–anal anastomosis, 2020 Cultural factors in biopsychosocial model, 340 in communication of psychological distress, 340 in pain behaviors, 340 Culture ascitic fluid, 1524 in bacterial overgrowth, 1775 in Helicobacter pylori infection, 840, 840t of intestinal bacteria, 1770 in small intestinal bacterial overgrowth, 1775 stool in acute diarrhea, 220 in amebiasis, 1909 in Campylobacter spp. infection, 1868 in chronic watery diarrhea, 222-223 in Clostridium difficile-associated diarrhea and colitis, 1896t, 1897 in enteric infection, 1849 in enterohemorrhagic Escherichia coli, 1856 in shigellosis, 1860 in typhoid fever, 1866 in typhoid fever, 1866 Cushing’s syndrome gastrointestinal manifestations of, 573t, 576 pancreatic endocrine tumors with, 493, 515 Cutaneous larva migrans, 367, 367f in hookworm infection, 1925-1927, 1926f Cutaneous manifestations. See Skin. Cutaneous porphyria, 1267-1268, 1267t Cutaneous vasculitis, in hepatitis A, 1282 Cutis laxa, X-linked, malabsorption in, 1759t-1762t CXC cytokines, in gut-associated lymphoid tissue, 29 Cyanoacrylate glue injection, for gastric variceal bleeding, 1512 Cycasin, hepatotoxicity of, 1455 Cyclic adenosine monophosphate (cAMP), in intestinal ion transport, 1682, 1682f, 1693, 1693f Cyclic guanosine monophosphate (cGMP), in intestinal ion transport, 1682, 1682f, 1693, 1693f Cyclin(s) in cell cycle regulation, 31, 32f in hepatic regeneration, 1213 Cyclin B, in esophageal cancer, 750 Cyclin D1 in cellular proliferation, 33 in esophageal cancer, 750

Cyclin-dependent kinase in hepatic regeneration, 1213 inhibitors of, 31-32 in regulation of cell cycle, 31, 32f Cyclizine, for vomiting, 207 Cyclooxygenase-1 (COX-1), in peptic ulcer disease, 863 Cyclooxygenase-2 (COX-2) in esophageal cancer, 750 in gastrointestinal tumors, 41 in peptic ulcer disease, 863 Cyclooxygenase-2 (COX-2) inhibitors cardiovascular risk of, 875-876 colorectal adenoma risk and, 2164 colorectal cancer and, 2196, 2198 in familial adenomatous polyposis, 2183-2184 hepatotoxicity of, 1437 NSAID enteropathy from, 2053 for NSAID ulcer prophylaxis, 874-875 after peptic ulcer bleeding, 303 ulcer healing and, 873 Cyclophilin inhibitor DEBIO-025, for hepatitis C, 1334 Cyclophosphamide, for ischemic bowel disease, 561 Cyclospora cayetanensis infection, 1916-1917 Cyclosporiasis, malabsorption in, 1829, 1829t Cyclosporine for celiac disease, 1817 for Crohn’s disease, 1967 hepatotoxicity of, 543 for liver transplantation, 1605-1606, 1606t nephrotoxicity of, 1609-1610 in pregnancy, 630 side effects of, 1998, 1998t for ulcerative colitis, 1998-1999, 1998t CYP2C9, 1417-1418 CYP2C19, 1417-1418 CYP2D6, 1417-1418 in autoimmune hepatitis, 1466 CYP2E1, 1418 in alcoholic liver disease, 1385 in ethanol metabolism, 1384 in halothane hepatitis, 1448 CYP3A4, 1418 CYP7A, 1418 Cyp7A1 in bile acid synthesis, 1076, 1078, 1705-1706 in gallstone disease, 1102, 1103t Cyst(s) choledochal, 1058-1059, 1058f-1059f, 1205 duplication. See Duplication. epidermoid, in familial adenomatous polyposis, 363, 2181-2182 gas-filled in pneumatosis coli, 2248-2250, 2249f-2250f in pneumatosis cystoides intestinalis, 240 hepatic, 1589-1590, 1589f hydatid, 1356t-1357t, 1363-1365, 1364f mesenteric, 620 mucus-filled, in colitis cystica profunda, 2252, 2252f neuroenteric, 1632 omphalomesenteric, 1629f, 1630 ovarian, ruptured, versus appendicitis, 2063t pancreatic. See also Pancreas, cystic neoplasms of; Pancreatic pseudocyst. congenital, 919 cyst fluid analysis in, 1028, 1028t differential diagnosis of, 1011, 1011t-1012t, 1012f endoscopic therapy for, 1041 peritoneal, benign, 619 Cystadenoma biliary, 1183-1184 serous pancreatic, 1027t-1028t, 1028-1030, 1030f versus pseudocyst, 1012t Cystic artery, 1049 Cystic duct anatomy of, 1048-1049, 1048f congenital anomalies of, 1050 Cystic duct remnant, after cholecystectomy, 1137

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxiii

xxiv

Index Cystic fibrosis, 940-949, 1641 atypical, 937 biliary tract in, 947-948 cancer risk in, 945 CFTR gene mutations in, 934f, 935-937, 936t clinical features of, 940, 941t diabetes mellitus in, 942 distal intestinal obstruction syndrome in, 944-945 gallbladder in, 947-948 gastroesophageal reflux disease in, 945 gastrointestinal manifestations of, 940, 941t genital abnormalities in, male, 948 intestinal manifestations of, 943 functional abnormalities in, 943 pathology of, 943 radiology of, 943 intussusception in, 945 jaundice in, 328 liver disease in, 945-947, 1277, 1277t clinical features of, 947 functional abnormalities in, 946-947 pathology of, 946, 946f radiologic features of, 946 treatment of, 947 liver transplantation for, complications after, 541 malnutrition in, 948-949 meconium ileus in, 943-944, 944f, 944t. See also Meconium ileus. nasal bioelectric response testing in, 940 nutrition in, 948-949 pancreatic manifestations of endocrine dysfunction in, 942 exocrine dysfunction in, 941-942 pancreatic enzyme replacement for, 942 pathology of, 936, 936f treatment of, 942-943 vitamin supplementation for, 942-943 pathology of, pancreatic, 940-941 prognosis in, 949 rectal prolapse in, 945 small intestinal bacterial overgrowth in, 1774 sweat testing in, 940 Cystic fibrosis transmembrane regulator (CFTR) in bicarbonate secretion, 922, 922f in chloride transport, 1683 gene for, 934f, 935-937, 936t. See also CFTR gene. Cystic vein, 1049-1050 Cysticercosis, in Taenia saginata/solium infection, 1932 Cystinuria, malabsorption in, 1713-1714, 1758-1763, 1759t-1762t Cytapheresis, for ulcerative colitis, 2001-2002 Cytochrome P450. See also CYP entries. in cirrhosis, 1418 developmental regulation of, 1418 in diabetes mellitus, 1418 in hepatic drug metabolism, 1417 induction of, 1418 nutritional effects on, 1418 pharmacogenetics and polymorphisms of, 1417-1418 Cytochrome P450 reductase, 1417 Cytokines. See also specific cytokines. in acute pancreatitis, 962-963 in alcoholic liver disease, 1387-1388, 1388f brain-gut interactions and, 346 in gastrointestinal tract, 13 in Helicobacter pylori infection, 837, 891 in hepatic regeneration, 1213-1214 in protein-energy malnutrition, 61, 62t in radiation injury, 639-640 in ulcerative colitis, 1980 Cytomegalovirus infection gastritis in, 849, 849f hepatitis in, 1347-1348, 1348f-1349f in HIV/AIDS diarrhea in, 527t, 528, 528f esophagitis in, 524, 524f hepatic, 533 after liver transplantation, 1608

Cytomegalovirus infection (Continued) mucous ulcers in, 357 after solid organ transplantation, 537-542, 539f Cytoprotective agents, for nonalcoholic fatty liver disease, 1410 Cytotoxic injection therapy, for esophageal cancer, 765 Cytotoxins bacterial, 1846 Clostridium difficile, 1896, 1896t Helicobacter pylori, 835, 838 Shigella, 1859 STX, in enterohemorrhagic Escherichia coli, 1856

D

D cells, 817-819, 915 antral enteroendocrine, 778 somatostatin-containing, 817-819, 820f Daclizumab, for ulcerative colitis, 2001 Dantrolene, hepatotoxicity of, 1436, 1438t Dapsone, for dermatitis herpetiformis, 1812 DCA. See Deoxycholic acid. DCC gene, in colorectal cancer, 2160, 2200t, 2202 DCT1, in zinc transport, 1725-1726 Death receptors, 32, 33f, 1214, 1420 Decongestants, colonic ischemia from, 2039 Deep venous thrombosis, in ulcerative colitis, 2012 Defecation. See also Fecal; Stool. anorectal motility during, 1670, 1670f colonic motility and, 1668-1670, 1669f denial of, 271 disorders of constipation and, 265-266, 265t-266t, 266f, 283 criteria for, 265, 266t operations for, 283 solitary rectal ulcer syndrome as, 2050 dyssynergic, 245 incomplete, fecal incontinence and, 245 normal, 264 physiology of, 264, 264f assessment of, 273-274 stimulation of, 264 stool size and consistency in, 264, 265f training in, for constipation, 282, 2293 Defecography in constipation, 273 in fecal incontinence, 248-249 in rectocele, 266 in solitary rectal ulcer syndrome, 2051 Deferoxamine, for hereditary hemochromatosis, 1246 Defibrillator, electrosurgery and, 656 Defibrotide, for sinusoidal obstruction syndrome, 1377 Deglutition. See Swallowing. Degos disease cutaneous manifestations of, 360-361, 361f gastrointestinal manifestations of, 562 vasculitis in, 2047 Dehydration. See also Fluid therapy. constipation and, 262 in infectious diarrhea, 1884-1885 Dehydroascorbate, 1718 Deletion, allelic, 38, 39f Dementia nutrition in, percutaneous endoscopic gastrostomy for, 90 in Whipple’s disease, 1837 Dendritic cells in food allergy, 140 immune function of, 29 Dentate line, 1622, 2257, 2258f Deoxycholic acid, 1093. See also Bile acid(s). accumulation of, 1086 bile lithogenicity and, 1101 synthesis and metabolism of, 1077f, 1077t, 1078

Deoxyribonucleic acid. See DNA. Depression, constipation in, 270 Dermatitis in glucagonoma, 506-507, 507f in hepatitis B, 1295 Dermatitis herpetiformis, 367-368, 368f celiac disease and, 1811-1812 gluten sensitivity in, 146 Dermatomyositis chronic intestinal pseudo-obstruction in, 2135 cutaneous manifestations of, 363-364, 364f gastrointestinal manifestations of, 558t, 561 Gottron’s papules in, 364, 364f Dermatosis-arthritis syndrome, in inflammatory bowel disease, 360 Des-γ-carboxy prothrombin, in hepatocellular carcinoma, 1572, 1572t Descending perineum syndrome constipation in, 266-267 fecal incontinence and, 245 Desflurane, hepatotoxicity of, 1449, 1449t Desipramine for bulimia nervosa, 134 for neuropathic pain, 2283t Desmoid tumors, in familial adenomatous polyposis, 2178f, 2180-2181, 2182t Desmosomes, 1208 in intestinal epithelium, 1676f, 1679 Detorsion, endoscopic, for gastric volvulus, 384-385 Developmental anomalies. See Congenital anomalies. Devon family syndrome, 2187 Dexamethasone for cholestasis of pregnancy, 631-632 for radiation-induced emesis, 642 for vomiting, 208 Dextrinase, α-limit, 1708, 1708f, 1708t Dextrins, α-limit, 1707, 1707f Dextropropoxyphene, cholestatic hepatitis with bile duct injury from, 1441 Diabetes, fibrocalculous pancreatic, 931 Diabetes mellitus abdominal pain in, unexplained, 575 acute cholecystitis in, 1131 after bariatric surgery, 117 celiac disease and, 1812 chronic intestinal pseudo-obstruction in, 2136, 2137f chronic pancreatitis and, 994, 997, 1010 constipation in, 268, 575 cystic fibrosis and, 942 cytochrome P450 expression in, 1418 diarrhea in, 574-575 diet for, 95-96 drug-induced liver disease in, 1417 esophageal hypomotility disorders in, 572 fecal incontinence in, 575 gallstone disease in, 1092, 1105 gastric dysfunction in, 572-574 gastroesophageal reflux disease in, 572 gastrointestinal manifestations of, 572-575, 573t gastroparesis in, 806-807, 807f glucagonoma and, 507 hepatitis C and, 1326 hepatobiliary disease in, 575 hereditary hemochromatosis and, 1242-1243 hereditary pancreatitis and, 950 insulin resistance in, regulation of, 19 after liver transplantation, 1608, 1608t, 1610 malabsorption in, 1766 megacolon in, 575 nonalcoholic fatty liver disease and, 1402 obesity and, 104, 106 pancreatic cancer and, 1018, 1020 pancreatic disease in, 575 small intestinal motility in, 1649-1650, 1657 somatostatinoma and, 512 tropical pancreatitis and, 952 Dialectical behavior therapy, for eating disorders, 132-133

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Dialysis. See also Hemodialysis. albumin for acute liver failure, 1568 for hepatic encephalopathy, 1546 for hepatorenal syndrome, 1549 ascites with, 1528, 1536 peritoneal, peritonitis with, 616, 1528 Diaphragm bacterial clearance through, 613, 613t crural antireflux function of, 707-708, 708f esophagogastric junction pressure and, 684 in swallowing, 682-683, 683f diseases of, 621-622 eventration of, 621 hiccups and, 621-622 NSAID-induced ulcer of, 2244-2245, 2245f tumors of, 621 Diaphragmatic breathing, for rumination, 204 Diaphragmatic hernia. See Hernia, diaphragmatic. Diaphragmatic hiatus, 683f Diarrhea, 211-232, 1673 acute, 216 differential diagnosis of, 216-217, 217t empiric therapy for, 226, 227t evaluation of patients with, 219-220, 220f in pregnancy, 1872 antibiotic-associated, 217-218, 229-230, 1889-1890, 1890t. See also Clostridium difficile-associated diarrhea and colitis. bile acid malabsorption causing, 226, 229, 1087 after bowel resection, 228-229 Brainerd, 231 in capillariasis, 1925 in carcinoid syndrome, 482 in celiac disease, 1804 chloride, congenital, 1641 after cholecystectomy, 229 chronic, 216 differential diagnosis of, 217, 217t empiric therapy for, 226-227, 227t evaluation of patients with, 220-226, 221f-222f in VIPoma, 509-510 clinical classification of, 215-216 Clostridium difficile-associated. See Clostridium difficile-associated diarrhea and colitis. complementary and alternative medicine for, 2293-2295, 2294t complex, 214-215, 214f-215f in Crohn’s disease, 1950, 1952-1953 definition of, 211 in diabetes mellitus, 574-575 drug-induced, 574 dietary history in, 219 differential diagnosis of, 216-218, 217t drugs and poisons causing, 217-218, 217t during enteral nutrition, 94-95, 230 epidemiologic features of, 216, 216t, 219 evaluation of patients with, 218-226 factitious, 230-231, 230t fatty, 216, 218, 225-226 fecal impaction and, 230 fecal incontinence and, 211, 246 functional, 219, 227 after gastric surgery, 228 in glucagonoma, 507 after hematopoietic stem cell transplantation, 550-552, 551t, 555 in hematopoietic stem cell transplantation candidates, 544 history in, 218-219 in HIV/AIDS, 526-530 differential diagnosis of, 526, 526t drug-induced, 530 evaluation and management of, 526-530, 527f-529f, 527t, 530t HIV as pathogen in, 529 in hospitalized patients, 229-230, 1872 in hyperthyroidism, 575-576

Diarrhea (Continued) after ileostomy, 229 infectious, 213 adenovirus in, 1870t, 1871-1872 Aeromonas spp. in, 1853-1854 antibiotics for, 1865t astrovirus in, 1870t, 1872 Bacillus cereus in, 1879t-1880t, 1882 bacterial classification of, 1846 pathogens in, 1843-1846 bacterial virulence factors in, 1844-1846, 1845f breast-feeding and, 1844 calicivirus in, 1870t, 1871 Campylobacter spp. in, 1865t, 1867-1868, 1868f in cholera, 214, 215f, 1850-1852. See also Cholera. classification of, 1846-1847, 1848t Clostridium difficile in, 230. See also Clostridium difficile-associated diarrhea and colitis. in cryptosporidiosis, 1915 in Cyclospora cayetanensis infection, 1916 diagnosis of, 222-223, 225, 1846-1849, 1847f, 1847t in elderly, 1875-1876 endoscopy in, 1849 Escherichia coli in, 1854-1857, 1854t. See also Escherichia coli. fecal leukocytes in, 1848-1849, 1848t gastric acidity and, 1844 in giardiasis, 1913 hospital-acquired, 1872 in hospitalized patients, 1872 host defense factors against, 1844 immune response in, 1844 inflammatory, 1848, 1848t intestinal microflora and, 1844 intestinal motility and, 1844 irritable bowel syndrome after. See Irritable bowel syndrome, postinfection. in Isospora belli infection, 1917 laboratory tests in, 1847f, 1849 mucus and, 1844 noninflammatory, 1847-1848 Norovirus in, 1870t, 1871 pathogens in, 216-217, 217t antimicrobial therapy based on, 1847t invasive, 1857-1869 toxigenic, 1849-1857 viral, 1869-1872 Plesiomonas shigelloides in, 1854 in pregnancy, 1872 probiotics for, 96 in proctitis syndrome, 1848 rotavirus in, 1869-1871, 1870f, 1870t Salmonella spp. in nontyphoidal, 1861-1864. See also Salmonellosis. typhoidal, 1864-1867. See also Typhoid fever. Shigella spp. in, 1857-1861. See also Shigellosis. stool culture/examination in, 1849 susceptibility to, 1843-1846 torovirus in, 1870t, 1872 in travelers, 1872-1875, 1873t-1875t. See also Traveler’s diarrhea. treatment of, 1884-1886 antimicrobial agents for, 1847t, 1885-1886 diet for, 1885 fluid therapy for, 1851t, 1884-1885 nonspecific, 1886, 1886t in Trichuris trichiura infection, 1927-1928 Vibrio spp. in, 1850-1853 viral pathogens in, 1869-1872, 1870t Yersinia enterocolitica in, 1865t, 1868-1869 inflammatory, 216 chronic, 218, 225, 225f in inflammatory bowel disease, 214-215 intestinal inflammation and, 1688

Diarrhea (Continued) in irritable bowel syndrome, 215, 219, 227, 2092 large-volume, 216 in laxative abuse, 230-231, 230t in malabsorption, 72, 218, 1745t in microscopic colitis, 228, 2241 in microsporidiosis, 1917 nocturnal, 218-219 of obscure origin, 231, 231t osmotic, 216 causes of, 212-213, 212t evaluation of patients with, 224, 224f palliative care for, 2285 pathophysiology of, 212-215, 212f, 212t physical examination in, 219 PINES acronym in, 214, 214f-215f postsurgical, 228-229 radiation-induced, 643 secretory, 216 in bile acid malabsorption, 1755-1756 causes of, 212t, 213-214 chronic, 217 evaluation of patients with, 222, 223f idiopathic, 218, 231 in short bowel syndrome, 1783 small-volume, 216 sodium, congenital, 213, 1641 after solid organ transplantation, 538t, 541-542 in somatostatinoma, 512 Starbucks, 217-218 traveler’s. See Traveler’s diarrhea. treatment of, 226-227, 227t tropical, enteropathogens causing, 1821-1822, 1822t in ulcerative colitis, 1983 in VIPoma, 509-510 watery, 216 chronic, 217-218, 217t, 222-224, 223f-224f in Whipple’s disease, 1836 in Zollinger-Ellison syndrome, 502 Diastasis recti, ventral hernia versus, 388 Diastolic dysfunction, in cirrhotic cardiomyopathy, 1553 Diazo reaction, in serum bilirubin measurement, 324 Diazoxide, for insulinoma, 498 Dichloroacetic acid, for anal warts, 2270, 2271t Diclofenac hepatotoxicity of, 1438t, 1439 for post-ERCP pancreatitis, 967 Dicyclomine, for gastric motility disorders, 812, 813t Didanosine, hepatotoxicity of, 1430 Dientamoeba fragilis infection, 1907f, 1914 Diet, 95-96. See also Nutrition; Nutritional therapy. in aphthous ulcers, 356 colonic transit time and, 262-263 colorectal adenoma risk and, 2163-2164 commercial programs for, 109 constipation and, 262, 271, 275-276, 275t Crohn’s disease and, 1946 in cystic fibrosis, 948-949 diabetic, 95-96 in diarrhea, 219 diverticulosis and, 2075-2076 in eosinophilic gastrointestinal disorders, 433-434, 433t in esophageal adenocarcinoma, 748 in esophageal squamous cell carcinoma, 747 exclusion in food allergy, 146-148 in irritable bowel syndrome, 2101 in fecal incontinence, 252 fiber- and residue-restricted, 95 in functional dyspepsia, 191 gallstone disease and, 1091 in gastric cancer, 892 in gastroesophageal reflux disease, 721 gastrointestinal cancer and, 2298 in gastroparesis, 574

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxv

xxvi

Index Diet (Continued) gluten-free in celiac disease, 84, 1813-1814, 1813t in dermatitis herpetiformis, 1812 in ulcerative enteritis, 2057 in glycogen storage disease type III, 1264 in hemorrhoids, 2260, 2263t high-carbohydrate, in short bowel syndrome, 1785 high-fiber, in irritable bowel syndrome, 2101 in hyperemesis gravidarum, 628 hyperphagia, in short bowel syndrome, 1784-1785, 1785t-1786t in infectious diarrhea, 1885 in insulinoma, 498 intestinal bacteria and, 1769-1770 lactose-restricted, 95 liquid clear, 95 full, 95 low-carbohydrate, 108-109 low-energy, 108 low-fat, 95, 108 in short bowel syndrome, 1785 low-sodium, in ascites, 1536 obesity and, 102, 108-109, 108t in oncogenesis, 41-42 oxalate-restricted, in short bowel syndrome, 1786, 1787t pancreatic cancer and, 1017-1018 peptic ulcer disease and, 863 peptide-based, in short bowel syndrome, 1786 postgastrectomy, 95 post–gastric bypass, 95 in protein-losing gastroenteropathy, 443 in radiation enteritis, 645-646 in short bowel syndrome, 1783-1786, 1785t-1787t soft, 95 specialty, 96 in tropical sprue, 1827-1828 in tyrosinemia, 1271 in urea cycle defects, 1273-1274 very low-calorie, 109 in bulimia nervosa, 134-135 Dietary supplements hepatotoxicity of, 1456-1458, 1457t malabsorption from, 1756, 1757t Diethylpropion, for obesity, 110t, 111 Dieulafoy-type colonic ulcer, 2245 Dieulafoy’s lesions, 602-604 bleeding in, 304, 314, 318, 602-604, 604f Diffusion, fluid and electrolyte transport by facilitated, 1680, 1680f simple, 1679-1680, 1680f Difluoromethylornithine (DFMO) colorectal adenoma risk and, 2164 colorectal cancer and, 2198-2199, 2199f, 2199t DiGeorge’s syndrome, malabsorption in, 1764 Digestion and absorption adaptive changes in, 1726-1729 bariatric surgery and, 1732-1733 developmental, 1729-1732, 1730f-1731f enzymatic, 1727, 1727f mucosal, 1726-1728, 1728f nutritional requirements and, 1728-1729, 1729f signaling for, 1729 therapeutic implications of, 1729 after bariatric surgery, 1732-1733 of carbohydrate, 1707-1712 amylase in, 1707-1708, 1707f brush border membrane hydrolases in, 1708, 1708f, 1708t disaccharidase in, 1708-1710, 1709f exit from epithelium in, 1711-1712 mucosal defects of, 1739 mucosal transport in, 1710-1711, 1710f in neonates and infants, 1732 of fat, 1699-1706 brush border membrane transfer in, 1702-1703, 1703f emulsification in, 1699, 1700f

Digestion and absorption (Continued) intestinal bacteria metabolic activity and, 1772 intracellular processing in, 1703-1705, 1704f-1705f intraluminal pH in, 1699t, 1701 lipase in, 1699-1701, 1700f liver X receptors in, 1705-1706, 1705t, 1706f micelles and lipid-containing particles in, 1700f, 1701 in neonates and infants, 1730-1732 water layer in, 1701-1702 fluid and electrolyte transport in, 1675-1694. See also Fluid and electrolyte transport. gastrointestinal integration in, 1695-1698, 1696f, 1698f jejunum in, short bowel syndrome and, 1780, 1780f of minerals, 1722-1726 in neonates, 1729-1732, 1730f-1731f noninvasive evaluation of, 1749-1754, 1750t oral, 1695-1696 of protein, 1712-1717 amino acid transport in, 1716-1717, 1716f, 1716t at brush border membrane and in cytoplasm, 1713-1715, 1713f, 1714t exit from epithelium in, 1717 in neonates and infants, 1729-1730, 1732 pancreatic proteases in, 1712-1713, 1713f, 1713t pepsins in, 1712 peptide transport in, 1715-1716, 1715f, 1715t of trace elements, 1725-1726 adaptive changes in, 1728-1729, 1729f of vitamins, 1717-1722 adaptive changes in, 1728-1729 fat-soluble, 1720-1722, 1721f, 1721t water-soluble, 1717-1720, 1718t Digestive system, blood flow to. See Splanchnic circulation. Digital clubbing in Crohn’s disease, 1953-1954 in hepatopulmonary syndrome, 1550 in immunoproliferative small intestinal disease, 1825f Digoxin, drug interactions of, with proton pump inhibitors, 871 Diloxanide furoate for amebiasis, 1910t for amebic liver abscess, 1369 Diltiazem hepatotoxicity of, 1433 for levator ani syndrome, 2273 topical, for anal fissure, 2265t, 2266 Dimethylformamide, hepatotoxicity of, 1452 Dipeptidases, 1713-1715, 1713f, 1714t Dipeptides, 1715-1716, 1715f, 1715t Diphenhydrinate, for vomiting, 207 Diphenoxylate with atropine for diarrhea, 226, 227t for fecal incontinence, 252-253 for infectious diarrhea, 1886 for malabsorption-related diarrhea, 72 for short bowel syndrome, 1784, 1784t Diphenylmethane derivatives, for constipation, 277t, 279 Diphyllobothrium spp. infection, 1931, 1931f Dipylidium caninum infection, 1933 Direct immunofluorescent antibody (DFA) microscopy in cryptosporidiosis, 1915 in giardiasis, 1913 Disaccharidases. See also Carbohydrate. biosynthesis and regulation of, 1708-1710, 1709f deficiency of in celiac disease, 1812-1813 malabsorption in, 1759t-1762t, 1763 hydrolysis of, 1708, 1708f, 1708t Disaccharides, nonabsorbable, for hepatic encephalopathy, 1545

Disc battery, ingestion of, 403 Discriminant analysis, for nutritional assessment, 68-69, 68t Disease-modifying antirheumatic drugs (DMARDs), gastrointestinal complications of, 559 Disease susceptibility, brain-gut interactions and, 345-346 Disinfection, in endoscopy, 655-656 Disse, space of, 1204, 1207, 1208f, 1490, 1490f Distal contractile integral, in esophageal pressure topography, 697, 699f Disulfiram, hepatotoxicity of, 1433 Diuretics for ascites, 1537-1538 colonic ischemia from, 2040 for protein-losing gastroenteropathy, 443 thiazide, malabsorption with, 1757t Divalent metal transporter 1 (DMT-1), 1724, 1724f in hereditary hemochromatosis, 1240 Diversion colitis, 2242-2243 Diverticula, 371-378 anatomy of, 2074, 2074f bleeding, 2085-2088, 2086f-2088f. See also Diverticulosis, colonic, bleeding in. duodenal, 376-377 extraluminal, 376, 376f intraluminal, 376, 377f esophageal, 373-374 clinical presentation and diagnosis of, 373-374, 373f epiphrenic, 373-374, 373f etiology and pathogenesis of, 373, 373f traction, 373 treatment and prognosis in, 374 false, 371. See also Pseudodiverticula. esophageal intramural, 374-375, 374f gastric, 375-376, 375f, 780t, 781-782 intramural, 375, 375f juxtacardiac, 375, 375f hypopharyngeal. See Diverticula, Zenker’s. ileal, 377-378 inverted, 2077 irritable bowel syndrome and, 2077 jejunal, 377-378 juxtapapillary, 376 Meckel’s, 1628-1630, 1629f bleeding in, 317-318, 1628-1629 intestinal obstruction in, 1629 nutritional therapy for, 83 perforation of, 2078, 2078t small intestinal, bleeding in, 318 true, 371 windsock, 376, 377f Zenker’s, 371-373 clinical presentation and diagnosis of, 371-372, 372t etiology and pathogenesis of, 371 oropharyngeal dysphagia in, 686, 689, 689f treatment and prognosis in, 372-373, 373f Diverticulectomy, for Zenker’s diverticula, 700 Diverticulitis, 2078-2083 abdominal pain in, 154t, 158 abscess in, 2080-2082, 2081f appendiceal, 2070 cecal, versus appendicitis, 2063t clinical features of, 2078-2079 colonic obstruction in, 2116-2117 complicated, 2081-2083 computed tomography in, 2079-2080, 2079f definition of, 2078 diagnosis of, 2079-2080 differential diagnosis of, 2079 in elderly, 2084 endoscopy in, 2080 fistula in, 2082 in immunocompromised patients, 2084 magnetic resonance imaging in, 2080 obstruction in, 2083 pathophysiology of, 2078, 2078t peritonitis/free perforation in, 2082 probiotics for, 96

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Diverticulitis (Continued) radiography in, 2079, 2079f recurrent, 2083 right-sided, 2084, 2084f sigmoid, versus appendicitis, 2063t treatment of, 2080-2081 ultrasonography in, 2080 uncomplicated, 2078-2081 in young patients, 2083 Diverticuloscope soft, 373 Weerda, 372-373, 373f Diverticulosis, 2073-2090 colonic, bleeding in, 311-312, 312f angiography in, 312 classification of, 311 endoscopic hemostasis in, 312 endoscopic stigmata of, 310, 312f surgery in, 312 complicated, 2078-2083. See also Diverticulitis. epidemiology of, 2073-2074, 2074t hemorrhage in, 2085-2088, 2086f-2088f jejunal, 2139 lower gastrointestinal bleeding in, 311-312, 312f pathogenesis of, 2075 pathologic anatomy of, 2074, 2074f prevalence of, 2073 risk factors for, 2074, 2074t segmental colitis associated with, 2084 treatment of, 2077-2078 uncomplicated, 2076-2078 asymptomatic, 2076 symptomatic, 2076-2078, 2076f DNA mutations of, detection of, molecular diagnostic strategies for, 43-44, 43t-44t nuclear, 1208 DNA microsatellite markers, in Wilson disease, 1255, 1255f DNA mismatch repair genes, 36t, 40 in hereditary nonpolyposis colorectal cancer, 2206 DNA stool testing in colorectal adenoma, 2167 in colorectal cancer, 2226 DNA tumor content, in colorectal cancer, 2215-2216 Docusate sodium, for constipation, 277t, 279 Dog tapeworm infection, 1933 Domperidone for functional dyspepsia, 191-193 for gastric motility disorders, 812, 813t for gastroparesis, 574 for irritable bowel syndrome, 2103 for vomiting, 207 Dopamine in emetic reflex, 197-198, 198f in gastrointestinal tract, 11 in intestinal ion transport, 1690 receptors for, 11 Dopamine D2 receptor antagonists, for vomiting, 207 Doppler ultrasonography in Budd-Chiari syndrome, 1373 in hepatocellular carcinoma, 1573 in peptic ulcer bleeding, 296 in portal vein thrombosis, 1378-1379, 1378f Double-bubble sign in annular pancreas, 787 in duodenal atresia/stenosis, 785-786, 786f in duodenal obstruction, 1632, 1632f Double-duct sign, in ampullary carcinoma, 1182, 1182f Down-regulation, 14 Down’s syndrome, duodenal atresia/stenosis and, 785 Doxorubicin, for metastatic pancreatic endocrine tumors, 520-521 Doxycycline esophagitis from, 737 prophylactic, for traveler’s diarrhea, 1875t for Whipple’s disease, 1841t DPC4 gene, in pancreatic cancer, 1020

DRB1 gene, in autoimmune hepatitis, 1463, 1464f, 1469 Dronabinol, for vomiting, 208 Droperidol, for vomiting, 207 Drotrecogin, for abdominal abscess, 419 Drug(s) abuse of. See Substance abuse. acute pancreatitis from, 965, 965t in children, 938, 938t adverse reactions to cross-sensitivity in, 1416 definition of, 1413 idiosyncratic, 1413-1414 reporting of, 1414 allergic reactions to, 1422, 1431-1434, 1431t bezoar formation from, 405, 405f, 405t cholestasis from. See Cholestasis, drug-induced. chronic intestinal pseudo-obstruction from, 2139 colitis from, 1991 colonic ischemia from, 2039, 2039t constipation from, 262, 263t, 270 diarrhea from, 217-218, 217t in diabetes mellitus, 574 in HIV/AIDS, 530 eosinophilia from, 432 fecal incontinence from, 245 gallstone disease from, 1092 gastropathy from, 856 hepatic angiosarcoma from, 1451-1452 hepatic elimination of, 1417 hepatic metabolism of, 1417-1419. See also Liver, drug metabolism in. hepatitis from. See Hepatitis, drug-induced. liver disease from, 1413-1446. See also Hepatotoxicity; Liver disease, drug-induced. malabsorption of, 1756, 1757t nausea and vomiting from, 199t, 200 nonalcoholic fatty liver disease from, 1402 obesity related to, 101-102, 101t peptic ulcer disease from, 862 in pregnancy, 626, 626t-627t sinusoidal obstruction syndrome from, 1445t, 1456 vomiting from, 199t, 200 weight gain from, 101-102, 101t Dual-energy x-ray absorptiometry (DEXA) scan, in eating disorders, 128 Dubin-Johnson syndrome, 1085t hyperbilirubinemia in, 1229 jaundice in, 326, 326t Duchenne muscular dystrophy chronic intestinal pseudo-obstruction in, 2138 gastrointestinal manifestations of, 581 Ductal obstruction hypothesis of chronic pancreatitis, 987 Ductal plate malformation, of liver, 1205-1206 Ductopenia, in primary biliary cirrhosis, 1480 Dukes classification of colorectal cancer, 2213-2215, 2214t Duloxetine, for neuropathic pain, 2283t Dumping syndrome, 810 after gastric surgery, 228 nutritional therapy for, 84 Dunbar syndrome, 608-609, 609f Duodenal bulb (cap), 779 Duodenal papilla, 779 Duodenal stump, difficult, closure of, 882 Duodenal ulcers. See also Peptic ulcer disease. antacids for, 869-870 bleeding. See Peptic ulcer bleeding. bulbar, surgical management of, 882 endoscopy in, 865f gastric acid secretion in, 828 gastric cancer and, 896-897 in gastrinoma, 864 Helicobacter pylori infection in, 838, 872, 889-890 in mastocytosis, 864 mucosal defenses against, 863

Duodenal ulcers (Continued) NSAID-associated. See also Peptic ulcer disease, NSAID-associated. prevention of, 873-876 treatment of, 872-877 obstructing, 884-885 perforated, 883-884, 884f in pregnancy, 629 prevalence of, 706, 706f refractory, 877 stress-related, 877-878 surgical management of, 882 Duodenoduodenostomy, for duodenal atresia/ stenosis, 786 Duodenogastric reflux, in gastroesophageal reflux disease, 713 Duodenojejunal flexure, 779 Duodenotomy in gastrinoma, 519 in Zollinger-Ellison syndrome, 505, 505f Duodenum. See also Small intestine. adenocarcinoma of, in familial adenomatous polyposis, 2179-2180 adenoma of, in familial adenomatous polyposis, 2179-2180 anatomy of, 1615 general, 779 microscopic, 779-780 annular pancreas and, 780t, 786-787 arterial supply to, 779 aspiration of in malabsorption, 1748 in small intestinal bacterial overgrowth, 1775 atresia and stenosis of, 780t, 785-787, 1632, 1632f congenital defects associated with, 785 presentation and diagnosis in, 785-786, 786f treatment of, 786 bicarbonate secretion in, in cystic fibrosis, 943 carcinoid tumors of, 477-480, 479t-480t congenital anomalies of, 780t, 785-787 development of, 779 diverticula of, 376-377 extraluminal, 376, 376f intraluminal, 376, 377f duplication cyst of, 780t, 787 embryologic development of, 1624 gastrinoma of, 499-501 hypersensitivity of to fatty acids, 799-800 in functional dyspepsia, 188 innervation of, 779 iron absorption in, 1240, 1241f lymphatic drainage of, 779 obstruction of in chronic pancreatitis, 1014 in malrotation, 2116 perforation of, in endoscopic retrograde cholangiopancreatography, 1197 scalloping of in celiac disease, 1809, 1810f in malabsorption, 1746, 1746f in tropical sprue, 1825, 1826f somatostatinoma of, 511-513 ulcers of. See Duodenal ulcers. varices of, 1513 venous drainage of, 779 villi of, 779-780 Duplication bile duct, 1050 colonic, 1632 duodenal, 780t, 787 esophageal, 670t, 671-672, 672f gastric, 780t, 782-783, 782f ileal, 1632 intestinal, 1631-1632 rectal, 1632 small intestinal, 1631-1632 Dwarf tapeworm infection, 1932-1933 Dwarfism, nutritional, in children, 62-63, 62t Dynamometry, fist-grip, in nutritional assessment, 66-67, 78f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxvii

xxviii

Index Dysautonomia, familial, gastrointestinal manifestations of, 580 Dysentery amebic. See Amebiasis, intestinal. bacillary. See Shigellosis. versus ulcerative colitis, 1876 Dysfibrinogemia, in cirrhosis, 1555 Dysgeusia, 354 Dyslipidemia. See also Hyperlipidemia. after bariatric surgery, 117 congenital, malabsorption in, 1759t-1762t, 1763 gallstone disease and, 1092 gastrointestinal manifestations of, 577-578 in liver disease, 1225 obesity and, 104 Dyspepsia, 183-196 in biliary tract disease, 184 definition of, 183 in esophageal cancer, 184 in food intolerance, 184 functional, 185-189 complementary and alternative medicine for, 2289-2291, 2290t definition of, 185 duodenal hypersensitivity in, 188 epidemiology of, 187 gastric accommodation and, 187-188 gastric distention hypersensitivity in, 188 gastric emptying and, 187 gastric motility disorders in, 809, 809t genetic factors in, 188 heartburn and, 187 Helicobacter pylori infection in, 188 eradication of, 191, 192t irritable bowel syndrome and, 187 pathogenic factors in, 188-189 pathophysiology of, 187-188 postinfection, 188 psychosocial factors in, 188-189 subgroups of, 185-187, 186f, 186t symptom complex of, 185-187, 186f, 186t treatment of, 191-193 general measures in, 191 pharmacologic, 191-193, 192t psychological, 193 recommendations for, 193-194, 194f-195f in gastric cancer, 184 in gastroesophageal reflux disease, 184 in irritable bowel syndrome, 2092 in medication intolerance, 184, 185t organic causes of, 183-185, 185t in pancreatic disorders, 184 in peptic ulcer disease, 184, 866-867, 866f-867f symptoms associated with, 183, 184t uninvestigated approach to, 189-191 classification of, 186f history in, 189 laboratory testing in, 189 management of, 189-191 empirical antisecretory therapy in, 190 Helicobacter pylori infection test and treat strategy in, 190 prompt endoscopy and direct treatment in, 189-190 recommendations for, 190-191 physical examination in, 189 Dysphagia, 173-176. See also Esophageal motility, disorders of. in achalasia, 691-692 in amyotrophic lateral sclerosis, 688 differential diagnosis of, 174-176, 174t-175t, 175f, 692-693 in distal esophageal spasm, 690-692 esophageal, 174-176, 175f, 175t versus oropharyngeal dysphagia, 691 evaluation of, 693-700 in gastroesophageal reflux disease, 714, 726 after hematopoietic stem cell transplantation, 550, 550f in HIV/AIDS, 524-526, 524f-525f, 525t malignant, 752-753 palliation of, 764-767

Dysphagia (Continued) in myasthenia gravis, 688-689 in myotonia dystrophy, 688 in oculopharyngeal dystrophy, 688 oropharyngeal, 174, 174t aspiration risk evaluation in, 700 in cricopharyngeal bar, 689, 689f cricopharyngeal myotomy for, 700 epidemiology of, 686-687 pathogenesis of, 687-691, 687t swallow therapy for, 700 treatment of, 700 in tumors, 688 in Zenker’s diverticula, 686, 689, 689f in Parkinson’s disease, 688 pathophysiology of, 173-176 in peptic esophageal strictures, 726 pill, odynophagia in, 176 in poliomyelitis, 688 stroke-related, 687-688 percutaneous endoscopic gastrostomy for, 90 Dysphagia lusoria, vascular anomalies of, 672, 673f Dysplasia-associated lesion or mass (DALM), in ulcerative colitis, 2006, 2007f Dysproteinemia, gastrointestinal manifestations of, 568-569 Dystonia, in Wilson disease, 1252

E

Eating disorders, 121-138. See also specific disorders, e.g., Anorexia nervosa. clinical features or complications of, 127, 127f-128f, 129t-130t constipation in, 271 diagnosis of, 122f, 122t, 123-128 differential diagnosis of, 125-126, 126t epidemiology of, 121 etiology of, 121-123 gastrointestinal manifestations of, 128-131, 129t-130t medical management of, 135-137, 136t laboratory evaluation of, 128 medical evaluation in, 127, 127f-128f not otherwise specified clinical features or complications of, 127, 129t-130t diagnosis of, 122f, 122t, 124-125 differential diagnosis of, 125-126, 126t nutritional assessment in, 125-126, 126t onset and course of, 123 pharmacotherapy for, 133-134 psychiatric treatment of, 131-134, 132f psychotherapy for, 132-133 treatment of, 131-137, 131f weight management for, 127, 134-135 Ebrotidine, hepatotoxicity of, 1433 Echinococcus, hepatic, 1356t-1357t, 1363-1365, 1364f Echinostoma spp. infection, 1934 Echocardiography in carcinoid heart disease, 483 in intrapulmonary shunt, 1550 “Ecstasy,” hepatotoxicity of, 1454 Ectasia. See Angioectasia; Gastric antral vascular ectasia (watermelon stomach); Lymphangiectasias; Telangiectasia. Ectopic pregnancy, ruptured, abdominal pain in, 154t Edema. See also Ascites. cerebral, in acute liver failure, 1561-1562, 1561t, 1565-1566 in Chagas’ disease, 1918 in protein-losing gastroenteropathy, 438440 Ehlers-Danlos syndrome, gastrointestinal manifestations of, 558t, 563 Ehrlichiosis, hepatic manifestations of, 1354 Eicosanoids, in intestinal ion transport, 1690

Elastase, 1712-1713, 1713f, 1713t fecal in chronic pancreatitis, 1000 in diarrhea, 226 functions of, 924 Elastase 1, fecal, 928t, 929 Elastography magnetic resonance, in hepatic fibrosis, 1235 transient in hepatic fibrosis, 1235 in hepatitis C, 1323, 1324t Elderly acute abdominal pain in, 161 angioectasia in, 595-596 appendicitis in, 2062 autoimmune hepatitis in, 1471 chronic pancreatitis in, 994 Clostridium difficile infection in, 1894-1895 constipation in, 260-262 Crohn’s disease in, 1972 diverticulitis in, 2084 fecal incontinence in, 244, 257 gastropathy in, 858 infectious diarrhea in, 1875-1876 liver transplantation in, 1597 malabsorption in, 1758 micronutrient requirements for, 58 nutrition in, 58 small intestinal bacterial overgrowth in, 1773 Electrical stimulation for fecal incontinence, 255 gastric for gastric motility disorders, 812, 813t for gastroparesis, 574 for vomiting, 209 sacral nerve colonic motility and, 1672 for constipation, 283 for fecal incontinence, 256, 257f Electrocardiography, in eating disorders, 128 Electrogastrography in functional vomiting, 202 in gastric motility assessment, 790f-791f, 799f, 804-805, 805f, 811-812, 811t treatment approaches based on, 811-812, 811t in vomiting, 206 Electrolytes absorption of, in short bowel syndrome, 1781-1782, 1781t, 1782f balance of, with ileostomy, 2016 in cholera diarrhea, 1851, 1851t concentrations of, in gastrointestinal fluids, 71t disturbances of in eating disorders, 128 in VIPoma, 509-510 with vomiting, 206 therapeutic. See Fluid therapy. transport of. See Fluid and electrolyte transport. Electromyography in constipation, 273 in fecal incontinence, 249 Electron microscopy, in Whipple’s disease, 1838, 1839f Electrosurgical burns in endoscopy, 656 in polypectomy, 660 Embolism. See also Thromboembolism. pulmonary, in ulcerative colitis, 2012 superior mesenteric artery, 2030, 2034-2035, 2034f Embolization for colonic angioectasia, 599 for diverticular bleeding, 2088, 2088f for gastrointestinal bleeding, 292 hepatic artery for carcinoid tumors, 488 for pancreatic endocrine tumors, 521 for hepatocellular carcinoma, 1579 for peptic ulcer bleeding, 302 Emesis. See Nausea; Vomiting. Emetic reflex, 197-198, 198f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Emphysematous cholecystitis, 1111-1112, 1111f, 1118, 1119t Emphysematous gastritis, 850 Emtricitabine, for hepatitis B, 1304 Emulsification, triglyceride, 1699, 1700f Enalapril, hepatotoxicity of, 1433 Encephalitozoon intestinalis infection, 19171918. See also Microsporidiosis. malabsorption in, 1829, 1829t Encephalomyopathy, mitochondrial neurogastrointestinal extraintestinal manifestations of, 2140 gastrointestinal manifestations of, 581 intestinal pseudo-obstruction in, 2133-2134, 2134f natural history of, 2141 Encephalopathy in acute pancreatitis, 982 hepatic. See Hepatic encephalopathy. portosystemic, test for, 1544-1545 Encopresis, fecal incontinence associated with, 266 Endocannabinoids, food intake effects of, 102-103 Endocarditis, in Whipple’s disease, 1837 Endocrine disorders. See also specific disorders, e.g., Diabetes mellitus. in cirrhosis, 1554 diarrhea from, 218 in eating disorders, 127, 129t-130t gastrointestinal manifestations of, 572-577, 573t in obesity, 105 Endocrine system, in protein-energy malnutrition, 63 Endocrine transmitter, 3-4, 4f Endocrine tumors classification of, 493 gastrointestinal. See Carcinoid tumors. molecular pathogenesis of, 494 pancreatic. See Pancreatic endocrine tumors. Endocytosis, in hepatocytes, 1209-1210 Endoglin, in esophageal cancer, 749-750 Endometrial cancer, obesity and, 105 Endometriosis, intestinal, 2253-2255, 2254f-2255f Endomicroscopy, confocal, in esophageal cancer, 755, 755f Endopeptidases, 1712-1713, 1713f, 1713t Endoplasmic reticulum granular, in acinar cells, 913, 913f of hepatocytes, 1209 rough, pancreatic enzyme synthesis in, 922f, 924 stress response of, in alcoholic liver disease, 1389 Endoprosthesis. See Stent(s). Endoscope, disinfection of, 655-656 Endoscopic band ligation. See Band ligation, endoscopic. Endoscopic gastrojejunostomy, percutaneous, 92, 92f Endoscopic gastrostomy, percutaneous. See Percutaneous endoscopic gastrostomy. Endoscopic hemostasis. See Hemostasis, endoscopic. Endoscopic mucosal resection for Barrett’s esophagus, 732 for esophageal cancer, 761, 761f-762f for gastric cancer, 902-903 Endoscopic retrograde cholangiopancreatography, 1191-1197 in acute pancreatitis, 973 in bile leaks, 1191-1192, 1193f in biliary strictures benign, 1193, 1195f indeterminate, 1193 malignant, 1193-1196, 1196f bleeding in, 660 in cholangiocarcinoma, 1174, 1175f cholangitis after, 661 in choledocholithiasis, 1117, 1191, 1192f in chronic pancreatitis, 1001-1003, 1002f, 1002t

Endoscopic retrograde cholangiopancreatography (Continued) in combination with percutaneous approach, 1197, 1197f complications of, 660-661, 1197 in Fasciola hepatica infection, 1935, 1935f in gallstone disease, 1108t, 1111, 1111f with guidewire cannulation, 968 in hepatobiliary ascariasis, 1922, 1923f interpretation of, variability in, 1002-1003 in intrahepatic cholangiocarcinoma, 1581 in jaundice, 332, 333t in malabsorption, 1749 in neonatal cholestasis, 1052 in pancreatic cancer, 1021, 1021f in pancreatic intraductal papillary mucinous tumor, 1031, 1032f pancreatitis after, 660, 660t, 967-968, 967t perforation in, 661 in primary sclerosing cholangitis, 1057, 1154, 1160, 1160f, 1192-1193, 1194f in recurrent pyogenic cholangitis, 1168f, 1169 in sphincter of Oddi dysfunction, 1069-1070, 1196-1197 with sphincterotomy, for recurrent pyogenic cholangitis, 1169 Endoscopic sphincterotomy. See Sphincterotomy, endoscopic. Endoscopic submucosal resection for esophageal cancer, 761-762 for gastric cancer, 903, 903f Endoscopic surgery, natural orifice transluminal, in appendicitis, 2068-2069 Endoscopic therapy for acute pancreatitis, 978 early, 1035 for local complications, 1035-1038, 1036f-1038f for recurrent disease, 1038-1039 for Barrett’s esophagus, 731-732 for bezoar removal, 405-406 for biliary tract disease, 1191-1197, 1192f-1196f for chronic pancreatitis, 1006-1007, 1039, 1039t, 1040f for colorectal cancer, 2237-2238, 2238f dilation with, complications of, 657 for esophageal cancer, 761-762, 761f-762f, 764-767 for foreign bodies, 400-401, 401t for gastric motility disorders, 813t, 814 for gastroesophageal reflux disease, 725 for hereditary hemorrhagic telangiectasia, 601 for pancreatic cancer, 1041 for pancreatic cysts, 1041 for pancreatic duct leaks, 1040-1041 for pancreatic pseudocyst, 981, 1011 for peptic ulcer bleeding, 879-881 combination methods in, 880-881 comparison of techniques in, 880 proton pump inhibitor use before, 881 repeat, 883 for portal hypertension, 1503-1504, 1503f, 1503t for primary sclerosing cholangitis, 1165-1166 for radiation enteritis, 646 for Zenker’s diverticula, 372-373, 373f, 700 Endoscopic ultrasonography. See Ultrasonography, endoscopic. Endoscopic variceal pressure, 1496, 1496t Endoscopy. See also Colonoscopy; Enteroscopy; Laparoscopy; Sigmoidoscopy. in ampullary carcinoma, 1182, 1182f of anorectum, 2259-2260 antibiotic prophylaxis for, 654, 654t anticoagulant/antiplatelet drug management for, 654 in Barrett’s esophagus, 727, 728f, 730 capsule in angioectasia, 597 in celiac disease, 1810 complications of, 658 in Crohn’s disease, 1958 in diarrhea, 223-224

Endoscopy (Continued) in gastroesophageal reflux disease, 716 in NSAID enteropathy, 2052, 2053f in obscure gastrointestinal bleeding, 291, 319-320 in small intestinal tumors, 2149-2150, 2149f in tropical sprue, 1827 in ulcerative enteritis, 2055, 2056f in carcinoid tumors, 485 in caustic injury, 407, 407f, 407t complication(s) of, 653-662 abdominal distention as, 656 cardiac, 654-655 electrosurgical hazards as, 656 infectious, 655-656 medicolegal considerations for, 656-657 respiratory, 654-655 sedation-related, 654-655, 655t severity of, 656 timing of, 656 in constipation, 272 in Crohn’s disease, 1958, 1958f in diarrhea, 223-224 in diverticulitis, 2080 in dysphagia, 176 in enteric infection, 1849 in eosinophilic esophagitis, 430-431, 431f in eosinophilic gastroenteritis, 431 in eosinophilic proctitis, 431 in esophageal cancer chromo-, 754-755, 754f diagnostic, 754-755, 754f-755f high-resolution, 755, 755f screening, 756 staging, 756 surveillance, 756 in esophageal motility disorders, 693 in food protein–induced enterocolitis, 431 in foreign bodies, 400 in gastric cancer, 899-900, 899f in gastroesophageal reflux disease, 716, 716t, 717f in gastrointestinal bleeding, 289-291 in Helicobacter pylori infection, 839, 840f, 840t history in, 653-654 informed consent for, 654 in malabsorption, 1746, 1746f monitoring during, 655 in obscure gastrointestinal bleeding, 319-320 capsule, 291, 319-320 deep, 320, 320t intraoperative, 319 patient preparation for, 653-654 in peptic ulcer bleeding, 878-879, 879f, 879t repeat, to confirm healing, 303 second-look, 301-302 in peptic ulcer disease, 864-865, 864t, 865f physical examination in, 653-654 in pill esophagitis, 736, 736f in pneumatosis coli and, 2249, 2249f in portal hypertension, 1496-1497, 1497f in pregnancy, 626-627 sedation for, complications of, 654-655, 655t small bowel. See also Enteroscopy. complications of, 658 therapeutic. See Endoscopic therapy. in ulcerative colitis, 1985-1987, 1986f upper, complication(s) of, 657-658 bleeding as, 657 perforation as, 657 respiratory, 657 topical anesthesia–related, 657 in upper gastrointestinal bleeding, 293-294 in vomiting, 204-205 Endosonography. See Ultrasonography, endoscopic. Endothelial cells, sinusoidal, 1203-1204, 1210 Endotheliitis, in liver transplant recipients, 1607, 1607f Endothelin-1 in hepatic vasoconstriction, 1490-1491 in portal hypertension, 1493 in portopulmonary hypertension, 1550

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxix

xxx

Index Endotoxin, in nonalcoholic fatty liver disease, 1404 Endotracheal intubation in acute liver failure, 1565 in Zenker’s diverticula, 371-372 Endovascular aneurysm repair, for abdominal aortic aneurysm, 607 Enema aminosalicylate, in ulcerative colitis, 1995 barium. See Barium enema. bismuth carbomer, in pouchitis, 2010 budesonide, in ulcerative colitis, 1996 butyrate, in ulcerative colitis, 2000 “cleansing,” chemical colitis from, 2247-2248, 2247t, 2248f in constipation, 277t, 279-280 epidermal growth factor, in ulcerative colitis, 2001 short-chain fatty acid, in diversion colitis, 2243 Energy expenditure disease severity and, 78, 79t, 85, 85t of physical activity, 48, 49t resting, 48, 48t-49t systemic inflammatory response and, 60 total daily, 47-48 Energy medicine, 2287 Energy metabolism, 47-49, 48t-50t in liver, 1216-1225 Energy requirements equations to predict, 78 evaluation of, 79, 79f for hospitalized patients, 48-49, 49t-50t resting, 48, 48t-49t Energy stores, 47, 48t in eating disorders, 123 Enkephalins, in intestinal ion transport, 1690 Entamoeba coli, 1907f Entamoeba dispar, 1909 Entamoeba gingivalis, 1907f Entamoeba hartmanni, 1907f Entamoeba histolytica, 1905-1911, 1907f. See also Amebiasis. life cycle of, 1367f, 1368, 1906, 1906f trophozoite of, 1906, 1907f-1908f, 1909 Entamoeba polecki, 1907f Entecavir for hepatitis B, 1303 in pregnancy, 1306 resistance to, 1303 HBV DNA polymerase mutation and, 1292, 1292f Enteral access procedures, complications of, 658 Enteral nutrition access methods for, 87-93, 88t advancement of, 94 for alcoholic liver disease, 1395, 1395f blenderized formulas for, 94 bolus, 93 complications of, 94-95, 658 continuous, 93 diarrhea during, 230 direct percutaneous jejunostomy for, 92-93, 92f fluoroscopic percutaneous access for, 93 formulas for, 94, 94t in gastrointestinal fistula, 421-422 immune-enhancing formulas for, 94 initiation of, 93 intermittent, 93 intolerance of, 93 for malabsorption, 72-74 modular formulas for, 94 nasoenteric tube for, 87-89, 88t-89t oral rehydration therapy for. See Oral rehydration solution. in patients undergoing major surgery, 70 percutaneous endoscopic gastrojejunostomy for, 92, 92f percutaneous endoscopic gastrostomy for, 89-91, 90f-91f predigested formulas for, 94 for severe acute pancreatitis, 1035 specialty formulas for, 94, 94t supplemental regimens during, 94

Enteral nutrition (Continued) surgical access for, 93 water requirements during, 93 Enteric microbiota. See Bacteria, intestinal (commensal). Enteric nervous system, 819-821, 820f, 1623, 1623f afferent neurons of, 1646, 1662-1663 in colonic motility, 263-264, 1662-1664, 1663f congenital anomalies of, 1636-1640 development of, 1626, 2124 efferent (motor) neurons of, 1646, 1663 excitatory, 1663 inhibitory, 1663 in fluid and electrolyte transport, 1686f, 1687-1688 integrative control by, 1648-1649 interneurons of, 1646-1647, 1663-1664 interstitial cells of Cajal and, relationship between, 794 in intestinal motility, 2121, 2122f-2123f, 2123-2124 intestinofugal neurons of, 1646-1647 organization of, 5-6, 5f plasticity of, 1649 in small intestinal motility, 1646-1647, 1646f transmitters in, 10-13, 12f, 1646, 1662 Enteritis. See also Gastroenteritis. versus appendicitis, 2063t infectious. See Diarrhea, infectious. lupus, 2136 radiation. See Radiation enteritis. in trichinosis, 1930 ulcerative. See Ulcerative enteritis. Enteritis necroticans, 1882 Enterobius vermicularis infection, 1928-1929, 1928f Enterochromaffin (EC) cell(s), 778, 817-819, 1618-1619, 1620f Enterochromaffin (EC) cell carcinoid tumors, 476-477, 477f appendiceal, 480-481 duodenal, 480 gastric, 478 Enterochromaffin-like (ECL) cells, 817-819 in gastrinoma, 499 Enteroclysis barium in obscure gastrointestinal bleeding, 318 in small intestinal obstruction, 2109-2110, 2112f in small intestinal tumors, 2147f, 2149 computed tomography in gastrointestinal bleeding, 318 in small intestinal obstruction, 2109, 2111f in small intestinal tumors, 2150, 2150f magnetic resonance, in malabsorption, 1749 small bowel, in malabsorption, 1749 Enterocolitis. See also Colitis. bacterial, complications of, 1859-1860, 1859t food protein–induced, 145, 146t, 429 endoscopy in, 431 treatment of, 433-435, 433t neutropenic, 2252-2253, 2253f nongranulomatous chronic idiopathic, malabsorption in, 1764-1765, 1765f pseudomembranous, 1890-1891 Clostridium difficile-associated, 1890-1891, 1897, 1897f, 1899-1900 in leukemia, 567 severe or fulminant, treatment of, 1899-1900 in Yersinia enterocolitica infection, 1869 Enterocytozoon bieuneusi infection, 1917-1918. See also Microsporidiosis. malabsorption in, 1829, 1829t Enteroendocrine cells. See Neuroendocrine (enteroendocrine) cells. Enterogastrone in gastric acid secretion, 823 secretin as, 8 Enterography computed tomography in Crohn’s disease, 1957, 1957f in gastrointestinal bleeding, 318

Enterography (Continued) in ulcerative enteritis, 2055 in vomiting, 204-205 magnetic resonance, in Crohn’s disease, 1957, 1957f Enterohepatic circulation, 1079-1080, 1079f disorders of, 1084-1087, 1085t transport proteins in, 1079f, 1081, 1082f, 1083t Enteroinsular axis, 18-19, 19t Enterokinase, 1697 congenital absence of, 956 deficiency of, malabsorption in, 1759t-1762t, 1763 Enterolithotomy, for gallstone ileus, 2115-2116 Enteropathy. See also Gastroenteropathy. AIDS, 529 apoptotic, 217-218 autoimmune, malabsorption in, 1765 in dermatitis herpetiformis, 1812 gluten-sensitive. See Celiac disease. NSAID, 2052-2054, 2053f portal, 602, 603f protein-induced, 145, 146t. See also Proteinlosing gastroenteropathy. tropical, 1828 tufting, 1640-1641 congenital, malabsorption in, 1759t-1762t Enteropathy-associated T cell lymphoma, 2054-2058 background on, 2054 in celiac disease, 457-458, 458f, 1798, 1818 clinical presentation in, 2055 definition of, 2054 diagnosis of, 2055-2056, 2057f pathology of, 2056, 2058f treatment of, 2056-2058, 2057f Enteroplasty, serial transverse, for short bowel syndrome, 1791-1793, 1793f Enteroscopy double balloon complications of, 658 in obscure gastrointestinal bleeding, 320, 320t in small intestinal tumors, 2150 in ulcerative enteritis, 2055-2056 push, complications of, 658 Enterotoxins Bacillus cereus, 1882 bacterial, 1845-1846, 1845f clostridial, 1881 Clostridium difficile, 1893-1894, 1893f Escherichia coli, 1845-1846, 1855 in intestinal ion transport, 1688t, 1691-1692 Enterovirus, gastritis in, 849 Environmental factors in cancer development, 41-42 in celiac disease, 1800-1802, 1801f in Crohn’s disease, 1946 in Helicobacter pylori infection, 834 in obesity, 101-102 in pancreatic cancer, 1017-1018 in ulcerative colitis, 1978-1979 Enzyme-coupled receptors, 14-15 Enzyme-linked immunosorbent assay (ELISA) in amebiasis, 1909-1910 in Clostridium difficile-associated diarrhea and colitis, 1896-1897, 1896t in cryptosporidiosis, 1915 in giardiasis, 1913 in schistosomiasis, 1937-1939 in strongyloidiasis, 1924 Eosinophil(s) in gastrointestinal tract, 425-427, 426f immune function of, 425-427 Eosinophilia in celiac disease, 433 in connective tissue disorders, 432 drug-induced, 432 in helminthic infection, 431-432 in hypereosinophilic syndrome, 432 in inflammatory bowel disease, 432-433 in intestinal polyps, 432

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Eosinophilia (Continued) after organ transplantation, 433 in vasculitis, 432 Eosinophilic esophagitis, 427-428 allergic, 143-144, 146t clinical presentation in, 427 differential diagnosis of, 431-433 epidemiology of, 427 evaluation of, 430-431, 431f food impaction in, 401 versus GERD, 428, 431 natural history of, 428 pathogenesis of, 425-427, 426f treatment of, 433-435, 433t Eosinophilic gastritis, 854-855, 854f Eosinophilic gastroenteritis, 428-429, 590 allergic, 144, 146t in Ancylostoma caninum infection, 1927 areas affected by, 428 diagnosis of, 428 differential diagnosis of, 431-433 evaluation of, 430-431 gastric outlet obstruction in, 428-429, 428f mucosal, 428 muscular, 428-429, 428f natural history of, 429 serosal, 429 string sign in, 428-430, 428f treatment of, 433-435, 433t Eosinophilic gastrointestinal disorders, 425-436 classification of, 425, 426t clinical entities in, 427-430 differential diagnosis of, 431-433 evaluation of, 430-431, 431f pathogenesis of, 425-427, 426f treatment of, 433-435, 433t Eosinophilic proctitis, 429-430, 429f endoscopy in, 431 treatment of, 433-435, 433t Eosinophilic proctocolitis, allergic, 144-145, 146t Eotaxin, in eosinophilic esophagitis, 427 EOX regimen, for gastric cancer, 904 Epidermal growth factor in esophageal cancer, 749-750 in gastrointestinal tract, 16 receptors for, 16 topical, for ulcerative colitis, 2001 Epidermal transglutaminase, in dermatitis herpetiformis, 1812 Epidermoid cyst, in familial adenomatous polyposis, 363, 2181-2182 Epidermolysis bullosa, 358 acquired, 358 esophagus in, 358 skin fragility in, 358, 358f Epidural analgesia, ileus and, 2126 Epigastric hernia, 390 Epigastric pain syndrome, 185-187, 186f, 186t Epigenetic modifications, in alcoholic liver disease, 1389 Epigenetic silencing, in colorectal cancer, 2203-2204 Epilepsy abdominal, 578-579 in celiac disease, 1806 Epinephrine, in intestinal ion transport, 1690 Epinephrine injection therapy for colonic diverticular bleeding, 312 for diverticular bleeding, 2085-2087 for gastrointestinal bleeding, 292 for peptic ulcer bleeding, 297-298, 298f-300f, 880 Epiploic appendagitis, 621 Episcleritis in Crohn’s disease, 1954-1955 in ulcerative colitis, 2011 Episiotomy, fecal incontinence after, 244 Epithelial cells. See also Epithelium. of bile duct. See Cholangiocytes. in Helicobacter pylori infection, 835-836 intestinal, 25-26 antigen processing and presenting by, 26-27, 27f

Epithelial cells (Continued) antigen trafficking across, 25-26 in food allergy, 140-141 as physical barrier, 25 recognition of pathogen-associated molecular patterns by, 26 types of, 1618 in ulcerative colitis, 1980 unstirred water layer on, in digestion and absorption, 1701-1702 of liver. See Hepatocytes. Epithelial-mesenchymal transition, in metastasis, 42, 42f Epithelial Na+ channel (ENaC), 1681, 1681f Epithelioid gastrointestinal stromal tumors, 463, 463f, 468 Epithelioid hemangioendothelioma, hepatic, 1583 Epithelium. See also Epithelial cells. esophageal, 667, 668f defenses of, 711-712, 712f dilated intercellular spaces of, in gastroesophageal reflux disease, 712, 712f intestinal, 1618 architecture of, 1675-1677, 1676f crypt-villus axis of, 1677-1678, 1678f development of, 1618 homocellular regulation in, 1693-1694 movement across, 1676f, 1678-1679 permeability and polarity of, 1676f, 1677 tight versus leaky, 1676f, 1678-1679 transport in. See Fluid and electrolyte transport. permeability and polarity of, 1677 Epping jaundice, 1454 Epstein-Barr virus infection gastritis in, 849 hepatitis in, 1347 ERCP. See Endoscopic retrograde cholangiopancreatography. Ergocalciferol. See Vitamin D. Ergosterol, 1722 Erlotinib, for esophageal cancer, 764 Eructation, repetitive, intestinal gas causing, 237 Erythema, necrolytic migratory, 365, 365f, 369, 506-508, 507f Erythema multiforme, oral, 358 Erythema multiforme major. See Stevens-Johnson syndrome. Erythema nodosum in Crohn’s disease, 1954 in inflammatory bowel disease, 359 in ulcerative colitis, 2010 Erythromycin for Campylobacter spp. infection, 1865t, 1868 for gastric motility disorders, 812, 813t for gastroparesis, 574 for irritable bowel syndrome, 2103 in upper gastrointestinal bleeding, 290 for vomiting, 208-209 for Whipple’s disease, 1841t Erythrophagocytosis, of Kupffer cells, in sickle cell anemia, 570-571, 570f Erythropoietic porphyria, 363 Erythropoietic protoporphyria, 1267t, 1268 Erythropoietic regulator, in iron absorption, 1725 Erythropoietin, tumors secreting, 515 Escherichia coli, 1854-1857, 1854t in abdominal abscess, 412, 418 enteroadherent, 1844 enteroaggregative, 1854t, 1857 enterohemorrhagic, 1854t, 1855-1857 clinical features of, 1856, 1856f diagnosis of, 1856 in elderly, 1876 epidemiology of, 1855-1856 treatment of, 1856-1857, 1865t virulence factors of, 1856 enteroinvasive, 1854t, 1855, 1865t enteropathogenic, 1854, 1854t, 1865t adherence factors of, 1844-1845, 1845f, 1854 intestinal ion transport in, 1692

Escherichia coli (Continued) enterotoxigenic, 1854-1855, 1854t, 1865t adherence factors of, 1845 toxins of, 1845-1846, 1855 in food poisoning, 1879t-1880t O157:H7 serotype of, 1855-1856 pedestal formation by, 1844, 1845f pili (fimbriae) of, 1845 shiga toxin–producing, 1856 in traveler’s diarrhea, 1873 Esomeprazole adverse effects of, 871 drug interactions with, 871 for gastroesophageal reflux disease, 723, 725 mechanisms of action of, 870 for NSAID ulcer prophylaxis, 874 for NSAID ulcers, 873 for peptic ulcer disease, 870-871 pharmacokinetics of, 870-871 Esophageal biopsy in esophageal cancer, 753-754, 754f in gastroesophageal reflux disease, 716, 717f Esophageal cancer, 745-767 adenocarcinomatous alcohol consumption in, 748 in Barrett’s esophagus, 727, 748-749 cholecystectomy and, 749 diet and nutrition in, 748 dysplasia and, 752 etiology of, 748-749 gastroesophageal reflux disease in, 748 genetic factors in, 749 Helicobacter pylori infection and, 728, 749, 838 medication use and, 749 obesity and, 728, 748 pathology of, 746f, 752 preexisting diseases associated with, 749 tobacco use in, 748 adenoid cystic, 767 after alkaline caustic injury, 408 barium esophagography in, 753, 753f basaloid, 767 carcinosarcomatous, 767 chest radiography in, 753 clinical features of, 752-753 computed tomography in, 753, 756-758, 758f diagnosis of, 753-755 dyspepsia in, 184 dysphagia in, 175-176, 752-753 dysplasia in, endoscopic detection of, 754-755, 754f-755f endoscopic ultrasonography in, 757-760, 760f endoscopy in chromo-, 754-755, 754f diagnostic, 754-755, 754f-755f high-resolution, 755, 755f screening, 756 staging, 756 surveillance, 756 epidemiology of, 745-746 FDG-PET in, 757-759, 759f genetic factors in, 749 imaging studies in, 753-754, 753f-754f laboratory testing in, 753 lymphoma as, 768 malignant melanoma as, 767-768 metastatic, 768 molecular biology of, 749-751 pathology of, 751-752 prevention of, 760 restaging of, 758-760 sarcoma as, 768 screening for, 756 small cell, 767 spectroscopy in, 755 squamous cell achalasia and, 703, 747-748 alcohol consumption in, 747 candidiasis and, 747-748 in caustic injury, 747-748 depth of invasion in, 751-752, 752f diet and nutrition in, 747

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxxi

xxxii

Index Esophageal cancer (Continued) dysplasia and, 746f, 751 epidemiology of, 745-746 esophageal preexisting diseases associated with, 747-748 etiology of, 747-748 genetic factors in, 749 in human papillomavirus infection, 748 pathology of, 746f, 751-752, 752f in Plummer-Vinson syndrome, 748 radiation therapy and, 748 strictures and, 747-748 tobacco use in, 747 tylosis and, 747-748 variants of, 767 staging of, 756-760 computed tomography, 756-757, 758f endoscopic, 756 endoscopic ultrasonographic, 757-758, 760f FDG-PET, 757, 759f minimally invasive, 758 after neoadjuvant therapy, 758-760 outcome and, 758 system for, 756, 757t surveillance for, 756 treatment of, 760-767 for early disease, 761-763 chemotherapy in, 762-763 endoscopic, 761-762, 761f-762f radiation therapy in, 762-763 surgical, 762 for locally advanced cancer, 763-764 chemoradiotherapy in, 763-764 radiation therapy in, 763 surgical, 763 for unresectable cancer, 764-767 analgesics in, 767 argon plasma coagulation in, 765 brachytherapy in, 767 chemotherapy in, 764 cytotoxic injection therapy in, 765 dilation in, 765 endoscopic palliative, 764-767 laser therapy in, 764-765, 765f multidisciplinary care in, 767 nutritional therapy in, 767 photodynamic therapy in, 765, 765f stents in, 765-767, 766f verrucous, 767 Esophageal dilation for eosinophilic esophagitis, 435 for esophageal cancer, 765 Esophageal dysphagia, 174-176, 175f, 175t Esophageal glands, secretions of, esophageal acid clearance and, 711 Esophageal manometry, 694-698, 699t conventional, 694, 695f in gastroesophageal reflux disease, 719 high-resolution. See Esophageal pressure topography, high resolution. of upper esophageal sphincter function, 678-679, 679f in vomiting, 205 Esophageal motility, 677-686. See also Esophageal sphincter(s); Swallowing. disorders of, 686-704 in achalasia, 686-687. See also Achalasia. barium swallow in, 693-694, 694f Chicago classification of, 698, 699t clinical presentation in, 691-692 contrast imaging in, 693-694, 694f definition of, 686 in diabetes mellitus, 572 diagnosis of, 693-700 differential diagnosis of, 692-693 dysphagia in, 174, 686-687, 691. See also Dysphagia, oropharyngeal. endoscopy in, 693 epidemiology of, 686-687 esophageal hypersensitivity reduction for, 703-704 intraluminal impedance measurement in, 695f, 698-699

Esophageal motility (Continued) manometry in, 694-698, 699t. See also Esophageal manometry. pathogenesis of, 687-691, 687t sensory testing in, 699-700 treatment of, 700-704 after variceal sclerotherapy, 738 in pregnancy, 625 in systemic lupus erythematosus, 560-561 Esophageal pressure topography, high-resolution, 682, 694, 696f of achalasia subtypes, 695, 697f, 701 classification of esophageal motility disorders based on, 698, 699t contractile front velocity in, 695-696, 696f, 698f of deglutitive esophagogastric junction relaxation, 694-695, 696f distal contractile integral in, 697, 699f Esophageal spasm diffuse, dysphagia in, 174 distal clinical presentation in, 692 differential diagnosis of, 693 dysphagia in, 690-692 pathogenesis of, 690-691 treatment of, 703 Esophageal sphincter(s) lower anatomy of, 665, 667f antireflux function of, 707, 708f pressures of in gastroesophageal reflux disease, 707, 708f, 708t modulators of, 684, 708t relaxation failure of. See Achalasia. swallow-induced relaxation of, 684 in gastroesophageal reflux disease, 709, 709f in swallowing, 682-685, 683f tone of, resting, 683-684 transient relaxation of, 684-685, 685f in gastroesophageal reflux disease, 708, 709f-710f upper anatomy of, 665, 667f disorders of, 686-704. See also Dysphagia, oropharyngeal. in Zenker’s diverticula, 371 function of, 679 in globus sensation, 176 manometric evaluation of, 678-679, 679f muscular components of, 678 opening of, 679-680, 680f in swallowing, 677-679, 678f-679f Esophageal stent in cancer, 765-767, 766f types of, 766 Esophageal strictures after caustic injury, 406f, 408 dysphagia in, 175-176 in epidermolysis bullosa, 358 after nasogastric intubation, 739, 739f radiation-induced, 641 squamous cell carcinoma and, 747-748 after variceal sclerotherapy, 738 Esophageal tumors, 745-770 classification of, 746t epithelial benign, 768, 768f malignant. See Esophageal cancer. fibrovascular, 769, 770f granular cell, 769 in HIV/AIDS, 525 nonepithelial benign, 768-770, 769f-770f malignant, 768 Esophageal ulcers in HIV/AIDS bacterial, 525 nonspecific, 524-525, 525f medication-related, 735-736, 736f, 738

Esophageal varices, 1494-1495, 1508-1510 bleeding from, 306-308 acute control of, 1509-1510, 1510f-1511f primary prophylaxis of, 1508-1509, 1509f secondary prophylaxis of, 1510, 1511f detection of, 1496-1498, 1497f-1498f endoscopic grading of, 1496-1497 natural history of, 1508 in pregnancy, 637 red color signs in, 1496-1497 Esophagitis bleeding in, 303 candidal, 354-355, 741-742, 742f in Chagas’ disease, 743 dyspepsia in, 184 in eating disorders, 130, 137 eosinophilic. See Eosinophilic esophagitis. erosive, 720 esophageal biopsy in, 716, 717f herpes simplex, 742 in HIV/AIDS, 524-526, 524f-525f, 525t human papillomavirus, 742-743 infectious, 741-743, 741t medication-induced, 735-739, 736f, 737t clinical presentation in, 736 diagnosis of, 736, 736f drugs associated with, 737-739, 737t management of, 736-737 mechanisms of, 735-736 Mycobacterium tuberculosis, 743 prevalence of, 705-706, 706f radiation-induced, 640-641, 640f reflux. See also Gastroesophageal reflux disease. bleeding from, 303 endoscopic grading systems for, 716, 716t, 717f endoscopy in, 716, 716t, 717f in syphilis, 743 Trypanosoma cruzi, 743 from vomiting, 206 Esophagogastric junction, 667, 668f, 773 adenocarcinoma of, pseudoachalasia in, 693 anatomy of, 682 antireflux function of, 707-708, 708f deglutitive relaxation of, 694-695, 696f high-pressure zone of, 682-685, 683f venous drainage zones of, 1494 Esophagogastroduodenoscopy in bariatric surgery, 115 in gastric cancer, 899-900, 899f Esophagography, barium. See Barium esophagography. Esophagorespiratory fistula, stent therapy for, 765-767, 766f Esophagus abdominal, 666f abnormal contractions of, chest pain in, 177-178 achalasia of. See Achalasia. acid clearance by, 710-713, 711f in gastroesophageal reflux disease, 710-713, 711f-712f adenoma of, 768 anatomy and histology of, 665-667, 666f-668f anatomy of, 680-682 antireflux barriers of, 707-708, 708f atresia of, 668-670, 669f, 669t-670t balloon distention studies of, 699 Barrett’s. See Barrett’s esophagus. biopsy of in esophageal cancer, 753-754, 754f in gastroesophageal reflux disease, 716, 717f black, 743 in candidal esophagitis, 741-742 body of, 665, 667f cancer of. See Esophageal cancer. carcinoid tumors of, 478 caustic injury to. See Caustic injury. cervical, 666f. See also Esophageal sphincter(s), upper. circulation of, 667 Crohn’s disease of, 1950

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Esophagus (Continued) deglutitive inhibition in, 681 devascularization of, for portal hypertension– related bleeding, 1507 developmental anomalies of, 668-675, 669t-670t dilatation of. See Esophageal dilation. distention of, chest pain in, 178 diverticula of, 373-374 clinical presentation and diagnosis of, 373-374, 373f epiphrenic, 373-374, 373f etiology and pathogenesis of, 373, 373f traction, 373 treatment and prognosis in, 374 duplication cyst of, 670t, 671-672, 672f embryology of, 668, 669f in epidermolysis bullosa, 358 epithelium of, 667, 668f defenses of, 711-712, 712f dilated intercellular spaces of, in gastroesophageal reflux disease, 712, 712f fistula of, 668-670, 669f, 669t-670t foreign bodies in, 397-401, 398f-399f gastric mucosa in, heterotopic, 670t, 675, 675f gastrointestinal stromal tumors of, 768-769, 769f hamartoma of, 769 hemangioma of, 769-770 hematoma of, spontaneous, 741 after hematopoietic stem cell transplantation, 555 hypersensitivity of, treatment of, 703-704 inflammation of. See Esophagitis. inflammatory fibroid polyp of, 768 injury to blunt, 739 medication-induced, 735-739, 736f, 737t penetrating, 739 traumatic, 739-740 tube-related, 739, 739f with vomiting, 206 innervation of, 665-666, 668f, 680-681 lamina propria of, 667, 668f in lichen planus, 359 lipoma of, 770 lymphatic system of, 667 motor function of, 677-686. See also Esophageal motility; Esophageal sphincter(s); Swallowing. mucosa of, 667, 668f musculature of, 665, 667f, 680 necrosis of, acute, 743 nutcracker, spastic, 697, 699f perforation of in Boerhaave’s syndrome, 740-741 in gastroesophageal reflux disease, 720 from trauma, 739-740 tube-related, 739 peristalsis of, 681-682, 683f. See also Esophageal motility; Esophageal sphincter(s); Swallowing. acid clearance during, 710-711, 711f pH of in extraesophageal GERD, 180 in gastroesophageal reflux disease, 717-719, 718f in rumination, 204 resection of for Barrett’s esophagus, 731 for esophageal cancer, 762-763 ringed, corrugated, 674 ring(s) of A (muscular), 670t, 672 B (Schatzki’s), 175, 670t, 672-674, 673f scleroderma of, 560 dysphagia in, 174 sensory function of, 685-686 sensory testing of, 699-700 after solid organ transplantation, 541 spasm of. See Esophageal spasm.

Esophagus (Continued) sphincters of. See Esophageal sphincter(s). spinal afferents to, 685-686 squamous papilloma of, 768, 768f staple transection of, for portal hypertension– related bleeding, 1506 stenosis of, congenital, 670t, 671, 671f strictures of. See Esophageal strictures. stromal tumors of, 465 submucosa of, 666f, 667 tear of in Boerhaave’s syndrome, 740-741 in Mallory-Weiss syndrome, 740 thoracic, 666f tubular duplication of, 671-672 tumors of, 745-770. See also Esophageal tumors. ulcers of. See Esophageal ulcers. vagal afferents to, 686 varices of. See Esophageal varices. vascular anomalies of, 670t, 672, 673f webs of, 670t, 674-675, 674f dysphagia in, 175 Estrogen gallstone disease and, 1092 hyperemesis gravidarum and, 628 jaundice from, 328-329 in primary biliary cirrhosis, 1485-1486 telangiectasia and, 316-317 Estrogen-progestin compounds, for gastric vascular ectasia, 1515-1516 Etanercept, for primary sclerosing cholangitis, 1165 Ethanol. See Alcohol. Ethanol injection, percutaneous, for hepatocellular carcinoma, 1579 Ethionamide, hepatotoxicity of, 1434 Ethnicity autoimmune hepatitis and, 1469-1470 constipation and, 262 Crohn’s disease and, 1943 ulcerative colitis and, 1976 Ethylene dibromide, hepatotoxicity of, 1452 Etoposide, for metastatic pancreatic endocrine tumors, 521 Etoricoxib, for NSAID ulcer prophylaxis, 874-875 Etretinate, hepatotoxicity of, 1433, 1438t Exenatide, for obesity, 113 Exercise abdominal distention and bloating and, 239 colorectal cancer and, 2194 constipation and, 262 energy expenditure of, 48, 49t excessive, in bulimia nervosa, 124 fecal incontinence and, 252 in obesity, 102, 109-110 Exocytosis, 924-925 in hepatocytes, 1209-1210 Exopeptidases, 1712-1713, 1713f, 1713t Extracellular matrix, of liver cell interactions with, 1211-1212 components of, 1212 Extracorporeal liver support, for acute liver failure, 1568 Extracorporeal shock wave lithotripsy for choledocholithiasis, 1124 for gallstone disease, 1122-1124, 1123t, 1124f for pancreatic duct stones, 1006-1007 Extrapyramidal disorders, gastrointestinal manifestations of, 579t, 580-581 Exudate/transudate determination, in ascites, 1523 Eye disorders in Crohn’s disease, 1954-1955 in ulcerative colitis, 2010t, 2011 in Whipple’s disease, 1837-1838 Ezetimibe, gallstone disease and, 1092

F

Fabry’s disease, gastrointestinal manifestations of, 577 Facies, dysmorphic, in Alagille’s syndrome, 1061

Factitious diarrhea, 230-231, 230t Factor VIIa, recombinant, for variceal bleeding, 307 Faget’s sign, in typhoid fever, 1866 Failure to thrive, in celiac disease, 1804 Famciclovir for herpes zoster ulcers, 357 for orolabial herpes, 357 Familial adenomatous polyposis, 2177-2184 ampullary carcinoma in, 1181 APC gene in, 2177-2178, 2178f, 2180-2182 attenuated, 2177t, 2184, 2185f cancer risks and screening recommendations for, 2181t clinical features of, 2178-2182 colonic, 2178-2179, 2179f extraintestinal, 2178f, 2180-2182 upper gastrointestinal, 2179-2180, 2180f colon carcinogenesis in, 2160 colorectal cancer in, 2208-2209, 2226-2227 congenital hypertrophy of retinal pigmented epithelium in, 2178f, 2180 cutaneous manifestations of, 363 dental abnormalities in, 2180-2182 desmoid tumors in, 2178f, 2180-2181, 2182t diagnosis and screening in, 2182-2183 duodenal adenocarcinoma in, 2179-2180 duodenal adenoma in, 2179-2180 epidermoid (inclusion) cysts in, 363, 2181-2182 extracolonic screening in, 2181t, 2184 Gardner’s variant, 2177t, 2180-2182. See also Gardner’s syndrome. gastric adenoma in, 2179 gastric cancer in, 893, 2179 gastric polyps in, 2179, 2180f genetic testing in, 2182, 2226 genetics of, 2177-2178, 2178f genotype-phenotype correlations in, 2178f, 2182 hepatoblastoma in, 1582 ileal adenoma in, 2180 jejunal adenoma in, 2180 mandibular osteomas in, 2180-2182 oral manifestations of, 363 pouchitis in, 2021 screening sigmoidoscopy in, 2182-2183 small intestinal adenoma in, 2148 treatment of, 2183-2184 medical, 2183-2184 surgical, 2183 Turcot’s variant, 2177t, 2184 variant, 2177t, 2184-2185 Familial atypical mole-malignant melanoma (FAMMM) syndrome, pancreatic cancer in, 1017 Familial Mediterranean fever gastrointestinal manifestations of, 558t, 563-564 peritonitis in, 618 Familial tooth agenesis, 2177t, 2184 Family factors in biopsychosocial model, 340 in functional abdominal pain syndrome, 169 Famotidine adverse effects of, 870 drug interactions with, 870 hepatotoxicity of, 1433 mechanisms of action of, 870 for NSAID ulcer prophylaxis, 873-874 for peptic ulcer disease, 870 pharmacokinetics of, 870 Fanconi-Bickel syndrome, 1711 FAP. See Familial adenomatous polyposis. Fas, as death receptor, 1214 Fasciculi longitudinales, 1049, 1067, 1068f Fasciola gigantica infection, 1935 Fasciola hepatica infection, 1356t-1357t, 1362-1363, 1935, 1935f Fasciolopsis buski infection, 1923f, 1933 Fast-track surgery for acute colonic pseudo-obstruction prophylaxis, 2129-2130, 2130t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxxiii

xxxiv

Index Fast-track surgery (Continued) for postoperative ileus prophylaxis, 2127-2128, 2128t Fasting in bulimia nervosa, 124 drug-induced liver disease and, 1417 Fat(s). See also Lipid(s). diarrhea from, 216 dietary, 51-52, 79 colorectal cancer and, 2193-2194 intake of, 1698-1706, 1699t in malabsorption, 73 restriction of, 95, 108 digestion and absorption of, 1699-1706 brush border membrane transfer in, 1702-1703, 1703f emulsification in, 1699, 1700f intestinal bacteria metabolic activity and, 1772 intracellular processing in, 1703-1705, 1704f-1705f intraluminal pH in, 1699t, 1701 lipase in, 1699-1701, 1700f liver X receptors in, 1705-1706, 1705t, 1706f micelles and lipid-containing particles in, 1700f, 1701 in neonates and infants, 1730-1732 orlistat and, 829 water layer in, 1701-1702 fecal, 928t, 929 analysis of, 222 qualitative, 1751 quantitative, 1750-1751, 1750f semiquantitative, 1751 in chronic pancreatitis, 1000 in sphincter of Oddi dysfunction, 1069 hepatic metabolism of, 1218f, 1220-1225, 1223f malabsorption of. See also Steatorrhea. in adrenal insufficiency, 1765 in autoimmune thyroid disease, 1766 breath tests for, 1751 in chronic pancreatitis, 997 colon in, 1741 congenital disorders of, 1759t-1762t diarrhea from, 218, 225-226 in elderly persons, 1758 fecal fat analysis for, 1750-1751, 1750f after gastric resection, 1756 in hyperthyroidism, 1765-1766 mechanisms of, 1737-1738, 1738t serum tests for, 1751 in short bowel syndrome, 1782-1785 in small intestinal bacterial overgrowth, 1774 tests for, 1750-1751, 1750f in tropical sprue, 1825-1827 stool assay for, 222 subcutaneous, necrosis of, in pancreatitis, 970 visceral, 103, 103t disorders related to, 104-105 Fat cells as endocrine cell, 103, 103f enlarged, disorders related to, 104-105 Fat mass, increased, disorders related to, 105-106 Fat necrosis in acute pancreatitis, 982 subcutaneous, in pancreatic disorders, 365 Fat wrapping, in Crohn’s disease, 1949 Fatigue, in primary biliary cirrhosis, 1479 Fatty acid(s) beta oxidation of mitochondrial, 1220 peroxisomal, 1220-1221 in colon carcinogenesis, 2194 dietary, 1698-1706, 1699t essential, 51-52 deficiency of, 51-52, 368-369 excess of hepatotoxicity of, 1403 in nonalcoholic fatty liver disease, 1403 free, in adipose tissue, release of, 1216-1217 gastric emptying of, duodenum and, 799-800 hydrogenation of, 1699

Fatty acid(s) (Continued) in intestinal phase of pancreatic enzyme secretion, 926 long-chain, 51 bacterial modification of, 1702 medium-chain, 1705 metabolism of in brain, food intake and, 103 hepatic, in nonalcoholic fatty liver disease, 1402-1403 monounsaturated, 1698, 1699t oxidation defects of, in acute fatty liver of pregnancy, 635 polyunsaturated, 1698, 1699t saturated, 1698, 1699t short-chain deficiency of, in diversion colitis, 2243 enemas with, for diversion colitis, 2243 intestinal transport of, 1685-1686 synthesis of, hepatic, 1220 trans-, 1699 Fatty acid ethanol esters, in alcoholic pancreatitis, 987 Fatty acid transport proteins, 1702 Fatty acid–binding proteins, 1703 Fatty liver disease alcoholic, 1383-1384, 1384f. See also Alcoholic liver disease. focal, 1410-1411, 1410f hepatitis C and, 1326 nonalcoholic, 1401-1412 versus alcoholic hepatitis, 1408-1409 versus alcoholic liver disease, 1391 antihyperlipidemic and cytoprotective agents for, 1410 antioxidants for, 1409-1410 after bariatric surgery, 118 cirrhosis in, 1401, 1406-1407 clinical features of, 1405, 1405t diagnosis of, 1407-1408, 1407f, 1408t epidemiology of, 1401 etiology of, 1401-1402, 1402t hepatic fibrosis in, 1404, 1408 histopathology of, 1405-1407, 1406f, 1407t imaging of, 1405, 1406f insulin-sensitizing agents for, 1410 iron reduction for, 1410 laboratory findings in, 1405, 1405t liver biopsy in, 1407-1408, 1408t liver transplantation for, 1410, 1603 natural history of, 1408-1409 obesity and, 104-105, 1401-1402 pathogenesis of, 1402-1405, 1404f risk factors for, 1408t, 1409 small intestinal bacterial overgrowth and, 1775 treatment of, 1409-1410, 1409t pregnancy-related, 201, 634-636, 635f diagnosis of, 635, 635f HELLP syndrome and, 633 jaundice in, 329 management of, 635-636 pathogenesis of, 635 FDG-PET. See Fluorodeoxyglucose positron emission tomography (FDG-PET). Fecal. See also Defecation; Stool. Fecal incontinence, 241-258 anal canal imaging in, 248, 248f anatomic considerations in, 242-243, 242f anorectal manometry in, 247-248, 247f, 251t balloon expulsion test in, 249 in children, 257-258 clinical features of, 245-246, 245t costs of, 241-242 defecography in, 248-249 in diabetes mellitus, 575 diagnostic testing in, 246-250, 251t diarrhea and, 211, 246 in elderly, 244, 257 epidemiology of, 241 evaluation of patient with, 245-250 grading system for, 246 history in, 245-246, 245t

Fecal incontinence (Continued) with ileal pouch–anal anastomosis, 2020 neurophysiologic testing in, 249, 249f-250f from obstetric injury, 243-244, 248 passive, 246 pathophysiology of, 242-245, 242f-243f, 243t physical examination in, 246 rectal sensory testing in, 248 saline infusion test in, 250 in spinal cord injury, 256-257 treatment of, 250-258, 252f, 257t anal plugs in, 254-255 anal sphincter bulkers in, 255 colostomy in, 256 electrostimulation in, 255 Malone enema in, 256 neuromuscular training in, 253-254, 253f, 254t pharmacologic, 252-253 radiofrequency energy in, 256 sacral nerve stimulation in, 256, 257f supportive measures in, 250-252 surgery in, 255-256, 255t urge, 246 Fecal occult blood test (FOBT) in colorectal adenoma, 2165 in colorectal cancer, 2221-2225, 2221t-2223t in diarrhea, 222 in gastrointestinal bleeding, 320-321 performance recommendations for, 2222, 2222t slide guaiac test in, 2222, 2222t trials of, 2222, 2223t Fecal retention, functional, 258, 266 Fecal seepage, 246, 257 Fecalith in appendicitis, 2060-2061 diverticulitis and, 2078 Fedotozine, for functional dyspepsia, 193 Feeding. See also Diet; Nutrition. acinar cell response to, 912 postoperative, ileus and, 2127 sham pancreatic enzyme secretion after, 926 thermic effect of, 48 tube. See Enteral nutrition. Felbamate, hepatotoxicity of, 1432 Felty’s syndrome gastrointestinal manifestations of, 558 hepatic involvement in, 559 Femoral hernia, 385-388 clinical manifestations and diagnosis of, 386-387, 387f complications and recurrence of, 388 etiology and pathogenesis of, 385-386, 386f incidence and prevalence of, 386 treatment and prognosis in, 387-388 Fenestration, for polycystic liver disease, 1590 Fennel, for colonic health, 2294t Fentanyl, in palliative care, 2281, 2281t Fenugreek, for colonic health, 2294t Ferrireductase, 1724 Ferritin, serum in hereditary hemochromatosis, 1244, 1244t in iron deficiency anemia, 321 in nonalcoholic fatty liver disease, 1405 Ferroportin, in hereditary hemochromatosis, 1240 Ferroportin 1, 1724-1725 Ferrous sulfate, for malabsorption, 74 Fertility in celiac disease, 1806 in Crohn’s disease, 1972 ileal pouch–anal anastomosis and, 2023 obesity and, 105 Fetal risks from drugs, 626, 626t-627t from radiation exposure, 627, 627t Fever in Crohn’s disease, 1953 familial Mediterranean gastrointestinal manifestations of, 558t, 563-564 peritonitis in, 618

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Fever (Continued) Katayama, in schistosomiasis, 1936 Oroya, hepatic manifestations of, 1353 paratyphoid, hepatic manifestations of, 1352 Q, hepatic manifestations of, 1353 Rocky Mountain spotted, hepatic manifestations of, 1353-1354 typhoid. See Typhoid fever. Fiber dietary, 1707 appendicitis and, 2061 colon cancer and, 42 colorectal adenoma risk and, 2163 colorectal cancer and, 2194-2195, 2197, 2199f, 2199t constipation and, 262, 275-276, 275t for diverticular disease, 83 diverticulosis and, 2075-2076 in hemorrhoids, 2260, 2263t restriction of, 95 malabsorption with, 1757t supplementation of, 275-276 commercial products for, 275t, 276 in diarrhea during enteral nutrition, 95 in diverticulosis, 2077 in familial adenomatous polyposis, 2183-2184 in irritable bowel syndrome, 2101 side effects of, 275-276 Fibrin adhesive for fistula-in-ano, 2267 for gastrointestinal fistula, 423 Fibrin injection therapy, for peptic ulcer bleeding, 880 Fibroblast growth factor, in gastrointestinal tract, 16 Fibrocalculous pancreatic diabetes, 931 Fibrocystic liver disease, 1500, 1589-1590, 1589f Fibroid polyp, inflammatory eosinophils in, 432 esophageal, 768 Fibrolamellar hepatocellular carcinoma, 1575, 1602 Fibromuscular dysplasia, vasculitis in, 2046-2047 Fibronectin, in ascites, 1526 FibroScan test, 1323, 1324t in hepatic fibrosis, 1235 Fibrosis in chronic pancreatitis, 987 in Crohn’s disease, 1949 cystic. See Cystic fibrosis. hepatic. See Hepatic fibrosis. periportal (“pipestem”), in schistosomiasis, 1362, 1362f retroperitoneal, 620-621, 621f FibroTest (FibroSURE), 1323, 1324t in hepatic fibrosis, 1235 in nonalcoholic fatty liver disease, 1408 Fibrovascular tumors, of esophagus, 769, 770f Fine needle aspiration in cholangiocarcinoma, 1162-1163 complications of, 661 in pancreatic cancer, 1022-1023 Fish oil for inflammatory bowel disease, 2293 for ulcerative colitis, 2000 Fish poisoning, 1884 Fish tapeworm infection, 1931, 1931f Fissure, anal. See Anal fissure. Fist-grip dynamometry, in nutritional assessment, 66-67, 78f Fistula anal. See Anal fistula. anocutaneous, 1634-1635 anovaginal, 2269 aortoenteric, 306, 608, 608f arteriovenous, splanchnic, portal hypertension in, 1501 atrial-esophageal, after radiofrequency catheter ablation, 739 biliary, bile acid circulation disturbances in, 1087

Fistula (Continued) cholecystoenteric, in gallstone disease, 1118-1119, 1119t colocutaneous, after percutaneous endoscopic gastrostomy, 90-91 colovaginal, in diverticulitis, 2082 colovesicular in Crohn’s disease, 1951 in diverticulitis, 2082 in Crohn’s disease, 422-423, 1947-1951, 1947f, 1951f in diverticulitis, 2082 enteric paraprosthetic, 608, 608f small intestinal bacterial overgrowth from, 1773 enterocutaneous in Crohn’s disease, 1951 infliximab for, 1967 enterovaginal, in Crohn’s disease, 1951 enterovesicular, in Crohn’s disease, 1951 gastrointestinal, 419-424. See also Gastrointestinal fistula. in imperforate anus, 1634f pancreatic, in chronic pancreatitis, 1014 perianal, in Crohn’s disease, 1950-1951, 1951f rectourethral, 1634f, 1635 rectovaginal, in Crohn’s disease, 1951 rectovesicular, 1634f, 1635 spontaneous closure of, prognostic factors in, 421, 422t tracheoesophageal, 668-670, 669f, 669t-670t distal, 669f, 670, 670t H-type, 669f, 670, 670t vestibular, 1635 Fistula-in-ano, 2267-2268, 2268f Fistulotomy, for fistula-in-ano, 2267 Fitz-Hugh–Curtis syndrome, 617, 617f gonococcal infection in, cholecystitis versus, 1115 hepatic, 1352-1353 Flamingo Wallstent, for esophageal carcinoma, 766 Flatus. See also Gas, intestinal. colonic hydrogen absorption and, 1741, 1741f excessive, 218-219 gases in, 235 intestinal bacteria metabolic activity and, 1772 normal passage of, 236 odoriferous, 236 excessively, 238 voluminous, 237-238 Flora. See Bacteria, intestinal (commensal). Floxuridine hepatotoxicity of, 1441 intrahepatic infusion of, for colorectal cancer, 2236-2237 Flucinonide, for aphthous ulcers, 356 Flucloxacillin, hepatotoxicity of, 1441 Fluconazole for candidal esophagitis, 526, 742 hepatotoxicity of, 1435 Fluid(s). See also Water. daily requirements for, 85 during enteral nutrition, 93 disturbances of, with vomiting, 206 intake of, constipation and, 262, 275 overload of ileus and, 2125-2126 in refeeding after nutritional deficiency, 71 restriction of, in ascites, 1536 therapeutic. See Fluid therapy. transport of. See Fluid and electrolyte transport. Fluid and electrolyte transport, 1675-1694 absorptive stimuli in, 1687t, 1690 active primary, 1680f, 1681 secondary, 1680-1681, 1680f of bicarbonate, 1685 bile acids in, 1692 carrier-mediated, 1680, 1680f channel-mediated, 1680, 1680f of chloride, 1682-1684, 1683f

Fluid and electrolyte transport (Continued) by diffusion facilitated, 1680, 1680f simple, 1679-1680, 1680f direction of movement in, 1676f, 1678 enterotoxins in, 1688t, 1691-1692 epithelial cell model for, 1676f, 1677 guanylin in, 1691 intestinal architecture related to, 1675-1677, 1676f ion transporters in, 1681-1686, 1681f-1683f spatial distribution of, 1677-1678, 1678f nitric oxide in, 1691 overview of, 1675, 1676f of potassium, 1684-1685 pump-mediated, 1680f, 1681 reactive oxygen metabolites in, 1691 regulation of extracellular, 1687, 1687t-1688t homocellular, 1693-1694 immunologic, 1688-1689 intracellular, 1692-1693, 1693f neural, 1686f, 1687-1688 osmotic, 1689-1690 paracrine immunoneuroendocrine system (PINES) in, 1686, 1686f systemic, 1689 secretory stimuli in, 1687t-1688t, 1690-1692 segmental heterogeneity of, 1677-1678, 1678f serotonin in, 1691 of short-chain fatty acids, 1685-1686 of sodium, 1681-1682, 1681f tight and leaky epithelia in, 1676f, 1678-1679 trans epithelial, 1676f, 1678-1679 transpleural, 1679 of water, 1679 Fluid therapy in acute pancreatitis, 977 in cholera, 1851, 1851t in esophageal variceal bleeding, 1509 in gastric outlet obstruction, 884 in gastrointestinal fistula, 420-421 in infectious diarrhea, 1851t, 1884-1885 in palliative care, 2284 in peritonitis, 615 in shigellosis, 1860 in short bowel syndrome, 1784, 1784t in small intestinal obstruction, 2110 in VIPoma, 510 in vomiting, 207 Flukes, 1933-1939 blood, 1935-1939, 1936f-1937f, 1938t. See also Schistosomiasis. hepatic, 1356t-1357t, 1361-1363, 1361f, 1934-1935, 1934t, 1935f intestinal, 1923f, 1933-1934 Flumazepil, for hepatic encephalopathy, 1546 Fluorescein dilaurate test, for pancreatic function testing, 929 Fluorescence in situ hybridization (FISH) in cholangiocarcinoma, 1174-1175, 1176f in Whipple’s disease, 1838, 1839f Fluorescence microscopy, in Cyclospora cayetanensis infection, 1916 Fluoride, dietary, 56t-57t Fluorodeoxyglucose positron emission tomography (FDG-PET) in carcinoid tumors, 485-486 in esophageal cancer, 757-759, 759f in gastrointestinal stromal tumors, 466-468, 467f, 471 in ulcerative enteritis, 2055 Fluoroquinolones for Campylobacter spp. infection, 1865t, 1868 in cirrhotic patients with variceal bleeding, 307 for infectious diarrhea, 1885-1886 prophylactic, for traveler’s diarrhea, 1875t resistance to, Campylobacter spp., 1868 for salmonellosis, 1864, 1865t for shigellosis, 1860-1861, 1865t for traveler’s diarrhea, 1874 for typhoid fever, 1867

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxxv

xxxvi

Index Fluoroscopy in percutaneous enteral access, 93 in small intestinal motility measurement, 1652-1653 5-Fluorouracil for anal warts, 2270, 2271t for colorectal cancer, 2234-2235, 2235f plus leucovorin, 2232-2233, 2235, 2235f plus leucovorin and irinotecan, 2232-2233, 2235 plus leucovorin and oxaliptan, 2232-2233, 2235 plus levamisole, 2232 plus radiation therapy, 2233-2234 for esophageal cancer, 764 for metastatic pancreatic endocrine tumors, 521 radiation enteritis and, 644-645 for small intestinal adenocarcinoma, 2152 Fluoxetine for anorexia nervosa, 133 for bulimia nervosa, 133 for obesity, 111-112 Flushing in carcinoids, 365, 482, 482f, 484 in VIPoma, 509 Fluticasone, for eosinophilic gastrointestinal disorders, 434 Focal nodular hyperplasia of liver, 1587-1588, 1588f in pregnancy, 637 Folate conjugase, 1714, 1714t Folate/folic acid, 81t, 82 absorption of, 1718, 1718t deficiency of after bariatric surgery, 118 in celiac disease, 1810-1811 in elderly persons, 1758 serum test for, 1752 in tropical sprue, 1826 dietary, 53t-55t colorectal adenoma risk and, 2164 colorectal cancer and, 2197, 2199f, 2199t reference nutrient intake for, 1718t sources of, 1718 esophageal cancer and, 747 malabsorption of, 1739 hereditary, 1759t-1762t supplementation of, for colorectal cancer prophylaxis, 2008 Foley catheter, for foreign body removal, 400 Follicle-associated epithelium, overlying Peyer’s patches, 25 Follicular colonization, in gastric MALT lymphoma, 449 Follicular lymphoma, 445-446 small intestinal, 453-454, 454f Fontan procedure, protein-losing gastroenteropathy after, 441 Food adverse reaction to, 139 causing heartburn, 178 contaminated, hepatotoxicity of, 1453-1454 hypersensitivity to. See Food allergy. impaction of, esophageal, 397-401, 398f-399f intake of. See also Satiety. carbohydrates in, 1706-1707 fats in, 1698-1706, 1699t protein in, 1712 regulation of, 1696 by gastrointestinal peptides, 18, 18t stomach and, 802-803, 803f intestinal gas production and, 238 intolerance of. See also Food allergy. definition of, 139 differential diagnosis of, 142t dyspepsia in, 184 irritable bowel syndrome and, 2098 oxalate content of, 1787t poisoning from. See Food poisoning. reflux and, 721 solid, gastric emptying of, 794-797, 795f-797f toxic reaction to, 139

Food allergy, 139-148 background on, 139 clinical features of, 142-146, 142t, 146t definition of, 139 diagnosis of, 142t, 146-148, 146t differential diagnosis of, 142t foods responsible for, 146t gastrointestinal IgE-mediated, 141-143 mixed IgE- and non–IgE-mediated, 143-145 non–IgE-mediated, 145-146, 146t natural history of, 148 oral tolerance and, 141-142 pathogenesis of, 139-142, 140t, 141f prevalence of, 139 radiologic findings in, 142 treatment of, 148 Food bolus, impacted, 401 Food journal, for eating disorders, 131-132 Food poisoning Bacillus anthracis in, 1879t-1880t, 1883-1884 Bacillus cereus in, 1879t-1880t, 1882 bacterial, 1878-1884, 1879t-1881t Clostridium botulinum in, 1879t-1880t, 1882-1883 Clostridium perfringens in, 1879t-1880t, 1881-1882 Listeria in, 1879t-1880t, 1882 meat-related, 1929-1930 parasitic, 1879t rates of, 1879t Salmonella spp. in, 1878-1879, 1879t-1880t seafood-related, 1884, 1930 Staphylococcus aureus in, 1879-1881, 1879t-1880t viral, 1879t Food protein–induced enterocolitis, 145, 146t, 429 endoscopy in, 431 treatment of, 433-435, 433t Foodstuffs, hepatotoxicity of, 1454-1455, 1457t Foramen of Winslow hernia, 392-395 Foreign bodies, 397-404 blunt objects as, 402-404 colorectal, 398-399, 404, 404f complications of, procedure-related, 404 diagnosis of, 399-400, 399f endoscopy for diagnostic, 400 therapeutic, 400-401, 401t, 658 epidemiology of, 397-398, 398f food bolus as, 401 history and physical examination in, 399 long objects as, 402 narcotic packets as, 404 passage of, points of impedance to, 398, 398f pathophysiology of, 398-399, 398f radiographic studies in, 399-400, 399f rectal, 398-399, 404, 404f, 2274 sharp and pointed objects as, 402, 402f-403f Forest classification of peptic ulcer bleeding, 295 Formulas, nutritional enteral, 94, 94t blenderized, 94 immune-enhancing, 94 modular, 94 predigested, 94 specialty, 94, 94t infant, for glycogen storage disease type I, 1264 parenteral central, 85, 85t peripheral, 85-86, 86t Foscarnet, for cytomegalovirus colitis, in HIV/ AIDS, 528 Fosinopril, hepatotoxicity of, 1433 Foveolar hyperplasia, in reactive gastropathy, 856, 856f Fractalkine, in gut-associated lymphoid tissue, 29 Fracture, proton pump inhibitors and, 871 Frey procedure, for chronic pancreatitis, in pain control, 1008

Fructose absorption of, 1706, 1711 dietary, 1706 hepatic metabolism of, 1219 intolerance of, 1755 malabsorption of, 1711, 1755 irritable bowel syndrome and, 2098 Fructose-6-phosphate, in hepatic glucose metabolism, 1217-1218 Fruits, gastric cancer and, 892 Fucosylated alpha fetoprotein, in hepatocellular carcinoma, 1572, 1572t Fumagillin, for Enterocytozoon bieneusi infection, 1918 Fumarylacetate hydrolase deficiency, in tyrosinemia, 1269, 1270f Functional abdominal pain syndrome, 165-170 biopsychosocial model of, 165-166, 166f central nervous system in, 165-168, 166f-167f clinical features of, 168-169 clinical implications of, 168 diagnosis of, 169 differential diagnosis of, 169 epidemiology of, 165 history in, 168 pathophysiology of, 165-168, 166f-167f patient behavior in, 168 physical examination in, 168-169 Rome II diagnostic criteria for, 165, 165t treatment of, 169-170 patient-physician relationship in, 169 pharmacologic, 169-170 plan for, 169 psychologic, 170 visceral pain transmission in ascending, 166-167, 166f descending modulation of, 167, 167f visceral sensitization in, 167 Functional constipation, 264, 265t Functional dyspepsia. See Dyspepsia, functional. Functional fecal retention, 258, 266 Fundectomy, gastric motility disorders after, 808 Fundic gland. See Oxyntic (fundic) gland. Fundic relaxants, for functional dyspepsia, 193 Fundoplication for Barrett’s esophagus, 730 failure of, 725 gastric motility disorders after, 808 for gastroesophageal reflux disease, 723f, 724-725 pseudoachalasia after, 693 Fungal infection in acute liver failure, 1562 gastritis in, 851-852 hepatic, 1365-1366 in hematopoietic stem cell transplantation candidates, 544 in HIV/AIDS diarrhea from, 526t-527t, 530 hepatic, 534 intestinal, after transplantation, 541-542 after liver transplantation, 1609 peritonitis in, 617 after solid organ transplantation, 539 Fungal liver abscess, after hematopoietic stem cell transplantation, 547t, 549, 554 Fungal superinfection, in abdominal abscess, 418 Furazolidone, for giardiasis, 1914 Furosemide for ascites, 1537 colonic ischemia from, 2040 Fusidic acid, for Clostridium difficile-associated diarrhea and colitis, 1899

G

G phase, cell cycle, 31, 32f G protein(s) activation of, 13-14, 13f classification of, 13, 14t effector systems of, 14, 14t membrane-associated, oncogenes and, 36-37

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index G protein–coupled receptor(s) in cellular proliferation, 33 classification of, 13, 14t desensitization of, 14 ligand binding to, 13-14, 13f resensitization of, 14 signal transduction of, 13, 13f structure of, 13, 13f GABA. See Gamma-aminobutyric acid (GABA). Gabapentin for irritable bowel syndrome, 2103 for neuropathic pain, 2283t Gabexate, for acute pancreatitis, 967-968, 980 Gag reflex, 687 Galactose hepatic metabolism of, 1219 malabsorption of, 1711 congenital, 1641 Galactose elimination capacity test, 1235 Gallbladder adenocarcinoma of. See Gallbladder carcinoma. adenoma of, 1149, 1150t adenomyoma of, 1146, 1147f, 1149, 1150t, 1183-1184 adenomyomatosis of, 1146-1149, 1147f-1148f agenesis of, 1050 anatomy of, 1048f, 1049-1050 arterial supply to, 1049 carcinoma of. See Gallbladder carcinoma. cholesterolosis of, 1144-1146, 1144f-1145f. See also Cholesterolosis. congenital anomalies of, 1050 development of, 1046f, 1047 in digestion and absorption, 1697 disorders of in cystic fibrosis, 947-948 in infants and children, 1063-1066 obesity and, 104 in somatostatinoma, 512 double, 1050 dyskinesia of, 1067 in children, 1066 emptying of, delayed, 1067 gallstone growth and, 1101 folded, 1050 Hartmann’s pouch of, 1048f, 1049 heterotopia of, 1183-1184 hydrops of, 1065-1066, 1113-1114 hypomotility of, in gallstone disease, 1099-1100 inflammation of. See Cholecystitis. innervation of, 1049-1050 kinked, 1050 lymphatics of, 1049-1050 malpositions of, 1050 mesenchymal tumors, 1183 mucosa of, 1050 multiple, 1050 papilloma of, 1145-1146, 1149, 1150t polyps of, 1149-1151 adenomatous, gallbladder carcinoma and, 1178 clinical features of, 1150-1151, 1151f definition of, 1149 diagnosis of, 1150-1151, 1151f epidemiology of, 1149 inflammatory, 1149, 1150t natural history of, 1151 pathology of, 1149-1150, 1150t in primary sclerosing cholangitis, 1151 treatment of, 1151 porcelain, in gallstone disease, 1119, 1119t in pregnancy, 626 pseudotumors of. See Gallbladder, cholesterolosis of. removal of. See Cholecystectomy. small, in cystic fibrosis, 947-948 strawberry, 1145 varices of, 1499, 1499f venous drainage of, 1049-1050 wall of, thickened, in acute acalculous cholecystitis, 1142-1143, 1143f

Gallbladder carcinoma, 1177-1181 adenomyomatosis and, 1148, 1148f clinical features and diagnosis of, 1179-1180, 1179f epidemiology of, 1177 etiology of, 1177-1178, 1178t gallstone disease and, 1137, 1177-1178 identified during laparoscopic cholecystectomy, 1181, 1181f pathogenesis of, 1178-1179 pathology of, 1178 progression to, 1149, 1178-1179 risk factors for, 1177-1178, 1178t staging of, 1180, 1180t treatment of, 1180-1181, 1181f Gallbladder ejection fraction, in acalculous biliary pain, 1141 Gallbladder sludge in acute acalculous cholecystitis, 1141-1142 prevention of, 1144 Gallstone disease, 1089-1120 adenomyomatosis and, 1147-1148 age and, 1090-1091 asymptomatic, 1104-1105 in children, 1064 cholecystectomy for, 1129 treatment of, 1129-1130 after bariatric surgery, 118-119, 1106 biliary pain in, 1107t, 1112-1113, 1130 biliary sludge and, 1091-1092 black pigment in bile duct, 1116 in children, 1063-1064 composition of, 1089 pathogenesis of, 1104 brown pigment, 329 in bile duct, 1116 in children, 1064 composition of, 1089 pathogenesis of, 1104, 1105f cancer and, 1105-1106, 1137 ceftriaxone-induced, in children, 1064 in children, 1063-1064, 1132 clinical features of, 1064 conditions associated with, 1063, 1063t pathogenesis of, 1063 treatment of, 1064 cholangitis in, 1107t, 1117-1118 cholecystitis in acute, 1107t, 1113-1116, 1114f, 1130-1132, 1130t, 1131f chronic, 1112-1113 emphysematous, 1111-1112, 1111f, 1118, 1119t, 1131-1132 cholecystoenteric fistula in, 1118-1119, 1119t cholescintigraphy in, 1108t, 1110-1111, 1110f cholesterol, 329 in children, 1064 composition of, 1089 clinical features of, 1105-1118, 1106f, 1107t complications of, uncommon, 1118-1119, 1119t computed tomography cholangiography in, 1108t, 1111-1112, 1111f in Crohn’s disease, 1101, 1955 definition of, 1089 in diabetes mellitus, 1092, 1105 diagnosis of, 1105-1118, 1106f, 1107t diet and, 1091 distribution of, 1090, 1091f drug-induced, 1092 dyspepsia in, 184 endoscopic retrograde cholangiopancreatography in, 1108t, 1111, 1111f endoscopic ultrasonography in, 1108t, 1109-1110, 1110f epidemiology of, 1089-1093, 1090f-1091f gallbladder carcinoma and, 1177-1178 gender and, 1090-1091 genetics of, 1101-1102, 1103t heart transplantation and, 1106 after hematopoietic stem cell transplantation, 554-555

Gallstone disease (Continued) in hematopoietic stem cell transplantation candidates, 545 in ileal disease, 1093 imaging studies in, 1106-1112, 1108t incidence of, 1089-1090 intrahepatic, 1089 in recurrent pyogenic cholangitis, 1167-1170, 1168f lipid abnormalities and, 1092 magnetic resonance cholangiography in, 1108t, 1111-1112, 1112f Mirizzi’s syndrome in, 1114, 1119, 1119t, 1132-1133 natural history of, 1104-1105 obesity and, 104, 1092 oral bile acid dissolution therapy for, 1121-1122, 1122t, 1123f oral cholecystography in, 1108t, 1110 pancreatitis in. See Pancreatitis, gallstone. pathophysiology of, 1097-1101, 1097f antinucleating/pronucleating imbalance in, 1098-1099 cholesterol hypersecretion in, 1097 cholesterol nucleation and crystallization in, 1098 gallbladder dysfunction in, 1099-1100 intestinal factors in, 1100-1101 stone growth in, 1101 pigment stones in, pathogenesis of, 1102-1104, 1105f porcelain gallbladder in, 1119, 1119t in pregnancy, 630, 1091, 1132 prevalence of, 1090, 1090f in primary sclerosing cholangitis, 1162 risk factors for, 1090-1093 in short bowel syndrome, 1788-1789 in sickle cell disease, 1106 in spinal cord injury, 1093 symptomatic, 1105 total parenteral nutrition and, 1091 treatment of, 1121-1138 in children, 1132 choice of, 1129 dissolution therapy in, 1121-1122, 1122t, 1123f extracorporeal shock wave lithotripsy in, 1122-1124, 1123t, 1124f indications for, 1129-1133 medical, 1121-1124 in pregnancy, 1132 surgical, 1124-1129. See also Cholecystectomy. types of, 1089 ultrasonography in, 1106-1109, 1108t, 1109f weight loss–induced, 1091 in Wilson disease, 1252 Gallstone ileus, 1119, 1133, 2115-2116, 2115f GALT. See Gut-associated lymphoid tissue (GALT). Gamma-aminobutyric acid (GABA) in gastrointestinal tract, 5 in hepatic encephalopathy, 1544 Gamma globulin, in autoimmune hepatitis, 1468, 1469t Gammaglutamyl transpeptidase, 1714, 1714t in alcohol abuse, 1389, 1390f elevation in, for alcohol use ranges, 1232 Ganciclovir, for cytomegalovirus colitis, in HIV/ AIDS, 528 Ganglia (ganglion cells), intestinal, 1623, 1623f, 1646. See also Myenteric plexus; Submucosal plexus. absence of. See Hirschsprung’s disease. hyperplasia of, 1639-1640, 1639f Gangliocytic paraganglioma, duodenal, 478, 480 Ganglioneuromatosis, intestinal, 2187 Ganglionitis chronic intestinal pseudo-obstruction in, 2132, 2140 idiopathic myenteric, 2140

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxxvii

xxxviii

Index Gangrene appendiceal, 2060 colonic, 2041, 2041f, 2043 gas, 1351-1352 gastric, 849-850 Gap junctions, 1644 in acinar cells, 913, 921 in hepatocytes, 1208 in intestinal epithelium, 1676f, 1679 Gardner’s syndrome, 2177t, 2180-2182 ampullary carcinoma in, 1181 Garlic gas derived from, excretion of, 236 for gastrointestinal cancer, 2298 for hepatopulmonary syndrome, 1552 Gas intestinal, 233-240 abdominal distention and bloating from, 238-240, 239f air swallowing causing, 233, 235f, 237 anal excretion of, 236. See also Flatus. bacterial consumption of, 235-236 breath excretion of, 236 carbohydrate malabsorption and, 234-235, 234t, 237 carbon dioxide in, 233, 235 in colon, 234-236, 234f-235f composition of, 233, 234f disorders of, 237-240 elimination of, 236 foods affecting, 238 hydrogen in, 233-235 in irritable bowel syndrome, 238-240, 239f metabolism and excretion of, 233-236, 235f methane in, 233-236 odoriferous, 236 in pneumatosis cystoides intestinalis, 240 propulsion of, through intestinal tract, 236-237, 237f retention of, 237-238 in small intestine, 233-234, 234f-235f sulfur-containing, 236, 238 tolerance of, 236-237, 237f volume of, 233, 234f presence of, in acute mesenteric ischemia, 2031, 2031f propulsion/expulsion abnormalities of, in irritable bowel syndrome, 2096-2097 Gas challenge test, 236-237, 237f Gas gangrene, 1351-1352 Gastrectomy diet after, 95 for gastric cancer, 902 gastric cancer after, 896 for gastric MALT lymphoma, 451 malabsorption after, 1756-1757 for peptic ulcer bleeding, 882 in Zollinger-Ellison syndrome, 504 Gastric. See also Stomach. Gastric accommodation functional dyspepsia and, 187-188 in gastric motility, 794-796, 796f-797f Gastric acid drug reduction of. See Antacids; Proton pump inhibitors. enteric infection and, 1844 reflux of. See Gastroesophageal reflux disease. regurgitation of. See Heartburn (pyrosis). secretion of after antrectomy, 828 basal, 827, 827t cholecystokinin in, 823 development of, 817 in duodenal ulcer, 828 endocrine, paracrine, and neurocrine cells in, 817, 818f, 821-827, 821f-823f enterogastrones in, 823 in gastric ulcer, 828 in gastrinoma, 828-829, 828f in gastroesophageal reflux disease, 713 ghrelin in, 823 Helicobacter pylori infection and, 713, 826-827, 826f, 836

Gastric acid (Continued) in hypercalcemia, 828 increased, diseases associated with, 828-829, 828f inhibition of, somatostatin in, 817-819, 820f-822f, 823 intracellular signaling pathway in, 821f, 823-824, 824f in mastocytosis, 828 maximal, 827, 827t meal-stimulated, 825-826 measurement of, 828 measurement of, 827-828 indications for, 827 methods for, 827, 827t orexin-A in, 823 peak, 827, 827t peptic ulcer disease and, 863 proton pump in, 821f, 823-824, 824f regulation of, 817, 818f, 821-827, 821f sham feeding–stimulated, measurement of, 828 stimulation of acetylcholine in, 821f-822f, 822-823 gastrin in, 821-822, 821f-822f histamine in, 821, 821f-822f in Zollinger-Ellison syndrome, 499, 503-505, 828 Gastric adenoma, in familial adenomatous polyposis, 2179 Gastric antral vascular ectasia (watermelon stomach), 601-602, 602f, 1514-1516, 1515f bleeding in, 305-306, 305f diagnosis of, 1514, 1515f diffuse, 1514 after hematopoietic stem cell transplantation, 549-550 management of, 1515-1516, 1515f versus portal hypertensive gastropathy, 602, 1515, 1515t Gastric artery, 775 left, 2027, 2028f Gastric atresia, 780-781, 780t Gastric banding gastroesophageal reflux disease after, 715 laparoscopic adjustable, 115, 116t-117t. See also Bariatric surgery. Gastric biopsy in gastritis, 846 in Helicobacter pylori infection, 839-840, 840f, 840t indications and sites for, 845, 846f Gastric bypass diet after, 95 malabsorption after, 1758 Roux-en-Y, 115, 116t-117t. See also Roux-en-Y gastric bypass. Gastric cancer, 887-906 alcohol and, 892 autoimmune metaplastic atrophic gastritis and, 848 barium radiography in, 900 cancer stem cell model for, 895 classification of, 888, 889f, 899-902, 899f-900f, 901t clinical features of, 898-899 computed tomography gastrography in, 900 computed tomography in, 901 diagnosis of, 899-900, 899f dietary factors in, 892 diffuse type of, 888, 889f, 893 dyspepsia in, 184 E-cadherin in, 893-894 early versus advanced, 900 endoscopic ultrasonography in, 901, 901f endoscopy in, 899-900, 899f epidemiology of, 887, 888f, 888t-889t etiology and pathogenesis of, 888-893, 889f-890f in familial adenomatous polyposis, 893, 2179 familial clustering of, 892-893 genetic abnormalities in, 893-895, 894t Helicobacter pylori infection and, 835, 838

Gastric cancer (Continued) Helicobacter pylori infection in, 888-893, 890f in hereditary nonpolyposis colorectal cancer, 893 inflammation and, 888-889, 890f interleukin-1β in, 893 interleukin-10 in, 893 intestinal type of, 888, 889f-890f in juvenile polyposis, 893 magnetic resonance imaging in, 901-902 metastasis of, 899 multistep model of, 888-889, 890f obesity and, 892 positron emission tomography in, 902 premalignant condition(s) in, 895-897 chronic atrophic gastritis as, 895 dysplasia as, 895-896, 895t, 896f gastrectomy history as, 896 gastric polyps as, 896 gastric ulcers as, 896-897 intestinal metaplasia as, 895 Menetrier’s disease as, 897 prevention of, 897-898 antioxidants for, 898 aspirin/NSAIDS for, 898 green tea for, 898 Helicobacter pylori eradication for, 897-898 prognosis in, 902, 902t restaging of, 902 risk factors for, 889t screening for, 897-898 serum markers in, 900 smoking and, 892 staging of, 900-902, 900f-901f, 901t treatment of, 902-904 chemoradiotherapy for, 903-904 chemotherapy for, 903-904 endoscopic, 902-903, 903f intraperitoneal chemotherapy for, 904 surgical, 902 for unresectable disease, 904 tumor necrosis factor-α in, 893 Gastric dysplasia, gastric cancer and, 895-896, 895t, 896f Gastric dysrhythmias assessment of, 805, 805f treatment approaches based on, 811-812, 811t in functional dyspepsia, 809, 809t in gastroesophageal reflux disease, 809 in idiopathic gastroparesis, 808-809 in irritable bowel syndrome, 809 Gastric electrical stimulation for gastric motility disorders, 812, 813t for gastroparesis, 574 for vomiting, 209 Gastric emptying. See also Gastric motility. assessment of, 803-804 breath tests in, 804 capsule technology in, 797f, 804 computed tomography in, 804 magnetic resonance imaging in, 804 scintigraphy in, 796f, 803-804 treatment approaches based on, 811-812, 811t ultrasonography in, 798f, 804 in vomiting, 205-206 of carbohydrates, 800, 800f delayed. See Gastroparesis. in eating disorders, 130, 135 of fatty acids, duodenum and, 799-800 food intake regulation and, 802-803, 803f functional dyspepsia and, 187 gender differences in, 801 of liquids, 798, 798f-799f in lower gastrointestinal bleeding, 290-291 pylorus in, 797 rapid, in dumping syndrome, 810 regulation of, 798-801, 800f, 800t, 1696, 1696f by gastrointestinal peptides, 18-19, 19t of solids, 794-797, 795f-797f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Gastric gland, 776, 777f. See also Oxyntic (fundic) gland. Gastric inhibitory polypeptide. See Glucosedependent insulinotropic polypeptide (GIP). Gastric lavage, in upper gastrointestinal bleeding, 290 Gastric lipase, 829, 830f, 1699 Gastric lymphoma. See Lymphoma, gastric. Gastric motility, 789-816. See also Gastric emptying. assessment of, 803-806 antroduodenal manometry in, 795f, 804 barostat tests in, 796f, 805 breath tests in, 804 capsule technology in, 797f, 804 computed tomography in, 804 electrogastrography for, 811-812, 811t electrogastrography in, 790f-791f, 799f, 804-805, 805f gastric emptying tests in, 811-812, 811t histopathologic studies in, 805-806 magnetic resonance imaging in, 804 nutrient drink test in, 805 scintigraphy in, 796f, 803-805 single photon emission computed tomography in, 805 ultrasonography in, 798f, 804 water-load test in, 805 development of, 803 disorders of, 806-810 delayed gastric emptying as. See Gastroparesis. diagnosis of, 810-812 diagnostic testing in noninvasive, 811-812, 811t standard, 811 in dumping syndrome, 810 in functional dyspepsia, 809, 809t in gastroesophageal reflux disease, 809 Helicobacter pylori infection and, 810 histopathologic studies of, 805-806 in intestinal pseudo-obstruction, 809-810 in irritable bowel syndrome, 809 in neurologic disorders, 810 paraneoplastic, 564 physical examination in, 811 spectrum of, 806, 806f symptoms of, 810-811 treatment of, 811t, 812-814, 813t dietary, 813t, 814 electrical, 812-814, 813t endoscopic, 813t, 814 pharmacologic, 812, 813t vomiting in, 201, 810 electrophysiologic basis of, 789-798, 790f-791f in fasted state, 794, 795f in fed state, 794-798 after liquid ingestion, 798, 798f-799f after solid food ingestion, 794-797, 795f-797f migrating motor complex in, 794, 795f neural modulation of, 794 in pregnancy, 625-626 receptive relaxation and accommodation in, 794-796, 796f-797f regulation of, 798-801, 800f, 800t slow waves in, 789-791, 790f-791f Gastric outlet obstruction in eosinophilic gastroenteritis, 428-429, 428f gastroparesis in, 808 malignant, treatment of, 904 in peptic ulcers, 868, 884-885 endoscopic therapy for, 884-885, 885f medical management of, 884 surgical management of, 885 string sign in, 428-430, 428f vomiting in, 199-201 Gastric pacemaker, 789-790, 790f-791f, 792 Gastric pacing, for gastric motility disorders, 812-813, 813t Gastric peristaltic waves, 789-790, 791f, 792, 796, 796f-797f

Gastric polyps in familial adenomatous polyposis, 2179, 2180f fundic gland, 896 gastric cancer and, 896 hyperplastic, 896 Gastric secretion, 817-832. See also specific biochemicals, e.g., Gastric acid. Gastric surgery, diarrhea after, 228 Gastric tonometry exercise testing, in intestinal angina, 2045 Gastric ulcers. See also Peptic ulcer disease. bleeding. See Peptic ulcer bleeding. endoscopy in, 864-865, 865f gastric acid secretion in, 828 gastric cancer and, 896-897 in gastrinoma, 864 Helicobacter pylori infection in, 838, 872, 889-890 high, 882-883 in mastocytosis, 864 mucosal defenses against, 863 NSAID-associated. See also Peptic ulcer disease, NSAID-associated. prevention of, 873-876 treatment of, 872-877 obstructing, 884-885 perforated, 883-884 in pregnancy, 629 radiation-induced, 857 refractory, 877 after solid organ transplantation, 541 stress-related, 877-878 surgical management of, 882-883 Gastric varices, 1494-1495, 1510-1513 bleeding from, 306-308 acute control of, 1511-1512, 1512f portosystemic shunts for, 1513 primary prophylaxis of, 1511 secondary prophylaxis of, 1512-1513 classification of, 1510-1511 detection of, 1496-1498 natural history of, 1511 transvenous obliteration of, 1513 Gastric vascular ectasia, antral. See Gastric antral vascular ectasia (watermelon stomach). Gastric vein, 775 left, anatomy of, 1489, 1490f Gastric volvulus, 383-385, 780t, 783 clinical manifestations and diagnosis of, 384, 385f etiology and pathogenesis of, 383-384, 384f-385f incidence and prevalence of, 384 mesenteroaxial, 384, 384f organoaxial, 384, 384f treatment and prognosis in, 384-385 vomiting in, 199-200 Gastric wall, layers of, 776 Gastrin, 6-7 clinical applications of, 7 excess of, 7 forms of, 7 in gastric acid secretion, 821-822, 821f-822f growth-promoting effects of, 16-17 in mucosal proliferation, 822 production of, 6-7 receptors for, 7 release of, 7 serum, in gastrinoma, 502-503 Gastrin/CCK receptors, 821-822 Gastrin provocation tests, in Zollinger-Ellison syndrome, 503 Gastrin-releasing peptide in cephalic phase of pancreatic enzyme secretion, 926 growth-promoting effects of, 17 as satiety signal, 1696 Gastrin-secreting (G) cells, of pyloric gland, 778, 778f Gastrinoma, 498-505 clinical features of, 499t, 501-502, 502t definition of, 498-499 diagnosis of, 500f, 502-503

Gastrinoma (Continued) duodenal, 477-480, 499-501 gastric, 478 gastric acid secretion in, 499, 503-505, 828-829, 828f gastric folds in, 499, 502 gastrin in, 502-503 growth patterns of, 500-501 lymph node, 499 malignant, 500 metastatic, 500-501, 501f natural history of, 501, 501f pancreatic, 499-501 pathology and pathophysiology of, 499-501 surgical management of, 505, 505f treatment of, 503-505 ulcer disease in, 864 Gastrinoma triangle, 499 Gastritis, 845-860 in actinomycosis, 850 allergic, 856 in anisakiasis, 852 antral, 846 diffuse, 846f in ascariasis, 852 in aspergillosis, 852 atrophic. See also Anemia, pernicious. autoimmune metaplastic, 846f, 847-848 chronic, gastric cancer and, 895 corporal, 895 diffuse corporal, 846f environmental metaplastic, 846-847 multifocal, 846f, 895 bacterial, 849-851, 851f biopsy in, 846 in candidiasis, 851 chronic nonspecific, 846-848, 846f-847f classification of, 845 collagenous, 853, 853f in Crohn’s disease, 852-853, 852f, 855-856 in cryptosporidiosis, 852 in cytomegalovirus, 849, 849f in diabetes mellitus, 574 differential diagnosis of, 859 emphysematous, 850 in enterovirus, 849 eosinophilic, 854-855, 854f in Epstein-Barr virus, 849 erosive. See Gastropathy, reactive. fungal, 851-852 in graft-versus-host disease, 856 granulomatous, 852-853, 852f in Helicobacter pylori infection, 838, 846, 846f-847f, 863 antrum-predominant, 863 lymphocytic, 854 pan-, 863 protein-losing gastroenteropathy in, 440 in herpes simplex, 849 in histoplasmosis, 851-852 in hookworm, 852 hypertrophic, giant. See Ménétrier’s disease. infectious, 849-852 in inflammatory bowel disease, 855-856 lymphocytic, 853-854, 854f in measles, 849 in Mycobacterium avium-intracellulare, 850 in Mycobacterium tuberculosis, 850 necrotizing, acute, 849-850 parasitic, 852 phlegmonous (suppurative), 849-850 in phycomycosis, 852 radiation-induced, 641-642 in sarcoidosis, 853 in strongyloidiasis, 852 in syphilis, 850-851, 851f treatment of, 859 in ulcerative colitis, 855-856 in varicella-zoster virus, 849 viral, 849, 849f vomiting in, 200 xanthogranulomatous, 853 from zinc, 1256-1257

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xxxix

xl

Index Gastritis cystica profunda, 855-856, 855f Gastroduodenal artery, 779, 911 Gastroduodenal Crohn’s disease, 1950 Gastroduodenal junction, 773 Gastroduodenal region, symptoms originating from, 183. See also Dyspepsia. Gastroenteritis. See also Diarrhea, infectious; Enteritis. abdominal pain in, 154t eosinophilic. See Eosinophilic gastroenteritis. in salmonellosis, 1863, 1863t vomiting in, 200 Gastroenteropathy. See also Enteropathy. allergic, 440 protein-losing. See Protein-losing gastroenteropathy. Gastroepiploic arteries, 775 Gastroepiploic veins, 775 Gastroesophageal junction. See Esophagogastric junction. Gastroesophageal reflux disease, 705-726. See also Esophagitis, reflux. acid reflux in mechanisms of, 708-710, 709f-710f prevalence of, 705 acid suppression trial in, 715-716 age and, 705, 706f asthma in, 714 asymptomatic, 714 after bariatric surgery, 117-118 barium esophagogram in, 719 Barrett’s esophagus and, 728, 728t, 730. See also Barrett’s esophagus. bleeding in, 720 chest pain in, 177-178, 714, 725 clinical course of, 719-720 clinical features of, 713-715 complications of, 720, 720f conditions associated with, 715 in cystic fibrosis, 945 in diabetes mellitus, 572 diagnosis of, 715-719, 716t differential diagnosis of, 715 dyspepsia in, 184 dysphagia in, 714, 726 ear, nose, and throat disease in, 714-715, 715f in eating disorders, 137 endoscopy in diagnostic, 716, 716t, 717f therapeutic, 725 versus eosinophilic esophagitis, 428, 431 epidemiology of, 705-707, 706f erosive, 720 in esophageal adenocarcinoma, 748 in esophageal atresia, 670 esophageal biopsy in, 716, 717f esophageal manometry in, 719 esophageal perforation in, 720 esophageal pH monitoring in, 717-719, 718f extraesophageal symptoms of, 179-181, 180f, 180t gastric motility disorders in, 809 genetics of, 707 health care impact of, 707 heartburn in, 705, 713, 726 Helicobacter pylori infection and, 838 impedance and acid testing in, 719 in irritable bowel syndrome, 2092 laryngitis in, 714-715, 715f mechanisms of, 708-710, 709f-710f nocturnal, 179 nonerosive (endoscopy-negative), 719-720 obesity and, 105, 706-707 odynophagia in, 176, 714 pathogenesis of, 707-713 acid clearance and, 710-713, 711f-712f antireflux barriers in, 707-708, 708f delayed gastric emptying in, 713 duodenogastric reflux in, 713 gastric acid secretion in, 713 gastric factors in, 713 hiatal hernia in, 710, 710f

Gastroesophageal reflux disease (Continued) hypotensive lower esophageal sphincter pressure in, 709-710, 710f swallow-induced lower esophageal sphincter relaxations in, 709, 709f transient lower esophageal sphincter relaxations in, 708, 709f-710f peptic strictures in, 720, 720f, 726 in pregnancy, 628-629, 715 prevalence of, 705-707, 706f regurgitation in, 714 in scleroderma, 715 sleep disturbances in, 725 after solid organ transplantation, 540-541 strain-induced reflux in, 709-710, 710f treatment of in Barrett’s esophagus, 730 for complications, 725-726 endoscopic, 725 histamine H2 receptor antagonists for, 721-723, 722f lifestyle modifications for, 721 maintenance therapy in, 723-724 nonprescription therapy for, 721 over-the-counter medications for, 721 prokinetic drugs for, 721-722 proton pump inhibitors, 715-716, 721-723, 722f surgery for, 723f, 724-725 for uncomplicated disease, 720-725 ulcers in, 720 water brash in, 714 in Zollinger-Ellison syndrome, 501-502, 502t, 715 Gastroesophageal varices in portal hypertension, 1494 in primary biliary cirrhosis, 1482 Gastrografin studies, in small intestinal obstruction, 2109-2110 Gastrography, computed tomography, in gastric cancer, 900 Gastrohepatic omentum, 612 Gastroileal (gastrocolic) reflex, 264 Gastrointestinal allergy, 143 Gastrointestinal bleeding, 285-322 in acute liver failure, 1561t, 1562, 1566 in acute pancreatitis, 982 in angioectasia, 593-594, 599 in chronic pancreatitis, 1011-1013, 1013f in Dieulafoy’s lesions, 602-604, 604f endoscopic hemostasis in, 291-292 endoscopy in, 289-291 after hematopoietic stem cell transplantation, 549-550, 550f history in, 285-286, 287t in HIV/AIDS, 531-532, 532t hospitalization for, 288 initial assessment and management of, 285-293, 286f, 289f-291f laboratory evaluation of, 288 localization of, 288 lower, 308-315 algorithm for, 289f-290f, 290-291 in anal fissures, 314 angiography in, 310 anoscopy in, 309 barium enema in, 311 causes of, 308, 308t, 311-315 in colitis, 312-313 in colon polyps and cancer, 313 in colonic angioectasias, 314 in colonic ulcer, 2244 colonoscopy in, 310 in colorectal cancer, 313, 2218, 2218f computed tomography in, 310 diagnostic approach to, 309-311 in Dieulafoy-type colonic ulcer, 2245 in diverticulosis, 311-312, 312f in hemorrhoids, 314 localization of, 308, 308f mortality in, 309 NSAIDS and, 309 after polypectomy, 313, 313f

Gastrointestinal bleeding (Continued) in radiation proctitis, 313-314, 314f in rectal Dieulafoy’s lesions, 314 in rectal ulcers, 314-315, 315f in rectal varices, 314 risk factors and risk stratification in, 309, 309t scintigraphy in, 310 sigmoidoscopy in, 309-310 in small intestinal tumors, 2148-2149 surgical management of, 311 medical therapy for, initial, 288-289, 289f obscure, 315-320 algorithm for, 291, 291f in angioectasia, 315-317, 317f in blue rubber bleb nevus syndrome, 317 causes of, 315-318, 315t definition of, 285 diagnostic testing in, 318-320 in Dieulafoy’s lesions of small intestine, 318 endoscopy in, 319-320 capsule, 319-320 deep, 320, 320t intraoperative, 319 in hereditary hemorrhagic telangiectasia, 317 in Meckel’s diverticulum, 317-318 NSAIDS and, 318 overt, approach to, 320 push enteroscopy in, 319 radiological studies in, 318-319 in small intestinal diverticula, 318 in small intestinal erosions and ulcers, 318 in small intestinal neoplasms, 318, 318f occult, 320-321 definition of, 285 fecal occult blood test in, 320-321 iron deficiency anemia in, 321. See also Iron deficiency anemia. palliative care for, 2286 in peptic ulcers. See Peptic ulcer bleeding. physical examination in, 286 in pseudoxanthoma elasticum, 362 radiologic imaging in, 292-293 resuscitation in, 288 as risk factor for spontaneous bacterial infection, 1531 severe, definition of, 285 site of, localization of, 288 after solid organ transplantation, 538t, 540, 543 surgical consultation in, 293 upper, 293-308 in aortoenteric fistula, 306 in Cameron’s lesions, 305, 305f causes of, 293, 293t, 294f in Dieulafoy’s lesions, 304 in endoscopic retrograde cholangiopancreatography, 660 endoscopy in, 293-294 epidemiology of, 293, 293t in esophagitis, 303 in gastric antral vascular ectasia, 305-306, 305f in gastroesophageal reflux disease, 720 in hemobilia, 306, 306f in hemosuccus pancreaticus, 306 in hospitalized patients, 303-304 in Mallory-Weiss tears, 304-305, 304f in peptic ulcer disease, 294-303. See also Peptic ulcer bleeding. after percutaneous transhepatic cholangiography, 1188 in portal hypertensive gastropathy, 306 risk factors and risk stratification in, 293, 294t after sphincterotomy, 306 in upper endoscopy, 657 in upper gastrointestinal cancer, 305 in varices, 306-308 Gastrointestinal cancer alarm symptoms of, 864t, 867, 867f complementary and alternative medicine for, 2297-2298 cutaneous manifestations of, 363-365

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Gastrointestinal cancer (Continued) metastatic, 564, 564f protein-losing gastroenteropathy in, 441 after solid organ transplantation, 538t, 543 upper, bleeding in, 305 Gastrointestinal disorders in celiac disease, 1804-1805 dyspepsia in, 184-185 in eating disorders, 128-131, 129t-130t, 135-137, 136t in HIV/AIDS, 523-524, 524f in refeeding syndrome, 71 Gastrointestinal fistula, 419-424 classification of, 419, 419t in Crohn’s disease, 1951 definition of, 419 diagnosis of, 420, 420f-421f management of, 420-424, 422t, 423f adjunctive therapies in, 423 in Crohn’s disease, 422-423 drainage in, 421 fluid and electrolyte therapy in, 420-421 nutritional therapy in, 421-422 somatostatin analogs in, 422 surgical, 423-424 nonhealing, conditions associated with, 420t outcome in, 424 pathophysiology of, 419-420, 420t Gastrointestinal fluids, electrolyte concentrations in, 71t Gastrointestinal function brain function and. See Brain-gut interactions. stress and, 342-343 Gastrointestinal hormones. See Hormone(s), gastrointestinal; specific hormone, e.g., Gastrin. Gastrointestinal ischemia, 2027-2048, 2028t. See also Intestinal ischemia; Mesenteric ischemia. Gastrointestinal lymphoma. See Lymphoma, gastrointestinal. Gastrointestinal polyposis, 2176-2188, 2177t inherited, 2176-2187, 2177t, 2181t, 2185t noninherited, 2188 Gastrointestinal stromal tumors (GISTs), 461-474, 904, 905f benign, 464 biopsy in, 468 in Carney triad, 473-474 CD117 in, 463 in children, 473 clinical practice guidelines for, 464-468 clinical presentation in, 462 colorectal, 465-466 computed tomography in, 466, 466f differential diagnosis of, 468 endoscopic ultrasonography in, 466 epithelioid, 463, 463f, 468 esophageal, 465, 768-769, 769f familial, 473 gastric, 465, 465f imaging studies in, 466-468, 466f-467f incidence of, 461 KIT in, 462-464, 463f location of, 461 magnetic resonance imaging in, 466 metastatic, 462, 468-472, 470t molecular pathogenesis of, 462-464, 463f nomenclature for, 462 pathology of, 462-464, 463f platelet-derived growth factor receptor-alpha in, 464 positron emission tomography in, 466-468, 467f postoperative, 471 as sarcoma, 462 site-specific diagnosis of, 465-466, 465f-467f small intestinal, 465 spindle cell, 463, 463f, 465f, 468 treatment of adjuvant, 473 in advanced disease, 468-472, 470t chemotherapy in, 468-469, 473

Gastrointestinal stromal tumors (Continued) imatinib mesylate in, 469-471, 470t, 473 in primary localized disease, 472-473 radiation therapy in, 469 sunitinib in, 471-472 surgical, 472-473 unresectable, 468-472, 470t Gastrointestinal tract cellular communication in, 3-5, 4f, 4t embryology of, 668, 669f fluid loads along, 212, 212f innervation of, 1617 neural regulation of, 5-6, 5f nonpeptide chemical messengers of, 10-13, 12f paraneoplastic syndromes affecting, 564 peptide hormones of, 6-10, 6f, 8f regulation of growth in, 15-17, 15f signal transduction in, 13-15, 13f, 14t Gastrointestinal tuberculosis, 1877-1878, 1877f-1878f. See also Tuberculosis, intestinal. Gastrojejunostomy, percutaneous, 92-93, 92f Gastroparesis, 806-809 in chronic pancreatitis, 1014-1015 in diabetes mellitus, 572-574, 806-807, 807f diet for, 814t dyspepsia and, 185 factors influencing, 1696, 1696f in functional dyspepsia, 809, 809t in gastroesophageal reflux disease, 713, 809 idiopathic, 807f, 808-809 in intestinal pseudo-obstruction, 809-810 in irritable bowel syndrome, 809 ischemic, 808 obstructive, 807f, 808 postsurgical, 807-808 postviral, 808-809 in scleroderma, 560 after solid organ transplantation, 540-541 symptoms of, 810 vomiting in, 201 Gastropathy. See also Gastritis. aging, 858 alcohol, 856-857, 857f bile reflux, 857-858 cocaine-induced, 857 differential diagnosis of, 859 hyperplastic, 858-859, 859f ischemic, 858 medication-induced, 856 in Ménétrier’s disease, 858-859, 859f portal hypertensive, 601-602, 603f, 857, 1514-1516 bleeding in, 306 diagnosis of, 1514, 1514f versus gastric vascular ectasia, 1515, 1515t management of, 1515-1516, 1515f prolapse, 858 radiation-induced, 857 reactive, 856-858, 856f-857f stress-related, 581, 857 toxin-induced, 856 treatment of, 859 uremic, 578 Gastroplasty, vertical banded, 115, 116t-117t. See also Bariatric surgery. Gastroschisis, 1628, 1628f Gastrostomy fluoroscopic percutaneous, 93 percutaneous endoscopic. See Percutaneous endoscopic gastrostomy. surgical, 93 Gaucher’s disease, gastrointestinal manifestations of, 577 Gaviscon, for gastroesophageal reflux disease, 721 GBV-C virus infection, 1343-1346, 1345f clinical features of, 1344 diagnosis of, 1344 epidemiology of, 1344 with HIV/AIDS, 1344-1346, 1345f treatment of, 1346 virology of, 1343-1344

Gemcitabine for cholangiocarcinoma, 1177 for pancreatic cancer, 1026-1027, 1026f Gemfibrozil, for nonalcoholic fatty liver disease, 1410 Gemtuzumab, hepatotoxicity of, 545 Gender alcoholic liver disease and, 1383, 1388 constipation and, 260 Crohn’s disease and, 1942 gallstone disease and, 1090-1091 gastric emptying and, 801 hepatitis C and, 1317, 1325 hepatocellular carcinoma and, 1570 irritable bowel syndrome and, 2093 Gene(s) allelic deletions in, 38, 39f loss of heterozygosity in, 38, 39f metastasis-associated, 43 neoplasia-associated, 35-41, 36t DNA mismatch repair genes as, 36t, 40 oncogenes as, 35-37, 36t tumor suppressor genes as, 36t-37t, 37-40, 38f-39f Gene expression hepatic, 1214-1216, 1215t nuclear receptors and, 1216 regulation of, 1215 Gene therapy for α1-antitrypsin deficiency, 1262 for glycogen storage disease type I, 1264 Genetic polymorphisms, in alcoholic liver disease, 1389 Genetic testing in celiac disease, 1809 of children, 951-952 in colorectal cancer, 2226 in familial adenomatous polyposis, 2182, 2226 in gastrointestinal cancers, 43-44, 43t-44t in hereditary nonpolyposis colorectal cancer, 2226 in pancreatitis, 951-952 Genitourinary system abnormalities of in cystic fibrosis, 948 in eating disorders, 129t-130t examination of, in acute abdominal pain, 156 Genomic instability, in oncogenesis, 34-35, 35f Geographic tongue, 354 GERD. See Gastroesophageal reflux disease. Germander, hepatotoxicity of, 1456-1457 Ghrelin, 10 in eating disorders, 123 in gastric acid secretion, 823 in hunger sensation, 802 in Prader-Willi syndrome, 10 regulation of satiety and hunger by, 18 tumors secreting, 515 for vomiting, 209 Giant hypertrophic gastropathy. See Ménétrier’s disease. Giardia intestinalis, 1911-1914, 1912f Giardiasis, 1911-1914 clinical presentation in, 1913, 1913t diagnosis of, 1912f, 1913 epidemiology of, 1911-1913, 1912f immune response in, 1913 malabsorption in, 1746-1747, 1747f, 1828-1829, 1829t pathogenesis of, 1913 prevention and control of, 1914 treatment of, 1914 vaccine for, 1914 Gilbert’s syndrome hyperbilirubinemia in, 1228-1229 jaundice in, 325, 326t Ginger for colonic health, 2294t for nausea and vomiting, 2288, 2289t Gingivitis, ulcerative, acute necrotizing, 357 Gingivostomatitis, primary herpetic, 356-357, 356t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xli

xlii

Index GISTs. See Gastrointestinal stromal tumors. Glasgow alcoholic hepatitis score, in alcoholic liver disease, 1393 Gliadin, in celiac disease, 1798, 1801-1802, 1808 Glibenclamide, hepatotoxicity of, 1435-1436 Glimepiride, hepatotoxicity of, 1435-1436 Glioma-polyposis, 2177t, 2184 Glissonian sheath, 1202-1203 Globus sensation, 176-177 pathophysiology of, 176-177 Glomerulonephritis, in hepatitis B, 1296 Glossitis, 353, 354f atrophic, 353 benign migratory, 354 median rhomboid, 353 taste disorders associated with, 354 Glossodynia, 353-354 Glottic spasm, with vomiting, 206 Glove powder granuloma, 617-618 Glucagon, 8 for foreign body removal, 400 pancreatic, 8 plasma, in glucagonoma, 508 post-translational modification of, 8, 8f Glucagon-like peptide(s) clinical applications of, 8 deficiency of, in short bowel syndrome, 1782 in digestion and absorption, 1697 production of, 8, 8f Glucagon-like peptide-1 in insulin release, 18, 19t in satiety, 18, 110, 802, 1696 Glucagon-like peptide 2, for short bowel syndrome, 1793 Glucagonoma, 505-508 clinical features of, 506-507, 506t, 507f definition of, 505-506 diagnosis of, 507-508 metastasis of, 506 necrolytic migratory erythema in, 365, 365f, 506-507, 507f pathophysiology and pathology of, 506 treatment of, 508 Glucoamylase, 1708, 1708f, 1708t Glucocorticoids for alcoholic liver disease, 1396-1397, 1396f-1397f for allergic eosinophilic esophagitis, 144 for ascariasis, 1923 for autoimmune hepatitis, 1467, 1471-1475, 1472t, 1474f for autoimmune pancreatitis, 991-992 for celiac disease, 1815, 1817 for Crohn’s disease, 1954, 1962t, 1963-1964 for eosinophilic gastrointestinal disorders, 434 for esophageal ulceration in HIV/AIDS, 526 for HELLP syndrome, 633-634 in intestinal ion transport, 1690 malabsorption with, 1757t in pregnancy, 630 for primary biliary cirrhosis, 1484 for primary sclerosing cholangitis, 1164-1165 side effects of, 1995t for typhoid fever, 1867 for ulcerative colitis, 1995-1996, 1995t for ulcerative enteritis, 2057 for vomiting, 208 withdrawal of, hepatitis B flares and, 1298 Glucokinase, in hepatic glucose metabolism, 1217 Glucose. See also Hyperglycemia; Hypoglycemia. in ascitic fluid, 1525 control of in glycogen storage disease type I, 1263 in intensive care unit, 49 during parenteral nutrition, 86 dietary, 1706 hepatic metabolism of, 1217-1220, 1218f intolerance of in glucagonoma, 507 in nonalcoholic fatty liver disease, 1402 in refeeding syndrome, 70-71

Glucose (Continued) malabsorption of, 1711 congenital, 1641 regulation of by gastrointestinal peptides, 18-19, 19t by hepatocytes, 1217-1219, 1218f transport of, 1710-1712, 1710f Glucose-6-phosphatase, 1217 deficiency of, 1262-1264 Glucose-6-phosphate formation of, 1217 metabolism of, 1217-1219 Glucose-6-phosphate translocase, deficiency of, 1262-1264 Glucose-dependent insulinotropic polypeptide (GIP), 8-9 in insulin release, 18, 19t receptors for, 8-9 Glucose-galactose malabsorption, 1759t-1762t, 1763 Glucose hydrogen breath test, for small intestinal bacterial overgrowth, 1776, 1776t Glucose transporter(s), 1217, 1681, 1681f Glucose transporter-2, 1711 deficiency of, 1711 Glucose transporter-5, 1711 Glutamine parenteral, 49-50 for short bowel syndrome, 74-75 synthesis of, hyperammonemia and, 1544 Glutaraldehyde, chemical colitis from, 2247, 2248f Glutathione in alcoholic liver disease, 1386 in drug-induced liver disease, 1419 reduced, secretion of, 1080 Glutathione peroxidase, 1419 Gluten, 1800-1801 hidden sources of, 1813t sensitivity to. See also Celiac disease. in dermatitis herpetiformis, 146 Gluten challenge, in celiac disease, 1810 Gluten-free diet in celiac disease, 84, 1813-1814, 1813t in dermatitis herpetiformis, 1812 in ulcerative enteritis, 2057 Glycerin suppository colonic ischemia from, 2040 for constipation, 277t, 280 Glycerophosphate pathway, in triglyceride digestion and absorption, 1704 Glycine, absorption of, 1713f Glycine-extended gastrin, 7 Glycogen dietary, 1707 as energy store, 47, 48t hepatic synthesis of, 1219 synthesis of, 1262, 1262f Glycogen-debranching enzyme, deficiency of, 1264 Glycogen storage disease, 1262-1265 type I, 1262-1264 type III, 1264 type IV, 1264-1265, 1265f Glycogenolysis, 1262, 1262f Glycolytic-gluconeogenic pathways, hepatic, 1219 Glycoprotein(s) in extracellular matrix, 1212 in gallstone disease, 1099 Glycoprotein-2, for short bowel syndrome, 83 Glycosaminoglycans, in extracellular matrix, 1212 Glycosphingolipidoses, neutral, gastrointestinal manifestations of, 577-578 Glycosylation, congenital disorders of, 12651266, 1265f Goblet cell(s), intestinal, 1617f-1618f, 1618 Goblet cell carcinoid tumors, appendiceal, 481, 481f Gold, gastrointestinal complications of, 559 Golgi complex in acinar cells, 913-914, 914f of hepatocytes, 1209 in pancreatic enzyme transport, 924

Gonococcal infection in Fitz-Hugh–Curtis syndrome, cholecystitis versus, 1115 hepatic manifestations of, 1352 Gottron’s papules, in dermatomyositis, 364, 364f Gr cells, 817-819 Graciloplasty, dynamic, for fecal incontinence, 255-256, 255t Graft nonfunction, primary, after liver transplantation, 1606-1607 Graft-versus-host disease gastritis in, 856 after hematopoietic stem cell transplantation cholestatic injury and, 548 diarrhea and, 550-552 gastrointestinal bleeding and, 549-550 intestinal, 552 liver disease and, 547t, 553-554 nausea, vomiting, and anorexia and, 545-546, 546f jaundice in, 328 after solid organ transplantation, 541 Grains, cereal, taxonomic relationships among, 1801, 1801f Gram stain, in ascitic fluid analysis, 1525-1526 Granisetron, for gastric motility disorders, 812, 813t Granular cell tumors biliary, 1183 esophageal, 769 Granulocyte colony-stimulating factor, for glycogen storage disease type I, 1264 Granulocyte count, in appendicitis, 2063 Granulocyte-macrophage colony-stimulating factor, recombinant, for Crohn’s disease, 1970 Granulocyte/monocyte apheresis, for ulcerative colitis, 2002 Granuloma in colonic tuberculosis, 1877, 1877f in Crohn’s disease, 1948, 1948f glove powder, 617-618 hepatic, 588, 589t drug-induced, 1437-1438, 1437t in HIV/AIDS, 533 in sarcoidosis, 588, 589f, 589t tuberculous, 1354 sarcoid, 588-590, 589f, 589t in schistosomiasis, 1936, 1937f Granulomatosis, Wegener’s, gastrointestinal manifestations of, 562 Granulomatous angiitis, allergic. See ChurgStrauss syndrome. Granulomatous cheilitis, 359 Granulomatous disease, chronic, malabsorption in, 1764 Granulomatous gastritis, 852-853, 852f Granulomatous liver disease, 1233, 1233t Green tea for gastric cancer prevention, 898 for gastrointestinal cancer, 2297 Grey Turner’s sign, in pancreatitis, 969, 970f GRF (growth hormone–releasing factor), 513 GRFoma, 513-514 clinical features of, 513-514 definition of, 513 diagnosis and differential diagnosis of, 514 pathophysiology and pathology of, 513 treatment of, 514 Groin hernia, 385-388. See also Femoral hernia. Growth factors. See also specific growth factors. in cellular proliferation, 32 in esophageal cancer, 749-750 in gastric cancer, 894 in gastrointestinal tract, 15-17 in hepatic regeneration, 1213-1214 peptide, oncogenes and, 36 receptors for, 15-16, 15f transcription factors and, 16 Growth hormone intestinal adaptation and, 1729 plasma, in GRFoma, 514 for short bowel syndrome, 1793

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Growth hormone–releasing factor (GRF), 513. See also GRFoma. Growth retardation, in celiac disease, 1806 Grynfeltt’s hernia, 392, 392f Guanosine monophosphate, cyclic, in intestinal ion transport, 1682, 1682f, 1693, 1693f Guanylate cyclase, receptor, 15 Guanylin, in intestinal ion transport, 1691 Guar gum, for constipation, 276 Gulf War syndrome, 345 Gum chewing, ileus and, 2127 Gut-associated lymphoid tissue (GALT). See also Immune system, mucosal. chemokines in, 29-30 components of, 445, 446f controlled/physiologic inflammation in, 21-22, 22f in food allergy, 139-142, 140t, 141f functional anatomy of, 25-26, 26f immunoglobulins of, 23-24, 24f physiology of, 24-25 Gut-brain interactions. See Brain-gut interactions. Gut-brain-liver axis, in glucose homeostasis, 1217 Gynecologic problems, in celiac disease, 1806

H

HAART. See Antiretroviral therapy, highly active (HAART). Hair bezoar of, 404-406 loss of, in Cronkhite-Canada syndrome, 364 in protein-energy malnutrition, 64 Hairy leukoplakia, in HIV/AIDS, 355, 355f HALO system, for Barrett’s esophagus, 731-732 Haloalkane compounds, hepatotoxicity of, 1450-1451, 1451t Halogenated aromatic compounds, hepatotoxicity of, 1452 Halothane hepatitis, 1447-1449, 1448t Hamartoma biliary, 1183 esophageal, 769 Hamartoma tumor syndromes, PTEN, 2181t, 2185t, 2187 Hamartomatous polyposis syndromes, 2148, 2185-2187, 2185t. See also individual syndromes, e.g., Peutz-Jeghers syndrome. Hamartomatous polyps, in tuberous sclerosis, 2186 Hannington-Kiff sign, in obturator hernia, 391 Harris-Benedict equation, for energy requirements, 78 Hartmann procedure, for diverticulitis, 2082 Hartmann’s pouch of gallbladder, 1048f, 1049 Hartnup’s disease, 1713-1714, 1717 malabsorption in, 1758-1763, 1759t-1762t Haustra, colonic, 1615-1616, 1660 loss of, in ulcerative colitis, 1987-1988, 1988f HCT. See Hematopoietic stem cell transplantation. Head injury, gastrointestinal manifestations of, 579 Health care seeking, in irritable bowel syndrome, 2094 Health care workers, hepatitis C in, 1318 Heart. See also Cardiac. Heart disease in acute pancreatitis, 977 carcinoid, 482-483, 483f in Crohn’s disease, 1955 in eating disorders, 127, 129t-130t endoscopy and, 654-655 gastrointestinal manifestations of, 584 in hereditary hemochromatosis, 1243 liver transplantation and, 1596 obesity and, 104 in refeeding syndrome, 70 Heart failure ascites in, 1517 in congestive hepatopathy, 1380 gastrointestinal manifestations of, 584 in infantile hemangioendothelioma, 1587

Heart transplantation complications of, 540-541 gallstone disease and, 1106 hepatitis C after, 1327 Heartburn (pyrosis), 178-179 in achalasia, 692 approach to, 179 versus dyspepsia, 187 functional dyspepsia and, 187 in gastroesophageal reflux disease, 713, 726 pathophysiology of, 179 in pregnancy, 628-629 prevalence of, 705 synonyms for, 178 Heavy chain disease, gastrointestinal manifestations of, 569 Hedgehog proteins, in pancreatic development, 916-917 Helicobacter pylori infection, 833-844 activating protein-1 in, 836 age and, 833 in aphthous ulcers, 356 in autoimmune metaplastic atrophic gastritis, 848 Barrett’s esophagus and, 838 cag pathogenicity island in, 835 CagA in, 836 carditis in, 848 chronic, natural history of, 889-890 colonization factors in, 834-835 conditions associated with, 838-839 cytokines in, 837 diagnostic testing for, 839-841 endoscopic, 839, 840f, 840t indications for, 839, 839t nonendoscopic, 840-841, 840t, 841f in duodenal ulcers, 838 in dyspepsia, test and treat strategy for, 190 environmental factors in, 834 in environmental metaplastic atrophic gastritis, 846-847 eosinophilic gastroenteritis with, 432 epidemiology of, 833-834 epithelial cell response to, 835-836 esophageal adenocarcinoma and, 749, 838 in functional dyspepsia, 188, 191, 192t gastric acid secretion and, 826-827, 826f, 836 in gastric cancer, 888-893, 890f gastric cancer and, 835, 838, 897-898 in gastric diffuse large B cell lymphoma, 451 in gastric MALT lymphoma, 448 gastric motility disorders and, 810 in gastric ulcers, 838 gastritis in, 838, 846, 846f-847f, 863 antrum-predominant, 863 lymphocytic, 854 pan-, 863 protein-losing gastroenteropathy in, 440 gastroesophageal reflux disease and, 838 genetic factors in, 834 hyperemesis gravidarum and, 628 immune response to, 835-838, 890, 893 Lewis antigens in, 834-835 major histocompatibility complex molecules in, 835 Ménétrier’s disease and, 440 nuclear factor kB in, 836 outer inflammatory protein A in, 836 oxidative stress in, 836 pathogenesis of, 834 peptic ulcer bleeding and, 294-295, 300, 303 in peptic ulcer disease. See Peptic ulcer disease, Helicobacter pylori–associated. in perforated peptic ulcers, 884 proton pump inhibitors and, 724 socioeconomic status and, 833-834 Toll-like receptors in, 835 transmission of, 834 treatment of, 841-843 antibiotic resistance and, 843 empirical, 866, 866f regimens for, 841-842, 842t

Helicobacter pylori infection (Continued) rescue, 842, 843t side effects of, 843 vacuolating cytotoxin in, 835, 838 virulence factors in, 834-835 Heliox, for pneumatosis cystoides intestinalis, 240 Heller myotomy for achalasia, 702 gastroesophageal reflux disease after, 715 HELLP syndrome, 632-634, 632t, 633f Helminthic infection. See also specific diseases and helminths. eosinophilia in, 431-432 hepatic, 1356t-1357t, 1359-1365 intestinal, 1921-1940 tropical malabsorption from, 1829-1830, 1829t Hemangioendothelioma epithelioid, hepatic, 1583 infantile, hepatic, 1587 Hemangioma, 604-605 capillary, 604 cavernous, 604, 605f hepatic, 1586-1587, 1586f rectal, 604, 605f colonic, 604 esophageal, 769-770 hepatic cavernous, 1586-1587, 1586f in pregnancy, 637 Hemangiomatosis, intestinal, diffuse, 605 Hemangiosarcoma, hepatic, 1582-1583 Hematemesis bleeding site in, 288 definition of, 285 Hematin, for porphyria, 1269 Hematochezia bleeding site in, 288 definition of, 285 severe algorithm for, 289f-290f, 290-291 causes of, 308t sources of, 308f Hematocrit in acute pancreatitis, 975 in gastrointestinal bleeding, 288 Hematologic malignancies, hepatic involvement in, 564-567, 565t Hematoma esophageal, spontaneous, 741 hepatic, in pregnancy, 634, 634f intramural bowel, in hemophilia, 569, 569f, 569t Hematopoietic stem cell transplantation abdominal pain after, 552, 553t anorexia after, 545-546, 546f, 555 bacterial infection after, 549 biliary obstruction after, 547t, 549 candidates for chronic liver disease in, 545 diarrhea in, 544 fungal liver infections in, 544 gallstones and bile duct stones in, 545 intestinal ulcers and tumors in, 544 iron overload in, 545 perianal pain in, 544 recent liver dysfunction in, 545 for celiac disease, 2057-2058 cholecystitis after, 549 cholestatic liver disease after, 547t, 548, 548f cirrhosis after, 554 complications of, 544-555 long-term, 553-555 pre-transplant evaluation for, 544-545 from transplant through day 200, 545-552 diarrhea after, 550-552, 551t, 555 donors for, viral hepatitis in, 544-545 drug-induced liver injury after, 547t, 554 dysphagia after, 550, 550f esophageal disorders after, 555 fungal liver abscess after, 547t, 549, 554 gastric antral vascular ectasia after, 549-550 gastrointestinal bleeding after, 549-550, 550f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xliii

xliv

Index Hematopoietic stem cell transplantation (Continued) gastrointestinal disorders after, 555 graft-versus-host disease after cholestatic injury and, 548 diarrhea and, 550-552 gastrointestinal bleeding and, 549-550 intestinal, 552 liver disease and, 547t, 553-554 nausea, vomiting, and anorexia and, 545-546, 546f hepatitis after acute, 547t, 548-549, 548f chronic, 547t, 554 hepatitis B after, 548-549, 554 hepatitis C after, 548-549, 554, 1327 hepatobiliary disorders after, 554-555 hyperammonemia and coma after, 549 iron overload after, 547t, 554 jaundice after, 546-549, 547t, 548f liver cancer after, 554 lymphoproliferative disorders after, 459, 549 nausea and vomiting after, 545-546, 546f, 555 pancreatic disorders after, 552, 555 perianal pain after, 552 sinusoidal obstruction syndrome after, 546-548, 547t, 548f, 1376 weight loss after, 555 Heme degradation of, to bilirubin, 323 hepatic synthesis of, 1266, 1266f iron transport by, 1724, 1724f Heme-carrier protein 1, 1724, 1724f Hemobilia, upper gastrointestinal bleeding in, 306, 306f Hemoccult test. See Fecal occult blood test (FOBT). Hemochromatosis cutaneous manifestations of, 365 hepatocellular carcinoma in, 1577 hereditary, 1239-1248 alcoholic liver disease versus, 1391 C282Y homozygosity in, 1242, 1243t, 1244 clinical features of, 1242-1243, 1243t computed tomography in, 1245-1246, 1246f diagnosis of, 1243-1246, 1244t, 1245f-1246f divalent metal ion transporter 1 in, 1724 family screening for, 1247 hepatic iron index in, 1245, 1245f hepatocellular carcinoma in, 1246 hepcidin in, 1240-1241, 1241f HFE mutation analysis in, 1244-1245, 1245f HFE-related, 1239 intestinal iron absorption in, 1240, 1241f iron-induced liver damage in, 1242 iron measurements in, 1244-1245, 1244t versus iron overload syndromes, 1239-1240, 1240t juvenile, 1239 liver biopsy in, 1244, 1245f liver transplantation for, 1246 non–HFE-related, 1239 pathophysiology of, 1240-1242, 1241f prognosis in, 1246 treatment of, 1246-1247, 1246t jaundice in, 327 liver transplantation for, 1603 portal hypertension in, 1498-1499 Hemoclips for gastrointestinal bleeding, 292 for peptic ulcer bleeding, 296-298, 298f-300f, 880 Hemodiadsorption systems, for acute liver failure, 1568 Hemodialysis chronic pancreatitis and, 994 hepatitis C and, 1318, 1326-1327 Hemoglobin breakdown of, 1227 fecal, tests for. See Fecal occult blood test (FOBT). Hemojuvelin gene, in hereditary hemochromatosis, 1241 Hemolysis, in HELLP syndrome, 632

Hemolytic anemia, salmonellosis and, 1862 Hemolytic uremic syndrome in enterohemorrhagic Escherichia coli infection, 1856-1857 gastrointestinal manifestations of, 569-570 pancreatitis in, in children, 939 Hemophagocytic syndrome, ascites in, 1528 Hemophilia, gastrointestinal manifestations of, 569, 569f, 569t Hemorrhage. See also Bleeding. in diverticulosis, 2085-2088, 2086f-2088f gastrointestinal. See Gastrointestinal bleeding. mesenteric, 620 retroperitoneal, 620 Hemorrhagic acute pancreatitis, 960-961 Hemorrhagic colitis, 2039 Hemorrhagic telangiectasia. See Telangiectasia, hereditary hemorrhagic. Hemorrhoid(s), 2260-2263 anticoagulation and, 2263 bleeding in, 314 cryotherapy for, 2261, 2263t diet therapy for, 2260, 2263t external, 2262-2263, 2264f-2265f in HIV/AIDS, 2261, 2263 internal, 2260-2262, 2262f, 2263t photocoagulation for, infrared, 2261, 2263t in pregnancy, 2263 procedure for prolapse of, 2262, 2262f, 2263t rubber band ligation for, 2261, 2263t sclerosing agents for, 2261, 2263t surgical management of, 2261-2262, 2262f, 2263t thrombosis of, 2262, 2264f Hemorrhoidal arteries, 1617, 2257 Hemorrhoidal plexus, 2257, 2258f Hemorrhoidectomy, 2262, 2263t fecal incontinence after, 244 Hemorrhoidopexy, stapled, 2262, 2262f, 2263t Hemosiderosis, after hematopoietic stem cell transplantation, 547t, 554 Hemostasis, endoscopic in angioectasia, 316 in colonic diverticulosis, 312 complications of, 657 in Dieulafoy’s lesions, 304 in gastric antral vascular ectasia, 305 in gastrointestinal bleeding, 291-292 in inpatient ulcer hemorrhage, 304 in Mallory-Weiss tear, 304-305 in peptic ulcer bleeding, 296-297, 297t. See also Peptic ulcer bleeding, endoscopic hemostasis in. Hemosuccus pancreaticus, bleeding in, 306 Henoch-Schönlein purpura cutaneous manifestations in, 360, 360f gastrointestinal manifestations of, 562 vasculitis in, 2047 Heparan sulfate proteoglycans, in protein-losing gastroenteropathy, 437-438, 439f Heparin low-molecular-weight for sinusoidal obstruction syndrome, 1377 for ulcerative colitis, 2000 for mesenteric venous thrombosis, 2037 for protein-losing gastroenteropathy, 443 for sinusoidal obstruction syndrome, 1377 for ulcerative colitis, 2000 Hepatectomy for gallbladder carcinoma, 1180-1181 native, in liver transplantation, 1604-1605 Hepatic. See also Liver. Hepatic arterioles, terminal, 1203 Hepatic artery, 1202-1203 anatomy of, 1489-1490 aneurysm of, 1381-1382 common, 2027, 2028f embolization of for carcinoid tumors, 488 for pancreatic endocrine tumors, 521 thrombosis of after liver transplantation, 1606-1607 after solid organ transplantation, 540

Hepatic duct. See also Bile duct(s). common, 1203 anatomy of, 1048, 1048f left, 1048, 1203 right, 1048, 1203 Hepatic encephalopathy, 1543-1546 in acute liver failure, 1561-1562, 1561t, 1565-1566 in alcoholic liver disease, 1390, 1390t, 1393, 1393f clinical features of, 1544-1545, 1545t diagnosis of, 1544-1545, 1545t minimal, 1544-1545 palliative care for, 2280, 2286 pathophysiology of, 1544, 1550f scoring system for, 1544-1545, 1545t treatment of, 1545-1546 Hepatic fibrosis in alcoholic liver disease, 1389 congenital, portal hypertension in, 1500 in congestive hepatopathy, 1380, 1381f drug-induced, 1441-1444 in hepatitis C progression to, 1324-1327, 1324f, 1325t staging of, 1322-1324, 1322f-1323f, 1322t, 1324t liver biochemical tests in, 1234-1235 in nonalcoholic fatty liver disease, 1404, 1408 in primary biliary cirrhosis, 1480-1481, 1481f stellate cell activation in, 1211 Hepatic growth factor, in hepatic regeneration, 1214 Hepatic hydrothorax in cirrhotic ascites, 1535 liver transplantation and, 1597 Hepatic iron index, in hereditary hemochromatosis, 1245, 1245f Hepatic triglyceride lipase, 1222 Hepatic vein, 1202-1203 thrombosis of. See Budd-Chiari syndrome. Hepatic vein pressure gradient, 1495, 1496t Hepatic venous hormone sampling, in pancreatic endocrine tumors, 517-519, 519f Hepatic venule, terminal, obstruction of, 1375-1377. See also Sinusoidal obstruction syndrome. Hepaticojejunostomy, for recurrent pyogenic cholangitis, 1169 Hepatis peliosis, 1381, 1381f Hepatitis alcoholic cholestasis in, 329 histology of, 1383-1384, 1384f jaundice in, 327 versus nonalcoholic fatty liver disease, 1408-1409 optimal management of, 1399 prognosis in, 1393, 1393f autoimmune, 1461-1476 antibodies in, 1461, 1470 asymptomatic, 1471 autoantibody-negative, 1467, 1467t with cholestasis, 1467, 1467t classification of, 1465-1466, 1466t clinical features of, 1470-1471, 1470t diagnosis of, 1461-1463, 1462f, 1463t in elderly, 1471 epidemiology of, 1468 ethnicity and, 1469-1470 fulminant, 1471 after hepatitis A, 1282 with hepatitis C, 1467-1468, 1468f hereditary, 1471 histologic findings in, 1461, 1462f, 1469, 1469t HLA (human leukocyte antigen) phenotypes in, 1465, 1469, 1469t immune diseases associated with, 1471 laboratory indices for, 1468-1469, 1469t MELD score in, 1470 panacinar or lobular, 1461, 1462f pathogenesis of, 1463-1465, 1464f portal hypertension in, 1498-1499

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Hepatitis (Continued) with primary biliary cirrhosis, 1466, 1467t versus primary sclerosing cholangitis, 1057, 1155 with primary sclerosing cholangitis, 1466-1467, 1467t prognostic indices for, 1468-1470, 1469t relapse of, 1475 scoring system for, 1462-1463, 1463t serologic markers of, 1470 severe, 1471 susceptibility alleles in, 1463, 1464f treatment of, 1471-1475 drug actions in, 1471-1472 drug toxicity and, 1474-1475 end points for, 1473-1474, 1474f failure of, 1473-1475 incomplete response to, 1474-1475 indications for, 1471, 1472t liver transplantation for, 1475-1476, 1603 regimens for, 1471, 1472t results of, 1474-1475 side effects of, 1472-1473, 1475 type 1, 1465, 1466t type 2, 1465-1466, 1466t variant forms of, 1466-1467, 1467t versus Wilson disease, 1251 versus cholecystitis, 1115 cholestatic from amoxicillin-clavulanic acid, 1440 with bile duct injury, 1441 from chlorpromazine, 1440 drug-induced, 1440 chronic from diclofenac, 1438t, 1439 from minocycline, 1438t, 1439 cytomegalovirus, 1347-1348, 1348f-1349f drug-induced acute, 1431-1437, 1431t chronic, 1438-1439, 1438t gender and, 1415-1416 granulomatous, 1437-1438, 1437t Epstein-Barr virus, 1347 GB agent, 1343-1346, 1345f. See also GBV-C virus infection. after hematopoietic stem cell transplantation acute, 547t, 548-549, 548f chronic, 547t, 554 in hematopoietic stem cell transplantation donors, 544-545 herpes simplex virus, 1348-1349, 1349f hypoxic, 1379 interface, 1461, 1462f, 1480-1481 ischemic, 1379-1380, 1380f, 1450t jaundice in, 327 rheumatoid arthritis with, 559 Sanban virus, 1346-1347 SARS, 1349 SEN virus, 1346-1347 syphilitic, 1354 TT virus, 1346 TTV-like mini-virus, 1346-1347 varicella-zoster virus, 1349 viral, 1233-1234 cholestasis in, 329 in hematopoietic stem cell transplantation donors, 544-545 jaundice in, 326-327 Yonban virus, 1346-1347 Hepatitis A virus, 1279-1280, 1280f Hepatitis A virus infection, 1279-1286 autoimmune hepatitis after, 1282 cell surface receptor for, 1279 clinical features of, 1281-1282 diagnosis of, 1281t, 1282-1283, 1283f epidemiology of, 1280-1281 extrahepatic manifestations of, 1282 fulminant, 1280, 1282, 1560 in HIV/AIDS, 533 immunization against, 1283-1285, 1283t-1284t in chronic liver disease, 1284-1285 liver failure in, 1558t-1559t, 1559f, 1560 pathogenesis of, 1281

Hepatitis A virus infection (Continued) prevention of, 1283-1285, 1283t-1284t prodromal symptoms of, 1282 relapsing, 1282 risk factors for, 1280 transmission of, 1280-1281 virology of, 1279-1280, 1280f Hepatitis B immunoglobulin, 1307, 1309t in liver transplantation, 1599-1600 Hepatitis B vaccine, 1307-1309, 1308t-1309t hepatocellular carcinoma and, 1575-1576, 1579-1580 Hepatitis B virus, 1289-1292 basal core promoter mutation in, 1291 carcinogenicity of, 1575-1576 core mutation in, 1291 DNA polymerase mutation in, 1291-1292, 1292f drug-resistant, 1291-1292, 1292f genome of, 1289, 1289f genotypes of, 1290, 1290t life cycle of, 1289-1290, 1290f mutations in, 1291-1292, 1292f in polyarteritis nodosa, 562 precore mutation in, 1291 replication of, 1289 in sickle cell anemia, 570-571 Hepatitis B virus infection, 1287-1309 active carriers of, 1288 acute clinical features of, 1294-1295 clinical sequelae of, 1288 treatment of, 1306 alanine aminotransferase (ALT) in, 1294-1295 anti-HBe in, 1293, 1293f, 1298 arthritis-dermatitis in, 1295 aspartate aminotransferase (AST) in, 1295 chronic acute flares in, 1292t, 1297-1298 antiviral therapy–induced, 1297-1298 genotypic variation and, 1298 hepatotropic virus superinfection and, 1298 immunosuppressive therapy–induced, 1297 spontaneous, 1297 clinical features of, 1295-1297 clinical sequelae of, 1288-1289 histopathology of, 1296-1297, 1296f cirrhosis in, 1288, 1293-1294, 1306 clinical features of, 1294-1298 clinical sequelae of, 1288-1289 cryoglobulinemia in, 1296-1297 cutaneous manifestations of, 366 definitions in, 1288 diagnosis of, 1298-1300 drug-induced liver disease in, 1417, 1425-1426 drug-resistant, 1304-1305, 1305f epidemiology of, 1287-1288 escape mutants in, 1309 extrahepatic manifestations of, 1295-1297 fulminant, 1288, 1295 geographic distribution of, 1287-1288 glomerulonephritis in, 1296 HBeAg in, 1290-1291, 1293, 1293f, 1299 HBsAg in, 1290-1291, 1295-1298 HBV DNA in, 1299 complications and, 1294 with HCV and HDV infection, 1311 with HCV infection, 1298, 1307 with HDV infection, 1298, 1309-1312. See also Hepatitis D virus infection. coprimary, 1311, 1311f superinfection, 1311, 1311f in hematopoietic stem cell transplant donors, 544-545 after hematopoietic stem cell transplantation, 548-549, 554 hepatic decompensation in, 1294 hepatocellular carcinoma in, 1288, 1294, 1575-1577 histopathology of, 1296-1297, 1296f in HIV/AIDS, 533, 1306-1307

Hepatitis B virus infection (Continued) immune clearance phase of, 1293, 1293f immune response to, 1292-1293 immune tolerance phase of, 1293, 1293f immunization against, 1307-1309, 1308t-1309t in liver transplantation, 1599-1600 inactive carriers of, 1288, 1293, 1293f incidence and prevalence of, 1288 liver failure in, 1558t-1559t, 1559f, 1560 molecular biology of, 1289-1292, 1289f-1290f, 1290t, 1292f natural history of, 1293-1294, 1293f pathogenesis of, 1292-1293 perinatal prophylaxis of, 1309, 1309t polyarteritis nodosa in, 1295-1296 postexposure prophylaxis of, 1309, 1309t precore mutation in, 1298 in pregnancy, 636, 1306 prevention of, 1307-1309, 1308t-1309t reactivation of, 1293, 1293f transmission of, 1287 treatment of, 1300-1307 antiviral agents for, 1302-1307 hepatocellular carcinoma and, 1580 choice of agents for, 1300, 1300t combination interferon and nucleoside analogues for, 1305-1306 disease flares and, 1297-1298 guidelines for, 1300-1302, 1301t interferon therapy for, 1300t, 1302 liver transplantation for, 1599-1600 nucleoside and nucleotide analogues for, 1300t, 1301-1305, 1305f combination, 1304-1305 resistance to, 1304-1305, 1305f in special populations, 1306-1307 in pregnancy, 1306 response to, definitions of, 1300, 1300t unresolved issues in, 1307 virologic endpoints in, 1300, 1300t Hepatitis C virus, 1313-1317 core protein of, 1315-1316 E1/E2 proteins of, 1316 genomic organization of, 1314 genotypes of, 1316-1317 life cycle and replication of, 1314-1316, 1314f-1315f NS proteins of, 1316 P7 protein of, 1316 polyprotein of, 1315-1316, 1315f quasi-species of, 1317 structure of, 1313 Hepatitis C virus infection, 1313-1336 acute, 1319-1321, 1320f, 1320t, 1327-1328 alanine aminotransferase (ALT) in, 1320, 1320f, 1326 alcoholic liver disease and, 1392, 1392f anti-HCV in, 1319-1320, 1320f autoantibodies in, 1321 autoimmune hepatitis with, 1467-1468, 1468f biopsy in, 1322-1324, 1322f-1323f, 1322t from breast-feeding, 1318 chronic, clinical features of, 1320-1321, 1320f, 1320t clinical features of, 1319-1321, 1320f cryoglobulinemia in, 1320 cutaneous manifestations of, 366, 366f diabetes mellitus in, 1326 diagnosis of, 1321-1322 in blood banks, 1322 direct assays for, 1321-1322 genotypes in, 1321-1322 indirect assays for, 1321 after known exposure, 1322 test selection in, 1322 drug-induced liver disease in, 1417, 1425-1426 epidemiology of, 1317-1318 extrahepatic manifestations of, 1320-1321, 1320t fibrosing cholestatic, 1319 fibrosis in progression to, 1324-1327, 1324f, 1325t staging of, 1322-1324, 1322f-1323f, 1322t, 1324t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xlv

xlvi

Index Hepatitis C virus infection (Continued) fulminant, 1325 with HBV infection, 1298, 1307 HCV RNA in, 1319-1320, 1320f in health care workers, 1318 in hematopoietic stem cell transplant donors, 544-545 after hematopoietic stem cell transplantation, 548-549, 554 in hemodialysis patients, 1318, 1326-1327 hepatitis A prophylaxis in, 1284-1285 hepatocellular carcinoma in, 1325, 1577 with HIV/AIDS, 533-534, 1327, 1333-1334 immune response in, 1319 immunization against, 1328 in immunocompromised patients, 1326-1327 incidence and prevalence of, 1317 in injection drug users, 1317-1318 insulin resistance in, 1321, 1326 after kidney transplantation, 1326-1327 after liver transplantation, 1326, 1334, 1600-1601, 1600t, 1601f, 1607-1608, 1609t lymphoma in, 1321 natural history of, 1324-1327, 1324f, 1325t pathogenesis of, 1318-1319 perinatal, 1318 porphyria cutanea tarda and, 1267-1268 postexposure prophylaxis of, 1327 in pregnancy, 636-637 prevention of, 1327-1328 progression of, 1324-1327, 1324f, 1325t factors influencing, 1325-1326, 1325t in immunocompromised patients, 1326-1327 in patients with normal ALT level, 1326 recurrence of, after liver transplantation, 1600-1601, 1600t, 1601f screening for, 1328 sexual transmission of, 1318 in sickle cell anemia, 570-571 after solid organ transplantation, 539 sporadic, 1318 transfusion-related, 1317 transmission of, 1317-1318 transplant-related, 1326 treatment of, 1328-1335 algorithm for, 1330f antiviral agents for, 1329 hepatocellular carcinoma and, 1580 contraindications to, 1332, 1332t efficacy of, 1329-1331, 1329f end points of, 1328 future directions in, 1334-1335 genotype and, 1330-1331 goals of, 1328 in HIV/AIDS, 1333-1334 indications for, 1331-1332 in liver transplant recipients, 1334 liver transplantation for, 1600-1601, 1600t, 1601f monitoring during, 1332-1333 for nonresponders, 1333, 1333f response to, predictive factors in, 1331 safety of, 1332-1333 virology of, 1313-1317, 1314f-1315f Hepatitis D virus, 1310 genotypes of, 1310 with HBV infection, 1298, 1309-1312 Hepatitis D virus infection, 1309-1312 anti-HDV in, 1310 clinical features of, 1311 diagnosis of, 1310 epidemiology of, 1309-1310 with HBV and HCV infection, 1311 with HBV coprimary infection, 1311, 1311f with HBV superinfection, 1311, 1311f HDAg in, 1310 HDV RNA in, 1310 in HIV/AIDS, 533 natural history of, 1310-1311, 1311f pathogenesis of, 1310 in pregnancy, 636 prevention of, 1312

Hepatitis D virus infection (Continued) treatment of, 1311-1312 virology of, 1310 Hepatitis E virus, 1337, 1338f, 1338t Hepatitis E virus infection alanine aminotransferase (ALT) in, 1340, 1340f clinical features of, 1340-1341, 1340t diagnosis of, 1341 epidemiology of, 1337-1340, 1339f, 1339t HEV Ag in, 1340, 1340f HEV RNA in, 1340, 1340f histopathology of, 1340 immunization against, 1341 pathogenesis of, 1340, 1340f in pregnancy, 636, 1341 prevention of, 1341-1342 seroepidemiology of, 1339-1340 transmission of, 1338-1339 treatment of, 1341-1342 virology of, 1337, 1338f, 1338t Hepatitis G virus infection, 1343-1346, 1345f. See also GBV-C virus infection. Hepatobiliary disorders in diabetes mellitus, 575 after hematopoietic stem cell transplantation, 554-555 after solid organ transplantation, 538t, 543-544 in ulcerative colitis, 2010t, 2012 Hepatobiliary scintigraphy. See Cholescintigraphy. Hepatoblast, 1201 proliferation of, 1045-1046 Hepatoblastoma, 1582 liver transplantation for, 1602 Hepatocavopathy, obliterative, 1372 Hepatocellular adenoma, 1584-1586 clinical features of, 1585 diagnosis of, 1585 epidemiology and pathogenesis of, 1584-1585, 1585f in glycogen storage disease type I, 1263 liver transplantation for, 1603-1604 pathology of, 1585-1586, 1585f in pregnancy, 637 treatment and prognosis in, 1586 Hepatocellular carcinoma, 1569-1580 abdominal pain in, 1570 aflatoxin B1 and, 1577 age and, 1570 versus alcoholic liver disease, 1392 alpha fetoprotein in, 1571-1572, 1572t angiography in, 1573 ascites in, 1517-1518, 1526, 1570-1571 in autoimmune hepatitis, 1469 in cirrhosis, 1577 clinical features of, 1570-1571, 1570t computed tomography in, 1573, 1573f des-γ-carboxy prothrombin in, 1572, 1572t diagnosis of, 1571-1572 diffuse, 1574 epidemiology of, 1569-1570, 1570f fibrolamellar, 1575, 1602 in pregnancy, 637 fucosylated alpha fetoprotein in, 1572, 1572t gender and, 1570 geographic distribution of, 1569, 1570f gross appearance of, 1574 after hematopoietic stem cell transplantation, 554 in hemochromatosis, 1246, 1577 in hepatitis B, 1288, 1294, 1575-1577 in hepatitis C, 1325, 1577 in HIV/AIDS, 1577 hypercalcemia in, 1571 hypoglycemia in, 1571 imaging studies in, 1572-1573, 1573f-1574f laparoscopy in, 1573 magnetic resonance imaging in, 1573, 1574f massive, 1574 MELD score in, 1602 in membranous obstruction of inferior vena cava, 1577 metastasis of, 1575

Hepatocellular carcinoma (Continued) microscopic appearance of, 1574-1575 moderately differentiated, 1574-1575 molecular pathways in, 1577, 1577t natural history of, 1578 nodular, 1574 oral contraceptives and, 1577 paraneoplastic manifestations of, 1571 pathogenesis of, 1575-1577, 1575t pathology of, 1574-1575 pityriasis rotunda in, 1571 polycythemia in, 1571 in pregnancy, 637 prevention of, 1579-1580 progenitor cell, 1575 rash in, 1571 risk factors for, 1575-1577, 1575t screening for, 1579, 1580t serum tumor markers of, 1571-1572, 1572t smoking and, 1577 staging of, 1575, 1576f trabecular, 1574 treatment of, 1576f, 1578-1579, 1578t chemoembolization for, 1579 chemotherapy for, 1579 liver transplantation for, 1578-1579, 1596, 1602 local ablation techniques for, 1579 percutaneous ethanol injection for, 1579 radiofrequency ablation for, 1579 surgical resection for, 1578 ultrasonography in, 1572-1573 undifferentiated, 1575 well-differentiated, 1574 in Wilson disease, 1577 Hepatocyte nuclear factor-6, 1201-1202 Hepatocytes, 1207-1210, 1208f cell junctions of, 1208 copper transport by, 1249, 1250f cytoskeleton of, 1208 differentiation of, 1201 biliary, 1201-1202 endocytosis in, 1209-1210 endoplasmic reticulum of, 1209 exocytosis in, 1209-1210 fat accumulation within. See Steatosis, hepatic. glucose regulation by, 1217-1219, 1218f Golgi complex of, 1209 ground-glass, in hepatitis B, 1296-1297, 1296f lysosomes of, 1209 mitochondria of, 1209 nucleus of, 1208-1209 peroxisomes of, 1209 plasma membrane of, 1207-1208 regenerative response of, 1212-1214, 1213f transplantation of for acute liver failure, 1568 for glycogen storage disease type I, 1264 transport proteins of, 1209 Hepatoerythropoietic porphyria, 363, 1267t, 1268 Hepatolithiasis, 1167-1170, 1168f Hepatomegaly in alcoholic liver disease, 1390, 1390t in glycogen storage disease type I, 1263 in HIV/AIDS, 532-535, 532t in leukemia, 567 in nonalcoholic fatty liver disease, 1405 Hepatopathy, congestive, 1380, 1381f Hepatoportal sclerosis, 1500 Hepatoportoenterostomy, Kasai for biliary atresia, 1054-1055, 1055f outcome after, 1056 Hepatopulmonary syndrome, 1549-1552 clinical features and diagnosis of, 1550-1551, 1551f liver transplantation for, 1596-1597 pathophysiology of, 1549-1550, 1550f screening for, 1551, 1551f treatment of, 1551-1552 Hepatorenal syndrome, 1546-1549 in alcoholic liver disease, 1393, 1393f cardiac output in, 1547 classification of, 1547-1548

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Hepatorenal syndrome (Continued) diagnosis of, 1547, 1547t pathophysiology of, 1546-1547, 1547f prevention of, 1548 prognosis in, 2280 renal arterial vasoconstriction in, 1546-1547, 1547f splanchnic arterial vasodilatation in, 1546, 1547f treatment of, 1548-1549, 1548t type 1, 1547-1548 type 2, 1548 Hepatotoxicity. See also Liver disease, drug-induced. of acetaminophen, 1427-1429, 1427t, 1428f of adulterated cooking oils, 1453-1454 of amiodarone, 1442 of amoxicillin-clavulanic acid, 1440 of anabolic steroids, 1440 of anesthetic agents, 1447-1449, 1448t-1450t of antidepressants, 1436 of antidiabetic drugs, 1435-1436 of antifungal agents, 1434-1435 of antiretroviral therapy, 1429-1430 of antituberculosis drugs, 1434-1436 of aspirin, 1430-1431 of azathioprine, 1445 of black cohosh, 1458 of carbon tetrachloride, 1450-1451, 1451t of chaparral, 1457 of chemical agents, 1450-1453, 1450t-1451t of Chinese herbal medications, 1457-1458 of chlorpromazine, 1440 of contaminated food, 1453-1454 of copper, 1453 of desflurane, 1449, 1449t of dextropropoxyphene, 1441 of diclofenac, 1438t, 1439 of drugs of abuse, 1454 of flucloxacillin, 1441 of foodstuffs, 1454-1455, 1457t of germander, 1456-1457 of HAART, 532, 1429-1430 of haloalkane compounds, 1450-1451, 1451t of halogenated aromatic compounds, 1452 of halothane, 1447-1449, 1448t of herbal medicines, 1456-1458, 1457t of Herbalife, 1458 of iron, 1453 of isoflurane, 1449, 1449t of isoniazid, 1434 of kava kava, 1458 of metals, 1453 of methotrexate, 1443-1444, 1443t of methoxyflurane, 1449, 1449t of minocycline, 1438t, 1439 of monoamine oxidase inhibitors, 1436 of mushrooms, 1450t, 1454-1455, 1457t of neurologic drugs, 1436-1437 of niacin, 1429 of nicotinic acid, 1429 of nitroaliphatic compounds, 1452 of nitroaromatic compounds, 1452 of nitrofurantoin, 1431-1432 of non-nucleoside reverse transcriptase inhibitors, 1430 of nonhalogenated organic compounds, 1452 of nonsteroidal anti-inflammatory drugs, 1437 of nucleoside (or nucleotide) reverse transcriptase inhibitors, 1430 of nutritional supplements, 1456-1458, 1457t of oral contraceptives, 1440 of pennyroyal, 1457 of pesticides, 1452-1453 of phosphorus, 1450t, 1453 of polychlorinated biphenyls, 1452 of protease inhibitors, 1430 of pyrrolizidine, 1456 of selective serotonin reuptake inhibitors, 1436 of sevoflurane, 1449, 1449t of tamoxifen, 1442-1443 of tetracycline, 1429 of thiazolidinediones, 1435 of thorium dioxide, 1453

Hepatotoxicity (Continued) of tricyclic antidepressants, 1436 of trinitrotoluene, 1452 of valproic acid, 1429 of vinyl chloride, 1451-1452 of vitamin A, 1455-1456 of weight loss products, 1458 Hepcidin, in hereditary hemochromatosis, 1240-1241, 1241f Hephaestin, 1724-1725 in hereditary hemochromatosis, 1240 HER2/neu gene, in esophageal cancer, 749-750 Herbal medicines, 2288t Chinese hepatotoxicity of, 1457-1458 for irritable bowel syndrome, 2103, 2291, 2292t for liver disease, 2295t, 2296-2297 for colonic health, 2293-2295, 2294t for diarrhea, 227 for functional dyspepsia, 193, 2289-2291, 2290t hepatotoxicity of, 1456-1458, 1457t for liver disease, 2295-2297, 2295t safety and regulation of, 2298-2299 Herbalife, hepatotoxicity of, 1458 Herbicides, hepatotoxicity of, 1452-1453 Hereditary hemorrhagic telangiectasia. See Telangiectasia, hereditary hemorrhagic. Hereditary nonpolyposis colorectal cancer (HNPCC) syndrome, 2204-2207, 2206t-2207t, 2207f Bethesda guidelines for, 2206t clinical features of, 2206, 2207t colon carcinogenesis in, 2160-2161 definition of, 2206, 2206t DNA mismatch repair genes in, 2206 gastric cancer in, 893 genetic testing in, 2226 microsatellite instability in, 40, 43-44, 2200t, 2203, 2203f, 2204t, 2206 small intestinal tumors in, 2148 type a (Lynch’s syndrome I), 2206 type b (Lynch’s syndrome II), 2206, 2207f Hernia, 379-396 abdominal wall, 379 in ascites, 1535 Bochdalek clinical manifestations and diagnosis of, 381 etiology and pathogenesis of, 379, 380f incidence and prevalence of, 380-381 treatment and prognosis in, 383 diaphragmatic, 379-383 clinical manifestations and diagnosis of, 381-382, 381f-382f congenital clinical manifestations and diagnosis of, 381, 382f etiology and pathogenesis of, 379, 380f incidence and prevalence of, 380-381 treatment and prognosis in, 383 etiology and pathogenesis of, 379-380, 380f gastric volvulus with, 383-384, 385f incidence and prevalence of, 380-381 mixed clinical manifestations and diagnosis of, 381 etiology and pathogenesis of, 379, 380f incidence and prevalence of, 380 post-traumatic, 379-383 treatment and prognosis in, 382-383 vomiting in, 199-200 epigastric, 390 femoral, 385-388 clinical manifestations and diagnosis of, 386-387, 387f complications and recurrence of, 388 etiology and pathogenesis of, 385-386, 386f incidence and prevalence of, 386 treatment and prognosis in, 387-388 foramen of Winslow, 392-395 Grynfeltt’s, 392, 392f

Hernia (Continued) hiatal in gastroesophageal reflux disease, 710, 710f Mallory-Weiss tear and, 740 nonreducing, 710 sliding Cameron ulcer in, 381, 381f, 383 clinical manifestations and diagnosis of, 381, 381f etiology and pathogenesis of, 379 incidence and prevalence of, 380 treatment of, 382-383 incarcerated, 379 incisional, 389-390 clinical manifestations and diagnosis of, 389 etiology and pathogenesis of, 389 incidence and prevalence of, 389 treatment and prognosis in, 389-390 inguinal, 385-388 and benign prostatic hypertrophy, 388 clinical manifestations and diagnosis of, 386-387, 387f and colorectal cancer screening, 388 complications and recurrence of, 388 direct versus indirect, 386, 386f etiology and pathogenesis of, 385-386, 386f incidence and prevalence of, 386 treatment and prognosis in, 387-388 internal, 392-395 clinical manifestations and diagnosis of, 394-395 etiology and pathogenesis of, 392-394, 393f incidence and prevalence of, 394 treatment and prognosis in, 395 lumbar, 392, 392f mesenteric, 392-395, 393f Morgagni clinical manifestations and diagnosis of, 381, 382f etiology and pathogenesis of, 379, 380f incidence and prevalence of, 380-381 treatment and prognosis in, 383 obturator, 391-392 paraduodenal, 392-395 paraesophageal clinical manifestations and diagnosis of, 381 etiology and pathogenesis of, 379, 380f gastric volvulus with, 383-384, 385f incidence and prevalence of, 380 treatment of, 382-383 parastomal, in Brooke ileostomy, 2017 pelvic and perineal, 391-392 Petit’s, 392, 392f retroanastomotic, 393-395 Richter’s, 379, 386 sciatic foramen, 391-392 small intestinal obstruction in, 2105-2106 Spigelian, 386f, 390-391 strangulated, 379 supravesicular, 392-395 trocar site, 389, 2106 umbilical, 390 congenital, 1626-1628, 1627f physiologic, 1624-1625, 1625f Herpes simplex virus infection in achalasia, 690 esophagitis in, 742 in HIV/AIDS, 525, 525f gastritis in, 849 gingivostomatitis in, 356-357, 356t hepatitis in, 1348-1349, 1349f in HIV/AIDS, 357, 525, 525f orolabial ulcers in, 356t, 357 in pregnancy, 636 after solid organ transplantation, 539, 539f Herpes zoster infection gastritis in, 849 hepatitis in, 1349 mucocutaneous ulcers in, 357 HES genes, in pancreatic development, 917 Heterophyes spp., 1933-1934

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xlvii

xlviii

Index Heterotopia gallbladder, 1183-1184 gastric, 670t, 675, 675f Heterozygosity, loss of, 38, 39f Hexachlorobenzene, hepatotoxicity of, 1452, 1454 Heyde’s syndrome, 316, 595 HFE gene in hereditary hemochromatosis, 1239, 1244-1245, 1245f in porphyria cutanea tarda, 1267-1268 HFE protein, in hereditary hemochromatosis, 1241-1242, 1242f Hiatal hernia. See Hernia, hiatal. Hibiscus, for colonic health, 2294t Hiccups, 177, 621-622 Hidradenitis suppurativa, 2273, 2273f High-density lipoprotein (HDL), 1221 ApoB-containing, 1224 gallstone disease and, 1092 receptors for, 1224 Hilar plate, 1048, 1048f Hindgut, 1624f, 1625 Hip fracture, proton pump inhibitors and, 724 Hirschsprung’s disease, 1636-1639 clinical features of, 1637-1638 congenital anomalies and syndromes associated with, 1637, 1637t constipation in, 269-270 diagnosis of, 1638-1639, 1638f genetics of, 1637, 1637t intestinal pseudo-obstruction in, 2134, 2144 management of, 1639 meconium plug and, 1639 pathogenesis of, 1636-1637, 1637t with Waardenburg syndrome, 2134 Histamine, 11-12 in gastric acid secretion, 821, 821f-822f gastrin effects on, 821-822 in gastrointestinal tract, 11-12 receptors for, 11-12 Histamine H1 receptor, in emetic reflex, 197-198 Histamine H2 receptor antagonists adverse effects of, 870 drug interactions with, 870 for functional dyspepsia, 191 for gastroesophageal reflux disease, 721-723, 722f malabsorption with, 1757t mechanisms of action of, 870 for NSAID ulcer prophylaxis, 873-874 for NSAID ulcers, 872-873 for peptic ulcer bleeding, 300, 881 for peptic ulcer disease, 870 pharmacokinetics of, 870 for stress ulcers, 303-304, 878 in Zollinger-Ellison syndrome, 504 Histiocytosis, Langerhans’ cell, in infants and children, 1057 Histoplasmosis gastritis in, 851-852 hepatic, 1365-1366 in HIV/AIDS, 527t, 530 HIV/AIDS, 523-536 abdominal pain in, 161, 530-531, 530t-531t anal warts in, 2271 anorectal disease in, 531, 531t appendicitis in, 2062 bacillary angiomatosis and, 1353 bacillary peliosis hepatis in, 534 biliary tract disease in, 534, 534f Campylobacter in, diarrhea in, 527t, 529 candidiasis in, 524, 526 cholangiopathy in, 1154 cholecystitis in, acalculous, 534 Clostridium difficile infection in, 527t, 529, 1895 coccidioidomycosis in, 527t, 530 cryptococcus in, 527t, 530 cryptosporidiosis in, 526, 527t cytomegalovirus infection in diarrhea in, 527t, 528, 528f esophagitis in, 524, 524f hepatic, 533

HIV/AIDS (Continued) diarrhea in, 526-530 differential diagnosis of, 526, 526t drug-induced, 530 evaluation and management of, 526-530, 527f-529f, 527t, 530t HIV as pathogen in, 529 drug-induced liver disease in, 1417 dysphagia in, 524-526, 524f-525f, 525t esophageal neoplasms in, 525 esophageal ulceration in bacterial, 525 nonspecific, 524-525, 525f esophagitis in, 524-526, 524f-525f, 525t gastrointestinal bleeding in, 531-532, 532t gastrointestinal symptoms in, evaluation of, 523-524, 524f GBV-C virus infection with, 1344-1346, 1345f HAART for. See Antiretroviral therapy, highly active (HAART). hemorrhoids in, 2261, 2263 hepatitis A in, 533, 1285 hepatitis B in, 533, 1306-1307 hepatitis C in, 533-534, 1327, 1333-1334 hepatitis D in, 533 hepatocellular carcinoma in, 1577 hepatomegaly in, 532-535, 532t herpes simplex virus infection in, 357, 525, 525f histoplasmosis in, 527t, 530 human papillomavirus infection in, anorectal, 531, 531t isosporiasis in, 527, 527t jaundice in, 534-535 Kaposi’s sarcoma in, 534 liver test abnormalities in, 532-535, 532t lymphoma in, 356, 459-460 hepatic, 534 microsporidiosis in, 527-528, 527t mucocutaneous manifestations of, 355-356, 355f Mycobacterium avium complex in diarrhea in, 527t, 529, 529f hepatic granuloma in, 533 odynophagia in, 524-526, 524f-525f, 525t opportunistic infections in, CD4 count and, 523-524, 524f pancreatitis in, 938-939 papillary stenosis in, 534, 534f peritonitis in, 617 Pneumocystis jirovecii hepatitis in, 534 salmonellosis in, 527t, 529 shigellosis in, 527t, 529 small bowel bacterial overgrowth in, 527t, 529 tropical malabsorption and, 1830 tuberculosis in diarrhea in, 527t, 530 hepatic, 533 intestinal, 1878 HLA (human leukocyte antigen) amebiasis and, 1906-1907 in autoimmune hepatitis, 1465, 1469, 1469t in microscopic colitis, 2241 in primary biliary cirrhosis, 1479 in primary sclerosing cholangitis, 1156 in ulcerative colitis, 1977-1978 HLA-DQ2/DQ8 haplotypes, in celiac disease, 1798, 1802, 1803f, 1809 HLA-DR3, in autoimmune hepatitis, 1465, 1469, 1469t HLA-DR4, in autoimmune hepatitis, 1465, 1469, 1469t hMLH1 gene, in hereditary nonpolyposis colorectal cancer syndromes, 40, 43-44 hMSH2/hMSH6 genes, in hereditary nonpolyposis colorectal cancer syndromes, 40, 43-44 Hodgkin’s disease, hepatic involvement in, 565, 565t Homeobox, in pancreatic development, 916-917 Homeopathy, 2288t for irritable bowel syndrome, 2292, 2292t for postoperative ileus, 2295

Homocysteine, in alcoholic liver disease, 1386-1387 Hookworm infection gastritis in, 852 immune response modulation in, 1926 intestinal, 1925-1927, 1926f-1927f iron deficiency in, 1926, 1926t serpiginous rash in, 1925-1927, 1926f stool examination in, 1926, 1927f Hormonal therapy for colonic angioectasia, 599 for intestinal endometriosis, 2255 Hormone(s), gastrointestinal, 3, 4t. See also Peptide(s). as chemical messengers, 6-10, 6f, 8f regulation of, by intraluminal releasing factors, 17, 17f regulation of growth by, 15-17, 15f regulation of insulin and glucose by, 18-19, 19t regulation of satiety and hunger by, 18, 18t in small intestinal motility, 1648 Hormone replacement therapy colonic ischemia from, 2040 colorectal adenoma risk and, 2164 colorectal cancer and, 2197-2198 Hospice care, 2277-2278. See also Palliative care. Hospitalized patients diarrhea in, 229-230, 1872 energy requirements for, 48-49, 49t-50t Host defense factors, against enteric infection, 1844 Hourglass stomach, in syphilis, 850, 851f Howel-Evans syndrome, in gastrointestinal malignancies, 364 Howship-Romberg sign, in obturator hernia, 391 Hox genes, in intestinal morphogenesis, 1623-1624 HspE7, for anal warts, 2271, 2271t Human chorionic gonadotropin in carcinoid tumors, 485 hyperemesis gravidarum and, 628 in pancreatic endocrine tumors, 493 Human herpesvirus-8 infection, primary effusion lymphoma in, 460 Human immunodeficiency virus infection. See HIV/AIDS. Human leukocyte antigen (HLA). See HLA (human leukocyte antigen). Human papillomavirus infection anal warts in, 2270-2271, 2271t esophageal squamous cell carcinoma in, 748 esophagitis in, 742-743 in HIV/AIDS, 531, 531t Huntington’s chorea, gastrointestinal manifestations of, 580 Hyalinosis, small vessel, 590 Hyaluronan, in hepatic fibrosis, 1234-1235 Hycanthone, hepatotoxicity of, 1431 Hydatid cyst, hepatic, 1356t-1357t, 1363-1365, 1364f Hydralazine, granulomatous hepatitis from, 1437t Hydration. See also Fluid therapy. in nutritional assessment, 65 Hydrazine, hepatotoxicity of, 1452 Hydrochloric acid secretion, by parietal cells, 821f, 823-824, 824f Hydrochlorofluorocarbons, hepatotoxicity of, 1451 Hydrocodone, in palliative care, 2280-2281, 2281t Hydrocortisone, for Crohn’s disease, 1963 Hydrogen, in intestinal gas, 233-235 Hydrogen breath test. See Breath test, hydrogen. Hydrogen peroxide, chemical colitis from, 2247, 2248f Hydrogen sulfide, odoriferous gas from, 236 Hydrolases, brush border membrane, 1708, 1708f, 1708t Hydromorphone, in palliative care, 2280-2281, 2281t Hydrops, gallbladder, 1065-1066, 1113-1114 Hydrothorax, hepatic, liver transplantation and, 1597

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index 3b-Hydroxy-C27-steroid dehydrogenase deficiency, 1085, 1275 Hydroxycut, hepatotoxicity of, 1458 5-Hydroxyindoleacetic acid, in carcinoids, 483-484, 483f, 484t Hygiene theory of appendicitis, 2061 Hymenolepis nana/diminuta infection, 1932-1933 Hyoid bone, 678, 678f Hyperammonemia coma from, in urea cycle defects, 1274 differential diagnosis of, 1544, 1545t glutamine synthesis and, 1544 after hematopoietic stem cell transplantation, 549 in hepatic encephalopathy, 1544, 1546 in urea cycle defects, 1272-1273 Hyperbaric oxygen therapy, for radiation enteritis, 647 Hyperbetalipoproteinemia, gastrointestinal manifestations of, 577 Hyperbilirubinemia in Alagille’s syndrome, 1061 in autoimmune hepatitis, 1470 conjugated or mixed, jaundice in, 326, 326t evaluation of, 1228-1229, 1229f, 1229t postoperative, 1449t in sepsis, 582 in sickle cell anemia, 571 treatment of, 335 unconjugated, jaundice in, 325-326, 326t Hypercalcemia acute pancreatitis in, 966 constipation in, 268 in esophageal cancer, 753 familial hyperparathyroidism with, pancreatitis in, 956 gastric acid secretion in, 828 in hepatocellular carcinoma, 1571 in parathyroid hormone–related peptide– secreting tumors, 515 in VIPoma, 509-510 Hypercholesterolemia, in glycogen storage disease type I, 1263 Hyperchylomicronemia, gastrointestinal manifestations of, 577 Hypercoagulability, in ulcerative colitis, 2012 Hypercortisolism, 573t, 576 Hyperemesis gravidarum, 201, 628 jaundice in, 329 Hypereosinophilic syndrome, 432 Hypergammaglobulinemia, in autoimmune hepatitis, 1470 Hyperganglionosis, congenital, constipation in, 270 Hypergastrinemia. See also Gastrinoma. causes of, 7 in gastrinoma, 502-503 from proton pump inhibitors, 871 in short bowel syndrome, 1782 Hyperglucagonemia, 508 Hyperglycemia in acute pancreatitis, 982 in cirrhosis, 1219-1220 gastric emptying in, 800, 800f, 806 in glucagonoma, 506 insulin resistance in, 1219-1220 in intensive care unit patients, 49 nonalcoholic fatty liver disease and, 1402 from parenteral nutrition, 86 in VIPoma, 509-510 Hyperinsulinism. See also Insulinoma. with enteropathy and deafness, malabsorption in, 1759t-1762t in insulinoma, 495-496 in nonalcoholic fatty liver disease, 1403 Hyperlipidemia. See also Dyslipidemia. acute pancreatitis in, 970 familial, pancreatitis in, 956 after liver transplantation, 1610 nonalcoholic fatty liver disease and, 1402 in primary biliary cirrhosis, 1486 Hyperlipoproteinemia, gastrointestinal manifestations of, 577

Hypermagnesemia, from magnesium supplementation, 276 Hypermethylation, in colorectal cancer, 2203-2204 Hyperoxaluria, in short bowel syndrome, 1790 Hyperparathyroidism familial, pancreatitis in, 956 gastrointestinal manifestations of, 573t, 576 hypercalcemia from, pancreatitis and, 966 in MEN I syndrome, 494 in Zollinger-Ellison syndrome, 503-505 Hyperphagia diet, in short bowel syndrome, 1784-1785 Hypersensitivity, visceral. See Visceral sensitization. Hypersensitivity reaction acute pancreatitis in, 965 food-related. See Food allergy. Hypersensitivity vasculitis, 2047 Hypertension after bariatric surgery, 117 after liver transplantation, 1609-1610 obesity and, 104 portal. See Portal hypertension. portopulmonary, 1549-1552 weight loss and, 106 Hyperthyroidism gastrointestinal manifestations of, 573t, 575-576 malabsorption in, 1765-1766 Hypertonic saline, for intracranial hypertension, 1565-1566 Hypertrichosis lanuginosa, in gastrointestinal malignancies, 365 Hypertriglyceridemia acute pancreatitis in, 965-966 in glycogen storage disease type I, 1263 Hyperuricemia, in glycogen storage disease type I, 1263 Hypnotherapy, 2288t for functional abdominal pain syndrome, 170 for irritable bowel syndrome, 2292, 2292t Hypoalbuminemia. See Protein-losing gastroenteropathy. Hypoaminoacidemia, in glucagonoma, 506-507 Hypobetalipoproteinemia, malabsorption in, 1759t-1762t, 1763 Hypocalcemia in acute pancreatitis, 982 after bariatric surgery, 118 Hypochloremic alkalosis, from vomiting, 206 Hypochlorhydria in somatostatinoma, 512 in VIPoma, 509-510 Hypoganglionosis, in colon, congenital, constipation in, 269-270 Hypogeusia, 354 Hypoglycemia in acute liver failure, 1561t causes of, 497-498, 497t in eating disorders, 128 fasting, 497, 497t in glycogen storage disease type I, 1263 in hepatocellular carcinoma, 1571 in insulinoma, 496-497. See also Insulinoma. postprandial, 497, 497t in shigellosis, 1859-1860 Hypoglycemic agents, oral, hepatotoxicity of, 1435-1436 Hypoglycin A, hepatotoxicity of, 1455 Hypogonadism, in cirrhosis, 1554 Hypokalemia in eating disorders, 128 in VIPoma, 509-510 from vomiting, 206 Hypomagnesemia, with secondary hypocalcemia, malabsorption in, 1759t-1762t Hyponatremia in ascites, 1536 from vomiting, 206 Hypoparathyroidism chronic intestinal pseudo-obstruction in, 2139 gastrointestinal manifestations of, 573t, 576-577

Hypopharynx anatomy of, 677, 678f diverticula of. See Diverticula, Zenker’s. Hypoproteinemia, 437-440. See also Proteinlosing gastroenteropathy. Hyposplenism, in celiac disease, 1813 Hypotension in acute liver failure, 1561t, 1562-1563 during endoscopy, 655 Hypothalamic-pituitary-adrenal (HPA) axis, in stress-immune response, 345-346 Hypothermia, for intracranial hypertension, 1565-1566 Hypothyroidism chronic intestinal pseudo-obstruction in, 2139 constipation in, 268 gastrointestinal manifestations of, 573t, 576 Hypoventilation, during endoscopy, 655 Hypovolemia, in small intestinal obstruction, 2106 Hypoxemia in hepatopulmonary syndrome, 1550 in liver disease. See Hepatopulmonary syndrome. Hypoxia in alcoholic liver disease, 1385-1386 in intestinal ischemia, 2029

I

Icterus. See Jaundice. Ileal pouch–anal anastomosis, 2018-2023 advantages of, 2018 anatomy of, 2018, 2019f cancer risk in, 2022-2023 clinical results of, 2019-2020 complications of, 2019, 2019f Crohn’s disease in pouch after, 2020 cuffitis in, 2020 double-stapled versus hand-sewn, 2022 fecal incontinence with, 2020 fertility and, 2023 in indeterminate colitis, 2023 infection after, 2019, 2019f, 2023 infliximab preoperative therapy and, 2023 introduction of, 2015 laparoscopic, 2024 long-term results of, 2018-2022 pelvic sepsis after, 2019, 2019f pouch failure in, 2021 pouchitis in, 2008-2010, 2020-2021, 2020t etiologies of, 2021 sequelae of, 2021 treatment of, 2021 proctocolectomy with, for ulcerative colitis, 2004-2005 quality of life in, 2021 risk-benefit analysis in, 2025, 2025t sexual dysfunction and, 2021-2022 strictures of, 2019, 2019f temporary ileostomy with, 2019, 2022 Ileitis differential diagnosis of, 1956t prestomal, in Brooke ileostomy, 2017 Ileocecal junction, in colonic motility, 1666, 1667f Ileocecal valve, 1615 loss of, short bowel syndrome in, 1783 resection of, small intestinal bacterial overgrowth after, 1773 Ileocolonic anastomosis, 1780f Ileorectal anastomosis, colectomy with. See Ileorectostomy. Ileorectoplasty, for fecal incontinence, 255t, 256 Ileorectostomy, 2015, 2023-2024 complications of, 2024 for constipation, 283 in familial adenomatous polyposis, 2183 patient selection for, 2024 physiology of, 2024 risk-benefit analysis in, 2025, 2025t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

xlix

l

Index Ileostomy, 2015-2026. See also Colectomy; Ileorectostomy; Proctocolectomy. Brooke, 2015, 2016f complications of, 2017 ileitis in, prestomal, 2017 parastomal hernias in, 2017 postoperative care in, 2017 quality of life with, 2021 risk-benefit analysis in, 2025, 2025t stomal obstruction in, 2017 for constipation, 283 continent (Kock pouch), 2015, 2017-2018, 2018f, 2025, 2025t continent (T pouch), 2018 diarrhea after, 229 electrolyte balance with, 2016 historical aspects of, 2015 laparoscopic, 2024 output of, 2016 quality of life with, 2016-2017 temporary, in ileal pouch–anal anastomosis, 2019, 2022. See also Ileal pouch–anal anastomosis. Ileum. See also Small intestine. adenoma of, in familial adenomatous polyposis, 2180 anatomy of, 1615 bile acid malabsorption in, steatorrhea from, 1755-1756 bile acid reabsorption in, 1083-1084 carcinoid tumors of, 480, 480t in Crohn’s disease, 1949-1950 diverticula of, 377-378 duplications of, 1632 gallstone disease in, 1093 nutrient presence in, gastrointestinal transit and, 1697 resection of bile acid circulation disturbances after, 1087 diarrhea after, 228-229 limited, short bowel syndrome in, 1783-1784 Ileus definition of, 2121 etiology of, 2124t gallstone, 2115-2116, 2115f postoperative, 2123 clinical features of, 2126 epidemiology of, 2123 homeopathy for, 2295 pathophysiology of, 2124-2125 prevention of, 2126-2127 fast-track surgery for, 2127-2128, 2128t in intraoperative period, 2126 medications for, 2127 multimodal approach to, 2127-2128, 2128t in postoperative period, 2126-2127, 2126t treatment of, 2126-2128 versus small intestinal obstruction, 2107, 2108f Imatinib mesylate adverse effects of, 470 for gastrointestinal stromal tumors, 469-471, 470t, 473 hepatotoxicity of, 545 resistance to, 471 Imerslund-Gräsbeck syndrome, malabsorption in, 1759t-1762t, 1763 Iminoglycinuria, malabsorption in, 1758-1763, 1759t-1762t Imipramine for bulimia nervosa, 134 for esophageal hypersensitity, 703-704 Imiquimod, for anal warts, 2270, 2271t Immune response modulation in Crohn’s disease, 1962t, 1966-1967 in hookworm infection, 1926 in pinworm infection, 1929 in ulcerative colitis, 1996-1999, 1997t-1998t Immune system in alcoholic liver disease, 1387-1388, 1388f in celiac disease, 1802-1804, 1803f in cholera, 1851 in Crohn’s disease, 1946-1947, 1947f in enteric infection, 1844

Immune system (Continued) in graft-versus-host disease. See Graft-versushost disease. in Helicobacter pylori infection, 835-838 in hepatitis B, 1292-1293 intestinal bacteria interactions with, 1771 in intestinal ion transport, 1688-1689 mucosal, 21-30. See also Gut-associated lymphoid tissue (GALT). antigen processing and antigen presenting cells in, 26-27, 27f barrier defenses in, 24-25, 140, 140t dendritic cells in, 29, 140 development of, 140 epithelial cells in, 25-26, 140-141 in food allergy, 139-142, 140t, 141f immunoglobulins in, 23-24, 24f intraepithelial lymphocytes in, 28 lamina propria mononuclear cells in, 28 oral tolerance in, 22-23, 23f Peyer’s patches and M cells in, 25, 26f T cell differentiation in, 28-29 T cells in, 140 in pregnancy, 626 in protein-energy malnutrition, 63 stress and, 345-346 in Whipple’s disease, 1835-1836 Immunization. See Vaccine. Immunocompromised patients acute abdominal pain in, 161 appendicitis in, 2062 diverticulitis in, 2084 salmonellosis in, 1862-1863 Immunodeficiency acquired. See HIV/AIDS. common variable hepatitis C in, 1326 malabsorption in, 1764, 1829t, 1830 in infants and children, sclerosing cholangitis in, 1057 lymphoma related to, 459-460 malabsorption in, 1763-1764 severe combined, malabsorption in, 1764 Immunoglobulin(s) of gut-associated lymphoid tissue, 23-24, 24f hepatitis B, 1307, 1309t in liver transplantation, 1599-1600 hepatitis C, 1328 intravenous in postexposure hepatitis A prophylaxis, 1283-1284 for pseudomembranous enterocolitis, 1900 Immunoglobulin A deficiency of in celiac disease, 1813 malabsorption in, 1763 dimeric, 23-24, 24f linear deposits of, 1812 secretory, 23-24, 24f Immunoglobulin E antibody to, for eosinophilic gastrointestinal disorders, 434-435 in gastrointestinal food allergy, 141-143 in gut-associated lymphoid tissue, 24 Immunoglobulin G in bile, 1093 in gut-associated lymphoid tissue, 24 in ulcerative colitis, 1979 Immunoglobulin G antitoxin A antibody, in Clostridium difficile-associated diarrhea and colitis, 1894, 1894f Immunoglobulin G4 in autoimmune pancreatitis, 991 in cholangitis, 1155 Immunoglobulin M in bile, 1093 in gut-associated lymphoid tissue, 24 Immunologic tests, in nutritional assessment, 78 Immunoproliferative small intestinal disease. See Small intestinal immunoproliferative disease. Immunosuppressive therapy for celiac disease, 1817 diarrhea from, 542

Immunosuppressive therapy (Continued) drug interactions with, 1606t hepatitis B flare from, 1297 for liver transplantation, 1605-1606, 1606t for primary sclerosing cholangitis, 1164-1165 Immunotargeted therapy and immunotherapy, for colorectal cancer, 2237 Impedance testing in esophageal motility disorders, 695f, 698-699 in gastroesophageal reflux disease, 719 in nutritional assessment, 66 in small intestinal motility disorders, 1652-1653 Imperforate anus, 1634f, 1635 Implantable ports, in parenteral nutrition, 87 Importins, 1209 Inborn errors of metabolism, 1259-1278 in α1-antitrypsin deficiency, 1260-1262 in bile acid synthesis defects, 1274-1276, 1275t in bile acid transport defects, 1275t, 1276-1277 clinical features of, 1259, 1260t in copper overload. See Wilson disease. in cystic fibrosis, 1277, 1277t in glycogen storage disease, 1262-1265 in glycosylation disorders, 1265-1266, 1265f in iron overload. See Hemochromatosis. liver transplantation for, 1603 mitochondrial, 1277-1278 in porphyrias, 1266-1269, 1266f, 1267t screening for, 1259, 1260t in tyrosinemia, 1269-1271, 1270f in urea cycle defects, 1271-1274, 1272f, 1273t Incision and drainage, of anorectal abscess, 2266-2267 Incisional hernia, 389-390 clinical manifestations and diagnosis of, 389 etiology and pathogenesis of, 389 incidence and prevalence of, 389 treatment and prognosis in, 389-390 Incisura angularis, of stomach, 775, 775f Incontinence, fecal. See Fecal incontinence. Incretins, 18-19 Indian hedgehog protein in intestinal development, 1618 in pancreatic development, 916 Indigestion. See Heartburn (pyrosis). Indinavir, hepatotoxicity of, 1430 Indocyanine green clearance test, 1235 Infant formula, for glycogen storage disease type I, 1264 Infants. See also Children; Neonates. acute abdominal pain in, 161 appendicitis in, 2062 bile duct perforation in, spontaneous, 1056 bile plug syndrome in, 1056 biliary atresia in. See Biliary atresia. biliary tract disease in, 1050-1052, 1051t-1052t botulism in, 1883 celiac disease in, 1804 choledochal cysts in, 1058-1059, 1058f-1059f cholestasis in, 1050-1052, 1051t-1052t. See also Cholestasis, pediatric. colic in, IgE-mediated hypersensitivity in, 143 digestion and absorption in, 1729-1732, 1730f-1731f gallbladder disease in, 1063-1066 glycogen storage disease in, 1262-1265 hepatic hemangioendothelioma in, 1587 hydrops of gallbladder in, 1065-1066 hypertrophic pyloric stenosis in, 783-785, 784f interlobular bile duct paucity in, 1206, 1206f nonsyndromic, 1060 syndromic, 1060-1062, 1061f-1062f intrahepatic bile duct in, congenital dilatation of, 1059-1060, 1060f premature, lactase deficiency in, 1754 primary sclerosing cholangitis in, 1056-1058 Infection. See also Abscess; Bacterial infection; Fungal infection; Peritonitis; Viral infection. in acute liver failure, 1561t, 1562, 1566 in acute pancreatitis, 963, 966, 977-978 in children, 938-939, 938t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Infection (Continued) catheter-related, 87, 1788f after endoscopic retrograde cholangiopancreatography, 1197 after endoscopy, 656 extrahepatic, liver transplantation and, 1597 after ileal pouch–anal anastomosis, 2023 after liver transplantation, 1607, 1607f, 1608t liver transplantation and, 1597 Infectious diarrhea. See Diarrhea, infectious. Inferior mesenteric artery anastomotic circulation of, 2028, 2029f anatomy of, 2028, 2029f Inferior mesenteric vein, anatomy of, 1489, 1490f Inferior vena cava, membranous obstruction of, 1371-1372, 1372f, 1374. See also BuddChiari syndrome. hepatocellular carcinoma in, 1577 Infertility. See Fertility. Infiltrative diseases, gastrointestinal manifestations of, 584-590 Inflammation controlled/physiologic, in gut-associated lymphoid tissue, 21-22, 22f Crohn’s disease and, 1947-1949, 1947f gastric cancer and, 888-889, 890f in irritable bowel syndrome, 2097-2098, 2097f in postoperative ileus, 2125 Inflammatory bowel disease. See Crohn’s disease; Ulcerative colitis. versus appendicitis, 2063t Inflammatory markers in Crohn’s disease, 1960 in ulcerative colitis, 1984 Inflammatory mediators, in intestinal ion transport, 1688-1689 Inflammatory pseudotumor, hepatic, 1367, 1588-1589 Infliximab for Crohn’s disease, 422-423, 1962t, 1967-1969 for perianal Crohn’s fistula, 2269 in pregnancy, 630 structure of, 2000-2001, 2001f for ulcerative colitis, 2000-2001 postoperative infection and, 2023 for ulcerative enteritis, 2057 Informed consent, for endoscopy, 654 Infrared photocoagulation, for hemorrhoids, 2261, 2263t Inguinal hernia. See Hernia, inguinal. Injection drug users, hepatitis C in, 1317-1318 Injection therapy. See also Sclerotherapy. for abdominal wall pain, 165 for colonic diverticular bleeding, 312 for diverticular bleeding, 2085-2087 for esophageal cancer, 765 for gastrointestinal bleeding, 292 for hepatocellular carcinoma, 1579 for peptic ulcer bleeding, 297-298, 298f-300f, 880 Inlet patch, 670t, 675, 675f Insecticides, hepatotoxicity of, 1452-1453 Insula-acinar portal system of, 914, 914f Insulin dosing of, during parenteral nutrition, 86 radioimmunoassays specific to, 497 release of adiponectin and, 103 regulation of, by gastrointestinal peptides, 18-19, 19t synthesis of, 496 Insulin-like growth factor in gastrointestinal tract, 16 receptors for, 16 Insulin resistance in diabetes mellitus, 19 in hepatitis C, 1321, 1326 in nonalcoholic fatty liver disease, 1403 Insulin-sensitizing agents, for nonalcoholic fatty liver disease, 1410 Insulinoma, 495-498 clinical features of, 496, 496t definition of, 495

Insulinoma (Continued) diagnosis of, 496-498, 497t localization of, 517-519, 518f-519f pathophysiology and pathology of, 495-496 treatment of, 498 medical, 498 surgical, 498 Integrins, in eosinophilic gastrointestinal disorders, 426, 426f Intensive care unit patients gastrointestinal complications in, 581-584, 582f hyperglycemia in, 49 Interferon(s) in gastrointestinal tract, 13 for hepatitis B, 1300t, 1302, 1307 plus nucleoside analogues, 1305-1306 hepatitis B flares and, 1298 for hepatitis D, 1311-1312 for HIV/HCV coinfection, 534 pegylated for hepatitis B, 1302, 1307 plus lamivudine, 1305-1306 for hepatitis C, 1329 plus ribavirin, 1320 for hepatitis D, 1311-1312 Interferon-alpha for anal warts, 2271, 2271t for carcinoid tumors, 489 for hepatitis C, in liver transplant recipients, 1600-1601 for pancreatic endocrine tumors, 521-522 Interferon-gamma oral tolerance and, 23 in protein-energy malnutrition, 61, 62t Interleukin(s), in gastrointestinal tract, 13 Interleukin-1, in protein-energy malnutrition, 61, 62t Interleukin-1β, in gastric cancer, 893 Interleukin-2 receptor, antibody to, for ulcerative colitis, 2001 Interleukin-2 receptor blocker, for liver transplantation, 1606t Interleukin-5 antibody to, for eosinophilic gastrointestinal disorders, 434 in eosinophilic esophagitis, 427 Interleukin-6 in acute pancreatitis, 975 in cholangiocarcinoma, 1173 in hepatic regeneration, 1213 in protein-energy malnutrition, 61, 62t Interleukin-10, in gastric cancer, 893 Interleukin-12/23 antibody to, for Crohn’s disease, 1970 in Crohn’s disease, 1944t-1945t, 1945-1946 in ulcerative colitis, 1977, 1980 Interleukin-13, in eosinophilic esophagitis, 427 Interleukin-15, in celiac disease, 1798, 1804, 2054 Interleukin-16, in Whipple’s disease, 1836 Intermediate filaments, of hepatocytes, 1208 International normalized ratio in cirrhosis, 1554-1555 in gastrointestinal bleeding, 288 Interpersonal therapy, for eating disorders, 132-133 Intersphincteric abscess, 2266, 2267f Intersphincteric groove, 2258 Interstitial acute pancreatitis, 960f, 961, 972 Interstitial cells of Cajal in colonic motility, 1661-1662, 1661f deep muscular, 1645 development of, 803 in enteric mesenchymyopathies, 2133 gastric, 791-792, 793f GIST cells and, 462 intestinal, 1619, 1621f intramuscular, 1645, 1661-1662, 1661f loss of, 806 myenteric, 1645, 1661, 1661f in slow-transit constipation, 264 in small intestinal motility, 1644-1645, 1644f submucosal, 1661, 1661f

Intestinal adaptation, 1726-1729 bariatric surgery and, 1732-1733 developmental, 1729-1732, 1730f-1731f enzymatic, 1727, 1727f mucosal, 1726-1728, 1728f nutritional requirements and, 1728-1729, 1729f to resection, 1783 in short bowel syndrome, pharmacologic enhancement of, 1793 signaling for, 1729 therapeutic implications of, 1729 Intestinal bacteria. See Bacteria, intestinal (commensal). overgrowth of. See Bacterial overgrowth, small intestinal. Intestinal failure. See Short bowel syndrome. Intestinal flukes, 1923f, 1933-1934 Intestinal ischemia, 2027-2048, 2028t anatomy of, 2027-2028, 2028f-2029f diarrhea from, 230 meandering artery in, 2028, 2029f mesenteric. See Mesenteric ischemia. pathophysiology and pathology of, 2028-2029 reperfusion injury in, 2029 in small intestinal obstruction, 2106 vasculitis in, 2046-2047 Intestinal lengthening procedures, for short bowel syndrome, 1791, 1791f-1792f Intestinal motility. See also Anorectal motility; Colonic motility; Small intestinal motility. in digestion and absorption, 1697 disorders of in chronic pancreatitis, 1014-1015 diarrhea from, 213-214, 218 in gallstone disease, 1101 ileus as. See Ileus. paraneoplastic, 564 pseudo-obstructive. See Pseudo-obstruction. small intestinal bacterial overgrowth in, 1773 enteric infection and, 1844 neural control system for, 2121, 2122f-2123f, 2123-2124 paracrine immunoneuroendocrine system (PINES) in, 1686 in pregnancy, 625-626 in scleroderma, 560 Intestinal obstruction, 2105-2120. See also Pseudo-obstruction. in carcinoid tumors, 476 closed loop, 2105 colonic, 2116-2120. See also Colonic obstruction. in colorectal cancer, 2216 distal, in cystic fibrosis, 944-945 manometry in, 2142 palliative care for, 2285 partial versus complete, 2105 simple versus strangulated, 2105 small intestinal, 2105-2116. See also Small intestinal obstruction. vomiting in, 198-201 Intestinal permeability test, for malabsorption, 1753-1754 Intestinal protozoa, 1905-1920 Intestinal T cell lymphoma, enteropathy-type. See Enteropathy-associated T cell lymphoma. Intestinal tapering procedure, for short bowel syndrome, 1791, 1793f Intestinal transit time in eating disorders, 130, 135-137 malabsorption and, 1741 Intestinal trefoil factor, in gastrointestinal tract, 16 Intestinal tuberculosis. See Tuberculosis, intestinal. Intestine(s). See also Colon; Rectum; Small intestine. absorptive cells of, 1617f, 1618 adaptation of. See Intestinal adaptation. adventitia of, 1619 aganglionosis of. See Hirschsprung’s disease. amebiasis of. See Amebiasis, intestinal. angina of, 2044-2046, 2045f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

li

lii

Index Intestine(s) (Continued) atresia of, 1632-1633, 1632f-1633f anomalies associated with, 1633 differential diagnosis of, 1633 congenital anomalies of, 1626-1641, 1627t duplications of, 1631-1632 electrolyte transport in. See Fluid and electrolyte transport. embryology of, 1623-1626, 1624f-1626f endometriosis of, 2253-2255, 2254f-2255f epithelium of. See Epithelial cells, intestinal; Epithelium, intestinal. fluid transport in. See Fluid and electrolyte transport. gas in. See Gas, intestinal. goblet cells of, 1618, 1618f interstitial cells of Cajal of, 1619, 1621f. See also Interstitial cells of Cajal. ion transport in. See Fluid and electrolyte transport. ischemia of. See Intestinal ischemia; Mesenteric ischemia. lymphatic development in, 1626 M cells of, 1619 metaplasia of classification of, 895 gastric cancer and, 895 mononuclear cells of, 28-29 morphogenesis of, genetic control of, 1623-1624, 1624f motility of. See Intestinal motility. muscularis of, 1619, 1621f nervous system of. See Enteric nervous system. neuroendocrine cells of. See Neuroendocrine (enteroendocrine) cells. neuronal dysplasia of, 1639-1640, 1639f constipation in, 270 obstruction of. See Intestinal obstruction. pacemaker cells of. See Interstitial cells of Cajal. Paneth cells of, 1618, 1619f perforation of in colorectal cancer, 2216 in HIV infection, 161 PINES regulatory system in, 214, 214f, 1686, 1686f pseudo-obstruction of. See Pseudo-obstruction. radiation damage to. See Radiation enteritis. resection of, types of, 1780f rotation of anomalies of, 1630-1631, 1631f normal, 1625, 1626f serosa of, 1619 stem cells of, 1618 submucosa of, 1619 transplantation of for chronic intestinal pseudo-obstruction, 2143 complications of, 541 for short bowel syndrome, 1791-1793, 1792t vascular development in, 1625-1626 wall of, layers of, 1663f Intestinofugal neuron, in small intestinal motility, 1646-1647 Intimin, 1844 Intra-abdominal pressure, elevated, in abdominal compartment syndrome, 160 Intracranial hypertension in acute liver failure, 1565 treatment of, 1565-1566 Intracranial pressure monitoring, in acute liver failure, 1565 Intracrine signaling, 15-16 Intraganglionic laminar endings, in small intestinal motility, 1647 Intraluminal impedance measurement, in esophageal motility disorders, 695f, 698-699 Intraluminal releasing factors, regulation of gastrointestinal hormones by, 17, 17f Intramuscular arrays, in small intestinal motility, 1647 Intraperitoneal chemotherapy, for gastric cancer, 904

Intrapulmonary shunt, in hepatopulmonary syndrome, 1550 Intrauterine factors, in obesity, 101-102 Intrinsic factor cobalamin binding to, 1719, 1719f deficiency of, malabsorption in, 1759t-1762t, 1763 gastric secretion of, 829-830 Intubation endotracheal in acute liver failure, 1565 in Zenker’s diverticula, 371-372 nasogastric esophageal injuries related to, 739, 739f gastroesophageal reflux disease after, 715 Intussusception in cystic fibrosis, 945 small intestinal obstruction in, 2114-2115, 2114f Iodine absorption of, 1726 dietary, 56t-57t, 80t, 81 Iodoalkanes, hepatotoxicity of, 1452 Iodoquinol for amebiasis, 1910t for amebic liver abscess, 1369 for Blastocystis hominis infection, 1914 for Dientamoeba fragilis infection, 1914 Ion channel, 1680, 1680f Ion channel–coupled receptor(s), 15 Ion pump, 1680f, 1681 Ion transporters. See also Fluid and electrolyte transport, intestinal. overview of, 1681-1686, 1681f-1683f spatial distribution of, 1677-1678, 1678f types of, 1679-1681, 1680f IPAA. See Ileal pouch–anal anastomosis. Ipecac syrup, risks associated with, 124-125 IPEX syndrome, malabsorption in, 1759t-1762t, 1764 Iproniazid, hepatotoxicity of, 1436 Irbesartan hepatotoxicity of, 1433 for portal hypertension, 1502-1503 Irinotecan for colorectal cancer, 2235, 2236f plus 5-fluorouracil and leucovorin, 2232-2233, 2235 for small intestinal adenocarcinoma, 2152 Iron absorption of, 1723-1725, 1724f adaptive changes in, 1728, 1729f in hereditary hemochromatosis, 1240, 1241f in short bowel syndrome, 1781t deficiency of. See also Iron deficiency anemia. in celiac disease, 1810-1811 after gastric bypass, 1758 in hookworm infection, 1926, 1926t in tropical sprue, 1826 dietary, 56t-57t, 80t, 81, 1723 gastropathy from, 856 hepatic storage of, in hepatitis C, 1326 hepatotoxicity of, 1453 malabsorption of, 1740 after gastric resection, 1757 measurement of, in hereditary hemochromatosis, 1244-1245, 1244t overload of. See Iron overload. reduction of, for nonalcoholic fatty liver disease, 1410 Iron deficiency anemia differential diagnosis of, 321 iron-refractory, malabsorption in, 1759t-1762t occult gastrointestinal bleeding and, 321 Iron overload African, 1240 in alcoholic liver disease, 1391 causes of, 1239-1240, 1240t after hematopoietic stem cell transplantation, 547t, 554 in hematopoietic stem cell transplantation candidates, 545 inherited forms of. See Hemochromatosis, hereditary.

Iron overload (Continued) parenteral, 1240 secondary, 1240, 1240t Iron regulatory proteins, 1725 Irritable bowel syndrome, 1673, 2091-2104 abdominal distention in, 238-240, 239f, 2092 abdominal pain in, 2092, 2095 abdominal surgery and, 2094 abuse history in, 167-168 alarm features in, 2099, 2099t after bariatric surgery, 117-118 biopsychosocial model of, 2095, 2095f celiac disease and, 2100 chronicity of, 2092-2093 clinical features of, 2092-2093 colonic ischemia in, 2044 complementary and alternative medicine for, 2291-2292, 2292t constipation in, 2092 costs of, 2094-2095 definition of, 2091, 2092t diagnosis of, 2099-2100, 2099t, 2100f diarrhea in, 215, 219, 227, 2092 diverticula and, 2077 epidemiology of, 2093-2094 functional dyspepsia and, 187 gastric motility disorders in, 809 gender and, 2093 health care seeking in, 2094 history in, 2092-2093 incidence of, 2094 inflammatory bowel disease and, 2092 intestinal gas in, 238-240, 239f noncolonic symptoms of, 2092 organic diseases associated with, 2099, 2099t pathophysiology of, 2095-2099, 2095f bacterial overgrowth in, 2096-2098 central dysregulation in, 2098 colonic flora abnormalities in, 2098 food intolerance and allergy in, 2098 gas propulsion/expulsion abnormalities in, 2096-2097 genetic factors in, 2099 inflammation in, 2097-2098, 2097f motility abnormalities in, 2095, 2096f psychological factors in, 2098-2099 visceral sensitization in, 2095-2096, 2097f physical examination in, 2093 postinfection, 2094, 2097, 2097f in Clostridium difficile-associated diarrhea and colitis, 1900 in shigellosis, 1861 in traveler’s diarrhea, 1875 prevalence of, 2093 probiotics for, 96 quality of life in, 2094 race and, 2093 risk factors for, 2094-2095 small intestinal bacterial overgrowth and, 1774-1775 small intestinal motility in, 1656 subgroups in, 2093-2094, 2093f treatment of, 2100-2103, 2101t antibiotics for, 2102 anticholinergics for, 2102 antidepressants for, 2102 antidiarrheal agents for, 2102 antispasmodics for, 2102 complementary and alternative medicine for, 2103 diet in, 2101 laxatives for, 2102 patient education and support in, 2100 probiotics for, 2102-2103 psychological therapy for, 2103 serotonin receptor drugs for, 2102 versus ulcerative colitis, 1991 Ischemia colonic, 2038-2044. See also Colon ischemia. hepatic, jaundice in, 327 intestinal, 2027-2048, 2028t. See also Intestinal ischemia. mesenteric. See Mesenteric ischemia.

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Ischemic gastroparesis, 808 Ischemic gastropathy, 858 Ischemic heart disease, gastrointestinal manifestations of, 584 Ischemic hepatitis, 1379-1380, 1380f, 1450t Ischiorectal abscess, 2266, 2267f Islet cell tumors. See Pancreatic endocrine tumors. Islets of Langerhans, 912f, 914 Isocarboxazid, hepatotoxicity of, 1436 Isoflurane, hepatotoxicity of, 1449, 1449t Isomaltase deficiency, congenital, 1641 Isoniazid hepatotoxicity of, 1434, 1438t for intestinal tuberculosis, 1878 Isoprostane, in alcoholic liver disease, 1385 Isosorbide dinitrate for achalasia, 701 for Chagas’ disease, 1919 for esophageal variceal bleeding, 1510 for portal hypertension, 1502 Isospora belli infection, 1917 Isosporiasis in HIV/AIDS, 527, 527t malabsorption in, 1829, 1829t Ispaghula, for constipation, 276 Ito cells. See Stellate cells, hepatic. Itopride, for functional dyspepsia, 193 Itraconazole, hepatotoxicity of, 1435 Ivermectin for strongyloidiasis, 1361, 1924-1925 for Trichuris trichiura infection, 1928 Iverson’s syndrome, pancreatic insufficiency in, 955

J

Jagged1 (JAG) gene, in Alagille’s syndrome, 1060-1062, 1206 Jaundice, 323-336. See also Hyperbilirubinemia. in acute hepatocellular dysfunction, 326-327 in alcoholic liver disease, 1390, 1390t in autoimmune pancreatitis, 991 in bile duct compression, 330 in bile duct diseases, 329-330 in biliary atresia, 1054 in biliary tract obstruction, 325t, 329, 334-335 in cholangiocyte injury, 328 in choledocholithiasis, 329 in cholestasis, 327-329, 1050-1052, 1051t1052t. See also Cholestasis. in chronic hepatocellular dysfunction, 327 in conjugated or mixed hyperbilirubinemia, 326, 326t in critically ill patients, 329 diagnostic approach to, 330-334, 330f history and physical examination in, 330-331, 331t imaging studies in, 332-334, 333t initial laboratory studies in, 331, 331t liver biopsy in, 334, 334f overall approach to, 332 serologic testing in, 334 differential diagnosis of, 325-330, 325t Epping, 1454 after hematopoietic stem cell transplantation, 546-549, 547t, 548f in hepatitis A, 1282 in hepatitis E, 1340-1341 in HIV/AIDS, 534-535 in infiltrative liver disease, 327-328 in liver disease, 325t, 326-329 in neonates, 326 palliative care for, 2285-2286 in pancreatic cancer, 1020, 1025 postoperative, 583-584, 583t, 1449, 1449t in pregnancy, 329 sepsis and, 582-583, 582f, 1353 serum bilirubin level in, 324 therapeutic options for, 334-335 in unconjugated hyperbilirubinemia, 325-326, 326t

JC virus, colorectal adenoma in, 2165 Jejunectomy, intestinal adaptation to, 1783 Jejunitis, ulcerative. See Ulcerative enteritis. Jejunocolonic anastomosis, 1780f Jejunoileal atresia, 1632-1633, 1633f Jejunoileal bypass, intestinal adaptation to, 1783 Jejunoileal Crohn’s disease, 1950 Jejunoileitis, ulcerative. See Ulcerative enteritis. Jejunostomy care of, 89 direct percutaneous, 92-93, 92f end-, 1780f malabsorption with, 1780 nutrient, 1780-1781, 1780f water and electrolyte, 1781-1782, 1781t, 1782f surgical, 93 Jejunum. See also Small intestine. adenoma of, in familial adenomatous polyposis, 2180 anatomy of, 1615 carcinoid tumors of, 480, 480t in digestion and absorption, short bowel syndrome and, 1780, 1780f diverticula of, 377-378 diverticulosis of, 2139 Jin bu huan, hepatotoxicity of, 1457-1458 Johanson-Blizzard syndrome, pancreatic insufficiency in, 954 Joint disease. See Arthritis; Arthropathy. Juvenile colonic polyps, 2174-2175, 2175f Juvenile hereditary hemochromatosis, 1239 Juvenile polyposis, 2185t, 2186-2187 cancer risks and screening recommendations for, 2181t colorectal cancer in, 2209 eosinophils in, 432 gastric cancer in, 893 Juvenile rheumatoid arthritis adult-onset, gastrointestinal and hepatic manifestations of, 559-560 Reye’s syndrome in, 1430-1431

K

K-ras oncogene in adenoma-carcinoma hypothesis, 2160 in colorectal cancer, 2200-2201, 2200t diversity of signaling through, 41, 41f in gastrointestinal tumors, 37 in pancreatic cancer, 1019-1020 Kaposi’s sarcoma, in HIV/AIDS, 355-356, 355f, 534 Karnofsky Performance Status, 2278, 2280t Karyopherins, 1209 Kasai hepatoportoenterostomy for biliary atresia, 1054-1055, 1055f outcome after, 1056 Katayama fever, 1936 Katayama syndrome, 1362. See also Schistosomiasis. Kato-Katz thick smear, for schistosomiasis, 1937 Kava kava, hepatotoxicity of, 1458 Kawasaki’s disease pancreatitis in, 939 vasculitis in, 2047 Kayser-Fleischer rings, in Wilson disease, 1252-1253 Kegel exercises in fecal incontinence, 253-254, 253f, 254t in rectocele, 266 Keratosis palmaris et plantaris, in gastrointestinal malignancies, 364 Ketoconazole drug interactions of, with proton pump inhibitors, 871 hepatotoxicity of, 1434-1435 Ketogenesis, during starvation, 59-60 Ketorolac, ileus and, 2127 Ketotifen, for eosinophilic gastrointestinal disorders, 434

Kidney. See also Renal. bile acid transport in, 1083-1084 disease of. See Hepatorenal syndrome; Renal disease. failure of. See Renal failure. in hepatocellular carcinoma, 1579 in protein-energy malnutrition, 64 transplantation of, 539-540 hepatitis C after, 1326-1327 Kidney-pancreas transplantation, 539-540 Kidney stones calcium oxalate, in short bowel syndrome, 1790, 1790f in Crohn’s disease, 1955 Killian’s triangle, 677-678, 678f herniation through, in Zenker’s diverticula, 371 KIT in gastrointestinal stromal tumors, 462-464, 463f tyrosine kinase inhibitor targeting, 469-471, 470t Klatskin tumor, 1171, 1172f, 1174f, 1175-1176, 1190, 1580-1582 Klebsiella oxytoca infection, antibiotic-associated diarrhea in, 1889 Klippel-Trenaunay-Weber syndrome, 606 Knudson’s hypothesis, 38, 38f Kock pouch, 2015, 2017-2018, 2018f, 2025, 2025t Kohlmeier-Degos disease. See Degos disease. Kupffer cells in alcoholic liver disease, 1387-1388, 1388f in nonalcoholic fatty liver disease, 1404 in sickle cell anemia, 570-571, 570f sinusoidal, 1203-1204, 1211 Kwashiorkor, in children, 61-62, 62t

L

Labetalol, hepatotoxicity of, 1433 Lactase, 1708, 1708f, 1708t age-related changes in, 1730 deficiency of, 1709. See also Lactose, malabsorption of. acquired primary, 1754, 1754t congenital, 1641, 1754, 1759t-1762t in cystic fibrosis, 943 diarrhea from, 213, 224 transient, in premature infants, 1754 persistent activity of, 1754, 1754t Lactase gene, 1709 Lactase–phlorizin hydrolase, 1730 developmental downregulation of, 1730 polymorphisms of, 1730, 1731f, 1754 Lactate, in ascites, 1526 Lactate dehydrogenase in ascitic fluid, 1525 in ischemic hepatitis, 1379, 1380f Lactic acidosis from nucleoside analogue reverse transcriptase inhibitors, 532-533 in short bowel syndrome, 1790-1791 Lactobacillus biotherapy for antibiotic-associated diarrhea, 1890 for Clostridium difficile-associated diarrhea and colitis, 1901 for diarrhea, 2293 Lactose in breast milk, 1732 dietary, restriction of, 95, 1755 intolerance of, 1709, 1751, 1754-1767 hydrogen breath test for, 1754-1755, 1755f irritable bowel syndrome and, 2098 symptoms of, 1754-1755 malabsorption of, 1754-1767, 1754t diarrhea due to, 213, 224 in elderly persons, 1758 hydrogen breath test for, 1751, 1754-1755, 1755f tolerance testing for, 1751

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

liii

liv

Index Lactulose for constipation, 276-278, 277t for hepatic encephalopathy, 1545 for urea cycle defects, 1273-1274 Ladd procedure, 2116 Ladd’s bands, 1630-1631, 1631f, 2116 Laimer’s triangle, 680 Lamina propria, 1617-1618 in celiac disease, 1800 of esophagus, 667, 668f mononuclear cells of, 28 of stomach, 778 Lamivudine for hepatitis B, 1302-1303 with adefovir, 1305 in liver transplantation, 1599-1600 peginterferon with, 1305-1306 hepatotoxicity of, 1430 in pregnancy, 1306 resistance to, 1302-1303 HBV DNA polymerase mutation and, 1291-1292, 1292f Langerhans’ cell histiocytosis, in infants and children, 1057 Lanreotide, for carcinoid tumors, 489 Lansoprazole adverse effects of, 871 drug interactions with, 871 hepatotoxicity of, 1433 mechanisms of action of, 870 for NSAID ulcer prophylaxis, 874 for NSAID ulcers, 873 for peptic ulcer disease, 870-871 pharmacokinetics of, 870-871 Laparoscopic adhesiolysis, for chronic abdominal pain, 170 Laparoscopic adjustable gastric banding, 115, 116t-117t. See also Bariatric surgery. Laparoscopic cholecystectomy. See Cholecystectomy, laparoscopic. Laparoscopy in acute mesenteric ischemia, 2031 in appendicitis, 2066 in ascites of unknown origin, 622 for gastric cancer, 902 in hepatocellular carcinoma, 1573 for ileal pouch–anal anastomosis, 2024 for ileostomy, 2024 ileus after, 2126 in pancreatic cancer, 1023 for perforated peptic ulcers, 884, 884f in peritoneal disease, 622, 622f for peritonitis, 615-616 for small intestinal obstruction, 2111-2112 staging, 622 Laparotomy hernia following, 389 for peritonitis, 615-616 Laplace, law of, 1495 Larazotide acetate, for celiac disease, 1820 Large intestine. See Colon; Rectum. Larva currens, 367 Larva migrans cutaneous, 367, 367f, 1925-1927, 1926f visceral, 1359 Laryngitis, reflux, 714-715, 715f Larynx reflux-induced changes in, 180-181 in swallowing, 678 Laser therapy for colorectal cancer, 2237 for esophageal cancer, 764-765, 765f Lateral sclerosis, amyotrophic dysphagia in, 688 gastrointestinal manifestations of, 581 Latex agglutination assay, for C. difficile, 1896t, 1897 Laugier-Hunziker syndrome, 363 Laxatives abuse of cathartic colon and, 2245-2246 constipation in, 137

Laxatives (Continued) diarrhea in, 230-231, 230t in eating disorders, 135, 136t anthraquinone, pseudomelanosis coli and, 2246f, 2247 colonic ischemia from, 2040 colonic motility and, 1671-1672 complementary and alternative medicine types of, 2293-2295, 2294t for constipation, 275t, 276-282, 277t ileus and, 2127 for irritable bowel syndrome, 2102 malabsorption with, 1757t in palliative care, 2284, 2285t stimulant, for constipation, 277t, 279 Lead, hepatotoxicity of, 1453 Lecithin for alcoholic liver disease, 1398 in bile, 1093 Lecithin-cholesterol acyltransferase, synthesis of, 1222 Leflunomide, gastrointestinal complications of, 559 Legionellosis, hepatic manifestations of, 1352-1353 Leiomyoma mesenteric, 620 small intestinal, 2145-2146 Leiomyomatosis peritonealis disseminata, 620 Leishmaniasis, visceral hepatic manifestations of, 1356t-1357t, 1358-1359 malabsorption in, 1829 Lenalidomide, for Crohn’s disease, 1970 Leptin, 10 in eating disorders, 123 food intake effects of, 10, 102 in nonalcoholic fatty liver disease, 1403 receptors for, 10 in satiety, 18, 802, 1696 Leptin gene, in obesity, 102 Leptospirosis, anicteric, hepatic manifestations of, 1354 Leucovorin, for colorectal cancer plus 5-fluorouracil, 2232-2233, 2235, 2235f plus 5-fluorouracil and irinotecan, 2232-2233, 2235 plus 5-fluorouracil and oxaliptan, 2232-2233, 2235 Leukemia gastrointestinal manifestations of, 566, 566t hepatic involvement in, 565t, 567 Leukoplakia hairy, in HIV/AIDS, 355, 355f oral, 355, 355f Leukotrienes, in intestinal ion transport, 1690-1691 Leuprolide for functional abdominal pain syndrome, 170 for irritable bowel syndrome, 2103 Levamisole, for colorectal cancer, plus 5-fluorouracil, 2232 Levator ani. See Puborectalis. Levator ani syndrome, 2259, 2272-2273 Levofloxacin, for Helicobacter pylori infection, 842, 843t Lewis antigens, in Helicobacter pylori infection, 834-835 Lexipafant, for acute pancreatitis, 980 Lichen planus esophagus in, 359 in hepatitis C, 366 oral manifestations of, 359, 359f Licorice, for liver disease, 2295t, 2297 Lidocaine methemoglobinemia from, 657 for neuropathic pain, 2283t Lidocaine metabolite formation test, 1235 Lifestyle modifications for alcoholic liver disease, 1394-1395 for constipation, 274-275

Lifestyle modifications (Continued) for gastroesophageal reflux disease, 721 for obesity, 109 Limbic system, in functional abdominal pain syndrome, 166, 166f-167f Linaclotide, for constipation, 281-282 Linton-Nachlas tube, for gastric variceal bleeding, 1512 Lipase congenital absence of, 955-956 developmental changes in, 1729, 1730f gastric, 829, 830f, 1699 in infants, 1699t milk-derived, 1729-1731, 1730f pancreatic, 1699-1700 functions of, 924 in triglyceride digestion, 1699-1701, 1699t, 1700f in pancreatic enzyme preparations, 1005t, 1009 secretion of in acinar cell carcinoma, 1033 secretin- or CCK-induced, 999-1000 serum in acute pancreatitis, 971 in chronic pancreatitis with steatorrhea, 997 Lipid(s). See also Fat(s). in bile hepatic secretion of, 1095-1097, 1096f physical states of, 1093-1094 disorders of. See Dyslipidemia; Hyperlipidemia. duodenal hypersensitivity to, in functional dyspepsia, 188 metabolism of, hepatic, 1218f, 1220-1225, 1223f Lipid droplets, small intestinal, in Whipple’s disease, 1838, 1838f Lipid-lowering therapy gallstone disease and, 1092 for nonalcoholic fatty liver disease, 1410 Lipid rafts, 1207 Lipid transport proteins, 1222 Lipokinetix, hepatotoxicity of, 1458 Lipolysis, decreased, 1738 Lipolytic enzymes, 1222 Lipoma colonic, 2176 esophageal, 770 small intestinal, 2145-2146, 2147f Lipomatosis, pelvic, 619 Lipoprotein classification of, 1221 high-density (HDL), 1221 ApoB-containing, 1224 gallstone disease and, 1092 receptors for, 1224 low-density (LDL), 1221 receptors for, 1224 metabolism of, hepatic, 1221-1222, 1223f receptors for, 1224 transport of, 1222-1224, 1223f very-low-density (VLDL), 1221, 1704 ApoB-containing, 1223-1224 receptors for, 1224 Lipoprotein lipase, synthesis of, 1222 Liquids, gastric emptying of, 798, 798f-799f Lisinopril, hepatotoxicity of, 1433 Listeriosis food-borne, 1879t-1880t, 1882 hepatic manifestations of, 1352 LITH genes, in gallstone disease, 1102, 1103t Lithocholic acid, 1093. See also Bile acid(s). synthesis and metabolism of, 1077f, 1077t, 1078 Lithotripsy, extracorporeal shock wave. See Extracorporeal shock wave lithotripsy. Liu-jun-zi-tang (TJ-43), for functional dyspepsia, 2290, 2290t Liv52, for liver disease, 2297 Liver. See also Hepatic; Liver disease. abscess of, 1366-1369 amebic, 1368-1369, 1368f, 1369t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Liver (Continued) fungal, after hematopoietic stem cell transplantation, 547t, 549, 554 pyogenic, 1366-1368, 1366t, 1367f, 1369t adenoma of. See Hepatocellular adenoma. anatomy of, 1202-1203, 1203f acinus model of, 1204, 1204f lobule model of, 1204, 1205f segmental, 1202, 1203f, 1604, 1604f angiosarcoma of, 1582-1583 drug-induced, 1451-1452 apoptosis in, 1214 bile acid transport in, 1079f, 1080-1083, 1082f, 1083t bioartificial, for acute liver failure, 1568 biochemical tests of. See Liver biochemical tests. biopsy of. See Liver biopsy. blood flow in, 1202-1203, 1489-1491, 1490f carbohydrate metabolism by, 1217-1220, 1218f carcinoma of. See Hepatocellular carcinoma. cells of functional integration of, 1211-1212 matrix interactions with, 1211-1212 regeneration and apoptosis of, 1212-1214, 1213f types and organization, 1207-1211, 1208f cholangiocarcinoma of, 1171, 1172f, 1173, 1175f, 1176, 1580-1582. See also Cholangiocarcinoma. cholesterol regulation by, 1095-1097, 1096f cirrhosis of. See Cirrhosis. congenital anomalies of, 1205-1206, 1205f-1206f cysts of, 1589-1590, 1589f hydatid, 1356t-1357t, 1363-1365, 1364f decompensation of, in hepatitis B, 1294 development of, 1045-1047, 1046f drug metabolism in, 1417-1419 cytochrome P450 enzymes in, 1417 injury with immunologic mechanisms of, 1422 toxic mechanisms of, 1419-1422 mitochondrial electron transport in, 1418 phase 1 (oxidation) pathways in, 1417-1418 phase 2 (conjugation) pathways in, 1418 phase 3 (elimination) pathways in, 1418-1419 ductal plate malformation of, 1205-1206 embryology of, 1201-1202, 1202f energy metabolism in, 1216-1225 enlarged. See Hepatomegaly. extracellular matrix of cell interactions with, 1211-1212 components of, 1212 fatty. See Fatty liver disease. fibrosis of. See Hepatic fibrosis. fructose metabolism by, 1219 galactose metabolism by, 1219 gene expression in, 1214-1216, 1215t granuloma of. See Granuloma, hepatic. hemangioendothelioma of epithelioid, 1583 infantile, 1587 hemangioma of cavernous, 1586-1587, 1586f in pregnancy, 637 hemangiosarcoma of, 1582-1583 heme synthesis in, 1266, 1266f. See also Porphyria. hepatocytes of. See Hepatocytes. hilum of, 1048, 1048f histology of, 1203-1205, 1203f infarction of, 1382 in preeclampsia, 634 infection of, 1351-1370. See also Liver, abscess of. bacterial, 1351-1355 fungal, 1365-1366 in hematopoietic stem cell transplantation candidates, 544 parasitic, 1355-1365, 1355t-1357t

Liver (Continued) inflammation of. See Hepatitis. inflammatory pseudotumor of, 1367, 1588-1589 injury to drug-induced, after hematopoietic stem cell transplantation, 547t, 554 in hereditary hemochromatosis, 1242 postoperative, 1449, 1449t-1450t innervation of, 1203 ischemia of. See Hepatitis, ischemic. Kupffer cells of, 1490, 1490f lipid metabolism by, 1218f, 1220-1225, 1223f lobes of, 1202 lobules of, 1204, 1205f lymphatics of, 1203 lymphoma of, 459, 565-566, 565t macroregenerative nodules of, 1588 mass lesion of, approach to, 1590-1592, 1591f metabolic function of, inherited disorders of. See Inborn errors of metabolism. metastasis to, 1583-1584 ascites in, 1517 in carcinoid tumor, 488 in colorectal cancer, 2231-2232 in hematologic malignancies, 564-567, 565t nodular hyperplasia of, 590-591 focal, 1587-1588, 1588f in pregnancy, 637 lobar or segmental, 591 regenerative, 591, 591f, 1588 oxygen supply to, 1202-1203 parenchyma of cells types in, 1207-1210, 1208f organization of, 1204-1205, 1204f-1205f partial nodular transformation of, portal hypertension in, 1500 perisinusoidal cells of, 1211 physiology of, 1207-1226 pit cells of, 1211 in pregnancy, 626 protein synthesis and degradation in, 1214-1216, 1215t regeneration of, 1212-1214, 1213f apoptosis in, 1214 cell cycle genes in, 1213 cytokine and growth factor integration in, 1213-1214 delayed early genes in, 1213 gene expression during, 1212-1213 hepatic growth factor and c-met in, 1214 immediate early genes in, 1212-1213 retransplantation of, 1611-1612, 1611f, 1611t Riedel’s lobe of, 1205 rupture of, in preeclampsia, 634, 634f sarcoidosis of, 588-590. See also Sarcoidosis. sarcomas arising in, 1583 sinusoids of anatomy of, 1203-1204, 1489-1490, 1490f in bile acid transport, 1081-1083 development of, 1202 dilatation of, in drug-induced liver disease, 1445t endothelial cells of, 1203-1204, 1210, 1490, 1490f Kupffer cells of, 1203-1204, 1211 obstruction of, 1375-1377. See also Sinusoidal obstruction syndrome. space of Disse of, 1204, 1207, 1490, 1490f space of Mall of, 1204 steatosis of. See Steatosis, hepatic. stellate cells of, 1203-1204, 1211, 1490, 1490f in portal hypertension, 1492 transplantation of. See Liver transplantation. tumor-like lesions of, 1587-1589 tumors of algorithm for, 1590-1592, 1591f benign, 1584-1587 drug-induced, 1445 malignant, 1569-1583 metastatic, 1583-1584 vascular development of, 1202 vasoregulation of, 1490-1491

Liver (Continued) von Meyenburg complexes of, 1590 zones of, 1204 Liver biochemical tests, 1227-1238 in cystic fibrosis, 946-947 for drug hepatotoxicity screening, 1236 in hepatic fibrosis, 1234-1235 of hepatic synthetic function, 1234 in HIV/AIDS, 532-535, 532t in methotrexate hepatotoxicity, 1444 in nonalcoholic fatty liver disease, 1405 for organ allocation, 1236-1237 in pregnancy, 626 in primary biliary cirrhosis, 1480 quantitative, 1235 for surgical candidacy, 1236-1237 in ulcerative colitis, 1984 Liver biopsy in acute fatty liver of pregnancy, 635, 635f in alcoholic liver disease, 1384f, 1390 in amyloidosis, 586 in α1-antitrypsin deficiency, 1261 in congestive hepatopathy, 1380, 1381f in HELLP syndrome, 633, 633f in hepatitis C, 1322-1324, 1322f-1323f, 1322t in hereditary hemochromatosis, 1244, 1245f in HIV/AIDS, 535 in jaundice, 334, 334f in lymphoma, 566 in methotrexate hepatotoxicity, 1444 in mitochondrial liver disease, 1278 for monitoring drug toxicity, 366-367, 367t in neonatal cholestasis, 1052 in nonalcoholic fatty liver disease, 1407-1408, 1408t in peliosis hepatis, 1381, 1381f in primary biliary cirrhosis, 1481 in primary sclerosing cholangitis, 1161, 1161f in schistosomiasis, 1937 in sickle cell anemia, 570, 570f Liver disease. See also Hepatobiliary disorders. alcoholic. See Alcoholic liver disease. in amyloidosis amyloid deposits in, 586, 586f clinical and laboratory findings in, 587 diagnosis of, 587-588 treatment and prognosis in, 588 in α1-antitrypsin deficiency, 1260-1262 autoimmune, in pregnancy, 637 bile acid levels in, 1236 in bile acid synthesis defects, 1274-1276, 1275t in bile acid transport defects, 1275t, 1276-1277 cholestatic. See Cholestasis. chronic, in hematopoietic stem cell transplantation candidates, 545 complementary and alternative medicine for, 2295-2297, 2295t cutaneous manifestations of, 365-366, 366f in cystic fibrosis, 945-947, 1277, 1277t clinical features of, 947 functional abnormalities in, 946-947 pathology of, 946, 946f radiologic features of, 946 treatment of, 947 drug-induced, 1413-1446. See also Hepatotoxicity. acute liver failure in, 1558-1559, 1558t-1559t, 1559f age and, 1415 alanine aminotransferase (ALT) and, 1415 alcohol and, 1416 from botanical and environmental hepatotoxins, 1454-1455, 1457t case definition of, 1414 from chemical agents, 1450-1453, 1450t-1451t classification of, 1422-1423, 1423t clinical features of, 1424 clinicopathologic features of, 1422-1424, 1423t dechallenge in, 1424 definitions in, 1413-1414

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lv

lvi

Index Liver disease (Continued) diagnosis of, 1424-1426 chronologic relationships in, 1425 extrahepatic features, 1425 liver biopsy in, 1425 physician awareness and, 1424-1425 viral hepatitis and, 1425-1426 dose-dependent, 1427-1431 epidemiology of, 1414-1417, 1416t frequency of, 1414-1415 gender and, 1415-1416 genetic factors in, 1415 hepatitis in. See Hepatitis, drug-induced. from herbal medicines, 1456-1458, 1457t histopathology of, 1424 HLA haplotype and, 1415 immunoallergic, 1431-1434, 1431t importance of, 1413-1415 latent period in, 1424 liver biopsy in, 1425 liver tests for, 1236 management of, 1426-1427 metabolic idiosyncrasy in, 1429, 1431t, 1434-1437 from metals, 1453 mortality in, 1413-1414 multiple drug use and, 1416 nutritional status and, 1417 pathophysiology of, 1417-1422 apoptosis in, 1420, 1421f biochemical mechanisms in, 1419-1420 cytochrome P450 enzymes in, 1417 for direct hepatotoxins, 1419 endothelial cells in, 1422 glutathione system in, 1419 immunologic mechanisms in, 1422 Kupffer cells in, 1422 necrosis in, 1420-1421, 1421f oxidative stress in, 1419, 1421 phase 1 (oxidation) pathways in, 1417-1418 phase 2 (conjugation) pathways in, 1418 phase 3 (elimination) pathways in, 1418-1419 reactive metabolites in, 1419 stellate cells in, 1422 toxic mechanisms in, 1419-1422 preexisting liver disease and, 1417 preexisting nonhepatic disease and, 1417 prevention of, 1426-1427 previous drug reactions and, 1416 reactive metabolite syndrome in, 1424 rechallenge in, 1424 risk factors for, 1415-1417, 1416t in skin disease, 366-367, 367t vascular disorders in, 1444-1445, 1445t encephalopathy in. See Hepatic encephalopathy. end-stage, prognostic criteria for, 2278-2279, 2280t fatty. See Fatty liver disease. in Felty’s syndrome, 559 fibrocystic, 1589-1590, 1589f in glycogen storage disease, 1262-1265 in glycosylation disorders, 1265-1266, 1265f granulomatous. See Granuloma, hepatic. in hematologic malignancies, 564-567, 565t after hematopoietic stem cell transplantation, 546-549, 547t, 548f hematopoietic stem cell transplantation candidates and, 545 in Hodgkin’s disease, 565, 565t infiltrative, jaundice in, 327-328 inflammatory. See Hepatitis. ischemic. See Hepatitis, ischemic. jaundice in, 325t, 326-329 in leukemia, 565t, 567 lipid abnormalities in, 1225 in lymphoma, 565-566, 565t metabolic. See Inborn errors of metabolism. mitochondrial, 1277-1278 in myeloproliferative disorders, 568 nutritional therapy for, 83

Liver disease (Continued) obesity and, 104-105 operative risk in, assessment of, 1236-1237 parenteral nutrition causing, 86 polycystic, 1589-1590, 1589f liver transplantation for, 1603-1604 in porphyrias, 1266-1269, 1266f, 1267t postoperative, 1449, 1449t pruritus in, 366 pulmonary syndromes in, 1549-1552. See also Hepatopulmonary syndrome. radiation-induced, 650-651 recurrent, after solid organ transplantation, 540, 540f renal syndromes in, 1546-1549. See also Hepatorenal syndrome. in rheumatoid arthritis, 558-559 in sarcoidosis, 588-590 clinical manifestations of, 589-590 granulomas in, 588, 589f, 589t severity of, scoring systems for, 1236 in short bowel syndrome, 1790 in sickle cell anemia, 570-572, 570f-571f in tyrosinemia, 1269-1271, 1270f in urea cycle defects, 1271-1274, 1272f, 1273t vascular, 1371-1382. See also specific disorders, e.g. Budd-Chiari syndrome. veno-occlusive. See Sinusoidal obstruction syndrome. Liver enzymes in eating disorders, 130-131, 137 in HELLP syndrome, 632, 632t in sickle cell anemia, 571, 571f in systemic lupus erythematosus, 561 Liver failure acute, 1557-1568 acetaminophen-induced, 1558-1559, 1558t-1559t, 1559f, 1563, 1563t-1564t cerebral edema in, 1561-1562, 1561t, 1565-1566 clinical features of, 1559t, 1561-1563, 1561t coagulopathy in, 1561t, 1562, 1566 definition of, 1557-1558 drug-induced, 1558-1559, 1558t-1559t, 1559f etiology of, 1558-1560, 1558t extracorporeal liver support for, 1568 gastrointestinal bleeding in, 1561t, 1562, 1566 hepatic encephalopathy in, 1561-1562, 1561t, 1565-1566 in hepatitis A, 1558t-1559t, 1559f, 1560 in hepatitis B, 1558t-1559t, 1559f, 1560 hepatocyte transplantation for, 1568 histology in, 1563, 1564f hypoglycemia in, 1561t hypotension in, 1561t, 1562-1563 indeterminate, 1558t-1559t, 1559f, 1560 infection in, 1561t, 1562, 1566 initial evaluation and management of, 1564-1565 liver transplantation for, 1567, 1567t, 1594, 1601, 1601t auxiliary, 1567-1568 criteria for, 1563, 1563t live donor, 1567-1568 multiple organ failure syndrome in, 1561t, 1562, 1566-1567 N-acetylcysteine for, 1565 neonatal, in mitochondrial liver disease, 1278 outcomes of, 1559f pancreatitis in, 1561t, 1562-1563 in pregnancy, 1560 prognosis in, 1563, 1563t-1564t, 1564f renal failure in, 1561t, 1563 respiratory failure in, 1561t, 1563 in sickle cell anemia, 571-572 treatment of, 1563-1567, 1564f in tyrosinemia, 1270 viral causes of, 1558t-1559t, 1559f, 1560 in Wilson disease, 1251-1252, 1560 fulminant in Budd-Chiari syndrome, 1373 in hepatitis A, 1280, 1282

Liver failure (Continued) in hepatitis B, 1288, 1295 in hepatitis C, 1325 in short bowel syndrome, 1790 Liver flukes, 1356t-1357t, 1361-1363, 1361f, 1934-1935, 1934t, 1935f Liver-kidney microsomal antibody (LKM) type 1, in autoimmune hepatitis, 1461, 1465 Liver transplantation, 1593-1612 for acetaminophen poisoning, 1428-1429 acute cellular rejection in, 1607-1608, 1607f versus recurrent hepatitis C, 1608, 1609t for acute liver failure, 1567, 1567t, 1594, 1601, 1601t auxiliary, 1567-1568 criteria for, 1563, 1563t live donor, 1567-1568 age and, 1597 for Alagille’s syndrome, 1062 for alcoholic liver disease, 1398-1399, 1399f, 1598-1599 for amyloidosis, 588 anastomotic stricturing after, 1608-1609 for α1-antitrypsin deficiency, 1261 for ascites, 1538 ascites after, 1528 for autoimmune hepatitis, 1475-1476, 1603 for biliary atresia, 1056 biliary complications of, 1608-1609 biliary strictures after, 1188 for Budd-Chiari syndrome, 1375, 1603 cancer after, de novo, 1610 for cholangiocarcinoma, 1176-1177, 1596, 1602-1603 for cholestatic liver disease, 1601-1602 complications of, 540, 540f, 1606-1609, 1607f, 1608t-1609t contraindications to, 1595-1598, 1596t coronary artery disease and, 1596 cytomegalovirus infection after, 1608 diabetes mellitus after, 1608, 1608t, 1610 donor for, 1604 evaluation process for, 1598, 1598t fungal infections after, 1609 for glycogen storage disease type I, 1264 for hemochromatosis, 1603 hepatic artery thrombosis after, 1606-1607 for hepatitis B, 1599-1600 for hepatitis C, 1600-1601, 1600t, 1601f hepatitis C after, 1326, 1334, 1600-1601, 1600t, 1601f, 1607-1608, 1609t for hepatoblastoma, 1602 for hepatocellular adenoma, 1603-1604 for hepatocellular carcinoma, 1578-1579, 1596, 1602 for hepatopulmonary syndrome, 1552, 1596-1597 for hepatorenal syndrome, 1549 for hereditary hemochromatosis, 1246 hyperlipidemia after, 1610 hypertension after, 1609-1610 immunosuppressants for, 1605-1606, 1606t indications for, 1594-1595, 1594f, 1594t disease-specific, 1598-1604 infection status and, 1597 infectious complications of, 1607, 1607f, 1608t listing for, 1595, 1598 live-donor, 1605, 1605t long-term management of, 1609-1612 general preventive medicine in, 1609-1610 immunizations and bacterial prophylaxis in, 1610-1611 retransplantation in, 1611-1612, 1611f, 1611t when to call transplant center in, 1611 lymphoproliferative disorders after, 1610 MELD score in, 1595, 1598, 1599f, 1602 for metabolic disorders, 1603 native hepatectomy in, 1604-1605 neurologic dysfunction after, 1608, 1608t for nonalcoholic fatty liver disease, 1410, 1603 nutritional status and, 1597-1598 opportunistic infections after, 1609 osteopenia after, 1610

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Liver transplantation (Continued) for pancreatic endocrine tumors, 522 pancreatitis after, 968 pediatric indications for, 1594 lymphoproliferative disorders after, 1610 Pneumocystis jirovecii prophylaxis after, 1609 for polycystic liver disease, 1603-1604 for portal vein thrombosis, 1379 for portopulmonary hypertension, 1552, 1596-1597 postoperative course in, 1606-1609 following discharge from hospital, 16081609, 1609t initial phase to discharge from hospital, 1606-1608, 1607f, 1608t pregnancy after, 638 for primary biliary cirrhosis, 1487, 1595, 1601-1602 primary graft nonfunction after, 1606-1607 for primary sclerosing cholangitis, 1163, 1166-1167, 1595, 1601-1602 pulmonary disease and, 1596-1597 renal insufficiency and, 1597, 1608, 1608t for sinusoidal obstruction syndrome, 1603 surgical aspects of, 1604-1605, 1604f T-tube removal after, 1609 for tyrosinemia, 1271 UNOS policies for, 1595 for urea cycle defects, 1274 vascular disease and, 1597 for vascular disorders, 1603 weight gain after, 1610 in well-compensated cirrhosis, 1594-1595, 1595f for Wilson disease, 1603 Liver X receptors, in fat digestion and absorption, 1705-1706, 1705t, 1706f Lobectomy, hepatic, for cholangiocarcinoma, 1176 Locus ceruleus, in stress-immune response, 345-346 Long objects, as foreign bodies, 402 Loperamide for diarrhea, 226, 227t for fecal incontinence, 252-253 for infectious diarrhea, 1886 intestinal dysmotility from, 2139 for irritable bowel syndrome, 2102 for malabsorption-related diarrhea, 72 for shigellosis, 1861 for short bowel syndrome, 1784, 1784t for traveler’s diarrhea, 1874 Lorazepam, for gastric motility disorders, 812, 813t Lord’s procedure, for hemorrhoids, 2261 Losartan hepatotoxicity of, 1433 for portal hypertension, 1502-1503 Lovastatin, hepatotoxicity of, 1433 Low-density lipoprotein (LDL), 1221 receptors for, 1224 Low-density lipoprotein scavenger receptor, 1224 Lowe oculocerebral syndrome, malabsorption in, 1759t-1762t Lubiprostone colonic motility and, 1672 for constipation, 277t, 280 ileus and, 2127 for irritable bowel syndrome, 2102 Lugano staging system, for gastrointestinal lymphoma, 447, 447t Lugol’s iodine, in esophageal cancer, 754 Lumbar hernia, 392, 392f Lumiracoxib hepatotoxicity of, 1437 for NSAID ulcer prophylaxis, 874-875 Lundh test meal, 928t, 929 Lung. See also Pulmonary. in Crohn’s disease, 1955 in hepatopulmonary syndrome, 1550 infection of. See Pneumonia. metastasis to, in colorectal cancer, 2232

Lung (Continued) in schistosomiasis, 1937 transplantation of complications of, 540-541 lymphoproliferative disorders after, 459 Lupus enteritis, 2136 Lupus erythematosus chronic intestinal pseudo-obstruction in, 2136 cutaneous manifestations of, 363 gastrointestinal manifestations of, 558t, 560-561 malabsorption in, 1758 pancreatitis in, in children, 939 protein-losing gastroenteropathy in, 440-441, 441f Reye’s syndrome in, 1430-1431 vasculitis in, 2047 Lupus serositis, ascites in, 1536 Luschka, ducts of, 1050 Luteinizing hormone, tumors secreting, 515 LXR response element, in fat digestion and absorption, 1705, 1706f Lycopene, 1722 Lye ingestion, esophageal cancer after, 408, 747-748 Lyme disease, hepatic manifestations of, 1354-1355 Lymphadenectomy, for gastric cancer, 902 Lymphadenopathy, in Whipple’s disease, 1836 Lymphangiectasias, intestinal, protein-losing gastroenteropathy in, 441, 441f, 443 Lymphangiogenesis, in metastasis, 43 Lymphatic disorders, intestinal, protein-losing gastroenteropathy in, 438t, 441, 441f Lymphatic system anal, 2257 biliary, 1049 colonic, 1617, 1623 duodenal, 779 esophageal, 667 gallbladder, 1049-1050 gastric, 775 hepatic, 1203 intestinal, 1617, 1623 development of, 1626 pancreatic, 912 small intestinal, 1617, 1623 in Whipple’s disease, 1836-1838 Lymphocytes B. See B cell(s). intraepithelial, 28 in celiac disease, 1798, 1800, 1803-1804 lamina propria, 28 migration of, 25, 26f T. See T cell(s). Lymphocytic colitis, 2239-2242 background on, 2239 clinical and laboratory features of, 2241 differential diagnosis of, 2241-2242 epidemiology of, 2240 etiology and pathogenesis of, 2240-2241 pathology of, 2240, 2240f treatment of, 2242, 2242f Lymphocytic gastritis, 853-854, 854f Lymphoid colonic polyps, benign, 2176 Lymphoid hyperplasia in appendicitis, 2060-2061 nodular, 2188 Lymphoid infiltration, diffuse, chronic intestinal pseudo-obstruction in, 2139-2140 Lymphoid tissue, gut-associated. See Gutassociated lymphoid tissue (GALT). Lymphoma with anti-TNF agents, 1968-1969 appendiceal, 2070 background on, 445-446, 446f biliary, 1184 Burkitt’s, small intestinal, 454-455, 455f in celiac disease, 1818 colorectal, 459 in Crohn’s disease, 1972-1973 effusion, in HHV-8, 460 esophageal, 768

Lymphoma (Continued) follicular, 445-446 small intestinal, 453-454, 454f gastric, 447-453, 904 diffuse large B cell, 450f, 451-453, 452t clinical features of, 452 epidemiology of, 451 etiology and pathogenesis of, 451 pathology of, 451-452 treatment of, 452-453, 452t marginal zone B cell (MALT), 447-453 antigen-driven B cell proliferation in, 448 clinical features of, 449-450 diagnosis of, 449-450, 450f epidemiology of, 447-448 etiology and pathogenesis of, 448-449 genetic studies in, 448 gross appearance and location of, 449 Helicobacter pylori infection in, 448 histology of, 449, 449f immunophenotype in, 449 monoclonality tests in, 449 pathogenesis of, model for, 448-449 pathology of, 449 staging and prognostic assessment of, 447t, 450 treatment of, 450-451, 450t T cell, 453 gastrointestinal, 445-460 diagnosis of, 446-447 overview of, 445-446, 446f sites of, 446, 446t staging and prognostic assessment of, 447, 447t treatment of, 447 hepatic, 459, 534, 565-566, 565t in hepatitis C, 1321 hepatosplenic, 566 in HIV/AIDS, 356, 459-460, 534 Hodgkins’s. See Hodgkin’s disease. immunodeficiency-related, 459-460 mantle cell, 445-446, 2188 small intestinal, 453, 454f pancreatic, 459, 1033 small intestinal, 453-459, 2145, 2146f Burkitt’s, 454-455, 455f diffuse large B cell, 453 follicular, 453-454, 454f immunoproliferative, 455-457 clinical features of, 456-457 diagnosis and staging of, 457 epidemiology of, 455 etiology and pathogenesis of, 455 malabsorption in, 1829t, 1830-1831 pathology of, 455-456, 456t treatment of, 457 mantle cell, 453, 454f marginal zone B cell (MALT), 453 non-immunoproliferative, 453 T cell enteropathy-type. See Enteropathyassociated T cell lymphoma. natural killer type, 459 tropical malabsorption and, 1830-1831 Waldeyer’s ring, 459 Whipple’s disease and, 1841 Lymphomatous polyposis, 2188 in mantle cell lymphoma, 453, 454f Lymphoplasmacytic sclerosing pancreatitis, 986-987, 991, 992f Lymphoproliferative disorders after hematopoietic stem cell transplantation, 549 after liver transplantation, 1610 after solid organ transplantation, 539, 539f, 543 Lynch’s syndrome I, 2206 Lynch’s syndrome II, 2206, 2207f Lysinuria, isolated, malabsorption in, 1759t-1762t Lysinuric protein intolerance, malabsorption in, 1759t-1762t, 1763 Lysosomal proteolysis, 1216

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lvii

lviii

Index Lysosomes, 924-925 of hepatocytes, 1209

M

M cells intestinal, 1619 oral tolerance and, 23, 23f in Peyer’s patches, 25, 26f M phase of cell cycle, 31, 32f Ma-huang, hepatotoxicity of, 1458 Macroamylasemia, 971 Macroglossia, in amyloidosis, 361-362 Macroglycogen, 1219 Macrophages, in malakoplakia, 2250 Macroregenerative hepatic nodules, 1588 MADH4 (SMAD4) gene, 40 in colorectal cancer, 2200t, 2202 in juvenile polyposis, 2187 Magnesium. See also Hypomagnesemia. absorption of, 1723 for constipation, 276, 277t diarrhea from, 212-213, 224 dietary, 52t, 80, 1723 intestinal ion transport and, 1689 for malabsorption, 74 malabsorption of, 1740, 1759t-1762t in refeeding syndrome, 70 in short bowel syndrome, 1781t, 1782, 1784 Magnesium citrate for colonic health, 2294t colonic ischemia from, 2040 Magnesium hydroxide, for constipation, 276, 277t Magnesium sulfate, for constipation, 276, 277t Magnetic resonance angiography, in portal vein thrombosis, 1378-1379, 1378f Magnetic resonance cholangiography, in gallstone disease, 1108t, 1111-1112, 1112f Magnetic resonance cholangiopancreatography, 1185-1186, 1186f in cholangiocarcinoma, 1174, 1175f in chronic pancreatitis, 1001 contraindications to, 1185-1186 contrast-enhanced, 1185-1186 in intrahepatic cholangiocarcinoma, 1581 in jaundice, 332, 333t in liver transplant recipients, 1608-1609 mangafodipir trisodium in, 1185-1186 in neonatal cholestasis, 1052 in primary sclerosing cholangitis, 1154, 1155f in recurrent pyogenic cholangitis, 1169 in sphincter of Oddi dysfunction, 1069 Magnetic resonance elastography, in hepatic fibrosis, 1235 Magnetic resonance enteroclysis, in malabsorption, 1749 Magnetic resonance enterography, in Crohn’s disease, 1957, 1957f Magnetic resonance imaging in abdominal abscess, 414-415 in acute pancreatitis, 972-973 in adenomyomatosis, 1148-1149 in Budd-Chiari syndrome, 1373 in carcinoid tumors, 485 in cavernous hemangioma of liver, 1586, 1586f in celiac disease, 1811 in cholangiocarcinoma, 1174 in cystic fibrosis–related liver disease, 946 in diverticulitis, 2080 in fecal incontinence, 248 in gastric cancer, 901-902 in gastric motility assessment, 804 in gastrointestinal stromal tumors, 466 in hepatic encephalopathy, 1545 in hepatocellular carcinoma, 1573, 1574f in malabsorption, 1749 in nonalcoholic fatty liver disease, 1405, 1406f in pancreatic cancer, 1022 in pancreatic endocrine tumors, 515-516, 516t in portal hypertension, 1498 in pregnancy, 627

Magnetic resonance imaging (Continued) in protein-losing gastroenteropathy, 442 in small intestinal motility measurement, 1652-1653 Magnetic resonance pancreatography, in sphincter of Oddi dysfunction, 1069 Magnetic stimulation, translumbar, anal motor evoked potentials after, 249, 250f Magnets, ingestion of, 403 Major histocompatibility complex (MHC). See also HLA (human leukocyte antigen). in Helicobacter pylori infection, 835 in primary sclerosing cholangitis, 1156 Malabsorption, 71-75, 1735-1768. See also Digestion and absorption; Short bowel syndrome. in amino acid transport defects, 1758-1763, 1759t-1762t in amyloidosis, 1756 assessment of, 71-72, 71t after bariatric surgery, 1733, 1758 bile acid. See Bile acid(s), malabsorption of. carbohydrate. See Carbohydrate, malabsorption of. clinical features of, 1741-1754, 1742t-1743t colon in, 1740-1741, 1740f-1741f in common variable immunodeficiency, 1764 congenital defects causing, 1758-1763, 1759t-1762t in connective tissue disorders, 1758 in Cronkhite-Canada syndrome, 2188 definition of, 1735 in diabetes mellitus, 1766 diarrhea in, 218 dietary supplements causing, 1756, 1757t in disaccharidase deficiency, 1759t-1762t, 1763 drug-induced, 1756, 1757t in elderly persons, 1758 in enteric anendocrinosis, 1765, 1766f in enteroendocrine deficiency, 1765, 1766f etiology, 1736-1737, 1736t evaluation of, 1741-1754 abdominal imaging in, 1749 anatomic investigations in, 1746 aspiration in, 1748 biopsy in, 1746-1747, 1747f-1748f, 1747t, 1766f blood tests in, 1742, 1743t carbon-13 breath tests for, 1754 clinical clues in, 1742-1746, 1743t computed tomography in, 1749 endoscopic retrograde cholangiopancreatography in, 1749 endoscopic ultrasonography in, 1749 endoscopy in, 1746, 1746f history in, 1742, 1742t intestinal permeability test for, 1753-1754 laboratory findings in, 1742, 1743t-1745t magnetic resonance imaging in, 1749 noninvasive absorptive and digestive functional, 1749-1754, 1750t pancreatic enzymes in, 1742, 1742t radiography in, 1749 signs and symptoms in, 1742t small bowel enteroclysis in, 1749 small bowel radiologic series in, 1749 ultrasonography in, 1749 videocapsule endoscopy in, 1748 xylose test for, 1753 fat. See Fat(s), malabsorption of. after gastric bypass, 1758 after gastric resection, 1756-1757 in giardiasis, 1746-1747, 1747f, 1913 in hyperthyroidism and autoimmune thyroid disease, 1765-1766 with ileostomy, 2016 in immunodeficiency syndromes, 1763-1764 in immunoglobulin A deficiency, 1763 in intestinal enterokinase deficiency, 1759t-1762t, 1763 intestinal transit time and, 1741 in IPEX syndrome, 1764 in lipid congenital disorders, 1759t-1762t, 1763

Malabsorption (Continued) in lupus erythematosus, 1758 management of, 72-75 diarrhea control in, 72 enteral feeding for, 72-74 fat intake in, 73 general approach to, 1767 mineral supplementation in, 74 oral rehydration therapy in, 73-74, 73t parenteral feeding in, 74 predigested formulas in, 73 vitamin supplementation in, 74 mechanisms compensating for, 1740-1741 micronutrient requirements in, 58-59 in neurofibromatosis type 1, 1764 in nongranulomatous chronic idiopathic enterocolitis, 1764-1765, 1765f in osteomalacia, 1766-1767 in osteoporosis, 1766-1767 pathophysiology of, 1736-1737, 1737t postinfection, 1823-1824 protein. See Protein(s), malabsorption of. in scleroderma, 560, 1758 in short bowel syndrome, 1779-1783, 1780f in small intestinal bacterial overgrowth, 1774 tropical. See Tropical malabsorption. in Whipple’s disease, 1746-1747, 1747f, 1836 in X-linked infantile agammaglobulinemia, 1764 Malakoplakia, colonic, 2250-2251, 2251f Malaria hepatic manifestations of, 1355-1358, 1356t-1357t hyperreactive splenomegaly in, 1358 pathobiology of Plasmodium life cycle in, 1355, 1355f Maldigestion. See also Malabsorption. in chronic pancreatitis, 996-997, 1009-1010 definition of, 1735 diarrhea from, 225 micronutrient requirements in, 59 Mall, space of, 1204 Mallory bodies, 1208 Mallory-Weiss syndrome, 740 Mallory-Weiss tear upper gastrointestinal bleeding in, 304-305, 304f from vomiting, 206 Malnutrition. See also Nutritional deficiency. in alcoholic liver disease, 1395 in chronic intestinal pseudo-obstruction, 2140 in Crohn’s disease, 1953 in cystic fibrosis, 948-949 drug-induced liver disease and, 1417 protein-calorie. See Protein-energy malnutrition. Malone enema, for fecal incontinence, 256 Malonyl-CoA, in hepatic fatty acid synthesis, 1220 Malrotation, 780t, 787, 2116, 2116f MALT. See Mucosa-associated lymphoid tissue (MALT). Maltase, 1708, 1708f, 1708t Maltose, 1707, 1707f Maltotriose, 1707, 1707f Malunions, pancreaticobiliary, anomalous, 919 Mandibular osteoma, in familial adenomatous polyposis, 2180-2182 Mangafodipir trisodium, in magnetic resonance cholangiopancreatography, 1185-1186 Manganese dietary, 56t-57t, 80t, 81 excess of, 58 Manipulative and body-based practices, 2287 Mannitol for constipation, 277t, 278 for intracranial hypertension, 1565-1566 Mannose, 1265, 1265f Manometry anorectal in constipation, 273 in fecal incontinence, 247-248, 247f, 251t in Hirschsprung’s disease, 1638-1639, 1638f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Manometry (Continued) antroduodenal, for gastric motility assessment, 795f, 804 esophageal, 694-698, 699t conventional, 694, 695f in gastroesophageal reflux disease, 719 high-resolution. See Esophageal pressure topography, high-resolution. of upper esophageal sphincter function, 678-679, 679f in vomiting, 205 gastrointestinal in chronic intestinal pseudo-obstruction, 2142 in functional vomiting, 202 in mechanical obstruction, 2142 in rumination, 204 in vomiting, 206 small intestinal, 1652-1653 interdigestive, 1655f, 1656 postprandial, 1655f, 1656 sphincter of Oddi, 1070-1071, 1070f in visceral myopathy/neuropathy, 2142 Mantle cell lymphoma, 445-446, 2188 small intestinal, 453, 454f Marasmus, in children, 61, 62t Marfan’s syndrome, gastrointestinal manifestations of, 558t, 563 Marshmallow, for colonic health, 2294t Mast cell(s), in irritable bowel syndrome, 2097-2098 Mast cell inhibitors, for eosinophilic gastrointestinal disorders, 434 Mast cell mediators, in intestinal ion transport, 1689 Mastocytosis cutaneous manifestations of, 362-363, 362f gastric acid secretion in, 828 gastrointestinal manifestations of, 567-568, 568f gastrointestinal ulcers in, 864 Matrix metalloproteinases, in colorectal cancer, 2204 McBurney’s point, in appendicitis, 2060 MDC-CCL2, in gut-associated lymphoid tissue, 29 MDR1 gene, in ulcerative colitis, 1977 MDR3 disease, 1276 Mean arterial pressure, in acute liver failure, 1566 Mean arterial pulmonary pressure, in portopulmonary hypertension, 1551 Mean corpuscular volume (MCV) in alcohol abuse, 1389, 1390f in gastrointestinal bleeding, 288 Meandering artery, 2028, 2029f Measles, gastritis in, 849 Meat consumption, colorectal cancer and, 2194 Mebendazole for ascariasis, 1360, 1922-1923 for capillariasis, 1925 for hookworm infection, 1926-1927 for pinworm infection, 1929 for trichinosis, 1361, 1930 for Trichuris trichiura infection, 1928 Mechanical ventilation in acute liver failure, 1565 gastrointestinal complications in, 581-584, 582f Meckel’s diverticulum, 1628-1630, 1629f bleeding in, 317-318, 1628-1629 intestinal obstruction in, 1629 Meckel’s scans, 292 Meconium ileus, 1633, 1641 in cystic fibrosis, 943-944, 944f, 944t Meconium plug, Hirschsprung’s disease and, 1639 Meconium plug syndrome, 1641 Medicare Hospice Benefit, 2277-2278 Medication intolerance, dyspepsia in, 184, 185t Medicolegal considerations, in endoscopy, 656-657 Meditation, 2288t Mediterranean fever, familial gastrointestinal manifestations of, 558t, 563-564 peritonitis in, 618

Medulloblastoma oropharyngeal dysphagia in, 688 in Turcot’s syndrome, 2184 Megacolon, 2143-2144 acquired, 2144 in Chagas’ disease, 1918, 1919f congenital, 2144. See also Hirschsprung’s disease. in diabetes mellitus, 575 toxic necrotizing colitis with, in amebic dysentery, 1909 in ulcerative colitis, 1992, 2005 Megarectum, 2143-2144 Megestrol acetate, for anorexia-cachexia syndrome, 2284 Meissner’s plexus. See Submucosal plexus. Melanin-concentrating hormone, food intake effects of, 102-103 Melanocortin-4 receptor gene, in obesity, 102 Melanoma anal, 2270 malignant biliary, 1184 esophageal, 767-768 metastasis of to gastrointestinal tract, 564, 564f to small intestine, 2153, 2153f Melatonin, in gastrointestinal tract, 11 MELD. See Model for End-stage Liver Disease (MELD). Melena bleeding site in, 288 definition of, 285 Melioidosis, hepatic manifestations of, 1353 Melkersson-Rosenthal syndrome, 359 Meltzer-Lyon test, in acalculous biliary pain, 1141 Ménétrier’s disease gastric cancer in, 897 gastropathy in, 858-859, 859f protein-losing gastroenteropathy in, 437, 440 treatment of, 443 Menin in carcinoid tumors, 478 in multiple endocrine neoplasia, 494 Meningismus, in shigellosis, 1859-1860 Menkes disease gene for, 1249 malabsorption in, 1759t-1762t Menstrual irregularities, in eating disorders, 128 Mental health referral, in functional abdominal pain syndrome, 170 Meperidine angioectasia endoscopic appearance and, 597-598 in palliative care, 2281t Mepolizumab, for eosinophilic gastrointestinal disorders, 434 6-Mercaptopurine for Crohn’s disease, 1962t, 1964-1966 lymphoma and, 1965-1966 metabolism of, 1964-1965, 1964f, 1996 for perianal Crohn’s fistula, 2269 in pregnancy, 630 side effects of, 1997-1998, 1997t for ulcerative colitis, 1996-1998, 1997t Mesalamine for Crohn’s disease, 1961, 1962t hepatotoxicity of, 1432 in pregnancy, 630 for ulcerative colitis, 1993-1995, 1993f, 1994t Mesenchymal tumors, biliary, 1183 Mesenchymyopathies, chronic intestinal pseudo-obstruction in, 2133 Mesenteric adenitis, versus appendicitis, 2063t Mesenteric adipose tissue, hypertrophy of, in Crohn’s disease, 1949 Mesenteric angioplasty, percutaneous transluminal, for intestinal angina, 2045-2046 Mesenteric artery inferior anastomotic circulation of, 2028, 2029f anatomy of, 2028, 2029f

Mesenteric artery (Continued) superior, 1617 anastomotic circulation of, 2028, 2029f anatomy of, 2027, 2028f embolus of, 2030, 2034-2035, 2034f thrombosis of, 2030, 2035 Mesenteric artery syndrome, superior, 608, 608f in anorexia nervosa, 131 vomiting in, 203, 203f Mesenteric hernia, 392-395, 393f Mesenteric ischemia acute, 2029-2036, 2030t abdominal pain in, 154t, 159-160, 2030 clinical features of, 2030 complications of, 2035 diagnosis of, 2030-2032, 2031f incidence of, 2030, 2033f laboratory findings in, 2030-2032 nonocclusive, 2030, 2035, 2035f results in, 2035-2036 superior mesenteric artery emboli in, 2030, 2034-2035, 2034f superior mesenteric artery thrombosis in, 2030, 2035 treatment of, 2032-2036, 2033f types of, 2030t, 2034-2035 chronic, 2044-2046, 2045f gastroparesis and, 808 Mesenteric panniculitis, 620-621 Mesenteric vein, inferior, anatomy of, 1489, 1490f Mesenteric venous thrombosis, 2036-2038 acute, 2033f, 2036-2038, 2037f chronic, 2036-2038 clinical features of, 2036 diagnosis of, 2036-2037, 2037f incidence of, 2036 pathophysiology of, 2036 predisposing conditions in, 2036, 2036t prognosis in, 2038 subacute, 2036 treatment of, 2037-2038 Mesenteritis, retractile, 620-621 Mesenteroaxial gastric volvulus, 384, 384f Mesentery anatomy of, 612 Castleman’s disease of, 620 cysts of, 620 diseases of, 619-621 embryologic development of, 1625 hemorrhage of, 620 inflammatory and fibrotic conditions of, 620-621 leiomyoma of, 620 tumors of, 620 Mesh placement, for radiation enteritis, 647 Mesh plug technique, for groin hernia repair, 387 Mesh repair of groin hernia, 387 of incisional hernia, 389-390 Mesoappendix, 2060 Mesocaval shunt, for Budd-Chiari syndrome, 1375 Mesothelioma, of peritoneum, 619 Metabolic acidosis in glycogen storage disease type I, 1263 in short bowel syndrome, 1790-1791 Metabolic disorders in acute pancreatitis, 965-966, 977, 982 in eating disorders, 129t-130t liver transplantation for, 1603 vomiting in, 200 Metabolic idiosyncrasy, in drug-induced liver disease, 1429, 1431t, 1434-1437 Metabolic liver disease. See Inborn errors of metabolism. Metabolic rate after feeding, 48 resting, 48, 49t Metabolic stress body compartment losses in, 61t and energy requirements for hospitalized patients, 48-49, 49t-50t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lix

lx

Index Metabolic syndrome criteria for, 103, 103t, 107 nonalcoholic fatty liver disease and, 1402 Metagenomics, of intestinal bacteria, 1771 Metagonimus yokogawai infection, 1933-1934 Metallothionein, 1725 copper sequestration by, 1249 Metals, hepatotoxicity of, 1453 Metaplastic atrophic gastritis autoimmune, 846f, 847-848 environmental, 846-847 Metastasis suppressor gene, 43 Metastatic disease angiogenesis in, 42-43 cutaneous, in gastrointestinal cancer, 365 epithelial-mesenchymal transition in, 42, 42f in esophageal cancer, 768 gastric, 904-905 in gastric cancer, 899 in gastrinoma, 500-501, 501f gastrointestinal, 564, 564f in gastrointestinal cancer, 564, 564f in gastrointestinal stromal tumors, 462, 468-472, 470t genes associated with, 43 in glucagonoma, 506 hepatic, 1583-1584 ascites in, 1517 in carcinoid tumor, 488 in colorectal cancer, 2231-2232 in hematologic malignancies, 564-567, 565t in hepatocellular carcinoma, 1575 lymphangiogenesis in, 43 in pancreatic cancer, 1026-1027, 1026f, 1027t in pancreatic endocrine tumors. See Pancreatic endocrine tumors, metastatic. peritoneal. See Peritoneum, carcinomatosis of. pulmonary, in colorectal cancer, 2232 small intestinal, 2153, 2153f in somatostatinoma, 511 tumor biology in, 42-43, 42f Meteorism, 237 Metformin diarrhea from, 574 malabsorption with, 1757t for nonalcoholic fatty liver disease, 1410 for obesity, 112-113 Methadone, for neuropathic pain, 2283t Methamphetamine, peptic ulcer disease and, 862 Methane, in intestinal gas, 233-236 Methanethiol, odoriferous gas from, 236 Methemalbumin, in acute pancreatitis, 971 Methemoglobinemia, from topical anesthesia, 657 Methionine malabsorption of, 1759t-1762t metabolism of, in alcoholic liver disease, 1386-1387, 1386f Methotrexate for Crohn’s disease, 1962t, 1966-1967 hepatic fibrosis from, 1443-1444 clinicopathologic features of, 1444 outcome and prevention in, 1444 risk factors for, 1443-1444, 1443t liver function monitoring in, 366-367, 367t malabsorption with, 1757t in pregnancy, 630 for primary biliary cirrhosis, 1484-1485 for primary sclerosing cholangitis, 1165 for ulcerative colitis, 1999 Methoxyflurane, hepatotoxicity of, 1449, 1449t Methylation, aberrant, in colorectal cancer, 2203-2204 Methylcellulose, for constipation, 276 Methyldopa hepatotoxicity of, 1432, 1438t malabsorption with, 1757t Methylene blue, in esophageal cancer, 754 Methylnaltrexone for constipation, 281 ileus and, 2127 for opioid-induced constipation, 2284

Methylprednisolone for acute cellular rejection of liver graft, 1607 for alcoholic liver disease, 1396, 1396f for Crohn’s disease, 1963 for inflammatory visceral neuropathy, 2143 Metoclopramide for abdominal distention and bloating, 239 for gastric motility disorders, 812, 813t for gastroesophageal reflux disease, 721-722 in gastrointestinal bleeding, 310 for gastroparesis, 574 in lower gastrointestinal bleeding, 290-291 for vomiting, 207 Metoprolol, hepatotoxicity of, 1433 Metronidazole for abdominal abscess, 417-418 for amebiasis, 1910, 1910t for amebic liver abscess, 1369 for Blastocystis hominis infection, 1914 for Clostridium difficile-associated diarrhea and colitis, 1898-1901, 1898t for Crohn’s disease, 1961, 1962t for Dientamoeba fragilis infection, 1914 for giardiasis, 1914 for Helicobacter pylori infection, 842, 842t for NSAID enteropathy, 2054 for pneumatosis coli, 2250 for pouchitis, 2009, 2021 resistance to, 1898 Micelles in bile, 1093-1095, 1095f in triglyceride digestion, 1700f, 1701 Michaelis-Gutmann bodies, in malakoplakia, 2251, 2251f Microbiota, enteric. See Bacteria, intestinal (commensal). Microcoil embolization, for colonic angioectasia, 599 Microfilaments, of hepatocytes, 1208 Microfold (M) cells intestinal, 1619 oral tolerance and, 23, 23f in Peyer’s patches, 25, 26f Microgastria, 780t, 781 Micronutrients colorectal cancer and, 2195 dietary, 52-59, 53t-59t requirements for factors affecting, 58-59, 59t in malabsorption, 58-59 MicroRNAs, in colorectal cancer, 2204 Microsatellite instability, 38, 39f in colorectal cancer, 34-35, 35f, 2160-2161, 2200t, 2202-2203, 2203f, 2204t, 2233, 2233f in esophageal cancer, 751 in gastric cancer, 894 in hereditary nonpolyposis colorectal cancer, 40, 43-44, 2200t, 2203, 2203f, 2204t, 2206 in oncogenesis, 34-35, 35f Microscopic colitis. See Colitis, microscopic (collagenous and lymphocytic). Microsporidiosis, 1917-1918 in HIV/AIDS, 527-528, 527t, 1917 malabsorption in, 1829, 1829t after solid organ transplantation, 541-542 Microtubules, of hepatocytes, 1208 Microvilli acinar cell, 913, 913f in small intestine, 1617f, 1618 Microvillus inclusion disease, 1640-1641, 1640f malabsorption in, 1759t-1762t Midgut formation of, 1624-1625, 1624f-1626f volvulus of, 780t, 787, 2116, 2116f Midodrine, for hepatorenal syndrome, 1548-1549, 1548t Migraine abdominal, 578 cyclic vomiting syndrome and, 202 Migrating motor complex. See also Anorectal motility; Colonic motility; Intestinal motility; Small intestinal motility.

Migrating motor complex (Continued) in gastric motility, 794, 795f in intestinal motility, 2121-2122 Milk breast fat content of, 1729 lactose in, 1732 oral tolerance and, 141-142 proteases in, 1729-1730, 1732 cow’s, allergy to, 146 Milk protein sensitivity versus celiac disease, 1811 definition of, 429 Milk thistle for alcoholic liver disease, 1398 hepatotoxicity of, 1456 for liver disease, 2295-2296, 2295t Milwaukee classification of sphincter of Oddi dysfunction, 1068-1069, 1068t Mind-body medicine, 2287 Mineral(s). See also specific minerals. digestion and absorption of, 1722-1726 major, 52, 52t, 80 malabsorption of, 1739-1740 congenital disorders of, 1759t-1762t in refeeding syndrome, 70 supplementation of in celiac disease, 1814-1815 in malabsorption, 74 trace, 52-58, 56t-57t, 80-81, 80t, 1725-1726, 1728-1729, 1729f in malabsorption, 74 Mineral oils, for constipation, 277t, 279 Mineralocorticoids, in intestinal ion transport, 1690 Minnesota tube, for gastric variceal bleeding, 1512 Minocycline, hepatotoxicity of, 1433, 1438t, 1439 Mirizzi’s syndrome, 1114, 1119, 1119t in gallstone disease, 1132-1133 Misoprostol for constipation, 281 for irritable bowel syndrome, 2103 for NSAID ulcer prophylaxis, 874 for NSAID ulcers, 873 for peptic ulcer disease, 872 Mitochondria acinar cell, 913 in apoptosis, 1420-1421 fatty acid beta oxidation in, 1220 fatty acid elongation in, 1220 hepatocyte, 1209 in alcoholic liver disease, 1385 in nonalcoholic fatty liver disease, 1404 in Wilson disease, 1252-1253 Mitochondrial liver disease, 1277-1278 Mitochondrial neurogastrointestinal encephalomyopathy extraintestinal manifestations of, 2140 gastrointestinal manifestations of, 581 intestinal pseudo-obstruction in, 2133-2134, 2134f natural history of, 2141 Mitogen-activating protein kinase signaling pathway, in cholangiocarcinoma, 1173 MLN-02, for ulcerative colitis, 2001 MMR genes, in colorectal tumorigenesis, 2160-2161, 2163 Model for End-stage Liver Disease (MELD) in alcoholic cirrhosis, 1394 in alcoholic liver disease, 1393 in autoimmune hepatitis, 1470 in hepatocellular carcinoma, 1602 for liver disease severity assessment, 1236 in liver transplantation, 1595, 1598, 1599f, 1602 in palliative care, 2279 in transjugular intrahepatic portosystemic shunt (TIPS), 1506 Molecular Adsorbents Recirculating System (MARS) for acute liver failure, 1568 for hepatic encephalopathy, 1546 for hepatorenal syndrome, 1549

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Molecular biology, in cancer detection, 43-44, 43t-44t Molecular mimicry, in primary biliary cirrhosis, 1479 Molybdenum, 56t-57t Moniliasis, 354-355, 354f Monoamine oxidase inhibitors, hepatotoxicity of, 1436 Monocarboxylate transporters (MCTs), 1685-1686 Monoclonality tests, in gastric MALT lymphoma, 449 Monoethylglycinexylidide test, 1235 Mononuclear cells, intestinal, 28-29 Montelukast for eosinophilic gastrointestinal disorders, 434 hepatotoxicity of, 1433-1434 Morgagni hernia clinical manifestations and diagnosis of, 381, 382f etiology and pathogenesis of, 379, 380f incidence and prevalence of, 380-381 treatment and prognosis in, 383 Morphine in cholescintigraphy, 1143 intestinal dysmotility from, 2139 for neuropathic pain, 2283t in palliative care, 2280-2281, 2281t Mosapride, for functional dyspepsia, 193 Motilin, 9-10 in hunger sensation, 802 in migrating motor complex, 794 Motilin agonists for functional dyspepsia, 193 for gastric motility disorders, 9-10 Motilin antagonists, for vomiting, 208-209 Motility. See Anorectal motility; Colonic motility; Intestinal motility; Small intestinal motility. Motor evoked potentials, anal, in fecal incontinence, 249, 250f Mouth. See Oral cavity. Movement disorders, in Wilson disease, 1252 MRCP. See Magnetic resonance cholangiopancreatography. MRI. See Magnetic resonance imaging. mTOR inhibitors for carcinoid tumors, 489 for metastatic pancreatic endocrine tumors, 522 Mucin in colon carcinogenesis, 2194 in gallstone disease, 1098-1099 gastric secretion of, 831-832 regulation of, 1099 Mucin genes, 1099 Mucinous carcinoma appendiceal, 2070 colorectal, 2211, 2213f, 2215 Mucinous pancreatic tumors cystic, 1027t-1028t, 1029, 1029f intraductal papillary, 1027t-1028t, 1030-1032, 1031f-1032f, 1031t-1032t Mucoceles, of appendix, 2070 Mucocutaneous disorders, 354-357 in candidiasis, 354-355, 354f in Crohn’s disease, 1954 in HIV infection, 355-356, 355f Mucocutaneous pigmentation, in Peutz-Jeghers syndrome, 363, 363f, 2186, 2186f Mucocutaneous ulcers, 356-357, 356f, 356t Mucosa adaptive changes in, 1726-1728, 1728f bacterial invasion of, 1846 of colon, 1617-1619, 1618f of esophagus, 667, 668f of gallbladder, 1050 immune system of. See Immune system, mucosal. of stomach, 776 heterotopic, in esophagus, 670t, 675, 675f peptic ulcer defenses of, 863 prolapse of, 858 stress-related injury to. See Stress ulcers.

Mucosa-associated epithelial chemokine, in gut-associated lymphoid tissue, 29 Mucosa-associated lymphoid tissue (MALT). See also Gut-associated lymphoid tissue (GALT). components of, 445, 446f definition of, 21 lymphoma of. See Lymphoma, gastric, marginal zone B cell (MALT). Mucosal addressin cellular adhesion molecule (MadCAM-1), in Crohn’s disease, 1946-1947 Mucosal resection, endoscopic for Barrett’s esophagus, 732 for esophageal cancer, 761, 761f-762f for gastric cancer, 902-903 Mucous neck cells, of oxyntic (fundic) glands, 777-778 Mucous surface cells, gastric, 776, 776f Mucus in colitis cystica profunda, 2252, 2252f in enteric defense, 1844 gastric secretion of, 831-832 Muir-Torre syndrome, 2185 colorectal cancer in, 2208-2209 cutaneous manifestations of, 363 Multidetector computed tomography, in gastric varices, 1497 Multidrug resistance protein(s), in hepatic drug elimination, 1419 Multidrug resistance protein-1, in GIST cells, 468 Multidrug resistance protein-2 in bilirubin transport, 323-324 in Dubin-Johnson syndrome, 326 Multiple endocrine neoplasia-I pancreatic endocrine tumors in, 494-495, 495t Zollinger-Ellison syndrome in, 502, 505 Multiple endocrine neoplasia-IIa, 494 Multiple endocrine neoplasia-IIb, 494, 2134, 2135f Multiple intraluminal impedance, in small intestinal motility measurement, 1652-1653 Multiple myeloma amyloidosis in, 568-569 gastrointestinal manifestations of, 568-569 Multiple organ failure syndrome, in acute liver failure, 1561t, 1562, 1566-1567 Multiple sclerosis, gastrointestinal manifestations of, 581 Murphy’s sign, in cholecystitis, 1114, 1142-1143 Muscarinic receptors, 10-11 in emetic reflex, 197-198 Muscle area, in nutritional assessment, 65-66, 66t Muscle contraction recording, in small intestinal motility measurement, 1652 Muscle function, in nutritional assessment, 65, 78, 78f Muscular dystrophy constipation in, 269 gastrointestinal manifestations of, 581 Muscularis mucosae, 1617-1618 Muscularis propria, of stomach, 776 Musculoskeletal system in Crohn’s disease, 1953-1954 in ulcerative colitis, 2010t-2011t, 2011-2012 in Whipple’s disease, 1837 Mushrooms, hepatotoxicity of, 1450t, 1454-1455, 1457t MUTYH polyposis, 2177t, 2184, 2185f Myasthenia gravis dysphagia in, 688-689 gastrointestinal manifestations of, 581 Mycobacterium avium complex infection, in HIV/AIDS diarrhea in, 527t, 529, 529f hepatic granuloma in, 533 Mycobacterium avium-intracellulare infection, gastritis in, 850 Mycobacterium bovis infection, 1877 Mycobacterium paratuberculosis infection, Crohn’s disease and, 1942-1943 Mycobacterium tuberculosis infection. See Tuberculosis.

Mycophenolate mofetil abdominal pain from, 542-543 for Crohn’s disease, 1967 diarrhea from, 542 for liver transplantation, 1606t for ulcerative colitis, 1999 Mycotic aneurysm, 607, 607f Myelofibrosis, gastrointestinal manifestations of, 568 Myeloma, multiple amyloidosis in, 568-569 gastrointestinal manifestations of, 568-569 Myeloproliferative disorders gastrointestinal manifestations of, 568 hepatic involvement in, 568 Myenteric ganglionitis, idiopathic, chronic intestinal pseudo-obstruction in, 2140 Myenteric plexus, 665-666, 668f, 681, 819-821, 820f, 1623, 1646, 1662, 1663f esophageal peristalsis control originating in, 681-682 Myenteric potential oscillations, in colonic motility, 1660 Myocardial infarction COX-2 inhibitors and, 875 vomiting in, 200 Myofascial pain syndromes, abdominal pain in, 164-165 Myopathy. See also Cardiomyopathy; Encephalomyopathy, mitochondrial neurogastrointestinal. of anal sphincter, constipation in, 269 AST/ALT ratio in, 1230 of colon, constipation in, 269 visceral chronic intestinal pseudo-obstruction in, 2133, 2142 childhood, 2134-2135, 2135t familial, 2133-2134, 2134f familial, 2140 manometry in, 2142 Myositis, in trichinosis, 1930 Myotomy cricopharyngeal for oropharyngeal dysphagia, 700 for Zenker’s diverticula, 700 endoscopic stapler-assisted, for Zenker’s diverticula, 372-373 for esophageal diverticula, 374 Heller for achalasia, 702 gastroesophageal reflux disease after, 715 for infantile hypertrophic pyloric stenosis, 784 Myotonic dystrophy chronic intestinal pseudo-obstruction in, 2138 dysphagia in, 688

N

Nabilone, for vomiting, 208 Nadolol for esophageal variceal bleeding, 1508 for portal hypertension, 1502 NAFLD. See Fatty liver disease, nonalcoholic. NAFLD Fibrosis Score, 1408 Nail changes, in Cronkhite-Canada syndrome, 364 Naloxone angioectasia endoscopic appearance and, 597-598, 598f ileus and, 2127 for pruritus, in primary biliary cirrhosis, 1486-1487, 1486t Naltrexone for alcohol abstinence, 1394-1395 for bulimia nervosa, 134 for pruritus, 1063 Naproxen, cardioprotective effect of, 877 Narcotic bowel syndrome, in functional abdominal pain syndrome, 169-170 Narcotic packets, as foreign bodies, 404 Narcotics. See Opioid(s). Nasal alar hypoplasia, 954

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxi

lxii

Index Nasal bioelectric response testing, in cystic fibrosis, 940 Nasogastric lavage, in gastrointestinal bleeding, 286 Nasogastric tube decompression with for ileus, 2126 for small intestinal obstruction, 2110-2111 for enteral nutrition, 87-89, 88t-89t esophageal injuries related to, 739, 739f gastroesophageal reflux disease and, 715 Nasojejunal tube endoscopic placement of, 89, 89t for enteral nutrition, 87-89, 88t-89t fluoroscopy-assisted placement of, 89 self-propelled, 89 Nasopharynx anatomy of, 677, 678f in swallowing, 677 Natalizumab for Crohn’s disease, 1962t, 1969-1970 progressive multifocal leukoencephalopathy with, 1970 for ulcerative colitis, 2001 Natural killer cells, in ulcerative colitis, 1980 Natural orifice transluminal endoscopic surgery (NOTES), in appendicitis, 2068-2069 Nausea in acute pancreatitis, 969 in appendicitis, 2061 chronic, differential diagnosis of, 809t, 810 complementary and alternative medicine for, 2288-2289, 2289t in Crohn’s disease, 1953 definition of, 197 dietary therapy for, 813t-814t, 814 in emetic reflex, 198 after hematopoietic stem cell transplantation, 545-546, 546f, 555 in irritable bowel syndrome, 2092 palliative care for, 2282, 2284t postoperative, risk factors for, 2126-2127, 2126t in pregnancy, 201, 628 radiation-induced, 642 after solid organ transplantation, 538t, 541 Navaho neurohepatopathy, 1278 NBT-PABA test, for pancreatic function testing, 929 Necator americanus infection, 1925-1926, 1926f, 1926t Necrolysis, toxic epidermal, oral manifestations of, 358 Necrolytic migratory erythema, 365, 365f, 369, 506-508, 507f differential diagnosis of, 507-508 in glucagonoma, 365, 365f, 506-507, 507f Necrosis in acute pancreatitis. See Pancreatitis, acute, necrotizing. cell death by, 1420-1421, 1421f in chronic pancreatitis, 987 esophageal, acute, 743 fat in acute pancreatitis, 982 subcutaneous, in pancreatic disorders, 365 hepatic, drug-induced, 1420-1421, 1421f Necrotizing cholecystitis, 1141 Necrotizing colitis, with toxic megacolon, in amebic dysentery, 1909 Necrotizing gastritis, 849-850 Necrotizing ulcerative gingivitis, 357 Nefazodone, hepatotoxicity of, 1436 Nematode (roundworm) infection. See also specific disorders, e.g., Strongyloidiasis. cutaneous, 367, 367f hepatic, 1356t-1357t, 1359-1361 intestinal, 1921-1930 Neoangiogenesis, in esophageal cancer, 750 Neomycin for hepatic encephalopathy, 1545-1546 malabsorption with, 1757t Neonates. See also Children; Infants. acute liver failure in, in mitochondrial liver disease, 1278

Neonates (Continued) bile plug syndrome in, 1056 biliary atresia in. See Biliary atresia. cholestasis in, 1050-1052, 1051t-1052t. See also Cholestasis, pediatric. in cystic fibrosis, 947 digestion and absorption in, 1729-1732, 1730f-1731f jaundice in, 326 oral tolerance in, 22 sclerosing cholangitis in, primary, 1057 Neoplasia. See also Cancer. in Barrett’s esophagus, 729, 729f gastrointestinal, 31-44 diagnostic strategies in, current and future approaches to, 43-44, 43t-44t genes associated with, 35-41, 36t DNA mismatch repair genes as, 36t, 40 oncogenes as, 35-37, 36t. See also Oncogene(s). tumor suppressor genes as, 36t-37t, 37-40, 38f-39f. See also Tumor suppressor gene(s). inherited, features common to, 37 oncogenesis associated with, mechanisms of, 34-35, 35f tumor metastasis and, biologic features of, 42-43, 42f Neostigmine for abdominal distention and bloating, 239 for acute colonic pseudo-obstruction, 2130 Nephrolithiasis in Crohn’s disease, 1955 in short bowel syndrome, 1790, 1790f Nephropathy, in schistosomiasis, 1937 Nephrotic syndrome, ascites in, 1517, 1528 Nesidioblastosis, 495-496 in pancreatic endocrine tumors, 493-494 Neurocrine transmitter, 3, 4f Neuroendocrine (enteroendocrine) cells, 1618 development of, 1619 hyperplasia of, in postinfection irritable bowel syndrome, 2097, 2097f microscopic appearance of, 1618-1619, 1620f neurosecretory granules in, 1619 of oxyntic gland, 778 types of, 1619, 1620t Neuroendocrine tumors classification of, 493 gastrointestinal. See Carcinoid tumors. molecular pathogenesis of, 494 pancreatic. See Pancreatic endocrine tumors. Neuroenteric cysts, 1632 Neurofibromatosis-1 chronic intestinal pseudo-obstruction in, 2137 cutaneous manifestations of, 362, 362f duodenal somatostatinoma in, 513 gastrointestinal involvement in, 2185t, 2187 gastrointestinal stromal tumors and, 473-474 malabsorption in, 1764 pancreatic endocrine tumors in, 495 Neurogastrointestinal encephalomyopathy, mitochondrial. See Mitochondrial neurogastrointestinal encephalomyopathy. Neuroglycopenia, in insulinoma, 496 Neurohepatopathy, Navaho, 1278 Neurohormonal stress mediators, in postoperative ileus, 2125 Neurokinin-1 receptor antagonists, for vomiting, 208 Neurokinin A or B, 9 receptors for, 9 Neurologic disorders in celiac disease, 1806 disabling, percutaneous endoscopic gastrostomy for, 90 dysphagia in, 687-691, 687t in eating disorders, 127, 129t-130t gastric motility disorders in, 810 gastrointestinal manifestations of, 578-581, 579t after liver transplantation, 1608, 1608t in tyrosinemia, 1270

Neurologic disorders (Continued) vomiting in, 200-201 in Whipple’s disease, 1836-1837 Neurologic drugs, hepatotoxicity of, 1436-1437 Neuromuscular disorders dysphagia in, 687-691, 687t gastrointestinal manifestations of, 579t, 581 Neuromuscular training, for fecal incontinence, 253-254, 253f, 254t Neuron-specific enolase, in carcinoid tumors, 485 Neurons, enteric afferent, 1646, 1662-1663 efferent, 1646, 1663 excitatory, 1663 inhibitory, 1663 Neuropathic abdominal pain, 578-579, 2280, 2283t peripheral, 579 Neuropathy autonomic in amyloidosis, 1756 gastrointestinal manifestations of, 579t, 580 constipation in, 270 pudendal nerve, fecal incontinence and, 244 visceral. See Visceral neuropathy. Neuropeptide Y (NPY), 9 in gastrointestinal tract, 5 in postoperative ileus, 2125 Neurophysiologic testing, in fecal incontinence, 249, 249f-250f Neurotensin, in intestinal ion transport, 1691 Neurotransmitters in brain-gut interactions, 343 gastric, 819-821, 820f gastrointestinal, 3-5, 4t, 10-13, 12f as hormones, criteria for, 4-5 peptides and, coexistence of, 5 Neurotrophins, for constipation, 281 Neutropenia in enterocolitis, 2252-2253, 2253f in glycogen storage disease type I, 1263 in Shwachman-Diamond syndrome, 953-954 Neutrophils in appendicitis, 2063 in ascitic fluid, 1521 in intestinal ischemia, 2029 Nevirapine, hepatotoxicity of, 1430 Niacin (vitamin B3), 53t-55t, 81-82, 81t absorption of, 1718t, 1720 deficiency of, cutaneous manifestations of, 368 hepatotoxicity of, 1429 reference nutrient intake for, 1718t Niclosamide, for Dipylidium caninum infection, 1933 Nicotine, for ulcerative colitis, 2000 Nicotinic acid, hepatotoxicity of, 1429 Nicotinic receptors, 10-11 Niemann-Pick C1-like 1 protein, in intestinal cholesterol transport, 1703 Niemann-Pick disease, gastrointestinal manifestations of, 577-578 Nifedipine for achalasia, 701 hepatotoxicity of, 1433 for hypertension in liver transplant recipients, 1609-1610 for sphincter of Oddi dysfunction, 1071 for thrombosed hemorrhoids, 2262 Nifurtimox, for Chagas’ disease, 1919 Night eating syndrome diagnosis of, 122f, 125 differential diagnosis of, 125-126, 126t epidemiology of, 121 Nimesulide, hepatotoxicity of, 1437 Nitazoxanide for amebiasis, 1910, 1910t for cryptosporidiosis, 1915-1916 for giardiasis, 1914 Nitrates gastric cancer and, 892 in oncogenesis, 42 for portal hypertension, 1502 for sphincter of Oddi dysfunction, 1071

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Nitric oxide colonic motility and, 1672 in esophageal peristalsis, 682 in gastrointestinal tract, 12, 12f in hepatic vasodilation, 1490-1491 in hepatopulmonary syndrome, 1549 in intestinal ion transport, 1691 in lower esophageal sphincter relaxation, 684-685 in portal hypertension, 1492-1494 in postoperative ileus, 2125 types of, 12 Nitroaliphatic compounds, hepatotoxicity of, 1452 Nitroaromatic compounds, hepatotoxicity of, 1452 Nitrofurantoin, hepatotoxicity of, 1431-1432, 1438t Nitrogen balance, 50-51 Nitroglycerin hepatotoxicity of, 1452 for portal hypertension, 1502 topical for anal fissure, 2265t, 2266 for levator ani syndrome, 2273 N-Nitroso compounds, esophageal cancer and, 747 Nizatidine adverse effects of, 870 drug interactions with, 870 mechanisms of action of, 870 for peptic ulcer disease, 870 pharmacokinetics of, 870 NK-κB, in gastric MALT lymphoma, 448 Nm-23, as metastasis suppressor gene, 43 Nociception. See Pain. Nocturnal diarrhea, 218-219 Nocturnal reflux, 179 Nocturnal sleep-related eating disorder, 121, 122f, 125 NOD2/CARD15 gene, in Crohn’s disease, 1943-1945, 1944t-1945t Nodular hepatocellular carcinoma, 1574 Nodular hyperplasia hepatic. See Liver, nodular hyperplasia of. lymphoid, 2188 Nodule, Sister Mary Joseph’s, 365 Non-Hodgkin’s lymphoma. See Lymphoma. Non-nucleoside reverse transcriptase inhibitors, hepatotoxicity of, 1430 Nonalcoholic fatty liver disease. See Fatty liver disease, nonalcoholic. Nonsteroidal anti-inflammatory drugs (NSAIDs) cardiovascular risk of, 875-877 colitis from, 2040 colonic strictures from, 2244-2245, 2245f colonic ulcers from, 2052-2054 colorectal adenoma risk and, 2163-2164 colorectal cancer and, 2196, 2198, 2199f, 2199t Crohn’s disease and, 1946 diaphragmatic ulcers from, 2244-2245, 2245f diverticular bleeding and, 2085 dyspepsia from, 184 enteropathy from, 2052-2054, 2053f esophagitis from, 738 in familial adenomatous polyposis, 2183 for gastric cancer prevention, 898 gastrointestinal complications of, 559 gastropathy from, 856 hepatotoxicity of, 1414-1415, 1437 ileus and, 2127 lower gastrointestinal bleeding and, 309 microscopic colitis and, 2240-2241 obscure gastrointestinal bleeding and, 318 peptic ulcer bleeding and, 294-295, 303 peptic ulcers from. See Peptic ulcer disease, NSAID-associated. Norepinephrine in gastrointestinal tract, 11 for hepatorenal syndrome, 1548t, 1549 receptors for, 11 in stress-immune response, 345-346 Norfloxacin, prophylactic, for spontaneous bacterial peritonitis, 1534, 1535t

Norovirus infection, 1870t, 1871 in elderly, 1876 Nortriptyline, for neuropathic pain, 2283t Norwalk-like virus infection, 1870t, 1871 Notch genes in bile duct development, 1201-1202 in intestinal development, 1618 in pancreatic development, 917 NOTCH2 receptor, in Alagille’s syndrome, 1060-1061 Notochord, in pancreatic development, 916-917 NSAIDs. See Nonsteroidal anti-inflammatory drugs (NSAIDs). NTBC, for tyrosinemia, 1271 Nuclear factor kB (NF-kB) in Crohn’s disease, 1946 in esophageal cancer, 750 in Helicobacter pylori infection, 836 in hepatic regeneration, 1213 in ulcerative colitis, 1979 Nuclear hormone receptor agonist, for ulcerative colitis, 2002 Nuclear oligomerization domain 1, 2 (NOD1, 2), recognition of pathogen-associated molecular patterns by, 26 Nuclear oncogenes, 37 Nuclear pore complex, 1209 Nuclear receptors, gene expression and, 1216 Nucleoside and nucleotide analogs for hepatitis B, 1300t, 1301-1305, 1305f combination, 1304-1305 plus interferon, 1305-1306 resistance to, 1304-1305, 1305f in special populations, 1306-1307 hepatitis B flares and, 1298 for hepatitis C, 1334 in pregnancy, 1306 resistance to, HBV DNA polymerase mutation and, 1291-1292, 1292f Nucleoside and nucleotide reverse transcriptase inhibitors hepatotoxicity of, 1430 lactic acidosis syndrome from, 532-533 Nucleus ambiguus, 678 Nutcracker esophagus, spastic, 697, 699f Nutraceuticals, for gastric motility disorders, 813t, 814 Nutrient drink test, in gastric motility assessment, 805 Nutrition, 47-76. See also Diet. aging and, 58 basic concepts of, 47-52 carbohydrates in, 51, 79 in cystic fibrosis, 948-949 energy metabolism and, 47-49, 48t-50t energy stores and, 47, 48t enteral. See Enteral nutrition. essential fatty acids in, 51-52 fat in, 79 lipids in, 51-52 macronutrients in, 79-80 in malabsorption, 71-75, 71t micronutrients in, 52-59, 53t-59t minerals in major, 52, 52t, 80 trace, 52-58, 56t-57t, 80-81, 80t nitrogen balance in, 51 parenteral. See Parenteral nutrition. protein in, 49-51, 50t, 79-80 status of, liver transplantation and, 1597-1598 vitamins in, 52, 53t-55t, 81-82, 81t Nutritional assessment, 64-69, 77-79 adductor pollicis electrical stimulation in, 67 anthropometry in, 65-66, 65t-66t, 77, 78f biochemical measures in, 67-69, 77-78 bioimpedance analysis in, 66 body compartment analysis in, 65, 66t body mass index in, 65, 65t-66t caloric assessment in, 78-79, 79t creatinine-height index for, 68, 68t discriminant analysis for, 68-69, 68t in eating disorders, 125-126, 126t fist-grip dynamometry in, 66-67, 78f

Nutritional assessment (Continued) functional measures in, 66-67 history in, 64-65, 77 hydration status in, 65 immunologic tests in, 78 indirect calorimetry in, 79, 79f mini-, 69 muscle area in, 65-66, 66t muscle function in, 65, 78, 78f overview of, 64 physical examination in, 65, 77 protein assessment in, 79 rapid screening tools for, 69 research and investigational tools for, 78 respiratory muscle strength analysis in, 67 serum proteins in, 67-68, 67t skinfold thickness in, 65-66, 66t of specific nutrient deficiencies, 65 subjective global assessment in, 69, 69t, 78 tissue depletion in, 65 weight loss in, 64 Nutritional deficiency, 60-69. See also Proteinenergy malnutrition. assessment of, 64-69. See also Nutritional assessment. after bariatric surgery, 118 in celiac disease, 1804, 1814-1815 in Crohn’s disease, 1953 in primary sclerosing cholangitis, 1162, 1165 starvation and, 59-60 with vomiting, 206 Nutritional dwarfism, in children, 62-63, 62t Nutritional therapy. See also Diet. for acute pancreatitis, 978-979 aggressive complications of, 70-71. See also Refeeding syndrome. indications for, 69-70 for alcoholic liver disease, 70, 1395-1396, 1395f for cancer, 84 for celiac disease, 84 for chronic intestinal pseudo-obstruction, 2143 for Crohn’s disease, 83, 1970-1971 for diverticular disease, 83 for dumping syndrome, 84 for eating disorders, 127, 134-135 gastrointestinal function after, 135, 136t for esophageal cancer, 767 for gastric motility disorders, 813t-814t, 814 for gastrointestinal fistula, 421-422 for glycogen storage disease type I, 1263 for liver disease, 83 for malnourished patients undergoing major surgery, 69-70 for obesity, 84-85, 84t oral, 95-96. See also Diet. in palliative care, 2284 for pancreatitis, acute, 83 for peritonitis, 616 for radiation therapy patients, 70 refeeding complications in. See Refeeding syndrome. for short bowel syndrome, 82-83 for ulcerative colitis, 2000 in vomiting, 207 Nystatin for candidal esophagitis, 742 for oral candidiasis, 354

O

Oasthouse syndrome, malabsorption in, 1759t-1762t Oat bran, for colonic health, 2294t Oats, 1801, 1801f tolerance of, in celiac disease, 1801-1802 Obesity, 99-114 abdominal paracentesis in, 1520 adiposity rebound and, 102 alcoholic liver disease and, 1392-1393, 1395 ascites in, 1518-1519

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxiii

lxiv

Index Obesity (Continued) Barrett’s cancer and, 728 body mass index in, 100, 106 cancer and, 105 cardiovascular disease and, 104 central adiposity in, 100, 103, 103t, 106-107 colorectal cancer risk and, 2194 complications of, 103-106 costs of, 101 diabetes mellitus and, 104, 106 diet and, 102 drug-induced, 101-102, 101t dyslipidemia and, 104 endocrine changes in, 105 environmental agents in, 101-102 epidemiologic model of, 101, 101f esophageal adenocarcinoma and, 748 etiology of, 101-103, 101f evaluation of, 106-107, 107f gallbladder disease and, 104 gallstone disease in, 1092 gastric cancer and, 892 gastroesophageal reflux disease and, 105, 706-707 genetic causes of, 102 history in, 106 host agents in, 102-103 hypertension and, 104 intrauterine factors in, 101-102 laboratory examination in, 107 liver disease and, 104-105 drug-induced, 1417 after liver transplantation, 1610 morbid bariatric surgery for, 115, 116t-117t. See also Bariatric surgery. epidemiology of, 115 mortality and, 103-104 neonatal environment and, 102 neurophysiologic factors in, 102-103 nonalcoholic fatty liver disease and, 1401-1402 nutritional therapy for, 84-85, 84t obstructive sleep apnea and, 105-106 osteoarthritis and, 105 pathology and pathophysiology of, 103, 103f physical activity and, 102 physical examination in, 106-107 pneumonia and, 105 prevalence of, 99-101, 100f prevention of, 108 psychosocial dysfunction in, 106 renal disease and, 104 rheumatoid arthritis and, 105 skin changes in, 105-106 smoking and, 102 starvation in, 60 treatment of, 108-113 bariatric surgery for, 84 bupropion in, 112 cannabinoid CB1 antagonists in, 111 diet in, 108-109, 108t exenatide in, 113 exercise in, 109-110 fluoxetine in, 111-112 lifestyle modifications in, 109 metformin in, 112-113 orlistat in, 110t, 111 pharmacologic, 110-113, 110t pramlintide in, 113 silbutramine in, 110t, 111 surgical, 113, 115, 116t-117t. See also Bariatric surgery. sympathomimetic amines in, 110t, 111 topiramate in, 112 zonisamide in, 112 viruses and, 102 weight loss benefits in, 106 Obliterative hepatocavopathy, 1372 Obstetric injury, fecal incontinence from, 243-244, 248 Obstructive sleep apnea after bariatric surgery, 118 obesity and, 105-106

Obturator hernia, 391-392 Obturator sign, in appendicitis, 2062 Occipital horn syndrome, malabsorption in, 1759t-1762t Octreotide for carcinoid crisis prophylaxis, 484 for carcinoid tumors, 489 for chronic intestinal pseudo-obstruction, 2143 for chronic pancreatitis, 1006 for diabetic diarrhea, 575 for diarrhea, 227, 227t, 2285 gallstone disease and, 1092 for gastrointestinal fistula, 422 for glucagonoma, 508 for GRFoma, 514 for hepatorenal syndrome, 1548-1549, 1548t for insulinoma, 498 in intestinal ion transport, 1690 for malabsorption-related diarrhea, 72 for peptic ulcer bleeding, 301 for portal hypertension, 1502 for portal hypertensive gastropathy, 602 for protein-losing gastroenteropathy, 443 for short bowel syndrome, 1784, 1784t for small intestinal bacterial overgrowth, 1777 for somatostatinoma, 513 for variceal bleeding, 288-289, 289f, 307 for VIPoma, 510-511 Ocular symblepharon, 357 Oculofacial skeletal myorhythmia, in Whipple’s disease, 1837 Oculomasticatory myorhythmia, in Whipple’s disease, 1837 Oculopharyngeal dystrophy, dysphagia in, 688 Oddi, sphincter of. See Sphincter of Oddi. Odynophagia, 176, 176t in gastroesophageal reflux disease, 714 in HIV/AIDS, 524-526, 524f-525f, 525t in pill esophagitis, 736 Ofloxacin, for spontaneous bacterial peritonitis, 1533 Ogilvie’s syndrome, 2128-2130, 2128t-2129t, 2129f Ohm’s law, 1491-1492 OKT3 for acute cellular rejection of liver graft, 1607 diarrhea from, 542 for liver transplantation, 1606t Olestra, malabsorption with, 1757t Oligopeptides, decreased absorption of, 1738 Olsalazine side effects of, 1994 for ulcerative colitis, 1993, 1993f, 1994t Omalizumab, for eosinophilic gastrointestinal disorders, 434-435 Omental bursa, formation of, 774f Omentum anatomy of, 612 diseases of, 619-621 greater, 612, 774f infarction of, 621 lesser (gastrohepatic), 612 tumors of, 620 Omeprazole adverse effects of, 871 drug interactions with, 871 for gastroesophageal reflux disease, 715-716, 721, 723 hepatotoxicity of, 1433 mechanisms of action of, 870 for NSAID ulcer prophylaxis, 874 for peptic ulcer bleeding, 301 for peptic ulcer disease, 870-871 pharmacokinetics of, 870-871 for Zollinger-Ellison syndrome, 503-504 Omphalocele, 1626-1628, 1627f Omphalomesenteric band, 1630 Omphalomesenteric cyst, 1629f, 1630 Omphalomesenteric duct. See Vitelline duct. Oncofetal proteins, for cancer detection, 44 Oncogene(s), 35-37 activation of, 35 cellular (proto-oncogene), 35

Oncogene(s) (Continued) conversion of gene to, 35 in gastrointestinal tumors, 36t growth factor–related, 36 nuclear, 37 in pancreatic cancer, 1019-1020 products encoded by, 35 protein kinase–related, 36 signal transduction–related, 36-37 viral, 41 Oncogenesis chemical carcinogenesis in, 41 clonal expansion in, 34 dietary factors in, 41-42 environmental factors in, 41-42 genomic instability in, 34-35, 35f multistep nature of, 34-35, 35f signaling pathways in, 40-41, 41f Ondansetron for gastric motility disorders, 812, 813t for hyperemesis gravidarum, 628 for pruritus, in primary biliary cirrhosis, 1486-1487, 1486t for radiation-induced emesis, 642 for vomiting, 207-208 Onshido, hepatotoxicity of, 1458 Opioid(s) for acute abdominal pain, 161 in chronic pancreatitis, 1004 colonic motility and, 1672 for diarrhea, 226-227, 227t for functional abdominal pain syndrome, 169-170 ileus and, 2125 intestinal dysmotility from, 2139 in intestinal ion transport, 1690 for malabsorption-related diarrhea, 72 in palliative care, 2280-2282, 2281f, 2281t Opioid agonists, kappa, for functional dyspepsia, 193 Opioid antagonists ileus and, 2127 peripheral mu-, for constipation, 281 Opioid-sparing analgesia, ileus and, 2127 Opisthorchiasis, hepatic, 1356t-1357t, 1363 Opisthorchis felineus infection, 1934-1935 Opisthorchis viverrini infection, 1934-1935, 1934t Opium poppy, for postoperative ileus, 2295 Opium tincture, for short bowel syndrome, 1784, 1784t Opportunistic infections in HIV/AIDS, CD4 count and, 523-524, 524f after liver transplantation, 1609 in Whipple’s disease, 1836 Optical coherence tomography, in esophageal cancer, 755 Oral allergy syndrome, 143 Oral cavity disorders of, 353-354, 354f in amyloidosis, 361-362 in candidiasis, 354-355, 354f in eating disorders, 129t-130t in epidermolysis bullosa, 358 in erythema multiforme, 358 in inflammatory bowel disease, 359-360, 359f in Kaposi’s sarcoma, 355-356, 355f in lichen planus, 359, 359f in lymphoma, 356 in pemphigoid, 357 in pemphigus vulgaris, 357-358 in Sjögren’s syndrome, 353 in Stevens-Johnson syndrome, 358 leukoplakia of, 355, 355f ulcers of aphthous, 356, 356f, 356t in herpes simplex, 356-357, 356t in herpes zoster, 357 Oral cholecystography in adenomyomatosis, 1148, 1148f in biliary pain, 1113 in gallstone disease, 1108t, 1110

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Oral contraceptives colonic ischemia from, 2040 Crohn’s disease and, 1946 focal nodular hyperplasia of liver and, 1588 hepatocellular adenoma and, 1584 hepatocellular carcinoma and, 1577 hepatotoxicity of, 1440 jaundice from, 328-329 malabsorption with, 1757t Oral rehydration solution characteristics of, 73t for cholera, 1851 composition of, 1852t, 1885 for diarrhea, 226 for infectious diarrhea, 1851, 1885 for malabsorption, 73-74, 73t rice-based, 1885 for short bowel syndrome, 1781-1782, 1782f, 1784 Oral tolerance food allergy and, 141-142 in gut-associated lymphoid tissue, 22-23, 23f induction of, commensal bowel flora in, 140-141 Orbital purpura, in amyloidosis, 361-362 Oregon grape, for colonic health, 2294t Orexin, food intake effects of, 102-103 Orexin-A, in gastric acid secretion, 823 Organ allocation, liver biochemical tests for, 1236-1237 Organ failure in acute pancreatitis, 974-975, 982 multiple, in acute liver failure, 1561t, 1562, 1566-1567 Organ transplantation. See Transplantation, solid organ. Organic anion transport protein (OATP), 1081-1082, 1082f, 1083t in bilirubin transport, 323-324 Organic compounds, nonhalogenated, hepatotoxicity of, 1452 Organic solute transporter (OST), in bile acid transport, 1082f, 1083-1084, 1083t Organoaxial gastric volvulus, 384, 384f Oriental flush syndrome, with alcohol, 1384-1385 Orlistat lipid digestion and, 829 malabsorption with, 1757t for obesity, 110t, 111 Ornithine decarboxylase, in colonic carcinogenesis, 2199-2200 l-Ornithine-l-aspartate, for hepatic encephalopathy, 1546 Ornithine transcarbamylase, deficiency of, 1271-1274, 1272f, 1273t Oropharyngeal dysphagia. See Dysphagia, oropharyngeal. Oropharynx anatomy of, 677, 678f anthrax infection of, 1884 in swallowing, 677-679, 678f Oroya fever, hepatic manifestations of, 1353 Orthostatic hypotension, postprandial, gastrointestinal manifestations of, 580 Osler-Weber-Rendu disease. See Telangiectasia, hereditary hemorrhagic. Osmolality intestinal ion transport and, 1689-1690 of oral rehydration solution, 1885 Osmotic gap, in stool water, calculation of, 221-222, 222f Osmotic laxatives, for constipation, 276, 277t Osteoarthritis, obesity and, 105 Osteoma, mandibular, in familial adenomatous polyposis, 2180-2182 Osteomalacia, in malabsorption, 1745t, 1766-1767 Osteonecrosis, in ulcerative colitis, 2012 Osteopenia in celiac disease, 1806, 1814-1815 in chronic pancreatitis with steatorrhea, 997 in Crohn’s disease, 1954 after liver transplantation, 1610

Osteopenia (Continued) in malabsorption, 1745t in primary biliary cirrhosis, 1485-1486 Osteoporosis in cirrhosis, 1554 in Crohn’s disease, 1954 fractures related to, from proton pump inhibitors, 871 in malabsorption, 1766-1767 in primary biliary cirrhosis, 1485-1486 in primary sclerosing cholangitis, 1162 Outer inflammatory protein A, in Helicobacter pylori infection, 836 Ovarian cancer, small intestinal obstruction and, 2113 Ovarian cyst, ruptured, versus appendicitis, 2063t Ovarian overstimulation syndrome, ascites in, 1528 Ovarian syndrome, polycystic, obesity and, 105 Ovarian torsion versus appendicitis, 2063t vomiting in, 200 Overweightness, 100. See also Obesity. “Owl’s eye” inclusions, in cytomegalovirus hepatitis, 1348, 1349f Oxalate dietary restriction of, in short bowel syndrome, 1786, 1787t excess of, in short bowel syndrome, 1790 renal excretion of, bile acid malabsorption and, 1087 Oxalate stones, in Crohn’s disease, 1955 Oxaliplatin for colorectal cancer, 2235, 2236f plus 5-fluorouracil and leucovorin, 2232-2233, 2235 for esophageal cancer, 764 Oxidative stress in alcoholic liver disease, 1385 in Helicobacter pylori infection, 836 in nonalcoholic fatty liver disease, 1403 Oxmetidine, hepatotoxicity of, 1433 Oxycodone for neuropathic pain, 2283t in palliative care, 2280-2281, 2281t Oxygen radicals in alcoholic liver disease, 1387 in intestinal ion transport, 1691 in intestinal ischemia, 2029 Oxygen therapy for hepatopulmonary syndrome, 1552 hyperbaric, for radiation enteritis, 647 for pneumatosis coli, 2250 for pneumatosis cystoides intestinalis, 240 Oxyntic (fundic) gland, 776-777, 777f, 817-819, 819f-820f chief cells of, 778, 817-819 endocrine (enteroendocrine) cells of, 778 mucous neck cells of, 777-778 parietal cells of, 777, 777f, 817-819 polyps of, proton pump inhibitors and, 724 Oxyntomodulin, as satiety signal, 1696

P

p-ANCA in Crohn’s disease, 1959 pouchitis and, 2009, 2020 in primary sclerosing cholangitis, 1160 in ulcerative colitis, 1979 P-glycoprotein, in GIST cells, 468 P16 in esophageal cancer, 750 in gastric cancer, 893-894 in pancreatic cancer, 1020 in regulation of cell cycle, 31-32 P21 in esophageal cancer, 750 in regulation of cell cycle, 31-32 P27 in gastric cancer, 893-894 in regulation of cell cycle, 31-32

p53, 40 in apoptosis, 32, 33f in colorectal cancer, 2200t, 2202 in esophageal cancer, 750 in gallbladder carcinoma, 1178-1179 in gastric cancer, 893-894 in pancreatic cancer, 1020 point mutations in, 40 in radiation-induced apoptosis, 639 Pacemaker deactivation of, before electrosurgery, 656 gastric, 789-790, 790f-791f, 792 Pacemaker cells. See also Interstitial cells of Cajal. of colon, 1661-1662, 1661f of small intestine, 1645 Pacing, gastric, for gastric motility disorders, 812-813, 813t Paget’s disease, anal, 2270 Pain abdominal. See Abdominal pain; Functional abdominal pain syndrome. biliary. See Biliary pain. cancer, complementary and alternative medicine for, 2298 chest. See Chest pain. epigastric, 185-187, 186f, 186t medications for. See Analgesics. parietal, 152, 612 perianal after hematopoietic stem cell transplantation, 552 in hematopoietic stem cell transplantation candidates, 544 unexplained, 2272-2273 referred, 152-153, 153f, 163 somatic-parietal, 152 swallowing-related. See Odynophagia. transmission of, afferent (ascending) pathways in, 151-152, 153f, 166-167, 166f, 343, 343f visceral. See Visceral pain. Pain behaviors, culture and, 340 Palliative care, 2277-2286 for abdominal pain, 2280-2282, 2281f, 2281t, 2283t for anorexia and cachexia, 2282-2284 for ascites, 2280, 2286 for constipation, 2284, 2285t definition of, 2277, 2278t for diarrhea, 2285 for gastrointestinal bleeding, 2286 goals of, 2277-2278, 2279t for hepatic encephalopathy, 2280, 2286 versus hospice care, 2277-2278 for intestinal obstruction, 2285 for jaundice, 2285-2286 for nausea and vomiting, 2282, 2284t prognostication in, 2278-2280, 2280t Pancreas, 909-920 abscess of, 960-961, 1038 acinar cell carcinoma of, 1027t, 1033 acini of, 912-915, 912f-914f. See also Pancreas, exocrine. protein synthesis in. See Pancreatic enzymes. agenesis of, 919, 955 anatomy of, 909-912, 910f-911f functional, 921, 922f annular, 780t, 786-787, 917-918, 917f blood supply to, 911 cancer of. See Pancreatic cancer. carcinoid tumors of, 479, 515. See also Pancreatic endocrine tumors. congenital anomalies of, 917-919 in cystic fibrosis. See Cystic fibrosis, pancreatic manifestations of. cystic neoplasms of, 1027-1033 diagnostic imaging of, 1028-1029, 1028f differential diagnosis of, 1011, 1011t-1012t, 1012f, 1027-1028, 1027t epidemiology of, 1027t mucinous, 1027t-1028t, 1029, 1029f cysts of. See also Pancreatic pseudocyst. congenital, 919

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxv

lxvi

Index Pancreas (Continued) cyst fluid analysis in, 1028, 1028t differential diagnosis of, 1011, 1011t-1012t, 1012f endoscopic therapy for, 1041 development of, 915-916, 915f signaling and growth factors in, 916-917 in digestion and absorption, 1697 dorsal, agenesis of, 955 duct system of. See Pancreatic duct(s). endocrine, 912 A cells of, 915 B cells of, 915 D cells of, 915 embryology of, 916 insula-acinar portal system of, 914, 914f islets of Langerhans of, 912f, 914 PP (pancreatic polypeptide) cells of, 915 tumors of. See Pancreatic endocrine tumors. exocrine acinar cells of, 912-915, 912f-914f, 921 centroacinar cells of, 912, 913f-914f, 914, 921 duct cells of, 914, 921 embryology of, 915-916, 915f functional anatomy of, 921, 922f genetic disorders of, 939-954, 940t. See also specific disorders, e.g., Cystic fibrosis. peri-insular acini of, 914 secretory physiology of, 1697 cholecystokinin in, 927, 927f duct cell ion transport in, 922, 923f inorganic constituents in, 921-922, 922f-923f organic constituents in, 923, 923t, 924f. See also Pancreatic enzymes. secretin in, 922, 923f fistula of, in chronic pancreatitis, 1014 fluid collections in, endoscopic complications related to, 1038 gastrinoma of, 499-501 glucagon in, 8 hereditary diseases of, 931-958. See also specific disorders, e.g. Cystic fibrosis. clinical aspects of, 939-954, 940t definitions and terminology of, 931-932 models of, 932-933, 933f modifier genes in, 937 susceptibility genes in, 933-937 histology of, 912-915, 912f-914f innervation of, 912 intraductal papillary mucinous tumor of, 1027t-1028t, 1030-1032, 1031f-1032f, 1031t-1032t lymphatic drainage of, 912 lymphoma of, 459, 1033 pseudocyst of. See Pancreatic pseudocyst. resection of for chronic pancreatic pain, 1008 for pancreatic cancer, 1024-1025, 1024f, 1025t serous cystadenoma of, 1027t-1028t, 10281030, 1030f short, congenital, 919 solid pseudopapillary tumor of, 1027t, 1032-1033, 1033f stellate cells of, fibrosis associated with, 987 transplantation of, complications of, 539-540 trauma to, pancreatitis from, 938, 966-967 tumors of classification of, 1018t endocrine. See Pancreatic endocrine tumors. nonendocrine, 1017-1034 ventral, agenesis of, 955 Pancreas divisum, 918, 918f, 1039 acute pancreatitis in, 968 chronic pancreatitis and, 993 Pancreas transcription factor-1, in pancreatic enzyme synthesis, 925 Pancreatic artery caudal, 911 dorsal, 911

Pancreatic ascites, 1517, 1527-1528, 1536 Pancreatic cancer, 1017-1027 acute pancreatitis and, 1020 chronic pancreatitis and, 1014, 1017 clinical features of, 1020, 1020t computed tomography in, 1020-1021, 1021f with cystic degeneration, 1027t in diabetes mellitus, 575 diagnosis of, 1020-1023, 1024f endoscopic retrograde cholangiopancreatography in, 1021, 1021f endoscopic ultrasonography in, 1022, 1022f environmental factors in, 1017-1018 epidemiology of, 1017-1018 fine needle aspiration cytology in, 1022-1023 genetic factors in, 1017, 1018f-1019f, 1018t-1019t, 1019-1020 incidence of, 1017 magnetic resonance imaging in, 1022 metastatic, 1026-1027, 1026f, 1027t molecular pathology of, 1019-1020, 1019f, 1019t pancreatitis and, 950-951 pathology of, 1018-1020, 1018t positron emission tomography in, 1022 prognosis in, 1025 risk factors for, 1017, 1018f, 1018t screening for, 1017 serum markers of, 1023 staging of, 1023-1024, 1023t, 1024f treatment of, 1024-1027 adjuvant chemoradiotherapy in, 1025-1026, 1026f chemotherapy in, 1026-1027, 1026f, 1027t endoscopic therapy for, 1041 palliative procedures in, 1025 surgical, 1024-1025, 1024f, 1025t targeted therapy in, 1027, 1027t Pancreatic cholesterol esterase, 1702 Pancreatic duct(s) abnormalities of, on endoscopic ultrasonography, 1003, 1003t, 1004f accessory, 911, 911f adenocarcinoma of, 1018-1020, 1019f, 1019t. See also Pancreatic cancer. anatomy of, 910-911, 911f cells of, 914, 921 chain-of-lakes appearance of, 1001, 1002f drainage of, 1008 in elderly, 1002 endotherapy of, 1039, 1040f hyperplasia of, in pancreatic endocrine tumors, 493-494 intraductal papillary mucinous tumor of, 1027t-1028t, 1030-1032, 1031f-1032f, 1031t-1032t leaks of, endoscopic therapy for, 1040-1041 main (of Wirsung), 910-911, 911f narrowing of, in autoimmune pancreatitis, 990f, 991 obstruction of, chronic pancreatitis and, 992-993 pressure in, in chronic pancreatitis, 995 sphincterotomy of, for chronic pancreatitis, 1006 stone removal from, 1006-1007 for chronic pancreatitis, 1006-1007 strictures of, in chronic pancreatitis, 995 Pancreatic endocrine tumors, 491-522 ACTH-secreting, 515 advanced, new treatments for, 522 in carcinoid syndrome, 515 chromogranins produced by, 493 classification of, 492-493, 492t computed tomography in, 515-516, 516t cystic, 1027t erythropoietin-secreting, 515 functional, 492, 492t gastrin-secreting, 498-505, 499t. See also Gastrinoma. general considerations in, 491-494 ghrelin-secreting, 515

Pancreatic endocrine tumors (Continued) glucagon-secreting, 505-508. See also Glucagonoma. growth hormone–releasing factor–secreting, 513-514. See also GRFoma. hepatic venous hormone sampling in, 517-519, 519f histologic features of, 492-493, 493f historical aspects of, 491-492 human chorionic gonadotropin in, 493 incidence and prevalence of, 492, 492t inherited disorders associated with, 493 insulin-secreting, 495-498, 496t. See also Insulinoma. localization of, 515-519, 516t luteinizing hormone–secreting, 515 magnetic resonance imaging in, 515-516, 516t metastatic, 519-522 chemotherapy for, 520-521 hepatic artery embolization and chemoembolization for, 521 imaging of, 516, 516t interferon-alpha for, 521-522 liver transplantation for, 522 radiofrequency ablation of, 521 somatostatin analogs for, 521 somatostatin receptor–directed radiotherapy for, 522 surgery for, 521 tumor biology, prognostic factors, and survival in, 519-520, 520t molecular pathogenesis of, 494 in multiple endocrine neoplasia, 494-495, 495t in neurofibromatosis-1, 495 nomenclature for, 493 nonfunctional, 492, 492t, 514-515 clinical features and diagnosis of, 514-515 definition of, 514 pathophysiology and pathology of, 514 treatment of, 515 origin of, 492-493 pancreatic polypeptide–secreting, 514-515 parathyroid hormone–related peptide– secreting, 515 pathophysiology of, 493-494 positron emission tomography in, 519 renin-secreting, 515 somatostatin receptor scintigraphy in, 516, 516t, 517f-518f, 519 somatostatin-secreting, 511-513, 511t. See also Somatostatinoma. in tuberous sclerosis, 495 ultrasonography in, 515-516, 516t endoscopic, 516t, 517, 518f intraoperative, 516t, 519 vasoactive intestinal peptide–secreting, 508-511. See also VIPoma. in von Hippel-Lindau disease, 495 Pancreatic enzymes, 914. See also specific enzymes, e.g., Amylase. activation of, 923, 924f in acute pancreatitis, 970-971 co-localization of, 962 functions of, 923-924 isolated defects of, 955-956 in malabsorption, 1742, 1742t proteolytic, developmental changes in, 1732 secretion of, 924-925 cellular regulation of, 925-926, 925f cephalic phase of, 926 diet-related changes in, 1727, 1727f digestive, 926-927, 927f disorders of. See Cystic fibrosis; Pancreatic insufficiency. feedback regulation of, 927-928 gastric phase of, 926 interdigestive, 926 intestinal phase of, 926, 927f tests of, 928-930, 928t synthesis of, 922f, 923-925, 923t, 924f regulation of, 925

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Pancreatic enzymes (Continued) therapeutic in chronic pancreatitis for maldigestion and steatorrhea, 1005t, 1009-1010 for pain control, 1005-1006, 1005t after pancreatic surgery, 1009 in cystic fibrosis, 942 in diarrhea, 226-227 lipase content in, 1005t, 1009 in Shwachman-Diamond syndrome, 954 Pancreatic function tests, 928-930, 928t in chronic pancreatitis, 998t, 999-1000 in diarrhea, 226 direct, 928-929, 928t, 999-1000 indirect, 928t, 929-930, 1000 in malabsorption, 1753 Pancreatic insufficiency. See also Pancreatic enzymes. in cystic fibrosis endocrine dysfunction in, 942 exocrine dysfunction in, 941-942 pancreatic enzyme replacement for, 942 pathology of, 936, 936f treatment of, 942-943 vitamin supplementation for, 942-943 in diabetes mellitus, 575 after hematopoietic stem cell transplantation, 552, 555 in Iverson’s syndrome, 955 in Johanson-Blizzard syndrome, 954 in Pearson’s marrow-pancreas syndrome, 955 in Shwachman-Diamond syndrome, 953, 953f in tropical sprue, 1825-1826 Pancreatic intraepithelial neoplasia, 1019, 1019f Pancreatic lipase inhibitor (orlistat) lipid digestion and, 829 malabsorption with, 1757t for obesity, 110t, 111 Pancreatic lipase–related proteins, in triglyceride digestion, 1701 Pancreatic phospholipase A1, 1697 Pancreatic phospholipase A2, 1702 Pancreatic polypeptide, 9 plasma, in PPoma, 515 as satiety signal, 1696 tumors secreting, 514-515 Pancreatic polypeptide family, 9 Pancreatic pseudocyst acute, 960-961, 980-981, 981f drainage of, 1036-1037, 1036f-1037f etiology of, 1035-1036, 1036f infected, 960-961 chronic, 1010-1011, 1011t-1012t bleeding from, 1012 versus cystic neoplasm, 1011, 1011t-1012t, 1012f, 1027-1028, 1028t drainage of, 1010-1011, 1039, 1040f endoscopic therapy for, 1011 pseudoaneurysm with, 1012-1013, 1013f surgery for, 1011 Pancreatic secretory trypsin inhibitor, 923 gene for. See SPINK1 gene. Pancreatic stent, for post-ERCP pancreatitis, 968 Pancreatic tissue, ectopic, 918-919, 918f Pancreatic vein, 911 Pancreaticobiliary unions/malunions adenomyomatosis and, 1146-1147 anomalous, 919 gallbladder carcinoma and, 1178 Pancreaticoduodenal artery, 911 Pancreaticoduodenal vein, 779, 911 Pancreaticoduodenectomy. See Whipple procedure. Pancreatitis acute, 959-984 abdominal pain in, 154t, 158-159, 969 abdominal radiography in, 971-972 abdominal ultrasonography in, 972 abscess in, 960-961 acute respiratory distress syndrome in, 963

Pancreatitis (Continued) alcoholic, 964-965, 970 chronic pancreatitis in, 989 versus gallstone pancreatitis, 973 with hyperlipidemia, 966 amylase in, serum, 970-971 APACHE-II scoring system for, 974 versus appendicitis, 2063t in ascariasis, in children, 938-939 autoimmune, 969 biliary sludge and, 963-964 BISAP scoring system for, 974-975 blood tests for, 971 blood urea nitrogen in, 974 burden of disease in, 959 after burn injury, 969 C-reactive protein in, 975 cardiovascular complications of, 977 in celiac disease, 969 chest radiography in, 972, 976 in children clinical features of, 939 etiology of, 938-939, 938t recurrent, 939 versus cholecystitis, 1115 cholesterolosis and, 1146 chronic pancreatitis and, 987, 989 clinical presentation in, 969-970, 970f coagulopathy in, 982 complications of local, 980-982, 980t systemic, 980t, 982 computed tomography in, 960f, 972, 973f, 975-976, 975t in cystic fibrosis, 941-942 cytokines in, 962-963 definition of, 931-932, 959-961, 960t in diabetes mellitus, 575 differential diagnosis of, 969t, 970 drug-induced, 965, 965t in children, 938, 938t endoscopic retrograde cholangiopancreatography in, 973 endoscopic therapy for, 978 early, 1035 for local complications, 1035-1038, 1036f-1038f for recurrent disease, 1038-1039 endoscopic ultrasonography in, 972 familial metabolic syndromes associated with, 956 fat necrosis in, 982 gallstone versus alcoholic pancreatitis, 973 in children, 938 criteria for, 960, 960t pathogenesis of, 962 pathophysiology of, 963 treatment of, 979 gastrointestinal bleeding in, 982 hematocrit in, 975 after hematopoietic stem cell transplantation, 552, 555 in hemolytic uremic syndrome, in children, 939 hemorrhagic, 960-961 hereditary, 949-950, 950f history in, 969 in hypercalcemia, 966 in hyperlipidemia, 970 in hypertriglyceridemia, 965-966 incidence of, 959 infection and, 963, 966, 977-978 in children, 938-939, 938t in inflammatory bowel disease, 969 interleukin-6 in, 975 interstitial, 960f, 961, 972 intestinal compression or fistulization in, 982 in Kawasaki disease, in children, 939 laboratory diagnosis of, 970-971 laboratory markers of, 975 lipase in, serum, 971

Pancreatitis (Continued) magnetic resonance imaging in, 972-973 metabolic complications of, 977, 982 in metabolic disorders, 965-966 methemalbumin in, 971 mild, 960 natural history of, 961 necrotizing, 960-961, 960f, 972, 994 infected, 960-961 pseudocyst with, 981 treatment of, 979, 979f, 981-982 walled-off, 960-961, 979, 979f, 981-982, 1037-1038, 1037f-1038f obstructive causes of, 963-964 organ failure in, 974-975, 982 in pancreas divisum, 968 pancreatic cancer and, 1020 pancreatic enzymes in, 970-971 co-localization of, 962 pancreatitis-associated protein in, 971 pathogenesis of, 961-962 pathophysiology of, 962-963 peritoneal lavage in, 975 phospholipase A2 in, 975 physical examination in, 969-970 polymorphonuclear leukocyte elastase in, 975 post-ERCP, 967-968, 967t postoperative, 968 predisposing conditions in, 963-969, 963t procalcitonin in, 975 in protein-calorie malnutrition, in children, 939 pseudocyst in, 960-961, 980-981, 981f drainage of, 1036-1037, 1036f-1037f etiology of, 1035-1036, 1036f infected, 960-961 Ranson’s criteria for, 960, 960t, 974 recurrent, 994 in children, 939 endoscopic therapy for, 1038-1039 idiopathic, sphincter of Oddi dysfunction in, 1072 renal complications of, 982 respiratory complications of, 977, 982 in Reye’s syndrome, in children, 939 scoring systems for, 960t, 974 severe, 994 criteria for, 960, 960t enteral feeding for, 1035 severity of, predictors of, 973-976 smoking and, 969 in sphincter of Oddi dysfunction, 969 SPINK1 gene in, 961-962 splenic complications of, 982 from structural abnormalities, in children, 938 systemic inflammatory response syndrome (SIRS) in, 963 in systemic lupus erythematosus, in children, 939 toxin-induced, 965 traumatic, 938, 966-967 treatment of, 976-980, 976f antibiotics in, 977-978 cardiovascular care in, 977 continuous regional arterial infusions in, 980 endoscopic, 978. See also Pancreatitis, acute, endoscopic therapy for. fluid resuscitation in, 977 general considerations in, 976-977 metabolic, 977 nutritional therapy for, 83 nutritional therapy in, 978-979 respiratory care in, 977 surgical, 979-980 trypsin activation in, 961-962 trypsinogen activation peptide in, 975 tumors and, 964 in vascular disease, 966 in acute liver failure, 1561t, 1562-1563

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxvii

lxviii

Index Pancreatitis (Continued) alcoholic. See Alcoholic pancreatitis. autoimmune, 1155 CFTR gene in, 962 in children, 937-939, 938t chronic, 985-1016 abdominal pain in, 995-996, 1039 alcohol cessation for, 1005 analgesics for, 1004 antioxidant therapy for, 1005 endoscopic therapy for, 1006-1007 from increased pressure and ischemia, 995-996 nerve block and neurolysis for, 1009 from neurologic alterations, 996 octreotide for, 1006 pancreatic duct sphincterotomy for, 1006 pancreatic duct stone removal for, 1006-1007 pancreatic enzymes for, 1005-1006, 1005t smoking cessation for, 1005 stent therapy for, 1006, 1007f surgery for, 1007-1009 treatment of, 1004-1009, 1005t abdominal radiography in, 1000 acute pancreatitis and, 987, 989 alcoholic, 964-965, 988-989 pathophysiology of, 987-988 prognosis in, 989 asymptomatic, 994 autoimmune, 986-987, 990-992, 990f, 992f, 993t big-duct, 998 bile duct obstruction in, 1012f, 1013-1014 biliary strictures in, 1039 CFTR gene in, 987-990 classification of, 988-995, 988t, 998 clinical features of, 995-997 complications of, 1010-1015 computed tomography in, 1000t, 1001, 1001f in cystic fibrosis, 941-942 definition of, 931-932, 985 diabetes mellitus in, 994, 997, 1010 diagnosis of, 998-1004 functional tests for, 998t, 999-1000 strategy for, 1003-1004 structural tests for, 998t, 1000-1003 duct-centric, idiopathic, 986-987, 991 ductal obstruction hypothesis of, 987 duodenal obstruction in, 1014 dysmotility in, 1014-1015 in elderly, 994 epidemiology of, 985-986 ERCP in, 1001-1003, 1002f, 1002t etiology of, 988-995, 988t fat maldigestion in, 997 gastrointestinal bleeding in, 1011-1013, 1013f gastroparesis in, 1014-1015 genetic factors in, 987-990 hemodialysis and, 994 hereditary, 950, 950f, 989-990 histology of, 986-987, 986f idiopathic, 994-995 early-onset, 994-995 late-onset, 995 incidence of, 986 magnetic resonance cholangiopancreatography in, 1001 maldigestion in, 996-997, 1009-1010 mortality in, 986 necrosis-fibrosis hypothesis of, 987 obstructive, 992-993 pancreas divisum and, 993 pancreatic cancer and, 1014, 1017 pancreatic fistula in external, 1014 internal, 1014 pathology of, 986-987, 986f pathophysiology of, 987-988 physical examination in, 997-998 pseudoaneurysm in, 1012-1013, 1013f pseudocyst in, 1010-1011, 1011t-1012t, 1039, 1040f

Pancreatitis (Continued) quality of life in, 986 recurrent or severe acute pancreatitis and, 994 SAPE (sentinel acute pancreatitis event) paradigm for, 987-988 small-duct (minimal-change), 998 small intestinal bacterial overgrowth in, 1774 smoking and, 989 sphincter of Oddi dysfunction in, 993, 1072 SPINK1 gene in, 987-990 splenic vein thrombosis in, 1013 steatorrhea in, 996-997, 1009-1010 toxic-metabolic hypothesis of, 987 treatment of, 1004-1010 for abdominal pain, 1004-1009, 1005t for diabetes mellitus, 1010 endoscopic therapy for, 1039, 1039t, 1040f for maldigestion and steatorrhea, 1009-1010 tropical, 989 ultrasonography in abdominal, 1000-1001, 1000t endoscopic, 1003, 1003t, 1004f variceal bleeding in, 1013 after double balloon enteroscopy, 658 after endoscopic retrograde cholangiopancreatography, 661, 1197 from endoscopic retrograde cholangiopancreatography, 660, 660t familial, 931, 952 in familial hyperlipidemia, 956 in familial hyperparathyroidism with hypercalcemia, 956 gallstone acute versus alcoholic pancreatitis, 973 criteria for, 960, 960t endoscopic therapy for, 1035 pathogenesis of, 962 pathophysiology of, 963 treatment of, 979 treatment of, 1132 genetics of CASR (calcium-sensing receptor) gene in, 934-935, 934f CFTR gene in, 934f-935f, 935-937, 936t. See also Cystic fibrosis. PRSS1 (cationic trypsinogen) gene in, 933-934, 934f, 989-990 PRSS2 (anionic trypsinogen) gene in, 934 SBDS gene in, 937 SPINK1 gene in, 937 hereditary, 949-952, 989-990 clinical features of, 937, 949-951 definition of, 931 diabetes mellitus and, 950 diagnosis of, 951 genetic testing in, 951-952 pancreatic cancer and, 950-951, 1017 in HIV/AIDS, 938-939 lymphoplasmacytic sclerosing, 986-987, 991, 992f models of, as complex disorder, 932-933, 933f in pregnancy, 630-631 recurrent, familial metabolic syndromes associated with, 956 in scleroderma, 560 susceptibility genes for, 932 traumatic, 938, 966-967 tropical, 952-953, 989 calcific, 931, 952 chronic, 989 definition of, 931 malabsorption and, 1831 subtypes of, 931 trypsin in, 932, 934f vomiting in, 200 Pancreatitis-associated protein, in acute pancreatitis, 971 Pancreatocholangitis, sclerosing, 1155

Pancreatoduodenectomy. See Whipple procedure. Pancreatography endoscopic retrograde. See Endoscopic retrograde cholangiopancreatography. magnetic resonance. See Magnetic resonance cholangiopancreatography. Paneth cells, intestinal, 1618, 1619f Panhypopituitarism, gastrointestinal manifestations of, 573t, 576 Panniculitis, mesenteric, 620-621 PANT mnemonic, in appendicitis, 157 Pantoprazole adverse effects of, 871 drug interactions with, 871 hepatotoxicity of, 1433 mechanisms of action of, 870 for peptic ulcer disease, 870-871 pharmacokinetics of, 870-871 Pantothenic acid (vitamin B5), 53t-55t, 81t, 82 absorption of, 1718t, 1720 reference nutrient intake for, 1718t Papain, 400 Papaverine for acute mesenteric ischemia, 2032 for nonocclusive mesenteric ischemia, 2035 for superior mesenteric artery embolus, 2034 Papaya, for colonic health, 2294t Papilla, duodenal, 779 Papillary mucinous tumor, intraductal, pancreatic, 1027t-1028t, 1030-1032, 1031f-1032f, 1031t-1032t Papillary stenosis, in HIV/AIDS, 534, 534f Papillitis, in recurrent pyogenic cholangitis, 1168 Papilloma gallbladder, 1145-1146, 1149, 1150t squamous, esophageal, 768, 768f Papillomavirus infection. See Human papillomavirus infection. Papulosis, malignant atrophic cutaneous manifestations of, 360-361, 361f gastrointestinal manifestations of, 562 vasculitis in, 2047 Para-aminosalicylate, malabsorption with, 1757t Paracellular transport, model of, 1676f, 1678-1679 Paracentesis for ascites, 1519-1521 ascitic fluid infection after, 1530-1531 contraindications to, 1519 diagnostic, 1520 follow-up, 1533-1534 indications for, 1519 needle type and entry site for, 1519-1520 patient position in, 1519-1520 for peritoneal carcinomatosis, 618-619 serial, 1538-1539 techniques in, 1520-1521 therapeutic, 1520-1521 Paracrine immunoneuroendocrine system (PINES), 214, 214f, 1686, 1686f Paracrine transmitter, 3-4, 4f Paraduodenal hernia, 392-395 Paraesophageal hernia clinical manifestations and diagnosis of, 381 etiology and pathogenesis of, 379, 380f gastric volvulus with, 383-384, 385f incidence and prevalence of, 380 treatment of, 382-383 Paraganglioma, biliary, 1183 Paraneoplastic acrokeratosis of Bazex, 365 Paraneoplastic pemphigus, 358 Paraneoplastic syndromes gastrointestinal, 564 in hepatocellular carcinoma, 1571 neurologic, chronic intestinal pseudoobstruction in, 2137-2138 Paraneoplastic visceral neuropathy chronic intestinal pseudo-obstruction in, 2137-2138 constipation in, 270 Paraprosthetic enteric fistula, 608, 608f Paraquat, hepatotoxicity of, 1452

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Parasitic disease. See also Helminthic infection; Protozoal infection. cutaneous, 367, 367f food-borne, 1879t gastritis in, 852 hepatic, 1355-1365, 1355t-1357t intestinal, 1921-1940 peritonitis in, 617 recurrent pyogenic cholangitis and, 1167 tropical malabsorption from, 1828-1829, 1829t Parastomal hernia, in Brooke ileostomy, 2017 Parasympathetic nervous system in colonic motility, 1662f, 1664 in small intestinal motility, 1647-1648 Parathyroid disease. See also Hyperparathyroidism; Hypoparathyroidism. gastrointestinal manifestations of, 573t, 576-577 Parathyroid hormone–related peptide, tumors secreting, 515 Parathyroidectomy, in Zollinger-Ellison syndrome, 503-505 Paratyphoid fever, hepatic manifestations of, 1352 Parecoxib, cardiovascular risk of, 875 Parenteral iron overload, 1240 Parenteral nutrition, 78 in acute pancreatitis, 979 administration of, 86 central, formula for, 85, 85t gallstone disease and, 1091 home, for short bowel syndrome, 1785-1788, 1788f-1789f laboratory testing in, 86 in malabsorption, 74 metabolic complications of, 86-87 order for, 85, 85t in patients undergoing major surgery, 69-70 peripheral, formula for, 85-86, 86t in radiation enteritis, 645-646 total, in gastrointestinal fistula, 421-422 for ulcerative colitis, 2000 vascular access devices for, 87 Parietal cells ion transport in, 821f, 823-824, 824f of oxyntic (fundic) glands, 777, 777f, 817-819 proton pump of, 821f, 823-824, 824f Parietal pain, 152, 612 Parietal peritoneum, 611-612 Parkes Weber syndrome, 606 Parkinson’s disease chronic intestinal pseudo-obstruction in, 2136 constipation in, 268 dysphagia in, 688 gastrointestinal manifestations of, 580-581 Paromomycin for amebiasis, 1910, 1910t for amebic liver abscess, 1369 for Dientamoeba fragilis infection, 1914 for giardiasis, 1914 Parotid hypertrophy, from chronic vomiting, 127f Paroxetine, for functional dyspepsia, 193 Paterson-Kelly syndrome, cervical esophageal web in, 675 Pathogen-associated molecular patterns, recognition of, 26 Patient-controlled analgesia, ileus and, 2127 Patient-physician relationship in biomedical model, 338, 338f in functional abdominal pain syndrome, 169 in treatment based on biopsychosocial model, 348 Pattern recognition receptors, recognition of pathogen-associated molecular patterns by, 26 Pearson’s marrow-pancreas syndrome, pancreatic insufficiency in, 955 Pectinate line, 1622, 1622f Pedestal formation, by Escherichia coli, 1844, 1845f Pefloxacin, for shigellosis, 1860-1861

Peginterferon for hepatitis B, 1302, 1307 plus lamivudine, 1305-1306 for hepatitis C, 1329 plus ribavirin, 1320 for hepatitis D, 1311-1312 Peliosis hepatis, 1381, 1381f in drug-induced liver disease, 1445t Pelvic abscess, 416, 416f Pelvic examination, in acute abdominal pain, 156 Pelvic floor abnormalities of, in fecal incontinence, 243-245, 243f, 243t dyssynergia of, 265 muscle strengthening exercises for, 266 Pelvic hernia, 391-392 Pelvic inflammatory disease, abdominal pain in, 154t Pelvic lipomatosis, 619 Pelvic muscles, strengthening of, for fecal incontinence, 253-254, 253f, 254t Pelvic sepsis, after ileal pouch–anal anastomosis, 2019, 2019f Pemphigoid, 357 bullous, 357 cicatricial, 357 Pemphigus, paraneoplastic, 358 Pemphigus vulgaris, 357-358 Penicillamine adverse effects of, 1256 in pregnancy, 637 for primary sclerosing cholangitis, 1165 for Wilson disease, 1255-1256, 1256t Penicillin G, for Whipple’s disease, 1841t Penicillin VK, for Whipple’s disease, 1841t Pennyroyal, hepatotoxicity of, 1457 Pentagastrin, clinical applications of, 7 Pentoxifylline for alcoholic liver disease, 1396-1397, 1397f for hepatopulmonary syndrome, 1552 for primary sclerosing cholangitis, 1165 Peppermint for functional dyspepsia, 2290-2291, 2290t for irritable bowel syndrome, 2291, 2292t Peppermint oil, for abdominal distention and bloating, 240 Pepsin esophageal injury from, 713 in protein digestion, 1712 Pepsinogen in gastric cancer, 897, 900 gastric secretion of, 829 Peptic strictures, in gastroesophageal reflux disease, 720, 720f, 726 Peptic ulcer bleeding, 294-303, 867 algorithm for, 289f aspirin and, 294-295, 303 Doppler ultrasonography in, 296 endoscopic hemostasis in, 296-297, 297t for active bleeding, 296-298, 298f for adherent clots, 297-298, 300f, 879, 879f for clean-based ulcers, 297, 300 immediate management after, 302 for nonbleeding visible vessel, 296-298, 299f for oozing of blood without other stigmata, 298-299 rebleeding after, 302, 302t endoscopic risk stratification in, 295-296, 295f-296f, 296t, 878-879, 879f, 879t endoscopic therapy for, 879-881 combination methods in, 880-881 comparison of techniques in, 880 proton pump inhibitor use before, 881 repeat, 883 endoscopy in repeat to confirm healing, 303 second-look, 301-302 epidemiology of, 294, 294f Helicobacter pylori infection and, 294-295, 300, 303 hemoclips for, 880 histamine H2 receptor antagonists for, 881

Peptic ulcer bleeding (Continued) initial management of, 878 injection therapy for, 880 NSAIDs and, 294-295, 303 pathogenesis of, 294-295 proton pump inhibitors for, 881 recurrent prevention of, 303 surgery versus endoscopic retreatment after, 883 risk factors for, 294, 294t risk stratification for, 878-879, 879f, 879t stigmata of, 295-296, 295f-296f stress-related mucosal injury versus, 877-878 surgical management of, 881-883 choice of operation in, 882 in difficult ulcers, 882-883 indications for, 881 for recurrent bleeding, 883 timing in, 881-882 thermal therapy for, 880 treatment of, 878-883 angiography with embolization in, 302 endoscopic, 296-297, 297t, 879-881 pharmacologic, 300-301, 881 surgical, 302 Peptic ulcer disease, 861-868. See also Helicobacter pylori infection. alarm symptoms in, 864t, 867, 867f α1-antitrypsin deficiency and, 581 aspirin and, 862 Helicobacter pylori infection and, 876 in chronic obstructive pulmonary disease, 581 clinical features of, 864, 865f complications of, 867-868, 878-883 diagnosis of, 864-865 endoscopy in, 864-865, 864t, 865f radiography in, 865, 866f drug-induced, 862 duodenal. See Duodenal ulcers. dyspepsia in, 184, 866-867, 866f-867f epidemiology of, 861-862 gastric. See Gastric ulcers. gastric acid secretion and, 863 gastric cancer and, 896-897 Helicobacter pylori–associated, 862, 862f, 869 maintenance therapy for, 872 NSAID use and, 862 patterns of, 863 smoking and, 862-863 test and treat strategy for, 866, 866f treatment of, 872 Helicobacter pylori–negative, NSAID-aspirin– negative, 864, 869 hemorrhage in. See Peptic ulcer bleeding. mucosal defenses against, 863 NSAID-associated, 862, 862f healing of, COX-2 inhibitors and, 873 Helicobacter pylori infection and, 862, 876 pathogenesis of, 863 prevention of, 873-877, 876t antacids for, 873 COX-2 inhibitors for, 874-875 histamine H2 receptor antagonists for, 873-874 misoprostol for, 874 proton pump inhibitors for, 874 recommendations for, 876-877, 876t treatment of, 872-877 histamine H2 receptor antagonists for, 872-873 misoprostol for, 873 proton pump inhibitors for, 873 sucralfate for, 873 obstruction from, 868, 884-885 pathogenesis of, 863-864 penetrating, 867-868 perforation in, 867-868 abdominal pain in, 154t, 159, 159f treatment of, 883-884, 884f population-based studies of, 861 in pregnancy, 629 refractory, management of, 877

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxix

lxx

Index Peptic ulcer disease (Continued) risk factors for, 862-863, 862f stress-related, 863 time trends for, 861-862 treatment of, 869-886 antacids for, 869-870 bismuth for, 871-872 histamine H2 receptor antagonists in, 870 overview of, 869-872 prostaglandin E analogs for, 872 proton pump inhibitors for, 870-871 sucralfate for, 871 vomiting in, 199-200 Zollinger-Ellison syndrome and, 501-502, 502t Peptidases, 1713-1715, 1713f, 1714t Peptide(s) defective hydrolysis of, 1738 diarrhea from, 213, 224 gastrointestinal, food intake effects of, 103 hormonal and neuronal, 3-5, 4f, 4t overview of, 6-10 regulation of insulin and glucose by, 18-19, 19t regulation of satiety and hunger by, 18, 18t synthesis, post-translational modification, and secretion of, 6, 6f therapeutic uses of, 6 neurotransmitters and, coexistence of, 5 transport of, 1715-1716, 1715f, 1715t Peptide receptor radionuclide therapy, for carcinoid tumors, 489 Peptide YY (PYY), 9 after bariatric surgery, 1733 in eating disorders, 122-123 in satiety, 18, 802, 1696 in short bowel syndrome, 1782 Percutaneous cholecystostomy, for acute acalculous cholecystitis, 1143 Percutaneous endoscopic gastrojejunostomy, 92, 92f Percutaneous endoscopic gastrostomy, 89-91, 90f-91f for cancer, 90 complications of, 90-91, 658 continuous, 90 for dementia, 90 for disabling neurologic conditions, 90 dislodgement of, 91 indications for, 89-90 kits for, 90-91, 90f low-profile devices for, 91, 91f placement of, 89-91 removal of, 90 replacement of, 91, 91f for stroke-related dysphagia, 90 venting, for gastric motility disorders, 813t, 814 Percutaneous gastrojejunostomy, 92-93, 92f Percutaneous jejunostomy, direct, 92-93, 92f Percutaneous transhepatic cholangiography, 1186-1190 for acute cholecystitis, 1190f, 1191 for bile duct injury, 1188-1189 for bile leaks, 1188 bleeding after, 1188 in cholangiocarcinoma, 1174 in choledocholithiasis, 1117 in combination with endoscopic approach, 1197, 1197f contraindications to, 1187-1188 in jaundice, 332-333, 333t for malignant biliary obstruction, 1189-1190 in neonatal cholestasis, 1052 for postoperative biliary strictures, 1188 for primary sclerosing cholangitis, 1166, 1188, 1189f in primary sclerosing cholangitis, 1154 technique for, 1186-1188, 1187f Percutaneous transluminal mesenteric angioplasty, for intestinal angina, 2045-2046 Perianal abscess, 2266, 2267f

Perianal pain after hematopoietic stem cell transplantation, 552 in hematopoietic stem cell transplantation candidates, 544 unexplained, 2272-2273 Periappendiceal abscess, 413-414, 413f, 416, 417f Pericarditis gastrointestinal manifestations of, 584 in ulcerative colitis, 2012 Peridiverticular abscess, 413-414, 414f, 416 Perihepatitis. See Fitz-Hugh–Curtis syndrome. Perineal hernia, 391-392 Perineum descending constipation in, 266-267 fecal incontinence and, 245 examination of, in constipation, 271 physical examination of, 246 Periodontal disease, in HIV/AIDS, 355 Periportal (“pipestem”) fibrosis, in schistosomiasis, 1362, 1362f Peristalsis. See Anorectal motility; Colonic motility; Gastric motility; Intestinal motility; Small intestinal motility. Peristomal varices, in primary sclerosing cholangitis, 1164 Peritoneal dialysis, peritonitis with, 616, 1528 Peritoneal lavage in acute abdominal pain, 156 in acute pancreatitis, 975 in peritonitis, 615 Peritoneovenous shunt, for ascites, 1540 Peritoneum, 611-624 bacterial clearance in, 613, 613t blood supply of, 612 carcinomatosis of, 618-619 ascites in, 1517, 1526, 1536 clinical features of, 618 prognosis in, 619 pseudomyxoma peritonei, 619 tissue of origin in, 618f treatment of, 618-619 intraperitoneal chemotherapy for, 619 in ovarian versus nonovarian cancer, 619 paracentesis for, 618-619 surgery for, 619 cysts of, benign, 619 defense mechanisms of, 411, 412t, 613, 613t gross anatomy of, 611-612 innervation of, 612 laparoscopic evaluation of, 622, 622f ligaments of, 611-612 mesenteries of, 612 mesothelioma of, 619 metastasis to. See Peritoneum, carcinomatosis of. microscopic anatomy of, 612 parietal, 611-612 physiology of, 612 regeneration of, 612 tumors of, 618-619 visceral, 611-612 Peritonitis abdominal abscess in, 412-413. See also Abdominal abscess. antibiotics for, 161 in appendicitis, 2061-2062 bacterial secondary bacteriology of, 1530, 1531t diagnosis of, 1529, 1531-1532 pathogenesis of, 1529f, 1530 signs and symptoms of, 1530, 1530t treatment of, 1532-1534, 1532t spontaneous bacteriology of, 1530, 1531t clinical setting for, 1529 diagnosis of, 1522f, 1523-1526, 1528, 1531-1532 follow-up paracentesis for, 1533-1534 hepatorenal syndrome and, 1547-1548 pathogenesis of, 1529, 1529f

Peritonitis (Continued) prevention of, 1534, 1535t prognosis in, 1534 versus secondary bacterial peritonitis, 1525, 1525f, 1531 signs and symptoms of, 1530, 1530t versus surgical peritonitis, 1527 treatment of, 1532-1534, 1532t chlamydial, 617, 617f from diverticulitis, 2082 fungal, 617 in HIV/AIDS, 161, 617 intra-abdominal adhesions after, 618 nonsurgical, 616-618, 616t parasitic, 617 with peritoneal dialysis, 616, 1528 primary, 616 rare causes of, 618 secondary, 412, 417-418 starch, 617-618 surgical, 612-616 computed tomography in, 614-615 diagnosis of, 615 etiology and pathogenesis of, 612-614 flora in, 613 killing and sequestration mechanisms in, 613-614, 613t history in, 614 laboratory tests in, 614-615 physical examination in, 614 prognosis in, 616 treatment of, 615-616 antibiotics for, 615 surgical, 615-616 tertiary, 413, 418, 615 tuberculous, 616-617 ascites in, 1517-1518, 1526-1527, 1536 Pernicious anemia autoimmune metaplastic atrophic gastritis in, 847-848 cobalamin deficiency in, 830 Peroxisome proliferator-activated receptor (PPAR)-γ, for ulcerative colitis, 2002 Peroxisomes biogenesis disorders of, 1275, 1275t fatty acid beta oxidation in, 1220-1221 hepatocyte, 1209 Personality, gastrointestinal disorders and, 341 Pesticides, hepatotoxicity of, 1452-1453 Petit’s hernia, 392, 392f Peutz-Jeghers syndrome, 2181t, 2185-2186, 2185t cancer in, 2181t, 2186 colorectal cancer in, 2209 mucocutaneous pigmentation in, 363, 363f, 2186, 2186f pancreatic cancer in, 1017 polyps of, 2175, 2175f, 2185t, 2186 small intestinal adenocarcinoma in, 2148 Peyer’s patches, 25, 26f, 1615 T cell differentiation in, 28-29 pH ascitic fluid, 1526 esophageal in extraesophageal GERD, 180 monitoring of catheter-free, 718 in gastroesophageal reflux disease, 717-719, 718f indications for, 719 limitations of, 718 in rumination, 204 gastric, somatostatin secretion and, 823, 823f intraluminal, in triglyceride digestion and absorption, 1699t, 1701 stool water, in diarrhea, 222 Phagocytosis, 1210 Pharynx. See also Hypopharynx; Nasopharynx; Oropharynx. anatomy of, 677, 678f cartilages of, 678, 678f innervation of, 678 musculature of, 677-678, 678f in swallowing, 679-680, 680f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Phencyclidine, hepatotoxicity of, 1454 Phendimetrazine, for obesity, 110t, 111 Phenelzine, hepatotoxicity of, 1436 Phenformin, malabsorption with, 1757t Phenobarbital, hepatotoxicity of, 1432 Phenolphthalein, for constipation, 277t, 279 Phenols, urinary, in protein malabsorption, 1752 Phenothiazines for gastric motility disorders, 812, 813t intestinal dysmotility from, 2139 for vomiting, 207 Phentermine, for obesity, 110t, 111 Phenylbutazone, granulomatous hepatitis from, 1437t Phenylbutyric acid, for α1-antitrypsin deficiency, 1262 Phenytoin enteral nutrition and, 95 hepatotoxicity of, 1432 for irritable bowel syndrome, 2103 malabsorption with, 1757t Pheochromocytoma, gastrointestinal manifestations of, 573t, 576 Phlebotomy for hereditary hemochromatosis, 1246 for porphyria, 1269 Phlegmon, definition of, 960-961 Phlegmonous (suppurative) gastritis, 849-850 Phosphate absorption of, adaptive changes in, 1728, 1729f for constipation, 276, 277t depletion of, in refeeding syndrome, 70 diarrhea from, 212-213 intestinal ion transport and, 1689 Phosphate enema, for constipation, 277t, 280 Phosphatidylcholine, digestion and absorption of, 1702 Phosphoenolpyruvate carboxykinase, 1218-1219 Phospholipase A2 in acute pancreatitis, 975 functions of, 924 Phospholipid in bile, 1093 dietary, intake of, 1698-1699 low, cholestasis and, 1085t structure of, 1698, 1698f synthesis of, 1096 Phospholipidosis, amiodarone-induced, 1442 Phosphomannomutase, deficiency of, 1265, 1265f Phosphomannose isomerase, deficiency of, 1265-1266, 1265f Phosphoproteins, in intestinal ion transport, 1692 Phosphorus dietary, 52t, 80 hepatotoxicity of, 1450t, 1453 supplemental, for cholestasis in children, 1062 Photocoagulation, infrared, for hemorrhoids, 2261, 2263t Photodynamic therapy for Barrett’s esophagus, 731 for cholangiocarcinoma, 1177 for colorectal cancer, 2237-2238 for esophageal cancer, 762, 765, 765f for hilar cholangiocarcinoma, 1196 Photosensitivity, in erythropoietic protoporphyria, 1268 Phycomycosis, gastritis in, 852 Phylloquinone. See Vitamin K. Physical activity abdominal distention and bloating and, 239 colorectal cancer risk and, 2194 constipation and, 262 energy expenditure of, 48, 49t excessive, in bulimia nervosa, 124 fecal incontinence and, 252 in obesity, 102, 109-110 Physician-patient relationship in biomedical model, 338, 338f in functional abdominal pain syndrome, 169 in treatment based on biopsychosocial model, 348 Phytates, malabsorption with, 1757t Phytobezoar, 404

Picrorrhiza kurroa, for liver disease, 2295t, 2297 Pigbel infection, 1882 Pigmentation disorders in Cronkhite-Canada syndrome, 364 in gastrointestinal stromal tumors, 473 in hereditary hemochromatosis, 1243 in Peutz-Jeghers syndrome, 363, 363f in tropical sprue, 1824, 1825f in Whipple’s disease, 1837 Pili (fimbriae), of Escherichia coli, 1845 Pill dysphagia, odynophagia in, 176 Pilocarpine, xerostomia, 353 Pilonidal disease, 2273-2274, 2274f PINES (paracrine immunoneuroendocrine system), 214, 214f, 1686, 1686f Pinocytosis, 1210 Pinworm infection, 1928-1929, 1928f Pioglitazone hepatotoxicity of, 1435 for nonalcoholic fatty liver disease, 1410 Pit cells, hepatic, 1211 Pituitary disease, gastrointestinal manifestations of, 573t, 576 Pityriasis rotunda, in hepatocellular carcinoma, 1571 Plasma membrane of hepatocytes, 1207-1208 proteins of, 1207-1208 Plasmodium spp., 1355, 1355f. See also Malaria. Plasticity, of enteric nervous system, 1649 Platelet(s) abnormalities of. See Thrombocytopenia. count of in gastrointestinal bleeding, 288 in HELLP syndrome, 632t, 633 Platelet-derived growth factor in gastrointestinal tract, 16 in intestinal development, 1618 Platelet-derived growth factor receptor-alpha, in gastrointestinal stromal tumors, 464 Plectin, 1208 Plesiomonas shigelloides infection, 1854 Pleural effusion, in ascites, 1535 Plicae circulares, 1615 Plummer-Vinson syndrome cervical esophageal web in, 675 esophageal squamous cell carcinoma in, 748 Pneumatic dilation for achalasia, 701-702 for distal esophageal spasm, 703 Pneumatosis coli, 2248-2250 clinical features and diagnosis of, 2249, 2249f etiology of, 2248-2249 pathology of, 2249-2250, 2250f treatment of, 2250 Pneumatosis cystoides intestinalis, 240, 2176 in chronic intestinal pseudo-obstruction, 2141, 2141f Pneumatosis intestinalis, after solid organ transplantation, 543 Pneumatosis linearis, 2248 Pneumocystis jirovecii infection hepatitis from, in HIV/AIDS, 534 prevention of, after liver transplantation, 1609 Pneumonia aspiration in enteral nutrition, 658 after upper endoscopy, 657 with vomiting, 206 obesity and, 105 proton pump inhibitors and, 724 Pneumoperitoneum after percutaneous endoscopic gastrostomy, 90-91 in perforated peptic ulcer, 159, 159f Podophyllin, for anal warts, 2270, 2271t POEMS syndrome, ascites in, 1528 Pointed objects, as foreign bodies, 402, 402f-403f Poiseuille, law of, 1492 Poisoning diarrhea in, 217-218, 217t food. See Food poisoning. Poliomyelitis, dysphagia in, 688

Pollen-food allergy syndrome, 143 Polyamines, in mucosal hypertrophy, 1727 Polyarteritis nodosa cutaneous manifestations of, 360 gastrointestinal manifestations of, 558t, 562-563, 562f in hepatitis B, 366, 1295-1296 vasculitis in, 2047 Polybrominated biphenyls, hepatotoxicity of, 1452 Polychlorinated biphenyls, hepatotoxicity of, 1452 Polycystic kidney disease autosomal dominant, polycystic liver disease in, 1589 autosomal recessive, fibrocystic liver disease in, 1590 portal hypertension in, 1500 Polycystic liver disease, 1589-1590, 1589f liver transplantation for, 1603-1604 Polycystic ovarian syndrome, obesity and, 105 Polycythemia, in hepatocellular carcinoma, 1571 Polyenylphosphatidylcholine (lecithin) for alcoholic liver disease, 1398 in bile, 1093 Polyethylene glycol for constipation, 277t, 278 for fecal impaction, 258 for irritable bowel syndrome, 2102 Polyethylene glycol purge, in gastrointestinal bleeding, 290-291, 310 Polyflex stent, for esophageal carcinoma, 766-767 Polyhydramnios, in duodenal atresia/stenosis, 785 Polymerase chain reaction assay (PCR) in Clostridium difficile-associated diarrhea and colitis, 1896t, 1897 in gastrointestinal cancers, 43, 43t in Whipple’s disease, 1840 Polymorphisms, single nucleotide, 38 Polymorphonuclear leukocyte elastase, in acute pancreatitis, 975 Polymorphonuclear leukocytes in ascitic fluid, 1522f, 1523 in intestinal ion transport, 1689 Polymyositis chronic intestinal pseudo-obstruction in, 2135 gastrointestinal manifestations of, 558t, 561 Polyp(s). See also Polyposis. colonic. See Colonic polyps. adenomatous. See Colorectal adenoma. esophageal, inflammatory fibroid, 768 gallbladder. See Gallbladder, polyps of. gastric in familial adenomatous polyposis, 2179, 2180f fundic gland, 896 gastric cancer and, 896 hyperplastic, 896 hamartomatous, in tuberous sclerosis, 2186 inflammatory colonic, 2175-2176 fibroid eosinophils in, 432 esophageal, 768 gallbladder, 1149, 1150t Peutz-Jeghers, 2185t, 2186 Polypectomy bleeding after, 658, 659f colonoscopy after, 2172, 2172t colorectal cancer incidence after, 2171-2172 electrosurgical burns after, 660 lower gastrointestinal bleeding after, 313, 313f for malignant colonic polyp, 2169 polyp recurrence after, 2171, 2171f in ulcerative colitis, 2008 Polypoid appearance, of solitary rectal ulcer, 2051, 2051f Polyposis. See also Polyp(s). adenomatous. See Familial adenomatous polyposis. cap, 2176 cutaneous manifestations of, 363-364, 363f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxi

lxxii

Index Polyposis (Continued) gastrointestinal, 2176-2188, 2177t inherited, 2176-2187, 2177t, 2181t, 2185t. See also Familial adenomatous polyposis. noninherited, 2188 glioma-, 2177t, 2184 hamartomatous, 2148, 2185-2187, 2185t. See also individual syndromes, e.g., Peutz-Jeghers syndrome. hyperplastic, 2188 juvenile, 2185t, 2186-2187 cancer risks and screening recommendations for, 2181t colorectal cancer in, 2209 eosinophils in, 432 gastric cancer in, 893 lymphomatous, 2188 in mantle cell lymphoma, 453, 454f mixed, hereditary, 2187 MUTYH, 2177t, 2184, 2185f Polysaccharides, nonstarch, dietary, 1707 Porcelain gallbladder, in gallstone disease, 1119, 1119t Pork tapeworm infection, 1931-1932 Porphobilinogen, in porphyrias, 1268 Porphyria, 1266-1269 acute, 1267, 1267t acute intermittent, 363, 1267, 1267t cutaneous, 363, 1267-1268, 1267t diagnosis of, 1267t, 1268 erythropoietic, 363 congenital, 1267t, 1268 hepatic complications of, 1268-1269 hepatoerythropoietic, 363, 1267t, 1268 pathophysiology of, 1266-1267, 1266f treatment of, 1269 variegate, 363, 1267, 1267t Porphyria cutanea tarda, 1267-1268, 1267t in hepatitis C, 366, 366f Portacaval shunt for Budd-Chiari syndrome, 1375 for portal hypertension–related bleeding, 1507-1508 Portal circulation, 1489-1491, 1490f Portal colopathy, 602, 603f Portal enteropathy, 602, 603f Portal hypertension, 1489-1516 in alcoholic liver disease, 1498-1499 ascites and, 1517 in autoimmune hepatitis, 1498-1499 bleeding in, 1501-1508 cancer and, 1501 causes of, 1498-1501, 1499t in cirrhosis, 1498-1499 classification of, 1493f, 1498 collateral circulation in, 1494-1495 computed tomography in, 1497, 1498f endoscopic ultrasonography in, 1498 endoscopy in, 1496-1497, 1497f in extrahepatic portal vein thrombosis, 1499-1500, 1499f in fibropolycystic liver disease, 1500 gastric antral vascular ectasia and, 602 in hemochromatosis, 1498-1499 hemodynamic alterations in, 1491-1495, 1491f-1493f in hereditary hemorrhagic telangiectasia, 1501 hyperdynamic circulation in, 1491-1494 idiopathic, 1500 intrahepatic resistance in, 1492-1493, 1493f drugs that decrease, 1502-1503 magnetic resonance imaging in, 1498 in nodular regenerative hyperplasia, 1500 noncirrhotic, in drug-induced liver disease, 1445t in partial nodular transformation of liver, 1500 portal pressure measurement in, 1495-1496, 1496t in primary biliary cirrhosis, 1482, 1498-1499 in primary sclerosing cholangitis, 1498-1499 pulmonary hypertension with. See Portopulmonary hypertension.

Portal hypertension (Continued) in sarcoidosis, 590, 1500-1501 in schistosomiasis, 1499 serum-ascites albumin gradient in, 1523-1524 in splanchnic arteriovenous fistula, 1501 treatment of, 1501-1508 endoscopic, 1503-1504, 1503f, 1503t pharmacologic, 1501-1503, 1501t portosystemic shunts for, 1507-1508, 1507f sites of action for, 1491f for specific lesions, 1508-1516 surgery for, 1506-1508, 1507f transjugular intrahepatic portosystemic shunt (TIPS) for, 1504-1506, 1504f, 1505t-1506t ultrasonography in, 1497 varices in, 1494-1495 anorectal, 1494, 1513, 1513f detection of, 1496-1498, 1497f-1498f ectopic, 1513-1514, 1513f-1514f esophageal, 1494-1495, 1508-1510 gallbladder, 1499, 1499f gastric, 1494-1495, 1510-1513 gastroesophageal, 1494 retroperitoneal, 1494 treatment of, 1508-1516 umbilical, 1494 vasoactive mediators in, 1491f-1492f, 1492 Portal hypertensive colopathy, 1514 Portal hypertensive gastropathy, 601-602, 603f, 857, 1514-1516 bleeding in, 306 diagnosis of, 1514, 1514f versus gastric vascular ectasia, 1515, 1515t management of, 1515-1516, 1515f Portal pressure, measurement of, 1495-1496, 1496t Portal pyelophlebitis, in appendicitis, 2067 Portal tract, 1203, 1203f Portal vein, 911, 1202-1203 anatomy of, 1489, 1490f development of, 1202 increased flow in, in portal hypertension, 1493-1494 septic thrombophlebitis of, in appendicitis, 2067 Portal vein pressure, 1496, 1496t Portal vein thrombosis, 1377-1379 in Budd-Chiari syndrome, 1373 cavernous transformation (portal cavernoma) in, 1377, 1378f in cirrhosis, 1377 clinical features and diagnosis of, 1378-1379, 1378f etiology of, 1377-1378, 1378t extrahepatic, portal hypertension in, 14991500, 1499f natural history of, 1379 treatment of, 1379 ultrasonography in, 1497 in umbilical sepsis, 1377 variceal bleeding in, 1379 Portal venules, terminal, 1203 Portopulmonary hypertension, 1549-1552 clinical features and diagnosis of, 1551 liver transplantation for, 1596-1597 pathophysiology of, 1550 treatment of, 1552 Portosystemic encephalopathy syndrome test, 1544-1545. See also Hepatic encephalopathy. Portosystemic shunt for Budd-Chiari syndrome, 1375 in cystic fibrosis, 947 for gastric variceal bleeding, 1513 for portal vein thrombosis, 1379 surgical for gastric variceal bleeding, 1513 for portal hypertension–related bleeding, 1507-1508, 1507f for stomal varices, 1513-1514 transjugular intrahepatic. See Transjugular intrahepatic portosystemic shunt (TIPS). for variceal bleeding, 308

Positron emission tomography in carcinoid tumors, 485-486, 486f fluorodeoxyglucose. See Fluorodeoxyglucose positron emission tomography (FDG-PET). in gallbladder carcinoma, 1179-1180 in gastric cancer, 902 in gastrointestinal stromal tumors, 466-468, 467f in pancreatic cancer, 1022 in pancreatic cystic neoplasms, 1028-1029 in pancreatic endocrine tumors, 519 Positron emission tomography/computed tomography, in cholangiocarcinoma, 1174 Postcholecystectomy syndrome, 1135-1137, 1137t. See also Sphincter of Oddi, dysfunction of. Postprandial distress syndrome, 185-187, 186f, 186t Potassium. See also Hypokalemia. absorption of, in short bowel syndrome, 1781, 1781t depletion of, in refeeding syndrome, 70 dietary, 52t intestinal transport of, 1684-1685 Potassium channels, 1685 Potassium chloride esophagitis from, 738 gastropathy from, 856 Pouch. See Ileal pouch–anal anastomosis; Kock pouch. Pouchitis chronic, 2021 in ileal pouch–anal anastomosis, 2008-2010, 2020-2021, 2020t etiologies of, 2021 sequelae of, 2021 treatment of, 2021 in ulcerative colitis, 2008-2010 PP (pancreatic polypeptide) cells, 915 PPoma, 514-515 clinical features and diagnosis of, 514-515 definition of, 514 pathophysiology and pathology of, 514 treatment of, 515 Prader-Willi syndrome ghrelin in, 10 obesity in, 102 Pramlintide, for obesity, 113 Pravastatin, hepatotoxicity of, 1433 Praziquantel for clonorchiasis or opisthorchiasis, 1363 for cysticercosis, 1932 for Dipylidium caninum infection, 1933 for Echinostoma spp. infection, 1934 for fish tapeworm infection, 1931 for Hymenolepis nana/diminuta infection, 1933 for liver fluke infection, 1934-1935 for recurrent pyogenic cholangitis, 1169 for schistosomiasis, 1362 Prazosin, for portal hypertension, 1502-1503 Prealbumin, serum, in nutritional assessment, 67-68, 67t, 78 Prebiotics, for ulcerative colitis, 1999-2000 Prednisolone for celiac disease, 1817 for Crohn’s disease, 1963 for inflammatory visceral neuropathy, 2143 in pregnancy, 630 for primary biliary cirrhosis, 1484 Prednisone for alcoholic liver disease, 1396, 1397f for aphthous ulcers, 356 for autoimmune hepatitis, 1471-1475, 1472t, 1474f for autoimmune pancreatitis, 991-992 for Crohn’s disease, 1963 for eosinophilic gastrointestinal disorders, 434 for esophageal ulceration in HIV/AIDS, 526 for graft-versus-host disease, 545-546 for irritable bowel syndrome, 2103 for liver transplantation, 1606t for primary biliary cirrhosis, 1484

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Prednisone (Continued) side effects of, 1472 for typhoid fever, 1867 for ulcerative colitis, 1995 Preeclampsia with acute fatty liver of pregnancy, 635 with HELLP syndrome, 632-634, 632t, 633f jaundice in, 329 liver disease in, 632 Pregabalin for irritable bowel syndrome, 2103 for neuropathic pain, 2283t Pregnancy, 625-638 acute abdominal pain in, 161 acute fatty liver of, 201, 634-636, 635f diagnosis of, 635, 635f HELLP syndrome and, 633 jaundice in, 329 management of, 635-636 pathogenesis of, 635 acute liver failure in, 1560 amebiasis in, 1910 anti-TNF agents in, 1969 appendicitis in, 161, 630, 2062 ascariasis in, 1922-1923 cholestasis of, 631-632 cirrhosis in, 637 Crohn’s disease in, 1969, 1972 drug safety in, 626, 626t-627t in eating disorders, 127 ectopic, versus appendicitis, 2063t endoscopy in, 626-627 gallstone disease in, 1091, 1132 gallstones in, 630 gastroesophageal reflux disease in, 628-629, 715 gastrointestinal function in, 625-626 HELLP syndrome in, 632-634, 632t, 633f hemorrhoids in, 2263 hepatic function in, 626 hepatic neoplasia and mass lesions in, 637 hepatic rupture, hematoma, and infarct in, 634, 634f hepatic vein thrombosis in, 637-638 hepatitis B in, 636, 1306 hepatitis C in, 636-637 hepatitis D in, 636 hepatitis E in, 636, 1341 herpes simplex virus in, 636 hyperemesis gravidarum in, 628 ileal pouch–anal anastomosis and, 2023 infectious diarrhea in, 1872 inflammatory bowel disease in, 629-630 intrahepatic cholestasis in, 329, 1085t jaundice in, 329 liver disease in autoimmune, 637 chronic, 637 of preeclampsia, 632 liver problems complicating, 636-638 liver problems unique to, 631-636 liver tests in, 626 after liver transplantation, 638 nausea and vomiting in, 201, 628 obesity and, 105 pancreatitis in, 630-631 peptic ulcer disease in, 629 portal hypertension in, 637 preeclampsia in with acute fatty liver of pregnancy, 635 with HELLP syndrome, 632-634, 632t, 633f hepatic manifestations of, 632 small intestinal motility in, 1657 Wilson disease in, 637 Premature infants, lactase deficiency in, 1754 Preprogastrin, 6-7 Preprohormone, 6, 6f Prerenal azotemia, hepatorenal syndrome and, 1546 Primary biliary cirrhosis. See Biliary cirrhosis, primary. Primary sclerosing cholangitis. See Sclerosing cholangitis, primary.

PRKCSH gene, in polycystic liver disease, 1590 Probiotics, 96 for acute pancreatitis, 980 for antibiotic-associated diarrhea, 1890 for Clostridium difficile-associated diarrhea and colitis, 1901-1902 colonic motility and, 1672 for diarrhea, 227, 2293 for diverticulitis, 96 for diverticulosis, 2077-2078 for hepatic encephalopathy, 1546 for infectious diarrhea, 96 for inflammatory bowel disease, 96, 2292-2293 for irritable bowel syndrome, 96, 2102-2103, 2291-2292, 2292t mechanisms of action of, 96 oral tolerance and, 22 for pouchitis, 2010 for small intestinal bacterial overgrowth, 1777 for ulcerative colitis, 1999-2000 Procalcitonin, in acute pancreatitis, 975 Procarcinogen chemical, 41 dietary, 41-42 Proctalgia fugax, 2272-2273 Proctitis in Crohn’s disease, 1950 eosinophilic, 429-430, 429f endoscopy in, 431 treatment of, 433-435, 433t radiation, bleeding in, 313-314, 314f syndromic, diagnosis of, 1848 Proctocolectomy. See also Ileostomy. clinical consequences of, 2016-2017 in familial adenomatous polyposis, 2183 fecal output after, 2016 functional sequelae of, 2016 with ileal pouch–anal anastomosis, for ulcerative colitis, 2004-2005 Proctocolitis, allergic eosinophilic, 144-145, 146t Proctoscopy, rigid, 2259-2260 Proctosigmoidoscopy, in colorectal cancer, 2221t, 2223-2224 Proenzymes, 923t, 924f Progastrin, 7, 821 Proglycogen, 1219 Prognostic indices, for nutritional assessment, 68-69, 68t Prognostication, in palliative care, 2278-2280, 2280t Progressive systemic sclerosis. See Scleroderma. Proinsulin, 496 plasma, in insulinoma, 497-498 Prokinetic drugs for chronic intestinal pseudo-obstruction, 2143 for constipation, 280-281 in eating disorders, 137 for functional dyspepsia, 191-193 for gastric motility disorders, 812, 813t for gastroesophageal reflux disease, 721-722 for gastroparesis, 574 ileus and, 2127 for vomiting, 208-209 Prolapse gastric, 858 of hemorrhoids, 2262, 2262f, 2263t rectal constipation in, 267-268 in cystic fibrosis, 945 solitary rectal ulcer syndrome and, 2050-2051 Promethazine, for vomiting, 207 Pronucleating proteins, in gallstone disease, 1098-1099 Propane, hepatotoxicity of, 1452 Propofol, complications of, 654 Propoxyphene, in chronic pancreatitis, 1004 Propranolol for esophageal variceal bleeding, 1508 for portal hypertension, 1502 for portal hypertensive gastropathy, 602 Propylthiouracil, for alcoholic liver disease, 1398

Prorelaxant drugs, for gastric motility disorders, 812, 813t Prostaglandin(s) in acute acalculous cholecystitis, 1142 in acute calculous cholecystitis, 1113 in intestinal ion transport, 1688-1690 as radioprotector, 649 Prostaglandin E analogs, for peptic ulcer disease, 872 Prostatic hypertrophy, benign, inguinal hernia and, 388 Prosthesis. See Stent(s). Prostones, 1672 Protease(s) in breast milk, 1729-1730, 1732 functions of, 924 pancreatic, in protein digestion, 1712-1713, 1713f, 1713t Protease inhibitors, hepatotoxicity of, 532, 1430 Proteasome dysfunction, in alcoholic liver disease, 1386 Protein(s), 1712-1717. See also Amino acids. antinucleating, in gallstone disease, 1098-1099 catabolism of pathways for, 1216 systemic inflammatory response and, 60 dietary, 49-51, 50t, 79-80 in cirrhosis, 83 disease severity and, 85, 85t enterocolitis syndrome from, 145, 146t, 429 endoscopy in, 431 treatment of, 433-435, 433t enteropathy from, 145, 146t. See also Protein-losing gastroenteropathy. intake of, 1712 requirements for, 50-51, 50t restriction of for hepatic encephalopathy, 1545 for urea cycle defects, 1273-1274 digestion and absorption of, 1712-1717 amino acid transport in, 1716-1717, 1716f, 1716t at brush border membrane and in cytoplasm, 1713-1715, 1713f, 1714t exit from epithelium in, 1717 in neonates and infants, 1729-1730, 1732 pancreatic proteases in, 1712-1713, 1713f, 1713t pepsins in, 1712 peptide transport in, 1715-1716, 1715f, 1715t endogenous sources of, 1712 folding of, 1216 liquid, for gastric motility disorders, 814 malabsorption of congenital disorders of, 1759t-1762t after gastric resection, 1756 mechanisms of, 1738 in small intestinal bacterial overgrowth, 1774 tests for, 1752 oncofetal, for cancer detection, 44 scaffolding, in intestinal ion transport, 1676f, 1679, 1692-1693 serum enteric loss of, 437. See also Protein-losing gastroenteropathy. hepatic synthesis of, 1214-1216, 1215t in nutritional assessment, 67-68, 67t, 79 in pregnancy, 626 total, in ascitic fluid, 1524-1525, 1525f transporter. See Ion transporters. Protein-coupled amino acid transporter 1 (PAT1), 1716, 1716f, 1716t Protein-energy malnutrition, 60-63. See also Nutritional deficiency. in adults, 62-63 assessment of, 64-69. See also Nutritional assessment. body composition and, 61, 61f, 61t in cancer patients, 84 in Child-Turcotte-Pugh class A alcoholic cirrhosis, 64 in children, 61-63, 62t cytokines in, 61, 62t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxiii

lxxiv

Index Protein-energy malnutrition (Continued) in neonatal cholestasis, 1062 pancreatitis in, in children, 939 in patients undergoing major surgery, 69-70 physiologic impairments from, 63-64 primary, 60 secondary, 60 systemic inflammatory response and, 60 Protein kinase(s) in intestinal ion transport, 1692 oncogenes and, 36 Protein kinase C (PKC), activation of, in colon carcinogenesis, 2194 Protein-losing gastroenteropathy, 437-444 in allergic gastroenteropathy, 440 α1-antitrypsin clearance in, 442 approach to patient with, 442-443, 442f clinical manifestations of, 438-440, 440t definition of, 437 diagnosis of, 441-443 in Helicobacter pylori gastritis, 440 in intestinal lymphangiectasias, 441, 441f in intestinal lymphatic disorders, 438t, 441, 441f laboratory tests in, 441-442 in Ménétrier’s disease, 440 in mucosal diseases with ulceration and erosions, 438t, 441 in mucosal diseases without ulceration and erosions, 438t, 440-441, 441f pathophysiology of, 437-438, 438t, 439f in systemic lupus erythematosus, 440-441, 441f treatment of, 443 Proteoglycans, in extracellular matrix, 1212 Proteolysis, 1712-1713, 1713f, 1713t defective, 1738 Prothrombin time in cirrhosis, 1554-1555 in cystic fibrosis, 947 in gastrointestinal bleeding, 288 in jaundice, 331 as marker of hepatic synthetic function, 1234 Proto-oncogenes, 35, 2200-2201 Proton pump, in gastric acid secretion, 821f, 823-824, 824f Proton pump inhibitors, 824-825 adverse effects of, 871 for Barrett’s esophagus, 730 for bleeding ulcers, 288-289, 289f cobalamin deficiency and, 829-830 drug interactions with, 871 for dyspepsia, 190 enteric infections and, 724 for esophageal chest pain, 178 for extraesophageal GERD, 180 fracture and, 724 for functional dyspepsia, 191, 192t fundic gland polyps and, 724 for gastroesophageal reflux disease, 715-716, 721-723, 722f for globus sensation, 177 for Helicobacter pylori infection, 841-842, 842t-843t Helicobacter pylori infection and, 724 hepatotoxicity of, 1433 hypergastrinemia from, 502 malabsorption with, 1757t for Mallory-Weiss tear, 305 mechanisms of action of, 824-825, 825f, 870 for NSAID ulcer prophylaxis, 874 for NSAID ulcers, 873 for peptic esophageal strictures, 726 for peptic ulcer bleeding, 300-301, 881 for peptic ulcer disease, 870-871 pharmacokinetics of, 870-871 pneumonia and, 724 in pregnancy, 629 safety concerns with, 724 for stress ulcer bleeding, 878 structure of, 824-825, 824f for Zollinger-Ellison syndrome, 503-504 Protoporphyria, erythropoietic, 1267t, 1268 Protoporphyrin IX, 1266, 1266f

Protozoal infection. See also specific diseases and protozoa. hepatic, 1355-1359, 1356t-1357t HIV/AIDS diarrhea in, 526, 526t-527t, 527f intestinal, 1905-1920 tropical malabsorption from, 1828-1829, 1829t PRSS1 (cationic trypsinogen) gene mutations of, testing for, 951-952 in pancreatitis, 933-934, 934f, 989-990 PRSS2 (anionic trypsinogen) gene, in pancreatitis, 934 Prucalopride, for constipation, 280-281 Pruritus in cholestasis of pregnancy, 631 in liver disease, 366 in primary biliary cirrhosis, 1479-1480, 1486-1487, 1486t in primary sclerosing cholangitis, 1161-1162, 1165 Pruritus ani, 2271-2272 pS2 peptide, in gastrointestinal tract, 16 Psammoma bodies, in duodenal somatostatinoma, 511-512 Pseudoachalasia differential diagnosis of, 693 postsurgical, 693 tumor-related, 693 Pseudoaneurysm, in chronic pancreatitis, 1012-1013, 1013f Pseudocyst, pancreatic. See Pancreatic pseudocyst. Pseudodiverticula, esophageal intramural, 374-375, 374f Pseudoephedrine, colonic ischemia from, 2039 Pseudolithiasis, biliary, in children, 1064 Pseudomelanosis coli, 279, 2246-2247, 2246f-2247f Pseudomelanosis colitis, protein-losing gastroenteropathy in, 441 Pseudomembranous enterocolitis, 1890-1891 Clostridium difficile-associated, 1890-1891, 1897, 1897f, 1899-1900. See also Clostridium difficile-associated diarrhea and colitis. in leukemia, 567 severe or fulminant, treatment of, 1899-1900 Pseudomyxoma peritonei, 619 Pseudo-obstruction, 2121-2144 colonic acute, 2121, 2128-2130, 2128t-2129t, 2129f. See also Megacolon. clinical manifestations of, 2129, 2129f epidemiology of, 2123 pathophysiology of, 2128-2129, 2129t prevention of, 2129-2130, 2130t prognosis in, 2129 treatment of, 2130, 2131f after hematopoietic stem cell transplantation, 552 definition of, 2121 epidemiology of, 2123 familial, 2133-2134 intestinal chronic, 1640, 2130-2143 acute subocclusive episodes in, 2142-2143 in amyloidosis, 2138-2139 in anorexia nervosa, 2140 in bulimia, 2140 in celiac disease, 2139 in Chagas’ disease, 2139 classification of, 2132t clinical manifestations of, 2140 complications of, 2140-2141 constipation in, 2142 in dermatomyositis, 2135 in diabetes mellitus, 2136, 2137f diagnosis of, 2141-2142 in diffuse lymphoid infiltration, 2139-2140 drug-induced, 2139 in Duchenne’s muscular dystrophy, 2138 epidemiology of, 2123 etiology of, 2131-2133, 2132t familial, 2133-2134

Pseudo-obstruction (Continued) in hypoparathyroidism, 2139 in hypothyroidism, 2139 idiopathic, 2133 in idiopathic myenteric ganglionitis, 2140 in jejunal diverticulosis, 2139 laboratory findings in, 2141-2142 malnutrition in, 2140 manometry in, 2142 in mesenchymyopathies, 2133 in myotonic dystrophy, 2138 natural history of, 2141 in neurofibromatosis, 2137 nutritional therapy for, 2143 pain control in, 2142 in Parkinson’s disease, 2136 pharmacotherapy for, 2143 pneumatosis cystoides intestinalis in, 2141, 2141f in polymyositis, 2135 primary, 2133-2135 in radiation injury, 2139 radiographic findings in, 2141 in scleroderma, 2135, 2136f secondary, 2135-2140, 2142 small bowel bacterial overgrowth in, 2140 in spinal cord injury, 2136-2137 surgical management of, 2143 in systemic lupus erythematosus, 2136 total parenteral nutrition–related disorders in, 2140 treatment of, 2142-2143 in visceral myopathies, 2133-2135, 2134f, 2135t, 2142 in visceral neuropathies, 2131-2134, 2132t, 2135f, 2137-2138, 2142 gastric motility disorders in, 809-810 paraneoplastic, 564 vomiting in, 201 neural pathways in, 2122f-2123f, 2123-2124 Pseudopapillary tumor, solid, pancreatic, 1027t, 1032-1033, 1033f Pseudopolyps colonic, 2175-2176 inflammatory colitis with, in schistosomiasis, 1936, 1936f in ulcerative colitis, 1981, 1981f, 1985-1988, 1988f Pseudoterranova decipiens infection, 1930 Pseudotumor of gallbladder. See Gallbladder, cholesterolosis of. inflammatory, of liver, 1367 Pseudovomitus, 198 Pseudoxanthoma elasticum, 362, 362f Psoas sign, in appendicitis, 2062 Psoriatic arthritis, gastrointestinal manifestations of, 563 Psychological disorders constipation and, 270-271 gastrointestinal disorders and, 341 Psychological distress communication of, culture and, 340 gastrointestinal disorders and, 341-342 visceral pain and, 344, 345f Psychological factors in biopsychosocial model, 341-342 in globus sensation, 176-177 in ulcerative colitis, 1980-1981 Psychological support, for constipation, 275 Psychological therapy for eating disorders, 131-134, 132f for esophageal hypersensitity, 704 for functional dyspepsia, 193 for irritable bowel syndrome, 2103 Psychosocial environment, in biopsychosocial model, 340-342 Psychosocial factors in biopsychosocial model, 348. See also Biopsychosocial model. in functional dyspepsia, 188-189 in obesity, 106 in visceral pain, 344-345

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Psychotherapy for eating disorders, 132-133 for functional abdominal pain syndrome, 170 for gastrointestinal symptoms, 349 Psyllium for colonic health, 2294t for constipation, 276 for diarrhea, 227, 227t for irritable bowel syndrome, 2291, 2292t PTEN hamartoma tumor syndromes, 2181t, 2185t, 2187 PTEN/P13K pathway, in intestinal development, 1618 PTHC. See Percutaneous transhepatic cholangiography. Puborectalis abnormalities of, fecal incontinence and, 244 electromyography of, in constipation, 273 Pudendal nerve in continence mechanism, 242 motor latency testing of, in fecal incontinence, 249, 249f neuropathy of, fecal incontinence and, 244 Puestow procedure, modified, for chronic pancreatitis, in pain control, 1008 Pulmonary. See also Lung. Pulmonary angiography, for hepatopulmonary syndrome, 1552 Pulmonary disease chronic obstructive, peptic ulcers in, 581 gastrointestinal manifestations of, 581-584 with liver disease. See Hepatopulmonary syndrome. liver transplantation and, 1596-1597 in ulcerative colitis, 2012 Pulmonary embolism, in ulcerative colitis, 2012 Pulmonary hypertension, with portal hypertension. See Portopulmonary hypertension. Purging in anorexia nervosa, 124 complications of, 127f-128f in eating disorders, 122t, 124, 126t Purging disorder, diagnosis of, 125 Purpura Henoch-Schönlein cutaneous manifestations in, 360, 360f gastrointestinal manifestations of, 562 vasculitis in, 2047 orbital, in amyloidosis, 361-362 thrombotic thrombocytopenic, gastrointestinal manifestations of, 570 Purtsher’s retinopathy, in acute pancreatitis, 982 Push enteroscopy, in obscure gastrointestinal bleeding, 319 Pyelonephritis, versus appendicitis, 2063t Pyelophlebitis, portal, in appendicitis, 2067 Pyloric gland, 776, 817, 819, 819f gastrin-secreting (G) cells of, 778, 778f Pyloric sphincter, contraction of, 791 Pyloromyotomy, for infantile hypertrophic pyloric stenosis, 784 Pylorospasm, gastroparesis in, 808 Pylorus atresia of, 780-781, 780t electrophysiologic characteristics of, 790-791, 792f in gastric emptying, 797 hypertrophic stenosis of adult, 785 infantile, 783-785, 784f metaplasia of, in Crohn’s disease, 1949 obstruction of. See Gastric outlet obstruction. strictures of, after caustic injury, 408 Pyoderma gangrenosum in Crohn’s disease, 1954 in inflammatory bowel disease, 359-360, 360f Pyoderma vegetans, in inflammatory bowel disease, 359 Pyogenic cholangitis, recurrent, 1167-1170, 1168f Pyostomatitis vegetans in inflammatory bowel disease, 359, 359f in ulcerative colitis, 2011

Pyrantel pamoate, for ascariasis, 1360 Pyrazinamide hepatotoxicity of, 1434 for intestinal tuberculosis, 1878 Pyridostigmine for abdominal distention and bloating, 239 for chronic intestinal pseudo-obstruction, 2143 for irritable bowel syndrome, 2103 Pyridoxine (vitamin B6), 53t-55t, 81t, 82 absorption of, 1718t, 1720 aging and, 58 for nausea and vomiting, 2288-2289, 2289t reference nutrient intake for, 1718t Pyrimethamine, malabsorption with, 1757t Pyrosis. See Heartburn (pyrosis). Pyrrolizidine, hepatotoxicity of, 1456 Pyruvate kinase, in hepatic glucose metabolism, 1218-1219 Pyruvate-phosphoenol pyruvate pathway, in hepatic glucose metabolism, 1218-1219

Q

Q fever, hepatic manifestations of, 1353 Quadriplegia, gastrointestinal problems in, 580 Quinacrine, for giardiasis, 1914 Quinidine, esophagitis from, 738 Quinine for babesiosis, 1358 granulomatous hepatitis from, 1437t

R

Rabeprazole adverse effects of, 871 drug interactions with, 871 mechanisms of action of, 870 for peptic ulcer disease, 870-871 pharmacokinetics of, 870-871 Race alcoholic pancreatitis and, 988-989 carcinoid tumors and, 475-476 hepatitis C and, 1317, 1325 irritable bowel syndrome and, 2093 Racecadotril, for diarrhea, 226 Radiation enteritis, 642-647 after chemoradiotherapy, 644-645 chronic, 644, 644t clinical manifestations of, 642-645, 643f diagnosis of, 645, 645t, 646f dose and, 644 risk factors for, 644 small intestinal bacterial overgrowth in, 1774 treatment and prevention of, 645-647, 645t Radiation exposure, in pregnancy, 627, 627t Radiation injury, 639-652 to anus, 650 apoptosis in, 639 after chemoradiotherapy colonic, 647-650, 648f esophageal, 640-641 gastric, 642 small intestinal, 644-645 chronic intestinal pseudo-obstruction in, 2139 cytokines in, 639-640 to esophagus, 640-641, 640f to large intestine, 647-650, 648f to liver, 650-651 molecular mechanisms of, 639-640 reduction of, by improved therapeutic techniques, 650-651 to small intestine, 642-647. See also Radiation enteritis. to stomach, 641-642 Radiation proctitis, chronic, 644t Radiation therapy. See also Chemoradiotherapy. anal fistula after, 2269 for Buschke-Lowenstein tumors, 2271 for colorectal cancer, 2237 plus 5-fluorouracil, 2233-2234 for esophageal cancer, 762-763

Radiation therapy (Continued) esophageal squamous cell carcinoma and, 748 for gastric diffuse large B cell lymphoma, 452 for gastric MALT lymphoma, 451 for gastrointestinal stromal tumors, 469 gastropathy from, 857 hypogeusia after, 354 injury from. See Radiation injury. intensity-modulated, 651 intestinal effects of. See Radiation enteritis. near-infrared, for levator ani syndrome, 2273 nutritional support during, 70 proctitis from, bleeding in, 313-314, 314f somatostatin receptor–directed, for pancreatic endocrine tumors, 522 Radiculopathy, thoracic nerve, abdominal pain in, 165 Radioallergosorbent test, in eosinophilic esophagitis, 430 Radiofrequency ablation atrial-esophageal fistula after, 739 for Barrett’s esophagus, 731-732 for colorectal cancer metastasis, 2232 of pancreatic endocrine tumors, 521 Radiography abdominal in abdominal abscess, 414 in acute mesenteric ischemia, 2030-2031, 2031f in acute pancreatitis, 971-972 in appendicitis, 2064 in chronic pancreatitis, 1000 in colonic obstruction, 2117, 2118f-2119f in constipation, 272-273, 273f in inguinal hernia, 387, 387f in neutropenic enterocolitis, 2253, 2253f in perforated peptic ulcer, 159, 159f in small intestinal obstruction, 2107-2108, 2107f-2108f in ulcerative colitis, 1987-1988, 1987f in vomiting, 204 in amyloidosis, 586, 586f barium. See Barium radiography. chest in abdominal abscess, 414 in acute pancreatitis, 972, 976 in esophageal cancer, 753 in chronic intestinal pseudo-obstruction, 2141 in diverticulitis, 2079, 2079f in duodenal atresia/stenosis, 785-786, 786f in eosinophilic esophagitis, 430 in eosinophilic gastroenteritis, 428f, 430 of foreign bodies, 399-400, 399f in infantile hypertrophic pyloric stenosis, 784, 784f in malabsorption, 1749 in peptic ulcer disease, 865, 866f Radionuclide imaging. See Scintigraphy. Ragged red fibers, in mitochondrial neurogastrointestinal encephalomyopathy, 2133, 2134f Raloxifene hepatotoxicity of, 1443 for osteoporosis, 1485-1486 Ramipril, hepatotoxicity of, 1433 Ranitidine. See also Histamine H2 receptor antagonists. adverse effects of, 870 drug interactions with, 870 for gastroesophageal reflux disease, 723 hepatotoxicity of, 1433 mechanisms of action of, 870 for NSAID ulcers, 872-873 for peptic ulcer disease, 870 pharmacokinetics of, 870 in pregnancy, 629 Ranson’s criteria for pancreatitis, 159, 960, 960t, 974 ras oncogene(s), 37. See also K-ras oncogene. Rash. See also Skin. in hepatocellular carcinoma, 1571 in hookworm infection, 1925-1927, 1926f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxv

lxxvi

Index Rash (Continued) in strongyloidiasis, 1924 in ulcerative colitis, 2010 Raspberry, for colonic health, 2294t Reactive arthritis, gastrointestinal manifestations of, 558t, 563 Reactive gastropathy, 856-858, 856f-857f Reactive oxygen species in alcoholic liver disease, 1387 in intestinal ion transport, 1691 in intestinal ischemia, 2029 Receptor(s) acinar cell, 925, 925f in bile acid transport, 1078, 1078t in cellular proliferation, 32-33 death, 32, 33f, 1214, 1420 enzyme-coupled, 14-15 G protein–coupled. See G protein–coupled receptor(s). growth factor, 15-16, 15f guanylate cyclase, 15 ion channel–coupled, 15 serine/threonine kinase, 15 taste, 17 Toll-like in Helicobacter pylori infection, 835, 890 intestinal expression of, 26 tyrosine kinase, 14 in cellular proliferation, 33 oncogenes and, 36 tyrosine phosphatase, 15 Rectal. See also Anorectal. Rectal agnosia, 245-246 Rectal motor complexes, in colonic motility, 1665 Rectal tube, for sigmoid volvulus decompression, 2120 Rectoanal inhibitory reflex, 242-243, 1669 Rectocele, constipation in, 266, 267f Rectourethral fistula, 1634f, 1635 Rectovesicular fistula, 1634f, 1635 Rectum. See also Anorectum. adenoma of. See Colorectal adenoma. agenesis of, 1635 anatomy of, 242-243, 242f, 1616 augmentation of, for fecal incontinence, 255t, 256 biopsy of in Hirschsprung’s disease, 1639 in solitary rectal ulcer syndrome, 2051, 2051f bleeding from. See also Gastrointestinal bleeding, lower. in Dieulafoy’s lesions, 314 in Meckel’s diverticulum, 1628-1629 in rectal ulcers, 314-315, 315f in ulcerative colitis, 1983 cancer of. See Colorectal cancer. carcinoid tumors of, 481-482, 486, 487f-488f cavernous hemangioma of, 604, 605f compliance of calculation of, 248 loss of, fecal incontinence in, 244 Dieulafoy’s lesions of, bleeding in, 314 digital examination of, 246 in constipation, 271, 272t duplications of, 1632 emptying of. See Defecation. enlarged, 2143-2144 examination of, in constipation, 271, 272t fecal storage in, 1669-1670 filling of, 1669-1670 foreign bodies in, 398-399, 404, 404f, 2274 hypersensitivity of, in irritable bowel syndrome, 2095, 2097f inflammation of. See Proctitis. lymphoma of, 459 motility in. See Anorectal motility. palpation of, 2259 prolapse of constipation in, 267-268 in cystic fibrosis, 945 solitary rectal ulcer syndrome and, 2050-2051

Rectum (Continued) sensation in assessment of in constipated patient, 273-274 in fecal incontinence, 248 constipation in, 267 stapled transanal resection of, for defecatory disorders, 283 stromal tumors of, 465-466 ulcers of bleeding, 314-315, 315f solitary, 267-268, 2050-2052, 2051f varices of, bleeding in, 314 Red blood cell count, in ascitic fluid, 1521 Red blood cell dyscrasia, gastrointestinal manifestations of, 570-572 Red pepper, for functional dyspepsia, 193 Redox alteration, in alcoholic liver disease, 1385 Refeeding, for eating disorders, 134 Refeeding syndrome, 70-71, 86 in anorexia nervosa, 134 cardiovascular complications of, 70 clinical recommendations for, 71 gastrointestinal dysfunction in, 71 glucose intolerance in, 70-71 mineral depletion in, 70 Reflex(es) abdominal accommodation, 237 anocutaneous, 246 anorectal inhibitory, 242-243, 1669 emetic, 197-198, 198f gag, 687 gastroileal (gastrocolic), 264 Reflexology, 2288t Reflux, gastroesophageal. See Gastroesophageal reflux disease. Reflux esophagitis. See Esophagitis, reflux. Regurgitation acid. See Heartburn (pyrosis). approach to, 179 differential diagnosis of, 179 in gastroesophageal reflux disease, 714 pathophysiology of, 179 versus rumination, 179 Relaxation therapy, 2288t for nausea and vomiting, 2289, 2289t Renal. See also Kidney. Renal arterial vasoconstriction, in hepatorenal syndrome, 1546-1547, 1547f Renal colic, 200, 2063t Renal disease in acute pancreatitis, 982 versus cholecystitis, 1115 chronic angioectasia in, 316 gastrointestinal manifestations of, 578-581 small intestinal bacterial overgrowth in, 1774 in cirrhosis. See Hepatorenal syndrome. in Crohn’s disease, 1955 in eating disorders, 129t-130t gastrointestinal manifestations of, 578 in glycogen storage disease type I, 1263 after liver transplantation, 1608, 1608t liver transplantation and, 1597 obesity and, 104 polycystic autosomal dominant, polycystic liver disease in, 1589 autosomal recessive, fibrocystic liver disease in, 1590 portal hypertension in, 1500 in schistosomiasis, 1937 in tyrosinemia, 1270 Renal failure in acute liver failure, 1561t, 1563 angiodysplasia in, gastrointestinal, 578 in cirrhosis. See Hepatorenal syndrome. gastrointestinal ulcers in, 864 Rendezvous procedure, 1197, 1197f Renin, tumors secreting, 515 Reovirus, in biliary atresia, 1053 Reperfusion injury, in intestinal ischemia, 2029 Repifermin, for ulcerative colitis, 2001

Respiratory chain disorders, mitochondrial, 1278 Respiratory failure, in acute liver failure, 1561t, 1563 Respiratory luminal devices, esophageal injuries related to, 739 Respiratory muscle strength analysis, in nutritional assessment, 67 Respiratory system in acute pancreatitis, 977, 982 embryology of, 668, 669f endoscopy-related disorders of, 654-655, 657 in protein-energy malnutrition, 63 Restriction fragment length polymorphism (RFLP) analysis, 38, 39f Resuscitation. See also Fluid therapy. in acute pancreatitis, 977 in gastrointestinal bleeding, 288 in small intestinal obstruction, 2110 RET gene, in Hirschsprung’s disease, 1637 Retching. See also Vomiting. definition of, 197 in emetic reflex, 198 versus vomiting, 197 Reticulothalamic tract, in visceral pain transmission, 343 Retinal pigmented epithelium, congenital hypertrophy of, in familial adenomatous polyposis, 2178f, 2180 Retinoblastoma, in esophageal cancer, 750 Retinol. See Vitamin A (retinol). Retinol-binding protein, serum, in nutritional assessment, 67-68, 67t Retinopathy, Purtsher’s, in acute pancreatitis, 982 Retroanastomotic hernia, 393-395 Retroperitoneum fibrosis of, 620-621, 621f hemorrhage of, 620 inflammatory and fibrotic conditions of, 620-621, 621f varices of, 1494 Revascularization, surgical for acute mesenteric ischemia, 2032 for intestinal angina, 2045-2046 Reye’s syndrome aspirin and, 1430 jaundice in, 327 pancreatitis in, in children, 939 Reynold’s pentad, in cholangitis, 1118 Rhabdomyosarcoma, embryonal, biliary, 1184 Rheumatoid arthritis drug-induced liver disease in, 1417 drug-related complications in, 559 gastrointestinal manifestations of, 558-559, 558t hepatic involvement in, 558-559 with hepatitis, 559 juvenile adult-onset, gastrointestinal and hepatic manifestations of, 559-560 Reye’s syndrome in, 1430-1431 NSAID therapy for, ulcers in, 559 obesity and, 105 with Sjögren’s syndrome, 559 small intestinal bacterial overgrowth in, 1774 Rheumatoid vasculitis, 2047 Rheumatologic diseases, gastrointestinal manifestations of, 557-564, 558t Rhizomelic chondrodysplasia punctata, 1275 Rhubarb, for colonic health, 2294t Rib, slipping, abdominal pain in, 165 Ribavirin, for hepatitis C, 1329 Riboflavin (vitamin B2), 53t-55t, 81, 81t absorption of, 1718t, 1720 reference nutrient intake for, 1718t Richter’s hernia, 379, 386 Rickets, vitamin D–resistant, malabsorption in, 1759t-1762t Rickettsial infections, hepatic manifestations of, 1353-1354 Riedel’s lobe, 1205 Rifabutin, for Helicobacter pylori infection, 842, 843t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Rifampin hepatotoxicity of, 1434 for intestinal tuberculosis, 1878 jaundice from, 325 in primary biliary cirrhosis, 1486, 1486t for pruritus, 1063 for Whipple’s disease, 1841t Rifaximin for Crohn’s disease, 1963 for diverticulosis, 2077 for hepatic encephalopathy, 1545-1546 for irritable bowel syndrome, 2102 prophylactic, for traveler’s diarrhea, 1874-1875, 1875t for small intestinal bacterial overgrowth, 1777 Riley-Day syndrome, gastrointestinal manifestations of, 580 Riluzole, hepatotoxicity of, 1436-1437 Rimonabant, for obesity, 111 Ritonavir, hepatotoxicity of, 532, 1430 Rituximab for cryoglobulinemia, 1320 for gastric diffuse large B cell lymphoma, 452 for gastric MALT lymphoma, 451 RNA ribosomal, 1208 transfer, 1208 Rockall Score for upper gastrointestinal bleeding, 293, 294t, 878 Rocky Mountain spotted fever, hepatic manifestations of, 1353-1354 Rodent tapeworm infection, 1932-1933 Rofecoxib cardiovascular risk of, 875 for NSAID ulcer prophylaxis, 874-875 Rokitansky-Aschoff sinuses, 1050, 1146, 1147f, 1148-1149 Rome II diagnostic criteria, for functional abdominal pain syndrome, 165, 165t Rosiglitazone hepatotoxicity of, 1435 for nonalcoholic fatty liver disease, 1410 for ulcerative colitis, 2002 Rosuvastatin, hepatotoxicity of, 1433 Rotavirus infection, 1691-1692, 1869-1871, 1870f, 1870t Rotor’s syndrome hyperbilirubinemia in, 1229 jaundice in, 326, 326t Roundworms. See Nematode (roundworm) infection. Roux-en-Y gastric bypass, 115, 116t-117t. See also Bariatric surgery. gastric motility disorders after, 808 hernia after, 394-395 iron-deficiency anemia after, 321 Rovsing’s sign, in appendicitis, 2062 Rubber band ligation, for hemorrhoids, 2261, 2263t Rumination, 203-204, 810 versus regurgitation, 179 RUNX3, in gastric cancer, 894 Rye, 1801, 1801f

S

S-adenosyl-l-methionine (SAMe) for alcoholic liver disease, 1397 for cholestasis of pregnancy, 631-632 deficiency of, in alcoholic liver disease, 1386-1387, 1386f for liver disease, 2295t, 2296 S cells, 7 Saccharomyces boulardii biotherapy for antibiotic-associated diarrhea, 1890 for Clostridium difficile-associated diarrhea and colitis, 1902 for diarrhea, 2293 for inflammatory bowel disease, 2293 for ulcerative colitis, 2000 Sacral nerve damage to, constipation in, 269

Sacral nerve (Continued) stimulation of colonic motility and, 1672 for constipation, 283 for fecal incontinence, 256, 257f Sacroiliitis, in ulcerative colitis, 2011 Saline enema, for constipation, 277t, 280 Saline infusion test, in fecal incontinence, 250 Saline transfusion, for gastrointestinal bleeding, 288 Salivation, esophageal acid clearance and, 711 Salmonella enterica, 1861 Salmonella spp., 1861 carriers of, 1863 in food poisoning, 1878-1879, 1879t-1880t mucosal invasion and intramucosal multiplication of, 1846 pathogenicity of, 1862 in traveler’s diarrhea, 1873 typing of, 1861 Salmonella typhi, 1864. See also Typhoid fever. carriers of, 1866-1867 Salmonellosis, 1861-1864 clinical features of, 1863, 1863t colitis in, 1863 in elderly, 1876 epidemiology of, 1861-1862, 1862f hepatic manifestations of, 1352 in HIV/AIDS, 527t, 529 microbiology of, 1861 pathogenesis of, 1862 predisposing conditions in, 1862-1863, 1862t in pregnancy, 1872 treatment of, 1863-1864, 1864t-1865t typhoidal. See Typhoid fever. Salpingitis, acute, versus appendicitis, 2063t Salt-losing state, after colectomy with ileostomy, 2016 Sanban virus infection, 1346-1347 Santorini, duct of, 911, 911f SAPE (sentinel acute pancreatitis event) paradigm, for chronic pancreatitis, 987-988 Sarcoidosis cutaneous manifestations of, 365, 366f gastritis in, 853 gastrointestinal manifestations of, 590 hepatic, 588-590 clinical manifestations of, 589-590 granulomas in, 588, 589f, 589t jaundice in, 327-328 portal hypertension in, 1500-1501 Whipple’s disease and, 1840 Sarcoma esophageal, 768 gastrointestinal stromal tumors as, 462 hepatic, 1583 Kaposi’s, in HIV/AIDS, 355-356, 355f, 534 Sarcopenia, in Crohn’s disease, 1954 Sargramostim, for Crohn’s disease, 1970 Satiety in eating disorders, 122-123 regulation of, 1696 by gastrointestinal peptides, 18, 18t stomach and, 802-803, 803f Scaffolding proteins, in intestinal ion transport, 1676f, 1679, 1692-1693 Schabadasch’s plexus, 1662 Schatzki’s ring, 175, 670t, 672-674, 673f Schilling test, for vitamin B12 malabsorption, 1752 Schistosoma spp., life cycle of, 1361-1362, 1361f, 1935-1936 Schistosomiasis central nervous system manifestations of, 1937 clinical features of, 1936-1937, 1936f-1937f diagnosis of, 1937-1939, 1937f, 1938t epidemiology of, 1935 hepatic, 1361-1362, 1362f hepatosplenic, 1936-1937, 1938t, 1939 intestinal, 1935-1939, 1936f-1937f Katayama fever in, 1936 nephropathy in, 1937 portal hypertension in, 1499

Schistosomiasis (Continued) pulmonary, 1937 treatment of, 1939 Sciatic foramen hernia, 391-392 Scintigraphy in abdominal abscess, 414 in appendicitis, 2064-2065 in Crohn’s disease, 1957-1958 in diverticular bleeding, 2087-2088, 2087f in gastric motility assessment, 796f, 803-805 in gastrointestinal bleeding, 292 hepatobiliary. See Cholescintigraphy. in jaundice, 333 in lower gastrointestinal bleeding, 310 in obscure gastrointestinal bleeding, 319 in protein-losing gastroenteropathy, 442 in small intestinal motility assessment, 1653 somatostatin receptor in carcinoid tumors, 485, 485f in pancreatic endocrine tumors, 516, 516t, 517f-518f, 519 Scirrhous carcinoma colorectal, 2211, 2215 hepatocellular, 1574-1575 Scleroderma candidal esophagitis in, 741-742 chronic intestinal pseudo-obstruction in, 2135, 2136f constipation in, 269 cutaneous, 363 esophageal, 560 dysphagia in, 174 gastroesophageal reflux disease in, 715 gastrointestinal, 558t, 560 malabsorption in, 560, 1758 vascular lesions in, 601 Sclerosing cholangitis autoimmune, 1467 classification of, 1153, 1154t primary, 1153-1167, 1233 asymptomatic, 1158 versus autoimmune hepatitis, 1057, 1155 autoimmune hepatitis with, 1466-1467, 1467t biliary epithelial cells in, 1157 biliary surgery for, 1166 cholangiocarcinoma in, 1154-1155, 1162-1163, 1163f, 1166, 1172, 1176 cholestasis in, 1161-1162 clinical features of, 1159-1160, 1159t colonic neoplasia in, 1163-1164 complications of, 1161-1165 in Crohn’s disease, 1156, 1955 in infants and children, 1057 diagnosis of, 1153-1155 differential diagnosis of, 1154-1155, 1154t diseases associated with, 1153, 1154t endoscopic retrograde cholangiopancreatography in, 1057, 1154, 1160, 1160f, 1192-1193, 1194f epidemiology of, 1155-1156 etiology and pathogenesis of, 1156-1157 gallbladder carcinoma and, 1178 gallbladder polyp in, 1151 gallstone disease in, 1162 genetic factors in, 1156 imaging studies in, 1160, 1160f immunologic factors in, 1156-1157 in infants and children, 1056-1058 infectious agents in, 1157 in inflammatory bowel disease, 1154, 1156 laboratory findings in, 1160 liver transplantation for, 1163, 1166-1167, 1595, 1601-1602 natural history of, 1157-1159 nutritional complications of, 1162, 1165 osteoporosis in, 1162 pathology of, 1160-1161, 1161f percutaneous transhepatic cholangiography for, 1188, 1189f peristomal varices in, 1164 physical findings in, 1159 portal hypertension in, 1498-1499 prognostic models of, 1158-1159, 1159t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxvii

lxxviii

Index Sclerosing cholangitis (Continued) pruritus in, 1161-1162, 1165 recurrent, 1166-1167 after solid organ transplantation, 540, 540f small-duct, 1153, 1158 symptomatic, 1158 symptoms of, 1159, 1159t toxins in, 1157 treatment of, 1164-1167 endoscopic, 1165-1166 medical, 1164-1165, 1164t percutaneous, 1166 surgical, 1166-1167 in ulcerative colitis, 1156, 2012 in infants and children, 1057 vascular injury in, 1156-1157 secondary, 1153, 1154t Sclerosing pancreatitis, lymphoplasmacytic, 986-987, 991, 992f Sclerosing pancreatocholangitis, 1155 Sclerosis amyotrophic lateral dysphagia in, 688 gastrointestinal manifestations of, 581 hepatoportal, 1500 multiple, gastrointestinal manifestations of, 581 systemic, progressive. See Scleroderma. tuberous hamartomatous polyps in, 2186 pancreatic endocrine tumors in, 495 Sclerotherapy complications of, 657 esophageal injury from, 738-739 for gastric variceal bleeding, 1512 for hemorrhoids, 2261, 2263t for variceal bleeding, 307, 1503, 1503t Scombroid poisoning, 1884 Scurvy, in vitamin C deficiency, 368f, 369 Scybalum, stercoral ulcers and, 2052 Seafood poisoning, 1884 Second messengers, in intestinal ion transport, 1682, 1682f, 1692-1693, 1693f Secretin, 7-8, 922, 923f actions of, 7-8 clinical application of, 8 feedback regulation of, 928 in pancreatic enzyme secretion, 926, 1697 receptors for, 7 regulation of, by intraluminal releasing factors, 17 Secretin test, 928-929, 928t in chronic pancreatitis, 999-1000 in diarrhea, 226 in gastrinoma, 503 in sphincter of Oddi dysfunction, 1069 Sedation for endoscopy, complications of, 654-655, 655t in pregnant patients, 627 Segmentectomy, hepatic, for cholangiocarcinoma, 1176 Seizure in celiac disease, 1806 in shigellosis, 1859-1860 Selenium absorption of, 1726 dietary, 56t-57t, 80t, 81 esophageal cancer and, 747 in short bowel syndrome, 1786t supplementation of, colorectal adenoma risk and, 2164 Selenium-75–labeled homotaurocholic acid test, for bile acid malabsorption, 1753 Self-expanding metallic stents covered, 1190 for distal bile duct strictures, 1193-1194, 1196f Self-help, for eating disorders, 132-133 Semaxanib, for hemangioma, 605 SEN virus infection, 1346-1347 Senescence, replicative, 32, 33f Sengstaken-Blakemore tube, for gastric variceal bleeding, 1512

Senna for colonic health, 2294t for constipation, 277t, 279 for opioid-induced constipation, 2284 Sensitization peripheral, in small intestinal motility, 1650 visceral. See Visceral sensitization. Sensory function of esophagus, 685-686 of stomach, 801-802, 801f Sensory neurons, extrinsic primary, 5 Sensory testing in esophageal motility disorders, 699-700 rectal, in fecal incontinence, 248 Seprafilm, postoperative abscess and, 412 Sepsis. See also Bacterial infection. versus acute liver failure, 1557-1558 cholestasis of after hematopoietic stem cell transplantation, 547t, 548, 548f jaundice in, 329, 1353 gastrointestinal complications in, 581-584, 582f hepatic dysfunction during, 582-583, 582f hyperbilirubinemia in, 582 jaundice and, 582-583, 582f, 1353 pelvic, after ileal pouch–anal anastomosis, 2019, 2019f after rubber band ligation of hemorrhoids, 2261 umbilical, portal vein thrombosis in, 1377 Septic joint, in Crohn’s disease, 1954 Sequential neural electrical gastric stimulation, for gastric motility disorders, 813t, 814 Serine/threonine kinase, receptor, 15 in cellular proliferation, 33 Serologic testing in amebiasis, 1909 in celiac disease, 1807-1809, 1807t, 1809f in Crohn’s disease, 1959 for Helicobacter pylori infection, 840-841, 840t in jaundice, 334 in primary biliary cirrhosis, 1480 Seronegative spondyloarthropathies, gastrointestinal manifestations of, 558t, 563 Serotonin, 11 actions of, 11, 12f in carcinoid heart disease, 482-483 in carcinoid syndrome, 483-484, 483f in carcinoid tumors, 484, 484t colonic ischemia and, 2040 colonic motility and, 1672 in eating disorders, 122-123 in emetic reflex, 197-198, 198f in gastrointestinal tract, 5, 11 in intestinal ion transport, 1691 in irritable bowel syndrome, 2098 in pancreatic enzyme secretion, 1697 receptors for, 11 in visceral sensitization, 167 Serotonin/norepinephrine reuptake inhibitors, for gastrointestinal symptoms, 349 Serotonin receptor antagonists clinical applications of, 11 for vomiting, 207-208 Serotonin reuptake inhibitors, selective for esophageal hypersensitity, 704 for functional abdominal pain syndrome, 170 for gastrointestinal symptoms, 349 hepatotoxicity of, 1436 for irritable bowel syndrome, 2102 Sertraline for anorexia nervosa, 133 for pruritus, in primary biliary cirrhosis, 1486-1487, 1486t Serum-ascites albumin gradient, 1522f, 1523-1524, 1523t corrected, 1524 high, 1536-1538 low, 1536 Serum markers, in gastric cancer, 900 Sessile serrated adenoma, 2174, 2174f Seton, for anal fistula, 2268-2269, 2268f

Severe acute respiratory syndrome (SARS) gastrointestinal manifestations of, 581 hepatitis in, 1349 Sevoflurane, hepatotoxicity of, 1449, 1449t Sexual abuse. See Abuse history. Sexuality in Crohn’s disease, 1972 after ileal pouch–anal anastomosis, 2021-2022 Sharp objects, as foreign bodies, 402, 402f-403f Shenxiahewining, for functional dyspepsia, 2290t, 2291 Shiga-like toxins (STX cytotoxins I and II), in enterohemorrhagic Escherichia coli, 1856 Shiga toxin, 1859 Shigella spp., 1857-1861 carriers of, 1860-1861 cytotoxins of, 1859 in food poisoning, 1879t-1880t mucosal invasion and intramucosal multiplication of, 1846 pathogenicity of, 1858, 1858f subgroups of, 1857-1858 in traveler’s diarrhea, 1873 Shigellosis, 1857-1861 clinical features of, 1859-1860, 1859t cytotoxins in, 1859 diagnosis of, 1858f, 1860 epidemiology of, 1858 hepatic manifestations of, 1352 in HIV/AIDS, 527t, 529 versus intestinal amebiasis, 1909, 1909t microbiology of, 1857-1858 pathogenesis of, 1858, 1858f in pregnancy, 1872 treatment of, 1860-1861, 1865t Short bowel syndrome, 1779-1796 calcium oxalate kidney stones in, 1790, 1790f in children, 1779, 1780t complications of, 1788-1791 diarrhea in, 228-229, 1783 etiology of, 1779, 1780t gallstones in, 1788-1789 incidence and prevalence of, 1779 intestinal adaptation in, 1783 pharmacologic enhancement of, 1793 lactic acidosis in, 1790-1791 liver disease in, 1790 malabsorption in, 1779-1783, 1780f bile acid, 1782 colonic carbohydrate salvage in, 1780-1781, 1781f fat, 1782-1785 medication, 1785 vitamin B12, 1782 water and electrolyte, 1781-1782, 1781t, 1782f pathophysiology of, 1779-1783, 1780f absorptive surface area loss in, 1780-1782, 1780f-1781f ileocecal valve loss in, 1783 site-specific endocrine cell and gastrointestinal hormone loss in, 1782 site-specific transport process loss in, 1782 water and electrolyte malabsorption in, 1781-1782, 1781t, 1782f quality of life in, 1795 small intestinal bacterial overgrowth in, 1783 survival rate for, 1795 treatment of algorithm for, 1794f diet in, 1783-1786, 1785t-1787t in extensive small intestinal resection and partial colectomy, 1784-1786 fluid and electrolytes in, 1784, 1784t glucagon-like peptide 2 in, 1793 glutamine in, 74-75 growth hormone in, 1793 home parenteral nutrition in, 1785-1788, 1788f-1789f intestinal lengthening procedures in, 1791, 1791f-1792f intestinal tapering procedure in, 1791, 1793f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Short bowel syndrome (Continued) intestinal transplantation in, 1791-1793, 1792t in limited ileal resection, 1783-1784 medical, 1783-1786 nutritional therapy in, 74-75, 82-83 oral rehydration solution in, 1781-1782, 1782f, 1784 pharmacologic, 1783-1784, 1784t, 1793 support group in, 1786 surgical, 1791-1793 Short-chain fatty acids, intestinal transport of, 1685-1686 Short stature, in celiac disease, 1804 Shunt(s) arteriovenous, in parenteral nutrition, 87 intrapulmonary, in hepatopulmonary syndrome, 1550 peritoneovenous, for ascites, 1540 portosystemic. See Portosystemic shunt. Shwachman-Bodian-Diamond syndrome (SBDS) gene in pancreatitis, 937 in Shwachman-Diamond syndrome, 954 Shwachman-Diamond syndrome, 953-954 bone marrow dysfunction in, 953-954 clinical features of, 953-954, 953t growth and development in, 954 molecular defects in, 954 neutropenia in, 953-954 pancreatic insufficiency in, 953, 953f treatment of, 954 Sickle cell crisis, 570 Sickle cell disease gallstone disease in, 1106 hepatic manifestations of, 570-572, 570f-571f salmonellosis and, 1862 Sigmoid colon, 1616 diverticulitis of, versus appendicitis, 2063t volvulus of, 2116-2117, 2118f-2119f, 2120 Sigmoid flotation maneuver, in diverticulosis, 2077 Sigmoidoscopy. See also Endoscopy. bleeding in, 658-659, 659f in Clostridium difficile-associated diarrhea and colitis, 1897 in colorectal adenoma, 2166 in colorectal cancer, 2221t, 2223-2224 complications of, 658-660 in constipation, 272 in diarrhea, 220, 223-225 in familial adenomatous polyposis, 2182-2183 flexible, 2260 in lower gastrointestinal bleeding, 309-310 perforation in, 659-660, 659f for sigmoid volvulus decompression, 2120 in ulcerative colitis, 1985 Signal transduction in gastrointestinal tract, 13-15, 13f, 14t oncogenes and, 36-37 Signaling pathways in cellular proliferation, 32-33, 34f in gastric acid secretion, 821f, 823-824, 824f in intestinal adaptation, 1729 of intestinal (commensal) bacteria, 1771-1772 in oncogenesis, 40-41, 41f Signet ring colorectal cancer, 2215 Silbutramine, for obesity, 110t, 111 Sildenafil for achalasia, 701 for portopulmonary hypertension, 1552 Silymarin for alcoholic liver disease, 1398 hepatotoxicity of, 1456 for liver disease, 2295-2296, 2295t Simethicone for abdominal distention and bloating, 239 for functional dyspepsia, 193 Simvastatin hepatotoxicity of, 1433 for portal hypertension, 1502-1503 for primary biliary cirrhosis, 1486 Single nucleotide polymorphisms (SNPs), 38

Single photon emission computed tomography, in gastric motility assessment, 805 Sinus tracts, in Crohn’s disease, 1949 Sinusoidal obstruction syndrome, 1375-1377 clinical features of, 1376 diagnosis of, 1376 drug-induced, 1445t, 1456 etiology of, 1376 after hematopoietic stem cell transplantation, 546-548, 547t, 548f liver transplantation for, 1603 pathology of, 1376-1377 prevention and treatment of, 1377 Sinusoids, hepatic. See Liver, sinusoids of. Sipple’s syndrome, 494 Sirolimus gastrointestinal side effects of, 541 for liver transplantation, 1606t Sister Mary Joseph’s nodule, 365 Sitosterolemia, 1085t, 1703 Sjögren’s syndrome gastrointestinal manifestations of, 558t, 561 oral manifestations of, 353 rheumatoid arthritis with, 559 Skin amebiasis of, 357 in candidiasis, 354-355, 354f in celiac disease, 1806 in connective tissue disorders, 363 in Crohn’s disease, 1950, 1954 in dermatomyositis, 364, 364f in eating disorders, 129t-130t in gastrointestinal malignancies, 363-365 in hepatitis, 366, 366f in hepatitis A, 1282 in HIV infection, 355-356, 355f in inflammatory bowel disease, 359-360, 359f in Kaposi’s sarcoma, 355-356, 355f larva migrans of, 367, 367f, 1925-1927, 1926f in liver disease, 365-366, 366f in lymphoma, 356 in metastatic gastrointestinal disease, 365 in obesity, 105-106 in parasitic disease, 367, 367f pigmentation of in Cronkhite-Canada syndrome, 364 in gastrointestinal stromal tumors, 473 in hereditary hemochromatosis, 1243 in Peutz-Jeghers syndrome, 363, 363f in tropical sprue, 1824, 1825f in Whipple’s disease, 1837 in polyposis syndromes, 363-364, 363f in porphyria, 1267-1268, 1267t in protein-energy malnutrition, 63-64 in vascular disease, 360-363 in vitamin deficiencies, 368-369, 368f Skin tags, anal, 2262-2263, 2265f Skinfold thickness, in nutritional assessment, 65-66, 66t SLC1A1 amino acid transporter, 1716-1717, 1716t SLC4 bicarbonate transporter, 1685 SLC10A2 sodium–bile acid cotransporter, in bile acid malabsorption, 1756 SLC26 multifunctional anion exchanger, 1685 SLC36A1 amino acid transporter, 1716, 1716f, 1716t Sleep colonic motility and, 1671, 1671f gastroesophageal reflux disease and, 725 obesity and, 102 Sleep apnea, obstructive after bariatric surgery, 118 obesity and, 105-106 Sleep-related eating disorder, nocturnal, 121, 122f, 125 Slipping rib syndrome, abdominal pain in, 165 Slow wave colonic, 1660 gastric, 789-791, 790f-791f small intestinal, 1654, 1654f SMAD, in esophageal cancer, 749-750

SMAD4 (DPC4) gene, 40 in colorectal cancer, 2200t, 2202 in juvenile polyposis, 2187 Small bowel bacterial overgrowth. See Bacterial overgrowth, small intestinal. Small bowel feces sign, in small intestinal obstruction, 2108-2109, 2110f Small cell carcinoma, esophageal, 767 Small intestinal adenocarcinoma, 2148-2152 clinical presentation in, 2148-2149 in Crohn’s disease, 1972-1973 diagnosis of, 2147f, 2149-2151, 2149f-2150f, 2149f incidence of, 2147 in Peutz-Jeghers syndrome, 2148 prognosis in, 2152 risk factors for, 2147 sites of, 2145, 2146f staging of, 2146, 2146t treatment of, 2151-2152 Small intestinal biopsy in celiac disease, 1799-1800, 1799f, 1809-1810, 1810f in diarrhea, 223, 225-226 in giardiasis, 1912f, 1913 in malabsorption, 1746-1747, 1747f-1748f, 1747t, 1766f PAS-positive cells in, differential diagnosis of, 1840 in tropical sprue, 1827 in Whipple’s disease, 1838-1840, 1838f-1839f Small intestinal enteroclysis, in malabsorption, 1749 Small intestinal immunoproliferative disease, 455-457 clinical features of, 456-457 diagnosis and staging of, 457 epidemiology of, 455 etiology and pathogenesis of, 455 malabsorption in, 1829t, 1830-1831 pathology of, 455-456, 456t treatment of, 457 Small intestinal manometry, 1652-1653 interdigestive, 1655f, 1656 postprandial, 1655f, 1656 Small intestinal motility, 1643-1658, 2121-2122. See also Intestinal motility. anatomy and, 1644 disorders of in acute illness, 1656-1657 approach to patient with, 1657 in diabetes mellitus, 1649-1650, 1657 diseases and clinical settings associated with, 1650t, 1656-1657 extrinsic afferent dysfunction in, 1650-1651 intrinsic neural dysfunction in, 1649-1650 in irritable bowel syndrome, 1656 mechanisms underlying, 1649-1651 in metabolic disorders, 1657 in pregnancy, 1657 small intestinal bacterial overgrowth in, 1773 smooth muscle dysfunction in, 1649-1651 enteric (intrinsic) neurons in afferent, 1646 efferent, 1646 interneurons and, 1646-1647 extrinsic neurons in afferent, 1647 efferent, 1647-1648 gastrointestinal hormones in, 1648 integrative regulation of, 1648-1649 interdigestive (fasting) motor cycle in, 1648-1649 interstitial cells of Cajal in, 1644-1645, 1644f in irritable bowel syndrome, 2095, 2096f measurement of, 1651-1653 breath tests for, 1652-1653 fluoroscopy for, 1652-1653 magnetic resonance imaging for, 1652-1653 manometry for, 1652-1653 multiple intraluminal impedance for, 1652-1653 muscle contraction recording for, 1652

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxix

lxxx

Index Small intestinal motility (Continued) scintigraphy for, 1653 single cell functional evaluation for, 1651-1652 spatiotemporal resolution for, 1651, 1651f transit studies for, 1652-1653 ultrasonography for, 1652 wall motion studies for, 1652 neural control system for, 1645-1648, 1646f, 2121, 2122f-2123f, 2123-2124 central connections of, 1648 extrinsic, 1647-1648 intrinsic, 1646-1647 normal, 1654-1656 fasting pattern of, 1654-1656, 1655f fed pattern of, 1654-1656, 1655f at fixed point, 1654, 1654f during migration, 1654 patterns of, 1654-1656, 1655f patterns of, 1654-1656, 1655f peristalsis pattern of, 1648, 1655 smooth muscle in, 1644, 1644f-1645f Small intestinal obstruction, 2105-2116 abdominal pain in, 154t, 158 abdominal radiography in, 2107-2108, 2107f-2108f in ascariasis, 1922, 1922f barium radiography in, 2109-2110, 2112f clinical presentation in, 2106-2107 computed tomography enteroclysis in, 2109, 2111f computed tomography in, 2106t, 2108-2109, 2108f-2111f constipation in, 269 etiology of, 2105-2106, 2106t in gallstone ileus, 2115-2116, 2115f Gastrografin studies in, 2109-2110 in hernias, 2105-2106 history in, 2106-2107 in intra-abdominal adhesions, 2105, 2107f, 2108, 2112 in intussusception, 2114-2115, 2114f laboratory findings in, 2107 malignant, 2106, 2113-2114 in midgut volvulus, 2116, 2116f morbidity and mortality of, 2112, 2113t pathophysiology of, 2106 physical examination in, 2107 postoperative, early, 2113 recurrent, 2112 small bowel feces sign in, 2108-2109, 2110f in small intestinal tumors, 2148 strangulated, 2109-2110, 2110f treatment of, 2110-2112 medical, 2110-2111 results of, 2112, 2113t surgical, 2111-2112 tumor-related, 2106, 2113-2114 whirl sign in, 2108-2109, 2110f Small intestinal transit time, 1652-1653 interdigestive, 1656 meal-related, 1655-1656 Small intestinal tumors, 2145-2154 adenocarcinoma as. See Small intestinal adenocarcinoma. benign, 2146, 2151 bleeding in, 318, 318f epidemiology of, 2145 etiology of, 2147 metastatic, 2153, 2153f pathology of, 2145-2147 risk factors for, 2147-2148, 2148t small intestinal obstruction in, 2106 treatment of, 2151-2152 types of, 2147t Small intestinal ulcers bleeding in, 318 causes of, 2050t diffuse, 2054-2058. See also Ulcerative enteritis. isolated, 2049-2054 nonspecific or idiopathic, 2049-2050, 2050t NSAID-induced, 2052-2054, 2053f

Small intestine. See also Duodenum; Ileum; Jejunum. absorptive capacity of, in pregnancy, 626 adenocarcinoma of. See Small intestinal adenocarcinoma. adenoma of in familial adenomatous polyposis, 2148 pathology of, 2146 progression of, 2148 adventitia of, 1619 anatomy of, 1615, 1621-1622, 1621f, 1644, 1644f atresia of, 1632. See also Intestine(s), atresia of. bacterial overgrowth of. See Bacterial overgrowth, small intestinal. basolateral membrane of carbohydrate transfer across, 1711-1712 iron transport across, 1724-1725 protein transfer across, 1717 bile acid transport in, 1083-1084 biopsy of. See Small intestinal biopsy. Brunner’s glands of, 1621, 1621f brush border membrane of carbohydrate transfer across, 1708, 1708f, 1708t hydrolases of, 1708, 1708f, 1708t protein transfer across, 1713-1715, 1713f, 1714t triglyceride transfer across, 1702-1703, 1703f cancer of. See Small intestinal adenocarcinoma. carcinoid tumors of, 480, 480t, 486, 487f-488f, 2145 obstruction in, 2106 cells of, 1621-1622 Crohn’s disease of. See Crohn’s disease. crypts of Lieberkühn of, 1619f, 1621-1622 aberrant, colorectal adenoma and, 2159 in celiac disease, 1800 in malabsorption, 1746-1747, 1748f in radiation enteritis, 642-643, 643f radiation-induced apoptosis of, 639 in tropical sprue, 1825, 1826f in ulcerative colitis, 1982, 1982f Dieulafoy’s lesions of, bleeding in, 318 in digestion and absorption, 1780, 1780f diverticula of, bleeding in, 318 double-balloon enteroscopy of, lesions on, 320, 320t duplications of, 1631-1632 fluid and electrolyte transport in, 1675-1694. See also Fluid and electrolyte transport. foreign bodies in, 399 immunoproliferative disease of. See Small intestinal immunoproliferative disease. innervation of, 1623, 1645-1648, 1646f intestinal gas in, 233-234, 234f-235f ischemia of, focal segmental, 2038 leiomyoma of, 2145-2146 lipoma of, 2145-2146, 2147f lymphatic drainage of, 1617, 1623 lymphoma of. See Lymphoma, small intestinal. microvilli in, 1617f, 1618 motility in. See Small intestinal motility. mucosa of, 1617-1619, 1617f-1621f in cystic fibrosis, 943 diet-related changes in, 1727-1728, 1728f gluten-free diet effects on, 1799f, 1800 postresection changes in, 1726-1727 muscularis of, 1619, 1621f neuroendocrine (enteroendocrine) cells of. See Neuroendocrine (enteroendocrine) cells. obstruction of. See Small intestinal obstruction. plicae circulares of, 1615 pseudo-obstruction of. See Pseudo-obstruction, intestinal. radiation injury to, 642-647. See also Radiation enteritis. resection of, short bowel syndrome after, 1784-1786. See also Short bowel syndrome. serosa of, 1619

Small intestine (Continued) slow wave of, 1654, 1654f smooth muscle of, 1644, 1644f-1645f dysfunction of, 1649-1651 fasting motor cycle of, 1648-1649, 1655f, 1656 fed motor cycle of, 1655-1656, 1655f fixed-point contractions of, 1654, 1654f innervation of, 1645-1648, 1646f migrating contractions of, 1654 stromal tumors of, 465 submucosa of, 1619 transplantation of, for complications with parenteral nutrition, 86-87 tuberculosis of, 1877 ulcers of. See Small intestinal ulcers. vascular supply to, 1617, 1622-1623 villi of, 1621, 1621f. See also Villus (villi). wall of, layers of, 1644, 1644f Small left colon syndrome, 1633 Small vessel hyalinosis, 590 Smoking acute pancreatitis and, 969 alcoholic liver disease and, 1392-1393 alcoholic pancreatitis and, 988-989 cessation of for chronic pancreatitis, 1005 for gastroesophageal reflux disease, 721 chronic pancreatitis and, 989 Crohn’s disease and, 1946 esophageal cancer and, 747-748 gastric cancer and, 892 hepatitis C and, 1325-1326 hepatocellular carcinoma and, 1577 hereditary pancreatitis and, 951 obesity and, 102 pancreatic cancer and, 1017 peptic ulcer disease and, 862-863 ulcerative colitis and, 1978-1979 Smooth muscle colonic, 1660-1661, 1661f ion channels in, 1661 nocturnal suppression of, 1671, 1671f disorders of, constipation in, 269 gastric, 790-791, 792f-793f small intestinal, 1644, 1644f-1645f. See also Small intestine, smooth muscle of. Smooth muscle antibodies, in autoimmune hepatitis, 1461, 1465 Snare cautery, for colorectal cancer, 2237, 2238f Snuff use, gastric cancer and, 892 Soapsuds enema, for constipation, 277t, 280 Social support, in biopsychosocial model, 342 Socioeconomic status constipation and, 262 Helicobacter pylori infection and, 833-834 ulcerative colitis and, 1976 Sodium. See also Hyponatremia. absorption of colonic, 263 in short bowel syndrome, 1781, 1781t balance of, with ileostomy, 2016 congenital diarrhea and, 213, 1641 dietary, 52t gastric cancer and, 892 restriction of, for ascites, 1536 in oral rehydration solution, 1885 transport of apical sodium channel for, 1681, 1681f sodium glucose transporters for, 1681, 1681f sodium-hydrogen exchangers in, 1681-1682, 1681f solute-coupled, 1681, 1681f urinary excretion of, and ascites, 1536-1537 Sodium benzoate for hepatic encephalopathy, 1546 for urea cycle defects, 1273-1274 Sodium bicarbonate, for ascites, 1538 Sodium cromoglycate, for eosinophilic gastrointestinal disorders, 434 Sodium-glucose cotransporters (SGLTs), 1681, 1681f, 1710-1711

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Sodium-hydrogen exchanger (NHE), 1681-1682, 1681f regulation of, 1682, 1682f Sodium phenylacetate for hepatic encephalopathy, 1546 for urea cycle defects, 1273-1274 Sodium phosphate, colonic ischemia from, 2040 Sodium picosulfate, for constipation, 277t, 279 Sodium polystyrene sulfonate, colonic ischemia from, 2040 Sodium-potassium-chloride cotransporter(s) (NKCCs), 1683, 1683f Sodium pump (Na+,K+-ATPase) in glucose transport, 1710, 1710f, 1782f in intestinal transport, 1676f, 1677 Sodium-taurocholate cotransporting polypeptide (NTCP), in bile acid transport, 1081, 1082f, 1083t Sodium–bile acid cotransporter (SLC10A2), in bile acid malabsorption, 1756 Sodium-potassium ratio, urine, and ascites, 1537 Solitary rectal ulcer syndrome, 267-268, 2050-2052, 2051f Solubilization of fat, reduced, 1738, 1738t malabsorption and, 1736 Soluble liver antigen antibodies, in autoimmune hepatitis, 1470 Solutes, bile acid transport of, 1080 Somatic-parietal pain, 152 Somatostatin, 9 actions of, 5 for acute pancreatitis, 968, 980 clinical applications of, 9 in gastric acid secretion, 817-819, 820f-822f, 823 in gastrointestinal tract, 9 inhibitory action of, 5 in intestinal ion transport, 1690 for peptic ulcer bleeding, 301 for portal hypertension, 1502 receptors for, 9 for variceal bleeding, 307 Somatostatin analogs. See also Octreotide. for carcinoid tumors, 489 for gastrointestinal fistula, 422 for insulinoma, 498 malabsorption with, 1757t for pancreatic endocrine tumors, 521 for VIPoma, 510-511 Somatostatin receptor scintigraphy in carcinoid tumors, 485, 485f in pancreatic endocrine tumors, 516, 516t, 517f-518f, 519 Somatostatin receptor–directed radiotherapy, for pancreatic endocrine tumors, 522 Somatostatinoma, 511-513 clinical features of, 511t, 512 definition of, 511 diagnosis and differential diagnosis of, 512-513 duodenal, 478-480, 511-513 metastasis of, 511 pathophysiology and pathology of, 511-512 treatment of, 513 Somatostatinoma syndrome, 511-513, 511t Sonic hedgehog protein in intestinal development, 1618 in pancreatic development, 916 Sorafenib, for hepatocellular carcinoma, 1579 Sorbitol, 1707 for constipation, 277t, 278 malabsorption of, irritable bowel syndrome and, 2098 Soy protein intolerance, versus celiac disease, 1811 Space of Disse, 1204, 1207, 1208f, 1490, 1490f Space of Mall, 1204 Spanish toxic oil syndrome, 1453-1454 Spasm esophageal. See Esophageal spasm. glottic, with vomiting, 206 Spasmolysin, in gastrointestinal tract, 16 Spastic nutcracker esophagus, 697, 699f

Spectroscopy, in esophageal cancer, 755 Sphincter anal. See Anal sphincter. artificial bowel, for fecal incontinence, 255-256, 255t esophageal. See Esophageal sphincter(s). pyloric, contraction of, 791 Sphincter ampullae, 1049, 1067, 1068f Sphincter choledochus, 1049, 1067, 1068f Sphincter of Oddi anatomy of, 1048f, 1049, 1067, 1068f dysfunction of, 1067-1072 acute pancreatitis in, 969 biliary, 1068-1069, 1068t after cholecystectomy, 1137 in chronic pancreatitis, 993 classification of, 1068t, 1069 clinical features of, 1068-1069, 1069t definition of, 1067-1068 diagnosis of, 1069-1071 endoscopic retrograde cholangiopancreatography in, 1069-1070, 1196-1197 epidemiology of, 1068 manometry in, 1070-1071, 1070f Milwaukee classification of, 1068-1069, 1068t pancreatic, 1069, 1072 in recurrent pyogenic cholangitis, 1168 sphincterotomy for, 1071-1072 failure of, 1072-1073, 1072t treatment of, 1071-1072 physiology of, 1067 pressures in, 1070-1071, 1070f relaxation of, for post-ERCP pancreatitis, 967 Sphincter pancreaticus, 1049, 1067, 1068f Sphincteroplasty anal, for fecal incontinence, 255-256, 255t balloon, for choledocholithiasis, 1191 Sphincterotomy bleeding after, 660, 1197 endoscopic for bile leaks, 1191-1192, 1193f for choledocholithiasis, 1191, 1192f for sphincter of Oddi dysfunction, 1071-1072 failure of, 1072-1073, 1072t internal for anal fissure, 2265t, 2266 for hemorrhoids, 2261 pancreatic duct, for chronic pancreatitis, 1006 upper gastrointestinal bleeding after, 306 Spigelian hernia, 386f, 390-391 Spinal afferents colonic, 1664 esophageal, 685-686 small intestinal, 1647 splanchnic, 801-802, 801f Spinal cord injury cervical, gastrointestinal problems after, 579-580 chronic intestinal pseudo-obstruction in, 2136-2137 constipation in, 268-269 fecal incontinence in, 256-257 gallstone disease in, 1093 gastrointestinal problems after, 579-580, 579t Spinal nerves, in somatic-parietal pain transmission, 152 Spindle cell gastrointestinal stromal tumors, 463, 463f, 465f, 468 SPINK1 gene in acute pancreatitis, 961-962 in chronic pancreatitis, 987-990 in familial pancreatitis, 952 in pancreatitis, 937 in tropical pancreatitis, 952-953 Spinoreticular tract, in visceral pain transmission, 343 Spinothalamic tract, in visceral pain transmission, 343 Spirochetal infection, of liver, 1354-1355 Spironolactone, for ascites, 1537 Spirulina, for colonic health, 2294t

Splanchnic (spinal) afferent neurons, 801-802, 801f Splanchnic arterial vasodilatation, in hepatorenal syndrome, 1546, 1547f Splanchnic arteriovenous fistula, portal hypertension in, 1501 Splanchnic circulation anatomy of, 2027-2028, 2028f-2029f collateral, 2028, 2029f vasculitis of, 2046-2047 Splanchnic vasodilation, in portal hypertension, 1493-1494 Splanchnicectomy, thoracoscopic, for chronic pancreatitis, in pain control, 1009 Spleen in acute pancreatitis, 982 atrophy of, in celiac disease, 1813 enlarged in hyperreactive malarial splenomegaly, 1358 in tropical splenomegaly syndrome, 1358 Splenectomy, for congenital erythropoietic porphyria, 1269 Splenic artery, 911, 2027, 2028f Splenic pulp pressure, 1496, 1496t Splenic vein, 911 Splenic vein thrombosis in chronic pancreatitis, 1013 ultrasonography in, 1497 Splenorenal shunt, distal, for portal hypertension–related bleeding, 1507, 1507f Spondylitis, ankylosing gastrointestinal manifestations of, 563 in ulcerative colitis, 2011 Spondyloarthropathies, seronegative, gastrointestinal manifestations of, 558t, 563 Spontaneous bacterial peritonitis. See Peritonitis, bacterial, spontaneous. Sport drinks, for diarrhea, 226 Sprue celiac. See Celiac disease. collagenous, 1811, 1817 tropical. See Tropical sprue. Squamous cell carcinoma achalasia and, 703, 747-748 esophageal. See Esophageal cancer, squamous cell. in Zenker’s diverticula, 371 Squamous papilloma, esophageal, 768, 768f Staphylococcus aureus infection antibiotic-associated diarrhea in, 1889 in food poisoning, 1879-1881, 1879t-1880t toxic shock syndrome from, hepatic manifestations of, 1351 Stapled transanal rectal resection (STARR), for defecatory disorders, 283 Starbucks diarrhea, 217-218 Starch. See Carbohydrate. Starch peritonitis, 617-618 Starvation, 59-60 body compartment losses in, 61t long-term, 60 short-term, 59-60 terminal phase of, 60 Statins gallstone disease and, 1092 hepatotoxicity of, 1433 Stauffer’s syndrome, jaundice in, 329 Stavudine, hepatotoxicity of, 1430 Steakhouse syndrome, 175, 401 Steatohepatitis, nonalcoholic, 1401. See also Fatty liver disease, nonalcoholic. bacterial overgrowth and, 1775 drug-induced, 1441-1444 histopathology of, 1405-1407, 1406f Steatorrhea. See also Fat(s), malabsorption of. in chronic pancreatitis, 996-997, 1009-1010 diarrhea and, 222, 225 in ileal bile acid malabsorption, 1755-1756 in malabsorption, 1745t in neonatal cholestasis, 1062 in primary biliary cirrhosis, 1487 in small intestinal bacterial overgrowth, 1774 in somatostatinoma, 512

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxxi

lxxxii

Index Steatosis, hepatic, 1221. See also Fatty liver disease. Stellate cells hepatic, 1203-1204, 1211 in alcoholic fatty liver disease, 1389 in hepatic fibrosis, 1211 in nonalcoholic fatty liver disease, 1404 pancreatic, fibrosis associated with, 987 Stem cells intestinal, 1618 transplantation of. See Hematopoietic stem cell transplantation. Stenosis ampullary, 1067-1068 anal, 1635, 2272 constipation in, 269 duodenal. See Duodenum, atresia and stenosis of. esophageal, congenital, 670t, 671, 671f papillary, in HIV/AIDS, 534, 534f pyloric, hypertrophic, adult, 785 Stent(s) biliary. See Biliary stent. for chronic pancreatic pain, 1006, 1007f for colonic obstruction, 2117-2120 for colorectal cancer, 2238 endoscopic, complications of, 658 esophageal in cancer, 765-767, 766f types of, 766 for gastric cancer, 904 for intrahepatic cholangiocarcinoma, 1581-1582 for pancreatic fistula, 1014 for post-ERCP pancreatitis, 968 self-expanding metallic covered, 1190 for distal bile duct strictures, 1193-1194, 1196f for transjugular intrahepatic portosystemic shunt, 1504 Stercoral ulcer, colonic, 2052 Steroid(s). See Anabolic steroids; Glucocorticoids. Stethoscope sign, in functional abdominal pain syndrome, 168-169 Stevens-Johnson syndrome, oral manifestations of, 358 Stierlin’s sign, in intestinal tuberculosis, 1878 Stiff-man syndrome, gastrointestinal manifestations of, 581 Still’s disease adult, gastrointestinal and hepatic manifestations of, 559-560 Reye’s syndrome in, 1430-1431 Stomach. See also Gastric. accommodation response of functional dyspepsia and, 187-188 in gastric motility, 794-796, 796f-797f adenocarcinoma of. See Gastric cancer. adenoma of in familial adenomatous polyposis, 2179 malignant transformation of, 896 anatomy of, 773-775, 817-821, 819f general, 773-775, 774f-775f microscopic, 776-778, 776f-778f antrum of, 775, 775f. See also Antrum. arterial supply to, 775 atresia of, 780-781, 780t bicarbonate secretion in, 830-831 bile reflux into, 857-858 biopsy of. See Gastric biopsy. body of, 775, 775f carcinoid tumors of, 478-479, 479t, 486-488, 487f-488f, 904, 905f autoimmune metaplastic atrophic gastritis and, 848 cardia of, 775-776, 775f, 817 in swallowing, 682-683 caustic injury to. See Caustic injury. congenital anomalies of, 780-785, 780t contractions of, assessment of, 795f, 804

Stomach (Continued) corpus of, electrophysiologic characteristics of, 790-791, 792f development of, 774f dilation of, acute, in eating disorders, 137 distention of hypersensitivity to, in functional dyspepsia, 188 in lower esophageal sphincter relaxation, 685 diverticula of, 375-376, 375f, 780t, 781-782 intramural, 375, 375f juxtacardiac, 375, 375f duplication of, 780t, 782-783, 782f electrophysiology of, 789-798 emptying of. See Gastric emptying. folds of in gastrinoma, 499, 502 hyperplastic, 858-859, 859f fundus of, 775, 775f electrophysiologic characteristics of, 790-791, 792f relaxation and accommodation of, 794-796, 796f, 805 gangrene of, 849-850 glandular anatomy of, 817-819, 819f-820f. See also Oxyntic (fundic) gland; Pyloric gland. hourglass, in syphilis, 850, 851f incisura angularis of, 775, 775f infectious diseases of, 849-852 inflammation of. See Gastritis. injury to, with vomiting, 206 innervation of, 776, 794, 819-821, 820f interstitial cells of Cajal in, 791-792, 793f intrinsic factor secretion in, 829-830 intrinsic primary afferent neurons in, 801, 801f ischemia of, 858 lamina propria of, 778 lymphatic drainage of, 775 lymphoma of. See Lymphoma, gastric. metastasis to, 904-905 motility of. See Gastric motility. mucosa of, 776 heterotopic, in esophagus, 670t, 675, 675f peptic ulcer defenses of, 863 prolapse of, 858 mucus secretion in, 831-832 muscularis propria of, 776 neurotransmitters in, 819-821, 820f outlet obstruction of. See Gastric outlet obstruction. paralysis of. See Gastroparesis. pepsinogen secretion in, 829 polyps of. See Gastric polyps. pylorus of. See Pylorus. radiation injury to, 641-642, 857 regions of, 775, 775f removal of. See Gastrectomy. sensory function of, 801-802, 801f serosa of, 776 smooth muscle cells in, 790-791, 792f-793f splanchnic (spinal) afferent neurons in, 801-802, 801f stromal tumors of, 465, 465f, 904, 905f submucosa of, 776 teratoma of, 780t, 783 tissue layers of, 776 tumors of, 887-906. See also Gastric cancer; Lymphoma, gastric; Stomach, carcinoid tumors of. ulcers of. See Gastric ulcers. vagal afferent neurons in, 801-802, 801f, 819-821, 820f varices of. See Gastric varices. venous drainage of, 775 volvulus of. See Gastric volvulus. watermelon. See Gastric antral vascular ectasia (watermelon stomach). Stomal obstruction, in Brooke ileostomy, 2017 Stomal placement, hernia following, 389 Stomal varices, 1513-1514

Stones cholesterol, in children, 1064 common bile duct. See Choledocholithiasis. gallbladder. See Gallstone disease. hepatic duct, 1167-1170, 1168f kidney calcium oxalate, in short bowel syndrome, 1790, 1790f in Crohn’s disease, 1955 pancreatic duct, extracorporeal shock wave lithotripsy for, 1006-1007 Stool. See also Defecation; Fecal. acid steatocrit test of, 1751 antigen test of, for Helicobacter pylori infection, 840t, 841 Ascaris lumbricoides in, 1922, 1923f bile acids in, 1753 blood in, 218-219, 222. See also Fecal occult blood test (FOBT). Bristol Scale for, 264, 265f, 2092, 2093f Capillaria (Paracapillaria) philippinensis in, 1925 characteristics of in diarrhea, 211, 218-219 in fecal incontinence, 245 choleric, electrolyte composition of, 1851, 1851t chymotrypsin in, 928t, 929 in chronic pancreatitis, 1000 in diarrhea, 226 consistency of, 264, 265f culture of in acute diarrhea, 220 in amebiasis, 1909 in Campylobacter spp. infection, 1868 in chronic watery diarrhea, 222-223 in Clostridium difficile-associated diarrhea and colitis, 1896t, 1897 in enteric infection, 1849 in enterohemorrhagic Escherichia coli, 1856 in shigellosis, 1860 in typhoid fever, 1866 DNA analysis of in colorectal adenoma, 2167 in colorectal cancer, 2226 elastase in in chronic pancreatitis, 1000 in diarrhea, 226 evacuation of. See Defecation. evaluation of in acute diarrhea, 220 in amebiasis, 1909 in celiac disease, 1810 in chronic diarrhea, 220-226, 222f in Clostridium difficile-associated diarrhea and colitis, 1896-1897, 1896t in enteric infection, 1849 in hookworm infection, 1926, 1927f in laxative abuse, 222 in shigellosis, 1860 timed, 220 fat in. See Fat(s), fecal. floating, 218-219, 236 in gastrointestinal bleeding, 285, 288 immunochemical testing of in colorectal adenoma, 2165-2166 in colorectal cancer, 2221t, 2223 impaction of in children, 258 diarrhea and, 230 polyethylene glycol for, 278 maroon-colored, 288, 291 mutagens in, 2195 osmotic gap of, 221-222, 222f output of, after proctocolectomy, 2016 pH of, in diarrhea, 222 retention of, functional, 258, 266 seepage of, 246, 257 size of, 264, 265f, 2093f Strongyloides stercoralis in, 1924 Trichuris trichiura in, 1923f, 1928 watery, 218-219 weight of, 262

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Stool (Continued) white blood cells in assay for, 220 in enteric infection, 1848-1849, 1848t Stool softeners and emollients, for constipation, 277t, 279 Stoppa procedure, 387 Stores regulator, in iron absorption, 1725 Strawberry gallbladder, 1145 Streptococcal infection, group A, toxic shock syndrome from, hepatic manifestations of, 1351 Streptococcus bovis bacteremia, colorectal adenoma in, 2165 Streptokinase, for superior mesenteric artery embolus, 2034-2035 Streptozotocin, for metastatic pancreatic endocrine tumors, 520-521 Stress in biopsychosocial model, 341 brain-gut interactions and, 342-343 colonic motility and, 1671 Crohn’s disease and, 346, 1946 gastrointestinal function and, 342-343 gastropathy and, 581, 857 immune response and, 345-346 metabolic body compartment losses in, 61t and energy requirements for hospitalized patients, 48-49, 49t-50t oxidative in alcoholic liver disease, 1385 in Helicobacter pylori infection, 836 in nonalcoholic fatty liver disease, 1403 peptic ulcer disease and, 863 ulcerative colitis and, 346 Stress mediators, neurohormonal, in postoperative ileus, 2125 Stress ulcers, 863. See also Peptic ulcer disease. bleeding, 303-304, 877-878 Stretch receptors, visceral, 152 Strictures anal, in Crohn’s disease, 1950 bile duct. See Biliary strictures. colonic. See Colonic strictures. in Crohn’s disease, 1951-1952, 1952f esophageal. See Esophageal strictures. gastric after caustic injury, 408, 408f in gastroesophageal reflux disease, 720, 720f pancreatic duct, in chronic pancreatitis, 995 in ulcerative colitis, 2005 Stridor, during endoscopy, 657 String sign in Crohn’s disease, 1952, 1952f in gastric outlet obstruction, 428-430, 428f Stroke dysphagia after, 687-688 percutaneous endoscopic gastrostomy for, 90 gastrointestinal manifestations of, 579 Strongyloides stercoralis, 1924-1925 Strongyloidiasis autoinfection in, 1924 cutaneous, 1924 fulminant, 1924 gastritis in, 852 hepatic, 1356t-1357t, 1360-1361 intestinal, 1924-1925 malabsorption in, 1829, 1829t STW 5 (Iberogast) for functional dyspepsia, 2290t, 2291 for irritable bowel syndrome, 2291, 2292t Subclavian artery, right, aberrant, dysphagia lusoria in, 672, 673f Subclavian vein, catheterization of complications of, 87 for parenteral nutrition, 87 Subjective global assessment, in nutritional assessment, 69, 69t, 78 Submucosa of esophagus, 666f, 667 of intestines, 1619

Submucosa (Continued) of small intestine, 1619 of stomach, 776 Submucosal plexus, 665-666, 668f, 681, 819-821, 820f, 1623, 1623f, 1646, 1662, 1663f Submucosal resection, endoscopic for esophageal cancer, 761-762 for gastric cancer, 903, 903f Subphrenic abscess, 416, 416f Subserosal plexus, 1623 Substance abuse colonic ischemia in, 2040 hepatotoxicity related to, 1454 liver transplantation and, 1596 Substance P, 9 in intestinal ion transport, 1693f receptors for, 9 Sucralfate for NSAID ulcers, 873 for peptic ulcer disease, 871 Sucrase, deficiency of, congenital, 1641 Sucrase-isomaltase, 1708, 1708f, 1708t biosynthesis of, 1708-1709, 1709f congenital deficiency of, 1709 deficiency of, malabsorption in, 1759t-1762t, 1763 regulation of, 1709-1710 Sugar alcohols for constipation, 277t, 278 diarrhea from, 213, 224 Sugar malabsorption, irritable bowel syndrome and, 2098 Sulfasalazine for Crohn’s disease, 1960-1961, 1962t hepatotoxicity of, 1432 malabsorption with, 1757t for NSAID enteropathy, 2054 side effects of, 1994, 1994t for ulcerative colitis, 1993-1994, 1993f, 1994t Sulfate diarrhea from, 212-213 intestinal ion transport and, 1689 Sulfonamides hepatotoxicity of, 1432 malabsorption with, 1757t Sulfur, intestinal gas containing, 236, 238 Sulindac in colorectal cancer, 2198 in familial adenomatous polyposis, 2183-2184 Sumatriptan, colonic ischemia from, 2040 “Sunflower cataract,” in Wilson disease, 1252 Sunitinib for gastrointestinal stromal tumors, 471-472 for metastatic pancreatic endocrine tumors, 522 Superior mesenteric artery, 1617 anastomotic circulation of, 2028, 2029f anatomy of, 2027, 2028f embolus of, 2030, 2034-2035, 2034f thrombosis of, 2030, 2035 Superior mesenteric artery syndrome, 608, 608f in anorexia nervosa, 131 vomiting in, 203, 203f Support group, for short bowel syndrome, 1786 Suppositories aminosalicylate, for ulcerative colitis, 1995 for constipation, 277t, 279-280 glycerin, colonic ischemia from, 2040 Supralevator abscess, 2266, 2267f Supravesicular hernia, 392-395 Swallow therapy, for dysphagia, 700 Swallowing, 677-686 bolus transport in, 680 deglutitive inhibition and, 681 difficult. See Dysphagia. esophageal peristalsis and, 681-682, 683f lower esophageal sphincter in, 682-685, 683f oral phase of, 679 oropharynx and upper esophageal sphincter in, 677-679, 678f-679f painful. See Odynophagia. pharyngeal, 679-680, 680f reflux episodes during, 709, 709f

Sweat testing in cystic fibrosis, 940 indications for, 944t in meconium ileus, 944 Sweet’s syndrome, in ulcerative colitis, 2011 Swimmer’s itch, 1936 Symblepharon, ocular, 357 Symmers’ pipe stem fibrosis, in schistosomiasis, 1936-1937 Sympathetic nervous system in colonic motility, 1662f, 1664 in postoperative ileus, 2125 in small intestinal motility, 1647-1648 Sympathomimetic amines, for obesity, 110t, 111 Symporter, 1680-1681 Synbiotics, for ulcerative colitis, 1999-2000 Syndecans, in protein-losing gastroenteropathy, 437-438, 439f Syo-saiko-to, hepatotoxicity of, 1458 Syphilis esophagitis in, 743 gastritis in, 850-851, 851f hepatic manifestations of, 1354 hepatitis in, 1354 Systemic inflammatory response in acute pancreatitis, 963 definition of, 973 energy expenditure and, 60 protein-energy malnutrition and, 60 Systemic lupus erythematosus. See Lupus erythematosus. Systemic sclerosis, progressive. See Scleroderma.

T

T cell(s) CD4+ count of, in AIDS definition, 523 in Helicobacter pylori infection, 891 in oral tolerance, 23 CD8+, in oral tolerance, 23 in celiac disease, 1800, 1803, 2054, 2056 differentiation of, 28-29 in food allergy, IgE-mediated, 141 in Helicobacter pylori infection, 891 intraepithelial, 28 migration of, 25, 26f in mucosa-associated lymphoid tissue, 445, 446f regulatory in food allergy, 140 in oral tolerance, 23 in ulcerative colitis, 1979-1980 T cell lymphoma gastric, 453 hepatosplenic, with anti-TNF agents, 1969 small intestinal enteropathy-type. See Enteropathy-associated T cell lymphoma. natural killer type, 459 T pouch, 2018 T-tube removal, after liver transplantation, 1609 Tachykinin(s), 9 Tacrine, hepatotoxicity of, 1436 Tacrolimus for Crohn’s disease, 1967 gastrointestinal side effects of, 541 hepatotoxicity of, 543 for liver transplantation, 1605-1606, 1606t nephrotoxicity of, 1609-1610 for primary sclerosing cholangitis, 1165 for ulcerative colitis, 1999 Taenia saginata/solium infection, 1931-1932 Takayasu’s disease, vasculitis in, 2047 Tamoxifen, steatohepatitis from, 1442-1443 Tangier disease, 577, 1221 Tap water enema, for constipation, 277t, 280 Tapeworms. See Cestode (tapeworm) infection. Taste disorders, 354 Taste receptors, 17 14 C-Taurocholate bile acid test, for bile acid malabsorption, 1753

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxxiii

lxxxiv

Index TD-5108, for constipation, 281 Technetium pertechnetate scintigraphy, in gastrointestinal bleeding, 292 Technetium sulfur colloid tracer, in gastrointestinal bleeding, 292 Teeth. See Tooth (teeth). Tegaserod for abdominal distention and bloating, 239 for constipation, 280 for functional dyspepsia, 193 for gastric motility disorders, 812, 813t for vomiting, 208 Teicoplanin, for Clostridium difficile-associated diarrhea and colitis, 1899 Telangiectasia definition of, 593 estrogen and, 316-317 hereditary hemorrhagic, 600-601 cutaneous manifestations of, 361, 361f, 600, 601f gastrointestinal bleeding in, 317, 600 portal hypertension in, 1501 Telaprevir, for hepatitis C, 1334 Telbivudine for hepatitis B, 1303 in pregnancy, 1306 resistance to, 1303 Telomerase, in esophageal cancer, 750 Tenascin, in collagenous colitis, 2240 Tenofovir disoproxil fumarate for hepatitis B, 1303-1304 in pregnancy, 1306 Tense ascites, 1535, 1539 Teratoma, gastric, 780t, 783 Terbinafine, hepatotoxicity of, 1435 Terlipressin for ascites, 1540 for hepatorenal syndrome, 1548, 1548t for portal hypertension, 1502 Terminal illness, palliative care in. See Palliative care. Testicular torsion, vomiting in, 200 Tetracycline for cholera, 1852 for Dientamoeba fragilis infection, 1914 esophagitis from, 736f, 737 for Helicobacter pylori infection, 842, 842t hepatotoxicity of, 1414-1415, 1429, 1433 malabsorption with, 1757t for tropical sprue, 1827-1828 for Whipple’s disease, 1841, 1841t Thalassemia, hepatic manifestations of, 572 Thalidomide for aphthous ulcers, 356 for colonic angioectasia, 599 for Crohn’s disease, 1970 for esophageal ulceration in HIV/AIDS, 526 in pregnancy, 630 Theophylline, for esophageal hypersensitity, 704 Thermal therapy for Barrett’s esophagus, 731 for colonic angioectasia, 599, 600f for gastric antral vascular ectasia, 602 for gastrointestinal bleeding, 291-292 for peptic ulcer bleeding, 297-298, 298f, 880 Thermic effect of feeding, 48 of physical activity, 48, 49t Thiamine (vitamin B1), 53t-55t, 81, 81t absorption of, 1718t, 1720 deficiency of after bariatric surgery, 118 in refeeding syndrome, 70 dietary sources of, 1720 reference nutrient intake for, 1718t Thiazides, malabsorption with, 1757t Thiazolidinediones hepatotoxicity of, 1435 for nonalcoholic fatty liver disease, 1410 Thiopental, for intracranial hypertension, 1565-1566

Thiopurine agents, for Crohn’s disease, 1962t, 1964-1966, 1964f Thoracic nerve radiculopathy, abdominal pain in, 165 Thorium dioxide, hepatotoxicity of, 1453 Thrombin injection therapy, for peptic ulcer bleeding, 880 Thromboangiitis obliterans, vasculitis in, 2046 Thrombocytopenia in cavernous hemangioma of liver, 1586 in cirrhosis, 1555 Thromboembolism in ulcerative colitis, 2012 venous gastrointestinal manifestations of, 570 in glucagonoma, 507 Thrombolytic therapy for Budd-Chiari syndrome, 1374 for superior mesenteric artery embolus, 2034-2035 Thrombosis catheter-related, 87 in Crohn’s disease, 1955 deep venous, in ulcerative colitis, 2012 of hemorrhoids, 2262, 2264f hepatic artery after liver transplantation, 1606-1607 after solid organ transplantation, 540 hepatic vein. See Budd-Chiari syndrome. mesenteric vein. See Mesenteric venous thrombosis. portal vein. See Portal vein thrombosis. splenic vein in chronic pancreatitis, 1013 ultrasonography in, 1497 superior mesenteric artery, 2030, 2035 Thrombotic thrombocytopenic purpura, gastrointestinal manifestations of, 570 Thrush, 354-355, 354f Thumbprinting, in colonic ischemia, 2041, 2041f-2042f Thymic extract, for liver disease, 2295t, 2297 Thyroid carcinoma, medullary, gastrointestinal manifestations of, 573t, 576 Thyroid disease. See also Hyperthyroidism; Hypothyroidism. autoimmune in celiac disease, 1812 malabsorption in, 1765-1766 in cirrhosis, 1554 Thyropharyngeus, 677-678, 678f Thyrotoxicosis, apathetic, 575 Ticlopidine, hepatotoxicity of, 1433-1434 Tight junctions in acinar cell, 913, 921 in intestinal barrier, 25, 1676f, 1678-1679 Tinidazole for amebiasis, 1910, 1910t for giardiasis, 1914 for Helicobacter pylori infection, 842, 842t TIPS. See Transjugular intrahepatic portosystemic shunt (TIPS). Tissue culture cytotoxicity assay, in Clostridium difficile-associated diarrhea and colitis, 1896, 1896t Tissue depletion, in nutritional assessment, 65 Tissue plasminogen activator, for sinusoidal obstruction syndrome, 1377 Tissue transglutaminase, in celiac disease, 1802-1803, 1803f Tizanidine, hepatotoxicity of, 1436-1437 TJ-9, for liver disease, 2295t, 2296 TLR5, recognition of pathogen-associated molecular patterns by, 26 TNM staging system for esophageal cancer, 756, 757t for gastric cancer, 900-901, 900f, 901t for gastrointestinal lymphoma, 447, 447t Tobacco use. See Smoking. Tolazamide, hepatotoxicity of, 1435-1436 Tolbutamide, hepatotoxicity of, 1435-1436 Tolcapone, hepatotoxicity of, 1436

Tolevamer, for Clostridium difficile-associated diarrhea and colitis, 1901 Toll-like receptors in Helicobacter pylori infection, 835, 890 intestinal expression of, 26 Toluene, hepatotoxicity of, 1452 Tongue black hairy, 354 geographic, 354 hairy leukoplakia involving, 355, 355f inflammation of, 353, 354f pain in, 353-354 Tooth (teeth) abnormalities of, in familial adenomatous polyposis, 2180-2182 erosions of from chronic vomiting, 127f in gastroesophageal reflux disease, 715 familial agenesis of, 2177t, 2184 Topiramate for bulimia nervosa, 134 hepatotoxicity of, 1432 for obesity, 112 Toremifene, hepatotoxicity of, 1443 Torovirus infection, 1870t, 1872 Toxic epidermal necrolysis, oral manifestations of, 358 Toxic megacolon necrotizing colitis with, in amebic dysentery, 1909 in ulcerative colitis, 1992, 2005 Toxic-metabolic hypothesis, of chronic pancreatitis, 987 Toxic shock syndrome, hepatic manifestations of, 1351 Toxin(s). See also Cytotoxins; Enterotoxins. acute pancreatitis from, 965 botulinum, 1883 cholera, 214, 1845, 1845f, 1850 Clostridium difficile, 1893-1894, 1893f diarrhea from, 217-218, 217t gastropathy from, 856 nonalcoholic fatty liver disease from, 1402 in primary sclerosing cholangitis, 1157 Shiga, 1859 Shiga-like, in enterohemorrhagic Escherichia coli, 1856 vomiting from, 199t, 200 Toxocariasis, hepatic manifestations of, 1356t-1357t, 1359-1360 Toxoplasmosis, hepatic manifestations of, 1356t-1357t, 1359 TP53 gene in apoptosis, 32, 33f in colorectal cancer, 2200t, 2202 in esophageal cancer, 750 in gallbladder carcinoma, 1178-1179 in gastric cancer, 893-894 in pancreatic cancer, 1020 TPMT gene polymorphism, azathioprine metabolism and, 1996-1997 Trabecular hepatocellular carcinoma, 1574 Tracheobronchial remnants, esophageal stenosis from, 671 Tracheoesophageal fistula, 668-670, 669f, 669t-670t distal, 669f, 670, 670t H-type, 669f, 670, 670t Tramadol in chronic pancreatitis, 1004 for neuropathic pain, 2283t Transcobalamin II, deficiency of, malabsorption in, 1759t-1762t, 1763 Transcription factors in cellular proliferation, 33 growth factors and, 16 Transcytosis, 1210 Transferrin, serum, in nutritional assessment, 67-68, 67t, 78 Transferrin receptor, HFE binding to, 1241-1242 Transferrin saturation index in hereditary hemochromatosis, 1244, 1244t in iron deficiency anemia, 321

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Transforming growth factor-α in esophageal cancer, 749-750 in gastrointestinal tract, 16 Transforming growth factor-α in cellular proliferation, 33 in Crohn’s disease, 1949 in esophageal cancer, 749-750 in gastrointestinal tract, 16 in oral tolerance, 22-23 in radiation injury, 639-640 receptors for, 15f, 16 serine/threonine kinase region on, 15 Transglutaminase, epidermal, in dermatitis herpetiformis, 1812 Transient elastography (FibroScan) in hepatic fibrosis, 1235 in hepatitis C, 1323, 1324t Transitional zone, anal canal, 2257, 2258f Transjugular intrahepatic portosystemic shunt (TIPS), 1504-1506 for ascites, 1540 for Budd-Chiari syndrome, 1374-1375 complications of, 1504, 1505t for ectopic varices, 1514 follow-up for, 1505-1506 for gastric antral vascular ectasia, 602 for gastric variceal bleeding, 1512-1513 for hepatopulmonary syndrome, 1552 for hepatorenal syndrome, 1549 patient selection for, 1506, 1506t for portal hypertension, 1504-1506, 1504f, 1505t-1506t for portal hypertension–related bleeding, 1505 for portal hypertensive gastropathy, 602 procedure for, 1504, 1504f for sinusoidal obstruction syndrome, 1377 stents for, 1504 for variceal bleeding, 308 Translumbar magnetic stimulation, anal motor evoked potentials after, 249, 250f Transmitters. See also Neurotransmitters. autocrine, 3, 4f gastrointestinal, 3-5, 4f, 4t as hormones, criteria for, 4-5 paracrine, 3-4, 4f Transplantation heart complications of, 540-541 gallstone disease and, 1106 hepatitis C after, 1327 hepatitis C after, 1326 intestinal for chronic intestinal pseudo-obstruction, 2143 complications of, 541 for short bowel syndrome, 1791-1793, 1792t kidney, 539-540 kidney-pancreas, 539-540 liver. See Liver transplantation. lung complications of, 540-541 lymphoproliferative disorders after, 459 small intestinal, for complications with parenteral nutrition, 86-87 solid organ abdominal pain after, 538t, 542-543 anorexia after, 538t, 541 biliary tract disease after, 540, 544 complications of diagnosis of, 541-544 overview of, 537-544, 538t, 539f constipation after, 541-542 cytomegalovirus after, 537-542, 539f diarrhea after, 538t, 541-542 eosinophilia after, 433 esophageal symptoms after, 541 fungal infections after, 539 gastric ulcers after, 541 gastroesophageal reflux disease after, 540-541 gastrointestinal bleeding after, 538t, 540, 543 gastrointestinal cancer after, 538t, 543 gastroparesis after, 540-541

Transplantation (Continued) graft-versus-host disease after, 541 hepatic artery thrombosis after, 540 hepatitis C after, 539, 1326 hepatobiliary complications of, 538t, 543-544 herpes simplex virus after, 539, 539f lymphoproliferative disorders after, 459, 539, 539f, 543 nausea and vomiting after, 538t, 541 pneumatosis intestinalis after, 543 stem cell. See Hematopoietic stem cell transplantation. Transport, fluid and electrolyte. See Fluid and electrolyte transport. Transporter proteins. See Ion transporters. Transthoracic echocardiography, in carcinoid heart disease, 483 Trauma diaphragmatic hernia with, 379-383 esophageal, 739-740 foreign body. See Foreign bodies. head, 579 pancreatic, 938, 966-967 spinal cord. See Spinal cord injury. Traveler’s diarrhea, 1872-1875 chronic, 1875, 1875t clinical features of, 1873-1874, 1874t epidemiology of, 1873 intestinal ion transport in, 1691 microbiology of, 1870t, 1873 prevention of, 1874-1875, 1875t treatment of, 1874 tropical, 1821-1822 Trefoil factors in gastric cancer, 894 gastric secretion of, 831 in gastrointestinal tract, 16 Trehalase, 1708-1709, 1708f, 1708t deficiency of, malabsorption in, 1759t-1762t Trematodes. See Flukes. Triamterene, malabsorption with, 1757t Triangle of Calot, 1125 Trichinella spp., 1929-1930, 1929f Trichinosis hepatic, 1361 intestinal, 1929-1930, 1929f Trichloroacetic acid, for anal warts, 2270, 2271t Trichobezoar, 405 Trichobilharzia ocellata infection, 1936 Trichuris muris infection, 1928 Trichuris suis biotherapy for Crohn’s disease, 1970 for ulcerative colitis, 2000 Trichuris trichiura infection, 1923f, 1927-1928, 1927f Triclabendazole, for fascioliasis, 1363, 1935 Tricyclic antidepressants. See Antidepressants, tricyclic. Trientine in pregnancy, 637 for Wilson disease, 1256, 1256t Trigger points, in myofascial pain syndromes, 164 Triglyceride(s), 1221 in ascitic fluid, 1521, 1526 digestion of, 1699-1702 brush border membrane transfer in, 1702-1703, 1703f emulsification in, 1699, 1700f intracellular processing in, 1703-1705, 1704f-1705f intraluminal pH in, 1699t, 1701 lipase in, 1699-1701, 1699t, 1700f micelles and lipid-containing particles in, 1700f, 1701 in neonates and infants, 1729-1732 water layer in, 1701-1702 elevated. See Hypertriglyceridemia. as energy store, 47, 48t formation of, 1220 gallstone disease and, 1092

Triglyceride(s) (Continued) medium-chain, 79 for cholestasis in children, 1062 for malabsorption, 73, 1767 for short bowel syndrome, 1783-1785 resynthesis of, 1703, 1705f structure of, 1698, 1698f Triglyceride lipase functions of, 924 pancreatic, 1699-1700 Trimethoprim-sulfamethoxazole for Cyclospora cayetanensis infection, 1916 for Isospora belli infection, 1917 prophylactic for spontaneous bacterial peritonitis, 1534, 1535t for traveler’s diarrhea, 1875t for traveler’s diarrhea, 1874 for Whipple’s disease, 1841, 1841t Trinitrotoluene, hepatotoxicity of, 1452 Tripeptides, 1715-1716 Tripeptides, in protein absorption, 1715f, 1715t Trisomy, duodenal atresia/stenosis and, 785 Trituration, of solid food, 1696 Trocar site hernia, 389, 2106 Troglitazone hepatotoxicity of, 1435 for nonalcoholic fatty liver disease, 1410 Tropheryma whippelii, 1834-1835, 1834f, 1841. See also Whipple’s disease. Tropical diarrhea, enteropathogens causing, 1821-1822, 1822t Tropical enteropathy, 1828 Tropical malabsorption. See also Tropical sprue. approach to patient with, 1831 in celiac disease, 1829t, 1830 in common variable immunodeficiency, 1829t, 1830 in Crohn’s disease, 1830 etiology of, 1822t helminthic infections causing, 1829-1830, 1829t HIV/AIDS and, 1830 in immunoproliferative small intestinal disease, 1829t, 1830-1831 in intestinal tuberculosis, 1829t, 1830 lymphoma and, 1830-1831 pancreatitis and, 1831 protozoal infections causing, 1828-1829, 1829t subclinical, 1823f, 1828 treatment of, 1829t Tropical pancreatitis, 952-953, 989 calcific, 931, 952 chronic, 989 definition of, 931 malabsorption and, 1831 subtypes of, 931 Tropical splenomegaly syndrome, 1358 Tropical sprue, 1822-1828 clinical manifestations of, 1824-1825, 1824t, 1825f definition of, 1822 diagnosis of, 1826-1827, 1827f, 1831 endemic, 1823 epidemic, 1823 epidemiology of, 1823, 1823f etiology of, 1823-1824 fat malabsorption in, 1825-1827 histopathology of, 1825, 1826f historical aspects of, 1822-1823 pathophysiology of, 1825-1826 treatment of, 1827-1828 versus tropical enteropathy, 1828 in visitors, 1823-1824 vitamin B12 deficiency in, 1824, 1825f, 1826-1827 Trousseau’s syndrome, 570 Truelove and Witts classification of ulcerative colitis, 1991, 1991t Trypanosomiasis, American. See Chagas’ disease. Trypsin, 1712-1713, 1713f, 1713t functions of, 924 in pancreatitis, 932, 934f, 961-962 regulation of, 927-928, 933, 934f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxxv

lxxxvi

Index Trypsinogen, serum, in chronic pancreatitis, 1000 Trypsinogen activation peptide, in acute pancreatitis, 975 Trypsinogen gene anionic (PRSS2), in pancreatitis, 934 cationic (PRSS1) mutations of, testing for, 951-952 in pancreatitis, 933-934, 934f, 989-990 TT virus infection, 1346 TTV-like mini-virus infection, 1346-1347 Tube feeding. See Enteral nutrition. Tuberculosis, 1877 esophagitis in, 743 gastritis in, 850 hepatic granulomas in, 1355 in HIV/AIDS diarrhea in, 527t, 530 hepatic, 533 intestinal, 1878 intestinal, 1877-1878, 1877f-1878f classification and distribution of, 1877 clinical features of, 1877 diagnosis of, 1877-1878, 1878f malabsorption in, 1829t, 1830 pathogenesis of, 1877 pathology of, 1877, 1877f treatment of, 1878 jaundice in, 327-328 peritonitis in, 616-617 ascites in, 1517-1518, 1526-1527, 1536 smear and culture for, in ascites, 1526 Tuberous sclerosis hamartomatous polyps in, 2186 pancreatic endocrine tumors in, 495 Tubo-ovarian abscess, versus appendicitis, 2063t Tufting enteropathy, 1640-1641, 1759t-1762t Tumor. See Cancer; Neoplasia; specific tumors and anatomic sites. Tumor-like lesions, of liver, 1587-1589 Tumor necrosis factor in alcoholic liver disease, 1387-1388, 1388f in Crohn’s disease, 1946 in gastrointestinal tract, 13 Tumor necrosis factor-α in apoptosis, 1420 in gastric cancer, 893 in hepatopulmonary syndrome, 1549 in nonalcoholic fatty liver disease, 1403 in protein-energy malnutrition, 61, 62t Tumor necrosis factor-α antagonists lymphoma with, 1968-1969 in pregnancy, 1969 Tumor suppressor gene(s), 37-40 in adenoma-carcinoma hypothesis, 2160 APC, 38-40 in gastrointestinal tumors, 36t-37t, 38f-39f inactivation of, 38, 39f Knudson’s hypothesis of, 38, 38f loss of heterozygosity or allelic deletion of, 38, 39f in pancreatic cancer, 1020 SMAD4, 40 TP53, 40 Turcot’s syndrome, 2177t, 2184 colorectal cancer in, 2208 Turmeric for functional dyspepsia, 2290t, 2291 for inflammatory bowel disease, 2293 Two-needle stick technique, for direct percutaneous jejunostomy, 92-93, 92f Tylosis, esophageal squamous cell carcinoma and, 747-748 Typhlitis, 2252-2253, 2253f after hematopoietic stem cell transplantation, 544 in leukemia, 567 Typhoid fever, 1864-1867 clinical features of, 1865-1866 diagnosis of, 1866 epidemiology of, 1864-1865 hepatic manifestations of, 1352 microbiology of, 1864 pathogenic mechanisms in, 1864

Typhoid fever (Continued) treatment of, 1865t, 1866-1867 vaccines for, 1867 Tyrosine, metabolism of, 1269, 1270f Tyrosine kinase nonreceptor, 15 oncogenes and, 36 receptor, 14 in cellular proliferation, 33 oncogenes and, 36 Tyrosine kinase inhibitor multiple, for gastrointestinal stromal tumors, 471-472 targeting KIT, 469-471, 470t Tyrosine phosphatase, receptor, 15 Tyrosinemia, 1269-1271 clinical features of, 1270 diagnosis of, 1270-1271 pathophysiology of, 1269, 1270f treatment of, 1271

U

Ubiquitin-ligase E3, in Johanson-Blizzard syndrome, 954 Ubiquitin/proteasome pathway, 1216 UDCA. See Ursodeoxycholic acid. Ulcer(s) in amebic colitis, 1908f, 1909 aphthous in Crohn’s disease, 359, 359f, 1948, 1954 oral, 356, 356f, 356t Cameron, in sliding hiatal hernia, 381, 381f, 383 colonic. See Colonic ulcers. duodenal. See Duodenal ulcers. esophageal. See Esophageal ulcers. gastric. See Gastric ulcers. in gastroesophageal reflux disease, 720 in intestinal tuberculosis, 1877-1878 marginal, after bariatric surgery, 118 mucocutaneous, 356-357, 356f, 356t oral aphthous, 356, 356f, 356t in herpes simplex, 356-357, 356t in herpes zoster, 357 peptic. See Peptic ulcer bleeding; Peptic ulcer disease. rectal bleeding, 314-315, 315f solitary, 267-268, 2050-2052, 2051f small intestinal. See Small intestinal ulcers. Ulcerative colitis, 1975-2014 abdominal pain in, 1983 acute pancreatitis in, 969 age and, 1976 anemia in, 2012 appendectomy and, 1979, 2069 versus appendicitis, 2063t arthropathy in, 2011, 2011t bone mineral density in, 2012 in celiac disease, 1813 chronic quiescent, 1982, 1982f classification of, 1991, 1991t clinical features of, 1982-1984 Clostridium difficile infection in, 1895 colectomy for, 1984-1985, 1984f, 2003-2004 colorectal cancer in, 2005-2008, 2006f-2007f, 2209-2210, 2209f-2210f cumulative risk of, 2209, 2209f dysplasia in, 2209-2210, 2210f screening for, 2227, 2229f colostomy for, 2024 complementary and alternative medicine for, 2292-2293 complications of, 2005-2010 versus Crohn’s disease, 1958-1959, 1959t, 1988-1989, 1989t-1990t crypt distortion in, 1982, 1982f cutaneous manifestations of, 359-360, 359f-360f, 2010-2011, 2010t diagnosis of, 1985-1988

Ulcerative colitis (Continued) abdominal radiography in, 1987-1988, 1987f barium enema in, 1987-1988, 1988f endoscopy in, 1985-1987, 1986f diarrhea in, 214-215, 1983 differential diagnosis of, 1988-1991, 1989t-1990t disease activity in, 1991-1992, 1991t-1992t versus dysentery, 1876 dysplasia in, 2005-2008, 2006f-2007f environmental factors in, 1978-1979 eosinophilia in, 432-433 epidemiology of, 1975-1976, 1976t epithelial cells in, 1980 erythema nodosum in, 2010 ethnicity and, 1976 etiology and pathogenesis of, 1977-1981 environmental factors in, 1978-1979 genetic factors in, 1977-1978 immune factors in, 1979-1980 psychogenic factors in, 1980-1981 extraintestinal manifestations of, 2010-2012, 2010t-2011t familial, 1977 fulminant, 1992 gastritis in, 855-856 genetic findings in, 1977-1978 geographic variation in, 1976 haustration loss in, 1987-1988, 1988f hematologic manifestations of, 2010t, 2012 hepatobiliary manifestations of, 2010t, 2012 HLA (human leukocyte antigen) alleles in, 1977-1978 hypercoagulability in, 2012 immunopathogenesis of, 1979-1980 incidence and prevalence of, 1975-1976, 1976t versus infectious colitis, 1989-1990, 1990t inflammatory pseudopolyps in, 1981, 1981f, 1985-1988, 1988f interleukin-12/23 in, 1977, 1980 intestinal microbiota and, 1978 irritable bowel syndrome and, 2092 laboratory findings in, 1984 MDR1 gene in, 1977 mortality in, 1985 musculoskeletal manifestations of, 2010t-2011t, 2011-2012 natural history and prognosis in, 1984-1985, 1984f-1985f ocular manifestations of, 2010t, 2011 osteonecrosis in, 2012 pathology of, 1981-1982, 1981f-1982f perforation in, 2005 pouchitis in, 2008-2010 in pregnancy, 629-630 primary sclerosing cholangitis in, 1156, 2012 in infants and children, 1057 probiotics, prebiotics, and synbiotics for, 96, 1999-2000 proctocolectomy for, 2004-2005 protein-losing gastroenteropathy in, 441 psychogenic factors in, 1980-1981 psychosocial factors in, 346 pyoderma gangrenosum in, 2010-2011 rectal bleeding in, 1983 in shigellosis, 1858, 1858f versus shigellosis, 1858f, 1860 signs of, 1983-1984 smoking and, 1978-1979 socioeconomic status and, 1976 stress and, 346 strictures in, 2005 T cells in, 1979-1980 toxic megacolon in, 1992, 2005 treatment of, 1992-2005 aminosalicylates for, 1993-1995, 1993f, 1994t antibiotics for, 1999 biologic therapy for, 2000-2001 cytapheresis for, 2001-2002 glucocorticoids for, 1995-1996, 1995t heparin for, 2000 immunomodulators for, 1996-1999, 1997t-1998t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Ulcerative colitis (Continued) medical, 1992-2002, 1992t in mild to moderately active disease, 2002f nicotine for, 2000 nuclear hormone receptor agonist for, 2002 nutritional therapy for, 2000 in severe disease, 2003f surgical, 2002-2005, 2003t Ulcerative enteritis, 2054-2058 background on, 2054 celiac disease and, 1817-1818, 1817f clinical presentation in, 2055 definition of, 2054 diagnosis of, 2055-2056, 2055f-2058f pathology of, 2056, 2058f in shigellosis, 1858, 1858f treatment of, 2056-2058, 2057f Ulcerative gingivitis, acute necrotizing, 357 Ulcerative jejunoileitis. See Ulcerative enteritis. Ultraflex stent, for esophageal carcinoma, 766 Ultrasonography in abdominal abscess, 414, 414f in acute abdominal pain, 156 in acute acalculous cholecystitis, 1142-1143, 1143f in acute calculous cholecystitis, 1115 in acute pancreatitis, 972 in adenomyomatosis, 1148-1149 in appendicitis, 2064, 2064f in biliary pain, 1113 of biliary tract, 1185 in choledochal cyst, 1059, 1059f in choledocholithiasis, 1117 in chronic pancreatitis, 1000-1001, 1000t, 1003, 1003t, 1004f in Crohn’s disease, 1957-1958 in cystic fibrosis, 946 in diverticulitis, 2080 Doppler in Budd-Chiari syndrome, 1373 in hepatocellular carcinoma, 1573 in peptic ulcer bleeding, 296 in portal vein thrombosis, 1378-1379, 1378f duplex, in intestinal angina, 2045 endoscopic in acute pancreatitis, 972, 1039 in adenomyomatosis, 1148-1149 biliary, 1197 in carcinoid tumors, 485 in cholangiocarcinoma, 1174 in chronic pancreatitis, 1003, 1003t, 1004f complications of, 661 in Crohn’s disease, 1957-1958 in esophageal cancer, 755, 757-760, 760f in fecal incontinence, 248, 248f fine needle aspiration in cholangiocarcinoma, 1162-1163 complications of, 661 in gallbladder polyp, 1150-1151 in gallstone disease, 1108t, 1109-1110, 1110f in gastric cancer, 901, 901f in gastrointestinal stromal tumors, 466, 2151f in intrahepatic cholangiocarcinoma, 1581 in jaundice, 333, 333t in malabsorption, 1749 in pancreatic cancer, 1022, 1022f in pancreatic cystic neoplasms, 1028, 1028f in pancreatic endocrine tumors, 516t, 517, 518f in portal hypertension, 1498 in small intestinal tumors, 2151, 2151f in solitary rectal ulcer syndrome, 2051 in gallbladder carcinoma, 1179-1180, 1179f in gallbladder polyp, 1150-1151, 1151f in gastric motility assessment, 798f, 804 in hepatic mass lesions, 1591-1592 in hepatocellular carcinoma, 1572-1573 in infantile hypertrophic pyloric stenosis, 784, 784f in intussusception, 2114-2115 in jaundice, 332, 333t laparoscopic, in choledocholithiasis, 1117 in malabsorption, 1749

Ultrasonography (Continued) in neonatal cholestasis, 1052 in nonalcoholic fatty liver disease, 1405, 1406f in pancreatic endocrine tumors, 515-517, 516t, 518f, 519 in portal hypertension, 1497 in schistosomiasis, 1938t, 1939 in small intestinal motility measurement, 1652 in sphincter of Oddi dysfunction, 1069 Umbilical hernia, 390 congenital, 1626-1628, 1627f physiologic, 1624-1625, 1625f Umbilical sepsis, portal vein thrombosis in, 1377 Umbilical varices, 1494 Umbilical vein, development of, 1202 Underfeeding, permissive, during parenteral nutrition, 86 Unfolded protein response, in alcoholic liver disease, 1389 Uniporter, 1680-1681 UNOS policies, for liver transplantation, 1595 Urea breath tests, for Helicobacter pylori infection, 840t, 841, 841f Urea cycle defects in, 1271-1274 clinical features of, 1272-1273 diagnosis of, 1273, 1273t laboratory values in, 1273t pathophysiology of, 1271-1272 treatment of, 1273-1274 steps of, 1271, 1272f Urease test, for Helicobacter pylori infection, 839, 840t Ureterosigmoidostomy, colorectal adenoma and, 2164 Urge incontinence, fecal, 246 Uric acid stones, in Crohn’s disease, 1955 Urinary bladder catheter, in ascites, 1537 Urine cholyl-PABA in, in small intestinal bacterial overgrowth, 1777 copper in, in Wilson disease, 1253t, 1254 culture of, in typhoid fever, 1866 phenols in, in protein malabsorption, 1752 sodium in, in ascites, 1536-1537 sodium:potassium ratio in, in ascites, 1537 trypsinogen activation peptide in, in acute pancreatitis, 975 Ursodeoxycholic acid for alcoholic liver disease, 1398 for antimitochondrial antibody–negative primary biliary cirrhosis, 1488 for autoimmune hepatitis, 1475 for cholestasis, 335 for cholestasis of pregnancy, 631-632 for colorectal cancer prophylaxis, 2008 in cystic fibrosis, 947 in gallstone disease, 1099 high-dose, for primary sclerosing cholangitis, 1164 for nonalcoholic fatty liver disease, 1410 for oral dissolution therapy, 1122 for primary biliary cirrhosis, 1483-1484, 1484f in primary biliary cirrhosis, 1486-1487, 1486t for primary sclerosing cholangitis, 1164, 1467 for pruritus, 1063 for sinusoidal obstruction syndrome, 1377 synthesis and metabolism of, 1077f, 1077t, 1078 therapeutic, 1087-1088 in ulcerative colitis, 2008 Urticaria, in parasitic diseases, 367 Urticaria pigmentosa, in mastocytosis, 362-363, 362f Usnic acid, hepatotoxicity of, 1458 Ustekinumab, for Crohn’s disease, 1970 Uterus, broad ligament of, hernia through, 395 Uveitis in Crohn’s disease, 1954-1955 in ulcerative colitis, 2011 in Whipple’s disease, 1837-1838

V

VacA gene, in Helicobacter pylori infection, 891-892 Vaccine cholera, 1852 Clostridium difficile, 1902 giardiasis, 1914 hepatitis A, 1283-1285, 1283t-1284t, 1309 hepatitis B, 1284, 1284t, 1307-1309, 1308t-1309t hepatitis C, 1328 hepatitis E, 1341 after liver transplantation, 1610-1611 typhoid fever, 1867 Vacuum-assisted closure device, in abdominal abscess drainage, 417, 421 Vagal afferent neurons in colonic motility, 1664 in small intestinal motility, 1647 in stomach, 801-802, 801f, 819-821, 820f Vagal nerves esophageal peristalsis control by, 681 in gastric emptying, 800-801 Vagotomy gastric motility disorders after, 807 for peptic ulcer bleeding, 882 in Zollinger-Ellison syndrome, 504 Valacyclovir for herpes zoster ulcers, 357 for orolabial herpes, 357 Valdecoxib, cardiovascular risk of, 875-876 Valerian, for colonic health, 2294t Valganciclovir, for cytomegalovirus colitis, in HIV/AIDS, 528 Valproic acid, hepatotoxicity of, 1429 Valsartan, hepatotoxicity of, 1433 Vancomycin, for Clostridium difficile-associated diarrhea and colitis, 1898-1901, 1898t Variceal pressure, endoscopic, 1496, 1496t Varicella-zoster infection. See Herpes zoster infection. Varices anorectal, 1494, 1513, 1513f bleeding from in chronic pancreatitis, 1013 endoscopic band ligation for. See Band ligation, endoscopic. in portal vein thrombosis, 1379 esophageal. See Esophageal varices. gallbladder, 1499, 1499f gastric. See Gastric varices. gastroesophageal in portal hypertension, 1494 in primary biliary cirrhosis, 1482 peristomal, in primary sclerosing cholangitis, 1164 in portal hypertension. See Portal hypertension, varices in. rectal, 314 retroperitoneal, 1494 umbilical, 1494 Vas deferens, congenital absence of, in cystic fibrosis, 948 Vascular access devices, for parenteral nutrition, 87 Vascular anomalies, of esophagus, 670t, 672, 673f Vascular disease. See also Ischemia; Thrombosis. acute pancreatitis in, 966 cutaneous, 360-363 in drug-induced liver disease, 1444-1445, 1445t gastrointestinal, 593-610, 594t hepatic, 1371-1382 liver transplantation and, 1597, 1603 Vascular ectasia colonic. See Angioectasia, colonic. gastric antral. See Gastric antral vascular ectasia (watermelon stomach). gastrointestinal bleeding from, 314-317, 317f Vascular elastosis, in carcinoid tumors, 476 Vascular endothelial growth factor in angioectasia, 596-597 in gastroesophageal varices, 1494-1495 in hepatopulmonary syndrome, 1550

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxxvii

lxxxviii

Index Vascular endothelial growth factor (Continued) in hereditary hemorrhagic telangiectasia, 600 inhibitors of. See Bevacizumab; Semaxanib. Vascular endothelial growth factor A (VEGF-A), in metastasis, 42-43 Vasculitides, gastrointestinal manifestations of, 558t, 562-563, 562f Vasculitis eosinophilia in, 432 in Henoch-Schönlein purpura, 2047 hypersensitivity, 2047 immune complex, of small vessels, 360, 360f in intestinal ischemia, 2046-2047 in Kawasaki’s disease, 2047 in lupus erythematosus, 2047 in malignant atrophic papulosis, 2047 in polyarteritis nodosa, 2047 rheumatoid, 2047 splanchnic, 2046-2047 in systemic lupus erythematosus, 561 in Takayasu’s disease, 2047 in thromboangiitis obliterans, 2046 Vasoactive intestinal peptide (VIP), 8. See also VIPoma. actions of, 8 in cephalic phase of pancreatic enzyme secretion, 926 in gastrointestinal tract, 5 in intestinal ion transport, 1691 in lower esophageal sphincter relaxation, 684 in postoperative ileus, 2125 in VIPoma, 509-510 Vasopressin for colonic angioectasia, 599 for portal hypertension, 1502 Vasopressor therapy, for hepatorenal syndrome, 1548, 1548t Vegetables, gastric cancer and, 892 Vena cava, inferior, membranous obstruction of, 1371-1372, 1372f, 1374. See also BuddChiari syndrome. hepatocellular carcinoma in, 1577 Venlafaxine for functional dyspepsia, 193 for neuropathic pain, 2283t Veno-occlusive disease, hepatic. See Sinusoidal obstruction syndrome. Venography, in Budd-Chiari syndrome, 1374 Venous thromboembolism gastrointestinal manifestations of, 570 in glucagonoma, 507 Venous thrombosis, deep, in ulcerative colitis, 2012 Ventilation, mechanical in acute liver failure, 1565 gastrointestinal complications in, 581-584, 582f Ventricular hypertrophy, left, in cirrhotic cardiomyopathy, 1553 Verapamil hepatotoxicity of, 1433 intestinal dysmotility from, 2139 Vermiform appendix, 1616 Verner-Morrison syndrome. See VIPoma. Verrucous squamous cell carcinoma, esophageal, 767 Vertical banded gastroplasty, 115, 116t-117t. See also Bariatric surgery. Very-low-density lipoprotein (VLDL), 1221, 1704 ApoB-containing, 1223-1224 receptors for, 1224 Vesicles, in bile, 1094 Vesiculobullous disease, 357-358 Vestibular fistula, 1635 Vibrio alginolyticus infection, 1853, 1865t Vibrio cholerae infection. See also Cholera. O1 (classic and El Tor), 1850-1852 non-O1 (O139 Bengal), 1852-1853 Vibrio damsela infection, 1853 Vibrio fluvialis infection, 1853 Vibrio furnissii infection, 1853 Vibrio hollisae infection, 1853 Vibrio metschnikovii infection, 1853 Vibrio mimicus infection, 1853

Vibrio parahaemolyticus infection, 1852-1853 clinical features of, 1853 epidemiology of, 1853 in food poisoning, 1879t-1880t intestinal ion transport in, 1691 treatment of, 1853, 1865t Vibrio vulnificus infection, 1853, 1865t Villus (villi) atrophy of in celiac disease, 1799-1800, 1799f in malabsorption, 1746-1747, 1746f, 1748f in tropical sprue, 1825, 1826f in ulcerative enteritis, 2056, 2058f of duodenum, 779-780 of small intestine, 1621, 1621f in Whipple’s disease, 1838, 1838f Vinyl chloride angiosarcoma from, 1451-1452 hepatotoxicity of, 1451-1452 VIP. See Vasoactive intestinal peptide (VIP). VIPoma, 508-511 clinical features of, 509-510, 509t definition of, 508-509 diagnosis and differential diagnosis of, 510 diarrhea in, 509-510 electrolyte disturbances in, 509-510 pathophysiology and pathology of, 509 treatment of, 510-511 Viral infection acute, versus alcoholic liver disease, 1392 acute liver failure and, 1558t-1559t, 1559f, 1560 diarrhea and, 1869-1872, 1870t in food poisoning, 1879t gastritis from, 849, 849f hepatic. See Hepatitis. HIV/AIDS diarrhea from, 526t-527t, 528, 528f obesity and, 102 Viral oncogenes, 41 Virulence factors bacterial, 1844-1846, 1845f of enterohemorrhagic Escherichia coli, 1856 of Helicobacter pylori, 834-835 Visceral analgesics, for functional dyspepsia, 193 Visceral fat, 103, 103t disorders related to, 104-105 Visceral functions, conditioning of, 339-340 Visceral larva migrans, 1359-1360 Visceral leishmaniasis, hepatic manifestations of, 1356t-1357t, 1358-1359 Visceral myopathy chronic intestinal pseudo-obstruction in, 2133, 2142 childhood, 2134-2135, 2135t familial, 2133-2134, 2134f familial, 2140 manometry in, 2142 Visceral neuropathy chronic intestinal pseudo-obstruction in, 2131-2133, 2132t, 2142 familial, 2134, 2135f paraneoplastic, 2137-2138 sporadic nonfamilial, 2133 inflammatory, 2131-2132, 2143 manometry in, 2142 markers for and targets of, 2132t paraneoplastic chronic intestinal pseudo-obstruction in, 2137-2138 constipation in, 270 Visceral pain, 151-152, 152f-153f, 163, 612 afferent (ascending) transmission of, 151-152, 153f, 166-167, 166f, 343, 343f amplification of. See Visceral sensitization. descending modulation of, 167, 167f, 344, 344f narcotic-induced potentiation of, 169-170 psychological distress and, 344, 345f psychosocial distress and, 344-345 regulation of, brain-gut interactions in, 343-345, 343f-345f Visceral peritoneum, 611-612

Visceral sensitization, 167, 343-344 anterior cingulate cortex in, 344-345 in functional abdominal pain syndrome, 166-167 in functional dyspepsia, 188 in irritable bowel syndrome, 2095-2096, 2097f psychological distress and, 344, 345f psychosocial factors in, 189 Visceral stretch receptors, 152 Viscus, ruptured, ascites in, 1528, 1531 Vitamin(s), 81-82, 81t colorectal cancer and, 2195 deficiency of, after bariatric surgery, 118 dietary, 52, 53t-55t, 81-82, 81t digestion and absorption of, 1717-1722 adaptive changes in, 1728-1729, 1729f fat-soluble, 1720-1722, 1721f, 1721t water-soluble, 1717-1720, 1718t fat-soluble, 52, 82 deficiency of, in primary biliary cirrhosis, 1486 digestion and absorption of, 1720-1722, 1721f, 1721t malabsorption of, 74, 1739, 1745t congenital disorders of, 1759t-1762t supplementation of in celiac disease, 1814-1815 in cystic fibrosis, 942-943 for malabsorption, 74 water-soluble, 52, 81-82 Vitamin A (retinol), 53t-55t, 81t, 82, 1721f, 1722 absorption of, 1721-1722, 1721t for cholestasis in children, 1063 deficiency of in cystic fibrosis, 942-943 in primary biliary cirrhosis, 1486 in small intestinal bacterial overgrowth, 1774 hepatotoxicity of, 1455-1456 for malabsorption, 74 malabsorption of, 1739 reference nutrient intake for, 1721t in short bowel syndrome, 1786t structure of, 1721f Vitamin B1. See Thiamine (vitamin B1). Vitamin B2. See Riboflavin (vitamin B2). Vitamin B3. See Niacin (vitamin B3). Vitamin B5. See Pantothenic acid (vitamin B5). Vitamin B6. See Pyridoxine (vitamin B6). Vitamin B7. See Biotin (vitamin B7). Vitamin B9. See Folate/folic acid. Vitamin B12 (cobalamin), 53t-55t, 81t, 82 absorption of, 1718-1720, 1718t, 1719f intrinsic factor and, 829-830 aging and, 58 deficiency of after bariatric surgery, 118 in elderly persons, 1758 in fish tapeworm infection, 1931 after gastric resection, 1756-1757 serum test for, 1752 in small intestinal bacterial overgrowth, 1774 in tropical sprue, 1824, 1825f, 1826-1827 in Zollinger-Ellison syndrome, 504 dietary sources of, 1718-1719 malabsorption of, 59 in autoimmune thyroid disease, 1766 congenital, 1759t-1762t, 1763 pathogenesis of, 1739 Schilling test for, 1752 in short bowel syndrome, 1782 reference nutrient intake for, 1718t serum, measurement of, 1752 in short bowel syndrome, 1786t supplementation of, in short bowel syndrome, 1783-1784 Vitamin C (ascorbic acid), 53t-55t, 81, 81t absorption of, 1717-1718, 1718t deficiency of, scurvy in, 368f, 369 dietary sources of, 1717 in familial adenomatous polyposis, 2183-2184 for gastrointestinal cancer, 2298 reference nutrient intake for, 1718t in short bowel syndrome, 1786t

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

Index Vitamin D, 53t-55t, 81t, 82 absorption of, 1721t, 1722 aging and, 58 for cholestasis in children, 1062 colorectal cancer and, 2195-2197, 2199f, 2199t deficiency of after bariatric surgery, 118 in chronic pancreatitis with steatorrhea, 997 in cystic fibrosis, 942-943 in primary biliary cirrhosis, 1486 in small intestinal bacterial overgrowth, 1774 for malabsorption, 74 malabsorption of, 1739 in metabolic bone disease, 1766 reference nutrient intake for, 1721t in short bowel syndrome, 1786t structure of, 1721f supplemental in celiac disease, 1814 in Crohn’s disease, 1970 in primary biliary cirrhosis, 1485-1486 in ulcerative colitis, 2012 Vitamin D–resistant rickets, malabsorption in, 1759t-1762t Vitamin E, 53t-55t, 81t, 82 absorption of, 1721t, 1722 for alcoholic liver disease, 1398 for cholestasis in children, 1063 deficiency of in cystic fibrosis, 942-943 after gastric resection, 1756 in primary biliary cirrhosis, 1486 in small intestinal bacterial overgrowth, 1774 in familial adenomatous polyposis, 2183-2184 for gastrointestinal cancer, 2298 for malabsorption, 74 malabsorption of, 1739 for nonalcoholic fatty liver disease, 1409 reference nutrient intake for, 1721t in short bowel syndrome, 1786t structure of, 1721f Vitamin K, 53t-55t, 81t, 82 absorption of, 1721t, 1722 for cholestasis in children, 1063 cutaneous reactions to, 366 deficiency of in cystic fibrosis, 942-943 in primary biliary cirrhosis, 1486 in short bowel syndrome, 1785, 1786t for malabsorption, 74 malabsorption of, 1739 reference nutrient intake for, 1721t structure of, 1721f Vitelline cyst, 1629f, 1630 Vitelline duct anomalies of, 1628-1630, 1629f blood vessel remnants in, 1630 patent, 1630 Vitelline vein, development of, 1202 Volume effect, in bariatric surgery, 119 Volvulus colonic, 2116-2117, 2118f-2119f, 2120 gastric. See Gastric volvulus. malrotation with, 1630, 1631f midgut, 2116, 2116f Vomiting, 197-210. See also Nausea. acute, 198-200, 204, 205f in acute pancreatitis, 969 in appendicitis, 2061 aspiration pneumonia with, 206 in Bacillus cereus infection, 1882 bilious, 198 chronic, 200-201, 204-206 complications of, 127f-128f differential diagnosis of, 809t, 810 clinical characteristics of, 198 coffee-ground, bleeding site in, 288 complementary and alternative medicine for, 2288-2289, 2289t complications of, 206 in Crohn’s disease, 1953 cyclic syndrome of, 202, 810 definition of, 197

Vomiting (Continued) diagnostic evaluation of, 204-206, 205f in duodenal atresia/stenosis, 785 esophageal injuries with, 206 etiology of, 198-204, 199t extraintestinal causes of, 200 fluid and electrolyte disturbances with, 206 functional, 201-202 gastric injuries with, 206 in gastric motility disorders, 201 in gastric outlet obstruction, 199-201 glottic spasm with, 206 after hematopoietic stem cell transplantation, 545-546, 546f, 555 infectious causes of, 200 in intestinal infarction, 200 in intestinal obstruction, 198-201 in metabolic disorders, 200 in neurologic disorders, 200-201 nutritional deficiencies with, 206 palliative care for, 2282, 2284t pathophysiology of, 197-198, 198f postoperative, 200 risk factors for, 2126-2127, 2126t in pregnancy, 201, 628 projectile, 198 radiation-induced, 642 relapsing, 200-201, 204-206 versus rumination, 203-204 after solid organ transplantation, 538t, 541 in superior mesenteric artery syndrome, 203, 203f toxin- or drug-induced, 199t, 200 treatment of dietary therapy for, 813t-814t, 814 fluid and electrolyte therapy in, 207 gastric stimulation in, 209 pharmacologic, 207-209 prokinetic agents in, 208-209 Vomitus content of, 198 odor of, 198 Von Hippel-Lindau disease, pancreatic endocrine tumors in, 495 Von Meyenburg complexes, 1590 Von Recklinghausen’s disease. See Neurofibromatosis-1. Von Willebrand disease, angioectasia in, 316 VSL #3 probiotic, for inflammatory bowel disease, 2293 Vulvovaginitis, in pinworm infection, 1928

W

Waardenburg syndrome, intestinal pseudoobstruction in, 2134 Waist circumference. See Central adiposity. Waldenström’s macroglobulinemia, gastrointestinal manifestations of, 569 Waldeyer’s ring lymphoma, 459 Wall motion studies, in small intestine evaluation, 1652 Walled off pancreatic necrosis, 960-961, 979, 979f, 981-982, 1037-1038, 1037f-1038f Wallstent II, for esophageal carcinoma, 766 Warfarin enteral nutrition and, 95 for mesenteric venous thrombosis, 2037-2038 Wartime history, visceral hypersensitivity and, 345 Warts (condylomata acuminata), 2270-2271, 2271t Water. See also Fluid(s). absorption of colonic, 263 intestinal, 212, 212f in short bowel syndrome, 1781-1782, 1781t, 1782f daily requirements for, 85 during enteral nutrition, 93 fecal contamination of, hepatitis E and, 1338-1339

Water (Continued) intestinal transport of, 1679. See also Fluid and electrolyte transport. Water brash, 179, 714 Water-load test, in gastric motility assessment, 805 Waterlow classification of protein-energy malnutrition, 61, 62t Watermelon stomach. See Gastric antral vascular ectasia (watermelon stomach). Weerda diverticuloscope, 372-373, 373f Wegener’s granulomatosis, gastrointestinal manifestations of, 562 Weight (body) excessive concern with in bulimia nervosa, 125 in eating disorders, 126t expected, in anorexia nervosa, 127 gain of. See also Obesity. drug-induced, 101-102, 101t after liver transplantation, 1610 ideal and energy requirements for hospitalized patients, 49, 50t in nutritional assessment, 65 loss of in anorexia-cachexia syndrome, 2282-2284 for ascites, 1537-1538 after bariatric surgery, 1732-1733 benefits of, 106 in celiac disease, 1805-1806 in chronic pancreatitis with steatorrhea, 997 in Crohn’s disease, 1953 in functional dyspepsia, 185 in glucagonoma, 507 after hematopoietic stem cell transplantation, 555 in malabsorption, 1745t in nutritional assessment, 64 plateau effect in, 99 rapid, gallstone disease after, 1091 in somatostatinoma, 512 in Whipple’s disease, 1836 management of, in eating disorders, 127, 134-135 Weight loss operations. See Bariatric surgery. Weight loss products, hepatotoxicity of, 1458 Weil’s syndrome, hepatic manifestations of, 1354 Wermer’s syndrome. See Multiple endocrine neoplasia-I. Wheat, 1801, 1801f adverse effects of. See Celiac disease. genetic modification of, for celiac disease, 1820 intolerance or allergy to, irritable bowel syndrome and, 2098 Whipple procedure for ampullary carcinoma, 1182-1183 for chronic pancreatitis, in pain control, 1008 for intraductal papillary mucinous tumor, 1031-1032, 1032t for pancreatic cancer, 1024, 1024f, 1025t Whipple’s disease, 1833-1842 cardiovascular system in, 1837 central nervous system in, 1836-1837 clinical features of, 1836-1838, 1837f diagnosis of, 1840 differential diagnosis of, 1840-1841 epidemiology of, 1833-1834 genomics of, 1834-1835 history of, 1833 immune defects in, 1835-1836 intestinal features of, 1836-1838, 1837f lymphatic system in, 1836-1838 lymphoma and, 1841 malabsorption in, 1746-1747, 1747f microbiology of, 1834-1835, 1834f musculoskeletal system in, 1837 neurologic disease in, 1836-1837 ocular manifestations of, 1837-1838 pathogenesis of, 1835-1836 pathology of extraintestinal, 1839-1840 small intestinal, 1838-1839, 1838f-1839f

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300

lxxxix

xc

Index Whipple’s disease (Continued) sarcoilosis and, 1840 skin hyperpigmentation in, 1837 treatment and prognosis in, 1841, 1841t Tropheryma whippelii in, 1834-1835, 1834f, 1841 Whipple’s triad, in insulinoma, 496 Whipworm infection, 1923f, 1927-1928, 1927f Crohn’s disease and, 1970 Whirl sign, in small intestinal obstruction, 2108-2109, 2110f White blood cells count of in acute cholecystitis, 1114 in appendicitis, 2063 in ascitic fluid, 1522f, 1523 in Clostridium difficile-associated diarrhea and colitis, 1897 in gastrointestinal bleeding, 288 in peritonitis, 614 fecal assay for, 220 in enteric infection, 1848-1849, 1848t Wilson disease, 1249-1258 acute liver failure in, 1560 alkaline phosphatase in, 1231 clinical features of, 1251-1252 copper pathway in, 1249-1251, 1250f diagnosis of, 1253-1255, 1253t family screening for, 1257 gallstone disease in, 1252 hepatic presentation in, 1251-1252 hepatocellular carcinoma in, 1577 hepatocellular mitochondria in, 1252-1253 jaundice in, 327 liver transplantation and, 1257, 1603 molecular defect in, 1249-1251, 1251f mutation analysis in, 1254-1255, 1255f neurologic presentation in, 1252 ocular signs of, 1252 pathology of, 1252-1253 in pregnancy, 637, 1257 presymptomatic diagnosis of siblings with, 1255, 1255f prognosis in, 1257 psychiatric presentation in, 1252 treatment of, 1255-1257, 1256t Wilson disease gene, 1249-1251, 1251f, 1254-1255, 1255f Wirsung, duct of, 911, 911f. See also Pancreatic duct(s).

Wnt pathway in cellular proliferation, 33, 34f in colorectal cancer, 2202, 2202f in intestinal development, 1618 in oncogenesis, 40-41 Wolman’s disease, malabsorption in, 1759t-1762t, 1763 Worms. See specific diseases and helminths. Wound healing, in protein-energy malnutrition, 63 Wound infections, peristomal, after percutaneous endoscopic gastrostomy, 90-91

X

X-linked adrenoleukodystrophy, 1275 X-linked cutis laxa, malabsorption in, 1759t-1762t X-linked infantile agammaglobulinemia, malabsorption in, 1764 Xanthelasmas, in primary biliary cirrhosis, 1486 Xanthine oxidase, oxygen radical production by, in intestinal ischemia, 2029 Xanthogranulomatous gastritis, 853 Xenobiotics, in primary biliary cirrhosis, 1479 Xerostomia, 353 Xylene, hepatotoxicity of, 1452 Xylose breath test, for small intestinal bacterial overgrowth, 1776t, 1777 Xylose test, for malabsorption, 1753

Y

Yellow dock, for colonic health, 2294t Yersinia enterocolitica infection, 1865t, 1868-1869 in food poisoning, 1879t-1880t Yersinia spp. infection, 1868-1869 hepatic manifestations of, 1352 Yonban virus infection, 1346-1347 Yu-cheng, 1454 Yusho oil disease, 1454

Z

Z-stent, for esophageal carcinoma, 766 Zafirlukast, hepatotoxicity of, 1433-1434 Zalcitabine, hepatotoxicity of, 1430

Zellweger spectrum disorders, 1275 Zellweger’s syndrome, 1275 Zenker’s diverticula, 371-373 clinical presentation and diagnosis of, 371-372, 372t etiology and pathogenesis of, 371 oropharyngeal dysphagia in, 686, 689, 689f treatment and prognosis in, 372-373, 373f, 700-701 Zidovudine, hepatotoxicity of, 1430 Zinc absorption of, 1725-1726 adaptive changes in, 1728, 1729f in short bowel syndrome, 1781t deficiency of, 58 cutaneous manifestations of, 368-369, 368f dietary, 56t-57t, 80t, 81, 1725 esophageal cancer and, 747 gastritis from, 1256-1257 for hepatic encephalopathy, 1546 for malabsorption, 74 malabsorption of, 1740 in pregnancy, 637 in short bowel syndrome, 1786t for Wilson disease, 1256, 1256t Zinc transporters, 1725 ZIP proteins, in zinc transport, 1725-1726 Zollinger-Ellison syndrome. See also Gastrinoma. clinical features of, 499t, 501-502, 502t definition of, 498-499 diagnosis of, 500f, 502-503 diarrhea in, 502 duodenotomy in, 505, 505f gastric acid hypersecretion in, 499, 503 pharmacologic treatment of, 503-504 surgical treatment of, 504-505 gastroesophageal reflux disease and, 501-502, 502t gastroesophageal reflux disease in, 715 in multiple endocrine neoplasia-I, 502, 505 natural history of, 501, 501f pathology and pathophysiology of, 499-501 peptic ulcer disease and, 501-502, 502t treatment of, 503-505 Zolpidem, hepatotoxicity of, 1436 Zona adherens, 1676f, 1679 Zona occludens, 1676f, 1679 Zonisamide, for obesity, 112 Zoster. See Herpes zoster infection. Zymogen granules, 913f, 914

Volume 1:  pages 1-1198  •  Volume 2:  pages 1199-2300